Interleukin-36y-producing macrophages drive IL-17-mediated fibrosis.

Biomaterials induce an immune response and mobilization of macrophages, yet identification and phenotypic characterization of functional macrophage subsets in vivo remain limited. We performed single-cell RNA sequencing analysis on macrophages sorted from either a biologic matrix [urinary bladder matrix (UBM)] or synthetic biomaterial [polycaprolactone (PCL)]. Implantation of UBM promotes tissue repair through generation of a tissue environment characterized by a T helper 2 (Th2)/interleukin (IL)–4 immune profile, whereas PCL induces a standard foreign body response characterized by Th17/IL-17 and fibrosis. Unbiased clustering and pseudotime analysis revealed distinct macrophage subsets responsible for antigen presentation, chemoattraction, and phagocytosis, as well as a small population with expression profiles of both dendritic cells and skeletal muscle after UBM implantation. In the PCL tissue environment, we identified a CD9 hi+ IL-36y + macrophage subset that expressed Th17-associated molecules. These macrophages were virtually absent in mice lacking the IL-17 receptor, suggesting that they might be involved in IL-17–dependent immune and autoimmune responses. Identification and comparison of the unique phenotypical and functional macrophage subsets in mouse and human tissue samples suggest broad relevance of the new classification. These distinct macrophage subsets demonstrate previously unrecognized myeloid phenotypes involved in different tissue responses and provide targets for potential therapeutic modulation in tissue repair and pathology. Overall design: Macrophages were sorted from mice quadriceps 1 week after undergoing volumetric muscle loss surgery with treatment of urinary bladder matrix, polycaprolactone, or saline. After sorting, cells were encapsulated using the 10x Chromium single cell RNA sequencing platform. Library prep followed the 10x 3' v2 protocol and were sequenced using an Illumina HiSeq 2000. Alignment was performed using 10x recommendations with the CellRanger software package to generate a counts matrix. Clusters of cells were determined using Seurat. In particular, data from the three conditions were pooled, normalized, and scaled. PCA was performed and used for clustering with a graph based method (shared nearest neighbor with the Louvaine method for community detection).