THE CANARYPOX VIRUS ALVAC AS A VECTOR lN CANCER GENE THERAPY

Abhijit Ghose

A thesis submitted in conformiv with the requirements for the degree of Master of Science

Graduate Department of lnstitute of Medical Science University of Toronto

O Copyright by Abhijit Ghose (1999)

National Library 1+1 dcanacia Bibliotheque nationale du Canada

Acquisitions and Acquisitions et Bibliographie Services senrices bibliographiques

395 Weilirigton Sireet 395. rue Wdligtori Ot!awaO(J K l A W -ON K l A W canada Canada

The author has granted a non- L'auteur a accordé une licence non exclusive Licence aiiowing the exclusive permettant à la National Library of Canada to Bibliothèque nationale du Canada de reproduce, loan, distriiute or seii reproduire, prêter, distribuer ou copies of this thesis in microform, vendre des copies de cette thèse sous paper or electronic formats. la forme de microfiche/film, de

reproduction sur papier ou sur format électronique.

The author retains ownership of the L'auteur conserve la propriété du copyright in this thesis. Neither the droit d'auteur qui protège cette thèse. thesis nor substantid extracts f b m it Ni la thèse ni des extraits substantiels may be printed or otherwise de celle-ci ne doivent être imprimés reproduced without the author's ou autrement reproduits sans son permission. autorisation.

The Canarypos Virus ALVAC as a Vector in Cancer Genc Therapy

By Abhijit Ghose

A Thesis submitîed in codormity with the requirements for the degree of Master of Science,

Graduate Department of Insitute of Medicai Science, University of Toronto, 1 999.

Abstract

The immunogenicity of recombinant canarypox (ALVAC) viral vectors within whole ce11 turnor

vaccines was evaluated using the early T-ce11 thymic Lymphoma STFl O. STFl O cells were

modi fied with the recombinant ALVAC vectors and injected into syngeneic mice. ControI mice

receiving unmodified cells developed twnors, while mice injected with STFIOIALVAC,

STF 1 WALVAC B7-1 or STF I O/ALVAC 87- 1 or STF 1 OIALVAC B7- l/ALVAC IL- 12 compktefy

rejected their tumors. Rechalleage at day 55 showed that none of these groups were

si gni ficantly protected However, modi fied regimens incorporating two vaccines induced a

protective effect in al1 vaccinateci groups. Notably, the parental ALVAC virus was equivalent

to al1 other recombinant ALVAC viwes in conferring antitumor immunity. Tumorigenicity

experirnents in nude mice revealed that the effector mechanism mediating rejection of tumor

cells bean'ng ALVAC vectors is multifactorid. Finally, in vilro experirnents revealed that

cytotoxic T-cells specific for parental STFlO cells cwld be generated as a result of in vivo

immunization with STF 1 O/ALVAC vaccines..

Acknowledgments

First and foremost, 1 would like to thank Dr. Neil Berinsteio for introducing me to the

exciting field of cancer immunology, and providing the knowledge and inspiration

necessary for any discipline that proposes to look into the nature of life- His unbridled

enthusiasm for learning and exploring infused us dl , and 1 can only hope that in fÙture, 1

can approach science with the same passion.

My sincerest thanks to my program advisors, Dr. Andre Schuh and Dr. David Spaner,

who shaped my project into what it becarne, provided me with academic guidance, and

chailenged me to learn about subjects that were completely alien to me. Thank you to

Jillian Haight, Rosemary Ahrens and Beverly McLaughlin, without them nothing would

have been accomplished. Many thanks to Monica Ghose for sharing her exemplary word

processing talents, this thesis would be quite unpresentable otherwise.

To Jennifer Tomlin and Jane Gillis, thanks for making me feel welcome; to Laurent

Verkoczy, 1 have yet to become a master such as yourself; to Nancy Pennell, th& for

always being there and having the solution to every problem; to Barbara Guinn, thank

you for doubling as a supervisor every now and then and showing me the ropes.

Thanks to Ivan Fong, who taught me what it meant and what it took to be a reai graduate

student; thanks to Maria Lorenzo and Ken Luk, their fnendshîp kept me sane, Thank you

to Elena Iakhnina, who throughout my studentship, was a good friend. excellent teacher

and a worthy tennis opponent. Finaily, my heartfelt thanks to AL Zarrin without whose

mentorship, my faith in science wodd be uncertain. He kept reminding me of the things

that are t d y important in life, and showed through exarnple how to become a scientist,

artist and philosopher at the sanie tirne.

Table of Contents

Absn-acr Acknowledgmen ts Table of Contents LLÎf of Tables a d Figures Lisr ofAbbreviations

Introduction Models in Antitumor Immunity - Recombinant Costirnulatory Molecuie Expression

B7- 1 B 7-2 ICAM 1 4-IBBL

Models in Antitumor Immunity - Recombinant Cytokine Expression

IL-2 I L 4 I L 6 IL-'? IFN-y TNF-a GM-CSF IL- 12

Combination Therapies Utilizing IL- 12 and 87- 1

Gene Transfer Methodologies Physicai Methods of Gene Transfer

Plasmid DNA I n m ~ t i o n Virai Methods of gene Transfer

Retrovirus Adenovins Poxvinises Canarypox V h (ALVAC)

Generation of Recombinant ALVAC viruses Titration of ALVAC v i w e s The ALVAC V h s in Cancer Therapy

Active Imrnunotherapy with tumors Genetically modified via ALVAC Vectors

.. 11 ..- 111

v vii xiv

Current Experimental Approach

Material and Methoth Cell Lines Infection of Ce11 Lines via ALVAC vectots Flow Cytometric Assessrnent of Ce11 Surface Marker Expression Irnmunoassay for IL- 12 production Reverse Transcription (cDNA synthesis) PCR for Gene Expression Growth Kinetics of ALVAC-Mected STF 1 O cells Stability of ALVAC Encoded products over time in ST Tumorigenicity in Wild Type BALB/c mice: Vaccination and Challenge Statisticd Analysis Cutotoxic T-lymphocyte Assays

Preparation of Stimulator Celis Preparation of Effector Cells Preparatoin of Target Cells CTL Setup

Differential Effects of in vitro stimulation on STF1 0 Lysis

Results Heterogeneity in T r o p i d i r a 1 Gene Expression Surface Marker Phenotype of STF 10 Growth Kinetics of ALVAC-Mected STF 10 cells MHC-class 1 expression on ALVAC lnfected STF 1 O cells Stability of ALVAC encoded Products over T h e Decreased Tumorigenicity of STF 1 O cells Modified by ALVAC Vectors

Tumorigenicity in Nude Mice Boosting with whoIe ce11 vaccines bearïng ALVAC vectors confers antitumoral immunity In Vitro Antitumoral Reactivity

Discussion Immunogenicity of the ALVAC Vector Antitumoral Protection following Vaccine Boost Protection against Tumor Epitopes as a Result of Vird Immunogenicity T-ce11 Independent Antitunior Mechanisms In Vitro CTL Activity

Ftrture Directions

References

Tables and Figures

Table 1: Level of IL-12 secreted by STFlO vaccines prior to h u n i z a t i o n of mice

Figure 1: Genomic Organization of recombinant ALVAC viruses

The ALVAC virus vector consists of a Linear double stranded DNA molecule measining

about 325 Kbp. The multiple cloning sites CS and C6 have been inserted as s h o w

above. The ALVAC B7-1 construct consists of a human B7-1 cDNA driven by the

vaccinia H6 promoter inserted into the C6 MCS, The ALVAC IL-12 constmct consists

of two p35 subunit cDNA sequences each driven by one E3L vaccinia promoter, and a

p40 subunit cDNA driven by the entomopox 42K promoter. The p35 and p40 subunit

constructs are inserted into the CS and C6 loci, respectively.

Figure 2: Heterogeneity in ALVAC Viral Transduction

To determine the tropism exhibited by the Canarypox virus ALVAC, ce11 lines of

different lineages and stages were assayed for the ability to successfdly take up the virus

and express the virally encoded protein B7-1, Clear histograms represent cells that were

specifically stained by anti-human B7-1 antibody, while shaded histograms represent

staining via an isotype matched non-specific antibody. Results indicate that recombinant

B7- 1 is expressed only in select tissues represented by NFS-70, B 16, ST, as well as two

vii

ST subclones STFlO and STG4. A20, K465, CI498 and P8 15 are completely negative

for ALVAC-encoded products.

Figure 3: Surface Marker Phenotype of STFlO

STF 1 O cells were stained with antibodies specific for the surface-expressed proteins

indicated, and then assayed by ceil cytometry as outlined in Materials and Methods-

Clear histograms represent specific staining for the marker indicated, while shaded

histograms represent irrelevant stahing of STF 10 via a non-specific isotype matched

antibody. STF10 cells are CD44+, CD25-, HSA+, CD4 low, CD8 Iow, and CD3-.

Figure 4: RT-PCR analysis of ST-FlO, and STFlO infected with ALVAC, for TdT,

RAG-1, RAG-2, and HPRT.

ST-F 1 O yields the expected size PCR product for RAG-1. RAG-2, as well as TdT. The

positive control for PCR ampification of TdT consisted of a genomic sample fiom mouse

tail-DNA. Positive controls for amplification of RAG-1, RAG-2 and HPRT consisted of

various cDNA preps of the mouse Pro-B Ce11 Iine NFS-70.

Figure 5: Growth Characte~tics of STF10 Cells af'ter infection Ma ALVAC vectors

STF I O cells were infected with ALVAC B7-1 as outlined in Materials and Methods, and

then replated in culture media at a concentration of 20,000 cellslml. Growth was

viii

monitored over a penod of 4 days, and was compared to the growth rate of a STFlO

culture that was plated at the same concentration, but was uninfected with ALVAC B7- 1.

Resuits indicate that infection of cells via recombinant ALVAC vectors does not alter the

growth characterstics of the STFl O ceils as the rate of expansion of the two cultures is

almost identical.

Figure 6: Effect of ALVAC viral lnfection on MHC class I expression via STFlO

cells

STF 1 O cells were infected with ALVAC B7- 1 as outlined in Materials and Methods, and

assayed for the expression of MHC 1 at tirnepoints corresponding to 6 hours into viral

infection, and 24 hours following viral infection. Results indicate that there is no

downregulation of classical MHC class I ( H - ~ K ~ and i 3 - 2 ~ ~ ) as the geomtric mean

fluorescence of infected cells is 17 at 6 hours, and 19 at 24 hows, compared to unidected

celk which display a geometric mean fluorescence of 14 at 6 hours and 1 5 at 24 hours.

Figure 7: Stability of ALVAC Encoded Products Over Time

Because the ALVAC viral vector is non-replicative in mammalian tissues, the expression

of ALVAC-encoded products will be transient as the population of cells divide and

ALVAC proteins approach their haif-life in situ. The SCID Thymoma ce11 line ST was

used as an indicator of ALVAC product stability: SCID Thymoma bulk culture was

transduced with ALVAC-B7-1 and half the culture was irradiated to stop it fiom dividing

further, while the other half was kept in log phase. Both cultures were assayed for

ALVAC-encoded 87-1 expression at discrete tirnepoints. Results indicate that in a non

dividing culture, GLVAC-encoded products can be detected by immunostaining upto

four days (> 96 hours) post infection, whereas for a rapidy dividing culture, expression is

optimal for about two days (48 hours), after which expression levels start declining, and

is Iost completely at four days.

Figure 8: Survival of Mice Post STFlO Tumor Variant Vaccination

STF 1 O tumor cells were infected with a) No virus b) ALVAC c) ALVAC-B7- 1 d)

ALVAC-IL- 12 or e) ALVAC-B7- 1 and ALVAC-IL- 12, and injected subcutaneously

into mice (using 5 mice per vaccine). Mice receiving cells alone succumbed to twnor,

whereas mice that received cells containing either parental or recombinant ALVAC

vectors al1 rejected tbeir himors at a fiequency of 1 00%.

Figure 9: Lack of Systemic Protection against Wild Type Challenge After a Single

Vaccination

Subsequent to vaccination of mice with STF 10 transduced with a) No Virus b) ALVAC

c) ALVAC-B7-1 d) ALVAC-IL42 or e) ALVAC-B7-1 and ALVAC-IL-12, ail

surviving mice were challenged at day 55 with a subcutaneous injection of wild type

(unrnodified) STF 1 O cells on the opposite (left) lateral flank. Previously unvaccinated

(naive) mice were also injected with wild type SFlO cells as a positive controi for

tumorigenicty. Survival of ùnmunized mice is not statistically improved for any group

denoted above (p > O.OS), even though there appears to be a delay in twnor formation in

mice previously immunized with STF 1 O/ALVAC-IL- 12.

Figure 10: Tumorigeniciîy of STFlO Variants in Nude Mice

STF 1 O cells were infected with parental or recombinant ALVAC vinses as desctïbed in

Materials and Methods, and injected subcutaneousIy into nude mice. All anirnals

receiving cells alone succumbed to tumor in both experiments (A & B). Animais

receiving STF 10 celis containing either parental or recombinant ALVAC vectors

displayed tumor rejection at statisitically significant (p < 0.05) fiequencies, relative to

mice receiving cells alone, in both independent experiments.

Figure 11: Survival of Mice Post Wild Type STFlO Challenge Followuig Vaccine

Boost

Mice were prîmed and boosted with cellular vaccines of a) STFlO / ALVAC b) STF 1 O /

ALVAC-B7 1 c) S E 10 / ALVAC-IL- 12 d) STF 1 O/ALVAC-B7- 1 and STF I O / ALVAC-

IL-12. In addition, one group of mice was only prirned with STF 10 / ALVAC-IL-12, but

left unboosted. Vaccinations and boosts were administered subcutaneously, on the right

flank, with the boost taking place 35 days after the initial vaccination. 21 days post the

boost, mice were challenged subcutaneously on the left flank with wild type SV10 cells

and assayed for survival. Naive controls that had not been previously vaccinated al1

developed tumors, while al1 other vaccinated groups displayed improved survival relative

to naive mice @ < 0.05 for al1 groups relative to naive mice).

Figure 12: Specific Activity of Splenocytes from STFlO/ALVAC-IL42 immunized

mice

Mice were immunized with STF 1 OIALVAC-IL- 1 2 and spleens were harvested on days

1 3, 1 7, 27 and 3 5. Splenocytes were stimulated with STF 1 O/ALVAC and specific killing

against STF 1 O/ALVAC-IL- 12 (A & B) or wild type STF 1 O (C & D) cells was evaluated.

Cytolytic activity was observed on day 27 against STFlO/ALVAC-IL42 cells (B) and

not STFlO cells (D). No cytolytic activity against either type of cells was observed on

day 1 3 (not shown), day 1 7 (A & C) and day 35 (not shown).

Figure 13: Effects of Differential Stimulation on Splenocyte Reactivity against

STFlO

Mice were imrnunized with STFlO/ALVAC and then boosted with the same vaccine. 21

days following the boost, splenocytes fiom an immunized mouse and a naïve control

were harvested and stimulated with STFlO (A & B) or STFIO/ALVAC (C). Results

indicate that cytolytic precursors specific for STFlO are indeed present in the splenocyte

culture (A), in addition to those specitic for STF 1 O/ALV AC (B); however, their detection

xii

is dependent on stimulation with wild type STF 10. Stimulation via STF 1 O/ALVAC does

not yield efficient cytotoxicity against wild type STF10, relative to naive controls (C).

xiii

Ab breviations

APC P-gai bp CEA CEF GY CHO CLMF CM CNS CON-A CTL DNA DTH D n : FCS GP GM-CSF HPRT IFN IL IRES Kb LCMV LTR MBP MCS Me MHC pCi MOI MS NK NKSF NTP PBS PBMC Pfù PHA PLP RAG R-EAE RG RNA

Antigen Presenting Ceil beta-galactosidase base pair Carcinoembryonic Antigen Chicken Embryo Fibroblast centigray Chinese Hamster Ovary C ytotoxic T-lyrnphoc yte Maturation Factor Culture Media Centrai Newous System Concanaval in-A Cytotoxic T-lymphocyte Deoxyribonucleic Acid Delayed Type Hypersensitivity DithioxytreitoI Fetai Calf Senun Glycoprotein Granulocyte Macrophage Colony Stimulating Factor Hypoxanthine Phosphoribosyl Transferase Interferon Interleukin Intemal Ribosome Entry Site Kilo base Lymp hocytic Choriomeningitis Vins Long Terminal Repeat Myelin Basic Protein Multiple Cloning Site Mercaptoe thano 1 Major Histocompatibility Complex microcurie Multiplicity of infection Multiple Sclerosis Naturd Killer Natural Killer Stimulatory Factor Nucleoside Triphosphate Phosphate BufSered Saline Peripheral Blood Mononuclear Ce11 Plaque Forming Unit P hytohemagglutinin Proteolipid Recombination Activating Gene Relapsing-Rernitting Experimental Autoimmune Encephalomyelitis Rabies Glycoprotein Ribonucleic Acid

xiv

r W SCID TdT TMEV TNF VSV

recombinant Vaccinia Virus Severe Combined Immunodeficiency TenninaI deoxy nucleo tidy k Transferase Theiler's Murine Encephalomyelitis Virus Turnor Necrosis Factor Vesicular Stomatitis V h

Introduction

Advances in our understanding of antigen presentation and effector activation have

provided the basis for new and exciting strategies for the therapy of cancer. It is widely

believed that spontaneous tumors arise due to an acquired ability of cancer cens to evade

immune surveillance, a process whereby immune effectors specifically recognize and

eliminate tumor ceiis fiom the body (1).

Normally turnor cells express various protein epitopes (tumor rejection antigens) on their

ce11 surface via theu MHC class I and class II molecules, which can be specifically

recognized by T-cells (2). In the course of a progressive cancer, the interaction between

the cancer ce11 and the T-cell. which shodd cause specific activation of the T-cell, does

not occur. This could be induced by several factors including immunological ignorance

of the tumor cell, down-regdation of MHC molecules expressing the tumor antigen (3),

down-regulation of costimulatory or adhesion molecules, loss of tumor antigen due to

genetic ùistability (4), and expression of immunosuppressive cytokines (5).

Thus, rnany biological therapies against cancer are aimed toward reactivation of these

tumor antigen specific T-cells. One such approach in active irnrnunotherapy involves the

modification of an MHC expressing ce11 iine by transfection of a gene encoding a

costimulatory molecule (6). This costimulatory molecule provides the second activating

signal to the T-cell, the fist signal comprising of activation via the T-ce11 receptor as it

recognizes a cognate MHC-peptide complex on the antigen presenting ce11 (7). Once a

cytotoxic T-ce11 is activated it can specifically recognize and kiil other cells with or

without receiving the costimulatory signal,

An alternative strategy is to introduce into the tumor ce11 a gene encoding a cytokine (8).

Cytokines can have a wide array of targets and effects, and generally they serve either to

activate immune effectors like T-cells, NK ceUs and macrophages directly, or recmit

professionai antigen presenting cells (APC). These APC take up tumor antigens, migrate

to the lymph node and then activate T-cells more potently due to a greater number of

costirnuIatory molecules present on their surface (9).

Models in Antitumor Immunity - Recombinant Costirnulatory Molecule Expression

In order for T-cells to be activated against an antigenic epitope they must receive two

signals from the antigen presenting ce11 (10): signal 1, arising fiom cognate interactions

between the MHC-peptide cornplex on the APC and a complementary T-ce11 receptor on

the T-cell, and signal 2, arising from a costirnutatory molecule on the APC and a

complementary receptor on the T-cell. Various ce11 SUfface receptors have been shown to

ddiver costimulatory signals to the T-celis including B7-1 and B7-2, 41BB-Ligand,

CD40-Ligand, ICAM 1,2, and 3, and LFA 3 (1 1).

Expression of costimulatory molecules has also been used in active immunotherapy

models, whereby tumor cells have been genetically modified to express genes which

provide this second activating signal, and manipulate the immune system to mount a

response against the tumor cells,

B 7-1

In a landmark paper, Chen et al. (6) showed that transfection of a murine melanoma

KI 735-M2 with an antigen E7 and the costimulatory molecule B7-1 induced complete

tumor rejection in 100% of mice, after an initial phase of tumor growth. This rejection

was dependent on the presence of the antigen E7, and was abrogated in nude mice,

showing that the in vivo response was T-ce11 mediated. More specifically, depletion

experiments showed that CD8+ cytotoxic T-cells were responsible for tumor rejection

and CD3+ T-helper cells could be removed without changing the antitumor response.

Furthemore, wild type tumors could be rejected with concomitant injection of B7-I

expressing tumors on the opposite flank, and 4 day established lung metastases could be

cured in 40% of mice with injection of 87-1 expressing tumor cells.

Townsend and Allison (12) performed a sirnilar experiment where they showed that

KI 735 cells were rejected at a fiequency of 90% if expressing recombinant B7- and that

this was an effective vaccine in mice, as 90% of the survivors were protected upon

rechallenge. Once again CD8+ T-cells were responsible.

Subsequently, however, Chen et al. (13) found that the inherent immunogenicity of the

tumor ce11 may be a factor in B7-1 mediated antitumor irnmunity. They assayed four

immunogenic lines RMA, EL-4, P8 15 and E6B2, and four non immunogenic lines MCA

101, MCA 102, Ag104 (sarcomas) and B16 melanoma and found that in the

immunogenic tumors, recombinant 37-1 expression induced 100% rejection, while in al1

the non-irnrnunogenic tumors, B7-1 expression had no effect. Furtherrnore, in v i m

anaiysis showed that the ody two immunogenic tumors tested, EL-4 and P8 15, induced a

CTL response fiom irnmunized mice, while the non-immunogenic tumors MCA 102 and

B 16 did not.

B 7-2

A second B7 molecule, designated 87-2 was cloned in 1993 by Freeman et al. (1 4 ) ,

based on its ability to induce T-cells to produce IL-2 and proliferate, and the abrogation

of this activity by CTLA4-Ig but not anti-B7-l antibody. Subsequently, B7-2 was also

tried in several active immunotherapy models. La Motte et al. (1 5 ) showed that P8 15

mastocytoma cells expressing B7-2 were just as effective as B7-1 expressing cells in

inducing an antitumoral response, and 87-1 and 87-2 tumor variants were equally

effective in retarding the growth of established tumors. Yang et al. (16) also

demonstrated that P8 15 cells expressing B7-2 were immunogenic, and protected mice

against wild type challenge through specific recruitment of CD8+ T-cells only.

[CAM 1

Initial active irnmunotherapy studies with ICAM 1 showed that ICAM 1 gene

transfection into the mwine fibrosarcoma MCA 105 significantly slowed tumor growth

(1 7, 18). Subsequent work by Uzendoski et al. (19) revealed that with MC38 tumor

cells, expression of ICAM 1 stimulated concanavalin A-activated T-cells to secrete TNF-

a. As well, a specific alloreactive cytotoxic T-ce11 reaction was observed against MC38

cells that were expressing recombinant ICAM 1. in vivo MC38 cells that were

expressing ICAM 1 were rejected at a fiequency of 100% and hirthermore, dl surviving

mice were protected from M e r tumor challenge, even when the dose of the challenge

was twice that of the immunogen.

4-1 BBL

4-1 BB ligand (4-IBBL) is a type II surface glycoprotein which belongs to the tumor

necrosis factor (TNF) family, and is expressed on antigen presenting cells such as

activated B-cells, macrophages and splenic dendritic cells (20, 2 1 ). It is a costirnulatory

molecule which can activate T-cells independently of CD28-engagement by either 87-1

or B7-2, but can also synergize with the latter two molecules, especially at low levels of

CD3 activation (22). in vivo studies have shown that like B7-1, expression oE4-1BBL on

tumor cells can effectively activate T-cells against the respective tumor, and in the long

mn, confer antitumor immunity. Guinn et al. (23) demonstrated that while A20 cells

expressing B7-1 were 100% tumorigenic, and those expressing B7-2 were more

immunogenic, A20 cells expressing CBBL in conjunction with B7-2 were completely

rejected in syngeneic mice. Furthemore, challenge of surviving mice with wild type

A20 revealed protective effects in 90-100% of rnice, depending on the clone that was

used as the immunogen Similarly, Melero et al. (24) showed that P8 15 cells expressing

4-1BBL were rejected in >90% of mice, and al1 surviving mice were protected against

systemic challenge. Antibody depletion experiments showed that only CD8+ T-cells

were required for tumor rejection when tumor cells themselves were expressing 4- 1 BBL,

and both CD4+ and CD8+ T-cells were needed if wild type tumors were already

established (24,25).

Models in Antitumor Immunity - Recombinant Cytokine Ex~ression

Cytokines are key modulators of host immune and idammatory responses. The transfer

and expression of cytokine expressing genes into tumor cells has yielded a novel strategy

to augment antitumor reactivity in various murine models. These gene modified tumors

have been shown to induce a rejection of the -or variants themselves, and in many

cases, induce systemic irnrnunity against the parental tumor cells.

Toward the goal of reducing tumongenicity via stimulation of local immune and

inflarnrnatory reactions, a number of cytokine genes have been used to genetically

rnodi* tumor cells (26,27). These include IL-1 (28), IL-2 (29,30), IL-4 (3 1,32), IFN-y

(33-39, IL-6 (36, 37), IL-7 (38, 39), IL-12 (40,41), TNF-a (42, 43), and GM-CSF (8).

The pleiotropic effect of cytokines on the various arms of the innate and adaptive

immune system poses a problem in establishing a definite mechanism by which tumor

cells are rejected. Because different cytokines ultimately affect different downstrearn ce11

types, different models of antihunor immunity have evolved.

IL-2

A variety of tumor models utilizing expression of IL-2 have been tested, including a

colon carcinoma CT26 and a melmoma Bl6 (291, a fibrosarcoma CMS-5 (30), a

mastocytoma (44), a mammary adenocarcinorna TSA (49, a fibrosarcoma MCA 102

(46) as well as bladder and lung carcinomas (47,48).

Predorninantly, the ttunor rejection or onset delay that is observed upon injection of

tumor cells expressing IL-2 is mediatec! by cytotoxic T-cells, though in some cases,

neutrophils and NK cells have been implicated (45, 46). A protective immunity toward

parental challenge was also found in a nunber of these cases- In a mode1 of an

established tumor, Comor et al- (47) grew a 7 day bladder carcinoma MBT-2 tumor in

mice, and observed a decreased tumor size with concomitant injection of radiation killed

IL-2 secreting cells.

I L 4

Unlike studies with IL-2, -or rejection models utilizing IL-4 secreting cells have

shown a number of different effector mechanisms to be responsible, depending on the

particular turnor line studied. Using a renal carcinoma RENCA, it was shown that local

rejection was actually T-ce11 independent and mediated by eosinophils; however,

systernic irnmunity toward wild type challenge was conferred and this was mediated only

through CDS+ T-cells (3 1). Studies with the J558L plasmacytoma showed eosinophils

and macrophages to be the major effector in confemng rejection of tumor cells, as

defined by blocking with specific antibodies against granulocytes (32).

IL-6

IL-6 is usually classified as a cytokine that is involved in uiflammatory responses and it

is produced by macrophages, monocytes, as well as activated T-cells (49). It has also

been shown to be able to increase cytotoxic activity of T-cells in vitro (50). IL-6 has

been implicated in both nonspecific as well as specific antitumoral responses. Sun et al.

(37) working with the B16 melanoma, showed that the antitumor responses were due to

mainly nonspecific proinflanmatory effectors like Macrophages and neutrophils.

Porgador et al. (36), working with the Lewis Lung Carcinoma D122, however, showed

that IL-6 conferred a reduction in both tumorigenicity as well as metastatic cornpetence

that was a result of T-ceIl mediated function. In addition, IL-6 irradiated cells were a

potent vaccine toward protecting mice against a subsequent wild type tumor challenge'

and moreover, a reduction in established tumor growth could also be obsewed with

concomitant injection of inactivated IL-6 transfectants.

IL- 7

IL-7 is a cytokine known to have a critical role in the development and maturation of

lymphocytes (7). However, this cytokine was also used in the context of antitumoral

immunity by Hock et al. (38) who showed that the plasmacytoma J558L was completely

rejected upon injection of IL-7 expressing tumors into mice. frnmunohistochemical

analysis showed infiltration of macrophages and CD4+ as well as CD8+ T-cells within

the tumor, while subset depletion experirnents showed an absolute dependence on CD4+

T-cells as well as macrophages, but not CD8+ T-cells. In contrast, McBride et al. (39)

showed an increase in the number of infiltrating CDS+ T-cells relative to CD4+ T-cells

within FSA fibrosarcoma turnors expressing IL-7, and splenocytes fiom animals that had

been vaccinated with FSA expressing IL-7 were able to specificaily lyse parental FSA

tumor cells in vitro, giving evidence for CDS+ T-celi mediated effects.

Various murine -or modeis have used EN-y expressing tumor cells to induce rejection

of tumor variants and induce and subsequent protection toward wild type challenge.

Predominantly, CD8+ T-cells have been implicated as the effector subset in the

antitumorai response, in models including a neuroblastoma C 1300 (5 1 ), a fibrosarcoma

CMS-5 (33), an adenocarcinoma SPI ( 3 9 , a colon carcinoma CT26 ( 3 9 , a lung

carcinoma 3LL (34) and a bladder carcinoma MBT-2 (47). In a number of these studies,

MHC class 1 upregulation was observed on the tumor cells themselves. upon expression

of the recombinant cytokine.

Working with a weakly immunogenic fibrosarcoma CMS-5, Gansbacher et al. (33)

showed that upon injection of mice with IFN-y expressing tumor variants, splenocytes

exhibited specific cytotoxicity against parental cells over a set of discrete tirne points.

Furthemore, an in vivo protective effect was seen upon injection of survivors with

parental unmodified CMS-5 cells. Porgador et al. (34) observed that EN-y expression

via retroviral infection of two poorly immunogenic clones of the lung carcinoma 3LL

turned the clones into high expressors of H - ~ K ~ and caused a specific decrease in

turnorigenicity and metastatic growth. In addition, irradiation of IF%y expressing clones

and their subsequent injection into mice induced significant protection against parental

cells, while injection of IFN-y expressing clones almost cured mice canying already

established micrometastases. Esumi el al. (35) showed also that while IFN-y induced

MHC 1 upregulation is necessary for tumor rejection, MHC upregulation by itself is not

sufficient and other downstream effects of IFN-y are integral in making tumor cells more

immunogenic.

ThrF- a

The primary source of TNF-a is usualiy considered to be the macrophage / monocyte, but

various other ce11 types aiso secrete it, including neutrophils, T and B lymphocytes, NK

cells as well as endotheliai cells (49). TNF-a exerts a wide variety of effects on diverse

ce11 types. and plays an important role in defense against infection and tumor growth (7).

Asher et al. (42), working with the MCA 205 sarcoma, showed that turnors expressing

TNF-a regressed in animals after an initial phase of tumor growth and that this regression

was mediated by CD4+ and CD8+ T-cells. Blankenstein et al. (43) demonstrated that

J558L plasmacytoma tumor establishment was significantly delayed with TNF-a

expression; however, in this case, this effect was blocked by an antibody to type 3

complement receptor (CR3) which blocks migration of inflammatory cells like

Macrophages.

GM-CSF

A number of early studies involving vaccination models with cytokine gene modified

cells showed that expression of certain cytokines were capable of augmenting T-ce11

rnediated immunity, defked by systernic protection of vaccinated animais after challenge

with wild type cells. Because it is known that some tumors are inherently immunogenic,

it has been postulated that instead of just having a specific effect of T-ce11 activity,

cytokines might aiso act nonspecificdy by killing the tumor cells and rendering them

immunogenic.

Dranoff et al. (8) addressed this issue speciflcally for the BI6 melanoma twnor mode1

where they made stable expressors of IL-2, 4, 5, 6, GM-CSF, IFN-y and TNF-a, and

vaccinated mice with live tumors expressing one of these cytokines. Delays in tumor

formation were observed with expression of IL-4,6, IFN-y and TNF-a. while only cells

expressing IL-2 were completely rejected. Expression of IL-6 eventually resulted in

death of animals, as did expression of IL-5 and GM-CSF.

Interestingly, while challenge of mice protected via vaccination with IL-2 expressing

tumor variants resulted in 100% mortality, challenge of mice vaccinated with variants

expressing iL-2 and GM-CSF induced systemic protection. This led to a postulation that

while IL-2 expression mediated a local antitumor response, expression of GM-CSF

mediated a more systemic effect. Using irradiated B 16 cells, Dranoff et ai. showed that

vaccination via GM-CSF expression did in fact lead to systemic ïmmunity against wild

type celis, while irradiated cells by themselves did not. Thus, irradiated cells conferred a

level of protection equivalent to other cytokines, while only GM-CSF induced complete

protection. Depletion via antibody administration showed CD4+ and CD8+ celis to be

primarily responsible for tumor rejection.

IL-12

IL- 1 2, originally called cytotoxic T-lymphocyte Maturation Factor (CLMF) is a

heterodimerïc cytokine cornposed of two subunits of 45 KDa and 35 KDa. It was

originally isoIated by Stem et al. (52) on the basis of its ability to synergize with IL-2 to

facilitate cytotoxic lymphocyte reactions, and act as a growth factor for PHA activated

human T-lymphoblasts. At about the same time, Kobayashi et al- (53) had also identified

the IL-1 2 activity which they designated Natural Killer Ce11 Stimulatory Factor W S F )

on the basis of its ability to induce production of IFN-y by resting PBMC and act as a

proliferative stimulus in combination with PHA and phorbol esters.

Afier the murine cDNA for IL-12 was cloned (54), Gately et al. (55) showed that in vivo

administration of recombinant IL-12 caused a dose dependent enhancement of NK ce11

lytic activity when assayed in vitro. In addition, IL-12 treated mice had elevated Levels of

s e m IFN-y. Lastly, rnice immunized with allogeneic splenocytes displayed enhanced

CTL activity against these cells when IL-12 was administered concomitantly with the

irnmunogen. This showed that IL-12 can enhance NK and CTL activity, and induce IFN-

y expression in vivo. Gennann et al. (56) also demonstrated that IL-12 could induce

proliferation of Thl cells activated via IL-2 or anti CD-3, but not Th2 cells.

One of the initial cancer therapy studies with IL-12 was performed by Zitvogel et al. (40)

who showed that injection of IL42 expressing fibroblasts at the site of an established

sarcoma MCA 207 could suppress tumor growth in a dose dependent manner, or

eliminate it, and induce long term irnrnunity towards challenge. Furthemore, IL-12

delivery by irradioted fibroblasts at a distant site led to efficient rejections of established

tumors. Immunohistochemical staining revealed that for mice treated with IL-12

expressing fibroblasts an enhanced number of CD4+ T-cells, CDS+ T-cells and

Macrophages were present at the tumor site.

Tahara et al. (41) subsequently used the MCA 207 sarcoma and showed that genetic

modification of the tumor ceil via retroviral expression of IL-12 also induced tumor

rejection when injected intradennally in mice. Furthemore, a protective immunity was

observed in most anirnals challenged on the opposite flank with wild type cells. Three

day old established tumors could also be cured by IL-12 expressing tumor cells

inoculated at a distal site. Depletion experiments showed both CD4+ and CDS+ T-cells

to be important for long term imrnunity.

Combination Therapies UtiliWig; IL- 1 2 and B7- 1

The synergistic effects of the costimulatory molecule B7- 1 and the cytokine IL- 12 were

initially observed by Kubin et al. in a senes of in vitro experbents involving both

mitogen activated, as well as peripheral blood T-cells. In this report (57) Kubin et al-

show firstly, that 5 days following activation via PHA, culture of T-ce11 blasts with IL42

and anti-CD28 induced proliferation that was far superior to that induced by IL- 12 alone.

Secondly, cytokine profiles were significantly altered as PHA blasts which were

stimulated with IL42 and CHO cells expressing B7-1 showed increased IFN-y

production relative to cells that were stimulated with only one of these agents, or with

anti-C D3 in conj unction with 87- 1. Thudly, human peripheral blood T-ly mphocytes

which were incubated with anti CD28 and IL-12 showed hi& IFN-y production, while

those incubated with either one showed baseline levels. Finally, addition of CTLA-4 Ig

to both PBMC as weil as PHA blasts reduced the level of IFN-y produced after

incubation with IL- 12 and CHO cells expressing B7-1.

Initial antitumor models utilinng IL-1 2 and B7- 1 were established by Coughlin et al. (58)

who showed that while SCK mammary carcinoma cells expressing B7- 1 were rejected in

only 28% of mice, and intraperitoneal administration of IL- 12 only induced a delay in

wild type tumor onset, the combination of rIL-12 administration with B7-1 gene

modification in tumors induced rejection in 92% of mice. Further, while wild type cells

could be effectively eradicated by administration of SCK-B7- 1 cells on the opposite flank

along with rIL-12 administration, rIL-12 by itself had no effect. Antibody depletion

experiments showed CD4+ and CD8+ T-cells to be responsibIe for wild type tumor

rejection.

Established turnor modeis were more successfully treated by Rao et al. (59) who showed

firstly, that even though 3 or 6 day old lung metastases of CT26.CL25 (adenocarcinorna)

could be effectively treated by administration of recombinant vaccinia virus (rVV)

encoding the mode1 tumor antigen P-gal, addition of rIL-12 greatly increased the

therapeutic effect. A M e r significant delay in w o r onset was observed if B7-1 was

also expressed by the r W expressing P-gal, and rL-12 was adMnistered in mcie with

established tumor burdens. This showed that B7-1 synergized with I L 4 2 in effectively

activating T-ceUs against pre-existing turnors expressing the p-gal model antigen.

Augmentation of B7- 1 mediated immunogenicity via intratumorai IL- 12 gene expression

was demonstrated by Pizzoferrato et al. (60) who showed that while A20 cells expressing

B7-1 were delayed in tumor formation (relative to wild type cells), A20 celis expressing

both IL- 12 and B7-1 were completely rejected in wiid type mice. Furthemore, 70% of

the surviving animals displayed systemic immunity in that they rejected a challenge of

wild type A20 cells on the opposite flank. Interesthgly, in this case, vaccination with

A20 cells expressing IL- 12 only also led to 100% tumor rejection, and an equivalent level

of protection against wild type challenge was observed. A synergistic effect between IL-

12 and B7-1 was observed in the established tumor model, whereby wild type tumors

could be cured by A20/IL-12 or A20/B7-1 AL-12 when the vaccines were injected into

the tumor site, but only by A20AL-12A37-1 when they were injected at a distal site.

Tumorigenicity experiments in nude mice demonstrated that T-cells were responsible for

tumor rejection, as A20 cells expressing IL-12 or IL-12 and B7-1 lost their

immunogenicity in these immunocompromised mice.

Gene Transfer Methodologies

A major issue in designing a gene-modified tumor vaccine, or any other DNA-based

vaccine is the type of vector to be used in the expression of recombinant constructs. The

choice of vector often depends on the approach of therapy and the type of response that is

desired. The methods of recombinant gene expression can be broadly broken up into

viral and non-viral (or physical), as discussed below.

Phvsical Methods of Gene Trader

Plasmid DNA Immunirarion

Many of the original studies examining the effects of in vitro genetic modification of

tumor cells on tumorigenesis and subsequent immunity, utilized plasmid DNA as the

vector expressing the gene of interest. These genes included IL-2 (44,45), IL4 (3 l), and

IFN-y (5 1) as well as costirnulatory molecules such as B7- 1 (6), B7-2 (1 9, and 4- 1BBL

(23). In addition, a study comparing the efficacy of 87-1 and IL- 12, both of which were

intratumorally encoded by plasmid vectors, was cmied out by Fallarino et al. (61). They

showed that while immunization with irradiated P8 15 variants expressing either IL- 12 or

B7-1 protected against wild type challenge, established tumors could only be cured by

injection of variants expressing IL-12 and not B7-1.

Because plasmids are not as constrained as most viruses in the exact arnount of DNA that

they need to incorporate, they are good candidates for gene transfer strategies utilizing

varying lengths of cDNA9s. As well, these vectors eliminate the need for working with

live viruses, minimizing health risks, as well as non specific immune responses in the

immunized host. The disadvantage associated with plasmids as vectors is the relatively

low eficiency of ceil transfection. This is particularly significant in the context of cancer

vaccination where stable ce11 luies cannot always be established, and primary cells need

to be transfected at eficiencies as high as possible.

To circumvent the problern of low transfection efficiency and longer periods of time

required to establish variants, an alternative strategy using a "gene gun" has been

developed (50). This strategy involves the coating of the Foreign DNA of interest

through precipitation, and bombarding the tissue of interest with these particles through

use of a helium powered accelerator. Following delivery, the DNA dissolves within the

aqueous environment of the cytoplasm and is available for expression.

In vitro modification of tumor cells via plasmid acceleration through a gene gun was

demonstrated by Mahvi et al. (62). In this report, Mahvi et al. initially bombarded an

adherent layer of B16 melanoma with DNA encoding GM-CSF, and obtained high

expression levels of this cytokine (1000 - 4000 ng/milIion celis/24 hrs). Further, a

vaccination of mice by these cells protected 58% of mice against challenge with wild

type B 16 cells. Finally, heterogeneous samples of primary human turnor ceils could be

successfi.dly transfected with hGM-CSF, upon homogenization of the tumor samples and

subjecting to DNA bombardment within 24 hours of surgical removal. Expression levels

of GM-CSF varied, depending on the patient and type of cancer, and ranged fiom

negligible to 140 ng/million celld24 hours. Lower levels of expression were attributed to

the clona1 nature of B16 (and an inherent ability to express cytokines at high levels),

greater ce11 death in primary cells, differential transfection efficiencies dependhg on

tumor type and stroma1 elements, as well as cytoplasmic size whch restricts the size of

the gold particles that can be incorporated.

Thus, genetic manipulation via plasmid DNA is an attractive approach for in vitro

modification of tumor cells, though traditional methods involving transfection may not be

as efficient. Further, approaches using DNA vaccines that have been implemented

include direct in vivo injection via gene p in murine models (63) and encapsulation

within lipids and direct intratumoral injection into patients (64,65).

Viral Methods of Gene Tmnsfer

Retrovirus

Retroviruses are double stranded RNA viruses which make specific contact with target

cells through surface envelope proteins on the viral capsid. Following initial binding, the

viral envelope fuses with the ce11 membrane and the RNA molecules enter the cell. The

encapsulated virai RNA polymerase then copies the RNA genome into a cDNA

molecule, which integrates into the cellular genome to fonn a provirus (7). The viral

genorne can be read in overiapping fiames, and the main products are gag (structural

protein), pol (reverse transcriptase), and env (trammembrane glycoprotein). The virai

DNA is flanked at either end by long terniinal repeats (LTR) and has a RNA polymerase

II promoter site at the 5' LTR (50).

Because retroviruses in their natural form are infectious and pathogenic, a two

component approach is used for gene therapy purposes. The recombinant gene is cloned

into the sites for the gag, pol, and env proteins, and the cDNA of interest is driven by

elements in the 5' LTR (50). Polycistronic messages can also be cloned, utilizing an

interna1 ribosome entry site (IRES) within the message (66). The recombinant retrovirus

can then be introduced into a packaging ce11 line which then provides the proteins that

cannot be encoded by the recombinant virus (66). This yields functional retroviruses

which are able to infect target cells, integrate into the host genome and express the

foreign product, but are unable to complete the lytic life cycle.

Many of the initial murine studies discussed above, involving cytokine gene-modified

tumor cells, used retrovimses as the vector expressing the gene of interest. These genes

included IFN-y (33), TNF-a (43), and GM-CSF (8). Subsequently, tumor cells

engineered to express retrovually encoded IL- 12 were also employed as a murine vaccine

by Tahara et al. (41). This vector encoded recombinant p40 and p35 subunits of IL-12 as

well as neomycin as a polycistronic message within the MCA 207 and MCA 102

sarcomas. Animals receiving this tumor vaccine rejected their tumor and were

significantly protected against challenge, while animals bearing tumor cells with

retrovirus encoding only neomycin succumbed to tumor formation.

Thus, recombinant retrovinises are aitractive as vectors for active immunotherapy as they

have higher fiequency of ce11 transformation than plasmids, and can generate high levels

of protein expression. In addition, the viral vector itself is non irnrnunogenic, and

therapeutic effects are specifically a fhction of the recombinant genes expressed.

The retrovirus is integrative within the host genome, and while this has the advantage of

conferring stable expression for the duration of the treatment period, it carries the

disadvantage of coderring constitutive, and even toxic levels of expression if the cells

carrying the virus are not eventually destroyed. Furthermore, because of this in tept ive

property, the retrovinis is mutagenic to the ce11 line it is introduced into. Cells also need

to be mitotically active in order to be infected by retroviral vectors. Finally because of

the compact nature of the retroviral genome, recombinant genes only up to about 9 Kb

c m be inserted (50).

Adenovirus

The limitations of the retroviral vector, namely, constitutive gene expression, requirement

for mitotically active cells, and mutagenicity toward host cells, have led to the

development of another viral vector that circumvents these disadvantages. Adenoviruses

have a 36 Kb DNA genome that encodes four early proteins (El to E4) and five late

proteins (L 1 to L5) @O), and like the retrovirus can accommodate approxirnately 8 Kb of

recombinant genetic material (67). Adenoviruses can be rendered replication defective

by deletion of the E 1 region of the genome (68), thus requiring a complementary ceil line

(293) to provide the necessary protein products in tram. Further deletion of the E3

region allows the incorporation of additional recombinant sequences.

Unlike the retrovinis, adenoviruses can be grown to hi& titers (69) and they c m infect

many cell types regardless of mitotic status (70). Furthermore. the virus exists

episomally, eliminating the risk of mutagenesis in the host cell. Also, because the virus

does not integrate into the host genome, and is replication defective, the recombinant

genes are expressed transiently (71). This eliminates the chance of toxicity that may be

incurred by prolonged expression of molecules like cytokines. The only disadvantage

associated with Adenovirus is that E l deteted vectors have been s h o w to be

imrnunogenic by thernselves (72). Repeated administration thus becomes difficult as the

immune system reacts specifically to the virus and the advantageous effects of the

recombinant gene are not obtained.

Despite the one disadvantage, the Adenovirus is still an efficient vector for antitumor

therapy. Warnier et al. (73) showed that a specific CTL response could be elicited

against the murine mode1 tumor antigen P8 15A, after mice were immunized

intradermally with adenovirus encoding this molecule, and their splenocytes were

restimulated with cells expressing P8 1 SA and B7-1. Adenoviruses expressing the

irrelevant protein B-galactosidase did not cause mice to develop a CTL response, even

following in vitro stimulation.

Adenoviruses encoding the recombinant cytokine IL-12 were used in antitumor therapy

by Bramson et al. (74) who showed that direct intratumoral injection of a mammary

adenocarcinorna with Adenovirus-mIL-12 produced significant tumor regression, and

animais who completely rejected their tumors were immune to secondary challenge. A

transient expression of IL- 12 was observed as expected, with IL- 12 levels being highest

between two and three days p s t injection.

Poxviruses

Poxvimses comprise a large family of DNA Wuses that have a wide range of hosts, both

vertebrate and invertebrate (75). The most well known poxviruses are variola,

responsibie for smallpox, and vaccinia, the virus used for vaccination purposes a g a k t

the former, more pathogenic agent (76).

In contrast to retrovinses and adenoviruses, poxviruses have a large complement of

genetic material, approximating about 200 - 300 Kilobase pairs, depending upon the

particular virus (76). The genome of the rnost well characterized poxvims, vaccinia, has

been sequenced (77), and it has been postulated to encode about 260 distinct proteins,

rnost of which are of unknown iùnction, Other poxviruses including canarypox, cowpox

and fowlpox, are thought to encode comparable numbers of genes (76).

Morphologically, poxviruses are roughly brick shaped or oval shaped virions about 400

nrn in diameter (76). Structurally, viruses are encapsulated at the outermost surface by a

network of surface tubules, followed by an outer membrane (75). At the core of the

virion is a nucleoprotein complex, which is flanked by two lateral bodies of wiknown

fünction. The entire genome is comprised of one linear DNA molecule with the single

stranded ends ligated together to form a hairpin loop at either end of the molecule. Also

present within the vaccinia nucleoprotein core are intact enzymes, including a

multisubunit RNA polymerase, a transcription factor, a poly A polymerase, a capping

enzyme, a methyltransferase and a DNA topoisornerase (76,78)-

Cunaryp0.r Virus (AL VAC)

The canarypox virus is a member of a related genus of poxvimses known as

uvipomiruses. Like vaccinia, the canarypox virus also has a large genome, measuring

about 325 Kb (79). DeveIoped at Virogenetics Corp. (Troy, New York), canarypox virus

vectors are appealing as vectors for gene therapy due to a number of reasons. Firstly,

they are able to infect a wide variety of ce11 types (80) with very high efficiencies of

infection (8 1 ). Also, like other poxvinises, the canarypox vectors can accommodate large

genes of up to 25 Kb (82). In addition, poxviruses exist cytoplasmically and so are not

mutagenic toward the host cell that they infect (76, 78). Finally, the added advantage of

the canarypox virus over other poxviral vectors is that it is replication defective within

mamnialian cells (80). While the v i n s can be maintained and propagated eficiently in

chicken embryonic fibroblasts (CEF), the canarypox virus life cycle is aborted in

mamallian cells before viral DNA replication takes place (83). The advantage of being

naturally defective is two fold: firstly, the virus cannot be pathogenic to the entire host

system once it has been targeted to a specific ce11 type, and secondly, genes are only

expressed transiently as the virus is diluted out and eventually lost in continually dividing

cells such as tumors. This helps to avoid chronic toxicities associated with continued

presence of protein products like cytokines. In summary, the advantages of wide tropism,

larger genome, episomal expression and replication defect make ALVAC an attractive

vector in gene therapy as an alternative to other vectors. The disadvantages associated

with retroviruses, such as the requirement for cycling cells, host mutagenicity and

permanent expression are overcome by this vector. Furthemore. an added advantage of

using ALVAC over other established Adenovims vectors is that the size of the

recombinant inserts may be increased to 25Kbp compared to about 8Kbp in

Adenovinses.

Generation of Recombinant ALVAC viruses

The ALVAC virai genome consists of a double stranded linear DNA molecule measuring

apporximately 325Kb (79). Although the genome has been characterized via XhoI

restriction mapping, a complete sequence is yet to be published. Recombinant cDNA

sequences expressed via ALVAC vectors are driven by promoters that are derived fiom

well characterized vaccinia virus genes (80). Specifically, for ALVAC B7-1 the human

B7- 1 cDNA is driven by the vaccinia virus prometer H6, and the expression cassette is

cloned into the C6 multiple cloning site. For ALVAC IL-12, expression of the p35

subunit is driven by the vaccinia E3L promoter and the expression of p40 is driven by

the 42K entomopox promoter. The expression cassettes for p35 and p40 are cloned into

the C5 and C6 multiple cloning sites, respectively (see figure 1).

Titration of the ALVAC viruses

The procedure used to titer ALVAC viruses has been described in detail by Puisieux et

al. (79). In this procedure, dilutions of viral Iysates are added to monolayers of chicken

embryo fibroblast cells (CEF) and incubated at 37"C, 5% COz. Subsequently, viral

media is removed and media containing 1.2 % agarose is added. 3 days later, media

containing 1 -2 % agarose is added again, 24 hours after which plaques are visualized and

counted.

The ALVAC Virus in Cancer Therap

The use of recombinant canarypox virus in cancer immunotherapy is gradually becoming

widespread. Toso et ai. (84) demonstrated that the ALVAC virus was a suitable vector in

providing stimulation of tumor infiltrating lymphocytes ex vivo. In this study , kadiated

pexipheral blood mononuclear cells infected with ALVAC encoding the tumor antigen

MAGE-1 specifically stimulated tumor infiltrating lymphocytes fiom a breast cancer

patient to proliferate. In addition, these lymphocytes recognized an allogeneic B-cell

lymphoma expressing vaccinia encoded MAGE- 1, but not the same ce11 line bearïng wild

type vaccinia virus.

The fact that ALVAC encoded proteins could also serve as effective antihunoral vaccines

was first shown by Roth et al. (85). In this report, mice imrnunized subcutaneously with

ALVAC expressing mutant human p53, or wild type hurnan p53, were significantly

protected against challenge with a fibroblast ce11 line 1 0(3)273.lNTX, which expressed a

mutant human p53 protein. Interestingly, irnmunization with ALVAC encoding mutant

murine p53 or wild type murine p53 also induced significant protection upon challenge

with the same ce11 line. in either case, subcutaneous injection with the non recombinant

ALVAC vector did not confer any protection.

Subsequently, Hodge et al. (86), working with ALVAC recombinants expressing the

tumor antigen CEA showed that immunization of mice intramuscularly elicited

lymphoproliferative as weH as cytolytic T-ce11 responses. Furthemore. significant titers

of ant i CEA IgG was observed in senim- Finally, 70% of mice which were immunized

three times with ALVAC CEA rejected a tumor challenge of the MC38 colon carcinoma,

which was expressing CEA. Injection of this ce11 line in animals receiving an irrelevant

vaccine (ALVAC RG (rabies glycoprotein)) resulted in turnor formation.

Very recently, in a clinical trial conducted by Marshall et al. (87), it was shown that after

vaccination of patients with ALVAC-CEA, an increase in fiequency of cytotoxic T-ce11

precursors specific for CEA could be observed in 7 out of 9 patients. ïhese patients al1

carried the K A - A 2 allele, and afier three intramuscular vaccinations with ALVAC-CEA

their peripheral blood T-cells were able to recognize a HLA matched ce11 line which

expressed CEA.

Active Immunothemy with Turnors Geneticallv Modified via ALVAC Vectors

In an expcrimental approach similar to the one undertaken in this project, Kawakita et al.

(81) looked at the in vivo growth of the RMI murine prostate cancer model, after

infection with recombinant ALVAC vectors. Cells were infected with ALVAC IL-2,

ALVAC B7-1, ALVAC IFN-y, or ALVAC TNF-a, or no virus at all, and injected

subcutaneously. Tumor growth was unaltered relative to cells alone in al1 cases except

with ALVAC TNF-a, where a delay in tumor growth was observed. Complete tumor

rejection was only observed with a combination of viruses encoding MF-a and IL-2,

while other combinations with ALVAC TNF-a had no effect-

To determine whether this prirnary rejection with cells bearing ALVAC TNF-a and

ALVAC IL-2 conferred imrnunity systemicaily, animals were vaccinated with irradiated

cells bearing the two recombinant vimses. and then challenged with wild type cells d e r

ten days. No protection, or delay in tumor formation, was observed after chdlenge.

SCID mice, which have no mature T or B cells also rejected a tumorigenic dose of RMI

cells infected with ALVAC IL-2 and ALVAC TNF-a, indicating that primary tumor

rejection is not mediated by specific immune effectors. In parailel. no CTL activity

against RM 1, or recombinant ALVAC infected RM 1 -variants, was observed with

splenocytes fiom wild type mice that were vaccinated with RM-I bearing ALVAC IL-2

and ALVAC TNF-a.

More recently, Puisieux et al. (79) showed that infection of TSIA marnmary

adenocarcinorna cells with ALVAC recombinants generated some nonspecific immunity

when tumor variants were injected subcutaneously. Cells were infected with ALVAC B-

galactosidase, ALVAC IL- 12, ALVAC IL-2, ALVAC GM-CSF, or ALVAC IFN-y, or no

virus at all. Significant survival was observed only in mice receiving cells bearing

ALVAC IL-12 at 30 days post injection. Interestingly, there was also a delay in tumor

formation in animals receiving cells bearing ALVAC Pgal, relative to animais receiving

cells alone.

A subsequent approach involved multiple intratumoral injections of pre-established 7-

day tumors with ALVAC IL-12 virus, or ALVAC P-gal, or PBS alone. At 46 days,

animals being treated with PBS had dl died, while animals given ALVAC IL- 12 survived

at a frequency of 100%. Surprisingly there was dso a 30% survïval in mice receiving

ALVAC P-gal.

In conjunction with treatment of established tumors with recombinant ALVAC vectors,

mice that displayed a non-necrotic tumor were fürther challenged with a lethal dose of

wild type cells on the opposite flank. 70% of mice that were receiving ALVAC IL12

within their original tumor cleared the contralateral challenge completely. Again,

ALVAC P-gal also conferred some protection as 30% of mice receiving ALVAC P-gal

into the original himor cleared the contralateral challenge.

Irnrnunohistochernical stainuig of tumors treated with ALVAC p-gal and ALVAC IL- 12

revealed a dense infiltrate of MAC-1 positive cells, indicating a presence of monocytes or

macrophages. In addition, in tumors treated with ALVAC IL-12, CD4+ and CDS+ T-

cells were also observed.

Injection of TS/A twnor cells into Nude mice proved to be lethal in al1 three scenarios of

treatment consisting of injections with ALVAC IL-12, ALVAC B-gal or PBS. Thus, in

contrast to the study by Kawakita et al. cited above, this experiment showed a complete

dependence on T-cells for tumor rejection.

The focus of this project was to characterize the mode of action of the ALVAC virus

vector within a lymphoid tumor model. Generally, we airned to observe whether a live

tumor vaccine bearing ALVAC done, ALVAC B7-1, ALVAC I L 4 2 or both ALVAC

B7- 1 and ALVAC IL- 12 would induce an anti-tumoral reaction, through specific T-ce11

activation and if so, whether prirnary rejection of tumor variants would confer systemic

irnmunity toward wild type challenge. The approach specificaily incorporated the

molecules B7- 1 and I L 4 2 for two reasons: fustly, both IL- 12 as well as B7- 1 are known

to have potent effects in specifkally augmenting T-ce11 activation andior killing through

unique mechanisms. Activation of tumor specific T-cells via these molecules may in turn

Iead to the formation of T-ce11 memory against hunor cells, facilitating systemic

irnmunity. Secondly, previous models established in our laboratory have shown the

efficacy of tumor vaccines expressing both IL-12 and B7-1, through plasmid and

retroviral expression vectors. Using the sarne molecules in the context of a novel viral

vector allows us to assess the differential effects of the viral vector aione in augmenting

cancer therapy.

For the purpose of this experirnent, we used the early T-cell lymphoma model STFI O.

We hypothesized that the ALVAC encoded molecules of B7-1 or IL-12, or both, would

indeed render the tumor cells more immunogenic, while the vector aione would not affect

the intrinsic properties of the tumor cells. Our in vivo experùnents show that frrst of all,

the ALVAC virus by itself is 100% immunogenic in that STFlO tumor cells bearing the

non-recombinant virus, are completely rejected in immunocompetent mice. Secondly,

the parental as well us the recombinant ALVAC vectors expressing B7-1 or IL-12 can

induce a protective effect against wild type tumor cells, when the initial cellular vaccine

is boosted before challenge. Thirdly, cytotoxic effectors specific for wild type tumor

cells can be detected in mice that have been immunized and boosted with

STFI O/ALVAC. Finally in athymic nude mice, STFlO tumor cells bearing ALVAC

vectors are rejected to a significant extent, even though not completely as in wild type

rnice. This indicates that primary rejection of tumor variants is mediated by both T-cells

as well as T-ce11 independent mechanisms.

Material and Methods

Ce11 Lines

The mature B-ce11 lymphomas A20, K46J, M12, the melanoma B16-F 1, the pro B-ce11

line NFS-70 and the mastocytoma P8 15 were purchased fiom American Type Cuiture

Collection (ATCC). The SCID thymorna Iule ST was derived from a spontaneous tumor

in a C-B-17 SCID mouse and was kindly provided by Dr. Gillian Wu (Ontario Cancer

Institute, University of Toronto, Toronto, Canada). The acute myeloid leukemia ce11 line

C-1498 was kindly provided by Dr. Andre Schuh @epartment of Medical Biophysics,

University of Toronto, Toronto, Canada).

AI1 cell lines were maintained in culture media (CM) of RPMI 1640 + 10% Fetal Calf

Serum (FCS) + 50 u M 2-mercaptoethano1, at 37OC, in 5% COz, with the exception of C-

1498 and B6-FI which were maintained in DMEM + 10% FCS.

To establish subclones, SClD Thymoma (ST) cells were plated into 96 well plates at

concentrations of 30, 3 and 0.3 cells / well in complete media. Plates corresponding to

each dilution were set up in duplicate. After 14 days, 8 ST subclones were observed on a

plate corresponding to 0.3 cells/well, and these were individually expanded. Two

representative clones were designated STF IO and STG4, and maintained in conditions

identical to those for the parental culture. B 16-FI was similarly cloned via limiting

dilution, and one representative subclone was designated B 16-F 1.2. B 16-F 1.2 was

maintained under conditions identical to those for the parental culture.

Infection of ce11 lines via ALVAC vectors

Al1 parental and recombinant ALVAC vinises were acquired fiom Virogenetics Corp.,

Troy, New York. For infection, ce11 lines except 816-FI were plated at a concentration

of 5 X 10' cells / mL in RPMI 1640 media with 2% Fetal Caif Serum (FCS). C- 1498 was

plated at the same concentration, using DMEM + 2% FCS, while B16-F1 was plated at

50% confluency within DMEM + 2% FCS. ALVAC B7-2 (vCP 1334, Virogenetics,

New York) was added at a rnultiplicity of infection MOI) of 20 P h / celL Cells were

incubated for 10 hours at 37°C and 5% COî, after which they were washed in semm fiee

media, and recovered overnight in their respective culture medium (CM) at 37°C and 5%

coz.

Flow Cvtometric Assessrnent of Ce11 Surface Marker Emression

For indirect immunofluorescent analysis, 1 x 106 cells were centrifuged for 8 minutes at

800 rpm, washed twice in phosphate buffered saline (PBS), and then treated with 1 ug of

the indicated primary antibody. An equal number of control cells were treated with 1 ug

of isotypic control antibody. Cells were incubated on ice for 30 minutes, and then

washed twice in PBS. Subsequently, cells were treated with polyclonal fluorescein

isothiocyanate (FITC) - conjugated goat anti mouse IgM / IgG secondary antibody. Cells

were then washed twice with PBS and then resuspended into 400 pL of 2%

paraformaldehyde / PBS fixative.

For direct immunofluorescent analysis, ceIls were trested with the indicated

Phycoerythrin (PE)- or Fluorescein isothiocyanate (FITC)- conjugated antibody

(Pharmingen), and a corresponding PE- or FITC- conjugated anti TNP isotypic control

antibody. Again, cells were incubated on ice for 30 minutes, and then washed twice with

PBS, after which they were fixed in 400 pL of 20/0 pparaformaldehyde fixative.

Antibodies used for stalliing included purified hamster IgG anti CD3, rat (r) IgG2b anti

CD4, rIgG2a anti CD8, mouse (m) IgG2a anti H - ~ K ~ D ~ , mIgG2a anti 1 - A ~ , rIgG2a anti

B7- 1, mIgM anti h.7-1 , rIgG2a anti B7-2, polyclonal FiTC-conjugated goat anti mouse

IgM/IgG, or FITC-conjugated Goat anti hamster IgG. CD44, CD25 and HSA were

assayed for by treating cells with PE-conjugated dgG2b anti CD44, rIgG2b anti CD25, or

FITC conjugated rIgG2b anti HSA. Al1 antibodies as well as non-specific isotypic

controls were purchased fiom Phanningen, except for anti CD3, FITC-Goat anti hamster

IgG and the corresponding isotypic control, which were purc hased fiom Cedariane.

Fluorescence emitted by specifically bound antibody was measured by flow cytometry on

a FACScalibur cytometer (Becton Dickinson) and data were anaiyzed using CellQuest

Software eecton Dickinson).

Immunoassav for IL- 12 Production

Prior to in vivo inoculation of mice with STFlO vaccines bearing parental and

recombinant ALVAC vectors (see below) supernatant: from virally infected cells were

assayed for the presence of IL-12 via Enzyme Linked IrnrnunoSorbent Assay (ELISA),

using the OptEIA kit (Pharmingen) according to manufacturer's protocol. Al1 values

presented in Table 1 are expressed as ng/mL./million celW24 hours.

Reverse Transcri~tion (cDNA synthesis)

10' cells from indicated samples were spun d o m and cellular mRNA was extracted

using a RN-easy Kit (Qiagen) according to manufacturer's protocol. For reverse

transcription, 2ug of RNA was used fiom each sample.

5 pL (2 ug) of RNA was added to 1.5 p..L of random hexarner (300 ng / uL), along with

5.5 pL of sterile water, and incubated at 65°C for 5 minutes. Subsequently, 4 pL of 1"

strand 5 x buffer, 2 pL of 0.1M DTT, 1 pL of 20 mM (WTP, and 0.5 pL of RNase

Inhibitor (RNguard) was added on ice, and then incubated at room temperature for 15

minutes. Each sample was done in duplicate. Following incubation at room temperature,

one of the duplicates fkom each sample was treated with 1 pL of Superscript II reverse

transcriptase (for +RT reactions). Samples were incubated for 1 hour at 37°C followed

by 1 hour at 48OC. Superscript II was then deactivated by incubation at 95°C for 5

minutes. mase A kvas then added to each tube, and samples were incubated at 37°C for

30 minutes, followed by incubation at 95°C for 5 minutes. The final volume of each

sample was then brought up to 40 pL with steriie water.

PCR for aene - expression

2.5 pL of the sense primer, and 2.5 pL of the antisense primer (both initially at 2.5 pmol /

uL) tvere added to 3 pL of indicated cDNA template, along with 5pL of 10 x PCR buffer,

5 pL of 10 rnM dNTP and 0.25 pL of Taq polymerase (Qiagen)- For TdT cDNA

amplification, samples were denatured for 5 minutes at 94OC followed by 30 cycles of

denaturation at 94°C for 30 seconds, reannealling at 61°C for 30 seconds and extension at

72°C for 30 seconds. Subsequently, a h a 1 extension took place at 72OC. For RAG-1,

RAG-2 and HPRT amplification, samples were denatured at 94°C for 5 minutes,

followed by 3 5 cycles of denaturation at 94°C for 30 seconds, annealing at 58OC for 48

seconds, extension at 72°C for 45 seconds. Subseq~entiy~ a final extension at 72°C took

place for 10 minutes.

Murine TdT cDNA was amplified to a 124 bp product using sense primer 171 with

sequence 5' ATATGCTTGCCAGCGAAGAACC 3' and antisense primer 172, with

sequence 5' GAG ATT TCA GTA CAG AGG ACG C 3'. TdT Primers were kindly

provided by Dr. Gillian Wu. Murine RAG-1 cDNA was amplified to a 545 bp product

using a sense primer of sequence 5' CCAAGCTCGAGACA'ITCTAGCACTC 3' and an

antisense primer of sequence 5' CTGGATCCGGAAAATCCTGGCAATG 3'. Murine

RAG-2 cDNA was amplified to a 471 bp product using sense primer 5'

CACATCCACAAGCAGGAAGTACAC 3' and antisense primer 5 '

GGTTCAGGGACATCTCCTACTAAG 3'. Murine HPRT was amplified using sense

primer 5' GCTGGTGAAAAGGACCTCT 3' and antisense primer 5'

CACAGGACTAGAACACCTGC 3'.

Positive control for TdT amplification comprïsed of a murine genomic sample from

mouse tail DNA. Genomic Positive controls contained 1 or 8 pL of 25mM MgClt, while

negative controls contained water in place of template. Positive controls for RAG- 1,

RAG-2 and HPRT consisted of two cDNA preps of the pro-B ce11 line NFS-70, while

negative controls contained water in place of template.

20 pL of amplification products of TdT, RAG-1, RAG-2 and HPRT were mixed with 4

pL of 6x loading bmer and electrophoresed on 1.5% agarose gel at 90 Volts for one

hour. A 1 Kb reference marker (Gibco BRL) was electrophoresed in parallel.

Growth Kinetics of ALVAC-infected STFl O cells

To detemine whether infection via ALVAC vectors affected the growth rate of STFf O

cells, 1 million STFlO cells were infected with ALVAC B7-1 as outlined above, and

following recovery, were replated at 2 X 10' cellslmL in CM (t=O). As a control, 1

million STF l O cells were treated similady, without ALVAC B7- 1 infection. Population

density (concentration in vitro) was then assayed at timepoints of t = 24, 48, 72 and 96

hours via trypan blue exclusion assay. Further analysis was not undertaken due to ce11

death in both cultures due to overwhelming population density.

Stabilitv of ALVAC Encoded ~roducts over time in ST

S t a b i l i ~ of ALVAC encoded products over tirne was assayed in both a continually

dividing culture as well as a non-dividing culture. 6 million ST cells were infected with

ALVAC B7- 1, as above. Subsequently, half the culture (3 million cells) was separated

and irradiated at 10,000 cGy and replated at 5 x 10' cells / mL in fiesh media. Post

irradiation, aiiquots were taken at 24, 48, 72 and 96 hours, and assayed for B7-1 as

described above. The cells did not divide M e r afier irradiation and afier 96 hours,

were completely non-viable. In pardlel, after infection and recovery, the non irradiated

cells were replated in fiesh media at 5 x 10' cells f mL and aliquots were takea at 24,48,

72, and 96 hours to assay for B7-1 expression. Since this culture was continually

dividing, cells were split and maintauied at 5 x 10' cells / mL at each of these timepoints.

Two M e r timepoints of 120 hours and 144 hours were assayed for 87-1 expression,

with this culture.

Tumorkenicip in Wild T v ~ e BALB/c rnice: Vaccination and Challenge

8-10 week old BALBk female mice were purchased from Charles River Laboratories

(Canada) and housed at the animal facility of the Division of Comparative Medicine,

University of Toronto. Al1 experiments were conducted in accordance with the

University of Toronto Animal Care guidelines.

STF l O celis were plated at 5 x 1 o5 cells / mL in RPMI 1640 + 2% FCS and idected in

vitro (as above) with the indicated ALVAC parental or recombinant virus, and incubated

for 10 hours at 37OC, 5% COz. Al1 infections were carried out at an MOI of 20, except

when combining ALVAC IL-12 and ALVAC B7-1 where an MOI of 10 was used for

each respective virus. Following infection, cells were washed twice and recovered

overnight in complete medium (CM).

Following recovery, cells were washed 4 times in serum fiee media, resuspended into

semm fiee media at a concentration of 5 x 106 cells / mL, and injected subcutaneously

into the right Bank of five corresponding groups of mice. A dose of 1 x 106 cells was

administered per mouse, with five mice per group. Tumor growth in vivo was monitored,

and animais were sacrificed when maximum tumor diameter reached 2 cm.

Surviving animals were challenged at day 55 with 1 x 1 o6 wild type STF 1 O cells on the

opposite (Ieft) flank. As a positive control for tumor growth, naive age-matched female

BALBIc rnice were also injected with wild type STF 10 cells.

To determine whether a vaccination regimen incorporating a cellular boost would induce

greater survival &er wild type challenge, a protocol utilizing two vaccines before wild

type challenge was undertaken. Mice were immunized with the indicated celIular

vaccines consisting of STF1O cells infected with parental or recombinant ALVAC

vectors. In this experiment, instead of using cells doubly infected with ALVAC 87-1 and

ALVAC IL-12 as one of the vaccine groups, a mixture of equal nurnbers of ALVAC B7-

1-infected and ALVAC IL-12-infected celis was used, 8 female BALBIc mice were used

per group, except in the group receiving STFlO/ALVAC IL-12, where 16 mice were

used.

34 days post vaccination, a new batch of STFlO cells were infected with parental or

recombinant ALVAC vectors- On day 35, animais nirviving the initial vaccination were

boosted with the exact same tumor vaccine (as they had received initially) via

subcutaneous injection on the same (nght) lateral flank. A boost of the same vaccine was

administered to al1 animals except those that had been immunized with STFlO/ALVAC

IL-12, where only 8 out of 16 were boosted, and 8 were lefi unboosted. A positive

control for tumor formation was included by the injection of wild type STFI O cells into 8

age-matched, previously unvaccinated BALBIc female mice.

On day 56 post initiai vaccination (or day 2 1 post boost), al1 rnice surviving the initial

vaccine as well as the boost (where applicable) were injected on the left lateral flank with

wild type STFlO cells at a dose of one million cells per mouse. Once again, positive

controls for tumor formation were included via the injection of wild type cells into

previousl y unvaccinated B ALB/c mice.

In experiments using athymic nude mice 8-10 week old fernale BALB/c ndnu mice were

purchased from Charles River Laboratories and maintained as above, except under sterile

conditions. A similac tumongenicity assay was perforrned, whereby STF 1 O cells were

infected with the indicated ALVAC virus, and injected subcutaneously on the right lateral

flank, at a dose of 1 x 106 cells per mouse. Agah, five mice were used per group, and

animals were euthanized when maximum tumor diameter reached 2 cm.

A follow-up tumongenicity experiment with nude mice incorporated 8 rnice per group,

with one modification, in that instead of doubly infecting one group of cells with

ALVAC B7-1 and ALVAC IL-12, a mixture of an equal number of ALVAC B7-1-

infected cells and ALVAC IL-12 infected cells, was used.

Statistical Analy sis

Kaplan Meier survival curves were compared using the log-rank test in the SPSS 6.1

Macintosh version statisticd package. Differences are considered statistically sigaiflcant

if p < 0-05.

Cvtotoxic T-lvmphocvte Assavs

Preparation of St imulator Cells

STF 1 O cells were plated at a concentration of 5 x 10' cells / mL in RPMI 1640 + 10%

FCS + 50 u M 2-Me (CM) and infected with ALVAC at an MOI of 20. and then grown

overnight at 37OC, 5% COz. Following incubation, ceils were washed twice in serum fiee

RPMI and resuspended in culture media (CM). Subsequently cells were irradiated at

10,000 =Gy, washed, and resuspended at 1 x 10' cells / mL in CM. 1 mL was

subsequently plated out in 24-well tissue culture plates.

Preparation of Eflecfor Cells

Because STF 1 OIALVAC IL- 12 consistently induced greatest antinimor protection after

one vaccination, mice were immunized with STFlO / ALVAC IL-12 via subcutaneous

injections and one mouse was euthanized at each timepoint of 13, 17,27 and 35 days post

injection to examine in vitro reactivity. In paraIlel, a naive mouse was also euthanized at

each timepoint. Animals were dissected under sterile conditions and spleens were

removed and resuspended in serum fiee RPMI. Subsequently, a single cell splenocyte

suspension was created by homogenizhg spleens and passing the homogenate through a

0.22 um ce11 strainer. The strainer was then washed tw-ice with 7 rnL of serum free

RPMI, and the flow through was pooled with the splenocyte suspension.

The splenocyte suspension was then centrifùged and washed twice with s e m fiee

RPMI. M e r the last wash, spienocytes were resuspended in 1 mL of Tris-Mt-Cl, pH

7.2 and incubated at 37°C for 2 minutes to lyse red blood cells. Following RBC lysis, the

ceils were washed once again in senun fiee RPMX and resuspended in cdture media

(CM) of RPMI + 10% FCS + 50 u M 2-Me at a concentration of 5 x 106 cells / mL. 1 mL

of this solution was then added to the 1 mL of the stimulator culture that had been

previously set up as indicated above. Each stimulator - effector well was set up in

duplicate. Cells were stimulated in vitro for 5 days, at 37OC. 5% COz.

Prepurution of Target Cells

4 days afier setting up the stimulator - effector CO-culture, 4 x 1o6 STFlO cells were

plated at 5 x lo5 cells / rnL in culture media (CM). 2 x 106 of these cells were infected

overnight with ALVAC IL42 at MOI of 20 and 2 x 106 were grown uninfected, at 37°C

5% CO2. Following ovemight incubation, 10 uCi of 3 ~ - ~ h y m i d i n e was added per

million cells for each group and cells were further incubated for 4 hours at 37OC, 5%

C o l . Subsequent to labeling, cells were washed in serum fiee RPMI and resuspended in

RPMI + 5% FCS at a concentration of 1 x 10' cells / mL.

CTL Setup

Effector cells fiom immunized mice and naive rnice that were being stirnulated with

S V 1 O / ALVAC IL42 were removed fiom the stimulation plate and resuspended via

vigorous pipetting into WMI 1640 -+ 5% FCS at a concentration of I x 10' cells / mL.

150 pL samples containing 1.5 x 106 cells were pipetted into 96 well plates, in duplicate

for lysis reactions against STFl O and STF 1 O / ALVAC IL- 12 target cells. Effector cells

were then serially difuted 3 times by taking out 50 pL and resuspendhg into 100 pL of

CTL medium RPM 1640 + 5% FCS. 100 pL of targets (at 1 x 10' cells / mL) were then

aliquoted into corresponding wells to yield final effector : target ratios of 100: 1, 3 3 : 1,

1 1 3 and3:l.

Plates were centrifùged for 8 minutes at 800 rpm (without brakes) and incubated at 37"C,

5% COz, for 4 hours. Following incubation, labelled unlysed cells were hatvested ont0

Unicell plates using a ce11 harvester. The Unicell plates containing the udysed cells were

washed, dned and baked at 42°C for 40 minutes. Subsequently, the porous membranes

of the plates were sealed at the back, and 25 pL of scintillant was added to each well.

The tops of the plates were then heat sealed and counts were read on a Packard-Top

Count scintillation counter. % Specific Lysis was calculated as outlined by Matzinger et

ul. (88), using the formula:

% Specific Lysis = 100% x (spontaneous cpm - experimental cpm) / spontaneous cpm

Differential Effects of in vitro stimulation on STF10 Lvsis

To determine whether unmodified STF l O cells differentially activated STFl O (parental

tumor) specific splenocytes Erom Unmunized mice, stimulation CO-cultures were set up as

above, against both STFlO/ALVAC as well as unmodified STF 10 cells. Responder cells

were obtained fiom mice that had been immunized with STFlO/ALVAC and then

boosted with the sarne tumor vaccine. Mice were at 21 days post boost, at the time of

splenocyte harvest, Targets of STFIO or STFIO/ALVAC were set up in the manner

described.

Results

Heterogeneitv in Tro~ism Viral Gene Expression

To determine whether the tropism or the gene expression of ALVAC encoded genes was

restricted to various subsets of lymphoid cells, a range of tumor ce11 lines were tested for

their ability to take up the ALVAC virus and express the recombinant hurnan B7-1 gene.

The ce11 lines tested included an acute myeloid leukeniic line C 1498, a melanoma B 16,

mature B-ce11 lymphomas K46J and A20, a mastocytoma P8 15, a pro B-ce11 line NFS-70,

a thymic lymphoma ST, as well as two ST subclones STFIO and STG4. As figure 2

shows, MO, K46J, P8 15 and C 1498 cells were completely negative for recombinant gene

expression, while the B 16 ce11 line displayed a very high fiequency of celts positive for

human B7- 1. In addition, the level of B7-1 molecule expression in this ce11 line was also

very high. NFS-70 and ST were the only murine lymphoid ce11 lines that showed

positivity for B7-1 expression, at fiequencies of 83% and 94%, respectively. STFlO and

STG4, which were subcloned fkom the parental ST population, could also be successfUy

infected with ALVAC B7- 1. Even though the subclones exhibited a high fiequency of

cells positive for B7-1 expression, there was some heterogeneity in the level of B7-1

expression by individual cells in the population. Despite the heterogeneity, the clonal

STFlO was chosen as a twnor mode1 rather than the bulk ST to minimize the level of

cellular heterogeneity that may result from successive mutations that take place in

individual cancer cells within the original tumor bulk.

Surface Marker Phenotye of STFl O

We were interested in evaluating the role of ALVAC based vectors in mediating

immunogenicity in lymphoid malignancies. For this reason, the thymic lymphoma

subclone STFlO was chosen as a tumor model. Because the parental thymic lymphoma

(ST) arose from a SCID mouse, the ce11 Luie was thought to be derived fiom early T-cells.

To c o n f i the developmental stage fiom which this thymic lymphoma arase? the ce11

line was screened for the presence of early T-ce11 markers.

As figure 3 shows, STFlO was CD&, CD25-, HSA+, CD4 low, CD8 low, and CD3-.

In addition, the ce11 line was B7-1 low, B7-2 low, MHC class 1 ( l 3 - 2 ~ ~ ~ ~ ) + and MHC

class II (1-A~) low. Further analysis of the intracellular enzyme constitution of STFlO

revealed positivity for recombination activating gene 1 (RAG-1) and recombination

activating gene 2 (RAG-2), as well as terminal deoxynucleotidyl transferase (TdT)

(figure 4). These results con fmed that STFlO was derived fiom an early double

negative thymic precursor T-lymphocyte.

Growth Kinetics of ALVAC Infected STFl O cells

To determine whether ALVAC infection affected the growth rate of STFlO cells, 1

million STF l O cells were infected with ALVAC B7- 1 and subsequently, population

growth was monitored over tirne, starting fiom a density of 2 X 105 c e l l d d . Figure 4

shows that compared to an uninfected control population, STF 10 cells that have been

infected with ALVAC vectors do not display growth kinetics that are notably diffèrent.

Both survival curves have sirnilar trends indicating that infection via ALVAC vectors has

no effect on the survival and proliferation of STF 1 0 celis.

MHC-class 1 ex~ression on ALVAC infected STFl O cells

To observe the effect of the intracellular ALVAC virus on MHC expression, the tevels of

l 3 - 2 ~ ~ and H - ~ D ~ were assayed at tirnepoints of 6 hours post initial exposure, and 24

hours post infection. As shown in figure 5, there is no downregulation of MHC class I

molecule expression due to the presence of ALVAC. Specifically a slight hcrease is

observed after poxvird infection, in that the georntnç mean £iuorescence fiom antibody

tagged MHC molecules is 17 and 19 for infected cells at 6 hours and 24 hours

respectively, and 14 and 15 for uninfected cells at 6 and 24 hours respectively.

Stabilitv of ALVAC encoded Products over Time

Because the ALVAC virus is non-integrative into the host genome, and is replication

incompetent in mammalian tissue, recombinant products encoded by this v ins will be

transiently expressed. To determine the duration of expression of ALVAC encoded

products, in both a rapidly dividing population, as well as in a non dividing population,

the STFlO tumor mode1 was used to monitor ALVAC B7-1 gene expression over time.

The number of cells positive for human 87-1 was assayed via cell sorting of both a log

phase as weI1 as an irradiated, non dividing culture.

As shown in figure 7, in a continually dividing culture, ALVAC encoded B7-1

expression was maximal at two days post infection (48 hours), as al1 the cells originally

infected with ALVAC B7-1 were still positive for the recombinant molecule. However,

by 72 hours post infection, the expression of human B7- 1 is Iost, and expression levels of

the recombinant molecule return to baseline. in contrast, in a culture where replication of

the cells is hdted via irradiation, the presence of ALVAC encoded products can be

detected nght up to the point at which the cells undergo apoptosis (> 96 hours). In

addition, the number of cells encoding ALVAC B7-1 remains fairly constant, at between

70 - 80% for the duration of this assay, though the level of protein expression (assayed

by fluorescence intensity) gradually decreases over time. These cells undergo apoptosis

at this time presumably due to an excess of DNA damage as a result of prier irradiation.

Viral infection does not induce apoptosis by itself, as the population which was infected,

but not irradiated, expanded continuously, leading eventually to the dominance of virus-

free daughter cells over virally infected cells.

This shows that fustly, the repiïcative index of the chosen ce11 line is an important factor

in how long cells can express ALVAC encoded products, as a population that is not

replicating maintains expression fiom the ALVAC vins longest. Secondly, the temporal

regdation and the half-life of the actual product being encoded are important factors as

even though cells may be positive for ALVAC B7-1, and hence the ALVAC virus, the

level of the recombinant product declines over time.

Decreased Turnorigenicitv of STFlO cells Modified b~ ALVAC Vectors

To determine if genetic modification of STFlO tumor cells by introduction of genes

encoding IL-12 and / or 87-1 could enhance antitumor imrnunity, STFlO turnor cells

containing ALVAC B7-1 or ALVAC IL-12, or both ALVAC B7-1 and ALVAC IL42

were injected into syngeneic mice. As controls, STFlO tumor cells bearing non-

recombinant ALVAC virus, and STFlO tumor cells with no virus were also injected in

parallel into syngeneic mice. Al1 irnmunizations were via subcutaneous injection in the

right flank.

While various previous models have used irradiated c e k as the initial vaccine inoculum

in order to achieve protective immunity against wild type tumor cells, we wanted to

examine whether a primary resonse could be generated against the vaccine before we

exarnined the issue of pretective immunity. As a result, in al1 o u experiments,

vaccination is carried out using [ive tumor cells, as opposed to imdiated cells so that in

the absence of an effective response against the vaccine, a clear indication is obtained via

turnor formation.

Control mice receiving STFlO cells alone al1 succumbed to tumor within 20 days post

injection, while mice receiving STF 1 O/ALVAC B7- 1, STF 1 O/ALVAC IL- 12 or

STF 1 O/ALVAC B7- I/ALVAC IL- 12, al1 remained completely tumor free, upto 55 days

post injection. Surprisingly, mice that received STFlO cells infected with the non

recombinant ALVAC virus (STF 1 O/ALVAC) also rejected their tumors and remained

100% turnor fkee up until day 55 (Figue 8).

Since al1 animals that received STFlO cells containing either parental or recombinant

ALVAC virus rejected their tumors at 100% fiequency, al1 anirnals were challenged with

an injections of wild type STFlO cells in the opposite (left) lateral flank, to determine if

systernic irnmunity against the wild type cells was conferred. As controls, previously

unvaccinated (naive) mice were also injected with wild type STFl O tumor cells.

Following challenge, 4/5 control (previously rinimmunized) mice developed tumors.

Tumor formation was also observed in 5/5 mice immunized with ALVAC B7-1,4/5 mice

immunized with STF I O/ALVAC, 4/5 mice immunized with STF I O/ALVAC IL- 12, and

3/5 mice imrnunized with STF 1 O/ALVAC £37-l/ALVAC IL- 12 (Figure 9). Mice that

were vaccinated with STF lO/ALVAC IL-1 2 initially seemed to display a delay in tumor

formation when challenged with wild type cells. However, log-rank statistical andysis

revealed that compared to survival of naive mice, none of the STF10 vaccines (including

STF 1 O/ALVAC IL- 12) were efficacious in conferring a significantly greater level of

protection afier wild type challenge @ > 0.05 in al1 cases).

Tumorigenicitv in Nude Mice

To determine whether the rejection demonstrated in the above expenment was mediated

by T-cells, a tumorigenicity experiment was camed out in BALBlc nude mice, which

lack mature T-cells.

STFl O cells were infected as above with either a) ALVAC B7-1, b) ALVAC IL-12, c)

ALVAC B7-1 and ALVAC IL-12, d) ALVAC, or e) No virus, and injected

subcutaneously into nude mice. In contrast to results with wiId type mice, STFlO

variants bearing ALVAC 87- 1, or ALVAC IL4 2, or both were not 100% imrnunogenic,

and formed nimors in US, 415 and 2/5 mice, respectively. STF 1 O cells infected with non

recombinant ALVAC parental vector were 100% immunogenic in nude mice, as al1 mice

receiving this cellular variant rejected their tumor vaccine (Figure IOA). However,

relative to tumor formation via unrnodified STFl O tumor cells, tumor formation via

STFlO cells infected with either parental or recombinant ALVAC vectors was still

delayed to a significant extent @ < 0.05). There was no significant difference in the

rejection of STF 1 OIALVAC, relative tu STF 1 O/ALVAC B7- 1, STF I O/ALVAC IL-1 2, or

STF 1 OIALVAC B7-1IALVAC IL- 12.

Similar results were obtained upon repeating the expenment whereby mice received

either a) STF 1 O alone, b) STF 1 O/ALVAC, c) STF 1 OIALVAC B7- 1, d) STF 1 OIALVAC

IL- 12, or e) a mixture of STF I O/ALVAC B7-1 and STF 1 O/ALVAC IL- 12. In agreement

with the initial experiment, tumors resulting from STF IO cells bearing either parental or

recombinant ALVAC vectors were significantly delayed in formation. relative to those

resulting fkom uninfected STFIO cells (p < 0.05 in al1 cases) (Figure IOB). As well,

though there was a significant deIay in tumor formation when STFlO cells were modified

with ALVAC vimses, rejection was not 100% as seen with wild type micet suggesting

the possible role of T-cells in tumor variant rejection.

Taken together, these preliminary fmdings point toward three important conclusions.

Firstly, in contrast to previous assumptions, the ALVAC vector is by itself immunogenic,

in the context of a subcutaneously adrninistered whole ce11 tumor vaccine that has been in

vitro modified. Secondly, as nude mice reject the STFl O variants to some extent relative

to unmodified controls, but not as efficiently as wild type rnice, the effector mechanism

responsible for rejection of STFl O tumor variants is multifactorial. Finally, T-ceiis are

involved to some extent as there is greater mortality in nude mice receiving STFlO

vaccines; however, the vaccination regimen employed is ineffective in conferring

systemic antitumor immunity. Even though there appears to be a slight delay in tumor

formation in mice vaccinated with STF 1 OIALVAC IL- 12, this delay is not statistically

significant, with a sampie size of n=5.

Boosting with whole ce11 vaccines bearing ALVAC vectors confers antitumoral immunity

Because the initial vaccination regimen employing one tumor ce11 vaccine injection

followed by a lethal challenge at day 55 did not yield a statisticdly significant protective

response with any of the tumor vaccines, the protocol was modified in two main respects.

Firstly. the STF 1 OIALVAC B7-IIALVAC IL-1 2 vaccine was replaced with a vaccine

made up of a mixture of STF IOIALVAC B7- 1 cells and STFI OIALVAC IL- 12 cells at

equal arnounts. This was done to keep the infection level via each respective virus

consistent, as an infection procedure that subjects both viruses to the same culture at the

same tirne could not control for the infection of every ceLl by the two viruses at equimolar

ratios. Secondiy, in order to examine whether the vaccination regimen could be

improved to afford better protection upon challenge with wild type tumor cells, a prime

and boost approach was utilized.

Animals were injected at day O with either a) STF10, b) STFlOIALVAC, c)

STFI OIALVAC B7-1, d) STFlOIALVAC IL-12, or e) STFIOIALVAC B7-1 &

STF I OIALVAC IL- 12. Al1 anirnals survived except those that received STF 1 O cells alone

(not shown), and on day 35, al1 animals were boosted with the same respective vaccine.

As controls, naive mice were injected with STFlO cells alone, and half the mice which

previously had received STFIOIALVAC IL-12 were left unboosted. Again, naive mice

receiving STF 10 alone died by day 21 (not shown). Al1 other mice survived, as expected

upon vaccination with STFlO cells modified with parental and recombinant ALVAC

vinises.

21 days after boosting, ail animals were challenged on the opposite flank with wild type

STFlO cells to determine whether systemic immunity had been conferred. However 6

days after challenge, one mouse out of eight, which had been initially vaccinated with

STF IOIALVAC IL- 12 but Ieft unboosted succumbed to tumor formation on the right

flank. Because the challenge was administered on the lefi flank, and the kinetics of

tumor formation argued against this tumor being formed as a result of challenge, this

m o u e was taken out of the sample of survivors that were chdlenged.

As shown in figure 11, al1 naive animals that received STFlO ceils succumbed to tumor

with expected kinetics. Animals that had been vaccinated and boosted with

STF 1 OIALVAC, STF 1 O/ALVAC B7-1, STF 1 O/ALVAC IL- 12 or both STF 1OIALVAC

IL- 1 2 and STF I OIALVAC B7-1 showed significantly delayed tumor formation (relative

to naive mice) in al1 cases (p < 0.01). This showed that after boosting with whole ce11

vaccines bearing parental or recombinant ALVAC vectors, some measure of systemic

irnmunity against STF 1 O cells had been induced. Furthemore, survival fiequencies were

also increased relative to naive mice, as mice vaccinated with STFlO/ALVAC,

STF 1 O/ALVAC B7- 1, STF 1 O/ALVAC IL-1 2, or STFI O/ALVAC IL- 12 and

STFlO/ALVAC B7-1 showed overall survival rates of 50%, 12.5%, 25% and 25%

respectively. Notably, STF I O/ALVAC by itself was a potent immunogen, in that priming

and boosting with STFI O/ALVAC gave a level of protection that was comparable to that

with other vaccines incorporating recombinant ALVAC vectors in STF 10 celis.

Also of interest is the fact that even though a previous vaccination with STFIO/ALVAC

IL-12 (figure 9) failed to elicit a protective response after challenge, this time one

vaccination with STF 1 O/ALVAC IL- 12 was significantly protective against tumor

formation, relative to naive mice @ < 0.01) when the sample size was increased tn n=8.

In addition, a double vaccination with STF 1O/ALVAC IL- 12 does not confer a protective

effect that is superior to that mediated by a single vaccination with STFlO/ALVAC IL-

12. as rnice vaccinated twice displayed kinetics of tunior formation that were comparable

to that seen in mice vaccinated oniy once.

In Vitro Antitumoral Reactivitv

Mice were imrnunized with STFlO/ALYAC IL-12, and their splenocytes were assayed at

various tirnepoints for reactivity against both wild type STF IO and STF I O/ALVAC IL-

12. Spleens were taken at days 13, 17, 27 and 35 post immunization. and lymphocytes

were stimulated against STF 1 O/ALVAC IL-1 2 before they were incubated with STF l O

and STF 1 O/ALVAC IL-1 2 targets for lysis. Spleens fiom naive mice were also processed

similady, as negative controls.

Significant in vitra cytolytic activity could be observed fiom mice only on day 27, and

only against STFIO/ALVAC IL-12 targets, at effector : target ratios of 100:i and 33:l

(figure 12B). Specific lysis of wiinfected STF 1 O tumor cells was not observed at this

timepoint (figure 12D), indicating that STFl O specific CTL-precursors were either absent

or in very low fkquency in the population of splenocytes. Lymphocytes Fom naive

mice. stimulated similariy with STF lO/ALVAC IL-12 were not able to Lyse STFl O or

STF I OIALVAC IL-1 2.

Because in vivo experirnents had previously shown that immuniung and boosting mice

with the STF 1 O/ALVAC variant also increased antitumor responses against STF 1 O tumor

celk upon challenge at day 21, animals were immunized with STFlO/ALVAC, boosted

tvith the sarne respective tumor vaccine, and then assayed for splenic reactivity against

STF 1 O and STFl O/ALVAC at day 2 1. In addition, the differential in vitro effects of

stimulation were also examined in that lysis of the uninfected STFlO ce11 line was

assayed after in vitro stimulation of splenocytes with either wild type STFlO or

STF 1 O/ALVAC.

As shown in figure 13, splenocytes fiom naive mice were unable to lyse uninfected

STF 1 O cells when incubated with STFl O in viiro. However, specific cytotoxic activity

against uninfected STF 10 celis was obsewed fiom splenocytes of mice immunized with

STF 1O/ALVAC, after stimulation ex vivo with STFlO cells (figure 13A). In addition,

specific activity against STF 1 O/ALVAC could dso be elicited from splenocytes

stimulated with wild type STF10, with comparably higher kiiling at the lower effector :

target ratio of 30: 1 (figure 1 3B).

In contrast, stimulation of splenocytes with STF 1 O/ALVAC yielded results that were

simiiar to previous experiments, in that specific cytolytic activity of splenocytes fiom

imrnunized mice was not appreciably greater than that elicited fiom splenocytes of naive

mice, when assayed against STF10 cells (figure 13C). There was also a higher level of

background killing by splenocytes fiom naive mice.

Taken together, these data imply that STFlO specific CTL precursors are indeed present

at day 21, in the splenocyte population of mice that are imrnuaized and boosted with

STFlO/ALVAC. Their clonal expansion, and their subsequent detection is dependent on

the stimulus provided, Le., they are selected more eficiently by uninfected STFIO cells

and cannot be detected if splenocytes are stimulated against STF1 O/ALVAC.

Discussion

The most significant observation descnbed here is that the ALVAC vector itself is

immunogenic and cari elicit a systemic cellular immune response. That is, in the context

of whole ce11 tumor vaccines with STFlO cells, the ALVAC vector was 100%

immunogenic and the additional benefits fiom the encoded B7-1 or IL42 could aot be

detected. Initial experiments indicated that protection against challenge with STFl O

could be elicited at appreciable, though statistically insignificant levels, with STF 1 O cells

modified with ALVAC IL-12. Repeating the in vbo experiment with a greater sample

size indicated that vaccination with STF 1 O/ALVAC IL- 1 2 is indeed significantly

protective; however, boosting with this vaccine does not confer a greater protective

effect.

AIthough protection fkom ALVAC IL- 12 modified STF l O ceils was not enhanced by

boosting the initial immunization, tumor protection was enhanced by boosting using ail

the other STFIO/ALVAC immunogens. A vaccination and boost with STFIOIALVAC is

significantly protective against wild type challenge, at levels that are comparable to that

induced by vaccinating and boosting with STF1 OIALVAC IL- 12. Thus, STF 1 O/ALVAC

is a potent irnmunogen in eliciting an anti-Sm10 response by itself, and in this mode1

there is no added benefit of IL-12 or 87-1 expression.

Experirnents in nude mice indicate that rejection of ALVAC - infected STF IO variants is

mediated in part by T-celis, due to reduction in rejection eficiency, but other effector

mechanisms are also responsible as ALVAC - infected cells are not completely

turnorigenic in nude mice. Lastly, in mice immunized with STFlO/ALVAC, splenic

cytotoxic T-celfs specific for wild type STFlO tumor cells can be detected in vitro but

this detection is dependent on stimulation (ex vivo) with STF10, and is abrogated on

stimulation with STF lO/ALVAC.

Immunogenicitv of the ALVAC Vector

We have shown that both the ALVAC recombinants and the non-engineered ALVAC

vector can induce immunogenic responses in immune competent as well as athymic nude

mice. In wild type mice, tumor cells bearing either parentai or recombinant ALVAC

vectors were always rejected at a fiequency of 100% except in one case where one 1 out

of 16 mice injected with STF lO/ALVAC IL-1 2 developed a tumor. As well, in nude

mice, STFIO cells bearing parental or recombinant ALVAC vectors were rejected at

vary ing, though statistically significant, fkequencies relative to unmodified STF 1 O ce1 1s

which were 100% turnorigenic.

These results indicate that the ALVAC virus by itself is immunogenic and can elicit an

immune response without the added expression of recombinant genes like TL-12 and / or

87-1 when applied in a whole ce11 vaccine. Immunogenicity of the ALVAC virus in

antitumor models was also demonstrated by Puisieux et al. (79) who showed firstly, that

whole ce11 immunization with TS/A adenocarcinorna cells bearing ALVAC P-gal caused

a delay in turnor onset, but did not confer total rejection as did TS/A cells bearing

ALVAC IL12. In our case we saw complete tumor rejection with the ALVAC virus,

comparable to that with ALVAC IL-12. This discrepancy may be resolved by

considering the fact that the multiplicity of infection employed by Puisieux et ai. (79)

was 5 Pfû / celI, whereas the MOI used here was 20 P h / cell. A lower MOI may lead to

a lower efficiency of infection by ALVAC in viiro, providing fewer stimulating signals

and targets in vivo during turnor rejection. Cells bearing ALVAC IL12 wouid in this case

be better rejected because IL-12 is secreted and can modulate effectors in the local

environment to a greater extent than can the parental virus by itself,

Immunogenicity of the ALVAC virus was aiso demonstrated in an additional experiment

by Puisieux et al. (79), whereby intra~unoral injection with the non-specific virus

ALVAC P-gal showed increased survival of mice relative to those injected with PBS.

However, mice injected with ALVAC IL-12 showed increased survival relative to any

other group. Again, this discrepancy may be resolved by the mode of administration:

intraturnoral injection only allows the virus to access those cells that are in close

proximity to the site of injection. As a result, oniy a few cells receive the vector, and any

immune response directed towards the vector alone is directed only to those cells, and not

the majority of the tumor mas. Injection with ALVAC IL-12, however, may have a

better effect because in this case even though a few cells receive the vector, IL-12 c m

affect the entùe local environment once secreted, Effectors al1 around the tumor can be

activated even though only a subset of cells produces IL- 12. in our case, the tumor cells

are modified in vitro, where parental viruses can interact with tumor cells more

homogeneously, allowing most cells to be infected before injection into mice.

Having established that the ALVAC virus is immunogenic, it still remains to be

elucidated which aspect of the virus is providing the immune stimulus. The primary

immune response may be directed against the virions and the protein components

themselves, or may be a function of the viral DNA, which becomes exposed once the

virus uncoats intracetlularly. The related poxvirus, vaccinia, is naturally immunogenic,

and various antigenic determinants have been characterized (89). In this report, it was

shown that antibodies specific for polypeptides fÏom the viral core as weU as the virai

envelope could be detected in sera from Lnmunized humans. Because the poxviruses are

related in structure, it is possible that antigenic detenninants are present in ALVAC both

at the core, which consists of structural proteins and enzymes, as well as the viral coat.

While T-ce11 determinants have yet to be characterized for either vaccinia or ALVAC, it

is possible that initial antitumoral reactivity by T-cells can be a Function of activation

against viral epitopes. These viral epitopes may be presented either indirectly by APC

which pick up determinants fiom apoptotic / necrotic turnor cells, or directly by the hunor

cells that harbour the virus.

Altematively, the nucleotide sequence of the viral genome itself may contain elements

that are immunostimulatory. For example, unmethylated CpG sequences have been

shown to elicit responses fiom various arms of the immune system. f i e g et al. (90)

showed that unmethylated bacterial DNA, as well as synthetically manufactured

oligonucleotides bearing a motif of a CpG island flanked by two 5' purines and two 3'

pyrimidines induce murine B-ceils to activate and secrete immunoglobulin. The response

is potent in that close to all splenic B cells react in this marner. Stacey et al. (91) showed

that macrophages (both prirnary and ce11 line) also react to CpG motifs by secreting

NOS, and Ballas et al. (92) demonstrated enhanced NK killing in vivo and in viîro, in

response to CpG motifs. Finally, dendritic cells have also been implicated in the

response to CpG motifs in that murine fetal skin derived dendritic cells undergo

maturation, upregulate MHC class II and B7-2, and secrete IL42 at arnounts far greater

than when activated by LPS (93). This activation of dendntic cells as a result of CpG

treatment has been shown to be abie to provide a positive stimulus for allogeneic T-cells

to proliferate in vitro (93).

Thus, if the genome of the ALVAC virus contaias CpG motifs, then it may be a mode

through which it is exerting its immunogenicity. Antigen presenting cells consisting of

macrophages and dendritic cells may take up cellular debris fiom apoptotic tumor cells,

and instead of presenting viral epitopes, they are presenting of tumor antigen epitopes

with the concurrent activation of costirnulatory molecules. The upregulation of

costimuIatory molecules in conjunction with turnor antigen presentation may serve to

activate previously anergized T-cells against tumor epitopes. Once activated, CDS+ T-

cells can reject the twnor efficiently.

Antiturnoral Protection followine Vaccine Boost

We showed that systemic antitumor protection against wild type STF l O tumor cells could

be enhanced by boosting the initial STFlO cellular vaccine. This was true for

immunogens containing both parental or recombinant ALVAC vectors, except those

modified with ALVAC IL-12, which proved to be protective afier a single vaccination.

Thus, in a double vaccination protocol, STFlO/ALVAC is as protective as al1 other tumor

vaccines, and the protective effect can be attributed to the virus alone.

These results indicate that potent T-ce11 memory against STFIO can be induced by

vaccination with STFlO/ALVAC by itself; however, this is dependent on a boost with the

relevant tumor vaccine. This agrees with the view that T-cells need to be exposed to

tumor or virai antigens more frequently in order to maintain immunity against tumor

cells, Le., T-ce11 memory is dependent on the increased presence of turnor antigens over

tirne. I f the host encounters an antigen at day 0, clears it and does not see it again until

day 55, the subsequent response is not as potent as that when the antigen is encountered

again partway through this period at day 35. The issue of whether T-ceil memory is

mediated by long Lived lymphocytes that have encountered antigen once, versus those

that continually need to mount a low level immune response against persistent antigen to

maintain memory, was addressed by Kundig et al. (94). lmmunization with a completely

immunogenic Vesicular Stomatitis Virus carrying the nucleoprotein N (VSV-N), and

subsequent challenge with pathogenic vaccinia vîmes containing the same immunogenic

CDS+ restricted protein N (Vacc-N), caused mice to lose T-ce11 rnediated reactivity

against N protein (and hence Vaccinia) over time. Further, the level of imrnunity is a

function of immunization dose, in that at lower doses of the original immunogen, the

immunity is lost faster.

Further, the role of antigen persistence was demonstrated (94), by vaccinating with a low

dose of VSV-N and challenging with Vacc-N with or without boosting wîth EL-4 cells

carrying N protein. Without boosting, mice were cornpletely susceptible to challenge at

day 44; however a boost at day 40 abrogated this susceptibility, and that multiple boosts

before this time point were not necessary.

Thus, T-ce11 memory responses against peripheral antigens rnay indeed be a fiinction of

the time following initial antigen clearance as weli as fiequency of antigen encounter. If

the antigen is completely cleared from the periphery, then T-ce11 respnses against the

antigen can only be observed if splenic T-cells are reactivated. According to Kundig et

al, in order to maintain peripheral immunity, the immunogen must persist in low arnounts

such that at the time of challenge, T-cells are already present in the periphery and

activated. This was demonstrated by showing that LN lymphocyte reactivity against

antigen is lost over time, while splenic activity is not, and that LN reactivity can be

regained by a boost.

An analogous situation may be applicable in antitumor responses, especially in the mode1

that is presented here. Challenge with wild type STF 10 too long after initial vaccination

may result in tumor formation due to the fact that T-cells have already cleared the initial

vaccine and have left the periphery. If however, a boost is administered, an additional

stimulus is provided for traficking T-cells to remain activated in the periphery. At the

time of challenge, these activated T-cells may be sufficient to mediate antitumor activity.

Protection against Tumor Epitopes as a Result of Viral Irnmunogenicity

Since wild type STFlO cells can in fact be cleared after vaccinating and boosting with

STF I O cells bearing parental and recombinant ALVAC vectors, it can be concluded that a

memory response can be induced via the activation of STF I O rumor speci fic T-cells upon

systemic challenge. While experiments in nude mice confirm that there are additional T-

ce11 independent mechanisms responsible for STF l O variant clearance, the presence of a

protective immunity against unmodzjied tumor ceiis suggests T-ceii involvement, How a

T-ce11 response agaînst STFlO cells c m be gained as a result of a virally infected cellular

vaccine is still unclear, but a number of possibilities exist. If, as discussed above, the

viral DNA is irnrnunostirnulatory, then during rejection of STF 1 O cells bearing ALVAC

vectors, tumor antigen specific T-cells that recognize cellular epitopes are being

activated. This would require tumor antigen uptake, processing and presentation by APC

that have upregulated their costimuIatory molecule and cytokine expression levels, in

response to ALVAC sequences. ln this way, previousiy anergized or ignorant T-cells are

activated against turnor epitopes, and can eff~ciently reject even wild type tumor cells

upon challenge.

Alternatively however, the ALVAC protein epitopes may themselves be the primary

imrnunogenic stimulus that is king presented to T-cells. Regardless of this, what we are

observing is that T-ce11 activation against non-viral, tumor epitopes can also be induced.

This transference of specificity may be explained either by the action of cross reactive T-

cells that are activated by viral epitopes and can now recognize twnor epitopes, or by the

phenomenon of epitope / detenninant spreading.

The phenornenon of epitope spreading has usually been associated with autoimmune

disorders in animal modets of disorders such as uisulin dependent diabetes (95) as well as

relapsing rernitting experimentai autoimmune encephalomyelitis (R-EAE) (96), which is

a Th1 mediated demyelinating disease responsibte for chronic CNS damage. R-EAE can

be induced in mice by the administration of an immunodominant peptide epitope fiom

the proteolipid (PLP) molecde (96), or an epitope from myelin basic protein W P ) (97)-

The initial reports of epitope spreading came fiom experiments whereby immunkation of

mice with the MBP protein allowed the isolation of proliferative T-ce11 clones specific for

the irnmunodominant epitope Acl- 1 1, as weli as non cross reactive T-ce11 clones specific

for additional determinants in peptides 35-47, 8 1- 100 and 12 1 - 140 (97). Importantly

however, reactivity to these epitopes could also be observed upon immunizhg mice with

only MBP Ac 1 - 1 1 and not the whole protein. Determinant spreading across different

molecules was demonstrated by Cross et ai. (98) who injected mice with epitope p87-99

of MBP, and found non cross reactive LN lymphocyte reactivity against epitopes of the

PLP protein during acute or chronic stages of disease.

Both interrnolecular as well as intramolecular epitope spreading were demonstrated by

McRae et ai. (96) where firstly, adoptive transfer of T-cells specific for MBP epitope 84-

104 resulted in generation of T-cells in vivo that were specific for the PLP determinant

139- 15 1. Secondly, adoptive transfer of T-cells of mice primed with PLP 139-1 5 1,

resulted in reactivity against a non cross reactive intramolecular epitope PLP 178-191,

dunng acute phase of EAE.

While the above studies are examples of epitope spreading as a result of initial prirning

via "self' epitopes, pnming via viral epitopes has also been shown to give nse to

determinant spreading. Like R-EAE, multiple sclerosis (MS) is also a T-ce11 mediated

autoimmune demyelinating disease, and it has been found that infection of rnice with

Theiler's murine encephaiomyelitis virus (TMEV) induces a chronic CD4+ T-ce11

rnediated disease that models progressive MS (99). Miller et al. (99) have shown that

injection of TMEV intracerebrally results in disease at about 40-50 days. Starting about

7 days post injection, and for the duration of the disease, infected mice continue to

display T-ce11 proliferative behaviour as well as DTH responses against wal epitopes

only. In contrast, there are no DTH or proliferative responses against a panel of myelin

epitopes. However, at 30 days post onset (Le., 80 days post viral infection), T-ce11

responses to myelin epitopes appear sequentially. Fust, responses are seen against PLP

139-15 1, followed by responses against epitopes 56-70 and 178-191 by day 164.

According to Miller et al. (99), the kinetics of this differentiai specificity argue against a

cross reactivity of T-cells against both cellular and viral epitopes as a result of molecula.

mimicry. if indeed there was cross reactivity. then reactivity against myelin epitopes

should also have been detected at tirnepoints when reactivity against virus was detected.

The mode1 put forth by Vanderlugt et al, (100) explains epitope spreading as a cascade

resulting from recognition of a single epitope. Antigen presenting cells first present

either the administered peptide epitope or viral epitope to CD4+ heiper cells. These T-

ceIIs activate the APC's to upregulate costirnulatory signals and at the same time initiate

an inflarnrnatory response. This results in destruction of cells carrying these epitopes.

However, following tissue destruction, cellular debns is taken up by the same group of

activated APC's. Subsequent presentation of cellular epitopes by the activated APC in

turn can Iead to priming of T-cells against self-detenninants.

T-ceil mediated immunity against unmodified STF 1 O tumor cells may also arise from

determinant spreading afier initial pruning with ALVAC epitopes- M C ' S picking up

turnor debns can present ALVAC epitopes fiom the initial cellular vaccines to CD4+ Th1

helpers, leading to APC activation. The APC in tum could activate CD8+ T-ceiis against

cells bearing ALVAC epitopes, while Th1 helpers could initiate an inflammatory

response- The resdting destruction of ALVAC bearing tumor cells could lead to the

uptake of cellular debris by activated denciritic cells and B-celis. These APC's which still

express upregulated costirnulatory molecules can now prime T-cells against cellular

epitopes. This in turn can induce a protective antitumor response when animais are

chailenged wiîh wild type tumor cells.

T-ce11 Independent Antitumor Mechanisms

in the experiments presented here, mice were shown to be significantly protected fiom

challenge after two respective vaccines of STFlO celfs bearing parental or recombinant

ALVAC vectors. While this level of antitumor immunity implicates imrnunologicai

memory, and hence, T-ce11 activity, primary tumor vaccines consisting of STFlO cells

bearing parental and recombinant ALVAC vectors were also significantly rejected in

homozygous nude mice which lack mature T-cells. Again, the added expression of IL-12

or B7-1 or both did not confer any additionaiiy protective effect over that mediated by the

vector alone. That this rejection of STFlO variants is not as potent as in wild type mice is

evidence that T-cells are indeed involved in primary rejection; however, the significant

level of protection over rnice receiving unmodified cells suggests the added role of other

effectors, such as macrophages, NK cells or neutrophils.

One of the fust studies of T-ce11 independent reactivity against gene modified tumor

vaccines was carried out by Golumbek et ai. (31) who showed that while growth of

Renca cells in nude or SCID rnice resulted in 100% mortality within 20 days, Renca cells

expressing IL-4 did not grow at al1 in nude or SCID mice until after two months post

injections. Histologic examination of the injection site in wild type mice showed a

presence of macrophages, as detennined by both gross characteristics, as well as Mac-1

positivity. This showed that in some cases, the initial phases of tumor rejection may be

carried out by non specific effectors, though T-cells may be required for ultimate

clearance.

A more specific mode1 of TA3 - independent NK - mediated antitumoral reactivity was

demonstrated by Alosco et al. (1 0 1) where murine fibrosarcoma CMS-5 cells either did

not grow, or regressed in SCID mice. This turnor rejection was ablated when mice were

treated with anti-asialo GM-1, which reduces NK activity. In parallel, in vitro

cytotoxicity of splenocytes was reduced against al1 of CMS-5, CMS-5 + IL-2 and YAC-1

cells when mice were treated with anti-asialo GM-1.

T-ce11 independent antinimor reactivity as a result of non specific responses against a

viral vector was observed by Siders et al. (102) using a renal carcinoma metastasis

model. In this report, Siders et al. established a metastatic hepatic malignancy by

intrasplenic injection of Adenovhs expressing IL- 12, or P-gal. In wild type mice, the

number of metastases were cured by 94% with injection of Ad EL42 (relative to no

treatment), and by 25-40% with injection of Ad P-gal. Staining of hepatic sections

revealed the existence of CD3+ T-celis, as well as lysozyme containing cells (neutrophils

or macrophages) around hepatic blood vessels of animals treated with Ad vectors.

To clarify the mode of antitunoral reactivity, the experiment was repeated in SCID mice,

where it was observed that treatment with Ad P-gal reduced the number of metastases by

46% while treatment with Ad IL-12 reduced the number of metastases by 95%. This

showed that firstly, an antitumoral response could by elicited by administration of the

virus alone, although not as efficiently as with IL-12 expression, and that the response

may not be T or B ce11 mediated. NK-ce11 depletion in SCID mice abrogated the

antitumoral activity in response to Ad P-gal, but Ad-IL-12 still atlowed for efficient

rejection of turnors. This showed that in this scenario, IL-12 exerts a greater effect in

mediating tumor rejection that is greater than that af5orded by the vector alone. However,

even the effect that is mediated by the vector is significant, and may not be entirely

dependent on T-cells.

A sirnilar model may be applicable in rejection of tumor cells bearing ALVAC vectors.

While T-cells may be responsible for ultimate clearance of tumors, such as in the case

with Golumbek et ai. (3 l), other effectors such as macrophages, neutrophils or NK celis

may be responsible for the partial clearance that is observed in nude mice. Because there

is no enhancement of the antitumor response with ALVAC-IL- 12 and / or ALVAC B7-1

over ALVAC alone, the immunogenicity of the tumor cells in nude mice can be

attributed to the virus alone. In nude mice, non adaptive effectors such as macrophages

and NK cells may be recognizing the tumor cells as targets that have been rnodified by

the virus in such a way as to be more likely candidates for killing- The virai effects on

the tumor cells are unclear at this point, but in addition to possible T-ce11 stimulatory

effects as dîscussed above, they are dehitely immunostimulatory towards other effectors

as well.

In Vitro CTL Activi-

In the in viîro CTL experiments presented here, splenocytes fiom mice immunized with

STF1 O/ALVAC IL-12 react against STFlO tumor cells infected with ALVAC IL- 12 but

not STF 1 0 alone, when stimulated in vitro with STF 1 O/ALVAC IL- 12. Subsequently, a

side-by-side comparison of different in viîro stimuli of either STF 1 OIALVAC or STF 10,

showed that stimulation by STF 10 alone allowed generation of T-cells specific for both

STF 1 O/ALVAC as well as unmodified STF 10, while stimulation with STF 1 OIALVAC

reduced the amount of observed reactivity against unmodified STFIO. Thus, it can be

concluded that STF10-specific CTL's can be amplified in vitro from spleens of

immunized mice, however, their amplification is dependent on the mode of stimulation.

Stimulation via STF 1 O/ALVAC (or STF 1 O/ALVAC IL- 12) allows the generation of T-

cells specific only for STF 10 infected with ALVAC viruses, while stimulation via STF l O

allows amplification of C m ' s specific for either STF 1 O or STF 1 O/ALVAC. Incidentally,

cells that were stimulated with Sm 1 O o d y displayed higher killing of STF 1 O/ALVAC

targets, relative to STFlO targets, at lower effector to target ratios. This may be due to

the fact that these splenocytes consist of T-cells that are specific for STFlO epitopes, as

well as T-cells specific for viral epitopes, and wbat we are observing is a synergistic

effect of these two groupe of cells, during lysis of STFlO cells beaing ALVAC.

Alternaiively, however, there is an increase in MHC class 1 expression (around 33%) in

STF 1 O cells after infection via recombinant ALVAC vectors. The higher level of MHC I

could provide a greater number of targets for T-cells that are specific for tumor epitopes,

leading to greater overall killing.

The fact that wild type tumors can be killed by splenocytes after stimulation via STFlO

shows that splenic T-cells specific for STFlO are indeed present around day 25; however,

their in vitro amplification is somehow being blocked during stimulation with STFl O

cells that have been infected with ALVAC. This inhibition of STFlO-specific T-ce11

amplification may be explained either by an inefficient presentation of tumor epitopes

during stimulation via STFlO/ALVAC (and hence an inefficient expansion of STF IO-

specific T-cells), or an inherently low level of tumor ce11 specific T-cells in the original

splenic population.

Suboptimal levels of T-ce11 activation due to differential antigen presentation was

demonstrated by Gallimore et al. (103) using LCMV viral antigens as immunogens. In

this report, it was shown that iaimunization of vaccinia immune mice with vaccinia

encoding a subdominant LCMV glycoprotein epitope (GP117) protected mice when

challenged with LCMV bearing GP 1 1 7. However, the same mice were not protected

when they were challenged with LCMV bearing GPI 17 as well as two other

immunodominant epitopes GP33 and GP276. Stability assays showed the GP117 - H-2b

cornplex to have a half life of less than 2 hours, while the GP33 - H-2b and GP-276 - H-2b

complexes had half-lives greater than 4 hours. Furthemore, peptide elution experiments

showed that while MC57 cells infected with the LCMV variant bearing GP117 but not

GP33 and GP276 gave an active fraction for GP117 as expected, cells that had been

infected with wild type LCMV did not display GPI 17. This suggested that in the

presence of other immunodominant epitopes in the cell, subdominant epitopes may not be

presented as efficiently. Gallimore et ai. (103) propose a mode1 whereby the presentation

of a subdominant epitope may be "competed out'' either due to higher dissociation rate

fiom MHC molecules or due to differentid processing via the ER.

Sirniiarly, in STFlO cells that have been infected with ALVAC viruses, it may be

possible that viral epitopes are being presented to T-cells more efficiently d u ~ g ex vivo

stimulation of spienocytes h m immunized mice. This difference in efficiency may be

due to a number of reasons, including a higher affinity of the viral epitopes for MHC

class 1, more efficient processing of viral epitopes via the ER, or the presence of a greater

number of viral epitopes that can effectively compete out the cellular epitopes for binding

to MHC. Regardless, cellular epitopes are excluded to the end-effect that in vitro, STFlO

epitope specific T-cells are not being expanded when stimulated STFlO/ALVAC or

STF 1 O/ALVAC IL- 12.

Alternatively, în animals that have k e n vaccinated with STF IO infected with parental or

recombinant ALVAC vectors, cytotoxic T-celi clones specific for TRA epitopes may be

fewer in number than those specific for ALVAC epitopes, and hence are not being

effrciently expanded when stimulated with STFlO/ALVAC. A significant difference

between the total number of peripherai T-cells that are specific for one antigen versus

another antigen may arise due to CDS+ T-ce11 cornpetition, a phenornenon described by

Freitas et al. (104).

Freitas et al. have demonstrated that reconstitution of irradiated hosts with a mixture of

transgenic (Tg) and non-transgenic (non-Tg) lymphocytes allows establishment of

peripheral CD8+ T-celts at frequencies that are dependent on the specificity of the

transgenic K R . Generally, in chimeras reconstituted with mixtures of Tg and non Tg

cells at Iow ratios of Tg/ non Tg, mature transgenic CDS+ cells in the periphery were rare

or absent. In chimeras with higher initial frequencies of Tg bone marrow cells, the

percentage of CDS+ single positive Tg cells was significant in the peripheral pool, but the

frequency was dependent on the TCR specificity. T-cells that were specific for HY

antigen constituted less than 10% of the peripheral (splenic and LN) population (relative

to non Tg cells), and T-cells specific for LCMV P 14 antigen constituted greater than 50%

of cells in periphery, when reconstituted separately with non transgenic cells. A

subsequent experîment, which looked at the effect of mixing the two transgenic

lymphocyte precursors, showed that at low anti HY: anti Pl4 precursor ratios, anti HY T-

cells in the periphery were rare. In contrast at high ratios, anti HY cells constituted the

majority of peripheral CD8+ T-cells after the f h t weeks of reconstitution, but at 3 weeks,

both populations reached identicai sizes. Finally, at 16 weeks, P l4 transgenic cells

became the dominant population. In vivo BrdU staining reveaied that while proliferation

of peripheral CD8+ transgenic was poor relative to wild type cells overall, in chimeras

injected with both anti HY and anti P14, there was an increased accumulation of stained

anti P l 4 ceIIs relative to anti HY cells. Further, the majority of the anti P l4 cells are

activated (CD44t) while only I O - 1 5% of anti KY ceUs are activated. These results Led

to the postulation that cornpetition does exist in selection of T-cells and that cellular

dominance may occur by preferential ce11 activation. In other words, a state of activation

may confer a "selective advantageY7 in establishing a dominant population (1 04).

This mode1 may be applicable in explaining why in vitro stimulation of splenocytes from

immunized mice with STF 1 O/ALVAC preferentially amplifies cells specific for

STFIO/ALVAC and not STFlO. In the in vivo situation T-cells specific for tumor

epitopes may initially constitute a smaller proportion of the entire population relative to

other T-ce1ls which are specific for non self peptides such as ALVAC epitopes. During

immunization with STFlO/ALVAC, the anti-ALVAC T-cells may be activated more

efficiently than anti STFIO ceIIs due to reasons such as immunodominance. This in tuni

could result in the dominance of ALVAC specific T-cells over STF10 specific T-cells in

the spleen. When these splenocytes are harvested and stimulated with STFlO/ALVAC,

anti-ALVAC T-cells preferentially expand due to the selective advantage gained by their

superior numbers and activation stahis. This leads to masking of STF10 specific T-cells

as these cells cannot establish a dominant population. In contrast, when stimulating with

STFlO oniy, anti-ALVAC T-celb are not re activated ex vivo and this time, anti STFlO

T-cells may regain a selective advantage due to their preferential activation.

Future Directions

The usefùlness of the ALVAC vector in immunotherapy of cancer has been demonstrated

by other investigators through use of murine models. In previous reports, therapeutic

effects have been gained through the expression of recombinant genes via the ALVAC

vector, while the vector itself has been shown ta be moderately immunogeaic by itself.

This report established the inherent imrnunogenicity of the ALVAC virus vector M e r ,

in a mode1 whereby the virus itself is 100Y0 immunogenic, as well as protective towards

wiId type tumor challenge.

A natural progression of this work would be to extend this particular turnor mode1 in

order to make it more representative of actual human malignancies, where disease may

often by well established before gene therapy is applied. Thus, the efficiency of the

ALVAC vector has to be evaluated in the treatment of established turnors whereby wild

type STFlO cells are inoculated within the mode1 animal, and treatment begins at a

subsequent tirnepoint. The nature of the treatment could involve the intratumoral

injection of ALVAC virues bearing IL42 or B7-1, to directly modiQ the existing tumor

cells such that the local immune effectors are subsequently activated. Altemativeiy, the

treatrnent could consist of an injection of radiation-inactivated tumor cells expressing IL-

12 or B7-1, and in this scenario, effectors are activated initially by the gene rnodified

vaccines. If systemic immunity is conferred, a subsequent response may then be directed

against the established wiid type turnor.

To ultimately make this vector suitable for cancer therapy in humans, a balance has to be

established in minimizing the immunogenicity of the ALVAC virus, while maximizing

the therapeutic effects of the recombinant genes. Factors affecting the level of

immunogenicity may include attributes unique to the model itself, such as tumor type and

MHC background. but also general treatment approaches such as route of administration,

dose of vim, as well as model of tumor establishment and treatrnent, So fa. the cancer

treatment modeis utilizing ALVAC have not been established with the differential effects

of these characteristics in mind. Thus, while the ALVAC virus is still an attractive vector

for cancer gene therapy, the relative effects of each of these parameters should be

established with regard to ALVAC immunogenicity. While clinicd trials utilizing

ALVAC are already underway, parameters influencing viral immunogenicity need to be

better defined before the ALVAC virus vector is applied more extensively for therapeutic

purposes.

References

1. Burnet, F. M. 1 970. The concept of immunological surveillance. Prog EXp Turnor

Res 13:L

2. Boon, T., J. C. Cerottini, B. Van den Eynde, P. van der Bruggen. and A. Van Pel.

1994. Tumor antigens recognized by T lymphocytes. Annu Rev Imrnunol12:337.

3 . Bodmer, W. F., M. J. Browning, P. Krausa, A. Rowan, D. C. Bicknell, and J. G.

Bodmer. 1993. Twnor escape fiom immune response by variation in HLA expression and

other mechanisms. Ann N Y Acad Sci 690:12.

4. Maryanski, J. L., M. Marchand, C. Uyttenhove, and T. Boon. 1983- Immunogenic

variants obtained by mutagenesis of mouse mastocytoma P8 15. VI. Occasional escape

from host rejection due to antigen-loss secondary variants. Int J Cancer 3 1 : 11 9.

5. Torre-Arnione, G., R. D. Beauchamp, H. Koeppen, B. H. Park, H. Schreiber, H.

L. Moses, and D. A. Rowley. 1990. A highly immunogenic tumor transfected with a

murine transforming growth factor type beta 1 cDNA escapes immune surveillance. Proc

Nat2 Acad Sci U S A 87:1186.

6. Chen, L., S. Ashe, W. A. Brady, 1. Hellstrom, K. E. Hellstrom, J. A. Ledbetter, P.

McGowan, and P. S. Linsley. 1992. Costimulation of antitumor imrnunity by the B7

counterreceptor for the T lymphocyte molecules CD28 and CTLA-4. Cell 71.-1093.

7 . Janeway, C. A. 1997. hmunobiology: The hrnune System in Health and

Disease. Current Biology/Garland f ublishing Inc.

8. Dranoff, G., E. Jaffee, A. Lazenby, P. Golumbek, H. Levitsky, K. Brose, V.

Jackson, H. Hamada, D. Pardoll, and R. C. Mulligan. 1993. Vaccination with irradiated

tumor cells engineered to secrete murine granulocyte-macrophage colony-stimulating

factor stimulates potent, specific, and long-lasihg anti-tumor immunity. Proc Nati Acad

Sci U S A 90r3539.

9. Huang, A. Y., P. Golumbek, M. Ahmadzadeh, E. Jaffee, D. Pardoll, and H.

Levitsky. 1994. Role of bone marrow-derived cells in presenting MHC class 1-restricted

tumor antigens. Science 264: 961.

10. Mueller, D. L., M. K. Jenkins, and R. H. Schwartz 1989. Clonal expansion versus

fùnctional cional inactivation: a costimuiatory signalling pathway determines the

outcome of T ceil antigen receptor occupancy. Annu Rev lmmunol7:445.

1 1. Baskar, S. 1996. Gene-modified tumor cells as cellular vaccine, Cancer Imrnunol

lrnrnunother 43: 165.

12. Townsend, S. E., and J. P. Allison. 1993. Tumor rejection after direct

costirnulation of CDS+ T cells by 87- transfected melanoma cells [see comments].

Science 259:368.

13. Chen,L.,P.McGowan,S.Ashe,J.Johnston,Y.Li,I.Hellstrom,andK.E.

Hellstrom. 1994. Turnor immunogenicity determines the effect of B7 costimulation on T

cell-mediated turnor ünmunity. J E X p Med 1 79:%?3.

14. Freeman, G. J., J. G. Gribben, V. A. Boussiotis, J. W. Ng, V. A. Restivo, Jr., L. A.

Lombard, G. S. Gray, and L. M. Nadler. 1993. Cloning of B7-2: a CTLA-4 counter-

receptor that costirnulates human T ce11 proliferation [see comments]. Science 262,909.

15. La Motte, R. N., M. A. Rubin, E. Barr, J. M. Leiden, J. A. Bluestone, and M. B.

Mokyr. 1996. Therapeutic effectiveness of the imrnunity elicited by P8 15 tumor cells

engineered to express the B7-2 costimuiatory molecule. Cancer hmrinol lmrnunother

42.- 161.

16. Yang, G., K. E. Hellstrom, 1. Hellstrom, and L. Chen. 1995. Antitumor immunity

elicited by tumor cells transfected with B7-2, a second ligand for CD28KTLA-4

costirnulatory molecules. J lmrnunolI51:2794.

17. Sartor, W. M., N. Kyprianou, D. F. Fabian, and A. T. Lefor. 1995. Enhanced

expression of IC AM- 1 in a murine fibrosarcorna reduces tumor gro wth rate. J Surg Res

59: 66.

18. Brno. D. K., N. Kyprianou, W. M. Sartor, D. F. Fabian, J. Turner, TT. Vu, A-

Patel, C. Trirnbach, and A. T. Lefor. 1995. Transfection of a m u ~ e fibrosarcoma with

intercel lular adhesion molecule- l enhances the response to adoptive immunotherapy.

Surgery I l 8-23 7.

19. Uzendoski, K., J. A. Kantor, S. 1. Abrams, J. Schlom, and J. W. Hodge. 1997.

Construction and characterization of a recombinant vaccinia virus expressing murine

intercellular adhesion molecule- 1 : induction and potentiation of antitumor responses.

Hztm Gene Ther 8 :8 j 1.

20. Pollok, K. E., Y. J. Kim, J. Hurtado, 2. Zhou, K. K. Kim, and B. S. Kwon. 1994-

4-IBB T-ce11 antigen binds to mature B cells and macrophages, and costirnulates anti-

mu-primed splenic B cells. Eur J lmmunol24.-367.

21. DeBenedette, M. A., N. R. Chu, K. E. Pollok, J. Hurtado, W. F. Wade, B. S.

Kwon, and T. H. Watts. 1995. Role of 4-1BB ligand in costimulation of T lymphocyte

growth and its upregulation on Ml 2 B lymphomas by CAMP. J f ip Med 181:98j.

22. DeBenedette, M . A., A. Shahinian, T. W. Mak, and T. H. Watts. 1997.

Costimulation of CDî8- T lymphocytes by 4- 1 BB ligand. J Zmmunol I j 8 : j j l .

23. Guinn, B. A., M. A. DeBenedette, T. H. Watts, and N. L. Berinstein. 1999. 4-

IBBL cooperates with B7-1 and 87-2 in converting a B ce11 lymphorna cell Luie into a

long-lasting antitumor vaccine. J Imrnunol162:5003.

24. Melero, I., N- Bach, K. E. Hellstrom, A. A d o , R. S. Mitder, and L. Chen. 1998.

Amplification of tumor immunity by gene transfer of the CO-stimulatory 4-1BB ligand:

synergy with the CD28 CO-stimulatory pathway. Eur J Immun02 28: 11 16.

25- Melero, I., W- W- Shuford, S. A. Newby, A. A d o , J- A. Ledbetîer, K. E,

Hellstrom, R S. Mittler, and L. Chen. 1997. Monoclonal antibodies against the 4- 1 BB T-

ce11 activation molecule eradicate established tumors. Nar Med 3:682-

26. Colombo, M. P., and G. Fomi. 1994. Cytokine gene transfer in tumor inhibition

and tumor therapy: where are we now? Immunol Toduy 15:48.

27. Tepper, R. I., and J. J. Mule. 1994. Experiniental and clinical studies of cytokine

gene-modified hunor cells. Hum Gene Ther 5: 153.

38. Douvdevani, A., M. Huleihel, M. Zoller, S. Segai, and R. N. Apte. 1992. Reduced

tumorigenicity of fibrosarcomas which constitutively generate IL-1 alpha either

spontaneously or following IL- 1 alpha gene transfer. Inr J Cancer 5 1 :8X.

29. Fearon, E. R., D. M. Pardoll, T. Itaya, P. Golurnbek, H. 1. Levitsky, J. W. Simons,

H. Karasuyama, B. Vogelstein, and P. Frost. 1990. Interleukin-2 production by tumor

celIs bypasses T helper function in the generation of an antitumor response. Cell60:397.

30. Gansbacher, B., K. Zier, B. Daniels, K. Cronin, R. Bannerji, and E. Gilboa. 1990.

lnterleukin 2 gene transfer into tumor cells abrogates tumorigenicity and induces

protective irnmunity. J Exp Med 1 72: 12 1 7.

3 1. Golurnbek, P. T.3 A. J. Lazenby, H. 1. Levitsky, L. M. JaEee, K. Karasuyama, M.

Baker, and D. M. Pardoll. 1991. Treatment of established reaal cancer by tumor cells

engineered to secrete interleukin-4. Science 254: 713.

3 2 Tepper, R, I., P. K. Pattengale, and P. Leder. 1989. Murine interleukin-4 displays

potent anti-tumor activity in vivo. Ce22 5 7:jO3.

33. Gansbacher, B., R Banne rji, B. Daniels, K. Zier, K. Cronin, and E. Gilboa. 1990.

Retroviral vector-mediated gamma-interferon gene transfer into tumot cells generates

potent and long lasting antitumor immunity. Cancer Res 50: 7820.

34. Porgador, A., R. Banneji, Y. Watanabe, M. Feldman, E. Gilboa, and L.

Eisenbach. 1993. Antirnetastatic vaccination of tumor-bearing rnice with two types of

IFN- gamma gene-inserted tumor cells. J ïmmzrnal150: 11/58.

35. Esumi, N., B. Hunt, T. Itaya, and P. Frost. 199 1. Reduced tumorigenicity of

murine tumor cells secreting gamma-interferon is due to nonspecific host responses and

is unrelated to ctass I major histocompatibility complex expression. Cancer Res ji:lZ85.

36. Porgador, A., E. Tzehoval, A. Katz, E. Vadai, M. Revel, M. Feldman, and L.

Eisenbach. 1992. Interleukin 6 gene transfection into Lewis lung carcinoma turnor cells

suppresses the malignant phenotype and confers irnmunotherapeutic cornpetence against

parental metastatic cells. Cancer Res 52:3679.

37. Sun, W. H., R- A. Kreisle, A. W. Phillips, and W, B. Ershler. 1992. In vivo and in

vitro characteristics of interleukin 6-transfected B 16 melanoma cells. Cancer Res

52:5312.

38. Hock, H., M. Dorsch, T. Diamantstein, and T. Blankenstein. 199 1. Interleukin 7

induces CD4+ T cell-dependent tumor rejection. J E X p Med 1 74: 129 1.

39. McBride, W. H., 5. D. Thacker, S. Comora, J. S. Economou, D. Kelley, D. Hogge,

S. M. Dubinett, and G. J. Dougherty. 1992. Genetic modification of a murine

fibrosarcoma to produce interleukin 7 stimulates host ce11 infiltration and turnor

irnmunity. Cancer Res S2:393 1.

40. Zitvogel, L., H. Tahara, P. D. Robbins, W. J. Storkus, M. R. Clarke, M. A.

Nalesnik, and M. T. Lotze. 1995. Cancer immunotherapy of established turnors with IL-

12. EEective delivery by geneticaliy engineered fibroblasts. Jlmmunoll55:1393.

41. Tahara, H., L. Zitvogel, W. J. Storkus, H. J. Zeh, 3rd, T. G. McKimey, R. D.

Schreiber, U. Gubler, P. D. Robbins, and M. T. Lotze. 1995. Effective eradication of

established murine tumors with IL-12 gene therapy using a polycistronic retroviral

vector- J lmrnunol154:6466.

42. Asher, A. L., J. J. Mule, A. Kasid, N. P. Restifo, J. C. Salo, C. M. Reichert, G.

Jaffe, B. Fendly, M. Kriegler, and S. A. Rosenberg. 1991. Murine tumor cells transduced

with the gene for tumor necrosis factor- alpha. Evidence for paracrine immune effects of

tumor necrosis factor against tumors. J Immunol I46:322 7.

43. Blankenstein, T., Z. H. Qin, K. Uberla, W. Muller. H. Rosen, H. D. Volk, and T.

Diamantstein. 199 1. Tumor suppression after tumor cell-targeted tumor necrosis factor

alpha gene transfer. J &CI Med 173: IO.17.

44. Ley, V., P. Langlade-Demoyen, P. Kourilsky, and E. L. Larsson-Sciard. 1991.

Interleukin 2-dependent activation of tumor-specific cytotoxic T lymphocytes in vivo.

Ezrr J lmrnzrnol2 1 :8j 1.

45. Cavallo, F., M. Giovarelli, A. Gulino, A. Vacca, A. Stoppacciaro. A. Modesti, and

G. Forni. 1992. Role of neutrophils and CD4+ T lymphocytes in the primary and memory

response to nonimmunogenic murine mammary adenocarchoma made immunogenic by

IL-2 gene. J hrnunol149:3627.

46. Karp, S. E., A. Farber, J. C. Salo, P. Hwu, G. M e , A. L. Asher, E. Shiloni, N. P.

Restifo, J. J. Mule, and S. A. Rosenberg. 1993. Cytokine secretion by genetically

modified nonimmunogenic murine fibrosarcoma Tumor inhibition by IL-2 but not tunor

necrosis factor. J Immun01 IjO:896.

47. Cornor, J., R. Banne rji, S. Saito, W. Heston, W. Fair, and E. Gilboa, 1993.

Regression of bladder tumors in mice treated with interleukin 2 gene- modified tunor

cells [published erratum appears in J Exp Med 1993 Jun 1 : 177(6):following 183 11. J Erp

Med 177:1127.

48. Porgador, A., B. Gansbacher, R. Bannerji, E. Tzehoval, E. Gilboa, M. Feldman,

and L. Eisenbach. 1993. Anti-metastatic vaccination of tumor-bearing mice with IL-2-

gene- inserted tumor cells. Int J Cancer 53:471.

49. Cohen, M. C., and S. Cohen. 1996. Cytokine function: a study in biologic

diversity. Am J Clin Parhol105:589.

50. Berinstein, N. L. 1998. Biological Therapy of Cancer. In The Basic Science of

O n c o l o ~ . 1. F. Tannock, and R. P. Hill, eds. McGraw-Hill Health Professions Division.

New York.

51. Watanabe, Y., K. Kuribayashi, S. Miyatake, K. Nishihara, E. Nakayama, T.

Taniyama, and T. Sakata. 1989. Exogenous expression of mouse interferon gamma

cDNA in mouse neuroblastoma Cl300 cells results in reduced tumorigenicity by

augmented anti-tumor imrnunity. Pruc Nat1 Acad Sci U S A 86:9456.

52. Stem, A. S., F. J. Podlaski, J. D. Hulmes, Y. C. Pan, P. M. Quim, A. G.

Wolitzky, P. C. Familletti, D- L. Stremlo, T. Truitt, R. Chizzonite, and et al. 1990.

Purification to homogeneity and partial characterization of cytotoxic lymphocyte

maturation factor fiom human B-lymphoblastoid celts- Proc Natl Acad Sci U S A

8 7: 6808.

53. Kobayashi, M., L. Fitz, M. Ryan, R M. Hewick, S- C. Clark, S- Chan, R- Loudon,

F. Sherman, B- Perussia, and G- Truichieri- 1989- Identification and purification of

natural killer ceIl stimulatory factor (NKSF), a cytokine with multiple biologic effects on

human lymphocytes. J EXp Med I70:827,

54. Schoenhaut, D. S., A. O. Chua, A. G. Wolitzky, P. M. Quim, C. M. Dwyer, W.

McComas, P. C. Familletti, M. K. Gately, and U. Gubler. 1992. Cloning and expression

of murine IL- 12. J Immunol H8:3433.

55. Gately, M. K., R. R. Warrier, S. Honasoge, D. M. Carvajal, D. A. Faherty, S. E.

Co~aughton , T. D. Anderson, U. Sarmiento, B. R. Hubbard, and M. Murphy. 1994.

Administration of recombinant IL- 12 to normal mice enhances cytolytic lymphocyte

activity and induces production of IFN-gamma in vivo. Int Immunol6: 157.

56. Gerrnann, T., M. K. Gately, D. S. Schoenhaut, M. Lohoff, F. Mattner, S. Fischer,

S. C. Jin, E. Schmitt, and E. Rude. 1993. Interleukin-I2/T ce11 stimulating factor, a

cytokine with multiple effects on T helper type 1 (TH) but not on Th2 cells. Eur J

Imrnunol23: 1762.

57. Kubin, M., M. Kamoun, and G. Tnnchieri. 1994. Interleukin 12 synergizes with

B7/CD28 interaction in inducing efficient proMeration and cytokine production of human

T cells. J Exp Med 2 8O:2ll.

58. Coughlin, C. M., M. Wysocka, H. L. Kurzawa, W. M. Lee, G. Triachieri, and S.

L. Eck. 1995. B7-1 and interleukin 12 synergistically induce effective antitumor

immunity. Cancer Res 55.4980.

59. Rao, J. B., R S. Chamberlain, V. Bronte, M. W. Carroil, K. R Irvine, B. Moss, S.

A. Rosenberg, and N. P. Restifo. 1996. IL-12 is an effective adjuvant to recombinant

vaccinia virus-based tumor vaccines: enhancement by simultaneous B7-1 expression. J

lmmunol156:3357.

60. Pizzoferrato, E., N. R. Chu, T. S. Hawley, F. H. Lieu, B. H. Barber, R. G.

Hawley, T. H. Watts, and N. L. Berinstein. 1997. Enhanced immunogenicity of B ce11

lymphoma genetically engineered to express both B7- 1 and interleukin- 12. Hum Gene

Ther 8 : Z 1 7 .

6 1. Fallarino, F., A. Ashikari, T. Boon, and T. F. Gajewski. 1997. Antigen-specific

regression of established tumors induced by active immunization with irradiated IL-12-

but not B7-1-transfected tumor cells. Int Immunol9:1259.

62. Mahvi, D. M., J. K. Burkholder, J. Turner, J. Culp, J. S. Malter, P. M. Sondel, and

N. S. Yang. 1996. Particle-mediated gene transfer of granulocyte-macrophage colony-

stimulating factor cDNA to turnor cells: implications for a clinically relevant tumor

vaccine. Hum Gene Ther 7: 1535.

63. Irvine, K. R., J. B. Rao, S. A. Rosenberg, and N. P. Restifo. 1996. Cytokine

enhancement of DNA immunization leads to effective treatment of established

pulmonary metastases. J Immunol 1 56:238.

64. Nabel, G. J., E. G. Nabel, 2. Y. Yang, B. A. Fox, G. E. Plautz, X. Gao, L. Huang,

S. Shu, D. Gordon, and A. E. Chang. 1993. Direct gene transfer with DNA-liposome

complexes in melanoma: expression, biologic activity, and lack of toxicity in humans.

Proc Nat2 Acad Sci U S A 90: 1 130 7.

65. Nabel, G. J., D. Gordon, D. K. Bishop, B. J. Nickoloff, 2. Y. Yang, A. Aruga, M.

J. Cameron, E. G. Nabel, and A. E. Chang. 1996. h u n e response in human melanoma

after transfer of an allogeneic class 1 major histocompatibility complex gene with DNA-

liposome complexes. Proc Narl Acad Sci U S A 93: Ij388.

66. Hawley, R- G-, F. H- Lieu, A- Z- Fong, and T- S- Hawtey- 1994- Versatile

retroviral vectors for potential use in gene therapy. Gene Ther 1:136.

67. Hitt, M., A. J. Bett, L. Prevec, and F. L. Graham. 1994, Construction and

Propagation of Hurnan Adenovirus Vectors. In Ce21 Biology: A laboratory handbook. J.

E . Celis, ed. Academic Press, San Diego.

68. Bett, A. J., W. Haddara, LI Prevec, and F. L. Graham. 1994. An efficient and

flexible system for construction of adenovirus vectors with insertions or deletions in early

regions 1 and 3. Proc Natl Acad Sci U S A 91:8802.

69. Mittal, S. K., A. J. Bett, L. Prevec, and F. L. Graham. 1995, Foreign gene

expression by human adenovins type 5-based vectors studied using firefly luciferase and

bacterial beta-galactosidase genes as reporters. Virology 210:226.

70. Li, Q., M. A. Kay, M. Finegold, L. D. Stratford-Perricaudet, and S. L. Woo. 1993.

Assessment of recombinant adenoviral vectors for hepatic gene therapy. Hum Gene Ther

./:-/O3.

71. Bramson, J. L., F. L. Graham, and J. Gauldie. 1995. The use of adenoviral vectors

for gene therapy and gene transfer in vivo. Curr Opin Biotechno26:590.

72. Yang, Y., F. A- Nunes, K. Berencsi, E. E. Furth, E, Gonczol, and J. M. Wilson.

1 994. Cellular immunity to viral antigens lirnits E 1 -deleted adenoviruses for gene

therapy . Proc Natl Acad Sci U S A 91:4307.

73. Waniier, G., M. T. Duffour, C. Uyttenhove, T. F. Gajewski. C. Lurquin, H.

Haddada, M. Perricaudet, and T. Boon. 1996. Induction of a cytolytic T-ce11 response in

mice with a recombinant adenovirus coding for tumor antigen PSISA. Int J Cancer

67:303.

74. Bramson, J. L., M. Hitt, C. L. Addison, W. J. Muller, J. Gauldie, and F- L.

Graham. 1 996. Direct intratumoral injection of an adenovirus expressing interleukin- 1 2

induces regression and long-lasting immunity that is associated with highly localized

expression of interleukin- 12. Hum Gene Ther 7: 1995.

75. Fenner, F. 1996. Poxvinises. In Fields ViroZogy. B. N. Fields, D. M. Knipe, and P.

M. Howley, eds. Lippincott-Raven Press, Philadelphia.

76. Moss, B. 1990. Poxviridae and their Replication. In Virology. B. N. Fields, and D.

M. Knipe. eds. Raven Press Limited, New York.

77. Goebel, S. J., G. P. Johnson, M. E. Perkus, S. W. Davis, J. P- Winslow, and E.

Paoletti. 1990. The complete DNA sequence of vaccinia virus. Virotogv 1 79:247.

78. Moss, B. 1991. Vaccinia virus: a tool for research and vaccine development.

Science 252: 1662.

79. Puisieux, I., L. Odin, D. Poujoly P. Moingeon, J. Tartaglia, W. Cox, and M.

Favrot. 1998. Canarypox virus-mediated interleukin 12 gene transfer into murine

mammary adenocarcinorna induces tumor suppression and long-term antitumoral

irnmunity [see comments]. Hum Gene Ther 9:248l.

80. Tartaglia, J., M. E. Perkus, J. Taylor, E. K. Norton, J. C. Audomet, W. 1. Cox, S.

W. Davis, J. van der Hoeven, B. Meignier, M. Riviere, and et al. 1992. NYVAC: a highly

attenuated strain of vaccinia v h . Virology 18812 1 7.

81. Kawakita, M., G, S. Rao, J. K. Ritchey, D. KI Omstein, M. A. Hudson, J.

Tartaglia, E. Paoletti, P. A. Humphrey, T. J- Harrnon, and T. L. Ratliff- 1997. Effect of

canarypox virus (AL,VAC)-mediated cytokine expression on murine prostate turnor

growth [see comments] . J Nat2 Cancer I . 8 9 : 4 2 8 .

82. Roth, J. A., and R. J. Cristiano. 1997. Gene therapy for cancer: what have we

done and where are we going? J Nat1 Cancer lnst 8912 1.

83. Paoletti, E., J. Tartaglia, and W. 1. Cox. 1993. Immunotherapeutic strategies for

cancer using poxvirus vectors, Ann N Y Acad Sci 6901293.

84. Toso, J. F., C. Oei, F- Oshidari, J. Tartaglia, E. Paoletti, H. K. Lyerly, S. Talib,

and K. J. Weinhold. 1996. MAGE- 1 -specific precursor cytotoxic T-lymphocytes present

among turnor- infiltrathg lymphocytes fiom a patient with breast cancer: characterization

and antigen-specific activation. Cancer Res 56: 16.

85. Roth, J., D. Dittrner, D. Rea, J. Tartaglia, E. Paoletti, and A. J. Levine. 1996. p53

as a target for cancer vaccines: recombinant canarypox virus vectors expressing p53

protect mice against lethal turnor ce11 challenge. Proc Naft Acad Sci U S A 93.4 78 1.

86. Hodge, J. W., J. P. McLaughlin, J. A. Kantor, and JI Schlom. 1997. Diversified

prime and boost protocols using recombinant vaccinia virus and recombinant non-

rep licating avian pox virus to enhance T-cell imrnunity and antitumor responses. Vaccine

15: 759.

87. Marshdl, J. L., M. J. Hawkins, K. Y. Tsang, E. Richmond, J. E. Pedicano, M. 2.

Zhu, and J. Schlom. 1999. Phase I shidy in cancer patients of a replication-defective

avipox recombinant vaccine that expresses human carcinoembryonic antigen. J C h

Oncol 17:332.

88. Matzinger, P. 199 1. The JAM test. A simple assay for DNA fragmentation and

ce11 death. Jimmunol Merhods 14.5:185.

89. Demkowicz, W. E., J. S. Maa, and M. Esteban. 1992. Identification and

characterization of vaccinia virus genes encoding proteins that are highly antigenic in

animals and are immunodominant in vaccinated humans. J C/irol66:386.

90. Krieg, A. M., A. K. Yi, S. Matson, T. J. Waldschmidt, G. A. Bishop, R. Teasdale,

G. A. Koretzky, and D. M. Klinman. 1995. CpG motifs in bacterial DNA tngger direct B-

ce11 activation. Naiure 374346.

91. Stacey, K. J., M. J. Sweet, and D. A. Hume. 1996. Macrophages ingest and are

activated by bacterial DNA. J lmmrtnol157:2116.

92. Ballas, S . K., W. L. Rasmussen, and A. M. Krieg. 1996. Induction of NK activity

in murine and human cells by CpG motifs in oligodeoxynucleotides and bactenal DNA. J

lrnmunol157:1840.

93. Jakob, T., P. S. Walker, A. M. Krieg, M. C. Udey, and J. C. Vogel. 1998.

Activation of cutaneous dendritic cells by CpG-containing otigodeoxynucleotides: a role

for dendritic cells in the augmentation of Thl responses by immunostirnulatory DNA. J

lmmunol161:3042.

94. Kundig, T. M., M.F. Bachmann,S. Oehen, U. W. H o f i a m , J. J. Simard, C. P.

Kalberer, H. Pircher, P. S. Ohashi, H. Hengartner, and R. M. Zinkernagel. 1996. On the

role of antigen in maintainhg cytotoxic T-ce11 memory. Proc Natl Acad Sci U S A

93~9716.

95. KaufÏnan, D. L., M. Clare-Salzler, J. Tian, T. Forsthuber, G. S. Ting, P. Robinson,

M. A. Atkinson, E. E. Sercarz, A. J. Tobin, and P. V. Lehmann. 1993, Spontaneous loss

of T-ce11 tolerance to glutamic acid decarboxylase in murine insuiin-dependent diabetes

[see comments] . Nature 366169.

96- McRae, B. L.+ C. L. Vanderlugt, M. C. Dal Canto, and S. D. Miller. 1995.

Fwictional evidence for epitope spreading in the relapsing pathology of experimental

autoimmune encephalomyelitis. J Ejrp Med 182: 75.

97. Lehmann, P. V., T. Forsthuber, A- Miller, and E. E, Sercarz. 1992- Spreading of

T-ce11 autoimmunity to cryptic determinants of an autoantigen. Nature 358: 155.

98. Cross, A- H-, V. K. Tuohy, and C. S. Raine. 1993. Development of reactivity to

new myelin antigens during chronic relapsîng autoimmune demyelination. Cell Immunol

1-/6:261.

99. Miller, S. D., C. L. Vanderlugt, W. S. Begolka, W. Pao, R. L. Yauch, K. L.

Neville, Y. Katz-Levy, A. Carrizosa, and B. S. Kim. 1997. Persistent infection with

Theiler's virus leads to CNS autoimmunity via epitope spreading [see comments]. Nat

Med 3r1133.

100. Vanderlugt, C. L., W. S. Begoka, K. L. Neville, Y. Katz-Levy, L. M. Howard, T.

N. Eagar, J. A. Bluestone, and S. D. Miller. 1998. The functional sigaificame of epitope

spreading and its regulation by CO-stimulatory molecules. Immunol Rev 164.63.

101. Alosco, T., B. A. Croy, B. Gansbacher, H. Q. Wang, U. Rao, and R. Bankert.

1993. Antitumor response independent of functional B or T lymphocytes induced by the

local and sustained release of interleukin-2 by the tumor cells, Cancer immun01

Imrnunother 3 6:3 64.

102. Siders, W. M., P. W. Wright, J. A. Hixon, W. G . Alvord, T. C. Back, R. H.

Wiltrout, and R. G. Fenton. 1998. T cell- and NK cell-independent inhibition of hepatic

metastases by systemic administration of an IL- 12-expressing recombinant adenovulis. J

Imrnunol Z6O:S.f6.j.

103. Gallimore, A., J. Hombach, T. Dumrese, H. G. Rammensee, R. M- Zinkernagel,

and H. Hengartner. 1998. A protective cytotoxic T ce11 response to a subdominant epitope

is influenced by the stability of the MHC class Ilpeptide complex and the overall

spectnim of Wal peptides generated within uifected cells. Eicr J Irnmunol28:3301.

104. Freitas, A. A., F. Agenes, and G. C. Coutinho. 1996. Cellular cornpetition

modulates survival and selection of CD8+ T cells. Eur J Immun01 26:2640.

Table 1 : Level of IL- 1 2 secreted by STF 1 O cells pnor to immunization of mice

Experiment purp~se STF 10 vaccine constituent Level of IL- 12 (ng/mL)

Tumorigenicity in Wild Vaccination Type BALBlc

Turnorigenicityinwild Prime Type BALBlc

(prime / boost)

Tumorigenicity in Wild Boost Type BALB/c

(prime / boost)

Tumorigenicity in Nude Vaccination Mice

Tumorigenicity in Nude Vaccination Mice

(Repeat)

Uninfected ALVAC ALVAC B7-1 ALVAC IL-12 ALVAC B7- I/ALVAC CL- 12

Uninfectecf ALVAC ALVAC B7-1 ALVAC IL- 12

Uninfected ALVAC ALVAC B7-1 ALVAC IL-12

Uninfected ALVAC ALVAC B7-I ALVAC L-12 ALVAC B7-I/ALVAC IL-12

Uninfecteci ALVAC ALVAC B7-1 ALVAC IL-12

4.5 f 0.56 Not Assayed Not Assayed 97.3 f: 4.4 54.5 I 2 - 4

Al1 Values are expressed in ng/ml/million ceIld24 hours

I

ALVAC hB7- 1 (vCP 1334)

C6 MCS

1 I ALVAC mIL- 12 (vCP 1303)

C5 MCS C6 MCS C5 MCS

Figure 1: Genomic Organization of recombinant ALVAC viruses

The ALVAC virus vector consists of a linear double stranded DNA molecule meüsuriiig about 325 Kbp. The multiple cloning sites C5 md C6 have been inserted us shown above. The ALVAC 87- 1 construct consists of a Iiumun 87- 1 cDNA dnven by the vaccinia H6 promoter inserted into the C6 MCS. The ALVAC IL- 12 constnict consists of two p35 subunit cDNA sequences each clriven by one E3L vaccinia promoter, and a p40 subunit cDNA driven by the entomopox 42K proinoter. The p35 and p40 subunit constructs are inserted into the C5 and C6 loci, respectivety.

A20 NFS 70 Cl498 B16 K46J

F

Fluorcsccncc Intensity

Figure 2: Heterogeneity in ALVAC Viral Transduction

To determine the tropism exhibited by the Canarypox virus ALVAC, cell lines of different lineages and stages were assayed for the ability to successfully take up the vinis and express the virally encoded protein B7-1. Clear histograms represent cells that were specifically stained by anti-human B7-1 antibody, while shaded Iiistograms represent staining via iln isotype matched non specific antibody. Result indicûte that recombinant 87- 1 is expressed only in select tissues represented by NFS-70, B 16, ST, as well as two ST subclones STFI O and STG4. A20, K46J. C 1498 and P8 15 are comyletely negative for ALVAC encoded produc ts.

94

CD 44 CD 25 HSA

I

Fluorcsccncc lnlcnsily

Figure 3: Surface Marker Phenotype of STF10

STFlO cells were staincd with antibodies specific for the surke-expressed proteins indicated, and then assayed by cell cytornetiy as outlined in Materiils and Methods. Clear histograrns represent specific staining for the mürker indicated, while shaded histograms represent irrelevant staining of STF10 vin a non-specific isotype inatched ontibody.

Water

NFS-70 Prep 3

NFS-70 Prep 2

NFS-70 Prep 1

ST-FI0 + ALVAC *fV 1 +RT

100 bp marker

Water Mouse Genomic DNA

NFS-70 Prep 1 / +RT

ST-FI 0 + ALVAC 1 +RT

Water

NFS-70 Prep 3

NFS-70 Prep 2

NFS-70 Prep 1

100 bp marker

Water NFS-70 Prep 3

NFS-70 Prep 1

STf I O Growth Characteristics

Cell Line:

O 20 40 60 80 100

Time Aïter Infection (Hours)

Figure 5: Growth Characteristics of STFlO Cells alter infection via ALVAC vectors

STFlO cells were infected with ALVAC 87-1 as outlined in Materials and Methods, and then replated in culture media at a concentration of 20,000 cellslml. Growth was monitored over a period of 4 days, and was compared to the growth rate of a STFlO culture that was plated at the same concentration, but was uninfected with ALVAC B7-1. Results indicate that infection of cells via recombinant ALVAC vectors does not alter the growth characterstics of the STFI O cells as the rate of expansion of the two cultures is similar.

6 Ilouts inlo Viriil Iiifcctioii 24 Hours Foltowitig Virril Iiifirciioii

Figure 6: E(iect of ALVAC viral Infection on MHC class 1 expression via STFlO cells

STFlO cells were infected with ALVAC 87-1 es outlined in Materials and Methods, and assayed for the expression of MHC 1 at tirnepoints corresponding to 6 hours into viral infection, aiid 24 hours following viral infection. Uesults indicnte that there is iio downregulütion of classical MHC class 1 ( H - 2 ~ b n d H-2Dd) as the georntric mean fluoresceiice of infected cells is 17 at 6 hours, and 19 at 24 hours, complired to uninfected cells whicli display il geoinetric mcan fluorescence of 14 at 6 hours and 15 at 24 hours.

98

- . . -

Survival of Mice Post STF10 Variant Injection

100 @-* a&

O 2 0 40 6 0

Dayr ?out Injection

STÇ10 Tumor Vaccines:

-+- Uninfected + ALVAC

-t ALVACIB7- 1 1

Figure 8: Survival of Mice Post STF10 Tumor Variant Vaccination

STFlO cells were infected with one of a) No Virus b) ALVAC c) ALVAC 87- 1 d) ALVAC IL- 12 or e) ALVAC IL12 & ALVAC B7- 1, and injected subcutaneously into mice (using 5 mice per vaccine). Mice rcceiving cells alone succumbed to tumor, whereas mice that received cells coniaining eithcr parental or recombinant ALVAC vectors, al1 rejected their tumors at a frequency of 100%.

Percentage Survival

1 Survid of Nude Mire Following STFIO Tumor Variant InjDrtion I A i o o i

80

(a 6 0 .L 2

40 al m m 0 2 0 2 al a O

O

STFI 0 Vaccine:

Days Post Injection - . . . . - .

-&- Üninfected + ALVAC + ALVAC 87-1 + ALVAC IL-12 +ALVAC 87-1 / IL-12

.. . . . -.. - - - ..

STF10 Vaccine::

Days Post Injection . . -

+ Uninfected + ALVAC + ALVAC 87-1 * ALVAC IL- 1 2 *AL 87-1/lL-12

Figure 10: Tumorigenicity of STFlO Variants in Nude Mice

STF! O cells were infected with parental or recombinant ALVAC viruses as described in Materials and Methods, and injected subcutaneously into nude mice. All onimals receiving cells alone succumbed to tumor in botli experiments (A & B). Animals receiving STFlO cells containhg either parental or recombinant ALVAC vectors displayed tumor rejection at statisitically significant (p < 0.05) frequencies, relative to inice receiving cells alone, in both independent experiments.

102

Suwival ot Mice Following Wild Type STF10 Challenge

P

O 20 40 6 0

0 8 ~ 8 P08t Challenge

Initial Vaccination Status:

-+- Naïve Control

+ ALVAC

+ ALVAC 87-1

-Jt ALVAC IL- 1 2

+ ALVAC 1112 (Unboosted)

-+ ALB71IALIL12

Figure 11: Survival of Mice Post W U Type STPlO Challenge Following Vaccine Boost

Mice were primed and boosted with cellular vaccines of a) STFlO / ALVAC b) STFlO / ALVAC 87 1 c) STFlO / ALVAC I L 4 2 d) STF IO/ALVAC B7- I and STF 10 1 ALVAC IL- 12. In addition, one group of mice was only primed with STF 10 1 ALVAC IL- 12, but left unboosted. Vaccinations and boosts were administered subcutaneously, on the right flank, with the boost taking place 35 days after the initial vaccination. 21 days post the boost, mice were challenged subcutaneously on the left flank with wild type STF10 cells and assayed for survival. Naive controls that had not been previously vaccinited al1 developed tumors, while al1 other vaccinated groups displayed improved survival relative to naive mice (p < 0.05 for al1 groups relative to nnive mice).

1 Target : STFlO 1 ALVAC IL-12 Cells 1

A Specltk Lyds et Day 17

w -, 5 0 Immunization Status -' 40 -& STF1 OIALVAC @ 30 IL-12

P 20 . .

V) + Naïve

8 O b

O 50 100 150

Effector : Target Ratio

B Speclflc Lysls i t Day 27 6 0 1 1

2 5 0 lmmuniratlon Status

+ STF? OIALVAC IL- 12

U)

/ + NaTve

$ 1 0

O 50 1 O0 150 EHsctor : Target Ratlo

C Speclflc Lysk at Day 17

1 Targets: Wild Type STFlO Cells I

lmmunization Status

O 5 0 100 150

Effector : Target Ratlo

Speciflc Lysis at Dey 27

lrnmuniratlon Status - . - + STF1 OIALVAC

IL-12

-C- Nalve

O 1 O0 200

Effector : Target Ratio

13, 17,27 riiid 35. Splcnocylcs wcrc stimulatcd with STFIOIALVAC ILI 2 nnd spccific killing iigiiinst STFIOIALVAC IL-12 (A & B) or wild type STFIO (C & D) cells wris cvaluatcd. Cytolytiç activity w u obscrvcd on dny 27 agninsi STFIOIALVAC IL-12 cclls (B) and nui STFIO cclls (D). No cyiolyiic iiciivity agninsi cithcr iypc of cclls was obscrvcd on dav 13 (nui show). dav 17 (A & C) und dav 35 (nui shown).

1 O4

- . . - - . -. - 1 In Mtro Stimulation with wild Type STFlO 1

Figure 13: ECtccts of Diflerential Stimulation on Splcnocytt Rtrctivity igainst STF10: Mice were immunized with STFlOlALVAC and then boosted with the same vaccine, 21 days followeing the boost, splenocytcs from an immunired mouse and R naïve conrol were Iiarvesied and stimulated with STFlO (A Rt B) or STFIO/ALVAC (C). Results indicate that cytolytic pwursors specific for STFlO are indeed present in the splenocyte culture (A). in addition to those specific for STF l OlALVAC (B); however. their detection is dependent on stimulation with wild type STFI O. Stiniulation via STP 1 OlALVAC does not yield entcient cytotoxicity against wild type STFI 0, relative to naive controls (C).

105