Genetic susceptibility to cancer

991

Genetic Susceptibility to Cancer Margaret R. Spifz, M.D., M.P.H., and Melissa L. Bondy, Ph.D.

For any given level of exposure to a carcinogen, only a proportion of exposed individuals will develop cancer. Interindividual differences in susceptibility at some stage of the carcinogenic process must be postulated. One contributing factor is variation in the activity of metabo- lizing enzymes responsible for conversion of procarcino- gens to proximate carcinogens. There is also a wide spec- trum of DNA repair capability within the general popula- tion. At one end are the genetic instability syndromes characterized by extreme sensitivity to carcinogenic ex- posures and high rates of cancer in homozygotes of these traits. Less extreme differences in mutagen sensitivity can be demonstrated by a quantitative assay of chromo- some breaks induced by in vitro mutagen exposure. Two case-control studies of patients with previously un- treated upper aerodigestive tract cancers have demon- strated mutagen sensitivity to be an independent risk fac- tor for the disease after controlling for the effects of to- bacco and alcohol. Mutagen sensitivity also may have prognostic relevance. There was a fourfold elevated risk of developing multiple primary cancers in mutagen-sen- sitive patients. There are also data suggestive of familial aggregation of cancer in first-degree relatives of muta- gen-sensitive patients (twofold risk for having one first- degree relative with cancer and sixfold risk for having two or more relatives with cancer). The preventive im- plications of identifying markers of carcinogen sensitiv- ity are manifold. Cancer 1993; 72991-5.

Key words: genetic susceptibility, mutagen sensitivity, DNA repair.

Because only a fraction of carcinogen-exposed individ- uals develop cancer, interindividual differences in sus- ceptibility to these environmental exposures have been postulated. In fact, almost every phase of the multistage process of carcinogenesis may be modified by interindi- vidual differences in susceptibility-differences that may be inherited or acquired.’ Omenn has stated that

From the *Department of Epidemiology, The University of Texas M.D. Anderson Cancer Center, Houston, Texas.

Supported in part by National Institutes of Health grants R 0 3 CA50945 and R 0 1 CA55769.

Address for reprints: Margaret R. Spitz, M.D., M.P.H., Depart- ment of Epidemiology, Box 189,1515 Holcombe, Houston, TX 77030.

Accepted for publication March 26, 1993.

the term ”characterization of risk” is more appropriate than ”quantitative risk assessment,” because the former conveys the need to include qualitative infor- mation, such as interindividual susceptibility, with quantitative extrapolations.’ By definition, the study of gene-environment interactions (ecogenetics) depends upon a multidisciplinary approach that integrates bio- monitoring using cytogenetic, molecular, or biochemi- cal methods with epidemiologic research. These markers provide more precise laboratory-based mea- surements of carcinogenic dose, biologic response, and host susceptibility. In this paper, we briefly review some recent advances in the development of markers for genetic susceptibility to cancer and report our recent studies of a particular assay of DNA repair capability.

Metabolic Factors in Susceptibility

Polymorphic metabolism of xenobiotic chemicals is one determinant of variation in susceptibility. There are de- monstrable interindividual differences in the metabolic capacity to convert procarcinogens to carcinogens and to detoxify carcinogens. These genetically determined differences result in variations in susceptibility, whose expression depends, of course, on exposure to the rele- vant ~arcinogen.~

For example, the ability to N-acetylate arylamines is genetically regulated. N-acetyltransferase is involved in the detoxification of carcinogenic arylamines by acet- ylation, and N-acetyltransferase activity is determined by simple autosomal Mendelian inheritance of two co- dominant alleles at a single genetic 10cus.~ In human populations, there is polymorphism of acetylator phe- notypes with homozygous-rapid, heterozygous, and homozygous-slow acetylator genotypes. Bladder cancer patients are more likely to be slow acetylators than are normal individuals. Although the rate of in- traindividual N-acetyltransferase activity variability is low, there can be 100-fold differences in interindividual a~t ivi ty .~ This is a clear-cut example bf a susceptibility marker, because it is an indicator of risk for a cancer with known genetic and environmental components.

Interindividual variability in carcinogen metabo- lism can be determined also by genetic variation in sin-

992 CANCER Supplement August 2, 2993, Volume 72, No. 3

gle forms of the cytochrome P450 enzymes.6 Studies of these enzymes' polymorphic variability are currently the focus of considerable interest.' The cytochrome P450 enzyme superfamily has the unique capability to activate many xenobiotics, including the known pro- carcinogens in cigarette smoke.

The metabolism of the antihypertensive debriso- quine is one such example; the ability to extensively metabolize this agent is linked with lung cancer This autosomal trait may modify susceptibility in a sub- stantial portion of the population, because almost three-fourths of a control population exhibited this met- abolic capacity8-a characteristic associated with a six- fold increased risk of lung cancer. No procarcinogens have been clearly identified yet as substrates for this enzyme, and negative and ambiguous results have been reported. Molecular tests for characterizing the P450 genes are being developed.'

Another P450 enzyme, aryl hydrocarbon hydroxy- lase, which is associated with the cytochrome P4501A1 isoenzyme, also has been studied extensively." Aryl hydrocarbon hydroxylase is involved in the activation of many promutagenic and procarcinogenic aryl hydro- carbons, such as benzo(a)pyrene. Predominantly inter- mediate and high aryl hydrocarbon hydroxylase-induc- ibility phenotypes have been demonstrated in periph- eral lymphocyte assays of patients with lung cancer." The hypothesis advanced was that individuals with high inducibility could more readily activate the pro- mutagens and procarcinogens in cigarette smoke. The differences in inducibility are thought to be determined by a single gene." There were initial difficulties in re- producing these results, but they since have been cor- roborated." More recently, cDNA probes for the cy- tochrome P4501A1 have been isolated and character- ized; a positive association between cigarette smoking and cytochrome P4501A1 gene expression has been noted in normal human lung tissue, and altered gene regulation has been demonstrated in lung cancer."

There is also considerable interindividual variation in the detoxifying activity of the isozymes of the gluta- thione S-transferases, which metabolize the polycyclic aromatic hydrocarbon constituents of tobacco smoke. These enzymes catalyze the detoxification of the muta- genic electrophiles that are generated by the cy- tochrome P450-mediated oxidation reactions. One would predict, therefore, that differences in glutathione S-transferase expression probably would affect an indi- vidual's susceptibility to carcinogen exp~sure. '~ A re- cent study indicated that high or intermediate activity levels of glutathione S-transferases can have a moder- ate protective effect on persons heavily exposed to to- bacco smoke.14

DNA Repair Capability

Differences in DNA repair capability also can modify cancer risk. Much of the evidence linking DNA repair deficiency and cancer risk has been derived from the study of autosomal recessive genetic instability syn- dromes, which are characterized by high cancer rates in homozygotes of these traits, Despite the rarity of these diseases, their unique cytogenetic and biochemical characteristics have thrown much light on the issue of genetic predisposition to cancer. Ataxia telangiectasia (AT), which is produced by a single gene that has been localized to 1 lq23, is a unique human model for study- ing genetic susceptibility to cancer.15 AT homozygotes are three times more sensitive to ionizing radiation than normal individuals, Dose-response curves for heterozy- gotes are intermediate between those found for normal individuals and homozygotes. Obligate AT heterozy- gotes are said to be at a sixfold higher risk than normal individuals of developing breast cancer.I6 One study reported that 63% of AT heterozygotes' lymphocytes contained increased numbers of X-ray-induced breaks, a trait seen in only 13% of control subje~ts. '~ This latter estimate (13%) is higher than the estimated frequency (10%) for AT heterozygotes in the general population. AT cells, as well as cells from persons with Fanconi anemia, Bloom syndrome, and other cancer-prone syn- dromes, show a significantly higher incidence of radio- sensitivity during G2 irradiation than do cells from nor- mal individuals.18 This radiosensitivity, which is mani- fested as chromatid breaks, may be due to less effective mechanisms for coping with free hydroxy radicals or a deficiency in DNA repair.

Sister Chromatid Exchanges

Sister chromatid exchange (SCE) is a sensitive but non- specific marker of exposure to genotoxic agents. Both environmental factors (especially cigarette smoking) and genetic factors have been implicated in the interin- dividual variation observed in baseline SCE frequency. It has been estimated that smoking accounts for approxi- mately 20% of the interindividual variation in baseline SCE freq~ency.'~ Induction of aberrations with diepox- ybutane may make SCE analysis a more sensitive marker of susceptibility.20 Diepoxybutane sensitivity in human lymphocytes is bimodally distributed, and in- creased diepoxybutane sensitivity is associated with in- creased frequency of baseline SCEs.

Pero et al. examined two other estimates of DNA repair, unscheduled DNA synthesis and ADP-ribosyl- transferase activity,'l both of which are sensitive to reg- ulation by oxidative stress and suppressed in patients with a variety of cancers. Substantial interindividual

Genetic Susceptibility to Cancer/Spitz and Bondy 993

Table 1. Risk Estimates for Mutagen Sensitivity, Cigarette Smoking, and Alcohol Use

Study lZ5 Study 2 (n = 137) (n = 216)

Mutagen sensitivity > 0.8 break per cell Adjusted*

Cigarette smoking (no. per day)

1-14 15-24 25+

Alcohol use (beer, wine, hard liquor)

1-2 3-6 > 6

3.9 (1.6, 9.1) 4.3 (2.0, 10.2)

0.7 (0.2, 2.2) 2.8 (1.1, 6.9)

12.7 (13.8, 42.3)

1.9 (0.6, 5.7) 5.0 (1.4, 17.9)

44.5 (2.5, 793.9)

2.7 (1.1, 6.6) 2.2 (1.0, 5.1)

4.1 (1.3, 13.3) 7.8 (3.0, 20.2)

11.0 (4.1, 29.5)

1.0 (0.5, 2.2) 1.7 (0.7, 4.2)

14.0 (3.1, 62.3) Values in parentheses are 95% confidence limits. * Adiusted for ciearette smoking, alcohol, age, and education.

variation in ADP-ribosyltransferase activity also has been demonstrated. Per0 et al. suggest that both un- scheduled DNA synthesis and ADP-ribosyltransferase responses are sensitive to varying intracellular glutathi- one content.

Mutagen Sensitivity

Hsu developed an assay in which chromosomal break- age induced by in vitro exposure to the radiomimetic drug bleomycin is used as an indirect measure of DNA- repair capability.” Bleomycin was chosen as the test mutagen because it is radiomimetic, inducing single- and double-strand DNA breaks, which are expressed as chromatid breaks within a relatively short time.23 About 12% of a normal control population exhibits mutagen hypersensitivity (>1 .O break per cell).” A greater pro- portion of previously untreated patients with lung, co- lon, and head and neck cancers (>50%) are mutagen- sensitive (defined as >0.8 break per ell).*^,'^ It has been demonstrated that cigarette smoking, per se, does not increase ~ensitivity.’~ These findings suggest that sensi- tivity to genotoxic mutagens may be related also to sus- ceptibility to environmental carcinogenesis.

Because upper aerodigestive tract cancers are sen- tinel diseases of exposure to tobacco and alcohol, they constitute an ideal group for the study of gene-environ- ment interactions. Bleomycin-induced sensitivity and environmental exposure have been assessed in two case-control studies (one reported here for the first time) of patients with previously untreated upper aero- digestive malignancies and healthy control subjects. In the first study, increased mutagen sensitivity (>0.8

break per cell) was shown to be a strong and significant risk factor [adjusted odds ratio (OR) = 4.31.” The first study was limited by size (n = 75 cases, 62 controls) and design constraints. In particular, the control population for this preliminary study was less than optimal with respect to sex (overrepresentation of females), age (younger than cases), and selection (hospital employees and spouses were included).

The second study was our examination of a further 108 white patients with histologically confirmed, previ- ously untreated squamous cell carcinoma of the upper aerodigestive tract. The control population (n = 108), which had no cancer history, was frequency-matched to the cases by age (+5 years), sex, and ethnicity and was selected from blood and platelet donors to The Uni- versity of Texas M.D. Anderson Cancer Center Blood Bank. The magnitude of the univariate and adjusted risk estimates for mutagen sensitivity (OR = 2.2, P < 0.05) was lower in this larger study, but still statistically significant. Both studies also demonstrated dose-re- sponse relationships between cancer risk and cigarette smoking and alcohol use (Table 1).

To evaluate the independent effect of mutagen sen- sitivity and its interaction with cigarette smoking and alcohol consumption, risk estimates for various combi- nations of mutagen sensitivity and smoking or alcohol use were computed in stratified analyses (Table 2). These analyses were restricted to study participants for whom all relevant information was available. Referent categories were study participants who were not muta- gen sensitive and did not use either cigarettes or alco- hol. Mutagen sensitivity (>0.8 break per cell) was a risk factor in the absence both of smoking or alcohol use (threefold to fivefold), and there were significantly ele- vated risks associated with smoking and alcohol use in

Table 2. Risk Estimates for Strata of Cigarette Smoking, Alcohol, and Mutagen Sensitivity (> 0.8 break per cell)

Study lZ5 Study 2 (n = 137) (n = 216)

Smoking/rnutagen sensitivity

No/no No/yes Yes/no Yes/ yes

Alcohol /mutagen sensitivity

No/no No/yes Yes/no Yeshes

1.0 5.8 (1.3, 25.1) 5.4 (1.5, 19.5)

19.8 (4.6, 84.8)

1.0 3.6 (1.2, 11.1) 5.4 (1.6, 18.3)

17.1 (4.1, 70.6)

1.0 3.2 (0.6, 18.7) 8.1 (1.7, 37.7)

23.0 (5.0, 106.0)

1.0 3.0 (1.4, 6.4) 3.0 (1.2, 7.8) 5.8 (2.3, 14.2)

Values in parentheses are 95% confidence limits.

994 CANCER Supplement August 2, 2993, Volume 72, No. 3

Table 3. Risk of Multiple Primary Cancers in Patients with Head and Neck Cancer by Mutagen Hypersensitivity Statusz6

Second Mutagen No. of malignancies Relative 95% Confidence sensitivity patients (%) risk limits

> 1.0 break

5 1.0 break per cell 33 9 (27.3) 4.4 1.2, 15.8

oer cell 5 1 4 17.8) - -

non-mutagen-sensitive persons. The combined effects of cigarette smoking and mutagen sensitivity suggested a multiplicative effect.

Our study also examined whether patients with mutagen sensitivity are genetically predisposed to cancer because of their inability to repair induced chro- mosome damage. This study evaluated the self-re- ported cancer histories of 669 first-degree relatives of 108 patients with cancers of the upper aerodigestive tract. We found familial aggregation of cancer in muta- gen-sensitive patients (OR = 2.6, 95% confidence limits = 1.1, 6.5) for one first-degree relative with cancer and for two or more first-degree relatives with cancer (OR = 6.6, 95% confidence limits = 1.7,25.7). Further studies are needed to confirm this finding.

In a third study, 84 previously untreated head-and- neck cancer patients were evaluated for sensitivity to bleomycin at baseline and then followed longitudinally for multiple primary malignancies for an average of 20 months.26 Four of the 51 nonsensitive patients (8%) de- veloped second primary malignancies, compared with 9 of 33 (27%) hypersensitive patients (Table 3). The relative risk of multiple primary cancer in the latter group was 4.4 (95% confidence limits = 1.2, 15.8). This finding has clinical and prognostic relevance, because the development of second malignant tumors is the most common cause of mortality in early-stage disease.

In Vitro Studies

It can be postulated that chemopreventive agents could modulate mutagen sensitivity. The bleomycin assay has been used to study the protective effects of compounds with chemopreventive properties in both human lym- phoblastoid cell lines and freshly cultured lymphocytes from cancer patients and control subjects. The antioxi- dants a-tocopherol acid and a-tocopherol acid succi- nate both exhibited a dose-dependent protective effect in preventing bleomycin-induced chromosome damage in human lymphoblastoid cell lines and peripheral blood lymphocytes.*’ The concentrations of tocoph- erols used were in the pharmacologic range of vitamin E. These data are biologically plausible and confirm pre-

vious findings that vitamin E inhibits tumor cell growth in vitro and prevents the binding of active carcinogenic metabolites to cellular DNA and induced chromosome breakage.”

A protective effect of 13-cis-retinoic acid, a syn- thetic retinoid shown effectively to suppress premalig- nant oral cavity lesions and reduce risk of second pri- mary lesions in patients with head and neck cancer,29t3o similarly has been studied.31 This study also docu- mented a statistically significant reduction in induced breaks in cell lines and in lymphocyte cultures treated with re ti no id^.^^ (Physiologic concentrations of reti- noids were used in this and the previously cited study.) A similar phenomenon has been reported for ascorbic acid and n-a~etyl-l-cysteine.~~ The bleomycin assay could be used to study other chemopreventive com- pounds and as an intermediate marker in clinical che- moprevention trials.

Summary

Risk assessment is now recognized as a multidisciplin- ary process, extending beyond the scope of traditional epidemiologic methodology. Such an assessment must include biological evaluation of interindividual differ- ences in carcinogenic susceptibility, including measure- ments of carcinogen metabolic activation and DNA re- pair capability. These susceptibility markers will enable us to identify high-risk population subgroups that can be targeted for intensive primary and secondary pre- ventive strategies.

References

Harris CC. Interindividual variation among humans in carcino- gen metabolism, DNA adduct formation and DNA repair. Carci- nogenesis 1989;10:1563-6. Omenn GS. Genetic susceptibility and the estimation of risk. In: Gordis L, ed. Epidemiology and health risk assessment. New York Oxford University Press, 1988; 92-106. Bodmer WF. Cancer Genetics, In: Bodmer WF, ed. Inheritance of susceptibility to cancer in man. New York: Oxford University Press, 1983: 1-15.

Genetic Susceptibility to Cancer/Spitz and Bondy 995

4.

5 .

6.

7.

8.

9.

10.

11.

12.

13.

14.

15.

16.

17.

Hein DW. Genetic polymorphism and cancer susceptibility: evi- dence concerning acetyltransferase and cancer of the urinary bladder. Bioessays 1988;9:200-4. Cartwright RA. Epidemiological studies on N-acetylation and C-center ring oxidation in neoplasia: Banbury report no 16. In: Omenn GS, Gelboin HV, eds. Genetic variability in responses to chemical exposure. Cold Spring Harbor, NY: Cold Spring Har- bor Laboratory, 1986: 359-368. Vahakangas K, Pelkonen 0. Host variations in carcinogen me- tabolism and DNA repair. In: Lynch HT, Hirayama T., eds. Ge- netic epidemiology of cancer. Boca Raton, FL: CRC Press, 1989: 35-54. Caporaso NE, Tucker MA, Hoover RN, Hayes RB, Pickle LW, Issaq HJ, et al. Lung cancer and the debrisoquine metabolic phe- notype. ] Natl Cancer Znst 1990;82:1264-72. Ayesh R, Idle JR, Ritchie JC, Crothers MJ, Hetzel MR. Metabolic oxidation phenotypes as markers for susceptibility to lung cancer. Nature 1984;312:169-70. Sugimura H, Caporaso NE, Shaw GL, Modali RV, Gonzalez FJ, Hoover RN, et al. Human debrisoquine hydroxylase gene poly- morphisms in cancer patients and controls. Carcinogenesis

Kellerman G, Shaw CR, S Luyten-Kellerman M. Aryl hydrocar- bon hydroxylase inducibility and bronchogenic carcinoma. N Engl] Med 1973;289:934-7. Kouri RE, McKiney CE, Slomianry DT, Snodgrass DR, Wray NP, McLemore TL. Positive correlation between aryl hydrocarbon hydroxylase activity and primary lung cancer as analyzed in cryopreserved lymphocytes. Cancer Res 1982;42:5030-7. McLemore TL, Adelberg S, Liu MC, McMahon NA, Yu SJ, Hub- bard WC, et al. Expression of CYPlAl gene in patients with lung cancer: evidence for cigarette smoke-induced gene expres- sion in normal lung tissue and for altered gene regulation in primary pulmonary carcinoma. Natl Cancer Inst 1990;82:1333- 9. Wolf CR. Metabolic factors in cancer susceptibility. Cancer Sun, 1990;9:437-74. Heckbert SR, Weiss NS, Hornung SK, Eaton DL, Motulsky AG. Glutathione S-transferase and epoxide hydroxylase activity in human leukocytes in relation to risk of lung cancer and other smoking-related cancers. 1 Natl Cancer Znst 1992;84:414-422. Gatti RA. Localizing the genes for ataxia telangiectasia: a human model for inherited cancer susceptibility. Adv Cancer Res 1991;

Swift M, Morrell D, Massey RB, Chase CL. Incidence of cancer in 161 families affected by ataxia-telangiectasia. N Engl ] Med

Wiencke JK, Wara DW, Little JB, Kelsey KT. Heterogeneity in the clastogenic response to x-rays in lymphocytes from ataxia-telan- giectasia heterozygotes and controls. Cancer Causes Control

1990;11:1527-30.

56~77-104.

1991;325:1831-6.

1992;3:237-45.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

Gantt R, Parshad R, Price FM, Sanford KK. Biochemical evi- dence for deficient DNA repair leading to enhanced G2 chroma- tid radiosensitivity and susceptibility to cancer. Radiation Res

Kelsey KT, Christiani DC, Wiencke JK. Bimodal distribution of sensitivity to SCE induction by diepoxybutane in human lym- phocytes, 11. Relationship to baseline SCE frequency. Mutaf Res

Wiencke JK, Wrensch MR, Miike R, Petrakis NL. Individual sus- ceptibility to induced chromosome damage and its implications for detecting genotoxic exposures in human populations. Cancer Res 1991; 51:5266-9. Pero RW, Miller DG, Lipkin M, Markowitz M, Gupta S, Win- awer SJ, et al. Reduced capacity for DNA repair synthesis in patients with or genetically predisposed to colorectal cancer. ] Nut1 Cancer Inst 1983;70:867-75. Hsu TC. Genetic predisposition to cancer with special reference to mutagen sensitivity. In Vitro Cell Dev Biol 1987;23:591-603. Cherry LM, Hsu TC. Bleomycin-induced chromosome damage in lymphocytes of medullary thyroid carcinoma patients and their family members. Anticancer Res 1983;3:367-72. Hsu TC, Johnston DA, Cherry LM, Ramkissoon D, Schantz SP, Jessup JM, et al. Sensitivity to genotoxic effects of bleomycin in humans: possible relationship to environmental carcinogenesis. Int 1 Cancer 1989;43:403-9. Spitz, MR, Fueger JJ, Beddingfield NA, Annegers JF, Hsu TC, Newel1 GR, et al. Chromosome sensitivity to bleomycin induced mutagenesis: an independent risk factor for upper aerodigestive tract cancers. Cancer Res 1989;49:4626-8. Schantz SP, Spitz MR, Hsu TC. Mutagen-sensitivity in head and neck cancer patients: a biologic marker for risk of multiple pri- mary malignancies. ] Natl Cancer Inst 1990;82:1773-5. Trizna Z, Hsu TC, Schantz SP. Protective effects of vitamin E against bleomycin-induced genotoxicity in head and neck cancer patients in vitro. Anticancer Res 1992;12:325-8. Shamberger RL, Baughman FF, Kalchert SL, Willis CS, Hoffman GC. Carcinogen-induced chromosomal breakage decreased by antioxidants. Proc Natl Acad Sci U S A 1973;70:1461-3. Hong WK, Endicott, Itri LM, Doos W, Batsakis JG, Bell R, et al. 13-cis-retinoic acid in the treatment of oral leukoplakia. N EnglJ Med 1986;315:1501-5. Hong WK, Lippman SM, Itri LM, Karp DD, Lee JS, Byers RM, et al. Prevention of second primary tumors with isotretinoin in squamous-cell carcinoma of the head and neck. N Engl ] Med

Trizna Z, Hsu TC, Schantz SP, Lee JL, Hong WK. Anticlasto- genic effects of 13-cis-retinoid acid in vitro. Eur ] Cancer (in press). Trizna Z, Schantz SP, Hsu TC. Effects of N-acetyl-L-cysteine and ascorbic acid on mutagen-induced chromosomal sensitivity in patients with head and neck cancers. A m ] Surg

1986;108:117-26.

1991;248:27-33.

1990;323:795-801.

1991;162:294-8.