Loss of mucin-type O-glycans impairs the integrity of the glomerular filtration barrier in the mouse kidney

The kidney's filtration activity is essential for removing toxins and waste products from the body. The vascular endothelial cells of the glomerulus are fenestrated, flattened, and surrounded by podocytes, specialized cells that support glomerular endothelial cells. Mucin-type core 1–derived O-glycans (O-glycans) are highly expressed on both glomerular capillary endothelial cells and their supporting podocytes, but their biological role is unclear. Biosynthesis of core 1–derived O-glycans is catalyzed by the glycosyltransferase core 1 β1,3-galactosyltransferase (C1galt1). Here we report that neonatal or adult mice with inducible deletion of C1galt1 (iC1galt1−/−) exhibit spontaneous proteinuria and rapidly progressing glomerulosclerosis. Ultrastructural analysis of the glomerular filtration barrier components revealed that loss of O-glycans results in altered podocyte foot processes. Further analysis indicated that O-glycan is essential for the normal signaling function of podocalyxin, a podocyte foot process–associated glycoprotein. Our results reveal a new function of O-glycosylation in the integrity of the glomerular filtration barrier.

The kidney's filtration activity is essential for removing toxins and waste products from the body. The vascular endothelial cells of the glomerulus are fenestrated, flattened, and surrounded by podocytes, specialized cells that support glomerular endothelial cells. Mucin-type core 1-derived O-glycans (O-glycans) are highly expressed on both glomerular capillary endothelial cells and their supporting podocytes, but their biological role is unclear. Biosynthesis of core 1-derived O-glycans is catalyzed by the glycosyltransferase core 1 ␤1,3-galactosyltransferase (C1galt1). Here we report that neonatal or adult mice with inducible deletion of C1galt1 (iC1galt1 ؊/؊ ) exhibit spontaneous proteinuria and rapidly progressing glomerulosclerosis. Ultrastructural analysis of the glomerular filtration barrier components revealed that loss of O-glycans results in altered podocyte foot processes. Further analysis indicated that O-glycan is essential for the normal signaling function of podocalyxin, a podocyte foot process-associated glycoprotein. Our results reveal a new function of O-glycosylation in the integrity of the glomerular filtration barrier.
The kidney acts as an ultrafiltration unit to remove toxins and waste products from the body. Filtration in the kidney occurs via the glomerulus, which consists of a specialized capillary bed and support cells. The vascular cells of the glomerulus are fenestrated, flattened endothelial cells and are surrounded by podocytes, specialized support cells that envelop the glomerular capillaries by extending foot processes that interdigitate with foot process from adjacent podocytes (1). Glomerular cap-illary endothelial cells, podocytes, and the fused extracellular matrix of these cells, termed glomerular basement membrane (GBM), 3 together comprise the glomerular filtration barrier, which retains blood plasma proteins and results in a urinary product with only trace amounts of protein. A hallmark of compromised glomerular filtration barrier integrity is elevated levels of protein/albumin in the urine, a condition called proteinuria/albuminuria (2).
Glomerulosclerosis, which presents with proteinuria in over 70% of cases, is a major clinical problem because scarred glomeruli cannot be repaired; many patients experience progressive glomerulosclerosis and eventual kidney failure. Therefore, it is imperative to identify factors that perturb the integrity of the glomerular filtration barrier and participate in the etiology of glomerulosclerosis.
Mucin-type O-glycosylation is initiated by addition of a Gal-NAc to serine or threonine residues on a peptide backbone (3). Core 1-derived O-glycans (O-glycans), a major form of O-glycans, are formed by addition of a galactose to the GalNAc by core 1 ␤1,3-galactosyltransferase (C1galt1) (4). The core 1 structures can be further extended and modified to form sialylated complex O-glycans (3). O-glycans are highly expressed in different components of the glomerular filtration barrier, with unclear functions (5)(6)(7). In this study, we found that mice with deficiency of O-glycans exhibit spontaneous proteinuria and rapidly progressing glomerulosclerosis. Lack of O-glycosylation results in defective podocalyxin signaling function. These results reveal new function of O-glycans in homeostasis of the glomerulus.

Postnatal deletion of C1galt1 leads to spontaneous glomerulosclerosis
To determine whether O-glycans play a significant role in kidney function, we generated mice with a doxycycline-induc-  (8). We confirmed the C1galt1 deletion by immunohistochemical staining for Tn antigen, which is exposed in the absence of core 1 O-glycans. As expected, wild-type controls did not exhibit Tn signals within the glomerular capillary tuft, podocytes covering the tuft, parietal epithelial cells of Bowman's capsule, tubules, or the interstitial blood vessels. In contrast, iC1galt1 Ϫ/Ϫ mice exhibited significant Tn expression in all of these locations (Fig. 1, A and B). iC1galt1 Ϫ/Ϫ mice had significant proteinuria at the time of sacrifice (ϳ4 months old) ( Fig. 1, B and C). Gross analysis revealed that iC1galt1 Ϫ/Ϫ mice showed contracted, paler, and corrugated kidneys (Fig. 1D). Periodic acid-Schiff (PAS) staining, which highlights irregularities of the glomerular architecture, showed that iC1galt1 Ϫ/Ϫ mice had adhesion formation between the glomerular capillary tufts and Bowman's capsule, obliteration of glomerular capillaries with hyalinosis, and severe intra-and periglomerular immune cell infiltrations (Fig. 1E) relative to WT controls. Masson trichrome staining (MTS) further demonstrated glomerular scarring, tubular atrophy, proteinaceous casts, and interstitial fibrosis in iC1galt1 Ϫ/Ϫ mice (Fig. 1F).

Adult deletion of C1galt1 leads to spontaneous glomerulosclerosis
To corroborate our finding, we sought to investigate whether spontaneous glomerulosclerosis occurs in adult mice with deficiency of C1galt1. Global deletion of C1galt1 in adult mice was achieved by administration of doxycycline to 6-week-old C1galt1 f/f ;Rosa26-rtTA;tetO-Cre Tg mice for 2 weeks. The kidneys of iC1galt1 Ϫ/Ϫ mice expressed Tn antigen in endothelial cells of glomerular capillaries and interstitial blood vessels as well as in podocytes and tubular epithelial cells. Spontaneous albuminuria, glomerulosclerosis, and broad capsular adhesions were observed in these mice by 8 weeks post-induction (Fig. 2, A-C) (9). These results demonstrated that O-glycans play a major role in maintaining the glomerular filtration function in healthy adult mice.

Alterations of foot processes of podocytes occur prior to development of proteinuria in mice with C1galt1 deficiencies
To investigate how a lack of O-glycans leads to abnormal kidney architecture and function, we determined renal defects at early time points after induced deletion of O-glycans. 6-week-old C1galt1 f/f ;Rosa26-rtTA;tetO-Cre Tg mice and WT littermates were fed food containing doxycycline for 2 weeks. Minimal irregularities of glomerular filtration or architecture were observed in these mice another 2 weeks after doxycyclineinduced deletion of C1galt1 (Fig. 3, A and B). However, we found that the ultrastructures of the podocyte-endothelium interface were strikingly different in iC1galt1 Ϫ/Ϫ mice compared with age-matched controls. Early effects of loss of O-glycans were collapse of the space between foot processes and secondary processes and thickening of the glomerular basement membranes, whereas WT littermates displayed podocyte foot process interdigitations (Fig. 3, C and D). These results support the theory that changes in podocyte foot processes in the absence of O-glycans are a prerequisite for development of the renal phenotype.

O-glycosylation of podocalyxin is essential for glomerulus integrity
Podocyte foot processes associate with many O-glycosylated sialoproteins such as podoplanin and podocalyxin (10, 11).

O-glycosylation and glomerulosclerosis
Podoplanin is a mucin-type transmembrane glycoprotein with extensive O-glycosylation (10,11). Our laboratory has demonstrated previously that O-glycosylation is critical for the stability and function of podoplanin (8,12,13). However, we found that postnatal global deletion of podoplanin only resulted in mild renal disease by 4 months post-induction (Fig. 4). This led us to focus on another O-glycoprotein, podocalyxin, which is expressed predominately at secondary foot processes of podocytes (14). Genetic deletion of Podocalyxin causes complete flattening of foot processes (15), which is compatible with the phenotypes we observed in mice lacking O-glycans. We examined the effect of reduced O-glycosylation on podocalyxin expression and function. Podocalyxin displayed a slower electrophoretic mobility but comparable expression in kidneys from iC1galt1 Ϫ/Ϫ mice compared with WT littermate controls (Fig. 5, A and B). Immunoprecipitated podocalyxin from iC1galt1 Ϫ/Ϫ mice displayed a concomitant increase in VVL staining (binding to Tn antigen) but decreased MALII and SNA staining (binding to sialylated glycans) by Western blotting, suggestive of a dramatic decrease in sialylated O-glycans on podocalyxin (Fig. 5B). In vitro evidence demonstrated that deletion of O-glycosylation sites in the extracellular domain of podocalyxin affects its apical targeting. Podocalyxin and its cytoplasmic binding partner NHERF participate in the forma-  4). For quantitative analysis of kidney histology, 50 full-size glomeruli for each specimen were assessed on PAS-stained sections, and the level of glomerulosclerosis in each glomerulus was semiquantitatively scored as follows: 0, no sclerosis; 1, sclerosis less than 10% of glomeruli; 2, sclerosis 10% to ϳ25% of glomeruli; 3, sclerosis 25% to ϳ50% of glomeruli; 4, sclerosis more than 50% of glomeruli. Data are presented as mean Ϯ S.D. (n ϭ 50). To evaluate interstitial fibrosis, 20 fields for each section were assessed as follows: 0, no fibrosis; 1, fibrosis less than 10% of areas; 2, fibrosis 10% to ϳ25% of areas; 3, fibrosis 25% to ϳ50% of areas; 4, fibrosis more than 50% of areas. The averages of the glomerulosclerosis and interstitial fibrosis scores were calculated from the total evaluated glomeruli or interstitial lesions in each section. Data are presented as mean Ϯ S.D. (n ϭ 20). *, p Ͻ 0.05; **, p Ͻ 0.01. Scale bars ϭ 50 m.

O-glycosylation and glomerulosclerosis
tion of an apical scaffold (16 -18). Podocalyxin binds to ezrin, directly or indirectly via NHERF, to form a complex triggering RhoA-ROCK signaling, thus inducing actin cytoskeleton reorganization. The complex, in turn, is stabilized by elevated RhoA/ROCK-dependent phosphorylation of Ezrin (19,20). Given these findings, one plausible mechanism is that O-glycans are required for the correct distribution and signaling function of podocalyxin at the apical surface of podocytes. Intriguingly, we found that loss of O-glycosylation decreased both co-immunoprecipitation of podocalyxin by NHERF2 and co-localization of podocalyxin and pEzrin along the apical plasma membrane of podocytes (Fig. 5, C and D). These data suggest that loss of the podocalyxin interaction with NHERF2/ pEzrin is a potential effector mechanism of renal dysfunction following loss of O-glycans.

Podocyte-specific deletion of C1galt1 leads to spontaneous glomerulosclerosis
In humans, focal segmental glomerulosclerosis (FSGS) is a frequent and severe glomerular disease characterized by destabilization of the podocyte architecture (21). To evaluate a possible role for podocyte O-glycans in maintaining the glomerular filtration barrier, we determined whether a podocyte-specific deletion of C1galt mimics the renal defects in iC1galt1 Ϫ/Ϫ mice. We generated adult mice with an inducible podocytespecific deletion of C1galt1 by crossing C1galt f/f mice with podocin-iCreER(T2) mice, which express tamoxifen-inducible Cre in podocytes (22). We found that C1galt f/f ;podocin-iCre-ER(T2) mice treated with tamoxifen postnatally from day 21 for 6 months developed albuminuria spontaneously. The kidneys of tamoxifen-treated mice were analyzed in further detail by PAS and MTS, which revealed gross histological changes, beginning with severely distorted and sclerosed glomeruli, compared with littermate controls (Fig. 6). These findings indicate that O-glycans on podocytes play an essential role in glomerular filtration barrier maintenance.

Discussion
O-glycans are highly expressed on glomerular capillary endothelial cells, their supporting podocytes, and tubular cells in the kidney. We found that global loss of O-glycans in adult iC1galt1 Ϫ/Ϫ mice results in spontaneous proteinuria and glomerulosclerosis. Further analysis indicates that O-glycosylation of podocytes is essential to protect the integrity of the glomerulus. In particular, O-glycosylation is important for podocalyxin to interact with NHERF2/pEzrin, which contribute to glomerular filtration barrier maintenance.
Postnatal Tn exposure because of C1galt1 deficiency may elicit an immune response (23), which may lead to kidney glomerulosclerosis and proteinuria. However, we found that crossing our adult iC1galt1 Ϫ/Ϫ with Rag1 Ϫ/Ϫ mice, which lack T cells and B cells (24,25), did not ameliorate the renal defects of iC1galt1 Ϫ/Ϫ mice, indicating that an abnormal immune response does not contribute to the kidney phenotypes of C1galt1 deficiency.
Sialylation commonly occurs on O-glycans. Our iC1galt1 Ϫ/Ϫ mice show reduced sialylation. Previously reports show that the morphologic alterations of podocytes are associated with a reduction in the sialic acid content of podocalyxin in puromycin aminonucleoside-treated rats (26,27), and systemic administration of a sugar analog that eliminates sialylation causes profound kidney dysfunction (28). Together, these data support that loss of sialylation may alter the function of podocalyxin. The mouse genetic approach we used in this study results in prolonged impairment of sialomucins and, thus, convincingly validates the importance of sialomucins in the kidney. The readouts, changes in the distribution of podocalyxin, and changes in podocalyxin signaling are logical outcomes of the loss of C1galt1 function by cells that make podocalyxin.
Most of the data are derived from mice with inducible global deletion of C1galt1. Even though the proteinuria/glomerulosclerosis phenotype is largely reproduced in mice with podocyte-specific deletion of C1galt1, indirect contributions of deficiency of O-glycans in cell types other than podocytes might also contribute to these abnormalities. Future studies are needed to address this issue.
In humans, podocalyxin expression in glomeruli by immunohistochemical evaluation is found to be reduced or lost in FSGS (29). Indeed, we found that adult mice with loss of C1galt1 exhibited reduced podocalyxin levels. However, podocalyxin levels appeared to be normal in mice with earlier stages of renal disease following deletion of C1galt1 (30,31), suggesting that reduced levels of podocalyxin in glomeruli might be caused by loss of podocytes. On the other hand, excretion of podocalyxin in the urine demonstrates a prognostic value for proteinuria development and kidney function in FSGS (32).

O-glycosylation and glomerulosclerosis
Given that O-glycosylation functions as an essential regulatory mechanism of podocalyxin expression and function, further studies of how O-glycans regulates podocalyxin function will provide new insights into the pathogenesis of glomerulosclerosis, and an ability to detect aberrant O-glycosylation will shed new light on the prognosis of glomerulosclerosis.

Mice
We previously reported mice in which C1galt1 was flanked by loxP sites (C1galt1 f/f mice) (12). We also reported mice with doxycycline-inducible global deficiency of C1galt1 (iC1galt1 Ϫ/Ϫ ), which were generated by crossing C1galt1 f/f mice with Rosa26-rtTA, tetO-Cre Tg mice (8). To induce postnatal deletion of C1galt1, mice were fed food containing doxycycline starting from postnatal days 1 to 4 months of age. Adult deletion was accomplished by feeding 6-week-old mice for 2 weeks. Wildtype littermates treated with the same regimen were used as controls. Tamoxifen-inducible podocyte-specific C1galt1 knock-out mice were obtained by crossing C1galt1 f/f mice with podocin-iCreER(T2) mice that express Cre recombinase specifically in the podocytes (22). Deletion of C1galt1 in C1galt1 f/f ; podocin-iCreER(T2) adult mice was accomplished by administrating tamoxifen (1 mg per day) for 5 consecutive days beginning at P21 and then once a week thereafter. We previously reported a strategy to generate mice with global inducible deletion of podoplanin (PDPN f/f ;CagCre mice) (13). Briefly, Pdpn f/f mice were crossed with CAG-Cre-ER T2 Tg mice. To induce postnatal deletion of podoplanin, tamoxifen (20 g per day) was administered orally from P1-5. WT littermates (Pdpn f/w ;CagCre or Pdpn f/f ) treated with the same regimen were used as controls. All vertebrate procedures were performed in accordance with the Guide for the Care and Use of Laboratory Animals and approved by the Institutional Animal Care and Use Committee at the Oklahoma Medical Research Foundation.

O-glycosylation and glomerulosclerosis Biochemical measurements of urine albumin and urine creatinine
Urine samples were collected from knock-out mice and littermate controls, briefly centrifuged, diluted, and boiled in 2ϫ Laemmli sample buffer. Albuminuria was assessed by 10% SDS-PAGE followed by Coomassie Blue staining. Urine albumin levels were measured qualitatively and in duplicate using an albumin ELISA quantitation kit according to the protocol of the manufacturer (Exocell), and the absorbance was read at 450 nm. Urine creatinine was measured in duplicate for each sample with a modified Jaffe reaction at an absorbance of 500 nm (Bio-Rad microplate reader).

Histological and ultrastructural analysis
Mouse kidney tissues were harvested, fixed in 4% paraformaldehyde, and embedded in paraffin. All sections (3 m in thickness) were stained with H&E, PAS (Thermo Scientific) or MTS (Sigma) according to the standard procedures or the instructions of the manufacturer. For transmission EM, kidneys were fixed in Karnovsky fixative containing 2% paraformaldehyde and 2.5% glutaraldehyde in 0.1% cacodylate buffer (pH 7.2) for 2 h at room temperature, followed by post-fixation with 1% osmium tetraoxide in cacodylate buffer for 90 min at 4°C and mordanting with 1% tannic acid in cacodylate buffer. Ultrathin sections were contrasted with uranyl acetate and lead citrate and examined with a Hitachi H-7600 transmission electron microscope located in the Imaging Core Facility of the Oklahoma Medical Research Foundation. We followed histological scoring methods described previously (9).

Immunohistochemistry and immunofluorescent staining
The study was performed as described previously with modifications (33). For IHC staining, sections were deparaffinized with xylene and alcohol series. After inactivation of endogenous peroxidase with 0.3% hydrogen peroxide and blocking with normal blocking serum, the sections were incubated with glycosylation-independent goat anti-podocalyxin antibody (R&D Systems) or biotinylated mAb against Tn antigen (mouse IgM, clone Ca3638) overnight at 4°C. After washing, sections were incubated with appropriate secondary antibodies conjugated to horseradish peroxidase. Immunoreactivity was visualized using a peroxidase diaminobenzidine kit (Vector Laboratories). Sections were then washed, counterstained with hematoxylin, and mounted, and photomicrographs were obtained. For IF staining, sections were blocked as described above and then incubated with primary antibody against CD31 (clone MEC13.3, Pharmingen, after treatment with 0.1% trypsin), podocalyxin (R&D Systems), CD45 (clone I3/2.3, Abcam), phosphorylated Ezrin Thr-567 (pEzrin, Cell Signaling Technology), or biotinylated mAb against Tn antigen (mouse IgM, clone Ca3638) overnight at 4°C, followed by incubation with the respective secondary antibodies conjugated to fluorescent labels (Alexa Flour 594 or 488, 1:500, Invitrogen) for 1 h at room temperature. Slides were washed in PBS, stained with DAPI, and mounted, and immunofluorescence images were obtained using an Olympus fluorescence microscope equipped with Slidebook 5.0 analysis software.

Immunoprecipitation and Western blotting
For immunoprecipitation, kidney cortices were lysed with RIPA buffer (1% Triton X-100, 0.1% SDS, 0.5% sodium deoxycholic acid, 5 mM tetrasodium pyrophosphate, 50 mM sodium fluoride, 5 mM EDTA, 150 mM NaCl, 25 mM Tris (pH 7.5), 4 mM Na 3 VO 4 , 5 mM N-ethylmaleimide, and protease inhibitor mixture). Lysates were renatured using 9 volumes of ice-cold RIPA buffer and then prepared for immunoprecipitation as follows. Cell lysates were precleared with protein AϩG-Sepharose beads for 1 h at 4°C, followed by incubation with antibodies against podocalyxin (R&D Systems) or NHERF2 (Santa Cruz Biotechnology) for 4 h at 4°C. For negative controls, an equal concentration of goat IgG was added instead of specific antibodies. Precipitated proteins were eluted from beads using 2% SDS in 50 mM Tris (pH 7.5) and diluted 1:20 with RIPA buffer, followed by Western blotting. Proteins were resolved by SDS-PAGE (4 -15% acrylamide), followed by electroblotting to a polyvinylidene difluoride membrane and blocking with 5% milk (w/v). Primary antibodies against VVL, MALII, or SNA (Vector Laboratories) were incubated at 4°C overnight, followed by incubation at room temperature with the respective horseradish peroxidase-conjugated secondary antibodies (1:5000) for 1 h. The immunoreactive proteins were detected by enhanced chemiluminescence with autoradiography.

Statistical analysis
Statistical analysis was performed using Student's t test. Differences were considered statistically significant at p Ͻ 0.05.