Abstract
Isoflavones found in soy products are a promising class of nutrients that may have a positive effect on human health. In particular, the phytoestrogen metabolite equol is associated with a reduced risk of developing female hormone-related diseases. However, the effect of equol on estrogen remains unclear. Equol can modify blood and urinary estradiol (E2) levels. The aim of this cross-sectional study was to examine the associations between urinary estrogen levels, equol levels, and equol production status in Japanese women. We analyzed urine samples from 520 women by gas chromatography-mass spectrometry. Urinary E2 and 4-hydroxylated E2 levels were higher in equol producers (EQP) than in non-EQPs (P < 0.0001 and P=0.00112, respectively). After adjusting for age and tobacco use by analysis of covariance, the association remained significant (β = 0.299, P < 0.0001). Analysis of covariance demonstrated that equol levels in urine were also positively associated with urinary E2 (β = 0.597, P < 0.0001). The log equol concentration showed a significant, but moderate, negative association with the serum E2 concentration (β = − 0.0225, P = 0.0462). Our findings suggest that equol may promote urinary E2 excretion and modify blood E2 levels in women.
Similar content being viewed by others
Introduction
The relationship between nutrients and cancer is an important public health issue. Estrogen can promote the development of several cancers, such as endometrial cancer and breast cancer; therefore, phytoestrogens are considered to be promising nutrients for cancer prevention. Indeed, the incidence of breast cancer in Asian countries is lower than that in Western countries1, which has led researchers to investigate lifestyles and dietary habits among Asian populations. The consumption of large amounts of soybeans or soybean products shows a significant association with a reduced risk of breast cancer2. Soybeans contain isoflavones, including daidzein and genistein3, which exhibit weak estrogenicity or anti-estrogenicity.
While daidzein is a major component of soy isoflavones, it can be metabolized to equol or O-desmethylangolensin (O-DMA) by human intestinal bacteria4,5. A previous study identified NADP(H)-dependent daidzein reductase from Lactococcus, converting daidzein to dihydrodaidzein (DHD)6. Moreover, DHD reductase, DHD racemase, tetrahydrodaidzein reductase were involved in equol production7,8. O-DMA is produced from DHD via C-ring cleavage by several bacteria9, but the enzyme remained unidentified. Individuals who can produce equol are known as equol producers (EQPs)10. Equol and daidzein are present at comparable levels in both the urine and blood of EQPs11,12 while in non-EQPs, the levels of equol are much lower than those of daidzein without association13. These features have prompted researchers to explore the prevalence of EQP status. The percentage of EQPs in Asian populations is 50–60%14, which was determined in sera by reversed-phase high-performance liquid chromatography-multiple reaction ion monitoring-mass spectrometry method with detection limit at 0.5 ng/mL. On the other hand, the percentage in the Western world is approximately 25–30%15, which was examined in urine by ultrahigh-pressure liquid chromatography with a detection limit at 2.93 nM. EQP status is related to age and dietary habit of taking isoflavone, but not sex16. Smoking is negatively associated with EQP status in women17.
Equol has the strongest estrogenic18,19 and antioxidative effects of all the isoflavones10. It has also been reported that equol was negatively associated with the risk of breast cancer in postmenopausal4 and premenopausal20 women, the incidence of hot flashes in perimenopausal and postmenopausal women21, and positively correlated with bone mineral density preservation22. These effects are likely mediated by estrogen receptors23,24. A study showed that blood estradiol (E2) levels are inversely correlated with urinary equol levels in postmenopausal women25, but it is still controversial26,27. In premenopausal women, isoflavone-rich soy products seemed inversely correlate with follicle-stimulating hormone (FSH) and luteinizing hormone (LH) and might mask the relationship between equol and E227,28. Potential effects of equol on circulating hormones in reproductive women may influence their reproduction ability29. However, the mechanism of how equol reduces blood E2 levels remains unclear.
If urinary equol is negatively correlated to blood estradiol levels, there are several possibilities that equol reduces blood E2 levels. Equol may enhance metabolism of estrogens to hydroxylated estrogens, or equol may promote urinary excretion of estrogens. Therefore, we considered urinary and serum E2 as the main outcomes of such effects. In this observational study, we evaluated the associations between the concentrations of equol, estrogens, and their metabolites in the urine of postmenopausal and premenopausal Japanese women. In addition, the association between serum estradiol and urinary equol excretion was analyzed in select samples.
Methods
Study participants
In the original nationwide study, 13,910 participants aged 20–70 years were recruited through medical check-ups in 11 prefectures in Japan between 2000 and 200130. The participants’ age, parity, smoking habits, present and past disease histories and menstrual status (regular cycles, irregular cycles, menopause, experienced gynecological surgery) were recorded using a self-report questionnaire. Body mass index (BMI) was obtained from participants in one study area, Kyoto (n = 109). A regular cycle was defined as menstrual cycles of 25–38 days. An irregular cycle was defined as menstrual cycles that were not regular (i.e., 25–38 days) in the previous 3 months. Menopause was defined as no menstruation for 12 months. Those who experienced gynecological surgery were categorized to eliminate the potential effects on estrogen levels. The questionnaire did not include anthropometry, dietary intakes such as soybean consumption, and length of menstrual cycle. Urine samples were collected at a time of no menstruation and were stored at − 30 °C for later analysis at the Kyoto University Human Specimen Bank31,32. The protocols for the specimen bank and for this study were in accordance with Ethical Guidelines for Medical and Health Research Involving Human Subjects of Japan and approved by the Ethics Committee of Kyoto University Graduate School of Medicine and Faculty of Medicine and Hospital (Latest approval number R1478-4 on May 19th, 2020, ‘Human exposure monitoring and risk assessment’). Written informed consent was obtained from all participants before sample collection.
For this study, 520 non-pregnant women from five prefectures recruited between November 2000 and December 2001 were randomly selected. A total of 100–110 samples per prefecture from the original participant group were analyzed for urinary isoflavone levels. No stratification and blocking for random selection were employed. All participants were of Asian ethnicity. Urinary isoflavones are biomarkers of blood isoflavone concentrations17. Urinary E2 levels were also analyzed to evaluate the association between equol and estrogen levels in the urine in a cross-sectional manner. The levels of estrogens other than E2 were measured in selected samples to evaluate the possible relationship between equol and estrogens. To confirm the relationship between equol and serum E2, an additional 100 paired urine and serum samples from the same original population were analyzed. Blood sampling was conducted by venipuncture in the morning under fasting conditions (more than 6 h) at the same institutions as urine samples. Serum was separated by centrifugation 30 min after sampling at room temperature and divided into polypropylene cryotubes. Serum samples were stored in –80 °C freezer.
Analysis of isoflavones and estrogens
Target analytes in the urine were analyzed using a gas chromatography-mass spectrometer (model 6890GC/5973MSD; Agilent Technologies Japan Ltd., Tokyo, Japan) according to the extraction method described by Grace et al.33. Briefly, 0.5 mL of urine was mixed with 2,500 U of a glucuronidase/sulfatase solution and incubated for 12 h at 37 °C. The sample solution was then loaded onto a Sep-Pak Plus C18 solid-phase column (Waters, Milford, MA) and eluted with 3 mL of a 1:1 acetonitrile and ethyl acetate mixture. The eluates were dried under a nitrogen stream. Residues were derivatized to trimethylsilyl ethers in 100 μL of methyl t-butyl ether and 50 μL of N,O-bis(trimethylsilyl)trifluoroacetamide with 1% trimethylchlorosilane at 60 °C for 1 h. Derivatives were separated on a DB-5MS column (length 30 m, inner diameter 0.25 mm, film thickness 0.25 µm) (Supplementary Fig. 1), ionized by electron impact ionization, and quantified by single-ion monitoring using a mass spectrometer. We analyzed five estrogens (estrone (E1), estradiol (E2), estriol (E3), 4-hydroxy E1 (4-OHE1), and 4-hydroxy E2 (4-OHE2) and two isoflavones (daidzein and equol). 4-OHE1 is metabolized from E1, and 4-OHE2 is metabolized from E2. The detection limits for the analyses were 0.1 (4-OHE2) to 1 (E2) ng/mL for the estrogens, 20 ng/mL for equol, and 28 ng/mL for daidzein. Quantification was conducted with internal standard (d4-daidzein) and nine points calibration curves (0.1 ng/mL to 1000 ng/mL). Linearity of the calibration curves (r) ranged from 0.991 (E2) to 0.998 (daidzein). Samples that contained analytes over the range of calibration curve were analyzed again after dilution by distilled water. In every 10 samples, quality control sample was analyzed to confirmed validity of the analysis. Coefficient of variation in quality control samples were less than 9% for all analytes (N = 52). Serum E2 and progesterone were analyzed by ELISA in accordance with the manufacturer’s instructions (ADI-900-174 and ADI-900-011, Enzo Life Sciences, Inc., Farmingdale, NY).
To validate the stability of analytes in archived samples, urine samples obtained within one month (n = 30) were analyzed in the same method. We confirmed that distribution and trend of analytes were comparable between long-archived samples and recently collected samples. ANOVA was used to compare the mean estrogen and isoflavone concentrations in the urine samples (ANOVA: P > 0.05 for all analytes). The equality of variances for analytes in the archived samples and the recently collected samples was assessed by F-test (P > 0.05 for all analytes). Similarly, serum samples obtained within one week (n = 15) were simultaneously analyzed for E2. ANOVA and F-test showed no significant differences between recent and archived serum samples (P = 0.792 and 0.885, respectively). These results indicated that the analytes did not degrade in the archived samples.
Values under the detection limits were recorded as half of the detection limit. The urinary concentrations of the estrogens and isoflavones were normalized to urinary creatinine concentrations. Urinary creatinine concentration was assayed by using enzymatic method34. The individual differences were large for these biological concentrations, which ranged by several orders of magnitude. The concentrations were not normally distributed. To enable a statistical analysis comparison, the urinary and serum concentrations were transformed into common logarithmic values. The 4-OHE1/E1, 4-OHE2/E2, and E2/E1 ratios were also calculated to evaluate estrogen metabolism35,36. EQPs were defined according to the criteria proposed by Ideno et al. for postmenopausal and premenopausal women37, which are based on urinary daidzein and equol levels: log (equol/daidzein) ≥ − 1.42 indicates EQP status, and log (equol/daidzein) < − 1.42 indicates non-EQP status. Participants who had a progesterone concentration greater than 1 ng/mL were classified as the luteal-phase group.
Statistical analysis
Differences in the means of potential outcome variables (log urinary and serum E2 levels) and proportions of other characteristics between the EQP and non-EQP groups were examined using Student’s t-test and Fisher’s exact test, respectively (two-tailed). The urinary equol excretion was evidently different between the EQP and non-EQP groups; thus, we could hypothesize that equol might affect E2 metabolism. The urinary equol concentrations were then used as the exposure variable (independent variable). To adjust the potential confounders, we conducted multivariable regression analyses. The analyses include a categorical variable (three smoking statuses and four menstrual statuses), and we employed analyses of covariance. Analyses of covariance were used to examine the correlation between urinary estrogen and equol levels, adjusted for age and smoking status and other variables. In a previous report, estrogen was associated with age and isoflavones, while smoking was also associated with estrogen and isoflavones38. Although an association between EQP statuses and overweight was controversial39, a report has shown a moderate correlation between BMI and urinary estrogens in postmenopausal Japanese women40. For these reasons, we chose them as covariates. Stepwise selection of variables was not employed to avoid overfitting41. We conducted stratified analyses for several variables (menopausal status, disease history, and menstrual cycle). This was because the potential effects of those stratified variables were not clear and modeling in the analysis of covariance was difficult, such as parameterization of variables, interaction terms thereof, and so on. Records with missing values were excluded from the analyses. JMP Pro (ver. 14; SAS Institute, Cary, NC) was used to perform these calculations. The alpha level for all tests was 0.05. P values were expressed with three significant figures unless they were less than 0.0001. We calculated the statistical power of the analysis of covariance using pwr.f2.test included in the R pwr package. When the effect size (Cohen’s f2) was 0.025 in the three-parameter model with 80% power (alpha = 0.05, two-tailed), the required sample size was 436. We analyzed a slightly larger number of samples to ensure statistical power.
Results
Characteristics and variables
Table 1 shows the characteristics of the study subjects and their urinary estrogen and isoflavone levels. The average age of the participants was 49.2 years (range: 23–74 years). Approximately half of the participants reported having regular menstrual cycles, and 11.6% were smokers. Some of the participants had current disease and/or a history of disease: 2 (present history: 1) reported endometriosis, 10 (2) myoma of the uterus, 1 (1) diabetes mellitus, 25 (11) thyroid disease, 5 (5) hypertension, 5 (2) cancer, and 76 (20) other diseases. In addition, 22 of the participants had multiple diseases: 18 had two diseases, and 4 had three diseases. Based on a cut-off level of log (equol/daidzein) ≥ − 1.42, 254 of the 520 participants (48.85%) were classified as EQPs. The participants were divided into EQP and non-EQP groups to evaluate possible differences in characteristics. Fisher’s exact test was used to examine differences in smoking, menstrual cycle, and disease history between the EQP and non-EQP groups. Smoking was significantly negatively associated with EQP status (P = 0.00152), as reported earlier for the same population17.
The urinary estrogen concentration is expressed as μg/gCr, as described previously42, and was further common log-transformed to normalize the distributions. The urinary concentration of the estrogen metabolite 4-OHE1 was typically higher than that of 4-OHE2 within total samples. Student’s t-test was used to examine differences in estrogen levels and isoflavone levels between the EQP and non-EQP groups. Regarding the isoflavone concentrations, the log equol (μg/gCr) and log equol/daidzein values were significantly higher in the EQP group than in the non-EQP group (P < 0.0001). Daidzein levels were lower in the EQP group (P < 0.0001), likely owing to the increased metabolism of daidzein to equol by the intestinal microflora.
Next, Student’s t-test was used to examine the urinary estrogen levels. Log urinary E2 (μg/gCr) was significantly elevated in the EQP group (P < 0.0001) compared with the non-EQP group, while the log urinary E1 and log urinary E3 levels did not differ between the groups. As a result, the E2/E1 ratio was also increased in the EQP group (P < 0.0001). Regarding estrogen metabolism, the log urinary 4-OHE2 value (μg/gCr) was significantly higher in the EQP group than in the non-EQP group (P = 0.00112), while the 4-OHE2/E2 ratio was lower in the EQP group (P = 0.0385). Log serum E2 (pg/mL) was significantly higher in the non-EQP group than in the EQP group (P = 0.00517).
Estrogen and isoflavone levels are also shown in each subgroup (Table S1). Age, menstrual cycle and smoking habits showed significant differences in urinary equol between subgroups (P = 0.0128, 0.0225 and 0.00577 by ANOVA, respectively). Urinary E2 levels were different among subgroups of smoking habits (P = 0.00179 by ANOVA). Hence, the effects of these variables were adjusted in following analyses.
Analysis of the association between urinary estrogen and equol levels by analysis of covariance
The difference in urinary equol excretion between the EQP and non-EQP groups prompted us to hypothesize that equol could affect E2 metabolism. The urinary equol concentration was considered to be a more reliable indicator than EQP status. Smoking is significantly associated with log equol (μg/gCr), and age affects estrogen levels. Hence, analyses of covariance were employed to examine the correlation between urinary E2 and equol, including the effects of age and smoking (Table 2). Current smokers had lower urinary E2 than non-smokers (β = − 0.0908, P = 0.00985), and ex-smokers had higher E2 (β = 0.144, P = 0.00434). As in the bivariate analysis, the log equol value was significantly positively associated with the log urinary E2 (μg/gCr) value (β = 0.597, P < 0.0001), with an R2 value of 0.751. The parameter estimate corresponds to 5.74-fold change in urinary E2 by IQR increase in equol concentration (17.6 μg/gCr to 329 μg/gCr).
Regarding the other urinary estrogens, log equol (μg/gCr) was significantly positively associated with log urinary 4-OHE2 (μg/gCr) value (β = 0.441, P < 0.0001) (Supplementary Table S2). After adjustment for the covariables (age and smoking), the log urinary E1 and log urinary 4-OHE1 (μg/gCr) became significant (β = 0.0697, P = 0.0277 and β = 0.197, P = 0.0209, respectively), but their R2 values were low (Supplementary Table S2).
Additional analyses were conducted to examine the influence of substituting values that were less than the limit of quantification (LOQ). Instead of substituting with half the LOQ, substituting with zero or with the LOQ was employed in extreme cases. The results showed similar trends as observed in the original analysis (Supplementary Table S3). The number of samples varied owing to the substitution, particularly when zero was substituted (logarithms were not applicable), but the effect sizes (beta parameter of equol) were almost the same as in the original analyses.
The association between EQP status and estrogens was also investigated by analysis of covariance, including age and smoking (Table 3, Supplementary Table S4). EQP status also showed a significant positive association with log urinary E2 (μg/gCr) (β = 0.299, P < 0.0001, R2 = 0.304) and log urinary 4-OHE2 (μg/gCr) (β = 0.233, P = 0.000435, R2 = 0.197).
Association between urinary equol and serum E2
To further examine the effect of increased excretion of E2 on serum E2 level, we analyzed an additional 100 paired serum and urine samples (Table 1). In the analysis of covariance, the log equol concentration was significantly negatively associated with the serum E2 levels (Table 4). However, the association was moderate (β = − 0.0225, P = 0.0462) compared with the association observed for urinary E2 excretion. IQR increase in urinary equol (4.51 μg/gCr to 189 μg/gCr) was associated with 0.919-fold lower serum E2 level.
Verification of potential confounding effects by daidzein, menstrual status, and disease history
Equol levels are dependent on soy product consumption and the presence of intestinal bacteria. Urinary equol might be related to other factors that influence estrogen levels. To adjust for potential confounders related to soy intake, urinary daidzein levels were incorporated into the analysis of covariance model (Supplementary Table S5). As Supplementary Table S5 shows, the log equol (μg/gCr) value remained significantly positively associated with log urinary E2 (β = 0.599, P < 0.0001), even after adjustment based on the log daidzein value. It was possible that some potential EQPs were not identified because of a low soybean intake at the time of sampling, resulting in their misclassification as non-EQPs. To eliminate this possibility, those who had substantial soybean intake (urinary daidzein > 1000 μg/gCr, N = 327) were extracted. Analysis of covariance in the subpopulation showed comparable positive association between log urinary E2 and log urinary equol (β = 0.613, P < 0.0001, Supplementary Table S6).
The effect of equol on urinary E2 excretion may depend on menstrual status. Therefore, we performed a stratified analysis of premenopausal women and postmenopausal women (Supplementary Table S7). As shown in Supplementary Table S7, there was a positive association between log urinary equol and log urinary E2 levels in both premenopausal and postmenopausal women (β = 0.550, P < 0.0001; and β = 0.651, P < 0.0001, respectively). This analysis was also conducted for urinary E1 and E3, but there were no significant effects of equol on those excretions irrespectively of menstrual statuses (Supplementary Table S7).
Premenopausal women who were available for serum progesterone (N = 115) were classified into luteal phase (N = 63) and follicular phase (N = 52) groups based on the progesterone levels (Supplementary Table S8). There was a significant positive association between log urinary E2 and log equol in both groups (P < 0.0001 for both groups, Supplementary Table S8).
BMI has been suggested for association with estrogen metabolism39. In addition, previously, residential area was moderately associated with EQP statues17, and it could affect the association between urinary equol and E2. Effect of BMI on the association between urinary equol and E2 was examined among participants in one study area (N = 109, Supplementary Table S9). BMI was not associated with urinary E2 while log urinary equol concentration remained significant for urinary E2 (β = 0.631, P < 0.0001).
To eliminate possible bias from disease status on the association between the log urinary equol and log urinary E2 levels, analysis of covariance was performed for the participants without a self-reported history of disease (N = 363) (Supplementary Table S10). There was a significant positive association between the log equol and log urinary E2 values (β = 0.588, P < 0.0001), as was observed in the full dataset.
Finally, we conducted analysis of covariance for urinary E2 including all the potential factors (disease history, menstrual statuses, daidzein, smoking habit and age) (Table 5). The association between urinary E2 and equol was still significant (β = 0.602, P < 0.0001).
Discussion
In this observational study, urinary E2 excretion was more strongly positively associated with equol levels in the urine than other estrogens. This association was consistent with EQP status. The blood E2 levels were analyzed in this study, and there was a moderate negative association between serum E2 and urinary equol excretion. Physiologically, increased E2 excretion could correlate with decreased E2 levels in the blood. The majority of the E2 in the blood is bound to sex hormone-binding globulin (SHBG) and albumin, while a small portion of E2 is unbound and bioavailable. It has been previously shown that blood E2 levels are negatively associated with urinary equol levels in postmenopausal women25. However, we found that the concentration of daidzein, the precursor of equol, did not correlate with urinary E2 levels. This suggests that equol may play an important physiological role, and that equol-producing microflora may affect the risk of developing an estrogen-related disease. In addition, 4-OHE2 excretion was increased in EQPs. 4-OHE2 forms reactive quinone species and induces DNA damage35. Enhanced elimination of 4-OHE2 by equol may play a protective role in hormone-related carcinogenesis. On the other hand, as mentioned before, lowered estradiol levels in premenopausal women may affect their reproduction ability29. To verify the mechanism between equol and estrogens, intervention study will be needed to examine the causal relationship.
Although the mechanism by which equol affects E2 excretion was not investigated in this study, our findings still provide valuable insights. Urinary equol levels were more positively associated with urinary E2 and 4-OHE2 than with other estrogens. EQPs exhibited higher levels of E2 and 4-OHE2 excretion and lower 4-OHE2/E2 ratios compared with non-EQPs. Hence, 4-hydroxylation of E2 was not a dominant factor, but promoted clearance of E2 could be. Our findings demonstrated possible associations of varying strength between equol and urinary E2, 4-OHE2, and E1. Similarly, blood estrogens, and E2 in particular, bind to SHBG with different affinities43. Equol also binds to SHBG and non-competitively inhibits E2 binding to SHBG44. This could increase the fraction of free E2 in the blood, thereby making it available for metabolism45. These hypotheses should be investigated in future studies to provide biological evidence for the effects of equol. Effects of equol on estrogen syntheses such as 17beta-hydroxysteroid dehydrogenases (17β-HSD) or CYP19 aromatase are another possibility. However, in vitro study showed no significant effect of equol on aromatase activity46, and there has been no report on an inhibitory effect of equol on 17β-HSD47.
This was an observational study, and potential biases could have affected some of the associations that we observed. Our dataset did not include details of lifestyles, dietary habits, and BMI (except for one study area). These unmeasured variables may confound with urinary equol, E2 excretions and serum E2. While several studies have explored the relationship between various lifestyle factors and equol levels, few of these factors have been shown to have a significant association with urinary equol excretion. One exception is smoking, which is positively associated with EQP status17 and urinary E238. Approximately 10% of the women included in this study smoked, which is consistent with a previously published national survey48. Multivariable analysis including smoking status showed consistent association between urinary E2 and equol. Therefore, the smoking characteristics of our study sample were standard. Nagata et al. demonstrated that women who consumed more dairy products had lower equol excretion than women who did not49. In addition, one report showed familial segregation of EQP, suggesting shared genetic or environmental factors within families50. Furthermore, equol levels are dependent on the levels of precursor isoflavones, i.e. soy consumption. The dietary habits of the participants were not recorded for this study. In particular, dairy products contain estrogens and account for 60–80% of dietary estrogen intake51,52. While it is possible that the relationship between EQP status and dairy consumption affects estrogen levels in humans, our findings suggest that this is not likely, because, if EQPs consumed fewer dairy products, their dietary estrogen intake and urinary estrogen excretion would be lower than those of non-EQPs. The effects of soy consumption were also evaluated using an analysis of covariance model and stratified analysis (Table5, Supplementary Table S5 and S6), which showed that daidzein did not affect the association between equol and E2 excretion. In addition, after menopause, blood estrogen levels decrease significantly. Age does not influence EQP status, and in a subgroup analysis of postmenopausal women and premenopausal women, the association between E2 and equol was unchanged. In relation to menstrual status, in premenopausal women, regularity and length of menstrual cycle can vary among individuals. In this study, samples from premenopausal women were divided into luteal and follicular phase groups; the associations between urinary E2 and equol were similar in both groups. But in future study, equol effects on menstrual cycle via FSH and LH53 are worth investigation. Finally, disease status may modify estrogen levels because of the pathophysiology of the disease or treatment. However, analysis of samples from women with and without a history of disease showed similar associations (Supplementary Table S10). Indeed, while it is likely that multiple, as yet unidentified, factors determine EQP status, approximately 50% of the participants in this study were EQPs, which is consistent with previous reports54, suggesting that there was no selection bias in our study population. Furthermore, the observed association between urinary E2 and equol remained high even after adjustment for several variables, so any unidentified confounder, if it exists, would have to be strongly associated with urinary E2. Nonetheless, future study with comprehensive background information (BMI, food consumption, etc.) can help the confirmation of this finding.
In this study, urine samples were used to assess isoflavone levels. Equol excretion might be promoted by kidney conditions and affect E2 levels in urine. Generally, the level of chemicals in the urine varies owing to dilution, and normalizing to creatinine levels is not always an effective approach. Although consumption of soy products may differ from day to day, the biological half-life of equol is ca. 8 h, and equol concentrations in the urine are strongly correlated with blood equol concentrations10. Hence, urinary equol level is considered a valid biomarker to distinguish with non-EQP and EQP. On the other hand, biological half-life of daidzein is shorter than that of equol and urinary daidzein would partially reflect soy product consumption13,55. Nevertheless, presence of daidzein in urine indicated short-term exposure to soy products, and influence on direct interaction for equol-E2 association is less likely. However, detecting equol in the urine is difficult if people do not consume soybean products. Our full sample set analysis might underestimate the proportion of potential EQPs. The stratified analysis according to urinary daidzein concentration produced similar results as the total sample analysis (Supplementary Table S6). Therefore, the definition of EQP status based on urinary equol and daidzein concentrations is unlikely to affect the association between urinary E2 and equol.
The urinary E2 levels were analyzed in all participants, as the primary outcome in this study was urinary E2 excretion, but the levels of other estrogens were also analyzed in selected participants. Serum samples were analyzed in 100 participants. Thus, the statistical power differed among the target analytes, which may have affected the results. However, urinary E1, 4-OHE1, and E3 levels were very similar between EQPs and non-EQPs, and analyzing more samples would not have changed the results. EQPs were defined by the equol/daidzein ratio in the urine34, as equol concentration is affected by intake of its precursor, daidzein. Using a different definition of EQP may have resulted in different associations with estrogens. The analysis of covariance model showed that urinary E2 levels correlated with urinary equol levels in a dose-dependent manner. In addition, E1 and 4-OHE1 levels showed statistically significant associations with urinary equol levels, even in a limited number of samples (Supplementary Table S2). Their β parameters and model fitness values were lower than those of E2 or 4-OHE2, indicating that E2 is the primary target of the effects of equol. The further studies are needed to clear the causal effect of equol with internal E2 levels.
In conclusion, our findings suggest that equol may promote urinary E2 excretion and modify moderately blood E2 levels in women.
Data availability
All data generated or analyzed during this study are included in this published article.
References
Michelle, D. A., Jaclyn, M. D., William, F. A., Susan, S. D. & Louise, A. B. Global trends in breast cancer incidence and mortality 1973–1997. Int. J. Epidemiol. 34, 405–441 (2005).
Messina, M. & Messina, V. Increasing use of soyfoods and their potential role in cancer prevention. J. Am. Diet Assoc. 91, 836–840 (1991).
Coward, L., Barnes, N. C., Setchell, K. D. R. & Barnes, S. Genistein, daidzein, and their beta-glycoside conjugates: antitumor isoflavones in soybean foods from American and Asian diets. J. Agric. Food Chem. 41, 1961–1967 (1993).
Goodman, M. T. et al. Urinary phytoestrogen excretion and postmenopausal breast cancer risk: the multiethnic cohort study. Cancer Prev. Res. (Phila). 2, 887–894 (2009).
Rowland, I. et al. Bioavailability of phyto-oestrogens. Br. J. Nutr. 89, S45-58 (2003).
Shimada, Y. et al. Cloning and expression of a novel NADP(H)-dependent daidzein reductase, an enzyme involved in the metabolism of daidzein, from equol-producing Lactococcus strain 20–92. Appl. Environ. Microbiol. 76, 5892–5901 (2010).
Shimada, Y. et al. Identification of two novel reductases involved in equol biosynthesis in Lactococcus strain 20–92. J. Mol. Microbiol. Biotechnol. 21, 160–1672 (2011).
Schröder, C. A. et al. Identification and expression of genes involved in the conversion of daidzein and genistein by the equol-forming bacterium Slackia isoflavoniconvertens. Appl. Environ. Microbiol. 79, 3494–3502 (2013).
Hur, H. et al. Isolation of an anaerobic intestinal bacterium capable of cleaving the C-ring of the isoflavonoid daidzein. Arch. Microbiol. 178, 75 (2002).
Setchell, K. D., Brown, N. M. & Lydeking-Olsen, E. The clinical importance of the metabolite equol-a clue to the effectiveness of soy and its isoflavones. J. Nutr. 132, 3577–3584 (2002).
Talbot, D. C. et al. Monoclonal antibody-based time-resolved fluorescence immunoassays for daidzein, genistein, and equol in blood and urine: application to the Isoheart intervention study. Clin. Chem. 53, 748–756 (2007).
Setchell, K. D. & Cole, S. J. Method of defining equol-producer status and its frequency among vegetarians. J. Nutr. 136, 2188–2193 (2006).
Mathey, J. et al. Concentrations of isoflavones in plasma and urine of post-menopausal women chronically ingesting high quantities of soy isoflavones. J. Pharm. Biomed. Anal. 41, 957–965 (2006).
Akaza, H. et al. Comparisons of percent equol producers between prostate cancer patients and controls: case-controlled studies of isoflavones in Japanese, Korean and American residents. Jpn. J. Clin. Oncol. 34, 86–89 (2004).
Guadamuro, L. et al. Bacterial communities and metabolic activity of faecal cultures from equol producer and non-producer menopausal women under treatment with soy isoflavones. BMC Microbiol. 17, 93 (2017).
Iino, C. et al. Daidzein intake is associated with equol producing status through an increase in the intestinal bacteria responsible for equol production. Nutrients 11, 433 (2019).
Liu, W., Tanabe, M., Harada, K. H. & Koizumi, A. Levels of urinary isoflavones and lignan polyphenols in Japanese women. Environ. Health Prev. Med. 18, 394–400 (2013).
Jackson, R. L., Greiwe, J. S. & Schwen, R. J. Emerging evidence of the health benefits of S-equol, an estrogen receptor beta agonist. Nutr. Rev. 69, 432–448 (2011).
Setchell, K. D. et al. S-equol, a potent ligand for estrogen receptor beta, is the exclusive enantiomeric form of the soy isoflavone metabolite produced by human intestinal bacterial flora. Am. J. Clin. Nutr. 81, 1072–10729 (2005).
Duncan, A. M. et al. Premenopausal equol excretors show plasma hormone profiles associated with lowered risk of breast cancer. Cancer Epidemiol. Biomarkers Prev. 9, 581–586 (2000).
Crawford, S. L. et al. The impact of dose, frequency of administration, and equol production on efficacy of isoflavones for menopausal hot flashes: a pilot randomized trial. Menopause 20, 936–945 (2013).
Tousen, Y. et al. Natural S-equol decreases bone resorption in postmenopausal, non-equol-producing Japanese women: a pilot randomized, placebo-controlled trial. Menopause 18, 563–574 (2011).
Muthyala, R. S. et al. Equol, a natural estrogenic metabolite from soy isoflavones: convenient preparation and resolution of R- and S-equols and their differing binding and biological activity through estrogen receptors alpha and beta. Bioorg. Med. Chem. 12, 1559–1567 (2004).
Mueller, S. O. et al. Phytoestrogens and their human metabolites show distinct agonistic and antagonistic properties on estrogen receptor alpha (ERalpha) and ERbeta in human cells. Toxicol. Sci. 80, 14–25 (2004).
Low, Y. L. et al. Phytoestrogen exposure is associated with circulating sex hormone levels in postmenopausal women and interact with ESR1 and NR1I2 gene variants. Cancer Epidemiol. Biomarkers Prev. 16, 1009–1016 (2007).
Frankenfeld, C. L. et al. 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 (2004).
Liu, B. et al. Equol-producing phenotype and in relation to serum sex hormones among healthy adults in Beijing. Wei Sheng Yan Jiu. 40, 727–731 (2011).
Cassidy, A., Bingham, S. & Setchell, K. D. Biological effects of a diet of soy protein rich in isoflavones on the menstrual cycle of premenopausal women. Am. J. Clin. Nutr. 60, 333–340 (1994).
Wocławek-Potocka, I. et al. Diverse effects of phytoestrogens on the reproductive performance: cow as a model. Int. J. Endocrinol. 2013, 650984 (2013).
Ezaki, T. et al. No clear-cut evidence for cadmium-induced renal tubular dysfunction among over 10,000 women in the Japanese general population: a nationwide large-scale survey. Int. Arch. Occup. Environ. Health. 76, 186–196 (2003).
Koizumi, A. et al. Assessment of human exposure to polychlorinated biphenyls and polybrominated diphenyl ethers in Japan using archived samples from the early 1980s and mid-1990s. Environ. Res. 99, 31–39 (2005).
Koizumi, A. et al. present, and future of environmental specimen banks. Environ. Health Prev. Med. 14, 307–318 (2009).
Grace, P. B. et al. Quantification of isoflavones and lignans in urine using gas chromatography/mass spectrometry. Anal. Biochem. 315, 114–121 (2003).
Okabe, H. et al. A kinetic and enzymatic method for determination of creatinine in serum and urine. Jpn. J. Clin. Chem. 4, 337–342 (1975).
Yager, J. D. Mechanisms of estrogen carcinogenesis: the role of E2/E1-quinone metabolites suggests new approaches to preventive intervention–a review. Steroids 99(Pt A), 56–60 (2015).
El-Hefnawy, T., Hernandez, C. & Stabile, L. P. The endocrine disrupting alkylphenols and 4,4’-DDT interfere with estrogen conversion and clearance by mouse liver cytosol. Reprod. Biol. 17, 185–192 (2017).
Ideno, Y. et al. Optimal cut-off value for equol-producing status in women: The Japan Nurses’ Health Study urinary isoflavone concentration survey. PLoS One. 13, e201318 (2018).
Gu, F. et al. Urinary concentrations of estrogens and estrogen metabolites and smoking in Caucasian women. Cancer Epidemiol. Biomarkers Prev. 22, 58–68 (2013).
Yasui, T. et al. Variation of urinary follicle-stimulating hormone level after menopause: from the results of Japan Nurses’ Health Study. J. Med. Invest. 66, 297–302 (2019).
Takahashi, A. et al. Association of equol with obesity in postmenopausal women. Menopause 28, 807–810 (2021).
Heinze, G. & Dunkler, D. Five myths about variable selection. Transplant. Int. 30, 6–10 (2017).
Homma, K. et al. Reference values for urinary steroids in Japanese newborn infants: gas chromatography/mass spectrometry in selected ion monitoring. Endocr. J. 50, 783–792 (2003).
Dunn, J. F., Nisula, B. C. & Rodbard, D. Transport of steroid hormones: binding of 21 endogenous steroids to both testosterone-binding globulin and corticosteroid-binding globulin in human plasma. J. Clin. Endocrinol. Metab. 53, 58–68 (1981).
Martin, M. E. et al. Interactions between phytoestrogens and human sex steroid binding protein. Life Sci. 58, 429–436 (1996).
Plowchalk, D. R. & Teeguarden, J. Development of a physiologically based pharmacokinetic model for estradiol in rats and humans: a biologically motivated quantitative framework for evaluating responses to estradiol and other endocrine-active compounds. Toxicol. Sci. 69, 60–78 (2002).
Campbell, D. R. & Kurzer, M. S. Flavonoid inhibition of aromatase enzyme activity in human preadipocytes. J. Steroid. Biochem. Mol. Biol. 46, 381–388 (1993).
Deluca, D. et al. Inhibition of 17beta-hydroxysteroid dehydrogenases by phytoestrogens: comparison with other steroid metabolizing enzymes. J. Steroid. Biochem. Mol. Biol. 93, 285–292 (2005).
Ojima, M., Hanioka, T., Tanaka, K. & Inoshita, E. A. Relationship between smoking status and periodontal conditions: findings from national databases in Japan. J. Periodontal. Res. 41, 573–539 (2006).
Nagata, C. et al. Dietary and lifestyle correlates of urinary excretion status of equol in Japanese women. Nutr. Cancer. 60, 49–54 (2008).
Frankenfeld, C. L. et al. Familial correlations, segregation analysis, and nongenetic correlates of soy isoflavone-metabolizing phenotypes. Exp. Biol. Med. (Maywood). 229, 902–913 (2004).
Malekinejad, H. & Rezabakhsh, A. Hormones in dairy foods and their impact on public health-a narrative review article. Iran J. Public Health. 44, 742–758 (2015).
Hartmann, S., Lacorn, M. & Steinhart, H. Natural occurrence of steroid hormones in food. Food Chem. 62, 7–20 (1998).
Hooper, L. et al. Effects of soy protein and isoflavones on circulating hormone concentrations in pre- and post-menopausal women: a systematic review and meta-analysis. Hum. Reprod. Update. 15, 423–440 (2009).
Yuan, J. P., Wang, J. H. & Liu, X. Metabolism of dietary soy isoflavones to equol by human intestinal microflora–implications for health. Mol. Nutr. Food Res. 51, 765–781 (2007).
Suzuki, Y., Hisada, A. & Yoshinaga, J. Inter- and intra-individual variation in urinary excretion of daidzein and equol in female Japanese. Biomarkers 19, 407–410 (2014).
Acknowledgements
The authors would like to express their sincere appreciation to Prof. Dr. Akio Koizumi, professor emeritus of Kyoto University, founder of the Kyoto University Human Specimen Bank, and also to many contributors. We thank Emily Crow, PhD, and Katherine Thieltges, from Edanz Group (https://en-author-services.edanz.com/ac) for editing a draft of this manuscript. This study was supported by a Grant-in-Aid for Scientific Research from the Japan Society for the Promotion of Science (JSPS) (Grant No. 20H03930) and by a JSPS and NRF of Korea bilateral joint research project.
Author information
Authors and Affiliations
Contributions
Conceptualization, M.H. and K.H.H.; formal analysis, T.F. and Y.F.; investigation, Z.L. and K.H.H.; writing—original draft preparation, T.F.; writing—review and editing, Y.F., M.H. and K.H.H.; supervision, K.H.H.; project administration, K.H.H.; funding acquisition, K.H.H. All authors have read and agreed to the published version of the manuscript.
Corresponding author
Ethics declarations
Competing interests
The authors declare no competing interests.
Additional information
Publisher's note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Supplementary Information
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Fujitani, T., Fujii, Y., Lyu, Z. et al. Urinary equol levels are positively associated with urinary estradiol excretion in women. Sci Rep 11, 19532 (2021). https://doi.org/10.1038/s41598-021-98872-2
Received:
Accepted:
Published:
DOI: https://doi.org/10.1038/s41598-021-98872-2
- Springer Nature Limited
This article is cited by
-
Equol: a metabolite of gut microbiota with potential antitumor effects
Gut Pathogens (2024)
-
Interaction between gut microbiota and sex hormones and their relation to sexual dimorphism in metabolic diseases
Biology of Sex Differences (2023)
-
The Impact of 17β-estradiol (E2) on the Growth Profile of Environmental Enterobacteriaceae
Water, Air, & Soil Pollution (2023)