The ubiquitous occurrence of polycyclic aromatic hydrocarbons (PAHs) leads to constant human exposure at low levels. Toxicologically relevant are especially the high-molecular weight substances due to their (pro-)carcinogenic potential. Following ingestion or uptake, the eukaryotic phase I metabolism often activates these substances to become potent DNA binders, and unsurprisingly metabolism and DNA-adduct formation of model substances such as benzo[a]pyrene (B[a]P) are well studied. However, apart from being subjected to eukaryotic transformations PAHs are also carbon and energy sources for the myriads of commensal microbes inhabiting man’s every surface. Yet, we know little about the microbiome’s PAH-metabolism capacity and its potentially adverse impact on the human host. This study now shows that readily isolable skin commensals transform B[a]P into a range of highly cyto- and genotoxic metabolites that are excreted in toxicologically relevant concentrations during growth. The respective bacterial supernatants contain a mixture of established eukaryotic as well as hitherto unknown prokaryotic metabolites, the combination of which leads to an increased toxicity. Altogether we show that PAH metabolism of the microbiome has to be considered a potential hazard.Electronic supplementary materialThe online version of this article (doi:10.1007/s00204-017-1964-3) contains supplementary material, which is available to authorised users.

The skin`s microbiome is predominantly commensalic, harbouring a metabolic potential far exceeding that of its host. While there is clear evidence that bacteria-dependent metabolism of pollutants modulates the toxicity for the host there is still a lack of models for investigating causality of microbiome-associated pathophysiology or toxicity. We now report on a biologically characterised microbial–skin tissue co-culture that allows studying microbe–host interactions for extended periods of time in situ. The system is based on a commercially available 3D skin model. In a proof-of-concept, this model was colonised with single and mixed cultures of two selected skin commensals. Two different methods were used to quantify the bacteria on the surface of the skin models. While Micrococcus luteus established a stable microbial–skin tissue co-culture, Pseudomonas oleovorans maintained slow continuous growth over the 8-day cultivation period. A detailed skin transcriptome analysis showed bacterial colonisation leading to up to 3318 significant changes. Additionally, FACS, ELISA and Western blot analyses were carried out to analyse secretion of cytokines and growth factors. Changes found in colonised skin varied depending on the bacterial species used and comprised immunomodulatory functions, such as secretion of IL-1α/β, Il-6, antimicrobial peptides and increased gene transcription of IL-10 and TLR2. The colonisation also influenced the secretion of growth factors such as VFGFA and FGF2. Notably, many of these changes have already previously been associated with the presence of skin commensals. Concomitantly, the model gained first insights on the microbiome’s influence on skin xenobiotic metabolism (i.e., CYP1A1, CYP1B1 and CYP2D6) and olfactory receptor expression. The system provides urgently needed experimental access for assessing the toxicological impact of microbiome-associated xenobiotic metabolism in situ.

Exposure to xenobiotics has repeatedly been associated with adverse health effects. While the majority of reported cases relate to direct substance effects, there is increasing evidence that microbiome-dependent metabolism of xenobiotic substances likewise has direct adverse effects on the host.