J Clin Pathol 1991;44:844-848

Detection of Epstein-Barr virus genomes inHodgkin's disease: Relation to age

R F Jarrett, A Gallagher, D B Jones, F E Alexander, A S Krajewski, A Kelsey, J Adams,

B Angus, S Gledhill, D H Wright, R A Cartwright, D E Onions

Leukaemia ResearchFund Virus Centre,Department ofVeterinary Pathology,University of Glasgow,Bearsden Road,Glasgow G61 lQHR F JarrettA GallagherS GledhillD E OnionsUniversityDepartment ofPathology,Southampton GeneralHospital,SouthamptonD B JonesD H WrightLeukaemia ResearchFund Centre forClinical EpidemiologySouthampton Site,UniversityDepartment ofMedical Oncology,Royal SouthantsHospitalF E AlexanderDepartment ofPathology, UniversityMedical School,EdinburghA S KrajewskiDepartment ofHistopathology, RoyalManchester Children'sHospitalA KelseyJ AdamsSchool of PathologicalSciences, University ofNewcastle Upon TyneB AngusLeukaemia ResearchFund Centre forClinicalEpidemiology, LeedsR A CartwrightCorrespondence to:R F JarrettAccepted for publication22 May 1991

AbstractAn investigation as to whether any

particular subgroup of patients withHodgkin's disease was particularly likelyto be Epstein-Barr virus (EBV) genome

positive was made on samples from 95patients. These were grouped accordingto age and Hodgkin's disease subtype,and analysed using Southern blotanalysis. Most samples from children or

adults aged 50 years or over containeddetectable EBV genomes; samples fromyoung adults were only rarely positive.The differences in EBV positivity by age

were highly significant, but there was no

significant association between EBV andhistological subtype after allowing forthe effect of age.The results support the hypothesis that

Hodgkin's disease in different age groups

may have different aetiologies, andsuggest that EBV does have a patho-genetic role in Hodgkin's disease inchildren and older age groups.

Accumulating evidence suggests that theEpstein-Barr virus (EBV) has a role in thepathogenesis of Hodgkin's disease. Initialstudies found raised antibody titres to EBV insuch patients,'3 and later it was shown thatincreased antibody titres occurred beforeHodgkin's disease was diagnosed.4 Usingsensitive molecular probes, EBV genomeshave been detected in tumour biopsyspecimens from a proportion of cases."' TheEBV genomes contained within the tumourshave been shown to be clonal, indicating thatthe EBV infected population has arisen from a

single infected cell.5""'" Furthermore, theEBV genomes have been localised to Reed-Sternberg cells by in situ hybridisation,79 10 14

and recent studies have shown that the EBVlatent genes EBER 1 and LMP are expressedby Reed-Sternberg cells (unpublished obser-vations).'4 '5 These latter observations supportthe argument that EBV has an aetiological rolein Hodgkin's disease and is not simplyreactivated during the course of the disease.A consistent feature of the molecular

studies is that EBV is detectable in only a

proportion of cases."' This figure varies from18-41% using Southern blot analysis." '3The failure to detect EBV in all cases does not

detract from the importance of the viral asso-

ciation as the pathological and epidemiologicalfeatures of Hodgkin's disease suggest that it is

a heterogeneous condition. The disease is

usually subdivided histologically into the foursubtypes described in the Rye classification.'6Epidemiologically, however, three groups ofcases can be distinguished and these cor-respond to the age groups 0-14 years, 15-34years, and 50 years and over.'7 There issome concordance between these two methodsof subgrouping Hodgkin's-for example,nodular sclerosis Hodgkin's disease (HDNS)accounts for the peak incidence in the agerange 15-34 years and mixed cellularityHodgkin's disease (HDMC) shows an in-creasing incidence with increasing age.'8Most groups studying the role of EBV in

Hodgkin's disease have reported the subtypedistribution of their cases. Using Southernblot analysis, four groups, includingourselves, have found a lower proportion ofEBV genome positive cases within the HDNSsubtype when compared with HDMC or allthe other subtypes grouped together. Thesedifferences, however, have not reachedstatistical significance.56" "' Significantassociations may have been masked either bythe small numbers of cases included in somestudies or the subjectivity of the histologicalclassification system used. In a larger series ofcases Pallesen et al, using immunohisto-chemistry, found a significant associationbetween EBV positivity and HDMC."

In our initial study of the associationbetween EBV and Hodgkin's disease it wasfound tht EBV positivity was significantlyassociated with increasing age.'3 It was alsonoted that the only two cases aged youngerthan 15 years in our study were EBVpositive. 3

The purpose of the present study, there-fore, was to determine whether the associationbetween age and EBV genome positivity re-mained following the analysis of a largernumber of cases in the age groups 50 yearsand over and younger than 15 years. Cases ofHDMC disease were also positively selectedto ascertain whether EBV was significantlyassociated with this histological subtype. Wealso EBV subtyped the positive cases toextend our previous observation that the EBVgenomes were of subtype A,'3 which isthought to be the prevalent strain of EBV inthe United Kingdom.'9

MethodsSamples oflymph node or spleen from a total of95 cases of Hodgkin's disease were examined.Thirty five of these have been reportedpreviously.'3 The cases were grouped accord-

844

on April 8, 2020 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.44.10.844 on 1 October 1991. D

ownloaded from

EBV in Hodgkin's disease

EBV positivity by age and histological subtype

Age and histological subtype

15-34y >50y

Selection procedure Paediatric HDNS HDMC Other 35-49y HDNS HDMC Other

Non-selected 2/2 2/22 2/4 0/2 3/8 6/7 2/2Selected for age 5/11 10/16 7/11 2/2Selected for age and subtype 2/8Total 7/13 2/22 4/12 0/2 3/8 16/23 9/13 2/2

"Other" includes cases of lymphocyte predominance and lymphocyte depleted Hodgkin's disease. Only non-selected cases orthose selected on the basis of age were included in the age analysis. The histological subtypes of cases in the young adults and the35-49 year age brackets are not given as the numbers in these groups were small; positive results were obtained from both HDMCand HDNS cases.

ing to age at date of biopsy and histologicalsubtype (table). Forty seven of the sampleswere from non-selected patients withHodgkin's disease; 40 were selected on thebasis of age; and a further eight were selectedfor both age and subtype to increase thenumbers in appropriate groups. These par-ticular age groups were chosen on the basis ofthe epidemiological data cited above.'7 Casesaged 35 to 49 years were not specifically selectedas they are thought to represent an overlapbetween the young adult (15-34 years) andolder (50 years and over) age groups. Casesaged less than 15 years were classified aspaediatric. Control samples included 35 ran-dom lymph nodes, 12 reactive lymph nodesfrom paediatric patients, and 14 lymph nodesfrom patients with non-Hodgkin's lymphomaaged over 60; the latter samples have beenreported previously.'3DNA was extracted from all samples using

standard procedures, and Southern blothybridisation was performed as described else-where.20 Southern blot hybridisation waschosen as the method of analysis in thesestudies as EBV genomes are only rarely detec-ted in non-Hodgkin's lymphomas using thistechnique.62' Notable exceptions to the aboveare Burkitt's lymphoma and lymphomas in theimmunosuppressed.2226 To assess the sensi-tivity of our Southem blot technique weanalysed 10 jMg of serial 10-fold dilutions ofDNA from the RAJI cell line. This cell line isestimated to contain 45 copies of the EBVgenome.27Two EBV probes were used. All samples

were hybridised with the BamHI-W fragmentwhich represents the internal repeat of EBVand therefore provides a sensitive indicator forits presence.' For this analysis 10 jig ofDNAwere digested with the restriction endonucleaseBamHI.Twenty six of the positive samples, includ-

ing samples from all groups, were subsequentlyhybridised with the EcoRI-D fragment ofEBV.28 This probe contains the terminalrepeats ofEBV and by virtue ofthis can be usedto assess the clonality of the infected popula-tion.29 For this analysis 20 jig of DNA weredouble-digested with BamHI. and EcoRI.Following digestion of circular EBV genomeswith BamHI and EcoRI, the EcoRI-D probedetects two fragments. One is a fragment of 6 5kilobase pairs which is of constant size in all

EBV genomes; the second varies in size fromone viral genome to another, depending on thenumber of terminal repeat sequences which itcontains. If the cell population infected withEBV has arisen from a single cell, infected at alow multiplicity of infection, then all thegenomes will contain identical numbers ofterminal repeats.29 This will result in thevisualisation of a single, discrete band inaddition to the 6-5 kilobase pair band.

In the above analyses DNA from placentaand from the infected cell line, B95-8,were used as negative and positive controls,respectively.The polymerase chain reaction (PCR) was

used to determine whether the EBV genomesdetected were of subtype A or B. The primersused in this assay and the reaction conditionshave been described elsewhere.3 30

Logistic regression was applied usingGENSTAT V."' Statistical testing usedanalysis of deviance and the asymptotic x2distribution for the deviance difference underthe null hypothesis. Each analysis was appliedto the largest number of results which wereappropriate from the table. In some analysesHDMC cases were grouped together withlympocyte predominance and lymphocytedepleted cases in a category designated all othersubtypes (HDAOS).

ResultsSENSITIVITY OF SOUTHERN BLOT ANALYSIS FORTHE DETECTION OF EBVFollowing Southern blotting and hybridisationwith the BamHI-W probe, EBV DNA wasdetectable in 10 jig of the I0O dilution ofRAJIDNA (data not shown). This suggests that wecan detect a single EBV genome present in lessthan 0-45% of the cells in any 10 jg sample.This figure is only an approximate value as thiscalculation does not allow for the variation inthe number of BamHI-W repeat sequences(usually seven to 12) present in different viralgenomes.

EBV IN PATIENTS AND CONTROLSThe results are summarised in the table. A totalof43 positive samples were identified; in these afragment of about 3-1 kilobase pairs, whichhybridised with the Bam HI-W probe, wasdetectable by autoradiography following twodays of exposure. Seventeen of the 47 non-

845

on April 8, 2020 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.44.10.844 on 1 October 1991. D

ownloaded from

Jerrett, Gallagher, Jones, Alexander, Krajewski, Kelsey, Adams, Angus, Gledhill, Wright, Cartwright, Onions

Figure I Southern blotanalysis of representative,positive Hodgkin's diseasesamples using the BamHI-Wprobe. Case 37: HDMCaged 14; case 74: HDNSaged 82; case 80: HDMCaged 54; case 85: HDMCaged 75; case 86; HDMCaged 72; case 88: HDLPaged 53. Additional bandsvisualised in B95-8, thepositive control, representthefragmentsflanking theBamHI-W repeatsequence.

selected cases wereEBV genome positive (fig 1).Two of the random lymph nodes were EBVpositive, one was a secondary nasopharnygealcarcinoma and the other was a B-cellimmunoblastic lymphoma; both malignancieshave been associated with EBV.6 21-23 One of the14 samples from patients over the age of60 withnon-Hodgkin's lymphoma was positive. Thissample, which was also from a B cell immuno-blastic lymphoma, gave a weakly positiveautoradiographic signal following a 10 dayexposure of the Southern blot.'3 None of thereactive nodes from paediatric patientscontained detectable EBV DNA.

EFFECT OF AGEIn the oldest age group 27 ofthe 38 (71%) cases

of Hodgkin's disease were positive and in the

paediatric group seven ofthe 13 (54%) sampleswere EBV positive. In contrast to these results,only four of the 28 (< 15%) cases in the youngadults aged 15 to 34 were positive. Sampleswhich were selected on the basis of both ageand subtype were excluded from the aboveanalyses. EBV was detected in three ofthe eightsamples from patients in the 35-49 year old agerange; this proportion of EBV positive cases isintermediate between that for the young adultand older cases.

Statistical testing of the above results,performed without adjustment for disease sub-type, indicated a trend for increasing EBVpositivity with age: x2 = 15 7 on 1 degree offreedom; p < 0-001. The departure from lineartrend was tested by adding a quadratic term tothe model: x2 = 9 6 on 1 degree of freedom;p < 0-01. This indicates that the trend for anincrease in positivity with increasing age is notlinear and infers that as a group the paediatriccases are different. The age groups 15-34 and 50years and over were then compared afteradjustment for subtype. The finding ofincreased positivity in the older age group washighly significant: x2 = 24-8 on 1 degree offreedom; p < 0000 1.

EFFECT OF HISTOLOGICAL SUBTYPEThere was a slight excess ofpositive cases in theHDMC group (table). This was statisticallytested in the two age groups 15-34 and 50 yearsand over; all cases in these age ranges wereincluded and adjustment for age was made.The effect of subtype was not significant wheneither HDNS, HDMC, and other subtypeswere compared (X2 = 0-73 on 2 degrees offreedom; p > 0 5), or when HDNS andHDAOS cases were compared (X2 = 0.73 on 1degree of freedom; p > 0 5). In the 15-34 yearage group there was an excess of positive casesin the HDAOS group, but this did not reachsignificance (Fisher's exact test, 1 tail; p =0-14).

i

4

l~4 4

4~~~mtw

___* 4

.4do

Figure 2 Panel A: Southern blot analysis of representative EBVgenome positiveHodgkin's disease samples using the EcoRI-D probe. Case 37: HDMC aged 14; case

40: HDMC aged 5; case 45: HDNS aged 24; case 54: HDMC aged 31; case 74;HDNS aged 82; case 82: HDMC aged 78; case 87: HD lymphocyte predominanceaged 65. In EcoRI and BamHI double-digested samples containing EBVDNA an

invariant band of 6 5 kilobase pairs is seen (just visible in case 54). In the Hodgkin'sdisease samples there is a single additional band of variable size (arrowed). Panel B:Southern blot analysis of case 60, HDNS aged 59, using the EcoRI-D probe. In this

case only the 6-5 kilobase pairfragment is detected. The ladder offragments seen in the

positive control lanes (B95-8) indicates the presence of linear EB Vgenomes and

episomalforms with different numbers of terminal repeats.29

CLONALITY OF EBV GENOMESThe clonality of the EBV genomes in 26 of thepositive cases was assessed by hydridisationwith the Eco RI-D probe, as outlined above. In25 of the 26 cases examined a single discretefragment, which varied in size from case tocase, was detected in addition to the invariant6 5 kilobase pair fragment (fig 2A). This sug-gests that the cell population infected withEBV in these cases is a clonal expansion of asingle infected cell. In the remaining sample,which was from a 59 year old man with HDNS(case 60), an intense invariant band at 6 5kilobase pairs was present, but no additionalbands were visualised (fig 2B). A similar resultwas obtained when this sample was digestedwith BamHI alone; the latter observationindicates that the fragment containing theterminal repeats was not comigrating with the6-5 kilobase pair band in the original analysis.

EBV SUBTYPEThirty of the EBV positive samples were EBVsubtyped using a PCR strategy; in all cases thegenomes were subtype A (data not shown).

846

on April 8, 2020 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.44.10.844 on 1 October 1991. D

ownloaded from

EBV in Hodgkin's disease

DiscussionIn this study we detected EBV genomes intumour samples from most of the paediatricand older patients with Hodgkin's disease.This contrasted with the case in young adultsin which EBV genomes were detected in-frequently. Statistical analysis showed that theassociation between age and EBV was highlysignificant. Analysis of control samples did notshow an excess of EBV positivity in lymphnodes from paediatric or older cases. Noadditional significant association between EBVpositivity and Hodgkin's disease subtypeemerged.Most groups who have studied the associa-

tion between EBV and Hodgkin's disease usingSouthern blot analysis have not examined theassociation between age and EBV positivity.Two exceptions to this are Boiocchi et al" andLibetta et al.'2 Boiocchi et al did not find anassociation between age and EBV positivity,but only five older cases were examined out of atotal of 17. In the study by Libetta et al therewas an excess of positive cases in older patientscompared with children and young adults, butthis was not significant. The results ofthis latterstudy are difficult to interpret as in over half ofthe cases which they considered EBV positivethe EBV BamHI-W probe hybridised to aDNA fragment of unexpected size.The possibility of an association between

EBV and Hodgkin's disease subtype has beenaddressed by most studies. In keeping with ourresults three other groups have reported a

slight excess of positivity among the HDMCor HDAOS cases, but this did not reachsignificance.56 1 Pallesen et al, using immuno-histochemistry, did find a significant associationbetween expression ofEBV and HDMC.'5 Ourresults suggest that subtype should not beanalysed separately from age as age is likely tobe an important confounding variable. This isof particular relevance here as the relativeincidence ofHDMC is greater in older patientswith Hodgkin's disease.'8

In the present study Southern blot analysiswas used to detect EBV genomes. Recentlythe PCR has been used to look for EBV inHodgkin's disease tumours and this hasresulted in the detection of larger proportions(50% and 58%) of positive samples.32"3 UsingSouthern blot hybridisation and the BamHI-W probe we detected a single EBV genome inabout 0 45% of the sample; this techniqueshould therefore be sufficiently sensitive todetect EBV in most cases of Hodgkin's diseaseif the EBV genomes are present in Reed-Stern-berg cells and their mononuclear equivalents.Indirect support from this comes from theresults of an earlier study in which we did notfind a correlation between Reed-Sternberg cellnumbers and EBV positivity. " Our assay mightnot have been sensitive enough to detect EBVin samples with low numbers of Reed-Stern-berg cells, but the method used avoids thepossibility of detecting low levels ofEBVDNApresent in residual B cells in the biopsy tissue.We have shown that in 25 of the 26 cases

examined the EBV genomes, in each case,

contained terminal fragments of identical size.

This finding correlates with clonality andtherefore suggests that the infected populationhas arisen from a single infected cell. This, inturn, supports the idea that EBV is involved inthe pathogenesis of Hodgkin's disease in thesecases.

In one case we could not detect a singlediscreet fragment containing the terminalrepeat sequences. The most likely explanationfor this is that the infected cells (either B cells orReed-Sternberg cells) are not clonal withrespect to EBV.The EBV DNA sequences from 30 of the

positive cases were subtyped into type A ortype B viruses using the PCR. All were type Aviruses. This result indicates that type Bviruses, which have been detected morefrequently in patients with immunosuppres-sion than in normal controls,34 are notoverrepresented in Hodgkin's disease tumours.The results of this study support the hypo-

thesis that Hodgkin's disease is a heterogen-eous condition with different aetiologies indifferent age groups."7 The finding that EBV ispresent in tumours from most paediatric andolder cases strengthens the argument that EBVis involved in the pathogenesis of these cases.The pattern ofEBV infection is not necessarilyidentical in these two age groups. We speculatethat de novo infection is important in thepaediatric group; reactivation of latent infec-tion, possibly as a result of declining T cellimmunity, would seem more likely in the olderpatients. The observation that most Hodgkin'sdisease tumours from patients infected with thehuman immunodeficiency virus (HIV) are alsoEBV positive935 provides some support for anassociation between decreased T cell immunityand EBV positivity, although it is possible thatother interactions are involved in the HIVpositive cases.35The low level of EBV positivity in young

adults, particularly in those with HDNS, sug-gests that other factors are involved in thesecases. This is particularly exciting as it is in thisgroup that there is most epidemiologicalevidence for a transmissable agent. Risk factorssuggestive of lack of early social contact havebeen found in the young adult peak incidencegroup, suggesting that delayed exposure to acommon infectious agent may be involved.36Clustering of Hodgkin's disease cases is alsomost likely to occur in cases aged under 35years and in HDNS.'8 37 We are currentlyinvestigating the possibility that other virusesare involved in this group.The possibility that EBV genome positivity

is associated with an unfavourable clinicaloutcome should also be evaluated further as ageover 50 years, HIV positivity, and HDMC,rather than HDNS, are all indicators of a poorprognosis."84

This work was supported by the Leukaemia Research Fund. Wethank Kathy Higginson for help with the collection of clinicalmaterial and Maisie Riddell for assistance with preparation ofthe manuscript.

1 Henle W, Henle G. Epstein Barr virus related serology inHodgkin's disease. National Cancer Institute Monograph1973;36:79-84.

2 Hesse J, Levine PH, Ebbesen P, Connelly RR, Mordhorst

847

on April 8, 2020 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.44.10.844 on 1 October 1991. D

ownloaded from

Jerrett, Gallagher, Jones, Alexander, Krajewski, Kelsey, Adams, Angus, Gledhill, Wright, Cartwright, Onions

CH. A case control study on immunity to two Epstein-Barr virus-associated antigens, and to herpes sirnplexvirus and adenovirus in a population-based group ofpatients with Hodgkin's disease in Denrmark, 1971-73. IntJ Cancer 1977;19:49-58.

3 Evans AS, Gutensohn M. A population-based case-controlstudy of EBV and other viral antibodies among personswith Hodgkin's disease and their siblings. Int J Cancer1984;34: 149-57.

4 Mueller N, Evans A, Harris NL, et al. Hodgkin's disease andEpstein-Barr virus. Altered antibody pattern before diag-nosis. N Engl J Med 1989;320:689-95.

5 Weiss LM, Strickler JG, Warnke RA, Purtilo DT, Sklar J.Epstein-Barr viral DNA in tissues of Hodgkin's disease.Am JPathol 1987;129:86-91.

6 Staal SP, Ambinder R, Beschorner WE, Hayward GS,Mann R. A survey of Epstein-Barr virus DNA in lym-phoid tissue. Frequent detection in Hodgkin's disease. AmJ Clin Pathol 1989;91:1-5.

7 Weiss LM, Movahed LA, Warnke RA, Sklar J. Detection ofEpstein-Barr viral genomes in Reed-Sternberg cells ofHodgkin's disease. N Engl J Med 1989;320:502-6.

8 Herbst H, Tippelmann G, Anagnostopoulos I, et al.Immunoglobulin and T-cell receptor gene rearrange-ments in Hodgkin's disease and Ki-l-positive anaplasticlarge cell lymphoma: Dissociation between phenotype andgenotype. Leuk Res 1989;13:103-16.

9 Uccini S, Monardo F, Ruco LP, et al. High frequency ofEpstein-Barr virus genome in HIV-positive patients withHodgkin' disease. Lancet 1989;i:1458.

10 Anagnostopoulos I, Herbst H, Niedobitek G, Stein H.Demonstration ofmonoclonal EBV genomes in Hodgkin'sdisease and KI-1 -positive anaplastic large cell lymphomaby combined Southern blot and in situ hybridization.Blood 1989;74:810-16.

11 Boiocchi M, Carbone A,De ReV, Dolcetti R. Is the Epstein-Barr virus involved in Hodgkin's disease? Tumori1989;75:345-50.

12 Libetta CM, Pringle JH, Angel CA, CraftAW, Malcolin AJ,Lauder I. Demonstration of Epstein-Barr viral DNA informalin-fixed, paraffin-embedded samples of Hodgkin'sdisease. JPathol 1990;161:255-60.

13 Gledhill S, Gallagher A, Jones D, et al. Viral involvement inHodgkin's disease: Detection of clonal type A EBVgenomes in tumour samples. Br J Cancer (in press).

14 Wu TC, Mann RB, Charache P, et al. Detection ofEBV geneexpression in Reed-Sternberg cells of Hodgkin's disease.Int J Cancer 1990;46:801-4.

15 Pallesen G, Hamilton-Dutoit SJ Rowe M, Young LS.Expression of Epstein-Barr virus latent gene products intumour cells of Hodgkin's disease. Lancet 1991;337:320-2.

16 Lukes RJ, Craver LF, Hall TC, Rappaport H, Ruben P.Report of the nomenclature committee. Cancer Res1966;26:131 1.

17 MacMahon B. Epidemiology of Hodgkin's disease. CancerRes 1966;26:1 189-200.

18 Alexander FE, McKinney PA, Williams J, Ricketts TJ,Cartwright RA. Epidemiological evidence for the 'two-disease hypothesis' in Hodgkin's disease. Int J Epidemiol(In press).

19 Young LS, Yao QY, Rooney CM, et al. New type B isolatesof Epstein-Barr virus from Burkitt's lymphoma and fromnormal individuals in endemic areas. J Gen Virol1987;68:2853-62.

20 Jarrett RF, Gledhill S, Qureshi F, et al. Identificationof human herpesvirus 6-specific sequences in twopatients with non-Hodgkin's lymphoma. Leukemia 1988;2:496-502.

21 Andiman W, Gradoville L, Heston L, et al. Use of clonedprobes to detect Epstein-Barr viral DNA in tissues ofpatients with neoplastic and lymphoproliferative diseases.J Infect Dis 1983;148:967-77.

22 Epstein MA, Achong BA, eds. The Epstein-Barr virus: recentadvances. New York: John Wiley, 1986:1-11.

23 Frizzera G. The clinico-pathological expressions ofEpstein-Barr virus infection in lymphoid tissues. Virchows Arch(Pathol Anat) 1987;53:1-12.

24 Zutter MM, Martin PJ, Sale GE, et al. Epstein-Barr viruslymphoproliferation after bone marrow transplantation.Blood 1988;72:520-9.

25 Nalesnik MA, Jaffe R, Starzl TE, et al. The pathology ofposttransplant lymphoproliferative disorders occurring inthe setting of cyclosporine A-prednisone immuno-suppression. Am J Pathol 1988;133:173-92.

26 Hamilton-Dutoit SJ, Pallesen G, Karkov J, Skinhoj P,Franzmann MB, Pedersen C. Identification ofEBV-DNAin tumour cells of AIDS-related lymphomas by in situhybridisation. Lancet 1989;i:554-5.

27 Sternas L, Middleton T, Sugden B. The average number ofmolecules of Epstein-Barr nuclear antigen 1 per cell doesnot correlate with the average number of Epstein-Barrvirus (EBV) DNA molecules per cell among differentclones of EBV-immortalized cells. J Virol 1990;64:2407-10.

28 Arrand JR, Rymo L, Walsh JE, Bjorck E, Lindahl T, GriffinBE. Molecular cloning of the complete Epstein-Barr virusgenome as a set of overlapping restriction endonucleasefragments. Nucleic Acids Res 1981;9:2999-3014.

29 Raab-Traub N, Flynn K. The structure of the termini of theEpstein-Barr virus as a marker of clonal cellular prolifera-tion. Cell 1986;47:883-9.

30 Sample J, Young L, Martin B, et al. Epstein Barr virus type 1and 2 differ in their EBNA 3a, EBNA 3b and EBNA 3cgenes. J Virol 1990;64:4084-92.

31 Payne RW, Lane PW, Ainsley AE, et al. Genstat V:References manual. Oxford: Clarendon Press, 1987.

32 Bignon YJ, Bernard D, Cure H, et al. Detection of Epstein-Barr virus genomes in lymph nodes of Hodgkin's diseasepatients. Mol Carcinog 1990;33:9-1 1.

33 Herbst H, Niedobitek G, Kneba M, et al. High incidence ofEpstein-Barr virus genomes in Hodgkin's disease. Am JPathol 1990;137:13-18.

34 Sixbey JW, Shirley P, Chesney PJ, Buntin DM, Resnick L.Detection of a second widespread strain of Epstein-Barrvirus. Lancet 1989;ii:761-5.

35 Uccini S, Monardo F, Stoppacciaro A, et al. High frequencyof Epstein-Barr virus genome detection in Hodgkin'sdisease of HIV-positive patients. Int J Cancer 1990;46:581-5.

36 Gutensohn NM. Social class and age at diagnosis ofHodgkin's disease: New epidemiologic evidence for the"two-disease hypothesis". Cancer Treat Rep 1982;66:689-95.

37 Alexander FE, Williams J, McKinney PA, Ricketts TJ,Cartwright RA. A specialst leukaemia/lymphoma registryin the UK. Part 2: Clustering of Hodgkin's disease. Br JCancer 1989;60:948-52.

38 Gobbi PG, Cavalli C, Federico M, et al. Hodgkin's diseaseprognosis: A directly predictive equation. Lancet1988;i:675-8.

39 Walker A, Schoenfeld ER, Lowman JT, Mettlin CJ, Mac-Millan J, Grufferman S. Survival of the older patientcompared to the younger patient with Hodgkin's disease.Cancer 1990;65:1635-40.

40 Serrano M, Bellas C, Campo E, et al. Hodgkin's disease inpatients with antibodies to human immunodeficiencyvirus. Cancer 1990;65:2248-54.

848

on April 8, 2020 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.44.10.844 on 1 October 1991. D

ownloaded from