Kindlin-2 Association with Rho GDP-Dissociation Inhibitor α Suppresses Rac1 Activation and Podocyte Injury

Abstract

Alteration of podocyte behavior is critically involved in the development and progression of many forms of human glomerular diseases. The molecular mechanisms that control podocyte behavior, however, are not well understood. Here, we investigated the role of Kindlin-2, a component of cell-matrix adhesions, in podocyte behavior in vivo. Ablation of Kindlin-2 in podocytes resulted in alteration of actin cytoskeletal organization, reduction of the levels of slit diaphragm proteins, effacement of podocyte foot processes, and ultimately massive proteinuria and death due to kidney failure. Through proteomic analyses and in vitro coimmunoprecipitation experiments, we identified Rho GDP-dissociation inhibitor α (RhoGDIα) as a Kindlin-2–associated protein. Loss of Kindlin-2 in podocytes significantly reduced the expression of RhoGDIα and resulted in the dissociation of Rac1 from RhoGDIα, leading to Rac1 hyperactivation and increased motility of podocytes. Inhibition of Rac1 activation effectively suppressed podocyte motility and alleviated the podocyte defects and proteinuria induced by the loss of Kindlin-2 in vivo. Our results identify a novel Kindlin-2–RhoGDIα–Rac1 signaling axis that is critical for regulation of podocyte structure and function in vivo and provide evidence that it may serve as a useful target for therapeutic control of podocyte injury and associated glomerular diseases.

Glomerular diseases affect millions of people worldwide and are growing at an annual rate of 5%–8%.¹ Numerous evidence indicates that podocyte injury and loss are key factors in the development and progression of glomerular diseases including FSGS, diabetic nephropathy, and membranous glomerulopathy.^2,3 With respect to their cytoarchitecture, podocytes are divided into three structural and functional segments: cell body, major processes, and foot processes (FP). FP tightly enwrap the glomerular capillary walls and form unique cell-cell junctions known as silt diaphragms (SD).⁴ The structure of FP and SD is believed to be essential for the functions of podocytes, establishing the selective permeability of the glomerular filtration barrier.⁵ Recent studies indicated that at least two major causes are linked to podocytopenia: (1) direct podocyte loss due to podocyte apoptosis or detachment from glomerular basement membrane (GBM); and (2) FP effacement or flattening, including interference with the actin cytoskeleton and disruption of SD complex.⁶

Kindlin-2 is an important component of cell-matrix adhesions, which has been implicated in linking integrins to the actin cytoskeleton and transducing bidirectional signaling.^7–14 Kindlin-2 belongs to the Kindlin family which also includes Kindlin-1 and -3.^9,15–19 In contrast to Kindlin-1 and -3, which are expressed primarily in epithelial and hematopoietic/endothelial cells, respectively, Kindlin-2 is ubiquitously expressed.^15,20,21 Global deletion of Kindlin-2 in mice results in peri-implantation lethality due to extensive detachment of the endoderm and epiblast,^13,22 demonstrating a critical role of Kindlin-2 in early embryonic development. Recently, using a conditional knockout strategy, we have found that Kindlin-2 is essential for chondrogenesis.²³ The functions of Kindlin-2 in the homeostasis of many organs, however, are largely unknown.

We have previously found that Kindlin-2 is highly expressed in human podocyte cells in culture.¹¹ Depletion of Kindlin-2 in cultured podocytes reduced cell-matrix adhesion and fibronectin matrix deposition.¹¹ The role of Kindlin-2 in renal glomerular functions and homeostasis in vivo, however, is not known. In this study, we have employed a conditional knockout strategy to determine the function of Kindlin-2 in glomerular podocytes in vivo. Furthermore, we have investigated the molecular mechanism by which Kindlin-2 functions in control of podocyte behavior. Our results identify a novel Kindlin-2–RhoGDIα–Rac1 signaling axis that is critical for control of podocyte behavior in vivo and provide evidence that this signaling axis may serve as a useful target for therapeutic intervention of podocyte injury and associated glomerular diseases.

Results

Generation of Podocyte-Specific Kindlin-2^Neph2 cKO Mice

Because global Kindlin-2 gene depletion leads to embryonic lethal in mice,¹³ we have generated podocyte-specific Kindlin-2 knockout mice (referred to as Kindlin-2^Neph2 cKO hereafter) to facilitate the studies of Kindlin-2 in glomerular podocyte function. Kindlin-2^Neph2 cKO mice were generated using the Cre-Lox system (Neph2-Cre) that targets exons 5 and 6 of the Kindlin-2 allele (Figure 1A).²³ Mice of all genotypes were born at the expected Mendelian frequency. Kindlin-2^Neph2 cKO mice, as well as Kindlin-2^fl/+;Neph2-Cre, Kindlin-2^+/+;Neph2-Cre (referred to as WT hereafter), were confirmed by PCR analysis of tail genomic DNA (Figure 1B). Immunoblotting and immunofluorescence analyses demonstrated that Kindlin-2 expression was markedly diminished in the Dynabeads-isolated podocyte-enriched cell fractions obtained from the Kindlin-2^Neph2 cKO mice (Figure 1, C and D) compared with that from the WT littermates. Furthermore, Kindlin-2 deletion was confirmed by immunofluorescence labeling of kidney tissue samples. In WT mice, Kindlin-2 was expressed in cytoplasm of podocytes (Figure 1E) and partially colocalized with podocyte marker nephrin (Figure 1F). By contrast, Kindlin-2 expression was barely detectable in the podocytes of Kindlin-2^Neph2 cKO mice (Figure 1, E and F).

Figure 1.

Generation of podocyte-specific Kindlin-2^Neph2 cKO mice. (A) The diagram depicts the strategy for generation of Kindlin-2^Neph2 cKO (cKO). Mice expressing Neph2-Cre were bred with mice carrying floxed Kindlin-2 locus (exons 5 and 6). (B) Representative PCR analysis of extracted genomic DNA from tail clippings. PCR product bands of floxed (300 bp) and wild-type (200 bp) are shown. Cre PCR product (650 bp) is also indicated. (C) Immunoblotting analysis of Kindlin-2 expression in isolated primary podocytes from WT and Kindlin-2^Neph2 cKO mice. (D) Podocytes harvested from WT and Kindlin-2^Neph2 cKO mice were plated on laminin-coated glass coverslips and stained for WT1 (red) and for Kindlin-2 (green). Scale bars, 10 μm. (E and F) Double-immunofluorescence detection of Kindlin-2 and WT1 (E) or Kindlin-2 and nephrin (F) on kidney sections of WT and Kindlin-2^Neph2 cKO mice. Scale bars, 10 μm. The lower panels of (F) show higher magnification images of the areas outlined with white lines in the upper panels. The experiments in Figure 1, B–F are representative of 4–6 independent experiments. Arrows indicate WT1-positive podocytes. E, Exon; fl, flox; WT, wild-type.

Kindlin-2^Neph2 cKO Mice Develop Proteinuria and Kidney Failure

To assess the functional consequence of Kindlin-2 depletion in glomerular function, we first examined the gross phenotype of Kindlin-2^Neph2 cKO mice. Kindlin-2^Neph2 cKO mice appeared normal at birth. However, the body weight of Kindlin-2^Neph2 cKO mice began to decrease at 4 weeks of age and this growth retardation became more apparent by 8 weeks of age, compared with that of WT mice (Figure 2, A and B). Moreover, many Kindlin-2^Neph2 cKO mice died and the median age of Kindlin-2^Neph2 cKO mice at death is 4 weeks (Figure 2C). To determine the cause of early death, we isolated kidneys from 8-week-old Kindlin-2^Neph2 cKO mice and found that they were pale with firm appearance and a granular surface (Figure 2G), suggesting that the kidney function was impaired. Thus, we collected urine from WT and Kindlin-2^Neph2 cKO mice at 1, 2, 4, and 8 weeks of age for proteinuria screening. Although proteinuria was not detected in 1-week-old Kindlin-2^Neph2 cKO mice, 2-week-old Kindlin-2^Neph2 cKO mice began to show selective albuminuria and developed massive proteinuria around 4–8 weeks of age. In contrast, none of the WT mice showed proteinuria (Figure 2D). Consistent with this result, plasma albumin levels were dramatically reduced in Kindlin-2^Neph2 cKO mice at 2 weeks and after (Figure 2E). Quantitation of the urine albumin-to-creatinine ratio revealed an increase of several hundreds of magnitude in Kindlin-2^Neph2 cKO mice compared with WT mice beginning at 2 weeks (Figure 2F). Renal function was further determined by measuring plasma creatinine, and an elevation of plasma creatinine level was observed in Kindlin-2^Neph2 cKO mice at an age as early as 2 weeks compared with that of WT littermates (Figure 2H). Collectively, these results demonstrate that the kidney function in Kindlin-2^Neph2 cKO mice is severely impaired.

Figure 2.

Podocyte-specific Kindlin-2^Neph2 cKO mice develop proteinuria and kidney failure. (A) Kindlin-2^Neph2 cKO mice fail to gain weight by 6 weeks of age compared with WT mice. ***P<0.001 versus WT; n=6 mice at each time point. (B) Representative picture shows the phenotypic appearance of WT and Kindlin-2^Neph2 cKO mice at 8 weeks of age. (C) Survival curve of Kindlin-2^Neph2 cKO mice shows 100% mortality by 10 weeks of age. n=20 (WT), n=22 (cKO). (D and E) SDS-PAGE analysis of albumin levels in the urine (D) and plasma (E) at different time points in WT and Kindlin-2^Neph2 cKO mice. (F) Quantification of urinary albumin normalized to creatinine at 1, 2, 4, and 8 weeks of age. ***P<0.001 versus WT; n=6 mice at each time point. (G) Kidneys of Kindlin-2^Neph2 cKO mice are paler and smaller than those of controls and have an irregular appearance at 8 weeks of age. (H) Elevated plasma creatinine in Kindlin-2^Neph2 cKO mice at 2, 4, and 8 weeks of age. ***P<0.001 versus WT; n=6 mice at each time point. cKO, conditional knockout; W, week; WT, wild-type.

Loss of Kindlin-2 in Podocytes Results in Glomerular Fibrosis and Disruption of Podocyte FP Architecture

To further investigate the functions of Kindlin-2 in renal function, we examined renal histology of Kindlin-2^Neph2 cKO mice. The kidneys of Kindlin-2^Neph2 cKO mice displayed drastically increased segmental glomerulosclerosis, mesangial expansion, and dilated distal tubules containing proteinaceous cast by week 4 (Figure 3A; quantified in Figure 3, B and C). Analyses of glomeruli by scanning electron microscopy revealed abnormal ultrastructural architecture and microvillous transformation on the apical surface of podocytes of Kindlin-2^Neph2 cKO mice. By week 8, only a very small number of podocytes could be found in the glomeruli of Kindlin-2^Neph2 cKO mice (Figure 4A, Supplemental Figure 1, also see below).

Figure 3.

Podocyte-specific Kindlin-2^Neph2 cKO mice develop glomerulosclerosis and interstitial fibrosis. (A) Kidney sections from Kindlin-2^Neph2 cKO and control littermates at 2, 4, and 8 weeks were subjected to H&E, PAS, or trichrome staining. Arrows indicate diffuse tubular dilatation and proteinaceous casts. Asterisks indicate fibrotic areas. Higher magnification images of representative areas of trichrome staining (indicated by squares) with interstitial fibrosis are shown on the right. Scale bar, 40 μm. (B) Quantification of interstitial fibrosis at 2, 4, and 8 weeks of age. ***P<0.001 versus WT; n=10 mice at each time point. (C) Quantification of glomerulosclerosis at 2, 4, and 8 weeks of age. ***P<0.001 versus WT; n=10 mice at each time point. cKO, conditional knockout; WT, wild-type.

Figure 4.

Podocyte-specific deletion of Kindlin-2 causes ultrastructural changes of podocytes. (A) Low-magnification (left panel) and high-magnification (right panel) scanning electron microscopy analysis of podocyte ultrastructure at 1 and 8 weeks of age Kindlin-2^Neph2 cKO and WT mice. Podocyte FP and microvillous transformation on the apical surface of podocytes can been seen in Kindlin-2^Neph2 cKO mice. Scale bar, 5 μm. (B) Low-magnification (left panel) and high-magnification (right panel) TEM analyses identified abnormal FP and GBM structure in Kindlin-2^Neph2 cKO at various ages compared with WT mice. Arrow indicates fused FP, arrowhead denotes naked GBM. Scale bar, 5 μm (left panel), 0.5 μm (right panel). (C) Quantification of SD density (i.e., the number of FP per micrometer of GBM) in Kindlin-2^Neph2 cKO and WT mice at different time points. ***P<0.001 versus WT; n=3 mice at 1 and 2 weeks of age; n=5 mice at 4 and 8 weeks of age. (D) Quantification of GBM thickness in Kindlin-2^Neph2 cKO and WT mice at different time points. ***P<0.001 versus WT; n=3 mice at 2 weeks of age; n=5 mice at 4 and 8 weeks of age. cKO, conditional knockout; WT, wild-type.

Consistent with these findings, transmission electron microscopy (TEM) showed prominent FP effacement, GBM membrane irregularities, and invaginations in the glomeruli of Kindlin-2^Neph2 cKO mice at as early as 2 weeks (Figure 4B). By week 8, podocytes were detached from the GBM, leading to a denuded appearance in some areas (Figure 4B). We quantified the average number of FP per micrometer length of basement membrane and found that it was reduced by 60% in Kindlin-2^Neph2 cKO mice at as early as 2 weeks compared with that of WT littermates (Figure 4C). It is worth noting that the FP effacement occurred earlier than GBM thickening (Figure 4D), suggesting that Kindlin-2 deletion is detrimental to podocytes, which likely contributes to initial albuminuria in Kindlin-2^Neph2 cKO mice.

Loss of Kindlin-2 Induces Podocyte Detachment and Apoptosis

The reduction of podocyte number per glomerulus is a hallmark in the development of glomerulus failure. The phenotypes of Kindlin-2^Neph2 cKO mice prompt us to examine the effects of Kindlin-2 depletion on the fate of podocytes. We calculated the average numbers of podocytes per glomerulus by counting podocyte-specific transcription factor WT1-positive podocytes. Immunohistochemistry results showed that WT1-positive cells were dramatically reduced in 4-week-old Kindlin-2^Neph2 cKO mice compared with WT littermates. By 8 weeks of age, approximately 85% of WT1 expression was lost in Kindlin-2^Neph2 cKO mice, confirming the severe loss of podocytes (Figure 5A; quantified in Figure 5B). To test whether the loss of podocytes is due to podocyte apoptosis and/or detachment, we assessed apoptosis in mouse kidney tissue samples by staining with an antibody specific for activated caspase–3 (Figure 5C, Supplemental Figure 2A) and terminal deoxynucleotidyl transferase-mediated dUTP nick-end labeling (TUNEL) staining (Supplemental Figure 2B), respectively. As shown in Figure 5C, although apoptosis was barely detectable in the glomeruli of WT mice, we occasionally identified activated caspase–3–positive cells in the glomeruli or tubule lumens of Kindlin-2^Neph2 cKO mice. Moreover, we found a significant increase in the level of urinary nephrin, a marker of podocytes, in 4-week-old Kindlin-2^Neph2 cKO mice (Figure 5D), suggesting that Kindlin-2 deficiency may cause podocyte detachment as well as apoptosis. Because 2-week-old Kindlin-2^Neph2 cKO mice exhibited proteinuria and podocyte FP effacement, which occurred before the loss of podocyte was detected (Figures 2D, 4B, and 5, A and B), we postulated that there might be structural alterations other than podocyte loss that contribute to the initiation of proteinuria in Kindlin-2^Neph2 cKO mice.

Figure 5.

Podocyte-specific deletion of Kindlin-2 results in loss of podocytes. (A) Representative images of kidney sections stained with WT1, a podocyte-specific transcription factor, at 1, 2, 4, and 8 weeks Kindlin-2^Neph2 cKO and WT mice. Arrow denotes the glomerulus with reduced podocyte number in Kindlin-2^Neph2 cKO mice. (B) Quantification of WT1 staining per glomerulus was performed with Kindlin-2^Neph2 cKO and WT mice at different time points. ***P<0.001 versus WT; n=4 mice at each time point. At least 30 glomeruli/mouse were counted. (C) Kidney sections were subjected to anti-caspase3 staining to identify apoptotic cells. Arrows indicate apoptotic cells present in both glomerular and tubulointerstitial areas at 4 weeks Kindlin-2^Neph2 cKO mice. Caspase3-positive cells were only observed in tubulointerstitial areas at 8 weeks Kindlin-2^Neph2 cKO mice. No apoptotic cells were found in WT mice. n=4 mice at each time point. (D) Representative immunoblotting of urine samples from Kindlin-2^Neph2 cKO and WT mice at 4 weeks analyzed with an antibody for nephrin, a marker for podocytes. Note that nephrin is present in the urine of 4 weeks Kindlin-2^Neph2 cKO mice but not that of WT mice, indicating the presence of podocytes in the urine of Kindlin-2^Neph2 cKO mice. Original magnification, ×200 (in A and C). cKO, conditional knockout; G, glomerular; IB, immunoblotting; WT, wild-type.

Loss of Kindlin-2 Results in Aberrant Expression and Distribution of SD Proteins and Abnormal Actin Arrangement

It is known that dysregulation of SD proteins and/or actin cytoskeletal organization in podocytes can contribute to FP effacement and, ultimately, proteinuria and kidney failure.² To test whether loss of Kindlin-2 affects SD proteins and actin cytoskeleton organization, we compared the expression and localization of SD proteins (nephrin and ZO-1) and actin-associated proteins (synaptopodin and α-actinin4) in WT and Kindlin-2^Neph2 cKO mice, respectively. As shown in Figure 6A, in WT glomeruli, staining with antibodies for SD/actin-associated proteins revealed a linear pattern along the GBM, corresponding to their localization close to the SD. However, in glomeruli from Kindlin-2^Neph2 cKO mice, aberrant distribution and reduced expression of these SD/actin-associated proteins were observed (Figure 6A). In contrast, the expression of desmin and α–smooth muscle actin (α-SMA), proteins that are normally expressed primarily in mesenchymal cells, was markedly increased in Kindlin-2^Neph2 cKO mice glomeruli compared with that in WT glomeruli (Figure 6A), suggesting that podocytes display a fibroblastic phenotypic change upon Kindlin-2 depletion. To corroborate the immunofluorescence data, immunoblotting was performed using primary podocytes isolated from Kindlin-2^Neph2 cKO and WT mice. The results of these experiments were consistent with that of the microscopic analyses (Figure 6B). Collectively, these data demonstrate that podocyte-specific ablation of Kindlin-2 causes aberrant distribution and expression of SD/actin-associated proteins and induces epithelial-mesenchymal transition–like phenotypic changes in podocytes.

Figure 6.

Podocyte-specific deletion of Kindlin-2 results in dysregulation of SD and actin cytoskeleton. (A) Kidney sections from Kindlin-2^Neph2 cKO and WT mice were stained with various antibodies against SD/actin-associated proteins ZO-1, synaptopodin, α-actinin4, and nephrin as well as mesenchymal proteins desmin and α-SMA. Representative micrographs are shown. Scale bar, 10 μm. (B) Immnoblotting analysis reveals that knockout of Kindlin-2 in podocytes reduces ZO-1, α-actinin4, nephrin, and synaptopodin expression. (C and D) Representative photomicrographs of cell morphology and immunofluorescence staining for F-actin (green) (C) or Vinculin (green) (D) in primary podocytes isolated from Kindlin-2^Neph2 cKO and WT mice. Podocyte nuclei were visualized with WT1 (red). Scale bars, 50 μm (C) or 10 μm (D). Quantification data are shown in the right panel. ***P<0.001 versus WT (C); **P<0.01 versus WT (D); n=4 independent experiments. cKO, conditional knockout; WT, wild-type.

The actin filament bundles form the backbone of terminal FP, and disruption of this cytoskeletal assembly is thought to underlie FP effacement.^24–26 Thus, we next examined the effects of Kindlin-2 on actin cytoskeleton organization in podocytes. Consistent with previous studies,²⁷ wild-type podocytes showed well defined transcellular stress fibers (Figure 6C, left panel), whereas primary podocytes derived from Kindlin-2^Neph2 cKO mice displayed severe loss of stress fibers (Figure 6C, middle panel) with a significant number (20%) of cKO podocytes showing prominent changes in cell shape (e.g., more rounded and contraction, see Figure 6C, right panel). Kindlin-2 depletion also altered the number and shape of focal adhesions (Figure 6D), confirming an important role of Kindlin-2 in podocyte actin cytoskeleton organization.

Kindlin-2 Regulates Rac1 Activation through Association with RhoGDIα

To investigate the mechanism by which Kindlin-2 regulates actin cytoskeleton organization, we sought to examine the activity of Rho-GTPases, which are known to play an important role in regulation of actin cytoskeleton organization.^28,29 Because the number of podocytes was dramatically decreased in Kindlin-2^Neph2 cKO mice (Figure 5, A and B), it was technically difficult to isolate a sufficient number of primary Kindlin-2–deficient podocytes for the biochemical analyses of Rho-GTPases. Thus, we knocked down Kindlin-2 in immortalized human podocytes using two different shRNAs targeting Kindlin-2 mRNA (shRNA1 and shRNA2), and an shRNA that does not correspond to any known human gene was used as a control. Knockdown (KD) of Kindlin-2 was confirmed by immunoblotting, which showed that >70% of endogenous Kindlin-2 protein was depleted in cells transfected with either shRNA (Figure 7A). As expected, Kindlin-2 KD reduced actin stress fiber and focal adhesion formation (Supplemental Figure 3, A and B). We next analyzed the activation of Rac1 and RhoA, respectively, in both Kindlin-2 KD and control podocytes. We found that Rac1 was significantly activated in Kindlin-2 KD podocytes, whereas the activation of RhoA was not significantly altered (Figure 7B; quantified in Figure 7C). To further test this, we analyzed cell motility, which is known to be promoted by Rac1. Consistent with the increase of Rac1 activation, podocytes lacking Kindlin-2 exhibited a marked increase in cell motility compared with control podocytes (Figure 7, D and E). To further assess the effect of Kindlin-2 on podocyte motility, we tracked individual podocytes under time-lapse microscopy. The results showed that whereas control podocytes were of low motility (average velocity=0.49 µm/min), KD of Kindlin-2 with shRNAs significantly increased the migratory velocity of podocytes (average velocity was increased to 0.89 µm/min with shRNA1 and to 0.84 µm/min with shRNA2, see Figure 7F). Interestingly, depletion of Kindlin-2 in HT1080 fibrosarcoma cells reduced rather than increased cell motility (Supplemental Figure 4), indicating that not all Kindlin-2–deficient cells exhibit a hyper-motile phenotype.

Figure 7.

Podocyte-specific deletion of Kindlin-2 promotes Rac1 activation. (A) KD of Kindlin-2 in immobilized human podocytes. Quantification data are shown in the lower panel. ***P<0.001 versus control; n=8 independent experiments. (B) Representative immunoblotting analysis of Rac1 and RhoA activity in Kindlin-2 KD and control podocytes. (C) Quantification of Rac1 and RhoA activity. *P<0.02 versus control; **P<0.01 versus control; n=4 independent experiments. (D) KD of Kindlin-2 increases podocyte migration. Representative images (upper panel) and quantification analysis (lower panel) of transwell migration assay were shown. *P<0.02 versus control; **P<0.01 versus control; n=4 independent experiments. (E) Representative images (upper panel) and quantification analysis (lower panel) of transwell migration assay were shown. Primary podocytes isolated from Kindlin-2^Neph2 cKO and WT mice were used. ***P<0.001 versus WT; n=3 independent experiments. (F) Multiple tracks of individual podocytes (left panel) and quantification of the velocity (right panel) from Kindlin-2 KD and control podocytes were shown. n=3 independent experiments with >160 cell tracks. ***P<0.001 versus control. cKO, conditional knockout; WT, wild-type.

To understand mechanistically how Kindlin-2 promotes Rac1 activation, we performed proteomic analyses as an initial screen to identify novel proteins potentially associated with Kindlin-2. To do this, we immunoprecipitated endogenous Kindlin-2 from podocyte lysates with an anti–Kindlin-2 antibody and submitted the anti–Kindlin-2 immunoprecipitates for LC-MS/MS analysis. The results showed that Rho-GDP dissociation inhibitor α (RhoGDIα), which binds GDP-bound Rac1 and acts as an inhibitor of Rac1 activation,³⁰ associates with Kindlin-2 (Supplemental Table 1). To confirm this, we performed coimmunoprecipitation (co-IP) experiments with either anti-Kindlin-2 or anti-RhoGDIα antibodies. The data showed that endogenous RhoGDIα was readily co-IPed with Kindlin-2 (Figure 8A). Reciprocally, endogenous Kindlin-2 was co-IPed with RhoGDIα (Figure 8B), confirming that Kindlin-2 associates with RhoGDIα. The association of Kindlin-2 with RhoGDIα was also detected in other cell types such as BT-549 breast cancer cells (Supplemental Figure 5). To further test this association, we generated GST fusion proteins containing the full length, the N-terminal regulatory domain (residues 1–67), or the C-terminal IgG-like domain (residues 68–204) of RhoGDIα and analyzed their Kindlin-2–binding activity using pulldown assays. As expected, Kindlin-2 was readily pulled down by GST-RhoGDIα (Figure 8C). Furthermore, GST fusion protein containing the N-terminal regulatory domain, but not that containing the C-terminal IgG-like domain, was also able to pull down Kindlin-2 (Figure 8C). It is worth noting, however, that the amount of Kindlin-2 pulled downed by the RhoGDIα N-terminal regulatory domain was smaller than that pulled downed by the full length RhoGDIα (Figure 8C), suggesting that Kindlin-2 associates primarily with the N-terminal regulatory domain of RhoGDIα, albeit the presence of the C-terminal IgG-like domain may strengthen the association with Kindlin-2.

Figure 8.

Kindlin-2 regulates Rac1 activation through association with RhoGDIα. (A) Immobilized human podocyte lysates were immunoprecipitated with anti–Kindlin-2 antibody or normal IgG followed by immunoblotting with antibodies as indicated. The presence of Kindlin-2 and RhoGDIα in cell lysates is shown as input. (B) Immobilized human podocyte lysates were immunoprecipitated with anti-RhoGDIα antibody or normal IgG followed by immunoblotting with antibodies as indicated. (C) GST fusion proteins containing the full-length, N-terminal (residues 1–67) or C-terminal (residues 68–204) fragment of RhoGDIα were used to pull down Kindlin-2 from human podocytes (lower panel). Upper panel: schematic illustration of RhoGDIα fragments that were used in the GST pull-down assay. Note that Kindlin-2 was pulled down by the N-terminal (residues 1–67) but not the C-terminal (residues 68–204) fragment of RhoGDIα. (D) Immunoblotting analysis of RhoGDIα protein levels in primary podocytes isolated from Kindlin-2^Neph2 cKO and WT mice (left panel). Quantification data are shown in the right panel. ***P<0.001 versus WT; n=7 independent experiments. (E) RhoGDIα immunoprecipitates from control and Kindlin-2 KD podocytes were analyzed by immunoblotting with anti-Rac1 and anti-RhoGDIα antibodies. Normal IgG is used as negative control. n=4 independent experiments. IB, immunoblotting; IP, immunoprecipitation; cKO, conditional knockout; WT, wild-type.

Next, we examined RhoGDIα protein expression in Kindlin-2^Neph2 cKO podocytes. Immunoblotting showed that the level of RhoGDIα protein was modestly reduced in both Kindlin-2^Neph2 cKO podocytes and Kindlin-2 KD podocytes compared with that in control cells (Figure 8D, Supplemental Figure 6). For Rac1 activation to occur, GDP-bound Rac1 must first be dissociated from RhoGDIα³⁰. Thus, we tested whether Rac1 association with RhoGDIα is affected by the loss of Kindlin-2. To do this, we analyzed the amount of Rac1 associated with RhoGDIα by co-IP experiment and found that the association of Rac1 with RhoGDIα was drastically reduced in Kindlin-2 KD podocytes compared with that in control cells (Figure 8E). Collectively, these results revealed a novel Kindlin-2–RhoGDIα–Rac1 signaling axis, in which downregulation of Kindlin-2 causes reduction of the level of RhoGDIα and dissociation of Rac1 from RhoGDIα, resulting in hyperactivation of Rac1 and increased podocyte motility.

Inhibition of Rac1 Reverses the Podocyte Defects Induced by the Loss of Kindlin-2 In Vivo

To test whether the Kindlin-2–RhoGDIα–Rac1 signaling axis described above contributes to the dysfunction of podocytes and renal failure in vivo, we tested the effect of treatment of the mice with NSC23766, a Rac1-specific inhibitor,³¹ on the phenotype induced by the loss of Kindlin-2. To better assess this, we used two different age groups of Kindlin-2^Neph2 cKO and WT mice (2 and 6 weeks of age, respectively). For the first group (i.e., 2 weeks of age), 35 WT mice and 72 Kindlin-2^Neph2 cKO mice were used. The 72 Kindlin-2^Neph2 cKO mice were divided into 35 subgroups, in which 33 subgroups contain two Kindlin-2^Neph2 cKO mice and two subgroups contain three Kindlin-2^Neph2 cKO mice (all Kindlin-2^Neph2 cKO mice in the same subgroup were littermates). For each subgroup, one or two Kindlin-2^Neph2 cKO mice were treated with NSC23766 for 2 weeks and the other (littermate) was treated with PBS for the same length of time as a control. The results showed that NSC23766 treatment significantly increased the survival of Kindlin-2^Neph2 cKO mice. Specifically, whereas 16 out of 35 Kindlin-2^Neph2 cKO mice died at week 4 in the PBS-treated group, only six out of 37 Kindlin-2^Neph2 cKO mice died at week 4 in the NSC-treated group. All 35 WT mice were alive at week 4 (Figure 9A). Because six out of 37 Kindlin-2^Neph2 cKO mice treated with NSC23766 still died (none of the WT mice died), treatment with NSC23766 did not completely rescue the defects caused by the loss of Kindlin-2. However, this number was much smaller than that of Kindlin-2^Neph2 cKO mice that were not treated with NSC23766 (six out of 37 versus 16 out of 35), indicating that treatment with NSC23766 does alleviate the defects caused by the loss of Kindlin-2. For the 35 subgroups of Kindlin-2^Neph2 cKO mice, 19 subgroups remained at the end of the treatment (one or more of the mice of the other 16 subgroups of Kindlin-2^Neph2 cKO mice died before the end of the treatment). These 19 subgroups of Kindlin-2^Neph2 cKO mice were used for further analyses. Specifically, six subgroups of the Kindlin-2^Neph2 cKO mice and six WT mice were used for IHC assay and albumin-to-creatinine measurement, five subgroups of the Kindlin-2^Neph2 cKO mice and five WT mice were used for scanning electron microscopy/TEM and albumin SDS-PAGE analyses, and eight subgroups of the Kindlin-2^Neph2 cKO mice and eight WT mice were used for primary podocyte isolation. For the second group (i.e., 6 weeks of age), 30 WT mice and 65 Kindlin-2^Neph2 cKO mice were used. The 65 Kindlin-2^Neph2 cKO mice were divided into 30 subgroups, in which 25 subgroups contained two Kindlin-2^Neph2 cKO mice and five subgroups contained three Kindlin-2^Neph2 cKO mice (all Kindlin-2^Neph2 cKO mice in the same subgroup were littermates). Kaplan–Meier survival curve reveals that NSC23766 treatment also increases survival of 6-week-old Kindlin-2^Neph2 cKO mice (Figure 9A). Specifically, 16 out of 30 of Kindlin-2^Neph2 cKO mice treated with PBS died during the 2-week treatment period (i.e., from week 6 to week 8). By contrast, only five out of 35 Kindlin-2^Neph2 cKO mice treated with NSC23766 were dead during the same period of time. All 30 WT mice were alive at week 8. These results confirm that treatment with Rac1 inhibitor NSC23766 alleviated—albeit it did not completely rescue—the defects induced by the loss of Kindlin-2. For the 30 subgroups of the Kindlin-2^Neph2 cKO mice, 12 subgroups of Kindlin-2^Neph2 cKO mice remained at the end of the treatment. Seven out of 12 subgroups of the Kindlin-2^Neph2 cKO mice and seven WT mice were used for IHC assay and albumin-to-creatinine measurement, and five out of 12 subgroups of the Kindlin-2^Neph2 cKO mice and five WT mice were used for scanning electron microscopy/TEM and albumin SDS-PAGE analyses. The results from these experiments showed that administration of NSC23766 partially restored the alterations of the gross appearance of kidneys resulting from the loss of Kindlin-2 (Figure 9B) and significantly reduced (albeit did not completely eliminate) albuminuria (Figure 9, C and D). Consistent with this, hematoxylin-eosin (H&E) and fibronectin staining analyses revealed that the distal tubule dilation, and glomerular and interstitial fibrosis, in 4-week-old and 8-week-old Kindlin-2^Neph2 cKO were significantly ameliorated after treatment with Rac1 inhibitor (Figure 9, E and F, Supplemental Figure 7). Moreover, scanning electron microscopy and TEM studies showed that podocyte ultrastructural architecture was partially restored by the treatment with Rac1 inhibitor (Figure 9, G and I, Supplemental Figure 8). Concordantly, treatment with NSC23766 reversed, at least in part, the reduction of RhoGDIα protein, Rac1 hyperactivation, and the increased podocyte motility caused by the loss of Kindlin-2 (Figure 9, J and L, Supplemental Figure 9). Treatment with NSC23766 did not reverse the decrease in podocyte-matrix adhesion (Figure 9M). Additionally, overexpression of Kindlin-2 or RhoGDIα also reduced Rac1 hyperactivation in Kindlin-2 KD podocytes (Supplemental Figure 9). Collectively, these results demonstrate that (1) the Kindlin-2–RhoGDIα–Rac1 signaling axis is critical for maintaining normal podocyte structure and function, and (2) increased activation of Rac1 is responsible, at least in part, for the albuminuria and renal failure induced by the loss of Kindlin-2.

Figure 9.

Treatment with NSC23766 (NSC) reverses the phenotypes induced by the loss of Kindlin-2 in mice. (A) Kaplan–Meier survival curves of 2-week-old (left panel) and 6-week-old (right panel) Kindlin-2^Neph2 cKO mice treated with NSC23766 (cKO+NSC) or PBS (cKO) for 14 days or WT controls. Note that NSC23766 treatment increases the mortality of Kindlin-2^Neph2 cKO mice in each age group (P<0.001, log-rank test). For the group of 2 weeks of age, 35 WT, 35 cKO, and 37 cKO+NSC mice were analyzed. For the group of 6 weeks of age, 30 WT, 30 cKO, and 35 cKO+NSC mice were analyzed. (B) Kidneys of WT, cKO, and cKO+NSC mice at 4 weeks of age. Three sets of the mice were analyzed and similar results were obtained. (B) Results from one of the representative sets. (C and D) Coomassie blue staining of SDS-PAGE analyses of urinary proteins (C) and quantification of albumin-to-creatinine levels in the urine (D) from 4 and 8 weeks of age WT, cKO, and cKO+NSC mice. ***P<0.001 versus other groups; n=6 mice for 4 weeks of age group; n=7 mice for 8 weeks of age group. (E) Renal tissues from WT, cKO, and cKO+NSC mice at 4 weeks were subjected to H&E staining (upper panel), and glomerular (middle panel) and tubulointerstitial (lower panel) fibronectin staining. Original magnification, ×200 for the upper and lower panels; ×400 for the middle panel. (F) Quantification of IHC scores for fibronectin staining within the glomeruli and the tubulointerstitial areas. ***P<0.001 versus other groups; n=6 mice for 4 weeks of age group; n=7 mice for 8 weeks of age group. (G and H) Low-magnification (left panel) and high-magnification (right panel) scanning electron microscopy (G) and TEM (H) analyses of podocyte ultrastructure in WT, cKO, and cKO+NSC mice at 4 weeks of age. Note that abnormal podocyte ultrastructure induced by the loss of Kindlin-2 was prevented to a great extent by NSC treatment. Scale bars, 5 μm (scanning electron microscopy and TEM, left panel) and 1 μm (TEM, right panel). (I) Quantification of the number of FP per micrometer of GBM in WT, cKO, and cKO+NSC mice at different time points. ***P<0.001 versus other groups; n=5 mice for each age group. (J) Immunoblotting analysis of RhoGDIα protein levels in primary podocytes isolated from WT, cKO, and cKO+NSC mice (upper panel). Quantification data are shown in the lower panel. *P<0.05 versus other groups; n=5 independent experiments. (K) Activity of Rac1 was determined by G-LISA assay. The primary podocytes isolated from WT, cKO, and cKO+NSC mice were used. ***P<0.001 versus other groups; n=8 independent experiments. (L) Cell migration. The migration of the control and Kindlin-2 KD podocytes treated with or without NSC was analyzed as described in the Concise Methods section. Representative images are shown in the left panel and quantification analyses are shown in the right panel. *P<0.02 versus control; ***P<0.001 versus control; n=4 independent experiments. (M) Cell adhesion. The adhesion of the control and Kindlin-2 KD podocytes treated with or without NSC to laminin was analyzed as described in the Concise Methods section. Representative images are shown in the left panel and quantification analyses are shown in the right panel. ***P<0.001 versus control; n.s. P>0.05 for NSC versus vehicle; n=4 independent experiments. cKO, conditional knockout; WT, wild-type.

Discussion

Alterations of the actin cytoskeleton in podocytes are critically involved in the development and progression of many human glomerular diseases. However, the signaling mechanisms that control actin cytoskeleton organization in podocytes are not well understood. In this study, we have employed a conditional knockout strategy to selectively ablate Kindlin-2 in podocytes in mice. Our results demonstrate a critical role of Kindlin-2 in regulation of podocyte actin cytoskeleton in vitro and in vivo. Loss of Kindlin-2 alters podocyte morphology, increases podocyte motility, and impairs glomerular filtration barrier function in vivo, resulting in massive proteinuria, renal failure, and ultimately death of the mice.

How does Kindlin-2 regulate the behavior of podocytes? Our aforementioned data demonstrate that FP effacement occurred earlier than podocyte detachment and apoptosis, suggesting that the initial effects of Kindlin-2 deletion on albuminuria may result from, at least in part, alterations of the actin cytoskeleton and SD.³² The Rho small G protein family members are well known to regulate various actin cytoskeleton–dependent cellular functions, including cell motility and morphologic changes. In mice, RhoA activity mediated stationary podocyte phenotype; either hyperactivation or reduction of RhoA induced proteinuria.^33,34 In contrast, podocyte-specific deletion of Rac1 failed to affect renal function, whereas hyperactivation of Rac1 induced FP effacement and proteinuria.^35,36 Interestingly, our results indicate that loss of Kindlin-2 in podocytes, which under normal physiologic conditions attach stably to the basement membrane, significantly increases Rac1 activation and podocyte motility. Notably, not all Kindlin-2–deficient cells exhibit a hyper-motile phenotype, because depletion of Kindlin-2 in some other cells types such as HT1080 fibrosarcoma cells (Supplemental Figure 4) reduced rather than increased cell motility. Importantly, inhibition of Rac1 effectively suppressed podocyte motility, reduced the albuminuria, and ameliorated the renal damage induced by the loss of Kindlin-2 in vivo, suggesting that Rac1 likely functions as a key downstream effector of Kindlin-2 in regulation of podocyte behavior, albeit we cannot completely rule out the possibility that systemic effects of the Rac1-specific inhibitor may also contribute to the benefits in vivo.

The finding that inhibition of Rac1 was able to significantly reverse albuminuria and renal damage induced by the loss of Kindlin-2 was somewhat surprising, because it has been widely believed that the primary function of Kindlin-2 is to regulate integrin-mediated cell-matrix adhesion. Indeed, loss of Kindlin-2 results in podocyte detachment in vivo and in vitro. Although the podocyte detachment probably also contributes to the renal phenotype that we found in Kindlin-2^Neph2 cKO mice, the fact that inhibition of Rac1 was able to partially reverse albuminuria and renal damage induced by the loss of Kindlin-2 in vivo suggests that Kindlin-2–mediated suppression of Rac1 and podocyte motility represents an important function of Kindlin-2 in podocytes.

How does Kindlin-2 regulate Rac1 activation? To address this question, we performed IP-MS and followed up co-IP and GST fusion protein pulldown experiments and identified RhoGDIα as a novel Kindlin-2–associated protein. RhoGDIα is the most abundant isoform of RhoGDIs and is ubiquitously expressed. It sequesters RhoGTPases in their inactive forms, and regulates both the GDP/GTP exchange cycle and the membrane association/dissociation cycle.^30,37 The RhoGDIα-GTPase complex is a major converging point for regulation of RhoGTPase activity and functions. Thus, proteins that regulate the RhoGDIα-GTPase complex formation are key regulators of RhoGTPase and actin cytoskeleton dynamics. Systemic knockout of RhoGDIα in mice resulted in severe proteinuria and renal failure.³⁸ TEM results showed that podocytes were severely injured and the FP were disrupted in RhoGDIα KO mice.³⁸ Shibata et al.³⁶ showed that Rac1 but not RhoA or Cdc42 was activated in the kidney of RhoGDIα KO mice and proteinuria was ameliorated by inhibition of Rac1, suggesting that Rac1 hyperactivation is responsible for proteinuria in RhoGDIα KO mice. Interestingly, clinical studies in human patients have shown that ARHGDIA, the gene encoding RhoGDIα, is involved in the pathogenesis of autosomal recessive congenital nephrotic syndrome.^39,40 Furthermore, three mutations in ARHGDIA have recently been reported to correlate with heritable nephrotic syndrome and all of these three mutations cause hyperactivation of Rac1, but not that of RhoA and Cdc42.⁴¹ Our studies show that loss of Kindlin-2 reduces the level of RhoGDIα, facilitates the release of Rac1 from RhoGDIα, increases Rac1 activation, and consequently promotes actin dynamics and podocyte motility. In addition, we have found that loss of Kindlin-2 reduced the expression of SD protein expression, which may also be explained by increased activation of Rac1, because hyperactivation of Rac1 is known to cause reduced expression of SD protein expression, such as nephrin and podocin.⁴² Recent studies suggest that Rac1 can also be activated in response to podocyte injury.^43–45 Thus, podocyte injury or proteinuria may also contribute to Rac1 activation seen in the Kindlin-2^Neph2 cKO mice, resulting in a positive feedback cycle and severe damage to podocytes and glomerular function. Taken together, the genetic, cellular, and biochemical studies described in this paper have revealed a novel signaling axis comprising of Kindlin-2, RhoGDIα, and Rac1 that is critical for control of podocyte motility and glomerular structure and function (Figure 10). Our findings suggest that therapeutic approaches targeting this signaling axis may benefit human patients with podocytopathy associated with aberrant Rac1 activation.

Figure 10.

A working model of Kindlin-2–RhoGDIα–Rac1 signaling in podocytes. In WT mouse podocytes, Kindlin-2 interacts with RhoGDIα, and sequesters Rac1 in its inactive form. In Kindlin-2^Neph2 cKO mice, loss of Kindlin-2 stimulates release of Rac1 from RhoGDIα for Rac1-GEF-effector interaction which leads to actin stress fiber disruption and SD protein reorganization, thereby resulting in podocyte foot effacement, enhanced motility, podocyte dysfunction, and ultimately proteinuria and renal failure. cKO, conditional knockout; WT, wild-type.

Concise Methods

Mice and Genotyping

Generation of Kindlin-2^fl/fl mice, in which LoxP sites were inserted from exons 5 and 6 at the Kindlin-2 locus through homologous recombination, was described in our previous study.²³ For selective deletion of Kindlin-2 in glomerular podocytes, Kindlin-2 ^fl/fl mice were crossed with Neph2-cre mice (from the Jackson Laboratory) to generate podocyte-specific Kindlin-2 knockout mice (Kindlin-2^Neph2 cKO). Tail genotyping was performed by routine PCR protocol. The following PCR primers were used for analyzing (1) Cre transgene: 5′-gat ctc cgg tat tga aac tcc agc-3′ and 5′-gct aaa cat gct tca tcg tcgg-3′, which generated a 650 bp fragment; and (2) Kindlin-2 genotyping: 5′-tgt gtt tca aag gta ctg gtc a-3′ and 5′-aca atg gtg ctt tgc cta ca-3′, which yielded 300 and 200 bp bands for the floxed and wild-type alleles. All animals were born normally at the expected Mendelian frequency. All control mice displayed normal phenotype. For treatment of mice with Rac 1 inhibitor NSC23766 (10 mg/kg per day, dissolved in PBS) (Calbiochem, San Diego, CA), we subcutaneously infused NSC23766 or PBS (as specified in each experiment) to 107 2-week-old mice, in which 35 were WT and 72 (35 subgroups) were Kindlin-2^Neph2 cKO mice (all Kindlin-2^Neph2 cKO mice in the same subgroup were littermates), and 95 6-week-old mice, in which 30 were WT and 65 (30 subgroups) were Kindlin-2^Neph2 cKO mice (all Kindlin-2^Neph2 cKO mice in the same subgroups were littermates), for 2 weeks following a protocol that was described previously.³⁶ At the end of NSC23766 treatment, the mice were analyzed by various assays as specified in the Results section. Animal experiments were approved by the Institutional Animal Care and Use Committee at the Southern University of Science and Technology of China.

Glomerular Isolation

Isolation of primary podocytes from WT or Kindlin-2^Neph2 cKO mice was performed as described previously.⁴⁶ Briefly, mice were anesthetized by intraperitoneal injection of pentobarbital (50 mg/kg) and perfused with 8×10⁷ Dynabeads M-450 (Dynal Biotech ASA, Oslo, Norway) diluted in 5 ml of PBS by the abdomen artery. After perfusion, kidneys were removed and cut into 1-mm³ pieces and digested in collagenase A (1 mg/ml) at 37°C for 30 minutes with gentle shaking. After digestion, the tissue was pressed gently through a 100-μm cell strainer (Falcon; BD Biosciences) and plated on collagen type I–coated dishes at 37°C in RPMI 1640 medium supplemented with 5% FBS and antibiotics (100 U/ml penicillin, 100 μg/ml streptomycin). After 3 days in culture, nonadherent glomeruli were removed by aspiration, and fresh media were reintroduced. Under such conditions, podocytes could be seen to begin budding out from glomerular cores within 24 hours. After a sufficient number of individual cells had expanded on coverslips or dishes, cells were processed for further study.

Histologic Examination and Immunohistochemical Staining

Paraffin-embedded mouse kidney sections (4 µm thickness) were prepared using a routine procedure. Tissues sections were stained with H&E, periodic acid–Schiff (PAS), and Masson trichrome using a standard protocol. Twenty glomeruli were randomly selected from each mouse and the glomerular damage was quantified by grading the severity of the sclerosis in the glomeruli on PAS-stained sections, on a scale of 0–4, as described previously.^47,48 Tubulointerstitial lesions score was evaluated in 20 randomly selected cortical areas per sample observed at ×200 magnification, using the method of Yamamoto et al.⁴⁹ Immunohistochemical staining (IHC) were stained with different antibodies against WT1 (Santa Cruz Biotechnology), active-caspase3 (Cell Signaling Technology), and fibronectin (BD Biosciences). After incubation with primary antibodies at 4°C overnight, the slides were then stained with horseradish peroxidase–conjugated secondary antibody (DAKO, Denmark). Nonimmune normal IgG was used to substitute primary antibodies as a negative control. Slides were viewed under a Leica TCS SP8 confocal microscope. Glomerular IHC score and tubulointerstitial IHC score for fibronectin staining were semiquantitated by scoring 20 randomly selected glomeruli under ×400 magnification and 20 interstitial fields under ×200 magnification per section on a 0–4 scale, as described previously.^47,49,50

Immunofluorescence Staining

Kidney cryosections were fixed with 4% paraformaldehyde for 30 minutes at room temperature. After blocking with 10% donkey serum for 1 hour, the slides were immunostained with primary antibodies against nephrin (Fitzgerald Industries International), WT1, synaptopodin (Santa Cruz Biotechnology), ZO-1 (Invitrogen), α-actinin4 (Abcam), desmin (Santa Cruz Biotechnology), α-SMA (Sigma-Aldrich), and Kindlin-2 (Millipore), respectively. The slides were then stained with Alex488 or Alex594-conjugated secondary antibody (Invitrogen Molecular Probes). Slides were viewed under a Leica TCS SP8 confocal microscope.

Isolated primary podocytes were cultured and fixed in 4% paraformaldehyde in 1× PBS for 20 minutes, permeabilized with 0.1% Triton X-100 in 1× PBS for 10 minutes, blocked with 3% BSA for 2 hours, and then incubated with appropriate primary and secondary antibodies as described above. Images were acquired and processed as described above.

Urinary Albumin and Creatinine Assay

For determination of albumin in urine, mice were placed in metabolic cages and urine was collected for 24 hours. Urine albumin was measured in duplicate using a mouse albumin ELISA quantitation kit, according to the manufacturer’s protocol (Bethyl Laboratories).⁵¹ Plasma and urine creatinine were measured in duplicate using commercial ELISA quantitation kits following the manufacturer’s protocol (Bioassay Systems).^46,52

Electron Microscopy

Electron microscopy of kidney samples was carried out as described previously.^46,53 Briefly, mouse kidneys were perfusion-fixed with 2.5% glutaraldehyde in PBS by left cardiac ventricular injection and postfixed in aqueous 1% OsO4. Specimens were dehydrated through an ethanol series, infiltrated in a 1:1 mixture of propylene oxide-Polybed 812 epoxy resin (Polysciences, Warrington, PA), and then embedded. Ultrathin sections were stained with 2% uranyl acetate, followed by 1% lead citrate. Sections were observed and photographed using a JEOL JEM 1210 transmission electron microscope (JEOL, Peabody, MA). For scanning electron microscopy, kidneys were perfusion-fixed in 2.5% glutaraldehyde in PBS as for TEM, dissected lengthwise, then postfixed in 1% OsO4. After dehydration through an ethanol series, kidney samples were critical point–dried, mounted on aluminum stubs, sputtered with gold-coat, and examined under JEOL 6335F field emission gun scanning electron microscope.

Quantitative analysis of SD density was performed as described previously.⁵² In brief, the number of podocyte FP present in each micrograph was divided by the total length of GBM in each image to determine the average density of podocyte FP.

Apoptosis Assays

Apoptotic cell death was determined using TUNEL staining with DeadEnd Fluorometric Apoptosis Detection System (Promega, Madison, WI), as previously reported.⁵⁴ Additionally, paraffin-embedded kidney sections were stained with anti-caspase3 antibodies (Cell Signaling Technology).

Cell Culture and shRNA Transfection

Conditionally immortalized human glomerular podocytes were propagated under permissive condition at 33°C as previously described.⁵⁵ To induce differentiation, cells were switched to 37°C (nonpermissive condition) for 10–14 days. To generate Kindlin-2–silencing stable cell lines, we transfected pGFP-V-RS–Kindlin-2 shRNAs or control vectors into immortalized human podocyte cells using Lipofectamine 3000 (Invitrogen). GFP-V-RS were obtained from Origene. The targeted sequences of Kindlin-2 shRNA1 and shRNA2 are 5′-aac agc gag aat ctt gga ggc cca tca ga-3′ and 5′-gct taa gct ggt gga gaa act cga tgt aa-3′, respectively. An irrelevant RNA (5′-acg cat gca tgc ttg ctt t-3′), which does not correspond to any known human genes, was used as control. To evaluate the effect of NSC23766 in vitro, human podocyte cells grown in nonpermissive condition were treated with 50 µM NSC23766 (dissolved in PBS) for 24 hours, then cells were collected for further analyses as specified in each experiment.

Immunoblotting Analysis

Primary podocytes from mice or human immobilized podocyte cells were pooled and lysed in SDS sample buffer. Protein expression was analyzed by Immnoblotting analysis as described previously.⁵⁶ The primary antibodies used were as follows: anti-WT1 (Santa Cruz Biotechnology), anti-nephrin (Millipore), synaptopodin (Santa Cruz Biotechnology), ZO-1, α-actinin4, anti-GAPDH (Abmart), anti–Kindlin-2 (Proteintech and Millipore), anti-talin (Millipore), anti-RhoGDIα (Santa Cruz Biotechnology), and anti-Rac1 (BD Biosciences). Full scans of original immunoblot data presented in this study were shown in the Supplemental Information.

Rac and Rho Activation Assay

Human immortalized podocytes were lysed following the manufacturer’s protocol (product code v4xc-2012; Thermo Fisher Scientific) and incubated with GST-PAK-agarose (which binds only to Rac-GTP) or GST-Rhotekin-agarose (which binds only Rho-GTP) for 2 hours, followed by separation by SDS-PAGE and immunoblotting with the appropriate primary and secondary antibodies.

Rac1 activation in isolated primary podocytes was determined using a Rac G-LISA Activation Assay Biochem kit (Cytoskeleton, Denver, CO) according to the manufacturer’s instructions as described previously.⁵⁷

Co-IP

Cell lysates were prepared using RIPA lysis buffer (Beyotime, Shanghai, China) supplemented with protease and phosphatase inhibitors (Roche). Samples were incubated for 30 minutes on ice and centrifuged (12,000 × g for 15 minutes at 4°C). For immunoprecipitation experiments, equal amounts of total lysates (2 mg) were incubated overnight at 4°C with 30 µl of protein A–Sepharose beads (precleaned lysates). In parallel, 50 µl of protein A–Sepharose beads was incubated for 2 hours at 4°C with antibodies of interest (2 µg) to generate the immunobeads, which were subsequently mixed with precleared lysates and incubated overnight at 4°C. The next day, beads were rinsed three times in PBS with 1 mM PMSF. Proteins were eluted from Sepharose beads by mixing with 60 µl of 1× loading buffer (containing 10% β-mercaptoethanol). Subsequently, samples were processed for immunoblotting.

GST Fusion Protein Pull-Down Assay

For generation of GST fusion proteins containing full length or various domains of RhoGDIα, cDNAs encoding RhoGDIα or its fragments were cloned into pGEX-4T-1 vector. Escherichia coli strain BL21 were then transformed with the expression vectors. GST and GST fusion proteins were purified from E. coli BL21 using Glutathione-Sepharose 4B matrix (GE Healthcare) according to the manufacturer’s instructions. Purified proteins were resolved by SDS-PAGE to verify their size and purity. In pull-down assays, GST or GST fusion proteins were bound to Glutathione-Sepharose, mixed with podocyte cell lysates, and incubated overnight at 4°C. Subsequently, the beads were washed three times with 1 ml of PBS containing 0.2% Triton X-100. GST and GST fusion proteins bound to the beads were eluted and analyzed by immunoblotting.

Cell Migration Assays

Transwell cell migration assays were performed in triplicate using cell culture inserts with 8.0-μm pore size membranes (Becton Dickinson, Franklin Lakes, NJ) following the manufacturer’s protocol. Podocyte cells (1×10⁵) were plated in the top chamber in DMEM. In the bottom chamber, culture medium containing 10% FBS was used as a chemoattractant. After 24 hours’ incubation at 37°C, media were removed and membranes were washed with 1× PBS three times; cells in the membranes were fixed in 4% formaldehyde and then stained with 5000× Hochest. Cells were counted under a Nikon T1-SAM microscope. For single-cell migration assay, podocyte cells (5×10³) were seeded in a culture dish coated with fibronectin (10 mg/ml) and placed in a heated and air-humidified chamber built in a Nikon TE2000E inverted microscope. Phase-contrast time-lapse imaging of a field containing 7–10 cells at a 4-minute interval for 6 hours was captured on the microscope with a 10× Ph1 objective, perfect focus system, and a Hamamatsu C-11440–22CU camera controlled by NIS-elements software (Nikon). Cell motility was tracked by using Image J (National Institutes of Health) with plug-in “Manual Tracking” (Fabrice Cordelières, Institut Curie, Orsay, France). Data were analyzed by the Chemotaxis-and-Migration-Tool (ibidi, http://ibidi.com/xtproducts/en/Software-and-Image-Analysis/Manual-Image-Analysis/Chemotaxis-and-Migration-Tool). In each experiment, >50 randomly selected cells were manually tracked to measure migration velocity.

Nano–LC-MS/MS Analysis

LC-MS/MS analysis was carried out as described previously.^58,59 Briefly, human podocyte cell lysates were immunoprecipitated with anti–Kindlin-2 or mouse control IgG. Each immunoprecipitated sample was then resolved in SDS loading buffer and denatured at 95°C for 5 minutes. After protein separation by 10% SDS-PAGE and staining with Brilliant Blue G-250, each lane of the gel was excised, cut into six slices, and incubated with 500 µl 50 mM ammonium bicarbonate (ABC) and 50% (vol/vol) acetonitrile (ACN) for destaining. The gel slices were dehydrated and rehydrated with 200 µl 100% ACN. Disulfide bonds were reduced with 200 µl 10 mM DTT. The proteins were alkylated with 100 mM iodoacetamide and 50 mM ABC for 30 minutes at room temperature in the dark. The gel pieces were dried in a speed vacuum and incubated overnight at 37°C with sequencing-grade modified trypsin (Promega, Fitchburg, WI) at an enzyme-to-protein ratio of 1:100 (wt/wt). The resulting peptides were extracted sequentially from the gel slices with 200 µl of 25 mM ABC and 200 µl 5% (vol/vol) formic acid (FA), 50% ACN by sonication for 20 minutes at each stage. After all supernatants were combined, the peptides were dried in a speed vacuum. The dried peptide mixtures were dissolved in 100 µl 1% FA and desalted using homemade C18-StageTips as described.⁵⁸ Finally, the eluted peptide samples were lyophilized to dryness and redissolved in 10 μl of 0.1% (vol/vol) FA in water for nano–LC-MS/MS analysis. Peptide mixtures were analyzed by an Orbitrap Fusion mass spectrometer (Thermo Fisher Scientific) coupled with an Easy-nLC 1000 (Thermo Fisher Scientific) ultrahigh-pressure liquid chromatography pump. The LC separation system consisted of a trap column (100 μm i.d. ×4 cm) and an analytic column with integrated spray tip (75 μm i.d. ×20 cm), both packed with 3 μm/120 Å ReproSil-Pur C₁₈ resins (Dr. Maisch GmbH, Ammerbuch, Germany). The buffers used for separation were 0.1% FA in water and 0.1% FA in ACN. Half of the obtained samples were first loaded onto the trap column at a flow rate of 2 μl/min and then separated by the analytic column at a flow rate of 300 nl/min. The gradient was set as follows: from 3% to 7% ACN in 2 minutes, from 7% to 22% ACN in 50 minutes, from 22% to 35% ACN in 10 minutes, from 35% to 90% ACN in 2 minutes, holding at 90% ACN for 6 minutes, declining to 3% ACN in 2 minutes, and holding at 3% ACN for 13 minutes. Full MS scans were performed in an Orbitrap mass analyzer over m/z range of 350–1550 with a mass resolution of 120,000.

Data Processing

MS data analysis was performed as described previously.⁵⁸ The raw data were searched against the human UniProt FASTA database (68,485 entries, downloaded on May 14, 2015) using Sequest HT⁶⁰ node integrated within the Proteome Discoverer software (Version 1.4; Thermo Fisher Scientific). The precursor and fragment mass tolerances were set to 10 ppm and 0.6 D, respectively. A maximum of two missed cleavages was allowed for trypsin digestion. Cysteine carbamidomethylation was set as fixed modification, whereas methionine oxidation, asparagine, and glutamine deamidation were set as variable modifications. False discovery rate of peptide spectrum matches and identified peptides were determined by searching the forward and reverse database with Sequest HT node, and were validated by the Percolator algorithm⁶¹ at 1% on the basis of q values. False discovery rate on the basis of posterior error probability was determined by searching a reverse database and was set to 0.01 for proteins and peptides.

Statistical Analyses

All data are presented as mean±SEM. The two-tailed t test (for parametric data) or Mann–Whitney test (for nonparametric data) were used to compare two groups of samples according to distribution. One-way ANOVA (for parametric data) and Kruskal–Wallis test (for nonparametric data) were used to account for multiple comparisons. Survival analysis was carried out using the log-rank test. P values <0.05 were considered significant. R-project software was used for evaluating normal distribution and Prism 5 (GraphPad) was used for statistical analysis.

Disclosures

None.