Hum. Reprod. Advance Access originally published online on August 18, 2006
Human Reproduction 2007 22(1):83-91; doi:10.1093/humrep/del318
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Testosterone levels in relation to oral contraceptive use and the androgen receptor CAG and GGC length polymorphisms in healthy young women
1 Department of Oncology and 2 Department of Cancer Epidemiology, Clinical Sciences, Lund University, Lund, Sweden
3 To whom correspondence should be addressed at: Helena Jernström, Department of Oncology, Clinical Sciences, Lund University, Barngatan 2:1, SE-221 85 Lund, Sweden. E-mail: helena.jernstrom{at}med.lu.se
 |
Abstract |
|---|
 Top Abstract IntroductionMaterial and methodsResultsDiscussionAcknowledgementsReferences |
|---|
BACKGROUND: The combined effect from the androgen receptor (AR) CAG and GGC length polymorphisms on testosterone levels has not been studied in young women. METHODS: Testosterone levels were measured in blood drawn on both menstrual cycle days 5–10 and 18–23 in 258 healthy women, aged
40 years, from high-risk breast cancer families. CAG and GGC length polymorphisms were analysed by PCR and fragment analyses. All women completed a questionnaire including information on oral contraceptive (OC) use and reproductive factors. RESULTS: OC users had lower median testosterone levels than non-users during cycle days 5–10 and 18–23 (P 0.005 for both). The BRCA mutation status was associated neither with testosterone levels nor with CAG or GGC length polymorphism. The CAG length polymorphism was not associated with testosterone levels. The cumulative number of long GGC alleles (
17 repeats) was significantly associated with lower testosterone levels in OC users during cycle days 5–10 (Ptrend =0.014), but not during cycle days 18–23 or in non-users. The interaction between the GGC length polymorphism and OC status was highly significant during cycle days 5–10 (P = 0.002) and near-significant during days 18–23 (P = 0.07). Incident breast cancer was more common in women with two short GGC alleles (log-rank P = 0.003). CONCLUSION: The GGC repeat length was the only significant genetic factor modifying the testosterone levels in current OC users from high-risk families. Homozygosity for the short GGC allele may be linked to the increased risk of early-onset breast cancer after OC exposure in high-risk women.
Key words: androgen receptor polymorphism/breast cancer/oral contraceptives/premenopausal women/testosterone levels
|
Introduction |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsDiscussionAcknowledgementsReferences |
|---|
Testosterone levels have been linked to several hormone-dependent diseases in women, such as polycystic ovarian syndrome (PCOS), hirsutism, androgenic alopecia, acne, type II diabetes mellitus and breast cancer (Redmond, 1998
; Westberg et al., 2001; Yu et al., 2003a; Kaaks et al., 2005). The role that testosterone might play in breast cancer pathogenesis has become the subject of increasing attention over the past several years (Westberg et al., 2001; Yu et al., 2003a; Kaaks et al., 2005). In women, 
25% of testosterone is produced in the ovaries and 25% in the adrenals and the remaining 50% is derived from peripheral conversion of proandrogens [dehydroepiandrosterone (DHEA) sulphate, DHEA and androstenedione] (Somboonporn and Davis, 2004). High levels of testosterone have been correlated with an increased risk of both pre- and post-menopausal breast cancers (Thomas et al., 1997; Key et al., 2002; Yu et al., 2003a,b; Micheli et al., 2004; Kaaks et al., 2005). The effects of testosterone on mammary tissue growth can be anti-estrogenic and thus anti-mitotic as well as indirectly estrogenic and stimulatory (Birrell et al., 1998; Yu et al., 2003a; Somboonporn and Davis, 2004). These effects depend on several factors, such as the type and dose of androgen, the breast cancer cell line, the structure of the androgen receptor (AR), the expression of co-regulators and co-factors (Somboonporn and Davis, 2004) and a possible interaction between testosterone and estrogen, progesterone and other growth-regulating hormones and their receptors (Birrell et al., 1998; Yu et al., 2003a).
Breast cancer is the most common type of cancer affecting women worldwide. Approximately 1 in 10 women in Sweden is diagnosed with breast cancer during her lifetime (data from the Swedish Cancer Registry). A positive family history of breast cancer approximately doubles the risk of the malignancy (Anderson et al., 2000). Hereditary factors are believed to play a role in 5–10% of the breast cancers (Narod and Foulkes, 2004; Dumitrescu and Cotarla, 2005; Lacroix and Leclercq, 2005). The most notable of the known mutations appears in the dominant high-penetrance genes BRCA1 and BRCA2 and confers a substantially increased risk of developing breast cancer and ovarian cancer at a young age (McPherson et al., 2000; Narod and Foulkes, 2004; Lacroix and Leclercq, 2005). However, a significant proportion of the breast cancer families with a dominant pattern of inheritance are non-BRCA1/2 families, also referred to as BRCAX families. Their pathogenesis is believed to depend on polymorphisms in several low-penetrance and high-prevalence genes working in combination with environmental factors (Narod and Foulkes, 2004; Lacroix and Leclercq, 2005). One such modifying gene may be the AR (Suter et al., 2003; Lillie et al., 2004; Lacroix and Leclercq, 2005), which is present in normal breast epithelium and in 40–85% of tumours (Lea et al., 1989; Soreide et al., 1992; Rody et al., 2005).
The AR is a regulator of androgen signalling in steroid hormone-sensitive cells, such as the breast epithelium and the hypothalamus. The AR gene is located on the X chromosome, and the AR protein functions as a transcription factor (Lundin et al., 2003). The N-terminal contains two functionally polymorphic microsatellites, a polyglutamin tract encoded by CAG repeats and a polyglycine tract encoded by (GGT)3GGG(GGT)2(GGC)n repeats (Faber et al., 1989; Lumbroso et al., 1997; Lundin et al., 2003). The transcriptional potential of the AR has so far been studied only in relation to the CAG repeat length in men (Krithivas et al., 1999) and in relation to CAG and GGC repeat lengths in vitro (Chamberlain et al., 1994; Gao et al., 1996; Irvine et al., 2000). Long repeats are associated with a weaker receptor activity, a weaker transcriptional potential and, possibly, compensatorily higher testosterone levels in men (Brufsky et al., 1997; Krithivas et al., 1999).
A few studies have investigated testosterone levels in relation to the different AR CAG genotypes in women (Westberg et al., 2001; Ibanez et al., 2003; Brum et al., 2005). To our knowledge, there are no studies on AR GGC genotypes in relation to testosterone levels in women. Short CAG repeat lengths have been reported to be associated with higher testosterone levels in young girls with ovarian hyperandrogenism (Ibanez et al., 2003), as well as in premenopausal (Westberg et al., 2001) and post-menopausal women (Brum et al., 2005). This suggests an increased androgen sensitivity (Ibanez et al., 2003) or a stimulatory effect of the AR on androgen production in women (Westberg et al., 2001), in contrast to the inhibitory effect in men (Brufsky et al., 1997). Thus, the results from studies in men cannot be extrapolated to women.
The different repeat length polymorphisms of AR have also been studied in relation to breast cancer risk in the general population (Giguere et al., 2001; Liede et al., 2003; Suter et al., 2003) and in women from high-risk breast cancer families (Rebbeck et al., 1999; Haiman et al., 2002). However, the risk with the different allele lengths may be modified by oral contraceptive (OC) use (Suter et al., 2003). OCs are also known to lower testosterone levels in most users (Carr et al., 1995; Jernström et al., 1997). Furthermore, OC use in young women has been linked to increased breast cancer risk in the general population (Collaborative Group on Hormonal Factors in Breast Cancer, 1996; Jernström et al., 2005) as well as in BRCA1/2 mutation carriers (Ursin et al., 1997; Narod et al., 2002).
To our knowledge, no one has studied the combined effects of both the AR CAG and the GGC repeat lengths and OC use on testosterone levels. The first aim of this study was to elucidate how the testosterone levels are influenced by the CAG and GGC genotypes in young women with and without current OC use. The second aim was to examine whether testosterone levels or CAG and GGC genotypes differ by BRCA mutation status. The third aim was to investigate whether testosterone levels or AR genotypes were associated with incident breast cancer.
|
Material and methods |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsDiscussionAcknowledgementsReferences |
|---|
Two hundred and fifty-eight young Swedish women from high-risk breast cancer families volunteered to participate in this study. They did not have any type of cancer at the time of enrolment between 1996 and 2002. Only women who were menstruating and who had not undergone a prophylactic mastectomy were eligible to participate. Potential participants were identified from charts and pedigrees from the Lund Oncogenetic Clinic. Individuals who themselves had been to the clinic were contacted, first by a letter including brief information on the study, then by telephone. Index individuals who were not eligible for the study were asked whether they were willing to inform relatives of the study and then inform us as to whether we might contact their relatives directly. The Ethics Committee of Lund University approved the study. All women signed a written informed consent.
A letter including an extensive epidemiologic questionnaire and a written consent form was mailed to women who verbally agreed to participate. The questionnaire included questions on reproductive factors, the use of OCs and other medications, smoking etc. A trained research nurse collected blood samples and body measurements once during the follicular phase, i.e. menstrual cycle days 5–10 (except for four women who had their follicular-phase samples drawn on day 4, and two women on day 11) and once again in the luteal phase, i.e. 5–10 days before the predicted onset of the following menstrual period, i.e. cycle days 18–23 in most women. All women were asked to call back with the date of the first day of their next menstrual period. Body measurements included height, weight, waist and hip circumferences and breast volume. The plasma and blood cells were separated and frozen at –70°C at our laboratory at the Department of Oncology, Lund.
Hormone analysis
Total testosterone in EDTA plasma was measured by electrochemiluminescent immunoassay with Roche Elecsys 1010/2010 and Modular analytics E170 analyzer (Roche Diagnostics, Mannheim, Germany). This system is based on the competitive binding analysis principle using monoclonal antibodies against testosterone. The sample was first incubated with biotinylated testosterone antibodies and testosterone derivate labelled with ruthenium complex (Elecsys Testosterone reagent kit nr 11776061–100 tests, Roche Diagnostics). The formed immune complexes were then incubated with streptavidin-coated microparticles, and the entire complexes were fixed to the solid phase via the interaction of biotin and streptavidin. The whole sample was then suctioned to measure cells, where the microparticles were fixed magnetically to the electrode, and the unbound particles were washed away. The following chemiluminescent reaction was measured in a photomultiplier. The measured light units have an association with the testosterone concentration in the sample. The maximal allowed immunoassay variation was 20%, and the detection limits were 0.069–52.00 nmol/l (Sanchez-Carbayo et al., 1998).
Genotyping
AR CAG
Genomic DNA was extracted from 300 µl of peripheral blood using Wizard, Genomic DNA Purification Kit (Promega, Madison, WI). The polymorphism in the AR gene is a trinucleotide repeat, ranging in size from 11 to 32 repeats. PCR primers 5'-GCGCGAAGTGATCCAGAA-3' (forward) and 5'-GTTGCTGTTCCTCATCCA-3' (reverse) were used, where the forward primer was fluorescently labelled with FAM (MWG-Biotech AG, Ebersberg, Germany) (Lubahn et al., 1989). PCR was performed in 15 µl reactions using 25 ng of DNA, 0.4 µM of each primer, 0.1 mM of each deoxynucleotide (Amersham Biosciences, Buckinghamshire, UK), 5% dimethylsulphoxide (DMSO, Sigma, St Louis, MO), 2.5 mM MgCl2 (Applied Biosystems, Foster City, CA), 1x PCR Gold buffer (Applied Biosystems) and 1 U AmpliTaq Gold (Applied Biosystems). The PCR product was analysed in an ABI3100 Genetic Analyzer (Applied Biosystems), and the results were evaluated using Genescan software. The number of repeats was determined by sequencing samples of varying size (Big Dye, Terminator Cycle Sequencing, Applied Biosystems) and using them as standards in fragment size analysis. For quality control, we re-analysed 20 samples. Nineteen samples were successfully analysed in the first attempt, and the concordance rate was 100%. We did not re-analyse the remaining sample.
AR GGC
Genomic DNA was extracted from 300 µl of peripheral blood using Wizard, Genomic DNA Purification Kit. The polymorphism in the AR gene is a trinucleotide repeat, ranging in size from 10 to 21 repeats. PCR primers 5'-TCCTGGCACACTCTCTTCAC-3' (forward) and 5'-GTTTCTTGCCAGGGTACCACACATCAGGT-3' (reverse) were used, where the forward primer was fluorescently labelled with HEX (MWG-Biotech AG) (Edwards et al., 1999).
PCR was performed in 15 µl reactions using 25 ng of DNA, 0.7 µM of each primer, 0.3 mM each of dATP, dCTP and dTTP (Amersham Biosciences), 0.06 mM of dGTP (Amersham Biosciences) and 0.24 mM of 7-Deaza-dGTP (Roche Diagnostics), 5% DMSO, 3.5 mM MgCl2, 1x PCR Gold buffer, 0.8 U AmpliTaq Gold and 0.1 U Cloned Pfu DNA Polymerase (Stratagene).
The PCR product was analysed in an ABI3100 Genetic Analyzer, and the results were evaluated using Genescan software. The number of repeats was determined by sequencing samples of varying size and using them as standards in fragment size analysis. For quality control, we re-analysed 20 samples and the concordance rate was 100%.
Mutation testing of the BRCA1 and BRCA2 genes was not performed as part of this study, and carrier status was obtained from the clinical records. BRCA1 and BRCA2 mutation testing is offered at the Oncogenetic Clinic of the Department of Oncology in Lund to women belonging to a high-risk breast cancer family—i.e. if an individual has at least three first-degree relatives with breast cancer and one diagnosed before age 50; two first-degree relatives with breast cancer and one diagnosed before age 40; or one first-degree relative with breast cancer diagnosed before age 30. Women are usually not offered testing before the age 25 years. BRCA gene mutation carriers included only those with confirmed deleterious alterations, i.e. nonsense or frameshift indel mutations that cause protein truncation, or known disease-associated missense mutations.
The participating women were classified into five different categories:
- Non-carriers belonging to BRCA1 and BRCA2 families, i.e. ‘BRCA1/ 2 negative’ (n = 55)
- BRCA1 mutation carriers (n = 23)
- BRCA2 mutation carriers (n = 7)
- Women from high-risk breast cancer families where no BRCA1 or BRCA2 mutations could be detected, i.e. members of a ‘BRCAX family’ (n = 111)
- Untested women from high-risk breast cancer families (n = 62)
Almost all women (n = 254) approved the analysis of the AR polymorphisms.
Follow-up
Women were followed until the development of a first breast cancer according to the regional cancer registries, until the date of a self-reported prophylactic mastectomy or until 17 May 2006, whichever came first. No woman had undergone a prophylactic oophorectomy before undergoing a prophylactic mastectomy. The report rate of cancer diagnoses to the Swedish Cancer Registries is close to a 100%. The clinical follow-up of high-risk women includes annual mammograms, ultrasounds and magnetic resonance images (MRIs) of the breasts in addition to a physical examination and annual follow-ups of the ovaries by ultrasound, CA-125 measurement and a gynaecological examination.
Data analyses
The Statistical Package for the Social Sciences 11.0.2 software was used for all statistical analyses. Mann–Whitney U-test was used to compare median testosterone levels in relation to OC status. Testosterone levels were not normally distributed, and the values were transformed using the natural logarithm (ln) to obtain a better distribution. Paired t-test was used to compare ln-transformed testosterone levels between cycle days 5–10 and 18–23. Linear regression models were used to compare circulating ln-transformed testosterone levels in relation to the cumulative number of long CAG and GGC alleles as well as to OC status. To test for interaction, we created an interaction term between OC status and cumulative number of long GGC alleles. The multivariate models were adjusted for age and day of the menstrual cycle when the blood samples were obtained. Log-rank tests were used to analyse the risk of incident breast cancers in relation to CAG and GGC genotypes. Cox-regression was used to analyse the hazard of developing breast cancer in relation to GGC genotype, adjusting for age and BRCA1 mutation status. A P-value of <0.05 was taken to be significant. All P-values were two-sided.
|
Results |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsDiscussionAcknowledgementsReferences |
|---|
The characteristics of the 258 women are summarized in Table I. Two hundred and fifty-four women were successfully genotyped for the CAG repeat lengths and 253 women for the GGC repeat lengths. Allele sizes varied between 11 and 31 repeats in the CAG repeat and between 10 and 21 in the GGC repeat (Tables II and III). The mean number of CAG repeats was 25.9, and the most common allele was the 20-repeat allele. The mean number of GGC repeats was 15.5, and the most common allele was the 17-repeat allele. The CAG and GGC repeat length distribution was similar in each of the groups of BRCA mutation status (BRCA1/2negative/BRCA1positive/BRCA2positive/BRCAX family/untested; data not shown). As per previous studies, the CAG allele with <22 repeats was classified as short (S) and with
22 repeats as long (L). The GGC allele with <17 repeats was classified as short and the GGC allele with 17 repeats as long (Dunning et al., 1999; Suter et al., 2003).
|
|
|
Testosterone levels in relation to OC use, time of the menstrual cycle and BRCA status
We then examined total testosterone levels in relation to the use of combined estrogen plus progestin OCs and different AR genotypes. We excluded women who were breastfeeding at the time of blood sample drawing (n = 4), women using hormonal contraceptives other than combined OCs, e.g. progesterone-only pills, injectable, intrauterine and implantable hormonal contraceptives (n = 19) or both (n = 1), as well as one woman who had not answered the question on hormonal contraceptive use (n = 1). Two hundred and thirty-three women were included in the analysis, 88 current OC users and 145 non-users.
Overall, median total testosterone levels were significantly lower in current OC users than in non-users both during cycle days 5–10 (1.24 versus 1.45 nmol/l; P = 0.005) and cycle days 18–23 (1.02 versus 1.65 nmol/l; P < 0.0001). Among non-users, testosterone levels were higher during the luteal phase than in the follicular phase (1.65 versus 1.45 nmol/l; P = 0.0001). In contrast to the non-users, current OC users had significantly lower median testosterone levels during cycle days 18–23 compared with cycle days 5–10 (1.02 versus 1.24 nmol/l; P < 0.0001). The BRCA mutation status did not modify the association between OC use and testosterone levels in either current OC users or non-users.
Testosterone levels in relation to CAG repeat lengths
The CAG alleles (S/S), (S/L), and (L/L) were not associated with testosterone levels in non-users or current OC users either during cycle days 5–10 or during cycle days 18–23 (Figure 1A and B). This fact remained after adjustment for age, nulliparity, exact day of the menstrual cycle when the blood sample was obtained, OC status, weight, waist-to-hip ratio and cumulative number of long CAG alleles.
|
Testosterone levels in relation to GGC repeat lengths
There was a significant interaction between the cumulative number of long GGC alleles (
17 repeats) and OC status on testosterone levels during cycle days 5–10 (P = 0.002) and near-significant interaction during cycle days 18–23 (P = 0.07) (Figure 2A and B). The increasing number of long GGC alleles was associated with lower testosterone levels in current OC users, but not in non-users. We stratified the model on OC status. Among non-users, there was no significant association between GGC and testosterone. Among current OC users, the testosterone levels decreased significantly with the number of long GGC alleles during cycle days 5–10 (Ptrend = 0.014), but not significantly during cycle days 18–23 (Ptrend = 0.10), adjusted for age, nulliparity, weight, waist-to-hip ratio and exact day of the menstrual cycle when the blood sample was obtained.
|
Because the frequency distribution of the GGC allele appeared to be biphasic, we also repeated our analyses defining <15 GGC repeats as short. The results remained similar but became insignificant, mainly because there were very few women with the new (S/S) GGC genotype.
Combined effects of CAG and GGC on testosterone levels
No linkage disequilibrium was observed between the frequencies of the different CAG and GGC alleles in a chi-square model (P = 0.24). We then selected 92 women who were homozygous for both the CAG and the GGC S or L alleles. Three women had four short CAG and GGC alleles, 41 women had two short CAG and two long GGC alleles, four women had two long CAG and two short GGC alleles and 44 women had four long CAG and GGC alleles. The combined effect of four short alleles was associated with the lowest testosterone levels in the three non-users with this genotype in both cycle phases (P = 0.03 and P = 0.04, respectively, adjusted for age, nulliparity, menstrual cycle day and CAG and GGC genotypes). None of the other combinations was significantly associated with the testosterone levels either in current OC users or in non-users. We then calculated the mean CAG and GGC repeat lengths for each woman. We did not observe any correlation between the mean lengths of the CAG or GGC repeats and testosterone levels among OC users or non-users.
New breast cancers in relation to CAG and GGC genotypes
To date, nine women have developed breast cancer, five with bilateral disease. The median follow-up time of the cohort was 4.53 years (range 0.07–10.06 years). Fourteen women have undergone a prophylactic mastectomy without the detection of cancer after inclusion in this study, and follow-up times for these women were censored at these time points. Five of the breast cancer patients were known BRCA1 mutation carriers, two belonged to BRCAX families, while two women belonged to untested families and have not undergone testing after their breast cancer diagnosis. The median age of diagnosis was 39 years (range 28–48 years). All women diagnosed with breast cancer were current or former OC users. We analysed the association of breast cancer incidence with AR polymorphisms. New breast cancers were not associated with CAG repeat lengths. (S/S) GGC genotype was associated with breast cancer during follow-up (log-rank 11.61, df 2; P = 0.003; Figure 3). No difference was seen between women with one or two long GGC alleles (P = 0.77). The increased hazard for early-onset breast cancer in women with the (S/S) GGC genotype remained significant after adjustment for age at inclusion and BRCA1 mutation status Hazard ratio 7.3 (95% confidence interval 1.3–39.3; P = 0.02). We did not observe any significant associations between testosterone levels in either cycle phase and breast cancer risk.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsDiscussionAcknowledgementsReferences |
|---|
The main finding in this study was that the GGC repeat length was the only significant genetic factor associated with testosterone levels, and only among current OC users during cycle days 5–10. There was a significant interaction between the GGC genotype and OC use on testosterone levels. The increasing number of long GGC alleles was associated with lower testosterone levels in current OC users but not in non-users. The CAG repeat length did not modify the testosterone levels in OC users or non-users. Current OC users had lower testosterone levels than non-users with almost all AR genotypes. New breast cancers were associated with the GGC but not with the CAG allele sizes. This is the first study to examine testosterone levels in relation to both CAG and GGC allele sizes in women and how the testosterone levels are affected by OC use in women with different CAG and GGC allele sizes.
In men, the AR with long CAG repeat lengths, which has lower activity, is associated with presumably compensatorily higher testosterone levels (Brufsky et al., 1997; Krithivas et al., 1999).
Short CAG repeat lengths have been linked with a modest increase in prostate cancer risk (Ntais et al., 2003; Singh et al., 2005). In contrast, the effect of the AR on testosterone levels in women is thought to be stimulatory rather than inhibitory (Westberg et al., 2001). In this study, we did not observe any association between the CAG repeat length and testosterone levels in this group of premenopausal women. The AR is located on the X chromosome, which is randomly inactivated in women (Brum et al., 2005). In this study, we have not performed X-chromosome inactivation analysis. However, high testosterone levels have been reported in post-menopausal women with short CAG repeat lengths of the AR (Westberg et al., 2001; Ibanez et al., 2003), even when taking into consideration the X-chromosome inactivation patterns (Brum et al., 2005).
Lower testosterone levels have been linked to OC use in premenopausal women (Carr et al., 1995; Jernström et al., 1997). OC use at a young age or before the first child has also been associated with an increased breast cancer risk (Collaborative Group on Hormonal Factors in Breast Cancer, 1996; Jernström et al., 2005). The exogenous estrogens in OCs may reduce LH-driven ovarian stromal steroidogenesis, which leads to lower total testosterone levels (Carr et al., 1995). Low testosterone levels have also been linked to an earlier age at breast cancer diagnosis and a more aggressive disease (Bulbrook and Thomas, 1989) as well as to a positive family history of breast cancer (Jernström et al., 1997). However, in more recent studies, high levels of testosterone have been linked to an increased risk of both pre- and post-menopausal breast cancers (Thomas et al., 1997; Key et al., 2002; Yu et al., 2003a,b; Micheli et al., 2004; Kaaks et al., 2005).
In this study, we did not observe any significant association between testosterone levels and incident breast cancers, but we only had a small number of cases of which one used OCs and another an intrauterine device with progestin at the time of blood sampling. The inconsistencies between studies on premenopausal testosterone levels in relation to breast cancer may depend on the menstrual cyclic variation of sex hormone levels, including testosterone (Micheli et al., 2004). One of the strengths in the present study is that testosterone levels were measured in a standardized way twice during the menstrual cycle in all women. Furthermore, we adjusted for the exact day of the menstrual cycle, taking into account days from the onset of the next period from samples obtained during cycle days 5–10, as well as the actual onset of the next menstrual period for samples obtained during cycle days 18–23. Our findings are consistent with earlier studies showing that the use of OC is associated with decreased testosterone levels in most women. We also show that the strength of the associations between the GGC allele sizes and testosterone levels varies within a menstrual cycle, even when we adjust for the exact menstrual cycle day. It is possible that this variation is partly because we did not obtain all blood samples on the exact same cycle day in all women, but it was not feasible for us to do so for practical reasons.
To date, nine women in this prospective cohort have developed breast cancer, five with bilateral disease, after the inclusion. All were current or former OC users. In the present study, OC users had lower testosterone levels than non-users with all AR genotypes, except for the (S/S) GGC genotype. We therefore examined the risk for incident breast cancer in relation to the AR genotypes. We found an increased risk associated with the (S/S) GGC genotype, which was associated with high testosterone levels in OC users. This finding is consistent with the earlier studies indicating a higher risk of breast cancer with higher testosterone levels (Thomas et al., 1997; Key et al., 2002; Yu et al., 2003a,b; Micheli et al., 2004; Kaaks et al., 2005). Conversely, other studies report that GGC repeat lengths are not associated with higher breast cancer risk (Dunning et al., 1999; Kadouri et al., 2001; Suter et al., 2003). One study reported a risk reduction with this genotype in combination with OCs (Suter et al., 2003). In this study, CAG polymorphisms were not associated with higher breast cancer risk. The findings are in line with some earlier studies, which reported that CAG is not associated with higher breast cancer risk in women from the general population (Dunning et al., 1999; Kadouri et al., 2001) or from high-risk families (Kadouri et al., 2001; Spurdle et al., 2005), but not with other studies that reported an association with higher breast cancer risk and long CAG repeat lengths in women from the general population and from the high-risk families (Giguere et al., 2001; Haiman et al., 2002; Liede et al., 2003; Suter et al., 2003; Lillie et al., 2004).
This is an early report from a prospective cohort of young healthy women from high-risk breast cancer families, and our results should therefore be considered preliminary. The median time of follow-up is now 4.53 years or 1519 person-years. This gives an annual breast cancer incidence in our cohort of 0.59%. This overall annual incidence rate is lower than that in other cohorts of high-risk women consisting of BRCA1 and BRCA2 carriers (Easton et al., 1995; Tonin et al., 1995; Ford et al., 1998). However, the present cohort includes 55 women who are known to not carry the respective BRCA1 or BRCA2 mutations that segregate in their families, as well as 111 women from BRCAX families and 62 women who are untested. Five of the nine cases are known BRCA1 mutation carriers, two cases belonged to BRCAX families and two cases were untested. We cannot exclude the possibility that the untested cases also carry BRCA1 or BRCA2 mutations. Two of the six cases were related, maternal aunt and niece. They both had the (L/L) GGC genotype and were BRCA1 mutation carriers. The association with the GGC genotype and breast cancer remained after adjustment for BRCA1 mutation status and when we re-analysed the data excluding either one of them. The prophylactic mastectomies are self-reported, and it is possible that other women have undergone the operation without notifying us. However, given that high-risk women undergo annual clinical examinations, it is unlikely that we missed more than one or two mastectomies, if any.
The BRCA status was not associated with the AR genotype or testosterone levels in our study, which is in line with the findings of others (Kadouri et al., 2001; Spurdle et al., 2005). The CAG distribution and peaks were similar to those seen in the general population of premenopausal Swedish women in a previous study by Westberg et al. (2001). CAG and GGC alleles were generally longer than in previous studies (Kadouri et al., 2001; Suter et al., 2003), a possible explanation being the differing ethnic backgrounds of the study populations, which can influence the AR repeat lengths (Lundin et al., 2003; Lillie et al., 2004), or non-standardized reference samples (Delmotte et al., 2001; Boorman et al., 2002). In this study, we used a ladder created with the actual size obtained by sequencing (Rodriguez et al., 2006).
To our knowledge, this is the first study to examine how the testosterone levels are influenced by the CAG and GGC genotypes in young women with and without current OC use. Our results showed that the testosterone levels were lower in most OC users and were modified by the GGC repeat lengths, but not the CAG repeat lengths, of the AR. GGC repeat length was the only significant genetic factor significantly associated with testosterone levels and only during menstrual cycle days 5–10. There was a significant interaction between the GGC genotype and OC use on testosterone levels. Because BRCA mutation status showed no association with the AR genotypes or the testosterone levels, we believe that our results may also apply to women in the general population. Furthermore, our data suggest that high-risk women who are homozygous for the short GGC repeat allele may be at an increased risk of the early-onset breast cancer after OC exposure regardless of BRCA mutation status, but this finding needs confirmation in other studies.
|
Acknowledgements |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsDiscussionAcknowledgementsReferences |
|---|
This study was supported by grants from Grönbergska fonden, Vetenskapsrådet (The Swedish Research Council, K2001–27GX–14120–01A), the Medical Faculty in Lund, the Mrs Berta Kamprad’s Foundation, The South Swedish Health Care Region (Region Skåne) and the Lund Hospital Fund. Helena Jernström’s position is funded by the Grönbergska foundation, The Swedish Research Council (K2002–27GP–14104–02B). We thank our research nurses Kerstin Nilsson, Monica Pehrsson and Karin Henriksson for their assistance with body measurements and blood drawing, and Johanna Wagenius, Johanna Frenander, Helen Sundberg, Malin Sternby and Susanna Holmquist for their assistance with recruitment. We thank Erika Bågeman for the validation of the genetic analyses. We also thank Dr Eric T.Dryver for proofreading the manuscript.
|
References |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsDiscussionAcknowledgementsReferences |
|---|
Anderson H, Bladström A, Olsson H, Möller TR. (2000) Familial breast and ovarian cancer: a Swedish population-based register study. Am J Epidemiol 152:1154–1163.
Birrell SN, Hall RE, Tilley WD. (1998) Role of the androgen receptor in human breast cancer. J Mammary Gland Biol Neoplasia 3:95–103.[CrossRef][ISI][Medline]
Boorman DW, Guo Y, Visvanathan K, Helzlsouer K, O’Brien TG. (2002) Automated fragment analysis method for determining androgen receptor CAG repeat length. Biotechniques 33:140–143.[ISI][Medline]
Brufsky A, Fontaine-Rothe P, Berlane K, Rieker P, Jiroutek M, Kaplan I, Kaufman D, Kantoff P. (1997) Finasteride and flutamide as potency-sparing androgen-ablative therapy for advanced adenocarcinoma of the prostate. Urology 49:913–920.[CrossRef][ISI][Medline]
Brum IS, Spritzer PM, Paris F, Maturana MA, Audran F, Sultan C. (2005) Association between androgen receptor gene CAG repeat polymorphism and plasma testosterone levels in postmenopausal women. J Soc Gynecol Investig 12:135–141.[ISI][Medline]
Bulbrook RD and Thomas BS. (1989) Hormones are ambiguous risk factors for breast cancer. Acta Oncol 28:841–847.[ISI][Medline]
Carr BR, Breslau NA, Givens C, Byrd W, Barnett-Hamm C, Marshburn PB. (1995) Oral contraceptive pills, gonadotropin-releasing hormone agonists, or use in combination for treatment of hirsutism: a clinical research center study. J Clin Endocrinol Metab 80:1169–1178.[Abstract]
Chamberlain NL, Driver ED, Miesfeld RL. (1994) The length and location of CAG trinucleotide repeats in the androgen receptor N-terminal domain affect transactivation function. Nucleic Acids Res 22:3181–3186.
Collaborative Group on Hormonal Factors in Breast Cancer. (1996) Breast cancer and hormonal contraceptives: collaborative reanalysis of individual data on 53 297 women with breast cancer and 100 239 women without breast cancer from 54 epidemiological studies. Lancet 347:1713–1727.[CrossRef][ISI][Medline]
Delmotte F, Leterme N, Simon JC. (2001) Microsatellite allele sizing: difference between automated capillary electrophoresis and manual technique. Biotechniques 31: 810,814–810,816,818.
Dumitrescu RG and Cotarla I. (2005) Understanding breast cancer risk – where do we stand in 2005? J Cell Mol Med 9:208–221.[ISI][Medline]
Dunning AM, McBride S, Gregory J, Durocher F, Foster NA, Healey CS, Smith N, Pharoah PD, Luben RN, Easton DF, et al. (1999) No association between androgen or vitamin D receptor gene polymorphisms and risk of breast cancer. Carcinogenesis 20:2131–2135.
Easton DF, Ford D, Bishop DT. (1995) Breast and ovarian cancer incidence in BRCA1-mutation carriers. Breast Cancer Linkage Consortium. Am J Hum Genet 56:265–271.[ISI][Medline]
Edwards SM, Badzioch MD, Minter R, Hamoudi R, Collins N, Ardern-Jones A, Dowe A, Osborne S, Kelly J, Shearer R, et al. (1999) Androgen receptor polymorphisms: association with prostate cancer risk, relapse and overall survival. Int J Cancer 84:458–465.[CrossRef][ISI][Medline]
Faber PW, Kuiper GG, van Rooij HC, van der Korput JA, Brinkmann AO, Trapman J. (1989) The N-terminal domain of the human androgen receptor is encoded by one, large exon. Mol Cell Endocrinol 61:257–262.[CrossRef][ISI][Medline]
Ford D, Easton DF, Stratton M, Narod S, Goldgar D, Devilee P, Bishop DT, Weber B, Lenoir G, Chang-Claude J, et al. (1998) Genetic heterogeneity and penetrance analysis of the BRCA1 and BRCA2 genes in breast cancer families. The Breast Cancer Linkage Consortium. Am J Hum Genet 62:676–689.[CrossRef][ISI][Medline]
Gao T, Marcelli M, McPhaul MJ. (1996) Transcriptional activation and transient expression of the human androgen receptor. J Steroid Biochem Mol Biol 59:9–20.[CrossRef][ISI][Medline]
Giguere Y, Dewailly E, Brisson J, Ayotte P, Laflamme N, Demers A, Forest VI, Dodin S, Robert J, Rousseau F. (2001) Short polyglutamine tracts in the androgen receptor are protective against breast cancer in the general population. Cancer Res 61:5869–5874.
Haiman CA, Brown M, Hankinson SE, Spiegelman D, Colditz GA, Willett WC, Kantoff PW, Hunter DJ. (2002) The androgen receptor CAG repeat polymorphism and risk of breast cancer in the Nurses’ Health Study. Cancer Res 62:1045–1049.
Ibanez L, Ong KK, Mongan N, Jaaskelainen J, Marcos MV, Hughes IA, De Zegher F, Dunger DB. (2003) Androgen receptor gene CAG repeat polymorphism in the development of ovarian hyperandrogenism. J Clin Endocrinol Metab 88:3333–3338.
Irvine RA, Ma H, Yu MC, Ross RK, Stallcup MR, Coetzee GA. (2000) Inhibition of p160-mediated coactivation with increasing androgen receptor polyglutamine length. Hum Mol Genet 9:267–274.
Jernström HC, Olsson H, Borg Å. (1997) Reduced testosterone, 17 beta-oestradiol and sexual hormone binding globulin, and increased insulin-like growth factor-1 concentrations, in healthy nulligravid women aged 19–25 years who were first and/or second degree relatives to breast cancer patients. Eur J Cancer Prev 6:330–340.[CrossRef][ISI][Medline]
Jernström H, Sandberg T, Bageman E, Borg A, Olsson H. (2005) Insulin-like growth factor-1 (IGF1) genotype predicts breast volume after pregnancy and hormonal contraception and is associated with circulating IGF-1 levels: implications for risk of early-onset breast cancer in young women from hereditary breast cancer families. Br J Cancer 92:857–866.[CrossRef][ISI][Medline]
Kaaks R, Berrino F, Key T, Rinaldi S, Dossus L, Biessy C, Secreto G, Amiano P, Bingham S, Boeing H, et al. (2005) Serum sex steroids in premenopausal women and breast cancer risk within the European Prospective Investigation into Cancer and Nutrition (EPIC). J Natl Cancer Inst 97:755–765.
Kadouri L, Easton DF, Edwards S, Hubert A, Kote-Jarai Z, Glaser B, Durocher F, Abeliovich D, Peretz T, Eeles RA. (2001) CAG and GGC repeat polymorphisms in the androgen receptor gene and breast cancer susceptibility in BRCA1/2 carriers and non-carriers. Br J Cancer 85:36–40.[CrossRef][ISI][Medline]
Key T, Appleby P, Barnes I, Reeves G. (2002) Endogenous sex hormones and breast cancer in postmenopausal women: reanalysis of nine prospective studies. J Natl Cancer Inst 94:606–616.
Krithivas K, Yurgalevitch SM, Mohr BA, Wilcox CJ, Batter SJ, Brown M, Longcope C, McKinlay JB, Kantoff PW. (1999) Evidence that the CAG repeat in the androgen receptor gene is associated with the age-related decline in serum androgen levels in men. J Endocrinol 162:137–142.[Abstract]
Lacroix M and Leclercq G. (2005) The ‘portrait’ of hereditary breast cancer. Breast Cancer Res Treat 89:297–304.[CrossRef][ISI][Medline]
Lea OA, Kvinnsland S, Thorsen T. (1989) Improved measurement of androgen receptors in human breast cancer. Cancer Res 49:7162–7167.[ISI][Medline]
Liede A, Zhang W, De Leon Matsuda ML, Tan A, Narod SA. (2003) Androgen receptor gene polymorphism and breast cancer susceptibility in The Philippines. Cancer Epidemiol Biomarkers Prev 12:848–852.
Lillie EO, Bernstein L, Ingles SA, Gauderman WJ, Rivas GE, Gagalang V, Krontiris T, Ursin G. (2004) Polymorphism in the androgen receptor and mammographic density in women taking and not taking estrogen and progestin therapy. Cancer Res 64:1237–1241.
Lubahn DB, Brown TR, Simental JA, Higgs HN, Migeon CJ, Wilson EM, French FS. (1989) Sequence of the intron/exon junctions of the coding region of the human androgen receptor gene and identification of a point mutation in a family with complete androgen insensitivity. Proc Natl Acad Sci USA 86:9534–9538.
Lumbroso R, Beitel LK, Vasiliou DM, Trifiro MA, Pinsky L. (1997) Codon-usage variants in the polymorphic (GGN)n trinucleotide repeat of the human androgen receptor gene. Hum Genet 101:43–46.[CrossRef][ISI][Medline]
Lundin KB, Giwercman A, Richthoff J, Abrahamsson PA, Giwercman YL. (2003) No association between mutations in the human androgen receptor GGN repeat and inter-sex conditions. Mol Hum Reprod 9:375–379.
McPherson K, Steel CM, Dixon JM. (2000) ABC of breast diseases. Breast cancer-epidemiology, risk factors, and genetics. BMJ 321:624–628.
Micheli A, Muti P, Secreto G, Krogh V, Meneghini E, Venturelli E, Sieri S, Pala V, Berrino F. (2004) Endogenous sex hormones and subsequent breast cancer in premenopausal women. Int J Cancer 112:312–318.[CrossRef][ISI][Medline]
Narod SA and Foulkes WD. (2004) BRCA1 and BRCA2: 1994 and beyond. Nat Rev Cancer 4:665–676.[CrossRef][ISI][Medline]
Narod SA, Dube MP, Klijn J, Lubinski J, Lynch HT, Ghadirian P, Provencher D, Heimdal K, Moller P, Robson M, et al. (2002) Oral contraceptives and the risk of breast cancer in BRCA1 and BRCA2 mutation carriers. J Natl Cancer Inst 94:1773–1779.
Ntais C, Polycarpou A, Tsatsoulis A. (2003) Molecular epidemiology of prostate cancer: androgens and polymorphisms in androgen-related genes. Eur J Endocrinol 149:469–477.[Abstract]
Rebbeck TR, Kantoff PW, Krithivas K, Neuhausen S, Blackwood MA, Godwin AK, Daly MB, Narod SA, Garber JE, Lynch HT, et al. (1999) Modification of BRCA1-associated breast cancer risk by the polymorphic androgen-receptor CAG repeat. Am J Hum Genet 64:1371–1377.[CrossRef][ISI][Medline]
Redmond GP. (1998) Androgens and women’s health. Int J Fertil Womens Med 43:91–97.[ISI][Medline]
Rodriguez G, Bilbao C, Ramirez R, Falcon O, Leon L, Chirino R, Falcon O Jr, Diaz BP, Rivero JF, Perucho M, et al. (2006) Alleles with short CAG and GGN repeats in the androgen receptor gene are associated with benign endometrial cancer. Int J Cancer 118:1420–1425.[CrossRef][ISI][Medline]
Rody A, Diallo R, Poremba C, Wuelfing P, Kissler S, Solbach C, Kaufmann M, Jackisch C. (2005) Androgen receptor expression in ductal carcinoma in situ of the breast: not a helpful marker for classification such as estrogen receptor alpha and progesterone receptor. Appl Immunohistochem Molecul Morphol 13:25–29.[CrossRef]
Sanchez-Carbayo M, Mauri M, Alfayate R, Miralles C, Soria F. (1998) Elecsys testosterone assay evaluated. Clin Chem 44:1744–1746.
Singh AS, Chau CH, Price DK, Figg WD. (2005) Mechanisms of disease: polymorphisms of androgen regulatory genes in the development of prostate cancer. Nat Clin Pract Urol 2:101–107.[CrossRef][ISI][Medline]
Somboonporn W and Davis SR. (2004) Testosterone effects on the breast: implications for testosterone therapy for women. Endocr Rev 25:374–388.
Soreide JA, Lea OA, Varhaug JE, Skarstein A, Kvinnsland S. (1992) Androgen receptors in operable breast cancer: relation to other steroid hormone receptors, correlations to prognostic factors and predictive value for effect of adjuvant tamoxifen treatment. Eur J Surg Oncol 18:112–118.[ISI][Medline]
Spurdle AB, Antoniou AC, Duffy DL, Pandeya N, Kelemen L, Chen X, Peock S, Cook MR, Smith PL, Purdie DM, et al. (2005) The androgen receptor CAG repeat polymorphism and modification of breast cancer risk in BRCA1 and BRCA2 mutation carriers. Breast Cancer Res 7:R176–R183.[CrossRef][ISI][Medline]
Suter NM, Malone KE, Daling JR, Doody DR, Ostrander EA. (2003) Androgen receptor (CAG)n and (GGC)n polymorphisms and breast cancer risk in a population-based case-control study of young women. Cancer Epidemiol Biomarkers Prev 12:127–135.
Thomas HV, Reeves GK, Key TJ. (1997) Endogenous estrogen and postmenopausal breast cancer: a quantitative review. Cancer Causes Control 8:922–928.[CrossRef][ISI][Medline]
Tonin P, Ghadirian P, Phelan C, Lenoir GM, Lynch HT, Letendre F, Belanger D, Monte M, Narod SA. (1995) A large multisite cancer family is linked to BRCA2. J Med Genet 32:982–984.[Abstract]
Ursin G, Henderson BE, Haile RW, Pike MC, Zhou N, Diep A, Bernstein L. (1997) Does oral contraceptive use increase the risk of breast cancer in women with BRCA1/BRCA2 mutations more than in other women? Cancer Res 57:3678–3681.
Westberg L, Baghaei F, Rosmond R, Hellstrand M, Landen M, Jansson M, Holm G, Bjorntorp P, Eriksson E. (2001) Polymorphisms of the androgen receptor gene and the estrogen receptor beta gene are associated with androgen levels in women. J Clin Endocrinol Metab 86:2562–2568.
Yu H, Shu XO, Li BD, Dai Q, Gao YT, Jin F, Zheng W. (2003a) Joint effect of insulin-like growth factors and sex steroids on breast cancer risk. Cancer Epidemiol Biomarkers Prev 12:1067–1073.
Yu H, Shu XO, Shi R, Dai Q, Jin F, Gao YT, Li BD, Zheng W. (2003b) Plasma sex steroid hormones and breast cancer risk in Chinese women. Int J Cancer 105:92–97.[CrossRef][ISI][Medline]
Submitted on May 15, 2006; resubmitted on June 15, 2006; resubmitted on July 5, 2006; accepted on July 10, 2006.
CiteULike 
Connotea 
Del.icio.us What's this?
| |||||||||||||||