Tobacco Documents Online

Medical Hypotheses 14: 181-198, 1984 18 CANCER COMMUNICABLB!• Henry T. Lynch,l t3uy S. Schuelkell and MiehaGl K. O7IaraZ Creighton Uni~versity Sdibol of Medicine Dapnrtments of Preventive Medicine/Public Healthl and Medical MlcrobiologyZ Omaha, NE 68178 ABSTRACT I INTRODUCf1ON Recent developments in canaer epidsmiology have led to the passibility of an exceedingly complex communicable factor(s) In cancer etiology. The transmission of such an agent(s) may require a susceptible genotype end/or other promotional events. . Likely candidates which support this supposition Include: 1) Spstein-Berr virus (nasopharpngeal carcinoma, Burkitt's lymphoma, salivary gland tumor among Eskimos, X-ltNcdnd lymphoproliferative syndrome of Purtllo)i 2) human T-ceU leukemia vlrus (adult T-eeII leukemia); 3) aeqaired immune deficienty syndrome (AIDS), complicated by Kaposi's sarcoma (etiologie agent remains elusive, though epidemiology suggests possible infeetious transmission)= 4) abnormal immune phenomena In households of Hodgkin's disease patients= and 5) clustering of tlrarious types of cancer in speuses, the general population, and families. We have selectively reviewed the literature and evolved an etiologic hypothesis which Integrates a communicable agent(s) In concert with genetic andror environmentsl carcinogenic interaetion which could conceivably explain a significant fraction of the total cancer burden. The etiology of cancer remains elusive. Nevertheless, when meticulous clinical laboratory-erlenied medie,4l-genetic%pidemiologie atudies are performed on large series of patients with apecitie histolog(e forms of cancer, subsets emerge which lend support to a variety of etiologic hypotheses. At one extreme there are patients r4ho show findings which support clear-cut hereditary etiology. as in autos4mal dominantly inherited familial multiple adenomatous polyposis eolt, retinoblestoma, the multiple nevoid basal cell carcinoma syndrome, the autosomal reeeesive chromosome breakage syndromes (e.g., ataxia telangleetasia, Petqeoai's eplostie anemia, and Bloom's syndrome), the X-linked lymphoproliferative syndrome, ahd an increasing variety of newly described autosomal dominant eaneer-prona pedigrees (Figure 11 see next page) (1). eReprint requests to Dr. H.T. Lynch 181

P, f N FIGURE [. PEDIGREE OF A KINDRED lt/UtIFESTING THE CANCER FAMILY SYNDROME t 4 sl C-49 1 I `d.t9 d. 43 d.]9 cgf t-LSt4 s • cta) 0. ~s 1 it-t5 2t-1s 39.26 34.31 77-l0 uRNA Caea 9sltla.ttow fete.er 10 -r.ebslew. ®-tlcah t.nttlctec. 4) •rWey 111aoey.0 -/Yttlvlo tri..rl % •PeNoM. O.IltI..O•r.al. i-Itaat 11-1ledder IT-It.fe twot Gca1o. trtwletsetlw 6t-tlf. St-fteottb ftnhotteto lrtrae.u 0

Alternatively, cancer-affected patients with 'etioldgies consonant with strong environmental influences will be even more freQuently encountered. A typical example of the latter Is eigarette smoking habituation In asbestos workers who harbor more than an 80-fold excess caneee risk (lung, pleura, colon) Ahen their nonsmoking, nonasbestos-working cohorts. However, In this example, as well as in other strongly "environmental" circumstances (the classicel evidence of serotal cancer In chimney sweeps and benzo(a)pyrene in urinary bladder cancer), one must still be cognizant of possible hereditary susceptibility to the specific carcinogenAc agents (2).. This, supposition has been supported by remarkable familial clusterings of inesotheliomas and asbestos exposure (3,4). A phenomenon recently recognized, In 'genetie" ss well as "environmental" etiologic settings is that of horizontal as well as vertical transmission of cancer through a complex Interaction p+ith a putative communieable agent (5). Such an infectious agent could conceptually act in concert with genetic and/or environmental carcinogenic agents (6-8). HYPOTHESIS We propose a cancer etiologic hypothesis which integrates a communicable agent in concert with genetic and/ar environmental carcinogenic interactions. In any given cancer occurrence, at least one of these three factors will be of paramount importance. However, the majority of elinieal cancers probably represent a complex interaction between these respective factors. Such an hypothesis would provide a plausible etiologic explanation for a large fraction of the total cancer burden. Should this hypothesis be confirnied, It would have a major bearing on strategies employed for cancer control as well as for, the experimental design of studies in the search for clues to cancer etiology. For example, presumed "spouse controls' might become suspect In certain circumstances, such as those where connubiality may represent a non-chance pheinom~on. Although a very small fraction of occurrences of any specific form of cancor, such as multiple myeloma (9), may exhibit connublaliity, one cannot reject the argument that it a transmissible factor were etiologic, then additional cannubial occurrences should necessarily be found. We make this statement since there is substantial evidence for extant ~hetdrog~eneity of all forms of, eanqer (1), including of course multiple myelom- a: -~, certain of the connubial occurrences could be due to a form of cancer (including multiple my,elomp) which is strongly Influenced by a transmissible agent (onebgenie virus4). After prolonged contact (respiratory, venereal), a patient harboring a cancer susceptible genotype, may beeome "Infected" from his affectbd partner, thereby explaining its manifestation in spouses. Conversely, a genetically cancer resistant spouse might become susceptible due to an excessively large and/or protracted exposure. The capability of msny human and animal viruses to maintain a long-term association with their hosts (6.g., the lateht and persistent nsture of herpetoviridae infection) is consistent with these possibilities. Human date which support this hypothesis are summarized in Table 1. K.a-o 1g3

TABLE 1. Literature evidence for genotypic ieterretloro vdith environmental faotoes In the evoluticn of human maltgnaney. I SEZRlNa MALIGNANCY MALIGNANCY OR RELATED OBSERVED QR FINDINGS REPERENCES spousos several typo of cancer In different spouae peie earvleal-penlte multiple mqeloma Kapost's sarcomat AIDS immunologieal sensitization maltgnant melanoma HoQgkin's disease elevated CEA levels in CPS Idndreds non-Hodglctn'a lymphoma aeute leukemia colanle carcinoma 4,10,11,15 30 12 13 populatian eeneer adult T-eell leucemfa 35,36 olustering Burtdtt's lqmphoma 31-34,30 nasopheryngeal oaralnonoa 37 and salivary gland earclaoma multiple myeloma 39 cancer patient eontaets oeteoaarcoma 44,45 breast cancer 42 sarcoma 40,41 adult T-ee11 leukemia 35,46 neueobl®stoma 43,44 Itmg cancer 48 Hodgldn's dtsease 49 masothelioma 4 melanoma 47 leukemia aed lymphoma 80 femiAal a8gregations mesothelloma 3 hepatoeenuter carcinoma 77,79 T-eell leukemla-Iqmphofia 46,80 troeopharqngeal eareinoma 66,67 Burldtt's lymphoma 58,66,67 ovarian oaneer 81,83 twin eeneordenee fer emeer 56,81 leukemia and lymphome 56,57,64,05, 69,70 malignant melanoma 55,83 Hodgkin's disease 61,68,73,75 184

noril-Hodgkin's lymphomas 69 asteogdade sarcoma 71,74 Kspcel's sarcoma 69,64,72 X-Fltnkad lymphoproli(erative 78,82 sqndrohne genetic associations overq 64 melenoma 90 HLA 86,87•89,93 Gm loci 99 glu4amate-pyruvate 86 traeeaminase locus CANCER HYPOTHESIS BAS$D UPON COMMUNICABLB AGENTS I etiologic meohenimns. Evidence whidi supports our communicable hypothesis will be reviewed. This evide4ee emanates from ttNe disciplines of genetics, immunology, virology, and environmental earelnogenes#s ®nd was derived from clinical settings involving a broad spectrum of cancers with a diverse hrray of potential SO=" Table I summarizes much of the available human cancer-related oDservations which support an environntental-genetie Interactive etiology. Most suggestive of a strong environmental Influence on eaneer etio1ogy are those instanoas wherein genetieailq dissimiior Individuals develop virtually identieal cancers and/or manifest cancer-related phenomena. In Table 1, this genetic diseihtllaritq requirement is demonstrated In the instances oft 1) Identical cancers oeeurcLtg ln both spouses (9-16)s 2) Wdcs between aervieel and penile eancera (17-4sh 3) multiple myeloma, including two eases of myeloma in the second spouse of Individuals whose first epouse had died of myeloma (8,Z3h 4) cancer ln the partner of a homosexual with ICapoet's sareoma (74)t 5) acquired immune deficiency (A1DS) In female se:ual partners of males with AIDS (25h 6) lmmune sensitization tQ tumor antigens correlated with length of spouse contact with patients (14,46,Z7)z and 7) elevated CEA levels In spouses of cancer-affected or high cancer risk lndlvlduis from Cancer Family Syndrome (CFS) families (28). Population Clusterhvt fipidemiologieal evidence of various types of cancers clustering In 185

genetically mixed populations is another observetion compatible with a prominent environmental influence. Cancers in eertain of these settiry'gs are candidetes for etiologic involvement of a communicable oncogenic virus (31-39). Oncogenicity In certain af these circumstances may be modulated by fectors Influencing host susceptibility to viruses and/or carcinogenic exposures (5). Mechanlstieaft, variations In susceptibility might, for example, involve enzyme systems where variations In , the metabolic activity of enzymes which convert carcinogenm to their active form have been found (2). Cancer Patient Contacts Research dealing with variable forms of contact with cancer patients has also disclosed eeveral Iines of evidence suggesting aprominent environmental effect. This evidence includess 1) Immunological sensitization of household contacts (40-47), and even leboratory/medical personnel (48), to tumor antigens; , 2) abnormal immune responses in spouses, relatives, and household contacts (47,48)s 3) social links between leukemia and lymphoma patients (58r 4) inductibn of a common sarcoma antigen on tissue culture cella by e cell-free tumor eell extraet (40h and 5) home exposure to asbestos in the special case of inesothelioma (4,51). Familial/Genetic AaQreeations In those cancers which have been observed to aggregrate In families (1,52-83), either primary genetic factors and/or a shared environment could account for these eancer clusterings. Where inheritnnee patterns are evident, as in the kindred presented in Figure 1, genetic factbrs are of paramount etiologic Importance. Nevertheless, genes do tat operate In a vacuum. Such families should therefore provide valuable resources for identifying putative• environmental agents with carcinogenic potential in a genetically susceptible host. This supposition is supported by the X~-linked lymphoproliferative syndrome wherein a genetic dsfect In Immunity predisposes to malignancy when the Epstein-Barr virus (EBV) is encountered (78,82). Possible genetic susceptibility to infection by human T-cell leukemie/lyrnphoma virus (HTLV) (84) may provide a further example of a putative oneogenie virus which Is also highly interactive with the immune system (35). lmmune abnormalititss have been a common finding in kindreds with multiple types of cancers showing aggregation (8S). Reeent evidence suggests that other cancers may have a primary genetic basis (83,86) and by implieation there Is the possibility of oncogenie viral involvement (and/or other transmissible oneogenie agents) in confunction with a cancer susceptible genotype. Genetie Cancer Associations In the General PoDul®tion Population studies also support a primary genetic etiologic role in specific I 186

eanoer oceurrenees. These inolude asaoQlatlons of 8LA .otigena (87r81),,, Gm loai (92,93), and eniymae with aaneer riak (94,95). The moohsniams whereby these loci may be linked to barmer rlsk are net kn4ean at pxeaent. In at lerat rome such e.aeer aseooiation3, )inkage with k cancer (or virus4) susoeptibility gene Is a posalbWty. Table !(see next pagb) delineat'®s viruses which are etlologlaally respans::ile for eaneer in IntraAumen systems (9A-120). , Giren knowledge of evolutionary prinaipies, it would appear unilkely tMt man wrould prove to be an eseepttan to vlral lnduetion of eaneer. In these inirahuman systems, host geneties In kaown to play a bruotal role ln the events suEbequent to contact with an onoogeale rirus W. Moreover, experlmental manlpulations oae effect the outcome of Infection. Thus, these experimental modbm probebly reflect events occurring in man where extremely complex and uncontrolled environmental taators Interact with genotypes of variable eenw prmemness. Dl$CtTSSfON Given the rapidly emerging evidence in support of In oncogenie viral etiology for a iraetion of Infrahuman cancer (Table 2), coupled with solid findings inerlminating human oneogenlc viruses (i.e., human T-eell leukamie/lymphoma virus (KTLV) In adult T-eeR leukemle-lymphoma (ATL) (35), and Epstein-Barr (8BY) virus etiology in var~p~ adaples of Surkitt'a 1ym phoma, nasopheryngesl oarelnome, and %1tNted raliferative oyadiome (ZT,119)), one finds strong support tor a eommunlcable teetot (oneogenle vitust) in en anknown tra,etlon of human cancer occurrences. This complex problem will undoubtedly center around the Issue of Integration of bueh an °inteo%tous" agent with one cr more raqulrements for onoogentcity to occurs l.e., 1) cancer suseeptlble penotype(initiatar) (7)1 2) promotor effects ,n terms of speeltle eevircnmental oareinogeeaa 3) ege at first Infection of the putative oneogenie rirusE 4) the vleal route of infeetiont and 8) its repetitiveness In terms of contact end/ar reactivation. A further requirement will be en optimum milieu for maximum efficiency of the initatlon-promotlon effect, sasumtng a two-hlt .Iar multiple extra hits?) mutatlbnal hypothesis (8). COMMONICABIIdTY IN NONCANCBR CHRONIC DE4HA8ES Evidence ot eommunienbillty Is emerging for a variety of ehronie dise+ases. Thua, we see examples of certain nonoancer rirally-lndueed neurological diseases such as Creutzfel"ekob disease and Kuru (121). Reeently, Interest was even extended for an Infeetlous etiology In a csrtain fraction of pallents with schizophrenia (122). 8peeifieally, lt was postulated that this etiologle phenomenon could be due to a transmissible virus which Infects eaetieally-pred osed individuals. There was a latency period of approximately ~ months. N~ed tamililial concordances for schizophrenia could not be explained on a purely genetic basi9. Rather, the data were consistent with the possibility that physieal, proximity of susceptible patients to afteeted (^infected") contacts was of maior etiologie Importance. Twin studies of tubercWosis and poliomyelltls suggest lhat there Is also a genetic influence comparable to that In schizophrenia In these particular Infectious diseases (122). It was speculated that, "..-in schizophrenia, exposure to an infectious agent end genetic predisposition may both be necessary for 187

'1'ABLE 2 91eos- papova.irtdae 6ovtse paptilaoa virus Slwpe papilloaa virus poirows.ires 8V40 Adeaovirldse Adeaovtraoes (buoae) Boviae adeao.lrus Siaiao adenovirus Berpetovtridaw 8erpes.lrpe-satair!- 9e atetes 8. pspio 6uioea pie berpessiree R. srlvtlasus Marek's dlseaee rtres Lscke Aerpesoirns posviridae Tabs monkey virus abepe :ibrema virus IIraamtosis virus Retroriridae Aous sarcoma virus Avlan lenkewta virus Others (100) mtocelisseow lfoodebsek hepstitte tpvlueor.x VIRAL ONCOM ta x.eeaac uost CMMM- AeaoClAT1os- ~r~exrg eo.npe Caresea.. ee Rabbit Carciooma 07 Ueose ddsoocarcioowa !n waltiple Ys tlsales fioakep ttbreorooes. sliem, leakemia, 99 l1spbe0la. esteesarcoma. and retienlvo eell sarcoma (baostsr) Ifao Traesforratloo aod ooeoteee.is 100.101 Bo.foe (tlbroeareowa) demonstrated for 102 Moaker saar adeerovaroses !a several 103 species of target ealis and/or bosta, respectively 8qnfrrel-aronkep- Leokeala. 1"mkers-(oar.osat) 1af- $pider monkey (s-1r.pbotroptal 1os Baboon Lyspboprolilerative (mar.oset) 108.107 Golnea p!s Leukesia, treasformatioo 1e8.1e9 Qottoetali rabbit Lympboprol/terattvs, 1# epbo.o 110 ChlChee LMiMwa 111 LeopaM iro` Reeal adeeoasreleo.a 112,113 Rbesns oonkeq Superficial ts.ors of il.be 114 wklcb regress Cottoatail rabbit fibrea (beslsa. sooiovasivs) 115 Rabbit ribro.a (beo/ae. aoalovui.e) lle Chickee garesa 117 Chicken Leukemia lie Iloltftple spesios Variety of tsspra 11e Woodebaek 8epatoeellslar earsieoma 120 eetro.irusse are widely distributed 1n nature (beisg taoed 1s spscles rangiaI from .orais to ®as) ssd great Sa saober (i100). Coessqsentl'. they wtll sot be listed here. "s reader is referred to tte eeal.r Blei ! Mwr 1/leo.es RpA itsor Yirvses. 1982 (sdq. &. Tsies, ~f. 7sieb, g. Valiss, ae0 i. Cotfls), Cold Sprle` arboT boratorles.

disease expression, snd in this respect, sehizophrenia dres®mbles dlseases in which the rate of Infeetion Is established." COidCUSIONS It would appear that we are mnldra[ a quantum leap i,n our oommunieable hypothesis In going from aehizephrenla to cancer. However, we purposely selected this example to support our hypothesis In order to demonstrate another disease wherein communicable factors may relate to the etiology of diseases which, heretofore, have been considered to have eithet a strong genetic predispositlon snd/ar Interaction with behavloral/eultural intluenaas. To investigate our hypothesis, It wiri be mandatory that we shift our emphasis from traditional cancer epidem(ologie approaches and focus upon a potentially broad spectrum of contributory events; e.g., time-space clusterings of cancer (comparable to ATL In selected populations, [apasi's sarooma snd -acquired immune deficiency among homosexuals), geneties, immunology-seroepidemiology, and chemical eareinogenesis; Thus, in the "eW of a putative oneogeato vieaa or other carcinogenic agent(s), all men are not equal in their susceptibility to the development of eaneer. pertaeuler emplmsis must be placed upon the manner in which these myriad factors may interact with each other along with due concern to the profound latency of eancer expresdon. We must therefore employ all possible research deedg+o and espe:hnental modals,raong from those of family studies to the' investigation of problems such as AIDS and the molecular biology of cancer (e.g., cellular/vlral ae~eogene priority. es~resetonl. Markers of susceptibility should receive high research For example, Ewing et a1 (14S) have, by electron micrroswpy, tomd ummual structuree (termed "vesieular rosettes°) In the cytoplasm ot lymphocytes from lymph nodes In 17 of 18 homosexual patients with lymphadenopethy, 3 of Q Individuals who died from AIDS, and only 2 of 31 eontrols; It was postulated that the vesicular rosettes "...may be dtsociated with an etiologic agent common to unexplained lymphadenopathy, AIDS, and sonbe lymphomas " To date, no such agent has been isolated and no vlral particles were observed In that report. However, the epidemiologie evidence, ineluding oaae elustering, as well as the occurrence of the disease among people receiving blood products, drug users, and children in households prone to AID8 (184), b eonslsitent with a transmissible agent. Therefore, it Ie possible that ".»theye rrbsettes eopld be a manifestation of airal tnfeet/on since It Ie known that viruses 4an have various morpholoqie effects on host oells without repDaatir,g." Ubiquitous viruses have also been implieated in this syndrome (125). Should these preliminary observations be verified among other patients with chronic unexplained lymphadenopathy, AIDS, and lymphoma, then the rosettes emuld,signity a factor (marker) In common, such as a virus. This could have etiolegic as well as (14thogenotit 6 endemle~o8courrenees 147,1g8) the and nmodulntion of worldwide spr~ead of AI of homosexual activity practiced (1s7). The ~e ~ua non for~o~ur pos~would be Identification of animal camterparts ot'apee`7tlo~aime~i malignant neoplastte lesions so that indlvidual components of our communicable cancer hypothesis might be tested In accord with Koeh's postulates. Availability of such models would provide a basl, for studies into possible methody of cancer prevention In man (e.g., retlnoids (130,131), tsolatlon of patients with associated human T-eeil leukemia/lymphoma virus (132), and other mathods). 189 iee . ..

ACKNOWLEDGEMENT We are grateful for support from the Council for Tobacco Research, USA, Grant No. 1297-AR1. REFERENCES 1. Lynch HT. Cancer Genetics. Charles C. Thomas Co., Springfield, 1976. 2. Harris CC, Mulvih/ll JJ, Thorgeirson SS, and Minna JP. Individual differences in cancer suseeptibility. Ann int Med 82:808, 1880. 3. Risberg B, Nickels J, and Wagermark J. Familial clustering of malignant mesothelioma. Cancer 45:2422, 1980. 4. LI FP, Lokich J, Lapey J, Neptune WB, and Wilkins EW. Familial mesothelloma after dntense asbestos exposure at home. JAMA 240:467, 1978. 5. T-W-Flennes RN. The viral theory of cancer. p 21 In Infectious Cancers of Animals and Man. Academic Press, New York, 1982. 6. Zur Hausen H. Human genital cancer: synergism between a two virus infect;on and initiating events? Lancet ii:1370, 1882. 7. Lynch HT and Guirgis HA. Childhood cancer and the SBLA syndrome. Med,Hypoth S:IS, 1979. 8. Knudson AG;. Genetic influences in human tumors. p 65 in Cancer - A Comprehensive Treatise (FF Becker, ed). Plenum, New York, 1975. 9. Kyle RA and Gretpp PR. Multiple onyeloma: houses and spouses. Cancer 51:735, 1983. 10. Cioeco A. On the mortality in brother-sister and husband-wife pairings. Human Bio 13:188, 1941. 11. Chen WY, Crittenden LB, Mantel N, and Cameron R. Site distribution' of cancer deaths in husband-wife and sibling pairs. JNCI 37:875, 1961. 12. Amos DA, Wellman WE, Bowie EJW, and Linman JW. Acute leukemia in a husband and wife. Mayo Clin Proc 42:468, 1861. 13. Law IP and Larson A. Colonic careinoma in a married couple. N Eng J Med Z97:1353, 1977. 14. Mintzis MJ, Berger AP, Greenwald 8, Greenwald L, and Golomb F. Malignant melanoma in spouses. Cancer 42:804, 1978. 15. Enek RE. Cancer in married couples. Milty Med 144:803, 1979. 16. Rus4 JE and Scanlon EF. Identical cancer in husband and wife. Surg Gyn Ob 180:664, 1980. 190