METTL3 attenuates proliferative vitreoretinopathy and epithelial‐mesenchymal transition of retinal pigment epithelial cells via wnt/β‐catenin pathway

Abstract Proliferative vitreoretinopathy (PVR) is a refractory vitreoretinal fibrosis disease, and epithelial‐mesenchymal transition (EMT) of retinal pigment epithelial (RPE) cells is the key pathological mechanism of PVR. However, few studies focused on the role of METTL3, the dominating methyltransferase for m6A RNA modification in PVR pathogenesis. Immunofluorescence staining and qRT‐PCR were used to determine the expression of METTL3 in human tissues. Lentiviral transfection was used to stably overexpress and knockdown METTL3 in ARPE‐19 cells. MTT assay was employed to study the effects of METTL3 on cell proliferation. The impact of METTL3 on the EMT of ARPE‐19 cells was assessed by migratory assay, morphological observation and expression of EMT markers. Intravitreal injection of cells overexpressing METTL3 was used to assess the impact of METTL3 on the establishment of the PVR model. We found that METTL3 expression was less in human PVR membranes than in the normal RPE layers. In ARPE‐19 cells, total m6A abundance and the METTL3 expression were down‐regulated after EMT. Additionally, METTL3 overexpression inhibited cell proliferation through inducing cell cycle arrest at G0/G1 phase. Furthermore, METTL3 overexpression weakened the capacity of TGFβ1 to trigger EMT by regulating wnt/β ‐catenin pathway. Oppositely, knockdown of METTL3 facilitated proliferation and EMT of ARPE‐19 cells. In vivo, intravitreal injection of METTL3‐overexpressing cells delayed the development of PVR compared with injection of control cells. In summary, this study suggested that METTL3 is involved in the PVR process, and METTL3 overexpression inhibits the EMT of ARPE‐19 cells in vitro and suppresses the PVR process in vivo.


| INTRODUC TI ON
Proliferative vitreoretinopathy (PVR) is a vitreoretinal fibrosis disease, which causes repeated retinal detachment and eventually leads to blindness. PVR is characterized by the formation of proliferative membranes. 1 Currently, surgical removal of proliferative membranes is the main treatment of PVR, which is still unable to prevent cell proliferation. 2 The pathological response of retinal pigment epithelium (RPE) cells plays a pivotal role in the PVR process. 3 After interacting with various cytokines, RPE cells are triggered into epithelial-mesenchymal transition (EMT) process, and eventually convert into myofibroblasts, which become the dominant cells for contraction of proliferative membranes. 4 Among those cytokines, transforming growth factor-beta (TGFβ) is a pivotal growth factor known to induce EMT of RPE cells, and is present at high levels in PVR patients. 5 The EMT of RPE cells is considered as the key pathologic mechanism of PVR. 3 Upon the initiation of the EMT process, RPE cells lose their polarity as well as tight junctions, and acquire stronger ability of proliferation and migration, gradually evolving into mesenchymal cells in morphology and biological behaviour. During EMT process, epithelial markers including ZO-1 are down-regulated, while mesenchymal markers such as N-cadherin and alpha smooth muscle actin (αSMA) are up-regulated. 6 As a marker of myofibroblasts, αSMA often serves as an indicator of the occurrence of EMT. 7 Detachment of RPE cells from the basement membrane and cell migration are considered to occur in an early stage of PVR pathology, which are meditated by wnt/β-catenin pathway. As a highly conserved pathway through evolution, wnt/β-catenin pathway is reported to modulate tissue movement and participate in EMT process in several kinds of cells including RPE cells. 1,[8][9][10] Although the process of EMT has been investigated previously, further studies are needed to elucidate the underlying molecular regulatory mechanism.
It was reported that epigenetic modifications including DNA methylation and histone acetylation, could regulate the EMT of RPE cells. 11,12 However, recent studies showed that epigenetic modifications of mRNA also play a key role in modulating several biological processes, and whether it also takes a part in EMT of RPE cells remained unknown. N6-methyladenosine (m6A) is the most common epigenetic modification of mRNA, which mediates more than 80% of RNA methylation. 13,14 The m6A modification has reversible and dynamic properties, whose balance is orchestrated by three different types of protein complex including 'writer' methyltransferases, 'eraser' demethylases, and 'reader' proteins that recognize m6A-modified mRNA sites. 15 However, as the critical catalytic subunit in m6A 'writer' methyltransferases, METTL3 plays a key role in generating m6A. 16,17 In recent years, scientists have found that METTL3 plays different roles in suppressing or promoting EMT in different cancer cells, [18][19][20] which indicates that the regulatory effect of METTL3 is cell-type specific. 21,22 However, to further understand the regulatory mechanism of PVR, it is essential to clarify the role of METTL3 in RPE cells. The crosstalk between METTL3 and the TGFβ signalling pathway in cancer cells also implies a correlation between METTL3 and EMT. [23][24][25] And it is a remarkable fact that in addition to tumour metastasis, EMT also occurs during organ fibrosis, wound healing and embryonic development. 26,27 However, little is known about the role of METTL3 in these processes, and even less is known about the relationship between METTL3 and proliferative vitreoretinopathy (PVR).
In this study, we found differences in METTL3 expression between human PVR membranes and normal RPE layers, and then we investigated the effect of METTL3 overexpression on EMT of ARPE-19 cells and wnt/β-catenin pathway. Additionally, to determine the role of METTL3 in the EMT of ARPE-19 cells in vivo, we injected cells overexpressing METTL3 into rat vitreous cavities and assessed their capabilities to form PVR membranes. Our data uncovered the critical function of METTL3 in regulation of proliferative vitreoretinopathy, and the elucidation of this regulatory mechanism will provide new ideas for the treatment of PVR.

| Human tissue collection and immunofluorescence analysis
The study protocol involving human patients was approved by the The nuclei were labelled with DAPI (Thermo Fisher Scientific, Rockford, US), and images were taken under a confocal laser scanning microscope (Carl Zeiss, Oberkochen, Germany). Sections incubated in the buffer containing no primary antibodies were used as a negative control to examine the specific staining of target proteins. where S is the amount of input sample RNA in ng.

| Immunofluorescence staining for cells
Immunofluorescence staining for cells was performed as described previously. 29 The following was used as the primary antibody:

| Apoptosis, cell cycle detection and MTT assay
Apoptosis, cell cycle detection and MTT assay were performed as described previously 29

| Transwell assay
The transwell assay was conducted using a 24-well transwell plate with a polycarbonate membrane (pore size = 8 μm; Corning at room temperature for 15 min. Images were captured using a light microscope (Carl Zeiss, Oberkochen, Germany). Cell migration ability was assessed based on the cell numbers that passed through the polycarbonate membrane, which were analysed and calculated using the Image J software.

| Wound healing assay
ARPE-19 cells were seeded in 6-well plates. When the cell density reached 80%, scratches were made by a p200 pipette tip. Cells were observed and photographed under an optical microscope at 0, 24 and 48 h. The cell migration ability was reflected by the wound healing area, which was analysed using the Image J software and calculated as follows: A 0 is the area of the wound at 0 h, and A 24/48 is the area of the wound at 24/48 h.

| Western blot analysis
Western blot analysis was performed as described previously, 31 The following are used as the primary antibodies:

| PVR induction and identification in rat eyes
All animal experiments in this study have been approved by the

| Statistical analysis
Statistical analysis was carried out using IBM SPSS 21.0. All data are presented as mean ± standard deviation (SD) of at least three independent experiments unless otherwise specified. Unpaired twotailed Student's t-test was used for comparison between two groups, and one-way ANOVA followed by Bonferroni's post hoc test was applied for multiple comparisons. Mann-Whitney U test was used for ordinal data suitable for non-parametric tests. *P <.05, **P <.01, ***P <.001, ****P <.0001, #P >.05.

| METTL3 showed differential expression between normal RPE layers and PVR membranes
Immunofluorescence staining showed that METTL3 was positively expressed on PVR membranes including ERMs and SRMs, and coexisted with cytokeratin, which is a marker of RPE cells ( Figure 1A). However, it seemed that the fluorescence of METTL3 was less overlapping with αSMA, which is a marker of myofibroblasts ( Figure 1B). The specific staining of cytokeratin, αSMA and METTL3 was confirmed by the absence of the primary antibodies ( Figure S1A). Results of qRT-PCR suggested that in normal RPE layer, METTL3 expression was much higher compared with ERMs, while αSMA expression showed the opposite trend ( Figure 1C,D). Then we analysed the relationship of expression between METTL3 and αSMA in ERMs, and it appeared that the expression of METTL3 was negatively correlated with the expression of αSMA ( Figure 1E).

| m6A abundance and METTL3 expression were down-regulated after the EMT of ARPE-19 cells
We used TGFβ1 to induce EMT in ARPE-19 cells and then detected total abundance of m6A, which was found to be down-regulated after 48h of EMT induction (Figure 2A). To identify the proteins responsible for the decline in total m6A level, the expression of m6A writers (METTL3 and WTAP), eraser (ALKBH5) and reader (YTHDF2) were detected by qPCR, and the most remarkable change was observed in METTL3 ( Figure 2B, Figure S1B).
Meanwhile, METTL3 expression also declined in protein level after EMT, with the concentration of TGFβ1 at 10 ng/ml being more pronounced ( Figure 2C).

| METTL3 overexpression inhibited proliferation of ARPE-19 cells through inducing G0/ G1 arrest
The expression of METTL3 was significantly elevated at the protein level after lentiviral transfection with METTL3 overexpression plasmid ( Figure 2D). Immunofluorescence staining showed that METTL3 was primarily expressed in the cell nucleus in both the vector group and METTL3 overexpression group ( Figure 2E), and the total m6A abundance increased along with the overexpression of METTL3 ( Figure 2F).
We assessed toxic effects of METTL3 overexpression on ARPE-19 cells through apoptosis assay and observed no significant difference between the METTL3 overexpression group and the vector group ( Figure 2G,H). However, MTT assay showed that METTL3 inhibited the proliferation of ARPE-19 cells on 48h and 72h after cells inoculation ( Figure 2I). Furthermore, we detected the ratio of cells in the G0/ G1 phase was higher in the METTL3 overexpression group, while the ratio of cells in the S phase showed the opposite trend, which implied that overexpressing METTL3 induced cell cycle arrest at the G0 / G1 phase, which might be partly responsible for the suppressed proliferation by METTL3 overexpression (Figure 2J,K).

| Effects of METTL3 on wnt/β-catenin signalling pathway in EMT of ARPE-19 cells
It has been well established that wnt/β-catenin signalling pathway is a highly conserved pathway through evolution, who plays a crucial part in cell migration. We have demonstrated that METTL3 inhibited migration of ARPE-19 cells together with down-regulating the expression of MMP9, the downstream gene of wnt/β-catenin pathway.
Furthermore, we explored the role of METTL3 in the wnt/β-catenin pathway and it was indicated by western blot that the protein level of pGSK3β, β-catenin and cyclinD1 was strikingly down-regulated in METTL3 overexpression cells, even in the absence of TGFβ1 ( Figure 4A,B). In addition, SKL2001, an agonist of wnt/β-catenin signalling pathway, abolished the effect of METTL3 on EMT and cell migration, indicating that METTL3 exerted its regulating role through the wnt/β-catenin pathway ( Figure 4C,D,E,F).

| METTL3 knockdown facilitated cell proliferation and EMT of ARPE-19 cells
As METTL3 overexpression suppressed cell proliferation and TGFβ1-induced EMT, we next investigated whether knockdown of METTL3 could facilitate proliferation and EMT in ARPE-19 cells.

F I G U R E 3
Effects of METTL3 overexpression on the EMT of ARPE-19 cells in the presence or absence of TGFβ1. A, Migration ability of ARPE-19 cells after seeding with or without TGFβ1 for 12 or 24 h, as measured by the transwell assay. Scale bar: 100μm. C, Statistical analysis of the number of cells that migrated through the upper chamber. B, Wound-repair ability of ARPE-19 cells was determined by the scratch assay with or without TGFβ1 for 0, 24 and 48 h. Scare bar: 100μm. D, Statistical analysis of the rate of wound closure. E, Cell morphology of the METTL3 overexpression and vector groups seeded with or without TGFβ1 (10 ng/ml) for 48 h. F, Western blot analysis of the epithelial marker ZO-1 and mesenchymal markers, N-cadherin, αSMA and MMP9, after treatment with TGFβ1 for 48 h. G, Statistical analysis of the relative protein density, which was normalized to GAPDH and β-actin. Data represent mean ± SD (n = 3). Statistical significance was determined by the unpaired two-tailed Student's t-test for pairwise comparisons and by one-way ANOVA, followed by Bonferroni's post hoc test for multiple comparisons (*P <.05, **P <.01, ***P <.001, ****P <.0001) We down-regulated the expression of METTL3 effectively with 2 separate, specific shRNAs ( Figure 6A). Apoptosis assay showed no significance between vector group and the other two groups transfected with METTL3 shRNAs, indicating that the down-regulation of METTL3 did not induce cell toxicity ( Figure 6B, 6C). METTL3 knockdown significantly accelerated the proliferation of ARPE-19 cells in 24h, 48h and 72h after cells were seeded in the 96-well plate ( Figure 6F). Meanwhile, we found in cell cycle assay that knockdown F I G U R E 4 Effects of METTL3 overexpression on wnt/β-catenin pathway and effects of activation of wnt/β-catenin pathway on EMT process of cells overexpressing METTL3. A, Western blot analysis of the proteins of wnt/β-catenin pathway, including pGSK3β, GSK3β, β-catenin and cyclinD1, with or without treatment with TGFβ1 for 48 h B, Statistical analysis of the relative protein density, which was normalized to GAPDH. C, Western blot analysis of the epithelial marker ZO-1 and mesenchymal marker N-cadherin after treatment with TGFβ1 for 48 h. D, Statistical analysis of the relative protein density, which was normalized to GAPDH. F, Migration ability of ARPE-19 cells after seeding with or without TGFβ1 for 24 h, as measured by the transwell assay. Scale bar: 100μm. E, Statistical analysis of the number of cells that migrated through the upper chamber. Statistical significance was determined by one-way ANOVA, followed by Bonferroni's post hoc test for multiple comparisons (*P <.05, **P <.01, ****P <.

CO N FLI C T S O F I NTE R E S T
The authors confirm that there are no conflicts of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data supporting the findings of this study are available from the corresponding author upon reasonable request. H, Statistical analysis of the relative protein density, which was normalized to GAPDH. Data represent mean ± SD (n = 3). Statistical significance was determined by one-way ANOVA, followed by Bonferroni's post hoc test for multiple comparisons (*P <.05, **P <.01, ***P <.001, ****P <.0001, #P >.05)