Proteoglycans (PGs) are important for the glomerular barrier, for cell signaling, and for the anchorage of cells to the glomerular basement membrane. They are, however, complex macromolecules, and their production has not yet been thoroughly investigated in podocytes. In the present study, we studied the biosynthesis of PGs by highly differentiated human podocytes and in rats. The cells were treated with puromycin aminonucleoside (PAN; a nephrosis-inducing agent), steroids (used as primary treatment for nephrotic syndrome), or both. Analysis was made by TaqMan real-time PCR, Western blotting, and by metabolic labeling with 35S and 3H. We found that podocytes produce versican, syndecan-1, decorin, and biglycan together with the previously known PG syndecan-4, glypican, and perlecan. PAN treatment downregulated the mRNA and the protein expression of both versican (by 24 ± 6%, P < 0.01, for mRNA and by 50% for protein) and perlecan (by 14 ± 5%, P < 0.05, for mRNA and by 50% for protein). The decreased expression was confirmed by studying the glomerular gene expression in rats treated with PAN during a time course study. In addition, puromycin decreased the expression of enzymes involved in the glycosaminoglycan biosynthesis. Steroid treatment decreased perlecan (by 24 ± 3%, P < 0.01) and syndecan-1 expression (by 30 ± 4%, P < 0.01) but increased the expression of decorin 2.5-fold. The observed alterations of PG synthesis induced by PAN may lead to decreased glomerular anionic charge and disturbed podocyte morphology, factors that are important for the development of a nephrotic syndrome.
- glomerular barrier
- glycosaminoglycan biosynthesis
a growing amount of evidence supports the idea of the glomerular filtration barrier as an integrative structure (7, 15). It is known that the barrier is selective to solutes passing through it depending on their shape, size, and charge (7, 30). Most of the size-selective properties have lately been ascribed to the podocytes and their slit diaphragm, where several novel proteins have been found during the last decade (14, 38, 41), whereas the site for charge selectivity seems to be situated in several levels of the barrier (7, 15). The glomerular basement membrane is the structure that has been ascribed the most charge-selective properties (21, 41), but it has recently been questioned (4). We and others previously showed that the glomerular endothelial glycocalyx consists of negatively charged components such as proteoglycans (PGs) (3, 20, 34, 39) and that it has charge-selective properties (20, 40, 42). Another possible site for charge interaction is the podocyte and there are previous reports concerning the podocyte cell surface and the involvement of negatively charged molecules such as podocalyxin (23) and the PG agrin (32). PGs are built up by a core protein with one or more polyanionic galactose or glucosaminoglycan (GAG) chains. The GAG chains are involved in cell signaling through the binding and storing of several different growth factors (such as TGF and FGF), cytokines, and enzymes (6). PGs are divided into different classes and found extracellularly, intercalated in cell membranes or attached via GPI anchors and intracellularly in storage granules (3).
In the present study, we investigated PG and GAG synthesis in well-characterized, highly differentiated human podocytes, with or without puromycin aminonucleoside (PAN; nephrosis-inducing agent) and in combination with steroid treatment. In addition, we looked at the glomerular gene expression as well as physiological parameters in rats treated with PAN during a time course study of 7 days. In particular, we were interested in investigating which class of core proteins was expressed and which key transferases were involved in the synthesis of PGs and their GAGs. Our hypothesis is that these molecules are involved in glomerular function and the development of proteinuria.
MATERIALS AND METHODS
Differentiated, growth-arrested podocytes were derived from a cell line which was conditionally transformed using a temperature-sensitive mutant of SV-40 T antigen. Additionally, the cells have been stably transfected with the catalytic subunit of the human telomerase gene (29). These cells have previously been described and characterized elsewhere (35). At the “permissive” temperature of 33°C, the SV-40 T antigen was active and allowed the cells to proliferate rapidly. Thermo switching the cells to the “nonpermissive” temperature of 37°C inactivated the T antigen, and the cells became growth arrested and expressed markers of differentiated podocytes in vivo (synaptopodin, CD2AP, and α-actinin-4).
Experimental Animals and Data Analysis
Female Sprague-Dawley rats (∼200 g, Harlan) were used for in vivo measurements of renal function and gene expression. PAN (Sigma, St. Louis, MO) was administered to the rats in a single intraperitoneal injection, in a dose of 150 mg/kg, while the control group received an equal dose of saline solution. The experiments were performed as a time course study (7 days) starting on day 1 after PAN injection. In addition were new rats treated 5 days with PAN alone or in combination with Solumidrol (3 mg/100 g, at days 0, 2, and 4 after the PAN injection). The experiments were approved by the local ethical committee. The protocol for the physiological measurement has been described previously (16). Glomeruli were isolated from the renal cortex by gradual sieving, and RNA was isolated by using the Qiagen Mini protocol (VWR, Stockholm, Sweden). Plasma and urine samples were analyzed for 51Cr-EDTA concentrations using a gamma counter. Albumin concentrations of urine and plasma samples were detected by size-exclusion chromatography by UV detection at 226 nm.
Cell Culture and Stimulation with PAN and Dexamethasone
Human podocytes were seeded into 25-cm2 flasks in RPMI 1640 medium with added insulin, transferrin, and selenite (Sigma), PSA solution (penicillin, streptomycin, and amphotericin B, Cascade Biologics, Portland, OR), and 10% FCS. The cells were allowed to differentiate during 16 days at 37°C before the experiment was started. Four different groups of cells were used: 1) cells were stimulated with PAN at a concentration of 1 μM; 2) stimulation with a combination of PAN and dexamethasone (1 μM each, Sigma); 3) treatment with dexamethasone (1 μM) alone; and 4) nonstimulated control cells grown under normal conditions (i.e “starved” in standard media with 1% FCS). After 48 h of stimulation, preparation of total RNA (Qiagen) and protein [lysed in 1% SDS, Complete mini protease inhibitor (Roche) and 10 mM Tris, pH 7.4] was performed.
Reverse transcription of RNA from the podocytes (four groups) and rat glomeruli (two groups) were performed using a standard protocol at a final cDNA concentration 50 ng/μl. The mRNA level of each target gene was quantified by real-time PCR on the ABI Prism 7900 Sequence Detection system [TaqMan, Applied Biosystems (ABI), Foster City, CA], as previously described (3). The genes for the PG core proteins and enzymes used are described in Tables 1 and 2. All of the primer probe pairs were ordered from ABI. We used both standard 96-well plates and “Low-density array microfluidic cards” to analyze the different genes. The microfluidic card is a new PCR application where very small sample volumes may be used. One sample (50 ng cDNA in 100 μl) is divided into 46 compartments of 2 μl by centrifugation. The samples are then run in duplicate for 23 different genes (including several endogenous control genes) in one run. The “96-well PCR” was carried out in a 50-μl reaction mix containing 50 ng sample cDNA/well in a 50-μl mix of probe, primers, and TaqMan universal PCR master mix (ABI; containing MgCl2, dNTPs, Taq Gold polymerase, and AmpEraseUNG). In both applications were the samples denatured at 95°C for 10 min and then subjected to 40 cycles of two-step PCR (15 s at 95°C, 1 min at 60°C). All samples run in 96-well plates were amplified simultaneously in triplicate in one assay run. Standard curves were computed for all genes from a series of twofold serial template dilutions from 3,125 through 200 ng (7 concentrations). For each sample, the amount of each target gene and endogenous control was determined from the corresponding standard curves. Log input amount was calculated according to the following formula: 10 − [(cell containing CT value) − b]/m; where b = y-intercept of standard curve line, and m = slope of standard curve line. Validation of amplification efficiency was made for every primer/probe set and was calculated for each run (Table 1).
For the “low-density array,” samples were run in duplicate and the comparative ΔΔCT method of relative quantification was used to calculate the differences in gene expression between the control and the three treatment groups. As endogenous controls, we used 18S and GAPDH (predesigned assay reagent applied by ABI, external run) to correct for potential variation in RNA loading or efficiency of the amplification reaction. The endogenous control genes were chosen after running of 11 housekeeping genes to find the most stable genes in the material investigated (“human endogenous control plate,” ABI). As a negative control, we used mRNA samples without performing reverse transcription.
Protein concentrations from cells and cell media were determined by using a BCA-Protein Assay Reagent kit (Pierce, Rockford, IL). Equal amounts of total protein were separated on Novex (San Diego, CA) precast 4–12% Bis-Tris gels for decorin and syndecans (-1 and -4), or 10% Tris·HCl gels (Bio-Rad, Hercules, CA) for versican and perlecan, and transferred to either polyvinylidene difluoride membrane (decorin and syndecans) or nitrocellulose membranes (versican and perlecan) (3). The following primary antibodies were used: rabbit anti-syndecan-1 and anti-syndecan-4 (Zymed, San Francisco, CA), rabbit anti-decorin (Abcam, Cambridge, UK), mouse anti-perlecan (Zymed), and rabbit anti-versican [a kind gift from Dr. D. Zimmermann, Dept. of Pathology, Zurich, Switzerland (44)]. We used mouse anti-β-actin (Sigma) as a loading control for perlecan and versican; the other PGs analyzed had bands that were the same size as the control antibody, ∼40 kDa. After being washed, membranes were incubated with the appropriate horseradish peroxidase (HRP)-conjugated secondary antibody (Amersham Life Science, Amersham, UK). Immunoreactive bands were visualized using ECL+ (Amersham) and a CCD camera (LAS1000; Fujifilm, Tokyo, Japan). The intensity of the bands was measured by densitometry (Image Gauge V.3.45) and compared between the groups.
Immunolocalization for the PG versican (rabbit anti-human, 1:100, Zimmermann), syndecan-1 (rabbit anti-human, 1:200, Zymed), and perlecan (mouse anti-human, 1:100, Zymed) were performed on human cortex. Sections (4 μm) were fixated in PBS containing: 2% FCS, 1% bovine serum albumin, 0.1% Triton X-100, and 0.05% Tween 20 for 30 min. Each primary antibody was incubated for 1 h at room temperature (RT), followed by washing steps and incubation with either goat anti-mouse IgG-Texas Red (1:500, Abcam) or sheep anti-rabbit IgG-FITC (1:200, Abcam). Staining of podocytes was made with either rabbit anti-podocin (1:1,000, Sigma, for perlecan) or mouse anti-synaptopodin (1:100, Abcam, for versican and syndecan-1) for 1 h at RT. After being washed with PBS-Tween, the proper secondary antibody, either Texas red or FITC (described above), was used.
Biolabeling of GAG
Colabeling with sulfate and tritiated glucosamine.
Human podocytes, in four flasks for each group, were stimulated and grown under the same conditions as described above. Stimulation was performed for 48 h at 37°C in a humidified, 5% CO2 atmosphere. Metabolic colabeling with [D-6-3H]-glucosamine (20 μCi/ml) and sulfate-35 (30 μCi/ml; Amersham Biosciences, Uppsala, Sweden) was done to measure both the length and the sulfate content of the GAG chains. This procedure labels only the GAGs synthesized during the 48-h incubation period. After stimulation, the conditioned media were collected and further analyzed by chromatography.
Analysis of PG Synthesized and Released to the Media
Ion exchange and molecular sieve chromatography for colabeled molecules (35S and 3H).
The cell media were separated in a serial HPLC step (Gynkotek, Germering, Germany). First, the samples were applied to an ion-exchange column (TSK gel, DEAE-5PW, 7.5 mm × 7.5 cm, Tosoh Bioscience, 50 mM phosphate buffer, pH 7.0), where negatively charged PGs bind to the column and other molecules were wasted. The bound PGs were eluted by a step increase in NaCl concentration from 0 to 1 M and transferred onto a size-exclusion column (Bio Sep-SEC-S3000, 300 × 7.8 mm, in 50 mM phosphate buffer, pH 7.0, 0.5 ml/min). Separated PGs were collected in 1-ml fractions and aliquots were quantified by scintillation counting. The system was calibrated by using LMW and HMW kits from Amersham Bioscience.
Results are presented as means ± SE, and differences were tested using the Wilcoxon rank sum test. For the time course study in rats, ANOVA and regression analysis were used.
Human Podocyte Cell Culture
The cells were of podocyte origin as judged by light microscopy, by initial characterization (35), and by positive PCR for synaptopodin, α-actinin-4, and CD2AP (data not shown). No significant difference in cell number, amounts of RNA or protein were seen between the groups after the stimulation period, confirming that there is no cytotoxic effect by PAN (24, 39) or dexamethasone at the concentration used.
Gene Expression Analysis for Podocytes
Several PGs were tested in the TaqMan real-time PCR analysis system and were found to be expressed by the human podocytes: syndecan-1 and -4, versican, glypican, perlecan, decorin, and biglycan. The expression of versican was downregulated by 24.3 ± 6% (means ± SE) in the PAN-treated group compared with the control (P < 0.01, n = 8). PAN treatment also downregulated perlecan by 14.4 ± 5% (P < 0.05, n = 8). We could not see any significant difference in expression between the groups (control and PAN) for the membrane-bound PGs syndecan-1 and -4 or glypican-1, or the secreted SLRPs (small leucine-rich PG; Fig. 1A). Treatment with PAN in combination with dexamethasone decreased the perlecan expression by 32 ± 3% compared with control (P < 0.01, n = 8). Steroid treatment alone downregulated the expression of both perlecan (23 ± 3%, P < 0.01, n = 8) and syndecan-1 (30 ± 4%, P < 0.01, n = 8) significantly. In addition, decorin was upregulated 2.3-fold (P < 0.01, n = 8; Fig. 1A).
Enzymes involved in the GAG chain modification (sulfate content and length of chain) were also analyzed by real-time PCR. In galactosaminoglycan assembly, chondroitin 4-O-sulfotransferase was downregulated by PAN by 23.4 ± 6% (P < 0.05, n = 8). In addition, two of the glucosaminoglycan enzymes important for the heparan sulfate (HS) biosynthesis were downregulated: EXT-1 (extoses) by 44.7 ± 6% (P < 0.05, n = 6) and N-deacetylase N-sulfotransferase (NDST) by 24.7 ± 5% (P < 0.05, n = 8; Fig. 1B). Treatment with dexamethasone alone decreased the expression of EXT-1, NDST, and heparan sulfate 3-O-sulfotransferase (HS3ST; Fig. 1B).
Validation of amplification efficiency was made for the primer/probe sets used in the 96-well plates and calculated for each run from the corresponding standard curve (see Table 1). For quantification of the results from the low-density array cards, we used comparative CT (ΔΔCT) relative quantification method (see materials and methods).
Protein Expression Analyzed by Western Blot
Syndecan (-1 and -4), decorin, versican, and perlecan were expressed by the podocytes as revealed by Western immunoblotting (Fig. 2). The versican antibody detects the V0 and V1 splice variants (∼370 and ∼260 kDa). When comparing the density of the sample bands for perlecan and versican, we could see a 40–50% lower intensity in the PAN-treated cells compared with control. In the syndecan blots, we could only see small or no differences in intensity between the two groups. The combination of dexamethasone and PAN did not improve the expression of versican. The protein expression of decorin was increased by the dexamethasone treatment in an analysis of both the cell media and the cell protein content (Fig. 2, E and F).
Immunohistochemistry performed on sections of human cortex shows that versican, syndecan-1, and perlecan colocalize with the podocyte-specific proteins podocin and synaptopodin (Fig. 3). The photos also demonstrate that these PGs are expressed by other cells in the glomeruli as well.
Secreted PG Synthesized Into the Culture Media
Colabeling with sulfate and tritiated glucosamine.
Labeled secreted PGs in the cell medium from both control and PAN-treated podocytes were separated in two steps, first by charge (on a DEAE column) and then by size-exclusion chromatography. The amounts of radioactive incorporation into large molecules (>73 Å) were reduced by ∼50% in the PAN-treated group compared with control, indicating perlecan and/or versican. For molecules of the sizes 55–73 Å, a ∼10% decrease for both 35S and 3H was seen, whereas there were increased amounts of incorporated sulfate by 28% for 27–55 Å and increased amounts of tritium for the size range 27–40 Å. Treatment with dexamethasone increased the amount of isotope incorporation threefold in the fractions of smaller PGs (<55 Å), compared with control, indicating decorin (data not shown).
Glomerular Filtration Rate and Fractional Clearance for Albumin
Values for glomerular filtration rate (GFR) and fractional clearance for albumin are presented in Fig. 4. The GFR for the PAN-treated rats was reduced over time from 0.9 ml·min−1·g wet wt−1 for day 1 to 0.075 ml·min−1·g wet wt−1 for day 7. The GFR for the control group was unaffected 0.84 ± 0.05 ml·min−1·g wet wt−1.
The fractional clearance was 0.00022 ± 0.00006 in controls. In day 3 after PAN administration, we could detect an increased fractional clearance with 0.00042, which reached 0.071 at day 7.
Gene Expression Analysis for Rat Glomeruli
As shown in Figs. 5 and 6, we could see an altered mRNA expression over time for several of the genes analyzed. The membrane-bound PG syndecan-4 (by 34 ± 7.5%, P < 0.001, n = 18; Fig. 5A) and the secreted versican (by 60 ± 6.5%, P < 0.001, n = 18; Fig. 6A) had significantly decreased expression during all 7 days with PAN. For two of the secreted PGs, biglycan (P < 0.001) and perlecan (P < 0.01), the expression was initially decreased with a subsequent increase over time (Fig. 6, B and C). At days 4-6 after PAN, NDST-1 was decreased by 52 ± 13.3% (P < 0.01, n = 9; Fig. 6D) while syndecan-1 was significantly increased by 210 ± 38.8% (P < 0.01, n = 8, days 5-7; Fig. 5C). We could not detect any significant differences or time-dependent trends for the other genes tested (glypican-1, agrin, HS3ST, and CS-6-ST3).
When further analyzing the gene expression in rat glomeruli at day 5 after PAN injection, with or without steroids, we found that the increased expression of biglycan [by 41% (+14, −12.5), P < 0.01, n = 8] and syndecan-1 [by 34% (+13.5, −12.2), P < 0.05, n = 8], as well as the decreased expression of HS3ST-1 [by 28% (+6, −5.5), P < 0.05, n = 8] seen in PAN were abolished by steroids (Fig. 7). The decreased expression of versican [by 46% (+10,−8.5), P < 0.05, n = 7] and the enzyme NDST [by 36.5% (+8,−7), P < 0.05, n = 8] were not reversed by the steroid treatment, the NDST expression was in fact even more decreased [by 61.5% (+6.5, −5.5), P < 0.01, n = 8]. We could not detect any significant difference for agrin, perlecan, syndecan-4, glypican, or decorin at this time point.
The glomerular expression of syndecan-4 (Fig. 5B) was decreased for days 4-7 after PAN treatment. Syndecan-1 (Fig. 5D) was gradually increased from days 5 to 7 after administration of PAN, confirming the RNA data. Due to low availability of rat antibodies, no Western blots were run for the other PGs analyzed above.
Negatively charged sulfate-containing molecules are of importance for the charge selectivity and morphology of the glomerular barrier. In the present study, we explored a large number of PG core proteins and enzymes involved in GAG biosynthesis in differentiated podocytes as well as in isolated glomeruli. The nephrosis-inducing agent PAN alters the expression of a few PGs and has a severe effect on different enzymes involved in GAG synthesis. Different PGs have different functions and relevance for maintaining the permselective properties of the filtration barrier. In vivo, PAN treatment altered the level of expression for different PGs over time (Figs. 5 and 6). For some genes, a decrease in expression was seen before any physiological effects were detected. It could be that the decrease in PGs is an initial step that will affect anionic charge in the barrier, and it could also be partly responsible for the altered podocyte morphology since the presence of PGs is important for cell signaling, function, and morphology. One should also have to consider the fact that these molecules are to some extent expressed by other cell types in the glomeruli (3, 31, 37). The effects seen in vivo therefore could be a result of compensatory up- or downregulation by mesangial or endothelial cells. Treatment with steroids reversed the PAN effect for two PGs and one enzyme. In most cases, however, the steroid treatment increased the effects seen after PAN and we could not detect any beneficial effects of steroids on the physiological parameters such as GFR and proteinuria.
After stimulation with PAN for 2 days, we could see a downregulation of versican and perlecan in both podocytes and isolated rat glomeruli. Versican belongs to the family of hyaluronan-binding PGs that is collectively termed “hyalectans” that are extracellular matrix components (19, 25). Pericellular matrix expansion involves the interaction of versican with several binding proteins and can be stimulated by growth factors such as PDGF and TGF-β1 (43). It has previously been shown that versican is expressed by glomerular endothelial cells and that the synthesis is sensitive to PAN treatment (3). Perlecan is together with agrin the predominant basement membrane PG in the glomerular basement membrane (11). Perlecan mediates cell attachment (podocyte-GBM anchorage via α3β1-integrin and dystroglycans) and it binds to both cytokines and growth factors (32). A decrease in GBM heparan sulfate has been reported to be associated with proteinuria (5), and a recent publication suggested GAG chains of perlecan to be involved (26). Interestingly, PAN treatment decreased the podocyte expression and activity of three enzymes important for the GAG chain modifications. EXT-1 is a member of the extoses which are glucosyltransferases that add alternating GlcNAc and GluA residues to the growing heparan sulfate chain. N-deacetylase N-sulfotransferase (NDST) is an enzyme essential for the heparan sulfate polymer alteration and involved in an early step in the GAG modification (36). In rats, the downregulated synthesis of NDST correlated with the time point of decreased GFR and induction of proteinuria (days 4-6). Nakayama et al. (28) previously reported a downregulation of this enzyme in both rat tissue (after 10 days) and cultured rat glomerular epithelial cells (24 h) after PAN treatment. Regarding the chondroitin sulfate biosynthesis PAN downregulated the 4-O-sulfotransferase in podocytes, an effect previously described in PAN-treated glomerular endothelial cells (3). In vivo PAN treatment decreased the expression of HS-3-sulfotransferase, an enzyme that adds O-sulfate to glucoronic acid or iduronate. The PAN-induced downregulations reported would probably affect the anionic charge, structure, and function of the GAG chains.
The membrane-bound PGs analyzed in this study, syndecan and glypican, are involved in cell-cell adhesion, binding of growth factors (such as FGF), and maintenance of epithelial morphology (2, 8). The presence of these molecules in podocytes has previously been described in rat tissue by in situ hybridization (31). Their core structure and GAG chains may contribute to the anchorage of the podocyte foot processes to the GBM since syndecan-1 is known to polarize to the basolateral surface of cultured epithelial cells (33). Moreover, it has recently been described that syndecan-4 interacts with α-actinin in a β-integrin-independent manner (10). The modification of the GAG chains described above could result in disrupted cell-basal membrane anchorage (1). In vivo, the expression of syndecan-4 was decreased from day 1 to 7. For syndecan-1, an upregulation at days 5-7 was detected, i.e., an increase when proteinuria and reduction of GFR had occurred. Steroid treatment resulted in a syndecan-1 expression similar to control levels.
Decreased concentrations of anionic sites and heparan sulfate PG (HSPG) in GBM have been demonstrated in both experimental (27) and different human glomerular diseases such as diabetic (17), membranous, and minimal change nephropathy (13). Our results confirm and extend previous data demonstrating a reduction of sulfate incorporation in glomeruli during the early onset of PAN-induced nephrosis (12), as well as in cell cultures of epithelial (9, 22) and endothelial cells (3, 39). After 48-h incubation with PAN, we could see an effect both in amount of large metabolic labeled PG and in the length of the GAG chains attached. The time period is in line with earlier reports describing morphological effects of PAN on the podocyte foot processes in vivo. Inokuchi et al. (18) saw a shortened length of foot processes after 2 days of PAN treatment (by scanning electron microscopy). Retraction of foot processes gives flat or cuboidal podocytes which evenly cover the entire GBM. Because this change in conformation has been attributed to a reduction of surface charge, it is necessary to consider the disposition and composition of the podocyte cell membrane. The altered gene expression seen on both core proteins and enzymes will have effects on the composition of the GBM, cell-GBM contact as well as the pericellular matrix, and may cause proteinuria.
In conclusion, we present a description of the production of negatively charged PGs by highly differentiated podocytes in vitro as well as in vivo in the rat. We show that PG and GAG synthesis is sensitive to PAN treatment, resulting in a decreased amount of negative charge and stabilizing structures in the glomerular filtration barrier.
This study was supported by Swedish Research Council Grants 9898 and 14764, the Inga-Britt and Arne Lundberg Research Foundation, the Wilhelm and Martina Lundgren Research Foundation, and Sahlgrenska University Hospital Grant LUA-S11733.
The technical assistance by E. Roos is gratefully acknowledged.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2006 the American Physiological Society