37. Characterization of Fibroblast Heterogeneity Across Embryonic Origins of Skin Using Single Cell Rna Sequencing Reveals the Regenerative Capacity of Robo2 Fibroblasts in Skin Repair

Purpose: Human skin scarring varies according to anatomical location, with the face healing at differing rates to scalp, ventrum and dorsum. Fibroblasts are the key cells involved in wound healing originating from four embryonic origins. Fibroblasts from the face, scalp, ventrum and dorsum originate from the neural-crest, cephalic mesoderm, lateral plate mesoderm and somite mesoderm, respectively. Here, we use single-cell RNA sequencing to understand fibroblast heterogeneity across mouse embryonic origins. Methods: Stented 2.5mm excisional wounds were created in the facial, scalp, ventrum and dorsum regions in C57/BL6J mice. Wounds were harvested at post-operative day (POD)-7 (midday through healing) and POD-14 (wound re-epithelialization), with unwounded skin serving as controls. Wounds were analyzed for histology, fluorescence-activated cell sorting (FACS), and confocal microscopy. Picrosirius red PSR staining was also performed to evaluate the extracellular matrix structure (ECM) of the four regions following wounding. At POD-7 and POD-14 wounds were digested and fibroblasts were sorted using FACS and processed for droplet based microfluidic single-cell RNA analysis (scRNAseq). Results: Facial wounds derived from the neural-crest displayed significantly accelerated wound closure, decreased scar thickness and collagen deposition, compared to dorsal, ventral, and scalp wounds (*P<0.05) (n=10). Using scRNAseq we identified 6 distinct clusters of fibroblasts at POD-14 across the four embryonic regions. Two subpopulations were abundant in facial derived fibroblasts enriched for Robo2, CD200, Ncam1, and Jag1. Pathway analysis using gene set enrichment analysis showed upregulation of known neural-crest signalling pathways including Wnt canonical-signalling, transforming growth factor-β regulation and hedgehog signalling in facial clusters, indicating that the embryonic origins of fibroblasts contributed to the outcome of skin scarring. Pseudotime analysis also revealed significant gene expression of Robo2 in the face wounds compared to the scalp, ventrum, and dorsum wounds. FACS and histology analysis of POD-14 wounds confirmed the significant infiltration of Robo2 positive fibroblasts in facial wounds compared to scalp, ventrum, and dorsum derived wounds (n=6) (*P<0.05). To evaluate the regenerative potential of Robo2 positive fibroblasts in skin healing, Robo2 positive fibroblasts were isolated using FACS and transplanted into dorsal wounds. Scar thickness was significantly decreased in dorsal wounds treated with Robo2 positive fibroblasts compared to wounds with Robo2 negative fibroblasts at POD-14 (*P<0.05). Wounds with Robo2 positive fibroblasts transplants also demonstrated decreased collagen type 1 and α-smooth muscle actin gene expression compared to wounds treated with Robo2 negative fibroblasts by RT-qPCR (*P<0.05). Finally, automated PSR analysis further confirmed that wounds treated with Robo2 positive fibroblasts displayed an ECM structure to more closely resemble facial wounds. Conclusion: Our study shows there is a high degree of fibroblast heterogeneity across embryonic regions, which may contribute to the diversity in skin scarring across the body.Neural crest derived facial fibroblasts show a decreased scarring response. Neural crest fibroblasts show high levels of Robo2 signaling. Understanding the mechanism by which Robo2 fibroblasts decreases dermal fibrosis may provide future therapeutic targets to overcome skin scarring.