Mutagen Sensitivity Exhibits a Dose-Response Relationship in Case-Control

We have been quantitating, as a marker of cancer susceptibility, induced chromatid breaks in lymphocyte cultures exposed to chemical mutagens. This report highlights the consistency of the results from two casecontrol studies, using different methods of presenting the data. In both the lung cancer case-control study, which used bleomycin, a radiomimetic agent, as the test mutagen, and the melanoma study, which used 4-nitroquinoline-oxide, an UV-mimetic agent, the mean number of breaks/cell was significantly higher in the cases compared with the controls. When the data were dichotomized at the 75th percentile of breaks in the control populations, significantly elevated adjusted odds ratios (3.7 and 5.0, respectively) were detected. Dose-response relationships were evident in both studies when the data were categorized by quartiles of breaks/cell in the controls, with highest risk estimates being in the top quartile of induced breaks. The potential for extending this assay to other cancer sites, using a variety of test mutagens, is exciting. Introduction In the I 970s and early l980s, studies on xeroderma pigmentosum and other genetic instability syndromes indicated that in the human population there are individuals who are genetically deficient in one system of DNA repair or another. Some investigators hinted that a gradation with respect to repair capability might exist in the “normal” human population, but no one had tested this idea experimentally. In 1983, one ofus (I) proposed a working hypothesis that links sensitivity to environmental mutagens (hence, susceptibility to carcinogenesis) and defective DNA repair and developed an in vitro assay to evaluate this association. In our previous studies, we used both the radiomimetic antibiotic bleomycin and the UV-mimetic carcinogen 4NQO3 to treat human lymphocytes in primary blood cultures and to record the number of chromatid breaks induced by either of these mutagens (2, 3). These experiments were designed to find Received 1 1/28/95; revised 3/6/96; accepted 3/15/96. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertise,ne,tt in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. I Supported by Research Grant I from the John S. Dunn Research Foundation, Houston, Texas, and by Grant CA 55769 from the National Cancer Institute. 2 To whom requests for reprints should be addressed, at Department of Cell Biology, Box 181, M. D. Anderson Cancer Center. Houston, TX 77030. The abbreviations used are: 4NQO. 4-nitroquinoline-l-oxide; b/c, breaks per cell; OR. odds ratio; CI. confidence interval. out: (a) whether the working hypothesis had some validity: and (h) if so, whether such differences were associated with cancer susceptibility, especially in reference to environmentally related carcinogenesis. Materials and Methods The assay methods used primary cultures of human peripheral lymphocytes for mutagen exposure and enumerated the number of chromatid breaks induced as an estimate of mutagen sensitivity. From 335 normal blood donors assayed for bleomycin clastogenicity (2), the mean number ofb/c was 0.60, with a SD of0.35. We used 1.00 b/c as the arbitrary demarcation line for separating individuals who were hypersensitive to bleomycin from those who were less so. Using this breakpoint, less than 13% of normal individuals could be classified as hypersensitive, whereas approximately 48% of patients with head and neck cancers were in this category (2). Similarly, from 103 normal individuals assayed for 4NQO sensitivity, the mean + 1 SD value was 0.80 (3). Approximately 15% of normal individuals were hypersensitive, whereas 44% of melanoma patients were hypersensitive (3). Because responses to mutagen-induced chromatid breakage rates represented a continuous variable, dichotomizing the breakage data may not be the optimal approach. One of the reasons for this inadequacy is inherent to the method employed; namely, instead of using the entire cell population to estimate the samples for mutagen sensitivity, we usually examine only 50 metaphases per sample. Lee et a!. (4) have shown that scoring 50 metaphases yields adequate statistical reliability. However, one or two metaphases containing a high number of chromatid breaks (e.g. ,seven or eight breaks each) or having no breaks could skew the data considerably. If mutagen sensitivity values could be stratified into more quantitative grades to represent degrees of susceptibility to environmental carcinogenesis, the test results would be more meaningful. In the present report, we calculated ORs to assess the dose-response relationship for chromatid breakage frequencies of both bleomycin and 4NQO exposure as an improvement to our previous dichotomous definition of mutagen sensitivity. For the bleomycin assay, the cases and controls were derived from a molecular epidemiology study of lung cancer in minority populations (Africanand Mexican-Americans) descnibed previously (5, 6). There were I I 3 of the former and 67 of the latter with newly diagnosed, previously untreated lung cancer and 270 controls matched for age, ethnicity, and sex. Epidemiological data were collected by personal interview with informed consent. For the 4NQO assay, there were 7 1 melanoma cases and 137 controls. The majority of cases were referred by Dr. Lynn Feun (University of Miami Sylvester Comprehensive Cancer Center, Miami, FL), and the control individuals were healthy volunteers, most of whom were spouses of patients with head and neck cancers. The blood culture, mutagen treatment, and cell harvest procedures have been reported previously (2). Briefly, primary on September 8, 2017. © 1996 American Association for Cancer Research. cebp.aacrjournals.org Downloaded from Table I Analyses of bleomycin sensitivity from a lung cancer case-control study Table 2 Analyses of 4-NQO sensitivity from a case-control study of melanoma