CYP17 genotype predicts serum hormone levels among pre-menopausal women

Chanley M. Small1,5, Michele Marcus1, Stephanie L. Sherman2, Amy K. Sullivan2, Amita K. Manatunga3 and Heather Spencer Feigelson4

1 Department of Epidemiology, 2 Department of Human Genetics and 3 Department of Biostatistics, Emory University, Atlanta, Georgia and 4 American Cancer Society, Atlanta, Georgia, USA

5 To whom correspondence should be addressed at: Department of Epidemiology, Emory University, 1518 Clifton Rd., Atlanta, GA 30322, USA. Email: csmall{at}sph.emory.edu


    Abstract
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Acknowledgements
 References
 
BACKGROUND: CYP17, which encodes cytochrome P450c17{alpha}, mediates both steroid 17{alpha}-hydroxylase and 17,20-lyase activities, and is essential for the production of glucocorticoids and sex steroids. There is evidence that a common polymorphism in CYP17 (T27C) is associated with estrogen levels, making it a potential marker of disease risk. METHODS: This is the first study to examine the relationship between CYP17 and estradiol (E2) using serum sampled exclusively from the early follicular phase of the menstrual cycle. We assessed the relationship between CYP17 and serum hormone levels, menstrual cycle length, bleed length, and age at menarche in 164 pre-menopausal women. RESULTS: Among women with body mass index (BMI) ≤25 kg/m2, those with the TC and CC genotypes had 19 and 42% higher E2 (P for trend 0.007) and 14 and 30% higher dehydroepiandrosterone sulphate respectively (P for trend 0.10) than women with the TT genotype. Androstenedione levels did not differ between genotypes. Among women with BMI >25 kg/m2, hormone levels did not differ by genotype. Women with the C allele were also more likely to have menstrual cycle lengths <27 days [odds ratio (OR) for TC=2.36, 95% confidence interval (CI)=1.24–4.52; OR for CC=5.59, 95% CI=1.53–20.43 compared to TT]. CYP17 genotype was not associated with menstrual bleed length or age at menarche. CONCLUSION: The CYP17 T27C polymorphism may be a marker of endocrine function.

Key words: CYP17/estradiol/menstrual cycle/P450(c17)/sex hormones


    Introduction
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Acknowledgements
 References
 
CYP17, which encodes cytochrome P450c17{alpha}, mediates both steroid 17{alpha}-hydroxylase and 17,20-lyase activities, and is essential for the production of glucocorticoids and sex steroids (Figure 1). In women, CYP17 is expressed in the adrenals, adipose tissue, ovarian theca cells, and the corpus luteum (Sasano et al., 1989; Hanukoglu, 1992Go; Puche et al., 2002Go). Mutations in CYP17 cause 17-hydroxylase deficiency (17OHD), a rare form of congenital adrenal hyperplasia (Costa-Santos et al., 2004Go). Approximately 40 different mutations in CYP17 have been reported to cause 17OHD that is characterized by hypertension, hypokalaemia, and sexual infantilism (Martin et al., 2003Go; Costa-Santos et al., 2004Go). The absence of 17,20-lyase activity in these individuals results in impaired production of gonadal sex steroids and high levels of basal progesterone (which is a substrate of 17{alpha}-hydroxylase) (Martin et al., 2003Go). These severe phenotypes demonstrate the importance of CYP17 in the synthesis of androgens and estrogens.



View larger version (15K):
[in this window]
[in a new window]
 
Figure 1. Steroid hormone pathway.

 
Carey et al. (1994)Go first identified a single nucleotide polymorphism (T27C) in the 5’ untranslated region (UTR) of CYP17 and hypothesized that the C allele might up-regulate gene expression leading to increased levels of serum hormones, including androstenedione and estradiol (E2). This polymorphism is common; the CC genotype is present in 11–19% of North American whites and 6–16% of African-Americans (Sharp et al., 2004Go). Several studies support the hypothesis that the CYP17 C allele may be a marker of increased steroidogenesis. It has been associated with increased risk of recurrent spontaneous abortion (Sata et al., 2003Go), a condition associated with high androstenedione levels (Li et al., 2002Go). It has also been associated with decreased risk of intrauterine growth retardation (Yamada et al., 2004Go) which is consistent with the observed positive correlation between pregnancy estrogen levels and fetal size (Peck et al., 2003Go). Further, the C allele has been associated with increased risk of uterine leiomyomas (Amant et al., 2004Go) which is consistent with the proliferative effect of estrogen on the endometrium (Deligdisch, 2000Go).

The T27C polymorphism could affect estrogen levels through several postulated mechanisms. Carey et al. (1994)Go suggested that the addition of an Sp-1 type (CCACC box) recognition site may up-regulate gene transcription and increase circulating hormones, but molecular evidence does not support this (Nedelcheva Kristensen et al., 1999Go). Other possible mechanisms, including transcriptional elongation or effects on mRNA translation rates, have not been tested experimentally. Alternatively, up-regulation of gene expression may result from another functional site in linkage disequilibrium with T27C. Regardless of the mechanism for up-regulation, this polymorphism may prove to be a significant marker of disease risk if it is associated with estrogen levels in pre-menopausal women.

Few studies have directly assessed whether the T27C polymorphism is associated with hormone levels in pre-menopausal women, and the results have been inconsistent. Feigelson et al. (1998)Go first reported increased E2 and progesterone associated with the C allele, but two subsequent studies (Garcia-Closas et al., 2002Go; Travis et al., 2004Go) were unable to confirm this finding. These studies did not restrict serum collection to a limited range of cycle days as done by Feigelson et al. In all three studies, hormones were measured predominantly during the late follicular and luteal phases. Because menstrual cycle length, particularly follicular phase length, varies among women, measurement of hormones during the late follicular and luteal phases may be confounded by cycle length. The goal of this study was to examine whether differences in hormone levels by genotype are present in the early follicular phase. Because previous studies have shown differential tissue-specific regulation of CYP17 in the ovary and adipose tissue (Puche et al., 2002Go), we also explored whether the effect of CYP17 on hormone levels is modified by body mass index (BMI).


    Materials and methods
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Acknowledgements
 References
 
Study population
CYP17 study participants were drawn from a larger study examining the relationship between the repeat size polymorphism in the 5’ UTR of the fragile X mental retardation (FMR1) gene and female reproductive ageing. All of the women included in this analysis had FMR1 repeat size ≤150. Specific details on the population recruitment are reported elsewhere (Sullivan et al., 2005Go). Briefly, women were enrolled between June 2000 and January 2004 from the general population and through families with a relative with fragile X syndrome. A total of 507 women aged 18–75 years completed a reproductive history questionnaire and gave buccal and/or blood samples. In the reproductive history questionnaire, women self-reported usual menstrual cycle and bleed length, and age at menarche. Women who were post-menopausal, pregnant, amenhorrhoeic, or who had hysterectomy, bilateral oophorectomy, or a history of radiation or chemotherapy were excluded from this study. Women with unclear cycling status, or who were taking HRT or fertility medications also were excluded. The final study sample consists of 164 women. The analysis of menstrual cycle length and bleed length does not include women using oral contraceptives (n=53). The study was approved by the Institutional Review Board at Emory University, Atlanta, GA.

Hormone analysis
Blood samples were drawn between day 2 and day 7 of the menstrual cycle. Ninety-three per cent of the samples were collected between day 2 and day 5. After collection, blood was stored on ice, and processing occurred within 3 h. Serum was stored at –80°C until the hormone assays were carried out. The Laboratory of Behavioural Neuroscience and Neuroimmunology at Emory University performed radioimmunoassays for E2, androstenedione (ASD) and dehydroepiandrosterone sulphate (DHEA-S) (Diagnostics Systems Laboratories Inc.). The ASD assay had standards ranging from 0.05 to 10 ng/ml. The DHEA-S assay had standards ranging from 25 to 8000 ng/ml. The E2 assay had standards ranging from 0.75–150 pg/ml (DSL 3rd generation E2 assay). Hormones were assayed in duplicate, and the duplicate values were averaged for our analysis. For our samples, the within-assay coefficient of variation (CV) was 3.2% for ASD, 1.9% for DHEA-S, and 3.8% for E2. Samples were run in three batches, and the between-assay CV ranged from 2.6 to 7.5% depending on the hormone measured and the level of the internal control. DHEA-S for one woman could not be obtained because of insufficient serum.

Genotyping analysis
DNA was extracted from buffy coats of peripheral blood samples. CYP17 genotyping was performed, as described in detail elsewhere (Carey et al., 1994Go). Briefly, a fragment containing the T27C substitution was amplified with the following primers: CYP-1 (forward), 5’-CATTCGCACTCTGGAGTC-3’ and CYP-2 (reverse), 5’-AGGCTCTTGGGGTACTTG-3’. PCR reactions were carried out in 22 µl aliquots containing ~100 ng of DNA, 0.5 µl of each primer, and 10xreaction buffer (containing 500 mmol/l Tris, 100 mmol/l KCl, 2 µl MgCl2, 1 IU of Taq polymerase, and 1.25 mmol/l deoxynucleotide triphosphates). An initial denaturation step was run at 94°C for 5 min, followed by 30 cycles of amplification with denaturation at 94°C for 1 min, annealing at 57°C for 1 min, and extension at 72°C for 1 min. This was followed by a final extension at 72°C for 5 min. The MspA1 restriction enzyme was used to digest the PCR fragments at 37°C for 2 h. Fragments were separated by gel electrophoresis and stained with ethidium bromide to identify the base pair substitution.

Approximately 5% of the samples were repeated for quality control of genotyping. The repeat analyses were run blind and the agreement of the repeated samples was 100%.

Statistical analysis
Unadjusted mean values and SE were calculated for the six outcomes of interest: E2, DHEA-S, ASD, cycle length, bleed length, and age at menarche. These results were stratified by CYP17 genotype.

Eighteen per cent of the women in our sample were related. To account for the dependency of related women, generalized estimating equations (GEE) (Zeger and Liang, 1986Go) were used to evaluate the relationship between CYP17 genotype and hormone levels or reproductive characteristics. We compared these results to models including only unrelated women. We coded CYP17 genotype ordinally based on the number of C alleles, and tested the statistical significance of the trend.

Prior to modelling, hormone levels were log-transformed to improve normality. Using GEE models, we estimated the adjusted geometric mean hormone levels and SE. Potential confounders, including age (continuous), BMI (kg/m2; continuous), race (white, African-American, Hispanic, Asian, other), current smoking (yes/no), parity (continuous), current oral contraceptive use (yes/no) and cycle day of blood collection (continuous) were chosen a priori based on our understanding of endocrinology and CYP17 biology. Age and race were included in all models. Other potential confounders were retained if their removal changed the parameter estimate for the association between CYP17 genotype and hormone concentration by >10%.

Because of differences in hormone synthesis between the ovaries and adipose tissue, we evaluated the interaction between BMI and CYP17. BMI was categorized as ≤25 or >25 kg/m2, based on the World Health Organization definition of overweight and obesity, and separate models were constructed for each BMI category. We also examined whether there was effect modification by race by including an interaction term between CYP17 genotype and race.

The self-report of reproductive characteristics showed strong digit preference. Therefore, we categorized these outcomes and modelled the probability of menstrual cycle length <27 days, bleed length <5 days, and menarche prior to 12 years. These thresholds reflect the lowest 25% of each distribution. Using GEE, we generated odds ratios (OR), representing the odds of being in the lowest 25% for the TC and CC genotype as compared to the reference group, the TT genotype, and 95% confidence intervals (CI). We compared these analyses with results from logistic regression models including only unrelated women.

The analysis of menstrual cycle length and bleed length was limited to women not using hormonal contraceptives. Age (continuous), BMI (kg/m2; continuous), race (white/other), current smoking (yes/no), and parity (continuous) were considered potential confounders. Racial groups of Hispanic, Asian, African-American, and other were combined into a single category because of small numbers. Age and race were included in the final models. Other potential confounders were retained if their removal changed the parameter estimate by >10%. The analysis of age at menarche only included age and race as confounders. BMI and smoking were not included as covariates since we did not have information on these variables at the time of menarche. All analyses were completed with SAS Version 9 (Carey, NC, USA).


    Results
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Acknowledgements
 References
 
Table I shows characteristics of the study population stratified by genotype. The study included 164 pre-menopausal women aged 18–52 years (mean 32). The mean BMI was 26 kg/m2. Fifty-three per cent of the women had BMI of ≤25 kg/m2. The women predominantly identified themselves as white (62.8%) and African-American (29.9%). Fifty-three women were current oral contraceptive users while 111 were not. The oral contraceptive users tended to be younger than the non-users (mean age 28.9 versus 33.6 years), but did not differ by race or BMI. Genotype distribution was 37% TT, 50% TC and 13% CC, and the allele frequencies were in Hardy–Weinberg equilibrium ({chi}2=0.351, df=2, P=0.82).


View this table:
[in this window]
[in a new window]
 
Table I. Characteristics of 164 pre-menopausal women according to cytochrome P450c17{alpha}(CYP17) genotype

 
Unadjusted analyses showed that the association between CYP17 genotype and hormone level did not differ between oral contraceptive users and non-users, so these results were combined. The unadjusted results in Table II show that women with the C allele had increased estradiol and DHEA-S, and decreased cycle length, bleed length, and age at menarche. Hormone levels in women were correlated: the correlation between E2 and DHEA-S was 0.23 (P=0.004); between E2 and ASD was 0.21 (P=0.008); and between DHEA-S and ASD was 0.57 (P < 0.0001).


View this table:
[in this window]
[in a new window]
 
Table II. Unadjusted meana of reproductive characteristics by CYP17 genotype

 
Table III shows geometric mean hormone level by CYP17 genotype and BMI adjusted for age, race, current smoking and oral contraceptive use. Among women with BMI ≤25, those with the TC and CC genotypes had 19 and 42% higher serum E2 respectively than women with the TT genotype (P for trend 0.007). Women with the TC and CC genotypes had 14 and 30% higher serum DHEA-S respectively than women with the TT genotype (P for trend 0.10). ASD levels did not differ between genotypes. Among women with BMI >25 kg/m2, hormone levels did not differ by genotype. These results did not change when the analysis was limited to women who were not using oral contraceptives. Further, adjusting for cycle day of serum collection and parity did not alter the findings, and there was no effect modification by race (data not shown).


View this table:
[in this window]
[in a new window]
 
Table III. Adjusted geometric meana hormone levels by CYP17 genotype and BMI

 
As shown in Table IV, among women who do not use oral contraceptives, those who have the C allele were more likely have cycle lengths <27 days controlling for age, race and BMI (OR for TC=2.36, 95% CI=1.24–4.52; OR for CC=5.59, 95% CI=1.53–20.43; P for trend=0.009). When this threshold was changed to <28 days, the association was similar (OR for TC=2.34, 95% CI=1.18–4.64; OR for CC=5.46 95% CI=1.38–21.56; P for trend=0.02, data not shown). These results were consistent with the decrease in mean cycle length among women with the C allele shown in the unadjusted results (Table II). Adjusting for current smoking and parity did not alter the findings. No association was seen between CYP17 genotype and bleed length or age at menarche.


View this table:
[in this window]
[in a new window]
 
Table IV. Adjusted odds ratio of reproductive characteristics by CYP17 genotype

 
The results of the models in Tables III and IV were similar whether the analysis accounted for the dependency of related individuals, assumed independence between women, or was restricted to unrelated individuals (data not shown).


    Discussion
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Acknowledgements
 References
 
Among pre-menopausal women with a low to moderate BMI, the CYP17 C allele was associated with serum E2 in a dose-dependent manner providing evidence that E2 may be up-regulated as a result of this polymorphism. E2 production is downstream of CYP17 activity. The production of DHEA and ASD occurs upstream and in closer proximity to CYP17 activity (Figure 1). Therefore, if CYP17 up-regulates E2, one may expect increases in DHEA-S (the reservoir for DHEA) and ASD as well. We found a non-significant trend for a dose-dependent increase in DHEA-S, but no significant increase in ASD among women with low to moderate BMI.

This is the first study to document the modifying effect of BMI on the relationship between CYP17 and serum hormone levels. Increases in E2 and DHEA-S associated with the C allele were limited to women with BMI ≤25 kg/m2. This suggests that the effect of the C allele may be limited to ovarian rather than adipose production of E2. There are at least two possible explanations for this. Differential tissue-specific regulation of CYP17 in the ovary and adipose tissue has been suggested by recent molecular investigations (Puche et al., 2002Go). Alternatively, the C allele may up-regulate steroid production in all women, but the relative proportion of end-products may vary among those with BMI >25 kg/m2 due to differences downstream in steroid biosynthesis. Corbould et al. (2002)Go found that the ratio of 17{beta}-hydroxysteroid dehydrogenase3 mRNA to CYP19 mRNA in adipose tissue was positively correlated with obesity. This suggests that steroid production in adipose tissue of obese women may be turned towards the production of testosterone rather than estrone and E2. This may also explain the result that low BMI was associated with higher E2 levels, which is consistent with some previous studies of pre-menopausal women (Manson et al., 2001Go; Baghaei et al., 2002Go; Randolph et al., 2003Go).

The relationship between CYP17 and E2 was first documented among young (18–33 years), nulliparous women with low BMI (mean=22.8 kg/m2, interquartile range 20.1–24.5) (Feigelson et al., 1998Go). Despite differences in the age and parity of our populations, the percentage increases in E2 in our early follicular phase samples (19 and 42% for TC and CC as compared to TT, respectively, among BMI ≤25 kg/m2) are similar to those seen on day 11 by Feigelson et al. (11 and 57%). Two other studies of pre-menopausal women found no relationship between the CYP17 genotype and serum E2 (Garcia-Closas et al., 2002Go; Travis et al., 2004Go). However, serum was not collected within a limited range of cycle days. Rather, the analysis was stratified by cycle phase (without a biological measure of ovulation) and adjusted by self-reported cycle day. This may have resulted in misclassification and an attenuated effect. Furthermore, neither study assessed the potential effect of BMI.

Five studies of post-menopausal women did not find an association between CYP17 genotype and E2 (Haiman et al., 2001Go; Berstein et al., 2002Go; Dunning et al., 2004Go; Travis et al., 2004Go; Tworoger et al., 2004Go). The only study to restrict their analysis to postmenopausal women with BMI<25 found the CC genotype associated with higher urinary estrone and estradiol (Onland-Moret et al., 2005). Furthermore, both the levels of endogenous estrogens and the mechanisms for estrogen production differ between pre- and post-menopausal women, possibly explaining the discrepant results between studies (Speroff et al., 1999Go).

Our analyses of reproductive characteristics provides indirect evidence that the T27C polymorphism regulates endocrine function. Women with the C allele were more likely to have cycle lengths <27 days. Higher estrogen metabolite levels have been associated with shorter cycles (Windham et al., 2002Go) and shorter follicular phase length (Harlow et al., 2000Go), suggesting a mechanism for this association. Our study participants self-reported menstrual cycle length. Because error in recall is likely distributed evenly across genotypes, error would result in bias toward the null.

Given that our population was drawn from a reproductive study of FMR1 repeat size, it may not reflect characteristics of the general population. However, controlling for FMR1 repeat size in our analyses or limiting our analyses to women with FMR1 repeat <60 (the more common repeat sizes in the general population) did not alter our findings.

This is the first study to examine the relationship between CYP17 and E2 using serum sampled exclusively from the early follicular phase of the menstrual cycle. E2 levels vary greatly during the menstrual cycle, and such variation is closely tied to the timing of ovulation, and follicular and luteal phase length. Determining the day of ovulation requires a labour-intensive protocol of collecting biological samples throughout the majority of the menstrual cycle. In the absence of these data, we suggest that a more consistent measure of E2 levels across women can be obtained during the early follicular phase. Michaud et al. (1999)Go found that early follicular phase serum E2 had better reproducibility across cycles than luteal phase E2. Although Ahmad et al. (2002)Go found that E2 levels from day 9 to day 11 were more consistent, their population was restricted to women with regular cycles. Furthermore, Ahmad et al. (2002)Go found that early follicular phase serum E2 had the highest correlation to average E2 across the entire menstrual cycle. Thus, the measurement of serum in the early follicular phase is the most consistent among women with varying cycle lengths.

Our findings, in concert with the earlier results of Feigelson et al. (1998)Go, highlight the need to examine the relationship between CYP17 and the entire downstream steroid hormone pathway in a large population-based study of young women. The optimal study design would include assessment of hormone levels early in the follicular phase with highly sensitive and specific assays and would carefully control for BMI and central adiposity. Genes involved in steroid biosynthesis and metabolism may play important roles in disease risk directly through influencing hormone levels and indirectly through altering other reproductive endpoints. Polymorphisms in these genes, such as T27C, may help to determine women at increased risk.


    Acknowledgements
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Acknowledgements
 References
 
We would like to thank Dr Celia Dominguez for her review of the endocrinology and Milburn Emery for his technical assistance. We would especially like to thank the volunteers and their families whose participation made this work possible. This work was supported by the National Institutes of Health grants, HD40637 and HD29909, and MO-1-RR00039 (General Clinical Research Center), and an Achievement Rewards for College Scientists (ARCS) Foundation Fellowship.


    References
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Acknowledgements
 References
 
Ahmad N, Pollard T and Unwin N (2002) The optimal timing of blood collection during the menstrual cycle for the assessment of endogenous sex hormones: can interindividual differences in levels over the whole cycle be assessed on a single day? Cancer Epidemiol Biomarkers Prev 11, 147–151.[Abstract/Free Full Text]

Amant F, Dorfling CM, de Brabanter J, Vandewalle J, Vergote I, Lindeque BG and van Rensburg EJ (2004) A possible role of the cytochrome P450c17alpha gene (CYP17) polymorphism in the pathobiology of uterine leiomyomas from black South African women: a pilot study. Acta Obstet Gynecol Scand 83, 234–239.[CrossRef][ISI][Medline]

Baghaei F, Rosmond R, Westberg L, Hellstrand M, Landen M, Eriksson E, Holm G and Bjorntorp P (2002) The lean woman. Obes Res 10, 115–121. [published erratum appears in Obes Res 10(3) (March 2002), following table of contents].[Abstract/Free Full Text]

Berstein LM, Imyanitov EN, Gamajunova VB, Kovalevskij AJ, Kuligina E, Belogubova EV, Buslov KG, Karpova MB, Togo AV, Volkov ON et al. (2002) CYP17 genetic polymorphism in endometrial cancer: are only steroids involved? Cancer Lett 180, 47–53.[CrossRef][ISI][Medline]

Carey AH, Waterworth D, Patel K, White D, Little J, Novelli P, Franks S and Williamson R (1994) Polycystic ovaries and premature male pattern baldness are associated with one allele of the steroid metabolism gene CYP17. Hum Mol Genet 3, 1873–1876.[Abstract]

Corbould AM, Bawden MJ, Lavranos TC, Rodgers RJ and Judd SJ (2002) The effect of obesity on the ratio of type 3 17beta-hydroxysteroid dehydrogenase mRNA to cytochrome P450 aromatase mRNA in subcutaneous abdominal and intra-abdominal adipose tissue of women. Int J Obes 26, 165–175.[CrossRef][ISI]

Costa-Santos M, Kater CE and Auchus RJ (2004) Two prevalent CYP17 mutations and genotype-phenotype correlations in 24 Brazilian patients with 17-hydroxylase deficiency. J Clin Endocrinol Metab 89, 49–60.[Abstract/Free Full Text]

Deligdisch L (2000) Hormonal pathology of the endometrium. Mod Pathol 13, 285–294.[CrossRef][ISI][Medline]

Dunning AM, Dowsett M, Healey CS, Tee L, Luben RN, Folkerd E, Novik KL, Kelemen L, Ogata S, Pharoah PD et al. (2004) Polymorphisms associated with circulating sex hormone levels in postmenopausal women. J Natl Cancer Inst 96, 936–945.[Abstract/Free Full Text]

Feigelson HS, Shames LS, Pike MC, Coetzee GA, Stanczyk FZ and Henderson BE (1998) Cytochrome P450c17alpha gene (CYP17) polymorphism is associated with serum estrogen and progesterone concentrations. Cancer Res 58, 585–587.[Abstract]

Garcia-Closas M, Herbstman J, Schiffman M, Glass A and Dorgan JF (2002) Relationship between serum hormone concentrations, reproductive history, alcohol consumption and genetic polymorphisms in pre-menopausal women. Int J Cancer 102, 172–178.[CrossRef][ISI][Medline]

Haiman CA, Hankinson SE, Colditz GA, Hunter DJ and De Vivo I (2001) A polymorphism in CYP17 and endometrial cancer risk. Cancer Res 61, 3955–3960.[Abstract/Free Full Text]

Hanukoglu I (1992) Steroidogenic enzymes—structure, function, and role in regulation of steroid-hormone biosynthesis. J Steroid Biochem Mol Biol 43, 779–804.[CrossRef][ISI]

Harlow SD, Baird DD, Weinberg CR and Wilcox AJ (2000) Urinary oestrogen patterns in long follicular phases. Hum Reprod 15, 11–16.[Free Full Text]

Li TC, Tuckerman EM and Laird SM (2002) Endometrial factors in recurrent miscarriage. Hum Reprod Update 8, 43–52.[Abstract/Free Full Text]

Manson JM, Sammel MD, Freeman EW and Grisso JA (2001) Racial differences in sex hormone levels in women approaching the transition to menopause. Fertil Steril 75, 297–304.[CrossRef][ISI][Medline]

Martin RM, Lin CJ, Costa EM, de Oliveira ML, Carrilho A, Villar H, Longui CA and Mendonca BB (2003) P450c17 deficiency in Brazilian patients: biochemical diagnosis through progesterone levels confirmed by CYP17 genotyping. J Clin Endocrinol Metab 88, 5739–5746.[Abstract/Free Full Text]

Michaud DS, Manson JE, Spiegelman D, Barbieri RL, Sepkovic DW, Bradlow HL and Hankinson SE (1999) Reproducibility of plasma and urinary sex hormone levels in premenopausal women over a one-year period. Cancer Epidemiol Biomarkers Prev 8, 1059–1064.[Abstract/Free Full Text]

Nedelcheva Kristensen V, Haraldsen EK, Anderson KB, Lonning PE, Erikstein B, Karesen R, Gabrielsen OS and Borresen-Dale AL (1999) CYP17 and breast cancer risk: the polymorphism in the 5' flanking area of the gene does not influence binding to Sp-1. Cancer Res 59, 2825–2828.[Abstract/Free Full Text]

Onland-Moret NC, Van Gils CH, Roest M, Grobbee DE and Peeters PHM (2005) CYP17, Urinary sex steroid levels and breast cancer risk in postmenopausal women. Cancer Epidemiol Biomarkers Prev 14, 815–820.[Abstract/Free Full Text]

Peck JD, Hulka BS, Savitz DA, Baird D, Poole C and Richardson BE (2003) Accuracy of fetal growth indicators as surrogate measures of steroid hormone levels during pregnancy. Am J Epidemiol 157, 258–266.[Abstract/Free Full Text]

Puche C, Jose M, Cabero A and Meseguer A (2002) Expression and enzymatic activity of the P450c17 gene in human adipose tissue. Eur J Endocrinol 146, 223–229.[Abstract/Free Full Text]

Randolph JF Jr, Sowers M, Gold EB, Mohr BA, Luborsky J, Santoro N, McConnell DS, Finkelstein JS, Korenman SG, Matthews KA et al. (2003) Reproductive hormones in the early menopausal transition: relationship to ethnicity, body size, and menopausal status. J Clin Endocrinol Metab 88, 1516–1522.[Abstract/Free Full Text]

Sasano H, Okamoto M, Mason JI, Simpson ER, Mendelson CR, Sasano N and Silverberg SG (1989) Immunolocalization of aromatase, 17 alpha-hydroxylase and side-chain-cleavage cytochromes P-450 in the human ovary. J Reprod Fertil 85, 163–169.[ISI][Medline]

Sata F, Yamada H, Yamada A, Kato EH, Kataoka S, Saijo Y, Kondo T, Tamaki J, Minakami H and Kishi R (2003) A polymorphism in the CYP17 gene relates to the risk of recurrent pregnancy loss. Mol Hum Reprod 9, 725–728.[Abstract/Free Full Text]

Sharp L, Cardy AH, Cotton SC and Little J (2004) CYP17 gene polymorphisms: prevalence and associations with hormone levels and related factors. a HuGE review. Am J Epidemiol 160, 729–740.[Abstract/Free Full Text]

Speroff F, Glass RH and Kase NG (1999) Clinical Gynecologic Endocrinology and Infertility. Lippincott-Williams &Wilkins, Philidelphia, PA.

Sullivan AK, Marcus M, Epstein MP, Allen EG, Anido AE, Paquin JJ, Yadav-Shah M and Sherman SL (2005) Association of FMR1 repeat size with ovarian dysfunction. Hum Reprod 20, 402–412.[Abstract/Free Full Text]

Travis RC, Churchman M, Edwards SA, Smith G, Verkasalo PK, Wolf CR, Wolf H and Key TJ (2004) No association of polymorphisms in CYP17, CYP19, and HSD17-B1 with plasma estradiol concentrations in 1,090 British women. Cancer Epidemiol Biomarkers Prev 13, 2282–2284.[Free Full Text]

Tworoger SS, Chubak J, Aiello EJ, Ulrich CM, Atkinson C, Potter JD, Yasui Y, Stapleton PL, Lampe JW, Farin FM et al. (2004) Association of CYP17, CYP19, CYP1B1, and COMT polymorphisms with serum and urinary sex hormone concentrations in postmenopausal women. Cancer Epidemiol Biomarkers Prev 13, 94–101.[Abstract/Free Full Text]

Windham GC, Elkin E, Fenster L, Waller K, Anderson M, Mitchell PR, Lasley B and Swan SH (2002) Ovarian hormones in premenopausal women: variation by demographic, reproductive and menstrual cycle characteristics. Epidemiology 13, 675–684.[CrossRef][ISI][Medline]

Yamada H, Sata F, Kato EH, Saijo Y, Kataoka S, Morikawa M, Shimada S, Yamada T, Kishi R and Minakami H (2004) A polymorphism in the CYP17 gene and intrauterine fetal growth restriction. Mol Hum Reprod 10, 49–53.[Abstract/Free Full Text]

Zeger SL and Liang KY (1986) Longitudinal data analysis for discrete and continuous outcomes. Biometrics 42, 121–130.[ISI][Medline]

Submitted on January 25, 2005; resubmitted on April 4, 2005; accepted on April 7, 2005.