Podocytes play a pivotal role in maintaining glomerular filtration function through their interdigitated foot processes. However, the mechanisms that govern the podocyte cytoskeletal rearrangement remain unclear. Through analyzing the transcriptional profile of renal biopsy specimens from patients with diabetic nephropathy (DN) and control donors, we identify SLIT-ROBO ρGTPase-activating protein 2a (SRGAP2a) as one of the main hub genes strongly associated with proteinuria and glomerular filtration in type 2 DN. Immunofluorescence staining and Western blot analysis revealed that human and mouse SRGAP2a is primarily localized at podocytes and largely colocalized with synaptopodin. Moreover, podocyte SRGAP2a is downregulated in patients with DN and db/db mice at both the mRNA and the protein level. SRGAP2a reduction is observed in cultured podocytes treated with tumor growth factor-β or high concentrations of glucose. Functional and mechanistic studies show that SRGAP2a suppresses podocyte motility through inactivating RhoA/Cdc42 but not Rac1. The protective role of SRGAP2a in podocyte function also is confirmed in zebrafish, in which knockdown of SRGAP2a, a SRGAP2 ortholog in zebrafish, recapitulates podocyte foot process effacement. Finally, increasing podocyte SRGAP2a levels in db/db mice through administration of adenovirus-expressing SRGAP2a significantly mitigates podocyte injury and proteinuria. The results demonstrate that SRGAP2a protects podocytes by suppressing podocyte migration.
Diabetic nephropathy (DN) is one of the leading causes of chronic kidney disease (1–3). Hypothesis-driven studies in the DN animal model have led to many insights about the development and progression of DN; however, the findings may not to be analogous to human patients. In the past decade, broad-based transcriptomics studies in DN with a human kidney tissue component were performed to uncover the pathogenic mechanisms of DN (4–6). Several previous studies revealed various expressed genes altered in DN, including the upstream regulatory factors and the enriched signaling pathways, but the causative effects of these genes and signal pathways in DN have not been well characterized. Further integration of gene expression data with matching clinical features and identification of the possible gene network responsible for the phenotype under disease conditions are necessary.
That glomerular podocytes play a pivotal role in the pathogenesis of DN is well documented (7,8). Podocyte depletion and loss generally are found in the early stages of DN (9,10), and such podocyte injuries are accompanied by a gradual decline of glomerular filtration rate (GFR) (11,12) and the initiation of proteinuria (13). As terminally differentiated cells residing on the outer surface of the glomerular basement membrane, podocytes play a critical role in maintaining the structure and function of the glomerular filtration barrier (GFB) through their interdigitated foot processes. This specialized function of podocytes depends on their unique cytoarchitecture, particularly the interdigitating foot-like actin-rich processes that arise from podocyte cell bodies and surround glomerular capillary walls. At the interface of the interdigitating foot processes and the capillary wall, the unique junction (also termed the slit diaphragm) allows ultrafiltration of serum. To withstand high pressure in the capillaries and to maintain intact and exact filtration properties, the podocyte must possess a dynamic contractile apparatus and precisely arrange its cytoskeleton spatially and temporally (14). The mechanisms that govern the dynamic arrangement of podocyte cytoskeleton, however, remain unclear.
Accumulating evidence suggests that actin filaments associated with a unique assembly of linker and adaptor molecules (15) are the predominant cytoskeletal components of podocyte foot processes (16). Besides acting as a scaffold for submembrane protein complexes, the cortical actin cytoskeleton also provides a tensile architectural support for podocyte cellular extensions. When podocytes undergo foot process spreading and retraction through remodeling of their cytoskeletal architecture and intercellular junctions, abnormal filter barrier function can occur (17). Buvall et al. (18) reported that EGFR/Src-mediated tyrosine phosphorylation of the actin-organizing protein synaptopodin in podocytes promotes binding to the serine/threonine phosphatase calcineurin, leading to the enhanced Rac1 signaling and ultimate loss of stress fibers in podocytes. Other important molecular switches that regulate the podocyte actin cytoskeleton are the prototypical members of the Rho family of GTPases Cdc42, Rac1, and RhoA (19–23). Previous studies showed that aberrant Rho-GTPase signaling is associated with podocyte mobility, leading to proteinuria (24,25). As a member of SLIT-ROBO ρGTPase-activating proteins (SRGAPs) (26), SRGAP2 belongs to the large Rho family of GTPases. SRGAP2 has been shown to display distinct expression patterns in the central nervous system where it regulates neuronal cell migration (27–29). However, although SRGAP2 was recently predicted as an enriched protein in podocytes (30), little information about its expression pattern and function in the kidney is available. How SRGAP2 modulates the integrity of the podocyte actin cytoskeleton when podocytes undergo foot process effacement under diabetic conditions remains unclear.
In the current study, we used multiple strategies to reveal the key molecules that change the dynamics of podocyte cytoskeleton under diabetic conditions. By analyzing the gene coexpression network and its association with baseline proteinuria and estimated GFR (eGFR), we demonstrate that SRGAP2a, an important component of the SLIT/ROBO signaling pathway during neuronal development, is primarily located at podocytes and functions as a hub gene that is tightly associated with proteinuria and eGFR in patients with DN. By using both in vitro and in vivo systems, we characterize the protective role of SRGAP2a in podocyte structure and function. The results show that podocyte SRGAP2a is downregulated in patients with DN and db/db mice, whereas increasing podocyte SRGAP2a level in db/db mice reverses podocyte cytoskeleton arrangement and thus mitigates podocyte injury and proteinuria. Furthermore, the mechanistic studies demonstrate that podocyte SRGAP2a maintains the normal structure and function of podocytes through suppressing RhoA/Cdc42 activities.
Research Design and Methods
Forty-one patients with type 2 DN by renal biopsy at the National Clinical Research Center of Kidney Diseases, Jinling Hospital, Nanjing University, were enrolled. The clinical characteristics of DN are detailed in Supplementary Table 1. Twenty healthy control glomerular samples were obtained from surgical nephrectomies. The protocol for the use of human samples was approved by the Human Subjects Committee of Jinling Hospital, Nanjing University (2013KLY-013-01), and a signed consent form was obtained from each patient and control donor.
Glomerulus Genome-Wide Gene Expression Profiling and Gene Network/Function Analysis
Microdissection of glomeruli was performed at 4°C. The isolated glomeruli were subject to RNA extraction followed by cDNA synthesis and quantitative PCR (qPCR) assay (QIAGEN, Valencia, CA). Genome-wide gene expression profiling was performed by using the Affymetrix microarray platform (Human Transcriptome Array 2.0). For identifying the expression pattern of various groups, weighted gene coexpression network analysis (WGCNA) was used to cluster coexpressed genes (Gene module). Kyoto Encyclopedia of Genes and Genomes pathway analysis (KEGGEST and GeneAnswers packages in R) was performed to identify enriched pathways in the gene module. The WGCNA was constructed by using the WGCNA package in R (31,32).
The use of animals was approved by the Institutional Animal Care and Use Committee at Jinling Hospital. Diabetic db/db mice on a C57BL/6 background and littermate db/m mice were obtained from The Jackson Laboratory. Body weight and fasting blood glucose levels were monitored weekly. Mouse urinary albumin and creatinine levels were measured by using Albuwell M (Exocell) and The Creatinine Companion Kit (BioAssay Systems).
To investigate the effect of SRGAP2a in mouse kidney, mice were transfected with an SRGAP2a-expressing adenovirus (Ad-SRGAP2a-GFP). An adenovirus expressing green fluorescent protein (Ad-GFP) served as a mock control. Ad-SRGAP2a-GFP and Ad-GFP were purchased from HanBio (Shanghai, China). Briefly, 50 μL of adenovirus (∼1011 plaque-forming units/mL) expressing GFP alone (mock) or SRGAP2a-GFP were injected into mice through the tail vein.
Podocyte Count, Stable Transfection With Lentivirus-Based SRGAP2a Short Hairpin RNAs, or SRGAP2a R527A
Human and mouse podocytes were counted as previously described (33). Human podocytes (34) initially were cultured in RPMI medium containing 10% FBS and insulin-transferrin-selenium (Gibco) at 33°C and then at 37°C for 10–14 days to allow cell differentiation. To knock down endogenous srGAP2, lentivirus (LV)-based short hairpin RNA (shRNA) srGAP2a was used to transfect the podocytes. The paired oligonucleotides targeting human srGAP2a gene was synthesized and annealed into a pHBLV-U6-Scramble-ZsGreen-Puro vector through digestion sites of BamH I and EcoR I. Proliferative podocytes (33°C) stably transfected with SRGAP2a shRNA (>90% proliferative podocytes were SRGAP2a shRNA/GFP positive) were generate through multiple rounds of selection against puromycin treatment. Proliferative podocytes were then induced to differentiated podocytes at 37°C. The primers for shRNA SRGAP2a and SRGAP2a R527A are listed in Supplementary Table 2.
Pulldown Assay for Small GTPase Activities and Immunoprecipitation
RhoA, Rac1, and Cdc42 activities were determined by measuring Rhoketin or PAK1 pulled down by GTP-Rho, GTP-Rac1, and GTP-Cdc42, respectively (35). The small GTPases were separated on 12% SDS-PAGE by following the manufacturer’s instructions (BK030; Cytoskeleton, Denver, CO). Purified His-tagged RhoA/Cdc42/Rac1 was also from the BK030 kit. Protein A/G agarose (Santa Cruz), His-tagged Dynabead (Invitrogen), and anti-SRGAP2a antibodies (Ab121977; Abcam) were used in the coimmunoprecipitation assay.
Data are presented as mean ± SD. Comparisons between groups were made by using a two-tailed unpaired Student t test or one-way ANOVA with Bonferroni post hoc test. Mann-Whitney nonparametric U test was used to analyze data in abnormal distribution. P < 0.05 was considered statistically significant. GraphPad Prism version 6 (GraphPad Software) and SPSS version 22 (IBM Corporation) statistical software were used for the data analysis.
Reduction of SRGAP2a Accounts for Proteinuria and the Aberrant eGFR Observed in Patients With DN
Because animal models of DN do not completely mimic the histological and functional changes of human DN (36,37), we used kidney tissue from patients with DN. To identify key genes associated with podocyte injury and proteinuria in DN, a sequential strategy from transcriptomic analysis to validation study was used (Supplementary Fig. 1). First, we performed genome-wide gene expression profiling on the Affymetrix microarray platform to determine differentially expressed genes in glomeruli between patients with DN and control donors (Gene Expression Omnibus accession no. GSE96804). WGCNA of glomeruli identified 18 gene coexpression modules (Supplementary Fig. 2A and B). Among these gene coexpression modules, the turquoise module, which includes 1,810 transcribed genes, exhibited the highest correlation with proteinuria (R = −0.79; P = 10−13) and baseline eGFR (R = 0.63; P = 10−7) (Supplementary Fig. 2C). Gene function analysis also indicated that the differential expression of genes in the turquoise module are significantly involved in cytoskeletal protein binding (P = 4.37 × 10−13) and cytoskeleton structure (P = 9.29 × 10−12) (Supplementary Table 3). As shown, the turquoise module contained 30 hub genes, and the majority of these genes were involved in cell cytoskeletal organization (Supplementary Fig. 2D). By analyzing the gene-enriched pathway, we also found that proteins involved in the axon guidance signaling pathway were associated with genes encoded by the turquoise module (Supplementary Fig. 2E). The analysis identified that SRGAP2 was one of the hub genes that had the strongest association with baseline proteinuria (R = −0.81; P = 9.61 × 10−15) and eGFR (R = 0.58; P = 1.60 × 10−6) in patients with DN (Supplementary Table 4), suggesting that the absence of SRGAP2 mediates aberrant renal function in diabetes. As shown in Fig. 1, whole-genome gene expression profiling of glomeruli from patients with DN and control donors revealed differential expression of SRGAP family proteins in DN (Fig. 1A). The qPCR analysis of glomeruli from patients with DN (n = 20) and db/db mice (n = 6) showed a significant downregulation of SRGAP2 in both (Fig. 1B). The level of SRGAPs was tightly correlated with proteinuria, plasma creatinine, and declining eGFR (Fig. 1C). To preclude the possibility that the SRGAP2a reduction was due to podocyte loss in these patients, we measured the SRGAP2a mRNA levels between patients with DN and control donors after adjusting for podocyte density. The SRGAP2a mRNA level showed a 31% reduction (P < 0.05) in glomeruli of patients with DN compared with control donors (Fig. 1D). This observation of a negative correlation between levels of SRGAP2 and proteinuria in Chinese patients with DN agrees with Nephroseq (http://www.nephroseq.org) data, which show SRGAP2 downregulation in glomeruli of Woroniecka diabetes (6) and Ju chronic kidney disease (30) (Supplementary Fig. 3).
SRGAP2, a GTPase-activating protein, was originally identified in neurons where it regulates cell migration and neurite outgrowth (26,38). However, SRGAP2 expression is not limited to neuronal cells. The expression of SRGAP2 has been predicted to be expressed in podocytes (30). On the basis of our transcriptomic analysis and previous findings, we postulated that SRGAP2 is a key regulator of the podocyte cytoskeleton and is essential for glomerular filtration function. Given that human SRGAP2 contains three paralogs (SRGAP2a, SRGAP2b, and SRGAP2c), we further analyzed which SRGAP2 paralogs are associated with the development of proteinuria in patients with DN. In agreement with the Higgins Normal Tissue Panel (http://www.nephroseq.org), our microarray results showed expression of SRGAP2a and SRGAP2c but not SRGAP2b in human glomeruli, excluding the role of SRGAP2b in regulating human glomerular function. Moreover, we found no significant reduction of SRGAP2c in glomeruli of patients with DN compare with that of control donors (Supplementary Fig. 4). Structural analysis also revealed that SRGAP2c only contain the F-BAR domain but no RhoGAP and SH3 domain (39). On the basis of these findings, the current study only focused on the expression and function of SRGAP2a.
Podocyte-Specific SRGAP2a Was Downregulated in Patients With DN and db/db Mice
To explore the function of SRGAP2a protein, we first determined where SRGAP2a expression is located in the kidney. Immunofluorescence staining of SRGAP2a in human kidney tissue revealed that SRGAP2a localizes in the glomeruli and colocalizes with synaptopodin but not the glomerular basement membrane marker collagen IV (Fig. 2A and B), suggesting that SRGAP2a is a podocyte-specific protein. The enrichment of SRGAP2a in glomeruli also was confirmed in mice. By trapping magnetic particles within the glomeruli followed by magnetic separation to isolate glomeruli (40), we obtained glomeruli from 8-week-old control and diabetic db/db mice with purity >95%. Western blot results also showed that SRGAP2a was predominantly expressed in the isolated glomerular fraction (Supplementary Fig. 5). Next, we determined srGAP2a level in patients with DN. In agreement with microarray data, the reduction of SRGAP2a in glomeruli of the current patients with DN was confirmed by qRT-PCR assay of samples from an independent group of patients with patients (Fig. 2D).
Downregulation of Podocyte SRGAP2a by Hyperglycemia- or Tumor Growth Factor-β1–Disrupted Podocyte Cytoskeleton
Given that both hyperglycemia and tumor growth factor-β1 (TGF-β1) contribute to the development of DN (41), we next determined whether these factors are involved in downregulating podocyte SRGAP2a. As shown in Fig. 3A and B, after treatment with a high concentration of glucose (30 mmol/L) or TGF-β1 (5 ng/mL) for 24 h, podocyte cytoskeletons staining by fluorescein-conjugated phalloidin showed that a high concentration of glucose or TGF-β1 significantly disrupted the podocyte cytoskeleton in a time-dependent manner. The same treatment also significantly reduced the SRGAP2a expression in a similar time-dependent manner (Fig. 3C). To test whether stable expression of SRGAP2a can rescue podocyte cytoskeleton disruption by TGF-β1 or high-concentration glucose, we examined the fluorescein-conjugated phalloidin staining in podocytes stably transfected with Ad-SRGAP2a-GFP or mock. As shown in Fig. 3D–F, although overexpression of SRGAP2a did not significantly alter the distribution of F-actin stress fiber in untreated podocytes, SRGAP2a overexpression largely rescued the disruption of F-actin stress fiber in podocytes caused by 24-h treatment with TGF-β1 or high-concentration glucose.
The Role of SRGAP2a in Controlling Podocyte Migration
To determine the effect of SRGAP2a downregulation on podocyte function, we knocked down SRGAP2a in cultured podocytes by using LV-delivered shRNA (Fig. 4A). As shown in Fig. 4B and E, the number of stress fibers in podocytes was significantly reduced by SRGAP2a knockdown. We next examined the effect of SRGAP2a knockdown on podocyte migration by using a wound closure assay. As shown in Fig. 4C and D, when confluent monolayers of differentiated podocytes were scratched to initiate the wound-healing process, podocytes with SRGAP2a knockdown showed significantly accelerated wound closure kinetics compared with untreated podocytes. The migration of podocytes was remarkably enhanced by SRGAP2a knockdown (Fig. 4F). These results suggest that SRGAP2a is involved in maintaining podocyte membrane stability.
SRGAP2a Inactivates RhoA/Cdc42 but Not Rac1
A previous study showed that SRGAP2 regulates the dynamics of the cytoskeleton/membranes through the larger F-BAR/RhoGAP/SH3-SRGAP module (27). We also sought to identify the small GTPases that interact with SRGAP2a in podocytes. We performed a proximity ligation assay to detect in situ protein-protein interactions of SRGAP2a with RhoA, Rac1, or Cdc42 in human renal biopsy specimens. As shown in Supplementary Fig. 6, the interactions of SRGAP2a with RhoA and Cdc42 but not Rac1 were detected in renal specimens from both patients with DN and control donors. However, a marked reduction of SRGAP2a-RhoA and SRGAP2a-Cdc42 protein complexes was observed in specimens from patients with DN compared with control donors. These results confirmed an impairment of SRGAP2a-Rho-GTPase interaction in renal DN biopsy specimens.
We next examined the interactions of SRGAP2a with RhoA and Cdc42 in cultured human podocytes. As shown in Fig. 5A and 5B, downregulation of SRGAP2a through LV-SRGAP2a shRNA reduced the binding of RhoA and Cdc42 with SRGAP2a, suggesting that the role of SRGAP2a in stabilizing podocyte cytoskeleton is likely through its interaction with RhoA or Cdc42. The coimmunoprecipitation assay in cultured podocytes also confirmed such SRGAP2a-RhoA and SRGAP2a-Cdc42 interactions (Fig. 5C–E). As expected, SRGAP2a knockdown in cultured human podocytes increased the GTP-bound forms of Cdc42 and RhoA but not Rac1 (Fig. 5C and D). By contrast, SRGAP2a overexpression in podocytes decreased the GTP-bound forms of Cdc42 and RhoA. To further test the specific binding of SRGAP2a with Cdc42 and RhoA, we mutated the potential Cdc42/RhoA-binding site on SRGAP2a to generate SRGAP2a R527A mutant (42). After overexpressing SRGAP2a R527A mutant in podocytes in which SRGAP2a was knocked down, we found that SRGAP2a R527A displayed significantly less binding to RhoA and Cdc42 (Fig. 5F). In agreement with this, exogenous-purified His-RhoA and His-Cdc42 also showed little interaction with SRGAP2a R527A expressed in podocytes (Fig. 5G). Taken together, the interactions of SRGAP2a with RhoA/Cdc42 play a critical role in regulating podocyte cytoskeleton stability: When SRGAP2a is available to bind to RhoA or Cdc42, these small GTPases remain in inactive states, binding to GDP instead of GTP.
SRGAP2a Knockdown in Zebrafish Leads to Podocyte Developmental Defects
Zebrafish kidney shares many features with human kidney, and thus, studies of zebrafish pronephros development and function can provide valuable knowledge for understanding human renal biology (43). Therefore, morpholino oligonucleotide (MO) embryo injection for podocyte function study has been well recognized (44,45). Although zebrafish have two SRGAP2 orthologs, SRGAP2a and SRGAP2b, only SRGAP2a shares high homology with human SRGAP2a (Supplementary Figs. 7–8). The current study thus focused on zebrafish SRGAP2a function.
To reduce the cell death induced by off-target effects of MO, p53 MO were coinjected with SRGAP2a splicing MO targeting exon1 and intron1 (e1i1). As shown in Fig. 6A–C, we observed 85.3% total pericardial, periorbital, and body edema of zebrafish larvae at 3.5 days postfertilization (dpf), and injection of p53 MO alone did not result in any phenotype. Next, we performed transmission electron microscopic imaging of zebrafish pronephric glomeruli, and compared with p53 MO alone, knockdown of SRGAP2a led to podocyte foot process effacement, a disrupted slit diaphragm, and a disorganized GFB (Fig. 6D and F). We also investigated the knockdown effect of SRGAP2a in transgenic embryos Tg (pod:GFP). The confocal imaging showed that knockdown of SRGAP2a resulted in diminished expression of GFP (Fig. 6E). Finally, to confirm the SRGAP2a MO-induced–specific phenotype, zebrafish embryos were injected with SRGAP2a mRNA after the MO treatment. The result clearly showed that injection of SRGAP2a mRNA rescued the MO-induced phenotype and reduced the total edema to 45.1% (Fig. 6G). The results show that loss of SRGAP2a results in zebrafish podocyte foot process effacement and GFB disruption.
Exogenous SRGAP2a Attenuates Podocyte Injury in db/db Mice
We next used db/db mice to explore the protective role of SRGAP2a in podocyte structure and function. In this experiment, the glomeruli were isolated from 6- to 20-week-old db/db mice. As shown in Fig. 7, after 16 weeks, the mice developed extracellular matrix deposition and massive foot process effacement (Fig. 7A and D), elevated blood glucose level and body weight (Fig. 7B), and significant albuminuria (Fig. 7C). Immunofluorescence staining also showed that podocyte SRGAP2a expression was markedly decreased after 12 weeks of age followed by a reduction of other podocyte-specific proteins synaptopodin and Wilms tumor 1 (WT-1) at ∼20 weeks of age (Fig. 7A). Immunoblot analysis confirmed that reduction of SRGAP2a in glomeruli from 12-week-old mice was significantly earlier than the reduction of synaptopodin and WT-1 (Fig. 7E). Furthermore, the levels of GTP-bound RhoA and Cdc42 in isolated glomeruli were markedly increased after 16 weeks (Fig. 7F).
To test whether stabilizing levels of podocyte SRGAP2a in db/db mice can protect mice from proteinuria, we injected 12-week-old db/db mice with Ad-SRGAP2a-GFP (Ad-SRGAP2a) (Fig. 8A). As shown in Fig. 8B and Supplementary Fig. 9, SRGAP2a levels in the glomeruli and tubular cells were significantly increased after the injection. The db/db mice treated with Ad-SRGAP2a exhibited a decreased foot process effacement (Fig. 8C), extracellular matrix deposition (Fig. 8D), and albuminuria (Fig. 8E) compared with the db/db mice treated with Ad-GFP. The mice administered Ad-SRGAP2a also displayed more stable synaptopodin expression and more WT-1–positive cells in the kidney (Fig. 8F and G). Furthermore, the mice administered Ad-SRGAP2a significantly inhibited the activities of RhoA and Cdc42 (Fig. 8H and I). These results indicate that SRGAP2a can protect podocytes from foot process effacement and detachment, thus reducing glomerular abnormalities, including sclerosis, in db/db mice.
In the current study, we integrated renal biopsy specimen–derived transcriptional profiles from patients with type 2 DN with matching clinical information to identify the critical gene responsible for proteinuria and eGFR. Through analyzing genome-wide transcriptional profiles from patients with type 2 DN and healthy control donors as well as from functional assays in both in vitro and in vivo systems, we identified SRGAP2a as a major player in protecting podocytes from foot process effacement under chronic diabetic conditions. Mechanistically, SRGAP2a decreases podocyte motility by suppressing Cdc42/RhoA activities.
Through comparing the renal transcriptional profiles between a DN and a control group, most previous studies have identified many genes that are differentially expressed during podocyte dysfunction (5,6). Because matching gene expression data with clinical features and identifying the possible gene network responsible for the phenotype under disease conditions provide an ideal approach to reveal disease mechanisms and potential drug targets of DN, the current study analyzed the gene networks responsible for the observed clinical features of patients with DN. We first performed a gene analysis that was based on genome-wide transcriptional profiles from both patients with DN and control donors to obtain the gene coexpression module and identified the hub genes that are the key connection for various signaling pathways. Next, we correlated expression or loss of these hub genes with clinical features of patients with DN, particularly proteinuria and eGFR. Through these analyses, we identified loss of expression of SRGAP2a as a pivotal factor in renal dysfunction in patients with diabetes.
This study shows that SRGAP2a is significantly downregulated in DN. Given that SRGAP2a knock down resulted in an increase of podocyte motility and developmental defects, whereas exogenous SRGAP2a attenuated podocyte injury, the results strongly argue that SRGAP2a is a key molecule required for maintaining the proper function of podocytes. By using intravital and kidney slice two-photon imaging of the three-dimensional structure of mouse podocytes, Brähler et al. (46) showed that normal podocytes remain in a nonmotile state in vivo but shift to a dynamic state during foot process effacement. The current results of a specific expression of SRGAP2a in podocytes and its role in suppressing podocyte migration agree with the notion that a decrease in podocyte motility is critical for maintaining normal podocyte function.
The function of SRGAP2 has been extensively studied in neurons, revealing that it interacts with members of the Rho family of small GTPases. The Rho family of GTPases, particularly Cdc42, RhoA, and Rac1, play a critical role in the dynamics of the podocyte actin cytoskeleton (47). On one hand, deficiency of Cdc42 or RhoA results in heavy proteinuria, foot process effacement, and glomerulosclerosis accompanied by podocyte cytoskeleton injury (20,22,48). On the other hand, the exceeding activation of RhoA in adult podocytes facilitates foot process effacement and proteinuria through disturbing actin cytoskeleton and focal adhesion (47,49). Therefore, tightly controlled Rho-GTPase activities are essential for maintaining stable podocyte actin cytoskeletons and normal podocyte function. Previous studies also have shown that podocytes can switch from a RhoA-dependent stable state to a Cdc42- and Rac1-dependent migratory state in response to stress (50). In certain states, RhoA and Rac1 were mutually antagonistic (47,51). To identify which small Rho-GTPase interacts with SRGAP2a in podocytes, we analyzed the interactions of SRGAP2a with RhoA, Rac1, and Cdc42 in human renal biopsy specimens. A significant reduction in SRGAP2a-RhoA and SRGAP2a-Cdc42 protein complexes were detected in renal samples from patients with DN compared with those from control donors, which suggests that SRGAP2a executes its function through modulating RhoA and Cdc42 activity. We also observed that knockdown of SRGAP2a increases the active forms of Cdc42 and RhoA but not Rac1 in cultured human podocytes, whereas upregulating podocyte SRGAP2a in vitro also reduces the activity of RhoA and Cdc42. Given that SRGAP2a could inhibit RhoA/Cdc42 activation and that patients with DN displayed an impaired interaction of SRGAP2a with RhoA/Cdc42 in renal tissues, these results support the notion that under the stress conditions of hyperglycemia and TGF-β stimulation in which podocyte RhoA and Cdc42 were overactivated, suppressing RhoA and Cdc42 activity by SRGAP2a is critical to stabilizing the podocyte cytoskeleton and maintaining normal podocyte function. In other words, loss of SRGAP2a in podocytes under such injury cues (hyperglycemia or TGF-β) leads to a greater podocyte migration and renal injury.
Although the data show that increasing SRGAP2a levels protects the podocyte F-actin stress fiber from disruption by hyperglycemia or TGF-β, overexpression of SRGAP2a under normal conditions did not enhance the level of F-actin stress fibers in untreated podocytes (Fig. 3). This finding suggests that SRGAP2a only affects RhoA and Cdc42, which are overactivated by hyperglycemia or TGF-β. In other words, SRGAP2a plays an important role in balancing the activity of the Rho family of GTPases and thus the stabilization of the podocyte cytoskeleton network. Contrary to the results that podocyte SRGAP2a inactivates RhoA and Cdc42 under hyperglycemic conditions or when cells are treated with TGF-β, Guerrier et al. (28) showed that neuronal cell SRGAP2a inhibits Rac1 activity but induces Cdc42 activity. Given that the alteration of cellular F-actin by SRGAP2 is through its interaction with the Rho family of GTPases, these results suggest that the interactions of SRGAP2a with Cdc42, RhoA, or Rac1 and their functions may vary in different cell types and under different pathological conditions. Studies by Fan and colleagues (52,53) showed that SRGAP1, another SRGAP family member, is expressed on the basal surface of podocytes and promotes podocyte detachment through interacting with myosin II regulatory light chain. Different from the function of SRGAP2a that we observed to stabilize podocyte F-actin fibers, they showed that SRGAP1 actually decreased polymerization of F-actin. Two possible explanations exist for this discord. First, SRGAP2a and SRGAP1 may bind to different Rho family of small GTPases under various conditions, thus mediating different signals downstream. For example, Coutinho-Budd et al. (54) showed that the F-BAR domains from SRGAP1, SRGAP2, and SRGAP3 regulate cell membrane deformation differently. In addition, Yamazaki et al. (55) showed that instead of inactivating Cdc42 or RhoA, SRGAP1 inactivates Rac1 and that depletion of SRGAP1 leads to Rac1 overactivation but RhoA inactivation. In combination with the current observation that SRGAP2a inactivates RhoA and Cdc42 but not Rac1, these results support the notion that SRGAP1 and SRGAP2 play opposite roles in terms of activating/inactivating the Rho family of GTPases. Second, the role of Robo2 signaling mediated by the same SRGAP protein in modulating F-actin polymerization/depolymerization can be completely different in various biological set-ups, and the current data support that rescue of F-actin disruption by overexpression of SRGAP2a only occurs in injured podocytes (by hyperglycemia or TGF-β) instead of normal podocytes.
In summary, this study has identified the SRGAP2a-mediated SLIT/ROBO signal pathway as a novel mechanism regulating the adhesion and motility of podocytes. Through this signaling pathway, SRGAP2a inactivates RhoA and Cdc42 in podocytes and thus decreases podocyte motility and foot process effacement. Given that podocyte SRGAP2a level was reduced in patients with DN and that enhancing podocyte SRGAP2a expression could rescue the impairment of podocyte function in db/db mice, the findings suggest podocyte SRGAP2a as a potential therapeutic target in mitigating podocyte injury and proteinuria in DN.
See accompanying article, p. 550.
Acknowledgments. The authors thank Jill Littrell (Georgia State University) for critical reading and editing of the manuscript. All the human samples were from Renal Biobank of National Clinical Research Center of Kidney Diseases.
Funding. This work is supported by grants from the National Key Research and Development Program of China (2016YFC0904100), National Key Technology R&D Program (2015BAI12B02), Major International (Regional) Joint Research Project (81320108007), Key Research and Development Program of Jiangsu Province (BE2016747), National Natural Science Foundation of China (No. 81500548, 81500556, and 81300652), and the Innovation Capability Development Project of Jiangsu Province (BM2015004).
Duality of Interest. No potential conflicts of interest relevant to this article were reported.
Author Contributions. Y.P., S.J., Q.H., D.Q., J.S., L.W., Z.C., M.Z., A.D., W.Q., K.Z., and Z.L. read and approved the final version of the manuscript. Y.P., S.J., K.Z., and Z.L. designed the study and wrote the manuscript. Y.P., S.J., and J.S. participated in data bioinformatics analysis. S.J., Q.H., D.Q., J.S., L.W., Z.C., M.Z., and A.D. performed the experiments. Y.P. and S.J. prepared the figures. Z.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.