Single-cell transcriptomics of Hic1 lineage cells recruited during regenerative and scar-forming skin wound healing.

Adult mammalian skin wound healing is typically accompanied by fibrotic scar that impairs normal skin function and regeneration of skin appendages. Interestingly, however, in adult mice, large severe skin injuries exhibit de novo formation of HFs following severe skin injuries (a phenomenon termed wound-induced HF neogenesis, WIHN). Understanding the competent cell types and molecular mechanisms that enable regenerative wound healing will be critical for developing treatments that restore skin function after injury. We described the existence of an adult bipotent hair follicle dermal stem cell (hfDSC) that functions to regenerate the connective tissue sheath and to populate the DP with new cells (Rahmani et al., 2014). Based on this, we hypothesized that the mesenchymal cells comprised within the neogenic HFs might originate from hfDSCs. To test this, we employed αSMACreERT2:ROSAYFP and Hic1CreERT2:TDTmt mice to examine the contribution of hfDSCs or hfDSCs and reticular/hypodermal progenitors, respectively, to the formation of neodermis and regeneration of de novo HFs. Mice received full-thickness excision wounds (>1 cm2) and then harvested at 18-140 days post-wounding (dpw). αSMA+ve and Hic1-lineage cells were activated upon wounding, migrated into the wound, and contributed to both DP and DS in almost all de novo-formed HFs. Surprisingly, hfDSCs contributed only a minority of cells (20%) to nascent DP cells, whereas Hic1-lineage cells generated >90% of the neogenic DP cells. In both cases, cells integrating into neogenic HF mesenchyme appeared to restore the hfDSC pool, since they repopulated the neogenic mesenchyme over successive regenerative hair cycles. Finally, using an ex vivo HF formation assay, we found that prospectively isolated extrafollicular Hic-lineage cells could participate in HF formation when exposed to a permissive environment. Our data reveal that despite their origin in the reticular/hypodermis, Hic1-lineage dermal progenitors are able to adopt a regenerative response during wound healing if provided with a permissive local environment. To isolate cells comprising the wound neodermis, we created small (8 mm diameter) and large (1.5 cm diameter) full-thickness skin wounds on P.24-26 Hic1Cre:tdTomato+ve mice treated with tamoxifen as pups (P.4-5). We FACS collected viable Hic1tdTomato+ve cells 14 dpw from four separate cell populations in parallel. These included: 1) cells recruited in small (scar-forming) wounds, 2) cells recruited to outside (scar-forming portion) of large wound, 3) cells recruited to center (regenerative portion) of large wounds, and 4) cells within the intact (uninjured) skin. All four samples were processed according to 10X Genomics ChromiumTM Single Cell 3’ Reagent Guidelines v2 Chemistry as per the manufacturer’s protocol. In brief, single cells were sorted into 0.1% BSA–PBS and partitioned into Gel Bead-In-EMulsions (GEMs) using 10xTM GemCodeTM Technology. This process lysed cells and enabled barcoded reverse transcription of RNA, generating full-length cDNA from poly-adenylated mRNA. DynaBeads® MyOneTM Silane magnetic beads were used to remove leftover biochemical reagents, then cDNA was amplified by PCR over 10 cycles. Quality control size gating was used to select cDNA amplicon size prior to library construction. Read 1 primer sequences were added to cDNA during GEM incubation. P5 primers, P7 primers, i7 sample index, and Read 2 primer sequences were added during library construction. Quality control and cDNA quantification was performed using Agilent High Sensitivity DNA Kit. Sequencing was performed first using Illumina MiSeq SR50 to approximate the number of recovered cells in each sample. We recovered 4061, 3776, 2604, and 3293 cells for Samples 1-4, respectively, with an estimated doublet rate of ≈3%. Based on this, we determined lane distributions for sequencing using Illumina HiSeq 4000 PE (75 bp paired-end reads) with a targeted sequencing depth of ~115,000 reads/cell.