Immunological studies Burkitt's lymphoma · Immunologicalstudies onBurkitt's lymphoma...

Postgraduate Medical Journal (March 1971) 47, 141-155.

Immunological studies on Burkitt's lymphoma

Leukemia Research Fund Lecture

Delivered at the Institute of Child Health, London, on 28 September, 1970

GEORGE KLEIN

Department of Tumor Biology, Karolinska Institutet, Stockholm, Sweden

A FEW years ago, it became clear, in the progress ofexperimental tumor immunology'-7 that all virus-induced tumors carry a common, group-specifictransplantation antigen, capable of inducing rejec-tion reactions in syngeneic hosts. The question arosewhether it would be possible to utilize this fact forthe search after possibly similar, virally determinedtransplantation antigens in tumors of unknownetiology, particularly human. Our attention turnedto Burkitt's lymphoma for three reasons:

(a) its postulated viral etiology8;(b) clinical observations'-" strongly suggesting

that host-defense reactions may plan an importantrole in this disease and decisively influence the out-come of chemotherapy;

(c) among the virally induced experimentaltumors, lymphomas lend themselves particularlywell for the demonstration of virally determinedcommon antigens, partly because it is easy to pre-pare free-cell suspensions, with a high frequency ofviable cells, and partly because humoral antibodiesare more easily demonstrable, as a rule, againstsurface antigens carried on lymphoma cells than onlarge carcinoma or sarcoma cells.The first question we asked appeared quite

straightforward: was it possible to obtain by testingthe reactivity of live Burkitt lymphoma (BL) cells,with patients' sera, and in comparison with appro-priate controls, evidence that would indicate theexistence of characteristic cell-membrane-associatedantigens, in analogy with the virally induced murineleukemias and, if so, could this information help toelucidate the etiology of the disease, as well as thepossible role of host defense reactions that mayinfluence its clinical course? To approach thisproblem, we chose the technique that was most

The studies of the author and his associates have beensupported by the Swedish Cancer Society, by Grant CA-04747and Contract No. NIH-69-2005 from the National Institutesof Health and a grant from the Damon Runyon Foundation.

sensitive in the experimental leukemia studies,'2 13viz. membrane immunofluorescence with viabletarget cells."4 The findings, summarized briefly in thefollowing chapters, essentially confirmed the ex-pectations, but they have also led to many un-expected observations and raised new dilemmas.Some of them may serve to exemplify the problemsencountered during the transition from the experi-mental to the human situation.

Studies on BL biopsy cellsDuring the first phase of this work, fresh BL

biopsy cells were exposed to the sera of BL patientsand various other donors and we were looking forattached immunoglobulins by the indirect membranefluorescence technique.101 102, 11O The sera of BLpatients reacted more frequently than African con-trol sera from donors with other neoplastic ornon-neoplastic diseases. The possibility that the re-activity of the BL sera was due to isoantibodiesbecame unlikely when it was found that autoch-thonous serum-cell combinations gave positivereactions in five of six cases where this could betested. It turned out, furthermore, that the mostregularly positive sera have been derived frompatients whose tumors have gone to total regressionafter chemotherapy. For this reason, the autoch-thonous target cell was frequently unavailable fromhighly positive serum donors. To exclude isoanti-bodies, such sera were tested in parallel series againstlymphoma cells and normal bone marrow cellsderived from the same allogeneic BL donor.Lymphoma cells, but not bone marrow cells,reacted regularly in such tests, thus increasing theprobability that the reactivity of the BL serum-cellcombinations could not be simply due to the presenceof isoantibodies. This was further reinforced by thefinding that lymphoid cells of normal donors and ofdonors with different types of leukemias and otherlymphoreticular diseases also failed to react.

Protected by copyright.

on Decem

ber 22, 2020 by guest.http://pm

j.bmj.com

/P

ostgrad Med J: first published as 10.1136/pgm

j.47.545.141 on 1 March 1971. D

ownloaded from

http://pmj.bmj.com/

George Klein

While this was encouraging, further studies on thespecificity of the reaction were hampered by thegreat variability of the biopsy preparations. Onemajor source of the difficulty was the variable degreeof immunoglobulin coating on the surface of thebiopsy cells. This coating was detected by directmembrane immunofluorescence with conjugatedanti-immunoglobulin reagents. It could be of twobasically different kinds: IgM and/or IgG, showingnot only a difference in class specificity but also adifference in behaviour in relation to the course ofthe disease."5 16 In the cases where the cell surfacereacted with anti-IgM conjugates, reactivity wasusually expressed on 100% or nearly 100% of thecells. When such cells were converted into estab-lished lines in vitro, their 'IgM-ring' was maintainedduring long-term propagation. The membrane-IgMreactive lines did not secrete IgM into the medium. 17Preliminary characterization of the reactive sub-stance in membrane fractionation experiments18indicates, at least as far as one cell of this type isconcerned, that 7S-size IgM subunits with [L- andK-chain specificities are integrated into the cellmembrane. Conceivably, this is the neoplasticvariety of a normal lymphoid cell that incorporatesmolecules of this type into its plasma membrane aspart of its normal differentiation. Lymphoid cells ofthis type have been postulated to play an importantrole in immunological memory and/or delayed hyper-sensitivity.19 The lymphoma cells may represent theneoplastic variant in the same way as myeloma cellsproject normal immunoglobulin-secreting plasmacells into a magnified, neoplastic image. Thephenomenon is not exclusive for BL; a Swedish caseof chronic lymphatic leukemia has been found withthe same cellular characteristics. 20 Quantitative com-parisons between different lymphoma lines of thistype showed'8 that the different lines carry verydifferent amounts of IgM-kappa reactive materialon their membrane. Although an extensive searchhas been made on more than sixty biopsies andderived lines, so far no line has been found thatwould carry other types of membrane-associatedimmunoglobulins than IgM-kappa.Whatever the nature of the cell that carries mem-

brane-associated immunoglobulins, it is importantto note in relation to the present discussion that thisproperty has always behaved as a cell marker whenrepeated biopsies were taken from the same patient.If it was present on the cells of a given tumor, it wasmaintained unchanged in the course of repeatedbiopsies; if it was absent, it remained absent. It wasalso maintained following successful heterograftingof a membrane-IgM-positive BL cell to the rat.2'The IgG coat behaved quite differently. It was rarelypresent on untreated BL biopsy cells, but it tended toappear if the tumor persisted in spite of treatment.

It accumulated following a recurrence that waspresumably due to the selection of a tetraploid,probably immunoresistant, BL cell variant.16' 22 'Self-enhancement', i.e. the accumulation of 'blocking'antibodies that prevent the access of immunelymphoid cells7 is an obvious possibility.Whether these considerations are realistic or not,

the changing pattern ofIgG coating with time and itsfailure to persist on derived in vitro lines,'5' 23 indi-cate that it is due to coating from the outside, unlikethe membrane-associated IgM, that appears torepresent a special type ofproduction from the inside.The presence of preformed immunoglobulin on

the cell surface may interfere with the indirect mem-brane immunofluorescence reaction and, whenpresent in subliminal degree, it probably explainssome of the variability encountered when biopsycells are used as targets. In order to avoid thisvariability, we started looking for more standardizedtarget cells and turned to established culture lines.

Experiments with established tissue culture linesA number of BL-derived lymphoblastoid cell lines

growing in stationary suspension cultures were testedagainst BL sera that reacted regularly with BLbiopsy cells, and were free of demonstrable isoanti-bodies.24 The pattern appeared strange but interest-ing. Four BL-derived lines gave positive membraneimmunofluorescence reactions in the indirect test,after exposure to the reference serum 'Mutua'(derived from a BL patient in long-term regression)whereas three BL-derived lines were negative.Eight control lines derived from various leukemiasand, in one case, from a normal donor, werenegative as well. At first, we could not understandthis pattern. A clue was obtained, however, whenthese results were compared with the reactivity ofthe same cell lines in the Henle test25 known todetect EB-viral (probably nucleocapsid) antigens.In carrier cultures, these antigens are present in asmall frequency of the cells, as a rule. These cellsshow degenerative features and, when simultaneousimmunofluorescence and electron microscopy areperformed26, 27 turn out to contain herpes-typeEpstein-Barr (EB) virus28 particles. The first com-parison of the membrane and EBV test revealed29that the four membrane-positive lines containedEBV-antigens in more than 1% of the cells whereasthe membrane-negative lines were either EBV-negative or contained a very small frequency of posi-tive cells (less than 1%).

This suggested that the membrane antigen de-tected by this reference serum may be determined bythe genome of the EB virus. More conclusive evidencewas obtained in a prospective study.229 Fourteen newlines were established from biopsies received fromNairobi, and the frequency of EBV-positive and of

142P

rotected by copyright. on D

ecember 22, 2020 by guest.

http://pmj.bm

j.com/

Postgrad M

ed J: first published as 10.1136/pgmj.47.545.141 on 1 M

arch 1971. Dow

nloaded from

http://pmj.bmj.com/


membrane-reactive cells was determined in parallel,on coded specimens, at two different laboratories.The same relationship was found as in the prelimi-nary retrospective study: only the lines that carrieda relatively high 'EBV-load' showed a positivemembrane antigen reactivity. In the reactive lines,the frequency of membrane-positive cells wasapproximately ten times higher than the frequencyof EBV-positive cells. The biopsies from which thelines were derived were membrane-positive, but EBV-negative, as a rule. EBV reactivity appeared duringthe first week in culture. This suggests that the pro-duction of the viral nucleocapsid antigen is sup-pressed in the tumor cell in vivo. The suppressivefactor could be antibody, but there are many otherpossibilities. Another curious observation was thatrepeated establishment of parallel lines from thesame patient, derived from successive biopsies, ledto lines with fairly similar EBV levels, whereas linesderived from different patients were quite different.30This suggests that the viral 'load' per cell, or theactivatability of the virus, or both, are characteristicfor the individual tumor. Since the membrane-associated IgM marker, mentioned above, andanother study with G6PD-isozyme markers31 stronglyindicate that the BL process has a clonal origin, thismay reflect the virus-cell relationship that charac-terizes a particular clone. This is also suggested bythe closely similar levels of EBV-DNA hybridizablecellular DNA in repeated biopsies taken from mul-tiple tumors from the same BL patient.32The postulate that the membrane antigen is deter-

mined by the EB virus, was directly confirmed whenit was found that it can be induced to appear inEBV-negative lines by the admixture of heavilyirradiated EBV-carrying cells,11' or by infectionwith EBV concentrates.33 3 In the infected cells,membrane antigen appeared after 20-24 hr. DNA-inhibitors such as cytosine arabinoside (Ara C) orIUDR did not prevent its appearance, whereas puro-mycin inhibited it completely. It behaves, in otherwords, like an 'early' product of the viral genome,not requiring viral DNA synthesis. In this respect, itshows certain parallels with membrane and Tantigens found in experimental oncogenic DNAvirus systems.35' 36, 6Although the relationship between EBV and the

membrane antigen was clarified by these studies,this applies only to the EBV-carrier cultures in vitroand it must be kept in mind that similar compellingevidence is lacking about the connection between themembrane antigens detected on the biopsy cells andthe virus, although there are strong indications thatthe biopsy cells probably express the same membraneantigen as the carrier cultures.37The antigenic components entering the EBV-

determined membrane and the intracellular nucleo-

capsid complex, respectively, differ with regard toimmunological specificity. By absorbing sera thatreacted with the membrane and the intracellularEBV complex as well, with large numbers of intact,viable membrane antigen-positive cells it was pos-sible to remove the membrane-reactive antibodies,with only a minor reduction in the anti-EBV titer.38Moreover, some sera could be found with antibodiesagainst the membrane antigen, or the EBV antigen,but not both. Although such 'discordant' sera werein minority, their existence is in line with theimmunological distinctness of the two antigen types.

Further analysis of the two antigen systems re-vealed95 that both the membrane and the intra-cellular antigens must be regarded as antigen com-plexes, with several distinct subcomponents. Serathat contain antibodies against several subcom-ponents of the intracellular EBV complex also tendto carry, as a rule, several antibody componentsagainst various parts of the membrane-antigen com-plex, but the relationship is not absolute and manycombinations can be found. Patients with large,persisting tumors frequently had a larger number ofserum antibody components against both antigencomplexes than sera from healthy, EBV-positiveindividuals, or convalescent sera from donors afterinfectious mononucleosis, or sera from BL patientswhose tumors have gone to long-term regressionfollowing chemotherapy.The nature of the membrane antigen, particularly

its specification by the viral or the cellular genomeremains to be clarified. Recently, indirect evidencehas accumulated suggesting that it may represent aviral envelope component. The ability of differentsera to neutralize an artificial EBV infection of EBV-negative culture lines (such as Raji or 6410) wasrelated to the titer of membrane-reactive antibody,and not to the anti-EBV titer.39 This was particularlyapparent when a series of sera were tested that werediscordant with regard to their anti-EBV and mem-brane reactivity. In another series of tests, the sera ofrabbits immunized with EBV concentrates were ableto block the membrane-antigen reaction specifically.40This indicated that the membrane-antigen was pre-sent in the immunizing material, either as a con-stituent or as a contaminant of the viral particles.

It has been shown that herpes simplex virus (HSV)is capable of including new membrane-antigens inthe cells it infects.4' Viral mutants with differentenvelope characteristics induce different membranechanges, in a way that closely parallels their envelopeproperties.4' It has been concluded that the appear-ance of viral envelope material is responsible for thechanges in the cell membrane. Presumably, the viruschanges the cellular membrane in order to facilitatethe process of its own envelopment. In view of theparallel between EBV neutralization and membrane-

143P



http://pmj.bm

j.com/

Postgrad M


arch 1971. Dow

nloaded from

http://pmj.bmj.com/

George Klein

reactive antibody levels, it is conceivable that theEBV-associated membrane-antigen represents viralenvelope material as well. This is of interest, because,for HSV, a relationship has been demonstrated be-tween the changed 'social behavior' of infectedcells and their membrane modifications after expo-sure to different mutants of HSV.42 The under-standing of the role EBV-induced membrane changesmay play for cell behavior may elucidate the re-lationship between this agent and the neoplasticdiseases with which it is most regularly associated.

Disease-related serological patternsThese can be discussed at two levels: (a) the

relationship between EBV-associated serological re-activity and the clinical and pathological diagnosis;(b) changes in EBV-related serological patterns dur-ing the clinical course of EBV-associated diseases.Concerning (a), it can be first stated that the sero-

logical anti-EBV reactivity, as determined by theHenle test25 is extremely widespread in all humanpopulations. If the level of significant reactivity isset at a 1: 10 serum dilution, as customary,43 thelarge majority of adult populations is EBV-positive.It may be questioned whether the 10-15Y. negatives(with titers <10) are real or spurious. Specific anti-bodies may occur at titers below 10 and may bemissed, due to the various test artefacts that arise athigh serum concentrations 43. On the other hand,whereas some of the <10 'negatives' may hidespecific reactivity, at least part of them must be realnegatives in the biological sense. A prospectivestudy" has shown that EBV-positive young adultsare protected from infectious mononucleosis,whereas a significant proportion of the 'EBV-negatives' (i.e. < 10) developed the disease andbecame EBV-positive, in the course of a 2-4 yearsobservation period.The causal relationship between EBV and at least

one form of infectious mononucleosis45 is mostclearly established by this prospective study. If thisis accepted, it immediately leads to the questionwhether EBV plays any role in other diseases andparticularly the neoplastic diseases with which it ismost regularly associated.The serological patterns that are now known can

be evaluated in different ways. In the Henle typeanti-EBV test, Burkitt lymphoma (BL) and naso-pharyngeal carcinoma (NPC) are distinguished byoutstandingly high anti-EBV titers, so far un-paralleled amongst other lymphoproliferativediseases and other carcinomas of the head and neckregion.43' 46-48 The geometric mean anti-EBV titer ofBL patients was eight times higher than in variouscontrol groups. There were no significant differencesbetween control sera collected from areas with ahigh or a low incidence of Burkitt's lymphoma.

With the exception of a few moribund cases, low( <1 : 80) anti-EBV titers were very rare among BLpatients and there are no histologically confirmedcases with negative (<1 :10) titers. Occasionally,long-term regression cases tended to show fallingtiters after some years, but this was by no means therule. The serological behavior of BL and NPC isalso exceptional in the precipitin test developed byOld, Oettgen et al.,49' 50 performed against a solubleantigen extracted from the EBV-carrying P3J line.NPC sera from African and American patients werepositive in 85-87%Y and 59yO of the African BL seragave positive precipitin reactions. Other neoplasms,including lymphoproliferative diseases and carci-nomas of different kinds gave a much lower incidenceof positives, with the exception of chronic lymphaticleukemia and lymphosarcoma that came close to thereactivity of the BL sera. Two distinct precipitin lines(B and P) could be identified regularly but there wasno obvious disease-related difference between thetwo.The antibodies against the EBV-associated mem-

brane-antigens can be most easily evaluated by theblocking of the direct membrane fluorescence re-action, obtained with specific reference con-jugates.'1' 52 When the Mutua conjugate was used,already referred to in the previous paragraph,Burkitt's lymphoma55 and nasopharyngeal carci-noma sera47 showed a high blocking activity, where-as the sera of normal African controls, Burkittpatients' relatives and African tonsillitis patientsshowed mostly negative reactions, even thoughoccasional positives were encountered. Head andneck tumors, other than Burkitt's lymphoma andnasopharyngeal carcinoma were also largely negativebut occasional highly positive sera have been en-countered in this material as well. The differencebetween the regularly high-reactive African orChinese nasopharyngeal carcinomas and the pre-dominantly low-reactive Indian hypopharyngeal andoropharyngeal carcinomas was particularly re-markable.47

In a 'tripartite' study, the anti-EBV (Henle) test,the blocking of direct membrane fluorescence andthe precipitin reaction have been compared with 151coded sera."" There was a clear overrepresentationof BL and NPC sera within the 'triple-high' re-activity group and they were virtually absent fromthe 'triple-low' group. The opposite was true for thetwo main groups of control sera, derived fromhealthy relatives of BL patients and from donorswith head and neck tumors other than BL and NPC.One interesting question concerns the relationship

between geographical localization and serologicalreactivity. Nasopharyngeal carcinomas are moreeasy to evaluate in this respect, since they representa clear pathological entity and are not readily

144P



http://pmj.bm

j.com/

Postgrad M


arch 1971. Dow

nloaded from

http://pmj.bmj.com/


confused with other conditions. The EBV-associatedserological reactivity of African, Swedish, French,Chinese and American cases was uniformly high andappeared to be characteristic for the anaplastic orpoorly differentiated type.46' 47, 50

The evaluation of Burkitt's lymphoma outsideAfrica presents a more difficult problem, because thepathological picture alone does not permit a sharpdistinction against other lymphomas. The combinedclinical and pathological picture has readily recog-nizable features in the high-endemic areas, but theyare less characteristic in other regions and theclassification becomes more arbitrary. If one never-theless examines the data on the serological behaviorof 'Burkitt-like' lymphomas outside Africa, itappears that the results are partly in line with theAfrican Burkitt lymphomas,58' 54 and partly differfrom the African cases, i.e. have no distinctivelyhigh EBV-associated reactivity (defined as highanti-EBV titer and/or high membrane-blockingindex) and thus resemble ordinary lymphosarcomasrather than 'true' Burkitts. This picture cannot beinterpreted meaningfully at present since serologycannot serve as the basis for classification if theproblem is to decide whether non-African cases havean African-Burkitt-like serology or not; the argu-ment becomes circular. Speculatively at least, onemay nevertheless consider the possibility that thenon-African Burkitt-like cases are heterogeneous.Some of them would be 'true Burkitts', i.e. have thesame etiology as the African cases, whereas otherswould be different and comparable to 'ordinary'lymphosarcomas. Whether this classification can bebased on the EBV-associated serological patternswill, of course, depend on the question whether therelationship of the EB virus to Burkitt's disease is ofan essential or of an accidental nature.

(b) Another approach to the study of disease-related EBV patterns is to follow the antibodytiters against the various EBV-associated antigenshorizontally, during the course of 'EBV-associated'diseases, such as BL and NPC. For comparison, onemay choose EBV-positive individuals with more orless related neoplastic diseases that are not regu-larly associated with high anti-EBV titers. Studiesof this type are now becoming feasible; some pre-liminary information is already available. In BL,the indirect membrane test, performed with biopsycell targets has indicated at an early stage101' 102 thatthe most highly reactive sera can be found in patientswhose tumors have gone to long-term regression.Later, when the more specific and sensitive blockingof direct membrane fluoresence replaced the indirecttest as the main method to detect antibodies againstthe EBV-associated membrane-antigen complex onestablished culture lines, it turned out5,532' ^5 thatnearly all histologically confirmed African BL sera

have a high blocking activity, i.e. show a completeor nearly complete cross-reactivity with the referenceconjugate. The few exceptions have come from mori-bund patients. This monotonously uniform blockingactivity, obtained with the undiluted sera hides largequantitative differences, however. When comparedby serial titration against the same reference con-jugate, the blocking titer of various BL patients' sera(taking a blocking index of 0'5 as the endpoint) canvary between 1: 1 and 1: 600.55 In the individualpatient, the titers may change considerably in thecourse of the disease, but, as a rule, they remainwithin the same order of magnitude: most changesare restricted to relatively few dilution steps up ordown and the patients can be therefore classifiedinto groups of low, medium and high reactivity.Our preliminary findings indicate that the block-

ing titer differences between patients, as well as thehorizontal changes, are influenced by a number offactors. In the course of rapid and extensive tumorgrowth, antibody levels probably fall, due to adsorp-tion to tumor cell membranes. When the patientreceives chemotherapy and the tumor regresses,there is often an increase in titer. At first sight, thismay seem paradoxical, in view of the immuno-suppressive effect of chemotherapy. It is known,however, that chemical immunosuppression inhibitsnew primary antibody responses against antigensadministered after the drug, but is much less efficientagainst immune reactions established before treat-ment.An increase in blocking antibody titers was par-

ticularly apparent in BL and NPC patients whoreceived local radiotherapy,' including cases wheretherapy did not lead to complete tumor regression.In view of the fact that X-irradiated tumor cells arerelativelygoodiimmunogens in experimental systems57this is of considerable interest. It may also berelevant that in EBV-carrier cultures with a relativelylow membrane-antigen reactivity, X-rays can inducethe appearance of the membrane-antigen on a largefraction of the cells."8

In BL patients with recurring tumors that con-tinue to grow in spite of chemotherapy, blockingantibody levels that have fallen to low levels at oraround the time of recurrence can rise again.16Subsequently, the lymphoma cells become coatedwith IgG, as a rule, if the tumor persists. It is con-ceivable that such cells represent immunoresistantvariants, similar to what has been found in experi-mental systems.59 This is supported, indirectly atleast, by the history of two patients whose tumorsrecurred after several years of total regression andcontained a high frequency of near-tetraploid cells,22in contrast to more than twenty other BL biopsiesexamined,0 with a shorter clinical history, that wereall in the near-diploid range. Tetraploid cells can

145P



http://pmj.bm

j.com/

Postgrad M


arch 1971. Dow

nloaded from

http://pmj.bmj.com/

George Klein

frequently outgrow host responses that efficientlyreject diploid cells of the same lineage.1The immunoglobulin coat acquired by the tumors

that persist in spite of therapy may be the equivalentof enhancing antibody or of blocking antibody inHellstrbm's sense.7 This is not necessarily an alterna-tive to the possibility that membrane-reactive anti-bodies may have a growth-inhibitory action butrather another facet of the same complex picture.An antibody that has cytotoxic or growth-inhibitoryproperties against immunosensitive cells may exertan enhancing effect (i.e. protect the target cell againstthe cell-mediated immune response) when it inter-acts with an immunoresistant cell without killing it.In addition, different antibodies no doubt differ;some can be cytotoxic and others enhancing towardsthe same target cell. In the course of chemotherapythat falls short of a total tumor kill, and the sub-sequent regrowth of the residual tumor with moreantigen release and antibody binding, the immuno-sensitivity of the tumor cell population and thekilling vs. enhancing power of the antibody popula-tion must obviously change in a complex way. Thiswould require a multicomponent experimentalanalysis, that is not yet within reach.

In addition to the changes in membrane-reactiveantibody levels brought about by the tumor itself(i.e. changes due to absorption, antigen release,effect oftumor growth on the immune response, etc.),the antibody titer may change for other, tumor-unrelated reasons, and this may, in turn, influencetumor growth. This possibility has been brought intofocus by the history of a BL patient16 who was intotal tumor regression for a period of 41 years andsubsequently developed widespread abdominal meta-stases. Her membrane-reactive antibody level, deter-mined by the blocking test, fell markedly more than6 months prior to recurrence, at a time when therewas no reason to suspect the presence of any meta-stases. When the abdominal recurrence becamemanifest 6 months later, the membrane-reactiveantibody level was still low, and the tumor cells werenot yet coated with IgG. In the course of the sub-sequent 2 months, the serum antibody level increasedagain and the lymphoma cells became IgG-coated.This secondary sequence of events decreases theprobability that the fall of the antibody level thatpreceded recurrence by 6 months was due to absorp-tion to an as yet cryptic tumor, because, in that case,a period of slow tumor growth would have followedduring the subsequent 6 months period, and thesecondary increase in antibody level, as well as thecoating of the lymphoma cells with immunoglobulinswould be expected to have occurred in the interim,already appearing at the time of clinical recurrence.Indirect as this reasoning is, it has neverthelessraised the question whether a fall in antibody levels

may be sometimes the cause, rather than merelythe consequence, oftumor recurrence, and whether itcould act by facilitating the outgrowth of 'dormant'neoplastic cells.There is some preliminary evidence indicating

that the antibodies against the soluble EBV-associated antigens detected by immunoprecipitationshow a different disease-related pattern, appearingat the time of progressive tumor growth andfrequently absent in patients whose tumors are incomplete regression.16 53 Although there arenumerous exceptions to this, a relationship of thistype appeared clearly when the horizontal history ofthe patient already mentioned above, was followedduring long-term regression and subsequent re-currence.16 This may also explain why high anti-EBV titered NPC sera are more frequently precipitin-positive than BL sera with comparably high titers:in NPC, the serum material is mainly derived frompatients with residual or progressively growingtumors, while a collection of BL sera include pro-gressor and regressor sera as well.

Recently, Henle et al. have described yet anotherEBV-associated antigen, designated as EA (earlyantigen),96 detected in EBV-infected Raji or 6410-cells. The sera of some EBV-positive donors, but notof others, contained antibodies against it. EA pre-cedes the classical Henle-type 'EBV-antigen' duringthe infectious cycle; the latter has been renamed to'VCA' (viral capsid antigen). The appearance of EAis readily prevented by puromycin but not by DNAinhibitors.34 The behaviour of EA is thus not unlikewhat has been mentioned above for MA, with oneimportant exception: whereas MA is compatiblewith continued cell growth and DNA synthesis, EAinhibits host macromolecular synthesis, as shownby a combination of immunofluorescence and auto-radiography63 and thus probably signals the entryof the cell into the lytic cycle. In contrast to EA,VCA is dependent on DNA synthesis and thereforeprobably represents a 'late' viral product.A study ofanti-EA titers in BL patients withknown

anti-VCA titers showed that tumor-bearing patientshave anti-EA antibodies more frequently and inhigher titers than patients whose tumors have goneto total regression. Also, whereas both BL, NPC andacute infectious mononucleosis (IM) patients haveanti-EA antibodies, anti-VCA-positive healthydonors only exceptionally show any anti-EAreactivity at all and, if they do, they have muchlower titers than tumor-bearing patients. In BL, arelationship could be demonstrated between thepresence of anti-EA antibodies and the probabilityof recurrence. In untreated patients and in patientstested at the time of regression, high anti-EA titersshowed a certain correlation with a poor prognosis.64Although there is thus a superficial similarity

146P



http://pmj.bm

j.com/

Postgrad M


arch 1971. Dow

nloaded from

http://pmj.bmj.com/


between the disease-related patterns shown by theanti-EA antibodies and the precipitating antibodylevels against the soluble antigen of Old et al.,mentioned above, it is also clear that the two anti-gens are different. Although the reactivity againstthem is correlated in the majority of the sera com-pared, a sufficient number of 'discordant sera' existto prove that the antigenic specificities are not iden-tical.64 Both represent intracellular, and probablynon-structural, early proteins, however.The tumor-related presence of precipitating and

of anti-EA antibodies is reminiscent of some DNA-virus-induced experimental tumor systems, andparticularly the case of antibodies against 'T-antigens'. In polyoma, SV40 and adenovirus-inducedtumors, it is the 'tumored hamster', i.e. the host ofnon-virus producing, T-antigen positive tumors thattends to develop antibodies against T-antigens. Theantibody levels usually fall when the tumor is re-moved or rejected and eventually disappear. T-anti-gens are intracellular, like the soluble antigens in thepresent system. They are also 'early' components ofthe viral cycle, independent ofDNA synthesis35 andthey are regularly present in transformed cells. Inthe former respect (independence ofDNA synthesis)they are similar to, in the latter respect (compati-bility with cell multiplication) they are different fromthe EA antigens. As far as the T-antigens are con-cerned, it is not known whether they are releasedfrom growing tumor cells by some kind of secretoryprocess, or are only liberated from dead and dyingcells. A release of the EA antigen may well be re-lated to an abortive virus cycle, perhaps induced intumor cells on their way to necrosis, i.e. under cir-cumstances where the virus cycle cannot proceed tocompletion. The frequent absence of anti-T and anti-EA antibodies after tumor removal is in sharpcontrast to the membrane-reactive antibodies, whichtend to remain high, or even increase following re-jection in experimental systems and as far as this hasbeen studied, probably in the BL system as well.This may be related to the continued presence of'dormant' tumor cells with a preserved, virally deter-mined, foreign membrane antigen, held in check,but not killed by the host immune mechanism.

Further clarification of the relationship betweenthe dynamics of antibody formation against differentEBV-determined antigens and the clinical course ofthe 'high-EBV-associated diseases', such as BL andNPC, may be helpful in elucidating important virus-tumor-host relationships, particularly if comparedwith EBV-positive sera from patients with othertumors that are not characterized by a regularly highEBV-association.

Nucleic acid hybridization studiesRecent work of zur Hausen et al.,32 65 demon-

strating the presence of cellular DNA capable ofspecific hybridization with purified EB-virus DNA isof considerable interest in the present context andmay help to understand the serological results. Bythis technique, it was shown5 that the BL-derivedRaji line does contain EBV-DNA, in spite of thefact that none of the three EBV-associated systemsMA, EA and VCA, can be demonstrated by immuno-fluorescence.65 Since VCA and EA are not demon-strable in BL biopsy specimens30 either, as a rule,although MA antigens are,24 a recent study con-cerning the presence of EBV-DNA-hybridizablecellular DNA in biopsy specimens is of interest.Thirteen biopsies from ten BL donors all showedthe presence of EBV-DNA, with 2-26 genome equi-valents per cell.32 In three donors who yielded doublebiopsies, there was a remarkable agreement betweenthe two tumors examined, although they were lo-cated in different anatomical sites (2-2, 7-8 and 21-26 approximate EBV genome equivalents per cell).In ten nasopharyngeal carcinoma biopsies, EBV-DNA was present in all, with 1-19 genome equiva-lents per cell. In ten tumors of other, miscellaneouskinds, taken from EBV-seropositive donors, therewas no demonstrable EBV-DNA. The same wastrue for a number ofother controls, including variousestablished cell lines of human origin, cytomegalo-virus-infected lines, Marek's tumor in chickens etc.These studies thus show, in agreement with the

serology, that there is a more intimate EBV-tumorassociation in BL and NPC than in other tumorsoccurring in anti-EBV seropositive donors.

Implications and dilemmasFour main dilemmas arise from this pattern of

findings; they are interrelated but all have theirspecific aspects. They can be briefly stated as follows:

(a) The etiological dilemma. Can the occurrence ofdistinctive, tumor-associated antigens give any cluesabout the etiology of the disease?

(b) The problem of neoplastic behavior. Are thechanges in the composition of the cell membrane, orother cellular organelles, as reflected by the appear-ance of new antigenic specificities, fundamentallyinvolved in the neoplastic behavior of the cell, or,in other words, does the unresponsiveness of the cellto growth control depend on the change in com-position or structure that is revealed by the im-munological tests?

(c) The therapeutic problem. Can any of the im-munological reactions now identified serve to mea-sure the patient's reactivity to his own tumor, inconnection with various therapeutic procedures, in-cluding attempts at immunotherapy?

(d) Are there any prevention approaches in sight?Concerning the etiological dilemma, it is a useful

point of departure that all virally induced experi-

147P



http://pmj.bm

j.com/

Postgrad M


arch 1971. Dow

nloaded from

http://pmj.bmj.com/

George Klein

mental tumors share the same antigen, as long as theyare induced by the same virus, at least as far as thetransplantation-type, membrane-associated antigensare concerned. The reverse, the assumption of acommon viral etiology on the basis of commonantigens found in tumors of unknown origin is notnecessarily justified, however. It has been shown""-70that virally induced new antigens can be made toappear by superinfecting normal cells or tumors ofunrelated etiology with oncogenic and even with somenon-oncogenic viruses. The only difference betweenthis secondary 'antigenic conversion' and theprimary event that occurs in direct relation to tumorinduction, is the lesser stability of the former,particularly in immune hosts,4 as well as a moreirregular association between antigen and tumor, de-pending on the accidental nature of superinfection.As discussed above, high anti-EBV titers and high

antibody levels against EBV-associated membraneantigens and soluble antigens are regularly asso-ciated with at least two neoplastic diseases: Burkitt'slymphoma and nasopharyngeal carcinoma. Fornasopharyngeal carcinoma, it is clear that this sero-logical pattern is independent of geographical orethnic origin. A similar situation may exist forBurkitt's lymphoma, but the lack of reliable criteriaby which the identity of the disease can be establishedoutside the endemic areas and distinguished fromordinary lymphomas makes a similar evaluation ofthe non-African cases more difficult.

It is important to stress that the main differencebetween BL and NPC and other normal or neoplasticserum donor categories investigated is not EBV-positivity, nor the occurrence of high-titered re-actions in occasional donors, since such donors maybe found in most other categories as well, but theregular and consistent association of high-titeredreactions according to all three tests. Looking at itfrom this angle, BL and NPC are unique. One mayquestion, however, whether this angle can be justi-fied or, more specifically, what it implies.As a starting point, we may take the convincing

demonstration that EBV is causally related to at leastone form of infectious mononucleosis.44 45 Thisform afflicts EBV-seronegative adolescents, as arule, is frequently positive for heterophile antibodiesand is regularly accompanied by seroconversion toanti-EBV positivity. As indicated by a prospectivestudy,44 anti-EBV individuals are apparently pro-tected from the disease.The serological screening ofmany different human

populations also showed33' 43 that there is another,' early' seroconversion to anti-EBV positivity, cul-minating around 4 years of age, and particularlyfrequent in low socioeconomic groups. This earlyinfection does not lead to infectious mononucleosisor any other disease entity so far recognized.

Viewed against this background, the relationshipofEBV infections to BL and NPC may be consideredin terms of the following alternatives:

(a) The virus that causes infectious mononucleosisis also responsible for these two tumors; if this istrue, intrinsic or extrinsic co-factors have to bepostulated to explain the malignant conversion (the'co-factor hypothesis').

(b) Different virus subtypes are responsible forthe different clinical entities (the 'multiple virushypothesis').

(c) The virus is a relatively harmless inhabitant oflymphoid tissues, although it may cause temporaryproliferation (mononucleosis) under certain con-ditions. When lymphoid tissues proliferate for otherreasons, e.g. in malignancies due to other, unrelatedcauses, the virus travels along as a passenger, withincreased antigen production and high-titered anti-body formation as a result. This 'passenger hypo-thesis' is the logical analogue of the 'antigenic con-version' of established tumors by etiologically un-related viruses, discussed above. In view of the highregularity of association, a requirement for a par-ticular trophic relationship between EBV and thetarget (lymphoid) tissue may be added in thepresent case.The passenger hypothesis cannot be excluded at

present, but it appears less likely in view of the factthat lymphoproliferative diseases other than BL andanaplastic carcinomas other than NPC do not showa regular high-titered EBV-association. This state-ment includes malignances occurring in the same orclosely adjacent anatomical areas, such as reticulumcell sarcoma, lymphosarcoma, etc., and carcinomasthat arise in or close to the tissues of the Waldeyerring, such as the hypopharynx, oropharynx, thetonsil, base of the tongue, soft palate, etc. Carcinomaof the maxilla is a possible exception, but largergroups remain to be investigated. Hodgkin's diseaserepresents a very interesting case in itself. Recentlyit has been found that the sarcomatous form, i.e.the lymphocyte-poor type with the worst prognosis,shows a high anti-VCA and anti-MA reactivity,quite comparable with BL and NPC, whereas thelymphocyte-rich and relatively more benign para-granulomatous form is low-reactive in both tests andthus resembles the control material.48 The granulo-matous form was intermediate, both with regard tohistological type and serological reactivity. Thismeans, as far as the serology is concerned, that itrepresents a mixture of high and low-reactive cases.Whereas it is thus possible that EBV plays somespecial role in the etiology of Hodgkin's sarcoma,the inverse correlation with lymphocytic predomi-nance would not be in line with a simple passengerhypothesis.None of this reasoning excludes the passenger

148P



http://pmj.bm

j.com/

Postgrad M


arch 1971. Dow

nloaded from

http://pmj.bmj.com/


hypothesis conclusively, of course, but quite anumber of ad hoc assumptions have to be made tomaintain it in face of this evidence. One would haveto postulate some specific trophic relationship be-tween the virus and the kind of lymphocyte thatgives rise to BL and is particularly abundant inNPC, that would not apply to the lymphoid cellsthat proliferate in the various other malignancies,used as controls. No valid objection can be raisedagainst such a hypothesis, but it appears rather far-fetched in view of the fact that EBV-carrying blastoidcell lines can be regularly isolated from EBV-positive individuals, including donors with lympho-reticular malignancies of the 'control' type, i.e.diseases that do not show a consistently high EBV-positive serology. The sarcomatous form of Hodg-kin's disease is also very hard to explain.The possibility that EBV acts together with some

co-factor in causing neoplastic disease or, to phrasethe same thesis differently, that it acts by increasingthe likelihood of neoplastic transformation broughtabout by other factors, has been recently proposedby Burkitt71 as far as the etiology of BL is con-cerned. In order to fit the geographic distribution ofthe disease with an ubiquitous virus, Burkitt pro-posed that an insect-transmitted co-factor is respon-sible for the malignant manifestation and specifiedit as chronic holoendemic malaria. This was based onthe absence of BL from certain areas where malariacontrol has been enforced for some time and itspresence in adjacent regions where malaria controlwas not regularly practiced.

It may be agreed that interactions between virusesand other agents, capable ofstimulating the prolifera-tion of a target tissue may lead to malignant trans-formation in experimental systems where neitherthe virus nor the other agent is oncogenic per se.72Since chronic malaria exerts a strong proliferativestimulus on the RES, Burkitt's modified theory isreasonable, although it may be objected that thesame picture would result from the transmission ofany etiological factor or co-factor mediated by theappropriate insect, and this includes other viruses.Recently, some preliminary evidence has been ob-tained concerning the frequency of the sickling traitin BL patients,73 however, that indicates a possiblerole of malaria in the causation of the disease.The third possibility, the multiple virus hypo-

thesis, implies the existence of closely related butbiologically different EBV-viruses with differencesin their oncogenic power and their target tissuepreference. In light of the information derived fromexperimental oncogenic viruses, this is a realisticalternative as well. As far as leukemia viruses of theRNA type are concerned, it will be recalled thatprior to the discovery of the interference test foravian leukosis virus classification,74 it was not

possible to distinguish by morphological or im-munological means between the viruses that wereresponsible for the different lympho- and myelo-proliferative diseases or for fowl sarcoma. It is nowknown that the avian leukosis-sarcoma virus grouphas many closely related members; some inducesolid tumors with highly distinctive properties,others are responsible for myeloid or erythro-myeloid leukemia, or lymphomatosis, and stillothers cause no recognizable disease at all. Aclosely similar development can be noticed in themurine leukosis-sarcoma field. The Friend, Moloney,Rauscher, Gross, Kaplan, Rich, Graffi, Mazurenko,etc. agents are similar antigenically and indistin-guishable by ultrastructure, but they induce distinctand characteristic clinical and pathological diseaseentities, specific for the viral agent.75 In the DNAfield, a possibly relevant example is the series ofherpes simplex mutants, studied by Roizmann andhis colleagues.42 Although this is not known to be anoncogenic system, it is important that different viralmutants induce different membrane changes in in-fected cells and, concomitantly, the cells are alteredin their 'social behaviour' in ways that are charac-teristic for the virus mutant. Although a lytic virusobviously cannot transform its targets, the cellularchanges are nevertheless concerned with intercellularrelationships. Conceivably, other, non-lytic virusesof the same family might induce membrane changescompatible with cellular viability and reproductiveintegrity, and a social behaviour changed in thedirection of disobedience to growth regulation-or,in other words, neoplasia. It may be recalled in thisconnection that the agents of at least two neoplasticdiseases, Marek's neurolymphomatosis in thechicken76 and Luck6's carcinoma in the frog77 wererecently identified as herpes-type viruses. A simianlymphoma is also probably due to a herpes-typevirus (Herpesvirus saimiri).78The immunological tests so far performed on

EBV-associated antigens, including those referredto in the previous sections, are not necessarily com-petent to reveal finer differences between closelyrelated but biologically different agents with cross-reactive or overlapping antigenic components. Apreliminary study of the membrane-antigens carriedon EBV-positive lymphoblastoid cell lines derivedfrom BL and NPC did not show any difference inthe reactivity patterns,79 but this may simply reflectthe insufficient discriminating ability of the test.

Further studies are needed to distinguish betweenthese possibilities. In order to narrow down thepassenger hypothesis, more extensive tests are de-sirable on tumor categories where occasional seragave high EBV-associated reactivity but only limitedsamples have been tested. It is also desirable to con-duct nucleic acid hybridization studies on the corre-

149P



http://pmj.bm

j.com/

Postgrad M


arch 1971. Dow

nloaded from

http://pmj.bmj.com/

George Klein

sponding tumors. More refined analytical methodsare needed for the attempts to dissect different virusvariants. The recent developments in the herpessimplex field suggest that biochemical studies onviral envelopes and altered cell membranes may beparticularly rewarding.As far as the sero-epidemiology of EBV-infections

is concerned, it has to be emphasized that there wereno significant differences between low and high BL-endemic areas with regard to the distribution of anti-EBV titers in relation to age;43 a comparison oftiters in normal populations is therefore unlikely toelucidate the possible role of this virus in Burkitt'slymphoma. A prospective sero-epidemiological studymay be more rewarding. It may be relevant in thisconnection that BL is essentially a childhood disease,with a peak incidence between 4 and 7 years. Thisfact, together with the clinical and serologicalevidence indicating a relatively high antigenicity inthe autochthonous host would speak for a shortlatency period during the oncogenic process. Inexperimental systems, highly antigenic tumors arisewith short latency periods, as a rule; or to put it inother words, highly antigenic tumors cannot escaperejection unless they grow out rapidly after theirinception."1 3, 80

If this reasoning is essentially correct and thelatency period of BL is relatively short, a prospectivesero-epidemiological study may be decisive. In thepopulations of risk within the high endemic areas,as in other populations, there is a relatively smallminority of EBV-negative children and another, evensmaller minority with high anti-EBV titers. Themajority consists of low-titered positives.43 Thequestion is whether BL develops in one of the twominority groups or at random and irrespectively ofanti-EBV titer. Provided that a sufficiently largenumber of sera could be collected and stored underappropriate identification, tests may become feasibleon pre-disease sera from individuals who developBurkitt's lymphoma within a few years' time. Itmight be objected that the sensitivity threshold ofthe anti-EBV test may be too high (1 : 10) and afraction of anti-EBV positive sera may be classifiedas false negatives. Since more concentrated seracannot be tested safely due to the non-specific arte-facts that tend to appear, this is probably true. It isalso clear, however, that at least a substantial partof the 'anti-EBV negative' donors, as defined by the1: 10 threshold, must be negative in the biologicalsense, since the prospective study on infectiousmononucleosis has clearly shown44 that the diseasedevelops only in this group, and not in persons classi-fied as anti-EBVpositiveaccording to the samecriteria.

Concerning the relationship between EBV andnasopharyngeal carcinoma, the same types of hypo-theses can be discussed as for BL. The multiple virus

hypothesis would imply an NPC-specific EBV-variant. The co-factor hypothesis would lead to aconsideration of both genetic and environmentalfactors, in view of the information on the incidenceof the disease in migrant high-risk populations.8' Aprospective serological study of this question wouldbe very difficult at the present time, since NPC,unlike BL, occurs over an extremely wide age-range.The possible significance, for the understanding of

neoplastic cell behaviour, of the cell membranechanges reflected by the appearance of new antigenscannot be assessed at present, but it may be pertinentto point out that cell membrane changes are amongthe most significant parameters of neoplastic be-haviour at present. They are almost invariably foundwhen comparable normal and transformed cells arestudied in parallel. They may concern changes inbehaviour, such as contact inhibition,82 or alteredexpression of phytoagglutinin receptors83 84 thatmay reflect a change in the synthesis of certainglycolipids,85 and are perhaps linked to the appear-ance of new surface antigens.88 Membrane-antigenchanges have been demonstrated in all experimentaltumors that have been thoroughly studied'-7 andalthough the details concerning antigenic strengthand patterns of cross-reactivity vary from system tosystem, they must reflect some remodelling of themembrane structure. Growth-regulating mechanisms,including both the forces that act via long-range,humoral arms and the short-range, contactual signalsas well, must transmit their message to the target cellvia receptors on the outer membrane. Nonlytic virus-cell interactions may result in the incorporation ofvirally determined (or virally derepressed) com-ponents into the membrane that render the appro-priate receptors insensitive to- regulation and, if thisis compatible with continued cell growth anddivision, it may trigger neoplastic development.Since infection with potentially oncogenic virusesand the concomitant surface antigenic changes arenot limited to the oncogenic target tissue but canoccur in other cells as well that remain normal (i.e.subject to regulation), a tissue or cell type-specificitymust be added to explain transformation. Sincedifferent tissues must obey different types of growthregulation, this is not surprising. Also, virally deter-mined antigens may be retained while in vivo tumori-genic properties decrease or are lost from cell hybridlines 87-89 or from the 'revertant' forms that may arisefrom transformed cultures in vitro.80-92 Furtherstudies on such systems will be most interesting, notonly for the understanding of neoplastic behaviourand the possible role of membrane changes in it, butalso for the understanding of normal growth-responsiveness at the cell level.

Meanwhile, the question whether EBV-associatedmembrane-antigens are essential for the neoplastic

150P



http://pmj.bm

j.com/

Postgrad M


arch 1971. Dow

nloaded from

http://pmj.bmj.com/


behavior of BL and NPC cells is not clear. As faras NPC is concerned, such membrane-antigens havebeen demonstrated on derived lymphoblastoid celllines79 but it is not known whether they are presenton the surface of the carcinoma cells. Establishedculture lines of BL cells carry EBV, as a rule, al-though at very different levels.29' 30 The membrane-antigen can only be demonstrated in lines with arelatively high 'EBV-load'29 and is subject to en-vironmental fluctuations.93 There is at least one BLline (Raji) which contains no EBV antigen demon-strable by immunofluorescence or virus particles94although it carries DNA that hybridizes specificallywith EBV-DNA, as already mentioned.66 Since theRaji line can be superinfected with EB virus,'6 theabsence of virus production is presumably not dueto repressors. If it carries genetic information de-rived from EBV, it is probably a defective viralgenome, lacking the cistrons that specify the mem-brane, capsid and early protein antigens. If therewould be any assurance that the Raji line representsa neoplastic cell, this would imply that the membraneantigen is not required for neoplastic behavior.Since this question cannot be tested directly with ahuman cell, however, a conclusive answer is notavailable Further studies on the presence of viralDNA and virus-specific mRNA in BL-derived lines,in comparison with EBV-carrying blastoid cell linesof other origin may prove very informative. In thisconnection, it is interesting that IM-derived lines arereportedly more prone to lose their EBV than BL-derived lines.33 Thus, whereas EBV is clearly helpfulin inducing lymphoblastoid transformation andfacilitates the establishment of stationary suspensioncultures,33' 97-9 there is no doubt that blastoid celllines can exist without the expression of a productiveEBV infection.Turning now to the therapeutic problems, it seems

clearly established that the host immune-responseplays an important part in Burkitt's lymphoma.This is indicated by the documented occurrence ofspontaneous regression,2 by the substantial fractionof long-term survivors, sometimes after only mildchemotherapy,9 11' 100 by the reactivity of the autoch-thonous host against its own tumor cells, indicatedby the presence of humoral antibodies reacting withthe surface of viable cells,101' 102 by the positiveCl-a fixation test,103 and by the transformation ofhost lymphocytes when confronted with mitomycin-treated autochthonous lymphoma cells in the mixedlymphocyte-target cell interaction test. 104 In addition,the processive accumulation of an IgG coating onthe cell surface of tumors that persist in spite oftherapy36 16 together with the tetraploid (immuno-resistant?) constitution of tumors that have recurredafter long-term regression22 suggests that the dyna-mics of immunoselection may also apply to this

human system as they do for experimental tumors.6'Immunoresistance may be as important as drugresistance, if not more so, in frustrating therapy.The host response to an autochthonous tumor is

no less complex than other immune responsesagainst viable cells. Different effector componentsinteract in ways that rejection, or its opposite,enhancement, will dominate the eventual outcome.Humoral antibodies are cytotoxic in some situationswhereas in others they lack demonstrable growthinhibitory effects but nevertheless manage to attachand thereby prevent the access of host lymphoidcells.3' 105 Recent evidence indicates that such'blocking antibody' may play an important role incounteracting the cell-mediated host responsein experimental0 and human107 tumors as well.The main therapeutic dilemma is what the proper

stimuli are, specific or non-specific, and how theyare best administered to the immune system in orderto achieve the objective, rejection, and avoid itsopposite, enhancement. The rationale of introducingimmune stimuli at a time when the tumor load con-fronting the host is minimal, i.e. after regression hasbeen induced by chemotherapy, is obvious.1081 09but the optimal form of stimulus and the best modeand timing of its administration is not. No a prioriguidance can be given from experimental studies,because the same mode of administration, dosage,vehicle, etc, of the same preparation may favourrejection in one system and enhancement in another,and the differences depending on host species, tumortype and individual characteristics of the tumor lineare immense. Ideally, it would be desirable to de-velop methods that allow the quantitative assess-ment of cell-bound immunity and the synergistic orantagonistic action of humoral antibodies in relationto it, in each untreated patient and follow it sub-sequently during treatment. While this should befeasible, at least in principle, its practical applica-tion is still in the future. Meanwhile, an empiricalapproach, based on as much rational reasoning asthe experimental models will allow, may yieldimportant information, as the work of Mathe andhis group clearly indicates.109

Obviously, the prevention approach will have toawait the further clarification of the relationshipbetween serum conversion and tumor development,preferably from a prospective study. A discussion ofthis beyond the general statement that the ultimategoal of an immunological approach must be pre-vention rather than therapy appears premature atthe present time.

AcknowledgmentsMost of the work reported here has been carried out in

collaboration with Drs Peter Clifford, W. Henle, G. Henle,J. Yata, J. Stjernsward, E. Klein, T. Tachibana, K. Nishioka,B. G. Goldstein, G. Pearson, L. Gergely, A. de Schryver,

151P



http://pmj.bm

j.com/

Postgrad M


arch 1971. Dow

nloaded from

http://pmj.bmj.com/

152 George Klein

B. Johansson,P. Gunven, Nina and Jerzy Einhorn, as reportedin the technical publications listed.

References1. KLEIN, G. (1969) Experimental studies in tumor im-

munology. Federation Proceedings, 28, 1739.2. KLEIN, G. (1966) Tumor antigens. Annual Review of

Microbiology, 20, 223.3. OLD, L.J. & BOYSE, E.A. (1964) Immunology of experi-

mental tumors. Annual Review of Medicine, 15, 167.4. SJOGREN, H.O. (1965) Transplantation methods as a

tool for detection of tumor specific antigens. Progressin Experimental Tumor Research, 6, 289.

5. PASTERNAK, G.I. (1969) Antigens induced by the mouseleukemia viruses. Advances in Cancer Research, 12, 1.

6. DEICHMANN, G.1. (1969) Immunological aspects ofcarcinogenesis by deoxyribonucleic acid tumor viruses.Advances in Cancer Research, 12, 101.

7. HELLSTROM, K.E. & HELLSTROM, 1. (1969) Cellularimmunity against tumor antigens. Advances in CancerResearch, 12, 167.

8. BURKITT, D. (1963) A lymphoma syndrome in tropicalAfrica. International Review ofExperimental Pathology,2, 69.

9. CLIFFORD, P. (1966) Further studies in the treatment ofBurkitt's lymphoma. East African Medical Journal, 43,179.

10. BURKITT, D. (1967) Chemotherapy of jaw tumors.Treatment of Burkitt's Tumour, UICC MonographSeries (Ed. by J. H. Burchenal), Vol. 8, p. 94.Springer Verlag, Heidelberg.

11. NGU, V.A. (1965) The African lymphoma (Burkitttumours): Survivals exceeding two years. BritishJournal of Cancer, 19, 101.

12. KLEIN, E. & KLEIN, G. (1964) Antigenic properties oflymphomas induced by the Moloney agent. Journal ofthe National Cancer Institute, 32, 547.

13. KLEIN, G., KLEIN, E. & HAUUHTON. G. (1966) Varia-tion of antigenic characteristics between different mouselymphomas induced by the Moloney virus. Journal ofthe National Cancer Institute, 36, 607.

14. MOLLER, G. (1961) Demonstration of mouse isoanti-gens at the cellular level by the fluorescent antibodytechnique. Journal of Experimental Medicine, 114, 415.

15. KLEIN, E., KLEIN, G., NADKARNI, J.J., NADKARNI, J.S.,WIGZELL, H. & CLIFFORD, P. (1968) Surface IgM-kappaspecificity on a Burkitt lymphoma cell in vivo and inderived culture lines. Cancer Research, 28, 1300.

16. KLEIN, G., CLIFFORD, P., HENLE, G., HENLE, W., OLD,L.J. & GEERING, L. (1969) EBV-associated serologicalpatterns in a Burkitt lymphoma patient during regres-sion and recurrence. International Journal of Cancer,4, 416.

17. VAN FURTH, R., KLEIN, E., NADKARNI,J.J. & NADKARNI,J.S. Personal communication.

18. KLEIN, E., ESKELAND, T., INOUE, M., STROM, R. &JOHANNSON, B. (1970) Surface immunoglobulin-moieties on lymphoid cells. Experimental Cell Research,62, 133.

19. SINGHAL, S.K. & WIGZELL, H. (1970) Cognition andrecognition of antigen by cell associated receptors.Progress in Allergy, 15. (In press.)

20. JOHANSSON, B. & KLEIN, E. (1970) Cell surface localizedIgM-kappa immunoglobulin reactivity in a case ofchronic lymphocytic leukaemia. Clinical and Experi-mental Immunology, 6, 421.

21. LEVIN, A.G., FRIBERG, S. & KLEIN, E. (1969) Xeno-transplantation of a Burkitt lymphoma culture linewith surface immunoglobulirl specificity. Nature, 222,997.

22. CLIFFORD, P., GRIPENBERG, N., KLEIN, E., FENYO,E.M. & MANOLOV, G. (1968) Treatment of Burkitt'slymphoma Lancet, ii, 517.

23. NADKARNI, J.S., NADKARNI, J.J., CLIFFORD, P., MANO-LOV, G., FENYO, E.M. & KLEIN, E. (1969) Characteris-tics of new cell lines derived from Burkitt lymphomas.Cancer, 23, 64.

24. KLEIN, G., CLIFFORD, P., KLEIN, E., SMITH, R.T.,MINOWADA, J., KOURILSKY, F.M. & BURCHENAL, J.H.(1967) Membrane immunofluorescence reactions ofBurkitt lymphoma cells from biopsy specimens andtissue cultures. Journal of the National Cancer Institute,39, 1027.

25. HENLE, G. & HENLE, W. (1966) Immunofluorescence incells derived from Burkitt's lymphoma. Journal ofBacteriology, 91, 1248.

26. HENLE, G. & HENLE, W. (1967) Immunofluorescence,interference, and complement fixation techniques inthe detection of herpes-type virus in Burkitt tumor celllines. Cancer Research, 27, 2442.

27. ZUR HAUSEN, H., HENLE, W., HUMMELER, K., DIEHL,V. & HENLE, G. (1967) Comparative study of culturedBurkitt tumor cells by immunofluorescence, auto-radiography and electron microscopy. Journal ofVirology, 1, 830.

28. EPSTEIN, M.A., ACHONG, B.G. & BARR, Y.M. (1964)Virus particles in cultured lymphoblasts from Burkitt'slymphoma. Lancet, i, 702.

29. KLEIN, G., PEARSON, G., NADKARNI, J.S., NADKARNI,J.J., KLEIN, E., HENLE, G., HENLE, W. & CLIFFORD, P.(1968) Relation between Epstein-Barr viral and cellmembrane immunofluorescence of Burkitt tumor cells.I. Dependence of cell membrane immunofluorescenceon presence of EB virus. Journal of ExperimentalMedicine, 128, 1011.

30. NADKARNI, J.S., NADKARNI, J.J., KLEIN, G., HENLE,W., HENLE, G. & CLIFFORD, P. (1970) EB viral antigensin Burkitt tumor biopsies and early cultures. Interna-tional Journal of Cancer, 6, 10.

31. FIALKOW, P.J., KLEIN, G., GARTLER, S.M. & CLIFFORD,P. (1970) Clonal origin for individual Burkitt tumours.Lancet, ii, 384.

32. ZUR HAUSEN, H., SCHULTE-HOLTHAUSEN, H., KLEIN,G., HENLE, W., HENLE, G., CLIFFORD, P. & SANTESSON,L. (1970) EB-virus DNA in biopsies of Burkitt tumoursand anaplastic carcinomas of the nasopharynx.Nature, 228, 1016.

33. HENLE, W. & HENLE, G. (1969) Proceedings of theInternational Symposium of Comparative LeukemiaResearch, Cherry Hill. Karger, Basle.

34. GERGELY, L., KLEIN, G. & ERNBERG, 1. (1971) Appear-ance of EBV-associated antigens in infected Raji cells.Virology. (In press.)

35. RAPP, F., BUBTEL, J.S., FELDMAN, L.A., KITAHARA, T.& MELNICK, J.L. (1965) Differential effects of inhibitorson the steps leading to the formation of SV40 tumorand viral antigens. Journal of Experimental Medicine,121, 935.

36. MEYER, M.G., BIRG, F. & BONNEAU, M.H. (1969)Cinetique de l'antigene de membrane dans le systemevirus polyome-hamster. Comptes rendus hebdomadairesdes seances de l'Academie des sciences, 268, 2848.

37. SMITH, R.T., KLEIN, G., KLEIN, E. & CLIFFORD, P.1967) Studies of the membrane phenomenon in culturedand biopsy cell lines from the Burkitt lymphoma.Advances in Transplantation (Ed. by J. Dausset, R.Hamburger and G. Math6), Vol. 779, p. 484. Munks-gaard, Copenhagen.

38. PEARSON, G., KLEIN, G., HENLE, G., HENLE, W. &CLIFFORD, P. (1969) Relation between Epstein-Barrviral and cell membrane immunofluorescence in Burkitt


on Decem


j.bmj.com

/P


j.47.545.141 on 1 March 1971. D

ownloaded from

http://pmj.bmj.com/

Immunological studies on Burkitt's lymphoma 153

tumor cells. IV. Differentiation between antibodiesresponsible for membrane and viral immunofluo-rescence. Journal of Experimental Medicine, 129, 707.

39. PEARSON, G., DEWEY, S., KLEIN, G., HENLE, G. &HENLE, W. (1970) Correlation between antibodies toEpstein-Barr virus (EBV)-induced membrane antigensand neutralization of EBV infectivity. Journal of theNational Cancer Institute, 45, 989.

40. BREMBERG, S., KLEIN, G. & EPSTEIN, A. (1969) Directmembrane fluorescence reaction of EBV-carryinghuman lymphoblastoid cells: Blocking tests with xeno-geneic antisera. International Journal of Cancer, 4, 761.

41. ROIZMAN, B. & SPRING, S.B. Alteration in immunologicspecificity of cells infected with cytolytic viruses.Proceedings of the Conference on Cross Reacting Anti-gens and Neo-antigens (Ed. by J. J. Trentin), p. 85.Williams & Wilkins, Baltimore.

42. ROIZMAN, B. (1969) Herpes viruses, membranes, andthe social behaviour of infected cells. Proceedings ofthe International Symposium on Applied and MedicalVirology, Fort Lauderdale, Florida.

43. HENLE, G., HENLE, W., CLIFFORD, P., DIEHL, V.,KAFUKO, G.W., KIRYA, B.G., KLEIN, G., MORROW,R.H., MANUBE, G.M.R., PIKE, P., TUKEI, P.M. &ZIEGLER, J.L. (1969) Antibodies to Epstein-Barr virusin Burkitt's lymphoma and control groups. Journal ofthe National Cancer Institute, 43, 1147.

44. NIEDERMAN, J.C., EVANS, A.S., SUBRAHMANYAN, L. &MCCOLLUM, R.W. (1970) Prevalence, incidence andpersistence of EB virus antibody in young adults. NewEngland Journal of Medicine, 282, 361.

45. HENLE, G., HENLE, W. & DIEHL, V. (1968) Relation ofBurkitt's tumor-associated herpes-type virus to infec-tious mononucleosis. Proceedings of the NationalAcademy of Sciences of the United States of America,59, 94.

46. HENLE, W., HENLE, G., BURTIN, P., CACHIN, Y.,CLIFFORD, P., DE SCHRYVER, A., DE THE, G., DIEHL, V.,Ho, H.C & KLEIN, G. (1970) Antibodies to Epstein-Barr virus in nasopharyngeal carcinoma, other headand neck neoplasms, and control groups. Journal ofthe National Cancer Institute, 44, 225.

47. DE SCHRYVER, A., FRIBERG, S. Jr, KLEIN, G., HENLE,W., HENLE, G., DE THt, G., CLIFFORD, P. & Ho, H.C.(1969) Epstein-Barr virus-associated antibody patternsin carcinoma of the post-nasal space. Clinical andExperimental Immunology, 5, 443.

48. JOHANSSON, B., KLEIN, G., HENLE, G. & HENLE, W.(1970) Epstein-Barr virus (EBV)-associated antibodypatterns in malignant lymphoma and leukemia. I.Hodgkin's disease. International Journal of Cancer,6, 450.

49. OLD, L.J., BOYSE, E.A., OETTGEN, H.F., DE HARVEN, E.,GEERING, G., WILLIAMSON, B. & CLIFFORD, P. (1966)Precipitating antibody in human serum to an antigenpresent in cultured Burkitt's lymphoma cells. Proceed-ings of the National Academy of Sciences of the UnitedStates of America, 56, 1699.

50. OLD, L.J., BOYSE, E.A., GEERING, G. & OETTGEN, H.F.(1968) Serologic approaches to the study of cancer inanimals and in man. Cancer Research, 28, 1288.

51. GOLDSTEIN, G., KLEIN, G., PEARSON, G. & CLIFFORD,P. (1969) Direct membrane immunofluorescence re-action of Burkitt's lymphoma cells in culture. CancerResearch, 29, 749.

52. KLEIN, G., GEERING, G., OLD, L.J., HENLE, G., HENLE,W. & CLIFFORD, P. (1969) Relation between Epstein-Barr viral and cell membrane immunofluorescence inBurkitt tumor cells. III. Comparison of blocking ofdirect membrane immunofluorescence. Journal ofExperimental Medicine, 129, 697.

53. KLEIN, G., GEERING, G., OLD, L.J., HENLE, G., HENLE,W. & CLIFFORD, P. (1970) Comparison of the anti-EBVtiter and the EBV-associated membrane reactive andprecipitating antibody levels in the sera of Burkittlymphoma and nasopharyngeal carcinoma patients andcontrols. International Journal of Cancer, 5, 185.

54. AHLSTROM, C.G., ANDERSSON, T., KLEIN, G. &AKERMAN, M. (1967) Malignant lymphoma of 'Burkitttype' in Sweden. International Journal of Cancer, 2, 583.

55. GUNVtN, P., KLEIN, G., HENLE, G., HENLE, W. &CLIFFORD, P. (1970) Antibodies to Epstein-Barr virus(EBV) associated membrane (MA) and viral capsid(VCA) antigens in African Burkitt lymphoma patientsand controls. Nature, 228, 1053.

56. EINHORN, N., KLEIN, G. & CLIFFORD, P. (1970)Increase in antibody titer against the EBV associatedmembrane antigen complex in Burkitt's lymphoma andnasopharyngeal carcinoma after local irradiation.Cancer, 26, 1013.

57. KLEIN, G., SJOGREN, H.O., KLEIN, E. & HELLSTROM,K.E. (1960) Demonstration ofresistance against methyl-cholanthrene-induced sarcomas in the primary autoch-thonous host. Cancer Research, 20, 1561.

58. YATA, J., KLEIN, G., HEWETSON, J. & GERGELY, L.(1970) Effect of metabolic inhibitors on membraneimmunofluorescence reactivity of established Burkittlymphoma cell lines. International Journal of Cancer, 5,394.

59. FENYO, E.M.. KLEIN, E., KLEIN, G. & SWIECH, K.(1968) Selection of and immunoresistant Moloneylymphoma subline with decreased concentration oftumor specific surface antigens. Journal of NationalCancer Institute, 40, 69.

60. MANOLOV, G. Personal communication.61. HAUSCHKA, T.S., KVEDAR, B.J., GRINNEL, S.T. & AMos,

D.B. (1956) Immuno-selection of polyploids from pre-dominantly diploid cell populations. Annals of the NewYork Academy of Sciences, 63, 683.

62. BURKITT, D. & KYALWAZI, S.K. (1967) Spontaneousremission of African lymphoma. British Journal ofCancer, 21, 14.

63. GERGELY, L., KLEIN, G. & ERNBERG, 1. (1971) Effect ofEBV-induced early antigens on host cell macromolecu-lar synthesis, studied by combined immunofluorescenceand radioautography. Virology. (In press.)

64. HENLE, G., HENLE, W., KLEIN, G., GUNVtN, P., CLIF-FORD, P., MORROW, R.H. & ZIEGLER, J.L. (1971) Anti-bodies to early EBV-induced antigens in Burkitt'slymphoma. Journal of the National Cancer Institute.(In press.)

65. ZUR HAUSEN, H. & SCHULTE-HOLTHAUSEN, H. (1970)Presence of EB virus nucleic acid homology in a 'virus-free' line of Burkitt tumor cells. Nature, 227, 245.

66. MATHt, G. (1967) Antigenicit6 nouvelle (demontr~epar isogreffe) d'un fragment de peau de souris infect~epar un virus leucdmogene. Comptes rendus hebdoma-daires des seances de l'Acadjmie des sciences, 264, 2702.

67. PASTERNAK, G. (1965) Serologic studies on cells ofGraffi virus-induced myeloid leukemia in mice.Journal of the National Cancer Institute, 34, 371.

68. SJOGREN, H.O. & HELLSTROM, 1. (1965) Induction ofpolyoma specific transplantation antigenicity in Molo-ney leukemia cells. Experimental Cell Research, 40, 208.

69. STUCK, B., OLD, L.J. & BOYSE, E.A. (1964) Antigenicconversion of established leukaemias by an unrelatedleukaemogenic virus. Nature, 202, 1016.

70. SVET-MOLDAVSKY, G.J., MKHEIDZE, D.M. & LIOZNER,A.L. (1967) Phenomena associated with skin grafting.Two phenomena associated with skin grafting fromtumor-bearing syngeneic donors. Journal ofthe NationalCancer Institute, 38, 933.


on Decem


j.bmj.com

/P


j.47.545.141 on 1 March 1971. D

ownloaded from

http://pmj.bmj.com/

154 George Klein

71. BURKITT. D. (1969) Etiology of Burkitt's lymphoma-an alternative hypothesis to a vectored virus. Journal ofthe National Cancer Institute, 42, 19.

72. SOUTHAM, C.M., TANAKA, S., ARATA, T., SIMKOVIC, D.,MIURA, M. & PEPTIOPULES, S.F. (1969) Enhancementof responses to chemical carcinogens by nononcogenicviruses and antimetabolites. Progress in ExperimentalTumor Research, 11, 194.

73. KAFUKO, G.W. & BURKITT, D.P. (1970) Burkitt'slymphoma and malaria. International Journal ofCancer, 6, 1.

74. RUBIN, H. (1966) A virus in chick embryos which in-duces resistance in vitro to infection with Rous sarcomavirus. Proceedings of the National Academy of Sciencesof the United States of America, 46, 1105.

75. RICH, M.A. (1968) Virus-induced murine leukemia.Experimental Leukemia (Ed. by M. A. Rich), p. 1.North Holland Publishing Co., Amsterdam.

76. CHURCHILL, A.E. (1968) Herpes-type virus isolated incell culture from tumors of chickens with Marek'sdisease. 1. Studies in cell culture. Journal of the NationalCancer Institute, 41, 939.

77. MIZELL, M., TOPLIN, 1. & ISAACS, J.J. (1969) Tumorinduction in developing frog kidneys by a zonal centri-fuge purified fraction of the frong herpes-type virus.Science, 165, 1134.

78. HUNT, R.D., MEL&NDEZ, L.V., KING, N.W., GILMORE,C.E., DANIEL, M.D., WILLIAMSON, M.E. & JONES, T.C.(1970) Morphology of disease with features of malig-nant lymphoma in marmosets and owl monkeys inocu-lated with herpesvirus saimiri. Journal of the NationalCancer Institute, 44, 447.

79. DE SCHRYVER, A., KLEIN, G. & DE THi, G. (1970)Surface antigens on lymphoblastoid cells derived fromnasopharyngeal carcinoma. Clinical and ExperimentalImmunology, 7, 161.

80. PREHN, R.T. (1963) The role of immune mechanisms inthe biology of chemically and physically inducedtumors. Conceptual Advances in Immunology and On-cology, p. 475. Hoeber, New York.

81. MUIR, C.S. & SHANMUGARATNAM, K. (1967) Cancer ofthe Nasopharynx, p. 1. UICC Monograph series.Munksgaard, Copenhagen.

82. ABERCROMBIE, M. (1966) Contact inhibition: Thephenomenon and its biological implications. TheSecond Decennial Review Conference on Cell Tissueand Organ Culture, Bedford, Penn., p. 249.

83. BURGER, M.M. (1968) Isolation of a receptor complexfor a tumour specific agglutinin from the neoplastic cellsurface. Nature, 219, 499.

84. INBAR, M. & SACHS, L. (1969) Structural differences insites on the surface membrane of normal and trans-formed cells. Nature, 223, 710.

85. HAKAMORI, S.I. & MURAKAMI, W.T. (1967) Glycolipidsof hamster fibroblasts and derived malignant-trans-formed cell lines. Proceedings ofthe National Academy ofSciences of the United States of America, 59, 254.

86. MORA, P.T., BRADY, R.O., BRADLEY, R.M. & McFAR-LAND, V.W. (1969) Gangliosides in DNA virus-trans-formed and spontaneously transformed tumorigenicmouse cell lines. Proceedings of the National Academyof Sciences of the United States of America, 63, 1290.

87. KLEIN, G., BREGULA, U., WIENER, F. & HARRIS, H.(1971) The analysis of malignancy by cell fusion. I. Hy-brids between tumour cells and L cell derivatives.Journal of Cell Science. (In press.)

88. WIENER, F., KLEIN, G. & HARRIS, H. (1971) Theanalysis of malignancy by cell fusion. IlI. Hybridsbetween diploid fibroblasts and other tumor cells.Journal of Cell Science. (In press.)

89. KLEIN, G. & HARRIS, H. (1971) To be published.

90. RAB1NOWITZ, Z. & SACHS, L. (1970) The formation ofvariants with a reversion of properties of transformedcells. Virology, 40, 193.

91. MACPHERSON, 1. (1970) The characteristics of animalcells transformed in vitro. Advances in Cancer Research,13, 169.

92. POLLACK, R.E., GREEN, H. & TODARO, G.J. (1968)Growth control in cultured cells: Selection of sublineswith increased sensitivity to contact inhibition and de-creased tumor-producing ability. Proceedings of theNational Academy of Sciences of the United States ofAmerica, 60, 126.

93. YATA, J. & KLEIN, G. (1969) Some factors affectingmembrane immunofluorescence reactivity of Burkittlymphoma tissue culture cell lines. InternationalJournal of Cancer, 4, 767.

94. EPSTEIN, M.A., ACHONG, B.G., BARR, Y.M., ZAJAC,B., HENLE, G. & HENLE, W. (1966) Morphological andvirological investigations on cultured Burkitt tumorlymphoblasts (strain Raji). Journal of the NationalCancer Institute, 37, 547.

95. SVEDMYR, A., DEMISSIE, A., KLEIN, G. & CLIFFORD, P.(1970) Antibody patterns in different human seraagainst intracellular and membrane-antigens andneutralization ofEBV infectivity. Journal ofthe NationalCancer Institute, 44, 595.

96. HENLE, W., HENLE, G., ZAJAC, B.A., PEARSON, G.,WAUBKE, R. & SCRIBA, M. (1970) Differential reactivityof human sera with EBV-induced 'early antigens'.Science, 169, 188.

97. HENLE, W, DIEHL, V., KOHN, G., ZUR HAUSEN, H. &HENLE, G. (1967) Herpes-type virus and chromosomemarker in normal leukocytes after growth with irra-diated Burkitt cells. Science, 157, 1064.

98. POPE, J.H., HORNE, M.K. & SCOTT, W. (1969) Identi-fication of the filtrable leukocyte-transforming factor ofQIMR-WIL cells as herpes-like virus. InternationalJournal of Cancer, 4, 255.

99. GERBER, P., WHANG-PENG, J. & MONROE, J.H. (1969)Transformation and chromosome changes induced byEpstein-Barr virus in normal human leukocyte cultures.Proceedings of the National Academy of Sciences of theUnited States of America, 63, 740.

100. BURKITT, D. (1967) Chemotherapy of jaw tumors.Treatment of Burkitt's Tumours, UICC MonographSeries (Ed. by J. H. Burchenal), Vol. 8, p. 2. SpringerVerlag, Heidelberg.

101. KLEIN, G., CLIFFORD, P., KLEIN, E. & STJERNSWARD,J. (1966) Search for tumor specific immune reactionsin Burkitt lymphoma patients by the membraneimmunofluorescence reaction. Proceedings of theNational Academy of Sciences of the United States ofAmerica, 55. 1628.

102. KLEIN, G., CLIFFORD, P., KLEIN, E. & STJERNSWARD, J.(1967) Search for tumor specific immune reactions inBurkitt lymphoma patients by the membrane immuno-fluoresence reaction. Treatment of Burkitt's Tumour,UICC Monograph Series (Ed. by J. H. Burchenal),Vol. 8, p. 209. Springer Verlag, Heidelberg.

103. NISHIOKA, K., TACHIBANA, T., KLEIN, G. & CLIFFORD,P. (1968) Complementological studies on tumor im-munity. Measurement of C1 bound to tumor cells andimmune adherence with Burkitt lymphoma cells. GannMonograph, 7, 49.

104. STJERNSWXRD, J., CLIFFORD, P., SING, S. & SVEDMYR,E. (1968) Indications of cellular immunological re-actions against autochthonous tumor in cancerpatients studied in vitro. East African Medical Journal,45, 484.

105. KLEIN, G. (1966) Humoral and cell-mediated mech-anisms for host defense in tumor immunity. Viruses


on Decem


j.bmj.com

/P


j.47.545.141 on 1 March 1971. D

ownloaded from

http://pmj.bmj.com/

Immunological studies on Burkitt's lymphoma 155

Inducing Cancer, Implications for Therapy (Ed. byW. J. Burdette), p. 323. University of Utah Press, SaltLake City.

106. HELLSTR6M, I. & HELLSTROM, K.E. (1970) Colonyinhibition studies on blocking and non-blocking serumeffects on cellular immunity to Moloney sarcomas.International Journal of Cancer, 5, 195.

107. HELLSTR6M, I., HELLSTROM, K.E., EVANS, C.A.,HEPPNER, G.H., PIERCE, E.G. & YANG, J.P.S. (1969)Serum-mediated protection of neoplastic cells frominhibition by lymphocytes immune to their tumor-specific antigens. Proceedings of the National Academyof Sciences of the United States of America, 62, 362.

108. SKIPPER, H.E. (1967) Destruction of leukemia cells in

animals. Criteria associated with destruction of leu-kemia and solid tumor cells in animals. Cancer Research,27, 2636.

109. MATHf, G. (1969) Approaches to the immunologicaltreatment of cancer in man. British Medical Journal,4, 7.

110. KLEIN, E., CLIFFORD, P., KLEIN, G. & HAMBERGER,C.A. (1967) Further studies on the membrane immuno-fluorescence reaction of Burkitt lymphoma cells.International Journal of Cancer, 2, 27.

111. KLEIN, G., KLEIN, E. & CLIFFORD, P. (1967) Search forhost defenses in Burkitt lymphoma: membrane immuno-fluorescence tests on biopsies and tissue culture lines.Cancer Research, 27, 2510.


on Decem


j.bmj.com

/P


j.47.545.141 on 1 March 1971. D

ownloaded from

http://pmj.bmj.com/

Immunological studies Burkitt's lymphoma · Immunologicalstudies onBurkitt's lymphoma...

Documents

Transcript of Immunological studies Burkitt's lymphoma · Immunologicalstudies onBurkitt's lymphoma...