Congenital Nephrotic Syndrome in Mice Lacking CD2-Associated Protein

See allHide authors and affiliations

Science  08 Oct 1999:
Vol. 286, Issue 5438, pp. 312-315
DOI: 10.1126/science.286.5438.312


CD2-associated protein (CD2AP) is an 80-kilodalton protein that is critical for stabilizing contacts between T cells and antigen-presenting cells. In CD2AP-deficient mice, immune function was compromised, but the mice died at 6 to 7 weeks of age from renal failure. In the kidney, CD2AP was expressed primarily in glomerular epithelial cells. Knockout mice exhibited defects in epithelial cell foot processes, accompanied by mesangial cell hyperplasia and extracellular matrix deposition. Supporting a role for CD2AP in the specialized cell junction known as the slit diaphragm, CD2AP associated with nephrin, the primary component of the slit diaphragm.

CD2AP is an adapter protein that interacts with the cytoplasmic domain of CD2 (1). CD2, a T cell and natural killer cell membrane protein, facilitates T cell adhesion to antigen-presenting cells. CD2AP enhances CD2 clustering and anchors CD2 at sites of cell contact (1). As CD2AP is widely expressed, it may play roles in other tissues.

Mice lacking CD2AP were generated by replacing the exon encoding the first SRC homology 3 (SH3) domain of CD2AP with a neomycin-resistance gene (Fig. 1A) (2). Two independent homologous recombinant clones were injected into blastocysts to generate chimeric mice (3). Homozygous CD2AP knockout (KO) mice were generated from heterozygous mice with a normal frequency of 25%. The genotypes of the mice were confirmed by Southern (DNA) blotting analysis (Fig. 1B). Immunoblotting (4) demonstrated loss of CD2AP protein in KO animals and reduced levels of CD2AP in heterozygotes (Fig. 1C).

Figure 1

(A) Strategy for homologous replacement of the first SH3 domain of CD2AP with a neomycin-resistance cassette (NEO). Restriction sites in the flanking introns around the targeted exon and the predicted sizes of the resulting fragments from Ssp I digestion are shown. (B) Southern blot analysis of wild-type (WT), heterozygous (HT), and homozygous CD2AP (KO) mice. (C) CD2AP immunoblotting of thymus, kidney, liver, and spleen from WT, HT, and KO mice. (D) SDS-PAGE analysis of urinary protein in WT, HT, and KO mice. Two microliters of urine from 1-, 2-, and 3-week-old mice was analyzed by SDS-PAGE and Coomassie stained. Genotypes were confirmed by immunoblotting with antibodies to CD2AP (bottom). One microgram of bovine serum albumin was run as a control (lane 1).

At about 3 weeks of age, CD2AP KO animals began to exhibit substantial growth retardation, and most KO mice were dead by 6 to 7 weeks of age. Postmortem examination of these mice revealed cardiac hypertrophy, splenic and thymic atrophy, and ascites. Histological examination revealed evidence of severe kidney pathology (Fig. 2D) (5). This pathology correlated with proteinuria, elevated blood urea nitrogen and creatinine concentrations, and reduced serum albumin concentrations, all signs of kidney dysfunction (6). Proteinuria was first detectable around 2 weeks of age (Fig. 1D) (7).

Figure 2

Severe glomerular disease in CD2AP KO mice. Kidneys from wild-type (7-day-old) and from 7-, 14-, and 28-day-old KO animals were analyzed by hematoxylin and eosin staining. (A) Kidney section of a 7-day-old wild-type mouse. (B) Kidney section of a 7-day-old CD2AP KO mouse demonstrating glomerular hypercellularity. (C) Kidney section from a 14-day-old CD2AP KO mouse demonstrating glomerular hypercellularity and mesangial deposits. (D) Kidney section from a 28-day-old CD2AP KO mouse demonstrating glomerulosclerosis and kidney tubule dilatation. Scale bar, 50 μm.

Microscopic examination revealed that the predominant kidney pathology involved the glomerulus, a specialized collection of capillary loops that constitutes the filtration apparatus of the kidney. The filtration barrier consists of the fenestrated capillary endothelial cells, the glomerular basement membrane (GBM), and the foot processes of the glomerular epithelial cells or podocytes (Fig. 3A). The spaces between the podocyte foot processes are regular (∼35 nm) and contain a specialized junctional structure called the slit diaphragm.

Figure 3

Electron microscopic analysis of CD2AP KO glomeruli. (A) Ultrastructural analysis of a glomerular capillary wall from a 7-day-old wild-type mouse demonstrates the normal morphology of the glomerular filtration barrier with normal fenestrated endothelial cells (E), glomerular basement membrane (GBM), and normal podocyte foot processes (FP). (B and C) Analysis of glomerular filtration barrier from a 7-day-old KO mouse shows foot process effacement [arrows in (B)], but normal endothelial cells and basement membranes. In (C), normal foot processes (arrows) adjacent to the damaged foot processes are shown. (D) Ultrastructural analysis from a 28-day-old mouse demonstrating massive extracellular matrix deposition (*) surrounding a mesangial cell. Scale bars, 1 μm.

Compared with the glomeruli of age-matched wild-type mice (Fig. 2A), some of the glomeruli in 1-week-old KO mice showed increases in size and cellularity, suggesting a congenital defect (Fig. 2B). Kidney tubules appeared normal. At 2 weeks of age, almost all glomeruli were affected, and many had evidence of mesangial deposits (Fig. 2C). By 4 weeks, glomeruli were sclerotic, with increased deposits and distended capillary loops (Fig. 2D). There was never any evidence of inflammation.

Electron microscopic (EM) examination of the kidneys indicated that the initial defect involved the podocytes (8). Kidneys from 1-week-old animals showed loss of foot process integrity with obliteration of the spaces between the foot processes (Fig. 3, B and C). These changes were found in all glomeruli and in all capillary loops. However, at this stage, normal foot processes were also present (Fig. 3C). No anomalies of the GBM or endothelial cells were detected. Older animals showed worsening of the foot process damage. In addition, mesangial deposits began to accumulate starting around 4 weeks of age (Fig. 3D). In some areas, these deposits were extensive, encroaching upon the lumen of the capillary loops. There were no subendothelial or subepithelial deposits indicative of immune complex pathology.

Deposits were composed of extracellular matrix normally secreted by mesangial cells. Deposits were strongly positive for fibronectin, collagens α1 and α2(IV), perlecan, and the laminin α1, α2, α5, β1, and γ1 chains, indicating the presence of laminins-1, -2 and -10 (Fig. 4A) (6,9). The GBM contained the usual laminin (laminin-11) and collagen IV isoforms (α3, α4, and α5 chains), and these isoforms were not found in the expanded mesangium (Fig. 4A) (6), suggesting that the deposits were of purely mesangial cell origin. In addition, integrin α3 was properly localized adjacent to the GBM on the basal face of podocytes (Fig. 4A), indicating that the podocytes maintained their proper polarity despite the multiple insults to the glomerulus.

Figure 4

Immunofluorescence studies of wild-type and CD2AP KO glomeruli. (A) The expanded mesangium of mutant glomeruli contains laminin (Lam) α2, laminin γ1, and perlecan (perl; shown in green) as well as other matrix proteins normally secreted by mesangial cells (see text). The composition of the GBM (arrows) is not affected by the mesangial defect; it contains the laminin α5, β2, and γ1 (shown) chains and the collagen (Col) α3, α4 (shown), and α5(IV) chains. Integrin α3 (Int α3; shown in red) appears properly localized juxtaposed to the GBM on the basal surface of podocytes in mutant glomeruli. (B) In glomeruli, CD2AP (shown in red) is expressed exclusively by podocytes, as determined by double-labeling with synaptopodin (Synpo, shown in green), a podocyte marker. Scale bar, 50 μm.

Immunofluorescence studies of wild-type kidney demonstrated that CD2AP was expressed primarily in podocytes (9). CD2AP exhibited an overlapping pattern of expression with synaptopodin, a podocyte foot process marker (Fig. 4B) (10). No CD2AP staining was detected in mesangial cells, but a subset of tubules did stain. CD2AP staining was absent in the kidneys of the KO mice (Fig. 4B).

Recently, mutations in the nephrin gene were identified as the cause of congenital nephrotic syndrome of the Finnish type (11). Nephrin, an immunoglobulin superfamily member, is expressed exclusively in podocytes and is thought to be the major component of the slit diaphragm (12). Because the function of nephrin is reminiscent of that of CD2, we tested whether CD2AP associates with nephrin. CD2AP and nephrin, however, could not be solubilized from purified glomeruli with nonionic detergents, so we could not determine whether the two proteins associate in the podocyte. Therefore, we generated a chimeric protein containing the extracellular and transmembrane domain of vesicular stomatitis virus (VSV) G protein fused to the cytoplasmic domain of nephrin. The fusion protein was expressed with myc-tagged CD2AP, and coimmunoprecipitation was assessed by immunoblotting VSV G immunoprecipitations with antibodies to myc (Fig. 5) (13). The nephrin fusion protein coimmunoprecipitated CD2AP at levels similar to those of the CD2 fusion protein (1). This was specific because a VSV G molecule lacking its cytoplasmic domain (G/T) did not coimmunoprecipitate CD2AP. In addition, nephrin and CD2AP interacted in vitro with the use of purified proteins and by yeast two-hybrid analysis (14).

Figure 5

Association of CD2AP with nephrin. A myc-tagged form of CD2AP was expressed either in HeLa cells alone (lane 1) or with a chimeric protein containing the nephrin cytoplasmic domain (VSV G/nephrin, lane 4), a chimeric protein known to interact with CD2AP (VSV G/CD2, lane 3), or the VSV G protein lacking a cytoplasmic domain (VSV G/T, lane 2). Cell lysates were immunoprecipitated with antibody to VSV G and immunoblotted with antibody to myc (top). Immunoblotting of whole-cell lysates with antibody to myc (middle) or with antibody to VSV G (bottom) was performed to demonstrate similar levels of expression.

Given the role of CD2AP in T cells, we also examined T cell function. Stimulation of KO T cells with antibodies to CD3 or the lectin concanavalin A (Con A) demonstrated impaired T cell function (15). To rule out a relation between the kidney and immune dysfunction, we transplanted bone marrow from KO animals into irradiated wild-type animals (16). Transplanted mice still demonstrated T cell deficits but showed no evidence of kidney dysfunction, demonstrating that the defect is intrinsic to the kidney.

Here, we demonstrated that CD2AP is critical to the integrity of the renal glomerulus. Disease progression appears to begin with epithelial cell injury leading to a mesangial reaction consisting of hyperplasia and massive matrix deposition. This conclusion is supported by the specific expression of CD2AP in podocytes and EM studies demonstrating that damage to podocyte foot processes is the initial lesion. Coimmunoprecipitation studies suggest that CD2AP associates with nephrin, a protein critical for podocyte function. As foot processes can apparently develop in the absence of CD2AP, we suspect that the association of CD2AP with nephrin mainly functions to anchor nephrin to the cytoskeleton.

The role of CD2AP in the renal glomerulus may be similar to its role in the T cell. CD2 and nephrin are both immunoglobulin superfamily proteins involved in forming specialized cell adhesions. In the T cell, this adhesive complex is the immunological synapse (1); in the kidney, it is the slit diaphragm. Supporting a general role for CD2AP in specialized cell contacts, CD2AP was recently shown to associate with the focal adhesion protein, p130CAS(17). Given the phenotype of the mouse, it will be important to determine whether CD2AP plays a role in the pathogenesis of human kidney diseases.

  • * These authors contributed equally to this work.

  • To whom correspondence should be addressed. E-mail: shaw{at}


View Abstract

Navigate This Article