M. Griffin1, D. Li1, J. Parker1, J. Guo1, C. Valencia1, M. Kuhnert1, A. Morgan1, J. Lu1, K. Erich Bauer-Rowe1, C. Yao1, M. Januszyk1, M. Longaker1, D. Wan1 1Stanford University, Palo Alto, CA, USA
Introduction: Progress towards the development of treatments aimed at reducing scars has been hampered by our limited understanding of specific fibroblast populations responsible for fibrotic versus regenerative healing. Increasing evidence for fibroblast heterogeneity suggests there may be fibroblast-intrinsic properties underlying differential scar formation. Dermal fibroblasts have different embryonic origins depending on their location: facial, scalp, ventral, and dorsal fibroblasts derive primarily from the cranial neural crest, cephalic mesoderm, lateral plate mesoderm, and paraxial mesoderm, respectively. These fibroblast origins may in turn influence fibrotic behaviour in postnatal life. Here, we conducted molecular investigations to understand the signaling pathways responsible for regenerative facial healing and strategies to utilize these pathways to overcome fibrotic healing.
Methods: To determine if the embryonic origins of fibroblasts affect scar formation, we developed a four-site wounding model in which we created 2.5 mm full-thickness excisional skin wounds on the face, scalp, ventrum, and dorsum of the mouse. Whole tissues were analyzed for histology and confocal microscopy at Post operative Day 14. Wounds from the same timepoints were also prepared for Chromium Single Cell RNA sequencing (scRNAseq) and ATAC sequencing, spatial transcriptomics, and CODEX spatial proteomics.
Results: H&E and Masson’s trichome staining revealed that facial wounds had less scar formation, with significantly decreased dermal thickness and collagen deposition, especially when compared to dorsal wounds (Fig. 1A). scRNA-seq revealed distinct Robo2+ fibroblasts subpopulations in regenerating facial wounds compared to dorsal, ventral, or scalp wounds (Fig. 1B). ATAC-seq further demonstrated epigenetic modification of the fibroblast landscape specifically in facial wounds through Robo2 signaling (Fig. 1C). Spatial phenotyping by CODEX and VISIUM demonstrated that Robo2+ fibroblasts exist in wounds that regenerate (Fig. 1D). Using small molecule inhibitors and transgenic genetic lentivirus models, we showed the upregulation of Robo2 signaling in dorsal wounds led to a transcriptomic and histological phenotype that mimicked facial wounds (Fig. 1E).
Conclusions: Using our novel four-site murine dermal injury model, we showed that the ontology of fibroblasts also matters for adult wound healing: Neural-crest -derived facial fibroblasts differ from fibroblasts of other body sites by their intrinsic capacity to promote healing with less fibrosis. Further, we demonstrated that dorsal fibroblasts can be modulated towards a neural-crest-like state to produce a less fibrotic phenotype, highlighting their plasticity despite their embryologically determined fibrogenic potential.