Cancer and normal ageing

Mechanisms of Ageing and Development, 25 (1984) 269--283

269

Elsevier ScientificPublishers Ireland Ltd.

CANCER AND NORMAL AGEING

PETER EBBESEN The Institute of Cancer Research, Radiumstationen, DK-SO00Aarhus C (Denmark)

(Received May 10th, 1983) (Revision receivedNovember20, 1983) SUMMARY Cancer has a monoclonal origin in a pre-mitotic cell and usually a multistep pathogenesis. The initiation of tumor development and most stages in tumor progression involve point mutations, chromosomal rearrangements, and often changes in gene dosage. Simultaneously, there is continuing selection of the cell clones most resistant to growth-regulating substances and/or lacking specific immunologic markers. Normal ageing is partly pre-programmed as demonstrated by the constancy of the maximum survival time for a species under various external conditions resulting in different mean survival times. Emerging evidence of characteristic DNA changes in post-mitotic cells of old individuals may turn out to be part of the programmed changes. However, random accumulation of mutational defects in both pre- and post-mitotic cells is'an unavoidable consequence of physics and therefore contributes to normal ageing and the accompanying increase in cell diversity. Cancer incidence increases with age. Firstly, because extended exposure increases the risk of inflicting the DNA changes prerequisite to oncogenesis; secondly, because the progression from one malignant cell to detectable, tumor is a matter of 10-30 per cent of a species maximum life span; and thirdly, because some alterations characteristic of normal ageing increase the susceptibility to carcinogens. There probably is an overlap of etiologic/accelerating factors for cancer and ageing. Such aspects of normal ageing such as decline in DNA repair capacity and decline in cellular immune reactivity should facilitate induction and early growth of neoplasia. Age changes that counteract cancer development include (hormonal) loss of proliferative stimulation and depletion of the pool of immature cells at greatest risk.

K e y words: Aging; Cancer; Clonality; Diversity; Pre-programmed; Multistep

hypothesis 0047-6374/84/$03.00 Printed and Publishedin Ireland

© 1984ElsevierScientificPublishers Ireland Ltd.

270 INTRODUCTION

The risk of developing adulthood cancer increases rapidly with age. In most industrialized countries more than half the cancers are diagnosed in persons above 70 years of age. There are several reasons for this statistical correlation between cancer and advanced age. In the present survey we will discuss the probable mechanism of cancer development and those aspects of "normal" ageing we believe to be most relevant to the etiologic and pathogenetic bonds between these two biological processes. CANCER PATHOGENESIS AND ETIOLOGY

The pathogenesis of cancer is reasonably well studied. Although rarely inherited [1] its fundamental structural basis is alterations of the chromosomes. The malignant phenotype is transmitted from mother to daughter cells, morphologic alterations of chromosomes are demonstrable in most cases [2], and transforming D N A sequences have been identified in many cases. Tumor cell morphology [3] and tumor cell production of embryonic antigens [4] suggest that tumors arise from rather undifferentiated cells. Most, but not all neoplasms, are consistent with a monoclonal origin. Our best evidence of this is the uniform isoenzyme pattern for the X-linked enzyme glucose 6-phosphate dehydrogenase in tumors developed in females heterozygous for this enzyme. Normal somatic cells in such individuals are a mosaic, some expressing A and some B-type isoenzyme [5]. Most carcinogens are mutagenic in at least one test system [6] and often they are known to bind covalently to D N A [7]. In some cases the carcinogenic potential correlates inversely with the cells' ability to excise damaged bases from D N A [8]. This, together with the monoclonality favours the somatic mutation (lasting change in genotype) model for cancer initiation. An as yet unproven hypothesis is that human chromosomes may harbour movable genetic elements like those present in bacteria and that rearrangement of these elements may trigger malignancy [9]. It is even possible that such movable sequences in humans may express themselves as C-type virus [10]. As judged from chemical analysis, malignant transformation results in very few new proteins being produced [11]. This suggests that the important changes occur with the regulatory genes in contrast to changes of the structural genes directing synthesis of a particular protein. Production of carcinoembryonic antigens [4] and expression of differentiation antigens on some tumor cell membranes [12], furthermore, suggests that in particular the genes active in normal differentiation may be affected. There are a few facts that do not easily fit in with the somatic mutation hypothesis. For example, transfer of embryonic cells into an abnormal location may lead to development of teratocarcinomas [13] and transfer of teratocarcinoma cells into normal blastocysts results in development of tumor-free animals [14]. Normal differentiation, furthermore, may unleash the malignant phenotype

271 of cells that have previously acquired DNA lesions at an earlier stage of differentiation [15]. Normal differentiation may also suppress existing inalignant phenotypes, e.g. the rare spontaneous regression of certain childhood noeplasias [16]. The later normalization of phenotype is also possible in experimental systems by adding differentiation-inducing factors [17]. Reversible events thus clearly may determine the presence/absence of malignant phenotype. We are presently not well informed on the question of which noxious agents are of quantitative importance for cancer initiation. Unavoidable errors of normal mitosis and producion of free radicals such as superoxide OF during normal energy metabolism [18] are likely to play a role, since radical scavengers may reduce the tumor incidence in animals [19,20], but the big changes in cancer incidences experienced by migrant populations [21,22] suggest that external milieu factors are of paramount importance. Ultraviolet (UV) light and smoking contribute to about one fourth of cancers in industrialized countries, but inhaled chemicals and food also play a role [23]. The initial irreversible (initiating) change in tumor pathogenesis [24] seems to include a fixation of the cells at a low level of differentiation with concomitant preservation of a proliferative potential [25]. After initiation with a chemical or irradiation it usually takes 10 to 20 per cent of a species' maximum lifespan before a one gram tumor of 109 cells has emerged. This was most convincingly demonstrated among the survivors of Hiroshima and Nagasaki [26]. During this period of multiplication of abnormal cells, heterogeneity emerges in the tumor cell population due to new mutations to which initiated cells are prone [27], and in some cases differentiation of some cells [28]. The speed and course of tumor progression are influenced by compounds (promotors) that stimulate cell division and differentiation [29]. As a result of the emerging cellular diversity, a selection takes place which favours subpopulations non-responsive to growth-inhibitory mechanisms [30] and devoid of novel surface antigenic characteristics [31,32] that elicit an immune response to the tumor cells. This tendency to convergence of cell characteristics during rumor progression also applies to DNA. It seems that there are specific fine chromosome defects [2] which are very favourable in a particular tissue and therefore stable when first affected by chance. Examples are the reciprocal 8:14 translocation emerging in many human B iymphomas irrespective of the etiologic factor [33] and the increase of gene dosage (trisomy of chromosome 15) in murine lymphomas [34]. Regarding how many intermediary steps take place during cancer development, the classical approach has been to derive such information from epidemioiogy. The incidence rate of all cancers lumped together increases roughly with a power of age. A simplified analysis goes like this: Provided cancer develops in discrete steps occurring independently of each other, the chance of step one occurring within the period t years (age) is a constant P~ times t (P1 x t). The chance of all the required n steps taking place within the period t is (Pj x t)(P2 x

272

t ) . . . ( P n x t)= Kit'. The incidence rate I is obtained by differentiation with respect to t, I = K2 x t ~-~. Take the logarithm and we get log I = K3 × log t +/(4 (K = constant). The slope of the incidence curve in the log-log plot thus should give the number of putative pathogenetic steps minus 1 (usually 5 or 6 in all) [35]. If one includes in the mathematical formula a factor to account for exponential growth of already initiated (premalignant) cells one gets formulas that can also fit the curves of tumors with an incidence peak at early age [36]. However, using this approach the number of steps derived from the formula is as low as two. Direct experimental work with tumor induction in animals with chemicals or irradiation also points to induction through only a few steps (initiation, promotion) [24,37]. Recent results in molecular biology are also consistent with a few-step pathogenesis. There is evidence that all normal cells harbour DNA sequences (oncogenes) which upon activation give a cell a malignant phenotype [38,39]. According to one assumption these oncogenes are activated by alterations in nearby normal regulatory DNA sequences. Various types of carcinogens acting at random on DNA thus should unleash a malignant transformation when they happen to act on these regulatory sequences adjacent to the oncogenes (promotor effect) [40]. The rate of increase of cancer incidence declines at extreme high age. This is usually considered a result of an age-dependent change in an intrinsic factor, most likely a decline in the number of cells at risk [41]. AGE AND SUSCEPTIBILITY TO CARCINOGENS

There is general agreement that fetuses after completion of organogenesis, children and adolescents are more susceptible to cancer induction by some agents than are adults. Examples are adenocarcinomas of the vagina in offspring of mothers treated with stilbestrol during pregnancy [42], leukemia after X-ray irradiation of children [43] and breast cancer also after irradiation where the adolescent girl is most at risk [44]. Rapid cell proliferation in these age groups is likely to be a major reason for the high susceptibility to carcinogens at that age. For the adult person, it is less clear whether normal ageing influences susceptibility to the tumor-inducing effect of a carcinogen to which one has not previously been exposed. The human environment is so complex that we can hardly ever be sure of non-exposure to a certain chemical, yet several epidemiologic studies have dealt with the question. Most studies indicate that the risk of cancer development increases as a potens function of the person's age at first exposure to carcinogens, the risk being directly proportional to age raised to a potens: bladder cancer in dyestuff workers [45], skin warts alter tar exposure [46,47], lung cancer in asbestos workers [48], and sinus cancer in nickel workers [49]. Direct measurements of chromosome aberrations also show that more abnormalities are produced by in vitro carcinogen treatment of leukocytes from old than from young human donors [50]. The same relation to age for cancer development is reported for adult humans exposed to irradiation. The excess

.i

?"

-

273

cancer cases caused by the irradiation increase with age at first treatment. The ratio of observed cancer incidence to expected incidence, however, does not necessarily change with age, as the spontaneous incidence of many tumors also increases with age. Examples are cancer following treatment of ankylosing spondylitis [51,52] and uterine cancer after irradiation treatment of metropathia hemorrhagica [53]. Bizzozero et al. [26] first reported on leukemia in A-bomb survivors and found the incidence unrelated to age up to 60 years. However, subsequent studies by Jablon and Kato [54] and Beebe et al. [55] clearly demonstrated an increase in risk with increase in age at time of detonation. A lone example of another age-susceptibility relationship is the incidence of cancer of the lung in nickel workers which increases with age at first exposure up to 25 years of age and then, for unknown reasons, falls off again [49]. Turning now to animal experiments, the conditions are more well defined with regard to what chemical and what doses are involved. However, the dosage necessary to induce a few tumors using a manageable number of animals far exceeds the exposures humans encounter. The high carcinogen doses used in the animal experiments may easily conceal what differences in susceptibility there may exist when the carcinogen doses do not exceed detoxification and/or repair capacities [56]. With this in mind we will look at available experimental evidence. When senescent BALB/c mouse skin is exposed to small doses of carcinogens one finds an increase in susceptibility as compared to younger adult skin. This is the case both when tumors are induced with a small dose of 7-12-dimethylbenz(a)anthracene [57],/3-irradiation [58], and UV light [59]. Skin grafted from middle-aged to young recipients developed the same increased susceptibility to carcinogens with further ageing as non-grafted skin. The high susceptibility to carcinogens of senescent skin thus must derive from local, autonomously developing alteration in the skin itself. In studies with carcinogen doses inducing high incidence of colon carcinomas in C57B1 mice, Defries et al. [60] reported the same incidence in 6- and 22-month-old mice treated with 1,2-dimethylhydrazine. However, chemically induced colon cancer in mice, once established, advances more rapidly in old than in young animals [61]. Injecting 4-fluoronebiphenylacetamide resulted in kidney tumors preferentially in the older test animals [62]. When administered to young animals, polycyclic aromatic hydrocarbons mostly cause mammary tumors in rats [63]. As the same relation to age of treatment was found with N-nitroso-N-methylurea, which does not require metabolic activation, the age-related changes of importance probably occur within the mammary glands themselves [64]. The finding for chemical induction of rat mammary tumors is reminiscent of the pattern for human mammary tumor development following irradiation as already mentioned. Summerhayes and Franks [65] found in vitro transformation of bladder epithelium easier with bladder epithelium cultures established from old than with those established from young mouse donors. Also, in vitro ageing can be accompanied by increased susceptibility to a transforming chemical [66] but this model system has not been studied in much detail.

274

NORMAL AGEING

The mechanism of normal ageing is a subject much less extensively studied than the etiology and pathogenesis of cancer. By normal ageing we here understand those gradual changes in the adult organism which in a stable external environment will increase the risk of dying for all members of a species. This definition excludes specific disease entities because they are neither common to all members of a species nor necessarily gradual in development [67]. That it is reasonable to establish normal ageing as a biologic concept independent of disease is derived from many observations, among which are: (A) the uniformity of many age-related changes in individuals developing different chronic diseases, (B) the fixed maximum survival time of each species irrespective of external milieu [67], (C) the insignificant influence of the presence/absence of detectable chronic disease on the survival time of genetically closely related animals [68], and (D) the existence of genetic syndromes with some aspects of accelerated ageing [69]. Death intensities (L age incidence rates, risk of dying in a coming short time interval provided one is alive at start of that interval) are per definition a measure of normal ageing and by far the best studied signs of ageing. Usually the Markham Gompertz exponential equation is used to describe death intensities: I = Kt" e kt + K2, where t is age, e is the natural logarithm (about 2.7) and K~ and K2 are constants [70]. However, a good approximation to the observed death intensities is also obtainable by a power function I = K. t n, where t is age and both K and n are constants [71]. For both total mortality and mortality from one of the major diseases the incidence curves run nearly parallel to the incidence curves for cancer in semi-log graphs [72]. The epidemiology of ageing therefore can be used in favour of a multistep development of normal ageing in much the same way as cancer can be viewed. This interpretation of ageing "pathogenesis" is consistent with the gradual change of morphology occurring with age. Ageing at the cellular level is primarily a DNA-based phenomenon as demonstrated by nuclear exchange between old and young cells [73]. Ageing includes a component of pre-programmed development. The maximum lifespan of members of each species is fixed and the probability that a given in vitro grown normal fibroblast in Gm phase will go into a new cell cycle (transition probability) declines with increasing number of in vitro doublings in a manner characteristic of each species [74,75]. In vivo survival of erythrocytes, furthermore, very roughly correlates with maximum lifespan of the whole organism [76]. Drosophilia melanogaster exhibits age-dependent activity of the flight muscle. This may be another example of an intrinsically regulated "pre-programmed" event since non-flying mutant flies with rudimentary wings have identical changes in their enzyme activity throughout their life [77]. Cellular ageing, like most normal differentiation (cellular specialization), appears to be irreversible. Transfer of old cell to young environment only influences ageing of cells or tissues to a limited extent [57,78-80].

275 What processes in the DNA govern normal ageing is largely unknown. Maybe normal differentiation tells something about ageing. For the B lymphocytes the process of differentiation is known in some detail; there recombinations accompanied by deletion of DNA segments take place with resulting irreversible alteration in DNA [81]. Whether other gene systems change in the same manner during normal differentiation of other cell types is, however, not known although supportive evidence of such rearrangement of genetic material is recent evidence of variation in DNA primary structure between different tissues of the same individual [82]. The principle of rearranging the order of the available DNA sequences by recombinations between non-homologous sequences or inserting new copies, transposons, of already present DNA sequences [83] may well be general for differentiation and ageing. Circulating quiescent T lymphocytes contain D N A strand breaks and these breaks rejoin after phytohemagglutinin stimulation [84]. This has lead to the fascinating hypothesis that cycles of DNA breaking and rejoining are part of normal differentiation in all types of caryotic cells [85]. Transient breaks could change the degree of supercoiling in specific parts of the DNA and thus determine which genes are transcribed [86]. Another possibility is that these break-rejoinings are steps in the rearrangements of genetic material mentioned above. DNA methylation is another process that may occur during normal cellular differentiation [87], and then again decrease during ageing [88]. The primary complete transcript (copy) of the chromosome DNA is edited with deletion and sometimes rearrangements before production of the secondary R N A copy, which then governs the subsequent protein synthesis. Differential use of the RNA transcripts by regulation at the level of R N A splicing also seems to be involved in differentiation [81]. We must, therefore, still consider normal differentiation a result of both irreversible and potentially reversible genetic alterations. One D N A change that appears to accompany normal ageing is a depletion of the number of DNA templates for ribosomal RNA in post-mitotic cells. The number of copies of certain genes determines the rate at which a given protein can be synthesized. During early development gene amplification ensuring a high rate of protein synthesis takes place in some species [89]. During normal ageing the opposite happens. Of several hundred copies of the DNA templates for ribosomal R N A in each post-mitotic cell (brain, heart) up to 70 per cent may be lost in senescence [90]. An apparently similar depletion of two endonuclease restriction fractions of D N A was observed during in vitro growth (ageing) of three strains of normal human fibroblasts [91]. A change in number of copies of a mitocondrial DNA sequence also seems essential for senescence (arrest of vegetative growth) in at least one fungus species [92]. One does not know if this probably very important phenomenon of DNA loss is a pre-programmed or a special type of mutationally inflicted lesion or just misrepair. In evaluating the various D N A changes one should bear in mind that changes in gene expression with age that are not deleterious to the organism by definition, are not part of normal ageing but could be labelled protracted maturation. If high

276

age is shown to be accompanied by beneficial (compensatory) alterations in gene expression, we may one day decide upon a broader definition of normal ageing. In addition to what takes place of pre-programmed processes in the DNA, random damage to DNA is unavoidable during normal cell division and as a consequence of the production of free radicals during normal metabolism [93]. To the extent this is not counteracted by DNA repair, it should also contribute to normal ageing and indeed some human repair-deficient syndromes seem to be linked with accelerated ageing [94]. The mutation hypothesis of ageing in fact once was predominant, supported as it was by the accumulation of chromosome aberrations occurring during life [95] and the appearance of many signs of ageing in young animals exposed to such mutagenic treatments as chemical mutagens and X-irradiation [96]. However, there is no clear correlation between mutagenic and "ageing" affect and only some characteristics of ageing were accelerated by a certain treatment [97]. Mutation thus clearly contributes to ageing, but only within limits set by the genetic make up of each species. All non-repaired mutations will increase the diversity of the cellular genotype and to the extent the affected genes are expressed also increase the phenotypic diversity. The impact of each mutation in a single post-mitotic cell is negligible, and even mutation of one dividing cell is self-limiting by the normal regulation of the number of cells derived from a stem cell present at any given time but the accumulation of mutations contributes to the increasing diversity among old individuals even when belonging to an inbred strain [98]. To what extent the increase in cellular diversity is countered by surveillance mechanisms is not known. Specific immune responses are at work as demonstrated by increasing incidence of autoimmune reactions with age [99]. The activity of cytotoxic T iymphocytes declines at great age [100] but the antibody-mediated surveillance for old erythrocytes detected by Kay's group [101] seems to remain active in old animals. Non-specific surveillance such as by NK lymphocytes is also a possibility, but NK cells primarily seem cytotoxic to stem cells and tumor cells, and not to specialized normal cells [102]. In addition to whatever changes may characterize the cells of ageing individuals, there is an on-going regular depletion of post-mitotic brain cells and, more interestingly (in relation to cancer), a depletion of various types of undifferentiated cells in reproductive organs [103] and bone marrow [104], and probably other places as well. POSSIBLE I N T E R R E L A T I O N S BETWEEN C A N C E R A N D N O R M A L A G E I N G

First let us reconsider two generalizations derived from epidemioiogy. Firstly, an occupational group with an above average cancer incidence nearly always shows a high incidence of non-malignant diseases (cardiovascular, cerebral insults, kidney failure, suicides) and below average mean survival time [105,106]. As this is the case with many different occupations, a sharing of etiologic factors between

277 malignant and degenerative lesions appears most likely. Among the supposedly shared factors are inducers of free radicals [107,108,109]. Secondly, cancer and ageing as measured by age-specific death rates have roughly the same increase with age. Both, therefore, could have a multistep development. However, after cancer initiation a cell clone keeps dividing at a certain low level of differentiation permitting emergence of new mutations and subsequent selection of those cells best fitted for continued mitosis. Normal ageing, in contrast, as one of its manifestations has accumulation of mutationally induced cellular changes not leading to continued proliferation although multifocal limited proliferation often occurs [110]. Even if the mutations accelerating ageing did not include any of the steps necessary for a cell to become malignant, the changes would still influence cancer initiation or growth in several ways. DNA-damaged cells might be more susceptible to further damage that could induce cancer, e.g. due to deficient repair capacity, and mutations might accelerate age decline in the immune system. The age-dependent (programmed?) loss of repetitive sequences from postmitotic cells is not important for cancer. Tumors rarely if ever develop from normal post-mitotic cells of brain or heart muscles. However, if dividing cell types such as leukocytes also lose DNA sequences during ageing this could be important. Furthermore, the methylation level of the repeated sequences of at least some mitotic cells (cow leukocytes) also decrease with age, and in these animals spontaneous leukosis is also accompanied by a decrease in methylation of these DNA sequences [88]. A correlation does not establish causality, but it is a tempting hypothesis that demethylation is an example of a part of normal ageing which facilitates accumulation of further DNA alterations leading to malignancy. The normal mid-life decline in number of some types of immature cells as part of normal reproduction physiology and the decline in immature cells at high age must influence cancer incidences downward at these points as cancer arises from not too differentiated cells. The loss of the immature cells may be reflected in the low susceptibility of senescent mammary tissue to tumor induction and the tendency of the very old to have tumors that are rather highly differentiated [111]. The depletion of immature cells should be a restricting factor on cancer initiation. It has, however, also been postulated that depletion of bone marrow hematopoietic cells may clear the way for leukemia development as leukemia development in estrogenized and irradiated mice can be abrogated by grafting of syngeneic lymphoid cells [112-114]. At the tissue level, one reason for a higher susceptibility to carcinogens in some senescent tissues could be increased production of mutagenic metabolites of carcinogens, such as described for senescent rodents [115, 116]. Much has also been made of the decline in specific cellular immune reactivity with age [117]. The low antigenicity of spontaneous tumors big enough to be detected [32] casts doubt upon the importance of loss of immune capacity. However, in the absence of markers for the pre-neoplastic cells and the first cells of a new tumor one cannot decide if immune response inhibits growth of cells in the early stages of

278 malignant transformation. Also the possible importance of declining non-specific cellular defence (NK cells) against all kinds of abnormal cells including cancer cells [118, 119, 120] is unresolved. RESEARCH AREAS THAT NEED ATTENTION All normal mammalian cells seem to harbour D N A sequences that upon activation give a cell a malignant phenotype [38]. Attempts to breed these genes out of inbred strains would cast light on their influence on normal development and ageing. It is likely that deletion of some oncogenes will be lethal due to their importance for normal development. Normal cell D N A sequences can be cut out with restriction enzymes and identified by hybridization with appropriate templates. By in oitro transfection to other normal cells of such well-characterized D N A pieces the sequences important for malignant transformation have been identified [121]. A similar identification of sequences that determines the number of doublings left for a fibroblast culture would be a major breakthrough. Experimental induction of lymphomas with X-rays or estrogen is preventable by transferring syngeneic lymphoid cells to the treated animals as mentioned before. The effect of such graftings to middle-age animals on the spontaneous cancer (lymphoma) at higher age should be studied especially since storage of human cells for later use by the donor is technically possible. The rise in incidence of many chronic diseases in parallel with the death intensity rate offers the possibility that the diseases are largely incidental to a common underlying mechanism of normal ageing. Working on that assumption, a multistep hypothesis of normal dying could be subject to animal experiments. CONCLUDING REMARKS The present level of knowledge suggests that the best we can do with regard to ageing is to push the appearance of severe debilitation closer to the fixed maximum age obtainable by members of our species. That ageing has to be linked to a high incidence of cancer is, however, not that sure. It may depend upon the degree to which the two biologic processes share intermediary steps in their development. ACKNOWLEDGEMENTS I am grateful to Professor B.L. Strehler, University of Southern California, Professor A. Viidik, Institute of Anatomy, University of Aarhus, Denmark, Professor L. Packer, Lawrence Berkeley Laboratory, University of California, Berkeley, and Dr. R.J. Biggar, National Institutes of Health, Bethesda, Maryland, for critical advice with the manuscript.

279 REFERENCES 1 P. Greenwald, Assessment of risk factors for cancer. Prey. Med., 9 (1980) 260-263. 2 J.J. Yunis, Specific fine chromosomal defects in human cancer: an overview. Hum. Pathol., 12 (1981) 503-515. 3 G.B. Pierce, P.K. Nakane, A. Martinez-Hernandez and J.M. Ward, Ultrastructural comparison of differentiation of stem cells of murine adenocarcinomas of colon and breast with their normal counterparts. J. Natl. Cancer Inst., 58 (1977) 1329-1345. 4 N. Zarncheck and G. Pusztaszeri, CEA, AFP, and other potential tumor markers. CA, 25 (1975) 204-214. 5 P.J. Fialkow, CIonal origin of human tumors. Biochim. Biophys. Acta, 458 (1976) 28.3-321. 6 L.A. Poirier and E.K. Weisburger, Selection of carcinogens and related compounds tested for mutagenic activity. J. Natl. Cancer Inst., 62 (1979) 833--840. 7 J.A. Miller and E.C. Miller, Ultimative chemical carcinogens as reactive mutagenic electrophiles. In H.H. Hiatt, J.D. Watson and J.H. Winstein (eds.), Origins of the Cancer, Book B, Cold Spring Harbor, 1977, pp. 605-628. 8 P. Kleihues and J. Bucheler, Long-term persistence of O6-methyl-guanine in rat brain DNA. Nature, 269 (1977) 625-626. 9 K. Shimotono and H.M. Temin, Evolution of retroviruses from cellular movable genetic elements. Cold Spring Harbor Syrup. Ouant. Biol., 46 (1981) 719--730. 10 B.J. Poiesz, F.W. Ruscetti, A.F. Gazdar, P.A. Bunn, J.D. Minna and R.C. Gallo, Detection and isolation of type C retrovirus particles from fresh and cultured lymphocytes of a patient with cutaneous T-cell lymphoma. Pr0c. Natl. Acad. Sci. USA, 77 (1980) 7415-7419. 11 R. Bravo and J.E. Celis, Gene expression in normal and vitally transformed mouse 3T3B and hamster BHI 21 cells. Exp. Cell Res., 127 (1980) 249-260. 12 H.F. Oettgen, Immunologic aspects of cancer. Hosp. Pract., 16 (1981) 85r-88. 13 G.R. Martin, Teratocarcinomas and mammalian embryogenesis. Sc/ence, 209 (1980) 768-776. 14 K. lllmensee and B. Mintz, Totipotent and normal differentiation of single teratocarcinoma cells cloned by injection into blastocysts. Proc. Natl. Acad. Sci. USA, 73 (1976) 549-553. 15 A. Anders and F. Anders, Etiology of cancer as studied in the platyfish-swordtail system. Biochim. Biophys. Acta, 516 (1978) 61-95. 16 G.M. Pierce and C. Wallace, Differentiation of malignant to benign cells. Cancer Res., 31 (1971) 127-134. 17 E. Fibach, T. Landau and L. Sachs, Normal differentiation of myeloid leukemia cells, induced by differentiation-inducing protein. Nature, 237 (1972) 276-278. 18 J.R. Totter, Spontaneous cancer and its possible relationship to oxygen metabolism, Proc. Natl. Acad. Sci. USA, 77 (1980) 1763--1767. 19 R.A. Franklyn, Butylated hydroxytoluene in sarcoma-prone dogs. Lancet, June 12 (1976) 1296. 20 D. Medina, H. Lane and F. Shepherd, Effects of selenium on mouse mammary tumorigenesis and glutathione peroxydase activity. Anticancer Res., I (1981) 377-382. 21 W. Haenszel, Cancer mortality among the foreign born in the United States. J. Natl. Cancer Inst., 26 (1961) 127-132. 22 National Institutes of Health, Publication 1981. 23 S.S. Epstein and J.B. Swartz, Fallacies of lifestyle cancer theories: Nature, 289 (1981) 127-130. 24 I. Berenblum and P. Shubik, A new, quantitative, approach to the study of the stages of chemical carcinogenesis in the mouse's skin. Br. J. Cancer, I (1947) 383-391. 25 H. Kaji, Studies on T-chain biosynthesis in tumours producing immunogiobulins in NZB mice. Biochem. J., 156 (1976) 681-684. 26 O.J. Bizzozero, Jr., K.G. Johnson and A. Ciocci, Radiation-related leukemia in Hiroshima and Nagasaki 1946-1964. I. Distribution, incidence, and appearance time. New Engi. J. Med., 274 (1966) 1095-1101. 27 M.S. Fisher and M.A. Cifore, Enhanced metastatic potential of murine fibrosarcomas treated in vitro with ultraviolet radiation. Cancer Res., 41 (1981) 3018-3023. 28 L. Foulds, The natural history of cancer. J. Chron. Dis., 8 (1958) 2-37.

280 29 P.C. Nowell, The cional evolution of tumor cell population. Science, 194 (1976) 23-28. 30 G. Klein, Immune and non-immune control of neoplastic development. Cancer, 45 (1980) 2486-2499. 31 K.G. Brand, L.C. Buoen and F. Brand, Antigen-deficient cell variants in preneoplastic foreign body reaction of mice. J. Natl. Cancer Inst., 49 (1972) 459--465. 32 H.B. Hewitt, E.R. Blake and A.S. Walder, A critique of the evidence for active host de.fence against cancer, based on personal studies of 27 routine tumors of spontaneous origin. Br. J. Cancer, 33 (1976) 241-259. 33 F. Miteiman, M. Andersson-Anvret, L. Brandt, D. Catovsky, G. Klein, G. Manolov, Y. Manolova, E. Mark-Vendel and P.G. Nielsson, Reciprocal 8:14 translocation in EBV-nogative B-ceU acute lymphocyte leukemia with Burk!tt-type cells. Int. J. Cancer, 24 (1979) 27-33. 34 R. Dofuku, J.L. Biedler, B.A. Spengler and L.S. Old, Trisomy of chromosome 15 in spontaneous leukemia in AKR mice. Proc. Natl. Acad. Sci. USA, 72 (1975) 1515-1517. 35 P. Armitage and R. Doll, The age distribution of cancer and a multistage theory of carcinogenesis. Br. £. Cancer, 8 (1954) 1-2. 36 S.H. Moolgawkar and A.G. Knudsen, Jr., Mutation and cancer: A model for human carcinogenesis, g. Natl. Cancer Inst., 66 (1981) 103%1052. 37 R.E. Rowland, P.M. Favilla, A.T. Klare and A.F. Stehrey, Tumor incidence for the radium patients. I. Radiological Physics Division Annual Report ANL-7860. Part II. Biology in Medicine, Argonne National Laboratory, Argonne, Illinois, 1971, pp. 1-8. 38 M.A. Martin, T. Bryan, S. Rasheed and A.S. Khan, Identification and cloning of endogenous retroviral sequences present in human DNA. Proc. Natl. Acad. SCi. USA, 78 (1981) 4892-4896. 39 E. Alessandra, L.C. Robbins and P.R. Andersen, Cellular genes analogous to retroviral one genes are transcribed in human turnout cells. Nature, 295 (1982) 116-119. 40 W.S. Haywards, B.G. Neel and S.M. Astrin, Activation of a cellular oncogene by promotor insertion in ALV-induced lymphoid leukosis. Nature, 290 (1981) 475--480. 41 P.R.J. Burch, The Biology of Cancer, a New Approach, MTP Press, Edinburgh, 1976. 42 A.L. Herbst, D.C. Poskanzer, S.J. Robboy, L. Friedlander and R.E. Scully, Prenatal exposure to stilbestrol. A Prospective comparison of exposed female offspring with unexposed controls. N. Engl. J. Med., 292 (1975) 334-339. 43 E.E. Schwartz and A.C. Upton, Factors influencing the incidence of leukemia: Special consideration of the role of ionizihg radiation. Blood, 13 (1958) 845--864. 44 D.H. McGregor, C.E. Land, K. Chor, S. Tokvuku, P.I. Lie, T. Watabayashi and G.W. Beebe, Breast cancer incidence among atomic bomb survivors, Hiroshima and Nagasaki, 1950-1969. J. Natl. Cancer lltst., 59 (1977) 799-811. 45 R.A.M. Case, M.E. Hoster, D.B. McDonald and J.T. Pearson, Tumours of the urinary bladder in workmen engaged in the manufacture and use of certain dyestuff intermediates in the British chemical industry. Br. J. Ind. Med., 11 (1954) 75-104. 46 R.E.W. Fisher, The effect of age upon the incidence of tar warts. In Twelfth International Congress on Occupational Health, Valtaneuvoston Kirjapaino, Helsinki, 1958, p. 402. 47 R. Doll, Epidemiological observations on susceptibility to cancer in man with special reference to age. Acta Unio Int. Contra Cancrum, 20 (1964) 747-752. 48 J.F. Knox, S. Holmes, R. Doll and I.D. Hill, Mortality from lung cancer and other causes among workers in an asbestos textile factory. Br. J. Ind. Med., 25 (1968) 293-303. 49 R. Doll, L. Morgan and F.E. Speitzer, Cancers of the lung and nasal sinuses of nickel workers. Br. J. Cancer, 24 (1970) 623--632. 50 M.P.. Bochov and N.P. Kulestrov, Age sensitivity of human chromosomes to alkylating agents. Murat. Res., 14 (1972) 345-353. 51 B.W.M. Court and R. Doll, Mortality from cancer and other causes after radiotherapy for ankylosing spondylitis. Br. Med. J., ii (1965) 1327-1332. 52 P.G. Smith and R. Doll, Age- and time-dependent changes in the rates of radiation-induced cancers in patients with ankylosing spondylitis following a single course of X-ray treatment. IAEA-SM, 224 (1982) 205-218. 53 P.G. Smith and R. Doll, Late effects of X irradiation in patients treated for metropathia hemorragica. Br. J. Radiol., 49 (1976) 224-232.

281 54 S. Jablon and H. Kato, Studies of the mortality of A-bomb survivors. Radiat. Res., 50 (1972) 649-408. 55 G.W. Beebe, A. Kato and C.E. Land, Studies of the mortality of A-bomb survivors. 8. Mortality experience of A-bomb survivors 1950-1974. R E R F Technical Report, Radiation Effects Research Foundation, Hiroshima, 1977, pp. 1-77. 56 A.E. Pegg, J.W. Nicoll, P.N. Magee and P.F. Swann, Importance of DNA repair in an organ. Specificity of tumour induction by N-nitroso carcinogens. Proc. Eur. Soc. Toxicol., 17 (1976) 39-54. 57 P. Ebbesen, Aging increases susceptibility of mouse skin to DMBA carcinogenesis independent of general immune status. Sc/ence, 183 (1974) 217-218. 58 P. Ebbesen, Increased susceptibility to ~-irradiation of senescent mouse skin irrespective of grafting to young recipients. Int. J. Radiat. Biol., 37 (1980) 563-567. 59 P. Ebbesen and M.L. Kripke, The influence of age and anatomical site on ultraviolet carcinogenesis in BALB/c mice. J. Natl. Cancer Inst,, 68 (1982) 691--694. 60 E.A. Defries, C. Rowlatt and M.V. Sheriff, The effects of age on tumor induction in C57BL mice using 1,2-dimethylhydrazine. Toxicol. Lett., 8 (1981) 87-88. 61 N.K. Clapp, E.H. Perkins, W.C. Klima and L.H. Cacheiro, Temporal advancement of diethylnitrosamine carcinogenesis in aging mice. J. Gerontol., 36 (1981) 158-163. 62 M.D. Reuber, Hyperplastic and neoplastic lesions of the kidney in buffalo rats of varying ages ingesting N-4(4-fluorobiphenyl)acetamide. I. Natl. Cancer Inst., 54 (1975) 42?-429. 63 C. Huggins, L.C. Grand and F.P. Brillantes, Mammary cancer induced by a single feeding of polynuclear hydrocarbons, and its suppression. Nature, 189 (1961) 204-20?. 64 J. Grubbs Clinton, J. Peckham and K.D. Cato, Mammary carcinogenesis in rats in relation to age at time of N-nitroso-N-methylurea administration. J. Natl. Cancer Inst., 70 (1983) 209-212. 65 I.C. Summerhayes and L.M. Franks, Effects of donor age on neoplastic transformation of adult mouse bladder epithelium in vitro. J. Natl. Cancer Inst., 62 (1979) 1017-1023. 66 C. Lasne, A. Gentil and I. Chouroulinkov, Two-stage malignant transformation of rat fibroblasts in tissue culture. Nature, 247 (1974) 490--491. 67 B.L. Strehler, Time, Cells and Aging, 2nd ed., Academic Press, New York, 1977. 68 R. Rask-Nielsen and P. Ebbesen, Spontaneous reticular neoplasms in (BA × DBA/2)Fl mice, with special emphasis on the occurrence of plasma cell neoplasms. J. Natl. Cancer Inst., 43 (1969) 553--564. 69 G.M. Martin, Genetic syndromes in man with potential relevance for the pathobiology of aging. Birth Defects, Orig. Attic. Set., 14 (1978) 5-39. 70 R. Henderson, Mortality Laws and Statistics, 1st edn., Wiley, New York, 1915. 71 B. Rosenberg, G. Kemeny, L.G. Smith, I.D. Skurnick and M.J. Bandurski, The kinetics and thermodynamics of death in multicellular organisms. Mech. Ageing Dev., 2 (1973) 275-293. 72 P.R.J. Bureh, D. Jackson, C.G. Fairpo and J.J. Murray, Gingival recession ("getting long in the tooth"), colorectal cancer, degenerative and malignant changes as errors of growth control. Mech. Ageing Dev., 2 (1973) 251-273. 73 A.L. Muggleton-Harris and L. Hayflick, Cellular aging studied by the reconstruction of replicating cells from nuclei and cytoplasms isolated from normal human diploid cells. Exp. Cell Res., 103 (1976) 321-330. 74 R. Haytiiek, Current theories of biological ageing. Fed. Proc., 34 (1975) 9-13. 75 J.R. Smith and R.T. Whitney, Intraelonal variation in proliferative potential of human diploid fibroblasts: Stochastic mechanisms for cellular ageing. Sc/ence, 207 (1980) 2---84. 76 D. R6hme, Evidence for a relationship between longevity of mammalian species and life spans of normal fibroblasts in vitro and erythrocytes in vivo. Proc. Natl. Acad. Sci. USA, 78 (1981) 5009-5013. 77 G.T. Baker, Age-dependent arginine phosphokinase activity changes in male vestigial and wildtype Drosophilia melanogaster. Gerontologia, 21 (1975) 203-210. 78 T. Makinodan, Immunobiology of aging. J. Am. Geriatr. Soc., 24 (1978) 249-252. 79 D.A. Ogden and H.S. Micklem, The fate of marrow cell populations from young and old donors. Transplantation, 12 (1976) 287-293. 80 C.W. Daniel, Finite growth span of mouse mammary gland serially propagated in vivo. Experient/a, 29 (1973) 1422-1424.

282 81 S. Tonegawa, H. Sakano, R. Mati, A. Traunecker, G. Heinrich, W. Roeder and Y. Kurosawa, Somatic reorganisation of immunoglobulin genes during lymphocyte differentiation. Cold Spring Harbor Syrup. Ouant. Biol., 46 (1981) 839-858. 82 B. Calabretta, D.C. Robberson, H.A. Barrera-Saidafia, T.P. Lambrou and G.T. Saunders, Genome instability in a region of human DNA enriched in A lu repeat sequences. Nature, 296 (1982) 219-225. 83 P. Nisen, R. Medford, J. Mansorer, M. Purucker, A. Skalka and L. Shapiro, Cell-cycle-associated rearrangement of inverted repeat DNA sequences. Proc. Natl. Acad. Sci. USA, 76 (1979) 624O--6244. 84 A.P. Johnstone and G.T. Williams, Role of DNA breaks and ADP-ribosyl transferase activity in eukaryotid differentiation demonstrated in human iymphocytes. Nature, 300 (1982) 368-370. 85 F. Farzaneh, S. Sholl and R. Zalin, in M, Smulson and T. Sugimura (eds.), Novel ADPribosylations of Regulatory Enzymes, Elsevier, Amsterdam, 1980, pp. 217-225. 86 A. MacDermott, Beyond the double helix. New Sci., 95 (1982) 27.8-231. 87 A. Razin and A.D. Riggs, DNA methylation and gene function. Sc/ence, 210 (1980) 604-610. 88 G.A. Romanov and B.F. Vanyvshin, Intragenome specificity of DNA methylation in animals. Tissue specificity and variation in the methylation of repeated sequences due to aging, carcinogenesis and hormonal induction. Mol. Biol. (USSR), 14 (1980) 357-368. 89 D.D. Brown, B.D. Igor and I.B. Dawid, Specific gene amplification in oocytes. Sc/ence, 160 (1968) 272-280. 90 B.L. Strehler and M.P. Chang, Loss of hybridizable ribosomal DNA from human postmitotic tissues during aging: II. Age-dependent loss in human cerebral cortex-hippocampal and somatosensory cortex comparison. Mech. Ageing Dev., 11 (1979) 379--382. 91 R.J. Shmookler Reis and S. Goidstein, Loss of reiterated DNA sequences during serial passage of human diploid fibroblasts. Ceil, 21 (1981) 739-749. 92 C. Viern~y, A.M. Keller, O. Begel and L. Belcour, A sequence of mitochondrial DNA is associated with the onset of senescence in a fungus. Nature, 297 (1982) 157-159. 93 D. Harman, The aging process. Proc. Natl. Acad. Sci. USA, 78 (1981) 7124-7128. 94 R.W. Hart, K.Y. Hall and F.B. Daniel, DNA repair and mutagenesis in mammalian cells. Photochem. Photobiol., 28 (1978) 131-155. 95 C. Crawley and H.J. Curtis, The development of somatic mutations in mice with age. Proc. Natl. Acad. Sci. USA, 49 (1963) 626-628. 96 F.M. Sinex, The mutation theory of aging. In M. Rockstein (ed.), Theoretical Aspects of Ageing, Academic Press, New York, 1974. 97 P. Alexander, Is there a relationship between aging, the shortening of life-span by radiation and the induction of somatic mutations? In N.W. Shock (ed.), Perspectives in Experimental Gerontology, C.C. Thomas, Springfield, 1966, p. 79. 98 C.F. Hollander, Age-associated pathology in the rat: Its consequence for gerontological research. In W.E.G. Mfiiler and J.W. Rohen (eds.), Biochemical and Morphological Aspects of Ageing, Steiner, Wiesbaden, 1981, pp. 220-226. 99 H.M. Hallgren, C.E. Buckley, III, V.A. Gilbertson and E.J. Yunis, Lymphocyte phytohemagglutinin responsiveness, immunogiobulins and autoantibodies in aging humans. J. Immunol., 3 (1973) 1101-1107. 100 R.A. Miller and O. Stutman, Decline, in aging mice, of the anti-2,4,6~trinitrophenyl (TNP) cytotoxic T-cell response attributable to loss of Lyt-2-, interleukin 2-producing helper cell function. Eur. J. Immunol., 11 (1981) 751-756. 101 G.D. Bennett and M.M.B. Kay, Homeostatic removal of senescent murine erythrocytes by splenic macrophages. Exp. Hematol., 9 (1981) 297-307. 102 M. Gidlund, A. ~ ' n , P.K. Pattengale, M. Jansson, H. Wigzeli and K. Nilsson, Natural killer cells kill tumor cells at a given stake of differentiation. Nature, 292 (1981) 848--850. 103 T. Krarup, T. Pedersen and M. Faber, Regulation of oocyte growth in the mouse ovary. Nature, 224 (1969) 187-188. 104 M.B. Chen, Age related changes in hematopoietic stem cell populations of a long-lived hybrid mouse. J. Cell. Physiol., 78 (1971) 225-232.

283 105 H.B. Jones, The relation of human health to age, place, and time. In J.E. Birren (ed.), Handbook of Aging and the Individual, University of Chicago Press, Chicago, 1959, pp. 336-369. 106 Dmtmatks $ ~ k , 1979. 107 K.K. Carroll, Experimental evidence of dietary factors and hormone-dependent cancers. Cancer Res., 35 (1975) 3374-3383. 108 L. Packer and J. Walton, Do free radicals speed and antioxidants slow aging and cancer? US Energy Res. Dev. Administr. under contract W-7405-ENG-48, University of California, Berkeley, 1976. 109 M.G. Cook and P. McNamara, Effect of dietary vitamin E on dimethylhydrazine-induced colonic tumors in mice. Cancer Res., 40 (1980) 1329-1331. 110 G.M. Martin, Proliferative homeostasis and its age-related aberrations. Mech. Ageing Dev., 9 (1979) 3S5-391. 111 T. Ishii and N.H. Sternby, Pathology of centenarians. Ill. Osseous system, malignant lesions, and causes of death. J. Am. Get/art. Soc., 26 (1978) 529-533. 112 P. Ebbesen and M.J. Dc~nhoff, Abrogated thymoma development and increased amyloid development in estrogenized mice grafted spleen and bone marrow cells. Proc. Soc. Exp. Biol. bled., 138 (1971) 850--855. 113 G.P. Hirsch, Somatic mutations and aging. In E.L.S. Schneider (ed.), The Genelics of Aging, Plenum Press, New York, 1978, pp. 91-134. 114 E.A. Perkins and L.C. Cacheiro, Immundepression, a noncontributor in radiation-induced leukemogenesis of the RFM m u s e . Cancer Res., 40 (1980) 357-361. 115 M.B. Baird and L.S. Birnbanm, Increased production of mutagenic metabolites or carcinogens by tissues from senescent rodents. Cancer Res., 34 (1979) 4752-4755. 116 L.S. Birnbaum and M.B. Baird, Senescent changes in rodent hepatic epoxide metabolism, Chem. Biol. Interact., 26 (1979) 245-256. 117 M.H. l-Iallgren and E.J. Yunis, Suppressor lymphocytes in young and aged humans. J. lmmunol., 118 (1977) 2004-2008. 118 S. Rockwell, Effect of host age on the transplantation, growth, and radiation response of TMT6 tumors. Cancer Res., 41 (1981) 527-531. 119 O. Hailer, M. Hansson, R. Kiesling and H. Wigzell, Nonconventional natural killer may play a decisive role in providing resistance against syngeneic tumour cells in vioo. Nature, 270 (1977) 609. 120 M. Onsrud, Age dependent changes in some human lymphocyte sub-populations. Change in natural killer cell activity. Acta PathoL Microbiol. Scand., 84 (1981) 55-62. 121 C. Shih, L.C. Pedhy, M. Murray and R.A. Weinberg, Transforming genes of carcinomas and neuroblastomas introduced into mouse fibroblasts. Nature, 290 (1981) 261-264.