Phytoestrogens exert pleiotropic effects on cellular signaling and show some beneficial effects on estrogen-dependent diseases. However, due to activation/inhibition of the estrogen receptors ERalpha or ERbeta, these compounds may induce or inhibit estrogen action and, therefore, have the potential to disrupt estrogen signaling. We performed a comprehensive analysis and potency comparison of phytoestrogens and their human metabolites for ER binding, induction/suppression of ERalpha and ERbeta transactivation, and coactivator recruitment in human cells. The soy-derived genistein, coumestrol, and equol displayed a preference for transactivation of ERbeta compared to ERalpha and were 10- to 100-fold less potent than diethylstilbestrol. In contrast, zearalenone was the most potent phytoestrogen tested and activated preferentially ERalpha. All other phytoestrogens tested, including resveratrol and human metabolites of daidzein and enterolactone, were weak ER agonists. Interestingly, the daidzein metabolites 3',4',7-isoflavone and 4',6,7-isoflavone were superagonists on ERalpha and ERbeta. All phytoestrogens tested showed reduced potencies to activate ERalpha and ERbeta compared to diethylstilbestrol on the estrogen-responsive C3 promoter compared to a consensus estrogen response element indicating a degree of promoter dependency. Zearalenone and resveratrol were antagonistic on both ERalpha and ERbeta at high doses. The phytoestrogens enhanced preferentially recruitment of GRIP1 to ERalpha similar to 17beta-estradiol. In contrast, for ERbeta no distinct preference for one coactivator (GRIP1 or SRC-1) was apparent and the overall coactivator association was less pronounced than for ERalpha. Due to their abundance and (anti)-estrogenic potencies, the soy-derived isoflavones, coumestrol, resveratrol, and zearalenone would appear to have the potential for effectively functioning as endocrine disruptors.

Curcumin and its natural congeners are of current interest because of their putative anti-inflammatory and anticarcinogenic activities, but knowledge about their metabolic fate is scant. In the present study conducted with precision-cut liver slices from male and female Sprague-Dawley rats, five reductive but no oxidative metabolites of curcumin and its demethoxy and bis-demethoxy analogues were observed and identified by HPLC and GC-MS analysis, mostly by comparison with authentic reference compounds. The major reductive metabolites were the hexahydrocurcuminoids in both male and female rat liver slices, whereas male rats formed more octahydro than tetrahydro metabolites and female rats more tetrahydro- than octahydrocurcuminoids. Tetrahydro, hexahydro, and octahydro metabolites were predominantly present as glucuronides, but a significant proportion of sulfate conjugates was also observed. The lack of formation of oxidative metabolites of curcumin and the ready generation of reductive metabolites were confirmed using rat liver microsomes and cytosol, respectively. Results of enzymatic hydrolysis studies conducted under various conditions revealed that curcumin and demethoxycurcumin are chemically less stable than bis-demethoxycurcumin, whereas the reductive metabolites of all three curcuminoids are stable compounds. This is the first report on the metabolism of demethoxycurcumin and bis-demethoxycurcumin. In view of the chemical instability of the parent curcuminoids, it is proposed to use their major phase I metabolites, that is, the stable hexahydro products, as biomarkers for exposure in clinical studies.

The soy isoflavones daidzein and genistein are found in high concentrations in human plasma and urine after soy consumption. However, in vitro and in vivo data regarding the oxidative metabolism of isoflavones in humans are scarce. Therefore, we have studied the oxidative metabolites of these compounds formed in human liver microsomes and excreted in urine of male and female humans ingesting soy products for 2 days. Human liver microsomes transformed the soy isoflavone daidzein to three monohydroxylated and three dihydroxylated metabolites according to GC/MS analysis. On the basis of a previous study with rat liver microsomes and with the help of reference substances, these metabolites were identified as 6,7,4'-trihydroxyisoflavone, 7,3',4'-trihydroxyisoflavone, 7,8,4'-trihydroxyisoflavone, 7,8,3',4'-tetrahydroxyisoflavone, 6,7,8,4'-tetrahydroxyisoflavone, and 6,7,3',4'-tetrahydroxyisoflavone. Significant amounts of the same metabolites except 6,7,8,4'-tetrahydroxyisoflavone were also found in urine of female and male volunteers after soy intake. Genistein was metabolized by human liver microsomes to six hydroxylation products. The main metabolites were the three aromatic monohydroxylated products 5,6,7,4'-tetrahydroxyisoflavone, 5,7,8,4'-tetrahydroxyisoflavone and 5,7,3',4'-tetrahydroxyisoflavone. The aliphatic monohydroxylated metabolite 2,5,7,4'-tetrahydroxyisoflavone and two aromatic dihydroxylated metabolites, 5,7,8,3',4'-pentahydroxyisoflavone and 5,6,7,3',4'-pentahydroxyisoflavone, were formed in trace amounts. The same hydroxylated genistein metabolites except the aliphatic hydroxylated one could also be detected in human urine samples. Methylated forms of the catechol metabolites, which were generated by incubations with catechol-O-methyltransferase in vitro could be detected only in trace amounts in the urine samples. This implies that this reaction does not play a major role in the biotransformation of the hydroxylated daidzein and genistein metabolites in vivo. Most of these oxidative metabolites are described as human in vivo metabolites for the first time. Their biological significance remains to be established.

The European Commission asked EFSA to update their 2006 opinion on ochratoxin A (OTA) in food. OTA is produced by fungi of the genus Aspergillus and Penicillium and found as a contaminant in various foods. OTA causes kidney toxicity in different animal species and kidney tumours in rodents. OTA is genotoxic both in vitro and in vivo; however, the mechanisms of genotoxicity are unclear. Direct and indirect genotoxic and non‐genotoxic modes of action might each contribute to tumour formation. Since recent studies have raised uncertainty regarding the mode of action for kidney carcinogenicity, it is inappropriate to establish a health‐based guidance value (HBGV) and a margin of exposure (MOE) approach was applied. For the characterisation of non‐neoplastic effects, a BMDL10 of 4.73 μg/kg body weight (bw) per day was calculated from kidney lesions observed in pigs. For characterisation of neoplastic effects, a BMDL10 of 14.5 μg/kg bw per day was calculated from kidney tumours seen in rats. The estimation of chronic dietary exposure resulted in mean and 95th percentile levels ranging from 0.6 to 17.8 and from 2.4 to 51.7 ng/kg bw per day, respectively. Median OTA exposures in breastfed infants ranged from 1.7 to 2.6 ng/kg bw per day, 95th percentile exposures from 5.6 to 8.5 ng/kg bw per day in average/high breast milk consuming infants, respectively. Comparison of exposures with the BMDL10 based on the non‐neoplastic endpoint resulted in MOEs of more than 200 in most consumer groups, indicating a low health concern with the exception of MOEs for high consumers in the younger age groups, indicating a possible health concern. When compared with the BMDL10 based on the neoplastic endpoint, MOEs were lower than 10,000 for almost all exposure scenarios, including breastfed infants. This would indicate a possible health concern if genotoxicity is direct. Uncertainty in this assessment is high and risk may be overestimated.

No abstract