Excessive mitochondrial fission has been identified as the central pathogenesis of diabetic kidney disease (DKD), but the precise mechanisms remain unclear. Disulfide-bond A oxidoreductase-like protein (DsbA-L) is highly expressed in mitochondria in tubular cells of the kidney, but its pathophysiological role in DKD is unknown. Our bioinformatics analysis showed that tubular DsbA-L mRNA levels were positively associated with eGFR but negatively associated with Scr and 24h-proteinuria in CKD patients. Furthermore, the genes that were coexpressed with DsbA-L were mainly enriched in mitochondria and were involved in oxidative phosphorylation. In vivo, knockout of DsbA-L exacerbated diabetic mice tubular cell mitochondrial fragmentation, oxidative stress and renal damage. In vitro, we found that DsbA-L was localized in the mitochondria of HK-2 cells. High glucose (HG, 30 mM) treatment decreased DsbA-L expression followed by increased mitochondrial ROS (mtROS) generation and mitochondrial fragmentation. In addition, DsbA-L knockdown exacerbated these abnormalities, but this effect was reversed by overexpression of DsbA-L. Mechanistically, under HG conditions, knockdown DsbA-L expression accentuated JNK phosphorylation in HK-2 cells. Furthermore, administration of a JNK inhibitor (SP600125) or the mtROS scavenger MitoQ significantly attenuated JNK activation and subsequent mitochondrial fragmentation in DsbA-L-knockdown HK-2 cells. Additionally, the down-regulation of DsbA-L also amplified the gene and protein expression of mitochondrial fission factor (MFF) via the JNK pathway, enhancing its ability to recruit DRP1 to mitochondria. Taken together, these results link DsbA-L to alterations in mitochondrial dynamics during tubular injury in the pathogenesis of DKD and unveil a novel mechanism by which DsbA-L modifies mtROS/JNK/MFF-related mitochondrial fission.
Diabetic kidney disease (DKD) is a common microvascular complication of diabetes mellitus and the leading cause of end-stage renal disease [1,2]. Emerging evidence has shown that chronic hyperglycemia induces irreversible cell injury in the kidney and plays a pivotal role in the progression of DKD . Recent findings have extended our understanding of the pathogenesis of DKD beyond the glomeruli to the tubular compartment and support a role for tubular involvement in DKD [4–7], but the mechanisms initiating and driving tubular injury during this process have not been thoroughly explored.
Proximal tubular epithelial cells (PTECs) contain more mitochondria than any other cell type in the kidney due to the high energy demand . Hence, PTECs are vulnerable to impaired mitochondrial function . Besides, mitochondria are dynamic organelles that continuously undergo fission and fusion through a process called mitochondrial dynamics . Growing evidence from others and our group has suggested that disruption of the dynamic balance, especially a shift toward fission, contributes to mitochondrial dysfunction in tubular cells in DKD [11–16]. Moreover, increased mitochondrial reactive oxygen species (mtROS) may cause mitochondrial dysfunctional signaling resulting in changes in mitochondrial dynamics . Therefore, identifying molecules specifically in tubular cells that inhibit mitochondrial fragmentation may permit the development of new strategies to prevent diabetic tubulopathy.
DsbA-L, originally named glutathione S-transferase (GST)-k , is a 25 kDa protein that is expressed highly in adipocytes, the liver and the kidney [18,19] and its localization is consistent with the region rich in mitochondria . Liu et al. demonstrated that DsbA-L participates in adiponectin multimerization and degradation [20–22]. Besides, DsbA-L is highly expressed in the tubular cells of murine and human kidneys, especially in proximal tubules [18,23]. However, the pathophysiological role of DsbA-L in the kidney and mitochondria remains unclear. Our recent study found that DsbA-L was significantly decreased in the kidney in high fat diet combined with streptozotocin (STZ)-induced diabetic mice, and knockout of DsbA-L further aggravated renal ectopic lipid disposition and lipid-related tubular injury in DKD . On the other hand, consistent with its subcellular distribution in tubular cells, prior studies have suggested that DsbA-L deficiency leads to mitochondrial dysfunction, including an increase in mtROS generation, mitochondrial DNA release and a concurrent decrease in ATP production, in hepatocytes and adipocytes [24,25]. However, the role of DsbA-L in mitochondrial dysfunction in renal tubular cells in DKD and its underlying mechanisms remain to be elucidated.
DRP1 is essential for mitochondrial fission in mammalian cells . Our previous studies reported that activation of DRP1/FIS1-mediated mitochondrial fission contributes to diabetic tubular injury [13,15]. In addition to Fis1-mediated fission, mitochondrial fission is also regulated by mitochondrial fission factor (MFF) . Phosphorylation of MFF recruits DRP1 assembly onto the surface of mitochondria ensuring mitochondrial fission. This process is required for mitochondrial fragmentation-mediated mitochondrial dysfunction in renal mesangial cells, podocytes and cardiac endothelial cells [28–30], but its role in tubular cells remains elusive. Besides, it has been reported that the JNK pathway is involved in modulating mitochondrial morphology mainly via MFF [28,30–32]. Of note, Liu et al. also found that DsbA-L down-regulation activated the JNK pathway through mtROS in hepatocytes . Based on this evidence, we hypothesized that DsbA-L may modulate mitochondrial dysfunction through ROS/JNK/MFF-mediated fission.
In the present study, we found that the expression of DsbA-L was decreased in the kidneys of STZ-induced diabetic mice. Furthermore, DsbA-L deficiency accentuated tubule mitochondrial fragmentation and tubular damage under a diabetic state. At the molecular level, DsbA-L was localized in the mitochondria. Knockdown of DsbA-L expression increased mtROS generation, which ultimately activated JNK/MFF-mediated mitochondrial fission, consequently inducing mitochondrial dysfunction and driving the development and progression of DKD.
Materials and methods
Antibodies, plasmids and other reagents
The following antibodies were used: Phosphorylated-JNK (CST, #9251), JNK (Abcam, ab76125), MFF (CST, #84580), phosphorylated-MFF (CST, #49281), DRP1 (Abcam, ab56788), DsbA-L (GeneTex, GTX82705), DsbA-L (Abcam, ab134173), KIM-1 (R&D Systems, AF1817), β-actin (Abcam, ab8226), MFN2 (Proteintech, 12186-1-AP), MFN1 (Proteintech, 13798-1-AP), FIS1 (Proteintech, 10956-1-AP). Calcein-acetoxymethyl ester (Calcein-AM) was purchased from Invitrogen. Cobalt chloride (CoCl2) was obtained from Aladdin (Shanghai, China). The DsbA-L plasmid was obtained from Dr Liu Feng at Central South University, Changsha, China. The DsbA-L siRNA and scramble control were purchased from RiboBio (Guangzhou, China). The sequences for DsbA-L siRNA and scramble control were 5′-CCTGTGCCGGTATCAGAAT-3′ and 5′-TCAGCACGGTGACTGAGAC-3′, respectively. Lipofectamine® 3000 was purchased from Invitrogen (Carlsbad, CA, U.S.A.).
To ensure that the role of DsbA-L was relevant to human kidney disease, we examined the correlations between the gene expression of DsbA-L and renal function index (eGFR, Scr and 24 h-proteinuria) by taking advantage of publicly available data sets from the Nephroseq database (https://www.nephroseq.org/resource/login.html, V5), which is a platform of comprehensive renal disease gene expression data sets . Within Nephroseq, the renal DsbA-L distribution pattern was based on the Higgins et al. study, which contained 5 normal kidney tissue samples that were obtained from areas of the nephrectomy specimens that were uninvolved in neoplasm . Correlation analysis between tubular DsbA-L mRNA expression and renal function index, such as estimated glomerular filtration rate (eGFR), serum creatinine (Scr) and 24-h proteinuria, were based on the following datasets. The dataset of Woroniecka et al. contained 10 tubular tissue samples from DKD patients. The dataset from the studies by Reich and Ju contained 24 and 25 tubular tissue samples from patients with IgA nephropathy (IgAN), respectively. The dataset of the ERCB study contained 7 tubular tissue samples from patients with lupus nephritis (LN) and 12 samples from patients with focal segmental glomerular sclerosis (FSGS). The detailed clinical features of the above patients are available from the Nephroseq database.
For gene coexpression analysis, normal human renal RNAseq data were extracted from Genotype-Tissue Expression (GTEx)  and applied for gene coexpression network construction . Correlations with Pearson r > 0.6 and P < 0.05 were subjected to GO analysis via DAVID online tools (Version 6.8) . The goal was to link specific transcripts (i.e. DsbA-L) to their coregulated gene set to implicate their molecular functions in the kidney.
Animal experiments and procedures
All animal procedures used in the present study were approved by the Animal Care and Use Committee of Central South University (Ethical Approval No. 2017-S112). DsbA-L−/− mice were generated in our laboratory as previously described . Animals were housed in a specific pathogen-free animal facility at Central South University, with a 12-h light/12-h dark cycle and ad libitum access to food and water. DsbA-L−/− mice and their wild-type (WT) littermates were injected intraperitoneally with STZ (50mg/kg body weight in citrate buffer, pH 4.5; Sigma-Aldrich, St. Louis, MO, U.S.A.) at 8 weeks of age for 5 days to induce hyperglycemia. One week after the injection, mice with blood glucose levels >16.7 mmol/l were selected for the experiment and sacrificed after 12 weeks. At the end of the last week, all mice were anesthetized with an intraperitoneal injection of 50 mg/kg body weight sodium pentobarbital and killed by exsanguination. The urinary β-NAG levels were measured with a NAG assay kit (Jiancheng Bioengineering, China).
Morphological analysis of the kidney
Mice renal biopsy tissues were fixed in 4% paraformaldehyde and embedded in paraffin. Three-μm-thick sections were used for periodic acid Schiff (PAS) and Masson staining. Glomerular and tubular damage were scored as previously described : (0), no glomerular/tubular damage; (1) 25% of the glomerular/tubular area affected; (2) 25–50% of the glomerular/tubular area affected; (3) 50% of the glomerular/tubular area affected.
Immunofluorescence staining and TUNEL assay
The frozen sections were fixed with 4% paraformaldehyde for 10 min at room temperature. Next, the sections were blocked with 0.1%TritonX-100 and 5% BSA in PBS buffer for 60 min at room temperature and then incubated with primary antibody overnight at 4°C. After washing with PBS, the sections were re-incubated with secondary antibodies conjugated with Alexa Fluor 488 (green) or 594 (red) for 1 h. The slides were counterstained with DAPI to visualize the nuclei and viewed with an LSM 780 META laser scanning microscope (Zeiss, Thornwood, NY). To assess the extent of apoptosis, TUNEL staining was performed on cryostat kidney sections using an In-Situ Cell Death Detection Kit (Roche Applied Science). Ten random fields of cells were counted to determine the number of cells undergoing apoptosis.
Transmission electron microscopy and analysis of mitochondrial morphology
A tissue block of approximately 1 mm3 from each kidney was fixed in 4% paraformaldehyde and 2.5% glutaraldehyde in phosphate buffer (pH 7.4) and postfixed with 1% osmium tetroxide. Then, the fixed tissue blocks were dehydrated and embedded in Epon 812. Ultrathin sections were stained with uranyl acetate and lead citrate and subsequently examined under a transmission electron microscope (TEM). Six images of proximal tubule epithelial cells per kidney section were randomly selected for each mouse (n = 5 per group) at ×15,000 magnification. The morphology of each mitochondrion was described by the aspect ratio (AR) using ImageJ (National Institutes of Health, Bethesda, MD). The aspect ratio was calculated as the lengths of major axes/lengths of minor axes. A decrease in the AR of mitochondria indicates mitochondrial fragmentation.
Cell culture studies
The renal human proximal tubular epithelial cell line HK-2 was obtained from ATCC and cultured in low glucose DMEM/F12 supplemented with 5% FBS (Gibco), 100 units/ml penicillin, and 100 μg/ml streptomycin. Media containing 30 mM d-glucose were used to mimic the high glucose damage. SP600125 (SP, 10 μM, Selleck Chemicals) were used 2 h before HG treatment to suppress JNK pathway. Mitoquinone mesylate (MitoQ, 1 μM, MCE) was used to scavenge the HG-induced mtROS production.
Western blot assay
Western blot analysis of specific protein expression was performed according to an established procedure . Briefly, kidney tissues or cultured cells were lysed in RIPA buffer (CoWin Biosciences, China). The homogenates were centrifuged at 4°C, 13000 rpm for 15 min and then the supernatants were collected. BCA Protein Assay Kit (CoWin Biosciences, China) was used to determine the protein concentration. Then 5× SDS loading buffer (CoWin Biosciences, China) was added and boiled for 10 min, cooled on ice, and 20 μg protein was subjected to SDS-PAGE. Quantification was performed by measuring the intensity of the bands with ImageJ software.
Quantitative real-time PCR
Total RNA was extracted from HK-2 cells using TRIzol™ reagent (Invitrogen) according to the manufacturer’s protocol. The following primers were used: MFF (human): sense: 5′-GGAGAGGATTGTTGTAGCAGGA-3′, antisense: 5′-TGTTGTGGCGATGGTGTCA-3′; β-actin (human): sense: 5′-TCGTGCGACATTAAGGAG-3′, antisense: 5′-GATGTCCACGTCACACTTCA-3′. Relative expression levels of MFF were calculated with the 2−ΔΔCt method and normalized to that of β-actin.
Assessment of mitochondrial morphology and mitochondrial colocalization
Briefly, live cells were incubated with MitoTracker Red (500 nM, Molecular Probes, Invitrogen) at 37°C for 20 min and examined by confocal microscopy. For each treated group, 10 fields of cells were randomly selected (100 cells/group) to delineate the shape of mitochondria. Mitochondria were subjected to the “Analyze Particles” function of ImageJ for acquiring the aspect ratio. Lower values of AR represent mitochondrial fragmentation under high glucose conditions. For mitochondrial colocalization assessment, cells were first incubated with MitoTracker Red and then were fixed and permeabilized as described above. They were successively incubated with primary antibody and secondary antibodies conjugated with Alexa Fluor 488 (green) or 594 (red). The degree of colocalization was calculated by the colocalization highlighter plugin and Mander’s coefficient plugin (ImageJ).
Measurement of mitochondrial permeability transition pore (mPTP)
mPTP opening was assessed using calcein acetoxymethyl ester (Calcein-AM) staining with CoCl2 quenching assay according to a previous study . Briefly, HK-2 cells were incubated with 1 μM Calcein-AM and 4 μM CoCl2 at 37°C for 15 min. Images were examined by confocal microscopy. Calcein-acetoxymethyl ester (Calcein-AM) dye is loaded into the cytosol and mitochondria via diffusion and is activated upon ester bond cleavage by esterase (green staining). Under normal conditions, CoCl2 enters the cell but not the mitochondrial matrix, and thereby quenches calcein fluorescence in cytosol. However, when mPTP opening is increased, CoCl2 enters the mitochondrial matrix, ultimately leading to the dissipation of mitochondrial calcein fluorescence. The mean fluorescence intensity (MFI) of Calcein-AM staining was calculated by ImageJ and reflects the mPTP opening, and reduced calcein MFI indicated increased mPTP opening.
Mitochondria isolation and detection of ROS
Mitochondria were isolated from HK-2 cells using a Mitochondria Isolation Kit (Beyotime Biotechnology, China) and all isolation steps were performed on ice. A fraction of the mitochondria was subjected to protein extraction and SDS-PAGE. Cells were also incubated with MitoSox (5 μM, Molecular Probes, Invitrogen) for 15 min at 37°C to assess mtROS production, and then examined by confocal microscopy. Intracellular ROS production in renal tissues was assessed using 6-μm-thick unfixed cryostat sections and stained with cell-permeable agent dihydroethidium (1 μM, DHE, Sigma-Aldrich) in the dark for 20 min at room temperature. For both cells and tissue sections (glomeruli excluded), 10 randomly selected fields were photographed, and the MFI was calculated by ImageJ.
Statistical analyses were performed using SPSS (21.0) and GraphPad Prism software (8.0). Data were expressed as the mean ± SD. Differences between two groups were tested using a two-tailed Student’s t-test. One-way analysis of variance (ANOVA) was used for multiple independent sample analysis. Comparisons between groups were performed with Tukey’s test post hoc analysis. Correlation analyses were carried out using Spearman’s correlation analysis. For all tests, P value less than 0.05 was considered statistically significant.
Reduced gene expression of DsbA-L is associated with CKD progression
The Nephroseq online tool was used to assess the clinical significance of the DsbA-L gene in various kinds of chronic kidney disease (CKD). As shown in Supplementary Figure S1A, DsbA-L mRNA was more highly expressed in the renal cortex, where the proximal tubule accounts for most of the cortical mass , in normal human kidneys. Further correlation analysis showed that the expression levels of tubular DsbA-L mRNA negatively correlated with 24 h-proteinuria in patients with focal segmental glomerulosclerosis (FSGS) (r = -0.422, P < 0.05) and lupus nephritis (LN) (r = -0.808, P = 0.0228) (Supplementary Figure S1B) and was negatively associated with serum creatinine (Scr) in patients with IgA nephropathy (IgAN) (r = -0.526, P = 0.0069) (Supplementary Figure S1C). In contrast, tubular DsbA-L mRNA levels were positively correlated with the estimated glomerular filtration rate (eGFR) in patients with DKD (r = 0.692, P = 0.0265), IgAN (r = 0.502, P = 0.0124) and LN (r = 0.953, P = 0.0009), respectively (Supplementary Figure S1D). Taken together, these results indicate that reduced tubular DsbA-L expression participates in the progression of CKD.
GO enrichment analyses of genes that are coexpressed with DsbA-L in normal human renal tissues
To better understand the role of the DsbA-L gene in human renal tissue under physiological conditions, we applied a coexpression strategy using the GTEx database. To this end, 1787 genes that were coexpressed with DsbA-L (r > 0.6, P < 0.05) were captured for GO term analysis using the DAVID online database. GO cellular component analysis showed that 1421 of the 1787 identified genes were predicted to be mitochondrial proteins (Supplementary Figure S1E,a). GO analysis of coexpressed genes demonstrated enrichment in pathways that were mainly related to metabolic pathways and oxidative phosphorylation (Supplementary Figure S1E,b). We further annotated these genes based on biology process. As expected, an important proportion of the identified genes were involved in mitochondrial translation, oxidation-reduction process and mitochondrial respiratory chain assembly (Supplementary Figure S1E,c). These findings collectively suggest that DsbA-L might play a critical role in maintaining mitochondrial homeostasis.
DsbA-L deficiency worsens streptozotocin (STZ)-induced diabetic kidney injury
To investigate the consequences of DsbA-L down-regulation in DKD, we generated DsbA-L knockout (DsbA-L−/−) mice and intraperitoneally injected them with STZ to induce a diabetic state. Confocal microscopic images delineated that the expression of DsbA-L was mainly localized in cortical tubular cells but not in glomeruli, and decreased fluorescence intensity of DsbA-L was observed in diabetic mice compared with that of wild-type (WT) mice (Figure 1A, c vs a). In addition, no positive staining was observed in DsbA-L−/- mice (Figure 1A, b and d), and Western blot analysis further validated the above findings (Figure 1B,C), confirming the efficient depletion of DsbA-L. As shown in Figure 1D–F, little difference in body weight, blood glucose and urine β-NAG level was observed between DsbA-L−/− mice and WT mice under physiological conditions, but the body weight of WT and DsbA-L−/− mice was higher than that of STZ-induced diabetic mice, and this effect was further accentuated in diabetic DsbA-L−/− mice. In contrast, the blood glucose and urine β-NAG levels increased in both WT and DsbA-L−/- diabetic mice, and the increases in blood glucose and β-NAG levels were greater in diabetic DsbA-L−/−mice. In addition, diabetic conditions induced significantly increased levels of kidney injury molecule-1 (KIM-1), a biomarker of tubular damage, in DsbA-L−/− mice compared with those of WT mice (Figure 1G,H), although KIM-1 expression in the WT and DsbA-L−/− control groups was low and comparable. Moreover, histopathological analysis showed that DsbA-L−/− mice exhibited normal renal morphology under light microscopy compared with their WT littermates (Figure 1I, a vs b and e vs f). However, notable tubular epithelial disruption, hypertrophy of glomeruli, increased mesangial matrix and tubulointerstitial fibrosis were observed in the diabetic mice (Figure 1I, c and g), and these changes were further exacerbated in the kidneys of diabetic DsbA-L−/− mice (Figure 1I, d and h). Quantitative analysis of kidney damage scores indicated that diabetic DsbA-L−/- mice had slightly increased glomerular damage (Figure 1J) but a notable increase in tubular damage compared with the diabetic mice (Figure 1K). Taken together, these in vivo results support a renoprotective role of DsbA-L in diabetic tubular injury
Effects of DsbA-L deficiency on renal function and pathologic changes in the kidneys of STZ-induced diabetic mice
DsbA-L deficiency accentuates mitochondrial fragmentation in the renal tubules of STZ-induced diabetic mice
Next, we examined the changes in mitochondrial morphology in tubular cells. TEM analysis revealed that the mitochondria appeared in tubular shape in the tubules of WT and DsbA-L−/- mice (Figure 2A, a and b). However, the mitochondria in tubular cells of the diabetic mice exhibited round and scattered shapes, indicating mitochondrial fragmentation (Figure 2A, c). This fragmented appearance was confirmed by assessment of the mitochondrial aspect ratio (AR), which was decreased to approximately 2.4 (Figure 2B). This phenotypic change was further accentuated in diabetic DsbA-L−/− mice (Figure 2A, d), and the AR value of the fragmented mitochondria was decreased to 1.4 (Figure 2B). These findings suggested that loss of DsbA-L resulted in an accumulation of fragmented mitochondria in renal tubular cells in a diabetic state. Mitochondrial morphology is governed by several mitochondrial fission and fusion proteins . To elucidate the regulation of DsbA-L, we first examined its effects on several well-documented mitochondrial shape proteins. As shown in Figure 2C, the mitochondrial profusion proteins MFN1 and MFN2 were significantly reduced, and profission proteins DRP1, FIS1, and MFF were notably increased in diabetic mice compared with WT and DsbA-L−/− mice. Moreover, knockout of DsbA-L further decreased the expression of MFN2 and increased the expression of DRP1 and MFF in diabetic mice but had little effect on the MFN1 and FIS1 expression. Additionally, we found that DsbA-L deficiency up-regulated the expression of phosphorylated-MFF (p-MFF) in diabetic mice. In addition, we observed JNK activation in diabetic mice as evidenced by the increased levels of phosphorylated JNK, and the loss of DsbA-L further exacerbated JNK phosphorylation in the diabetic state (Figure 2C, f). Collectively, we hypothesize that DsbA-L deficiency promotes mitochondrial fission in the diabetic milieu through the concordance of DRP1 and MFF.
DsbA-L deficiency accentuates mitochondrial fragmentation, oxidative stress and tubular apoptosis in the kidneys of STZ-induced diabetic mice
Since ROS cause mitochondrial damage and dysfunctional signaling, resulting in changes in mitochondrial dynamic and cellular apoptosis , we determined ROS and apoptosis levels within kidney tissues by dihydroethidium (DHE) and TUNEL staining, respectively. Under normal conditions, both WT and KO mice had low and comparable intensities of DHE staining (Figure 2D, a vs b) and the number of TUNEL positive cells in tubules (Figure 2D, e vs f). However, KO and WT mice showed a dramatic increase in DHE signals (Figure 2D, c and d) and the number of TUNEL positive cells (Figure 2D, g and h) following STZ induction, and quantification analysis demonstrated that the DsbA-L−/- mice had significantly stronger DHE signals (Figure 2E) and more apoptotic cells (Figure 2F) than WT mice under diabetic conditions.
DsbA-L knockdown promotes mitochondrial ROS generation and mitochondrial fragmentation, which are dampened by DsbA-L overexpression in HK-2 cells subjected to high glucose
DsbA-L is localized in mitochondria . By confocal imaging, we observed a noticeable colocalization of DsbA-L with the mitochondrial marker MitoTracker (Red) in HK-2 cells, as highlighted by the white color, under LG ambience. However, HG (30 mM D-glucose) treatment for 24 h led to a decrease in fluorescence intensity of DsbA-L (Figure 3A) and further reduced the colocalization of DsbA-L with mitochondria, as shown by a significant reduction in the Mander’s colocalization coefficient (Figure 3B). To further confirm the subcellular localization of DsbA-L, we isolated the mitochondrial fraction from HK-2 cells. Western blot assays showed that DsbA-L was found in mitochondria and was decreased following HG treatment; the mitochondrial-specific marker complex IV (COX IV) served as a loading control (Figure 3C,D). Given the previous bioinformatics analysis showing that DsbA-L plays a key role in mitochondrial homeostasis, we aimed to determine whether knockdown or overexpression of DsbA-L in HK-2 cells affected mitochondrial ROS (mtROS) production, which is a biomarker of mitochondrial dysfunction in diabetic kidneys . The transfection efficiency was determined by Western blot assay (Figure 3C,D). As expected, a notable increase in mtROS generation was observed in HK-2 cells subjected to HG, as indicated by MitoSox staining. In addition, we observed that upon stimulation with HG, mitochondria became short and fragmented as indicated by decreased aspect ratio values, whereas knockdown of DsbA-L expression significantly increased mtROS levels (Figure 3E (a) and F), concurrent with a significant increase in mitochondrial fragmentation (Figure 3E (b) and G), in hyperglycemic conditions. We further observed that increased mtROS generation and the fragmented mitochondrial phenotype were reversed by overexpression of DsbA-L (DsbA-L O/E) in HK-2 cells (Figure 3E–G). Since mtROS are a key regulator of mitochondrial dynamics, we next tested the possibility that DsbA-L affects mitochondrial fission by promoting mtROS generation using a specific mtROS scavenger MitoQ. Simultaneous analysis of measured mitochondrial superoxide production and mitochondrial aspect ratio revealed that treatment with MitoQ reduced mtROS overproduction (Figure 3E (a) and F) and subsequently rescued the mitochondrial fragmentation (Figure 3E (b) and G) in DsbA-L knockdown HK-2 cells. Taken together, these results suggest that the effect of DsbA-L on mitochondrial fragmentation is dependent on the overgeneration of mtROS.
DsbA-L is localized in mitochondria and mediates HG-induced mitochondrial ROS production and fragmentation in HK-2 cells
Knockdown of DsbA-L exacerbates mitochondrial fragmentation by activating the mtROS/JNK pathway in HK-2 cells subjected to HG
Next, we examined whether the activation of JNK was mediated by mtROS as previously reported in hepatocytes . We pretreated DsbA-L knockdown cells with MitoQ, a selective mtROS scavenger. JNK activation was partially abolished by MitoQ (Figure 4A,B), indicating that DsbA-L deficiency induced JNK activation was mediated by mtROS. Furthermore, to explain whether JNK activation was implicated in DsbA-L deficiency-mediated mitochondrial fission, a JNK inhibitor SP600125 (SP) was used. Subsequently, mitochondrial morphology was observed. Mitochondria in DsbA-L knockdown HK-2 cells further divided into several fragments compared with those of the HG group, and this effect was negated by the JNK inhibitor (Figure 4C,D). These results indicate that quenching mtROS production could successfully alleviate JNK activation in DsbA-L-knockdown tubular cells and improve mitochondrial fragmentation.
Knockdown of DsbA-L promotes mitochondrial fragmentation via the mtROS-JNK pathway in HK-2 cells subjected to high glucose
DsbA-L modulates MFF-mediated mitochondrial fragmentation via JNK
To investigate the mechanisms by which DsbA-L/JNK regulates mitochondrial fission, we focused on MFF, based on our in vivo results. Consistent with this, we found that HG induced a significant increase in MFF expression and that knockdown of DsbA-L further increased MFF expression, whereas inhibition of JNK via SP reversed the up-regulation of MFF expression in DsbA-L knockdown HK-2 cells under HG ambience (Figure 5A,C). Activation of JNK contributes to MFF mRNA transcription in cardiocyte under ischemia-reperfusion conditions , and we next examined MFF expression at the transcriptional level. The qPCR results shown in Figure 5E demonstrated that MFF mRNA expression was significantly increased in response to HG stimulation and that DsbA-L knockdown further increased MFF mRNA levels. However, this effect was mostly rescued by treatment with the JNK inhibitor, suggesting that JNK activation contributes to MFF transcription and that DsbA-L deficiency under HG conditions exacerbates mitochondrial fragmentation that occurred at least partially through JNK-mediated MFF transcription.
Knockdown of DsbA-L promotes MFF-dependent mitochondrial fission through the JNK pathway in HK-2 cells subjected to high glucose
Notably, it has been reported that phosphorylation of MFF increases its interactions with DRP1 . Western blot assays showed that HG treatment increased p-MFF expression and that knockdown of DsbA-L further increased HG-induced MFF phosphorylation, whereas inhibition of JNK by SP abrogated this effect (Figure 5A,B). Considering that the p-MFF levels were unchanged relative to the MFF levels (Figure 5D), we believe that the increase in p-MFF was attributed to the increased expression of total MFF. Through confocal immunofluorescence assay, we found that HG treatment markedly enhanced the interactions between DRP1 and its receptor p-MFF, which was concomitant with increased translocation of DRP1 from the cytoplasm to mitochondria (Figure 5F,G). In addition, accumulation of DRP1 in the mitochondrial fraction was also validated by Western blot analysis (Figure 5H,I). These effects were further exacerbated in DsbA-L knockdown HK-2 cells and were alleviated by JNK inhibitor (SP) treatment (Figure 5F–I). Taken together, these results indicated that the negative effect of down-regulation of DsbA-L on mitochondrial fragmentation may be through the upregulation of MFF and consequentially, increased DRP1 and p-MFF interactions.
We provide in vivo and in vitro evidence that DsbA-L down-regulation in tubular cells impairs mitochondrial function. First, we found that reduced tubular DsbA-L expression was associated with the progression of CKD in patients. Second, DsbA-L expression was decreased in the kidneys of diabetic mice and the absence of DsbA-L in mice enhanced diabetic tubular injury and mitochondrial dysfunction. Third, our in vitro study showed that DsbA-L was localized in the mitochondria; down-regulation of DsbA-L expression promoted mtROS-induced JNK activation, which contributed to MFF gene transcription, consequently leading to mitochondrial fragmentation in tubular cells. These results indicate that DsbA-L plays a key role in maintaining mitochondrial homeostasis and participates in the tubular damage in DKD.
The importance of tubular injury has been increasingly emphasized, and damage to tubular cells has been strongly associated with a reduction in kidney function and the development of DKD [4,6,7]. In addition, in the clinic, a portion of DKD patients with progressive renal insufficiency do not exhibit microalbuminuria, which may be due to selective tubular injury . Based on these perspectives, it is believed that tubular damage plays a key role in DKD-associated damage. Mitochondrial dysfunction, which is characterized by excessive mitochondrial fission, has been postulated to play a central role in tubular damage in DKD. Thus, identifying molecules that modulate mitochondrial dynamics is imperative.
DsbA-L is an abundant tubular antioxidant protein, particularly in proximal tubular cells, and is located in mitochondria [18,23]. However, the role of DsbA-L in the pathogenesis of DKD is less well-known. In the present study, we first observed that the reduced tubular DsbA-L mRNA expression was strongly associated with the renal functional index (i.e. eGFR, Scr and 24-h proteinuria) in patients with CKD, including DKD (Supplementary Figure S1). Furthermore, we confirmed that DsbA-L was decreased in the kidneys of diabetic mice, and DsbA-L deficiency aggravated renal damage (Figure 1). More remarkably, diabetic DsbA-L−/− mice had increased levels of tubular damage biomarkers (i.e. β-NAG and KIM-1) and notable increases in tubular damage scores (Figure 1). These data, together with DsbA-L specific tubular localization, indicated that DsbA-L is as an important player in mediating diabetic tubular injury. These findings also confirmed the role of tubular damage in the development and progression of DKD.
Notably, we found that the blood glucose levels were triflingly but significantly increased in the DsbA-L -/- diabetic group compared with those in the WT diabetic group (Figure 1E), indicating that a hyperglycemia-dependent mechanism of kidney damage is induced by systemic DsbA-L deficiency. However, the increase in blood glucose levels between WT mice and diabetic mice (8.2 ± 0.3 vs 20.4 ± 3.0) was significantly higher than that between diabetic WT mice and diabetic DsbA-L−/− mice (20.4 ± 3.0 vs 22. 7 ± 4. 2), thus we hypothesized that there existed a hyperglycemia-independent mechanism of kidney damage is induced by systemic DsbA-L deficiency. Recent studies reported by Liu and colleagues found that conditional knockout of DsbA-L in hepatocytes and adipocytes led to mtROS overproduction, mtDNA release, and reduced ATP production [24,25]. Each of these processes is strongly affected by alterations in the balance of mitochondrial dynamics, especially mitochondrial fragmentation. Thus, we hypothesized that renal DsbA-L protects against renal damage in DKD by ameliorating mitochondrial fragmentation. To test this hypothesis, we applied a coexpression strategy using the GTEx database and DsbA-L knockout mice to evaluate the functional role of DsbA-L in the kidney. We found that 1421 of the 1787 identified genes that were coexpressed with DsbA-L were predicted to be mitochondrial proteins and were mainly related to metabolic pathways and oxidative phosphorylation (Supplementary Figure S1E). In vivo, knockout of DsbA-L enhanced tubular cell mitochondrial fragmentation under the diabetic state (Figure 2). In vitro experiments further showed that knockdown of DsbA-L expression exacerbated HG-induced mitochondrial fragmentation, and this effect was reversed by DsbA-L overexpression in HK-2 cells (Figure 3). These data suggest that loss of DsbA-L under diabetic conditions leads to renal damage that occurs at least partially through perturbation in mitochondrial function. Besides, electron microscopy analysis of the kidney showed no changes in mitochondrial morphology with DsbA-L deficiency under normal conditions (Figure 2A, b). In addition, we found that the kidneys of DsbA-L−/- mice were phenotypically normal under light microscopy (Figure 1I, b and f), and renal oxidative stress and apoptosis were not evident (Figure 2D, b and f). These observations were consistent with the consequences of the normal mitochondrial phenotype that was observed in the kidneys of DsbA-L−/− mice. These findings also received support from the other DsbA-L global knockout mice and liver specific knockout mice [19,24]. Based on these observations, we hypothesized that DsbA-L is a stress-response protein that alters the injury response and is responsible for DKD development.
Next, we explored the mechanisms by which DsbA-L regulates mitochondrial fission under HG ambience. ROS cause structural and functional damage to mitochondria, including fragmentation. The redox enzymes p66Shc or MIOX triggers mitochondrial fragmentation via control of mtROS generation [13,15]. Consistent with this, in present study, we observed that knockdown DsbA-L expression accentuated HG-induced mtROS generation and mitochondrial fragmentation; blocking mtROS overproduction by MitoQ rescued mitochondrial fragmentation in DsbA-L knockdown HK-2 cells (Figure 3), suggesting that DsbA-L modulates mitochondrial fission via mtROS. Additionally, elevated mtROS activates the JNK pathway in hepatocyte DsbA-L deficient mice . In the present study, our data also demonstrated that under a diabetic state, DsbA-L deficiency activated the JNK pathway in the renal cortex (Figure 2). In vitro, knockdown of DsbA-L expression further increased JNK activation in HK-2 cells and quenching mtROS production with MitoQ inhibited JNK activation and subsequent mitochondrial fragmentation; likewise, this trend was also observed by inhibition of JNK activation (Figure 4). Collectively, these data revealed that mtROS overgeneration induced by DsbA-L down-regulation was responsible for JNK activation-mediated mitochondrial fragmentation.
MFF-mediated fission mainly relies on JNK activation in various tissues [30,32]. In the present study, we found that JNK activation increased MFF expression at the gene and protein level (Figure 5A,E). These data were consistent with the results observed in cardiomyocytes as reported by Zhou and colleagues. They found that phosphorylated JNK translocated into the nucleus to promote MFF gene expression . Thus, we believe that JNK activation is at least involved in the modulation of MFF expression at the transcriptional level. Further studies are needed to investigate which specific JNK-targeted transcription factor initiates transcription of the MFF gene in tubular cells, since JNK is a member of the MAP kinase family, which phosphorylates a number of transcription factors, primarily components of AP-1 such as JUN, IRS1 and ATF2 and thus regulates AP-1 transcriptional activity [44,45]. In addition, in a recent investigation, Feng et al. showed that JNK-mediated MFF phosphorylation was required for mitochondrial fission in human renal mesangial cells that were exposed to HG ambience . Danesh and colleagues found that DRP1/MFF-mediated mitochondrial fission contributed to diabetic podocyte injury . These data suggest that MFF is a key effector of mitochondrial fission in the mesangial cells and podocytes. Here, in the renal proximal tubular cells, we observed that HG stimulation increased MFF phosphorylation and its interaction with DRP1 on the mitochondrial surface. Knockdown of DsbA-L in HK-2 cells further increased MFF phosphorylation and its interaction with DRP1 while pretreatment with the JNK inhibitor reversed these effects (Figure 5). These data suggest that MFF-mediated mitochondrial fission plays a common and important role in mitochondrial dysfunction in DKD, and DsbA-L down-regulation mediates tubular mitochondrial dysfunction in DKD at least by activating the mtROS-JNK-MFF pathways. However, we cannot exclude the possibility that additional mechanisms may account for DsbA-L-regulated mitochondrial fragmentation. For example, in in vivo study, we observed that DsbA-L knockout decreased expression of pro-fusion protein (MFN2) and increased expression of profission protein (DRP1), suggesting that DsbA-L gene deletion impairs the balance between mitochondrial fusion and fission, resulting in increased fragmentation.
Finally, dysregulated mtROS level is a hallmark of mitochondrial dysfunction in DKD . Elevated renal ROS levels were found both in mouse model and patients with DKD [13,15,46]. However, the underlying mechanisms of mtROS overgeneration in states of hyperglycemia in tubular pathobiology are not clearly defined. In the present study, we found that DsbA-L deficiency aggravated tubular cell oxidative stress, as indicated by DHE staining, in diabetic mice (Figure 2); In in vitro study, knockdown of DsbA-L expression further aggravated mtROS generation, as indicated by MitoSox staining, in HK-2 cells, while DsbA-L overexpression reversed it (Figure 3). Based on these findings and the fact that mitochondria are the main source of ROS in PTECs , we postulated that DsbA-L is responsible for mtROS production under HG ambience. mtROS formation is related to the functionality of the mitochondrial potential permeability transition pore (mPTP) . In the present study, we found that HG treatment promoted mPTP opening in HK-2 cells, which was consistent with our previous studies . Knockdown of DsbA-L expression further increased HG-induced mPTP opening, and this effect was abrogated by DsbA-L overexpression (Supplementary Figure S2). These results indicate that DsbA-L ameliorates mtROS overproduction at least partially by improving mPTP function. In addition, the release of cytochrome c from mitochondria induces the generation of ROS [50,51]. Interestingly, a previous study identified potential DsbA-L interacting proteins, including cytochrome c, by LC-ESI-MS/MS analysis , and in the present study, we found that DsbA-L was localized to mitochondria in HK-2 cells (Figure 3). This supports the idea that DsbA-L may modify mtROS production by interacting with cytochrome c. If this is correct, we hypothesize that DsbA-L interacts with cytochrome c to prevent mtROS overproduction through two independent mechanisms: (1) by serving as a molecular chaperone in mitochondria and interacting with cytochrome c to block cytochrome c release under HG ambience, since DsbA-L has been reported to function as a molecular chaperone to facilitate adiponectin assembly in adipocytes ; and (2) by protecting cytochrome c against redox enzymes, such as p66Shc , which oxidizes under HG ambience. Further investigation will be needed to validate these possibilities.
In summary, our study highlights that DsbA-L down-regulation induces alterations in mitochondrial dynamic via inhibition of mitochondrial fusion by decreasing MFN2 expression and promotes mitochondrial fission by activating the mtROS–JNK–MFF pathway (Figure 6). These results enhance our understanding of the relevance of DsbA-L-mediated ROS generation and its interplay with the mitochondrial fusion-fission machinery. Therefore, the manipulation of DsbA-L expression may be a therapeutic strategy to ameliorate mitochondrial dysfunction in diabetic tubulopathy. We recognize the pitfalls associated with the use of a STZ-treated diabetic mice model. The issue is that STZ has been reported to cause nephrotoxicity in addition to pancreatic β-cells , and this raises the possibility that DsbA-L deficiency may enhance renal susceptibility to STZ toxicity, which will be a focus of future investigations. Besides, the in vivo experiments were performed in systemic DsbA-L knockout mice. Thus, the conditional deletion of DsbA-L in proximal tubular cells is required in the future to support our findings.
The potential pathways by which DsbA-L modulates mitochondrial dynamics in renal tubular cells under hyperglycemia
Growing evidence has suggested that disruption of the dynamic balance, especially a shift toward fission, contributes to tubular damage in DKD. Therefore, identifying molecules in tubular cells that inhibit mitochondrial fragmentation in DKD would be imperative.
The present study showed that DsbA-L deficiency accentuated tubule mitochondrial fragmentation and subsequent tubular damage under a diabetic state. At the molecular level, down-regulation of DsbA-L expression increased mtROS generation, which ultimately activated JNK–MFF mediated mitochondrial fission, consequently inducing mitochondrial dysfunction and driving the development and progression of DKD.
We hypothesize that manipulation of DsbA-L expression would suggest a therapeutic strategy to ameliorate diabetic tubulopathy.
The authors declare that there are no competing interests associated with the manuscript.
This study was funded by grants from the National Natural Sciences Foundation of China [grant number 81730018]; and the National Key R&D Program of China [grant number 2018YFC1314002].
P.G. and L.S. designed the study; J.H.L. performed the bioinformatic analysis. P.G., M. Y. and X.H.C. performed the animal experiments. P.G., M.Y., X.H.C. and S.X. performed the in vitro experiments; Figure preparation, original draft wiring, editing and review were performed by P.G. and L.S.; All authors have revised and approved the submitted manuscript.
We would like to thank the Department of Nephrology, Second Xiangya Hospital, Central South University and the Hunan Key Laboratory of Kidney Disease and Blood Purification in Changsha, China. We would like to thank the Nephroseq database and Genotype-Tissue Expression (GTEx) project which provided us with comprehensive public resources to study tissue-specific gene expression.
chronic kidney disease
- COX IV
diabetic kidney disease
disulfide-bond A oxidoreductase-like protein
estimated glomerular filtration rate
focal segmental glomerulosclerosis
kidney injury molecule-1
mitochondrial fission factor
mean fluorescence intensity
mitochondrial permeability transition pore
proximal tubular epithelial cell
reactive oxygen species