FGF13-Sensitive Alteration of Parkin Safeguards Mitochondrial Homeostasis in Endothelium of Diabetic Nephropathy

Chronic hyperglycemia contributes to both microvascular and macrovascular complications of diabetes, which are typical manifestations of endothelial dysfunction in the glomerulus (1). Glomerular endothelial cells (GECs) are highly specialized cells with fenestrae and a charged luminal glycocalyx layer that functionally contribute to the filtration barrier (2,3). The severity of diabetic nephropathy (DN) positively correlates with endothelial dysfunction in type 2 diabetes (4,5). Furthermore, reduction of GEC fenestrae correlates with an increase in the level of albuminuria and the loss of glomerular filtration in type 2 diabetes (6).

Mitochondrial dysfunction has been linked to DN. Various novel therapies that target mitochondrial dysfunction are being investigated for the treatment or prevention of DN (7). As kidney damage progresses in diabetes, the mitochondrial membrane potential in endothelial cells, podocytes, and proximal tubules decreases (8–10). In experimental models of DN, renal endothelial cells have mitochondrial structural abnormalities (8). Interestingly, accumulation of 8-oxoguanine, a marker of mitochondrial dysfunction, is detectable in GECs, but not in podocytes, of patients with DN histopathology (8). Therefore, targeting mitochondrial homeostasis of GECs may be a promising new therapeutic strategy to prevent or ameliorate DN.

Fibroblast growth factor 13 (FGF13), which belongs to the FGF homologous factor (FHF) subfamily, is the ancestral gene of the FGF family (11). Unlike many canonical FGFs, FHFs lack a recognizable secretory signal sequence and do not activate any known FGF receptors. Rather, FHFs remain intracellular, where they are multifunctional. Members of the FHF subfamily have been reported to be associated with kidney injury. FGF11 participates in renal injury during DN progression by promoting proliferation and fibrosis of mesangial cells (12). Perturbation of the miR-342-3p/FGF11 cascade by hyperinsulinemia contributes to vascular dysfunction in type 2 diabetes (13). Moreover, FGF13 expression is significantly increased in vascular smooth muscle cells and aortas of db/db mice (14). FGF13 is also upregulated in patients with septic shock–associated acute kidney injury relative to the patients without this injury (15). mRNA abundance of FGF13 in the kidneys is increased at all time points after ischemia-reperfusion injury (16). However, it is unknown whether FGF13 plays an essential role in the pathogenesis of type 2 DN (T2DN).

Here, we found that FGF13 expression positively correlates with T2DN progression. On the basis of transcriptomics analyses, we surmised that FGF13 modulates mitochondrial homeostasis by targeting Parkin. Parkin, a RING-between-RING E3 ubiquitin ligase, is inhibited by several types of kidney injury, including T2DN. It participates in complex cellular processes, including mitophagy, apoptosis, immune signaling, and differentiation (17). Both mitophagy and apoptosis are critical for maintaining proper mitochondrial homeostasis when mitochondria are damaged by various stresses. Therefore, we sought to determine whether FGF13 contributes to the intracellular regulatory mechanisms that coordinate T2DN progression through a Parkin-mediated prosurvival function.

The animals were handled according to the guidelines and principles published in the National Institutes of Health Guide for the Care and Use of Laboratory Animals. The study was approved by Wenzhou Medical University animal care and welfare committee (protocol no. wydw2021-0307) (Supplementary Material).

Hematoxylin-eosin (H-E), and Masson trichrome staining were used to examine kidney sections. The deposition of collagen was assessed by evaluation of the area of Masson trichrome–positive and nucleus-free segments (containing glomerular areas and tubulointerstitial areas). Investigators were blind to these studies, and kidney pathology was assessed by experienced pathologists.

After the indicated treatment, the cells or tissue blocks of ∼3 mm³ (including a portion of the renal cortex and outer medulla) were fixed in 2.5% glutaraldehyde dissolved in 0.1 mol/L PBS and prepared for routine electron microscopy, then processed for EPON embedding using standard protocols. Samples were examined with an H-7500 transmission electron microscope.

RNA isolation, semiquantitative RT-PCR (sqRT-PCR) and quantitative real-time RT-PCR were performed as described previously (18). Gene-specific primer sequences used for PCR analyses are listed in Supplementary Table 2.

Samples were fixed with 4% paraformaldehyde and washed with PBS, then the apoptotic cells were detected using an in situ cell death detection kit (Promega). After the TUNEL reaction, samples were stained with DAPI to detect nuclei and imaged using confocal microscopy.

Mitochondrial fractionation was performed using the Mitochondrial Isolation Kit for Cultured Cells and Tissues (Thermo Fisher Scientific) according to the manufacturer’s instructions.

After the indicated treatment, the cells were collected using TRIzol reagent (Takara, Dalian, China) and stored at −80°C for RNA extraction. Novogene (Beijing, China) detected the purity and concentration of RNA samples, constructed libraries, and performed the sequencing using the Illumina NovaSeq 6000 (Lianchuan Sciences, Hangzhou, China) sequencing platform (Supplementary Material). Processed and raw data are available through the Gene Expression Omnibus under the accession no. GSE192889.

All the data were generated from at least three independent experiments. Statistical analysis of the experimental data were performed using GraphPad Prism 8.0.2 (GraphPad Software, San Diego, CA). Unpaired Student t test or one-way ANOVA with Tukey multiple comparison test was performed relative to data type. Values were considered statistically significant if P < 0.05. All biological and technical experiments were performed at least three times, and the data are presented as mean ± SEM, unless noted otherwise.

The data sets generated and/or analyzed during the current study are available from the corresponding authors upon reasonable request.

We first determined whether FGF13 is associated with T2DN. Typical renal fibrosis in glomeruli was observed in patients with T2DN compared with those without T2DN and was worse in patients with severe T2DN than in those with mild T2DN, as measured by H-E and Masson trichrome staining (Fig. 1A). In parallel, immunofluorescence (IF) analysis confirmed that FGF13 was mainly expressed in glomeruli, and the expression was much lower in patients with mild T2DN than in those with severe T2DN (Fig. 1A). Moreover, we found that FGF13 expression progressively increased in db/db mice (Fig. 1B–D). IF staining of mouse renal tissues and human samples showed that FGF13 was significantly increased in CD31⁺ GECs under diabetic conditions (Supplementary Fig. 1A–D). By contrast, FGF13 did not markedly colocalize with WT1, a marker of podocytes, or PDGFRβ, a marker of mesangial cells (Supplementary Fig. 1A and B). Consistently, the mRNA expression of Fgf13 was upregulated in GECs of db/db mice compared with db/m mice, but was not significantly changed in the other cell types in glomeruli (Fig. 1E). We then confirmed the expression of FGF13 in vitro. Human renal GECs ((HRGECs) incubated with 30 mmol/L glucose (high glucose [HG]) and 0.1 mmol/L palmitate (PA) (HG-PA) exhibited a time-dependent increase of FGF13 expression and endothelial dysfunction, as reflected by decreased vascular endothelial (VE) cadherin expression and increased N-cadherin expression (Fig. 1F). In addition, sqRT-PCR analysis showed that mRNA expression of FGF13 isoforms Fgf13-S, Fgf13-V, Fgf13-Y, and Fgf13-VY, but not Fgf13-U, differed in GECs between db/m and db/db mice (Supplementary Fig. 1E and F).

To elucidate the biological significance of endothelial FGF13 expression in T2DN in vivo, Cdh5-CreER^T2 mice were cross bred with Fgf13^flox/Y or Fgf13^flox/flox mice to generate endothelial cell–specific Fgf13 knockout (Fgf13^ECKO) mice, which was confirmed by tail genotyping (Fig. 2A). Additionally, IF staining of FGF13 and CD31 in the aorta, heart, and kidney confirmed that FGF13 was specifically ablated in endothelial cells (Supplementary Fig. 2A). No significant difference in tissue phenotypes was found between control mice and Fgf13^ECKO mice of either sex under basal conditions (Supplementary Fig. 2B). In this study, the T2DN model was generated using male mice because female mice may be less sensitive to islet cell toxins (19). IF staining demonstrated that FGF13 expression in GECs was markedly lower in Fgf13^ECKO mice than in control mice under both basal and diabetic conditions (Fig. 2B). Consistently, the protein level of FGF13 was significantly lower in glomeruli isolated from Fgf13^ECKO mice and was slightly increased in high-fat diet plus streptozotocin (HFD + STZ)–induced diabetic Fgf13^ECKO mice (Fig. 2C). H-E and Masson trichrome staining revealed that glomerular fibrosis was ameliorated in Fgf13^ECKO diabetic mice compared with control diabetic mice (Fig. 2D and F). Furthermore, blood urea nitrogen (BUN), as well as albuminuria and cystatin C levels, were significantly lower in Fgf13^ECKO diabetic mice than in control diabetic mice (Fig. 2G). Of note, there was no significant difference in hemoglobin A_1c between the control diabetic and the Fgf13^ECKO diabetic groups (Supplementary Fig. 2C). Moreover, transmission electron microscopy analyses revealed that diabetic mice exhibited a marked reduction in the density of glomerular endothelial fenestrations compared with control mice. However, fenestrations were easily observed in Fgf13^ECKO diabetic mice compared with diabetic mice, and FGF13 deletion alone did not alter the density of fenestrations (Fig. 2H). Next, we generated an HRGEC line in which Fgf13 gene expression was stably depleted using lentivirus-mediated RNA interference (L-shFgf13) (Fig. 2I). Both basal and maximal respiration, as well as the ATP production, were higher in HG-PA–treated Fgf13 knockdown (KD) HRGECs than in HG-PA–treated control HRGECs, indicating that mitochondrial respiration was increased by Fgf13 KD (Fig. 2J). Also, the gradually disappearing mitochondrial cristae in HG-PA–treated HRGECs was restored by Fgf13 KD (Fig. 2K). VE-cadherin was continuously detected by IF around the entire periphery of cells cultured with normal glucose, whereas HG-PA–treated cells exhibited discontinuous and unstable junctions with intercellular gaps (Fig. 2L, white arrows). Of note, Fgf13 KD rescued the perturbed barrier function of HG-PA–treated HRGECs, as reflected by their ability to form continuous cell-cell contacts (Fig. 2L). The level of FGF13 was not obviously changed in kidneys from the type 1 diabetic model compared with control mice but significantly increased in T2DN (Supplementary Fig. 2D). Furthermore, the protein level of FGF13 was increased in HRGECs treated with PA or HG-PA while showing no significant change in HG-treated HRGECs (Supplementary Fig. 2E). The expression of FGF11, another member of the FHF subfamily, was increased in HG-treated HRGECs and decreased in HRGECs treated with PA or HG-PA (Supplementary Fig. 2E). Similar results were obtained in mice (Supplementary Fig. 2D). We suspect that FGF13 is more driven by lipids than HG, while FGF11 is more driven by HG in various renal injuries.

To elucidate the mechanism by which FGF13 modulates T2DN, RNA sequencing was conducted to explore the transcriptional profiles of HRGECs treated with HG-PA alone or together with L-shFgf13. After gene selection based on fold changes (>1.5 or <0.5) and P values (<0.05), 207 (including upregulated and downregulated) genes were found to be differentially expressed between HG-PA–treated HRGECs transfected with L-Scramble (L-Scr) and HG-PA–treated HRGEC transfected with L-shFgf13 (Supplementary Fig. 3A). Kyoto Encyclopedia of Genes and Genomes pathway analysis of dysregulated genes revealed that many key mitochondrial pathways (20), such as mitophagy, apoptosis, citrate cycle (tricarboxylic acid cycle), Parkinson disease, and cGMP-PKG signaling pathway, were affected in HG-PA–treated Fgf13 KD HRGECs in T2DN (Supplementary Fig. 3B and C), although the effects were not statistically significant for some pathways. Interestingly, FGF13 exhibited an obvious mitochondrial localization in HRGECs (Supplementary Fig. 3D).

Increasing evidence suggests that glomerular endothelial mitochondrial dysfunction is the predominant characteristic of DN susceptibility (8). Given the potential relevance of FGF13 to diabetic mitochondrial dysfunction, we assayed mitophagy and apoptosis in HRGECs. Colocalization between LC3B and mitochondria was evaluated by blocking autophagy flux using bafilomycin A1 to assess mitophagy. This colocalization was reduced in HG-PA–treated HRGECs but was preserved by Fgf13 KD (Fig. 3A). Consistently, mitochondria-enriched subcellular fractions exhibited increased LC3B-II/LC3B-I and decreased TOM20 protein levels in L-shFgf13–infected, but not L-Scr–infected, cells treated with HG-PA (Fig. 3B). We then sought to characterize the role of FGF13 in apoptosis. Remarkably, Fgf13 KD reduced apoptosis induced by HG-PA, as reflected by decreases in cleaved caspase-3 expression, the Bax/Bcl-2 ratio, and TUNEL⁺ cells (Fig. 3B and C). To exclude the possibility that the upregulation of FGF13 is a compensatory mechanism in T2DN, we constructed an FGF13 overexpression (OE) system (Ad-Fgf13 OE) in HRGECs (Supplementary Fig. 4A). Compared with HG-PA–treated HRGECs, Fgf13 OE significantly downregulated mitophagy flux and LC3B-II/LC3B-I levels and increased the TOM20 protein level in mitochondria (Supplementary Fig. 4B and C). Apoptosis was exacerbated in Fgf13 OE HRGECs treated with HG-PA, as determined by immunoblotting and the TUNEL assay (Supplementary Fig. 4B and D). In addition, VE-cadherin staining exhibited more discontinuous and unstable junctions with intercellular gaps (Supplementary Fig. 4E, white arrows), indicating the detrimental role of FGF13 in GEC injury.

We further confirmed the effect of FGF13 on mitophagy and apoptosis in vivo. The results showed that colocalization of endothelial mitochondria and autophagic puncta was increased, while the ratio of LC3B-II/LC3B-I was increased and TOM20 expression decreased in Fgf13^ECKO diabetic mice compared with HFD + STZ–treated control mice (Fig. 3D and Supplementary Fig. 3E). Moreover, endothelial-specific deletion of Fgf13 ameliorated HFD + STZ–induced apoptosis in glomeruli, as demonstrated by a decrease in TUNEL⁺ cells (Fig. 3E). Similarly, cleaved caspase-3 expression and the Bax/Bcl-2 ratio were reduced (Supplementary Fig. 3E).

Parkin is associated with Parkinson disease and determines cell fate in response to mitochondrial damage. FGF13 was previously reported to be associated with Parkinson disease (21). Hence, we determined whether a Parkin-mediated prosurvival function participates in the regulatory role of FGF13 in T2DN. To clarify the role of Parkin in mediating the effects of FGF13 on diabetes in vitro, we performed RNA sequencing of HG-PA–treated Fgf13 KD HRGECs transfected with or without Prkn-targeting siRNA (siPrkn) (Fig. 4A). Transfection of siPrkn abrogated Fgf13 KD–mediated upregulation of mitophagy and downregulation of prominent proapoptotic initiators in HG-PA–treated HRGECs (Fig. 4B). Functional clustering analysis highlighted that mitophagy-related pathways (Parkinson disease and phagosome) and apoptosis signaling were highly affected (Fig. 4C). Parkin-mediated mitophagy is considered a prosurvival mechanism in response to mitochondrial stress (22). Parkin has also been reported to prevent apoptosis in a variety of settings (23–25). We thus evaluated whether disruption of Parkin abrogated the protective effects of Fgf13 depletion on mitochondrial homeostasis in vitro. Transmission electron microscopy indicated that Prkn KD abrogated the improvement of mitochondrial morphology induced by Fgf13 KD in HG-PA–treated HRGECs, as reflected by broken mitochondrial cristae (Fig. 4D). Furthermore, Prkn KD diminished basal oxygen consumption, the ATP-linked oxygen consumption rate, and maximal respiration in HG-PA–treated Fgf13 KD HRGECs (Fig. 4E). Immunoblotting and IF assays demonstrated that Prkn KD blocked the effects of Fgf13 KD on the mitophagosomal machinery, as reflected by increased TOM20 expression and decreased mitochondrial LC3B-II/LC3B-I levels (Fig. 4F) as well as decreased colocalization of LC3B and mitochondria in HG-PA–treated Fgf13 KD HRGECs (Fig. 4G). In addition, Prkn KD in HG-PA–treated Fgf13 KD HRGECs increased Bax/Bcl-2–dependent cleavage of caspase-3 (Fig. 4F) and the number of TUNEL⁺ cells (Fig. 4H), accompanied by dysregulated endothelial barrier integrity (Fig. 4F and I).

An immunohistochemical assay confirmed that expression of Parkin was much lower in T2DN samples than in control samples of both renal biopsy tissues and mouse kidneys (Fig. 5A). The endogenous FGF13-Parkin complex was detected in immunoprecipitates of endogenous FGF13 from HRGEC lysates (Fig. 5B). The interaction between FGF13 and Parkin was also observed in IB4⁺ cells in mouse glomeruli (Fig. 5C). Therefore, we explored whether FGF13 affects the subcellular localization of Parkin in HRGECs. Under resting conditions, Fgf13 KD increased the mitochondrial localization and decreased the cytoplasmic localization of Parkin in HRGECs, while Fgf13 OE decreased the mitochondrial localization of Parkin. Under diabetic conditions, Parkin immunostaining was increased in all the subcellular fractions in Fgf13 KD HRGECs, while the opposite was found in Fgf13 OE HRGECs (Fig. 5D and E). Moreover, immunoblotting showed that Fgf13 KD attenuated HG-PA–induced downregulation of Parkin expression in the cytosolic, mitochondrial, and nuclear fractions of HRGECs (Fig. 5F). Of note, Fgf13 KD increased the mitochondrial localization of Parkin in HRGECs under resting conditions (Fig. 5F). In parallel, downregulated Parkin expression in diabetic mouse glomeruli was rescued in Fgf13^ECKO mice (Fig. 5G).

Autophagy has been implicated as a protagonist of apoptosis in stress models (26–28). Hence, we investigated whether crosstalk occurs between mitophagy and apoptosis in Fgf13 KD HRGECs treated with HG-PA. Interestingly, suppression of mitophagy by Atg7 KD did not obviously affect the downregulation of apoptosis induced by Fgf13 deficiency in HG-PA–treated HRGECs (Fig. 6A and C). Atg7 KD partially suppressed the protective effects of Fgf13 KD on mitochondrial homeostasis and endothelial permeability in HG-PA–treated HRGECs, as determined by mitochondrial respiration and VE-cadherin staining (Fig. 6A, D, and E). Additionally, Fgf13 KD HRGECs were treated with PETCM, a pharmacological caspase-3 activator. At the concentrations used, mitophagy was not significantly affected by PETCM, while caspase-3 activity was increased in HG-PA–treated Fgf13 KD HRGECs (Fig. 6A–C). Increased caspase-3 activity in Fgf13 KD cells partly affected the VE-cadherin and N-cadherin expression, mitochondrial respiration, ATP content, and endothelial barrier integrity (Fig. 6A, D, and E). By contrast, the protective effects of Fgf13 KD on mitochondrial homeostasis and endothelial barrier integrity in HG-PA–treated HRGECs were dramatically abolished in Atg7-deficient cells treated with PETCM (Fig. 6A–E), indicating that FGF13 affects mitophagy and apoptosis through different mechanisms during T2DN progression.

Parkin recruits itself onto mitochondria through self-association, which is essential to initiate mitophagy (29,30). Here, PINK1 was used as a positive control to drive Parkin self-association. We found that Parkin self-association was markedly decreased in HEK293T/17 cells overexpressing Fgf13 (Fig. 6F), suggesting that FGF13 regulates mitophagy by inhibiting the self-association of Parkin. On the other hand, a previous study demonstrated that Parkin functions as a transcriptional repressor of p53 (31), which induces apoptosis by upregulating the transcription of many proapoptotic genes, including Bax (32). Therefore, we investigated whether Parkin-p53-Bax signaling correlates with the effects of FGF13 on apoptosis. Notably, Fgf13 KD downregulated the expression of p53, the level of which was enhanced by Prkn KD (Fig. 6G). In addition, chromatin immunoprecipitation experiments showed that FGF13 markedly reduced the interaction between Parkin and the p53 promoter in HRGECs (Fig. 6H). Importantly, p53 depletion prevented the Prkn KD–mediated increase of apoptosis in Fgf13 KD HRGECs treated with HG-PA (Fig. 6I and J), suggesting that Parkin-p53-Bax signaling participates in FGF13-mediated apoptosis under diabetic conditions.

To determine whether FGF13-sensitive alteration of Parkin safeguards mitochondrial homeostasis in the endothelium of T2DN in vivo, we used Fgf13 and Prkn double-KO (DKO) mice. Individual KO or DKO of Fgf13 and Prkn was confirmed in the kidneys of mice with different genotypes (Fig. 7A). Prkn KO under resting conditions showed a similar phenotype as the control mice (Supplementary Fig. 5A–E). Also, no significant difference in hemoglobin A_1c levels was observed in all diabetic groups (Supplementary Fig. 5F). In contrast, the renoprotection induced by Fgf13 deficiency in T2DN was significantly reversed in DKO mice (Fig. 7B–E). Notably, DKO mice treated with HFD + STZ exhibited a substantial increase in BUN, albuminuria, and cystatin C, along with an obvious reduction in glomerular endothelial fenestrations, in contrast with the protected phenotype of Fgf13^ECKO diabetic mice (Fig. 7D and F–H). Furthermore, we examined mitophagy and apoptosis in HFD + STZ–treated DKO mice. Compared with Fgf13^ECKO mice under diabetic conditions, fewer LC3B foci colocalized with mitochondria in GECs from DKO mice, which correlated with decreased LC3B-II/LC3B-I and increased TOM20 protein levels (Fig. 7I and J and Supplementary Fig. 6A). Also, apoptosis was markedly higher in DKO mice than in Fgf13^ECKO mice with T2DN, as evidenced by assessment of TUNEL⁺ cells and immunoblotting (Fig. 7K and L and Supplementary Fig. 6A). By contrast, renal-specific OE of Parkin in Fgf13^ECKO mice by in situ injection of adeno-associated virus 2 at the cortex of the left-side kidney induced a similar renal phenotype as individual depletion of Fgf13 (Supplementary Figs. 6B and 7), suggesting that FGF13 and Parkin act in the same pathway to regulate T2DN.

The kidney is arguably the most important target of microvascular damage in diabetes (33). The presence and severity of chronic kidney disease identify individuals who are at increased risk of adverse health outcomes and premature mortality. Consequently, preventing and managing chronic kidney disease is now the key aim of the overall management of patients with diabetes. In the current study, we provided multiple lines of evidence that FGF13 is a previously unappreciated nonsecretory protein that regulates endothelial mitochondrial homeostasis in T2DN. First, FGF13 expression was induced in GECs during T2DN progression, and endothelial-specific deletion of Fgf13 alleviated T2DN progression while Fgf13 OE had the opposite effect. Second, a Parkin-mediated prosurvival function participated in the regulatory role of FGF13 in mitochondrial homeostasis. These findings establish that FGF13-sensitive alteration of Parkin safeguards mitochondrial homeostasis in the endothelium of T2DN.

The involvement of canonical secretory FGF signaling in pathological development of DN is well documented. In this regard, FGF21, a member of the endocrine FGF subfamily whose expression is regulated by peroxisome proliferator–activated receptor c in adipose tissue, is an effective glucose-lowering agent in rodents (34,35). FGF1, the prototype of the FGF protein family, was recently found to ameliorate DN by countering inflammatory signaling cascades in injured renal tissue (36). However, little is known about the function of nonsecretory proteins of the FHF subfamily in DN. FGF13 is a noncanonical FGF with identified roles in neuronal development, pain sensation, and cardiac physiology. Our laboratory and others have characterized the role of FGF13 in cardiac hypertrophy (18,37). Nevertheless, the role of FGF13 in kidney diseases has heretofore received less attention. In this study, we used RNA sequencing to elucidate the mechanism by which FGF13 modulates T2DN. The results suggested that mitochondrial homeostasis is involved in FGF13-regulated progression of T2DN. Furthermore, this study demonstrated the functional role of FGF13 in mitophagy and apoptosis under diabetic conditions for the first time. Endothelial autophagy is an endogenous protective system, and impairment of autophagy in some settings, such as upon exposure to oxidative stress, can jeopardize endothelial integrity and lead to organ dysfunction (38). In addition, apoptosis is an important contributor to damage and loss of renal microvascular endothelial cells and development of microvascular rarefaction and may increase the risk of renal injury progression (39). Here, we found that Fgf13 deficiency activates mitophagy and prevents apoptosis, leading to mitochondrial homeostasis.

We further established the mechanistic basis for renal protection upon Fgf13 KD. Parkin, a critical mediator of mitochondrial quality control, has been implicated in determining cell fate upon mitochondrial damage (40). Our RNA sequencing analyses showed that the mitophagy-related pathways (Parkinson disease and phagosome) and apoptosis signaling were highly affected in HG-PA–treated Fgf13 KD HRGECs transfected with siPrkn. Furthermore, transfection of siPrkn abrogated Fgf13 KD–mediated mitochondrial homeostasis in HG-PA–treated HRGECs, indicating that Parkin is an important target of FGF13. Accordingly, we determined how FGF13 modulates Parkin. Expression of Parkin was much lower in T2DN samples than in control samples of both renal biopsy tissues and mouse kidneys. Also, the endogenous FGF13-Parkin interaction was observed in HRGECs and mouse glomeruli. Notably, under resting conditions, Fgf13 KD increased the mitochondrial localization and decreased the cytoplasmic localization of Parkin in HRGECs, while Fgf13 OE decreased the mitochondrial localization of Parkin. Under diabetic conditions, Parkin expression was decreased in the cytosolic, mitochondrial, and nuclear fractions of HG-PA–treated HRGECs, and its expression in these fractions was dramatically restored by Fgf13 KD. These data suggest that FGF13-sensitive alteration of Parkin safeguards mitochondrial homeostasis in the endothelium of DN. Endothelial-specific DKO of Fgf13 and Prkn abrogated the renal protective effects observed upon individual depletion of Fgf13 in vivo. By contrast, Fgf13^ECKO mice overexpressing Prkn had a similar renal phenotype as Fgf13^ECKO mice. These data further confirm that FGF13-sensitive alteration of Parkin safeguards mitochondrial homeostasis in the endothelium of DN. Nevertheless, additional studies are needed to define how FGF13 affects Parkin expression in order to regulate the dynamic equilibrium between mitophagy and apoptosis in T2DN.

Some researchers have implicated autophagy and mitophagy as regulators of apoptosis in stress models or during myoblast differentiation (28,30,31), while others have suggested that regulation of mitophagy and apoptosis is mechanistically distinct (41–43). Based on our results, regulation of mitophagy and apoptosis by FGF13 through Parkin in GECs appears to be mechanistically distinct. Inhibition of mitophagy did not affect apoptosis, and activation of apoptosis by PETCM did not significantly affect mitophagy in HG-PA–treated Fgf13 KD HRGECs, suggesting that mitophagy and apoptosis may occur through different mechanisms in distinct cell types and according to the unique bioenergetics of each model system. Initiation of mitophagy is mediated by translocation of Parkin to the mitochondrial outer membrane. Importantly, Parkin recruits itself onto mitochondria through self-association (29,30), and FGF13 regulates Parkin-dependent mitophagy by inhibiting self-association of Parkin, indicating that FGF13 has a direct role in regulation of mitophagy in GECs. Prior work suggested that the antiapoptotic effects of Parkin are mediated by cytosolic Parkin, which is the major pool of the protein at rest (24,44). Parkin has been suggested to reduce apoptosis by inhibiting proapoptotic Bax by way of limiting recruitment of cytosolic Bax to mitochondria or promoting degradation of dysregulated or mutated Bax (28,41,44). Furthermore, some studies assumed that Parkin prevents apoptosis by inhibiting the interaction between Bax and VDAC2 or inhibiting the activity of BAK (28). In this study, the inhibitory effects of Parkin on apoptosis induced by Fgf13 deficiency directly suppressed transcription of several proapoptotic genes, including Bax, indicating that Parkin indirectly prevents apoptosis in HG-PA–treated Fgf13 KD HRGECs. A previous study delineated another function of Parkin as a transcriptional repressor of p53 (31). p53 induces apoptosis by upregulating the transcription of many proapoptotic genes, including Bax (32). Here, we confirmed that Parkin-p53-Bax signaling participates in FGF13-mediated apoptosis under diabetic conditions.

In summary, the bifunctional role of Fgf13 deficiency in inhibiting apoptosis and promoting mitophagy through Parkin may shape mitochondrial homeostasis regulation in GECs and T2DN progression. This pathway can be targeted to prevent and control T2DN and may be relevant to the pathogenesis of other FGF13- and Parkin-associated diseases.

Funding. This work was supported by the National Natural Science Foundation of China (grants 82070507, 81900240, 81770498, and 81773346), Natural Science Foundation of Zhejiang Province (LZ21H020002), and the Natural Science Foundation of Ningbo (202003N4070).

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

Author Contributions. J.S. wrote the original draft of the manuscript. J.S., L.J., and W.C. reviewed and edited the manuscript. J.S. and W.C. contributed to the conceptualization. X.G. and C.N. contributed to the methodology. C.N., Y.C., L.J., and W.C. acquired funding. P.C., Y.L., and X.W. contributed to the validation. L.L., M.L., Y.S., X.H., Y.C., J.Z., and J.F. contributed to the investigation. X.L. and L.J. contributed resources. L.J. and W.C. supervised the project. W.C. contributed to the project administration. W.C. 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.