Fig 1: Epidermal SEMA3C is impaired in human diabetic wounds, but its topical reintroduction restores re-epithelization of diabetic skin ulcers in vivo.(A–D) SCT-normalized mean expression of (A) SEMA3C and (B) SOX6 in human keratinocytes, and (C) NRP1 and (D) NRP2 in human endothelial cells (GSE166120). Data shown for healthy skin, diabetic skin without ulcers, healing diabetic foot ulcers, and non-healing diabetic foot ulcers. Each dot represents a sequencing run. Healthy control n = 11, diabetic non-ulcer n = 15, diabetic healing ulcer n = 9, diabetic non-healing ulcer n = 4. Statistical significance was determined using a Wilcoxon rank-sum test (p < 0.05 = *, p < 0.01 = **).(E) Schematic overview of the large diabetic wound ulcer model. 1-cm full-thickness wounds were generated in Leprdb/J diabetic mice before topical treatment with recombinant SEMA3C protein (human or mouse) or vehicle control at D0, D2, and D4 post-wounding. Wounds were harvested for closure and revascularization analysis at D14PW.(F–H) Representative dorsal wound images from (F) vehicle, (G) human SEMA3C, and (H) mouse Sema3c treatment groups at D4PW, D7PW, and D14PW. White dashed lines mark initial wound boundaries; yellow lines indicate wound diameter at imaging analysis. Images represent three independent biological replicates per condition (n=3).(I) Quantification of mean wound closure rate in WT and Leprdb/J mice treated with vehicle, human SEMA3C, or mouse Sema3c. Measurements taken at days D0, D2, D4, D7, D9, and D14PW. Data represents the mean of three independent biological replicates per condition (n=3), besides Leprdb/J – Vehicle which is two biological replicates. Statistical significance was determined by two-way ANOVA (p < 0.05 = *, p < 0.01 = **). Significance is marked between Leprdb/J -mSema3c treated vs Leprdb/J – Vehicle.(J) Quantification of D7PW wound closure rate in WT and Leprdb/J mice treated with vehicle, human SEMA3C, or mouse Sema3c. Data represents the mean of three independent biological replicates per condition (n=3), besides Leprdb/J – Vehicle which is two biological replicates. Statistical significance was determined by two-way ANOVA (p < 0.05 = *, p < 0.01 = **).(K-M) Representative ventral-view images of subcutaneous wounds in Leprdb/J mice at D14PW, treated with (K) vehicle, (L) human SEMA3C, or (M) mouse Sema3c for each condition. Boxed region highlights subcutaneous wound bed and periwound region. (Right) Enlarged view of the boxed region. White dashed lines mark the original wound boundaries, and yellow lines indicate the central wound bed. Yellow arrowheads indicate regions with active revascularization.Images represent three independent biological replicates per condition (n=3).(N-O) Periwound revascularization coverage in Leprdb/J mice at D14PW after treatment with vehicle, human SEMA3C, or mouse Sema3c. (O) Proportion of periwound revascularization at D14PW in each treatment group. Data represents the mean of three independent biological replicates per condition (n=3). Statistical significance was determined by two-way ANOVA (p < 0.05 = *, p < 0.01 = **).(P) Proposed model summarizing the role of the Sema3c-Nrp1/Nrp2 signaling axis in healthy versus diabetic skin wound healing. In healthy wounds, activation of Sox6+ Basal keratinocytes promote Sema3c secretion and paracrine signaling to Nrp1/Nrp2+ endothelial cells, facilitating re-epithelialization and revascularization of the wound bed. This circuit is disrupted in diabetic wounds, leading to reduced Sox6+ Basal keratinocyte activation and decreased Sema3c expression. This results in impaired re-epithelization and periwound vascular repair. However, topical reintroduction of Sema3C restores periwound revascularization and wound re-epithelization in diabetic skin ulcers.
Fig 2: Sox6+ basal keratinocytes are master controllers driving wound re-epithelization through dynamic epidermal–neurovascular crosstalk.(A) Schematic illustration of cross-tissue wound signaling between epidermal Central Orchestrators and deep tissue population populations.(B–E) Visium HD spatial localization of Basal IV keratinocytes (blue) and proliferative endothelial cells (yellow) across the healing time course (B) UW, (C) D4PW, (D) D7PW, (E) D30PW. Images represent one biological replicate, with one technical replicate shown for UW. Red arrowheads indicate the suprabasal wound edge; red guidelines denote subcutis wound boundaries. Yellow boxes indicate magnified insets highlighting Basal IV–endothelial interactions at the wound edge. Scale bar, 1 mm.(F–I) Quantification of anatomical distance between Basal IV and proliferative endothelial cells along the anterior–posterior wound axis (1 a.u. = 9 mm) at (F) UW, (G) D4PW, (H) D7PW, (I) D30PW. Data represent two technical replicates from one biological sample per timepoint. Significance was assessed using a variance-based localization test with Benjamini–Hochberg correction (* p < 0.05; ** p < 0.01; *** p < 0.001).(J–K) CellChat-inferred signaling network (J) and corresponding ligand–receptor pathways (K) transmitted between Basal IV keratinocytes and endothelial cells. Edge width denotes interaction strength.(L) Visium HD spatial transcriptomic expression of Sema3c pathway components (Sema3c, Nrp1, Nrp2, Plxna4) overlaid on corresponding H&E sections at D7PW. Scale bar, 1 mm.(M–N) Single-molecule RNA FISH (smFISH) immunofluorescence showing Sox6 (magenta) and Sema3c (yellow) mRNA transcripts with DAPI (white) in (M) UW and (N) D4PW skin. Boxed regions highlight an Sox6-high epidermal zone. (Right) zoomed in images of the Sox6-high zone with single-channel panels shown. Wound edge (w.e.) indicated by red arrowheads and subcutis wound boundaries indicated by red lines. Scale bars, 100 μm (full images) or 10 μm (zoomed insets).(O–P) Boxplots quantifying mean fluorescence intensity of (O) Sox6 and (P) Sema3c smFISH signals across timepoints (UW, D4PW, D7PW), comparing unwounded, distal, and proximal wound regions. UW includes three biological replicates, D4PW two biological replicates, and D7PW one biological replicate. Statistical testing performed using a two-sided Wilcoxon test. Variability is represented using the interquartile range (IQR). Statistical significance was determined using a Wilcoxon rank-sum test (p < 0.05 = *, p < 0.01 = **).(Q) UMAP visualizations and corresponding pseudotime ordering of UW IFE keratinocyte subclusters from OWHA snRNA-seq. n = two biological replicates.(R) Heatmap of top 100 pseudotime-associated genes in UW snRNA-seq keratinocytes ordered by cluster and pseudotime. Arrowheads mark genes enriched in early (purple) vs. late (yellow) pseudotime.(S) UMAP visualizations and corresponding pseudotime ordering of D4PW–D7PW IFE keratinocyte subclusters showing branching into re-epithelization (red) and neurovasculogenesis (green) lineage trajectories. n = two biological replicates per timepoint.(T) Heatmap of the top 100 pseudotime-associated genes expressed along the neurovasculogenesis pseudotime trajectories in D4PW–D7PW keratinocytes, ordered by cluster and timepoint. Arrows indicate early (purple) and late (yellow) pseudotime gene signatures.(U–W) Mean pseudotime expression profiles of (U) Sox6, (V) Sema3c, and (W) Krt6a along the D4PW–D7PW proliferative pseudotime trajectory.(X) Gene Ontology terms among genes upregulated at early pseudotime stages of the neurovasculogenesis trajectory, subdivided into molecular function (MF), biological process (BP), and cellular component (CC) categories.
Fig 3: Sox6+ Basal IV keratinocytes are active in human wound healing but entirely missed by conventional scRNA-seq capture.(A) UMAP embedding of the fully integrated cross-species wound-healing atlas (COWA) containing 236,930 single cells. COWA integrates the multimodal OWHA dataset with the human wound healing dataset (GSE241132), generating a total of 40 sequencing runs (28 murine, 12 human).(B) Annotation legend displaying all 107 fine-grained subcluster identities represented in COWA.(C–D) UMAP projections showing (C) human (n=12) and (D) mouse (n=28) wound datasets downsampled to 43,406 cells each for comparable visualization of interspecies cell capture differences.(E) Sankey plot showcasing proportional representation of major cell types between mouse and human samples across metaclusters, highlighting shared and species-specific tissue compositions. Percentages display proportion of each respective metacluster assigned to integrated cross-species Harmony clusters.(F) MILO differential abundance analysis of select IFE keratinocyte, endothelial, pericyte, and Schwann cell subclusters states significantly enriched in mouse (blue) or human (red) samples (FDR = 0.15). Mouse-enriched (>0 LFC) states represent populations preferentially detected in by multimodal murine atlasing.(G–H) UMAP feature plots depicting (G) mouse Sox6 and (H) human SOX6 expression.(I–J) UMAP feature plots of (I) mouse Sema3c and (J) human SEMA3C, showing substantially reduced SEMA3C detection in human wound-healing scRNA-seq datasets.(K–L) Dot plots of Basal IV–associated markers (Sox6, Cdh13, Sema3c) and the wound-induced keratinocyte gene Krt16 in IFE keratinocytes from (K) the murine OWHA dataset and (L) the human wound dataset. Dot size indicates the percentage of expressing cells; color intensity denotes mean expression level.(M) Beeswarm plot of MILO differential abundance comparing whole-cell (blue) versus single-nucleus (red) RNA-seq, highlighting modality-specific biases in cell type capture (FDR = 0.15).(N) Visium spatial transcriptomic feature plot of D7PW human wound tissue (GSE241132), annotated by regional labels as defined Liu et al. 2024. Red arrowheads mark the wound edge.(O–P) Quantification of Basal IV module scores in (O) unwounded and (P) D7PW human tissue sections (calculated using AddModuleScore()). Red box highlights wound edge region zoomed-in in the right panel. Red arrowheads denote the wound edge; white dashed lines outline the epidermal-dermal boundary.(Q–R) Quantification of SEMA3C pathway signature scores in (Q) unwounded and (R) D7PW human tissue sections. Red box highlights wound edge region zoomed-in in the right panel. Red arrowheads denote the wound front; white dashed lines indicate the epidermal-dermal boundary.(S–T) Mean expression of (S) SEMA3C in keratinocytes and (T) NRP1 and NRP2 in endothelial cells at the human wound edge across healing timepoints (D0PW, D1PW, D7PW, D30PW). Each dot represents the mean signal per sequencing run. Statistical significance was assessed using a Wilcoxon rank-sum test relative to D0/UW (* p < 0.05; ** p < 0.01). n= 4.
Supplier Page from Abcam for Recombinant Human Semaphorin 3c protein