CD2-associated Protein (CD2AP) Expression in Podocytes Rescues Lethality of CD2AP Deficiency*

Mice born without CD2-associated protein (CD2AP) develop renal failure and nephrotic syndrome about 4 weeks after birth and die around 6 weeks of age. Although CD2AP is widely expressed, the severity of the renal failure precludes a clear determination of the role of CD2AP in other tissues. Here we generated transgenic mice expressing CD2AP using a podocyte-specific promoter. Podocyte-specific expression of CD2AP prevented the development of proteinuria, demonstrating that the renal failure is solely due to loss of CD2AP in podocytes and not in other renal or in immune cells. CD2AP-deficient mice are long-lived and appear phenotypically normal. Histological analysis demonstrated testicular abnormalities that were age-related. CIN85, a paralog of CD2AP, is poorly expressed in both the podocyte and the basal seminiferous tubule, suggesting that the loss of CD2AP in specific tissues may be compensated for by CIN85.

CD2-associated protein (CD2AP) 1 was originally cloned as a protein that interacts with the cytoplasmic domain of CD2, a T lymphocyte-and natural killer-specific membrane protein (1). Although originally cloned from a lymphocyte library, CD2AP is broadly expressed and found in all tissues except brain. Recently, a gene related to CD2AP was cloned. The protein encoded by this gene, CIN85, is also widely expressed and found in most tissues including brain (2)(3)(4)(5). Both molecules contain three amino-terminal SH3 domains followed by a proline-rich region and a coiled-coil domain at the carboxyl terminus. CIN85 was originally cloned as an interactor of c-Cbl (3) and the p85-␣ subunit of phosphatidylinositol-3 kinase (4).
The broad expression of CD2AP in most tissues suggested that CD2AP would have functions in cells outside of the immune system. The phenotype of CD2AP-deficient mice, which develop massive proteinuria around 3 weeks of age, develop nephrotic syndrome around 4 weeks of age, and ultimately die from renal failure at 6 -10 weeks of age, suggested an important role for CD2AP in the kidney (6). In the kidney, CD2AP is expressed mainly by the podocytes, proximal tubule, and collecting duct (7). Because the first detectable defect was podocyte foot process fusion, we postulated that podocyte dysfunction was the underlying cause of the renal failure, although the major histological feature of the lesion is significant mesangial expansion (6). In addition, whether CD2AP plays important roles in other organs was potentially masked by the severe renal failure in CD2AP-deficient animals.
Here we report that transgenic mice expressing CD2AP in podocytes are completely rescued from the renal phenotype of CD2AP deficiency. No proteinuria was detected, and the mice appeared to have a life span similar to wild-type animals. This demonstrates that the most important functions of CD2AP are in the kidney. Interestingly, transgenic/CD2AP-deficient male animals are infertile, and this is supported by histological studies. This suggests that CD2AP also plays an important role in the development of sperm in the testes.

Generation of Transgenic Mice and Genotyping-A cDNA encoding
Myc-tagged CD2AP was ligated into a vector containing the nephrin promoter (kind gift of S. Quaggin) (8). The transgenic construct was purified away from vector backbone and used for microinjection into mouse pro-nuclei in the Department of Pathology Transgenic Core Facility at Washington University Medical School. Founders were identified using a PCR-based strategy and were bred to animals heterozygous for the cd2ap-null allele to establish a colony of breeders positive for the transgene and heterozygous at the cd2ap locus. Transgene genotyping PCR was performed with primers (forward, 5Ј-GGCAC-CAAAGAATGTTCCTGG-3Ј, and reverse, 5Ј-CATCCTTGGGTTCT-GTCTCTCG-3Ј) that amplify the region of the transgene joining the mouse nephrin promoter to the CD2AP cDNA and yield a 750-bp product. PCR was performed using standard conditions. Animals selected for experiments were euthanized according to approved protocols.
Immunoblotting-Immortalized podocytes were obtained as a generous gift from Peter Mundel (9). Whole kidneys, organs, or isolated podocytes were rinsed in phosphate-buffered saline and homogenized in lysis buffer containing 10 mM Tris, pH 7.5; 150 mM NaCl; 1 mM EDTA; 1% Nonidet P-40; 5 mM aprotonin; and 5 mM leupeptin. Protein concentrations were estimated using the BCA method (Pierce), and equivalent amounts were fractionated by denaturing SDS-PAGE. Protein was then transferred to nitrocellulose membrane and probed with rabbit anti-CD2AP and rabbit anti-Erk2 (Santa Cruz Biotechnology) or rabbit anti-CD2AP or mouse anti-synaptopodin (generous gift of Peter Mundel) followed by horseradish peroxidase-conjugated donkey antirabbit secondary antibodies (Jackson ImmunoResearch).
Histology-After dissection, kidneys and testes were rinsed in phosphate-buffered saline and fixed overnight in 4% paraformaldehyde (Sigma) at 4°C. Tissues were then rinsed in phosphate-buffered saline and taken to the Histopathology Laboratory at Washington University Medical School for paraffin-embedding, sectioning, and staining with eosin and hematoxylin.
Immunohistochemistry-After dissection, kidneys and testes were washed with phosphate-buffered saline and immediately frozen in tissue molds containing OCT compound (Tissue Tek). Sections were collected onto fibronectin-coated slides and stained with rabbit anti-CD2AP and rat anti-entactin or goat anti-vimentin (C-20, Santa Cruz Biotechnology) antibodies or rabbit anti-CIN85 and rat anti-entactin antibodies followed by Cy3-conjugated donkey anti-rabbit (Jackson ImmunoResearch), fluorescein isothiocyanate-conjugated goat anti-rat (Jackson ImmunoResearch), or fluorescein isothiocyanate-conjugated donkey anti-goat (Jackson ImmunoResearch) secondary antibodies.
* The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Lymphocytes were collected on poly-L-lysine-coated slides and stained with fluorescein isothiocyanate-conjugated anti-mouse T cell receptor-␤ monoclonal antibody (BD Biosciences).

RESULTS
Generation of Podocyte-specific CD2AP Transgenic Animals-The CD2AP cDNA was appended with sequences encoding a Myc epitope and ligated into an expressing vector containing the mouse nephrin promoter that directs expression specifically in podocytes (8). Transgenic mice were generated using standard microinjection techniques, and two founder animals (line 21 and line 396) were identified. Both founders were bred to CD2AP heterozygous mice to generate transgenic, CD2AP-deficient mice (6). These mice were grossly normal and have life spans that are not noticeably different from wild-type animals with survival at least to 15 months.
Immunoblot analysis of whole kidney lysates with a rabbit anti-CD2AP serum confirmed transgene expression in both founder lines ( Fig. 1, top immunoblot, lanes 3 and 4). Immunoblots of various organs revealed no ectopic expression of the transgene except in muscle (Fig. 1, bottom immunoblot). Transgene expression levels in the kidney appeared to be considerably lower than the levels detected in a CD2AP heterozygous control mouse (Fig. 1, compare lanes 3 and 4 with lane 1), but CD2AP is expressed in tubules and collecting duct in addition to its expression in podocytes (7). Immunofluorescence staining was therefore used to assess the specificity of transgene expression as well as relative levels of CD2AP as compared with heterozygous controls. In control CD2AP heterozygous animals, CD2AP antibodies stained glomeruli in the typical, ribbon-like pattern consistent with podocyte expression (Fig. 2, A and B). CD2AP expression was also clearly detectable in tubular epithelia. CD2AP knock-out animals were used as negative controls (Fig. 2, C and D). Staining of transgenic KO mice from line 21 (Fig. 2, E and F) and line 396 (Fig. 2, G and H) demonstrated podocyte-specific expression of the CD2AP without detectable transgene expression in tubular epithelium. The staining intensity of CD2AP suggests that podocyte expression levels of the transgene were comparable with levels observed in heterozygous controls. (Fig. 2, compare E and G with A). Transgenic expression of CD2AP was not seen in T cells (Fig. 2, I-L). Immunoblots of lysates from multiple tissues, including brain, intestine, and liver, were negative for Myc-tagged CD2AP (data not shown).
CD2AP is enriched in the podocyte (Fig. 3, A-C). However, its paralog, CIN85, is more strongly expressed in the mesangial cell rather than the podocyte (Fig. 3, D-F). The intraglomerular staining of CIN85 did not change in the CD2AP KO (Fig. 3, G-I). In immunoblots from lysates of immortalized podocyte cell lines, however, there is weak expression of CIN85 (Fig. 3J).
Phenotypic Characterization of Transgenic, CD2AP-deficient Mice-Transgenic CD2AP-deficient animals were tested for proteinuria. Although CD2AP-deficient animals develop proteinuria around 3 weeks of age, transgenic/CD2AP-deficient animals did not exhibit proteinuria at any time, and these studies were followed up to 14 months. In addition, serum chemistries were found to be within normal ranges and comparable with wild-type mice as tested out to 10 months, including blood urea nitrogen (KO, 22 mg/dL, standard deviation 2.0; wild type, 19 mg/dL, standard deviation 3.3), creatinine (KO 0.6 mg/dL, standard deviation 0; wild type, 0.6 3 mg/dL, standard deviation 0.05), and glucose (103 mg/dL, standard deviation 12). Hematoxylin and eosin staining of formalin-fixed kidney sections showed no apparent differences between wild-type and transgenic CD2AP-deficient mice (Fig. 4). Thus, podocytespecific expression of CD2AP completely ameliorated the renal defects of the CD2AP-deficient animals.
Since CD2AP and CIN85 are proposed to be involved in the down-regulation of growth factor receptors (10 -12), we analyzed the propensity of transgenic/CD2AP-deficient mice to develop tumors. Histological analysis of multiple organs from 10.5-month-old transgenic/CD2AP-deficient animals did not reveal any obvious neoplastic processes (data not shown). Fur- thermore, breeding onto a tumorigenic background (p53 heterozygosity) did not increase the incidence of tumors as compared with p53 single-null animals (data not shown).
Since CD2AP has also been implicated in down-regulation of T cell receptor signal transduction, we also looked for evidence of autoimmunity (12). There was no histological evidence of any chronic inflammatory process up to 10 months of age. Fluorescence-activated cell sorter analysis of splenocytes from wild-type and transgene-positive, CD2AP-deficient mice demonstrated normal ratios of B cells to T cells without elevation of CD44ϩ T cells in transgenic, KO mice, indicating that the basal state of T lymphocyte activation is normal in these animals.
Infertility in Transgenic/CD2AP-deficient Mice-During the course of establishing and expanding both transgenic lines, we found that transgene-positive, CD2AP-deficient male mice were infertile. Histological analysis on testes from KO and wild-type mice showed decreased numbers of mature spermatids in transgenic, CD2AP KO mice as compared with littermate controls (Fig. 5). Although there were some hypocellular seminiferous tubules at 5 weeks of age, the majority of the seminiferous tubules were histologically normal (Fig. 5, A and  D). The hypocellularity and sperm loss became more apparent and more widespread as the mice aged (Fig. 5, B, C, E, and F). To confirm that CD2AP is expressed in testes, we stained sections of testes with CD2AP antibodies. We found that CD2AP is specifically expressed in vascular endothelial cells, in Leydig cells, and in the basal portion of the seminiferous tubule (Fig. 6A). Co-staining with anti-vimentin antibody, which in the seminiferous tubule is a marker for Sertoli cells but not for germ cells, revealed partial co-localization of CD2AP with Sertoli cells (Fig. 6, D-F). In Sertoli cells, CD2AP was localized primarily to the basal-lateral surface. Interestingly, when we stained with antibodies to the CD2AP paralog, CIN85, we found a distinct non-overlapping pattern of expression with staining restricted to areas adjacent to and within the lumens of the seminiferous tubules (Fig. 6C). The staining pattern for CIN85 in the wild-type testes was identical in CD2AP KO animals (data not shown). The immunofluorescence findings suggested non-redundant patterns of expression of CD2AP and CIN85 in testis and suggested that CD2AP plays an important role in the supporting cells of the seminiferous tubule and also potentially in immature spermatocytes and spermatogonia. DISCUSSION Although CD2AP is widely expressed, CD2AP-deficient animals die of kidney failure at 6 -8 weeks of age (6). Although the pathology of the kidney is most notable for mesangial changes with the deposition of significant amounts of mesangial matrix by 4 weeks of age, the specific expression of CD2AP in podocytes, proximal tubule, and collecting duct led us to propose that the absence of CD2AP in podocytes was responsible for the massive renal failure. Because the mice begin developing proteinuria around 3 weeks of age and die around 6 weeks of age (6), it was not clear whether other tissues might be affected by the absence of CD2AP. To answer these questions, we generated a transgenic mouse expressing CD2AP on a podocyte promoter (8). When bred to the CD2AP knock-out, we found that the transgene could completely rescue the kidney phenotype of the CD2AP-deficient mouse. We found some ectopic expression of transgenic CD2AP in muscle; however, this should impact neither the kidney phenotype nor the immune phenotype. Importantly, there has been no report of nephrin expression in skeletal muscle in mouse (13) or human (14). However, despite expression of nephrin in the mouse hindbrain (13), there is no evidence of brain transgene brain expression by immunoblot.
Surprisingly, given the wide expression of CD2AP, the CD2AP-deficient, transgenic animals have a normal life span and did not exhibit any obvious abnormalities. This suggests that the place where CD2AP is most important is in the podocyte or that other related molecules can compensate for CD2AP in other tissues. There is one CD2AP paralog, CIN85, which is also broadly expressed in multiple tissues (2,15).
We have now identified two areas, the podocyte and the basal seminiferous tubule, where CD2AP is highly expressed, whereas CIN85 is weak or absent. In the seminiferous tubules, antibodies to CD2AP reacted mainly with the seminiferous epithelium, which is composed of Sertoli cells and immature spermatocytes. Antibodies to CIN85, in contrast, only stained the more apical portions of the seminiferous tubule, where mature spermatocytes exist. This suggested that the loss of CD2AP is having a specific effect on the seminiferous epithelium.
These two tissues with a mutually exclusive pattern of expression of CD2AP and CIN85, the glomerulus and testes, also developed the most severe phenotype in the setting of CD2AP deficiency, suggesting that these two proteins may be somewhat redundant and that loss of one protein may be compensated for by the presence of the paralog. It will be interesting to evaluate the phenotype of a true CIN85 knock-out animal, and eventually, an animal deficient in both of these proteins.
The Sertoli cells form the blood-testis barrier and provide an environment that supports the development of spermato-cytes. Junctional complexes between Sertoli cells prevent the egress of large proteins from the serum into the luminal space of the tubule. This is proposed to function to protect immature spermatocytes from the immune system as they develop as well as to optimize the microenvironment for proper sperm maturation (reviewed in Ref. 16). The stem cells, the spermatogonia, sit in a position at the base of the epithelium and are outside of the blood-testis barrier. As FIG. 6. CD2AP and CIN85 are differentially expressed in wildtype mouse testis. Sections of mouse testis were stained with rabbit anti-CD2AP on wild-type testes (A, oblique cut) with background staining on CD2AP knock-out testes used as a negative control (B). Anti-CIN85 antibodies (C, red) were co-stained with anti-vimentin, which is used as a marker for Sertoli cells but not germ cells (C, green). Anti-CD2AP (D) was co-stained with anti-vimentin (E). The merge (F) reveals that CD2AP partially co-localizes with vimentin at the basal portion of the seminiferous tubules with apparent staining of both Sertoli cells and spermatogonia. CIN85, in contrast, localizes within and adjacent to the lumen of the seminiferous tubules (C). spermatocytes develop (and begin to express new antigens), the spermatocytes move toward the lumen by traversing the space between the Sertoli cells. To do this, they must break and reform junctional complexes to reach the lumen of the tubule (16). Similar to its postulated role in the kidney, CD2AP may play an important role in the biogenesis or stability of junctional complexes in the testis.
Since the testes look histologically normal in young animals, CD2AP is not required to form a normal testis architecture. We suspected that the absence of CD2AP leads either to Sertoli cell loss or to dysfunction with a concomitant breakdown of the bloodtestis barrier as the mice age. In this way, the phenotype of the testis is similar to the kidney phenotype of CD2AP-deficient mice. The kidneys of CD2AP-deficient animals appear histologically normal until around 2-3 weeks of age when proteinuria is first detected (6). We have postulated a role for CD2AP in the function of the slit diaphragm and suggested a breakdown in this specialized junctional complex as blood begins to flow through the glomerular capillaries at high pressure soon after the mice are born. Breakdown of the junctional complexes could result in both podocyte and Sertoli cell loss as well as access of the luminal contents to self-reactive antibodies.
Because CD2AP-deficient T cells show hyper-reactive responses (12), we were interested to see whether CD2APdeficient animals might have an increased incidence of autoimmunity or cancer. Work from several laboratories has suggested that CD2AP and CIN85 are involved in downregulation of tyrosine kinase-based receptors such as epidermal growth factor receptor, hepatocyte growth factor receptor, and the T cell receptor (10 -12). This is supported by reported associations of CD2AP and CIN85 with endophilin, synaptojanin, and phosphatidylinositol-3 kinase (4,10,11,17). Our own work suggests that the defect is not on internalization but rather is related to multivesicular body formation (18). Surprisingly, we found no evidence of autoimmunity or cancer in animals examined up to 1 year of age. This could be related to increased levels of apoptosis (19) or could be due to differences of responses in vivo. Nevertheless, we are currently testing to see whether these transgenic, CD2AP-deficient animals are more susceptible to the induction of autoimmune diseases.
Lastly, given the hyperproliferative phenotype of CD2APdeficient T cells and the deposition of immunoglobulin into the basement membranes of CD2AP heterozygous mice, it was possible that the kidney disease was related to immunologic damage rather than primary podocyte dysfunction. The suc-cessful rescue of the kidney phenotype of CD2AP-deficient animals with podocyte-specific expression of CD2AP convincingly demonstrates that the kidney dysfunction is primary to the podocyte and is not mediated by an immunological process. Supporting this, CD2AP-deficient RAG-deficient mice, which lack T and B cells, develop kidney failure at a rate that is similar to CD2AP-deficient mice alone. 2 Thus, our studies with mice demonstrated that podocyte dysfunction is solely responsible for the kidney failure and death of CD2AP-deficient animals. The additional phenotype of infertility and testicular damage suggested potential roles of CD2AP in other specialized junctional complexes.