METTL3 enhances NSD2 mRNA stability to reduce renal impairment and interstitial fibrosis in mice with diabetic nephropathy

The inclusion criteria of patients were as follows: (1) patients were diagnosed with diabetes: fasting blood sugar (FBG) ≥ 126 mg/dL (7.0 mmol/L) (American Diabetes Association Standards, 2013 [23]); (2), the renal biopsy examination results accorded with the diagnosis standard of DN. The following patients were excluded: (1) patients with type I diabetes; (2) patients with other diseases or complications, such as primary or secondary renal diseases, heart failure, malignant hypertension, cardiovascular accident, infection, liver dysfunction or aberrantly high alanine transaminase level; regnant and lactating women; (3) pregnant and lactating women; (4) patients with cancer and (5) patients without complete clinical information.

The FBG, serum creatinine (SCr), serum cystatin C (S-Cys-C), the 24-h urine protein (24-h U-protein) and urine cystatin C (U-Cys-C) were analyzed using an automatic biochemical analyzer (DxC700AU, Beckman Coulter, Fullerton, CA, USA).

Total RNA from serum samples of patients, mouse kidney tissues, and from SV40-ES-13 cells was extracted using the TRIzol Reagent (Invitrogen, Thermo Scientific Pierce, Rockford, IL, USA). A PrimeScript™ RT reagent kit (Takara Holdings Inc., Kyoto, Japan) was used for cDNA synthesis. Gene expression was quantified using TB Green® Premix ExTaq^TMII (Takara) on a CFX96 PCRsystem (Bio-Rad Inc., Hercules, CA, USA). Relative gene expression normalized to GAPDH was calculated by the 2^−ΔΔCt method. The primers are listed in Table 1.

Male C57BL/6 mice (7 weeks old) procured from Vital River Co., Ltd. (Beijing, China) were fed in a 12-h dark/light cycle with free access to feed and water. This study was approved by the Animal Care and Ethics Committee of Liyang People’s Hospital and carried out in compliance with the Animals in Research: Reporting In vivo Experiments (ARRIVE) guidelines. All animal procedures were in accordance with the Guide for the Care and Use of Laboratory Animals (NIH Publication No. 85–23, revised 1996). Great efforts were made to minimize the pain of animals.

A murine model of DN was induced by the concomitant administration of unilateral nephrectomy (UN), high-fat-diet (HFD) and streptozotocin (STZ) as previously reported [24]. In short, the mice subjected UN after one week of acclimation to have the renal artery, renal vein and nephrotoxin of the left kidney ligated and resected. In general, mice underwent UN would not show significant symptoms, and the procedure would not affect the risk of developing hyperglycemia, hypertension or nephrotoxin. After two days, the mice were fed with normal feed (10% kcal from fat; Control group; n = 15) or HFD (70% kcal from HFD; HFD group; n = 90). All feed was procured from Trophic Animal Feed High-Tech Co., Ltd. (Nantong, Jiangsu, China).

After three weeks, the mice given HFD were intraperitoneally injected with STZ (100 mg/kg) to induce DN. The mice were continually fed for five weeks. Mice with DN were further assigned into six groups (n = 15 in each): DN group; adenovirus vector (AAV)-negative control (NC) group; AAV-NSD2 group; AAV-NSD2 + Glucagon group; AAV-METTL3 group and AAV-METTL3 + AAV-shRNA group after corresponding treatments with AAV-NC, AAV-NSD2, AAV-NSD2 + glucagon, AAV-METTL3 and AAV-METTL3 + AAV-shRNA, respectively.

The AAV overexpressing NSD2 (AAV-NSD2) and METTL3 (AAV-METTL3), and the vector silencing NSD2 (AAV-shRNA) were all procured from GenePharma Co., Ltd. (Shanghai, China). The empty AAV vector (AAV-NC; GenePharma) was used as control. The virus titer was 1 × 10⁹ TU/mL. Mice subjected to AAV treatment were injected with the corresponding AAV at one week after STZ injection. The AAV was injected at 20 μL each time, and the injection was conducted once a week for four weeks. The clinical dose of glucagon for human is 0.5 U/70 kg. The corresponding dose for mice (0.5/70) × 9.1 = 0.065 (U/kg), which was used to treat the mice.

The paraffin-embedded tissues were cut into 5-μm sections. The sections were dewaxed and rehydrated. The pathological changes in kidney tissues were determined using a hematoxylin and eosin (HE) staining kit (G1120, Solarbio Science & Technology Co., Ltd., Beijing, China) or a Masson's trichrome staining kit (G1340, Solarbio) following the manufacturer’s protocol. All images were photographed using a camera-equipped light microscope (Nikon Instruments Inc., Tokyo, Japan).

A mouse mesangial cell line SV40-MES-13 (CL-0470) was procured from Procell Life Science & Technology Co., Ltd. (Hubei, China). The cells were cultured in Dulbecco's modified Eagle's medium (DMEM; Gibco Company, Thermo Fisher Scientific) containing 10% fetal bovine serum (FBS) and 5.6 mM glucose at 37 °C with 5% CO₂.

The pEXP-RB-Mam plasmid-based overexpressing vector-NSD2, vector-METTL3 and the control vector NC used for cell transfection were procured from RiboBio Co., Ltd. (Guangzhou, Guangdong China). Short hairpin (sh) RNAs of YTHDF1 and NSD2(sh-YTHDF1 and sh-NSD2) and the control sh-NC were procured from Genepharma. All transfection was conducted using the Lipofectamine 2000 (Thermo Fisher Scientific) according to the instruction manual. The cells cultured in DMEM containing 5.6 nM glucose were set as normal-glucose (NG) group, while those cultured in the medium containing 30 nM glucose were set as high-glucose (HG) group.

Total protein in SV40-MES-13 cells or kidney tissues was extracted using the radio-immunoprecipitation assay cell lysis buffer (Beyotime Biotechnology Co. Ltd., Shanghai, China). After protein determination using a Pierce™ BCA kit (Thermo Fisher Scientific), an equal amount of protein was run on 10% SDS-PAGE and transferred loaded onto polyvinylidene fluoride membranes (Invitrogen). The membranes were blocked in 5% non-fat milk for 2 h and incubated with the primary antibodies anti-NSD2 (1:1,000, ab259940, Abcam), anti-Fibronectin (1:1,000, ab2413, Abcam), anti-collagen type I alpha 1 chain (COL1A1; 1:1,000, ab260043, Abcam), anti-E-cadherin (1:1,000, #3195, Cell Signaling Technology (CST), Beverly, MA, USA), anti-YTHDF1 (1:1,000, ab252346, Abcam) anti-METTL3 (1:1,000, ab195352, Abcam) and the internal control anti-GAPDH (1:1,000, #5174, CST) at 4℃ overnight, followed by incubation with horseradish peroxidase (HRP)-conjugated IgG (1:1,000, ab6721, Abcam) at 25℃ for 2 h. The protein signals were developed using the Pierce™ ECL system (Thermo Fisher Scientific). Relative protein level was examined using the Image J.

Proliferation of the cells was examined using a BeyoClick™ EdU-647 kit (Beyotime). In short, the cells were sorted in 6-well plates. After adherence, the medium was renewed, and each well was loaded with 10 μM EdU solution and the cells were incubated at 37℃ for 2.5 h. Next, the cells were fixed in 4% PFA (Beyotime) for 15 min, permeabilized in 0.3% Triton X-100 (Elabscience Biotechnology Co., Ltd., Wuhan, Hubei, China) for 8 min, and then added with 500 μL Apollo staining buffer in the dark for 40 min. The nuclei were stained with 4', 6-diamidino-2-phenylindole (DAPI) for 10 min. The staining was observed under a fluorescence microscope (Zeiss, Oberkochen, Germany) and quantified using the Image J software.

The cells were fixed in 4% PFA for 10 min and incubated in 1% BSA/10% normal goat serum/0.3 M glycine-contained 0.1% PBS-Tween for 1 h to permeate the cells and block the non-specific protein–protein interaction. After that, the cells were incubated with anti-Fibronectin (1:250, ab2413, Abcam), anti-COL1A1 (1:250, ab270993, Abcam) and anti-E-cadherin (1:50, #3195, CST) at 4℃ overnight and then with fluorescence-conjugated IgG H&L (Alexa Fluor® 488; 1:500, ab150077, Abcam) or IgG H&L (Alexa Fluor® 647; 1:500, ab150079, Abcam) at 25℃ for 2 h. The nuclei were stained with DAPI (blue). The cells were observed under the fluorescence microscope. The mean fluorescence intensity (MFI) of proteins was analyzed using the Image J software.

Total m6A RNA methylation was determined using an EpiQuik kit (Epigentek, NY, USA) according to the manufacturer’s protocol. Total RNA was isolated using the TRIzol reagent as well. Each well was loaded with 80 μL binding solution. The RNA samples, positive control and negative control were loaded into the designative wells and allowed to stand at 37 °C for 90 min. After being washed in washing buffer for three times, the samples were incubated with 50 μL diluted capture antibody at 22℃ for 60 min, and with 50 μL detection antibody and 50 μL diluted enhancer solution for 30 min. After that, 100 μL developer solution was further added for a 10-min incubation to produce color change. The reaction was terminated by 100 μL stop solution. The optical density at 450 nm was read using a microplate reader. The m6A RNA quantity was evaluated according to the standard curve. The ratio of the amount of m6A RNA methylation to the amount of total input RNA was examined.

Total RNA from cells was extracted using a miRNeasy Mini kit (Qiagen GmbH, Hilden, Germany) through DNA digestion using the RNase-free DNase Set (Qiagen). Magnetic beads (Millipore Corp., Billerica, MA, USA) were coated with 5 μg anti-m6A (ab208577, Abcam) or anti-IgG (Abcam) at 22℃ for 30 min. The antibody-coated beads were incubated with 50 µg total RNA in RNase-inhibiting IP buffer at 4 °C overnight. After digestion in proteinase K, the m6A-bound RNA was precipitated by the phenol–chloroform RNA extraction method. The abundance of m6A was examined by RT-qPCR.

RIP was performed according to the instructions of a Magna RIP™ kit (Millipore). The magnet beads were coated with 5 μg primary antibodies (anti-YTHDF1, anti-YTHDF2, anti-YTHDF3 or anti-IgG; all acquired from Abcam) and incubated at 22℃ for 30 min. The antibody-coated magnet beads were cultured with the cells at 4 °C overnight. Then, the magnet-protein-RNA complexes were washed with RIP washing buffer for six times and then incubated with proteinase K digestion buffer at 55 °C for 30 min. The RNA was extracted using the phenol–chloroform extraction method. Then, the abundance of NSD2 was determined by RT-qPCR.

The cells were treated with 5 µg/mL actinomycin D (Sigma-Aldrich Chemical Company, St Louis, MO, USA) to suppress the transcription of whole mRNA. After incubation for indicated time, the cells were collected, and the RNA sample was collected for RT-qPCR to examine the transcription level of NSD2.

Prism 8.02 (GraphPad, La Jolla, CA, USA) was applied for data analysis. Data from at least three independent experiments were exhibited as the mean ± standard deviation (SD). Differences were analyzed by the t test (two groups) or one- or two-way analysis of variance (ANOVA) (multiple groups) followed by Tukey’s post-hoc test. Correlations between variables were analyzed by Pearson’s correlation analysis. p < 0.05 was set as the cut-off value for significant difference.

The FBG, SCr, S-Cys-C, 24-h U-protein, and U-Cys-C levels in the serum samples of the patients were examined. Compared to the healthy controls, the serum levels of FBG, SCr, S-Cys-C, 24-h U-protein, and U-Cys-C were significantly elevated in patients with DN (Fig. 1A). Thereafter, the RT-qPCR results indicated that the expression of NSD2 in serum samples of patients with DN was significantly reduced (Fig. 1B). The Pearson’s correlation analyses showed that the serum NSD2 level was inversely correlated with the FBG, SCr, S-Cys-C, 24-h U-protein, and U-Cys-C levels in patients with DN (Fig. 1C).

A murine model of DN was established. The model mice were administrated with AAV-NSD2 overexpressing NSD2 to examine the function of NSD2 in DN. Since NSD2 has been reported to participate in beta-cell proliferation and result in decreased glucose concentration [14], we therefore further treated the AAV-NSD2-treated mice with glucagon to evaluate whether the effect of NSD2 on DN is glucose-dependent.

The kidney/body weight and the changes of physiological and pathological indicators were analyzed. Compared to the control mice, the kidney/body weight of DN mice was significantly enhanced with time (Fig. 2A), along with incrementally increased FBG levels and SBP as well as renal impairment-related indicators SCr, S-Cys-C, 24-h U-protein, and U-Cys-C (Fig. 2B). Importantly, AAV-NSD2 treatment significantly reduced the kidney/body weight of mice and decreased the levels of FBG, SBP, SCr, S-Cys-C, 24-h U-protein, and U-Cys-C. The additional glucagon treatment significantly blocked the suppressive effects of AAV-NSD2 on FBG levels, and it also partly counteracted the suppressive effects of AAV-NSD2 on SBP, SCr, S-Cys-C, 24-h U-protein, and U-Cys-C (Fig. 2A and B).

Meanwhile, the levels of oxidative stress- and inflammation-related markers (SOD and MDA; IL-6 and MCP-1), and the contents of collagenous tissue-specific amino acid hydroxyproline in mouse kidney were determined using ELISA kits. We found that the SOD contents were decreased whereas the MDA, IL-6, MCP-1, and hydroxyproline levels were significantly elevated in DN mice. It was observed that the AAV-NSD2 treatment significantly elevated the SOD levels and suppressed the MDA, IL-6, MCP-1, and hydroxyproline levels in mouse kidneys, whereas the effect of AAV-NSD2 was partially diminished by glucagon (Fig. 2C). Moreover, the levels of NSD2 and the interstitial fibrosis-related proteins COL1A1, Fibronectin and E-cadherin were examined using western blot analysis. It was found that the levels of COL1A1 and Fibronectin in mouse kidney tissues were incrementally elevated whereas the levels of NSD2 and E-cadherin were declined in DN mice. Further AAV-NSD2 treatment significantly elevated the NSD2 and E-cadherin levels but reduced the COL1A1 and Fibronectin levels in mice. The glucagon treatment slightly counteracted the effects of AAV-NSD2, though, there showed no statistical significance (Fig. 2D). In addition, the HE staining results indicated that the mice with DN showed significant pathological symptoms, such as glomerular dilatation, glomerulosclerosis and diffuse mesangial proliferation. But the pathological changes were alleviated upon NSD2 overexpression, and the glucagon treatment blocked the effect of NSD2 in a certain extent (Fig. 2E). The Masson’s trichrome staining showed that the collagen deposition area in kidney tissues was incrementally increased in DN mice. Overexpression of NSD2 significantly decreased the collagen deposition and reduced fibrosis. However, the anti-fibrotic role of NSD2 was partially blocked by glucagon as well (Fig. 2F). Collectively, it can be inferred that NSD2 can alleviate interstitial fibrosis in the kidney tissues of mice with DN, which is partially attributed to its regulation on glucose suppression.

SV40-MES-13 cells were treated with HG to mimic a DN-like condition in vitro. The RT-qPCR showed that the NSD2 expression in cells was decreased after HG treatment in a time-dependent manner (Fig. 3A). After 48 h, the NSD2 expression was declined by half. Therefore, 48 h of HG treatment was applied in the subsequent experiments.

DNA vectors overexpressing NSD2 were transfected into SV40-MES-13 cells, and the successful transfection was confirmed by RT-qPCR (Fig. 3B). Thereafter, the cells were treated with HG. EdU labeling was performed to examine proliferation activity of cells. It was found that the DNA replication ability of SV40-MES-13 cells was significantly enhanced by HG treatment but reduced upon NSD2 overexpression (Fig. 3C). Moreover, the immunofluorescence staining results indicated that HG treatment significantly reduced the expression of E-cadherin but elevated the expression of COL1A1 and Fibronectin. But these changes were prevented by NSD2 treatment (Fig. 3D-F).

The upstream mechanism responsible for HG-induced NSD2 downregulation was explored. HG treatment has been previously reported to reduce m6A modification to affect gene expression [19]. According to the m6A quantification assay, the m6A modification level in HG-treated SV40-MES-13 cells was significantly reduced (Fig. 4A). In addition, the subsequent MeRIP-qPCR assay indicated that HG treatment significantly reduced m6A modification of NSD2 mRNA (Fig. 4B), which might be the cause of NSD2 downregulation in the HG-treated cells.

As a “writer” of m6A modification, the expression of METTL3, which is necessary for the insulin secretion, was suggested to be reduced in the peripheral blood samples of diabetic patients [20]. We then explored the METTL3 expression in the serum samples of the patients with DN. The RT-qPCR suggested that the serum level of METTL3 was significantly reduced in patients with DN (Fig. 4C), which presented a positive correlation with the NSD2 expression (Fig. 4D). We therefore speculated that METTL3 downregulation in DN reduces m6A modification of NSD2 mRNA to induce NSD2 downregulation. Thereafter, overexpression vector of METTL3 was transfected into SV40-MES-13 cells, and the successful upregulation of METTL3 was examined by RT-qPCR again (Fig. 4E). Thereafter, it was found that METTL3 overexpression significantly elevated the expression of NSD2 in the HG-treated SV40-MES-13 cells (Fig. 4F-G).

The m6A level in SV40-MES-13 cells was further examined after METTL3 overexpression. It was found that METTL3 overexpression significantly restored the m6A level in cells blocked by HG treatment (Fig. 5A). Overexpression of METTL3 also restored the m6A modification level on NSD2 mRNA after HG treatment (Fig. 5B). Moreover, METTL3 overexpression enhanced the NSD2 mRNA stability (Fig. 5C).

Thereafter, RIP and RT-qPCR assays were performed to examine the levels of common m6A modification “readers” (YTHDFs) on the NSD2 mRNA. It was found that METTL3 overexpression significantly elevated the YTHDF1 modification level on the NSD2 mRNA (Fig. 5D). Silencing of YTHDF1 was introduced in SV40-MES-13 cells. The transfection of sh-YTHDF1 significantly reduced the expression of NSD2 (Fig. 5E-F) and reduced the stability of NSD2 mRNA (Fig. 5G).

To identify the interaction between METTL3 and NSD2 and their function in SV40-MES-13 cells, the cells transfected with vector-METTL3 were further transfected with shRNA of NSD2. Still, the successful transfection was confirmed by RT-qPCR (Fig. 6A). The cells were subjected to HG treatment. The EdU labeling assay suggested that the METTL3 overexpression blocked the DNA replication ability of SV40-MES-13 cells induced by HG, but that of cells was rescued when NSD2 was artificially suppressed (Fig. 6B). Likewise, the immunofluorescence staining results indicated that METTL3 overexpression led to an increase in E-cadherin expression while a decline in the expression of COL1A1 and Fibronectin. However, further silencing of NSD2 blocked the role of METTL3 and restored the expression of COL1A1 and Fibronectin whereas reduced the expression of E-cadherin (Fig. 6C-E).

AAV-METTL3 overexpressing METTL3 and the AAV-shRNA silencing NSD2 were administrated into mice with DN to validate their interaction and function in vivo. Compared to AAV-NC, AAV-METTL3 significantly reduced the kidney/body weight of mice. But the concomitant administration of AAV-shRNA increased the kidney/body weight of mice with DN again (Fig. 7A). AAV-METTL3 also significantly reduced the levels of FBG, SCr, S-Cys-C, 24-h U-protein, U-Cys-C and SBP in mice, but the levels of these indicators were restored by AAV-shRNA (Fig. 7B). Moreover, the ELISA results showed that the AAV-METTL3 treatment elevated the SOD level whereas suppressed the IL-6, MCP-1 and hydroxyproline levels. Likewise, the function of AAV-METTL3 was suppressed by NSD2 silencing (Fig. 7C). In addition, the protein levels of METTL3 and NSD2 and the interstitial fibrosis-related proteins COL1A1, Fibronectin and E-cadherin in the kidney tissues were examined. The western blot analysis indicated that AAV-METTL3 treatment significantly elevated the protein levels of METTL3, NSD2 and E-cadherin in mouse kidney tissues and reduced the levels of COL1A1 and Fibronectin. As expected, further administration of AAV-shRNA reduced the levels of NSD2 and E-cadherin, but restored the levels of COL1A1 and Fibronectin in mouse kidney tissues (Fig. 7D). The HE staining results also revealed that overexpression of METTL3 mitigated the pathological changes in mice. But the DN-related symptoms in mouse kidney was aggravated again upon NSD2 silencing (Fig. 7E). The Masson’s trichrome staining suggested that the collagen deposition in mouse kidney tissues was significantly suppressed by METTL3 overexpression. However, the protective role of METTL3 was blocked by NSD2 silencing (Fig. 7F).