Arecoline induces the release of exosomes from epithelial cells
Exosomes, which are mediators between cell-cell communication by the transfer of their cargos and involved in extracellular matrix homeostasis.28 Thus, we predicted that arecoline (20µg/ml, 48h) can induce the release of exosomes from epithelial cells and targets the functional phenotype of fibroblasts. Exosomes from PBS and arecoline treated epithelial cells medium were isolated using differential centrifugations, then characterized, quantified by TEM, NTA and immunoblotting (Fig. S1A-C). Under the TEM, nanovesicles were seen as a round shape with bilayered membranes, which consistent with the characteristic shape and size range of exosomes. NTA revealed that the major vesicle sizes secreted by PBS and arecoline treated epithelial cells were 139 nm and 140 nm, respectively. By immunoblotting of lysates from purified nanovesicles, the exosomal markers, including CD9, CD63, CD81 (tetraspanins), TSG101 (ESCRT-I complex subunit), and Hsp70 was highly expressed in exosomes isolated from PBS and arecoline treated epithelial cells. Calnexin (endoplasmic reticulum membrane) and GAPDH were nearly no expressed in exosomes. Moreover, treatment of epithelial cells with GW4869, an inhibitor of exosome biosynthesis, can reduced arecoline induced exosomes release from epithelial cells (Fig. S1D). These results indicated that these nanovesicles isolated from PBS and arecoline treated epithelial cells presented typical characteristics of exosomes.
Epithelial cell-derived exosomes are responsible for the myofibroblast differentiation phenotype
Then, we hypothesized that epithelial cell-derived exosomes induced by arecoline might participate in fibroblast differentiation. In order to visualize exosomes transfer, we first incubated fibroblasts with epithelial cell-derived exosomes labeled by DiI stain for 24h and evaluated the exosomes uptake levels by measuring the red signal in the fibroblasts (Fig. 1A). The result shows that a robust exosomes signal in the cytoplasm of fibroblasts after incubation of DiI-labeled exosomes, suggesting that exosomes was successfully taken up by fibroblasts. The levels of collagen type I and α-SMA were higher in fibroblasts exposed to exosomes derived from arecoline treated epithelial cells (Fig. 1B, 1C) and dose-dependent with exosomes concentration (Fig. 1D, 1E). Meanwhile, we verified that the protein content of collagen type I and α-SMA was equivalent in PBS-, arecoline- and GW4869- exosomes of epithelial cell, excluding the regulation of protein level in exosomes (Fig. S1E). In addition, we collected the culture supernatants of fibroblasts and found that the total collagen content was significantly increased in fibroblasts group by incubated with arecoline treated epithelial cells exosomes (Fig. 1F). The migration and contraction abilities of fibroblasts were also significantly upregulated by treated with epithelial cell-derived exosomes (Fig. 1G, 1H). Consistent with the above results, GW4869 inhibited the expression of collagen type I and α-SMA, collagen secretion, contraction and migration abilities of fibroblasts after treated with epithelial cells derived exosomes (Fig. 1A-H). These data suggested that exosomes secreted from arecoline treated epithelial cells could modulating the phenotypic transformation of fibroblasts.
miR-17-5p is enriched in OSF epithelial cells-derived exosomes
Next, we compared the differences miRNAs expressed in PBS and arecoline (20µg/ml, 48h) treated epithelial cells exosomes by RNA-seq. We further analysed the expression of top 8 up-regulated miRNAs in both two groups (Fig. 2A, 2B). By qRT-PCR, we found that miR-17-5p was the most differentially expressed miRNA in exosomes (Fig. 2C). Moreover, the expression of miR-17-5p was up-regulated in epithelial cells (Fig. 2D) and exosomes derived from corresponding supernatant in a arecoline dose-dependent manner (Fig. 2E). Interestingly, we found that treatment with GW4869 inhibitor resulted in the accumulation of miR-17-5p in epithelial cells (Fig. 2F). On the other hand, incubation with arecoline derived exosomes similarly upregulated 17-5p expression in fibroblasts, which was blocked by GW4869 (Fig. 2G). To find out whether miR-17-5p associated with progression of fibrosis, miR-17-5p expression levels were analysed by OSF tissues (Fig. 2H). Compared with the normal tissues, the expression of miR-17-5p was higher in fibrosis tissues on average than in normal tissues.
We then assessed the miR-17-5p levels in a mouse model of OSF. After 6 weeks of exposure to Bleomycin (Fig. S2A), the mice developed a connective tissue remodeling phenotype. Scanning electron microscopy (SEM) showing that, compared with PBS group, the structural of collagen fiber accumulation, dense, and arranged disorderly in Bleomycin group (Fig. S2B). HE, Masson and Sirus red staining confirm that collagen deposition, thickened and blood vessels were reduced in oral mucosa lamina propria (Fig. S2C, 2D). The hydroxyproline experiment showed that the collagen content in the local tissues of the oral mucosa was significantly increased (Fig. S2E). Immunoblotting revealed that the protein levels of collagen type I and α-SMA (myofibroblasts marker) were upregulated in the mucosa of mice exposed to Bleomycin (Fig. S2K). These results indicated that Bleomycin successfully induced OSF in mice, which was consistent with previous research. qRT-PCR analysis demonstrated that, in Bleomycin treated oral mucosa and blood-derived exosomes samples, miR-17-5p was up-regulated relative to the control. Interestingly, these results were also validated in a mouse skin fibrosis model (Fig. S2). These results indicated that up-regulation of miR-17-5p may be involved in progression of fibrosis.
Exosomal miR-17-5p is required for the myofibroblast differentiation phenotype
To determine whether miR-17-5p is involved in the dysfunctional cross-talk of the epithelial cells and fibroblasts, a co-culture cell model was developed. Briefly, we plated epithelial cells transfected with the Cy3-miR-17-5p mimic in the upper chamber and fibroblasts in the lower chamber (Fig. 3A). The co-culture system was separated by 0.4 µm pore, which allowing the transmission of microparticles, such as exosomes, but inhibiting direct contact between cells. After 36 h, we found strong red fluorescence in fibroblasts and low red fluorescence co-culture with GW4869 (Fig. 3B). This phenomenon proved that miR-17-5p might be transferred from epithelium to fibroblasts through exosomes. When epithelial cells transfected with miR-17-5p (mimic) or anti-miR-17-5p (inhibitor) and fibroblasts were placed in co-culture, miR-17-5p levels were increased in epithelium and fibroblasts transfected with miR-17-5p but decreased by transfection with anti-miR-17-5p (Fig. 3C, 3D). Similarly, the levels of collagen I and α-SMA were upregulated in fibroblasts co-cultured with miR-17-5p transfected epithelial cells or downregulated with anti-miR-17-5p transfected group (Fig. 3E, 3F). Then, we investigated the role of miR-17-5p derived from epithelium in the myofibroblast differentiation phenotype. The levels of collagen type I and α-SMA were higher in fibroblasts transfected miR-17-5p (Fig. 3G, 3H). The collagen secretion (Fig. 3I), contraction (Fig. 3J) and migration abilities (Fig. 3K) of fibroblasts were up-regulated or down-regulated after co-cultured with epithelium transfected miR-17-5p or anti-miR-17-5p. This phenomenon proved that miR-17-5p from epithelium exosomes might participate in epithelium-fibroblasts communication and inducing phenotype transformation of fibroblast.
Exosome-mediated transfer of miR-17-5p targets Smad7 directly promote the myofibroblast differentiation phenotype
To further explore the mechanism through which exosomes and miR-17-5p induced myofibroblast differentiation, we investigated the target gene involved in mediating the effect of miR-17-5p by TargetScan (www.targetscan.org), MiRDB (www.mirdb.org) and MiRBase (www.mirbase.org). We found that Smad7 was a target of miR-17-5p in 3′-UTR area (Fig. 4A). Luciferase reporter assays were conducted by use of plasmids containing wild or mutant Smad7 3’UTR with miR-17-5p binding sites. Following co-transfection of 293T with a plasmid containing wild or mutant Smad7 3’UTR and miR-17-5p control (NC) or mimic. There was a significant reduction in luciferase activity when miR-17-5p and wild Smad7 3’UTR was co-transfection in 293T cells. However, as it did not affect the luciferase activity when the binding site was mutated (Fig. 4B). Consistently, exosomes derived from arecoline treated epithelial cells also decreased the luciferase activity of Smad7 3’UTR, but not that for PBS and exosomes inhibition group (Fig. 4C). Then, we co-culturing miR-17-5p transfected epithelial cells with fibroblasts and finding that co-localization of miR-17-5p and target Smad7 was observed in fibroblasts (Fig. 4D). Remarkably, the expression level of Smad7 was down-regulated and inversely correlated with miR-17-5p expression in tissues of clinical patient (Fig. 4E, 4F) and OSF model mice (Fig. 4H, 4I). Furthermore, after fibroblasts transfected with miR-17-5p for 48h demonstrated a decrease of Smad7 levels, whereas inhibition of miR-17-5p reversed this process and an increase of TGFBR1 (TGF-β Receptor I) and p-Smad2 levels were evident (Fig. 4G). In addition, we observed that miR-17-5p resulted in increased levels of fibronectin, collagen I and α-SMA expression in fibroblasts.
To determine whether exosomal miR-17-5p derived from arecoline treated epithelium promotes fibroblasts differentiation through regulation of Smad7, control, anti-miR-17-5p or Smad7 siRNA-transduced fibroblasts were exposed to exosomes derived from arecoline treated epithelium. The phenotype of fibroblasts and protein levels of α-SMA and collagen type I were determined. The results showed that downregulation of miR-17-5p blocked the changes of fibrosis markers expression (Fig. 5A, 5B) and prevented the collagen contraction (Fig. 5C), secretion (Fig. 5D) and migration abilities (Fig. 5E) of fibroblasts in fibroblasts caused by exosomes derived from arecoline treated epithelial cells, but knockdown of Smad7 attenuated the effect of the miR-17-5p inhibitor.
miR-17-5p promotes ubiquitin-mediated Smad7 protein degradation
Smad7 has been identified as a negative regulator that ameliorates TGF-β mediated fibrogenesis. However, it is not clear whether miRNA interfere with the function of Smad7-mediated TGF-β signaling. As a result, we transfected miR-17-5p into HEK293T and fibroblasts for 24 h, then the cells were treated with 100 µg/mL cycloheximide (CHX, a protein synthesis inhibitor). The cell lysates were then collected within a specified time period and analysed by Immunoblotting. Compared to control group, miR-17-5p overexpression evidently promoted Smad7 protein degradation, thus shortening its half-life (Fig. 6A, 6B). We also treated the cells with 10 µm MG132 (a specific inhibitor of a ubiquitin-binding protein) to explore the effect of miR-17-5p on Smad7 protein degradation. The reduction of Smad7 protein by miR-17-5p overexpression was restored by MG132 (Fig. 6C). The levels of ubiquitinated Smad7 protein were increased when miR-17-5p was overexpressed in the presence of ubiquitin (Fig. 6D). These results indicated that miR-17-5p regulates Smad7 protein stability via enhancing its ubiquitination dependent degradation.
Numerous studies have confirmed that the mechanism of samd7 inhibits TGF-β signaling is by recruiting E3 ligase to TGFBR1, leading to ubiquitination and degradation of the receptor.29 Then, we speculated that miR-17-5p may destabilize Smad7 protein by ubiquitination system to promote TGFBR1 expression. To this end, we then transfected HEK293T cells with Smad7 and monitored the half-life of TGFBR1 protein after CHX treatment. Compared to control group, the protein level of TGFBR1 in the Smad7 transfected group decreased at each time point with CHX treatment (Fig. 6E). In order to further evaluate the effect of miR-17-5p on the expression of Smad7 and TGFBR1, we overexpressed miR-17-5p or Smad7 in HEK293T and fibroblasts to detect the expression of protein levels. Compared with the control group, miR-17-5p reduced the expression of Smad7 and increased the expression level of TGFBR1, whereas overexpression of Smad7 showed the opposite result (Fig. 6F). In the end, we performed Co-IP assay to further explore whether miR-17-5p acts as a scaffold to enhance the binding of Smad7 with TGFBR1 and found that the association between Smad7 and TGFBR1 was enhanced in HEK293T and fibroblasts by miR-17-5p overexpression or inhibition (Fig. 6G).
Upregulation of miR-17-5p promotes OSF progress in mice
To confirm the role of miR-17-5p in Bleomycin induced OSF in mice, miR-17-5p agomir was local injected in the oral mucosa, or SB525334 (a TGF-β inbibitor) was intraperitoneal injected simultaneous with the Bleomycin exposure (Fig. 7A). The results of histopathological section staining (Fig. 7B) and hydroxyproline (Fig. 7C) assay showed that mice receiving miR-17-5p agomir aggravated the accumulation and volume of tissue collagen. However, SB525334 could effectively reverse the amplification effect of fibrosis induced by miR-17-5p. Immunoblotting showed that, in miR-17-5p agomir mice, up-regulated of miR-17-5p induced the expression of fibronectin, collagen type I and α-SMA expression, and activated TGF-β pathway (Fig. 7D). Similarly, there were higher expressions of exosomal miR-17-5p in the sera of the miR-17-5p agomir mice compared with the other groups (Fig. 7E). Moreover, the expression of smad7 was decreased in mice treated with miR-17-5p agomir, whereas SB525334 could reverse its expression level. The above experimental results were also validated in a mouse skin fibrosis model (Fig. S3). These data provided support for our hypothesis that miR-17-5p can promote the development of fibrosis through the TGF-β pathway in vitro and in vivo.