Daidzein-metabolizing phenotypes in relation to serum hormones and ...

Cancer Causes Control (2008) 19:1085–1093 DOI 10.1007/s10552-008-9172-3

ORIGINAL PAPER

Daidzein-metabolizing phenotypes in relation to serum hormones and sex hormone binding globulin, and urinary estrogen metabolites in premenopausal women in the United States Charlotte Atkinson Æ Katherine M. Newton Æ Frank Z. Stanczyk Æ Kim C. Westerlind Æ Lin Li Æ Johanna W. Lampe

Received: 13 November 2007 / Accepted: 27 April 2008 / Published online: 14 May 2008 Ó Springer Science+Business Media B.V. 2008

Abstract Objective Blood and urine concentrations of hormones are implicated in the etiology of some cancers. Small studies have assessed relationships between production of the daidzein metabolites equol and O-desmethylangolensin (ODMA) and hormones, but findings are unclear. We evaluated relationships between daidzein-metabolizing phenotypes and follicular phase concentrations of estrogens, androgens, sex hormone binding globulin (SHBG), and urinary estrogen metabolites in premenopausal women. This work was supported by the National Institute of Health (R01CA97366 and U01CA63731). C. Atkinson L. Li Division of Public Health Sciences, Fred Hutchinson Cancer Research Center, Seattle, WA 98109, USA K. M. Newton J. W. Lampe Department of Epidemiology, University of Washington, Seattle, WA 98195, USA K. M. Newton Group Health Center for Health Studies, Seattle, WA 98101, USA F. Z. Stanczyk Keck School of Medicine, University of Southern California, Los Angeles, CA 90089, USA K. C. Westerlind Division of Endocrinology, Metabolism and Diabetes, University of Colorado Health Sciences Center, Denver, CO 80045, USA

Methods Two-hundred women collected a first-void urine sample after a 3-day soy challenge, and 191 and 193 provided fasting blood and spot urine samples, respectively, during days 5–9 of their menstrual cycle. Soy challenge urines were analyzed for isoflavones; serum was analyzed for estrogens, androgens, and SHBG; spot urines were analyzed for 2-hydroxyestrone and 16a-hydroxyestrone. Data were log-transformed and multiple regression analyses were conducted to assess relationships between daidzein-metabolizing phenotypes and hormones and SHBG. Data from 187 and 189 women were included in analyses of serum and urine hormones, respectively. Results 55 (27.5%) and 182 (91%) of the 200 women who provided a soy challenge urine sample were equoland ODMA-producers ([87.5 ng/ml urine), respectively. In unadjusted analyses, equol-producers (n = 52) had lower free testosterone than equol non-producers (n = 137, p = 0.02). In adjusted analyses, there were no differences between producers and non-producers of either daidzein metabolite. Conclusions In the absence of a soy intervention, we found no difference in serum or urine hormone concentrations between producers and non-producers of equol or ODMA. Keywords Androgens Equol Estrogens Isoflavones O-desmethylangolensin

Introduction J. W. Lampe (&) Division of Public Health Sciences, Fred Hutchinson Cancer Research Center, P.O. Box 19024, M4-B402, Seattle, WA 98109, USA e-mail: [email protected]

Circulating estrogens and androgens and urinary estrogen metabolites have been implicated in the etiology of certain cancers (e.g., breast and endometrial) and other conditions

123

1086

with high morbidity (e.g., uterine fibroids and endometriosis) arising in hormone-sensitive tissues [1–7]. As such, the determinants of circulating concentrations and urinary excretion of hormones may be associated with susceptibility to a number of hormone-dependent conditions. Soy is a rich source of isoflavones. Studies assessing the effects of soy or isoflavone consumption in relation to circulating concentrations of sex hormones in pre- and post-menopausal women have reported inconsistent findings [8, 9], potentially due to the failure to distinguish equol-producers from non-producers [10]. Equol and O-desmethylangolensin (ODMA) are produced by intestinal bacteria in approximately 30–50% and 80–90% of individuals, respectively [11–17]. Equol is more biologically active than daidzein in vitro [18], and equol and ODMA can inhibit enzymes involved in steroid hormone metabolism, such as aromatase, 5a-reductase, and 17bhydroxysteroid dehydrogenase [19–21]. Thus, it has been suggested that producers and non-producers of equol and ODMA may respond differentially to soy or isoflavone interventions [10, 18]. Given that intestinal bacteria can metabolize hormones [22–25], an alternative hypothesis is that daidzein-metabolizing phenotypes are markers of intestinal bacterial profiles that are themselves associated with hormone concentrations [18]. Findings from studies investigating associations between urine or blood concentrations of equol and breast cancer risk have been mixed [26–28]. A soy intervention study in premenopausal women reported that, irrespective of isoflavone dose, equol-producers (but not non-producers) had circulating concentrations of hormones that were likely to be associated with a reduced risk of breast cancer [29]. This suggested a potential role for the equol-producer phenotype that was independent of soy or isoflavone dose. However, this was not shown in a subsequent study in premenopausal women [30]. Few studies have examined the potential health effects associated with the production of ODMA. If the intestinal bacteria responsible for, or associated with, equol and ODMA production are involved in determining hormone concentrations, differences between producers and non-producers would be apparent irrespective of soy consumption. Furthermore, if such differences do exist, the daidzein-metabolizing phenotypes could ultimately be used as early markers of risk of hormonedependent conditions and may lead to strategies to alter colonic bacterial populations toward a more low-risk environment. Our objectives were to determine, in the absence of a soy intervention, relationships between daidzein-metabolizing phenotypes and serum estrogens, androgens, and SHBG, and urinary estrogen metabolites and their ratio in premenopausal women living in the United States.

123

Cancer Causes Control (2008) 19:1085–1093

Materials and methods Recruitment Recruitment of participants for this study has been described in detail elsewhere [31]. Briefly, women of ages 40–45 years who had undergone a screening mammogram in the previous 10 months at Group Health (GH) were identified from the GH Breast Cancer Screening Program [32]. An initial contact letter was mailed to potential participants, and followed up with a phone call to screen individuals for interest and eligibility. Eligibility criteria were established to include premenopausal women and to exclude women taking exogenous hormones and those who had taken antibiotics in the 3 months prior to participation in the study. For full details of exclusion criteria, see [31]. All study procedures were approved by the Institutional Review Boards of the Fred Hutchinson Cancer Research Center and GH, and all study participants provided written informed consent. Data and sample collection Two hundred and three women attended a study clinic visit. We aimed to schedule the clinic visit between days 5–9 of a woman’s menstrual cycle, and 198 women (98%) attended their clinic visit during this time frame. A health and demographics questionnaire was mailed to each participant prior to their appointment and they were asked to complete it and bring it to their clinic visit. During the clinic visit, the participants’ weight, height, and waist and hip circumferences were measured, and a fasting blood sample and spot urine sample was obtained. Blood tubes and spot urine samples were transferred into coolers with ice packs and transported to the Fred Hutchinson Cancer Research Center Specimen Processing Laboratory. Tubes were centrifuged at 3,005 or 3,280g for 10 min, usually within 2 h of collection. Serum was removed and stored at -70°C prior to analysis. An aliquot of plain urine was stored for creatinine analysis, and the remaining urine was supplemented with one part ascorbic acid solution (100 mM solution of USP ascorbic acid, using 17.6 g/l deionized water) to 30 parts urine to prevent oxidation of labile estrogen metabolites. Samples were stored at -70°C until analysis. Assessment of equol and ODMA producer phenotypes Women were phenotyped for equol- and ODMA-producer status using a soy challenge as described previously [31]. Briefly, women were asked to consume one soy bar (or one-third of a bag of soy nuts) on three consecutive days and to collect a first-void urine sample on the morning of


the fourth day. Urine samples were analyzed for isoflavonoids by gas chromatography–mass spectrometry (GC– MS) as described elsewhere [31], and equol- and ODMAproducers were defined as individuals with detectable urinary concentrations of equol (87.5 ng/ml, or 362 nmol/l) and ODMA (87.5 ng/ml, or 339 nmol/l). The intra-assay CVs for isoflavonoids in the quality-control sample, measured in duplicate for each batch, were \9%. The interassay CVs were \13%. Urinary creatinine concentrations were measured to ensure that urine samples were sufficiently concentrated ([80 mg/l; 0.71 mmol/l), as described previously [31]. Serum hormones and SHBG Serum samples were measured for estrone (E1), E1-sulfate (E1S), estradiol (E2), dehydroepiandrosterone (DHEA), DHEA-sulfate (DHEAS), androstenedione (A), testosterone (T), and sex hormone binding globulin (SHBG). After organic solvent extraction and Celite column partition chromatography, E1, E2, T, A, and DHEA were quantified by sensitive and specific radioimmunoassays [33, 34]. Chromatographic separation of the steroids was achieved by use of different concentrations of toluene in isooctane and ethyl acetate in isooctane. SHBG and DHEAS were quantified by chemiluminescent immunoassay on the Immulite Analyzer (Siemens Medical Solutions Diagnostics, Los Angeles, CA). E1S was quantified by a highly specific direct RIA (Diagnostic Systems Laboratories, Webster, TX). Concentrations of free and bioavailable (non-SHBGbound) E2 and T were calculated using the measured total E2 and total T, respectively, and SHBG concentrations and an assumed constant for albumin [35–37]. Two blinded quality control (QC) samples were included in each of five batches. The % coefficient of variation (CV) across all QC samples (n = 10) ranged from 6.2% for E2 to 17.2% for E1S. Mean intra-assay CVs (each CV was based on two QC samples per batch) ranged from 3.1% for SHBG to 14.9% for E1S. Urinary estrogen metabolites Concentrations of 2-hydroxyestrone (2-OHE1) and 16ahydroxyestrone (16a-OHE1) were measured in spot urine samples using a commercially available competitive, solid-phase enzyme-linked immunoassay (ESTRAMET, ImmunaCare Corp., Bethlehem, PA) and the ratio of 2/16 computed. Urine samples were thawed and brought to room temperature before analysis. Based on measured creatinine values, samples were diluted 1:2 or 1:4 prior to analysis with manufacturer-supplied diluent. In urine, 2-OHE1 and 16a-OHE1 are found in the glucuronide conjugate form and require removal of the sugar moiety before

1087

recognition by the monoclonal antibodies. Samples were deconjugated with b-glucoronidase and arylsulphatase, and then neutralized. Samples were incubated for 3 h at room temperature then kinetically read every 2 min for 20 min using a Thermomax Microplate Reader (Molecular Devices, Sunnyvale, CA). Estrogen metabolite values were determined from a calibration curve derived from six standards supplied with the kit (0.625–15.0 ng/ml). All samples, controls, and standards were assayed in triplicate. In-house and manufacturer-supplied controls were included in each of the assays performed. Any sample outside of the range of the standard curve or with a coefficient of variation greater than 10% was re-assayed (n = 22). Intra-assay and inter-assay CV’s for 2-OHE1 were 4.4% and 8.8%, respectively; for 16a-OHE1, they were 5.1% and 9.2%, respectively. Data analysis Of the 203 women who attended a clinic visit, 200 returned a soy challenge urine sample, and of these, 199 completed a health and demographics questionnaire, and 191 and 193, respectively, provided a fasting blood and spot urine sample during days 5–9 of their menstrual cycle. Four women had estradiol concentrations [400 pg/ml, which are considered periovulatory values, and were excluded from all hormone and SHBG analyses. In addition, one woman had a value of zero for serum DHEA concentration and was excluded from analyses on DHEA. All hormone and SHBG data were skewed and were log transformed (natural log) for analyses. Differences between producers and non-producers of equol and ODMA in demographic and lifestyle factors were assessed using t-tests and MannWhitney-U tests (continuous data) and Chi square and Fishers exact tests (categorical data). Multiple regression analyses were conducted to assess relationships between equol and ODMA production and hormones and SHBG. Potential variables for adjustment were based on a priori knowledge of factors that have previously been shown to be associated with hormones and SHBG, urinary estrogen metabolites, and equol and ODMA-producer phenotypes. Variables included in each model were those which were significantly associated with the dependent variable or equol/ODMA-producer status in univariate analyses. The variables considered in univariate analyses were age, body mass index (BMI), waist-to-hip ratio (WHR), height, smoking status, family history of breast and/or ovarian cancer, race, ethnicity, education, prior hormone use, age at menarche, ever pregnant, age at first pregnancy, number of pregnancies resulting in live births, number of induced abortions, ever breastfed a child/children, total number of months breastfed a child/children, day of menstrual cycle on which the blood/urine sample was obtained, and percent

123

1088

mammographic breast density. Data were analyzed using SAS (version 9.1; SAS Institute, Cary, NC), and a twosided p-value of B0.05 was considered statistically significant.

Results Of the 200 women who returned a soy challenge urine sample, 55 (27.5%) were classed as equol-producers, and 182 (91%) were classed as ODMA-producers. Data on demographics, anthropometrics, and reproductive factors are shown in Table 1. More equol-producers than nonproducers were Hispanic or Latino and had received C17 years of education, but fewer equol-producers than non-producers had received 16 years of education. ODMAproducers were taller and less likely to be Asian than nonproducers. There were no differences between producers and non-producers of either daidzein metabolite for any of the reproductive characteristics measured using the health and demographics questionnaire (Table 1). In unadjusted analyses, compared to non-producers, equol-producers had lower serum concentrations of free testosterone (p = 0.02; Table 2). In adjusted analyses, there were no differences between producers and nonproducers of either daidzein metabolite (Tables 2 and 3).

Discussion In this predominantly Caucasian population of premenopausal women in the US, the prevalence of equol- and ODMA-producers was similar to that reported in other Western populations [14, 17, 38]. We observed some differences between daidzein-metabolizing phenotypes in demographic factors, but no differences in circulating concentrations of hormones and urinary estrogen metabolites. Associations between equol-producer status and education and ethnicity, and between ODMA-producer status and race and height, have been discussed elsewhere [31]. In agreement with our previous study in postmenopausal women [17], reproductive characteristics were similar between producers and non-producers of equol and ODMA. Because it has been shown in vitro that intestinal bacteria can metabolize hormones [22–25], our hypothesis was that the daidzein-metabolizing phenotypes might be markers of intestinal bacterial profiles that can affect hormone concentrations via unique metabolic actions of either the daidzein-metabolizing bacteria, or other bacteria that are associated with their presence in the intestines [18]. As such, differences in hormone concentrations between

123


phenotypes would be apparent irrespective of soy consumption. We found no difference in serum hormone concentrations between equol-producers and non-producers, which suggests that the intestinal bacteria responsible for, or associated with, equol and ODMA production are not involved in determining hormone concentrations. Our findings are somewhat in contrast to those of a small isoflavone intervention in premenopausal women [29]. In that study, irrespective of isoflavone dose, equol producers (n = 5), compared to non-producers (n = 9), had lower serum concentrations of a number of estrogens, androgens, and SHBG during the early follicular and mid luteal phases of the menstrual cycle. This suggested a protective effect, in terms of breast cancer risk, of being an equol-producer that was independent of isoflavone intake. However, the low soy dose provided 10 mg of isoflavones per day, which is higher than mean intakes in Western populations [39, 40], and it is possible that low levels of equol may be capable of eliciting biological responses. In subsequent, relatively small, studies in pre- [30] and postmenopausal [17] women, there were no significant differences between producers and non-producers of equol and ODMA (postmenopausal women only), in circulating concentrations of estrogens or androgens in the absence of prolonged soy consumption. In our study we did not observe any statistically significant differences between producers and nonproducers of ODMA in circulating concentrations of estrogens, androgens, or SHBG. To our knowledge, no other observational studies in premenopausal women have investigated relationships between daidzein-metabolizing phenotypes and urinary estrogen metabolite concentrations. In a soy intervention study in postmenopausal women [38] there was an increase in 2-OHE1 and the 2-OHE1: 16a-OHE1 ratio with soy consumption, but only among equol-producers. In a study in which postmenopausal women were phenotyped for equol- and ODMA-producer status using a soy challenge [17], the 2-OHE1: 16a-OHE1 ratio was non-significantly higher among equol-producers than non-producers, and 2-OHE1 was significantly higher in ODMA-producers compared with non-producers. In contrast to these findings, we did not observe any differences in urinary estrogen metabolite concentrations between producers and nonproducers of either daidzein metabolite. There are a number of strengths of this study. We employed relatively stringent eligibility criteria for recruitment and collected a large amount of information from the women, which has resulted in a very wellcharacterized study population. Blood and urine samples were collected within a specified phase of the menstrual cycle, and, to our knowledge, this is the largest study to date to have assessed hormones and SHBG in relation to daidzein-metabolizing phenotypes in premenopausal


1089

Table 1 Study participant demographics and lifestyle factors [mean (standard deviation) or n (%)] by equol-producer and O-desmethylangolensin (ODMA)-producer statusa Equol producers Equol non-producers p (n = 55) (n = 144)

ODMA producers ODMA non-producers p (n = 181) (n = 18)

Age (years)

42.6 (1.2)

42.3 (1.4)

0.15

42.5 (1.4)

42.1 (1.1)

Height (m)

1.65 (0.07)

1.65 (0.07)

0.82

1.66 (0.07)

1.61 (0.06)

0.003

Weight (kg) Body Mass Index (kg/mb)

70.5 (14.7) 25.8 (5.1)

71.1 (14.2) 26.0 (4.9)

0.78 0.83

71.3 (14.2) 26.0 (4.9)

67.1 (15.6) 25.8 (5.0)

0.23 0.90

Waist:Hip ratiob

0.79 (0.05)

0.79 (0.06)

0.69

0.79 (0.06)

0.80 (0.06)

0.30

White

49 (91)

125 (87)

0.93§

163 (91)

11 (65)

Asian

3 (6)

10 (7)

9 (5)

4 (24)

Other

2 (4)

8 (6)

8 (4)

2 (12)

5 (9)

1 (1)

5 (3)

1 (6)

50 (91)

142 (99)

176 (97)

16 (94)

4 (7)

8 (6)

12 (7)

0 (0)

0.23

Racec 0.006§

Ethnicity Hispanic or Latino Not Hispanic or Latino

0.007§

0.42§

Education B12 years 13–15 years 16 years C17 years

0.002§

12 (22)

40 (28)

45 (25)

7 (39)

8 (15)

53 (37)

56 (31)

5 (28)

68 (38)

6 (33)

7 (4)

1 (6)

0.58§

31 (56)

43 (30)

Smoker Current

1 (2)

7 (5)

Past

18 (33)

38 (26)

52 (29)

4 (22)

Never

36 (65)

99 (69)

122 (67)

13 (72)

Yes

38 (69)

103 (72)

130 (72)

11 (61)

No

17 (31)

41 (28)

51 (28)

7 (39)

2.65 (1.40)

2.90 (1.64)

0.40

2.9 (1.6)

2.6 (1.3)

0.66

Number of pregnancies resulting in 1.97 (0.93) live birthse

2.08 (0.88)

0.56

2.1 (0.9)

2.0 (1.2)

0.87

Number of induced abortionsf

0.72 (1.23)

0.63

0.64 (1.13)

0.64 (0.81)

0.62

30 (81) 7 (19)

85 (84) 16 (16)

0.80§

106 (83) 21 (17)

9 (82) 2 (18)

1.00§

21.2 (20.6)

22.8 (19.2)

0.69

22.1 (19.2)

25.2 (23.8)

0.65

Yes

42 (76)

100 (69)

0.33

129 (71)

13 (72)

0.93

No

13 (24)

44 (31)

52 (29)

5 (28)

13.0 (1.2)

12.8 (1.3)

12.8 (1.3)

13.2 (1.4)

0.51§

0.65§

Ever been pregnant

How many times pregnant?d

0.45 (0.65)

0.74

0.34

Did you breastfeed your child/children Yes No Length of time breastfed child/ children (months)g Ever used hormones

Age at menarche (years)

0.38

0.22

p values calculated using t-tests, Chi square test, Mann–Whitney U-test (denoted by), and Fishers exact test (denoted by §); percentages may not add to 100% due to rounding

a

b

Equol-producer (EP) n = 53 and equol non-producer (ENP) n = 141, and ODMA-producer (OP) n = 176

c

EP n = 54 and ENP n = 143, and OP n = 180 and ODMA non-producer (ONP) n = 17

d

EP n = 37 and ENP n = 103, and OP n = 129 and ONP n = 11

e


f


g


123

1090


Table 2 Unadjusted and adjusted geometric mean (95% confidence interval) serum estrogen, androgen, and sex hormone binding globulin, and urine estrogen metabolite concentrations by equol-producer statusa Measure

Unadjustedb,c Equol producers (n = 52)

Adjustedc,d Equol non-producers (n = 135)

p

Equol producers (n = 52)

Equol non-producers (n = 133)

p

73.11 (62.85,85.04)

0.24

1.49 (1.25,1.77)

0.82

Serum Estrone (pg/ml) Estrone-sulfate (ng/ml) Estradiol (pg/ml) Free estradiol (pg/ml) Testosterone (ng/dl)

66.05 (60.60,71.98)

72.02 (68.15,76.11)

1.46 (1.33,1.61)

1.50 (1.40,1.60)

0.10 68.53 (59.22,79.29) 0.70

1.51 (1.28,1.78)

102.03 (87.76,118.63) 99.33 (90.77,108.69) 0.76 91.60 (75.28,111.46) 85.43 (73.00,99.98) 2.33 (1.99,2.72)

2.25 (2.03,2.49)

0.72

2.17 (1.75,2.70)

0.21

27.85 (23.85,32.51)

0.25

23.78 (21.69,26.07)

25.90 (24.50,27.38)

Free testosterone (pg/ml)

3.93 (3.58,4.31)

4.51 (4.23,4.79)

0.02

4.20 (3.58,4.94)

4.72 (4.00,5.59)

0.06

Androstenedione (ng/ml)

0.90 (0.83,0.97)

0.96 (0.91,1.01)

0.15

0.94 (0.83,1.07)

0.98 (0.86,1.12)

0.40

0.80

3.95 (3.23,4.84)

3.81 (3.09,4.70)

0.64

Dehydroepiandrosterone (ng/ml)

0.12 26.07 (22.42,30.31)

0.41

1.93 (1.63,2.29)

3.69 (3.33,4.09)

3.62 (3.35,3.92)

Dehydroepiandrosterone-sulfate (lg/dl)

81.14 (70.27,93.70)

76.29 (69.52,83.73)

0.49 92.17 (72.16,117.72) 81.73 (63.51,105.18) 0.21

Sex hormone binding globulin (nmol/l)

56.84 (51.03,63.32)

51.53 (47.87,55.47)

0.16 62.99 (52.32,75.84)

58.34 (47.89,71.07)

2-Hydroxyestrone (ng/mg Cr)

16.89 (14.64,19.49)

14.82 (13.32,16.49)

0.19 16.22 (12.95,20.31)

13.96 (11.70,16.66)

0.12

16a-Hydroxyestrone (ng/mg Cr)

11.42 (10.36,12.59)

10.43 (9.68,11.24)

0.19 10.49 (8.90,12.36)

9.27 (8.15,10.56)

0.08

1.48 (1.32,1.66)

1.42 (1.30,1.56)

0.64

1.44 (1.23,1.69)

0.86

0.23

Urine

2:16 ratio a

1.46 (1.19,1.79)

Multiple regression analyses were used to assess relationships between equol production and hormones and SHBG

b

Equol-producer (EP) n for dehydroepiandrosterone analyses = 51

c

Equol non-producer (ENP) n for urine estrogen metabolite analyses = 137 (unless stated otherwise for adjusted analyses)

d

Adjustment variables included in each model were Hispanic/Latino, education, and the following: Estrone: day of cycle, ever breastfed a child/ children (yes/no), ever pregnant (yes/no), EP n = 51, and ENP n = 131; Estrone-sulfate: day of cycle; Estradiol: BMI, day of cycle, ever breastfed a child/children (yes/no), number of months breastfed a child/children, race (note: analysis not adjusted for Hispanic/Latino), EP n = 51, and ENP n = 130; Free estradiol: BMI, number of live births, number of months breastfed a child/children, day of cycle, race (note: analysis not adjusted for Hispanic/Latino), EP n = 49, and ENP n = 127; Testosterone: age, day of cycle, income; Free testosterone: age, BMI, WHR, day of cycle, EP n = 50, ENP n = 130; Androstenedione: age, BMI; Dehydroepiandrosterone: EP n = 51; Dehydroepiandrosteronesulfate: age, EP n = 52, and ENP n = 133; Sex hormone binding globulin: BMI, WHR, family history of breast or ovarian cancer, income, age at menarche; 2-hydroxyestrone: BMI, WHR, day of cycle, age at menarche, number of months breastfed a child/children, race (note: analysis not adjusted for Hispanic/Latino), EP n = 49, ENP n = 130; 16a-hydroxyestrone: day of cycle, race (note: analysis not adjusted for Hispanic/ Latino); 2:16 ratio: BMI, WHR, ever pregnant (yes/no), number of months breastfed a child/children, race (note: analysis not adjusted for Hispanic/Latino), EP n = 49, ENP n = 130)

women. All women underwent a soy challenge to determine daidzein-metabolizing phenotypes, which is essential in Western populations with low soy intakes. Given that this study was conducted in a Western population and that relatively few women reported regularly consuming soy [31], our findings provide information on the effects of the daidzein-metabolizing phenotypes per se on hormones and SHBG, rather than the combined effect of soy intakes and daidzein-metabolizing phenotypes. There are several limitations of our study. Given the widespread ability to produce ODMA, the assessment of differences between ODMA-producers and non-producers was limited due to the relatively small number of ODMA non-producers. In addition, most women were Caucasian and well-educated, and our findings may be generalizable only to similar populations of women. Although we

123

recruited women who reported regular menstrual periods, some of the women in our study may have been perimenopausal. However, median age at natural menopause is approximately 49.6–51.5 years, with an average duration of perimenopause of 4 years [41, 42], and estradiol levels do not decline until the late perimenopause, or within a year of the cessation of menses [43, 44]. This suggests that the age range of the women in this study was sufficient to capture predominantly premenopausal women, and that any impact of early perimenopause on estradiol levels would be minimal. An additional limitation of this study is that we may not have adequate statistical power to detect small differences between daidzein-metabolizing phenotypes. Post-hoc power calculations, based on the hormone distributions observed and sample sizes achieved in this study, revealed that we had between 27% (for 2-OHE1) and 92%


1091

Table 3 Unadjusted and adjusted geometric mean (95% confidence interval) serum estrogen, androgen, and sex hormone binding globulin, and urine estrogen metabolite concentrations by O-desmethylangolensin (ODMA)-producer statusa Measure

Unadjustedb,c ODMA producers (n = 171)

Adjustedc,d ODMA non-producers (n = 16)

p

ODMA producers (n = 171)

ODMA non-producers (n = 16)

p

65.35 (55.61,76.81)

0.91

1.69 (1.39,2.04)

0.40

Serum Estrone (pg/ml) Estrone-sulfate (ng/ml) Estradiol (pg/ml) Free estradiol (pg/ml) Testosterone (ng/dl)

70.51 (67.08,74.12)

68.14 (60.50,76.75)

1.48 (1.39,1.56)

1.65 (1.43,1.90)

0.69 66.00 (59.78,72.87) 0.26

1.55 (1.38,1.74)

99.96 (92.06,108.54) 101.24 (82.44,124.33) 0.93 85.88 (73.65,100.14) 96.92 (75.24,124.84) 0.36 2.26 (2.06,2.48)

2.38 (2.00,2.84)

0.74

1.93 (1.63,2.28)

0.38 24.10 (21.42,27.12)

2.34 (1.76,3.11)

0.20

25.13 (23.90,26.42)

27.12 (23.07,31.88)

25.63 (21.47,30.58)

0.49

Free testosterone (pg/ml)

4.32 (4.09,4.57)

4.50 (3.73,5.44)

0.67

4.32 (3.84,4.86)

4.23 (3.49,5.12)

0.83

Androstenedione (ng/ml)

0.94 (0.90,0.98)

0.98 (0.86,1.11)

0.63

0.90 (0.82,0.98)

0.93 (0.80,1.07)

0.67

Dehydroepiandrosterone (ng/ml)

3.65 (3.42,3.90)

3.54 (2.73,4.58)

0.78

3.68 (3.20,4.24)

3.60 (2.84,4.56)

0.86

Dehydroepiandrosterone-sulfate (lg/dl)

77.72 (71.56,84.40)

76.50 (60.06,97.43)

0.91 89.06 (75.15,105.53) 77.23 (58.04,102.76) 0.33

Sex hormone binding globulin (nmol/l)

52.70 (49.46,56.15)

55.77 (44.15,70.45)

0.61 52.39 (44.84,61.20)

58.81 (47.48,72.84)

0.24

2-Hydroxyestrone (ng/mg Cr)

15.38 (14.09,16.78)

15.24 (10.14,22.91)

0.96 13.22 (11.14,15.68)

14.15 (10.55,18.99)

0.65

16a-Hydroxyestrone (ng/mg Cr)

10.62 (9.96,11.32)

11.54 (9.56,13.92)

0.45

8.97 (7.91,10.18)

10.34 (8.33,12.83)

0.20

1.45 (1.35,1.56)

1.32 (0.89,1.95)

0.49

1.43 (1.22,1.66)

1.36 (1.05,1.75)

0.71

Urine

2:16 ratio a

Multiple regression analyses were used to assess relationships between ODMA production and hormones and SHBG

b

ODMA-producer (OP) n for dehydroepiandrosterone analyses = 170

c

OP n for urine estrogen metabolite analyses = 173 (unless stated otherwise for adjusted analyses)

d

Adjustment variables included in each model were height, race, and the following: Estrone: day of cycle, ever breastfed a child/children (yes/ no), ever pregnant (yes/no), OP n = 167; Estrone-sulfate: day of cycle; Estradiol: BMI, day of cycle, ever breastfed a child/children (yes/no), number of months breastfed a child/children, OP n = 165; Free estradiol: BMI, number of live births, number of months breastfed a child/ children, day of cycle, OP n = 161 and ODMA non-producer (ONP) n = 15; Testosterone: age, day of cycle, income, OP n = 170; Free testosterone: age, BMI, WHR, day of cycle, education, OP n = 165; Androstenedione: age, BMI, OP n = 170; Dehydroepiandrosterone: OP n = 170; Dehydroepiandrosterone-sulfate: age, OP n = 170; Sex hormone binding globulin: BMI, WHR, family history of breast of ovarian cancer, income, age at menarche, education, OP n = 165; 2-hydroxyestrone: BMI, WHR, day of cycle, age at menarche, number of months breastfed a child/children, OP n = 163; 16a-hydroxyestrone: day of cycle; 2:16 ratio: BMI, WHR, ever pregnant (yes/no), number of months breastfed a child/children, OP n = 163

(for androstenedione) statistical power and between 13% (for 2-OHE1) and 56% (for androstenedione) statistical power to detect a 15% difference in hormone concentration between equol-producers and non-producers and between ODMA-producers and non-producers, respectively. Finally, a large number of statistical comparisons were made, suggesting that some of the statistically significant findings may have occurred by chance alone. Overall, our findings suggest no differences in serum and urine hormone concentrations between daidzeinmetabolizing phenotypes. It is possible that harboring the intestinal bacteria that produce equol or ODMA in addition to soy exposure may be more important than the phenotypes alone. This study was conducted in a Western population with relatively low soy consumption patterns, and future studies should assess interactions between equol- and ODMA-producer status and soy consumption in

relation to circulating hormones and SHBG and urinary estrogen metabolites. Acknowledgments We wish to thank Kathy Plant, Kelly Ehrlich, and the GH group for screening interviews, clinic visits and study coordination, Wendy Thomas for isoflavone analyses, JoAnn Prunty for creatinine analyses, and all of the study participants.

References 1. Persson I (2000) Estrogens in the causation of breast, endometrial and ovarian cancers—evidence and hypotheses from epidemiological findings. J Steroid Biochem Mol Biol 74:357–364. doi: 10.1016/S0960–0760(00)00113-8 2. Stewart EA (2001) Uterine fibroids. Lancet 357:293–298. doi: 10.1016/S0140-6736(00)03622-9 3. Giudice LC, Kao LC (2004) Endometriosis. Lancet 364: 1789–1799. doi:10.1016/S0140-6736(04)17403-5

123

1092 4. Muti P, Bradlow HL, Micheli A et al (2000) Estrogen metabolism and risk of breast cancer: a prospective study of the 2:16alpha-hydroxyestrone ratio in premenopausal and postmenopausal women. Epidemiology 11:635–640. doi:10.1097/0000 1648-200011000-00004 5. Kabat GC, Chang CJ, Sparano JA et al (1997) Urinary estrogen metabolites and breast cancer: a case–control study. Cancer Epidemiol Biomarkers Prev 6:505–509 6. Meilahn EN, De Stavola B, Allen DS et al (1998) Do urinary oestrogen metabolites predict breast cancer? Guernsey III cohort follow-up. Br J Cancer 78:1250–1255 7. Zheng W, Dunning L, Jin F, Holtzman J (1998) Correspondence re: Kabat GC et al., Urinary estrogen metabolites and breast cancer: a case–control study. Cancer Epidemiol Biomark Prev 6:505–509, 1997. Cancer Epidemiol Biomarkers Prev 7:85–86 8. Kurzer MS (2002) Hormonal effects of soy in premenopausal women and men. J Nutr 132:570S–573S 9. Kurzer MS (2000) Hormonal effects of soy isoflavones: studies in premenopausal and postmenopausal women. J Nutr 130:660S–661S 10. Setchell KD, Brown NM, Lydeking-Olsen E (2002) The clinical importance of the metabolite equol-a clue to the effectiveness of soy and its isoflavones. J Nutr 132:3577–3584 11. Rowland IR, Wiseman H, Sanders TA, Adlercreutz H, Bowey EA (2000) Interindividual variation in metabolism of soy isoflavones and lignans: influence of habitual diet on equol production by the gut microflora. Nutr Cancer 36:27–32. doi:10.1207/S15327914 NC3601_5 12. Akaza H, Miyanaga N, Takashima N et al (2002) Is daidzein nonmetabolizer a high risk for prostate cancer? A case–controlled study of serum soybean isoflavone concentration. Jpn J Clin Oncol 32:296–300. doi:10.1093/jjco/hyf064 13. Kelly GE, Joannou GE, Reeder AY, Nelson C, Waring MA (1995) The variable metabolic response to dietary isoflavones in humans. Proc Soc Exp Biol Med 208:40–43 14. Lampe JW, Karr SC, Hutchins AM, Slavin JL (1998) Urinary equol excretion with a soy challenge: influence of habitual diet. Proc Soc Exp Biol Med 217:335–339 15. Hutchins AM, Slavin JL, Lampe JW (1995) Urinary isoflavonoid phytoestrogen and lignan excretion after consumption of fermented and unfermented soy products. J Am Diet Assoc 95: 545–551. doi:10.1016/S0002-8223(95)00149-2 16. Arai Y, Uehara M, Sato Y et al (2000) Comparison of isoflavones among dietary intake, plasma concentration and urinary excretion for accurate estimation of phytoestrogen intake. J Epidemiol 10:127–135 17. Frankenfeld CL, McTiernan A, Tworoger SS et al (2004) Serum steroid hormones, sex hormone-binding globulin concentrations, and urinary hydroxylated estrogen metabolites in post-menopausal women in relation to daidzein-metabolizing phenotypes. J Steroid Biochem Mol Biol 88: 399–408. doi:10.1016/j.jsbmb. 2004.01.006 18. Atkinson C, Frankenfeld CL, Lampe JW (2005) Gut bacterial metabolism of the soy isoflavone daidzein: exploring the relevance to human health. Exp Biol Med (Maywood) 230:155–170 19. Evans BA, Griffiths K, Morton MS (1995) Inhibition of 5 alphareductase in genital skin fibroblasts and prostate tissue by dietary lignans and isoflavonoids. J Endocrinol 147:295–302 20. Pelissero C, Lenczowski MJ, Chinzi D, Davail-Cuisset B, Sumpter JP, Fostier A (1996) Effects of flavonoids on aromatase activity, an in vitro study. J Steroid Biochem Mol Biol 57: 215–223. doi:10.1016/0960-0760(95)00261-8 21. Adlercreutz H, Bannwart C, Wahala K et al (1993) Inhibition of human aromatase by mammalian lignans and isoflavonoid phytoestrogens. J Steroid Biochem Mol Biol 44:147–153. doi: 10.1016/0960-0760(93)90022-O

123

Cancer Causes Control (2008) 19:1085–1093 22. Adlercreutz H, Martin F, Pulkkinen M et al (1976) Intestinal metabolism of estrogens. J Clin Endocrinol Metab 43:497–505 23. Ja¨rvenpa¨a¨ P, Kosunen T, Fotsis T, Adlercreutz H (1980) In vitro metabolism of estrogens by isolated intestinal micro-organisms and by human faecal microflora. J Steroid Biochem 13:345–349. doi:10.1016/0022-4731(80)90014-X 24. Ja¨rvenpa¨a¨ P (1990) In vitro metabolism of catechol estrogens by human fecal microflora. J Steroid Biochem 35:289–292. doi: 10.1016/0022-4731(90)90286-2 25. Lombardi P, Goldin B, Boutin E, Gorbach SL (1978) Metabolism of androgens and estrogens by human fecal microorganisms. J Steroid Biochem 9:795–801. doi:10.1016/0022-4731(78)90203-0 26. Ingram D, Sanders K, Kolybaba M, Lopez D (1997) Case–control study of phyto-oestrogens and breast cancer. Lancet 350:990– 994. doi:10.1016/S0140-6736(97)01339-1 27. Zheng W, Dai Q, Custer LJ et al (1999) Urinary excretion of isoflavonoids and the risk of breast cancer. Cancer Epidemiol Biomarkers Prev 8:35–40 28. Grace PB, Taylor JI, Low YL et al (2004) Phytoestrogen concentrations in serum and spot urine as biomarkers for dietary phytoestrogen intake and their relation to breast cancer risk in European prospective investigation of cancer and nutrition-norfolk. Cancer Epidemiol Biomarkers Prev 13:698–708 29. Duncan AM, Merz-Demlow BE, Xu X, Phipps WR, Kurzer MS (2000) Premenopausal equol excretors show plasma hormone profiles associated with lowered risk of breast cancer. Cancer Epidemiol Biomarkers Prev 9:581–586 30. Bonorden MJ, Greany KA, Wangen KE et al (2004) Consumption of Lactobacillus acidophilus and Bifidobacterium longum do not alter urinary equol excretion and plasma reproductive hormones in premenopausal women. Eur J Clin Nutr 58:1635–1642. doi:10.1038/sj.ejcn.1602020 31. Atkinson C, Newton KM, Aiello-Bowles EJ, Yong M, Lampe JW (2008) Demographic, anthropometric, and lifestyle factors and dietary intakes in relation to daidzein-metabolizing phenotypes among premenopausal women in the United States. Am J Clin Nutr 87:679–687 32. Taplin SH, Ichikawa L, Buist DS, Seger D, White E (2004) Evaluating organized breast cancer screening implementation: the prevention of late-stage disease? Cancer Epidemiol Biomarkers Prev 13:225–234. doi:10.1158/1055-9965.EPI-03-0206 33. Goebelsmann U, Bernstein GS, Gale JA et al (1979) Serum gonadotrophin, testosterone, estradiol and estrone levels prior to and following bilateral vasectomy. In: Lepow IH, Crozier R (eds) Vasectomy: Immunologic and pathophysiologic effects in animals and man. New York, Academic Press, p 165 34. Probst-Hensch NM, Ingles SA, Diep AT et al (1999) Aromatase and breast cancer susceptibility. Endocr Relat Cancer 6:165–173. doi:10.1677/erc.0.0060165 35. Vermeulen A, Verdonck L, Kaufman JM (1999) A critical evaluation of simple methods for the estimation of free testosterone in serum. J Clin Endocrinol Metab 84:3666–3672. doi: 10.1210/jc.84.10.3666 36. Sodergard R, Backstrom T, Shanbhag V, Carstensen H (1982) Calculation of free and bound fractions of testosterone and estradiol-17 beta to human plasma proteins at body temperature. J Steroid Biochem 16:801–810. doi:10.1016/0022-4731(82) 90038-3 37. Rinaldi S, Geay A, Dechaud H et al (2002) Validity of free testosterone and free estradiol determinations in serum samples from postmenopausal women by theoretical calculations. Cancer Epidemiol Biomarkers Prev 11:1065–1071 38. Nettleton JA, Greany KA, Thomas W, Wangen KE, Adlercreutz H, Kurzer MS (2005) The effect of soy consumption on the urinary 2:16-hydroxyestrone ratio in postmenopausal women

Cancer Causes Control (2008) 19:1085–1093 depends on equol production status but is not influenced by probiotic consumption. J Nutr 135:603–608 39. Mulligan AA, Welch AA, McTaggart AA, Bhaniani A, Bingham SA (2007) Intakes and sources of soya foods and isoflavones in a UK population cohort study (EPIC-Norfolk). Eur J Clin Nutr 61:248–254. doi:10.1038/sj.ejcn.1602509 40. Keinan-Boker L, van Der Schouw YT, Grobbee DE, Peeters PH (2004) Dietary phytoestrogens and breast cancer risk. Am J Clin Nutr 79:282–288 41. McKinlay SM, Brambilla DJ, Posner JG (1992) The normal menopause transition. Maturitas 14:103–115. doi:10.1016/03785122(92)90003-M

1093 42. McKinlay SM (1996) The normal menopause transition: an overview. Maturitas 23:137–145. doi:10.1016/0378-5122(95) 00985-X 43. Burger HG, Dudley EC, Hopper JL et al (1999) Prospectively measured levels of serum follicle-stimulating hormone, estradiol, and the dimeric inhibins during the menopausal transition in a population-based cohort of women. J Clin Endocrinol Metab 84:4025–4030. doi:10.1210/jc.84.11.4025 44. Burger HG, Dudley EC, Robertson DM, Dennerstein L (2002) Hormonal changes in the menopause transition. Recent Prog Horm Res 57:257–275. doi:10.1210/rp.57.1.257

123