Knockdown of HSPA13 Inhibits TGFβ1-Induced Epithelial-Mesenchymal Transition of RPE by Suppressing the PI3K/Akt Signaling Pathway

Abstract

Purpose: This study aimed to explore the impact of HSPA13 on epithelial-mesenchymal transition (EMT) in retinal pigment epithelial (RPE) cells and proliferative vitreoretinopathy (PVR) development, along with its associated molecular mechanisms.

Methods: HSPA13 expression was evaluated in epiretinal membranes (ERMs) from patients with PVR using immunohistochemistry. The effects of HSPA13 knockdown on TGFβ1-induced EMT in hESC-RPE cells were studied through quantitative PCR (qPCR), Western blot, and wound healing assays. Intracellular Ca²⁺ levels were measured using Fluo-8/AM incubation. A rat PVR model was induced by the intravitreal injection of RPE cells combined with platelet-rich plasma (PRP). RNA-seq was applied to study the molecular mechanism of HSPA13 knockdown-mediated EMT inhibition.

Results: HSPA13 was found in human ERMs and its expression increased with TGFβ1 treatment in hESC-RPE cells. Knockdown of HSPA13 inhibited TGFβ1-induced EMT and migration. In the PVR rat model, HSPA13 was expressed in the ERMs and its knockdown in RPE cells reduced the development of PVR. Consistent with these observations, RNA-seq showed a global suppression of TGFβ1-induced EMT and migration by shHSPA13 in RPE cells. Mechanistically, TGFβ1 treatment increased intracellular Ca²⁺ levels, leading to an upregulation of HSPA13 expression. Downregulation of HSPA13 hindered the phosphorylation of PI3K/Akt in TGFβ1-induced RPE cells.

Conclusions: Our study revealed the involvement of HSPA13 in PVR development, as well as in TGFβ1-induced EMT of RPE through the PI3K/Akt signaling pathway. Targeting HSPA13-related pathways involved in regulating EMT in RPE cells could serve as a novel therapeutic approach for patients with PVR.

Proliferative vitreoretinopathy (PVR) is a common complication that occurs after rhegmatogenous retinal detachment (RRD) and is a major cause of surgical failure.¹ The epiretinal membranes (ERMs) could develop on the surfaces of the detached retina and the posterior hyaloid. The contraction of ERM in PVR causes further retinal detachment, which often leads to vision loss.² Surgery is the primary approach for treating PVR.³ Despite the advancements in techniques, the functional outcomes of PVR surgery have still been unsatisfactory.^4,5 Therefore, investigating the pathogenesis of PVR and developing an effective drug that prevents PVR formation are needed. 

The retinal pigment epithelial (RPE) layer, located between the photoreceptors and the choriocapillaris, plays a crucial role in maintaining retinal function. Despite being usually considered to be terminally differentiated, clinical observations suggest that RPE cells exhibit reduced differentiation and undergo epithelial-to-mesenchymal transition (EMT) in various retinal degenerative disorders.⁶ In PVR, RPE cells are identified as the primary cells of the ERMs, and the EMT in RPE cells is a main contributor to the progression of PVR.^7,8 During the process of EMT, RPE cells lose their epithelial characteristics and acquire mesenchymal properties via breakdown in cell-cell adhesion and contact, cytoskeletal rearrangement, and an increase in cell mobility.⁹ EMT is regulated by various signaling pathways, including transforming growth factor beta (TGFβ),¹⁰ ERK1/2,¹¹ Wnt/β-catenin,¹² Notch,¹³ and PI3K-Akt.¹⁴ TGFβ is highly expressed in the retinal microenvironment in PVR.¹⁵ Thus, TGFβ has a crucial role in triggering EMT in PVR. 

Heat shock proteins (HSPs) are key regulators of cell homeostasis and are engaged in various cellular processes, such as fibrosis.¹⁶ Nevertheless, the contribution of HSPs to fibrotic processes appears to be contingent upon the specific HSP involved. For example, HSP70 protects against pancreatic fibrosis¹⁷ and pulmonary fibrosis,¹⁸ while reducing HSP90 ameliorates pancreatic fibrosis¹⁹ and pulposus fibrosis.²⁰ In addition, HSPs have a significant impact on the TGFβ1 pathway.¹⁶ Given the well-documented role of the TGFβ pathway in EMT of RPE, targeting HSPs could offer a novel treatment strategy for retinal degeneration diseases caused by EMT in RPE cells, such as PVR. 

The stress 70 protein chaperone (STCH), also known as HSPA13, belongs to the HSP70 superfamily.²¹ The human HSPA13 gene encodes a 60 kDa peptide²² and its expression is triggered by Ca²⁺ stress, not heat shock.²¹ Although the stress-inducible Hsp70 has been extensively researched, there is a lack of studies on HSPA13. Several studies indicate that HSPA13 plays a role in the process of protein import.²³ In addition, the expression of HSPA13 is much higher in tumor tissues compared to nontumor tissues.²⁴ Recently, HSPA13 was reported to promote hepatocellular carcinoma (HCC) migration and invasion.²⁵ Considering that fibrogenesis and cancer share several properties, including the occurrence of the EMT process,²⁶ we aim to investigate the effect of HSPA13 on the EMT of RPE cells. 

Our study first revealed the presence of HSPA13 in ERMs of patients with PVR. Further, we observed that treatment with TGFβ1 caused an increase in intracellular Ca²⁺ levels, leading to an upregulation of HSPA13 in human embryonic stem cell derived RPE (hESC-RPE) cells and ARPE-19 cells. Suppression of HSPA13 hindered the TGFβ1-induced EMT and cell migration in vitro, and restored the retinal morphology in PVR rats. Mechanistically, knockdown of HSPA13 reduced the phosphorylation levels of PI3K/Akt induced by TGFβ1. These findings suggest that targeting HSPA13 could serve as a potential therapeutic option for PVR. 

Materials and Methods

Cell Culture

Human embryonic stem cells (hESCs) were differentiated into RPE cells using our previously described protocol.²⁷ The hESC-RPE cells were cultured onto 1% Matrigel-coated plates at a density of 1.5 × 10⁶ per well in 6 well plates in RPE medium, which consisted of Dulbecco's Modified Eagle's Medium: Nutrient Mixture F-12 (DMEM/F-12; Gibco, USA), 10% knockout serum replacement (KSR; Gibco, USA), 1% MEM nonessential amino acids (MEM NEAA; Gibco, USA), 2 mM L-glutamine (Gibco, USA), 1% penicillin-streptomycin (Gibco, USA), and 0.1 mM b-mercaptoethanol (Gibco, USA). The medium was changed every 2 to 3 days. Cells were stained with 0.4% trypan blue dye, and the cell diameter and the number of viable cells were determined with a Countstar cell counter instrument (ALIT Life Science, China). 

To induce EMT, hESC-RPE cells were seeded in 1% Matrigel-coated 12-well plates at a density of 5 × 10⁵ cells/well. Following 24 hours of cell plating, the cells underwent 24-hour KSR deprivation. Subsequently, the cells were subjected to incubation with a concentration of 10 ng/mL of recombinant human TGFβ1 for 48 hours. 

Antibodies and Reagents

Antibodies against E-cadherin (E-CAD; CST, Cat# 14472, USA), FN1 (Proteintech, Cat# 66042, China), HSPA13 (Proteintech, Cat# 12667, China), α-SMA (Abcam, Cat# ab7817, UK), Cytokeratin 8 (CK8; Abcam, Cat# ab9023, UK), PI3K p85 (CST, Cat# 4292, USA), p-PI3K p85/p55 (Beyotime, Cat# AF5905, China), AKT (CST, Cat# 4691, USA), p-AKT (CST, Cat# 13038, USA), and ZO-1 (Thermo, Cat# 617300, USA) were used for Western blot or immunostaining. The BAPTA/AM was acquired from Selleck Inc. (Houston, TX, USA) and administered to hESC-RPE cells at a dose of 10 µM. Recombinant human TGFβ1 was purchased from Sino Biological Inc. (Beijing, China). 

Immunohistochemistry

The ERMs from patients with PVR were obtained from the Department of Ophthalmology of Tongji Hospital. The ERMs were stained with 0.1% toluidine blue and subsequently fixed in 4% paraformaldehyde (PFA, Sangon Biotech, China). After dehydration treatment with a sucrose gradient, these tissues were then embedded in optimal cutting temperature compound (OCT; compound 4583; Tissue-Tek). These samples were sectioned into 10 µm thickness, and permeabilized with 0.3% Triton X-100 for 15 minutes. After washing three times with phosphate-buffered saline (PBS), these sections were blocked in 3% bovine serum albumin (BSA, Sangon Biotech, China) for 45 minutes at room temperature. Next, these sections were incubated with rabbit anti-HSPA13 antibody (diluted 1:200 in the block solution) and mouse anti-CK8 antibody (diluted 1:100 in the block solution) or anti-α-SMA antibody (diluted 1:100 in the block solution) at 4°C overnight. Following a 1-hour incubation with the corresponding secondary antibody mixture, the sections were subsequently counterstained with 4',6-diamidino-2-phenylindole (DAPI) for 2 minutes to visualize the nuclear morphology. DAPI is a fluorescent stain that binds to DNA in the nucleus. After washing three times with PBS, the coverslips were mounted on glass slides. These sections were observed using a Nikon A1R fluorescence microscope. 

For immunohistochemical examination of the eyeballs, rats were euthanized 4 weeks post-intravitreal injection. The eyeballs were then enucleated immediately and fixed in 4% PFA. The subsequent procedures are identical to those used for the immunohistochemistry of ERMs. 

Immunocytochemistry

Cells were cultured on coverslips placed in 24-well plates. After the necessary treatments, cells were rinsed once with PBS, fixed in 4% PFA for 10 minutes, and permeabilized in 0.1% Triton X-100 for 10 to 15 minutes. The coverslips were washed 3 times with PBS, blocked in PBS containing 3% BSA for 30 minutes at room temperature, and stained with primary antibodies at 4°C overnight. Next, the coverslips were washed and incubated with secondary antibodies. Nuclei were confirmed by staining with DAPI for 2 minutes. After washing three times with PBS, the coverslips were mounted on glass slides and photographed using a Nikon A1R microscope. 

RNA Extraction and Quantitative Real-Time PCR

Total RNA was extracted from cells using TRIzol reagent (Takara, Japan) according to the manufacturer's instructions. The RNA concentration was determined with a Nanodrop 2000 spectrophotometer (Thermo Fisher Scientific, USA), and cDNA was synthesized using a reverse transcription kit (Takara, Japan). The quantitative PCR (qPCR) was performed using the SYBR Green Master Mix (Tiangen, China) on a Bio-Rad Chromo4 thermal cycler instrument (Bio-Rad, USA). The gene expression levels were standardized by comparing them to the expression of the housekeeping gene GAPDH and calculated using the 2^-ΔΔCT method. The untreated control group, specifically the RPE cells or shScramble-RPE cells, serves as the relative reference. The specific primer sequences for HSPA13, and RPE/EMT-related genes are listed in Supplementary Table S1. 

Western Blot Analysis

Total proteins from cells were extracted in ice-cold RIPA buffer (Beyotime, China) supplemented with protease inhibitor cocktails (TargetMol, USA). The samples were then centrifuged at 13,200 g at 4°C for 20 minutes. The supernatant was collected and the protein concentration was determined by a BCA protein assay kit (Pierce, USA). Then, 10 to 20 µg of the protein lysate was separated by 10% SDS-PAGE and transferred onto PVDF membranes (Millipore, USA). After blocking with 5% BSA, the membranes were incubated with primary antibodies in 5% BSA overnight at 4°C, followed by incubation with the corresponding secondary antibodies for 1 hour at room temperature, and an enhanced chemiluminescence system (Pierce, USA) was used for band detection. 

Plasmid Construction and Lentivirus Infection

The short interfering oligonucleotides designed to target HSPA13 with the sequence GATCCCCGGCTGACGTCTTCCACGTC were synthesized by Sangon Biotech, China, and then subsequently inserted into the pLVXpuro vector. A scrambled sequence was used as a negative control. The lentivirus-shHSPA13 and lentivirus-shScramble were produced within 293T cells. Subsequently, hESC-RPE cells were infected with shScramble or shHSPA13 lentivirus. After 36 hours of infection, the cells underwent screening with 1 ng/mL puromycin for 1 week to generate a stable cell line. 

Live-Cell Ca²⁺ Imaging

The hESC-RPE cells were cultured in 48-well plates and incubated with TGFβ1 alone or in combination with BAPTA/AM for 48 hours. Then, the cells were loaded with the fluorescence Ca²⁺ probe Fluo-8/AM (2 µM, for 45 minutes loading in KSR-free culture medium; Beyotime, China). The Ca²⁺ imaging in hESC-RPE cells was performed within 60 minutes of dye loading using a Nikon A1R fluorescence microscope. 

RNA-Sequencing

Gene expression profiles were analyzed by high throughput RNA sequencing (BGI, Shenzhen, Guangdong, China), and further analysis was performed using the related software Dr. Tom (BGI, Shenzhen, Guangdong, China). The mRNAs with a fold change >2 and a P < 0.05 were defined as differentially expressed genes (DEGs) with statistical significance. The GSEA software was used to perform gene set enrichment analysis (GSEA). The RNA sequencing data were deposited in the NCBI's Gene Expression Omnibus database (GSE264212). 

Scratch Wound Healing Assay

The scratch wound healing assay was performed as described in our previous work.²⁷ Briefly, the hESC-RPE monolayers were scratched with a sterile pipette tip, gently washed twice with sterile PBS to remove the resulting debris, and incubated in KSR-free medium for 24 hours. The cells and wound were imaged at 0 and 24 hours, and the area of wound closure was calculated with ImageJ. 

Establishment of the PVR Rat Model

Six-week-old Sprague-Dawley (S-D) rats were obtained from SLACCAS (Shanghai, China) and housed in the Tongji University Animal Center. All animals used in this study were maintained in specific pathogen-free conditions in micro isolator cages and were treated according to the guidelines provided in the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research and the Guides for the Care and Use of Animals (National Research Council and Tongji University). The protocol was approved by the Committee on the Ethics of Animal Experiments of Tongji University (Permit Number: TJAA09620204). A total of 40 S-D rats were randomly assigned to 4 groups: (1) the PBS group, (2) the hESC-RPE + platelets rich plasm (PRP) group, (3) the shScramble-hESC-RPE + PRP group, and (4) the shHSPA13-hESC-RPE + PRP group. Each group consisted of 10 rats. The PVR rat model was established as previously reported.²⁸ Briefly, the rats were anesthetized and their pupils were dilated before vitreous injection. To induce PVR, a mixture of hESC-RPE cells (4 µl, 1 × 10⁵ cells per µl) and 4 µl PRP were injected into the vitreous of S-D rats. 

Hematoxylin and Eosin Staining

The eyeballs were promptly enucleated and fixed in 4% PFA. Following gradient dehydration, the eyes were embedded in paraffin and sectioned into continuous slices of 5 µm thickness. After deparaffinization and rehydration, the sections were stained using a Hematoxylin and Eosin Staining Kit (Beyotime, China). The stained eyeball slides were mounted with neutral balsam and images were captured with a light microscope. 

Statistical Analysis

Each experiment was repeated at least three times. Data are presented as the mean ± SEM (standard error of the mean) of at least three independent experiments. The Statistical Package GraphPad Prism version 9.0 (GraphPad Software, San Diego, CA, USA) was used to analyze the data. Statistical analysis involved a Student's t-test, and 1-way and 2-way ANOVA with Tukey's multiple comparison test as appropriate. A P value of less than 0.05 was considered to be statistically significant. 

Results

HSPA13 is Expressed in ERMs From Patients With Clinical PVR

To investigate the potential involvement of HSPA13 in the pathogenesis of PVR, we detected the expression of HSPA13 in ERMs collected from patients with PVR by immunostaining. As shown in Figure 1, the HSPA13 showed strong immunoreactivity within the ERM. Double-staining further revealed that HSPA13 co-localized with both the epithelial marker CK8 and the EMT marker α-SMA. As RPE cells are the only kind of epithelial cells seen in the ERM,²⁹ this finding suggests that HSPA13-positive cells are derived from RPE cells undergoing EMT. However, there is a variability in staining patterns among the cases, potentially attributable to a range of factors, including variations in ERMs size, tissue section orientation, and EMT degree. Actually, the cell count in the Cases 1 and 2 is significantly lower than that in the Cases 3 and 4. Despite this variability, all the results consistently reveal the co-localization of HSPA13 with the epithelial marker CK8 and the EMT marker α-SMA. 

Figure 1.

 

HSPA13 expression in ERMs from patients with clinical PVR. Representative immunofluorescence staining of HSPA13 (red), α-SMA (green), or CK8 (green) and DAPI (blue) in ERM tissue of four patients with clinical PVR. The expression of HSPA13 colocalized with α-SMA or CK8 (yellow) in ERM tissue. Scale bar = 50 µm.

TGFβ1 Enhanced the Expression of HSPA13 Depending on Intracellular Ca²⁺

The expression pattern of HSPA13 in PVR membranes indicates that HSPA13 might play a role in PVR pathogenesis. As the EMT of RPE cells plays a crucial role in the pathogenesis of PVR, we next investigated whether HSPA13 contributes to the EMT process in RPE cells. To stimulate the EMT of RPE cells in vitro, we used TGFβ1-induced EMT model. In this model, we treated hESC-RPE cells with 10 ng/mL TGFβ1 for 48 hours. The selection of this duration and concentration was determined by time course and dose curve studies (Supplementary Figs. S1, S2). As expected, TGFβ1 treatment successfully induced a morphological transformation of hESC-RPE cells from polygonal to elongated spindle-shaped cells (Fig. 2A). Consistent with the change in morphology, the expression of the epithelial and specific RPE markers E-CAD, RPE65, and BEST was downregulated (Fig. 2B), and the expression of the mesenchymal markers FN1 (Figs. 2C, 2E, 2F) and PAI-1 (Fig. 2C) were upregulated. Meanwhile, we observed an increase in both the mRNA (Fig. 2D) and protein (Figs. 2E, 2G) levels of HSPA13 in RPE cells treated with TGFβ1. A similar result was also observed in ARPE-19 cells and primary mouse RPE (mRPE) cells treated with TGFβ1 (Supplementary Fig. S3). Additionally, double-immunofluorescence labeling revealed an elevated expression of both HSPA13 and α-SMA in the TGFβ1-treated RPE cells (Fig. 2H), further confirming the upregulation of HSPA13 in cells undergoing EMT. On the contrary, high expression of HSPA13 was accompanied by low and cytoplasmic expression of E-cadherin, suggesting that cells with high HSPA13 expression experience a loss of cell contact (Supplementary Fig. S3). 

Figure 2.

 

TGFβ1 treatment upregulated HSPA13 expression in hESC-RPE cells, which is dependent on intracellular Ca²⁺. (A) Representative images of the cell morphology change after the treatment with TGFβ1 for 48 hours. Scale bar = 20 µm. (B, C, D) The qPCR detection of the expression of the RPE markers E-cadherin (E-CAD), RPE65, BEST (B), the EMT markers FN1 and PAI-1 (C), and HSPA13 (D) in hESC-RPE cells with or without TGFβ1 challenge. (E) Representative Western blots and (F, G) quantification of the expression of FN1 and HSPA13 in hESC-RPE cells with or without TGFβ1 challenge. The β-actin was used as a loading control. (H) Double-immunostaining of HSPA13 and E-cadherin in hESC-RPE cells treated with or without TGFβ1 (10 ng/mL) for 48 hours. Scale bar = 20 µm. (I) The hESC-RPE cells were treated with TGFβ1 alone or in combination with BAPTA/AM (10 µM) for 48 hours. The Fluo8/AM dye was utilized to detect intracellular Ca²⁺ signaling (green). DMSO was used as a negative control. Scale bar = 20 µm. (J) Quantification of the fluorescent intensity of the Ca²⁺ staining in I. (K) Representative Western blots and (L) quantification of the expression of HSPA13 in hESC-RPE cells treated with TGFβ1 or in combination with BAPTA/AM. The β-actin was used as a loading control. Full, uncropped Western blots are presented in Supplementary Figure S8. NC = negative control. (M) The qPCR detection of the expression of HSPA13 in hESC-RPE cells treated with TGFβ1 alone or in combination with BAPTA/AM. Data are expressed as the means ± SEM of three independent experiments. *P < 0.05. **P < 0.01. ***P < 0.001. ****P < 0.0001.

A previous study showed that treatment with TGFβ1 in megakaryocytes resulted in an increased level of intracellular Ca²⁺.³⁰ In line with this result, we found that TGFβ1 treatment also increased intracellular Ca²⁺ in hESC-RPE cells (Figs. 2I, 2J). Because HSPA13 is a Ca²⁺ stress-induced member of the HSP70 superfamily, we then investigated whether the increased HSPA13 expression induced by TGFβ1 is dependent on intracellular Ca²⁺. To this end, we used BAPTA/AM to chelate intracellular Ca²⁺ and then examined the impact of TGFβ1 on the expression of HSPA13. As shown in Figures 2K–M, chelating calcium prevented TGFβ1-induced HSPA13 upregulation in hESC-RPE cells. A similar result was also observed in ARPE-19 cells (Supplementary Fig. S2). To further confirm the role of Ca²⁺ in the upregulation of EMT and HSPA13, we conducted immunofluorescence co-staining to detect the expression of α-SMA and HSPA13 under chelated calcium conditions. Our findings revealed that RPE cells treated with BAPTA and TGFβ1 exhibited reduced expression of HSPA13 and α-SMA compared to cells treated with TGFβ1 alone (Supplementary Fig. S4). Taken together, these data demonstrate that TGFβ1 can induce HSPA13 expression in hESC-RPE cells, and such induction effect is dependent on intracellular Ca²⁺. 

Knockdown of HSPA13 Attenuated TGFβ1-Induced EMT in hESC-RPE Cells

To elucidate the function of HSPA13 in the EMT of RPE cells, we established a stable HSPA13 knockdown hESC-RPE cell line by infecting cells with lenti-shHSPA13 virus. Lenti-shScramble was used as a negative control. The mesenchymal morphology induced by TGFβ1 was suppressed in shHSPA13-RPE cells (Fig. 3A). The shHSPA13 markedly reduced the level of HSPA13 in hESC-RPE cells and weakened the expression of EMT marker genes triggered by TGFβ1, compared to the shScramble group (Fig. 3). Moreover, as anticipated, the immunofluorescence analysis targeting the specific tight junction protein ZO1 and the EMT marker α-SMA showed a marked reduction or even loss of tight junctions between cells treated with TGFβ1, concurrent with elevated α-SMA expression (Fig. 3G). Importantly, suppression of HSPA13 restored the peripheral localization of ZO1 and attenuated the upregulation of α-SMA, which aligns with the morphological improvements depicted in Figure 3A. These findings reinforce the notion that downregulation of HSPA13 can counteract TGFβ1-induced EMT in hESC-RPE cells. We also observed similar results in ARPE-19 cells (Supplementary Fig. S5). Notably, HSPA13 downregulation on its own (shHSPA13 without TGFβ1) could increase the expression of RPE-related genes (Figs. 3B, 3D, 3F), consistent with the improved epithelial morphology observed in shHSPA13-RPE cells (Fig. 3A). Given that hESC-RPE cells spontaneously undergo EMT during in vitro culture, this finding suggests that downregulating HSPA13 helps maintain an RPE identity and subsequently hinders the occurrence of EMT. 

Figure 3.

 

HSPA13 knockdown attenuated TGFβ1-induced EMT of hESC-RPE cells. (A) Representative images of the cell morphology change after the treatment with TGFβ1 for 48 hours. Scale bar = 20 µm. (B) Representative images of Western blots and quantification of the expression of HSPA13 (C), E-cadherin (D), and FN1 (E) in HSPA13-knockdown (shHSPA13) hESC-RPE cells with or without TGFβ1 challenge. Scramble (shScramble) served as a negative control. The β-actin was used as a loading control. Full, uncropped Western blots are presented in Supplementary Figure S9. (F) The expression of HSPA13, the RPE markers E-cadherin (E-CAD), RPE65, BEST, and the EMT markers FN1 and PAI-1 was detected by qPCR. (G) Immunofluorescence staining of ZO1 (purple) and α-SMA (red) in both shScramble- and shHSPA13-hESC-RPE cells, with or without TGFβ1 challenge. The RPE cells infected with either shScramble or shHSPA13 lentivirus were labeled in green. DAPI = nuclei. Scale bar = 20 µm. Data are expressed as the means ± SEM from three independent experiments. *P < 0.05. **P < 0.01. ***P < 0.001. ****P < 0.0001.

To further investigate the effect of HSPA13 on TGFβ1-induced EMT of RPE cells, we performed bulk RNA sequencing (RNA-seq). The volcano plots were used to show the DEGs between different compared groups (Figs. 4A–C). The Gene Ontology (GO) analysis of the DEGs between the shHSPA13 + TGFβ1 group and the shScramble + TGFβ1 group revealed that these changed genes were mainly involved in extracellular matrix, cell adhesion, and visual perception (Figs. 4D, 4E). Gene set enrichment analysis (GSEA) revealed that gene signatures associated with EMT were significantly downregulated in the shHSPA13 + TGFβ1 group relative to the shScramble + TGFβ1 group (Fig. 4F). The top EMT-related DEGs were presented in a clustering heatmap (Fig. 4G). 

Figure 4.

 

Bulk RNA-Seq analysis of the effect of HSPA13 knockdown on the gene expression profile in TGFβ1-induced hESC-RPE cells. (A, B, C) Pairwise volcano plot of differentially expressed genes (DEGs) between different treatments: shScramble + TGFβ1 and shScramble (A), shHSPA13 and shScramble (B), shHSPA13 + TGFβ1 and shScramble + TGFβ1 (C). (D, E) Gene Ontology (GO) analysis of the DEGs between shHSPA13 + TGFβ1 and shScramble + TGFβ1: top 10 cellular components (D) and top 10 biological processes (E). (F) Gene set enrichment analysis (GSEA) shows a global suppression of EMT-related genes in the shHSPA13 + TGFβ1 group. (G) Heatmap shows the top EMT-related DEGs that downregulated by HSPA13 knockdown (fold change ≥ 2, P < 0.05).

RPE migration has been implicated in the development of the PVR.³¹ A wound-healing assay was conducted to investigate the effect of HSPA13 on RPE cell migration. As shown in Figures 5A and 5B, shScramble-RPE cells treated with TGFβ1 exhibited greater motility compared to untreated cells. However, shHSPA13-RPE cells treated with TGFβ1 showed delayed migration when compared with TGFβ1-treated shScramble-RPE cells. Once more, the knockdown of HSPA13 resulted in the suppression of TGFβ1-induced migration of ARPE-19 cells, as demonstrated by wound healing assay and Transwell migration assay (Supplementary Fig. S6). Further, RNA-seq results revealed that HSPA13 knockdown attenuated the expression of migration-related genes (Figs. 5C, 5D). 

Figure 5.

 

Effects of HSPA13 knockdown on the migration ability of EMT cells induced by TGFβ1. (A) Representative images of the wounded area at 0 hours and 24 hours after scratching. Scale bar = 100 µm. (B) Statistical analysis of the percentages of the wound area (n = 4). Data are expressed as the means ± SEM from three independent experiments. *P < 0.05. (C) GSEA shows a global suppression of cell migration-related genes in the shHSPA13 + TGFβ1 group. (D) Heatmap shows the signature genes enriched in cell migration through GSEA analysis.

shHSPA13 Inhibits PI3K/Akt Pathway in TGFβ1-Induced RPE Cells

To investigate the mechanism by which HSPA13 knockdown inhibits the EMT of RPE cells, we further analyzed our RNA-seq data. GO enrichment analysis revealed that upregulated DEGs in the shHSPA13 + TGFβ1 group compared with the shScramble + TGFβ1 group were mainly involved in visual perception, and downregulated DEGs were focused on cell adhesion and extracellular matrix organization (Figs. 6A, 6B). The Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis indicated that the downregulated DEGs were predominantly associated with the PI3K/Akt signaling pathway (Fig. 6C). The PI3K/Akt-related DEGs were shown in the heatmap (Fig. 6D). Western blots were subsequently performed to verify the regulatory mechanisms of HSPA13 on the PI3K/Akt signaling pathway. The results showed that HSPA13 knockdown markedly downregulated the levels of phosphorylated PI3K (p-PI3K) and phosphorylated Akt (p-Akt), whereas there was no change in the expression of total PI3K and Akt (Figs. 6E–G). To determine the specificity of HSPA13 toward the PI3K/Akt pathway, we further analyzed 2 additional critical signaling mechanisms: the Smads signaling, the primary signaling induced by TGFβ, and the Wnt/β-catenin signaling, another crucial regulator of EMT. The results revealed that the downregulation of HSPA13 did not significantly influence either of these two signaling pathways (Supplementary Fig. S7). This finding implies that HSPA13 exhibits a distinct specificity in regulating EMT processes, particularly for its lack of impact on the Smads and Wnt signaling pathways. 

Figure 6.

 

HSPA13 knockdown inhibits TGFβ1-induced PI3K/Akt phosphorylation in hESC-RPE cells. (A, B) Gene Ontology (GO) analysis of the upregulated DEGs (A) and the downregulated DEGs (B) between shHSPA13 + TGFβ1 and shScramble + TGFβ1 in biological processes. (C) KEGG pathway enrichment analysis of downregulated DEGs between shHSPA13 + TGFβ1 and shScramble + TGFβ1. A second signaling pathway is the PI3K/Akt pathway. (D) Heatmap shows KEGG pathways that were significantly enriched in PI3K/Akt-associated genes. (E) Western blot assays were performed to examine the effect of knockdown of HSPA13 on the PI3K/Akt pathway activity induced by TGFβ1. The β-actin was used as a loading control. Full, uncropped Western blots are presented in Supplementary Figure S10. (F, G) Quantification of the blots in E. Data are expressed as the means ± SEM of three independent experiments. ns = not significant, *P < 0.05, ***P < 0.001.

Knockdown of HSPA13 Inhibits EMT In Vivo

To investigate the effect of HSPA13 on EMT in vivo, we used a rat model of PVR established by intravitreal injection of hESC-RPE cells and PRP. Following 4 weeks of intravitreal injection, the hESC RPE + PRP group displayed a pronounced disruption in the retinal structure, particularly in the inner and outer retinal layers (Fig. 7A). Importantly, a remarkable co-expression of α-SMA and HSPA13 was evident within the epiretinal region of this group (Fig. 7A). This finding aligns with our earlier observations of HSPA13 and α-SMA co-localization within ERMs from patients with PVR (Fig. 1), thereby further substantiating the expression of HSPA13 in the ERMs. To further validate the inhibitory effect of shHSPA13 on EMT, we separately injected shScramble hESC RPE + PRP and shHSPA13-hESC RPE + PRP into the vitreous cavity. Hematoxylin and eosin (H&E) staining revealed significant disparities between the two groups: the shScramble group exhibited dense ERMs (denoted by asterisks) and convoluted retinal tissue, whereas the shHSPA13 group presented with only mild retinal folds (Fig. 7B). Immunofluorescence staining further illuminated that the injected shScramble-hESC-RPE cells (GFP-labeled) developed distinct ERMs with strong α-SMA expression (Fig. 7C). In contrast, the shHSPA13-hESC RPE + PRP group displayed a reduced number of RPE cells in the epiretinal region, accompanied by absent or low levels of α-SMA expression. These data suggest that HSPA13 knockdown effectively inhibits EMT in vivo. 

Figure 7.

 

Effect of HSPA13 on EMT in vivo. (A) Representative images of α-SMA (red) and HSPA13 (green) immunofluorescence staining in the rat retina 4 weeks after intravitreal injection with hESC-RPE and platelets rich plasm (PRP). The white asterisk denotes the presence of epiretinal membranes (ERMs). (B) Representative images of hematoxylin and eosin (H&E) stained sections of eyeballs collected from the indicated groups at 4 weeks after intravitreal injection. The black asterisk indicated ERMs. (C) Representative images of α-SMA (red) immunoreactivity and GFP expression in rat retina tissue 4 weeks after intravitreal injection. The cell nuclei were stained with DAPI (blue). GFP was used to fluorescently label the injected cells. The white asterisk indicated ERMs. Scale bar = 50 µm.

Discussion

RPE cells have been reported to undergo EMT in various retinal degenerative disorders, such as PVR and age-related macular degeneration (AMD).⁶ However, there is still no effective pharmacological treatment for these diseases. In this study, for the first time, we demonstrated that HSPA13 was involved in TGFβ1-induced EMT of RPE cells and implicated in the pathogenesis of PVR. We found that TGFβ1 can induce HSPA13 expression in various RPE cells, including hESC-RPE cells, ARPE-19 cells, and mRPE cells. Intriguingly, this induction effect was dependent on intracellular Ca²⁺. More importantly, knockdown of HSPA13 suppressed the EMT of RPE cells in vitro and in vivo. Mechanistically, HSPA13 knockdown inhibited activation of the PI3K/Akt signaling pathway induced by TGFβ1. These findings suggest that HSPA13 may serve as a new target for PVR treatment. 

It is well known that Ca²⁺ influx into cells via plasma membrane protein channels is tightly regulated to maintain cellular homeostasis.³² Previous studies have shown that the epithelial cell marker E-cadherin and the mesenchymal cell marker vimentin are regulated by calcium.^33,34 Furthermore, emerging evidence from tumor studies indicated a strong correlation between EMT inducers and intracellular Ca²⁺ levels.^35,36 TGFβ-driven EMT of MCF7 breast cancer cells was found to be correlated with an elevated influx of Ca²⁺ into the cells.³⁵ The transient receptor potential melastatin-like 7 (TRPM7) triggers the process of EMT in breast cancer by transiently increasing levels of Ca²⁺. Blocking this channel reduces the effects of hypoxia and epidermal growth factor (EGF) on EMT.³⁶ Consistent with these studies conducted on tumor cells,^35,36 our results revealed an increased Ca²⁺ level in TGFβ1-driven EMT in hESC-RPE cells (see Figs. 2I, 2J) and ARPE-19 cells (see Supplementary Fig. S4), suggesting that the TGFβ-Calcium-EMT axis might play an important role in the development of PVR. 

HSPA13, a unique member of the HSP70 superfamily, is induced by Ca²⁺ stress rather than heat shock.²¹ To date, there is relatively little research on HSPA13. Several reports have shown that HSPA13 is highly expressed in tumors and involved in tumor proliferation, migration, and invasion.^24,25 Given that cell migration is a characteristic of EMT, the tumor migration promoted by HSPA13 implies its potential role in the progression of EMT. Here, we demonstrated, for the first time, that HSPA13 was expressed in the ERMs of patients with PVR (see Fig. 1). Besides, the expression of HSPA13 was significantly elevated in TGFβ1-induced EMT of RPE cells (see Fig. 2, Supplementary Fig. S3). More importantly, HSPA13 knockdown inhibited EMT of RPE cells in vitro (see Fig. 3, Supplementary Fig. S5) and in vivo (see Fig. 7). RPE cell migration is an important step in the development of PVR. We also found that knockdown of HSPA13 effectively inhibited TGFβ1-induced RPE cell migration (see Fig. 5, Supplementary Fig. S6). In support of these findings, our RNA-seq data revealed that genes related to EMT and cell migration were suppressed by HSPA13 knockdown (see Figs. 4, 5). Collectively, these findings indicate that HSPA13 plays an important role in the TGFβ1-induced EMT process in RPE cells and consequently, in the development of PVR. 

The KEGG enrichment analysis demonstrated a significantly correlation with the PI3K/Akt pathway (see Fig. 6C), a critical component in EMT transformation, metastasis, and survival of tumor cells.³⁷ This finding implies that the PI3K/Akt pathway participate in the HSPA13-regulated EMT process in RPE cells. Further elucidation through Western blot experiments indicated that the downregulation of HSPA13 specifically suppresses TGFβ-induced PI3K activation (see Fig. 6), without affecting Smad signaling, the canonical pathway triggered by TGFβ, or Wnt/β-catenin signaling, a crucial regulator of EMT (see Supplementary Fig. S7). 

Our study primarily concentrated on the suppressive impact of HSPA13 knockdown on TGFβ1-induced EMT. Notably, downregulation of HSPA13 alone, in the absence of TGFβ1 enhances the expression of RPE-specific markers (E-cadherin, RPE65, and BEST; see Fig. 3). This observation suggests that reduced HSPA13 levels contribute to the preservation of RPE characteristics, thereby hindering spontaneous EMT in vitro. Critically, we found no significant modulation of Smad2/3 signaling upon HSPA13 downregulation, indicating that the inhibition of TGFβ1-induced EMT through HSPA13 knockdown is independent of the canonical TGFβ pathway. This implies a general inhibitory effect of HSPA13 downregulation on EMT, irrespective of the specific triggering stimulus. 

Additionally, it is remarkable that hESC-derived RPE cells exhibit the ability to traverse the retina (see Fig. 7C). Intriguingly, these RPE cells that infiltrate the retina appear not to undergo EMT, evidenced by their lack of α-SMA expression. Notably, an increased proportion of shHSPA13-RPE cells was observed infiltrating the retina, suggesting a potential mechanism by which shHSPA13 inhibits EMT in RPE cells. This intriguing observation warrants further exploration. 

Collectively, our findings provide novel evidence that HSPA13 is involved in the EMT of RPE cells and ERMs formation in PVR. Mechanistically, stimulation with TGFβ1 treatment resulted in increased intracellular Ca²⁺ levels, leading to upregulation of HSPA13 expression, which in turn activated the PI3K/Akt pathway, facilitating EMT. Targeting HSPA13-related pathways involved in regulating EMT in RPE cells may offer a promising therapeutic strategy for patients with PVR. 

Acknowledgments

Supported by the Fundamental Research Funds for the Central Universities (22120220621), and the National Natural Science Foundation (81670867, 81372071, and 81770942). 

Statement of Ethics: This study involving human participants was approved by the Ethics Committee of Shanghai Tongji Hospital affiliated with Tongji University and complied with the Declaration of Helsinki. 

Author Contributions: Furong Gao and Mengwen Li performed the experiments, analyzed the data, and wrote the main manuscript text. Lixia Lu, Guo-tong Xu, and Wei Xu, conceived the experiments, analyzed the data, and revised the manuscript. Lilin Zhu, Juan Wang, Haibin Tian, Caixia Jin, Jingying Xu, and Qingjian Ou participated in the analysis and discussion. Jiao Li, Jie Xu, and Song Jia carried out the cell experiments. All authors reviewed the manuscript and had final approval for the submitted version. 

Disclosure: F. Gao, None; M. Li, None; L. Zhu, None; J. Li, None; J. Xu, None; S. Jia, None; Q. Ou, None; C. Jin, None; H. Tian, None; J. Wang, None; J. Xu, None; W. Xu, None; G.-T. Xu, None; L. Lu, None