c i g o l o hematologyvolume 85supplementto no. 12december...

Men

sile

–Sp

ed.A

bb.P

ost.

–45

%ar

t.2,

com

ma

20B,

Legg

e66

2/96

-Fili

ale

diPa

via.

Ilm

itten

tech

iede

lare

stitu

zione

deif

asci

coli

non

cons

egna

tiim

pegn

ando

sia

paga

rele

tass

edo

vute

hjournal ofhematology

volume 85supplement

to no. 12december 2000

ISSN 0390-6078

haema

tologica


volume 85supplement to no. 12

december 2000

ISSN 0390-6078

editor-in-chiefedoardo ascari

published by theFerrata Storti Foundation

established in 1920

haema

tologica

Haem

atolog

ica–

Vol.

85, s

uppl

emen

t to

n. 1

2, D

ecem

ber

200

0

Editorial CommitteeTiziano Barbui (President of the Italian Society of Hematology, Bergamo); CirilRozman (Representative of the Spanish Association of Hematology and Hemotherapy, Barcelona); Pier Mannuccio Mannucci (Italian Society of Hemostasis and Thrombosis, Milan)

Associate EditorsCarlo Brugnara (Boston), Red Cells & Iron. Federico Caligaris Cappio (Torino), Lymphocytes & Immunology. Carmelo Carlo-Stella (Milano), Hematopoiesis &Growth Factors. Paolo G. Gobbi (Pavia), Lymphoid Neoplasia. Franco Locatelli(Pavia), Pediatric Hematology. Francesco Lo Coco (Roma), Translational Research.Alberto Mantovani (Milano), Leukocytes & Inflammation. Giuseppe Masera (Monza), Geographic Hematology. Cristina Mecucci (Perugia), Cytogenetics. Pier Giuseppe Pelicci (Milano, Italian Society of Experimental Hematology), Molecular Hematology. Paolo Rebulla (Milano), Transfusion Medicine. MiguelAngel Sanz (Valencia), Myeloid Neoplasia. Salvatore Siena (Milano), MedicalOncology. Jorge Sierra (Barcelona), Transplantation & Cell Therapy. VicenteVicente (Murcia), Hemostasis & Thrombosis

Editorial BoardAdriano Aguzzi (Zürich), Adrian Alegre Amor (Madrid), Claudio Anasetti (Seattle),Jeanne E. Anderson (San Antonio), Nancy C. Andrews (Boston), William Arcese(Roma), Andrea Bacigalupo (Genova), Carlo Balduini (Pavia), Luz Barbolla (Madrid),Giovanni Barosi (Pavia), Javier Batlle Fourodona (La Coruña), Yves Beguin (Liège),Marie Christine Béné (Nancy), Yves Beuzard (Paris), Andrea Biondi (Monza), MarioBoccadoro (Torino), Niels Borregaard (Copenhagen), David T. Bowen (Dundee),Ronald Brand (Leiden), Salut Brunet (Barcelona), Ercole Brusamolino (Pavia), ClaraCamaschella (Torino), Dario Campana (Memphis), Maria Domenica Cappellini(Milano), Angelo Michele Carella (Genova), Gian Carlo Castaman (Vicenza), MarcoCattaneo (Milano), Zhu Chen (Shanghai), Alan Cohen (Philadelphia), Eulogio CondeGarcia (Santander), Antonio Cuneo (Ferrara), Björn Dahlbäck (Malmö), RiccardoDalla Favera (New York), Armando D’Angelo (Milano), Elisabetta Dejana (Milano),Jean Delaunay (Le Kremlin-Bicêtre), Consuelo Del Cañizo (Salamanca), Valerio DeStefano (Roma), Joaquin Diaz Mediavilla (Madrid), Francesco Di Raimondo(Catania), Charles Esmon (Oklahoma City), Elihu H. Estey (Houston), Renato Fanin(Udine), José-María Fernández Rañada (Madrid), Evarist Feliu Frasnedo (Barcelona),Jordi Fontcuberta Boj (Barcelona), Francesco Frassoni (Genova), Renzo Galanello(Cagliari), Arnold Ganser (Hannover), Alessandro M. Gianni (Milano), NorbertC. Gorin (Paris), Alberto Grañena (Barcelona), Eva Hellström-Lindberg (Huddinge),Martino Introna (Milano), Rosangela Invernizzi (Pavia), Achille Iolascon (Bari),Sakari Knuutila (Helsinki), Myriam Labopin (Paris), Catherine Lacombe (Paris),Francesco Lauria (Siena), Mario Lazzarino (Pavia), Roberto M. Lemoli (Bologna),Giuseppe Leone (Roma), A. Patrick MacPhail (Johannesburg), Ignazio Majolino(Palermo), Patrice Mannoni (Marseille), Guglielmo Mariani (Palermo), EstellaMatutes (London), Alison May (Cardiff), Roberto Mazzara (Barcelona), GiampaoloMerlini (Pavia), José Maria Moraleda (Murcia), Enrica Morra (Milano), KazumaOhyashiki (Tokyo), Alberto Orfao (Salamanca), Anders Österborg (Stockholm), Pier Paolo Pandolfi (New York), Ricardo Pasquini (Curitiba), Andrea Pession(Bologna), Franco Piovella (Pavia), Giovanni Pizzolo (Verona), Domenico Prisco(Firenze), Susana Raimondi (Memphis), Alessandro Rambaldi (Bergamo), FernandoRamos Ortega (León), José Maria Ribera (Barcelona), Damiano Rondelli (Bologna),Giovanni Rosti (Ravenna), Bruno Rotoli (Napoli), Domenico Russo (Udine), StefanoSacchi (Modena), Giuseppe Saglio (Torino), Jesus F. San Miguel (Salamanca),Guillermo F. Sanz (Valencia), Hubert Schrezenmeier (Berlin), Mario Sessarego(Genova), Pieter Sonneveld (Rotterdam), Yoshiaki Sonoda (Kyoto), Yoichi Takaue(Tokyo), José Francisco Tomás (Madrid), Giuseppe Torelli (Modena), Antonio Torres(Cordoba), Pinuccia Valagussa (Milano), Andrea Velardi (Perugia), Ana Villegas(Madrid), Françoise Wendling (Villejuif), Pier Luigi Zinzani (Bologna)

Publication Policy CommitteeEdoardo Storti (Chair, Pavia), John W. Adamson (Milwaukee), Carlo Bernasconi(Pavia), Gianni Bonadonna (Milano), Gianluigi Castoldi (Ferrara), Albert de laChapelle (Columbus), Peter L. Greenberg (Stanford), Fausto Grignani (Perugia),Lucio Luzzatto (New York), Franco Mandelli (Roma), Massimo F. Martelli (Perugia),Emilio Montserrat (Barcelona), David G. Nathan (Boston), Alessandro Pileri (Torino), Vittorio Rizzoli (Parma), Eduardo Rocha (Pamplona), Pierluigi Rossi Ferrini(Firenze), George Stamatoyannopoulos (Seattle), Sante Tura (Bologna), Hermanvan den Berghe (Leuven), Zhen-Yi Wang (Shanghai), David Weatherall (Oxford),Soledad Woessner (Barcelona)

Editorial OfficeIgor Ebuli Poletti, Paolo Marchetto, Michele Moscato, Lorella Ripari, Rachel Stenner


haema

tologica

Official Organ of the Italian Society of Hematology, the Italian Society of Experimental Hematology,the Spanish Association of Hematology and Hemotherapy, the Italian Society of Hemostasis and Thrombosis,andthe Italian Society of Pediatric Hematology/Oncology

Editor-in-ChiefEdoardo Ascari (Pavia)

Executive EditorMario Cazzola (Pavia)

Editorial policyHaematologica – Journal of Hematology (ISSN 0390-6078) is owned bythe Ferrata Storti Foundation, a non-profit organization created throughthe efforts of the heirs of Professor Adolfo Ferrata and of ProfessorEdoardo Storti. The aim of the Ferrata Storti Foundation is to stimulateand promote the study of and research on blood disorders and theirtreatment in several ways, in particular by supporting and expandingHaematologica.The journal is published monthly in one volume per year and has both apaper version and an online version (Haematologica on Internet, website: http://www.haematologica.it). There are two editions of the printjournal: 1) the international edition (fully in English) is published by theFerrata Storti Foundation, Pavia, Italy; 2) the Spanish edition (the inter-national edition plus selected abstracts in Spanish) is published by Edi-ciones Doyma, Barcelona, Spain. The contents of Haematologica are protected by copyright. Papers areaccepted for publication with the understanding that their contents, allor in part, have not been published elsewhere, except in abstract form orby express consent of the Editor-in-Chief or the Executive Editor. Furtherdetails on transfer of copyright and permission to reproduce parts ofpublished papers are given in Instructions to Authors. Haematologicaaccepts no responsibility for statements made by contributors or claimsmade by advertisers.Editorial correspondence should be addressed to: Haematologica JournalOffice, Strada Nuova 134, 27100 Pavia, Italy (Phone: +39-0382-531182– Fax: +39-0382-27721 – E-mail: [email protected]).

Year 2001 subscription information

International editionAll subscriptions are entered on a calendar-year basis, beginning in Jan-uary and expiring the following December. Send subscription inquiriesto: Haematologica Journal Office, Strada Nuova 134, 27100 Pavia, Italy(Phone: +39-0382-531182 Fax: +39-0382-27721 - E-mail:[email protected]). Payment accepted: major credit cards (Ameri-can Express, VISA and MasterCard), bank transfers and cheques. Subscription rates, including postage and handling, are reported below.Individual subscriptions are intended for personal use. Subscribers to theprint edition are entitled to free access to Haematologica on Internet,the full-text online version of the journal. Librarians interested in gettinga site licence that allows concurrent user access should contact theHaematologica Journal Office.

Rates Institutional Personal

Print editionEurope Euro 300 Euro 150Rest of World (surface) Euro 300 Euro 150Rest of World (airmail) Euro 350 Euro 200Countries with limited resources Euro 35 Euro 25

Haematologica on InternetWorldwide Free Free

Spanish print editionThe Spanish print edition circulates in Spain, Portugal, South and Cen-tral America. To subscribe to it, please contact: Ediciones Doyma S.A.,Travesera de Gracia, 17-21, 08021 Barcelona, Spain (Phone: +34-93-414-5706 – Fax +34-93-414-4911 – E-mail: [email protected]). Sub-scribers to the Spanish print edition are also entitled to free access tothe online version of the journal.

Change of addressCommunications concerning changes of address should be addressed tothe Publisher. They should include both old and new addresses andshould be accompanied by a mailing label from a recent issue. Allow sixweeks for all changes to become effective.

Back issuesInquiries about single or replacement copies of the journal should beaddressed to the Publisher.

AdvertisementsContact the Advertising Manager, Haematologica Journal Office, StradaNuova 134, 27100 Pavia, Italy (Phone: +39-0382-531182 – Fax: +39-0382-27721 – E-mail: [email protected]).

16 mm microfilm, 35 mm microfilm, 105 mm microfiche and articlecopies are available through University Microfilms International, 300North Zeeb Road, Ann Arbor, Michigan 48106, USA.


haema

tologica

Associated with USPI, Unione Stampa PeriodicaItaliana. Premiato per l’alto valore culturale dalMinistero dei Beni Culturali ed Ambientali

Haematologica publishes monthly Editorials, Original Papers, Reviews and Scientific Cor-respondence on subjects regarding experimental, laboratory and clinical hematology.Editorials and Reviews are normally solicited by the Editor, but suitable papers of this typemay be submitted for consideration. Appropriate papers are published under the head-ings Decision Making and Problem Solving and Molecular Basis of Disease.Review and Action. Submission of a paper implies that neither the article nor any essen-tial part of it has been or will be published or submitted for publication elsewhere beforeappearing in Haematologica. Each paper submitted for publication is first assigned by theEditor to an appropriate Associate Editor who has knowledge of the field discussed in themanuscript. The first step of manuscript selection takes place entirely inhouse and hastwo major objectives: a) to establish the article’s appropriateness for Haematologica’sreadership; b) to define the manuscript’s priority ranking relative to other manuscriptsunder consideration, since the number of papers that the journal receives is greater thanwhich it can publish. Manuscripts that are considered to be either unsuitable for the jour-nal’s readership or low-priority in comparison with other papers under evaluation will notundergo external in-depth review. Authors of these papers are notified promptly; with-in about 2 weeks, that their manuscript cannot be accepted for publication. The remain-ing articles are reviewed by at least two different external referees (second step or clas-sical peer-review). After this peer evaluation, the final decision on a paper’s acceptabil-ity for publication is made in conjunction by the Associate Editor and one of the Editors,and this decision is then transmitted to the authors.Conflict of Interest Policies. Before final acceptance, authors of research papers orreviews will be asked to sign the following conflict of interest statement: Please provideany pertinent information about the authors’ personal or professional situation that mightaffect or appear to affect your views on the subject. In particular, disclose any financial sup-port by companies interested in products or processes involved in the work described. A notein the printed paper will indicate that the authors have disclosed a potential conflict ofinterest. Reviewers are regularly asked to sign the following conflict of interest statement:Please indicate whether you have any relationship (personal or professional situation, in par-ticular any financial interest) that might affect or appear to affect your judgment. Researcharticles or reviews written by Editorial Board Members are regularly processed by the Edi-tor-in-Chief and/or the Executive Editor.Time to publication. Haematologica strives to be a forum for rapid exchange of newobservations and ideas in hematology. As such, our objective is to review a paper in 4weeks and communicate the editorial decision by fax within one month of submission.However, it must be noted that Haematologica strongly encourages authors to send theirpapers via Internet. Haematologica believes that this is a more reliable way of speedingup publication. Papers sent using our Internet Submission page will be processed in 2-3weeks and then, if accepted, published immediately on our web site. Papers sent via reg-ular mail or otherwise are expected to require more time to be processed. Detailed instruc-tions for electronic submission are available athttp://www.haematologica.it.

Submit papers to:http://www.haematologica.it/submission

orthe Editorial Office, Haematologica, Strada Nuova 134, 27100 Pavia, Italy

Manuscript preparation. Manuscripts must be written in English. Manuscripts withinconsistent spelling will be unified by the English Editor. Manuscripts should be preparedaccording to the Uniform Requirements for Manuscripts Submitted to Biomedical Journals,N Engl J Med 1997; 336:309-15; the most recent version of the Uniform Requirements canbe found on the following web site:http://www.ama-assn.org/public/peer/wame/uniform.htm

With respect to traditional mail submission, manuscripts, including tables and figures,should be sent in triplicate to facilitate rapid reference. In order to accelerate processing,author(s) should also enclose a 3.5” diskette (MS-DOS or Macintosh) containing the man-uscript text; if the paper includes computerized graphs, the diskette should contain thesedocuments as well. Computer programs employed to prepare the above documents shouldbe listed. Title Page. The first page of the manuscript must contain: (a) title, name and surname ofthe authors; (b) names of the institution(s) where the research was carried out; (c) a run-ning title of no more than 50 letters; (d) acknowledgments; (e) the name and full postaladdress of the author to whom correspondence regarding the manuscript as well asrequests for abstracts should be sent. To accelerate communication, phone, fax numberand e-mail address of the corresponding author should also be included.Abstract. The second page should carry an informative abstract of no more than 250 wordswhich should be intelligible without reference to the text. Original paper abstracts mustbe structured as follows: background and objectives, design and methods, results, inter-pretation and conclusions. After the abstract, add three to five key words.Editorials should be concise. No particular format is required for these articles, whichshould not include a summary.Original Papers should normally be divided into abstract, introduction, design and meth-ods, results, discussion and references.The section Decision Making and Problem Solving presents papers on health decision sci-ence specifically regarding hematologic problems. Suitable papers will include thosedealing with public health, computer science and cognitive science. This section may alsoinclude guidelines for diagnosis and treatment of hematologic disorders and positionpapers by scientific societies.Reviews provide a comprehensive overview of issues of current interest. No particular for-mat is required but the text should be preceded by an abstract which should be struc-tured as follows: background and objective, evidence and information sources, state ofart, perspectives. Within review articles, Haematologica gives top priority to: a) paperson molecular hematology to be published in the section Molecular basis of disease; b)papers on clinical problems analyzed according to the methodology typical of Evidence-Based Medicine.Scientific Correspondence should be no longer than 500 words (a word count should beincluded by the authors), can include one or two figures or tables, and should not con-tain more than ten strictly relevant references. Letters should have a short abstract (≤ 50words) as an introductory paragraph, and should be signed by no more than six authors.Correspondence, i.e. comments on articles published in the Journal will only appear in

hhaematologica Instructions to Authors

For additional information, the scientificstaff of Haematologica can be reachedthrough:mailing address: Haematologica, StradaNuova 134, I-27100 Pavia, Italy. Tel. +39.0382.531182. Fax +39.0382.27721.e-mail: [email protected]: http://www.haematologica.it

our Internet edition. Pictures of particular interest will be published in appropriate spaceswithin the journal.Tables and Illustrations. Tables and illustrations must be constructed in consideration ofthe size of the Journal and without repetitions. They should be sent in triplicate with eachtable typed on a separate page, progressively numbered with Arabic numerals and accom-panied by a caption in English. All illustrations (graphs, drawings, schemes and pho-tographs) must be progressively numbered with Arabic numerals. In place of original draw-ings, roentgenograms, or other materials, send sharp glossy black-and-white photographicprints, ideally 13 by 18 cm but no larger than 20 by 25 cm. In preparing illustrations, thefinal base should be considered the width of a single column, i.e. 8 cm (larger illustrationswill be accepted only in special cases). Letters and numbers should be large enough toremain legible (> 1 mm) after the figure has been reduced to fit the width of a single col-umn. In preparing composite illustrations, each section should be marked with a small let-ter in the bottom left corner. Legends for illustrations should be typewritten on a sepa-rate page. Authors are also encouraged to submit illustrations as electronic files togeth-er with the manuscript text (please, provide what kind of computer and softwareemployed).Units of measurement. All hematologic and clinical chemistry measurements should bereported in the metric system according to the International System of Units (SI) (AnnIntern Med 1987; 106:114-29). Alternative non-SI units may be given in addition. Authorsare required to use the standardized format for abbreviations and units of the Interna-tional Committee for Standardization in Hematology when expressing blood count results(Haematologica 1991; 76:166).References should be prepared according to the Vancouver style (for details see:http://www.ama-assn.org/ public/peer/wame/uniform.htm or also N Engl J Med 1997;336:309-15). References must be numbered consecutively in the order in which they arefirst cited in the text, and they must be identified in the text by Arabic numerals (in paren-theses). Journal abbreviations are those of the List of the Journals Indexed, printed annu-ally in the January issue of the Index Medicus [this list (about 1.3 Mb) can also be obtainedon Internet through the US National Library of Medicine website, at the following world-wide-web address: http://www.nlm.nih.gov/tsd/serials/lji.html). List all authors when six or fewer; when seven or more, list only the first three and add etal. Examples of correct forms of references follow (please note that the last page must beindicated with the minimum number of digits):

Journals (standard journal article,1,2 corporate author,3 no author given,4 journal supple-ment5):

1. Najfeld V, Zucker-Franklin D, Adamson J, Singer J, Troy K, Fialkow PJ. Evidence for clon-al development and stem cell origin of M7 megakaryocytic leukemia. Leukemia 1988;2:351-7.

2. Burgess AW, Begley CG, Johnson GR, et al. Purification and properties of bacteriallysynthesized human granulocyte-macrophage colony stimulating factor. Blood 1987;69:43-51.

3. The Royal Marsden Hospital Bone-Marrow Transplantation Team. Failure of syngene-ic bone-marrow graft without preconditioning in post-hepatitis marrow aplasia. Lancet1977; 2:242-4.

4. Anonymous. Red cell aplasia [editorial]. Lancet 1982; 1:546-7.

5. Karlsson S, Humphries RK, Gluzman Y, Nienhuis AW. Transfer of genes into hemopoi-etic cells using recombinant DNA viruses [abstract]. Blood 1984; 64(Suppl 1):58a.

Books and other monographs (personal authors,6,7 chapter in a book,8 published proceed-ing paper,9 abstract book,10 monograph in a series,11 agency publication12):

6. Ferrata A, Storti E. Le malattie del sangue. 2nd ed. Milano: Vallardi; 1958.

7. Hillman RS, Finch CA. Red cell manual. 5th ed. Philadelphia: FA Davis; 1985.

8. Bottomley SS. Sideroblastic anaemia. In: Jacobs A, Worwood M, eds. Iron in biochem-istry and medicine, II. London: Academic Press; 1980. p. 363-92.

9. DuPont B. Bone marrow transplantation in severe combined immunodeficiency with anunrelated MLC compatible donor. In: White HJ, Smith R, eds. Proceedings of the thirdannual meeting of the International Society for Experimental Hematology. Houston:International Society for Experimental Hematology; 1974. p. 44-6.

10.Bieber MM, Kaplan HS. T-cell inhibitor in the sera of untreated patients with Hodgk-in’s disease [abstract]. Paper presented at the International Conference on MalignantLymphoma Current Status and Prospects, Lugano, 1981:15.

11.Worwood M. Serum ferritin. In: Cook JD, ed. Iron. New York: Churchill Livingstone; 1980.p. 59-89. (Chanarin I, Beutler E, Brown EB, Jacobs A, eds. Methods in hematology; vol1).

12.Ranofsky AL. Surgical operation in short-stay hospitals: United States-1975. Hyattsville,Maryland: National Center for Health Statistics; 1978. DHEW publication no. (PHS)78-1785, (Vital and health statistics; series 13; no. 34).

References to Personal Communications and Unpublished Data should be incorporated inthe text and not placed under the numbered References. Please type the references exact-ly as indicated above and avoid useless punctuation (e.g. periods after the initials ofauthors’ names or journal abbreviations).

Galley Proofs and Reprints. Galley proofs should be corrected and returned by fax orexpress delivery within 72 hours. Minor corrections or reasonable additions are permitted;however, excessive alterations will be charged to the authors. Papers accepted for publi-cation will be printed without cost. The cost of printing color figures will be communicat-ed upon request. Reprints may be ordered at cost by returning the appropriate form sentby the publisher.Transfer of Copyright and Permission to Reproduce Parts of Published Papers. Authorswill grant copyright of their articles to the Ferrata Storti Foundation. No formal permis-sion will be required to reproduce parts (tables or illustrations) of published papers, pro-vided the source is quoted appropriately and reproduction has no commercial intent. Repro-ductions with commercial intent will require written permission and payment of royalties.

hhaematologica Instructions to Authors

For additional information, the scientificstaff of Haematologica can be reachedthrough:mailing address: Haematologica, StradaNuova 134, I-27100 Pavia, Italy. Tel. +39.0382.531182. Fax +39.0382.27721.e-mail: [email protected]: http://www.haematologica.it

hhaematologica Table of Contents

2000; vol. 85; supplement to no. 12

(indexed by CurrentContents/Life Sciences and in FaxonFinder and Faxon XPRESS, also avail-able on diskette with abstracts)

Papers

CD34-positive cells: biology and clinical relevanceCarmelo Carlo Stella, Mario Cazzola, Paolo De Fabritiis,Armando De Vincentiis, Alessandro Massimo Gianni,Francesco Lanza, Francesco Lauria, Roberto M. Lemoli, Corrado Tarella, Paola Zanon, Sante Tura.........................................1

Peripheral blood stem cells in acute myeloid leukemia:biology and clinical applicationsMassimo Aglietta, Armando De Vincentiis, Luigi Lanata,Francesco Lanza, Roberto Massimo Lemoli, GiacomoMenichella, Agostino Tafuri, Paola Zanon, Sante Tura ....19

Peripheral blood stem cell transplantation forthe treatment of multiple myeloma: biologicaland clinical implicationsFederico Caligaris Cappio, Michele Cavo,Armando De Vincentiis, Luigi Lanata, Roberto MassimoLemoli, Ignazio Majolino, Corrado Tarella, Paola Zanon,Sante Tura ......................................................................................................................32

Allogeneic hematopoietic stem cells from sources otherthan bone marrow: biological and technical aspectsFrancesco Bertolini, Armando De Vincentiis, Luigi Lanata,Roberto Massimo Lemoli, Rita Maccario, Ignazio Majolino,Luisa Ponchio, Damiano Rondelli, Antonio Tabilio, Paola Zanon, Sante Tura..................................................................................49

Clinical use of allogeneic hematopoietic stem cells fromsources other than bone marrowWilliam Arcese, Franco Aversa, Giuseppe Bandini, Armando De Vincentiis, Michele Falda, Luigi Lanata,Roberto Massimo Lemoli, Franco Locatelli, Ignazio Majolino, Paola Zanon, Sante Tura................................69

Ex vivo expansion of hematopoietic cellsand their clinical useMassimo Aglietta, Francesco Bertolini, CarmeloCarlo-Stella, Armando De Vincentiis, Luigi Lanata, RobertoMassimo Lemoli, Attilio Olivieri, Salvatore Siena, PaolaZanon, Sante Tura ..................................................................................................92

Cell therapy: achievements and perspectives Claudio Bordignon, Carmelo Carlo-Stella, Mario PaoloColombo, Armando De Vincentiis, Luigi Lanata, RobertoMassimo Lemoli, Franco Locatelli, Attilio Olivieri,Damiano Rondelli, Paola Zanon, Sante Tura ...........................117

Antitumor vaccination: where we stand Monica Bocchia, Vincenzo Bronte, Mario Paolo Colombo,Armando De Vincentiis, Massimo Di Nicola, Guido Forni,Luigi Lanata, Roberto Massimo Lemoli, Massimo Massaia,Damiano Rondelli, Paola Zanon, Sante Tura ...........................156

haematologica vol. 85(suppl. to n. 12):December 2000

CD33– HLA-DR–, seem to be more homogeneous in size(small lymphocyte-like cells) and lack cytoplasmic gran-ules and prominent nucleoli. Again, the morphology ofthis cellular population appears to reflect the functionalcharacteristics of these cells (e.g. low protein synthe-sis, very low proliferative activity with predominantlyG0 phase).

Several monoclonal antibodies (MY10, 12.8, B1-3C5,115.2, ICH3, TUK3, etc.) raised against the leukemic celllines KG1 or KG1a and an anti-endothelial cell anti-body (QBEND10) assigned to the CD34 cluster havebeen shown to identify a transmembrane glycoproteicantigen of 105-120 kD expressed on 1-3% of normalbone marrow cells, 0.01-0.1% peripheral blood cellsand 0.1-0.4% cord blood cells.5 Different antibodiesrecognize distinct epitopes of the same antigen. CD34antigen expression is associated with concomitantexpression of several other markers that can be classi-fied as lineage non-specific markers (Thy1, CD38, HLA-DR, CD45RA, CD71) and lineage specific markers,including T-lymphoid (TdT, CD10, CD7, CD5, CD2), B-lymphoid (TdT, CD10, CD19), myeloid (CD33, CD13) andmegakaryocytic (CD61, CD41, CD42b) markers.5 Theexpression of lineage non-specific markers allows theheterogeneous CD34+ population to be divided into twodistinct subpopulations characterized, respectively, bylow or high expression of Thy1, CD38, HLA-DR, CD45RA,CD71. These two cell subpopulations contain early andlate hematopoietic progenitor cells, respectively.6-8

In addition to conventional immunological markersclassified on the basis of their assignation to specificclusters of differentiation, CD34 cells express receptorsfor a number of growth factors. Two distinct families of

The CD34-positive cell: definition andmorphology

Cellular expression of the CD34 antigen identifies amorphologically and immunologically heterogeneouscell population that is functionally characterized by thein vitro capability to generate clonal aggregates derivedfrom early and late progenitors and the in vivo capac-ity to reconstitute the myelo-lymphopoietic system ina supralethally irradiated host.1-3

Immunohistochemical studies have demonstratedthat the CD34 antigen is stage but not lineage specif-ic. In fact, independently of the differentiative lineage,it is expressed only by ontogenetically early cells.4 Foryears a major obstacle to the morphological identifica-tion of putative hematopoietic stem cell has been thedifficulty in separating them from their direct progeny.The use of CD34 and other suitable cell surface mark-ers (i.e. CD33, CD38, HLA-DR antigens) in fluorescence-activated cell-sorting techniques or other cell separa-tion methods has allowed considerable progress in thisfield.

Positively selected, lineage committed CD34+ cellsand more immature, lineage negative CD34+ CD33–

HLA-DR– cells are shown in Figure 1 and Figure 2,respectively. On May-Grünwald-Giemsa stained prepa-rations, CD34+ cells are medium sized cells having largenuclei, eccentrically surrounded by narrow rims of deepblue cytoplasm occasionally containing cytoplasmicgranules. Some normal CD34+ cell nuclei show one ormore pale blue nucleoli. Taken together, these findingsreflect the heterogeneous proliferative status and pro-tein synthesis of this cell population. Conversely, earli-er hematopoietic progenitors, identified as CD34+

review

CD34-positive cells: biology and clinical relevance

haematologica 2000; 85(supplement to no. 12):1-18

CARMELO CARLO STELLA, MARIO CAZZOLA,*PAOLO DE FABRITIIS,° ARMANDO DE VINCENTIIS,#

ALESSANDRO MASSIMO GIANNI,@ FRANCESCO LANZA,^FRANCESCO LAURIA,** ROBERTO M. LEMOLI,°°CORRADO TARELLA,## PAOLA ZANON,@@ SANTE TURA°°Cattedra di Ematologia, Università di Parma, Parma; *Diparti-mento di Medicina Interna e Terapia Medica, Sezione di Med-icina Interna e Oncologia Medica, Università di Pavia e IRCCSPoliclinico S. Matteo, Pavia; °Dipartimento di BiopatologiaUmana, Sezione di Ematologia, Università “La Sapienza”,Roma; #Dompé Biotec SpA, Milano; @Centro Trapianto MidolloOsseo “Cristina Gandini”, Divisione di Oncologia Medica C,Istituto Nazionale Tumori, Milano; ^Istituto di Ematologia eFisiopatologia dell’Emostasi, Università degli Studi di Ferrara,Ferrara; **Istituto di Scienze Mediche, Università degli Studidi Milano, Milano; °°Istituto di Ematologia “Lorenzo e AriostoSeragnoli”, Università degli Studi di Bologna, Bologna; ##Cat-tedra di Ematologia, Università di Torino, Torino; @@AmgenS.p.A., Milano, Italy

Correspondence: Prof. Sante Tura, Istituto di Ematologia L. e A. Serà-gnoli, Policlinico S. Orsola, via Massarenti 9, 40138 Bologna, Italy.Acknowledgments: the preparation of this manuscript was supportedby grants from Dompé Biotec SpA, Milano, and Amgen S.p.A., Milano,Italy. Received October 18, 1994; accepted May 2, 1995.

related receptors have been identified: (i) tyrosine kinasereceptors, including the stem cell factor receptor (SCF-R, CD117) and the macrophage colony-stimulating fac-tor receptor (M-CSF-R, CD115); (ii) hematopoieticreceptors not containing a tyrosine kinase domain, suchas the granulocyte-macrophage colony-stimulating fac-tor receptor (GM-CSF-R, CDw116).5,9

The identification of new markers selectivelyexpressed on primitive lymphohematopoietic cells(CD34+ CD38–) represents a stimulating research field. Inthis context, stem cell tyrosine kinase receptors (STK),such as STK-1, a human homologue of the murine Flk-2/Flt-3, are of particular relevance.10-12 The ligands forthese receptors might represent new factors able toselectively control stem cell self-renewal, proliferationand differentiation.13-15

Clonogenic and biologic activityThe structural and functional integrity of the

hematopoietic system is maintained by a relatively smallpopulation of stem cells, located mainly in the bonemarrow, that can (i) undergo self-renewal to producemore stem cells or (ii) differentiate to produce progenywhich are progressively unable to self-renew, irreversiblycommitted to one or another of the various hematopoi-etic lineages, and able to generate clones of up to 105

lineage-restricted cells that mature into specializedcells.16-18

The decision of a stem cell to either self-renew or dif-ferentiate and the selection of a specific differentiationlineage by a multipotent progenitor during commitmentare intrinsic properties of stem cell progenitor cells andare regulated by stochastic mechanisms.19 Survival andamplification of hematopoietic progenitors are con-trolled by a number of regulatory molecules (hematopoi-etic growth factors) interacting according to complexmodalities (synergism, recruitment, antagonism).19 A fur-ther level of hematopoietic control is exerted by nucleartranscription factors that activate lineage-specific genesregulating growth factor responsiveness and/or the pro-liferative capacity of hematopoietic cells.20

Detection of the most primitive hematopoietic celltypes is now possible due to the technique of long-termbone marrow culture. In the case of human bone mar-row, a 5- to 8-week time period between initiating cul-tures and assessing clonogenic progenitor numbersallows quantification of a very primitive cell in the start-ing population, the so-called long-term culture-initiat-ing cell (LTC-IC).21 Committed progenitors of the varioushematopoietic cell classes can be quantitated by a num-ber of short-term culture clonogenic assays.19 CD34 anti-gen expression associated with low CD38 and CD45RAexpression and variable HLA-DR expression is a typicalfeature of LTC-IC, CFU-Blast, CFU-T, CFU-B. In contrast,CD34 antigen expression associated with CD38 and HLA-DR expression is a typical feature of multipotent (CFU-GEMM) and lineage-restricted (CFU-GM, CFU-G, CFU-M,BFU-E, CFU-E, BFU-Meg, CFU-Meg) hematopoietic prog-enitor cells5 (Figure 3). Recently reported data haveshown that low or absent expression of the Thy1 or SCFreceptor can be efficiently used to enrich primitivehematopoietic progenitors from the heterogeneousCD34+ cell population.7,9 Although the CD34 antigen isvirtually expressed by all progenitor cells, the percentageof CD34+ cells with assayable in vitro clonogenic activ-ity ranges from 10 to 30%. The problem of non-clono-genic CD34+ cells is still open and not adequatelyexplained by the presence of lymphoid progenitors whichare not assayable with current in vitro systems. Non-proliferating CD34+ cells might represent a subpopula-tion that is not responsive to conventional myeloidhematopoietic growth factors. The non-proliferatingCD34+ subset might require the presence of co-factors,such as the ligand of STK-1 or the hepatocyte growthfactor, able to activate stem cell-specific genes whoseexpression is a prerequisite for acquiring responsivenessto conventional growth factors.14,22,23

In lethally irradiated non-human primates, both autol-ogous and allogeneic CD34+ cells have been shown tohave the capacity to reconstitute the myelo-lymphopoi-etic system, thus suggesting that the stem cell respon-sible for hematopoietic reconstitution is CD34+.24,25 It

C. Carlo Stella et al.2


Figure 1. May-Grünwald-Giemsa (MGG) staining of cytospinpreparation of enriched CD34+ cells, highly purified byavidin-biotin immunoaffinity.

Figure 2. MGG staining of cytospin preparation of enrichedCD34+ CD33– HLA-DR– cells. The CD34+ cell fraction wasfurther depleted of CD33+ HLA-DR+ cells by immunomag-netic separation.

CD34+ cells: biology and clinical relevance 3


has also been shown that human CD34+ HLA-DR– cellstransplanted in utero in the fetus of sheep initiate andsustain a chimeric hematopoiesis producing humanprogenitor cells of all differentiative lineages.26 Autolo-gous CD34+ cells enriched by avidin-biotin columns havebeen shown to be able to reconstitute myelo-lym-phopoiesis in patients receiving high-dose chemoradio-therapy.27 The results of studies using CD34+ bone mar-row cells in the allogeneic setting in patients receivingboth related as well as unrelated allogeneic marrowtransplants will soon be available.27 In addition, trials areplanned that will use allogeneic peripheral blood CD34+

cells either alone or with marrow.27

Characterization and function of the CD34cell surface molecule

The CD34 cell surface molecule has been biochemi-cally characterized and both the human cDNA and genehave been cloned and sequenced in the last fewyears.28,29

CD34 is a one-pass type I transmembrane glycopro-tein with a molecular weight of 105-120 kDs in eitherthe reduced or unreduced form28 (Figure 4). CD34 pro-tein is not homologous to any other known protein. Theminimum size of the CD34 protein is 354-amino acidsand contains nine sites for N-glycosylation and a sev-eral for O-glycosylation that are essential constituentsof the three epitopes of the molecule; this molecule isalso rich in sialic acid. Its biochemical composition sug-

gests a mucin-like structure and in some respectsresembles leucosialin (CD43). Sequence comparisonsbetween human and mouse CD34 show a very low lev-el of identity in the glycosylated region, 70% identity inthe globular domain, and 92% in the transmembraneand cytoplasmic regions.

Using a KG1 cell line library, it has been shown thatthe human CD34 gene is located on chromosome 1, andrecent studies with in situ hybridization have assignedits localization to band 1q32, in close proximity to oth-er genes that encode growth factors or function mole-cules such as CD1, CD45, TGF2, laminin, LAM/GMP,etc.28

Seven CD34 monoclonal antibodies (MoAbs) wereclustered at the 4th Workshop on Leukocyte Differenti-ation Antigens (Vienna, 1988),30,31 and another 15MoAbs were verified as recognizing the CD34 moleculeduring the 5th International Workshop on LeukocyteDifferentiation Antigens (Boston, 1983), the most directevidence being reactivity with cells transfected withCD34 cDNA and binding to CD34 protein.32,33 The epi-tope specificity of the CD34 antibodies was classifiedinto three distinct groups according to the sensitivity ofthe epitopes to enzymatic cleavage (which was per-formed using neuraminidase, chymopapain and glyco-protease from Pasteurella haemolytica), reactivity withfibroblasts and high endothelial venules, and crossblocking experiments (Table 1).34,35 We know, in fact,that glycoprotease from Pasteurella haemolytica specif-ically cleaves only proteins containing sialylated O-linked glycans.34 Based on these data, it can be furtherpostulated that class III epitopes are more proximal tothe extracellular side of the cell membrane than class Iand class III epitopes.

Furthermore, for most CD34 MoAbs (with few excep-tions) cross blocking experiments are in agreement withthe classification based on enzymatic cleavage of theCD34 protein. In other words, using a cocktail of CD34MoAbs, CD34 reactivity is blocked only in the case thatMoAbs belonging to the same CD34 epitope are simul-taneously employed. On the contrary, the combined useof MoAbs recognizing class II and III or class I and II orclass III epitopes does not affect cell reactivity. More-over, CD34 MoAbs defining class III epitopes are unableto react with CD34 glycoprotein in Western blotsbecause this epitope is sensitive to denaturation.

The pattern of expression of CD34 antibodies exhib-ited by CD34+ acute leukemias is partially in accordancewith that derived from epitope mapping based on thedifferential sensitivity of CD34 to enzymatic treatment.However, about one third of CD34 MoAbs do not seemto belong to any of these subgroups and for this pecu-liar pattern of expression are referred to as atypicalCD34 reagents. The widest variation in CD34 MoAbreactivity has been demonstrated in acute myeloidleukemia (AML) samples, allowing us to postulate theoccurrence of aberrant antigens or of distinct epitopesin subgroups of leukemias. Alternatively, it can behypothesized that the expression of different antibod-Figure 3. Cellular organization of the hematopoietic system.

ies could reflect the degree of maturation of leukemiccells. The presence of new, distinct non overlapping epi-topes could be proposed on the basis of the data pub-lished so far in the literature. As far as the expression ofCD34 in normal and leukemic cells is concerned, it hasbeen calculated by flow cytometry that the number ofmolecular equivalents of soluble fluorochrome (MESF)expressed by leukemic and normal progenitors rangesfrom 18,200 to 322,000 and from 8,000 to 124,000,respectively.

Recent data collected by Lanza et al.36 seem to indi-

cate that class I-type MoAbs are more sensitive to freez-ing procedures than class II and III MoAbs, since theepitope is not identifiable following a frozen/thawedmethodology.

The function of CD34 surface glycoprotein in hemato-poietic stem and progenitor cells is still the object ofdebate. In light of recent findings, it would seem to playa relevant role in modulating cell adhesion.36 Further-more, it has been demonstrated that CD34 probably actsas an adhesive ligand for L-selectin. It has been furtherpostulated that the CD34 molecule could play a protec-tive role against proteolytic enzyme-mediated damagedue to its high number of O-glycosylation sites. The cyto-plasmic domain contains two sites for protein kinase Cphosphorylation and one for tyrosine phosphorylation.28

Given the discordant reactivity of these molecules,the choice of the CD34 MoAb to employ may be impor-tant when analyzing cell positivity for the CD34 mole-cule in both leukemic and normal samples, and may beresponsible for the differences reported by variousauthors in the literature concerning the prognostic roleplayed by this antigen in acute myeloid leukemias.37 Thetype of CD34 MoAb used to enumerate progenitor cellsis probably relevant in the peripheral blood stem cellautograft setting as well, since both early and lateengraftment following transplantation are, to someextent, related to the number of hemopoietic stem cellscollected at the time of blood or bone marrow harvests,and to the degree of progenitor cell maturation relatedto the expression of lineage markers such as HLA-DR,CD71, CD38, CD33, and myeloperoxidase.

Techniques for CD34+ cell separationA number of different techniques have been proposed

for separating CD34+ cells. The common aim is to pro-duce a cell population with optimal purity and viabilityby means of a low cost, rapid and simple separationtechnique. The first separation techniques exploitedparameters such as size and cell density and were rep-resented by Ficoll-Hypaque and Percoll density gradi-



Fig. 4

GPI anchor

Lipid bilayer

N-glycosylation site

O-glycosylation site

Figure 4. Schematic representation of the CD34 cell surfacemolecule.

Epitope class Clones CD34 reactivity

paraffin frozen westernsection section blotting

I. Sensitive to neuraminidase (from Vibriocholerae), chymopapain, glycoprotease (fromPasteurella haemolytica)

14G3, BI3C5, My10,* 12.8,*ICH3,* Immu-133, Immu-409

positive negative positive

II. Resistant to neuraminidase. Glycoprotease,and chymopapain sensitive

43A1, MD34.3,MD34.1, MD34.2, QBend10,4A1, 9044, 9049

positive positive positive

III. Resistant to neuraminidase, chymopapainand glycoprotease

CD34 9F2,HPCA2,581, 553.563, Tuk 3, 115.2

negative positive negativeTable 1. Epitope specificity ofCD34 MoAbs as assessed bytheir differential sensitivity.

*Incomplete digestion by neuroaminidase.

ents. In the last two decades, the development of mon-oclonal antibodies has allowed a more specific and care-ful cell selection by identifying surface antigens used astargets for cell separation (Table 2).

FACS (Fluorescence Activated Cell Sorter)Flow cytometry is able to physically separate differ-

ent populations after incubation of cells with fluo-rochrome-conjugated monoclonal antibodies. This cellsorting technique can yield a highly purified (> 98%)cell population. In addition, the use of electronic gatesallows selection and recovery of several subpopulationsaccording to antigenic expression and different charac-teristics such as size and cytoplasmic granularity. Thistechnique has been very useful for studying CD34+ sub-populations but cannot be employed to select largenumbers of cells due to its complexity and low recov-ery. The recent development of high-speed cell sorting,however, might allow clinical utilization of this tech-nique.

PanningAnti-CD34 monoclonal antibodies bound to one of

the surfaces of cell culture flasks were recently used toselect CD34+ cells. When a cell suspension is introducedinto the flask, the positive population is blocked on theplastic surface, while CD34 negative cells remain in sus-pension and can be easily eliminated. Adherent cellsshould contain the CD34+ population with a viability> 90%.38

Immunomagnetic systemsImmunomagnetic beads are uniform, super-para-

magnetic, polystyrene beads with affinity purified anti-mouse Ig covalently bound to the surface. They areequally suited for negative and positive cell separation;the rosetted target cell can easily be isolated by apply-ing a magnet on the outer wall of the test tube for 1-2minutes. Immunomagnetic beads coupled with CD34monoclonal antibodies can be utilized for positive selec-tion of CD34+ cells to obtain a population with > 80%viable CD34+ cells.39 Similarly, immunomagnetic beadscan be employed for negative depletion with mono-clonal antibodies binding lineage-specific antigens.

High-affinity chromatography based onbiotin-avidin interaction

This method is based on an immunoadsorption tech-nique that relies on the high affinity interaction betweenthe protein avidin and the vitamin biotin. Avidin-biotinimmunoadsorption has been employed for both positiveselection and depletion of specific cell populations.40

The instrument includes a set of non-reusable productsincluding biotinylated anti CD34 antibody, plastic bags,filters and a column of avidin-biotin beads. An auto-mated version controlled by a computer which guaran-tees reproducibility of operation and reduces risks ofoperator errors has been developed for clinical use. Itscapacity has recently been increased so that a singlecolumn can process more than 50×1010 bone marrow orperipheral blood cells in 1-2 hours and sustain bone

marrow engraftment in patients submitted to auto-graft.41

CD34-positive subpopulations: phenotypic and functional analysis

The normal CD34+ cell population likely containsprogenitors of all human lympho-hematopoietic lin-eages, including stem cells capable of hematopoieticreconstitution after bone marrow transplantation.1 Lev-els of CD34 expression decline with differentiation; con-sequently, the earliest clonogenic cells (CFU-blast, LTC-IC, etc.) express the highest levels of CD34, while themost differentiated (CFU-G, CFU-M, CFU-E, CFU-Meg)express only low levels of CD34 (Figure 5). The CD34antigen has been used to identify, enumerate and iso-late cells from different lympho-hematopoietic lineages,as well as develop in vitro tests for indirect evaluationof cells with different functional and clonogenic capac-ity.42

Pre-clinical studiesSeveral animal-human systems have been created to

utilize chimeras for the study of lympho-hematopoiesisin vivo. The first experiments demonstrated the feasibil-ity of transplanting human fetal stem cells to sheepfetuses; the postnatal presence of human cells in thesheep was documented at several points in time.43 Fur-thermore, some early CD34+ subpopulations were ableto repopulate sheep bone marrow; animals were trans-planted in utero with CD34+/DR– cells and the presenceof a chimeric population with the functional character-istics of hemopoietic progenitors was demonstrated inthe marrow and peripheral blood in a percentage of cas-es.44 Berenson et al.45 also showed how hematopoieticprogenitors (positive for the Ia antigen and subsequentlyfor the CD34 antigen) could reconstitute the marrow oflethally irradiated dogs. Of the seven animals treated, allshowed complete marrow engraftment after reinfusionof Ia-positive cells; only one dog died from infection.45

Similarly, marrow cells from 5 primates (baboons) weretreated in vitro with a biotinylated anti-CD34 antibodyand then passed through a column of avidin; after auto-graft, all the animals showed marrow engraftment fol-lowed by hematologic reconstitution comparable to thatof control animals.46 Furthermore, the demonstrationthat allogeneic CD34+ cells can reconstitute the hema-topoietic system in lethally irradiated baboons con-firmed that this cell population includes pluripotentstem cells.25

Lymphoid precursor cellsThe CD34+ cell compartment contains all the cells

expressing terminal deoxynucleotidyltransferase (TdT),which is an intranuclear enzyme expressed in early lym-phoid cells undergoing immunoglobulin or T-cell recep-tor gene rearrangement. Flow cytometry has shown thatthe great majority of TdT+ cells in the marrow coexpressCD34, CD19 and CD10 (B-cell precursors), as well as T-cell differentiation antigens such as CD7, CD5 and CD2.



A small proportion of CD34+/TdT+ cells coexpress CD10,which might represent a common lymphoid progenitorfor both B and T lineages.47 Recently, Miller et al. report-ed that CD34+ cells may also generate NK cells in vitro.48

Granulocyte-macrophage precursorsMarrow erythroid progenitors lack specific markers

and therefore are difficult to identify. Glycophorin A-directed monoclonal antibodies recognize all hemoglo-binized cells, but this molecule is expressed in only asmall proportion of CD34+ cells and is absent in clono-genic cells.49

High levels of CD45 are present on BFU-E, but thisantigen is lost by the CFU-E stage;50 however, theCD45RO isoform is well represented on earlier erythroidprogenitors, while the CD45RA isoform is present oncommitted myeloid progenitors. The expression of CD71(transferrin receptor) is currently considered to be thespecific antigen for the CD34+ erythroid population.CD71 is present at high levels on erythroid progenitorcells and at very low levels in all the other progenitorcells.51 Expression of CD71 increases from stem cells toBFU-E, then declines during erythroid maturation. Inaddition, marrow CD34+ erythroid cells might express IL-3, GM-CSF (CD116) and erythropoietin receptors, basedon the action of these growth factors on CFU-erythroidcells.52

Myeloid precursor clonogenic cells (CFU-GM, CFU-G,CFU-M, CFU-MK) coexpress CD34, HLA-DR, CD117 (c-kit), CD45RA, CD33 and CD13; CD15 is present at lowlevels on CFU-G, while CFU-M specifically expressCD115 (M-CSF receptor). CFU-MK are the only CD34+

cells which express the platelet glycoproteins identifiedby the CD61, CD42 and CD41 monoclonal antibodies.5Dendritic cells also originate from bone marrow, but theconditions that direct their growth and differentiationare still poorly characterized. GM-CSF stimulates thegrowth of dendritic cells from mouse peripheral blood;however, it was recently reported that CD34+ cells maygive rise to dendritic/Langherans cells after stimulationwith GM-CSF and tumor necrosis factor-α.53

Multilineage progenitors and stem cellsCFU-GEMM contain precursor clonogenic cells of both

myeloid and erythroid lineages and express CD34, HLA-DR, CD38, CD117 and CD45RA. They also express lowamounts of CD33, but not CD13. In a hypothetical dif-ferentiation scheme involving pluripotent stem cells, thelympho-hemopoietic compartment is the next cell typeand can be identified in vitro with CFU-Blast and LTC-IC.

The lack of expression of CD38 is the most importantcharacteristic of these early progeny, which represent1% of CD34 and less than 0.01% of mononuclear cells.The lack of CD38 allows separation of committed prog-enitors (CD34+/CD38+) from earlier compartments(CD34+/CD38–) by a single marker combination.54

Furthermore, the earliest CD34+ cells coexpress lowlevels of CD45RO6 and are negative for staining with thefluorescent dye rhodamine 123. The role of HLA-DR indefining earlier cell types is still controversial; a series

of evidence indicates that the stem cell compartmenthas the CD34+/CD38–/HLA-DR+ phenotype.8,54 Recentworks, however, have not found HLA-DR in the earliestcells.55 These data were confirmed by the possibility thatHLA-DR expression may discriminate the Ph-positiveleukemic compartment (HLA-DR+) from normal residualcells (HLA-DR–) in chronic myeloid leukemia.56

Adhesion molecules and cytokinereceptors

A number of molecules within the integrin family havebeen shown to mediate interactions between earlyCD34-positive cells and bone marrow stromal cells.These include VLA-4/VCAM-1, VLA-5/fibronectin, LFA-1or ICAM, and several others. Each adhesion moleculeappears to mediate a specific cell interaction.Hematopoietic growth factors may be active in a solu-ble form or in a membrane-anchored form; adhesionmolecules may be crucial for allowing anchored growthfactors to bind the target cell. Table 3 lists the mainadhesion molecule and growth factor receptorsexpressed on CD34+ cells.

CD34 expression on normal and neo-plastic cells

It has been known that CD34 monoclonal antibodiesbind specifically to vascular endothelium ever since a110 kd protein extracted from freshly isolated umbili-cal vessel endothelium was identified with CD34 anti-bodies in Western blots and in Northern blot analysis.57,58

CD34 molecules have a striking ultrastructural localiza-tion on endothelial cells: they are concentrated primar-ily on the luminal side, in particular on membraneprocesses, many of which interdigitate between adja-cent endothelial cells.57,58 Since this region is an impor-tant site for leukocyte adhesion and transendothelial



Table 2. Recovery of CD34-positive cells after different separation tech-niques.

Enrichment Recovery Large-scaleseparation

(% CD34+ (% CD34+in the recovered of the original

population) sample)Negative depletionby immunomagnetic 20-60 30-60 nobeadsPositive selection byimmunomagnetic 60-80 30-60 yesbeadsFluorescence activatedcell sorter (FACS) > 95 30 - 50 time consumingPanning 50 - 80 30 - 60 yes

Ceprate SC® 50 - 80 40 - 70 yes

traffic, in contrast to previous experiences,33 it has beenhypothesized that CD34 may be antagonistic orinhibitory to the adhesive functions of vascular endo-thelium. This was supported by the demonstration thatCD34 gene expression at the mRNA level is reciprocal-ly down-regulated when adhesion molecules ICAM-1and ELAM-1 are up-regulated by IL-1 during inflamma-tory skin lesions associated with leukocyte infiltra-tion.33,59 Furthermore, additional studies conducted onboth paraffin embedded and cryopreserved sectionsdemonstrated that fibroblasts also react with anti-CD34MoAbs.60 However, it should be noted that while CD34MoAbs reacted with all classes of epitopes on cryopre-served sections, class III epitopes were not recognized byspecific anti-CD34 MoAbs on paraffin embedded sec-tions. Levels of CD34 expression, highest in immaturehematopoietic precursor cells, decrease progressivelywith cell maturation.

Regarding hematologic malignancies, CD34 isexpressed in a large percentage of acute leukemias.61

The fluorescence intensity of CD34 expression is variableand higher in acute lymphoblastic (ALL) than in acutemyeloblastic leukemia (AML). In these latter patients,the CD34 antigen is found on 40-60% of leukemic blastsand is most frequently associated with M0, M1 and M4FAB cytotype, secondary leukemias, karyotypic abnor-malities involving chromosome 5 or 7, P170 expressionand poor prognosis.61-64 Thus, CD34 expression may beconsidered the most predictive negative factor in AML

patients strictly correlated with intensity of expres-sion.65,66 In ALL, CD34 is expressed in 70% of patients,particularly in those with a B-lineage phenotype. Inthese patients, unlike AML cases, the clinical relevanceof CD34 expression is controversial; however, accordingto the findings of a Pediatric Oncology Group,67 itsexpression in B-lineage cases was associated withhyperdiploidy, lower frequency of initial central nervoussystem (CNS) leukemia and a favorable prognosis. In T-cell ALL cases, on the other hand, CD34 expressionshowed a positive correlation with initial CNS leukemiaand CD10 negativity, but not with any presenting favor-able-risk characteristics.67,68

Lastly, CD34 and HLA-DR expression may be very use-ful in discriminating between the very few benign prim-itive hematopoietic progenitors and their malignantcounterparts in patients with chronic myeloid leukemia(CML). In fact, it has been demonstrated that normalprogenitor cells are CD34+ and HLA-DR–, while malig-nant progenitor cells, which exhibit Ph and bcr/ablrearrangement, express HLA-DR antigens.56

Positive and negative regulators ofhematopoietic progenitor cells

The hematopoietic stem cell is defined by its exten-sive self-renewal capacity, multilineage differentiationpotential and capacity for of long-term reconstitutionof normal marrow function in lethally irradiated ani-mals.68 Transplantation of retrovirally marked murine



Figure 5. Functional differentiation and antigenic expression of hematopoietic cells.

stem cells has shown that only a few multilineage prog-enitor cells induce the repopulation of engraftedhematopoietic tissue,69 suggesting that hematopoiesismay be supported by a succession of short-lived clones.Moreover, experimental evidence indicates that theprocesses of self-renewal, differentiation and selectionof lineage potentials are intrinsic properties of the stemcells and occur according to a stochastic (random) mod-el.70 In the steady state, most of the stem cells are qui-escent (G0 phase) and begin active cycling randomly.71

Conversely, survival and proliferation of hematopoieticprogenitors are regulated by cytokines, which also actin preventing apoptosis.72 According to this model, theinduction of differentiation by a cytokine may be con-sidered as the consequence of the proliferation of a spe-cific population stimulated by that factor. The broadterm cytokines includes growth factors such as fibrob-last growth factor, colony stimulating factors (CSFs) (e.g.granulocyte-CSF, granulocyte-macrophage-CSF), andinterleukins (ILs). Depending on their target cells anddifferent proliferative potential, cytokines can be divid-ed into three categories:71 1) late-acting, lineage-spe-cific factors; 2) intermediate-acting, lineage-nonspecif-ic factors; 3) early-acting growth factors inducing therecruitment of dormant early progenitors in the cellcycle (Figure 6).71

The majority of late-acting factors support the prolif-eration and maturation of lineage-committed progeni-tors and the functional properties of terminally differ-entiated cells. Erythropoietin (Epo) is the physiologicregulator of erythropoiesis73 and thrombopoietin has thesame role in thrombopoiesis,74 while M-CSF and IL-5are considered specific for macrophages and eosinophils,respectively. G-CSF exerts its activity on committedneutrophil precursors, although it has been shown to

be a synergistic factor for primitive hematopoieticcells.75

Intermediate-acting, lineage-nonspecific factorsinclude IL-3, IL-4 and GM-CSF. Their activity is mainlydirected toward progenitor cells in the intermediatestages of hematopoietic development. However, theyinteract with later-acting growth factors for the pro-duction of more mature cells,76 as well as with thecytokines capable of triggering stem cell cycling. Ontheir own, they act as survival factors for stem cells andappear to stimulate early hematopoietic precursors onlyafter their exit from G0.77

Several cytokines have recently been identified fortheir capacity to stimulate the proliferation of the ear-liest hematopoietic cells. Mapping studies of normalblast cell colony formation from single progenitors haveshown that IL-6, G-CSF, IL-11, stem cell factor (SCF), IL-12 and leukemia inhibitory factor (LIF) recruit dormantstem cells into the cell cycle that are then able torespond to additional growth factors.77 Whereas the per-missive factors retain limited proliferative potential bythemselves, they strongly enhance the stimulatory activ-ity of IL-3, GM-CSF, G-CSF and EPO on CD34+ and moreimmature CD34+ lineage-progenitors.78 In addition topositive interaction with intermediate-and late-actinggrowth factors, SCF synergistically or additively aug-ments the colony-forming ability of other early-actinggrowth factors, such as IL-11, IL-12 and IL-6, on myeloidand bilineage (i.e. lymphomyeloid) primitive cells.79

Recently, the ligands for the STK-1 or FLT3 receptor, andthe hepatocyte growth factor were shown to be able tostimulate very primitive hematopoietic progenitor cells.However, their biological activity is still under investi-gation.

The main sources of both positive and negative reg-ulatory proteins are accessory myeloid cells and thestromal component of the bone marrow. In general,microenvironment cells do not constitutively producecytokines, rather transcription and/or translationprocesses are rapidly induced by cytokines such as IL-1,TNF. The extracellular matrix also participates in theregulation of hematopoiesis by binding growth factorsand presenting them in a biologically active form tobone marrow progenitor cells.80

Most data suggest that stem cells express low levelsof growth factor receptors and require multiple prolif-erative stimuli to enter into the cell cycle, while com-mitted progenitor cells can be effectively stimulated byindividual cytokines.23 Combinations of two or moregrowth factors81 can stimulate hematopoietic cellseither by amplifying the progeny cell production of sin-gle precursors (synergy) or by inducing additional clono-genic cells to proliferate (recruitment). Examples ofthese two types of enhancement are given in Figures 7and 8, where CD34+ CD33– DR– cells are simultaneous-ly stimulated by two or three growth factors. The mol-ecular basis regulating the complex interplay betweencytokines is still largely unknown. However, one pro-posed mechanism for growth factor synergism is the



Table 3. Adhesion molecule and growth factor receptorsexpressed on CD34-positive cells.

Antigen name CD Ligand

GlyCAM-L/selectin none adhesion molecule

ICAM1,2,3 CD11a/CD18 adhesion molecule

H-CAM CD44 adhesion molecule

VLA-4 CD49d/CD29 adhesion molecule

VLA-5 CD49d/CD29 adhesion molecule

FGF-R none growth factor

M-CSF CD115 growth factor

GM-CSF-R CD116 growth factor

SCF (c-kit) CD117 growth factor

Interferon γ-R CD119 growth factor

IL 7-R CD127 growth factor

induction of CSF receptors on early hematopoietic prog-enitor cells.82 This appears to be a coordinate cascadetransactivation via up-modulation of growth factorreceptors that leads to proliferation and differentiationof human marrow cells.83 Conversely, the structuralhomology between some growth factors,84 the presenceof shared receptor subunits on the cell membrane85 andcommon signal-transduction proteins86 provide a poten-tial explanation for their functional similarities and theapparent redundancy of their activity.

The growth of hematopoietic progenitor cells is alsoregulated by soluble negative regulators such asmacrophage inflammatory protein-1a (MIP-1a), tumornecrosis factor-α (TNF-α), interferons (IFNs), prosta-glandins and transforming growth factor-β (TGF-β). TheTGF-β family of proteins includes at least five isoforms(TGF-β1-5) which are encoded by different genes87 andproduced by stromal cells, platelets and bone cells.Moreover, a subset of very primitive murine hematopoi-etic cells has been shown to secrete active TGF-β1 by anautocrine mechanism.88 TGF-β1 and 2 isoforms arebimodal regulators of murine and human hematopoiet-ic progenitor cells, and their activity is based upon thedifferentiation state of the target cells and the pres-ence of growth factors.89 In the human system, forinstance, CFU-GEMM derived from purified CD34+

CD33– cells are inhibited by TGF-β1, whereas more com-mitted progenitors such as CFU-G or CFU-GM are notaffected. In addition, high proliferative potential-CFC(HPP-CFC) responsive to a combination of CSFs aremarkedly inhibited by TGF-β1, while more mature CFU-

GM are actually enhanced when GM-CSF is used ascolony forming factor.78 TGF-β1 and 2-induced myelo-suppression is partially counteracted by early-actinggrowth factors such as G-CSF, IL-6 and fibroblastgrowth factor.90 On the other hand, TGF-β3 has beenshown to be a more potent suppressor of human BMprecursors and its activity on hematopoiesis is onlyinhibitory,89 although the synergistic growth factors IL-11 and IL-9 seem to be able to oppose the negative reg-ulation of TGF-β3 on human CD34+ cells.91 Severalpotential modes of action of the TGF-β family have beensuggested, including down-modulation of cytokinereceptors,92 interaction with the underphosphorylatedform of the retinoblastoma gene product in late G1phase,93 and alteration of gene expression.94 Early stud-ies have shown that TGF-β1 and 3 exert their activity onnormal and leukemic cells by lengthening or arrestingthe G1 phase of the cell cycle,95 and this effect is func-tional to protect normal CD34 positive cells from thetoxicity of alkylating agents in vitro.96 More recentinvestigations have demonstrated that TGF-β regulatesthe responsiveness of mice hematopoietic cells to SCF,which is known to be the main synergistic factor forboth murine and human stem cells, through a decreasein c-kit mRNA stability that leads to decreased cell-sur-face expression.97

Similarly to TGF-β, TNF-α has been reported to haveboth inhibitory and stimulatory effects on hematopoi-etic progenitor cells. TNF-α potentiated the IL-3 andGM-CSF-mediated growth of human CD34+ cells inshort-term liquid culture assay. However, it inhibited



AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA

AAAAAAAAA

AAAAAAAAAAAAAAAAAAAAAAAAAAAAAA







STEM CELLS PROGENITORCELLS

MATURINGCELLS

Neutrophils

Red cells

Platelets

SCFIL-11IL-12IL-6G-CSF

IL-1GM-CSFIL-3IL-9

IL-4G-CSFGM-CSFIL-3

SCFM-CSFIL-5

GM-CSFIL-3EPO

IL-9SCF

SCFIL-3GM-CSF

IL-6IL-11

G-CSFGM-CSF

M-CSFIL-5

EPO

IL-6MEG-CSF

EPO

Figure 6. Cytokines exert their activity at different levels of the hematopoietic differentiation pathway. Each progenitor cell isconcurrently affected by multiple regulators.

the growth promoting activity of G-CSF.98 Two TNFreceptors with molecular weights of 55 and 75 kd,respectively, have recently been identified.99 Whereasthe p55 receptor mediates TNF-α effects on committedprogenitor cells, the p75 receptor is involved in signal-ing the inhibition of murine primitive cells.

Furthermore, it was recently shown that TNF-α iscapable of inhibiting the multi-growth factor (GM-CSF,IL-3, G-CSF, SCF, IL-1)-dependent growth of humanHPP-CFC derived from CD34+ cells through interactionwith both p55 and p75 receptors, while the p55 recep-tor exclusively mediates the bifunctional activity of TNF-α on more mature BM precursors responsive to singlecytokines.100

MIP-1a is a peptide of 69-amino acids with a mole-cular weight of 7.8 kd produced by activated macro-phages, T-cells and fibroblasts.101 It belongs to a largefamily of putative cytokines that includes MIP-1β, MIP-2 and IL-8 (chemokines). Biologic characterization hasshown that MIP-1α enhances the M-CSF- and GM-CSF-dependent growth of CFU-GM, while it inhibits thecolony-forming ability of hematopoietic precursors pre-sent in a cell population enriched for CD34++ DR+ cellsstimulated with erythropoietin, IL-3 and GM-CSF.102

Taken together, these results indicate that a complexinterplay between positive and negative regulatory pro-teins determines the proliferation or inhibition of earlyhematopoietic progenitor cells (Figure 9). In general, theactivity of inhibitors of hemopoiesis appears to bereversible, lineage-nonspecific and directed at the ear-ly stages of differentiation. In addition, TGF-β has shownsome degree of differential activity between normal andneoplastic lymphoid cells.96 Thus, negative regulatorsmay be clinically relevant to the protection of the hema-topoietic stem cell compartment from the dose-limitingtoxicity of neoplastic disease therapy.96,103,104

Collection of CD34+ cellsBone marrow and peripheral blood are the only

sources of immature hemopoietic precursors identifiedas CD34+ cells. A diagnostic marrow sample containsonly a few CD34+ cells, while even fewer of them arepresent in peripheral blood samples taken under steadystate conditions. Large quantities of CD34+ cells can becollected with massive marrow harvests, such as fortransplantation purposes. Nevertheless, marrow CD34+

cells are somewhat elusive due to their scatteringamong the predominant CD34+ hematopoietic popula-tion.105 Recently developed cell separation proceduresallow collection of highly enriched CD34+ cell popula-tions. However, this is generally accomplished throughaspecific and often unacceptable cell loss.106 So far thelimited number of immature precursors commonlyobtained from both bone marrow and peripheral bloodhas been the major obstacle to a simple identificationand analysis of marrow CD34+ cells. Indeed, the grow-ing interest in CD34+ cells is primarily the result of thedevelopment of new therapeutic modalities that alloweasy access to large quantities of hemopoietic precur-sors through the peripheral blood. The key role in theseinnovative approaches is represented by the introduc-tion of hemopoietic growth factors for clinical use.107

At present GM-CSF and G-CSF are the most exten-sively employed and studied hemopoietic growth factorsin the clinical setting. From the very beginning, it wasobserved that GM-CSF or G-CSF administration wasassociated with an increase in circulating hemopoieticprogenitors.108,109

Later on, it was demonstrated that this phenomenoncould be extensively and reproducibly amplified by com-bining growth factor administration with high-dosechemotherapy.110-112

Indeed hemopoietic progenitors are massively, thoughtransiently, mobilized into peripheral blood duringhemopoietic recovery following high-dose chemother-apy given with growth factor support. Such an abun-dance of immature hemopoietic cells makes them eas-ily recognizable by cell sorting techniques that employanti-CD34 MoAbs.113 Under optimal conditions, the pro-portion of CD34+ cells may reach values as high as 20-30% of the total leukocyte count. In steady state con-ditions CD34+ cells do not exceed 4% of the total mar-row population, while they are undetectable by cell sort-



Figure 7. Human CD34+ CD33- DR- cells were stimulated byIL-9 (A) or IL-9 and SCF (B) in the presence of EPO. The addi-tion of SCF induced the growth of large multicentered BFU-E colonies containing > 10,000 cells. The different size ofthe colonies indicates amplification of stem cell progeny(synergy).

A

B

ing techniques in the peripheral blood. Chemotherapyand growth factors do not merely induce a relativeincrease of immature hemopoietic cells; their absolutenumber is amplified several times over basal conditions.This allows collection of sufficient amounts of hemo-poietic progenitors for autografting purposes by meansof a few leukapheresis procedures.110,113,114

Values of 10-20×104 CFU-GM/kg represent the min-imal required dose for marrow engraftment with periph-eral blood progenitors.115-117 In fact, much higher quan-tities of circulating progenitors can be collected usingappropriate mobilization procedures. Under optimalconditions, circulating CD34+ cell peak values may rangebetween 150 and 700/µL on days of maximal mobiliza-tion. As a rule, at least 8×106 CD34+ cells/kg or more canthus be collected with 1 to 3 leukapheresis proce-

dures.114 These huge quantities, approximately corre-sponding to 50×104 CFU-GM/kg, must be consideredthe ideal threshold dose of peripheral blood progenitorsfor autografting purposes. Indeed values of 8×106 CD34+

cells/kg or more guarantee a rapid and durable hemo-poietic recovery when circulating progenitors are usedas the sole source for marrow reconstitution followingsubmyeloablative treatment.118-121

Thus far, massive CD34+ cell mobilization has beenmost commonly observed when growth factor is admin-istered following high-dose cyclophosphamide, given at7 g/m2. Indeed chemotherapy that induces profoundleukocytopenia seems to be crucial for optimal mobi-lization; for instance, cyclophosphamide at doses low-er than 7 g/m2 produces a reduced mobilizing stimu-lus.111,122 Several other chemotherapy schedules havealso been successfully employed for mobilization pur-poses. The principal chemotherapy protocols reportedto be highly effective in inducing CD34+ cell mobiliza-tion are summarized in Table 4.110, 111, 115, 117, 122-127

As stressed earlier, hemopoietic growth factors play akey role in mobilization. This has been clearly docu-mented with cyclophosphamide. A median peak value of75 CD34+ cells/µL has been recorded following high-dose cyclophosphamide alone, whereas 420 and 500/µLare the median values recorded when GM-CSF or G-CSF, respectively, are added to cyclophospha-mide.112,113,128,129 Extensive growth factor-induced mobi-lization is further substantiated by the possibility of col-lecting sufficient CD34+ cells using growth factoralone.130,131 In this setting the most promising experi-ences have been produced with G-CSF. The resultsreported indicate new opportunities for the utilizationof mobilized progenitors in allogeneic transplantationprocedures.132,133 Combined chemotherapy and growth



Figure 8. The addition of SCF to IL-11, in the presence ofEPO, results in a much higher number of colony-formingcells derived from CD34+ CD33– DR– progenitors (right), ascompared to the IL-11 and EPO combination (left) (recruit-ment).

Differentiation

Self renewal

Negative Positive Survival

IFN a, b, gTGF-b3

TGF-b1-2

TNF-a

MIP-1a

SCFIL-11IL-12IL-6G-CSFIL-1IL-9(?)

IL-11GM-CSF

Figure 9. Interplay between pos-itive and negative regulatory pro-teins acting on hematopoieticstem cells.

factors remain the most efficient mobilization proce-dure at this time. However, ongoing studies are direct-ed at improving mobilization efficiency with growth fac-tors alone. For example, G-CSF at doses higher than 5µg/kg/day up to 20 µg/kg/day may have higher mobi-lization capacity.134 Further improvement might derivefrom the clinical availability of new cytokines, such asIL-3, SCF and others, to be employed alone or in com-bination with G- or GM-CSF.135-137 However, the poten-tially high mobilizing activity of such new cytokine com-binations must be accompanied by no or very few sideeffects in order to be considered for a wide clinicalapplicability.

Mobilizing protocols generally include daily deliveryof growth factors, starting 1 to 3 days after chemother-apy administration and continuing until harvesting pro-cedures are completed. Growth factor is usually admin-istered for a total of 7-12 consecutive days. The mostconvenient route of delivery is subcutaneous, with 1 to2 doses per day. Progenitor cell harvests are performedduring hemopoietic recovery, provided that progenitorcell mobilization is documented. Indeed various para-meters have been considered as an indirect indication ofprogenitor mobilization, including an increase in WBC or,alternatively, in monocytes, basophils or platelets.138-140

However, CD34+ cell evaluation remains mandatory foran accurate definition of the extent of progenitor mobi-lization.141,142

CD34+ cells should be monitored daily from the early

stages of hemopoietic recovery. Detection of circulatingCD34+ cells is not sufficient reason for starting leuka-pheresis procedures; an adequate number of CD34+ cells(>20-50/µL) and WBC > 1.0×109/L and platelet count> 30×109/L are required for safe and effective progeni-tor cell harvesting.114,142 Leukaphereses are performedusing continuous-flow blood cell separators. A completeleukapheresis procedure generally takes 2-3 hours andthe total blood volume processed ranges from 6 to 10liters.114,140,143,144 The harvested cells are resuspended infreezing medium and then cryopreserved for subsequenttransplantation. One to 3 leukaphereses repeated onconsecutive days usually provide large amounts of prog-enitor cells capable of rapid engraftment after submye-loablative treatments.

Factors and procedures favoring CD34+ cell mobiliza-tion have been well established. However, other condi-tions adversely influencing the mobilization phenome-non should be considered. A major limitation is repre-sented by impaired marrow function, as can occur inpreviously treated patients.111 In fact, patients at firstrelapse following a single treatment protocol maintainan adequate mobilization capacity;145 however, few ifany mobilized CD34+ cells can be obtained from heavi-ly treated patients previously exposed to multiplechemotherapy courses. Mobilization is also profoundlyreduced and often totally abolished in patients previ-ously exposed to radiotherapy, especially if it was deliv-ered to the pelvis or to extended vertebral areas.117,123

Lastly, marrow invasion by tumor cells may negativelyaffect mobilization capacity. This is typically reported inmyeloma patients in whom the extent of CD34+ cellsoften correlates with the degree of marrow involvementby tumoral plasma cells.122 In conclusion, several newfindings have dramatically improved the procedures forcollection of large amounts of CD34+ cells. However,optimal marrow function is a prerequisite for exploitingfully the activity of all those factors known for theirmobilization-inducing capacity.

CD34+ positive cells in umbilical cord bloodSignificant numbers of human hematopoietic stem

cells can be found in umbilical cord blood and can beused for allogeneic bone marrow transplantation.Mayani et al.146 found that 1-2% of the total number ofcord blood-derived low-density cells express high levelsof the CD34 antigen and low or undetectable levels ofthe antigens CD45RA and CD71. These populations werehighly enriched in clonogenic cells (34%), in particularin multipotent progenitors (12%). By culturing thesecells at low concentration for 8-10 days in highlydefined serum-free liquid cultures supplemented withvarious hematopoietic cytokines, it was possible toachieve a significant expansion (about 100-fold) of theCD34+ cell population.

These last results were recently confirmed by Traycottet al.147 in an elegant study showing that umbilical cordblood CD34+ cells rapidly exit G0-G1 phases and start tocycle in response to stem cell factor. This feature would



Table 4. Main chemotherapy protocols employed in hematopoietic prog-enitor mobilization.

Authors Protocol Drug(s) Dosagecharacteristics employed

Gianni et al.110 single agent cyclophosphamide highTarella et al.123 single agent etoposide highGianni et al.124

Kotasek et al.121 single agent cyclophosphamide intermediateTarella et al.122

Dreyfus et al.125 single agent cytarabine highSchimazaki et al.126 multiple drugs etoposide high

cytarabineKawano et al.115 multiple drugs daunorubicin intermediate

cyclophosphamideetoposide

Pettengell et al.127 multiple drugs doxorubicin intermediatecyclophosphamideetoposide

Hass et al.117 multiple drugs cytarabine highmitoxantrone

make cord blood CD34+ cells more suitable candidatesthan bone marrow cells for ex vivo expansion.

It is clear that in vitro expansion and maturation ofhematopoietic progenitor cells might be of particularrelevance for transplantation of cord blood hematopoi-etic cells; however, information about the effects ofsuch an expansion on the cells required for long-termhematopoietic reconstitution is highly desirable.

The dual role of peripheral bloodhematopoietic progenitor cells in onco-hematology

CD34+ progenitor cells circulating in the peripheralblood represent an enriched and easily accessible sourceof two distinct cell populations: committed progenitorcells and hematopoietic stem cells.

Although circulating progenitors are commonly calledstem cells, the presence of circulating stem cells hasbeen formally proved only in mice.148 In humans thepresence of hematopoietic stem cells in the peripheralblood is almost certain (see below), but to call reinfu-sion of circulating progenitors transplantation ofperipheral blood stem cells (PBSC or similar acronyms)is hardly appropriate. This terminology overlooks the factthat the tremendous interest in peripheral blood auto-grafting is not (at least so far) a consequence of its beinga simple surrogate for autologous bone marrow trans-plantation (as the term PBSC would suggest), but ratherderives from the unique property of this procedure toreduce the duration of the severe pancytopenia that fol-lows submyeloablative treatments from two-threeweeks (when bone marrow cells are used) to a few daysonly.110,149 The reason for the rapid recovery which occursafter circulating progenitor autotransfusion (CPAT) hasnot been formally proved, but it is most likely a conse-quence of the much larger (10- to 100-fold higher)amount of committed progenitors reinfused when apatient is autografted with blood-derived (as opposed tomarrow-derived) cells. The most likely hypothesis is thatthese late progenitors (post-progenitors) of granulo-cytes and platelets are capable of giving rise to matureprogeny within a few days, thus allowing submyeloab-lated patients to survive the initial aplastic phase.

The fundamental role of committed progenitor cellswas elegantly proved in mice by Jones et al.150 None ofthe lethally irradiated animals transplanted with a purepopulation of stem cells free of more mature progeni-tors (CFU-GM, CFU-S) survived the initial aplasia. Amore recent paper151 challenged this ‘conventional wis-dom’ model, and maintained that peripheral blood stemcells per se are capable of rapidly maturing in vivo.

Other authors, using a very similar approach, reachedan opposing conclusion.152 In humans the most con-vincing, albeit indirect, evidence in favor of the role ofcommitted progenitors in accelerating post-transplantrecovery was provided by Robertson et al.,153 who doc-umented a significantly prolonged hematopoietic recov-ery for patients undergoing autologous bone marrow

transplantation purged ex vivo with anti-CD33 mono-clonal antibodies. This result, which occurred after atreatment that selectively kills the most mature prog-enitor cells (expressing the CD33 surface antigen), rep-resents convincing indirect proof of the role of com-mitted progenitors in early engraftment.

The clinical role of committed progenitors in reduc-ing the morbidity and mortality of high-dose therapyhas expanded since hemopoietic growth factors havebecome clinically available. In fact, infusion of rhGM-CSF or rhG-CSF, in particular after administration ofmyelotoxic doses of certain stem cell-sparing agents,allows easy collection of an amount of CFU-GM/kg bodyweight 10 to 100 times higher than that contained in abone marrow harvest.110

As already documented by a large number of papers,the use of circulating progenitors has brought about adramatic change in the perspectives of high-dose sub-myeloablative regimens. In fact, these once specialized,expensive and highly toxic treatments are now well tol-erated, easy to administer, and clinically useful (costeffective). As an example, it is worth mentioning theinitial Milan Cancer Institute experience in a group ofover 50 poor-risk breast cancer patients who receivedhigh-dose melphalan plus an optimal amount of circu-lating progenitors (≥ 5×104 CFU-GM/kg body weight).These patients required a median of 10.5, 11 and 12.5days to score > 0.5, 1.0 and 2.0×109/L neutrophils/µL,respectively. Moreover, more than half of them did notrequire platelet transfusions, while the remaining onesneeded only one or two transfusions during the firstweek after autografting. Such mild to moderate toxici-ty was never described before the clinical availability ofcommitted progenitor cells.

In conclusion, born as a compassionate surrogate ofbone marrow autografting, today CPAT is rapidly replac-ing ABMT. In fact, today it is the latter that should beconsidered a compassionate need procedure, useful inthose few patients unable to mobilize a sufficient num-ber of circulating progenitor cells.

The second, distinct role of circulating progenitors isrelated to the presence of stem cells, i.e. totipotent pre-cursors responsible for durable reconstitution of all lym-pho-hematopoietic lineages following marrow ablativetherapy. Since virtually no high-dose treatment that canbe safely administered to humans is genuinely mye-loablative, formal proof of the existence of circulatingstem cells must await either stable transduction of DNAmarkers into autografted cells or the use of this cellpopulation for allografting.

These experiments are presently underway in severallaboratories,154,155 and preliminary data do confirm thepresence of stem cells in the peripheral blood of humans.Their utilization is expected to revolutionize fields likeallogeneic bone marrow transplantation and somaticgene therapy, whenever the target of genetic manipu-lations is the hematopoietic cell.156



References

1. Civin CI, Strauss LC, Brovall C, Fackler MJ, SchwartzJF, Shaper JH. Antigenic analysis of hematopoiesis III.A hematopoietic progenitor cell surface antigendefined by a monoclonal antibody raised against KG-1a cells. J Immunol 1984; 133:157-65.

2. Sutherland RD, Keating A. The CD34 antigen: struc-ture, biology and potential clinical applications. JHematother 1992; 1:115-29.

3. Andrews RG, Singer JW, Bernstein ID. Precursors ofcolony-forming cells in humans can be distinguishedfrom colony-forming cells by expression of the CD33and CD34 antigens and light scatter properties. J ExpMed 1989; 169:1721-31.

4. Andrews RG, Singer JW, Bernstein ID. Monoclonalantibody 12.8 recognizes a 115-kd molecule presenton both unipotent and multipotent hematopoieticcolony-forming cells and their precursors. Blood1986; 67:842-52.

5. Civin CI, Gore SD. Antigenic analysis of hemato-poiesis: a review. J Hematother 1993; 2:137-44.

6. Lansdorp PM, Sutherland HJ, Eaves CJ. Selectiveexpression of CD45 isoforms on functional subpopu-lations of CD34+ hematopoietic cells from humanbone marrow. J Exp Med 1990; 172:363-6.

7. Baum CM, Weissman IL, Tsukamoto AS, Buckle AM.Isolation of a candidate human hematopoietic stem-cell population. Proc Natl Acad Sci USA 1992; 89:2804-8.

8. Huang S, Terstappen LWMM. Lymphoid and myeloiddifferentiation of single human CD34+, HLA-DR–,CD38– hematopoietic stem cells. Blood 1994; 83:1515-26.

9. Gunji Y, Nakamura M, Osawa H, et al. Human prim-itive hematopoietic progenitor cells are more enrichedin KITlow cells than in KIThigh cells. Blood 1993;82:3283-9.

10. Rosnet O, Marchetto S, de Lapeyriere O, Birnbaum D.Murine Flt3, a gene encoding a novel tyrosine kinasereceptor of the PDGFR/CSF1R family. Oncogene1991; 6:1641-50.

11. Matthews W, Jordan CT, Wiegand GW, Pardoll D,Lemischka IR. A receptor tyrosine kinase specific tohematopoietic stem and progenitor cell-enriched pop-ulations. Cell 1991; 65: 1143-52.

12. Small D, Levenstein M, Kim E, et al. STK-1, the humanhomologue of Flk-2/Flt-3, is selectively expressed inCD34+ human bone marrow cells and is involved inthe proliferation of early progenitor/stem cells. ProcNatl Acad Sci USA 1994; 91:459-63.

13. Lyman SD, James L, van den Bos T, et al. Molecularcloning of a ligand for the flt3/flk-2 tyrosine kinasereceptor: a proliferative factor for primitive hemato-poietic cells. Cell 1993; 75: 1157-67.

14. Lyman SD, James L, Johnson L, et al. Cloning of thehuman homologue of the murine flt3 ligand: a growthfactor for early hematopoietic progenitor cells. Blood1994; 83:2795-801.

15. Hannum C, Culpepper J, Campbell D, et al. Ligand forFLT3/FLK2 receptor tyrosine kinase regulates growthof haematopoietic stem cells and is encoded by vari-ant RNAs. Nature 1994; 368:643-8.

16. McCulloch EA. Stem cells in normal and leukemichemopoiesis (Henry Stratton lecture, 1982). Blood1983; 62:1-13.

17. Metcalf D. The molecular control of cell division, dif-ferentiation commitment and maturation in haemo-poietic cells. Nature 1989; 339:27-30.

18. Orlic D, Bodine DM. What defines a pluripotent

hematopoietic stem cell (PHSC): will the real PHSCstand up! Blood 1994; 84:3919-94.

19. Ogawa M. Differentiation and proliferation of hema-topoietic stem cells. Blood 1993; 81:2844-53.

20. Tsai FY, Keller G, Kuo FC, et al. An early haematopoi-etic defect in mice lacking the transcription factorGATA-2. Nature 1994; 371:221-6.

21. Sutherland HJ, Eaves CJ, Eaves AJ, Dragowskas W,Lansdorp PM. Characterization and partial purifica-tion of human marrow cells capable of initiating long-term hematopoiesis in vitro. Blood 1989; 74:1563-70.

22. Kmiecik TE, Keller JE, Rosen E, van de Woude GF.Hepatocyte growth factor is a synergistic factor forthe growth of hematopoietic progenitor cells. Blood1992; 80:2454-7.

23. Metcalf D. Hematopoietic regulators: redundancy orsubtlety? Blood 1993; 82:3515-23.

24. Berenson RJ, Andrews RG, Bensinger WI, et al. CD34+marrow cells engraft lethally irradiated baboons. J ClinInvest 1988; 81:951-9.

25. Andrews RG, Bryant EM, Bartelmez SH, et al. CD34marrow cells, devoid of T and B lymphocytes, recon-stitute stable lymphopoiesis and myelopoiesis inlethally irradiated allogeneic baboons. Blood 1992;80:1693-9.

26. Srour EF, Zanjani ED, Cornetta K, et al. Persistence ofhuman multilineage, self-renewing lymphohema-topoietic stem cells in chimeric sheep. Blood 1993;82:3333-42.

27. Berenson R. Transplantation of hematopoietic stemcells. J Hematother 1993; 2:347-9.

28. Greaves MF, Brown J, Molgaard HV, et al. Molecularfeatures of CD34: a hemopoietic progenitor cell-asso-ciated molecule. Leukemia 1992; 6:31-6.

29. Silvestri F, Banavali S, Baccarani M, Preisler HD. Prog-enitor cell associated antigen CD34: biology and clin-ical applications. Haematologica 1992; 77:265-72.

30. Peschle CH, Köller U. Cluster report: CD34. In: KnappW, Dörken B, Gilks WR, Rieber EP, Schmidt RE, SteinH, von dem Borne AEGKr, eds. Leukocyte typing IV:white cell differentiation antigens. Oxford:Oxford Uni-versity Press, 1989: 817-8.

31. Civin CI, Trishmann TM, Fackler MJ, et al. Report onthe CD34 cluster workshop. In: Knapp W, Dörken B,Gilks WR, Rieber EP, Schmidt RE, Stein H, von demBorne AEGKr, eds. Leukocyte typing IV: white cell dif-ferentiation antigens. Oxford:Oxford University Press,1989: 818-25.

32. Lanza F, Moretti S, Papa S, Malavasi F, Castoldi G.Report on the Fifth international Workshop onHuman Leukocyte Differentiation Antigens, Boston,November 3-7, 1993. Haematologica 1994; 79:374-86.

33. Greaves MF, Colman SM, Bühring HJ, et al. Report onthe CD34 cluster workshop. In: Schlossman SF,Boumsell L, Gilks W, et al, eds. Leukocyte typing V:White cell differentiation antigens. Oxford:OxfordUniversity Press, 1995, in press.

34. Sutherland DR, Marsh JCW, Davidson J, Baker MA,Keating A, Mellors A. Differential sensitivity of CD34epitopes to cleavage by Pasteurella haemolytica gly-coprotease: implications for purification of CD34-positive progenitor cells. Exp Hematol 1992; 20:590-9.

35. Lanza F, Castoldi GL. Large scale enrichment ofCD34+ cells by Percoll density gradients: a CML-basedstudy design. In: Wunder E, Serke S, Solovat H, HenonP, eds. Hematopoietic stem cells. Dayton:AlphaMedPress, 1993:255-70.

36. Lanza F, Bi S, Castoldi GL, Goldman JM. Abnormalexpression of N-CAM (CD56) adhesion molecule on





myeloid and progenitor cells from chronic myeloidleukemia. Leukemia 1993; 7:1570-5.

37. Lanza F, Rigolin GM, Moretti S, Latorraca A, Castol-di GL. Prognostic value of immunophenotypic char-acteristic of blasts cells in acute myeloid leukemia.Leuk Lymphoma 1994; 13(Suppl 1):81-5.

38. Morecki S, Topalian SL, Myers WW, Okrongly D,Okarma TB, Rosenberg SA. Separation and growthof human CD4+ and CD8+ tumor infiltrating lym-phocytes and peripheral blood mononuclear cells bydirect positive panning on covalently attached mon-oclonal antibody coated-flasks. J Biol Resp Modifiers1990; 9:463-74.

39. Civin CI, Strauss LC, Fackler MJ, Trismann TM, WileyJM, Loken RL. Positive stem cell selection. Basic sci-ence. In: S Gross, A Gee, DA Worthington White, eds.Bone Marrow Purging and Processing. New York:Wiley-Liss, 1990:387-402.

40. Berenson RJ, Bensinger WI, Kalamasz D. Eliminationof Daudi lymphoblasts from human bone marrowusing avidin-biotin immunoadsorption. Blood 1986;67:509-15.

41. Berenson RJ, Bensinger WI, Hill RS, et al. Engraftmentafter infusion of CD34+ marrow cells in patients withbreast cancer or neuroblastoma. Blood 1991; 77:1717-22.

42. Carlo-Stella C, Mangoni L, Piovani G, Garau D, Almi-ci C, Rizzoli V. Identification of Philadelphia-negativegranulocyte-macrophage colony-forming units gener-ated by stroma-adherent cells from chronic myeloge-nous leukemia patients. Blood 1994; 83:1373-80.

43. Zanjani ED, Pallavicini MG, Ascensao JL, et al.Engraftment and long-term expression of human fetalhemopoietic stem cells in sheep following transplan-tation in utero. J Clin Invest 1992; 89:1178-88.

44. Srour EF, Zanjani ED, Brandt JE, et al. Sustainedhuman hematopoiesis in sheep transplanted in uteroduring early gestation with fractionated adult humanbone marrow cells. Blood 1992; 79:1404-12.

45. Berenson RJ, Bensinger WI, Kalamasz D, et al. Engraft-ment of dogs with Ia-positive marrow cells isolated byAvidin-Biotin immunoadsorption. Blood 1987; 69:1363-7.

46. Berenson RJ, Andrews RG, Bensinger WI, et al. Anti-gen CD34+ marrow cells engraft lethally irradiatedbaboons. J Clin Invest 1988; 81:951-5.

47. Gore SD, Kastan MB, Civin CI. Normal human bonemarrow precursors that express terminal deoxynu-cleotidyl transferase include T-cell precursors and pos-sible lymphoid stem cells. Blood 1991; 77:1681-90.

48. Miller JS, Verfaillie C, McGlave P. The generation ofhuman natural killer cells from CD34+/DR– primitiveprogenitors in long term bone marrow culture. Blood1992; 80:2182-7.

49. Loken MR, Civin CI, Bighee WL, Langlois RG, JensenRH. Coordinate glycosylation and cell surface expres-sion of glycophorin A during normal human erythro-poiesis. Blood 1987; 70:1959-61.

50. Shaw VO, Civin CI, Loken MR. Flow cytometric analy-sis of human bone marrow. IV. Differential quantita-tive expression of T-200 common leukocyte antigenduring normal hematopoiesis. J Immunol 1988;140:1861-7.

51. Loken MR, Shaw VO, Dattilio K, Civin CI. Flow cyto-metric analysis of human bone marrow. I. Normal ery-throid development. Blood 1987; 69:255-63.

52. Sawada K, Krantz SB, Kans JS, et al. Purification ofhuman colony-forming units-erythroid and demon-stration of specific binding of erythropoietin. J ClinInvest 1987; 80:357-66.

53. Caux C, Dezutter-Dambuyant C, Schmitt D, Ban-chereau J. GM-CSF and TNF-α cooperate in the gen-

eration of dendritic Langherans cells. Nature 1992;360:258-61.

54. Terstappen LWMM, Huang S, Safford M, Lansdorp P,Loken MR. Sequential generation of hematopoieticcolonies derived from single nonlineage-committedCD34+/CD38– progenitor cells. Blood 1991; 77:1218-27.

55. De Fabritiis P, Dowding C, Bungey J, et al. Phenotyp-ic characterization of normal and CML CD34-posi-tive cells: only the most primitive CML progenitorsinclude Ph-neg cells. Leuk Lymphoma 1993; 11:51-61.

56. Verfaillie C, Miller WJ, Boylan K, McGlave PB. Selec-tion of benign primitive hematopoietic progenitors inchronic myelogenous leukemia on the basis of HLA-DR antigen expression. Blood 1992; 79:1003-10.

57. Beschoner WE, Civin CI, Strauss LC. Localization ofhemopoietic progenitor cells in tissue with the anti-My-10 monoclonal antibody. Am J Pathol 1985;119:1-6.

58. Fina L, Molgaard HV, Robertson D, et al. Expressionof the CD34 gene in vascular endothelial cells. Blood1990; 75: 2417-26.

59. Delia D, Lampugnani MG, Resnati M, et al. CD34expression is regulated reciprocally with adhesion mol-ecules in vascular endothelial cells in vitro. Blood1993; 81:1001-8.

60. Majdic O, Johannes Stockl, Pickl WF, et al. Signalingand induction of enhanced cytoadhesiveness via thehematopoietic progenitor cell surface molecule CD34.Blood 1994; 83: 1226-34.

61. Borowitz MJ, Gockerman JP, Moore JO, et al. Clinico-pathologic and cytogenetic features of CD34 (MY10)-positive acute nonlymphocytic leukemia. Am J ClinPathol 1989; 91:265-70.

62. Campos L, Guyotat D, Archimbaud E, et al. Surfacemarker expression in adults with acute myelocyticleukemia: correlations with initial characteristics, mor-phology and response to therapy. Br J Haematol1989; 72:161-6.

63. Geller RB, Zahurak M, Hurwitz CA, et al. Prognosticimportance of immunophenotyping in adults withacute myelocytic leukaemia: the significance of thestem-cell glycoprotein CD34 (My10). Br J Haematol1990; 76:340-7.

64. Myint H, Lucie NP. The prognostic significance of theCD34 antigen in acute myeloid leukaemia. Leuk Lym-phoma 1992; 7:425-9.

65. Lauria F, Raspadori D, Ventura MA, et al. CD7 expres-sion does not affect the prognosis in acute myeloidleukemia. Blood 1994; 83:3097-8.

66. Borowitz Mj, Shuster JJ, Civin CI, et al. Prognostic sig-nificance of CD34 expression in childhood B-precur-sor acute lymphocytic leukemia: a Pediatric Oncolo-gy Group Study. J Clin Oncol 1990; 8:1389-96.

67. Pui CH, Hancock ML, Head DR, et al. Clinical signif-icance of CD34 expression in childhood acute lym-phoblastic leukemia. Blood 1993; 82:889-98.

68. Till JE, McCulloch EA. Hemopoietic stem cell differ-entiation. Biochim Biophys Acta 1980; 605:431-59.

69. Van Zant G, Chen JJ, Scott-Micus K. Developmentalpotential of hematopoietic stem cells determinedusing retrovirally marked allogenic marrow. Blood1991; 77:756-63.

70. Till JE, McCulloch EA, Siminovitc L. A stochastic mod-el of stem cell proliferation, based on the growth ofspleen colony-forming cells. Proc Natl Acad Sci USA1964; 51:29-34.

71. Ogawa M. Differentiation and proliferation ofhematopoietic stem cells. Blood 1993; 81:2844-53.

72. Bergamaschi G, Rosti V, Danova M, Lucotti C, Caz-zola M. Apoptosis: biological and clinical aspects.



Haematologica 1994; 79:86-93.73. Barosi G. Control of erythropoietin production in

man. Haematologica 1993; 78:77-9.74. Cazzola M. The end of a long search: at last throm-

bopoietin. Haematologica 1994; 79:397-9.75. Ikebuchi K, Clark SC, Ihle JN, Souza LM, Ogawa M.

Granulocyte colony-stimulating factor enhances inter-leukin-3-dependent proliferation of multipotentialhemopoietic progenitors. Proc Natl Acad Sci USA1988; 85:3445-9.

76. Sonoda Y, Yang YC, Wong GG, Clark SC, Ogawa M.Analysis in serum-free culture of the targets of recom-binant human hemopoietic factors: interleukin-3 andgranulocyte/macrophage colony-stimulating factorare specific for early developmental stages. Proc NatlAcad Sci USA 1988; 85: 4360-6.

77. Leary AG, Zeng HQ, Clark SC, Ogawa M. Growth fac-tor requirements for survival in G0 and entry into thecell cycle of primitive human hematopoietic progeni-tors. Proc Natl Acad Sci USA 1992; 89:4013-7.

78. Moore MAS. Clinical implications of positive and neg-ative hematopoietic stem cell regulators. Blood 1991;78:1-19.

79. Hyrayama F, Shih JP, Awgulewitsch A, Warr GW,Clark SC, Ogawa M. Clonal proliferation of murinelymphohematopoietic progenitors in culture. ProcNatl Acad Sci USA 1992; 89: 5907-11.

80. Gordon MY, Riley GP, Watt SM, Greaves MF.Compartimentalization of a haematopoietic growthfactor (GM-CSF) by glycosaminoglycans in the bonemarrow microenvironment. Nature 1987; 326:403-5.

81. Montagna C, Massaro P, Morali F, Foa P, Maiolo AT,Eridani S. In vitro sensitivity of human erythroid prog-enitors to hemopoietic growth factors: studies on pri-mary and secondary polycythemia. Haematologica1994; 79:311-8.

82. Jacobsen SEW, Ruscetti FW, Dubois CM, Wine J,Keller JR. Induction of colony-stimulating factor recep-tor expression on hematopoietic progenitor cells: Pro-posed mechanism for growth factor synergism. Blood1992; 80:678-87.

83. Testa U, Pelosi E, Gabbianelli M, et al. Cascade oftransactivation of growth factor receptors in earlyhuman hematopoiesis. Blood 1993; 81:1442-56.

84. Hirano T, Yasukawa K, Harada H, et al. Complemen-tary DNA for a novel human interleukin (BSF-2) thatinduces B lymphocytes to produce immunoglobulin.Nature 1986; 324: 73-6.

85. Kitamura T, Sato N, Arai K, Miyajima A. Expressioncloning of the human IL-3 receptor cDNA reveals ashared b subunit for the human IL-3 and GM-CSFreceptors. Cell 1991; 66: 1165-9.

86. Yin T, Taga T, Lik-Shing Tsang M, Yasukawa K, Kishi-moto T, Yang YC. Involvement of IL-6 signal trans-ducer gp130 in IL-11-mediated signal transduction. JImmunol 1993; 151: 2555-61.

87. Massague J. The transforming growth factor-family.Ann Rev Cell Biol 1990; 6:597-641.

88. Ploemacher RE, van Soest PL, Boudewijn A. Autocrinetransforming growth factor 1 blocks colony forma-tion and progenitor cell generation by hemopoieticstem cells stimulated with steel factor. Stem Cells1993; 11:336-47.

89. Jacobsen SEW, Keller JR, Ruscetti FW, Kondaiah P,Roberts AB, Falk LA. Bidirectional effects of trans-forming growth factor β (TGF-β) on colony-stimulat-ing factor-induced human myelopoiesis in vitro: dif-ferential effects of distinct TGF- isoforms. Blood 1991;78:2239-47.

90. Keller JR, Jacobsen SEW, Dubois CM, Hestdal K,Ruscetti FW. Transforming growth factor-α: a bidi-rectional regulator of hematopoietic cell growth. Int J

Cell Cloning 1992; 10:2-11.91. Lemoli RM, Fogli M, Fortuna A, Tura S. Interleukin-11

(IL-11) and IL-9 counteract the inhibitory activity oftransforming growth factor β3 (TGF β3) on humanprimitive hemopoietic progenitor cells. Haematolo-gica 1995; 80:5-12.

92. Jacobsen SEW, Ruscetti FW, Dubois CM, Lee J, BooneTC, Keller JR. Transforming growth factor-β trans-modulates the expression of colony stimulating factorreceptors on murine hematopoietic progenitor celllines. Blood 1991; 77:1706-16.

93. Laiho M, Di Caprio JA, Ludlow JW, Livingstone DM,Massague J. Growth inhibition by TGF-β linked to sup-pression of retinoblastoma protein phosphorylation.Cell 1990; 62: 175-85.

94. Takehara K, Le Roy EC, Grotendorst GR. TGF-β inhi-bition of endothelial cell proliferation: alteration ofEGF binding and EGF-induced growth-regulatory(competence) gene expression. Cell 1987; 49:415-22.

95. Strife A, Lambek C, Perez A, et al. The effects of TGF-β3 on the growth of highly enriched hematopoieticprogenitor cells derived from normal human bonemarrow and peripheral blood. Cancer Res 1991; 15:4828-36.

96. Lemoli RM, Strife A, Clarkson BD, Haley JD, GulatiSC. TGF-β3 protects normal human hematopoieticprogenitor cells treated with 4-hydroperoxycyclo-phosphamide in vitro. Exp Hematol 1992; 20: 1252-6.

97. Dubois CM, Ruscetti FW, Stankowa J, Keller JR. Trans-forming growth factor-β regulates c-kit message sta-bility and cell-surface protein expression in hemato-poietic progenitors. Blood 1994; 83:3138-45.

98. Caux C, Saeland S, Favre C, Duvert V, Mannoni P,Bancherau J. Tumor necrosis factor-a strongly poten-tiates interleukin-3 and granulocyte-macrophagecolony-stimulating factor-induced proliferation ofhuman CD34+ hematopoietic progenitor cells. Blood1990; 75:2292-8.

99. Schall TJ, Lewis M, Koller KJ, et al. Molecular cloningand expression of a receptor for human tumor necro-sis factor. Cell 1990; 61:361-4.

100. Rusten LS, Jacobsen FW, Lesslauer W, Loetscher H,Smeland EB, Jacobsen SEW. Bifunctional effects oftumor necrosis factor-α (TNF-α) on the growth ofmature and primitive human hematopoietic progeni-tor cells: involvement of p55 and p75 TNF receptors.Blood 1994; 83:3152-9.

101. Graham GJ, Wright EG, Hewick R, et al. Identifica-tion and characterization of an inhibitor of hemato-poietic stem cell proliferation. Nature 1990; 344:442-4.

102. Broxmeyer HE, Sherry B, Lu L, et al. Enhancing andsuppressing effects of recombinant murine macro-phage inflammatory proteins on colony formation invitro by bone marrow myeloid progenitor cells. Blood1992; 76:1110-6.

103. Dunlop DJ, Wright EG, Lorimore S, et al. Demon-stration of stem cell inhibition and myeloprotectiveeffects of SCI/MIP1α in vivo. Blood 1992; 79:2221-5.

104. Moreb J, Zucali JR, Rueth S. The effects of tumornecrosis factor-α on early human hematopoietic prog-enitor cells treated with 4-hydroperoxycyclophos-phamide. Blood 1990; 76:681-9.

105. Andrews RG, Singer JW, Bernstein ID. Precursors ofcolony forming cells in human can be distinguishedfrom colony forming cells by expression of the CD33and CD34 antigens and light scatter properties. J ExpMed 1989; 169:1721-31.

106. Gabbianelli M, Sargiacomo M, Pelosi E, Testa U, Isac-chi G, Peschle C. Pure human hematopoietic progen-itors: permissive action of basic fibroblast growth fac-



tor. Science 1990; 249:1561-4. 107. Peters W, Rosner G, Ross M, et al. Comparative

effects of granulocyte-macrophage colony-stimulat-ing factor (GM-CSF) and granulocyte colony-stimu-lating factor (G-CSF) on priming peripheral bloodprogenitor cells for use with autologous bone mar-row transplantation after high-dose chemotherapy.Blood 1993; 81:1709-19.

108. Duhrsen U, Villeval JL, Boyd J, Kannourakis G,Morstyn G, Metcalf D. Effects of recombinant humangranulocyte colony-stimulating factor on hematopoi-etic progenitor cells in cancer patients. Blood 1988;72:2074-81.

109. Socinski MA, Elias A, Schnipper L, Cannistra SA,Antman KH, Griffin JD. Granulocyte-macrophagecolony-stimulating factor expands the circulatinghaemopoietic progenitor cell compartment in man.Lancet 1988; i:1194-8.

110. Gianni AM, Siena S, Bregni M, et al. Granulocyte-macrophage colony-stimulating factor to harvest cir-culating haematopoietic stem cells for autotrans-plantation. Lancet 1989; ii:580-5.

111. Kotasek D, Shepherd KM, Sage RE, et al. Factorsaffecting blood stem cell collections following high-dose cyclophosphamide mobilization in lymphoma,myeloma and solid tumors. Bone Marrow Transplant1992; 9:11-7.

112. Teshima T, Harada M, Takamatsu Y, et al. Cytotoxicdrug and cytotoxic drug/G-CSF mobilization ofperipheral blood stem cells and their use for auto-grafting. Bone Marrow Transplant 1992; 10:215-20.

113. Siena S, Bregni M, Brando B, Ravagnani F, Bonadon-na G, Gianni AM. Circulation of CD34-positivehematopoietic stem cells in the peripheral blood ofhigh-dose cyclophosphamide treated patients:enhancement by intravenous recombinant humanGM-CSF. Blood 1989; 74:1905-14.

114. Ravagnani F, Siena S, Bregni M, Sciorelli G, GianniAM, Pellegris G. Large scale collection of circulatinghaematopoietic progenitors in cancer patients treat-ed with high-dose cyclophosphamide and recombi-nant human GM-CSF. Eur J Cancer 1990; 26:562-4.

115. Kawano Y, Takaue Y, Watanabe T, et al. Effects ofprogenitor cell dose and preleukapheresis use ofhuman recombinant granulocyte colony-stimulatingfactor on the recovery of hematopoiesis after bloodstem cell autografting in children. Exp Hematol 1993;21:103-8.

116. Bensinger W, Singer J, Appelbaum F, et al. Autolo-gous transplantation with peripheral blood mononu-clear cells collected after administration of recombi-nant granulocyte stimulating factor. Blood 1993; 81:3158-63.

117. Haas R, Mohle R, Fruhauf S, et al. Patient character-istics associated with successful mobilizing and auto-grafting of peripheral blood progenitor cells in malig-nant lymphoma. Blood 1994; 83:3787-94.

118. To LB, Dyson PG, Juttner CA. Cell-dose effect in cir-culating stem-cell autografting. Lancet 1986; ii:404-5.

119. Gianni AM, Tarella C, Siena S, et al. Durable and com-plete hematopoietic reconstitution after autograftingof rhGM-CSF exposed peripheral blood progenitorcells. Bone Marrow Transplant 1991; 6:143-5.

120. Hohaus S, Goldschmidt H, Ehrhardt R, Haas R. Suc-cessful autografting following myeloablative condi-tioning therapy with blood stem cells mobilized bychemotherapy plus rhG-CSF. Exp Hematol 1993; 21:508-14.

121. Siena S, Bregni M, Di Nicola M, et al. Durability ofhematopoiesis following myeloablative cancer thara-py and autografting with peripheral blood hematopoi-etic progenitors. Ann Oncol 1994; 5:935-41.

122. Tarella C, Boccadoro M, Omedè P, et al. Role ofchemotherapy and GM-CSF on hemopoietic progen-itor cell mobilization in multiple myeloma. Bone Mar-row Transplant 1993; 11: 271-7.

123. Tarella C, Ferrero D, Siena S, et al. Conditions influ-encing the expansion of the circulating hemopoieticprogenitor cell compartment. Haematologica 1990;75:11-4.

124. Gianni AM, Bregni M, Siena S, et al. Granulocyte-macrophage colony stimulating factor or granulocyte-colony stimulating factor infusion makes high-doseetoposide a safe outpatient regimen that is effective inlymphoma and myeloma patients. J Clin Oncol 1992;10:1955-62.

125. Dreyfus F, Leblond V, Belanger C, et al. Peripheralblood stem cell collection and autografting in highrisk lymphoma. Bone Marrow Transplant 1992; 10:409-13.

126. Shimazaki C, Oku N, Ashihara E, et al. Collection ofperipheral blood stem cells mobilized by high-doseAra-C plus VP-16 or aclarubicin followed by recom-binant human granulocyte-colony stimulating factor.Bone Marrow Transplant 1992; 10:341-6.

127. Pettengell R, Testa NG, Swindell R, Crowther D, Dex-ter TM. Transplantation potential of hematopoieticcell released into the circulation during routinechemotherapy for non-Hodgkin’s lymphoma. Blood1993; 82:2239-48.

128. Tarella C, Ferrero D, Bregni M, et al. Peripheral bloodexpansion of early progenitor cells after high-dosecyclophosphamide and rhGM-CSF. Eur J Cancer1991; 27:22-7.

129. Bender JG, Lum L, Unverzagt KL, et al. Correlation ofcolony-forming cells, long-term culture initiating cellsand CD34+ cells in apheresis products from patientsmobilized for peripharal blood progenitors with dif-ferent regimens. Bone Marrow Transplant 1994;13:479-85.

130. Haas R, Ho AD, Bredthauer U, et al. Successful autol-ogous transplantation of blood stem cells mobilizedwith recombinant human granulocyte-macrophagecolony-stimulating factor. Exp Hematol 1990; 18:94-8.

131. Sheridan WP, Begley CG, Juttner C, et al. Effect ofperipheral blood progenitor cells mobilized by fil-grastim (G-CSF) on platelet recovery after high-dosechemotherapy. Lancet 1992; i:640-4.

132. Weaver CH, Buckner CD, Longin K, et al. Syngeneictransplantation with peripheral blood mononuclearcells collected after the administration of recombi-nant human granulocyte colony-stimulating factor.Blood 1993; 82:1981-4.

133. Matsunaga T, Sakamaki S, Ohi S, Hirayama Y, NiitsuY. Recombinant human granulocyte colony-stimulat-ing factor can mobilize sufficient amounts of periph-eral blood stem cells in healthy volunteers for allo-geneic transplantation. Bone Marrow Transplant1993; 11:103-8.

134. Schwinger W, Mache Ch, Urban Ch, Beaufort F, Togl-hofer W. Single dose of filgrastim (rhG-CSF) increas-es the number of hematopoietic progenitors in theperipheral blood of adult volunteers. Bone MarrowTransplant 1993; 11:489-92.

135. Brugger W, Bross K, Frisch J, et al. Mobilization ofperipheral blood progenitor cells by sequential admin-istration of interleukin-3 and granulocyte-macrophagecolony-stimulating factor following polychemothera-py with etoposide, ifosfamide and cisplatin. Blood1992; 79:1193-200.

136. D’Hondt V, Guillaume T, Humblet Y, et al. Toleranceof sequential or simultaneous administration of IL-3and G-CSF in improving peripheral blood stem cell



harvesting following multi-agent chemotherapy: apilot study. Bone Marrow Transplant 1994; 13:261-4.

137. de Revel T, Appelbaum FR, Storb R, et al. Effects ofgranulocyte colony-stimulating factor and stem cellfactor, alone and in combination, on the mobiliza-tion of peripheral blood cells that engraft lethally irra-diated dogs. Blood 1994; 83:3795-9.

138. Liu KY, Akashi K, Harada M, Takamatsu Y, Niho Y.Kinetics of circulating haematopoietic progenitorsduring chemotherapy-induced mobilization with orwithout granulocyte colony-stimulating factor. Br JHaematol 1993; 84:31-8.

139. Bishop MR, Anderson JR, Jackson JD, et al. High-dosetherapy and peripheral blood progenitor cell trans-plantation: effects of recombinant human granulo-cyte-macrophage colony-stimulating factor on theautograft. Blood 1994; 83: 610-6.

140. Jansen WE. Peripheral blood and bone marrowhematopoietic stem cells: are they the same? SeminOncol 1993; 20:19-27.

141. Kessinger A, Armitage JO. The evolving role of autol-ogous peripheral stem cell transplantation followinghigh-dose therapy for malignancies. Blood 1991;77:211-3.

142. Siena S, Bregni M, Brando B, et al. Flow cytometry forclinical estimation of circulating hematopoietic prog-enitors for autologous transplantation in cancerpatients. Blood 1992; 77:400-9.

143. Menichella G, Pierelli L, Foddai ML, et al. Autologousblood stem cell harvesting and transplantation inpatients with advanced ovarian cancer. Br J Haema-tol 1991; 79:444-50.

144. Pettengell R, Morgenstern GR, Woll PJ, et al. Periph-eral blood progenitor cell transplantation in lym-phoma and leukemia using a single apheresis. Blood1993; 82:3770-7.

145. Caracciolo D, Gavarotti P, Aglietta M, et al. High-dose sequential (HDS) chemotherapy with blood andmarrow cell autograft as salvage treatment in verypoor prognosis, relapsed non-Hodgkin lymphoma.Bone Marrow Transplant 1993; 12:621-5.

146. Mayani H, Dragowska W, Lansdorp PM. Cytokine-induced selective expansion and maturation of ery-throid versus myeloid progenitors from purified cordblood precursor cells. Blood 1993; 81:3252-8.

147. Traycott CM, Abboud MR, Laver J, Clapp DW, SrourEF. Rapid exit from G0-G1 phases of cell cycle inresponse to stem cell factor confers on umbilical cordblood CD34+ cells an enhanced ex vivo expansionpotential. Exp Hematol 1994; 22:1264-72.

148. Molineux G, Pojda Z, Hampson IN, Lord BI, DexterTM. Transplantation potential of peripheral bloodstem cells induced by granulocyte colony-stimulatingfactor. Blood 1990; 76:2153-8.

149. Gianni AM, Bregni M, Siena S, et al. Rapid and com-plete hemopoietic reconstitution following combinedtransplantation of autologous blood and bone mar-row cells. A changing role for high dose chemo-radio-therapy? Hematol Oncol 1989; 7:139-48.

150. Jones RJ, Wagner JE, Celano P, Zicha MS, Sharkis SJ.Separation of pluripotent hematopoietic stem cellsfrom spleen colony forming cells. Nature 1990;347:188-9.

151. Uchida N, Aguila HL, Fleming WH, Jerabek L, Weiss-man IL. Rapid and sustained hematopoietic recoveryin lethally irradiated mice transplanted with purifiedThy-1.1loLin-Sca-1+ hematopoietic stem cells. Blood1994; 83:3758-79.

152. Bradford G, Williams N, Barber L, Bertoncello I. Tem-poral thrombocytopenia after engraftment withdefined stem cells with long-term marrow reconstitu-tion activity. Exp Hematol 1993; 21:1615-20.

153. Robertson MJ, Soiffer RJ, Freedman AS, et al. Humanbone marrow depleted of CD33-positive cells medi-ates delayed but durable reconstitution of hemato-poiesis: clinical trial of MY9 monoclonal antibody-purged autografts for the treatment of acute myeloidleukemia. Blood 1992; 79:2229-36.

154. Dreger P, Suttorp M, Haferlach T, et al. Allogeneicgranulocyte colony-stimulating factor-mobilized peri-pheral blood progenitor cells for treatment of engraft-ment failure after bone marrow transplantation.Blood 1993; 81:401-9.

155. Brenner MK, Rill DR, Moen RC, et al. Gene markingto trace origin of relapse after autologous bone-mar-row transplantation. Lancet 1993; i: 85-6.

156. Bregni M, Magni M, Siena S, Di Nicola M, Bonadon-na G, Gianni AM. Human peripheral bloodhematopoietic progenitors are optimal targets ofretroviral mediated gene transfer. Blood 1992; 80:1418-22.


review

Peripheral blood stem cells in acutemyeloid leukemia: biology and clinicalapplications


MASSIMO AGLIETTA, ARMANDO DE VINCENTIIS, LUIGI LANATA,FRANCESCO LANZA, ROBERTO M. LEMOLI, GIACOMO MENICHELLA,AGOSTINO TAFURI, PAOLA ZANON, SANTE TURA

Clinica Medica, Università di Torino, Novara; Dompé BiotecSpA, Milano; Amgen Italia SpA; Cattedra di Ematologia, Uni-versità di Ferrara, Ferrara; Istituto di Ematologia Seràgnoli,Università di Bologna, Bologna; Servizio di Ematologia edEmotrasfusione, Università Cattolica del Sacro Cuore, Roma;and Cattedra di Ematologia, Università La Sapienza, Roma

Correspondence: Prof. Sante Tura, Istituto di Ematologia L. e A.Seràgnoli, Policlinico S. Orsola, via Massarenti 9, 40138 Bologna,Italy.Received September 22, 1995; accepted November 27, 1995.

Ackowledgments: preparation of this manuscript was supported by agrant from Dompé Biotec SpA and Amgen Italia SpA, Milan, Italy.

the expression of a limited self-renewal potential.6 Lapi-dot et al. provided the most convincing evidence of thestem cell role of CD34+ cells in AML by showing thatonly the CD34+/CD38– cell fraction was capable of gen-erating acute leukemia when transplanted into SCIDmice.7

These observations indicate the relevance of definingthe growth and receptor expression pattern of leukemicCD34+ cells, their response to CSFs as well as theirkinetic status compared to their normal counterparts.Among the different cytokines involved in the regula-tion of hemopoiesis, a key role in the pathogenesis ofthe leukemic growth is probably played by stem cellfactor (SCF), interleukin 3 (IL-3), granulocyte-macro-phage CSF (GM-CSF) and G-CSF.8-12

SCF receptor (c-kit) is expressed by the vast majori-ty of AML.8,9 Both high and low affinity receptors havebeen demonstrated (Table 1). c-kit shares structuralsimilarities with the receptors for M-CSF and PDGF. Alinear correlation between the percentage of CD34+

cells and c-kit expression has been documented, thusindicating that CD34+ AML express high levels of c-kit.In adult patients, the presence of a high number ofCD34+ cells has been shown to correlate with a badprognosis.

c-kit activation plays a foundamental role in the reg-ulation of the early phases of CD34+ cell stimulation.The interaction of SCF with its ligand exerts a modestproliferative stimulus on immature quiescent cells andup-regulates the expression of receptors for othergrowth factors. While in normal hematopoiesis this trig-gers myeloid differentiation, in AML it may activateself-renewal and expansion of the leukemic popula-tion.11-14

High affinity receptors for GM-CSF and IL-3 (Table 1)are expressed by nearly all AML, irrespective of the FABsubtype.15,16 IL-3, GM-CSF (and IL-5) receptors consist ofan α subunit (ligand specific) and a shared β subunit.While the α subunit has a low affinity for the ligand andalone is incapable of transducing the signal, the asso-ciation of the two subunits gives rise to a functioning

Clinical application of circulating stem cells forautologous transplantation is steadily expanding.1It has become increasingly clear that mobilized

peripheral blood progenitor cells (PBSC) induce fasterhematopoietic recovery, fewer febrile days, lower trans-fusion requirement and shorter hospitalization thanbone marrow (BM)-derived cells.2,3 More recently, rapidand sustained engraftment has also been reported usinggranulocyte colony-stimulating factor (G-CSF)-mobi-lized allogeneic PBSC following myeloablative therapy.4

In contrast to solid tumors and many hematologicalmalignancies, PBSC transplantation is not widely usedfor acute myeloblastic leukemia (AML) patients. In thissetting there are still unanswered questions such as therole of autologous stem cell transplantation in post-remission therapy, as well as major issues concerningPBSC mobilization and collection: the expression ofCD34 antigen on leukemic stem cells as compared totheir normal counterparts, the biologic significance ofCD34+ AML, the response of leukemic cells to CSFs usedto optimize PBSC harvest, the potential contaminationof PBSC grafts by residual AML cells and the role of ex-vivo purging of leukemic cells.

This review analyzes the most recent advances in thisfield, addressing clinical and biological issues relevantto the use of autologous PBSC for AML patients.

Growth factor receptor expression andresponse of leukemic cells to human CSFs

The CD34 antigen is a 105-120 KD glycoproteinexpressed on the cell surface of hematopoietic progen-itors and stem cells, but it is not expressed on latehematopoietic cells or on many tumor cells.1 CD34+ cellsare responsible for the self-renewal and the expansionof the large majority of AML. It has recently been shownthat most of the clonogenic cells in AML derive from theCD34+ cell fraction as opposed to CD34– cells.5,6 More-over, CD34+ cells co-expressing differentiation markers(CD33, CD38) have a reduced proliferative potentialsince in vitro they give rise to small colonies unable tooriginate secondary clones. This phenomenon is likely

20


high affinity receptor which is a type I receptor devoidof endogenous tyrosine-kinase activity. β chain activa-tion induces several tyrosine-kinases like Fyn, Lyn, Fps,Jaks, which transduce a signal common to IL-3 and GM-CSF,17 whereas a subunit activation induces ligand spe-cific pathways. In the majority of AML cases IL-3 andGM-CSF induce the proliferation of CD34+, although tovariable extents. Correlations have been observedbetween responses to different factors but no significantadditive effects have been noted.

Exposure of leukemic cells to GM-CSF or IL-3 in vitrocan give rise to a generation of mature cells. However,the persistence of blast cells capable of secondaryleukemic colony formation indicates that the differen-tiation potential of IL-3 and GM-CSF is negligible andthat they are unable to abolish the self-renewal of theleukemic population.18-20 SCF synergizes with IL-3 andGM-CSF in inducing large clones primarily composed ofundifferentiated cells.

G-CSF receptor is expressed by nearly all AML (Table1).16,21 However, M2 and M3 AML appear to express thehighest number of receptors. In vitro growth stimulationis not consistent except for M2 and M3 AML; G-CSFaction is additive or synergic with that of IL-3, SCF and,to a lesser extent, with that of GM-CSF. G-CSF alsoinduces some degree of differentiation of leukemicCD34+ cells, and the presence of the growth factoraffects the formation of secondary colonies. In addition,CSF treatment seems to prevent cell death in AML.22

The in vivo use of growth factors in AML patientsderives from contrasting hypotheses:

a) use of growth factors before and during cytostatictreatment to induce the proliferation of quiescentleukemic progenitors. The increased proliferative rateand, possibly, the intracellular accumulation of somecytotoxic drugs (i.e. Ara-CTP) should increase the frac-tion of cells killed.23,24 This approach has never been test-ed in randomized trials specifically addressing this issue.It seems, however, to be of modest value with G-CSF orGM-CSF.25,26 It remains to be seen if this approach wouldbe more useful with molecules such as SCF that are par-ticularly active on leukemic CD34 cells;

b) use of growth factors as differentiating agents withthe aim of exhausting the self-renewal potential of theleukemic progenitors. On the basis of in vitro and pre-liminary (although still to be confirmed) in vivo data, G-CSF seems the most promising molecule;27

c) use of growth factors for accelerating the recoveryof residual normal progenitors after induction chemo-therapy. This approach has been pursued with G-CSFand GM-CSF in AML patients > 60 years of age, forwhom the pancytopenia following cytotoxic treatmentis particularly profound and long-lasting and carries arelevant risk of life-threatening infections. In this set-ting, both G-CSF and GM-CSF given after inductionchemotherapy reduce the duration of neutropenia with-out affecting the rate of severe infections.28,29 Moreover,G-CSF, but not GM-CSF, appears to increase the com-plete remission rate. Both cytokines, however, have noimpact on the survival rate. Of interest, no evidence ofaccelerated growth of residual leukemic cells has beenobserved.

All these data demonstrate how controversial the useof hemopoietic growth factors in the treatment of AMLis, although the most recent results suggest the safetyof G-CSF administration following induction-consoli-dation treatment.29

Stem cell kinetics in AMLThe hematopoietic cell renewal process is supported

by a small population of bone marrow cells termedhematopoietic stem cells. They are defined as cells capa-ble of long-term hematopoietic reconstitution and dif-ferentiation into multiple hematopoietic lineages. It isgenerally held that, in the steady state, the majority ofnormal stem cells are dormant in the cell cycle and onlya few of them supply all the hematopoietic cells at a giv-en time. More than thirty years ago, stem cell kineticstudies30 proposed the concept of a true resting stateand coined the term G0 as the state from which stemcells randomly move to the active cell cycle.

Subsequent studies31 confirmed Lajtha’s observationsby showing that brief in vitro exposure of bone marrowcells to highly specific radioactive thymidine does notreduce the number of multipotential progenitors. Asshown in Figure 1, most normal bone marrow CD34+

progenitor cells are indeed quiescent in G0. Culture ofenriched human progenitors documented that theyremain as single cells for as long as 2 weeks in cultureand begin proliferation upon stimulation with combi-nations of cytokines.32 Based on mathematical studies,stem cell function was seen as a model in which thedecision to self-renew and differentiate followed a sto-chastic process.33 By replating individual blast cellcolonies, Till and coworkers showed that the productionof secondary blast cell colonies is a self-renewal processand that the generation of secondary multilineagecolonies is differentiation. Thus the self-renewal processis associated with renewed dormancy in the cell cyclewhile the differentiation process is characterized bycontinuous cell doubling.

M. Aglietta et al.

Table 1. High affinity receptors for hematopoietic growth factors in AML.

Affinity, kd No. of receptors Reference (range) per cell

SCF 16-158 pmol/L 200-8000 14G-CSF 36-130 pmol/L 55-1200 16

GM-CSF 64-404 pmol/L 40-1263 15,16IL-3 26-467 pmol/L 21-145 15,16

21


Similar work was performed on leukemic stem cells byseveral authors and two fundamental antithetical mod-els were proposed, based on the presence of quiescentprogenitor cells in human leukemia.

It was postulated that leukemic progenitors were pre-dominantly involved in rapid cell cycling as judged bymeasuring the proportion of S-phase cells using H3-TdR or hydroxyurea.34 However, this was in contrast withprevious observations obtained in vivo by continuousinfusions of 3H-thymidine for 8-10 days.35 In thoseexperiments, 88 to 93% of the leukemic cells werelabeled at the end of infusion, whereas almost all thesmallest leukemic cells were not, suggesting that theywere in an extended G0. Using in vivo pulse labeling withtritiated thymidine in AML patients, it was furthershown that blasts with a high proliferative rate do notbehave as a pool of normal self-maintaining cells, butrather as a normal multiplication-maturation compart-ment.36 Data obtained with different techniques (label-ing and cell culture methods) and the heterogeneity ofstudy cell populations may represent the reason forthese discordant results. The hypothesis that AML prog-enitors are characterized by a substantial number ofnonproliferating or very slowly proliferating blast cells(lower RNA content)37,38 was the rationale for differentapproaches to AML treatment. For instance, the com-bined use of cytokines and chemotherapy to recruit qui-escent cells into the cell cycle, enhancing the cytotox-icity of cycle-specific agents.39,40

Raymakers et al.41 have studied the proliferativecapacity of the bone marrow fraction double stainedfor CD34 and CD33 in AML patients. The cloning effi-cacy was highly variable in different AML samples, withpredominant cluster growth. Cluster and colony growthwas similar between CD34–/CD33+ and CD34+/CD33+,in contrast to what is observed in normal bone marrow.The most primitive CD34+/CD33– fraction was found inhighly proliferative colony growth. When this analysiswas extended to AML with a more mature phenotype(small fraction of CD34+/33–), the highly proliferativecolonies deriving from the CD34+/33– fraction werefound to be disomic by in situ hybridization in allpatients who were characterized by chromosomalabnormalities. Nevertheless, the authors could notexclude the presence of leukemic stem cells kineticallycharacterized by low or no proliferation under theirexperimental culture conditions.

A further study42 aimed at evaluating the specificactivity of SCF on enriched CD34+ in suspension cultureby measuring Ki67 expression and flow cytometric DNAcontent showed no difference in cell cycle distributionamong progenitors obtained from normal bone marrow,umbilical cord blood and chronic myeloid leukemiaCD34+ peripheral blood stem cells.

Further investigations on the role of a family of pro-teins recently identified as cell cycle regulators, such ascyclin A, B, D, E and of their catalytic subunits, the cyclin-dependent kinases cdk2, cdk4, cdk6 and cdc2, may helpto identify kinetic features and fine differences between

normal and leukemic hemopoietic stem cells, as well asevents involved in neoplastic transformation.43

In conclusion, the kinetic characteristics of leukemicstem cells have still not been defined, mainly becausedifferent experimental conditions allow evaluation ofprogenitors with different degrees of maturation andtherefore with different proliferative characteristics. Theheterogeneity among different leukemia subtypesshould also be taken into account.

Expression of CD34 antigen in AML andCD34+ leukemias: clinical and biologicalsignificance

Based on current information, there is no doubt thata substantial number of acute leukemias express theCD34 antigen on the cell membrane of blast cells. How-ever, the incidence of such expression in AML has beenfound to be highly variable (25-64% of the patientsexamined), depending on a number of factors, as shownin Table 2.

The variability in the reported incidence of CD34+ AMLhas also influenced the prognostic relevance of CD34expression in AML. Most authors found a clear associ-ation between CD34+ AML and a lower incidence ofcomplete remission following induction therapy. Inaddition, the relapse rate was higher in AML showingpositivity for the CD34 antigen compared to that of theCD34– group.44-53 However, other authors did not con-firm these results and found no significant difference inthe complete remission rate or overall survival of CD34+

and CD34– AML patients.54-60

Expression of the CD34 antigen in AML and its asso-ciation with different survival rates could be due to a

Peripheral blood stem cells in acute myeloid leukemia

Figure 1. DNA/RNA cellular content (acridine orange) ofnormal enriched CD34+ cells. Cell cycle measurements con-firms that the majority of progenitors are quiescent (G0)with only few going into cycle (G1).

22


number of factors. First of all, the cut-off point for CD34positivity that should be used to decided whether anAML sample is carrying this antigen. Since the propor-tion of CD34+ cells is around 1% of normal bone mar-row mononuclear cells and 0.01-0.1% of peripheralblood leukocytes, many authors have considered 5% asthe optimal cutoff level for classifying CD34+ AML.Nowadays, most authors agree that the cutoff point forCD34 should be 20% in order to avoid misinterpretationof the data coming from surface marker analysis. How-ever, there is no scientific basis for considering a sam-ple with 20% positive cells as positive, while anotherspecimen with 19% positivity as negative, since the lev-el of CD34 expression in a substantial number ofpatients is characterized by a continuous spectrum.53Furthermore, it must be kept in mind that the choice ofa cutoff level of 5% could give rise to erroneous results,since it can be influenced by the methods used to detectantigen expression, which are characterized by differentlevels of sensitivity and specificity. Indirect immunoflu-orescence staining is more sensitive, although less spe-cific, while the opposite is true for the direct technique.As far as the instrumentation is concerned, it must beunderlined that modern flow cytometers are highly sen-sitive in detecting surface marker positivity with respectto microscope analysis and immunoenzymatic tech-niques such as APAAP, PAP, etc. In addition, wheneverpossible immunophenotype analysis should be prefer-entially performed on fresh, not cryopreserved bonemarrow samples, and if this is not possible the number

of blasts present in the specimen analyzed should becarefully evaluated.61-65

A recent report showed that CD34 antigen expressionin AML samples having a marked heterogeneity in cellsize was found preferentially on small leukemic cellswith little or no side scatter. This feature was also asso-ciated with shorter remission duration and survival, sug-gesting that this morphological heterogeneity couldreflect a peculiar biological behavior of AML.66,67

Moreover, discrimination of blast cells from residualnormal nucleated cells is less likely to be obtained inAML cells by looking at light scattering properties (for-ward and side scatter) and expression of the CD45 anti-gen. For this reason, a multiparametric approach usingtwo-three-colour analysis is strongly recommended inorder to define the predominant leukemic population aswell as minor pathological clones or subclones. In addi-tion, CD34 positivity has to be evaluated solely on theblast population in order to avoid misinterpretation ofthe data. In fact, the percentage of blasts could varyfrom 30% to 99% in the bone marrow, and from 1 to99% in the peripheral blood.

Another point which deserves careful discussion isrepresented by the level of expression for CD34 in AML.In normal hemopoiesis, the CD34 antigen is expressedon virtually all colony forming cells (CFU) and lympho-cyte progenitors of either T or B lineage. However, with-in the progenitor cell compartment the degree of posi-tivity for CD34 decreases with cell differentiation (max-imum for multipotent cells and minimum for unipotentcells), and disappears in morphologically identifiablebone marrow precursors. Studies performed at the VInternational Workshop on Leukocyte DifferentiationAntigens (Boston, 1993) recognized three main subsetsof CD34+ normal bone marrow cells with CD34 antigendensities: low (2,000-5,000 binding sites per cell-ABC),medium (10,000-20,000 ABC), high intensities (25,000-40,000 ABC). This heterogeneity in CD34 antigen expres-sion in normal progenitors makes it difficult to use thismolecule for the monitoring of minimal residual disease(MRD) in AML patients treated with chemotherapyand/or bone marrow transplantation.68 Flow cytometryallows the recognition of a subset of CD34+ AML char-acterized by bright expression for CD34 (> 50,000 ABC),which could therefore be easily recognized even whenpresent in a very low percentage (< 0.1% of nucleatedcells). This subset represents about 20-30% of CD34+

AML, so the remaining AML patients should be checkedfor MRD by using alternative ways (strategical double ortriple staining: CD34/CD56; CD34/CD65/TdT; cytoge-netics, molecular biology, etc.).68

Another source of variability in detecting CD34+ AMLis represented by the type of CD34 monoclonal anti-body used for immunophenotypic analysis. It has beendemonstrated that at least three distinct CD34 epitopesexist, based on their differential sensitivity to enzymat-ic cleavage (using neuroaminidase, chymopapain andglycoprotease), Western blotting analysis, cell reactivi-ty studies, and cross blocking experiments.1,69-75 So far

Table 2. Possible explanations for the differences reported in the litera-ture concerning the incidence of CD34+ AML.

1. Cut-off levels for the discrimination of positive and negative cases2. Detection systems employed (flow cytometry, type of flow cytometer,

immunoenzymatic technique-APAAP, immunogold, PAP-immuno-fluorescence microscope)

3. Specimen analyzed (bone marrow, peripheral blood)4. Percentage of leukemic cells present in the sample examined5. Use of cryopreserved rather than fresh cells6. Use of different CD34 antibodies recognizing distinct

CD34 epitopes7. Percentage value and level of intensity for CD348. Light scattering properties of CD34+ cells9. Patients analyzed (de novo AML or secondary AML)10. Biologic characteristics of AML cells (chromosome aberrations,

gene abnormalities, immunophenotypic profile of CD34+ AML blasts)11. Type of chemotherapy regimen employed

M. Aglietta et al.

23


at least twenty-two CD34 monoclonal antibodies(McAbs) have been shown to recognize the CD34 mol-ecule, the most direct evidence being reactivity withcells transfected with CD34 cDNA and binding to CD34glycoprotein.72 Recently, it has been reported that anumber of AML cases are positive for some CD34 McAbsand negative for others (especially if they belong to adifferent epitope class), confirming the necessity ofusing the same CD34 McAbs in order to achieve com-parable results between different centers.71,72,74

Another point which needs to be considered whenevaluating the incidence of CD34+ AML is represented bypatient characteristics at diagnosis. The number of CD34+

cases is higher in secondary than in newly diagnosedAML. If we consider that both the biology and the clini-cal pattern of secondary AML are quite different fromthose of de novo AML, one may argue that including ofboth types of leukemias in a clinical setting could inter-fere significantly with the prognostic relevance of CD34expression on AML blast cells. In this context, the type oftreatment utilized by various authors (chemotherapy reg-imen, allogeneic and autologous bone marrow trans-plantation), which can influence the outcome of the dis-ease, is also of some relevance.

CD34+ AML are characterized by a higher incidence ofchromosomes abnormalities involving chromosomes 5(–5, 5q–), 7 (–7.7q–), and to a lesser extent chromo-somes 16 (16q), 17 (17p), 11 (trisomy 11), or multiplechromosomes at the same time, generating the so-called major karyotypic abnormalities.44,46,48,52,57,76 Recentstudies have found a close relationship between CD34expression in AML and previous exposure to chemother-apy, radiotherapy and/or pesticides.47 CD34+ AML arealso associated with trilineage myelodysplasia, dys-granulopoiesis, and/or abnormalities of the p53 tumorsuppressor gene.77

The correlation between CD34+ AML and FAB sub-types is illustrated in Table 3. On the other hand, mostbiphenotypic acute leukemias (BAL) show positivity forthe CD34 antigen.

Finally, the antigenic profile of CD34+ AML is ratherheterogeneous, depending essentially on the morpho-logical subtype and to a lesser extent on the differenti-ation stage of the leukemic clone. The large majority ofCD34+ AML coexpress a number of antigens which arenot associated with cell commitment, such as HLA-DR,CD38, CD45RO, CD45RA, CD71, and IL3 receptor. Inaddition, some surface and cytoplasmic glycoproteinsexpressed by committed myeloid cells were found to bepositive in CD34+ AML, i.e. CD33, CD13, CD117 (stemcell factor receptor), CDw116 (GM-CSF receptor), G-CSFreceptor, myeloperoxidase, lysozyme78-82 (Figures 2 and3). The stem cell- associated antigen Thy-1 (CD90) isnegative in this AML subtype, while nuclear TdT is some-times positive. CD34+ cells also express high levels of P-glycoprotein, which is the product of the multiple drugresistance (MDR) gene.

Post-remission therapy of acute myeloidleukemia and potential role of autologousstem cell transplantation

About two thirds of previously untreated adults withprimary acute myeloid leukemia enter complete remis-sion (CR) after induction therapy based on cytarabineand an anthracycline.83 However, long-term disease-freesurvival occurs in a minority of cases since most subjectsrelapse from proliferation of occult residual leukemiccells. Following conventional consolidation treatmentless than 25% of patients remain in complete remissionat four years.84,85

In order to eradicate residual AML cells and improvedisease-free survival, three approaches have beenemployed in the last ten years: (a) intensive postremis-sion chemotherapy; (b) allogeneic bone marrow trans-plantation (BMT), and (c) myeloablative conditioningregimens followed by autologous BMT as supportivetherapy.

Intensive postremission chemotherapy is essentiallybased on the use of high-dose cytarabine (1 to 3 g/m2

3 6 to 12 doses), either alone or in combination withother agents. Results of uncontrolled studies performedin the late ’80s and early ’90s (reviewed by Cassileth etal.85) indicated that intensive post-remission therapyachieves long-term disease-free survival in 25-30% ofpatients in first CR. Mayer et al.83 recently reported aprospective study aimed at evaluating the effect of theintensity of postremission chemotherapy on survival ofleukemic patients. Acute leukemia individuals in first CRwere randomly treated with four courses of cytarabineat one of three doses: 100 mg/m2 per day for 5 days bycontinuous infusion; 400 mg/m2 per day for 5 days bycontinuous infusion, or 3 g/m2 in a 3-hour infusion every


Table 3. Clinical and biological characteristics of CD34+ AML.

Incidence: 30-50% (de novo AML); 50-70% (secondary AML)History: previous exposure to chemo-radiotherapy or pesticidesCorrelation with FAB subtypes: M0,M1,M5 (70-90%); M2,M4 (20-60%); M3:

1-5%; M6, M7: 20-50%Prognosis: poor (mean survival rates less than 12-18 months)Cellular density for CD34: variable from case to case (range: 3,000- 100,000

per blast cell)Immunophenotypic profile: in most cases CD45+, HLA-DR+, CD38+, CD33+,

CD117+, Thy1– Chromosome abnormalities: -5, -7, 5q-, 7q-, 16q, 17p, major karyotype

aberrationsTherapy: to be defined (more aggressive chemotherapy regimens?)

24


12 hours on days 1, 3 and 5. In patients 60 years of ageor younger the probability of remaining disease-freeafter four years correlated with the postremissioncytarabine dose: 24% for the 100-mg group, 29% forthe 400-mg group, and 44% for the 3-g group (p =0.002). In patients older than 60 the probability ofremaining disease-free after four years was 16% or lessin each of the three postremission cytarabine groups,with no significant difference between groups. It shouldbe noted that a significant proportion of AML patientsachieving CR cannot proceed to intensive chemothera-py due to persistent bone marrow aplasia.

Allogeneic bone marrow transplantation offers manyadvantages, including the graft-versus-leukemia effect;however, the availability of a histocompatible siblingdonor is restricted to approximately 25% of potentialcandidates. Published studies report disease-free sur-vival rates at four years ranging from 45 to 58%.86-88

These figures should be considered with caution sincethey are biased by the exclusion of patients whorelapsed before allogeneic BMT. The new approachrecently described by Aversa et al.,89 i.e. a strongimmunosuppressive and myeloablative conditioningregimen followed by transplantation of a large numberof haploidentical stem cells depleted of T lymphocytes,may open new perspectives for allogeneic bone marrowtransplantation in AML patients. In this setting, themobilization and collection of allogeneic PBSC is cru-cial to overcoming HLA-disparity.89

Pilot studies on the use of autologous BMT as postre-mission therapy90,91 indicate that disease-free survival atfour years is on the order of 50% (i.e. comparable to thatof allogeneic BMT). Autologous BMT has the advantageof lower procedure-related mortality than allogeneic BMT(approximately 10-15%), but involves a high risk ofleukemic relapse (about one half treatment failures). Arecent trial by Zittoun et al.92 showed that both autolo-gous and allogeneic BMT performed in first CR resultedin a significantly better disease-free survival than inten-sive consolidation chemotherapy. The projected rate atfour years was 55% for allogeneic BMT, 48% for autol-ogous BMT and 30% for intensive chemotherapy with nodifferences between allogeneic and autologous BMT.

Taken together, these results demonstrate the poten-tial benefit of autologous stem cell transplantation forleukemic patients. However, the high incidence ofleukemic relapse and delayed hematological recoveryafter ABMT have prompted several authors to investi-gate the use of mobilized PBSC.

CD34+ mobilized hematopoietic cells forsupport of intensive postremissionchemotherapy of AML

As stated above, autologous transplantation of mobi-lized hematopoietic progenitor cells has been shown toreconstitute hematopoiesis more efficiently than BM-derived grafts.2 Moreover, early studies have failed todetect neoplastic cells in the peripheral blood of AMLpatients during the early recovery phase following

Figure 2. Dual color fluorescence analysis with a flow cytometer(CD34/FITC; c-kit-CD117-PE) in a patient with AML FAB M4. Bone mar-row sample shows a blast percentage of 84%. A) Light scattering properties of blast cells (forward scatter= cell vol-

ume; side scatter= cell granularity).B)Contour plot diagram showing 83% of CD34+ cells, 71% of cccc----kkkkiiiitttt+

cells, and 68.5% of cells co-expressing CD34 and cccc----kkkkiiiitttt.C)Histogram distribution of cells stained for CD34 monoclonal anti-

body.D)Histogram distribution of cells labelled with CD117 monoclonal anti-

body.

Figure 3. Discordant reactivity between different epitope class anti-bodies is shown in a patient with AML M1 (% blasts = 90%); 36.5% ofblasts are positive for class I, 0.1% for class II, and 47% for class III.The intensity of fluorescence varied significantly from antibody to anti-body even within the same epitope class.

M. Aglietta et al.

25


remission induction/consolidation chemotherapy.93,94

Based on these reports, several investigators haveaddressed the question of whether the use of PBSCmight result in a more rapid engraftment and a lowerrisk of relapse in AML patients. Relevant issues includethe level of malignant cell contamination, the thresh-old dose of hematopoietic precursors (i.e. CFU-GM andCD34+ cells), the optimal timing for stem cell collectionand the potential benefit derived from the use of select-ed cytokines to improve PBSC harvest.

To et al.93 studied leukemia-associated cytogeneticabnormalities in myeloid colonies derived from earlyremission PBSC. At a sensitivity level of 2:100 cells, theywere not able to detect the t(8;21) in 293 samplesexamined. Recently, the more sensitive nested reversetranscriptase polymerase chain reaction (RT-PCR) wasused to monitor minimal residual disease in the BM andperipheral blood of leukemic patients considered inremission by morphologic analysis.95 In that study, theauthors found no differences between PBSC collectionsand simultaneous BM harvests. However, the degree ofleukemic contamination may have been differentamong the PCR-positive groups because the two-stepPCR is highly sensitive but not quantitative.

The issue of leukemia-free autograft was recentlyunderscored by gene-marking studies showing thatresidual contaminating AML cells contribute to relapsewhen reinfused into patients.96 In this regard, leukemicrecurrence remains the most frequent cause of treat-ment failure in AML patients97,98 and preliminary non-randomized clinical studies have not reported anyadvantage for ABMT over PBSC99-101 in terms of disease-free survival and overall survival rate (Table 4). More-over, it could be argued that the interval between com-plete remission and myeloablative therapy may beshorter for PBSC patients, since the exclusion rate ishigher for ABMT patients.101 Thus, randomized studiesare warranted to rule out selection bias. Reinfusion ofPBSC markedly shortened the period of marrow aplasiacompared to purged and unpurged ABMT99-102 andreduced morbidity and resource utilization (Table 4). Inparticular, To et al.102 demonstrated an advantage of 11and 19 days in the median time to achieve 0.5×109 neu-trophils/L and 50×109 platelets/L, respectively, in favorof PBSC, whereas two other studies99,100 showed a morerapid neutrophil engraftment (28 days and 12 days,respectively) but not a highly significant faster plateletrecovery. Most likely the acceleration of hematopoiet-ic reconstitution derives from the reinfusion of a high-er number of early pluripotent precursors and largeamounts of committed progenitor cells which requireless time to reach maturation.103

Early studies in acute leukemia indicated that an opti-mal CFU-GM dose of 50×104/kg body weight is requiredfor complete and sustained reconstitution of BM func-tion.104 Other reports99-101 and our own preliminary expe-rience (Tables 4 and 5) have demonstrated similarresults with lower CFU-GM numbers. These differencesare probably related to different assay methods, where-

as the more reproducible evaluation of progenitor cellsby flow cytometry has suggested a threshold dose of2×106 CD34+ cells/kg.105 Interestingly, the number ofprogenitor cells infused is only indirectly related to long-term engraftment, suggesting that additional measure-ments of CD34+ cell fraction subsets may be helpful inpredicting sustained recovery of hematopoiesis. More-over, a transient secondary fall in neutrophil andplatelet count has been described104 during the 3rd-8th

weeks after transplantation, indicating a time lagbetween exhaustion of the committed progenitor cellpool, which is responsible for early engraftment, andexpansion of the more immature pluripotent stem cellcompartment.

Lastly, the timing of BM harvest or PBSC collection inthe autologous setting plays an important role in thequality of the autograft. It has been clearly establishedthat the amount and the length of previous therapyaffects the number of circulating CD34+ cells;106 how-ever, reinfusion of PBSC collected in the recovery phaseof induction therapy has resulted in a higher relapserate101,107 indicating that returning a larger quantity ofcells to patients without careful analysis of minimalresidual disease may increase the probability of trans-plantation of leukemic cells. Thus, at least one consol-idation cycle of treatment should be performed in the


Table 4. Reported studies on autologous stem cell transplantation in AMLpatients. PBSC indicates peripheral stem cell transplantation, while ABMTrefers to autologous bone marrow transplantation.

PBSC/ABMT PBSC/ABMT PBSC/ABMT

Reference # 99 102 100Pts 28/683 38/13° 20/23*CFU-GM reinfused NR 86.6/12.1** 2.3/0.1(x104/Kg) (0.2-4.1/0-1)Median time to:> 0.5 x 109 PMN/L 15.5/27 11/22 14/42(9-60/9-389) (9-17/12-35) NR> 50 x 109 PLT/L 58.5/50 13.5/32 30/46(11-713/10-700) (9-NA/21-NA) NRPLT transfusion NR 5.4/8.8** NRDays on antibiotics NR 9.3/14.3** NRHospitalization NR 27.5/35.1** 45/73(days)Relapse rate 57%/48% NR NRDFS 39%/42% NR 35%/51%° AML pts = 19 in PBSC group and 1 in ABMT group.* Comparison was made with 23 pts receiving purged marrow. ** Results expressed asthe mean. Abbreviations: NR, not reported; NA, not achieved; DFS, disease free survival;PMN, neutrophil; PLT, platelet.

26


case of stem cell mobilization to take advantage of invivo purging, coupled with a low number of apheresisprocedures. To this end, it has been shown108 thatadministration of G-CSF during the recovery phase ofconsolidation chemotherapy in acute leukemia patientsincreased the peak level of CFU-GM and CFU-MIX by 5.8and 4.3 times, respectively, compared to cycles were G-CSF was not used, and significantly prolonged the peri-od of mobilization of stem cells. Although the role ofcytokine treatment in AML patients is still under eval-uation, preliminary data from a large cohort of leukemicpatients suggest that G-CSF administration does notaffect either the remission or relapse rate,109,110 where-as a protective effect regarding relapse was shown in arandomized study.29

In practice, leukapheresis sessions should be startedafter a careful evaluation of CD34+ cells in the periph-eral blood by flow cytometry, at time of hematopoieticrecovery from transient myelosuppression.

When adequate mobilization of progenitors (CD34+

cells > 10-15/µL) occurs, daily leukaphereses should beperformed until the collection of a minimum number of2×106/kg CD34+ cells. The leukapheresis products shouldbe evaluated for the presence of residual contaminat-ing leukemic cells by immunophenotyping, karyotypicanalysis and RT-PCR-based molecular analysis in thosesamples deriving from patients who had shown a spe-cific phenotypic and/or molecular marker at diagnosis.

Moreover, because the content of circulating progen-itors is generally low (< 1% of the mononuclear cellfraction), blood cell separator efficiency must be opti-mized. Collection efficiency (CE) is the percentage ofcells entering the system that are eventually collected:

CE (%) =No. harvested cells

× processed blood volumeNo. of cells in preapheresis

blood unit volume

Acceptable CE should not be lower than 50%. CE is auseful parameter for evaluating blood cell separatoreffectiveness in harvesting PBSC, independently of thepatient’s clinical condition.

Purging in AMLConsidering the possibility of relapse from minimal

residual disease (MRD) derived from autologous graft,several investigators addressed the issue of ex vivo purg-ing of leukemic cells prior to stem cell reinfusion. Usingthe Brown Norway myelocytic leukemia rat model, ithas been shown that injection of 25 leukemic cellsinduces leukemia in 50% of recipients.111 By applyingthe same mathematical model to humans, it has beensuggested that reinfusion of 10,000 residual leukemiccells may result in a relapse rate as high as 50%.111 Morerecently, the role of residual tumor cells in clinicalrelapse after autograft was indicated by a clinical study

involving 114 B-cell lymphoma patients with t(14;18)who received autologous marrow treated with a com-bination of monoclonal antibodies directed against B-cell associated antigens plus complement.112 Followingpurging, no lymphoma cells could be detected by PCRamplification of the bcl-2 gene in the marrow of 57patients. Disease-free survival was increased in theseindividuals with respect to that of patients whose mar-row contained detectable tumor cells. Moreover, theability to remove lymphoma cells was the most impor-tant prognostic factor for predicting relapse (39% ver-sus 5% of purged patients after a median follow-up of23 months).112 Lastly, genetic marking of marrow cellswith the neomycin-resistance gene has provided theformal proof that reinfusion of residual leukemic cells inAML patients contributes to a recurrence of the dis-ease.113 Taken together, these findings demonstrate theneed for effective ex vivo treatments to improve theoutcome of autologous transplantation.

Among the many purging methods proposed for theelimination of MRD,114 cyclophosphamide (Cy) deriva-tives are the most widely used agents in AML patients,since preclinical models demonstrated that these com-pounds were able to eliminate residual BM neoplasticcells in the Brown Norway rat system.115 The main mech-anism of action of Cy active metabolites is based uponmarked inhibition of leukemic progenitor cell (CFU-L)growth115 while sparing normal primitive hematopoieticcells.116 Furthermore, these alkylating agents seem toinduce apoptotic death of leukemic cells117 as well as theactivation of immune mechanisms capable of control-ling leukemic cell proliferation.118 Combinations of 4-hydroperoxycyclophosphamide (4-HC) or nitrogen mus-tard and VP-16 have also been proposed to increase theselective toxicity of pharmacologic purging towardsneoplastic cells.119 Several monoclonal antibodies that

M. Aglietta et al.

Table 5. Experience of the Institute of Hematology “Seragnoli”, Bolognaon autologous stem cell transplantation in AML patients. The results areexpressed as median (range) and refer to AML (n=7) and RAEB-T (n=2)patients in I CR. PBSC collections were carried out following consolida-tion treatment. PMN and PLT recovery was recorded as such when the PBcount was > 0.5 and 20 x 109/L, respectively.

Apheresis productsPts PBSC MNC CFU-GM CD34+

collections (108/Kg) (104/Kg) (106/Kg)9 3 (2-3) 7 (2.8-11.6) 11.8 (2.8-78.2) 7 (3.1-17.5)Hematological reconstitutionPts day to PMN day to PLT PLT RBC

Hospitalrecovery recovery transfusion transfusion

discharge8 14 (11-34) 18 (10-NR) 2.5 (0->10) 4 (1-9) 18 (14-38)Abbreviations: MNC, mononuclear cells; RBC, red blood cells.

27


recognize tumor-associated or cell-differentiation anti-gens not expressed by primitive cells responsible forhematopoietic engraftment have been selected for clin-ical trials after in vitro studies demonstrated a highpurging efficacy with the use of complement, toxins andradioactive molecules. However, the heterogeneity ofantigen expression on neoplastic cells and the lack oftumor-specific determinants may greatly affect the effi-ciency of antibody-based strategies for depletion ofleukemic cells. In this context, the combination of twopurging techniques has been explored with promisingresults.120 Other approaches include the use of pho-toactive compounds which sensitize leukemic cells andspecifically damage their cell membranes upon exposureto light,121 and biological methods based on the differ-ent proliferative patterns of leukemic cells and their nor-mal counterparts when cultured ex vivo for several daysin the presence of stromal cells.122

Clinical trialsClinical retrospective data supporting the beneficial

effect of purging have been progressively accumulating.The Baltimore team provided indirect evidence in favorof purging by correlating effective CFU-GM colony elim-ination with a significant decrease in relapse.123 Fur-thermore, the same authors associated the sensitivity to4-HC of CFU-L grown in remission with the posttrans-plant outcome.124 The 3 most recent surveys of theLeukemia Working Party of the EBMT group have con-sistently reported lower relapse rates following reinfu-sion of BM purged with mafosfamide,91,125,126 especiallyin patients transplanted within 6 months of CR and inslow remitters (> 40 days to achievement of CR), twopatient populations considered at high risk of diseaserecurrence. In fact, the proportion of patients relapsingin the purged and unpurged groups was 29±5% vs50±4%, respectively, following a conditioning regimenthat included total body irradiation (p < 0.0001). Morestriking differences were found when considering onlythose patients autografted early after CR (16±6% vs60±6%) and patients with an interval from diagnosis toCR greater than 40 days (20±8% vs 61±6%). Gulati etal. and Laporte et al.127, 128 reported disease free survivalwhich approximated 60% in AML patients in I CR rein-fused with autologous marrow treated with 4-HC andVP-16 and mafosfamide. In the same paper by the Parisgroup,151 it was suggested that the higher the initialcontent of BM CFU-GM, the lower the risk of transplantrelated mortality and the higher the chance of curingthe disease.

Despite these results in favor of ex vivo elimination ofcontaminating leukemic cells, this procedure is not rou-tinely performed in the majority of transplant centers.The main reasons might be: 1) the delay in hematolog-ical recovery after reinfusion of purged autografts; 2)the increase in the cost of ABMT; 3) the need for tech-nical training and, most of all, 4) the lack of resultsderived from prospective clinical studies demonstrating

the effects of purging. The feasibility of such trials israther limited due to the high number of patients need-ed to obtain adequate statistical power.

So far, no data are available on purging protocols forPBSC collections; however, there are several reasons forproposing purging strategies for autograft of circulat-ing autologous stem cells. Unlike solid tumors andmalignant lymphomas, acute leukemias easily involvePB. Moreover, the number of hematopoietic progenitorcells (and possibly leukemic precursors) in PB auto-transplants is usually at least 10 times higher than thatof ABMT. Finally, as discussed above, the intervalbetween CR and autotransplant may be shorter for PBSCpatients, who may be thus considered high risk patientsfor relapse. Critical issues for designing experimentalstudies in this setting would include the proper assess-ment of MRD before and after purging, the establish-ment of reproducible technical protocols (cell concen-tration, RBC contamination, etc.), and careful evaluationof the toxicity of purging agents on PB progenitors, sincethe kinetic status of circulating stem cells followingmobilization protocols (especially if CSFs are used) maybe different from BM stem cells.129

ConclusionsAutologous BMT has been widely used as consolida-

tion therapy in AML patients in first or second remission;however, delayed hematopoietic engraftment occurs ina substantial proportion of patients resulting in signif-icant morbidity and mortality. This is mainly due to theadverse effects of prior intensive chemotherapy on BMharvest, a decrease in the normal stem cell pool inleukemic patients and, perhaps, toxic damage to themarrow microenvironment. Thus, several groups haveinvestigated the use of circulating progenitor cells withthe twofold aim of reducing transplant-related toxicityand widening the number of potential candidates formyeloablative therapy with the support of autologousstem cells.

As for hematopoietic reconstitution, previous studieshave provided evidence that PBSC transplantation mayoffer some advantages over BM autografting. However,crucial issues such as asynchronous mobilization of nor-mal vs leukemic cells and potential contamination ofPBSC collections, timing of PBSC harvest, detection ofminimal residual disease, and the role of growth factorsto accelerate hematological recovery and optimize stemcell collection have not been fully addressed.

In the present paper, the latest advances in this fieldhave been reviewed with special focus on the biology ofputative leukemic stem cells; operative guidelines havealso been provided for those investigators who wish todesign proper clinical trials on PBSC autotransplantationin acute leukemia.

Definitive answers regarding the role of PBSC will becoming from a large European randomized trial whichis currently comparing peripheral blood stem cell andBM-derived graft.


28


References

1. Carlo Stella C, Cazzola M, De Fabritiis P, et al. CD34-positive cells: biology and clinical applications.Haematologica 1995; 80:367-87.

2. Siena S, Bregni M, Di Nicola M, et al. Durability ofhematopoiesis following autografting with peripheralblood hematopoietic progenitors. Ann Oncol 1994;5:935-41.

3. To LB. Is our current strategy in manipulating hemo-poiesis in autologous transplantation correct? StemCells 1993; 11: 283-9.

4. Majolino I, Aversa F, Bacigalupo A, Bandini G, ArceseW, Reali G. Allogeneic transplants of the rhG-CSF-mobilized peripheral blood stem cells (PBSC) fromnormal donors. Haematologica 1995; 80:40-3.

5. Yin M, Silvestri FF, Banavali SD, et al. Clonogenicpotential of myeloid leukemia cells in vitro is restrict-ed to leukemia cells expressing the CD34 antigen. EurJ Cancer 1993; 29A: 2279-83.

6. Carlo-Stella C, Mangoni L, Almici C, Frassoni F, FiersW, Rizzoli V. Growth of CD34+ acute myeloblasticleukemia colony-forming cells in response to recom-binant hemopoietic growth factors. Leukemia 1990;4:561-6.

7. Lapidot T, Sirard C, Vormoor J, et al. A cell initiatinghuman acute myeloid leukaemia after transplantationinto SCID mice. Nature 1994; 367:645-8.

8. Ikeda H, Kanakura Y, Tamaki T, et al. Expression andfunctional role of the proto-oncogene c-kit in acutemyeloblastic leukemia cells. Blood 1991; 78:2962-8.

9. Broudy VC, Smith FO, Lin N, Zsebo K, Egrie J, BersteinID. Blasts from patients with acute myelogenousleukemia express functional receptors for stem cell fac-tor. Blood 1992; 80:60-7.

10. Motoji T, Watanabe M, Uzumaki H, et al. Granulo-cyte colony-stimulating factor (G-CSF) receptors onacute myeloblastic leukemia cells and their relation-ship with the proliferative response to G-CSF in clono-genic assay. Br J Haematol 1991; 77:54-9.

11. Budel LM, Touw IP, Delwel R, Clark SC, Lowenberg B.Interleukin-3 and granulocyte-monocyte colony stim-ulating factor receptors on acute myelocytic leukemiacells and relationship to the proliferative response.Blood 1989; 74:565-71.

12. Pietsch T, Kyas U, Steffens U, et al. Effects of humanstem factor (c-kit ligand) on proliferation of myeloidleukemia cells: heterogeneity in response and synergywith other hematopoietic growth factors. Blood 1992;80:1199-206.

13. Piacibello W, Sanavio F, Bresso P, et al. Stem cell fac-tor improvement of proliferation and maintenance ofhemopoietic progenitors in myelodysplastic syn-drome. Leukemia 1994; 8:250-7.

14. Carlo Stella C, Rizzoli V. Stem cells and stem cell fac-tor(s). Haematologica 1995; 80:1-4.

15. Miyajima A, Mui ALF, Ogorochi T, Sakamaki K.Receptors for granulocyte-macrophage colony stimu-lating factor, interleukin-3 and interleukin-5. Blood1993; 82:1960-74.

16. Lowenberg B, Touw IP. Hemopoietic growth factorsand their receptors in acute leukemia. Blood 1993;81:281-92.

17. Mui ALF, Miyajima A. Interleukin-3 and granulocyte-macrophage colony-stimulating factor receptor sig-nal transduction. Proc Soc Exp Biol Med 1994; 206:284-8.

18. Begley CG, Metcalf D, Nicola NA. Primary humanmyeloid leukemia cells: comparative responsivenessto proliferative stimulation by GM-CSF or G-CSF and

membrane expression of CSF receptors. Leukemia1987; 1:1-8.

19. Aglietta M, DeFelice L, Stacchini A, et al. Effect ofhemopoietic growth factors on the proliferation ofacute myeloid and lymphoid leukemias. Leuk Lym-phoma 1990; 2:207-14.

20. Damiani D, Michieli M, Revignas MG, et al. Prolifer-ation and maturation effects of in vivo granulocyte-macrophage colony stimulating factor in acute non-lymphocyte leukemia. Haematologica 1992; 77: 25-9.

21. Kondo S, Okamma S, Asano Y, Harada M, Niko Y.Human granulocyte colony stimulating factor recep-tors in acute myelogenous leukemia. Eur J Haematol1991; 46:223-30.

22. Sachs L, Lotem J. Control of programmed cell deathin normal and leukemic cells: new implications fortherapy. Blood 1993; 82:15-21.

23. Aglietta M, De Felice L, Stacchini A, et al. In vivo effectof granulocyte-macrophage colony-stimulating factoron the kinetics of human acute myeloid leukemia cells.Leukemia 1991; 5:979-84.

24. Bettelheim P, Valent P, Andreeff M, et al. Recombinathuman granulocyte-macrophage colony stimulatingfactor in combination with standard inductionchemotherapy in de novo acute myeloid leukemia.Blood 1991; 77:700-11.

25. Estey E, Thall P, Andreeff M, et al. Use of granulocytecolony-stimulating factor before, during and after flu-darabine plus cytarabine induction therapy of newlydiagnosed acute myelogenous leukemia or myelodys-plastic syndromes: comparison with fludarabine pluscytarabine without granulocyte colony-stimulatingfactor. J Clin Oncol 1994; 12:671-8.

26. Ohno R, Naoe T, Kanamaru A, et al. A double-blindcontrolled study of granulocyte colony-stimulatingfactor started two days before induction chemother-apy in refractory acute myeloid leukemia. Blood 1994;83:2066-92.

27. Giralt S, Escudier S, Kantarjian H, et al. Preliminaryresults of treatment with filgrastim for relapse ofleukemia and myelodysplasia after allogenic bonemarrow transplantation. N Engl J Med 1993; 329:757-61.

28. Stone RM, Berg DT, George SL, et al. Granulocyte-macrophage colony stimulating factor after initialchemotherapy for elderly patients with primary acutemyelogenous leukmia. N Engl J Med 1995; 332:1671-7.

29. Dombret H, Chastang C, Fenaux P, et al. A controlledstudy of recombinant human granulocyte colony stim-ulating factor in elderly patients after treatment foracute myelogenous leukemia. N Engl J Med 1995;332:1678-83.

30. Lajtha LG. On the concept of the cell cycle. J CellComp Physiol 1963; 62:143-5.

31. Hodgson GS, Bradley TR. Properties of haematopoi-etic stem cells surviving 5-fluorouracil treatment: evi-dence for a pre-CFU-S cell? Nature 1979; 281:381-2.

32. Leary AG, Hirai Y, Kishimoto T, Clark SG, Ogawa M.Survival of hemopoietic progenitors in G0 does notrequire early hemopoietic regulators. Proc Natl AcadSci USA 1989; 86:4535-9.

33. Till JE, McCulloch EA, Siminovitch L. A stochasticmodel of stem cell proliferation, based on the growthof spleen colony-forming cells. Proc Natl Acad Sci USA1964; 51:29-34.

34. Minden MD, Till JE, McCulloch EA. Proliferative stateof blast cell progenitors in acute myeloblasticleukemia (AML). Blood 1978; 52:592-6.

35. Clarkson B, Fried J, Striff A, Saki Y, Ota K, Ohkita T.Studies of cellular proliferation in human leukemia.III. Behavior of leukemic cells in three adults with acute

M. Aglietta et al.

29


leukemia given continuous infusions of 3H-thymidinefor 8-10 days. Cancer 1970; 25:1237-60.

36. Gavosto F, Pileri A, Gabutti V, Masera P. Non self-maintaining kinetics of proliferating blasts in humanacute leukaemia. Nature 1967; 216:188-9.

37. Andreeff M, Darzynkiewicz Z, Sharpless TK, ClarksonBD, Melamed MR. Discrimination of humanleukemia subtypes by flow cytometric analysis of cel-lular DNA and RNA. Blood 1980; 55:282-93.

38. Raza A, Maheshwari Y, Preisler HD. Differences in cellcycle characteristics among patients with acute non-lymphocytic leukemia. Blood 1987; 69:1647-53.

39. Lista P, Brizzi MF, Rossi M, Resegotti L, Clark SC,Pegoraro L. Different sensitivity of normal andleukaemic progenitor cells to Ara-C and IL-3 com-bined treatment. Br J Haematol 1990; 76:21-6.

40. Tafuri A, Andreeff M. Kinetic rationale for cytokine-induced recruitment of myloblastic leukemia followedby cycle-specific chemotherapy in vitro. Leukemia1990;12:826-34.

41. Raymakers R, Wittebol S, Pennings A, Linders E, Pod-dighe P, De Witte T. Residual normal, highly prolifer-ative progenitors can be isolated from the CD34+/33–fraction of AML with a more differentiated phenotype(CD33+). Leukemia 1995; 9:450-7.

42. Gore SD, Amin S, Weng LJ, Civin C. Steel factor sup-ports the cycling of isolated human CD34+ cells inthe absence of other growth factors. Exp Hematol1995; 23:413-21.

43. Hunter T, Pines J. Cyclins and Cancer II: cyclin D andCDK inhibitors come of age. Cell 1994; 79:573-82.

44. Vaughan W, Civin C, Welsenburger D, et al. Acuteleukemia expressing the normal human haematopoi-etic stem cell membrane glycoprotein CD34+ (MY10).Leukemia 1988; 2:661-6.

45. Campos L, Guyotat D, Archimbaud E, et al. Surfacemarker expression in adult acute myeloid leukemia:correlations with initial characteristics, morphologyand response to therapy. Br J Haematol 1989; 72:161-6.

46. Borowitz M, Gockerman J, Moore J, et al. Clinico-pathologic and cytogenetic features of CD34+(MY10) positive acute nonlymphocytic leukemia. AmJ Clin Pathol 1989; 91:265-70.

47. Geller RB, Zahurak M, Hurwitz C. Prognostic impor-tance of immunophenotyping in adults with acutemyelocytic leukemia: the significance of the stem cellglycoprotein CD34+ (MY10). Br J Haematol 1990;76:340-7.

48. Guinot M, Sanz G F, Sempere A, et al. Prognostic val-ue of CD34 expression in de novo acute myeloblasticleukemia. Br J Haematol 1991; 78:533-4.

49. Lee E, Yang J, Leavitt R. The significance of CD34+and TdT determinations in patients with untreatedde novo acute myeloid leukemia. Leukemia 1992; 6:1203-9.

50. Solary E, Casasnovas R, Campos L and the Grouped’Etude Immunologique des Leucemies. Surfacemarkers in adult acute myeloblastic leukemia: corre-lation of CD19+, CD34+ and CD14+/DR+ pheno-type with shorter survival. Leukemia 1992; 6:93-9.

51. Myint H, Lucie N. The prognostic significance of theCD34+ antigen in acute myeloid leukemia. Leuk Lym-phoma 1992; 7:425-9.

52. Fagioli F, Cuneo A, Carli M, et al. Chromosome aber-rations in CD34+ positive acute myeloid leukemia:correlations with clinicopathologic features. CancerGenet Cytogenet 1993; 71:119-24.

53. Lanza F, Rigolin M, Moretti S, Latorraca A, CastoldiG. Prognostic value of immunophenotypic character-istics of blast cells in acute myeloid leukemia. Leuk

Lymphoma 1994; 13:81-5.54. Smith FO, Lampkin B, Versteeg C, et al. Expression of

lymphoid-associated cell surface antigens by child-hood acute myeloid leukemia cells lacks prognosticsignificance. Blood 1992; 79:2415-22.

55. Kuerbit SJ, Civin CI, Krisher JP, et al. Expression ofmyeloid-associated and lymphoid-associated cell sur-face antigens in acute myeloid leukemia of childhood:a pediatric oncology group study. J Clin Oncol 1992;10:1419-29.

56. Selleri C, Notaro R, Catalano L, Fontana R, Del Vec-chio L, Rotoli B. Prognostic irrelevance of CD34+ inacute myeloid leukemia. Br J Haematol 1992; 82:479-82.

57. Sperling C, Buchner T, Sauerland C, Fonatsch C, ThielE, Ludwig W. CD34 expression in de novo acutemyeloid leukemia. Br J Haematol 1993; 85:635-7.

58. Ciolli S, Leoni F, Caporale R, Pascarella A, Salti F,Rossi-Ferrini P. CD34+ expression fails to predict theoutcome in adult acute myeloid leukemia. Haemato-logica 1993; 78:151-5.

59. Del Poeta G, Stasi R, Venditti A, et al. Prognostic val-ue of cell marker analysis in de novo acute myeloidleukemia. Leukemia 1994; 8:388-94.

60. Sperling C, Buchner T, Creutzig U. Clinical, morpho-logic, cytogenetic and prognostic implications ofCD34+ expression in childhood and adult de novoAML. Leuk Lymphoma 1995; 17:417-26.

61. Terstappen L, Safford M, Konemann S, et al. Flowcytometric characterization of acute myeloidleukemia. Part II. Phenotypic heterogeneity at diag-nosis. Leukemia 1991; 9:757-67.

62. van’t Veer M, Kluin-Nelemans J, van Der Schoot C,van Putten W, Adriaansen H, van Wering E. Qualityassessment of immunological marker analysis and theimmunological diagnosis in leukemia and lymphoma:a multi-centre study. Br J Haematol 1992; 80:458-65.

63. Brando B, Sommaruga E. Nationwide quality controltrial on lymphocyte immunophenotyping and flowcytometer performance in Italy. Cytometry 1993; 14:294-306.

64. Castoldi G, Liso V, Fenu S, Vegna L, Mandelli F.Reproducibility of the morphological diagnostic cri-teria for acute myeloid leukemia: the GIMEMA groupexperience. Ann Haematol 1993; 66:171-4.

65. Drexler H, Eckhard T, Wolf-Dieter L. Acute myeloidleukemias expressing lymphoid-associated antigens:diagnostic incidence and prognostic significance.Leukemia 1993; 7:489-98.

66. Kawada H, Ichikawa Y, Watanabe S, Nagao T, Ari-mori S. Flow cytometric analysis of cell surface anti-gen expressions on acute myeloid leukemia cell pop-ulations according to their cell size. Leuk Res 1994;18:29-35.

67. Syrjala M, Ruutu T, Jansson SE. A flow cytometricassay of CD34-positive cell populations in the bonemarrow. Br J Haematol 1994; 88:679-84.

68. Campana D, Pui CH. Detection of minimal residualdisease in acute leukemia: methodologic advancesand clinical significance. Blood 1995; 85:1416-34.

69. Greaves MF, Brown J, Molgaard HV, et al. Molecularfeatures of CD34: a hemopoietic progenitor cell-asso-ciated molecule. Leukemia 1992; 6:31-6.

70. Sutherland DR, Marsh JCW, Davidson J, Baker MA,Keating A, Mellors A. Differential sensitivity of CD34epitopes to cleavage by Pasteurella haemolytica gly-coprotease: implications for purification of CD34-positive progenitor cells. Exp Hematol 1992; 20:590-9.

71. Egeland T, Gaudernack G. Immunomagnetic isola-


30


tion of CD34+ cells: methodology and monoclonalantibodies. In: Wunder S, Serke S, Solovat H, HenonP, eds. Hematopoietic stem cell: the Mulhouse man-ual. Dayton: AlphaMed Press, 1994:141-8.

72. Greaves MF, Colman SM, Böhring HJ, et al. Report onthe CD34 cluster workshop. In: Schlossman SF,Boumsell L, Gilks W et al, eds. Leucocyte typing V:White cell differentiation antigens. Oxford: OxfordUniversity Press, 1995:840-7.

73. Silvestri F, Banavali S, Baccarani M, Preisler HD. Prog-enitor cell associated antigen CD34: biology and clin-ical applications. Haematologica 1992; 77:265-72.

74. Lanza F, Castoldi G. Large scale enrichment of CD34+cells by percoll density gradients. A CML-based studydesign. In: Wunder E, Serke S, Sovalat H, Henon P,eds. Hematopoietic stem cells: the Mulhouse manu-al. Dayton: Alpha Med Press, 1994:255-70.

76. Slovak ML, Traweek ST, Willman CL, et al. Trisomy11: an association with stem/progenitor cell immuno-phenotype. Br J Haematol 1995;90:266-73.

77. Lai J, Preudhomme C, Zandecki M, et al. Myelodys-plastic syndromes and acute myeloid leukemia with17p deletion. An entity characterized by specific dys-granulopoiesis and a high incidence of p53 mutations.Leukemia 1995; 9:370-81.

78. Pierelli L, Teofili L, Menichella G, et al. Further inves-tigations on the expression of HLA-DR+, CD33+ andCD13+ surface antigens in purified bone marrow andperipheral blood CD34+ haematopoietic progenitorcells. Br J Haematol 1993; 84:24-30.

79 . Craig W, Kay R, Cutler R, Lansdorp P. Expression ofTHY-1 on human haematopoietic progenitor cells. JExp Med 1993; 177:1331-42.

80. Holyoake T, Alcorn M. CD34+ positive hemopoieticcells: biology and clinical applications. Blood Rev1994; 8:113-24.

81. Lanza F, Moretti S, Castagnari B, et al. CD34+leukemic cells assessed by different CD34 monoclon-al antibodies. Leuk Lymphoma 1995 (in press).

82. Meckenstock G, Heyll A, Schneider E, et al. Acuteleukemia coexpressing myeloid, B- and T-lineage asso-ciated markers: multiparameter analysis of criteriadefining lineage commitment and maturational stagein a case of undifferentiated leukemia. Leukemia1995; 9:260-4.

83. Mayer RJ, Davis RB, Schiffer CA, et al. Intensivepostremission chemotherapy in adults with acutemyeloid leukemia. N Engl J Med 1994; 331:896-903.

84. Cassileth PA, Linch E, Hines JD, et al. Varying intensi-ty of postremission therapy in acute myeloid leukemia.Blood 1992; 79:1924-30.

85. Cassileth PA, Nowell PC, Larson RA. Acute leukemia.In: McArthur JR, Schrier S, eds. Hematology - 1993.St. Louis: The American Society of Hematology,1993:38-48.

86. Clift RA, Buckner CD, Appelbaum FR, et al. Allogeneicbone marrow transplantation in patients with acutemyeloid leukemia in first remission: a randomized tri-al of two irradiation regimens. Blood 1990; 76:1867-71.

87. Young JW, Papadopoulos EB, Cunningham I, et al. T-cell-depleted allogeneic bone marrow transplantationin adults with acute nonlymphocytic leukemia in firstremission. Blood 1992; 79:3380-7.

88. Bortin MM, Horowitz MM, Rowlings PA, et al. 1993Progress report from the International Bone MarrowTransplant Registry. Bone Marrow Transplant 1993;12:97-104.

89. Aversa F, Tabilio A, Terenzi A, et al. Successful engraft-ment of T-cell-depleted haploidentical three-lociincompatible transplants in leukemia patients byaddition of recombinant human granulocyte colony-

stimulating factor-mobilized peripheral blood prog-enitor cell to bone marrow inoculum. Blood 1994;84:3948-55.

90. Cassileth PA, Andersen J, Lazarus HM, et al. Autolo-gous bone marrow transplant in acute myeloidleukemia in first remission. J Clin Oncol 1993; 11:314-9.

91. Gorin NC, Labopin M, Meloni G, et al. Autologousbone marrow transplantation for acute myeloblasticleukemia in Europe: further evidence of the role ofmarrow purging by mafosfamide. Leukemia 1991;5:896-904.

92. Zittoun RA, Mandelli F, Willemze R, et al. Autologousor allogeneic bone marrow transplantation comparedwith intensive chemotherapy in acute myelogenousleukemia. N Engl J Med 1995; 332:217-23.

93. To LB, Russel J, Moore S, et al. Residual leukemia can-not be detected in very early remission peripheralblood stem cell collection in acute non-lymphoblas-tic leukemia. Leuk Res 1987; 11:327-9.

94. Juttner CA, To LB, Haylock DN, et al. Circulatingautologous stem cells collected in very early remissionfrom acute non-lymphoblastic leukemia produceprompt but incomplete hemopoietic reconstitutionafter high dose melphalan or supralethal chemora-diotherapy. Br J Haematol 1985; 61:739-45.

95. Nagafuji K, Harada M, Takamatsu Y, et al. Evaluationof leukemic contamination in peripheral blood stemcell harvests by reverse transcriptase polymerase chainreaction. Br J Haematol 1993; 85:578-83.

96. Brenner MK, Rill DR, Moen RC, et al. Gene-markingto trace the origin of relapse after autologous bone-marrow transplantation. Lancet 1993; 341:85-6.

97. Gorin NC, Dicke K, Lowenberg B. High dose therapyfor acute myelocytic leukemia treatment strategy:what is the choice? Ann Oncol 1993; 4 (Suppl. 1):59-80.

98. Bassan R, Barbui T. Remission induction therapy foradults with acute myelogenous leukemia: towards theICE age? Haematologica 1995; 80:82-90.

99. Reiffers J, Korbling M, Labopin M, et al. Autologousblood stem cell transplantation versus autologousbone marrow transplantation for acute myeloidleukemia in first complete remission. Int J Cell Cloning1992; 7(Suppl. 1):111.

100. Korbling M, Fliedner TM, Holle R, et al. Autologousblood stem cell (ABSCT) versus purged bone marrowtransplantation (pABMT) in standard risk AML: influ-ence of source and cell composition of the autografton hemopoietic reconstitution and disease free sur-vival. Bone Marrow Transplant 1991; 7:343-9.

101. Sanz MA, de la Rubia J, Sanz GF, et al. Busulfan pluscyclophosphamide followed by autologous bloodstem cell transplantation for patients with acutemyeloblastic leukemia in first complete remission: areport from a single institution. J Clin Oncol 1993;11:1661-7.

102. To LB, Roberts MM, Haylock DN, et al. Comparisonof hematological recovery times and supportive carerequirements of autologous recovery phase peripher-al blood stem cell transplants, autologous bone mar-row transplants and allogeneic bone marrow trans-plants. Bone Marrow Transplant 1992; 9:277-84.


104. To LB, Haylock DN, Dyson PG, Thorp D, RobertsMM, Juttner CA. An unusual pattern of hemopoieticreconstitution in patients with acute myeloid leukemiatransplanted with autologous recovery phase periph-eral blood. Bone Marrow Transplant 1990; 6:109-14.

M. Aglietta et al.

31


105. Bender JG, To LB, Williams S, Schwartzberg LS. Defin-ing a therapeutic dose of peripheral blood stem cells.J Hematother 1992; 1:329-42.

106. Kotasek D, Sheperd KM, Gage RE, et al. Factorsaffecting blood stem cell collections following high-dose cyclophosphamide mobilization in lymphoma,myeloma and solid tumors. Bone Marrow Transplant1992; 9:4-17.

107. Laporte JP, Gorin NC, Feuchtenbaum J, et al. Relapseafter autografting with peripheral blood stem cells[letter]. Lancet 1987; 2:1393.

108. Yan Liu K, Akashi K, Harada M, Takamatsu Y, Niho Y.Kinetics of circulating haematopoietic progenitorsduring chemotherapy-induced mobilization with orwithout granulocyte colony-stimulating factor. Br JHaematol 1993; 84:31-8.

109. Onho R, Tomonaga M, Kobayashi T, et al. Effect ofgranulocyte colony-stimulating factor after intensiveinduction in relapsed or refractory acute leukemia. NEngl J Med 1990; 323:871-7.

110. Onho R, Hiraoka A, Tanimoto M, et al. No increaseof leukemia relapse in newly diagnosed patients withacute myeloid leukemia who received granulocytecolony-stimulating factor for life-threatening infectionduring remission induction and consolidation thera-py. Blood 1993; 81:561-2.

111. Hagenbeek A, Schultz FW, Martens ACM. In vitro orin vivo treatment of leukemia to prevent relapse afterautologous bone marrow transplantation. In: DickeKA, Spitzer G, Jagannath S, eds. Autologous bonemarrow transplantation: Proceedings of the FourthInternational Symposium. MD Anderson HospitalPubl., 1989:107-12.

112. Gribben JH, Freedman AS, Neuberg D, et al. Immuno-logic purging of marrow assessed by PCR before autol-ogous bone marrow transplantation for B-cell lym-phoma. N Engl J Med 1991; 325:1525-33.

113. Brenner MK, Rill DR, Moen RC, et al. Gene-markingto trace origin of relapse after autologous bone mar-row transplantation. Lancet 1993; 341:85-6.

114. Gulati SC, Lemoli RM, Acaba L, Igarashi T, WasserheitC, Fraig M. Purging in autologous and allogeneic bonemarrow transplantation. Curr Opin Oncol 1992; 4:543-50.

115. Sharkis SJ, Santos GW, Colvin MO. Elimination ofacute myelogenous leukemia cells from marrow andtumor suspensions in the rat with 4-hydroperoxycy-clophosphamide. Blood 1980; 55:521-3.

116. Siena S, Castro-Malaspina H, Gulati SC, et al. Effectsof in vitro purging with 4-hydroperoxycyclophos-phamide on the hematopoietic and microenviron-ment elements of human bone marrow. Blood 1985;65:655-62.

117. Bullock G, Tang C, Tourkina E, et al. Effect of com-bined treatment with interleukin-3 and interleukin-6on 4-hydroperoxycyclophosphamide induced pro-

grammed cell death or apoptosis in human myeloidleukemia cells. Exp Hematol 1993; 21:1640-4.

118. Skorski T, Kawalec M, Hoser G, Ratajczac M, Gna-towski B, Kawiak J. The kinetic of immunologic andhematologic recovery in mice after lethal total bodyirradiation and reconstitution with syngeneic bonemarrow cells treated or untreated with mafosfamide(Asta Z 7654). Bone Marrow Transplant 1988; 3:543-51.

119. Lemoli RM, Gulati SC. In vitro cytotoxicity of VP-16-213 and nitrogen mustard: agonistic on tumor cellsbut not on normal human bone marrow progenitors.Exp Hematol 1990; 18:1008-12.

120. Lemoli RM, Gasparetto C, Scheinberg DA, et al.Autologous bone marrow transplantation in acutemyelogenous leukemia: in vitro treatment withmyeloid-specific monoclonal antibodies and drugs incombination. Blood 1991; 77:1829-36.

121. Lemoli RM, Igarashi T, Knizewski M, et al. Dye-medi-ated photolysis is capable of eliminating drug-resis-tant tumor cells. Blood 1993; 31:793-800.

122. Chang J, Morgenstern GR, Coutinho LH, et al. Theuse of bone marrow cells grown in long-term culturefor autologous bone marrow transplantation in acutemyeloid leukemia: an update. Bone Marrow Trans-plant 1989; 4:5-10.

123. Rowley SD, Jones RJ, Piantadosi S, et al. Efficacy of exvivo purging for autologous bone marrow transplan-tation in the treatment of acute non lymphoblasticleukemia. Blood 1989; 74:501-6.

124. Miller CB, Zenhbauer BA, Piantadosi S, Rowley SD,Jones RJ. Correlation of occult clonogenic leukemiadrug sensitivity with relapse after autologous bonemarrow transplantation. Blood 1991; 78:1125-31.

125. Gorin NC, Aegerter P, Auvert P, et al. Autologousbone marrow transplantation for acute myelocyticleukemia in first remission. A European survey of therole of marrow purging. Blood 1990; 75:1606-14.

126. Labopin M, Gorin NC. Autologous bone marrowtransplantation in 2502 patients with acute leukemiain Europe: a retrospective study. Leukemia 1992; 6(Suppl. 4):95-9.

127. Gulati SC, Acaba L, Yahalom J, et al. Autologous bonemarrow transplantation for acute myelogenousleukemia using 4-hydroperoxycyclophosphamide andVP-16 purged bone marrow. Bone Marrow Transplant1992; 10:129-34.

128. Laporte JP, Douay L, Lopez M, et al. One hundredtwenty-five adult patients with primary acute leukemiaautografted with marrow purged with mafosfamide:a 10-year single institution experience. Blood 1994;84:3810-8.

129. Roberts AW, Metcalf D. Noncycling state of periph-eral blood progenitor cells mobilized by granulocytecolony-stimulating factor and other cytokines. Blood1995; 86:1600-5.



The aim of this review is to define the role ofperipheral blood stem cell transplantation for thetreatment of multiple myeloma. Therefore, we

first review our present knowledge of this disease andthen analyze the clinical trials based on the use ofautologous bone marrow or peripheral stem celltransplantation. Optimal methods for peripheralblood stem cell transplantation will also be discussed.

MyelomagenesisMultiple myeloma (MM), the prototype plasma cell

malignancy, is characterized by the uncontrolledaccumulation of plasma cells that replace normalbone marrow (BM) and by the overproduction ofmonoclonal immunoglobulins (Ig) and cytokines. Anumber of observations provided both by basic sci-ences and by clinical investigation allow us to placethe disease and its unusual features in a more coher-ent perspective and to discuss new therapeuticoptions properly.

EpidemiologyThe reported incidence of MM is available for the

years up to 1982 and varies substantially in differentcountries.1 A striking increase in the incidence of MMhas been noticed in the last thirty years and is onlypartially2 explained by amelioration of diagnosticcapabilities.3 Between 1973 and 1990 an increase of40% among people over 65 and of almost 15%among people under 65 has been recorded in US Can-cer Death rates.4 Ethnic differences are apparent: theincidence is twice as high and the mortality rate hasquadrupled in blacks, while doubling in whites.4 Bycontrast, rates among Asians are lower than those ofwhites living in the same geographic area.5

Both genetic and environmental factors can beinvoked to explain these ethnic differences. A signif-icant increase has been detected in first-degree rel-atives of patients.5 Moreover, an increased risk has

been observed to be associated with occupationaland environmental elements that include farmingexposure to pesticides, exposure to ionizing radia-tions, petroleum and rubber processing, as well aspersistent (viral) infections.3 The main conclusion thatcan be drawn from a large body of observations isthe necessity of discriminating the genetic roots fromthe environmental links of the disease. As a corollary,it may be asked which elements (genetic vs. environ-mental) are associated with the development of mon-oclonal gammopathy of undetermined significance(MGUS) and how they relate to the progression ofMGUS to overt MM.

Cytogenetics and molecular biologyTwo major pieces of information have emerged

from cytogenetic studies. The first is that no consis-tent (yet not random) chromosome abnormalitieshave been detected in MM.6 The second is thatnumeric chromosome abnormalities are shared byMGUS and MM.7,8 Both facts lead us to ask what theprerequisite is and what the additional events are inthe development of plasma cell malignancies. We stilldo not know the prerequisite events that lead toMGUS, to MM or to the evolution of MGUS into MM,or how they differ from collateral elements that sim-ply favor the malignant process. Along the same vein,it is interesting that no known specific oncogene hasyet been related to the development of MM or to thetransition from MGUS to overt MM. The genes mostcommonly implicated in MM, like N-RAS, P53 andretinoblastoma gene (RB), are all involved in the latestages of the disease.9

If the same cytogenetic abnormalities are shared bytwo clinical situations as different as MGUS and overtMM, a patrolling role for the immune system can beenvisaged in the natural history of plasma cell disor-ders. It is not unreasonable to suspect that if theimmune system is able to keep a malignant clone

review

Peripheral blood stem celltransplantation for the treatment of multiple myeloma: biological and clinical implications


FEDERICO CALIGARIS CAPPIO, MICHELE CAVO,ARMANDO DE VINCENTIIS, LUIGI LANATA, ROBERTO MASSIMO

LEMOLI, IGNAZIO MAJOLINO, CORRADO TARELLA, PAOLA ZANON,SANTE TURA

Cattedra di Immunologia Clinica, Dipartimento di ScienzeBiomediche e Oncologia Umana, Università di Torino, Turin;Istituto di Ematologia ed Oncologia Medica“Lorenzo e AriostoSeràgnoli”, Università di Bologna, Bologna; Dompé BiotecSpA, Milan; Dipartimento di Ematologia, Unità Trapianti diMidollo Osseo, Ospedale Cervello, Palermo; Divisione Univer-sitaria di Ematologia, Università di Torino, AziendaOspedaliera S. Giovanni, Turin; Amgen Italia SpA, Milan; Italy

Acknowledgments: preparation of this manuscript was supported bygrants from Dompé Biotec SpA and Amgen Italia SpA, Milan, Italy.

Correspondence: Prof. Sante Tura, Istituto di Ematologia ed Oncolo-gia Medica “Seràgnoli”, Policlinico S. Orsola, via Massarenti 9,40138 Bologna, Italy. Received January 11, 1996; accepted June 4, 1996.

33


under control, a benign MGUS is the resulting disease;the breakdown of this control would lead to MM. Littledirect, but much indirect evidence is available in murinemodels to suggest the immunomodulation of myelomacell growth by host effector cells.10

Immunochemistry and B cell differentiationstudies

Three major findings have been obtained throughimmunochemistry and by a more proper understandingof the differentiation processes of B lineage cells. First,MM paraproteins may be directed against a wide vari-ety of infectious agents, suggesting that MM develop-ment and antigen (Ag) stimulation may be causally relat-ed.11-13 Second, the Ig isotype of MM plasma cells is gen-erally IgG or IgA, demonstrating that the predominantphenotype of MM tumor cells is post-switch.9 Third,clonal proliferation involves a cell population that hasalready passed through the stage of Ig genes somatichypermutation.14,15 Since this process occurs in the ger-minal centers (GC) of secondary follicles,16 its presenceis a clear marker of the differentiative and functionallevel reached by the cell population being analyzed.

By and large, the observation that MM is a neoplasmof plasma cells that have a post-switch phenotype,show somatic mutations and may produce monoclonalIg with targeted antibody (Ab) activity leads to the con-clusion that MM is an Ag-driven process, even if thespecific causal Ag is generally unknown. This assump-tion has to be confronted with the simple, though basic,lesson from clinical medicine that MM is a BM disorder.In contrast with the distribution of normal plasma cells,MM plasma cells localize uniquely within the BM.9Although the lamina propria of the intestine containsmore Ig-producing cells than all other tissues in thebody, it is never a site where MM develops, not evenIgA1- and IgA2-producing MM.17 Likewise, involvementof the spleen and/or lymph nodes, though typical ofWaldenström’s macroglobulinemia, is very unusual inMM.17 The exclusive BM localization of MM plasma cellsappears to conflict with the extensive somatic hyper-mutations of the Ig they produce, which indicate aperipheral origin of malignant cells. However, while thesteps of Ag processing and presentation that lead tothe generation of somatically mutated IgG and IgA plas-ma cells occur only in secondary lymphoid follicles, theBM is a major site of IgG and IgA production in T-cell-dependent secondary immune responses.18-20 Plasma cellprecursors with specific traffic commitments originatefrom secondary lymphoid organs and migrate to the BMa few days after the Ag challenge (Figure 1).20,21

The issue whether MM plasma cell precursors are ear-ly BM stem cells or late peripheral B cells is misleading.The cell whose original transformation has ultimatelygenerated the malignant plasma cell progeny that wesee in MM cannot be equated with the B cell popula-tion that disseminates the disease throughout the axi-al skeleton.21 The identity of the hypothetical MM stemcell is unknown, i.e. we do not know either the cellular

target of the primary transforming event or where,when and how the unknown cellular target was hit bythe transforming event. By contrast, the informationavailable on the B cell population that feeds the down-stream compartment of plasma cells and disseminatesthe disease indicates that this population has been gen-erated in peripheral lymphoid organs during secondaryT-cell-dependent Ab response, is programmed to hometo the BM, and is committed to differentiate in closeassociation with the BM microenvironment (Figure2).21,22 On the basis of existing data, the most likely can-didate for the physiological B lymphocyte equivalent ofthe MM plasma cell precursor is either a B memory cellor a plasma blast (Figure 1).14,15,23,24

Microenvironment and cytokines It is assumed that BM-seeking plasma cell precursors

receive a differentiation signal after contact with theBM stromal microenvironment (Figure 2).25,26 Microen-vironmental stromal cells play an essential role in thegrowth of plasma cell tumors both in mice27 and inhumans.28 MM BM stromal cells are well equipped witha large series of adhesion and extracellular matrix mol-ecules that mediate homotypic and heterotypic inter-actions and provide anchorage sites to cells selectivelyexposed to locally released growth factors.22,29,30 MMBM stromal cells produce cytokines like IL-6 known toplay a crucial role in the evolution of the disease bothin experimental systems, including IL-6 transgenic mice,

Stromal cell

extracellular matrix proteins stromal cell released cytokines (IL-6)

centroblast centrocyte

B memory cellsplasma blast

Plasma cell

BONE MARROW

LYMPHOID FOLLICLES

B memory cell

Figure 1. Plasma cell precursors generated in peripherallymphoid organs differentiate in contact with bone marrowstromal cells.

PBSC transplantation in multiple myeloma

34


F. Caligaris Cappio et al.

and in vivo.31-34 High levels of IL-6 are observed in thesera of patients with aggressive or progressive MM,35

and infusion of anti-IL-6 antibodies in patients withplasma cell leukemia or MM refractory to therapy hasdecreased the size of the plasma cell pool and ham-pered the proliferative activity of plasma cells.36

Malignant MM plasma cells are not inert vehicles ofmonoclonal Ig. They also produce a number of cytokines,including interleukin (IL)-1b, tumor necrosis factor(TNF)-b and monocyte-macrophage colony stimulatingfactor (M-CSF), that activate stromal and accessorycells, aa well as having significant osteoclast activatingfactor (OAF) activity.22,37 A minority of human MM celllines autonomously produce small amounts of IL-6, butit is unclear whether fresh MM plasma cells can alsoproduce IL-6.34 IL-6, besides promoting B cell prolifer-ation and differentiation, has recently been shown tohave important OAF activity.32,33

These experimental findings linked to clinical obser-vations lead to the attractive hypothesis (Figure 2) thata self-maintaining series of mutual interactions betweenthe malignant B cell clone and the BM microenviron-ment may explain the progression of MM22 through theproduction of ever-increasing amounts of cytokinescapable of recruiting and activating several microenvi-ronmental cells, including osteoclasts.

The role of autologous transplantationin the treatment of multiple myeloma

Investigations into the use of myeloablative therapyfor the management of MM were pionereed in the mid-1980s and were stimulated by a persistent lack ofprogress in prognosis with conventional chemothera-py.38,39 As is the case with any experimental approach,initial trials were restricted to the treatment of patientswith advanced refractory or relapsing disease and werefocused mainly on defining the feasibility and toxicity ofthe procedure. These preliminary experiences were per-formed without the support of hemopoietic stem cellsand demonstrated that high-dose melphalan (HDM), giv-en intravenously (i.v.) at doses ranging between 100 and140 mg/m2, yielded an increase in the complete remis-sion (CR) rate, albeit at the expense of prolonged mar-row aplasia and an unacceptably high early mortalityrate.40-42 On the basis of these observations later stud-ies with chemotherapeutic agents administered at mye-loablative doses, and possibly added total body irradia-tion (TBI), were carried out with the support of autolo-gous BM and/or peripheral blood hemopoietic stem cells(PBSC).43 Demonstration of the safety and relative effi-cacy of autotransplants in refractory MM41,44-46 encour-aged subsequent application of this procedure in earli-er phases of the disease45,46 and, more recently, in new-ly diagnosed patients as well.47,48 Over the past decadeinterest in this new treatment strategy has progressive-ly grown, and the number of reported patients receivingautologous hemopoietic stem cell-supported myeloab-lative therapy is now approximately one thousandworldwide.

What lessons have we learned from this collectiveexperience? It is difficult to draw firm conclusions frompublished trials since none of them were controlled andpatient populations were different, as were the prepar-ative treatments and the criteria used for evaluatingtumor response. In addition, the bias introduced bypatient selection and, in most of the cases, the lack ofan adequate follow-up also helped complicate correctinterpretation of the data. As a consequence, the exactrole of autotransplantation in the management of MMstill remains poorly defined and could be properlyaddressed only in controlled clinical studies comparingautografting and conventional chemotherapy. There areat least several such trials in progress at the moment inEurope and the United States. Data reported at the lastASH meeting in Seattle (1995) by the IntergroupeFrançais du Myelome are promising and suggest anadvantage for autografted patients in terms of increasedCR rate and extended survival duration.49

Obviously, these results warrant confirmation in larg-er independent series. For this reason, similar investi-gations are currently being conducted in the UnitedStates under the auspices of the National Cancer Insti-tute. While the conclusions of these studies are beingawaited, analyses of available transplant data have pro-vided the following important information.

Fibroblasts

T lymphocytes

Macrophages

Bone Marrow

Osteoclasts

Precursors

Plasma cells

IL-1bb, TNF-bb, M-CSF

IL-1bb, IL-3, GM-CSF,TNFaa IL-6, IL-7,stem cell factor

Microenvironment cells

Fig. 2

ReciprocalActivation

Figure 2. Model of multiple myeloma growth and progres-sion based upon a series of mutual interactions between theB-cell clone and the bone marrow microenvironment.

35


Transplant-related mortalityTransplantation of autologous hemopoietic stem cells

following myeloablative therapy has greatly improvedthe tolerance to this modality of treatment and reducedthe frequency of procedure-related mortality to lessthan 5-10%50-52 (Tables 1, 2). More recently, with thecombined support of BM and PBSC followed by post-transplant administration of hemopoietic growth fac-tors, early mortality was further decreased to approxi-mately 1-2%.53

Tumor response and overall survivalIncreased tumor response, as recognized by an

increase in the CR rate, has been reported by manygroups following myeloablative treatments (Table1).45–48,54,55 Basically, criteria for CR included both thedisappearance of monoclonal plasma cells in the bonemarrow, as evaluated on cytological smear examinationor on flow cytometric analysis of DNA and cytoplasmicimmunoglobulins, and no detectable M component byroutine electrophoresis (later immunofixation wasadded). As would be logically expected, the CR rate var-ied in different studies, with a range between 20% and80%, mainly depending on the use of more or less strin-gent definition criteria and the status of the disease attransplant (Tables 1 and 2). Moreover, the length of sur-vival was generally extended after autotransplant, up toa median of approximately 3 to 5 years (Tables 1 and2).48,50-52

Choice of myeloablative therapyHistorically, the autotransplant experience in MM can

be divided into two groups of studies: the ones using andthose not using TBI as part of the conditioning regimen.With few exceptions,55 HDM, administered at dosesranging between 140 and 200 mg/m2 has been themainstay of both chemo-radiotherapy45,47,49,50,52,56 andradiation-free regimens48,53,54,57 for the following rea-sons: it shows a close dose relationship, is not cross-resistant with other alkylating agents and compared tocyclophosphamide, seems to offer a better chance ofovercoming chemotherapy resistance.58 In the absence ofcontrolled clinical studies comparing different prepara-tive treatments in specific subgroups of patients, it ishard to draw any meaningful conclusion concerning thebest conditioning treatment. The impression from thedata available in the literature is that no particular reg-imen demonstrated clear-cut superiority over the oth-ers. Therefore the choice of treatment to be used aspreparation for autotransplant should ultimately takeinto account the ability to perform TBI, patient eligibil-ity for TBI (those previously irradiated on the spine can-not, in fact, be candidates for radiation), and the expect-ed toxicity. HDM at 200 mg/m2 probably has less acuteextrahematological toxicity than regimens including TBI,a finding that formed the basis for exploring repeatedadministrations of this drug with tandem (or double)autotransplant programs.53,59


Group No. % Source % % Median mos.pts. sens. % BM % PB ED CR PFS Surv.

EBMT 130 68 63 25 6 48 17 27

Univ. Arkansas (USA) 287 60 unknown <5 27 (IF) 22 35

French Registry 133 77 61 38 4 37 33 46

Abbreviations: EBMT, European Group for Blood and Marrow Transplantation; Sens., responsive to conventional chemotherapy; BM, bone mar-row; PB, peripheral blood; ED, early death; CR, complete remission; IF, immunofixation analysis; PFS, progression-free survival.

Table 1. Results of autotrans-plants for multiple myeloma.

Author No. Median mos BM/PB TBI % % IFN-a Median mos.pts. to transpl. ED CR PFS Surv.

Jagannath 14 <12 +/- + 0 36 (IF) – 16 33+ (86%)

Attal 35 9 +/- + 3 43 + 33+ (53%) 41+ (81%)

Cunningham 53 <12 +/- – 2 75 – 23 54+ (63%)

Harosseau 103 7.5 +/+ + 4 33 ± 37 54

Barlogie 89 <12 +/+ +/- 0 46 (IF) + 37 71+

Abbreviations: BM, bone marrow; PB, peripheral blood; ED, early death; CR, complete remission; IF, immunofixation analysis; IFN-a, interferon-a; PFS, progression-free survival.

Table 2. Results of autotrans-plants for recently diagnosedMM patients with chemosensi-tive disease.

36



Remission durationAs previously emphasized, myeloablative therapy

requiring autologous hemopoietic stem cell support pro-vides substantial antitumor response, especially inpatients with good prognosis (see below). However, evenin this favorable condition, a considerable relapse rate,approaching 60% at 3 years, is reported after auto-transplant and no plateau is yet apparent on relapse-freesurvival curves.50-52 These results contrast with the 30%probability of long-term unmaintained remissions (andpossible cures) reported by several groups for patientsreceiving allogeneic transplantation.60 It has been sug-gested that the lack of an immunological effect by thedonor’s marrow T lymphocytes on the residual myelomacells (i.e. graft-versus myeloma)61,62 and/or possibletumor reseeding may account for the apparently lessdurable duration of disease control following autolo-gous as opposed to allogeneic transplantation. For thisreason, important issues currently under clinical inves-tigation in the autografting setting include furtherincreases in the cytotoxic dose intensity level and deple-tion of tumor cells from the graft (see below).

Prognostic variablesSeveral important variables affecting the outcome of

autologous transplantation have been identified (Table3), including b2-microglobulin (b2-M) levels,45,47,50-52,56

pre-transplant disease status,45,51,52 age,45,51,52 perfor-mance status,45 Ig isotype45,51,52 and response to mye-loablative therapy (e.g. attainment or non-attainment ofCR).47,48 In particular, at multivariate regression analysisearly mortality was reported to be highest among resis-tant relapsing patients, who also had the poorestresponse to myeloablative therapy and the shortestrelapse-free survival duration.45 In contrast, low serumb2-M levels, both at diagnosis and before autografting,and prior responsiveness to conventional chemotherapyconferred the highest CR rate, as well as prolongedrelapse-free and overall survival durations.45,47,50-52,56 Inaddition, the timing of autotransplant also emerged as animportant and independent prognostic parameter.56,64

This observation, on the one hand, was related to thegenerally reported improved outcome of patients trans-planted earlier and, on the other hand, reflected theacquisition of multiple biological abnormalities inadvanced phases of the disease63 that ultimately led torefractoriness even to high-dose therapy.64 Conversely,retaining sensitivity to high-dose therapy in earlier phas-es of MM assured better results, even in patients withprimary refractory disease.45,47,65

New perspectives under clinical investiga-tion

Based on the assumption that the failure of the con-ditioning regimen to eradicate the myeloma clone con-tributes most to post-transplant relapse, attempts toincrease the intensity, and possibly the efficacy, oftreatment by means of repeated courses of myeloab-lative therapy have recently been undertaken.46,53 Themore rapid recovery of hemopoiesis assured by the

combined use of PBSC and post-transplant adminis-tration of hemopoietic growth factors59 made the dou-ble transplant strategy feasible for approximately 60%of patients within one year.53 Results of pilot trials inprimary refractory MM indicated that such an approachprovided superior antitumor effect with improvedevent-free and overall survival durations with respectto a single transplant.53

A controlled clinical study comparing in a random-ized fashion single vs. double autografting in newlydiagnosed patients is currently underway in France. Asimilar trial is already in the early accrual stage in Italy.These studies will clarify in the next several yearswhether double transplant is associated with betterprognosis. Alternatively, efforts to improve the clinicalimpact of autotransplant have been carried out by sev-eral groups and have included depletion of tumor cellsfrom autografts by both negative selection of myelo-ma cells and positive selection of CD34+ hemopoieticstem cells,66,67 as well as post-transplant immunomo-dulation with interferon-a (IFN-a).47,49,68

In summary, hemopoietic stem cell-supported mye-loablative therapy holds the promise of being a safe andeffective treatment modality for MM. It yields betteroverall response and CR rates than conventionalchemotherapy and may prolong the duration of sur-vival.49

These conclusions, while encouraging, have beendrawn mainly from uncontrolled studies carried out inselect groups of patients and obviously warrant confir-mation in controlled clinical trials which are currentlyunder way. Therefore the next several years will clarifywhether newly diagnosed patients with symptomaticMM can be routinely offered a single or double auto-transplant as first-line or early salvage therapy for theirdisease.

Table 3. Variables affecting the outcome of autotransplantsfor multiple myeloma.@

D i s e a s e s t a t u s

Variable Refractory Refractory + Responsive

CR RFS Surv. CR RFS Surv.

Low b2M – +* +* + +* +*

Early transplant + + + + +* +*

CR achievement +* +*

Double transplant +* +*

CT responsiveness + + +

Younger age – + + + + +

Non IgA isotype – + + – + +

*In multivariate analyses. Abbreviations: CT, conventional chemotherapy;RFS, relapse-free survival. @Ref.: 45,47,48,50,51,52,56,63,64,65.

37


While the results of these studies are being awaited,wider application of myeloablative therapy should prob-ably be encouraged. Less heavily pretreated patientswho did not respond to prior conventional chemother-apy are more likely to benefit primarily from autotrans-plant. In addition, data available from the literature dosuggest that a superior outcome of this procedure canbe anticipated in patients with chemosensitive diseaseand low tumor burden at diagnosis. Hence, ongoingclinical trials aimed at comparing conventional versusmyeloablative therapy will also address the importantissue of the role of autotransplant as early consolida-tion therapy in patients with intrinsically good progno-sis. However, even in this favorable situation, recurrenceof the underlying malignant disease remains a majorproblem and is the most common cause of treatmentfailure. For this reason, attempts to improve the clini-cal impact of autografting are under active clinicalinvestigation.

In addition, many other problems regarding autolo-gous transplantation for MM are still unresolved andshould be formally addressed in future clinical trials.The most important of these issues include the choiceof the best conditioning regimen, the optimal source ofhemopoietic stem cells, the nature of relapse after auto-grafting, the benefit from purging techniques and, final-ly, the likelihood of long-term disease control, espe-cially for patients with molecularly defined CR.69

Advantages offered by the use ofPBSCs in the treatment of multiplemyeloma

The use of PBSC in support of high-dose chemo-radiotherapy (peripheral blood stem cell transplanta-tion) (PBSCT) is a valid alternative to autologous bonemarrow transplantation (ABMT) in the treatment ofboth hematologic and non-hematologic neoplastic dis-orders.70-72 The growing interest in this procedure can beexplained by: i) the possibility of mobilizing and col-lecting large amounts of hemopoietic progenitors,73,74

and ii) the rapid hemopoietic recovery observed follow-ing PBSCT.70,71,74-78

Progenitor collection represents the critical step in theprocedure. Daily monitoring of circulating CD34+ cells isan essential assay in predicting the number and timingof leukaphereses.79,80 Under proper conditions, only a fewleukapheresis procedures are required to collect enoughprogenitor cells for marrow reconstitution after mye-loablative treatments. Indeed, when circulating CD34+

cells rise to >50/µL, 1-2 leukaphereses may yield morethan 503104/CFU-GM/kg or 83106/CD34+ cells/kg,which are considered the ideal values for optimalengraftment.80-82 In addition, it has been shown thatlarge quantities of very immature elements, identified aslong-term culture-initiating cells (LTC-IC), are mobilizedas well.83-85

Inclusion in the harvested material of very immatureelements is responsible for the stable and durable mar-

row reconstitution observed in patients autografted withcirculating progenitors.77,83 Thus the term PBSC, nowcommonly employed to identify mobilized hemopoieticprogenitors, relies on both biological and clinical obser-vations. As previously emphasized, the rapidity ofengraftment is the major advantage offered by PBSC.Nevertheless, some authors argued that BM cells stim-ulated by growth factor administration might be at leastas efficient as mobilized progenitors in ensuring rapidengraftment following myeloablative treatment.87-89

However, it has recently been shown that both com-mitted and early progenitors are by far more frequentin PB than in BM during maximal mobilization.90 Thisconclusive observation points toward the preferentialuse of PBSC as the hemopoietic cell source for graft-ing purposes.

Since its introduction into clinical practice, PBSCThas been considered a promising approach for MMpatients.91,92 Several studies have been designed in thelast few years.

Reported results have shown a significant decrease inhemopoietic toxicity following this procedure as com-pared to ABMT, with recovery of granulocytes >0.53109/Land plateles >25-303109/L within approximately 2 weeksafter autograft (Table 4)42,44-48,52,54,56,93-98 This was paral-leled by rather good tolerability with rare early fatalevents.52,56,97,98

In addition, hemopoietic reconstitution by PBSCseems to be long lasting. MM patients may requirerepeated exposure to high-dose cytotoxic therapy.Reducing hemopoietic toxicity might be critical for theultimate treatment outcome. Therefore, also for itslong-term effect, PBSCT may have a positive impact onthe life expectancy of those patients who are suitablefor intensified chemo-radiotherapy treatments.99

PBSC mobilization and collection inmultiple myeloma

PBSC mobilization in myeloma patients PBSC collection presents specific problems in patients

with MM, where a decrease of progenitors in the bonemarrow is due in part to a defect of the mono-cyte/macrophage activation pathway. In fact, CD34+ cellsfrom MM patients grow normal numbers of colonieswhen stimulated by normal monocytes, while normalCD34+ cells have a reduced growth rate with MM mono-cytes.100 Another aspect is prior treatment. Repeatedcourses of chemo-radiotherapy are able to exhaust thepool of pluripotent stem cells,101 resulting in insufficientprogenitor cell harvests.59,102-105 Studies specificallyaddressed at MM patients show that melphalan106 andtreatment-free interval prior to PBSC mobilization107 alsohave an influence on the release of progenitors into theperipheral blood, while the value of BM plasmacytosis asan independent factor is more questionable.108,109 As aconsequence of these and other unknown factors, prog-enitor yields in MM are often unpredictable and lowerthan those observed in other malignant disorders.110


38


Nonetheless, cell harvests sufficient for one or two sub-sequent autografts are usually obtained,59,97,108,111-113 evenin patients with markedly infiltrated marrow or primaryresistant disease.109 To avoid the adverse influence of pre-mobilization treatment, PBSC collection in MM patientsshould be planned as early as possible in the course ofdisease, and alkylating drugs should be omitted in theprimary treatment. It should also be kept in mind thatheavily pre-treated patients require more leukapheresesand show slower platelet recovery after autograft.109 Thekey issues in the apheretic harvest of PBSCs in MM arepresented in Table 5.

PBSC also may be collected from patients with malig-nancies in steady state conditions;114 however, multipleaphereses are required with this method. Mobilizationof progenitors with cytotoxic chemotherapy, hemopoi-etic growth factors, or a combination of the two istherefore generally preferred. The hematopoietic recov-ery that occurs after cytotoxic chemotherapy is accom-panied by a PBSC rise that is proportional to the inten-sity of myelosuppression.102, 108

In MM, chemotherapy alone with either HDM,115 orCHOP-like regimens112,113,116 or intermediate- to high-dose cyclophosphamide (Cy)97,117,118 has been used tomobilize PBSC. However, the failure rate, defined as thepercentage of patients with a low progenitor cell peakin the blood or poor collections at the end of theapheresis program, was relatively high, ranging from 20to 30%. Moreover, when using high-dose therapy pro-tocols without growth factor support, one should con-sider that this implies an undue risk of severe toxicity.118

G-CSF73,119-121 and GM-CSF,75,112 as well as othercytokines are able to promote a dramatic rise of prog-enitors in the circulation. In a study of MM patients,administration of G-CSF at 10 µg/kg alone for six daysinduced a considerable increase in CFU-GM and CD34+

cells,111 with rapid recovery of counts after autograft.However, the use of growth factors alone in patientswith neoplastic disorders produces little enthusiasmamong hematologists. In fact, the spike of progenitorcells can be further amplified by combining growth fac-tors with chemotherapy.71 Together with the demon-stration that tumor cells are also mobilized by growthfactors,123 this fact makes the combination of

chemotherapy with G-CSF or GM-CSF the most reliableapproach.86,109,112,113

In MM as in other diseases,74,77 the use of growth fac-tors following cytotoxic treatment proves to be superi-or to chemotherapy alone in terms of progenitor cellyield,108,112 and significantly contributes to minimizingtreatment toxicity.112,124 High progenitor peak levels arereported108 with high-dose chemotherapy, namely Cy at7 g/m2 or etoposide (VP16) at 2 g/m2 followed by G-CSFor GM-CSF, and results seem to compare favorably withintermediate-dose Cy with or without G-CSF or GM-CSF. In conclusion, the optimal schedule for PBSC mobi-lization in MM has not yet been defined, though themost experience is with Cy at 7 g/m2 followed by G-CSFor GM-CSF. A review of the mobilization schedulesreported so far in MM patients is presented in Table 6.

Target of collections and cell monitoringCD34+ cell number and CFU-GM dose are both reliable

predictors of engraftment time.125-129 The amount ofPBSC necessary for engraftment is not clearly defined,but values of 10 to 203104/kg CFU-GM represent a rea-sonable minimal dose.110,120 Irrespective of disease, rapidneutrophil engraftment has been reported with203104/kg CFU-GM or 23106/kg CD34+ cells.125,130,131

However, a higher dose may be necessary for rapid andfull platelet engraftment.105,132 In a recent study of MMautografts, Tricot et al.59 found that platelet engraft-ment is influenced by previous history and cell dose. Inpatients with more than 24 months of chemotherapybefore the autograft, they found a dose ≥53106/kg tobe required for rapid and full platelet recovery postgraft. This number of CD34+ cells may be obtained with1 or 2 apheretic runs, and only a minority of patients,namely those with prolonged pre-mobilization treat-ment, need a higher number of apheretic procedures.The number of cells needed is obviously greater when adouble autograft is planned. When this is the case, sincerecovery after a second autograft is influenced by thesame factors as the first,59 the number of CD34+ cells tobe collected simply has to be doubled.

CD34+ cell monitoring in blood and collection prod-ucts is undoubtely the most reliable and rapid methodfor apheresis planning,131,133-135 though the assay


T ime t o r ecove r y f r om°Intensified treatments* leukopenia thrombocytopenia Treatment-related References

(days) (days) deaths (%)

Without autograft 28# 27 17 42,46,54,93-95

With BMT 20 26 7 44,45,47,48,96

With PBSCT 14 18 3.7 52,56,97,98

*intensified treatments consisted of HDM (60-200 mg/sqm) in most studies; the association of TBI/HDM was also used in some programs with auto-graft; °time to recovery from leukopenia and thrombocytopenia was reported as days to reach >0.5x109 ANC/µL and > 25x109 platelets/L, respective-ly, in nearly all studies; #table data have been calculated as medians from median values of hemopoietic recovery and from percentages of treatment-related deaths reported in each quoted study.

Table 4. Toxicity of intensifiedtreatments with or withoutautologous stem cell support inmultiple myeloma patients.

39


requires skillful personnel and carries a substantial cost.The issue has been reviewed extensively by Rowley.136

Siena et al.133 initially suggested starting the collectionprogram as soon as CD34+ cells were detectable in theperipheral blood. However, in terms of efficiency, thebest collections are performed when CD34+ cells are attheir peak. In practice, aphereses should be started assoon as the CD34+ cells in the blood exceed a given lev-el. We suggest a value of 20 CD34+ cells/µL combinedwith a WBC level >1.03109/L and a platelet count>303109/L before starting collections.82,133 Mononu-clear cells (MNC) in DNA synthesis also predict a goodyield when their level in the blood is >5% (or>250/µL).137

Few studies report detailed data on apheretic PBSCcollection in MM. Dimopoulos et al.109 began theaphereses when the MNC count went above 0.33109/L,having as target the collection of > 23106/kg CD34+

cells. They were able to collect >3.03106/kg CD34+ cellsdaily in patients with ≤ 4 months of prior chemothera-py, but the mean daily yield was uniformly lower(<13106 CD34+ cells/kg) in patients with more than 12months of chemo-radiotherapy. Tricot et al.59 initiatedcollections upon recovery of a WBC count > 0.53109/L,and assumed a target of > 63,108/kg MNC to supporttwo autografts. In a recent study113 aphereses werestarted as soon as the WBC count exceeded 53109/Lafter a CHOP-like regimen followed by G-CSF, and>63106/kg CD34+ cells were collected from all patientsin 1 to 3 aphereses.

A predictive test with G-CSF, a single dose of 10mcg/kg, followed by CD34+ cell monitoring on days 4and 5 has been proposed.138 The study included patientswith MM, but the sample was too small to draw anyconclusions. Steady-state CD34+ cell counts seem topredict the yield of PBSCs after mobilization withchemotherapy and G-CSF,139 but not after G-CSFalone.140 Table 7 shows the first apheresis day reportedwith different mobilization methods.98,108,111,112,117,141 Itis clear that the CD34+ cell peak occurs very early(approximately day 5 or 6) during mobilization withgrowth factors alone. When chemotherapy is included

in the mobilization schedule, the CD34+ cell peak dayoccurs later (approximately day 20), but the subsequentuse of growth factors will shorten it by a week or so.

To conclude, we suggest (Table 8) mobilizing PBSCwith the combination of chemotherapy and growth fac-tors (G-CSF or GM-CSF), and performing serial deter-minations of CD34+ cells in the blood. Aphereses shouldbe started as soon as the level of CD34+ cells exceeds20/µL, and collections should be performed daily withtwice the blood volume processed each time. Continu-ous-flow separators are to be preferred. As target forcollections, the figure of 23106/kg CD34+ cells per sin-gle autograft should be adopted for patients with < 24months of prior chemotherapy, while a greater number(> 53106/kg) should be collected in patients with alonger treatment history.

Assessment of myeloma cells in theperipheral blood and role of ex vivopurging

PBSC collections are generally believed to have low-er tumor cell contamination than BM harvests in can-cer patients eligible for autografting. Moreover, the useof circulating progenitor cells has shown more rapidhematopoietic reconstitution than reinfusion of BM-derived cells, thus reducing the incidence of seriousinfections and virtually eliminating mortality.142 Conse-quently, PBSCT is widely used after myeloablative ther-apy for the treatment of myeloma patients.53,56,59 How-ever, myeloma-related B-cells bearing the same idio-typic determinant as the neoplastic plasma cells havebeen identified in the blood of MM patients understeady-state conditions,143-149 and they may play a cru-cial role in the pathogenesis of the disease.144,147 There-fore in this chapter we will review the published dataconcerning: i) the presence of MM elements in PB andtheir kinetics in response to mobilization protocols; ii)methods for myeloma cell assessment; iii) methods forex vivo removal of contaminating tumor cells and therole of purging with respect to disease relapse.


Issue Related aspects

Dysregulated or suppressed hematopoiesis Decreased rate of progenitors,* defective monocyte activation,* pri-or chemotherapy

Methods for mobilization Type (and doses) of chemotherapy, use of growth factors

Toxicity of mobilization therapy Fever, allergy, infections, thrombosis

Kinetics of recovery after mobilization Timed and asynchronous use of WBC, monocytes and platelets

Prediction of harvest Prior chemotherapy, G-CSF test

Progenitor cell assays CD34+ cells, CFU-C

Target of collections Need for >53106/kg CD34+ cells in heavily pre-treated patients*

Apheresis method Cell separator, volume processed, schedule of aphereses

Tumor contamination of harvest Purging technique

Table 5. Key issues in mobi-lization and collection ofPBSCs.Note. Most aspects areshared with other malignantdisorders, and only a fewmay specifically affect multi-ple myeloma. These latterare marked with an *.

40


Identification of circulating myeloma cells Circulating B-cells belonging to the malignant clone

were originally thought to be pre-B-cells on the basis ofthe surface expression of the CD10 (CALLA) Ag,150 anendopeptidase present on all fetal pre-B and B-cells, onadult pre-B-cells and their neoplastic counterparts.151

However, the CD10 Ag has also been found on activatedB-cells151 and does not seem to be restricted to the ear-ly stages of B-lineage differentiation. Moreover, PBabnormal B-lymphocytes express plasma cell markerssuch as PCA-1 and PC-1 and the CD45RO Ag isoform,which is typical of late B-cells.145 Thus phenotypic analy-sis of circulating CD19+ cells indicates a heterogeneous,continuously differentiating B-lineage.145 By physicalparameters, CD19+ cells include a small and a large sub-set that are mainly late B-cells (pre-plasma cells) coex-pressing CD20, CD10, PCA-1, CD45RO and CD24 Ag.148

The majority of large B-cells also express the CD56 Agand high density CD38, whereas small lymphocytes showonly minor expression of these 2 antigenic determinants.This phenotypic profile (i.e. CD19+ CD20+ CD38++ CD56+)is not found in normal resting B-cells. Interestingly,malignant cells were detected at diagnosis, irrespectiveof tumor burden and stage of disease,148 and treatmenthad no detectable effect on the large B-cell subset. Con-versely, a significant decrease in the number of small B-lymphocytes followed chemotherapy, although thesecells returned to baseline value once the therapy wasdiscontinued. In this regard, it was previously shown thatcirculating CD19+ cells in MM express the functionalmultidrug transporter p-glycoprotein,147,169 thus sug-gesting that blood B-cells include a highly drug-resistantsubset capable of inducing disease recurrence in myelo-ma patients. However, it should be noted that matureplasma cells do not always express the CD19 Ag, where-as the presence of the CD56 Ag discriminates clonal plas-ma cells from normal ones.152 In addition, the recentlydescribed monoclonal antibody B-B4152 seems to be high-ly specific for BM and circulating terminal plasma cells.

More recently, the issue of myeloma cell contaminationin leukapheresis products and the kinetics of circulatingtumor cells in response to mobilization protocols havebeen addressed.67,69,153-155 These studies have consistent-ly shown that the majority of PBSC collections, if not all,are contaminated by myeloma cells, which represent upto 10% of PB mononuclear cells by immunophenotyp-ing and molecular analysis using polymerase chain reac-tion (PCR) with consensus oligonucleotides to the Igheavy chain complementary determining region III (CDRIII) (see below).155 The same pattern of contaminationhas been shown following high-dose Cy and either G- orGM-CSF,67,69,155 as well as after G-CSF alone,154 suggest-ing that growth factors for stem cell mobilization,regardless of the use of chemotherapy, may influencethe expression of adhesion molecules associated withthe myeloma cell membrane. Notably, kinetic analysishas demonstrated that following high-dose Cy and G-CSF, the concomitant mobilization of plasma cells andhematopoietic progenitor cells in the PB takes place withthe maximum peak of neoplastic elements occurringwithin the optimal time period for collection of circu-lating CD34+ cells.67 Conversely, GM-CSF seems toreduce asynchronous mobilization of neoplastic ele-ments and hematopoietic stem cells into PB, so that thecontamination of actively proliferating myeloma cells isminimal in the first two days of apheresis.156

Methods for assessment of minimal residualdisease

A number of methods have been proposed to detectmalignant cells in the blood of myeloma patients, includ-ing immunologic assessment by monoclonal antibodies,flow cytometry analysis of DNA and cytoplasmic Ig,studies on gene rearrangement. Each of these techniqueshas limitations in sensitivity and, in some cases, speci-ficity. For instance, analysis of the hypervariable regionof the Ig heavy chain (IgH) gene using a set of family-specific primers (IgH fingerprinting) requires 0.1% mon-


Authors No. pts Treatment Growth factor Day of Peaked Peaked Notesprogenitor peak CD34+/µL CFU-GM/mL

Reiffers117 15 Cy 7 g/m2 no nr nr nr 5/13 failures

Jagannath97 36 Cy 6 g/m2 no nr nr nr better with GM-CSF

39 Cy 6 g/m2 GM-CSF 17 nr nr

Tarella108 11 Cy 7 g/m2 or VP16 2 g/m2 GM-CSF 15 (13-16) 126 6432

4 Cy 231.2 g/m2 no 16 (16-18) 31 462

4 Cy 231.2 g/m2 GM-CSF 14 (14-15) 77 2588

Ossenkoppele111 6 no G-CSF36 gg 6 845

Majolino112 7 VCAD no 20 (17-30) 622

7 VCAD G-CSF 13 (9-17) 22 893

Vasta113 6 VCED G-CSF 13 (12-15) 70 2391

Legend. Cy: cyclophosphamide; VCAD: vincristine 1 mg, cyclophosphamide 4x500 mg/m2, adriamycin 2x50 mg/m2, dexamethasone 4x40 mg. VCED was identical to VCAD except that epirubicin 2x60 to 80

mg/m2 was substituted for adriamycin. nr: not reported.

Table 6. PBSC mobilization schedules in multiple myeloma.

PBSC transplantation in multiple myeloma 41


oclonal cells157 and may produce false positive results.Conversely, dual-parameter flow cytometric analysis(e.g. CD19/monoclonal light chain) and evaluation ofintracytoplasmic monoclonal heavy or light chain arehighly specific and allow detection as low as 0.1%67 (Fig-ure 3); however, they only assess mature Ig+ B-cells.Recently, several laboratories have described applica-tions of PCR techniques to increase significantly the sen-sitivity and specificity of detection of minimal residualdisease (MRD). Both consensus oligonucleotides (ODN)146

and family-specific primers69 have been used to ampli-fy the CDRIII of rearranged heavy chain alleles (Figure 4)from myeloma samples. From the sequence of the ampli-fied products, allele-specific (tumor-specific) oligonu-cleotides (ASO) were synthesized and used directly inPCR amplification reactions (ASO-PCR) for each patientsample to detect the malignant clone. The sensitivity ofthis method is 1:105 normal cells and a quantitativeanalysis can be performed by generating titrationscurves of tumor cells. Alternatively, direct fingerprintingof CDRIII IgH gene rearrangement may be used, althoughthe sensitivity is 1:104 normal cells.67

The biological and prognostic significance of cancercells present in autologous grafts is still unknown andcirculating myeloma cells may only reflect advancedstages of the disease; therefore relapse may be causedby regrowth of residual clonogenic cells in vivo. How-ever, considering that MM is a disease intrinsic to BMand recent studies clearly show that reseeding of rein-fused malignant cells contributes to relapse,158 severalattempts have been made to remove myeloma cellsfrom BM or PBSC autografts using different strategies.

Ex vivo purging of myeloma cellsOf the purging methods proposed for the elimination

of MRD, the cyclophosphamide derivative 4-hydroperoxy-cyclophosphamide (4-HC) was the first used,159 on thebasis of in vitro models demonstrating that this com-pound was able to eliminate BM-infiltrating MM celllines.160 The main mechanism of action of 4-HC is basedon a marked inhibition of myeloma cell growth, whereasit spares normal primitive hematopoietic cells.161 More-over, this alkylating agent seems to induce the apoptot-ic death of tumor cells162 as well as activate immunemechanisms capable of controlling malignant cell prolif-eration.163 Because 4-HC does not affect surface antigenexpression of myeloma cells, it is also a potential candi-date for combined treatment with monoclonal antibod-ies (MoAbs), and preliminary in vitro data confirm theadditive effect of these two purging techniques.160 Sev-eral MoAbs directed against tumor-associated or cell dif-ferentiation antigens not expressed by primitive cellsresponsible for hematopoietic engraftment have beenselected for clinical trials after in vitro studies demon-strated high purging efficacy with the use of comple-ment,164,165 toxins166,167 or immunoaffinity columns.168

Gobbi et al. developed a series of MoAbs that recognizemature plasma cells as well as B-cell precursors. One ofthem (8A) was conjugated with the ribosome-inactivat-

Table 8. Recommendations for PBSC mobilization and theirapheretic harvest in patients with multiple myeloma.

• Mobilization with chemotherapy + growth factors (G-CSF or GM-CSF)

• Serial CD34+ determinations according to institutional protocol

• Start apheresis when CD34+ cells in blood >203106/L

• Continuous flow separator, volume processed 3 2 blood volume per run

• Collect at least 23106/kg CD34+ cells in patients with < 24 months priorchemothrapy, at least 53106/kg CD34+ cells in patients with > 24 months prior chemotherapy

Figure 3. Circulating monoclonal B-lymphocytes and plasmacells assessed by double fluorescence immunostaining:intracytoplasmic Ig (green)/nuclear BRDU (red). Bromod-eoxyuridine (BRDU) is incorporated in actively proliferatingcells.

Table 7. Day of cell peak andof first apheresis after PBSCmobilization in patients withMM. The addition of G-CSFor GM-CSF shortens thetime to progenitor peak andconsequently the time toapheresis. Mean number ofapheretic procedures waslower when growth factorswere employed.

Regimen Day of Day of No. apheresis Referencesprogenitor peak first apheresis

G-CSF 10 mcg/kg/day 3 6 d 6 6 phlebotomy 3 2 112

Cy 7 g/m2 nr 20 6 117

Cy 7 g/m2 + GM-CSF 15 nr 4 98-108

Cy 7 g/m2 + G-CSF 15 14 2-3 141

VCAD 20 14 6 112

VCAD + G-CSF 13 12 2-3 112

Legend: nr: not reported.

F. Caligaris Cappio et al.42


ing toxin momordin and clinically tested in 8 advancedstage MM patients to eliminate, ex vivo, contaminatingmyeloma cells prior to ABMT.166 Although a marked tumorreduction was observed in all evaluable patients, none ofthem achieved CR and hematopoietic reconstitution fol-lowing the myeloablative conditioning therapy was sig-nificantly delayed in 3 patients. These preliminary resultsshowed the feasibility of this purging approach despitethe poor selection of patients.

The same MoAbs were also employed in vitro to removemyeloma cells through the avidin-biotin immunoabsorp-tion technique, and the result was a greater than 3 logreduction in tumor cells with acceptable recovery of BMprogenitors.168 More recently, Goldmacher et al.167 report-ed the development of an anti-CD38 immunotoxin capa-ble of killing 4-6 logs of human myeloma and lymphomacell lines. The immunotoxin was composed of an anti-CD38 antibody conjugated to a chemically modified ricinmolecule (blocked ricin). However, the CD38 Ag may notbe the proper target for purging because it is stronglyexpressed on myeloma plasma cells (see above) and oncommitted hematopoietic progenitor cells,169 which arethought to be essential for rapid BM reconstitution. Morespecific antibodies directed either toward B-cells (anti-CD10 and CD20) or mature plasma cells (PCA-1) and com-plement were used to deplete tumor cells from the graftbefore ABMT by Anderson et al.165 Following a TBI-con-

taining conditioning regimen, a neutrophil count greaterthan 0.53109/L and an unsupported platelet countgreater than 203109/L were reached at a median of 21days (range 12-46) and 23 days (range 12-53), respec-tively. Similarly, immunologic reconstitution was not dif-ferent from that commonly observed in cancer patientsreceiving unmanipulated autograft. This study docu-mented that high-dose chemo-radiotherapy can producea high response rate in pretreated patients with sensitivedisease, and MoAb-based purging methods do not preventrapid and sustained engraftment. However, the occur-rence of relapses post-ABMT and partial responses willnot define the need, if any, for marrow purging until moreeffective ablative strategies are developed. Taken togeth-er, these data demonstrate that the heterogeneity of Agexpression on neoplastic cells and the lack of true tumor-specific determinants may greatly influence the efficacyof antibody-based strategies for the depletion of myelo-ma cells. Alternatively, long-term Dexter-type marrowcultures have been used to select normal myeloid prog-enitors from heavily infiltrated myeloma BM, on the basisof the selective growth advantage of benign cells overmalignant cells in this system.170

Enrichment of hematopoietic CD34+ cells has latelybeen shown to be an alternative approach to myelomacell removal with a limited loss of normal stem cells. TheCD34 Ag is a 110-120 kD glycoprotein that is mainly

Sequencing

ASO probe construction

clonal CDR-III PCR product at diagnosis

ASOhybridization negative

ASOhybridization positive

CDRIII junctional region

non clonal CDRIII PCR product= molecular remission

clonal CDRIII PCR product= molecular persistance of minimal residual disease

VH N DH N JH

VH N DH N JHVH N DH N JH

VH N DH N JH

(allele specific oligonucleotide probe)

VH familyconsensus oligo

VH consensus oligo CDRIII

JH consensus oligo

FRI CDRI FRII CDRII FRIII

Figure 4. Schematic representation of the genomic region of rearranged CDRIII of IgH gene and further utilization of the PCRproduct for detection of MRD. For further details see text.

43


expressed on the earliest identifiable precursor cells andcommitted myeloid progenitors.169 In normal individuals,CD34+ cells represent 1% to 4% of the mononuclear cellfraction in the BM, whereas they are barely detectable inthe PB.169 In addition, the CD34 Ag is not expressed onthe surface of mature plasma cells in MM, although thepossibility that this glycoprotein may be present on clon-ally less differentiated B-lymphocytes is still a matter ofdebate. As reported above, recent data support thehypothesis that MM originates in the later stages of B-cell differentiation when B lymphocytes have lost theCD34 Ag,171 whereas other studies have found CD34+

cells to be part of the neoplastic clone.172,173 It should beunderlined, however, that reverse transcription-PCR,which was used to detect MRD in those studies, is anextremely sensitive technique, and the potential conta-mination of the CD34+ cell fraction by unwanted cellsshould be carefully avoided.

In this respect, Vescio et al.171 did not find IgH geneclonal rearrangement in collections of 99.99% pureCD34+ cells obtained after using a combination of twomethods of stem cell purification. Schiller et al.66 andLemoli et al.67 reported the first studies on purified, CD34-selected PBSCT conducted in patients with advanced MM.A median of 4.65 and 43106 CD34+ cells/kg were rein-fused in the two trials with a median purity of 77% and88.5%, respectively. The median time to neutrophil andplatelet recovery was 12 days and 10 and 11 days, respec-tively, with no difference with respect to a group ofpatients receiving unmanipulated PBSCs.67

Both reports utilized rigorously quantitative immuno-fluorescence and/or IgH gene rearrangement analysis,and tumor cell depletion ranging from 2.5 to 4.5 logs wasachieved. However, the persistence of myeloma cells inthe CD34+ cell fraction was documented by sensitive PCRassay in all cases heavily contaminated before positiveselection of CD34+ cells. Thus an additional purging stepmay be necessary to achieve a virtually tumor-free auto-graft.

In this regard, studies aimed at optimizing myelomacell depletion by positive selection of primitive CD34+

Lin–Thy+ cells have already been performed155 and clini-cal trials are currently in progress.

In summary, all these studies show the capacity ofpurging techniques to eliminate a substantial proportionof the myeloma cells from autologous grafts withoutaffecting their engraftment potential. The clinical impactof purging on disease relapse remains to be determinedin future randomized trials.

Post-transplant (immuno)therapyIn MM as well as in other hematologic malignancies,

the primary objective of high-dose therapy with hemo-poietic stem cell support is to prolong survival and pos-sibly to cure an otherwise incurable disease. The aim ofpost-transplant therapy is to prevent recurrence of thedisease while assuring good quality of life. From thislatter point of view, there is no room for additionalchemotherapy as a preventive means. In addition, high-

dose chemotherapy itself involves a risk of secondarymyelodysplastic syndrome or acute myeloid leukemia.This risk is apparently related to prolonged alkylatingagent therapy prior to transplantation and wouldundoubtedly increase with additional post-transplantchemotherapy.

In the past few years interferon-a (IFN-a) has beenextensively evaluated in the management of MM, eitheras part of the induction program or as maintenance ther-apy.173 Although controversial findings were frequentlyreported, several clinical trials showed a prolongation ofthe remission phase, and even of the survival duration,for patients receiving IFN-a after a favorable responseto conventional chemotherapy.174,175 These results sug-gested that IFN-a might be particularly useful in patientswith low tumor burden or minimal residual disease, andled to clinical investigations of this agent in the auto-graft setting.

The European Group for Blood and Marrow Transplan-tation (EBMT) has recently presented a retrospective studyof a large series of MM patients treated with autologousstem cell transplantation.50 Interestingly, post-transplanttreatment with IFN-a was independently associated withextended survival of responding patients, i.e. those achiev-ing either CR or partial remission. Moreover, Powels etal.176 designed a randomized clinical trial aimed at com-paring maintenance IFN-a therapy with no maintenanceafter HDM and ABMT.175 The authors found that IFN-aprolonged remission and improved the survival after auto-transplant, and that this effect was particularly marked inthe group of patients achieving CR.

Maintenance IFN-a is usually started three monthsafter transplant and is given sc at a dosage of 33106

U/m2, 3 times weekly. This dose usually induces mildhematological and non-hematological toxicity, thusallowing good quality of life. Available data indicate thatabout 50% of the MM patients who achieve CR and arethen treated with IFN-a remain in remission four yearsafter transplantation.

Alternatively, maintenance treatments aimed at pro-longing the duration of disease control after transplan-tation may also include the administration of interleukin2 (as nonspecific immunotherapy)177 or humanized anti-idiotype monoclonal antibodies, which could allow selec-tive killing of myeloma cells and might be particularlyuseful for controlling minimal residual disease.

References

1. Herrinton LJ, Weiss NS, Olshan AF. The epidemiologyof myeloma. In: Malpas JS, Bergsagel DE, Kyle RA, eds.Myeloma, biology and management, Oxford:OxfordMedical Publ, 1995:127-68.

2. Turesson I, Zettervall O, Cuzick J, Waldenström JB,Velez R. Comparison of trends in the incidence of mul-tiple myeloma in Malmo, Sweden and other countries.N Engl J Med 1984; 310:421-4.

3. Obrams GI, Potter M. Epidemiology and biology ofmultiple myeloma. Berlin:Springer-Verlag, 1991.

4. Evaluating the National Cancer Program: an ongoing


44


process. National Cancer Institute, Bethesda, Md,USA, 1994.

5. Riedel DA, Pottern LM. The epidemiology of multiplemyeloma. Hematol Oncol Clin N Am 1992; 6:225-47.

6. Sawyer JR, Waldron JA, Jagannath S, Barlogie B. Cyto-genetic findings in 200 patients with multiple myeloma.Cancer Genet Cytogen 1995; 82:41-9.

7. Zandecki M, Obein V, Bernardi F, et al. Monoclonalgammopathy of undetermined significance: chromo-some changes are a common finding within bone mar-row plasma cells. Br J Haematol 1995; 90:693-6.

8. Drach J, Angerler J, Schuster J, et al. Interphase fluo-rescence in situ identifies chromosomal abnormalitiesin plasma cells from patients with monoclonal gam-mopathy of undetermined significance. Blood 1995;86:3915-21.

9. Barlogie B. Plasma cell myeloma. In: Beutler E, Licht-man MA, Coller BS, Kipps TJ, eds. Williams’ Hematol-ogy, 5th ed. New York:1995: 1109-26.

10. Hoover RG, Kornbluth J. Immunoregulation of murineand human myeloma. Hematol Oncol Clin N Am1992; 6:407-24.

11. Potter M. Myeloma proteins (M-components) withantibody-like activity. N Engl J Med 1971; 284:831-8.

12. Seligmann M, Brouet JC. Antibody activity of humanmyeloma globulins. Semin Hematol 1973; 10:163-77.

13. Konrad RJ, Kricka LJ, Goodman DBP, Goldman J, Sil-berstein LE. Myeloma-associated paraprotein directedagainst the HIV-1 p24 antigen in an HIV-1-seropositivepatient. N Engl J Med 1993; 328:1817-9.

14. Bakkus MHC, Heirman C, Van Riet I, Van Camp B,Thielemans K. Evidence that multiple myeloma Ig heavychain VDJ genes contain somatic mutations but showno intraclonal variation. Blood 1992; 80:2326-35.

15. Vescio RA, Cao J, Hong CH, et al. Myeloma Ig heavychain V region sequences reveal prior antigenic selec-tion and marked somatic mutation but no intraclonaldiversity. J Immunol 1995; 155:2487-97.

16. Kelsoe G. B cell diversification and differentiation inthe periphery. J Exp Med 1994; 180:5-6.

17. Caligaris-Cappio F, Gregoretti MG. Basic concepts:plasma cells in multiple myeloma. In: BGM Durie, G.Gahrton, eds. Multiple Myeloma. London:E. ArnoldPublisher, in press.

18. Benner R, Hijmans W, Haajman JJ. The bone marrow:the major source of serum immunoglobulins, but stilla neglected site of antibody formation. Clin ExpImmunol 1981; 46:1-8.

19. DiLosa RM, Maeda K, Masuda A, Szakal AK, Tew JG.Germinal center B cells and antibody production in thebone marrow. J Immunol 1991; 146:4071-7.

20. Liu YJ, Johnson GD, Gordon J, MacLennan ICM. Ger-minal centres in T-cell dependent antibody responses.Immunol Today 1992; 13:17-21.

21. MacLennan ICM. In which cells does neoplastic trans-formation occur in myelomatosis? Curr Top Microbi-ol Immunol 1992; 182:209-13.

22. Caligaris-Cappio F, Gregoretti MG, Ghia P, Bergui L. Invitro growth of human multiple myeloma: implicationsfor biology and therapy. Hematol Oncol Clin N 1992;6:257-71.

23. Corradini P, Boccadoro M, Voena C, Pileri A. Evidencefor a bone marrow B cell transcribing malignant plas-ma cell VDJ joined to Cm sequence in IgG and IgAsecreting multiple myelomas. J Exp Med 1993; 178:1091-6.

24. Billadeau D, Ahmann G, Greipp P, Van Ness B. Thebone marrow of multiple myeloma patients contains Bcell populations at different stages of differentiationthat are clonally related to the malignant plasma cell.J Exp Med 1993; 178:1023-31.

25. Rolink A, Melchers F. Generation and regeneration ofcells of the B-lymphocyte lineage. Curr Opin Immunol1993; 5:207-17.

26. Kinkade P, Lee G, Pietrangeli CE, Hayashi SH, GimbleJ. Cells and molecules that regulate B lymphopoiesis inbone marrow. Annu Rev Immunol 1989; 7:111-43.

27. Degrassi A, Hilbert DM, Rudikoff S, Anderson AO, Pot-ter M, Coon HG. In vitro culture of primary plasmacy-tomas requires stromal cell feeder layers. Proc NatlAcad Sci USA 1993; 90:2060-4.

28. Caligaris-Cappio F, Bergui L, Gregoretti MG, et al. Roleof bone marrow stromal cells in the growth of humanmultiple myeloma. Blood 1991; 77:2688-93.

29. Thiery JP, Boyer B. The junction between cytokines andcell adhesion. Current Opin Cell Biol 1992; 4:782-92.

30. Hynes RO. Integrins: versatility, modulation and sig-nalling in cell adhesion. Cell 1992; 69:11-25.

31. Klein B, Zhang XG, Jourdan M, et al. Paracrine ratherthan autocrine regulation of myeloma-cell growth anddifferentiation by interleukin-6. Blood 1989; 73:517-26.

32. Klein B, Zhang XG, Lu ZY, Bataille R. Interleukin-6 inhuman multiple myeloma. Blood 1995; 85:863-72.

33. Kishimoto T, Akira S, Narazaki M, Taga T. Interleukin-6 family of cytokines and gp130. Blood 1995; 86:1243-54.

34. Jernberg H, Pettersson M, Kishimoto T, Nilsson K.Heterogeneity in response to interleukin 6 (IL-6),expression of IL-6 and IL-6 receptor mRNA in a panelof established human multiple myeloma cell lines.Leukemia 1991; 5:255-65.

35. Bataille R, Jourdan M, Zhang XG, Klein B. Serum lev-els of interleukin-6, a potent myeloma cell growth fac-tor, as a reflection of disease severity in plasma celldyscrasias. J Clin Invest 1989; 84:2008-11.

36. Klein B, Wijdenes J, Zhang XG, et al. Murine anti-inter-leukin-6 monoclonal antibody therapy for a patientwith plasma cell leukemia. Blood 1991; 78:1198-204.

37. Bataille R, Chappard D, Klein B. Mechanisms of bonelesions in multiple myeloma. Hematol Oncol Clin NAm 1992; 6:285-95.

38. Bersagel DE. The role of chemotherapy in the treat-ment of multiple myeloma. Baillière Clin Haematol1995; 8:783-94.

39. Gregory WN, Richards MA, Malpas JS. Combinationchemotherapy versus melphalan and prednisolone inthe treatment of multiple myeloma: an overview ofpublished trials. J Clin Oncol 1992; 10:334-42.

40. McElwain TJ, Powles RL. High-dose intravenous mel-phalan for plasma-cell leukemia and myeloma. Lancet1983; 1:822-4.

41. Barlogie B, Hall R, Znader A, et al. High-dose melpha-lan with autologous bone marrow transplantation formultiple myeloma. Blood 1986; 67:1298-301.

42. Selby PJ, McElwain TJ, Nandi AC, et al. Multiple myelo-ma treated with high dose intravenous melphalan. BrJ Haematol 1987; 66:55-62.

43. Cavo M, Benni M, Gozzetti A, et al. The role of haema-topoietic stem cell-supported myeloablative therapyfor the management of multiple myeloma. Baillière ClinHaematol 1995; 8:795-813.

44. Barlogie B, Alexanian R, Dicke KA, et al. High-dosechemoradiotherapy and autologous bone marrowtransplantation for resistant multiple myeloma. Blood1987; 70:869-72.

45. Jagannath S, Barlogie B, Dicke K, et al. Autologousbone marrow transplantation in multiple myeloma:identification of prognostic factors. Blood 1990; 76:1860-6.

46. Harosseuau JL, Milpied N, LaPorte JP, et al. Double-intensive therapy in high-risk multiple myeloma. Blood


45


1992; 79: 2827-33.47. Attal M, Huguet F, Schlaifer D, et al. Intensive com-

bined therapy for previously untreated aggressivemyeloma. Blood 1992; 79: 1130-6.

48. Cunningham D, Paz-Arez L, Milan S, et al. High-dosemelphalan and autologous bone marrow transplanta-tion as consolidation in previously untreated myelo-ma. J Clin Oncol 1994; 12:759-63.

49. Attal M, Harousseau JL, Stoppa AM, et al. High dosetherapy in multiple myeloma: final analysis of aprospective randomized study of the “IntergroupeFrançais du Myelome” (IFM 90). Blood 1995; 86 (sup-pl. 1): Abstract n° 485.

50. Björkstrand B, Goldstone AH, Ljungman P, et al. Prog-nostic factors in autologous stem cell transplantationfor multiple myeloma: An EBMT registry study. LeukLymphoma 1994; 15:265-72.

51. Barlogie B, Jagannath S, Vesole D, et al. Autologousand allogeneic transplants for multiple myeloma.Semin Hematol 1995; 32:31-44.

52. Harousseau JL, Attal M, Divine M, et al. Autologousstem cell transplantation after first remission induc-tion treatment in multiple myeloma: a report of theFrench registry on autologous transplantation in mul-tiple myeloma. Blood 1995; 85:3077-85.

53. Vesole DH, Barlogie B, Jagannath S et al. High-dosetherapy for refractory multiple myeloma: improvedprognosis with better supportive care and doubletransplants. Blood 1994; 84:950-6.

54. Gore ME, Selby PJ, Viner C, et al. Intensive treatmentof multiple myeloma and criteria for complete remis-sion. Lancet 1989; 2:879-81.

55. Dimopoulos MA, Alexanian R, Przepiorka D, et al.Thiotepa, busulfan, and cyclophosphamide: a newpreparative regimen for autologous marrow or bloodstem cell transplantation in high-risk multiple myelo-ma. Blood 1993; 82:2324-8.

56. Fermand JP, Chevret S, Ravaud P, et al. High-dosechemoradiotherapy and autologous blood stem celltransplantation in multiple myeloma: results of a phaseII trial involving 63 patients. Blood 1993; 82:2005-9.

57. Alegre A, Lamana M, Arranz R, et al. Busulfan and mel-phalan as conditioning regimen for autologous periph-eral blood stem cell transplantation in multiple myelo-ma. Br J Haematol 1995; 91:380-6.

58. Barlogie B, Jagannath S, Dixon DO, et al. High-dosemelphalan and granulocyte-macrophage colony-stim-ulating factor for refractory multiple myeloma. Blood1990; 76:677-80.

59. Tricot G, Jagannath S, Vesole D, et al. Peripheral bloodstem cell transplants for multiple myeloma: identifica-tion of favourable variables for rapid engraftment in225 patients. Blood 1995; 85:588-96.

60. Cavo M, Benni M, Cirio TM, et al. Allogeneic bonemarrow transplantation for the treatment of multiplemyeloma. An overview of published trials. Stem Cells1995; 13 (Suppl. 2): 121-6.

61. Tricot G, Vesole DH, Jagannath S, Hilton J, Munshi N,Barlogie B. Graft-versus-myeloma effect: proof of prin-ciple. Blood 1996; 87:1196-8.

62 Verdonck LF, Lokhorst HM, Dekker AW, NieuwenhuisHK, Petersen EJ. Graft-versus-myeloma effect in twocases. Lancet 1996; 347:800-1.

63. Joshua DE, Gibson J, Brown RD. Mechanisms of theescape phase of myeloma. Blood Rev 1994; 8:13-20.

64. Alexanian R, Dimopoulos M, Smith T, et al. Limitedvalue of myeloablative therapy for late multiple myelo-ma. Blood 1994; 83:512-6.

65. Alexanian R, Dimopoulos MA, Smith T, et al. Earlymyeloablative therapy for multiple myeloma. Blood1994; 84:4278-82.

66. Schiller G, Vescio R, Freytes C, et al. Transplantation ofCD34+ peripheral bood progenitor cells after high-dosechemotherapy for patients with advanced multiplemyeloma. Blood 1995; 86:390-7.

67. Lemoli RM, Fortuna A, Motta MR, et al. Concomitantmobilization of plasma cells and hemopoietic progen-itors into peripheral blood of multiple myelomapatients: positive selection and transplantation ofenriched CD34+ cells to remove circulating tumourcells. Blood 1996; 87:1625-34.

68. Cunningham D, Powles R, Viner C, et al. High-dosechemotherapy and autologous bone marrow trans-plantation in multiple myeloma. Abstract Book IVInternational Workshop on Multiple Myeloma(Rochester) 1993; p. 102.

69. Corradini P, Voena C, Astolfi M, et al. High-dosesequential chemoradiotherapy in multiple myeloma:residual tumor cells are detectable in bone marrow andperipheral blood cell harvests and after autografting.Blood 1995; 85:1596-602.

70. Gianni AM, Bregni M, Siena S, et al. Rapid and com-plete hemopoietic reconstitution following combinedtransplantation of autologous blood and bone marrowcells. A changing role for high-dose chemotherapy?Hematol Oncol 1989; 7:139-48.

71. Gianni AM, Bregni M, Siena S, et al. Granulocyte-macrophage colony stimulating factor to harvest cir-culating hemopoietic stem cells for autotransplanta-tion. Lancet 1989; 2:580-5.

72. Kessinger A, Armitage JO. The evolving role of autolo-gous peripheral stem cell transplantation followinghigh dose therapy for malignancies. Blood 1991; 77:211-2.

73. Sheridan WP, Begley CG, Juttner C, et al. Effect ofperipheral blood progenitor cells mobilized by filgras-tim (G-CSF) on platelet recovery after high dosechemotherapy. Lancet 1992; 1:640-4.

74. Tarella C, Ferrero D, Bregni M, et al. Peripheral bloodexpansion of early progenitor cells after high dosecyclophosphamide and rhGM-CSF. Eur J Cancer 1991;27:22-7.

75. Haas R, Ho AD, Bredthauer U, et al. Successful autol-ogous transplantation of blood stem cells mobilizedwith recombinant human granulocyte macrophagecolony-stimulating factor. Exp Hematol 1990; 18:94-8.

76. Shimazaki C, Oku N, Ashihara E, et al. Collection ofperipheral blood stem cells mobilized by high-doseAra-C plus VP-16 or aclarubicin followed by recombi-nant human granulocyte colony-stimulating factor.Bone Marrow Transplant 1992; 10:341-6.

77. Teshima T, Harada M, Takamatsu Y, et al. Cytotoxicdrug and cytotoxic drug/G-CSF mobilization of periph-eral blood stem cells and their use for autografting.Bone Marrow Transplant 1992; 10:215-20.

78. Brugger W, Birken R, Bertz H, et al. Peripheral bloodprogenitor cells mobilized by chemotherapy plus gran-ulocyte-colony stimulating factor accelerate both neu-trophil and platelet recovery after high-dose VP16, ifos-famide and cisplatin. Br J Haematol 1993; 84:402-7.


80. Siena S, Bregni M, Brando B, et al. Practical aspects offlow cytometry to guide large-scale collection of circu-lating hematopoietic progenitors for autologous trans-plantation in cancer patients. Int J Cell Cloning 1992;10:26-9.

81. Siena S, Bregni M, Di Nicola M, et al. Durability ofhematopoiesis following myeloablative cancer therapy


46


and autografting with peripheral blood hematopoieticprogenitors. Ann Oncol 1994; 5:935-41.

82. Ravagnani F, Siena S, Bregni M, Sciorelli G, Gianni AM,Pellegris G. Large cell collection of circulatinghaematopoietic progenitors in cancer patients treatedwith high-dose cyclophosphamide and recombinanthuman GM-CSF. Eur J Cancer 1990; 26:562-4.

83. Tong J, Hoffman R, Siena S, et al. Characterization andquantitation of primitive hemopoietic progenitor cellspresent in peripheral blood autografts. Exp Hematol1994; 21:1016-24.

84. Bender JG, Lum L, Unverzagt KL, et al. Correlation ofcolony-forming cells, long-term culture-initiating cellsand CD34+ cells in apheresis products from patientsmobilized for peripheral blood progenitors with differ-ent regimens. Bone Marrow Transplant 1994; 13:479-85.

85. Pettengell R, Luft T, Henschler R, et al. Direct com-parison by limiting dilution analysis of long-term cul-ture-initiating cells in human bone marrow, umbilicalcord blood, and blood stem cells. Blood 1994; 84:3653-9.

86. Gianni AM, Tarella C, Siena S, et al. Durable and com-plete hematopoietic reconstitution after autograftingof rhGM-CSF exposed peripheral blood progenitorcells. Bone Marrow Transplant 1990; 6:143-5.

87. Johnsen HE, Hansen PB, Plesner T, et al. Increased yieldof myeloid progenitor cells in bone marrow harvestedfor autologous transplantation by pretreatment withrecombinant human granulocyte-colony stimulatingfactor. Bone Marrow Transplant 1992; 10:229-34.

88. Janssen WE. Peripheral blood and bone marrowhematopoietic stem cells: are they the same? SeminOncol 1993; 20 (suppl. 6):19-27.

89. Eaves CJ. Peripheral blood stem cells reach newheights. Blood 1993; 82:1957-9.

90. Tarella C, Benedetti G, Caracciolo D, et al. Both earlyand committed haematopoietic progenitors are morefrequent in peripheral blood than in bone marrow dur-ing mobilization induced by high-dose chemotherapy+ G-CSF. Br J Haematol 1995; 91:535-43.

91. Reiffers J, Marit G, Boiron JM. Autologous blood stemcell transplantation in high-risk multiple myeloma. BrJ Haematol 1989; 72:296-7.

92. Pileri A, Tarella C, Bregni M, et al. GM-CSF exposedperipheral blood progenitors as sole source of stemcells for autologous transplantation in two patientswith multiple myeloma. Haematologica 1990; 75:79-82.

93. Barlogie B, Alexanian R, Smallwood L, et al. Prognos-tic factors with high-dose melphalan for refractory mul-tiple myeloma. Blood 1988; 72:2015-9.

94. Lokhorst HM, Meuwissen OJA, Verdonk LF, Dekker A.High-risk multiple myeloma treated with high-dosemelphalan. J Clin Oncol 1992; 10:47-51.

95. Cunningham D, Paz-Ares L, Gore ME, et al. High-dosemelphalan for multiple myeloma: long-term follow-updata. J Clin Oncol 1994; 12:764-8.

96. Mansi J, da Costa F, Viner C, Judson I, Gore ME,Cunningham D. High-dose busulfan in patients withmyeloma. J Clin Oncol 1992; 10:1569-73.

97. Jagannath S, Vesole DH, Glenn L, Crowley J, BarlogieB. Low-risk intensive therapy for multiple myeloma withcombined autologous bone marrow and blood stemcell support. Blood 1992; 80:1666-72.

98. Gianni AM, Tarella C, Bregni M, et al. High-dosesequential (HDS) chemo-radiotherapy, a widely applic-able regimen, confers survival benefit to patients withhigh-risk multiple myeloma. J Clin Oncol 1994;12:503-9.

99. Tricot G, Jagannath S, Vesole DH, Crowley J, Barlogie

B. Relapse of multiple myeloma after autologous trans-plantation: survival after salvage therapy. Bone MarrowTransplant 1995; 16:7-11.

100. Wunder E. Stimulation of granulocyte macrophageprogenitors via monocyte/macrophage activation: afundamental regulatory pathway of terminal differen-tiation. In: Wunder EW, Henon PR, eds. Peripheralblood stem cell autografts. Berlin:Springer-Verlag,1993:58-66.

101. Mauch P, Ferrara J, Hellman S. Stem cell self-renewalconsiderations in bone marrow transplantation. BoneMarrow Transplant 1989; 4:601-7.

102. Kotasek D, Sheperd KM, Sage RE, et al. Factors affect-ing blood stem cell collections following high-dosecyclophosphamide mobilization in lymphoma, myelo-ma and solid tumors. Bone Marrow Transplant 1992;9:11-7.

103. Haas R, Möhle R, Fruehauf S, et al. Patient character-istics associated with successful mobilizing and auto-grafting of peripheral blood progenitor cells in malig-nant lymphoma. Blood 1994; 83:3787-94.

104. Schneider JG, Crown JP, Wasserheit C, et al. Factorsaffecting the mobilization of primitive and committedhematopoietic progenitors into the peripheral bloodof cancer patients. Bone Marrow Transplant 1994;14:877-84.

105. Bensinger WI, Longin K, Appelbaum F, et al. Peripher-al blood stem cells (PBSCs) collected after recombi-nant granulocyte colony stimulating factor (rhG-CSF):an analysis of factors correlating with the tempo ofengraftment after transplantation. Br J Haematol1994; 87:825-31.

106. Laporte JP, Gorin NC, Dupuy-Montbrun MC, et al.Failure to collect sufficient amount of peripheral bloodstem cells (PBSC) for autografting in patients with end-stage multiple myeloma. Bone Marrow Transplant1988; 3(suppl 1):89.

107. Tarella C, Caracciolo D, Gavarotti P, et al. Circulatingprogenitors following high-dose sequential (HDS)chemotherapy with G-CSF: short intervals betweendrug courses severely impair progenitor mobilization.Bone Marrow Transplant 1995; 16:223-8.

108. Tarella C, Boccadoro M, Omedé P, et al. Role ofchemotherapy and GM-CSF on hemopoietic progeni-tor cell mobilization in multiple myeloma. Bone Mar-row Transplant 1993; 11: 271-7.

109. Dimopoulos MA, Hester J, Huh Y, Champlin R, Alex-anian R. Intensive therapy with blood progenitor trans-plantation for primary resistant multiple myeloma. BrJ Haematol 1994; 87:730-4.

110. Haas R, Möhle R, Murea S, et al. Characterization ofperipheral blood progenitor cells mobilized by cyto-toxic chemotherapy and recombinant human granu-locyte colony-stimulating factor. J Hematother 1994;3:323-30.

111. Ossenkoppele GJ, Jonkhoff AR, Huijgens PC, et al.Peripheral blood progenitors mobilised by G-CSF (fil-grastim) and reinfused as unprocessed autologouswhole blood shorten the pancytopenic period follow-ing high-dose melphalan in multiple myeloma. BoneMarrow Transplant 1994; 13:37-41.

112. Majolino I, Marcenò R, Buscemi F, et al. Mobilizationof circulating progenitor cells in multiple myeloma dur-ing VCAD therapy with or without rhG-CSF. Haema-tologica 1995; 80:108-14.

113. Vasta S, Majolino I, Morabito F, et al. The associationof vincristine, cyclophosphamide, epirubicin and dex-amethasone (VCED) followed by rhG-CSF may providea good PBSC mobilization in patients with multiplemyeloma. Submitted for publication.

114. Kessinger A, Armitage JO, Smith DM, et al. High-dose


47


therapy and autologous peripheral blood stem celltransplantation for patients with lymphoma. Blood1989; 74:1260-5.

115. Hénon P, Beck G, Debecker A, Eisenman JC, Lepers M,Kandel G. Autograft using peripheral blood stem cellscollected after high-dose melphalan in high-risk multi-ple myeloma. Br J Haematol 1988; 70:254-5.

116. Fermand JP, Levy Y, Gerota J, et al. Treatment of aggres-sive multiple myeloma by high-dose chemotherapy andtotal body irradiation followed by blood stem cellsautologous graft. Blood 1989; 73:20-3.

117. Reiffers J, Marit G, Boiron JM. Peripheral blood stem-cell transplantation in intensive treatment of multiplemyeloma. Lancet 1989; ii:1336.

118. Indovina A, Majolino I, Scimé R, et al. High dosecyclophosphamide: stem cell mobilizing capacity in 21patients. Leuk Lymphoma 1994; 14:71-7.

119. Duhrsen V, Villeval JL, Boyd J, Kannourakis G, MoerstynG, Metcalf D. Effects of recombinant human granulo-cyte colony-stimulating factor on hematopoietic prog-enitor cells in cancer patients. Blood 1988; 72:2074-81.

120. Bensinger W, Singer J, Appelbaum F, et al. Autologoustransplantation with peripheral blood mononuclearcells collected after administration of recombinantgranulocyte stimulating factor. Blood 1993; 81:3158-63.

121. Majolino I, Buscemi F, Scimè R, et al. Treatment ofnormal donors with rhG-CSF 16 µg/kg for mobiliza-tion of peripheral blood stem cells and their aphereticcollection for allogeneic transplantation. Haemato-logica 1995; 80:219-26.

122. Socinski MA, Cannistra SA, Elias A, Antman KH,Schnipper L, Griffin JD. Granulocyte-macrophagecolony stimulating factor expands the circulatinghaemopoietic progenitor cell compartment in man.Lancet 1988; i:1194-8.

123. Brugger W, Bross KJ, Glatt M, Weber F, MertelsmannR, Kanz L. Mobilization of tumor cells and hematopoi-etic progenitor cells into the peripheral blood ofpatients with solid tumors. Blood 1994; 83:636-40.

124. Gianni AM, Bregni M, Siena S, et al. Granulocyte-macrophage colony-stimulating factor or granulocytecolony-stimulating factor infusion makes high-doseetoposide a safe outpatient regimen that is effective inlymphoma and myeloma patients. J Clin Oncol 1992;10:1955-62.


126. Smith R, Sweetenham JW. A mononuclear cell dose of3x108/kg is sufficient to predict early multilineageengraftment in patients undergoing high dose therapyand transplantation with cryopreserved peripheralblood progenitor cells for malignant lymphoma. Blood1994; 84 (Suppl. 1):364.

127. Schwartzberg L, Birch R, Blanco R, et al. Rapid andsustained hematopoietic reconstitution by peripheralblood stem cell infusion alone following high-dosechemotherapy. Bone Marrow Transplant 1993; 11:369-74.

128. Pierelli L, Iacone A, Quaglietta AM, et al. Haemopoi-etic reconstitution after autologous blood stem celltransplantation in patients with malignancies: a mul-ticentre retrospective study. Br J Haematol 1994; 86:70-5.

129. Buscemi F, Indovina A, Scimè R, et al. CD34+ cell sub-sets and platelet recovery after PBSC autograft. BoneMarrow Transplant, in press.

130. Schiller G, Rosen L, Vescio R, et al. Threshold dose ofautologous CD34 positive peripheral blood progenitor

cells required for engraftment after myeloablative treat-ment for multiple myeloma. Blood 1994;84 (suppl 1):207.

131. Zimmerman TM, Lee WJ, Bender JG, Mick R, WilliamsSF. Quantitative CD34 analysis may be used to guideperipheral blood stem cell harvests. Bone MarrowTransplant 1995; 9:439-44.

132. Indovina A, Majolino I, Buscemi F, et al. Engraftmentkinetics and long-term stability of hematopoiesis fol-lowing autografting of peripheral blood stem cells.Haematologica 1995; 80:115-22.

133. Siena S, Bregni M, Brando B, et al. Flow cytometry forclinical estimation of circulating hematopoietic prog-enitors for autologous transplantation in cancerpatients. Blood 1991; 77: 400-9.

134. Hénon PR, Wunder E, Zingsem J, Lepers M, Siegert W,Eckstein R. Collection of peripheral blood stem cells.Apheresis monitoring and procedure. In: Wunder EW,Henon PR, eds. Peripheral blood stem cell autografts.Berlin:Springer-Verlag, 1993:185-93.

135. Buscemi F, Fabbiano F, Felice R, et al. Use of largepolygonal contiguous gates for flow cytometry analy-sis of circulating progenitor cells. Bone Marrow Trans-plant 1993; 12:305.

136. Rowley S. Analysis of the collected product. In:Kessinger A, McMannis JD, eds. Practical considera-tions of apheresis in peripheral blood stem cell trans-plantation. Lakewood, Co: Cobe BCT, Inc., 1994:35-51.

137. Legros M, Fleury J, Curé H, et al. New method for stemcell quantification: applications to the management ofperipheral blood stem cell transplantation. Bone Mar-row Transplant 1995; 15:1-8.

138. Mijovic A, Mufti GK. Single dose of filgrastim (rhG-CSF) to predict mobilization of hemopoietic progeni-tors in patients with hematologic malignancies. BoneMarrow Transplant 1995; 15:813-4.

139. Fruehauf S, Haas R, Conradt C, et al. Peripheral bloodprogenitor cell (PBPC) counts during steady-statehematopoiesis allow to estimate the yield of mobilizedPBPC after filgrastim (R-metHuG-CSF)-supportedcytotoxic chemotherapy. Blood 1995; 85:2619-26.

140. Roberts AW, Begley CG, Grigg AP, Basser RL. Dosteady-state peripheral blood progenitor cell (PBPC)counts predict the yield of PBPC mobilized by filgras-tim alone? Blood 1995; 86:2451.

141. Cavo M, Gozzetti A, Lemoli RM, et al. High-dosecyclophosphamide (7 g/m2) for the treatment of new-ly diagnosed and refractory multiple myeloma patients.Submitted for publication.

142. To LB. Is our current strategy in manipulating hemo-poiesis in autologous transplantation correct? StemCells 1993; 11: 283-9.

143. Berenson J, Wong R, Kim K, Brown N, Lichstentein A.Evidence of peripheral blood B lymphocyte but not Tlymphocyte involvement in multiple myeloma. Blood1987; 70: 1550-3.

144. Bergui L, Schena M, Gaidano GL, Riva M, Caligaris-Cappio F. Interleukin-3 and interleukin-6 synergistical-ly promote the proliferation and diffentiation of malig-nant plasma cell precursors in multiple myeloma. J ExpMed 1989; 170:613-8.

145. Pilarski LM, Jensen GS. Monoclonal circulating B cellsin multiple myeloma: a continuously differentiatingpossibly invasive population as defined by expressionof CD45 isoforms and adhesion molecules. HematolOncol Clin N Am 1992; 6:297-332.

146. Billadeau D, Quam L, Thomas W, et al. Detection andquantitation of malignant cells in the peripheral bloodof multiple myeloma patients. Blood 1992; 80:1818-24.


48


147. Pilarski LM, Belch AR. Circulating monoclonal B cellsexpressing P glycoprotein may be a reservoir of mul-tidrug-resistant disease in multiple myeloma. Blood1994; 83:724-36.

148. Bergsagel LP, Masellis Smith A, Szczepek A, Mant MJ,Belch AR, Pilarski LM. In multiple myeloma, clonotyp-ic B lymphocytes are detectable among CD19+ periph-eral blood cells expressing CD38, CD56, and mono-typic Ig light chain. Blood 1995; 85: 436-47.

149. Witzig T, Kyle R, O’Fallon W, Greipp P. Detection ofperipheral blood plasma cells as a predictor of diseasecourse in patients with smouldering multiple myelo-ma. Br J Haematol 1994; 87:266-72.

150. Grogan TM, Durie BGM, Lomen C, et al. Delineationof a novel pre-B cell component in plasma cell myelo-ma: immunochemical, immunophenotypic, genotyp-ic, cytologic, cell culture, and kinetic features. Blood1987; 70:932-42.

151. LeBien TW, McCormack RT. The common acutelymphoblastic leukemia antigen (CD10). Emancipa-tion from a functional enigma. Blood 1989; 73:625-35.

152. Pellat-Deceunyck C, Bataille R, Robillard N, et al.Expression of CD28 and CD40 in human myelomacells: a comparative study with normal plasma cells.Blood 1994; 84:2597-603.

153. Witzig TE, Gertz MA, Pineda AA, Kyle RA, Greipp PR.Detection of monoclonal plasma cells in the peripher-al blood stem cell harvests of patients with multiplemyeloma. Br J Haematol 1995; 89:640-2.

154. Vora A, Toh C, Peel J, Greaves M. Use of granulocytecolony-stimulating factor (G-CSF) for mobilizingperipheral blood stem cells: risk of mobilizing clonalmyeloma cells in patients with bone marrow infiltra-tion. Br J Haematol 1994; 86:180-2.

155. Gazitt Y, Reading C, Hoffman R, et al. Purified CD34+

Lin– Thy+ stem cells do not contain clonal myelomacells. Blood 1995; 86:381-9.

156 Gazitt Y, Tian E, Barlogie B, et al. Differential mobi-lization of myeloma cells and normal hematopoieticstem cells in multiple myeloma after treatment withcyclophosphamide and granulocyte-macrophage-colony-stimulating-factor. Blood, 1996; 87:805-11.

157. Björkstrand B, Ljungman P, Bird JM, Samson D,Garthon G. Double high-dose chemoradiotherapy withautologous stem cell transplantation can induce mol-ecular remissions in multiple myeloma. Bone MarrowTransplant 1995; 15:367-71.

158. Brenner MK, Rill DR, Moen RC, et al. Gene-marking totrace origin of relapse after autologous bone marrowtransplantation. Lancet 1993; 341:85-6.

159. Reece DE, Barnett MJ, Connors JM. Treatment of mul-tiple myeloma with intensive chemotherapy followedby autologous BMT using marrow purged with 4-hydroperoxycyclophosphamide. Bone Marrow Trans-plant 1993; 11:139-46.

160. Gulati SC, Shimazaki C, Lemoli RM, Atzpodien J, Clark-son BD. Ex vivo treatment of myeloma cells by 4-HC,VP-16, LAK cells and antibodies. Eur J Haematol 1989;43(Suppl. 51):164-72.

161. Siena S, Castro-Malaspina H, Gulati SC, et al. Effectsof in vitro purging with 4-hydroperoxycyclophos-phamide on the hematopoietic and microenvironmentelements of human bone marrow. Blood 1985; 65:655-62.

162. Bullock G, Tang C, Tourkina E, et al. Effect of com-

bined treatment with interleukin-3 and interleukin-6on 4-hydroperoxycyclophosphamide induced pro-grammed cell death or apoptosis in human myeloidleukemia cells. Exp Hematol 1993; 21:1640-4.

163. Skorski T, Kawalec M, Hoser G, Ratajczcac M, Gna-towski B, Kawiak J. The kinetic of immunologic andhematologic recovery in mice after lethal total bodyirradiation and reconstitution with syngeneic bonemarrow cells treated or untreated with mafosphamide(Asta Z 7654). Bone Marrow Transplant 1988; 3:543-51.

164. Tong AW, Lee JC, Fay JW, Stone MJ. Elimination ofclonogenic stem cells from human multiple myelomacell lines by a plasma cell-reactive monoclonal anti-body and complement. Blood 1987; 70:1482-9.

165. Anderson KC, Andersen J, Soiffer R, et al. Monoclonalantibody-purged bone marrow transplantation thera-py for multiple myeloma. Blood 1993; 82:2568-76.

166. Gobbi M, Cavo M, Tazzari PL, et al. Autologous bonemarrow transplantation with immunotoxin-purgedmarrow for advanced multiple myeloma. Eur J Haema-tol 1989; 43 (Suppl. 51):176-81.

167. Goldmacher VS, Bourret LA, Levine BA, et al. Anti-CD38-blocked ricin: an immunotoxin for the treatmentof multiple myeloma. Blood 1994; 84:3017-25.

168. Lemoli RM, Gobbi M, Tazzari PL, et al. Bone marrowpurging for multiple myeloma by avidin-biotinimmunoadsorption. Transplantation 1989; 47:385-7.

169. Carlo Stella C, Cazzola M, De Fabritiis P, et al. CD34-positive cells: biology and clinical relevance. Haema-tologica 1995; 80: 367-87.

170. Visani G, Lemoli RM, Dinota A, et al. Evidence thatlong-term bone marrow culture of patients with mul-tiple myeloma favors normal hemopoietic prolifera-tion. Transplantation 1989; 48:1026-31.

171. Vescio RA, Hong CH, Cao J, et al. The hematopoieticstem cell antigen, CD34, is not expressed on the malig-nant cells in multiple myeloma. Blood 1994; 84:3283-90.

172. Takishita M, Kosaka M, Goto T, Saito S. Cellular ori-gin and extent of clonal involvement in multiple myelo-ma: genetic and phenotypic studies. Br J Haematol1994; 87:735-42.

173. Avvisati G, Petrucci MT, Mandelli F. The role of bio-therapies (interleukins, interferons and erythropoietin)in multiple myeloma. Baillière Clinical Haematol 1995;8:815-29.

174. Mandelli F, Avvisati G, Amadori S, et sl. Maintenancetreatment with recombinant interferon alpha-2b inpatients with multiple myeloma responding to con-ventional induction chemotherapy. N Engl J Med 1990;322:1430-4.

175. Westin J, Rodjer S, Turesson I et Al. Interferon alpha-2b versus no maintenance therapy during the plateauphase in multiple myeloma: a randomized study. Br JHaematol 1995; 89:561-8.

176. Powels R, Cunningham D, Malpas JS, et al. A ran-domized trial of maintenance therapy with Intron-Afollowing high-dose melphalan and ABMT in myeloma.Blood 1994; 84 (Suppl. 1):535a.

177. De Laurenzi A, Iudicone P, Zoli V, et al. Recombinantinterleukin-2 treatment before and after autologousstem cell transplantation in hematologic malignancies:clinical and immunologic effects. J Hematother 1995;4:113-20.



review

Allogeneic hematopoietic stem cellsfrom sources other than bone marrow:biological and technical aspects


FRANCESCO BERTOLINI, ARMANDO DE VINCENTIIS, LUIGI LANATA,ROBERTO M. LEMOLI, RITA MACCARIO, IGNAZIO MAJOLINO,LUISA PONCHIO, DAMIANO RONDELLI, ANTONIO TABILIO, PAOLA ZANON, SANTE TURA

Division of Oncology, IRCCS Fondazione Maugeri, Pavia;Dompé Biotec SpA, Milan; Amgen Italia SpA, Milan; Istitutodi Ematologia ed Oncologia Medica “Seràgnoli”, University ofBologna, Bologna; Department of Pediatrics, IRCCS Policlini-co S. Matteo di Pavia, Pavia; Department of Hematology andBMT Unit, Ospedale “V. Cervello”, Palermo; Department ofInternal Medicine, University of Pavia and IRCCS Policlinico S.Matteo, Pavia; Institute of Hematology, University of Perugia,Italy

Background and Objectives. Identification and charac-terization of hematopoietic stem cells in peripheralblood (PB) and cord blood (CB) have suggested feasi-ble alternatives to conventional allogeneic bone mar-row (BM) transplantation. The growing interest in thisuse of allogeneic stem cells has prompted the WorkingGroup on CD34-positive Hematopoietic Cells to reviewthis subject by analyzing its biological and technicalaspects.

Evidence and Information Sources. The method usedfor preparing this review was informal consensus devel-opment. Members of the Working Group met threetimes, and the participants at these meetings exam-ined a list of problems previously prepared by thechairman. They discussed the individual points in orderto reach an agreement on the various concepts, andeventually approved the final manuscript. Some of theauthors of the present review have been working in thefield of hematopoietic stem cell biology and process-ing, and have contributed original papers published inpeer-reviewed journals. In addition, the material exam-ined in the present review includes articles andabstracts published in journals covered by the ScienceCitation Index® and Medline®.

State of Art. Several studies have now shown thathematopoietic stem cells collected from peripheralblood after the administration of G-CSF, or from cordblood upon delivery, are capable of supporting rapidand complete reconstitution of BM function in allo-geneic recipients. Perhaps more importantly, reinfusionof large numbers of HLA-matched T-cells from PB col-lections or T-cells with various degrees of HLA disparityfrom CB did not result in a higher incidence or greaterseverity of acute graft-versus-host disease than expect-ed with BM. Based on the data reviewed, operativeguidelines for mobilization, collection and graft pro-cessing are provided.

Perspectives. It should be remembered that despitethe growing interest, these procedures must be stillconsidered as advanced clinical research and shouldbe included in formal clinical trials aimed at demon-strating their definitive role in stem cell transplantation.In this regard, a large European randomized study is

currently comparing PB and BM allografts. However,the possibility of collecting large quantities ofhematopoietic progenitor-stem cells, perhaps withreduced allo-reactivity, offers an exciting perspectivefor widening the number of potential stem cell donorsand greater leeway for graft manipulation than is possi-ble with BM.©1997, Ferrata Storti Foundation

Key words: hematopoietic stem cells, bone marrow, cord blood,peripheral blood, allogeneic transplantation, graft-versus-host dis-ease

Allogeneic bone marrow transplantation has pro-gressed from a highly experimental procedure tobeing accepted as the preferred form of treat-

ment for a wide variety of diseases.1 There have beenimpressive improvements in this therapeutic procedurein the last two decades, but the most importantadvances probably took place in the last few years andconcern the source of hematopoietic stem cells itself.Whereas this had always been by definition the bonemarrow since the very beginning, identification of stemcells in peripheral and cord blood has now provideduseful alternatives.

In 1994 the growing interest in the use of peripher-al blood stem cells (PBSC) in the setting of allogeneicbone marrow transplantation induced the GITMO(Gruppo Italiano Trapianto di Midollo Osseo) to promotea Study Committee for evaluating the key aspects ofallogeneic PBSC collection and transplantation. ThisCommittee produced a list of recommendations thatwere published as a position paper in this Journal at thebeginning of 1995.2 In summary, the authors stronglyrecommended the use of allogeneic PBSC in experi-enced centers, in well-defined clinical settings, and pos-sibly – for the time being – in patients with advanceddisease.

As the use of PBSC expanded both in the autologousand the allogeneic setting, expression of the CD34 anti-gen became increasingly important for their character-

Correspondence: Prof. Sante Tura, Istituto di Ematologia ed OncologiaMedica “Seràgnoli”, Policlinico S. Orsola, via Massarenti 9, 40138Bologna, Italy. Received August 22, 1996; accepted January 20,1997.

50


F. Bertolini et al.

ization. An ad hoc working group reviewed the biologyand clinical relevance of CD34-positive cells in this jour-nal in 1995.3 In particular, techniques for CD34-positivecell separation and procedures for their collection fromperipheral blood were analyzed. A brief chapter was alsodevoted to CD34-positive cells in cord blood.3

The working group on CD34-positive hematopoieticcells subsequently reviewed the use of PBSC in acutemyeloid leukemia4 and multiple myeloma.5 The growinginterest in the use of PBSC and cord blood stem cells inthe setting of allogeneic transplantation has nowprompted the working group to review this subject byanalyzing its biological and technical aspects.

PBSC mobilization and collection innormal donors

Until recently, the collection of hematopoietic cells forallogeneic transplantation has required general or spinalanesthesia and multiple punctures of iliac bones. How-ever, marrow harvesting is not completely devoid ofcomplications, side effects or patient discomfort. In areport on 1270 harvest procedures in Seattle,6 6 donorssuffered life-threatening complications and 10 showedsignificant operative site morbidity. As many as 10 per-cent of donations were associated with fever, andincreasing donor age was significantly linked to poorcell harvest. In a different survey, 10 percent of donorsrecovered completely from marrow donation only morethan 30 days after the procedure.7

PBSC transplantation represents an alternativeapproach. In autologous transplantation peripheralblood is now replacing bone marrow as a source of prog-enitor cells.8 The advantage is quicker hematopoieticrecovery9,10 with consequently fewer complications andshorter hospital stay.

In the autologous setting, PBSC can be collected aftermobilization with chemotherapy,11,12 growth factors,13

or a combination of the two.14 In a randomized study,leukaphereses created less anxiety and pain than bonemarrow harvest.15

In allogeneic transplantation, the use of PBSC hasbeen somewhat delayed by a possible increase in graft-versus-host disease (GVHD) as a consequence of themuch higher number of lymphocytes in the graft inocu-lum, and by the need for a mobilization treatment forhealthy individuals in order to obtain a good cell yield.However, the clinical experience of the last two yearssuggests that the incidence of acute GVHD is notincreased with PBSC as compared to marrow, and thatin healthy donors a sufficient cell number can beobtained by using growth factors alone, in particular G-CSF.16-20 As a consequence, the number of allogeneicPBSC transplants is increasing rapidly. The EuropeanBlood and Marrow Transplant Group (EBMT) registeredonly 12 PBSC allografts in 1993, but their numberincreased to 180 in 1994 and to 537 in 1995 (Gratwohl,personal communication).

Collection of PBSC in normal donorsOn biological grounds, there are several means of

mobilizing progenitor cells into the peripheral blood, buttheir ultimate modality of action is always detachmentof the CD34+ progenitor cell from marrow stroma andendothelium, to which it is normally bound by interac-tions with different integrin-adhesion molecules.3,21 Wemay induce detachment either by an inhibition of thelink between CD34+ cells and stroma, or by inducing astress to the hematopoietic system capable of favoringthe egress of progenitor cells from marrow to circula-tion. The former is obtained by means of monoclonalantibodies directed against adhesion molecules,22 whilethe latter is based on the use of a drug or a combina-tion of drugs. Richman et al.23 demonstrated for the firsttime in man that chemotherapy-induced cytopenia isfollowed by a substantial increase of CFU-C in blood.

In normal donors, however, the use of chemotherapyis ethically unacceptable, and only growth factors mustbe employed. Though a number of other cytokines areable to induce an increase of PBSC, only G-CSF and GM-CSF have been utilized in clinical practice. G-CSF in par-ticular has an excellent mobilizing effect when usedalone.13,24-29

The pilot experience with stem cell mobilization innormal donors is the one reported by the Seattle group.They administered G-CSF 300 µg/day or 6 µg/kg/day toincrease WBC levels in apheresis collections from gran-ulocyte donors.30 A number of different schedules werelater applied to mobilize PBSC for allogeneic transplan-tation. The results in terms of CD34+ cell collection arereported in Table 1. In most of the studies the G-CSFdose ranged from 10 to 16 µg/kg/day. With 16 µg/kg/dayfor 5 days, Weaver et al.31 collected 1.6 to 12.6 (median9.6)×106/kg CD34+ cells with two aphereses. All trans-plants were syngeneic, and recovery of 0.5×109/L gran-ulocytes and 20×109/L platelets occurred on day 13 and10, respectively. With the same dose administered for 4days, Majolino et al.27 were able to mobilize (median)147×106/L CD34+ cells on day 4, a 65-fold increase overthe baseline level. The median collection was 754×106

Authors (ref.#) DonorsNo.

G-CSF,dose/kg and

days ofadministration

CD34+ collected

x 106

ApheresisNo.

Weaver (31) 4 16 µg, 4 d 9.6/kg 2

Korbling (16) 9 12 µg, 7 d 13.1/kg 3

Bensinger (18) 8 16 µg, 6 d 13.1/kg 2

Schmitz (17) 8 5-10 µg, 5-6 d 6.7/kg 1-3

Russell (34) 9 6-8 µg, 2-4 d 4.7/kg 1-2

Majolino (27) 56

10 µg, 5 d16 µg, 4 d

754789

22

Tabilio (1996) 39 12 µg, 4-7 d 132.6 2-4

Table 1. Relationship between CD34+ cell yield and G-CSFdose in allogeneic PBSC donors. Only clinical experiencesare reported.

51


CD34+ cells and 270×108 CD3+ cells with 2 aphereses.With 12 µg/kg for 6 days, Körbling et al.16 collected amean of 13.1×106/kg CD34+ cells with 3 aphereses, andtheir patients recovered >0.53109/L granulocytes on day10 and >20×109/L platelets on day 14. However, theirshort recovery times were also influenced by the absenceof methotrexate from GVHD prophylaxis.

The relationship between G-CSF dose and CD34+ cellmobilization is supported in part by the study by Dregeret al.,26 who compared 5 µg/kg/day and 10 µg/kg/day G-CSF in normal volunteers. They found 10 µg/kg to besuperior in terms of progenitor cell yield. With higherdoses the advantage seems to decline, and no statisti-cal difference was found between 10 µg/kg for 5 daysand 16 µg/kg for 4 days in a retrospective non-ran-domized study (Figure 1).32 Dührsen33 has suggested thatthe maximal effect in terms of progenitor cell increaseis that obtained at a dose level of 10 µg/kg/day. This isalso the dose recommended by the GITMO in its recent-ly published guidelines.2

The number of apheretic procedures necessary for agood collection is critical for the donor, and may varywith the dose and schedule of G-CSF as well as with thevolume processed.

Bensinger et al.28 routinely employ a schedule of 16µg/kg/day for 5-6 days in an effort to minimize thenumber of apheretic procedures. With this dose, a medi-an of approximately 7×106/kg CD34+ cells are obtainedwith a single apheresis performed on day 5. At the M.D.Anderson Cancer Center in Houston29 a schedule of 12µg/kg/day for 4-6 days is used. With a single large vol-ume apheresis the target CD34+ cell dose of > 4×106/kgis reached in nearly 80% of the donors. Russell et al.34

mobilized their donors with 6-8 µg/kg/day for 2-4 days.By daily monitoring of CD34+ levels, the target of2.5×106/kg CD34+ cells was achieved with a single 2-4hour harvest in 12 out of 14 donors. With 24 µg/kg/day

G-CSF for 4 days Waller et al.35 were able to collect13×106/kg CD34+ with a single apheresis; however, onedonor suffered severe side effects and the G-CSF dosehad to be halved.

The mobilization kinetics of PBSC under low daily dos-es of G-CSF has also been investigated. With 2.5µg/kg/day G-CSF on days 1 to 6 followed by 5.0µg/kg/day on days 7 to 10, a CFU-GM peak was obtainedon day 6, but continuing G-CSF administration at 5µg/kg/day did not increase the level of circulating CFU-GM.36

With a single G-CSF dose of 15 µg/kg a significant risein CD34+ cells, CFU-GM and BFU-E was obtained,37 butthe reported counts of 250/mL, 3.2×103/mL and1.75×103/mL, respectively, are not comparable withthose obtained with prolonged administration sched-ules. Bishop et al.38 reported their experience with G-CSFat 5 µg/kg/day. Aphereses began on day 4 of G-CSFadministration. However, the target cell CD34+ dose of>1×106/kg required 3 to 4 aphereses. With this method,median time to ANC > 0.5×109/L was 12 days but allpatients received G-CSF after the allograft.

With G-CSF doses ranging from 10 to 16 µg/kg/day,the progenitor cell peak occurs on day 4 or 5.2,20,27,28,32,39

Since the CD34+ cell level rapidly declines after growthfactor withdrawal, it is highly recommended that itsadministration be continued till the end of aphereticharvests.

In both the Seattle and the GITMO experiences, theWBC peak occurred approximately the same day as theCD34+ cell peak. In the GITMO study,39 the level of CD34+

cells reached a peak of (mean) 135.5×106/L CD34+ cells,a 19-fold increase over the mean baseline level.

Lymphocytes also increased, doubling their counts onday 5. A number of CD34+ cells >4×106/kg was collect-ed in 51% of donors with a single apheresis, in 85%with two. Optimal collections are obtained on days 4and 5 of G-CSF administration.28,39 It is likely that start-ing PBSC collection on day 4 is best when using 16µg/kg/day, whereas day 5 is better when lower doses areemployed (Figure 2).

In normal volunteers GM-CSF has found applicationless frequently than G-CSF. Lane et al.40 studied G- andGM-CSF alone and a combination of the two. The totalnumber of CD34+ cells collected from the G-CSF groupwith a single apheresis was 119×106, and was not sig-nificantly different from that collected from the grouptreated with G- and GM-CSF (101×106 CD34+ cells), butboth were greater than that from the group treated withGM-CSF (12.6×106). However, a higher fraction of anearly CD34+/HLA-DR–/CD38– cell population was foundamong the CD34+ cells after GM-CSF administration.Whether this early fraction is associated with more rapidengraftment is presently unknown.

Predictive factors for progenitor cell yield have notbeen studied in normal volunteers. Though there is anec-dotal experience of donors failing to respond, only agewas reported to influence the quality of collections in asingle study.26

Non bone marrow allogeneic hemopoietic stem cells

Figure 1. Schematic representation of timing in PBSC mobi-lization in healthy donors. G-CSF is administered at a dailydose of 10 µg/kg, apheretic harvest is performed on day 5(and 6). Day 5 collection cells are stored at 4°C till the fol-lowing day, when they are infused together with day 6 cells.Parentheses indicate that G-CSF is given and apheresesperformed only if the target number of CD34+ cells is notreached with the day 5 apheretic run.

52


The number of PBSC necessary for rapid and stableengraftment is unknown. In the autologous setting adose of >2×106/kg CD34+ cells has been suggested,41

but the requirement might be higher in allogeneic trans-plantations as a consequence of the immunologicalmechanisms involved. In Seattle28 4 out of the 53 nor-mal donors yielded only 0.6, 1.49, 1.55 and 1.74×106

CD34+ cells/kg. Despite the low cell numbers, success-ful engraftment was achieved in all cases. We suggestthat collection of >4×106/kg CD34+ cells is the target fora safe allogeneic transplantation. Lower doses, howev-er, may be sufficient. A lower limit of >23106/kg CD34+

cells would be reasonable for those patients whosedonors respond poorly to cell mobilization.

There are currently no contraindications to cryop-reservation of cells for later use after thawing. Thoughmost centers currently infuse freshly collected aphere-sis products, one may consider the advantage of sepa-rating the mobilization/collection phase from trans-plantation in terms of logistics and patient safety.

Side effects and toxicity of the procedureEarly toxic effects of G-CSF in healthy donors are now

well known. The GITMO survey39 on 76 healthy subjectsaged 6 to 67 years receiving G-CSF for PBSC mobiliza-tion reveals that the side effects of G-CSF administra-tion are acceptable, the only problem being moderate tosevere bone pain in 13% of donors (Table 2).

Twenty-three percent of donors also said theapheretic procedures were demanding. Comparable sideeffects are reported in other studies.34,28 Additional prob-lems could include pneumothorax due to jugular veincannulation and paresthesia.34 Nonetheless, donors whohad previously given marrow mostly agreed that theypreferred blood cell mobilization and collection to mar-row harvest.34-39 A good policy would be to avoid the useof venous catheters. In autologous PBSC harvestingwhere mobilization treatment often includes chemo-therapy, central venous catheter (CVC) occlusion neces-sitating thrombolytic therapy was the most commonly

observed complication, occurring in 15.9% of CVC-aid-ed collections.42

Variations in blood counts mainly consist of a pro-nounced WBC increase, a moderate thrombocytopeniaand a slight decrease of hematocrit values. In the Ital-ian survey, thrombocytopenia from mild (< 70×109/L) tomoderate (< 50×109/L) followed PBSC harvests in 40%and 10% of cases, respectively. WBC counts exceeded50×109/L in 40% of cases, and 70×109/L in 8%. Bensingeret al.28 report their experience with 124 donors treatedwith G-CSF at various doses and scheduling. Forty-onewere granulocyte donors, while 13 were PBSC syngene-ic and 63 allogeneic donors. One donor had a myocar-dial infarction after the first apheresis, but he had a pre-vious history of infarction. Thrombocytopenia was in partrelated to G-CSF dosage, in part to the volume of bloodprocessed. A count <100×109/L never occurred in gran-ulocyte donors receiving 4 to 12 µg/kg/day and multipleaphereses.

With higher doses of G-CSF, thrombocytopeniaoccurred in 5% of donors undergoing 1-2 aphereses andin 100% of those collected for 4 days. When the 4-daycollection donors received their platelets back by a sec-ond spin of the apheretic product, the incidence ofthrombocytopenia fell to 25%.

Biochemical abnormalities follow G-CSF administra-tion and consist primarily of mild elevation of ALT, LDHand alkaline phosphatase.39 This last is directly relatedto the action of G-CSF on the granulocytic lineage.43

These abnormalities have no clinical effects.At present, we have little data concerning the late

effects of G-CSF administration in healthy donors;44

however, the growing interest in PBSC allogeneic trans-plantation causes the need for prospective studies onthe donor population. This kind of study is difficult forstatistical reasons. Recently, Hasenclever and Sextro45

presented a preliminary study on long-term risks ofgrowth factor administration to healthy donors. In orderto demonstrate a tenfold increase in leukemia risk, morethan 2000 healthy PBSC donors should be followed forover 10 years. A control group of BMT donors of equalsize would be necessary. Such a study can only be doneon a multi-national basis. However, it is mandatory tofollow the PBSC donors regularly, and to register care-fully any variation in their blood counts.

F. Bertolini et al.

% donors

Absent Mild Moderate Severe

Bone pain 27.4 59.6 9.6 3.2

Arthralgias 54.8 32.2 11.2 1.6

Headache 70.9 20.9 8 0

Fatigue 74.5 22 3.3 0

Fever 91.5 6.7 1.6 0

Table 2. Incidence and grading of side effects reported dur-ing G-CSF administration in 76 healthy donors from the GIT-MO.39

Figure 2. Variations of blood cell counts in normal donorsduring G-CSF treatment and apheretic collection of mononu-clear cells. The curves represent mean values. Data arethose of 76 normal donors.39

53


The studies on leukemia development after G-CSFtreatment must be considered with caution. A Japanesegroup46 reported data on 170 children with aplastic ane-mia. Eleven out of the 108 receiving G-CSF had a trans-formation to MDS or leukemia, while this evolutionoccurred in none of the 62 patients not receiving G-CSF. Another study47 reports the evolution toMDS/leukemia in 13 patients with congenital neu-tropenia treated with G-CSF, with the occurrence ofmonosomy 7 in 10 of them. However, as suggested bySmith et al.48 in a study on the leukemic evolution ofKostmann’s disease, the fact that congenital neutrope-nia may evolve into MDS and AML under G-CSF treat-ment has no implication for normal donors, since it isthe underlying hemopoietic defect that represents a pre-leukemic condition. In fact, in Kostmann’s disease not allthe chromosomal aberrations involve chromosome 7,and when other abnormalities are detected leukemiadoes not develop.

The use of G-CSF for mobilization of PBSC in childrenshould be considered with more attention. Though theremay be a specific advantage in collecting PBSC fromchildren in the case of considerable disparity in bodyweight with the recipient, the GITMO stated that sucha practice should be discouraged in standard allogene-ic transplants.2 This is also the opinion of the ItalianAssociation of Pediatric Hemato-Oncology (AIEOP).

PBSC have also been employed for allogeneic engraft-ment in MUD transplants. A small series was presentedby Ringdén et al.49 in Geneva. Six patients with high riskhematological malignancies received PBSC from unre-lated donors, 4 of them as primary treatment and 2 fortreatment of graft failure. For PBSC mobilization thedonors received G-CSF 5 to 12 µg/kg and leukaphereseswere performed using continuous flow devices. Onedonor complained of rib pain and one of nausea, dizzi-ness and anxiety.

One advantage of using PBSC for MUD transplantscould derive from the higher number of progenitor cells,with better engraftment and reduction of failures. Forthe donor, the chance of obtaining stem cells for unre-lated transplants without the need for general anesthe-sia is certainly appealing, and would probably encouragemore volunteers to donate stem cells. It would also beeasier to expand especially the number of donors belong-ing to ethnic minorities. Apheresis-derived mononuclearcells might be stored in liquid nitrogen and shipped whenneeded. Age limit for donors could be expanded. How-ever, because of the limited experience with G-CSFmobilization in normal donors, National Marrow DonorRegistries have not approved the use of PBSC as firstchoice. We expect this will remain the case in the fore-seeable future.

In general, the use of growth factors for any purposein healthy subjects should be considered experimental.The donor should be informed of the potential short andlong-term risks of growth factors and leukapheresis, aswell as of anesthesia, and he should be given the pos-sibility of choosing. Donor consent should be asked on

the basis of a protocol previously approved by an offi-cial ethical committee.2

Characterization of CD34+ hematopoieticprogenitor cells mobilized into peripheralblood of normal donors by rHG-CSF

The preliminary results of clinical trials on allogeneicPBSC transplantation have demonstrated the capacity ofG-CSF to mobilize true stem cells capable of long-termreconstitution of marrow function. Moreover, similarlyto autografting, the most striking finding of PBSC trans-plantation has been the faster recovery of hematopoiesisafter myeloablative conditioning regimens as comparedto transplantation of BM-derived stem cells. Thus, clin-ical investigators asked the question of whether circu-lating progenitor cells may differ from their BM coun-terparts with respect to kinetic status, immunopheno-type, frequency of both committed and primitive pre-cursors, and their proliferative response to colony stim-ulating factors (CSFs).

One early report50 has shown a high expression ofmyeloid antigens on PB CD34+ cells (i.e. CD33, CD13) atthe expense of B-lineage-associated antigens (i.e. CD10,CD19, CD20), coupled with a high colony-formingcapacity of G-CSF-stimulated apheresis products. More-over, Roberts and Metcalf51 have clearly shown in ananimal model and in humans that only a small minori-ty of mobilized PBSC undergo active DNA synthesis,whereas BM cells contain more than 30% of S-phaseclonogenic progenitors. This finding, coupled with thelack of expression of CD71 antigen (transferrin receptor)and the Rhodamine 123 dull status52 observed in CD34+

cells from cancer patients mobilized with G-CSF, hassuggested that PB progenitors may be functionally inac-tive since they are in deep G0-phase of the cell cycle.

However, these results are somewhat in contrast withclinical data indicating rapid BM recovery after autolo-gous and allogeneic transplantation and the experi-mental evidence that circulating CD34+ cells representan optimal target for efficient retroviral infection requir-ing cell cycling for integration.53 In addition, it is veryimportant to assess the kinetic profile of the CD34+ cellfractions which are believed to ensure permanentengraftment after PBSC allografting, such as cells phe-notypically identified as CD34+/CD38–, CD34+/CD33–/HLA-DR– or very primitive progenitor cells capable ofgenerating clonogenic precursors in secondary semisol-id assay after 5 or more weeks of liquid culture, long-term culture-initiating cells (LTC-IC). In this regard,defective long-term repopulating activity of early BMcells induced to S-phase by cytokines has recently beenshown.54

To further elucidate the phenotypic profile and func-tional and kinetic characteristics of G-CSF-mobilizedhematopoietic progenitor cells, highly purified CD34+

cells from the apheresis products of normal individualsundergoing PBSC collections for allogeneic transplanta-tion were recently analyzed. The results were then com-


54


pared with those obtained on CD34+ cells enriched fromthe BM of the same donors under steady-state condi-tions and after G-CSF administration on the same day asPBSC harvest.55 The results confirmed the expression ofCD33 and CD13 antigens on a higher percentage of cir-culating CD34+ cells compared to BM cells (91±31% SDand 85.3±10% SD versus 51.1±21% SD and 64.6±25%SD, respectively; p <0.05) and the significantly lowerexpression of the B-cell associated antigen CD19(1.3±0.9% SD in PB and 12.4±12% SD in BM). Howev-er, a small but consistent proportion of very immatureCD34+/CD38– and CD34+/HLA-DR– cells was readily iden-tified in PB that was no different from BM-derived cells.When we compared primed PB CD34+ cells with those ofsteady-state BM, we reported the same frequency ofcolony-forming unit cells (CFU-C). However, bothmyeloid (CFU-GM) and erythroid (BFU-E) circulating pre-cursors showed increased responsiveness to singlegrowth factors (e.g. IL-3) or combinations of G-CSF/SCFor IL-3/SCF. Analysis of cell-cycle distribution of PB andBM CD34+ cells (Figure 3) demonstrated a negligible pro-portion of mobilized CD34+ cells in S/G2M phase. How-ever, the vast majority of circulating CD34+ cells werefound to be actually cycling, being in G1-phase with atendency, although not statistically significant, towardthe recruitment of primed CD34+ cells out of G0-phase.Moreover, it was observed that G-CSF treatment pro-vided CD34+ cells with a little, yet significant, protectionfrom programmed cell death.

Functional characterization of G-CSFmobilized primitive cells

Using the LTC-IC assay, which allows the detection ofvery primitive progenitors, it was found that PBSC gen-erate a higher number of CFC after 5 and 8 weeks oflong-term culture than their bone marrow or cord bloodcounterparts. Also, the frequency of 5-week-old cob-blestone area-forming cells (CAFC), a surrogate of LTC-IC measurement, within PBSC is similar at week 5 tothat of BM and cord blood and higher than the fre-

quency in the latter tissues at week 8.56 This suggeststhat PBSC contain either an adequate (or even a high-er) number of primitive progenitors (on a per cell basis)or a higher number of very primitive and therefore verypotent cells able to give rise to a high number of daugh-ter cells.

The leukapheresis product in fact is enriched in cellswith a very primitive phenotype, i.e. CD34+ Lin– Thy-1+,and contain cells able to repopulate SCID-hu mice, thatrepresent an in vivo model for studying the hematopoi-etic reconstitutive ability of a given population of cells.57

Nevertheless, even though in a cohort of heavily treat-ed cancer patients58 the number of LTC-IC was found tobe 2-10-fold higher after chemotherapy + GM-CSF thanin steady-state collections; a high interpatient variabil-ity was observed and the proliferative potential of mobi-lized LTC-IC (measured as the number of CFC producedby single LTC-IC) was lower than BM or steady-state cir-culating LTC-IC, suggesting that mobilized LTC-IC areless potent progenitors than their bone marrow andblood counterparts. No correlation was found betweenthe number of LTC-IC in the graft and the number ofCFC or CD34+ cells, or with the speed of engraftment. Allthese findings together show that chemotherapy +cytokine treatment allows the mobilization of progeni-tors with short- and long-term reconstitutive ability, butit is also evident that previous radiotherapy or stem cell-toxic drugs tend to significantly reduce the number ofCFC and LTC-IC that can be harvested by apheresis, eventhough they do not alter the ability of PBPC toengraft.58,59

Increasing lines of evidence suggest that the fasterengraftment after PBSC infusion might be related toboth the proliferative status of mobilized progenitorsand to the high number of committed progenitorsinfused. Both in the murine and in the human model,short-term G-CSF treatment increases the proportionof actively proliferating progenitor cells in the bonemarrow but not in the blood, where CFC and CD34+ cellsappear to be mostly in G1-phase but are easilyrecruitable into S-phase.55

To directly quantitate the proportion of cycling LTC-IC from the blood of cancer patients undergoingchemotherapy +G-CSF, mobilized CD34+ cells wereexposed to tritiated thymidine (3H-Tdr) in a 16-24-hrsuicide assay.60 At the end of the incubation periodaliquots of cells were tested for surviving progenitors ina LTC-IC assay. After 16 hrs of incubation in serum freemedium containing growth factors (Steel factor, G-CSFand IL-3) and in the absence of 3H-Tdr, the number LTC-IC remained at input level (panel A). A lower proportionof mobilized LTC-IC is initially quiescent in comparisonto normally circulating LTC-IC (% survival: 30±7, n=10,and 81±8, n=20, respectively), showing a cycling statusvery similar to that of BM LTC-IC (% survival: 21±6,n=11) (Figure 4). Similar data were found on PBSC col-lections from normal donors.55 In fact, similarly to moremature progenitor cells, very few circulating LTC-IC werefound in S-phase (1±3% SEM as compared to 21±8%

F. Bertolini et al.

Figure 3. Analysis of cell-cycle distribution of PB and BMCD34+ cells.

Cell-

cycl

e di

stri

butio

n of

CD

34+

cells

(%)

55


SEM of BM), whereas the proportion of LTC-IC cyclingwas superimposable on that of BM. The frequency wasnot different in the two compartments (48.2±35 SEMand 62.5±54 SEM for 104 CD34+ cells in PB and BM,respectively; p = ns). Thus these data, coupled with pre-vious observations from nonhuman primates on cell-cycle status and response to CSF of cytokine-mobilizedCD34+ cells,61 suggest that many circulating progeni-tors are not deeply quiescent in G0-phase. Rather, theyare actively cycling and their high clonogenic efficien-cy and prompt proliferative response to CSFs may indi-cate a faster progression through cell-cycle mediated,perhaps, by G-CSF priming. A kinetic and functional pat-tern of CD34+ cells similar to that observed in normalPBSC donors has been found in acute leukemia and mul-tiple myeloma patients mobilized with chemotherapyand G-CSF, and in lymphoma individuals receiving G-CSF alone (Lemoli, unpublished observations). Thus,regardless of the mobilization protocol, the administra-tion of G-CSF and/or the change of compartment (i.e.egress into peripheral blood) induces a profound effecton the characteristics of hematopoietic progenitor cells.Further studies are presently directed toward investi-gating the modulation of the expression of integrin

adhesion molecules critical for mobilization and relat-ed to cytokine-induced cell-cycle transit.62 Moreover,pharmacological doses of steel factor determine a redis-tribution of stem cells in mice63 and reduce the avidityof a4b1 and a5b1 integrins on the MO7e cell line, witha consequent inhibition of the specific cell adhesion ofMO7e cells to VCAM-164. Other adhesion molecules,like L-selectin and VLA4, might play a role in the mobi-lization of hematopoietic progenitors in primates.65

Stem cells from umbilical cord blood:biological aspects

More than 20 years ago it was described in this Jour-nal that hematopoietic progenitors circulate betweenthe fetus and the placenta during gestation.66 Howev-er, placental/umbilical cord blood (CB) from humannewborns was not considered as a source of stem cellsfor clinical use until Broxmeyer et al.67 enumerated thenumber of CFU-GM that could be collected from the CBremaining in the placenta after birth and suggested thatthe total number was sufficient for transplantation inpediatric patients. The Fanconi anemia patient who in1988 first received a CB transplant from his HLA-matched sibling68 is still alive at the present time andcured from the hematological point of view, thusdemonstrating the long-term engraftment capability ofCB-derived stem cells.

In the past five years interest in the biological aspectsand clinical applications of CB has grown since largeCB banks for unrelated stem cell transplantation havebeen implemented in the USA and Europe, and morethan 300 CB transplants have been performed. Howev-er, many aspects of the properties of CB stem cells arestill elusive. It is remarkable that a unit of CB used fortransplantation contains 1-83106 CD34+ cells and 10-1203103 CAFC/LTC-IC,56,69-70 i.e. 1-2 logs fewer than thetotal number of CD34+ cells and CAFC usually infusedinto recipients of allogeneic BM or PBSC. On average,recipients of CB transplants are given 0.05-0.5×106

CD34+ cells/kg b.w., while it has been suggested thatrecipients of allogeneic PBSC must receive at least 2.5-5×106 CD34+ cells/kg b.w. to obtain safe hematopoiet-ic engraftment.71 On the other hand, despite the delayin the reconstitution of the megakaryocytic lineage, therate of engraftment failure in CB transplant recipientsis similar to that observed after BM or PBSC trans-plants.72 These observations have prompted a number ofinvestigators to focus on the proliferation potential ofCB stem cells. In an elegant study, Lansdorp et al.73 sort-ed CB-derived, CD34+CD45RAloCD71lo cells, defined asstem cell candidates. In liquid cultures supplementedwith SCF, IL-3,-6 and Epo, these purified cells generat-ed a number of CD34+ and mature cells significantlygreater than that obtained in cultures of purified CD34+

CD45RAlo CD71lo cells obtained from adult donors. Thisadvantage was clearly ontogeny-related, since the pro-liferative potential of purified CD34+ CD45RAlo CD71lo

cells collected from fetal liver was superior to that ofCB-derived cells. In this context, Hows et al.74 demon-

Figure 4. Recovery and proliferative status of CFC and LTC-IC in steady-state normal blood and bone marrow and in theleukapheresis products of cancer patients obtained afterchemotherapy + G-CSF. The cells were cultured for 16 hrsin a medium containing serum substitutes, SF (100ng/mL), IL-3 (20 ng/mL) and G-CSF (20 ng/mL) in the pres-ence or absence of 3H-Tdr (panel B and A, respectively).


56


strated in long-term stroma culture that both progeni-tor cell cultures and the lifespan of cultures were greaterin CB than in adult BM, and Payne et al.75 showed thatthe proportion of CD34+ cells that are CD38– (lin–) issignificantly higher in CB than in other stem cell sources.In contrast to what has been documented in adult BM,Traycoff et al.76 demonstrated that LTC-IC and cells pre-sumably capable of in vivo engraftment reside in theCD34+HLA-DR+Rh123dull fraction. The cycling status ofCB progenitors is still a matter of debate. In fact, someinvestigators using the tymidine suicide techniquereported a higher frequency of ceIls in S phase,77 where-as others did not find any difference between CB andadult BM in the frequency of actively cycling progeni-tors.78 More insight into this area is especially necessarysince CB is a very attractive target for the transfer ofgenes able to correct inherited or non-inherited diseasessuch as thalassemia, Fanconi anemia, ADA-deficiency,etc,79 and since entering S phase is required for genetransfer through safety modified retroviruses. Interest-ingly, a higher efficiency of retrovirus-mediated genetrasfer has been reported in CB than in BM progeni-tors.78,80,81 One possible explanation is the particularlyrapid exit from the G0/G1 phases of the cell cycle inresponse to cytokines described by Traycoff et al.82 inCB-derived CD34+ progenitors, which might also justi-fy the ontogeny-related advantage in proliferativepotential.

The functional meaning of these differences in the invitro behavior of phenotypically defined CB and BM pop-ulations is not yet fully understood, but these findingsrepresent an interesting parameter to consider whenassessing the suitability of a CB unit for transplantationin pediatric or adult patients. So far, in fact, most CBtransplant recipients have been pediatric patientsweighing less than 50 kg. Sporadic reports of CB trans-plants in adult recipients have indicated that the timeto myeloid lineage engraftment is comparable to that ofBM recipients, whereas the delay in platelet reconstitu-tion seems to be more pronounced than in pediatric CBtransplant recipients.83 As described in the CB process-ing section, ex vivo expansion of CB progenitors prior totrasplantation might be useful to hasten hematopoieticengraftment; however, since the long-term engraftmentcapability of ex vivo cultured cells might be lost orimpaired,84 more work seems necessary to reach thisimportant goal.

CB collectionEstablished advantages of CB banks over BM donor

registries include the immediate availability of the frozenCB unit, minimal donor attrition, the presence of CBdonors from minority groups that are poorly represent-ed in BM donors registeries, and a much lower incidenceof CMV infected donors. In fact, the time from therequest for a CB unit to finding a matched donor is onaverage less than 2 months, and less than 1% of CB unitsare contaminated by CMV.85

Worldwide, the creation of large CB banks has prompt-

ed investigators to improve the methods for CB collec-tion and fractionation. The first method described byBroxmeyer et al.67 included CB collection in heparinizedtubes by gravity. Further studies69,86 indicated that thisopen system is much more prone to bacterial contami-nation than closed systems based upon CB collection inbags, as first proposed by Gluckman et al.87 Anotherapproach, proposed by Turner et al.,88 includes catheter-ization of the umbilical vein. However, in a recent reportthis procedure was found to cause significant contami-nation of the CB collection by maternal cells, includingpotentially harmful T cells.89 It has been demonstrated,moreover, that CPD/CPD-A have an advantage over ACDand heparin because the former can anticoagulate bloodover a wider volume range. Figure 5 describes resultsobtained using the method of CB collection in closedbags while the placenta is still in utero. Briefly, after thebirth the umbilical cord is doubly clamped 1-2 cm fromthe newborn and transected before the newborn isremoved from the operative field. The free end of thecord must be accurately disinfected before CB collectionby venepuncture of the umbilical vein. As shown in Fig-ure 5, there is a strict correlation between the time ofumbilical cord clamping, the volume and the total num-ber of nucleated cells collected. If the clamping procedureis delayed to the second minute after birth, it seems dif-ficult to collect systematically a number of nucleatedcells sufficient for clinical use of the CB unit. In fact,after the birth the newborn is frequently positionedbelow the level of the uterus, and this determines the so-called transfusion effect from the placenta to the new-born.90,91 Interestingly, when newborns are delivered fol-lowing the Leboyer method there is no transfusion effectsince the newborn is kept above the level of uterus.92

Under these circumstances, early clamping of the cord isnot required for collection of CB for clinical use. At thepresent time there is no consensus among neonatolo-gists and pediatricians about the more appropriate tim-ing of umbilical cord clamping. However, cord clampingin the first 30-60” after birth seems adequate to mostreviewers,93-97 and recent data on the immediate follow-up of newborns who underwent early clamping of thethe umbilical cord and CB collection support the safetyof this procedure.98 In this retrospective study, none ofthe newborns who had CB collected were reported tosuffer from weight loss, fatigue while feeding, tachyp-nea and tachycardia, hypoxia, or cardiac or pulmonarydisease with reduced arterial oxygen saturation. The solesignificant difference between the group of 59 CB donorsand the control group was a slight reduction of Hb val-ues, which corresponded to a loss of about 15 to 20 mgof iron.

CB processingAs mentioned above, a CB unit to be used for trans-

plantation contains remarkably fewer CD34+ progenitorcells than BM or PBSC collections. For this reason, dur-ing CB manipulation the loss of progenitor cells must becarefully avoided. In pilot projects for large scale bank-

F. Bertolini et al.

57


ing, CB was in fact stored as unmanipulated wholeblood. The high cost of this procedure, which requireslarge liquid nitrogen space, has fueled intense researchto concentrate CB nucleated cells in a reduced volume.Among the methods recently proposed for CB process-ing, however, some included density separation byFicoll99 or red cell sedimentation by means of animalgelatin,69,100 i.e. reagents that are currently not (andprobably will never be) licensed for use in humans byregulatory agencies like the the FDA. Consequently, pro-cedures involving the use of licensed products likeHES101,102 should be recommended and a maximum lossof 10-15% of progenitors accepted. In this context, itmust be noted that a number of patients have alreadybeen successfully transplanted with red cell-depletedCB.101,103 Conversely, data on the engraftment potentialof purified, CB-derived CD34+ cells are still poorly repro-ducible in the SCID-hu mouse model104 and totally lack-ing in humans, so for the time being the storage of puri-fied CD34+ CB progenitors for clincal use is not recom-mended. However, it has been shown in vitro that theproliferation potential of purified CD34+ cells is marked-ly superior to that of the low density or Ficoll frac-tion.70,104 This finding has major implications for the pos-sible ex vivo expansion of an aliquot of the CB unit pri-or to transplantation. Two different strategies have beenevaluated: the goal of some authors was to obtain mul-tiple lineage expansion of progenitors by means ofcytokine combinations including SCF, FLT-ligand IL-1,IL-3, IL-6, IL-11, G- and GM-CSF,104-106 while others wereinterested in selected-lineage expansion of themyeloid107 or megakaryocytic lineage.108

Rubinstein et al.101 have recently proposed a new pro-cedure for washing the CB unit prior to transplantation.Advantages of this approach include removal of free Hbfrom lysed red cells and a significant reduction in theDMSO infused, a molecule which is particularly toxicfor pediatric transplant recipients.109 In vitro data indi-

cate that the washing procedure may improve theengraftment potential of the transplanted cells, but thisfinding should be futher confirmed in an in vivo model.

Immunological features of cord bloodlymphocytes

The immune system which develops during fetal lifeis not fully competent at birth and continues the dif-ferentiatation process after birth in response to variousantigenic challenges. At least three important elementscontrol the development of the immune system duringfetal life, thus determining the peculiar characteristicsof the cord blood lymphocyte (CBL)-mediated immuneresponse: i) limited or even absent antigenic experience,ii) immaturity of the majority of lymphocyte popula-tions, and iii) feto-maternal immunological interac-tion.110,111 These elements are believed to influence theimmunological features most strictly related to cordblood transplantation (CBT) and, in particular, thecapacity to develop alloantigen-directed reactivity, anti-microbial immunity and anti-tumor immune surveil-lance.

As a consequence of poor antigenic experience duringpregnancy, the majority of CBL are naive cells express-ing the RA isoform of the CD45 molecule.112 The mostprominent immunoregulatory function of CD45RA+ Tlymphocytes is suppressor activity.113 These peculiar fea-tures of neonatal lymphocytes explain their incapacity todevelop, both in vivo114 and in vitro,115 an immuneresponse directed towards recall antigens (i.e. tetanustoxoid, influenza virus).

The immaturity of the CBL population is also a directconsequence of its poor antigenic experience. Comparedto adult blood, the distribution of the most relevant CBLsubpopulations is characterized by a reduced percent-age of CD3+ mature T lymphocytes and by the presenceof immature T and NK lymphocyte subsets which arenot detectable in adult peripheral blood.116,117

B lymphocytes are present in a normal or even aug-mented percentage in cord blood as compared with adultblood, even though immunoglobulin production is lim-ited to the IgM class.118 Other CBL peculiar features relat-ed to their immaturity are low expression of adhe-sion/costimulation molecules such as CD11a (LFA-1),CD18, CD54 (ICAM-1), CD58 (LFA-3) antigens,113,117,119

poor expression of CD40 ligand on activated T lympho-cytes,120 and reduced ability to secrete some cytokines(i.e. g-interferon, tumor necrosis factor and interleukin-4).117

Spontaneous NK activity is reduced in cord blood ascompared to adult blood, possibly because of the lowexpression of adhesion molecules, known to be useful inpromoting the capacity of NK lymphocytes to adhere totarget cells.121 On the other hand, antibody-dependentcell cytotoxicity (ADCC) and lymphokine activated killer(LAK) activity of cord blood reach values comparable toor even higher than those observed in adult blood.121

Moreover, recent data demonstrate that the innate

Figure 5. Effect of the time of umbilical cord clamping on themean (±1 SD) volume and nucleated cell count of cord bloodcollections (n=67). As in most European OB/GYN units,soon after birth the newborn is kept below the level of theuterus, and delayed clamping of the umbilical cord is asso-ciated with a reduction of cord blood volume and cellularcontent.


58


F. Bertolini et al.

immunity directed towards Epstein-Barr virus-infectedcells is remarkably high in CBL collected from the major-ity of neonates.122 The capacity of cord blood NK cells tobe promptly activated in vitro suggests that innateimmunity plays a key role in immune surveillance dur-ing fetal and perinatal life, as long as specific T-cellmediated adaptive immunity can be generated.

From an immunological viewpoint, pregnancy can beconsidered as a successful HLA-mismatched transplan-tation. It is well known that the feto-maternal, anato-mo-functional barrier allows the reciprocal transfer oflymphocyte subpopulations. It is thus conceivable tohypothesize that a very effective immunological net-work acts to prevent fetal rejection and graft-versus-host reactions (GVHR).111,123 Several lines of clinical andexperimental evidence support this hypothesis, in par-ticular: i) CBL preferentially display suppressor ratherthan helper immunological functions111,113 and ii) bothB and T lymphocytes maintain a state of hyporespon-sivity towards noninherited maternal HLA molecules fora long time after birth.124,125 A further confirmation ofthis peculiar state of tolerance derives from a recentlyreported observation126 on the occurrence of acuteGVHD in patients given CBT from donors who were dis-parate for the noninherited paternal allele, and on theabsence of significant acute GVHD in recipients whosedonors were disparate for the noninherited maternalallele.

The fetal/neonatal period has been postulated to rep-resent a crucial time in ontogeny, during which T and Blymphocytes learn to discriminate between self and non-self through the development of a state of tolerancetoward antigens they encounter.127 The concept of neo-natal tolerance was recently re-examined in mice128-130

and it was demonstrated that induction of this phe-nomenon may depend on several elements, includingthe nature of the antigen-presenting cells,128 the doseof antigen administered,129 and the mode of immuniza-tion.130

Poor antigenic experience, immaturity of lymphocytesubpopulations, feto-maternal immunological interac-tions, and neonatal tolerance may, altogether, con-tribute to the generation of a suppressive effect on CBLalloreactivity, thus permitting the use of HLA-partiallymatched donors for CBT. In agreement with this hypoth-esis, reduced proliferative and cytotoxic activity towardsalloantigens was reported by several authors to be pre-sent in cord blood as compared with adult peripheralblood.117,131-133 However, normal CBL alloreactivity hasbeen documented in other studies.134,135 The discrepan-cies observed between the above mentioned reports maydepend on the high interindividual variability in the dis-tribution of cord blood T and NK lymphocyte subpopu-lations. Interestingly, it has been recently reported that,even though proliferative response to alloantigen in aprimary mixed lymphocyte culture (MLC) is comparablein adult and cord blood, restimulation in secondary MLCinduces increased specificity and activity of adult allore-active lymphocytes and a state of unresponsiveness inCBL.136 These data strongly suggest that repeated in vit-ro stimulation with allogeneic cells amplifies the specificimmune response of adult lymphocytes, while the sameprocedure induces a state of anergy in neonatal cells.

Several clinical and experimental data obtained in thesetting of allogeneic bone marrow transplantation(BMT) demonstrate that there is a strong correlationbetween GVHD and graft-versus-leukemia (GVL) effect.Thus, due to their low alloreactive capacity (responsiblefor the reduced GVHR),126,135 CBL could be less efficientin mediating a GVL effect. As far as we know, no dataconcerning the identification of cord blood T lympho-cytes capable of mediating specific anti-leukemic activ-ity have been reported in the literature. This lack ofinformation is not surprising since it is well known thatthe frequency of these cells is extremely low, even in theperipheral blood of healthy adult donors. However, somestudies have recently demonstrated that innate anti-leukemic activity mediated by LAK cells and measured

Method of T-Cell Depletion Cells Removed T-Cell Depletion(x log10)

SBA lectin and E-rosette depletion T and B lymphocytes, monocytes,neutrophils 2.5 - 3.0

Multiple E-rosette depletions T lymphocytes 2.0Mouse MoAb (anti-CD2, CD8) + rabbit C’ T lymphocytes 2.0Mouse MoAb (anti CD6) + human C’ T lymphocytes 1.5-2.0Rat MoAb(CAMPATH-1) + human C’ T and B lymphocytes, monocytes, 2.5

Anti-CD5 immunotoxin-Ricin A Immunomagnetic separation (anti-CD3, CD8)

T lymphocytes 2.0

SBA lectin + immunomagnetic separation T and B lymphocytes, monocytes,neutrophils

3.1

AIS CD%/8T-CELLector T lymphocytes 2.5Autologous Immunorosettes(anti-CD2 and CD3 tetrametric complexes)

T lymphocytes 2-3

Table 3. Methods of T-Cell depletion in clinical trials.

59


against long-term tumor cell lines, is comparable inadult and cord blood.119,137 Even though the above men-tioned studies are interesting, more experimental dataand clinical observations are required to better definethe potential GVL effect of CBL.

PBSC processing

T-cell depletionThe role of T lymphocytes in bone marrow transplan-

tation is very complex. They are in fact responsible forGvHD, a major contributing factor correlated with mor-bidity and mortality in allogeneic bone marrow trans-plantation.138,139 Indeed when T-cells are removed fromthe graft before transplantation, the incidence of GvHDdecreases sharply.71,140-143 On the other hand, the possi-ble beneficial roles of T-lymphocytes include sustainingengraftment144 and preventing relapses through thegraft-versus-leukemia effect.145,146 They are also crucialin immune-hematological reconstitution after trans-plantation because slow or deficient reconstitution maylead to a high incidence of viral infections or otherinfectious complications. Ex vivo manipulation of the T-lymphocyte content is easier and T-cell depleted allo-geneic transplants may in the future be followed byinfusion of non-alloreactive T-lymphocytes or of specif-ically engineered lymphocyte clones exerting an anti-viral or anti-neoplastic effect.

Standard T-cell depletion techniquesOver the past 15 years, several techniques have been

developed for depleting T-cells from human marrowallografts.147 Table 3 summarizes the principles on whichthey are based and the degree of T-cell depletion eachprovides. Unfortunately, these methods may be time-consuming, cumbersome and difficult to standardize indifferent transplantation centers. Results are thereforeoften variable and no general consensus has emerged onthe use and benefit of bone marrow T-cell depletion.

Very few data are available on the use of standard T-cell depletion methods in heterogeneous nucleated cellpopulations collected by leukaphereses from the periph-eral blood of donors previously stimulated by hemopoi-etic growth factors.

Kessinger et al.148 first reported allogeneic transplan-tation utilizing T-cell depleted peripheral bloodmononuclear cells and a sheep erythrocyte rosettingtechnique. Engraftment was rapid and grade II acuteGvHD was observed.

In a preliminary study after monocyte lysis with L-phenylalanine methyl ester, Suzue et al.149 depleted T-lymphocytes from apheresis products harvested afterstimulation of healthy donors with G-CSF using both E-rosettes with sheep red blood cells and panning withflasks coated with anti CD5/CD8 monoclonal antibod-ies. They reported an unsatisfactory depletion of T-cells(99.5%) and a stem cell recovery of only 7.5%.

Aversa et al.71 employed soybean lectin agglutinationfollowed by 2 to 4 rounds of E-rosetting with sheep redblood cells on the leukapheresis product from donorsstimulated with G-CSF. This approach achieves approx-imately 3×log10 T-lymphocyte depletion, as measuredby cytofluorimetric assays. The main drawbacks are itscomplexity and lengthy laboratory times.

Stem cell positive selectionIn principle, reducing T lymphocytes in the leuka-

pheresis product by positively selecting CD34+ hemo-poietic progenitors appears to be a valid technical alter-native. Table 4 shows the basis of the main techniquesfor positive selection of hemopoietic progenitor cells.All use one monoclonal antibody which identifies anepitope on the human CD34 antigen. Separation iseffected by collecting the antibody-sensitized cells ontoa solid phase such as magnetic beads, plastic plates orcolumns of non magnetic particles, while non-targetcells remain in suspension. Systems that utilize highspeed flow cytometry to sort CD34+ cell populationshave also been developed.150

The CD34+ stem cell selection systems adopted inclinical use are based on immunoadsorption and indirectimmunomagnetic beads.

Most clinical trials to date have been carried out witha ®Ceprate Stem Cell Concentrator (CellPro Inc., Bothell,WA, USA), which employs biotinylated 12.8 monoclon-al antibody. The sensitized cells are applied to a columnof avidin-coated polyacrylamide beads. Cells expressing

Company Method Antibody Detachment

CellPro Immunoadsorption 12.8 MechanicalBaxter Magnetic beads indirect 9C5 ChymopapainBaxter Magnetic beads indirect 9C5 PR34+™ oligopeptideDynal Magnetic beads direct BI3C5 Anti-Antibody

(anti Fab of mouse MoAb)AIS Panning ICH3 MechanicalImmunotech Magnetic latex beads direct QBEND10 Not requiredMilteny Magnetic colloid indirect QBEND10 Not requiredTerry Fox Laboratory Magnetic Colloid Indirect 8G12 Not requiredSystem x FACS (high speed) Various Not required

Table 4. Methodsavailable for stemcell positive selec-tion.


60


F. Bertolini et al.

the CD34 antigen are retained and unlabelled cellswashed through the column with gentle mechanicalagitation. The CD34+ cells are then removed from thebeads and collected.

Using this system on the leukapheresis product, Linket al.151 recovered a mean of 30% CD34+ cells, with apurity of 70%. Peripheral blood CD3+ cells were reducedby 3 logs. Other investigators have reported similarresults.152-155 This degree of T-cell depletion is known toprevent severe GvHD in severe combined immune defi-ciency (SCID) patients after matched or mismatchedbone marrow transplantation.140 In leukemia patientsundergoing matched transplants it may not be enoughto eliminate GvHD completely without the concomitantadministration of immunosuppressive drugs. The thresh-old number of clonable T-lymphocytes in the inoculumshould be below 1×105/kg b.w.,156 which is difficult toachieve with one-step positive selection of hemopoiet-ic stem cells. For a successful mismatched bone marrowtransplant the T-lymphocyte threshold must be < 3-5×104/kg b.w.157 in the inoculum because of the greaterlikelihood and increased severity of GvHD in thesepatients. On the other hand, infusing a number wellbelow the threshold value could expose the patient toa high risk of graft failure.

Because T-cell depletion with the ®Ceprate systemapplied directly on the leukapheresis product does notreduce T-lymphocyte content from the graft by morethan 3 logs, an additional T-cell depletion step isrequired.

In 10 patients with different types of leukemias, Aver-sa et al.157 employed an E-rosetting procedure beforepositively selecting hemopoietic progenitors with the®Ceprate system. This combined method yielded a T-celldepletion of 4.3 logs in the graft and a mean CD34+

recovery of 50-60%.158

In small-scale experiments, Fernandez et al.159 appliedthe E-rosetting procedure after positive selection ofCD34+ cells with this same ®Ceprate system to obtain amean log10 T-cell depletion of 4. Slaper-Cortenbach etal.160 achieved a median recovery of 42.7% CD34+ cellsand a T-lymphocyte reduction of 2-3 logs in 13 hap-loidentical transplants for SCID and in leukemia patientsby employing autologous immunorosettes after positiveselection of CD34+ cells.

CD34+ progenitor cell immunomagnetic selection(Baxter, Irvine, CA, USA) achieved a 3 log T-cell deple-tion in preclinical experiments.40,161 However, the mainproblem with this methodology was bead release fromthe target CD34+ cells. In fact release mediated by chy-mopapain may cause intractable cell clumping, partic-ularly when a large number of cells are processed.Recently the PR34+TM stem cell releasing agent, anoligopeptide competing with the anti-CD34 monoclon-al antibody for the release of the CD34+ cells from themagnetic beads, has also been proposed.161 Preliminaryresults showed a reduction of non-target T cells by afactor of 2-3 logs with yields of CD34+ cells rangingfrom 31.1 to 85%.162

In conclusion, positive selection of CD34+ cells withthe ®Ceprate system reduces the graft T-lymphocytecontent under the threshold of risk for GvHD only whencombined with standard T-depletion techniques such asE-rosetting with sheep red blood cells or autologousimmunorosetting. Indirect immunomagnetic systemshave to be evaluated more precisely for the use as a T-cell depletion system.

Immunogenic activity of CD34+

hematopoietic cellsAutologous transplantation of selected CD34+ cells

induces rapid and complete hematologic reconstitutionin myeloablated patients. In addition, isolation of CD34+

cells can be considered as an ex vivo means of purgingneoplastic cells from the marrow or peripheral blood ofpatients with solid tumors or hematologic malignan-cies.163-165

In the allogeneic setting, selection of CD34+ cells maybe aimed at depleting donor T-cells and professionalantigen-presenting cells (APC) such as monocytes, acti-vated B-cells and dendritic cells, which are very potentstimulators of T-cell responses. Dendritic cells constitu-tively express the B7-2 (CD86) costimulatory moleculeand upregulate B7-1 (CD80), B7-2 (CD86) and othermolecules upon activation.166-171 Furthermore, they wererecently shown to derive from CD34+ marrow or periph-eral blood cells, and can be rapidly generated in vitro inthe presence of a specific combination of growth fac-tors.172-177 Since it has been demonstrated that T-cellreceptor (TCR): antigen interaction, in the absence of

Authors TNC(x 108/kg)

CD34+

(x 106/kg)CD3+

(x 106/kg)

NK(x 106/kg)

Dreger et al. (26) 13.52 8.16 404 N.R.

Weaver et al. (191)* 20.53 9.6 450 N.R.

Körbling et al. (16) 16.5 10.7 300 64.3

Schmitz et al. (17) 8.6 13.1 340 94.0

Bensinger et al. (18) 10.6 13.1 385 N.R.

Majolino et al. (27) 9 6.84 250 27

Rambaldi et al. (203) 8 6.9 279 N.R.

Table 5. Median value of nucleatedcells, CD34+ cells, CD3+ cells andnatural killer cells infused inpatients undergoing allogeneicperipheral blood stem cell trans-plantation.Legend. *Syngeneic transplants.N.R.=Not reported.

61



appropriate costimulation, may induce T-cell unrespon-siveness or even apoptotic deletion,171,178-181 the alloanti-gen presenting function of CD34+ marrow cells wasrecently investigated to evaluate whether transplanta-tion of purified CD34+ cells could minimize the immunesensitization of an allogeneic receipient.182 CD34+ mar-row cells have been purified to >98% by a two-stepprocedure consisting of a first enrichment on animmunoaffinity chromatography column, followed byfluorescence activated cell sorting. Cytofluorimetricanalysis of purified CD34+ marrow cells revealed theexpression of HLA-DR and CD86 on >95% and 6% ofthe cells, respectively. Primary mixed leukocyte culturesdemonstrated that irradiated CD34+ marrow cellsinduce brisk proliferation of allogeneic T-cells isolatedfrom HLA-DR incompatible donors. On the basis of pre-vious reports,183,184 expression of CD18, the commonchain of a family of leukointegrins, was investigated onCD34+ marrow cells and CD34+/CD18– cells were sortedto investigate whether this cell population was enrichedin early hemopoietic precursors incapable of immunos-timulating activity.

On average, 25% of CD34+ marrow cells were CD18–

by direct immunofluorescent analysis. Purified CD34+,CD34+/CD18+ and CD34+/CD18– marrow subsets weretested in bulk MLC with allogeneic T-cells, and it wasobserved that CD34+, CD34+/CD18+ and unseparatedmarrow mononuclear cells have a similar capacity tostimulate a T-cell response. Conversely, CD34+/CD18–

cells do not elicit any T-lymphocyte proliferation. More-over, limiting dilution assay (LDA) experiments showed,on a per cell basis, that CD34+/CD18– and CD34+/CD86–

marrow cells have a very poor ability to induce a T-cellresponse, as opposed to CD34+/CD18+ and CD34+/CD86+

marrow cells. Since most marrow LTC-IC were includ-ed in the CD34+/CD18– cell fraction, it was concludedthat CD34+/CD18–, or CD34+/CD86– marrow cells, mayrepresent a useful source of progenitor cells for allo-geneic transplantation because of their high stem cellactivity combined with reduced immunogenicity. Dataon normal human G-CSF mobilized CD34+ peripheralblood (PB) cells show that on average 30% of the cellsexpress CD18 and only 3% express CD86, while func-tional in vitro results are consistent with what was pre-viously observed in marrow. Thus, CD34+ PB cells canpotently stimulate T cells, likely through the B7:CD28pathway, and CD34+/CD18– PB cells still have very weakimmunostimulating activity.

In a preliminary study sibling baboons were fullyengrafted with allogeneic CD34+ marrow cells withoutGVHD, after receiving total body irradiation as condi-tioning regimen and standard GVHD prophylaxis.184

Development of mobilization regimens capable ofincreasing the number of peripheral blood hemopoieticstem cells in normal healthy donors allowed sufficientamounts of CD34+ PB cells to be harvested for allogeneictransplantation in humans. In fact, transplantation ofenriched populations of G-CSF mobilized CD34+ cellsresulted in rapid engraftment, similar to that observed

in allogeneic PBSC transplants.185-190 Purification ofCD34+ cells on the Ceprate column obtains on averagea 3 log depletion of CD3+ T cells in the graft; however,several studies reported contrasting rates of acuteGVHD. In particular, > 80% of the patients transplant-ed with CD34+ PB cells in Seattle experienced aGVHDgrade II-III after receiving a median number of 0.7×106

T-cells/kg in the graft and GVHD prophylaxis withcyclosporin-A (CsA)± methotrexate (MTX).187 Anotherstudy reported 2 cases out of 5 who died from aGVHD.188

By contrast, other groups reported a very low incidenceof GVHD.189,190 One of the reasons for these disparitiesmay be that small numbers of patients, often with dif-ferent malignancies and clinical characteristics, areincluded in these studies. Nevertheless, two hypothesescould be addressed: the first one suggests that infusionof as little as 0.5-1×106 CD3+ T cells/kg could be poten-tially capable of initiating GVHD, which would be pre-vented by further steps in T-cell depletion.158 The secondhypothesis, still to be tested, is whether APC in marrowor peripheral blood could play a role in the developmentof GVHD by presenting allogeneic peptides to donor T-cells.

In this regard, a subset of CD34+ cells in the graft mayinduce the activation and proliferation of a limited num-ber of T cells, such as those still present after CD34purification.

Peripheral blood stem cells: immunological aspects

Very few data are available on the effects of hemo-poietic growth factors used to mobilize PBSC on periph-eral blood lymphocytes.

Weaver et al.191 analyzed the influence of G-CSF onperipheral blood lymphocytes from 13 individuals (11autografts and 2 normal donors). In all cases theyobserved a slight increase in CD3, CD4, CD8, CD19 andCD20-positive lymphocytes, with a return to pretreat-ment values by days 4 and 5 of G-CSF administration.The change in the CD4/CD8 ratio was not statisticallysignificant.

The expression of CD2, CD3, CD4, CD7, CD8, CD20,CD25, CD57 and HLA-DR antigens was evaluated dur-ing administration of G-CSF (12 ug/kg/day for 5-7 days)to healthy donors. No significant variations wereobserved in the different lymphocyte subsets, in theCD4/CD8 ratio or in the expression of CD25 and HLA-DRantigens (unpublished data). G-CSF administration doesnot cause direct activation of T lymphocytes in vivo. Thismight be expected because lymphocytes do not possessthe G-CSF receptor.192 However, it is possible that acti-vation could be caused by cytokine release from cellsstimulated by G-CSF.

Other important aspects of the PBSC allograft includethe lymphocyte content, particularly T lymphocytes andnatural killer cells, in the apheresis product. Table 5reports data on the total number of CD3+ lymphocytesderived from peripheral blood stem cells that wereinfused for allogeneic transplants. The number of

62


infused T lymphocytes was always 1.5-2 logs greaterthan that derived from bone marrow.193

The exact relationship between the T-lymphocytecontent in the graft and the development and severityof GvHD remains unclear. A linear relationship betweenthe number of T lymphocytes infused and the develop-ment of GvHD has long been hypothesized,139,194,195 butthis correlation has not always been confirmed.196,197

Findings in allogeneic peripheral blood stem cell trans-plantation seem to suggest that the number of T lym-phocytes is less important than donor-cell specificity intriggering GvHD.16,17

Another aspect of peripheral blood stem cell trans-plantation concerns the number of natural killer cells(NK) infused (Table 5) since they are important effectorcells in graft-versus-leukemia activity.198 The number ofinfused NK cells is about 20 times greater in an allo-geneic peripheral blood stem cell transplant than in abone marrow graft.16,17 The question of whether this willtranslate into more potent GvL activity in patients allo-grafted with peripheral blood stem cells compared tounmanipulated bone marrow cannot be answered atthis time, but needs further study. However, preliminarydata from a murine model demonstrated strong GvLactivity for allogeneic NK cells without the induction ofGvHD.199

An important technical point is the effect of freezingand thawing of the graft or of keeping the apheresisproduct at 4°C overnight on T-lymphocytes inactiva-tion. Van Bekkum200 described selective elimination ofimmunologically competent cells from bone marrowafter storage at 4°C. Eckardt et al.201 also noted that cry-opreservation of allogeneic marrow may reduce the riskof acute GvHD. Selective depletion or induction of aner-gy in GvHD-inducing cells was hypothesized.201 Storageof the apheresis product at 4°C overnight does not mod-ify the surface expression of the CD3, CD4, CD8, or theCD57 antigens in a significant manner (Tabilio, 1996,unpublished results); however, how cryopreservationaffects the alloreactivity of peripheral blood stem cellsneeds to be investigated further.

ConclusionsSeveral studies have now shown that hematopoietic

stem cells collected from PB after the administration ofG-CSF, or from CB upon delivery, are capable of sup-porting rapid and complete reconstitution of BM func-tion in allogeneic recipients.16-20,204,205 The faster recov-ery of hematopoiesis as compared to BM-derived allo-grafts, together with a lower incidence of aGVHD thanexpected with BM transplantation, raises the questionof whether PBSC collections may differ from conven-tional BM harvests with respect to the number of stemcells and their functional characteristics, lymphoid cellcomposition and T-cell reactivity. Moreover, recent clin-ical studies on transplantation of CB-derived cells fromunrelated HLA-mismatch donors support the notion thatsources of hematopoietic stem cells other than BM rep-resent a feasible alternative to conventional transplan-

tation. In this paper, the phenotypic, functional andkinetic features of circulating and CB hematopoieticcells were reviewed. We also emphasized the technicalaspects of CB collection and processing, as well as theprotocols for PBSC mobilization and collection from nor-mal donors. Notably, novel data on the immunogenicand kinetic profile of BM and PB CD34+ cells may shednew light on stem cell biology and may help clinicalinvestigators to design future trials on transplantationof purified hematopoietic progenitors.

It should be remembered that despite growing inter-est these procedures must still be considered asadvanced clinical research and should be included informal clinical trials aimed at demonstrating their defin-itive role in stem cell transplantation. In this regard, alarge European randomized study is currently compar-ing PBSC and BM allografts. However, the possibility ofcollecting a large quantity of hematopoietic progenitorstem cells from PB, perhaps with reduced allo-reactivi-ty, offers an exciting perspective for widening the num-ber of potential stem cell donors and greater leeway forgraft manipulation than is possible with BM.

AcknowledgmentsPreparation of this manuscript was supported by

grants from Dompé Biotec SpA and Amgen Italia SpA,Milan, Italy.

This review article was prepared by a group of expertsdesignated by Haematologica and by representatives oftwo pharmaceutical companies, Amgen Italia SpA andDompé Biotec SpA, both from Milan, Italy. This co-oper-ation between a medical journal and pharmaceuticalcompanies is based on the common aim of achieving anoptimal use of new therapeutic procedures in medicalpractice. In agreement with the Journal’s Conflict ofInterest Policy, the reader is given the following infor-mation. The preparation of this manuscript was sup-ported by educational grants from the two companies.Dompé Biotec SpA sells G-CSF and rHuEpo in Italy, andAmgen Italia SpA has a stake in Dompé Biotec SpA. Thispaper has undergone a regular peer-review process andhas been evaluated by two outside referees.

References

1. Thomas D. Bone marrow transplantation: past, pre-sent and future. Haematologica 1991; 76:353-6.

2. Majolino I, Aversa F, Bacigalupo A, Bandini G, ArceseW, Reali G. Allogeneic transplants of rhG-CSF-mobi-lized peripheral blood stem cells (PBSC) from normaldonors. Haematologica 1995; 80:40-3.

3. Carlo-Stella C, Cazzola M, De Fabritiis P, et al. CD34-positive cells: biology and clinical applications.Haematologica 1995; 80:367-87.

4. Aglietta M, De Vincentiis A, Lanata L, et al. Peripher-al blood stem cells in acute myeloid leukemia: biolo-gy and clinical applications. Haematologica 1996;81:77-92.

5. Caligaris-Cappio F, Cavo M, De Vincentiis A, et al.Peripheral blood stem cell transplantation for thetreatment of multiple myeloma: biological and clini-

F. Bertolini et al.

63



cal implications. Haematologica 1996; 81:356-75.6. Buckner CD, Clift RA, Sanders JE, et al. Marrow har-

vesting from normal donors. Blood 1984; 64:630-4.7. Stroncek DF, Holland PV, Bartch G, et al. Experiences

of the first 493 unrelated marrow donors in thenational marrow donor program. Blood 1993;81:1940-6.

8. Gratwohl A, Schmitz N. Introduction to First Interna-tional Symposium on allogeneic peripheral blood pre-cursor cell transplants. Bone Marrow Transplant1996; 17(suppl 2):1-3.

9. To LB, Roberts MM, Haylock DN, et al. Comparisonof haematological recovery times and supportive carerequirements of autologous recovery phase peripher-al blood stem cell transplants, autologous bone mar-row transplants and allogeneic transplants. BoneMarrow Transplant 1992; 9:277-84.

10. Indovina A, Majolino I, Buscemi F, et al. Engraftmentkinetics and long term stability of hematopoiesis fol-lowing autografting of peripheral blood progenitorcells. Haematologica 1995; 80:115-22.

11. To LB, Sheppard KM, Haylock DN, et al. Single highdoses of cyclophosphamide enable the collection ofhigh numbers of haematopoietic stem cells from theperipheral blood. Exp Hematol 1990; 18:442-7.

12. Indovina A, Majolino I, Scimè R, et al. High dosecyclophosphamide: stem cell mobilizing capacity in21 patients. Leuk Lymphoma 1994; 14:71-7.

13. Sheridan WP, Begley CG, Juttner CA, et al. Effect ofperipheral blood progenitor cells mobilized by fil-grastim (rhG-CSF) on platelet recovery after high-dosechemotherapy. Lancet 1992; 1:640-4.

14. Gianni AM, Bregni M, Siena S, et al. Granulocyte-macrophage colony-stimulating factor to harvest cir-culating hemopoietic stem cells for autotransplanta-tion. Lancet 1989; 2:580-5.

15. Auquier P, Macquart-Moulin G, Moatti JP, et al. Com-parison of anxiety, pain and discomfort in two proce-dures of hematopoietic stem cell collection: leukacy-tapheresis and bone marrow harvest. Bone MarrowTransplant 1995; 16:541-7.

16. Körbling M, Przepiorka D, Huh YO, et al. Allogeneicblood stem cell transplantation for refractoryleukemia and lymphoma: potential advantage ofblood over marrow allografts. Blood 1995; 85:1659-65.

17. Schmitz N, Dreger P, Suttorp M, et al. Primary trans-plantation of allogeneic peripheral blood progenitorcells mobilized by filgrastim (granulocyte colony-stim-ulating factor). Blood 1995; 85:1666-72.

18. Bensinger WI, Weaver CH, Appelbaum FR, et al.Transplantation of allogeneic peripheral blood stemcells mobilized by recombinant human granulocytecolony stimulating factor. Blood 1995; 85:1655-8.

19. Bacigalupo A, Majolino I, Van Lint MT, et al. Trans-plantation of rhG-CSF mobilized allogeneic peripher-al blood cells from HLA identical sibling donors[abstract]. Bone Marrow Transplant 1995; 15(Suppl.2):5.

20. Majolino I, Saglio G, Scimé R, et al. High incidence ofchronic GVHD after primary allogeneic peripheralblood stem cell transplantation in patients with hema-tologic malignancies. Bone Marrow Transplant 1996;17:555-60.

21. Krause DS, Fackler MJ, Civin CI, May WS. CD34:structure, biology, and clinical utility. Blood 1996;87:1-13.

22. Papayannopoulou T, Nakamoto B. Peripheralizationof hemopoietic progenitors in primates treated withanti-VLA4 integrin. Proc Natl Acad Sci USA 1993;90:9374-8.

23. Richman CM, Weiner RS, Yankee RA. Increase in cir-

culating stem cells following chemotherapy in man.Blood 1976; 47:1031-4.

24. Bungart B, Loeffler M, Goris H, Diehl V, Nijhof W.Differential effects of recombinant human colonystimulating factor (rh G-CSF) on stem cells in mar-row, spleen and peripheral blood in mice. Br J Haema-tol 1990; 76:174-9.

25. Sheridan WP, Begley CG, To LB, et al. Comparison ofautologous filgrastim (rhG-CSF)-mobilized peripher-al blood progenitor cells to restore hemopoiesis afterhigh-dose chemotherapy for lymphoid malignancies.Bone Marrow Transplant 1994; 14:105-11.

26. Dreger P, Haferlach T, Eckstein V, et al. rhG-CSF-mobilized peripheral blood progenitor cells for allo-geneic transplantation: safety, kinetics of mobiliza-tion, and composition of the graft. Br J Haematol1994; 87:609-13.

27. Majolino I, Buscemi F, Scimé R, et al. Treatment ofnormal donors with rhG-CSF 16 µg/kg for mobiliza-tion of peripheral blood stem cells and their aphereticcollection for allogeneic transplantation. Haemato-logica 1995; 80:219-26.

28. Bensinger WI, Buckner CD, Rowley S, Storb R, Appel-baum FR. Treatment of normal donors with recom-binant growth factors for transplantation of allo-geneic blood stem cells. Bone Marrow Transplant1996; 1(Suppl. 2):19-21.

29. Anderlini P, Miller P, Sundberg J, et al. “High-dose” G-CSF (filgrastim) for stem cell mobilization in normaldonors: a prospective study. ASCO Proc 1996;15:269.

30. Bensinger WI, Prince TH, Dale DC, et al. The effectsof daily recombinant human granulocyte colony stim-ulating factor administration on normal granulocytedonors undergoing leukapheresis. Blood 1993; 81:1883-8.


32. Majolino I, Scimé R, Vasta S, et al. Mobilization andcollection of PBSC in healthy donors. Comparisonbetween two schemes or rhG-CSF administration. EurJ Haematol 1996; 57:214-21.

33. Dührsen U, Villeval JL, Boyd J, Kannourakis G,Morstyn G, Metcalf D. Effects of recombinant humangranulocyte colony-stimulating factor on hematopoi-etic progenitors cells in cancer patients. Blood 1988;72:2074-81.

34. Russell JA, Luider J, Weaver M, et al. Collection ofprogenitor cells for allogeneic transplantation fromperipheral blood of normal donors. Bone MarrowTransplant 1995; 15:111-5.

35. Waller CF, Bertz H, Engelhardt M, et al. Mobilizationof peripheral blood progenitor cells (PBPC) for allo-geneic peripheral blood progenitor cell transplanta-tion (allo-PBPCT): efficacy and toxicity of a high doserhG-CSF regimen. Bone Marrow Transplant 1996; 17(Suppl. 2):71.

36. Matsunaga T, Sakamaki S, Kohgo Y, Ohi S, HirayamaY, Niitsu Y. Recombinant human granulocyte colony-stimulating factor can mobilize sufficient amounts ofperipheral blood stem cells in healthy volunteers forallogeneic transplantation. Bone Marrow Transplant1993; 11:103-8.

37. Schwinger W, Mache C, Urban C, Beaufort F, Togl-hofer W. Single dose of filgrastim (rhG-CSF) increas-es the number of hematopoietic progenitors in theperipheral blood of adult volunteers. Bone MarrowTransplant 1993; 11:489-92.

38. Bishop MR, Tarantolo SR, Schmit-Pokorny K, et al.

64


F. Bertolini et al.

Mobilization of blood stem cells from HLA-identicalrelated donors with low-dose granulocyte colony-stimulating factor for allogeneic transplantation.Blood 1995; 86(Suppl. 1):463.

39. Majolino I, Cavallaro AM, Bacigalupo A, et al. Mobi-lization and collection of PBSC in healthy donors: aretrospective analysis of the Italian Bone MarrowTransplantation Group (GITMO). Haematologica1997; 81:47-52.

40. Lane TA, Law P, Maruyama M, et al. Harvesting andenrichment of hematopoietic progenitor cells mobi-lized into the peripheral blood of normal donors bygranulocyte-macrophage colony-stimulating factor(GM-CSF) or G-CSF: potential role in allogeneic mar-row transplantation. Blood 1995; 85:275-82.


42. Goldberg SL, Mangan KF, Klumpp TR, et al. Compli-cations of peripheral blood stem cell harvesting:review of 554 PBSC leukaphereses. J Hematother1995; 4:85-90.

43. Fossa SD, Poulsen JP, Anders A. Alkaline phosphataseand lactate dehydrogenase changes during leukocy-tosis induced by G-CSF in testicular cancer. Lancet1992; 340:1544.

44. Sakamaki S, Matsunaga T, Hirayama Y, Kuga T, Niit-su Y. Haematological study of healthy volunteers 5years after G-CSF. Lancet 1995; 346:1432-3.

45. Hasenclever D, Sextro M. Safety of alloPBPCT donors:biometrical considerations on monitoring long termrisks. Bone Marrow Transplant 1996; 17(suppl 2):28-30.

46. Ohara A, Kojima S, Tsuchida M, et al. Evolution ofacquired severe aplastic anemia to MDS/leukemia inchildhood: a retrospective strudy on 170 severe aplas-tic anemia child patients. Blood 1995; 86(suppl 1):1333.

47. Kalra R, Dale D, Freedman M, et al. Monosomy 7 andactivating RAS mutations accompany malignanttransformation in patients with congenital neutrope-nia. Blood 1995; 86:4579-86.

48. Smith OP, Reeves BR, Kempski HM, Evans JP. Kost-mann’s disease, recombinant HuG-CSF, monosomy 7and MDS/AML. Br J Haematol 1995; 91:150-3.

49. Ringdén O, Potter MN, Oakhill A, et al. Transplanta-tion of peripheral blood progenitor cells from unre-lated donors. Bone Marrow Transplant 1996; 17(Sup-pl 2):62-4.

50. Tjonnfjord GE, Steen R, Evensen SA, et al. Character-ization of CD 34+ peripheral blood cells from healthyadults mibilized by recombinant human granulocytecolony-stimulating factor. Blood 1994; 84:2795-801.

51. Roberts AW, Metcalf D. Noncycling state of periph-eral blood progenitor cells mobilized by granulocytecolony-stimulating factor and other cytokines. Blood1995; 86:1600-5.

52. To LB, Haylock DN, Dowse T, et al. A comparativestudy of the phenotype and proliferative capacity ofperipheral blood (PB) CD34+ cells mobilized by fourdifferent protocols and those of steady-state PB andbone marrow CD34+ cells. Blood 1994; 84:2930-9.

53. Bregni M, Magni M, Siena S, et al. Human peripher-al blood hematopoietic progenitors are optimal tar-gets of retroviral-mediated gene transfer. Blood 1992;80:1418-22.

54. Peters SO, Kittler ELW, Ramshaw HS, et al. Ex-vivoexpansion of murine marrow cells with interleukin-3(IL-3), IL-6, IL-11 and stem cell factor leads toimpaired engraftment in irradiated host. Blood 1996;87:30-7.

55. Lemoli RM, Tafuri A, Fortuna A, et al. Cycling status

of CD 34+ cells mobilized into peripheral blood ofhealthy donors by recombinant human granulocytecolony-stimulating factor. Blood 1997; in press.

56. Pettengel R, Luft T, Henschler R, et al. Direct com-parison by limiting dilution analysis of long-term cul-ture-initiating cells in human bone marrow, umbelicalcord blood, and blood stem cells. Blood 1994; 84:3653-9.

57. Murray L, Chen B, Galy A, et al. Enrichment of humanhematopoietic stem cell activity in the CD34+ Thy-1+Lin– subpopulation from mobilized peripheral blood.Blood 1995; 85:368-78.

58. Sutherland HJ, Eaves CJ, Lansdorp PM, Phillips GL,Hogge DH. Kinetics of committed and primitive bloodprogenitor mobilization after chemotherapy andgrowth factor treatment and their use in autotrans-plants. Blood 1994; 83:3808-14.

59. Dreger P, Klöss M, Petersen B, et al. Autologous prog-enitor cell transplantation: prior exposure to stem cell-toxic drugs determines yield and engraftment ofperipheral blood progenitor cell but not of bone mar-row grafts. Blood 1995; 86:3970-8.

60. Ponchio L, Conneally E, Eaves CJ. Quantitation of thequiescent fraction of longterm culture initianting cells(LTC-IC) in normal human blood and marrow andthe kinetics of their growth factor-stimulated entryinto S-phase in vitro. Blood 1995; 86:3314-21.

61. Donahue RE, Kirby MR, Metzger ME, et al. Peripher-al blood CD 34+ cells differ from bone marrowCD34+ cells in Thy-1 expression and cell cycle statusin nonhuman primates mobilized or not mobilizedwith granulocyte colony-stimulating factor. Blood1996; 87:1644-53.

62. Becker PS, Li Z, Quesenberry PJ. Cytokine regulationof cell adhesion receptor expression in hematopoiet-ic cells. Blood 1994; 84:280a (Suppl. 1).

63. Fleming WH, Alpern EJ, Uchida N, Ikuta K, WeissmanIL. Steel factor influences the distribution and activi-ty of murine hematopoietic stem cells in vivo. ProcNatl Acad Sci USA 1993; 90:3760-4.

64. Kovach NL, Lin N, Yednock T, Harlan JM, Broudy VC.Stem cell factor modulates avidity of a4b1 and a5b1integrins expressed on hematopoietic cell lines. Blood1995; 85:159-67.

65. Möhle R, Murea S, Kirsch M, Haas R. Differentialexpression of L-selectin, VLA-4, and LFA-1 on CD34+progenitor cells from bone marrow and peripheralblood durin G-CSF-enhanced recovery. Ex Hematol1995; 23:1535-42.

66. Gabutti V, Foà R, Mussa F, Aglietta M. Behavior ofhuman hematopoietic stem cells in cord and neona-tal blood. Haematologica 1975; 60:492.

67. Broxmeyer HE, Douglas GW, Hancgoc G, et al.Human umbilical cord as a potential source of trans-plantable hematopoietic stem/progenitor cells. ProcNatl Acad Sci USA, 1989; 86:3828.

68. Gluckman E, Broxmeyer HE, Auerbach DA, et al.Hematopoietic reconstitution in a patient with Fan-coni’s anemia by means of umbilical cord blood froman HLA-identical sibling. N Engl J Med 1989; 321:1174-8.

69. Bertolini F, Lazzari L, Lauri L, et al. A comparativestudy of different procedures for the collection andbanking of umbilical cord blood. J Hematother 1995;4:29-36.

70. Bertolini F, Soligo D, Lazzari L, Corsini C, Servida F,Sirchia G. The effect of interkleukin 12 in ex-vivoexpansion of human hematopoietic progenitors. Br JHaematol 1995; 90:935-8.

71. Aversa F, Tabilio A, Terenzi A, et al. Successful engraft-ment of T-cell-depleted haploidentical “three loci”incompatible transplants in leukemia patients by

65


addition of recombinant human G-CSF-mobilizedperipheral blood progenitor cells to bone marrowinocolum. Blood 1994; 84:3948-55.

72. Wagner JE, Kernan NA, Steinbuch M, Broxmeyer HE,Gluckman E. Allogeneic sibling umbilical-cord-bloodtransplantation in children with malignant and non-malignant disease. Lancet 1995; 346:214-9.

73. Lansdorp PM, Dragowska W, Mayani H. Ontogeny-related changes in proliferative potential of humanhematopoietic cells. J Exp Med 1993; 178:787-91.

74. Hows JM, Bradley BA, Marsh JCW, et al. Growth ofhuman umbilical-cord blood in longterm haemopoi-etic cultures. Lancet 1992; 340:73-6.

75. Payne TA, Traycoff CM, Laver J, Xu F, Srour EF,Abboud MR. Phenotypic analysis of early hematopoi-etic progenitors in cord blood and determination oftheir correlation with clonogenic progenitors: rele-vance to cord blood stem cell transplantation. BoneMarrow Transplant 1995; 15:187-92.

76. Traycoff CM, Abboud MR, Laver J, et al. Evaluation ofthe in vitro behavior of phenotypically defined popu-lations of umbilical cord blood hematopoietic prog-enitor cells. Exp Hematol 1994; 22:215-22.

77. Gabutti V, Timeus F, Ramenghi U, et al. Expansion ofcord blood progenitors and use for hemopoieticreconstitution. Stem Cells 1993; 11(Suppl. 2):105-12.

78. Moritz T, Keller DC, Williams DA. Human cord bloodcells as targets for gene transfer. Potential use in genet-ic therapies of SCID. J Exp Med 1993; 178:529-36.

79. Kohn DB, Weinberg KI, Nolta JA, et al. Engraftmentof gene-modified umbilical cord blood cells inneonates with adenosine deaminase deficiency.Nature Med 1995; 1:1017-23.

80. Bertolini F, De Monte L, Corsini C, et al. Retrovirus-mediated transfer of the multidrug resistance geneinto human haematopoietic progenitors. Br J Haema-tol 1994; 88:318-24.

81. Bertolini F, Battaglia M, Corsini C, et al. Engineeredstromal layers and continuous flow culture enhancemultidrug resistance gene transfer in hematopoieticprogenitors. Cancer Res 1996; 56:2566-72.

82. Traycoff CM, Abboud MR, Laver J, Clapp DW, SrourEF. Rapid exit from G0/G1 phases of cell cycle inresponse to stem cell factor confers on umbilical cordblood CD34+ cells an enhanced ex vivo expansionpotential. Exp Hematol 1994; 22:1264-72.

83. Lambertenghi-Deliliers G, Bertolini F, Della Volpe A,et al. Unrelated mismatched cord blood transplanta-tion in an adult patient with secondary AML. BoneMarrow Transplant 1997; in press .

84. Traycoff CM, Cornetta K, Yoder MC, et al. Ex vivoexpansion of murine hematopoietic progenitor cellsgenerates classes of expanded cells possessing differ-ent levels of bone marrow repopulating potential. ExpHematol 1996; 24:229-306.

85. Rubinstein P, Rosenfield RE, Adamson JW, StevensCE. Stored placental blood for unrelated bone mar-row reconstitution. Blood 1993; 7:1679-90.

86. Kögler G, Somville Th, Adams O, et al. Critical stan-dards for stem cell preparations from unrelated cordblood within Eurocord. Exp Hematol 1995; 22:898.

87. Gluckman E, Devergie A, Thierry D, et al. Clinicalapplication of stem cells transfusion from cord bloodand rationale for cord blood banking. Bone MarrowTransplant 1992; 9 Suppl 1:114-7.

88. Turner CW, Luzins J, Hutcheson CA. A modified har-vest technique for cord blood hematopoietic stemcells. Bone Marrow Transplant 1992; 10:89.

89. Abecasis MM, Machado AM, Boavida G, et al. Hap-loidentical cord blood transplant contaminated withmaternal T cells in a patient with advanced leukemia.

Bone Marrow Transplant 1996; 17:891-5.90. Yao AC, Moinian M, Lind J. Distribution of blood

between infant and placenta after birth. Lancet 1969;2:871.

91. Yao AC, Lind J. Effect of gravity on placental transfu-sion. Lancet 1969; 2:508.

92. Nelle M, Zilow EP, Kraus M, Bastert G, Linderkamp O.The effect of the Leboyer delivery on blood viscosityand other hemorheologic parameters in termneonates. Am J Ob Gyn 1993; 169:189-93.

93. Sinclair JC, Bracken MB. eds. Effective care of new-born infant. Oxford: Oxford Medical Publ.; 1992.

94. Roberton NRC, ed. Textbook of neonatology. Lon-don: Churchill Livingstone; 1992.

95. Pritchard JA, McDonald PC, Grant NF, eds. William’sObstetrics. Norwalk: Appelton-Century-Crofts; 1985.

96. Taeush HW, Ballard RA, Avery ME. eds. Schaffer andAvery’s diseases of the newborn. Philadelphia :WBSaunders; 1991.

97. Polin RA, Fox WW, eds. Fetal and neonatal physiolo-gy. Philadelphia: WB Saunders; 1992.

98. Bertolini F, Battaglia M, De Iulio C, et al. Placentalblood collection: Effects on newborns. Blood 1995;86:4699.

99. Harris DT, Schumacher MJ, Rychlik S, et al. Collec-tion, separation and cryopreservation of umbilicalcord blood for use in transplantation. Bone MarrowTransplant 1994; 13:135-43.

100. Almici C, Carlo-Stella C, Mangoni L, et al. Densityseparation of umbilical cord blood and recovery ofhemopoietic progenitor cells: Implications for cordblood banking. Stem Cells 1995; 13:533-40.

101. Rubinstein P, Dobrila L, Rosenfield RE, et al. Pro-cessing and cryopreservation of placental/umbilicalcord blood for unrelated bone marrow reconstitution.Proc Natl Acad Sci USA 1995; 92:10119-22.

102. Bertolini F, Battaglia M, Zibera C, et al. A new methodfor placental/cord processing in the collection bag.Analysis of factors involved in red blood cell removal.Bone Marrow Transplant 1996; in press.

103. Pahwa RN, Fleischer A, Than S, Good RA. Successfulhematopoietic reconstitution with transplantation oferythrocyte-depleted allogeneic human umbilical cordblood cells in a child with leukemia. Proc Natl AcadSci USA 1994; 91:4485-8.

104. Di Giusto DL, Lee R, Moon J, et al. Hematopoieticpotential of cryopreserved and ex vivo manipulatedumbilical cord blood progenitor cells evaluated in vit-ro and in vivo. Blood 1996; 87:1261-71.

105. Petzer AL, Hogge DE, Lansdorp PM, Reid DS, EavesCJ. Self-renewal of primitive hemetopoietic cells (long-term culture-initiating cells) in vitro and their expan-sion in defined medium. Proc Natl Acad Sci USA1996; 93:1470-4.

106. Brugger W, Heimfeld S, Berenson RJ, Mertelsmann R,Kanz L. Reconstitution of hematopoiesis after high-dose chemotherapy by autologous progenitor cellsgenerated ex vivo. N Engl J Med 1995; 333:283-7.

107. Williams SF, Lee WL, Bender JG, et al. Selection andexpansion of peripheral blood CD34+ cells in autolo-gous stem cell transplantation for breast cancer.Blood 1996; 87:1687-91.

108. Choi ES, Nichol JL, Hokom MM, et al. Platelets gen-erated in vitro from proplatelet-displaying humanmegakaryocytes are functional. Blood 1995; 85:402-13.

109. Okamoto Y, Takaue Y, Saito S, et al. Toxicities asso-ciated with cryopreserved and thawed peripheralblood stem cell autografts in children with active can-cer. Transfusion 1993; 33:578-81.

110. Burgio GR, Hanson LA, Ugazio AG, eds. Immunolo-gy of the neonate. Berlin: Springer-Verlag; 1987. p.


66


F. Bertolini et al.

188.111. Jacoby DR, Olding LB, Oldstone MBA. Immunologic

regulation of fetal-maternal balance. Adv Immunol1984; 35:157-208.

112. Picker LJ, Treer JR, Ferguson-Darnell B, Collins PA,Buck D, Terstappen LWMM. Control of lymphocyterecirculation in man. J Immunol 1993; 150:1105-21.

113. Clement LT, Vink PE, Bradley GE. Novel immunoreg-ulatory functions of phenotypically distinct subpopu-lations of CD4+ cells in the human neonate. JImmunol 1990; 145:102-8.

114. Burgio GR, Curtoni E, Genova R, Magrini U. Skin reac-tivity in childhood: phytohemagglutinin skin test andstreptokinase skin test. Ped Res 1971; 5:88-93.

115. Clerici M, De Palma L, Roilides E, Baker R, ShearerGM. Analysis of T helper and antigen-presenting cellfunctions in cord blood and peripheral blood leuko-cytes from healthy children of different ages. J ClinInvest 1993; 91:2829-36.

116. Maccario R, Nespoli L, Vitiello A, Ugazio AG, BurgioGR. Lymphocyte subpopulations in the neonate: iden-tification of an immature subset of OKT8-positive,OKT3-negative cells. J Immunol 1983; 130:1129-31.

117. Harris DT, Schumacher MJ, Locascio J, et al. Pheno-typic and functional immaturity of human umbilicalcord blood T lymphocytes. Proc Natl Acad Sci USA1992; 89:10006-10.

118. Moller G. Ontogeny of human lymphocyte function.Immunol Rev 1981; 57:161.

119. Keever CA, Abu-Hajir M, Graf W, et al. Characteriza-tion of alloreactivity and anti-leukemia reactivity ofcord blood mononuclear cells. Bone Marrow Trans-plant 1995; 15:407-19.

120. Brugnoni D, Airo P, Graf D, et al. Ineffective expres-sion of CD40 ligand on cord blood T cells may con-tribute to poor immunoglobulin production in thenewborn. Eur J Immunol 1994; 24:1919-24.

121. Montagna D, Maccario R, Ugazio AG, Mingrat G,Burgio GR. Natural cytotoxicity in the neonate: highlevels of lymphokine activated killer (LAK) activity. ClinExp Immunol 1988; 71:158-65.

122. Moretta A, Comoli P, Montagna D, et al. High fre-quency of EBV-lymphoblastoid cell line-reactive lym-phocytes in cord blood: evaluation of cytolytic activi-ty and IL-2 production. Clin Exp Immunol 1996; inpress.

123. Moller G. Immunology of feto-maternal relationship.Immunol Rev 1983; 75:175.

124. Claas FHJ, Gijbels Y, van der Velden-de Munck J, vanRood JJ. Induction of B cell unresponsiveness to non-inherited maternal antigens during fetal life. Science1988; 241:1815-7.

125. Zhang L, van Bree S, van Rood JJ, Claas FHJ. The influ-ence of breast-feeding on the T cell allorepertoire.Transplantation 1991; 52:914-6.

126. Wagner JE, Kernan NA, Steinbuch M, Broxmeyer HE,Gluckman E. Allogeneic sibling umbilical-cord-bloodtransplantation in children with malignant and non-malignant disease. Lancet 1995; 346:214-9.

127. Burnet FM. The clonal selection theory of acquiredimmunity. Nashville: Vanderbilt Univ. Press; 1959.

128. Ridge JP, Fuchs EJ, Matzinger P. Neonatal tolerancerevisited: turning on newborn T cells with dendriticcells. Science 1996; 271:1723-6.

129. Sarzotti M, Robbins DS, Hoffman PM. Induction ofprotective CTL responses in newborn mice by a murineretrovirus. Science 1996; 271:1726-8.

130. Forsthuber T, Yip HC, Lehmann PV. Induction of TH1and TH2 immunity in neonatal mice. Science 1996;271:1728.

131. Risdon G, Gaddy J, Stehman FB, Broxmeyer HE. Pro-liferative and cytotoxic responses of human umbilical

cord blood T lymphocytes following allogeneic stim-ulation. Cell Immunol 1994; 154:14-24.

132. Harris DT, Schumacher MJ, LoCascio J, Booth A, BardJ, Boyse EA. Immunoreactivity of umbilical cord bloodand post-partum maternal peripheral blood withregard to HLA-haploidentical transplantation. BoneMarrow Transplant 1994; 14:63-8.

133. Harris DT, LoCascio J, Besencon FJ. Analysis of thealloreactive capacity of human umbilical cord blood:implications for graft-versus-host disease. Bone Mar-row Transplant 1994; 14:545-53.

134. Deacock SJ, Schwarer AP, Bridge J, Batchelor JR, Gold-man JM, Lechler RI. Evidence that umbilical cordblood contains a higher frequency of HLA class II-spe-cific alloreactive T cells than adult peripheral blood.Transplantation 1992; 53:1128-34.

135. Apperley JF. Umbilical cord blood progenitor celltransplantation. Bone Marrow Transplant 1994; 14:187-96.

136. Risdon G, Gaddy J, Horie M, Broxmeyer HE. Alloanti-gen priming induces a state of unresponsiveness inhuman umbilical cord blood T cells. Proc Natl AcadSci USA 1995; 92:2413-7.

137. Harris DT. In vitro and in vivo assessment of the graft-versus-leukemia activity of cord blood. Bone MarrowTransplant 1995; 15:17-23.

138. Butturini A, Franceschini F, Gale RP. Critical analysisof T-cell depletion in man. In: Martelli MF, GrignaniF, Reisner Y, eds. T-cell depletion in allogeneic bonemarrow transplantation. Rome: Serono SymposiaReview; 1988. p. 1-13.

139. Ferrara JLM, Deeg HJ. Graft-versus-host disease. NEngl J Med 1991; 324:667-74.

140. Reisner Y, Kapoor N, Kirkpatrick D, et al. Transplan-tation for severe combined immunodeficiency withHLA-A, B, D, Dr incompatible parental marrow frac-tioned by soybean agglutinin and sheep red bloodcells. Blood 1983; 61:341-8.

141. O’ Reilly RJ, Kernan N, Cunningham I, et al. Soybeanlectin agglutination and E-rosette depletion forremoval of T-cells from HLA-identical and non-iden-tical marrow grafts administered for the treatment ofleukemia. In: Martelli MF, Grignani F, Reisner Y, eds.T cell depletion in allogeneic bone marrow transplan-tation. Rome: Serono Symposia Review; 1988. p. 123-9.

142. Ash RC, Casper JT, Chitamba CR, et al. Successfulallogeneic transplantation of T-cell depleted bonemarrow from closely HLA-matched unrelated donors.N Engl J Med 1990; 322:485-94.

143. Drobyski WR, Ash RC, Casper JT, et al. Effect of T-celldepletion as graft-versus-host disease prophylaxis onengraftment, relapse, and disease-free survival in unre-lated marrow transplantation for chronic myeloge-nous leukemia. Blood 1994; 83:1980-7.

144. Kernan NA, Flomenberg N, Dupont B, et al. Graftrejection in recipients of T-cell-depleted HLA-non-identical marrow transplants for leukemia. Trans-plantation 1987; 43:842-7.

145. Goldman JM, Gale RP, Horowitz MM, et al. Bonemarrow transplantation for chronic myelogenousleukemia in chronic phase: increased risk of relapseassociated with T cell depletion. Ann Intern Med1988; 108:806-14.

146. Horowitz MM, Gale RP, Sondel PM, et al. Graft-ver-sus-leukemia reactions after bone marrow transplan-tation. Blood 1990; 75:555-62.

147. O’Reilly R. T-cell depletion and allogeneic bone mar-row transplantation. Semin Hematol 1992; 29(Sup-pl. 1):20-6.

148. Kessinger A, Smith DM, Strandjord SE, et al. Allo-geneic transplantation of blood derived, T-cells

67


depleted hemopoietic stem cells after myeloablativetreatment in a patient with acute lymphoblasticleukemia. Bone Marrow Transplant 1989; 4:643-6.

149. Suzue T, Kawano Y, Takaue Y, et al. Cell processingprotocol for allogeneic peripheral blood stem cellsmobilized by granulocyte colony-stimulating factor.Exp Hematol 1994; 22:888-92.

150. Keij JF, Groenewegen AC, Visser JWM. High-speedphotodamage cell sorting: an evaluation of the ZAP-PER prototype. In: Darzynkiewicz Z, Robinson JP,Crissman HA, eds. Methods in cell biology. New York:Academic Press; 1994. p. 371-85.

151. Link H, Arseniev L, Bahre O, et al. Combined trans-plantation of allogeneic bone marrow and CD34+blood cells. Blood 1995; 86:2500-8.

152. Arseniev L, Tischler HJ, Battmer K, et al. Treatment ofpoor graft function with allogeneic CD34+ cellsimmunoselected from G-CSF-mobilized peripheralblood progenitor cells of the marrow donor. BoneMarrow Transplant 1994; 14:791-7.

153. Di Persio J, Martin B, Abbond C, et al. Allogeneic BMTusing bone marrow and CD34-selected mobilizedPBSC; comparison to BM alone and mobilized PBSCalone. Exp Hematol 1994; 22:697a.

154. Lemoli RM, Tazzari PL, Fortuna A, et al. Positive selec-tion of hematopoietic CD34+ stem cells provides“indirect purging” of CD34– lymphoid cells and thepurging efficiency is increased by anti CD2 and anti-CD30 immunotoxins. Bone Marrow Transplant 1994;13:465-71.

155. Behringer D, Bertz H, Hardung-Backes M, et al. Allo-geneic peripheral blood progenitor cell transplanta-tion with or without CD34+ stem cell selection: followup and immune reconstitution. Bone Marrow Trans-plant 1996; 17 (Suppl. 1):298a.

156. Kernan NA, Collins NH, Juliano L, et al. Clonable Tlymphocytes in T cell-depleted bone marrow trans-plants correlate with development of graft-vs-host dis-ease. Blood 1986; 68:770-3.

157. Aversa F, Terenzi A, Tabilio A, et al. Addition of PBSCsto the bone marrow inoculum allows engraftment ofT-cell-depleted HLA-incompatible transplants. Br JHaematol 1996; 93 (Suppl. 2):146.

158. Tabilio A, Falzetti F, Aversa F, et al. T cell depletion ofperipheral blood stem cells by CD34+ cell selectionand E-rosettes. Submitted for publication.

159. Fernandez JM, Yan Y, Bleans S, et al. T-cell depletionof peripheral blood progenitor cells (PBPC) by eitherCD34+ selection and E-rosette depletion or SBAagglutination and E-rosette depletion: comparison ofT-cell and hematopoietic cell progenitor recovery.Blood 1995; 86 (Suppl. 1):626a.

160. Slaper-Cortenbach ICM, Wijngaarden-du Bois MJGJ,de Vriesvan Rossen A, et al. New immunorosette tech-nique for the depletion of T cells from allogeneic stemcell transplantation. Br J Haematol 1996; 93(Suppl.2):170a.

161. Marolleau JP, Dal Cortivo L, Robert J, et al. The newclinical grade peptide PR34+TM stem cell releasingagent. In: Latest advancements in immunomagneticCell selection for grafts engineering in autologous andallogeneic stem cell transplantation. EBMT Baxtersatellite symposium. Vienna, March 3rd, 1996.

162. Kunkel L, Mills B, Burgess J, et al. Early experienceswith selection of CD34+ cells and peptide releaseusing a fully automated selection process. In: Latestadvancements in immunomagnetic Cell selection forgrafts engineering in autologous and allogeneic stemcell transplantation. EBMT Baxter satellite sympo-sium. Vienna, March 3rd, 1996.

163. Berenson RJ, Bensinger WI, Hill RS, et al. Engraftmentafter infusion of CD34+ marrow cells in patients with

breast cancer or neuroblastoma. Blood 1991; 77:1717-22.

164. Gorin NC, Lopez M, Laporte JP, et al. Preparation andsuccesful engraftment of purified CD34+ bone mar-row progenitor cells in patients with non Hodgkin’slymphoma. Blood 1995; 85:1647-54.

165. Lemoli RM , Fortuna A, Motta MR, et al. Concomi-tant mobilization of plasma cells and hematopoieticprogenitors into peripheral blood of multiple myelo-ma patients: positive selection and transplantation ofenriched CD34+ cells to remove circulating tumorcells. Blood 1996; 87:1625-34.

166. Steinman RM. The dendritic cell system and its role inimmunogenicity. Annu Rev Immunol 1991; 9:271-96.

167. Caux C, Vanbervliet B, Massacrier C, et al. B70/B7-2is identical to CD86 and is the major functional ligandfor CD28 expressed on human dendritic cells. J ExpMed 1994; 180:1841-7.

168. Sallusto F, Lanzavecchia A. Efficient presentation ofsoluble antigen by cultured human dendritic cells ismaintained by granulocyte/macrophage colony-stim-ulating Factor plus Interleukin 4 and downregulatedby tumor necrosis factor. J Exp Med 1994; 179:1109-18.

169. Romani R, Gruner S, Brang D, et al. Proliferating den-dritic cell progenitors in human blood. J Exp Med1994; 180:83-93.

170. Hart DN, Starling GC, Calder VL, Fernando NS.B7/BB-1 is a leukocyte differentiation antigen onhuman dendritic cells induced by activation. Immu-nology 1993; 79:616-20.

171. Guinan EC, Gribben JG, Boussiotis VA, Freeman GJ,Nadler LM. Pivotal role of the B7:CD28 pathway intransplantation tolerance and tumor immunity. Blood1994; 84:3261-82.

172. Reid CDL, Stackpoole A, Meager A, Tikerpae J. Inter-actions of tumor necrosis factor with granulocyte-macrophage colony-stimulating factor and othercytokines in the regulation of dendritic cell growth invitro from early bipotent CD34+ progenitors inhuman bone marrow. J Immunol 1992; 149:2681-8.

173. Szalbocs P, Moore MAS, Young JW. Expansion ofimmunostimulatory dendritic cells among the myeloidprogeny of human CD34+ bone marrow precursorscultured with c-kit ligand, GM-CSF and TNF-α. JImmunol 1995; 154:5851-61.

174. Young JW, Szalbocs P, Moore MAS. Identification ofdendritic cell colony-forming units among normalhuman CD34+ bone marrow progenitors that areexpanded by c-kit ligand and yeld pure dendritic cellcolonies in the presence of granulocyte/macrophagecolony-stimulating factor and tumor necrosis factor-α. J Exp Med 1995; 182:1111-20.

175. Siena S, Di Nicola M, Bregni M, et al. Massive ex vivogeneration of functional dendritic cells from mobi-lized CD34+ blood progenitors for anticancer thera-py. Exp Hematol 1995; 23:1463-71.

176. Rosenzwaig M, Canque B, Gluckman JC. Human den-dritic cell differentiation pathway from CD34+ hema-topoietic precursor cells. Blood 1996; 87:535-44.

177. Strunk D, Rappesberger K, Egger C, et al. Generationof human dendritic cells/Langerhans cells from circu-lating CD34+ hematopoietic progenitor cells. Blood1996; 87:1292-302.

178. Schwartz RH. A cell culture model for T lymphocyteclonal anergy. Science 1990; 48:1349-56.

179. Tan P, Anasetti C, Hansen JA, et al. Induction ofalloantigen-specific hyporesponsiveness in human Tlymphocytes by blocking interaction of CD28 with itsnatural ligand B7/BB1. J Exp Med 1993; 177:165-73.

180. DeSilva DS, Urdhal KB, Jenkins MK. Clonal anergy isinduced in vitro by T cell receptor occupancy in the


68


absence of proliferation. J Immunol 1991; 147:2461-6.

181. Jenkins MK. The ups and downs of T cell costimula-tion. Immunity 1994; 1:443-6.

182. Rondelli D, Andrews RG, Hansen JA, Ryncarz R, Faer-ber MA, Anasetti C. Alloantigen presenting function ofnormal human CD34+ hematopoietic cells. Blood1996; 88:2619-25.

183. Gunji Y, Nakamura M, Yanigisawa M, Miura Y, SudaT. Expression and function of adhesion molecules onhuman hematopoietic stem cells: CD34+LFA-1- cellsare more primitive than CD34+LFA-1+ cells. Blood1992; 80:429-36.

184. Krensky AM, Mentzer SJ, Clayberger C, et al. Herita-ble lymphocyte function-associated antigen-1 defi-ciency: abnormalities of cytotoxicity and proliferationassociated with abnormal expression of LFA-1. JImmunol 1985; 135:3102-8.

185. Andrews RG, Bryant EM, Bartelmez SH, et al. CD34+marrow cells, devoid of T and B lymphocytes, recon-stitute stable lymphopoiesis and myelopoiesis inlethally irradiated allogeneic baboons. Blood 1992;80:1693-701.

186. Bensinger WI, Rowley S, Appelbaum FR, et al. CD34selected allogeneic peripheral blood stem cell (PBSC)transplantation in older patients with advancedhematologic malignancies. Blood 1995; 86 (Suppl.1):376.

187. Schiller G, Rowley S, Buckner CD, et al. Transplanta-tion of allogeneic CD34+ peripheral blood stem cells(PBSC) in older patients with advanced hematologicmalignancy. Blood 1995; 86 (Suppl.1):1545.

188. Link H, Arseniev L, Bahre O, et al. Transplantation ofallogeneic hematopoiesis by selected CD34+ bloodcells. Blood 1995; 86 (Suppl.1):1150.

189. Urbano-Ispizua A, Rozman C, Marìn P, et al. Selectionof CD34+ peripheral blood progenitor cells (PBPC)for allogeneic transplantation. Blood 1995; 86(Sup-pl.1):901.

190. Holland HK, Bray AM, Geller RB, et al. Transplanta-tion of HLA-identical positively selected CD34+ cellsof peripheral blood (PBSC) and marrow (BM) fromrelated donors results in prompt engraftment and lowincidence of GVHD in recipients undergoing allo-geneic transplantation. Blood 1995; 86(Suppl.1):1543.

191. Weaver CH, Longin K, Buckner CD, et al. Lymphocytecontent in peripheral blood mononuclear cells col-lected after the administration of recombinant humangranulocyte colony-stimulating factor. Bone MarrowTransplant 1994; 13:411-5.

192. Nicola NA. Why do hemopoietic growth factor recep-tors interact with each other? Immunol Today 1987;8:134-40.

193. Noga SJ, Davis JM, Vogelsgang GB, et al. The com-bined use of elutriation and CD8/magnetic bead toengineer the bone marrow allograft. In: Worthington-White DA, Gee AP, Gross S, eds. Advances in bonemarrow purging and processing. New York: Wiley-Liss;1992. p. 411.

194. Owens AH, Santos GW. The induction of graft versushost disease in mice treated with cyclophosphamide.J Exp Med 1968; 128:277-91.

195. van Bekkum DW, de Vries MJ. Radiation chimaeras.New York: Academic Press; 1967.

196. Atkinson K, Farrell C, Chapman G, et al. Female mar-row donors increase the risk of acute graft-versus-host-disease: effect of donor age and parity and analy-sis of cell subpopulations in the donor marrow inocu-lum. Br J Haematol 1986; 63:231-9.

197. Jansen J, Goselink HM, Veenhof WFJ, et al. The impactof the composition of the bone marrow graft onengraftment and graft-versus-host-disease. Exp Hema-tol 1983; 11:967-73.

198. Murphy WJ, Reynolds CW, Tiberghien P, Longo DL.Natural killer cells and bone marrow transplantation.J Natl Cancer Inst 1993; 85:1475-82.

199. Zeis M, Uharek L, Glass B, et al. Natural killer cellsgiven after allogeneic BMT induce strong graft-vs-leukemia (GVL) effects. Br J Haematol 1996; 93 (sup-pl 2):76a.

200. van Bekkum DW. The selective elimination of immu-nologically competent cells from bone marrow andlymphatic cell mixtures. Effect of storage at 4°C.Transplantation 1964; 2:393-8.

201. Eckardt JR, Roodman GD, Boldt DH, et al. Compar-ison of engraftment and acute GvHD in patientsundergoing cryopreserved or fresh allogeneic BMT.Bone Marrow Transplant 1993; 11:125-31.

202. Dreger P, Suttorp M, Haferlach T, et al. Allogeneicgranulocyte colony-stimulating factor mobilized peri-pheral blood progenitor cells for treatment of engraft-ment failure after bone marrow transplantation.Blood 1993; 81:1404-7.

203. Rambaldi A, Viero P, Bassan R, et al. G-CSF-mobi-lized peripheral blood progenitor cells for allogeneictransplantation of resistant or relapsing acuteleukemias. Leukemia 1996; 10:860-5.

204. Kurtzberg J, Laughlin M, Graham ML, et al. Placentalblood as a source of hematopoietic stem cells fortransplantation into unrelated recipients. N Engl JMed 1996; 335:157-66.

205. Laporte JP, Gorin NC, Rubinstein P, et al. Cord-bloodtransplantation from an unrelated donor in an adultwith chronic myelogenous leukemia. N Engl J Med1996; 335:167-70.

F. Bertolini et al.


review

Clinical use of allogeneichematopoietic stem cells fromsources other than bone marrow


WILLIAM ARCESE,* FRANCO AVERSA,° GIUSEPPE BANDINI,#

ARMANDO DE VINCENTIIS,@ MICHELE FALDA,^ LUIGI LANATA,@

ROBERTO M. LEMOLI,# FRANCO LOCATELLI,§ IGNAZIO MAJOLINO,**

PAOLA ZANON,°° SANTE TURA#

*Department of Cellular Biotechnology and Hematology, Univer-sity “La Sapienza”, Rome; °Department of Clinical Medicine,Pathology and Pharmacology, Section of Clinical Hematologyand Immunology, University of Perugia, Perugia; #Institute ofHematology and Medical Oncology "L. & A. Seragnoli", Universi-ty of Bologna, Bologna; @Dompé Biotec SpA, Milan; ^Division ofHematology, Ospedale S. Giovanni Battista, Turin; §Departmentof Pediatrics, University of Pavia and IRCC Policlinico S. Matteo,Pavia; **Department of Hematology and BMT Unit, Ospedale “V.Cervello”, Palermo; °°Amgen Italia SpA, Milan, Italy

Correspondence: Prof. Sante Tura, Istituto di Ematologia ed Oncolo-gia Medica “L. & A. Seràgnoli”, Policlinico S. Orsola, via Massarenti9, 40138 Bologna, Italy.

Background and Objectives. Peripheral blood stem cells(PBSC) are being increasingly used as an alternative toconventional allogeneic bone marrow (BM) transplanta-tion. This has prompted the Working Group on CD34-Positive Hematopoietic Cells to evaluate current utiliza-tion of allogeneic PBSC in clinical hematology.

Evidence and Information Sources. The methodemployed for preparing this review was that of infor-mal consensus development. Members of the WorkingGroup met three times, and the participants at thesemeetings examined a list of problems previously pre-pared by the chairman. They discussed the singlepoints in order to reach an agreement on differentopinions and eventually approved the final manuscript.Some of the authors of the present review have beenworking in the field of stem cell transplantation andhave contributed original papers in peer-reviewed jour-nals. In addition, the material examined in the presentreview includes articles and abstracts published injournals covered by the Science Citation Index® andMedline®.

State of the Art. Review of the current literature showsthat unmanipulated allogeneic PBSC give prompt andstable engraftment in HLA-identical sibling recipients.Despite the much higher number of T-cells infused, theincidence and severity of acute GVHD after PBSC trans-plant seems comparable to that observed with bonemarrow (BM) cells. In comparison to the latter, PBSCprobably ensure faster immunologic reconstitution inthe early post-transplant period. Controversial resultson the incidence and severity of acute-GVHD have beenreported when CD34+ selection methods are used.Prospective randomized trials are underway to comparethe results of PBSC and BM allogeneic transplantation.In mismatched family donor transplants, T-cell deplet-ed PBSC successfully engraft immune-myeloablatedrecipients through a mega-cell-dose effect able to over-come the HLA barrier. Experience with PBSC in the con-text of unrelated donor transplants is currently anecdo-tal and prospective trials should be completed beforethat practice becomes routine. Finally, there is alsolimited evidence that, following induction chemo-therapy, the addition of PBSC to donor lymphocyteinfusion (DLI) for treatment of leukemia relapse after

BMT may improve the safety and effectiveness of DLIitself. Concerning cord blood (CB) transplants, themost interesting aspects are the ease of CB collectionand storage, the low risk of viral contamination and thelow immune reactivity of CB cells. This last property hasits clinical counterpart in an apparently reduced inci-dence and severity of acute GVHD both in sibling andunrelated CB transplants, probably making the level ofdonor/recipient HLA disparity acceptable a greaterdegree with respect to what is required for transplantsfrom other sources. ©2000, Ferrata Storti Foundation

Key words: hematopoietic stem cells, bone marrow, cord blood, periph-eral blood, allogeneic transplantation, graft-versus-host disease

In the field of allogeneic transplantation the use ofalternative sources of stem cells, namely peripheralblood stem cells (PBSC)1 and placental cord blood (CB)

stem cells,2 is rapidly expanding. The European Group forBlood and Marrow Transplantation (EBMT) registeredonly 12 allogeneic PBSC transplants in 1993, but thisnumber increased to 180 in 1994 and to 571 in 1995.3Concerning cord blood (CB) transplants, following initialattempts4,5 considerable experience has now beenachieved in the USA and Europe so that this modality isentering a phase of extensive clinical application, withhundreds of procedures registered both from sibling andunrelated donors,6,7 The ease of collection and storage ofCB stem cells and the apparent tolerance-inducing prop-erty of CB CD8+ suppressor cells8 are the most interest-ing aspects of this latter source.

Stem cells in peripheral blood and CB both possess amigratory status and differ in part from those found inbone marrow with respect to their biological and func-tional properties. However, while PBSC are envisaged asa means of improving results by increasing the numberof cells available, placental CB stem cells open the real-istic perspective of increasing the number of transplantsthanks to the availability of thousands of cord samplesfor patients who lack a compatible donor among family

70


members.This search for new stem cell sources also arises from

the fact that allogeneic bone marrow transplantation stillcarries a high procedure-related mortality and diseaserecurrence rate. Cell dose has an influence on engraft-ment and chance of survival. In a retrospective studyBacigalupo et al.9 showed that patients with hematolog-ic malignancies who receive allogeneic bone marrowgrafts with higher CFU-GM numbers have significantlyhigher platelet counts on day +80 and a lower mortalityrate than those who receive fewer CFU-GM. The effect ofCMV infection on platelet counts also appears to be lesspronounced when the number of progenitor cells is high-er. The use of more hematopoietic progenitors would thenresult in improved transplant outcome.

The growing interest in allogeneic PBSC induced theItalian Bone Marrow Transplant Group (GITMO) to drawup a list of recommendations that were originally pub-lished in 199510 and recently revised in light of theincreasing experience gained worldwide during the lasttwo years.11 Moreover, in a previous issue of this Journal12

a review article analyzed the biological and technicalaspects of PB and CB stem cells. The key aspects dealtwith were the mobilization and collection methods, thecapacity for stable hemopoietic reconstitution, kineticcharacteristics and immunological features.

Historical background Interest in allogeneic transplantation of PBSC began

thirty years ago. In the late sixties, based on an earlierdemonstration that autologous PBSC were capable ofrestoring irradiation-myeloablated hematopoiesis,13-15

the Seattle group reported the first successful attemptsat allogeneic PBSC transplantation in dogs16,17 and non-human primates.18 However, due to the high GVHD inci-dence, those experiments were unable to demonstratelong-term stability of the graft. Only a decade later didpurification of PBSC and application of cytogeneticmethods allow a group of German investigators19,20 todocument in dogs the stability of donor-derived hemo-poietic function for more than ten years after PBSC allo-geneic transplantation.

A key issue at that time was the low number of prog-enitors in steady-phase peripheral blood, and clinicalapplication of PBSC was limited to autologous trans-plantation in CML,21 where hematopoietic progenitors,mostly of the leukemic counterpart, circulate in highnumbers and can easily be collected by apheresis with-out any prior stimulation. A further step was the demon-stration that the PBSC level increases dramatically dur-ing the post-chemotherapy recovery phase;22 however, itwas the advent of G-CSF and GM-CSF that provided therapid expansion of PBSC technology and led to their usein allogeneic transplantation as well. The pioneer workof Socinski et al.23 and Gianni et al.24 established the abil-ity of hematopoietic growth factors to expand the cir-culating progenitor cell pool either when used alone orin conjunction with chemotherapy. However, transferringgrowth-factor PBSC mobilization strategy from auto-

graft patients to normal donors took some years. In fact,the safety of growth factors and the clinical applicabil-ity of allogeneic PBSC in terms of GVHD incidence andlong-term engraftment represented serious reasons forcaution. Clinical PBSC allogeneic transplantation beganin 1989, when Kessinger et al.25 reported the first attemptin an HLA-matched recipient with ALL. The patient wasan 18-year-old man in third remission after CNS andtesticular relapse. His sibling female donor preferred todonate PBSC rather than bone marrow, and she under-went 10 apheretic procedures without any mobilizationtreatment. The apheresis product was T-depleted bysheep erythrocyte rosetting and infused after condition-ing with high-dose Ara-C and TBI. The patient achievedfull donor engraftment as demonstrated by cytogeneticstudies but died on day +32, and sustained engraftmentcould not be demonstrated.

Four years later, in 1993, Russell et al.26 reportedanother transplant in a patient whose sibling donor pre-sented an increased risk of complications from anes-thesia. In this case, 10 mg/kg/day of G-CSF were givento mobilize PBSC. The cells were collected at two leuka-phereses containing 36.8×104/kg CFU-GM and infusedwithout any prior manipulation. Engraftment occurredrapidly and GVHD did not develop despite the high T-cellcontent of the graft sample. The same year, a group ofinvestigators from Kiel University27 also successfullyemployed allogeneic PBSC. A 47-year old AML patientwho failed to engraft after bone marrow transplantationfrom an HLA-identical sibling donor was infused withthe unmanipulated product of 3 leukaphereses per-formed after treating the donor with 6 mg/kg/day G-CSF. Engraftment occurred on day +14, with moderateacute GVHD that responded to immunosuppressionstarting on day +18. Restriction fragment length poly-morphism (RFLP) typing demonstrated full donorengraftment up to 60 days following transplantation.

Another important step was the five PBSC transplantsfrom syngeneic donors performed in Seattle and report-ed in 1993:28 with a median of 9.6×106/kg CD34+ cellsinfused, the patients engrafted 0.5×109/L granulocyteson day 13 and 20×109/L platelets on day 10. In 1995three separate reports appeared in the same issue ofBlood, one from Seattle,29 a second from Houston30 andthe other from Kiel;31 a total of 25 patients were allo-grafted with PBSC from their HLA-identical siblingdonors. Acute GVHD was apparently not increased inthose series. Molecular analysis of engraftment30,31 fur-nished definitive proof of the experimental data20 sug-gesting that allogeneic PBSC contain true long-termrepopulating stem cells. The high engraftment potentialof PBSC was exploited by the Perugia team33 to suc-cessfully transplant leukemia patients from their hap-loidentical, three-loci-incompatible family donorsthrough T-cell depletion. Finally, Ringdén et al.34 recent-ly reported the use of allogeneic PBSC in selected unre-lated donor transplants.

The kinetics of PBSC under cytokine mobilization wasextensively analyzed in a previous review published in

W. Arcese et al.

71


this Journal.12 PBSC mobilization in healthy donors isbest accomplished with G-CSF. The aspect of donorsafety was analyzed in a cooperative GITMO study35 in

which short-term side effects were shown to be mini-mal. Ten µg/kg/day of G-CSF for 5 days enabled the col-lection of >4×106/kg CD34+ cells with two aphereses in85% of donors. Variations in blood counts included asharp elevation of WBC and CD34+ cells and a moder-ate transitory thrombocytopenia. One problem, howev-er, is the lack of data on the late effects of G-CSF. At theGeneva conference on allogeneic PBSC, Hasenclever andSextro36 presented a feasibility study of long-term riskanalysis. In order to demonstrate a tenfold increase inleukemia risk, more than 2000 healthy PBSC donorswould have to be followed for over 10 years. A controlgroup of BMT donors of equal size would also be nec-essary. Such a study could only be carried out on a mul-ti-national basis.

Transplantation of allogeneic PBSCfrom HLA-identical siblings

Conditioning regimens and GVHD prophylaxis Conditioning regimens employed in PBSC transplanta-

tion are the same as those used for bone marrow trans-plantation (BMT). As listed in Table 1, the majority ofpatients received CY-TBI or BU-CY with CY at 120 or 200mg/kg. Indeed, 33.8% and 24.5% of reported patientswere conditioned with CY-TBI and BU-CY, respectively.Analogously to BM transplantation, patients with SAAreceived cyclophosphamide alone or in association withATG.

Recently, some innovative regimens have been devel-oped in order to: i) increase antitumor activity; ii) reducetreatment-related toxicity. For instance, high dose Ara-C or VP16 has been employed for patients with moreadvanced disease. Others considered thiotepa, a potentmyeloablative drug first introduced in conditioning bythe Perugia group.37 This latter compound was used alongwith classical BU-CY2 in a large series at the M.D. Ander-son Cancer Center38,39 or was associated with cyclophos-phamide.40 In this study, thiotepa was introduced in thehope of reducing the liver toxicity of busulfan.

It is interesting to note the recent introduction of flu-darabine, a purine analogue initially proposed at con-ventional dosage by the M.D. Anderson group41 and ata higher dose by the Perugia group in HLA-mismatchedtransplants.42 Fludarabine has been associated with sev-eral other drugs or drug combinations includingcyclophosphamide plus cis-platinum, high-dose (HD)Ara-C, idarubicine plus Ara-C or melphalan.41,43 Thesefludarabine containing regimens were followed by fullengraftment with complete donor chimerism in theabsence of severe aplasia. Basically, these new regimensallow allograft even in older or medically infirm patients,since they reduce the toxicity but still maintain an effec-tive graft versus leukemia reaction. For the same reasonAdkins et al.44 combined low-dose TBI (550 rad) at ahigh dose rate (30 cGy/min) with cyclophosphamide 120

mg and methylprednisolone 2 g over two days.A non-myeloablative regimen with busulfan and

methylprednisolone combined with immunosuppressionwith the CD3 monoclonal antibody was used by Tan etal.,45 while Slavin et al.46 employed fludarabine and ATGas intensive immunosuppression associated with busul-fan at 4 mg/kg/day over 2 days. Although these regimensare not specifically designed for PBSC transplantation,the high number of inoculated PB stem cells overcomesgraft rejection, favoring rapid and stable chimerism.

The large amount of stem cells in the inoculum mightexplain the full engraftment observed in a patient whocould not complete the whole regimen and receivedbusulfan alone as preparation.47

The large number of CD3+ve cells present in the PB-derived inoculum has raised some concerns about theseverity of GVHD following PBSC allograft. However,GVHD prophylaxis has not been substantially modifiedfrom the standard regimens used for bone marrow trans-plants. Cyclosporin A (CsA) has been used alone (2.9%)or in association with either methotrexate (MTX) (49.8%)or methylprednisolone (MP) (21.2%) in 71% of patients(Table 1).

New immunosuppressive regimens including tacro-limus (FK506) in association with methylprednisolone,methotrexate39 or monoclonal antibodies48 have beenexplored in 17% of cases.

Finally, some centers have developed techniques for invitro stem cell enrichment. However, using Ceprate-cellseparation, a 2-3 log depletion of T-lymphocytes is notenough to avoid the risk of GVHD and further immuno-suppression is generally required.

In some institutions cryopreservation of collected PBSCis preferred to freshly collected material for several rea-sons. First, cryopreservation allows precise evaluation ofthe hemopoietic progenitor content in the harvestedmaterial. Second, the allograft may be scheduled at theproper time once adequate quantities of PBSC are col-lected. Finally, although still unproven, a reduced risk ofacute GVHD in patients transplanted with cryopreservedBM cells has been suggested.49

Clinical use of allogeneic hematopoietic stem cells

Table 1. Conditioning regimens and GVHD prophylaxis: 301patients.

Regimens No. % GVHD prophylaxis No. %

CYTBI 102 33.8% CsA-MTX 150 49.8%

BUCY 68 24.5% CsA-MP 82 21.2%

TioBUCY 67 22.2% FK506-MTX 7 2.3%

VP16TBI 10 3.3% FK506-MP 44 14.6%

TioCY 8 2.6% CsA 15 2.9%

CYATG 3 0.9% MTX 1 0.3%

Others 43 16% Others 2 0.6%

72


Engraftment Engraftment kinetics following PBSC allograft has been

extensively investigated. None of the studies includepatients dying before day 21. The median time to reachan absolute neutrophil count(ANC) above 0.5×109/Lranges between 10-16 days. The reported incidence ofgraft failure is definitely low: 1/59 in the EBMT survey,50

1/41 in the MD Anderson series,51 1/26 in the Canadianexperience.52 The few rejections occurred in transplantswith 1-2 antigen disparity. Platelet engraftment is alsoprompt, with median time to achieve an absolute plateletcount (APC) of 20×109/L ranging between 10 and 18 daysin the reported series.

Platelet more than neutrophil engraftment may beaffected by acute GVHD or CsA toxicity as well as by ear-ly relapse or progressive disease.50 In a recent report bythe Genoa group a second infusion of PBSC without con-ditioning was required to achieve full engraftment ofplatelets in three out of thirty-one patients.40

The prompt engraftment offered by PBSC implies areduction in transfusion need. The reported transfusionrequirements range from 2 to 10 packed red cell units andfrom 3 to 12 platelet units.

Furthermore, GVHD prophylaxis may adversely influ-ence the time to engraftment. In particular, methotrex-ate given for GVHD prophylaxis delays neutrophil andplatelet engraftment.53,54

Growth factors, mainly G-CSF, have been employed inseveral studies to speed-up engraftment. Urbano-Ispizua reported that G-CSF given to patients not receiv-ing methotrexate accelerates neutrophil recovery(p=0.001); median time to >20×109/L platelets was sig-nificantly delayed (p 0=0.01), although the time to reach503109/L platelets was not affected. No difference inengraftment kinetics was seen between cryopreservedand fresh PBSC when G-CSF was administered follow-ing transplantation.50,55

The studies carried out so far are not sufficient todraw definitive conclusions about engraftment withPBSC as compared to engraftment with BM. Random-ized studies are still in progress and results are not yetavailable. Most information comes from comparisons ofPBSC results with historical data from BM transplants.

A highly informative study was reported by the Seat-tle group, which compared 37 PBSC transplantedpatients with a historical group of 37 bone marrowrecipients.56 Patients were well matched for diagnosis,disease stage, age and graft versus host prophylaxis.Faster neutrophil engraftment, 14 versus 16 days toreach more than 0.5×109/L (p=0.0063), and earlierachievement of platelet transfusion independence, 11versus 15 days (p=0.0014), were observed in PBSC recip-ients compared to the BM control group. Consequent-ly, the median number of platelet units transfused was24 versus 118 (p=0.0001) and the median number of redblood cell units transfused was 8 versus 17 (p=0.0005)in the PBSC group and in the BM group, respectively.

Similar results have been reported by Russel et al.:52

duration of aplasia for both neutrophils (p=0.0002) and

platelets (p=0.0003) was significantly reduced inpatients receiving PBSC compared to BM recipients.Interestingly, the advantage of PBSC was also main-tained if methotrexate was used as GVHD prophylaxis.

A recent report by Rosenfeld et al.57 evaluated 19patients transplanted with PBSC. No growth-factor wasemployed in the post-transplant phase. Significantlyfaster neutrophil recovery was observed in PBSC trans-planted patients compared to historical control grouptransplanted with BM (p=0.01). However, the differencewas not significant when the PBSC group was comparedto BM recipients given G-CSF in the post-transplantphase.

More recently, a prospective non-randomized studywas carried out by the M.D. Anderson group.39 The studyincluded 74 adults transplanted with HLA-matchedrelated donors. Thiotepa, busulfan and cyclophos-phamide were employed as preparative regimen. Thepatients were divided into 3 cohorts: Group 1 receivedBMT using CsA and MTX as GVHD prophylaxis, Group 2received marrow using CsA and MP, and Group 3received PBSC with CsA and MP. All patients were giv-en G-CSF post-transplant. Median time to neutrophils> 0.5×109/L was 17, 9 and 10 days, and to platelets> 20×109/L was 32, 25 and 18 days in Groups 1, 2 and3, respectively. The use of CsA and MP for GVHD pro-phylaxis, rather than the source of engrafted cells wasshown to be the most important factor for rapid neu-trophil and platelet recovery. Provided that CsA/MP wasused for GVHD prophylaxis, platelet transfusion require-ment was found to be significantly lower in PBSC thanin BMT recipients (p=0.04). Significant differences con-cerning regimen-related toxicity were seen for grade 2-4 stomatitis only between the BMT group using MTX inGVHD prophylaxis and the PBSC group using MP.

Correlation between engraftment kinetics and quan-tity of PB cells infused is still an open question. Theabsolute number of nucleated PB cells or CD34+ cells didnot correlate with time to neutrophils > 0.5×109/L orwith time to platelets > 20, > 50 or > 100×109/L in astudy by Rosenfeld et al.57 Similarly, Urbano-Ispizua etal. did not find any correlation using several cut-off val-ues of CD34+ cells at 2.5, 3, 4, 5.5 and 7×106/kg. In con-trast, Roy et al.58 reported a correlation between CD34+

cells infused and engraftment using a mobilization reg-imen with G-CSF at a dose of 5 µg/kg. In a large seriespublished by the M.D. Anderson group,37 in univariateanalysis of patients not given MTX prophylaxis the num-ber of total nucleated cells infused positively affectedANC recovery. Moreover, platelet recovery was positivelyinfluenced by the number of CD34+ cells, as well as byyoung age and sex mis-matching.

Immune reconstitution after transplantationof peripheral blood stem cells

Patients undergoing allogeneic BMT experience a pro-longed period of profound cellular and humoral immun-odeficiency, mainly due to complete pre-transplantdestruction of the host lymphohemopoietic system, the

W. Arcese et al.

73


use of immunosuppressive drugs for GVHD prophylaxisand the development of GVHD.59-61 This immunodefi-ciency lasts until stem cells and mature lymphocytescontained in the transplanted marrow repopulate andreconstruct the hematopoietic and lymphopoietic sys-tems which had been destroyed by the pre-transplantconditioning regimen. In particular, immunologicalreconstitution after BMT is considered to be dependenton two distinct phenomena.59,60 In the early post-trans-plant period, there is an expansion of mature donor-derived lymphocytes transferred with the graft, aprocess influenced by both the recipient’s environmentand the cytokine storm62 related to the transplant pro-cedure. Thereafter, naive lymphocytes derived from thedifferentiation of donor hematopoietic stem cells colo-nize the lymphoid organs of the recipient and sustainthe late immune response.

The crucial role of the first step in immunologicalrecovery is demonstrated by the observations thatpatients receiving a T-cell depleted transplant are atparticular risk for infections and that patients trans-planted using donors either recently vaccinated againstor immune to a certain pathogen usually have a morerapid recovery of specific T-cell response than ones whoreceived bone marrow from unprimed donors.63-65 For-mal proof of the contribution of transferred donor-derived lymphocytes to recipient immune reconstitu-tion has been recently reported.62 In fact, using the com-bination of a cell culture method and a PCR amplifiedtechnique to study tetanus toxoid (TT)-specific T-cellsclones, it was possible to demonstrate that patients afterBMT display a small response that can be accounted forby a few donor-derived clones and that the T-cell clonestransferred with the transplant were still detectablewithin the donor polyclonal T-cell lines for up to at least5 years after BMT. Moreover, the vaccination of donorswith TT before BMT resulted in a more relevant transferof antigen-experienced T-cells.66

The expansion of mature donor-derived lymphocytestransferred with the graft in recipients of peripheralblood stem cell (PBSC) transplantation could be expect-ed to be more efficient than patients given BMT, in viewof the higher number of donor lymphocytes transferred.However, at present, few reports specifically addressingthe question of immune recovery after transplantationof PBSC are available.

Ottinger et al.67 demonstrated that, compared to BMTrecipients, patients who were given a PBSC transplanthad a more rapid recovery of both naive and memoryCD4+ cells (expressing the RA and RO isoforms of theCD45 molecule, respectively) whose counts significant-ly exceeded those observed following marrow trans-plantation. This determined that in patients receivingPBSC transplantation the characteristic inversion of theCD4+/CD8+ ratio observed after BMT was not encoun-tered. Furthermore, the B-cell levels and, at least for thefirst 2 months after transplantation, the monocytecounts were augmented. Since monocytes of granulocytecolony-stimulating factor (G-CSF)-mobilized donors

have been demonstrated to reduce the responsiveness ofalloantigen specific T-cells, the increase in their countcould contribute to the low incidence and reduced sever-ity of acute GVHD reported after transplantation ofPBSC.68 Moreover, it must be noted that there is a promptrecovery of the lymphocyte counts after transplant ofPBSC coupled with an enhanced in vitro response of lym-phocytes to aspecific polyclonal activators (phyto-hemagglutinin and pokeweed mitogen) and to recallantigens (TT, Candida). The most likely hypothesis forexplaining this accelerated recovery of helper T cells, Blymphocytes and monocytes is that the number of lym-phocytes infused for each subset is more that one mag-nitude higher in recipients of PBSC transplant than inpatients given BMT. However, alternative mechanismscannot be excluded.

Similar results in terms of more rapid recovery of CD4+

cells have also been reported by Bacigalupo et al.40 inadults with advanced leukemia who received high-dosechemotherapy followed by G-CSF mobilized PBSC. Morerecently, two additional reports have further confirmedthat recipients of PBSC transplants have a faster recov-ery of both naive and memory helper T cells.69,70 More-over, one of these studies documented that patientsexperiencing a more rapid recovery of the lymphocytecount had a significantly better probability of survivalafter transplantation.69

Whether the improved immune reconstitutionobserved after transplantation of PBSC is associated witha lower incidence of infectious complications stillremains to be documented. In one of the previously men-tioned studies,69 the actuarial risk of reactivation ofhuman cytomegalovirus (HCMV) infection in patientsgiven a PBSC transplant was comparable to thatobserved in a historical control group of BMT recipients.This could be attributed to a greater viral load infusedwith the graft and correlated with the very large num-ber of nucleated cells that can harbor HCMV transfused.Nonetheless, since the use of donor-derived adoptiveimmune therapy has been shown to be able to cure orprevent HCMV-related interstitial pneumonia and EBV-induced lymphoproliferative disorders,71-73 it can behypothesized that patients given PBSC transplants, witha more efficient transfer of antigen-experienced lym-phocytes, may have a reduced incidence and/or reducedseverity of infectious complications. Support for this the-ory is provided by the study reported by Bensinger etal.74 in which a lower number of deaths from infectiouscomplications was observed in patients given PBSC ascompared to a historical group of BMT recipients.

Acute and chronic GVHDPBSC collections contain a large number of T-cells –

approximately 10 times more than unmanipulated mar-row grafts.75 Therefore concern for increased incidenceand severity of GVHD after their infusion into an allo-geneic host has been and still is a major issue after PBSCtransplantation. Here we analyze the results reportedso far in the most recent peer-reviewed studies pub-


74


W. Arcese et al.

lished. Because of the relatively short follow-up of thesestudies, the assessment of chronic GVHD (cGVHD) is lesscomplete and less accurate than that of the acute form.Some of the studies in fact do not address the problemof cGVHD. Acute GVHD, on the other hand, can now beevaluated in a rather significant number of patients. Weshall look first at the characteristics of the studies, thenanalyze acute and cGVHD separately and finally makecomparisons between marrow and PB blood transplants.A set of tentative comments will be made at the end ofthe chapter.

Type of studies. Selected studies of PBSC transplanta-tion for hematological malignancies are reported inTables 2 and 3. Several of them have been analyzed inthe section on hematological recovery. Table 2 focuseson the main demographic characteristics, while Table 3gives details of the transplant procedure and results ofacute and chronic GVHD where applicable. Six studiesare from a single institution,39,40,55,75-77 while three arefrom several centers;50,52,54 one study from the EBMTGroup,50 multicentric in nature, also reports severalpatients included in four of the other studies – a typi-cal example of double reporting – so that its resultsreinforce what has already been observed. The figuresfrom this last study were not calculated in any furtherstatistical analysis in order to avoid the error of count-ing some of the patients twice. However, they are use-ful for comparisons and have been left in the tables. Thetotal number of patients is 212. None of the studies isprospective or randomized, but four52,55,75,77 compare theresults of PBSC with those of marrow, although usingdifferent methods. We shall have to wait some time

before seeing the results of the two prospective ran-domized studies comparing marrow and PBSC trans-plantation which are now in progress in Europe and theUS; for the moment, the reports analyzed here representthe best we have. The 8 studies took place recently,between late 1993 and 1995, and mostly dealt withadults (median age 38 yrs, with a range from 1-57), butsome included pediatric patients. Transplants were fromfully HLA-identical siblings in 96% of the cases, but aminority received cells from family donors mismatchedfor one HLA antigen; a minority of patients (5 to 10%)also received a second allo transplant, usually from theoriginal sibling who had donated the marrow. Patientsshowed a typical spectrum of hematological malignan-cies for which transplant is indicated. The majority(median 83%) were in advanced phases of their dis-eases, although definitions are quite variable with theterm high-risk being used as a synonym for advancedphase, but 17% had early phase or low-risk disease atthe time of transplant. These proportions differed wide-ly within studies, some including 100% advanced dis-eases and others only 60%, with many more early phasepatients. Pre-transplant regimens were obviously dif-ferent, but despite their apparent disparities they can begrouped into those based on busulphan (54%) or TBI(31%). Only two studies differ considerably from therest of the series: the Genoa group40 purposely employeda low intensity regimen based on thiotepa andcyclophosphamide to reduce toxicity in a rather oldpatient population. The MD Anderson Hospital,39,55 onthe other hand, used a very intensive regimen combin-ing busulphan, thiotepa and cyclophosphamide in a pop-

Table 2. Main characteristics of the studies reviewed.

Ref Type Of study Period of study Median N° pts Median 2nd transplants Hla family Phase of the diseaseFollow-up days (range) age yrs (range) mismatches

74 SC 12/93-11/95 nr 37 38 (20-52) NO NO Advanced 100%

52 MC 5/93-6/95 nr 26 40 (1-54) 3 (11%) 6 (23%) High risk 23 (88%)Standard risk 3 (12%)

40 SC nr 136 (6-228) 31 44 (19-55) NO 3 (10%) Advanced 28 (90%)Early 3 (10%)

55 SC nr 270 (180-600) 25 43 (17-57) 1 (4%) NO Relapse 21 (84%)Remission 4 (16%)

76 MC 3/94-7/96 nr 24* 37 (16-57) 1 (4%) 1 (4%) Early 10 (42%)Advanced 14 (58%)

54 SC 1/94-4/95 nr 33 36 (12-53) 8 (24%) NO Early 12 (36%)Advanced 21 (64%)

39 SC 32 months nr 19 Not detailed 1 (5%) NO Early 18%Advanced 82%

77 SC 3/94-4/95 111 (15-402) 17 33 (16-52) NO NO Early 6 (35%)Advanced 11 (65%)

50 MC 1994 nr 51° 39 (2-54) NO NO Early 15 (25%)Advanced 44 (75%)

LEGEND: SC = single center; MC = multicenter; n.r. = not reported; *includes 1 pt with SAA; °includes patients from studies #40, 76 and 54.

75


ulation of similar age. Another difference is represent-ed by the processing of the collected PBSC: in 130 cas-es (61%) they were infused fresh and in 82 cases (39%)they were cryopreserved instead until infusion. Finally,GVHD prophylaxis was not uniform: it was based on acombination of CsA and short-course methotrexate in130 (62%) of the cases, and on CsA plus prednisone in42 cases (20%); only one study38 reports 19 patients(9%) who received a combination of tacrolimus andprednisone. CsA alone was used in three patients (1.4%).

Acute GVHD. The incidence of aGVHD, grade II to IV,was about 40% on average. The Genoa study40 reporteda 55% incidence, but it also included the oldest patientsin the series; the lowest incidence, 22%, was reportedin the series from the MD Anderson Hospital wheretacrolimus was used for GVHD prophylaxis.39 The inci-dence of severe aGHVD, i.e. grade III and IV, was on aver-age 16% (range 11-24%). An interesting point is thefact that while most studies showed a direct correlationbetween the overall incidence of GVHD and severeGVHD – the latter being about half of the former – oth-ers did not. Two studies from Italy40,76 which reportedan overall incidence of over 50% also showed a lowincidence of severe GVHD, which means that grade IIaccounted for the majority of the cases. No correlationcould be made from the existing data between the vari-ables known to influence aGVHD78 and the results either

within the single studies as discussed by their authorsor by combining data as in this review. Of interest, onthe specific issue of PBSC, no correlation was found withthe number of T-lymphocytes infused or with the use offresh or cryopreserved cells. However, it should be not-ed that the highest incidences of severe aGVHD (22-24%) were reported when prophylaxis was based onCsA/prednisone,55 which is perhaps less effective thanCsA/MTX or in the Spanish study which reported datafrom multiple institutions54 with a good proportion ofpatients receiving CsA/prednisone for GVHD prophylax-is. Nevertheless, a high incidence was also reported in astudy from Brazil where CsA/MTX was used in all cas-es.77 Mortality from aGVHD is not reported in all stud-ies, as shown in Table 3; those giving the causes of deathoften do not mention, when infection was the maincause, whether GHVD was associated. However, consid-ering the causes of death of 47 events analyzable indetail, GVHD was the main cause in 12 (25%). Inciden-tally, this figure is higher than the overall incidence ofgrade III-IV GVHD, but data on death are reported forfewer patients than data on GVHD.

Chronic GVHD. The number of patients analyzable forcGHVD is smaller than for aGVHD; survival > 90, 100 or150 days is the requisite for evaluation. In addition,some studies give many details on cGVHD while othersdo not address the issue55 or mention it very briefly.39,52,54


Table 3. Transplant modalities, aGVHD and cGVHD.

Ref Fresh or Conditioning G-CSF GVHD Acute GVHD Chronic GVHDcryo- regimen* post TX prophylaxis n. grade grade GVHD related/ n. limited extensive

preserved cells N° pts N° pts evaluable II-IV III-IV total deaths evaluable

74 F TBI 32 no CsA/MTX 19 35 13 (37%) 5 (14%) 1/15 17 3 (18%) 4 (24%)Bus 5 CsA/PDN° 18

52 C TBI 18 no CsA/MTX 26 nr 37% nr nr nr 53% overall

40 F Thio-CTX 31@ no CsA/MTX 31 31 17 (55%) 4 (13%) 4/12 28 15 (53%) 7 (25%)

55 C Bus 25 yes CsA/PDN 25 25 11 (42%) 6 (22%) 3/7 nr nr nr

76 F^ Bus 22 no Csa/MTX 23 22 10 (45%) 2 (9%) 0/7 16 1 (5%) 9 (50%)Other 3 CsA 1

54 F 21 TBI 17 yes CsA 2C 1 Bus 16 (11 pts) CsA/MTX 22 32 11 (34%) 7 (22%) nr 11 4 (36% overall)

CsA/PDN 9

39 C Bus 19 yes FK-506/PDN 19 nr 22% 11% nr nr nr nr

77 F Bus 17 no CsA/MTX 17 10 3 (33%) 4 (24%) 4/4 10 3 (33%) 0

50 F 49 TBI 22 CsA/PDN 7Bus 22 yes CsA 6 57 30 (50%) 14 (23%) 7/29 49 17 (35%) 13 (26%)

C 10 Thiotepa 11 (14 pts) Other 9Other 4

*Conditioning regimen mainly based on; @Regimen not including TBI or busulphan; CTX = cyclophosphamide; ^cells infused over 2 days: apheresis of day 1 stored at 4°C until infu-sion; °PDN=prednisone; n.r.=not reported

76


We calculate that slightly more than 110 patients areevaluable. The overall incidence ranged from 36% to78%; considering the four studies which give detailedinformation, the extensive form of the disease occurredwith nearly the same frequency or less than the limitedform in three studies,40,56,77 while one series reported astriking incidence of the extensive form, with the limit-ed one being only minimally represented.76 In that studycGVHD developed de novo in 5 out of 10 patients, atvariance with the low incidence of aGVHD observed ear-lier. Another interesting observation is contained in astudy from Seattle:56 of 10 patients at risk, who hadreceived CsA/prednisone for prophylaxis, 6 developedcGVHD, while only 1 of 7 given CsA/MTX did so.

Comparisons between marrow and PBSC transplanta-tion. Four studies52,55,56,77 compared the incidence of GVHDafter PBSC or marrow transplantation. These studies werecarried out using matched-pair analysis with a historicalcontrol group of marrow recipients who were matchedfor diagnosis, disease and disease phase at transplant,age, GVHD prophylaxis56 or age and disease status.52 Onestudy does not give the characteristics of the marrowrecipients.77 The Seattle study found a lower incidence ofgrade II-IV aGVHD for PBSC recipients – 37% vs 56% –severe GVHD (grade III-IV) was even more impressivelylower in PBSC recipients, 14% vs 33% of marrow trans-plants. However, due to the small number of patientsthese data are not statistically significant. The overallincidence of cGVHD was similar in the two groups, witha tendency toward more severity in the PBSC than in themarrow group (42% vs 26% for any grade of clinicalcGVHD), but again this was not statistically significant.The striking effect of MTX in the GVHD prophylaxis reg-imen observed in this study has already been mentioned.The multicenter Canadian study52 reported a higher inci-dence of both aGVHD and cGVHD for the PBSC groupthan for the marrow recipients: 37% vs 21% grades II-IV aGVHD and 53% vs 48% for cGVHD, respectively. Adifferent kind of comparison can be made in a study fromthe MD Anderson Hospital,55 where patients withadvanced hematological malignancies received, over a 3-year period, the same conditioning protocol but differentforms of GVHD prophylaxis and different sources of allo-geneic stem cells. The PBSC patients could be comparedto the marrow group who received CsA/prednisone asGVHD prophylaxis: severe aGVHD was slightly less in thePBSC group, 22% vs 33%, but this difference was not sta-tistically significant. No data on cGHVD were provided.The Brazilian study77 reported more aGVHD in the PBSCthan in the marrow group, grade III-IV 4/17 vs 3/21, butagain this was not statistically significant. No compara-tive data on cGVHD were given.

Comments. The fear of an unacceptably high rate ofsevere and perhaps uncontrollable acute GVHD after alloPBSC can now be allayed with a certain degree of con-fidence on the basis of the data analyzed here. In a pop-ulation of adults with mainly advanced hematologicalmalignancies who sometimes received second trans-plants or not fully HLA-identical grafts and, most impor-

tant, were often not given the best GVHD prophylaxisavailable, acute GVHD was no more than what is expect-ed with marrow transplants. Similar conclusions hadbeen reached in an earlier review on the subject,56

although on a much smaller number of patients. It ismore difficult to say whether GVHD is slightly moresevere than after conventional marrow grafting, consid-ering the wide variation in its occurrence, due to themultitude of factors which influence it. A broad com-parison of the published data indicate that GVHDobserved after PBSC is higher than in the best marrowseries79 but not worse than what is described in the largereports from registries.80 Four studies have attempted acomparison with retrospective marrow transplants andnone found significantly increased aGVHD incidence orseverity. It is interesting to speculate why after infusionof 1 log more T-lymphocytes as compared to marrowaGVHD is not increased. One explanation is that anynumber of T-cells, once the 105/kg threshold has beensurpassed,81 is already high enough to cause GVHD andeven a 10-fold increase, with respect to the marrow,does not make any difference. Perhaps an extraordinar-ily large number of T-cells infused, that is usually notreached after G-CSF mobilization, for example 3 or 4logs, could be the next threshold above which moresevere GVHD would regularly occur. Data from a studyin Seattle, where 1500×106/kg donor buffy-coatmononuclear cells were intentionally infused soon afterBMT in patients with advanced disease to enhance agraft- versus-leukemia effect did show that acute, severeGVHD was indeed increased; in that study, unfortunate-ly, GVHD prevention was based on MTX only so compar-isons with today’s practices are not possible.82 Otherexplanations for why aGVHD does not increase relate tothe possible biological modifications of lymphokine pro-duction induced by G-CSF. For example, in mice G-CSFhas been demonstrated to induce a polarization of T-lymphocytes towards the production of type-2 cytokines(namely IL-4 and IL-10), which display an anti-inflam-matory effect. Such polarization was shown to be long-lasting and was associated with a significant reductionin the severity of GVHD after transplantation of thesecells into allogeneic mice recipients.83 These results donot seem to be attributable to a direct effect of G-CSFon T-cells, since this subset of lymphocytes rarelyexpresses G-CSF receptors. Instead, these findings couldbe explained by the anti-inflammatory effects of G-CSF;in fact, administration of G-CSF decreases tumor necro-sis factor (TNF) secretion.84 Moreover, in normal subjectsG-CSF is able to increase the production of two impor-tant cytokine antagonists such as soluble TNF-receptorand IL-1 receptor antagonists.85

Finally, leukapheresis products from G-CSF-mobilizeddonors contain a large number of monocytes. These cellshave been demonstrated to significantly reduce thealloantigen specific proliferative response of T-lympho-cytes.67 Also, monocytes from subjects treated with G-CSFor GM-CSF can induce the apoptosis of T-lymphocytes viathe interaction of the FAS molecule with its ligand.86 The

W. Arcese et al.

77


use of G-CSF may inhibit the function of monocytes asantigen presenting cells and this, in turn, may explainthe ability of this cytokine to polarize T-cells towards ananti-inflammatory cytokine profile. These findings couldalso contribute to explaining the unexpectedly low inci-dence and reduced severity of GVHD after transplantationof PBSC. However, more studies on the characterizationof G-CSF mobilized lymphocytes are needed.

With regard to cGHVD, it appears from this analysisthat there is a trend toward a slightly increased incidenceafter PBSC, although not all individual studies had thesame results. The clinical presentation of cGVHD wasreported as peculiar in two studies, with many de novocases31,75 but also with a high response to treatment.49 Itshould be noted that early data from the M.D. Andersonon PBSC transplants also reported an increase in cGVHD,with more liver and gastrointestinal manifestations com-pared to the marrow.87 At the time of this writing, sever-al patients were still on immunosuppressive treatment sothe full magnitude of cGVHD will be appreciated only inthe future after withdrawal of CsA, which is a criticaltime for the development of the syndrome. Furthermore,in a study on aplastic anemia88 the infusion of donorbuffy coat cells was associated with a significant increaseof cGHVD. However, this increase of cGVHD was notobserved in malignancies in the study.82 Clearly a longerfollow-up of a much larger number of patients is need-ed to answer the question of cGVHD.

Transplantation of enriched allogeneicCD34+ cells

As reported above, allogeneic PBSC transplantationresults in the infusion of approximately 1 log more T-cells than conventional BM transplantation. Thus, in orderto reduce the potential risk of severe aGVHD, severalinvestigators have attempted to remove T-lymphocytesfrom allogeneic grafts. The only technique that has beenutilized so far for T-cell depletion has been the positiveselection of hematopoietic CD34+ cells.

The CD34 antigen is present on the earliest identifiableprogenitor cells and committed myeloid precursors,whereas it is not expressed on mature myeloid and B- andT-lymphoid cells.89 However, CD34+ cells co-expressingboth T-lymphoid (CD2, CD3, CD7) and B-lymphoid mark-ers (CD19) are likely to be the early precursors of the T-and B-lymphoid lineages.90 Preclinical studies have alsoshown the capacity of positive selection of CD34+ cells toeliminate 3 to 4 logs of T-cells, coupled with a substan-tial recovery of hematopoietic progenitors.91,92 Morerecently, transplantation of autologous CD34+ cells hasbeen proven to reconstitute normal hematopoiesis in can-cer patients treated with myeloablative regimens.93-96

Based on these premises, Link et al.97 transplanted 5patients with unmodified marrow and CD34+ selectedPBSC and 5 patients with enriched marrow and PB CD34+

cells. They concluded that hematopoietic recovery wasaccelerated with respect to marrow allografts withoutan apparent increase in aGVHD following conventionalCsA and MTX prophylaxis. In a subsequent study,98 the

same authors transplanted 10 individuals with positive-ly selected circulating CD34+ cells alone. The patientswere grouped according two different regimens ofaGVHD prophylaxis: CsA alone or CsA and MTX. The medi-an grades of aGVHD were 3 in group I (CsA) and 1 ingroup II (CsA plus MTX). Two patients in group I died fromaGVHD and 2 leukemic relapses occurred in group II.Complete and stable donor hematopoiesis was shown inall patients with a median follow-up of 370 days (range45-481). It was concluded that despite a 3-log reductionof T-cells by CD34+ cell enrichment, CsA alone was notsufficient to avoid severe aGVHD.

More recently, Bensinger et al.99 transplanted 16patients with advanced hematologic malignancies withHLA-identical highly enriched PB CD34+ cells. Prophy-laxis against aGVHD was CsA alone for 5 patients and CsAplus MTX for 11. A median of 8.96×106 CD34+ cells/kg ofpatient body weight were infused with a median purityof 62%. Positive selection of stem cells resulted in amedian 2.8-log reduction of T-cells. Despite the promptand sustained engraftment, 8 out of 16 patients diedbetween 3 and 97 days post-transplant of transplant-related causes and 1 of progressive disease. Grade 2-4aGVHD occurred in 86% of patients and 6 out of 8 evalu-able patients developed clinical chronic GVHD.

More promising results have been reported by Urbano-Ispizua et al. (1997), who recently transplanted 20 acuteand chronic leukemia patients with allogeneic CD34+

cells. The median number of CD34+ cells and CD3+ cellsinfused was 2.9×106/kg and 0.42×106/kg, respectively.The patients were conditioned with fractionated TBI (totaldose 13 Gy in 4 fractions) and cyclophosphamide 120mg/kg. Additional GVHD prophylaxis included CsA andmethylprednisolone. No patients developed grade II- IVaGVHD. The overall procedure was associated with lowmorbidity and no transplant-related deaths occurredwithin the first 100 days. Although the median follow-up (7.5 months) is rather short for a full evaluation ofcGVHD incidence and disease relapse, the absence ofextensive cGVHD and the low rate of disease recurrence(only 3 out of 20 patients relapsed) encourage furtherstudies in this direction. In comparison with previousstudies,99 it should be noted that the median age of thepatient population was 40 years and only 35% of theindividuals were older than 45 years. Moreover, themajority of the leukemic patients were transplanted inthe early phase of their disease. Both these parametersare generally associated with lower transplant-relatedmortality and a lower incidence of severe GVHD.

Although further studies involving larger numbers ofpatients are currently in progress, these results, takentogether, demonstrate that infusion of CD34+ selectedPBSC results in rapid and stable engraftment. However,transplantation of purified stem cells may induce a high-er rate of acute and chronic GVHD than expected, thusrequiring full GVHD prophylaxis. Therefore this approachfor T-cell depletion should be carefully evaluated in thesetting of HLA-identical PBSC transplantation andweighed against the potential increased risk of disease


78


relapse, and perhaps delayed immunological reconstitu-tion, and the increased cost of the procedure.

Allogeneic PBSC from haploidenticalfamilial donors: the mega-stem-celldose concept

Allogeneic BMT has been largely confined to patientswho are HLA-identical to their donors. At present, onlyabout 30-35% of patients who might benefit from allo-geneic BMT have an HLA-identical sibling. The estab-lishment of large registries of HLA-typed individualsduring recent years has led to a substantial increase intransplants from unrelated donors.100-102 Although 40 to50% of patients are successful in locating HLA-A, B,DR-matched unrelated donors, many patients still fail tofind an appropriate donor.103,104 In contrast, nearly allpatients have an HLA-haploidentical relative (parent,child, sibling,) who could serve as a donor.

The feasibility and safety of transplants from partial-ly matched family members have been investigated andthe results of these studies have demonstrated that HLAmatching is a critical and limiting factor in marrowtransplantation.105-107 In published works on mismatchedtransplants, there has been no large study involvingpatients mismatched with their donors by one full hap-lotype. These experiences have been limited because theproblems of transplant increase with the number ofantigenic disparities between donor and host.106 In 2- or3-antigen mismatched transplants, studies by the Seat-tle program105 and the International Bone Marrow Trans-plant Registry108 reported graft failure in 20 to 30% ofcases. The reported incidence of acute GvHD (grade II orgreater) varied from 34% to 100% overall, but in 2- and3-antigen mismatched patients the incidence was atleast 80%.105,106 Severe GvHD was a greater problemthan graft rejection, preventing more widespread use ofmismatched related transplants during the latter 1970sand 1980s.

By contrast, extensive experience in severe combinedimmunodeficiency (SCID) patients has shown that GVHDis largely preventable, even in 3-antigen mismatchedtransplants, when a 3-log T-cell depletion of the donorbone marrow is achieved.109,110 In 1981 Reisner et al.reported the first case of leukemia treated with a T-celldepleted marrow transplant from a haploidentical, 3-loci incompatible, parental donor.111 There was fullengraftment and no GVHD. Subsequently, clinical trialsin mismatched-sibling BMT for patients with leukemiawere begun using the lectin or other T-cell depletingmethods which included monoclonal antibodies withcomplement or conjugated to toxins, and counterflowcentrifugal elutriation (reviewed in ref. #112). It wasdetermined that the threshold dose below which GVHDwas not seen in matched patients was 2.0×105 Tcells/kg.113 However, early enthusiasm for all methodswas soon tempered by an increased incidence (> 50%)of graft rejection.114

In exploring the problem of failure in mismatched

grafts, the inadequacy of immunosuppression was doc-umented by the observation of residual host lympho-cytes in patients who failed to engraft after being con-ditioned with conventional preparative regimens andgiven T-cell depleted mismatched transplants.114 Workin esperimental models has shown that incompatible T-cell depleted transplants can be successfully performedby manipulating the conditioning regimen and/or thegraft composition.115 The immunologic response of theremaining host immune system against the graft canbe overcome by increasing the total dose of TBI116 or byadding selective anti-T measures with minimal toxicity,such as splenic irradiation117 or in vivo treatment withanti-T monoclonal antibodies.118 Engraftment is alsoimproved by increasing the myeloablative effect of theconditioning regimen through the use of dimethyl-myleran, busulfan or thiotepa, given with TBI.119,120

Different cytoreductive agents or radiation regimenswere therefore added to the basic conditioning proto-cols used for conventional BMT. Although a marked ben-eficial effect was found in recipients of T-cell depletedHLA-identical bone marrow upon adding ATG andthiotepa to TBI and cyclophosphamide,121 none of theseagents were found to be useful in recipients of T-celldepleted haploidentical 3-antigen incompatible trans-plants. Others, using the monoclonal antibody Campath-1G instead of ATG, have observed similarly disappoint-ing rejection rates.122

Concerning the composition of the graft, Lapidot et al.showed that megadoses of T-cell depleted incompatiblebone marrow inoculum could obtain full donor-typeengraftment in mice treated with sublethal irradiation,or presensitized with donor lymphocytes or partiallyreconstituted before the transplant by adding back acontrolled number of host-type mature thymocytes.123

The means of overcoming graft failure elucidated inthe experimental model can be applied in the clinical set-ting by combining approaches that increase both theconditioning of the host and the size of the stem cellinoculum. The major advance that finally made full hap-lotype-mismatched transplantation possible in leukemiapatients was the availability of rhG-CSF124 and the expe-rience in autologous transplants in which G-CSF was usedto mobilize high numbers of stem cells into the blood ofpatients without significant side effects.125 Their employ-ment made it feasible to increase the number of donorstem cells to a level which, in animal models, made trans-plantation across the histocompatibility barrier possi-ble.117 On the basis of these concepts, the BMT team atthe University of Perugia first introduced the megadosecell transplant in full haplotype-mismatched leukemiapatients.126 After a conditioning regimen which included8 Gy TBI in a single fraction at a fast dose rate (16 cGy/m),thiotepa (10 mg/kg), rabbit ATG (20 mg/kg in 4 days) andcyclophosphamide (100 mg/kg in 2 days), advancedleukemia patients, mostly adults, were given the combi-nation of marrow and G-CSF-mobilized blood stem cells.Donors compatible with the patients for only one haplo-type and 3-antigen disparate on the other haplotype

W. Arcese et al.

79


underwent bone marrow harvest followed within a fewdays by treatment with G-CSF (12 µg/kg/d ×7 days). Fourleukaphereses of progenitor cells were performed start-ing on the fourth day. The marrow as well as the leuka-pheresis product were each depleted of T-cells using soy-bean lectin agglutination and E-rosetting.127 Both theCD34+ cells and CFU-GM were increased 7- to 10-foldover bone marrow alone, and the average number of CD3+

cells infused was 2.2×105/kg recipient body weight. Fol-lowing conditioning and stem cell infusion, patientsreceived no additional GvHD prophylaxis. The results ofthe first 17 patients were reported in 1994126,128 and sub-sequently 27 additional leukemia patients, most of themin chemoresistant relapse at the time of transplant, weretreated. For the first time a very rapid hematopoieticengraftment was observed in more than 80% of patientsand, without any post-transplant prophylaxis, acuteGVHD occurred in only 27%, and there was no significantchronic GvHD. As for survival, 7 patients are currentlyalive and disease free at a median follow-up of more than3 years. The major complications observed in this pilotstudy were interstitial pneumonitis, which occurred in43%, and infections in the setting of GvHD. Both wereresponsible for the 60% transplant-related mortality.

This pilot experience showed that the megadose cellstrategy, together with a highly immunosuppressive andmyeloablative conditioning, resulted in a high incidenceof durable engraftment with significantly reduced GvHDcomparable to historical experience with unmanipulat-ed transplants. It also confirmed that in humans, as inmice, the stem cell dose plays a crucial role in over-coming HLA-histocompatibility barriers.129 This conceptis also supported by the recent work by Rachamin etal.130 demonstrating that purified CD34+ cells have avery powerful veto activity. They are able to specifical-ly reduce, in a mixed lymphocyte culture, the frequen-cy of CTL precursors against the stimulatory cells of thesame subject and thereby help to overcome allogeneicrejection and enhance their own engraftment.

The approach to haplotype-mismatched transplantshas evolved since Aversa et al.126,128 originally proposedthe megadose cell concept. With their initial protocolGvHD was decreased but not eliminated, and it con-tributed to transplant-related mortality, which was sig-nificantly greater than in matched patients receivingsimilar conditioning (Aversa et al., unpublished data).These remaining problems were addressed in a subse-quent trial, where it was possible to completely abrogateGVHD by improving the T-cell depletion method.131 Byprocessing the peripheral blood progenitor cells with aninitial debulking of both mononuclear and T-cells withone-round E-rosetting followed by positive selection ofCD34+ cells with the Ceprate stem cell concentrator(CellPro Inc. Bothell, WA, USA), it was possible to infusea median of 3×104 CD3+ cells/kg and 13×106 CD34+

cells/kg in 24 high-risk leukemia patients. Condition-ing-related toxicity was also reduced by modifying pre-transplant chemotherapy. As a substitute for cyclo-phosphamide, which was considered a possible factor in

the early mortality in the first pilot study, fludarabinewas tested. In fact, it had been shown to have power-ful immunosuppressive effect in patients treated forlymphoproliferative disorders, even at doses which werenot associated with significant extra-hematologic tox-icity.132 Furthermore, in a mouse model TBI+fludarabine(40 mg/m2/d × 5) was shown to provide an immuno-suppressive effect comparable to TBI+cyclophos-phamide.133

At present, a regimen including TBI in a single frac-tion, thiotepa, fludarabine and ATG followed by the infu-sion of T-depleted bone marrow plus T-depleted CD34-selected blood cells is being evaluated for toxicity andefficacy. The preliminary results of this study wererecently presented at the American Society of Hema-tology meeting in Orlando.134,135 As hoped, with thedecrease in the number of T-cells infused and the mod-ifications in conditioning, the problem of GvHD waslargely prevented (only 2 patients developed grade IIacute GvHD and one progressed to chronic GvHD); theengraftment rate was 95% and there was a decrease intransplant-related mortality to 29% compared to theprevious 60%.

A more recent update on 48 patients was presentedin Mannheim.136 The abstract reports that 22/28 patientswere in chemoresistant relapse at the time of trans-plant; age ranged from 4 to 53 years (median 27). Fortyof 48 patients engrafted, grade II-IV acute GvHDoccurred in only two patients and no one developedchronic GvHD. Twenty patients were alive and diseasefree at a median follow-up of 5 months (range 1-16).There were 11 relapses and 17 nonhematologic deaths.Transplant-related mortality was 35%.

An unsolved problem remains the slow immunologicrecovery of engrafted patients that is responsible forinfections. Counting of peripheral blood lymphocyteswhich exhibited a phenotype of NK cells (CD56+), helperT cells (CD3+/CD4+), cytotoxic T-cells (CD3+/CD4+), cyto-toxic T-cells (CD3+/CD8+) and B-cells (IgM+) revealed ear-ly recovery (within 2-4 weeks) of NK cells and extreme-ly delayed recovery of T cells. In particular, CD4+ cellsreached near normal values after 10 to 12 months.137 Inaddition, the frequencies of T-cells responding to poly-clonal activators in a sensitive limiting dilution assaywere approximately 1 in 100 within the first post-graft-ing month and 1 in 10 at 10 months post-transplant(control responder cell frequencies are in the range of 1in 2).137 The low number of T-cells, combined with theirfunctional peculiarities (i.e. failure to respond to TcRstimulation) are certainly implicated in the high fre-quency of infectious complications and are stronglyindicative of a markedly distorted T-cell maturationalprocess.

Interestingly, looking at post-transplant immunereconstitution, Albi et al. observed a large donor-typeTcR-ab+ CD8+ cell population that co-express NK-likereceptors for specific MHC class I alleles.137 NK cellsexpressed multiple, clonotypically distributed membranereceptors with different specificities for families of MHC


80


class I alleles (termed killer cell inhibitory receptors, KIR).The interaction between these receptors and the appro-priate alleles produces a signal which inhibits killing ofthe target cells. Analysis of more than 900 clonesrevealed that 40% to 80% of these KIR+ T-cells exhibitNK-like functions, i.e. they were able to lyse class I-neg-ative targets and were functionally blocked by theexpression of specific class I alleles on target cells. Fur-thermore, these cells do not lyse autologous hemo-poietic cells, but are able to lyse fresh leukemic cells.137

This might suggest that they could provide a graft-ver-sus-leukemia effect without causing GVHD.

In a period of twenty years transplants across the his-tocompatibility barrier have advanced from being exper-imentally to clinically possible. The principles outlinedat the beginning – adequate cell dose, adequateimmunosuppression and myeloablation, avoidance ofGvHD – have been successfully combined. Two othergroups have recently reported on successful engraftmentin haplotype mismatched transplants by combining bonemarrow and G-CSF-mobilized blood stem cells afterCD34-positive selection for patients with advancedleukemia.138,139 Refinements of this protocol should makehaplotype mismatched transplants an attractive thera-peutic option for patients with high-risk leukemia with-out a matched related or unrelated donor.

Furthermore, there are enormous potential applica-tions of the concept of the stem cell dose for the futuretreatment of non-neoplatic diseases like aplastic ane-mia, Fanconi’s anemia, SCID, thalassemia, and for induc-tion of tolerance in organ transplantation. This approachshould be applicable not only in mismatched transplantsbut also for overcoming problems which remain in thematched transplant setting, such as rejection in aplas-tic anemia, regimen-related toxicity in Fanconi’s anemiaand thalassemia.

Transplantation of allogeneic PBSCfrom unrelated donors

Two retrospective studies have recently suggestedthat the number of hematopoietic cells present in BMharvest correlates with the clinical outcome in the set-ting of stem cell transplantation from both HLA-identi-cal siblings and from HLA-matched unrelated donors.9,140

In the latter case, the number of cells infused has provento be the most potent prognostic factor for survival.Therefore, given the much higher number of progenitorcells collected in primed PB as compared to conven-tional BM harvest, the use of PBSC appears to be apromising alternative for improving the results of trans-plantation from unrelated donors. In this regard, Ring-den et al.34 and Stockschlader et al.141 recently reportedtheir preliminary experience with transplantation ofallogeneic PBSC from full-matched or 1-antigen mis-matched unrelated donors. In particular, Ringden et al.transplanted 6 patients with high-risk hematologic dis-ease. Four of them received allogeneic PBSC as primarytreatment while 2 others were treated after a BM graft

failure. Five PBSC collections were infused without anymanipulation; in 1 case Campath-1 monoclonal anti-body was used for T-cell depletion. In the German study,1 AML patient in 2nd CR received purified CD34+ cellsfrom an HLA-matched unrelated donor. The total num-ber of patients transplanted is too small and the follow-up too short to draw any conclusion; however, thesepreliminary data showing a rapid rate of engraftmentare encouraging, whereas the role of T-cell depletionremains to be clarified.

Transplantation of umbilical cord bloodprogenitor cells

The existence of hematopoietic progenitors circulat-ing between the fetus and the placenta during gestationwas first described in this Journal more than 20 yearsago,142 but their clinical application began only when itbecame evident that the progenitor cell content of CBwas sufficient for bone marrow repopulation in pediatricpatients given myeloablative chemo-radiotherapy.143 In1988, a patient affected with Fanconi’s anemia was firsttransplanted with CB progenitor cells from his HLA-matched healthy sibling.4 Subsequently, successful CBtransplants (CBT) were sporadically reported in patientsaffected by both malignant and nonmalignant disor-ders.5,144-147 The establishment of large CB banks inEurope and USA, and improvement of the methods ofcell collection, manipulation and freezing have permit-ted a rapidly increasing use of CB progenitor cells, whichare now extensively employed for allogeneic transplan-tation.148-150

The biological and functional characteristics of CBhematopoietic stem cells have been already reviewedby the Working Group.12

Clinical results after cord bloodtransplantation

As mentioned above, the use of human umbilical CBhematopoietic progenitors represents an alternativemodality of transplantation. Advantages of CBT includeease of hematopoietic stem cell collection, absence ofdonor risks, low risk of viral contamination (cytomegalo-virus, Epstein-Barr virus, etc.) and, for transplantationamong unrelated individuals, prompt availability ofhematopoietic stem cells. Over the past decade, pla-cental blood has been used to transplant hundreds ofpatients (mainly children) and information on the rateand kinetics of engraftment and on the risk of severeacute or chronic GVHD is now available for CBT recipi-ents from both related and unrelated donors.

In the two largest cohorts of patients transplantedfrom an HLA-identical sibling reported to date,7,148 theprobability of engraftment of donor hematopoiesis was79% and 85%, respectively, even though it must beunderlined that in the cohort analyzed by Wagner andcolleagues rejections were mainly observed in patientsaffected by bone marrow failure syndromes or hemo-globinopathies, which are diseases with a high risk ofgraft failure. In the cohort reported by Wagner et al., the

W. Arcese et al.

81


median time to achieve granulocyte (PMN >0.5×109/L)and platelet (PLT >50×109/L) recovery was 22 and 49days, respectively; these values were greater than thoseobserved with BMT. Comparable time for PMN and PLTrecovery were observed in the European experience.7 Inparticular, in this latter report, patients receiving a high-er number of nucleated cells (i.e. more than 37×106/kg)experienced faster engraftment than those given a low-er number of cord blood progenitors, suggesting thatthe number of cells infused is the main factor influenc-ing the rate of hematologic recovery. More prolongedperiods of profound leukopenia and thrombocytopeniahave also been described in children receiving CBT fromunrelated donors. In fact, in the first two series ofpatients transplanted from an unrelated donor,149,150

PMN recovery occurred in a median of 22 and 24 days,respectively, whereas the median time for PLT recoverywas 82 and 67 days, respectively. The importance of thenumber of cells infused on the kinetics of PMN and PLTengraftment in the Eurocord Transplant Group experi-ence was also observed in the group of patients givenan unrelated CBT. Moreover, unlike BMT, where the useof hematopoietic growth factors has been demonstrat-ed to hasten myeloid recovery significantly,151,152 admin-istration of these cytokines has produced conflictingresults in CBT recipients. In fact, in the cohort of patientsreceiving CBT from HLA-identical or disparate familydonors studied by Wagner et al., the use of G-CSF orGM-CSF did not influence the kinetics of PMN recon-stitution.142 In contrast, in a group of children trans-planted using unrelated CB units reported by the sameauthors, patients receiving hematopoietic growth fac-tors experienced faster myeloid recovery than those whowere not given the cytokines.150

The delayed rate of neutrophil engraftment and the

conflicting data mentioned above could be explained bythe infusion of fewer progenitor cells with CBT withrespect to BMT, as suggested by the European experi-ence, or, alternatively, by the particular characteristicsof the proliferative, self-renewing and differentiatingcapacity of CB cells. A practical consequence of theabove observation is that specific attention should bepaid to the risk of infectious complications in childrenreceiving CBT.

During the first few months after transplant CBTrecipients show a steady, impressive increase in HbFwhose values are significantly higher than thoseobserved in patients receiving BMT. Moreover, the sub-sequent decline is usually less pronounced than thatobserved in normal children in the first year of life (Fig-ure 1).153,154 This preferential production of g chains inerythroid progenitors seems to reproduce the normalontogeny of erythropoiesis, even though the persistenceof HbF levels higher than those observed in the first yearof age suggests a more delayed switch from fetal toadult hemoglobin synthesis.

The dose of CB progenitor cells necessary to ensureearly and sustained hematopoietic engraftment andfavorable clinical outcome has still not been preciselydefined. Wagner et al.148 claimed that the lowest doseof CB nucleated cells reported to be capable of yieldingcomplete and sustained engraftment is 1×107/kg ofrecipient body weight. However, as previously men-tioned, the Eurocord Transplant Group documented thata dose of nucleated cells available before thawing offewer than 3.7×107/kg recipient body weight was high-ly predictive of both graft failure and poor survival afterCBT.7 The importance of this value also emerges fromKurtzberg et al’s experience.149 Ten out of 13 patientsundergoing CBT from an unrelated donor and having

Figure 1. HbF levels observed in the 5 CBT recipients inthe first year after transplant. Normal percentiles of HbFvalues (dotted line, mean±2SD) observed in the first yearof age (Galanello et al., 1981) are also reported (F.Locatelli, personal data).


82


received fewer than 3.73107/kg nucleated cells failed tobenefit from the procedure. Although the importanceof this cellular dose appears evident from these tworeports, it should be noted that rarely is such a numberof cells available in the case of adult patients. In fact,since the average leukocyte count in placental blood isabout 10×106/mL and the average volume of donatedblood is about 80 mL, the average number of nucleatedcells before thawing in one cord blood unit may reach800×106. About 30% of nucleated cells are lost duringthe thawing and washing procedure, and even thoughthe loss mostly involves mature cells which have no rolein transplantation, it is reasonable to expect fewer than3.7×107/kg viable cells for patients with body weightgreater than 30-40 kg.

The reduced immune reactivity of cord blood cellsfound a clinical counterpart in 38 children reported byWagner et al.148 who received CBT from an HLA-identi-cal or 1-antigen mismatched sibling. In these patients,the incidence of grade II-IV acute GVHD and limitedchronic GVHD was 3% and 6%, respectively, with nopatient dying of GVHD. Confirmatory results wereobtained by the Eurocord Transplant Group, whichreported a 9% incidence of grade II-IV acute GVHD inCBT recipients from an HLA-identical relative. Howev-er, it is noteworthy that the same group documented a50% incidence of grade II-IV acute GVHD in patientstransplanted from an HLA nonidentical family donor.

In CBT performed between unrelated subjects with, insome cases, a disparity of 2-3 HLA antigens, the inci-dence of acute grade III-IV GVHD is reduced (approxi-mately 10-20%)7,149,150 with respect to that observedafter unmanipulated BMT between unrelated subjectsfor whom, notwithstanding complete HLA identitybetween recipient and donor, the observed risk of acutegrade III-IV GVHD reaches at least 30-40%. In particu-lar, in the cohort of patients given CBT from an unrelat-ed donor reported by Kurtzberg et al.,149 no patient devel-oped grade IV acute GVHD or experienced hepaticinvolvement or died of acute GVHD, and only 4 out of 65patients given an unrelated CBT reported by the EurocordTransplant Group showed grade IV acute GVHD.

From the data collected up to now, therefore, it clear-ly appears that CBT, from both familial and unrelateddonors, is associated with a reduced risk of acute andchronic GVHD.148-150 In view of this observation, differ-ent centers tend to adopt less intensive schemes ofGVHD prophylaxis. Typically, children transplanted witha CB unit collected from an HLA-identical sibling receiveGVHD prophylaxis consisting of CsA alone, whereas forpatients undergoing unrelated CBT the most widely usedregimens are those based on a combination of CsA witheither low- or high-dose steroids.149,150 The associationof CsA with short-term methotrexate as proposed bythe Seattle group in BMT recipients155 is not generallyemployed due to concerns about the prolongation oftime required for engraftment and possible damage tohematopoietic progenitors with a reduction in the

potential for marrow repopulation. Procedures involvingT-cell depletion of CB cells are also discouraged.

The reported low incidence of GVHD7,148-150 might, onthe other hand, be a major drawback to the use of CB asa source of stem cells for allogeneic transplantation inleukemic patients. In fact, since the role of allogeneiclymphocytes in the control and/or eradication of malig-nancy is well established, the potential absence of GVLactivity could represent a theoretical concern in leukemicsubjects given CBT. Currently available data do not con-clusively establish whether CBT really predisposespatients to an increased risk of leukemia relapse. How-ever, considering the concern mentioned above, thechoice of less intensive GVHD prophylaxis schemes couldrepresent a possible means for partially sparing theimmune-mediated GVL effect, which may significantlycontribute to preventing regrowth of leukemia cells.

Immunological reconstitution followingcord blood transplantation

Although immunological reconstitution after BMT hasbeen extensively studied,58,59 few data are available onthe kinetics of immune recovery in CBT recipi-ents.144,153,156 After CBT, recovery of T-cell immunity, aswell as that of natural killer subpopulations, mimickswhat is described in BMT recipients.153 In particular, inthe early post-transplant period recovery of CD8+ lym-phocytes seems to be faster than that of CD4+ cells,determining a characteristic inversion of the ratiobetween the two subpopulations during the first 6months after CBT, similarly to what is described in BMTrecipients. Considering the much lower number of lym-phoid cells transferred with CBT as compared to BMT,the recovery of T-lymphocyte number and functiontowards normal must be considered rapid. The promptrecovery of T-cell immunity following CBT could be pos-itively influenced by the reduced incidence and severi-ty of both acute and chronic GVHD, which per seadversely affects the acquisition of lymphocyte func-tion. However, it must be noted that the prompt recov-ery of lymphocyte function in vitro does not necessari-ly correlate with effective in vivo immunity. In fact, atpresent there are insufficient data to prove that thisrapid T-cell recovery translates into a low incidence ofviral and fungal infections after CBT.

In contrast to what is observed in BMT recipients,58,59

an impressive increase in the percentage and absolutenumber of B-lymphocytes, apparently not related toviral infections, has been documented in children receiv-ing CBT.153,157 Possible hypotheses to account for thisobservation could involve the physiological character-istics of B-cell ontogeny in the first year of life and/ordifferent distribution of mature memory lymphocytes inbone marrow and CB.158,159

Ethical problems of cord bloodtransplantation

Like any other innovative treatment, CBT also posessome ethical questions that have still not been com-pletely resolved. In particular, these ethical considera-

W. Arcese et al.

83


tions can be subdivided into those concerning trans-plants between HLA-compatible siblings and thoseregarding CBT from unrelated donors.

The two main ethical problems regarding transplanta-tion from a family donor are those of conceiving a sib-ling with the hope of producing a compatible donor fora previous child who requires transplantation of hema-topoietic stem cells, and of his/her HLA typing in utero.Of course, any decision to conceive a child for the solepurpose of making it become a cord blood donor entailsbelittling the value of the individual to be born. Howev-er, it cannot be ignored that it is extremely difficult toseparate the reasons that lead to conceiving a child sole-ly for the joy of procreating from those linked to thepossibility of saving a living, sick child. On the otherhand, even this last reason does not lessen the impor-tance of the future child who will bring happiness to thefamily in addition to being the person who, in the caseof successful transplantation, allowed the family to savethe life of a child who would have otherwise been lost.160

In the meantime, it is important to stress the inap-propriateness of performing HLA typing in utero; becauseof the increased abortion rate due to the procedure(about 1-2%), it entails the risk of causing the death ofa healthy human being and would in any case be deeplydespicable if it were used to dispose of a conceived childfound to be HLA-incompatible with the sick patient.From the point of view of the unborn child, HLA typingin utero quite obviously poses critical problems and offersno advantages, but only tangible risks for that unbornchild’s survival. HLA typing in utero should be carried outonly when other, far more important reasons (for exam-ple, advanced age of the mother with consequent high-er risk of chromosome-21 trisomy for the fetus) suggestperforming prenatal diagnostic procedures.

Since the donor is a newborn infant, the use of cordblood for an unrelated sick patient has raised many ques-tions of ethical interest. These ethical aspects go beyondthe scope of this review, but we would like to commentbriefly on some of them. Particular attention has beendevoted to the problems raised by tests required to deter-mine whether cord blood is suitable and usable withoutthe risk of transmitting to the recipient any disease car-ried by the donor cells (namely infectious diseases andgenetically transmissible disorders). In fact, for this spe-cific aspect the ethical question is: what kind of behav-ior should be adopted by the medical operator who workswith a woman (or with the parents of a child) if a dis-ease for which there is no therapy is detected in theinfant?161 Such possibly dramatic news must cause aslittle damage as possible. We must by all means preventour increasingly profound biological awareness of ourselves from leading to a culture of anguish. One mightrecall in this regard that, for example, it has been stat-ed that minors should not be tested for abnormal genesunless there is an effective curative or preventive treat-ment that must be instituted early in life.162

Another heavily debated problem regarding unrelat-ed transplants is the case of a cord blood unit assigned

to allotransplantation, making these cells unavailablein the case the donor needs them for an autologoustransplant. However, to ensure that every CB donor hasthe right to use the donated blood for himself if neces-sary, there would be no way to provide cord blood unitsfor allotransplants. Therefore the very nature of thetechnique originally conceived for allotransplants wouldbe profoundly transformed, and this would punish alldonation ethics at their very core.

Strictly linked to these considerations is the problemof private banking of cord blood cells.163 In any case, wefirmly believe that involving money-making aspects inCB transplantation technology is unacceptable. In par-ticular, as stated by other authors as well,164 no part ofthe human body should be commercialized and CBshould not be used for the benefit of financial specula-tors.

Rational use of PBSC in the treatmentof leukemic relapse after allogeneictransplant

The cure rate of patients receiving an allogeneic trans-plant for hematological malignancies is negativelyaffected by relapse. The incidence and time of diseaserecurrence depend on several factors such as diagnosis,disease phase at the time of transplant, conditioningregimen, GVHD prophylaxis and T-cell content of theinfused graft.165

The response rate and clinical outcome of relapsingpatients with acute leukemia treated with chemother-apy alone are extremely poor.166,167 Interferon therapysignificantly prolongs the survival of CML patientsrelapsed after transplant, but its benefit is not durableover long-term follow-up.168-170 Finally, a second trans-plant offers some possibilities of cure for relapsedpatients but it carries high morbidity and regimen-relat-ed mortality.171-173

During the past few years the therapeutic approach topost-transplant relapse has been substantially modified.Following its first report by Kolb et al.,174 donor lym-phocyte infusion (DLI) is currently being used as a formof adoptive immunotherapy for patients with hemato-logical malignancies who relapse after transplant175-188

or develop EBV lymphoproliferative disorders.72,189

A number of observations in allogeneic transplantssupport the evidence that a GVL effect, whether asso-ciated or not with GvHD and mediated by donorimmunocompetent cells, contributes to the eradicationof the neoplastic clone.82,190-195

The rationale for the use of DLI in post-transplantrelapse is based on two main factors: 1) the persistenceof an immunotolerant status versus donor cells in therelapsing host; 2) the cytotoxic activity exerted by HLA-unrestricted NK and LAK cells or by HLA-restricted T-cells of donor origin against host malignant cells.188,196

However, the effectiveness of DLI therapy is variablesince it greatly depends on the type of disease and itsstage at the time of relapse. Following DLI, a high pro-


84


portion of CML patients with molecular, cytogenetic orchronic phase hematologic relapse will likely experiencea long-term disease free survival, but the success rateis substantially lower in recurrent AML and virtuallyabsent in patients relapsing with ALL or blast crisisCML.186 Furthermore, the therapeutic success of DLI iscounteracted by related severe complications such asGVHD and myelosuppression, which occur in up to 90%and 50% of cases, respectively. The mortality due to DLImay approach 20%.

Therefore in order to optimize adoptive immunotherapywith DLI and to improve the general management of post-transplant relapse, several biological and clinical condi-tions should be considered.

One of the most important factors is the time requiredfor GVL to destroy the host neoplastic cells. In early stageCML this time seems to be enough to allow a GVL reac-tion to build up and eliminate residual CML cells. By con-trast, neoplastic growth is so fast in acute leukemia thatit may not be challenged by the GVL effect.

A further variable influencing the response to DLI is thepotential of the neoplastic clone to mature and differen-tiate into dendritic cells, which contribute to the GVL reac-tion by enhancing the antigen presentation capacity oftumor cells. This property is spontaneously attained byCML cells in chronic phase and, to some extent, by AMLcells, particularly when cell differentiation follows thetumor reduction induced by chemotherapy.197 Dendriticcells derived from bone marrow or produced in vitro byCD34+ cell cultures in the presence of cytokines198,199 canexert their action through an HLA restricted mechanism.200

Therefore, in treating leukemia relapse weakly expressingHLA or tumor-specific antigens, donor hematopoieticprogenitor cells may improve the immunologic effectmediated by DLI.

Finally, bone marrow chimerism detected by PCR priorto DLI may predict either response to treatment or theoccurrence of myelosuppression. Although long-term per-sistence of donor T-cells in the peripheral blood duringrelapse has been reported,184 this observation does notprovide any prognostic information on the post-DLI clin-ical outcome. Southern blot RFLP analysis, erythrocytephenotype and cytogenetics have been employed todetect residual donor cells, but no correlation was foundamong pre-DLI BM chimerism, response to treatment andthe risk of myelosuppression. However, pre-DLI BMchimerism assessed by quantitative PCR of VNTRsequences in relapsed CML patients is associated withcytogenetic and molecular remission and strongly pre-dicts the development of aplasia, thereby providing anearly indication for the reinfusion of PBSC from thedonor.201,202

Donor PBSC reinfusion has frequently been adopted asfrom rescue of DLI-associated myelosuppression. As to thecombined use of donor PBSC and DLI, the reported expe-riences are limited to a small number of patients whorelapsed with acute leukemia.203-205

In these studies, donors were stimulated with G-CSF atdoses ranging from 2.5 mg/kg for 10 days to 16 mg/kg for

5 days. The apheresis products obtained over 1 to 3 con-secutive days contained a median of 4×108/kg CD34+ cellsand a median of 3.5×108/kg CD3+ cells. All patientsreceived chemotherapy prior to PBSC infusion and mostof them achieved CR with prompt hematopoietic recon-stitution which in the cases analyzed originated fromdonor cells. The majority of patients developed acute orchronic GVHD and related complications. In one of thereported series, the median duration of CR after this com-bined treatment was longer than the median time fromtransplant to relapse.206 These results compare favorablywith those recently reported in patients receiving DLIalone for relapse of acute leukemia or myelodysplasia afterBMT.187 Of the eight patients receiving this treatment, onlyone achieved CR and 7 died of progressive disease.

In conclusion, these preliminary experiences suggestthat patients relapsing with acute leukemia or advancedphase CML after BMT should be treated with intensivechemotherapy regimens, not necessarily includingimmunosuppressive drugs, followed by donor mobilizedPBSC.

This approach might result in certain therapeuticadvantages such as: 1) reduction of the tumor burden; 2)slowing down of the neoplastic growth; 3) acceleration ofdonor hematopoiesis recovery and promotion of dendrit-ic cell differentiation; 4) the possibility for the immuno-competent donor cells to express their GVL activity to agreater extent. Whether the additional administration ofcytokines (IFN, IL-2, G-CSF, GM-CSF) would improve theefficacy of chemotherapy and PBSC is unknown at presentand awaits further investigation.

Finally, the GVL reaction exerted by donor lymphocytesagainst CML cells which retain biological features of ear-ly stage disease is potent enough that patients might bespared a repetition of previous chemotherapy. However,donor PBSC infusion should be considered for CMLpatients with either cytogenetic or chronic phase relapsewho show minimal (< 10%) or no BM chimerism.206 Insuch cases the use of donor PBSC is mainly indicated tocounteract the risk of severe BM aplasia following theinfusion of DLI alone.

Contributions and AcknowledgmentsAll authors equally contributed to the conception and

writing of this review article.

DisclosuresConflict of interest: this review article was prepared

by a group of experts designated by Haematologica andby representatives of two pharmaceutical companies,Amgen Italia SpA and Dompé Biotec SpA, both fromMilan, Italy. This co-operation between a medical jour-nal and pharmaceutical companies is based on the com-mon aim of achieving an optimal use of new therapeu-tic procedures in medical practice. In agreement with theJournal’s Conflict of Interest Policy, the reader is giventhe following information. The preparation of this man-uscript was supported by educational grants from thetwo companies. Dompé Biotec SpA sells G-CSF and rHuE-po in Italy, and Amgen Italia SpA has a stake in Dompé

W. Arcese et al.

85


Biotec SpA.Redundant publications: no substantial overlapping

with previous papers.

Manuscript processingReceived on July 29, 1997; accepted on November 28,

1997.

References

1. Majolino I, Cavallaro AM, Scimè R. Peripheral bloodstem cells for allogeneic transplantation. Bone Mar-row Transplant 1996; 18(suppl 2):171-4.

2. Rubinstein P, Rosenfield RE, Adamson JW, StevensCE. Stored placental blood for unrelated bone mar-row reconstitution. Blood 1993; 81:1679-90.

3. Gratwohl A, Hermans J, Baldomero H, for the Euro-pean Group for Blood and Marrow Transplantation(EBMT). Blood and marrow transplantation activityin Europe 1995. Bone Marrow Transplant 1997;19:407-19.

4. Gluckman E, Broxmeyer HE, Auerbach AD, et al.Hematopoietic reconstitution in a patient with Fan-coni’s anemia by means of umbilical-cord blood froman HLA-identical sibling. N Engl J Med 1989; 321:1174-8.

5. Vilmer E, Sterkers G, Rahimy C, et al. HLA-mis-matched cord blood transplantation in a patient withadvanced leukemia. Transplantation 1992; 53:1155-7.

6. Rubinstein P, Carrier C, Adamson J, et al. New YorkBlood Center’s program for unrelated place-ntal/umbilical cord blood (PCB) transplantation: 243transplants in the first 3 years [abstract]. Blood 1996;88 (suppl 1):142a.

7. Gluckman E, Rocha V, Boyer-Chammard A, et al. forthe Eurocord Transplant Group and the EuropeanBlood and Marrow Transplantation Group. Outcomeof Cord-Blood Transplantation from Related andUnrelated Donors. N Engl J Med 1997; 337:373-81.

8. Risdon G, Gaddy J, Horrie M, Broxmeyer HE. Allo-antigen priming induces a state of unresponsiveness inhuman umbilical cord blood T-cells. Proc Natl AcadSci USA 1995; 92:2413-7.

9. Bacigalupo A, Piaggio G, Podestà M, et al. Influenceof marrow CFU-GM content on engraftment and sur-vival after allogeneic bone marrow transplantation.Bone Marrow Transplant 1995; 15:221-6.

10. Majolino I, Aversa F, Bacigalupo A, Bandini G, ArceseW, Reali G. Allogeneic transplants of rhG-CSF-mobi-lized peripheral blood stem cells (PBSC) from normaldonors. Haematologica 1995; 80:40-3.

11. Majolino I, Aversa F, Bacigalupo A, Bandini G, ArceseW. Peripheral blood stem cells for allogeneic trans-plantation. Recommendations from the GITMO1996. Haematologica 1996; 82:529-31.

12. Bertolini F, De Vincentiis A, Lanata L, et al. Allogeneichematopoietic stem cells from sources other thanbone marrow: biological and technical aspects.Haematologica 1997; 82:220-38.

13. Goodman JW, Hodgson GS. Evidence for stem cells inthe peripheral blood of mice. Blood 1962; 19:702-14.

14. Thomas ED, Collins JA, Herman Jr EC, Ferrebee JW.Marrow transplants in lethally irradiated dogs givenmethotrexate. Blood 1962; 19:217-28.

15. Cavins JA, Scheer SC, Thomas ED, Ferrebee JW. Therecovery of lethally irradiated dogs given infusions ofautologous leukocytes preserved at –80°C. Blood1964; 21:38-43.

16. Epstein RB, Graham TC, Buckner CD, Bryant J,Thomas ED. Allogeneic marrow engraftment by crosscirculation in lethally irradiated dogs. Blood 1966;28:692-707.

17. Storb R, Epstein RB, Radge H, Bryant J, Thomas ED.Marrow engraftment by allogeneic leukocytes inlethally irradiated dogs. Blood 1967; 30:805-11.

18. Storb R, Graham TC, Epstein RB, Sale GE, ThomasED. Demonstration of hemopoietic stem cells in theperipheral blood of baboons by cross circulation.Blood 1977; 50:537-42.

19. Körbling M, Fliedner TM, Calvo W, Ross WM, Noth-durft W, Steinbach I. Albumin density gradient purifi-cation of canine hemopoietic blood stem cells(HBSC): long-term allogeneic engraftment withoutGVH-reaction. Exp Hematol 1979; 7:277-88.

20. Carbonell F, Calvo W, Fliedner TM, et al. Cytogeneticstudies in dogs after total body irradiation and allo-geneic transfusion with cryopreserved blood mononu-clear cells: observations in long-term chimeras. Int JCell Cloning 1984; 2:81-8.

21. Goldman JM, Catovsky D, Hows J, Spiers ASD, GaltonDAG. Cryopreserved peripheral blood cells function-ing as autografts in patients with chronic granulocyt-ic leukemia in transformation. Br Med J 1979; 1:1310-3.

22. Richman CM, Weiner RS, Yankee RA. Increase in cir-culating stem cells following chemotherapy in man.Blood 1976; 47:1031-4.

23. Socinski MA, Cannistra SA, Elias A, Antman KH,Schnipper L, Griffin JD. Granulocyte-macrophagecolony stimulating factor expands the circulatinghemopoietic cell compartment in man. Lancet 1988;i:1194-8.

24. Gianni AM, Siena S, Bregni M, et al. Granulocyte-macrophage colony-stimulating factor to harvest cir-culating hematopietic stem cells for autotransplanta-tion. Lancet 1989; 2:580-5.

25. Kessinger A, Smith DM, Strandjord SE, et al Allogeneictransplantation of blood-derived, T-cell depletedhemopoietic stem cells after myeloablative treatmentin a patient with acute lymphoblastic leukemia. BoneMarrow Transplant 1989; 4:643-6.

26. Russell NH, Hunter A, Rogers S, et al. Peripheralblood stem cells as an alternative to marrow for allo-geneic transplantation. Lancet 1993; 341:1482.

27. Dreger P, Suttorp M, Haferlach T, Loffler M, SchmitzN, Schroyens W. Allogeneic granulocyte colony-stim-ulating factor-mobilized peripheral blood progenitorcells for treatment of engraftment failure after bonemarrow transplantation. Blood 1993; 81:1404-7.


29. Bensinger WI, Weaver CH, Appelbaum FR, et al.Transplantation of allogeneic peripheral blood stemcells mobilized by recombinant human granulocytecolony stimulating factor. Blood 1995; 85:1655-8.

30. Körbling M, Przepiorka D, Huh YO, et al. Allogeneicblood stem cell transplantation for refractoryleukemia and lymphoma. Potential advantages ofblood over marrow allografts. Blood 1995; 85: 1659-65.

31. Schmitz N, Dreger P, Suttorp M, et al. Primary trans-plantation of allogeneic peripheral blood cells mobi-lized by filgrastim (granulocyte colony-stimulating fac-tor). Blood 1995; 85:1666-72.

32. Majolino I, Saglio G, Scime’ R, et al. High incidenceof chronic GVHD after primary allogeneic peripheralblood stem cell transplantation in patients with hema-


86


tologic malignancies. Bone Marrow Transplant 1996;17:555-60.

33. Aversa F, Tabilio A, Terenzi A, et al. Successful engraft-ment of T-depleted haploidentical three loci incom-patible transplants in leukemia patients by addition ofrecombinant human granulocyte-colony-stimulatingfactor mobilized peripheral blood progenitor cells tobone marrow inoculum. Blood 1994; 84:3948-55.

34. Ringdén O, Potter MN, Oakhill A, et al. Trans-plantation of peripheral blood progenitor cells fromunrelated donors. Bone Marrow Transplant 1996;17(suppl 2):62-4.

35. Majolino I, Cavallaro AM, Bacigalupo A, et al. Mobi-lization and collection of PBSC in healthy donors. Aretrospective analysis of the Italian Bone MarrowTransplantation Group (GITMO). Haematologica1997; 82:47-52.

36. Hasenclever D, Sextro M. Safety of alloPBPCT donors:biometrical considerations on monitoring long termrisks. Bone Marrow Transplant 1996; 17(suppl 2):28-30.

37. Aversa F, Pelicci PG, Terenzi A, et al. Results of T-depleted BMT in chronic myelogenous leukaemiaafter a conditioning regimen that included Thiotepa.Bone Marrow Transplant 1991; 7:24.

38. Körbling M, Mirza N, Thall P, et al. Experience with100 HLA-identical allogeneic blood stem cell trans-plantations: CD34 cell dose and engraftment.[abstract]. Blood 1996; 88 (Suppl.I Part 1):2451.

39. Przepiorka D, Ippoliti C, Khoury I, et al. Allogeneictransplantation for advanced leukemia: improvedshort-term outcome with blood stem cell grafts andtacrolimus. Transplantation 1996; 62:1806-10.

40. Bacigalupo A, Van Lint MT, Valbonesi M, et al. Thio-tepa cyclophosphamide followed by granulocytecolony-stimulating factor mobilized allogeneic periph-eral blood cells in adults with advanced leukemia.Blood 1996; 88:353-7.

41. Khouri I, Keating MJ, Przepiorka D, et al. Engraftmentand induction of GVL with Fludarabine (FAMP) basednon-ablative preparative regimen in patients withchronic lymphocytic leukemia (CLL) and lymphoma[abstract]. Blood 1996; 88 (Suppl. I Part 1):1194.

42. Martelli MF, Aversa F, Velardi A, et al. New tools forcrossing the HLA barrier: fludarabine and megadosestem cell transplants [abstract]. Blood 1996; 88 (Sup-pl 1):1924.

43. Giralt S, Estey E, van Besien K, et al. Induction of graft-versus-leukemia without myeloablative therapy usingallogeneic PBSC after purine analog containing regi-mens [abstract]. Blood 1996; 88 (Suppl.I Part1):2444.

44. Adkins D, Spitzer G, Brown R, et al. High-dose rate(30 cGy/min) low-dose (550 cGy) single exposuretotal body irradiation (TBI) as conditioning for allo-geneic peripheral blood stem cell (ALLOPBSC) trans-plantation results in complete donor chimerism andrapid engraftment without severe acute graft versushost disease (AGVHD) or toxicity [abstract]. Blood1996; 88 (Suppl.I Part 1):443.

45. Tan P, Yeow Tee G. Stem cell allografting using a non-myeloablative regimen (Abstract). Blood 1996; 88(Suppl.I Part 1):449.

46. Slavin S, Nagler A, Naperstek E, et al. Immunotherapyof leukemia in conjunction with non-myeloablativeconditioning: engraftment of blood stem cells anderadication of host leukemia with non-myeloablativeconditioning based on fludarabine and anti-thymo-cyte globulin (ATG) [abstract]. Blood 1996; 88 (Sup-pl.I Part 1):2443.

47. Majolino I, Cavallaro AM, Santoro A, et al. Success-ful engraftment of allogeneic PBSC after conditioning

with busulfan alone. Bone Marrow Transplant 1997;19:621-3.

48. Harada M, Kodera Y, Asano S. Clinical results of allo-geneic peripheral blood stem cell transplantation:Japanese survey 1995. Bone Marrow Transplant 1996;17(Suppl 2):S47.

49. Eckart JR, Roodman G, Boldt D, et al. Comparison ofengraftment and acute GVHD in patients undergoingcryopreserved or fresh allogeneic BMT. Bone MarrowTransplant 1993; 11:125-31.

50. Schmitz N, Bacigalupo A, Labopin M, et al. Trans-plantation of peripheral blood progenitor cells fromHLA-identical sibling donors. Br J Haematol 1996;95:715-23.

51. Körbling M, Huh YO, Durett A, et al. Allogeneic bloodstem cell transplantation: peripheralization and yieldof donor-derived primitive hematopoietic progenitorcells(CD34+ Thy-1 dim) and lymphoid subsets, andpossible predictors of engraftment and graft-versus-host disease. Blood 1996; 86:2842-8.

52. Russel JA, Brown C, Bowen T, et al. Allogeneic bloodcell transplants for hematological malignancy: pre-liminary comparison of outcomes with bone marrowtransplantation. Bone Marrow Transplant 1996;17:703-8.

53. Bensinger W, Buckner C, Demirez T, Storb R, Appel-baum F. Transplantation of allogeneic peripheralblood stem cells. Bone Marrow Transplant 1996; 17(Suppl 2):S56.

54. Urbano-Ispizua A, Solano C, Brunet S, et al. Allo-geneic peripheral blood progenitor cell transplanta-tion: analysis of short term engraftment and acuteGVHD incidence in 33 cases. Bone Marrow Trans-plant 1996; 18:35-40.

55. Przepiorka D, Anderlini P, Ippoliti C, et al. Allogeneicblood stem cell transplantation: reduction in earlytreatment-related morbidity and mortality for patientswith advanced hematologic malignancies. Bone Mar-row Transplant, in press.

56. Bensinger W, Clift R, Anasetti C, et al. Transplantationof allogeneic peripheral blood stem cells mobilized byrecombinant human granulocyte colony stimulatingfactor. Stem cells 1996; 14:90-105.

57. Rosenfeld C, Collins R, Piñeiro L, Agura E, NemunaitisJ. Allogeneic blood cell transplantation without post-transplant colony-stimulating factors in patients withhematopoietic neoplasm: a phase II study. J ClinOncol 1996; 14:1314-9.

58. Roy J, Delage R, Demers C, et al. Stem cells collectedwith intermediate dose G-CSF lead to prompt engraft-ment in allogeneic blood transplant recipients[abstract]. Blood 1996; 88 (Suppl.I Part 1):1603.

59. Atkinson K. Reconstitution of the haemopoietic andimmune systems after marrow transplantation. BoneMarrow Transplant 1990; 5:209-26.

60. Lum LG. The kinetics of immune reconstitution afterhuman marrow transplantation. Blood 1987; 69:369-80.

61. Voltarelli JC, Stites DP. Immunologic monitoring ofbone marrow transplantation. Diagn Immunol 1986;4:171-93.

62. Ferrara JLM, Deeg HJ. Graft-versus-host disease. NEngl J Med 1991; 310:667-74.

63. Kato S, Yabe M, Yabe M, et al. Studies on transfer ofvaricella-zoster-virus specific T-cell immunity frombone marrow donor to recipient. Blood 1990; 75:806-9.

64. Boland GJ, Vlieger AM, Ververs C, Gast GC. Evidencefor transfer of cellular and humoral immunity tocytomegalovirus from donor to recipient in allogene-ic bone marrow transplantation. Clin Exp Immunol1992; 88:506-11.

W. Arcese et al.

87


65. Rouleau M, Senik A, Leroy E, Vernant JP. Long-termpersistence of transferred PPD-reactive T cells afterallogeneic bone marrow transplantation. Trans-plantation 1993; 55:72-6.

66. Vavassori M, Maccario R, Moretta A, et al. RestrictedTCR repertoire and long-term persistence of donor-derived antigen-experienced CD4+ T-cells in allogeneicbone marrow transplantation recipients. J Immunol1996; 157:5739-47.

67. Ottinger HD, Beelen DW, Scheulen B, Schaefer UW,Grosse-Wilde H. Improved immune reconstitutionafter allotransplantation of peripheral blood stemcells instead of bone marrow. Blood 1996; 88:2775-9.

68. Mielcarek M, Martin P, Torok-Storb B. Suppression ofalloantigen-induced T cell proliferation by large num-bers of CD14+ cells in G-CSF mobilized peripheralblood mononuclear cells. Blood 1996; 88 (Suppl1):249a.

69. Pavletic ZS, Bishop MR, Joshi SS, et al. Immunologicreconstitution after allogeneic blood stem cell trans-plantation: fast lymphocyte recovery correlates withbetter survival. Blood 1996; 88 (Suppl 1):422a.

70. Donnenberg AD, Donnenberg VS, Meyer EM, PattonTJ, Griffin EL, Ball ED. Accelerated immune reconsti-tution after matched allogeneic PBSC transplantation.Blood 1996; 88 (Suppl 1):612a.

71. Walter EA, Greenberg PD, Gilbert MJ, et al. Recon-stitution of cellular immunity against cytomegalovirusin recipients of allogeneic bone marrow by transfer ofT-cell clones from the donor. N Engl J Med 1995;333:1038-44.

72. Papadopoulos EB, Ladanyi M, Emanuel D, et al. Infu-sions of donor leukocytes as treatment of Epstein-Barrvirus associated lymphoproliferative disorders com-plicating allogeneic marrow transplantation. N Engl JMed 1994; 330:1185-91.

73. Rooney CM, Smith CA, Ny CYC, et al. Use of gene-modified virus-specific T lymphocytes to controlEpstein-Barr-virus-related lymphoproliferation. Lancet1995; 245:9-13.

74. Bensinger W, Clift R, Martin P, et al. Allogeneicperipheral blood stem cell transplantation in patientswith advanced hematologic malignancies: a retro-spective comparison with marrow transplantation.Blood 1996; 88:2794-800.

75. Anderlini P, Przepiorka D, Champlin R, Korbling M.Biologic and clinical effects of granulocyte colonystimulating factor in normal individuals. Blood 1996;88:2819-25.

76. Majolino I, Scimé R, Cavallaro AM, et al. Trans-plantation of unmanipulated allogeneic PBSC: pre-liminary report on 24 patients. Leuk Lymphoma 1998;in press.

77. Azevedo VM, Aranha FJP, Gouvea JV, et al. Allogeneictransplantation with blood stem cells mobilized byrhG-CSF for hematological malignancies. Bone Mar-row Transplant 1995; 16:647-53.

78. Deeg HJ, Huss R. Acute graft-versus-host disease. In:Atkinson K, ed. Clinical bone marrow transplantation:a reference textbook. Cambridge: Cambridge Univer-sity Press 1994. p. 297-311.

79. Chao MJ, Schmidt GM, Niland JC, et al. Cyclo-phosphamide, methotrexate, and prednisone com-pared with cyclosporine and prednisone for prophy-laxis of acute graft-versus-host disease. N Engl J Med1993; 329:1225-30.

80. Gratwhol A, Hermans J, Apperley J, et al. Acute graft-versus-host disease: grade and outcome in patientswith chronic myelogenous leukemia. Blood 1995;85:813-8.

81. Kernan NA, Collins NM, Juliano M, Cartagenia T,

Dupont B, O’Reilly RJ. Clonable T lymphocytes in T-cell depleted bone marrow transplants correlate withdevelopment of graft-versus-host disease. Blood1986; 68:770-3.

82. Sullivan KM, Storb R, Buckner CD, et al. Graft-versus-host disease as adoptive immunotherapy in patientswith advanced hematologic neoplasms. N Engl J Med1989; 320:828-34.

83. Pan L, Delmonte J Jr, Jalonen CK, Ferrara JLM. Pre-treatment of donor mice with granulocyte-colonystimulating factor polarizes donor T lymphocytestoward type-2 cytokine production and reduces sever-ity of experimental graft versus host disease. Blood1995; 86:4422-8.

84. Gorgen I, Hartung T, Leist M, et al. Granulocytecolony-stimulating factor treatment protects rodentsagainst lipopolysaccharide-induced toxicity via sup-pression of systemic tumor necrosis factor-a. JImmunol 1992; 149:918-24.

85. Hartung T, Docke WD, Grabner F, et al. Effect of gran-ulocyte colony-stimulating factor treatment on ex vivoblood cytokine response in human volunteers. Blood1995; 85:2482-9.

86. Talmadge JE, Singh RK, Agetois A, Ozerol I, Ino K. Tcell apoptosis (mediated in part by FAS ligand) bymonocytes in cytokine mobilized stem cell products:resultant immmunosuppression. Blood 1996; 88(suppl 1):610 a.

87. Anderlini P, Przepiorka D, Khouri I, et al. Chronicgraft-vs-host disease after allogeneic marrow or bloodstem cell transplantation [abstract]. Blood 1995;86(suppl 1):109 .

88. Storb R, Doney KC, Thomas ED, et al. Marrow trans-plantation with or without donor buffy-coat cells for65 transplanted aplastic anemia patients. Blood1982; 59:236-46.

89. Carlo-Stella C, Tabilio A. Stem cells and stem celltransplatation. Haematologica 1996; 81:573-87.

90. Strauss LC, Rowley SD, Larussa VF, et al. Antigenicanalysis of hematopoiesis. V. Characterization ofMY10 antigen expression by normal lymphohemato-poietic progenitor cells. Exp Hematol 1986; 14:878-86.

91. Lemoli RM, Tazzari PL, Fortuna A, et al. Positive selec-tion of hematopoietic CD34+ stem cells provides“indirect purging” of CD34– lymphoid cells and thepurging efficiency is increased by anti-CD2 and anti-CD30 immunotoxins. Bone Marrow Transplant 1994;13:465-71.

92. Dreger P, Viehmann K, Steinmann J, et al. G-CSF-mobilized peripheral blood progenitor cells for allo-geneic transplantation: comparison of T cell deple-tion strategies using different CD34+ selection sys-tems or CAMPATH 1. Exp Hematol 1995; 23:147-54.

93. Shpall EJ, Jones RB, Bearman SI, et al. Transplantationof enriched CD34-positive autologous marrow intobreast cancer patients following high-dose chemo-therapy: Influence of CD34-positive peripheral bloodprogenitors and growth factors on engraftment. J ClinOncol 1994; 12:28-36.

94. Brugger W, Henschler R, Heimfeld S, et al. Positivelyselected autologous blood CD34+ cells and unsepa-rated peripheral blood progenitor cells mediate iden-tical hematopoietic engraftment after high dose VP16, ifosfamide, carboplatin, and epirubicin. Blood1994; 84:1421-6.

95. Gorin NC, Lopez M, Laporte JP, et al. Preparation andsuccessful engraftment of purified CD34+ bone mar-row progenitor cells in patients with non-Hodgkin’slymphoma. Blood 1995; 86:1647-54.

96. Lemoli RM, Fortuna A, Motta MR, et al. Concomitantmobilization of plasma cells and hematopoietic prog-


88


enitors into peripheral blood of multiple myelomapatients: positive selection and transplantation ofenriched CD34+ cells to remove circulating tumorcells. Blood 1996; 1625-34.


98. Link H, Arseniev L, Bähre O, Kadar J, Diedrich H, Poli-woda H. Transplantation of allogeneic CD34+ bloodcells. Blood 1996; 87:4903-9.

99. Bensinger W, Buckner C, Shannon-Dorcy K, et al.Transplantation of allogeneic CD34+ peripheralblood stem cells in patients with advanced hemato-logic malignancy. Blood 1996; 88:4132-8.

100. McCullough J, Hansen J, Perkins H, Stroncek D,Bartsch G. The National Marrow Donor Program:how it works, accomplishments to date. Oncology1989; 3:63-72.

101. Kernan NA, Bartsch G, Ash RC, et al. Retrospectiveanalysis of 462 unrelated marrow transplants facili-tated by the National Marrow Donor Program(NMDP) for treatment of acquired and congenital dis-orders of the lymphohematopoietic system and con-genital metabolic disorders. N Engl J Med 1993; 328:593-602.

102. Stroncek DF, Clift RA, Sanders JE, et al. Experiences ofthe first 493 unrelated marrow donors in the nation-al marrow donor program. Blood 1993; 81:1940-6.

103. Beatty PG, Dahlberg S, Mickelson EM, et al. Proba-bility of finding HLA-matched unrelated marrowdonors. Transplantation 1988; 45:714-8.

104. Anasetti C, Hansen J. Bone marrow transplantationfrom HLA-partially matched related donors and unre-lated volunteer donors. In :Forman SJ, Blume KG,Thomas ED, eds. Bone Marrow Transplantation.Boston: Blackwell Scientific, 1994:665-79.

105. Anasetti C, Amos D, Beatty PG, et al. Effect of HLAcompatibility on engraftment of bone marrow trans-plants in patients with leukemia or lymphoma. N EnglJ Med 1989; 320:197-204.

106. Anasetti C, Beatty PG, Storb R, et al. Effect of HLA-incompatibility on graft-versus-host disease, relapse,and survival after marrow transplantation for patientswith leukemia or lymphoma. Hum Immunol 1990;29:79-91.

107. Beatty PG, Clift RA, Mickelson EM, et al. Marrowtransplantation from related donors other than HLA-identical siblings. N Engl J Med 1985; 313:765-71.

108. Ash RC, Horowitz MM, Gale RP, et al. Bone marrowtransplantation from related donors other than HLA-identical siblings: effect of T-cell depletion. Bone Mar-row Transplant 1991; 7:443-52.

109. Reisner Y, Kapoor N, Kirkpatrick D, et al. Trans-plantation for severe combined immunodeficiencywith HLA-A, B, D, DR incompatible parental marrowfractionated by soybean agglutinin and sheep redblood cells. Blood 1983; 61:341-8.

110. Reisner Y, Lapidot T, Singer TS, Schwartz E. The hostbarrier in animal models of T-cell depleted allogeneicbone marrow transplantation. In: Martelli MF, Grig-nani F, Reisner Y, eds. T-cell depletion in allogeneicbone marrow transplantation. Serono SymposiaReview 1988; 13:37-47.

111. Reisner Y, Kirkpatrick D, Dupont B, et al. Transplanta-tion for acute leukaemia with HLA-A and B noniden-tical parental marrow cells fractionated with soybeanagglutinin and sheep red blood cells. Lancet 1981;2:327-31.

112. O’Reilly RJ, Kernan NA, Cunningham I, et al. T celldepleted marrow transplants for the treatment ofleukemia. In: Gale RP, Champlin RE, eds. Bone mar-row transplantation: current controversies. New York,

Alan R Liss, 1989. p. 477-93.113. Kernan NA, Collins NH, Juliano L, Cartagena BS,

Dupont B, O’Reilly RJ. Clonable T lymphocytes in T-cell depleted bone marrow transplants correlate withdevelopment of graft versus host disease. Blood 1986;68:770-73.

114. Kernan NA, Flomenberg N, Dupont B, O’Reilly RJ.Graft rejection in recipients of T-cell depleted HLAnonidentical transplants for leukemia: identificationof host derived anti-donor allocytotoxic T-lympho-cytes. Transplantation 1987; 43:482-7.

115. Reisner Y, Martelli MF. Bone marrow transplantationacross HLA barriers by increasing the number of trans-planted cells. Immunol Today 1995; 16:437-40.

116. Schwartz E, Lapidot T, Dalia G, Singer T, Reisner Y.Abrogation of bone marrow allograft resistance inmice by increased total body irradiation correlateswith eradication of host clonable T cells and allore-active cytotoxic precursors. J Immunol 1987; 138:460-5.

117. Lapidot T, Singer TS, Salomon O, Terenzi A, SchwartzE, Reisner Y. Booster irradiation to the spleen follow-ing total body irradiation: A new immunosuppressiveapproach for allogeneic bone marrow transplanta-tion. J Immunol 1988; 141: 2619-24.

118. Cobbold SP, Martin G, Quin S, Waldman H. Mono-clonal antobodies to promote marrow engraftmentand tissue graft tolerance. Nature 1986; 323:164-6.

119. Lapidot T, Terenzi A, Singer TS, Salomon O, ReisnerY. Enhancement by dimethyl myleran of donor typechimerism in murine recipients of bone marrow allo-grafts. Blood 1989; 73:2025-32.

120. Terenzi A, Lubin I, Lapidot T, et al. Enhancement ofT-cell depleted bone marrow allografts in mice bythiotepa. Transplantation 1990; 50:717-20.

121. Aversa F, Terenzi A, Carotti A, et al. Thiotepa improvesresults of T-cell depleted bone marrow transplants foracute leukemia. A seven year experience [abstract]. BrJ Haematol 1996; 93(Suppl 2):616.

122. Hale-G, Waldmann H. CAMPATH-1 monoclonalantibodies in bone marrow transplantation. J Hema-tother 1994; 3:15-31.

123. Lapidot T, Terenzi A, Singer TS, Salomon O, ReisnerY. Quantitative aspects of bone marrow allograftrejection in mice: new model with emphasis on rele-vance to leukemia patients. Bone Marrow Transplant1988; 3 (Suppl 1):18-22.

124. Groopman JE, Molina JM, Scadden DT. Hemato-poietic growth factors. Biology and clinical applica-tions. N Engl J Med 1989; 321:1449-59.

125. Gianni AM, Bregni M, Siena S, et al. Rapid and com-plete hematopoietic reconstitution following com-bined transplantation of autologous blood and bonemarrow cells. A changing role for high dosechemotherapy? Hematol Oncol 1989; 7:139-48.

126. Aversa F, Tabilio A, Terenzi A, et al. Successful engraft-ment of T-cell depleted “three loci” mismatched trans-plants in leukemia patients by addition of rhG-CSF-mobilized peripheral blood progenitor cells to bonemarrow inoculum [abstract]. Exp Hematol 1994;22:146.

127. Reisner Y, Friederich W, Fabian I. A shorter procedurefor preparation of E-rosette depleted bone marrowfor transplantation. Transplantation 1986; 42:312-6.

128. Aversa F, Tabilio A, Terenzi A, et al. Successful engraft-ment of T-cell depleted haploidentical “three loci”incompatible transplants in leukemia patients byaddition of recombinant human granulocyte colony-stimulating factor mobilized peripheral blood prog-enitor cells to bone marrow inoculum. Blood 1994;84:3948-55.

129. Bachar-Lustig E, Rachamim N, Hong-Wei Li, Lan F,

W. Arcese et al.

89


Reisner Y. Megadose of T cell depleted bone marrowovercomes MHC barriers in sublethally irradiatedmice. Nature Med 1995; 12:1268-73.

130. Rachamim N, Gan J, Segall H, et al. Human CD34hematopoietic stem cells exhibit potent veto activity.[Submitted for publication].

131. Tabilio A, Falzetti F, Aversa F, et al. T-cell depletion forperipheral blood stem cells by CD34+ cell selectionand E-rosettes. [Submitted for publication].

132. Keating MJ, Kantarjian H, Talpaz M, et al. Fludara-bine: a new agent with major activity against chroniclymphocytic leukemia. Blood 1989; 74:19-25.

133. Terenzi A, Aversa F, Perruccio K, et al. Efficacy of flu-darabine as immunosuppressor for bone marrowtransplantation conditioning [abstract]. Blood 1996;88 (Suppl 1):2373.

134. Martelli MF, Aversa F, Velardi A, et al. New tools forcrossing the HLA barrier: fludarabine and megadosestem cell transplants [abstract]. Blood 1996; 88 (Sup-pl 1):1924.

135. O’Reilly RJ, Hansen JA, Kurtzberg J, et al. Allogeneicmarrow transplants: approches for the patients lack-ing a donor. In: Schechter GP, McArthur JR, eds.Hematology 1996, Education Program of the Ameri-can Society of Hematology. Orlando: American Soci-ety of Hematology; 1996. p. 132-46.

136. Aversa F, Terenzi A, Velardi A, et al. Megadose stemcell transplants for crossing the HLA barrier[abstract]. J Mol Med 1997; 75:176.

137. Albi N, Ruggeri L, Aversa F, et al. Natural killer (NK)-cell function and anti-leukemic activity of a large pop-ulation of CD3+/CD8+ T-cells expressing NK recep-tors for major histocompatibility complex class I after“three loci” HLA-incompatible bone marrow trans-plantation. Blood 1996 ; 87:3993-4000.

138. Bacigalupo A, Mordini N, Pitto A, et al. CD34+ select-ed stem cell transplants in patients with advancedleukemia from 3-loci mismatched family donors[abstract]. Blood 1995; 86 (Suppl 1):937.

139. Kato K, Kojima S, Kondo M, Inaba J, Matsuyama.Allogeneic bone marrow and peripheral stem celltransplantation from a haplo-identical mother andCD34 positive selection for CML. Bone MarrowTransplant 1996; 18:449-52.

140. Sierra J, Storer B, Hansen JA, et al. Transplantation ofmarrow cells from unrelated donors for treatment ofhigh-risk acute leukemia: The effect of leukemic bur-den, donor HLA-matching and marrow cell dose.Blood 1997; 89: 4226-35.

141. Stockschlader M, Loliger C, Kruger W, et al. Trans-plantation of allogeneic rh-G-CSF mobilized periph-eral CD34+ cells from an HLA-identical unrelateddonor. Bone Marrow Transplant 1995; 16:719-22.

142. Gabutti V, Foà R, Mussa F, Aglietta M. Behavior ofhuman haematopoietic stem cells in cord and neona-tal blood [letter]. Haematologica 1975; 60: 492.

143. Broxmeyer HE, Gordon GW, Hangoc G, et al. Humanumbilical cord blood as a potential source of trans-plantable hematopoietic stem/progenitor cells. ProcNatl Acad Sci USA 1989; 86:3828-32.

144. Wagner JE, Broxmeyer HE, Byrd RL, et al. Trans-plantation of umbilical cord blood after myeloablativetherapy: analysis of engraftment. Blood 1992 79:1874-81.

145. Vowels MR, Lam-PO-Tang R, Berdoukas V, et al. Briefreport: correction of X-linked lymphoproliferative dis-ease by transplantation of cord-blood stem cells. NEngl J Med 1993; 1623-5.

146. Bogdanic V, Nemet D, Kastelal A, et al. Umbilical cordblood transplantation in a patient with Philadelphia-chromosome positive chronic myelogenous leukemia.Transplantation 1993; 58:477-9.

147. Laporte JP, Gorin NC, Rubinstein P, et al. Cord-bloodtransplantation from an unrelated donor in an adultwith chronic myelogenous leukemia. N Engl J Med1996; 335:167-70.

148. Wagner JE, Kernan NA, Steinbuch M, Broxmeyer HE,Gluckman E. Allogeneic sibling umbilical cord bloodtransplantation in forty-four children with malignantand non-malignant disease. Lancet 1995; 346:214-9.

149. Kurtzberg J, Laughlin M, Graham M, et al. Placentalblood as a source of hematopoietic stem cells fortransplantation into unrelated recipients. N Engl JMed 1996; 335:157-66.

150. Wagner JE, Rosenthal J, Sweetman R, et al. Success-ful transplantation of HLA-matched and HLA-mis-matched umbilical cord blood from unrelated donors:analysis of engraftment and acute graft-versus-hostdisease. Blood 1996; 88:795-802.

151. Gisselbrecht C, Prentice HG, Bacigalupo A, et al.Placebo-controlled phase I/II trial of lenograstim inbone-marrow transplantation. Lancet 1994; 343:696-700.

152. Locatelli F, Pession A, Zecca M, et al. Use of recom-binant human granulocyte colony-stimulating factorin children given allogeneic bone marrow transplan-tation for acute or chronic leukemia. Bone MarrowTransplant 1996; 17:31-7.

153. Locatelli F, Maccario R, Comoli P, et al. Haemato-poietic and immune recovery after transplantation ofcord blood progenitor cells in children. Bone MarrowTransplant 1996; 18:1095-101.

154. Galanello R, Melis MA, Ruggeri R, Cao A. Prospectivestudy of red blood cell indices, hemoglobin A2 andhemoglobin F in infants heterozygous for b-tha-lassemia. J Pediat 1981; 99:105-7.

155. Storb R, Deeg HJ, Whitehead J, et al. Methotrexateand cyclosporine compared with cyclosporine alonefor prophylaxis of acute graft-versus-host disease aftermarrow transplantation for leukemia. N Engl J Med1986; 314:729-35.

156. Sweetman R, Rosenthal J, Sender LS, et al. Successfulengraftment, delayed immune reconstitution and min-imal GVHD following unrelated cord blood trans-plantation in children and adults [abstract]. Blood1995; 86 (Suppl. 1):388a.

157. Locatelli F, Maccario R, Zecca M. Placental bloodtransplantation [letter]. N Engl J Med 1997; 336:69.

158. Vitetta ES, Berton MT, Burger C, Kepron M, Lee WT,Yin XM. Memory B and T cells. Ann Rev Immunol.1991; 9:193-217.

159. Gathings WE, Kubagawa H, Cooper MD. A distinctivepattern of B cell immunity in perinatal humans.Immunol Rev 1981; 57:107-26.

160. Burgio GR, Nespoli L, Locatelli F. Bone marrow trans-plantation in children: between primum non nocere(above all, do not harm) and primum adiuvare (aboveall, help). In: GR Burgio, JD Lantos, eds. Primum nonnocere today. A symposium on pediatric bioethics.Amsterdam:Elsevier, 1994. p. 77-93.

161. Silberstein LE, Jefferies LC. Placental-blood banking.A new frontier in transfusion medicine. N Engl J Med1996; 335:199-201.

162. Marshall E. Clinical promise, ethical quandary. Sci-ence 1996; 271:586-8.

163. Lind SE. Ethical considerations related to the collec-tion and distribution of cord blood stem cells fortransplantation to reconstitute hematopoietic func-tion. Transfusion 1994; 34:828-34.

164. Gluckman E, O’Reilly RY, Wagner J, Rubinstein P.Patents versus transplants. Nature 1996; 382:108.

165. Kumar L. Leukemia: management of relapse after allo-geneic bone marrow transplantation. J Clin Oncol1994; 12:1710-7.


90


166. Frassoni F, Barrett AJ, Granena A, et al. Relapse afterallogeneic bone marrow transplantation for acuteleukaemia: a survey by the E.B.M.T. of 117 cases. BrJ Haematol 1988; 70:317-20.

167. Mortimer J, Blinder MA, Schulman S, et al. Relapse ofacute leukemia after marrow transplantation: natur-al history and results of subsequent therapy. J ClinOncol 1989; 7:50-7.

168. Arcese W, Mauro FR, Alimena G, et al. Interferon ther-apy for Ph-1 positive CML patients relapsing after T-cell depleted allogeneic bone marrow transplantation.Bone Marrow Transplant 1990; 5:309-15.

169. Higano CS, Raskind W, Singer JW. Use of alpha-inter-feron for the treatment of relapse of chronic myel-ogenous leukemia after allogeneic bone marrow trans-plantation. Blood 1992; 80:1437-42.

170. Arcese W, Goldman JM, D’Arcangelo E, et al. Out-come for patients who relapse after allogeneic bonemarrow transplantation for chronic myeloid leukemia.Blood 1993; 82:3211-9.

171. Sanders JE, Buckner CD, Clift RA, et al. Second mar-row transplants in patients with leukemia who relapseafter allogeneic bone marrow transplantation. BoneMarrow Transplant 1988; 3:11-9.

172. Mrsic M, Horowitz MM, Atkinson K, et al. SecondHLA-identical sibling transplants for leukemia recur-rence. Bone Marrow Transplant 1992; 9:269-75.

173. Radich JP, Sanders J, Buckner C, et al. Second allo-geneic bone marrow transplantation for patients withrecurrent leukemia after initial transplant with totalbody irradiation containing regimens. J Clin Oncol1993; 11:304-13.

174. Kolb HJ, Mittermuller J, Clemm C, et al. Donor leuko-cyte infusion for treatment of recurrent chronic myel-ogenous leukemia in marrow transplant patients.Blood 1990; 76:2462-5.

175. Drobyski WR, Roth MS, Thibodeau SN, et al. Molec-ular remission occurring after donor leukocyte infu-sions for the treatment of relapse chronic myeloge-nous leukemia after allogeneic bone marrow trans-plantation. Bone Marrow Transplant 1992; 10:301-4.

176. Drobyski WR, Keever CA, Roth MS, et al. Salvageimmunotherapy using donor leukocyte infusions astreatment for relapsed chronic myelogenous leukemiaafter allogeneic bone marrow transplantation: effica-cy and toxicity of a defined T cell dose. Blood 1993;82:2310-8.

177. Helg C, Roux E, Beris P, et al. Adoptive immunother-apy for recurrent CML after BMT. Bone MarrowTransplant 1993; 12:125-30.

178. Hertenstein B, Wiesmeth M, Novotny J, et al. Inter-feron-a and donor buffy coat transfusions for treat-ment of relapsed chronic myeloid leukemia after allo-geneic bone marrow transplantation. Transplantation1993; 56:1114-8.

179. Szer J, Grigg AP, Phillips GL, et al. Donor leucocyteinfusions after chemotherapy for patients relapsingwith acute leukaemia following allogeneic BMT. BoneMarrow Transplant 1993; 11:109-11.

180. Bar BMAM, Schattenberg A, Mensink EJBM, et al.Donor leucocyte infusion for chronic myeloidleukemia relapsed after allogeneic bone marrow trans-plantation. J Clin Oncol 1993; 11:513-9.

181. Porter D, Roth M, McGarigle C, et al. Induction ofgraft-versus-host disease as immunotherapy forrelapsed chronic myeloid leukemia. N Engl J Med1994; 330:100-6.

182. van Rhee F, Lin F, Cullis JO, et al. Relapse of chronicmyeloid leukemia after allogeneic bone marrow trans-plant: the case for giving donor leukocyte transfusionsbefore the onset of hematologic relapse. Blood 1994;83:3377-83.

183. Ferster A, Bujan W, Mouraux T, et al. Complete remis-sion following donor leukocyte infusion in ALL relaps-ing after haploidentical bone marrow transplantation.Bone Marrow Transplant 1994; 14:331-2.

184. Mackinnon S, Papadopoulos EB, Carabasi MM, et al.Adoptive immunotherapy evaluating escalating dosesof donor leukocytes for relapse of chronic myeloidleukemia after bone marrow transplantation: separa-tion of graft-versus-leukemia responses from graft-ver-sus-host disease. Blood 1995; 86: 1261-8.

185. Giralt S, Hester J, Huh Y, et al. CD-8 depleted donorlymphocyte infusion as treatment for relapsed chron-ic myelogenous leukemia after allogeneic bone mar-row transplantation. Blood 1995; 86:4337-43.

186. Kolb HJ, Schattenberg A, Goldman JM, et al. Graft-versus-leukemia effect of donor lymphocyte transfu-sion in marrow grafted patients. Blood 1995;86:2041-50.

187. Porter DL, Roth MS, Lee SJ, et al. Adoptive immuno-therapy with donor mononuclear cell infusions totreat relapse of acute leukemia or myelodysplasia afterallogeneic bone marrow transplantation. Bone Mar-row Transplant 1996; 18:975-80.

188. Slavin S, Naparstek E, Nagler A, et al. Allogeneic celltherapy with donor peripheral blood cells and recom-binant human interleukin-2 to treat leukemia relapseafter allogeneic bone marrow transplantation. Blood1996; 87:2195-204.

189. Heslop HE, Brenner MK, Rooney CM. Donor T cellsas therapy for EBV lymphoproliferation post bonemarrow transplant. N Engl J Med 1994; 331:679.

190. Weiden PL, Sullivan KM, Fluornoy N, et al. Anti-leukemic effect of chronic graft-versus-host disease.Contribution to improved survival after allogeneicmarrow transplantation. N Engl J Med 1981; 304:1529-33.

191. Apperley JF, Mauro FR, Goldman JM, et al. Bone mar-row transplantation for chronic myeloid leukemia infirst chronic phase: importance of a graft-versus-leukemia effect. Br J Haematol 1988; 69:239-45.

192. Frassoni F, Sessarego M, Bacigalupo A, et al. Com-petition between recipient and donor cells after bonemarrow transplantation for chronic myeloid leukemia.Br J Haematol 1988; 69:471-3.


194. Higano CS, Brixey M, Bryant EM, et al. Durable com-plete remission of acute non-lymphocytic leukemiaassociated with discontinuation of immunosuppres-sion following relapse after allogeneic bone marrowtransplantation. A case report of a probable graft-ver-sus-leukemia effect. Transfusion 1990; 50:175-7.

195. Collins RH, Rogers ZR, Bennett M, et al. Hematolog-ic relapse of chronic myelogenous leukemia followingallogeneic bone marrow transplantation. Apparentgraft-versus-leukemia effect following abrupt discon-tinuation of immunosuppression. Bone MarrowTransplant 1992; 10:391-5.

196. Jiang YZ, Cullis JO, Kanfer EJ, et al. T-cell and NK cellmediated graft-versus-leukemia reactivity followingdonor buffy-coat transfusion to treat relapse aftermarrow transplantation for chronic myeloid leukemia.Bone Marrow Transplant 1993; 11:133-8.

197. Mittermuller J, Kolb HJ, Gerhartz HH, et al. In vivodifferentiation of leukemic blasts and effect of lowdose ARA-C in a marrow grafted patient with leukemicrelapse. Br J Haematol 1986; 62:757-62.

198. Caux C, Dezutter-Dambuyant C, Schmitt D, et al.GM-CSF and TNF-a cooperate in the generation ofdendritic Langerhans cells. Nature 1992; 360:258-61.

199. Romani N, Gruner S, Brang D, et al. Proliferating den-

W. Arcese et al.

91


dritic cell progenitors in human blood. J Exp Med1994; 180:83-93.

200. Huang AYC, Golumbeck P, Ahmadzadeh M, et al.Role of bone marrow-derived cells presenting MHCclass I-restricted tumor antigens. Science 1994;264:961-5.

201. Rapanotti MC, Arcese W, Buffolino S, et al. Sequen-tial molecular monitoring of chimerism in chronicmyeloid leukemia patients receiving donor lympho-cyte transfusion for relapse after bone marrow trans-plantation. Bone Marrow Transplant 1997; 19:703-7.

202. Keil F, Haas OA, Fritsch G, et al. Donor leukocyte infu-sion for leukemic relapse after allogeneic marrowtransplantation: lack of residual donor hematopoiesispredicts aplasia. Blood 1997; 89:3113-7.

203. Sica S, Di Mario A, Salutari P, et al. Chemotherapyand recombinant human granulocyte colony-stimu-

lating factor primed donor leukocyte infusion fortreatment of relapse after allogeneic bone marrowtransplantation. Bone Marrow Transplant 1995;16:483-5.

204. Metha J, Powles R, Shighal S, et al. Cytokine-mediat-ed immunotherapy with or without donor leukocytesfor poor risk acute myeloid leukemia relapsing afterallogeneic bone marrow transplantation. Bone Mar-row Transplant 1995; 16:133-7.

205. Gurman G, Arslan O, Koc H, Akan H. Donor leuko-cyte infusion for relapsed ANLL after allogeneic BMTand the use of interferon alpha to induce graft-versus-leukemia effect. Bone Marrow Transplant 1996;18:825-6.

206. Carella AM, Frassoni F, Melo J, et al. New insights inbiology and current therapeutic options for patientswith chronic myelogenous leukemia. Haematologica



Background and Objectives. Hematopoietic stem cellsare being increasingly used for treatment of malignantand nonmalignant disorders. Various attempts havebeen made in recent years to expand and manipulatethese cells in order to increase their therapeutic poten-tial. A Working Group on Hematopoietic Cells has ana-lyzed the most recent advances in this field.

Evidence and Information Sources. The method usedfor preparing this review was an informal consensusdevelopment. Members of the Working Group met threetimes, and the participants at these meetings exam-ined a list of problems previously prepared by thechairman. They discussed the single points in order toachieve an agreement on different judgments, andeventually approved the final manuscript. Someauthors of the present review have been working in thefield of stem cell biology, processing and transplanta-tion, and have contributed original papers in peer-reviewed journals. In addition, the material examined inthe present review includes articles and abstracts pub-lished in journals covered by the Science CitationIndex® and Medline®.

State of Art. Over the last decade, recombinant DNAtechnology has allowed the large scale production ofcytokines controlling the proliferation and differentia-tion of hemo-lymphopoietic cells. Thus, in principle, exvivo manipulation of hemopoiesis has become feasible.The present review covers three major area of interestin experimental and clinical hematology: manipulationof hematopoietic stem/progenitor cells, cytotoxic effec-tor cells and antigen presenting dendritic cells. Prelimi-nary data demonstrate the possibility of using, in aclinical setting, ex vivo expanded hematopoietic cellswith the aim of reducing, and perhaps abrogating, themyelosuppression after high-dose chemotherapy. Con-currently, other important potential applications for exvivo manipulation of hematopoietic cells have recentlybeen investigated such as the generation and expan-sion of cytotoxic cells for cancer immunotherapy, andthe large scale production of professional antigen pre-senting cells capable of initiating the process ofimmune response.

Conclusions and Perspectives. Present and future chal-lenges in this field are represented by the expansion oftrue human stem cells without maturation, to extendthis strategy to allogeneic stem cell transplantation aswell as the manipulation of cycling of primitive progen-itors for gene therapy programs. The selective out-growth of normal progenitor cells over neoplastic cellsto achieve tumor-free autografts may ameliorate theresults of autologous transplantation. The selective pro-duction of cellular subsets to manipulate the graft ver-sus-host and graft versus-tumor effects and anti-tumorvaccination strategies may be important to improvecellular adoptive immunotherapy.©1998, Ferrata Storti Foundation

Key words: hematopoietic stem cells, bone marrow, cord blood, dendritic cells, peripheral blood, allogeneic transplantation

Hematopoietic stem cellsFollowing the discovery that bone marrow trans-

plantation could be used to rescue irradiated mice, theidentification and characterization of the hematopoi-etic stem cell has become essential in order to achievenew developments in stem cell expansion and trans-plantation (SCT).1 The potential for using stem cells asvehicles for gene therapy has further increased theefforts of a number of research groups working on stemcell identification, characterization, cloning and manip-ulation.2

Self-renewal and differentiation Marrow and blood hematopoietic cells are heteroge-

neous and belong to different lineages at differentstages of maturity. The structural and functionalintegrity of the hematopoietic system is maintained bystem cells that, by definition, comprise a relatively smallcell population, located mainly in the bone marrow,which can (i) undergo self-renewal to produce stemcells or (ii) differentiate to produce progeny which isprogressively unable to self-renew, irreversibly com-mitted to one or other of the various hematopoietic lin-

review

Ex vivo expansion of hematopoieticcells and their clinical use


MASSIMO AGLIETTA,* FRANCESCO BERTOLINI,° CARMELO

CARLO-STELLA,# ARMANDO DE VINCENTIIS,@ LUIGI LANATA,^ROBERTO M. LEMOLI,§ ATTILIO OLIVIERI,** SALVATORE

SIENA,°° PAOLA ZANON,^ SANTE TURA§

*Hematology Oncology, Ospedale Mauriziano, University ofTurin, Turin; °Department of Oncology, IRCCS FondazioneMaugeri, Pavia; #Hematology and Bone Marrow Transplanta-tion, University of Parma, Parma; @Dompé Biotec SpA, Milan;^Amgen Italia SpA, Milan; §Institute of Hematology and Med-ical Oncology "L. & A. Seragnoli", University of Bologna,Bologna; **Division of Hematology, University of Ancona,Ancona; and °°Istituto Clinico Humanitas, Milan, Italy

Correspondence: Sante Tura, M.D., Istituto di Ematologia edOncologia Medica "L. & A. Seragnoli", Policlinico S. Orsola, via Massarenti 9, 40138 Bologna, Italy

93


Clinical use of hematopoietic stem cells

eages, and able to generate clones of up to 105 lineage-restricted cells that mature into specialized cells.3Although in recent years, (i) the development of in vit-ro and in vivo assays for hematopoiesis, (ii) the identifi-cation and characterization of hematopoietic growthfactors, and (iii) the development of strategies forenriching hematopoietic cells have expanded our knowl-edge and understanding of hematopoiesis, the definitionof the stem cell, originally proposed by Lajtha4 andMcCulloch,5 has not substantially changed.

In addition to self-renewal and differentiation, a num-ber of properties are ascribed to hematopoietic stemcells, including a high migratory potential, the ability toundergo asymmetric cell divisions, the capacity to existin a mitotically quiescent form and extensively regen-erate the different cell types that constitute the tissuein which they exist.6

The issues of asymmetrical and symmetrical cell divi-sions and the regulation of self-renewal/differentiationprocess are crucial when analyzing stem cell behaviorand the potential for stem cell manipulation. Asymmet-ric cell divisions produce one differentiated daughter(progenitor cell) and another daughter that is still a stemcell (Figure 1A). When all cell divisions are necessarilyasymmetric and controlled by cell-intrinsic mechanisms,no amplification of the stem cell size is possible.7 Asym-metric divisions are referred to unequally distributedtranscription factors in daughter cells,8,9 and have beenshown to be possible in hematopoietic progenitors byclone-splitting experiments.10 Symmetric cell divisionsproduce either two progenitor cells or two stem cellsaccording to a 0.5 probability of self-renewing versusdifferentiative divisions (Figure 1B). In this case, it can beassumed that the size of the stem cell pool can be mod-ified by factors affecting the 0.5 probability value, i.e.,factors that control the probability of self-renewing ver-sus differentiative divisions.7 A third model postulatesthat individual cell divisions can be, but not necessarilyare, asymmetric with respect to daughter cell fate (Fig-ure 1C). This model also implies that daughters behavedifferently due to different local environments. Although

it is not known whether a single cell can switch from anasymmetric to a symmetric mode of cell division, avail-able evidence in the hematopoietic stem cell systemfavors a predominance of symmetric cell divisions.

The decision of a stem cell to either self-renew or dif-ferentiate as well as the selection of a specific differen-tiation lineage by a multipotent progenitor during com-mitment have been proposed to be regulated accordingto either stochastic or deterministic (inductive) models.Based on computer simulation and the distributions ofcolony-forming units in spleen (CFU-S) in individualspleen colonies, stochastic models postulate that thedecision of a stem cell to self-renew (birth) or to differ-entiate (death) is randomly regulated by a probabilityparameter "p" which is equal to 0.5 in steady-state con-ditions.11 Deterministic models postulate the existenceof lineage-specific anatomic niches that direct the dif-ferentiation of uncommitted progenitors.12 There is exper-imental evidence suggesting that the hematopoietic sys-tem may employ both stochastic and deterministicstrategies, probably depending upon the stage of lineagedifferentiation.

Based on a number of studies performed in the lastthree decades, the regulation of self-renewal, commit-ment, proliferation, maturation, and survival can beassumed to reflect highly integrated processes under con-trol of extracellular mechanisms, including regulatorymolecules and microenvironment, as well as intracellu-lar mechanisms, including protooncogenes, cell cycle reg-ulators, tumor suppressor genes, transcription factors.Regulatory molecules include positive (growth factors)and negative (interferons, TGF-b, MIP-1a) factors whichinteract in complex ways (synergism, recruitment, antag-onism).13,14 Molecules that maintain the stem cell stateare beginning to be identified. These include ligands ofthe Notch family receptors that act from outside the cellas regulators of proliferation or maintenance of theundifferentiated state,15 and factors like PIE-1 that actfrom within the cell.16 However, despite the efforts whichhave been devoted to elucidating the issue of self-renew-al control, no factors have yet been identified that are

Figure 1. Possible patternsof stem cell division. “S”indicates a stem cell; “C”indicates a committed prog-enitor cell.

94


capable of maintaining self-renewing divisions and themolecular basis of self-renewal capacity remains to beelucidated. Growth factors so far identified more proba-bly act as regulators of proliferation and survival. Theo-retically, growth factors and cell-cell interactions caninfluence the outcome of fate decisions by stem andprogenitor cells in a selective or instructive manner.According to selective mechanisms, the stem cell com-mits to a particular lineage independently of the growthfactors and the factors act to control survival and prolif-eration of committed progenitors. In instructive mecha-nisms, growth factors cause the stem cell to choose onelineage at the expense of others. The relative contribu-tion of these two mechanisms to hematopoietic regula-tion remains controversial, but experimental evidencesuggests that at least some subsets of stem/progenitorcells can be instructed by growth factors to choose onedifferentiation pathway at the expense of others.17 In theabsence of still unidentified instructive signals, it can behypothesized that environmental signals may act byincreasing or decreasing the probability of choosing aparticular fate, rather than promoting or repressing it inan all-or-none manner.2

Although growth factors play a key role in stem/pro-genitor cell proliferation and differentiation it seemsimprobable that hematopoiesis is regulated by a randommix of growth factors and responsive cells. Indeed, it islikely that regulatory molecules and localization phe-nomena within marrow stroma are required to sustainand regulate hematopoietic function.18 Stromal cells ofthe hematopoietic microenvironment provide the physi-cal framework within which hematopoiesis occurs. Theyplay a role in directing the processes by synthesizing,sequestering or presenting growth-stimulatory andgrowth-inhibitory factors, and also express a broad reper-toire of adhesion molecules which mediate specific inter-actions with hematopoietic stem/progenitor cells.19 Dif-ferential expression of adhesion molecules could causedifferent stem cell subsets to home to different marrowmicroenvironments capable of differently affecting self-renewal.

While much progress has been made in identifyingcytokines and stromal factors, little is known about intra-cellular mechanisms regulating hematopoietic stem/pro-genitor cells self-renewal and differentiation.20 Struc-ture-function analysis of growth factor receptors as wellas identification of novel signal transduction moleculeshave provided new insights into the processes involved insignal transmission pathways. Post-translational modifi-cations of pre-existing proteins, in particular tyrosinephosphorylation, play a key role in transmitting signalsand thereby linking extracellular signals to the activationof nuclear effector molecules which govern gene expres-sion.20 Accumulating evidence points to transcription fac-tors such as AML-1, Ikaros, SCL/Tal-1, Rbtn-2, Tan-1,GATA-2, and HOX homeobox genes as important regula-tors of these processes. Overexpression of HOXB4 inmurine bone marrow cells markedly increases the regen-erative potential of long-term repopulating cells and

causes an expansion in clonogenic progenitor cell num-bers, without altering their ability to differentiate nor-mally into mature myeloid, erythroid and lymphoid cells.21

In contrast, overexpression of HOXB3 causes defectivelymphoid differentiation and progressive myeloprolifer-ation.22 The glucocorticoid receptor, in combination withan activated receptor tyrosine kinase, seems to be a keyregulator of erythroid self-renewal.23 Shc overexpressionincreases GM-CSF sensitivity and prevents apoptosis ofthe GM-CSF-dependent acute myeloid leukemia cell lineGF-D8, thus suggesting that Shc is an important regula-tor of cell survival and proliferation.24 Different levels ofprotein kinase C modulate progenitor cell phenotype byfavoring myelomonocytic or eosinophil differentiation.25

Recently, it has been shown that telomerase expressioncorrelates with hematopoietic self-renewal potential.Hematopoietic stem cells show decreasing telomerelength with increasing age.26 Thus, telomerase may reg-ulate self-renewal capacity by reducing the rate of DNAshortening. Overall, intracellular mechanisms of hemato-poietic control result in the repression or de-repressionof lineage-specific genes regulating growth factorresponsiveness and/or proliferation potential.27 The exactknowledge of these mechanisms will greatly modify ourapproach to stem/progenitor cell manipulation.

In summary, stem and progenitor cell behavior is theresult of highly integrated phenomena based on extra-cellular signals triggering intracellular transduction phe-nomena. The properties of self-renewal and differenti-ation give stem cells their remarkable ability to repop-ulate the hematopoietic tissue of lethally irradiated orgenetically defective recipients. Understanding theinterplay between extracellular and intracellular regu-latory factors in controlling lineage determinationremains an important challenge for the future clinicaluse of hematopoietic cells.

Stem cell antigen(s)CD34 is a surface glycophosphoprotein expressed on

early lympho-hematopoietic stem and progenitor cells,small-vessel endothelial cells, as well as embryonicfibroblasts.28 CD34+ hematopoietic cells are morpholog-ically and immunologically heterogeneous and func-tionally characterized by the in vitro capability to gen-erate clonal aggregates derived from early and lateprogenitors and the in vivo capacity to reconstitute themyelo-lymphopoietic system in a myeloablated host.29

The CD34+ cell population contains virtually all themyeloid and lymphoid progenitors as well as a smallsubset of cells that can initiate and maintain stromalcell-supported long-term cultures. Expression of theCD34 marker has dominated attempts to isolate, purifyand characterize human hematopoietic stem cells by avariety of immunologic means.

Several monoclonal antibodies (MoAbs) assigned tothe CD34 cluster identify a transmembrane glycoproteinantigen of 105-120 kd expressed on 0.5-2% normal BMcells, 0.01-0.1% peripheral blood cells and 0.1-0.4%cord blood cells.30 The function of the CD34 antigen is

M. Aglietta et al.

95


not yet known, although it seems that CD34 is involvedin stem/progenitor cell localization/adhesion in the mar-row.31 CD34 antigen expression is associated with theconcomitant expression of several markers, includingthe lineage non-specific markers Thy1, CD38, HLA-DR,CD45RA, CD71 as well as T-lymphoid, B-lymphoid,myeloid and megakaryocytic differentiation markers.30

Analysis of the expression of CD38, Thy-1, CD71, theisoforms of CD45, and uptake of rhodamine-123 haveresulted in a consensus stem cell phenotype which isCD34bright, Thy-1+, CD38–, CD45RA–, rh-123dull, Hoechst33342dull, Lin–. CD34+ cells also express receptors for anumber of growth factors classified as tyrosine kinasereceptors, such as the stem cell factor receptor (SCF-R)or the stem cell tyrosine kinase receptors (STK), andhematopoietic receptors, not containing a tyrosinekinase domain.32 Tyrosine kinase receptors are of par-ticular relevance since their ligands might represent newfactors able to selectively control stem cell self-renew-al, proliferation and differentiation.33 Recently, CD34–

cells have been shown to have functional characteris-tics associated with stem cells and differentiate, in vivo,to CD34+ cells.34

Stem/progenitor assays to evaluate engraftment potential

Different types of progenitors can be measured direct-ly by multiparameter phenotyping of CD34+ cells or sub-populations. Although this approach has the significantadvantage that the results may be quickly available andcan be used to guide clinical decisions, correlationsbetween progenitor cell phenotype and functional activ-ity are not yet refined enough to be clinically applicable.In addition, although CD34 antigen is expressed by vir-tually all progenitor cells, the percentage of CD34+ cellswith clonogenic activity in vitro ranges from 10 to 50%.Non-clonogenic CD34+ cells include lymphoid progenitorsas well as subsets of cells which are unresponsive to con-ventional growth factors and might require the presence

of still unknown factors able to activate stem cell specificgenes.14 Figure 2 shows a schema of the cellular organi-zation of hematopoiesis based on in vitro and in vivofunctional properties of progenitor cells.

Recently, a new human hematopoietic cell, termed theSCID repopulating cell (SRC), that is capable of extensiveproliferation and multilineage repopulation of the bonemarrow of non-obese diabetic (NOD)/SCID mice has beenidentified.35,36 The SRCs which are detected exclusively inthe CD34+ CD38– cell fraction have been shown to bebiologically distinct from CFC and most LTC-IC.35,36 Withthe exception of transplantation of human cells intoimmunodeficient mice, the identification of putativehuman stem cells has essentially relied on in vitro assays.Short-term in vitro assay systems require appropriatenutrients and growth factors and are particularly suitablefor measuring quantitative changes of the differentprogenitor cell types as well as for evaluating growthfactor responsiveness or investigating differential effectsof regulatory molecules on progenitors at differentstages of differentiation or on different hematopoieticpathways.37 Short-term assays are not suitable for ana-lyzing self-renewal or interactions of hematopoieticprogenitors with stromal cells.

By using the long-term culture (LTC) technique, a sus-tained production of myeloid cells can be readily achievedin vitro, provided that a stromal layer is present, whenmarrow (or blood) is placed in liquid culture at relativelyhigh cell concentration, with appropriate supplements,temperature and feeding conditions.37

The LTC system, based on the re-establishment in vit-ro of the essential cell types and mechanism responsiblefor the localized and sustained production of hematopoi-etic cells in the marrow in vivo, offers an approach ableto investigate not only the proliferative and differentia-tive events but also self-renewal of any clonogenic celltypes. A 5- to 8-week time period between initiatingcultures and assessing clonogenic progenitor numbersallows a very primitive, self-renewing human cell, the


Figure 2. Cellular organizationof hematopoiesis.

96


so-called long-term culture-initiating cell (LTC-IC) to bequantified.38,39 Limiting dilution assays allow the fre-quency of LTC-IC and their proliferative potential (num-ber of CFU-GM generated by each LTC-IC) to be calcu-lated. Another assay system, the cobblestone area-form-ing cell (CAFC), uses a pre-formed stroma as a supportfor hematopoiesis.40 In this system, the primitive cellsare measured directly by their ability to form character-istic colonies of cells resembling cobblestones.

With the possibility of studying not only the differen-tiation but also the self-renewal of primitive progenitors,the LTC system will play an increasingly important role inthe design and assessment of new strategies involvingthe genetic engineering of hematopoietic cells and mar-row stromal cells. Hematopoietic cells that can generateactive hematopoiesis for weeks in vitro or months in vivoafter transplantation are considered stem cells. This seemsa clinically useful criteria, because it characterizes thosecells which are important for sustained hematopoieticrecovery following SCT. However, it reflects an oversim-plification of the rather complex process of hematopoi-etic function. In fact, the ability of a cell to provide long-term hematopoietic activity can either be due to a longperiod of quiescence after the initiation of the culture orbe a function of the probability of stem cell self-renewalwhich influences the long-term survival of stem cellclones.41 Thus, the number of primitive cells measured inan LTC assay will be the product of the number of stemcells present at the onset of the culture and the proba-bility of stem cell self-renewal. Although LTC assays willlikely predict the in vivo repopulating activity of the graft,the clinical definition of a stem cell does not considerthose stem cells that differentiate and die soon aftertransplantation or initiation of a culture.

Preparation of hematopoietic cells forex vivo expansion

In the vast majority of cell culture systems, the pres-ence of inhibitory mature and accessory cells limits thedegree of ex vivo expansion of the progenitor cell com-partment. Thus, a higher production of total cells, clono-genic cells and more immature hematopoietic progeni-tors has been observed when purified progenitor cells(namely CD34+ cells) rather than the whole BM, cordblood.42 or peripheral blood stem cell (PBSC) collections43

are cultured ex vivo. Haylock et al.43 selected and cul-tured 1,000 CD34+ cells in presence of a 10-fold excessof contaminating CD34–, CD3–, CD14– cells and found nodifferences in total cell production after 14 days of cul-ture, as compared to the production of 1,000 CD34+ cellsgrown alone. However, when CD34+ cells were mixedwith increasing concentrations of CD3+ CD14+ cells, amarked decrease in the total cell output was observedafter two weeks of culture suggesting an inhibitory activ-ity of monocytes and T-cells. Despite the lack of infor-mation on CFU-C production, these results point out thatthe purity of the starting population is an important vari-able and it was recommended that at least 50% of cellsshould be CD34+ in the initial cellular input.43 There are

also practical advantages in using a purified stem cellpopulation as starting material for ex vivo manipulation,such as the ease of cell handling and the amount ofcytokines consumed in the culture. Other variables thatplay an important role in stem cell expansion, which willbe discussed in detail in the following paragraphs are:initial cell density, different cytokines used and their con-centration, the presence of stromal cells, the composi-tion of the culture medium and the refeeding schedule.Conversely, it has been recently demonstrated44 thatCD34+ cells can be safely and efficiently processed aftercryopreservation suggesting that the availability of freshhematopoietic cells may not be an essential prerequisitefor ex vivo expansion. Moreover, whereas the majority ofpreclinical and clinical studies45 have attempted to opti-mize the expansion of highly enriched CD34+ cells, undercertain circumstances it may be advisable to select ear-lier subfractions of progenitor cells such as CD34+ DR–

cells in chronic myelogenous leukemia (CML) or CD34+

Thy-1+ lin– cells in multiple myeloma (MM). In these dis-eases the CD34+ cell population is still contaminated, tovarious degrees, by malignant cells;46-49 therefore, isola-tion of primitive progenitors prior ex vivo expansion mayprovide a starting cell population with a high prolifera-tive potential free of tumor cells.

Methods for hematopoietic progenitor cell(HPC) enrichment

Several methods have been proposed for purification ofHPC (Table 1). Their final target is a cell population withoptimal purity, viability and high proliferative potentialobtained by means of a low cost, rapid and simple sepa-ration technique. Early attempts toward the purificationof HPC were based on the cell physical properties. Den-sity-gradient centrifugation, velocity sedimentation andelutriation are methods that separate cells based on cellsize and buoyant density. More recently, immunologicselection techniques which take advantage of the expres-sion of specific antigens on HPC membrane, have alloweda much better degree of enrichment. Specifically, the

M. Aglietta et al.

Table 1. Hematopoietic cell separation systems.

Recovery Purity Clinical grade(% of initial cells) (% CD34+ cells) device

Immunomagnetic beadsNegative selection 20-50 20-60 YesPositive selection 30-60 50-90 Yes

MACS >80 >90 Under evaluation

Panning 30-50 40-70 No

Avidin-biotin 40-60 50-90 Yesimmunoabsorption

FACS (high-speed cell sorting) 30-50 >95 Yes

97


demonstration of the presence of the CD34 antigen onHPC28 has led to a number of positive selection systemswhich use MoAbs to purify hematopoietic precursors.

Alternatively, CD34+ cells can be enriched by depletionof CD34– accessory and mature cells.

Fluorescence activated cell sorter (FACS)Flow cytometry can be used to separate HPC from a

heterogeneous population after incubation of cells withfluorochrome-conjugated MoAbs directed to cell-sur-face markers. Moreover, multiparameter enrichment canbe obtained by combining physical properties such ascell size and cytoplasmic granularity and intracellularcharacteristics indicating cellular function (e.g. propid-ium iodide to determine cell viability; rhodamine-123 toassess metabolic quiescence and nucleic acid dyes toevaluate cycling status). This cell sorting technique canyield a highly purified (> 99%) cell population combin-ing positive and negative markers for HPC using clini-cally-graded MoAbs. The main criticisms to the use offlow cytometry for selecting large numbers of cells arethe low recovery of target cells and the length of timerequired to process the whole BM harvest or the leuka-pheresis products. The development of multiparameterhigh-speed cell sorting has been described and recent-ly upgraded for clinical use.50,51 Viable cells have beensorted at rates as high as 40,000 cells/sec as comparedto 2,000-5,000 cells/sec of commercially available cellsorters. Thus, the sample processing time can now bereduced to 8-12 hours. The sorted cell fraction alsomaintains its hematopoietic potential based on the pres-ence of CFU-C, more immature CAFC and long-termrepopulating cells in mice.51 Presently, selection of HPC(i.e. CD34+ Thy-1+ lin– cells) from clinical samples isdirected toward the purification of cell populationshighly enriched for HPC free of contaminating malig-nant cells in myeloma patients.47

PanningAnti-CD34 MoAbs bound to the bottom of cell cul-

ture flasks have been used to select CD34+ cells.52 Thetarget cell population present in a heterogeneous cellsuspension is blocked on the plastic surface while CD34–

cells remain in the supernatant and can be easily elim-inated. Despite the good results reported, the availabil-ity of more efficient methods of cell separation havemade this technique largely redundant.

Immunomagnetic systemsA variety of magnetic cell-separation methods have

been described.53 Some of these systems are commer-cially available and have been used in clinical trials. Themain differences between the currently used magneticcell-separation methods are the composition and size ofthe magnetic particles used for labeling the cells and theseparation process. Superparamagnetic beads can beequally used for negative and positive cell separationdepending on the specificity of MoAbs. The rosetted tar-get cells can be easily isolated from unlabeled cells bya magnet applied on the outer wall of the test tube or

blood bag. Large magnetic beads (diameter >0.5 µm)have been used clinically for the purging of neuroblas-toma and lymphoma cells from stem cell harvests priorto autologous transplantation.53,54 Immunomagneticbeads coupled with anti-CD34 MoAb can be used forpositive selection of HPC.55 However, before the clinicaluse of the enriched cell fraction, the cell-bound parti-cles must be removed to avoid damage to the cellsand/or toxic events to the patient. Beads can be releasedusing chymopapain or a peptide competing with theCD34 Ab. In alternative, HPC can be enriched by nega-tive depletion of mature and accessory cells targetinglineage-specific antigens.

More recently, the magnetic cell-sorting (MACS) sys-tem has been proposed as an efficient and more man-ageable alternative to flow cytometry for cell separa-tion.56 It uses colloidal-sized superparamagnetic parti-cles made of dextran and iron oxide with 60 nm ofdiameter. The use of very small beads minimizes unspe-cific binding and allows the efficient isolation of rarecells. In addition, the magnetic particles are readilyinternalized by the labeled cells without affecting theirphysical, phenotypic and functional capacity.57 Table 2reports the results of a large number of experiments(=14) comparing the efficiency of the Mini-MACS sys-tem for selecting CD34+ cells from two different cellu-lar sources (R.M.L., unpublished observations).

High-affinity chromatography based onavidin-biotin immunoabsorption

This technique relies on the high affinity between theprotein avidin and the vitamin biotin whose interactionhas an extremely high dissociation constant (= 10-15 M).In this system, a heterogeneous cell population is incu-bated with a biotinylated antibody to the CD34 antigen.The cell mixture is then passed through a disposable col-umn containing avidin-coated polyacrylamide beads.CD34+ cells are retained on the column due to the highaffinity binding of biotin to avidin while negative cellsare washed away. Target cells are then recovered bymechanical agitation of the column which disrupts theantibody-antigen link. Thus, bound cells are eluted fromthe column mainly free of antibody. An automated ver-sion of the device controlled by a computer which guar-

Table 2. High efficiency of Mini-MACS separation system forthe enrichment of BM or circulating CD34+ cells.

EnrichmentSource Pre Post Recovery CE* LTC-IC

(%) (%) (%) (x104 CD34+)

BM 2.3±1 97±3 88±9 3.6±0.3 62.5±54

PB 0.7±0.4 98.9±1 90±8 3.9±0.4 48.2±35

*Abbreviations: CE, clonogenic efficiency; BM, bone marrow; PB, peripheral blood.The results are expressed as mean±SD.


98


antees reproducibility and reduces risks of operator errorshas been developed and used in the setting of autologousand allogeneic stem cell transplantation.46, 58

Dynamic systems In addition to positive or negative selection of CD34+

cells, effective ex vivo expansion of hematopoietic prog-enitor cells and, perhaps, putative stem cells can also beachieved in the presence of unselected cell populations.In this case the use of dynamic perfusion cultures isstrictly required. As stated above, the production ofinhibitory factor(s) by mature and accessory cells ratherthan the availability of growth promoting factors, isprobably the main limitation to successful stem cellexpansion. For instance, it is well known that subopti-mal cell expansion occurs when CD34+ cells are culturedat concentrations exceeding 104 cells/mL. Moreover, thepresence of stromal feeder-layer cells seems to beimportant for effective BM stem cell expansion inducedby exogenous cytokines. Therefore, an artificial capil-lary-perfusion system (Bioreactor) in which nutrientsconsumed by proliferating cells are continuously replen-ished by exchange of the nutrient-depleted mediumwith fresh culture medium, has been tested for ex vivoexpansion of hematopoietic cells co-cultivated withstroma cells or stromal cell lines.59 Cytokines such as IL-3, IL-6 and GM-CSF have been added to the culture tooptimize cell expansion and to provide growth factorswhich are not produced by stromal cells (i.e. IL-3). In thissystem, cultured cells are confined in a small compart-ment separated from a large medium reservoire. Medi-um exchange is optimized when there is a maximalmembrane surface area per unit volume across whichmedium and nutrients can pass by diffusion. This is bestachieved by a capillary-perfusion module. Three impor-tant requirements for optimal expansion of hematopoi-etic cells are: the capillary porous size, the capacity ofsupporting the attachment of stromal cells by the mod-ule and the minimal cell activation by the materials ofthe module. In fact, mature myeloid cells such as macro-phages release ,upon surface activation, tumor necrosisfactor(s) (TNFs), interferon(s) (IFNs) and other substanceswhich negatively affect stem cell expansion.

Bioreactors have induced a remarkable expansion ofcommitted hematopoietic progenitor cells coupled witha modest increase in the number of LTC-IC,59 a popula-tion of primitive cells which correlates most positivelywith the long-term reconstituting capacity of autolo-gous and allogeneic grafts.

Ex vivo expansion of myeloidprogenitor cells

Ex vivo expansion of hemopoietic progenitors mightresult in:• amplification of the population of committed pro-

genitors due to an extensive, although controlled, dif-ferentiation process;

• amplification of the stem cell pool through extensiveself-renewal of the early progenitor cell population.

Obviously, both processes can take place simultane-ously mimicking, in vitro, the complex interplay of reg-ulatory mechanisms that allows hemopoiesis in vivo. Thelatter situation, so far, has never been obtained in vitro,whereas different approaches have permitted the firstgoal to be reached (at least, to a certain extent) andsome recent data suggest that relevant self-maintain-ing processes can be triggered.

The clinical relevance of extended differentiation ver-sus self-renewal is obviously different. The induction ofan increased in vitro production of committed progeni-tors might hasten the early phases of hemopoietic recon-stitution which occur after myeloablative treatment andstem cell transplantation. Moreover an increased num-ber of infused cells might modulate graft versus-hostdisease (GVHD) intensity in the allogeneic setting.60

Although relevant, the potential clinical benefit oftechniques allowing only committed progenitor cellexpansion is outweighed by the possibility of triggeringthe self maintenance, and perhaps amplification, of ear-ly hemopoietic progenitors. In this situation, startingfrom a limited number of progenitors, long-term recon-stitution of hematopoiesis might become feasible.Moreover, ex vivo manipulation of primitive hemopoiet-ic cells could be performed under experimental condi-tions suboptimal for the growth of neoplastic cell con-taminating autologous grafts. Thus, a purging effectcould be obtained.

Several attempts of in vitro expansion of hemopoiet-ic progenitors have been published in the last years.Recent reviews61, 62 summarize early experiences.

The first studies on ex vivo generation of hematopoi-etic progenitors involved liquid culture in the presenceof cytokines such as SCF, IL-1, IL-3, IL-6, G-CSF, GM-CSFand Epo. These experiences showed that a relevantincrease (from 10 to 1,000 fold) of CD34+ cells and ofcommitted progenitors can be obtained. The expansionof committed progenitors does not mean, however, thatthe long-term reconstitution of hemopoiesis is achiev-able. Indeed, several data support the concept that anuncontrolled commitment decreases the stem cell pool.Yonemura et al.63 have reported that IL-3 or IL-1 abro-gates the reconstituting ability of hematopoietic stemcells. Furthermore, Peters et al.64 have demonstratedthat ex vivo expansion of murine marrow cells with IL-3, IL-6, IL-11 and SCF leads to impaired engraftment inirradiated hosts.

As indicated by Traycoff et al.65 ex vivo expansion ofhematopoietic cells using SCF, IL-1a, IL-3 and IL-6 gen-erates classes of cells possessing different levels of BMrepopulating potential based on their cycling status.Along this line, Young et al.66 correlated a higher prolif-erative capacity with a quiescent status after ex vivoexpansion in the presence of SCF, IL-3, IL-6 and LIF.Taken together, these data suggest that cultures in thepresence of cytokine combinations based upon SCF, IL-1 and IL-3 involve differentiation of BM or mobilizedCD34+ cells entering the S phase. Different results wereobtained by Di Giusto et al.,67 who found that ex vivo

M. Aglietta et al.

99


expanded cord blood CD34+ cells repopulated the mar-row of immunodeficient mice as well as non-expandedcells. However, it must be remembered that cord bloodis rich in hemopoietic progenitors68 that have anincreased proliferation potential.69

New strategies to induce the expansion of CD34+ cellswith little (or no) differentiation might involve differentapproaches. The use of stirred suspension70 or hollowfiber71 bioreactors has been proposed in order to growcells in a more physiologic environment, and inhibitorssuch as TGF-b and MIP-1a have been the object ofintense studies. Recently, MIP-1a has been found toexert a weak inhibitory effect on CD34+CD38– cells andto enhance the proliferation of CD34+CD38+ cells,whereas TGF-b strongly inhibits both cell popula-tions.72,73 The most promising results, however, havebeen obtained with the recent introduction of FL andTpo. FL, a recently discovered member of the class IIItyrosine kinase receptor family,74 is able to induce pro-liferation of very early hematopoietic progenitors thatare non-responsive to other early acting cytokines, andto improve the maintainance of progenitors in vitro.75-77

This is also supported by the finding that FL significant-ly reduced the number of cultured cells undergoingapoptosis.78 Analysis of the effects of 16 cytokines onCD34+ CD38– cells showed that FL, SCF and IL-3 pro-duced a 30-fold amplification of the input of LTC-IC.79

Yonemura et al.80 compared FL- and SCF-driven ex vivoexpansion. They reported that both cytokines, in com-bination with IL-11, enhanced the production of prog-enitors, but with different kinetics. In fact, the maximalexpansion by FL required a longer incubation than withSCF. Interestingly, in these studies the combination ofSCF/IL-11, together with IL-3, reduced the ability of cul-tured cells to reconstitute hematopoiesis in irradiatedhosts.80 Other recent data have compared the effect ofFL and SCF. FL acts as a self-renewal or prolifera-tion/expansion signal for CD34+-low cells while theeffect of SCF is more likely to transduce a differentia-tion signal, resulting in more rapid repopulation at theexpense of cell expansion.81 Gene transfer studies inmice have also demonstrated that FL maintains the abil-ity of human CD34+ cells to sustain long-term hemato-poiesis. In fact, incubation of CD34+ cells with FL beforetransduction was associated with long-term provirusexpression, whereas provirus expression declined inrecipients of CD34+ cells transduced in the absence ofFL.82 The expansion ex vivo of early progenitors seems tobe affected at the single cell level by changes in cytokineconcentrations. In a recent paper by the Vancouvergroup,83 maximal LTC-IC expansion was obtained in thepresence of 30 times more FL, SCF, IL-3, IL-6 and G-CSFthan could concomitantly stimulate the near-maximalamplification of CFC.

Tpo, the ligand of the mpl receptor expressed on bothearly and committed hematopoietic progenitors,84 isknown to support megakaryocytopoiesis.85 Moreover, ithas been shown to be capable of enhancing ex vivoexpansion of early/committed progenitor cells. As single

factors, FL and Tpo stimulated a net increase of LTC-ICgenerated from CD34+ CD38– cells within 10 days.79 Fur-thermore, as demonstrated in mice recipients of BMcells transduced with the mpl receptor,86 Tpo does notinduce lineage-restricted commitment of mpl-receptorpositive pluripotent progenitors but permits their com-plete erythroid and megakaryocytic differentiation. Tpohas also been found to increase the multilineage growthof CD34+ CD38– cells from 3%, in absence of thecytokine, up to 40% when Tpo is added to SCF and FL.87

The presence of additional cytokines such as IL-3, IL-6and Epo does not significantly enhance clonal growthabove that observed in response to Tpo, SCF and FL.87

lnterestingly, the soluble form of Tpo receptor and G-CSFreceptor directly stimulate the proliferation of primitivehematopoietic progenitors of mice in synergy with SCFand FL.88

A step toward extensive ex vivo amplification of ear-ly human progenitor cells has been reported by Piaci-bello et al.89 They first demonstrated that IL-3 inducesan early production of committed progenitors but is notable to sustain true self maintenance of hemopoieticstem cells even in the presence of other early actingcytokines (FL, Tpo, SCF). Afterwards, several combina-tions of early acting cytokines were tested for their abil-ity to sustain long-term hematopoiesis in stroma freecultures. Among the various combinations tested onpurified cord blood CD34+ cells, the mixture of Tpo + FLwas found to be able to maintain early progenitors upto six months.89 These data indicate the enormouspotential of cord blood progenitors and the key role ofTpo and FL in the regulation of early hematopoiesis.However, several issues remain to be clarified:• is it true self-renewal or a slow differentiation of cord

blood cells, which are rich in immature progenitors?• what is the in vivo repopulating capacity of ex vivo

expanded cells?• is such expansion possible using CD34+ cells obtained

from the marrow or peripheral blood of adult sub-jects?In this view, while a number of papers have already

reported that committed progenitors can be generatedand safely administered to transplant recipients, thereare no reports on expansion of cells with long-termrepopulating capacity in humans.

Brugger et al.45 reported the successful reconstitutionof hematopoiesis in ten cancer patients transplantedwith autologous cells generated from CD34+ cells cul-tured in the presence of SCF, IL-1b, IL-3, IL-6 and Epo.However, the conditioning regimen given to thesepatients was not fully myeloablative, and this studyoffered no insight into the long-term engraftmentpotential of cells generated in this fashion. A similarapproach was followed by Alcorn et al.44 In ten patientswith malignancy, an aliquot of the PBSC harvest wasrecovered from liquid nitrogen and CD34 were selected.Cells were cultured for 8 days in the presence of thesame cytokine combination. A mean of 379×106 expand-ed cells were reinfused in addition to unmanipulated


100


cells. The authors reported that the total LTC-IC num-ber was not increased and most of the CD34+ cells weredifferentiated in front of an average 15-fold CFU-GMexpansion (range 4-39). Similarly, averages of 71 (range27-151)-fold megakaryocytic cell expansion and 1,040(19-16.000)-fold erythroid cell expansion were report-ed. Although adverse reactions were not reported, nodifference in the kinetics of engraftment was observedin comparison with historical controls.

Different results come from preclinical studies wherethe infusion of large numbers of ex vivo expanded com-mitted hematopoietic progenitors, together with unma-nipulated cells might speed engraftment afterchemotherapy and/or total body irradiation (TBI). Datafrom Uchida et al.90 have suggested in the past that mostof the short- as well as the long-term engraftmentpotential resides in uncommitted progenitors. Morerecently, Szilvassy et al.91 have demonstrated that par-tially differentiated ex vivo expanded cells acceleratehematologic recovery in myeloablated mice transplant-ed with highly enriched long-term repopulating stemcells. In humans, Williams et al.92 reinfused 9 breast can-cer patients with unmanipulated apheresis productstogether with a mean of 44×106/kg mature CD15+ cellsgenerated ex vivo by CD34+ cells cultured in the pres-ence of PIXY-321. No toxicity was observed after rein-fusion, and time of white cell recovery was similar tothat observed in the retrospective control group. In amore recent study,78 megakaryocytic progenitors (MP)were obtained from CD34+ cells cultured in serum-freemedium in the presence of Tpo, FL, SCF, IL-3, -6, -11 andMIP-1a. Proliferation peaked on day 7 in culture, and a8±5-fold expansion of CD34+/CD61+ cells, a 17±5-foldexpansion of CFU-MK and a 58±14-fold expansion ofthe total number of CD61+ cells was obtained. Ten can-cer patients undergoing autologous PBPC transplantreceived MP generated ex vivo (range 1-21 CD61+ cells×105/kg) together with unmanipulated PBSC. Platelettransfusion support was not needed in 2 out of the 4patients receiving the highest dose of cultured MP andthis result compared favorably with a retrospective con-trol group of 14 patients, all requiring platelet transfu-sion support. A major concern is the potential expansionof contaminating tumor cells along with hematopoiet-ic progenitors. In fact, it has been demonstrated thatCD34+ cell selection decreases (but does not abrogate)neoplastic cell contamination from aphereses of myelo-ma patients.46 For instance, in the majority of B cell lym-phoma patients CD34+ cell selection does not eliminatecontaminating t(14;18)+ cells. However, during ex vivoexpansion residual lymphoma cells do not proliferateand become undetectable by molecular analysis in themajority of cases.93 Similarly, Vogel et al. recently indi-cated that exogenously mixed epithelial tumor cell linesmight have a relative disadvantage over CD34+ cells dur-ing ex vivo expansion.94

Future directionsFuture challenges in this field are represented by the

expansion of true human stem cells without matura-tion, to extend this strategy to allogeneic stem celltransplantation, and especially cord blood allograft, aswell as the manipulation of cycling of primitive prog-enitors for gene therapy programs.

Selective amplification of specific myeloid lineages (e.g.platelets or granulocytes) may improve the results ofautologous transplantation. Moreover, although earlyresults need confirmation, the amplification of early/com-mitted hematopoietic cells coupled with the removal ofneoplastic cells contaminating autologous grafts appearsto be feasible.

Expansion of cytotoxic effectorsHuman cytotoxic effector (CE) cells can be divided in

two major groups:1. cells requiring prior antigen sensitization, which rec-

ognize their target in the context of the major his-tocompatibility complex (MHC) molecules;

2. cells not requiring prior antigen sensitization beingspontaneously cytotoxic against tumor target cells(e.g. K-562 cell line) in a non-MHC restricted setting.

While the first group includes only some subsets of T-lymphocytes (CD8+ or CD4+ cells), the second one ismore heterogeneous and includes both T-cells and nat-ural killer (NK) cells, expressing the CD56 antigen.95

The so-called antibody-dependent cellular cytotoxic-ity (ADCC) can be mediated by cells expressing the Fcgreceptor II and the Fcg receptor III (e.g. NK cells andCD3+/CD16+ cells). Although this activity is not exhibit-ed by non MHC-restricted cells, it cannot be consideredaspecific and it is also exerted by monocytes.

The lymphokine-activated killer cells (LAK) are capa-ble of killing NK-resistant cellular targets (e.g. Daudi cellline). Although some tissue-resident lymphocytes mayhave spontaneous LAK activity, normal blood mono-nuclear cells do not show any LAK activity, which can beacquired after incubation with Interleukin-2 (IL-2).96

Therefore the LAK assay is a measure of the capacityof T and NK cells to become activated and to expresscytolytic function.

Killing mechanisms of cytolytic effectorsCytotoxic T-Lymphocytes (CTL) and NK cells possess at

least two distinct, fast-acting, lytic mechanisms:97, 98

1. the granule exocytosis pathway involves the secre-tion of perforin and granzymes which penetratethroughout the target cell pores, inducing cell death;

2. a non-secretory mechanism which is mediated bythe interaction between the Fas-ligand, expressedby the killer-cell and Fas (CD95) which triggers acascade of proteolytic enzymes leading to apoptosisof both the killer and the target cell.

A third cytolytic pathway, involving TNF, has recent-ly been described.99

A series of apoptosis-resistant clones of human lym-

M. Aglietta et al.

101


phoma cells has been described. These cells expressfas/APO-1 receptors lacking the intracytoplasmic sig-naling domain.100

NK cells, as well as CTL, can recognize MHC class Imolecules. However, recognition of MHC on target cellsdownregulates the NK cell function, suggesting thepresence of inhibitory receptors.101

LAK cells are not MHC-restricted and are also capa-ble of killing freshly isolated tumor cells. Similarly toCTL and NK cells, their activity is mediated by both lyt-ic pathways (perforin/granzyme) and Fas-mediatedapoptosis.

Cytokines involved in CE functionSeveral cytokines affect CTL and NK cell response. In

particular, IL-2 expands the pool of alloreactive CTL pre-cursors and IL-15 (produced by monocytes) mimics IL-2 action by inducing g-IFN production, the activation ofmemory T cells and CTL proliferation.102

In addition, GM-CSF can affect certain T-lymphocytefunctions by enhancing their cytotoxicity and g-IFNproduction. This multifunctional cytokine can also aug-ment NK cell function and the expression of adhesionmolecules on the surface of leukemic cells.103,104 More-over, it has been hypothesized that the association ofGM-CSF/IL-2 can also be useful for the activation ofcytotoxic effectors by circulating progenitors, preserv-ing the clonogenic potential of normal hemopoietic pre-cursors.105

IL-12 elicits the production of g-IFN by CTL thusenhancing their antineoplastic efficacy and promotes thedifferentiation of T-helper-1.106,107 Finally, IL-7 seems tobe critical for the development of CTL and for a fastimmune reconstitution after bone marrow transplanta-tion (BMT).108

In conclusion, these cytokines play a pivotal role in theimmune response against tumor cells by expanding, acti-vating and recruiting CE and secondary effector cells(macrophages) or by directly inhibiting tumor cell growth.

Role of cytotoxic effectors in immuno-surveillance

There are several in vitro and in vivo data supportingthe role of immunosurveillance in tumor growth con-trol.109 The graft-versus-leukemia (GVL) effect has beendemonstrated to play a critical role in the eradication ofminimal residual disease (MRD) after allogeneictrasplantation and there is evidence supporting the roleof both T cells and NK cells in preventing diseaserelapse.110

Today, there is no doubt that CTL have a major role inkilling allogeneic tumor cells in a MHC-restricted man-ner. For example, in CML, the higher relapse incidenceafter T-cell depleted allogeneic BMT111 and the dramat-ic effect of donor T-lymphocyte infusion after relapsefollowing BMT,112 strongly support the importance ofMHC-restricted GVL.

Unfortunately, a selective GVL effect (separated fromGVHD) can be obtained very rarely in patients receivingallogeneic BMT. Moreover, although animal models indi-cate that autologous GVHD exists and could generate asignificant antitumor effect, the high incidence of relapsein patients receiving autologous BMT demonstrates thatautologous GVL is often clinically ineffective.113

Rationale for cytotoxic effectors’expansion

There are many important reasons to increase thenumber, the efficacy and the specificity of cytolyticeffectors both in the allogeneic and in the autologoussetting. The main clinical goals are the following:1. to reduce the relapse-incidence after autologous

BMT, which is still relevant in acute leukemia, lym-phoma and breast cancer;

2. to cure diseases in which autologous BMT can onlyprolong the survival (multiple myeloma, CML, meta-static chemosensitive cancers);

3. to reduce the incidence of relapse after allogeneicBMT especially in patients transplanted with greattumor burden;

4. to accelerate the immune reconstitution after BMT,in order to reduce the morbidity and transplant-related-mortality caused by serious infections (e.g.CMV and systemic mycosis).

The CE which are investigated for in vivo or ex vivoexpansion are mainly NK cells in the autologous settingand CTL in the allogeneic setting.

Figure 3. NK cells differentiation pathway. Modified from Miller et al, Blood 1994; 83: 2594-601.


102


M. Aglietta et al.

Development and expansion of NK cells NK cells belong to the naive part of the immune sys-

tem and begin to appear into PB early after allo andauto-BMT. These cells express the N-CAM homologousCD56 antigen, but lack T-cell receptor a/b complexes.They are also characterized by low affinity receptors forIgG (CD16)114-116 and their binding to malignant cells ismediated by CD18 molecule.117 The most importantorgan for NK differentiation is the BM. Several data sug-gest that a common precursor of T cells and NK cellsdoes exist. Fetal NK cells express the TCR g, d, e, j sub-units while fetal B precursors do not express TCR sub-units (Figure 3).

These bipotential T/NK precursors do not have TCRrearrangement and share CD34/CD33/CD7 antigens.They differentiate to NK in the presence of SCF, IL-7, IL-2 and stromal feeder cells. The first step of NK differen-tiation is stroma-dependent while the second step of NKmaturation is stroma-independent and is stronglypotentiated by the association of IL-2 with IL-7.118 In allstudies IL-2 is required for NK cell differentiation fromCD34+ cells. This finding suggests that a fraction ofCD34+ cells expresses IL-2 receptors and that activatedT-cells (as IL-2 source) should be detectable in the cel-lular milieu. However, T-cell deficient mice have normalNK cell development and humans lacking the g-chainsubunit of IL-2R lack NK activity. These unexpectedobservations have recently been supported by thedemonstration that IL-15, produced by BM stromal cells,can directly induce CD34+ cells to differentiate intoCD3–/56+ NK cells in the absence of IL-2.119

The last step of NK development, after the expressionof CD16, is characterized by the appearance of the CD56molecule. The intensity of CD56 expression directly cor-relates with the proliferative potential and the killingability of NK cells.120

There is clear evidence that mature NK elements havea clonally-distributed ability to recognize their targetcell by class I MHC alleles and a precise correlation hasbeen established between the expression of p58 recep-tors on NK cell surface and class I MHC alleles. Thesereceptors transduce an inhibitory signal upon interac-tion with MHC class I antigens, to prevent NK cells fromkilling target cells expressing certain (self) HLA alleles.

These findings are consistent with a self-tolerancemechanism exerted by the NK population which can bedisrupted as a consequence of tumor transformation orviral infection or any other events inducing (or mask-ing) class I molecules.121,122

After incubation with IL-2, NK cells become LAK cellscapable of killing otherwise NK-resistant target cells.These (NK) activated cells express new markers such asCD25, MHC class II and fibronectin which can be usefulfor the evaluation of their functional state. The NK cellcompartment is heterogeneous and distinct subsetshave been characterized. The most informative func-tional differences are based on relative CD56 fluores-cence: only CD56+bright, but not CD56+dim NK cells, expressthe high-affinity IL-2 receptor. As a consequence, they

respond to low concentrations of IL-2 and expand 10times more than CD56+dim. This subset seems to be sig-nificantly reduced in leukemic patients. A remarkablereduction of CD56+bright NK cells has been observed inCML patients coupled with a significant decrease oftheir spontaneous cytotoxicity against the K-562 line.However, this defect was corrected by 18 hours incu-bation with 1000 U/mL of recombinant IL-2.120 Thesedata strongly suggest that during tumor progression,the NK compartment (and particularly the small fractionof NK-CD56+bright with high proliferative ability) is pro-gressively suppressed even though the exogenousadministration of IL-2 can partially reverse this phe-nomenon. The strong correlation between functionalcapacity of the NK cell compartment and tumor pro-gression has often been reported as well as the effica-cy of the administration of LAK cells plus IL-2 in restor-ing an anti-tumor response.123

Along this line, more than 90% of patients with acuteleukemia in complete remission do not show sponta-neous cytotoxicity against autologous blast cells. How-ever, ex vivo treatment with IL-2 restores cytolytic activ-ity in 37.5% of these patients.124

The first attempts to generate and expand LAK activ-ity either in vivo or in vitro (after ex vivo incubation withIL-2) have been clinically disappointing especially inpatients autotransplanted for acute lymphoblasticleukemia (ALL).125 Nevertheless, there is now a renewedinterest in the use of activated NK cells in hematologicmalignancies, based on the optimization of differentapproaches: • administration of IL-2 in vivo to expand functional-

ly active CE in patients with low tumor-burden, inorder to reach an optimal effector/target ratio;

• harvesting and culturing large amount of NK cells foradditional ex vivo expansion/activation with IL-2.Expanded cells should be reinfused in the early phaseafter BMT;

• sequential combination of both techniques (Figure4).126

The systemic administration of IL-2 for in vivo expan-sion and activation of the NK compartment may theo-retically have some advantages:1. high number of CE precursors in the body;2. possibility of activating CE residing in the tumor

bulk;3. more feasible and cheaper strategy than the gener-

ation and administration of LAK cells.On the other hand, the main drawbacks of this

approach are the high toxicity of systemic administra-tion of IL-2 and the high variability of anti-tumorresponse.

Sources of cytotoxic effectors andmodalities of NK cell expansion

Human NK progenitor cells can normally be found inthe BM,127,128 and originate from CD34+ hematopoieticprogenitors.129 So far, the generation of NK cells fromCD34+ precursors has been described on a small scale

103


basis but not in large scale experiments.130

High numbers of functionally active NK cells can beeasily demonstrated in mobilized cells from patientsreceiving chemotherapy plus G-CSF. Silva et al.131 haveshown a 5.4-fold expansion of NK cells from leuka-pheresis products incubated in the presence of IL-2 for6-8 days without affecting the CD34+ cell content.However, decreased function of NK cells has recentlybeen described in the PB of normal donors after G-CSFadministration.132 Circulating NK progenitors showed adecreased killing capacity and diminished proliferativeability in response to IL-2, as compared to theirunprimed BM counterpart.

Based on previously published LAK trials,123, 133 it can beestimated that about 1010-1011 activated NK cells areneeded to stimulate an anti-tumor response. A 100-foldex vivo expansion of these cells from a standard leuka-pheresis collection would, therefore be required to obtainsuch a high number of effectors.

Beaujean et al.134 reinfused autologous BM cells incu-bated for 10 days with IL-2 in 5 ALL patients followinga myeloablative treatment. This procedure resulted inan important loss of hemopoietic progenitors withdelayed engraftment. Moreover, in spite of this attemptto induce an autologous GVL, all patients eventuallyrelapsed.134

Wong et al.135 compared the ability of IL-2 alone orcombined with IL-7 or IL-12 to stimulate NK activity in

BM or PB samples. They found that IL-2/IL-12-activat-ed blood cells suppressed the growth of the leukemia cellline K-562 about eight-fold more efficiently than BMcells. They also found that cryopreservation and subse-quent stimulation of BM and PB cells did not significantlydecrease the activity of NK cells. Finally, the combina-tion of IL-2 and IL-12 showed a synergistic effect on bothBM and PB elements.135 Large-scale ex vivo expansionof NK cells for adoptive immunotherapy not onlyrequires an optimal source of precursors, but also clin-ically approved materials and procedures. In this con-text, Miller et al.136 obtained a 21-fold expansion of NKcells using a 21-day large scale NK culture performed ingas-permeable bags. Pierson et al.137 observed a 352-fold expansion of NK cells after 33 days of incubationin a bioreactor. Their starting population was NK pre-cursors enriched by negative panning with anti-CD5 andanti-CD8 antibodies. The activated NK population washighly purified (>90%) in CD56+ /CD3– cells and main-tained a powerful cytotoxicity against K-562 cells.

The use of a homogeneous NK cell fraction for cell-therapy programs seems to be advantageous becauseactivated NK cells have more specific lytic activity thanheterogeneous LAK populations.138 However, creatingsuch a fraction requires a first step of enrichment (e.g.by eliminating CD8+/CD5+ cells) and long-term culturescarry the risk of fungal or bacterial contamination.

Figure 4. Strategies for generation and activation of NK cells in vitro and in vivo. Modified from Klingemann et al., Exp Hema-tol 1993; 21:1263-70.

A

B


104


M. Aglietta et al.

Cytokine induced killer cells (CIK) expansion

In 1986 Lanier described a subset of CD3+ T cells co-expressing the CD56 antigen which is a typical NK mark-er.139 A remarkable expansion of this cellular subset hasrecently been obtained by Schmidt et al.140 following a16-day incubation in the presence of IFNg, IL-1, IL-2and a monoclonal antibody against CD3 as the mito-genic stimulus. The ability of CIK cells to depleteleukemic cells from CML marrow was then investigatedby the same group.141 While standard LAK cells were, inmost cases, unable to lyse CML cells, CIK cells were tox-ic to both autologous and allogeneic CML blasts with-out affecting normal hemopoietic progenitors.

CTL expansion for adoptive therapyIt is well known that the GVL effect can be trans-

ferred with donor-buffy-coat (BC) lymphocytes.142 Theantitumor effect of the CTL contained in the BC hasbeen shown to be more potent than that induced by NKcells, even though NK cells exert a GVL activity differ-ent from GVHD.143

In a murine model, a single dose of 23107 CTLs intumor bearing mice (DBA/2) resulted in the eradicationof primary cancer and metastases without causingsevere GVHD.144 Unfortunately, this GVL effect cannot beeasily separated from GVHD in humans. As a matter offact, in clinical studies the beneficial effect of CTLadministration is often offset by the severity of GVHD ormarrow aplasia.

Although leukemic cells share common antigens withother tissues of the host, there is also the chance thatdistinct leukemic antigens may be recognized by specificallogeneic CTL.145

Leukemia-specific T-cell clones have been isolatedfrom HLA-identical siblings146 and this finding mayexplain the high incidence of CR, without GVHD, inpatients with CML relapsed after allo-BMT and treatedwith donor BC.147

The subset of donor-lymphocytes involved in the GVLeffect is not entirely defined. Both CD4+ and CD8+ GVLeffectors have been described in animal models. Recentstudies in man suggest a prominent role of CD8+ cellsin acute leukemia and CD4+ cells in CML.148 The thera-peutic index of this approach may be increased bytreating the donor lymphocytes, previously activatedwith recipient PHA-stimulated blast, with anti-CD25ricin-conjugated antibodies.149 This procedure gives ori-gin to a CTL population which retains over 75% of itsantileukemic activity with only 10% of the initialresponsiveness against the non-leukemic cells of therecipient.

Strategies for generation and expansion ofspecific CTL

A very attractive system for generating and expand-ing CTL is based on the selection and isolation of tumor-specific peptides (e.g. those encoded by bcr-abl or PML-RARa fusion genes) and to presenting them to T-cells tostimulate a specific T-cell response. The responding T-

clones can then be amplified and selected by limitingdilution techniques. Unfortunately, in many cases thetumor-specific peptide is not presented by leukemic cellsmaking the generation of peptide-specific CTL useless.150

In this regard, the transfection in tumor cells of DNAsequences encoding for co-stimulatory molecules (B7-1)or cytokines such as GM-CSF has greatly enhanced theanti-tumor response of T-cells. Alternatively, the use ofprofessional antigen presenting cells (APC), primed withtumor specific antigens (e.g. tumor specific idiotype inlow-grade B-cell lymphoma) has proved to be effectivefor the generation of tumor-specific CTL clones in vivocapable of inducinga measurable anti-tumor response.151

Allo-CTL have also been used against EBV-related lym-phoma developed in allograft recipients and HIV patients.EBV infection is controlled in normal individuals by spe-cific CTL which lyse EBV-infected B-cells upon recogni-tion of viral peptides presented on the cell membrane inassociation with MHC class I molecules.

EBV-specific CTL have been isolated from normaldonor leucocytes and expanded ex vivo by Rooney etal.152 Following the reinfusion of 1.2×108 CTL/m2 into anallograft recipient, the complete resolution of an EBV-related immunoblastic lymphoma was observed.

Ex vivo expanded CTL can also be used to restore CMV-specific responses in immunodeficient individuals receiv-ing allogeneic BMT. Walter et al.153 treated 14 patientswith infusions of CD8+ CTL directed to CMV proteinsobtained from bone marrow donors. In this study, CMV-specific CTL were expanded by stimulation with anti-CD3antibodies coupled with autologous CMV-infectedfibroblasts in IL-2-containing culture. This approach toadoptive immunotherapy was well tolerated by the recip-ients and not associated with severe GVHD.

Future directionsPreliminary clinical data suggest that the efficacy of

donor BC infusion for the treatment of leukemic relapsecan be significantly improved by the administration of IL-2 in vivo after reinfusion and by a brief incubation of BCwith IL-2 before the reinfusion.154 The antitumor effica-cy of T-lymphocytes can also be enhanced by transfec-tion of cytokine genes or new receptors. An interestingapproach is represented by the binding of TCR to a spe-cific anti-tumor antibody Fab fragment or the use ofbispecific (anti-tumor and anti-CD3) antibodies capableof recruiting and expanding tumor-specific CTL at thetumor site.155 Recently, a significant autologous GVHDeffect has been obtained by the addition of g-IFN tocyclosporin-A in order to upregulate MHC class II mol-ecules .113 Alternatively, GM-CSF seems to enhance theanti-tumor response by stimulating professional APC(see below).156

In conclusion, there is clear evidence that CTL exerttheir cytotoxic effect through the recognition of minorHLA antigens.157 However, in some patients a very lowfrequency of specific antileukemic CTL, responsible fora GVL effect distinct from GVHD, have been isolated.Moreover, genetic approaches, such as the transduction

105


of donor lymphocytes with a suicide gene, have beenproposed to control GVHD occurring after CTL adminis-tration.158

Ex vivo generation of human dendritic APCClinical investigators are keenly interested in the role

of APC in the initiation of immune responses because ofthe potential to exploit these cells for immunotherapyof cancer and viral diseases. Pioneer studies in mammalsby Steinman et al.159 have demonstrated that the spe-cialized system of APC is constituted by BM-deriveddendritic cells (DC). DC are distinguished by their uniquepotency and ability to capture, process, and presentantigens into peptide-HLA complexes to naive T lym-phocytes and to deliver the co-stimulatory signals nec-essary for T lymphocyte activation and proliferation.

Here, we summarize the main experimental evidencesupporting the working hypothesis that individuals vac-cinated with DC expanded ex vivo and engineered topresent tumor associated antigen(s) can mount tumor-specific humoral and cellular responses. This can lead totumor regression as well as protective immunity againsttumor growth in vivo.159-162

Identification of dendritic cellsDC are leukocytes derived from hematopoietic stem

cells along the myeloid differentiation pathway (Figure5). The differentiation of DC is a stepwise process:163 orig-inating from myeloid progenitors in the BM, immature DCdistribute via blood to tissues where they have the capac-ity to take up and to process antigens. As migratory DC,they transit through the lymph or blood to lymphoidorgans, where they become mature DC, which lose anti-

gen-processing ability and acquire superior antigen-pre-senting capacity for T lymphocytes.163 In humans, DC atdifferent developmental stages circulate in PB and theyare found in virtually all tissues of the body where,depending on the location, they are referred to as inter-stitial DC (heart, kidney, gut, and lung), Langerhans cells(skin, mucous membranes), interdigitating DC (thymicmedulla, secondary lymphoid tissue); or veiled cells (lymph,blood).163 DC are regarded as distinct from mono-cytes/macrophages, although they share a common prog-enitor after the CFU-GM stage. However, this has beenquestioned as mixed colonies of dendritic cells andmacrophages are generated in vitro from single CD34+

hematopoietic progenitors more commonly163-175 thanpure DC colonies.176

DC can be distinguished from other APC by a) mor-phology; b) cell-surface membrane phenotype; and c)the strong capacity to present antigens to T cells, usu-ally assessed in the allogeneic mixed leukocyte culture(reviewed in ref. #163).

Cutaneous DC, as well as most of the DC generated exvivo from human CD34+ progenitors cells express highlevels of the surface membrane CD1a antigen (Figure 6).Although CD1a antigen can be found on cortical thy-mocytes and some B lymphocytes, its presence (noted byimmunofluorescence and flow cytometry) is the mostuseful way to quantify the ex vivo generation of DC fromearly precursors. In addition to the CD1a antigen, DCexpress peculiarly high levels of class I and class II his-tocompatibility complex structures, co-stimulatory mol-ecules for T-lymphocytes such as B7-1 (CD80) and B7-2 (CD86), and adhesion molecules such as ICAM-1(CD54) and ICAM-3 (CD50) which are involved in DC-

Figure 5. Scheme of DCdevelopment in BM, PBand thymus. Abbreviations: ly, lympho-cyte; NK, natural killercells; NSE, non-specificesterase; PPSC, pluripo-tent stem cells; CFU-DL,dendritic/Langherans cellcolony forming unit. Modi-fied from Reid CA, Br JHaematol 1997; 96:217-33.


106


dependent T-lymphocyte proliferation. DC lack mono-cyte/macrophage- and lymphocyte-lineage-restrictedantigens with the exception of the CD4 antigen.170 Asshown in Table 3, relevant co-stimulatory (B7-1 and B7-2) and adhesion molecules (ICAM-1) are expressed on allCD1a+ cells derived from CD34+ progenitors, but on few-er CD1a+ cells derived from monocytes.177,178 Morerecently, the CD83 cell surface antigen has been recog-nized as a valuable tool for detecting blood DC.179

Ex vivo expansion of dendritic cellsAlthough DC circulate in the PB and are found in vir-

tually all tissues of the body, it is difficult to obtainenough cells for ex vivo manipulation because of theirscattered locations and low number in the blood wherethey account for approximately 0.1% of all leukocytes.163

For this reason, it has been of crucial interest to knowthat: a) TNF-a cooperates with GM-CSF to generate DCfrom CD34+ hematopoietic progenitors from BM, cordblood or PB;166, 168-176 and b) IL-4 cooperates with GM-CSF in the development of DC from circulating mono-cytes.169,180 A detailed description of the methods utilizedto obtain human DC from myeloid precursors has beenrecently reported.163 However, in evaluating these meth-ods in view of a clinical trial, at least three issues shouldbe taken into consideration:a) the type of DC generated either from monocytes

(monocyte-derived DCs) or from CD34+ hematopoieticprogenitors (CD34+ derived DC);

b) the source of serum for DC growth in culture;c) the combination of cytokines required for optimal ex

vivo expansion of functional immunostimulatory DC.Monocyte-derived DC are being employed in patients

with advanced stage malignancies in phase I-II clinical tri-

als. The trials are particularly aimed at evaluating toxici-ty and immune responses after subcutaneous adminis-tration following DC pulsing ex vivo with either melanomatumor-associated peptides181-183 or with B-cell lymphomaand myeloma idiotype proteins from autologousserum.151,161 Early reports from clinical studies, in patientswith melanoma who are HLA-A1 positive and whosemalignant cells express the MAGE-1 gene, show that invivo immunization with autologous monocyte-derivedDC pulsed with MAGE-1 gene coded nonapeptide, is nottoxic and can induce peptide-specific autologousmelanoma reactive CD8+ cytotoxic T-lymphocyteresponses in situ at the vaccination site and at distanttumor sites181 as well as in PB.182 From a technical pointof view, in these studies the generation of DC from PBmononuclear cells is dependent on a culture mediumnecessarily containing GM-CSF without serum or withhuman pooled donor serum. Under these conditions theproduction of DC is quite scarce in comparison with thatachieved with fetal calf serum, as reported in early stud-ies.169 However, the presence of fetal calf serum in theculture medium induces undesired DC-mediated immuneresponses to xenogenic proteins as observed in murine184

and human preclinical studies.170 In this regard, experi-mental data suggest that CD34+ cell-derived DC can alsobe generated in the absence of serum if the culturemedium containing GM-CSF and TNF-a is supplement-ed with TGF-b1.185

Modalities for the large-scale procurement of func-tional DC from CD34+ hematopoietic progenitors inpatients with cancer have been evaluated.170 It wasfound that mobilized PB progenitors currently utilized inphase III trials186 include a fraction of CD34+ DC precur-sors which give origin ex vivo to a progeny with the char-

M. Aglietta et al.

Figure 6. Flow cytometry evaluation of cell surface phenotype of DC derived ex vivo from CD34+ progenitors on day 12 of cul-ture in the presence of GM-CSF, TNF-aa, SCF and FL.

107



acteristics of professional APC i.e., typical DC morphol-ogy and immunophenotype undistinguishable from cuta-neous Langerhans cells and DC from cord blood and BMCD34+ cells. Most importantly, these ex vivo generatedDC retained the capacity to process and present antigensto T lymphocytes as demonstrated by elicitation of HLAclass II and class I-restricted activation of CD4+ and CD8+

autologous T lymphocytes in response to xenogenic anti-gens of fetal calf serum170 or melanoma tumor-associ-ated antigen peptides,177, 178 respectively. Quantificationof progenitors of DC by limiting dilution analysis ofCD34+ cells sorted from blood cell autografts showedthat they are approximately 140-fold more numerousthan in steady-state control autograft. To obtain thisfavorable result, blood cell autografts were collected atthe time of maximal mobilization of CD34+ cells into PBas occurs after treatment with high-dose cyclophos-phamide and cytokines.

In a systematic search for culture conditions capableof ameliorating the ex vivo generation of DC dendriticcells, a variety of exogenous stimuli have been evaluat-ed as well as monocyte-derived versus CD34+ cell-derivedDC.170,177,178 In this respect, it has been shown that GM-CSF plus TNF-a-dependent generation of DC from mobi-lized CD34+ cells is 2.5 fold enhanced by either FL or SCF,and 5-fold enhanced by a combination of these growthfactors. In addition, autologous high-dose chemothera-py recovery phase serum rather than fetal calf serum orhuman donor pooled AB serum has been shown to be theoptimal serum for the generation of DC. Regardless ofthe precise mechanism of action of FL and SCF in asso-ciation with GM-CSF and TNF-a on the enhancement ofDC differentiation and proliferation, these findings haveprovided new advantageous tools for the large-scale gen-eration of DC from mobilized CD34+ cells in patientsundergoing cancer treatment. In fact, the stimulation ofCD34+ cells from an average blood cell autograft shouldpermit the generation of a median of 0.63109/kg DC from

an average 65 kg individual, i.e., almost 403109 DC. Incontrast, differentiation of DC from monocytes in thepresence of autologous high-dose chemotherapy recov-ery phase serum plus GM-CSF and IL-4 is not associatedwith a comparably high outgrowth of DC.177,178 Theseobservations, together with the weaker expression of co-stimulatory molecules in monocyte-derived DC in com-parison with CD34+ cell-derived DC187 may favor the uti-lization of the latter source of APC for the developmentof active immunization programs involving DC in humans.The comparative efficiency as APC of DC derived frommonocytes versus CD34+ hematopoietic progenitors hasrecently been studied with DC isolated from blood ofpatients with melanoma. In particular, it has been shownthat DC derived from G-CSF-mobilized CD34+ cells aremore efficient than those derived from monocytes ininducing melanoma tumor-associated antigen peptidespecific activation of autologous CD8+ cytotoxic T-lym-phocytes. Interestingly, in the same experiments the lat-ter cells were also capable of lysing a panel of melanomacell lines sharing the same HLA class I alleles with thepatients from whom CD8+ cytotoxic T-lymphocytes weregenerated with tumor-associated antigen peptide pulsedautologous DC.177,178 Moreover, CD34+ cells mobilized intoPB by G-CSF were shown to be capable of generating ahigher number of mature and fully functional DC thantheir BM counterparts.188 However, it should be pointedout that, at present, clinical studies utilizing DC as vehi-cles for anti-tumor vaccination have been carried outwith monocyte-derived DC either freshly isolated fromPB152 or cultured ex vivo.161, 181-183 In addition, culture con-ditions which allow the large scale production of termi-nally differentiated and fully functional monocyte-derived DC have recently been described.189

Dendritic cells for antitumor cell therapyAn extensive review on the clinical use of DC is beyond

the scope of this paper, however, a few remarks in thisregard are needed.

The goal of vaccination is the induction of protectiveimmunity. Originally, vaccinations were used in the set-ting of infectious diseases, but are now expanding toinclude the treatment of allergy, autoimmune diseases,and tumors. A rational approach to vaccination involves3 steps: a) the identification of the protective effectormechanisms, b) the choice of an antigen that can inducea response in all individuals, and c) the use of an appro-priate way to deliver the vaccine to induce the propertype of response.159-162

It has now been demonstrated that certain tumor cellsare antigenic by expressing tumor-associated antigensthat can be recognized by T lymphocytes in a syngeneichost. However, they are often poorly immunogenic, atleast in part because they lack the cellular armamentar-ium for specific T-lymphocyte recognition, activation,and co-stimulation typical of APC especially DC.190,191

Different mechanisms may account for the ability oftumor cells to evade immune responses. Tumor cells may

Table 3. Phenotype of CD1a+ dendritic cells generated frommobilized CD34+ progenitors or from monocytes.

CD1a+ cells derived fromAntigens CD34+ progenitors Monocytes

CD14 2% 3%CD80 100% 84%CD86 100% 67%CD54 100% 67%

MHC Class IHLA-A0201 100% 100%

MHC Class IIHLA-DR 100% 100%HLADQ 60% 89%

Modified from Mortarini et al., Cancer Res, in press.

108


display low immunogenicity through low MHC and/ortumor-speciifc antigen expression or may downregulateFas and constitutively express Fas-Ligand, which binds toFas on cytotoxic T-lymphocytes, resulting in apoptosis ofthe latter.192 Furthermore, it has recently been shownthat vascular endothelial growth factor produced bytumors inhibits the functional maturation of CD34+ cell-derived DC.193 When considering the use of the uniqueantigen-presenting capacity of DC to prime specific anti-tumor T-lymphocytes, this phenomenon should be tak-en into account, as it may result in poor recovery andfunction of DC directly recovered from the blood of can-cer patients. In contrast, dendritic APC expanded ex vivoin the presence of cytokines and in the absence ofinhibitory factors released by tumors would probably befunctional.193-195 However, it should also be consideredthat published data demonstrate that whereas in vivoadministration of DC loaded with low doses of tumorantigen enhances antitumor immunity, DC pulsed withhigh doses of antigen or high numbers of tumor antigen-pulsed DC may inhibit development of immunity.196 Thisfinding supports the notion that stimulation of DC-medi-ated antigen presentation in vivo may act in a tolero-genic or immunogenic fashion depending on a variety ofpartially understood factors.197

Basically, there are at least two approaches to tumorvaccination (Figure 7). The first is to identify a tumor-associated antigen to be used as a vaccine, the secondis to increase the immunogenicity of tumor cells and letthe immune system decide which antigen to target.Indeed, in experimental models with the appropriatemanipulation exploiting the physiologic function ofantigen-presenting DC, the immune system can beinduced to mount responses that can kill tumor cellsand also protect animals from subsequent challengeeven with a poorly immunogenic tumor.185,198-203

Given the richness of recently identified tumor asso-ciated antigens and their corresponding peptide epi-topes recognized by MHC-restricted CD8+ or CD4+ Tlymphocytes (Table 4), investigators are currently eval-uating the clinical efficacy of specific tumor-associat-ed antigen-based vaccines for the treatment of variousmalignancies. Recently, in a cooperative clinical trial itwas observed that partial tumor regressions can occurin HLA-A1+ patients with melanoma treated with anaked MAGE 3 peptide epitope vaccine even in theabsence of any engineering of antigen-presenting cellsor adjuvant cytokine(s).204 This clinical evidence inducesthe belief that the effectiveness of peptide-based vac-cines is likely to benefit further from administration ofappropriate cytokines156,205,206 or cellular adjuvants (e.g.DC) capable of promoting cellular immunity.

Among hematologic malignancies, CML is beingintensively evaluated as a possible target of dendriticAPC-based immunotherapy. It is well known that CMLis characterized by a specific translocation of the c-abloncogene (9q34) to the bcr region on chromosome 22(22q11). Alternative recombination sites involving eitherthe second or third exon of the bcr gene splicing to exon

2 of the abl gene yield two potential fusion gene tran-scripts, b2a2 and b3a2, respectively. The translated 210-kd bcr-abl fusion protein, which has abnormal tyrosinekinase activity, includes a new potentially antigenicsequence at the fusion site: a new amino acid is gener-ated at the junctional site by the fusion event; in theb2a2 fusion a glutamic acid (E) is encoded, whereas inthe b3a2 recombination event a lysine (K) is generated.Interestingly, a bcr-abl peptide from the b3a2 fusionregion has been found to be immunogenic in mice.207 Inhumans, binding of b3a2 peptides to various HLA classI alleles208 and priming of CD8+ cytotoxic T-lymphocytesin vitro has been described although the capacity ofthese peptide-specific CD8+ T-lymphocytes to lyse CMLcells has not been determined.209 In contrast, in a recentstudy it has been demonstrated that CD4+ T-lympho-cytes can be identified that proliferate in an HLA class-II restricted manner in response to a 11mer(GFKQSSKALQR) b3a2 peptide especially when the lat-ter is presented by purified CMRF-44+ blood DC.210 In thesame study the peptide-specific CD4+ T-lymphocyteswere able to respond to the whole protein in crudeextract from CML cells. Intriguingly, dendritic antigen-presenting cells in CML patients can be derived fromthe malignant clone and these malignant dendritic cellscan induce antileukemic reactivity in autologous T lym-phocytes without the necessity of additional exoge-nous antigens.211

Although, the above observations cannot be extra-polated a priori to other malignancies carrying specifictranslocations and corresponding fusion genes andproducts,212 further investigations on the possible clin-

M. Aglietta et al.

Table 4. Tumor antigens capable of eliciting T-lymphocyteresponses.

Activated oncogene productsMutated: Position 12 point mutation of 21ras

Rearranged: bcr-abl (b3a2 peptide)Overexpressed: HER-2/neu

Tumor suppressor gene productsp53 mutations

Reactivated embryonic gene productsMAGE family (at least 12 genes)BAGEGAGE

Melanocyte differentiation antigensTyrosinase proteinMelan-A/MART1gp100gp75

Viral gene productsHuman papilloma virus antigens (E6, E7)EBV EBNA-1 gene products

Idiotype epitopesIg and TCR hypervariable regions

109


ical application of bcr-abl peptide(s) presented by autol-ogous dendritic antigen-presenting cells are warranted.

The translation into the clinical setting of the aboveexperimental hints favoring the use of DC pulsed ex vivowith synthetic tumor-associated antigen peptides foreffective cell therapy in humans is likely to be hamperedby: a) the limited availability of patients with HLA typ-ing compatible for the utilized tumor-associated anti-gen peptide(s); b) the occurrence of independent mech-anisms of tumor escape in vivo such as loss of expres-sion of tumor-associated antigens or of HLA class I dur-ing tumor progression; and c) the short duration of theimmune responses thus requiring annoying boost vac-cinations. It has been suggested213 that these limita-tions may be overcome by transduction of genes encod-ing relevant proteins into DC or their progenitors so thatDC could tailor peptides to their own HLA moleculesthus obviating the need to synthesize tumor-specificpeptides most of which have stringent HLA restrictions.A further advantage of the transduction approach maybe the stable long-term expression of the antigen bythe DC, which would allow its presentation to theimmune system for longer periods without the concernsabout the turnover of preformed peptide/HLA complex-es in vivo after immunization.214

Based on the above experimental body of evidence, apioneer clinical trial has evaluated the ability of autolo-gous monocyte-derived DC pulsed ex vivo with nonHodgkin’s lymphoma-specific idiotype protein to stimu-late host immunity when infused as a vaccine.151 In thisstudy active immunotherapy of patients with B-cell lym-phoma against idiotypic determinants led to anti-tumorimmunity that correlated with improved clinical out-come in some patients.

Regardless of the type of cell manipulation (Figure 7)(ex vivo pulse with tumor-associated antigen peptidesversus transduction with tumor associated antigengenes versus immunization with fusions of dendritic andcarcinoma cells215) that will be successful in clinical

applications the necessity of methods of generatinglarge numbers of functional DC is implicit for the evo-lution of such studies.

Future directionsSince DC have been shown to be intimately involved

in the generation of CD4+ and CD8+ T-lymphocyte medi-ated tumor-specific immunity, it is attractive to specu-late that vaccination with DC pulsed or engineered exvivo to present tumor antigen(s) may be effective in gen-erating tumor immunity in vivo. Among the recentlyprospected sources of DC, namely BM, neonatal cordblood, and adult PB, the last is certainly the richest andmost accessible in all patients with cancer, although itremains to be confirmed whether functional differenceswill favor the utilization of monocyte- versus CD34+ cell-derived DC. Thus, in the clinical setting of adoptiveimmunotherapy for patients with malignancies, a ther-apeutic protocol could be envisioned involving the mobi-lization of CD34+ cells into PB with hematopoieticgrowth factors, with or without prior intensive chemo-therapy. Thus, enrichment of hematopoietic progenitorcells could be followed by ex vivo generation of DCpulsed or engineered to present tumor antigen(s), to bereinfused as a potential secondary tumor-specific immu-notherapy or vaccination. It remains to be establishedwhether the latter effect could be further enhanced byin vivo adjuvant cytokines such as GM-CSF and/or FL, asoccurs in murine models.

A potential advantage of ex vivo immune cell thera-py over direct in vivo immune intervention, is the lackof functional inhibition that may occur in vivo. Thishypothesis is based on a recently proposed mechanismof tumor escape/resistance from the host immune sys-tem, in which cancer cells produce a vascular endothe-lial growth factor that impairs antigen presentationrequired to induce specific antitumor immune respons-es in vivo.193


Figure 7. Sources of tumor anti-gens for DC-based cancer vaccines.

110


Contributions and AcknowledgmentsAll the authors equally contributed to the manuscript

and they are listed in alphabetical order.

DisclosuresConflict of interest. This review article was prepared by

a group of experts designated by Haematologica and byrepresentatives of two pharmaceutical companies, AmgenItalia SpA and Dompé Biotec SpA, both from Milan, Italy.This co-operation between a medical journal and phar-maceutical companies is based on the common aim ofachieving optimal use of new therapeutic procedures inmedical practice. In agreement with the Journal’s Con-flict of Interest policy, the reader is given the followinginformation. The preparation of this manuscript was sup-ported by educational grants from the two companies.Dompé Biotec SpA sells G-CSF and rHuEpo in Italy, andAmgen Italia SpA has a stake in Dompé Biotec SpA.

Manuscript processingManuscript received March 11, 1998; accepted June

10, 1998.

References

1. Scott MA, Gordon MY. In search of the haemopoieticstem cell. Br J Haematol 1995; 90:738-43.

2. Morrison SJ, Shah NM, Anderson DJ. Regulatorymechanisms in stem cell biology. Cell 1997; 88:287-98.

3. Metcalf D. The molecular control of cell division, dif-ferentiation commitment and maturation in haemo-poietic cells. Nature 1989; 339:27-30.

4. Lajtha LG. Haemopoietic stem cells. Br J Haematol1975; 29:529-35.

5. McCulloch EA. Control of hematopoiesis at the cellu-lar level. In: Gordon AS, ed. Regulation of hemato-poiesis (vol. 1). New York: Appleton, 1970. p 133-54.

6. Gordon MY. Physiological mechanisms in BMT andhaemopoiesis - revisited. Bone Marrow Transplant1993; 11:193-7.

7. Potten CS, Loeffler M. Stem cells: attributes, cycles,spirals, pitfalls and uncertainties. Lessons for and fromthe crypt. Development 1990; 110:1001-20.

8. Verdi JM, Schmandt R, Bashirullah A, et al. Mam-malian numb is an evolutionary conserved signalingadapter protein that specifies cell fate. Curr Biol 1996;6:1134-45.

9. Zhong WM, Feder JN, Jiang MM, Jan LY, Jan YN. Asym-metric localization of a mammalian numb homologduring mouse cortical neurogenesis. Neuron 1996;17:43-53.

10. Mayani H, Dragowska W, Lansdorp PM. Lineage com-mitment in human hemopoiesis involves asymmetriccell division of multipotent progenitors and does notappear to be influenced by cytokines. J Cell Physiol1993; 157:576-9.

11. Till JE, McCulloch EA, Siminovitch L. A stochasticmodel of stem cell proliferation, based on the growthof spleen colony-forming cells. Proc Natl Acad Sci USA1964; 51:29-33.

12. Trentin JJ. Influence of hematopoietic organ stroma(hematopoietic inductive microenvironments) on stemcell differentiation. In: Gordon AS. ed. Regulation ofHematopoiesis. vol. 1. New York: Appleton, 1970. p.

161-85.13. Ogawa M. Differentiation and proliferation of

hematopoietic stem cells. Blood 1993; 81:2844-53.14. Metcalf D. Hematopoietic regulators: redundancy or

subtlety? Blood 1993; 82:3515-23.15. Ellisen LW, Bird J, West DC, Soreng AL, Reynolds TC,

Smith SD, Sklar J. TAN-1, the human homologue ofthe Drosophila notch gene, is broken by chromoso-mal translocations in T lymphoblastic neoplasms. Cell1991; 66:649-61.

16. Seydoux G, Mello CC, Pettitt J, Wood WB, Priess JR,Fire A. Repression of gene expression in the embryon-ic germ lineage of C. elegans. Nature 1996; 382:713-6.

17. Metcalf D. Lineage commitment of hemopoietic prog-enitor cells in developing blast cell colonies: influenceof colony-stimulating factors. Proc Natl Acad Sci USA1991; 88:11310-4.

18. Pawson T. Protein modules and signalling networks.Nature 1995; 373:573-80

19. Bedi A, Sharkis SJ. Mechanisms of cell commitment inmyeloid cell differentiation. Curr Opin Hematol 1995;2:12-21.

20. Williamson EA, Boswell SH. Signal transduction dur-ing myeloid cell differentiation. Curr Opin Hematol1995; 2:29-40.

21. Sauvageau G, Thorsteinsdottir U, Eaves CJ, et al. Over-expression of HOXB4 in hematopoietic cells causes theselective expansion of more primitive populations invitro and in vivo. Gens Dev 1995; 9:1753-65.

22. Sauvageau G, Thorsteinsdottir U, Hough MR, et al.Overexpression of HOXB3 in hematopoietic cells caus-es defective lymphoid development and progressivemyeloproliferation. Immunity 1997; 6:13-22.

23. Wessely O, Deiner E-M, Beug H, von Lindern M. Thegluocorticoid receptor is a key regulator of the decisionbetween self-renewal and differentiation in erythroidprogenitors. EMBO J 1997; 16:267-80.

24. Dotti GP, Carlo-Stella C, Spinelli O, et al. Shc overex-pression induces selective hypersensitivity to GM-CSFand prevents apoptosis of the GM-CSF-dependentacute myelogenous leukemia cell line GF-D8. In:Hematology and blood transfusion. New York: Sprin-ger-Verlag, 1997. In press.

25. Rossi F, McNagny KM, Smith G, Frampton J, Graf T.Lineage commitment of transformed haematopoieticprogenitors is determined by the level of PKC activity.EMBO J 1996; 15:1894-901.

26. Vaziri H, Dragowska W, Allsopp RC, Thomas TE, HarleyCB, Lansdorp PM. Evidence for a mitotic clock inhuman hematopoietic stem cells: loss of telomeric DNAwith age. Proc Natl Acad Sci USA 1994; 91:9857-60.

27. Morrison SJ, Uchida N, Weissman IL. The biology ofhematopoietic stem cells. Annu Rev Cell Dev Biol1994; 11:35-71.

28. Carlo-Stella C, Cazzola M, De Fabritiis P, et al. CD34-positive cells: biology and clinical relevance. Haema-tologica 1995; 80:367-87.

29. Krause DS, Fackler MJ, Civin CI, Stratford May W.CD34: structure, biology, and clinical utility. Blood1996; 87:1-13.

30. Civin CI, Gore SD. Antigenic analysis of hematopoiesis:a review. J Hematother 1993; 2:137-44.

31. Baumhueter S, Singer MS, Henzel W, et al. Binding ofL-selectin to the vascular sialomucin CD34. Science1993; 262:436-41.

32. Gunji Y, Nakamura M, Osawa H, et al. Human prim-itive hematopoietic progenitor cells are more enrichedin KITlow cells than in KIThigh cells. Blood 1993;82:3283-9.

33. Small D, Levenstein M, Kim E, et al. STK-1, the humanhomologue of Flk-2/Flt-3, is selectively expressed in

M. Aglietta et al.

111


CD34+ human bone marrow cells and is involved inthe proliferation of early progenitor/stem cells. ProcNatl Acad Sci USA 1994; 91:459-63.

34. Zanjani ED, Almeida-Porada G, Leary AG, Ogawa M.Human bone marrow CD34- cells engraft in vivo andundergo multilineage expression including giving rise toCD34+ cells. Blood 1997; 90:252a.

35. Larochelle A, Vormoor J, Hanenberg H, et al. Identifi-cation of primitive hematopoietic cells capable ofrepopulating NOD/SCID mice: implications for genetherapy. Nature Med 1996; 2:1329-37.

36. Bhatia M, Wang JCY, Kapp U, Bonnet D, Dick JE.Purification of primitive human hematopoietic cellscapable of repopulating immunodeficient mice. ProcNatl Acad Sci USA 1997; 94:5320-5.

37. Eaves CJ, Cashman JD, Eaves AC. Methodology oflong-term culture of human hematopoietic cells. J TissCult Methods 1991; 13:55-62.

38. Sutherland HJ, Eaves CJ, Eaves AJ, Dragowskas W,Lansdorp PM. Characterization and partial purifica-tion of human marrow cells capable of initiating long-term hematopoiesis in vitro. Blood 1989; 74:1563-70.

39. Petzer AL, Hogge DE, Lansdorp PM, Reid DS, Eaves CJ.Self-renewal of primitive human hematopoietic cells(long-term-culture-initiating cells) in vitro and theirexpansion in defined medium. Proc Natl Acad Sci USA1996; 93:1470-4.

40. Breems DA, Blokland EAW, Neben S, Ploemacher RE.Frequency analysis of human primitive haematopoiet-ic stem cell subsets using a cobblestone area formingcell assay. Leukemia 1994; 8:1095-104.

41. Gordon MY, Amos TAS. Stochastic effects in hemo-poiesis. Stem Cells 1994; 12:175-9.

42. Briddell RA, Kern BP, Zilm KL, Stoney GB, McNieceIK. Purification of CD34+ cells is essential for optimalex vivo expansion of umbilical cord blood cells. JHematother 1997; 6:145-50.

43. Haylock D, Simmons P, To LB, Juttner C. Growth fac-tors and ex vivo expansion of hemopoietic progenitorcells. In: Morstyn G, Sheridan W. eds. Cell Therapy.Cambridge: Cambridge University Press, 1996. p. 221-37.

44. Alcorn MJ, Holyoake TL, Richmond L, et al. CD34-positive cells isolated from cryopreserved peripheral-blood progenitor cells can be expanded ex vivo andused for transplantation with little or no toxicity. J CIinOncol 1996; 14:1839-47.

45. Brugger W, Heimfeld S, Berenson RJ, Mertelsmann R,Kanz L. Reconstitution of hematopoiesis after high-dose chemotherapy by autologous progenitor cellsgenerated ex-vivo. N Engl J Med 1995; 333:283-7.

46. Lemoli RM, Fortuna A, Motta MR, et al. Concomitantmobilization of plasma cells and hematopoietic prog-enitors into peripheral blood of multiple myelomapatients: Positive selection and transplantation ofenriched CD34+ cells to remove circulating tumorcells. Blood 1996; 87:1625-32.

47. Gazitt Y, Reading C, Hoffman R, et al. Purified CD34+Lin- Thy+ stem cells do not contain clonal myelomacells. Blood 1995; 86:381-9.

48. Maguer-Satta V, Petzer AL, Eaves AC, Eaves CJ. BCR-ABL expression in different subpopulations of func-tionally characterized Ph+ CD34+ cells from patientswith chronic myeloid leukemia. Blood 1996; 88:1796-804.

49. Fogli M, Amabile M, Martinelli G, et al. Selectiveexpansion of normal haemopoietic progenitors fromchronic myelogenous leukaemia marrow. Br J Haema-tol 1998; 101:119-29.

50. Peters D, Branscomb E, Dean P, et al. The LLNL high-speed sorter: design, features, operational character-istics and biological utility. Cytometry 1985; 6:290-

301.51. Tsukamoto A, Sasaki D, Chen BP, Hoffman R. Char-

acterization and isolation of mobilized peripheralblood stem cells using a high-speed cell sorter. In:Morstyn G, Sheridan W, eds. Cell Therapy. Cambridge:Cambridge University Press, 1996. p. 183-98.

52. Okarma T, Lebkowski J, Schain L, et al. The AIS collec-tor: a new technique for stem cell purification. In: Wor-thington-White DA, Gee AP, Gross S, eds. Advances inbone marrow purging and processing. New York: WileyLiss, 1992. p. 449-59.

53. Kemshead JT, Ugelstad J. Magnetic separation tech-niques: their application to medicine. Mol Cell Bio-chem 1985; 67:11-8.

54. Kvalheim G, Sorensen O, Fodstad O, et al. Immuno-magnetic removal of B-lymphoma cells from humanbone marrow: A procedure for clinical use. Bone Mar-row Transpl 1988; 3:31-41.

55. Civin CI, Strauss LC, Facker MJ, Trismann TM, WileyJM, Loken RL. Positive stem cell selection. Basic sci-ence. In: Gross S, Gee A, Worthington-White DA, eds.Bone marrow purging and processing. New York.Wiley-Liss, 1990. p. 387-402.

56. Milteny S, Mueller W, Weichel W, Radbruch A. Highgradient magnetic cell separation with MACS. Cytom-etry 1990; 11:231-8.

57. Lemoli RM, Tafuri A, Fortuna A, et al. Cycling statusof CD34+ cells mobilized into peripheral blood ofhealthy donors by recombinant human granulocytecolony-stimulating factor. Blood 1997; 89:1189-96.


59. Koller MR, Emerson SG, Palsson BO. Large scaleexpansion of human stem and progenitor cells frombone marrow mononuclear cells in continuous perfu-sion cultures. Blood 1993; 82:378-84.

60. Reisner Y, Martelli MF. Bone marrow transplantationacross HLA barriers by increasing the number of trans-planted cells. Immunol Today 1995; 16:437-40.

61. Moore MAS: Expansion of myeloid stem cells in cul-ture. Semin Hematol 1995; 32:183-92.

62. Emerson GE. Ex vivo expansion of hematopoietic pre-cursors, progenitors and stem cells: the next generationof cellular therapeutics. Blood 1996; 87:3082-8.

63. Yonemura Y, Ku H, Hirayama F, Souza LM, Ogawa M.Interleukin 3 or interleukin 1 abrogates the reconsti-tuting ability of hematopoietic stem cells. Proc NatlAcad Sci USA 1995; 93:4040-4.

64. Peters SO, Kittler ELW, Ramshaw HS, Quesenberry PJ.Ex vivo expansion of murine marrow cells with IL-3, IL-6, IL-11 and SCF leads to impaired engraftment in irra-diated hosts. Blood 1996; 87:30-7.

65. Traycoff CM, Cornetta K, Yoder MC, Davidson A,Srour EF. Ex vivo expansion of murine hematopoieticprogenitor cells generates classes of cells possessing,different leveIs of bone marrow repopulating poten-tial. Exp Hematol 1990; 24:299-306.

66. Young JC, Varma A, DiGiusto D, Backer MP. Retentionof quiescent hematopoietic cells with high proliferativepotential during ex vivo stem cell culture. Blood 1996;87:545-56.

67. Di Giusto DL, Lee R, Moon K, et al. Hematopoieticpotential of cryopreserved and ex vivo manipulatedumbilical cord blood progenitor cells evaluated in vit-ro and in vivo. Blood 1996; 87:1261-71.

68. Gabutti V, Foà R, Mussa F, Aglietta M. Behaviour ofhuman haematopoietic stem cells in cord and neona-tal blood. Haematologica 1975; 60:4.

69. Lansdorp PM, Dragowska W, Mayani H. Ontogeny-related changes in proliferative potential of humanhematopoietic celIs. J Exp Med 1993; 178:787-91.


112


70. Zandstra PW, Eaves CJ, Piret JM. Expansion ofhematopoietic progenitor cell populations in stirredsuspension bioreactors of normal human bone mar-row cells. Biotechnology 1994; 12:909-14.

71. Bertolini F, Battaglia M, Corsini C, et al. Engineeredstromal layers and continuous flow culture enhancemultidrug resistance gene transfer in hematopoieticprogenitors. Cancer Res 1996; 56:2566-72.

72. Keller JR, Bartelmez SH, Sitnicka E, et al. Distinct andoverlapping direct effects of macrophage inflammato-ry protein 1a and transforming growth factor-b onhematopoietic progenitor/stem cell growth. Blood1994; 84:2175-81.

73. Van Reenst PCF, Snoeck HW, Lardon F, et al. TGF-band MIP-1a extert their main inhibitory activity on veryprimitive CD34++CD38– cells but show opposite effectson more mature CD34+CD38+ human progenitors.Exp Hematol 1997; 24:1509-15.

74. Matthews W, Jordan CT, Wiegand GW, Pardoll D,Lemischka IR. A receptor tyrosine kinase specific tohemopoietic stem cell-enriched populations. Cell1991; 65:1143-52.

75. Shah AJ, Smogorzewska EM, Hannum C, Crooks GM.Flt3 ligand induces proliferation of quiescent humanbone marrow CD34+CD38- cells and maintains prog-enitor cells in vitro. Blood 1996; 87:3563-70.

76. Piacibello W, Fubini L, Sanavio F, et al. Effect ofhuman FLT3 ligand on myeloid leukemia cells growth:heterogeneity in response and synergy with otherhematopoietic growth factors. Blood 1995; 86:4105-14.

77. Piacibello W, Garetto L, Sanavio F, et al. The effects ofhuman FLT3 ligand on in vitro human megakaryocy-topoiesis. Exp Hematol 1996; 24:340-6.

78. Bertolini F, Battaglia M, Pedrazzoli P, et al. Megakary-ocytic progenitors can be generated ex vivo and safelyadministered to autologous peripheral blood progen-itor cell transplant recipients. Blood 1997; 89:2679-88.

79. Petzer Al, Zandstra PW, Piret JM, Eaves CJ. Differentialcytokine effects on primitive (CD34+CD38-) humanhematopoietic cells: Novel responses to Flt3-ligandand thrombopoietin. J Exp Med 1996; 183:2551-8.

80. Yonemura Y, Ku H, Lyman SD, Ogawa M. In vitroexpansion of hematopoietic progenitors and mainte-nance ot stem cells: comparison between FLT3/FLK-2ligand and Kit ligand. Blood 1997; 89:1915-21.

81. Moore TA, Zlotnik A. Differential effects of Flk2/Flt-3ligand and stem cell factor on murine thymic progen-itor cells. J Immunol 1997; 158:4187-92.

82. Dao MA, Hannum CH, Kohn DB, Nolta JA. FLT3 lig-and preserves the ability of human CD34+ progenitorsto sustain long-term hematopoiesis in immune-defi-cient mice after ex vivo retroviral mediated transduc-tion. Blood 1997; 89:446-56.

83. Zandstra PW, Conneally E, Petzer AL, Piret JM, EavesCJ. Cytokine manipulation of primitive human hema-topoietic cell self-renewal. Proc Natl Acad Sci USA1997; 94:4698-703.

84. Kaushansky K. Thrombopoietin: The primary regulatorof platelet production. Blood 1995; 86:419-31.

85. Kaushanski K, Broudy VC, Lin N, et al. Thrombopoi-etin, the mpl ligand, is essential for full megakaryocytedevelopment. Proc Natl Acad Sci USA 1995; 92:3234-38.

86. Goncalves F, Lacout O, Villeval JL, Wendling F,Vainchenker W, Dumenil D. Thrombopoietin does notinduce lineage-restricted commitment of Mpl-Rexpressing pluripotent progenitors but permits theircomplete erythroid and megakaryocytic differentia-tion. Blood 1997; 89:3544-53.

87. Ramsfjell V, Borge OJ, Cui L, Jacobsen SEW. Throm-

bopoietin directly and potently stimulates multilineagegrowth and progenitor cell expansion from primitive(CD34+CD38-) human bone marrow progenitor cells.J Immunol 1997; 158:5169-77.

88. Ku H, Hirayama F, Kato I, et al. Soluble thrombopoi-etin receptor and granulocyte colony-stimulating fac-tor receptor directly stimulate proliferation of primitivehematopoietic progenitors of mice in synergy with steelfactor or the ligand for FIt3/Flk2. Blood 1996; 88:4124-31.

89. Piacibello W, Sanavio F, Garetto L, et al. Extensiveamplification and self-renewal of human primitivehematopoietic stem cells from cord blood. Blood1997; 89:2644-53.

90. Uchida N, Aguila HL, Fleming WH, Jerahek I, Weiss-man IL. Rapid and sustained hematopoietic recover inIethally irradiated mice transplanted with purifiedThy1.Low, Lin-Sca-1+ hematopoietic stem cells. Blood1994; 83:3758-79.

91. Szilvassy SJ, et al. Partially differentiated ex vivo expand-ed celIs accelerate hematologic recovery in myeloab-lated mice transplanted with highly enriched long-termrepopulating stem cells. Blood 1996; 88:3642-53.

92. Williams SF, Lee WJ, Bender JG, et al. Selection andexpansion of peripheral blood CD34+ cells in autolo-gous stem cell transplantation for breast cancer. Blood1996; 87:1687-91.

93. Widmer L, Pichert G, Jost LM, Stahel RA. Fate of con-taminating t(14,18)+ lymphoma cells during ex vivoexpansion of CD34-selected hematopoietic progeritorcells. Blood 1996; 88:3166-75.

94. Vogel W, Behringer D, Scheding S, Kanz L , Brugger W.Ex vivo expansion of CD34+ peripheral blood progen-itor cells: Implications for the expansion of contami-nating epithelial tumor cells. Blood 1996; 88:2707-13.

95. Whiteside TL, Rinaldo CR jr, Herberman RB. Cytolyt-ic cell functions. On "Manual of Clinical LaboratoryImmunology". 4th ed. Am Soc Microbiol WashingtonD.C., 1992. p. 220-30.

96. Torpey DJ III, Lindsley MD, Rinaldo CR Jr. HLA-restrict-ed lysis of herpes simplex virus-infected monocytes andmacrophages mediated by CD4+ and CD8+ T lym-phocytes. J Immunol 1989; 142:1325-32.

97. Kagi D, Vignaux F, Ledermann B, et al. Fas and perforinpathways as major mechanisms of T cell-mediatedcytotoxicity. Science 1994, 265:528-30.

98. Montel AH, Bochan MR, Hobbs JA, Lynch DH, Brah-mi Z. Fas involvement in cytotoxicity mediated byhuman NK cells. Cell Immunol 1995, 166:236-46.

99. Braun MY, Lowin B, French L, Acha Orbea H, TschoppJ. Cytotoxic T cell deficient in both functional Fas lig-and and perforin show residual cytolytic activity yetlose their capacity to induce lethal acute graft versushost disease. J Exp Med 1996, 183:657-61.

100. Cascino I, Papoff G, De Maria R, Testi R, Ruberti G.Fas/Apo-1 (CD-95) receptor lacking the intracyto-plasmic signaling domain protects tumor cells fromFas-mediated apoptosis. J Immunol 1996; 156:13-7.

101. Lanier LL, Phillips JH. Inhibitory MHC class I receptorson NK cells and T cells. Immunol Today 1996; 17:86-92.

102. Kanegane H, Tosato G. Activation of naive and mem-ory T cells by Interleukin-15. Blood 1996; 88:230-5.

103. Richard C, Alsar MJ, Calavia J, et al. Recombinanthuman GM-CSF enhances T cell mediated cytotoxicfunction after ABMT for hematological malignancies.Bone Marrow Trasplant, 1993; 11:473-8.

104. Bendall LJ, Kortlepel, Gottlieb DJ. GM-CSF enhancesIL-2-activated natural killer cell lysis of clonogenic AMLcells by upregulating target cell expression of ICAM-1.Leukemia 1995; 9:677-84.

M. Aglietta et al.

113


105. Cantori I, Olivieri A, Rupoli S, et al. The association ofIL-2 plus GM-CSF has protective effects and reducesthe apoptosis in liquid culture. Blood 1996; 88:112a.

106. Trinchieri G. Interleukin-12: a cytokine produced byantigen presenting cells with immunoregulatory func-tions in the generation of T-helper cells type 1 andcytotoxic lymphocytes. Blood 1994, 84:4008-27.

107. Gerosa F, Paganin C, Peritt D, et al. Intereleukin-12primes human CD4 and CD 8 T cell clones for highproduction of both interferon-g and interleukin-10. JExp Med 1996, 183:2559-69.

108. Abdul-Hai A, Or R, Slavis S, et al. Stimulation ofimmune reconstitution by interleukin-7 after syngene-ic bone marrow transplantation in mice. Exp Hema-tol, 1996; 24:1416-22.

109. Mavroudis D, Barret J. The graft-versus-leukemiaeffect. Current Opinion Hematol 1996: 3:423-29.

110. Glass B, Uharek L, Zeis M, et al. Graft-versus-leukemiaactivity can be predicted by natural cytotoxicity againstleukemia cells. Br J Haematol 1996; 93:412-20.

111. Goldman JM., Gale RP., Horowitz MM, et al. Bonemarrow transplantation for chronic myelogenousleukemia in chronic phase: increased risk for relapseassociated with T cell depletion. Ann Intern Med 1988;108:806-14.

112. Kolb HJ, Schattenberg A, Goldman JM, et al. Graft-versus-leukemia effect of donor lymphocyte transfu-sion in marrow grafted patients. Blood 1995; 86:2041-50.

113. Hess AD, Bright EC, Thoburn C, Vogelsang GB, JonesRJ, Kennedy MJ. Specificity of effector T lymphocytesin autologous graft-versus-host disease: Role of themajor histocompatibility complex class II invariantchain peptide. Blood, 1997; 89:2203-9.

114. Hokland M, Jacobsen N, Ellegaard J, Hokland P. Nat-ural killer function following allogeneic bone marrowtransplantation. Transplantation 1988; 45:1080-4.

115. Bengtsson M, Totterman TH, Smedmyr B, Festin R,Simonseen B. Regeneration of functional and activat-ed NK and T subset cells in the marrow and bloodafter autologous bone marrow transplantation. Aprospective phenotypic study with 2/3-color FACSanalysis. Leukemia 1989; 3:68-75.

116. Jorgensen H, Hokland P, Jensen T, Basse P, HoklandM. Natural killer cell in peripheral blood after autolo-gous bone marrow transplantation. Nature Immun1995; 14:164-72.

117. Timonen T, Maenpaa A, Helander T, Malygin A,Jaaskelainen J. Adhesion molecules in the binding andinfiltration of human natural killer cells. In: GahmbergCG, Mandrup-Poulsen T, Wognsen Bach L, Hokfelt B,eds. Leukocyte adhesion: basic and clinical aspects.proc 6th Novo Nordisk Foundation Symposium "Leu-kocyte adhesion" 1992. p. 353-60.

118. Miller JS, Alley KA, McGlave PB. Differentiation of nat-ural killer cells from human primitive marrow progen-itors in a stroma based long term culture system: iden-tification of a CD34+/CD7+ NK progenitor. Blood1994; 83:2594-601.

119. Mrozek E, Anderson P, Caligiuri MA. Role of inter-leukin-15 in the development of human CD56+ nat-ural killer cells from CD34+ hematopoietic progenitorcells. Blood 1996; 87:2632-40.

120. Pierson BA, Miller JS. CD56+bright and CD56+dimnatural killer cells in patients with chronic myeloge-nous leukemia progessively decrease in number,respond less to stimuli that recruit clonogenic naturalkiller cells, and exhibit decreased proliferation on a percell basis. Blood 1996; 88:2279-87.

121. Daniels B, Daniels B, Karlhofer FM, Seaman WE,Yokoama W. A natural killer cell receptor specific fora major histocompatibility complex class I molecule. J

Exp Med 1994; 180:687-92.122. Moretta A, Vitale M, Sivori S, et al. Human natural

killer cell receptors for HLA-class I molecules. Evidencethat the kp43 (CD94) molecule functions as receptorfor HLA-B alleles. J Exp Med 1994; 180:545-55.

123. Rosenberg SA, Lotze MT, Muul LM, et al. A progressreport on the treatment of 157 patients with advancedcancer using lymphokine-activated killer cells and inter-leukin-2 or high-dose interleukin-2 alone. N Engl J Med1987; 316:889-97.

124. Parrado A, Rodriguez-Fernadez JM, Casares S, et al.Generation of LAK cells in vitro in patients with acuteleukemia. Leukemia 1993; 7:1344-8.

125. Attal M, Blaise D, Marit G, et al. Consolidation treat-ment of adult acute lymphoblastic leukemia: aprospective, randomized trial comparing allogeneic ver-sus autologous bone marrow transplantation and test-ing the impact of recombinant interleukin-2 after autol-ogous bone marrow transplantation. Blood 1995;86:1619-28.

126. Klingemann HG, Deal H, Reid D, Eaves CJ. Design andvalidation of a clinically applicable culture procedurefor the generation of interleukin-2 activated naturalkiller cells in human bone marrow autografts. ExpHaematol 1993; 21:1263-70.

127. Koo GC, Peppard JR, Latime EC. Characterization ofcytotoxic cells generated from bone marrow culture.Cell Immunol 1986; 98:172-80.

128. Silva MRG, Hoffman R, Srour EF, Ascensao JL. Gener-ation of human natural killer cells from immatureprogenitors does not require marrow stromal cells.Blood 1994; 84:841-6.

129. Lotzova E, Savary CA, Champlin RE. Genesis of humanoncolytic natural killer cells from primitive CD34+CD33- bone marrow progenitors. J Immunol 1993;150:5263-9.

130. Miller JS, Verfaille C, McGlave PB. The generation ofhuman natural killer cells from CD34+/DR- primitiveprogenitors in long-term bone marrow culture. Blood1992; 80:2182-7.

131. Silva MRG, Parreira A, Ascensao JL. Natural killer cellnumbers and activity in mobilized peripheral bloodstem cell grafts: conditions for in vitro expansion. ExpHematol, 1995; 23:1676-81.

132. Miller JS, Prosper F, MC Cullar V. Natural Killer cellsare functionally abnormal and NK cell progenitors arediminished in granulocyte colony-stimulating factor-mobilized peripheral blood progenitor cell collections.Blood 1997; 90:3098-105.

133. Hercend T, Farace F, Baume D, et al. Immunotherapywith lymphokine-activated natural killer cells andrecombinant interleukin-2: a feasibility trial in meta-static renal cell carcinoma. J Biol 1990; 9:546-55.

134. Beaujean F, Bernaudin F, Kuentz M, et al. Successfulengraftment after autologous transplantation of 10-day cultured bone marrow activated by interleukin 2 inpatients with acute lymphoblastic leukemia. BoneMarrow Transplant 1995; 15:691-6.

135. Wong EK, Eaves C, Klingemann HG. Comparison ofnatural killer activity of human bone marrow andblood cells in culture containing IL-2, IL-7 and IL-12.Bone Marrow Transplant 1996; 18:63-71.

136. Miller JS, Klingsporn S, Lund J, et al. Large-scale ex vivoexpansion and activation of human natural killer cellsfor autologous therapy. Bone Marrow Transplant1994; 14: 555-62.

137. Pierson BA, Europa AF, Hu WS, Miller JF. Productionof human natural killer cells for adoptive immunother-apy using a computer-controlled stirred-tank bioreac-tor. J Hematother 1996; 5:475-83.

138. Vujanovic NL, Yasumura S, Hirabayashi H, et al. Anti-tumor activities of human IL-2 activated natural killer


114


cells in solid tumor tissues. J Immunol 1995; 154:281-9.

139. Lanier LL, Civin CI , Loken MR, Phillips JH. The rela-tionship of CD16 (Leu-11) and Leu-19 (NKH-1) anti-gen expression on human peripheral blood NK cellsand cytotoxic T lymphocytes. J Immunol 1986; 136:4480-6.

140. Schmidt-Wolf IGH, Lefterova P, Johnston V, Huhn D,Blume KG, Negrin RS. propagation of large numbersof T cells with natural killer cell markers. B J Haematol1994; 87:453-8.

141. Scheffold C, Brandt K, Johnston V, et al. Potential ofautologous immunologic effector cells for bone mar-row purging in patients with chronic myeloid leukemia.Bone Marrow Transplant 1995; 15:33-9.

142. Slavin S, Ackerstein A, Weiss L, Nagler A, Reuven O,Naparstek E. Immunotherapy of minimal residual dis-ease by immunocompetent lymphocytes and their acti-vation by cytokines. Cancer Invest 1992; 10:221-7.

143. Glass B, Uhrek L, Zeis M, Loefler H, Mueller-RuchholtzW, Gassmann W. Graft-versus-leukaemia activity canbe predicted by natural cytotoxicity against leukaemiacells. Br J Haematol 1996; 93:412-20.

144. Rocha M, Umansky V, Lee KH, Hacker HJ, Benner A,Schirrmacher V. Difference between graft-versus-leukemia and graft-versus-host reactivity. I. Interactionof donor immune T cells with tumor and/or Host cells.Blood 1997; 89:2189-202.

145. Van der Harst D, Goulmy E, Falkenburg JHF, et al.Recognition of minor histocompatibility antigens onlymphocytic and myeloid leukemic cells by cytotoxicT-cell clones. Blood 1994; 83:1060-6.

146. Hoffmann T, Theobald M, Bunjes D, Weiss M, Heim-pel H, Heit W. Frequency of bone marrow T cellsresponding to HLA-identical non-leukemic andleukemic stimulator cells. Bone Marrow Transplant1993; 12:1-8.

147. Kolb HJ, Mittermueller J, Clemm CH, et al. Donorleukocyte trasfusions for treatment of recurrent chron-ic myelogenous leukemia in marrow transplantpatients. Blood 1990; 76:2462-5.

148. Champlin R. Separation of graft-vs.-host disease andgraft-vs-leukemia effect against chronic myelogenousleukemia. Exp Hematol 1995; 23:1148-51.

149. Datta AR, Barrett AJ, Jang YZ, et al. Distinct T cell pop-ulations distinguish chronic myeloid leukaemia cellsfrom lymphocytes in the same individual: a model forseparating GVHD from GVL reactions. Bone MarrowTransplant 1994; 14:517-24.

150. Oettel KR, Wesly OH, Albertini MR, et al. AllogeneicT-cell clones able to selectively destroy Philadelphiachromosome-bearing (Ph1+) human leukemia linescan also recognize Ph1- cells from the same patient.Blood 1994; 83:3390-402.

151. Hsu FJ, Benike C, Fagnoni F, et al. Vaccination ofpatients with B-cell lymphoma using autologous anti-gen-pulsed dendritic cells. Nature Med 1996; 2:52-8.

152. Rooney CM, Smith CA, Loftin S, et al. Use of gene-modified virus-specific T lymphocytes to control Ep-stein-Barr-virus-related lymphoproliferation. Lancet1995; 345:9-13.

153. Walter EA, Greenberg PD, Gilbert MJ, et al. Reconsti-tution of cellular immunity against cytomegalovirus inrecipients of allogenic bone marrow by transfer of T-cell clones from the donor. N Engl J Med 1995; 333:1038-44.


155. Weiner GJ, De Gast GC. Bispecific monoclonal anti-

body therapy of b-cell malignancy. Leuk Lymphoma1995; 16:199-207.

156. Disis ML, Bernhard H, Shiota FM, et al. Granulocytemacrophage colony-stimulating factor: an effectiveadjuvant for protein and peptide-based vaccines.Blood 1996; 88:202-10.

157. Faber LM, Van Luxemburg-Heijs SAP, Veenhof WFJ,Willemze R, Falkenburg JHF. Generation of CD4+ cyto-toxic T-lymphocyte clones from patient with severegraft-versus-host disease after allogenic bone marrowtransplantation: implication for graft-versus-leukemiareactivity. Blood 1995; 86:2821-8.

158. Bonini C, Verzelletti S, Servida P, et al. Transfer of theHSV-TK gene into donor peripheral blood lymphocytesfor in vivo immunomodulation of donor antitumorimmunity after alloBMT. Blood 1994; 84:110a.

159. Steinman RM. Dendritic cells and immune-based ther-apies. Exp Hematol 1996; 24:859-62.

160. Lanzavecchia A. Identifying strategies for immuneintervention. Science 1993; 260:937-44.

161. Wong J. Dendritic cells reach out to the clinic. NatureMed 1997; 3:129.

162. Young JW, Inaba K. Dendritic as adjuvants for class Imajor histocompatibility complex-restricted antitumorimmunity. J Exp Med 1996; 183:7-11.

163. Reid CDL. The dendritic cell lineage in haemopoiesis.Br J Haematol 1997; 96:217-33.

164. Galy A, Travis M, Cen D, Chen B. Human T, B, natur-al killer, and dendritic cells arise from a common bonemarrow progenitor cell subset. Immunity 1995; 3:459-73.

165. Austyn JM. New insights into the mobilization andphagocytic activity of dendritic cells. J Exp Med 1996;183:1287-92.

166. Gluckman JC, Canque B, Chapuis F, Rosenzwajg. Invitro generation of human dendritic cells and cell ther-apy. Cytokin Cell Mol Ther 1997; 3:187-96.

167. Reid CDL, Stackpole A, Meager A, Tikerpae J. Interac-tions of tumor necrosis factor with granulocyte-macrophage colony-stimulating factor and othercytokines in the regulation of dendritic cell growth invitro from early bipotent CD34+ progenitors in humanbone marrow. J Immunol 1992; 149:2681-8.

168. Caux C, Dezutter-Dambuyant C, Schmitt D, Bancher-au J. GM-CSF and TNF-a cooperate in the generationof dendritic Langerhans cells. Nature 1992; 360:258-61.

169. Romani N, Gruner S, Brang D, et al. Proliferating den-dritic cell progenitors in human blood. J Exp Med1994; 180: 83-93.

170. Siena S, Di Nicola M, Bregni M, et al. Massive ex-vivogeneration of functional dendritic cells from mobilizedCD34+ blood progenitors for anticancer therapy. ExpHematol 1995; 23:1463-71.

171. Szabolcs P, Moore MAS, Young JW. Expansion ofimmunostimulatory dendritic cells among the myeloidprogeny of human CD34+ bone marrow precursorscultured with c-kit ligand, granulocyte-macrophagecolony-stimulating factor, and TNF-a. J Immunol1995; 154:5851-61.

172. Bernhard H, Disis ML, Heimfeld S, Hand S, GralowJR, Cheever JR. Generation of immunostimulatory den-dritic cells from human CD34+ hematopoietic prog-enitor cells of the bone marrow and peripheral blood.Cancer Res 1995; 55:1099-104.

173. Mackensen A, Herbst B, Kohler G, et al. Delineation ofthe dendritic cell lineage by generation of large num-bers of Birbeck granule-positive Langerhans cells fromhuman peripheral blood progenitors cells in vitro.Blood 1995; 86:2699-707.

174. Rosenzwajg M, Canque B, Gluckman JC. Human den-dritic cell differentiation pathway from CD34+ hema-

M. Aglietta et al.

115


topoietic precursor cells. Blood 1996; 87:535-44.175. Fisch P, Kohler G, Garbe A, et al. Generation of anti-

gen-presenting cells for soluble protein antigens ex-vivo from peripheral blood CD34+ hematopoieticprogenitor cells in cancer patients. Eur J Immunol1996; 26:595-600.

176. Szabolcs P, Avigan D, Gezelter S, et al. Dendritic cellsand macrophages can mature independently from ahuman bone marrow derived post-colony-forming unitintermediate. Blood 1996; 87:4520-30.

177. Mortarini R, Di Nicola M, Siena S, et al. Cytokinerequirements for eliciting melanoma specific cytotox-ic T lymphocytes by melanoma antigen peptidesloaded on CD34+ cell-derived versus monocyte-derived autologous dendritic cells. Haematologica1996; 81 (suppl. to 5):83.

178. Mortarini R, Anichini A, Di Nicola M, et al. Autologousdendritic cells derived from CD34+ progenitors andfrom monocytes are not functionally equivalent anti-gen presenting cells in the induction of Melan-A/MART-127-35-specific cytotoxic T-lymphocytesfrom peripheral blood lymphocytes of melanomapatients with low frequency of cytotoxic T-lymphocyteprecursors. Cancer Res 1998; in press.

179. Thurner M, Papesh C, Ramoner R, et al. In vitro gen-eration of CD83+ human blood dendritic cells foractive tumor immunotherapy. Exp Hematol 1997;25:232-7.

180. Sallusto F, Lanzavecchia A. Efficient presentation ofsoluble antigen by cultured human dendritic cells ismantained by granulocyte/macrophage colony-stimu-lating factor plus interleukin 4 and downregulated bytumor necrosis-a. J Exp Med 1994; 179:1109-18.

181. Mukherji B, Chakraborty NG, Yamasaki S, et al. Induc-tion of antigen-specific cytolytic T cells in-situ inhuman melanoma by immunization with syntheticpeptide-pulsed autologous antigen presenting cells.Proc Natl Acad Sci USA 1995; 92:8078-82.

182. Hu X, Chakraborty NG, Sporn JR. Enhancement ofcytolytic T lymphocyte precursor frequency in melano-ma patients following immunization with MAGE-1peptide loaded antigen presenting cell-based vaccine.Cancer Res 1996; 56:2479-83.

183. Nestle FO, Alijagic S, Gilliet M, et al. Vaccination ofmelanoma patients with peptide-or tumor lysate-pulsed dendritic cells. Nature Med 1998; 4:328-32.

184. Porgador A, Gilboa E. Bone marrow generated den-dritic cells pulsed a class I-restricted peptide are potentinducers of cytotoxic T-lymphocytes. J Exp Med 1995;182:255-60.

185. Strobl H, Riedl E, Scheinecker C, et al. TGF-b1 pro-motes in vitro development of dendritic cells fromCD34+ hemopoietic progenitors. J Immunol 1996;157:1499-507.

186. Gianni AM, Bregni M, Siena S, et al. High-dose chemo-therapy and autologous bone marrow transplantationcompared with MACOP-B in aggressive B-cell lym-phoma. N Engl J Med 1997; 336:1290-7.

187. Siena S, Di Nicola M, Mortarini R, et al. Expansion ofimmunostimulatory dendritic cells from peripheralblood of patients with cancer. The Oncologist 1997;2:65-9.

188. Ratta M, Rondelli D, Fortuna A, et al. Generation andfunctional characterization of human dendritic cellsderived from CD34+ cells mobilized into peripheralblood: Comparison with bone marrow CD34+ cells. BrJ Haematol 1998; in press.

189. Romani N, Reider D, Heuer M, et al. Generation ofmature dendritic cells from human blood. An improvedmethod with special regard to clinical applicability. JImmunol Methods 1996; 196:137-51.

190. Grabbe S, Beissert S, Schwarz T, Gransetin RD. Den-

dritic cells as initiators of tumor immune responses: Apossible strategy for tumor immunotherapy? ImmunolToday 1995; 16:117-21.

191. Knight SC. Dendritic cells as initiators of tumor immu-nity. Immunol Today 1995; 16:547.

192. Nagata S. Fas ligand and immune evasion. NatureMed 1996; 2:1306-7.

193. Gabrilovich DI, Chen HL, Girgis KR, et al. Productionof vascular endothelial growth factor by humantumors inhibits the functional maturation of dendrit-ic cells. Nature Med 1996; 2:1096-103.

194. Gabrilovich DI, Ciernik IF, Carbone DP. Dendritic cellsin antitumor immune responses. 1. Defective antigenpresentation in tumor-bearing hosts. Cell Immunol1996; 170:101-10.

195. Gabrilovich DI, Nadaf S, Corak J. Dendritic cells inantitumor immune responses. 2. Dendritic cells grownfrom bone marrow precursors, but not mature DCsfrom tumor-bearing mice, are effective antigen carriersin the therapy of established tumors. Cell Immunol1996; 170:111-9.

196. Steptoe RJ, Thomson AW. Dendritic cells and toler-ance induction. Exp Immunol 1996; 105:397-402.

197. Finkelman FD, Lees AL, Gause WC, Morris SC. Den-dritic cells can present antigen in vivo in a tolerogenicor immunogenic fashion. J Immunol 1996; 157:1406-14.

198. Grabbe S, Bruvers S, Gallo RL, Knisely TL, Nazareno R,Granstein RD. Tumor antigen presentation by murineepidermal cells. J Immunol 1991; 146:3656-61.

199. Cohen PJ, Cohen PA, Rosenberg SA, Katz SI, Mul JJ.Murine epidermal Langerhans cells and splenic den-dritic cells present tumor-associated antigens toprimed T cells. Eur J Cancer 1994; 24:315-9.

200. Flamand V, Sornasse T, Demanet C, et al. Murine den-dritic cells pulsed in-vitro with tumor antigen inducetumor resistance in-vivo. Eur J Immunol 1994; 24:605-10.

201. Mayordomo JI, Zorina T, Storkus WJ, et al. Bone mar-row-derived dendritic cells pulsed with synthetic tumorpeptides elicit protective and therapeutic antitumorimmunity. Nature Med 1995; 12: 1297-302.

202. Celluzzi CM, Mayordomo JI, Storkus WJ, Lotze MT,Falo LD. Peptide pulsed dendritic cells induce antigen-speciifc, CTL-mediated protective tumor immunity. JExp Med 1996; 183:283-7.

203. Paglia P, Chiodoni C, Rodolfo M, Colombo MP.Murine dendritic cells loaded in vitro with soluble pro-tein prime CTL against tumor antigen in vivo. J ExpMed 1996; 183:317-22.

204. Marchand M, Weynants P, Rankin E, et al. Tumorregression responses in melanoma patients treatedwith a peptide encoded by gene MAGE-3. Int J Cancer1995; 63:883-5.

205. Maraskovsky E, Brasel K, Teepe M, et al. Dramaticincrease in the numbers of functionally mature den-dritic cells in flt3 ligand-treated mice: Multiple den-dritic cell subpopulations identified. J Exp Med 1996;184:1953-62.

206. Lynch DH, Andreasen A, Maraskovsky E, Whitmore J,Miller RE, Schuh JCL. Flt3 ligand induces tumor regres-sion and antitumor immune responses in vivo. NatureMed 1997; 6:625-31.

207. Chen W, Peace DJ, Rovira DK, You S, Cheever MA. Tcell immunity to the joining region of p210 bcr-ablprotein. Proc Natl Acad Sci USA 1992; 89:1468-73.

208. Bocchia M, Wentworth PA, Southwood S, et al. Spe-cific binding of leukemia oncogene fusion protein pep-tides to HLA class I molecules. Blood 1995; 85:2680.

209. Greco G, Fruci D, Accapezzato D, et al. Two bcr-abljunction peptides bind HLA-A3 molecules and allowspecific induction of human cytotoxic T lymphocytes.


116


Leukemia 1996; 10:693-9.210. Mannering SI, McKenzie JL, Fearnley DB, Hart DNJ.

HLA-DR1-Restricted bcr-abl (b3a2)-specific CD4+ Tlymphocytes respond to dendritic cells pulsed withb3a2 peptide and antigen-presenting cells exposed tob3a2 containing cell lysates. Blood 1997; 90:290-7.

211. Choudhury A, Gajewski JL, Liang JC, et al. Use ofleukemic dendritic cells for the generation of anti-leukemic cellular cytotoxicity against Philadelphiachromosome-positive chronic myelogeneous leukemia.Blood 1997; 89:1133-42.

212. Dermine S, Bertazzoli C, Marchesi E, et al. Lack of T-cell mediated recognition of the fusion region of thepml/RAR-a hybrid protein by lymphocytes of acutepromyelocytic leukemia patients. Clin Cancer Res

1996; 2:593.213. Henderson RA, Nimgaonkar MT, Watkins SC, Finn O:

Human dendritic cells genetically engineered to expresshigh levels of the human epithelial tumor antigenmucin (MUC-1). Cancer Res 1996; 56:3763-70.

214. Aicher A, Westermann J, Cayeux S, et al. Successfulretroviral mediated transduction of a reporter gene inhuman dendritic cells: Feasibility of therapy with genemodified antigen presenting cells. Exp Hematol 1997;25:39-44.

215. Gong J, Chen D, Kashiwaba M, Kufe D. Induction ofantitumor activity by immunization with fusions ofdendritic and carcinoma cells. Nature Med 1997;5:558-61.

M. Aglietta et al.


review

Cell therapy: achievements and perspectives


CLAUDIO BORDIGNON,* CARMELO CARLO-STELLA,° MARIO PAOLO

COLOMBO,@ ARMANDO DE VINCENTIIS,^ LUIGI LANATA,^ ROBERTO

MASSIMO LEMOLI,§ FRANCO LOCATELLI,** ATTILIO OLIVIERI,°°DAMIANO RONDELLI,§ PAOLA ZANON,^ SANTE TURA§

*Institute of Hematology, S. Raffaele Hospital, Milan; °Division of Oncology, Istituto Nazionale dei Tumori, Milan; @Experimental Oncology D, Istituto Nazionale dei Tumori,Milan; ^Dompé Biotec, Milan; §Institute of Hematology andMedical Oncology, University of Bologna and Policlinico S.Orsola Bologna; °°Division of Hematology, University ofAncona, Ancona; **Department of Pediatrics, University ofPavia Medical School and IRCCS Policlinico S. Matteo, Pavia;and Amgen Italia, Milan, Italy

Background and Objectives. Cell therapy can be consid-ered as a strategy aimed at replacing, repairing, orenhancing the biological function of a damaged tissue orsystem by means of autologous or allogeneic cells. Therehave been major advances in this field in the last fewyears. This has prompted the Working Group onHematopoietic Cells to examine the current utilization ofthis therapy in clinical hematology.

Evidence and Information Sources. The method employedfor preparing this review was that of informal consensusdevelopment. Members of the Working Group met threetimes, and the participants at these meetings examined alist of problems previously prepared by the chairman. Theydiscussed the single points in order to reach an agree-ment on different opinions and eventually approved thefinal manuscript. Some of the authors of the presentreview have been working in the field of cell therapy andhave contributed original papers in peer-reviewed jour-nals. In addition, the material examined in the presentreview includes articles and abstracts published in jour-nals covered by the Science Citation Index and Medline.

State of the Art. Lymphokine-activated killer (LAK) andtumor-infiltrating lymphocytes (TIL) have been used sincethe ’70s mainly in end-stage patients with solid tumors,but the clinical benefits of these treatments has not beenclearly documented. TIL are more specific and potentcytotoxic effectors than LAK, but only in few patients(mainly in those with solid tumors such as melanoma andglioblastoma) can their clinical use be considered poten-tially useful. Adoptive immunotherapy with donor lympho-cyte infusions has proved to be effective, particularly inpatients with chronic myeloid leukemia, in restoring astate of hematologic remission after leukemia relapseoccurring following an allograft. The infusion of donor T-cells can also have a role in the treatment of patients withEpstein-Barr virus (EBV)-induced post-transplant lympho-proliferative disorders. However, in this regard, generationand infusion of donor-derived, virus specific T-cell lines orclones represents a more sophisticated and saferapproach for treatment of viral complications occurring inimmunocompromized patients. Whereas too few clinicaltrials have been performed so far to draw any firm conclu-sion, based on animal studies dendritic cell-basedimmunotherapy holds promises of exerting an effective

anti-tumor activity. Despite leukemic cells not beingimmunogenic, induction on their surface of co-stimulatorymolecules or generation of leukemic dendritic cells mayinduce antileukemic cytotoxic T-cell responses. Tumorcells express a variety of antigens and can be geneticallymanipulated to become immunogenic. The main in vitroand in vivo functional characteristics of marrow mes-enchymal stem cells (MSCs) with particular emphasis ontheir hematopoietic regulatory role are reviewed. In addi-tion, prerequisites for clinical applications using culture-expanded mesenchymal cells are discussed

Perspectives. The opportuneness of using LAK cells oractivated natural killer (NK) cells in hematologic patientswith low tumor burden (e.g. after stem cell transplanta-tion) should be further explored. Moreover the role of newcytokines in enhancing the antineoplastic activity of NKcells and the infusion of selected NK in alternative to CTLfor graft versus leukemia (GVL) disease (avoiding graftversus host disease (GvHD) seems very promising. Sepa-ration of GVL from GvHD through generation and infusionof leukemia-specific T-cell clones or lines is one of themost intriguing and promising fields of investigations forthe future. Likewise, strategies devised to improveimmune-reconstitution and restore specific anti-infectiousfunctions through either induction of unresponsiveness torecipient alloantigens or removal of alloreactive donor T-cells might increase the applicability and success ofhematopoietic stem cell transplantation. Cellularimmunotherapy with DC must be standardized and sever-al critical points, discussed in the chapter, have to beproperly addressed with specific clinical studies. Stimula-tion of leukemic cells via CD40 receptor and transductionof tumor cells with co-stimulatory molecules and/orcytokines may be useful to prevent a tumor escapingimmune surveillance. Tumor cells can be genetically mod-ified to interact directly with dendritic cells in vivo orrecombinant antigen can be delivered to dendritic cellsusing attenuated bacterial vectors for oral vaccination.MSCs represent an attractive therapeutic tool capable ofplaying a role in a wide range of clinical applications inthe context of both cell and gene therapy strategies.©1999, Ferrata Storti Foundation

Key words: cell therapy

Correspondence: Prof. Sante Tura, Istituto di Ematologia e OncologiaSeragnoli, Policlinico S. Orsola, Via Massarenti 9 40138 Bologna, ItalyPhone: +39-051-390413 – Fax: +39-051-398973 – E-mail:[email protected]

118


The role of lymphoid cells in rejecting solid tumorstransplanted into animal models was strongly sug-gested in the first decades of this century by J.B.

Murphy (1926)1 who, nonetheless, did not demonstrateit formally. Following his revolutionary findings on theimmunologic mechanisms of allogeneic skin toleranceand rejection, in 1958 P.B. Medawar2 coined the term“immunologically competent cell” to define a cell that is“fully qualified to undertake an immunologicalresponse”. Forty years after Medawar’s definition, thedevelopment of molecular and biological research hasenormously improved our understanding of the com-plex regulatory mechanisms of proliferation, differenti-ation and function of the cells involved in the immuneresponse. The concomitant evolution of biotechnologyhas also progressively given new opportunities to isolateand/or expand cell subsets, or to develop new mole-cules, in order to amplify or modify specific cell func-tions. Thus, the possibility of exploiting a specific cellfunction, in vivo or ex vivo, to obtain a therapeuticeffect, such as an anti-tumor cytotoxic activity, or com-plete immune reconstitution, is part of the definition ofcell therapy that is herein reviewed.

In a general context, cell therapy can be considered asa strategy aimed at replacing, repairing, or enhancing thebiological function of a damaged tissue or system bymeans of autologous or allogeneic cells. For instance, inthe hematopoietic system cell therapy may include: a)removal or enrichment of various cell populations; b)expansion of hematopoietic cell subsets; c) expansion oractivation of lymphocytes for immunotherapy; and d)genetic modification of lymphoid or hematopoietic cells,when these cells are intended to engraft permanently ortransiently in the recipient and/or be used in the treat-ment of a disease.

This review contains extensive considerations on theclinical use of lymphocytes and/or natural killer (NK) cellsas a strategic weapon in preventing or curing the neo-plastic relapse after chemotherapy and/or hematopoieticstem cell transplantation, the infusion of T-cell clones orlines able to restore a specific anti-viral activity, the invivo and ex vivo potential use of dendritic cells to gener-ate a tumor-specific cytotoxic activity, and the innova-tive use of donor stromal cells in conjunction with stemcell transplantation. Even tumor cells engineered toexpress cytokine or co-stimulatory molecules and repre-senting the entire antigenic repertoire of a certain neo-plasia can be used as a cancer vaccine. On the other hand,a broad definition of cell therapy at this time shouldinclude autologous and allogeneic transplants of puri-fied hematopoietic stem cells, which, however, have beenextensively reviewed in previously published reports.3-5

Tumor escape from immune surveillanceAlthough several mechanisms allowing tumor cells to

escape the host immune protection have been recentlydescribed, it is conceivable that others remain stillundiscovered. However, tumor cells often fail to inducespecific immune responses because of their inability to

function as competent antigen presenting cells (APC).Professional APC, in fact, are fully capable of deliveringtwo signals to T cells:6 the first is antigen (Ag) specificand is mediated by the interaction of MHC moleculescarrying antigenic peptides with the T-cell receptor(TCR), and the second signal, or co-stimulatory signal, isnot Ag-specific and is principally mediated by membersof the B7 family, namely B7-1 (CD80) and B7-2 (CD86),via their T-cell receptors CD28 and CTLA-4, and/or byCD40 via CD40L binding.7-8

The lack of a suitable tumor-associated antigen (TAA),910

or defective antigen processing,11 or production ofimmunologic inhibitors,12 or lack of co-stimulatory sig-naling by tumor cells,13 as other mechanisms, can all con-tribute to prevent or abrogate an anti-tumor immuneresponse. Moreover, neoplastic cells within the sametumor may show different reactivity with monoclonalantibodies (mAbs), cytotoxic T-lymphocyte (CTL) clonesand lymphokine-activated killer (LAK) or tumor infiltrat-ing lymphocyte (TIL) populations. Furthermore, despitemany tumors having TAA and potentially being capableof stimulating T cells, in some cases they fail to inducean adequate CTL frequency in vitro. In other cases theantigen loss can be one of the mechanisms for escapingimmune protection.14 Private TAA often result frommutated gene products15 and are potentially useful fordeveloping tumor vaccines. These Ags, however, can bedown-regulated or modified by point mutations, induc-ing a consistent reduction or the abrogation of peptide-binding by specific CTLs. Another critical issue for pre-venting immune responses is the absence, or the down-regulation of MHC molecules on neoplastic cells, asshown in animal models,16 or in human lung cancer.17

The pivotal role of B7 molecules in the immuneresponse has been demonstrated in a variety of experi-mental models showing that after TCR signaling, bind-ing of CD28 induces T-cell interleukin-2 (IL-2) secre-tion, proliferation and effector function, whereas pre-sentation of the antigen in the absence of co-stimuliinduces T-cell unresponsiveness either by anergy orclonal deletion. Therefore, since most neoplastic cellslack co-stimulatory molecules, it is likely that they candeliver the first signal through the MHC:TCR binding,but not the second one, thus driving host T-cells to tol-erate the tumor.18 Potential strategies to prevent or toreverse T-cell tolerance by CD28 or CD40L stimulation,or IL-2 receptor triggering, are under investigation.

Further mechanisms impairing immunologic respons-es include the suppression of cytotoxic activity by therelease of soluble factors or by direct cell-contact. Infact, tumor cells may secrete cytokines, such as MIP-1a,or TGF-b, or IL-10, that may be capable of inhibiting Tcell activity.12 Alternatively, tumor cells may induce T-cell apoptotic clonal deletion by increasing Fas:Fas-Lligation.19 A schematic example of the main defectsdescribed in the tumor cell: T cell interaction is shownin Figure 1.

Finally, since normal lymphocytes can bind to venu-lar endothelial cells through adhesion receptors, such as

C. Bordignon et al.

119


L-selectin or a/b integrins, and then by rolling out theycan reach tissues, lack of adhesion receptors on tumorvessels might prevent lymphocytic infiltration and con-tact with neoplastic cells.20 In this case even the beststrategies aimed at modifying the immunogenicity oftumor cells may not be successful at overcoming thelack of an antitumor immune surveillance.

Lymphokine-activated killer and tumor-infiltrating lymphocytes: past and present

Natural killer cells and lymphokine-activat-ed killer phenomenon

Since 1970 NK cells have been recognized as a func-tionally distinct subset of cytotoxic effectors (Table 1).NK cells from rodent or from human peripheral blood killa wide range of tumor cells and virus-transformed cellswithout the need for prior sensitization.

In 1975 Heberman et al.21 described a phenomenon ofnormal unstimulated lymphoid cells lysing culturedtumor-cell lines in a short in vitro assay. This cytolyticactivity was subsequently shown to be neither MHCrestricted nor mediated by the T-cell receptor complex.Such ability to eliminate tumor cells, but not normaltissues suggests that NK cells are not only involved inthe control of cancer, but also that their presence andstate of activation are important in the outcome of thedisease and finally in the treatment of tumors.22

Mature NK have a clonally-distributed ability to rec-ognize their target cell by class I MHC alleles. Karre etal.23 demonstrated in a murine model that leukemia cell

Cell therapy

Figure 1. Main mechanisms for tumor escape of immunesurveillance.

Table 1. Characteristics of cytotoxic effectors useful for adoptive immunotherapy of cancer.

Effector type CTLs TILs NK cells LAK cells CIKs

Source Peripheral Metastatic Peripheral blood NK cells Subset of blood lymphocytes lymph nodes and bone marrow and CTL activated T-lymphocytes

by IL-2 activated by cytokines

Culture conditions:Tumor stimulation Yes None None None Noneneed of IL-2 for response ++++ ++++ ++++ (CD56dim) - ++++,IFN-g,

+ (CD56bright) IL-12, anti-CD3 antibodyDuration of culture 6 weeks 4 weeks 2-3 weeks 2-5 days 2-5 weekstarget cells in vitro allogeneic cells autologous K-562 Raji, Daudi autologous and

tumoral cells allogeneic tumoral cells

In vitro cytolytic activity :MHC restricted to none: spontaneous none: lyse cytotoxic activityallogeneic cells lysis of virus-infected cells, a wide spectrum of superior to LAK;

Specificity Restricted to autologous tumoral cells, tumor cells lyse whetherMHC not restricted, autologous tumor allogeneic tumoral cells including cells CML autologoustoward opsonized (MHC and/or that are or allogeneic blasts

cells (ADCC) tumor associated) antibody-dependent resistant to NK; but do not lyseantigens) cell-mediated cytotoxicity normal hematopoietic

(ADCC) specificity progenitors

Effector phenotype -CD3+/4+ ,CD3+/8+ CD3+/8+/56+ CD3-/CD16+/CD56+ CD56+ CD3+/56+ -CD3+/8+/16+. CD25+

CTL: cytotoxic T-lymphocytes; TIL: tumor inflitrating lymphocytes; NK: natural killer; LAK: lymphokine-activated killer; CIK: citokine-activated killer.

120


lines lacking certain MHC class I molecules were killedby NK cells, while parental H-2 bearing line were not. Inhumans both NK and a subset of cytotoxic T-lympho-cytes express receptors for MHC HLA class I moleculeswhich exert an inhibitory effect on cell-mediated cyto-toxicity. These surface molecules, belonging to theimmunoglobulin superfamily, have been termed killer-cell inhibitory receptors (KIRs). Two distinct KIR familieshave been described: a) KIRs with IG-like domains, rec-ognizing HLA-A, B and C alleles; and b) the CD94/NKG2Asubtype, with a lectin domain, recognizing peptidesrelated to the HLA-E class I system.24 The interactionbetween KIRs and the corresponding MHC class I anti-gens prevent NK from killing target cells expressing selfHLA alleles.25 In addition some NK also express receptorsthat induce lysis of target cells expressing foreign HLAclass I alleles.26

These findings explain the mechanism of self-toler-ance in the NK population, which can be disrupted as aconsequence of tumor transformation or viral infectionor any other events inducing a loss or a substantial mod-ification of class I molecules. These transformed cellscan easily escape detection by T-lymphocytes by downregulating MHC antigens, but are normally destroyedby autologous NK cells.27

The NK cell compartment is heterogeneous and dis-tinct NK subsets have been characterized. The mostinformative functional differences are based on relativeCD56 fluorescence: only CD56+bright, but not CD56+dim NK,express the high-affinity IL-2 receptor, and respond tothe low IL-2 concentration. They also expand 10 timesmore than CD56+dim.28

NK progenitors differentiate into immature NK inpresence of SCF, IL-7, IL-2 and bone marrow stromalcells producing IL-15. This last cytokine can directlyinduce CD34+ cells to differentiate into NK cells in theabsence of IL-2.29 The second step of NK maturation isstroma-independent and is characterized by the appear-ance of CD56 molecules: the intensity of CD56 expres-sion reflects the proliferative potential and the killingability of the NK.30

The effects of IL-2 on NK precursors appears to bestage-specific, confirming that, while mature NK pre-cursors readily respond to IL-2, more immature progen-itors need complete mixtures of cytokines and stromalcells. NK cells, after incubation with IL-2, become lym-phokine-activated killer cells: LAK cells kill NK-resistantcell targets (e.g. Daudi cell line) and a wide spectrum ofdifferent fresh tumor cells in both autologous and allo-geneic settings, while fresh normal tissues are resistantto LAK-mediated lysis.31

Although some tissue-resident lymphocytes may havespontaneous LAK activity, normal blood mononuclearcells (MNC) do not show any LAK activity, which can beacquired only after incubation with interleukin-2.32

These NK activated cells express new markers such asCD25, MHC class II antigens and fibronectin.33 LAK activ-ity can be generated not only in peripheral blood MNC,but also in the thymus, spleen, bone marrow and in MNC

from lymph nodes. Many experimental data suggest thatmost LAK precursors are present in the null lymphocytepopulation.

In humans LAK activity was much more evident in theMNC population after depletion of macrophages, T andB-cells. Residual MNC were CD16+ and did not show T-cell markers.34

LAK cells: experimental observations andclinical trials

In animal models the combined administration of IL-2 and LAK has proved to be more efficacious than eithercomponent alone. In murine models the administrationof high-dose IL-2 alone or in conjunction with LAK cellsinduced the regression of lung, liver and subdermalmetastases. The antitumor effect correlated both withthe IL-2 dose and the number of LAK cells administered;finally at different doses of IL-2, the concomitantadministration of LAK cells resulted in increased reduc-tion in established metastases.35,36

LAK cells are capable of inhibiting acute myeloidleukemia (AML) progenitor growth, and leukemia inci-dence is higher in people with deficiency of NK cells.37

In the large majority of patients at diagnosis or in relapseblasts appear resistant to lysis by autologous LAK cells.Moreover, about 90% of patients with acute leukemia incomplete remission do not show spontaneous cytotoxi-city against autologous blast cells, but ex vivo treatmentwith IL-2 restores cytolytic activity in 37.5% of thesepatients.38 In a population of 42 patients with AML incomplete remission, LAK cytotoxicity against autologousleukemic blasts was not significantly different from LAKof normal subjects.39 However, multivariate analysis forprognostic factors showed that patients whose LAK hadmore lytic activity on leukemic blasts had significantlyless risk of relapse than patients with poor LAK activity.

In the first National Cancer Institute trial endstagecancer patients received high-dose bolus IL-2 therapyfor 3 to 5 days.35 Lymphocytes harvested during the sys-temic treatment with IL-2 were cultured in the presenceof IL-2 for 2 to 4 days, in order to expand the LAK cellnumber; autologous LAK cells were then reinfused intopatients in combination with the high-dose intravenousbolus IL-2 administration. Of 72 patients with renal can-cer who were treated, 33% obtained an overall response,8 with complete response (CR) and 17 with partialresponse (PR); of 48 patients with metastatic melanoma21% responded with 4 CR and 6 PR; responses werealso observed in patients with colorectal carcinoma andnon-Hodgkin’s lymphoma.40 The ILWG used the samestrategy, obtaining an overall response rate of 19% inpatients with melanoma and 16% in those with renalcarcinoma.41 After these initial trials the original schemaof the National Cancer Institute was modified with theuse of IL-2 in continuous infusion rather than bolusinjection in order to reduce the systemic toxicity.42

The first randomized study, comparing IL-2 alone toIL-2 plus LAK cells, was published by McCabe.43 This tri-al included patients with either renal carcinoma or

C. Bordignon et al.

121


Cell therapy

melanoma; no significant difference in response ratebetween the two groups was reported. A second ran-domized study at the National Cancer Institute followedthese pioneering experiences, comparing IL-2 alone toIL-2/LAK cells:44 181 patients were enrolled in this study(90 in the IL-2 plus LAK arm and 91 in the IL-2 alone).A total of 10 CR were observed in the IL-2/LAK arm ascompared to only 3 in the IL-2 alone arm. The overallresponse rates were similar, but there was a survivaltrend (p=0.07) in favor of the IL-2/LAK arm: the actu-arial survival for patients receiving IL-2/LAK was 31%compared to 17% for those receiving IL-2 alone. Toxic-ity was virtually equivalent in both arms and the major-ity of toxic effects were due to IL-2 administration,while the only complication associated with LAK ther-apy was transient hepatitis A, due to contamination ofthe culture medium.

A third randomized trial, comparing IL-2 alone versusIL-2/LAK therapy was published in 1995.45 In this studyonly patients with advanced renal carcinoma weretreated and IL-2 was administered as a continuous infu-sion rather than bolus injection. Seventy-one patients

entered (36 vs. 35) this trial and only 6% overallobtained a major response, with a median survival of 13months; the difference between the two groups was notsignificant. Therefore it may be concluded that LAK cellsdid not improve the activity of IL-2 in patients withadvanced renal carcinoma.

The last randomized trial published was conducted in174 primary lung carcinoma patients after surgery, com-paring the adjuvant treatment with IL-2 plus LAK (fortwo years) with conventional treatment.46 The 5- and 9-year survival rates were significantly superior in patientsreceiving IL-2/LAK therapy, but no comparison wasplanned between Il-2 alone and IL-2/LAK therapy. Theimpressive results obtained in terms of overall survivalalso in non-curative cases after surgery (OS: 52% at 5years) should probably be interpreted as due to fact thatin this study patients received the immunotherapy afterconsistent tumor debulking.

Other clinical trials (non-randomized) were conduct-ed with IL-2 with or without LAK cells, and the overallresponse rate was similar for both the immunotherapymodalities.47,48 The detailed review of other (non-ran-

Table 2. Clinical trials with LAK cells.

Treatment schedule

Author Year Patients Kind of tumor IL-2 (dose and schedule) LAK cells Response

Rosenberg 1987 157 Melanoma Randomize: IL-2 vs. IL-2+LAK CR: 2.2% vs 7.5%PR: 10.9% vs 14.2%mR: 2.2% vs 9.4%

West 1987 40 Miscellaneous 1-7x106 U/m2/day CI CR+PR: 22-28%

Yoshida 1988 23 Brain tumor Direct injection of LAK into recurrent tumor Regression: 26%cavity + IL-2 (50-400 U); multiple treat

Fisher 1988 29v Renal carcinoma 12.9 MIU/kg (median 10 doses) 7x1010 cells OR: 16%

West 1989 30 Renal carcinoma 3x106 U/m2/day CI NR 22-28%

Dutcher 1989 32 100,000 U/kg q8h 8.9x1010 CR+PR: 19%

Paciucci 1989 24 Miscellaneous 1-5x106 U/m2/day CI 5.6x109 CR+PR: 20.8%

Neqrier 1989 51 Renal carcinoma 3x106 U/m2/day CI 1.2x1010 CR+PR: 27%

Stahel 1989 23 Miscellaneous 3x104 U/kg q8h 5.1x1010 CR+PR: 17%

Rosenberg 1993 181 Metastatic cancer Randomized: IL-2 vs IL-2+LAK CR: 5% vs 11.76%PR: 15.2% vs 16.5%OS (3 yrs): 17% vs 31% (p2=0.089)

Bajorin 49 Renal carcinoma Randomized: IL-2 vs IL-2+LAK (73x109) No difference(3 MU/m2)

Keilholz 1994 9 Liver metastic carcinoma IL-2 CI into the splenic artery LAK transfer into the CR+PR: 33%or intravenous infusion portal vein or the

hepatic artery

Murray Law 1995 66 Renal carcinoma Randomized: IL-2 vs IL-2+LAK (NR) CR+PR: 9% vs 3% (3x106 U/m2/day) (p=0.61)

Kimura 1997 82,788 Resected lung carcinoma Randomized: IL-2+LAK vs. Standard therapy OS (5 yrs): 54.4% vs 52%(7x105 U/day x 3 days) (1-5x109 cell) OS (9 yrs): 33.4% vs 24.2%

122


domized) experiences using these two differentimmunotherapies suggests that LAK cell reinfusionslightly increased the number of CR and the duration ofresponse, especially in patients with metastaticmelanoma (Table 2).49,50

In hematologic malignancies the first attempts to gen-erate and expand LAK activity by using IL-2 in vivo wereclinically disappointing especially in patients autotrans-planted for ALL; after transplantation patients were ran-domly assigned to treatment with systemic IL-2 (with-out LAK cell administration) or no treatment, but thedisease-free survival was similar in the two arms.51 Theuse of LAK cells has also been proposed after autologoustransplantation for hematological malignancies, but thevery small series of patients reported does not allow anydefinitive conclusion to be drawn about its clinical ben-efit.52 Beaujean et al.53 reinfused, after myeloablativetherapy, BM incubated with IL-2 into 5 ALL patients,observing a very marked delay of the engraftment andthe recurrence of disease in all patients. Recently therehas been a report of 61 women with breast cancer auto-transplanted with IL-2 activated PBPC and treated withlow dose IL-2 starting from PBPC reinfusion, withoutgraft failures or major toxicity; there are no data con-cerning the outcome of patients and this experience onlyconfirms the feasibility of the approach.54

In a very preliminary experience a sustained majorcytogenetic response to immunotherapy with GM-CSF+IL-2 and LAK infusion was observed in chronicmyeloid leukemia (CML) patients after autologous trans-plantation.55 However, a renewed interest in thisapproach has led to new research pursuing differentdirections:a. selection of patients with low tumor-burden and

with significant in vitro LAK activity against autolo-gous tumor cells, in order to reach an optimal effec-tor/target ratio;

b. harvest of large amounts of NK cells (for additionalex vivo expansion/activation with IL-2) to be rein-fused in the early phase after BMT;56

c. direct activation of leukapheresis, after priming withchemotherapy followed by cytokines, in order toreinfuse, after HDT, a product richer in cytotoxiceffectors and probably less contaminated;57

d. identification and selection of more efficient NKprogenitors (e.g. adherent NK) by eliminating unde-sired accessory cells which could inhibit their killingand proliferative ability;58, 59

e. generation and expansion of other CE subsets withmore powerful activity against autologous tumorcells, e.g. cytokine-induced killer cells (CIK);60

f. use of other cytokines in association with IL-2, inorder to potentiate the activity and/or improve theselectivity of activated peripheral blood MNC.

Tumor infiltrating lymphocytesThe disappointing results of adoptive immunothera-

py with blood-derived LAK cells led to a search for morespecific CE cells. Tumor infiltrating lymphocytes (TIL) are

T-lymphocytes with unique tumor activity that infiltratesome tumors and can be expanded by long-term culturewith IL-2 at low-intermediate concentrations.61 Inmurine models TIL have exhibited a stronger anti-tumoreffect than LAK cells on a per-cell basis; in humans TILhave been isolated with variable frequency from differ-ent solid tumors and very often (about 30% of cases)from patients with melanoma. Phenotypic analysisshowed that TIL consisted mainly of CD4+ cells in colon,breast and urothelial tumors, while in melanoma CD8+

cells are prevalent.62,63 CD3– CD16+ NK cells have alsobeen isolated from several tumors, confirming the largeheterogeneity of tumor infiltrates.64 The mechanism ofthe antitumor action of TIL is unknown; there is someevidence that these cells secrete cytotoxins andcytokines which are capable of killing tumor cells andrecruiting other CE.

Experimental models and clinical trialsMice carrying spontaneous metastases, treated with

IL-2 plus tumor-derived T-cells, obtained from spleno-cytes after mixed lymphocyte-tumor cultures, had a bet-ter survival than those treated with LAK cells; previoustumor debulking (with chemotherapy and/or radiother-apy) was needed to maximize the efficacy of TIL-thera-py.65

Unfortunately large amounts of TIL can be collectedvery rarely, and the large scale expansion of this popu-lation is crucial in order to obtain relevant clinicalresponses; this step of ex vivo manipulation is notalways successful, because the need for prolonged cul-ture of TIL (from 6 to 8 weeks with IL-2) may abrogatethe selectivity against the tumor; moreover only a smallfraction of the readministered human TIL is able to con-centrate in the tumor sites.66

Wong et al.67 showed in a mouse model that TIL pref-erentially localize in the liver and lungs. In contrast traf-ficking studies employing TIL radiolabeled with In111,have shown that TIL do traffic to tumor sites;68 this hom-ing property should produce high concentrations of TIL,and probably their permanence, in the area of a tumor.

Human TIL transfected in vitro with the neomycin-resistance gene and reinfused intravenously, have beendetected by polymerase chain reaction (PCR) techniquesfrom 6 to 60 days in patients affected by metastaticmelanoma.69 Aebersold et al.70 observed a strong correla-tion between the tested tumor cytotoxicity in vitro andthe in vivo response, in a small cohort of patients withmetastatic melanoma. A similar relationship wasobserved in a murine model in which the in vivo thera-peutic effect of TIL correlated with secretion of IFNg andtumor necrosis factor (TFNa).71

In order to increase their specificity and potency, TILhave been engineered with genes encoding cytokines orcytotoxins such as TNF or IFN-g or IL-2.69,72 However,some experimental observations suggest that these highconcentrations of cytokines can cause systemic toxici-ty and in some cases could even make the tumor moreaggressive.73,74

C. Bordignon et al.

123


In addition to their potential therapeutic use ascytolytic effectors, the ability of some TIL to recognizeunique antigens on tumor cells has made the study of thebiologic characteristics of these antigens more feasible.Melanomas from different patients who share MHC anti-gens are often cross-recognized by allogeneic TIL, ascould be expected for an MHC-restricted T-cell response;the presence of shared antigens in different patients withmelanoma suggests the possibility of using these anti-gens in an active immunization program for this disease.75

When adoptively transferred into patients, TIL showedsignificant therapeutic efficacy in patients with advancedmelanoma, but not in renal carcinoma patients. In aphase II trial patients with malignant melanoma weretreated with IL-2 and TIL following chemotherapy:76 39%of them achieved some sort of response, including thosewho had previously experienced a failure of IL-2 thera-py. Kradin et al.77 treated some patients with a combina-tion of chemotherapy, IL-2 and TIL, obtaining 23% ofresponses in those affected by melanoma and 29% inthose with renal carcinoma, but none in patients withnon-small cell lung carcinoma.

A summary of most clinically relevant clinical trialswith TIL is given in Table 3.

The lack of important clinical trials with TIL is proba-

bly due to the difficulties in finding TIL at diagnosis andespecially because the techniques for TIL priming andexpansion are time-consuming and not completelystandardized. TIL therapy is still young, but its very inter-esting potential has not yet been thoroughly investi-gated.

New approaches with LAK or TIL cellsAllogeneic setting. Whereas it is widely accepted that

graft-versus-host disease (GvHD) is initiated by donor Tcells recognizing foreign host antigens, other factorsincluding toxicity of conditioning regimens and cytokinedysregulation are involved in the pathogenesis ofGvHD.78,79 Data from murine experiments show that NKcells play an active role both in GvHD and in garft-ver-sus-leukemia (GVL) events: in a recently published mod-el 100% of SCID mice bearing human leukemic cells,and transplanted with NK+ T-cells, died of acute GvHD;but while animals which received only T-cells developedclinical GVL associated with relevant chronic GvHD, NK-transplanted animals showed the same degree of pro-tection from leukemia, experiencing only mild-moder-ate acute GvHD without chronic GvHD.80 These datasuggest that in order to optimize the GVL effect whileminimizing the severity of acute GvHD, donor grafts

Cell therapy

Table 3. Clinical trials with LAK cells and IL-2.

Treatment schedule

Author Year Patients Kind of tumor IL-2(dose and schedule) TIL cells Response

Rosenberg 1988 20 metastatic melanoma 13105 U/kg every 8h; CPM 25 mg/kg 20.531010 cell Regress: 60%

Kradin 1989 38 miscellaneous 1-33106 U/m2 CIx 24h OR: 26%

Rosenberg 1990 5 metastatic melanoma TIL gene modified

Aoki 1991 10 advanced or recurrent OR: 70%ovarian cancer TIL after single CI CPM Long term: 57%

Dillman 1991 21v metastatic melanoma 183106 IU/m2/day CI 1011 cell OR: 24%expensive, difficult

Arienti 1993 12v metastatic melanoma 1303106 IU/m2/day CI 6.83109 cell RR:33%

Belldegrun 1993 10v metastatic renal 23106 IU/m2/day in 96h (IL-2) TIL CR: 30%cell carcinoma 63106 IU/m2/day (IFN-g)

Schwartz– 1994 41 melanoma IL-2 TIL CR+PR: 21.9%entruber

Pockaj 1994 38 metastatic melanoma 7.23105 IU/kg every 8h 1.3-2.231011 cell OR: 38.5%and CPM 25 mg/kg

Chang 1997 20v advanced melanoma and IL-2 anti-CD3 vaccine primed OR:33.3%renal cell cancer lymph node cells PR:9.1%

activated

Curti 1998 solid tumor and NHL 9x106 UI/m2/day x 7 days CI T CD4+ cell+ anti CD3 some tumor regression

Ridolfi 1998 32 miscellaneous 12-6 MIU/ day 5.8x1010 TIL no response in patients(West's schedule) with advanced cancer

ev: evaluable; PR: partial response; OR: overall response.

124


should be manipulated by adding a moderate dose of T-cells in the early phase and using purified NK cells in thelate phase after transplantation.

Preliminary data suggest that in normal donors, afterG-CSF mobilization, NK progenitors have decreasedkilling capacity and diminished proliferative ability inresponse to IL-2, compared to the unprimed bone mar-row counterpart.81 In contrast, after an HLA incompat-ible transplant, a progressive expansion of NK and CTLwith NK like function (CD3+/CD56+) has been observed;recipients received the T-cell depleted graft withoutdeveloping GvHD, but in most cases a significant GVLeffect could be demonstrated both in vitro and in vivo;these data support the critical role of CTL KIR+ in thisparticular subset of transplanted patients.82

Concerning the expanding role of cord blood trans-plantation, even though the content of NK in this sourceseems normal, the decreased IL-12 production by cordblood MNC, reducing IFN-g stimulation, may contributeto reduce NK and LAK cytotoxicity; these data suggestone possible explanation for cord blood immaturity andtheir clinical implications such as decreased GvHD andGVL, which could be enhanced by IL-12 administration.83

Autologous setting. Considering the impressive resultsobserved in the allogeneic setting using donor-buffy coatlymphocytes for treatment of relapse, CML seems anattractive field for testing the efficacy of adoptiveimmunotherapy in the autologous setting too; someexperimental data support this hypothesis. The MNC ofpatients with CML contain a population of benign NKcells which can be expanded and activated by IL-2, gen-erating a CE population capable of killing both NK-sen-sitive and NK-resistant tumor targets.84 Both numberand functional activity of activated NK (ANK) in CMLpatients decrease with the progression of the disease.85

In vitro data show that autologous ANK inhibit bothcommitted and very early Philadelphia positive progen-itors in a MHC-unrestricted manner.86 In these experi-ments CML progenitor cell killing by autologous and allo-geneic ANK (after T-cell depletion) was comparable.Finally the CML blast killing was not dependent of solu-ble factors because it was abrogated by a transwellmembrane, but was mediated by cell-to-cell contactbeing significantly blocked by anti-integrin antibodies.87

In 1986 Lanier and Phillips described a subset of CD3+

T cells co-expressing the CD56 antigen which is a typicalNK marker (CIK).88 More recently Schmidt-Wolf et al.89

obtained large expansion of this subset in a 16-day liq-uid culture containing IFN-g, IL-1, IL-2 and a monoclon-al antibody against CD3 as the mitogenic stimulus. Thesame group tested the ability of this population to purgebone marrow in patients with CML; they found that whilestandard LAK cells were in most cases unable to lyse CMLcells, CIK cells were able to lyse both autologous and allo-geneic CML blasts, without affecting normal hemopoiet-ic progenitors.90 Recently it has been reported that CIKadministration in SCID mice bearing human CML inducedthe disappearance of Ph’+ cells in the spleen of 12/14animals.91

Another interesting potential application of autolo-gous LAK is the treatment of EBV-related lymphomasarising in organ-transplanted patients; a preliminarydescription of four complete responses after treatmentwith autologous peripheral MNC incubated with IL-2seems very promising.92 Recently in thyroid cancerpatients Katsumoto et al.93 generated cytotoxic CD4+

lymphocytes from TIL after non-specific in vivo stimula-tion with OK-432 (which induces severe local inflam-mation in the draining lymph nodes) and low-dose IL-2,obtaining large amounts of cytotoxic CD4+ (Th1) cells,producing high levels of IFN-g and TNF-b in the super-natants. These CE lysed a wide spectrum of tumor celllines; anti-TCR antibodies did not inhibit their killingactivity, which was in favor of a non-MHC restrictedlysis, while antibodies anti-ICAM-1 completely inhibit-ed the activity.

Tsurushima et al.94 induced autologous CTLs directlyfrom peripheral blood MNC by preparing a co-culture ofminced tissue fragments of glioblastoma multiformewith a mixture of cytokines (IL-1, 2, 4, 6 and IFN-g) for2 weeks. At the end of culture the population containedmainly CD4+ and CD8+ lymphocytes able to kill 82 to100% of the glioblastoma cells while normal LAK cellskilled only 33%.

Finally, in follicular lymphomas freshly isolated TIL,normally lacking tumor-specific cytotoxicity, were stim-ulated with lymphoma cells, in the presence of IL-2 andCD40 ligand; these T-TIL were capable of proliferatingin response to follicular lymphoma cells; moreover TILcould be further expanded in the presence of IL-4, IL-7and IFN-g.95

The potential role of new cytokinesSeveral cytokines affect CTL and NK response: first of

all IL-2 which expands the precursor pool of alloreac-tive CTL; IL-15 (producted by monocytes) mimics IL-2action by inducing IFN-g production, T-cell memoryactivation and CTL proliferation.96 IL-12 shares certainfunctional properties with IL-2, but using a different,IL-2 independent pathway.97 In addition IL-12 enhancesthe lytic activity of human peripheral blood MNC againsta wide spectrum of tumors.98,99 Recently it has beenobserved that the combination of IL-2 and IL-12 is capa-ble of inducing lysis of blasts resistant to IL-2-activat-ed effectors, even in the autologous setting.100

Therefore the association of IL-2 plus IL-12 could poten-tially become an important tool to increase the antitumorefficacy both ex vivo, by generating large amounts ofCIK,101 and in vivo, by systemic administration.

GM-CSF is a cytokine capable of inducing a pleiotropicimmunostimulatory effect and also increases the immuno-genicity of tumors; in a model for ex vivo expansion of LAKcells from leukaphereses in order to obtain contempora-neously a decontaminated harvest and a large amount ofCE to reinfuse after myeloablative therapy, the associationGM-CSF+IL-2 obtained a 5-fold expansion of the NK com-partment while sparing the clonogenic potential of hemo-poietic progenitors.102

C. Bordignon et al.

125


Biodistribution and targeting of LAK and TILAt present adoptive immunotherapy with LAK or IL-2

activated TIL has had limited success in patients withadvanced cancer. Although a well-defined mechanismremains to be established, numerous in vitro findings andin vivo data suggest that the cancer-specific cytotoxic-ity of CE is obtained in multiple steps; a prerequisite,however, is optimal delivery of CE to the target tissueswhile minimizing systemic cytotoxicity. Two major areascurrently requiring investigation are the survival andlocalization of adoptively transferred CE in the tumor-bearing host, and the detailed mechanism of tumorregression. The major goals in this area concern the opti-mal administration of systemic cytokines together withCE, and (finally) the ways to enhance localization andtranscapillary migration of the infused cells.

Experimental evidence together with theoretical con-siderations based on CE functions indicate that the abil-ity of adoptive immunotherapy to eradicate an estab-lished tumor is quantitatively determined by the initialtumor burden, growth pattern, and the magnitude ofimmunologic response generated by CE and other acces-sory cells at the site of the tumor.103,104 Thus, to achievetumor eradication and minimize systemic toxicity, theexplanation of the mechanisms underlying lymphocytebiodistribution and the factors governing effector celluptake in tumor sites is critical, but unfortunately dataabout CE biodistribution in humans are scarce.

Although a physiologically based kinetic modelingapproach has been applied to the pharmacokinetics ofdrugs and antibodies, there has been no effort to extendthis approach to cell biodistribution, probably becauseof its complexity.

One interesting attempt to apply this method to adop-tive immunotherapy has, however, recently been pub-lished.105 The importance of lymphocyte infiltration fromsurrounding normal tissues into tumor tissue was foundto depend on lymphocyte migration rate, tumor size,and host organ.

It is likely that therapy with CE has not been as effec-tive as originally promised, in part because of the verylow CE concentration in the systemic circulation; thiswas mainly due to lung entrapment. Reducing this phe-nomenon by decreasing the attachment rate or adhesionsite density in the lung by 50%, the tumor uptake couldbe increased by 40% for TIL to 60% for adherent NKcells.

Theoretical models indicate that intra-arterial admin-istration has a dramatic advantage over intravenousdelivery, with more than a 1,000-fold higher CE accu-mulation in the tumor site. Indeed experiments in murinemodels show that it is possible to eliminate liver metas-tasis by loco-regional administration of human IL-2 ANKor by systemic adoptive transfer.106

Finally the differences in biodistribution between dif-ferent lymphocyte populations, mainly due to the dif-ferent attachment rates in the tumor and the lung,should be carefully considered. ANK cells are more eas-ily trapped than CTL in lung vessels due to their larger

diameter and greater rigidity.107 A greater accumulationof TIL was expected in the spleen as a result of theirstronger adhesion at this site through the lymphocytehoming receptor.108 Although this model has limitationsrelated to the sensitivity of analysis of parameters suchas adhesion-site density, lymphocyte attachment andarrest rate, it could be considered a useful basis fordesigning new experimental models to increase the con-centration and recirculation of CE in tumor sites, reduc-ing effector cell rigidity or blocking adhesion molecules.

The so-called antibody-dependent cellular cytotoxic-ity (ADCC) could be mediated by cells expressing Fcgreceptor II and Fcg receptor III (e.g. NK cells andCD3+/CD16+ cells). This kind of cytotoxicity, even thoughexhibited by non MHC-restricted cells, cannot be con-sidered aspecific and is also exhibited by monocytes.

LAK cells are extremely potent mediators of ADCC109

and thus the use of LAK plus IL-2 in combination withmonoclonal antibodies will probably become a power-ful tool for treating some immunogenic tumors. Thisapproach has been tested in patients with colorectalcancer,110 but could be also proposed for treating someimmunogeneic hematologic malignancies such as fol-licular lymphoma or multiple myeloma.

Donor lymphocyte infusion for treatmentof leukemia relapse and as a means foraccelerating immunologic reconstitutionin patients given transplantation ofhematopoietic progenitors

Manipulation of the immune system after hema-topoietic stem cell transplantation (HSCT) to reverseleukemia relapse or to reduce its incidence remains oneof the most fascinating, even though difficult, chal-lenges for successful cure of patients with hematolog-ic malignancies. In fact, over the last 10-15 years, evi-dence has emerged from clinical transplantations tosuggest that the anti-leukemia effect of allogeneic HSCTcannot merely be ascribed to the myeloablative thera-py employed during the preparative regimen, donor lym-phocytes playing a pivotal role in the eradication ofmalignant cells. Adoptive immunotherapy with donorlymphocyte infusion (DLI) in patients relapsing afterHSCT has provided one of the most effective demon-strations of the importance of the graft-versus-leukemiaeffect in the cure of patients with hematologic malig-nancies.111,112

Even though DLI may sometimes be burdened by com-plications that endanger the patient's life, mainly myelo-suppression and GvHD, in individuals with CML experi-encing relapse in chronic phase after an allograft approx-imately 70% complete remissions can be obtained withthis treatment.112-116 Most of these remissions are sus-tained over time, this proving the capacity of DLI to erad-icate clonogenic leukemia cells or control their re-growth. DLI has also been extensively employed toreverse relapse in patients with acute leukemia, non-Hodgkin’s lymphoma and multiple myeloma. However,

Cell therapy

126


the response rate of patients with other hematologicmalignancies, especially acute leukemia, is significantlylower.115, 116 In fact, only 20-30% of patients with AMLachieve a hematologic remission after DLI and the val-ue for patients with ALL is even lower. Patients withacute leukemia experiencing recurrence following anallograft have a higher probability of response with DLIif treated after having achieved a state of completeremission with chemotherapy, that is in a condition char-acterized by a limited tumor burden.117

The most important factor predicting response to DLIin patients with CML is the type of relapse. In fact, asalready mentioned, patients suffering from cytogeneticrelapse or hematologic relapse in chronic phase have ahigh probability of response to DLI, while patients withmore advanced disease (accelerated phase or blast cri-sis) respond less frequently (20-25% of cases).112-116

Relapse occurring in the first 1-2 years after allograft,115

little or no acute and chronic GvHD after transplantationor removal of T-lymphocytes before HSCT116 are alsoassociated with a higher probability of benefitting fromDLI. In patients with CML responding to DLI, the mediantime to obtain hematologic remission has been report-ed to be about 6-8 weeks,115 whereas a longer time (inthe order of 11 months) is needed for molecular remis-sion, this documenting that clearance of leukemia cellsis a dynamic, progressive phenomenon.118 The number ofT-cells to be infused and the best schedule of DLI foroptimal response without concurrent development ofsevere GvHD are still to be conclusively established sincethey depend on several variables, such as degree of HLA-compatibility between donor and recipient, original dis-order, and type of relapse.112 Some authors have claimedthat infusion of no more than 13107 donor-derived T-cells per kg of recipient body weight or CD8-depletedlymphocytes can induce a state of remission and sub-stantially prevent GvHD occurrence.119 However, recent-ly, the Hammersmith Hospital group reported that theresponse in CML patients relapsing after HSCT and giv-en graded increments of donor lymphocytes seems to beless sustained over time than that observed after infu-sion of a larger number (i.e. >13108/kg of recipient bodyweight) of T-cells (Dazzi F, personal communication,1999). Support to the importance of the number of cellsinfused is also given by the results of Lokhorst et al.,120

who observed that, in multiple myeloma, patients givenmore than 13108 T-cells/kg had the highest probabilityof benefitting from DLI. In some of these patients, theresponse was complete with disappearance of myelomaproteins.

The two major complications occurring after DLI aremyelosuppression and GvHD. Myelosuppression is expe-rienced by approximately 50% of the patients treatedwith DLI for CML in hematologic relapse, while it occursmuch less frequently in patients with cytogenetic recur-rence,112 this indicating that such a complication isobserved in situations characterized by a predominanceof host-type hematopoiesis. Therefore, myelosuppres-sion can be explained by a direct effect of the transfused

donor lymphocytes on hematopoietic cells of the recip-ient, similarly to that observed in transfusion-associat-ed GvHD. The majority of patients experiencing myelo-suppression after DLI recover a normal blood cell countspontaneously: nevertheless, myelosuppression may befatal in approximately 10% of patients, with death beingcaused by infection or bleeding.115,116 Infusion of a hugenumber of donor-derived peripheral blood hematopoi-etic progenitors, mobilized through hematopoieticgrowth factors such as granulocyte colony-stimulatingfactor (G-CSF), can alleviate the problem of pancytope-nia in some selected cases, hastening the recovery ofneutrophil and platelet counts.

Grade II-IV acute GvHD develops in almost half ofpatients given DLI,115,116 the highest incidence beingobserved when the donor is an unrelated volunteer.121

Incidence and severity of GvHD after DLI does not appearto correlate with GvHD after the original transplant andit may occur with a high incidence since donor lympho-cyte therapy involves the infusion of large numbers of T-cells, whose immunocompetence is not usually modulat-ed by cyclosporin A and/or methotrexate. Even thoughGvHD occurring after DLI is well-correlated with diseaseresponse as proved by the observation that most patientsobtaining a hematologic remission after this treatmentdeveloped acute and/or chronic GvHD, GvHD may not besufficient to induce GVL. Moreover, some patients notexperiencing GvHD after DLI achieve hematologic remis-sion, this indicating the existence of a GVL effect sepa-rate from development of GvHD.113,116,117,122

GVL effect occurring after HSCT and DLI is consideredto be mediated by HLA-unrestricted NK or LAK cells or byT-lymphocytes that recognize leukemia cells in an HLA-restricted fashion.123,124 In particular, when patient anddonor are HLA-identical, it is believed that recipient non-MHC-encoded minor histocompatibility antigens (mHAg)are recognized by donor CTL. While widely distributedmHAg account for the GVL effect associated to GvHD, tis-sue restricted or leukemia-specific antigens can elicit aspecific GVL reaction108-113 and it has been demonstratedthat both CD4+ and CD8+ CTL recognizing mHAg in a clas-sical MHC-restricted fashion can be generated in vit-ro.124,125 In particular, mHAg-specific CD8+ CTL can dis-play strong lysis of mature leukemia cells, as well as sup-press, together with CD4+ mHAg-specific CTL, the growthof clonogenic leukemia precursor cells.126,127 Productionof cytokines (such as g-interferon and tumour necrosisfactor a) able to induce the apoptotic death of leukemiacells can also contribute to the GVL effect.128,129 This said,it is not surprising that several efforts have been direct-ed towards the identification of strategies capable ofselecting and/or amplifying specific GVL response, notassociated with development of GvHD. Since it has beendocumented in humans that CTL directed against allo-geneic leukemic blasts can be detected in the peripheralblood of healthy donors130 and that CTL specifically reac-tive towards recipient leukemic blasts can emerge andpersist over time in children given allogeneic HSCT131 apossible intriguing approach is that of generating and

C. Bordignon et al.

127


expanding clones or cell lines that are leukemia-reactive.The first elegant demonstration of the feasibility and effi-cacy of this sophisticated strategy has been recentlyreported by Falkenburg et al.,132 who, through the infusionof donor-derived in vitro cultured CTL specifically recog-nizing leukemia progenitor cells, induced a completehematologic and cytogenetic response in a patient withCML who had relapsed after an allograft and was resis-tant to DLI treatment.

A diverse, but equally elegant, approach proposed toabrogate the DLI-associated GvHD and its relevant mor-bidity and mortality is the infusion of thymidine kinasegene-transduced DLI followed by treatment of the recip-ient with ganciclovir if GvHD occurs.133 In a study report-ed by Bonini et al.133 this strategy proved to be able tocontrol GvHD in 3 patients experiencing this complica-tion after DLI; two of them, who had achieved a com-plete hematologic remission before ganciclovir adminis-tration, remained in full remission after disappearance ofthe transduced lymphocytes. If confirmed in a largernumber of patients with a longer follow-up, geneticmanipulation of donor lymphocytes, through the trans-fer of a suicide gene for specific and selective elimina-tion of effector cells responsible for GvHD, could demon-strate the possibility of separating GvHD from GVL effect,thus sparing the anti-leukemia activity of DLI.

One of the most important, still unsolved problem ofDLI is that concerning the much lower efficacy of GVL inpatients with acute leukemia than in those with CML. Animmediate explanation for this observation may be thatthe more rapid growth kinetics of blast cells, whichoccurs in patients with acute leukemia during the lagperiod between leukocyte infusion and GVL development,may hamper the immune-mediated effect played bydonor lymphocytes in controlling disease progression. Infact, response to DLI occurs after weeks and hence theexponential expansion of leukemia cells in vivo mayexceed the immune response.112,113,124 The more encour-aging results obtained when DLI is used as consolidationtherapy for patients who have obtained a completeremission after chemotherapy provide support for thisinterpretation. However, other hypotheses, involving dif-ferent intrinsic susceptibility of acute leukemia to adop-tive immunotherapy must be considered. In particular,since patients with ALL have the lowest chance both ofresponding to DLI and of benefitting from the GVL effectafter bone marrow transplantation,134 a peculiar resis-tance of lymphoid leukemia to immunotherapy cannotbe excluded.

As peptides differentially expressed within thehematopoietic system can trigger and act as a target ofthe GVL reaction,112,124 it could be hypothesized that, forexample, the presence of these antigens on myeloidblasts, but not on lymphoid leukemia cells accounts forthe low response of ALL to donor lymphocytes. Thereported demonstration of CTL response directedtowards peptides derived from proteinase 3, which isexpressed by myeloid cells (including blast cells),135 is atypical example of the possible differential susceptibil-

ity to the immune-mediated anti-leukemia effect of dif-ferent types of hematologic malignancies.

Several other possibilities exist to explain why acuteleukemia (and in particular ALL) can escape the GVLeffect. For example, leukemia cells may have defectiveexpression of HLA-class I or II molecules on their surfacesuch that they do not present antigens or, alternatively,the mechanisms of antigen processing and transport maybe impaired.112,129 Moreover, leukemia blasts may prod-uct cytokines (such as transforming growth factor b, IL-10) capable of suppressing T-cell activation, expansionand effector function or may express on their cell sur-face molecules, such as FAS ligand, able to mediate T-cell apoptosis.112,129 One of the most interesting fields ofinvestigation for explaining why in some patients a sus-tained anti-leukemia response in vivo fails to be inducedis that of co-stimulatory molecules. As previouslydescribed, full activation of T-cells requires two distinctbut synergistic signals.136 In fact, in the absence of co-stimulatory signals, a T cell encountering an antigenbecomes unresponsive to the appropriate stimulation(anergic)137 or undergoes programmed cell death (apop-tosis).138 Leukemia cells lacking these co-stimulatorymolecules have a poor capacity of inducing a T-cell spe-cific immune response and induction of CD80 and CD86,by signalling through the CD40 molecule, is able torestore T-cell co-stimulation via CD28 and to generateboth allogeneic and autologous CTL, which could con-tribute to inducing or maintaining a state of hematologicremission.139,140

Some clinical strategies have been devised to improvethe efficacy of adoptive immunotherapy in patients withacute leukemia. An approach for ameliorating the effi-cacy of DLI which has produced interesting results is thatrecently reported by Slavin et al.,106 who documented thatthe success rate of this adoptive immune therapy may beincreased in patients with both acute and chronicleukemia by activation of donor peripheral blood lym-phocytes with IL-2 both in vivo and/or in vitro. In partic-ular, a relevant proportion of patients who had notresponded to DLI were induced into remission only afterin vivo administration of IL-2 or in vitro activation ofdonor lymphocytes. If further confirmed, the resultsobtained make it possible to hypothesize that this strat-egy could be employed as first-line treatment of patientswith acute leukemia relapsing after an allograft, sinceALL and to a lesser extent AML patients do not greatlybenefit from DLI alone. Another reasonable attempt forimproving the response to DLI in patients with acuteleukemia is to use this adoptive immunotherapy in indi-viduals with minimal residual disease, as determined bycytogenetic investigations or sensitive molecular tools,that is in conditions characterized by a limited tumorburden, in which the GVL effect has demonstrated itsgreatest efficacy.

Unmanipulated DLI may also provide a means of com-pensatory T-cell repletion for the prevention of leukemiarecurrence in patients given a T-cell depleted marrowtransplantation from a relative. This approach has been

Cell therapy

128


recently proposed141 and studies enrolling larger cohortsof patients are necessary to define whether this strate-gy can be useful to prevent the increased risk of relapseassociated with the removal of donor T-cells. However,the main indication of adoptive infusion of donorimmune cells to accelerate immune reconstruction afterHSCT is transplants from HLA-disparate family donors.Infusion of a high number of T-cell depleted, peripher-al blood hematopoietic progenitors from these donorshas been demonstrated to be associated with a highchance (>95%) of donor hematopoietic engraftment.142

The significant delay in immune reconstitution, duemainly to removal of mature T cells from donor marrowand HLA disparity between donor and recipient, remainsthe major problem of HSCT from HLA-disparate donors.In fact, it is responsible for the dramatic incidence ofleukemia relapse and life-threatening viral and fungalinfections observed after this type of HSCT. A possiblestrategy to improve the process of immune recovery isto infuse donor T-lymphocytes selectively rendered non-reactive towards alloantigens of the recipient, but main-taining the capacity to generate an immune responseagainst viruses, fungi and leukemia cells. In this regard,as previously mentioned, the manipulation of co-stim-ulatory molecules is an extremely promising field ofinvestigation, since the absence of a second signalinduces anergy rather than activation of T lymphocytes.Drugs and monoclonal antibodies blocking co-stimula-tory pathways have been demonstrated to be able toprevent T-cell activation in response to alloantigens andto induce a state of anergy.143 In particular, it wasrecently documented that the combination of mono-clonal antibodies blocking CD80/CD86 molecules andcyclosporin A was able to generate a state of selectivein vitro unresponsiveness of T cells towards allo-anti-gens, not reversed by adding IL-2.144 Since the inductionof this state of unresponsiveness was associated withthe maintenance of in vitro capacity to respond towardvirus antigens and leukemia cells,145 the relevance ofthis approach is evident for strategies of donor T-celladd-back after T-cell depleted transplant of hematopoi-etic progenitors from HLA-partially matched donorsaimed at accelerating the process of immune reconsti-tution.

A different, but equally promising, method of deletionof unwanted alloresponses is based on the eliminationof alloreactive T-cells after specific activation throughtheir killing146 or fluorescence-activated cell sorting,147

while sparing T cells with other functions. In a humanpre-clinical study, it was demonstrated that allospecif-ic T-cell depletion by using an immunotoxin directedagainst the p55 chain of IL-2 receptor, was feasible andspecific.146 The spared T-cells were still able to prolifer-ate against third-party cells, Candida and cytomegalo-virus antigens,148 as well as to kill both leukemia blastsand autologous EBV-B lymphoblastoid cell lines.149

Moreover, in vivo studies in a murine animal modelshowed that this particular T-cell depletion was effi-cient, at least partially, in preventing both graft rejec-

tion and GvHD in a complete haplotype mismatchedcombination.150

Finally a brief mention should be made of the gener-ation and infusion of T cells with suppressive and regu-latory activity. A particular subset of these cells calledTr1 has recently been described by Groux et al.,151 whoin an animal model demonstrated the ability of this pop-ulation to prevent, through their activity on naive cells,the occurrence of ovo-albumin induced inflammatorybowel disease. Whether these cells will have a role inpromoting a true state of tolerance in transplant ofhematopoietic progenitors or solid organs (in which theimmune response to alloantigens is mainly sustained bymemory cells) remains to be proved in specific pre-clin-ical and clinical studies currently underway.

Adoptive immunotherapy for the treatmentof viral infections in immunocompromisedpatients

Prevention or treatment of viral infections in immune-compromised patients through the infusion of specific T-cell lines or clones is one of the most sophisticatedexamples of adoptive immunotherapy approaches.152 Infact, it implies the elaboration of true cellular-engineer-ing strategies able to generate, select and expand lym-phocyte subsets, which display a specific function. Thefirst study in humans to evaluate the efficacy of adop-tively transferred T-cell clones for reconstitution of spe-cific immunity was performed in recipients of allogene-ic HSCT at risk of developing human cytomegalovirus(HCMV) infection and/or disease.153 Even though pre-emptive therapy of HCMV infection based on monitor-ing of antigenemia154 and prophylaxis of seropositiveHSCT recipients using antiviral drugs (i.e. ganciclovir andfoscarnet)155 have significantly reduced the number ofpatients experiencing HCMV disease, this viral infectionstill represents a major life-threatening complication ofstem cell allograft. The capacity to recover from a severeHCMV infection in transplanted patients is directly cor-related with the ability of the host to generate virus-specific class I HLA-restricted CD8+ cytotoxic cells andduring the first 100 days after HSCT approximately 50%of patients are persistently deficient in CD8+ cytotoxic T-lymphocytes specific for HCMV.156,157 It is not surprisingthat, to evaluate the efficacy of adoptive immunother-apy in this viral infection, HCMV-specific CD8+ T-cellclones of donor origin were generated and infused inHSCT recipients.153,158 These cells, generated through ahighly complex expansion strategy using irradiateddonor-origin skin fibroblasts infected with a strain ofHCMV, proved to be efficient in the prophylaxis againstHCMV infections that can complicate allogeneic HSCT.Moreover, the cloning strategy allowed selection of Tcells which lacked significant alloreactive capacity and,thus, did not cause clinically relevant GvHD or toxicity.These clones, directed towards either pp65 or pp150 (twoabundant viral tegument proteins presented for recog-nition by cytotoxic T-lymphocytes), restored HCMV-spe-

C. Bordignon et al.

129


cific cytotoxicity, which persisted for several weeks.158 Infact, through a PCR technique able to detect the Va andVb T-cell receptor rearrangements specific for the donorclones, it was possible to prove the donor origin of thesecells formally and to document the persistence of theadoptively transferred HCMV-specific T-cells for at least12 weeks. Unfortunately, these clones persisted in thecirculation at high levels only in patients experiencing anendogenous recovery of CD4+ virus-specific cells.158 Bycontrast, in patient lacking this spontaneous recovery ofHCMV-specific CD4+ lymphocyte, the donor-origin,adoptively transferred cytotoxic T-cell activity progres-sively declined and eventually disappeared. This obser-vation emphasizes the importance of CD4+ lymphocytesin promoting sustained restoration of antigen-specificimmunity and suggests that the use of polyclonal T-celllines containing both CD4+ and CD8+ cells could bepreferable to the infusion of cytotoxic T-cell clones.

In this regard, the use of T-cell lines for preventionand/or treatment of Epstein-Barr virus-induced lym-phoproliferative disorders (LPD) has represented a fur-ther, equally sophisticated, evolution of the approach-es of adoptive immunotherapy for the restoration ofvirus-specific immunity. EBV-LPD have emerged as asignificant complication for both HSCT and solid organtransplant recipients.159-161 In the former cohort, the useof HLA-partially matched family and unrelated donors,as well as selective procedures of T-cell depletion spar-ing B-lymphocytes, are risk factors for the developmentof EBV-LPD.160-162 In HSCT recipients these disorders areof donor origin and usually present in the first 4-6months after transplantation, whereas in patients giv-en a solid organ allograft they usually develop from therecipient B-lymphocytes months to years after trans-plantation.160,161 High levels of EBV-DNA in blood and invitro spontaneous growth of EBV-lymphoblastoid celllines predict development of these lymphoproliferativedisorders.163 They often present as high-grade diffuselarge cell B-cell lymphomas, which are oligoclonal ormonoclonal and express the full array of EBV antigensincluding EBNA-1 through EBNA-6 and the latencymembrane proteins LMP-1 and LMP-2.161 The lym-phomas which develop in immunocompromised hostsnot only invade the hematopoietic system, but also thelung, nasopharynx and central nervous system. The ther-apeutic approaches proposed to date (i.e. discontinua-tion of immunosuppression, a-IFN, antiviral agents andcytotoxic chemotherapy) have been applied with vary-ing, but overall unsatisfactory, results; moreover, graftrejection, GvHD and toxicity are frequent complicationsof these strategies, and mortality rate due to EBV-LPDremains high.160,161

Normal EBV seropositive individuals have a high fre-quency of circulating virus-specific cytotoxic T-lympho-cytes precursors, which control outgrowth of EBV-infect-ed B-cells. Since EBV-LPD in immunocompromised hostsappears to stem from a deficiency of virus-specific cyto-toxic activity, it is reasonable to hypothesize that anadoptive immunotherapy approach with donor-derived

T-lymphocytes could be able to prevent unchecked lym-phoproliferation and eradicate established disease. In1994, the Sloan Kettering group first demonstrated that,through the infusion of unselected peripheral bloodmononuclear cells from a donor, 5 patients given HSCTwith post-transplant EBV-LPD obtained remission of thedisease.164 However, this treatment was associated withdevelopment of clinically relevant GvHD and 2 patientsof inflammatory-mediated lung damage, leading to res-piratory failure.

A further refinement of this approach was achieved byRooney and colleagues, who generated EBV-specific T-cell lines from donor lymphocytes and infused them asprophylaxis against EBV-LPD in patients given T-celldepleted HSCT from HLA-disparate family or unrelateddonors, and, thus, considered at high risk for this dis-ease.165 The infusion of these polyclonal T-cell linesproved to be safe and effective in the prevention of EBV-LPD. Moreover, these cytotoxic cells may also have arole in the treatment of established disease.165 The mostrecent update of this experience confirms that the infu-sion of EBV-specific T-cell lines is highly effective for theprevention of EBV-LPD, since none of 39 patients givena T-cell depleted allograft and treated with this adop-tive immunotherapy developed the disease, as comparedto 7 out 61 transplanted patients not receiving the pro-phylactic treatment.166 Gene marking studies haveshown the persistence of these donor-derived EBV-spe-cific cell cytotoxic lines in patient’s peripheral blood formonths after infusion and their re-appearance afterperiods of apparent non-identifiability during episodesof viral reactivation, this further stressing the impor-tance of helper T-cell function in the persistence oftransferred CD8+ cells.167

The profound immunosuppression necessary for graftsurvival carries a well-recognized predisposition to thedevelopment of viral complications, in particular EBV-LPD, also in recipients of solid organ transplantation.159

An immunotherapy approach to EBV-LPD using autolo-gous in vitro generated EBV-specific cytotoxic lines couldbe an appealing strategy in this cohort of patients. Sup-port for this hypothesis is given by the recently described,although not unexpected, possibility of generating, frompre-transplantation blood samples of EBV-seropositivesolid organ transplant recipients, virus-specific T-celllines which are effective in controlling EBV replicationpost-transplantation.168 However, generation and stor-age of cytotoxic lines for each patient undergoing solidorgan transplantation requires enormous, unavailablelevels of funding, laboratory facilities and workforce. Amore rational strategy is to generate, expand and infuseautologous EBV-specific cytotoxic lines from the periph-eral blood of organ transplant patients presentingincreased EBV-DNA levels after transplantation, which, aspreviously mentioned, are a risk factor for EBV-LPD devel-opment. The feasibility of generating autologous EBV-specific cytotoxic lines from the peripheral blood of organtransplant patients receiving in vivo immunosuppressionfor prevention of graft rejection has been recently

Cell therapy

130


proved.169 Moreover, these cytotoxic T-lymphocytes weredemonstrated to be able to display EBV-specific killing invivo, as proved by prompt viral DNA clearance, withoutaugmenting the probability of graft rejection. A peculiarproblem, fortunately not particularly common, is that ofEBV-seronegative patients, who develop primary EBVinfection after solid organ transplantation. In fact, inthese patients, in vitro generation of virus-specific T-celllines able to control EBV-driven B-cell proliferation canbe particularly complicated, time-consuming and some-times unsuccessful.

Autologous EBV-specific cytotoxic lines with demon-strated anti-viral activity in vitro and in vivo may alsohave a role in the treatment of other EBV-associated pri-mary malignancies: for example, 40-50% of patientswith Hodgkin’s disease tumor cells are EBV-antigen pos-itive and may therefore be suitable targets for virus spe-cific cytotoxic lymphocytes.160,170 A recently reportedstudy provides further support for this possibility, docu-menting that, although more complicated than in nor-mal donors, generation of EBV-specific cytotoxic lines isfeasible in a relevant proportion of patients with EBV-positive Hodgkin’s disease.171 These lines retained theirpotent antiviral effects in vivo and persisted for morethan 13 weeks in patients with relapsed Hodgkin’s dis-ease.171 Whether this approach of adoptive immunother-apy will become an adjunctive treatment option forpatients failing to gain benefit from conventionalchemotherapy remains to be proved in prospective clin-ical trials.

Finally, it should be mentioned that adoptive transferof cytotoxic T-cell response could be of value also inthe prevention or treatment of other viral infections thatcause morbidity and mortality in immunocompromisedpatients. In this regard, pre-clinical studies are under-way to establish systems for generating cytotoxic T-cellresponses to adenovirus.160,172

Genetically engineered donor lympho-cyte infusion for treatment of leukemiarelapse and as a means of acceleratingimmunologic reconstitution in patientsgiven transplantation of hematopoieticprogenitors

Tumor recurrence is the major cause of treatment fail-ure of autologous bone marrow transplantation.173,174

Indeed, the rate of tumor relapse is lower when trans-plantation is performed between matched unrelated ormismatched family member donor and recipients. It isnow established that the curative potential of allo-BMTis represented by the additional effect of high dosechemo-radiotherapy in addition to the presence of allo-geneic T-lymphocytes that are responsible for theGVL.175,176 However, the therapeutic impact of allogene-ic BMT is limited by the inevitable occurrence ofGvHD.177 Severe GvHD can be circumvented by the in vit-ro removal of T-lymphocytes from the BMT.175 Howev-er, recipients of depleted marrow have delayed immune

recovery, and increased incidences of viral infectionsand tumor relapse.178,179

Recent studies have shown the clinical efficacy of theadoptive transfer of immune effectors specific for viralantigens153,195,167 in patients who underwent BMT. In thiscontext gene transfer of a marker gene provides a meansof evaluating the survival, homing and efficacy of theinfused cells.

In marrow-transplanted recipients, lymphoprolifera-tive disorders associated with EBV, a human herpes virusthat normally replicates in epithelial cells of the oropha-ryngeal tract, occurs in 5-30% of the treated patients.EBV-LPD are usually malignant B-cell lymphomas ofdonor origin, which may be either polyclonal or mono-clonal. The latter have a rapidly progressive, fulminatingand fatal course.180,181 The transformed B cells expressvirus-encoded latent cycle nuclear antigens, latent mem-brane proteins, and a number of cell adhesion molecules.Most of these viral proteins are recognized as antigensby the immune system of a normal individual.182 In thenormal host, in fact, EBV-induced lymphoid prolifera-tion is controlled by EBV-specific and MHC-restricted T-lymphocytes, MHC-unrestricted effectors and by anti-bodies directed toward specific viral antigens. Since alimited number of specific cytotoxic T-lymphocytes isrequired for controlling EBV-transformed B-lymphocytesin normal individuals, the administration of donor lym-phocytes for the occurrence of EBV-LPD in recipients ofT-cell depleted bone marrow transplantation could con-trol this severe complication by providing the patientwith donor immunity against EBV.164,183 Successfulregression of the disease, documented histologically andby full clinical remission, has been achieved by the infu-sion of unmanipulated donor leukocytes.164 However,acute or mild chronic GvHD developed in all the patientswho responded to the treatment.164

To prevent GvHD, Brenner’s group has evaluated theuse of EBV-specific CTL rather than unmanipulated Tcells. Donor derived EBV-specific CTL have been gener-ated in vitro by stimulation with irradiated donor-derived EBV-infected lymphoblastoid cell lines (LCL).184

The polyclonal effector populations were predominant-ly CD8+ with a varying number of CD4 and showed spe-cific cytotoxic activity toward the EBV-infected targetcells. In order to investigate the long-lasting survival ofthe injected cells, the anti-EBV effectors were markedwith the neo-gene before administration.

Neo-marked cells were detected in circulation for atleast 10 weeks after the injections.165 Moreover, theinfusions allowed the establishment of a population ofCTL precursors that could be activated to proliferate byin vivo or in vitro challenge with the virus.166 The authorsshowed that EBV-specific CTL lines expressing the neo-marker, could be derived from patient's peripheral bloodlymphocytes (PBL) for up to 18 months, by in vitro res-timulation with the autologous EBV-lines.166

These findings support a more widespread use of anti-gen-specific CTL in the treatment of infections and can-cer. Their use may extend in the near future to other dis-

C. Bordignon et al.

131


eases which express well-known antigens that couldserve as target of CTL therapy (e.g. Hodgkin’s disease andnasopharyngeal carcinoma).

The adoptive transfer of in vitro stimulated effectorsachieves clinical results without causing the appear-ance of GvHD. However, the application of this strate-gy to a large number of allo-BMT treated patients, espe-cially in prophylaxis protocols has some limitationsrelated to the in vitro manipulation necessary for thegeneration of specific effectors (e.g. availability ofdonor-EBV lines; in vitro stimulation and expansion ofantigen-specific effectors). An alternative approach wasproposed in 1994 by the S. Raffaele Hospital group.185,186

Their protocol was aimed at maintaining the potentialof the infusion of polyclonal cell lines while providing aspecific means to control acute GvHD. To this aim theytransduced donor lymphocytes by a retroviral vectorcontaining a suicide gene for in vivo selective elimina-tion of the infused lymphocytes.

It was previously shown that introduction of a geneencoding for a susceptibility factor, a so-called suicidegene, makes transduced cells sensitive to a drug notordinarily toxic.187,188 A series of retroviral vectors carry-ing a suicide gene for ganciclovir-mediated in vivo selec-tive elimination of the infused lymphocytes wasdesigned. The vectors carried either an HSV-thymidine-kinase-neo (Tk-neo) fusion gene, coding for a chimericprotein for both negative and positive selection, or theHSV-Tk gene alone.189

A crucial prerequisite for the application of this strat-egy in the clinical context is the transduction of allinfused donor lymphocytes. For this purpose, thedesigned retroviral vectors also carried a gene encodinga modified (non-functional) cell surface marker notexpressed on human lymphocytes. Positive immu-noselection of the transduced cells190 by the use of thecell surface marker resulted in virtually 100% gene-modified lymphocytes.

Based upon the preclinical data described above, aclinical protocol was developed186 for the use of donorlymphocytes transduced by the SFCMM-2 retroviral vec-tor for transfer and expression of two genes: the HSV-Tk gene that confers to the transduced PBL in vivo sen-sitivity to the drug ganciclovir, for in vivo specific elim-ination of cells potentially responsible for GvHD; and amodified (non-functional) form of the low affinityreceptor for the nerve growth factor gene (∆LNGFr), forin vitro selection of transduced cells and for in vivo fol-low-up of the infused donor lymphocytes.

Increasing doses (beginning at 13106/kg) of donorPBL were infused into several patients affected byhematologic malignancies who developed severe com-plications following a T-cell depleted BMT from theirHLA-identical related donors. After the infusion, thetransduced lymphocytes could be detected in the bloodof patients by cytofluorimetric and PCR analyses. In par-ticular one patient affected by an EBV-LPD, showed aprogressive increase in the number (up to 13.4% of thetotal PBL) of infused marked lymphocytes that was

accompanied by a complete clinical response. However,signs of acute GvHD, confirmed by skin biopsy, wereobserved approximately four weeks after the infusionof the transduced-donor lymphocytes. The intravenous(i.v.) administration of two doses of ganciclovir (10mg/kg/day) quickly resulted in elimination of markeddonor PBL, and near resolution of all clinical and bio-chemical signs of acute GvHD.187

As mentioned before, when comparable preparativeregimens are employed, the rate of tumor recurrencesafter autologous BMT is significantly higher than therate observed after allogeneic BMT. GvHD develops in50-70% of patients undergoing allogeneic BMT. Theeffectors of such response are thought to be maturedonor lymphocytes from the marrow graft that respondto the foreign major and/or minor histocompatibilityantigens of the recipient and also recognize and destroythe tumor cells. In fact, patients who underwent matureT-cell-depleted allogeneic BMT have a lower rate ofGvHD but also a higher rate of leukemia relapses.178,179

The infusion of donor lymphocytes, early after T-cell-depleted allogeneic BMT, increases the incidence ofGvHD without improving the control of leukemia.191

However, a delayed transfusion of donor lymphocytes,when graft tolerance is established, seems to be moreeffective in preventing and treating tumor relapses.

Indeed the delayed administration of donor lympho-cytes has recently become a new tool for treatingleukemic relapse after BMT. Patients affected by post-BMT recurrence of chronic myelogenous leukemia, acuteleukemia, lymphoma, and multiple myeloma couldachieve complete remission after the infusion of donorleukocytes without requiring cytoreductive chemother-apy or radiotherapy,106,192-194 even though the responserate of patients with acute leukemia, non-Hodgkin’slymphoma and multiple myeloma is significantly lowerthan that of patients affected by chronic myelogenousleukemia. Although the delay in the administration of Tlymphocytes is expected to reduce the risk of GvHD, thisrisk is still present at higher doses of donor T-cells.116

Therefore, as described above, a clinical protocol wasdeveloped, for the use of donor lymphocytes transducedby the SFCMM-2 retroviral vectors186 for transfer andexpression of the HSV-Tk gene, and the cell surfacemarker ∆LNGFr, for in vitro selection of 100% trans-duced cells and for in vivo follow-up of the infuseddonor lymphocytes.190

In a phase I-II study, eight patients affected by hema-tologic malignancies who developed severe complica-tions following an allogeneic T-cell depleted BMT,received escalating doses of donor PBL transduced bythe described retroviral vector.133 After gene transfer,transduced cells were selected for the expression of thecell surface marker ∆LNGFr by the use of specificimmunobeads and the proportion of transduced cellswas assessed by cytofluorimetric analysis.190 In this study,we made the following observations: 1) transduced cellssurvived long-term in vivo and were detectable by cyto-fluorimetric analysis and PCR in high proportions (up to

Cell therapy

132


13.4% of circulating PBL) and long-term (up to 6months); 2) three patients showed complete response,three patients had partial response, one progressed withno response, and one patient could not be evaluated; 3)three patients developed GvHD that required ganciclovirtreatment; 4) ganciclovir-mediated elimination of trans-duced cells resulted in near resolution of all clinical andbiochemical signs of acute GvHD. Data from this study133

indicate that genetically modified cells maintain their invivo potential to develop both anti-tumor and GvHDeffect, and may represent a new potent tool for exploit-ing anti-tumor and anti-host immunity, while providinga specific means for eliminating acute GvHD, in theabsence of any immunosuppressive drug.

A potential limitation of the clinical approachesdescribed could be the development of a specificimmune response against vector-encoded proteins,which might allow the selective elimination of thetransduced cells by the host immune system. For somegene products, such as the hygromycin-thymidinekinase (Hy-Tk) fusion protein, a specific immuneresponse, able to eliminate large numbers of transducedcells in less than 48 hours, has been described in HIV-patients.195

We observed that immune recognition and killing ofcells transduced by retroviral vectors is a more generalphenomenon related to the foreign nature of the pro-teins expressed by the injected cells. Indeed, cellsexpressing the widely used marker gene neo and theHSV-Tk gene are targets of a strong immune response,while the endogenous proteins (e.g. the cell surfacemarker ∆LNGFr) are not recognized, even if ectopicallyexpressed in a context which is otherwise extremelyimmunogenic.196 The relative immunogenicity detectedfor the three vector-encoded components (none by∆LNGFr, low by HSV-Tk, high by neo) clearly outlinedthe modifications of this type of gene therapy. Since neois the only component not-strictly necessary for thestrategy and can be efficaciously replaced by the surfacemarker for all in vitro handling and selection,189,197 theimmunogenicity of the new neo-less vectors should bereduced.

The clinical results obtained with gene modified donorlymphocytes, for the treatment of hematologic relaps-es and EBV-lymphoproliferative disorders, suggest thepotential use of this approach.133 The transfer of a sui-cide gene, that allows selective and specific eliminationof effector cells of GvHD may allow full advantage to betaken of the beneficial effect of allogeneic lymphocyteswith the possibility of eliminating all unwanted effectsof GvHD in the absence of toxic side effects. A largescale application of this strategy will increase the num-ber of patients who could potentially benefit from allo-geneic BMT by allowing the use of less compatible mar-row donors.

With regard to the immune recovery associated withthe genetically-engineered donor lymphocytes, our grouphas recently obtained in vitro data demonstrating thatgenetically-engineered donor T-cells maintain a normal

TCR Vb immune repertoire and retain antigen-specificlytic activity against an allogeneic target or an autolo-gous EBV cell line at cytotoxic T-cell precursor frequen-cies comparable to unmodified lymphocytes. In the lightof this in vitro evidence, and our previous clinical appli-cation,133 a clinical trial, based on the prophylactic infu-sion of 13107/kg HSV-Tk transduced T-cells six weeksafter T-cell-depleted bone marrow transplantation, wasdeveloped. In the first five treated patients we docu-mented the presence of various proportions of transducedcells in the peripheral blood. In particular, genetically-engineered donor lymphocytes were responsible for anti-viral immune reconstitution in one patient. CD3+ lym-phocytes began to appear in the circulation of thispatient two weeks after the infusion of HSV-Tk T-cells. Allthe CD3+ lymphocytes were genetically engineered asdocumented by the expression of the cell surface mark-er ∆LNGFr. These cells retained a polyclonal TCR reper-toire and were probably responsible for the clearance ofa persistent CMV antigenemia. Indeed, the CMV anti-genemia dropped below levels which could be detectedby PCR shortly after the appearance of circulating genet-ically-engineered CD3+ T cells in the absence of anyantiviral drug therapy.198 These data, if confirmed in alarger number of patients with longer follow-up, suggestthat in addition to the anti-tumor activity, the infusionof genetically-engineered donor lymphocytes may play arole in restoring immunity against opportunistic infec-tions early after allogeneic BMT.

Dendritic cells as natural adjuvants incancer immunotherapy

Among professional antigen presenting cells (APC),dendritic cells (DC) are specialized in capturing and pro-cessing antigens into peptide fragments that bind tomajor histocompatibility complex molecules. DC are themost potent stimulators of T-cell responses and theyare unique in that they stimulate not only memory butalso naive T-lymphocytes. Thus, DC appear critical(nature adjuvants) for the induction of B-and T-cell-mediated immune responses. Recent evidence in exper-imental models supports the role of DC for immuniza-tion strategies aimed at stimulating specific anti-tumorimmunity.

In this section we will briefly review:1. the biological characterization of DC;2. different strategies for ex vivo generation of DC;3. methods for the efficient delivery of tumor associat-

ed antigens (TAA) to DC;4. the use of DC for cellular immunotherapy.

Biological characterization of dendriticcells

DC are widely distributed in the body and are partic-ulary abundant in tissues that interface the environ-ment (i.e. Langerhans cells in the skin and mucous mem-branes) and in lymphoid organs (interdigitating DC)where they act as sentinels for incoming pathogens.

C. Bordignon et al.

133


Cell therapy

Inflammatory signals such as TNF-a and IL-1b as wellas bacteria, bacterial products (LPS) and viruses inducemigration of antigen-loaded DC from the peripheral tis-sues to secondary lymphoid organs. During migration,DC mature and upregulate MHC, adhesion and co-stim-ulatory molecules, thus strongly augmenting their abil-ity to prime T-cells.199-204

The functional activity of DC derives from a numberof properties of these cells (Figure 2). Their dendriticshape, along with the high level of expression of certainadhesion molecules and integrins (LFA-3, ICAM-1,ICAM-3), increases the area of contact with the effec-tor cells of the immune system.205 DC strongly expressthe HLA class II molecules -RD, -DQ and -DP and co-stimulatory molecules (CD80, CD86 and CD40) whichactivate their ligands on T-cells (CD28, CTLA-4 andCD40L), thus providing the second signal strictly neces-sary to induce a proliferative response, rather than tol-erance, upon antigen recognition.199 In addition, DC pro-duce a number of cytokines including IL-12 which pro-motes a cytotoxic immune response by inducing the dif-ferentiation of TH0 cells to IFN-g and IL-2-producingTH1 cells.206,207 It has recently been demonstrated thatupon Ag recognition, T-helper cells activate DC viaCD40-CD40L interaction and activated DC are then ableto trigger a cytotoxic response from T-killer cells.208-210

However, DC are present in peripheral tissues in animmature state unable to prime T-cells. At this stage of

differentiation, they can very efficiently take up solubleantigens, particles and micro-organisms by phagocytosis,macropinocytosis or by the macrophage mannose recep-tor, Fcg and Fce receptors,211 but they lack all the acces-sory signals for T-cell activation. Antigen uptake inducesDC to maturation by up-regulating MHC and co-stimu-latory molecules as well as DC-associated Ag (e.g. CD83and p55) whereas the capacity to capture and process Agis lost. However, full activation of DC is dependent uponthe contact with T-cells by the CD40-CD40L interactionwhich induces the production of IL-12. Thus, the keyfunctions of DC (antigen uptake, T-cell stimulation) arestrictly segregated to subsequent stages of differentiation(Figure 3). It is noteworthy that IL-10212 and vascularendothelial growth factor (VEGF), secreted by cancercells,213 prevent the maturation of DC thus inhibiting theefficient priming of T-cells.

Different strategies for the generation ofDC ex vivo

Circulating CD14+ monocytes represent the most read-ily available source of DC if incubated with appropriatecytokines such as GM-CSF, IL-4 and TNF-a.214, 215 More-over, DC precursors have been isolated within the CD34+

cell fraction in bone marrow, cord blood and steady stateor mobilized peripheral blood.216-221 Also in this case thedifferentiation of CD34+ cells into fully functional DC isstrictly dependent upon stimulation with certaincytokines such as GM-CSF, TNF-a, SCF, FLT3-L and IL-4.

Figure 2. Phenotypic and functional characteristics of dendritic cells. Modified from ref. #199 (Bancherau and Steinman,Nature, 1998).

134


An extensive review of the different types of human DCand their ex vivo generation is beyond the scope of thischapter. However, in view of the clinical use of DC a fewcritical points should be stressed. GM-CSF and IL-4induce the differentiation of non-proliferating CD14+

monocytes to immature DC with a low level of expressionof CD83 and p55 Ag and are largely incapable of prim-ing naive T-cells. These immature DC are not fully differ-entiated and revert to an adherent state if the cytokinesare removed from the culture medium.222,223 The additionof inflammatory cytokines such as TNF-a, IL-1b or PGE2for 1-2 days to the medium containing GM-CSF and IL-4 promotes the maturation of DC and increases the abil-ity of stimulating T-cells. A potential bias toward theclinical use of this culture system is the requirement offetal calf serum (FCS), a xenogenic protein that is con-traindicated for human use. An innovative culture systemfor the generation of mature and functional DC from cir-culating monocytes that uses FCS-free conditions hasrecently been described.222,223 In this system, adherentperipheral blood (PB) cells are cultured for 6-7 days withGM-CSF and IL-4 in the presence of FCS, which is thenwashed out, and subsequently exposed to macrophage-conditioned medium (Mo-CM) and 1-5% autologousplasma for 1-3 days. Mo-CM is very efficient in inducingthe terminal maturation of DC and is prepared by grow-ing T-cell-depleted PB cells on immunoglobulin (Ig)-coat-ed Petri dishes for 24 hours.

Taken together, these findings lead to the conclusionthat immature DC generated from CD14+ cells in thepresence of GM-CSF and IL-4 are well equipped for cap-turing and processing soluble TAA. However, they dorequire a further maturation stimulus (Mo-CM, TNF-a)to exert their stimulatory effect on T-cells. Immature DCare the ideal targets for genetic manipulation using viral

or bacterial vectors which infect non-replicating cells(see below). In this case, the modified pathogens caninduce by themselves the full maturation of DC. In alter-native, mature DC could be used in vaccination protocolsinvolving TA peptides as DC also prime T-cells to foreignAg that bind directly to MHC molecules without priorprocessing.224

As reported above, CD34+ cells can be induced to dif-ferentiate into fully functional DC which resemble cuta-neous Langherans cells.218 The issue of the large scaleproduction of DC from CD34+ precursors has been dis-cussed in detail elsewhere.5 However, very recently thephenotypic and functional characteristics of DC derivedfrom CD34+ cells mobilized into PB or from BM progen-itors have been formally compared.225 The publishedresults indicate that G-CSF mobilizes DC precursors (CFU-DC) with an increased frequency and a higher prolifera-tive capacity than their BM counterparts. This findingtranslates into a higher number of mature DC generatedin liquid culture. Despite pre-treatment with G-CSF, thesecells maintain the same functional capacity of stimulat-ing allogeneic T-cells as BM-derived DC. CD34+ cell-derived DC are also capable of processing and presentingsoluble Ag to autologous T-cells for both primary andsecondary immune responses. The potential clinical use-fulness of autologous serum in place of FCS220 was alsoconfirmed in the same study. Of note, IL-4 was shown tobe capable of modulating DC differentiation from bipo-tent CD34+ cells during the later stages of the culture aspreviously demonstrated for monocyte-derived DC.226

Thus, mobilized CD34+ cells may represent the optimalsource for the generation of DC for cancer immuno-therapy rather than BM precursors. Very recent data indi-cate the mobilization of large numbers of DC precursorsby GM-CSF227 and FLT-3L.228 However, it remains to be

C. Bordignon et al.

Figure 3. Functional propor-ties of dendritic cells at dif-ferent stages of differentia-tion. Pathogens or inflamma-tory cytokines induce thematuration of dendritic cellswhich become activatedupon interaction with T-cellsvia CD40-CD40L.

135


Cell therapy

established whether circulating CD34+ elements are anequivalent source of DC to CD14+ monocytes. In this view,it has recently been demonstrated229 that CD34+ cell-derived DC are more efficient than monocyte-derived DC,from the same patients, in stimulating a specific CTLresponse to Melan-A/Mart-1 peptides.

Delivery of TAA to DCSeveral methods for the efficient delivery of TAA to DC

have been described so far (Figure 4). Their rationale isbased on the finding that tumor cells are often poorlyimmunogenic due to the lack of T-cell recognition, acti-vation and co-stimulation typical of professional APC. Tothis end, Gong et al.230 fused murine DC with the carci-noma cell line MC38 to provide tumor cells with thefunctional characteristics of DC. The fusion cells showedall the phenotypic features of DC and were shown to becapable of preventing tumor growth when the micewere challenged with the cell line. Moreover, treatmentwith fusion cells induced the rejection of pulmonarymetastates.

Several TA peptides which are presented to T-cells inassociation with HLA class I molecules have beenrecently identified and proved to be useful in stimulat-ing an autologous CTL response in vitro and in vivo.However, pulsing DC with peptides may not be optimalfor clinical application because of the strict MHC restric-tion of the immune response and their limited stability.In addition, pulsing with peptides may not induce a T-cell response directed toward tumor cells expressing therelevant Ag. Although DC can be loaded with a cocktailof peptides from different Ag derived from the sametype of cancer (see below), this vaccination approach islikely to limit patient selection on the basis of HLA phe-notype. An attractive alternative is the use of unfrac-tionated tumor-derived proteins, when available (seebelow), apoptotic cells231 or tumor lysates. In the lastcase the obvious disadvantage is the possibility of induc-ing immune responses against self-Ag expressed in tis-sues other than tumor cells.

A further possibility is the transduction of DC withexpression vectors encoding for TAA genes (Figure 4). DCcan be engineered by different means which differ intheir capacity of targeting quiescent cells, stable inte-gration in the genome, infection efficiency and stimula-tion of anti-tumor immunity (Figure 5). Retrovirally-transduced DC constitutively express the relevantsequence and are potent stimulators of a specific T-cellresponse.232 However, retroviral vectors have a relativelylow efficiency of transduction, they can only infectactively replicating cells and carry the theoretical risk ofoncogenic transformation of target cells. Conversely, ade-noviruses infect both quiescent and proliferating cellsand do not integrate into DNA.233 Moreover, supernatantswith a high titer of the virus can be easily obtained.Recently, DC have been transduced with an adenoviruscombined with cationic liposomes showing an infectionefficiency close to 100%.234 The major limitation to theclinical use of adenoviruses is their high immunogenici-ty which induces the production of neutralizing antibod-ies and the rapid development of CTL directed at infect-ed cells.

Vaccinia virus vectors are not oncogenic, do not inte-grate into genome and can be manipulated to carrylarge fragments of heterologous DNA.235 However, theseviruses are toxic for target cells and the viability of DCis approximately 50%. Nonetheless, antigen-specificinhibition of tumor growth has been observed in murinemodels using vaccinia vectors encoding for CEA andMucin-1.236,237 Two phase I clinical trials have been con-ducted to assess the safety of vaccinia virus vectorsengineered to express HPV and CEA genes and to assestheir capacity of stimulating an immune response.238,239

More recently, maturation of DC with neo-biosynthesis,translocation and stabilization of MHC molecules onthe cell surface and efficient induction of both CD4 andCD8 T-cell activation has been induced by infection withbacterial vectors.240 As a result, a model Ag (ovalbumin)expressed on the surface of recombinant Streptococco

Figure 4.

136


Gordonii, is processed and presented on MHC class Imolecules 106 times more efficiently than soluble OVAprotein. Therefore, bacterial vectors are potentially use-ful means of delivering exogenous Ag to DC for stimu-lating a tumor-specific CTL response. A differentapproach has been taken by Boczkowsky et al.241 whotransfected DC with the total RNA extracted from tumorcells and combined it with cationic lipid to enhance theinfection efficiency. Similarly to the use of tumor lysates,this strategy can be applied in those situations in whicha tumor-specific antigenic marker is lacking; the majorconcern is the increased risk of autoimmune reactivity.

DC for cellular immunotherapyThe central role of DC in stimulating a tumor-specif-

ic immune response is well established in vitro and invivo in animal models.232,241-246 Whereas murine DCpulsed with TA-proteins or peptides or transduced withTAA genes have induced both the rejection of challengetumor cells and the regression of established cancers, itremains to be determined which of the several strate-gies proposed for cellular immunotherapy is the mostefficient. It may well be that different tumors requiredifferent approaches.

In humans, initial studies were performed in patientswith melanoma using DC pulsed with MAGE pep-tide.247,248 The infusion of loaded DC induced the migra-tion of MAGE-specific CTL to the site of injection andincreased the frequency of circulating tumor-specificCTL. More recently, Nestle et al.249 have treated advancestage melanoma patients with intranodal injection ofpeptides or tumor lysates-pulsed DC according to theHLA profile of the patient. The authors reported thestimulation of a peptide-specific T-cell response in all

cases. Moreover, in 5/16 patients an objective clinicalresponse was observed. In this study, DC were generat-ed ex vivo from monocyte precursors in the presence ofIL-4 and GM-CSF and directly injected into an inguinallymph node to reach T-cell rich areas.

Tumor-specific peptides (fragments of prostate spe-cific antigen, PSA) have also been used to pulse autolo-gous DC in prostate cancer patients refractory to hor-mone-therapy.250 Seven out of 51 patients showed a par-tial response while none of the patients in the controlgroup, injected with peptides alone, showed any clinicalbenefit. In B-cell malignancies, the patient-specific idio-type (Id) gene sequence and its protein product representthe optimal targets for vaccination strategies as previ-ously shown in murine models251,252 and humans.253 Hsuet al.254 have reported on the treatment of 4 patientswith low-grade non-Hodgkin’s lymphoma (NHL), resis-tant to conventional chemotherapy or who had relapsed,with DC pulsed with the Id as soluble antigen. A tumor-specific T-cell-response was observed in all cases cou-pled, in one case, with the regression of tumor burden.At the time of writing, 16 patients have been treatedand a tumor-specific cellular response has been found in8 individuals (R. Levy, personal communication). Thesame strategy of targeting the Id has been proposed bythe same group for inducing a T-cell immune responsein multiple myeloma patients.255

In contrast to the strategy used by Nestle et al.249 in thispreliminary trial DC were freshly isolated from the PB bysubsequent enrichment steps and were reinfused intra-venously. Although a much larger number of DC wereinjected in NHL patients compared to melanoma patients(3-203106 DC vs 13106), this approach raises concernsabout both the efficacy of uncultured PB DC of efficiently

C. Bordignon et al.

Figure 5.

137


stimulating T-cells and the capacity of Id-loaded APC toreach secondary lymphoid organs to prime T-cells, escap-ing the entrapment of the pulmonary apparatus.

Future directionsThe few clinical data available so far have barely pro-

vided the proof of principle that autologous DC gener-ated ex vivo and reinfused into cancer patients areeffective in stimulating an anti-tumor immune response.This is the result of the complexity of the interplaybetween different cellular populations involved in tumorimmunity. In addition, cellular immunotherapy with DChas yet to be standardized. As mentioned above, crucialissues such as 1) the choice of the most suitable TAA tostimulate an immune response; 2) the use of solubleproteins/peptides or DC engineered with expression vec-tors; 3) the optimal source for the generation of DC andthe number of APC needed to promote a clinical effect;and 4) the most effective route of administration of DC,are points which still need to be solved. At this stage,relying for the most part on animal studies, we can onlyconclude that DC-based immunotherapy holds promis-es of exerting a potent anti-tumor effect in humans.

Oral vaccination by in vivo targeting of DCA simple approach to targeting APC in vivo is to use

attenuated bacterial vectors, such as those commonlydeveloped to control infectious diseases. They usuallyenter the host through the oral route and selectivelyreplicate within macrophages and DC. Listeria monocy-togenes is a promising vaccine carrier that naturallyinfects APC, and may deliver immunogens to both MHC-I and II pathways of antigen processing and presenta-tion.256 Furthermore, this bacterium may constitute perse an excellent danger signal for the immune system,since it stimulates the innate immune response to pro-duce cytokines (e.g. IL-12) and mediators (e.g. nitricoxide) that enhance antigen presentation. In addition, itpromotes a TH1-type cellular response, which is main-ly associated with the eradication of tumors and intra-cellular parasites. Most of these features are also sharedby Salmonella typhimurium-based carriers.

The ideal vaccine carrier should maintain its immu-nogenicity intact, being attenuated enough to allow itsuse in humans. However, the safety profile of a vaccinedestined for human use also requires the absolute sta-bility of the mutant phenotype, which can only be guar-anteed by the generation of chromosomal deletionmutants. Furthermore, the release of recombinant micro-organisms under uncontrolled conditions makes the lackof antibiotic resistance markers essential. Mutation ofgenes involved in bacterial spread and survival are thebest targets for attenuation.

The recent progress in Listeria and Salmonella genet-ic manipulation and the availability of suitable in vitroand in vivo models, make these micro-organisms veryattractive vaccine delivery systems.

For example, attenuated Listeria monocytogenes car-rier strains expressing the b-galactosidase (b-gal) mod-el antigen can prevent outgrowth of an experimental

tumor in BALB/c mice by inducing a specific immuneresponse against the b-gal TAA.257 Similarly, a live atten-uated AroA- auxotrophic mutant of Salmonella typhi-murium (SL7207) has been used as a carrier for thepCMVbb vector that contains the b-gal gene under thecontrol of the immediate early promoter of cytomega-lovirus (CMV). After a primary immunization and threeorally administered boosts at 15-day intervals, a Sal-monella-based vaccine induced both cell-mediated andsystemic humoral responses to b-gal. These experimentssuggested that insertion of a plasmid containing anexpression cassette into a Salmonella-carrier allowedDNA immunization and specific targeting of antigenexpression to APC, in vivo, through oral immunization.To prove that the transgene was actually expressed byAPC cells as a function of a eukaryotic promoter thegreen fluorescent protein (GFP) was placed under thecontrol of either the eukaryotic CMV or a prokaryoticpromoter and spleen cells from treated mice were ana-lyzed by cytofluorometric analysis.

GFP was detectable in both macrophages and DC, butnot in other splenocytes, of mice treated with Salmo-nella containing the CMV-plasmid, 28 days after thefirst vaccine administration, whereas it was unde-tectable in spleen cells of mice receiving the Salmonel-la containing the constitutive prokaryotic promoterwhich directs GFP synthesis only within the carrier.258

GFP expression in DC highlights the possibility of load-ing DC without the need for ex vivo manipulations andopens up the possibility of administering a cancer vac-cine orally. Oral vaccination is viewed as an easier andmore acceptable strategy for patients especially in aphase in which they are disease-free.

Leukemic cells as antigen presentingcells

Tumors may escape immune detection and killingthrough a variety of mechanisms affecting the capacityof either presenting tumor antigens or fully activating T-cells.258,260 In particular, tumor cells are likely to preventa clinically evident cytotoxic T-cell response because ofthe absence of a specific antigenic tumor peptide, orbecause they lack HLA molecules, or co-stimulatory mol-ecules on their surface. In this last case the patient’s T-cells might become anergic and tolerate tumor cells.Alternatively, neoplastic antigens may induce a clonaldeletion of thymocytes,261 or tumor cells expressing Fasmolecule may be responsible for an apoptotic T-cell dele-tion through Fas:FasL interaction.262 So far, differentimmunologic strategies aimed at overcoming thesedefects by inducing or improving the antigen presentingfunction of tumor cells have been demonstrated in exper-imental models,263,264 and the hypothesis that leukemiccells may become efficient APC by changing their phe-notype or by differentiating into DC-like cells has beentested. A first example was shown in B-cell neoplasmssince it is well known that normal B-cells may presentantigen to T-cells265 and that cognate interactionsbetween B- and T-cells may induce either a T-cell prolif-

Cell therapy

138


eration and an enhanced T-helper activity to cytotoxic T-cells,266 or T-cell clonal unresponsiveness.267 The trigger-ing of the CD40 receptor on the surface of APC increas-es the expression of adhesion and co-stimulatory mole-cules both in vitro and in vivo.269,269 Thus, the possibilityof modifying the phenotype and the APC function ofCD40+-chronic lymphocytic leukemia B (CLL-B) cellsthrough the CD40:CD40L interaction was demonstratedshowing that this pathway induces the upregulation ofCD80 and CD86 on CLL-B cells and the triggering of a T-cell proliferative response.270,271 These results support theidea that induction of B7 molecules on CLL-B cells, eitherby T-cell-contact and growth factors,270, 272 or by genetransfer methods273 may be a potential clinical vaccine-therapy capable of eliciting efficient anti-leukemicimmune responses. Similar approaches may also apply toB non-Hodgkin’s lymphomas (B-NHL). Studies in experi-mental models indicated that CD40 stimulation mayresult in the inhibition of lymphoma cell growth in vivo,274

and in the up-regulation of adhesion receptors and co-stimulatory molecules on lymphoma cells in vitro.275,276

Interestingly, follicular B-NHL cells which express CD40and low levels of B7-2 fail to present alloantigen, butafter activation via CD40 they express higher levels of B7-1 and LFA-3 and alloreactive T-cells respond to tumorcells efficiently.276 Finally, encouraging results have alsobeen obtained in pre-B acute lymphoblastic leukemia139

in which approximately 50% of the cases blast cells havebeen reported to express CD86 but not to induce tumorrejection, and B7-blasts determine an immunologic tol-erance of the tumor. Nonetheless, this study showed thatpre-activation of blast cells via CD40, or cross-linkingCD28, or signaling through the common g chain of theIL-2 receptor on T-cells can prevent T-cell tolerance. Theauthors hypothesize at least two possible mechanisms toexplain the induction of lymphocyte unresponsiveness:first, they propose that at the time of initial transforma-tion, clonogenic pre-B acute leukemia cells may notexpress CD86 thus inducing a T-cell anergy that could notbe reversed by following expression of CD86 on a blastcell fraction; second, they suggest that marrow microen-viroment may play a role in modulating T-cell immunityby secreting negative regulators, as previously shown inexperimental models.277,278 However, after co-stimulationby either B7 transfectants or professional APC, autolo-gous antileukemic cytotoxic marrow T cells can be gen-erated upon contact with CD40-stimulated pre-B acuteleukemia cells.140

All these data on B-cell neoplasms strongly suggestthat poor tumor immunogenicity may depend on boththe quality and the quantity of accessory moleculesrequired for T-cell stimulation. However, future thera-peutic strategies aimed at stimulating the CD40 recep-tor, or at directly transducing B7 molecules on chronic oracute leukemia B-cells will facilitate the ex vivo expan-sion of specific anti-tumor cytototoxic T-cells. Normalmyeloid CD34+ progenitors include a small subset ofAPC279, 280 that are committed precursors of themacrophage/dendritic lineage.281 In fact, both marrow

and peripheral blood CD34+ cells, and circulating mono-cytes can be utilized to obtain large numbers of dendrit-ic cells in vitro. Due to the relevance of co-stimulatorymolecules on tumor cells for the generation of anti-tumorimmune responses, the hypothesis of whether even acuteor chronic myelogenous leukemic cells might differenti-ate into dendritic cells in vitro and become immunogenichas been addressed by several groups. Alternatively,transduction of co-stimulatory molecules on leukemicmyeloblasts has been attempted in experimental modelsto generate specific cytotoxic responses. Both theseapproaches require that TAA are expressed and exposedon HLA molecules, and it is likely that genetic alterations,such as chromosomic translocations, might result in theappearance of pathologic peptides, specific for eachacute or chronic leukemia and potentially immunogenic.Chronic myelogenous leukemia may represent an optimalcandidate for antitumor vaccine strategies since severalreports have shown that the bcr-abl fusion protein canbind to defined HLA class I and class II molecules282-286

and also that dendritic cells generated in vitro from CMLpatients still carry the t(9;22).287,288 In this latter study, infact, CML cells that were incubated with GM-CSF, IL-4and TNF-a developed DC phenotypic and functional char-acteristics inducing autologous cytotoxic T-cells capableof directly lysing leukemic cells and of inhibiting CMLcolony growth in vitro. Further studies suggested thatCML DC-stimulated anti-leukemic T-cell reactivity is dueto an oligoclonal T-cell response and develops in an HLA-restricted manner.289 Dendritic cells can be generatedeven from CD34+ CML marrow progenitors in the pres-ence of GM-CSF, TNF-a and IL-4, and after 7-10 days ofculture they are Ph+, express high levels of HLA mole-cules and co-stimulatory receptors and induce a T-cellproliferation 10-30 fold higher than unprocessed marrowcells.290 Nonetheless, it is likely that different culture sys-tems may be required for efficient in vitro generation ofDC when using CML-CD34+ cells rather than normalprogenitors, since the former show a lower DC clono-genic activity but both their expansion and their differ-entiation can be significantly improved by prolonging theduration of culture in the presence of specific growthfactors.291

When a neoplastic event affects undifferentiated ormore mature progenitors of the granulocytic and/ormacrophage lineage an AML develops, and we can dis-tinguish different subtypes of AML on the basis of mor-phologic and phenotypic characteristics. The identifica-tion of AML cells with some phenotypic affinities to DC,such as the expression of the CD1a marker,292 or derivingfrom a monocytic/dendritic cell progenitor,293 has beenattempted in the past. Indeed in this latter study, cellsfrom an AML, FAB M2 patient were shown to differenti-ate into terminal DC with potent alloantigen presentingcapacity after in vitro culture with GM-CSF, TNF-a, SCFand IL-6. Similar results were achieved by culturing fresh-ly isolated AML cells with GM-CSF, IL-4 and IL-13 for 7days.294 Alternatively, restoration of anti-tumor immunecontrol can be attempted by identifying peptides, such as

C. Bordignon et al.

139


PR-1 derived from proteinase 3,135 that could be capableof inducing HLA-restricted cytotoxic T-lymphocytes tolyse fresh leukemic cells, or by engineeing leukemic cellsto induce either the expression of co-stimulatory mole-cules or the production of cytokines. The role of B7-1 indeveloping protective immunity was initially tested in amouse model in which the injection of a myeloid cell linetransfected with the bcr/abl gene was rapidly lethal,while prolonged survival was observed only in mice thatreceived the cell line co-transfected with the B7-1gene.295 Moreover, the same model was used to test therole of both B7-1 and B7-2, suggesting that B7-1 may bemore effective than B7-2 in obtaining an efficient in vivoanti-leukemic response.296 The potential advantage of B7-transduced blasts was confirmed by using primary AMLcells instead of a cell line; a CD8+ T-cell dependent andB7:CD28-mediated anti-leukemia activity was docu-mented.297 A recent study compared the in vitro immuno-genic activity of human AML cells cultured with GM-CSF,IL-4 and TNF-a, or transfected with CD80.298 Both theseapproaches resulted in an enhanced T-cell response in amismatched primary MLR, however, only B7-1 trans-duced AML cells stimulated a strong immune response ofT-cells from an HLA identical bone marrow donor, andgenerated leukemia reactive CD4+ T-cell lines and clones.Interestingly, this model allowed the authors to observeCD80+AML-mediated T-cell responses that can be direct-ed against the patient’s minor histocompatibility anti-gens or tumor-specific antigens.

Although B7-1 and B7-2-engineered tumor cells

could play a pivotal role in anti-leukemia immuno-therapy strategies, there is evidence that transductionof other receptors299 or cytokines300-303 might, at least,co-operate with B7 molecules in the antigen present-ing capacity of neoplastic cells.

Genetically modified cells as vaccinefor the active immunotherapy of cancer

Non-specific approaches to cancer immunotherapyprobably date back to the beginning of the 18th centu-ry and originated from the observation of sporadic,spontaneous remission of tumors in patients who suf-fered severe bacterial infection. This observationprompted Dr. William B. Coley to begin, in 1891, to treatpatients with soft tissue sarcoma with a mixture ofGram positive and negative bacteria: Coley’s toxins.

This empirical approach was enforced by Shear’s dis-covery that endotoxins were active components respon-sible for tumor hemorrhagic necrosis. Furthermore, thefinding that bacillus Calmette-Guérin (BCG) increasedresistance to tumor transplants in mice led to clinicalapplication of BCG which, together with Streptococcus-derived OK-432, is a strategy used to this day.

The anti-tumor effects obtained by treatment with BCGand derivatives are largely dependent on indiscriminatenecrosis of tissues containing mycobacterium (the Kochphenomenon). The discovery of cytokines explained most ofthe phenomena induced by microbial products andcytokines were then used with the initial hope of copyingthe positive effects of such bacterial products while avoid-

Cell therapy

Figure 6.

Active immunization for cancer:strategies are based on whether tumor antigenshave been molecularly defined.

Known, limited repertoire of TAA

Large repertoire ofunknown antigens

Gene, protein/peptide Whole tumor cellsCell fragments RNA

Loading DC

Directly in vivoEx vivo

Gene

DNA vaccination

Bacterial vectors

Needs adjuvant:plasmid sequencesmuscle injury

140


C. Bordignon et al.

ing the negative ones.More recently, the discovery of Th1 and Th2 distinct

pathways of T-cell maturation helped to explain protec-tive and non-protective BCG-induced cell-mediatedimmune reactions in tuberculosis, phenomena that havecorrelates with protection against cancer. In the pres-ence of a Th1 deflected immune response, the effect ofTNF-a is not that of large necrosis which is, rather, thecharacteristic of inflamed tissues of a Th2 type ofresponse, in this case extremely sensitive to TNF-a.304

Cytokines deflecting the immune response to a Th1 orTh2 type of response may drive the type of immuneresponse to cancer cells, escaping the simple definition ofTh1 promoting and Th2 inhibiting anti-tumor immunity.Rather, a strong Th1 as well as a strong Th2 responsemay induce tumor destruction and immune memory withthe same efficacy although through different mecha-nisms (see below). Moreover, genetic background mayinfluence the ability to mount a Th1 or Th2 response, asshown in murine models.

Microbial products have mainly local effects whichmay be reproduced and improved by local injection ofrecombinant cytokines. Experiments in non-tumor sys-tems have shown that IL-2 offsets antigen recognitionand overcomes tolerance. Thus cytokines could be usednot only to stimulate tumor destruction but also toimpair tolerance and activate effective and specificimmune recognition of TAA.

Identification and cloning of the long elusive TAA,especially from human melanomas,305 pointed tumorimmunotherapy to a general systemic response and, thus,the use of cytokines shifted from that of being responsi-ble for local tumor debulking to that of being an aid totriggering and boosting the immune response to TAA.

In addition to antigens triggering the T-cell-receptor(TCR) of T-lymphocytes, optimal T-cell response alsorequires co-stimulatory molecules, as detailed above.

Cytokines, co-stimulatory molecules and severalcloned TAA are now available: how can we use them toprovide an effective immunotherapeutic approach tocancer patients?

Two major strategies are envisaged (see Figure 6): one,already described, takes advantage of antigen avail-ability in the forms of genes, proteins or peptides and ofthe standardized methods of obtaining DC from periph-eral blood in large quantities to be loaded with the anti-gen and reinfused in vivo; the other strategy still con-siders the tumor cells representative of the entire anti-genic repertoire of a certain neoplasia; such cells, genet-ically modified to produce cytokines and/or co-stimula-tory genes, could be injected into patients as a cellularvaccine. In the latter case a pool of cell lines derivedfrom different patients with the same type of tumorcould increase the antigenic repertoire and avoidimmunoselection that certain antigens may haveencountered in some patients. Unmatched MHA are nota problem in terms of antigen presentation since inject-ed cells are destroyed and represented by host APC.Moreover if different sets of alloantigens are selected

from different pools, the risk of repeated alloimmuniza-tion during booster vaccination would probably beavoided. The background and prospectives of genetical-ly modified tumor cell vaccines are presented below.

Cytokines at the tumor siteIn initial studies recombinant cytokines were inject-

ed at the tumor site or cytokine genes were insertedinto somatic cells to be injected at the tumor site. Allthese studies collectively established that most of thecytokines accumulated at the tumor site were able toinduce tumor destruction and the reaction they inducedwas sometimes strong enough to eradicate a tumorantigenically unrelated to the cytokine-releasing cells.The obtained tumor debulking was often followed by asystemic tumor-specific immune memory. It should beunderlined, however, that tumor debulking may occurthrough non-specific immune reactions or so fast as toprevent efficient T-cell priming, this being reminiscentof the dichotomy described for BCG: indiscriminatenecrosis versus protective immunity.

Engineered tumor cell vaccinesEngineering of tumor cells with the gene of a partic-

ular cytokine is an efficient way of ensuring that thiscytokine will be durably present at the tumor site.Repeated local injections would, of course, have thesame effect. Bolus administration, however, does notprovide a constant supply of cytokine. Its effects aremuch less evident than those achieved by the injectionof engineered tumor cells306 that can ensure the provi-sion of antigen and continued local accumulation of thecytokine until a physiologic or a pharmacologic thresh-old is reached, and the biological activity of the cytokinecan begin.

The immunogenicity that tumor cells can acquireupon cytokine-gene transduction may stem fromrecruitment by released cytokines, of particular reper-toires of inflammatory cells, whose differing abilities toinfluence TAA presentation and secrete secondarycytokines may shape both immunogenicity and deflec-tion of the ensuing immune memory towards a Th1 orTh2 type of response. A cytokine may be simultaneous-ly involved in tumor rejection, leukocyte recruitmentand activation of memory mechanisms.

Many experimental studies have been performed inmice over the last seven years and cytokine genes fromIL-1 to IL-18 have been tested. Most of those studiesdescribed whether a certain cytokine gene, upon trans-duction, can inhibit tumor growth in vivo; some alsodescribed whether the cytokine induced protectiveimmunization against challenge by parental cells where-as only a few studies described efficacy in a therapeu-tic setting. It is clear that the way cytokines modifytumor oncogenicity, immunogenicity and curative effectis not only dependent on the cytokine employed but alsoon the tumor model utilized. The immune mechanismsresponsible for inhibition of tumor growth may not bethe same as those required for immune memory or thosenecessary for eradication of an established tumor.

141


Translation of animal studies into a clinical settingfaces a substantial difference, that is the fast growth oftransplanted tumors and therefore the short time win-dow in which immunization can be performed beforethe animal’s death. In murine models, the so-calledestablished tumor is a tumor that has been injected oneto three days before the beginning of vaccination. Thiscontrasts with phase I/II clinical studies in which enrolledpatients have advanced disease. Clear evidence of ther-apeutic effects is not expected in these patients, there-fore tumors with antigens whose genes have been clonedand are recognized by CTL should be used to allow, atleast, an immunologic follow-up that could prove theeffect of vaccination. This confines the choice to thosecarrying the MAGE, GAGE and BAGE family genes and tomelanomas, which also express antigens of the mela-nocyte lineage, such as tyrosinase, gp100 and MART-1/Melan-A.305 The choice is further restricted by the dif-ficulty of obtaining cells and cell lines from tumors thatare not melanomas to be transduced and then employedfor immunologic evaluation. Melanoma is thus the tumormost frequently chosen for vaccination studies.

Nevertheless, vaccination with cytokine-transduced,freshly isolated cells, which should retain the tumor-antigen repertoire, could be a way of generating tumor-specific T-lymphocyte lines and clones with which toidentify antigens expressed by tumors other thanmelanomas.

In a few cases only, the antigens associated with themurine tumors employed in pre-clinical studies werecharacterized; the majority of studies designed to dis-cover the immunologic mechanisms associated withtumor rejection utilized proteins not classifiable astumor-associated antigens, such as b-galactosidase244

and influenza nucleoprotein.307 Most of these animalstudies were carried out in the syngeneic system, thatin humans corresponds to the autologous situation, inwhich a tumor cell line was both the cell vaccine and thetumor to be cured. Autologous application is actuallydifficult, since it requires tumor cell cultures from everypatient for both gene transduction and immunologicfollow-up. Each patient’s cell vaccine should then bechecked for safety, and a great variability in terms ofcytokine production other than adhesion molecules andantigenic phenotypes may exist between cell vaccines.The use of allogeneic cell lines, on the other hand, hasthe advantage of employing vaccines well-character-ized in terms of tumor antigen, MHC and adhesion mol-ecules, as well as the constant amount of cytokinereleased; these parameters in combination may providea standard reagent for clinical studies.

Both syngeneic and allogeneic tumor cells expressinga common TAA are processed by host APC such that TAAderived peptides are presented in association with hostMHC in either case.307 Nevertheless, in most clinical pro-tocols the expression of the MHC class I allele, whichpresents TAA derived peptide(s), on the immunizingtumor cells is preferred. If cross-priming occurs effi-ciently, this should not be necessary, but it is still unclear

whether vaccination with transduced tumor cells actu-ally primes the host or boosts already present activatedT-lymphocytes. This observation indicates that co-stim-ulatory molecules, such as B7, in addition to cytokinesmay be transduced in cell vaccines in order to amplifythe boosting effect, since is not clear whether B7 trans-duced cells prime the host directly.

Clinical vaccination protocols using IL-2 or IL-4 gene-transduced allogeneic melanoma cells have been per-formed at the Istituto Nazionale Tumori in Milan, Italy.An HLA-A2 melanoma cell line expressing Melan-A/MART-1, tyrosinase, gp100 and MAGE-3 has beentransduced and irradiated before the treatment ofadvanced HLA-A2+ melanoma patients.308 In the firstprotocol, patients were injected subcutaneously on days1, 13, and 26 with IL-2 gene-transduced and irradiatedmelanoma cells at doses of 5 (3 patients) and 15 (4patients) 3107 cells. Mixed lymphocyte-tumor cultures(MLTC) and limiting dilution analyses were performed tocompare pre- and post-vaccination PBL. While MLTCrevealed an increased but MHC-unrestricted cytotoxic-ity, in two cases the frequencies of melanoma-specificCTL precursors were clearly augmented by vaccination.In one patient, HLA class II-restricted effectors werefound to be involved in the recognition of autologoustumor. Which antigen(s) was involved in the recognitionby PBL of vaccinated patients remains unclear. In 3 outof 5 cases studied, pre- and post-vaccination PBL couldnot recognize any melanoma peptide tested or known tobe restricted by HLA-A2 allele.308 Among other possibleexplanations, this might be due to a tumor associatedantigenic repertoire that exceeds the limited number ofantigens whose genes have been cloned so far.

This indicates that vaccination with cell lines is advan-tageous because the cell lines stimulate the host withthe entire repertoire of known and unknown antigens. Inthe allogeneic system it is then easy to rotate the trans-duced cell line within the protocols and so maximize thechances that a relevant tumor antigen is present in thevaccine. Some antigens, in fact, may be negativelyselected and lost in one patient-derived line, but not inothers. In addition, selection of allogeneic cell lines dis-playing various MHC reduces the interference of repeat-ed boosting with strong alloantigens. Indeed vaccina-tion with a pool of three melanoma cell lines commencedbefore the cloning of known melanoma associated anti-gens, resulted in increased survival correlated with thelevel of antibody against the GM2 ganglioside, indicat-ing possible involvement of a humoral response; corre-lation with the CTL response was not investigated.309

Going back to animal studies in which vaccinationtherapy with cytokine-transduced tumor cells was suc-cessful, it should be underlined that it was not clearwhich of the measured immune responses was respon-sible for the therapeutic effect since, generally, induc-tion of cytotoxic T-lymphocytes was, per se, insufficientto produce a cure. In keeping with this statement, vac-cination of 13 evaluable patients with MAGE-3.A1 pep-tide resulted in 3 clinical regressions, although no CTL

Cell therapy

142


precursors were found in the PBL of these responders.310

Refined animal studies performed to identify whichimmune responses correlate with the therapeutic activ-ity indicated that both T- and B-cells should be proper-ly activated.311, 312

These observations may suggest that while a patientcould be immunized against a tumor, the immunity thusinduced might be insufficient to fight the establishedtumor growing within its own stroma. The combinationof poor immune function and large tumor burden makespatients with advanced disease dubious predictors ofclinical response.

The general idea is that cytokine engineered tumorcells should be used as vaccines in minimal disease set-tings.313 A new form of treatment would thus be avail-able for combination with conventional managementof patients after surgical removal of their tumor,patients with minimal residual disease, or patientsexpected to manifest tumor recurrence after a signifi-cant apparently disease-free interval. When comparedwith conventional forms of management, vaccination isa soft, non-invasive treatment, unlikely to cause partic-ular distress or side-effects, and could be administeredafter resection of a primary tumor when recurrence isexpected.

Use of mesenchymal cells for treatment ofneoplastic and non-neoplastic disorders

In addition to hematopoietic stem cells which can dif-ferentiate to produce progenitors committed to termi-nal maturation,314 human bone marrow also containsstem cells of non-hematopoietic tissues which are cur-rently referred to as mesenchymal stem cells (MSC),because of their ability to differentiate into cells thatcan roughly be defined as mesenchymal, or as marrowstromal cells because they appear to arise from the com-plex array of supporting structures found in marrow.315

Stromal cells of the marrow microenvironment includefibroblasts, endothelial cells, reticular cells, adipocytes,osteoblasts and macrophages, the last, although ofhematopoietic origin, being considered functional com-ponents of the regulatory stroma.316 The heterogeneouspopulations of mesenchymal cells and their associatedbiosynthetic products have the unique capacity to reg-ulate hematopoiesis.317

Environmental components can modify the prolifera-tive and differentiative behavior of hematopoietic cells bymeans of (i) cell-to-cell interactions, (ii) interactions ofcells with extracellular matrix molecules, and (iii) inter-actions of cells with soluble growth regulatory mole-cules.316 All these regulatory modalities participate instromal cell-mediated regulation of hematopoiesis. Infact, marrow stromal cells provide the physical frameworkwithin which hematopoiesis occurs, play a role in direct-ing the processes by synthesizing, sequestering or pre-senting growth-stimulatory and growth-inhibitory fac-tors, and also produce numerous extracellular matrix pro-teins and express a broad repertoire of adhesion mole-cules that serve to mediate specific interactions with

hematopoietic stem/progenitor cells of both myeloid andlymphoid origin.318, 319 Although growth factors play keyroles in stem/progenitor cell proliferation and differenti-ation it seems improbable that hematopoiesis is regulat-ed only by a random mix of growth factors and respon-sive cells. Rather, it is likely that regulatory moleculesand localization phenomena within marrow stroma arerequired to sustain and regulate the function of thehematopoietic system.320

Although it is commonly accepted that stem cells arecapable of homing to the marrow and docking at specif-ic sites, the exact role of microenvironmental cells, adhe-sion molecules and extracellular matrix molecules in reg-ulating the localization and spatial organization ofhematopoietic stem cells in the marrow and drivingmyeloid and lymphoid regeneration following stem celltransplantation remains a matter of hypothesis.321 Stud-ies in animals demonstrated that stem and progenitorcells have different distributions across the femoral mar-row cavity of mice, thus suggesting that marrow stromais organized into functionally discrete environments, suchas primary microenvironmental and secondary microen-vironmental areas, allowing distinct differentiation pat-terns of hematopoietic stem cells.322 The stem cell nichehypothesis, proposed by Schofield323 suggested that cer-tain microenvironmental cells of the marrow stromacould maintain the stem cells in a primitive, quiescentstate. Another mechanism supporting the concept of spe-cialized microenvironmental areas is stroma-mediated,compartimentalized growth factor production. Growthfactor produced locally by stromal cells may bind to theextracellular matrix and be presented to immobilized tar-get cells which recognize each growth factor throughspecific receptors.320 This mechanism may provide theopportunity for localizing distinct growth factors at rel-atively high concentrations to discrete sites. As yet, rel-atively little is known of the nature of the factor(s) pro-duced by different stromal cell types which modulate lin-eage development. However, a growing body of evidencesuggests that marrow stroma is involved not only in reg-ulating myeloid cell growth, but also in T- and B-cell lym-phopoietic development.324-327 Distinct adhesion mole-cules and cytokines are known to regulate stroma-depen-dent T- and B-lymphopoiesis,328, 329 suggesting that mar-row stroma may function as a site of T- as well as B-celllymphopoiesis.

The existence of self-renewing MSC is supported byseveral in vitro and in vivo data.330 At the functional lev-el, MSC residing within marrow microenvironment, estab-lish marrow stroma both in vitro and in vivo and havemultilineage differentiation capacity, being capable ofgenerating progenitors with restricted developmentpotential which include fibroblast, osteoblast, adipocyte,chondrocyte and myoblast progenitors (Figure 7).331-333

Putative stromal cell progenitors have been identified inhuman marrow by their ability to generate colonies offibroblast-like cells originating from single clonogenicprogenitors termed fibroblast colony-forming units (CFU-F).334 These progenitors, which belong to the osteogenic

C. Bordignon et al.

143


stromal lineage, play a central role in establishing themarrow microenvironment both in vitro and in vivo.335-337

Under appropriate culture conditions and supplementa-tion with specific stimuli, a proportion of marrow CFU-Fcan be induced to either adipogenesis333 or osteoblasto-genesis.338 Studies involving ectopic transplantation ofindividual fibroblastic clones grown in vitro from mousemarrow beneath the renal capsule of syngeneic hostsdemonstrated that approximately 15% produced a mar-row organ containing the full spectrum of stromal celltypes of hematopoietic microenvironment, thus suggest-ing that CFU-F have multilineage differentiation capaci-ty and supporting the stromal stem cell hypothesis.339

Based on these findings, CFU-F can be identified as mul-tipotent stromal progenitors rather than lineage-restrict-ed fibroblast progenitors.

CFU-F can be enriched from adult bone marrow bymeans of the STRO-1 monoclonal antibody that identi-fies essentially all assayable marrow CFU-F.340 STRO-1+

cells do not express the CD34 antigen and fail to gener-ate hematopoietic progenitors, thus facilitating a cleanseparation between hematopoietic and stromal progen-itors.341 Flow-sorted STRO-1+ cells grown under long-term culture conditions generate adherent stromal lay-ers consisting of fibroblasts, osteoblasts, smooth musclecells and adipocytes.340 These stromal layers are capableof supporting hematopoiesis in long-term cultures initi-ated with CD34+ cells. In addition to STRO-1, other mon-oclonal antibodies, such as SH-2, have been describedwhich specifically detect mesenchymal progenitors.342

In vivo data generated in animal models support thefunctional regulatory role of the marrow microenviron-ment. In the fetal sheep model of in utero stem celltransplantation, co-transplantation of stem cells withmarrow stromal cells has been shown to improve levels

of donor cell engraftment.343 In the NOD/SCID mousemodel of in utero stem cell transplantation, fetal stemcells have a nine times greater engraftment potentialbut this advantage is abrogated if the recipients are irra-diated prior to transplant, indicating that the marrowmicroenvironment is important in driving myeloid andlymphoid engraftment.344

The importance of stromal cells in hematopoiesis hasalso been demonstrated by several studies in humans.Despite normal peripheral blood counts, levels of primi-tive and committed progenitors in the bone marrow ofpatients who have received allogeneic stem cell trans-plantation remain subnormal for many years.345 Further-more, cultured stromal cells from patients who havereceived allogeneic stem cell transplant (SCT) show sig-nificant impairment in their ability to support the growthof hematopoietic progenitors from normal marrow.346

Decreased CFU-GM production and defective stroma pro-duction have been demonstrated following autologousSCT347 as well as after induction chemotherapy.348

The role of marrow stroma in hematopoietic regulationand the peculiar functional characteristics of stromalcells raise the possibility that the delivery of ex vivoexpanded marrow MSC into a hematopoietically-com-promised marrow might promote hematopoiesis. Bonemarrow stromal cells are a quiescent, non-cycling popu-lation with low cell turn-over, as demonstrated by theresistance to irradiation. Based on these characteristics,methods have been developed which allow for genedelivery into stromal cells.349 Since stromal cells are meta-bolically active they also provide a suitable means ofsecreting therapeutic proteins, including coagulation fac-tors or adenosine deaminase.350 Recent data showing thatMSC suppress allogeneic T-cell responses in vitro sug-gest a role for stromal cells in modulating allogeneic

Cell therapy

Figure 7.

144


transplant rejection and graft-versus-host disease.351

It must be emphasized that because of the limitedknowledge of MSC biology, clinical applications of stro-mal cells, although exciting, essentially remain a mat-ter of hypothesis to be carefully tested in the appropri-ate clinical setting. Essential prerequisites for clinicalapplications using culture-expanded mesenchymal cellsas a supplement for hematopoietic SCT are (i) the pos-sibility of isolating mesenchymal progenitors andmanipulating their growth under defined in vitro cultureconditions352 and (ii) the demonstration of the possibil-ity of efficiently introducing cultured stromal cells backinto patients.

Studies in rodents and dogs have clearly demonstrat-ed that if sufficient stromal cells are reinfused, they notonly seed the bone marrow but also enhancehematopoietic recovery.353-357 Although demonstratedin several mouse models, the transplantability of mar-row stromal elements remains a controversial issue inhumans.358, 359 The majority of data so far generated inrecipients of HLA-identical marrow transplants hasfailed to demonstrate any contribution of donor cells tomarrow stroma regeneration.358 Although many factorsmay affect the transplantability of stromal elements,the low frequency of stromal progenitors in conven-tional marrow harvests may explain the failure of mes-enchymal cell transplanttion in humans.

Indeed, during the last decade, SCT methodology haschanged substantially, particularly as a result of theincreasing use of peripheral blood transplants. The exis-tence of a circulating stromal progenitor has beendemonstrated by using a NOD/SCID model and this isextremely relevant to stromal cell therapy.360 By usingthe X-linked human androgen receptor (HUMARA) geneand fluorescent in situ hybridization analysis for the Ychromosome, the transplantability of stromal progeni-tors in a proportion of recipients of haploidentical HLA-mismatched T-cell-depleted allografts reinfused with acombination of bone marrow and mobilized peripheralblood cells has recently been demonstrated (Carlo-Stel-la and Tabilio, unpublished observations, 1999). Takentogether, these findings allow the hypothesis that MSCare transplantable in man provided that an adequate,but as yet unidentified, number of CFU-F is reinfused. Inaddition, these data allow the planning of clinical stud-ies using culture-expanded, gene-marked mesenchymalcells in order to investigate a number of issues, includ-ing (i) dose of marrow stromal progenitors necessary toachieve a transplant; (ii) duration of post-transplantmarrow stromal cell function; (iii) role of stromal cellsin myeloid, B- and T-lymphoid reconstitution followingSCT.

A limited number of clinical trials using ex vivo gen-erated MSC are currently underway. So far, the onlypublished phase I clinical trial using MSC reported thatthe systemic infusion of autologous MSC appears to bewell tolerated.361 MSC can be explored as vehicles forboth cell therapy and gene therapy (Table 4). MSC couldbe used to replace marrow microenvironment damaged

by high-dose chemotherapy in order to either improvehematopoietic recovery from myeloablative chemother-apy or to treat late graft failures or delayed plateletengraftment. Based on their functional characteristics,MSC are attractive vehicles for gene therapy in that theyare expected not to be lost through differentiation asrapidly as hematopoietic progenitors. Examples of dis-eases in which stromal cell-mediated gene therapymight be appropriate include factor VIII and factor IXdeficiencies and the various lysosomal storage diseases.Interestingly, compared to skin fibroblasts or leukocytes,marrow-derived mesenchymal cells produce signifi-cantly higher levels of a-iduronidase, an enzymeinvolved in type II mucopolysaccharidoses (Danesino andCarlo-Stella, unpublished data). In addition, stromal cellsmight also be transduced with cDNA of varioushematopoietic growth factors or cytokines. Thisapproach might allow high levels of compartimentalizedgrowth factor production and might be used (i) to stim-ulate hematopoiesis in patients with congenital oracquired hematopoietic defects, (ii) to improve B- andT-cell recovery following allogeneic SCT, (iii) to acceler-ate myeloid reconstitution in recipients of cord bloodtransplants.

In conclusion, MSC appear to be an attractive thera-peutic tool capable of playing a role in a wide range ofclinical applications in the context of both cell and genetherapy strategies. However, a number of fundamentalquestions about MSC still need to be resolved beforethey can be used for safe and effective cell and genetherapy.

ConclusionsAlthough most of the new therapeutic approaches of

cell therapy are experimental and have not yet been val-idated by phase III clinical trials, they appear to hold ahigh therapeutic potential. Separation of GVL from GvHDthrough generation and infusion of leukemia-specific T-cell clones or lines is one of the most intriguing andpromising fields of investigations for the future. Like-wise, strategies devised to improve immune reconstitu-tion and restore specific anti-infectious functionsthrough either induction of unresponsiveness to recipi-ent alloantigens or removal of alloreactive donor T-cells

C. Bordignon et al.

Table 4. Potential clinical applications of mesenchymalstem cells.

• Replacement of chemotherapy-damaged stroma• Enhancement of myeloid recovery following hematopoietic stem cell transplantation• Enhancement of T- and B-cell reconstitution following allogeneic stem cell transplantation• Compartimentalized growth factor/cytokine production• Modulation of GvHD• Delivery of exogenous gene products

145


might increase the applicability and success ofhematopoietic stem cell transplantation. Cellularimmunotherapy with DC must be standardized and sev-eral critical points, discussed in this review article mustbe properly addressed with specific clinical studies. Stim-ulation of leukemic cells via CD40 receptors and trans-duction of tumor cells with co-stimulatory moleculesand/or cytokines may be useful in preventing tumorescape from immune surveillance. Tumor cells can begenetically modified to interact directly with dendriticcells in vivo or recombinant antigens can be delivered todendritic cells using attenuated bacterial vectors by oralvaccination. MSC represent an attractive therapeutictool capable of playing a role in a wide range of clinicalapplications in the context of both cell and gene thera-py strategies.

Contributions and AcknowledgmentsAll authors gave substantial contributions to analysis

and interpretation of literature data and drafting thearticle or revising it critically. CB was primarily respon-sible for the section on genetically engineered donor lym-phocyte infusion for treatment of leukemia relapse, CCSfor the section on mesenchymal cells, MPC for the sec-tions on oral vaccination and engineered tumor vaccines;RML for the section on dendritic cells DR. FL was primar-ily responsible for sections on adoptive immunotherapy,AO for the sections on LAK and TIL and DR for the sectionon tumor escape from immune surveillance and onleukemic cells as APC. The authors are listed in alpha-betical order.

DisclosuresConflict of interest. This review article was prepared by

request from Haematologica. The authors were a group ofexperts and representatives of two pharmaceutical com-panies, Amgen Italia SpA and Dompé Biotec SpA, bothfrom Milan, Italy. This co-operation between a medicaljournal and pharmaceutical companies is based on thecommon aim of achieving optimal use of new therapeu-tic procedures in medical practice. In agreement with theJournal’s Conflict of Interest policy, the reader is giventhe following information. The preparation of this man-uscript was supported by educational grants from the twocompanies. Dompé Biotec SpA sells G-CSF and rHuEpo inItaly, and Amgen Italia SpA has a stake in Dompé BiotecSpA.

Redundant publications: no overlapping with previouspapers.

Manuscript processingManuscript received May 10, 1999; accepted August

30, 1999.

References

1. Murphy JB. Monography. Rockefeller Institute ofMedical Research 1926; 21:1-168.

2. Medawar PB. Croonian Lecture. Proc Royal Soc B1958; 148:159-61.

3. Bertolini F, de Vincentiis A, Lanata L, et al. Allogene-

ic hematopoietic stem cells from sources other thanbone marrow: biological and technical aspects. Hae-matologica 1997; 82:220-38.

4. Arcese W, Aversa F, Bandini G, et al Clinical use ofallogeneic hematopoietic stem cells from sources oth-er than bone marrow. Haematologica 1998; 83:159-82.

5. Aglietta M, Bertolini F, Carlo-Stella C, et al. Ex-vivoexpansion of hematopoietic cells and their clinical use.Haematologica 1998; 83:824-48.

6. Janeway CA Jr, Bottomly K. Signals and signs for lym-phocyte responses. Cell 1994; 76:275-85.

7. Thompson CB. Distinct roles for the costimulatoryligands B7-1 and B7-2 in T helper cell differentiation?Cell 1995; 81:979-82.

8. van Kooten C, Banchereau J. Functions of CD40 on Bcells, dendritic cells and other cells. Curr OpinImmunol 1997; 9:330-7.

9. Boon T, DePlaen E, Traversari C, et al. Identificationof tumor rejection antigens recognized by T lympho-cytes. Cancer Surv 1992; 13:23-37.

10. van der Bruggen P, Traversari C, Chomez P, et al. Agene encoding an antigen recognized by cytolytic Tlymphocytes on human melanoma. Science 1991;254:1643-7.

11. Restifo NP, Esquivel F, Kawakami Y, et al. Identifica-tion of human cancers deficient in antigen processing.J Exp Med 1993; 177:265-72.

12. Merogi AJ, Marrogi AJ, Ramesh R, Robinson WR, Fer-min CD, Freeman SM. Tumor-host interaction: analy-sis of cytokines, growth factors, and tumor-infiltratinglymphocytes in ovarian carcinomas. Hum Pathol1997; 28:321-31.

13. Guinan EC, Gribben JG, Boussiotis VA, Freeman GJ,Nadler LM. Pivotal role of the B7: CD28 pathway intransplantation tolerance and tumor immunity. Blood1994; 84:3261-82.

14. Maeurer MJ, Gollin SM, Martin D, et al. Tumor escapefrom immune recognition. J Clin Invest 1996; 98:1633-41.

15. Wolfel T, Hauer M, Schneider J, et al. A p16INK4a-insen-sitive CDK4 mutant targeted by cytolytic T lympho-cytes in a human melanoma. Science 1995; 269:1281-4.

16. Tanaka K, Yoshioka T, Bieberich C, Jay C. Role ofmajor histocompatibility complex class I antigens intumor growth and metastasis. Annu Rev Immunol1988; 6:359-80.

17. Korkolopoulou P, Kaklamanis L, Pezzella F, Harris AL,Gatter KC. Loss of antigen-presenting molecules(MHC class I and TAP-1) in lung cancer. Br J Cancer1996; 97:148-53.

18. Schultze J, Nadler LM, Gribben JG. B7-mediated co-stimulation and the immune response. Blood Rev1996; 10:111-27.

19. Strand S, Galle PR. Immune evasion by tumours:involvement of the CD95 (APO-1/Fas) system and itsclinical implications. Mol Med Today 1998; 4:63-8.

20. Onrust SV, Harti PM, Rosen SD, Hanahan D. Modu-lation of L-selection ligand expression during animmune response accompanying tumorigenesis intransgenic mice. J Clin Invest 1996; 97:54-64.

21. Herberman RB, Nunn ME, Lavrin DH. Natural cyto-toxic reactivity of mouse lymphoid cells against syn-geneic and allogeneic tumors. I. Distribution of reac-tivity and specificity. Int J Cancer 1975; 16: 216-29.

22. Barlozzari T, Reynolds CW, Herberman RB. In vivorole of natural killer cells: involvement of large granu-lar lymphocytes in the clearance of tumor cells in anti-asialo GM1-treated rats. J Immunol 1983; 131:1024-7.

23. Karre K, Ljunggren H-G, Piontek G, Kiessling R. Selec-

Cell therapy

146


tive rejection of H2-deficient lymphoma variants sug-gests alternative immune defense strategy. Nature1986; 319:675-8.

24. Braud VM, Allan DS, O’Callaghan CA, et al. HLA-Ebinds to natural killer cell receptors CD94/NKG2A, Band C. Nature 1998; 391; 795-9.

25. Ciccone E, Grossi CE, Velardi A. Opposing functionsof activatory T-cell receptors and inhibitory NK-cellreceptors on cytotoxic T cells. Immunol Today 1996;17:451-3.

26. Colonna M. Immunology: unmasking the killer'saccomplice. Nature 1998; 391:642-3.

27. Daniels B, Karlhofer FM, Seaman WE, YokoyamaWM. A natural killer cell receptor specific for a majorhistocompatibility complex class I molecule. J Ex Med1994; 180:687-92.

28. Miller JS, Alley KA, McGlave PB. Differentiation of nat-ural killer cells from human primitive marrow prog-enitors in a stroma based long term culture system:identification of a CD34+/CD7+ NK progenitor.Blood 1994; 83:2594-601.

29. Mrozek E, Anderson P, Caligiuri MA. Role of inter-leukin-15 in the development of human CD56+ nat-ural killer cells from CD34+ hematopoietic progenitorcells. Blood 1996; 87:2632-40.

30. Pierson BA, Miller JS. CD56+bright and CD56+dim nat-ural killer cells in patients with chronic myelogenousleukemia progessively decrease in number, respondless to stimuli that recruit clonogenic natural killercells, and exhibit decreased proliferation on a per cellbasis. Blood 1996; 88:2279-87.

31. Rayner AA, Grimm EA, Lotze MT, et al. Lymphokine-activated killer (LAK) cell phenomenon. IV. Lysis byLAK cell clones of fresh human tumor cells from autol-ogous and multiple allogeneic tumors. J Natl CancerInst 1985; 75:67-75.

32. Torpey DJ 3rd, Lindsley MD, Rinaldo CR Jr. HLA-restricted lysis of herpes simplex virus-infected mono-cytes and macrophages mediated by CD4+ and CD8+T lymphocytes. J Immunol 1989; 142:1325-32.

33. Jorgensen H, Hokland P, Jensen T, et al. Natural killercells in peripheral blood after autologous bone mar-row transplantation: a combined phenotypic andfunctional study. Nat Immunol 1995; 14:164-72.

34. Roberts K, Lotze MT, Rosenberg SA. Separation andfunctional studies of the human lymphokine-activat-ed killer cell. Cancer Res 1987; 47: 4366-71.

35. Rosenberg SA. Lymphokine-activated killer cells: a newapproach to immunotherapy of cancer. J Natl CancerInst 1985; 75:595-603.

36. Papa MZ, Mulé JJ, Rosenberg SA. The antitumor effi-cacy of lymphokine activated killer cells and recombi-nant IL-2 in vivo: successful immunotherapy of estab-lished pulmonary metastases from weakly immunogenicand non-immunogenic murine tumors of three distincthistological types. Cancer Res 1986; 46:4973-8.

37. Lotzova E, Savary CA, Herberman RB. Inhibition ofclonogenic growth of fresh leukemia cells by unstim-ulated and IL-2 stimulated NK cells of normal donors.Leukemia Res 1987; 11:1059-66.

38. Parrado A, Rodriguez-Fernadez JM, Casares S, et al.Generation of LAK cells in vitro in patients with acuteleukemia. Leukemia 1993; 7:1344-8.

39. Archimbaud E, Bailly M, Doré JF. Inducibility of lym-phokine activated killer (LAK) cells in patients withacute myelogenous leukaemia in complete remissionand its clinical relevance. Br J Haematol 1991; 77:328-34.

40. Rosenberg SA, Lotze MT, Muul LM, et al. Observa-tions on the systemic administration of autologouslymphokine-activated killer cells and recombinantinterleukin-2 to patients with metastatic cancer. N

Engl J Med 1985; 313:1485-92.41. Hawkins MJ. PPO Update IL2/LAK. Princ Prac Oncol

1989; 3:1-4.42. Weiss GR, Margolin KA, Aronson FR, et al. A ran-

domized phase II trial of conttinuous infusion inter-leukin-2 or bolus injection interleukin-2 plus lym-phokine-activated killer cells for advanced renal cellcarcinoma. J Clin Oncol 1992; 10:275-81.

43. McCabe MS, Stablein D, Hawkins MJ. The modifiedgroup C experience-phase III randomized trials of IL-2 versus IL-2/ LAK in advanced renal cell carcinomaand advanced melanoma [abstract]. Proceedings ofAmerican Society of Clinical Oncology; 1991; 10:213a.

44. Rosenberg SA, Lotze MT, Yang JC, et al. Prospectiverandomized trial of high-dose interleukin-2 alone or inconjunction with lymphokine-activated killer cells forthe treatment of patients with advanced cancer. J NatlCancer Inst 1993; 85:622-32.

45. Murray Law T, Motzer RJ, Mazumdar M, et al. PhaseIII randomized trial of interleukin-2 with or withoutlymphokine activated killer cells in the treatment ofpatients with advanced renal cell carcinoma. Cancer1995; 76: 824-32.

46. Kimura H, Yamaguchi Y: A phase III randomized studyof interleukin-2 lymphokine-activated killer cellimmunotherapy combined with chemotherapy orradiotherapy after curative or noncurative resectionof primary lung carcinoma. Cancer 1997; 80:42-9.

47. Boldt DH, Mills BJ, Gemlo BT, et al. Laboratory cor-relates of adoptive immunotherapy with recombinantinterleukin-2 and lymphokine-activated killer cells inhumans. Cancer Res 1988; 48:4409-16.

48. Margolin KA, Rayner AA, Hawkins MJ, et al. Inter-leukin-2 and lymphokine-activated killer cell therapyof solid tumors: analysis of toxicity and managementguidelines. J Clin Oncol 1989; 7:486-98.

49. Négrier S, Philip T, Stoter G, et al. Interleukin-2 withor without LAK cells in metastatic renal cell carcino-ma: a report of a European multicentre study. Eur JCancer Clin Oncol 1989; 25(suppl 3):S21-S28.

50. Escudier B, Farace F, Droz JP, et al. Abstract Pro-ceedings of American Society of Clinical Oncology1991; 10:527a.

51. Attal M, Blaise D, Marit G, et al. Consolidation treat-ment of adult acute lymphoblastic leukemia: aprospective, randomized trial comparing allogeneicversus autologous bone marrow trasplantation andtesting the impact of recombinant interleukin-2 afterautologous bone marrow transplantation. Blood1995; 86:1619-28.

52. Fefer A, Benyunes M, Higuchi C, et al. IL-2+/- lym-phokine activated killer cells as consolidative immuno-therapy after autologous bone marrow transplanta-tion for hematologic malignancies. Acta Hematol1993; 89:2-7.

53. Beaujean F, Bernaudin F, Kuentz M, et al. Successfulengraftment after autologous transplantation of 10-day cultured bone marrow activated by interleukin 2in patients with acute lymphoblastic leukemia. BoneMarrow Transplant 1995; 15:691-6.

54. Mechan KR, Verma UN, Cahill R, et al. Interleukin-2-activated hematopoietic stem cell transplantation forbreast cancer: investigation of dose level with clinicalcorrelates. Bone Marrow Transplant 1997; 20:643-51.

55. Olivieri A, Cantori I, Montanari M, et al. GM-CSF plusIL-2 administration associated with multiple autolo-gous LAK reinfusions can induce a major cytogeneticresponse in early relapsed CML after autologoustransplantation: a case report [abstract]. Bone Mar-row Transplant 1998; 21:S65.

56. Klingemann HG, Deal H, Reid D, Eaves CJ. Design

C. Bordignon et al.

147


and validation of a clinically applicable culture pro-cedure for the generation of interleukin-2 activatednatural killer cells in human bone marrow autografts.Exp Hematol 1993; 21:1263-70.

57. Silva MRG, Parreira A, Ascensao JL. Natural killer cellnumbers and activity in mobilized peripheral bloodstem cell grafts: conditions for in vitro expansion. ExpHematol 1995; 23:1676-81.

58. Miller JS, Klingsporn S, Lund J, et al. Large-scale exvivo expansion and activation of human natural killercells for autologous therapy. Bone Marrow Transplant1994; 14:555-62.

59. Vujanovic NL, Rabinowich H, Lee YJ Jost L, HerbermanRB, Whiteside TL. Distinct phenotypes and function-al characteristics of human natural killer cells obtainedby rapid interleukin-induced adherence to plastic. CellImmunol 1993; 151:133-57.

60. Sheffold C, Brandt K, Johnston V. Potential of autol-ogous immunologic effector cells for bone marrowpurging in patients with chronic myeloid leukemia.Bone Marrow Transplant 1995; 15:33-9.

61. Spiess PJ, Yang JC, Rosenberg SA. In vivo antitumoractivity of tumor-infiltrating lymphocytes expanded inrecombinant IL-2. J Natl Cancer Inst 1987; 79:1067-75.

62. Balch CM, Riley LB, Bae YJ, et al. Patterns of humantumour-infiltrating lymphocytes in 120 human can-cers. Arch Surg 1990; 12:200-5.

63. Haas GP Solomon D, Rosenberg SA. Tumour-infil-trating lymphocytes from nonrenal urological malig-nancies. Cancer Immunol Immunother 1990; 30:342-50.

64. Heo DS Whiteside TL, Johnson JT, Chen KN, BarnesEL, Herberman RB. Long term interleukin-2-depen-dent growth and cytotoxic activity of tumour-infil-trating lymphocytes from human squamous cell car-cinoma of the head and neck. Cancer Res 1987; 47:6353-62.

65. Rodolfo M, Salvi C, Bassi C, Parmiani G. Adoptiveimmunotherapy of a mouse colon carcinoma withrecombinant interleukin-2 alone or combined withlymphokine-activated killer cells or tumor-immunelymphocytes. Survival benefit of adjuvant post-surgi-cal treatments and comparison with experimentalmetastases model. Cancer Immunol Immunother1990; 31:28-36.

66. Griffin JD. Hemopoietins in oncology: factoring outmyelosuppression. J Clin Oncol 1989; 7:151-5.

67. Wong RA, Alexander RB, Puri RK, Rosenberg SA. Invivo proliferation of adoptively transferred tumor-infil-trating lymphocytes in mice. J Immunother 1991; 10:120-30.

68. Fisher B, Packard BS, Read EJ, et al. Tumor localiza-tion of adoptively transferred indium-111 labeledtumor infiltrating lymphocytes in patients withmetastatic melanoma. J Clin Oncol 1989; 7:250-61.

69. Rosenberg SA, Aebersold P, Cornetta K, et al. Genetransfer into humans: immunotherapy of patientswith advanced melanoma using tumor infiltratinglymphocytes modified by retroviral gene transfer. NEngl J Med 1990; 323:570-8.

70. Aebersold P, Hyatt C, Johnson S, et al. Lysis of autol-ogous melanoma cells by tumor-infiltrating lympho-cytes: association with clinical response. J Natl Can-cer Inst 1991; 83:932-7.

71. Barth RJ, Mulé JJ, Spiess PJ, Rosenberg SA. Interferongamma and tumor necrosis factor have a role intumor regressions mediated by murine CD8+ tumor-infiltrating lymphocytes. J Exp Med 1991; 173:647-58.

72. Treisman J, Hwu P, Minamoto S, et al. Interleukin-2transduced lymphocytes grow in an autocrine fashion

and remain responsive to antigen. Blood 1995; 85:139-45.

73. Bani MR, Garofalo A, Scanziani E, Giavazzi R. Effectof interleukin-1-b on metastases formation in differ-ent tumor systems. J Natl Cancer Inst 1991; 83:119-23.

74. Malik STA, Naylor S, EAST N, Oliff A, Balkwill FR.Cells secreting tumour necrosis factor show enhancedmetastases in nude mice. Eur J Cancer 1990; 26:1031-4.

75. Robbins PF, Kawakami Y. Human tumor antigens rec-ognized by T-cells. Curr Opin Immunol 1996; 8:628-36.

76. Rosenberg SA, Packard BS, Aebersold PM, et al. Useof tumor-infiltrating lymphocytes and interleukin-2 inthe immunotherapy of patients with metastaticmelanoma. A preliminary report. N Engl J Med 1988;319:1676-80.

77. Kradin RL, Lazarus DS, Dubinett SM, et al. Tumour-infiltrating lymphocytes and interleukin-2 in treatmentof advanced cancer. Lancet 1989; 1:577-9.

78. Deeg HJ, Spitzer TR, Cootler-Fox M, et al. Condition-ing-related toxicity and acute graft-versus-host diseasein patients given methotrexate/cyclosporine prophy-laxis. Bone Marrow Transplant 1991; 7:193-8.

79. Vogelsang GB, Hess AD. Graft-versus-host disease:new directions for a persistent problem. Blood 1994;84:2061-7.

80. Xun CQ, Thompson JS, Jennings CD, et al. The effectof human IL-2-activated natural killer and T-cells ongraft-versus-host disease and graft-versus-leukemia inSCID mice bearing human leukemic cells. Transplan-tation 1995; 6:821-7.

81. Miller JS, Prosper F, Mc Cullar V. Natural killer cells arefunctionally abnormal and NK cell progenitors arediminished in granulocyte colony-stimulating factor-mobilized peripheral blood progenitor cell collections.Blood 1997; 90:3098-105.

82. Albi N, Ruggeri L, Aversa F, et al. Natural killer (NK)function and antileukemic activity of a large popula-tion of CD3+/CD8+ T cells expressing NK receptorsfor major histocompatibility complex class I afterthree-loci HLA-incompatible bone marrow trans-plantation. Blood 1996; 87:3993-4000.

83. Lee S, Suen Y, Chang L, et al. Decreased interleukin-12 from activated cord versus adult peripheral bloodmononuclear cells and upregulation of interferon-g,natural killer, and lymphokine activated killer activityby IL-12 in cord blood mononuclear cells. Blood1996; 88:945-54.

84. Verfaillie C, Miller W, Kay N, McGlave PB. Adherentlymphokine activated killer (ALAK) cells in chronicmyelogenous leukemia: a benign cell population withpotent cytotoxic activity. Blood 1989; 74:793-7.

85. Hauch M, Gazzola MV, Small T, et al. Anti-leukemiapotential of interleukin-2 activated natural killer cellsafter bone marrow transplantation for chronic myel-ogenous leukemia. Blood 1990; 75:2250-62.

86. Miller JS, Verfaillie C, McGlave P. Expansion and acti-vation of human natural killer cells for autologoustherapy. J Hematother 1994; 3:71-4.

87. Cervantes F, Pierson BA, McGlave PB, Verfaillie CM,Miller JS. Autologous activated natural killer cells sup-press primitive chronic myelogenous leukemia prog-enitors in long-term culture. Blood 1996; 87:2476-85.

88. Lanier LL, Phillips JH. Inhibitory MHC class I receptorson NK cells and T cells. Immunol Today 1986; 17:86-92.

89. Schmidt-Wolf IGH, Lefterova P, Johnston V. Propa-gation of large numbers of T cells with natural killercell markers. Br J Haematol 1994; 87:453-8.

Cell therapy

148


90. Scheffold C, Brandt K, Johnston V. Potential of autol-ogous immunologic effector cells for bone marrowpurging in patients with chronic myeloid leukemia.Bone Marrow Transplant 1995; 15:33-9.

91. Hoyle C, Bangs CD, Chang P, Kamel O, Metta B,Negrin RS. Expansion of Philadelphia chromosome-negative CD3+ CD56+ cytotoxic cells from chronicmyeloid leukemia patients: in vitro and in vivo effica-cy in severe combined immunodeficiency disease mice.Blood 1998; 92:3318-27.

92. Nalesnik MA, Rao AS, Furukawa H, et al. Autologouslymphokine-activated killer cell therapy of Epstein-Barr virus-positive and -negative lymphoproliferativedisorders arising in organ transplant recipients. Trans-plantation 1997; 63:1200-5.

93. Katsumoto Y, Monden T, Takeda T, et al. Analysis ofcytotoxic activity of the CD4+ T lymphocytes generat-ed by local immunotherapy. Br J Cancer 1996; 73:110-6.

94. Tsurushima H, Qin Liu S, Tsuboi K, Yoshii Y, Nose T,Ohno T. Induction of human autologous cytotoxic Tlymphocytes against minced tissues of glioblastomamultiforme. J Neurosurg 1996; 84:258-63.

95. Schultze JL, Seamon MJ, Michalak S, Gribben JG,Nadler LM. Autologous tumor-infiltrating T cells cyto-toxic for follicular lymphoma cells can be expanded invitro. Blood 1997; 89:3806-16.

96. Kanegane H, Tosato G. Activation of naive and mem-ory T cells by interleukin-15. Blood 1996; 88:230-5.

97. Robertson M, Soffier R, Wolf S, et al. Response ofhuman natural killer (NK) cell to NK cell stimulatoryfactor (NKSF): cytolytic activity and proliferation ofNK cells are differentially regulated by NKSF. J ExpMed 1992; 175:779-88.

98. Rossi AR, Pericle F, Rashleigh S, Janiec J, Djeu J. Lysisof neuroblastoma cell lines by human natural killercells activated by interleukin-12. Blood 1994; 83:1323-8.

99. Kusher DI, Rashlegh SR, Endicott JN, Djeu J. Inter-leukin-2 and interleukin-12 activate killer cell cytolyt-ic response of peripheral blood mononuclear cellsfrom patients with advanced head and neck squa-mous cell carcinoma [abstract]. Proceedings of theAmerican Association of Cancer Research 1994; 35:3119a.

100. Vitale A, Guarini A, Latagliata R, Cignetti A, Foa R.Cytotoxic effectors activated by low-dose IL-2 plus IL-12 lyse IL-2-resistant autologous acute myeloidleukemia blasts. Br J Haematol 1998; 101:150-7.

101. Satoh M, Seki S, Hashimoto W, et al. Cytotoxic gd orab T cells with a natural killer cell marker, CD56,induced from human peripheral blood lymphocytesby a combination of IL-12 and IL-2. J Immunol 1996;157:3886-92.

102. Olivieri A, Cantori I, Provinciali M, et al. The associa-tion of GM-CSF plus IL-2 for cytotoxic cell expansion:influence of monocyte and GM-CSF concentration[abstract]. Blood 1997; 90:4302a.

103. Basse P, Goldfarb RH. Localization of immune effec-tor cells to tumor metastases. In: Goldfarb R, White-side T, eds. Tumor Immunology and Cancer Therapy,New York: Marcel Dekker, Inc., 1994. p. 149-58.

104. Kuznetsov VA, Makalkin IA, Taylor MA, Perelson AS.Nonlinear dynamics of immunogenic tumors: para-meter estimation and global bifurcation analysis. BullMath Biol 1994; 56:295-321.

105. Zhu H, Melder RJ, Baxter LT, Jain RK. Physiologicallybased kinetic model of effector cell biodistribution inmammals: implications for adoptive immunothera-py. Cancer Res 1996; 56:3771-81.

106. Okada K, Nannmark U, Vujanovic N, et al. Elimina-tion of established liver metastases by human inter-

leukin-2-activated natural killer cells after locoregion-al or systemic adoptive transfer. Cancer Res 1996;56:1599-608.

107. Sasaki A, Jain RK, Maghazachi AA, Goldfarb RH, Her-berman RB. Low deformability of lymphokine-acti-vated killer cells as a possible determinant of in vivobiodistribution. Cancer Res 1989; 49:3742-6.

108. Melder RJ, Jain RK. Kinetics of interleukin-2 inducedchanges in rigidity of human natural killer cells. CellBiophys 1992; 20:161-76.

109. Shiloni E, Eisenthal A, Sachs D, et al. Antibody-depen-dent cellular cytotoxicity mediated by murine lym-phocytes activated in recombinant interleukin-2. JImmunol 1987; 6:1992-8.

110. Eisenthal A, Cameron RB, Uppenkamp I, et al. Effectof combined therapy with lymphokine-activated killercells, interleukin-2 and specific monoclonal antibodyon established B16 melanoma lung metastases. Can-cer Res 1988; 48:7140-5.

111. Kolb HJ, Mittermuller J, Clemm C, et al. Donor leuko-cyte transfusions for treatment of recurrent chronicmyelogenous leukemia in marrow transplant patients.Blood 1990; 76:2462-5.

112. Locatelli F. The role of repeat transplantation of hae-matopoietic stem cells and adoptive immunotherapyin treatment of leukaemia relapsing following allo-geneic transplantation. Br J Haematol 1998; 102:633-8.

113. Mackinnon S, Papadopoulos EB, Carabasi MH, et al.Adoptive immunotherapy evaluating escalating dosesof donor leukocytes for relapse of chronic myeloidleukemia after bone marrow transplantation: separa-tion of graft-versus-leukemia responses from graft-ver-sus-host disease. Blood 1995; 86:1261-68.

114. van Rhee F, Lin F, Cullis JO, et al. Relapse of chronicmyeloid leukemia after allogeneic bone marrow trans-plant: the case for giving donor leukocyte transfusionsbefore the onset of hematologic relapse. Blood 1994;83:3377-83.

115. Collins-RH J, Shpilberg O, Drobyski WR, et al. Donorleukocyte infusions in 140 patients with relapsedmalignancy after allogeneic bone marrow transplan-tation. J Clin Oncol 1997; 15:433-44.

116. Kolb HJ, Schattenberg A, Goldman JM, et al. Graft-versus-leukemia effect of donor lymphocyte transfu-sions in marrow grafted patients. European Group forBlood and Marrow Transplantation Working PartyChronic Leukemia. Blood 1995; 86:2041-50.

117. Mehta J, Powles R, Treleaven J, et al. Outcome ofacute leukemia relapsing after bone marrow trans-plantation: utility of second transplants and adoptiveimmunotherapy. Bone Marrow Transplant 1997; 19:709-19.

118. Raanani P, Dazzi F, Sohal J, et al. The rate and kinet-ics of molecular response to donor leucocyte transfu-sions in chronic myeloid leukaemia patients treatedfor relapse after allogeneic bone marrow transplanta-tion. Br J Haematol 1997; 99:945-50.

119. Giralt S, Hester J, Huh Y, et al. CD8-depleted donorlymphocyte infusion as treatment for relapsed chron-ic myelogenous leukemia after allogeneic bone mar-row transplantation. Blood 1995; 86:337-43.

120. Lokhorst HM, Schattenberg A, Cornelissen JJ, ThomasLLM, Verdonck LF. Donor leukocyte infusions areeffective in relapsed multiple myeloma after allogene-ic bone marrow transplantation. Blood 1997; 90:4206-11.

121. Dazzi F, Raanani P, van Rhee P, et al. Donor lympho-cyte infusion (DLI) for relapse of CML after allo-BMT:comparison of two regimens. Bone Marrow Trans-plant 1998; 21(Suppl. 1):S70.

122. Slavin S, Naparstek E, Nagler A, et al. Allogeneic cell

C. Bordignon et al.

149


therapy with donor peripheral blood cells and recom-binant human interleukin-2 to treat leukaemia relapseafter allogeneic bone marrow transplantation. Blood1996; 87:2195-204.

123. Murphy WJ, Longo DL.The potential role of NK cellsin the separation of graft-versus-tumor effects fromgraft-versus-host disease after allogeneic bone mar-row transplantation. Immunol Rev 1997; 157:167-76.

124. Falkenburg, JH, Smit WM, Willemze R. Cytotoxic T-lymphocyte (CTL) responses against acute or chronicmyeloid leukemia. Immunol Rev 1997; 157:223-30

125. Goulmy E. Human minor histocompatibility antigens:new concepts for marrow transplantation and adop-tive immunotherapy. Immunol Rev 1997; 157:125-40.

126. Faber LM, van der Hoeven J, Goulmy E, et al. Recog-nition of clonogenic leukemic cells, remission bonemarrow and HLA identical donor bone marrow byCD8+ or CD4+ minor histocompatibility antigen-spe-cific cytotoxic T lymphocytes. J Clin Invest 1995; 96:877-83.

127. Faber LM, van Luxemburg-Heijs SAP, Veenhof WFJ,Willemze R, Falkenburg JHF. Generation of CD4+cytotoxic T-lymphocytes clones from a patient withsevere graft-versus-host disease after allogeneic bonemarrow transplantation: implications for graft-versus-leukemia reactivity. Blood 1995; 86:2821-8.

128. Jiang JYZ, Mavroudis DA, Dermine S, Molldrem J,Hensel NF, Barrett AJ. Preferential usage of T cellreceptor (TCR) V by allogeneic T cells recognizingmyeloid leukemia cells: implications for separatinggraft-versus-leukemia effect from graft-versus-host dis-ease. Bone Marrow Transplant 1997; 19:899-903.

129. Barrett AJ, Malkovska V. Graft-versus-leukemia:understanding and using the alloimmune response totreat haematological malignancies. Br J Haematol1996; 93:754-61.

130. Faber LM, van-Luxemburg Heijs SA, Willemze R,Falkenburg, JH. Generation of leukemia-reactive cyto-toxic T lymphocyte clones from the HLA-identicalbone marrow donor of a patient with leukemia. J ExpMed 1992; 176: 1283-9.

131. Montagna D, Locatelli F, Calcaterra V, et al. Does theemergence and persistence of donor derived leukemia-reactive cytotoxic T-lymphocytes protect patients giv-en an allogeneic BMT from recurrence? Results of apreliminary study. Bone Marrow Transplant 1998; 22:743-50.

132. Falkenburg JH, Wafelman AR, van Bergen CAM, et al.Leukemia-reactive cytotoxic T lymphocytes (CTL)induce complete remission in a patient with refracto-ry accelerated phase chronic myeloid leukemia (CML).Blood 1997; 90(Suppl. 1):589a.

133. Bonini C, Ferrari G, Verzelletti S, et al. HSV-TK genetransfer into donor-lymphocytes for control of allo-geneic graft-versus-leukaemia. Science 1997; 276:1719-24.


135. Molldrem J, Dermime S, Parker K, et al. Targeted T-celltherapy for human leukemia: cytotoxic T lymphocytesspecific for a peptide derived from proteinase 3 pref-erentially lyse human myeloid leukemia cells. Blood1996; 88:2450-7.


137. Gimmi CD, Freeman GJ, Gribben JG, Gray G, NadlerLM. Human T-cell clonal anergy is induced by antigenpresentation in the absence of B7 costimulation. ProcNatl Acad Sci USA 1993; 90:6586-90.

138. Noel PJ, Boise LH, Green JM, Thompson CB. CD28costimulation prevents cell death during primary T cellactivation. J Immunol 1996; 157:636-42.

139. Cardoso AA, Schultze JL, Boussiotis VA, et al. Pre-Bacute lymphoblastic leukemia cells may induce T-cellanergy to alloantigen. Blood 1996; 88:41-8.

140. Cardoso AA, Seamon MJ, Afonso HM, et al. Ex vivogeneration of human anti-pre-B leukemia-specificautologous cytolytic T cells. Blood 1997; 90:549-61.

141. Barrett AJ, Mavroudis D, Molldrem J, et al. Optimiz-ing the dose and timing of lymphocytes add-back inT-cell depleted BMT between HLA-identical siblings.Blood 1996; 88 (Suppl. 1): 460a.

142. Aversa F, Tabilio A, Terenzi A, et al. Successful engraft-ment of T-cell-depleted haploidentical “three-loci”incompatible transplants in leukemia patients byaddition of recombinant human granulocyte colony-stimulating factor-mobilized peripheral blood prog-enitor cells to bone marrow inoculum. Blood 1994;84: 3948-55.

143. Gribben JG, Guinan EC, Boussiotis VA, et al. Com-plete blockade of B7 family-mediated costimulation isnecessary to induce human alloantigen-specific aner-gy: a method to ameliorate graft-versus-host diseaseand extend the donor pool. Blood 1996; 87:4887-93.

144. Comoli P, Montagna D, Moretta A, Zecca M, Locatel-li F, Maccario R. Alloantigen-induced human lym-phocytes rendered nonresponsive by a combination ofanti-CD80 monoclonal antibodies and cyclosporin-Asuppress mixed lymphocyte reaction in vitro. JImmunol 1995; 155:5506-11.

145. Comoli P, Locatelli F, Montagna D, et al. Induction ofalloantigen-specific anergy do not impair cytolyticactivity of leukemia-reactive human T cells. Blood1997; 90(suppl. 1):535a.

146. Cavazzana-Calvo M, Fromont C, Le Deist F, et al. Spe-cific elimination of alloreactive T cells by an anti-inter-leukin-2 receptor B chain-specific immunotoxin.Transplantation 1990; 50:1-7.

147. Mickey B, Kohn C, Lowdell MW, Prentice HG. Selec-tive removal of alloreactive lymphocytes from periph-eral blood mononuclear cell preparations. Blood1996; 88(suppl 1):253a.

148. Valteau-Couanet D, Cavazzana-Calvo M, Le Deist F,Fromont C, Fisher A. Functional study of residual Tlymphocytes after specific elimination of alloreactiveT cells by a specific anti-interleukin-2 receptor B chainimmunotoxin. Transplantation 1993; 56:1574-6.

149. Montagna D, Yvon E, Calcaterra V, et al. Depletion ofalloreactive T cells by a specific anti-interleukin-2receptor p55 chain immunotoxin does not impair invitro antileukemia and antiviral activity. Blood 1999;93:3550-7.

150. Cavazzana-Calvo M, Stephan JL, Sarnacki S, et al.Attenuation of graft-versus-host disease and graftrejection by ex vivo immunotoxin elimination of allore-active T cells in an H-2 haplotype disparate mousecombination. Blood 1994; 83:288-98.

151. Groux H, O’Garra A, Bigler M, et al. A CD4+ T-cellsubset inhibits antigen specific T-cell responses andprevents colitis. Nature 1997; 389:737-42.

152. Riddell SR, Greenberg PD. Principles for adoptive Tcell therapy of human viral disease. Annu RevImmunol 1995; 13:545-86.

153. Riddell SR, Watanabe KS, Goodrich JM, Li CR, AghaME, Greenberg PD. Restoration of viral immunity inimmunodeficient humans by the adoptive transfer ofT cell clones. Science 1992; 257:238-41.

154. Locatelli F, Percivalle E, Comoli P, et al. Humancytomegalovirus infection in pediatric patients givenallogeneic bone marrow transplantation: role of ear-ly treatment of antigenemia on patients’ outcome. Br

Cell therapy

150


J Haematol 1994; 88:64-71.155. Goodrich JM, Bowden RA, Fisher L, Keller C, Schoch

G, Meyers JD. Ganciclovir prophylaxis to preventcytomegalovirus disease afetr allogeneic marrowtransplant. Ann Intern Med 1993; 118:173-8.

156. Reusser P, Riddell SR, Meyers JD, Greenberg PD. Cyto-toxic T-lymphocyte response to cytomegalovirus afterhuman allogeneic bone marrow transplantation: pat-tern of recovery and correlation with cytomegalovirusinfection and disease. Blood 1991; 78:1373-80.

157. Li CR, Greenberg PD, Gilbert MJ, Goodrich JM, Rid-dell SR. Recovery of HLA-restricted cytomegalovirus(CMV) specific T cell responses after allogeneic bonemarrow transplantation: correlation with CMV dis-ease and effect of ganciclovir prophylaxis. Blood 1994;83:1971-9.

158. Walter EA, Greenberg PJ, Gilbert MJ, et al. Reconsti-tution of cellular immunity against cytomegalovirusin recipients of allogeneic bone marrow by transfer ofT-cell clones from the donor. N Engl J Med 1995; 333:1038-44.

159. Nalesnik MA. Posttransplantation lymphoprolifera-tive disorders (PTLD): current perspectives. SeminThor Cardiov Surg 1996; 8:139-48.

160. Heslop EE, Rooney CM. Adoptive cellular immuno-therapy for EBV lymphoproliferative diseases.Immunol Rev 1997; 157:217-22.

161. O’Reilly R, Small TN, Papadopulos E, Lucas K, Lacer-da J, Koulova L. Biology and adoptive cell therapy ofEpstein-Barr-virus associated lymphoproliferative dis-orders in recipients of marrow allografts. ImmunolRev 1997; 157:195-216.

162. Hale G, Waldmann H. Risks of developing EpsteinBarr virus-related lymphoproliferative disorders afterT cell depleted marrow transplants. Blood 1998; 91:3079-83.

163. Rooney CM, Loftin SK, Holladay MS, Brenner MK,Krance RA, Heslop HE. Early identification of Epstein-Barr virus associated post-transplant lymphoprolifer-ative disorders. Br J Haematol 1995; 89:98-103.

164. Papadopoulos EB, Ladanyi M, Emanuel D, et al. Infu-sions of donor leukocytes as treatment of Epstein-Barrvirus associated lymphoprolipherative disorders com-plicating allogeneic marrow transplantation. N Engl JMed 1994; 330:1185-91.

165. Rooney CM, Smith CA, Ng CY, et al. Use of gene-mod-ified virus-specific T lymphocytes to control Epstein-Barr-virus-related lymphoproliferation. Lancet 1995;345:9-13.

166. Rooney CM, Smith CA, Ng CYC, et al. Infusion ofcytotoxic T cells for the prevention and treatment ofEpstein Barr virus-induced lymphoma in allogeneictransplant recipients. Blood 1998; 92:1549-55.

167. Heslop HE, Ng CY, Li C, et al. Long-term restorationof immunity against Epstein-Barr virus infection byadoptive transfer of gene-modified virus-specific Tlymphocytes. Nature Med 1996; 2:551-5.

168. Haque T, Amlot PL, Helling N, et al. Reconstitution ofEBV specific T cell immunity in solid organ transplantrecipients. J Immunol 1998; 160:6204-9.

169. Comoli P, Locatelli F, Gerna G, Grossi P, Viganò M,Maccario R. Autologous EBV-specific cytotoxic T cellsto treat EBV-associated post-transplant lymphopro-liferative disease (PTLD) [abstract]. Blood 1997; 90(Suppl.1):249a.

170. Sing AP, Ambinder RF, Hong DJ, et al. Isolation ofEpstein-Barr virus (EBV)-specific cytotoxic T lympho-cytes that lyse Reed-Sternberg cells: implications forimmune-mediated therapy of EBV+ Hodgkin disease.Blood 1997; 89:1978-86.

171. Roskrow MA, Suzuki N, Gan Y, et al. Epstein-Barr virus(EBV)-specific cytotoxic T lymphocytes for the treat-

ment of patients with EBV positive relapsed Hodgkin’sdisease. Blood 1998; 91:2925-34.

172. Smith CA, Woodruff LS, Kitchingman GR, RooneyCM. Adenovirus-pulsed dendritic cells stimulatehuman virus-specific T-cell responses in vitro. J Virol1996; 70:6733-40.

173. Spitzer G, Velasquez W, Dunphy FR, Spencer V. Autol-ogous bone marrow transplantation in solid tumors.Curr Opin Oncol 1992; 4:272-8.

174. Jones RJ. Autologous bone marrow transplantation.Curr Opin Oncol 1993; 5:270-5.

175. O'Reilly R. Bone marrow transplantation. Curr OpinHematol 1993; 1:221-2.

176. Thomas ED, Clift RA, Fefer A, et al. Marrow trans-plantation for the treatment of chronic myelogenousleukemia. Ann Intern Med 1986; 104:155-63.

177. Santos GW, Hess AD, Volgelsang GB. Graft-versus-host reactions and disease. Immunol Rev 1985; 88:169-92.

178. Kernan NA, Bordignon C, Collins NH, et al. Bone mar-row failure in HLA-identical T-cell depleted allogene-ic transplants for leukemia: I Clinical aspects. Blood1989; 74:2227-36.

179. Goldman JM, Gale RP, Horowitz MM, et al. Bonemarrow transplantation for chronic myelogenousleukemia in chronic phase. Ann Intern Med 1988;108: 806-14.

180. Zutter MM, Martin PJ, Sale GE, et al. Epstein-Barrvirus lymphoproliferation after bone marrow trans-plantation. Blood 1988; 72:520-9.

181. Shapiro RS, McClain K, Frizzera G, et al. Epstein-Barrvirus associated lymphoproliferative disorders follow-ing bone marrow transplantation. Blood 1988; 71:1234-43.

182. Tosato G. The Epstein-Barr virus and the immune sys-tem. Adv Cancer Res 1987; 49:75-125.

183. Heslop HE, Brenner MK, Rooney CM. Donor T cellsto treat EBV-associated lymphoma [letter]. N Engl JMed 1994; 331:679-80.

184. Smith CA, Heslop E, Hollyday MS, et al. Productionof genetically modified EBV-specific cytotoxic T cellsfor adoptive transfer to patient at high risk of EBV-associated lymphoproliferative disease. J Hematother1995; 4:473-9.

185. Servida P, Rossini S, Traversari C, et al. Gene transferinto peripheral blood lymphocytes for in vivo immu-nomodulation of donor anti-tumor immunity in apatient affected by EBV-induced lymphoma [abstract].Blood 1993; 82:214a.

186. Bordignon C, Bonini C, Verzeletti S, et al. Transfer ofthe HSV-tk gene into donor peripheral blood lym-phocytes for in vivo modulation of donor anti-tumorimmunity after allogeneic bone marrow transplanta-tion. Hum Gene Ther 1995; 6:813-9.

187. Moolten FL. Drug sensitivity (“suicide”) genes forselective cancer chemotherapy. Cancer Gene Ther1994; 1:279-87.

188. Tiberghien P, Reynolds CW, Keller J, et al. Ganciclovirtreatment of herpes simplex thymidine kinase-trans-duced primary T lymphocytes: an approach for spe-cific in vivo donor T-cell depletion after bone marrowtransplantation? Blood 1994; 84:1333-41.

189. Verzeletti S, Bonini C, Traversari C, et al. Transfer ofthe HSV-tk gene into donor peripheral blood lym-phocytes for in vivo immunomodulation of donoranti-tumor immunity after allo-BMT. Hum Gene Ther1998; 6:813-9.

190. Mavilio F, Ferrari G, Rossini S, et al. Peripheral bloodlymphocytes as target cells of retroviral vector-medi-ated gene transfer. Blood 1994; 83:1988-97.

191. Sullivan KM, Storb R, Buckner CD, et al. Graft-versus-host disease as adoptive immunotherapy in patients

C. Bordignon et al.

151


Cell therapy

with advance hematologic neoplasms. N Engl J Med1989; 320:828-34.

192. Porter DL, Roth MS, Mc Garigle C, Ferrara JLM, AntinJH. Induction of a graft-versus-host disease as immu-notherapy for relapsed chronic myeloid leukemia. NEngl J Med 1994; 330:100-6.

193. Drobyski WR, Keever CA, Roth MS, et al. Salvageimmunotherapy using donor leukocyte infusions astreatment for relapsed chronic myelogenous leukemiaafter allogeneic bone marrow transplantation: effica-cy and toxicity of a defined T-cell dose. Blood 1993;82:2310-8.

194. Tricot G, Vesole DH, Jagganath S, Hilton J, Munshi N,Barlogie B. Graft-versus-myeloma effect: proof of prin-ciple. Blood 1996; 87:1196-8.

195. Riddell SR, Elliott M, Lewinsohn DA, et al. T-cell medi-ated rejection of gene-modified HIV-specific cytotox-ic lymphocytes in HIV-infected patients. Nature Med1996; 2:216-23.

196. Bonini C, Verzelletti S, Servida P, et al. Immunityagainst the transgene product may limit efficacy ofHSV-tk-transduced donor peripheral blood lympho-cytes after allo-BMT [abstract]. Blood 1995; 84:628a.

197. Phillips K, Gentry T, McCowange G, Gilboa E, SmithC. Cell-surface markers for assessing gene transfer intohuman hematopoietic cells. Nature Med 1996; 2:1154-6.

198. Ciceri F, Marktel S, Bonini C, et al. HSV-TK genetical-ly engineered donor lymphocytes restore anti-viralimmunity early after T-depleted BMT [abstract].Blood 1998; 92:667a.

199. Bancherau J, Steinman, RM. Dendritic cells and thecontrol of immunity. Nature 1998; 392:245-52.

200. Guery JC, Adorini L. Dendritic cells are the most effi-cient in presenting endogenous naturally processedself-epitopes to class II-restricted T cells. J Immunol1995; 154:536-44.

201. Inaba K, Metlay JP, Crowley MT, Steinman RM. Den-dritic cells pulsed with protein antigens in vitro canprime antigen-specific MHC-restricted T cells in situ.J Exp Med 1990; 172:631-40.

202. Young JW, Steiman RM. Dendritic cells stimulate pri-mary human cytolytic lymphocyte responses in theabsence of CD4+ helper T cells. J Exp Med 1990; 171:1315-20.

203. Bhardwaj N, Bender A, Gonzales N, et al. Influenzavirus-infected dendritic cells stimulate strong prolifer-ative and cytolytic responses from human CD8+ Tcells. J Clin Invest 1994; 94:797-801.

204. Metha Damani A, Markowicz S, Engleman EG. Gen-eration of antigen-specific CD8+ CTLs from naive pre-cursors. J Immunol 1994; 153:996-1003.

205. Winzler C, Rovere P, Rescigno M, et al. Maturationstages of mouse dendritic cells in growth factor-dependent long-term cultures. J Exp Med 1997; 185:317-28.

206. Macatonia SE, Hosken NA, Litton M, et al. Dendriticcells produce IL-12 and direct the development ofTH1 cells from the naive CD4+ T cells. J Immunol1995; 154:5071-9.

207. Cella M, Scheidegger D, Palmer-Lehmann K, et al. Lig-ation of CD40 on dendritic cells triggers productionof high levels of interleukin-12 and enhances T cellstimulatory capacity: T-T help via APC activation. JExp Med 1996; 184:747-52.

208. Ridge JP, Di Rosa F, Matzinger P. A conditioned den-dritic cell can be a temporal bridge between a CD4+T-helper and a T-killer cell. Nature 1998; 394:474-8.

209. Bennet SRM, Carbone FR, Karamalis F, Flavell RA,Miller JFAP, Hearth WR. Help for cytotoxic T-cellresponses is mediated by CD40 signalling. Nature1998; 394: 478-80.

210. Schoenberger SP, Toes REM, van der Voort EIH,Offringa R, Melief CJM. T-cell help for cytotoxic T lym-phocytes is mediated by CD40-CD40L interactions.Nature 1998; 394:480-3.

211. Steinman RM, Swanson J. The endocytic activity ofdendritic cells. J Exp Med 1995; 182:283-8.

212. Koch F, Stanzl U, Jennewein P, et al. High level IL-12production by murine dendritic cells: upregulation viaMHC class II and CD40 molecules and down regula-tion by IL-4 and IL-10. J Exp Med 1996; 184:741-7.

213. Gabrilovich DI, Chen HL, Girgis KR , et al. Produc-tion of vascular endothelial growth factor by humantumors inhibits the functional maturation of dendrit-ic cells. Nature Med 1996; 2:1096-103.

214. Romani N, Gruner S, Brang D, et al. Proliferating den-dritic cell progenitors in human blood. J Exp Med1994; 180:83-93.

215. Sallusto F, Lanzavecchia A. Efficient presentation ofsoluble antigen by cultured human dendritic cells ismaintained by granulocyte/macrophage colony-stim-ulating factor plus interleukin-4 and down regulatedby tumor necrosis factor alpha. J Exp Med 1994; 179:1109-18.

216. Reid CDL, Stackpole A, Meager A, Tikepae J. Interac-tion of tumor necrosis factor with granulocyte-macrophage colony-stimulating factor and othercytokines in the regulation of dendritic cell growth invitro from early bipotent CD34+ progenitors inhuman bone marrow. J Immunol 1992; 149:2681-8.

217. Caux C, Dezutter-Dambuyant S, Schmitt D, Bancher-au J. GM-CSF and TNF-a cooperate in the generationof dendritic Langherans cells. Nature 1992; 360:258-61.

218. Strunk D, Rappersberger K, Egger C, et al. Generationof human dendritic cells/Langherans cells from circu-lating CD34+ hematopoietic progenitor cells. Blood1996; 87:1292-302.

219. Szabolcs P, Moore MAS, Young JW. Expansion ofimmunostimulatory dendritic cells among the myeloidprogeny of human CD34+ bone marrow precursorscultured with c-kit ligand, granulocyte-macrophagecolony-stimulating factor, and TNF-a. J Immunol1995; 154:5851-61.


221. Fisch P, Kohler G, Garbe A, et al. Generation of anti-gen-presenting cells for soluble protein antigens exvivo from peripheral blood CD34+ cells hematopoi-etic progenitor cells in cancer patients. Eur J Immunol1996; 26:595-600.

222. Romani N, Reider D, Heuer M, et al. Generation ofmature dendritic cells from human blood. Animproved method with special regard to clinicalapplicability. J Immunol Methods 1996; 196:137-51.

223. Bender A, Sapp M, Schuler G, Steinman RM, BhardwajN. Improved methods for the generation of dendriticcells from non-proliferating progenitors in humanblood. J Immunol Methods 1996; 196:121-35.

224. Bhardwaj N, Young JW, Nisanian AJ, Biggers J, Stein-man RM. Small amounts of superantigen, when pre-sented on dendritic cells, are sufficient to initiate Tcell responses. J Exp Med 1993; 178:633-42.

225. Ratta M, Rondelli D, Fortuna A, et al. Generation andfunctional characterization of human dendritic cellsderived from CD34+ mobilized into peripheral blood:comparison with bone marrow CD34+ cells. Br JHaematol 1998;101:756-65.

226. Rosenzwajg M, Canque B, Gluckman JC. Human den-dritic cell differentiation pathway from CD34+ hema-topoietic precursor cells. Blood 1996; 87:535-44.

152


227. Haug JS, Todd G, Bremer R, Link D, Brown R, DiPer-sio JF. Mobilization of CD80+ dendritic cells into theperipheral circulation by GM-CSF but not G-CSF[abstract]. Blood 1998; 92 (Suppl 1):444a.

228. Lebsack ME, Maraskowsky E, Roux E, et al. Increasedcirculating dendritic cells in healthy human volunteersfollowing administration of FLT3 ligand alone or incombination with GM-CSF or G-CSF [abstract].Blood 1998; 92 (Suppl 1):507a.

229. Mortarini R, Anichini A, Di Nicola M, et al. Autolo-gous dendritic cells derived from CD34+ progenitorsand monocytes are not functionally equivalent APC inthe induction of Melan-A/Mart-27-35-specific CTLfrom PBL of melanoma patients with low frequency ofCTL precursors. Cancer Res 1997; 57:5534-41.

230. Gong J, Chen D, Kashiwaba M, Kufe D. Induction ofantitumor activity by immunization with fusions ofdendritic and carcinoma cells. Nature Med 1997; 3:558-61.

231. Albert LM, Sauter B, Bhardwaj N. Dendritic cellsacquire antigen from apoptotic cells and induce classI-restricted CTLs. Nature 1998; 392:86-9.

232. Reeves ME, Royal RE, Lam JS, Rosenberg SA, Hwu P.Retroviral transduction of human dendritic cells witha tumor-associated antigen gene. Cancer Res 1996;56:5672-7.

233. Gong J, Chen L, Chen D, et al. Induction of antigen-specific antitumor immunity with adenovirus-trans-duced dendritic cells. Gene Ther 1998; 4:1023-8.

234. Dietz AB, Vuk-Pavlovic S. High efficiency adenovirus-mediated gene transfer to human dendritic cells.Blood 1998; 91:392-8.

235. Di Nicola M, Siena S, Bregni M, et al. Gene transferinto human dendritic antigen-presenting cells by vac-cinia virus and adenovirus vectors. Cancer Gene Ther1998; 5:350-6.

236. Akagi J, Hodge JW, McLaughlin JP, et al. Therapeuticantitumor response after immunization with anadmixture of recombinant vaccinia viruses expressinga modified MUC1 gene and the murine T-cell costim-ulatory molecule B-7. J Immunother 1997; 20:38-47.

237. McLaughlin JP, Schlom J, Kantor JA, et al. Improvedimmunotherapy of a recombinant carcinoembryonicantigen vaccinia vaccine when given in combinationwith interleukin-2. Cancer Res 1997; 56:2361-7.

238. Borysiewicz LK, Fiander A, Nimako M, et al. A recom-binant vaccinia virus encoding human papillomavirustypes 16 and 18, E6 and E7 proteins as immunother-apy for cervical cancer. Lancet 1996; 347:1523-7.

239. McAneny D, Ryan CA, Beazley RM, et al. Results of aphase I trial of a recombinant vaccinia virus thatexpresses carcinoembryonic antigen in patients withadvanced colorectal cancer. Ann Surg Oncol 1996; 3:495-500.

240. Rescigno M, Citterio S, Thery C, et al. Bacteria-induced neo-biosynthesis, stabilization, and surfaceexpression of functional class I molecules in mousedendritic cells. Proc Natl Acad Sci USA 1998; 95:5229-34.

241. Boczkowski D, Nair KS, Snyder D, Gilboa E. Dendrit-ic cells pulsed with RNA are potent antigen-present-ing cells in vitro and in vivo. J Exp Med 1996; 184:465-72.

242. Mayordomo JL, Zorina T, Storkus WJ, et al. Bone mar-row-derived dendritic cells pulsed with synthetictumor peptides elicit protective and therapeutic anti-tumor immunity. Nature Med 1995; 1:1297-302.

243. Celluzzi CM, Mayordomo JL, Storkus WJ, Lotze MT,Falo LD. Peptide-pulsed dendritic cells induce anti-gen-specific CTL-mediated protective tumor immuni-ty. J Exp Med 1996; 183:283-7.

244. Paglia P, Chiodoni C, Rodolfo M, Colombo MP.

Murine dendritic cells loaded in vitro with soluble pro-tein prime cytotoxic T lymphocytes against tumor anti-gen in vivo. J Expl Med 1996; 183:317-22.

245. Porgador A, Snyder D, Gilboa E. Induction of antitu-mor immunity using bone marrow-generated den-dritic cells. J Immunol 1996; 156:2918-26.

246. Flamand V, Sornasse T, Thielemans K, et al. Murinedendritic cells pulsed in vitro with tumor antigeninduce tumor resistance in vivo. Eur J Immunol 1994;24:605-10.

247. Mukherji B, Charkraborty NG, Yamasaki S, et al.Induction of antigen specific cytolytic T cells in situ inhuman melanoma by immunization with syntheticpeptide-pulsed autologous antigen presenting cells.Proc Natl Acad Sci USA 1995; 92:8078-82.

248. Hu X, Charkraborty NG, Sporn JR, et al. Enhancementof cytolytic T lymphocyte precursors frequency inmelanoma patients following immunization with theMAGE-1 peptide loaded antigen presented-based vac-cine. Cancer Res 1996; 56:2479-83.

249. Nestle FO, Alijagic S, Gilliet M, et al. Vaccination ofmelanoma patients with peptide-or tumor lysate-pulsed dendritic cells. Nature Med 1998; 4:328-32.

250. Salgaller ML, Tjoa BA, Lodge PA, et al. Dendritic cell-based immunotherapy of prostate cancer. Crit RevImmunol 1998; 18:109-19.

251. Tao MH, Levy R. Idiotype/granulocyte-macrophagecolony-stimulating factor fusion protein as a vaccinefor B-cell lymphoma. Nature 1993; 362:755-8.

252. Campbell MJ, Esserman L, Byars NE, et al. Idiotypevaccination against murine B cell lymphoma.Humoral and cellular requirements for the full expres-sion of antitumor immunity. J Immunol 1990;145:1029-36.

253. Hsu FJ, Caspar CB, Czerwinsky D, et al. Tumor-spe-cific idiotype vaccines in the treatment of patients withB-cell lymphoma. Long-term results of a clinical study.Blood 1997; 89:3129-35.

254. Hsu FJ, Benike C, Fagnoni F, et al. Vaccination ofpatients with B-cell lymphoma using autologous anti-gen-pulsed dendritic cells. Nature Med 1996; 2:52-8.

255. Liso A, Stockerl-Goldstein KE, Reichardt VL, et al. Idio-type vaccination using dendritic cells after autologousperipheral blood progenitor cell transplantation formultiple myeloma [abstract]. Blood 1998; 92(Suppl.1):105a.

256. Kaufmann SHE. Immunity to intracellular bacteria.Annu Rev Immunol 1993; 11:129-63.

257. Paglia P, Arioli I, Frahm N, Chakraborty T, ColomboMP, Guzman CA. The defined attenuated Listeriamonocytogenes ∆mp12 mutant is an effective oral vac-cine carrier to trigger a long-lasting immune responseagainst a mouse fibrosarcoma. Eur J Immunol 1997;27:1570-5.

258. Paglia P, Medina E, Arioli I, Guzman CA, ColomboMP. Oral DNA vaccination with Salmonellatyphimurium mediates gene transfer to antigen pre-senting cells and results in protective immunity againsta murine fibrosarcoma. Blood 1998; 92:3172-6.


260. Hellstrom KE, Hellstrom I, Chen L. Can co-stimulat-ed tumor immunity be therapeutically efficacious?Immunol Rev 1995; 145:123-45.

261. Lauritzsen GF, Hofgaard PO, Scenck K, Bogen B. Clon-al deletion of thymocytes as a tumor escape mecha-nism. Int J Cancer 1998; 78:216-22.

262. Nagata S. Fas ligand and immune evasion. Nat Med1996; 2:1306-17.

263. Chen L, Ashe S, Brady WA, et al. Costimulation of anti-tumor immunity by the B7 counter-receptor for the Tlymphocyte molecules CD28 and CTLA-4. Cell 1992;

C. Bordignon et al.

153


71:1093-102.264. Townsend SE, Allison JP. Tumor rejection after direct

costimulation of CD8+ T cells by B7-transfected mela-noma cells. Science 1993; 259:368-70.

265. Lanzavecchia A. Antigen-specific interactions betweenT and B cells. Nature 1985; 314:537-9.

266. Kurt-Jones EA, Liano D, Hayglass KA, Benacerraf B,MS Sy, Abbas AK. The role of antigen-presenting Bcells in T cell priming in vivo. Studies of B cell-deficientmice. J Immunol 1988; 140:3773-8.

267. Schwartz RH. Models of T cell anergy: is there a com-mon molecular mechanism? J Exp Med 1996;184:1-8.

268. Noelle RJ. CD40 and its ligand in host defense. Immu-nity 1996; 4:415-9.

269. Yang Y, Wilson JM. CD40 ligand-dependent T cell acti-vation: requirement of B7-CD28 signaling throughCD40. Science 1996; 273:1864-7.

270. Ranheim EA, Kipps TJ. Activated T cells induce expres-sion of B7/BB1 on normal or leukemic B cells througha CD40-dependent signal. J Exp Med 1993; 177:925-35.

271. Yellin MJ, Sinning J, Covey LR, et al. T lymphocyte Tcell-B cell-activating molecule/CD40-L moleculesinduce normal B cells to express CD80 (B7/BB-1) andenhance their costimulatory activity. J Immunol 1994;153:666-74.

272. Ranheim EA, Kipps TJ. Tumor necrosis factor-alphafacilitates induction of CD80 (B7-1) and CD54 onhuman B cells by activated T cells: complex regulationby IL-4, IL-10, and CD40L. Cell Immunol 1995; 161:226-35.

273. Cantwell MJ, Sharma S, Friedmann T, Kipps TJ. Ade-novirus vector infection of chronic lymphocyticleukemia B cells. Blood 1996; 88:4676-83.

274. Funakoshi S, Longo DL, Beckwith M, et al. Inhibitionof human B-cell lymphoma growth by CD40 stimula-tion. Blood 1994; 83:2787-94.

275. Vyth-Dreese FA, Dellemijn TAM, van Oostveen JW,Feltkamp CA, Hekman A. Functional expression ofadhesion receptors and costimulatory molecules byfresh and immortalized B-cell non-Hodgkin lym-phoma cells. Blood 1995; 85:2802-12.

276. Schultze JL, Cardoso AA, Freeman GJ, et al. Follicularlymphomas can be induced to present alloantigen effi-ciently: a conceptual model to improve their tumorimmunogenicity. Proc Natl Acad Sci USA 1995; 92:8200-4.

277. Inge TH, Hoover SK, Susskind BM, Barret SK, BearHD. Inhibition of tumor-specific cytotoxic T-lympho-cyte responses by transforming growth factor beta 1.Cancer Res 1992; 52:1386-92.

278. Harada M, Matsunaga K, Oguchi Y, et al. The involve-ment of transforming growth factor beta in impairedantitumor T-cell response at the gut-associated lym-phoid tissue (GALT). Cancer Res 1995;55:6146-51.

279. Rondelli D, Andrews RG, Hansen JA, Ryncarz R, Faer-ber MA, Anasetti C. Alloantigen presenting function ofnormal human CD34+ hematopoietic cells. Blood1996; 88:2619-25.

280. Rondelli D, Anasetti C, Fortuna A, et al. T cell allore-activity induced by normal G-CSF-mobilized CD34+blood cells. Bone Marrow Transplant 1998; 21:1183-91.

281. Ryncarz R, Anasetti C. Expression of CD86 on humanmarrow CD34+ cells identifies immunocompetentcommitted precursors of macrophages and dendriticcells. Blood 1998; 91:3892-900.

282. Bocchia M, Wentworth PA, Southwood S, et al. Spe-cific binding of leukemia fusion protein peptides toHLA class I molecules. Blood 1995; 85:2680-4.

283. Bocchia M, Korontsvit T, Xu Q, et al. Specific humancellular immunity to bcr-abl oncogene-derived pep-

tides. Blood 1996; 87:3587-92.284. Pawelec G, Max H, Halder T, et al. BCR/ABL leukemia

oncogene fusion peptides selectively bind to certainHLA-DR alleles and can be recognized by T cells foundat low frequency in the repertoire of normal donors.Blood 1996; 88:2118-24.

285. Mannering SI, McKenzie JL, Fearnley DB, Hart DNJ.HLA-DR-restricted bcr-abl (b3a2)-specific CD4+ Tlymphocytes respond to dendritic cells pulsed withb3a2 peptide and antigen-presenting cells exposed tob3a2 containing cell lysates. Blood 1997; 90:290-7.

286. Papadopoulos KP, Suciu-Foca N, Hesdorffer CS,Tugulea S, Maffei A, Harris PE. Naturally processedtissue- and differentiation stage-specific autologouspeptides bound by HLA class I and II molecules ofchronic myeloid leukemia blasts. Blood 1997; 90:4938-46.

287. Eibl B, Ebner S, Duba C, et al. Dendritic cells gener-ated from blood precursors of chronic myelogenousleukemia patients carry the Philadelphia translocationand can induce a CML-specific primary cytotoxic T-cell response. Genes Chromosomes Cancer 1997;20:215-23.

288. Choudhury A, Gajewski JL, Liang JC, et al. Use ofleukemic dendritic cells for the generation of anti-leukemic cellular cytotoxicity against Philadelphiachromosome-positive chronic myelogenous leukemia.Blood 1997;89:1133-42.

289. Choudhury A, Toubert A, Sutaria S, Charron D,Champlin RE, Claxton DF. Human leukemia-deriveddendritic cells: ex-vivo development of specificantileukemic cytotoxicity. Crit Rev Immunol 1998; 18:121-31.

290. Smit WM, Rijnbeek M, van Bergen CAM, et al. Gen-eration of dendritic cells expressing bcr-abl fromCD34-positive chronic myeloid leukemia precursorcells. Hum Immunol 1997; 53:216-23.

291. Carlo-Stella C, Garau D, Regazzi E, et al. Generationof BCR/ABL positive dendritic cells from chronic myel-ogenous leukemia CD34+ cells [abstract]. Blood1998;92 (Suppl.1):2590.

292. Misery L, Campos L, Dezutter-Dambuyant C, et al.CD1-reactive leukemic cells in bone marrow: presenceof Langherans cell marker on leukemic monocyticcells. Eur J Haematol 1992; 48:27-32.

293. Santiago-Schwarz F, Coppock DL, Hindenburg AA,Kern J. Identification of a malignant counterpart ofthe monocyte-dendritic cell progenitor in an acutemyeloid leukemia. Blood 1994; 84:3054-62.

294. Strunk D, Linkesch W. Acute myelogenous leukemiacells can be differentiated into dendritic cells in vitro[abstract]. Blood 1997; 90 (Suppl.1):829.

295. Matulonis UA, Dosiou C, Lamont C, et al. Role of B7-1 in mediating an immune response to myeloidleukemia cells. Blood 1995; 85:2507-15.

296. Matulonis UA, Dosiou C, Freeman G, et al. B7-1 issuperior to B7-2 costimulation in the induction andmaintenance of T-cell mediated anti-leukemia immu-nity. Further evidence that B7-1 and B7-2 are func-tionally distinct. J Immunol 1996; 156:1126-31.

297. Dunussi-Joannopoulos K, Weistein HJ, Nickerson PW,et al. Irradiated B7-1 transduced primary acute myel-ogenous leukemia (AML) cells can be used as thera-peutic vaccines in murine AML. Blood 1996; 87:2938-46.

298. Mutis T, Schrama E, Melief CJM, Goulmy E. CD80-transfected acute myeloid leukemia cells induce pri-mary allogeneic T-cell responses directed at patient spe-cific minor histocompatibility antigens and leukemia-associated antigens. Blood 1998; 92:1677-84.

299. Li Y, Hellstrom KE, Newby SA, Chen L. Costimulationby CD48 and B7-1 induces immunity against poorly

Cell therapy

154


immunogenic tumors. J Exp Med 1996; 183:639-44.300. Colombo MP, Ferrari G, Stoppacciaro A, et al. Gran-

ulocyte colony-stimulating factor gene transfer sup-presses tumorigenicity of a murine adenocarcinoma invivo. J Exp Med 1991; 173:889-97.

301. Zilocchi C, Stoppacciaro A, Chiodoni C, Parenza M,Terrazzini N, Colombo MP. Interferon gamma-inde-pendent rejection of interleukin 12-transduced carci-noma cells requires CD4+ T cells and granulo-cyte/macrophage colony-stimulating factor. J ExpMed 1998; 188:133-43.

302. Zitvogel L, Robbins PD, Storkus WJ, et al. Interleukin-12 and B7.1 co-stimulation cooperate in the induc-tion of effective antitumor immunity and therapy ofestablished tumors. Eur J Immunol 1996; 26:1335-41.

303. Dunussi-Joannopoulos K, Dranoff G, Weinstein HJ,Ferrara JL, Bierer BE, Croop JM. Gene immunothera-py in murine acute myeloid leukemia: granulocyte-macrophage colony-stimulating factor tumor cell vac-cines elicit more potent antitumor immunity com-pared with B7 family and other cytokine vaccines.Blood 1998; 91:222-30.

304. Grange JM, Stanford JL, Rook GA. Tuberculosis andcancer: parallels in host responses and therapeuticapproaches? Lancet 1995; 345:1350-2.

305. Boon T, van der Bruggen P. Human tumor antigensrecognized by T lymphocytes. J Exp Med 1996; 183:725-30.

306. Forni G, Giovarelli M, Cavallo F, et al. Cytokineinduced tumor immunogenicity: from exogenouscytokines to gene therapy. J Immunother 1993; 14:253-7.

307. Huang YC, Golumbeck P, Ahmadzadeh M, Jaffee E,Pardoll D, Levitsky H. Role of bone-marrow derivedcells in presenting MHC class I-restricted tumor anti-gens. Science 1994; 264:961-5.

308. Arienti F, Sulé-Suso J, Belli F, et al Limited antitumorT cell response in melanoma patients vaccinated withinterleukin-2 gene-transduced allogeneic melanomacells. Hum Gene Ther 1996; 7:1955-63.

309. Morton DL, Foshag LJ, Hoon DSB, et al. Prolongationof survival in metastatic melanoma after active specificimmunotherapy with a new polyvalent melanoma vac-cine. Ann Surg 1992; 216:463-82.

310. Marchand M, Weynants P, Rankin E.Tumor regres-sion responses in melanoma patients treated with apeptide encoded by gene MAGE-3. Int J Cancer 1995;63:883-5.

311. Rodolfo M, Melani C, Zilocchi C, et al. IgG2a inducedby IL-12-producing tumor cell vaccines but not IgG1induced by IL-4 vaccine are associated with the erad-ication of experimental metastases. Cancer Res 1998;58:5812-7.

312. Rodolfo M, Zilocchi C, Cappetti B, Parmiani G,Melani C, Colombo MP. Eradication of experimentalmetastases by IL-12-transduced tumor vaccine is asso-ciated with GM-CSF producing CD8 lymphocytes rec-ognizing tumor antigens that are not immunoselect-ed. Gene Therapy 1999, in press.

313. Colombo MP, Forni G. Cytokine gene transfer intumor inhibition and tentative tumor therapy: Whereare we now? Immunol Today 1994; 15:48-51.

314. Scott MA, Gordon MY. In search of the haemopoiet-ic stem cell. Br J Haematol 1995; 90:738-43.

315. Gronthos S, Simmons PJ. The biology and applica-tion of human bone marrow stromal cell precursors.J Hematother 1996; 5:15-23.

316. Dexter TM. Regulation of hemopoietic cell growth anddevelopment: experimental and clinical studies.Leukemia 1989; 3:469-74.

317. Morrison SJ, Shah NM, Anderson DJ. Regulatory

mechanisms in stem cell biology. Cell 1997; 88:287-98.

318. Trentin JJ. Influence of hematopoietic organ stroma(hematopoietic inductive microenvironments) onstem cell differentiation. In: Gordon AS, ed. Regula-tion of hematopoiesis. vol 1. New York: Appleton,1970. p. 161-185.

319. Simmons PJ, Zannettino A, Gronthos S, Leavesley D.Potential adhesion mechanisms for localisation ofhaemopoietic progenitors to bone marrow stroma.Leuk Lymphoma 1994; 12:353-63.

320. Gordon MY, Riley GP, Watt SM, Greaves MF. Com-partimentalization of a haemopoietic growth factor(GM-CSF) by glycosaminoglycans in the bone mar-row microenvironment. Nature 1987; 326:403-5.

321. Carlo-Stella C, Tabilio A. Stem cells and stem celltransplantation. Haematologica 1996; 81:573-87.

322. Gordon MY. Physiological mechanisms in BMT andhaemopoiesis - revisited. Bone Marrow Transplant1993; 11:193-7.

323. Schofield R. The relationship between the haemopoi-etic stem cell and the spleen colony-forming cell: ahypothesis. Blood Cells 1978; 4:7-25.

324. McGinnes K, Quesniaux V, Hitzler J, Paige C. HumanB-lymphopoiesis is supported by bone marrow-derived stromal cells. Exp Hematol 1991; 19:294-303.

325. Landreth KS, Dorshkind K. Pre-B cell generationpotentiated by soluble factors from a bone marrowstromal cell line. J Immunol 1988; 140:845-52.

326. Kierney PC, Dorshkind K. B lymphocyte precursorsand myeloid progenitors survive in diffusion chambercultures but B cell differentiation requires close asso-ciation with stromal cells. Blood 1987; 70:1418-24.

327. Touw I, Löwenberg B. Production of T lymphocytecolony-forming units from precursors in human long-term bone marrow cultures. Blood 1984; 64:656-61.

328. Dorshkind K, Johnson A, Collins L, Keller GM, PhillipsRA. Generation of purified stromal cell cultures thatsupport lymphoid and myeloid precursors. J ImmunolMethods 1986; 89:37-47.

329. Barda-Saad M, Rozenszajn LA, Globerson A, ZhangAS, Zipori D. Selective adhesion of immature thymo-cytes to bone marrow stromal cells: relevance to T celllymphopoiesis. Exp Hematol 1996; 24:386-91.

330. Prockop DJ. Marrow stromal cells as stem cells fornonhematopoietic tissues. Science 1997; 276:71-4.

331. Owen ME, Cave J, Joyner CJ. Clonal analysis in vitro ofosteogenic differentiation of CFU-F. J Cell Sci 1987;87:731-9.

332. Owen ME, Friedenstein AJ. Stromal stem cells: mar-row-derived osteogenic precursors. CIBA Found Symp1988; 136:42-60.

333. Bennet JH, Joyner CJ, Triffitt JT, Owen ME. Adipocytecells cultured from marrow have osteogenic potential.J Cell Sci 1991; 99:131-9.

334. Castro-Malaspina H, Gay RE, Resnick G, et al. Char-acterization of human bone marrow fibroblastcolony-forming cells (CFU-F) and their progeny. Blood1980; 56:289-301.

335. Friedenstein AJ, Chailakhjan RK, Lalykina KS. Thedevelopment of fibroblast colonies in monolayer cul-tures of guinea-pig bone marrow and spleen cells. CellTissue Kinet 1970; 3:393-403.

336. Ashton BA, Allen TD, Howlett CR, et al. Formation ofbone and cartilage by marrow stromal cells in diffu-sion chambers in vivo. Clin Orthop 1980; 151:294-307.

337. Friedenstein AJ, Chailakhyan RK, Latsinik NV, et al.Stromal cells responsible for transferring the microen-vironment of the hemopoietic tissues. Cloning in vit-ro and retransplantation in vivo. Transplantation1974; 17:331-40.

C. Bordignon et al.

155


338. Gronthos S, Graves SE, Ohta S, Simmons PJ. TheSTRO-1+ fraction of adult human bone marrow con-tains the osteogenic precursors. Blood 1994; 84:4164-73.

339. Owen M. Lineage of osteogenic cells and their rela-tionship to the stromal system. J Bone Miner Res1985; 3:1-12.

340. Simmons PJ, Torok-Storb B. Identification of stromalcell precursors in human bone marrow by a novelmonoclonal antibody, STRO-1. Blood 1991; 78:55-62.

341. Simmons PJ, Torok-Storb B. CD34 expression by stro-mal precursors in normal human adult bone marrow.Blood 1991; 78:2848-53.

342. Majumdar MK, Thiede MA, Mosca JD, Moorman M,Gerson SL. Phenotypic and functional comparison ofcultures of marrow-derived mesenchymal stem cells(MSCs) and stromal cells. J Cell Physiol 1998; 176:57-66.

343. Almeida-Porada G, Ascensao JL, Zanjani ED. The roleof sheep stroma in human haemopoiesis in thehuman/sheep chimaeras. Br J Haematol 1996; 93:795-802.

344. Gan OI, Murdoch B, Larochelle A, Dick JE. Differen-tial maintenance of primitive human SCID-repopu-lating cells, clonogenic progenitors, and long-termculture-initiating cells after incubation on humanbone marrow stromal cells. Blood 1997; 90:641-50.

345. Li S, Champlin R, Fitchen JH, Gale RP. Abnormalitiesof myeloid progenitor cells after "successful" bonemarrow transplantation. J Clin Invest 1984; 75:234-41.

346. O'Flaherty E, Sparrow R, Szer J. Bone marrow stro-mal function from patients after bone marrow trans-plantation. Bone Marrow Transplant 1995; 15:207-12.

347. Domenech J, Gihana E, Dayan A, et al. Haemopoiesisof transplanted patients with autologous marrowsassessed by long-term marrow culture. Br J Haematol1994; 88:488-96.

348. Carlo-Stella C, Tabilio A, Regazzi E, et al. Effect ofchemotherapy for acute myelogenous leukemia onhematopoietic and fibroblast marrow progenitors.Bone Marrow Transplant 1997; 20:465-71.

349. Keating A, Horsfall W, Hawley RG, Toneguzzo F.Effect of different promoters on expression of genesintroduced into hematopoietic and marrow stromalcells by electroporation. Exp Hematol 1990; 18:99-102.

350. van Beusechem VW, Kukler A, Heidt PJ, Valerio D.Long-term expression of human adenosine deaminasein rhesus monkeys transplanted with retrovirus-infect-ed bone-marrow cells. Proc Natl Acad Sci U S A 1992;

89:7640-4.351. Klyushnenkova E, Mosca JD, McIntosh KR, Thiede

MA. Human mesenchymal stem cells suppress allo-geneic T cell responses in vitro: implications for allo-geneic transplantation [abstract]. Blood 1998; 92(suppl 1):2652.

352. Gronthos S, Simmons PJ. The growth factor require-ments of STRO-1-positive human bone marrow stro-mal precursors under serum-deprived conditions invitro. Blood 1995; 85: 929-40.

353. Perkins S, Fleischman RA. Hematopoietic microenvi-ronment: origin, lineage, and transplantability of thestromal cells in long-term bone marrow cultures fromchimeric mice. J Clin Invest 1988; 81:1072-7.

354. Anklesaria P, Kase K, Glowacki J, et al. Engraftment ofa clonal bone marrow stromal cell line in vivo stimu-lates hematopoietic recovery from total body irradia-tion. Proc Natl Acad Sci USA 1987; 84:7681-5.

355. El-Badri NS, Wang BY, Cherry, Good RA. Osteoblastspromote engraftment of allogeneic hematopoieticstem cells. Exp Hematol 1998; 26:110-6.

356. Nolta JA, Hanley MB, Kohn DB. Sustained humanhematopoiesis in immunodeficient mice by co-trans-plantation of marrow stroma expressing human inter-leukin-3: analysis of gene transduction of long-livedprogenitors. Blood 1994; 83:3041-51.

357. Anklesaria P, Fitzgerald TJ, Kase K, Ohara A, Green-berger JS. Improved hematopoiesis in anemic Sl/Sldmice by splenectomy and therapeutic transplantationof a hematopoietic microenvironment. Blood 1989;74:1144-51.

358. Simmons PJ, Przepiorka D, Thomas ED, Torok-StorbB. Host origin of marrow stromal cells following allo-geneic bone marrow transplantation. Nature 1987;328:429-32.

359. Keating A, Singer JW, Killen PD, et al. Donor origin ofthe in vitro haemopoietic microenvironment aftermarrow transplantation in man. Nature 1982;298:280-3.

360. Henschler R, Junghahn I, Fichtner I, Becker M, GoanSR. Donor fibroblasts from human blood engraft inimmunodeficient mice [abstract]. Blood 1998; 92(suppl 1): 2417.

361. Lazarus HM, Haynesworth SE, Gerson SL, RosenthalNS, Caplan A. Ex vivo expansion and subsequent infu-sion of human bone marrow-derived stromal progen-itor cells (mesenchymal progenitor cells): implicationsfor therapeutic use. Bone Marrow Transplant 1995;16:557-64.

Cell therapy


Background and Objectives. Vaccination is an effectivemedical procedure of preventive medicine based onthe induction of a long-lasting immunologic memorycharacterized by mechanisms endowed with highdestructive potential and specificity. In the last fewyears, identification of tumor-associated antigens (TAA)has prompted the development of different strategiesfor antitumor vaccination, aimed at inducing specificrecognition of TAA in order to elicit a persistentimmune memory that may eliminate residual tumorcells and protect recipients from relapses. In thisreview characterization of TAA, different potentialmeans of vaccination in experimental models and pre-liminary data from clinical trials in humans have beenexamined by the Working Group on HematopoieticCells.

Evidence and Information Sources. The methodemployed for preparing this review was that of informalconsensus development. Members of the WorkingGroup met four times and discussed the single points,previously assigned by the chairman, in order toachieve an agreement on different opinions andapprove the final manuscript. Some of the authors ofthe present review have been working in the field ofantitumor immunotherapy and have contributed origi-nal papers to peer-reviewed journals. In addition, thematerial examined in the present review includes arti-cles and abstracts published in journals covered by theScience Citation Index and Medline.

State of the art. The cellular basis of antitumorimmune memory consists in the generation and extend-ed persistence of expanded populations of T- and B-lymphocytes that specifically recognize and reactagainst TAA. The efficacy of the memory can be modu-lated by compounds, called "adjuvants", such as cer-tain bacterial products and mineral oils, cytokines,chemokines, by monoclonal antibodies triggering co-stimulatory receptors. Strategies that have been shownin preclinical models to be efficient in protecting fromtumor engraftment, or in preventing a tumor rechal-lenge, include vaccination by means of soluble proteinsor peptides, recombinant viruses or bacteria as TAAgenes vectors, DNA injection, tumor cells genetically

modified to express co-stimulatory molecules and/orcytokines. The use of professional antigen-presentingcells, namely dendritic cells, either pulsed with TAA ortransduced with tumor-specific genes, provides a use-ful alternative for inducing antitumor cytotoxic activity.Some of these approaches have been tested in phaseI/II clinical trials in hematologic malignancies, such aslymphoproliferative diseases or chronic myeloidleukemia, and in solid tumors, such as melanoma,colon cancer, prostate cancer and renal cell carcinoma.Different types of vaccines, use of adjuvants, timing ofvaccination as well as selection of patients eligible forthis procedure are discussed in this review.

Perspectives. Experimental models demonstrate thepossibility of curing cancer through the active inductionof a specific immune response to TAA. However, whilepre-clinical research has identified several possible tar-gets and strategies for tumor vaccination the clinicalscenario is far more complex for a number of possiblereasons. Since experimental data suggest that vaccina-tion is more likely to be effective on small tumor bur-den, such as a minimal residual disease after conven-tional treatments, or tumors at an early stage of dis-ease, better selection of patients will allow more reli-able clinical results to be obtained. Moreover, a poorcorrelation is frequently observed between the ability ofTAA to induce a T-cell response in vitro and clinicalresponses. Controversial findings may also be due tothe techniques used for monitoring the immune status.Therefore, the development of reliable assays for effi-cient monitoring of the state of immunization of cancerpatients against TAA is an important goal that willmarkedly improve the progress of antitumor vaccines.Finally, given the promising results, identification ofnew or mutated genes involved in neoplastic eventsmight provide the opportunity to vaccinate susceptiblesubjects against their foreseeable cancer in the nextfuture.©2000, Ferrata Storti Foundation

Key words: antitumor vaccination.

review

Antitumor vaccination: where we stand


MONICA BOCCHIA,* VINCENZO BRONTE,° MARIO P. COLOMBO,#

ARMANDO DE VINCENTIIS,@ MASSIMO DI NICOLA,^

GUIDO FORNI,§ LUIGI LANATA,** ROBERTO M. LEMOLI,°°MASSIMO MASSAIA,## DAMIANO RONDELLI,°° PAOLA ZANON,**SANTE TURA°°*Department of Hematology, University of Siena; °Departmentof Oncology, University of Padova; #Division of Oncology, Isti-tuto Nazionale dei Tumori, Milan; @Dompè-Biotec, Milan; ̂ BoneMarrow Transplant Unit, Istituto Nazionale dei Tumori, Milan;§Department of Clinical and Biological Sciences, University ofTurin; **Amgen Italia, Milan; °°Institute of Hematology andMedical Oncology “Seràgnoli”, University of Bologna; ##Divisionof Hematology, University of Turin, Italy

Correspondence: Prof. Sante Tura, Istituto di Ematologia e Oncolo-gia Medica “Seràgnoli”, Università di Bologna, Policlinico S.Orsola-Malpighi, via Massarenti 9, 40138 Bologna, Italy. Phone: interna-tional +39-051-6363680 – Fax: international-39-051-6364037 –E-mail: [email protected]

157


Antitumor vaccines: the meaning As illustrated in a previous paper of this series, many

strategies are being used to try to cure cancer, each onebased on different theoretical and experimentalgrounds.1 Active immunotherapy strategies elicit specificor non-specific antitumor reactions by stimulating thepatient’s immune system. Alternatively, lymphocytescollected from patients are stimulated in vitro and re-injected into the patient (adoptive immunotherapy).Lastly, passive immunotherapy consists in the adminis-tration of antitumor antibodies to the patient. Howev-er, dealing with such a dramatic issue as is cancer, emo-tional empiricism spurs the adoption of distinct strate-gies or their mixing in an apprehensive pursuit of effi-cacy. Indeed, emotional empiricism has been and still isa deadly sin of tumor immunology. In the long run, onlyrational considerations lead to clinical progress.

The issueVaccination is an effective medical procedure charac-

terized by being a) predominantly a maneuver of pre-ventive medicine; b) based on the induction of a long-lasting immunologic memory that is c) characterized bymechanisms endowed with high destructive potentialand specificity.2 It rests on an artificial encounter of asham pathogen with the immune system. The shampathogen elicits a strong host reaction and leaves a per-sistent memory of this first artificial fight. If the realpathogen enters the immunized organism it becomes thetarget of a much stronger and precise reaction than thatput up by a non-immunized organism. If the pathogenhad a small possibility of escaping the reaction of a naiveimmune system, very seldom would it evade a memoryreaction.3 The cellular basis of immune memory consistsin the extended persistence of expanded populations ofT- and B- lymphocytes that specifically recognize andreact against the pathogen. Memory lymphocytes are alsoexperienced veterans, able to detect a pathogen prompt-ly and fight effectively against it.

There is a large universe of sham pathogens that areused for vaccination, named antigens or immunogens.A killed or inactivated pathogen, or a non-pathogenicorganism sharing critical molecules with the real thingcan be a good immunogen. Memory can also be inducedby a protein from the pathogen. In addition, a virus engi-neered with the DNA coding for a protein of thepathogen or even the mere naked DNA induces an effec-tive memory.4

The efficacy of the memory is modulated by numer-ous compounds, called adjuvants and danger signals.5,6

These provide additional activation signals, recruit reac-tive leukocytes at the immunization site and delay anti-gen catabolism. Bacterial products and mineral oils aretypical conventional adjuvants. Cytokines and chemo-kines also act as adjuvants. As will be discussed in detail,their use allows the induction of selective mechanismsof immune memory.7

Antitumor vaccination has a defined goal: to provokespecific recognition of tumor-associated antigens (TAA)

in order to elicit a persistent immune memory. Manyexperimental data have shown that following immu-nization, the growth of tumor cells expressing the sameTAA as the vaccine can be impaired. In patients, theimmune memory elicited by vaccines is sometimes fastand strong enough to hamper the growth of theirtumor.8

A brief historyInterest in antitumor vaccination arose around 1900

when a series of microbial vaccines proved to be effec-tive. The idea was straightforward: to apply the sameintervention to tumor. «If .........it is possible to protectsmall laboratory animals in an easy and safe way againstinfectious and highly aggressive neoplastic specimens,then it will be possible to do the same for humanpatients». These words of 1897 by Paul Ehrlich9 igniteda series of studies with transplantable mouse tumors.However the underlying issue turned out to be morecomplex than had first been presumed. More than onecentury was required to elucidate its molecular andgenetic features.

The first outcome was not a progress in anti-tumorvaccination, but instead the definition of a few rules ofallograft rejection. Transplanted tumors were rejected byimmunized host not because they expressed a particu-lar TAA but because they were from histoincompatiblemice and displayed normal allogeneic histocompatibil-ity antigens.10 Later studies with syngeneic mousestrains showed the feasibility of immunizing a mouseagainst a subsequent tumor challenge.11 However, thesuspicion that residual unnoticed histocompatibility dif-ferences were involved in these vaccination-rejectionstudies was not ruled out until experiments by GeorgeKlein in 1960.12 Carcinoma was induced by methyl-cholanthrene in syngeneic mice. The carcinoma wasthen surgically removed, and its cells were cultured invitro and used to repeatedly immunize the mouse inwhich the tumor had arisen. Finally the immunizedmouse and a few control syngeneic mice were chal-lenged with the carcinoma cells of the original tumormaintained in vitro. While these cells gave rise to a car-cinoma in control mice, they were rejected by the immu-nized mouse in which the carcinoma had arisen origi-nally. This evidence of the possibility of immunizingagainst a lethal dose of own tumor was of seminalimportance and had notable consequences. A very largeseries of subsequent studies established a few basicfoundations of tumor vaccination.13

Looking back at tumor immunology over the last 20-30 years, the importance of models in influencingimmunologic beliefs is strikingly evident. Inappropriateuse of an experimental model may produce wrongbeliefs, from which it is then very hard to escape. Virus-and chemically-induced tumors form highly immuno-genic models. Since they are easy to handle, these wereused to establish the rules of tumor vaccination. How-ever, it was disputable whether the information fromthese models was relevant to the situation of patients

Antitumor vaccination

158


with cancer. Using a series of murine spontaneoustumors, Hewitt concluded that it was not possible toimmunize against these tumors.14 This observation hadcrucial importance in shaping subsequent studies. Thepossibility that the experimental work done with highimmunogenic transplantable tumors had little relevanceto human tumors was a dark shadow that hindered theprogress of tumor immunology. Later, more careful useof these spontaneous tumors and more refined immu-nization techniques showed that Hewitt's conclusionswere wrong.15

Boon led the genetic and molecular identification ofa large series of TAA. Initially his studies were performedusing conventional transplantable mouse tumors.16

Then, tumor antigens were detected on the same spon-tanous tumors that had previously been classified asnon-immunogenic by Hewitt.15 Now many antigensassociated with human tumors have been identified.17,18

The targetsBoon and others18,19 provided an unambiguous defin-

ition of TAA, an important finding that definitively laidto rest the doubts on the foundations of tumorimmunology in man.20 In many cases TAA are peptidespresented by class I and class II glycoproteins of themajor histocompatibility complex (MHC). Things thatmay give rise to these tumor-associated peptides areenhanced or diminished expression of some normal pro-teins and the new expression of altered or normallyrepressed molecules. Less frequently these antigens aretumor-specific as they derive from mutated proteins.Lastly, various TAA are shared by tumors with distincthistology and origin (Table 1). Telomerase catalytic sub-unit looks like another widely expressed TAA recognizedby T-lymphocytes. It is markedly activated in mosthuman tumors while it is silent in normal tissues.21

Why?The central tenet of antitumor vaccination is that the

immune system is able to destroy tumor cells and toretain a long-lasting memory provided that TAA are firstefficiently recognized. While the studies aimed at thedefinition of TAA progressed quickly, investigations oflymphocyte receptors and their idiotype network, co-stimulatory molecules, and cytokines were leading to amore exact description of the requirements for theinduction of an immune response.22 Finally, technicalrefinement of genetic engineering is making the devel-opment of new cancer vaccines easier .23 The conver-gence of these issues is once again placing antitumorvaccination at the cutting edge of biological research.A survey by Science24 indicates that antitumor vaccina-tion is expected soon to become an established thera-peutic option.

When?Whereas individuals are immunized with microbial

vaccines prior to encountering the pathogen, cancerpatients have to be immunized when a tumor has beenalready detected. It is not yet possible to predict which

combination of gene mutations will give rise to cancer.Therefore, the common clinical setting is elicitation ofan immune response in a tumor-bearing patient, ratherthan prior to tumor development. The very concept ofvaccine is somewhat distorted since it has moved frombeing preventive to being therapeutic.23

The kind of patients who should be considered eligi-ble for tumor vaccination is not a minor issue. In manytrials patients with advanced diseases are enrolled bothfor compassionate reasons and because of the con-straints imposed by ethical considerations. But, do exper-imental data suggest that vaccination could be effectivein advanced stages of neoplastic progression? The exper-imental data provide an unambiguous picture of thepotentials and limits of vaccination. This picture is not,however, generally taken into account. Perhaps uncon-scious reasons lead to experimental data being assessedwith optimistic superficiality.25 Many experimental stud-ies have shown that an antitumor response can be elicit-ed by new vaccines. This results in strong resistance toa subsequent tumor challenge and inhibition of minimalresidual disease remaining after convention therapy. Thepitfall hidden in the evaluation of these vaccine-re-chal-lenge experiments is that successful immunization ofhealthy mice against a subsequent re-challenge withtumor cells does not demonstrate a true therapeuticeffect.26 Examination of more realistic studies of the abil-ity of vaccines to cure existing tumors shows that onlya minority of tumor-bearing mice could be cured. Fur-thermore, the limited therapeutic efficacy of vaccineswas lost when they were not given in the first few daysafter the implantation of tumor cells.26

A similar picture is emerging from phase I studies onvaccination of cancer patients. The vaccination is safe,but the results suggest that only a minority of patients(about 10%) display an objective response. The immuno-logic performance status of these patients is obviouslysuboptimal for this type of therapy. Even so, one wouldhave expected a greater number of responders to sup-port the promise of new sophisticated vaccines.23

Therapeutic vaccination has not had much success inthe management of infectious diseases. Its use againstthe progression of an established tumor is very chal-lenging, since it must secure an effective immuneresponse capable of getting the better of a well-estab-lished, proliferating tumor. Of the several objectives that

M. Bocchia et al.

Table 1. Cross-expression of some tumor-associated anti-gens among histologically different human tumors from dis-tinct organs.

bladder BAGE GAGEbreast BAGE MAGE CEA p53 ras MUC-1colon CEA p53 raslung BAGE CEA p53melanoma BAGE GAGE MAGE raspancreas CEA p53 ras MUC-1sarcomas GAGE

159


have been made approachable by antitumor vaccines,the cure of advanced tumors is both the most difficultand at the same time the most common in clinical tri-als. The strong emotions kindled by cancer suffering pro-vide the main justification for these attempts.27 Perhaps,major improvements in antitumor vaccination will makethis goal approachable. However, the data reviewedshow that this is far from the present reality.

Nevertheless it should be considered that most tumorlethality depends on a few neoplastic cells remainingafter surgical excision of a tumor mass or after havingescaped direct killing by chemo- and radiotherapy. Manyexperimental findings28-30 suggest that a stage of min-imal residual disease is one in which it is possible toforesee a significant cure by immunization. After suc-cessful conventional management the tumor burdenmay be low, and the tumor may reappear after a longdormancy. This is a realistic setting in which vaccinationcould lead to the induction of anti-tumor immunitycapable of extending the survival of patients. As thereare grounds for believing that antitumor vaccines couldbe used as an effective anticancer tool, the purpose ofthis review is to describe the types of vaccines that arebeing experimented, emerging clinical results and thenew perspectives opened by this scientific endeavor.

Common tools

Cytokines and cellular signalsThe immunologic attempt of the immune system to

prevent the development of a neoplastic disease maybe ineffective due to either a lack of immunogenicity oftumor cells, or to a weak reaction unable to contrast theneoplastic proliferation. In both cases, it is likely thatmost of the physiologic mechanisms of priming of theimmunologic effector cells may be impaired or absent.In fact, initiation of immune responses requires thatprofessional antigen-presenting cells (APC) deliver a firstsignal to T lymphocytes through the binding of the T-cellreceptor by the peptide enclosed in the HLA molecule,that is responsible for the specificity of the immuneresponse, and a second or co-stimulatory signal that isnot antigen-specific but it is required for T-cell activa-tion31,32 mainly through CD80 (B7-1) and CD86 (B7-2)binding to CD28 receptor, or the CD40:CD40L pathway.Moreover, the capacity of dendritic cells (DC) to activatenatural killer (NK) cells by ligation of the CD40 moleculewith its counter-receptor has recently been demon-strated.33,34 Immunocompetent cells may also determinethe type of immune response by the expression ofchemokines and by the release of pro-inflammatory, oranti-inflammatory cytokines which drive T-cells to dif-ferent activities or even to suppression.35

Therefore, given the complex network of regulatorysignals by professional APC and naive and memory lym-phocytes occurring in antigen-specific immune respons-es, it is not surprising that tumor cells may fail to induceefficient humoral and cellular immune reactions evenwhen a well known TAA is present. In this review, sev-

eral strategies to overcome the immune escape mech-anisms of tumor cells will be considered, such as thedirect use of TAA to elicit specific reactions, the use ofdendritic cells to present TAA in order to enhance theimmune response, and the use of tumor cells genetical-ly modified to function as professional APC or to releasesoluble factors. Animal models have been widely usedfor many years to demonstrate the effect of differentcytokines, added to or secreted by tumor cell-based vac-cines, in increasing the in vitro and in vivo cytotoxicityagainst tumor challenge. The role of the main cytokinesinvolved in activation of humoral and cellular immuneresponses is represented in Figure 1. On the basis ofcytokine functions it has been previously shown inexperimental models that: the number of APC in thesite of tumor infiltration can be increased by cytokinessuch as granulocyte-macrophage colony-stimulatingfactor (GM-CSF) and interleukin (IL)-4, which shouldallow also the differentiation of DC precursors; B- andT-cell responses are potently increased by IL-2, IL-12, orGM-CSF;36 and in particular T-cell cytotoxicity isenhanced by GM-CSF, IL-2, IL-12, interferon (IFN-γ),tumor necrosis factor (TNF-α), while NK activity isenhanced by IL-12 or FLT 3-L.36,37 However, differentmodels and different TAA resulted in controversial find-ings. Furthermore, since these cytokines are likely to bemore effective when released within the tumor area,the transduction of cytokine genes into tumor cells andtheir use as cellular vaccines after irradiation has beentested in animal models and in humans.38-40

Initial clinical experiences in patients with advancedmelanoma or renal cell carcinoma suggest that tumorcell-based vaccines, either engineered to produce GM-CSF or IL-12 or with exogenous GM-CSF, may facilitatemarked infiltration of DC and CD4+ and CD8+ T-lym-phocytes into tumor lesions, potentially improving theantitumor effect. These data provided evidence of thefeasibility of the approach but were unlikely to be ableto address the point of efficacy, due to the large tumorburden of these patients. Future immunotherapyattempts should, in fact, focus on the possibility of erad-icating minimal residual disease. More recent datademonstrated the role of GM-CSF as a useful adjuvantin peptide-based vaccines in ovarian and breast cancerand in follicular lymphoma, as will be described later inthis review.

Figure 1 also shows that cell-to-cell contact viaCD40:CD40L plays a pivotal role in activating specific T-cell, B-cell and NK-cell responses. On the other hand, T-cell tolerance can be obtained by blockade of CD40Lreceptor in non-human primates undergoing solid organallogeneic transplantation, and in mice receiving eitherallogeneic bone marrow or solid organ transplanta-tion.41-43

Recent experiments demonstrated that stimulation, viaan activating anti-CD40 antibody, resulted in the acti-vation of host APC and could convert lymphocytes ofmice treated with a tolerogenic peptide vaccine intocytotoxic T- cells. Moreover, this stimulation induced the


160


M. Bocchia et al.

regression of established tumors that had not beenaffected by previous vaccination alone,44,45 thus showingthat triggering the CD40 molecule may both overcome T-cell tolerance in a tumor-bearing animal and greatlypotentiate a peptide-based vaccine. Moreover, genetransfer by an adenovirus vector of CD40L in human B-cell chronic lymphocytic leukemia (B-CLL) allowed theactivation of bystander non-infected B-CLL cells thatupregulated co-stimulatory molecules such as CD80 andCD86 and stimulated autologous cytotoxic T-cells.46

Another important issue concerns the way T-cells areturned off by CTLA-4 receptor following activation viaCD28. In fact, both CD28 and CTLA-4 bind with highaffinity to CD80 and CD86 and CTLA-4 physiologicallyblocks T-cell activation. In the case of antitumor T-cellresponse it has been demonstrated that blockade of neg-ative regulatory signals by an anti-CTLA-4 monoclonalantibody may retard tumor growth in experimental sys-tems.47,48 More recent studies in mice suggested that thismolecule was extremely efficient in causing tumorregression when used in combination with subtherapeu-tic doses of melphalan, or with a GM-CSF-expressingtumor.49,50

Altogether, these studies strongly support the role ofcytokines or immunomodulatory molecules in anti-cancer vaccine strategies. However, they do not clarifywhether a strict T-helper (Th1) response is required toachieve tumor killing, or whether a humoral responseinduced by anti-inflammatory cytokines should also bepursued. Finally, future directions of anticancer vaccineresearch are likely to deal with monoclonal antibodiesenhancing or blocking specific receptors.

Dendritic cells as initiators of immuneresponse

Dendritic cells (DC) represent a heterogeneous popu-lation of leukocytes defined by morphologic, phenotyp-ic and functional criteria which distinguish them frommonocytes and macrophages.32 From among the pro-fessional APC, DC are the most potent stimulators of T-cell responses and play a crucial role in the initiation ofprimary immune responses.32

The DC system comprises at least three distinct sub-sets, including two within the myeloid or non-lymphoidlineage, and a third represented by lymphoid DC.32,51

There is also a continuum of differentiation within eachof these subsets, from precursors circulating throughblood and lymphatics, to immature DC resident inperipheral tissues, to mature or maturing forms in thethymus and secondary lymphoid organs. Recent studieshave focused on the different roles of lymphoid andmyeloid DC: more resident lymphoid DC induce toler-ance to self, whereas migratory myeloid DC, includingLangherans cells, are activated by foreign antigens in theperiphery and move to lymphoid organs to initiate animmune response.32

DC have always been described as having two distinctfunctional stages: 1) immature, with high antigenuptake and processing ability, and poor T-cell stimula-tory function; 2) mature, with high stimulatory functionand poor antigen uptake and processing ability. Bacte-rial products such as lipopolysaccharides, and inflam-matory cytokines such as IL-1, TNF-α, type I interferons(IFN α or β) and prostaglandin E2 (PGE2) stimulate DCmaturation, whereas IL-10 inhibits it.52 Interestingly,

Figure 1. Principal cytokinesinvolved in the antitumor immuneresponse.

161


human and murine DC upregulate the synthesis of HLAclass I and II molecules, and B7-1, B7-2 and CD40 mol-ecules, after ingestion of bacteria or bacterial products,such as lipopolysaccharides, can prime naive T-cells.

An emerging concept is that APC activate T-helper (Th)cells not only with antigen and co-stimulatory signals,but also with a polarizing signal (signal 3). This signalcan be mediated by many APC-derived factors, but IL-12and PGE2 seem to be of major importance. As for Th cells,APC can be functionally polarized. In vitro experimentswith monocyte-derived DC showed that the presence ofIFN-γ during activation of immature DC induces matureDC with the ability to produce large quantities of IL-12and, consequently, a Th1-driving capacity (APC1 or DC1).In contrast, PGE2 primes for a low IL-12 production abil-ity and Th2-driving capacity (APC2 or DC2).32,53 DC-stim-ulated CD4+ cells upregulate CD40L/ CD154 that recip-rocally activates DC via CD40. This renders DC morepotent stimulators of CD8+ cytotoxic T-cell (CTL).54 Thisnovel concept is in contrast to simultaneous stimulationof CD4+ and CD8+ T-cells by DC, whereby the CD4+ T-cells secrete helper lymphokines in support of CD8+ CTLdevelopment.55 Together with CD40 L and CD40, two dif-ferent groups have discovered another paired member ofthe TNF-TNF receptor family. This factor, termed TNF-related activation induced cytokine (TRANCE) or receptoractivator of NF-kB ligand (RANK-L), is expressed by Tcells.56 Its corresponding receptor, receptor activator ofNF-kB (RANK7) or TRANCE R, is expressed by mature DCsbut not on freshly isolated B cells, macrophages, or Tcells.57 Ligation to this receptor causes either activationof T-lymphocytes or enhancement of DC survival. In addi-tion, IL-12 is a critical mediator of DC-supported differ-entiation of naive, but not memory, B-cells,58 indicatingthat direct interactions occur between DC and B cells,apart from those that occur via cognate CD4+ T-cell help.

Lastly, it should be mentioned that the primary andsecondary B cell follicles contain another population ofDC, the follicular DC (FDC). The origin of these cells isnot clear, and most investigators believe that they arenot leukocytes. FDC trap and retain intact native antigenas immune complexes for long periods of time, presentit to B cells and are likely to be involved in the affinitymaturation of antibodies, the generation of immunememory and the maintenance of humoral immuneresponses.59

In conclusion, there is a general agreement in consid-ering DC as very important players in the game ofimmune responses against foreign antigens either ofinfectious agents or of neoplastic cells. Many developingimmunotherapeutic strategies against danger antigensare aimed at exploiting the powerful antigen-present-ing properties of DC by an in vivo or ex vivo engineeringof the DC system. In fact, subcutaneous or intramuscu-lar injection of antigens relies on the local recruitmentand activation of DC to capture and present antigens tothe immune system. Although the techniques for target-ing tumor antigens to DC in situ might eventually obvi-ate the need for ex vivo manipulation of DC, novel meth-

ods for ex vivo generation and activation of largeamounts of human DC have been developed.

Antitumor vaccination: types and formulations

Killing for priming and killing to destroyThe way cell vaccines die when injected in vivo influ-

ences DC loading. The initial activity of cytokines trans-duced into the cell vaccine is to select the leukocyte typerecruited and stimulated at the site of injection. Tumorcells are killed by effectors of the innate response; NKand polymorphonuclear cells also produce secondarycytokines that set up a local inflammation recruiting DC,whereas T-cell response is activated later. Of note, geneengineering allows the manipulation of the first phase ofthe process through the choice of cytokine and/or co-factor and by deciding the level of cytokine to be pro-duced. Once the infiltrate leukocytes are activated, theirresponse to the triggering cytokine is physiologic andindependent. They produce other cytokines thus ampli-fying the magnitude and the complexity of response. Theinitial stage is finalized to T- and B-cell activation, killingof cell vaccine is to provide the antigen(s) and theinflammatory response should provide the right envi-ronment for such activation.60

The final stage is aimed at the destruction of existingtumor. Specific immune response is often insufficientto fight solid tumor nodules. Among the possible caus-es, immunosuppression, low effector-target ratio, MHCdownregulation on tumor cells are the most common.Animal studies have shown that these problems can beovercome by general inflammation associated with neu-trophil influx. This combination may destroy the tumor-associated blood vessels61 in a way that may resembletissue damage in vasculitis. In this setting a specificimmune response is not directly responsible for tumorelimination but should be strong enough to at leastbegin and direct the inflammatory response to the tumorsite. In this perspective, tumor vessels are the main tar-get of a non-specific immune response, their function-al impairment increasing the effect of either T- or B-cell-mediated specific immune responses.

An add to DC-common link?Although DC have been indicated as central in alerting

and activating the immune system, it is now clear thatcertain peptides are not and can not be presented bymature DC. This observation, made by Van den Eynde andcolleagues,62 concerns autoantigens and T lymphocytesthat recognize them without being deleted in the thymusand normally without provoking autoimmunity. In fact,APC differ from other cell types by the proteosome thatdigests protein into immunogenic peptides to fit the MHCgroove. APC immunopreoteosomes have three catalyticsubunits substituted by those induced by IFN-γ, thus gen-erating slightly different peptides from those generated bynon-APC cells.

Several self-antigens identified as tumor-associatedbecause of CTL recognition may not be processed by


162


immunoproteosome. The implication is that such CTL werenot generated through DC presentation or at least notthrough DC processing unless this happened during tran-sition from immature to mature DC.63 Perhaps free pep-tides can be captured on the surface of DC for presenta-tion, or perhaps other as yet unknown mechanisms areinvolved.

Genetically modified tumor cell vaccinesOld and recent discoveries confirm the possibility of a

cancer vaccine made of tumor cells. An empirical approach, such as the use of allogeneic

whole-cell vaccine composed of 3 allogeneic melanomacell lines established in vitro, allowed a 3-fold increase ofthe five-year survival of patients with stage IV melanomaas reported by Morton et al.64 The most active componentof Morton’s vaccine has been identified by Livingston andcolleagues65 to be a ganglioside (GM2). Patients whodevelop antibody response to ganglioside showed a sig-nificant survival advantage. Whether a CTL response wasalso activated has not been investigated but does proba-bly exist. However, this finding prompted a phase III studyin patients with stage III disease.

In the autologous setting, irradiated melanoma cellswere modified with dinitrophenyl and used to treatpatients with metastatic disease. Clinical evidence ofinflammatory response to superficial metastases wasreported. The same treatment administered to phase IIIpatients who remain tumor-free after resection of lymphnode metastases has resulted in 50 and 60% 4-yearrelapse-free and overall survival, respectively.66 Immuno-staining, TCR repertoire analysis and functional data ofnode-metastases post-vaccination have shown that treat-ment with autologous dinitrophenyl-modified melanomacells can expand certain T-cell clones at the tumor site.67

The above results bring up two issues: autologous ver-sus allogeneic tumor cells and chemical or genetic (seebelow) modifications of tumor cells to be used as cancervaccines.

Before discussing these issues we should, however,address a more basic question, that is how to use cancercells as vaccine. Well before identification of tumor anti-gens, their existence was inferred in melanoma on thebasis of expansion and characterization of cytotoxic T-lymphocytes recognizing autologous tumor,68 whereasantibodies against a variety of tumor types were isolatedfrom patient sera. Antigens recognized by CTL can betumor-specific and even restricted to the autologoustumor or cross-react among different neoplasms of thesame or different tissue origin depending on the mecha-nism generating such an antigen: point mutation, incor-rect splicing, over-expression, translocation and other (seeTable 2). The number of tumor antigens is always increas-ing thus suggesting that our knowledge of the antigenicrepertoire of tumor cells is partial. In addition, which anti-gen among those already characterized should be used ina vaccine?

These problems can be solved altogether by using tumorcells that would represent the entire antigenic repertoire

with a single drawback, that is immunoselection of cer-tain antigens. Tumor cells from different patients mayundergo different processes of immunoselection andtherefore a pool of these tumors would be ideal for prepar-ing an allogeneic vaccine. The finding that antigensbelonging to allogeneic cells are processed by host anti-gen-presenting cells,69 so that they can be recognized byhost T-lymphocytes (a phenomenon called cross-priming),removes any conceptual obstacles to the use of allogene-ic cells. Allovaccines have several advantages over autol-ogous vaccines: cell lines that are extensively character-ized in vitro can be used to treat all the patients includedin a clinical study. Genetic modifications of tumor cellshave been widely studied as a way to increase immuno-genicity. These modifications can be easily made to a cellline, thus avoiding the need to isolate cells from everytumor. The rate of success in culturing tumor cells from aprimary tumor varies according to the type of tumor andexperience of the operator. Transduced cell lines can beselected for production of a certain amount of transgenethus assuring that the same vaccine will be given to allpatients. Which modification is most efficient in inducingtumor immunity has not been unequivocally determinedsince variability among different tumors has beendescribed. Chemical modification, also called hapteniza-tion, aims at adding helper determinant to tumor cells,70

although the exact mechanism and the downstream path-ways activated in this way are not clearly understood.Genetic modification is now preferred since the effect ofseveral cytokines and co-stimulatory molecules are knownat molecular level. Their genes can be transduced intotumor cells that acquire new immunoregulatory functions.In this way a desired immune response can be fine-tunedthrough gene dosage, recruitment of certain cell types,deflection of Th1 or Th2 type of response and several oth-ers mechanisms.60

To summarize the most recent and promisingapproaches we may consider two strategies aimed atfavoring vaccine interaction with DC or directly with Tlymphocytes. The two approaches can be viewed forpriming or boosting (Figure 2). A combination ofcytokine and co-stimulatory factors is likely to be syn-ergistic as generally occurs when soluble and cell-con-tact signals are given together. Tumor cells transducedwith both GM-CSF and CD40L have been shown to beheavily infiltrated by DC. GM-CSF induces proliferationand maturation of hematopoietic cells, and has beenshown to stimulate DC accessory properties andenhance the immune response initiated by these cells.71

The CD40/CD40L interaction plays a critical role incell-mediated immunity, and in proliferation and acti-vation of APC, as shown for B cells, monocytes and, morerecently, DC.72 Ligation of CD40 on monocytes and DCresults in the secretion of several cytokines, including IL-1, IL-6, IL-12, and TNF-α. The murine tumor transducedwith both GM-CSF and CD40L genes showed that tumorinfiltrating DC can take up cellular antigen from thetumor and can present it to T lymphocytes in vitro.73 Inthis setting, DC bridges cell vaccine-T-lymphocytes

M. Bocchia et al.

163


interaction and can be envisaged as the way to primepatients against weaker antigens otherwise ignored bythe immune system. A complementary approach wouldconsider the possibility of boosting the primed or theexisting immune response. In this case DC that re-encounter activated T-cells can be lysed and may not beappropriate for boosting. Boosting can be done usingtumor cells transduced with IL-2 and B7-1 such as toprovide both cell expansion and co-stimulation fromthe tumor cell vaccine directly to T lymphocytes (Figure2). The way cellular antigen can be captured by DC forpriming of T-lymphocytes depends on how tumor cellsdie after injection. Irradiation of the cell vaccine inducesapoptotic cell death and apoptotic bodies can be cap-tured by DC.74 Others suggest that peptides of cellularprotein complexed to heat shock protein (HSP), the nat-ural chaperon of peptide from proteosome to mem-brane-associated TAP, leave the cells upon necrosis tobe taken up by DC.75,76 DC loading must be followed byDC maturation and migration to the lymph node toensure correct antigen presentation.32

The way cytokines and co-signals modulate vaccine-host interactions may determine the extent and the effi-cacy of treatment. Systemic activation of the immuneresponse (T-cell cytotoxicity, antibodies) is easy to mea-sure but can not be used as a read-out system to predictwhether the tumor will be rejected. Tumor nodules mightbe reached by circulating lymphocytes which, however,may be neutralized because of either immunosuppres-sion or peripheral tolerance. The former is likely depen-dent on tumor size and is less expected when a smalltumor, minimal residual disease or prevention of recur-rence is the target to be treated. The latter could be sur-mounted by appropriate co-signals, for example CD40L.45

Soluble proteins as immunogensSoluble proteins derived from autologous cancer cells

are not as immunogenic as proteins derived from infec-tious agents. The degree of foreignness, which dependson the reciprocal distribution between epitopes subjectand not subject to self-tolerance, is low in TAA. More-over, TAA do not have the particulate nature of infec-tious agents that is a key factor in increasing theirimmunogenicity. Finally, infectious agents are rich inmolecules that are able to raise immune responsivenesswithout being themselves immunogenic. A significanteffort has been made over the last few years to trans-late the immunogenic properties of infectious agentsinto cancer vaccines for clinical use. The most suitableTAA should be directly involved in the malignant behav-ior of tumor cells; contain multiple immunodominant Bcell and T cell epitopes, including both helper and cyto-toxic T-cell epitopes; contain a very high degree of for-eignness; be unprotected by self-tolerance mechanisms.Of course, HLA haplotype remains a major constraint indetermining whether TAA-derived immunodominantpeptides can elicit tumor-specific immune responses ina given individual (see also below).

Building up immunogenicity of soluble pro-teins

Even the most suitable TAA is not immunogenic unlessit is processed and presented by professional APC to thehost immune system. To this aim, TAA should interactwith APC directly. The route of administration and adju-vants are key factors in determining the final outcome


Table 2a. Tumor-associated antigens recognized in class IHLA restriction.

Antigen defined Type of antigen Type of tumor HLA-restrictionand distribution allele

gp 100 Melanocyte diff. Melanoma A2, A3, A24, Cw8Melanocortin receptor 1 Melanocyte diff. MelanomaMART-1/MelanA Melanocyte diff. Melanoma A2, B45Tyrosinase Melanocyte diff. Melanoma A1, A2, A24, B44TRP-1 (gp 75) Melanocyte diff. Melanoma A31TRP-2 Melanocyte diff. Melanoma A2, A31, Cw8p15 Widely expressed Melanoma A24SART-1 Widely expressed Lung carcinoma A2601PRAME Widely expressed Melanoma, renal A24NAG-V* Melanoma sheared Melanoma A2.1β catenin Unique tumor specific Melanoma A24CDK4-Kinase Unique tumor specific Melanoma A2MUM-1 Unique tumor specific Melanoma B44TRP2/INT2 Unique tumor specific Melanoma A6801, Cw8CASPASE-8 Unique tumor specific Head/neck cancer B35HLA-A*201 mutated Unique tumor specific Renal cancer A2.1KIAA0205 Unique tumor specific Bladder cancer B44*03BAGE Cancer/ testis Melanoma Cw 1601GAGE-1/2 Cancer/ testis Melanoma Cw 6MAGE-1 Cancer/ testis Melanoma A1, Cw16MAGE-3 Cancer/ testis Melanoma A1, A2, B44RAGE Cancer/ testis Renal cancer B7NY-ESO-1 Cancer/ testis Melanoma, ovarian, A2

esophageal cancerK-RAS-D13 mutated Shared tumor-specific Colon cancer A2.1p53 mutated Shared tumor-specific Colon and lung A2.1Bcr/abl Shared tumor-specific CML A2.1, A3,

A11, B8 MUC-1 Shared tumor-specific Breast, colon, A11

pancreatic cancerHPV16E7 Viral related Cervical cancer A2.1

Table 2b. Human tumor antigens recognized by HLA classII-restricted CD4+ T cells.

Antigen Tissue distribution Class II restriction

Tyrosinase Melanoma/melanocytes DRb1*0401

Tyrosinase Melanoma/melanocytes DRb1*1501

Triosephosphate isomerase Melanoma, unique DRb1*0101mutated form

MAGE-3 Melanoma and other tumors, testis DRb1*1301

MAGE-1, -2 or -6 Melanoma and other tumors, testis DRb1*1301

MAGE-3 Melanoma and other tumors, testis DRb1*1101

RAS-D12 mutated Colon and pancreatic cancer DR1

HER-2/neu Breast and ovarian cancer DR11

PML/RARα Acute promyelocytic leukemia DR2

Bcr/abl Chronic myeloid leukemia DR1, DR4, DR11

CDC27 Melanoma DR4

*N-acetylglucosaminyl transferaseV. CML: chronic myeloid leukemia.

164


of immunization. Intravenous injection of soluble anti-gens induces tolerance, whereas subcutaneous or intra-muscular injection of the same antigens results in immu-nity because of the interaction with epidermal Langer-hans cells or dermal dendritic cells. Accordingly, deliveryof antigens via mucosal surfaces may result in immuni-ty because antigens may interact with the numerous DClocated just beneath the epithelium of mucosal lymphoidorgans. The association between soluble antigens andAPC is made much more intense by adjuvants. Adjuvantsact via different principles and pathways, but the com-mon goals are to prolong the interaction with APC bypromoting a slow antigen release, and functionally acti-vate APC themselves by delivering danger signals.Cytokines have also emerged as potent immunoadju-vants since they can influence the immune responses atdifferent levels (see above).

Particulation of soluble proteinsPrecipitation with aluminium hydroxide or aluminium

phosphate has been used to particulate antigens in diph-theria, tetanus, hepatitis B, and other vaccines. Thesetypes of vaccines induce antibody formation, but very lit-tle delayed cutaneous hypersensitivity (DTH) or cell-medi-ated cytotoxicity. In a pilot study, five stage I-III patientswith multiple myeloma were immunized with autologousidiotype (Id) precipitated in aluminium phosphate sus-pension. Three patients developed idiotype-specific T-and B- cell responses, but these responses were transient

and their amplitude was low.77

The use of immunostaining complexes (ISCOMs) isanother strategy that has been used to particulate anti-gens. ISCOMs are cagelike structures made of choles-terol, saponin, phospholipid, and viral envelope proteinsto which other proteins can be associated. Saponin is aplant derivative that is critical to the efficacy of ISCOMs.QS21 is the most effective fraction of saponin and iscurrently under clinical investigation in several trials.78

ISCOMs may reach the endocytic pathway and induceDTH and CTL responses other than antibody production.Liposomes, virosomes, and proteasomes are alternativestrategies to ISCOMS. Experimental data have recentlyshown in the 38C13 mouse B-cell tumor that liposomalformulation of autologous Id converts this weak self-antigen into a potent tumor rejection antigen.79

Promoting slow release of soluble antigensA slow release of antigen is one of the major goals of

adjuvants. Antigen polymerization and emulsifying agentshave been put together to achieve this goal. Polymeriza-tion can be obtained by association with non-ionic blockpolymers or by association with carbohydrate polymers.Non-ionic block polymers have been used as componentsof water-in-oil or oil-in-water emulsions. SAF-1 (SyntexAdjuvant Formulation-1) is an oil-in-water adjuvant for-mulation containing non-ionic block polymers, squalene,and Tween 80. SAF-1 has been used by Kwak et al. in theirpioneering study on idiotype vaccination in follicular lym-

M. Bocchia et al.

Figure 2. Priming of DC and boosting of T-cell responses by transduced tumor cells as vaccine.

165


phoma patients.80 Chemical immunomodulators such asderivatives of muramyl dipeptide, the smallest subunit ofthe mycobacterial cell wall that retains immunoadjuvantactivity, can be added to oil-in-water adjuvants.81 Thisapproach was used in the study by Hsu et al. in which idio-type/KLH conjugates were delivered to follicular lymphomapatients after mixing with SAF-165 that containedmuramyl dipeptide as immunomodulator.82 Lipid compo-nents of bacteria have also been used as adjuvants inexperimental models. These are portions of the LPS endo-toxins of Gram-negative bacteria. These molecules are,however, too toxic and chemically modified derivativeshave been developed for clinical use. So far, monophos-phoryl lipid A is the least toxic derivative capable of pro-moting cell-mediated immunity.83

Polysaccharide polymers have also been used as a sus-tained-release vehicle for TAA. Poly-N-acetyl glu-cosamine is a highly purified, biocompatible polysac-charide matrix that has recently become available forthis purpose.84

Peptide vaccinesT lymphocytes recognize small peptides that repre-

sent the degradation products of a complex intracellu-lar process and are presented on the cell surface com-plexed to 1 of 2 types of histocompatibility leukocyteantigen molecules (HLA classes I or II). CTLs (CD8+) main-ly recognize peptides of 8 to 10 amino acids derivedfrom intracellular or endogenous proteins and com-plexed to HLA class I molecules.85-89 CD4+ T- lympho-cytes recognize exogenous proteins which are ingestedby APC, degraded to peptides of 12-24 amino acids andcomplexed to HLA class II molecules.90,91

Class I and Class II peptides that are presented on thecell surface, although randomly derived from the origi-nal protein, must contain specific amino acids in 1 or 2critical positions in order to be able to bind the appro-priate HLA molecules. Thus, peptide binding is HLA-restricted. The amino acid motifs responsible for thespecific peptide-binding to HLA class I and class II mol-ecules have been determined for the common HLA typesby analyzing acid-eluted naturally processed peptidesand by using cell lines defective in intracellular peptideloading and processing.92-94

Several tumor-specific and some leukemia-specificpeptides have so far been identified, and studies aimedat evaluating the potential clinical benefit of peptide vac-cines in cancer patients have begun. Among the reasonsthat make a peptide vaccine strategy interesting are sev-eral unique advantages that peptide immunization offersover other vaccine approaches: 1) peptide vaccines per-mit specific targeting of the immune response against 1or 2 unique antigens (thus limiting the potential autoim-mune cross-reactivity or immunosuppressive activityoften observed with more complex immunogens); 2)emerging technology has made it simple, rapid and inex-pensive to sequence and prepare larger quantities oftumor antigen peptides for both laboratory and clinicaluse; 3) use of synthetic peptides greatly reduces the pos-

sible risk of bacterial or viral contamination that mightderive from autologous or allogeneic tissue for immu-nization. On the down side, the main disadvantages ofpeptide immunization are: 1) lack of universal applica-bility as each peptide is restricted to a single HLA mole-cule; 2) poor immunogenicity of most native peptides; 3)risk of inducing antigenic tolerance. Successful attemptsto enhance HLA binding affinity have been based on syn-thetically generating peptides with amino acid deletionsor substitutions while maintaining antigen specificity.95,96

In initial studies, synthetic substituted peptides appearedto enhance immunogenicity and also to overcome thehost immune tolerance that exists to native peptides.97,98

Conversely it has been reported that changes in the finespecificity of modified peptide-reactive T-cells followingvaccination may occur with subsequent loss of tumor cellrecognition.99

A peptide vaccine approach involves several stepswhich are aimed firstly at identifying the appropriatepeptide, secondly at checking for its immunogenicityand relevance as a tumor-associated antigen (TAA) invitro, and thirdly at formulating a safe product to beused clinically.

Identification of the appropriate peptide1) From protein to peptide. This approach involves the

screening of potentially HLA-binding peptides within thesequence of a known tumor-specific protein by using HLAanchor motifs and epitope selection. Peptides derived bymut RAS,100 melanoma-associated MAGE protein,101

prostate specific antigen102 and chronic myelogenousleukemia (CML) specific P210 were identified by thisapproach.103,104 In CML, for example, a possible total of 76peptides, 8 to 11 amino acids in length, spanning the b2a2and b3a2 junctional regions of bcr-abl were screened forHLA class I- binding motifs and tested for the effectivebinding property to purified HLA molecules. Four of them,all derived from the b3a2 breakpoint, were found to beable to bind with either intermediate or high affinity topurified HLA A3, A11 and B8.103 A similar approachallowed identification of class II b3a2 breakpoint peptidescapable of binding HLA DR11,105 DR4106 and DR1107 Thismethod for identifying tumor-specific peptides is rela-tively simple, fast and suitable for any known intracellu-lar protein that may be a potential TAA. Nevertheless, itbears the disadvantage that it cannot, by itself, predictwhether the identified peptide is found on HLA moleculesof the leukemia or cancer cells that contain the parentprotein.

2) From peptide to protein. Another strategy used toidentify suitable peptides for cancer vaccines involves thestructural analysis of naturally processed peptides (NPPs)bound to HLA class I and class II molecules of cancercells. NPPs were first isolated and sequenced by acid-elution from immunoaffinity purified HLA molecules andsubsequently compared with existing proteinsequences.108 Alternative approaches are to obtain NPPsby mechanically destroying and acid-treating wholetumor cells and/or by exposing living tumor cells to rapid


166


acid treatment.109 These procedures should characterizetumor-, differentiation stage- and tissue-specific self-antigen MHC-bound peptides as well as the naturallyprocessed proteins from which they are derived and usethem as tools for immunotherapy. The main disadvantageof this approach is that many tumor cells express lowlevels of HLA molecules and the yield of NPPs can bescarce. Nevertheless some immunogenic peptides derivedfrom wild-type p53 protein, melanoma associated MART-1 and gp100 proteins were identified by these meth-ods,110,111 and naturally processed peptides from acutemyeloblastic leukemia cells and CML blasts are nowunder evaluation.112,113 Advantages and disadvantages ofsynthetic versus natural tumor peptides have beenrecently reviewed.114

3) From tumor infiltrating lymphocytes to peptide.Probably the most clinically relevant tumor peptides arethose identified from the epitope analysis of tumor infil-trating lymphocytes (TIL).115-118 In preliminary experimentsHLA class I restricted CTL lines were derived by repetitivein vitro stimulation of TIL with autologous tumor cells.Subsequently, by transfection of a tumor cDNA libraryand in vitro sensitization assays, the peptide sequencesrecognized by the tumor-specific CTLs were identified aswere the parent proteins (i.e. peptides derived frommelanoma associated gp100, tyrosinase and the MAGEfamily). Most TIL-derived tumor peptides found in thepast few years are MHC class I-restricted; however, anovel melanoma antigen resulting from a chromosomalrearrangement and recognized by a HLA-DR1-restrictedCD4+-TIL has recently been identified.119

Checking for peptide immunogenicityExcept for TIL-derived peptides, all other tumor-spe-

cific, HLA binding, synthetic or naturally expressed pep-tides still need to be tested for immunogenicity. Theability of inducing CTL or specific CD4+ proliferation hasbeen evaluated for all tumor-specific peptides that weresubsequently used in clinical trials. P210-derived pep-tides, for example, were able to elicit peptide-specific T-cell immunity both in normal donors,105,120 and CMLpatients.121 Their relevance as TAAs has been furtherconfirmed by observing peptide-specific HLA restrictedCTLs and CD4+ cells able to mediated killing of b3a2-CML cells and proliferation in the presence of b3a2 con-taining cell lysates, respectively.106,107 The latter findingswere the indirect proof of natural processing of P210and of HLA presentation of breakpoint-derived peptides.Although strong peptide-specific CTL and CD4+ respons-es have been shown in vitro for most tumor peptides sofar identified, few data on T-cell induced immunity afterpeptide vaccination in patients have been generat-ed.100,122,123 Thus, strategies to improve peptide immuno-genicity, by using different adjuvants and delivery sys-tems, are currently under evaluation.

Peptide vaccine formulationThe goal of experimental clinical protocols using pep-

tide antigens for active vaccination is to induce a strong

CTL response against the immunizing antigen and there-by against tumor cells expressing the antigen. The modeof peptide-based cancer vaccination critically affects theclinical outcome. The synthesis of a peptide on a largescale, its purification and testing for common QualityControl/Quality Assurance compliance are simple andfast procedures but the choice of an effective deliverysystem for the peptide is crucial. Peptide vaccinationstrategies currently being evaluated include: 1) directpeptide vaccination with immunologic adjuvants and/orcytokines;124 2) lipopeptide conjugates;125 3) peptideloading onto splenocytes or DC;126 4) lysosomal com-plexes.127 Recently, a specific formulation of the poly-saccharide poly-N-acetyl glucosamine has been found tobe an effective vehicle for sustained peptide delivery ina murine vaccine model able to generate a primary CTLresponse with a minimal peptide dose.84 Finally, trigger-ing CD40 in vivo with an activating antibody consider-ably improved the efficacy of peptide-based anti-tumorvaccines in mice, converting a peptide with minimalimmunogenicity into a strong CTL inducer.44

Recombinant virusesThe molecular identification of the antigens on human

tumors recognized by T- and B-lymphocytes offers theopportunity to design novel cancer vaccines based onrecombinant forms of TAA. Genes coding for TAA can beinserted into the genome of attenuated micro-organ-isms such as bacteria and viruses. Viruses are amongthe most interesting vectors since they are able toinduce antibody, Th, and CTL responses in the absenceof co-stimulation.128 Their long-lasting cohabitationwith human beings has likely favored the evolution ofspecific patterns recognized by the innate immune sys-tem which create an immunostimulatory environmentfor optimal immune responses. The vector choice, how-ever, is limited by the possible disadvantages of recom-binant virus utilization, such as recombination withwild-type viruses, oncogenic potential, or virus-inducedimmunosuppression.

Vaccinia virus (VV) belongs to the Poxviridae family,and its worldwide use in the smallpox eradication cam-paign demonstrated that it was safe and very effective.To date, no other large-scale vaccination program hashad such an impact on human diseases, because small-pox has been virtually eliminated from the world popu-lation. Large amounts of foreign DNA can be stablyinserted into the VV genome by homologous recombi-nation.129 VV employs a built-in transcriptional and post-translational apparatus to produce large amounts of theprotein encoded by the inserted gene. VV sojourns with-in the host cell cytoplasm and does not integrate nor isit oncogenic.129 The induction of potent cellular andhumoral immune responses with recombinant (r)VV wasobserved in several tumor systems.130-132 Preclinical stud-ies in models of pulmonary metastatization caused bytumors bearing a prototype TAA have revealed some fea-tures of rVV that are important in determining success-ful therapy. In particular, TAA gene must be expressed

M. Bocchia et al.

167


under the control of a strong, early promoter whichallows its expression in professional APC such as DC.133

Moreover, while prevention from tumor challengerequires only the synthesis of the TAA by the recombi-nant virus, genes encoding immunostimulatory mole-cules inserted in the recombinant poxvirus,134 exogenouscytokines,131,135 or their combination136 are required toinduce eradication of established tumors.

Encouraging results have also been obtained in mousemodels more relevant to the therapy of human cancer.Immunization of mice transgenic for the human HLA-A*0201 allele with an rVV encoding a form of themelanoma antigen gp100, which had been modified toincrease epitope binding to the restricting class I mole-cule, elicited CD8+ T-lymphocytes specific for the epitopethat is naturally presented on the surface of an HLA-A*0201-expressing mouse melanoma.137 Repeated inoc-ulations of an rVV encoding the mouse tyrosinase-relat-ed protein-1 (TRP-1/gp75) caused autoimmune attack ofnormal melanocytes manifested by hair depigmentation(vitiligo) and CD4+-mediated melanoma destruction inmice.138 In a clinical trial, administration of VV encodinghuman carcinoembryonic antigen (CEA) proved effectivein inducing both humoral and cellular immune respons-es in patients with colorectal cancer.139

Poxviruses are not the only choice for tumor immu-nologists. Adenoviruses in which critical genes thatenable viral replication have been deleted and replacedby genes encoding heterologous antigens, have beengenerally used in gene therapy studies, but have alsoprovided antitumor activity when employed as immuno-gens.140 Liver toxicity was described following systemicadministration of high titers of first generation E1-deleted Adenovirus vectors optimized for gene thera-py.141 These side effects, which certainly raise some con-cerns about Adenovirus administration in patients,might not restrain their use in cancer immunotherapysince systemic delivery of high viral titers would cer-tainly not be the favored immunization route.

Initial clinical trials have unveiled some of the intrin-sic limitations of recombinant viruses. Many patients, infact, have high neutralizing antibodies against VV as aconsequence of its use as a vaccine for smallpox preven-tion, and against Adenoviruses, which cause upper res-piratory tract infections throughout life. High doses ofrecombinant adenoviruses expressing the humanmelanoma antigens MART-1 and gp100 could be safelyadministered to cancer patients, but the high levels ofneutralizing antibodies present in their sera likely impairthe ability of these viruses to immunize against themelanoma antigens.142

These results, largely expected, have not put an endto the clinical use of recombinant viruses, as variousstrategies have been exploited to overcome pre-existingimmunity. Several groups have engineered non-repli-cating viruses which normally do not infect humanbeings. The Avipoxviridae family comprises viruses, suchas fowlpox and canarypox, which can productively infectavian but not mammalian cells, and are not cross-reac-

tive with VV.143 A recombinant fowlpox virus expressinga model TAA was able to cure established tumors inmice;144 an important aspect of this study was the obser-vation that prior exposure to VV did not abrogate theimmune responses induced by the recombinant fowlpoxvirus. A different non-replicating virus, canarypox virus(ALVAC), has also been employed to elicit immuneresponses against a variety of antigens.145,146

A highly attenuated strain of VV, known as modified VVAnkara (MVA), has been inoculated as smallpox vaccineinto more than 120,000 recipients without causing anysignificant side effect.147 Replication of MVA is blocked atthe step of virion assembly and for this reason the MVAvectors produce recombinant proteins expressed underthe control of both early and late viral promoters, thusmimicking the expression in wild-type virus. An MVA vec-tor, and a fowlpox virus vector expressing a model TAAshowed better therapeutic effects on pulmonary metas-tasis than a VV encoding the same TAA.148 MVA is a verypromising vector for the development of recombinantvaccines for cancer, and can be efficiently used in com-bination with DNA vaccines.149

As an alternative approach, the mucosal route ofadministration was recently shown to overcome pre-existing immunity to VV. Intrarectal immunization ofvaccinia-immune mice with rVV expressing HIV gp160induced specific serum antibody and strong HIV-specificCTL responses in both mucosal and systemic lymphoidtissue, whereas systemic immunization was ineffectiveunder these circumstances.150

Direct immunization of mice with recombinant aden-oviruses resulted in the induction of high titers of neu-tralizing antibodies, which precluded a boost of CTLresponses after repeated inoculations. The presence ofneutralizing antibodies did not, however, affect theimmunogenicity of infected DC, as repeated administra-tion of virus-infected DC boosted the CTL response evenin mice previously infected with the recombinant vec-tor.151 It was also shown that protective immunity againstmouse melanoma “self” antigens, gp100 and TRP-2, couldbe obtained by DC transduced with Adenovirus vectorencoding the antigen. Importantly, immunization withAdenovirus-transduced DC was not impaired in mice thathad been pre-immunized with Adenovirus.152 With thehelp of these novel strategies, recombinant vectors couldbe used in the general population, including those indi-viduals previously exposed to the viruses. Gene transferto different human DC subpopulations by vaccinia andadenovirus vectors is a conceivable strategy for TAA load-ing.153

DNA vaccinesFollowing the first and somewhat shocking demon-

stration that the intramuscular injection of naked DNA(i.e., DNA devoid of a viral coat) encoding the influenzaA nucleoprotein could induce nuceloprotein-specificCTL, and protect mice from challenge with heterologousinfluenza strains,154 DNA immunization has become arapidly developing technology. This vaccination method


168


provides a stable and long-lasting source of antigen,and elicits both antibody- and cell-mediated immuneresponses. Compared to recombinant viruses, DNA vac-cines offer a number of potential advantages becausethey are cheap, easy to produce, and do not require spe-cial storage or handling. DNA vaccines express virtual-ly only the heterologous gene, therefore, they shouldinduce an immune response selective for the antigenand not the vector, thus supplying a source of antigensuitable for repeated boosting. DNA vaccination hasproven to be a generally applicable approach to variouspreclinical animal models of infectious and non-infec-tious diseases,155 and several DNA vaccines have nowentered phase I/II human clinical trials. Although theclinical application of DNA vaccines is a very youngpractice, some trials have already demonstrated that ispossible to elicit a specific CTL response against malar-ia and HIV proteins in human volunteers.156,157

This novel vaccination approach involves differentsteps: 1) cloning of a heterologous gene under the con-trol of a viral promoter (ordinarily derived from the CMVimmediate early region); 2) purification of the endotox-in-free DNA plasmid from bacteria factories; 3) admin-istration of the expression vectors by direct intramus-cular or intradermal injection with a hypodermic nee-dle or using a helium-driven, gene gun to shoot the skinwith DNA-coated gold beads. Heterologous DNA canalso be introduced into recombinant Salmonella,158,159

or Listeria strains160 that can be thus administered by amucosal route (Figure 3). In addition to these classicroutes of DNA delivery, plasmid-based gene expressionvectors have also been admixed with polymers andadministered with a needle-free injection device,achieving high and sustained levels of antigen-specificantibodies.161 The route of DNA delivery can profound-ly influence the type of immune response by preferen-tially activating different Th populations: gene gun bom-bardment elicits a Th2 response, while intramuscularinoculation induces Th1 activation, even though theantigen form (i.e., membrane-bound vs secreted) canalso exert some effect.162,163

The immunostimulatory activity of DNA vaccines hasbeen associated with the prokaryotic-derived portion ofthe plasmid, which contains a central CpG motif in thesequence PuPuCpGPyPy.164,165 In their unmethylatedform, these hexamers stimulate monocytes and macro-phages to produce different cytokines with a Th1 pro-moting activity including IL-12, TNF-α, and IFN-γ.166,167

A plasmid that incorporated several CpG islands in theprokaryotic ampicillin-resistance gene induced astronger immune response when compared to a secondplasmid carrying the kanamycin-resistance gene whichpossesses none.168 To date, it is not known whether theCpG motifs will have the same immunostimulatoryproperties when applied to vaccination of human beings.However, it was recently reported that CpG motifs canactivate in vitro subsets of freshly isolated human DC topromote Th1 immune responses.169

The mechanism of DNA-induced immunization has not

yet been fully elucidated. An exclusive role for the directtransfection of normal tissue cells, such as myocytes orkeratinocytes, has been debated because surgical ablationof the injected muscle within 1 minute of DNA inocula-tion did not affect the magnitude and longevity of DNA-induced antibodies.170 Moreover, studies with bone-mar-row chimeras clearly indicated that bone-marrow-derived APC, either transfected by the DNA plasmid orable to capture the antigen expressed by other trans-fected cells, were necessary to prime T- and B-lympho-cyte responses.171 Indeed, more recent evidence suggeststhat Th and CTL are activated by DC directly transfectedin vivo following DNA immunization.172,173

The first applications of DNA immunization to preclin-ical models of tumor growth revealed some interestingaspects. In general, the potency of naked DNA does notequal that of recombinant viruses, probably because DNAdoes not undergo a replicative amplification in the trans-fected cells, which in turn limits the amount of heterol-ogous antigen produced. Inflammatory responses causedby DNA inoculation are more contained than those occur-ring during infection with viruses; for this reason, repeat-ed inoculations of plasmid DNA, or the use of adjuvantssuch as cardiotoxin are generally required for the induc-tion of an optimal response. Another emerging issue isthat the efficacy of the vaccination approach dependsmore on the type of antigen than on the route of admin-istration (Table 3). Vaccines based on shared viral anti-gens, or model TAA artificially introduced in the experi-mental tumors can be used to induce a strong, and oftentherapeutic immune response. Using a gene gun for DNAimmunization, Irvine et al.174 observed effective treatmentof established pulmonary metastases, but recombinantcytokines were necessary to enhance the therapeuticeffects. Unlike viral and model TAA, self TAA fail to inducesterilizing immunity since therapy of established tumorshas been rarely reported, and prevention from challengeis often partial.175-178 Central and peripheral tolerance toself antigen has thus emerged as the main limitation tothe successful application of DNA vaccines to the thera-py of cancer. This conclusion seems to apply to severalmouse melanocyte differentiation antigens, a class ofmolecules that is expressed in both melanomas andmelanocytes and includes tyrosinase, TRP-1/gp75, TRP-2,and gp100/pmel 17. However, tolerance can be broken bythe use of a xenogeneic source of TAA. While immuniza-tion with mouse TRP-1/gp75 or TRP-2 antigens failed toinduce a detectable immune response, vaccination witha plasmid DNA encoding the human homologous anti-gens elicited autoantibodies and CTL in C56BL/6mice:178,179 immunized mice rejected metastatic melano-ma and developed patchy depigmentation of their coats(vitiligo). This “obligatory” association of vitiligo and anti-tumor response was recently questioned by a study show-ing that immunization with mouse TRP-2-encoding plas-mid could protect CB6 F1 mice in the absence of overtvitiligo, suggesting a role for the genetic background incontrolling both the extent and the consequences of theimmune activation against self TAA.180

M. Bocchia et al.

169


A novel vaccine that combines the properties of virus-es and DNA is based on antigen production in the con-text of an alphavirus replicon. These new vectors rely onthe ability of the alphavirus RNA replicase to drive thereplication of its own gene, as well as of subgenomicRNA encoding the heterologous antigen. This loop ofself-replication results in a several-fold amplification ofprotein production in infected cells.181 Compared withtraditional DNA vaccine strategies, in which vectors arepersistent and the expression constitutive, the expres-sion mediated by the alphaviral vector is transient andlytic, resulting in a decrease of biosafety risks as well asthe risk of inducing immunologic tolerance due to long-lasting antigen expression. A single intramuscular injec-tion of a self-replicating RNA immunogen at doses low-er than those required for standard DNA-based vaccineselicited antibodies, CD8+ T-cell responses, and prolongedthe survival of mice with established tumors.182 Inter-estingly, the enhanced immunogenicity of these vectorscorrelated with the apoptotic death of transfected cells,which facilitated their uptake and presentation by DC.

Methodology for ex vivo generation of DCInvestigators working in human and murine systems

have discovered culture conditions that use hematopoi-etic cytokines to support the growth, differentiation, andmaturation of large amounts of DC. Therefore, DC canalso be purified from peripheral blood after removal ofother defined T, B, NK and monocyte populations by usingantibodies and magnetic beads or a cell-sorter.183 How-ever, the very low frequencies of DC in accessible bodysamples, especially blood, limits the use of these DC forvaccination protocol. The ex vivo differentiation of DCprogenitors can be traced easily by monitoring changesin some key surface molecules such as CD1a (acquired byDC) and CD14 (expressed by monocytes and lost by DC).Furthermore expression of co-stimulatory molecules suchas CD40, CD80, CD86, as well as HLA antigens, can beused to evaluate the stage of differentiation and thedegree of maturation of DC during in vitro culture. Inaddition, two new markers, CD83 and p55, have beenshown to be selectively expressed by a small subset ofmature DC differentiated in in vitro culture.184,185 Accord-ing to the knowledge of DC ontogeny, two major strate-gies are used. The first is based on the ability of CD34+

progenitors isolated from bone marrow,186 peripheralblood,187 or neonatal cord blood188 to differentiate ex vivowithin 12-14 days into mature CD1a+/CD83± HLA-DR+

DC in the presence GM-CSF and TNF-α. Both stem cellfactor (SCF) and FLT3 ligand are able to augment the DCyield if these key factors are present in the culture.189

The maturation of DC from progenitors is influencednot only by cytokines, but also by extracellular matrix(ECM) proteins, such as fibronectin which has beenreported to enhance DC maturation by mediating a spe-cific adhesion through the a5b1 integrin receptor.190 Thechoice of culture conditions, especially the cytokine com-bination, will influence DC purity, maturation and func-tion, and this is a consideration of prime importance

before starting a DC-based immunotherapy strategy. Amore practical approach is the production of DC fromCD14+ monocytes, in the presence of GM-CSF and IL-4.

A future strategy for easy achievement of largeamounts of DC is the in vivo injection of the samecytokines utilized for ex vivo DC generation. In fact, theadministration of FLT3 ligand either in animals191 or inhumans192 results in a reversible accumulation of func-tionally active DC in both lymphoid and non-lymphoidtissues. Therefore, in murine models it has been demon-strated that FLT3 ligand caused the regression of varioustumors supporting the suggestion that DC may be direct-ly involved in the antitumor effect of FLT3 ligand.193,194

Strategies for delivery of TAA into DCSeveral approaches for delivery of TAA into DC have

been utilized. To date, in the clinical protocol of vacci-nation by DC both synthetic peptides corresponding toknown tumor antigens and tumor-eluted peptides havebeen used for DC-mediated antigen presentation.195

While synthetic peptides represent only the limited anti-genic repertoire of the presently known tumor antigens,tumor-eluted peptides, though originating fromunknown proteins, may reflect a wider antigenic spec-trum. Another potential disadvantage of using definedsynthetic peptides to activate tumor-reactive T-cells isthat the generated peptide-specific T-cells may not rec-ognize autologous tumor cells expressing the antigen ofinterest. Loading DC with cocktails of different synthet-ic peptides, corresponding to different tumor antigensexpressed by the same tumor, has been demonstrated tobe a clinically effective procedure. Nevertheless it is pos-sible that the synthetic peptide-approach will limitpatient selection, on the basis of the HLA phenotype,and will prevent the possibility of activating both CD4and CD8 T-cells directed to different epitopes of thesame antigen. To by-pass these disadvantages, severalalternative methodologies using a mix of TAA have beendeveloped. DC are able to internalize complete tumorlysates or apoptotic cells and to present derived antigenin an HLA I-restricted manner.196 In addition, DC secreteantigen-presenting vesicles, called exosomes, whichexpress functional HLA class I and class II, and T-cellco-stimulatory molecules. Tumor peptide-pulsed DC-derived exosomes prime specific cytotoxic T-lympho-cytes in vivo and eradicate or suppress growth of estab-lished murine tumors in a T-cell-dependent manner.197

However, a possible limitation of these approaches isthe need for large numbers of primary samples or tumorcell lines and the complete lack of control of the natureof the antigens that are being presented by the DC. Theuse of RNA instead of protein could constitute a goodalternative since it could be amplified in vitro. In addi-tion, substractive hybridization could allow the enrich-ment of tumor-specific RNA, thus limiting immuneresponse against self antigen.198 A further possibility isthe engineering of DC with expression vectors carryingTAA genes. Among the viral vectors, retroviral, adenovi-ral, and vaccinia vectors have been widely utilized to


170


transduce either monocyte or CD34+ cell-derived DC.Many authors have chosen retroviral vectors becauseretroviral transduced-DC should be able to constitu-tively express and process TAA to produce long-termantigen presentation in vivo. Specific CTLs against trans-duced-TAA are elicited by retrovirus-engineered DC.199

However, their low efficiency of transduction limits theclinical use of retroviruses.

In contrast, adenoviral vectors infect replicating andnon-replicating cells, are easy to handle, and supernatantwith clinical grade high titer is readily achievable. Bothmonocyte and CD34+ cell-derived DC can be transducedwith high efficiency by adenovirus combined with poly-cations.200,201 Moreover, DC transduced by adenovirusmaintain their APC functions.202 However, the use of ade-novirus vectors is hampered by their immunogenicitywhich causes the rapid development of a CTL responsethat eliminates virus-infected cells and generation ofneutralizing antibodies in recipients. Moreover, vacciniavirus which is a member of the poxvirus family, is notoncogenic, does not integrate into the host genome, iseasy to manipulate genetically and is capable of accept-ing large fragments of heterologous DNA.203 The trans-duction of CD34+ cell-derived DC is feasible but theiruse is limited by the narrow therapeutic index betweenoptimal transduction and target cell viability.153 The pres-ence of acquired genetic abnormalities in myeloid hema-tologic malignancies might represent the basis for inno-vative immune-based anti-leukemic strategies. In fact,intracellular proteins can be processed and presented onthe cell surface by HLA molecules indicating the possi-

bility that leukemia-specific genetic abnormalities maybe targets for cytotoxic T-cells. The generation of func-tional monocyte-derived DC carrying the specific genet-ic lesion has been reported for both acute myeloid andlymphoid leukemia (AML),204,205 as well as chronic myel-ogenous leukemia (CML).206 Current protocols for ex vivoDC generation from CD34+ cells does not allow largenumbers of leukemic DC to be produced, probablybecause of a defective proliferative and/or maturativecapacity of transformed CD34+ cells.

Recently, a protocol which allows the optimal gener-ation of BCR/ABL-positive DC from CML-derived CD34+

cells has been reported.207

Antitumor vaccination: emerging clinical results

Clinical trials in solid tumorsDespite the fact that the large number of ongoing clin-

ical trials which can be derived from the Physician DataQuery (PDQ) of NCI (Figure 4) suggests a diffuse interestin immunotherapy, there is still a strong need to definethe clinical impact of immunotherapy in the treatmentof solid tumors. Table 4 summarizes the already pub-lished vaccination trials carried out using a) autologousor allogeneic neoplastic cells, b) synthetic peptides cor-responding to defined TAA, alone or pulsed on autolo-gous monocytes or DC.

MelanomaMelanoma is the most striking example of a non-virus-

induced immunogenic tumor in man that is able to elic-

M. Bocchia et al.

Figure 3. DNA immunization protocols.

171


it T-cell-mediated antitumor immunity. The majority oftumor antigens defined by T-cells have been identifiedutilizing patients’ T-lymphocytes as effector cells andtumor cells obtained from autologous (metastatic) tumordeposits as the target.68 Several investigators have iso-lated cross-reactive tumor-specific CTLs from peripheralblood, lymphocytes, or tumor-infiltrating lymphocytes ofmelanoma patients, and these CTLs are able to recognizecommon tumor antigen expressed in melanomas thatshare the restricting HLA class I allele.19 Thus, large num-bers of ongoing or published clinical trials have been car-ried out on patients with metastatic melanoma. Experi-mental strategies are encouraged since with currentstandard therapies the prognosis of patients withmetastatic melanoma is poor with a median survival ofabout 6 months.208 Several approaches to induce antitu-mor immune response have been reported.

Irradiated tumor cells. The initial anti-melanoma vac-cines were similar in their preparation to the vaccinesagainst infectious diseases. Crude preparations ofhomogenized tumor cells were mixed with immune-stimulating adjuvants such as viral or bacterial particles.Some of these early vaccines were tested in clinical tri-als and induced objective clinical responses in about25% of the cancer patients.209 The most impressive trialwas reported by Morton et al.64 who administered, to136 stage IIIA and IV (American Joint Commitee on Can-cer, AJCC) melanoma patients, a polyvalent melanomacell vaccine (MVC) comprising 3 allogeneic melanomacell lines. Of 40 patients with evaluable disease, 9 (23%)had regressions (3 complete). Induction of cell-mediat-

ed and humoral immune responses to commonmelanoma-associated antigens present on autologousmelanoma cells was observed in patients receiving thevaccine. Survival correlated significantly with DTH andantibody response to MCV and there was a 3-foldincrease in 5-year survival of patients with stage IVmelanoma. Livingston et al.65 randomized 122 stage IIImelanoma patients free of disease after surgery toreceive treatment with the ganglioside GM2/BCG vac-cine or with BCG alone. All patients were pretreated withlow-dose cyclophosphamide. In most patients vaccinat-ed with GM2/BCG, an antibody production against gan-glioside was demonstrated and this was associated witha prolonged disease-free interval and survival, althoughthe improvement did not reach statistical significance.

Gene-modified tumor cell vaccine. Autologous or allo-geneic tumor or fibroblast cells have been modified toexpress cytokines and/or co-stimulatory molecules and/ora suicide compound. This genetic engineering is mainlyperformed ex vivo using retroviral vectors. In the pub-lished human trials, tumor cells have been transduced toexpress several cytokines. Arienti et al.210 described 12stage IV melanoma patients who underwent vaccinationwith HLA-A2-compatible allogeneic human melanomacells (5x107 or 15x107 cells) engineered to release IL-2.Little toxicity with three mixed clinical responses wasrecorded. Among the nine patients immunologically eval-uated, peripheral blood lymphocytes from three patientsdisplayed enhanced non-HLA-restricted cytotoxicity, andtwo of those individuals had an increased reactivityagainst tyrosinase peptide or gp-100 peptide after immu-


Table 3. DNA vaccines in active immunotherapy of experimental mouse tumors.

TAA Experimental tumor Route of inoculation Adjuvant Vaccine effects Reference

Beta-galactosidase adenocarcinoma gene gun bombardment cytokines treatment of 2-day-old (47)(model TAA) followed by cytokine i.p. (IL-2, IL-12) pulmonary metastases

gag from M-MuLV leukemia i.m., 3 inoculations every 10 days none complete protection (56)from challenge

HPV-E7 sarcoma 3 gene gun bombardments, none complete protection (48)every 2 weeks from challenge

Neu spontaneous mammalian i.m., 4 weekly inoculations IL-2-encoding plasmid partial protection (50)tumor from challenge

P1A mastocytoma i.m., 3 inoculations every 10 days none partial protection (49)from challenge

Idiotype/GM-CSF B-cell lymphoma i.m., 3 inoculation every 3 weeks none partial protection (57)fusion protein from challenge

mouse TRP-2 melanoma 3 gene gun bombardments, IL-12-encoding plasmid partial protection (48)every 2 weeks from challenge

human TRP-1 melanoma 5 gene gun bombardments, weekly none reduction of pulmonary (52)metastases upon challenge; vitiligo

human TRP-2 melanoma 1-4 gene gun bombardments GM-CSF (in therapy setting) reduction of pulmonary (51)metastases (preventionand therapy); vitiligo

mouse TRP-2 melanoma i.m. cardio-toxin protection from challenge; (53)no vitiligo

172


nization. In the only patient for whom the autologousmelanoma line was available, and following in vitro stim-ulation of the PBLs after vaccination, the frequency ofCTL precursor (CTLp) was significantly enhanced. Othergroups utilized a similar approach transducing differentcytokines, i.e. IL-12,40 IL-7,211 and GM-CSF.39 Despite doc-umented specific antitumor reactivity with an increasedfrequency of anti-melanoma cytolytic precursor cells,negligible clinical results were demonstrated. To summa-rize, the advantage of a genetically modified autologouscell vaccine is that it contains the whole collection oftumor proteins and therefore has the greatest chance ofinducing an immune response against relevant tumorantigens. However, growing autologous tumor cells in vit-ro to establish tumor cell lines is time-consuming andoften unsuccessful. These studies are relevant since theyshow that injection of gene-modified cells into a patienta) is safe; b) is followed by efficient and variable trans-duction rate of host tissues; c) is associated with trans-gene expression in the patient; d) is associated with bio-logical activity of the transgene product in most instances.

Synthetic and natural peptides. Phase I clinical trialshave been carried out using synthetic peptides corre-sponding to defined TAA. The clinical trial usingmelanoma differentiation antigen (MAGE-3.A1) byMarchand et al.,123 enrolling 39 chemoresistant stage IVmelanoma patients, was encouraging because monthlyinjection of 100-300 µg peptide alone was associatedwith tumor regression in 7 out of 26 patients whoreceived the complete treatment. All but one of theseregressions involved cutaneous metastases. No evidencefor CTL response was found in the blood of the 4 patientswho were analyzed, including 2 who displayed completetumor regression. In contrast, Jager et al.212 immunizedsimilar patients with gp100 peptide along with GM-CSFand, in some patients, were able to document anincrease in the specific CTL activity against the immu-nizing peptide. Rosenberg et al.213 reported that 31metastatic melanoma patients were immunized withthe modified g209-2M peptide in incomplete Freund’sadjuvant (IFA) along with IL-2 obtaining tumor regres-sion in 42% of patients. Peripheral blood mononuclearcells harvested from these patients after, but not before,immunization exhibited a high degree of reactivityagainst the native g209-217 peptide, as well as againstHLA-A2+ melanoma cells.

These studies indicate that vaccination with synthet-ic peptides is well tolerated, with occasional occurrenceof mild fever and inflammation at the site of injection.Nonetheless, it should be pointed out that there is usu-ally a poor correlation between induction of specific T-cells and the clinical response. The reasons for this dis-crepancy might be the selection of the patients enrolledin the trials since the majority of patients were in stageIV with large amounts of disease.

Dendritic cells. Autologous DC generated from periph-eral blood monocytes have been utilized as antigen-pre-senting cells after their loading with specific melanomaantigens. Chakraborty et al.214 found that intradermal

administration of DC pulsed with a MAGE-1 HLA classI-restricted peptide could elicit peptide and autologousmelanoma reactive-CTLs in patients with advancedmelanoma. However, despite the presence of these CTLpin the vaccination site, peripheral blood, and distanttumor sites, no significant therapeutic responses wereseen.

Nestle et al.195 recently described the immunization of16 melanoma patients using DC loaded with melanomapeptides or tumor lysates. DC were pulsed with a cock-tail of gp100, MART-1, tyrosinase, MAGE-1, or MAGE-3peptides chosen to suit the individual patient class I HLAmolecules. Four patients whose HLA haplotype was inap-propriate for peptide pulsing received DC pulsed withautologous tumor lysate. Keyhole limpet hemocyanin(KLH) was included during antigen pulsing. DC wereadministered by direct injection into uninvolved lymphnodes via ultrasound guidance to facilitate entry into thelymphatics and to minimize DC loss. Patients received 6-10 injections of 1x106 cells every 1-4 weeks. Toxicity waslimited to mild local reactions at the injection sites.Immunologic monitoring revealed DTH skin reactions topeptides in 11 cases, and peptide-specific CTLs could berecovered from the skin biopsies of some patients.Regression of tumor was seen in 5 out 16 patients,including 2 complete responses lasting over 15 months.Responding tumor sites included skin, lung, soft tissue,bone, and pancreas. Importantly, two of the respondingpatients received only tumor lysate-pulsed DC, suggest-ing an approach applicable to cancers lacking definedtumor antigens.

In a similar but smaller trial at the University of Pitts-burgh,215 6 HLA-A2+ patients with metastatic melanomareceived four weekly intravenous injections of 1-3x106

monocyte-derived DC pulsed with HLA-A2-restrictedpeptides derived from MART-1, gp100, and tyrosinase.Complete regression of a subcutaneous mass lastingmore than one year has been observed in one patient.

CCoolloonn ccaanncceerrColon cancer is potentially curable by surgery; the

cure rate is, however, moderate to poor depending onthe extent of disease. Adjuvant chemotherapy with 5-fluorouracil plus levamisole or folinic acid is the stan-dard treatment for stage III colon cancer based on theresults of numerous co-operative and intergroup clini-cal studies. In contrast, adjuvant chemotherapy for stageII disease has no benefit.216 Despite several immunother-apeutic approaches having been tested for colon can-cer patients, only one study has reported clinical results.In a prospective randomized study,217 254 patients withstage II or III post-surgery colon cancer were randomlyassigned to receive active specific immunotherapy,namely autologous tumor cell-bacille Calmette-Guèrin(BCG) or no adjuvant treatment. The immunotherapyprogram comprised three weekly intradermal injectionsstarting 4 weeks after surgery, with a booster vaccina-tion at 6 months with 107 irradiated autologous tumorcells. The first vaccination contained 107 BCG organ-

M. Bocchia et al.

173


isms. The 5-year median follow-up showed a 44%reduction of risk of recurrence in all patients receivingthe vaccinations. The major impact of immunotherapywas evident in patients with stage II disease, who hada significantly longer disease-free period and 61% riskreduction. In addition, no patient discontinued treat-ment early because of side effects.

Recently, Foon et al.218 generated anti-idiotype anti-body, designated CeaVac, that is an internal image ofCEA. Thirty-two patients with resected Dukes’ B, C, andD, and incompletely resected Dukes’ D disease weretreated with 2 mg of CeaVac every other week for fourinjections and then monthly until tumor recurrence orprogression. Fourteen patients were treated concur-rently with a 5-FU chemotherapy regimen. All 32patients entered into this trial generated a potent anti-CEA humoral and cellular immune response. Interesting,the 5-FU regimen did not affect the immune response.A phase III trial for patients with resected colon canceris ongoing.

PPrroossttaattee ccaanncceerrSeveral prostate-tissue-associated antigens, including

prostatic alkaline phosphatase (PAP), prostate-specificmembrane antigen (PSMA), and prostate-specific anti-gen (PSA), are now being explored as targets for prostatecancer immunotherapy.219-221 Valone et al.221 have carriedout a dose-escalation trial of peripheral blood DC pulsedwith recombinant PAP protein in 12 patients withadvanced prostate cancer. Intravenous administrationof 0.3, 0.6, and 1.2x109 pulsed cells/m2 monthly for threemonths resulted in T-cell proliferative responses againstPAP in all patients, the magnitude of which was relat-ed to cell dosage. Toxicity was limited to myalgias inthree patients. Clinical outcomes have not been report-ed.

Salgallar et al.222 have recently updated the results ofanother trial223 in which monocyte-derived DC pulsedwith HLA-A2-binding peptides derived from PSMA wereadministered to 82 patients with advanced, hormone-refractory prostate cancer. Six infusions of up to 2x107

DC were administered every six weeks, with half of thepatients also receiving systemic GM-CSF. The treatmentwas well tolerated. However, only two patients mount-ed T-cell responses against the PSMA peptides as mea-sured by enzyme-linked immunospot (ELISPOT) and DTHskin testing. Although four patients had reductions inserum tumor markers following vaccination, the con-current administration of radiotherapy and hormonaltherapy makes interpretation of the vaccine’s effectsdifficult.

RReennaall cceellll ccaarrcciinnoommaaRenal cell carcinoma accounts for 2% of all malig-

nancies and many patients have metastatic disease atdiagnosis and the prognosis is unfavorable. At present,neither chemotherapy nor radiation therapy has any sig-nificant influence on the course of disease or the sur-vival time. Immunotherapy using recombinant IL-2alone224 or combined with interferon-α225 is currentlythe standard therapy for metatastic renal cell carcino-ma. Cellular immunotherapy includes the adoptivetransfer of in vitro expanded tumor infiltrating lympho-cytes226 as well as active immunotherapy with an autol-ogous tumor cell vaccine engineered to secrete GM-CSF.38 Although each of these attempts generatedpromising results neither attempt met the expectations.Recently, Holtl et al.227 have administered, to 4 metasta-tic renal cell carcinoma patients, autologous monocyte-derived DC pulsed with autologous tumor cell lysate.Each patient received 3 monthly intravenous infusionswith the immunogeneic KLH. Initial results have shown


Figure 4. NCI-registered immunotherapeutic trials according to diseases and strategies.

174


that a potent immunologic response to KLH and, mostimportantly, against cell lysate could be measured invitro after the vaccinations. In addition, the treatmentwas well tolerated with moderate fever as the only sideeffect. In contrast, only one partial response after 2 vac-cinations was observed.

Recently, Kugler et al.228 vaccinated 17 patients withmetastatic renal cell carcinoma using hybrids of autol-ogous tumor and allogeneic DC generated by an elec-trofusion technique. After vaccination, and with a meanfollow-up time of 13 months, four patients completelyrejected all metastatic tumor lesions, one presented amixed response, and two had a tumor mass reduction ofgreater 50%. These promising data indicate that hybridcell vaccination is a safe and effective therapy for renalcell carcinoma and may provide a broadly applicablestrategy for other malignancies with unknown antigens.

Hematologic malignanciesTTuummoorr vvaacccciinneess iinn BB--cceellll llyymmpphhoopprroolliiffeerraattiivvee ddiissoorrddeerrss

Multiple myeloma (MM) and low-grade non-Hodgkin’slymphomas (NHL) are clonal expansions of lymphoid cellsthat have rearranged immunoglobulin (Ig) genes. Earlyduring development, pre-B-cells become committed tothe expression of a heavy and light chain Ig variableregion. The heavy chain derives from the recombination

of variable (V) with diversity (D) and joining (J) regiongenes with a constant region (C). The V-D-J joins occurwith a variable number of nucleotide insertions or dele-tions resulting in a unique sequence which creates thethird hypervariable region (CDR III) and contributes tothe antigen-binding site. These antigenic regions (idio-type; Id) are characteristic for any given Ig-producingtumor (e.g. MM and NHL) and can be recognized by animmune response consisting of anti-Id antibodies and/orby Id reactive T-cells.229-235 The tumor-derived Id is a selfprotein which, in most circumstances, is poorly immuno-genic. However, haptens and adjuvants, includingcytokines, have been used in several animal models toincrease Id immunogenicity and establish protective anti-Id-immunity.236 Lately, Id vaccines have come into med-ical use in patients with lymphoma and MM.

Idiotype vaccination in human lymphoma A pioneering study was carried out in 9 lymphoma

patients in CR or partial remission. They were immunizedwith subcutaneous injections of autologous Id, conju-gated to KLH and emulsified in an oil-in-water emulsioncontaining non-ionic block polymers.80 Specific anti-Idhumoral and/or cellular responses were observed in 7/9patients. Two patients with measurable disease showeda clinical improvement. These results have been con-

M. Bocchia et al.

Disease Stage Pts VaccineRoute of Admin. TA Ads Side Effects Clinical Results

Immunol Results Authors

Melanoma IIIA/IV 136 Melanoma Cell Vaccine (MCV) ID

Human Melanoma Cell

LinesYes/No Mild (Erithema,

Fever)9/40 evaluable pts

(3 CR, 6 PR)

↑Cell mediated and humoral responses to MCV ; ↑ Activation of TIL

Morton, 1992

IV 12 IL-2 Transduced-Melanoma Cell Line SC Melanoma Cell

Line No Mild (Erithema, Fever) 3 MR; 1 SD ↑ Melanoma-Specific

CTLp Arienti, 1996

IV 10IL-7 Transduced-

Autologous Melanoma Cells

SCAutologous Melanoma

CellsNo Mild (Fever) 2 MR 5 SD ↑ Melanoma-Specific

CTLp Moller, 1998

IV 21GM-CSF Transduced-

Autologous Melanoma Cells

ID and SCAutologous Melanoma

CellsNo Mild (Erithema,

Induration, Itching) 1 PR; 1 MR; 3 Minor Res↑ Melanoma-Specific CTLp; 80% Tumor Destruction into Metastases

Soiffer,1998

IV 31 Peptide+IL-2 SC modified g209-2M

IL-2 and IFA Mild (Erithema) 6 PR, 3 MR, 3 SD ↑ Melanoma-Specific

CTLp Rosenberg, 1998

IV 39 Peptide SC and ID MAGE-3 (HLA-A1) No Mild 3 CR; 4 PR

No Increase of anti MAGE-3 CTLs even

in RespondersMarchand,1999

IV 16Ag Pulsed Autologous

Monocyte- derived DCs Lymph nodes

Peptide Cocktail or Autologous

Tumor Lysate KLH Mild (Fever) 2 CR; 3 PR;1 MR

DTH to KLH in 16; DTH to peptide-pulsed DCs in 11

Nestle, 1998

IV 17Ag Pulsed Autologous

Monocyte- derived DCs

ID Autologous Tumor Lysate No No 1 PR

DTH to Vaccine in 9/17; CD8 Cells in

expanded-VILChakraborty, 1998

Colon Cancer II and III 254 Irradiated Autologous

Tumor Cells ID Autologous Tumor Cells BCG Mild

Stage II: 61% Risk Reduction for Recurrences;

Stage III: no Significant Benefits

DTH+: ≥90% Pts Vermorken,1999

Prostate Cancer

Locally Advanced 82

Ag Pulsed Autologous Monocyte- derived

DCs Lymph nodesPeptide

Cocktail or Autologous

Tumor Lysate KLH Mild (Fever) 2 CR; 3 PR; 1 MR

DTH to KLH in 16; DTH to peptide-pulsed DCs in 11

Salgallar, 1998

Renal Cell Cancer IV 12 Ag Pulsed Monocyte

derived-DCs IV Autologous Tumor Lysate KLH Mild (Fever) 1 PR DTH to KLH after 1°

and 2° Vaccination Holt, 1999

Pts: Patients; TA: Tumor Antigens; Ads: Adjuvants; SC: Subcutaneous; MR: Mixed Response; SD: Stable Disease; CTLp: Cytotoxic T Lymphocyte precursor; ID: Intradermic; PR: Partial Response; TIL: Tumor-Infiltrating LymphocytesIFA: Incomplte Freund's Adjuvant; KLH: Keyhole Lympocianin; DTH: Delayed Type Hypersensitivity; IV: Intravenous

Table 4. Phase I/II trials of vaccination in patients with solid tumor.

175


firmed in a larger series of patients.82 Following standardchemotherapy, 41 patients with B-cell lymphomareceived subcutaneous injections of autologous Id-KLHconjugates mixed with an oil-in-water emulsion con-taining non-ionic block polymers and threonyl-muramyldipeptide. Approximately 50% of the patients generatedspecific anti-Id responses and isolated tumor regressionswere observed. In particular, 11/16 patients had a signif-icant increase in the frequency of tumor-specific cyto-toxic T-lymphocytes precursor (CTLp).237 The medianduration of freedom from cancer-progression was sig-nificantly prolonged and this resulted in a survival advan-tage, especially in patients who generated cell-mediat-ed anti-Id immunity.

These pioneering studies did not prospectively inves-tigate the effect of Id vaccines on tumor burden, sincemost patients were already in clinical remission, andstandard tumor regression criteria could not be used. Arecent study has directly evaluated the ability of Id vac-cines to eradicate residual t(14;18)+ lymphoma cells in20 patients in first remission after ProMACE-basedchemotherapy.238 These patients received multiple injec-tions of Id /KLH conjugates in the presence of GM-CSF.Eight of eleven patients with detectable translocationsin the peripheral blood converted to a PCR negative sta-tus after vaccination. Tumor-specific cytotoxic CD8+ andCD4+ T-cells were uniformly seen in most patients. Anti-bodies to autologous Id were also detected, but theywere apparently not required for molecular remissionsince the latter was achieved in some patients withouta detectable antibody response. Clinical monitoring indi-cates a 90% disease-free survival after a median follow-up of 3 years. This is encouraging compared with the44% disease-free survival (after a median follow-up of3 years) in another series of patients treated at the sameInstitution with anti-B4-blocked ricin in first remissionafter ProMACE-based chemotherapy. The encouragingresults obtained in lymphoma patients have provided therationale for exploring the use of Id vaccination in MM.

Id vaccination in human MM MM has several biological features that can advanta-

geously be exploited in the setting of active specificimmunotherapy. Among others, pre-existing tumor-spe-cific T-cell immunity,229-232 the possibility of using clon-al markers to track the fate of residual tumor cells,239

and the preserved susceptibility of chemoresistantmyeloma cells to the effector mechanisms of cytolyticT-cells240 may, together, represent a favorable basis forthe efficacy of Id vaccines.

In a pilot study, five MM patients were repeatedlyimmunized with autologous Id precipitated in alumini-um phosphate suspension.77 Four patients were previ-ously untreated and one patient was in stable-partialremission following chemotherapy. Three patients devel-oped specific anti-Id T- and B-cell responses, but theseresponses were transient and their magnitude was low.A more effective immunization schedule was developedwith the goal of achieving long-lasting T-cell anti-Id

immunity. This effort was focused on early stage MM,based on the assumption that Id-specific T-cells are pre-sent at higher frequency mainly in patients with earlystage MM or MGUS.231 Most of these Id-reactive T-cellsare Th1-type cells in early stage MM, whereas they arepredominantly Th2-type in patients with advanced dis-ease. Thus, a more effective antitumor T-cell immuneresponse may be expected if vaccines are delivered whenthe frequency of T-cells with the potential to developcytotoxic activity is higher. In this series, patientsreceived subcutaneous injections of autologous Id pre-cipitated in aluminium phosphate suspension, togetherwith free GM-CSF.241 Long-lasting Id-specific T-cellresponses were induced in all five immunized patients.Moreover, one patient showed a decrease of circulatingId upon immunization.

A vaccination trial in which Id/KLH conjugates andGM-CSF were administered subcutaneously as a main-tenance treatment after high-dose chemotherapy andPBPC infusion has recently been published.242 Mostpatients generated Id-specific DTH reactions. DTH speci-ficity was confirmed in one patient by investigating thereactivity to synthetic peptides derived from the VDJsequence of the tumor-specific Ig heavy chain. In 3patients with minimal residual disease, the DTH skintests remained positive up to two years after the lastimmunization, but residual tumor cells were not elimi-nated by these long-lasting immune responses. Never-theless, these patients remained in clinical remissionwithout any further maintenance treatment. Thus, it ispossible to generate anti-Id immune responses that arenot potent enough to eliminate residual tumor cells, butare sufficient to hold the disease in check for extendedperiods.

These results have been confirmed in a preliminaryreport of 18 patients with MM receiving Id/KLH conju-gates and GM-CSF in first remission after high-dosechemotherapy and tandem transplantation followed byPBPC infusions. In particular, 50% of the patients gen-erated positive DTH reactions and 2 patients, in partialremission at the time of vaccination, achieved a com-plete remission following vaccination. A retrospectivepair-mate analysis has shown a trend for a better clini-cal outcome in MM patients receiving Id vaccines com-pared to those receiving IFN-α alone. This tendency isparticularly evident in patients who generated positiveId-specific T-cell responses. Although preliminary andretrospective, this is the first study providing clinical evi-dence that the generation of T-cell immune responseagainst tumor cells may positively influence the clinicaloutcome of MM patients in the remission phase (N.Munshi and L. Kwak , personal communications).

DC-based anti-Id vaccinationAs previously discussed, a proportion of lymphoma

and MM patients enrolled in clinical trials mounted anId-restricted antibody and T-cell response and some ofthem showed tumor regression. However, there areimportant differences between lymphoma cells and MM


176


plasma cells. In the case of lymphomas, the cells arecharacterized by high surface expression and little anti-body secretion whereas myeloma cells have very lowlevels of cell surface Ig with high levels of antibodysecretion. Thus, it is unlikely that the sole generation ofan antibody-based anti-Id immune response will bebeneficial for MM patients. In fact, anti-Id antibodiesmay be blocked from reaching the tumor cells by thehigh levels of circulating Id. Moreover, despite the exis-tence of a pre-plasma cell stem cell compartment inMM with a higher expression of surface Ig, it is possi-ble that tumor cells would not express enough targetprotein for the antibodies to be effective. Conversely,an Id specific T-cell response would not need to bind tocell surface Ig to be active. T-cells do not recognizeintact protein, but are specific for processed peptidefragments of the Id expressed on class I or II molecules.The advantage of a cytotoxic T-cell response is that itwould not be blocked by free circulating paraproteinand would not depend on the expression of the nativeprotein on the surface of tumor cells. Moreover, B-cells,including putative myeloma stem cells, are known toprocess and present peptides of the Ig on their mem-brane associated with class I and II molecules. Therefore,optimal strategies for Id vaccination may require theinduction of a T-cell-mediated immune response whichis best achieved by the use of APC. In this view, the rapidgeneration of a T-cell immunity in healthy volunteersafter a single injection of mature DC has recently beendescribed.243 Nine normal subjects were given subcuta-neous injections of monocyte-derived mature DCunpulsed or pulsed with KLH, tetanus toxoid (TT) or HLA-A*0201-positive restricted influenza matrix peptide.Four other individuals received these antigens withoutDC. Of note, administration of unpulsed DC or antigensalone failed to induce any T-cell response. Conversely, aCD4+ T-cell response was observed in 9/9 and 5/6 sub-jects injected with KLH or TT-pulsed DC, respectively.Moreover, a significant stimulation of effector andmemory CD8+ CTLs was also reported. This feasibility tri-al provides the first controlled evidence of the capacityof DC to stimulate T-cell immunity.

DC-based anti-Id vaccination has been reported In B-cell malignancies in a few papers. Hsu et al. have244

described 4 low-grade non-Hodgkin’s lymphoma (NHL)patients resistant to conventional chemotherapy orrelapsed who were injected intravenously with Id-pulsedDC freshly isolated from PB by subsequent enrichmentsteps. A tumor-specific T-cell response was observed inall cases associated, in one case, with tumor regression.Sixteen patients have been treated so far and an anti-Id restricted cellular response has been observed in 8subjects (R. Levy, personal communication). The samestrategy of targeting the Id has been applied by the samegroup to induce a T-cell immune response in MMpatients.245 Twelve patients were injected, 3 to 7 monthsafter autologous stem cell transplantation, with Id-pulsed DC followed by 5 subcutaneous boosts of Id/KLHadministered with adjuvant. Whereas 11/12 patients

developed a strong KLH-specific cellular proliferativeresponse, thus suggesting immunocompetence afterhigh-dose chemotherapy, only 2 individuals generatedan anti-Id restricted T-cell proliferation and only 1/3patients showed a transient Id-specific CTLs response.This approach raises concerns about the efficacy ofuncultured blood DC of stimulating efficiently T-cells,the capacity of Id-loaded DC to reach secondary lym-phoid tissues to prime T-cells escaping the entrapmentof the lungs and the role of Id-KLH boosts after DCadministration.

Wen et al.246 reported immunization of one MM patientinjected with Id-pulsed DC derived from adherentmononuclear cells in the presence of appropriatecytokines. In this paper, Id-specific T- cell proliferationand secretion of IFN-γ were reported, as were the pro-duction of anti-Id antibodies. Similar results have recent-ly been reported by Cull et al. who treated two patientswith advanced refractory myeloma with a series of fourvaccinations using autologous Id-protein pulsed DC com-bined with adjuvant GM-CSF.247 DC were derived fromadherent mononuclear cells. Both patients generated aspecific T-cell proliferative response that was associatedwith the production of IFN-γ, indicating a Th-1-likeresponse. However, no Id-specific cytotoxic T-cellresponse could be demonstrated. Lim and Bailey-Woodhave also treated 6 MM with DC generated from adher-ent mononuclear cells.248 DC were pulsed with the autol-ogous Id and KLH as a control vaccine. All patients devel-oped both B- and T- cell responses to KLH, suggesting theintegrity of the host immune system. Id-specific respons-es were also observed. In one patient, a modest but con-sistent drop in the serum Id level was observed. Lastly, avaccine formulation based on CD34 stem cell-derived DCpulsed with Id-derived peptides has recently been used in11 MM patients with advanced disease.249 Five patientsgenerated Id-specific immune responses and one patientshowed a decreased plasma cell infiltration in the bonemarrow.

New strategies in Id vaccinationThe generation of an effective antitumor response

greatly depends on the final activation of tumor-specificcytolytic CD8 cells. This is the final event resulting from aseries of cognate and non-cognate interactions occurringamong tumor cells, professional APC, CD4, and CD8 cells.Each of these cell populations plays a unique role, andmay represent a possible target of immune intervention toimprove the efficacy of vaccination. Malignant B-cells canbe modified to become efficient APCs themselves and pre-sent peptides from their own tumor-specific antigens toautologous T cells. To this end, a number of strategies arecurrently under preclinical and clinical evaluation. Onepossibility is to fuse tumor cells with dendritic cells. Thefusion product will combine the functional properties ofDC with the full antigenic repertoire of tumor cells.250 Asan alternative, malignant B-cells can be turned into effec-tive APC by stimulating cell surface CD40 with its specif-ic counter-receptor CD40 ligand.251,252 Genetic engineering

M. Bocchia et al.

177


is another approach that may turn malignant B-cells intoeffective APC. Transfection with immunologically relevantDNA sequences coding for cytokines or co-stimulatorymolecules greatly enhances the ability of malignant B-cells to activate antitumor immune responses. Thisapproach has recently been used in MM taking advantageof the selective expression of functional adenoviral recep-tors on the cell surface of myeloma cells.253

Interestingly, there has been a description254 of theimmunization of a matched related donor of allogeneicbone marrow with the myeloma derived Id (conjugatedwith KLH) isolated pretransplant from patients. Aftertransplantation, a CD4+ T-cell line was established fromthe peripheral blood of the recipient and found to be ofdonor origin and proliferate specifically in response tothe myeloma Id. Thus, this experience demonstrates theprinciple of transfer of donor immunity with the advan-tage of immunizing a tumor naive donor who may bemore likely to be able to generate an immune responseagainst the tumor Ig without interference or suppressionby the malignant cells. However, the lack of an anti-IdCTL response and ethical concerns on the immunizationof healthy donors with tumor-derived products, suggestthat in the future an alternative strategy based upon thegeneration, ex vivo, of Id-reactive CTL clones by means ofAPC, will be needed.

The prerequisite for any vaccine-based strategy is thepossibility of differentiating tumor cells from normal cells.Id is absolutely tumor-specific, but is not directly relat-ed to the malignant phenotype of myeloma cells, and isself-Ag. As such, it is protected by self-tolerance mech-anisms. There is a growing list of alternative tumor-spe-cific antigens that can be exploited as targets for activespecific immunotherapy. Among others, the core proteinof Muc-1,255,256 antigens encoded by MAGE-type genes,257

overexpressed or fusion proteins resulting from chromo-somal abnormalities may all represent alternativeimmunogens.258,259 Compared to Id, some of these anti-gens may be more intrinsically related to the malignantphenotype of tumor cells.

Polymorphism of the HLA molecules is a major obsta-cle in the outcome of vaccines aimed at triggeringcytolytic T-cell responses. By combining amino acidsequencing of tumor-specific antigens and HLA typing, itis now possible to predict whether the HLA alleles of agiven individual can bind tumor-derived peptides. Sever-al groups are planning to use Id vaccination only in thosepatients for whom preliminary sequencing demonstratesa compatible restriction between HLA and peptides.

Finally, a new generation of immunogens has beendeveloped using DNA-based technologies. The wholetumor-specific immunoglobulin or the variable regionsequences of both heavy- and light-chains have been usedas immunogens or to produce recombinant proteins inbacteria. However, it has soon become clear that nakedDNA is not immunogenic per se and additional sequencesare required to elicit protective immune responses.260

Sequences coding for cytokines, chemokines or xeno-geneic proteins have been included in these constructs

and used as adjuvants.261-264 Experimental data indicatethat these fusion genes have indeed the potential to raiseprotective immunity in both NHL and MM.

Vaccine trials in chronic myelogenousleukemia

Chronic myelogenous leukemia (CML) is a biphasicneoplastic disorder with a prolonged indolent phase last-ing an average of 4 years followed by an acute phase ofblastic transformation which inevitably leads rapidly todeath. There is no curative therapy for CML other thanallogeneic bone marrow transplantation, an option thatis available only to a small fraction of patients who haveboth a matched donor and are young enough to toler-ate the procedure.

Recently, IFN-α has been shown to induce hematolog-ic remissions in most CML patients, with a relevant portionof them also experiencing several degrees of cytogeneticresponse and ultimately a statistically significant prolon-gation of their chronic phase. However, still too few CMLpatients are long survivors if not cured regardless of thetreatment option they received. Because of the unique fea-tures of this disease, the hallmark translocation that char-acterizes all neoplastic cells, a therapeutic targetingapproach only the Ph+ clone could be a powerful tool inthe treatment of CML.

The first direct evidence of the immune system’s cru-cial role in recognizing and eliminating Ph+ CML cellscame from the demonstration that infusion of large dos-es of peripheral blood leukocytes from the marrow donorinduced durable remission in patients with CML who hadrelapsed following a T cell depleted marrow allograft.265,266

This latter finding proved that in CML the graft-versus-tumor effect is mediated by the cellular arm of theimmune system. While the nature of the response is like-ly to be largely allogeneic a possible role for specific anti-CML responses is suggested by the lack of this observa-tion in patients with other myeloid leukemias undergo-ing the same treatment.267

The hypothesis that CML cells could be recognized bythe immune system through the presentation of P210,the tumor-specific product of the bcr/abl hybrid gene,was first tested. The evidence that P210 b3a2-breakpointpeptides were able to bind HLA class I and HLA class IImolecules and to elicit specific T cell responses in normaldonors provided the rationale for a peptide vaccine inCML patients.103-106 Pinilla-Ibarz et al.124 have recentlycompleted a phase I dose escalation trial (5 doses over 10weeks) of a multivalent peptide vaccine (5 peptides) plusQS21 in patients with CML and b3a2 breakpoint. Patientcharacteristics included hematologic remission, IFN-αtherapy and no HLA restriction. In a preliminary report thepeptide vaccine appeared safe with patients experienc-ing only minimal discomfort at the site of injection.

With regards to the immune response, peptide-specif-ic delayed hypersensitivity (DTH) in vivo and peptide-spe-cific proliferation in vitro were shown but no peptide-specific CTL response was induced. Bocchia et al. are cur-rently conducting a multicenter phase I/II trial of a pen-


178


tavalent peptide vaccine plus QS21 and GM-CSF in b3a2-CML patients expressing any of HLA A3, A11 B8 or DR11.Patient characteristics also include major or completecytogenetic response with or without IFN-α maintainingtherapy. The protocol comprises 6 s.c. vaccinations withpeptides + QS21 at 2 weekly intervals, with GM-CSFinjected at the vaccine site for 4 consecutive days start-ing the day before each vaccination. Goals of the studyare the evaluation of the induction of peptide-specific T-cell response, of the role of GM-CSF as immunologicadjuvant in CML patients and the impact of the peptidevaccine on minimal residual disease.

As in other cancers, the use of DC as powerful induc-ers of an active specific immune response in CML is nowunder in vitro evaluation.107,268 Interestingly, most CML-derived DC carry the t(9;22) translocation and thereforecould naturally present P210 derived peptides. In fact,CML derived Ph+ DC were able to strongly stimulateautologous T-cells that displayed vigorous cytotoxicityactivity against autologous CML cells but low reactivityto HLA-matched normal bone marrow cells or autolo-gous remission state bone marrow mononuclear cells.206

P210-derived peptides could have a role in inducing thisleukemia-specific response and a vaccine strategy whichcombines Ph+ DC and breakpoint peptides will be inves-tigated.

ConclusionsThis review shows that there are many prospects of

curing cancer through the active induction of a specificimmune response to TAA. The terms of the matter arenow defined with molecular and genetic details formelanomas. Ongoing research is aimed at defining TAAon other forms of tumors. Indeed, experimental data andvery recent clinical evidence suggest that antitumor vac-cines will soon be a new form of tumor treatment thatwill be able to be adopted for the management ofdefined stages of neoplastic disease, in sequential asso-ciation with conventional treatments.269

Prediction of when the efficacy of antitumor vaccina-tion will be assessed and will become a routine proce-dure is beyond a simple scientific evaluation. While pre-clinical research has identified several possible targetsand strategies for tumor vaccination, the clinical sce-nario is far more complex and as yet no specific clue hasemerged to clearly envisage a clinical development strat-egy which could make biotechnology investments in thisarea attractive enough to pharmaceutical companies.Patent issue complexity further contributes to slowingdown the development of expensive clinical trial pro-grams. A cautious, yet attentive attitude seems, at themoment, to be the general behavior of the pharmaceu-tical industry.

At present peptide vaccination may appear of moreimmediate application. Several phase I clinical trials havealready been carried out using synthetic peptides fromdefined TAA. These peptides have been administeredalone123 or combined with adjuvants, or presented bymonocytes or DC.269 Nearly all studies indicate that this

form of vaccination is well tolerated, mild fever andinflammation at the site of injection being the only occa-sional side effects observed. Nonetheless it should bepointed out that there is usually a poor correlationbetween peptide ability to induce a T-cell response andclinical response. Among the several reasons that mayaccount for this discrepancy, the choice of the peptidesmay have a critical importance. Most peptide-based vac-cines have considered HLA class I restricted peptidesonly, whereas there is increasing evidence that tumor-specific CD4+ T-cells may be important in inducing aneffective antitumor immunity. The addition of peptidesthat bind class II HLA glycoproteins to peptide vaccinescould lead to an amplification of the immune responseas well as to better clinical effect.

A survey of the outcomes of vaccination trials showsthat the poor correlation between induction of immuno-logic responses and the clinical results is a consistentfinding, independently of the immunizing strategyadopted. Many factors may contribute to this poor cor-relation, e.g.:

a) the selection of the patients enrolled in the trial:tumor burden, stage of disease, and others, as discussedabove;

b) the techniques used for the immune monitoring invitro: most of the current studies evaluate T-cell induc-tion through in vitro peptide stimulation of PBMC, whilethe use of tetrameric soluble class I-peptide complex-es270 or reverse solid phase ELISPOT analysis271 may pro-vide complementary information;

c) the assays for immune monitoring in vivo: often apositive DTH test does not correlate with evident tumorregression.

Perhaps, fine needle biopsies at the site of regressingand non-regressing tumors could provide a more directinsight into the events associated with the clinical out-come. The immune pattern within the tumor should becompared to the one found in the skin. This should allowevaluation of the exact role of vaccine-induced T-cells.Which of the existing in vitro and in vivo assays corre-lates most accurately with clinical responses remains tobe established.

Several recent studies showing that the immune sys-tem recognizes TAA during tumor growth did not clari-fy whether such recognition was indeed associated withsubsequent tumor cell destruction. The development ofreliable assays for efficient monitoring of the state ofimmunization of cancer patients against TAA is as animportant goal that will markedly affect the progress ofantitumor vaccines. A major problem in testing the effi-cacy of antitumor vaccination in adjuvant settingsdepends on both the long period of time and large num-ber of patients required. The possibility of effectivelymonitoring the immune response induced acquires crit-ical importance since it may provide a much earlier sur-rogate end-point, predictive of the clinical outcome.

The downregulation of the expression of TAA repre-sents another crucial issue for vaccination therapies,since it may lead to the immunoselection of tumor cell

M. Bocchia et al.

179


clones that hide the target TAA. An ideal TAA is a pro-tein that is essential for sustaining the malignant phe-notype, and that is not stripped or downmodulated bythe immune reaction. Mutations that give rise to TAA ofthis kind have been described.272 However, they will bean appropriate target only for the tumors expressingthese particular mutations and will not be suitable formore general cancer vaccines. Improvements in theidentification of tumor-associated mutations that maybe potentially recognized by the immune system mayalso open up the possibility of tailoring individual can-cer vaccines. The recent report of the construction offusion cells composed of autologous tumor cells and DCor TAA pulsed-DC represents a step forward towards thequick manufacture of tumor-specific and individualizedvaccines. In contrast, the characterization of the telom-erase catalytic subunit (hTERT) expressed in more than85% of human cancers appears to open the way to anovel strategy for a general anticancer vaccination tar-geting a widely distributed TAA.273

The in vivo or ex vivo introduction of TAA genes into DCthrough recombinant viral vectors is still hampered bythe lack of an ideal viral vector and by the induction ofan immune response against the viral proteins. Nonethe-less, many viral vectors successfully used in animal mod-els and currently tested in clinical trials appear to be safevehicles for gene transfer, without any major toxicity. Tocontrol transgene expression levels better, investigatorsare exploiting tissue- or cell-specific regulatory elementssuch as cytomegalovirus promoters and enhancers thatare preferentially expressed in tumor cells.

Certainly the new prospects opened by antitumor vac-cines are fascinating. When compared with conven-tional cancer management, vaccination is a soft, non-invasive treatment free from particular distress andiatrogenic side effects. Antitumor vaccines can beexpected to have a considerable social impact, but afew large clinical trials enrolling the appropriate patientsare now necessary to assess their efficacy.

This review will end considering their use not in thetreatment of cancer patients but to prevent cancer inhealthy persons, a so far neglected prospect. Currentstudies are leading to the detection of gene mutationsthat predispose to cancer.274 Identification of the geneat risk and its mutated or amplified products would pro-vide the opportunity to vaccinate susceptible subjectsagainst their foreseeable cancer. Molecular characteri-zation of altered gene products predictably destined tobecome a tumor antigen will be the first step towardsthe engineering of effective vaccines to be used for thispurpose.25

An unrestrained imagination may picture an evenbroader application of antitumor vaccines, i.e. their useto prevent tumors in the general population. Molecularand genetic data suggest that the number of TAA is notendless. Several of the TAA detected so far are shared byhistologically distinct tumors arising in different organs(Table 1). The possibility of vaccinating against mostcommon human cancers by using not many more than

twenty TAA may perhaps be conceivable. Experimentaldata suggest that the immunity elicited by specific vac-cination is much more effective in the inhibition ofincipient tumors than in the cure of those that havealready progressed.275 The risk of inducing an autoim-mune disease would be a major worry since not rarelyantigens acting as TAA are expressed by normal tis-sues.276 This risk would be much harder to accept whentreating healthy individuals than in the vaccination ofcancer patients, where the scales of risk-benefit arebiased by a short life-expectancy. On the other handfailure to intervene when a disease so diffuse and dra-matic as cancer can be prevented could also be viewedas harmful.277 Lastly, it should be considered that thesame or even a higher risk of inducing autoimmunereactions is associated with many antimicrobial vac-cines. Fortunately, they started to be used when this riskwas not yet perceived.

In conclusion, even if cancer vaccines are an olddream,278 only recently has their design become a ratio-nal enterprise. There are now many ways of construct-ing vaccines able to elicit a strong protective immuni-ty. This progress is offering ground for optimism.

Contributions and AcknowledgmentsThe authors were a group of experts and representa-

tives of two pharmaceutical companies, Amgen ItaliaSpa and Dompé Biotech Spa, both from Milan, Italy. Thisco-operation between a medical journal and pharma-ceutical companies is based on the common aim ofachieving optimal use of new therapeutic procedures inmedical practice.

FundingThe preparation of this manuscript was supported by

educational grants from Dompé Biotech Spa and AmgenItalia Spa.

DisclosuresConflict on interest: Dompé Biotec Spa sells G-CSF and

rHuEpo in Italy, and Amgen Italia Spa has a stake in Dom-pé Biotec Spa.

Redundant publications: no substantial overlappingwith previous papers.

Manuscript processingManuscript received August 7, 2000; accepted October

24, 2000.

References

1. Bordignon C, Carlo-Stella C, Colombo MP, et al. Celltherapy: achievements and perspectives. Haemato-logica 1999; 84:1110-49.

2. Ada GL. The immunological principles of vaccination.Lancet 1990; 335:523-6.

3. Sprent J. Lifespan of naive, memory and effector lym-phocytes. Curr Opin Immunol 1993; 5:433-8.

4. Rabinovich NR, McInnes P, Klein DL, Hall BF. Vaccinetechnologies: view to the future. Science 1994;265:1401-4.

5. Medzhitov R, Janeway CA Jr. Innate immunity: thevirtues of a nonclonal system of recognition. Cell


180


1997; 91:295-8.6. Matzinger P. Tolerance, danger, and the extended

family. Annu Rev Immunol 1994; 12:991-1045.7. Musiani P, Modesti A, Giovarelli M, et al. Cytokines,

tumor-cell death and immunogenicity: a question ofchoice. Immunol Today 1997; 18:32-6.

8. Pardoll DM. Cancer vaccines: a road map for the nextdecade. Curr Opin Immunol 1996; 8:619-20.

9. Ehrlich P. In: Collected papers of Paul Ehrlich. Lon-don: Pergamon Press; 1960.

10. Snell GD. The histocompatibility systems. TransplantProc 1971; 3:1133-8.

11. Gross L. Intradermal immunization of C3H miceagainst a sarcoma that originated in an animal of thesame line. Cancer Res 1963; 3:326-33.

12. Klein G, Sjogren HO, Klein E. Demonstration of resis-tance against methylcholanthrene-induced sarcomasin the primary autochthonous host. Cancer Res 1960;20:1561-72.

13. Jaffe EM, Pardoll DM. Murine tumor antigens: is itworth the search? Curr Opin Immunol 1996; 8:622-7.

14. Hewitt HB, Blake ER, Walder AS. A critique of the evi-dence for active host defence against cancer, based onpersonal studies of 27 murine tumors of spontaneousorigin. Br J Cancer 1976; 33:241-59.

15. Boon T. Antigenic tumor cell variants obtained withmutagens. Adv Cancer Res 1983; 39:121-51.

16. De Plaen E, Lurquin C, Van Pel A, et al. Immunogenic(tum-) variants of mouse tumor P185: cloning thegene of tum- antigen P91A and identification of thetum- mutation. Proc Natl Acad Sci USA 1988;85:2274-8.

17. Traversari C, van der Bruggen P, Luescher IF, et al. Anonapeptide encoded by human gene MAGE-1 is rec-ognized on HLA-A1 by cytolytic T lymphocytes direct-ed against tumor antigen MZ2-E. J Exp Med 1992;176:1453-7.

18. Coulie PG. Human tumour antigens recognized by Tcells: new perspectives for anti-cancer vaccines? MolMed Today 1997; 3:261-8.

19. Robbins PF, Kawakami Y. Human tumor antigens rec-ognized by T cells. Curr Opin Immunol 1996; 8:628-35.

20. Nossal GJV. The case history of Mr. T.I. - terminalpatient or still curable. Immunol Today 1980; 1:5-6.

21. Vonderheide RH, Hahn WC, Schultze JL, Nadler LM.The telomerase catalytic subunit is a widely expressedtumor-associated antigen recognized by cytotoxic Tlymphocytes. Immunity 1999; 10:673-9.


23. Colombo MP, Forni G. Immunotherapy I: Cytokinegene transfer strategies. Cancer Metast Rev 1996;15:317-28.

24. Immunity future, Science, http://www.aaas.org/sci-ence/immunology.

25. Lollini PL, Forni G. Specific and nonspecific immuni-ty in the prevention of spontaneous tumours.Immunol Today 1999; 20:347-50.

26. Nanni P, Forni G, Lollini PL. Cytokine gene therapy:hopes and pitfalls. Ann Oncol 1999; 10:261-6.

27. Colombo MP, Forni G. Cytokine gene transfer intumor inhibition and tumor therapy: where are wenow ? Immunol Today 1994; 15:48-51.

28. Allione A, Consalvo M, Nanni P, et al. Immunizingand curative potential of replicating and nonreplicat-ing murine mammary adenocarcinoma cells engi-neered with interleukin (IL)-2, IL-4, IL6, IL-7, IL-10,tumor necrosis factor α, granulocyte-macrophagecolony-stimulating factor, and γ-interferon gene oradmixed with conventional adjuvants. Cancer Res

1994; 54:6022-6.29. Cavallo F, Signorelli P, Giovarelli M, et al. Antitumor

efficacy of adenocarcinoma cells engineered to pro-duce interleukin 12 (IL-12) or other cytokines com-pared with exogenous IL-12. J Natl Cancer I 1997;89:1049-58.

30. Cavallo F, Di Carlo E, Butera M, et al. Immune eventsassociated with the cure of established tumors andspontaneous metastases by local and systemic inter-leukin 12. Cancer Res 1999; 59:414-21.


32. Banchereau J, Steinman RM. Dendritic cells and thecontrol of immunity. Nature 1998; 392:245-52.

33. Kitamura H, Iwakabe K, Yahata T, et al. The naturalkiller T (NKT) cell ligand alpha-galactosylceramidedemonstrates its immunopotentiating effect by induc-ing interleukin (IL)-12 production by dendritic cellsand IL-12 receptor expression on NKT cells. J Exp Med1999; 189:1121-8.

34. Cayeux S, Richter G, Becker C, Pezzutto A, Dorken B,Blankenstein T. Direct and indirect T cell priming bydendritic cell vaccines. Eur J Immunol 1999; 29:225-34.

35. Sallusto F, Lanzavecchia A, Mackay CR. Chemokinesand chemokine receptors in T-cell priming andTh1/Th2-mediated responses. Immunol Today 1998;19:568-74.

36. Forni G, Cavallo F, Consalvo M, et al. Molecularapproaches to cancer immunotherapy. Cytokines MolTher 1995; 1:225-48.

37. Peron JM, Esche C, Subbotin VM, Maliszewski C,Lotze MT, Shurin MR. FLT3-ligand administrationinhibits liver metastases: role of NK cells. J Immunol1998; 161:6164-70.

38. Simons JW, Jaffee EM, Weber CE, et al. Bioactivity ofautologous irradiated renal cell carcinoma vaccinesgenerated by ex vivo granulocyte-macrophage colony-stimulating factor gene transfer. Cancer Res 1997;57:1537-46.

39. Soiffer R, Lynch T, Mihm M, et al. Vaccination withirradiated autologous melanoma cells engineered tosecrete human granulocyte-macrophage colony-stim-ulating factor generates potent antitumor immunity inpatients with metastatic melanoma. Proc Natl AcadSci USA 1998; 95:13141-6.

40. Sun Y, Jurgovsky K, Moller P, et al. Vaccination with IL-12 gene-modified autologous melanoma cells: pre-clinical results and a first clinical phase I study. GeneTher 1998; 5:481-90.

41. Kirk AD, Burkly LC, Batty DS, et al. Treatment withhumanized monoclonal antibody against CD154 pre-vents acute renal allograft rejection in nonhuman pri-mates. Nat Med 1999; 5:686-93.

42. Larsen CP, Elwood ET, Alexander DZ, et al. Long-termacceptance of skin and cardiac allografts after block-ing CD40 and CD28 pathways. Nature 1996; 381:434-8.

43. Blazar BR, Taylor PA, Noelle RJ, Vallera DA. CD4(+)T cells tolerized ex vivo to host alloantigen by anti-CD40 ligand (CD40L:CD154) antibody lose theirgraft-versus-host disease lethality capacity but retainnominal antigen responses. J Clin Invest 1998;102:473-82.

44. Diehl L, den Boer AT, Schoenberger SP, van der VoortEI, Schumacher TN, Melief CJ. CD40 activation in vivoovercomes peptide-induced peripheral cytotoxic T-lymphocyte tolerance and augments anti-tumor vac-cine efficacy. Nat Med 1999; 5:774-9.

45. Sotomayor EM, Borrello I, Tubb E, et al. Conversionof tumor-specific CD4+ T-cell tolerance to T-cell prim-ing through in vivo ligation of CD40. Nat Med 1999;

M. Bocchia et al.

181


5:780-7.46. Kato K, Cantwell MJ, Sharma S, Kipps TJ. Gene trans-

fer of CD40-ligand induces autologous immunerecognition of chronic lymphocytic leukemia B cells. JClin Invest 1998; 5:1133-41.

47. Leach DR, Krummel MF, Allison JP. Enhancement ofantitumor immunity by CTLA-4 blockade. Science1996; 271:1734-6.

48. Yang YF, Zou JP, Mu J, et al. Enhanced induction ofantitumor T-cell responses by cytotoxic T lymphocyte-associated molecule-4 blockade: the effect is mani-fested only at the restricted tumor-bearing stages.Cancer Res 1997; 57:4036-41.

49. van Elsas A, Hurwitz AA, Allison JP. Combinationimmunotherapy of B16 melanoma using anti-cyto-toxic T lymphocyte-associated antigen 4 (CTLA-4)and granulocyte/macrophage colony-stimulating fac-tor (GM-CSF)-producing vaccines induces rejectionof subcutaneous and metastatic tumors accompaniedby autoimmune depigmentation. J Exp Med 1999;190:355-66.

50. Hurwitz AA, Yu TF, Leach DR, Allison JP. CTLA-4blockade synergizes with tumor-derived granulocyte-macrophage colony-stimulating factor for treatmentof an experimental mammary carcinoma. Proc NatlAcad Sci USA 1998; 95:10067-71.

51. Young JW. Dendritic cells: expansion and differentia-tion with hematopoietic growth factors. Curr OpinHematol 1999; 6:135-44.

52. Rescigno M, Granucci F, Citterio S, Foti M, Ricciardi-Castagnoli P. Coordinated events during bacteria-induced DC maturation. Immunol Today 1999;20:200-3.

53. Kapsenberg ML, Hilkens CM, Wierenga EA, Kalinski P.The paradigm of type 1 and type 2 antigen presentingcells. Implications for atopic allergy. Clin Exp Allergy1999; 29 (Suppl) 2:33-6.

54. Cella M, Scheidegger D, Palmer-Lehmann K, Lane P,Lanzavecchia A, Alber G. Ligation of CD40 on den-dritic cells triggers production of high levels of inter-leukin-12 and enhances T cell stimulatory capacity:T-T help via APC activation. J Exp Med 1996;184:747-52.

55. Rissoan MC, Soumelis V, Kadowaki N, et al. Recipro-cal control of T helper cell and dendritic cell dif-ferentation. Science 1999; 283:1183-6.

56. Wong BR, Josien R, Lee SY, et al. TRANCE (tumornecrosis factor [TNF]-related activation-inducedcytokine), a new TNF family member predominantlyexpressed in T cells, is a dendritic cell-specific survivalfactor. J Exp Med 1997; 186:2075-80.

57. Anderson DM, Maraskovsky E, Billingsley WL, et al. Ahomologue of the TNF receptor and its ligandenhance T-cell growth and dendritic-cell function.Nature 1997; 390:175-9.

58. Dubois B, Massacrier C, Vanbervliet B, et al. Criticalrole of IL-12 in dendritic cell-induced differentiation ofnaive B lymphocytes. J Immunol 1998;161:2223-31.

59. Caux C, Liu YJ, Banchereau J. Recent advances in thestudy of dendritic cells and follicular dendritic cells.Immunol Today 1995; 16:2-4.

60. Cao L, Kulmburg P, Veelken H, et al. Cytokine genetransfer in cancer therapy. Stem Cells 1998; 16(Sup-pl 1):251-60.

61. Colombo MP, Lombardi L, Melani C, et al. Hypoxictumor cell death and modulation of endothelial adhe-sion molecules in the regression of granulocyte colony-stimulating factor-transduced tumors. Am J Pathol1996; 148:473-83.

62. Morel S, Lévy F, Burlet-Schiltz O, et al. Processing ofsome antigens by the standard proteasome but not bythe immunoproteasome results in poor presentation

by dendritic cells. Immunity 2000; 12:107-17.63. Groettrup M, Schmidtke G. Selective proteasome

inhibitors: modulators of antigen presentation ? DrugDiscov Today 1999; 4:63-71.

64. Morton DL, Foshag LJ, Hoon DS, et al. Prolongationof survival in metastatic melanoma after active specificimmunotherapy with a new polyvalent melanoma vac-cine. Ann Surg 1992; 216:463-82.

65. Livingston PO, Wong GY, Adluri S, et al. Improvedsurvival in stage III melanoma patients with GM2 anti-bodies: a randomized trial of adjuvant vaccinationwith GM2 ganglioside. J Clin Oncol 1994; 12:1036-44.

66. Berd D, Maguire HC Jr, Schuchter LM, et al. Autolo-gous hapten-modified melanoma vaccine as postsur-gical adjuvant treatment after resection of nodalmetastases. J Clin Oncol 1997; 15:2359-70.

67. Sensi M, Farina C, Maccalli C, et al. Clonal expansionof T lymphocytes in human melanoma metastasesafter treatment with a hapten-modified autologoustumor vaccine. J Clin Invest 1997; 99:710-7.

68. Boon T, van der Bruggen P. Human tumor antigensrecognized by T lymphocytes. J Exp Med 1996; 183:725-9.

69. Huang YC, Golumbeck P, Ahmadzadeh M, Jaffee E,Pardoll D, Levitsky H. Role of bone marrow-derivedcells in presenting MHC class I-restricted tumor anti-gens. Science 1994; 264:961-5.

70. Fujiwara H, Aoki H, Yoshioka T, Tomita S, Ikegami R,Hamaoka T. Establishment of a tumor-specificimmunotherapy model utilizing TNP-reactive helperT cell activity and its application to the autochtho-nous tumor system. J Immunol 1984; 133:509-14.

71. Sallusto F, Lanzavecchia A. Efficient presentation ofsoluble antigen by cultured human dendritic cells ismaintained by granulocyte/macrophage colony-stim-ulating factor plus interleukin 4 and downregulated bytumor necrosis factor α. J Exp Med 1994; 179:1109-18.

72. van Kooten C, Banchereau J. Functions of CD40 on Bcells, dendritic cells and other cells. Curr OpinImmunol 1997; 9:330-7.

73. Chiodoni C, Paglia P, Stoppacciaro A, Rodolfo M,Parenza M, Colombo MP. Dendritic cells infiltratingtumors cotransduced with granulocyte-macrophagecolony-stimulating factor (GM-CSF) and CD40 lig-and genes take up and present endogenous tumor-associated antigens, and prime naive mice for a cyto-toxic T lymphocyte response. J Exp Med 1999;190:125-33.

74. Albert ML, Bhardwaj N. Resurrecting the dead: DCscross-present antigen derived from apoptotic cells onMHC I. The Immunologist 1998; 5:194-8.

75. Melcher A, Todryk S, Hardwick N, Ford M, JacobsonM, Vile RG. Tumor immunogenicity is determined bythe mechanism of cell death via induction of heatshock protein expression. Nat Med 1998; 4:581-7.

76. Srivastava PK, Menoret A, Basu S, Binder RJ,McQuade KL. Heat shock proteins come of age: prim-itive functions acquire new roles in an adaptive world.Immunity 1998; 8:657-65.

77. Bergenbrant S, Yi Q, Osterborg A, Bjorkholm M, et al.Modulation of anti-idiotypic immune response byimmunization with the autologous M-componentprotein in multiple myeloma patients. Br J Haematol1996; 92:840-6.

78. Adluri S, Gilewski T, Zhang S, Ramnath V, RagupathiG, Livingston P. Specificity analysis of sera from breastcancer patients vaccinated with MUC1-KLH plus QS-21. Br J Cancer 1999; 79:1806-12.

79. Kwak LW, Pennington R, Boni L, Ochoa AC, RobbRJ, Popescu MC. Liposomal formulation of a self lym-


182


phoma antigen induces protective antitumor immu-nity. J Immunol 1998; 160:3637-41.

80. Kwak LW, Campbell MJ, Czerwinski DK, Hart S, MillerRA, Levy R. Induction of immune responses in patientswith B-cell lymphoma against the surface-immunoglobulin idiotype expressed by their tumors.N Engl J Med 1992; 327:1209-15.

81. Byars NE, Allison AC. Adjuvant formulation for use invaccines to elicit both cell-mediated and humoralimmunity. Vaccine 1987; 5:223-8.

82. Hsu FJ, Caspar CB, Czerwinski D, et al. Tumor-specificidiotype vaccines in the treatment of patients with B-cell lymphoma: long-term results of a clinical trial.Blood 1997; 89:3129-35.

83. Johnson AG, Tomai M, Solem L, Beck L, Ribi E. Char-acterization of a nontoxic monophosphoryl lipid A.Rev Infect Dis 1987; 9:S512-6.

84. Maitre N, Brown JM, Demcheva M, et al. Primary T-cell and activated macrophage response associatedwith tumor protection using peptide/poly-N-acetylglucosamine vaccination. Clin Cancer Res 1999;5:1173-82.

85. Yewdell JW, Bennink JR. Cell biology of antigen pro-cessing and presentation to major histocompatibilitycomplex class I molecule-restricted T-lymphocytes.Adv Immunol 1992; 52:1-123.

86. Unanue ER, Cerottini JC. Antigen presentation. FASEBJ 1989; 3:2496-502.

87. Howard JC. Supply and transport of peptides pre-sented by class I MHC molecules. Curr Opin Immunol1995; 7:69-76.

88. Rammensee HG. Chemistry of peptides associatedwith MHC class I and class II molecules. Curr OpinImmunol 1995; 7:85-96.

89. Braciale TJ, Braciale VL. Antigen presentation: struc-tural themes and functional variations. ImmunolToday 1991; 12:124-9.

90. Germain RN. MHC-dependent antigen processingand peptide presentation: providing ligands for T lym-phocyte activation. Cell 1994; 76:287-99.

91. Accolla RS, Adorini L, Santoris S, Sinigaglia F, Guardi-ola J. MHC: orchestrating the immune response.Immunol Today 1995; 16:8-11.

92. Rammensee HG, Friede T, Stevanovic S. MHC ligandsand peptide motifs: first listing. Immunogenetics1995; 1:178-228.

93. Falk K, Rotzschke O, Stevanovic S, Jung G, Ram-mensee HG. Allele-specific motifs revealed by sequenc-ing of self-peptides eluted from MHC molecules.Nature 1991; 351:290-6.

94. Stuber G, Modrow S, Hoglund P, et al. Assessment ofmajor histocompatibility complex class I interactionwith Epstein-Barr virus and human immunodeficien-cy virus peptides by elevation of membrane H-2 andHLA in peptide loading-deficient cells. Eur J Immunol1992; 22:2697-703.

95. Sette A, Vitiello A, Reherman B, et al. The relationshipbetween class I binding affinity and the immuno-genicity of potential cytotoxic T cell epitopes. JImmunol 1994; 153:5586-92.

96. Kawakami Y, Eliyahu S, Sakaguchi K, et al. Identifica-tion of the immunodominant peptides of the MART-1 human melanoma antigen recognized by the major-ity of HLA-A2-restricted tumor infiltrating lympho-cytes. J Exp Med 1994; 180:347-52.

97. Parkhurst MR, Salgaller ML, Southwood S, et al.Improved induction of melanoma-reactive CTL withpeptides from the melanoma antigen gp100 modifiedat HLA-A*0201-binding residues. J Immunol 1996;157:2539-48.

98. Valmori D, Gervois N, Rimoldi D, et al. Diversity of thefine specificity displayed by HLA-A*0201-restricted

CTL specific for the immunodominant Melan-A/MART-1 antigenic peptide. J Immunol 1998;161:6956-62.

99. Clay TM, Custer MC, McKee MD, et al. Changes inthe fine specificity of gp100(209-217)-reactive T cellsin patients following vaccination with a peptide mod-ified at an HLA-A2.1 anchor residue. J Immunol 1999;162:1749-55.

100. Khleif SN, Abrams SI, Hamilton JM, et al. A phase Ivaccine trial with peptides reflecting ras oncogenemutations of solid tumors. J Immunother 1999;22:155-65.

101. Manici S, Sturniolo T, Imro MA, et al. Melanoma cellspresent a MAGE-3 epitope to CD4(+) cytotoxic T cellsin association with histocompatibility leukocyte anti-gen DR11. J Exp Med 1999; 189:871-6.

102. Correale P, Walmsley K, Nieroda C, et al. In vitro gen-eration of human cytotoxic T lymphocytes specific forpeptides derived from prostate-specific antigen. J NatlCancer I 1997; 89:293-300.

103. Bocchia M, Wentworth PA, Southwood S, et al. Spe-cific binding of leukemia oncogene fusion protein pep-tides to HLA class I molecules. Blood 1995; 85:2680-4.

104. Greco G, Fruci D, Accapezzato D, et al. Two bcr-abljunction peptides bind to HLA-A3 molecules andallow specific induction of human cytotoxic T lym-phocytes. Leukemia 1996; 10:693-9.

105. Bocchia M, Korontsvit T, Xu Q, et al. Specific humancellular immunity to BCR-ABL oncogene-derived pep-tides. Blood 1996; 87:3587-92.

106. Bosh GJ, Joosten AM, Kessler JH, Melief CJ, LeeksmaOC. Recognition of BCR-ABL positive leukemia blastsby human CD4+ T cells elicited by primary in vitroimmunization with a BCR-ABL breakpoint peptide.Blood 1996; 88:3522-7.

107. Mannering SI, McKenzie JL, Fearnley DB, Hart DN.HLA-DR1-restricted bcr-abl (b3a2)-specific CD4+ Tlymphocytes respond to dendritic cells pulsed withb3a2 peptide and antigen-presenting cells exposed tob3a2 containing cell lysates. Blood 1997; 90:290-7.

108. Demotz S, Grey HM, Appella E, Sette A. Characteri-zation of a naturally processed MHC class II-restrict-ed T-cell determinant of hen egg lysozyme. Nature1989; 342:682-4.

109. Storkus WJ, Zeh HD, Salter RD, Lotze MT. Identifica-tion of T-cell epitopes: rapid isolation of class-I pre-sented peptides from viable cells by mild acid elution.J Immunother 1993; 14:94-103.

110. Papadopoulos KP, Hesdorffer CS, Suciu-Foca N, Hib-shoosh H, Harris PE. Wild-type p53 epitope natural-ly processed and presented by an HLA-B haplotypeon human breast carcinoma cells. Clin Cancer Res1999; 5:2089-93.

111. Skipper JC, Gulden PH, Hendrickson RC, et al. Mass-spectometric evaluation of HLA-A*0201-associatedpeptides identifies dominant naturally processedforms of CTL epitopes from MART-1 and gp100. IntJ Cancer 1999; 82:669-77.

112. Ostankovitch M, Buzyn A, Bonhomme D, et al.Antileukemic HLA-restricted T-cell clones generatedwith naturally processed peptides eluted from acutemyeloblastic leukemia blasts. Blood 1998; 92:19-24.

113. Papadopoulos KP, Suciu-Foca N, Hesdorffer CS, etal. Naturally processed tissue-and differentiationstage-specific autologous peptides bound by HLAclass I and II molecules of chronic myeloid leukemiablasts. Blood 1997; 90:4938-46.

114. Bellone M, Iezzi G, Imro MA, Protti MP. Cancerimmunotherapy: synthetic and natural peptides in thebalance. Immunol Today 1999; 20:457-62.

115. Kawakami Y, Nishimura MI, Restifo NP, et al. T-cell

M. Bocchia et al.

183


recognition of human melanoma antigens. JImmunother 1993; 14:88-93.

116. van der Bruggen P, Traversari C, Chomez P, et al. Agene encoding an antigen recognized by cytolytic Tlymphocytes on a human melanoma. Science 1991;254:1643-7.

117. Kawakami Y, Eliyahu S, Jennings C, et al. Recognitionof multiple epitopes in the human melanoma antigengp100 by tumor-infiltrating T lymphocytes associatedwith in vivo tumor regression. J Immunol 1995;154:3961-8.

118. Kang X, Kawakami Y, el-Gamil M, et al. Identificationof a tyrosinase epitope recognized by HLA-A24-restricted, tumor-infiltrating lymphocytes. J Immunol1995; 155:1343-8.

119. Wang RF, Wang X, Rosenberg SA. Identification of anovel major histocompatibility complex class II-restricted tumor antigen resulting from a chromoso-mal rearrangement recognized by CD4+ T cells. J ExpMed 1999; 189:1659-68.

120. Pawelec G, Max H, Halder T, et al. BCR/ABL leukemiaoncogene fusion peptides selectively bind to certainHLA-DR alleles and can be recognized by T cells foundat low frequency in the repertoire of normal donors.Blood 1996; 88:2118-24.

121. Yotnda P, Firat H, Garcia-Pons F, et al. Cytotoxic T cellresponse against the chimeric p210 BCR-ABL proteinin patients with chronic myelogenous leukemia. J ClinInvest 1998; 101:2290-6.

122. Goydos JS, Elder E, Whiteside TL, Finn OJ, Lotze MT.A phase I trial of a synthetic mucin peptide vaccine.Induction of specific immune reactivity in patientswith adenocarcinoma. J Surg Res 1996; 63:298-304.

123. Marchand M, van Baren N, Weynants P, et al. Tumorregressions observed in patients with metastaticmelanoma treated with an antigenic peptide encodedby gene MAGE-3 and presented by HLA-A1. Int J Can-cer 1999; 80:219-30.

124. Pinilla-Ibarz J, Cathcart K, Korontsvit T, et al. Vacci-nation of patients with chronic myelogenous leukemiawith bcr-abl oncogene breakpoint fusion peptidesgenerates specific immune responses. Blood 2000; 95:1781-7.

125. Steller MA, Gurski KJ, Murakami M, et al. Cell-medi-ated immunological responses in cervical and vaginalcancer patients immunized with a lipidated epitope ofhuman papillomavirus type 16 E7. Clin Cancer Res1998; 4:2103-9.

126. Mukherji B, Chakraborty NG, Yamasaki S, et al. Induc-tion of antigen-specific cytolytic T cells in situ inhuman melanoma by immunization with syntheticpeptide-pulsed autologous antigen presenting cells.Proc Natl Acad Sci USA 1995; 92:8078-82.

127. Shirai M, Pendleton CD, Ahlers J, et al. Helper-cyto-toxic T lymphocytes (CTL) determinant linkagerequired for priming of anti-HIV CD8+CTL in vivo withpeptide vaccine constructs. J Immunol 1994;152:549-56.

128. Bachmann MF, Zinkernagel RM, Oxenius A. Immuneresponses in the absence of costimulation: virusesknow the trick. J Immunol 1998; 161:5791-4.

129. Moss B, Carroll MW, Wyatt LS, et al. Host rangerestricted, non-replicating vaccinia virus vectors asvaccine candidates. Adv Exp Med Biol 1996; 397:7-13.

130. Hodge JW, Schlom J, Donohue SJ, et al. A recombi-nant vaccinia virus expressing human prostate-specif-ic antigen (PSA): safety and immunogenicity in a non-human primate. Int J Cancer 1995; 63:231-7.

131. Bronte V, Tsung K, Rao JB, et al. IL-2 enhances thefunction of recombinant poxvirus-based vaccines inthe treatment of established pulmonary metastases. J

Immunol 1995; 154:5282-92.132. Kantor J, Irvine K, Abrams S, Kaufman H, DiPietro J,

Schlom J. Antitumor activity and immune responsesinduced by a recombinant carcinoembryonic antigen-vaccinia virus vaccine. J Natl Cancer I 1992; 84:1084-91.

133. Bronte V, Carroll MW, Goletz TJ, et al. Antigen expres-sion by dendritic cells correlates with the therapeuticeffectiveness of a model recombinant poxvirus tumorvaccine. Proc Natl Acad Sci USA 1997; 94:3183-8.

134. Chamberlain RS, Carroll MW, Bronte V, et al. Cos-timulation enhances the active immunotherapy effectof recombinant anticancer vaccines. Cancer Res 1996;56:2832-6.

135. Rao JB, Chamberlain RS, Bronte V, et al. IL-12 is aneffective adjuvant to recombinant vaccinia virus-based tumor vaccines: enhancement by simultaneousB7-1 expression. J Immunol 1996; 156:3357-65.

136. Carroll MW, Overwijk WW, Surman DR, Tsung K,Moss B, Restifo NP. Construction and characteriza-tion of a triple-recombinant vaccinia virus encodingB7-1, interleukin 12, and a model tumor antigen. JNatl Cancer I 1998; 90:1881-7.

137. Irvine KR, Parkhurst MR, Shulman EP, et al. Recom-binant virus vaccination against "self" antigens usinganchor-fixed immunogens. Cancer Res 1999; 59:2536-40.

138. Overwijk WW, Lee DS, Surman DR, et al. Vaccinationwith a recombinant vaccinia virus encoding a "self"antigen induces autoimmune vitiligo and tumor celldestruction in mice: requirement for CD4(+) T lym-phocytes. Proc Natl Acad Sci USA 1999; 96:2982-7.

139. Tsang KY, Zaremba S, Nieroda CA, Zhu MZ, Hamil-ton JM, Schlom J. Generation of human cytotoxic Tcells specific for human carcinoembryonic antigen epi-topes from patients immunized with recombinant vac-cinia-CEA vaccine. J Natl Cancer I 1995; 87:982-90.

140. Chen PW, Wang M, Bronte V, Zhai Y, Rosenberg SA,Restifo NP. Therapeutic antitumor response afterimmunization with a recombinant adenovirus encod-ing a model tumor-associated antigen. J Immunol1996; 156:224-31.

141. O'Neal WK, Zhou H, Morral N, et al. Toxicologicalcomparison of E2a-deleted and first-generation ade-noviral vectors expressing alpha1-antitrypsin after sys-temic delivery. Hum Gene Ther 1998; 9:1587-98.

142. Rosenberg SA, Zhai Y, Yang JC, et al. Immunizingpatients with metastatic melanoma using recombi-nant adenoviruses encoding MART-1 or gp100melanoma antigens. J Natl Cancer I 1998; 90:1894-900.

143. Baxby D, Paoletti E. Potential use of non-replicatingvectors as recombinant vaccines. Vaccine 1992; 10: 8-9.

144. Wang M, Bronte V, Chen PW, et al. Active immuno-therapy of cancer with a nonreplicating recombinantfowlpox virus encoding a model tumor-associatedantigen. J Immunol 1995; 154:4685-92.

145. Cox WI, Tartaglia J, Paoletti E. Induction of cytotoxicT lymphocytes by recombinant canarypox (ALVAC)and attenuated vaccinia (NYVAC) viruses expressingthe HIV-1 envelope glycoprotein. Virology 1993; 195:845-50.

146. Perkus ME, Tartaglia J, Paoletti E. Poxvirus-based vac-cine candidates for cancer, AIDS, and other infectiousdiseases. J Leukocyte Biol 1995; 58:1-13.

147. Mahnel H, Mayr A. Experiences with immunizationagainst orthopox viruses of humans and animals usingvaccine strain MVA. Berl Munch Tierarztl 1994;107:253-6.

148. Carroll MW, Overwijk WW, Chamberlain RS, Rosen-berg SA, Moss B, Restifo NP. Highly attenuated mod-


184


ified vaccinia virus Ankara (MVA) as an effectiverecombinant vector: a murine tumor model. Vaccine1997; 15:387-94.

149. Schneider J, Gilbert SC, Blanchard TJ, et al. Enhancedimmunogenicity for CD8+ T cell induction and com-plete protective efficacy of malaria DNA vaccinationby boosting with modified vaccinia virus Ankara. NatMed 1998; 4:397-402.

150. Belyakov IM, Moss B, Strober W, Berzofsky JA.Mucosal vaccination overcomes the barrier to recom-binant vaccinia immunization caused by preexistingpoxvirus immunity. Proc Natl Acad Sci USA 1999;96:4512-7.

151. Brossart P, Goldrath AW, Butz EA, Martin S, BevanMJ. Virus-mediated delivery of antigenic epitopes intodendritic cells as a means to induce CTL. J Immunol1997; 158:3270-6.

152. Kaplan JM, Yu Q, Piraino ST, et al. Induction of anti-tumor immunity with dendritic cells transduced withadenovirus vector-encoding endogenous tumor-asso-ciated antigens. J Immunol 1999; 163:699-707.

153. Di Nicola M, Siena S, Bregni M, et al. Gene transferinto human dendritic antigen-presenting cells by vac-cinia virus and adenovirus vectors. Cancer Gene Ther1998; 5:350-6.

154. Ulmer JB, Donnelly JJ, Parker SE, et al. Heterologousprotection against influenza by injection of DNAencoding a viral protein. Science 1993; 259:1745-9.

155. Donnelly JJ, Ulmer JB, Shiver JW, Liu MA. DNA vac-cines. Annu Rev Immunol 1997; 15:617-48.

156. Wang R, Doolan DL, Le TP, et al. Induction of anti-gen-specific cytotoxic T lymphocytes in humans by amalaria DNA vaccine. Science 1998; 282:476-80.

157. Calarota S, Bratt G, Nordlund S, et al. Cellular cyto-toxic response induced by DNA vaccination in HIV-1-infected patients. Lancet 1998; 351:1320-5.

158. Medina E, Guzman CA, Staendner LH, Colombo MP,Paglia P. Salmonella vaccine carrier strains: effectivedelivery system to trigger anti-tumor immunity by oralroute. Eur J Immunol 1999; 29:693-9.

159. Paglia P, Medina E, Arioli I, Guzman CA, ColomboMP. Gene transfer in dendritic cells, induced by oralDNA vaccination with Salmonella typhimurium,results in protective immunity against a murinefibrosarcoma. Blood 1998; 92:3172-6.

160. Ikonomidis G, Paterson Y, Kos FJ, Portnoy DA. Deliv-ery of a viral antigen to the class I processing and pre-sentation pathway by Listeria monocytogenes. J ExpMed 1994; 180:2209-18.

161. Anwer K, Earle KA, Shi M, et al. Synergistic effect offormulated plasmid and needle-free injection forgenetic vaccines. Pharm Res 1999; 16:889-95.

162. Pertmer TM, Roberts TR, Haynes JR. Influenza virusnucleoprotein-specific immunoglobulin G subclassand cytokine responses elicited by DNA vaccinationare dependent on the route of vector DNA delivery. JVirol 1996; 70:6119-25.

163. Feltquate DM, Heaney S, Webster RG, Robinson HL.Different T helper cell types and antibody isotypes gen-erated by saline and gene gun DNA immunization. JImmunol 1997; 158:2278-84.

164. Krieg AM, Yi AK, Matson S, et al. CpG motifs in bac-terial DNA trigger direct B-cell activation. Nature1995; 374:546-9.

165. Klinman DM, Yi AK, Beaucage SL, Conover J, KriegAM. CpG motifs present in bacteria DNA rapidlyinduce lymphocytes to secrete interleukin 6, inter-leukin 12, and interferon gamma. Proc Natl Acad SciUSA 1996; 93:2879-83.

166. Halpern MD, Kurlander RJ, Pisetsky DS. BacterialDNA induces murine interferon-gamma productionby stimulation of interleukin-12 and tumor necrosis

factor-alpha. Cell Immunol 1996; 167:72-8.167. Klinman DM, Barnhart KM, Conover J. CpG motifs as

immune adjuvants. Vaccine 1999; 17:19-25.168. Sato Y, Roman M, Tighe H, et al. Immunostimulato-

ry DNA sequences necessary for effective intradermalgene immunization. Science 1996; 273:352-4.

169. Hartmann G, Weiner GJ, Krieg AM. CpG DNA: apotent signal for growth, activation, and maturationof human dendritic cells. Proc Natl Acad Sci USA1999; 96:9305-10.

170. Torres CA, Iwasaki A, Barber BH, Robinson HL. Dif-ferential dependence on target site tissue for gene gunand intramuscular DNA immunizations. J Immunol1997; 158:4529-32.

171. Iwasaki A, Torres CA, Ohashi PS, Robinson HL, Bar-ber BH. The dominant role of bone marrow-derivedcells in CTL induction following plasmid DNA immu-nization at different sites. J Immunol 1997; 159:11-4.

172. Porgador A, Irvine KR, Iwasaki A, Barber BH, RestifoNP, Germain RN. Predominant role for directly trans-fected dendritic cells in antigen presentation to CD8+T cells after gene gun immunization. J Exp Med 1998;188:1075-82.

173. Akbari O, Panjwani N, Garcia S, Tascon R, Lowrie D,Stockinger B. DNA vaccination: transfection and acti-vation of dendritic cells as key events for immunity. JExp Med 1999; 189:169-78.

174. Irvine KR, Rao JB, Rosenberg SA, Restifo NP. Cytokineenhancement of DNA immunization leads to effectivetreatment of established pulmonary metastases. JImmunol 1996; 156:238-45.

175. Tuting T, Gambotto A, DeLeo A, Lotze MT, RobbinsPD, Storkus WJ. Induction of tumor antigen-specificimmunity using plasmid DNA immunization in mice.Cancer Gene Ther 1999; 6:73-80.

176. Rosato A, Zambon A, Milan G, et al. CTL responseand protection against P815 tumor challenge in miceimmunized with DNA expressing the tumor-specificantigen P815A. Hum Gene Ther 1997; 8:1451-8.

177. Chen Y, Hu D, Eling DJ, Robbins J, Kipps TJ. DNA vac-cines encoding full-length or truncated Neu induceprotective immunity against Neu-expressing mamma-ry tumors. Cancer Res 1998; 58:1965-71.

178. Bowne WB, Srinivasan R, Wolchok JD, et al. Couplingand uncoupling of tumor immunity and autoimmu-nity. J Exp Med 1999; 190:1717-22.

179. Weber LW, Bowne WB, Wolchok JD, et al. Tumorimmunity and autoimmunity induced by immuniza-tion with homologous DNA. J Clin Invest 1998; 102:1258-64.

180. Bronte V, Apolloni E, Ronca R, et al. Genetic vacci-nation with "self" tyrosinase-related protein 2 causesmelanoma eradication but not vitiligo. Cancer Res2000; 60:253-8.

181. Liljestrom P. Alphavirus expression systems. Curr OpinBiotechnol 1994; 5:495-500.

182. Ying H, Zaks TZ, Wang RF, et al. Cancer therapy usinga self-replicating RNA vaccine. Nat Med 1999; 5:823-7.

183. O'Doherty U, Steinman RM, Peng M, et al. Dendriticcells freshly isolated from human blood express CD4and mature into typical immunostimulatory dendrit-ic cells after culture in monocyte-conditioned medi-um. J Exp Med 1993; 178:1067-76.

184. Zhou LJ, Schwarting R, Smith HM, Tedder TF. A nov-el cell-surface molecule expressed by human interdig-itating reticulum cells, Langerhans cells, and activat-ed lymphocytes is a new member of the IG superfam-ily. J Immunol 1992; 149:735-42.

185. Mosialos G, Birkenbach M, Ayehunie S, et al. Circu-lating human dendritic cells differentially express highlevels of a 55-kd actin-bundling protein. Am J Pathol

M. Bocchia et al.

185


1996; 148:593-600.186. Reid CD, Stackpoole A, Meager A, Tikerpae J. Inter-

actions of tumor necrosis factor with granulocyte-macrophage colony-stimulating factor and othercytokines in the regulation of dendritic cell growth invitro from early bipotent CD34+ progenitors inhuman bone marrow. J Immunol 1992; 149:2681-8.

187. Szabolcs P, Moore MA, Young JW. Expansion ofimmunostimulatory dendritic cells among the myeloidprogeny of human CD34+ bone marrow precursorscultured with c-kit ligand, granulocyte-macrophagecolony-stimulating factor, and TNF-alpha. J Immunol1995; 154:5851-61.

188. Caux C, Vanbervliet B, Massacrier C, et al. CD34+hematopoietic progenitors from human cord blooddifferentiate along two independent dendritic cellpathways in response to GM-CSF+TNF alpha. J ExpMed 1996; 184:695-706.


190. Staquet MJ, Jacquet C, Dezutter-Dambuyant C,Schmitt D. Fibronectin upregulates in vitro generationof dendritic Langerhans cells from human cord bloodCD34+ progenitors. J Invest Dermatol 1997; 109:738-43.

191. Maraskovsky E, Brasel K, Teepe M, et al. Drammaticincrease in the numberS of functionally mature den-dritic cells in Flt3 ligand-treated mice: multiple den-dritic cell subpopulations identified. J Exp Med 1996;184:1953-62.

192. Lebsack ME, Maraskovsky E, Roux E, et al. Increasedcirculating dendritic cells in healthy human volunteersfollowing administration of FLT3 ligand alone or incombination with GM-CSF or G-CSF. Blood 1998; 92(Suppl.1):2086a.

193. Lynch DH, Andreasen A, Maraskovsky E, Whitmore J,Miller RE, Schuh JC. Flt3 ligand induces tumor regres-sion and antitumor immune responses in vivo. NatMed 1997; 3:625-31.

194. Esche C, Subbotin VM, Maliszewski C, Lotze MT,Shurin MR. FLT3 ligand administration inhibits tumorgrowth in murine melanoma and lymphoma. CancerRes 1998; 58:380-3.

195. Nestle FO, Alijagic S, Gilliet M, et al. Vaccination ofmelanoma patients with peptide- or tumor lysate-pulsed dendritic cells. Nat Med 1998; 4:328-32.

196. Albert ML, Sauter B, Bhardwaj N. Dendritic cellsacquire antigen from apoptotic cells and induce classI-restricted CTLs. Nature 1998; 392:86-9.

197. Zitvogel L, Regnault A, Lozier A, et al. Eradication ofestablished murine tumors using a novel cell-free vac-cine: dendritic cell-derived exosomes. Nat Med 1998;4:594-600.

198. Boczkowski D, Nair SK, Snyder D, Gilboa E. Dendrit-ic cells pulsed with RNA are potent antigen-present-ing cells in vitro and in vivo. J Exp Med 1996; 184:465-72.

199. Reeves ME, Royal RE, Lam JS, Rosenberg SA, Hwu P.Retroviral transduction of human dendritic cells witha tumor-associated antigen gene. Cancer Res 1996;56:5672-7.

200. Dietz AB, Vuk-Pavlovic S. High efficiency adenovirus-mediated gene transfer to human dendritic cells.Blood 1998; 91:392-8.

201. Di Nicola M, Carlo-Stella C, Milanesi M, et al. Large-scale feasibility of gene transduction into humanCD34+ cell derived-dendritic cells by polication/ade-noviral complex. Br J Haematol 2000; (In Press).

202. Zhong L, Granelli-Piperno A, Choi Y, Steinman RM.Recombinant adenovirus is an efficient and non-per-

turbing genetic vector for human dendritic cells. Eur JImmunol 1999; 29:964-72.

203. Moss B. Vaccinia virus: a tool for research and vaccinedevelopment. Science 1991; 252:1662-7.

204. Choudhury BA, Liang JC, Thomas EK, et al. Dendrit-ic cells derived in vitro from acute myelogenousleukemia cells stimulate autologous, antileukemic T-cell responses. Blood 1999; 93:780-6.

205. Cignetti A, Bryant E, Allione B, Vitale A, Foa R, Cheev-er MA. CD34(+) acute myeloid and lymphoidleukemic blasts can be induced to differentiate intodendritic cells. Blood 1999; 94:2048-55.

206. Choudhury A, Gajewski JL, Liang JC, et al. Use ofleukemic dendritic cells for the generation ofantileukemic cellular cytotoxicity against Philadelphiachromosome-positive chronic myelogenous leukemia.Blood 1997; 89:1133-42.

207. Carlo-Stella C, Regazzi E, Garau D, et al. Ex vivo gen-eration of bcr/abl positive dendritic cells from CD34+chronic myeloid leukemia cells. Haematologica 1998;83 (Meeting Supplement):78-9.

208. Balch CM, Reintgen DS, Kirkwood JM, et al.Cuta-neous melanoma. Cancer Principles & Practice ofOncology, 5th Edition, De Vita VT, Hellman S, Rosen-berg SA (Eds.), Lippincot-Raven, 1997, p. 1947-94.

209. Linehan DC, Goedegebuure PS, Eberlein TJ. Vaccinetherapy for cancer. Ann Surg Oncol 1996; 3:219-28.

210. Arienti F, Sule-Suso J, Belli F, et al. Limited antitumorT cell response in melanoma patients vaccinated withinterleukin-2 gene-transduced allogeneic melanomacells. Hum Gene Ther 1996; 7:1955-63.

211. Moller P, Sun Y, Dorbic T, et al. Vaccination with IL-7 gene-modified autologous melanoma cells canenhance the anti-melanoma lytic activity in peripher-al blood of patients with a good clinical performancestatus: a clinical phase I study. Br J Cancer 1998;77:1907-16.

212. Jager E, Ringhoffer M, Dienes HP, et al. Granulocyte-macrophage-colony-stimulating factor enhancesimmune responses to melanoma-associated peptidesin vivo. Int J Cancer 1996; 67:54-62.

213. Rosenberg SA, Yang JC, Schwartzentruber DJ, et al.Immunologic and therapeutic evaluation of a syn-thetic peptide vaccine for the treatment of patientswith metastatic melanoma. Nat Med 1998; 4:321-7.

214. Chakraborty NG, Sporn JR, Tortora AF, et al. Immu-nization with a tumor-cell-lysate-loaded autologous-antigen-presenting-cell-based vaccine in melanoma.Cancer Immunol Immun 1998; 47:58-64.

215. Lotze MT, Hellerstedt B, Stolinski L, et al. The role ofinterleukin-2, interleukin-12, and dendritic cells incancer therapy. Cancer J Sci Am 1997; Suppl 1:S109-14.

216. Blieberg H, Rougier P, Wilke HJ. Management of colo-rectal cancer. Martin Dunitz Publisher, London, 1998.

217. Vermorken JB, Claessen AM, van Tinteren H, et al.Active specific immunotherapy for stage II and stageIII human colon cancer: a randomised trial. Lancet1999; 353:345-50.

218. Foon KA, John WJ, Chakraborty M, et al. Clinical andimmune responses in resected colon cancer patientstreated with anti-idiotype monoclonal antibody vac-cine that mimics the carcinoembryonic antigen. J ClinOncol 1999; 17:2889-95.

219. Tjoa B, Boynton A, Kenny G, Ragde H, Misrock SL,Murphy G. Presentation of prostate tumor antigens bydendritic cells stimulates T-cell proliferation and cyto-toxicity. Prostate 1996; 28:65-9.

220. Correale P, Walmsley K, Nieroda C, et al. In vitro gen-eration of human cytotoxic T lymphocytes specific forpeptides derived from prostate-specific antigen. J NatlCancer I 1997; 89:293-300.


186


221. Valone F, Small E, Peshwa MV, et al. Phase I trial ofdendritic cell-based immunotherapy with APC8015for hormone-refractory prostate cancer (HRPC). ProcAm Soc Cancer Res 1998; 39:173(Abstr).

222. Salgallar M, Lodge PA, Tjoa BA, et al. Monitoring ofprostate-specific membrane antigen (PSMA)-specificimmune response and prostate markers in a phase IIclinical trial with patients infused with dendritic cellspulsed with PSMA-derived peptides. Proc Am SocCancer Res 1998; 39:(Abstr).

223. Murphy G, Tjoa B, Ragde H, Kenny G, Boynton A.Phase I clinical trial: T-cell therapy for prostate cancerusing autologous dendritic cells pulsed with HLA-A0201-specific peptides from prostate-specific mem-brane antigen. Prostate 1996 ;29:371-80.

224. Rosenberg SA, Lotze MT, Muul LM, et al. A progressreport on the treatment of 157 patients with advancedcancer using lymphokine-activated killer cells andinterleukin-2 or high-dose interleukin-2 alone. N EnglJ Med 1987; 316:889-97.

225. Vogelzang NJ, Lipton A, Figlin RA. Subcutaneousinterleukin-2 plus interferon alfa-2a in metastatic renalcancer: an outpatient multicenter trial. J Clin Oncol1993; 11:1809-16.

226. Figlin RA, Pierce WC, Kaboo R, et al. Treatment ofmetastatic renal cell carcinoma with nephrectomy,interleukin-2 and cytokine-primed or CD8(+) selectedtumor infiltrating lymphocytes from primary tumor. JUrol 1997; 158:740-5.

227. Holtl L, Rieser C, Papesh C, et al. Cellular andhumoral immune responses in patients with metasta-tic renal cell carcinoma after vaccination with antigenpulsed dendritic cells. J Urol 1999; 161:777-82.

228. Kugler A, Stuhler G, Walden P, et al. Regression ofhuman metastatic renal cell carcinoma after vaccina-tion with tumor cell-dendritic cell hybrids. Nat Med2000; 6:332-6.

229. Dianzani U, Pileri A, Boccadoro M, et al. Activatedidiotype-reactive cells in suppressor/cytotoxic sub-populations of monoclonal gammopathies: correla-tion with diagnosis and disease status. Blood 1988;72:1064-8.

230. Osterborg A, Masucci M, Bergenbrant S, Holm G,Lefvert AK, Mellstedt H. Generation of T cell clonesbinding F(ab')2 fragments of the idiotypic immu-noglobulin in patients with monoclonal gammopa-thy. Cancer Immunol Immun 1991; 34:157-62.

231. Yi Q, Osterborg A, Bergenbrant S, Mellstedt H, HolmG, Lefvert AK. Idiotype-reactive T-cell subsets andtumor load in monoclonal gammopathies. Blood1995; 86:3043-9.

232. Yi Q, Eriksson I, He W, Holm G, Mellstedt H, Oster-borg A. Idiotype-specific T lymphocytes in monoclonalgammopathies: evidence for the presence of CD4+and CD8+ subsets. Br J Haematol 1997; 96:338-45.

233. Wen YJ, Lim SH. T cells recognize the VH comple-mentarity-determining region 3 of the idiotypic pro-tein of B cell non-Hodgkin’s lymphoma. Eur JImmunol 1997; 27:1043-7.

234. Wen YJ, Ling M, Lim SH. Immunogenicity and cross-reactivity with idiotypic IgA of VH CDR3 peptide inmultiple myeloma. Br J Haematol 1998; 100:464-8.

235. Fagerberg J, Yi Q, Gigliotti D, et al. T-cell-epitope map-ping of the idiotypic monoclonal IgG heavy and lightchains in multiple myeloma. Int J Cancer 1999; 80:671-80.

236. Bianchi A, Massaia M. Idiotypic vaccination in B-cellmalignancies. Mol Med Today 1997; 3:435-41.

237. Nelson EL, Li X, Hsu FJ, et al. Tumor-specific, cytotoxicT-lymphocyte response after idiotype vaccination forB-cell, non-Hodgkin's lymphoma. Blood 1996; 88:580-9.

238. Bendandi M, Gocke CD, Kobrin CB, et al. Completemolecular remissions induced by patient-specific vac-cination plus granulocyte-monocyte colony-stimulat-ing factor against lymphoma. Nat Med 1999; 5:1171-7.

239. Corradini P, Voena C, Astolfi M, et al. High-dosesequential chemoradiotherapy in multiple myeloma:residual tumor cells are detectable in bone marrowand peripheral blood cell harvests and after auto-grafting. Blood 1995; 85:1596-602.

240. Shtil A, Turner JG, Durfee J, Dalton WS, Yu H.Cytokine-based tumor cell vaccine is equally effectiveagainst parental and isogenic multidrug-resistantmyeloma cells: the role of cytotoxic T lymphocytes.Blood 1999; 93:1831-7.

241. Osterborg A, Yi Q, Henriksson L, et al. Idiotype immu-nization combined with granulocyte-macrophagecolony-stimulating factor in myeloma patientsinduced type I, major histocompatibility complex-restricted, CD8- and CD4-specific T-cell responses.Blood 1998; 91:2459-66.

242. Massaia M, Borrione P, Battaglio S, et al. Idiotypevaccination in human myeloma: generation of tumor-specific immune responses after high-dose chemother-apy. Blood 1999; 94:673-83.

243. Dhodapkar MV, Steinman RM, Sapp M, et al. Rapidgeneration of broad T-cell immunity in humans aftera single injection of mature dendritic cells. J Clin Invest1999; 104:173-80.

244. Hsu FJ, Benike C, Fagnoni F, et al. Vaccination ofpatients with B-cell lymphoma using autologous anti-gen-pulsed dendritic cells. Nat Med 1996; 2:52-8.

245. Reichardt VL, Okada CY, Liso A, et al. Idiotype vacci-nation using dendritic cells after autologous periph-eral blood stem cell transplantation for multiplemyeloma-a feasibility study. Blood 1999; 93:2411-9.

246. Wen YJ, Ling M, Bailey-Wood R, Lim SH. Idiotypicprotein-pulsed adherent peripheral blood mononu-clear cell-derived dendritic cells prime the immune sys-tem in multiple myeloma. Clin Cancer Res 1998; 4:957-62.

247. Cull G, Durrant L, Stainer C, Haynes A, Russell N.Generation of anti-idiotype immune responses fol-lowing vaccination with idiotype-protein pulsed den-dritic cells in myeloma. Br J Haematol 1999; 107:648-55.

248. Lim SH, Bailey-Wood R. Idiotypic protein-pulsed den-dritic cell vaccination in multiple myeloma. Int J Can-cer 1999; 83:215-22.

249. Titzer S, Christensen O, Manzke O, et al. Vaccinationof multiple myeloma patients with idiotype-pulseddendritic cells: immunological and clinical aspects. BrJ Haematol 2000; 108:805-16.

250. Gong J, Chen D, Kashiwaba M, Kufe D. Induction ofantitumor activity by immunization with fusions ofdendritic and carcinoma cells. Nat Med 1997; 3:558-61.

251. Schultze JL, Cardoso AA, Freeman GJ, et al. Follicularlymphomas can be induced to present alloantigen effi-ciently: a conceptual model to improve their tumorimmunogenicity. Proc Natl Acad Sci USA 1995;92:8200-4.

252. Schultze JL, Michalak S, Seamon MJ, et al. CD40-acti-vated human B cells: an alternative source of highlyefficient antigen presenting cells to generate autolo-gous antigen-specific T cells for adoptive immunother-apy. J Clin Invest 1997; 100:2757-65.

253. Teoh G, Chen L, Urashima M, et al. Adenovirus vec-tor-based purging of multiple myeloma cells. Blood1998; 92:4591-601.

254. Kwak LW, Taub DD, Duffey PL, et al. Transfer ofmyeloma idiotype-specific immunity from an actively

M. Bocchia et al.

187


immunised marrow donor. Lancet 1995; 345:1016-20.

255. Takahashi T, Makiguchi Y, Hinoda Y, et al. Expressionof MUC1 on myeloma cells and induction of HLA-unrestricted CTL against MUC1 from a multiplemyeloma patient. J Immunol 1994 ;153:2102-9.

256. Treon SP, Mollick JA, Urashima M, et al. Muc-1 coreprotein is expressed on multiple myeloma cells and isinduced by dexamethasone. Blood 1999; 93:1287-98.

257. Van Baren N, Brasseur F, Godelaine D, et al. Genesencoding tumor-specific antigens are expressed inhuman myeloma cells. Blood 1999; 94:1156-64.

258. Rao PH, Cigudosa JC, Ning Y, et al. Multicolor spec-tral karyotyping identifies new recurring breakpointsand translocations in multiple myeloma. Blood 1998;92:1743-8.

259. Chesi M, Nardini E, Brents LA, et al. Frequent translo-cation t(4;14)(p16.3;q32.3) in multiple myeloma isassociated with increased expression and activatingmutations of fibroblast growth factor receptor 3. NatGenet 1997; 16:260-4.

260. Stevenson FK, Zhu D, King CA, Ashworth LJ, Kumar S,Hawkins RE. Idiotypic DNA vaccines against B-celllymphoma. Immunol Rev 1995; 145:211-28.

261. Tao MH, Levy R. Idiotype/granulocyte-macrophagecolony-stimulating factor fusion protein as a vaccinefor B-cell lymphoma. Nature 1993; 362:755-8.

262. Syrengelas AD, Chen TT, Levy R. DNA immunizationinduces protective immunity against B-cell lymphoma.Nat Med 1996; 2:1038-41.

263. King CA, Spellerberg MB, Zhu D, et al. DNA vaccineswith single-chain Fv fused to fragment C of tetanustoxin induce protective immunity against lymphomaand myeloma. Nat Med 1998; 4:1281-6.

264. Biragyn A, Tani K, Grimm MC, Weeks S, Kwak LW.Genetic fusion of chemokines to a self tumor antigeninduces protective, T-cell dependent antitumor immu-nity. Nat Biotechnol 1999; 17:253-8.

265. Kolb HJ, Mittermuller J, Clemm CH, et al. Donorleukocyte transfusions for treatment of recurrentchronic myelogenous leukemia in marrow transplantpatients. Blood 1990; 76:2462-5.

266. Mackinnon S, Papadopoulos EB, Carabasi MH, et al.Adoptive immunotherapy evaluating escalating dosesof donor leukocytes for relapse or chronic myeloidleukemia after bone marrow transplantation: separa-tion of graft-versus-leukemia responses from graft-ver-sus-host disease. Blood 1995; 86:1261-8.


268. Nieda M, Nicol A, Kicuchi A, et al. Dendritic cells stim-ulate the expansion of bcr-abl specific CD8+ T cellswith cytotoxic activity against leukemic cells frompatients with chronic myeloid leukemia. Blood 1998;91:977-83.

269. Schadendorf D. Cancer Vaccines. Once more unto thebreach. The Economist 1994; 12:78-9.

270. Molldrem JJ, Lee PP, Wang C, Champlin RE, DavisMM. A PR1-human leukocyte antigen-A2 tetramercan be used to isolate low-frequency cytotoxic T lym-phocytes from healthy donors that selectively lysechronic myelogenous leukemia. Cancer Res 1999; 59:2675-81.

271. Pass HA, Schwarz SL, Wunderlich JR, Rosenberg SA.Immunization of patients with melanoma peptide vac-cines: immunologic assessment using the ELISPOTassay. Cancer J Sci Am 1998; 4:316-23.

272. Wolfel T, Hauer M, Schneider J, et al. A p16INK4a-insensitive CDK4 mutant targeted by cytolytic T lym-phocytes in a human melanoma. Science 1995;269:1281-4.

273. Vonderheide RH, Hahn WC, Schultze JL, Nadler LM.The telomerase catalytic subunit is a widely expressedtumor-associated antigen recognized by cytotoxic Tlymphocytes. Immunity 1999; 10:673-9.

274. Jonsen AR, Durfy SJ, Burke W, Motulsky AG. Theadvent of the 'unpatients'. Nat Med 1996; 2: 622-4.

275. Nanni P, Forni G, Lollini PL. Cytokine gene therapy:hopes and pitfalls. Ann Oncol 1999; 10:261-6.

276. Pardoll DM. Inducing autoimmune disease to treatcancer. Proc Natl Acad Sci USA 1999; 96: 5340-2.

277. Prevention of cancer in the next millenium: Report ofthe Chemoprevention Working Group to the Ameri-can Association for Cancer Research. Cancer Res1999; 59: 4743-58.

278. Ehrlich P. Ueber den jetzigen Stand der Karzinom-forschung. Ned Tijdschr Geneeskd 1909; 5:273-90.


Index of authors

Aglietta Massimo, 19, 92Arcese William, 69Aversa Franco, 69 Bandini Giuseppe, 69 Bertolini Francesco, 49, 92 Bocchia Monica, 156 Bordignon Claudio, 117 Bronte Vincenzo, 156 Caligaris Cappio Federico, 32 Carlo-Stella Carmelo, 1, 92, 117 Cavo Michele, 32 Cazzola Mario, 1 Colombo Mario P., 117, 156 De Fabritiis Paolo, 1 De Vincentiis Armando, 1, 19, 32, 49, 69, 92, 117, 156 Di Nicola Massimo, 156 Falda Michele, 69 Forni Guido, 156 Gianni Alessandro Massimo 1Lanata Luigi , 19, 32, 49, 69, 92, 117, 156 Lanza Francesco, 1, 19 Lauria Francesco, 1 Lemoli Roberto M., 1, 19, 32, 49, 69, 92, 117, 156 Locatelli Franco , 69, 117 Maccario Rita, 49 Majolino Ignazio, 32, 49, 69 Massaia Massimo, 156 Menichella Giacomo, 19 Olivieri Attilio, 92, 117 Ponchio Luisa, 49 Rondelli Damiano, 49, 117, 156 Siena Salvatore, 92 Tabilio Antonio, 49 Tafuri Agostino, 19Tarella Corrado, 1, 32 Tura Sante, 1, 19, 32, 49, 69, 92, 117, 156 Zanon Paola, 1, 19, 32, 49, 69, 92, 117, 156

Direttore responsabile: Prof. Edoardo Ascari

Autorizzazione del Tribunale di Pavia n. 63 del 5 marzo 1955

Composizione: = Medit – via gen. C.A. Dalla Chiesa, 22 – Voghera, Italy

Stampa: Tipografia PI-ME – via Vigentina 136 – Pavia, Italy

zzz

Printed in December 2000

In collaborazione con Dompé Biotec s.p.a.

c i g o l o hematologyvolume 85supplementto no. 12december...

Documents