cellular adaptation in the origin and development of cancer1 · cellular adaptation in the origin...

(CANCER RESEARCH 51, 2751-2761. June 1. I991|

Perspectives in Cancer Research

Cellular Adaptation in the Origin and Development of Cancer1

Emmanuel Farber and Harry RubinDepartments of Pathology and of Biochemistry, Medical Sciences Building, University of Toronto, Toronto. Ontario, Canada M5S IA8 [E. F./, and Department ofMolecular and Cell Biology and I'irus Laboratory, Stanley Hall, University of California, Berkeley, California 94720 [H. R.]

Introduction

A dominant view today in cancer research assumes that themajor need for our understanding of cancer resides in thethorough documentation of the many alterations in gene structure, organization, and arrangement that we know are presentin some malignant cell populations and in their immediateprecursors. While attractive from some points of view, especially in the hope of delineating a relatively simple cause-effectrelationship between the genome and cancer, this simple formulation largely ignores a fundamental property of all livingforms, the ability to respond and adapt to many environmentalperturbations. Given the widespread occurrence of cancer inmany species and given the unusually prolonged nature of theprocess of cancer development in most instances in humansand in animals, often lasting from one-third to two-thirds ofthe life span of the organism, it would indeed be unlikely andunbiological for cancer development to proceed without theparticipation of many active responses, including adaptive ones,to the various new cellular phemomena that are part of thecarcinogenic process. Immunological surveillance has been apopular speculation proposed to account for this prolongedprocess. However, immunological responses are only a smallpart of this spectrum of reactions, with considerable doubt (1)that they participate actively prior to the relatively late appearance of unequivocal cancer with its many new biological components such as tissue destruction with cell death, inflammation, ulcÃ©ration,etc.

All living forms have evolved in a largely unfriendly or evenhostile environment in which protective or adaptive responsesto many different forms of potential injury or harm have beenessential for survival and reproduction. In fact, it appears thatthe acquisition of mechanisms for many different adaptiveresponses could be considered to be just as important forsurvival and reproduction as the development of the essentialpathways for the basic physiological needs of living organismssuch as for energy and for the syntheses of the small and largemolecules that are the component parts of the pathways. It iswidely recognized that a variety of physiological adaptive responses to varying altitudes, other environmental influences,and hormonal modulations as well as in aging are fundamentalproperties of living organisms. In addition, the vast array ofdifferent xenobiotic chemicals and organisms as well as radiations to which all living forms have been exposed since the earlyforms evolved some 2 to 3 billion years ago would requirehighly versatile protective systems to be developed. Such protective or adaptive mechanisms are known to be presentthroughout the whole spectrum of living forms from single cellmicroorganisms to highly differentiated eukaryotes.

Received 9/18/90; accepted 3/25/91.' The authors' research included in this article was supported by grants from

the National Cancer Institute of Canada and the Medical Research Council ofCanada (MT-5994) (E. F.): and from National Institutes of Health. CA 15744.and grant 1948. Council for Tobacco Research. USA (H. R.).

Despite the obvious importance of such systems, their studyin disease has received little current emphasis, especially witha mechanistic orientation. Since disease processes are largelyexpressions of how living organisms react and respond toperturbations in the external and internal environments, adaptive or protective responses and their modulations and mechanisms would seem to be of the greatest concern in fundamentalstudies of disease pathogenesis. Although such considerationshave been the subject of occasional deliberations over the years(e.g., Refs. 2-8), they have received little attention in the

detailed studies of chronic disease in general and cancer inparticular, especially from the viewpoint of mechanism.

In the face of the many hazards, including some carcinogenicones, to which biological systems have been exposed for millions of years, it seems appropriate to ask the question,"What

mechanisms might have evolved to preserve biological continuity?" Probably many earlier species have been eliminated by

exposure to one or more of such hazards. In those that havesurvived and endured, what is the nature of the adaptationsthat have evolved and could a knowledge of these be useful indesigning new approaches to the prevention of cancer or evenits treatment (9)? What essential roles, if any, do these adaptiveprocesses play in the development of different cancers?

Adaptive Responses during Cancer Development

Cancer development with the different known etiologicalagents, chemicals, radiations, DNA viruses, and some RNAviruses without oncogenes is a very prolonged process as alreadymentioned. The only apparent exceptions to this are the specialcarcinogenic processes induced with oncogene-containing retro-viruses (10). The long period of development is characteristically associated with the presence of various types of focalproliferations of an entirely benign nature (some designated as"benign neoplasms") that may show a slow cellular evolution

to cancer (9).During this prolonged process, there appear to be a minimum

of three general types of cellular adaptive responses: type A,early acute transitory responses, usually of a diffuse, zonal, orregional nature, largely reversible; type B, clonal constitutiveadaptation with the development of a new cell population thatis associated with a resistance of the organ or tissue and of theorganism to environmentally induced cell damage; and type C,patterns of adaptation of cells to a variety of environmentalmodulations during neoplastic transformation. The main substance of this communication is to discuss briefly each of thesethree types of adaptations and to suggest how they might beimportant as one set of guidelines for cancer prevention andcancer therapy. This overview is limited to cellular responsesthat seem to participate actively in the step-by-step sequence ofevents leading to malignant neoplasia. Adaptive responses ofthe whole organism, especially the late ones in response to fully-formed cancers, are omitted.

2751

Research. on October 19, 2020. © 1991 American Association for Cancercancerres.aacrjournals.org Downloaded from

http://cancerres.aacrjournals.org/

CELLULAR ADAPTATION AND THE ORIGIN OF CANCER

Acute Transitory Adaptive Responses

The known responses in this category relate mainly to carci-nogenesis with radiations and with mutagenic chemicals. Although it is our conviction that similar adaptive responses mightbe found in cancer development with cancer-related DNA viruses and some RNA viruses (e.g., Refs. 11 and 12), data inthis area are scanty.

Reversible Inductions of Metabolizing Enzymes. One of themost general and basic forms of adaptation to xenobiotic chemicals, including, of course, the majority of drugs, is inductionof enzyme patterns that relate to the metabolism and possibledetoxification of the chemical (13-17). This type of response isseen throughout the living world including the simplest bacteria. In mammals, this often includes two major aspects ofdetoxification: metabolic activation to more polar derivatives,such as via the cytochrome P-450 system and other oxidativeas well as reductive pathways (phase I); and conjugation ofmetabolic products for excretion or other pathways of removal(phase II). The enzymes involved often show transitory increases in both transcription and translation and usually returnto their baseline levels after the exposure to the xenobiotic isover.

It is well documented that these acute responses are frequently associated with the more rapid excretion of the potentially toxic xenobiotic with an obvious protective result (13-20). Apparently, the multiple nature of the responses to somany xenobiotics facilitates both metabolism and excretion.

Included in the xenobiotic chemicals are a smaller group ofactual or potential mutagenic ones, genotoxic carcinogens, thatform a significant percentage of chemical carcinogens. Theseagents also induce a spectrum of enzymes in liver and otherorgans that encompass both phase I and phase II components(21, 22). The obvious roles of those components in the genesisof more toxic mutagenic derivatives, most often electrophilicreactants, are well established. However, any role of thesereactive derivatives in the survival of the organism is not evidentat this time.

Acute Reactive Proteins. Another reaction that is often seenin the higher organisms is the synthesis and secretion of acutereactive proteins, especially by the liver (23). These proteins aretransitorily elevated in the blood during an episode of injuryand return to their baseline values when their injury is over.They are considered to play a protective role in the acute injury.

Heat Shock Proteins. A third general adaptive response toheat, toxic chemicals and a wide spectrum of other environmental perturbations is the synthesis of "heat shock proteins"

(24-26). This response, like the xenobiotic-induced enzymeinduction, is present in Escherichia coli, in yeast, and in a widespectrum of organisms including insects and mammals. Thereis an overall agreement that these responses are often protectiveand should be considered as adaptive (26).

Thus, at least three different adaptive response patterns canbe distinguished on exposure to many different xenobiotics andat least two, Responses 1 and 3, are present in a wide range oforganisms, from single cell to highly complex multicellular.Responses that span such a broad spectrum of living organismsprobably play an important role in maintaining the integrityand function of the different living forms, especially as to howthey react to and survive environmental disturbances (27-30).

DNA Repair Enzymes. There is no doubt that the development of a variety of mechanisms for the repair of DNA, alteredby mutagenic chemicals and radiations and perhaps viruses, is

a major adaptive set of responses of many different organismsfrom single cell to the human (31-34). The mechanisms includephotoreversal of pyrimidine dimers, mismatch repair acting onincorrectly paired bases from unfaithful DNA replication, enzymatic alterations with recombinant events and excision repair. Interesting enzymes are the alkyltransferases which remove methyl or other alkyl groups from the O6 position ofguanine, the O* position of thymine, or the phosphate groups

of DNA and transfer the alkyl groups to a cysteine S in thetransferase itself. In E. coli, the transfer is part of a veryinteresting adaptive response to alkylating agents associatedwith the ada gene.

The DNA repair systems are very widespread in nature andtheir adaptive protective role is clearly seen in many systemsincluding, of course, humans. The susceptibility to skin cancerof humans with xeroderma pigmentosum, with their severaltypes of deficiencies in the repair enzyme systems, is a strikingexample of the importance of the DNA repair systems forsurvival.

Cianai Constitutive Adaptation

During a systematic study of the sequence of steps in thecarcinogenic process in the liver (35), it became evident thatanother form of adaptation exists, one that originates in raresingle hepatocytes scattered throughout the liver and that isconstitutive, rather than diffuse and readily reversible (9, 36).This response pattern has been seen with genotoxic chemicalcarcinogens. It involves the induction of a new resistance phe-notype in a rare hepatocyte and the expansion of these rarehepatocytes to form focal proliferations (hepatocyte nodules).The recognition of this new adaptive phenomenon hasimportant potential implications not only for a new conceptualframework for the development of some cancers but also intumor promotion, in cancer prevention, in the analysis ofresistance of many primary cancers to chemotherapy, and inthe analysis of how some living forms have adapted, presumablythrough evolution, for survival in a hostile environment.

Background. During the first few years of our study of themechanisms of chemical carcinogenesis in the liver, we gradually became impressed by five considerations, each of whichsuggested that some new approach was needed.

(a) Cancer development in humans and animals with chemicals, DNA viruses and some RNA viruses, radiations, and dietis an inordinately long process, requiring from one-third totwo-thirds of the life span of the organism (9, 37). This isinteresting and puzzling, since it can be clearly shown in experimental animals and in some humans that exposure to a carcinogen began or occurred many years prior to the first appearanceof cancer and was brief and transitory. Although a dominantrole for immunological surveillance was first suggested byThomas in 1959 (38, 39), the evidence for this during chemicalcarcinogenesis is unimpressive (1). Alternative hypotheses forthe very prolonged nature of the carcinogenic process are necessary if we are to develop rational approaches to cancerprevention.

(b) Autonomous or semiautonomous growth of initiated cellsis a property acquired late in the carcinogenic process. Initiationwith one of many different carcinogens is at no time followedby spontaneous or autonomous proliferation of any cells in theliver or in other organs. Only with large doses of carcinogensand periods of exposure much longer than are required forinitiation does one see focal lesions with autonomous cell

2752




proliferation. If the initiated altered cells induced by exposureto a carcinogen do not grow spontaneously, how might adifferential stimulation of their growth be created? The growthof the rare altered cell leading to focal proliferations ("benignneoplasms") is a key phenomenon in promotion in virtually all

experimental carcinogenesis and in many human systems ofcancer development.

(c) Virtually every chemical carcinogen is an inhibitor of cellproliferation (mitoinhibition) (see Ref. 40 for referencestherein). Haddow (41) suggested in 1938 that inhibition of cellproliferation could be an early effect of carcinogens and that,in such an environment, resistant cells might arise and beencouraged to proliferate. Recently, it has been reported thattwo non-carcinogenic promoting agents for hepatocarcinoge-nesis, phÃ©nobarbital(42-45) and orotic acid (46-48), also showconsiderable inhibition of hepatocyte proliferation, not unlikethat seen with the carcinogens.

(</) The response of tissues or organs during carcinogenesis,most relevant as potential precursors for cancer, is focal, notgeneral. In the majority of instances of cancer development inhumans or in animals in which a precursor cell population orlesion has been identified or proposed, the "preneoplastic" and"precancerous" changes are always focal, often clonal, involving

only a very small number of altered cells (9. 36, 37). A selectionpressure must be established if the altered cells are to proliferateselectively, relative to the surrounding cells (49).

(e) Carcinogens (chemicals, viruses, non-ionizing and ioniz

ing radiations) are widespread in nature. Humans and animalshave been exposed for millions of years to carcinogens, bothpresent in the external environment, including the food, andgenerated endogenously. How has nature handled carcinogensand carcinogenesis? What systems have evolved that couldallow for the relatively long life span of so many species? Couldany insight into these questions generate clues to mechanismsand interpretation of the carcinogenic processes?

The Resistant Hepatocyte as an Early Response to Carcinogens. With well over 75 different chemical carcinogens of quitedifferent chemical structure and properties as well as in hepa-tocarcinogenesis seen in rats fed a choline-devoid diet, a resist

ant hepatocyte is a common accompaniment of initiation (9,35). A seemingly similar phenotype appears in isolated hepa-tocytes on infection with v-H-ras and v-ra/(12) and in culturedrat liver epithelial cells during transformation induced by exposures to /V-methyl-yV'-nitro-A'-nitrosoguanidine (50, 51).

The resistance phenotype (52) is manifested at three levels oforganization: (a) physiological (behavior), resistance to inhibition of cell proliferation (mitoinhibition); (b) cellular, resistanceto visible cytological responses to some cytotoxic environments;and (c) biochemical-molecular, the presence of a biochemicalpattern that diminishes or counteracts possible toxic effects ofxenobiotics.

(a) At the physiological level, the key manifestation of theresistance phenotype in the rare hepatocyte is the ability toproliferate vigorously in response to a proliferative stimulus inan environment, such as that created by the presence of one ofmany carcinogens and some promoters, in which the vastmajority of hepatocytes, the nonresistant ones, are inhibited.This is a common basis for the promoting ability of manyhepatocarcinogens and may also be important with two non-carcinogenic promoting agents, the drug, phÃ©nobarbitaland thenormal metabolite and obligatory precursor of pyrimidine nu-cleotides, orotic acid (42-48).

(b) The hepatocyte nodules show resistance to cytotoxic

effects of the agents used in their generation and cross-resistance to other xenobiotics and to a dietary deficiency to whichthat rat was not exposed (35, 36).

(c) The resistant hepatocytes induced with quite differentregimens show an unusually common biochemical pattern (35,36). This pattern shows consistent decreases in several phase Icomponents including total cytochromes P-450 and severalmixed function oxygenases, consistent increases in many phaseII conjugating systems, characteristic alterations in glucosemetabolism, and a reproducible altered pattern of iron andheme metabolism (36).

It is important to emphasize that many different enzymesand components contribute to this phenotype. For example,the decrease in total cytochromes P-450 is so large (60-80%)that several different ones must be involved (53, 54). Thismultiplicity of active components in the resistant hepatocytescould readily allow for considerable variation in the levels ofone or more components in different clones derived from thesehepatocytes (55, 56) without compromising their overall resistant behavior in an appropriate selecting environment.

It is interesting and noteworthy that the resistance phenotypein the hepatocyte nodules has several resemblances to thephenotype of some human cancer cells that are resistant tochemotherapeutic agents (57-59).

Nodules of Resistant Hepatocytes as a Type of Adaptationwith Survival Value. What is the evidence that the nodules ofresistant hepatocytes which we designate as "hepatocyte nodules" are better considered as a form of physiological adaptive

response (9) rather than as collections of abnormal cells, namelybenign neoplasms, that are part of a pathological sequencespecifically related to the development of cancer? According tothe latter view, the nodules of hepatocytes derived by clonalexpansion of potential initiated hepatocytes are abnormal fromthe outset and represent foreign cell populations derived on thebasis of abnormal genes and gene products.

The most telling evidence that points overwhelmingly to thephysiological adaptive nature of the hepatocyte nodules existsat three levels of organization: (a) the cell and tissue architecture and organization; (b) the whole organism; and (c) thebiochemical-molecular.

(a) The hepatocytes in the nodules are organized differentlythan are the hepatocytes in the mature normal and surroundingliver. In contrast to the hepatocytes in the latter which arearranged predominantly as single cell plates, those in the nodules are in the form of double cell plates and acini (60).However, this difference in organization and architecture is notan expression of abnormality, since the hepatocytes in thenodules undergo spontaneously a radical differentiation withrestructuring and remodelling (61, 62). This restructuring,clearly genetically programmed, involves the cell to cell organization and the blood supply as well as the biochemical pattern.

This phenomenon is seen in several models of hepatocarcin-ogenesis. In every respect, this is truly a process of differentiation, analogous to other examples of differentiation in normaldevelopment.

(b) Rats with hepatocyte nodules are unusually resistant to alethal dose of a potent hepatotoxic agent, carbon tetrachloride(63). Rats with nodules show complete resistance to a dose ofCC14 that is lethal for 100% of normal rats.

This demonstration of survival value indicates that the genetic information necessary to develop hepatocyte nodules andto have them remodel by differentiation could have been favoredduring the long period of evolution.

2753




Another aspect, possibly related to survival, is the resistanceof the hepatocytes in nodules to the cytotoxic effects of somehepatotoxic xenobiotics in vitro and in vivo (see Refs. 35 and64).

(c) At the biochemical-molecular level, the biochemical pat

tern can readily account for the resistance of the nodules tocytotoxic effects of xenobiotics and for the more efficient excretion of at least one carcinogen, 2-acetylaminofluorene (65).A marked decrease in activation associated with the largedecrement in phase I components, coupled with the moreefficient conjugation and excretion of active moieties generated,associated with the several conjugating systems in phase II,match quite well (65).

The virtual certainty that this pattern is physiologically programmed is indicated by its manifold nature with at least 25 to30 different enzymes and other components and by the observations that some agents such as lead nitrate, an interferon,butylated hydroxyanisole, and butylated hydroxytoluene inducea biochemical pattern in normal livers very similar to that seenin the hepatocyte nodules but in a reversible or transient manner(36). This special biochemical pattern is not seen in fetal,neonatal, or regenerating liver but seems to be uniquely associated with the hepatocyte nodule induced by one of manydifferent carcinogens.

These results suggest that hepatocyte nodules represent a newstate of differentiation, better adapted to survive in a hostileenvironment. The resistance, coupled with the capacity of themajority of nodules to undergo spontaneous differentiation tonormal-appearing mature liver, indicate that the clonal expansion seen in the early phase of carcinogenesis is geneticallyprogrammed and is part of a normal physiological pattern ofadaptation to some types of xenobiotics (9, 35, 36).

Is Clonal Adaptation Also Probable in Other CarcinogenicSystems? Superficially, there are several similarities in theoverall patterns among many experimental and human systemsfor cancer development (9, 37). These include: (a) the very longso-called "incubation period" during which the cellular evolu

tion of malignancy from nonmalignant precursor lesions isoccurring; (b) the apparent reversibility of many putative pre-neoplastic or precancerous lesions; (c) the common biologicalbehavior and appearance of early focal proliferations such aspolyps, papillomas, and nodules with many different carcinogens and promoting environments. Interestingly, the colonsystem shows a very large number of changes in the mRNApatterns between normal mucosa, polyps, and multiple poly-posis (66). This is quite consistent with the suggestion thatpolyps may represent a different state of differentiation. Itwould seem prudent, therefore, to explore some of these systemsfrom the point of view of clonal adaptation. In our opinion, itis quite likely that this type of adaptation, as well as perhapsothers, might well be important in cancer development inseveral systems.

The most telling is in the case of malignant melanoma inhumans. Clark et al. (67) have shown that the genesis andbehavior of nevi and their role in the development of melanomahave many similarities to the hepatocyte nodules in rat livercarcinogenesis. The nevus shows at least two options, differentiation to nerve endings as a major pathway and persistencewith further cellular evolution to cancer as a minor one.Whether the resemblances between rat hepatocyte nodules andhuman nevi extend to biochemical-molecular mechanisms remains a fascinating problem for study.

Relevance of Clonal Adaptation to Carcinogenesis and Cancer.Tumor promotion, despite its frequent association with anelement of mystique, is basically clonal expansion of initiatedcells. In the liver, this leads to the genesis of hepatocyte nodules.The concept of clonal adaptation places tumor promotion inthe liver in quite a different perspective. This part of thecarcinogenic process is only in part and perhaps only peripherally related to the ultimate development of cancer but becomesan important way in which the living organism responds in anadaptive fashion to commonly occurring xenobiotics in theenvironment. On this basis, scientifically, the neoplastic derivative, the cancer, is perhaps only incidental while the adaptivecomponent is the most important from the perspective ofbiology. In this regard, the "turning on" of a new phenotype in

a rare cell during initiation becomes the key to what is seen inpromotion.

How a mutagenic carcinogen induces a new common constellation of biochemical components that go to make up theresistance phenotype is not understood. Although the inductionof mutations in a rare cell by mutagenic carcinogens or metabolites is well known, it remains to be established that this is themolecular mechanism for the induction of the resistant hepatocyte. Altered regulatory genes, rare gene rearrangements,gene translocations, or gene amplifications are some possibilities. Hypomethylation of some genes, the protein products ofwhich show considerable elevation in the hepatocyte nodules,has been found (68-72). Altered metabolic patterns [see Foxand Radicic (73) and Dean and Hinshelwood (5)] coupled with"progressive state selection" (see below) offer possible epige-

netic alternatives to the mutation hypothesis.It would appear that the study of the control of transcription

of several of the discrete components of the resistance pheno-types, such as glutathione 5-transferase 7-7 (P), 7-glutamyl-transpeptidase, D-T diaphorase (quinone reducÃase), and discrete relevant cytochromes P-450, to name but a few, could inturn suggest possible underlying mechanisms for clonal adaptation including altered methylation.

An interesting aspect of clonal adaptation of great practicalimportance relates to the resistance of many cancers to chemotherapy (74). It is often implied that resistance to chemotherapymight be a property acquired during the schedule of treatmentas a mutation or some other genomic alteration including geneamplification. The experience in the liver and the existence ofclonal adaptation suggest the possibility that resistance may bea primary property of the cancer in some instances. Given thederivation of cancer by cellular evolution from a nodule with aresistance phenotype that was used mainly for clonal expansion,it is certainly conceivable that many cancers in some tissuesmay begin with a resistance phenotype. The presence in cancerprecursor cell populations of P-glycoprotein, as the product ofmdr gene and other components of resistance to xenobioticsincluding chemotherapeutic agents (52, 57-59), would suggestthat some cancers may begin with a resistance phenotype andmay have to lose this phenotype as a prerequisite for a positivecytotoxic response to some chemotherapeutic agents.

In view of the widespread occurrence of many potentialcarcinogens in the environment, it remains important to understand how evolution has enabled organisms to adjust or toadapt to this ever-present hazard. The long life span of manyspecies in the face of this hazard attests to the efficiency withwhich evolution has succeeded. The long precancerous period,lasting as long as one-third to one-half or more of the life span,has largely relegated neoplasia to the postreproductive period.

2754




The acquisition of a mechanism to adapt, such as by clonaladaptation, would seem to be an important way in which thishas been accomplished. If this viewpoint is valid, one wouldhave to consider neoplasia as mainly a manifestation of theimperfection of the adaptive process and cancer as a deviant ofadaptation (75). Such a viewpoint might open several newavenues for the development of novel ways to interrupt thecarcinogenesis process and thereby prevent cancer.

Finally, the presence of a major adaptive component duringcarcinogenesis would offer an interesting new perspective to thedominant current paradigm of carcinogenesis, the progressiveacquisition of abnormal genes and gene products without anyregard to the adaptive ability and talent of biological systems.Our current view would envision the appearance of new physiological phenotypes during initiation either by mutation or byother mechanisms including epigenetic, and this phenotypewould play a central role in the long adaptive process duringcarcinogenesis. Once a malignant change supervenes as a rarederivative relatively late in the development of cancer, a sequence of altered genes could then begin which might be areasonable basis for some of the late changes seen in the furtherevolution of cancers. "Progressive state selection" (see below)

could play an important role in this progression as well as inits modulation. A formulation more consistent with moderndynamic biology, such as here proposed, not only seems attractive theoretically but also offers many possible practical approaches to the ultimate prevention and treatment of cancer.

Neoplastic Transformation of Cells in Culture

Initiation or the First Stage of Carcinogenesis in Culture. Theconcepts of initiation and promotion of neoplastic transformation and the associated two stage model of carcinogenesiswere developed in the early 1940s from studies of chemicallyinduced tumors in the skin of rabbits and mice (76, 77). In themouse, the first, or initiating, stage was induced by a singleapplication of carcinogens such as coal tars or purified polycy-clic hydrocarbons and persisted indefinitely thereafter in thetreated area, which generally retained its normal appearance.Repeated application to that area of a promoting agent such ascroton oil or TPA,2 which by themselves are not carcinogenic,

resulted in the second stage defined by the appearance oftumors. Largely because of the speed of induction of the initiated stage, its persistence after removal of the carcinogen, andthe focal origin of the ensuing tumor, the first stage wasgenerally considered to be mutational in origin (78, 79). Therequirement for repeated application of promoting agents andthe slowness of onset of the second stage led to the belief thatit was nonmutational, or epigenetic, in origin. In fact, however,whole animal systems were too complex and cumbersome toprovide definitive evidence of these widely accepted explanations for the two stages. To add to the complications, it hasbeen pointed out that the two-stage model refers only to thenumber of manipulations used and that many more than twosteps occur in the two stage system in animals (49).

The availability of cell culture systems that could be inducedto undergo neoplastic transformation by exposure to carcinogens offered a much simpler, easily manipulated, and readilyquantified means for settling the question of the nature of two-stage carcinogenesis. Primary cultures of Syrian hamster em-

2The abbreviations used are: TPA, IJ-O-tetradecanoylphorbol-IJ-acelate:DMBA. 7.12-dimethylbenz[a)anthracene.

bryos at cloning densities were the first to be used successfullyin studies of in vitro carcinogenesis by chemicals (80). Up to25% of the clones of these cells underwent morphologicaltransformation during a 9-day treatment with carcinogenicpolycyclic hydrocarbons, and some of the transformed cloneswent on to produce sarcomas when transplanted back intoSyrian hamsters. The high percentage of clones which becamemorphologically transformed indicated to the authors that thetransformation was not the result of "the usual type of randomlyoccurring mutation" (80). However, since the scoring of trans

formation was subjective, and only a single stage could bediscerned, the mechanism(s) of the two stage models could notbe resolved in this system.

In 1970, Mondai and Heidelberger (81) carried out a classicseries of experiments on chemical carcinogenesis in a line ofmouse prostate cells. They found that every cell exposed tomethylcholanthrene gave rise upon cloning to a population inwhich a small minority of the cells initiated transformed foci.By contrast, only 6% of control cells gave rise to clonal populations which had some transformed foci. The fact that 100%of carcinogen-treated cells produced progeny with an increasedprobability for transformation argued strongly against a mutational origin for the transformation.

The work in culture was continued with the C3H10T'/2 lineof mouse embryo cells, hereafter referred to as 10T'/2 cells. Thisline was widely used by many investigators to study carcinogenesis induced by a variety of chemical carcinogens (82) and byX-irradiation (83). A finding reported by all who used thesystem was that the frequency of transformation (the numberof transformed foci divided by the number of cells initiallyseeded) was inversely proportional to the number of cells seeded(82-88). One particular version of this type of experiment wasto administer 400 to 600 rads of X-irradiation to a low densityof 10T'/2 cells, allow them to grow to confluency, transfer them

on many dishes over a wide range of concentrations, and scorefoci 4 weeks after each cell concentration reached confluency(83). A consistent finding was that the number of foci per dish(usually less than one per culture) was independent of thenumber of cells seeded in the transfer. It was concluded thatmost, if not all of the cells originally exposed to the X-rays hadbeen altered by the treatment in such a manner as to transmitto their progeny an increased probability for transformation. Asimilar conclusion was reached for chemically induced transformation of the 10T'/2 cells (87) and of retrovirus-infected rat

cells (89).The production of heritable change in an entire population

is a radical departure from the quantitative expectation ofspecific genetic mutation, which usually occurs at a rate of 10~6/

cell division or less. Furthermore, this finding of high frequency,heritable increase in focus-forming capacity initiated by X-raysis consistent with other reports that most or all of the cellswhich survive X-irradiation but receive sublethal hits haveheritable changes in growth rate (90, 91) and sensitivity tomutagenic treatment (92). A correlated finding is that the totalsize of the transformation targets of ionizing radiation at highlinear energy transfer is as much as 1000 times higher than thatof the targets for specific gene mutations, which are themselveslarger than that for a single gene (93). Large as it is, the targetsize based on transformation may be an underestimate for theinitiation process, since only a small minority of initiated cellsundergo transformation (94).

Several types of experiments with tumor initiators in animalsshowed that enduring changes occur in a high proportion of

2755




carcinogen-treated normal cells just as they do in cell lines. Inone case, the skin of mice was treated with the carcinogenDMBA and 5 weeks later almost all the basal cells in theDMBA-treated area responded to croton oil by a faster onsetof DNA synthesis than did the basal cells of mice which hadnot been treated with the carcinogen (95). Since epidermis is arenewing tissue with continuous turnover of cells, the persistentDMBA-induced responsiveness indicated a heritable increasein the reactivity to promoting agents of all the basal cells in thearea. In another case, DMBA was repeatedly applied to hamstercheek pouches, resulting in the appearance of localized carcinomas (96). There was a large increase in proteolytic activityin the carcinomas and in the surrounding, normal-appearingtissues, indicating that a high frequency change had occurredover the entire exposed area. Several studies on tumor inductionin normal diploid cells show that heritable change occurs in ahigh proportion of them after a single dose of X-rays. X-irradiation of thyroid cells freshly removed from the rat converts 1 of 32 of the clonogenic cells into tumorigenic cells (97).A similar result was obtained with mammary epithelium removed from rats that had been treated 24 h earlier with achemical carcinogen (98). At least 1 in 250 of the cells initiateda mammary carcinoma when dispersed and grafted into the fatpads of isologous recipients. Taken together, these findingsclearly show that carcinogenic agents produce heritable changesin a much higher proportion of treated cells than would beexpected of conventional mutations. The implication is eitherthat damage to any one of a very large number of genes canproduce the neoplastic transformation or that the primarytargets are not genetic. An important finding favoring the latterinterpretation is that X-ray-initiated 10T'/2 cells are irreversibly

reverted to their original uninitiated condition by treatmentwith protease inhibitors begun as late as 10 days, or 13 celldivisions after irradiation (99). There is certainly no reason toexpect that protease inhibitors would revert genetic mutationsin a whole population of cells, particularly at so long a timeand so many divisions after X-irradiation.

Transformation or the Second Stage of Carcinogenesis in Culture. The promotion stage is defined as that which leads totumor formation. Transformation is the cell culture analogueof tumor formation and can be considered the second stage ofcarcinogenesis in vitro (94). Although all the X-irradiated 1OT'/2

cells appear to be initiated, only a small fraction of them (about10~6) actually produce transformed foci (94). It was concluded

that the second step in transformation is a very rare event.Fluctuation analysis of clone size distribution of transformedcells agreed with the hypothesis that the second event is aspontaneous mutation, with a constant small probability ofoccurring every time a cell divides (94). This conclusion was atodds with an earlier finding that no cells capable of producingfoci on transfer appeared in carcinogen-treated lOT'/z cultures

until about 3 weeks after they had become confluent and netgrowth had long since stopped (84). It also seems inconsistentwith the transforming effect of the tumor promoter, TPA, on10T'/2 cells exposed to 100 rads of X-rays, which is insufficientradiation to significantly increase the frequency of the rarefocus found among control cultures (100). The addition of TPAto the lightly X-rayed cells increased focus production to theextent that almost every culture arising from a single cell had1 or more foci. The effect of TPA was therefore similar toincreasing the dose of X-rays to 400-600 rads, in that itincreased the focus-forming capacity of almost every cell. Onthis basis, the promoting or second stage in these cells, like the

initiating or first stage, could be considered a high frequencyprocess, even though its visible manifestation occurs with lowprobability.

The question of the nature of the second or transformingstep has since been taken up in the NIH 3T3 cell system, whichcan be considered an initiated line, since foci of transformedcells develop if the cultures are left in the confluent state for 10days or longer in a given range of serum concentrations (101-103). This capacity for focus formation is highly dependent onthe passage history of the cells. If they are passaged frequentlyat low density in high serum concentration, their capacity fortransformation steadily diminishes to the point that no transformed cells arise, even after several successive postconfluentincubations.' Passage of the original cells at low density but in

low serum concentration gives rise to focus-forming cells, justas prolonged incubation at confluency does. Both conditionswhich evoke the transformation constrain the growth and overall metabolism of cells, and it was surmised that the transformation represents an adaptive response to that constraint.However, growth constraint imposed by the very differentmethod of limiting the availability of glutamine actually inhibitstransformation instead of evoking it (104). The observation thattransformation occurs only under certain specified physiological conditions shows that it is not a spontaneous event in thesense that genetic mutations are. The fact that it was favoredby physiological conditions which constrain rather than encourage multiplication was further evidence against a muta-tional origin of the transformation.

Although only a small fraction (usually less than 1%) of theNIH 3T3 cells became focus formers when exposed to growth-constraining conditions, the entire population actually underwent a heritable increase in growth capacity. This was expressedin a substantial increase in saturation density of the entire cellpopulation after 3 short passages in low serum concentrationand even greater increases in further passages (102). Clonalanalysis of a similarly treated population revealed that all thecells had acquired an increased capacity for focus formation,although less than 1% of the cells seeded from the unclonedparental population initiated foci in a direct assay (102, 105).There was an extremely wide range of capacities to producefoci among the clones, and some of them drifted up or down inthat capacity as they were repeatedly passaged and assayed forfocus formation (102). It is apparent that the population doesnot fall into two simple classes of transformed and nontrans-formed cells; rather it consists of a mosaic of cells with a verylarge range of capacities for transformation in confluent cultures. There may be an almost continuous spectrum of potentialities or probabilities for transformation among cells whenthe appropriate conditions are applied. In addition, the expression of the transformed state in the development of foci dependson the initial state of the cells, a variety of factors in the assay,and the differential in growth rates among the cells atconfluency.

These considerations indicate that there is an optimal degreeof constraint for eliciting the adaptive growth response thatcharacterizes transformation. If the serum concentration inconfluent cultures is too low or too high or if cells are seededat such high concentrations that they become crowded as soonas they attach and spread on the dish, the number of foci perseeded cell is reduced. It is important to take these factors intoaccount when comparing the results obtained in cell culturewith those obtained in the animal. In cell culture, sparse cell

3 H. Rubin, unpublished observations.

2756




populations in high serum concentrations are growing andmetabolizing at a maximal rate and the adaptive responseassociated with transformation is brought on when constraintsassociated with confluency or low serum concentrations areintroduced. In the animal, most cell types are normally underthe constraint introduced by the 3 dimensional contiguity ofthe organized tissue. For example, very few cells of the intactliver are synthesizing DNA at any moment, and less than 0.01 %are in mitosis (106). If the liver epithelial cells are dispersedand seeded at low density in culture, 80% of them enter DNAsynthesis and mitosis within 48 h (107). Therefore, in theanimal, the optimal level of growth and metabolism for neo-plastic transformation of initiated cells may come only whenthey are stimulated to multiply. This appears to be the case inliver carcinogenesis, where nodule formation by carcinogen-treated cells is brought on by compensatory growth (49). Bycontrast, transformation of rat liver cells in culture is enhancedby keeping the liver cells in the confluent, relatively inhibitedstate for 3 weeks at each month passage, instead of seriallypassaging them once a week just as they reach confluence (50).Thus, growth stimulation in vivo and growth inhibition in vitromay converge on the optimal condition for transformation.

Just as reversal was demonstrated for the initiated state, sohas it been shown for the transformed state. It was first reportedthat cultivation of transformed Syrian hamster cells at lowdensity reversed several of their in vitro growth properties andreduced their capacity for tumor formation (108). Then it wasfound that the same type of low density growth fully reversedthe morphological and tumor-producing properties of a cloneof 10T'/2 cells that had been transformed by X-irradiation (109).

Similar treatment of spontaneously transformed NIH 3T3 cellsresulted in a loss of their focus-forming capacity and of theirability to form tumors in nude mice (102, 110). In both the10T'/2 and NIH 3T3 cells, the transformed phenotype was

perpetuated by maintaining the cells at high density. The reversibility of the transformed state of a whole culture consideredalongside its induction under conditions of constraint, and thegraded response of the entire population leave little room fordoubt that the transforming step, like that of initiation, is in itsinception a nongenetic process.

Adaptive Origin of Spontaneous Transformation. The diverseevidence for the adaptive origin of "spontaneous" transforma

tion of NIH 3T3 cells can be summarized as follows:1. No cells are capable of producing dense foci under standard

conditions if the cells had been repeatedly transferred at lowdensity in high serum concentrations. Focus-forming cells canbe detected, however, by reassaying cells from the standardassay within a few days after the culture becomes confluent andnet increase in cell number has ceased. Cells capable of formingfoci in the standard assay also arise even if cultures are passagedat low density, if the serum concentration in the passages ismaintained at a low level. Unlike mutations, therefore, "spontaneous" transformation is not really spontaneous but occurs

only when the appropriate inducing conditions obtain.2. Although the capacity to form foci arises in only a fraction

(less than 1%) of the population repeatedly passaged at lowdensity in low serum, the bulk of the population gains thecapacity to multiply to higher saturation densities. Each cellalso gives rise to a clone with an increased probability to formfoci in the standard assay. The involvement of the bulk of thepopulation in these changes in growth capacity is inconsistentwith a mutational origin.

3. There are many degrees of adaptation to multiply in low

serum concentrations or at high cell densities. We cannot, infact, exclude the possibility that the variation in these cellularcapacities is continuous and therefore distinctive from the discrete nature of mutations.

4. The capacities for focus formation and tumor developmentare reversible for the entire population of transformed cells.

Although adaptation criteria 1 and 4 have been demonstratedin cells other than the NIH 3T3 line, it is only in this cell thatall 4 criteria have been demonstrated.

Speculative Mechanism of Carcinogenesis. Given the evidencethat the growth properties of a high proportion of cells treatedwith carcinogens are heritably altered, the question arises ofhow such pervasive change in whole populations might beelicited. First, however, it must be noted that heritability amongsomatic cells of the capacity for tumor formation differs inseveral respects from that associated with transmission of specific gene mutations in organisms. Not only does transformation in culture occur with high frequency in response to carcinogens but the phenotype is heterogeneous, involves the gradualloss of many of the differentiated properties of the treated cells,and can be prevented in the 10T'/2 line by protease inhibitors,which have no effect on mutation frequency in mutagen-treatedcells (111). It appears, therefore, that proteases play an essentialrole in the initiation of transformation in lOT'/z cells.

Protease activity in yeast cells responds to a change in substrates. Yeast cells multiply exponentially until they exhausttheir main substrate, glucose, and remain at a constant numberuntil they start utilizing another substrate (112). The secondphase of growth is much slower than the first, and it is accompanied by a 300-fold increase in protease activity. An inactivating mutation in a key protease results in a failure of otherproteolytic and hydrolytic enzymes to be processed to an activeform. If a mutant cell is exposed to wild type cytoplasm throughmating or heterokaryon formation, the mutant segregants canmaintain indefinitely the wild type content of active proteasesand hydrolases normally dependent on the now inactive protease (112). The yeast system demonstrates clearly that analtered phenotype can be transmitted to progeny by epigeneticmeans without the involvement of cytoplasmic genes (113). Thedemonstration of the epigenetic persistence of wild type protease content may have particular significance for neoplastictransformation since: (a) the initiated state in lOT'/z cells can

be reversed by protease inhibitors (99, 111); (b) there is a largeincrease in protease content in carcinomas and their surrounding normal appearing tissue of the carcinogen-treated hamstercheek pouch (96); (c) the addition of proteases to density-inhibited chick embryo cells stimulates them to multiply (114);(d) a serine protease that is assayed as an activator of plasmin-ogen is elevated in primary tumor cells, in cells transformed byviruses, and in a number of tumorigenic cell lines (115); (e) theinduction by TPA of secretion of the protease which activatesplasminogen has been used to order histologically distinctclasses of human colonie adenomas from the most benign tothe most advanced premalignant state (116). It seems plausible,therefore, that persistent perturbation of the concentration orlocalization in cells of proteases and other hydrolytic enzymesmight result in a transmissible defect of growth regulation. Itis not unlikely that such perturbation would increase the geneticinstability of the cells, introducing genetic changes which mightcontribute to tumor progression.

It remains a question what the target for carcinogens mightbe that could initiate and propagate such epigenetic changes incells. One possibility is carcinogen-induced damage to the mem-

2757




branes of lysosomes which contain most of the cellular hydro-lases. It is noteworthy that UV light, which is a common skincarcinogen, can in very small doses cause hemolysis of humanerythrocytes which are virtually without DNA or RNA (117).The hemolysis, which represents the leakage of hemoglobinfrom the erythrocyte, is apparently the result of a direct effectof UV light on its membrane, and it has been conjectured thatthe mechanism underlying the changes in skin during carcino-genesis by UV light is closely comparable to that which increases the permeability of the erythrocyte (117). Mitochondriaof yeast cells are also disrupted by UV light (118). This effectis of special interest in cancer where the most commonlyobserved biochemical change is increased glycolysis, which wasconsidered a regulatory response to defective respiration (119).A relation between damaged mitochondrial membranes anddefective respiration became quite plausible with the advent ofthe chemiosmotic model of oxidative phosphorylation and itsdependence on transmembrane gradients of electrons and protons (120). Nongenetic changes in surface structures can beperpetuated for hundreds of cell generations, as has been demonstrated by direct visualization in Paramecium (121). Othermethods must be developed to demonstrate long term membrane effects in metazoan cells; thus the proposal is as yethypothetical. It is an interesting fact, however, that the carcinogenic polycyclic aromatic hydrocarbons are highly lipophilicand probably localize to a large extent in membranes (122).Although speculative, the suggestion of carcinogenesis throughmembrane damage would appear to be testable by studies ofthe effects of carcinogens on, for example, membrane trafficking, cell permeability, erythrocyte hemolysis, protease distribution, and mitochondrial function. It is well to keep in mind,however, that the definitive evidence in cell culture against amutational origin of both the initiation and promotion stagescame from quantitative experiments on the growth behavior ofliving cells. Much remains to be done at that level to obtain afuller understanding of neoplastic transformation.

General Considerations. As noted earlier, the two stage modelof cancer refers to the manipulations used in experimentalcarcinogenesis, but many more than two steps are involved (49).Foulds (123) believed that "initiation establishes an expanse or

region of incipient neoplasia that is coextensive with the areaof exposure to carcinogenic action and has a permanent, repl-

icable new capacity for neoplastic development which can beaugmented by repetition of the carcinogenic stimulus." In this

view, the initiated cells, as observed in the classical skin carcinogenesis model (17, 78), are not dormant tumor cells but arecells with an increased reactivity to further applications ofcarcinogens or of promoters. The changes effected by initiationare not restricted to the places where tumors emerge at a latertime. Foulds' view coincides with the evidence presented here

that the carcinogenic stimulus affects the entire population ofexposed cells, although only a small fraction progress to theovertly transformed state. If the initiation step is seen as endowing cells with an increased reactivity, the promotional treatment may differ only in degree rather than in kind; i.e., promoting agents may be effective only on cells previously sensitized by a more potent, initiating agent. That this is the case issuggested by the previously mentioned observation that mouseepidermis initiated by treatment with DMBA enters DNAsynthesis more rapidly when treated with croton oil than doesnoninitiated epidermis (95). Similarly, TPA fails to induce thesecretion of plasminogen activator in primary cultures of normal human colon but does so in cultures of adenoma cells with

increasing efficiency depending on their progression to malignancy (116). It has also been reported that tumor promotersenhance the proliferation of preneoplastic rat liver cells to asignificantly greater extent than normal unaltered liver cells(124). Finally, there is the finding discussed earlier that TPAhas the same transforming effect on 10T'/2 cells that havereceived 100 rads of X-rays as does raising the X-ray dose to400-600 rads (100). Thus, promoting and initiating agents mayultimately produce the same result in cells, although the pathway through which they act may differ. That only a rare cell inan initiated field is the actual antecedent of the tumor wouldfollow from the intrinsic heterogeneity of growth capacities inthe cell population (102, 103, 125, 126).

An interpretation similar to that developed here was proposed by Blum to account for epidermal cancer induced in miceby UV light (117, 127). Results from a large series of carefullyquantitated experiments indicated that successive doses of UVlight progressively accelerate the proliferation of the epidermalcells. Blum concluded that the results could not be explainedby separable induction and growth phases or by somatic mutation. Rather, the tumor cells could be considered as differingfrom normal cells only in the quantitative sense of proliferatingmore rapidly in the regulated context of the epidermis witheach successive exposure to UV light. Although UV light undoubtedly damages cells, the major response of the epidermisis an increase in the rate of mitosis and a consequent increasein thickness of the skin (117). If the radiation is not repeated,the epidermis returns to its normal thickness. The increasedmitosis can be viewed as an overshoot of the adaptive responseto cell damage. If the radiation is repeated frequently, the skinremains thickened and carcinomas and sarcomas appear withina few months. Blum designated the cellular response to therepeated UV irradiation "progressive acceleration of growth."

He indicated that all the cells exposed to the treatment wereincrementally altered but noted that very slight differences intheir accelerated growth rates would, over a period of months,result in the progeny of one cell outgrowing the others to forma tumor. In fact, however, there were signs of the beginning ofseveral tumors on as small an irradiated expanse as the mouseear, and sometimes carcinomas and sarcoma tissue were foundin the same tumor mass (117).

Progressive acceleration of growth can account for the progressive increase in number and thickness of transformed fociof NIH 3T3 cells repeatedly passaged at high density (110) orin stepwise decreases in serum concentration (105). To thisshould be added the selective effect of maintaining cells underconditions of moderate growth constraint. It has been proposedthat such constraint selects not only for individual cells withincreased growth capacity but also for metabolic states withincells that are subject to continuous phenotypic fluctuation ingrowth capacity (102, 105, 110, 126). Successive repetitions ofthis process would lead to relatively stable increases in growthcapacity under conditions of moderate constraint, and the process was called "progressive state selection" (102). Those con

straints, which are known to lower metabolic rates as well asgrowth, could lead to an increase in protease, as occurs in yeastwith the transition from exponential growth to the stationaryphase, and with the switch from rich to minimal medium (128).Such changes in previously sensitized, i.e., initiated, cells wouldlead to increased growth capacity and heterogeneity, just asexposure to a carcinogen or promoter would, thereby resultingin repeated selection of cells with the greatest growth potentialand ultimately in neoplasia. The combined effects could be

2758




characterized as: destabilization + state selection â€”Â»progressivegrowth acceleration.

The adaptive nature of the response is illustrated in theprotective value of the nodules from a liver initiation-promotion

regimen against the lethal effect of xenobiotics in the rat (63)and in the protection against sunlight provided by the thickening of epidermis after UV irradiation (117). A generalizedmodel for adaptation can be inferred in the continued multiplication and function of focus-producing cells in confluent cultures where the rest of the cells become strongly inhibited, andmany die. Unfortunately, the adaptive response reveals its lessbeneficent aspect in neoplasia when the stress becomes toosevere or prolonged. As one prophetic investigator put it some15 years ago, "oncogenesis may be adaptive ontogenesis" (129).

Acknowledgments

We greatly appreciate the manuscript preparation by Lori Cutler andDawn Davidson.

References

1. Sim man. O. Immunodepression and malignancy. Adv. Cancer Res.. 22:261-422. 1975.

2. Welch. W. H. Adaptation in Pathological Processes. Baltimore: JohnsHopkins Press. 1937.

3. Forbus, W. D. Reaction to Injury, Vols. 1 and 2. Baltimore: Williams &Wilkins. 1943. 1952.

4. Nicolson, G. W. deP. Studies on Tumor Formation. London: Butterworth&Co., 1950.

5. Dean, A. C. R., and Hinshelwood, C. Growth, Function and Regulation inBacterial Cells. Oxford, United Kingdom: Clarendon Press, 1966.

6. Marker!. C. L. Neoplasia: a disease of cell differentiation. Cancer Res., 28:1908-1914, 1968.

7. Knox. W. E. Enzymes, adaptation and homeostasis. N. Engl. J. Med.. 295:784-785. 1976.

8. Potter. V. R. Physiological adaptation at the molecular level: the frontierwhere research on differentiation and malignancy meet. Perspect. Biol.Med., 24: 524-542. 1981.

9. Farber, E., and Cameron, R. The sequential analysis of cancer development.Adv. Cancer Res., 31: 125-226, 1980.

10. Reddy, E. P., Skalka. A. M., and Curran, T. The Oncogene Handbook.Amsterdam: Elsevier/North-Holland Biomedicai Press, 1988.

11. Gonzalez, F. J. The molecular biology of cytochrome P450s. Pharmacol.Rev.. 40: 243-288, 1989.

12. Burt. R. K.. Garfield. S.. Johnson. K., and Thorgeirsson. S. S. Transformation of rat liver epithelial cells with v-H-ras or v-ra/causes expression ofM DR-1, glutathione-.S'-transferase P and increased resistance to cytotoxicchemicals. Carcinogenesis (Lond.), 9: 2329-2332, 1988.

13. Conney, A. H. Pharmacological implications of microsomal enzyme induction. Pharmacol. Rev.. 19: 317-366. 1967.

14. Schulte-Hermann, R. Induction of liver growth by xenobiotic compoundsand other stimuli. CRC Crit. Rev. Toxicol., 3: 97-158, 1974.

15. Schulte-Hermann, R. Reactions of the liver to injury. Adaptation. In: E.Farber and M. M. Fisher (eds.). Toxic Injury of the Liver, pp. 385-444.New York: Marcel Dekker. Inc., 1979.

16. Talalay, P.. and Prochaska, H. J. Mechanisms of induction ofNAD(P)H:quinone reducÃase.Chem. Ser., 27A: 61-66, 1987.

17. Wattenberg, L. W. Chemoprevention of cancer. Cancer Res.. 45: 1-8, 1985.18. Kalant, H., Roschlau. W. H. E., and Sellers. E. M. Principles of Medical

Pharmacology, Ed. 4. pp. 47-53. Toronto, Ontario. Canada: University ofToronto Press, 1985.

19. Neben, D. W. The 1986 Bernard B. Brodie Award Lecture: the geneticregulation of drug-metabolizing enzymes. Drug. Metab. Dispos., 16: 1-7,1986.

20. Neben, D. Wâ€žand Gonzalez, F. J. P450 genes: structure, evolution andregulation. Annu. Rev. Biochem.. 56: 945-993, 1987.

21. Conney, A. H. Induction of microsomal enzymes by foreign chemicals andcarcinogenesis by polycyclic aromatic hydrocarbons. G. H. A. Clowes Memorial Lecture. Cancer Res.. 42: 4875-4917, 1982.

22. Guengerich. F. P. Roles of cytochrome P-450 enzymes in chemical carcinogenesis and cancer chemotherapy. Cancer Res.. 48: 2946-2954. 1988.

23. Kushner, I. The acute phase response: an overview. Methods Enzymol.,163: 373-383, 1988.

24. Schlesinger. M. J., Ashburner, M., and Tissieres, A. (eds.). Heat Shockfrom Bacteria to Man. Cold Spring Harbor, NY: Cold Spring HarborLaboratory. 1982.

25. Carr, B. J., Huang, T. H.. Buzin. C. H.. and Itakura. K. Induction of heatshock gene expression without heat shock by hepatocarcinogens and during

hepatic regeneration in the liver. Cancer Res., 46: 5106-5111, 1986.26. Schlesinger. M. J. Heat shock proteins: the search for function. J. Cell

Biol., 103: 321-325, 1986.27. Nover, L. (ed.). Heat Shock Response of Eukaryotic Cells. Berlin: Springer-

Verlag, 1984.28. Atkinson, B. G., and Waiden, D. B. (eds.). Changes in Eukaryotic Gene

Expression in Response to Environmental Stress. Orlando, FL: AcademicPress, 1985.

29. Sharma, R. N., Cameron, R. G., Farber. E., Griffin, M. J., Joly, J-G., andMurray, R. K. Multiplicity of induction patterns of rat liver microsomalmono-oxygenases and other polypeptides produced by administration ofvarious xenobiotics. Biochem. J., 182: 317-327, 1979.

30. Carr, B. K. Pleiotropic drug resistance in hepatocytes induced by carcinogens administered to rats. Cancer Res., 47: 5577-5583, 1987.

31. Bohr. V. A., Phillips. D. H., and Hanawalt, P. C. Heterogeneous DNAdamage and repair in the mammalian genome. Cancer Res., 47:6426-6436,1987.

32. Demple, B. Adaptive responses to genotoxic damage: bacterial strategies toprevent mutation and cell death. BioEssays. 6: 157-160, 1987.

33. Lindahl, T., Sedgewick. D., Sekiguchi, M., and Nakabeppu. Y. Regulationand expression of the adaptive response to alkylating agents. Annu. Rev.Biochem., 57: 133-157, 1988.

34. Sanear, A., and Sanear, G. B. DNA repair enzymes. Annu. Rev. Biochem.,57:29-67, 1988.

35. Farber, E., and Sarma, D. S. R. Hepatocarcinogenesis: a dynamic cellularperspective. Lab. Invest.. 56: 4-22, 1987.

36. Farber. E. Clonal adaptation during carcinogenesis. Biochem. Pharmacol.,39: 1837-1846, 1990.

37. Foulds, L. Neoplastic Development, Vol. 2. New York: Academic Press,1975.

38. Thomas, L. Discussion. In: H. S. Lawrence (ed.). Cellular and HumoralAspects of Hypersensitivity. pp. 529-532. New York: Harper and Row,1959.

39. Prehn, R. T. Tumor progression and homeostasis. Adv. Cancer Res., 23:203-236, 1976.

40. Farber, E. The pathology of experimental liver cell cancer. In: H. M.Cameron, D. A. Linsell. and C. P. Warwick (eds.). Liver Cell Cancer, pp.243-277. Amsterdam: Elsevier/North-Holland Biomedicai Press. 1976.

41. Haddow, A. Cellular inhibition and the origin of cancer. Acta UniÃ³Intern.Contra Cancrum, 3: 342-352, 1938.

42. Abanobi, S. E., Lombardi. B., and Shinozuka, H. Stimulation of DNAsynthesis and cell proliferation in the liver of rats fed a choline-devoid dietand their suppression by phÃ©nobarbital.Cancer Res., 42:412-415, 1982.

43. Barbasen, H.. Rassenfosse. C., and Betz, E. H. Promotion mechanism ofphÃ©nobarbitaland partial hepatectomy of DENA hepatocarcinogenesis cellkinetics effect. Br. J. Cancer, 47: 517-525, 1983.

44. Eckl, P. M., Meyer, S. A., Whitcomb, W. R., and Jirtle, R. L. PhÃ©nobarbitalreduces EGF receptors and the ability of physiological concentrations ofcalcium to suppress hepatocyte proliferation. Carcinogenesis (Lond.), 9:479-483, 1988.

45. Tatematsu, M., Aoki, T.. Kagawa, M., Mera, Yâ€žand Ito, N. Reciprocalrelationship between development of glutathione 5-transferase positive liverfoci and proliferation of surrounding hepatocytes in rats. Carcinogenesis(Lond.), 9:221-225, 1988.

46. Laconi, E., Li, F., Semple. E., Rao, P. M., Rajalakshmi, S., and Sarma, D.S. R. Inhibition of DNA synthesis in primary cultures of hepatocytes oforotic acid. Carcinogenesis (Lond.), 9: 675-677, 1988.

47. Pichiri-Coni. G., Coni. P.. Laconi, E., Schwarze. P. E.. Seglen, P. O., Rao,P. M.. Rajalakshmi. S.. and Sarma, D. S. R. Studies on the mitoinhibitoryeffect of orotic acid on hepatocytes in primary culture. Carcinogenesis(Lond.), //: 981-984, 1990.

48. Lea, M. A., Luke, A., Assad. S., and Ayyala. S. Inhibitory effects of orotateon precursor incorporation into nucleic acids. Chem.-Biol. Interact., 75:49-59, 1990.

49. Farber. E. Sequential events in chemical carcinogenesis. In: F. F. Becker(ed.), Cancer: a Comprehensive Treatise, Ed. 2, Vol. 1, pp. 485-506. NewYork: Plenum Publishing Corp., 1982.

50. Lee, L. W., Tsao, M-S., Grisham, J. W.. and Smith, G. J. Emergence ofneoplastic transformants spontaneously or after exposure to A'-methyl-A''-nitro-A'-nitrosoguanidine in populations of rat liver epithelial cells culturedunder selective and non-selective conditions. Am. J. Pathol., 135: 63-71.1989.

51. Tsao, M-S.. Shepherd. J., and Batist. G. Phenotypic expression in spontaneously transformed cultured rat liver epithelial cells. Cancer Res., 50:1941-1947, 1990.

52. Farber, E. Resistance phenotype in the initiation and promotion of chemicalhepatocarcinogenesis. Chem. Ser., 27A: 131-133. 1987.

53. Buchmann. A.. Kuhlmann, W.. Schwarz, M., Kunz, W., Wolf, C. R., Moll,E., Freidberg, T., and Oesch, F. Regulation of expression of four cyto-chromes P-450 isozymes, NADPH-cytochrome P-450 reducÃase,glutathione transferases B and C and microsomal epoxide hydrolase in preneoplasticand neoplastic lesions in rat liver. Carcinogenesis (Lond.), 6:513-521,1985.

54. Roomi, M. W.. Ho. R. K., Sarma, D. S. R., and Farber, E. A commonbiochemical pattern in preneoplastic hepatocyte nodules generated in fourdifferent models in the rat. Cancer Res., 45: 564-571, 1985.

55. Pilot. H. C.. Barsness. L.. Goldsworthy. T., and Kitagawa, T. Biochemicalcharacterization of stages of hepatocarcinogenesis after a single dose of

2759




diethylnitrosamine. Nature (Lond.). 271: 456-458. 1978.56. Ogawa. K.. Soll, D. B., and Farber, E. Phenotypic diversity as an early

properly of putative preneoplastic hepatocyte populations in liver carcino- 84.genesis. Cancer Res.. 40: 725-733, 1980.

57. Cowan, K. H., Batist, Gâ€žTulpule, H., Sinha. B.. and Myers, C. E. Similarbiochemical changes in human breast cancer cell and carcinogen-induced 85.resistance to xenobiotics. Proc. Nati. Acad. Sci. USA. 83:9328-9332, 1986.

58. Thorgeirsson, S. S., Huber, B. E.. Sorell, S.. Fajo. A., Pastan, L, and 86.Gottesman, M. M. Expression of the multidrug resistant gene in hcpatocar-cinogencsis and regulatory rat liver. Science (Washington DC), 236: 1120-1122, 1987. 87.

59. Fairchild, C. R., Ivy, S. P., Rushmore, T., Lee. G.. Koo, P.. Goldsmith, M.E.. Myers, C. E., Farber. E., and Cowan, K. H. Carcinogen-induced mdroverexpression in association with xenobiotie resistance in rat preneoplastic 88.liver nodules and hepatoccllular carcinomas. Proc. Nati. Acad. Sci. USA.Â«4:7701-7705. 1987.

60. Ogawa. K.. Mediine, A., and Farber, E. Sequential analysis of hepatic 89.carcinogenesis: the comparative architecture of preneoplastic. malignant,prenatal, postnatal and regenerating liver. Br. J. Cancer. 40:782-790. 1979.

61. Enomoto. K., and Farber, E. Kinetics of phenotypic maturation or remod- 90.cling of hyperplastic nodules during liver carcinogenesis. Cancer Res., 42:2330-2335. 1982. 91.

62. Tatematsu, M., Nagamine. Y.. and Farber. E. Redifferentiation as a basisfor remodeling of carcinogen-induced hepatocyte nodules to normal appear- 92.ing liver. Cancer Res., 43: 5049-5058. 1983.

63. Harris, L., Morris, L. E., and Farber. E. The protective value of a liver 93.initiation-promotion regimen against the lethal effect of carbon tetrachlo-ride. Lab. Invest.. 61: 467-470. 1989.

64. Gans. J. H. Diethylnitrosamine-induced changes in mouse liver morphologyand function. Proc. Soc. Exp. Biol. Med.. 153: 116-120. 1976. 94.

65. Rinaudo. J. A. S., and Farber, E. The pattern of metabolism of 2-acetyla-minofluorenc in carcinogen-induced hepatocyte nodules in comparison tonormal liver. Carcinogenesis (Lond.), 7: 523-528, 1986. 95.

66. Augenlicht. L. H., Wahrman, M. Z.. Halsey. H.. Anderson. L.. Taylor. J.,and Lipkin, M. Expression of cloned sequences in biopsies of human colonietissue and in colonie carcinoma cells induced to differentiate in vitro. Cancer 96.Res., 47:6017-6021. 1987.

67. Clark, W. H., Jr.. Elder. D. E.. and Guerry. D. IV. The pathogencsis andpathology of dysplastic nevi and malignant melanoma. In: E. Farmer and 97.A. Hood (eds.). The Pathology of the Skin. pp. 684-756. Englewood Cliffs.NJ: Appleton-Century-Crofts. 1990.

68. Williams, J. B., Lee, A. Y. H.. Cameron. R. G.. and Picket!. C. B. Rat liverNAD(P)H:quinone reducÃase.J. Biol. Chem.. 261: 5524-5528. 1986. 98.

69. Ding, V. D-H., Cameron, R., and Pickett, C. B. Regulation of microsomalxenobiotic epoxide hydrolase messenger RNA in persistent hepatocytenodules and hepatomas induced by chemical carcinogens. Cancer Res., 50: 99.256-260, 1990.

70. Rao, P. M.. Anthony. A.. Rajalakshmi, S., and Sarma. D. S. R. Studies onhypomethylation of liver DNA during early stages in chemical carcinogen- 100.esis in the rat liver. Carcinogenesis (Lond.). 10: 933-937. 1989.

71. Coni. P.. Pang. J.. Pichiri-Coni, G.. Hsu, S.. Rajalakshmi. S.. and Sarma.D. S. R. Hypomethylation of 3-hydroxy, 3-methylglutaryl coenzyme A 101.(HMG-CoA) reducÃasegene and its expression during hepatocarcinogenesis.Proc. Am. Assoc. Cancer Res.. 31: 162. 1990.

72. Koo. P.. Nagai, M.. Ghoshal. A., Rao. P.. and Farber. E. Comparison 102.between the regulation of the placenta! form of glutathione .V-transferase(GST-7-7) expression in rat liver in precancerous nodules and in lead treatedliver. Proc. Am. Assoc. Cancer Res.. 30: 187. 1989. 103.

73. Fox, M., and Radicic, M. Adaptational origin of some purine-analogueresistant phenotypes in cultured mammalian cells. MutÃ¢t.Res., 49: 275-296. 1978. 104.

74. Rothenbcrg. M.. and Ling. V. Multidrug resistance: molecular biology andclinical relevance. J. Nati. Cancer Inst.. 81: 907-910. 1989.

75. Farber, E. Cancer: a disease of adaptation? Proc. Am. Assoc. Cancer Res.. 105.50:672-673. 1989.

76. Rous, P.. and Kidd, J. G. Conditioned neoplasms and subthreshold neo-plastic states. J. Exp. Med.. 73: 365-390. 1941. 106.

77. Berenblum, I., and Shubik, P. The persistence of latent tumor cells inducedin the mouse skin by a single application of 9:10-dimethyl-l:2-ben7.anthra- 107.cene. Br. J. Cancer, 3: 384-386. 1949.

78. Berenblum. 1. Sequential aspects of chemical carcinogenesis: skin. In: F. F.Becker (ed.). Cancer: a Comprehensive Review, pp. 451-484. New York: 108.Plenum Publishing Corp.. 1982.

79. Yuspa. S. H.. and Poirier, M. C. Chemical carcinogenesis: from animalmodels to molecular models in one decade. Adv. Cancer Res., 50: 25-70, 109.1988.

80. Berwald, Y., and Sachs, L. In vitro transformation of normal cells to tumorcells by carcinogenic hydrocarbons. J. Nati. Cancer Inst., 35: 641-661. 110.1965.

81. Mondai. S., and Heidelberger. C. In vitro malignant transformation bymelhylcholanthrene of the progeny of single cells derived from C3H mouse 111.prostate. Proc. Nati. Acad. Sci. USA. 65: 219-229. 1970.

82. Reznikoff. C. A.. Bertram, J. S., Brankow. D. W.. and Heidelberger. C. 112.Quantitative and qualitative studies of chemical transformation of clonedC3H mouse embryo cells sensitive to postconfluencc inhibition of celldivision. Cancer Res.. 33: 3239-3249. 1973.

83. Kennedy, A. R., Fox, M.. Murphy, G.. and Little, J. B. Relationship between 113.

2760

X-ray exposure and malignant transformation in C3H lOT'/i cells. Proc.Nati. Acad. Sci. USA, 77: 7262-7266. 1980.Haber. D. A., Fox, D. A.. Dynan. W. S.. and Thilly. W. F. Cell densitydependence of focus formation in the C3H 10T1/; transformation assay.Cancer Res.. 37: 1644-1648. 1977.Terzaghi. M.. and Little, J. B. X-irradiation-induced transformation in aC3H mouse embryo-derived cell line. Cancer Res. 36: 1367-1374. 1976.Han, A., and Elkind, M. M. Transformation of mouse C3H/IOT'/2 cells bysingle and fractionated doses of \-rays and fission-spectrum neutrons.Cancer Res.. 39: 123-130. 1979.Fernandez. A.. Mondai. S., and Heidelberger. C. Probabalistic view of thetransformation of cultured C3H 10T'/Â¡mouse embryo fibroblasts by 3-methylcholanthrene. Proc. Nati. Acad. Sci. USA. 77: 7272-7276, 1980.Huband. J. C.. Abernethy. D. J.. and Boreiko. C. J. Potential role oftreatment artifact in the effect of cell density upon frequencies of C3H/10T'/2 cell transformation. Cancer Res.. 45: 6314-6321. 1985.Suk, W. A.. Poiley. J. A.. Raineri. R.. Steuer. A. F., and Tennant. R. W.Chemical enhancement of survival in aggregation of retroviral-infected ratcells: An interlaboratory comparison. Em. Mutagen., 7: 727-746, 1985.Sinclair, W. K. X-ray induced heritable damage (small colony formation) incultured mammalian cells. Radial. Res., 21: 584-611. 1964.Beer, J. Z. Heritable lesions affecting proliferation of irradiated mammaliancells. Adv. Radial. Biol.. X: 363-417. 1979.Frank. J. P.. and Williams, J. R. X-ray induction of persistent hypersensi-livity to mutation. Science (Washington DC). 216: 307-308. 1982.Goodhead. D. T. Deductions from cellular studies of inactivation. mutage-nesis and transformation. In: J. D. Boice Jr. and J. F. Fraumeni Jr. (eds.)Radiation Carcinogenesis: Epidemiology and Biological Significance, pp.369-385. New York: Raven Press. 1984.Kennedy, A. R., Cairns, J., and Little, J. B. Timing of the steps in transformation of C3H 10T'/2 cells by X-irradiation. Nature (Lond.). 307: 85-86,1984.Frankfurt, O. S., and Raitcheva, E. Fast onset of DNA synthesis stimulatedby tumor promoter in mouse epidermis at the initiation stage of carcinogenesis. J. Nati. Cancer Inst., 51: 1861-1864. 1973.Messadi, D. V., Billings, P., Shklar, G., and Kennedy, A. R. Inhibition oforal carcinogenesis by a protease inhibitor. J. Nati. Cancer Inst., 76: 447-452. 1986.\\atanabe, H., Tanner. M. A.. Domann. F. E.. Gould, M. N., and Clifton.K. H. Inhibition of carcinoma formation and of vascular invasion in graftsof irradiation-initiated thyroid clonogens by unirradiated thyroid cells. Carcinogenesis (Lond.). 9: 1329-1335. 1988.Zhang, R.. Haag, J.. and Gould, M. N. Quantitating the frequency ofinitiated cells in in situ NMU-exposed mammary gland. Proc. Am. Assoc.Cancer Res., 31: 138. 1990.Kennedy. A. The conditions for the modification of radiation transformationin vitro by a tumor promoter and protease inhibitors. Carcinogenesis(Lond.). 6: 1441-1445. 1985.Kennedy. A. R., and Little. J. B. Investigation of the mechanism forenhancement of radiation transformation in vitro by 12-O-letradecanoylphorbol-13-acetate. Carcinogenesis (Lond.). /: 1039-1047, 1980.Rubin. H.. and Xu. K. Evidence for the progressive and adaptive nature ofspontaneous transformation in the NIH 3T3 cell line. Proc. Nati. Acad. Sci.USA, 86: 1860-1864, 1989.Rubin, A. L., Yao, A., and Rubin. H. Relation of spontaneous transformation in cell culture to adaptive growth and clonal heterogeneity. Proc. Nati.Acad. Sci. USA, 87: 482-486, 1990.Grundel. R., and Rubin. H. The effect of interclonal heterogeneity on theprogressive, confluence-mediated acquisition of the focus forming phenotypein NIH 3T3 populations. Cancer Res.. 51: 1003-1013, 1991.Rubin. A. L. Suppression of transformation by and growth adaptation tolow concentrations of glutamine in NIH 3T3 cells. Cancer Res.. 50: 2832-2839. 1990.Yao. A.. Rubin. A., and Rubin. H. Progressive state selection of cells in lowserum promotes high density growth and neoplastic transformation in NIH3T3 cells. Cancer Res.. 50: 5171-5176. 1990.Bucher. N. L. Regeneration of mammalian liver. Int. Rev. Cytol., 15: 245-300. 1963.Michalopoulos, G., Gianculli, H. D., Novotny, A. R., Kligerman, A. D.,Strom, S. C.. and Jirtle. R. L. Liver regeneration studies with rat hepatocytesin primary culture. Cancer Res., 42: 4673-4682, 1982.Rabinowitz, Z., and Sachs, L. The formation of variants with a reversion ofproperties of transformed cells. II. In vitro formation of variants frompolyoma-transformed cells. Virology, 38: 343-346. 1969.Brouty-Boye. D.. Gresser, I., and Baldwin. C. Reversion of the transformedphenotype to the parental phenotype by subcultivation of X-ray transformedC3H/10T'/2 cells at low cell density. Int. J. Cancer. 24: 253-260, 1979.Rubin. A. L., Arnstein, P.. and Rubin. H. Physiological induction andreversal of focus formation and tumorigcnicity in NIH-3T3 cells. Proc. Nati.Acad. Sci. USA. 87: 10005-10009. 1990.Kuroki. T.. and Drevon, C. Inhibition of chemical transformation in C3H/lOT'/i cells by protease inhibitors. Cancer Res.. 39: 2755-2761, 1979.Jones. E. W'.. Moehle. C.. Kolodny, M.. Aynardi. M., Park, F.. Daniels, L.,

and Gartlow. S. Genetics of vacuolar proteases. In: 3. Hicks (ed.). YeastCell Biology. UCLA Symposia on Molecular and Cell Biology. New Series,pp. 505-518. New York: Alan R. Liss, Inc., 1986.Zubenko. G. S., Park. F. J.. and Jones. E. W. Genetic properties of the




PEP4 locus in Saccharomyces cereviseae. Genetics. 102: 649-690, 1982.114. Sefton, B. M.. and Rubin, H. Release from density dependent growth

inhibition by proteolytic enzymes. Nature (Lond.), 227: 843-845, 1970.115. Quigley, J. P. Phorbol ester-induced morphological changes in transformed

chick fibroblasts: evidence for direct catalytic involvement of plasminogenactivator. Cell. 17: 131-141. 1979.

116. Friedman, E.. Urmacher. C., and \\inawer, S. A model for human coloncarcinoma evolution based on the differential response of cultured preneo-plastie. premalignant and malignant cells to 12-0-tetradecanoylphorbol-13-acetate. Cancer Res., 44: 1568-1578, 1984.

117. Blum, H. F. Carcinogenesis by Ultraviolet Light. Princeton. NJ: PrincetonUniversity Press. 1959.

118. Saracheck, A., and Townsend, G. F. The disruption of mitochondria ofSaccharomyces by ultraviolet radiation. Science (Washington DC), 117:31-

33. 1953.119. Aisenberg, A. C. The Glycolysis and Respiration of Tumors. New York:

Academic Press, 1961.120. Mitchell, P. Coupling of phosphorylation to electron and hydrogen transfer

by a chemi-osmotic type of mechanism. Nature (Lond.). 191: 144-148,1961.

121. Bcisson. J., and Sonneborn, T. M. Cytoplasmic inheritance of the organi

zation of the cell cortex in Paramecium aurelia. Proc. Nati. Acad. Sci. USA,5^:275-282. 1965.

122. Willmer, E. N. Steroids and cell surfaces. Biol. Rev. (Cambridge), 36: 368-398, 1961.

123. Foulds. L. Neoplastic Development. Vol. 1. New York: Academic Press,1969.

124. Schulte-Hermann, R.. Onde, G., Schuppler. J., and Timmermann-Trosi-ener, 1. Enhanced proliferation of putative preneoplastic cells in rat liverfollowing treatment with the tumor promoters phÃ©nobarbital,hexachloro-cyclohexane. steroid compounds, and nafenopin. Cancer Res., 41: 2556-2562, 1981.

125. Grundel. R.. and Rubin. H. Maintenance of multiplication rate stability bycell populations in the face of heterogeneity among individual cells. J. Cell.Sci.. 91: 571-576. 1988.

126. Rubin. H. The significance of biological heterogeneity. Cancer MetastasisRev.. 9: 1-20. 1990.

127. Blum. H. F. On the mechanism of cancer induction by ultraviolet radiation.J. Nati. Cancer Inst.. //: 463-495, 1950.

128. Hansen, R. J., Switzer, R. L.. Hinze, Hâ€žand Holzer, H. Effects of glucoseand nitrogen source on the levels of proteinases, peptides and proteinaseinhibitors in yeast. Biochim. Biophys. Acta, 496: 103-114, 1977.

129. Nery, R. Carcinogenic mechanisms: a critical review and a suggestion thatoncogenesis may be adaptive ontogenesis. Chem.-Biol. Interact., 12: 145-169, 1976.

2761



1991;51:2751-2761. Cancer Res Emmanuel Farber and Harry Rubin Cellular Adaptation in the Origin and Development of Cancer

Updated version

http://cancerres.aacrjournals.org/content/51/11/2751.citation

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

[email protected] 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/51/11/2751.citationTo request permission to re-use all or part of this article, use this link



http://cancerres.aacrjournals.org/cgi/alerts

mailto:[email protected]



cellular adaptation in the origin and development of cancer1 · cellular adaptation in the origin...

Documents