Many studies suggest that the potential impact of bisphenol S (BPS) as an endocrine disruptor is comparable to that of bisphenol A (BPA). However, in vitro-to-in vivo and from animal to human extrapolations require knowledge of the plasma free fraction of the active endocrine compounds. The present study aimed to characterise BPA and BPS binding to plasma proteins both in humans and different animal species. The plasma protein binding of BPA and BPS was assessed by equilibrium dialysis in plasma from adult female mice, rats, monkeys, early and late pregnant women as well as paired cord blood, early and late pregnant sheep and foetal sheep. The fraction of free BPA was independent of plasma concentrations and ranged between 4% and 7% in adults. This fraction was 2 to 3.5 times lower than that of BPS in all species except sheep, ranging from 3% to 20%. Plasma binding of BPA and BPS was not affected by the stage of pregnancy, BPA and BPS free fractions representing about 4% and 9% during early and late human pregnancy, respectively. These fractions were lower than the free fractions of BPA (7%) and BPS (12%) in cord blood. Our results suggest that similarly to BPA, BPS is extensively bound to proteins, mainly albumin. The higher fraction of free BPS compared to BPA may have implications for human exposure assessment since BPS free plasma concentrations are expected to be 2 to 3.5 times higher than that of BPA for similar plasma concentration.

Ovarian cells are critical for reproduction and steroidogenesis, which are functions that can be impacted by exposure to xenobiotics. As in other extra-hepatic tissues, biotransformation events may occur at the ovarian level. Such metabolic events deserve interest, notably as they may modulate the overall exposure and toxicity of xenobiotics. In this study, the comparative metabolic fate of two bisphenols was investigated in ovarian cells. Bisphenol A (BPA), a model endocrine disruptor, and its major substitute bisphenol S (BPS) were selected. Bovine granulosa cells (primary cultures) and theca explants (ex vivo tissue) were exposed for 24 hr to tritium-labeled BPA, BPS and their respective glucuronides (i.e. their major circulating forms), at concentrations consistent with low-dose exposure scenarios. Mass balance studies were performed, followed by radio-HPLC profiling. The capability of both cell compartments to biotransform BPA and BPS into their respective sulfo-conjugates was demonstrated, with sulfation being the predominant metabolic route. In theca, there was a significantly higher persistence of BPA (compared to BPS) residues over 24 hr. Moreover, only theca explants were able to decon-jugate inactive BPA-glucuronide and BPS-glucuronide back into their biologically active aglycone forms. Deconjugation rates were demonstrated to be higher for BPS-G than for BPA-G. These findings raise concerns about the in situ direct release of bisphenols at the level of the ovary and demonstrate the relevance of exploring the biotransformation of bisphenols and their circulating metabolites in different ovarian cells with specific metabolic capabilities. This work also provides essential knowledge for the improved risk assessment of bisphenols.

Emerging mycotoxins are currently gaining more attention due to their high frequency of contamination in foods and grains. However, most data available in the literature are in vitro, with few in vivo results that prevent establishing their regulation. Beauvericin (BEA), enniatins (ENNs), emodin (EMO), apicidin (API) and aurofusarin (AFN) are emerging mycotoxins frequently found contaminating food and there is growing interest in studying their impact on the liver, a key organ in the metabolization of these components. We used an ex vivo model of precision-cut liver slices (PCLS) to verify morphological and transcriptional changes after acute exposure (4 h) to these mycotoxins. The human liver cell line HepG2 was used for comparison purposes. Most of the emerging mycotoxins were cytotoxic to the cells, except for AFN. In cells, BEA and ENNs were able to increase the expression of genes related to transcription factors, inflammation, and hepatic metabolism. In the explants, only ENN B1 led to significant changes in the morphology and expression of a few genes. Overall, our results demonstrate that BEA, ENNs, and API have the potential to be hepatotoxic.

Deoxynivalenol (DON) is one of the most common mycotoxins in cereals and their by-products. Its adverse effects on animal and human health have been extensively studied in the intestine, but little attention has been paid to another target organ for mycotoxins, the liver that is potentially exposed after intestinal absorption and enterohepatic circulation. To assess DON's toxicity in an ex vivo model structurally and physiologically closer to the whole liver, we developed a pig precision-cut liver slices (PCLS) model. PCLS contain all cell types and maintain intercellular and cell-matrix interactions, among other architectural features of the liver. The human HepG2 cell line was used for comparison. We observed that after a short exposure, DON reduced the cell viability of HepG2 cells and induced the expression of genes involved in apoptosis, inflammation and oxidative stress. When PCLS were exposed to DON, damage to the tissues was observed, with no changes in markers of liver function or injury. Exposure to the toxin also triggered liver inflammation and apoptosis, effects already observed in pigs fed DON-contaminated diets. Overall, these data demonstrate that DON had toxic effects on a liver cell line and on whole liver tissue, consistent with the effect observed during in vivo exposure. They also indicate that pig PCLS is a relevant and sensitive model to investigate the liver toxicity of food contaminants.

The toxicity of mycotoxins containing bisfuranoid structures such as aflatoxin B1 (AFB1) depends largely on biotransformation processes. While the genotoxicity and mutagenicity of several bisfuranoid mycotoxins including AFB1 and sterigmatocystin have been linked to in vivo bioactivation of these molecules into reactive epoxide forms, the metabolites of genotoxic and mutagenic AFB1 precursor versicolorin A (VerA) have not yet been characterized. Because this molecule is not available commercially, our strategy was to produce a library of metabolites derived from the biotransformation of in-house purified VerA, following incubation with human liver S9 fractions, in presence of appropriate cofactors. The resulting chromatographic and mass-spectrometric data were used to identify VerA metabolites produced by intestinal cell lines as well as intestinal and liver tis-sues exposed ex vivo. In this way, we obtained a panel of metabolites suggesting the involvement of phase I (M + O) and phase II (glucuronide and sulfate metabolites) enzymes, the latter of which is implicated in the detoxi-fication process. This first qualitative description of the metabolization products of VerA suggests bioactivation of the molecule into an epoxide form and provides qualitative analytic data to further conduct a precise metabolism study of VerA required for the risk assessment of this emerging mycotoxin.

The gut microbiota–intestine–liver relationship is emerging as an important factor in multiple hepatic pathologies, but the hepatic sensors and effectors of microbial signals are not well defined. By comparing publicly available liver transcriptomics data from conventional vs. germ-free mice, we identified pregnane X receptor (PXR, NR1I2) transcriptional activity as strongly affected by the absence of gut microbes. Microbiota depletion using antibiotics in Pxr+/+ vs Pxr-/- C57BL/6J littermate mice followed by hepatic transcriptomics revealed that most microbiota-sensitive genes were PXR-dependent in the liver in males, but not in females. Pathway enrichment analysis suggested that microbiota–PXR interaction controlled fatty acid and xenobiotic metabolism. We confirmed that antibiotic treatment reduced liver triglyceride content and hampered xenobiotic metabolism in the liver from Pxr+/+ but not Pxr-/- male mice. These findings identify PXR as a hepatic effector of microbiota-derived signals that regulate the host’s sexually dimorphic lipid and xenobiotic metabolisms in the liver. Thus, our results reveal a potential new mechanism for unexpected drug–drug or food–drug interactions.

Additional file 5 Gene lists and GO enrichment analyses for Pxr-/- vs Pxr+/+ transcriptome comparisons. (A) List of genes significantly up-regulated in males Pxr-/- compared to males Pxr+/+. (B) GO term enrichment analysis using the list of genes significantly up-regulated in males Pxr-/- compared to males Pxr+/+. Metabolic pathways (summary) were considered as significantly enriched when q-value < 0.05. (C) List of genes significantly up-regulated in females Pxr-/- compared to females Pxr+/+. (D) GO term enrichment analysis using the list of genes significantly up-regulated in females Pxr-/- compared to females Pxr+/+. Metabolic pathways (summary) were considered as significantly enriched when q-value < 0.05. (E) List of genes significantly down-regulated in males Pxr-/- compared to males Pxr+/+. (F) GO term enrichment analysis using the list of genes significantly up-regulated in males Pxr-/- compared to males Pxr+/+. No summary metabolic pathways (summary) were considered as significantly enriched with q-value < 0.05. (A) List of genes significantly down-regulated in females Pxr-/- compared to females Pxr+/+. (B) GO term enrichment analysis using the list of genes significantly down-regulated in females Pxr-/- compared to females Pxr+/+. Metabolic pathways (summary) were considered as significantly enriched when q-value < 0.05.

Skin metabolism is important to consider when assessing local toxicity and/or penetration of chemicals and their metabolites. If human skin supply is limited, pig skin can be used as an alternative. To identify any species differences, we have investigated the metabolism of 10 chemicals in a pig and human skin explant model. Phase I metabolic pathways in skin from both species included those known to occur via cytochrome P450s, esterases, alcohol dehydrogenases and aldehyde dehydrogenases. Common Phase II pathways were glucuronidation and sulfation but other conjugation pathways were also identified. Chemicals not metabolized by pig skin (caffeine, IQ and 4-chloroaniline) were also not metabolized by human skin. Six chemicals metabolized by pig skin were metabolized to a similar extent (percentage parent remaining) by human skin. Human skin metabolites were also detected in pig skin incubations, except for one unidentified minor vanillin metabolite. Three cinnamyl alcohol metabolites were unique to pig skin but represented minor metabolites. There were notable species differences in the relative amounts of common metabolites. The difference in the abundance of the sulfate conjugates of resorcinol and 4-amino-3-nitrophenol was in accordance with the known lack of aryl sulfotransferase activity in pigs. In conclusion, while qualitative comparisons of metabolic profiles were consistent between pig and human skin, there were some quantitative differences in the percentage of metabolites formed. This preliminary assessment suggests that pig skin is metabolically competent and could be a useful tool for evaluating potential first-pass metabolism before testing in human-derived tissues.

Being able to explore the metabolism of broad metabolizing cells is of critical importance in many research fields. This article presents an original modeling solution combining metabolic network and omics data to identify modulated metabolic pathways and changes in metabolic functions occurring during differentiation of a human hepatic cell line (HepaRG). Our results confirm the activation of hepato-specific functionalities and newly evidence modulation of other metabolic pathways, which could not be evidenced from transcriptomic data alone. Our method takes advantage of the network structure to detect changes in metabolic pathways that do not have gene annotations and exploits flux analyses techniques to identify activated metabolic functions. Compared to the usual cell-specific metabolic network reconstruction approaches, it limits false predictions by considering several possible network configurations to represent one phenotype rather than one arbitrarily selected network. Our approach significantly enhances the comprehensive and functional assessment of cell metabolism, opening further perspectives to investigate metabolic shifts occurring within various biological contexts.