Human Nutrition and Metabolism - UAB

Human Nutrition and Metabolism

Long-Term Dietary Habits Affect Soy Isoflavone Metabolism and Accumulation in Prostatic Fluid in Caucasian Men1,2 Tammy E. Hedlund,3 Paul D. Maroni,* Paul G. Ferucci,* Robert Dayton, Stephen Barnes,† Kenneth Jones,† Ray Moore,† Lorraine G. Ogden,** Kristiina Wa¨ha¨la¨,‡ Holly M. Sackett,** and Karen J. Gray†† School of Medicine, Department of Pathology, *Department of Surgery, Division of Urology, **Department of Preventive Medicine and Biometrics, and ††The University of Colorado Cancer Center, The University of Colorado Health Sciences Center, Denver, CO; †Department of Pharmacology & Toxicology, The University of Alabama at Birmingham, Birmingham, AL; and ‡Department of Chemistry, The University of Helsinki, Helsinki, Finland ABSTRACT The soy isoflavones daidzein and genistein are believed to reduce prostate cancer risk in soy consumers. However, daidzein can be metabolized by the intestinal flora to form a variety of compounds with different bioactivities. In the current study, we investigated the influence of long-term dietary habits on daidzein metabolism in healthy Caucasian men (19 – 65 y old). A secondary goal was to compare plasma and prostatic fluid concentrations of 5 isoflavonoids: genistein, daidzein, equol, dihydrodaidzein, and O-desmethylangolensin. Baseline plasma levels of isoflavonoids were quantitated in 45 men by HPLC-electrospray ionization-MS. Participants then consumed a soy beverage daily for 1 wk, and post-soy isoflavonoid levels were quantitated in plasma and prostatic fluid. Equol was the only metabolite that appeared to be influenced by routine dietary habits. Stratified analyses revealed that men who had consumed ⱖ30 mg soy isoflavones/d for at least 2 y had 5.3-times the probability of producing equol than men who had consumed ⱕ5 mg/d (P ⫽ 0.014). Additionally, those men who consumed animal meat regularly had 4.7-times the probability of producing equol than men who did not consume meat (P ⫽ 0.023). Equol production was not linked to age, BMI, or the consumption of yogurt, dairy, fruit, or American-style fast food. Daidzein and its metabolites (but not genistein) were typically present at higher levels in prostate fluid than plasma (median ⫽ 4 –13 times that in plasma). In conclusion, our data suggest that the ability of Caucasian men to produce equol is favorably influenced by the long-term consumption of high amounts of soy and the consumption of meat. Last, the high concentrations of isoflavonoids in prostatic fluid increases the potential for these compounds to have direct effects in the prostate. J. Nutr. 135: 1400 –1406, 2005. KEY WORDS:

●

soy isoflavones

●

Seventh Day Adventist

Soy consumption has been linked to reduced risks for prostate cancer in worldwide epidemiologic studies (1), as well as in case-controlled studies in Chinese (2), Japanese, AfricanAmerican, and Caucasian men (3). Strong evidence is also provided by a 16-y-long prospective health study involving 13,855 Seventh Day Adventist men from California. In that study, men who consumed ⬎1 glass of soy milk/d had a 70% lower risk of prostate cancer than men who consumed none

●

prostate cancer

(4). Studies such as these have prompted researchers to try to identify which nutritional components of soy are responsible for preventing cancer. Although the saponins (5), BowmanBirke inhibitor (6), and lunasin (7), may all have health benefits, the isoflavonoids are perhaps the most promising of compounds in soy for the prevention of prostate cancer (8). The soy isoflavonoids are a family of plant-derived polyphenolic compounds, commonly known as phytoestrogens. The isoflavonoids are good candidates for anticancer agents based on their antioxidant capacity (9), their favorable influence on steroid hormone biosynthesis (10) and transport (11), and their direct antiproliferative effects on many cell types in vitro (12). Nutritional studies reveal another level of complexity to the idea that soy prevents prostate cancer. The intestinal metabolism of the isoflavone daidzein differs significantly among people. Daidzein can be converted by the gut microflora to several compounds that differ in their bioactivities (9,13–15) and half-lives in the body (16). Three known me-

1 Additional supplemental data regarding time-dependent changes of isoflavonoids in plasma and prostatic fluid are available with the online posting of this article at www.nutrition.org. 2 Supported in part by awards to T.H. from the Grove Foundation and the San Francisco Foundation and grant #M01RR00051 from the National Institutes of Health to the General Clinical Research Center at the University of Colorado Health Sciences Center. The mass spectrometer was purchased by funds from a NIH Instrumentation Grant (S10RR06487) and from the University of Alabama at Birmingham. Operation of the Mass Spectrometry Shared Facility has been supported in part by a NCI Core Research Support Grant to the University of Alabama at Birmingham Comprehensive Cancer Center (P30 CA13148). 3 To whom correspondence should be addressed. E-mail: [email protected].

0022-3166/05 $8.00 © 2005 American Society for Nutritional Sciences. Manuscript received 6 January 2005. Initial review completed 24 January 2005. Revision accepted 19 March 2005. 1400

HABITUAL DIET AND SOY ISOFLAVONE METABOLISM

tabolites are dihydrodaidzein (DHD),4 O-desmethylangolensin (O-DMA), and equol (17). Several lines of evidence suggest that the metabolite equol has increased bioactivity compared with daidzein (9,14,15), and may be especially relevant to the prevention of prostate cancer. We found equol to be 10-fold more potent than daidzein at reducing the growth of normal and malignant prostatic epithelial cells in vitro (18); research from other laboratories supports this (19,20). Additionally, equol is commonly produced in Asians consuming soy-rich diets (21,22), although only 1 in 3 Caucasians produces it after eating soy (23,24). Morton et al. (25) were the first to show that equol can be highly concentrated within the prostatic fluid (PF) of Asians, at levels far exceeding that in plasma. However, 2 important issues remain unclear. First, it is not known whether other daidzein metabolites tend to be present at higher levels in PF than plasma. Second, it is not clear whether Caucasians also tend to have higher levels of isoflavonoids in their PF than plasma because the British plasma and PF samples in Morton’s study were not taken from the same group of individuals (25). Together, these data suggest that the low incidence of prostate cancer in Asians may be related not only to their regular consumption of soy, but also to their ability to convert daidzein to equol, and to concentrate it within prostatic secretions. Habitual diet is thought to affect the predominant strains of bacteria in the gut; therefore, it is likely to influence daidzein metabolism as well. However, the dietary habits that promote equol production remain unclear. A role for low-fat, high-fiber diets was suggested by 2 studies (23), although wheat bran alone did not measurably alter equol production over a 1-mo period (24). Another hypothesis is that the regular consumption of soy itself promotes equol production, but again, 1-mo intervention studies yielded conflicting results (26,27). We suspect that these dietary interventions have to be sustained over a longer period of time before consistent changes in daidzein metabolism can be observed. The current study was designed to help us gain a better understanding of the relation between habitual diet, isoflavone metabolism, and the potential prevention of prostate cancer. Because short-term intervention studies did not alter daidzein metabolism, we chose to investigate long-term dietary habits. Soy metabolism was compared in 2 study cohorts: long-term high-soy consumers and long-term low-soy consumers. These cohorts were comprised of Seventh Day Adventist (SDA) and non-SDA men with omnivorous, vegetarian, vegan, and macrobiotic diets. The plasma concentrations of 5 soy isoflavonoids (genistein, daidzein, dihydrodaidzein, equol, and ODMA) were determined at baseline and after participants consumed a soy beverage daily for 1 wk. Our objectives were 1) to determine whether long-term consumption (ⱖ2 y) of high amounts of soy is associated with an increased probability that a person will produce equol, 2) to determine whether the concentrations of soy isoflavonoids tend to be higher in PF than plasma of Caucasian men; and 3) to investigate potential correlations between equol production and other routine dietary habits. SUBJECTS AND METHODS Study approvals. This study was approved by the Colorado Multiple Institutional Review Board, and is compliant with the Health Insurance Portability and Accountability Act of 1998. Writ-

4 Abbreviations used: DHD, Dihydrodaidzein; HS, high soy; LS, low soy; O-DMA, O-desmethylangolensin; PF, prostatic fluid; PSA, prostate specific antigen; SDA, Seventh Day Adventist.

1401

ten informed consent was obtained from all participants before entry into the study. Recruitment. SDA men were targeted to facilitate recruitment of individuals who routinely consume high amounts of soy (ⱖ30 mg isoflavones/d). Healthy men between the ages of 19 and 65 y were recruited from 8 Colorado SDA churches as well as Avista Adventist Hospital (all located within 113 km of Denver, CO). Because very few SDA men consumed low amounts of soy (ⱕ5 mg/d), additional low soy consumers were recruited from the University of Colorado Health Sciences Center through the employee email list. Dietary survey. Participants completed a dietary survey developed by our laboratory for the primary purpose of estimating monthly consumption of 34 different soy foods, including those available at local markets and through SDA churches. The survey also estimated consumption of meat, dairy, fruit, yogurt, probiotics, alcohol, antacids, other sources of phytoestrogens, and American style fast-food (e.g., cooked foods coming from McDonald’s, Burger King, Wendy’s). Several questions were also included regarding prior diagnosis of prostate disease, gastrointestinal surgeries, and antibiotic usage. Participants described their usual diet as either omnivorous (eating animal and plant products), vegetarian (no animal meat), vegan (no animal meat or dairy products), or macrobiotic (no meat, dairy, or processed foods of any kind, such as those made with flour). Based on monthly estimates of food consumption, we calculated their mean daily isoflavone intake using the manufacturers’ reported isoflavone contents or the values presented in the 1999 USDA-Iowa State University Database (28). Participants were unaware of the final inclusion and exclusion criteria to avoid possible coercion. From the 108 completed dietary surveys, 45 men met the inclusion and exclusion criteria cited below, and had consumed either low or high amounts of soy, as defined below. Inclusion criteria. Participants were at least three-fourths Caucasian (based on their grandparents’ races), were in relatively good health, had resided in the United States for at least the past 2 years, and had normal prostate specific antigen (PSA) levels as assessed with their first blood draw at the University of Colorado Hospital. PSA was quantitated by the UCH Clinical Laboratories using the Axsym RIA (Abbott Laboratories). It should be noted that although high PSA levels are associated with several prostatic diseases (including prostatitis, benign prostatic hyperplasia, and prostate cancer), serum PSA levels tend to increase with age even in healthy people. On the basis of current recommendations (29,30), we chose the following age-defined upper limits as guidelines for normal total PSA: 1.5 ng/mL for men ⬍ 40 y old; 2.5 ng/mL for men 40 – 49 y old; 3.5 ng/mL for men 50 –59 y old; and 4.5 ng/mL for men 60 – 65 y old. Exclusion criteria. Men were excluded from participation in the metabolism portion of the study if they had ever been diagnosed with any prostate disease, had any surgical resection of their intestines or stomach (because this might affect metabolism and absorption), or had taken antibiotics within the last 2 mo (this destroys the gut flora). Study populations. 1) Long-term high soy consumers (n ⫽ 25) were defined as those men who had consumed ⱖ30 mg soy isoflavones/d for at least 2 y. All of these men were SDA. 2) Long-term low soy consumers (n ⫽ 20) were defined as those men who had consumed ⱕ5 mg soy isoflavones/d for at least 2 y; 10 of these men were SDA, and 10 were not. Dietary habits were compared among SDA and non-SDA low soy consumers to determine whether there were any significant differences that could affect the outcome of our study. The only difference we noted was in alcohol consumption; none of the SDA men consumed alcohol, consistent with the recommendations of their religion. Although 7 of 10 non-SDA low soy consumers reported drinking some alcohol, this never exceeded 1 alcoholic beverage/d. Participant procedures. This study required 2 visits to the Adult General Clinical Research Center at the University of Colorado Hospital, Denver, CO. At the first clinic visit, volunteers provided blood samples for baseline measurements of isoflavonoids and PSA. Each participant was then sent home with a 1-wk supply of the soy beverage Power DreamTM described in detail below. We rationalized that a 7-d dietary intervention should provide ample time for metabolites to be produced, absorbed into the blood stream, and to accumulate in the glandular secretions of the prostate. This rationale was

1402

HEDLUND ET AL.

based in part on reports of the time-dependent appearance of daidzein metabolites in blood after soy consumption (31), and the timedependent penetration of lipid soluble compounds into PF (32). Participants consumed 1 carton (330 mL) of Power DreamTM daily and recorded the exact time each was consumed. They were also asked to take their last soy drink 5– 6 h before their 2nd clinic visit, a time at which daidzein and genistein were reported to peak in plasma after soy milk consumption (33,34). Actual time spans, however, varied from 4.5 to 13 h, which significantly affected plasma and PF concentrations in several cases (see online supplemental data).1 Participants were also asked to refrain from any sexual activity that would induce ejaculation for 4 d before their last appointment. We found that this reduced the duration and discomfort of the prostate massage, while increasing the probability of our obtaining sufficient fluid for analysis. At the second clinic visit, post-soy blood samples were obtained and a trained urologist performed the rectal prostate massage to collect PF. Soy beverage description. Power DreamTM, kindly donated by Imagine Foods/Hain Celestial Group is an organic soy milk– based beverage made from whole pressed soy beans in single-serve, readyto-drink cartons. Total isoflavone contents ranged from 42 to 60 mg/serving (⬃43% daidzein, 53% genistein, and 4% glycitein), depending on the flavor. This small variation in isoflavone content was not a concern given the extensive differences in baseline soy consumption among participants. The purpose of the soy drink was not to normalize the soy intakes of the participants, but to provide a sufficient level of isoflavones in all men to enable metabolite detection. Plasma and serum preparation. For serum PSA assays, blood was collected in VacutainersTM without additives, and analyses were performed at the University of Colorado Hospital Clinical Laboratories. For plasma isoflavonoid analyses, blood was collected in VacutainersTM with heparin:lithium, and plasma was isolated by centrifugation at 1700 ⫻ g for 10 min at 4°C. Plasma was frozen at ⫺70°C until batch processed for isoflavonoid quantitation. PF preparation and analyses. Sufficient volumes of PF were obtained from 36 of 45 participants, and were prepared as described by Morton et al. (25). In brief, expressed PF (20 –300 ␮L) was collected in 1-mL sterile centrifuge tubes and frozen at ⫺70°C until batch processed. Samples were thawed and centrifuged at 3000 ⫻ g for 15 min to remove precipitate. Duplicate aliquots of the supernatants (15 ␮L) were then digested with Helix pomatia extract and analyzed for isoflavonoid contents as described below. Isoflavonoid analyses. The quantitation of isoflavonoids was performed at the University of Alabama at Birmingham under the direction of Dr. Stephen Barnes using reversed-phase HPLC with an electrospray ionization interface and mass spectrometry [HPLC-ESIMS] (35,36). Phenolphthalein glucuronide acid, apigenin, and 4-methylumbelliferyl sulfate were added as internal controls to each plasma sample before digestion with ␤-glucuronidase and aryl sulfatase (Helix pomatia extract, Sigma Aldrich) in 150 mmol/L ammonium bicarbonate buffer, pH 5. The samples were extracted with hexane to remove neutral lipids and then with diethyl ether to recover the isoflavonoid aglucones. The extracts were evaporated to dryness and later reconstituted in 80% aqueous methanol. Aliquots were analyzed by HPLC-ESI-MS using multiple reaction ion monitoring for specific detection of daidzein, equol, O-DMA, DHD, and genistein. Analyses were performed using a Shimadzu SIL-HT gradient HPLC and a Sciex API III triple quadrupole MS. Areas under each peak were determined using the program MacQuan provided by Sciex. They were corrected by the area of the internal standard apigenin and then compared with areas for known standards prepared fresh for each analysis. Mean concentrations and SD were calculated for duplicate aliquots from each sample. The limits of quantitative detection for each isoflavonoid ranged from 8 to 30 nmol/L starting with 250 ␮L plasma, depending on the specific compound. Because of the reduced volume of PF used (15 ␮L), the limits of quantitative detection in PF were 133–500 nmol/L. Statistical analyses. Data were analyzed by a biostatistician using SPSS statistical software. Descriptive statistics are presented as the median, minimum, and maximum values for continuous variables, and the percentage of cases for categorical variables. The median was

chosen as the most appropriate measure of central tendency based on the unique nature of our data. The distributions were highly skewed and often contained multiple 0 values because not all men produced each metabolite. Transformations of the data (e.g., square root or logarithmic transformation) did not remove the skew in the data due to the large number of zero values. Thus, data are presented as medians and/or percentage of men who produced a metabolite and nonparametric methods were used for the analysis. Baseline differences between high soy and low soy consumers were compared using Mann-Whitney U tests for continuous variables (e.g., age, BMI). Differences between proportions were compared using ␹2 or Fisher Exact Tests for univariate associations (e.g., comparing the proportion of high soy and low soy consumers who produced a metabolite). Cochran-Mantel-Haenszel tests were used to examine the relation between long-term soy consumption and metabolite production adjusting for other differences in baseline diet. All reported P-values are based on 2-sided tests, and differences were considered significant at P ⬍ 0.05.

RESULTS Dietary survey. To verify that our dietary survey sufficiently distinguished the high-soy (HS) consumers from the low-soy (LS) consumers, baseline total isoflavone levels were compared. Total isoflavone levels represent the sum of concentrations of the 5 isoflavonoids analyzed. Indeed, median baseline levels in the HS group were nearly 5 times those in the LS group (P ⫽ 0.006, Table 1). Furthermore, among HS consumers, the mean isoflavone intake estimated by the survey was correlated with the total isoflavonoid concentrations in plasma at baseline (r ⫽ 0.75, P ⬍ 0.001). This suggests that our survey did not miss any major sources of these compounds from their diets. Baseline characteristics of LS and HS groups. The LS and HS groups did not differ in age, BMI, or PSA levels (Table 1). However, several significant differences were noted. Vegetarianism was more common among HS than LS consumers based on 2 criteria; HS consumers were more likely to identify themselves as vegetarians/vegans (72 vs. 0%), and were less likely to eat beef, pork, or chicken (36 vs. 90%). HS consumers were also less likely to consume dairy at least 1 time/d (24 vs. 80%). There were few or no differences between the HS and LS groups in their consumption of yogurt, fruit, or American style fast-food (data not shown). The LS and HS groups differed in alcohol consumption because none of the SDA participants reported consuming alcohol, whereas several LS non-SDA participants did (Table 1). We do not think that the modest levels of alcohol consumed by men in this study (ⱕ1 alcoholic beverage/d) would greatly affect the gut flora, and hence daidzein metabolism. However, this has not been tested directly to our knowledge. The isoflavonoid profile in one participant. As an example of the type of data collected for each individual, the isoflavonoid profile is shown from participant #18, a HS consumer (Fig. 1). Although several isoflavonoids were detectable in plasma at baseline, the concentrations increased for each compound after consumption of the soy beverage. All compounds except genistein were present at much higher levels in his PF than in his plasma. Isoflavonoid results for HS and LS consumers. Baseline and post-soy isoflavonoid concentrations were measured in the plasma and PF of all participants (Table 2). These data are expressed as median, minimum, and maximum levels because the data are not normally distributed (see Methods for details). Median concentrations of equol are zero in each category because only 9 of 45 (or 20%) of the men in our study produced detectable levels. Therefore, median concentrations


1403

TABLE 1 Comparison of demographics, baseline dietary habits, and serum prostate specific antigen levels in LS and HS consumers1

n Soy intake, mg/d Baseline plasma total isoflavonoids, nmol/L Age, y BMI, kg/m2 Baseline PSA, ng/mL Post-soy PSA, ng/mL General diet, % Vegan Vegetarian Macrobiotic Omnivorous Meat consumption, % Never Few times/wk 1 time/d 2 times/d Dairy consumption, % Never Few times/wk 1 time/d 2 times/d ⱖ3 times/d Alcoholic drinks/wk,3 % 0 1–3 4–7 8–11 ⱖ12

LS Consumers

HS Consumers

20 0.0 (0.0; 4.6) 28.6 (10.8; 843.5) 43 (28; 60) 26.8 (21.0; 37.3) 0.9 (0.4; 2.0) 0.8 (0.4; 2.5)

25 50.8 (31.2; 162.3) 139.5 (15.7; 1731.3) 46 (32; 65) 25.8 (19.9; 38.3) 1.1 (0.3; 2.5) 1.0 (0.3; 2.6)

0 0 10 90

4 68 0 28

10 55 15 20

64 28 4 4

10 10 20 45 15

4 72 12 4 8

65 25 10 0 0

100 0 0 0 0

P-value2 ⬍0.001 0.006 0.293 0.424 0.559 0.341 ⬍0.001

0.003

⬍0.001

0.006

1 Values are reported as median (min; max), unless otherwise specified. 2 Mann Whitney U-tests for continuous variables, and ␹2 tests for proportions. 3 One alcoholic beverage was defined as any of the following: a bottle of beer (330 mL), a glass of wine (125 mL), fortified wine in drinks (60 mL),

or spirits in drinks (30 mL).

of equol are also presented separately for those 9 men who produced detectable levels (Table 2). Extremely high concentrations of daidzein, DHD, and equol were present in many PF samples (Table 2). Concentrates commonly exceeded 1000 nmol/L, and in some men,

FIGURE 1 The isoflavonoid profile of a long-term HS consumer. Concentrations are shown for 5 individual isoflavonoids as well as total levels in baseline plasma, post-soy drink plasma, and post-soy drink PF. Values are means ⫾ SD of duplicate measurements.

exceeded 10,000 nmol/L; 25% of the men had PF total isoflavonoid concentrations exceeding 10,000 nmol/L, with a value as high as 47,510 nmol/L in one participant. Frequency of metabolite production. We next calculated the percentages of men with detectable levels of each compound in their plasma after consuming soy (Table 3). The presence of daidzein and genistein in all plasma samples indicates compliance with soy drink consumption by all participants. Our finding that only 20% of men produced equol supports previous reports that equol is a rare metabolite of daidzein in Caucasians (23,24). In contrast, DHD and ODMA were commonly produced. Equol was the only metabolite that appeared to be produced less frequently in LS consumers than HS consumers. However, due to several dietary differences between LS and HS soy consumers, a more careful analysis of these data was necessary. This appears in a subsequent section. An important finding is that daidzein and its metabolites, but not genistein, were typically present at higher concentrations in the PF than in plasma. To better analyze this phenomenon we calculated the PF:plasma ratios of each isoflavonoid (Table 3). In men who produced equol, median PF levels were 12.7-times that found in plasma. Similarly, DHD, O-DMA, and daidzein were all typically present at higher concentrations in PF than plasma. In contrast to all other isoflavonoids, median levels of genistein were lower in PF than in plasma. The median ratios did not differ between LS and HS consumers.

HEDLUND ET AL.

1404

TABLE 2 Isoflavonoid concentrations in plasma and PF of LS and HS cohorts1 Baseline plasma

Post-soy plasma

Post-soy PF

nmol/L Daidzein Genistein DHD O-DMA Equol Equol in producers Total Isoflavonoids

LS HS LS HS LS HS LS HS LS HS LS HS LS HS

9 (ND2; 304) 33 (ND; 590) 19 (11; 473) 85 (16; 648) ND (ND; 77) ND (ND; 201) ND (ND; 66) 15 (ND; 374) ND (ND; 66) ND (ND; 67) 463 (25; 66) 26 (9; 67) 29 (11; 843) 139 (16; 1731)

438 (66; 1459) 544 (301; 1971) 409 (41; 2489) 720 (226; 2439) 71 (ND; 331) 97 (ND; 472) 35 (ND; 156) 99 (10; 367) ND (ND; 209) ND (ND; 352) 1193 (29; 209) 123 (89; 352) 1032 (387; 4028) 1680 (1047; 3619)

1878 (181; 6231) 2779 (295; 29,847) 401 (162; 4114) 480 (196; 5139) 201 (ND; 3023) 541 (ND; 10,598) ND (ND; 1182) 250 (ND; 1229) ND (ND; 347) ND (ND; 9780) 3473 (347; 347) 2664 (353; 9780) 3711 (343; 10,643) 6655 (887; 47,510)

1 Values are presented as median (min; max); LS consumers (n ⫽ 20); HS consumers (n ⫽ 25). 2 ND, not detectable. The minimum level of detection in 0.25 mL serum ranges from 8 –30 nmol/L, depending on the specific compound. Using

15 ␮L PF, the levels of detection are 133–500 nmol/L. 3 Only two LS consumers produced equol, and of these only one had sufficient PF for analysis.

Routine diet and the production of DHD and O-DMA. There was no relation between any dietary factor and the ability of men to produce the metabolites O-DMA or DHD. However, a greater sample number would be necessary to adequately test the relation of these metabolites to dietary habits other than soy consumption. In addition, because none of the SDA men drank alcohol, and a very small number took antacids or probiotics regularly, we were unable to determine whether these factors were linked to the production of ODMA, DHD or equol. Soy and meat consumption are associated with an increased probability of making equol. Associations between soy consumption and equol production were initially tested in univariate analyses, without consideration of other dietary differences that existed between HS and LS consumers: 28%

(7 of 25) of long-term HS consumers produced equol, compared with only 10% (2 of 20) of long-term LS consumers. Thus, HS consumers had 2.8-times the probability of producing equol than LS consumers, although the difference was not significant (P ⫽ 0.134, 95% CI: 0.6, 12.0). However, when HS consumers were subdivided into moderately HS consumers (30 – 49 mg/d) and very HS consumers (ⱖ50 mg/d), differences among the groups emerged (P ⫽ 0.042). Specifically, 43% (6 of 14) of men who had the greatest isoflavone intake (50 –162 mg/d) produced equol, compared with 9% (1 of 11) of men who consumed 30 – 49 mg/d, and 10% (2 of 20) of men who consumed 0 –5 mg/d. Thus, men who consumed at least 50 mg/d had 4.4-times the probability of making equol than men who consumed ⬍50 mg/d (P ⫽ 0.017, 95% CI: 1.3–15.2 times the probability).

TABLE 3 Percentages of men with detectable levels of isoflavonoids in plasma, and higher levels in PF than plasma in LS consumers and HS consumers1 Detectable plasma levels

Isoflavonoid

Higher levels in PF than plasma

Ratio of isoflavonoids in PF vs. plasma

1002 1002 803 1003 1003 893 88 90 56 44

12.2 (12.2; 12.2) 16.0 (3.8; 51.1) 6.93 (ND4; 28.0) 13.33 (1.1; 62.7) 2.73 (1.6; 3.9) 3.83 (0.5; 10.6) 4.0 (0.5; 32.6) 5.3 (0.6; 29.5) 1.4 (0.1; 9.9) 0.7 (0.2; 2.1)

% Equol DHD O-DMA Daidzein Genistein

LS HS LS HS LS HS LS HS LS HS

10 28 90 88 95 100 100 100 100 100

1 Values are percentages or median (min; max) from a total of 20 LS consumers and 25 HS consumers. 2 Of those who had detectable plasma levels of equol, 100% had higher levels in their PF. 3 These calculations were based on those men with plasma levels ⱖ100 nmol/L because DHD and O-DMA can be detected reliably only in the

PFs of these men. This correction was not necessary for the other more abundant isoflavonoids. 4 ND, not detectable.


FIGURE 2 The production of equol by men is associated with the long-term consumption of high amounts of soy, and the regular consumption of animal meat. *Significant difference between HS and LS consumers (95% CI: 1.2–23.8; P ⫽ 0.014), as well as meat consumers and nonmeat consumers (95% CI: 1.1–20.7; P ⫽ 0.023). Significance was determined using Cochran-Mantel-Haenszel estimates of common relative risk.

Age, BMI, and baseline dietary habits were not associated with equol production in univariate analyses (data not shown). However, because baseline dietary habits differed between HS and LS consumers, we performed stratified analyses to examine the effect of soy consumption on equol production while adjusting for baseline differences in dietary habits. Soy intake and meat consumption were the only factors associated with equol production in these analyses (Fig. 2); 56% (5 of 9) of HS consumers who also consumed beef, pork, or chicken, produced equol. In comparison, equol was produced in only 12% (2 of 16) of HS consumers who did not consume meat, and 11% (2 of 18) of meat consumers who did not eat soy. Only 2 subjects were LS consumers that did not consume meat (both of these men consumed macrobiotic/ vegetarian diets) and neither of them produced equol. Using Cochran-Mantel-Haenszel estimates of the common relative risk, HS consumers (using our original definition of ⱖ30 mg isoflavones/d) had 5.3 times the probability of producing equol than LS consumers (95% CI: 1.2–23.8, P ⫽ 0.014). Furthermore, those who ate meat had 4.7 times the probability of producing equol than those who did not eat meat (95% CI: 1.1–20.7, P ⫽ 0.023). DISCUSSION Although isoflavonoid concentrations have been measured in tissue homogenates from surgical prostatectomy specimens (37), this is not possible in studies with healthy men. Morton et al. (25) were the first to use PF as a less invasive way of estimating prostatic isoflavonoid levels. PF is obtained by the rectal massage of the prostate, and is considered to be a relatively pure source of glandular secretions, unlike ejaculate. Any compound in PF must have passed through the basal and secretory epithelial cells that are prone to malignancy, and thus may better represent what these epithelial cells are exposed to. This model suggests that isoflavonoids have the potential to act in at least 2 ways. They may have direct actions on the prostatic epithelium inhibiting cellular proliferation, or once concentrated in PF, they may act as antioxidants to slow the accumulation of toxic products of oxidation in luminal secretions that may accumulate over time, especially in less sexually active men. However, it is important to

1405

recognize that the actual concentrations of isoflavonoids within prostatic epithelial cells may be different from that in plasma or PF. The development of new technology will be necessary before this can be addressed directly. The data from the current study provide several opportunities for speculation. For example, the degree to which daidzein and its metabolites are concentrated in PF did not differ significantly between LS and HS consumers (Table 3). This is important because it suggests that long-term dietary differences do not alter the mechanisms responsible for the accumulation of these putative anticancer compounds within prostatic secretions. Another interesting observation relates to the extremely high levels of daidzein, DHD, and equol we found in PF (Table 2), commonly exceeding 1 ␮mol/L and in some cases exceeding 10 ␮mol/L. The highest level of equol (9.8 ␮mol/L) is in agreement with the highest level reported (14 ␮mol/L) in the PF of Chinese men eating traditional soy-rich diets (25). In contrast, the highest PF level of daidzein in our study was 15-fold greater than the highest level observed in these Chinese men (25). The nature of this difference is not clear. One possibility is that Chinese men more commonly convert daidzein to other metabolites. Another interesting observation was that median levels of genistein were actually lower in PF than in plasma, unlike all other isoflavonoids tested. It is possible that genistein is quickly metabolized once it enters the glandular epithelial cells. Such end-organ metabolism of genistein was found with human breast cancer cells in vitro (38). Another possibility is that genistein is not actively concentrated within prostatic secretions, as are other isoflavonoids. Additional research will be required to elucidate these possibilities. A secondary goal of the current study was to determine whether long-term HS consumers had significantly lower baseline PSA levels than LS consumers. This would be consistent with the lower risks for prostate cancer reported in soy consumers (1). However, baseline PSA levels did not differ between the HS and LS groups (Table 1). Similarly, PSA concentrations did not differ before and after a 1-wk period of consuming the soy drink in either group. These data are in agreement with larger studies indicating that although soy isoflavones can effectively stabilize the rate of PSA increase in prostate cancer patients (39), they have no apparent effect on PSA concentrations in healthy men (40). Previous studies found that equol production is inversely related to fat consumption (23,24) and directly related to vegetable fiber consumption (23,24). Our current finding of an association between meat consumption and equol production may initially appear to contradict this. However, among the HS consumers that produced equol, 2 of 7 never consumed meat, and 5 of 7 consumed meat only a few times/wk. This is far less meat than is consumed by the average American, and is closer to a vegetarian diet. However, given our small sample numbers, and the fact that meat consumption was not part of our initial hypothesis, this association should be tested in future studies. It will then be of interest to determine whether a HS diet, in combination with modest amounts of meat, selects for certain strains of bacteria that are involved in the conversion of daidzein to equol. This type of diet might also act by promoting adaptive changes within existing flora to enable equol production. Future studies should also address the specific components of animal meat that may be involved. Is it the animal tissue itself, or is it contaminating factors such as hormones, antibiotics, or bacteria? In conclusion, our data suggest that long-term dietary habits can significantly affect the intestinal metabolism of daidzein. Furthermore, it seems likely that the high concentrations

HEDLUND ET AL.

1406

of isoflavonoids in PF are related to the ability of soy to reduce prostate cancer risk. LITERATURE CITED 1. Hsing, A. W., Tsao, L. & Devesa, S. S. (2000) International trends and patterns of prostate cancer incidence and mortality. Int. J. Cancer 85: 60 – 67. 2. Lee, M. M., Gomez, S. L., Chang, J. S., Wey, M., Wang, R. T. & Hsing, A. W. (2003) Soy and isoflavone consumption in relation to prostate cancer risk in China. Cancer Epidemiol. Biomark. Prev. 12: 665– 668. 3. Kolonel, L. N., Hankin, J. H., Whittemore, A. S., Wu, A. H., Gallagher, R. P., Wilkens, L. R., John, E. M., Howe, G. R., Dreon, D. M., et al. (2000) Vegetables, fruits, legumes and prostate cancer: a multiethnic case-control study. Cancer Epidemiol. Biomark. Prev. 9: 795– 804. 4. Jacobsen, B. K., Knutsen, S. F. & Fraser, G. E. (1998) Does high soy milk intake reduce prostate cancer incidence? The Adventist Health Study (United States). Cancer Causes Control 9: 553–557. 5. Philbrick, D. J., Bureau, D. P., Collins, F. W. & Holub, B. J. (2003) Evidence that soyasaponin Bb retards disease progression in a murine model of polycystic kidney disease. Kidney Int. 63: 1230 –1239. 6. Kennedy, A. R. (1998) The Bowman-Birk inhibitor from soybeans as an anticarcinogenic agent. Am. J. Clin. Nutr. 68: 1406S–1412S. 7. Galvez, A. F., Chen, N., Macasieb, J. & de Lumen, B. O. (2001) Chemopreventive property of a soybean peptide (lunasin) that binds to deacetylated histones and inhibits acetylation. Cancer Res. 61: 7473–7478. 8. Adlercreutz, H. (1995) Phytoestrogens: epidemiology and a possible role in cancer protection. Environ. Health Perspect. 103 (suppl. 7): 103–112. 9. Arora, A., Nair, M. G. & Strasburg, G. M. (1998) Antioxidant activities of isoflavones and their biological metabolites in a liposomal system. Arch. Biochem. Biophys. 356: 133–141. 10. Adlercreutz, H., Bannwart, C., Wa¨ha¨la¨, K., Makela, T., Brunow, G., Hase, T., Arosemena, P. J., Kellis, J. T., Jr. & Vickery, L. E. (1993) Inhibition of human aromatase by mammalian lignans and isoflavonoid phytoestrogens. J. Steroid Biochem. Mol. Biol. 44: 147–153. 11. Adlercreutz, H., Hockerstedt, K., Bannwart, C., Bloigu, S., Hamalainen, E., Fotsis, T. & Ollus, A. (1987) 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. 27: 1135–1144. 12. Kim, H., Peterson, T. G. & Barnes, S. (1998) Mechanisms of action of the soy isoflavone genistein: emerging role for its effects via transforming growth factor beta signaling pathways. Am. J. Clin. Nutr. 68: 1418S–1425S. 13. Markiewicz, L., Garey, J., Adlercreutz, H. & Gurpide, E. (1993) In vitro bioassays of non-steroidal phytoestrogens. J. Steroid Biochem. Mol. Biol. 45: 399 – 405. 14. Verma, S. P. & Goldin, B. R. (1998) Effect of soy-derived isoflavonoids on the induced growth of MCF-7 cells by estrogenic environmental chemicals. Nutr. Cancer 30: 232–239. 15. Dubey, R. K., Gillespie, D. G., Imthurn, B., Rosselli, M., Jackson, E. K. & Keller, P. J. (1999) Phytoestrogens inhibit growth and MAP kinase activity in human aortic smooth muscle cells. Hypertension 33: 177–182. 16. Kelly, G. E., Joannou, G. E., Reeder, A. Y., Nelson, C. & Waring, M. A. (1995) The variable metabolic response to dietary isoflavones in humans. Proc. Soc. Exp. Biol. Med. 208: 40 – 43. 17. Kelly, G. E., Nelson, C., Waring, M. A., Joannou, G. E. & Reeder, A. Y. (1993) Metabolites of dietary (soya) isoflavones in human urine. Clin. Chim. Acta 223: 9 –22. 18. Hedlund, T. E., Johannes, W. U. & Miller, G. J. (2003) Soy isoflavonoid equol modulates the growth of benign and malignant prostatic epithelial cells in vitro. Prostate 54: 68 –78. 19. Adlercreutz, H. (2002) Phyto-oestrogens and cancer. Lancet Oncol. 3: 364 –373. 20. Mitchell, J. H., Duthie, S. J. & Collins, A. R. (2000) Effects of phytoestrogens on growth and DNA integrity in human prostate tumor cell lines: PC-3 and LNCaP. Nutr. Cancer 38: 223–228. 21. Adlercreutz, H., Honjo, H., Higashi, A., Fotsis, T., Hamalainen, E., Hasegawa, T. & Okada, H. (1991) Urinary excretion of lignans and isoflavonoid

phytoestrogens in Japanese men and women consuming a traditional Japanese diet. Am. J. Clin. Nutr. 54: 1093–1100. 22. Arai, Y., Uehara, M., Sato, Y., Kimira, M., Eboshida, A., Adlercreutz, H. & Watanabe, S. (2000) Comparison of isoflavones among dietary intake, plasma concentration and urinary excretion for accurate estimation of phytoestrogen intake. J. Epidemiol. 10: 127–135. 23. Rowland, I. R., Wiseman, H., Sanders, T. A., Adlercreutz, H. & Bowey, E. A. (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. 24. Lampe, J. W., Karr, S. C., Hutchins, A. M. & Slavin, J. L. (1998) Urinary equol excretion with a soy challenge: influence of habitual diet. Proc. Soc. Exp. Biol. Med. 217: 335–339. 25. Morton, M. S., Chan, P. S., Cheng, C., Blacklock, N., Matos-Ferreira, A., Abranches-Monteiro, L., Correia, R., Lloyd, S. & Griffiths, K. (1997) Lignans and isoflavonoids in plasma and prostatic fluid in men: samples from Portugal, Hong Kong, and the United Kingdom. Prostate 32: 122–128. 26. Lampe, J. W., Skor, H. E., Li, S., Wa¨ha¨la¨, K., Howald, W. N. & Chen, C. (2001) Wheat bran and soy protein feeding do not alter urinary excretion of the isoflavan equol in premenopausal women. J. Nutr. 131: 740 –744. 27. Lu, L. J. & Anderson, K. E. (1998) Sex and long-term soy diets affect the metabolism and excretion of soy isoflavones in humans. Am. J. Clin. Nutr. 68: 1500S–1504S. 28. U.S. Department of Agriculture (2002) USDA-Iowa State University Database on the Isoflavone Content of Foods., A.R.S., Release 1.3–2002. Nutrient Data Laboratory Web site: http://www.nal.usda.gov/fnic/foodcomp/Data/isoflav/ isoflav.html [last accessed March 2001] 29. Oesterling, J. E., Jacobsen, S. J., Chute, C. G., Guess, H. A., Girman, C. J., Panser, L. A. & Lieber, M. M. (1993) Serum prostate-specific antigen in a community-based population of healthy men. Establishment of age-specific reference ranges. J. Am. Med. Assoc. 270: 860 – 864. 30. Preston, D. M., Levin, L. I., Jacobson, D. J., Jacobsen, S. J., Rubertone, M., Holmes, E., Murphy, G. P. & Moul, J. W. (2000) Prostate-specific antigen levels in young white and black men 20 to 45 years old. Urology 56: 812– 816. 31. Setchell, K. D., Brown, N. M., Desai, P., Zimmer-Nechemias, L., Wolfe, B. E., Brashear, W. T., Kirschner, A. S., Cassidy, A. & Heubi, J. E. (2001) Bioavailability of pure isoflavones in healthy humans and analysis of commercial soy isoflavone supplements. J. Nutr. 131: 1362S–1375S. 32. Charalabopoulos, K., Karachalios, G., Baltogiannis, D., Charalabopoulos, A., Giannakopoulos, X. & Sofikitis, N. (2003) Penetration of antimicrobial agents into the prostate. Chemotherapy 49: 269 –279. 33. Setchell, K. D., Faughnan, M. S., Avades, T., Zimmer-Nechemias, L., Brown, N. M., Wolfe, B. E., Brashear, W. T., Desai, P., Oldfield, M. F., et al. (2003) Comparing the pharmacokinetics of daidzein and genistein with the use of 13C-labeled tracers in premenopausal women. Am. J. Clin. Nutr. 77: 411– 419. 34. Morton, M. S., Matos-Ferreira, A., Abranches-Monteiro, L., Correia, R., Blacklock, N., Chan, P. S., Cheng, C., Lloyd, S., Chieh-ping, W. & Griffiths, K. (1997) Measurement and metabolism of isoflavonoids and lignans in the human male. Cancer Lett. 114: 145–151. 35. Barnes, S., Coward, L., Kirk, M. & Sfakianos, J. (1998) HPLC-mass spectrometry analysis of isoflavones. Proc. Soc. Exp. Biol. Med. 217: 254 –262. 36. Urban, D., Irwin, W., Kirk, M., Markiewicz, M. A., Myers, R., Smith, M., Weiss, H., Grizzle, W. E. & Barnes, S. (2001) The effect of isolated soy protein on plasma biomarkers in elderly men with elevated serum prostate specific antigen. J. Urol. 165: 294 –300. 37. Hong, S. J., Kim, S. I., Kwon, S. M., Lee, J. R. & Chung, B. C. (2002) Comparative study of concentration of isoflavones and lignans in plasma and prostatic tissues of normal control and benign prostatic hyperplasia. Yonsei Med. J. 43: 236 –241. 38. Peterson, T. G., Coward, L., Kirk, M., Falany, C. N. & Barnes, S. (1996) The role of metabolism in mammary epithelial cell growth inhibition by the isoflavones genistein and biochanin A. Carcinogenesis 17: 1861–1869. 39. Hussain, M., Banerjee, M., Sarkar, F. H., Djuric, Z., Pollak, M. N., Doerge, D., Fontana, J., Chinni, S., Davis, J. et al. (2003) Soy isoflavones in the treatment of prostate cancer. Nutr. Cancer 47: 111–117. 40. Adams, K. F., Chen, C., Newton, K. M., Potter, J. D. & Lampe, J. W. (2004) Soy isoflavones do not modulate prostate-specific antigen concentrations in older men in a randomized controlled trial. Cancer Epidemiol. Biomark. Prev. 13: 644 – 648.