Distinct fibroblast progenitor subpopulation expedites regenerative mucosal healing by immunomodulation

Abstract

Injuries that heal by fibrosis can compromise organ function and increase patient morbidity. The oral mucosal barrier has a high regenerative capacity with minimal scarring, but the cellular mechanisms remain elusive. Here, we identify distinct postnatal paired-related homeobox-1+ (Prx1+) cells as a critical fibroblast subpopulation that expedites mucosal healing by facilitating early immune response. Using transplantation and genetic ablation model in mice, we show that oral mucosa enriched with Prx1+ cells heals faster than those that lack Prx1+ cells. Lineage tracing and scRNA-seq reveal that Prx1+ fibroblasts exhibit progenitor signatures in physiologic and injured conditions. Mechanistically, Prx1+ progenitors accelerate wound healing by differentiating into immunomodulatory SCA1+ fibroblasts, which prime macrophage recruitment through CCL2 as a key part of pro-wound healing response. Furthermore, human Prx1+ fibroblasts share similar gene and spatial profiles compared to their murine counterpart. Thus, our data suggest that Prx1+ fibroblasts may provide a valuable source in regenerative procedures for the treatment of corneal wounds and enteropathic fibrosis.

Graphical abstract

Figure 1.

Expedited wound healing in Prx1⁺ cell–enriched gingiva. (A) Left, macrophotograph of mouse palatal rugae numbered from anterior to posterior direction. Right, whole tissue imaging for red fluorescence in Prx1^CreERT:R26R^tdTomato mice that received tamoxifen injection for 5 d. Scale bar, 1 mm. N = 3 mice from two independent experiments. (B) Representative immunofluorescent images showing tdTomato⁺ signals in sagittal sections of R1–R3 (green box; magnified view is shown in inset, orange box) and R3–R6 (blue box). Animals were euthanized 2 wk after last tamoxifen injection. Scale bar, 250 μm. N = 3 mice from two independent experiments. (C) Macrophotograph of 1 mm palatal wounds created in Prx1⁺-enriched R2 or Prx1-deficient R4 after 2, 4, and 6 d of healing. Scale bar, 1 mm; M1, first molar tooth. Dashed line demarcates clinical wound outline. N = 7 mice for each time point, three independent experiments. (D) Masson trichrome-stained section images of R2 and R4 wounds. Scale bar, 0.5 mm. N = 7 mice for each time point, three independent experiments. (E and F) Quantification of open wound area percent (E) and epithelial gap closure distance (F) during wound healing. Each dot represents one wound location. Data represent mean ± SEM (N = 7 mice for each time point for a total of 21 mice, three independent experiments). Student’s t test; **P < 0.01, ***P < 0.001. (G) Quantification of new collagen formation by measuring blue-stained area in the wound bed from trichrome sections. Each dot represents one wound location. Data represent mean ± SEM (N = 7 mice for each time point for a total of 21 mice, three independent experiments). Student’s t test; *P < 0.05, ***P < 0.001. (H) Schematic diagram of gingival tissue engraftment from R2 to R4 to enrich Prx1⁺ cells or from R4 to R4 as control group in Prx1^{CreERT-eGFP mice}. Representative immunofluorescent images validating incorporation of GFP⁺ cells. Scale bar, 0.5 mm. N = 3 mice each from two independent experiments. (I) Representative trichrome-stained images of day 2 wounds after R2 or R4 engraftment. N = 7–8 mice each from three independent experiments. (J and K) Quantification of granulation tissue (J) and new collagen formation areas (K) in wounds created after R2 or R4 autograft. Each dot represents one mouse, and data represent mean ± SEM; N = 7–8 mice per group for a total of 15 mice, three independent experiments. Student’s t test; **P < 0.01. (L) Macrophotograph of day 4 wounds in experimental Prx1^CreERT:R26R^DTA/+ (DTA^Prx1) or control littermate mice (Prx1^CreERT). Scale bar, 1 mm; dashed line demarcates clinical wound outline. N = 6–7 mice each from three independent experiments. (M and N) Representative images of COL3⁺ stained area within the lamina propria (demarcated with dashed line) in control and DTA^Prx1 mice (M), and quantification of stromal healing by measuring COL3⁺ area in the wound bed at day 4 (N). Scale bar, 1 mm. Each dot represents one mouse, and data represent mean ± SEM; N = 6–7 mice per group for a total of 13 mice, three independent experiments. Student’s t-test; ***P < 0.001.

Figure S1.

Characterization of postnatal Prx1^enh+ cells and validation in DTA^Prx1 mice. (A) Left, immunofluorescence experiments using PRRX1 antibody to detect immunopositive cells in R2 and R4 of wildtype C57BL/6 (B6) mice. Scale bar, 100 μm. Right, quantification of PRRX1⁺ cells in R2 and R4 of B6 mice. Each dot represents one wound location, and data represent mean ± SEM. N = 6 mice, three independent experiments. Student’s t test; ***P < 0.001. (B) PRRX1 immunopositive cells detected by immunofluorescence in R2 of Prx1^CreERT-eGFP mice. Arrows point to double-positive PRRX1⁺GFP⁺ cells whereas arrowheads point to single PRRX1⁺ cells. Scale bar, 100 μm. Right, quantification of double PRRX1⁺GFP⁺, single PRRX1⁺, and single GFP⁺ cell percentage, normalized per total PRRX1⁺ cell count or per total GFP⁺ cell count. N = 3 mice, two independent experiments. (C) Left, graphical diagram of mouse palate from which anterior (R1–3) or posterior (R3–8) gingivae were collected for flow cytometry. Right, GFP⁺ cell percentage in single cell suspension prepared from anterior or posterior palate tissue. N = 3 mice, two independent experiments. (D) Left, paraffin-embedded sections from palatal gingiva corresponding to second rugae of Prx1^CreERT:R26R^tdTomato mice. Tissues were stained with RFP antibody and vimentin (green, top) or CD31 (green, bottom) antibody. Scale bar, 50 μm. Right, quantification of tdTomato⁺ cells that co-express vimentin or CD31. Each dot represents one mouse, and data represent mean ± SEM. N = 4 mice, two independent experiments. (E) Representative immunofluorescence images in buccal mucosa, jejunum, and colon from Prx1^eGFP mice. Cryosections were stained with antibody against GFP to detect postnatal Prx1⁺ cells and anti-EpCam antibody to distinguish intestinal epithelium in jejunum and colon specimens. Dotted lines demarcate epithelial–lamina propria–muscularis propria borders in buccal mucosa. Scale bar, 100 μm. N = 3 mice, two independent experiments. (F–I) Immunofluorescence experiments validating Prx1⁺ fibroblast ablation in experimental Prx1^CreERT:R26R^DTA/+ (DTA^Prx1) mice by examining unwounded palatal rugae. (F) Quantification of GFP⁺ Prx1⁺ cells in Prx1^CreERT-GFP or Prx1^CreERT-eGFP:R26R^DTA that received tamoxifen. Each dot represents one mouse, and data represent mean ± SEM. N = 6 mice per group for a total of 12 mice, three independent experiments. Student’s t test; ***P < 0.001. (G) Quantification of CD45⁺ leukocytes (red) in epithelial (K14⁺, green) or lamina propria compartments. Each dot represents one mouse, and data represent mean ± SEM. N = 6 mice per group for a total of 12 mice, three independent experiments. Student’s t test. (H) Quantification of CD31⁺ endothelial cells (red) and the number of blood vessels. Each dot represents one mouse, and data represent mean ± SEM. N = 6 mice per group for a total of 12 mice, three independent experiments. Student’s t test. (I) Quantification of αSMA⁺ pericytes (green). Each dot represents one mouse, and data represent mean ± SEM. N = 6 mice per group for a total of 12 mice, three independent experiments. Student’s t test. (J) Immunofluorescent signal for COL3 in littermate controls (Prx1^CreERT) and experimental DTA^Prx1 mice in day 4 wounds of the R4 area. Scale bar, 1 mm; dashed line marks COL3-immunopositive area in wound bed. N = 6 mice per group, three independent experiments. (K) Quantification of epithelial gap closure (left) and COL3⁺ area (right) in day 4 wounds. Each dot represents one mouse, and data represent mean ± SEM. N = 6 mice per group for a total of 12 mice, three independent experiments. Student’s t test.

Figure 2.

scRNA-seq analysis of oral fibroblasts and Prx1⁺ cells in steady state. (A) Experimental diagram for scRNA-seq analysis of oral mucosa derived from the hard palate of C57BL/6 mice. (B) Left, UMAP plot of CD45⁻ gingival cells isolated from 8-wk-old mice. Right, UMAP plot of oral fibroblast clusters, selected based on putative extracellular matrix gene expression. (C) Dot plot showing enriched expression of top five genes from each six oral fibroblast clusters. (D) GO analysis for fibroblast clusters 1–4, truncated to five biological processes. Full list is provided in Table S1. (E) Dot plot showing enrichment of mesenchymal progenitor cell–associated genes Prrx1, Mx1, Lepr, and Axin2 in oral fibroblast clusters. (F) Gene module scores for Wnt activator, inhibitor, and receptor genes. (G) Violin plots of Wnt-associated genes significantly upregulated in cluster 4, as determined by Wilcoxon rank-sum analysis. (H) Real-time qPCR of cluster 4–specific genes (Prrx1, Sfrp2, Wif1, Slpi) comparing GFP⁺ Prx1⁺ fibroblasts versus GFP⁻ fibroblasts sorted from the anterior palate of Prx1^eGFP mice. Data represent mean ± SEM. Fold changes in mRNA expression (N = 3 mice) from three independent experiments are shown. Student’s t test; **P < 0.01, ***P < 0.001.

Figure S2.

scRNA-seq analysis of murine CD45⁻ gingival cells. (A) Left, schematic diagram for scRNA-seq data quality control measurements. Right, mitochondrial percent, number of features, and UMI counts to exclude cell counts with outlier values (>threefold from median value). (B) Heatmap of differentially expressed genes from each cluster. Upregulated genes are indicated on the y axis. (C) Violin plots for genes known to be putatively expressed for specific cell type (Col3a1, Dcn: fibroblasts, Acta2: smooth muscle/pericytes, Acan: chondrocytes, Pecam1: endothelium, Lyve1: lymphatics, Bglap2: osteoblasts, Mpz: neuronal, Sntn: ciliated epithelia, H2-Ab1: phagocytic). (D) Immunophenotyping of lineage-negative (Lin⁻: CD45⁻CD31⁻Epcam⁻Ter119⁻) Prx1⁺ cells for the expression of pan-fibroblast marker (PDGFRA⁺) and mesenchymal progenitor markers (CD90, CD73, and CD105). Isotype IgG was used as negative control. Right, percent positive for each tested marker within GFP⁺ cell population. Each dot represents one mouse, N = 4 mice from two independent experiments are shown.

Figure 3.

Prx1⁺ fibroblasts function as fibroblast progenitors during homeostasis. (A) Pseudotime analysis of three major trajectories with Prrx1^high cluster seeded as starting root. (B) Ly6a, Pi16, or Acan expression paired with Prrx1 expression along calculated pseudotime of each trajectory. (C) Experimental scheme for in vivo lineage tracing of postnatal Prx1⁺ cells. (D) Immunofluorescent images in Prx1^CreERT-eGFP:R26R^tdTomato mice chased for 1, 4, and 16 wk after last tamoxifen (TAM) injection. Scale bar, 250 μm; inset scale bar, 50 μm. N = 5 mice per each endpoint for a total of 15 mice, three independent experiments per endpoint. (E) Quantification of tdTomato⁺ to GFP⁺ cell number ratio. Each dot represents one mouse, and data represent mean ± SEM; N = 5 mice per each end point for a total of 15 mice, three independent experiments. One-way ANOVA and Tukey’s post-hoc test; **P < 0.01, ***P < 0.001. (F) GFP⁺ cell number normalized to total nucleated cells in mice traced for 1, 4, and 16 wk after tamoxifen. Each dot represents one mouse, and data represent mean ± SEM; N = 5 mice per each end point for a total of 15 mice, three independent experiments. One-way ANOVA and Tukey’s post-hoc test. (G) Flow cytometry analysis of Prx1-lineage cells from Prx1^CreERT:R26R^tdTomato mice traced for 4 wk. N = 6 mice, three independent experiments. (H) Pie chart quantification of Prx1-lineage cell identity by the expression of SCA1, PDGFRA, and/or CD146. Average percent of tdTomato⁺ numbers ± SEM is shown from N = 6, three independent experiments.

Figure 4.

Preferential differentiation of Prx1⁺ progenitors toward SCA1⁺ oral fibroblasts during wound healing. (A) Flow cytometry analysis and quantification of GFP⁺ Prx1⁺ cells expressing proliferation marker Ki67 compared to GFP⁻ PDGFRA⁺ pan-fibroblasts. Each dot represents pooled cells from two mice, and data represent mean ± SEM; N = 4 data points from a total of eight mice, two independent experiments. Student’s t test; *P < 0.05. (B) Representative images for Prx1-lineage cells (red) and EdU incorporation (green) in day 4 wound from Prx1^CreERT:R26R^tdTomato mice. N = 3 mice from two independent experiments. Scale bar, 1 mm. Inset, arrows point to some EdU⁺tdTomato⁺ cells, dashed line demarcates epithelia–lamina propria border. (C) Quantification of tdTomato⁺ cells per mm² in day 4 wounds created in R2, R3, or R4 from Prx1^CreERT:R26R^tdTomato mice. Data represent mean ± SEM; N = 3 mice from two independent experiments. (D) Flow cytometry analysis of Prx1-derived lineage cells 4 d after wounding. Right, quantification of Prx1-lineage cell identity during wound healing by the expression of SCA1, PI16, PDGFRA, and/or CD146. FB, fibroblast; neg, negative. Each dot represents one mouse, and data represent mean ± SEM; N = 4 mice from two independent experiments. (E) Representative immunofluorescent images of day 4 wound bed showing SCA1⁺tdTomato⁺ fibroblasts (white arrowheads) and some SCA1⁻tdTomato⁺ cells (orange arrowheads). Scale bar, 50 μm; N = 3 mice from two independent experiments. (F) Flow cytometry analysis of myofibroblasts (lineage⁻αSMA⁺PDGFRA⁺) in day 4 wounds. Right, quantification of percent myofibroblasts that are tdTomato⁺ or tdTomato⁻. Each dot represents one mouse, and data represent mean ± SEM; N = 8 mice from three independent experiments. Student’s t test; ***P < 0.001. (G) Representative immunofluorescent images of day 2 wounds after R2 or R4 graft incorporation, stained with SCA1 and vimentin antibodies. Arrows point to initial wound edges. Scale bar, 0.5 mm. Inset, arrowheads point to SCA1⁺vimentin⁺ cells. N = 5–6 each from two independent experiments. (H) Quantification of cell numbers per mm² for pan-fibroblasts (pan-FB; vimentin⁺ spindle-shaped cells), SCA1⁺ FB (SCA1⁺vimentin⁺), PI16⁺ (PI16⁺vimentin⁺), and pericytes (CD146⁺CD31⁻). Each dot represents one mouse, and data represent mean ± SEM. N = 5–6 mice per group for a total of 11 mice from two independent experiments. Student’s t test; *P < 0.05, **P < 0.01.

Figure 5.

Identification of highly chemotactic Ly6a/SCA1⁺ oral fibroblasts. (A) Single-cell expression of Ly6a, Ccl2, Ccl7, and Cxcl1 in oral fibroblast subsets of the mouse palate. (B) Flow cytometry analysis of CCL2 expression in SCA1⁻ and SCA1⁺ fibroblasts (CD31⁻CD45⁻Epcam⁻Ter119⁻PDGFRA⁺) from unwounded and day 2 wounds in C57BL/6 mice. Right, quantification of fibroblast (FB) percent positive for CCL2. Each dot represents one mouse, and data represent mean ± SEM. N = 5 mice per endpoint for a total of 10 mice from three independent experiments. Two-way ANOVA and Tukey’s post-hoc test; ***P < 0.001. (C) Pie chart of cell percentages that express CCL2. Average percent ± SEM is shown from N = 5 mice for each endpoint, a total of 10 mice from three independent experiments. (D) Flow cytometry plot gated for live CD45⁺CD11b⁺ cells to quantify F4/80⁺ (macrophages) and Ly6g⁺ (neutrophils) cells in R2 and R4 wounds 2 d after wounding. N = 6–8 mice per endpoint from three independent experiments for each endpoint. (E and F) Quantification of neutrophil (E) and macrophage (F) infiltration to the R2 and R4 wounds during wound healing. Each dot represents one wound location. Data represent mean ± SEM (N = 6–8 mice for each time point for a total of 20 mice, three independent experiments). Student’s t test; *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. (G–J) Representative flow cytometry plots (G and I) and quantification of pro-inflammatory macrophages co-expressing IL1β (H) and pro-resolving macrophages co-expressing TGFβ1 (J) in healing R2 and R4 wounds. Each dot represents pooled cells from the designated wound location of one mouse. Data represent mean ± SEM (N = 6–8 mice for each time point for a total of 20 mice, three independent experiments). Student’s t test; *P < 0.05, ***P < 0.001.

Figure S3.

Validation of oral fibroblast heterogeneity in mouse palate. (A) H&E staining of mouse palatal rugae. (B) Immunofluorescence experiment with SCA1 (red) and PDGFRA (green) antibodies. N = 3 mice, two independent experiments. Scale bar, 100 μm; dashed lines demarcate borders between epithelium, papillary and reticular lamina propria, and submucosa. (C) Quantification of SCA1 mean fluorescent intensity (MFI) in the papillary, reticular, and submucosal layers. Each dot represents one mouse, and data represent mean ± SEM. N = 3 mice, two independent experiments. One-way ANOVA and Tukey’s post-hoc test; **P < 0.01. (D) Immunofluorescence experiment with PI16 (red) and PDGFRA (green) antibodies. N = 3 mice, two independent experiments. Scale bar, 100 μm; dashed lines demarcate borders between epithelium, papillary and reticular lamina propria, and submucosa. (E) Quantification of PI16 MFI in the papillary, reticular, and submucosal layers. Each dot represents one mouse, and data represent mean ± SEM. N = 3 mice, two independent experiments. One-way ANOVA and Tukey’s post-hoc test; ***P < 0.001. (F) Immunofluorescence experiments with anti-RFP (to detect mCherry, red) and K14 (epithelial marker, green) antibodies within the palatal gingiva of CCL2^mCherry mice. Representative image of N = 3 mice from two independent experiments is shown. Scale bar, 100 μm; dashed line demarcates submucosal layer. (G) Left, representative images of immunofluorescence stained with CXCL1 (green) and SCA1 (red) antibodies. Scale bar, 100 μm; dashed lines demarcate epithelial–lamina propria border. Right, CXCL1 MFI in epithelium and lamina propria. Each dot represents one mouse, and data represent mean ± SEM. N = 3 mice, two independent experiments. Student’s test; ***P < 0.001. (H) Flow cytometry analysis and gating strategy for CCL7⁺ expressing gingival cells. N = 4 mice from two independent experiments. (I) Left, percent of CCL7⁺ cells that are PDGFRA⁺ (fibroblast identity) or PDGFRA⁻ (non-fibroblast); right, percent of CCL7⁺ fibroblasts that are SCA1⁺ or SCA1⁻. Each dot represents one mouse, and data represent mean ± SEM. N = 4 mice, two independent experiments. Student’s test; **P < 0.01, ***P < 0.001. Epi, epithelium; PL, papillary layer; RL, reticular layer; SM, submucosa; AU, arbitrary unit.

Figure 6.

Heightened inflammation and delayed healing by Prx1⁺ cell ablation or CCL2 deletion. (A) Left, representative immunofluorescent images of SCA1 (green) and vimentin (red) staining in day 4 wound of DTA^Prx1 mice and control littermates. Arrowheads point to SCA1⁺vimentin⁺-immunopositive spindle-shaped cells. Right, quantification of SCA1⁺vimentin⁺ spindle-shaped cells in the wound bed per mm² area. Scale bar, 100 μm. Each dot represents one mouse, and data represent mean ± SEM. N = 6 mice per group for a total of 12 mice, three independent experiments. Student’s t test; ***P < 0.001. (B) F4/80⁺ macrophages (left) and Ly6g⁺ neutrophils (right) per live cells during wound healing in control and DTA^Prx1 mice. Each dot represents one mouse, and data represent mean ± SEM. N = 5–7 mice per group and time point for a total of 38 mice (14 for day 2, 13 for day 4, and 11 for day 6), three independent experiments. Student’s t test; *P < 0.05. (C and D) Percentage of macrophages that are pro-inflammatory in day 2 wound (C) or that are pro-resolving in day 6 wound (D) in control and DTA^Prx1 mice. Each dot represents one mouse, and data represent mean ± SEM. N = 5–7 mice per group for a total of 14 for day 2 (C) and 11 for day 6 (D), three independent experiments. Student’s t test; *P < 0.05, **P < 0.01. (E) Quantification of SCA1⁺ fibroblast percent that express CCL2 in Prx1^CreERT:Ikbkb^f/f littermate mice that received vehicle (corn oil, +veh) or tamoxifen (+tam). Analysis in day 2 wound is shown. Each dot represents one mouse, and data represent mean ± SEM. N = 6–7 mice per group for a total of 13 mice, three independent experiments. Student’s t test; ***P < 0.001. (F) F4/80⁺ macrophage infiltration to healing wounds in Prx1^CreERT:Ikbkb^f/f mice that received vehicle or tamoxifen. Each dot represents one mouse, and data represent mean ± SEM. N = 6–7 mice per group and time point for a total of 26 mice (14 for day 2 and 12 for day 6), three independent experiments. Student’s t test; **P < 0.01. (G and H) Quantification of pro-inflammatory macrophage percent in day 2 wound (G) and pro-resolving macrophages in day 6 wound (H) in Ikbkb-deleted mice. Each dot represents one mouse, and data represent mean ± SEM. N = 6–7 mice per group for a total of 14 mice for day 2 (G) and 12 for day 6 (H), three independent experiments. Student’s t test; *P < 0.05, ***P < 0.001. (I) Quantification of epithelial gap closure in day 2 wound (left) and COL3⁺ immunopositive area in day 4 wound bed (right) in Prx1^CreERT:Ikbkb^f/f mice that received vehicle or tamoxifen. Each dot represents one mouse, and data represent mean ± SEM. N = 5–7 mice per group for a total of 24 mice (12 for day 2 and 12 for day 4 wound), three independent experiments. Student’s t test; ***P < 0.001. (J) CCL2 expression in unwounded gingiva (R1) from control CCL2^f/f-mCherry littermates and experimental Prx1^CreERT:CCL2^f/f-mCherry (△CCL2^Prx1) mice 2 wk after tamoxifen injection (left). Right, quantification of CCL2⁺ fibroblasts in day 4 wounds by flow cytometry. Scale bar, 50 μm; SM, submucosa. CCL2 was visualized by staining with primary anti-RFP antibody followed by secondary antibody conjugated to AlexaFluor 647. N = 3–4 mice per group for a total of seven mice, two independent experiments. (K) Flow cytometry analysis of F4/80⁺ macrophage infiltration in day 2 wounds of control and △CCL2^Prx1 mice. Right, quantification of macrophages per total live cell. Each dot represents one mouse, and data represent mean ± SEM. N = 5–6 mice per group for a total of 11, three independent experiments. Student’s t test; *P < 0.05. (L) Quantification of percent pro-inflammatory macrophages (CD45⁺CD11b⁺F4/80⁺IL1β⁺), normalized to total macrophage cell numbers in △CCL2^Prx1 mice in day 2 wounds. Each dot represents one mouse, and data represent mean ± SEM. N = 5–6 mice per group for a total of 11, three independent experiments. Student’s t test; *P < 0.05. (M) Quantification of epithelial gap closure in day 2 wounds and COL3⁺ area in day 4 wounds in control, △CCL2^Prx1 mice, and △CCL2^Prx1 mice that received recombinant CCL2 (+rCCL2). Each dot represents one mouse, and data represent mean ± SEM. N = 5–11 mice per group for a total of 16 for day 2 wound and 27 for day 4 wound, three independent experiments. Student’s t test; *P < 0.05, **P < 0.01, ***P < 0.001.

Figure 7.

Improved skin wound healing by adoptive Prx1^enh+ cell transfer. (A) Schematic experimental approach using Prx1^CreERT:R26R^tdTomato mice as donor and recipient groups after tamoxifen administration. (B) Representative immunofluorescent images of unwounded and day 4 scalp skin wound in Prx1^CreERT:R26R^tdTomato mice that received saline (+vehicle) or Prx1^enh+ cell grafting from littermate Prx1^CreERT:R26R^tdTomato mice. N = 3 mice per treatment and endpoint for a total of 12 mice, two independent experiments. (C) Quantification of Prx1⁺-lineage tdTomato⁺ cells in unwounded and wounded scalp skin. Each dot represents one mouse, and data represent mean ± SEM. N = 3 mice per treatment and endpoint for a total of six for unwounded and six for day 2 wound, two independent experiments. (D) Representative microphotographs of 1 mm scalp wound in C57BL/6 mice that received either vehicle or Prx1^enh+ cells prior to wounding. Donor cells were derived from Prx1^CreERT-eGFP mice. Scale bar, 1 mm. N = 6–7 mice per treatment for a total of 13, three independent experiments. (E) Longitudinal quantification of open skin wound area during wound healing process. Each dot represents one mouse within time point, and data represent mean ± SEM. N = 6–7 mice per treatment for a total of 13 mice, three independent experiments. Student’s t test; **P < 0.01, ***P < 0.001. (F) Left, H&E-stained sections from day 6 skin wounds that received either vehicle (+veh) or cell transfer (+Prx1^enh+). Arrows designate original 1 mm wound length; scale bar, 0.5 mm. Right, quantification of hair follicle numbers per 1 mm length within wound bed. Each dot represents one mouse, and data represent mean ± SEM. N = 6–7 mice per treatment for a total of 13 mice, three independent experiments. Student’s t test; **P < 0.01. (G) Left, representative immunofluorescent images of skin wounds that were stained with antibody against αSMA. Dashed line demarcates epithelium–dermis border. Scale bar, 100 μm. Right, quantification of αSMA-immunopositive area in mm². Each dot represents one mouse, and data represent mean ± SEM. N = 6–7 mice per treatment for a total of 13 mice, three independent experiments. Student’s t test; **P < 0.01. (H) Representative flow cytometry plots pre-gated for live CD45⁺CD11b⁺ cell populations in day 6 skin wounds that received vehicle or Prx1^enh+ cells. N = 6–7 mice per treatment for a total of 13, three independent experiments. (I) Quantification of neutrophils (PMN, CD45⁺CD11b⁺Ly6g⁺) and macrophages (CD45⁺CD11b⁺F4/80⁺) in day 6 wounds, normalized to total CD45⁺ leukocyte population. Each dot represents one mouse, and data represent mean ± SEM. N = 6–7 mice per treatment for a total of 13 mice, three independent experiments. Student’s t test; *P < 0.05. (J) Quantification of pro-resolving macrophages that express TGFβ1 in day 6 wounds of vehicle or Prx1^enh+-treated mice, normalized to total macrophage cell numbers. Each dot represents one mouse, and data represent mean ± SEM. N = 6–7 mice per treatment for a total of 13 mice, three independent experiments. Student’s t test; *P < 0.05.

Figure 8.

Transcriptomic profile and spatial distribution of human Prx1⁺ oral fibroblasts. (A) Graphical diagram of human oral biopsy locations from the occlusal view of maxilla. (B) Representative immunofluorescent images of each oral biopsy specimen stained with PRRX1 (red) and vimentin (green) antibodies. Scale bar, 100 μm. N = 5–6 human specimen per group for a total of 26 specimens, three independent experiments. (C) Quantification of PRRX1⁺vimentin⁺ spindle-shaped cell percentage per total spindle-shaped fibroblasts. Each dot represents one human specimen, and data represent mean ± SEM. N = 5–6 specimens for a total of 26 samples, three independent experiments. One-way ANOVA and Tukey’s post-hoc test; *P < 0.05. (D) UMAP of single cells derived from the anterior rugae of human oral biopsy (left) and fibroblast sub-clusters (right). (E) Ridge plot of PRRX1 gene expression from oral fibroblast subclusters. (F) GO analysis for biological processes of PRRX1-enriched fibroblast cluster 1. Full list is provided in Table S4. (G) Gene module scores for Wnt activator, inhibitor, and receptor genes. (H) Violin plots of Wnt-associated genes significantly upregulated in cluster 1, as determined by Wilcoxon rank-sum analysis. (I) Monocle-based trajectory and pseudotime calculation of fibroblast subclusters with PRRX1^high selected as a starting node. (J) Feature plot of single-cell expression of CCL2 and CXCL1 in oral fibroblast clusters. (K) Normalized gene expression of PRRX1, SFRP2, CCL2, and CXCL1 along calculated pseudotime.

Figure S4.

scRNA-seq analysis of human anterior palatal gingiva. (A) UMAP of putative genes highly expressed in each cluster identified. FB, fibroblast. (B) UMAP of myeloid and fibroblast subsets. Mono, monocyte; DC, dendritic cells; Mac1, inflammatory macrophage. (C) CellChat analysis of single cells for CCL and CXCL signaling networks between fibroblast and myeloid subsets. (D) Relative contribution of CCL (left) and CXCL (right) ligand to known receptor interaction among fibroblast-myeloid cell subsets. (E) Violin plots of ligand and receptor genes (CCL2, ACKR1, CXCL8, and CXCR2) implicated as the highest contributors of CCL and CXCL signaling interactions.