Globotriaosylceramide Gb3 Influences Wound Healing and Scar Formation by Orchestrating Fibroblast Heterogeneity

Abstract

Cutaneous fibroblast heterogeneity is mechanistically linked to wound repair outcomes and fibrotic progression, with glycosphingolipid metabolism emerging as a critical determinant of physiological fibroblast diversity. Through integrative analysis of spatiotemporal omics, lipidomics, and single-cell RNA sequencing (scRNA-seq) coupled with histological evaluation of clinical specimens, the functional involvement of globotriaosylceramide (Gb3) in dermal regeneration processes is systematically investigated. Comparative profiling reveals significant upregulation of Gb3 biosynthesis in superficial second-degree burns (SSDB) relative to deep second-degree burn (DSDB) injuries. Hexosaminidase subunit beta (HEXB) is identified as the exclusive differentially expressed Gb3 synthase distinguishing these injury subtypes. Functional validation through in vitro and in vivo models demonstrates that pharmacological suppression of HEXB-mediated Gb3 synthesis exacerbates fibroblast-to-myofibroblast transdifferentiation, attenuated fibroblast growth factor 2 (FGF2) signal transduction, and ultimately potentiated fibrotic scarring. These findings establish a novel HEXB-Gb3-FGF2 regulatory axis governing fibroblast phenotypic plasticity in differential-depth skin injuries, providing mechanistic insights for developing targeted antifibrotic therapies.

Keywords: fibroblast heterogeneity; globotriaosylceramide; hexosaminidase subunit beta; scar formation; wound healing.

Figure 1

Combination of ST and metabolomics analysis revealed differentially expressed Gb3 between DSDB and SSDB. A) Study flow chart of Gb3 for scarless burn wound healing. Created in BioRender. B,C) Spatiotemporal Transcriptomic Analysis of Burn Wound Samples. Left panel: Hematoxylin and eosin (H&E) staining. Middle panel: Enrichment map for Sphingolipid Metabolism pathway. Right panel: Enrichment map for Myofibroblast Differentiation pathway. For each enrichment map (middle and right), the main panel displays raw expression values, while the inset in the lower‐right corner shows normalized expression values. B) Mixed superficial and deep second‐degree burn sample at postburn day 19. The region highlighted by the black box in the lower‐left corner of the H&E image represents the deep second‐degree burn (DSDB) area. Within both the Sphingolipid Metabolism and Myofibroblast Differentiation enrichment maps, the DSDB region exhibits a more yellowish hue, indicating higher expression levels of pathway‐associated genes compared to surrounding regions. C) Mixed superficial and deep second‐degree burn sample at postburn month 6. The region highlighted by the blue box on the left side of the H&E image represents the DSDB scar area. In both the Sphingolipid Metabolism and Myofibroblast Differentiation enrichment maps, the DSDB scar region shows a more yellowish coloration, signifying elevated expression levels of pathway‐associated genes relative to adjacent areas. D,E) Lipid metabolism suggested that Gb3 in lipids was significantly increased in SSDB (n = 3). The data are presented as the means ± SDs. **p < 0.01. F) Immunofluorescence suggested that the expression of Gb3 was higher in SSDB than that in DSDB.

Figure 2

Gb differentially located in dermal fibroblasts and mainly distributed in papillary instead of reticular fibroblasts. A) Spatial metabolomics analysis of normal skin displayed that Gb predominantly lay in small, spindle‐shaped papillary fibroblasts rather than large, stellate reticular fibroblasts, with spatial statistical significance. According to cytotoxin staining in a previous study, it was deciphered that Gb enriched in papillary fibroblasts. B) The scRNA‐seq analysis of burned and normal skin samples identified 9 cell clusters and the Cleveland plot showed cell types distribution in distinct samples as well as the accuracy of cell annotation with four canonical markers (PTPRC, VIM, KRT1, and PECAM1). C) GSVA quantitively identified that the activity of Gb synthesis metabolic pathway was significantly stronger in SSDB than that in DSDB, while other types of GSL born no significance. The data are presented as the means ± SDs. **p < 0.01, ***p < 0.001. D) Stereo‐seq analysis revealed predominant enrichment of the Gb metabolic pathway within the papillary dermis. The region superficial to the yellow dotted line denotes the epidermis, while the zone bounded by the yellow and brown dotted lines corresponds to the papillary dermis. The compartment deep to the brown dotted line represents the reticular dermis. Color intensity within each anatomical region indicates pathway expression levels. The lighter the color, the higher the expression level. E) Immunofluorescence verified that Gb mainly existed in papillary fibroblasts in both normal and burned skin samples.

Figure 3

HEXB differentially expressed in dermal of DSDB and SSDB with temporal variation. A) scRNA‐seq of fibroblasts in burned skin samples identified 11 subtypes. B) Intersection of GbRGs and DEGs of fibroblasts in DSDB and SSDB showed 6 significant GbRGs (HEXB, A4GALT, ST3GAL1, GLA, GBGT1, ST3GALT2). C,D) Cellular feature plots and violin plots of the 6 GbRGs indicated that HEXB had the most significant difference between DSDB and SSDB. E) Stereo‐seq elucidated that HEXB was mainly expressed in papillary dermis and varied with the development of wound healing. The region superior to the yellow dashed line denotes the epidermal layer, while the area delimited between the yellow dashed line and brown dashed line corresponds to the papillary dermis. The compartment inferior to the brown dashed line represents the reticular dermis. Color intensity within each anatomical region indicates HEXB expression levels. The lighter the color, the higher the expression level. F) Immunofluorescence revealed temporal variation of HEXB expression. G) Immunofluorescence found that Gb3 (green) and HEXB (red) costained cells (orange) existed in papillary dermis of burned skin samples.

Figure 4

Alteration of HEXB resulted in codirectional alterations of Gb3 in fibroblasts affecting cell heterogeneity. A) Verification of HEXB overexpression at the mRNA and protein level in dHFBs transfected with HEXB^OE, as detected by qRT‐PCR and western blot (n = 3). B) Representative immunofluorescence staining of Gb3 (red) in HEXB^OE and NC^OE. Quantitative analysis of Gb3 fluorescence intensity in the experiments was shown on the right (n = 3). C) Representative immunofluorescence staining of FAP (red) and CD90 (green) in HEXB^OE and NC^OE cells. Quantitative analysis of FAP+CD90‐ proportion in the experiments was shown on the right (n = 3). D) Western blot and quantification of HEXB^OE and NC^OE cells. Data were normalized against GAPDH (n = 3). E) Barplots of qRT‐PCR quantifying the mRNA levels of papillary fibroblast and reticular fibroblast related genes in HEXB^OE and NC^OE cells. Data were shown as log₂Fold Change (FC) over NC^OE cells (n = 3). F) CCK‐8 proliferation assay was used to demonstrate the proliferation of HEXB^OE cells (n = 3). G) Scratch test showed the migration of HEXB^OE cells (n = 3). p‐values were differences compared to: NC^OE. H) Verification of HEXB knock‐down at the mRNA and protein level in dHFBs transfected with shHEXB, as detected by qRT‐PCR and western blot (n = 3). I) Representative immunofluorescence staining of Gb3 (red) in shHEXB and shNC (scale, 100 µm). Quantitative analysis of Gb3 fluorescence intensity in the experiments was shown on the right (n = 3). J) Representative immunofluorescence staining of FAP (red) and CD90 (green) in shHEXB and shNC. Quantitative analysis of FAP‐CD90+ proportion in the experiments was shown on the right (n = 3). K) Western blot and quantification of shHEXB cell and shNC cell. Data were normalized against GAPDH (n = 3). L) Barplots of qRT‐PCR quantifying the mRNA levels of papillary fibroblast and reticular fibroblast related genes in shHEXB cells and shNC cells. Data were shown as log₂FC over shNC cells (n = 3). M) CCK‐8 proliferation assay was used to demonstrate the proliferation of shHEXB cells (n = 3). N) Scratch test showed the migration of shHEXB cells (n = 3). p‐values were differences compared to: shNC. The data are presented as the means ± SDs. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

Figure 5

Evaluating the efficacy of Gb3 in promoting wound healing and preventing scar in vivo. A) Screening optimal drug treatment concentration with the qRT‐PCR. 48‐h cellular culture with 5 µm HEXB inhibitor was selected for subsequent experiments (n = 3). B) Screening optimal drug treatment concentration with the qRT‐PCR. 48‐h cellular culture with 10 µm Gb3 inhibitor was selected for subsequent experiments (n = 3). C) An experimental scheme of burn wound model. D) Visual representations of burn wounds. A yellow dashed line delineates the scar area. E) A diagram of appearance healing. F) Wound closure progression (n = 6). G) Immunofluorescence images demonstrate the effects of Gb3 and HEXB inhibitor on HEXB expression at the wound site. White arrows denote positive signals. H–J) Representative images of dermis thickness, Masson's trichrome staining, and Sirius red staining (Collagen I marked by red or orange; Collagen III marked by green) on day 42 postburn. K–M) Quantitation of dermis thickness, Masson's trichrome staining, and Sirius red staining among each group (n = 6). Statistical analysis was conducted using one‐way ANOVA. The data are presented as the means ± SDs. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

Figure 6

Correlation Analysis of Gb3 Synthesis and FGF2 Signaling Pathway. A) Correlation analysis scatter plots indicated that the FGF2‐FGFR signaling pathway was significantly activated in fibroblasts with active Gb biosynthesis, and the two were significantly positively correlated. B) Verification of FGF2 expression at the mRNA and protein level in dHFBs transfected with HEXB^OE and shHEXB, as detected by qRT‐PCR and western blot (n = 3). The data are presented as the means ± SDs. **p < 0.01, ****p < 0.0001 versus NC^OE, ^## p < 0.01, ^#### p < 0.0001 versus shNC. C) Barplots of qRT‐PCR quantifying the mRNA levels of FGF2 signal pathway genes in HEXB^OE cells and shHEXB cells. Data were shown as log₂FC over untreated cells (n = 3). The data are presented as the means ± SDs. **p < 0.01, ****p < 0.0001 versus NC^OE, ^## p < 0.01, ^#### p < 0.0001 versus shNC. D) After 24 h of serum‐free starvation culture, shHEXB, and shNC cells were treated with FGF2 (MCE, HY‐P7330, 10 ng mL⁻¹) for 24 h. The expression of FGF2 target gene mRNA was detected by qRT‐PCR. The data are presented as the means ± SDs. ***p < 0.01, ****p < 0.0001 versus shHEXB, ^### p < 0.01, ^#### p < 0.0001 versus shNC. E,F) Western blot and quantitative real‐time PCR of cells treated as in (D). Data were normalized against GAPDH (n = 3). G) Visual representations of burn wounds. A yellow dashed line delineates the scar area. H) Wound closure progression. (n = 6) I) Representative images of dermis thickness, Masson's trichrome staining, and Sirius red staining (Collagen I marked by Red; Collagen III marked by green) on day 42. J) Quantitation of dermis thickness, Masson's trichrome staining, and Sirius red staining among each group n = 6. Statistical analysis was conducted using one‐way ANOVA. The data are presented as the means ± SDs. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

Figure 7

Recombinant FGF2 inhibits expression of pro‐fibrotic genes in dHFBs in an FGFR1‐dependent manner. A) mRNA from untreated early passage dHFBs was analyzed via qRT‐PCR for FGFR1, FGFR2, FGFR3, and FGFR4. ΔCt values were normalized to 18S. B) After 24 h of serum‐free starvation culture, dHFBs were treated with Gb3 (10 µm) for 48 h. The expression of FGFR1 target gene mRNA was detected by qRT‐PCR. C) Representative immunofluorescence staining in dHFBs showing colocalization of Gb3 (red) and FGFR1 (green). D) Interaction between Gb3 and FGFR1. Residues of Gb3 interacting with FGFR1 were labeled and shown in blue stick model. Hydrogen bond is shown with green dash. E) After 24 h of serum‐free starvation culture, dHFBs were treated with FGFR1‐specific tyrosine kinase inhibitor PD173074 (MCE, HY‐P7330, 1 nm) for 48 h. The expression of FGFR1 target gene mRNA was detected by qRT‐PCR. F) Barplots of qRT‐PCR quantifying the mRNA levels of FGF2 signal pathway genes in control, Gb3, control+FGFR1‐inhibitor, and Gb3+FGFR1‐inhibitor cells (n = 3). G, H) Barplots of qRT‐PCR quantifying the mRNA levels of papillary fibroblast and reticular fibroblast related genes in control, Gb3, control+FGFR1‐inhibitor, and Gb3+FGFR1‐inhibitor cells (n = 3). I) Western blot of cells treated as in (F). Data were normalized against GAPDH (n = 3). The data are presented as the means ± SDs. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.