Y. Harita, Department of Pediatrics, Graduate School of Medicine, University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-8655, Japan Fax: +81 3 3816 4108 Tel: +81 3 5800 8659 E-mail: firstname.lastname@example.org
The slit diaphragm (SD) is an intercellular junction between renal glomerular epithelial cells (podocytes) that is essential for permselectivity in glomerular ultrafiltration. The SD components, nephrin and Neph1, assemble a signaling complex in a tyrosine phosphorylation dependent manner, and regulate the unique actin cytoskeleton of podocytes. Mutations in the NPHS1 gene that encodes nephrin cause congenital nephrotic syndrome (CNS), which is characterized by the loss of the SD and massive proteinuria. Recently, we have identified the expression of the transmembrane glycoprotein signal regulatory protein α (SIRPα) at the SD. In the present study, we analyzed the expression of SIRPα in developing kidneys, in kidneys from CNS patients and in proteinuric rat models. The possibility that SIRPα interacts with known SD proteins was also investigated. SIRPα was concentrated at the SD junction during the maturation of intercellular junctions. In the glomeruli of CNS patients carrying mutations in NPHS1, where SD formation is disrupted, the expression of SIRPα as well as Neph1 and nephrin was significantly decreased, indicating that SIRPα is closely associated with the nephrin complex. Indeed, SIRPα formed hetero-oligomers with nephrin in cultured cells and in glomeruli. Furthermore, the cytoplasmic domain of SIRPα was highly phosphorylated in normal glomeruli, and its phosphorylation was dramatically decreased upon podocyte injury in vivo. Thus, SIRPα interacts with nephrin at the SD, and its phosphorylation is dynamically regulated in proteinuric states. Our data provide new molecular insights into the phosphorylation events triggered by podocyte injury.
The glomerular capillary wall of the kidney plays a crucial role as a filtration barrier that selectively prevents the leakage of plasma proteins into urine. Defects in this barrier result in nephrotic syndrome, which is characterized by massive loss of protein in the urine (proteinuria). The glomerular capillary wall consists of three structural layers: the fenestrated endothelial cells, the glomerular basement membrane (GBM) and the visceral epithelial cells, also called podocytes. Podocytes possess unique interdigitating cell extensions, or foot processes, that are bridged by the slit diaphragm (SD). The SD is a podocyte-specific intercellular junction with an electron-dense, zipper-like structure that functions as a major glomerular filtration barrier. The first SD molecule to be identified was the transmembrane protein nephrin, which is encoded by the NPHS1 gene [1–3]. Mutations of NPHS1 result in congenital nephrotic syndrome (CNS), which is characterized by loss of the SD and massive proteinuria. Inactivation of the gene in mice leads to the absence of the SD, massive proteinuria and neonatal death . Other membrane-associated proteins such as Neph1, podocin and CD2-associated protein (CD2AP) are crucial filtration components localized at the SD that form protein complexes with nephrin [4,5]. In the last decade, several other podocyte components have been reported to be associated with proteinuria , including glomerular epithelial protein (GLEPP) 1 , Wilms’ tumor 1 (WT1)  and podocalyxin .
The SD not only serves as a junctional structure but also functions as a signaling nexus . Several SD components are substrates for tyrosine kinases, and their phosphorylations are now known to play key roles in the regulation of podocyte function. The cytoplasmic domain of nephrin is transiently tyrosine phosphorylated by the Src-family tyrosine kinase (SFK) Fyn in developing or injured podocytes [11,12]. The Src homology 2 (SH2) domain of Nck binds to several phosphorylated tyrosines in nephrin, and these interactions regulate podocyte morphogenesis or response to injury by inducing actin polymerization [12,13]. Phosphorylation of a specific tyrosine residue of nephrin also recruits phospholipase C (PLC) γ, which triggers a Ca2+ response . Neph1 is also phosphorylated by Fyn and then recruits Grb2 [15,16]. This event is necessary for Neph1-induced actin polymerization at the plasma membrane . The critical role of tyrosine phosphorylation in the filtration barrier is also suggested by proteinuria and the effacement of foot processes in fyn-deficient mice [11,17].
Recently, we have reported that signal regulatory protein α (SIRPα) (also known as CD172a or Src homology 2 domain-containing protein tyrosine phosphatase substrate-1, SHPS-1) is expressed at the SD in podocytes . SIRPα is a transmembrane receptor that is primarily expressed in myeloid cells or neurons [19–21]. The extracellular domain of SIRPα consists of three immunoglobulin superfamily motifs. The widely expressed cell-surface protein CD47 (also known as integrin-associated protein) is a ligand of SIRPα and binds to its extracellular domain. This interaction provides an inhibitory effect on intracellular signaling in phagocytes or neurons. Soluble ligands such as surfactant proteins A and D (SP-A and SP-D) also bind to SIRPα, following which they block proinflammatory mediator production in the lung . The cytoplasmic region of SIRPα contains four tyrosine residues that loosely conform to the immunoreceptor tyrosine-based inhibition motif (ITIM) , which is characterized by a consensus amino acid sequence of I/V/L/SxYxxL/V . ITIMs are present in a large group of molecules that negatively regulate signaling pathways elicited by certain receptors , and they are often phosphorylated by SFK. SIRPα is tyrosine phosphorylated at these ITIMs, and these phosphorylated motifs recruit the cytoplasmic tyrosine phosphatases SH2-domain-containing protein tyrosine phosphatase 1 (SHP-1) and SHP-2 . Recruitment of SHP-1/2 in the vicinity of the cell membrane by phosphorylation of SIRPα is important for interrupting signaling from receptor tyrosine kinases by dephosphorylation of the receptor or other substrates . Because CD47 is exclusively localized in mesangial cells in glomeruli , and because mesangial cells and podocytes are separated by GBM, SIRPα may play unique role(s) at the SD. To date, alteration of SIRPα expression in proteinuric states has not been reported.
In the present study, we analyzed (a) SIRPα expression in developing kidney and proteinuric diseases including CNS carrying NPHS1 mutations, (b) the molecular link between SIRPα and SD components and (c) the expression and phosphorylation of SIRPα in normal kidney and two proteinuric models. Our findings suggest that SIRPα is associated with nephrin at the SD.
SIRPα expression in developing and diseased glomeruli
First, we evaluated SIRPα protein expression within developing glomeruli. Dual-labeling immunofluorescence analysis of SIRPα and zonula occludens 1 (ZO-1) was performed using frozen sections of newborn rat kidney. ZO-1, which is predominantly localized in the podocyte foot processes and concentrated at the insertion of the SD, is first expressed at the S-shaped body stage in developing glomeruli . SIRPα was first seen during the S-shaped body stage and detected broadly along the cell surface with strong signals at the basal margin of presumptive podocytes (Fig. 1A). In the capillary loop stage and mature glomeruli, SIRPα and ZO-1 were in close proximity at the basal margin of the presumptive podocytes (Fig. 1B). These results suggest that SIRPα becomes concentrated at the SD junction during the maturation of the intercellular junctions.
Next, we analyzed the expression of SIRPα and other SD components (nephrin, Neph1, podocin and ZO-1), as well as critical components for the barrier function of podocytes (GLEPP-1 and podocalyxin), and WT1, a component associated with podocyte differentiation (Fig. 2), in kidneys from CNS patients caused by NPHS1 mutations (CNS1, heterozygous missense mutation (P676R); CNS2, homozygous mutation (Q839RfsX849)). We also compared their expression in kidneys from a control case and in patients with other glomerular diseases such as IgA nephropathy and Alport syndrome (Fig. 2). In normal kidney, intense signals for SIRPα and other SD components were observed along with glomerular capillary loops. Clear signals demarcating glomerular capillaries for each SD component were also observed in glomeruli from IgA nephropathy and Alport syndrome patients. In contrast, signals for nephrin completely disappeared in the CNS patients. Signals for SIRPα and Neph1 were negligible along the glomerular capillaries in the CNS patients, and the signals for podocin were only slightly decreased and shifted to a discontinuous coarse granular pattern in CNS glomeruli. The staining patterns of ZO-1, GLEPP-1, WT1 and podocalyxin did not change significantly in glomeruli from CNS, IgA nephropathy and Alport syndrome patients. These results show that failure of SD structure formation by nephrin disrupts the expression of SIRPα and Neph1, suggesting that SIRPα and Neph1 are closely associated with nephrin expression.
SIRPα interacts with nephrin
We next sought to determine whether SIRPα interacts with other SD components at the podocyte intercellular junction. A series of co-immunoprecipitation experiments was performed using human embryonic kidney (HEK) 293T cells transiently co-expressing SIRPα and Neph1, podocin, transient receptor potential cation channel 6 (TRPC-6) or nephrin. While Neph1, podocin or TRPC-6 were not co-immunoprecipitated with SIRPα, nephrin was co-immunoprecipitated with SIRPα (Fig. 3A). Because the SIRPα 4YF mutant, in which four cytoplasmic tyrosine residues are substituted for phenylalanines, was also co-immunoprecipitated with nephrin, SIRPα interacts with nephrin independent of its cytoplasmic phosphorylation status (Fig. 3B). This interaction was not altered by co-transfection with Fyn, or by mutation of nephrin Y1204F which disrupts its interaction with Nck or PLC-γ  (Fig. 3C). This result suggests that nephrin phosphorylation also does not directly affect nephrin–SIRPα interaction. Glomerular endogenous nephrin was co-immunoprecipitated with endogenous SIRPα, suggesting that SIRPα forms a complex with nephrin in vivo (Fig. 3D).
To further confirm the physical interaction between SIRPα and nephrin, we examined their distributions in adult and developing kidney. Both SIRPα and nephrin immunoreactivity were observed along the glomerular capillary loop, and their expression patterns overlapped in mature human glomeruli (Fig. 4A,B). In developing rat glomeruli, SIRPα immunoreactivity was detected along the cell surface and basal margins of presumptive podocytes, where it colocalized with nephrin (Fig. 4C). These results support the physical interaction between SIRPα and nephrin at the SD in vivo.
Expression and phosphorylation status of SIRPα in proteinuric states
Because SIRPα is already known as a substrate of SFK , we examined whether SIRPα is tyrosine phosphorylated in glomeruli. The cytoplasmic domain of SIRPα contains four YXX(L/V/I) motifs (Y436, Y460, Y477 and Y501) that can recruit SHP-1/2 when tyrosine phosphorylated. We prepared a rabbit polyclonal antibody against a phosphopeptide surrounding Y501 as described in Materials and methods (anti-pY501). Amino acid sequences surrounding Y501 are conserved among rat, mouse and humans. To confirm the epitope specificity of this antibody, plasmids encoding wild-type SIRPα, SIRPα (Y501F) or SIRPα (4YF) (four tyrosine residues, Y436, Y460, Y477 and Y501, are replaced with phenylalanine) mutants were expressed in HEK293T cells together with a control vector or an expression vector for Fyn. The anti-phospho-SIRPα (pY501) IgG detected phosphorylation of SIRPα (Fig. 5A). The weak reactivity of anti-pY501 to Y501F may be due to the sequence similarity among the four ITIMs and their surrounding sequences. No signal was observed in lysates from cells expressing the SIRPα 4YF mutant and Fyn, indicating that anti-pY501 IgG specifically detects the phosphorylated form of SIRPα. Then, by using this antibody, we assessed whether SIRPα is tyrosine phosphorylated in glomeruli. Glomeruli isolated from normal rats were solubilized in 1% NP-40 lysis buffer and separated into soluble and insoluble fractions, followed by immunoblotting with anti-SIRPα IgG and anti-pY501 IgG. Phosphorylation of SIRPα was clearly detected in adult rat glomeruli (Fig. 5B).
Next, we investigated whether treatments that cause podocyte injury changed the expression and phosphorylation status of SIRPαin vivo. Puromycin aminonucleoside (PAN) nephropathy and the protamine sulfate (PS) perfusion model in rats are characterized by foot process effacement and proteinuria, which are typical features of nephrotic syndrome [12,14,16]. Immunohistochemical analysis showed that nephrin signals demarcating the capillary lumen decreased considerably in these models . In contrast, signals for SIRPα were conserved in both proteinuric models (Fig. 6A,B). Immunoblotting also confirmed that the amount of SIRPα protein in both models was not significantly altered. To analyze the phosphorylation status of SIRPα, anti-SIRPα immunoprecipitates from glomerular lysates of normal and PAN-treated rats were immunoblotted with anti-phosphotyrosine (Fig. 5C) or anti-phospho-SIRPα antibody (Fig. 5D). Tyrosine phosphorylation of SIRPα, which was clearly observed in the glomeruli of control rats, was decreased in day 2 in the PAN nephropathy model (Fig. 5C,D) and decreased by 78% 5 days after PAN injection (Fig. 5D). Phosphorylation of SIRPα was also downregulated by 88% after podocyte injury by PS treatment in vivo (Fig. 5E).
Tyrosine phosphorylation of SIRPα is associated with phosphorylation of nephrin
Next, we tested whether nephrin phosphorylation was affected by phosphorylation of SIRPα, or vice versa. In HEK293T cells expressing SIRPα or nephrin, weak but distinct tyrosine phosphorylation of both proteins was detected (Fig. 7A,B). SU6656, a specific SFK inhibitor, significantly reduced tyrosine phosphorylation of SIRPα and nephrin, suggesting that SFK is responsible for the phosphorylation of both proteins (Fig. 7A,B). Cotransfection of nephrin did not alter the phosphorylation status of SIRPα (data not shown). Notably, the co-expression of wild-type SIRPα with nephrin reduced nephrin phosphorylation (*P < 0.01 versus control), whereas this effect was abrogated by mutation of the four cytoplasmic tyrosine residues (Fig. 7C,D), suggesting that SIRPα is associated with nephrin phosphorylation in cultured cells in vitro.
SD proteins such as nephrin and Neph1 influence podocyte process patterning and cellular junction formation by assembling a protein complex that regulates actin cytoskeletal dynamics . In the process, tyrosine phosphorylation plays key roles in the formation and activation of adaptor proteins. In this study, we have shown that SIRPα interacts with the SD protein nephrin, and its phosphorylation is dynamically regulated following podocyte injury. Although CD47, a known ligand of SIRPα, has a single immunoglobulin V-set domain that binds to the N-terminal V-set Ig domain of SIRPα, nephrin has eight C2-set Ig domains. Therefore, SIRPα at the SD organizes a complex with a different structure to that previously reported in other cells.
Defects in SD complex components result in the failure of the glomerular filtration barrier and lead to proteinuria in humans and mice. During the formation of polarized epithelial structures, the junctional complex between immature podocytes migrates along the lateral membrane towards the basal membrane and finally forms a concentrated SD structure [28,29]. SIRPα is expressed in developing presumptive podocytes and is concentrated at the SD area in mature glomeruli, where it interacts with nephrin. In CNS kidneys with NPHS1 mutations, nephrin is not recruited to the cell membrane because of defective intracellular nephrin transport . Notably, the expression levels of SIRPα and Neph1 were significantly decreased in our CNS patients, indicating that nephrin forms the core of the SD complex, which consists of Neph1 and SIRPα.
SIRPα is highly tyrosine phosphorylated in adult podocytes, and its phosphorylation levels are decreased in proteinuric rat models. In contrast to podocytes, SIRPα is dephosphorylated in the resting state in neurons and immune cells, where its phosphorylation is induced by binding of its ligand or by various growth factors and events such as integrin-mediated cell adhesion to extracellular matrix proteins [31–34]. Currently, the regulatory mechanism(s) for continuous phosphorylation of SIRPα in mature podocytes remains unclear. Because the activation of nephrin by clustering  did not induce SIRPα phosphorylation (data not shown) and because co-transfection of nephrin did not alter the phosphorylation status of SIRPα (data not shown), nephrin does not seem to act upstream of SIRPα phosphorylation. Therefore, SIRPα may continuously sense intracellular signals or extracellular ligand(s) in resting podocytes via the following pathways: (a) stimulation by integrin-mediated cell adhesion to extracellular matrix proteins on GBM; (b) stimulation by interaction of SIRPα with other SD molecules except for nephrin; (c) stimulation by circulatory or growth factors produced by adjacent glomerular cells.
Although net phosphorylation levels in the SD should reflect a balance between phosphorylation and dephosphorylation, the roles of tyrosine phosphatases in SD signaling and in cell morphology have not been intensively investigated. Podocytes express several phosphatases (SHP-2, PTP-PEST PTP1B and PTP-36), and sodium vanadate, an inhibitor of protein tyrosine phosphatases, induces the reorganization of the podocyte actin cytoskeleton and focal contacts . Although SIRPα expression decreased nephrin phosphorylation in transfected 293T cells, this effect may not be attributed to SHP-1/2 because the involvement of SHP-2 in nephrin phosphorylation is not apparent in podocytes . Rather, protein tyrosine phosphatase 1B (PTP1B) binds to and dephosphorylates nephrin , suggesting that each SD component may be subject to specific phosphorylation and dephosphorylation and can be differentially regulated in developmental processes and in podocyte injury. Further studies including gene knockout are required to identify the downstream targets of SIRPα in podocytes in vivo.
Given the function of the SD as an essential framework for signaling and as a physical filtration barrier, further knowledge of the expression and phosphorylation of SIRPα should provide insight into the structural integrity of the SD and the molecular pathogenesis of proteinuria.
Materials and methods
Antibodies and reagents
The following antibodies were obtained commercially: mouse monoclonal anti-Flag IgG (M2; Sigma, St Louis, MO, USA), rabbit polyclonal anti-SIRPα IgG (Upstate, Lake Placid, NY, USA), mouse monoclonal anti-SIRPα IgG (Santa Cruz Biotechnology, Santa Cruz, CA, USA), mouse monoclonal anti-ZO-1 IgG (33-9100; Invitrogen, Carlsbad, CA, USA), mouse monoclonal anti-GLEPP-1 IgG (MU336-UC; BioGenex, San Ramon, CA, USA), mouse monoclonal anti-WT1 IgG (M3561; Dako, Carpinteria, CA, USA), mouse monoclonal anti-podocalyxin IgG (MAB430; Millipore, Bedford, MA, USA), rat monoclonal anti-HA IgG (3F10; Roche Diagnostics, Mannheim, Germany), mouse monoclonal anti-c-Myc IgG (sc-40; Santa Cruz) and mouse monoclonal anti-phosphotyrosine IgG (4G10; Upstate). Rabbit polyclonal anti-Neph1 IgG , rabbit polyclonal anti-nephrin IgG  and rabbit polyclonal phospho-specific anti-nephrin IgG (anti-pY1204) were previously described . Rabbit polyclonal phospho-specific anti-SIRPα IgG (anti-pY501) was raised against an HPLC-purified synthetic oligopeptide CPSFSEpYASVQVQ (the first cysteine is not part of the SIRPα sequence) coupled to keyhole limpet hemocyanin. The antiserum was affinity purified by the immunogen using a SulfoLink column (Pierce, Rockford, IL, USA) and absorbed with non-phosphorylated peptide, CPSFSEYASVQVQ. Rabbit polyclonal anti-podocin IgG was raised against a synthetic oligopeptide KPVEPLNPKKKDSPML. Western blotting was performed with these antibodies diluted at 1 : 2000. SU6656 (Merck KGaA, Darmstadt, Germany) was purchased.
Cell culture and transfection
HEK293T cells were purchased from the ATCC (Manassas, VA, USA) and maintained in DMEM containing 10% fetal bovine serum. Transfections were performed using Lipofectamine 2000 reagent (Invitrogen) following the manufacturer’s instructions.
Eukaryotic expression constructs
Using 5′-cgcgaattcccgccatggagcccgccggcccggc-3′, 5′-ccggtcgaccttcctctggacttggacactggc-3′, 5′-cccctcgagtaagcttgagcccaaatcttgtgacaaaac-3′ and 5′-aagggtaccttatcatttacccggagacagggaga-3′ as primers, a cDNA fragment coding for full-length rat SIRPα and podocin was amplified by PCR and cloned into a pCMV-Tag4A vector (Stratagene, Cedar Creek, TX, USA). SIRPα phenylalanine substitution mutants Y501F and 4YF (Y436F, Y460F, Y477F and Y501F) were prepared using standard PCR methods. Mammalian expression plasmids encoding Fyn (a gift from T. Tezuka, Tokyo University, Japan) , mouse TRPC6-HA (a gift from C. Hisatsune, RIKEN Brain Science Institute, Japan) , rat Neph1 , rat nephrin  and nephrin phenylalanine substitution mutant Y1204F  were as previously described. Restriction digestion and DNA sequencing were performed to validate all constructs.
Cells were lysed with IP buffer (20 mm Tris/HCl, pH 7.5, 150 mm NaCl, 1% NP-40, 1 mm phenylmethylsulfonyl fluoride, 10 mg·mL−1 antipain, 10 mg·mL−1 leupeptin, 100 U·mL−1 aprotinin, 50 mm NaF, 1 mm EDTA and 1 mm orthovanadate) for 15 min on ice. Lysates were clarified by centrifugation and incubated with beads conjugated with M2 anti-Flag IgG for 1 h at 4 °C. Beads were washed twice with IP buffer, and bound proteins were eluted with 100 mm glycine/HCl (pH 2.6).
The renal tissue from normal rats was quickly frozen in n-hexane and cooled to −70 °C. Frozen sections 3-μm thick were cut with a cryostat and fixed with periodate-lysine-paraformaldehyde for 1 min. The cryosections were rinsed with NaCl/Pi and blocked in blocking solution (1% BSA in NaCl/Pi). The sections were incubated with the primary antibodies and visualized with Alexa Fluor conjugated secondary antibodies (Invitrogen). Paraffin-embedded samples from human renal biopsy samples were deparaffinized in xylene and rehydrated through graded alcohols in H2O, followed by heat-induced epitope retrieval by incubating in a target retrieval solution (S1699; Dako) for 15 min at 121 °C. Sections were cooled to room temperature and incubated with the primary antibodies followed by incubation with Alexa Fluor conjugated secondary antibodies (Invitrogen). Images were obtained with an inverted microscope (model IX71; Olympus, Tokyo, Japan) and were processed using Adobe Photoshop CS3.
Kidney biopsy samples were obtained from patients admitted to Kidney Center, Tokyo Women’s Medical University. CNS patient 1 (CNS1) presented with nephrotic syndrome with low serum albumin and high cholesterol at birth. Unilateral nephrectomy was performed when the patient was 15 months old. CNS patient 2 (CNS2) presented with severe nephrotic syndrome within 6 weeks resistant to angiotensin II inhibition and underwent unilateral nephrectomy at 11 months. Mutation analysis of NPHS1 was performed, and heterozygous missense mutation (P676R) (CNS1) and homozygous mutation (Q839RfsX849) (CNS2) were found. Normal tissue samples adjacent to teratoma (2 years old) and biopsy specimens from adult donor kidneys were used as control. Biopsy samples were also obtained from IgA nephropathy (5 years old) and Alport syndrome (3 years old). Samples were collected after receipt of informed consent from the patients and this study was approved by the ethics committee of the Tokyo Women’s Medical University.
All experiments with animals were performed according to the guidelines set by the Animal Center of the Institute of Medical Science, University of Tokyo. Perfusion of rat kidneys with PS was performed essentially as previously described . Six-week-old male Wistar rats were purchased from Charles River Laboratories Japan Inc. (Atsugi, Japan). The rats were anesthetized with pentobarbital. Kidneys were perfused through the aorta at 5 mL·min−1 with Hank’s balanced salt solution (HBSS) for 20 min followed by PS solution (500 mg·mL−1 in HBSS) for 20 min. The cryostat sections for immunohistological study and glomerular lysates were prepared as previously described . Induction of PAN nephrosis was performed as described previously . Young male rats (Wistar) were injected intraperitoneally with 10 mg·(100 g body weight)−1 of PAN (Sigma). PAN nephrosis rats were sacrificed on days 2, 5, 7 and 11 for protein samples from glomeruli.
This study was partly supported by the Japan Foundation for Pediatric Research. The study was also supported in part by a Grant-in-Aid for Young Scientists (B) (20790991) (to H.Y.) and a Grant-in-Aid for Scientific Research (22390204) (to H.Y., H.S. and I.T.) from the Ministry of Education, Culture, Sports, Science and Technology of Japan. We thank T. Tezuka for providing plasmids. We also thank H. Kosako and K. Shirakabe for valuable comments.