VDR/Atg3 Axis Regulates Slit Diaphragm to Tight Junction Transition via p62-Mediated Autophagy Pathway in Diabetic Nephropathy

Diabetic nephropathy (DN) is one of the most common microvascular complications of patients with diabetes, and 30% of patients with type 1 diabetes and 40% of patients with type 2 diabetes will develop DN (1). The main pathophysiological change of DN is albuminuria, which has been attributed to podocytes injury and extensive foot effacement and ultimately leads to decline of renal function (2,3). Slit diaphragm (SD) has a unique zipper structure, mainly composed of nephrin, podocin, CD2AP, and other proteins, and its morphological and functional stability is the key to maintenance of normal glomerular filtration function (4). Maintaining the normal SD structure is therefore vital for the treatment of DN at early stage.

Tight junction (TJ) is a special structure formed by direct membrane contact that exists in glomeruli and participates in intercellular interaction and signal transduction (5). In podocytes, TJ, mainly composed of zona occluden-1 (ZO-1), junctional adhesion molecule A (JAM-A), and occludin, can regulate the filtration of water, solutes, and proteins (5,6). Under certain nephrotic conditions, the integrity and function of the glomerular SD structure were impaired but presumably replaced by the TJ structure, resulting in so-called SD-TJ transition (7–9). SD-TJ transition was manifested as foot processes effacement under electron microscopy (EM) (10). Previous studies have demonstrated that blocking SD-TJ transition is helpful for the stabilization of SD structure and attenuating proteinuria under nephrotic conditions, including DN (10,11). However, the molecular mechanism of SD-TJ transition in DN is still unknown.

Autophagy is a conservative degradation system that formed and developed gradually during evolution for physiological and pathological conditions (12). Previously, Kume and Koya (13) showed that insufficient podocytes autophagy is associated with severe podocytes injury, SD destabilization, and massive proteinuria in patients with diabetes. A body of evidence has also indicated that recovery of autophagic flux in podocytes can effectively alleviate podocytes injury and reduce foot processes effacement in DN (14–16). However, the cross talk between autophagy and SD-TJ transition has not yet been described. Interestingly, autophagy selectively reduces epithelial TJ permeability of ions and small molecules by autophagic degradation of the TJ protein claudin-2 (17). Hence, we hypothesized that impaired autophagic flux blocked TJ proteins degradation in podocytes, leading to SD-TJ transition in DN.

Vitamin D receptor (VDR) is a nuclear transcription factor that regulates the expression of many genes by specifically binding to its ligand 1,25-(OH)₂D₃, to regulate the progress of various life activities. VDR can promote the initiation, nucleation, maturation, and other processes of autophagy by regulating beclin-1, bcl-2, mTOR, intracellular calcium sample concentration, and lysosomal maturation (18). Current researches have demonstrated that VDR downregulation in human renal biopsy samples is associated with severity of albuminuria in patients with type 2 diabetes (19,20). In addition, VDR knockout mice showed more severe albuminuria and glomerulosclerosis due to the thickened glomerular basement membrane and podocytes effacement (21,22). However, whether VDR deficiency in podocytes was responsible for the impaired autophagic flux and/or SD-TJ transition in DN is largely unknown.

In the current study, we describe for the first time that impaired autophagic flux contributed to SD-TJ transition in human renal biopsy samples, diabetic animals, and cultured podocytes; we also provide the novel finding that the VDR/Atg3 axis was responsible for the SD-TJ transition via the p62-mediated autophagic pathway in DN.

Human kidney biopsy specimens were obtained from 24 patients who were diagnosed with DN at IIa, IIb, III, and IV stages and nontumor kidney tissues from 6 patients who had renal cell carcinoma and underwent nephrectomy were taken as the control group. All of the DN patients had been diagnosed with diabetes at least 5 years prior, with different levels of proteinuria. None of the patients had other autoimmune diseases.

Reagents, 3-methyladenine (3-MA), chloroquine (CQ), and streptozocin (STZ), were purchased from Sigma-Aldrich (St. Louis, MO). Antibodies included the following: podocin (ab50339; Abcam), VDR (ab3508; Abcam), p62 (ab56416; Abcam), LC3I and LC3II (4108S; Cell Signaling Technology), ZO-1 (ab190085; Abcam), Atg3 (3415; Cell Signaling Technology), Atg7 (2613; Cell Signaling Technology), and WT1 (ab89901; Abcam).

All animal experiments conformed to the National Institutes of Health Guide for the Care and Use of Laboratory Animals. A total of 27 male Sprague-Dawley rats (5–6 weeks old, ∼190 g) were purchased from Shanghai Slac Laboratory Animal (Shanghai, China). Rats were randomly separated into three groups: group 1, the control group (n = 9); group 2, DN model (n = 9); and group 3, DN and treated with vitamin D receptor agonist (VDRA) paricalcitol (n = 9). DN was induced by intraperitoneal injection of streptozocin (STZ) (Sigma-Aldrich), which is dissolved in 10 mmol/L citrate buffer (pH 4.2) at 58 mg/kg body wt, while control rats (n = 9) only received 10 mmol/L citrate buffer solution. Two weeks after STZ injection, paricalcitol was administered at 0.4 μg/kg i.p. daily (n = 9) (23). Body weights were assessed every 2 weeks. At the end of the experiment and prior to sacrifice, blood samples and 24-h urine samples were collected in metabolic cages for biochemical studies, and the kidneys were harvested for protein or RNA extraction and histological analyses.

The C57BL/KsJ db/db, 6-week-old, male mice and age-matched db/m mice were acquired from GemPharmatech Co., Ltd (Nanjing, China). The mice were separated into three groups: group 1, db/m mice (n = 9); group 2, db/db mice (n = 9); and group 3, db/db mice treated with paricalcitol (n = 9). db/db mice were treated with paricalcitol (0.3 μg/kg body wt i.p.) (24), three times per week over a total of 12 weeks. At the end of the experiment and prior to sacrifice, blood samples and 24-h urine samples were collected in metabolic cages for biochemical studies, and the kidneys were harvested for protein or RNA extraction and histological analyses.

Fixed kidney specimens were embedded in paraffin and sectioned (3 μm thickness). Renal sections were stained with periodic acid Schiff (PAS). Immunohistochemistry of the sections was conducted with a methodology previously described (25). Primary antibodies include podocin, VDR, P62, and ZO-1. For semiquantitative analysis of the above indicators, the optical density and area were calculated by ImageJ to evaluate the integral optical density and calculate mean optical density. For determination of the number of podocytes, serial kidney sections were stained with WT1 antibody (1:50) and WT1-positive cells were counted, with the number calculated with use of the dissector/fractionator combination method. Histological analyses were performed by three independent nephrologists in a blinded manner.

Kidney tissue was fixed in 2.5% glutaraldehyde. After thorough washing in PBS, the samples were exposed to 1% osmium tetroxide, dehydrated through a series of grade dethanol. Then tissues were embedded and polymerized. Samples was cut into ultrathin sections of 50–70 nm and stained with uranyl acetate and lead citrate before analysis. Images were obtained from a Tecnai 10 transmission electron microscope (JEM-1010; JEOL, Tokyo, Japan).

Conditionally immortalized mouse podocytes (MPC5) were kindly donated by Professor Chuan-ming Hao (Huashan Hospital, affiliated with Fudan University). MPC5 were cultured in RPMI-1640 medium with 10% FBS, penicillin (100 units/mL), and streptomycin (100 mg/mL) at 33°C for proliferation and then cultured at 37°C for 10–14 days without interferon-γ for differentiation. Differentiated podocytes were made quiescent in medium that contained 0.1% FBS for 24 h, and the cells were then exposed to treatment for the indicated time periods.

For knockdown experiments, podocytes were transiently transfected with siRNA specifically targeting VDR, Atg3, or a negative control siRNA, respectively, using Lipofectamine 3000 (Invitrogen, Carlsbad, CA) according to the manufacturer’s instructions. Cells were transfected with 20 nmol/L VDR, Atg3, or control siRNA for 24 h before further treatment. Cellular protein was extracted and subjected to Western blot analysis for detection of VDR and Atg3.

Autophagy flux was analyzed based on the distribution and alteration of mRFP-GFP-LC3 fluorescence signals under fluorescence microscopy. mRFP-GFP-LC3 adenoviruses were provided by Hanbio Biotechnology Co., Ltd (Shanghai, China). Podocytes were inoculated with mRFP-GFP-LC3 adenoviruses for 24 h. The cells were washed with PBS, fixed with 4% paraformaldehyde for 10 min, and permeabilized with 0.5% Triton X-100 (Sigma-Aldrich) for 10 min. After that, the cells were stained with DAPI (Sigma-Aldrich) for 6 min. After the designated treatments, the cells were fixed with 4% paraformaldehyde and examined under fluorescence microscopy (Olympus, Tokyo, Japan). The number of mRFP, GFP, and GFP-mRFP dots were counted in five microscopic fields.

Immunofluorescence staining of podocytes was performed after incubation of mixed TdT and dUTP for 2 h with antibody against ZO-1 and p62, followed by incubation with a secondary antibody (Sigma-Aldrich). Cell nuclei were stained with DAPI. And then samples were visualized viewed with a Nikon Eclipse 80i Epi-fluorescence microscope equipped with a digital camera (DS-Ri1; Nikon).

Flow cytometer was used to assess the apoptotic rate of podocytes. Various groups of cells were collected and resuspended with phosphate buffer solution. Annexin V FITC (5 μL) (Thermo Fisher Scientific) and Annexin V PI (10 μL) (Thermo Fisher Scientific) were added, mixed uniformly to react for 15 min, and placed on ice, and the apoptotic rate was detected within 1 h by flow cytometer (LSRFortessa X-20; BD).

Western blot analysis was performed as previously described. The treated cells were lysed with radioimmunoprecipitation assay buffer containing a protease and phosphorylase inhibitor on ice for 30 min. Total proteins were obtained by centrifugation for 10 min (12,000 rpm, 4°C). The protein concentration was quantified with the BCA Protein Assay Kit according to the manufacturer’s instructions (Thermo Fisher Scientific). Then, 20 μg protein samples were subjected to SDS-PAGE and transferred onto polyvinylidene fluoride membranes (Millipore, Burlington, MA). After blocking with 5% nonfat dried milk for 2 h, the membranes were incubated with antibodies against cleaved Atg 3 (1:1,000), Atg7 (1:1,000), microtubule-associated protein 1 light chain 3 (LC3I, LC3II) (1:1,000), VDR (1:1,000), p62 (1:1,000), and ZO-1 (1:1,000).

Caspase activities were measured by Caspase Colorimetric Activity Assay Kit (APT129; Sigma-Aldrich, Darmstadt, Germany) and Caspase Colorimetric Activity Assay Kit (APT165; Sigma-Aldrich). Procedures were followed according to the manufacturer’s instructions.

Data are expressed as means ± SD. Comparisons between two groups were performed with Student t test, and multiple groups were compared with one-way ANOVA followed by Dunnett multiple comparison test. P < 0.05 was considered statistically significant. Statistical analyses were carried out with Prism 7.0 (GraphPad Software, San Diego, CA).

All data generated or analyzed during this study are included in the published article and in Supplementary Material.

To examine whether SD-TJ transition is involved in the glomeruli injury in DN patients, we conducted immunostaining in kidney biopsy specimens from 24 DN patients at stages IIa, IIb, III, and IV, respectively, and six normal human kidney tissues formed the control group (Fig. 1A). Clinical characteristics of DN patients and the control group are described in Supplementary Table 1. Podocytes foot processes of normal kidney glomeruli were separated by the filtration slit above the glomerular basement membrane and appeared to be single cross strands in the EM (Fig. 1B [CON-i and CON-ii]). Multiple membrane “fusion” or “kissing” points between these foot processes are characteristic of the TJ structure. Neighboring foot processes membrane were closely anastomosed in focal segments of glomeruli of DN patients (Fig. 1B [DN1-i through DN1-4i, black box; DN1-ii through DN4-ii, white circle]). Meanwhile, the expression of podocin (SD protein) was downregulated in the glomeruli of DN patients, whereas the TJ protein ZO-1 increased with progression of DN compared with that in the glomeruli of normal human kidneys (Fig. 1C–F). Interestingly, ZO-1 only exhibited an increasing trend from stage IIa to III and started to decrease from stage III to IV, which maybe be due to the increased extent of podocytes injury and loss of podocytes at the late stages (III and IV stages) of DN patients. This could be explained by the fact that the number of Wilms tumor protein (WT-1), a surrogate marker for podocytes number, was decreased in DN patients at stage III to IV compared with the control group (Fig. 1G and H). This is also further confirmed by the results of an increasing trend in average content of ZO-1 in podocytes (the ratio of ZO-1 abundance to podocytes numbers) of DN patients with the pathological progression from stage IIa to IV (Supplementary Fig. 1A).

P62 is the substrate of autophagic degradation, the accumulation of which indicates the obstruction of autophagic flux (26). In our study, accumulation of p62 protein was significantly increased in the glomeruli of DN patients at stage IIa to III (Fig. 1I and J) and for average content of p62 in podocytes of DN patients there was an increasing trend with the pathological progression from stage IIa to IV (Supplementary Fig. 1B). The expression trend of p62 was positively correlated with the changes of ZO-1 but negatively correlated with the changes of podocin in DN patients from stage IIa to IV (Fig. 1K and L), which indicated that impaired autophagic flux may be related to the SD-TJ transition of podocytes in DN patients.

Starvation was the most commonly used way to stimulate autophagy (26). To elucidate whether autophagy has a role in the degradation of ZO-1, we first starved cultured podocytes for 6 h and found a more robust colocalization of ZO-1 with the p62 in podocytes after starvation treatment, indicating that ZO-1 was degraded by the p62-mediated autophagic pathway (Fig. 2A). Conversely, robust accumulation of ZO-1 in the membrane of cultured podocytes was observed after blocking of autophagic flux with 3-MA or CQ (Fig. 2A). Western blot experiments showed trends similar to the change of ZO-1 in cultured podocytes after the treatment of starvation, 3-MA, or CQ, respectively (Fig. 2B and C). To further confirm the degradation of ZO-1 mediated by p62, we knocked down p62 with a specific siRNA in cultured podocytes (Fig. 2D). It was shown that ZO-1 was accumulated in the membrane of cultured podocytes when p62 was knocked down compared with results in the control group (Fig. 2E). Finally, immunoprecipitation results proved that there is a direct interaction between p62 and ZO-1, which was enhanced after the activation of autophagy by starvation, whereas this interaction was attenuated by blocking autophagic flux with the treatment with 3-MA or p62 siRNA (Fig. 2F and G). The above suggested that p62 mediated the autophagic degradation of ZO-1 in cultured podocytes.

It was shown that the apoptotic podocytes were significantly increased after treatment with either of the autophagic inhibitors 3-MA or CQ and that this could be attenuated by p62 siRNA (Fig. 3A and B). As expected, the activity of caspase8 and caspase3 was significantly increased after treatment with 3-MA or CQ, which could be reversed when p62 was knocked down (Fig. 3C and D). Furthermore, knocking down p62 reversed the decrease of podocin expression induced by 3-MA or CQ in cultured podocytes (Fig. 3E and F). These data established a strong causal relationship of podocytes injury mediated by p62 with activation of caspase3 and caspase8 signaling pathways.

Previous studies had suggested that VDR was involved in the regulation of autophagic flux and that VDR deficiency might contribute to podocytes injury in DN (18,21,22). However, whether VDR deficiency was correlated with impaired autophagic flux in podocytes is still unknown. To address this question, we firstly measured the VDR expression in the DN patients cohort and found that VDR expression was downregulated with the pathological progression (Fig. 4A and B). Meanwhile, VDR was positively correlated with the podocin expression at all stages of DN (Fig. 4C). In addition, VDR is negatively correlated with ZO-1 expression in the glomeruli of DN patients with the pathological progression from stage IIa to IV (Fig. 4D and E).

To further verify the causative effect of VDR on impaired autophagic flux, we selectively knocked down the expression of VDR with specific siRNA in cultured podocytes (Fig. 5A). mRFP-GFP-LC3 adenovirus was used to evaluate the level of autophagic flux in cultured podocytes. Under fluorescence microscope, yellow spots indicates autophagosomes in overlay images. Because GFP fluorescent protein is sensitive to acid, when autophagy and lysosome fuse, GFP fluorescence is quenched and only red fluorescence can be detected, which indicates autolysosomes (26). Interestingly, the number of autophagosomes reduced significantly after silencing of VDR, while the number of autolysosomes showed no obvious change, suggesting that autophagic flux is suppressed at an early stage (Fig. 5B–D). In addition, the expression of LC3II/I was decreased, accompanied by p62 accumulation, further confirming that VDR knockdown impaired autophagic flux in cultured podocytes (Fig. 5E and F).

Our results showed VDR deficiency blocked autophagic flux by affecting the transformation from LC3I to LC3II (Fig. 5E), while Atg7 and Atg3 genes were responsible for regulating this process to initiating the autophagic flux (Fig. 5G). We supposed Atg7 and/or Atg3 were responsible for the impaired autophagic flux due to reduced VDR expression. Interestingly, we found that Atg3 rather than Atg7 was decreased when VDR was knocked down in cultured podocytes. In fact, Atg7 was significantly increased when VDR was knocked down (Fig. 5H). In addition, overexpression of Atg3 reversed the impaired autophagic flux induced by knocking down VDR (Fig. 5I–M). Collectively, Atg3 rather than Atg7 was responsible for the impaired autophagy flux due to reduced expression of VDR in cultured podocytes.

Consistent with the in vivo results, high glucose (HG) treatment impaired autophagic flux as indicated by the decrease of autophagosomes and autolysosomes in cultured podocytes (Fig. 6A–C). In addition, HG treatment led to the decrease of VDR, Atg3, and LC3II/LC3I and the accumulation of p62 (Fig. 6D and E). Furthermore, HG blocked p62-mediated autophagic degradation of ZO-1 (Fig. 6F). What’s more, HG increased podocytes apoptosis and caspase3 and caspase8 activity and reduced podocin expression in cultured podocytes (Fig. 6G–L). However, VDRA treatment partially restored all the changes mentioned above (Fig. 6). These results suggested that VDRA treatment restored HG-induced impairment of autophagic flux, SD-TJ transition, and cellular injury in podocytes.

To further verify the findings in animal models, we evaluated a rat model with STZ-induced DN. All physiologic and biochemical data are presented in Supplementary Table 2. Rats with STZ-induced diabetes exhibited elevated blood glucose levels compared with the control group (P < 0.05). However, there was no improvement after VDRA treatment (P > 0.05). Rats with STZ-induced diabetes gained less weight than the control groups at the end of the experiment (P < 0.05). Similarly, this change was not reversed by VDRA treatment. Albuminuria was increased 6.3-fold in comparison with the control rats; however, VDRA treatment effectively attenuated the proteinuria in comparisons with the rats with STZ-induced diabetes (P < 0.05). Serum concentrations of creatinine did not differ among the three groups (P > 0.05) (Supplementary Table 2). Interestingly, rats with STZ-induced diabetes showed foot processes effacement and TJ-like changes similar to those in DN patients (Fig. 7A and B). We further analyzed the expression of VDR, podocin, p62, and ZO-1 by immunohistochemistry. It could be clearly seen that VDR and podocin were decreased in kidneys of rats with STZ-induced diabetes, while p62 and ZO-1 were increased. However, VDRA could reverse all these changes (Fig. 7C–F). In addition, VDRA treatment restored Atg3 expression (Fig. 7G and H) and reduced the caspase3 and caspase8 activity in the glomeruli of DN rats (Fig. 7I and J). Taken together, VDRA treatment improved autophagic flux and attenuated SD to TJ transition and podocytes injury in the DN rat model.

We then further confirmed our in vivo results via another db/db mice DN model. All physiologic and biochemical data are presented in Supplementary Table 3. db/db mice exhibited elevated blood glucose levels compared with the db/m group (P < 0.05); however, there was no improvement after VDRA treatment (P > 0.05). db/db mice showed a higher weight than the db/m groups at the end of the experiment (P < 0.05). Serum concentrations of creatinine did not differ among the three groups (P > 0.05). Albuminuria in db/db mice was increased 9.1-fold in comparison with db/m mice; however, VDRA treatment effectively attenuated the proteinuria compared with db/db mice (Supplementary Table 3). Interestingly, similar to the changes in the STZ rat model and DN patients, db/db mice showed foot processes effacement and TJ-like changes, while foot processes were separated by SD in wild-type mice (Fig. 8A and B). In addition, the expression of VDR and podocin was decreased, accompanied by the increase of p62 and ZO-1 in db/db mice compared with the wild-type mice (Fig. 8C–F), while VDRA could reverse all these changes. Furthermore, VDRA treatment restored Atg3 expression (Fig. 8G and H) and reduced caspase3 and caspase8 activity in the glomeruli of db/db mice (Fig. 8I and J). Collectively, VDRA treatment improved autophagic flux and attenuated SD to TJ transition and podocytes injury in the DN mice model.

In this study, we demonstrated that downregulating VDR/Atg3 axis signaling led to podocytes injury and SD-TJ transition via the p62-mediated autophagic pathway in DN (as summarized in Supplementary Fig. 2). These findings provide a novel insight about the effacement of foot processes, a typical early pathophysiological feature of DN, and may identify a therapeutic implication for VDRA in the treatment of early DN.

DN is one of the common complications of diabetes (27). Podocytes injury, manifested as foot processes effacement, abnormality of podocytes cytoskeleton protein and changes of SD, and decrease in podocytes numbers, appears in all stages of DN (28). Morphologically, effacement of foot processes (early stage) and disappearance of foot processes (late stage) were observed under EM in the glomeruli of DN patients (11,14). At the molecular level, the expression of SD-related proteins such as nephrin and podocin was decreased, which indicated the impairment of SD structure (29,30). While the loss of the foot processes due to the shedding of podocytes injury at the late stage is easily understood, the exact mechanism of foot processes effacement at the early stage of DN is still unclear.

TJ structure was found in the SD between podocytes by EM in 1961 (31). SD is an indispensable zipper-like structure between normal podocytes. Previous research proved that ZO-1, JAM-A, and occludin were important for TJ formation (6). ZO-1 is necessary for podocytes to maintain normal SD structure and function (32,33). ZO-1 was also reported to represent an increasing trend in puromycin aminonucleoside treatment–induced podocytes injury, where it was concentrated along the newly formed TJ and the remaining SDs discontinuously (8). Of note, enhanced TJ function and SD-TJ transition may facilitate podocytes injury (10). Our results also showed obvious SD-TJ transition in the glomeruli of kidney tissues from DN patients and diabetic animals (Figs. 1, 7, and 8). Hence, the SD-TJ transition may suggest a new mechanism for foot processes effacement at the molecular level in DN.

Proteostasis is conducted via sophisticated networks of mechanisms that act to maintain the quality of proteins and the evolutionary diversity of the biological functions of proteins (34). Autophagy is an essential mechanism to maintain the homeostasis in podocytes, and improvement of autophagic flux can alleviate podocytes injury in DN (14). Previous studies demonstrated that hyperglycemia impaired autophagic activity in podocytes by various pathways (35,36). Since a previous study reported that claudin-2 (TJ protein) is degraded by the autophagic pathway in intestinal epithelial cells (17), we proposed the hypothesis that the increasing formation of TJ under DN conditions is due to the blocked degradation of ZO-1 caused by impaired autophagy flux in podocytes. As expected, p62 mediated the autophagic degradation pathway of ZO-1 in cultured podocytes (Fig. 2), which explained the positive correlation between the abundance of p62 and ZO-1 in the renal tissues from DN patients (Fig. 1). Our results may provide a novel insight about the reason why TJ formation is increased in DN.

There has been a growing body of evidence confirming that impaired autophagic flux leads to podocytes injury due to increased oxidative stress and endoplasmic reticulum stress (37). P62 protein is an autophagy-specific degradation substrate. When autophagic flux occurs normally, p62 binds to ubiquitination degradation target protein and delivers the degradation substrate to autophagy-lysosome for degradation, thereby achieving intracellular molecular homeostasis (38,39). Previous studies have also shown that elevated p62 levels per se can activate caspase3 and caspase8 to cause cellular damage (40). However, it is unclear whether the accumulation of p62 can directly activate caspase3 and caspase8 in podocytes under HG or autophagy blocking conditions. In this study, we found that p62 knockdown could effectively reduce caspase3 and caspase8 activity and podocytes injury induced by either HG or autophagy inhibitors, suggesting that p62 accumulation caused by impaired autophagic flux might be one of the main mechanisms of SD-TJ transition.

VDR deficiency has been demonstrated to participate in autophagy regulation in various steps, including initiation, nucleation, and maturation (18). However, whether VDR is involved in the regulation of autophagic flux in podocytes has not been studied. During autophagy induction, the C-terminal of LC3-I is affected by Atg4/Atg7/Atg3 and modified by phosphatidylethanolamine to convert into LC3-II. The formation of LC3 could be blocked by an oxidation effect of Atg3 and Atg7 to impair autophagic flux (41). In cultured podocytes, silenced VDR could inhibit autophagy, especially LC3-I to LC3-II transition, which is conducted by downregulating Atg3 instead of Atg7 (Fig. 5). Besides, VDR agonist can upregulate Atg3 and enhance autophagy to reduce the SD-TJ transition and of VDR signaling against podocytes injury in DN (21,22), the underlying mechanism is far from clarified. This study is the first to demonstrate that the VDR/Atg3 axis participates in the SD-TJ transition by regulating autophagic flux in pathological progress of DN, which may provide new evidence for applying VDRA to treat early DN.

In conclusion, we here identified that VDR/Atg3 axis deficiency caused SD-TJ transition through the p62-mediated autophagic pathway, which will provide a novel insight for understanding podocytes injury in DN, while early administration of VDRA might be an important strategy to prevent the development of DN by inhibiting foot processes effacement.

Funding. This work was supported by grants from the National Natural Science Foundation of China (81720108007, 8203000544, 81670696) to B.-C.L. and B.W. (81700618, 82070735) and the Natural Science Foundation of Jiangsu Province (BK20181487) to B.W. This research was supported by additional grants from the National Key Research Programme of Ministry of Science and Technology (2018YFC130046, 2018YFC1314000) to B.-C.L.

Duality of Interest. No potential conflicts of interest relevant to this article were reported.

Author Contributions. B.W. and B.-C.L. contributed to the conception of the study. J.-y.Q., T.-t.T., L.-l.L., N.Y., and H.-l.G. performed the experiment. B.W. and B.-C.L. contributed significantly to analysis and manuscript preparation. B.W., J.-y.Q., and T.-t.T. performed the data analyses and wrote the manuscript. L.-L.L., Y.W., Z.-L.L., M.W., and J.-Y.C. helped perform the analysis with constructive discussions. B.-C.L. reviewed the manuscript. B.-C.L. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.