940 Epithelial-microbial crosstalk modifies wound healing outcomes: from shotgun metagenomics to mechanistic understanding

Impaired wound healing is a major complication of diabetes and threatens to affect approximately one-quarter of the ∼400 million adults worldwide living with diabetes. Diabetic foot ulcer (DFU) complications result in two-thirds of all lower extremity amputations, with 5-year mortality rates surpassing many cancers. While DFUs are colonized with diverse microbes including pathogens, it is unclear how they impact clinical outcomes and the mechanisms by which they alter wound healing responses. To identify strain-level microbial biomarkers influencing clinical outcomes in a longitudinal prospective cohort of 100 subjects with DFU, we integrated metagenomic shotgun sequencing with clinical metadata in a machine learning framework. Specific strains of Staphylococcus aureus were associated with poor outcomes. Similarly, we identified skin commensals and other non-pathogenic bacteria (eg. Corynebacterium spp.) that associated with favorable outcomes. We used human diabetic ex vivo and murine diabetic in vivo model systems to demonstrate species- and strain-level influence of the corresponding cultured clinical wound isolates on keratinocyte and dermal fibroblast responses and wound closure. For example, S. aureus strains isolated from DFU with poor outcomes strikingly inhibited wound healing responses and closure compared to generalist S. aureus strains. Dual-RNAseq transcriptomic profiling of both host and microbe during in vivodiabetic wound healing revealed candidate mechanisms of interaction to explain differential healing responses and clinical outcomes. This work sheds light on host-microbe interactions that modify wound healing outcomes while revealing clinically significant biomarkers with the potential to guide management and treatment strategies of chronic wounds including DFU.