A multi-strain human skin microbiome model provides a testbed for disease modeling

The skin microbiome plays a critical role at the interface between the human epidermis and the environment, providing colonization resistance against pathogenic strains, training host immunity, and supporting epithelial turnover. Inversely, dysbiotic skin microbiome states are associated with skin disease, particularly inflammatory conditions such as atopic dermatitis and psoriasis. Current evaluation of human host and microbiome interactions relies on post hoc studies after disease onset. This limits the ability to evaluate the causal roles of host and microbe during disease progression. One approach to characterizing microbial and host biology in a controlled and reproducible context is to derive in vitro models of sufficient complexity and stability to support perturbation and response. Current tools for studying these processes are focused on testing antagonistic or synergistic relations between two or more strains for short (hours to days) culture durations, thereby precluding studies of relevant complexity and chronic disease states. Here, we present an in vitro model of the human skin microbiome comprising a six strain consortium colonizing primary human keratinocyte-derived tissue in Air-Liquid Interface for up to 7 days. We evaluated readouts of tissue health including histology, gene expression, and transepithelial electrical resistance (TEER), as well as relative strain abundance to characterize microbiome stability over time. Skin cells formed a complex tissue structure over two weeks and maintained stable or increasing TEER after 7 days of co-culture with the microbial consortium. Up to five of the six strains were viable on the skin tissue surface on day 7 as validated by custom qPCR assays, demonstrating a robust and stable testbed for microbiome studies. A remarkable feature of this model is the persistence of Cutibacterium acnes in an aerobic tissue culture environment, since C. acnes growth is typically demonstrated under anaerobic conditions, suggesting that the skin tissue model is conducive to more natural growth states of native skin strains. The addition of cytokines representative of atopic dermatitis elicited a marked decrease in tissue barrier by day 7 compared to healthy controls, irrespective of the microbiome presence. Furthermore, an alteration in relative strain abundance was observed in diseased model tissues, demonstrating capability to study the impact of disease states on the microbiome and vice versa. We envision this model system as a test bed to evaluate the influence of commensals on host biology, the influence of external environment on microbiome stability, and chronic diseases impacted by dysbiosis.