Male C57BL/6 mice weighing 16–20 g were purchased from Medical Experimental Animal Center of Xi’an Jiaotong University. All mice were housed in a temperature-controlled room (22–24°C) with a 12 h light/dark cycle and have free access to diets and water. Mice were classified into Control group (n = 8) and high-fat-diet group (HFD, n = 20) randomly according to body weight. Control mice were fed with normal chow diets, while HFD mice were fed with high-fat diets. After 8 weeks, following fasting for 12 h, mice in the Control group and HFD group were intraperitoneally injected with citrate buffer and 60 mg/kg body weight streptozotocin (STZ, Sigma-Aldrich, United States) respectively for three consecutive days. Three days after that, fasting blood glucose (FBG) was examined after fasting. DM model is successfully achieved if the FBG of the mice in the HFD group is 15–25 mmol/L for two consecutive days. Then the DM mice were randomly divided into Model group (n = 8) and EPO group (n = 8). The mice in EPO group were subcutaneously injected with 500 IU/kg body weight recombinant human EPO (3SBIOINC., China) 3 times a week, while the mice in Control group and Model group were subcutaneously injected with normal saline. The body weight and FBG of each mouse were measured weekly. After the administration for 6 weeks, mice were sacrificed and the serum and kidney tissues were obtained for subsequent experiments. All animal experimental procedures were approved by the Institutional Animals Care and Use Committee at Xi’an Jiaotong University and in accordance with the National Institutes of Health Guide for Care and Use of Laboratory Animals.

Human glomerular mesangial cells were kindly provided by Prof. Ren (School of Basic Medical Sciences, Xi’an Jiaotong University, China). Mesangial cells were cultured in low glucose Dulbecco’s Modified Eagle’s Medium (DMEM, Hyclone, United States) supplemented with 10% fetal bovine serum (Biological Industries, Israel), 100 U/ml penicillin plus 100 μg/ml streptomycin (Solarbio, China) under 5% CO2 at 37°C. Cells were treated with low glucose (Control, 5.5 mM) and high glucose (HG, 30 mM) with or without EPO (20 IU/ml) for 48 h. For all experiments, cells were cultured in serum-free medium for 12 h before HG and EPO administration.

For detecting biochemical changes of kidney tissues and serum, renal tissues were homogenized in phosphate-buffered saline (PBS). The levels of MDA and SOD activity of tissue homogenates as well as the levels of Creatinine (CRE) and Blood Urea Nitrogen (BUN) were measured by assay kits (Beyotime, China).

Pathological changes of kidney tissues were detected as described previously (Chen et al., 2018). In brief, paraffin-embedded renal tissue sections went through de-waxing in xylene and rehydration in decreasing grades of alcohol, and then tissue sections were stained by hematoxylin-eosin (HE) stain, Masson stain and periodic acid-Schiff stain according to manufacturer’s instructions. Results were observed by a light microscrope (Olympus, Japan).

Total protein of renal tissues and glomerular mesangial cells was extracted by RIPA lysis buffer supplemented with protease inhibitors (Solarbio, China). Protein concentrations were determined using the BCA Protein Kit (Solarbio, China). All samples were boiled at 95°C for 5 min. Equal amounts of proteins were separated in 8–12% SDS-PAGE gels and then transferred onto PVDF membranes (Millipore, United States). After blocked with 5% non-fat milk for 2 h, the membranes were incubated with various primary antibodies overnight at 4°C, namely rabbit anti-LC3B (1:1,500, GeneTex, United States), rabbit anti-P62 (1:1,000, Proteintech, China), rabbit anti-PINK1 (1:1,000, Abcam, United Kingdom), mouse anti-Parkin (1:200, Santa Cruz Biotechnology, United States), rabbit anti-Drp-1 (1:1,000, Cell Signaling Technology, United States), rabbit anti-Mfn-2 (1:1,000, Cell Signaling Technology, United States). And then the membranes were washed 3 times with Tris-buffered saline containing Tween-20 for around 15 min and subsequently incubated with corresponding secondary antibodies for 2 h. After washing 3 times with TBST for around 15 min, the bands were detected by using ECL reagents (Abbkine, United States) and their density was quantified by using ImageJ software.

Total RNA was isolated from renal tissues using RNA isolation kit (Takara, Japan). RNA was reverse transcribed into cDNA using qPCR-RT Master mix kit (Takara, Japan). Finally, cDNA was analyzed by real-time PCR using the SYBR Green qPCR Mix (Takara, Japan). Relative expression of target mRNA was calculated based on the 2^−ΔΔCt comparative method and normalized to that of the internal standard (β-actin). The primers are listed in Table 1.

Mesangial cells were seeded in 12-well plates and exposed to different treatments. As for the measurement of mitochondrial ROS, the treated cells were incubated with MitoSOX Red (Invitrogen, United States) and Mito-tracker Green (Beyotime, China) for 20 min according to manufacturer’s instructions. In terms of the detection of mitochondrial transmembrane potential, the treated cells were incubated with JC-1 working solution (Beyotime, China) for 30 min, and then the cells were washed by JC-1 washing buffer for 3 times. After that, fluorescence images were observed by a fluorescence microscope (Olympus, Japan).

In order to detect the level of mitophagy, the colocalization of mitochondria with autophagosome was achieved by GFP-LC3 lentivirus transfection and mitochondrial staining as reported previously (Jiang et al., 2020). Briefly, mesangial cells were infected with GFP-LC3 lentivirus (Genechem, China) at a suitable MOI of 50 in accordance with manufacturer’s instructions. After 72 h of transfection, the positive cells were screened out by the treatment of puromycin (5 ug/ml). The viable cells, after being exposed to HG and EPO for 48 h, were used to measure the colocalization of mitochondria with autophagosome by staining mitochondria through Mito-tracker Red (Beyotime, China). Fluorescence images were observed by a fluorescence microscope (Olympus, Japan).

Mesangial cells were cultured in 12-well plates with coverslips. After receiving various treatments for 48 h, cells were fixed with 4% paraformaldehyde at room temperature for 20 min and washed 3 times with PBS for 15 min. Then cells were permeabilized with 0.1% Triton X-100 for 30 min at 4°C. After being blocked with goat serum for 2 h, cells were incubated with primary antibodies overnight at 4°C, namely rabbit anti-LC3B (GeneTex, United States), rabbit anti-PINK1 (Abcam, United Kingdom), mouse anti-Parkin (Santa Cruz Biotechnology, United States). After that, cells were washed 3 times for 15 min and incubated with fluorescent secondary antibody (Zhuangzhi, China) for 2 h in the darkroom at room temperature. Eventually, the nuclei of cells were stained with DAPI and images were captured by a fluorescence microscope (Olympus, Japan).

PINK1 siRNA and relevant scrambled siRNA were purchased from GenePharm (Shanghai, China). Mesangial cells at 30–40% confluence, which were seeded in 6-well plates, were transfected with scrambled siRNA or PINK1 siRNA using lipofectamine 2000 according to the manufacturer’s instructions.

Cell apoptosis was detected as previously described (Tsai et al., 2020). Mesangial cells were seeded in 6-well plates. After being exposed to different treatments for 48 h, treated cells were incubated with Annexin V-FITC and PI solution (Beyotime, China) for 15 min in the darkroom at room temperature. The level of cell apoptosis was evaluated by flow cytometry (ACEA Biosciences, United States).

All data were obtained from at least three independent experiments and were presented as mean ± SEM. Results were subjected to a one-way analysis of variance (ANOVA) followed by least significant difference (LSD) test when comparing multiple groups by IBM SPSS 22.0. Values of p < 0.05 were considered significant.

Autophagy and the functions of mesangial cells are associated with the development of DN, and autophagic flux serves as an important index of the level of autophagy. So HG-injured mesangial cells were used as the in vitro DN model and key proteins related to autophagic flux were measured. As shown in Figure 1, compared with the control group, HG significantly reduced the ratio of LC3B-Ⅱ/LC3B-Ⅰ and increased P62 protein expression, resulting in the blockage of autophagic flux. However, EPO increased the ratio of LC3B-Ⅱ/LC3B-Ⅰ and decreased the level of P62 in a dose-dependent manner (Figure 1), thereby alleviating the blockage of autophagic flux in HG-injured mesangial cells. Therefore, we choose 20 IU/ml EPO treatment to conduct next experiments.

HG was reported to aggravate mitochondrial malfunction, as displayed by mitochondrial fragmentation as well as abnormal mitochondria membrane potential and mitochondrial ROS levels, and induce a higher level of apoptosis in DN models (Jiang et al., 2020). Consistently, compared with Control group, mesangial cells exposed to HG had a lot of fragmented mitochondria (Figure 2A), along with the up-regulated Drp-1 (Figure 2B) and down-regulated Mfn-2 protein levels (Figure 2C) which were indispensable for mitochondrial fission. Furthermore, HG induced the loss of mitochondrial membrane potential (Figure 2D), promoted the ectopic production of mitochondrial ROS (Figures 2E,G), and increased the apoptosis level (Figures 2F,H) in mesangial cells, leading to mitochondrial dysfunction. However, EPO partly restored the mitochondrial morphology as well as the loss of mitochondria membrane potential, and significantly decreased the level of mitochondrial ROS and apoptosis in HG-treated mesangial cells compared with HG group.

In order to explore the underlying mechanism, we then identified whether the protective effects of EPO attribute to the rescue of PINK1/Parkin-mediated mitophagy in mesangial cells. Mitophagy, especially PINK1/Parkin-mediated mitophagy, has been demonstrated to play an important role in DN. Many studies have shown that decreased level of PINK1/Parkin-mediated mitophagy is involved in the development of DN (Li et al., 2017; Jiang et al., 2020). Accordingly, our results showed that HG significantly decreased the protein expression of PINK1 (Figure 3A) and Parkin (Figure 3B) compared with the control group, indicating reduced level of PINK1/Parkin-mediated mitophagy, while EPO reversed the expression of PINK1 and Parkin and promoted PINK1/Parkin-mediated mitophagy. In PINK1/Parkin-mediated mitophagy, the interaction of LC3 and mitochondria and the translocation of Parkin from cytoplasm to mitochondria are essential. Compared with the control group, HG decreased the co-localization of LC3 with mitochondria (Figure 3C), Parkin with mitochondria (Figure 3D) as well as LC3 with Parkin (Figure 3E), thereby inhibiting the translocation of LC3 to mitochondria and the translocation of Parkin from cytoplasm to mitochondria. Also, reduced number of GFP-LC3 puncta was observed in HG group (Figure 5F), suggesting a decreased level of PINK1/Parkin-mediated mitophagy. In contrast, EPO increased the co-localization of LC3 with mitochondria (Figure 3C), Parkin with mitochondria (Figure 3D) and LC3 with Parkin (Figure 3E) as well as the number of GFP-LC3 puncta (Figure 3F), indicating that PINK1/Parkin-mediated mitophagy was involved in the effect of EPO on HG-injured mesangial cells. These results demonstrated that HG inhibited PINK1/Parkin-mediated mitophagy in mesangial cells, while EPO restored this change.

In order to confirm the critical role of PINK1/Parkin-mediated mitophagy in HG-injured mesangial cells, we further used PINK1 siRNA#2 to knockdown the expression of PINK1 in mesangial cells. The knockdown efficiency of PINK1 was verified by Western blotting (Figure 4A). Again, compared with the control group, HG decreased the co-localization of LC3 and Parkin and significantly reduced the number of GFP-LC3 puncta, and EPO treatment alleviated these changes (Figures 4B-D). However, the knockdown of PINK1 partly abrogated the effects of EPO on the enhanced co-localization of LC3 with Parkin and increased number of GFP-LC3 puncta (Figures 4B,C), demonstrating that the protective effects of EPO was dependent on PINK/Parkin-mediated mitophagy.

In vitro experiments indicated that EPO mitigated HG-injured mesangial cells. We then explored whether EPO alleviated DN in vivo experiments by establishing the DN mouse model through high-fat diet combined with intraperitoneally injection with STZ. During the administration of EPO or normal saline in three different groups for 6 weeks, we examined the body weight and FBG of mice every week. As shown in Figure 5A, there was no significant difference in the body weight between these three groups in the whole 6 weeks. In terms of FBG, mice of the Model group had a significant higher level of FBG than that of the Control group, while EPO significantly decreased the FBG levels of the Model mice (Figure 5B). Moreover, compared with mice in Control group, mice in the Model group had increased levels of serum creatinine (Figure 5C), blood urea nitrogen (Figure 5D), MDA (Figure 5F) and reduced SOD activity (Figure 5E), indicating decreased renal functions and increased oxidative stress of kidney tissues. However, EPO significantly reversed these parameters mentioned above (Figures 5C–F), thereby improving renal functions of the Model mice. Subsequently, we used histological staining, including HE, Masson and PAS staining, to examine kidney morphology, collagen deposition and basement membrane thickness respectively. As shown in Figure 5G, abnormal glomerular structure, renal fibrosis and matrix accumulation were observed in kidney tissues of mice in Model group compared with the mice of the control group, while EPO administration alleviated these histological changes.

According to the in vitro results, the protective effects of EPO on mesangial cells exposed to HG were associated with PINK1/Parkin-mediated mitophagy. So we further detected the mRNA and protein levels of genes in PINK1/Parkin-mediated mitophagy in the kidney tissues. Compared with mice in the control group, the ratio of LC3B-II/LC3B-I (Figures 6E,F), the mRNA and protein levels of PINK1 (Figures 6C,E,F), and Mfn-2 protein level (Figures 6G,H) in the kidney tissues of mice of Model group were significantly decreased, while P62 and Drp-1 were significantly increased in terms of the mRNA and protein levels (Figures 6B,E–H), resulting in the blockage of autophagic flux and inhibited level of PINK1/Parkin-mediated mitophagy. However, there was no difference in the mRNA expression of LC3 in these two groups (Figure 6A), and the mRNA and protein levels of Parkin were elevated in the model group (Figures 6D,G,H). Of note, EPO administration partly reversed these changes as indicated by the significantly increased ratio of LC3B-II/LC3B-I (Figures 6E,F) as well as expressions of PINK1 (Figures 6C,E,F) and Mfn-2 (Figures 6G,H), and decreased expressions of P62 (Figures 6B,E,F) and Drp-1 (Figures 6G,H). These results suggested that EPO could mitigate the blockage of autophagic flux and improve the level of PINK1/Parkin-mediated mitophagy.