target cells of epstein-barr-virus (ebv)-positive post ...williams.medicine.wisc.edu/ptldebv.pdf ·...

For personal use. Only reproduce with permission from The Lancet Publishing Group.

MECHANISMS OF DISEASE

Summary

Background Epstein-Barr virus (EBV)-associated post-transplant lymphoproliferative disease (PTLD) encompassesa histologically broad range of lesions, arising from theexpanded pool of EBV-infected B cells in theimmunocompromised host. Identification of the precisecellular origin of these tumours could clarify theirpathogenesis.

Methods Of 13 cases of EBV-positive cases of PTLDcharacterised by histological analysis, pattern of EBV geneexpression, and clinical course, 11 had monoclonal orbiclonal lesions in which we determined the progenitor Bcell by immunoglobulin heavy chain (IgH) genotyping.

Results Two tumours had a naive B cell genotype and twoshowed patterns of IgH somatic mutation typical of antigen-selected (post-germinal-centre) memory cells. All four aroseearly post-transplant and expressed the markers of EBVtransformation—Epstein-Barr nuclear antigen (EBNA) 2 andlatent membrane protein (LMP) 1. However, seven tumours,either of early or late onset and including some withdownregulated EBNA 2 and LMP 1, arose from post-germinal cells with randomly mutated or sterile IgHgenotypes usually incompatible with B-cell survival in vivo.

Interpretation PTLD can arise from a broad range of targetB cells and not only from the pool of antigen-selectedmemory cells that EBV generally colonises inimmunocompetent individuals. Tumour development seemsfrequently associated with the EBV-induced rescue andexpansion of B cells that have failed the physiologicalprocess of germinal centre selection into memory. Thisfinding shows an unexpected connection betweenpathogenesis of PTLD and that of EBV-positive Hodgkin’slymphoma, another B-cell malignancy of atypical post-germinal-centre cell origin.

Lancet 2003; 361: 217–23See Commentary page 192

Cancer Research UK Institute for Cancer Studies (J M Timms MBChB,

A Bell PhD, Prof A B Rickinson PhD) and Department of Pathology(J R Flavell BSc, P G Murray PhD, Prof H-J Delecluse MD), University ofBirmingham, Birmingham, UK; and Service d’AnatomiePathologique, Centre Hospitalier Lyon-Sud, Pierre-Bénite, France (A Traverse-Glehen MD, Prof F Berger MD)

Correspondence to: Dr Henri-Jacques Delecluse, Department ofPathology, University of Birmingham, Birmingham B15 2TT, UK(e-mail: [email protected])

IntroductionEpstein-Barr virus (EBV), a gammaherpesvirus withpotent B-cell-transforming activity, is causatively linkedto three B-cell malignant diseases: endemic Burkitt’slymphoma, Hodgkin’s lymphoma, and post-transplantlymphoproliferative disease (PTLD).1 All three types oftumour have distinct cellular phenotypes and patterns ofEBV gene expression, suggesting that the pathogeneticrole of the virus might be different in each case. Burkitt’slymphoma is a monoclonal tumour whose phenotype andcontinuing mutation of immunoglobulin variable genesequences suggest that it originates from germinal centrecells.2,3 Although the regular presence of the virus intumour cells strongly suggests a causative role,expression of EBV antigen in Burkitt’s lymphoma isusually restricted to Epstein-Barr nuclear antigen 1(EBNA1), and cell growth seems to be mainly driventhrough deregulated expression of the cellular C-MYConcogene.1

Hodgkin’s lymphoma consists of a monoclonal tumourin which the malignant cell phenotype has no obviousnormal cellular counterpart. However, immunoglobulingene analysis has shown the disease to be derived fromatypical post-germinal-centre cells with random,frequently inactivating mutations—quite unlike themutations seen in classic antigen-selected memory Bcells.4,5 Such atypical B cells would normally have died byapoptosis within the germinal centre6 but seem to havebeen rescued and neoplastically transformed by acombination of EBV infection and as yet poorly definedcellular genetic change;7 viral antigen expression inHodgkin’s lymphoma is limited to EBNA1 and the twolatent membrane proteins (LMP) 1 and 2.

Current knowledge of PTLD suggests a less complexpathogenesis than those outlined above. Many cases ofPTLD, especially those arising within 1–2 years ofallograft, seem to be expansions of cells directlytransformed by EBV and growing opportunistically inthe absence of T-cell surveillance.1 These lesions, whichmay be monoclonal or polyclonal, are often dominatedby cells expressing the full spectrum of EBV latentproteins (ie, EBNA1, 2, 3A, 3B, 3C, and LP, and LMP1and 2), which are similar to EBV-transformedlymphoblastoid cell-lines in vitro.8,9 In immuno-competent individuals, EBV infection seems mainlyconfined to classic, post-germinal-centre, memory cells.First, in the tonsils of patients with primary EBV-infection, some virus-infected naive B cells are present,but virus-driven clonal expansions preferentially involvecells with a classic antigen-selected genotype.10 Second,EBV-positive B cells in the blood of virus carriersexamined long after primary infection are positive forsurface IgM, IgG, or IgA, but not for IgD; thisphenotype characterises most of the memory B-cellsubset.11,12 We sought to determine the type or types of Bcell from which PTLD arises.

Target cells of Epstein-Barr-virus (EBV)-positive post-transplantlymphoproliferative disease: similarities to EBV-positiveHodgkin’s lymphomaJudith M Timms, Andrew Bell, Joanne R Flavell, Paul G Murray, Alan B Rickinson, Alexandra Traverse-Glehen, Françoise Berger, Henri-Jacques Delecluse

Mechanisms of disease

THE LANCET • Vol 361 • January 18, 2003 • www.thelancet.com 217


MethodsTissue samplesOf PTLD cases at the E Herriot Hospital, Lyon, France,between 1991 and 2001, DNA extracted from freshtumours was available for 13 patients, as was formalin-fixed and (in some cases) frozen tumour material. Alltumours were classified histopathologically with WHOcriteria.13 Patients gave informed consent for tissuesamples to be taken. Ethical approval was given by theComité consultative de protection des personnes dans larecherche biomedicale de lyon A, Hôpital Hotel-Dieu,Lyon.

Immunophenotyping and immunohistologyAt presentation, tumour cell suspensions were stainedwith monoclonal antibodies for surface B cell (CD20,CD38) and T cell (CD3) markers, and for surface andintracytoplasmic immunoglobulin heavy (IgH) and light(IgL) chain isotypes. EBV status was established by in-situ hybridisation for Epstein-Barr-encoded RNAs(EBERs)14 and expression of EBNA2 and LMP1 bystaining of frozen tissue sections, where available, withmonoclonal antibodies PE2 and CS1–4.14 Thedetectability of both antigens, especially EBNA2, isreduced by formalin fixation; therefore, when onlyformalin-fixed tissue was available, sensitivity wasenhanced with antigen retrieval and a peroxidase-coupled antimouse IgG.

Immunoglobulin gene amplification, sequencing, andanalysis The IgH sequences corresponding to FRAMEWORK REGION

FR1, COMPLEMENTARITY DETERMINING REGIONS CDR1and CDR2, FR2, FR3, and the joining sequence JH wereamplified from total genomic DNA with established

primer combinations (van Dongen JJM, personalcommunication).15 Monoclonal PCR products werepurified from an 8% polyacrylamide gel, ligated to thepGEM-T Easy vector (Promega, Madison, WI, USA)and transformed into JM109 Escherichia coli competentcells. At least five independent clones containing insertswithin the expected size range were PCR sequenced withan Applied Biosystems (Foster City, CA, USA) ABI3700 automated sequencer; for most tumours we usedtwo independent rounds of amplification and assessedthe products independently.

We analysed sequences with GCG (version 10.2),MacVector, and Assembly Lign software, and alignedwith the closest germline sequences selected fromNational Center for Biotechnology Information andVbase databases with the Blastn program. We assessedchanges in IgH sequence from germline with twomethods developed for B-cell lymphomas16,17 on the basisof the fact that in antigen-selected B cells, replacementmutations outnumber silent mutations in the CDRs thatinteract with antigen, whereas silent mutationspredominate in the FRs to avoid any overall alteration offramework Ig structure. One method16 uses amultinomial statistical analysis (available at www-stat.stanford.edu/immunoglobulin) and gives theprobability of the observed mutation pattern in FRs andCDRs having arisen by chance; probability values of lessthan 0·05 are regarded as significant and thereforestrongly indicative of an antigen-selected IgH genotype.The other method,17 which focuses only on FRsequences, calculates the ratio of the sum of replacementto the sum of silent mutations (R/S ratio) shown by atumour or a group of tumours. R/S ratios of 1·1 or less,which are seen in normal memory B cells, stronglysuggest an antigen-selected genotype, whereas thosegreater than 1·9, which are typical of germinal centrecentroblasts before antigen selection, strongly indicate anon-antigen selected genotype.

Role of the funding sourceThe sponsors of this work had no role in study design,data collection, data analysis, or writing of the report.

ResultsTable 1 summarises clinical features of the 13 cases ofPTLD. 12 tumours showed histological features typicalof PTLD.13 Thus, cases 2, 4, and 5, which arose 1, 3, and8 months post-transplant, respectively, were poly-morphic, whereas all the rest, including some of theearliest onset cases, were classified as monomorphicdiffuse large cell lymphomas, with varying amounts ofplasmacytoid differentiation. These 12 tumours were all


218 THE LANCET • Vol 361 • January 18, 2003 • www.thelancet.com

GLOSSARY

IMMUNOGLOBULIN COMPLEMENTARITY DETERMININGREGIONS (CDRS)

Localised hypervariable sequences within the V region of theimmunoglobulin molecule that constitute the antigen binding site andthus confer antigen specificity.

IMMUNOGLOBULIN FRAMEWORK REGIONS (FRS)

Sequences within the variable (V) region of the immunoglobulin moleculethat are conserved and provide the structural framework of the protein.

SOMATIC HYPERMUTATION

A process in germinal centres whereby the immunoglobulin V regiongenes acquire mutations and thus generate antibody diversity. Only cellswith intact FRs and with CDRs providing high antigen affinity will surviveand be selected into memory.

Case Transplanted organ Time post-transplant Site of tumour Pathology EBER status Treatment Outcome

1 Kidney 1 month Lymph node Monomorphic + RIS Regression2 Kidney/pancreas 1 month* Lymph node Polymorphic + RIS Regression3 Heart 2 months Lymph node MonomorphicDLCBL + RIS Regression4 Bone marrow 3 months Lymph node Polymorphic + RIS Died after 9 days5 Liver 8 months In graft Polymorphic + RIS Regression6 Kidney 14 months In graft Monomorphic DLBCL + RIS + CD20 mAb Died7 Kidney 12 years Lymph node Monomorphic B+T +† RIS + Chemo Died after 16 months8 Bone marrow 3 weeks Lymph node Monomorphic DLBCL + RIS + CD20 mAb Died after 1 month9 Liver 18 months Stomach Monomorphic DLCBL/PD + RIS + Surgery Died after 18 months10 Kidney 18 months Bladder Monomorphic DLBCL/PD + RIS + Chemo Died after 15 months11 Heart 6 months Lung Monomorphic DLBCL/PD + RIS Died after 2 months12 Kidney 6 years Nasopharynx Monomorphic DLBCL + RIS Died after 1 day13 Heart 2·5 years Lymph node Monomorphic DLCBL/PD + RIS + Chemo Died after 5 months

DLBCL=diffuse large B-cell lymphoma. B+T=mixed B and T cell lymphoma. PD=plasmacytic differentiation. RIS=reduction of immunosuppression. mAb=monoclonalantibodies. Chemo=chemotherapy. +=positive. *Patient received two transplants and PTLD arose 1 year after 1st graft or 1 month after 2nd graft. †EBER-positive in B cells only.

Table 1: Clinical features of PTLD cases


of B-cell origin, expressed surface or intracellularimmunoglobulin, and lacked CD3; although most wereCD20-positive. Four cases (9, 10, 11, and 13) that hadstrong plasmacytic differentiation were CD20-negativebut expressed the plasmacytoid marker CD38. Case 7was the exception, presenting unusually late at 12 years’post-transplant as a composite B and T lymphoma. In all13 cases, the malignant B-cell population was EBER-positive. The tumours had a variable response totreatment by reduced immune suppression (combined insome cases with chemotherapy, treatment with anti-CD20 monoclonal antibodies, or surgery); four tumours,all of which arose 1–8 months post-transplant, regressed;the remainder did not.

Two tumours (cases 1 and 2, which both arose within1 month of grafting; table 2) expressed EBNA2 andLMP1 and were polyclonal by IgH and IgL isotypestaining. Seven tumours (cases 3–8 and 11) could also be classified as EBV-transformed lesions fromEBNA2/LMP1 co-expression on frozen sections (whenavailable) or LMP1 expression accompanied byborderline to definite EBNA2 staining on formalin-fixedbiopsy material. On the basis of surface IgH/IgL isotypestaining, one of these tumours (case 4) was judged to bebiclonal, four (5–8) monoclonal, and one (11)monoclonal with intracytoplasmic IgL only. Four othertumours did not express detectable EBNA2 or LMP1 (inthree cases based on the more sensitive staining of frozensections); these included two tumours (9 and 10) thatwere monotypic for surface IgH/IgL isotype staining, andtwo (12 and 13) with only monotypic intracytoplasmicIgL. These EBNA2/LMP1-negative cases tended to beamong the tumours with pronounced plasmacyticdifferentiation and presenting late (18 months–6 years)post-transplant.

The two tumours (1 and 2) that seemed polyclonal byisotype staining also gave a diffuse smear of IgH

amplification products, indicative of a polyclonalpopulation. The remaining eleven all yielded either oneor two distinct amplified bands that could be sequencedand compared with germline IgH sequences (table 2).Tumours from cases 3 and 4 each yielded twoamplification products, both of which showed productiveIgH rearrangements, suggesting that both tumours werecomposed of two independent neoplastic clones. In case4 this finding was consistent with the immuno-phenotyping data, suggesting that this was a biclonaltumour; no immunophenotyping was available for case 3.The IgH sequences of both clones in tumour 3 (3i, 3ii)and of one of the clones in tumour 4 (4i) were identical topublished germline sequences, and therefore veryprobably of naive B-cell origin. The two nucleotidechanges in clone 4ii compared with published germlinesequences might be due to germline polymorphism, butwe cannot exclude the possibility that the IgH gene couldhave mutated, especially since immunophenotypingshowed that one of the constituent populations in tumour4 was IgA-positive; B cells with a switched Ig phenotypeare thought to represent memory rather than naive cells.6

By contrast, the remaining nine tumours showed clearevidence of IgH mutations ranging from 1·8% to 10% ofthe analysed sequence, indicating that these tumours hadarisen from a post-germinal-centre B cell. However, threedifferent patterns of mutation could be discerned. One,shown by cases 5 and 6, had substantial diversificationfrom the germline sequence (6% and 7·3% respectively)but with preservation of the reading frame and adistribution of mutations characteristic of those noted inantigen-selected memory cells (R/S ratio <1·1 in FRs,and p<0·05 that the excess of silent mutations in the FRswas due to chance rather than to antigen selection). Wetherefore believe that tumours 5 and 6 have arisen fromclassic memory cells. Case 6 gave a second amplified IgHproduct that was not only mutated but also had a



Case EBNA2 Immuno- IgH family Number IgH Functional R/S (ratio) in FR* p (FR)† p (CDR)† Ongoing B cell of origin,/LMP1 phenotype usage mutations (%) IgH mutations IgH gene status

1 +/+ Polyclonal n/a n/a n/a n/a n/a n/a n/a n/a2 +/+ Polyclonal n/a n/a n/a n/a n/a n/a n/a n/a3i‡ +/+§ Not done VH3/DP46JH3 0 + n/a n/a n/a – Naive3ii‡ +/+§ Not done VH3/DP50JH4 0 + n/a n/a n/a – Naive4i‡ +/+ IgM, A, �, � VH3/DP47JH4 0 + n/a n/a n/a – Naive4ii‡ +/+ IgM, A, �, � VH3/DP49JH4 2 (0·73%) + 2/0 0·843 0·675 – Naive or memory,

non-ag selected5 +/+§ IgG, � VH3/DP77JH4 16 (6%) + 2/2 (1·0) <0·001 <0·001 – Memory,

ag selected6 +/+§ IgA, � VH1/DP8.5JH4 20 (7·3%) + 4/7 (0·57) 0·003 0·243 – Memory,

VH4/DP65JH1-2-4 n/a ag selected7 +/+§ IgA, � VH4/DP70JH4 5 (1·8%) + 4/0 0·851 0·476 + Post GC,

non-ag selected8 +/+ IgG, � VH3/DP46JH6 11 (4%) + 5/2 (2·5) 0·243 0·237 + Post GC,

non-ag-selected9 –/– IgG, � VH4/4.35JH6 25 (9·1%) + 14/2 (7·0) 0·338 0·127 – Post GC,

non-ag selected10 –/– IgG, � VH4/DP65JH4 28 (10%) + 11/4 (2·75) 0·057 0·029 – Post GC,

non-ag selected11 +/+ � VH1/DP14JH6 15 (6%) + 5/3 (1·66) 0·087 0·243 – Post GC,

sterile12 –/–§ � VH3/DP31JH4 22 (7·4%) – n/a n/a n/a + Post GC,

+ stop codon sterile13 –/– � VH3/DP35JH4 Large deletion – n/a n/a n/a – Post GC,

sterile

ag=antigen. GC=germinal centre. n/a=not applicable. *Number of replacement (R) versus silent (S) mutations in FR is indicated as well as the ratio of these numberswhere appropriate. †p indicates the probability that mutations in the FRs and in the CDRs were random rather than ag-selected. ‡i and ii represent different clones fromthe same cases. §Staining on paraffin sections.

Table 2: IgH genotype and presumed cell of origin of PTLD


deletion that inactivated the reading frame; this mightrepresent the second IgH allele of the same tumourclone.

Tumours 7–11 seem to be derived from somaticallymutated but non-antigen-selected B cells. In these cases,mutations again left the reading frame intact butreplacement changes were concentrated in FRs; thispattern strongly suggested a random rather than anantigen-selected process. In cases 7–10, the R/S ratios formutations in the FRs exceeded 1·9, and the R/S ratio incase 11 was 1·66, above the 1·1 value typical of antigen-selected cells. The same conclusion was reached fromstatistical analysis of the distribution pattern of mutations(p>0·05) in FR in all five cases, and in CDR in all butcase 10. Four of these tumours expressed a viable IgHchain (IgG or IgA) at the cell surface, but the tumourcells in case 11 were IgH-negative, suggesting thatinactivating mutations might have occurred outside theamplified IgH fragment.

Case 11 might, therefore, be closer to the third patternof sequence change shown by cases 12 and 13, both ofwhich showed clear evidence of inactivation of the IgHlocus (figure 1). In case 12, a stop codon had been

introduced within CDR1, which would prevent IgHchain translation. In case 13, the IgH allele was renderednon-functional by a large deletion that removed CDR2,CDR3, the diversity (D) region, and part of JH4, whileplacing downstream sequences out of frame. Thesefindings were precisely in line with the immuno-phenotyping results, which showed that the tumour cellsin cases 12 and 13 did not make a viable IgH product.

In most tumours with evidence of somaticallymutated IgH genes, every independently clonedamplification product from the IgH locus showed anidentical pattern of mutation, which is consistent with aclonally expanding tumour population with no furtherIgH mutations—ie, a post-germinal-centre genotype.However, in some tumours mutation seemed to becontinuing, since many of the cloned sequences hadacquired individual mutations in addition to those seen inthe majority consensus sequence; these were cases 7 and8, both tumours with an intact but randomly mutatedconsensus sequence, and case 12 in which the consensussequence itself already contained a stop mutation. Anexample of such diversification is shown in figure 2 withdata from tumour 8. In this one case, not only the



A

B

CDR1

CDR1

CDR2

CDR3

CDR2

Case 12

Case 12

Case 12

Case 13

Case 13

Case 13

Case 13

Figure 1: Sequence analysis of the non-functional IgH rearrangements in tumours 12 and 13(A) Sequence comparison between the VH region of a representative IgH sequence from tumour 12 and the most homologous germline gene VH3 DP31.Nucleotide identities are shown by dots, replacement substitutions in upper case, and silent mutations in lower case. The stop codon TGA in the tumourIgH sequence is shown in bold; above and below are the deduced IgH aminoacid sequences and the positions of complementarity-determining regionsCDR1 and CDR2. (B) Sequence comparison between a representative IgH sequence from tumour 13 and the most homologous IgH mRNA sequenceSC77u-09 (VH3 DP35 JH4). The 187bp out of frame deletion in the tumour IgH sequence which removes CDR2 and most of CDR3 is shown by the greyboxes.


original tumour but also a tumour-derived cell line wasavailable. Both the tumour and the cell line gave the same majority consensus sequence by immunoglobulingenotype analysis, and both showed some continuingdiversification, although in each case the changes notedwere unique to every individual amplified sequence.

DiscussionThe link between EBV infection, immunosuppression,and PTLD is well documented18,19 but the precisecircumstances leading to tumour growth are poorlyunderstood. We reasoned that, just as in Burkitt’slymphoma2,3 and Hodgkin’s disease,4,5 identification ofthe types of B cell from which PTLD arises could help toclarify tumour pathogenesis. Our PTLD series includednine tumours that expressed the EBNA2 and LMP1markers that show direct EBV transformation, all butone of which arose in the first year or so post-transplant.Within this group were seven biclonal or monoclonallesions in which identity of the target cell could beestablished. Two such cases (3 and 4) clearly carriednon-mutated IgH genes. Although these tumours couldhave arisen from rare cells selected into memory withoutaffinity maturation,6,7 a much more likely explanation isthat they are of naive B-cell origin. We infer that,although EBV seems to preferentially expand andcolonise memory B cells in the context of naturalinfection,10-12 there is no fundamental barrier to the EBV-induced transformation of naive B cells in vivo.

Two other tumours (5 and 6) that were positive forEBNA2 and LMP1 seemed to be of classic memory B-cell origin, on the basis of the number and distributionof nucleotide changes. By contrast, three tumours that were positive for EBNA2 and LMP1 and of post-germinal-centre origin (cases 7, 8, and 11) hadmany replacement mutations in framework regions quite atypical of antigen-selected cells. Indeed,immunophenotyping suggested that the IgH gene incase 11 had been inactivated, or the IgH protein had been rendered unstable, by mutations outside the amplified sequence. Of these three tumours, two (8and 11) presented as classic PTLD within 1 and 6 months of transplantation, whereas the other (7) wasunusual, arising 12 years’ post-transplant and presentingas a composite lymphoma with separate B-cell and T-cell populations. This situation has parallels withangioimmunoblastic lymphadenopathy, a lesion mainlycomposed of malignant T cells but in which EBV-positive B-cell expansions also occur and sometimesprogress to B-cell lymphoma.20 These three cases ofatypical B-cell origin were all non-responsive to reducedimmunosuppression, raising the possibility that despitehaving LCL-like patterns of EBV antigen expression,these tumours were less susceptible to EBV-specific T-cell control. Such loss of responsiveness could belinked to acquired cellular genetic changes, especially atthe BCL6 locus, which seems to be an inadvertent targetof the SOMATIC HYPERMUTATION machinery.21



Figure 2: Sequence analysis of ongoing somatic mutation in tumour 8 and the corresponding tumour-derived cell line Comparison between five tumour-derived VH sequences (8i–8v), six cell line-derived VH sequences (8vi–8xi) and the most homologous germlinesequence VH3 DP46. Nucleotide identities are shown by dots, replacement substitutions in upper case, and silent substitutions in lower case. CDR1and CDR2 are also marked.

CDR1

CDR2

8i8ii8iii8iv8v8vi8vii8viii8ix8x8xi




Four other EBV-positive PTLD, which generally aroselate between 18 months and 6 years of transplant, didnot have detectable expression of EBNA2 or LMP1,implying a more complex multistep pathogenesis inwhich viral transforming functions had eventually beensubsumed by cellular genetic change.21,22 These tumourswere also derived from post-germinal B cells whose IgHgenotype was inconsistent with antigen selection. In twocases (9 and 10) there was heavy mutation and anaccumulation of replacement changes in the FRs, thoughboth tumours retained functional IgH genes and indeedexpressed IgG by immunophenotyping. In the other two(12 and 13), no IgH protein expression could be seen byimmunophenotyping and the IgH gene had clearly beenrendered non-functional by mutation. Early work9 onPTLD shows a number of cases without IgH chainreactivity by immunophenotypic staining.

Some PTLDs of atypical B-cell origin even gaveindications of continuing IgH mutation, although, unlikein classic germinal centre tumours,23 there were no obviouslineage relations between the individual cells of thetumour. A similar event has been noted in the EBV-positive B-cell compartment of T-cell dominated lesions inangioimmunoblastic lymphadenopathy 20 and one of thePTLDs with such activity (case 7) did arise in the settingof a coresident T-cell lymphoma, thereby recapitulatingthe situation seen in angioimmunoblastic lymphadenopa-thy. Though its basis remains to be determined, suchcontinuing mutational activity might even contribute to theprocess of neoplastic change.

We conclude that, besides its more conventional targets,EBV can rescue and clonally expand post-germinal-centreB cells with IgH genotypes that are usually incompatiblewith cell survival in vivo. These cells are often the sourceof PTLD, and grow out either as a direct result of EBVtransforming activity itself, or after the subsequentacquisition of cellular genetic change. Indeed, suchatypical post-germinal-centre cells, if rescued by EBVinfection, may be especially prone to additional geneticchange. These findings suggest an unexpected connectionbetween PTLD and EBV-positive Hodgkin’s disease,which is itself a tumour of atypical post-germinal-centre Bcell origin that eventually presents with complexcytogenetic changes, a unique cellular phenotype, and adistinct viral antigen profile characterised by strong LMP1expression in the absence of EBNA2.24,25 EBV-positiveHodgkin’s lymphoma sometimes follows PTLD (Berger F,unpublished)26,27 and, although the incidence of Hodgkin’sis not as high as that of PTLD in immunocompromisedpatients, virtually all cases arising in transplant recipientsare EBV genome-positive,28,29 compared with only 30–40%of such tumours in the general population in Westerncountries. Perhaps EBV-infection and the subsequentsurvival/proliferation of atypical post-germinal-centre Bcells represent a common initiating step in thepathogenesis of both diseases.

ContributorsJ M Timms and A Bell contributed equally to this investigation; they didIgH PCR, DNA sequencing, and comparative IgH sequence analyses, andassisted in preparation of the manuscript. J Flavell and P Murraycontributed to the immunohistological studies. A Traverse-Glehenobtained clinical data and contributed to the immunohistological studies. F Berger obtained tissue samples, supervised the immunohistologicalstudies, provided pathological diagnoses, and contributed to editing of the manuscript. A B Rickinson and H J Delecluse were the principalsupervisors responsible for the design of the study, statistical analysis,interpretation of the IgH sequence data, and writing of the manuscript.

Conflict of interest statementNone declared.

AcknowledgmentsWe thank J J M van Dungen and colleagues for providing PCR primers(BIOMED2 study PL96–3936), and the Functional Genomics Laboratory(supported by BBSRC grant 6/JIF13209) and the Glaxo WellcomeBiocomputing Laboratory (supported by MRC grant 4600017), School ofBiosciences, University of Birmingham. This work was funded by grantsfrom Cancer Research UK to ABR and HJD, by a University HospitalBirmingham NHS Trust to HJD, and by a Wellcome Trust ClinicalResearch fellowship to JMT.

References1 Rickinson AB, Kieff E. Epstein-Barr virus. In: Knipe DM,

Howley PM, eds. Fields Virology. 4th edn. Philadelphia: Lippincott,Williams and Raven 2001: 2575–627.

2 Chapman CJ, Mockridge CI, Rowe M, Rickinson AB, StevensonFK. Analysis of VH genes used by neoplastic B cells in endemicBurkitt’s lymphoma shows somatic hypermutation and intraclonalheterogeneity. Blood 1995; 85: 2176–81.

3 Tamaru J, Hummel M, Marafioti T, et al. Burkitt’s lymphomasexpress VH genes with a moderate number of antigen-selectedsomatic mutations. Am J Pathol 1995; 147: 1398–407.

4 Kuppers R, Rajewsky K, Zhao M, et al. Hodgkin disease: Hodgkinand Reed-Sternberg cells picked from histological sections showclonal immunoglobulin gene rearrangements and appear to bederived from B cells at various stages of development. Proc Natl Acad Sci USA 1994; 91: 10962–66.

5 Kanzler H, Kuppers R, Hansmann ML, Rajewsky K. Hodgkin andReed-Sternberg cells in Hodgkin’s disease represent the outgrowth ofa dominant tumor clone derived from (crippled) germinal center Bcells. J Exp Med 1996; 184: 1495–505.

6 Rajewsky K. Clonal selection and learning in the antibody system.Nature 1996; 381: 751–58.

7 Kuppers R, Rajewsky K. The origin of Hodgkin and Reed/Sternbergcells in Hodgkin’s disease. Annu Rev Immunol 1998; 16: 471–93.

8 Young L, Alfieri C, Hennessy K, et al. Expression of Epstein-Barrvirus transformation-associated genes in tissues of patients with EBVlymphoproliferative disease. N Engl J Med 1989; 321: 1080–85.

9 Thomas JA, Hotchin NA, Allday MJ, et al. Immunohistology ofEpstein-Barr virus-associated antigens in B cell disorders fromimmunocompromised individuals. Transplantation 1990; 49: 944–53.

10 Kurth J, Spieker T, Wustrow J, et al. EBV-infected B cells ininfectious mononucleosis: viral strategies for spreading in the B cellcompartment and establishing latency. Immunity 2000; 13: 485–95.

11 Babcock GJ, Decker LL, Volk M, Thorley-Lawson DA. EBVpersistence in memory B cells in vivo. Immunity 1998; 9: 395–404.

12 Babcock GJ, Hochberg D, Thorley-Lawson AD. The expressionpattern of Epstein-Barr virus latent genes in vivo is dependent uponthe differentiation stage of the infected B cell. Immunity 2000; 13:497–506.

13 World Health Organisation Classification of Tumours. Pathologyand Genetics of Tumours of Haemopoietic and Lymphoid Tissues.Lyon: IARC Press, 2001.

14 Delecluse HJ, Kremmer E, Rouault JP, et al. The expression ofEpstein-Barr virus latent proteins is related to the pathologicalfeatures of post-transplant lymphoproliferative disorders. Am J Pathol1995; 146: 1113–20.

15 Aubin J, Davi F, Nguyen-Salomon F, et al. Description of a novelFR1 IgH PCR strategy and its comparison with three other strategiesfor the detection of clonality in B cell malignancies. Leukemia 1995;9: 471–79.

16 Lossos IS, Tibshirani R, Narasimhan B, Levy R. The inference ofantigen selection on Ig genes. J Immunol 2000; 165: 5122–26.

17 Klein U, Goossens T, Fischer M, et al. Somatic hypermutation innormal and transformed human B cells. Immunol Rev 1998; 162:261–80.

18 Hanto DW, Birkenbach M, Frizzera G, et al. Confirmation of theheterogeneity of posttransplant Epstein-Barr virus-associated B cellproliferations by immunoglobulin gene rearrangement analyses.Transplantation 1989; 47: 458–64.

19 Starzl TE, Nalesnik MA, Porter KA, et al. Reversibility oflymphomas and lymphoproliferative lesions developing undercyclosporin-steroid therapy. Lancet 1984; 1: 583–87.

20 Brauninger A, Spieker T, Willenbrock K, et al. Survival and clonalexpansion of mutating “forbidden” (immunoglobulin receptordeficient) epstein-barr virus-infected b cells in angioblastic t celllymphoma. J Exp Med 2001; 194: 927–40.

21 Cesarman E, Chadburn A, Liu YF, et al. BCL-6 gene mutations inposttransplantation lymphoproliferative disorders predict response totherapy and clinical outcome. Blood 1998; 92: 2294–302.

22 Knowles DM, Cesarman E, Chadburn A, et al. Correlativemorphologic and molecular genetic analysis demonstrates threedistinct categories of posttransplantation lymphoproliferativedisorders. Blood 1995; 85: 552–65.




23 Levy S, Mendel E, Kon S, Avnur Z, Levy R. Mutational hot spots inIg V region genes of human follicular lymphomas. J Exp Med 1988;168: 475–89.

24 Pallesen G, Hamilton-Dutoit SJ, Rowe M, Young LS. Expression of Epstein-Barr virus latent gene products in tumour cells of Hodgkin’s disease. Lancet 1991; 337:320–22.

25 Herbst H, Dallenbach F, Hummel M, et al. Epstein-Barr viruslatent membrane protein expression in Hodgkin and Reed-Sternbergcells. Proc Natl Acad Sci USA 1991; 88: 4766–70.

26 Nalesnik MA, Randhawa P, Demetris AJ, et al. Lymphomaresembling Hodgkin disease after posttransplant lymphoproliferativedisorder in a liver transplant recipient. Cancer 1993; 72: 2568–73.

27 Bierman PJ, Vose JM, Langnas AN, et al. Hodgkin’s diseasefollowing solid organ transplantation. Ann Oncol 1996; 7: 265–70.

28 Rowlings PA, Curtis RE, Passweg JR, et al. Increased incidence ofHodgkin’s disease after allogeneic bone marrow transplantation. J Clin Oncol 1999; 17: 3122–27.

29 Garnier JL, Lebranchu Y, Dantal J, et al. Hodgkin’s disease aftertransplantation. Transplantation 1996; 61: 71–76.



A loose sternotomy wire

Ryoichi Kondo, Masayuki Haniuda, Hirofumi Nakano, Jun Amano

Clinical picture

Department of Surgery (R Kondo MD, M Haniuda MD, H Nakano MD, J Amano MD), Shinshu University School of Medicine, Matsumoto 390-8621, Japan

A 49-year-old Japanese man underwent total resectionof mature teratoma through a median sternotomy onDec 12, 2000. The postoperative course was smooth andpostoperative chest radiograph showed normal findings,except for a loose wire loop, which was used for closingthe sternum, projecting dorsally in the lateral view(figure, lower). On Dec 24, the patient suddenlyexperienced syncope and developed circulatory andrespiratory failure. Emergency chest computedtomogram revealed copious clots and effusion in thepericardial cavity, especially around the projected wireloop (figure, upper). Urgent reoperation showed thewire loop penetrating the pericardium and that it hadnicked the anterior wall of the aortic root. The adventitiaof the aorta revealed an abrasion with a diameter ofaround 2 cm caused by the loose wire loop. Under extra-corporeal circulation, the abrased aorta was repaired bymattress suture using felt strips and wrapped. Thepatient tolerated the procedure well and had anuncomplicated postoperative course. Our experiencewould draw the attention of thoracic and cardiovascularsurgeons to one potential risk associated with the use ofwire.

target cells of epstein-barr-virus (ebv)-positive post ...williams.medicine.wisc.edu/ptldebv.pdf ·...

Documents