Fibromodulin Is Essential for Fetal-Type Scarless Cutaneous Wound Healing

Abstract

In contrast to adult and late-gestation fetal skin wounds, which heal with scar, early-gestation fetal skin wounds display a remarkable capacity to heal scarlessly. Although the underlying mechanism of this transition from fetal-type scarless healing to adult-type healing with scar has been actively investigated for decades, in utero restoration of scarless healing in late-gestation fetal wounds has not been reported. In this study, using loss- and gain-of-function rodent fetal wound models, we identified that fibromodulin (Fm) is essential for fetal-type scarless wound healing. In particular, we found that loss of Fm can eliminate the ability of early-gestation fetal rodents to heal without scar. Meanwhile, administration of fibromodulin protein (FM) alone was capable of restoring scarless healing in late-gestation rat fetal wounds, which naturally heal with scar, as characterized by dermal appendage restoration and organized collagen architectures that were virtually indistinguishable from those in age-matched unwounded skin. High Fm levels correlated with decreased transforming growth factor (TGF)-β1 expression and scarless repair, while low Fm levels correlated with increased TGF-β1 expression and scar formation. This study represents the first successful in utero attempt to induce scarless repair in late-gestation fetal wounds by using a single protein, Fm, and highlights the crucial role that the FM-TGF-β1 nexus plays in fetal-type scarless skin repair.

Supplemental Figure S1

Fibromodulin (Fm) expression decreases in unwounded fetal rat skin during embryonic day 16.5 (E16)-to-embryonic day 18.5 (E18) transition. A: Scatterplot of combined corresponding dye-sway microarray experiments. Diagonal lines define the areas of twofold differential signal intensities (P < 0.01). B: Expression of small leucine-rich proteoglycan fibromodulin (fmod; orange), biglycan (bgn; white), decorin (dcn; gray), and lumican (lum; yellow) is also confirmed by real-time quantitative RT-PCR. Data are normalized to E16 rat skin samples. Data are expressed as means ± SD. n = 4. ^∗P < 0.05, ^∗∗P < 0.01.

Figure 1

Fibromodulin (Fm) plays a crucial role in rat fetal-type scarless wound healing. A: At 72 hours after injury, phosphate-buffered saline (PBS)- and IgG-treated embryonic day 16.5 (E16) wounds heal scarlessly and are indistinguishable from age-matched unwounded embryonic day 19.5 (E19) skin. However, anti-FM antibody–treated E16 wounds heal with scar formation. B and C: Hair follicle densities are analyzed based on hematoxylin and eosin (H&E) staining (B), while fractal dimension (Fd) and lacunarity (L) are analyzed based on Picrosirius red–confocal laser scanning microscopy (PSR-CLSM) imagery (C). D–F: PBS- and Vitrogen-treated embryonic day 18.5 (E18) wounds heal with scarring at 72 hours after injury. Vitrogen + Fm–treated E18 wounds heal without scar, and the regenerated tissue resembles unwounded embryonic day 21.5 (E21) skin (D), accompanied by hair follicle densities analysis (E) and Fd and L analyses (F). Blue arrowhead, hair follicle; black arrowhead, surgical dye. Data are expressed as means ± SD. n = 8 (A, anti-Fm antibody); n = 9 (A, PBS; D–F, Vitrogen + Fm); n = 10 (C–F, PBS); n = 13 (A, IgG); n = 15 (C–F, Vitrogen). ^∗P < 0.05 (two-sample t-test). Scale bar = 100 μm.

Figure 2

Fibromodulin (Fm) deficiency leads to scar formation in embryonic day 16.5 (E16) fetal mouse wounds. A: Unwounded embryonic day 19.5 (E19) wild-type (WT) and Fmod^−/− mouse skin shows no significant histologic difference. B: At 72 hours after injury, E16 fetal mouse wounds shows scarless repair, while E16 Fmod^−/− fetal mouse wounds present scar formation characterized by the absence of hair follicles and epidermal hypertrophy. C: Administration of FM with Vitrogen partially restores scarless wound healing in E16 Fmod^−/− wounds, while Vitrogen control alone fails to do so. Hair follicle densities are analyzed based on hematoxylin and eosin (H&E) staining (D), while fractal dimension (Fd) and lacunarity (L) are analyzed based on Picrosirius red–confocal laser scanning microscopy (PSR-CLSM) imagery (E). Blue arrowhead, hair follicle; black arrowhead, surgical dye. Data are expressed as means ± SD. n = 6. ^∗P < 0.05 (two-sample t-test). Scale bars: 50 μm (black); 25 μm (white).

Figure 3

Fibromodulin (Fm) reduces transforming growth factor (Tgf)-β1 expression in embryonic day 18.5 (E18) fetal rat wounds. A: Compared with Vitrogen alone, Vitrogen + Fm significantly reduces Tgf-β1 protein expression in E18 rat wounds at 72 hours after injury [inset depicts unwounded embryonic day 21.5 (E21) rat skin]. B and C: Quantification of protein expression (B) and real-time quantitative RT-PCR analyses (C) further reveal that Vitrogen control–treated E18 rat wounds approximately double transcription of Tgf-β1 compared with unwounded E21 rat skin. However, Vitrogen + FM application markedly reduces Tgf-β1 expression to the levels of age-matched unwounded E21 rat skin. A similar transcription pattern is also observed in Col1α1 (encoding α₁ chain of type I collagen) instead of Col3α₁. Real-time quantitative RT-PCR data are normalized to unwounded E21 rat skin. Data are expressed as means ± SD. n = 3 pools (3 fetal wounds per pool, 9 fetuses total; C and D); n = 9 fetuses (B). ^∗P < 0.05, ^∗∗P < 0.01 (two-sample t-test). Scale bar = 50 μm. IHC, immunohistochemistry.

Figure 4

Fibromodulin (Fm) deficiency leads to increased transforming growth factor (Tgf-β1) expression in embryonic day 16.5 (E16) fetal mouse wounds. A: Compared with E16 wild-type (WT) fetal mouse wounds, E16 Fmod^−/− fetal mouse wounds exhibit increased Tgf-β1 staining at 72 hours after injury. Furthermore, the increased Tgf-β1 expression in E16 Fmod^−/− fetal wounds are not weakened by Vitrogen collagen control but are decreased by Vitrogen + FM. B and C: These phenomena are confirmed by quantification of protein expression (B) and real-time quantitative RT-PCR analyses (C). D: A similar transcription pattern is also observed in Col1α1 (encoding α₁ chain of type I collagen) instead of Col3α₁. Blue arrowhead, hair follicle; and black arrowhead, surgical dye. Real-time quantitative RT-PCR data are normalized to unwounded embryonic day 19.5 (E19) mouse skin. Data are expressed as means ± SD. n = 3 pools (3 fetal wounds per pool, 9 fetuses total; C and D); n = 6 (B). ^∗P < 0.05, ^∗∗P < 0.01 (two-sample t-test). Scale bar = 50 μm.