En-1 expression is up regulated in mouse skin wound dermis and TGF-β1 activated mDFs
Using IHC staining, we demonstrated that the expression of EN-1 was significantly higher in mouse skin wound dermis than that in the normal skin (Fig. 1A). Since TGF-β1 is a key factor in fibroblasts activation in fibrotic diseases, leading to fibroblasts differentiation into myofibroblasts(Melchionna et al. 2021), we used the recombinant adenovirus Ad-TGF-β to infect mDFs with high efficiency (Fig. 1B, panel a), and demonstrated that TGF-β1 stimulated mDFs exhibited a high level of En-1 expression at 48h, compared with that treated with Ad-RFP as assessed by TqPCR (Fig. 1B, panel b). Thus, these results revealed that EN-1 was highly expressed in mouse skin wound and was upregulated by TGF-β1 in dermal fibroblasts.
Blockade of the YAP/Engrailed-1 pathway or silencing EN-1 expression suppresses the activation of mDFs through remodeling the ECM ultrastructure, and altering the major protein components and HA production of ECM
Even though the inhibitors of YAP-TEAD pathway such as Verteporfin were reported to block the activation of EN-1 in fibroblasts and inhibit scar formation in animal models(Jiang and Rinkevich 2021), their efficiency and underlying mechanisms remain to be fully understood. In this study, we first determined the effective concentrations of Verteporfin in mDFs by using WST-1 assay, and selected 0.5 µmol as the optimal concentration for further experiments (Fig. S1B, panels a & b).
Unlike wound repair in adult epithelia, no scarring is formed in fetal tissue that is characterized by rapid re-epithelialization, lack of inflammation, and restoration of normal tissue architecture(Ho et al. 2014). TqPCR analysis showed that Verteporfin significantly inhibited the expression of scar formation related genes in TGFβ1-activated mDFs, including Profilin1, Prdx1, Ltf, Lgals1 and Calr (Fig. 2A). During skin wound healing, fibroblasts transform into myofibroblasts and synthesize and secrete large amounts of ECM proteins, which accumulate in excessive amounts and often in aberrant and forms of ECM proteins, leading to distorted and non-functional aggregation of scar tissue. We found that F-actin was significantly down regulated after Verteporfin treated in mDFs by Phalloidin staining for cytoskeleton (Fig. 2B).
Through SEM ultrastructure analysis, we further found that the reticular fibers were significantly reduced and loosely organized after Verteporfin treatment in mDFs, compared with that in the DMSO control group (Fig. 2C, panel a). IF staining revealed that the expression of FN-1, FIB and TNC proteins was significantly downregulated in ECM of mDFs after Verteporfin treatment (Fig. 2C panel b) while no significant changes in expression levels were found for COL1A2, COL3A1, SPARC, and TSP, compared with the treatment group and the negative control (NC) as shown in Fig. S1C. HA is an ECM glycosaminoglycan (GAG) that is widely distributed throughout the body and plays important physiological functions in the body, such as regulating the permeability of blood vessel walls and the diffusion of hydrolysates, and promoting wound healing(Amorim et al. 2021). Using Alcian blue staining, we found that Verteporfin promoted HA production in the ECM (Fig. 2D, panel a). Furthermore, TqPCR analysis showed that HA catabolic related genes Hyal1, Hyal2, Hyal3, Ph-20 and Kiaa were significantly downregulated by Verteporfin treatment at 72h (Fig. 2D, panel b). IF staining further confirmed that Verteporfin reduced the expression of HYAL2 (Fig. 2D, panel c).
To specifically silence the expression of En-1 expression, we constructed a recombinant adenovirus to expressing siRNAs that target En-1 (namely Ad-simEn1). We showed that the Ad-simEn1 effectively transduced mDFs and knocked down En-1 expression (Fig. S1E, panels a & b). TqPCR analysis showed that Ad-simEn1-mediated silencing of En-1 expression in the Ad-TGF-β1 transduced mDFs significantly inhibited the expression of scar formation related genes, including Profilin1, Prdx1, Lgals1 and Calr (Fig. 3A). Furthermore, we found that F-actin was significantly down regulated after Ad-simEn1 treated in mDFs by Phalloidin staining (Fig. 3B).
Through SEM analysis, we further observed that the reticular fibers of the ECM derived from the Ad-simEn1-transduced TGFβ1-stimulated mDFs were significantly reduced and loosely organized (Fig. 3C, panel a). Furthermore, IF staining analysis revealed that the expression of FN-1, COL1A2, COL3A1 and FIB proteins was significantly downregulated in ECM of mDFs transduced by Ad-simEn1(Fig. 3C panel b), while no changes in the expression levels were found in SPARC, TSP, TNC (Fig. S1F). Alcian blue staining of the ECM showed that Ad-simEn1 promoted HA production (Fig. 3D, panel a). TqPCR analysis revealed that HA catabolic related genes Cd44, Hyal1, Hyal2, Hyal3, Ph-20 and Kiaa were significantly downregulated in the mDFs transduced with Ad-simEn1 at 72h (Fig. 3D, panel b). IF staining showed that silencing En-1 expression in mDFs by Ad-simEn1 reduced the expression of CD44 and HYAL2 (Fig. 3D, panel c). Taken together, these findings strongly suggest that inhibition of YAP/EN-1 signaling pathway or silencing EN-1 expression effectively suppresses the activation of mDFs in vitro, through inhibiting the expression of scar formation genes, altering protein components that shape ECM ultrastructure, and slowing down HA catabolism.
Inhibition of EN-1 functions suppresses keloid scar formation in the rabbit ear skin wound model through ECM remodeling
Since rodents usually do not undergo keloid scar healing of skin wound, we chose the commonly-used rabbit ear skin wound healing model(Dos Santos et al. 2015). Briefly, full-thickness wounds were created on ventral side of the ears in adult rabbits and divided into four groups, followed by treatments with Verteporfin or DMSO. Gross images of the healing wounds were taken at multiple time points, up to 2 weeks. Although there was no significant difference in wound healing rate among 2 groups, the surfaces of the healing sites in Verteporfin groups were significantly flatter, while apparent protruding scars were observed on the skin surfaces in DMSO groups (Fig. 4A, panels a & b), suggesting that inhibition of EN-1 functions may effectively suppress keloid scar formation in vivo. Both H & E and Masson’s Trichrome staining analyses revealed that much fewer dermal fibroblasts and lower collagen deposition above the cartilage layer in the Verteporfin group than that in the DMSO control group (Fig. 4B, panel a). Furthermore, we analyzed the low magnification images of multiple samples in each group and calculated the SEI values. We found that the SEI value in the Verteporfin group was within 2 to 3 and significantly lower than that of the control group (Fig. 4B, panel b), indicating that the scar formation can be effectively inhibited by Verteporfin at the skin wound sites. IHC staining analysis showed that EN-1, FN-1, FIB and COL3A1 were significantly down regulated when EN-1 functions were inhibited by Verteporfin (Fig. 4C).
We further build same rabbit ear skin wound healing model, followed by treatments with Ad-simEn1or Ad-RFP. Gross images of the healing wounds were taken at multiple time points, up to two weeks. The surfaces of the healing sites in Ad-simEn1 group were significantly flatter, while apparent protruding scars were observed on the skin surfaces in Ad-RFP group (Fig. 5A, panels a & b). Both H & E and Masson’s Trichrome staining analyses revealed that much fewer dermal fibroblasts and lower collagen deposition above the cartilage layer in the Ad-simEn1 group than that in the Ad-RFP control group (Fig. 5B, panel a). We also found that the SEI value in the Ad-simEn1 groups was almost 2 and significantly lower than that of the control group (Fig. 5B, panel b), indicating that the scar formation can be effectively inhibited by Ad-simEn1-mediated silencing of En-1 expression at the skin wound sites. IHC staining analysis showed that EN-1, FN-1, FIB, COL1A2 and COL3A1 were significantly down regulated when EN-1 functions were inhibited by Ad-simEn1-mediated silencing (Fig. 5C). Collectively, these results demonstrate that silencing EN-1 suppresses scar formation in rabbit hypertrophic scarring model through altering ECM major components, including FN-1, FIB, COL1A2 and COL3A1.