Nutritional Epidemiology
Polymorphisms in the CYP19 Gene May Affect the Positive Correlations between Serum and Urine Phytoestrogen Metabolites and Plasma Androgen Concentrations in Men1 Yen-Ling Low,* James I. Taylor,* Philip B. Grace,*† Mitch Dowsett,** Elizabeth Folkerd,** Deborah Doody,** Alison M. Dunning,‡ Serena Scollen,‡ Angela A. Mulligan,†† Ailsa A. Welch,†† Robert N. Luben,†† Kay-Tee Khaw,†† Nick E. Day,†† Nick J. Wareham,†† and Sheila A. Bingham*††2 *MRC Dunn Human Nutrition Unit, Cambridge, UK; †HFL, Newmarket Road, Fordham, UK; **Institute of Cancer Research, London, UK; ‡CR-UK Department of Oncology, Strangeways Research Laboratory, Cambridge, UK; and ††EPIC, Institute of Public Health and Strangeways Research Laboratory, Cambridge, UK ABSTRACT Phytoestrogens have been hypothesized to protect against prostate cancer via modulation of circulating androgen concentrations. We conducted a cross-sectional study of 267 men in the Norfolk arm of the European Prospective Investigation into Cancer and Nutrition (EPIC) cohort with 2 aims: first, to investigate the association between phytoestrogen exposure (measured from diet, urine, and serum) and plasma concentrations of sex hormone-binding globulin (SHBG), androstanediol glucuronide, testosterone and Free Androgen Index (FAI); and second, whether the association may be modified by polymorphisms in CYP19 and SHBG genes. Dietary daidzein and genistein intakes were obtained from food diaries and computed using an in-house food composition database. Urinary and serum concentrations of 3 isoflavones (daidzein, genistein, glycitein), 2 daidzein metabolites O-desmethylangolensin (O-DMA) and 2 lignan metabolites (enterodiol and enterolactone) were measured using mass spectrometry. There was no association between dietary, urinary, and serum phytoestrogens and plasma SHBG concentrations. Enterolactone was positively associated with plasma androstanediol glucuronide concentrations (urinary enterolactone: r ⫽ 0.127, P ⫽ 0.043; serum enterolactone: r ⫽ 0.172, P ⫽ 0.006) and FAI (urinary enterolactone: r ⫽ 0.115, P ⫽ 0.067; serum enterolactone: r ⫽ 0.158, P ⫽ 0.011). Both urinary and serum equol were associated with plasma testosterone (urinary equol: r ⫽ 0.332, P ⫽ 0.013; serum equol: r ⫽ 0.318, P ⫽ 0.018) and FAI (urinary equol: r ⫽ 0.297, P ⫽ 0.027; serum equol: r ⫽ 0.380, P ⫽ 0.004) among men with the TT genotype but not the CC or CT genotypes (r ⫽ – 0.029 to – 0.134, P ⫽ 0.091– 0.717) for the CYP19 3⬘untranslated region (UTR) T-C polymorphism. Urinary and serum enterolactone showed similar genotype-dependent associations with testosterone but not with FAI. In this first study on phytoestrogen-gene associations in men, we conclude that enterolactone and equol are positively associated with plasma androgen concentrations, and interactions with CYP19 gene may be involved. J. Nutr. 135: 2680 –2686, 2005. KEY WORDS:
●
phytoestrogens
●
biomarkers
●
polymorphism
●
androgens
O-desmethylangolensin (O-DMA)3 (3). The lignans include enterolactone and enterodiol, derived from colonic microbial fermentation of matairesinol and secoisolariciresinol, which are found in a wide variety of plant foods. One of the potential mechanisms by which phytoestrogens are thought to protect against prostate cancer is via modulation of circulating androgen concentrations. Prostate cancer is a hormone-responsive disease and androgen deprivation has been the mainstay of treatment for locally advanced and metastatic disease (4). Phytoestrogens were shown to inhibit
Animal and epidemiologic studies have largely pointed toward a protective effect of soy consumption on prostate cancer risk (1,2). The bioactive compounds in soy are thought to be phytoestrogens, which are naturally occurring plant compounds that are structurally similar to the hormone 17estradiol. Phytoestrogens in the human diet can be divided into 2 main groups, the isoflavones and the lignans. The isoflavones include glycitein, daidzein, and genistein found in legumes and especially soy, and their metabolites equol, and
1 Supported by the UK Medical Research Council, UK Food Standards Agency, Cancer Research UK, World Cancer Research Fund and the Agency for Science, Technology and Research, Singapore. 2 To whom correspondence should be addressed. E-mail:
[email protected].
3 Abbreviations used: EPIC, European Prospective Investigation into Cancer and Nutrition; SHBG, sex hormone-binding globulin; FAI, Free Androgen Index; O-DMA, O-desmethylangolensin; SNP, single nucleotide polymorphism; UTR, untranslated region.
0022-3166/05 $8.00 © 2005 American Society for Nutrition. Manuscript received 27 June 2005. Initial review completed 22 July 2005. Revision accepted 23 August 2005. 2680
PHYTOESTROGEN GENE ASSOCIATIONS WITH PLASMA ANDROGENS
androgen metabolism enzymes such as aromatase, 5␣-reductase, and 17-hydroxysteroid dehydrogenase in vitro (5–11). There is also some evidence that phytoestrogens can stimulate the production of sex hormone-binding globulin (SHBG) (12– 15), the main protein that binds to androgens and restricts their biological activities. The effects of phytoestrogens on sex hormone concentrations in men were investigated in several intervention studies (16 –25) and 2 cross-sectional studies (26,27) but results have been conflicting. All of the previous studies used dietary phytoestrogen intake as the exposure measure. However, there is considerable interindividual variation in phytoestrogen metabolism, which may determine its subsequent health effects. The effects of phytoestrogen metabolites such as equol, which is thought to be the most biologically active (28), can be captured only in studies using biomarker measurements. In addition, plasma sex hormone concentrations are regulated by a network of different enzymes. In a cross-sectional study in a large group of postmenopausal women, we found that single nucleotide polymorphisms (SNPs) in genes encoding aromatase (CYP19) and SHBG were significantly associated with differences in plasma concentrations of estradiol and SHBG (29). It is possible that these SNPs could also affect androgen concentrations in men. No previous study had considered whether such genetic polymorphisms may modify the association between phytoestrogens and circulating androgen and SHBG concentrations. Thus, we conducted a cross-sectional study of 267 men in the Norfolk arm of the European Prospective Investigation into Cancer and Nutrition (EPIC) cohort with 2 aims: first, to investigate the association between phytoestrogen exposure (measured from diet, urine, and serum) and plasma concentrations of androstanediol glucuronide, testosterone, Free Androgen Index (FAI), and SHBG; and second, whether the association may be modified by polymorphisms in the CYP19 and SHBG genes. SUBJECTS AND METHODS Study subjects. Subjects in this study were 267 men from a nested prospective case control study on prostate cancer in the EPIC-Norfolk cohort (unpublished data). In EPIC-Norfolk, men and women aged 45–75 y, resident in Norfolk, UK, were recruited in 1993–1997 using general practice age sex registers; 30,452 completed a health questionnaire and gave written informed consent. Permission for the study was obtained from The Norfolk and Norwich Hospital Ethics Committee. A total of 25,630 individuals attended a medical examination, and gave blood and an untimed spot urine sample at a clinic (30). All subjects were healthy at the time of recruitment and sample collection. The urine samples were stored at ⫺20°C until analyzed for creatinine and phytoestrogens. The serum and plasma samples were stored at ⫺40°C until analyzed for phytoestrogens and sex hormones, respectively. Dietary data. Dietary data were obtained using 7-d food diaries. These were given out to the subjects at the medical examination after instruction; the diaries were completed and returned by post (93% compliance). Details were described elsewhere (31). A total of 265 men completed the food diaries. Information from the 7-d dietary diaries was used to calculate dietary isoflavone intakes using a food composition database based on daidzein and genistein concentrations measured in 300 commonly eaten foods. Details on the sampling of foods, and analysis of daidzein and genistein and their contents in different foods were reported elsewhere (32–35). The isoflavone content of foods gathered from a literature search of published values was also incorporated into the food composition database for use in the analysis. This food composition database of isoflavones forms part of the VENUS (Vegetal Estrogens in Nutrition and the Skeleton) database, a regional food composition database established to facilitate the estimation of exposure levels to phytoestrogens in 4 Euro-
2681
pean countries, including Italy, the Netherlands, Ireland, and the United Kingdom (36,37). Urinary analysis. Urinary creatinine concentrations were measured based on a kinetic modification of the Jaffe´ reaction using the Roche Reagent for creatinine on a Roche Cobas Mira Plus chemistry analyser (Roche Products). Spot urine samples were analyzed for 3 isoflavones (daidzein, genistein, and glycitein), 2 metabolites of daidzein [O-desmethylangolensin (O-DMA) and equol] and 2 lignans (enterodiol and enterolactone). Triply 13C-labeled standards in methanol were added to a 200-L sample, and conjugates were hydrolyzed to aglycones by enzymatic hydrolysis. The aglycones were extracted on SPE cartridges and derivatized to trimethylsilyl derivatives for analysis using isotope dilution GC-MS. Details and information on quality assurance and methodology were reported elsewhere (38). Limits of detection ranged from 1.2 g/L (enterodiol) to 5.3 g/L (enterolactone). The mean intra-assay CV ranged from 1.8% (equol) to 6.5% (glycitein). The mean interassay CV for all analytes was ⬍9% except for O-DMA (20.2%) and glycitein (26.5%), both of which did not have a corresponding triply 13C-labeled standard at the time of analysis. Serum phytoestrogen analysis. Serum samples were analyzed for 3 isoflavones (daidzein, genistein, and glycitein), 2 metabolites of daidzein (O-DMA and equol), and 2 lignans (enterodiol and enterolactone). Triply 13C-labeled standards in methanol were added to a 200-L sample and conjugates were hydrolyzed to aglycones by enzymatic hydrolysis. The aglycones were extracted on SPE cartridges, then dried under nitrogen and redissolved in 40% methanol for analysis using isotope dilution LC-MS-MS. Details and information on quality assurance and methodology were reported elsewhere (39). Statistically calculated limits of detection range from 82 ng/L (daidzein) to 222 ng/L (equol). The mean intra-assay CV ranged from 2.8% (enterolactone) to 5.7% (glycitein). The mean interassay CV ranged from 3.0% (genistein) to 4.4% (O-DMA). Plasma sex hormone analyses. Plasma samples were analyzed for testosterone, androstanediol glucuronide, and SHBG. Testosterone and SHBG were measured by chemiluminescent immunoassays using an Immulite autoanalyzer from Diagnostic Products Corporation. The testosterone assay had a detection limit of 0.3 nmol/L and the intraand interassay CVs were 7.1 and 12.0%, respectively. The detection limit for SHBG was 0.2 nmol/L and the intra- and interassay CVs were 4.6 and 5.7%, respectively. Androstanediol glucuronide was measured using a RIA kit (DSL-6000) from Diagnostic Systems Laboratories. The detection limit was 0.2 nmol/L and the intra- and interassay CVs were 3.2 and 6.5% respectively. Genotype analyses. We genotyped one SNP in the CYP19 gene and 2 SNPs in the SHBG gene. All genotyping was carried out using end-point Taqman assays (Applied Biosystems) in 384-well arrays, which included blank wells as negative controls. Assays were run on MJ Tetrad thermal cyclers (Genetics Research Instrumentation), and genotypes were subsequently read on a 7900 Sequence Detector (Applied Biosystems) according to manufacturer’s instructions. Details on primer and probe sequences and assay temperatures were published elsewhere (29). An automated robotic high-throughput system in a low-volume 384-well format was used. Each assay was tested on a specific test set of 96 DNA samples (80 unique, 14 duplicates, and 2 no template controls), before use in the main study; all gave clear clustering and showed 100% concordance in the duplicates. Genotype data were obtained on 233 men for the SHBG 5⬘ untranslated region (UTR) G-A polymorphism; 232 men for the CYP19 3⬘UTR T-C (rs10046) polymorphism; and 230 men for the SHBG D356N (rs6259) polymorphism. 2 tests were used to determine that the genotype distributions conformed to those expected under Hardy-Weinberg equilibrium. Data analysis. Statistical analyses were performed using SPSS software version 12.0. Urinary excretion of phytoestrogens was expressed as nmol/mmol urinary creatinine. The Kruskal Wallis test was used to compare differences in dietary, urinary, and serum phytoestrogen concentrations among subjects with different genotypes for each SNP. FAI was calculated using the formula FAI ⫽ testosterone ⫻ 100/SHBG. All dietary, urinary, and serum phytoestrogen data and plasma androstanediol glucuronide, testosterone, SHBG, and FAI data were skewed. Therefore data were logarithmically transformed to
LOW ET AL.
2682
TABLE 1 Dietary intake, urinary excretion, serum concentrations of phytoestrogens, and plasma sex hormone concentrations of study subjects1 Mean
Range2
240.7 (223.8, 258.8) 302.5 (281.4, 325.3)
31.2–2880.4 36.2–5034.7
59.4 (52.0, 68.5) 32.6 (28.5, 37.0) 12.0 (10.6, 13.7) 12.4 (10.1, 15.1) 5.0 (4.1, 5.8) 19.5 (17.2, 21.9) 227.6 (199.7, 260.1)
ND–1128.0 ND–1866.3 ND–130.2 ND–98.4 ND–351.7 ND–743.7 ND–2948.3
9.4 (8.3, 11.0) 16.7 (14.4, 19.6) 0.5 (0.4, 0.6) 0.9 (0.7, 1.2) 0.9 (0.8, 1.1) 1.0 (0.9, 1.1) 12.8 (11.1, 14.8)
0.4–140.9 ND–1064.8 ND–18.0 ND–14.7 ND–55.0 ND–38.1 ND–162.1
12.2 (11.5, 13.0) 15.2 (14.6, 15.8) 42.3 (40.3, 44.5) 36.0 (34.5, 37.4)
1.4–53.0 4.4–38.2 2.0–131.0 6.6–220.0
n Dietary intake, g/d Daidzein Genistein Urinary excretion, nmol/ mmol Cr Daidzein Genistein Glycitein O-DMA Equol Enterodiol Enterolactone Serum concentrations, nmol/L Daidzein Genistein Glycitein O-DMA Equol Enterodiol Enterolactone Plasma concentrations Androstanediol glucuronide, nmol/L Testosterone, nmol/L SHBG, nmol/L FAI 1 2
265
267
267
264 266 266 266
Values are geometric means (95% CI). Based on untransformed values. ND, not detected.
obtain approximately normal frequency distributions and all subsequent statistical tests were performed on log-transformed data. Oneway ANOVA was used to compare differences in plasma androstanediol glucuronide, testosterone, SHBG, and FAI among subjects with different genotypes for each SNP. Trend tests were used to assess any linear associations between plasma testosterone, androstanediol glucuronide, SHBG, and FAI across common homozygotes, heterozygotes, and rare homozygotes for the respective SNPs. Hierarchical
multiple regression and partial correlations were used to assess the degree of association between dietary, urinary, and serum phytoestrogens and plasma androstanediol glucuronide, testosterone, SHBG, and FAI, controlling for age, weight, height, and time of venipuncture. Adjusting for subjects who subsequently developed prostate cancer did not affect the results. Hence, the results were presented for all men combined. To assess potential effect modification by polymorphisms in CYP19 and SHBG genes, we repeated the regression analyses, stratified according to men with different genotypes for each SNP. Interaction was then assessed by comparing the regression coefficients for men with different genotypes using the method described by Matthews and Altman (40,41). All P-values were 2-sided and values ⬍ 0.05 were considered significant.
RESULTS The subjects were (mean ⫾ SD) 67.5 ⫾ 6.3 y old. Dietary intakes of daidzein and genistein were low, with mean intakes of 240 and 300 g/d, respectively. Enterolactone was the main metabolite in urine, with mean concentrations several times higher than those of the other phytoestrogens (Table 1). Of 267 subjects, 221 (82.8%) had detectable concentrations of equol in their serum (ⱖ0.45 nmol/L) whereas 163 (61.0%) had detectable concentrations of urinary equol (ⱖ7.85 nmol/L). Serum equol was detected in all of the 163 subjects with detectable urinary equol (data not shown). Dietary, urinary, and serum phytoestrogen concentrations did not differ among men with different genotypes for each SNP except for urinary O-DMA. Urinary O-DMA was lower in men with the AA genotype (mean rank ⫽ 82.0) compared with the AG (mean rank ⫽ 114.0) and GG (mean rank ⫽ 121.9) genotypes for SHBG 5⬘ UTR G-A polymorphism (P ⫽ 0.043). Men with the AA genotype for the SHBG 5⬘ UTR G-A polymorphism had higher plasma testosterone (Pheterogeneity ⫽ 0.01; Ptrend ⫽ 0.005), higher SHBG (Pheterogeneity⬍0.001; Ptrend ⬍0.001), but lower FAI (Pheterogeneity ⫽ 0.035; Ptrend ⫽ 0.011) compared with men with the AG and GG genotypes. There was a significant trend toward higher testosterone concentrations in men carrying the N allele for the SHBG D356N polymorphism (Ptrend ⫽ 0.031). Androstanediol glucuronide, testosterone, SHBG, and FAI concentrations did not differ among men with different genotypes for the polymorphisms investigated (Table 2).
TABLE 2 Association between SNPs in genes involved in sex hormone metabolism and activity, and plasma sex hormone concentrations SNP
Genotype
n
% of total
Androstanediol glucuronide1
Testosterone1
SHBG1
FAI1
nmol/L CYP19 3⬘UTR T-C (rs10046) SHBG 5⬘UTR G-A SHBG D356N (rs6259)
CC CT TT AA AG GG
44 125 63 12 84 137
19.0 53.9 27.2 5.2 36.1 58.8
11.1 (9.4, 13.1) 13.1 (12.0, 14.3) 11.8 (10.5, 13.2) 15.1 (11.6, 19.8) 12.4 (11.1, 14.0) 12.0 (11.1, 13.0)
15.4 (14.0, 17.0) 15.3 (14.3, 16.3) 15.2 (14.1, 16.4) 19.5 (16.1, 23.6)* 15.8 (14.7, 17.0) 14.7 (13.9, 15.5)
40.1 (33.8, 47.6) 42.7 (39.9, 45.6) 44.6 (40.7, 48.8) 60.9 (47.2, 78.7)† 47.0 (43.5, 50.7) 39.0 (36.3, 42.0)
38.5 (34.0, 43.6) 35.8 (33.7, 38.1) 34.2 (31.7, 36.8) 32.0 (26.0, 39.4) 33.7 (31.2, 36.4) 37.6 (35.6, 39.8)
NN DN DD
5 52 173
2.2 22.6 75.2
14.3 (10.6, 19.2) 11.4 (10.0, 13.0) 12.6 (11.7, 13.6)
19.4 (15.3, 24.6)‡ 16.3 (15.0, 17.8) 15.0 (14.3, 15.8)
47.6 (38.9, 58.3) 45.7 (41.7, 50.0) 42.4 (40.0, 44.9)
40.7 (31.0, 53.6) 35.7 (33.1, 38.5) 35.5 (33.7, 37.4)
1 Values are geometric means (95% CI). * Different from the AG and GG genotypes; Pheterogeneity ⫽ 0.01; Ptrend ⫽ 0.005. † Different from the AG and GG genotypes: Pheterogeneity ⬍ 0.001; Ptrend ⬍ 0.001. Different from the AG and GG genotypes; Pheterogeneity ⫽ 0.035; Ptrend ⫽ 0.011. ‡ Significant trend across the NN, DN and DD genotypes; Pheterogeneity ⫽ 0.086; Ptrend ⫽ 0.031.
PHYTOESTROGEN GENE ASSOCIATIONS WITH PLASMA ANDROGENS
2683
TABLE 3 Association between dietary, urinary, and serum phytoestrogens and plasma androstanediol glucuronide and FAI levels in 266 EPIC-Norfolk men Association with plasma androstanediol glucuronide (nmol/L) Partial correlation coefficient2
R2-change1 Dietary intake Daidzein Genistein Urinary concentrations Daidzein Genistein Glycitein O-DMA Equol Enterodiol Enterolactone Serum concentrations Daidzein Genistein Glycitein O-DMA Equol Enterodiol Enterolactone
Association with FAI
P-value
R2-change1
Partial correlation coefficient2
P-value
0.000 0.000
0.010 0.019
0.874 0.767
0.000 0.000
⫺0.003 ⫺0.013
0.968 0.842
0.000 0.000 0.001 0.000 0.002 0.002 0.015
0.003 0.006 ⫺0.034 0.000 0.046 ⫺0.043 0.127
0.957 0.926 0.585 0.999 0.462 0.495 0.043
0.000 0.001 0.001 0.005 0.000 0.001 0.012
0.012 ⫺0.032 0.037 0.075 0.012 0.027 0.115
0.848 0.607 0.560 0.231 0.851 0.662 0.067
0.006 0.004 0.001 0.004 0.010 0.000 0.027
0.083 0.066 0.040 0.066 0.102 0.011 0.172
0.187 0.292 0.526 0.295 0.105 0.864 0.006
0.000 0.001 0.005 0.001 0.001 0.001 0.022
⫺0.004 ⫺0.036 0.074 0.028 ⫺0.034 0.041 0.158
0.947 0.569 0.235 0.652 0.587 0.514 0.011
1 Hierarchical linear regression model for log-transformed data of the particular phytoestrogen exposure and plasma sex hormone, adjusted for age, weight, height, and time of day at venipuncture. R2-change represents the proportion of variance explained uniquely by the particular phytoestrogen variable. 2 Pearson partial correlation for log-transformed data, adjusted for age, weight, height, and time of day at venipuncture.
After adjustment for age, weight, height, and time of venipuncture, enterolactone was positively associated with plasma androstanediol glucuronide concentrations and FAI, with enterolactone explaining up to 2.7% of the variance of these sex hormone concentrations (Table 3). There was no significant association between any other phytoestrogens and androstanediol glucuronide, testosterone, SHBG, and FAI. To investigate any potential effect modification by SNPs in CYP19 and SHBG genes, we studied the associations between phytoestrogen and sex hormone concentrations stratified by genotype for each of the 3 SNPs. For the 2 SHBG SNPs, the phytoestrogen-sex hormone associations were similar among men with different genotypes. However, for the CYP19 3⬘UTR T-C SNP, the associations between 2 phytoestrogen
metabolites (enterolactone and equol) and androstanediol glucuronide, testosterone, and FAI differed significantly among men with different genotypes. Both urinary and serum equol correlated positively with plasma androstanediol glucuronide concentrations and explained ⬃10% of the variance in androstanediol glucuronide among men with the CC genotype but not in men with the CT or TT genotypes (Table 4). Both urinary and serum equol showed clear positive associations with plasma testosterone and FAI among men with the TT genotype, but this was not the case for the CC or CT genotypes (Table 5). Similarly, enterolactone was positively associated with plasma testosterone concentration among men with the CC genotype but not the CT or TT genotypes. However, this enterolactone
TABLE 4 Association between equol and plasma androstanediol glucuronide concentrations among men with different genotypes for CYP19 3⬘UTR T-C polymorphism CC genotype (n ⫽ 43)
 (SE)
R -change
Partial correlation coefficient2
0.186 (0.092) 0.393 (0.173)
0.091 0.112
0.327 0.364
1
Urinary equol Serum equol
2
1
CT or TT genotype (n ⫽ 184)
P-value
 (SE)
0.052 0.029
⫺0.023 (0.047) ⫺0.032 (0.154)
1
R -change
Partial correlation coefficient2
P-value
P-value for heterogeneity3
0.001 0.000
⫺0.037 ⫺0.016
0.621 0.834
0.043 0.067
2
1
1 Hierarchical linear regression model for log-transformed data of the particular phytoestrogen exposure and plasma sex hormone, adjusted for age, weight, height and time of day at venipuncture. R2-change represents the proportion of variance explained uniquely by the particular phytoestrogen variable. 2 Pearson partial correlation for log-transformed data, adjusted for age, weight, height, and time of day at venipuncture. 3 P-value for test of heterogeneity between the  coefficients for the CC vs. the CT or TT genotypes.
LOW ET AL.
2684
TABLE 5 Association between phytoestrogens and plasma testosterone and FAI among men with different genotypes for CYP19 3⬘UTR T-C polymorphism TT genotype (n ⫽ 62)
Testosterone Urinary equol Serum equol Urinary enterolactone Serum enterolactone FAI Urinary equol Serum equol Urinary enterolactone Serum enterolactone
CT or CC genotype (n ⫽ 169)
Partial correlation coefficient2 P-value
 (SE)1
R2-change1
0.102 (0.040) 0.278 (0.114) 0.087 (0.024) 0.064 (0.024)
0.081 0.075 0.149 0.084
0.332 0.318 0.449 0.337
0.096 (0.042) 0.348 (0.116) 0.054 (0.027) 0.048 (0.026)
0.073 0.119 0.075 0.049
0.297 0.380 0.266 0.245
Partial correlation P-value for coefficient2 P-value heterogeneity3
 (SE)1
R2-change1
0.013 0.018 0.001 0.012
⫺0.023 (0.036) ⫺0.032 (0.088) ⫺0.008 (0.026) 0.018 (0.027)
0.002 0.001 0.001 0.003
⫺0.051 ⫺0.029 ⫺0.025 0.053
0.522 0.717 0.758 0.504
0.032 0.031 0.007 0.203
0.027 0.004 0.049 0.071
⫺0.019 (0.035) ⫺0.146 (0.086) 0.023 (0.025) 0.052 (0.026)
0.002 0.015 0.004 0.021
⫺0.042 ⫺0.134 0.068 0.157
0.595 0.091 0.390 0.047
0.035 0.0006 0.385 0.913
1 Hierarchical linear regression model for log-transformed data of the particular phytoestrogen exposure and plasma sex hormone, adjusted for age, weight, height, and time of day at venipuncture. R2-change represents the proportion of variance explained uniquely by the particular phytoestrogen variable. 2 Pearson partial correlation for log-transformed data, adjusted for age, weight, and height. 3 P-value for test of heterogeneity between the  coefficients for TT vs. CT or CC genotypes.
association diminished when FAI was studied instead of plasma testosterone (Table 5). DISCUSSION In this cross-sectional study of 267 men, we found significant differences in SHBG concentrations among men with different genotypes for the SHBG 5⬘UTR G-A polymorphism. Men with the AA genotype had the highest testosterone, highest SHBG concentration, and lowest FAI. These findings in men have not been reported previously. But these observations are consistent with that from a recent study of 1975 postmenopausal women in EPIC-Norfolk in which we found that women with the AA genotype had the highest SHBG and lowest estradiol:SHBG ratio (29). The 5⬘UTR change might affect the rate of production of SHBG, although we cannot rule out that it may not be directly functional itself but may be in linkage disequilibrium with another functional change. On the other hand, the SHBG D356N polymorphism is thought to affect the rate of clearance of SHBG from the circulation (42). Among postmenopausal women, the SHBG D356N polymorphism was also associated with significant differences in SHBG concentrations (29). However, we did not observe such differences in men. This may be due to the small number of rare NN homozygotes (n ⫽ 5) in this study. The geometric means of the dietary isoflavone intakes, urinary, and serum phytoestrogen concentrations of the men in this study were comparable to those found in women in EPIC-Norfolk (43). The mean intake of isoflavones was 540 g/d. This is most likely an underestimate because recent work by Clarke et al. (44) suggested that the estimated daily isoflavone intake for an average adult in United Kingdom was 3 mg/d, based on chemical analysis of food group composites from the 1998 UK Total Diet Study. The underestimation of isoflavone intake is due largely to the use of soy in processed foods that have not yet been captured in existing food composition databases (45). Nevertheless, these isoflavone intake levels in UK are still several times lower than that reported in Asian populations (46,47). The low habitual soy consumption in these EPIC-Norfolk men is reflected in their low serum
isoflavone concentrations. The sum of the mean serum concentration of daidzein and genistein was 26 nmol/L. This serum concentration is similar to that reported in 96 postmenopausal women in the United States (19 nmol/L) (48) but is much lower than the serum concentrations reported in 215 Japanese men and women (595 nmol/L) (47). Although phytoestrogens were reported to stimulate SHBG concentrations (12–15), we did not observe any association between any of the phytoestrogens and SHBG concentrations in this study. This null finding is unlikely to be due to the low phytoestrogen exposure in our population because 2 previous cross-sectional studies, which included high soy consumers, also did not find any association between soy intake and SHBG concentrations in men (26,27). Our finding adds to the body of evidence supporting a lack of association between phytoestrogen exposure and SHBG concentrations in men. At least 10 intervention studies (16 –25) and 2 cross-sectional studies (26,27) investigated the effect of phytoestrogens on sex hormone concentrations in men; 7 of the 10 interventions and both cross-sectional studies reported no effect on testosterone levels, whereas 3 intervention studies (16,18,24) reported decreased testosterone or FAI with phytoestrogen supplementation. All studies used either soy intake or dietary isoflavone intake as an assessment of exposure. None investigated the effect of lignans even though lignans are likely to be the more important group of phytoestrogens in Western populations, which have low habitual soy consumption. In addition, none of the previous studies were able to examine phytoestrogen metabolites. The main strength of this study lies in our comprehensive assessment of phytoestrogen exposure, utilizing dietary, urinary, and serum measures of phytoestrogens. This enabled us to study not only the effects of dietary phytoestrogens but also the effects of phytoestrogen metabolites such as equol and enterolactone, which are more biologically active than their precursors. We found significant positive associations between urinary and serum equol and plasma androstanediol glucuronide among men with the CC genotype and between equol and testosterone and FAI among men with the TT genotype
PHYTOESTROGEN GENE ASSOCIATIONS WITH PLASMA ANDROGENS
for the CYP19 3⬘UTR T-C polymorphism. These observations are suggestive of phytoestrogen-gene interaction. CYP19 codes for the aromatase enzyme, which catalyzes the irreversible conversion of androgens to estrogens. The haplotype block structure and tagging SNPs of the CYP19 gene were published (49). The entire coding region of CYP19 lies within a single block that can be divided into 4 common haplotypes in the U.S. Caucasian population. The T allele examined in this study uniquely tags the most common haplotype, whereas all 3 remaining haplotypes carry the C allele (29). Hence, men with the CC genotype may be a potentially heterogeneous group and any observed health effects would be harder to interpret compared with men with TT genotype. The T allele of the CYP19 3⬘UTR T-C polymorphism is associated with higher mRNA levels and thus higher aromatase activity (50) and has been associated with higher estradiol levels in postmenopausal women (29). In this study, men with the TT genotype appeared to be particularly responsive to equol. The reason for this is unclear. Although speculative, the explanation may lie in the complex feedback to the hypothalamus-pituitary axis on testosterone production that is due in part to aromatization of testosterone to estrogens in the hypothalamus. We found significant positive associations between enterolactone and androstanediol glucuronide and FAI. The positive association between enterolactone and androstanediol did not differ significantly between men with different genotypes in the stratified analyses. This suggests that this positive association may not involve interaction with any of the investigated polymorphisms. In contrast, the positive association between enterolactone and testosterone was modified by the CYP19 3⬘UTR T-C polymorphism (Table 4). The effect of the modification diminished when FAI was studied instead of testosterone. This may suggest that any effects of enterolactone-gene interaction on testosterone levels may be compensated by corresponding changes in SHBG levels. Therefore FAI remains unaffected so that the overall effect of the interaction would be less physiologically important. In this study, we used 3 different methods of assessing phytoestrogen exposure. In this group of men, urinary phytoestrogen concentrations correlated strongly with their serum concentration. Correlation coefficients ranged from 0.63 (glycitein) to 0.88 (daidzein) with P-values ⬍ 0.001. Dietary daidzein and genistein intakes also correlated significantly with their respective urinary and serum measurements. But correlation was weaker, with correlation coefficients ranging from 0.13 (serum daidzein vs. dietary daidzein) to 0.21 (urinary genistein vs. dietary genistein). The high degree of correlation between urine and serum phytoestrogens would account for the similar findings obtained using these 2 measures. Results obtained from dietary measures were less similar to those from urinary and serum biomarkers due to the weaker correlations. Nonetheless, daidzein and genistein were not significantly associated with any of the sex hormones in this study regardless of the measure used. The significant associations between phytoestrogens and plasma androgens in this study all tended to be positive. If higher androgen levels are associated with higher prostate cancer risk, then it would seem that the androgen-increasing effects of phytoestrogens may be adverse. However, epidemiologic and animal studies have generally found protective effects of phytoestrogens (1,2). The relation between circulating androgen levels and prostate cancer risk is still unclear because most prospective studies did not find any significant association between higher androgen levels and increased prostate cancer risk (51–53). Circulating androgens can be converted to estrogen, which has also been implicated in
2685
increasing prostate cancer risk (51,54). Until more is known about the role of different sex hormones in prostate cancer development, it is hard to predict how the androgen-increasing properties of phytoestrogens observed in this study would relate to cancer risk. Because of the known diurnal variation in androgen measurements, we adjusted for the time of day at venipuncture. Smoking was not associated with hormone levels in our subjects and was excluded from the model. But we cannot rule out that our results may be due to residual confounding from other unknown factors. We also cannot rule out that the positive results could have arisen by chance due to multiple testing in our stratified analysis. But we assessed interaction by comparing differences in effect sizes between subgroups as suggested by Matthews and Altman (40,41) rather than simply relying on the P-value in an attempt to reduce false positives. In this first study on phytoestrogen-gene associations in men, we found that enterolactone and equol were positively associated with plasma androgen levels, and interactions with the CYP19 gene may be involved. However, this study is exploratory and investigated polymorphisms in only 2 genes and plasma levels of androstanediol glucuronide, testosterone and SHBG in a relatively small group of men. More comprehensive studies will be required to investigate these phytoestrogen-gene interactions further. Such studies could measure additional hormones (e.g., androstenedione, estrone, estradiol), include other genes (e.g., CYP17, HSD17B1, SRD5A2) that are involved in the androgen metabolism pathway, and be conducted among larger study sets so as to facilitate investigation of rarer polymorphisms. Larger case control studies investigating the effects of phytoestrogens on prostate cancer risk in men with different genotypes would also be useful to determine whether phytoestrogen-gene interactions could result in differential cancer risks. ACKNOWLEDGMENT We thank Dr. Nigel Botting from University of St Andrews, UK for supplying triply 13C-labeled phytoestrogen standards for urinary and serum analyses of phytoestrogens.
LITERATURE CITED 1. Messina MJ. Emerging evidence on the role of soy in reducing prostate cancer risk. Nutr Rev. 2003;61:117–31. 2. Magee PJ, Rowland IR. Phyto-oestrogens, their mechanism of action: current evidence for a role in breast and prostate cancer. Br J Nutr. 2004;91:513– 31. 3. Committee on Toxicity of Chemicals in Food/Consumer Products and the Environment. COT Report—Phytoestrogens and Health. London: Food Standards Agency; 2003. 4. Tammela T. Endocrine treatment of prostate cancer. J Steroid Biochem Mol Biol. 2004 Nov;92(4):287–95. 5. Kirk CJ, Harris RM, Wood DM, Waring RH, Hughes PJ. Do dietary phytoestrogens influence susceptibility to hormone-dependent cancer by disrupting the metabolism of endogenous oestrogens? Biochem Soc Trans. 2001 May;29(Pt 2):209 –16. 6. Evans BA, Griffiths K, Morton MS. Inhibition of 5 alpha-reductase in genital skin fibroblasts and prostate tissue by dietary lignans and isoflavonoids. J Endocrinol. 1995 Nov;147(2):295–302. 7. Makela S, Poutanen M, Lehtimaki J, Kostian ML, Santti R, Vihko R. Estrogen-specific 17 beta-hydroxysteroid oxidoreductase type 1 (E.C. 1.1.1.62) as a possible target for the action of phytoestrogens. Proc Soc Exp Biol Med. 1995;208:51–9. 8. Makela S, Poutanen M, Kostian ML, Lehtimaki N, Strauss L, Santti R, Vihko R. Inhibition of 17beta-hydroxysteroid oxidoreductase by flavonoids in breast and prostate cancer cells. Proc Soc Exp Biol Med. 1998;217:310 – 6. 9. Krazeisen A, Breitling R, Moller G, Adamski J. Phytoestrogens inhibit human 17beta-hydroxysteroid dehydrogenase type 5. Mol Cell Endocrinol. 2001 Jan 22;171(1–2):151– 62. 10. Adlercreutz H, Bannwart C, Wahala K, Makela T, Brunow G, Hase T, Arosemena PJ, Kellis JT Jr, Vickery LE. Inhibition of human aromatase by mammalian lignans and isoflavonoid phytoestrogens. J Steroid Biochem Mol Biol. 1993;44:147–53
2686
LOW ET AL.
11. Wang C, Makela T, Hase T, Adlercreutz H, Kurzer MS. Lignans and flavonoids inhibit aromatase enzyme in human preadipocytes. J Steroid Biochem Mol Biol. 1994 Aug;50(3– 4):205–12. 12. Adlercreutz H, Mousavi Y, Clark J, Hockerstedt K, Hamalainen E, Wahala K, Makela T, Hase T. Dietary phytoestrogens and cancer: in vitro and in vivo studies. J Steroid Biochem Mol Biol. 1992 Mar;41(3– 8):331–7. 13. Mousavi Y, Adlercreutz H. Genistein is an effective stimulator of sex hormone-binding globulin production in hepatocarcinoma human liver cancer cells and suppresses proliferation of these cells in culture. Steroids. 1993;58: 301– 4. 14. Loukovaara M, Carson M, Palotie A, Adlercreutz H. Regulation of sex hormone-binding globulin production by isoflavonoids and patterns of isoflavonoid conjugation in HepG2 cell cultures. Steroids. 1995;60:656 – 61. 15. Adlercreutz H, Ho¨ckerstedt K, Bannwart C, Bloigu S, Ha¨ma¨la¨inen E, Fotsis T, Ollus A. Effect of dietary components, including lignans and phytoestrogens, on enterohepatic circulation and liver metabolism of estrogens and on sex hormone binding globulin (SHBG). J Steroid Biochem Mol Biol. 1987;27:1135– 44. 16. Dillingham BL, McVeigh BL, Lampe JW, Duncan AM. Soy protein isolates of varying isoflavone content exert minor effects on serum reproductive hormones in healthy young men. J Nutr. 2005;135:584 –91. 17. Goldin BR, Brauner E, Adlercreutz H, Ausman LM, Lichtenstein AH. Hormonal response to diets high in soy or animal protein without and with isoflavones in moderately hypercholesterolemic subjects. Nutr Cancer. 2005; 51(1):1– 6. 18. Gardner-Thorpe D, O’Hagen C, Young I, Lewis SJ. Dietary supplements of soya flour lower serum testosterone concentrations and improve markers of oxidative stress in men. Eur J Clin Nutr. 2003;57:100 – 6. 19. Lewis JG, Morris JC, Clark BM, Elder PA. The effect of isoflavone extract ingestion, as Trinovin, on plasma steroids in normal men. Steroids. 2002 Jan; 67(1):25–9. 20. Mitchell JH, Cawood E, Kinniburgh D, Provan A, Collins AR, Irvine DS. Effect of a phytoestrogen food supplement on reproductive health in normal males. Clin Sci (Lond). 2001;100:613– 8. 21. Teede HJ, Dalais FS, Kotsopoulos D, Liang YL, Davis S, McGrath BP. Dietary soy has both beneficial and potentially adverse cardiovascular effects: a placebo-controlled study in men and postmenopausal women. J Clin Endocrinol Metab. 2001 Jul;86(7):3053– 60. 22. Nagata C, Takatsuka N, Shimizu H, Hayashi H, Akamatsu T, Murase K. Effect of soymilk consumption on serum estrogen and androgen concentrations in Japanese men. Cancer Epidemiol Biomarkers Prev. 2001;10:179 – 84. 23. Higashi K, Abata S, Iwamoto N, Ogura M, Yamashita T, Ishikawa O, Ohslzu F, Nakamura H. Effects of soy protein on levels of remnant-like particles cholesterol and vitamin E in healthy men. J Nutr Sci Vitaminol (Tokyo). 2001 Aug;47(4):283– 8. 24. Habito RC, Montalto J, Leslie E, Ball MJ. Effects of replacing meat with soyabean in the diet on sex hormone concentrations in healthy adult males. Br J Nutr. 2000 Oct;84(4):557– 63. 25. Mackey R, Ekangaki A, Eden JA. The effects of soy protein in women and men with elevated plasma lipids. Biofactors. 2000;12(1– 4):251–7. 26. Allen NE, Appleby PN, Davey GK, Key TJ. Soy milk intake in relation to serum sex hormone levels in British men. Nutr Cancer. 2001;41(1–2):41– 6. 27. Nagata C, Inaba S, Kawakami N, Kakizoe T, Shimizu H. Inverse association of soy product intake with serum androgen and estrogen concentrations in Japanese men. Nutr Cancer. 2000;36(1):14 – 8. 28. Setchell KD, Brown NM, Lydeking-Olsen E. The clinical importance of the metabolite equol-a clue to the effectiveness of soy and its isoflavones. J Nutr. 2002;132:3577– 84. 29. Dunning AM, Dowsett M, Healey CS, Tee L, Luben R, Folkerd E, Novik KL, Kelemen L, Ogata S, et al. Polymorphisms associated with circulating sex hormone levels in postmenopausal women. J Natl Cancer Inst. 2004;96(12):936 – 45. 30. Day N, Oakes S, Luben R, Khaw K-T, Bingham S, Welch A, Wareham N. EPIC-Norfolk: study design and characteristics of the cohort. Br J Cancer. 1999; 80(Suppl. 1):95–103. 31. Bingham SA, Welch AA, McTaggart A, Mulligan A, Runswick SA, Luben R, Oakes S, Khaw KT, Wareham N, Day NE. Nutritional methods in the European Prospective Investigation of Cancer in Norfolk. Public Health Nutr. 2001;4(3):847– 58.
32. Liggins J, Bluck LJC, Coward WA, Bingham SA. Extraction and Quantification of Daidzein and Genistein in Food. Anal Biochem. 1998;264:1–7. 33. Liggins J, Bluck LJC, Runswick S, Atkinson C, Coward WA, Bingham SA. Daidzein and genistein content of fruits and nuts. J Nutr Biochem. 2000;11:326 – 31. 34. Liggins J, Bluck LJC, Runswick S, Atkinson C, Coward WA, Bingham SA. Daidzein and genistein contents of vegetables. Br J Nutr. 2000;84:717–25. 35. Liggins J, Mulligan A, Runswick S, Bingham SA. Daidzein and genistein content of cereals. Eur J Clin Nutr. 2002;56:961– 6. 36. Kiely M, Faughnan MS, Wa¨ha¨la¨ K, Brants H, Mulligan A. Phyto-estrogen levels in foods: the design and construction of the VENUS database. Br J Nutr. 2003;89(Suppl. 1):S19 –23. 37. van Erp-Baart MA, Brants HA, Kiely M, Mulligan A, Turrini A, Sermoneta C, Kikkinen A, Valsta LM. Isoflavone intake in four different European countries: the VENUS approach. Br J Nutr. 2003;89(Suppl. 1):S25–30. 38. Grace PB, Taylor JI, Botting NP, Fryatt T, Oldfield MF, Bingham SA. Quantification of isoflavones and lignans in urine using gas chromatography/ mass spectrometry. Anal Biochem. 2003;315:114 –21. 39. Grace PB, Taylor JI, Botting NP, Fryatt T, Oldfield MF, Al-Maharik N, Bingham SA. Quantification of isoflavones and lignans in serum using isotope dilution liquid chromatography/ tandem mass spectrometry. Rapid Commun Mass Spectrom. 2003;17:1350 –7. 40. Matthews JN, Altman DG. Statistics notes. Interaction 2: Compare effect sizes not P values. Br Med J. 1996 Sep 28;313(7060):808. 41. Matthews JN, Altman DG. Interaction 3: how to examine heterogeneity. Br Med J. 1996 Oct 5;313(7061):862. 42. Cousin P, Dechaud H, Grenot C, Lejeune H, Pugeat M. Human variant sex hormone-binding globulin (SHBG) with an additional carbohydrate chain has a reduced clearance rate in rabbit. J Clin Endocrinol Metab. 1998;83:235– 40. 43. Low YL, Taylor JI, Grace PB, Dowsett M, Scollen S, Dunning AM, Mulligan AA, Welch AA, Luben RN, et al. Phytoestrogen exposure correlation with plasma estradiol in postmenopausal women in European Prospective Investigation of Cancer and Nutrition-Norfolk may involve diet-gene interactions. Cancer Epidemiol Biomarkers Prev. 2005;4:213–20. 44. Clarke DB, Lloyd AS. Dietary exposure estimates of isoflavones from the 1998 UK Total Diet Study. Food Addit Contam. 2004;21:305–16. 45. Clarke DB, Barnes KA, Lloyd AS. Determination of unusual soya and non-soya phytoestrogen sources in beer, fish products and other foods. Food Addit Contam. 2004;21:949 – 62. 46. Seow A, Shi CY, Franke AA, Hankin JH, Lee H-P, Yu MC. Isoflavonoid levels in spot urine are associated with frequency of dietary soy intake in a population-based sample of middle-aged and older Chinese in Singapore. Cancer Epidemiol Biomarkers Prev. 1998;7(2):135– 40. 47. Yamamoto S, Sobue T, Sasaki S, Kobayashi M, Arai Y, Uehara M, Adlercreutz H, Watanabe S, Takahashi T, et al. Validity and reproducibility of a self-administered food-frequency questionnaire to assess isoflavone intake in a Japanese population in comparison with dietary records and blood and urine isoflavones. J Nutr. 2001;131:2741–7. 48. Frankenfeld CL, Patterson RE, Horner NK, Neuhouser ML, Skor HE, Kalhorn TF, Howald WN, Lampe JW. Validation of a soy food-frequency questionnaire and evaluation of correlates of plasma isoflavone concentrations in postmenopausal women. Am J Clin Nutr. 2003;77:674 – 80. 49. Haiman CA, Stram DO, Pike MC, Kolonel LN, Burtt NP, Altshuler D, Hirschhorn J, Henderson BE. A comprehensive haplotype analysis of CYP19 and breast cancer risk: the Multiethnic Cohort. Hum Mol Genet. 2003 Oct 15;12(20): 2679 –92. 50. Kristensen VN, Harada N, Yoshimura N, Haraldsen E, Lonning PE, Erikstein B, Karesen R, Kristensen T, Borresen-Dale AL. Genetic variants of CYP19 (aromatase) and breast cancer risk. Oncogene. 2000 Mar;219(10):1329 –33. 51. Platz EA, Giovannucci E. The epidemiology of sex steroid hormones and their signaling and metabolic pathways in the etiology of prostate cancer. J Steroid Biochem Mol Biol. 2004 Nov;92(4):237–53. 52. Hsing AW, Reichardt JK, Stanczyk FZ. Hormones and prostate cancer: current perspectives and future directions. Prostate. 2002 Sep 1;52(3):213–35. 53. Bosland MC. The role of steroid hormones in prostate carcinogenesis. J Natl Cancer Inst Monogr. 2000(27):39 – 66. 54. Risbridger GP, Bianco JJ, Ellem SJ, McPherson SJ. Oestrogens and prostate cancer. Endocr Relat Cancer. 2003 Jun;10(2):187–91.
