Centers for Disease Control and Prevention
Centers for Disease Control and Prevention
Centers for Disease Control and Prevention CDC Home Search CDC CDC Health Topics A-Z    
Office of Genomics and Disease Prevention  
Office of Genomics and Disease Prevention

This HuGE Review was published in the American Journal of Epidemiology 2004; 160(8):729-740
PubMed ID: 15466495

CYP17
Gene Polymorphisms: Prevalence and Associations with Hormone Levels and Related Factors. A HuGE Review

L. Sharp, AH Cardy, SC Cotton and J. Little

From the Epidemiology Group, Department of Medicine and Therapeutics
University of Aberdeen
Aberdeen, United Kingdom

Received for publication: January 20, 2004
Accepted for publication:
May 4, 2004


Download print version




 ABSTRACT

The cytochrome P-450c17alpha (CYP17) gene, located on chromosome 10q24.3, encodes the enzyme cytochrome P-450c17alpha, which functions at key branch points in steroid hormone biosynthesis. Three polymorphisms have been described, but only the single base-pair change in the 5'-untranslated region (5'-UTR) has been investigated to any great extent. In single studies, the variant was associated with reduced messenger RNA level in ovarian cells but not with messenger RNA level in breast tissue. Homozygosity for the 5'-UTR variant is most common in East Asian (32%) and Japanese (22%) populations and is less common among White (mainly European and North American (14%)) and Black (mainly African-American (13%)) populations, but selection biases are likely to have affected these frequency estimates. Genotype appears to influence circulating estrogen levels in premenopausal women, while studies of relations with hormone levels in men have produced inconclusive results. However, relatively few studies have been conducted. Seven of 11 retrospective studies suggested a modest association between genotype and age at menarche. Random error in recall of age at menarche is likely to have attenuated this relation. Associations between genotype and postmenopausal estrogen use and bone mass have been observed in single studies. Further investigation of relations between genotype and hormone levels, exogenous hormone use, and markers of hormonal status may advance understanding of hormonally mediated diseases.

Key Words: CYP17 ; epidemiology; genetics; gonadal steroid hormones; hormones; menarche; polymorphism (genetics); steroid 17-alpha-hydroxylase

Abbreviations: CI, confidence interval; CYP17, cytochrome P-450c17alpha; HRT, hormone replacement therapy; mRNA, messenger RNA; 5'-UTR, 5'-untranslated region.

 
 GENES

The cytochrome P-450c17alpha (CYP17) gene, located on chromosome 10q24.3, encodes the enzyme cytochrome P-450c17alpha, which functions at key branch points in steroid hormone biosynthesis in the adrenal gland, ovary, and gonads (1,2). Specifically, cytochrome P-450c17alpha mediates both steroid 17alpha-hydroxylase activity, which coverts pregnenolone to dehydroepiandrosterone, and 17,20-lyase activity, which generates androstenedione from progesterone. These androgens may then be converted to estrone, testosterone, and estradiol (3) (Figure 1). The discovery that the CYP17 gene is polymorphic prompted investigation of the role of gene variants in the etiology of diseases and conditions in which estrogens or androgens play an important role, notably breast cancer, polycystic ovary syndrome, endometrial cancer, and prostate cancer. Unraveling whether genotype influences steroid hormone metabolism and/or is related to markers of endogenous hormonal state would aid in the understanding of hormonally mediated conditions. In this paper, we review and synthesize the findings of studies on the prevalence and effects of polymorphisms in the CYP17 gene.

 Figure 1
 Pathway of steroid hormone synthesis and metabolism, including the role  of the cytochrome P-450c17alpha (CYP17) gene.
 
 GENE VARIANTS

There are numerous mutations in the CYP17 gene, the majority of which are extremely rare (4,5). Three common polymorphisms have been described (6-8), but only one, in the 5'-untranslated region (5'-UTR), has been extensively investigated. This polymorphism involves a single base-pair change (T-->C) in the promoter region, 34 base pairs upstream from the initiation of translation and downstream from the transcription start site (6). The variant creates a recognition site for the MspAI restriction enzyme. In line with the literature, we will refer to the common allele as A1 and the variant allele, –34T-->C, as A2. Thus, we will refer to homozygosity for the common allele as "A1A1" rather than use the base-pair designate CC, heterozygosity as "A1A2" (rather than CT), and homozygosity for the variant allele as "A2A2" (rather than TT); this nomenclature was used in all of the literature reviewed.

The A2 variant is thought to create an additional Sp-1-type (CCACC box) promoter site. Because the number of promoter elements may correlate with promoter activity (9), it has been postulated that the variant might result in increased transcription (6,10). However, in the only molecular transcription study carried out to date, an additional binding site created by the polymorphism did not seem to be utilized (11).

Miyoshi et al. (12) studied messenger RNA (mRNA) expression in tumor tissue from 67 women with breast cancer and in normal breast tissue from another 51 women who also had breast cancer. While the mean CYP17 mRNA level was significantly higher in the tumor samples than in the normal breast tissue, there were no significant differences in mRNA levels between carriers and noncarriers of the variant allele in either the normal breast tissue group or the tumor tissue group. However, intratumoral estradiol levels were significantly higher in carriers than in non-carriers. Daneshmand et al. (13) measured CYP17 mRNA expression in ovarian thecal cells (connective tissue cells from the Graafian follicle) obtained from 51 premenopausal women undergoing hysterectomy and bilateral oophorectomy for reasons unrelated to ovarian disorders. Mean mRNA levels decreased with the number of variant alleles; compared with levels in cells from persons homozygous for the common A1 allele, levels in heterozygotes were approximately halved, and levels in A2 homozygotes were approximately half those observed in heterozygotes (13). These functional effects require confirmation.

Two further polymorphisms in CYP17 have been reported. Miyoshi et al. (8) identified a G-->A polymorphism at nucleotide 1951 in the promoter region. Crocitto et al. (7) reported that intron 6 contains a C-->A transition at nucleotide 5471. Information is lacking on the functional effects of these polymorphisms, if any, and on relations with the 5'-UTR polymorphism.

 

back to top
 
 POPULATION FREQUENCIES

We searched MEDLINE and EMBASE for papers published from 1990 to June 2003, using the Medical Subject Headings "cytochrome P-450" and "steroid 17 alpha-monooxygenase" and the text words "CYP17" and "P-450c17." We identified further relevant studies by hand-searching reference lists in published articles. Once studies were identified, they were reviewed, and those including at least 50 subjects without cancer or other diseases were included in our analyses. Several studies were studies of men diagnosed as having benign prostatic hyperplasia or prostate enlargement or men who had lower urinary tract symptoms consistent with this diagnosis; these studies were excluded, since the results might not be generalizable. Studies of prostatic tissue obtained from men undergoing radical prostatectomy were excluded for the same reason. From the remaining studies, we extracted data on numbers and sources of subjects and numbers of subjects with each genotype. We calculated frequencies and associated 95 percent confidence intervals among heterozygotes and homozygotes for all of the studies identified. We assessed Hardy-Weinberg equilibrium of genotype frequencies for individual studies using the Pearson chi2 test. A significance level of p < 0.05 was used to reject the null hypothesis that the genotype frequencies were in equilibrium.

For ethnic groups for which there were several studies, we pooled the data to compute a more precise estimate of the homozygous variant frequency. Articles in which the ethnicity of the study subjects was described as "mixed" were excluded from this analysis. Data from African-American series were combined with data on African Blacks; data from Japanese-American series were combined with data from Japan; and data from Asian-American series were combined with data from studies conducted elsewhere in Asia (i.e., Singapore and Taiwan). Series were considered White if the sample was described as "predominantly White" or included at least 85 percent White subjects.

5'-UTR polymorphism
Web table 1 shows findings from 50 reports containing data on the frequency of heterozygous and homozygous variants (8, 10, 11, 13–59). (This information is described in the first of two supplementary tables; each is referred to as "Web table" in the text and is posted on the website of the Human Genome Epidemiology Network, as well as on the Journal’s website.) Some of the series in the table overlap, but the extent of this is not clear. Several of the studies suffered from one or more of the following limitations: 1) the study had a small sample size; 2) there was a lack of information on the source of the subjects and/or the criteria used to select subjects; and 3) the study was not population based, comprising instead hospital patients, persons who had undergone medical investigations, or volunteers, without a clearly defined sampling frame. This makes it difficult to determine whether apparent geographic or ethnic variation reflects true differences or simply selection or participation bias. In addition, these issues may hinder generalizability.

Figure 2 shows the pooled frequency of A2 homozygosity across ethnic groups. A2 homozygosity is most common among persons of East Asian origin. When we pooled data from three studies from Taiwan (21-23), one study from Singapore (20), and one study of persons of Asian or Indian-Pakistani origin living in Canada (36), the frequency of A2 homozygosity was 32 percent (95 percent confidence interval (CI): 29.8, 35.1). In the Japanese series, which included six studies from Japan (8,15-19) and a study of Japanese Americans (44), A2 homozygotes comprised 22 percent of the subjects (95 percent CI: 20.4, 24.1).

  Figure 2
  Pooled frequency of the cytochrome P-450c17alpha (CYP17) A2A2 genotype,   by ethnic group. Bars, 95% confidence interval. Studies included in the   pooled estimate for each ethnic group are as follows—Black: Kittles et al.   (14), Lunn et al. (22), Lai et al. (36), Feigelson et al. (44), and Weston   et al. (48); White: Kristensen et al. (11), Kittles et al. (14), Lunn et al.   (22), Mitrunen et al. (23), Diamanti-Kandarakis et al. (25),   Gudmundsdottir et al. (26), Marszalek et al. (27), Kuligina et al. (28),   Bergman-Jungeström et al. (29), Wadelius et al. (30), Allen et al. (31),   Techatraisak et al. (32), Dunning et al. (33), Young et al. (34), Lai et al.   (36), Haiman et al. (38), Haiman et al. (39), Chang et al. (40), Feigelson   et al. (41), Feigelson et al. (44), Garner et al. (47), Weston et al. (48),     McCann et al. (50), Ambrosone et al. (51), García-Closas et al. (52),   Zmuda et al. (53), Stanford et al. (54), Spurdle et al. (56), and Cui et al.   (57); Latino: Feigelson et al. (44) and Weston et al. (48); Japanese:   Miyoshi et al. (8), Huang et al. (15), Habuchi et al. (16), Kado et al. (17),   Hamajima et al. (18), Gorai et al. (19), and Feigelson et al. (44); Other   East Asia: Wu et al. (20), Huang et al. (21), Lunn et al. (22), Yu et al.   (23), and Lai et al. (36).

The pooled frequency of A2 homozygosity for White populations from North America, Europe, and Australia was considerably lower than the frequency for Asians: 14 percent (95 percent CI: 13.5, 15.0). Within North America, the prevalence in White populations was 11–19 percent. In the studies from Australia, the frequencies were 11 percent and 15 percent (55-57). Within Europe, there was no evidence of a strong geographic distribution, but most of the data were from Nordic countries, Iceland, and the United Kingdom, where the frequency ranged from 7 percent to 17 percent. Few data from other parts of Europe were available. In St. Petersburg, Russia, the frequency was 23–25 percent (28). In a small series from Greece, no A2 homozygotes were observed (25), but the genotype frequencies were not in Hardy-Weinberg equilibrium.

Data for other ethnic groups and geographic areas were limited. In two studies of Hispanic-American women, 18 percent (95 percent CI: 13.6, 14.8) were A2 homozygotes (44,48). In African Americans, the frequency of A2A2 homozygosity was 6–16 percent. In the single study from Africa (14), 9 percent of Nigerian men were A2A2 homozygous. The pooled estimate for all Black populations was 13 percent (95 percent CI: 10.8, 16.1).

Other polymorphisms
From one study (7), the frequency of the variant allele of the intron 6 polymorphism was estimated to be 21 percent among Asians and 5 percent among African Americans. With regard to the polymorphism in the promoter region, the variant A allele was present in 5 percent (10 of 184) of volunteers in Osaka, Japan (8).

 
  ASSOCIATIONS BETWEEN THE 5'-UTR POLYMORPHISM AND HORMONE LEVELS OR MARKERS OF ENDOGENOUS HORMONAL STATE

Genetic variation in CYP17 has been postulated to cause differences in circulating hormone levels. This is a possible mechanism by which the gene might influence risk of disease. We reviewed studies of the association between CYP17 genotype and circulating hormone levels, markers of hormonal status (e.g., age at menarche), or phenotypes associated with an abnormal hormonal milieu (e.g., male pattern baldness). We identified these studies using the search strategy described above, but because there were relatively few such studies, no restriction by number of subjects was applied.

Studies in females
Hormone concentrations
The relation between the 5'-UTR polymorphism and endogenous hormone concentrations has been investigated in six studies of disease-free women (13, 25, 35, 39, 41, 52), all but one of them studies of premenopausal women (Table 1). The hormones measured differed between studies. Interpretation of the findings of these studies is complicated by methodological difficulties in adequately characterizing hormones that vary in level over the menstrual cycle. A single measurement taken at any time during the cycle appears to be reasonably reliable for the determination of levels of dehydroepiandrosterone sulfate, sex hormone-binding globulin, and androstenedione (60, 61). For estradiol and progesterone, and possibly for free androgens, a single measurement is less informative but appears to be most consistent if taken on specific days of the cycle (estradiol: days 9–11; progesterone: days 17–21; free androgen: days 12–15) (61).

 Table 1
 Findings from studies of the relation between polymorphisms in the
 5'-untranslated region of the cytochrome P-450c17alpha (CYP17) gene and  hormone levels in women

A priori, it might be expected that the A2 allele would be associated with higher estrogen levels via its postulated effect on promoter activity and transcription. The first study in this area was consistent with this hypothesis (41). Feigelson et al. (41) recruited 83 Latina, African-American, and Asian volunteers aged 18–33 years who had never had a pregnancy lasting 12 weeks or more, were not obese, had not taken oral contraceptives or hormonal medications in the previous 3 months, and whose current menstrual cycle was ovulatory and expected to be 25–32 days in length. Mean serum estradiol levels on day 11 were 11 percent higher in heterozygotes and 57 percent higher in A2 homozygotes than in A1 homozygotes (p for trend = 0.04). Day 22 estradiol concentrations were also elevated in A2 carriers, but to a lesser extent (heterozygotes: +7 percent; homozygotes: +28 percent; p for trend = 0.06). Day 22 progesterone levels were 24 percent and 30 percent higher in heterozygotes and A2 homozygotes, respectively, than in A1A1 homozygous women (p for trend = 0.04). Similarly, among 256 premenopausal women of mixed ethnicity, Daneshmand et al. (13) observed 72 percent higher mean estradiol levels in heterozygotes than in A1 homozygotes. Furthermore, in a study of urine levels of (biologically inactive) 2-hydroxyestrone and (biologically active) 16alpha-hydroxyestrone among 494 nulliparous Canadian women aged 17–35 years, the 2-hydroxyestrone:16alpha-hydroxyestrone ratio was significantly lower in A2 homozygotes than in women with other genotypes; this finding was adjusted for oral contraceptive use, body mass index, smoking status, and ethnic group (35).

In contrast, García-Closas et al. (52), who measured estradiol and progesterone in premenopausal women according to phase of the menstrual cycle (early follicular: 0–9 days from the onset of last menses (n = 52); late follicular: 10–14 days from onset (n = 54); luteal: >15 days from onset (n = 112)), found that differences in estradiol and progesterone levels by genotype varied according to cycle phase. However, none of the differences were statistically significant.

Theoretically, CYP17 genotype could affect androgen levels either directly, because of the enzyme’s role in the generation of dehydroepiandrosterone and androstenedione, or indirectly, through the enzyme’s effects on estrogen levels and the link between androstenedione, testosterone, estrone, and estradiol (Figure 1). With regard to premenopausal women, Daneshmand et al. (13) found that androstenedione concentrations were 40 percent higher in heterozygotes than in A1 homozygotes but testosterone levels did not differ. In a small study carried out in Greece, the mean level of serum total testosterone was 15 percent higher in heterozygotes than in A1A1 homozygotes (25). García-Closas et al. (52) found testosterone to be 4 percent higher in heterozygotes and 12 percent higher in A2 homozygotes than in A1 homozygotes, but this finding was not statistically significant. In that study, androstenedione and dehydroepiandrosterone levels were not associated with genotype.

Haiman et al. (39) measured serum hormone fractions in 462 postmenopausal women from the Nurses’ Health Study who were without cancer and were not taking postmenopausal hormones. Compared with women with the A1A1 genotype, A2A2 subjects had slightly higher levels of estrone (+11 percent; p = 0.05) and estrone sulfate (+8 percent; p = 0.17). Estradiol levels were slightly lower (–12 percent; p = 0.13). Concentrations of testosterone, androstenedione, dehydroepiandrosterone, and dehydroepiandrosterone sulfate varied little between A2 and A1 homozygotes. These results differed from those previously published for a subgroup of the subjects (38).

Age at menarche, onset of puberty, age at menopause, and years of menstruation
Age at menarche is at least partly hormonally mediated (62). Eleven studies have investigated associations between the 5'-UTR polymorphism and age at menarche in women without cancer or other diseases (10, 18-20, 33, 36, 38, 43, 46, 48, 51). All of these studies involved asking adult women to recall their age at menarche, an approach that is likely to result in random error; bias seems unlikely. In a small longitudinal study of women in their twenties in whom menarche and age had been prospectively recorded, the misclassification rate was 25 percent when the women were grouped into three categories based on recalled age at menarche (<12.0, 12.0–12.9, or >13.0 years) (63). These results may provide some indication of the possible extent of misclassification in the studies of CYP17 and age at menarche; the effect of such misclassification would be to reduce observed differences between genotypes.

Overall, seven of the 11 studies were compatible with a modest association between A2A2 genotype and earlier menarche. Five of the 11 studies summarized mean age at menarche according to genotype (18, 19, 33, 36, 48). Four of these observed a slightly lower age at menarche in A2A2 women than in A1A1 women, but the differences were small and not statistically significant (18, 19, 36, 48). The remaining six studies presented data on proportions of women who had experienced menarche above or below a particular age; three were consistent with a lower age at menarche among A2 carriers (10, 20, 51) and three were not (38, 43, 46).

Kadlubar et al. (49) assessed associations between breast development, a marker of puberty onset, and genotype in 137 US girls with a mean age of 9.5 years. Consistent with the observations related to age at menarche, a greater proportion of girls with the A2A2 genotype had experienced the onset of puberty (66 percent) in comparison with girls with other genotypes (51–52 percent), although this finding was based on small numbers (18 girls with the A2A2 genotype) and the association was not statistically significant.

Gorai et al. (19) observed that mean age at natural menopause did not differ significantly by genotype among 333 postmenopausal Japanese women. Heterozygous women had a slightly greater average number of years of menstruation than either group of homozygotes (p = 0.090) because they had a younger mean age at menarche. Ambrosone et al. (51) observed that 62 percent of A1A1 women had experienced menopause at age 48 years or younger, as compared with 53 percent of women with the A2 allele. Feigelson et al. (10) reported that age at menopause was not associated with CYP17 genotype but did not show their results.

Use of postmenopausal hormones
Feigelson et al. (44) investigated 749 postmenopausal women with no current or prior history of breast, ovarian, endometrial, or cervical cancer from a large multiethnic cohort established using driver’s license files in Hawaii and Los Angeles, California, and found that A2A2 women were half as likely as A1A1 women to be current users of hormone replacement therapy (HRT) (odds ratio = 0.52, 95 percent CI: 0.31, 0.86). This intriguing association was present in all four ethnic groups studied. In four smaller studies carried out in the United States, where postmenopausal hormone use has been relatively high (64,65), data were available on frequencies of HRT use by genotype among women who served as control subjects for cancer cases (39,46,51,66). Three of these control series were nested within cohort studies (39, 46, 66) and the other was population based (51). In a study conducted by Helzlsouer et al. (46) among 115 women aged 60.2 years (standard deviation, 11.5) who were primarily of European origin, only 14 percent of A2A2 homozygous women had ever used HRT, as reported retrospectively by self-completed questionnaire, as compared with 31 percent of heterozygotes and 34 percent of women with the A1A1 genotype. This is consistent with the observation of Feigelson et al. (44). In contrast, in an analysis of 554 predominantly White women from the Nurses’ Health Study, there were no differences by genotype in the frequencies of never, past, and current use of HRT, as reported on questionnaires returned by subjects at 2-year intervals during the period 1976–1996 (39). Ambrosone et al. (51) also found no difference in frequencies of HRT use (assessed through retrospective interview) by genotype among 188 women of European origin, but they did not distinguish between A1A2 and A2A2 subjects. Using data nested in the same cohort as that of Feigelson et al. (44) but from a different random sample, comprising 391 postmenopausal women with no history of endometrial, breast, or ovarian cancer, McKean-Cowdin et al. (66) found that a slightly higher proportion of A2 carriers (25 percent) had used estrogen replacement therapy without progestins at baseline (at ages 45–75 years) than A1A1 women (19 percent).

A possible concern is publication bias in the reporting of data on the relation between the CYP17 polymorphism and HRT use. The study that had findings consistent with the report of Feigelson et al. (44) was substantially smaller than the Nurses’ Health Study, in which no relation was found. Investigation of this association was not the purpose of any of these studies and was not commented on by the authors.

The ways in which CYP17 might influence HRT use are not clear, but it is possible that they could be related to differences in levels of circulating hormones between women with different genotypes. For example, one possibility is that A2A2 women may have higher estrogen levels prior to menopause and thus may suffer less from menopausal symptoms than women with other genotypes, who initially have lower estrogen levels.

Breast density
Because breast density has been shown to be a strong independent predictor of breast cancer risk, Haiman et al. (67) investigated whether the extent of breast density measured by mammogram was related to genotype. A total of 152 African-American women and 244 White women who had been diagnosed with breast cancer, and for whom information on mammographic density in the contralateral (cancer-free) breast was available, were evaluated. Mean percentage breast density (percentage of the breast comprising mammographically dense tissue) was associated with age, menopausal status/HRT use, body mass index, parity, age at first birth, and age at menarche. Among African-American women, there was no difference in breast density (adjusted for the other predictors of density) according to CYP17 genotype. In White women, those who were A2A2 homozygous had higher density than women with other genotypes, but this finding was not statistically significant. It is not clear whether these results would be generalizable to the population of women at risk of breast cancer (as opposed to those with the disease).

Studies in males
Hormone concentrations
Four studies (31, 37, 42, 53) have evaluated the relation between genotype and hormone concentrations in nondiseased men (Table 2). Regarding the accuracy of the hormone levels assessed in these studies, a single measurement of testosterone and the metabolite dihydrotestosterone is thought to reliably reflect mean levels over a period of 1 year (68), whereas little is known about the reliability of a single measurement of the other hormones investigated.

  Table 2
  Findings from studies of the relation between polymorphisms in the  
  5'-  untranslated region of the cytochrome P-450c17alpha (CYP17) gene and   hormone levels in men

No association between genotype and mean total tes-tosterone level has been observed (31, 37, 53). In one study of White males in the United States (53), median levels of bioavailable testosterone were significantly higher in A2 homozygotes than in A1 homozygotes, but this pattern was not observed among vegetarians in the United Kingdom (31). The single study of dihydrotestosterone found no association with genotype (37).

In a study described in abstract form only (42), a statistically significant trend of decreasing levels of androstanediol glucuronide, a marker of 5alpha-reductase activity, with increasing number of A2 alleles was observed, but this was not confirmed in two other studies (31,37). No clear associations with estradiol (37, 53), sex hormone-binding globulin (31,53), or luteinizing hormone (31) were observed in the few studies in which these factors were investigated (Table 2).

Markers of hormonal status
With regard to markers of hormonal status in men, male pattern baldness, waist:hip ratio, stature, and bone mass have been investigated in single studies (37, 53, 69, 70). The findings are summarized in Web table 2 (posted on the website of the Human Genome Epidemiology Network , as well as on the Journal’s website). The only statistically significant associations observed were between the A2A2 genotype and stature and the cross-sectional area of the femoral neck (53).

 
  LABORATORY TESTS

The 5'-UTR polymorphism is detected most easily using polymerase chain reaction amplification followed by restriction fragment length polymorphism analysis involving digestion with MspAI. Several different primer sets have been developed (6, 10, 32). The promoter region polymorphism was identified by single-strand conformation polymorphism (8). Since it destroys a recognition site for NaeI, it could be detected routinely using restriction fragment length polymorphism methods. The single study that has reported on the intron 6 polymorphism detected it using single-strand conformation polymorphism methods (7).

Few studies have explicitly reported data on the proportion of specimens that were successfully analyzed and assigned a CYP17 genotype. In two articles in which these data were shown, Dunning et al. (33) successfully genotyped 97 percent of blood samples provided by breast cancer cases for the 5'-UTR polymorphism and Helzlsouer et al. (46) assigned genotypes to 97 percent of blood samples obtained from breast cancer cases and controls combined. Few data are available on the success rate for DNA extraction or the repeatability of genotyping, which, together with the percentage of specimens successfully genotyped, are important indicators of the analytical validity of the methods used (71). Information is lacking on the sensitivity and specificity of the restriction fragment length polymorphism methods used for detecting the true underlying genotype.

back to top
 
  CONCLUSIONS AND PRIORITIES FOR RESEARCH

Geographic and ethnic variation in prevalence
We have observed distinct geographic and ethnic variation in the prevalence of the A2 variant of the 5'-UTR polymorphism. Homozygosity for the variant appears to be more common in Japanese and other East Asian populations than in White populations in Europe and North America. The prevalence in Blacks was similar to that in Whites, but the available data on Blacks pertain almost exclusively to African Americans. In the single study of African Blacks (14), the frequency was lower than that in African Americans. Data on Hispanic populations are limited to two small US series (44, 48) in which the frequency was intermediate between that for Whites and that for East Asian populations.

Although 50 published reports located contained data on the A2A2 genotype frequency, many studies were of series that were unlikely to be representative of the general population. Thus, it not clear to what extent the observed variation might be due to the selected nature of many of the study series. Moreover, several ethnic groups (e.g., African Blacks; Asian populations other than those in Japan, Singapore, and Taiwan) have been relatively little studied. Therefore, further large population-based studies are needed in order to clarify the extent to which the prevalence of the 5'-UTR polymorphism varies by geographic area, ethnic group, and age. The functional consequences of the polymorphism are still unclear, and further investigation is needed to clarify them.

As regards the intron 6 and promoter region G-->A polymorphisms, almost nothing has been published on genotype frequency, gene function, or gene-disease associations. Already compiled population-based series or case-control or cohort studies might provide a relatively rapid route for the documentation of genotype frequencies and investigation of genotype-disease associations. If either polymorphism proved to be relatively common and/or to be related to disease, functional studies would be warranted.

Hormone levels
Interest in CYP17 in the etiology of particular diseases arose because of the key role of the gene in estrogen biosynthesis, and genetic variation, once discovered, was postulated to cause interindividual differences in circulating hormone levels. Some (but not all) of the relevant studies are consistent with the hypothesis that the A2 variant, by altering gene function in some way, alters hormone concentrations—at least, estrogen levels in premenopausal women. For postmenopausal women, the evidence is limited to one study (39) that included women not using HRT and that had null findings. The dominant sites for steroid hormone production in pre- and postmenopausal women are the ovaries and adrenal glands, respectively, and it has been suggested that CYP17 expression may differ between these sites (39). For men, the findings are inconsistent and inconclusive.

Methodological issues
Interpretation of studies of hormones in women is complicated by the methodological difficulties involved in adequately characterizing hormone levels, which vary over the menstrual cycle. For men, for hormones other than testosterone and dihydrotestosterone, little is known about the reliability of a single measurement. Hormone concentrations can vary by ethnic group, and this means that findings for one ethnic group may not be generalizable to another (72, 73). Additional difficulties are caused by several other factors: 1) hormone levels change with age (52, 74); 2) levels of some hormones are associated with anthropometric factors such as body mass index and waist:hip ratio (75–77); and 3) hormone levels may be influenced by behaviors such as tobacco smoking, alcohol consumption, and physical activity (52, 76–79). This suggests that genotype-hormone analyses should be adjusted for a range of possible confounders. Some hormones are interrelated: For example, the level of sex hormone-binding globulin determines levels of free estradiol and testosterone. This presents another issue of confounding. In addition, it is not clear whether what is biologically relevant with regard to disease risk is the total amount of a particular hormone or the free amount.

Therefore, further investigation is required to resolve the question of whether CYP17 genotype influences hormone concentrations in both sexes. Studies of different ethnic groups are needed. Both total and free hormone levels should be determined where relevant. Possible confounding of the association between genotype and one particular hormone by other hormones should be investigated. For women, repeated measurements should be taken at different points in the menstrual cycle, and pre- and postmenopausal women should be investigated separately. The effects (where relevant) of gravidity, use of exogenous hormones (e.g., oral contraceptives), and exposure to other lifestyle factors that can modulate hormone levels also require consideration.

Interactions
Other genes are in operation in the androgen and estrogen pathways (including SRD5A2, HSD17B, COMT, CYP1A1, and CYP19), and it seems likely that hormone levels, to the extent that they are under genetic control, are determined by the relative balance in the activities of the range of enzymes controlled by these genes. Thus, investigation of only a single polymorphism in a single gene (i.e., CYP17) is likely to be overly simplistic. There is currently a paucity of data on the effects of combinations of polymorphisms on hormone levels; therefore, the joint effects of CYP17 and other polymorphic genes involved in steroid biosynthesis require elucidation.

Age at menarche and other hormonally mediated factors
The evidence on whether CYP17 genotype affects hormonally mediated factors such at age at menarche is inconsistent and/or limited. The findings of most studies are compatible with the existence of a modest association between A2A2 genotype and earlier menarche. While this might be interpreted as consistent with the postulated effect of the A2 allele on estrogen levels, the associations have generally not been statistically significant. Synthesis of the findings of these studies is complicated by the fact that different authors have presented their results differently. Pooled analysis of findings from the current studies, with adjustment for years of recall, might help to clarify this matter in the short term. Stratification by ethnic group might also be helpful. The single study of breast development in young girls (49) supported an association between puberty and genotype. Further studies in girls and young women should be undertaken, where ethically possible, to complement the retrospective studies carried out in adult women.

Potential associations between CYP17 and age at menopause, breast density, and (in both sexes) bone mass/stature require further study. The current evidence is very limited.

HRT and use of other exogenous hormones
The association between HRT and CYP17 reported by Feigelson et al. (44)—A2 homozygotes were half as likely to use HRT as A1 homozygotes—is intriguing. Findings from the other four studies with available data on this issue (39, 46, 51, 66), all of which were smaller, are inconsistent. If it is confirmed, the association suggests that HRT may be a confounder of some gene-disease associations (e.g., the association with breast cancer), and hence genotype relative risks for these diseases should be adjusted for HRT use.

The possible CYP17-HRT association is consistent with the postulated effects of CYP17 on circulating estrogen levels. Associations between genotype and lifestyle factors have been observed for other polymorphic genes. For example, the CYP2A6 genotype has been associated with the likelihood of taking up tobacco smoking (80), and the ALDH2 genotype has been associated with whether people consume alcohol (81). Further investigations of the relation of CYP17 genotype to HRT and use of other exogenous hormones (e.g., oral contraceptives) might be illuminating. Use of exogenous hormones in a population is clearly influenced by culture, attitudes, medical prescribing practices, and press coverage of the results of research studies and medical opinion. Thus, studies carried out in populations where patterns of use differ from those in the United States would be very informative. Particularly valuable would be exploration of reasons for use or nonuse (poor tolerance, side effects, etc.). In the longer term, this might lead to the formulation of exogenous hormones tailored to specific genetic makeups. In the context of the current controversy surrounding the relative benefits and adverse effects of HRT (82,83), such research would be most valuable.

  Appendix
  Internet sites providing genetic information
back to top
 
 NOTES

Correspondence to Linda Sharp, National Cancer Registry Ireland, Elm Court, Boreenmanna Road, Cork, Ireland (e-mail: linda.sharp@ncri.ie)

back to top
TABLES AND FIGURES
back to top
 REFERENCES
  1. Picado-Leonard J, Miller WL. Cloning and sequence of the human gene for P450c17 (steroid 17 alpha-hydroxylase/17,20 lyase): similarity with the gene for P450c21. DNA 1987;6:439–48.
  2. Brentano ST, Picado-Leonard J, Mellon SH, et al. Tissue- specific, cyclic adenosine 3',5'-monophosphate-induced, and phorbol ester-repressed transcription from the human P450c17 promoter in mouse cells. Mol Endocrinol 1990;4:1972–9.
  3. Utiger RD. Insulin and the polycystic ovary syndrome. (Editorial). N Engl J Med 1996;335:657–8.
  4. Yanase T. 17 alpha-Hydroxylase/17,20-lyase defects. J Steroid Biochem Mol Biol 1995;53:153–7.
  5. Franks S, White D, Gilling-Smith C, et al. Hypersecretion of androgens by polycystic ovaries: the role of genetic factors in the regulation of cytochrome P450c17 alpha. Baillieres Clin Endocrinol Metab 1996;10:193–203.
  6. Carey AH, Waterworth D, Patel K, et al. Polycystic ovaries and premature male pattern baldness are associated with one allele of the steroid metabolism gene CYP17. Hum Mol Genet 1994;3:1873–6.
  7. Crocitto LE, Feigelson HS, Yu MC, et al. Short report on DNA marker at candidate locus: a polymorphism in intron 6 of the CYP17 gene. Clin Genet 1997;52:68–9.
  8. Miyoshi Y, Iwao K, Ikeda N, et al. Genetic polymorphism in CYP17 and breast cancer risk in Japanese women. Eur J Cancer 2000;36:2375–9.
  9. Kadonaga JT, Jones KA, Tijian R. Promoter-specific activation of RNA polymerase II transcription by SP1. Trends Biochem Sci 1986; 11:20–3.
  10. Feigelson HS, Coetzee GA, Kolonel LN, et al. A polymorphism in the CYP17 gene increases the risk of breast cancer. Cancer Res 1997;57:1063–5.
  11. Kristensen VN, Haraldsen EK, Anderson KB, et al. CYP17 and breast cancer risk: the polymorphism in the 5' flanking area of the gene does not influence binding to Sp-1. Cancer Res 1999;59:2825–8.
  12. Miyoshi Y, Ando A, Ooka M, et al. Association of CYP17 genetic polymorphism with intra-tumoral estradiol concentrations but not CYP17 messenger RNA levels in breast cancer tissue. Cancer Lett 2003;195:81–6.
  13. Daneshmand S, Weitsman SR, Navab A, et al. Overexpression of theca-cell messenger RNA in polycystic ovary syndrome does not correlate with polymorphisms in the cholesterol side-chain cleavage and 17alpha-hydroxylase/C 17-20-lyase promoters. Fertil Steril 2002;77:274–80.
  14. Kittles RA, Panguluri RK, Chen W, et al. CYP17 promoter variant associated with prostate cancer aggressiveness in African Americans. Cancer Epidemiol Biomarkers Prev 2001;10: 943–7.
  15. Huang J, Ushiyama T, Inoue K, et al. Possible association of CYP17 gene polymorphisms with the onset of rheumatoid arthritis. Clin Exp Rheumatol 1999;17:721–4.
  16. Habuchi T, Liqing Z, Suzuki T, et al. Increased risk of prostate cancer and benign prostatic hyperplasia associated with a CYP17 gene polymorphism with a gene dosage effect. Cancer Res 2000;60:5710–13.
  17. Kado N, Kitawaki J, Obayashi H, et al. Association of the CYP17 gene and CYP19 gene polymorphisms with risk of endometriosis in Japanese women. Hum Reprod 2002;17:897–902.
  18. Hamajima N, Iwata H, Obata Y, et al. No association of the 5' promoter region polymorphism of CYP17 with breast cancer risk in Japan. Jpn J Cancer Res 2000;91:880–5.
  19. Gorai I, Tanaka K, Inada M, et al. Estrogen-metabolizing gene polymorphisms, but not estrogen receptor-alpha gene polymorphisms, are associated with the onset of menarche in healthy postmenopausal Japanese women. J Clin Endocrinol Metab 2003;88:799–803.
  20. Wu AH, Seow A, Arakawa K, et al. HSD17B1 and CYP17 polymorphisms and breast cancer risk among Chinese women in Singapore. Int J Cancer 2003;104:450–7.
  21. Huang C-S, Chern H-D, Chang K-J, et al. Breast cancer risk associated with genotype polymorphism of the estrogen- metabolizing genes CYP17, CYP1A1, and COMT: a multigenic study on cancer susceptibility. Cancer Res 1999;59:4870–5.
  22. Lunn RM, Bell DA, Mohler JL, et al. Prostate cancer risk and polymorphism in 17 hydroxylase (CYP17) and steroid reductase (SRD5A2). Carcinogenesis 1999;20:1727–31.
  23. Yu MW, Yang YC, Yang SY, et al. Hormonal markers and hepatitis B virus-related hepatocellular carcinoma risk: a nested case-control study among men. J Natl Cancer Inst 2001;93:1644–51.
  24. Mitrunen K, Jourenkova N, Kataja V, et al. Steroid metabolism gene CYP17 polymorphism and the development of breast cancer. Cancer Epidemiol Biomarkers Prev 2000;9:1343–8.
  25. Diamanti-Kandarakis E, Bartzis MI, Zapanti ED, et al. Polymorphism T-->C (–34 bp) of gene CYP17 promoter in Greek patients with polycystic ovary syndrome. Fertil Steril 1999;71:431–5.
  26. Gudmundsdottir K, Thorlacius S, Jonasson JG, et al. CYP17 promoter polymorphism and breast cancer risk in males and females in relation to BRCA2 status. Br J Cancer 2003;88: 933–6.
  27. Marszalek B, Lacinski M, Babych N, et al. Investigations on the genetic polymorphism in the region of the CYP17 gene encoding 5'-UTR in patients with polycystic ovarian syndrome. Gynecol Endocrinol 2001;15:123–8.
  28. Kuligina ES, Togo AV, Suspitsin EN, et al. CYP17 polymorphism in the groups of distinct breast cancer susceptibility: comparison of patients with the bilateral disease vs. monolateral breast cancer patients vs. middle-aged female controls vs. elderly tumor-free women. Cancer Lett 2000;156:45–50.
  29. Bergman-Jungeström M, Gentile M, Lundin AC, et al. Association between CYP17 gene polymorphism and risk of breast cancer in young women. Int J Cancer 1999;84:350–3.
  30. Wadelius M, Andersson SO, Johansson JE, et al. Prostate cancer associated with CYP17 genotype. Pharmacogenetics 1999;9:635–9.
  31. Allen NE, Forrest MS, Key TJ. The association between polymorphisms in the CYP17 and 5 alpha-reductase (SRD5A2) genes and serum androgen concentrations in men. Cancer Epidemiol Biomarkers Prev 2001;10:185–9.
  32. Techatraisak K, Conway GS, Rumsby G. Frequency of a polymorphism in the regulatory region of the 17alpha-hydroxylase-17,20-lyase (CYP17) gene in hyperandrogenic states. Clin Endocrinol (Oxf) 1997;46:131–4.
  33. Dunning AM, Healey CS, Pharoah PD, et al. No association between a polymorphism in the steroid metabolism gene CYP17 and risk of breast cancer. Br J Cancer 1998;77:2045–7.
  34. Young IE, Kurian KM, Annink C, et al. A polymorphism in the CYP17 gene is associated with male breast cancer. Br J Cancer 1999;81:141–3.
  35. Jernström H, Vesprini D, Bradlow HL, et al. Re: CYP17 promoter polymorphism and breast cancer in Australian women under age forty years. (Letter). J Natl Cancer Inst 2001;93: 554–5.
  36. Lai J, Vesprini D, Chu W, et al. CYP gene polymorphisms and early menarche. Mol Genet Metab 2001;74:449–57.
  37. Haiman CA, Stampfer MJ, Giovannucci E, et al. The relationship between a polymorphism in CYP17 with plasma hormone levels and prostate cancer. Cancer Epidemiol Biomarkers Prev 2001;10:743–8.
  38. Haiman CA, Hankinson SE, Spiegelman D, et al. The relationship between a polymorphism in CYP17 with plasma hormone levels and breast cancer. Cancer Res 1999;59:1015–20.
  39. Haiman CA, Hankinson SE, Colditz GA, et al. A polymorphism in CYP17 and endometrial cancer risk. Cancer Res 2001;61:3955–60.
  40. Chang BL, Zheng SL, Isaacs SD, et al. Linkage and association of CYP17 gene in hereditary and sporadic prostate cancer. Int J Cancer 2001;95:354–9.
  41. Feigelson HS, Shames LS, Pike MC, et al. Cytochrome P450c17alpha gene (CYP17) polymorphism is associated with serum estrogen and progesterone concentrations. Cancer Res 1998;58:585–7.
  42. Makridakis N, Reichardt J, Pike M, et al. Genetic control of human androgen metabolism by the CYP17 and SRD5A2 genes. (). Am J Hum Genet 1997;61(suppl):A257.
  43. Goodman MT, McDuffie K, Guo C, et al. CYP17 genotype and ovarian cancer: a null case-control study. Cancer Epidemiol Biomarkers Prev 2001;10:563–4.
  44. Feigelson HS, McKean-Cowdin R, Pike MC, et al. Cytochrome P450c17alpha gene (CYP17) polymorphism predicts use of hormone replacement therapy. Cancer Res 1999;59:3908–10.
  45. Feigelson HS, McKean-Cowdin R, Coetzee GA, et al. Building a multigenic model of breast cancer susceptibility: CYP17 and HSD17B1 are two important candidates. Cancer Res 2001;61:785–9.
  46. Helzlsouer KJ, Huang HY, Strickland PT, et al. Association between CYP17 polymorphisms and the development of breast cancer. Cancer Epidemiol Biomarkers Prev 1998;7:945–9.
  47. Garner EI, Stokes EE, Berkowitz RS, et al. Polymorphisms of the estrogen-metabolizing genes CYP17 and catechol-O-methyltransferase and risk of epithelial ovarian cancer. Cancer Res 2002;62:3058–62.
  48. Weston A, Pan C, Bleiweiss IJ, et al. CYP17 genotype and breast cancer risk. Cancer Epidemiol Biomarkers Prev 1998;7:941–4.
  49. Kadlubar FF, Berkowitz GS, Delongchamp RR, et al. The CYP3A4*1B variant is related to the onset of puberty, a known risk factor for the development of breast cancer. Cancer Epidemiol Biomarkers Prev 2003;12:327–31.
  50. McCann SE, Moysich KB, Freudenheim JL, et al. The risk of breast cancer associated with dietary lignans differs by CYP17 genotype in women. J Nutr 2002;132:3036–41.
  51. Ambrosone CB, Moysich KB, Furberg H, et al. CYP17 genetic polymorphism, breast cancer, and breast cancer risk factors. Breast Cancer Res 2003;5:R45–51.
  52. García-Closas M, Herbstman J, Schiffman M, et al. Relationship between serum hormone concentrations, reproductive history, alcohol consumption and genetic polymorphisms in pre-menopausal women. Int J Cancer 2002;102:172–8.
  53. Zmuda JM, Cauley JA, Kuller LH, et al. A common promoter variant in the cytochrome P450c17alpha (CYP17) gene is associated with bioavailable testosterone levels and bone size in men. J Bone Miner Res 2001;16:911–17.
  54. Stanford JL, Noonan EA, Iwasaki L, et al. A polymorphism in the CYP17 gene and risk of prostate cancer. Cancer Epidemiol Biomarkers Prev 2002;11:243–7.
  55. Spurdle AB, Chen X, Abbazadegan M, et al. CYP17 promotor polymorphism and ovarian cancer risk. Int J Cancer 2000;86:436–9.
  56. Spurdle AB, Hopper JL, Dite GS, et al. CYP17 promoter polymorphism and breast cancer in Australian women under age forty years. J Natl Cancer Inst 2000;92:1674–81.
  57. Cui JS, Spurdle AB, Southey MC, et al. Regressive logistic and proportional hazards disease models for within-family analyses of measured genotypes, with application to a CYP17 polymorphism and breast cancer. Genet Epidemiol 2003;24:161–72.
  58. Ribeiro ML, Santos A, Carvalho-Salles AB, et al. Allelic frequencies of six polymorphic markers for risk of prostate cancer. Braz J Med Biol Res 2002;35:205–13.
  59. dos Santos A, Ribeiro ML, Mesquita JC, et al. No association of the 5' promoter region polymorphism of CYP17 gene with prostate cancer risk. Prostate Cancer Prostatic Dis 2002;5:28–31.
  60. Muti P, Trevisan M, Micheli A, et al. Reliability of serum hormones in premenopausal and postmenopausal women over a one-year period. Cancer Epidemiol Biomarkers Prev 1996;5:917–22.
  61. Ahmad N, Pollard TM, Unwin N. 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 2002;11:147–51.
  62. Vihko RK, Apter DL. The epidemiology and endocrinology of the menarche in relation to breast cancer. Cancer Surv 1986;5:561–71.
  63. Apter D, Reinila M, Vihko R. Some endocrine characteristics of early menarche, a risk factor for breast cancer, are preserved into adulthood. Int J Cancer 1989;44:783–7.
  64. Keating NL, Cleary PD, Rossi AS, et al. Use of hormone replacement therapy by postmenopausal women in the United States. Ann Intern Med 1999;130:545–53.
  65. Banks E, Barnes I, Baker K, et al. Use of hormonal therapy for menopause in nine European countries. (IARC Scientific Publication no. 156). Lyon, France: International Agency for Research on Cancer, 2002:301–3.
  66. McKean-Cowdin R, Feigelson HS, Pike MC, et al. Risk of endometrial cancer and estrogen replacement therapy history by CYP17 genotype. Cancer Res 2001;61:848–9.
  67. Haiman CA, Bernstein L, Van Den Berg D, et al. Genetic determinants of mammographic density. (Electronic article). Breast Cancer Res 2002;4:R5.
  68. Vermeulen A, Verdonck G. Representativeness of a single point plasma testosterone level for the long term hormonal milieu in men. J Clin Endocrinol Metab 1992;74:939–42.
  69. Khaw KT, Barrett-Connor E. Lower endogenous androgens predict central adiposity in men. Ann Epidemiol 1992;2:675–82.
  70. Simon D, Charles MA, Nahoul K, et al. Association between plasma total testosterone and cardiovascular risk factors in healthy adult men: The Telecom Study. J Clin Endocrinol Metab 1997;82:682–5.
  71. Little J, Bradley L, Bray MS, et al. Reporting, appraising, and integrating data on genotype prevalence and gene-disease associations. Am J Epidemiol 2002;156:300–10.
  72. Randolph JF Jr, Sowers M, Gold EB, et al. Reproductive hormones in the early menopausal transition: relationship to ethnicity, body size, and menopausal status. J Clin Endocrinol Metab 2003;88:1516–22.
  73. Heald AH, Ivison F, Anderson SG, et al. Significant ethnic variation in total and free testosterone concentration. Clin Endocrinol (Oxf) 2003;58:262–6.
  74. Vermeulen A, Goemaere S, Kaufman JM. Testosterone, body composition and aging. J Endocrinol Invest 1999;22:110–16.
  75. Gapstur SM, Gann PH, Kopp P, et al. Serum androgen concentrations in young men: a longitudinal analysis of associations with age, obesity, and race. The CARDIA Male Hormone Study. Cancer Epidemiol Biomarkers Prev 2002;11:1041–7.
  76. Daniel M, Martin AD, Faiman C. Sex hormones and adipose tissue distribution in premenopausal cigarette smokers. Int J Obes Relat Metab Disord 1992;16:245–54.
  77. Verkasalo PK, Thomas HV, Appleby PN, et al. Circulating levels of sex hormones and their relation to risk factors for breast cancer: a cross-sectional study in 1092 pre- and postmenopausal women
    (United Kingdom). Cancer Causes Control 2001;12:47–59.
  78. Oncken C, Prestwood K, Cooney JL, et al. Effects of smoking cessation or reduction on hormone profiles and bone turnover in postmenopausal women. Nicotine Tob Res 2002;4:451–8.
  79. Wu F, Ames R, Evans MC, et al. Determinants of sex hormone-binding globulin in normal postmenopausal women. Clin Endocrinol (Oxf) 2001;54:81–7.
  80. Tyndale RF, Sellers EM. Variable CYP2A6-mediated nicotine metabolism alters smoking behavior and risk. Drug Metab Dispos 2001;29:548–52.
  81. Lin YP, Hsieh HI, Chen YC, et al. Association between smoking, acetaldehyde dehydrogenase-2 1-1 status, and alcohol drinking among Taiwanese polyvinyl chloride workers. J Occup Environ Med 2001;43:701–5.
  82. Stephenson J. FDA orders estrogen safety warnings: agency offers guidance for HRT use. JAMA 2003;289:537–8.
  83. Office of Research on Women’s Health, National Heart, Lung, and Blood Institute. Women’s health and menopause: a comprehensive approach. Bethesda, MD: National Heart, Lung, and Blood Institute, 2002. (NIH publication no. 02-3284).
back to top
 
 INTERNET SITES
 
 MEDICAL LITERATURE SEARCH
Last Updated October 25, 2004