MicroRNA-204-5p Regulates Epithelial-to-Mesenchymal Transition during Human Posterior Capsule Opacification by Targeting SMAD4

Abstract

Purpose.: To investigate the role of microRNA (miRNA) in regulating epithelial-to-mesenchymal transition (EMT) during human posterior capsule opacification (PCO).

Methods.: A miRCURY LNA microRNA array was used to evaluate the miRNA profiles of human PCO tissues and normal attached lens epithelial cells (LECs). An in vitro human donor capsular bag model was used to investigate the role of miRNAs in the EMT during PCO. The expression of SMAD4, phospho-SMAD2/3, and a panel of EMT markers was detected by Western blot and quantitative RT-PCR.

Results.: The results of miRNA profiling in human PCO tissues and normal attached LECs demonstrated that, among other miRNAs, miR-204-5p expression was down-regulated. Using bioinformatics, we identified SMAD4, one of the mediators of TGF-β/SMAD signaling, as a predicted target of miR-204-5p. Overexpression of miR-204-5p in primary LECs increased E-cadherin expression and decreased the expression of vimentin and alpha smooth muscle actin. Furthermore, miR-204-5p overexpression enhanced the repression of TGF-β2–induced EMT in the presence of SMAD4 small interfering RNA.

Conclusions.: Our data provide firm evidence of a role for miR-204-5p in the direct regulation of EMT through its targeting of SMAD4 and, consequently, TGF-β signaling. Because of its ability to repress the EMT, miR-204-5p may be a novel target for PCO therapeutic intervention.

Introduction

A secondary cataract, also termed as posterior capsular opacification (PCO), is the most common complication of cataract surgery. ¹ Following the insult of surgery, residual lens epithelial cells (LECs) rapidly grow at the equator and under the anterior lens capsule. These cells proliferate and migrate onto the posterior capsule, where light scattering changes can give rise to secondary visual loss. ² This condition is known as PCO. The remaining LECs undergo an epithelial-to-mesenchymal transition (EMT), resulting in the formation of fibroblasts instead of normal differentiation into lens fiber cells. 

EMT is a process during which epithelial cells lose their differentiated phenotypes and change their morphology and transcriptional program to those characteristic of mesenchymal cells. ³ At present, three types of EMT have been classified based on the phenotype of output cells. Previous studies determined that LECs undergo a transition into mesenchymal cells and differentiate into lens fiber cells, characteristic of type 2 EMT. ^4,5 Published studies suggest that EMT is a crucial event in the transformation of LECs during wound healing after cataract surgery ^6–8 and is therefore a prominent part of the recovery process after cataract surgery and one of the leading causes of PCO. Importantly, EMT has been observed at the initial stages of recovery in mice undergoing regeneration, indicating that the process includes an initial phase of repair and lens differentiation. ⁹ A series of parallel and convergent signaling pathways that lead to the common phenotypes of EMT in PCO are activated in LECs, such as the transforming growth factor-beta/mothers against decapentaplegic homolog (TGF-β/SMAD) signaling pathway, ^6,10 the Src family kinase/focal adhesion kinase (SFK/FAK) signaling pathway, ¹¹ and the ubiquitin proteasome pathway. ¹² Recent studies suggest that TGF-β induces LECs to undergo EMT and implicate the TGF-β/SMAD signaling pathway as a central factor in the pathologic fibrosis of PCO. ^13,14 SMADs are intracellular proteins that transduce extracellular signals from TGF-β ligands to the nucleus, where they activate downstream TGF-β gene transcription. ^15,16 SMAD4 is a central signaling component of the SMAD pathway. ¹⁷ Upon TGF-β ligand stimulation, SMAD4 forms a heteromeric complex with SMAD2 and SMAD3 (R-Smads), which are phosphorylated by an activated TGF-β receptor complex consisting of two type I and two type II transmembrane serine threonine kinase receptors. ¹⁸  

Recently, the microRNA (miRNA)-dependent posttranscriptional regulation of genes associated with lens development was proposed to also play a role in lens regeneration and cataract-associated EMT. The differential expression of miRNAs is often associated with pathophysiological disorders. Therefore, PCO is likely to be regulated by miRNAs, and the identification of miRNAs that are differentially expressed between human normal lens capsule samples and PCO samples may help to identify miRNAs that are involved in PCO and establish the basis for unraveling their pathogenic roles. We previously showed that the miR-184 and miR-204-5p competitive RNA network was implicated in the control of mouse lens regeneration models that mimic PCO. ¹⁹ In the present study, we used miRNA microarray technology and human lens capsule bag cultures to identify the specific miRNA expression pattern in PCO tissues and ascertain the potential role of miR-204-5p in PCO-associated EMT. 

Materials and Methods

Patient Tissue Sample Collection

This study was conducted according to the principles outlined in the Declaration of Helsinki and was approved by the institutional ethics review board at the Shandong Eye Institute. Fresh PCO tissues were obtained from the Shandong Eye Institute (Qingdao, China). Written informed consent for the collection of the samples and subsequent analysis was obtained from each patient. Normal attached LEC samples from organ donors were provided by the International Federation of Eye Banks at the Eye Bank of Shandong, China (Qingdao, China). 

RNA Extraction and miRNA Microarray Study

Total RNA was isolated using the NucleoSpin RNA II System (Macherey-Nagel, Düren, Germany), and first-strand cDNA was synthesized using a MMLV Reverse Transcriptase 1st-Strand cDNA Synthesis Kit (Epicentre Biotechnologies, Madison, WI) according to the manufacturer's protocol. The miRNA microarray analysis, including labeling, hybridization, scanning, normalization, and data analysis, was performed by KangChen Bio-tech (Shanghai, China) using a miRCURY LNA microRNA Array (v.16.0, Exiqon, Vedbaek, Denmark). 

In Vitro Human Donor Capsular Bag Model

Capsular bags were obtained from human eyes donated for corneal transplantation from the International Federation of Eye Banks at the Eye Bank of Shandong, China (Qingdao, China). The model used was previously described by Liu et al. ²⁰ The first part of this study was designed to detect the expression of EMT markers, including E-cadherin, vimentin, and α-SMA, in the capsular bag model. The capsular bags were maintained in Dulbecco's modified Eagle's medium/nutrient mixture F12 (DMEM/F12, Life Technologies, Gaithersburg, MD) with 10% fetal bovine serum alone for 0, 3, 7, or 14 days (n = 3 for each time point). The second part of this study, which did not include TGF-β2 treatment, utilized four groups: group 1, capsular bags that were cultured for 3 days were used as the control group; group 2, cells were transfected with a miR-204-5p mimic control; group 3, cells were transfected with a miR-204-5p mimic; and group 4, cells were transfected with a miR-204-5p inhibitor. Each group contained three samples. The expression of SMAD4, phospho-SMAD2/3, and a panel of EMT markers were detected by Western blot or quantitative (q)RT-PCR. The third part of this study employed a 24-hour TGF-β2 treatment (10 ng/mL), followed by cotransfection of the cells with or without SMAD4 small interfering RNA (siRNA) and a miR-204-5p mimic to investigate the effect of miR-204-5p on the LEC EMT. Capsular bags that were cultured for 3 days were also used as a control, and each group contained three samples. 

Statistical Analysis

The results of luciferase activity assay, qPCR, and Western blot were expressed as the mean ± standard deviation. Statistical analysis was performed using the one-way ANOVA, comparing the groups with the Student-Newman-Keuls test and the least significant difference procedure using commercial software (SPSS 11.5; SPSS, Chicago, IL). A value of P < 0.05 was considered statistically significant. 

Information on the qRT-PCR analysis, transfection of capsular bags and efficiency detection, luciferase activity assay, Western blot analysis, and immunohistochemistry is available in the Supplemental Materials and Methods. (see Supplementary Material and Supplemental Materials and Methods file). The primers used in qRT-PCR for miRNAs detection were listed in Supplementary Table S2 (see Supplementary Material and Supplementary Table S2) and the primers used in qRT-PCR for EMT markers detection were listed in Supplementary Table S3 (see Supplementary Material and Supplementary Table S3). Antibodies used in Western blot were listed in Supplementary Table S4 (see Supplementary Material and Supplementary Table S4). 

Results

Epithelial-to-Mesenchymal Transition Occurs during Human Posterior Capsular Opacification

To investigate whether EMT occurred in human PCO, Western blot and immunohistochemistry analyses were performed to detect E-cadherin, vimentin, and alpha smooth muscle actin (α-SMA) protein expression in the human PCO samples. Western blot was used to quantify the changes in protein expression (Fig. 1A, top panel shows data from the gels, and the lower panel shows normalization to GAPDH). E-cadherin expression was undetectable in the PCO tissues (P < 0.05, n = 3). Vimentin is often used as a marker for cells undergoing EMT during both normal development and metastatic progression. The expression of vimentin was higher in the PCO tissues than in the normal samples (P < 0.05, n = 3). Moreover, the expression of α-SMA, which is a late marker of the EMT, was also up-regulated in the PCO tissues (P < 0.05, n = 3). The results of the immunohistochemistry analysis are shown in Figure 1B. The normal attached LECs demonstrated high E-cadherin expression and little or no expression of vimentin and α-SMA. The expression of E-cadherin was significantly down-regulated in PCO tissues; whereas, the expression of vimentin and α-SMA was up-regulated. The results of both the Western blot and immunohistochemistry analyses support the hypothesis that EMT is a significant event during human PCO. 

Figure 1. 

 

Expression of EMT markers in human PCO tissues. (A) Changes in protein expression as determined by Western blot. Expression of E-cadherin is lost in the PCO tissues (P < 0.05, n = 3). Expression of vimentin was higher in the PCO samples than in the normal attached LECs (P < 0.05, n = 3). The expression of α-SMA was also up-regulated in the PCO tissues (P < 0.05, n = 3) (top, data from the gels; bottom, normalization to GAPDH). Significant differences between the PCO tissues and normal attached LECs are indicated by an asterisk (*P < 0.05). (B) Results of the immunohistochemistry analysis. Normal attached LECs demonstrated high E-cadherin expression and little or no expression of vimentin and α-SMA. In contrast, E-cadherin was down-regulated, and vimentin and α-SMA were up-regulated in the PCO tissues.

Differential miRNA Expression between Normal Attached LECs and PCO Tissues

To investigate whether miRNAs are differentially expressed in normal attached LECs and PCO tissues, we collected six normal attached LEC samples from an eye bank and six PCO tissues from six different patients with PCO. The samples were pooled into four groups (n1, n2, p1, and p2; N = 2). Group n1 comprised three normal attached LEC samples, and group n2 included another three remaining normal attached LEC samples. Group p1 comprised three PCO tissues, and group p2 included another three remaining PCO tissues. Since the N was only equal to 2, statistical analysis couldn't be done. Therefore, we chose 122 human miRNAs that were found to be differentially expressed by more than 2.0-fold (mean value). Among these miRNAs, 10 miRNAs were identified as being up-regulated more than 2.0-fold in the PCO tissues compared with those in the normal attached LECs. In contrast, 112 different miRNAs were down-regulated more than 2.0-fold in the PCO tissues compared with those in the normal attached LECs. Significantly, the expression of miR-204-5p and miR-204-3p were down-regulated 66.8-fold and 12.1-fold, respectively. The results of the cluster analysis are shown in Figure 2A. 

Figure 2.

 

miRNA profiles in normal attached LECs and PCO tissues. (A) Hierarchical clustering was performed with the normalized miRNA data (greater than 2-fold change). A total of 122 miRNAs were identified as having significantly altered expression between the PCO tissues and normal attached LECs. This figure depicts some of the 122 regulated miRNAs (mean values). Rows, miRNA; Column, normal attached LECs (n1, n2) and PCO tissues (p1, p2). For each miRNA, the red color indicates genes with high expression, and the green color denotes genes with low expression (B and C). Validation of select microarray data by qRT-PCR using the same extracted total RNA used for the microarray analysis. We also collected another one PCO tissue to group p3 and pooled another three normal attached LECs to group n3. Groups n1, n2, and n3 and p1, p2, and p3 were used. Down-regulated (B) and up-regulated (C) miRNAs in PCO tissues. Triplicate assays were performed for each sample, and the relative amount of each miRNA was normalized to the U6 snRNA. Significant differences between the PCO tissues and normal attached LECs are indicated by an asterisk (*P < 0.05).

To confirm whether the expression of these miRNAs was significantly different in the normal attached LECs and PCO tissues, qRT-PCR was performed to compare expression levels using the same extracted total RNA used for the microarray analysis. We also collected another one PCO tissue (group p3) and pooled another three normal attached LECs (group n3). Groups n1, n2, and n3 and p1, p2, and p3 were used for qRT-PCR analysis. Individual miRNA levels in each sample were quantified and normalized to U6 expression. Of the filtered miRNAs, six different miRNAs (hsa-miR-31, −184, −204-5p, −498, −4279, and −1469) were validated as showing significant differences between the normal attached LECs and PCO tissues (P < 0.05). The relative expression levels of hsa-miR-31, hsa-miR-184, and hsa-miR-204-5p were down-regulated significantly in PCO tissues compared with normal attached LECs (Fig. 2B). In contrast, hsa-miR-498, hsa-miR-4279, and hsa-miR-1469 were up-regulated significantly in PCO tissues compared with normal attached LECs (Fig. 2C). One of the down-regulated miRNAs, miR-204-5p, merited further investigation because it was predicted to target SMAD4, which plays an important role in EMT. A literature search did not reveal any known correlation between miR-204-5p and the clinico-pathological features of PCO. 

Human Donor Capsular Bags as a PCO Model In Vitro

After demonstrating the association of EMT with the onset of human PCO and identifying differential miRNA expression between PCO tissues and normal attached LECs, we investigated the role and function of miRNA hsa-miR-204-5p, which was significantly decreased in the PCO tissues, using a human donor capsular bag explant culture as a suitable model to study the EMT in PCO. ²⁰ The general appearance of representative capsular bags at different time points ranging from 0 to 14 days after pinning is shown in Figure 3A. Six pins were used to preserve the natural circular outline of the bag. The extent of the capsulorhexis could be observed clearly by phase-contrast microscopy. Early wrinkles began to form within 3 to 7 days of culture (Figs. 3A-ii and 3A-iii, white arrow). Wrinkles then increased in number and magnitude, with the formation of major wrinkles observed by 14 days (Fig. 3A-iv, white arrows). The major wrinkles were regions of light scatter, which could be observed clearly by dark-field micrograph (Fig. 3A-iv). As shown in Figure 3B, the migration of primary human LECs in the capsular bags was observed under a light microscope at different time points. Cell growth was observed on the anterior capsule during the first day of culture (Fig. 3B-ii). Then, the cells progressed beyond the rhexis and grew towards the center during days 3 to 7 in culture (Figs. 3B-iii and 3B-iv). By day 14, the cultures became confluent on the posterior capsule, and the cells acquired a regular appearance (Fig. 3B-v). 

Figure 3.

 

General appearance of a cultured human donor capsular bag. (A) Low-power, dark-field views of capsular bag preparations immersed in culture medium at various stages of growth: (i) immediately after pinning, (ii) after 3 days, (iii) after 7 days, and (iv) after 14 days. Light-scattering areas arise from the debris left after the operation. The disc-shaped opening in the anterior capsule can be observed clearly, revealing the posterior capsule beneath. Early wrinkles began to form within 3 to 7 days of culture (ii, iii, white arrow), and the wrinkles increased in number and magnitude over time. By 14 days in culture, major wrinkles were formed (iv, white arrow). The major wrinkles are regions of light scattering, which could be observed clearly in the dark-field micrograph (iv). (B) Low-magnification phase-contrast micrographs of cells growing on the capsular bag after (i) 0, (ii) 1, (iii) 3, (iv) 7, or (v) 14 days. The images show human lens epithelium growing on the anterior (AC) and posterior (PC) areas of the lens capsule. In each image, the AC can be observed on the left, and the central PC is on the right; the edge of the AC is indicated by the black arrows. Cell regrowth was observed on the AC during the first day of culture (ii). The cells progressed beyond the rhexis and grew toward the center within 3 to 7 days of culture (iii, iv). By day 14, the cells occupied the entire PC and acquired a regular appearance (v).

The expression of EMT markers was determined in our in vitro PCO model (Fig. 4). LECs produced E-cadherin, vimentin, and α-SMA proteins in the capsular bag PCO model (Fig. 4A). The staining intensity of E-cadherin was moderate and appeared to be greater after 1 day of culture (Fig. 4A-i) and then decreased within 3 to 14 days of culture (Figs. 4A-ii, 4A-iii, and 4A-iv). The capsular bag LECs appeared to have little or no α-SMA or vimentin protein expression at day 3 of culture (Figs. 4A-v and 4A-ix). However, the staining intensity of both α-SMA and vimentin increased within 3 to 7 days of culture (Figs. 4A-vi, 4A-vii, 4A-x, and 4A-xi) and then decreased again by 14 days of culture (Figs. 4A-viii and 4A-xii). Because the immunocytochemistry results demonstrated that EMT could be observed within 3 to 14 days, we used the lens capsular bag at 3 days of culture as a control. Immunocytochemistry is not quantitative, so we performed qRT-PCR to quantify the mRNA expression of select EMT markers. The mRNA levels of E-cadherin, α-SMA, and vimentin were evaluated in the PCO model in vitro (Figs. 4B, 4C, and 4D). Expression of E-cadherin was highest when initially cultured (day 0) and decreased 3.35-fold by day 3, remaining at similar levels through day 14 (Fig. 4B). In contrast, expression of vimentin and α-SMA increased over the duration of the capsular bag culture. The highest expression levels of α-SMA were observed throughout the first 3 days of culture, at which point the reduction of α-SMA indicated the beginning of LECs transdifferentiation. Due to the complete transdifferentiation of most LECs by 7 days of culture, the expression of α-SMA gradually decreased by day 14 (Fig. 4C). The highest expression of vimentin was also observed at 3 days of culture and was greater than 2-fold relative to the levels at day 0. Starting at day 7, the expression of vimentin gradually decreased (Fig. 4D). These results are consistent with the data we collected when we compared fresh normal attached LECs with PCO tissues (Fig. 1), indicating that the capsular bag cultures faithfully recapitulated the conditions observed in vivo. 

Figure 4.

 

Expression of EMT markers in the PCO model in vitro. (A) Results of the immunohistochemistry analysis. The LECs produced E-cadherin, vimentin, and α-SMA protein in the capsular bag PCO model. The staining intensity of E-cadherin appeared to be greater at 1 day of culture (i) and then decreased from 3 days to 14 days (ii–iv). The LECs in the capsular bag did not appear to express α-SMA or vimentin protein at day 1 of culture (v, ix). However, the staining intensity of both α-SMA and vimentin increased within 3 to 7 days of culture (vi, vii, x, xi) and then decreased after 14 days of culture (viii, xii). (B–D) Results of the qRT-PCR analysis. The expression of E-cadherin in LECs at 3 days of culture decreased 3.35 -fold compared with the expression at 0 days of culture and continue to decrease at later stages (B). Although both α-SMA (C) and vimentin (D) were low at day 0, their expression levels increased, especially at day 3 of the capsular bag culture, which coincides with the beginning of LEC transdifferentiation. Significant differences are indicated by an asterisk (*P < 0.05).

MiR-204-5p Directly Targets and Down-Regulates Human SMAD4 and SMAD4 Protein Levels

Given that the biological significance of miRNA deregulation relies on the target genes, we performed computational predictions of the target genes for the most significantly down-regulated miRNA in our study, hsa-miR-204-5p.²¹ We found that the TGF-β/SMAD signaling pathway was significantly enriched among the predicted targets of miR-204-5p (P = 0.035) (Supplementary Table S1). From this bioinformatics database, miR-204-5p was predicted to target the SMAD4 gene at a site in its 3′-UTR, a 7mer located at position 2880–2887 (Fig. 5A). To test the possibility of a direct link between miR-204-5p and human SMAD4, we performed a dual luciferase reporter assay in primary human LECs from the capsular bag model. The results are shown in Figure 5B. A significant decrease in relative luciferase activity was observed when pGL3-SMAD4-3′-UTR was cotransfected with a miR-204-5p mimic. The results of transfection efficiency detection were shown in Supplementary Figure 1 (see Supplementary Material and Supplementary Fig. S1). In the presence of the miR-204-5p mimic, expression of the renilla luciferase reporter was repressed 2.5-fold compared with the vector-only control. Significantly, partial deletion of the perfectly complementary sites in the 3′-UTR of SMAD4 abolished the suppressive effect due to the disruption of the interaction between miR-204-5p and SMAD4. Furthermore, SMAD4 protein expression was decreased by transfection with the miR-204-5p mimic but increased by transfection with a miR-204-5p inhibitor in LECs (Fig. 5C). However, SMAD4 mRNA levels were not significantly influenced by the overexpression or inhibition of miR-204-5p (data not shown), suggesting that SMAD4 expression was primarily inhibited by miR-204-5p at the translational level. Phospho-SMAD2/3 protein expression was also detected by Western blot, as shown in Figure 5C, and its protein level was not significantly influenced by the overexpression or inhibition of miR-204-5p. Together, these results confirmed that SMAD4 is a direct target of miR-204-5p and is regulated by miR-204-5p. 

Figure 5.

 

MiR-204-5p directly targets and down-regulates human SMAD4 and SMAD protein levels. (A) Position of the miR-204-5p target sequence in the 3′-UTR region of SMAD4 mRNA. Note that the continuous 7mer is 100% complementary. (B) A dual luciferase reporter assay in primary human lens epithelial cells from the capsular bag model. A significant decrease in relative luciferase activity was observed when the pGL3-SMAD4-3′-UTR was cotransfected with the miR-204-5p mimic. In the presence of the miR-204-5p mimic, the expression of the renilla luciferase reporter was repressed 2.5-fold compared with the empty vector control. However, partial deletion of the perfectly complementary site in the 3′-UTR of SMAD4 abolished the suppressive effect due to the disruption of the interaction between miR-204-5p and SMAD4. The data represent the mean value ± SD of four different experiments. Significant differences (*P < 0.05) are identified by asterisks. (C) SMAD4 protein levels are down-regulated by miR-204-5p (top, data from the gels; bottom, normalization to GAPDH). There were four groups in this part: 1, capsular bags cultured for 3 days were used as the control group; 2, the cells were transfected with an miR-204-5p mimic control; 3, the cells were transfected with an miR-204-5p mimic; and 4, the cells were transfected with an miR-204-5p inhibitor. Each group contained three samples. SMAD4 protein expression was decreased less than 2-fold (asterisk) by transfection with the miR-204-5p mimic but restored by transfection with the miR-204-5p inhibitor in LECs. However, the SMAD4 mRNA level was not significantly influenced by the overexpression of miR-204-5p (data not shown), suggesting that SMAD4 expression is primarily inhibited by miR-204-5p at the translational level. Phospho-SMAD2/3 protein expression was not significantly influenced by the overexpression or inhibition of miR-204-5p. Together, these results confirmed that SMAD4 is a direct target of and is regulated by miR-204-5p.

Up-Regulation of miR-204-5p Repressed the EMT in an In Vitro PCO Model

To further investigate the role of miR-204-5p in the development and progression of PCO by its ability to repress EMT, Western blot, and qRT-PCR analyses were performed to detect the protein and mRNA levels of E-cadherin, α-SMA, and vimentin, which are markers of EMT, in an in vitro PCO model. The results shown in Figure 6A illustrate that E-cadherin protein expression increased when the LECs were transfected with the miR-204-5p mimic but decreased when the LECs were transfected with the miR-204-5p inhibitor. In contrast, the protein expression of both α-SMA and vimentin decreased when the LECs were transfected with the miR-204-5p mimic but increased when LECs were transfected with the miR-204-5p inhibitor. Similarly, E-cadherin mRNA levels were increased by the miR-204-5p mimic but decreased by the miR-204-5p inhibitor (Fig. 6C), while α-SMA and vimentin mRNA levels were decreased in LECs transfected with the miR-204-5p mimic but increased in LECs transfected with the miR-204-5p inhibitor (Figs. 6B and 6D). These results clearly show that the up-regulation of miR-204-5p represses EMT in an in vitro PCO model. 

Figure 6.

 

Expression of EMT markers when miR-204-5p is overexpressed in LECs. This experiment was conducted using treatment with 10 ng/mL TGF-β2 for 24 hours, and the cells were then cotransfected in the presence or absence of SMAD4 siRNA and the miR-204-5p mimic to investigate the effect of miR-204-5p on the LEC EMT. Capsular bags cultured for 3 days were used as the control group, and each group contained three samples. (A) Protein expression levels of the EMT markers. E-cadherin protein expression was increased when the LECs were transfected with the miR-204-5p mimic but decreased when the LECs were transfected with the miR-204-5p inhibitor. In contrast, the expression of both α-SMA and vimentin was decreased when the LECs were transfected with the miR-204-5p mimic but increased when the LECs were transfected with the miR-204-5p inhibitor. Top, the gels; bottom, the normalization graphs. (B–D) Up-regulation of miR-204-5p repressed the mRNA expression of EMT markers. The abundance of mRNAs encoding α-SMA (B) and vimentin (C) was reduced in LECs transfected with the miR-204-5p mimic but increased in LECs transfected with the miR-204-5p inhibitor. However, the abundance of E-cadherin mRNA was increased by the miR-204-5p mimic and decreased by the inhibitor. These observations suggest that the up-regulation of miR-204-5p represses the EMT in an in vitro PCO model.

The finding that SMAD4 is directly targeted by miR-204-5p indicates that this important pathway can be negatively regulated by miR-204-5p. To verify our data, we examined whether silencing SMAD4 would enhance the effect of miR-204-5p when EMT was induced by TGF-β2 in our capsular bag model. Within 3 days of culture, capsular bag cells were starved for a further 24 hours and then treated with 10 ng/mL TGF-β2 (Fig. 7). After 24 hours of TGF-β2 stimulation, the levels of SMAD4, E-cadherin, vimentin, and α-SMA were quantified by Western blot. In the presence of TGF-β2, LEC expression of SMAD4, vimentin, and α-SMA increased 2.98 ± 0.031, 2.83 ± 0.026, and 4.15 ± 0.038 fold, respectively, relative to LECs without TGF-β2. In contrast, E-cadherin expression decreased 0.82 ± 0.05 fold relative to LECs without exposure to TGF-β2. These results indicate the induction of EMT. When SMAD4 was inhibited by RNA interference, the expression of SMAD4, vimentin, and α-SMA were decreased 1.03 ± 0.025, 1.78 ± 0.028, and 2.11 ± 0.022 fold, respectively, compared with TGF-β2 stimulation, but the expression of E-cadherin was increased 1.50 ± 0.013 fold, suggesting that SMAD4 is involved in EMT during the development of PCO. We next assessed whether miR-204-5p overexpression enhanced its inhibitory effect on TGF-β2–induced EMT in the presence of SMAD4 siRNA. When cells were cotransfected with SMAD4 siRNA and the miR-204-5p mimic, the expression of SMAD4, vimentin, and α-SMA was decreased 0.89 ± 0.017, 0.79 ± 0.021, and 1.15 ± 0.030 fold, respectively, but the expression of E-cadherin was increased 1.65 ± 0.013 fold. Our results clearly indicate that in our model up-regulation of miR-204-5p affected the level of SMAD4 and, as a consequence, reduced the levels of EMT markers, leading to the inhibition of EMT. 

Figure 7.

 

Silencing of SMAD4 expression enhanced the repressive effect of miR-204-5p in TGF-β2–induced EMT in a capsular bag model. After 24 hours of TGF-β2 stimulation, the levels of SMAD4, E-cadherin, vimentin, and α-SMA were quantified by Western blot. In the presence of TGF-β2, the expression levels of SMAD4, vimentin, and α-SMA in the LECs increased 2.98 ± 0.031, 2.83 ± 0.026, and 4.15 ± 0.038 fold, respectively, relative to the LECs that were not exposed to TGF-β2. In contrast, the expression of E-cadherin decreased 0.82 ± 0.05 fold relative to the LECs that were not exposed to TGF-β2. When SMAD4 was silenced by RNA interference, the expression of SMAD4, vimentin, and α-SMA decreased 1.03 ± 0.025, 1.78 ± 0.028, and 2.11 ± 0.022 fold, respectively, but the expression of E-cadherin increased 1.50 ± 0.013 fold compared with TGF-β2 stimulation. When the cells were cotransfected with SMAD4 siRNA and the miR-204-5p mimic, the expression of SMAD4, vimentin, and α-SMA decreased 0.89 ± 0.017, 0.79 ± 0.021, and 1.15 ± 0.030 fold, respectively, but the expression of E-cadherin increased 1.65 ± 0.013 fold compared with TGF-β2 stimulation.

Discussion

In this study, we identified potential miRNA modulators of EMT during human PCO. Of the regulated miRNAs, miR-204-5p piqued our interest because it was down-regulated approximately 67-fold. This observation suggested that it was a potential regulator of EMT during PCO, and thus miR-204-5p was selected for further analysis. miR-204 has previously been shown to be highly expressed in fetal human retinal pigment epithelium and zebrafish lenses. ²² In addition, miR-204 is uniformly expressed in the epithelia of the anterior region of the mouse lens, the nonpigmented epithelium of the ciliary body, and the corneal anterior surface. ^23,24 Functional studies of miR-204 have revealed that it can act as an important endogenous negative regulator of Runx2, which inhibits osteogenesis and promotes the adipogenesis of mesenchymal progenitor cells and bone marrow stromal cells. ²⁵ Additionally, the dysregulation of miRNA-204 mediates the migration and invasion of endometrial cancer by regulating FOXC1. ²⁶ Several targets of miR-204 previously validated in other cell types have profound roles in TGF-β signaling, including prostaglandin-endoperoxide synthase 2. ²⁷ Wang et al. ²⁸ found that TGF-β receptor 2 and Snail2 are direct targets of miR-204 and that a reduction in miR-204 expression leads to reduced expression of claudins 10, 16, and 19 (message/protein). They reported that a relatively high expression of miR-204 can preserve the epithelial phenotype. In the present study, we found that miR-204-5p was significantly down-regulated in human PCO samples. These results are in agreement with the results from a recently published study in a mouse model by our group. ¹⁹ We previously showed that miR-184 and miR-204 play a significant role in the control of PCO formation in mice; however, the responsible mechanisms were not delineated. Based on the significant down-regulation of hsa-miR-204-5p in human PCO, we undertook detailed mechanistic studies. We showed that hsa-miR-204-5p targets the 3′-UTR of SAMD4 and inhibits EMT in the LECs of a human donor capsular bag model in vitro. This model recapitulates many features of EMT observed during human PCO development in vivo. Our data suggest that regulation of miR-204-5p may be SMAD4-dependent because SMAD4 expression was significantly suppressed when miR-204-5p was overexpressed. Moreover, the expression of phospho-SMAD2/3 was unaffected in the presence of either miR-204 mimic or inhibitor. This result indicated that SMAD4, rather than SMAD2/3, is a specific target gene of miR-204-5p in the regulation of LEC EMT. 

We next investigated whether PCO formation could be successfully attenuated by regulating miR-204-5p in a human donor capsular bag model. Common established markers of the EMT are the loss of E-cadherin expression and the gain of α-SMA and vimentin expression. During human PCO, the LECs transdifferentiate into fibroblastic cells through an EMT process. ^29,30 These cells become myofibroblast-like, express mesenchymal markers, and exhibit a contractile phenotype contributing to the wrinkling and fibrosis of the lens capsule. The expression of α-SMA is characteristic of myofibroblasts, ^31,32 and this molecule is also one of the established markers of EMT in LECs. ^7,10 Activation of α-SMA–positive myofibroblasts is considered a key event in the progression of lens fibrosis. Significantly increased expression of α-SMA is induced initially (3 days), suggesting activation of EMT and transdifferentiation of LECs into myofibroblasts. In the present study, we observed that these EMT-associated markers were perturbed by miR-204-5p overexpression or knockdown. The results revealed that PCO-associated EMT was inhibited, since the expression of both the transdifferentiation marker α-SMA and mesenchymal marker vimentin were inhibited, and the epithelial marker E-cadherin was restored when LECs were transfected with the miR-204-5p mimic. However, when the LECs were transfected with the miR-204-5p inhibitor, changes in the expression level of these three EMT markers were reversed. Saika et al. ³³ and Sato et al. ³⁴ reported that TGF-β/SMAD3 signaling is critical in the induction of EMT in lens or renal epithelium. In addition, gene ablation of SMAD3 blocks EMT is found in mammary gland epithelial cells. ³⁵ Saika et al. found that transient adenoviral gene transfer of SMAD7 ³⁶ or other anti-SMAD genes (Id2/3) ³⁷ to the lens also prevents or attenuates EMT of LECs, and inhibition of p38MAP kinase suppresses fibrotic reaction of retinal pigment epithelial cells. ³⁸ In the present study, we found miR-204-5p–associated regulation of the EMT might prevent the activation of a TGF-β/SMAD4–dependent pathway and the development of associated secondary cataract characteristics, such as fibrosis. Because SMAD4 is a target gene of miR-204-5p and SMAD4 is critical in TGF-β-induced EMT, we tested and determined that down-regulation of SMAD4 expression also occurs by the cotransfection of miR-204-5p mimic and silenced SMAD4. The silencing of SMAD4 directly represses α-SMA and vimentin transcription and has been suggested to directly increase E-cadherin transcription. These results indicate that silencing the expression of SMAD4 could enhance the repressive effect of miR-204-5p in the TGF-β–induced EMT. 

In conclusion, we have identified a molecular mechanism involving miR-204-5p and SMAD4 in the EMT during the development of secondary cataracts. The potential therapeutic role of miRNAs in the prevention of PCO remains a major clinical challenge. The presented evidence underscores the importance of miR-204-5p as a novel target for therapeutic intervention and indicates that this miRNA merits further investigation as a promising gene therapy target for the treatment of PCO. In addition, our findings provide the impetus to study the role of miRNAs in EMT observed in other diseases, such as renal fibrosis and cancer. ^39,40  

Supplementary Materials

pdf S1 

pdf S2 

pdf S3 

pdf S4 

pdf S5 

Acknowledgments

We thank Meifang Dai (KangChen Bio-tech, Shanghai, China) for her excellent technical assistance. 

Footnotes

 Supported by the National Natural Science Foundation of China (30600698 and 30901637) and Shandong Province Natural Science Foundation (ZR2012HM009 and BS2012YY030).

Footnotes

 Disclosure: Y. Wang, None; W. Li, None; X. Zang, None; N. Chen, None; T. Liu, None; P.A. Tsonis, None; Y. Huang, None

Footnotes

⁴  These authors contributed equally to the work.