To investigate whether TGF-β1 or -β3 treatment modulates the expression of genes involved in wound healing and fibrosis, we performed a RT² profiler PCR array analysis of the mRNA obtained from 4-week hCF 3D constructs after TGF-β1 or -β3 treatment. A total of 86 genes (Table A1—Appendix A) were measured, and their association with distinct pathways was mapped (Figure 1A). Genes with a fold change (<−1.0 or >1.0) between mRNA profiles from untreated 3D constructs vs. treated (TGF-β1 or -β3) were confirmed. Whilst the magnitude of change was dissimilar for many genes, the presented heatmap (Figure 1B) narrows that number and lists 9 genes—AGT, GREM1, ITGAV, ITGB1, JUN, MAPK14, MMP14, PDGFR, and SRC—that expressed a >1.0-fold increase in TGF-β1 treatment and <−1.0-fold decrease in TGF-β3 treatment that were significantly (p < 0.05) different compared to untreated, as well as between the two different TGF-β treatments. Furthermore, we extended our observations to include integrins ITGAV and ITGB1 (Figure 1), as well as ITGB3, ITGB5, and ITGB6 (Table A1—Appendix A), which were elevated with TGF-β1 treatment compared to TGF-β3. These genes are involved in the profibrotic, extracellular matrix, and cell adhesion pathways, as shown in Figure 1A. Also included in the heatmap is ACTA2, which, following both TGF-β1 and -β3 treatments, was significantly (p < 0.05) upregulated compared to untreated controls; however, TGF-β3 treatment did not augmented ACTA2 expression to the same levels as seen with TGF-β1 treatment. Interestingly, these mRNA profiles highlight examples of dissimilarities, which show that TGF-β1 and -β3 impact distinct effects on genes involved in wound healing and fibrosis in hCF 3D constructs. In addition, the fact that ACTA2 gene expression was upregulated after treatment with TGF-β1 and -β3 to varying degrees, supports previous data that these isoforms can have differential effects on corneal fibrosis.

It is well established that TGF-β1 induces the key phenotypic myofibroblast marker, αSMA and exhibits activation of FAK, as FAK signaling is implicated in myofibroblast differentiation. Since ACTA2 was significantly increased in the 3D constructs after TGF-β1 and -β3 treatment in the array analysis (Figure 1B), we further investigated the opposing mechanisms of fibrosis by examining gene and protein expression of ACTA2 (gene)/αSMA (protein) and FAK in hCF 3D constructs after TGF-β1 or -β3 treatment. Similar to the array data, the qRT-PCR results showed that TGF-β1 and -β3 treatment increased ACTA2 gene expression as compared to control (untreated); however, unlike the array data, TGF-β1 significantly upregulated ACTA2 (* p < 0.05), whereas TGF-β3 did not (Figure 2A). At the protein level, however, both TGF-β1 and -β3 increased αSMA expression when compared with control (**** p < 0.0001). Interestingly, even though both TGF-β1 and -β3 treatments were significant, TGF-β1 increased αSMA ~2-fold higher than TGF-β3 (**** p < 0.0001) (Figure 2B). Similarly, TGF-β1 significantly increased both mRNA and protein FAK expression as compared to control (* p < 0.05), but TGF-β3 did not (Figure 2C,D), even though there was a slight increase in ACTA2 in the TGF-β3 treated samples. These data indicate that TGF-β1 significantly enhanced both ACTA2/αSMA and FAK gene and protein expression, whereas TGF-β3 did not augment their expression to a similar degree, thus showing differences in their activation capacity.

Considering the differences in FAK expression between TGF-β1 and -β3 treatment, we next challenged the TGF-β-mediated myofibroblast differentiation by interfering with FAK signaling. Here, an FAK inhibitor (FAKi; CAS 4506-66-5) was used to evaluate the contribution of TGF-β-mediated myofibroblast differentiation in hCF 2D cell and 3D construct models. Using immunofluorescent staining, we examined αSMA localization in hCF that had been left untreated (control) or stimulated with TGF-β3, TGF-β1, or TGF-β1 + FAKi. In 2D culture, TGF-β1 increased αSMA expression, as did TGF-β3 treatment (Figure 3A). In contrast, FAKi was effective in abrogating the increase in αSMA localization induced by TGF-β1, maintaining a similar amount of localization as seen with untreated control. We compared and quantified the αSMA localization relative to untreated controls to show that both TGF-β1 and -β3 drove myofibroblast differentiation (* p < 0.05), but FAKi treatment was similar to untreated control (Figure 3C). In 3D constructs, TGF-β1 strongly induced αSMA expression, as expected, and αSMA was less pronounced in TGF-β3 conditions. Impressively, the expression of TGF-β1-induced αSMA was blunted by FAKi (Figure 3B). Furthermore, we quantified αSMA localization relative to untreated controls to show significant enhancement in αSMA in TGF-β1 treatment (*** p < 0.001) (Figure 3D). Similarly, the αSMA expression levels of TGF-β3 and TGF-β1 + FAKi treatments were significantly reduced compared to TGF-β1 treatment (** p < 0.01). These data indicates that interfering with FAK can attenuated TGF-β1-mediated myofibroblast differentiation in 3D construct models.

Considering the effect that FAK-inhibition has upon attenuating TGF-β1-mediated αSMA localization, we further explored whether FAKi inhibited αSMA at the molecular level. In 2D culture, qRT-PCR data indicated that TGF-β1 treatment significantly increased ACTA2 gene expression compared to that of control (* p < 0.05) (Figure 4A), while TGF-β3 treatment only showed a modest increase. With the introduction of FAKi, TGF-β1 + FAKi only showed a modest decrease in ACTA2 gene expression as compared to TGF-β1 treatment alone. At the protein level, however, both TGF-β1 and -β3 significantly increased αSMA protein expression compared with control (** p < 0.01) (Figure 4B), which agrees with the immunofluorescent data (Figure 3A). No significant difference in αSMA expression was observed between TGF-β1 and -β3 treatments. Interestingly, TGF-β1 + FAKi significantly reduced αSMA protein expression compared with that of TGF-β1 alone (** p < 0.01) (Figure 4B). In the 3D constructs, TGF-β1 significantly increased ACTA2/αSMA mRNA and protein expression compared with control (**** p < 0.0001 and * p < 0.05, respectively) (Figure 4C,D); however, despite showing significant differences compared to that of the control (* p < 0.05), TGF-β3 treatment only modestly increased ACTA2/αSMA expression. Interestingly, TGF-β1 + FAKi treatment significantly attenuated ACTA2/αSMA mRNA and protein expression when compared with TGF-β1 alone (**** p < 0.0001 and ** p < 0.01, respectively) (Figure 4C,D). When FAK mRNA and protein were analyzed under these same conditions, FAK expression was found to be significantly upregulated with TGF-β1 treatment when compared to that of untreated control (* p < 0.05), but only modestly increased with TGF-β3 (Figure 4E,F). Strikingly, FAK mRNA and protein significantly decreased with TGF-β1 + FAKi treatment compared to that of TGF-β1 alone (* p < 0.05).