Consumption of Lactobacillus Acidophilus and Bifidobacterium Longum Do Not Alter Urinary Equol Excretion and Plasma Reproductive Hormones in Premenopausal Women

Home , Equol, Isoflavonoid

ORIGINAL COMMUNICATION Consumption of Lactobacillus acidophilus and Bifidobacterium longum do not alter urinary equol excretion and plasma reproductive hormones in premenopausal women

MJL Bonorden1, KA Greany1, KE Wangen1, WR Phipps2, J Feirtag1, H Adlercreutz3 and MS Kurzer1*

1Department of Food Science and Nutrition, University of Minnesota, St. Paul, MN, USA; 2Department of Obstetrics and Gynecology, University of Rochester, Rochester, NY, USA; and 3Division of Clinical Chemistry, University of Helsinki, Helsinki, Finland

Objective: To confirm the results of an earlier study showing premenopausal equol excretors to have hormone profiles associated with reduced breast cancer risk, and to investigate whether equol excretion status and plasma hormone concentrations can be influenced by consumption of probiotics. Design: A randomized, single-blinded, placebo-controlled, parallel-arm trial. Subjects: In all, 34 of the initially enrolled 37 subjects completed all requirements. Intervention: All subjects were followed for two full menstrual cycles and the first seven days of a third cycle. During menstrual cycle 1, plasma concentrations of estradiol (E2), estrone (E1), estrone-sulfate (E1-S), testosterone (T), androstenedione (A), dehydroepiandrosterone-sulfate (DHEA-S), and sex-hormone-binding globulin (SHBG) were measured on cycle day 2, 3, or 4, and urinary equol measured on day 7 after a 4-day soy challenge. Subjects then received either probiotic capsules (containing Lactobacillus acidophilus and Bifidobacterium longum) or placebo capsules through day 7 of menstrual cycle 3, at which time both the plasma hormone concentrations and the post-soy challenge urinary equol measurements were repeated. Results: During menstrual cycle 1, equol excretors and non-excretors were not significantly different with respect to subject characteristics, diet, or hormone concentrations. Significant inverse correlations were found between E2 and body mass index (BMI) (P ¼ 0.02), SHBG and BMI (P ¼ 0.01), DHEA-S and dietary fiber (P ¼ 0.04), and A and protein:carbohydrate ratio (P ¼ 0.02). Probiotic consumption failed to significantly alter equol excretor status or hormone concentrations during menstrual cycle 3, although there were trends towards decreased concentrations of T (P ¼ 0.14) and SHBG (P ¼ 0.10) in the probiotic group. Conclusions: We were unable to verify a previously reported finding of premenopausal equol excretors having plasma hormone concentrations different from those of nonexcretors. Furthermore, a 2-month intervention with probiotic capsules did not significantly alter equol excretion or plasma hormone concentrations. European Journal of Clinical Nutrition (2004) 58, 1635–1642. doi:10.1038/sj.ejcn.1602020 Published online 23 June 2004

Keywords: phytoestrogen; isoflavone; equol; hormone; women; premenopausal; probiotic

*Correspondence: MS Kurzer, Department of Food Science and Nutrition, University of Minnesota, St. Paul, Minnesota 55108, USA. Introduction E-mail: [email protected] Several lines of evidence suggest that consumption of Contributors: MJLB was the primary author and study coordinator isoflavones, phytoestrogens abundant in soy foods, may and assisted with subject recruiting; KAG was a study co-coordinator and assisted with subject recruiting; KEW helped with study design; decrease the risk of breast cancer, possibly on the basis of WRP assisted with study design and provided medical consulting; JF associated changes in sex steroid action. Such evidence assisted with study design and provided microbiological consulting; includes case–control studies showing that increased soy HA developed and performed the equol analysis; MSK was the intake is associated with reduced breast cancer risk (Lee et al, principal investigator, who supervised all aspects of the study. All authors contributed to the data interpretation and manuscript 1991; Wu et al, 1996; Zheng et al, 1999; Dai et al, 2002). In preparation. particular, the isoflavone-metabolite equol has been of Probiotics, phytoestrogen metabolism, and hormones in women MJL Bonorden et al 1636 interest, and in a case–control study of women consuming a sumption of bacteria with low b-glucuronidase activity Western diet, those in the highest quartile of urinary equol would increase production of equol and reduce plasma excretion had an adjusted odds ratio of 0.27 (95% con- concentrations of reproductive hormones. fidence interval 0.10–0.69) compared to those in the lowest quartile of excretion (Ingram et al, 1997). Equol itself is not present in soy foods, but is produced by Materials and methods bacterial metabolism of the isoflavone daidzein. Interestingly, Subjects however, only about 30–40% of humans excrete more than Premenopausal women, 18–36 y old, were selected from the trace amounts of equol, regardless of isoflavone intake Minneapolis/St. Paul metropolitan area after a telephone (Cassidy et al, 1994; Hutchins et al, 1995; Kelly et al, 1995; screening and a detailed self-administered questionnaire. Lampe et al, 1998; Duncan et al, 2000; Rowland et al, 2000). Exclusionary criteria included BMI o18 or 430 kg/m2, Furthermore, in a soy-feeding study in premenopausal menstrual cycles o24 or 432 days, antibiotic or exogenous women, equol excretors had lower concentrations of sex hormone use within the last 3 months, probiotic supple- steroids compared to nonexcretors, regardless of isoflavone ment use within the last 30 days, alcohol consumption dose (Duncan et al, 2000). These results suggest that the greater than seven drinks/week, chronic health problems, association of equol excretion and decreased breast cancer regular medication use including aspirin, weight change of risk may largely reflect the tendency of equol excretors to more than 10 pounds within the previous year or more than have more favorable hormonal profiles from the standpoint five pounds within the previous 2 months, soy allergy, of breast cancer risk, as opposed to merely reflecting increased current pregnancy or lactation. Of the women evaluated for isoflavone intake. If this were the case, equol may be a marker baseline analysis, 72% were Caucasian, 6% were African for the presence of colonic bacterial enzymatic activity that American, 10% were Asian, 6% were Hispanic, and 6% from both allows for equol to be produced as well as for a favorable South America were multi-racial. In all, 34 subjects com- sex steroid concentration profile, presumably on the basis of pleted the entire study. changes in the enterohepatic circulation of steroids. Owing to the potential biological importance of equol, researchers have investigated whether equol production Experimental design status can be altered. In fact, it is known that certain dietary The University of Minnesota Institutional Review Board, changes can alter the bacterial profile of the intestine (Rao, Human Subjects Committee, approved the study protocol. 1995) and may also affect isoflavone metabolism. Two Two-thirds of the subjects were randomized to receive reported attempts have been made to increase equol probiotic and one-third randomized to receive placebo excretion through dietary means. Lu et al (1996) reportedly capsules for approximately 2 months (5777 days) in a increased equol excretion in six premenopausal women single-blinded, parallel-arm design. Subjects remained free- given soy milk for 1 month, while Lampe et al (2001) were living and supplemented their normal diets with a 4-day soy not able to change equol-excretor status in premenopausal challenge on days 4–7 of menstrual cycles one and three. women after giving soy protein or wheat bran for 1 month. Beginning on day 8 of menstrual cycle one, all subjects Effects of intestinal bacteria on enterohepatic circulation consumed probiotic or placebo capsules daily through day 7 of steroid hormones may also be relevant to cancer of menstrual cycle three. prevention. Lactobacilli and bifidobacteria, two predominant families of intestinal flora (Gibson et al, 1995), have low b-glucuronidase activity (Hawksworth et al, 1971), and may Experimental diet reduce enterohepatic circulation of plasma hormones by Subjects consumed their habitual diets, but were provided decreasing hydrolysis of glucuronide-conjugated hormones with detailed dietary instructions to minimize the consump- in the intestine (Rowland et al, 1999). Furthermore, con- tion of phytoestrogens and fermented foods. Subjects were sumption of lactobacilli and bifidobacteria has been shown also instructed to consume three placebo or probiotic- to significantly increase b-glucosidase activity in humans containing capsules daily on an empty stomach. Each (Marteau et al, 1990), an activity necessary to convert the probiotic capsule contained a total of 109 colony-forming isoflavone glycoside to an aglycone, and specific strains of units of Lactobacillus acidophilus and Bifidobacterium longum bifidobacteria have been shown to increase equol synthesis strain DDSs þ 1 and 10–15 mg fructooligosaccharide (UAS in vitro (Tsangalis et al, 2003). Laboratories, Minnetonka, MN, USA). The placebo capsules We hypothesized that the ability to convert daidzein to (Gallipots Inc., St. Paul, MN, USA) contained rice starch and equol is characteristic of bacteria that also reduce enterohe- were of similar size, shape, and packaging as the probiotic patic circulation of reproductive hormones and may explain capsules. For the soy challenge, subjects consumed soy the inverse association between equol excretion and breast protein isolate (The Solae Company, St. Louis, MO, USA) cancer risk. The objectives of this study were to verify the providing 0.49 g protein and 0.84 mg isoflavones (aglycone cancer-preventive reproductive hormone profiles of preme- units) per kilogram body weight per day (3276 g protein and nopausal equol excretors and to determine whether con- 55711 mg isoflavones/day) for 4 days. The isoflavone

European Journal of Clinical Nutrition Probiotics, phytoestrogen metabolism, and hormones in women MJL Bonorden et al 1637 breakdown was 57% genistein, 35% daidzein, and 8% into two distinct categories. Equol excretors excreted greater glycitein. Given known information about the pharmacoki- than 900 nmol per 24 h and nonexcretors excreted less than netics of isoflavones, provision of this dose for 4 days is 400 nmol per 24 h. No subjects excreted between 400 and sufficient to separate equol excretors from nonexcretors. 900 nmol per 24 h. Statistical analysis was performed using the Statistical Study procedures Analysis System (SAS Institute, Inc., Cary, NC, USA) version Body weight was obtained in a hospital gown and fasting 8.2. Comparisons of baseline anthropometric measures, diet, blood was drawn in heparinized tubes on days 2, 3 or 4 of and hormones between placebo and control groups and menstrual cycles one and three. Plasma was separated, between equol excretors and nonexcretors were made using ascorbic acid and sodium azide were added to a final two-sample t-tests. Within-subject comparisons of baseline concentration of 0.1%, and aliquots were frozen at À701C and end-of-intervention hormone, diet, and equol excretion until analysis. were made using paired t-tests. Spearman correlation Urine collections (24 h) were started in the morning of day coefficients, r, were calculated to assess the relationships 7 of menstrual cycles one and three, the final day of each 4- among baseline BMI, diet, and hormones. Data are presented day soy challenge. Urine was collected in plastic containers as mean7standard deviation (s.d.). Equol and DHEA-S containing 1 g of ascorbic acid per liter and separated into measurements were log transformed for statistical analysis aliquots after the addition of sodium azide to a final and are presented as geometric mean (95% confidence concentration of 0.1%. Aliquots were frozen at –201C until interval). P-values were considered significant at less than analysis. Urinary creatinine was measured to assess the 0.05. completeness of collection. Dietary data were collected at baseline and following the 2- month intervention using the Healthy Habits and History Results Questionnaire (Block et al, 1986), a food frequency ques- Of the 37 women who started the study, 34 completed all tionnaire containing 63 items, which incorporates the study requirements. Three subjects dropped out during the portion size of foods consumed into the final analysis. The course of the study, but provided baseline plasma and urine questionnaire included consumption of food items such as samples as well as dietary data. One subject had a 20-day yogurt and milk substitute, which would help assess subject menstrual cycle during the course of the study and was compliance to dietary restrictions. removed from all analyses. Therefore, 36 women were included in baseline statistics and correlations and 33 were included in evaluation of the dietary intervention. Analytical methods Subject characteristics, baseline dietary data, and baseline All plasma samples were analyzed for estradiol (E ), estrone 2 plasma hormones are shown in Table 1. There were no (E ), estrone-sulfate (E -S), testosterone (T), androstenedione 1 1 significant differences in age, body weight, body mass index (A), dehydroepiandrosterone-sulfate (DHEA-S), and sex hor- (BMI), menstrual cycle length, or intake of total energy, mone binding globulin (SHBG). All samples were analyzed for each hormone in one batch by double antibody RIA, using 125I-labeled hormone (Diagnostics Systems Labora- tories, Inc., Webster, TX, USA). Intra-assay variabilities were Table 1 Baseline characteristics, daily dietary intake and early follicular o4% for all analytes. plasma hormone concentrations of subjects Urine was analyzed for equol by competitive time-resolved Probiotic (n ¼ 24) Placebo (n ¼ 12) fluoroimmunoassay as previously described (Brouwers et al, 2003). Intra-assay coefficients of variation were between 3.4 Age (y) 27752675 and 6.9%, and inter-assay coefficients of variation were Body wt (kg) 67.0713 60.7711 2 7 7 between 7.4 and 7.7%, at three different concentration BMI (kg/m ) 23.8 1.5 23.0 4.1 MC length (days) 28.071.5 27.871.4 levels. The values were corrected by a formula to correspond Energy (kJ) 608271612 573571047 to values obtained by reference gas chromatographic-mass Protein (g) 60.3721.4 57.8714.2 spectrometric methods and mean recoveries were 94.3% Carbohydrate (g) 165749 165744 7 7 (Brouwers et al, 2003). Fat (g) 60.7 17 52.2 13 Dietary fiber (g) 12.274.7 12.674.5 Dietary information was analyzed with DietSys Software E2 (pmol/l) 158741 174736 version 4.0 (National Cancer Institute, 1997). Diets were E1 (pmol/l) 74.0729 74.477.4 7 7 evaluated for energy, dietary fiber, macronutrient, and E1-S (nmol/l) 22.2 7.2 19.7 7.4 7 7 micronutrient content. T (nmol/l) 1.28 0.4 1.21 0.4 A (nmol/l) 6.8474.0 6.5672.1 DHEA-S (nmol/l) 4526 (3442, 5935) 4390 (2954, 6531) 7 7 Statistical analysis SHBG (nmol/l) 31.4 14 35.7 17 Although there are no established levels of equol excretion All values are mean7s.d. with the exception of DHEA-S, which is shown as that separate equol excretors from nonexcretors, subjects fell geometric mean (95% CI). MC ¼ menstrual cycle.

European Journal of Clinical Nutrition Probiotics, phytoestrogen metabolism, and hormones in women MJL Bonorden et al 1638 protein, carbohydrate, fat, fiber or micronutrients between although there was a trend toward decreased SHBG in the the probiotic and placebo groups at baseline. There were also probiotic group (P ¼ 0.10). There were also no significant no significant differences in plasma hormones between the differences in pre-post hormone changes between the two probiotic and placebo groups at baseline. groups although the probiotic group showed a tendency At baseline, significant inverse correlations were found toward decreased T relative to the placebo group (P ¼ 0.14).

between E2 and BMI (r ¼À0.39, P ¼ 0.02), SHBG and BMI Baseline subject characteristics, dietary intake data, and (r ¼À0.41 P ¼ 0.01), DHEA-S and dietary fiber (r ¼À0.34, plasma hormone concentrations of equol excretors and P ¼ 0.04), and between A and protein:carbohydrate ratio nonexcretors are shown in Table 4. Of the 36 women (r ¼À0.40, P ¼ 0.02). No other significant associations were included in the analysis at baseline, 13 (36%) were equol found between baseline BMI or diet and plasma hormone excretors. There were no significant differences in age, body concentrations; however, A tended to decrease with increas- weight, BMI, cycle length, or consumption of energy, ing BMI (r ¼À0.27 P ¼ 0.11) and T tended to decrease with macronutrients, micronutrients or dietary fiber between increasing protein:carbohydrate ratio (r ¼À0.25, P ¼ 0.14). equol excretors and nonexcretors. There were also no There were no significant changes in body weight or intake significant differences between these two groups with

of total energy, carbohydrate, fat, fiber or micronutrients respect to E2,E1,E1-S, T, A, DHEA-S or SHBG, although the (calcium, phosphorous, iron, sodium, potassium, vitamin A, excretors showed a nonsignificant trend toward lower E1,E1- vitamin B1, vitamin B2, vitamin B3, vitamin C, fiber, folate, S and T, and higher SHBG. vitamin E, zinc, vitamin B6, and magnesium) in either group Excretion of equol (expressed as nmol/24 h) did not during the intervention (Table 2, micronutrient data not change significantly in excretors or nonexcretors following shown). However, the probiotic group had a significant drop the 2-month intervention in either the treatment or placebo in protein intake (P ¼ 0.006). group (Table 5). Results were the same when equol excretion Plasma reproductive hormone data are presented in was expressed per mg of urinary creatinine. Several indivi- Table 3. When comparing baseline to end-of-intervention duals increased excretion during the study although in- values within probiotic or placebo groups, no significant creases in those receiving probiotic capsules were similar in changes occurred in any of the hormones measured, magnitude to those receiving placebo.

Table 2 Daily dietary intake and body weight of subjects at baseline and post-intervention

Probiotic (n ¼ 22) Placebo (n ¼ 11)

Baseline Post-intervention Baseline Post-intervention

Energy (kJ) 601971665 558071826 55927971 530471584 Protein (g) 60.1722 50.6720* 56.7714 51.8718 Carbohydrate (g) 162750 152752 159740 156753 Fat (g) 60.5718 56.2721 51.8714 47.0716 Fiber (g) 12.274.9 11.975.4 12.574.7 13.274.6 Body wt (kg) 67.3710 67.1710 59.3714 59.0714

Mean7s.d. based on respective food frequency questionnaires. *Significantly different from baseline (P ¼ 0.006).

Table 3 Early follicular hormones and equol excretion at baseline and post-intervention

Probiotic (n ¼ 22) Placebo (n ¼ 11)

Baseline Post-Intervention Baseline Post-Intervention

E2 (pmol/l) 159743 151737 181731 167728 E1 (pmol/l) 74.0729 75.9726 79.6732 79.2732 E1-S (nmol/l) 22.177.7 22.078.2 20.976.6 22.079.3 T (nmol/l) 1.3270.4 1.2570.4 1.2170.4 1.3970.4 A (nmol/l) 6.9174.2 7.1673.5 6.6372.1 6.9873.5 DHEA-S (nmol/l) 4661 (3686, 5881) 4851 (3848, 6152) 4092 (2656, 6938) 3984 (2412, 6558) SHBG (nmol/l) 32.3714 29.1712 36.8719 35.0718 Equol among excretors (nmol/24-h) 3281 (2440, 4412) 3130 (2046, 4790) 2465 (1255, 4844) 3537 (2217, 5641) Equol among nonexcretors (nmol/24-h) 185 (148, 232) 187 (143, 244) 124 (67, 227) 99 (59, 167)

All values are mean7s.d. except DHEA-S and equol, which are shown as geometric mean (95% CI). Within the probiotic group, n ¼ 7 excretors and n ¼ 15 nonexcretors; within the placebo group, n ¼ 6 excretors and n ¼ 5 nonexcretors.

European Journal of Clinical Nutrition Probiotics, phytoestrogen metabolism, and hormones in women MJL Bonorden et al 1639 Discussion On the other hand, a recent report showed no hormonal

Although we found a tendency toward lower E1,E1-S and T, effects in premenopausal women consuming isoflavone and higher SHBG concentrations in premenopausal equol supplements (Maskarinec et al, 2002), although, in this latter excretors, none of these differences were statistically sig- study, the number of equol excretors was unusually low, nificant. Thus, previously reported hormone differences suggesting that their method of equol detection was not very between premenopausal equol excretors and nonexcretors sensitive and many of the equol excretors may have gone (Duncan et al, 2000) were not confirmed in this study. undetected. It is also possible that isoflavone effects differ However, a major difference between the two studies should when provided as purified supplements, as opposed to those be noted. In the Duncan et al study, subjects consumed soy in soy protein isolate, as in our study. protein isolate providing between 10 and 128 mg isofla- We also found no significant differences between equol vones/day. Although the researchers intended the lowest excretors and nonexcretors with respect to dietary intake. dose to be a control against which the higher doses could be Although equol excretion has been positively correlated with compared, their lowest isoflavone dose (10 mg/day) ap- total fat and meat intake and fat-to-fiber ratio (Adlercreutz proaches current estimates of typical intake in Japan. As et al, 1991), processed and high-fat meat (Lampe et al, 1999) suggested by these researchers, it is possible that the equol and carbohydrate, fiber, plant protein, and percent of energy synthesized from this low isoflavone dose may have been as carbohydrate (Lampe et al, 1998; Rowland et al, 2000), our sufficient to alter plasma hormones in equol excretors. In findings agree with other research showing no unique our study, subjects were required to abstain from all soy dietary patterns in equol excretors (Hutchins et al, 1995; products and were exposed to negligible quantities of Duncan et al, 2000; Lampe et al, 2001). Our results showing isoflavones. Taken together, these two studies suggest that no differences in age, body weight, BMI, and cycle length equol excretors show lowered hormone concentrations only between excretors and nonexcretors also agree with pre- when exposed to small quantities of isoflavones (10 mg/day). viously published results (Duncan et al, 2000; Lampe et al, 2001). However, these results must be interpreted with caution, as they are limited by small sample sizes, homo- genous populations, and insensitive dietary assessment Table 4 Baseline characteristics, daily dietary intake and early follicular tools. plasma hormone concentrations of equol excretors and nonexcretors Consumption of probiotic capsules did not alter equol Excretors (n ¼ 13) Non-excretors (n ¼ 23) excretor status in this study even though the dose of probiotic provided was similar in dose to other studies in 7 7 Age (y) 27 6265 which the enzyme activity of the intestinal flora was shown Body wt (kg) 62.0712 66.5713 BMI (kg/m2) 22.873.8 24.073.6 to be altered (Goldin and Gorbach, 1984; Benno and MC length (days) 27.771.8 28.171.3 Mitsuoka, 1992). It is possible that probiotic consumption Energy (kJ) 570171796 611271214 altered the intestinal environment but not enough to 7 7 Protein (g) 57.7 24 60.5 16 significantly increase equol production. This lack of effect Carbohydrate (g) 160758.1 168740.1 Fat (g) 54.9717 59.5716 of isoflavone hydrolysis on increasing isoflavone-metabolite Dietary fiber (g) 13.075.3 11.974.1 excretion is consistent with another study in which 7 7 E2 (pmol/l) 164 51 164 33 consumption of a beverage containing isoflavone aglycones 7 7 E1 (pmol/l) 68.5 21 77.3 35 did not increase urinary equol excretion in comparison with E1-S (nmol/l) 20.375.8 22.078.0 T (nmol/l) 1.2570.31 1.2870.42 consumption of the same beverage containing conjugated A (nmol/l) 6.7772.4 6.7473.9 isoflavones (Richelle et al, 2002). DHEA-S (nmol/l) 4499 (3469, 5854) 4282 (3117, 5881) Only one report exists of nonexcretors gaining the ability 7 7 SHBG (nmol/l) 33.5 12 32.4 17 to excrete measurable amounts of equol following a 1-month Values are mean7s.d. except for DHEA-S, which is shown as geometric mean soy intervention (Lu et al, 1996), although this study (95% CI). MC ¼ menstrual cycle. was limited by initial labeling of equol excretors and

Table 5 Equol excretion at baseline and post-intervention (nmol/24-h)

Probiotic Placebo

Baseline Post-Intervention Baseline Post-intervention

Excretors 3281 (2440, 4412) 3130 (2046, 4790) 2465 (1255, 4844) 3537 (2217, 5641) Nonexcretors 185 (148, 232) 187 (143, 244) 124 (67, 227) 99 (59, 167)

Values are geometric mean (95% CI). Within the probiotic group, n ¼ 7 excretors and n ¼ 15 nonexcretors; within the placebo group, n ¼ 6 excretors and n ¼ 5 nonexcretors.

European Journal of Clinical Nutrition Probiotics, phytoestrogen metabolism, and hormones in women MJL Bonorden et al 1640 nonexcretors following only one dose of soymilk. A more carbohydrate diet was consumed (Bhathena et al, 1989). recent study using a soy or wheat-bran intervention was Most studies of diet and hormones have generally evaluated

unable to alter excretion from baseline (Lampe et al, 2001). T, A, E1 or E2 and have not measured DHEA-S (Ha¨ma¨la¨inen Probiotic consumption did not significantly reduce plasma et al, 1983; Goldin and Gorbach, 1994). reproductive hormones, although the bacteria provided were A strength of this study was the ethnic diversity of subjects similar in dose and bacterial species to those previously who participated, though the actual number of subjects in shown to significantly lower fecal b-glucuronidase activity in each ethnic group was small. Out of the four Asian subjects, humans (Goldin and Gorbach, 1984; Benno and Mitsuoka, three were recent immigrants and only one had lived mainly 1992). It is possible that the microflora were altered in our in North America. Interestingly, all the three Asian-born subjects, but b-glucuronidase was not changed dramatically subjects who had lived mainly in Asia were equol excretors. enough to significantly reduce enterohepatic recycling of Compared with the typical 30–40% rate of equol excretors in these hormones. A second possibility is that the strains we western populations, Adlercreutz et al (1991) reported that used were not able to decrease b-glucuronidase activity. 47% of Japanese men and women were equol excretors. A Differences between strains of the same bacterial species more recent study (Morton et al, 2002) of Japanese men and have been shown regarding resistance to stomach and bile women found the prevalence of female excretors close to acids as well as other factors that may jeopardize probiotic those found in Western populations, but the prevalence survival through the gastrointestinal tract (Bezkorovainy, among males was 58%. These results suggest that the 2001). prevalence of equol excretion in Asia is higher than that in It is possible to also evaluate the relationships the West. Perhaps the microflora necessary to metabolize among baseline characteristics, dietary factors, and repro- daidzein to equol is more predominant in individuals ductive hormones, although these analyses are limited consuming a traditional Asian diet. by the relatively small sample size and must be interpreted In conclusion, our study of healthy premeno- with caution. The inverse association between follicular- pausal women showed significant inverse correlations at

phase E2 and BMI is similar to previous reports from studies baseline between E2 and body mass index (BMI), SHBG and performed in massively obese women (Grenman et al, 1986), BMI, DHEA-S and dietary fiber, and between A and despite the fact that women in our study were within protein:carbohydrate ratio. A 2-month consumption of a normal range of BMI (18–31 kg/m2). Also consistent with Lactobacillus acidophilus and Bifidobacterium longum did previous reports (Key et al, 2001), we found an not significantly decrease plasma reproductive hormones inverse correlation between SHBG and BMI. The inverse from baseline, although trends were found toward relationship found between SHBG and BMI is likely decreased T and SHBG. We confirmed reports that equol explained by the effect of BMI increasing insulin levels with excretors do not show a unique dietary pattern and found subsequent inhibition of liver production of SHBG (Plymate that probiotic consumption did not increase equol excre- et al, 1988; Franks et al, 1991). Although the inverse tion. We did not confirm our previous finding of breast relationship between A and BMI in our cohort was only of cancer-preventive hormone profiles in premenopausal equol borderline significance, it is consistent with other reports excretors. Given that in our previous study, the women were (Grenman et al, 1986), and may be explained by the consuming at least 10 mg isoflavones/day in soy protein increased metabolic clearance rate in obesity (Kirschner isolate, these data suggest that a minimum quantity of et al, 1983). isoflavones taken in soy protein may be required to alter The inverse association between A and protein:carbohy- endogenous hormones in equol excretors. Future research drate ratio has not been reported previously. Adlercreutz et al should determine the effects of very low doses of soy (1989) have previously reported an inverse association isoflavones on plasma hormones in equol excretors and between A and carbohydrate intake, while testosterone, but nonexcretors, as well as the specific bacterial species and not A, concentrations have been inversely related to the strains that may alter intestinal microflora and increase protein-to-carbohydrate ratio in other studies (Anderson et al, equol production. 1987; Volek et al, 1997). Our data showed a trend toward this direction but results did not reach statistical significance. The mechanism for this relationship is not known; however, Acknowledgements it has been hypothesized that dietary protein may alter liver This research was supported by the General Clinical Research enzyme systems that control metabolism of steroids and Center Grant MO1-RR00400 from the National Center for consequently circulating concentrations (Anderson et al, Research Resources and the Minnesota Agricultural Experi- 1987). ment Station. We would like to thank the study participants The inverse correlation between DHEA-S and dietary fiber for their time and dedication, as well as the staff at the is not consistent with some previous reports in premeno- General Clinical Research Center, University of Minnesota. pausal (Dorgan et al, 1996) or postmenopausal (Adlercreutz We are grateful to Jennifer Nettleton, Kris Greany, and Jill et al, 1989) women, yet one study in premenopausal women Hamilton-Reeves for their help with laboratory assays and found DHEA-S to be significantly lower when a low-fat, high- Will Thomas for his help with statistical analysis. The soy

European Journal of Clinical Nutrition Probiotics, phytoestrogen metabolism, and hormones in women MJL Bonorden et al 1641 protein isolate and probiotic capsules were generously Ha¨ma¨la¨inen EK, Adlercreutz H, Puska P & Pietinen P (1983): Decrease donated by The Solae Company, St. Louis, MO, and UAS of serum total and free testosterone during a low-fat high-fibre Laboratories, Minnetonka, MN, respectively. diet. J. Steroid Biochem. 18, 369–370. Hawksworth G, Drasar BS & Hill MJ (1971): Intestinal bacteria and the hydrolysis of glycosidic bonds. J. Med. Microbiol. 4, 451–459. Hutchins AM, Slavin JL & Lampe JW (1995): Urinary isoflavonoid References phytoestrogen and lignan excretion after consumption of fermen- Adlercreutz H, Hamalainen E, Gorbach SL, Goldin BR, Woods MN & ted and unfermented soy products. J. Am. Diet. Assoc. 95, 545–551. Dwyer JT (1989): Diet and plasma androgens in postmenopausal Ingram D, Sanders K, Kolybaba M & Lopez D (1997): Case–control vegetarian and omnivorous women and postmenopausal women study of phyto-oestrogens and breast cancer. Lancet 350, 990–994. with breast cancer. Am. J. Clin. Nutr. 49, 433–442. Kelly GE, Joannou GE, Reeder AY, Nelson C & Waring MA (1995): The Adlercreutz H, Honjo H, Higashi A, Fotsis T, Hamalainen E, Hasegawa variable metabolic response to dietary isoflavones in humans. T & Okada H (1991): Urinary excretion of lignans and isoflavonoid Proc. Soc. Exp. Biol. Med. 208, 40–43. phytoestrogens in Japanese men and women consuming a Key TJ, Allen NE, Verkasalo PK & Banks E (2001): Energy balance and Am. J. Clin. Nutr. traditional Japanese diet. 54, 1093–1100. cancer: the role of sex hormones. Proc. Nutr. Soc. 60, 81–89. Anderson KE, Rosner W, Khan MS, New MI, Pang SY, Wissel PS & Kirschner MA, Samojlik E & Silber D (1983): A comparison of Kappas A (1987): Diet–hormone interactions: protein/carbohy- androgen production and clearance in hirsute and obese women. drate ratio alters reciprocally the plasma levels of testosterone and J. Steroid Biochem. 19, 607–614. cortisol and their respective binding globulins in man. Life Sci. 40, Lampe JW, Gustafson DR, Hutchins AM, Martini MC, Li S, Wahala K, 1761–1768. Grandits GA, Potter JD & Slavin JL (1999): Urinary isoflavonoid Benno Y & Mitsuoka T (1992): Impact of Bifidobacterium longum on and lignan excretion on a Western diet: relation to soy, vegetable, human fecal microflora. Microbiol. Immunol. 36, 683–694. and fruit intake. Cancer Epidemiol. Biomarkers Prev. 8, 699–707. Bezkorovainy A (2001): Probiotics: determinants of survival and Lampe JW, Karr SC, Hutchins AM & Slavin JL (1998): Urinary equol growth in the gut. Am. J. Clin. Nutr. 73, 399S–405S. excretion with a soy challenge: influence of habitual diet. Proc. Bhathena SJ, Berlin E, Judd J, Nair PP, Kennedy BW, Jones J, Smith Soc. Exp. Biol. Med. 217, 335–339. PM, Jones Y, Taylor PR & Campbell WS (1989): Hormones Lampe JW, Skor HE, Li S, Wahala K, Howald WN & Chen C (2001): regulating lipid and carbohydrate metabolism in premenopausal Wheat bran and soy protein feeding do not alter urinary excretion women: modulation by dietary lipids. Am. J. Clin. Nutr. 49, 752– of the isoflavan equol in premenopausal women. J. Nutr. 131, 740– 757. 744. Block G, Hartmann AM, Dresser CM, Carroll MD, Gannon J & Lee HP, Gourley L, Duffy SW, Esteve J, Lee J & Day NE (1991): Dietary Gardner L (1986): A data-based approach to diet questionnaire effects on breast-cancer risk in Singapore. Lancet 337, 1197–1200. design and testing. Am. J. Epidemiol 124, 453–469. Lu LJ, Lin SN, Grady JJ, Nagamani M & Anderson KE (1996): Altered Brouwers E, L’Homme R, Al-Maharik N, Lapcik O, Hampl R, Wahala K, Mikola H & Adlercreutz H (2003): Time-resolved fluoroimmu- kinetics and extent of urinary daidzein and genistein excretion in noassay for equol in plasma and urine. J. Steroid Biochem. Mol. Biol. women during chronic soya exposure. Nutr. Cancer 26, 289–302. 84, 577–588. Marteau P, Pochart P, Flourie B, Pellier P, Santos L, Desjeux JF & Cassidy A, Bingham S & Setchell KD (1994): Biological effects of a Rambaud JC (1990): Effect of chronic ingestion of a fermented diet of soy protein rich in isoflavones on the menstrual cycle of dairy product containing Lactobacillus acidophilus and Bifidobacter- premenopausal women. Am. J. Clin. Nutr. 60, 333–340. ium bifidum on metabolic activities of the colonic flora in humans. Dai Q, Franke AA, Jin F, Shu XO, Hebert JR, Custer LJ, Cheng J, Gao Am. J. Clin. Nutr. 52, 685–688. YT & Zheng W (2002): Urinary excretion of phytoestrogens and Maskarinec G, Williams AE, Inouye JS, Stanczyk FZ & Franke AA risk of breast cancer among Chinese women in Shanghai. Cancer (2002): A randomized isoflavone intervention among premeno- Epidemiol. Biomarkers Prev. 11, 815–821. pausal women. Cancer Epidemiol. Biomarkers Prev. 11, 195–201. Dorgan JF, Reichman ME, Judd JT, Brown C, Longcope C, Schatzkin Morton MS, Arisaka O, Miyake N, Morgan LD & Evans BA (2002): A, Forman M, Campbell WS, Franz C, Kahle L & Taylor PR (1996): Phytoestrogen concentrations in serum from Japanese men and Relation of energy, fat, and fiber intakes to plasma concentrations women over forty years of age. J. Nutr. 132, 3168–3171. of estrogens and androgens in premenopausal women. Am. J. Clin. Plymate SR, Matej LA, Jones RE & Friedl KE (1988): Inhibition of sex Nutr. 64, 25–31. hormone-binding globulin production in the human hepatoma Duncan AM, Merz-Demlow BE, Xu X, Phipps WR & Kurzer MS (Hep G2) cell line by insulin and prolactin. J. Clin. Endocrinol. (2000): Premenopausal equol excretors show plasma hormone Metab. 67, 460–464. profiles associated with lowered risk of breast cancer. Cancer Rao AV (1995): Effect of dietary fiber on intestinal microflora and Epidemiol. Biomarkers Prev. 9, 581–586. health. In Dietary Fiber in Health & Disease eds D Kritchevsky and C Franks S, Kiddy DS, Hamilton-Fairley D, Bush A, Sharp PS & Reed MJ Bonfield, pp 257–266. St. Paul: Eagan Press. (1991): The role of nutrition and insulin in the regulation of sex Richelle M, Merten-Pridmore S, Bodenstab S, Enslen M & hormone binding globulin. J. Steroid Biochem. Mol. Biol. 39, Offord EA (2002): Hydrolysis of isoflavone glycosides to aglycones 835–838. by beta-glycosidase does not alter plasma and urine isoflavone Gibson GR & Roberfroid MB (1995): Dietary modulation of the pharmokinetics in postmenopausal women. J. Nutr. 132, human colonic microbiota: introducing the concept of prebiotics. 2487–2592. J. Nutr. 125, 1401–1412. Rowland I, Wiseman H, Sanders T, Adlercreutz H & Bowey E (1999): Goldin BR & Gorbach SL (1984): The effect of milk and lactobacillus Metabolism of oestrogens and phytoestrogens: role of the gut feeding on human intestinal bacterial enzyme activity. Am. J. Clin. microflora. Biochem. Soc. Trans. 27, 304–308. Nutr. 39, 756–761. Rowland IR, Wiseman H, Sanders TA, Adlercreutz H & Bowey EA Goldin BR & Gorbach SL (1994): Hormone studies and the diet and (2000): Interindividual variation in metabolism of soy isoflavones breast cancer connection. Adv. Exp. Med. Biol. 364, 35–46. and lignans: influence of habitual diet on equol production by the Grenman S, Ronnemaa T, Irjala K, Kaihola HL & Gronroos M (1986): gut microflora. Nutr. Cancer 36, 27–32. Sex steroid, gonadotropin, cortisol, and prolactin levels in healthy, Tsangalis D, Ashton JF, McGill AEJ & Shah NP (2003): Biotransforma- massively obese women: correlation with abdominal fat cell size tion of isoflavones by bifidobacteria in fermented soymilk and effect of weight reduction. J. Clin. Endocrinol. Metab. 63, 1257– supplemented with D-glucose and L-cysteine. J. Food Sci. 68, 1261. 623–631.

European Journal of Clinical Nutrition Probiotics, phytoestrogen metabolism, and hormones in women MJL Bonorden et al 1642 Volek JS, Kraemer WJ, Bush JA, Incledon T & Boetes M (1997): breast cancer in Asian-Americans. Cancer Epidemiol. Biomarkers Testosterone and cortisol in relationship to dietary nutrients and Prev. 5, 901–906. resistance exercise. J. Appl. Physiol. 82, 49–54. Zheng W, Dai Q, Custer LJ, Shu XO, Wen WQ, Jin F & Franke AA Wu AH, Ziegler RG, Horn-Ross PL, Nomura AM, West DW, Kolonel (1999): Urinary excretion of isoflavonoids and the risk of breast LN, Rosenthal JF, Hoover RN & Pike MC (1996): Tofu and risk of cancer. Cancer Epidemiol. Biomarkers Prev. 8, 35–40.

European Journal of Clinical Nutrition