LOSS OF MFAP5 AND ITS EFFECT ON SKIN HOMEOSTASIS AND WOUND HEALING

MFAP5, also known as microfibrillar-associated glycoprotein-2 (MAGP2), may influence parameters of skin wound healing related to scar formation. To further elucidate its role in skin wound healing, we assessed skin wound repair in Mfap5-/- mice. Loss of MFAP5 significantly reduced wound closure rates and angiogenesis while enhancing neutrophil and macrophage influx into wounds. Loss of MFAP5 also reduced the deposition of total and mature collagen in uninjured normal skin (NS), but not in wounds. Furthermore, NS dermis of Mfap5-/- mice was thinner without any reduction in tensile strength. Single-cell RNA-sequencing of NS and wounds from Mfap5+/+ and Mfap5-/- mice revealed two fibroblast subclusters that express MFAP5 more highly than other subclusters. Enrichment analysis of the differentially expressed genes (DEGs) in these two subclusters suggests these fibroblasts engage in extracellular matrix (ECM) deposition and angiogenesis. Mfap5+/+ and Mfap5-/- fibroblasts also exhibit transcriptomic differences throughout in vivo wound healing, though as healing progressed, fewer differences were evident. To examine the direct effect of MFAP5 on fibroblasts outside of the wound space, fibroblasts were isolated from Mfap5+/+ and Mfap5-/- mice for in vitro analysis. MRNA-sequencing of Mfap5+/+ and Mfap5-/- fibroblasts found genes involved in cellular migration and proliferation, ECM synthesis, and angiogenesis to be downregulated in Mfap5-/- fibroblasts vs Mfap5+/+ fibroblasts. Functionally, Mfap5-/- fibroblasts exhibited reduced migration, contractility, proliferation, and ECM deposition. Our findings indicate that MFAP5 is a multifunctional glycoprotein in skin wound healing as it promotes angiogenesis and collagen deposition, inhibits inflammatory cell influx, and promotes pro-scarring fibroblast behavior. Overall design: Wound tissues were collected from male and female Mfap5+/+ or Mfap5-/- mice (n=4-5 animals for each sex and each experimental group) on days 0 (D0), 3 (D3), and 7 (D7) post-wounding. A total of 48-60 mice were used for single-cell RNA-sequencing. Collected wound tissues were pooled separately based on sex, experimental group, and time post-wounding. Tissue collected from D0 was considered NS. After collection and pooling, tissues were micro-dissected, minced, and incubated at 37°C in a dispase (Worthington Biochemical Corporation, Lakewood, NJ, Catalog # 9001-92-7)/liberase (Roche, Basel, Switzerland, Catalog # 5401135001) solution supplemented with DNAse I (Sigma-Aldrich, Catalog # D5025) and penicillin-streptomycin (Gibco, Waltham, MA, Catalog # 15-140-122) for 60 min with constant rotation. The final concentrations for each solution component were 3.6 mg/mL dispase, 30 µg/mL liberase, 100 µg/mL DNAse I, and 100 Units/mL-100 µg/mL penicillin-streptomycin. Post-incubation, cell suspensions were passed through a 40 µM cell strainer (ThermoFisher Scientific, Waltham, MA). Single-cell suspensions were treated with 1x RBC lysis buffer (Biolegend, San Diego, California, Catalog # 420301), washed, and resuspended in cold PBS before debris removal using the Debris Removal Solution (Miltenyi Biotec, North Rhine-Westphalia, Germany, Catalog # 130-109-398) according to the manufacturer's protocol. All cells were then frozen in a mixture of 70% DMEM (Corning Inc, Corning, NY), 30% Fetal Bovine Serum (FBS) (GeminiBio, Sacremento, CA), and 10% DMSO (Sigma-Aldrich) before being sent to Novogene (Novogene America, Sacramento, CA). Thawed cell suspensions underwent dead cell removal, and only samples with >100,000 cells, >70% viability, and minimal to no cell clumps were used for single-cell RNA sequencing. GEM generation, barcoding, post-GEM-RT cleanup, cDNA amplification, and cDNA library construction were performed using Single-cell 3' v3 chemistry (10X Genomics, Pleasanton, California). Libraries were then sequenced on an Illumina NovaSeq 6000 (Illumina, San Diego, CA). Dead cell removal, cell counting and viability testing, GEM generation, barcoding, post-GEM-RT cleanup, cDNA amplification, library preparation, quality control, and sequencing were all performed by Novogene (Novogene America). Transcripts were mapped to the mouse reference genome (GRCm38/mm10) using Cell Ranger Version 7.0.0. Before downstream analyses, low-quality cells were removed and data integration was performed using the RPCA method. Quality control metrics included keeping genes present in at least 10 cells and cells with at least 200 genes, while removing cells with either a mitochondrial gene expression greater than 10% or a multiplet likelihood score greater than 0.25. Post-quality control, 72,876 cells remained for downstream bioinformatic analyses. All data quality control and downstream analyses were done by us using R version 4.4.1 with Bioconductor v3.20. Clustering of cells was performed using the Seurat R package. In brief, digital gene expression matrices were column-normalized and log-transformed. Principal component analysis (PCA) was first performed on the list of highly variable genes to identify cell clusters. Genes were selected for inclusion in PCA with an average expression > 0.01 and dispersion > 1.0. We used the Jackstraw method in Seurat to identify significant principal components (PCs). The top 10 PCs were used for clustering with the Louvain modularity-based community detection algorithm to generate cell clusters (FindClusters function, 18 clusters with resolution = 0.2). Cell cluster marker genes were identified using the FindAllMarkers() function according to the following criteria: adjusted p-value (p.adj) < 0.05 and log(fold-change) (Log2FC) > 0.25. To present high-dimensional data in a two-dimensional space, we performed UMAP analysis using the results of PCA with significant PCs as input. Differentially expressed marker genes used for cell type identification were selected using the following criteria: p.adj < 0.05, log(fold-change) > 2, pct.1 > 0.40 (percentage of cells where the gene is detected within the cluster), and pct.2 < 0.20 (percentage of cells where the gene is detected in other clusters). Differentially expressed marker genes meeting these cut-offs were input into EnrichR, and likely cell types were identified using the top 5 enrichments from the following cell marker curated databases: 1) CellMarker 2024, Tabula Muris, and PanglaoDB. Known marker genes used for skin cell types in previously published single-cell RNA-sequencing studies were also used. We followed the same procedures as above for the sub-clustering analysis of fibroblasts. The top 15 PCs were used for clustering, and 6 subclusters (subclusters 0-5) were obtained with a resolution = 0.1. Fibroblast subcluster marker genes were identified using FindAllMarkers() function according to the following criteria: p.adj < 0.05 and log(fold-change) > 0.25. To present high-dimensional data in two-dimensional space, we performed UMAP analysis using the results of PCA with significant PCs as input. DEG signatures for each fibroblast subcluster were analyzed to determine if a unique fibroblast subcluster differentially expressed MFAP5 as a marker gene in NS and during wound healing in vivo. Gene ontology (GO) enrichment for biological processes (BP) terms and Reactome pathway enrichment analysis was performed on DEGs with a p.adj < 0.05 and log(fold-change) > 1 for the fibroblasts subclusters that differentially expressed MFAP5 relative to other fibroblast subclusters via the EnrichR package in R. GO BP terms with a p.adj < 0.05 following Benjamini and Hochberg's approach for controlling the false discovery rate (FDR) were considered statistically significantly enriched terms, and only the top 25 significantly enriched terms based on p.adj were included in our analyses. We used the FindAllMarkers() function to identify DEGs between fibroblasts isolated from Mfap5+/+ or Mfap5-/- mice at each time point. We identified DEGs between Mfap5+/+ or Mfap5-/- fibroblasts based on the following criteria: p.adj < 0.01, average Log2FC> 0.25, and the percentage of cells where the gene is detected in either Mfap5+/+ or Mfap5-/- fibroblasts (pct.1) is more than 25% (pct.1 > 0.25). GO BP and Reactome pathway term enrichment analysis was performed on the list of differentially upregulated and downregulated genes in Mfap5-/- fibroblasts relative to Mfap5+/+ fibroblasts via the EnrichR package in R version 4.4.1. GO terms with a p.adj < 0.05 following Benjamini and Hochberg's approach for controlling the FDR were considered statistically significantly enriched, and only the top 25 significantly enriched terms based on p.adj were included in our analysis.