FGF2 antagonizes aberrant TGFβ regulation of tropomyosin: role for posterior capsule opacity

Abstract Transforming growth factor (TGF) β2 and fibroblast growth factor (FGF) 2 are involved in regulation of posterior capsule opacification (PCO) and other processes of epithelial–mesenchymal transition (EMT) such as cancer progression, wound healing and tissue fibrosis as well as normal embryonic development. We previously used an in vivo rodent PCO model to show the expression of tropomyosin (Tpm) 1/2 was aberrantly up‐regulated in remodelling the actin cytoskeleton during EMT. In this in vitro study, we show the Tpms family of cytoskeleton proteins are involved in regulating and stabilizing actin microfilaments (F‐actin) and are induced by TGFβ2 during EMT in lens epithelial cells (LECs). Importantly, we found TGFβ2 and FGF2 played contrasting roles. Stress fibre formation and up‐regulation of α‐smooth muscle actin (αSMA) induced by TGFβ2 could be reversed by Tpm1/2 knock‐down by siRNA. Expression of Tpm1/2 and stress fibre formation induced by TGFβ2 could be reversed by FGF2. Furthermore, FGF2 delivery to TGFβ‐treated LECs perturbed EMT by reactivating the mitogen‐activated protein kinase (MAPK)/ extracellular signal‐regulated kinase (ERK) pathway and subsequently enhanced EMT. Conversely, MEK inhibitor (PD98059) abated the FGF2‐mediated Tpm1/2 and αSMA suppression. However, we found that normal LECs which underwent EMT showed enhanced migration in response to combined TGFβ and FGF2 stimulation. These findings may help clarify the mechanism reprogramming the actin cytoskeleton during morphogenetic EMT cell proliferation and fibre regeneration in PCO. We propose that understanding the physiological link between levels of FGF2, Tpm1/2 expression and TGFβs‐driven EMT orchestration may provide clue(s) to develop therapeutic strategies to treat PCO based on Tpm1/2.


Introduction
Age-related cataract, a chronic disorder of ageing, is the main cause of blindness worldwide. PCO is a common, significant complication following cataract surgery. Advances in surgical techniques, intraocular lens materials and designs have reduced the PCO rate, but it remains a significant problem worldwide, even in young and infant patients [1,2]. After cataract surgery, aberrant cell growth across the lens capsule often leads to fibrosis and secondary visual loss, known as PCO, so-called secondary or after cataracts [3]. EMT of LECs is the main cause [4][5][6][7]. EMT is related to other eye diseases such as pterygium and glaucoma and the wound healing process after eye surgery [8][9][10]. To regulate EMT is important for treatment of many eye diseases including PCO. The current treatment of PCO is YAG laser capsulotomy. However, this method can lead to uveitis, cystoid maculae oedema, elevation of intraocular pressure, retinal detachment and intraocular lens damage. Further, in paediatric cataract, it is clear that cataract extraction and correction of aphakia should be performed as soon as possible during the key period of vision [11]. However, PCO is a common complication of the surgery and also leads to amblyopia. YAG laser capsulotomy is not difficult to perform, but it needs patients' cooperation and eye fixation [11]. Secondary capsulotomy surgery is sometimes required in children [12]. We believe that it is important to study on the prevention of PCO for child patients. Further, it is important to regulate EMT and PCO in LECs for the clinical treatments using accommodative lens refilling [13] and for the regeneration of clear lens in vivo in future.
Aberrant TGF b signalling plays a central role in the pathobiology of cells or tissues by dysregulating extracellular matrix (ECM)-related genes in LECs, akin to the development of human anterior subcapsular cataract [14][15][16][17] and PCO [5,[17][18][19][20]. Moreover, TGFbs are involved in induction of tissue fibrosis, myofibroblast formation and apoptosis [21][22][23] by up-regulating genes encoding ECM proteins including aSMA, types I and III collagens. Previous studies suggest FGF may contribute to PCO development. FGF2 is expressed in human LECs [24] and involved in lens development [25] regulating cell proliferation and migration. This molecule is involved in stimulation of lens fibre differentiation in a dose-dependent manner [26,27] and also activates LEC mitosis increasing the formation of collagen [28]. FGF2 was shown to reduce the contraction of a collagen gel in bovine LECs and the proportion of cells expressing aSMA [29], indicating that FGF2 acts contrary to TGFb. Moreover, EMT in PCO was reported to be regulated by ECM components and soluble growth factors or cytokines, including epidermal growth factor, FGFs and TGFbs [26,[28][29][30][31][32], of which TGFb and FGF are key mediators of EMT and are frequently and abundantly expressed in PCO tissues. Despite many studies on the activities and roles of TGFb in lens, it remains unknown how TGFbs and FGF 2 synergistically act in EMT and how these molecules affect gene expressions including aSMA and tropomyosins (Tpms) during PCO development.
A number of target genes of FGF2 and TGFb have been identified whose expression is activated by them, some of which are overly stimulated and implicated in EMT process including aSMA, fibronectin and Tpms, and thereby PCO formation [2,[33][34][35]. More recently, our group identified modulation in expression of Tpms, specifically high molecular weight Tm isoforms from Tpm1 and Tpm2 genes, in a rodent model of PCO and in LECs obtained from cataractous humans of various ages [36]. We previously reported expression of Tpm1/2 was minimal in rat LECs, and expression of Tpm1/2 that increased selectively during EMT was linked to fibrosis in PCO [36]. Other cellular abnormalities, particularly in aberrant expression of cytoskeleton and ECM proteins, are induced because of overshooting of cellular signalling mediated by reactive oxygen species (ROS) [37]. It is known that ROS-induced damage to cells is related to ROS-driven overstimulation TGF-b1-mediated signalling [38,39] leading to overmodulation of certain genes expression, including aSMA and TGFb-induced protein (big-h3). Overexpression of those genes was involved in cataractogenesis, PCO and pathophysiological disorders of cells and tissues [31,39,40].
Previously, we showed that LECs deficient in peroxiredoxin 6 (Prdx6) display increased expression of ROS, phenotypic changes, a characteristic of terminal cell differentiation and EMT [39]. Prdx6 provides cytoprotection against internal and external environmental stresses and plays a role in cellular signalling by detoxifying ROS thereby controlling gene regulation [39,[41][42][43]. Using proteomic analysis of Prdx6-deficient (Prdx6 À/À ) mouse LECs, we found that such cells displayed elevated expression of cytoskeleton proteins Tpm1, Tpm2 and vimentin [44]. Therefore, we posit that because Tpms are implicated in regulation of cellular activities by stabilizing ECM proteins (specifically actin microfilaments), aberrant expression of Tpm1 and Tpm2 genes is likely to be involved in the phenotypic alteration of Prdx6 À/À LECs in mice.
Furthermore, Tpms are recognized as actin filament stabilizing proteins, regulating the dynamics and structural properties of the filaments by controlling the interaction of the filaments with actinbinding proteins [45][46][47]. The human tropomyosin genes should be known as TPM1 through TPM4 (Tpm1 through Tpm4 for mouse and rat tropomyosin) to be consistent with other gene nomenclatures [48]. The various isoforms generated via alternative exon splicing are listed under each gene [47]. The balance between levels of isoforms in a given cell determines the cell's Tpm functions [49][50][51][52]. Several TGFb target genes, including Tpm1, Tpm2, a-actinin1 and calponin2-encoding actin-binding proteins, were implicated in the assembly of stress fibres [51,53], and Tpms played a crucial role in stabilizing actin filaments [54]. TGFb specifically up-regulates expression of Tpm1 and Tpm2 genes but has no effect on regulation of Tpm3 and Tpm4 genes, which encode low molecular weight Tpms [51,53]. In addition, our group demonstrated an increased abundance of Tpm1/2 in differentiating LECs, suggesting involvement of TGFb-induced deleterious signalling in the induction of Tpm1/2 [55]. Given the above scenario, we surmised that over expression and activation of TGFb induced by both surgical stress and ROS during cataract surgery induces and accelerates EMT by up-regulating Tpm1/2 genes, leading to PCO [55] and the process may be modulated by FGF2 in cellular microenvironment. In this study, we determined the role and potential of FGF2 to act as an antagonist of aberrant TGFb signalling and its inducible genes/proteins such as Tpm1/2 and aSMA that play a part in EMT process and stress fibre formation in mouse and human LECs. In addition, we show the involvement of FGF2mediated MAPK/ERK 1/2 signalling on TGF-b2-induced EMT. Loss of Tpms in TGFb-evoked EMT by FGF2 was significantly linked to MAPK/ERK1/2 pathway. We provide evidence for a regulatory role (s) of Tpms in the EMT process leading to PCO, and how the process may be restored by FGF2 in a concentration-dependent manner. Understanding the underlying mechanisms of EMT, a cause of PCO, and its regulation by TGFb, FGF2 and Tpms is important to develop new treatments to inhibit EMT and postpone PCO, secondary cataract.

RNA interference
siRNAs were transfected into cells according to the protocol recommended for Lipofectamine â RNAi MAX reagent (ThermoFisher Scientific Japan Ltd., Tokyo, Japan). MLEC and HLECs were transiently transfected with siRNAs against a mixture of mouse, and human Tpm1 (Silencer â Select pre-designed siRNA, ID:s75390, ThermoFisher Scientific Japan Ltd.) and mouse and human Tpm2 (Silencer â Select pre-designed siRNA, ID:s75393) or negative control (Silencer â Select Negative control#1 siRNA: NC-siRNA). The final concentrations of the siRNAs were 10 nM.

Real-time reverse transcriptase-polymerase chain reaction (RT-PCR)
Total RNA from the MLECs or HLECs was extracted using an RNeasy Mini Kit (Qiagen, Valencia, CA, USA) per manufacturer's instructions. To measure the expression of mouse and human Tpm mRNAs, we conducted relative quantification of mRNA using a Prism7300 (Applied Biosystems â , ThermoFisher Scientific Japan Ltd., Tokyo Japan). PCR amplification was performed using a TaqMan Universal Master Mix and pre-developed mouse Tpm2 probe mix (Applied Biosystems â ), which recognize Tpm1 isoform, and mouse Tpm1 probe mix, which recognizes Tpm2, 3 and 5 isoforms. The relative quantity of Tpm1/2 and Tpm2 mRNA was determined using the comparative Ct method and then normalized using a pre-developed TaqMan ribosomal RNA control reagent VIC probe as an endogenous control (Applied Biosystems â ).

ERK-MAPK signalling pathway assay
The activities of ERK kinase were examined by Western blot analyses using antibodies against their phosphorylated forms. For the analysis of ERK activity, MLECs were treated with TGFb2 (10 ng/ml) and/or FGF2 (10 ng/ml) for 10 and 60 min. To analyse the effect of a pharmacological inhibitor of FGF receptor (FGFR) (SU5402: SU, Sigma-Aldrich) and a potent and selective inhibitor of MAP kinase (also known as MAPK/ERK kinase or MEK kinase) (PD98059: PD, Sigma-Aldrich) on FGF2-induced ERK phosphorylation, MLECs were pretreated with PD or SU with/without TGFb2 (10 ng/ml) and/or FGF2 (10 ng/ml) for 60 min and 24 hrs.

Cell migration assay
Cell migration assays were performed using Radius TM 24-Well Cell Migration Assay Kit (Cell Biolabs, Inc., San Diego, CA, USA) per manufacturers' protocol. Briefly, MLECs were grown on pre-treated Radius TM 24-well cell migration plate (Cell Biolabs, Inc.) at 2.0 9 10 5 /well for 9 hrs. Cells do not attach in the Radius TM gel spot area (~0.68 mm). Plates were treated with Radius TM gel removal solution to expose the cell-free area to cell migration. After removal of gel, cells were treated with the presence or absence of 10.0 ng/ml FGF2 or 10 ng/ml TGFb2 in DMEM containing 0.1% BSA for 24 hrs to observe their effect on cell migration.

Statistical analysis
Data were reported as means AE S.D.s and analysed by one-way ANOVA, followed by a t-test when appropriate, with P < 0.05 deemed significant.

Results
Effect of TGFb and/or FGF2 on morphological changes and the expression of Tpm1/2 and aSMA, EMT markers in MLECs in vitro To examine whether MLECs treated with TGFb or TGFb plus FGF2 showed phenotypic change, we selected TGFb2 as the major isoform expressed in eyes, which plays a role in PCO [59], and FGF2 which is the most potent in FGF family. We performed phase contrast microscopic observation to analyse phenotypes of MLECs with/without addition of TGFb2 (10 ng/ml) and/or FGF2 (10 ng/ml) for 48 hrs. Untreated MLECs were cuboidal shaped and organized in compact islets (Fig. 1A). After 48 hrs with TGF-b, cells in these islets showed a spindle-shaped morphology, elongated and underwent EMT-like change (Fig. 1B) in response to TGFb2 compared to untreated control (Fig. 1A). Many rounded dead cells were observed. FGF2-treated MLECs were more proliferated than control (Fig. 1C). MLECs treated with TGFb2 plus FGF2 (10 ng/ml each) were elongated showing fibroblastic-like changes (Fig. 1D). These results prompted us to examine the contribution of TGFb alone or in combination with FGF2 in mobilizing LECs towards EMT or vice versa. Figure 1 shows TGFb2, and TGFb2 in combination with FGF2 differentially affected LECs phenotype. Next we examined levels of genes which have been reported to be involved in regulation of shape and geometric features of cells including aSMA and Tpms. Western blotting analysis of levels of aSMA and Tpm1/2 protein in MLECs in response to TGFb1 and b2 (10 ng/ml, each) using antibody specific to aSMA and Tpm1/2 revealed that TGFb1-treated cells displayed enhanced expression of Tpm1 and aSMA proteins on day 4 ( Fig. 2A; **P < 0.016, ****P < 0.05). Similarly, TGFb2-treated MLECs showed increased abundance of Tpm1 and aSMA proteins ( Fig. 2A; *P < 0.0002, ***P < 0.0015, Fig. 3A and B; *P < 0.0042, ***P < 0.0028). TGFb2 was more potent in up-regulating expression. To determine whether increased expression of aSMA and Tpm1 by TGFb was transcriptional level, we performed real-time RT-PCR.
FGF2-treated cells also exhibited changes as shown in Figure 1. MLECs treated with FGF2 for 2 and 4 days showed a marked reduction in the expressions of Tpm1 and aSMA protein ( Fig. 4A; *P < 0.0004, **P < 0.002, *** P < 0.00001). To investigate the effect of FGF2 concentration on levels of Tpm1 or aSMA expression in MLECs, cells were treated with different concentrations of FGF2 for 2 days. We found that reduction of Tpm1 expression in MLECs was inversely related to the concentration of FGF2 (Fig. 4B; *P < 0.007, **P < 0.0005, ***P < 0.0000001). Unlike TGFb, FGF2 suppressed the expression of Tpm1 and aSMA in MLECs suggesting that FGF2 acts as an antagonist of TGFb signalling.
Next, to examine the combined effect of TGFb2 and FGF2 on Tpm1 and aSMA expression in MLECs, cells were stimulated with FGF2 (10 ng/ml) and TGFb2 (10 ng/ml) for 2 and 4 days. Based on the observed EMT-like changes in MLECs (Fig. 1D), we speculated that FGF2 and TGFb2 may induce Tpm1 or Tpm2 expression. However, FGF2 showed an opposing action against TGFb2 0 s adverse effects, with significantly reduced expression of Tpm1 and aSMA proteins ( Fig. 5A; *P < 0.0006, **P < 0.002). Similarly, in HLECs treated with TGFb2 and FGF2 in combination, the expression of Tpm1 and Tpm2 protein was increased in response to TGFb2 and decreased in response to FGF2 and TGFb2 at day 4 ( Fig 5B; P < 0.0008, **P < 0.013, ***P < 0.005). Taken together, these findings indicate that TGFb2 induces expression of Tpm1, Tpm2 and aSMA, and in contrast, FGF2 acts as antagonist and reverses the TGFb2-mediated aberrant activation of Tpm1, Tpm2 and aSMA expression in MLECs.
Effect of TGFb2 and/or FGF2 on stress fibre formation in Tpm1/2-siRNA-transfected HLECs The actin cytoskeleton plays a crucial role in regulation of cellular processes including proliferation, apoptosis, cell migration and invasion [54,60]. TGFb induces rapid reorganization of the actin cytoskeleton, whereas prolonged incubation with TGFb induces stress fibres [61,62]. Stress fibres, which are contractile bundles of actin filaments and actomyosin, are essential for cell adhesion, migration and maintenance of cell shape [63]. To examine stress fibre formation in HLECs, we performed F-actin staining with phalloidin as described in the 'Materials and Methods' section. Microscopic examination revealed an absence of stress fibres ( Fig. 6A-a, left)   HLECs, and Tpm1/2 was faintly immunolabelled in the perinuclear area of untreated HLECs (Fig. 6A-a, right). In contrast, after 48 hrs with TGF-b, F-actin was assembled into thick parallel bundles, or actin stress fibres, traversing the ventral cell surface (Fig. 6A-b, left) and Tpm1/2 was localized at these stress fibres indicating involvement of TGFb2 in their formation ( Fig. 6A-b, right). In HLECs treated with Fig. 3 Expression of Tpm1/2 and aSMA in response to TGFb2 in MLECs transfected with siRNA against Tpm1/2 and negative control. Cultured MLECs were plated in 35 mm dishes at a density of 1 9 10 5 in DMEM with 10%FBS for 24 hrs. MLECs were transfected with siRNA against Tpm1/2 and negative control. At 24 hrs after transfection, LECs were treated with 10 ng/ml of TGFb2 in DMEM containing 2%FBS for 2 days. A: Cell lysates were prepared, and Western blotting analysis was performed. aSMA was used as the marker of EMT. GAPDH was used for control of protein concentration on Western blot analysis. Data were from three experiments and were reported as means AE S.D.s. FGF2, stress fibre formation was markedly reduced (Fig. 6A-c, left) and levels of Tpm1/2 were reduced. In HLECs treated with TGFb2 and FGF2, stress fibre formation was reduced and located in the cellular periphery ( Fig. 6A-d, left) and Tpm1/2 was immunolabelled in the perinuclear area of HLECs. This result suggests that stress fibre formation in TGFb2-treated HLECs was blocked because of the presence of FGF2. Further, siRNA against Tpm1/2 was transfected in HLECs. Immunolabelling of Tpm1/2 was reduced in Tpm1/2-siRNA- transfected HLECs (Fig. 6B a-d, right). In Tpm1/2-siRNA-treated HLECs, stress fibres were not induced after treatment with or without TGFb2 and/or FGF2 (Fig. 6B b-d, right). These results suggest that FGF2 and knock-down of Tpm1/2 may inhibit TGFb2-mediated stress fibre formation in LECs. Higher expression of Tpm1/2 may induce stress fibre formation in LECs.

Effect of TGFb2 and/or FGF2 on cell migration in MLECs
To determine the effect of TGFb2 with/without FGF2 on migration, we conducted a cell migration assay as described in the 'Materials and Methods' section. We found that treatment with TGFb2 induced MLECs migration (Fig. 7A and B; *P < 0.01), which was further augmented, when MLECs were treated with both TGFb2 and FGF2 in combination ( Fig. 7A and B; *P < 0.01), indicating TGFb2 in combination with FGF2 accelerated the migration process of MLECs.
Further, treatment with FGF2 alone did not induce the migration of MLECs ( Fig. 7A and B).
FGF2 inhibits TGFb2-induced Tpm1 and aSMA gene, but promotes TGFb2-induced cell migration via activation of MAPK/ERK pathway We next examined the involvement of signalling pathway(s) responsible for FGF2-mediated down-regulation of Tmp1a expression and promotion of cell migration in the presence or absence of FGF2 with TGFb2. To investigate the effect of FGF2 with/without TGFb2 stimulation on phosphorylation of ERK1/2 in MLECs, MLECs were treated with FGF2 and/or TGFb2. Cellular extracts were immunoblotted using antibody as indicated in Fig. 8. FGF2 induced the ERK phosphorylation within/at 10 min and remained for 60 min with/without TGFb2 stimulation. These results suggest that FGF2 activated ERK pathway, as it stimulated ERK phosphorylation, and wherein treatment of TGFb2 had no effect on ERK phosphorylation ( Fig. 8; *P < 0.00002, **P < 0.0003). For validation, we used MECK inhibitor (PD) to block the ERK pathway, and FGFR antagonist (SU) to inhibit FGF2 stimulation. As shown in Fig. 9, FGF2-stimulated phosphorylation of ERK was inhibited by PD and SU at 60 min (*P < 0.0005) confirming the existence of ERK pathway in MLECs, which was activated by FGF2.
MAPK has been reported to be involved in regulation of Tpms [64]. To investigate the MAPK/ERK pathway involved in the repression of Tpm gene, we tested the effects of MECK inhibitor and FGF2 antagonist on FGF2-induced repression of Tpm1 and aSMA expression (Fig. 10). We observed PD inhibited the repression of Tpm1 induced by FGF2 at 24 hrs of treatment ( Fig. 10; *P < 0.05), and both PD and SU inhibited the repression of aSMA expression induced by FGF2 at 24 hrs of treatment ( Fig. 10; *P < 0.05, **P < 0.002).
To determine whether MAPK/ERK pathway was involved in the FGF2 and TGFb2-induced cell migration, we examined the effect of PD on migration of MLECs treated with TGFb2 and FGF2. Cell migration was induced by treatment with TGFb2 and FGF2, and PD inhibited such migration ( Fig. 11; *P < 0.02, **P < 0.05) suggesting that both factors are functionally involved in LECs' migration.

Discussion
This study provides evidence that Tpm1/2 plays an important role in TGFb2-induced EMT, stress fibre formation and cell migration in LECs. Knock-down of Tpm1/2 by siRNA blocks the elevation of aSMA and formation of actin stress fibre. The process of EMT is implicated in cancer progression, wound healing and tissue fibrosis as well as normal embryonic development [65][66][67] (Thiery, 2003;Lee et al., 2006) [68]. In cancer, EMT leads to generation of more aggressive and invasive carcinoma cells as well as cancer stem cells [68]. EMT involves disassembly of the polarized epithelial architecture and remodelling of the cell cytoskeleton, including intermediate and actin filaments. The present study revealed a dynamic process in the initiation and progression of EMT that occurs in PCO. We demonstrated that the effect of aberrant TGFb2 signalling on LECs migration, proliferation, stress fibre and EMT is influenced by FGF2 growth factor in the cellular microenvironment. Undoubtedly, this effect of FGF2 can be associated with its concentration. Interestingly, we found that FGF2 and knock-down of Tpm1/2 antagonize the effect of TGFb2 on EMT features (Figs 2, 5 and 6) by repressing the expression of Tpms, aSMA and stress fibres. In contrast, this molecule synergistically acted with TGFb2 and promoted cell migration (Fig. 7). This function of FGF2 may be related to FGF2 concentration in the cellular microenvironment. We posit that during the development of EMT, FGF2 and TGFb2 differentially affect the EMT process in a manner linked to their differential concentrations. Tpms are actin-stabilizing proteins that play a major role in maintaining cellular integrity. By assessing expression of Tpm1 and Tpm2 genes, in MLECs and HLECs, we found that growth and differentiation of LECs were differentially regulated by TGFb and FGF2 according to differential expression of Tpm genes (Figs 2-5, 7). TGFb induces epithelial to myofibroblastic transition (EMyoT) which was accompanied with Tpm and aSMA expression (a maker for EMT). Importantly, Tpms may promote the formation of stress fibres undergoing EMyoT which expresses F-actin. However, we found FGF2 suppresses the TGFb2-induced up-regulation of Tpm and aSMA (Fig. 5). Reduction of Tpm by FGF2 suppresses the formation of stress fibres and thereby activates fibroblastic LECs which induce cell migration as we observed (Figs 1 and 6). This result indicates that LECs in PCO that contain abundant FGF2 are less differentiated than those expressing predominant TGFb2 in the cellular microenvironment. We previously reported increased expression of Tpm1/2 during EMT and demonstrated that selective elevation of Tpm1/2 in rat LECs was correlated with fibrosis observed in PCO using in vivo rat model [55]. In addition, we showed that expression of Tpm1/2 was induced/ elevated in transdifferentiated multi-layered and spindle-shaped LECs in a rat model of PCO and human cataracts with anterior subcapsular cataract including in differentiated HLECs in a dislocated lens capsule [55]. Data from these previous results strongly suggested that expression of Tpm1/2 is linked with progression of PCO [55]. Several TGF-b target genes, including Tpm1 and Tpm2, have been implicated in the assembly of stress fibres [51,53]. Of these, Tpms in particular have been shown to play a crucial role in stabilizing actin filaments [54]. Furthermore, we have also reported up-regulation of Tpm1/2 expression in differentiating LECs (in the presence of TGFb), demonstrating involvement of TGFb-induced deleterious signalling in the induction of Tpm1/2. In the past, several in vitro and in vivo studies have examined the role of TGFb2 in EMT and wound healing processes using LECs [5,6,69]. Transdifferentiation analogous to that observed in PCO can be induced by exposing LECs to TGFb [5,6,70,71].
Moreover, a major contribution of FGF2 in PCO development has been reported [1,27,72]. FGF2 is a potent mitogenic growth factor which is present in the eye lens environment [27], and concentration of FGF2 may increase after cataract surgery [72]. For instance, Wallentin et al. [73] showed that aqueous FGF2 was increased in rabbits up to 30 days following cataract extraction. Furthermore, in a study using a rabbit PCO model, the level of active TGFb decreased and increased FGF2 stimulated cell proliferation, immediately after surgery, and at around 2 weeks after surgery, active TGFb returned to normal level stimulating EMT [35]. TGFb2 has been shown to inhibit the proliferative effect of FGF2 on rabbit LECs growth [73]. FGF2 reduced the contraction of a collagen gel by bovine LECs and the proportion of cells expressing aSMA [29], indicating that FGF2 has opposing actions to TGFb2. Furthermore, simultaneous treatment of rat LECs explants with TGFb2 and FGF2 were more affected by TGFb2 than corresponding high cell coverage explants, showing greater cell loss, more marked formation of spindle cells and expression of aSMA [71]. Excessive deposition of ECM and formation of plaques of swollen cells, also features of PCO [1,71], occurred only when TGFb was supplemented with FGF2 [70,71].
Our study found cotreatment with TGFb2-and FGF2-induced spindle-shaped fibroblastic formation of HLECs and accelerated the EMT process. Furthermore, we found that TGFb2 and FGF2 in combination suppressed the formation of stress fibres in LECs and the expression of Tpm and aSMA. Importantly, Tpm1/2 knock-down inhibited the TGFb2-induced formation of stress fibres. Tpms have been shown to stabilize actin filaments [54]. Thus, our study revealed that Tpms may induce stress fibre in response to TGFb2 in the EMT process.
We found that TGFb2 induced differentiation of LECs into two types of mesenchymal cells: one was Tpm1-and aSMA-positive myofibroblastic cells generated through EMyoT by TGFb2 alone, and the other was activated Tpm1-and aSMA-negative fibroblastic cells generated through EMT by both TGFb2 and FGF2 in combination. Furthermore, our result also indicated that TGFb2 enhanced cell migration compared with non-treated cells, and the co-addition of FGF2 further promoted cell migration (Fig. 7). Based on these data, we believe that Tpm abundance may inhibit cell migration and reduced level of Tpm may induce cell migration as we showed that FGF2 markedly suppressed the expression of Tpms in LECs.
Moreover, FGF has been shown to be a major activator of the ERK-MAPK pathway in the eye lens in vivo [74,75]. Further, FGF2 activation of ERK1/2 signalling could influence TGFb2 induction of Tpm1 gene expression. The present study showed the inducible expression of Tpm1 was attenuated in the presence of PD, a specific inhibitor of MAPK/ERK signalling, and SU, an inhibitor of FGFR2. Also, cell migration induced by TGFb2 and FGF2 was inhibited in the presence of PD. These data support the notion that ERK signalling affects cell migration and the formation of stress fibres by suppressing the expression of Tpm.
In summary, we demonstrated both independent and combined roles of TGFb2 and FGF2 in the differential regulation of EMT and how this process is associated with Tpm expression including the implications for PCO development. Importantly, we found that FGF2 acts as an antagonist against TGFb-mediated EMT progression. We demonstrated that FGF2-induced ERK signalling attenuates aberrant expression of Tpm1 and cell migration. Our results provide a novel insight into the regulation and function of Tpms during LEC differentiation, which is influenced profoundly by growth factors. These findings reveal that opposing effects of FGF2 and TGFb2 on Tpm1/2 gene expression control the phenotypic plasticity of LECs on PCO progression. Further, Tpm1/2 may regulate other eye diseases such as pterygium and glaucoma and wound healing processes related to EMT.
We hope this study provides clues to develop new therapies of PCO and other EMT-related diseases targeting the balance regulation Fig. 9 Effect of FGFR antagonist (SU) and MECK inhibitor (PD) on FGF2-induced activation of ERK pathway. MECK inhibitor (PD) to block ERK pathway and FGFR antagonist (SU) to inhibit FGF2 stimulation were used. To evaluate the effect of FGF2 with/without TGFb2 stimulation on phosphorylation of ERK1/2, MLECs were treated with 0 or 10 ng/ml of TGFb2 and/or FGF2 in DMEM containing 0.1% BSA with/without SU or PD for 10 min or 24 hrs. Cell lysates were prepared, and Western blotting analysis was performed using anti-rabbit p44/42 MAPK (Erk1/2) monoclonal Ab or antiphospho-rabbit p44/42 MAPK (Erk1/2) monoclonal Ab. Data were from three experiments and were reported as means AE S.D.s. Fig. 10 Effect of MECK inhibitor (PD) and FGF2 antagonist (SU) on repression of Tpm1 and aSMA expression in response to FGF2. Cultured MLECs were plated in 35 mm dishes at a density of 1 9 10 5 in DMEM with 10%FBS for 24 hrs. MLECs were treated with 0 or 10 ng/ml of TGFb2 plus FGF2 in DMEM containing 0.1% BSA with/without SU or PD for 60 min and 24 hrs. Cell lysates were prepared, and Western blotting analysis was performed using anti-Tpm1/2 Ab and anti aSMA Ab, with GAPDH used for control of protein concentration. Data were from three experiments and were reported as means AE S.D.s.

925
of Tpms by growth factors. However, further in-depth studies will be required to fully clarify the underlying mechanism of Tpm1/2 involvement in EMT process. Fig. 11 Effect of MECK inhibitor (PD) on migration of MLECs treated with TGFb2 and FGF2. MLECs were plated in 24-well plates, pre-coated with collagen type I, at a density of 1 9 10 5 in DMEM with 10%FBS for 24 hrs. After hydrogel removal to expose the cell-free region, MLECs were treated with 0 or 10 ng/ml of TGFb2 plus FGF2 with/without PD in DMEM containing 0.1% BSA for 24 hrs. Phase contrast micrographs were then taken with a digital camera. Data were from three experiments and were reported as means AE S.D.s. Scale bar, 180 lm.