VOLUME 23 SEPTEMBER 1963 NUMBER 8

The Complex Etiology of Cancer*

CHESTER M. SOUTHAM

(Division of Clinical Chemotherapy, Sloan-Kettering Institute; and Departments of Medicine, Memorial andJames Ewing Hospitals and Cornell University Medical College,New York, New York)

SUMMARY

The recognized causes of cancer are briefly reviewed, and the inadequacy of presentknowledge to explain cancer incidence and distribution is pointed out. New knowledgeand concepts concerning carcinogenesis are then discussed, emphasizing the combinedeffect of mutiple factors; recent advances in the knowledge of the nature, growth, andpathogenicity of viruses; and the role of host factors. The concept of a complex etiology—the concurrent or sequential action of multiple etiologic factors (co-carcinogenesis inits broadest definition)—is suggested as a plausible explanation for the incidence patterns of clinical cancer.

The etiology of cancer can be considered at twolevels : (a) the fundamental derangement of normal life processes which is disease or (b) the factors which, by acting upon the organism, lead tothis derangement. In considering the first, it mustbe admitted that the cause of cancer is unknown,although recent advances in molecular biology andcytogenetics nourish the hope that an understanding of ultimate causes may be attainable. At thesecond level our ignorance is of a different sort,for we recognize a confusing multiplicity of factorswhich are indisputably associated with the development of neoplastic disease in man.

THE CAUSES OF CANCER IN MANIonizing irradiation is perhaps the most widely

recognized cause of human cancer, as tragicallyproved by the fate of early radiotherapists who un

* Based on a talk presented at the Division of Medicinal

Chemistry, Annual Meeting of the American Chemical Society,Chicago, Illinois, September 5, 1961. The author's researchwork is supported by grants from the American Cancer Society,the New York City Health Research Council (under contract#U-1096 II CS), the National Institutes of Health of the UnitedStates Public Hea1t1@Service, and the Phoebe Waterman Fund.

Received for publication January 18, 1963.

wittingly suffered over-exposure to roentgen raysand radium (115, 119, l4@), the ill-fated radiumdial painters of 1918 (70), the consumers of radiumwater in the days when it was considered a tonic(4@,45), and recipients of thorotrast for roentgenographic visualization of the liver (8).

Ultraviolet irradiation is a definite factor in thecausation of cancer as evidenced by the high frequency of skin carcinoma which develops insailors' and farmers' skin (70).

Chemicals have been recognized as causes ofcancer since the classic observations of Pott onchimney sweeps' cancer over @0Oyears ago (106),and even earlier in the writings of Paracelsus concerning the high incidence of lung disease in theminers of Eastern Europe (1@4) who we nowknow were exposed to dust high in carcinogenicmetals (46, 70, 108) as well as radioactivity. Workers in petroleum and coal tar industries (70), heavysmokers (4, v.8, 57, 150), and patients treated withhigh doses of arsenic (70) are recognized as havinga high risk of cancer. This knowledge is also reflected in our current, perhaps exaggerated, concern with cranberry sprays, wax containers, lipsticks (36), kippered herrings (30), and aged wines(79).

1105

CANCER RESEARCH

Thj@ One

II//III/I/IIIIIIII!IIllhIIllhIII@I@Ii@Ii@I@I11111EGBE.-KA2-5g4Q

Research. on December 11, 2020. © 1963 American Association for Cancercancerres.aacrjournals.org Downloaded from

Cancer Research1106 Vol. @3,September 1963

The influence of hormones on the growth ofcertain tumors is common knowledge, and thereis also evidence that hormones are implicated inthe origin of human cancer. Adrenal corticosteroidsappear to enhance metastasis of cancer (58) . Sexsteroids influence the growth of cancer of theprostate (71), breast (103, 107), and uterus (77),Sex is related (possibly by way of hormones) to theoccurrence of melanoma, leukemia, lymphoma,and cancer of the gastrointestinal tract.

A carcinogenic influence of ovarian steroids issuggested by several clinical observations. Observations by Corscaden (@) suggested that evidenceof high endogenous estrogen during menopausewas related to the occurrence of endometrial cancer. Gusberg (56) has reported an association between administration of estrogens and incidenceof endometrial carcinoma in post-menopausalwomen. The sustained estrogenic stimulationwhich occurs in patients with polycystic ovariesis sometimes accompanied by adenomatous hyperplasia of the endometrium or carcinoma in situ,and endometrial carcinoma has been reported atan unusually high frequency and early age in suchpatients (7@, 1@5). Women whose ovarian function was curtailed by surgical castration or irradiation have a slightly lower incidence of breast cancer than the general population (65).

In mice, sustained high-level stimulation bypituitary gonadotropins appears to be the cause ofadenocarcinoma in ovaries which are transplantedinto the spleen (15) and in the adrenals of ovariectomized mice (144).

Genetic factors also influence cancer genesis.Although most epidemiologic studies fail to showany familial tendency in human cancer and mostexamples of “cancer families― are readily explained as pure chance (100), a genetic influenceon human cancer incidence has been shown incomparative studies of cancer concordance insiblings (85) and in monozygous and dizygoustwins (60) and by utilization of genetic markerssuch as blood groups (6@). Leukemia is nearlytwice as frequent in boys as in girls, even in theyoungest age groups when hormonal differences areminimal (1@9), suggesting that the sex chromosomes might carry genes which influence leukemogenesis. Recent demonstrations of chromosomalaberrations in chronic myeloid and acute leukemia(61, 99, 118) indicated abnormalities at the cytogenetic level which are either the cause or result ofcancer.

There is little evidence that mechanical factorsplay any causal role in human cancer, but theycannot be completely dismissed. The relation oftrauma to cancer is a constantly recurring medico

legal problem. Certainly most if not all clinical reports of trauma being followed by tumors can bebetter interpreted as coincidental trauma whichdirects attention to an existing but previously unnoticed lesion (5, 134) . An occasional case, however, cannot be so readily dismissed (10@2),and itis possible that trauma is a factor in the greaterincidence of cervix cancer in multiparous patients(149). The frequentassociationof bladdercancerwith bilharziasis (41) is presumably due tochronic irritation from the mechanical or chemicaleffects of the schistosome. Laboratory studies, tobe discussed below, also provide evidence of a rolefor trauma in carcinogenesis.

Aberrant embryologic development is the cvident cause of such tumors as teratomas, hamartomas, and Wilm's embryonal nephroma. Thesame terminology might be used for choriocarcinoma in the female, although in this example thevictim is of the generation antecedent to thetumor cells. However, the embryonic rest concept(v.3) is a superficial explanation, because it ignoresthe fact that there had to be an initiating cause ofthe abnormality in the embryo. It does not explainwhy the embryonic rest occurred or why it becamemalignant.

I have not yet mentioned the word “virus,―because there is as yet no proof of a viral etiology ofany human cancer. However, suggestive evidenceand unsubstantiated reports (1, 13, @5,@27,8@2,86,1@1, 1@7, 13@, 140, 141) are numerous, and although unconvincing are not necessarily wrong.More important is the fact that viruses cause cancers in many animal species, and present knowledge of the behavior of these oncogenic virusesleaves us with no basis for denial of the possibility(I personally would say the probability) thatviruses could also be etiologic agents of humancancer.

GAPS IN PRESENT KNOWLEDGEIn seeking to understand the etiology of cancer,

we are faced not only with this embarrassingmultiplicity of undeniable cancer causes, but anequally embarrassing awareness that no one ofthese causes is adequate to explain or even to beconsistent with the patterns of cancer occurrencein man. The epidemiology of human cancerpresents many perplexing questions. Do the manydifferent types of cancer have the same etiology,or does each have a different cause? Why don'twe all get cancer? Amongst the victims of atomicirradiation, amongst sailors and farmers exposedto excessive ultraviolet irradiation, amongst workers in the aniline dye and petroleum industrieslarge numbers of individuals must have nearly

Research. on December 11, 2020. © 1963 American Association for Cancercancerres.aacrjournals.org Downloaded from

SOUTHAM—The Complex Etiology of Cancer 1107

identical exposures to recognized carcinogenic influences, and yet (although cancer incidence is in

creased in these populations) it is still only an occasional individual who develops cancer. Why isthere such a long incubation period between cxposure to a recognized carcinogenic stimulus andthe development of overt disease? What happensto the affected cell or the causative agent in thislong interval?

This then is the dilemma—a plethora of causesand an unpredictability of responses. Where do wego from here?

FRESH CONCEPTS CONCERNINGCARCINOGENESIS

Three areas will be considered : (a) the conceptof co-carcinogenesis, (b) new knowledge in virology, and (c) the role of factors intrinsic to the host.

The broad concept of co-carcinogenesis.—Cocarcinogenesis is the phenomenon of an additiveor synergistic effect of two agents, such that thecombination causes cancer in a manner thatneither can do alone. Awareness of this phenomenon is not new, but it deserves re-emphasis because it is a broad concept, applicable to any cornbination of factors which act jointly in carcinogenesis. The familiar examples of co-carcinogenesiswith two chemical agents should be recognized asspecial examples of a general phenomenon. Fundamental observations were published by Shear(1@) and by Rous and Kidd in 1938 (11@2,113).They were confirmed and emphasized by Berenblum (9) and Shubik (192), and subsequently bynumerous investigators.

A classic example of co-carcinogenesis is primary application of methylcholanthrene (at a dosewhich causes few or no tumors) to mouse orrabbit skin, followed by applications of croton oil,which if given alone causes no tumors. The combined treatment leads to an early and high incidence of skin cancer. Berenbium (10, 11) conceivesof an essentially instantaneous and irreversibleinitiation of a change in the affected cell by themethyicholanthrene (or other initiator) such thatthenceforth this change, although morphologicallyundetectable, can readily be promoted to the fullyneoplastic state by the croton oil (or other promotor). Although many investigators do not accept completely the concept of stepwise initiationand promotion or the all-or-none nature of theinitiation step, the fact of a co-carcinogenic synergism of various chemical agents is irrefutable.

Chemical agents other than the traditional carcinogens may influence carcinogenesis. Horton(68) has shown that the choice of solvents for carcinogens greatly influences the results. Different

solvents, although having no independent carcinogenic effects, will variously accelerate the effect ofthe recognized carcinogen.

Estrogenic steroids have a co-carcinogenic effectwith certain carcinogenic hydrocarbons (47, 81).A possible mechanism is suggested by the observation that estrogens increase the pick-up of particulate matter by cells in tissue culture (80). In contrast, Leiter and Shear (83) and Homberger andTregier (66) found female mice less susceptiblethan males to induction of sarcomas by dibenzpyrene, but it is possible that these results weredue to sex-linked genetic effects rather than hormonal effects, since the same differences werefound between castrated male and female mice(66). Cortisonehasbeenvariouslyreportedasenhancing (135) and diminishing (7, 98) tumorgenesis when administered in conjunction withchemical carcinogens.

Urethan augments the leukemogenic effects ofx-ray or estrogens or methylcholanthrene in mice(75, 76).

Menkin reported an augmentation of methylcholanthrene carcinogenesis by certain growth factors found in inflammatous exudates (95).

Gross (53) and Kaplan (73) have demonstratedthat x-radiation can activate leukemogenic virusesin mice.

Co-carcinogenic action of chemicals plus viruseshas been observed in several experimental systems.The first, and still the most convincing, experiments were those of Rous who, with several coworkers (109, 111—13),showed that the combination of Shope papilloma virus plus tar or methylcholanthrene in various sequences acceleratedpapilloma formation in rabbits and increased thetendency of these papillomas to transform to carcinoma. Rowson et cii. (116) observed enhancementof tumor incidence in mice given both polyomavirus and chemical carcinogens. F. Duran-Reynals(33) for many years championed the concept thatlatent infections of presumably nononcogenicviruses such as fowl pox viruses in mice and rabbitsand chickens increased the carcinogenic effect ofmethyicholanthrene. Although his experimentalresults were often consistent with this hypothesis,they were not generally accepted as conclusive.M. L. Duran-Reynals has recently published impressive evidence (34) that, if methylcholanthreneis applied to mice during primary infection byvaccinia virus, tumor incidence is increased. Confirmatory studies were reported by Martin (90),who also reported similar results with poliovirus.Wiseley and co-workers (147) demonstrated thatmice subjected to artificial smog plus intranasaladministration of myxo viruses (influenza and

Research. on December 11, 2020. © 1963 American Association for Cancercancerres.aacrjournals.org Downloaded from

1108 Cancer Research Vol. 923, September 1963

Newcastle disease) had a greater frequency ofbronchial adenomas and carcinomas than micesubjected to either smog or virus alone. Tanakaand Southam observed an earlier appearance andgreater incidence of skin papillomas in mice infected with West Nile virus or Herpes simplexvirus (139) concurrently with methylcholanthreneor benzpyrene painting than occurred in animalsreceiving the chemical carcinogens alone.

Mazurenko (91) reported that the administration of vaccinia virus and cortisone to mice was followed frequently by the development of leukemiaand that the leukemia was thenceforth transmissible by a leukemia virus distinctly different

CHART 1.—Schematic representation of the possible courses

of virus infection in the human or animal host.

from vaccinia virus. This suggests a joint effect oftwo viruses, only one of which is known to haveoncogenic capacity. Wyburn-Mason (148) has suggested on the basis of clinical observations thathuman cancer of the lip might be related to Herpessimplex infection.

New concepLi in virology.—Since virology is currently an area of great interest and research activity and is the most controversial, at least insofaras human carcinogenesis is concerned, it is pertinent to examine recent experimental facts and resulting theoretical considerations which may bearon the problem of carcinogenesis by viruses.

It is a common misconception that virus infection means disease. Obviously, much disease iscaused by viruses, but infection without disease ismuch more common. The ratio of clinical diseaseto inapparent infection is less than 1: 100—probably 1: 1000—for poliovirus (93), Japanese B encephalitis virus (1920, 1926),and St. Louis encephalitis virus (19) in man, and many viruses which frequently infect man are not known ever to causedisease (94). This is true for many of the ECHOviruses which were dubbed “orphanviruses― because they had no disease to call their own.

There is a further erroneous assumption that,if one escapes serious illness during a virus infection, that is the end of the infection. How often

this is true we do not know; but it is not necessarily true, and we may learn that persistent (although asymptomatic) infection is the rule ratherthan a rarity. Chart 1 is a diagrammatic representation of the possible courses of virus infection.The possibilities of inapparent infections, persistent infections with or without eventual excacerbations, and prolonged incubation time (fromentry of virus to appearance of disease) should benoted. The well known vagaries of Herpes simplexand viral hepatitis in man illustrate these variouscourses of infection. It will be recognized that theoccurrence of late exacerbations necessarily meansthat some alteration must occur in the relationship between host and parasite. Such alterationsmight result from changes intrinsic to the parasiteor the host, or might be caused by extrinsic influences.

Recognition that inapparent infections are commonplace raises the possibility that the range of

CHART @.—Schematic representation of possible courses of

virus infection of a cell. Small black dots indicate viral nucleicacid, and surrounding circle indicates viral protein antigen.(Many possible relationships—including intranuclear virus,unassembled viral components, additional viral componentssuch as lipides, and induced biochemical changes in cell protoplasm—are omitted for simplicity.)

susceptibility of different species to viruses may bemuch broader than we now believe, because almost all virus studies are based on an interest inthe disease as it occurs naturally or as it appears inan experimental system. Little work has been addressed to the question of nonpathogenic infections in man or experimental animals.

If we consider virus infection at the level of thecell, we also find phenomena which are influencing

•0

Research. on December 11, 2020. © 1963 American Association for Cancercancerres.aacrjournals.org Downloaded from

SOUTHAM—The Complex Etiology of Cancer 1109

our thinking about viruses and cancer. Chart 92is adiagrammatic representation of some virus-cellrelationships. This illustrates that cell destructionis not an inevitable consequence of cell infection,that cell-to-cell transmission may occur withoutexposure of the virus to extracellular fluids (andtherefore without exposure to serum antibody),that virus may exist as a relatively simple nucleicacid lacking its antigen protein coat or as cornplete (antigen-coated) virus. If virus propagatesin the incomplete (nucleic acid) form, it wouldneither elicit nor be affected by antibody.

The range of viral infectibility is determined inpart by compatibility between receptors on cellmembranes and on the protein coat of the virus.Alterations in either may alter what is otherwisean absolute specificity. Thus poliovirus nucleic acidwill infect rat and chicken cells with propagationof complete virus particles (105), although cornplete poliovirus can infect only human and simiancells. Bacterial cells devoid of cell walls (protoplasts) are susceptible to infection by phages towhich the unaltered cells are resistant (130). Lessdramatic but basically similar is the heightenedsusceptibility to infection of mammalian cellstreated with hypertonic salt (3, 40). At presentthese are laboratory tricks, but they suggest thatviruses have versatility which will further stretchour concepts of host specificity.

Vertical transmission (592), the transmission ofvirus from one generation to another, in contrastto horizontal transmission from neighbor to neighbor, is the usual manner of transmission of animaloncogenic viruses such as Gross's leukemia, polyoma, Bittner's milk agent, and lymphomatosis offowl. Vertical transmission may occur throughthe mother's milk (or yolk of chickens) or throughthe placenta, and can probably also occur throughintracellular infection of the germinal cell—eitherovum or sperm (16).

The disappearance of virus after it has infectedcells and produced a persistent change has beennoted in many experimental situations and can bevariously explained. A classic example is the apparent disappearance of Shope papilloma virusfrom the papillomas which it induces in domesticrabbits. Syverton (138) used the term “masking―for this phenomenon to express the concept thatthe virus was still present in a disguised form.Whether the virus is masked or has truly disappeared is still unresolved. Sarcomas can be induced in turkeys by Rous virus, but the virus cannot be demonstrated in such tumors by the usualmethods of chick or egg inoculation. Recentstudies by Bergs and Groupé (14) have demonstrated that this apparent absence of virus is due

to its neutralization in tissue homogenates byantibody or other serum factors. If the cells weregrown for several days in antibody-free tissue calture, the virus became detectable. Another disappearing act is the phenomenon of lysogeny inphage. In this situation the virus is present in thebacterial cells at all times but is in a chemicallyand morphologically incomplete form and doesnot cause destruction of the bacterial host cell. Itcan, however, be made to reveal itself by subjecting the cells to various conditions that upset thebalanced growth of virus and cell (63). Recentlydisappearance of infective virus or even of extractable infective nucleic acid has been reportedin hamster cells which had become neoplastic because of infection by polyoma virus (146). The evidence to date is consistent with the suggestion thatthe virus has truly induced a somatic mutationsuch that no further influence of the virus is required to maintain the cell in its neoplastic state.The possibility still remains, however, that eventhis virus persists undetected within transformedcells. A new twist on the relation of virus to tumoris the report by Pitelka (104) that virus-like partides are readily visualized by the electron microscope in certain breast tumors arising in inbredlines of mice which are apparently free of theBittner milk factor. This may be yet another example of an altered form of a virus which nevertheless has a continuing causal relationship to theneoplastic process, although the possibility thatthis is merely a contaminating virus cannot be dismissed.

Many different types of tumor may be causedby a single kind of virus. Polyoma virus in mice isthe most startling example—capable of causingmore than a score of different histologic forms oftumors in mice (1392), as well as causing tumors inrats and hamsters (38, 39) . Gross has made thefascinating observation that, whereas inoculationof mice with his leukemia virus ordinarily causesleukemia of the lymphoid type arising from cells ofthe thymus gland, if the gland is removed frominfected animals at an early age lymphatic leukemia does not develop, but at a later datemyelogenous leukemia arises from bone marrowcells (54). This virus also infects and causes leukemia in rats (55). The Luckévirus is well knownas a cause of kidney adenocarcinoma in frogs(35, 84) but on transfer to salamanders it reportedly caused rhabdomyosarcoma—a tumor of skeletalmuscle (110). Even the well known Rous sarcomavirus has multiple manifestations. It causes a nonneoplastic hemorrhagic disease in young chicksand chick embryos (31), and recent reports fromRussia (136, 151), Sweden (92), and Czecho

Research. on December 11, 2020. © 1963 American Association for Cancercancerres.aacrjournals.org Downloaded from

Cancer Research Vol. 23, September 19631110

slovakia (137) and this laboratory (97) indicatethat it can also cause sarcomas and cystic tumorsin rats. A virus, SV 40, discovered in tissue culturesof monkey kidney cells has no presently recognizedpathogenic effect in monkeys but causes cancersin hamsters (37). Trentin has reported thatAdenovirus Type 192,previously known only as aprobable cause of mild or asymptomatic respiratory infections in humans, causes pulmonary andsubcutaneous cancers in hamsters (143).

Viruses not only have a native potentiality fordiverse pathogenic effects but also are themselvessubject to mutation. The earliest reports were byMundry and Gierer (96) demonstrating mutationof tobacco mosaic virus following exposure tonitrous acid. More recently mutations have beeninduced in animal viruses (17, 50). Dulbecco (929)has reported spontaneous mutation of polyomavirus during growth in hamster cell tissue cultures.The transformation phenomenon of Avery (6)—the alteration of genetic characteristics of an organism by treatment with nucleic acid from adifferent organism—has been reported for viruses.Mixtures of live myxoma virus and noninfectiousnucleic acid extracts of fibroma virus resulted infibromas rather than myxomas when introducedinto rabbits (78). Transformation resulting in hybrid viruses has been reported to occur when mixtures of two strains of influenza virus were cultivated together (921).

Host factors in carcinogenesis.—Thehost rarelyif ever plays a passive role in disease processes. Incarcinogenesis and tumor pathogenesis there areseveral situations in which an influence of immunologic and other host factors is recognized orsuspected.

The concept of specific immune tolerance (920,59, 992) has implications for oncology. The current concept is that an animal during the prenatalor neonatal period has no discrimination forforeignness but accepts anything present withinits body during this period as being normal or“self.―Recognition of things as foreign or “notself―is achieved shortly after birth. Thus, if a neoplastic mutation should occur in, or if a virusshould gain entry into, an animal during this“tolerogenic―period, it would not be recognizedas foreign. No cellular reactions would occur. Noantibodies would be formed. This concept raisesan intriguing question as to the outcome of infection in a specifically tolerant host. If a virus wereaccepted as “self,―would unimpeded viral propagation lead to great pathogenesis, or might thehost somehow be so adapted to the virus that noharm would ensue? The latter is clearly the casewith lymphocytic choriomeningitis virus in mice

(69), whereneonatal infection is persistentbutnonpathogenic, whereas infection initiated a fewdays later excites a host reaction and ends ineither elimination of the virus or (more often)death of the host. A similar phenomenon has beendemonstrated by Rubin (117) for chicken lymphomatosis virus. If these two examples shouldbe representative of a common phenomenon, it isclear that undetected latent virus infection mightbe well nigh universal and without importanceuntil some new factor intervened to upset thebalance and lead to disease.

The possible relationship of tolerance to cancergenesis is not restricted to virus infections. Since,according to present concepts, the development ofan abnormal cell during fetal life would not berecognized and resisted, such a cell would persistand multiply. This would account for the persistence of teratomas and other embryonic rests (a!though additional mechanisms must be postulatedfor the initial deviation of embryonic developmentand the later progression into neoplasia) . In ananalogous manner, other states of diminished immuno!ogic reactivity might allow undesirable cellswhich normally would be resisted to propagatewithout hindrance. Since carcinogens may depressreticuloendothelial system functions, it seems possible that even such classical carcinogens asmethy!cholanthrene might act in part by depressing host resistance to the growth of spontaneousmutants, rather than solely by direct mutagenicaction on the target cell. Although dose and timefactors for methylcholanthrene carcinogenesismake this seem unlikely, the concept cannot bearbitrarily dismissed and is particularly appealingas a possible mechanism in carcinogenesis by polyfunctional alkylating agents (18, 64).

Maini et at. (87, 88) observed that abnormalmitoses were common in regenerating rat liverpost-hepatectomy and that concurrent treatmentwith the carcinogen 3'-Me-DAB favored propagation of the abnormal clones. They suggested thatthis was probably attributable, at least in part, toan effect on host factors rather than to a direct effect on the hepatic cells.

Extension of these immunologic considerationshas led several authors (924,51, 74, 145) to proposethat immunologic reactions play a more direct rolein carcinogenesis. According to this hypothesis,through mutation of unspecified cause there occurs in certain cells a genetic loss which is equatedto an antigenic loss. This lost antigen is nowforeign to the affected cell, which consequentlyrecognizes the normal cells of the body as beingforeign because they contain the antigen which theabnormal cells now lack. This recognition of

Research. on December 11, 2020. © 1963 American Association for Cancercancerres.aacrjournals.org Downloaded from

SOUTHAM—The Complex Etiology of Cancer ‘In

foreignness leads the antigen-deficient cells to react in the way that normal lymphoid tissue reactswhen faced with a foreign antigen—that is, byhyperplasia and production of antibody-like factors that have deleterious effects against the normal cells, in the same manner as is postulated tooccur in anto-immune diseases. These two features, toxic effect on normal tissues and excessivegrowth of one abnormal cell type, thus resemblethe picture of cancer. This hypothesis has a certhin plausibility if the affected cell is of the typewhich is capable of immune responses—i.e., a cellof the reticuleoendothelial system—and thus it dcserves consideration and investigation as an explanation for such diseases as leukemia, lymphoma, Hodgkin's disease, and multiple myeloma.However, it seems inconceivable that this conceptcould be extended to neoplasms of epithelial cellsor neoplasms of most connective tissue cells fromwhich arise the great preponderance of carcinomasand sarcomas.

Cortisone and related adrenal steroids inhibitthe host defense mechanism. They also affect carcinogenesis and tumor growth (58, 98), possiblybecause of their effects on specific and nonspecificimmunologic reactions.

Other studies suggest that such nonspecific hostfactors as fibroblasts and wandering cells may influence carcinogenesis. Orr (101) has presentedevidence that the connective tissue stroma subjacent to areas of carcinogen application plays apart in the development of epidermal carcinomas.Subcutaneous injection of diatomaceous earth inchickens and skin incisions in mice (procedureswhich stimulate connective tissue growth) increasesusceptibility to Rous sarcoma virus (114) andchemical carcinogenesis (926,49, 1923),respectively.Clinical observations and some laboratory studies(43, 48, 1928) suggest that healing wounds offer abetter site for tumor growth than normal tissues,and that some tissues or organs provide a “bettersoil―for the implantation of tumor metastases.Even the more neglected intercellular cement mayinfluence carcinogenesis and pathogenesis, sincevariations in the physical state of this substanceinfluence the diffusion of viruses and chemicals(392) and probably metastases of transplantedtumors (44, 89).

THE CONCEPT OF A COMPLEXETIOLOGY

From all the foregoing one is led to the conceptthat in carcinogenesis we are faced not simply withmultiple etiologies but rather with complex etiologies; that there are not merely a score of diverseagents, each capable of causing cancer, but a num

her of agents and accessory factors which if theychance to act in appropriate combination or sequence may lead to cancer. The mechanisms bywhich such combinations of agents enhance cancerincidence might be various and unrelated. Severalpossibilities seem reasonable : (a) Penetration oflocally acting carcinogenic chemicals into tissue orinto cells might be increased by the action of so!vents, irritants, or viruses. (b) Altered metabolismor defective detoxification of systemically actingchemical carcinogens might result from geneticdefects or hepatic or renal damage by chemicalsor viruses. These two possibilities visualize thechemical carcinogen as the prime etiologic agentand all other influences as co-carcinogenic factors.(c) Infectivity or pathogenicity of a virus mightbe enhanced or altered by changes in the penetrability or metabolism of the infected cells or bychanges of immunologic reactivity of the host induced by chemicals, hormones, irradiation, or irritation. (d) Local or systemic alterations in immunological or hormonal conditions might favorsurvival and growth of mutant cells that would diein a “normal―tissue environment. These two possibilities visualize a virus or mutagenic force suchas x-ray as the primary agent of carcinogenesiswhile hydrocarbons, hormones, and cytotoxicagents would be accessory factors. (e) A series ofalterations of intrinsic cellular characteristicsmight be caused successively by the various carcinogenic influences, each alteration approachingbut no one step being by itself the complete neoplastic transformation. This is in essence theclassical concept of co-carcinogenesis as championed by Berenblum but is not restricted tochemical carcinogens. It visualizes each contribution to the total change as necessary, and hence noone cause can be singled out as the primary agent.

Thus, depending on which of the above is mostnearly true, the relative importance of individualfactors in the fundamental mechanisms of carcinogenesis would differ vastly. However, in anotherand more practical sense, each factor, no matterwhat its mode of action, would be of major importance because its absence would diminish theprobability that a neoplastic cell would arise andgrow into clinical cancer. Any method of avoidingor counteracting any of these factors would be efficacious prophylaxis against cancer.

This concept of a complex etiology of cancerseems more compatible with the epidemiology ofhuman cancer than any concept of singly actingetiologic agents, because if a final result requiresthe action of multiple factors which are acting withvarious frequencies in a population, then the frequency with which the total etiologic require

Research. on December 11, 2020. © 1963 American Association for Cancercancerres.aacrjournals.org Downloaded from

11192 Cancer Research Vol. 23, September 1963

ments will be met is the product of their individualfrequencies. To make a purely imaginary example,let us suppose that one-tenth of a given populationhas influenza at some time during their lives, thatone-tenth of the individuals are cigarette smokers,and that one-tenth have a genetic susceptibility tocancer. If we then assume that each of these threeconditions operates as an etiologic factor in cancerof the lung, but that no two factors alone are sufficient to bring on the disease, then the incidenceof lung cancer would be 1 : 1000. If any two of thesefactors were sufficient to cause lung cancer, the incidence would be 1 : 100, and if still a fourth factorwith the same frequency of exposure were required,then the incidence would be only 1 : 10,000. Thisconcept allows for the relative infrequency ofspecific types of cancer in a population that has ahigh frequency of exposure to presumed causativefactors and the fact that epidemiologic studies ofcancer etiology often reveal suggestive correlationsbut seldom provide convincing evidence of causalrelationship.

CONCLUSIONIn conclusion, it seems worth while to consider

the hypothesis that most cancer is due not to anysingle causative agent but rather to a complex ofmultiple factors each of which contributes something to the total process. These several factorsmay act simultaneously or sequentially; continuously, repeatedly, or rarely. They may persistor disappear from the affected individual or cell.They may be additive or synergistic or even antagonistic. They may cause persistent or transienteffects They may act directly on the potentiallyneoplastic cell or indirectly by affecting othertissues of the host. They may bring about a biochemical change which is characteristic and fundamental to all neoplastic cells or may cause changeswhich are only superficially similar.

Such a concept adds nothing to our understanding of the fundamental intracellular change whichinitiates the neoplastic process. (For discussions ofthis area see references [67, 1311.) It is annoyingbecause it promises no easy solution. Nevertheless,it merits attention because it fits the facts of cancer incidence better than any concept of a singleetiology, and it is appealing because it offers thepossibility that attack upon or removal of any oneof the postulated multiple factors might significantly reduce cancer incidence.

REFERENCES1. AGEENKO, A. I. Leukaemias Induced in Rats by Extracts

from Human Leukaemic Tissue. Neoplasma, 8:131-43,1961.

@. AHLSTROM, C. G., and JONSSON, N. Induction of Sarcoma

in Rats by a Variant of Roes Virus. Acta PathoL MicrobioLScand., 54: 145—fl,196g.

S. ALEXANDER,H. E. ; KocH, G. ; MOUNTAIN, I. M. ; andVAN D@aass, 0. Infectivity of RNA from Poliovirus in

Human Cell Monolayers. J. Exp. Med., 108:493-506,1958.

4. AUEEBACH, 0.; S@roirr, A. P.; HAMMOND,E. C.; and G@aFINKEL, L Changes in Bronchial Epithelium in Relation

to Cigarette Smoking and in Relation to Lung Cancer.New Eng. J. Med., 265:9.53-67, 1961.

5. Auwrm@,L S. The Role of Trauma in Oncogenesis: AJudicial Consideration. J.A.M.A., 175:946—50, 1961.

6. Avsmy, 0. T.; MACLEOD, C. M.; and McC@uiTY, M.Studiesofthe ChemicalNature ofthe Substance InducingTransformation of Pneumococcal Types. Induction ofTransformation by a Desoxyribonucleic Acid FractionIsolated from Pneumococcus Type ifi. J. Exp. Med.,79:137-57, 1944.

7. BASERGA, R., and Sinmix, P. The Action of Cortisone onTransplanted and Induced Tumors in Mice. Cancer Res.,14:1@—16,1954.

8. BASERGA, R.; Yoxoo, H.; and HENEGAR, G. C. Thoratrast-induced Cancer in Man. Cancer, 13: 10@1-S1, 1960.

9. Bnax@reLuas, I. The Mechanism of Carcinogenesis. AStudy of the Significanceof Cocarcinogenesisand RelatedPhenomena. Cancer lies., 1:807-14, 1941.

10. . Carcinogenesis and Tumor Pathogenesis. Adv.Cancer Res., 2:1@9—75,1954.

11. . Circumstantial Evidence Pointing to Differencesbetween Cancers in Terms of Etiologic Factors. CancerRes, 16:675—80,1956.

1@. BERENBLtTM, I., and Siiuimc, P. The Role of Croton OilApplications AssOCiated with a Single Painting of a Carcinogen in Tumour Induction of the Mouse's Skin. Brit.J. Cancer, 1:379—82, 1947.

13. BERGOIJrS, V. M. Experimental Studies of the Etiology ofLeukaemia in Men. A Review. Neoplasma, 5:337-47,1958.

14. BERGS,V. V., and GROUP@,V. Critical Factors in Successful Recovery of Roes Sarcoma Virus from Turkey Tumors. Science, 134:19@, 1961.

15. Biaxuro, M. S., and Brsic.n@, G. S. Development ofTumors in the Rat Ovary after Transplanting into theSpleen. Proc. Soc. Exp. BioL Med., 55: 176—80,1944.

16. BITTNER,J. J. Influence of the Mammary-Tumor Agentin the Genesis of Mammary Cancer in Agent-free Miceafter Male Transmission. J. NatL Cancer Inst., 25:177—

@00,1960.17. BOEY@, A. Induction of a Mutation in Poliovirus by

Nitrous Acid. Virology, 9:691—700, 1959.18. BOYLAND,E., and HORNING,E. S. The Induction of

Tumors with Nitrogen Mustards. Brit. J. Cancer, 3:118-@S,1949.

19. BRODY,J. A.; Bunxs, K. F.; BROWNING,G.; and Sca&'r'rNER, J. D. Apparent and Inapparent Attack Rates for St.

Louis Encephalitis in a Selected Population. New Eng. J.Med., 261:644—47,1959.

@0.Buiusrr, F. M. The New Approach to Immunology. NewEng. J. Med., 264:@4—34,1961.

@1.Bunsiur, F. M., and Lnm, P. E. A Genetic Approach toVariation in Influenza Viruses. 3. Recombination ofCharacters in Influenza Virus Strains Used in Mixed Infections. J. Gen. Microbiol., 5:59, 1951.

%@.CORSCADEN, J. A., and GUSBERG, S. B. The Backgroundof Cancer of the Corpus. Am. J. Obst. GynecoL,53:419—31, 1947.

@S.COWDRY,E. V. Cancer Cells. Philadelphia: Saunders,1955.

@4.DA@sdHEK, W., and SCHWARTZ,R. S. Leukemia andAutoimmunization—Some Possible Relationships. Blood,14:1151—58,1959.

Research. on December 11, 2020. © 1963 American Association for Cancercancerres.aacrjournals.org Downloaded from

SOUTHAM—The Complex Etiology of Cancer 1113

@5.DE CARVALHO, S. Leukemogenic and CarcinogenicActivity of Human Leukemic and Tumoral RNA in Mice.Proc. Am. Assoc. Cancer Res., 3: 104, 1960.

@6.DEELMAN, H. T. Die Entstehung des experimentellenTeerkrebes mid die Bedeutung der Zellenregeneration. Z.Krebsforsch., 21:@0—@6, 19@4.

@7.DMOCHOWSKI, L. Viruses and Tumors in the Light ofElectron Microscope Studies: A Review. Cancer Res., 20:

977—1015,1960.@8.Dou@, R., and Hiu@, A. B. Lung Cancer and Other Causes

of Death in Relation to Smoking. Brit. M. J., 2:1071—81,1956.

@9.DULBECCO, R. Viral Carcinogenesis. Cancer Res., 21:975—80,1961.

30. DUNGAL, N. The Special Problem of Stomach Cancer inIceland. J.A.M.A., 178:789—98, 1961.

31. DURAN-REYNALa, F. A Hemorrhagic Disease Occurringin Chicks Inoculated with the Roes and Fujinami Viruses.Yale J. Biol. Med., 13:77—98, 1940.

3@. . Tissue Permeability and the Spreading Factors inInfection. A Contribution to the Host: Parasite Problem.Bact. Rev., 6: 197—@5g,194g.

33. . Studies on the Combined Effects of ChemicalCarcinogens, Hormones, and Virus Infection. Texas Rep.BioL Med., 15:306—33, 1957.

34. DURAN-REYNALS, M. L., and STANLEY, B. VacciniaDermal Infection and Methylcholanthrene in Cortisone

treated Mice. Science, 134: 1984—85,1961.35. DURYEE, W. R. The Mechanism of Virus-induced Trans

formations in Tubules of the Frog Kidney. Acts UnioInternat. contra Cancrum, 15:587—94, 1959.

36. DUSHANE, G. Risk and Benefit. (Editorial.) Science,131:1579, 1960.

37. EDDY, B. E.; BORMAN, G. S.; GRUBBS, G. E.; and You@o,R. D. Identification of the Oncogenic Substance inRhesus Monkey Kidney Cell Cultures as Simian Virus 40.Virology, 17:65—75, 196fl.

38. EDDY, B. E.; STEWART, S. E.; STANTON, M. F.; and MaR

COTTE, J. M. Induction of Tumors in Rat by Tissue Cul

ture Preparations of SE Polyoma Virus. J. Natl. CancerInst.,22:161—7fl,1959.

39. EDDY, B. E. ; STEWART,S. E. ; YOUNG, R. ; and Mmun, B.Neoplasms in Hamsters Induced by Mouse Tumor AgentPassed in Tissue Culture. J. NatJ. Cancer Inst., 20:747—6@, 1958.

40. Eu.ras, K. A. 0., and COLTER, J. S. The Interaction ofInfectious RNA with a Mammalian Cell Line. I. Relationship between the Osmotic Pressure of the Medium and theProduction of Infectious Centers. Virology, 11:484-43,1960.

41. FERGUSON, A. R. Associated Bitharziosis and PrimaryMalignant Disease of the Urinary Bladder, and Observationson a Seriesof40 Cases.J.Pathol.Bacteriol.,16:76—94, 1911.

4@. FERNAu, A.,and SMERECKER, H. tYberdas Verbleibenradioaktiver Substanz in Organismus bei RadiumEmanationstrinkkuren. Strahientherapie, 46:365—73,193g.

43. FISHER,B.,and FISHER,E. R. ExperimentalStudiesofFactorsInfluencingHepatic Metastases.IlL EffectofSurgical Trauma with Special Reference to Liver Injury.Ann. Surg., 150:731—44, 1959.

44. GAsic, G.; LOEBEL, F.; and BADINEz, 0. Cementing Substances in Metastasizing and Non-metastasizing Transplantable Tumours in Mice. Nature, 185:864—65, 1960.

45. GKrmEII, A. 0., and NoRms, C. Poisoning from DrinkingRadium Water. J.A.M.A., 100:400-40g, 1933.

46. GILMAN,J. P. W., and RUCKERBAUER,G. W. Metal Carcinogenesis. I. Observations on the Carcinogenicity of a

Refinery Dust, Cobalt Oxide and Colloidal Thorium Dioxide. Cancer Res., 22: 15@—57,196g.

47. GILMoui@, M. D. Investigation into the Influence ofEstrone on the Growth and Genesis of Malignant Cells.J. Pathol. Bacteriol., 45: 179—87,1937.

48. GouIE, H., and Fm.ix, M. D. Growth Characteristics ofFree Tumor Cells Transferred Serially in the PeritonealFluid of the Mouse. Cancer Res., 11:73—80, 1951.

49. GOTFFRIED,B.; Moi@carui@,N.; and PAm, J. Effect ofRepeated Surgical Trauma on Chemical Carcinogenesis.Cancer Res., 21:658—60, 1961.

50. GRANOFF,A. Induction of Newcastle Disease Virus Mutants with Nitrous Acid. Virology, 13:40@—8,1961.

51. GREEN, H. N. Immunological Aspects of Cancer. In:Cancer, 3: 1-41. London: Butterworth & Co., 1958.

@g•GROSS, L. Viral Etiology of “Spontaneous―Mouse Leukemia: A Review. Cancer Res., 18:871—81, 1958.

53. . Serial Cell-free Passage of a Radiation Activated Mouse Leukemia Agent. Proc. Soc. Exp. BioLMed., 100:10@Z—5,1959.

54. . Development of Myeloid Leukemia in Thymectomized Mice Following Inoculation of LymphaticLeukemia Virus. Proc. Am. Assoc. Cancer Res., 3:115,1960.

55. . Induction of Leukemia in Rats with Mouse Loukemia (Passage A) Virus. Proc. Soc. Exp. Biol. Med.,106:890—93, 1961.

56. GUSBERG, S. B. Precursors of Corpus Carcinoma. Estrogens and Adenomatous Hyperplasia. Am. J. Obst.GynecoL, 54:965—@7,1947.

57. HAMMOND,E. C., and Homi, D. Smoking and DeathRates—Report on 44 Months of Follow-up of 187,783Men. J.A.M.A., 166:1g94—1308, 1958.

58. HAnTs&ANN, W. H., and SRmu@ocx, P. GastroduodenalMetastases from Carcinoma of the Breast. An AdrenalSteroid-induced Phenomenon. Cancer, 14:436—31, 1961.

59. HASEK, M.; LENGEROVA,L.; and HRABA, T. Transplantation Immunity and Tolerance. Adv. ImmunoL, 1:1-66,196g.

60. HAUGE, M., and H@&nv@w,B. Malignant Growth inTwins. Acta Genet., 11:37g—78,1961.

61. HAUSCHKA, T. S. The Chromosomes in Ontogeny andOncogeny. Cancer Res., 21:957—74, 1961.

6g. HELMBOLD, V. W. Sammelstatistik zur Prufung aufKorrelationen zwischen dem weiblichen Genitalcarcinommid dem ABO- mid Rhesus-System. Acta Genet., 11:29—51, 1961.

63. HERSHEY,A. D. Bacteriophages as Genetic and Biochemical Systems. Adv. Virol., 4:25—62, 1957.

64. HESTON, W. E. Occurrence of Tumors in Mice InjectedSubcutaneously with Sulfur Mustard and Nitrogen., J.Nati. Cancer Inst. 14:131—40, 1953.

65. HIRAYAMA,T., and Wrnuza, E. A Study of Epidemiologyof Cancer of the Breast. II. The Influence of Hysterectomy. Cancer, 15:28—38, 1962.

66. H0MBURGER, F., and TREGIER, A. Modifying Factors in

Carcinogenesis.ProgressExp. Tumor lIes.,1:311—28,1960.

67. HORSFALL,F. L., JR. Heritance of Acquired Characters.Science, 136:472—76, 1962.

68. HORTON, A. W.; DENas@tN,D. T.; and TRosSlrr,R. P.Carcinogenesis of the Skin. II. The Accelerating Properties of Aliphatic and Related Hydrocarbons. Cancer Res.,17:758—66, 1957.

69. HoTcIm'@, J., and WEIGAND, H. Studies of LymphocyticChoriomeningitis in Mice. L The Relationship betweenAge at Inoculation and Outcome of Infection. J. ImmunoL, 88:392—400, 1961.

Research. on December 11, 2020. © 1963 American Association for Cancercancerres.aacrjournals.org Downloaded from

Cancer Research Vol. 23, September 19631114

70. HUEPER, W. C. Carcinogens in the Human Environment.A.M.A. Arch Pathol., 71:287-07, 355—80,1961.

71. HUGOINS, C.; STEVENS, R. E., Ja.; and HODGES, C. V.Studies on Prostatic Cancer. IL The Effects of Castrationon Advanced Carcinoma of the Prostate Gland. Arch.Surg., 43:209—23, 1941.

72. JACKSON, R. L., and DOCKERTY, M. B. The SteinLeventhal Syndrome: Analysis of 43 Cases with SpecialReference to Association with Endometrial Carcinoma.Am. J. Obst. Gynecol., 73:161—73, 1957.

73. KAPLAN, H. S. ExperimentalAlteration of Host Responseto an Occult Tumor Virus. Induction of Thymic Lymphosarcoma in X-irradiated Mice. National Cancer InstituteMonograph, 4: 141—50,1960.

74. KAPLAN, H. S., and SssrrHnns, D. W. Autoimmunity inMan and Homologous Disease in Mice in Relation to theMalignant Lymphomas. Lancet, 2: 1—4,1959.

75. KAWAMOTO,S.; IDA, N.; KIRSCHBAUM, A.; and TAYLOR,G. Urethan and Leukemogenesis in Mice. Cancer lIes.,18:725—29, 1958.

76. K@&w@uiovo,S. ; KIRSCHBAUM, A. ; IBANEZ, M. L. ; TRENTIN, J. J. ; and TAYLOR, H. G. Influence of Urethan on theDevelopment of Spontaneous Leukemia and on the Induction of Hemangiomas in the AKR and C58 Strains ofMice. Cancer Res., 21:71—74, 1961.

77. KELLEY, R. M., and B@nn, W. H. ProgestationalAgents in the Treatment of Carcinoma of the Endometrium. New Eng. J. Med., 264:216—22, 1961.

78. KILHAM, L. The Fibroma-myxoma Virus Transformation. Adv. Virus lIes., 7: 103—29,1960.

79. KORPASSY, B. Tannins as Hepatic Carcinogens. ProgressExp. Tumor Res., 2:245—90, 1961.

80. Ko'rIN, P. ; EDMONDSON, H. A. ; and FAi@x, H. L. TheInhibiting Effect of Estrogens on the Development ofCirrhosis and Hepatocellular Carcinoma in Rats on aCholine-deficient Diet. Am. J. Pathol., 32:638, 1957.

81. KOTIN, P. ; Wisms@, D. V.; and GoimoN, I. Productionof Lung Cancer in Mice by Inhalation Exposure to Influenza Virus and Aerosols of Hydrocarbons. ProgressExp. Tumor Res., 3:186, 1963.

82. LACOUR, F. ; LAcoua, J. ; HAREL, J. ; and HUPPERT, J.Transplantable Malignant Tumors in Mice Induced byPreparations Containing Ribonucleic Acid Extractedfrom Human and Mouse Tumors. J. Natl. Cancer Inst.,24:301, 1960.

83. LEITER, J., and SHEAR, M. J. Quantitative Experimentson Production of Subcutaneous Tumors in Strain A Micewith Marginal Doses of 8,4-Benzpyrene. J. Nati. CancerInst., 3:455—77,1948,

84. Lucxi@, B. A Neoplastic Disease of the Kidney of theFrog Rana pipien3. Am. J. Cancer, 20:352—79, 1934.

85. MACRUN, M. T. Inheritance of Cancer in Man. Schweiz.z. Pathol.Bacteriol.,18:463—71,1955.

86. MAGRASSI, F. Etudes sur le problème étiopathogeniquedes leucémieshumaines. Proc. 3d. International Cong.—International Society of Hematology (Cambridge), pp.248—51,1951.

87. MAINI, M. M., and STICH, H. F. Chromosomes of TumorCells. II. Effect of Various Liver Carcinogens on Mitosisof Hepatic Cells. J. NatI. Cancer Inst,, 26:1413—24, 1961.

88. M.UNI, M. M.; STICH, H. F.; and McCuiLocH, E. A.Changes in Cell Populations in Precancerous Lesions ofRats. Proc. Am. Assoc. Cancer lIes., 3:341, 1962.

89. MaISKH, I. N., and KOSLOVA, N. A. The Effect of Antihyaluronidase Serum of Goats upon the Metastases ofBrown-Pearce Carcinoma in Rabbits. Boll. Exp. Biol.Med., 50:101—5, 1960 (Russian text with English summary).

90. MARTIN, C. M.; MAGNUSSON, S.; Goscissisxx, P. J.; and

HANSEN, G. F. Common Human Viruses as CarcinogenicVectors. Science, 134: 1985—86, 1961.

91. MAZURENKO, N. P. Induction of Leukoses in Mice withInfectious Virus, and the Significance of the Latter in theEtiology of Disease. Prob. Oncol., 6:878, 1960.

92. MEDAWAR, P. B. A Discussion of Immunological Toterance. Introductory remarks. Proc. Royal Soc. London,s.B.,146:1—8,1956.

93. MELNICK, J. L. Application of Tissue Culture Methods toEpidemiological Studies of Poliomyelitis. Am. J. Pub.Health, 44:571—80, 1954.

94. . Problems Associated with Viral Identificationand Classification in 1956. Ann. N.Y. Acad. Sci., 67:363—82, 1957.

95. MENKIN, V. Biologic Cocarcinogens. Progress Exp. Tumor Res.,1:279—310,1960.

96. MUNDRY, K. W., and GIERER, A. Die Erzeugung vonMutationen des Tabakmosaikvirus durch chemischeBehandlung seiner Nucleinsaure in Vitro. Z. Vererbungslehre,89:614—30,1958.

97. MUNROE, J. S., and Sou'rH@a, C. M. Studies of Rous Sarcoma Virus in Rats. Federation Proc., 21:453, 1962.

98. N@x@u, T. Influences of Small Doses of Various Cm-ticosteroids on the Incidence of Chemically Induced Subcutaneous Sarcomas in Mice. Cancer Res., 21 :1961.

99. NOWELL, P. C., and HUNGERFORD, D. A. ChromosomeStudies on Normal and Leukemic Leukocytes., J. NatI.Cancer Inst., 25:85—109, 1960.

100. OLIVER, C. P. Genetic Studies of Families with High Cancer Incidence. In: Genetics and Cancer, pp. 426-38.Austin: University of Texas Press, 1959.

101. ORR, J. Stromat Changes during Carcinogenesis. In:Ford Hospital Symposium on Biological Interactions inNormal and Neoplastic Growth, pp. 321—29. Boston:Little, Brown & Co., 1962.

log. PACK, G. T. Tumors of Soft Somatic Tissues. New York:Hoeber, 1958.

103. PEARSON, 0. H.; WERT, C. D.; HOLLANDER, V. P.; andTREVES, N. E. Evaluation of Endocrine Therapy for Advanced Breast Cancer. J.A.M.A., 154:234—39, 1954.

104. PITELKA, D. R.; DEOME, K. B.; and BEEN, H. A.Viruslike Particles in Precancerous Hyperplastic Mammary Tissues of C3H and C311f Mice. J. Natl. Cancer

Inst., 25:753—77, 1960.105. PLOTKIN, S. A.; CARP, R. I.; and GRAHAM, A. S. The

Polioviruses of Man. Ann. N.Y. Acad. Sci., 101:857—89,1962.

106. PoTT, P. Chirurgicat Observations Relative to the Catsmet, the Polypus of the Nose, the Cancer of the Scrotum,

the Different Kinds of Rupture, and the Mortification ofthe Toes and Feet, p. 208. 1st ed. London: Printed byT. J. Carnegie for Hawes, Clarke and Collins, 1775.(“CancerScroti,― pp. 63—68).

107. RAWSON,R. W. The Relationship of Hormonal Environmerit to the Genesis and to the Inhibition of NeoplasticGrowth: Is Cancer Autonomous? Am. J. Obst. Gynecol.,66:999—1014,1953.

108. ROCKSTROH,H. Zur Atiotogie des Bronchiatkrebses inArsenverarbeitenden Nickeihutten. Beitrag zur Syncarcinogenesis des Berufskrelses. Arch. Geschwulstforsch,14:151—62, 1959.

109. ROGERS, S., and Rous, P. Joint Action of Chemical Carcinogen and a Neoplastic Virus to Induce Cancer inRabbits. Results of Exposing Epidermal Cells to a Carcinogenic Hydrocarbon at Time of Infection with ShopePapilloma Virus. J. Exp. Med., 93:459—88, 1951.

110. RosE, S. M., and ROSE, F. C. Tumor Agent Transformations in Amphibia. Cancer Res., 12:1, 1952.

Research. on December 11, 2020. © 1963 American Association for Cancercancerres.aacrjournals.org Downloaded from

SO1JTHAM—The Complex Etiology of Cancer 1115

in Tissue Culture. J. Natl. Cancer Inst.,20: 1223—44,1958.138. STEWART, S. E., and IRWIN, M. L. Cellular Proliferation

in Primary Tissue Cultures Induced with Substance Derived from Cell-free Concentrates from Human Neo

plastic Material. Cancer Rca., 20:766—77, 1960.134. STOLL, H. L., JR., and CRISSEY, J. T. Epithelioma from

Single Trauma. N.Y. State Med. J., 62:496—500, 1962.135. SULZBERGER, M. B. ; HERRMAN, F.; PICCAGLI, R.; and

FRANK L. Incidence of Epidermal MethylcholanthreneTumors in Mice after Administration of Cortisone. Proc.Soc. Exp. Biol. Med., 82:673—75, 1953.

136. SVET-MOLDAVSKY,G. J. Development of Multiple Cystsand Haemorrhagic Affections of Internal Organs inAlbino Rats Treated during the Embryonic or NewbornPeriod with Rous Sarcoma Virus. Nature, 180:1299—1300, 1957.

137. SVOBODA,J. Presence of Chicken Tumor Virus in the Sarcoma of the Adult Rat Inoculated after Birth with RousSarcoma Tissue. Nature, 186:980—81, 1960.

138. SYVERTON,J. T. ; WEu.s, E. ; KOOMEN, J., JR. ; DASCOMB,H. E.; and BERRY, G. P. The Virus-induced RabbitPapilloma-to-Carcinoma Sequence. III. ImmunologicalTests for Papilloma Virus in Cottontail Carcinomas. CancerRes.,10:474—82,1950.

139. TANAKA, S., and SOUTHAM, C. M. Joint Action of WestNile Virus and Chemical Carcinogens in Production ofPapillomas in Mice. J. Natl. Cancer Inst., 29:711—22,1962.

140. TIMOFEYEVSKI, A. D. Role of Viruses in the Etiology ofTumors. Moscow: Medgiz, 1961 (Russian text).

141. Tooi@&N,H. W. A Virus Associated with TransplantableHumanTumors.Bull. N.Y. Acad. Med., 37:305—10, 1961.

142. TOTTEN, R. S.; ANTYPAS, P. G.; DUPERTUIS, S. M.;GAINSFORD, J. C.; and WHrrs'., W. L. Pro-existingRoentgen-Ray Dermatitis in Patients with Skin Cancer.

Cancer,10:1024—30,1957.143. TRENTIN, J. J. ; YABE, Y.; and TAYLOR, G. The Quest for

Human Cancer Viruses. Science, 137:885—41, 1962.144. TULLOS, H. S. ; KIRSCHBAUM, A. ; and TRENTIN, J. J.

Role of Gonadotropic Hormone in the Initiation andProgression of Adrenal Tumors in Ovariectomized Mice.Cancer lIes., 21:730—34, 1961.

145. TYLER, A. A Developmental Immunogenetic Analysis ofCancer Centering about Anti-Host Immunologic Response of Neoplastic Cells. In: Ford Hospital Symposiumon Biological Interactions in Normal and NeoplasticGrowth, pp. 533—72.Boston: Little, Brown & Co., 1962.

146. VOOT, M. E., and DULBECCO, R. Studies on Cells Rendered Neoplastic by Polyoma Virus: The Problem of thePresence of Virus-related Materials. Virology, 16:41—51,1962.

147. WISELEY, D. V.; Ko@rIN,P.; FOWLER, P. R.; and TRIVEDI,J. The Combined Effect of Repeated Viral Infection andExposure to Carcinogenic Aerosols on Pulmonary TumorInduction in C57 Black Mice. Proc. Am. Assoc. CancerRes.,3:278,1961.

148. Wmunn-MAsoN, R. Malignant Change FollowingHerpes Simplex. Brit. Med. J., 2:615—16, 1957.

149. WYNDER, E. L. ; CoRrmiiw, J.; SHROFF, P. D.; andDORAISWAMI, K. R. A Study of Environmental Factorsin Carcinoma of the Cervix. Am. J. Obst. GynecoL, 68:1016—52,1954.

150. WYNDER, E. L., and HOFFMANN, D. Some Practical As

pects of the Smoking-Cancer Problem. New Eng. J.Med., 262:540—45, 1960.

151. ZILBER, L. A., and KRYUKOVA, I. N. Haemorrhagic Diseaseof Rats Caused by Rous Sarcoma Virus.Problemsof Virology, 2:247—52, 1957.

111. Rous, P., and FRIEDEWALD, W. F. The Effect of ChemicalCarcinogens on Virus-induced Rabbit Papillomas. J.Exp. Med., 79:511—38,1944.

112. Rous, P., and KIDD, J. G. The Carcinogenic Effect of aPapilloma Virus on the Tarred Skin of Rabbits. I. Description of the Phenomenon. J. Exp. Med., 67:399—428,

1938.113. . The Activating Transforming and Carcinogenic

Effects of the Rabbit Papilloma Virus (Shope) upon Implanted Tar Tumors. Ibid., 71:787—812, 1940.

114. Rous, P.; MURPHY, J. B. ; and TYTLER, W. H. The Roleof Injury in the Production of a Chicken Sarcoma by aFilterable Agent; and A Filterable Agent the Cause of aSecond Chicken Tumor, an Osteochondrosarcoma.J.A.M.A.,58:1751,1912;59:1793—94,1912.

115. ROWNTREE, C. X-ray Cancer. Brit. M.J., 2: 1111—12,1922.

116. RowsoN, K. E. K.; ROE, F. J. C.; BALL, J. K.; and SAi@&MAN, M. H. Induction of Tumors by Polyoma Virus: En

hancement by Chemical Agents. Nature, 191:893—95,1961.

117. RuniN, H. The Nature of a Virus-induced Cellular Resistance to Rous Sarcoma Virus. Virology, 13:200—206,1961.

118. SANDBERG,A. A.; ISHIHARA, T.; MIwA, T.; and HAUSCHKA, T. S. The in Vivo Chromosome Constitution of Mar

row from 34 Human Leukemias and 60 Non-leukemicControls. Cancer lIes., 21:678—89, 1961.

119. SAUNDERS, T. S., and MONTGOMERY,H. Chronic Roentgen and Radium Dermatitis. J.A.M.A., 110:23—28, 1938.

120. SCHERER, W. F. ; KITOAKA, M., OKusso, T. ; and OGATA,T. Ecologic Studies of Japanese Encephalitis Virus inJapan. VII. Human Infection. Am. J. Trop. Med.Hygiene,8:144—59,1959.

121. SCHWARTZ,S. 0.; SCHOOLMAN,H. M.; SZANTO,P. B.; andSPURRIER, W. Studies in Leukemia. VI. The Induction of

Leukemia in AKR Mice by Means of Cell-free BrainFiltrates of Humans Who Died of Leukemia. Cancer Res.,17:218—21,1957.

122. SHEAR, M. J. Studies in Carcinogenesis. V. Methyl Derivatives of 1,2-benzanthracene. Am. J. Cancer, 33:499—537, 1938.

123. SHunIK, P. Carcinogenesis in the Syrian hamster. Proc.7th International Cancer Congress (London), p. 48, 1958.

124. SIGERIST, H. E. Four Treatises of Theophrastus vonHohenheim, called Paracelsus. Baltimore: Johns HopkinsPress, 1941.

125. SOUTHAM,A. L. The Natural History of Menstrual Disorders. Ann. N.Y. Acad. Sci., 75:840—54, 1959.

126. . Serological Studies of Encephalitis in Japan. II.Inapparent Infections by Japanese B Encephalitis Virus.J. Infect. Dis., 99:163—69, 1956.

127. . The Role of Viruses in Neoplasia, with Emphasison Human Leukemia. J. Ped., 63: 138-57, 1963.

128. SOUTHAM,C. M., and BABCOCK, V. I. Tissue Preferencesof “Metastases―of HEp 2 Cells in Newborn Rats (inpreparation).

129. SOUTHAM,C. M.; CRAVER, L. F.; DARGEON, H. W.; andBURCHENAL, J. H. A Study of Natural History of Acute

Leukemia with Special Reference to Duration of theDisease and the Occurrence of Remissions. Cancer, 4:39—59, 1951.

130. SPIzIzEN, J. Infection of Protoplasts by Disrupted T2Virus. Proc. NatI. Acad. Sci., 43:694—701, 1957.

131. STEINER, P. E. Additional Comments on Differences between Cancers in Terms of Etiologic Factors. CancerRes.,16:681—83,1956.

132. STEWART, S. E.; EDDY, B. E.; and BORGESSE, N. G. Neoplasms in Mice Inoculated with a Tumor Agent Carried

Research. on December 11, 2020. © 1963 American Association for Cancercancerres.aacrjournals.org Downloaded from

1963;23:1105-1115. Cancer Res   Chester M. Southam  The Complex Etiology of Cancer

  Updated version

  http://cancerres.aacrjournals.org/content/23/8_Part_1/1105

Access the most recent version of this article at:

   

   

   

  E-mail alerts related to this article or journal.Sign up to receive free email-alerts

  Subscriptions

Reprints and

  .pubs@aacr.orgDepartment at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

  Permissions

  Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://cancerres.aacrjournals.org/content/23/8_Part_1/1105To request permission to re-use all or part of this article, use this link

Research. on December 11, 2020. © 1963 American Association for Cancercancerres.aacrjournals.org Downloaded from