CYP1A1 Gene Polymorphism and Risk of Epithelial Ovarian Neoplasm

Gynecologic Oncology 86, 124 –128 (2002) doi:10.1006/gyno.2002.6720

CYP1A1 Gene Polymorphism and Risk of Epithelial Ovarian Neoplasm Dilek Aktas,* ,1 Inci Guney,† Mehmet Alikasifoglu,* Kunter Yu¨ce,† Ergul Tuncbilek,* and Ali Ayhan† *Department of Genetics and †Department of Obstetrics & Gynecology, Hacettepe University Medical School, 06100, Sıhhıye, Ankara, Turkey Received December 21, 2001

Objective. Gene– environment interactions have been the focus of a number of recent studies of the occurrence of human cancers, and an association between the risk and the CYP1A1*3 polymorphism has been noticed for several cancers. Previous studies suggest that estrogens are involved in the etiology of ovarian cancer. The cytochrome P450 1A1 (CYP1A1) gene polymorphism may play role in the development of epithelial ovarian neoplasm by detoxification of polycyclic hydrocarbons and other compounds and the concentration of estrogens and their metabolites. Therefore, we assessed the association of CYP1A1 gene polymorphism in patients with epithelial ovarian neoplasm in the Turkish populations through a case-control study. Methods. Using an allele-specific polymerase chain reaction (PCR)-based method, CYP1A1*3 polymorphism, in exon 7 of the gene, was analyzed in 117 epithelial ovarian neoplasm patients and 202 control subjects. Results. The CYP1A1 Ile/Val genotype significantly increased the risk for patients with epithelial ovarian neoplasm (OR 5.7, 95% CI 3.34 –9.76). Furthermore, there were statistical differences in the distribution of CYP1A1 Val/Val genotype among all patients (OR 5.85, 95% CI 2.40 –14.25). In other words, the presence of the Val allele significantly increased the risk of epithelial ovarian neoplasm. Among benign tumors, the frequency of Ile/Val and Val/Val genotypes was found to be statistically significant with an ORs of 6.01 and 4.38 (95% CI 2.61–13.84 and 1.04 –18.38, respectively). In the benign serous ovarian tumors, patients with Ile/Val and Val/Val revealed a 7.2- and 10.5-fold higher risk of having ovarian carcinoma (95% CI 2.22–23.40 and 2.16 –51.19), respectively. In the benign mucinous ovarian carcinoma patients, the frequency of Ile/Val was found to be statistically significant with an OR of 5.15 (95% CI 1.75–15.16). However, no patient with Val/Val genotype was observed in this group and no statistical distribution was performed. Among borderline tumors, CYP1A1 Ile/Val genotype significantly increased the risk for patients (OR 5.15, 95% CI 1.75–15.16). However, only one patient was observed with the Val/Val allele and the frequency of this genotype was not found to be statistically different with an OR of 2.50 (95% CI 0.27–22.64). Among ovarian cancer patients, there were statistically differences in the distribution of CYP1A1 Ile/Val and Val/ Val genotypes (OR 5.73, 95% CI 3.04 –10.76; and OR 7.42, 95% CI 2.80 –19.66), suggesting that patients carrying these genotypes were at increased risk for ovarian carcinoma. In serous carcinoma, 1

To whom correspondence and reprint requests should be addressed. Fax: ⫹90 312 3115522. E-mail: [email protected]. 0090-8258/02 $35.00 © 2002 Elsevier Science (USA) All rights reserved.

124

patients with CYP1A1 Ile/Val and Val/Val revealed a 6.5- and 10-fold higher risk of having ovarian cancer (OR 7.09, 95% CI 3.30 –15.22; and OR 8.77, 95% CI 2.83–27.14). In mucinous carcinoma, patients with CYP1A1 Ile/Val and Val/Val also revealed a 5.4 and 10.5 times higher risk of having ovarian cancer. There were no statistical significance in the distribution of Val allele among endometroid-type cancer patients. Conclusions. Our data, although based on a small number of subjects, suggest that variant alleles of CYP1A1 gene in ovarian epithelial cells, directly or through other components, may contribute to initiation of ovarian carcinogenesis. © 2002 Elsevier Science (USA)

Key Words: ovarian carcinoma; epithelial neoplasm; CYP1A1; polymorphism; estrogen.

INTRODUCTION Ovarian carcinoma is the fourth most common cause of cancer death in women and the epidemiological evidence strongly suggests that ovarian cancer may be caused primarily by environmental factors; but these associations have not been firmly established [1]. Epithelial ovarian neoplasms are believed to arise in the ovarian surface epithelium or its derivates, epithelial crypts, and inclusion cysts. The “incessant ovulation” and “estrogen-mediated proliferation” models hypothesize propose that premalignant and malignant ovarian surface epithelium cells with estrogen receptors proliferate in response to elevated hormone levels present within the microenvironment of the ovary immediately following ovulation; however, no direct evidence to support these models has been described [2– 4]. Estrogens are associated with carcinogenic events and they can play in the induction of cancer: the generation of electrophilic species that can covalently bind to DNA to induce oncogenic mutations and stimulation of cell proliferation [5–7]. Estrogens are mainly metabolized via two main pathways: 16␣-hydroxylation and formation of 2- and 4-catechol estrogens (2- and 4-OH (E1)E2) [8 –10]. The formation of 2-OH (E1)E2 is predominantly mediated by cytochrome A1 (CYP1A1) and A2, which also catalyze 4-hydroxylation to a lesser extent [11]. Unless detoxified, catechol estrogens may be oxidized to electrophilic metabolites, catechol estrogen qui-

CYP1A1 GENE POLYMORPHISM AND RISK OF EPITHELIAL OVARIAN NEOPLASM

nones that can react with DNA to form depurinating and stable adducts. These adducts, particularly depurinating adducts, can be potential endogenous initiators of cancer and lead to oncogenic mutations and may subsequently initiate many human cancer [5, 12, 13]. Individual susceptibility to cancer may be partly explained by genetics variability in metabolic activities related to enzymes on cytochrome P450 1A1 (CYP1A1) or enzymes on glutathione S-transferase M1 (GSTM1) [14]. CYP1A1 is key to the metabolic activation of polycyclic hydrocarbons. It has provided a basis for numerous assessments of its role in chemical carcinogenesis by enhancing susceptibility of individuals, possibly through bioactivation of carcinogens [15–17]. The CYP1A1 gene is polymorphic in human and show interethnic differences [15]. In the 3⬘-noncoding region of CYP1A1, an MspI restriction length polymorphism (RFLP) segragates in linkage disequilibrium with a polymorphism (CYP1A1*3) in codon 462 of exon 7 that results in a Ile to Val replacement in the catalytic region [18]. Among races, the large difference in the CYP1A1*3 polymorphism has been noticed [18 –20]. To determine whether CYP1A1 played a role in the development of epithelial ovarian neoplasms (EON), we assessed the association of this polymorphism in patients in comparison to control subjects in Turkish population.

MATERIAL AND METHODS Patients and controls. The study was designed as a casecontrol study. The patients and control subjects attended the outpatient clinics at the Hacettepe University Hospital, Ankara, Turkey, between January 1999 and July 2001. The study group consisted of 117 Turkish epithelial ovarian neoplasm patients. Among these patients, 31 (26.5%) had benign, 16 (13.7%) had borderline tumors, and 70 (59.8%) had ovarian cancer. The mean age was 49.9 years: 44.4 years for benign tumors, 49.3 for borderline tumors, and 56.1 for ovarian cancer. Histopathological diagnosis and clinical staging were classified according to the criteria of the International Federation of Gynecology and Obstetrics (FIGO) and the most common histologic tumor type was serous among benign tumors and ovarian cancer (Table 1). The patients and control individuals, living in the different areas of Anatolia, were admitted to our reference hospital, and they represent a combined data for Turkish population. Informed consent was obtained from control subjects and patients, and all procedures were approved by the institute’s responsible committee. The control group consisted of 202 individuals who had no history of any malignancy. Controls were recruited from the Hacettepe University Hospital subjects and randomly selected individuals were analyzed to compare the distributions of CYP1A1 genotypes Genotyping. Blood samples from epithelial ovarian neoplasms and control groups were collected and stored at ⫺20°C.

125

Genomic DNA was extracted from blood using the method described by Miller et al. [21]. The genotyping of CYP1A1 was determined by PCR described by Hayashi et al. [18]. Each DNA sample was amplified in two separate reaction using one of two 5⬘ primers: 5⬘-AAG ACC TCC CAG CGG GCA AT-3⬘ or 5⬘-AAG ACC TCC CAG CGG GCA AC-3⬘. All reactions included the 3⬘ primer: 5⬘-GAA AGG CTG GGT CCA CCC TCT-3⬘ [22]. After an initial denaturation at 94°C for 2 min, followed by 28 cycles at 94°C for 60 s, 65° C for 45 s, and a final extension at 72°C 60 s. To visualize CYP1A1 gene PCR products were run on 2% agarose gel with ethidium bromide. Statistical analysis. All analyses were performed using the Statistical Package for the Social Sciences program (SSPS Inc., Chicago, IL). For statistical analysis, patients were divided into those with benign and borderline tumors and ovarian cancer patients. The means were tested using Fisher exact tests. To describe the strength of the association, crude odds ratios (OR) were calculated and given within 95% confidence intervals. All tests were two sided, and the criterion of probability was at 0.05. Cases and controls were unmatched. RESULTS By using the PCR method, CYP1A1*3 polymorphism was analyzed in 117 patients with epithelial ovarian neoplasm and 202 control subjects. Table 1 shows the distribution of the CYP1A1 gene in patients and controls. A total of 202 controls were studied of whom 158 (78.2%) were homozygous wildtype (Ile/Ile), 35 (17.3%) were heterozygous (Ile/Val), and 9 (4.4%) were the minor valine substituted allele homozygous (Val/Val). The mean age of control population was 61.83 years. Genotype-specific ORs were estimated assuming that the genotype frequencies in controls were consistent with Hardy-Weinberg equilibrium [23]. Among the total of 117 epithelial ovarian neoplasm patients, the frequency of individuals carrying Ile/Val allele was significantly higher compared to controls: When adjusted odds ratio were calculated patients with Ile/Val genotype revealed a 5.7fold higher risk of having ovarian neoplasm than these with Ile/Ile (95% CI 3.34 –9.76). The higher frequency of patients with Val/Val genotype was also found in all patients (OR 5.85, 95% CI 2.40 –14.25), indicating that persons carrying this genotype are increased risk for epithelial ovarian neoplasm (Table 1). In the group of the benign ovarian tumors, there was a higher frequency of the Ile/Val genotype in patients with benign tumor than in healthy controls (OR 6.0, 95% CI 3.34 –9.76). Moreover, Val/Val genotype tended to be more common among patients with benign tumor than among controls (OR 4.3, 95% CI 1.04 –18.38). In the group of benign serous tumor, patients with Ile/Val and Val/Val genotype revealed a 7.2- and 10.3-fold higher risk than these with Ile/Ile (95% CI 2.22– 23.40 and 2.16 –51.19), respectively. As for the benign muci-

126

AKTAS ET AL.

TABLE 1 Distribution of CYP1A1*3 Polymorphism in Control Subjects and Patients with Epithelial Ovarian Neoplasm lle/lle Controls EON OR b 95% CI d ␹2 P Benign tumors OR b 95% CI d ␹2 P Serous OR a 95% CI d ␹2 P Mucinous OR b 95% CI d ␹2 P Borderline tumors OR b 95% CI d ␹2 P Ovarian cancer OR b 95% CI d ␹2 P Serous OR b 95% CI d ␹2 P Mucinous OR b 95% CI d ␹2 P Endometroid OR b 95% CI d ␹2 P

158 (78.22) a 45 (38.4) 1c

12 (38.8) 1c

5 (31.2) 1c

7 (46.7) 1c

7 (43.7) 1c

26 (37.1) 1c

14 (32.5) 1c

5 (35,7) 1c

8 (61.5) 1c

lle/Val

Val/Val

Any Val

35 (17.33) 57 (48.7) 5.718 3.347–9.769 44.307 0.000 16 (51.6) 6.019 2.616–13.848 20.960 0.000 8 (50.0) 7.223 2.229–23.408 13.892 0.001 8 (53.3) 5.159 1.755–15.168 10.515 0.004 8 (50.0) 5.159 1.755–15.168 10.515 0.004 33 (47.1) 5.730 3.048–10.769 32.765 0.000 22 (51.2) 7.094 3.306–15.223 29.975 0.000 6 (42.3) 5.417 1.564–18.757 8.591 0.010 4 (30.8) 2.257 0.644–7.916 1.694 0.284

9 (4.45) 15 (12.9) 5.852 2.402–14.254 17.955 0.000 3 (9.6) 4.389 1.408–18.384 4.771 0.064 3 (18.8) 10.533 2.167–51.196 12.325 0.001 — —

44 (21.78) 72 (60.6) 5.745 3.484–9.476 50.604 0.000 19 (61.2) 5.686 2.564–12.605 21.264 0.000 11 (68.8) 7.80 2.607–23.939 17.337 0.000 8 (53.3) 4.104 1.411–11.940 7.629 0.001 9 (56.3) 4.617 1.627–13.098 9.572 0.004 44 (62.9) 6.077 3.373–10.949 40.075 0.000 29 (67.5) 7.438 3.620–15.282 35.335 0.000 9 (63.8) 6.464 2.061–20.274 12.774 0.001 5 (38.5) 2.244 0.699–7.204 1.931 0.178

1 (6.3) 2.508 0.278–22.641 0.716 0.382 11 (15.8) 7.427 2.805–19.661 20.294 0.000 7 (16.3) 8.778 2.839–27.142 18.708 0.000 3 (21.5) 10.533 2.167–51.196 12.329 0.011 1 (7.7) 2.194 0.247–19.504 0.522 0.417

Note. These data were available for 202 control subjects and for 117 EON (epithelial ovarian neoplasm). a Numbers in parentheses denote percentages. b OR, represents the relation risk for EON patients with lle/Val and Val/Val genotype relative to those with lle/lle. c Reference category. d CI, confidence interval.

nous tumors, the higher frequency of patients with Ile/Val genotype was found among patients and there was a significant increase in relative risk association with this genotype (OR 5.1, 95% CI 1.75–15.16). However, no patient with Val/Val genotype was observed in this group.

In the group of the borderline tumors, there was a higher frequency of Ile/Val genotype in patients with borderline tumors than in healthy control (OR 5.15, 95% CI 1.75–15.16). In this group, only a patient had Val/Val genotype and no relation was observed.

CYP1A1 GENE POLYMORPHISM AND RISK OF EPITHELIAL OVARIAN NEOPLASM

Among the ovarian cancer patients, the Ile/Val genotype revealed a 5.7-fold higher risk of having ovarian cancer than those with Ile/Ile (95% CI 3.04 –10.76). Furthermore, there was statistical significant increase in relative risk association with Val/Val genotype between these cancer patients and controls (OR 7.4, 95% CI 2.80 –19.66), suggesting that individuals carrying this genotype are increased risk for developing ovarian carcinoma. In the group of serous ovarian carcinoma, there were higher frequencies of the Ile/Val and Val/Val genotypes in patients than in healthy controls (OR 7.0, 95% CI 3.30 – 15.22; and OR 8.7, 95% CI 2.83–27.14), respectively. In the group of mucinous ovarian carcinoma, the higher frequencies of patient with Ile/Val and Val/Val genotypes were also found in these patients (OR 5.4, 95% CI 1.56 –18.75; and OR 10.6, 95% CI 2.16 –51.19), respectively. However, no relation was observed with Ile/Val and Val/Val genotypes in the group of endometroid cancer patients (Table 1). Although small number of subjects had been observed in subgroups, some of them reached a statistically significant increase in relation to risk association. On the other hand, the frequency of individuals carrying any Val allele (Ile/Val and Val/Val genotype) was compared to controls and is shown Table 1. DISCUSSION The large difference in the CYP1A1 frequencies has been noticed among races. The CYP1A1 Val/Val genotype is most frequent in Asians than in the Caucasian population [15, 24, 25]. We have previously demonstrated the CYP1A1*3 polymorphism in Turkish population [20]. In that study, we have found that the proportion was 4.4% for the Val/Val genotype, which was much lower than in the Asian population. However, the prevalence of this allele was higher than in the German (2.8%) and Poland (2.2%) populations [24, 25]. In the present study, we observed that women with Ile/Val and Val/Val genotype were at 5.7 and 5.8 times the risk of epithelial ovarian neoplasm. Furthermore, Ile/Val and Val/Val genotypes were six- and four-fold higher than in benign tumor patients with Ile/Ile genotypes. Our finding of an association of the CYP1A1 Val allele indicates that individuals carrying this allele may have a higher risk of developing benign ovarian tumors. We also demonstrated a statistically significant association between Ile/Val genotype and borderline tumors. In the group of ovarian carcinomas, patients with Ile/Val and Val/Val allele were at 5.7 and 7.4 times the risk of ovarian carcinoma compared with Ile/Ile genotypes, implying that patients carrying any Val allele are at a greater risk of developing epithelial ovarian neoplasm. A nonsignificant effect between CYP1A1*3 polymorphism and endometroid cancer was observed. On the other hand, two valine alleles should be imply a greater risk for ovarian carcinomas than one allele (Table 1); we would suggest that this finding be viewed as a research area requiring further study, ideally with a larger sample size.

127

To our knowledge, there has been only one previous study evaluating the association between CYP1A1 polymorphism and ovarian cancer risk [26]. In the case-control study that consisted of 36 Caucasian, 58 Asian, and 35 “other” (mainly Native Hawaiian) ovarian cancer cases and 42 Caucasian, 69 Asian, and 33 “other” controls, Goodman et al. has reported some ethnic heterogeneity in the frequency of the variant alleles. The distributions of CYP1A1 polymorphism in ovarian cancer patients and controls in that report were 4 versus 5% for Val/Val allele and 23 versus 24% for Ile/Val allele, and no association of ovarian cancer patients with the CYP1A1 genotype has been observed [26]. Polymorphism may be associated in different ways with each other and with other unknown polymorphisms in various ethnic groups. In the present report, the genotype frequencies in controls were consistent with the Hardy-Weinberg equilibrium and all of our patients and control subjects were from the Turkish population as previously stated. Furthermore, the results of our study clearly demonstrated strong associations between Ile/Val and Val/Val genotype in epithelial ovarian neoplasm. Due to the high incidence of ovarian cancer and the lack of clear definitive associations with many of the proposed risk factors, the study of genetic susceptibility to environmental and endogenous factors such as estrogens may help to define individuals at higher risk of ovarian cancer. The CYP1A1 is expressed in a number of steroid hormone-responsive tissues including ovary, breast, brain, kidney, and prostate [9]; it activates polycyclic aromatic hydrocarbons of estradiol (E2) at the C-2, C-6␣, and C-15␣ positions in several extrahepatic tissues including epithelial cells [9, 27]. In this report, we found a positive association between CYP1A1 Val allele with the risk of epithelial ovarian cancer. Recently, the different steroid hydroxylation activities have been reported in all allelic variants of CYP1A1 [28]; we suggest here that these differences in CYP1A1*3 polymorphism may be a marker of altered estradiol metabolism and of increased susceptibility to ovarian cancer. In conclusion, our data demonstrate the novel finding of an increase the frequency of the Val allele in a series of epithelial ovarian neoplasm patients. The results suggest that variant alleles of CYP1A1 gene in ovarian epithelial cells, directly or through other components, may contribute to initiation of ovarian carcinogenesis. Although there is still much to understand concerning both the molecular and the cellular events, the data present interesting and valuable information for future research. REFERENCES 1. Young RC, Perez CA, Hoskins WJ. Cancer of ovary. In: DeVita VT Jr, Hellman S, Rosenberg, SA, editors. Cancer: principles and practice of oncology. Philadelphia: Lippincott Williams & Wilkins, 2000:1226 – 63. 2. Cramer DW, Welch WR. Determinants of ovarian cancer risk. II. Inference regarding pathogenesis. J Natl Cancer Inst 1983;71:7171–721.

128

AKTAS ET AL.

3. Clinton GM, Hua W. Estrogen action in human ovarian cancer. Crit Rev Oncol Hematol 1997;25;1–9. 4. Riman T, Persson I, Nilsson S. Hormonal aspects of epithelial ovarian cancer: review of epidemiological evidence. Clin Endocrinol (Oxford) 1998;49:695–707. 5. Cavalieri EL, Stack DE, Devanesan PD, Todorovic R, Dwivedy I, Higginbotham S, Johansson SL, Patil KD, Gross ML, Gooden JK, Ramanathan R, Cerny RL, Rogan EG. Molecular origin of cancer: catechol estrogen-3,4-quinones as endogenous tumor initiators. Proc Natl Acad Sci USA 1997;94:10937– 42. 6. Chakravarti D, Mailander P, Cavalieri EL, Rogan EG. Estrogen-DNA damage in mouse skin H-ras gene is mutated by error-prone repair. Proc Am Assoc Cancer Res. 2000;41:107. 7. Nandi S, Guzman RC, Yang J. Hormones and mammary carcinogenesis in mice, rats and humans: a unifying hypothesis. Proc Natl Acad Sci USA 1995;92:3650 –7. 8. Ball P, Knuppen R. Catechol estrogens (2-and 4-hydroxyestrogens): chemistry, biogenesis, metabolism, occurrence and physiological significance. Acta Endocrinol Suppl 1980;93:1–127. 9. Martucci CP, Fishman J. P450 enzymes of estrogen metabolism. Pharmacol Ther 1993;57:237–57. 10. Zhu BT, Conney AH. Functional role of estrogen metabolism in target cells: review and perspective. Carcinogenesis 1998:19:1–27. 11. Chai XM, Conney AH, Zhu BT. 17␤-Estradiol metabolism by selectively expressed human cytochrome P450. Proc Am Assoc Cancer Res 1998;39: 2626. 12. Stack D, Byun J, Gross ML, Rogan EG, Cavalieri E. Molecular characteristics of catechol estrogen quinones in reactions with deoxyribonucleosides. Chem Res Toxicol 1996;9:851–9. 13. Li KM, Devanesan PD, Rogan EG, Cavalieri EL. Formation of the depurinating 4-hydroxyestradiol (4-OHE2)-1-N7Gua and 4-OHE2– 1.N3Ade adducts by recation of E2–3,4-quinone with DNA. Proc Am Assoc Cancer Res 1998;39:636. 14. Nebert DW, McKinnon RA, Puga A. Human drug-metabolizing enzyme polymorphisms: effect on risk of toxicity and cancer. DNA Cell Biol 1996;15:273– 80. 15. Kawajiri K, Nakachi K, Imai K, Yoshii A, Shinoda N, Watanabe J. Identification of genetically high risk individuals to lung cancer by DNA polymorphism of cytochrome P4501A1 gene. FEBS Lett 1990;63:131–3. 16. Kawajiri K, Nakachi KI, Imai K, Watanabe J, Hayashi S. The CYP1A1 gene and cancer susceptibility. Crit Rev Oncol Hematol 1993;14:77– 87.

17. Hamada GS, Sugimura H, Suzuki I, Nagura K, Kiyokawa E, Iwase T, Tanaka M, Takahashi T, Watanabe S, Kino I. The heme-binding region of CYP1A1 rather than the RsaI polymorphism of CYP2E1, is associated with lung cancer in Rio de Janeiro. Cancer Epoidemiol Biomark Prev 1995;4:63–7. 18. Hayashi SI, Watanabe J, Nakachi K. PCR detection of an A/G polymorphism within exon 7 of the CYP1A1 gene. Nucleic Acids Res 1991;19: 4797. 19. Hirvonen A, Husgafvel PK, Karjalainen A, Anttila S, Vainio H. Point mutational MspI and Ile-Val polymorphism closely linked in the CYP1A1 gene: lack of association with susceptibility to lung cancer in a Finnish study population. Cancer Epidemiol Biomark Prev 1992;1:485–9. 20. Aktas D, Hascicek M, Sozen S, Ozen H, Tuncbilek E. CYP1A1 and GSTM1 polymorphic genotypes in benign prostatic hyperplasia and prostate cancer patients. Eur J Hum Genet 2001;9(Suppl 1):111. 21. Miller SA, Dykes DD, Polesky HF. A simple out procedure for extracting DNA from human nucleated cells. Nucleic Acids Res 1988;16:1215. 22. Jaiswal AK, Gonzalez FJ, Nebert DW. Human P1-450 gene sequence and correlation of mRNA with genetic differences in benzo(a) pyrene metabolism. Nucleic Acids Res 1985;13:4503–20. 23. Lathrop G. Estimating genotype relative risks. Tissue Antigens 1983;22: 160 – 6. 24. Cascorbi I, Brocmoller J, Roots I. A C4887A polymrophism in exon 7 of human CYP1A1: population frequency, mutation linkages, and impact on lung cancer susceptibility. Cancer Res 1996;56:4965–9. 25. Mrozikiewicz PM, Landt O, Cascorbi I, Roots I. Peptide nucleic acidmediated polymerase chain reaction clamping allows allelic allocation of CYP1A1 mutations. Anal Biochem 1997;250:256 –7. 26. Goodman MT, McDuffie K, Kolonel LN, Terada K, Donlon TA, Wilkens LR, Gua C, Marchand LL. Case-control study of ovarian cancer and polymorphism in genes involved in catecholestrogen formation and metabolism. Cancer Epidemiol Biomark Prev 2001;10:209 –16. 27. Spink DC, Eugster HP, Lincoln DW, Schuetz JD, Schuetz EG, Johnson JA, Kaminsky LS, Gierthy JF. 17␤-Estradiol hydroxilation catalzed by human cytochrome P4501A1: a comparison of the activities induced by 2,3,7,8tetrachlorodibenzo-p-dioxin in MCF-7 cells with those from heterolous expression of the cDNA. Arch Biochem Biophys 1992;293:342– 8. 28. Schwarz D, Kisselev P, Schunck WH, Chernogolov A, Boidol W, Cascorbi I, Roots I. Allelic variants of human cytochrome P4501A1 (CYP1A1): effect of T461N and I462V subctitutions on steroid hydroxylase specificity. Pharmacogenetics 2000;10:519 –30.