SMAD specific E3 ubiquitin protein ligase 1 accelerates diabetic macular edema progression by WNT inhibitory factor 1

Abstract

BACKGROUND

Diabetic macular edema (DME) is the most common cause of vision loss in people with diabetes. Tight junction disruption of the retinal pigment epithelium (RPE) cells has been reported to induce DME development. SMAD-specific E3 ubiquitin protein ligase (SMURF) 1 was associated with the tight junctions of cells. However, the mechanism of SMURF1 in the DME process remains unclear.

AIM

To investigate the role of SMURF1 in RPE cell tight junction during DME.

METHODS

ARPE-19 cells treated with high glucose (HG) and desferrioxamine mesylate (DFX) for establishment of the DME cell model. DME mice models were constructed by streptozotocin induction. The trans-epithelial electrical resistance and permeability of RPE cells were analyzed. The expressions of tight junction-related and autophagy-related proteins were determined. The interaction between insulin like growth factor 2 mRNA binding protein 2 (IGF2BP2) and SMURF1 mRNA was verified by RNA immunoprecipitation (RIP). SMURF1 N6-methyladenosine (m6A) level was detected by methylated RIP.

RESULTS

SMURF1 and vascular endothelial growth factor (VEGF) were upregulated in DME. SMURF1 knockdown reduced HG/DFX-induced autophagy, which protected RPE cell tight junctions and ameliorated retinal damage in DME mice. SMURF1 activated the Wnt/β-catenin-VEGF signaling pathway by promoting WNT inhibitory factor (WIF) 1 ubiquitination and degradation. IGF2BP2 upregulated SMURF1 expression in an m6A modification-dependent manner.

CONCLUSION

M6A-modified SMURF1 promoted WIF1 ubiquitination and degradation, which activated autophagy to inhibit RPE cell tight junctions, ultimately promoting DME progression.

Key Words: Diabetic macular edema; Retinal pigment epithelium cells; Autophagy; SMAD specific E3 ubiquitin protein ligase 1; WNT inhibitory factor 1; N6-methyladenosine modification; Vascular endothelial growth factor signaling pathway

Core Tip: In this study, we demonstrated that insulin like growth factor 2 mRNA binding protein 2 upregulates SMAD-specific E3 ubiquitin protein ligase (SMURF) 1 expression in an N6-methyladenosine-dependent manner, and subsequently SMURF1 activates the Wnt/β-catenin-vascular endothelial growth factor signaling pathway by promoting ubiquitination and degradation of WNT inhibitory factor 1 to promote autophagy, which ultimately disrupts the tight junctions of the retinal pigment epithelium cells and thus led to diabetic macular edema.

INTRODUCTION

Diabetic macular edema (DME) is a common complication of diabetes mellitus and a leading cause of vision loss in people with diabetes[1]. It is reported that approximately 6.81% of people with diabetes mellitus are affected by DME[2]. Anti-vascular endothelial growth factor (VEGF) therapy and laser photocoagulation are currently the main treatment options for DME[3]. Nevertheless, non-response, resistance and relapse to anti-VEGF therapy still occur in some patients[4]. Consequently, it is crucial to explore new therapeutic targets for DME.

Retinal pigment epithelium (RPE), composed of RPE cells, is an important component of the blood-retinal barrier (BRB), where RPE cell tight junctions disruption leads to BRB dysfunction, which triggers DME[5]. For example, upregulation of apelin-13 expression disrupts retinal microvascular endothelial cell tight junctions thereby promoting DME progression[6]. Adenylate cyclase activating polypeptide 1 inhibits DME by promoting tight junctions of human RPE cells[7]. Autophagy inhibits cellular tight junctions thereby disrupting the blood-brain barrier and intestinal barrier[8-10]. It has also been reported that monounsaturated oleic acid (OA) promotes autophagy in human RPE cells thereby inducing DME[11]. However, whether autophagy affects the tight junctions of RPE cells is unclear. Therefore, exploring the mechanism of autophagy in RPE cells is essential to improve the treatment of DME.

SMAD-specific E3 ubiquitin protein ligase (SMURF) 1 mediates a variety of biological processes involving bone formation, embryonic development, and cell growth[12]. SMURF1 is associated with activation of autophagy. SMURF1 promotes autophagy by activating NFE2-like bZIP transcription factor 2 (Nrf2)[13]. SMURF1 promoted autophagy by controlling transcription factor EB (TFEB) nuclear entry to activate lysosomal biogenesis[14]. It has also been reported that SMURF1 is associated with cell tight junctions. SMURF1-mediated ubiquitination of TNF receptor associated factor 4 (TRAF4) inhibited the tight junctions of epithelial cells[15]. SMURF1 targets the ubiquitination of p120-catenin leading to dissociation of the adhesion complex[16]. Activation of autophagy disrupts tight junctions in capillary endothelial cells and downregulates tight junction proteins expression in brain endothelial cells[8,9]. It was worthwhile to explore whether SMURF1 inhibits intercellular tight junctions by activating autophagy. More importantly, SMURF1 promotes the development of diabetes and its complications. For instance, SMURF1 induced diabetes by inhibiting β-cell proliferation through mediating signal transducer and activator of transcription 3 ubiquitination[17]. SMURF1 promotes G protein-coupled bile acid receptor 1 degradation thereby contributing to diabetic nephropathy[18]. SMURF1 accelerates renal fibrosis in diabetic mice by inhibiting activation of the Nrf2/antioxidant response element pathway[19]. SMURF1 has been shown to exacerbate retinal damage from oxidative stress, thereby inducing the development of dry age-related macular degeneration[20]. This suggests that SMURF1 is a viable target for the treatment of retinal ROS injury. However, the role of SMURF1 in DME is unclear.

The molecular mechanisms of SMURF1 downstream regulation were further explored. SMURF1 has been reported to modulate disease progression by enhancing ubiquitination and degradation of target proteins[21]. Wnt inhibitory factor 1 (WIF1) is a secreted antagonist that binds to Wnt proteins to prevent the Wnt/β-catenin signaling pathway[22]. WIF1 is negatively correlated with autophagy in lung cancer[23]. Analogously, WIF1 preserves photoreceptor cells in diabetic retinopathy by blocking the Wnt/β-catenin pathway[24]. Additionally, WIF1 attenuates angiogenesis and improves visual function in retinopathy mice by suppressing the Wnt/β-catenin-VEGF signaling pathway[25]. It has even been suggested that VEGF activates cellular autophagy[26-28]. Therefore, we hypothesized that WIF1 may inhibit autophagy by suppressing VEGF. Ubibrowser bioinformatics analysis predicted that WIF1 is a ubiquitination substrate for SMURF1. It is worthwhile to investigate whether SMURF1 regulates autophagy in DME by mediating the ubiquitination of WIF1. The N6-methyladenosine (m6A) modification of RNA, mediated by m6A methyltransferases, demethylases and m6A binding proteins, plays an important regulatory role under physiological and pathological conditions[29]. Insulin like growth factor 2 mRNA binding protein 2 (IGF2BP2) is an RNA binding protein involved in the modulation of mRNA location, translation and stability through m6A modification[30]. IGF2BP2 induces pyroptosis of RPE cells by stabilizing HOXD cluster antisense RNA 1 (HAGLR) through the m6A modification to accelerate the pathological process of diabetic retinopathy[31]. We used the Starbase database prediction to show that there is a binding relationship between IGF2BP2 and SMURF1. In addition, we identified the presence of m6A modified sites on SMURF1 mRNA by SRAMP database prediction. However, the connection between IGF2BP2 and SMURF1 remains obscure.

In the present study, we hypothesized that IGF2BP2 upregulated SMURF1 expression in an m6A-dependent manner, followed by upregulated SMURF1 promoting ubiquitination and degradation of WIF1. In addition, the degradation of WIF1 promoted autophagy through activation of the Wnt/β-catenin-VEGF signaling pathway, which disrupted RPE cell tight junctions and ultimately contributed to the development of DME. Our study revealed the molecular mechanism by which the IGF2BP2/SMURF1/WIF1 axis promotes autophagy in RPE cells via the Wnt/β-catenin-VEGF signaling pathway, providing a theoretical basis and novel target for the treatment of DME.

MATERIALS AND METHODS

Clinical sample collection

We recruited 15 DME patients, 12 type 2 diabetes mellitus (T2DM) patients, and 20 idiopathic macular hole (IMH) patients treated at Zhujiang Hospital, Southern Medical University Outpatient Clinic. All patients were between 30 and 80 years of age. They underwent a detailed physical and ophthalmologic examination. Inclusion criteria for patients with DME were: Meeting diagnostic criteria for DME, central macular thickness > 300 μm with significant visual loss (best corrected visual acuity < 0.8), and no previous history of intravitreal drug injections, vitrectomy, or laser treatment. Exclusion criteria for DME were glaucoma, uveitis or ocular trauma, infection, gestational diabetes, acute complications of diabetes (e.g., diabetic ketoacidosis), liver disease or renal failure, and other chronic conditions. Inclusion criteria for T2DM were: Fulfillment of diagnostic criteria for T2DM, such as fasting blood glucose concentration > 7 mmol/L and concomitant symptoms of polyuria, thirst and weight loss. Patients with T2DM were excluded from retinopathy and other chronic diseases such as hypertension, chronic liver disease and chronic kidney disease. Patients with self-occurring macular hole without significant primary pathology in the eye, such as ocular trauma, refractive error, and other retinal pathologies, were included as IMH criteria. The exclusion criteria for IMH patients were: Hypertension, endocrine and metabolic diseases. All patients included had no previous history of medication.

The aqueous humor (50-100 μL per patient) was collected by clear corneal puncture, then centrifuged to remove cellular debris, followed by RNA extraction. The study was approved by the Ethics Committee of Zhujiang Hospital, Southern Medical University (No. 2024-KY-274-01) and all patients signed an informed consent form.

Cell culture

Human RPE cells (ARPE-19) were obtained from the National Collection of Authenticated Cell Cultures (Shanghai, China). Cells were cultured in DMEM/F-12 (1:1) medium (Gibco; California, United States) containing 100 U/mL penicillin/streptomycin (Gibco) and 10% fetal bovine serum (Gibco) and incubated at 37°C and 5% CO₂. For normal concentration glucose (NG) group and high glucose (HG) groups, when cell density was approximately 80%, cells were cultured in serum-free DMEM/F-12 medium for 24 hours. The cells were cultured with medium containing NG (5.5 mmol/L D-glucose; Sigma; Missouri, United States) or HG (30 mmol/L D-glucose) for 5 days. For the HG+ desferrioxamine mesylate (DFX) group, after HG treatment, the cells were treated with 0.1 mmol/L DFX (Sigma) for 24 hours. All cells were characterized by short tandem repeat and certified as uncontaminated by mycoplasma testing before use.

Cell transfection

Cells (1 × 10⁵) were inoculated in six-well plates (Corning; NY, United States) and cultured. Transfection was performed at 50% cell density. The plasmids sh-NC, sh-WIF1, sh-IGF2BP2, and HA-ub were synthesized by Sangon Biotech (Shanghai, China). Opti-MEM medium (Gibco) was used to mix the plasmid and Lipofectamine 3000 (Invitrogen; CA, United States) for 15 minutes. The mixture was subsequently placed in the medium and cultured for 48 hours. To obtain the SMURF1 knockdown cell line, the cells were incubated with SMURF1-knockdown adenovirus (5 × 10⁷ pfu titer; Kingsley; Nanjing, China) and culture medium for 8 hours, and then replaced with fresh medium to incubate the cells. Cells were screened using puromycin (3 μg/mL; MedChemExpress; NJ, United States) until a stable knockdown SMURF1 cell line was successfully constructed.

Immunofluorescence

Cells were fixed with 4% paraformaldehyde (Beyotime; Shanghai, China), permeabilized with 0.5% Triton X-100 (Beyotime), and then blocked with 2% bovine serum albumin (Sigma). Cells were incubated with anti-LC3 antibody (1:2000; ab51520; Abcam) overnight at 4 °C, followed by 45 minutes incubation with secondary antibody (1:500; ab96885; Abcam). Cell nuclei were stained with DAPI (Sigma), followed by sealing, and then observed and photographed under a fluorescence microscope (Leica; Vizsla, Germany).

Trans-epithelial electrical resistance detection

Cells were inoculated into a 0.4 μm pore size Transwell insert (Corning) and cultured until the cells grew to a monolayer. Electrical resistance values were measured using a millicel electrical resistance system (Millipore; MA, United States). The measurement was repeated three times. There were no cells in the blank group. The trans-epithelial electrical resistance (TEER) value was calculated as follows: (experimental group resistance value - blank group resistance value)/membrane area of the insert.

Detection of cell permeability

Transwell inserts containing 1% FBS medium were utilized to culture cells and a single cell layer was formed after 14 days. As described previously, Fluorescein isothiocyanate-dextran (MedChemExpress) was added to the insert and incubated for 2 hours to assess cell permeability[32]. The medium at the bottom was collected and the sample was added to a 96-well plate (Corning). Finally, the absorbance at 485 nm and 528 nm was measured in an enzyme-labeled instrument (Thermo Fisher; MA, United States).

Construction of the streptozotocin mouse model

C57BL/6J mice (aged 6-7 week-old, weighting 20-23 g, Cyagen; Suzhou, Jiangsu, China) were injected intraperitoneally with 55 mg/kg of streptozotocin (STZ; Merck; Darmstadt, Germany) daily for 5 days. Meanwhile, the control group was injected with citrate buffer. After the last injection of STZ, mice were fasted for 6 hours, followed by obtaining blood samples from the tails of mice, and then blood glucose concentration using a blood glucose level test kit (D799406; Sangon Biotech; Shanghai, China) once a week for four weeks. As described in the previous study, the mice with blood glucose concentrations above 250 mg/dL was considered to be diabetic mice as reported previously[33]. We also examined retinal vascular permeability in mice with Evans Blue. Briefly, 20 mg/mL of Evans Blue (Sigma) was injected into mice tail vein, and the retinas were gently removed after 10 minutes. Tissues were fixed in 4% paraformaldehyde. Retinal blood vessels were then observed and photographed using a laser scanning confocal microscope (Leica). The excitation wavelength was 546 nm. To obtain SMURF1-knockdown mice, vectors or SMURF1-knockdown adenovirus (5 × 10⁷ pfu titer; Kingsley) were injected into the eyes of mice using a Hamilton syringe (MPI Corporation; Taiwan, China). The subsequent construction of the STZ mouse model was performed after 48 hours. The study was approved by the Ethics Committee of Zhujiang Hospital, Southern Medical University (No. LAEC-2024-114).

Co-immunoprecipitation

Cells were collected and lysed. A portion of the cell lysate (100 μL) was used as Input, and the remainder (400 μL) was incubated with protein A/G agarose beads (Thermo Fisher) and anti-IgG antibody (ab172730; Abcam) or anti-SMURF1 antibody (WH0057154M1-100UG; Sigma) or anti-WIF1 antibody (YA2664; MedChemExpress) overnight. It was followed by washing the magnetic beads with PBS buffer and then detecting the levels of the indicated proteins. The IgG group was used as a negative control.

RNA immunoprecipitation

The binding relationship between IGF2BP2 and SMURF1 mRNA was examined using a RNA immunoprecipitation (RIP) kit (KT102-01; Saicheng Bio; Guangzhou, China). Specifically, cells were lysed, then cell lysates were incubated with IGF2BP2 antibody (03-251; Sigma)-magnetic beads or IgG antibody-magnetic bead complexes, followed by RNA purification, and finally SMURF1 mRNA levels were detected. RNA concentration and purity were determined by a NanoDrop spectrophotometer (Thermo Fisher), and the OD₂₆₀/OD₂₈₀ range was 1.8-2.0.

Methylated RIP

SMURF1 m6A levels were measured using the Methylated RIP assay kit (17-10499; Sigma). Specifically, the cells were lysed and then the RNAs were fragmented. The anti-m6A antibody was then incubated with the RNA fragments, followed by immunoprecipitation of the antibody-RNA complexes using magnetic beads. Finally, the m6A enrichment was assayed by quantitative real-time PCR (qRT-PCR) method.

qRT-PCR

Cells were collected, and RNA was purified using an RNA extraction kit (DP419; TIANGEN; Beijing, China), followed by reverse transcription to cDNA utilizing a reverse transcription kit (K1691; Thermo Fisher). The levels of the target gene mRNAs were then detected using a SYBR Green Realtime PCR kit (QPK-201T; TOYOBO; Osaka, Japan). GAPDH was adopted as the standard internal reference. The primers involved in this method were presented in Table 1.

Table 1 The primer sequences of quantitative real-time PCR in study.

Gene	Forward	Reverse
h-SMURF1	TGAGGAGTCTTACCGCCAGA	GACAAGTGGTCGGGGTTGAT
h-VEGFA	TCAAACCTCACCAAGGCCAG	GCTCCAGGGCATTAGACAGC
h-IGF2BP2	TGGAAGCGCATATCAGAGTG	AGTGCCCGATAATTCTGACG
h-GAPDH	CTGACTTCAACAGCGACACC	GTGGTCCAGGGGTCTTACTC
m-SMURF1	GCTGAAGCCCTTTGACCAGA	GCGCCTGTAGAGCCTTGC
m-GAPDH	AGCCCAAGATGCCCTTCAGT	CCGTGTTCCTACCCCCAATG

SMURF1: SMAD-specific E3 ubiquitin protein ligase 1; VEGFA: Vascular endothelial growth factor A; IGF2BP2: Insulin like growth factor 2 mRNA binding protein 2.

Western blot

Cells were collected and lysed, and then protein concentrations were measured with a BCA kit (Beyotime). Subsequently, the proteins (50 μg per group) were separated on SDS-PAGE and transferred to a PVDF membrane (Merck). The membrane was then blocked with 5% skimmed milk and followed by incubation with primary and secondary antibodies. Finally, the bands were visualized using an ECL luminescent solution (Beyotime), followed by observation and photography in a chemiluminescent system (Bio-rad; CA United States). The antibodies involved in this method were as follows: Anti-SMURF1 antibody (1:2000 dilution; PA5-11943; Invitrogen), anti-WIF1 antibody (1:1000 dilution; MA5-35666; Invitrogen), anti-IGF2BP2 antibody (1:1000 dilution; ab128175; Abcam), anti-occluding antibody (1:1000 dilution; ab216327; Abcam), anti-ZO-1 antibody (1:1000 dilution; 5406; Cell Signaling Technology; MA, United States), anti-claudin-1 antibody (1:1000 dilution; 4933; Cell Signaling Technology), anti-Beclin-1 antibody (1:1000 dilution; HY-P80568MedChemExpress), anti-LC3II/LC3I antibody (1:500 dilution; ABC929; Sigma), anti-p62 (dilution; P0067; Sigma) antibody, anti-Wnt1 antibody (1:1000 dilution; MABD168; Sigma), anti-VEGF antibody (1:1000 dilution; ab214424; Abcam), anti-HA antibody (1:1000 dilution; 3724; Cell Signaling Technology), anti-β-catenin antibody (1:1000 dilution; PLA0230; Sigma), Goat anti-Rabbit IgG antibody (1:5000 dilution; A32732; Invitrogen), and Goat anti-Mouse IgG antibody(1:5000 dilution; A32723; Invitrogen).

Statistical analysis

The experiment was repeated at least three times. Data were expressed as mean ± SD and analyzed with SPSS 20.0 software (IBM; NY, United States). Differences between the two groups were compared using student's t-test and differences between multiple groups were compared using one-way ANOVA after Tukey's test. The normality of data was assessed by the Shapiro-Wilk test, and there were no significant deviations from the normality of all data (P > 0.05). P < 0.05 represented statistical significance.

RESULTS

Expression of SMURF1 and VEGFA was upregulated in DME

SMURF1 induces dry age-related macular degeneration by exacerbating oxidative stress[20]. To investigate the role of SMURF1 in DME, we collected aqueous humor from patients with DME, T2DM, or IMH, respectively. SMURF1 expression was upregulated in T2DM and more significantly in DME compared with IMH (Figure 1A). VEGFA, a member of the VEGF family, plays a key role in DME development by promoting angiogenesis[34,35]. Expression of VEGFA was increased in T2DM and further increased in DME in comparison with the IMH group (Figure 1A). There was a positive correlation between the expression of SMURF1 and VEGFA in DME patients (Figure 1B). It has been reported that DFX induces osmotic membrane damage and that HG/DFX treatment of cells can mimic the DME environment of RPE cells in vitro[36]. HG treatment upregulated SMURF1 and VEGFA expression in RPE cells compared to the NG group, and HG in combination with DFX treatment accelerated this effect (Figure 1C and D). The above results indicated that the expression of SMURF1 and VEGFA were correlated with DME.

Figure 1 Expression of SMAD-specific E3 ubiquitin protein ligase 1 and vascular endothelial growth factor A was upregulated in diabetic macular edema. A: The aqueous humor was collected from diabetic macular edema patients (n = 15), type 2 diabetes mellitus patients (n = 12), and idiopathic macular hole patients (n = 20), SMAD-specific E3 ubiquitin protein ligase (SMURF) 1 and vascular endothelial growth factor A (VEGFA) expression was detected by quantitative real-time PCR (qRT-PCR); B: Correlation analysis between SMURF1 expression and VEGFA expression; C: ARPE-19 cells were treated with normal concentration of glucose (NG), high concentration of glucose (HG) or desferrioxamine mesylate (DFX), and the mRNA levels of SMURF1 and VEGFA were then detected by qRT-PCR; D: NG, HG or DFX treated ARPE-19 cells were collected, and the protein levels of SMURF1 and VEGFA were detected by western blot and the gray scale of the bands was quantified. The experiments were repeated three times. ^aP < 0.05, ^bP < 0.01, ^cP < 0.001. VEGFA: Vascular endothelial growth factor A; SMURF 1: SMAD-specific E3 ubiquitin protein ligase 1; IMH: Idiopathic macular hole; HG: High glucose; DFX: Desferrioxamine mesylate.

SMURF1 knockdown promoted RPE cell tight junctions and inhibited autophagy in DME

Tight junctions between RPE cells are essential for maintaining retinal function; however, autophagy and disruption of RPE tight junction led to abnormal BRB function and induce macular degeneration[37,38]. To evaluate the function of SMURF1 in the tight junctions and autophagy of RPE cells, SMURF1 was knocked down in RPE cells (Figure 2A and B). It was shown that HG treatment reduced the TEER of RPE cells compared to the NG group, and HG combined with DFX treatment further reduced the TEER of the cells, while SMURF1 knockdown attenuated these effects (Figure 2C). In addition, HG treatment increased the permeability of RPE cells which was further increased by HG combined with DFX treatment, whereas SMURF1 knockdown partly abrogated these effects (Figure 2D). The tight junction proteins and autophagy-related proteins expression in RPE cells was then examined. It was found that HG treatment resulted in reduced expression of occluding, ZO-1 and claudin-1, which was further reduced by HG combined with DFX treatment, and these reduced expressions of occluding, ZO-1 and claudin-1 were reversed by SMURF1 knockdown (Figure 2E). Furthermore, HG treatment increased Beclin-1 and LC3II/I expression while decreasing p62 expression compared to the NG group, and HG combined with DFX treatment accelerated these effects, which were attenuated by SMURF1 knockdown (Figure 2F and G). These results suggested that SMURF1 knockdown promoted RPE cell tight junctions and inhibited autophagy in DME.

Figure 2 SMAD-specific E3 ubiquitin protein ligase 1 knockdown promoted retinal pigment epithelium cell tight junctions by inhibiting autophagy in diabetic macular edema. A: The SMAD-specific E3 ubiquitin protein ligase (SMURF) 1 knockdown construct or vector were transfected in ARPE-19 cells, and the mRNA level of SMURF1 was then detected by quantitative real-time PCR; B: ARPE-19 cells transfected with the vector or SMURF1 construct were collected, followed by western blot to detect SMURF1 protein levels; C: Vector or SMURF1 knockdown ARPE-19 cells were treated with normal concentration of glucose (NG), high concentration of glucose (HG) or desferrioxamine mesylate (DFX), followed by determination of cellular trans-epithelial electrical resistance; D: Vector or SMURF1 knockdown ARPE-19 cells were treated with NG, HG or DFX, and the detection of cell permeability was performed; E: Vector or SMURF1 knockdown ARPE-19 cells were treated with NG, HG or DFX, and then the expression of occluding, ZO-1, and claudin-1 was detected by western blot; F: Vector or SMURF1 knockdown ARPE-19 cells were treated with NG, HG or DFX, and the expression of Beclin-1, LC3I, LC3II, and p62 was then detected by western blot; G: Vector or SMURF1 knockdown ARPE-19 cells were treated with NG, HG or DFX, followed by immunofluorescence analysis of LC3. The experiments were repeated three times. ^aP < 0.05, ^bP < 0.01, ^cP < 0.001. SMURF1: SMAD-specific E3 ubiquitin protein ligase 1; NG: Normal concentration of glucose; HG: High concentration of glucose; DFX: Desferrioxamine mesylate.

SMURF1 activated the Wnt/β-catenin-VEGF signaling pathway by regulating ubiquitination and degradation of WIF1

SMURF1 was reported to function as a ubiquitin ligase that mediated the ubiquitination and degradation of target proteins[39]. Ubibrowser bioinformatics analysis predicted WIF1 as a potential ubiquitination substrate for SMURF1, and then the interaction between SMURF1 and WIF1 was confirmed (Figure 3A). In addition, SMURF1 knockdown decreased the ubiquitination of WIF1 while increasing the protein level of WIF1 (Figure 3B). Moreover, SMURF1 knockdown significantly inhibited the degradation of WIF1 protein (Figure 3C). Meanwhile, HG treatment downregulated WIF1 expression compared to the NG group, which was further downregulated by HG combined with DFX treatment, whereas it was reversed by SMURF1 knockdown (Figure 3D). The Wnt/β-catenin-VEGF pathway has been reported to be highly dysregulated in retinal vascular disease[40]. Nevertheless, the regulatory mechanism of this pathway in DME remains to be elucidated. It was found that HG treatment increased the expression of the Wnt/β-catenin-VEGF signaling pathway proteins, including Wnt1, β-catenin, and VEGF and that HG combined with DFX treatment accelerated these effects, whereas SMURF1 knockdown abolished them (Figure 3E). Therefore, the results showed that SMURF1 activated the Wnt/β-catenin-VEGF signaling pathway via promoting WIF1 ubiquitination and degradation.

Figure 3 SMAD-specific E3 ubiquitin protein ligase 1-mediated ubiquitination degradation of WNT inhibitory factor 1 activated the Wnt/β-catenin-vascular endothelial growth factor signaling pathway. A: ARPE-19 cells were collected, followed by the interaction between SMAD-specific E3 ubiquitin protein ligase (SMURF) 1 and WNT inhibitory factor 1 (WIF1) was verified by the Co-immunoprecipitation (Co-IP) method; B: The SMURF1 knockdown construct or vector was transfected into ARPE-19 cells, followed by the Co-IP method for detecting the ubiquitination level of WIF1; C: The SMURF1 knockdown construct or vector was transfected into ARPE-19 cells, and then cells were treated with Cycloheximide (50 ng/mL) for 0, 2, 4, and 8 hours, followed by the protein level detection of WIF1 by western blot; D: The SMURF1 knockdown construct or vector was transfected into ARPE-19 cells, and cells were then treated with normal concentration of glucose (NG), high concentration of glucose (HG) or desferrioxamine mesylate (DFX), followed by western blot for the detection of WIF1 expression; E: The SMURF1 knockdown construct or vector was transfected into ARPE-19 cells, and cells were then treated with NG, HG, or DFX, followed by western blot for the detection of Wnt1, β-tcatenin, and vascular endothelial growth factor. The experiments were repeated three times. ^aP < 0.05, ^bP < 0.01, ^cP < 0.001. SMURF1: SMAD-specific E3 ubiquitin protein ligase 1; WIF1: WNT inhibitory factor 1; NG: Normal concentration of glucose; HG: High concentration of glucose; DFX: Desferrioxamine mesylate.

SMURF1 regulated HG/DFX-induced tight junction inhibition and autophagy in RPE cells via WIF1

It was further investigated whether SMURF1-mediated WIF1 expression regulates tight junctions and autophagy in RPE cells. Knockdown of SMURF1 was found to increaseWIF1 protein expression in RPE cells under HG/DFX induction, which was reversed by knockdown of WIF1 (Figure 4A). Additionally, WIF1 knockdown decreased the TEER and promoted the RPE cell permeability, which reversed the effect of SMURF1 knockdown (Figure 4B and C). Moreover, occluding, ZO-1, and Claudin-1 expression was down-regulated after WIF1 knockdown, which weakened the role of SMURF1 knockdown (Figure 4D). Consistently, WIF1 knockdown promoted the expression of Beclin-1 and LC3II/LC3I and reduced p62 expression in RPE cells, which reversed the effect of SMURF1 knockdown (Figure 4E and F). Hence, these results suggested that SMURF1 promoted autophagy by downregulating WIF1 expression thereby disrupting the tight junctions of RPE cells.

Figure 4 SMAD-specific E3 ubiquitin protein ligase 1 regulated high concentration of glucose/desferrioxamine mesylate-induced tight junction inhibition and autophagy in retinal pigment epithelium cells via WNT inhibitory factor 1. A: ARPE-19 cells were transfected with the SMAD-specific E3 ubiquitin protein ligase (SMURF) 1 or WNT inhibitory factor (WIF) 1 knockdown construct or vector, followed by treatment of the cells with high concentration of glucose or desferrioxamine mesylate, and then the protein level of WIF1 was detected by western blot; B: ARPE-19 cells were transfected with the SMURF1 or WIF1 knockdown construct or vector, followed by the measurement of cellular trans-epithelial electrical resistance; C: ARPE-19 cells were transfected with the SMURF1 or WIF1 knockdown construct or vector, followed by the detection of cell permeability; D: SMURF1, WIF1 or vector knockdown ARPE-19 cells were collected, and the expression of occluding, ZO-1 and claudin-1 was detected by western blot; E: SMURF1, WIF1 or vector knockdown ARPE-19 cells were collected, and the expression of Beclin-1, LC3I, LC3II and p62 was detected by western blot; F: ARPE-19 cells were transfected with the SMURF1 or WIF1 knockdown construct or vector, followed by the immunofluorescence analysis of LC3. The experiments were repeated three times. ^aP < 0.05, ^bP < 0.01, ^cP < 0.001. SMURF1: SMAD-specific E3 ubiquitin protein ligase 1; WIF1: WNT inhibitory factor 1; HG: High concentration of glucose; DFX: Desferrioxamine mesylate.

IGF2BP2 upregulated SMURF1 expression in an m6A modification-dependent manner

Further studies were conducted on the upstream regulatory mechanism of SMURF1. We found that IGF2BP2 is a potential RNA-binding protein for SMURF1 by Starbase database prediction. IGF2BP2 has been reported to promote diabetic retinopathy by stabilizing HAGLR via m6A modification[31]. It was predicted that there were m6A modification sites on SMURF1 mRNA by the SRAMP database (Figure 5A). In addition, compared to the NG group, HG treatment elevated the m6A level of SMURF1 in RPE cells, which was further elevated by HG combined with DFX treatment (Figure 5B). Consistently, HG treatment increased the protein level of IGF2BP2 relative to the NG group, which was further increased by HG combined with DFX treatment (Figure 5C). Subsequently, the interaction between IGF2BP2 protein and SMURF1 mRNA was confirmed, and IGF2BP2 knockdown accelerated the degradation of SMURF1 mRNA (Figure 5D and E). Furthermore, IGF2BP2 knockdown downregulated the expression level of SMURF1 (Figure 5F and G). Therefore, our results indicated that IGF2BP2 upregulates SMURF1 expression by mediating the m6A modification of SMURF1.

Figure 5 Insulin like growth factor 2 mRNA binding protein 2 upregulated SMAD-specific E3 ubiquitin protein ligase 1 expression in an N6-methyladenosine modification-dependent manner. A: The N6-methyladenosine (m6A) modification site of SMAD-specific E3 ubiquitin protein ligase (SMURF) 1 was predicted by the SRAMP database; B: ARPE-19 cells were treated with normal concentration of glucose (NG) or high concentration of glucose (HG), and the m6A level of SMURF1 was then detected by methylated RNA immunoprecipitation (RIP); C: ARPE-19 cells were treated with NG or HG, and then the expression of insulin like growth factor 2 mRNA binding protein 2 (IGF2BP2) was detected by western blot; D: The interaction between IGF2BP2 protein and SMURF1 mRNA was verified by RIP method; E: ARPE-19 cells were transfected with IGF2BP2 knockdown construct or vector, and then cells were treated with dactinomycin (10 µg/mL) for 0, 2, 4, or 6 hours, followed by quantitative real-time PCR (qRT-PCR) to detect the stability of SMURF1 mRNA; F: IGF2BP2 or vector knockdown ARPE-19 cells were collected, and then the mRNA levels of IGF2BP2 and SMURF1 were detected by qRT-PCR; G: ARPE-19 cells were transfected with IGF2BP2 knockdown construct or vector, followed by western blot to detect the protein levels of IGF2BP2 and SMURF1. The experiments were repeated three times. ^aP < 0.05, ^bP < 0.01, ^cP < 0.001. SMURF1: SMAD-specific E3 ubiquitin protein ligase 1; IGF2BP2: Insulin like growth factor 2 mRNA binding protein 2; HG: High concentration of glucose; DFX: Desferrioxamine mesylate; NG: Normal concentration of glucose.

SMURF1 knockdown improved retinal damage in DME mice

We developed the STZ mouse model to evaluate SMURF1 function in vivo. In comparison to the sham group, the fasting blood glucose concentrations of the STZ mice were all greater than 250 mg/dL, and the retinal vessels of these mice were vascularized with leakage (Figure 6A and B). SMURF1 expression was upregulated in STZ mice compared with the Sham group, and SMURF1 knockdown abolished this effect (Figure 6C and D). Moreover, tight junction proteins expression of including claudin-1 occluding and ZO-1 was downregulated in STZ mice, which was reversed by knockdown of SMURF1 (Figure 6E). Meanwhile, Beclin-1 and LC3II/LC3I expression levels were elevated and p62 expression was decreased in STZ mice, while knockdown of SMURF1 counteracted these effects (Figure 6F). Finally, WIF1 expression was downregulated in STZ mice compared to the Sham group, and the Wnt/β-catenin-VEGF signaling pathway was activated, whereas these effects were eliminated by SMURF1 knockdown (Figure 6G). In summary, our study demonstrated that knockdown of SMURF1 ameliorated retinal damage in DME mice.

Figure 6 SMAD-specific E3 ubiquitin protein ligase 1 knockdown improved retinal damage in diabetic macular edema mice. A: The Streptozotocin (STZ) mouse model was constructed, and then the blood glucose concentration of the mice was examined; B: The STZ mouse model was constructed, and the permeability of mouse retinal blood vessels was then examined by Evans Blue assay; C: STZ model mice were injected intraocularly with adenoviruses of sh-NC or sh-SMAD-specific E3 ubiquitin protein ligase (SMURF) 1, and then the expression of SMURF1 was detected by quantitative real-time PCR; D: STZ model mice were injected intraocularly with adenoviruses of sh-NC or sh-SMURF1, and the expression of SMURF1 was then detected by western blot; E: STZ model mice were injected intraocularly with adenoviruses of sh-NC or sh-SMURF1, followed by western blot to detect the expression of occluding, ZO-1, and claudin-1; F: STZ model mice were injected intraocularly with adenoviruses of sh-NC or sh-SMURF1, and then the expression of Beclin-1, LC3I, LC3II, and p62 was detected by western blot; G: STZ model mice were injected intraocularly with adenoviruses of sh-NC or sh-SMURF1, followed by western blot to detect the expression of WIF1, Wnt1, β-catenin, and vascular endothelial growth factor was detected by western blot. n = 6. ^aP < 0.05, ^bP < 0.01, ^cP < 0.001. SMURF1: SMAD-specific E3 ubiquitin protein ligase 1; VEGF: Vascular endothelial growth factor; STZ: Streptozotocin.

DISCUSSION

DME is a leading cause of vision loss in people with diabetes[2]. Although anti-VEGF agents have shown some efficacy in the treatment of DME, there are a number of patients who continue to experience persistent edema after treatment[5]. Currently, there is little known about DME and its pathogenesis has not been fully elucidated. Herein, we revealed the molecular mechanism by which IGF2BP2 upregulates SMURF1 expression in an m6A modification-dependent manner, thereby promoting the ubiquitination and degradation of WIF1, which activates the Wnt/β-catenin-VEGF signaling pathway, thereby contributing to DME development. Our study has provided a theoretical basis as well as potential targets for the treatment of DME.

The BRB plays an important role in maintaining normal retinal function, consisting of tight junctions between RPE cells[41]. Hyperglycemia induces the disruption of RPE cell tight junctions, leading to the formation of subretinal fluid and ultimately DME[42]. Interestingly, autophagy may be involved in the development of DME. For example, activation of protein kinase AMP-activated catalytic subunit alpha 1 reversed methylglyoxal-induced macular edema by inhibiting autophagy[43]. Monounsaturated OA promoted autophagy in RPE cells thereby increasing the risk of DME[11]. Sustained oxidative stress induced autophagy in RPE cells, leading to Age-related macular degeneration[44]. In addition, it has been reported that autophagy disrupts cellular tight junctions. For example, Aβ1-42 oligomers disrupted the tight junctions of capillary endothelial cells by activating autophagy[9]. Human immunodeficiency virus protein tat-induced autophagy down-regulated tight junction protein expression in brain endothelial cells[8]. Furthermore, SMURF1 disrupts the tight junctions of mammary epithelial cells by mediating TRAF4 ubiquitination and thus disruption[15]. It has also been suggested that SMURF1 promotes autophagy by controlling the nuclear entry of TFEB and activating Nrf2[13,14]. SMURF1 promotes dry age-related macular degeneration through acceleration of oxidative stress[20]. In our study, SMURF1 was observed for the first time to be upregulated in DME. More interestingly, SMURF1 knockdown inhibited autophagy-related proteins expression, promoted tight junction-related proteins expression, and ameliorated retinal damage in DME mice. Mechanistically, SMURF1 disrupted tight junctions in RPE cells by inducing autophagy, thereby accelerating DME formation.

SMURF1 mediates the ubiquitination of target proteins to induce diabetes and diabetic complications[17,18]. It is well known that WIF1 inhibits activation of the Wnt/β-catenin signaling pathway by binding to Wnt[45]. The Wnt/β-catenin signaling pathway has been implicated in the pathogenesis of DME, and inhibiting this pathway appears to be detrimental to the development of DME[46]. Inhibition of the Wnt/β-catenin-VEGF pathway improves visual function in mice[25]. Furthermore, a recent study demonstrated that astaxanthin exerts anti-VEGF efficacy in in vitro and in vivo models of DME by down-regulating the expression of the transcription factors hypoxia inducible factor 1 subunit alpha and X-box binding protein 1[47]. Importantly, targeting VEGF inhibits the progression of DME[48]. Moreover, the anti-VEGF therapy is considered a first-line treatment for DME[49]. Interestingly, WIF1 has been reported to promote autophagy in non-small cell lung cancer cells[50]. However, WIF1 expression has also been reported to be negatively correlated with autophagy[23]. In our study, we found that SMURF1 activated the Wnt/β-catenin-VEGF signaling pathway by promoting WIF1 ubiquitination and degradation, which induced autophagy and thus disrupts the tight junctions of RPE cells. Taken together, the present study is the first to elucidate the molecular mechanism by which SMURF1 modulates the progression of DME by modulating WIF1 ubiquitination and degradation.

m6A is a well-characterized post-transcription modification of RNA and functions as a key regulator in several physiological and pathological processes[29]. IGF2BP2 is a key m6A binding protein that regulates gene expression through recognition of m6A modifications and binding to target mRNAs[30]. It has been reported that IGF2BP2 enhances the expression of the autophagy-related gene RB1CC1 in an m6A-dependent manner[51]. It has also been reported that IGF2BP2 upregulates RB1 inducible coiled-coil 1 expression in an m6A-dependent manner thereby triggering RPE cells pyroptosis[31]. We found for the first time that IGF2BP2 binds to SMURF1 and upregulates SMURF1 expression in an m6A-dependent manner. In summary, this study elucidated the molecular mechanism by which the IGF2BP2/SMURF1/WIF1 axis disrupts the tight junctions of RPE cells by inducing autophagy in RPE cells, which ultimately promotes the development of DME.

Since there is no recognized and mature method, the current construction of cellular and animal models of DME is mainly based on the diabetes model construction. It has been reported that vascular permeability damage is closely related to the development of DME, and that increased vascular permeability leads to retinal thickening and thus DME[52]. Interestingly, it has been shown that DFX induces membrane permeability damage[53,54]. Moreover, several studies have demonstrated that treatment of RPE cells with HG/DFX is able to mimic the DME environment in vitro[36,55]. Therefore, in this study, we constructed the in vitro DME model using HG/DFX treatment of RPE cells. However, the appropriateness of this method for DME still needs to be further explored. In addition, this study is a basic research, and our clinical experiments did not involve treatment or interventions for the patients, so the number of clinical samples collected in our study was small. In the future, we will expand the number of clinical samples in clinical intervention trials and use power analysis for further studies when conditions permit. However, whether the current sample size is sufficient for replication in a wider population is uncertain and remains to be further explored. It has been shown that measuring the permeability of retinal blood vessels in mice can be used to assess alterations in retinal structure[56]. Our study investigated the effect of SMURF1 knockdown on the expression of tight junction-related proteins of the BRB in STZ mice. However, given the restrictive conditions, we did not detect the effect of SMURF1 knockdown on retinal vascular permeability in DME mice, which is a challenge we need to address in the future. The SMURF1 inhibitor A01 has been reported to exert a therapeutic benefit in the cataract mouse model[57]. A specific inhibitor of SMURF1 (Lead molecule 38) has also been designed and synthesized, and it showed excellent oral pharmacokinetics in rats[58]. Therefore, SMURF1 inhibitors have great potential for clinical applications. However, whether SMURF1 exerts long-term efficacy in the in vivo treatment of DME and the study of SMURF1 inhibitors in the clinic are our future aims. Furthermore, in this study, we demonstrated that SMURF1 is a potential target for DME treatment. Mechanistically, SMURF1 regulates autophagy in RPE cells by modulating WIF1 ubiquitination and degradation thereby affecting Wnt/β-catenin-VEGF signaling. Significantly, knockdown of SMURF1 directly affected the expression of tight junction-associated proteins and VEGFA in RPE cells and retinas of STZ mice, which suggests that targeting SMURF1 may enhance the efficacy of anti-VEGF therapy in DME. However, it remains to be further explored whether intervening SMURF1 directly affects VEGF action. In addition, considering that various signaling pathways act relatively rapidly in vivo. Therefore, whether the complex regulation of VEGFA by SMURF1 will inhibit the therapeutic applicability of this pathway still needs to be further verified by a large number of experiments. Available studies have shown that treatment of cells with 25 mmol/L D-glucose can be used to construct an in vitro model of DME to mimic a high-concentration glucose environment[55]. In addition, 25 or 30 mmol/L D-glucose is commonly used in the construction of cellular models of diabetes-related diseases[47]. Therefore, we concluded that 25-30 mmol/L D-glucose could be used as a HG treatment concentration for in vitro models of diabetes-related diseases. In the present study, we found that 30 mmol/L D-glucose induced damage of ARPE-19 cells, which led to the determination of this concentration as a high-glucose treatment concentration for the in vitro model of DME in subsequent experiments. The injury of ARPE-19 cells caused by HG treatment suggests that the injury may also have occurred in diabetic patients. However, considering the complexity of the in vivo environment, blood glucose concentrations in diabetic patients may be affected by many other factors. Therefore, whether the in vitro HG treatment concentration can realistically mimic the blood glucose level of diabetic patients needs to be further explored in the future.

CONCLUSION

Our study demonstrates that IGF2BP2 upregulates SMURF1 expression in an m6A-dependent manner, and SMURF1 activates the Wnt/β-catenin-VEGF signaling pathway by facilitating WIF1 ubiquitination and degradation, ultimately leading to DME development. Our study provides the theoretical basis and identifies potential molecular targets for the treatment of DME. We will further explore the small molecule drugs which target the IGF2BP2/SMURF1/WIF1 axis.