Debrisoquine Hydroxylase (CYP2D6) and Prostate Cancer'

Vol. 7, 1075-1078, December 1998 Cancer Epidemiology, Biomarkers & Prevention /075

Debrisoquine Hydroxylase (CYP2D6) and Prostate Cancer’

Phillip G. Febbo, Philip W. Kantoff,2 cological agents (Table 1 ; Refs. 3 and 4). Individuals who are Edward Giovannucci, Myles Brown, Gloria Chang, PMs3 are very sensitive to the drugs hydroxylated by this Charles H. Hennekens, and Meir Stampfer enzyme and can have life-threatening side effects from them Lank Center for Genitourinary Oncology, Dana Farber Cancer Institute. (5). Normal individuals are known as RMs. There is a third Harvard Medical School, Boston. Massachusetts 021 15 IP. G. F., P. W. K., group of patients recently identified as ultrarapid metabolizers M. B., G. Cl; and Division of Preventive Medicine IC. H. H.] and Channing who have multiplication of the active CYP2D6 gene (6). Laboratory EE. G., M. S.], Department of Medicine, Brigham and Women’s Hospital and Harvard Medical School, Physician’s Health Study, and The variation in individual susceptibility to carcinogens, Departments of Epidemiology IC. H. H.] and Nutrition IE. G., M. S.], Harvard such as those found in tobacco, stimulated interest in associa- School of Public Health, Boston. Massachusetts 021 15 tions between polymorphic p450 expression and carcinogenesis (7). Initial studies, using metabolic ratios of excreted un- changed debrisoquine and its 4-hydroxy metabolite to identify Abstract the PM phenotype, demonstrated that PM individuals are less The p450 hepatic microsomal enzyme system metabolizes likely to develop lung cancer (7, 8). Similar phenotypic analysis exogenous dnigs and carcinogens. Debrisoquine hydroxylase of the debrisoquine polymorphism suggested that they also (CYP2D6), one member of the p450 hemoproteins, has have a decreased risk of aggressive bladder cancer (9). It has polymorphic expression leading to poor metabolism of been proposed that this association is the result of decreased debrisoquine and similar compounds in approximately 7% activation of procarcinogens by the CYP2D6 enzyme (1). of Caucasians. The genetic locus for this enzyme has been The CYP2D6 locus has been mapped to 22q I 3. 1 ( 10, 1 1) characterized, and the mutations responsible for the slowed and includes the intact, active gene referred to as CYP2D6 as metabolism have been identified. Epidemiological studies of well as two pseudogenes CYP2D7P and CYP2D8P, which the CYP2D6 phenotype suggest an association between the contain inactivating mutations ( 12). The most common muta- normal or rapid metabolism phenotype and increased risk tions responsible for the PM phenotype have been identified of lung and bladder cancer. Preliminary data have also within the CYP2D6 locus (4, 1 1, 13). A characteristic G to A suggested an association with prostate cancer (CaP). We transition mutation at thejunction of exon 3 and intron 4, which used a PCR-based assay to investigate possible associations leads to an early termination codon and defective mRNA, between the CYP2D6 B allele, the most common genetic accounts for approximately 80% of mutant alleles ( 1 1). This mutation responsible for the poor metabolism phenotype, has been designated as the B allele of CYP2D6. Deletion of the and CaP. Using genomic DNA isolated from peripheral entire CYP2D6 gene (the D allele) accounts for another 15% of blood, we genetically typed 571 men with CaP and 767 PM alleles, and most of the remaining PM alleles are accounted matched controls, all participants in the Physician’s Health for by a combination of three specific mutant alleles (14). Study. Relative to men homozygous for the wild-type allele, Individuals who have the PM phenotype are most com- heterozygotes for the B allele have an odds ratio of 1.19 monly homozygous for inactivating mutations. However, there (95% confidence interval, 0.94-151) for CaP, and men is a spectrum of debrisoquine metabolism ranging from mdi- homozygous for the B allele have an odds ratio of 1.37 (95% viduals with two wild-type alleles and rapid metabolism, to confidence interval, 0.86-2.20). When analyzed as a trend heterozygotes with a single mutant allele and an intermediate over zero, one, or two copies of the B allele, there emerges a metabolic rate, to individuals homozygous for mutant alleles possible association between the B allele and an increased who are PMs (15). Although there is considerable overlap in risk of CaP of borderline statistical significance (P 0.07). phenotype between individuals homozygous and those heterozygous for the wild-type allele, a statistically significant Introduction “gene dose” effect has been observed ( I 6). The cytochrome p450 enzymes metabolize a wide range of Assays using PCR and RFLPs have supported earlier exogenous and endogenous organic compounds (1). Many phenotypic studies associating the PM phenotype with a de- pharmacological agents used in medical practice are also sub- creased risk of lung cancer ( 1 7, 18) and suggest a decreased risk strates for p450 enzymes (2). Approximately 7% of Caucasians of liver cancer ( 19) have a specific p450 enzyme defect that leads to slow metab- Only sparse data link CYP2D6 and CaP. In a preliminary olism of debrisoquine as well as more than 30 other pharma- report of patients with a variety of malignancies, 9.3% of patients with CaP (n = 54) were homozygous for a mutant

allele compared with 4.3% of individuals without cancer (ii = 720; Refs. 20 and 21), for a relative risk of 2.3. However, Received 1/8/98: revised 8/27/98; accepted 9/14/98. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. I Supported by NIH/National Cancer Institute Grants CA42182, CA58684, and CA72036 and by the Dana Farber Prostate Cancer Research Fund. 3 The abbreviations used are: PM, poor metabolizer: CaP. prostate cancer: 2 To whom requests for reprints should be addressed. at Dana Farber Cancer CYP2D6, debrisoquine hydroxylase; RM. rapid metabolizer; NNK. 4-(methylni- Institute, 44 Binney Street, Boston, MA 021 15. trosamino)- I-(3-pyridyl)-I -butanone: OR, odds ratio: CI. confidence interval.

Table I Known subs trates for CYP2D6 and Gleason score were recorded. Cases without pathological staging were classified as indeterminate stage, unless there was Debnsoquine Timolol clinical evidence of distant metastases. “Aggressive” cases Sparteine Haloperidol were defined as those diagnosed either with stage C or D Encainide Perhexiline (extraprostatic) CaP or with poor histological differentiation or Bufuralol Clozapine Metoprolol Deprenyl Gleason score 7. Thus, cases with clinical stage A or B, no Amitriptyline Ethylmorphine pathological staging, and moderate or better histological grade Dextromethorphan Guanoxan were classified as indeterminate aggressiveness. Aminflamine Lignocaine For each case of CaP identified, one or two controls were cndeine NNK selected at random from men of the same age and smoking Ecstasy (MDMA) status who had provided blood, had not had a previous prosta- tectomy, and had not reported a diagnosis of CaP at the time of diagnosis that was reported by the case. Two very elderly cases were matched by age within 2 years. We initially matched 589 perhaps due to the small number of cases, this was not statis- cases to 590 controls by age and smoking (a case was subse- tically significant. quently found not to have CaP and was excluded accounting for We used PCR to identify the presence of the CYP2D6 B the extra control). Subsequently, 204 additional controls were allele in men with CaP, and matched controls within the Phy- included in the study. These controls had been matched with sician’s Health Study to explore the association between the 204 of the original cases as part of an independent study PM phenotype and the risk of CaP. requiring the availability of serum hormone levels. Thus, although most of the cases were matched to a single control, 204 Materials and Methods cases were matched to two controls. Study Population. The Physician’s Health Study is a random- Laboratory Analysis. Coded whole blood samples from cases ized, double-blind, placebo-controlled trial investigating long- and matched controls were received from the Physician’s term use of aspirin and /3-carotene among 22,07 1 American, Health Study, with the laboratory investigators blinded to the predominantly white (97% Caucasian), male physicians, ages name and case status of each sample. Genomic DNA was 40-84 years, in 1982. Men were excluded if they reported a obtained from 500 1 of the thawed whole blood using a prior history of myocardial infarction, stroke, transient ische- commercially available kit (QIAamp DNA extraction kit; Qia- mic attacks, unstable angina, cancer (except for nonmelanoma gen, Inc., Chatsworth, CA). DNA concentration and purity skin cancer), present renal or liver disease, peptic ulcer disease were determined by UV absorbance on a Beckman DU640 or gout, contraindications to use of aspirin, or present use of spectrophotometer. Each sample was diluted to a final concen- aspirin or other platelet-active agents or vitamin A supple- tration of 20 ng/l and stored at -20#{176}Cuntil analysis. ments. The aspirin arm was terminated in January 1988 due to To identify the polymorphic foci for each case and control, 80 a statistically extreme reduction in the risk of a first myocardial ng of sample DNA was added to the PCR reaction mixture, which infarction among those in the aspirin group (22). included primers 5’-GCTTCGCCAACCACTCCG-3’ and 5’-A- Blood Collection. Study participants completed two mailed AATCCFGCTCTfCCGAGGC-3’ (Gough et a!., 1990) at a con- questionnaires before randomization in 1982 and additional ques- centration of 0.25 LM each along with 50 m’vi KCL, 1.5 mM tionnaires at 6 months, 12 months, and annually thereafter. Before MgC12, 250 pM each dN’FP, and 1 .0 units of AmpliTaq (Perkin- randomization, blood kits were sent to all participants with instruc- Elmer Corp.) in a final volume of 22 tl. All amplifications were lions to have their blood drawn into vacutainer tubes containing performed using MicroAmp tubes (Perkin-Elmer Corp.). EDTA, centrifuged, and to have the plasma and a vial of whole A Perkin-Elmer GeneAmp PCR System 9600 thermocy- blood returned (by overnight prepaid courier) in polypropylene cler was programmed for a two-step PCR. After 4 mm at 94#{176}C, cryopreservation vials. The kit included a cold pack to keep the samples underwent 35 cycles with a melting step at 94#{176}Cfor specimens cool until receipt the following morning, when they 30 s and an annealing and elongation step at 68#{176}Cfor1 mm and were aliquoted and stored at -82#{176}C.Specimens were received 30 s. There was a final elongation step for 8 mm at 72#{176}C,and from 14,916 (68%) of the randomized physicians; over 70% be- samples were then cooled to 4#{176}C. tween September and November 1982. After amplification, 12 tl of amplified product was digested Case Identification and Control Matching. When a partici- with 4 units of BstOl according to the manufacturer’s recommen- pant reported a diagnosis of CaP on any follow-up question- dations (New England Biolabs, Beverly, MA). Digested product naire, relevant medical records, including pathology reports, was separated using a I .6% agarose gel containing ethidium bro- were obtained and reviewed by study physicians from the End mide. Genotype was based on banding pattern (See Fig. 1). Points Committee. Data Analysis. We conducted analyses to determine whether By 1996, 589 cases of CaP were confirmed among the polymorphisms in CYP2D6 were related to CaP risk. In addi- men who had provided blood in 1982. After 10 years of follow- tion to total CaP, we conducted analyses of tumors that pos- up, over 99% of the study participants were still reporting sessed a more aggressive phenotype, as determined by histo- morbidity events, and vital status was ascertained for 100%. logical grade and tumor stage. In the initial set of matched cases Overall mortality in the cohort was low, minimizing censoring and controls, results were similar using conditional logistic due to loss-to-follow-up or death. regression and unconditional logistic regression controlling for A study physician reviewed medical records for each case age and smoking status. Hence, we used unconditional logistic to determine the stage at diagnosis, tumor grade, Gleason score, regression, controlling for age and smoking, the matching van-

type of presentation (e.g. , symptoms, screening rectal exam, ables, to compute the relative risk (estimated by the OR) of CaP etc.), and treatment modalities. Stage was recorded according to and 95% CIs. By using unconditional logistic regression, we the modified Whitemore-Jewett classification scheme. If mul- were able to use the entire set of controls for analyses that used

tiple tissue samples were examined, the highest reported grade only a subset of the cases (e.g. , aggressive cancers), which

Size 1 2 3 Size Table 2 Prevalence of CYP2D6 genotypes among CaP patients and controls Marker wt/wt wI/mt mt/mt Marker Hornozygous for Heterozygous Homozygous for the RM allele (n) fl) PM allele (ii)

Patients with CaP 62.2’7 (355) 31.4% (179) fi5(7 37) ! (ml 571) Controls 66.2% (511) 283rk (217) SIck (39) (‘1 767)

Fig. 1. PCR products of representative genotypes after digestion with BstOI. number of CYP2D6 B alleles in a small number of men with The size marker used is HaeIII-digested PhiX; Lane 1, homozygous wild type prostate cancer when compared with individuals without cancer. (wtlwt): Lane 2, heterozygous (wt/mt); Lane 3. homozygous mutant (PM: mt/mi). An increased risk of cancer was also associated with the PM phenotype when women with breast cancer were studied (23). In vitro studies have recently demonstrated that a nitrosamine found in tobacco smoke, NNK, is metabolized by maximized our power. All reported Ps are two-sided. The trend CYP2D6 (among other p450 hemoproteins) when the cDNA is across categories of RM (homozygous), RM (heterozygous), cloned into and expressed by a B-lymphoblastoid cell line (24). and PM was tested by using a variable with the values 1 , 2, and This nitrosamine is a known rodent pulmonary carcinogen (25), 3, respectively, in a logistic regression model. and altered metabolism of NNK may help explain the association between lung and bladder cancer and CYP2D6. However, Results no definite carcinogen for prostatic epithelium has been iden- rifled (26). Thus, the significance of poor metabolism of an Of the 1383 samples, 1338 were successfully amplified. The exogenous compound such as NNK is unclear in CaP. A per- inability to amplify the 45 samples was attributed to poor quality sonal history of smoking did not seem to influence risk of CaP DNA based on 0D260:280 ratios, but we cannot exclude the according to the CYP2D6 genotype. possibility that some of the samples that failed to amplify were Testosterone and its 5-hydroxy metabolite, dihydrotestos- homozygous for large deletions of the CYP2D6 foci. These dele- terone, play a key role in the development ofCaP (27). Changes tions are uncommon; homozygosity for such an abnormality has in the metabolism of androgens have the potential to affect the an estimated frequency of approximately 0.0015 and probably risk of CaP. There is no direct evidence supporting CYP2D6 accounts for very few of the samples failing to amplify. hydroxylase having androgens as substrates. However, dark The overall percentage of poor metabolizers among con- DA rats, deficient in CYP2D6 activity, have an associated trols was 5.1% (39 of 767), which is what is expected of a decrease in hepatic hydroxylation of testosterone and proges- mostly Caucasian population. Among cases, the proportion was terone (28). Alterations in the hydroxylation of testosterone is slightly higher at 65% (37 of 57 1 ; Table 1). We found a greater a potential mechanism for the increased risk of CaP with the proportion of heterozygotes (OR, 1.19; CI, 0.94-1 .5 1 ; P = PM phenotype. The PM phenotype also is associated with an 0.16) and homozygotes (OR, 1.37; CI, 0.86-2.20; P = 0.19) for increased risk of breast cancer (23), another cancer with hor- the mutant allele among CaP patients (Table 2). We found a monal dependence. Decreased hydroxylation of androgens and trend (P = 0.07) of borderline statistical significance for in- estrogens may lead to delayed clearance and result in higher creasing risk with an increasing number of mutant alleles across exposure of the prostate and breast to these compounds. How- categories of RM (homozygous), RM (heterozygous), and PM ever, in the subset of subjects with available testosterone levels, genotypes. The relative risk for aggressive CaP was similar to we found no clear correlation between the CYP2D6 genotype the overall risk of CaP among patients who were homozygous and androgen levels (data not provided). for the PM allele (OR, 1 .32; CI, 0.72-2.44); Table 3). Another possible explanation for an association between the Smoking history was examined in relation with the B allele of CYP2D6 and CaP is in linkage disequilibrium with an CYP2D6 genotype and CaP. Of the 1338 men, 729 had been or were presently smoking tobacco cigarettes. The relative risk for unidentified gene that alters the risk of CaP. Recent linkage analysis places the CYP2D6 gene on the 22 chromosome in the region CaP among men homozygous for the PM allele of CYP2D6 of 22q13.l (29). This is in close proximity to the platelet-derived who never smoked was 1 .29 (CI, 0.66-2.5 1), similar to the 1.48 growth factor subunit gene (maximum sex-average lod score, (CI, 0.76-2.85) for those with a smoking history. /3 0.80; 0 = 0.036) and to translocation sites for acute lymphocytic leukemia, t(9;22); chronic myelocytic leukemia, t(9;22); and Ew- Discussion ing’s sarcoma, t( 1 1;22; Ref. 29). The participation of genes in this These data show no overall statistically significant association region with regard to CaP development is unknown. between the B allele and CaP, but a small increase in the risk The mutation in CYP2D6 analyzed here is the most com- of CaP in men with one or two copies of the B allele of mon. but not the only mechanism leading to the autosomal

CYP2D6, of borderline significance (P = 0.07), was observed. recessive trait associated with poor metabolism of debriso- This analysis is based on a trend in metabolic phenotype from quine. The PCR misclassifies approximately 20-30% of alleles

RM (homozygote, wild type), RM (heterozygote), and PM leading to the PM phenotype (I 1 , 14). Assuming an incidence (homozygote, mutant; Ref. 15). of 7% PM phenotype in Caucasian populations, the frequency Although earlier studies of the CYP2D6 phenotype and can- of alleles leading to PM of debrisoquine would be 0.26. By cer suggested a protective effect of PMs on the risk of lung, using an assay that correctly identifies 80% of these alleles, the bladder, and liver cancer, our study suggests that the B allele observed allele frequency would be 20% and the observed leading to the PM phenotype is associated with an increased risk homozygotes for the allele is reduced to 4%. This assumes that of CaP. This is in accordance with the preliminary studies by Wolf all mutant alleles not correctly identified are considered normal. et a!. (20) and Dale Smith et a!. (21) that found an increased This misclassification of exposure status moderately reduces

Table 3 ORs and 95% CIs for risk of CaP according to CYP2D6 genotype

Patient population RM (homozygous) RM (heterozygous) PM P

Risk ofall caP l.0(reference) 1.19(0.94-1.51) 1.37 (0.86-2.20) 0.07 (No. cases/no. control) (355/5 1 1 ) ( 179/217) (37/39) Risk of aggressive CaP I .0 (reference) I .24 (0.91-1 .69) 1 .32 (0.72-2.44) 0.14

( No. cases/no. control ) ( I 62/5 I I ) (85/2 17 ) ( I 6/39)

our statistical power, but should not affect the validity of the 10. Eichelbaum, M., Baur, M., Osikowska-Evers, B., Zekom. C., and Rittner, C. association between CaP and CYP2D6. There is no reason to Chromosomal assignment of human cytochrome P450 (debrisoquine/sparteine type) to chromosome 22. Br. J. Clin. Pharmacol., 23: 455-458, 1987. suspect that other mutations not identified would bias the associ- II. Gough. A., Miles. J.. Spun’. N.. Moss. J.. Gaedigk, A., Eichelbaum, M., and ation suggested in this study. Some of the samples that fail to Wolf. C. Identification of the primary gene defect at the cytochrome P450 CYP2D amplify may contain other CYP2D6 mutations, including deletion locus. Nature (Lond.), 347: 773-776, 1990. of the entire gene, but there is no reason to suspect that they would 12. Kimura, S., Umeno, M., Skoda, R., Meyer, U.. and Gonzalez, F. The human bias the association we observed with the mutations we assessed. debrisoquine 4-hydroxylase (CYP2D) locus: sequence and identification of the More accurate identification of each individual’s pheno- polymorphic CYP2D6 gene, a related gene. and a pseudogene. Am. J. Hum. type may strengthen the association between CYP2D6 and CaP. Genet., 45: 889-904, 1989. Expanding the genetic analysis to identify the most common I 3. Heim, M., and Meyer. U. Genotyping of poor metabolisers of debrisoquine by allele-specific PCR amplification. Lancet, 336: 529-532, 1990. mutant alleles leading to the PM phenotype can achieve a 95% 14. Nebert, D., McKinnon, R.. and Puga, A. Human drug-metabolizing enzyme concordance between identified genotype and expressed phe- polymorphisms: effects on risk of toxicity and cancer. DNA Cell Biol., 15: notype (13, 19). 273-280. 1996.

Because the polymorphism is relatively uncommon, this 15. Mura, C.. Panserat. S., Vincent-Viry. M.. Galteau. M.. Jacqz-Aigrain. E., and study has only modest power to detect relatively small to Krishnamoorthy. R. DNA haplotype dependency of debrisoquine 4-hydroxylase moderate effects. Although neither relative risk estimate for (CYP2D6) expression among extensive metabolisers. Hum. Genet.. 92: 367-372, heterozygotes or homozygotes was statistically significant, the 1993. 95% CI (0.94-I .5 1 for heterozygotes) indicate the range of 16. Broly. F., Gaedigk, A., Heim, M., Eichelbaum, M., Morike, K., and Meyer, U. Debrisoquine/sparteine hydroxylation genotype and phenotype: analysis of relative risk estimates that are compatible with the data. common mutations and alleles of CYP2D6 in a european population. DNA Cell These data are compatible with a possible modest associ- Biol., 10: 545-558, 1991. ation between the most common mutant allele leading to poor I 7. Gonzalez, F., and Meyer, U. Molecular genetics of the debrisoquin-sparteine metabolism of debrisoquine and CaP of borderline significance. polymorphism. Clin. Pharmacol. Ther.. 50: 233-238, 1991. Although there is no known biological mechanism for such an I 8. Dolzan, V., Rudolf, Z., and Breskvar, K. Human CYP2D6 gene polymor- association, metabolism of exogenous toxins, endogenous an- phism in Slovene cancer patients and healthy controls. Carcinogenesis (Lond.), drogens, or linkage to an unidentified locus affecting an mdi- 16: 2675-2678, 1995. vidual’s risk of CaP are possibilities. Additional studies ad- 19. Aqundez. J.. Ledesma, M., Benitez, J.. Ladero, J., Rodriquez-Lescure, A., Diaz-Rubio, F., and Diaz-Rubio, M. CYP2D6 genes and risk of liver cancer. dressing the association between mutations of the CYP2D6 Lancet, 345: 830-831, 1995. gene and the PM phenotype and risk for CaP are needed. 20. Wolf, C.. Dale Smith. C.. Gough, A.. Moss, J., Vallis, K., Howard, G., Carey, F., Mills, K., McNee, W., Carmichael. J., and Spun. N. Relationship between the Acknowledgments debrisoquine hydroxylase polymorphism and cancer susceptibility. Carcinogen- esis (Lond.), /3: 1035-1038, 1992. We thank Unsed Ajani for statistical analysis and program review. 21 . Dale Smith, C., Moss, J.. Gough, A., Spun’. N., and Wolf, C. Molecular genetic analysis ofthe cytochrome p450-debrisoquine hydroxylase locus and association with References cancer susceptibility. Environ. Health Perspect., 98: 107-1 12. 1992. 22. The Steering Committee for the Physician’s Health Study. Preliminary re- I . Smith, G.. Stanley. L. A.. Sim. D.. Strange. R. C.. and Wolf, C. R. Metabolic port: findings from the aspirin component of the ongoing Physician’s Health polymorphisms and cancer susceptibility. Cancer Surv.. 25: 27-63, 1995. Study. N. EngI. J. Med.. 3/8: 262-264. 1988. 2. Idle, J., and Smith, R. Polymorphisms of oxidation at carbon centers of drugs and their clinical significance. Drug Metab. Rev.. 9: 301-317, 1979. 23. Ladero, J., Benitez, J., Jara, C., Llerena, A., Valdivielso, M., Munoz, J., and Vargas. E. Polymorphic oxidation of debrisoquine in women with breast cancer. 3. Mahgouh. A.. Idle. J.. Dring, L.. Lancaster, R., and Smith. R. Polymorphic Oncology,48: 107-110, 1991. hydrozylarion of debrisoquine in man. Lancet, ii: 584-586, 1977. 24. Crespi. C.. Penman, B., Gelboin, H., and Gonzalez, F. A tobacco smoke- 4. Gonzalez. F.. Skoda, R., Kimura, S., Umeno, M., Zanger. U. Nebert. D.. derived nitrosamine, 4-(methylnitrosamino-I -(3-pyridyl)- 1 -butanone, is activated Gelboin. H.. Hardwick, J., and Meyer. U. Characterization of the common genetic by multiple human cytochrome P450s including the polymorphic human cyto- defect in humans deficient in debrisoquine metabolism. Nature (Lond.), 331: chrome P4502D6. Carcinogenesis (Lond.), 12: 1 197-1201, 1991. 442-446. 1988. 25. Hoffman, D.. Castonguay, A., Rivenson, A., and Hecht, S. Comparative 5. Eichelbaurn, M. Polymorphic drug oxidation in humans. Fed. Proc.. 43: 2298-2302. 1984. carcinogenicity and metabolism of 4-(methylnitrosamino- I -(3-pyridyl)- I - butanone and N’-nitrosonomicotine in Syrian golden hamsters. Cancer Res., 41: 6. Johansson. I., Lundqvist. F.. Bertilsson. L.. DahI, M., Sjoqvist, F.. and Ingle- 2386-2393, 1981. man-Sundberg, M. Inherited amplification of an active gene in the cytochrome P450 CYP2D locus as a cause of ultrarapid metabolism of debrisoquine. Proc. 26. Giovannucci, E. Epidemiologic characteristics of prostate cancer. Cancer NatI. Acad. Sci. USA. 90: 1 1825-1 1829, 1993. (Phila.), 75: 1766-1777. 1995. 7. Ayesh. R.. Idle. J.. Ritchie. J., Crothers. M.. and Hetzel. M. Metabolic 27. Coffey, D. Physiological control of prostatic growth: an overview. In: Pros- oxidation phenotypes as markers for susceptibility to lung cancer. Nature (Lond.), tate Cancer: UICC Technical Report Series, pp. 4-23. Geneva: International 3/2: 169-171. 1984. Union Against Cancer. 1979. 8. Sugirnura. H., Caporaso. N., and Shaw, G. Human debrisoquine hydroxylase 28. Kupfer. A., and Preisig. R. Inherited defects of hepatic drug metabolism. gene polymorphism in cancer patients and controls. Carcinogenesis (Lond.). //: Semin. Liver Dis., 3: 341-354. 1983. 1527-1530. 1990. 29. Gough, A., Dale Smith, C., Howell, S., Wolf, C., Bryant, S., and Spurt. N. 9. Kaisary. A.. Smit. P.. Jaczq. E.. McAllister. C.. Wilkinson. G., Ray. W.. and Localization of the CYP2D gene locus to human chromosome 22ql3.l by Branch. R. Genetic predisposition to bladder cancer: ability to hydroxylate polymerase chain reaction. in sitim hybridization. and linkage analysis. Genomics, debrisoquine and mephenytoin as risk factors. Cancer Res.. 47: 5488-5493. 1987. 15: 430-432. 1993.

P G Febbo, P W Kantoff, E Giovannucci, et al.

Cancer Epidemiol Biomarkers Prev 1998;7:1075-1078.

Updated version Access the most recent version of this article at: http://cebp.aacrjournals.org/content/7/12/1075

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cebp.aacrjournals.org/content/7/12/1075. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.