Glomerular capillary filtration barrier characteristics are determined in part by the slit-pore junctions of glomerular podocytes. Protein tyrosine phosphatase receptor-O (PTPro) is a transmembrane protein expressed on the apical surface of podocyte foot processes. Tyrosine phosphorylation of podocyte proteins including nephrin may control the filtration barrier. To determine whether PTPro activity is required to maintain glomerular macromolecular permeability, albumin permeability (Palb) was studied after incubation of glomeruli from normal animals with a series of monoclonal (mAb) and polyclonal antibodies. Reagents included mAbs to rabbit and rat PTPro and polyclonal rabbit immune IgG to rat PTPro. mAb 4C3, specific to the amino acid core of PTPro, decreased its phosphatase activity and increased Palb of rabbit glomeruli in a time- and concentration-dependent manner. In contrast, mAb P8E7 did not diminish phosphatase activity and did not alter Palb. Preincubation of 4C3 with PTPro extracellular domain fusion protein blocked glomerular binding and abolished permeability activity. In parallel experiments, Palb of rat glomeruli was increased by two mAbs (1B4 and 1D1) or by polyclonal anti-rat PTPro. We conclude that PTPro interaction with specific antibodies acutely increases Palb. The identity of the normal ligand for PTPro and of its substrate, as well as the mechanism by which phosphatase activity of this receptor affects the filtration barrier, remain to be determined.
- slit-pore junction
- filtration barrier
protein tyrosine phosphatase receptor type O (PTPro), originally called glomerular epithelial protein 1 (GLEPP1), is a receptor-like membrane protein tyrosine phosphatase located on podocyte foot processes and the apical cell surface in rabbits, rats, mice, and humans (9, 10, 25, 29, 31). Expression of PRPro has been demonstrated in developing glomeruli and is lost in several models of glomerular injury (9, 10, 21, 32, 34). Mice deficient in PTPro have abnormal podocyte shapes with shortened foot processes and develop hypertension and decreased GFR after uninephrectomy (30). The podocyte is essential to maintaining the glomerular filtration barrier as evidenced by the finding that genetic defects in junctional proteins, cytoskeletal elements, or in a calcium channel in podocytes lead to proteinuria and progressive glomerular injury in humans and in animal models (33).
Tyrosine phosphorylation plays a major role in cell signaling, control of paracellular permeability, and remodeling of the actin cytoskeleton (14, 23, 26). Specifically, injection with protamine causes increased tyrosine phosphorylation of ZO-1 in the podocyte (11), the ZO-1, CD2AP, and CASK complex with nephrin (12), and paracellular permeability of monolayer cultured MDCK cells is increased by inhibition of tyrosine phosphatase (14). Tyrosine phosphorylation of nephrin by Src kinase alters its association with other slit-pore proteins while the nephrin ectodomain controls Src kinase activation and actin polymerization (28). Similarly, the interaction between nephrin and phosphoinositide 3-kinase appears to be important in nephrin-mediated actin cytoskeletal rearrangement (35). These and other data support the model that homophilic interactions of the nephrin ectodomain result in phosphorylation of specific nephrin cytoplasmic domain tyrosine residues, recruitment of the adapter protein Nck, and consequent assembly of actin filaments (8, 13, 28). Events occurring at the slit-junction may well be essential to the control of the filtration barrier.
We postulated that PTPro activity may play a role in determining glomerular macromolecular permeability. Its location on the apical surface of the podocyte and its role in tyrosine phosphorylation make it an attractive candidate. We investigated the effect of antibodies against the extracellular domain (ECD) of PTPro on albumin permeability (Palb) using isolated glomeruli, (18). Certain species-specific monoclonal (mAb) and polyclonal antibodies increased Palb in both rat and rabbit glomeruli. One mAb that increased Palb also decreased phosphatase activity while another had no effect on either phenomenon. We conclude that the interaction of certain antibodies against ECD epitopes of PTPro acutely increases glomerular macromolecular permeability, possibly through inactivation of phosphatase activity. Thus tonic activity of PTPro appears to play a role in maintaining normal function of the glomerular filtration barrier.
MATERIALS AND METHODS
Production and characterization of the rabbit PTPro reagents.
mAbs to PTPro (4C3 and P8E7) as well as to podocalyxin (5F7) were produced following immunization of mice with isolated rabbit glomeruli (25). 4C3 appears to bind to the amino acid core of the extracellular domain of PTPro since it binds to the denatured PTPro molecule under reducing conditions and to the nonglycosylated form of the PTPro molecule expressed by Escherichia coli as a fusion protein and in the cDNA expression system. Cloning data indicate that the 4C3 binding epitope is within fibronectin repeat 3. In contrast, P8E7 mAb does not recognize PTPro under reducing conditions on Western blotting, and it does not recognize the PTPro extracellular domain in bacterial expression systems. We have interpreted these findings as evidence that P8E7 recognizes a conformational epitope on the PTPro extracellular domain. The identity of the epitope to which P8E7 binds is not known. A control mAb (BB5) was produced by immunizing mice with human glomerular basement membrane; this antibody does not bind to either rabbit or rat tissues. The rabbit PTPro extracellular glutathione-S-transferase (GST) fusion protein was produced as previously described (25).
Animal care was in accordance with National Institutes of Health guidelines, and experimental protocols were approved by the Institutional Animal Care and Use Committee at MCW and the University of Michigan.
Cloning and sequencing of rat PTPro and production of rat PTPro extracellular domain GST fusion protein.
Rat PTPro cDNA was isolated using standard cloning techniques from a rat kidney cDNA library. The full-length rat PTPro sequence was obtained. A fusion protein corresponding to fibronectin repeats 2–5 was prepared by PCR and cloned into a vector for expression of the GST fusion protein. The fusion protein was purified by glutathione affinity chromatography and used for preparation of both mAbs and polyclonal antibodies in mice.
Antibody production and purification.
For polyclonal antibody production, four rabbits were immunized using 100 μg of the affinity-purified PTPro ECD GST fusion protein. Immune rabbit IgG was purified by ammonium sulfate fractionation followed by affinity purification on a rat PTPro ECD GST Sepharose 4B column and eluted with glycine HCl, pH 2.5. For mAb production, two mouse mAbs recognizing rat PTPro extracellular domain fusion protein were generated using standard methods. These antibodies were of the IgG2a isotype. As a control, a mAb designated L11–135, which recognizes rabbit class II (DQ) but not rat class II proteins, was class switched from IgG1 to IgG2a by clonal selection using limiting dilution and an IgG2a-specific ELISA.
Characterization of antibody binding to rat and rabbit glomeruli.
Rabbit and rat kidney cryostat sections fixed with methanol were used for assaying inhibition of antibodies following incubation with species-specific extracellular domain fusion proteins. To confirm that antibodies were specific, we preincubated mAbs with fusion proteins (10-fold excess of fusion protein by weight for rabbits and equal amount of fusion protein for rats) for 30 min at 20°C. Following a blocking step using 10% human serum antibody, preparations were added at 2 μg·100 μl−1·section−1 and incubated for 30 min. Sections were then washed, and the secondary antibody (fluorescein-labeled goat anti-mouse) that had been preabsorbed with the relevant species fusion protein was added. Sections were again washed and mounted for viewing.
Immunoprecipitation and Western blotting.
Rabbit glomeruli were isolated by cold perfusion and iron embolization as previously described (25). Isolated glomeruli were suspended in Ringer buffer containing 4% BSA and 25 μg of immunopurified mAbs (P8E7, 4C3, or BB5) for 15 min at 37°C in a shaking water bath. Glomeruli were then washed three times with cold TBS to remove BSA and free antibody. The glomeruli (50,000/ml) were then extracted with 1% Triton buffer containing inhibitors (2 mM PMSF, 5 mM N-ethylmaleimide, 2 mM EDTA). The glomerular extracts were then incubated with 100 μl of suspended protein G beads (Sigma) for 16 h at 4°C on a rotor. Beads were then washed four times with TBS containing inhibitors. They were then suspended in a 100-μl volume and divided into four aliquots for subsequent assay. One aliquot of beads was boiled in SDS/reducing agent for Western blotting under reducing conditions using mAb 4C3 to confirm that an equal amount of PTPro protein was present in each immunoprecipitate. Separate aliquots were assayed for phosphatase activity using two different methods: 1) the synthetic substrate p-nitrophenylphosphate, and 2) the 32P-labeled myelin basic protein as substrate as previously described (25).
Antibody effect on phosphatase activity of PTPro.
Briefly, rabbit glomeruli were isolated following cold perfusion and iron embolization as previously described (25). Purified glomeruli were suspended in DMEM containing 100 μg/ml of affinity-purified mouse IgG1 mAb (4C3, P8E7, 5F7, and BB5) and incubated at 37°C for 10 min. Glomeruli were then extracted with Triton X-100-containing Tris buffer. The glomerular extracts were then incubated with 100 μl of suspended protein G beads (Sigma) for 16 h at 4°C on a rotor. Beads were then washed four times with TBS containing protease inhibitors. They were then suspended in a 100-μl volume, divided into aliquots, and assayed for phosphatase activity using p-nitrophenylphosphate as a substrate (25). As described above, the same extracts were also analyzed by Western blotting using mAb 4C3 to quantify the amount of PTPro protein in the immunoprecipitates to confirm that mAbs 4C3 and P8E7 precipitated the same amount of PTPro protein.
The in vitro method by which Palb of isolated glomeruli is calculated has been previously described (18). Glomeruli from normal adult New Zealand White rabbits and normal adult Sprague-Dawley rats were incubated with experimental reagents and subjected to a Palb assay.
To document that the observed variation in expansion of capillary volume in response to an albumin oncotic gradient was due to a change in the albumin reflection coefficient (σalb) and in Palb rather than to changes in other physical characteristics of the glomerulus, we performed experiments using high-molecular-weight (HMW) dextran (200 kDa) as an impermeable oncotic agent. In these experiments, glomeruli were isolated in 4% BSA medium and observed using videomicroscopy while medium was replaced with isoncotic HMW dextran solution. Capillaries of normal glomeruli show no change in isoncotic HMW dextran. Capillaries of glomeruli with decreased σalb undergo partial collapse, and there is a decrease in glomerular volume in isoncotic HMW dextran solution (18). The medium was then replaced with 1% dextran to demonstrate the influx of fluid into the capillaries. Capillaries of normal glomeruli and of those with preserved σdextran respond with an increase in glomerular volume regardless of changes in σalb. Comparable responses of control and experimental glomeruli to a dextran-dextran gradient indicate comparable exchangeable capillary volume and capillary compliance.
Experimental reagents used in Palb assay.
To test the effect and specificity of anti-PTPro antibodies on glomerular Palb, glomeruli of each species were incubated with several experimental and control reagents for periods of 10–120 min. Reagents, concentrations, and incubation times are listed in Table 1. Irrelevant antibodies used in studies of rabbit glomeruli included monoclonal BB5, an antibody to human GBM that binds to laminin and 5F7, an antibody to rabbit podocalyxin. Each of these mAbs was of isotype IgG1. BB5 did not bind to rabbit glomeruli, while 5F7 bound in a podocyte pattern. Dose-response (5–25 μg/ml) and time-response (5, 10, 25, and 60 min) studies were carried out with 4C3. mAbs P8E5 and BB5 were used at 25 μg/ml, and 5F7 was used at 11 and 55 μg/ml. Incubations were carried out for 60 or 120 min.
Reagents used for the study of anti-PTPro in rat glomeruli included polyclonal antibodies from immunized rabbits and mAbs 1B4 and 1D1. These antibodies were of the IgG2a isotype. The antibodies stained glomeruli in an epithelial cell pattern. The epitope to which each binds is unknown. Irrelevant antibodies included those contained in rabbit serum before immunization with PTPro (preimmune serum), 4C3 and L11–135. The polyclonal antibodies were used at 5 μg/ml, and the mAbs to PTPro at 100 μg/ml. Concentrations of irrelevant antibodies are indicated in Table 1. Incubations were carried out for 60 or 120 min.
Values for Palb are expressed as average ± SD, and n indicates the number of experimental animals studied, unless indicated otherwise. Comparisons among groups of animals were made using ANOVA. A P value < 0.05 was accepted as significant.
Specificity of binding of antibodies directed against the ECD of rabbit and rat glomerular PTPro.
mAb 4C3 against rabbit PTPro bound to rabbit tissue but not to rat tissue. Binding was blocked by preincubation with rabbit ECD fusion protein. The control mAb BB5 did not bind to rabbit kidney cortex sections. Rat mAb to PTPro ECD bound to rat sections; binding was inhibited by rat ECD fusion protein. These results are shown in Fig. 1.
Effect of mAb binding on PTPro phosphatase activity.
The immunoprecipitates of rabbit glomerular proteins prepared using 4C3 and P8E7 contained the same amount of the PTPro protein. However, the 4C3 immunoprecipitate showed approximately half the phosphatase activity compared with the P8E7 immunoprecipitate (Fig. 2). The average phosphatase activity in the 4C3 immunoprecipitate was 45 ± 8% of that in the P8E7 immunoprecipitate (n = 8, P < 0.01). Thus binding of mAb 4C3 to the ECD of PTPro reduces phosphatase activity.
Effect of antibodies against the ECD of PTPro on Palb of isolated rabbit glomeruli.
Incubation with 25 μg/ml of 4C3 for 10 min increased Palb, an effect seen for up to 120 min, the longest time point studied (Fig. 3). Incubation with 5 or 10 μg/ml 4C3 for up to 60 min failed to affect Palb (data not shown). Average Palb after incubation with 4C3 was 0.59 ± 0.08 (n = 7, P < 0.01 vs. control). These values represent calculations using average observed glomerular volume increase of 3.4% after 4C3 incubation and 8.4% for control glomeruli. Palb was not significantly increased by P8E7 (0.12 ± 0.25, n = 4) or by a mAb to podocalyxin (BB5, 0.13 ± 0.15, n = 9) or laminin (5F7, −0.39 ± 0.14, n = 3).
Incubation of 4C3 with rabbit PTPro ECD fusion protein before glomerular incubation blocked binding of PTPro to glomeruli and ameliorated the increase in Palb (0.26 ± 0.06, n = 4), as shown in Fig. 3. In contrast, incubation of 4C3 with rabbit podocalyxin fusion protein did not inhibit glomerular binding and did not inhibit the increase in Palb (0.66 ± 0.16, n = 2, data not shown).
Effect of polyclonal antibodies and mAbs directed against the ECD of PTPro on Palb of isolated rat glomeruli.
Anti-rat PTPro polyclonal antibody bound to rat glomeruli and increased Palb (0.56 ± 0.05, n = 2), as seen in Fig. 4, while BB5 and 4C3, which did not bind to rat glomeruli, did not increase Palb (−0.07 ± 0.10, n = 4 and −0.15 ± 0.06, n = 4, respectively, data not shown). Incubation of anti-rat PTPro IgG with rat PTPro ECD GST fusion protein prevented both glomerular binding and increase in Palb (−0.20 ± 0.20, n = 4) (Fig. 4). Rat mAbs1B4 and 1D1 each increased Palb (0.54 ± 0.15, n = 3, and 0.44 ± 0.27, n = 3, respectively, P < 0.01), while control protein L11–135 did not affect Palb (−0.08 ± 0.12, n = 3), as seen in Fig. 4. Preimmune rabbit serum had a small but statistically significant effect on Palb (0.28 ± 0.11, n = 11, P < 0.05 vs. control). This effect was not concentration dependent and was comparable in magnitude to the changes induced by incubation with complement-containing serum in prior studies of complement-mediated injury (19).
To verify that altered volumetric responses were the result of altered permeability rather than changes in exchangeable volume or in capillary compliance, we performed additional studies of rat glomeruli treated for 60 min with polyclonal anti-rat PTPro Ab. In these studies, the volumetric response to replacing medium containing BSA with that containing isoncotic HMW dextran (MW 200) was assessed using established protocols (18). Results are expressed as means ± SD for five glomeruli. Control glomeruli showed no change in volume when 4% BSA medium was replaced by dextran solution (ΔV = −0.03 ± 0.45). This finding confirms that σalb and σdextran did not differ. In contrast, antibody-treated glomeruli showed a decrease in volume (ΔV = −2.9 ± 1.6%). This volume decrease confirms that anti-rat PTPro decreased the ratio of σalb to σdextran. Subsequent replacement of 4% dextran medium with 1% dextran led to comparable volume increase in control and experimental glomeruli (ΔV = 4.6 ± 1.2 and 5.9 ± 0.8%, respectively). The equivalent glomerular volume responses of control and treated glomeruli to HMW dextran gradients is evidence that glomerular physical characteristics and σdextran are not altered by antibody treatment. Taken together, these results provide strong evidence of decreased σalb and increased Palb after treatment with anti-PTPro. Results are shown in Fig. 5.
We have demonstrated that antibodies to the ECD of PTPro increase Palb of isolated glomeruli, an effect seen with mAbs and polyclonal antibodies and in both rabbit and rat glomeruli. Effective antibodies to PTPro included mAb 4C3, which interacts with the fibronectin domain 3 of rabbit ECD and decreases phosphatase activity of PTPro, and antibodies to rat ECD, including 1B4 and 1D1. Irrelevant mAbs and mAbs that lacked glomerular binding had no effect on Palb. Fusion proteins composed of species-specific ECD interacted with the relevant antibodies and prevented glomerular binding as well as the increase in Palb. These findings support the interpretation that the effects of antibodies on the permeability barrier are specific consequences of antibody interaction with the PTPro ECD.
In contrast to the effects of the other mAb and polyclonal Abs, mAb P8E7 had no effect on Palb. P8E7, like 4C3, bound to glomeruli in a podocyte pattern and recognized PTPro protein on Western blotting under nonreducing conditions (25). However, the two mAbs differed in their specific binding to PTPro and in their effect on its phosphatase activity. Each recognized PTPro under nonreducing conditions, but under reducing conditions 4C3 antibody recognized a 205-kDa region representing PTPro while P8E7 did not. 4C3 binding was blocked by a fusion protein containing PTPro fibronectin repeat 3 while P8E7 binding was not blocked by this fusion protein (25). PTPro phosphatase activity was diminished by 4C3 compared with that seen with P8E7. Taken together, these results suggest that P8E7 recognizes a conformational epitope that is destroyed by denaturation and that fibronectin domain 3 may participate in homophilic PTPro interactions. In addition, the association between diminished phosphatase activity and increased Palb suggests that tonic PTPro activity is required to maintain the normal characteristics of the glomerular filtration barrier.
The measurement of Palb in vitro was first described in studies of permeability induced by incubation with the polycation protamine, and the relevant principles are described in detail in that report (18). Since then, Palb has been measured in a wide variety of experimental models of glomerular injury (reviewed in Ref. 20). Under the standard conditions that we have used, calculation of Palb is based on observations of glomerular volumetric responses to albumin oncotic gradients. The rationale for Palb measurements was based on prior observations that the glomerular capillary volume of intact isolated glomeruli varied directly with the oncotic composition of the medium in which they were suspended. Changes in glomerular volume with changes in bathing medium result from capillary expansion or collapse and is directly proportional to the effective oncotic gradient (17). The oncotic gradient, in turn, represents the product of the chemical oncotic gradient, ΔΠ, and the reflection coefficient, σ, of the solute employed. In normal glomeruli σalb is ∼1. With injury to the filtration barrier, decreased σalb decreases the effective oncotic gradient, diminishes capillary expansion, and results in a smaller steady-state capillary and total glomerular volume.
The finding that Palb is increased by an antibody that inhibits PTPro phosphatase while PTPro knockout animals fail to manifest proteinuria is in agreement with our findings in several models of renal disease (1, 7, 15, 16, 19). We have found that increased Palb is more sensitive and permits earlier identification of glomerular injury than does increased urinary protein in studies of antibodies to anti-Fx1A (19), of antibodies to β1-integrin (1), and of Dahl salt-sensitive rats (7) and of rats transgenic for the RF-2 gene of fawn-hooded hypertensive rats (16). Recent studies in PAN nephrosis also show increased Palb before detectable proteinuria (15). The current findings of increased Palb during experimental treatment are consistent with the concept that increased Palb provides a sensitive and early measurement of glomerular dysfunction.
PTPro has a large extracellular domain containing eight fibronectin type III-like repeats, a hydrophobic transmembrane segment, and a single PTPase domain (10, 25, 29). The knockout of PTPro caused changes in morphology of primary podocyte processes and decreased the total length of interdigitating tertiary processes and of the slit-diaphragm between them (30). PTPro (−/−) mice exhibit hypertension and elevated serum creatinine only after uninephrectomy or subtotal nephrectomy. These findings suggest that PTPro plays a role in the regulation of foot processes. They are, however, in apparent contradiction to the current findings that an antibody-induced decrease in phosphatase activity increases Palb. It is possible that the glomeruli of PTPro (−/−) mice have abnormal Palb without proteinuria, a finding that could result from tubular reabsorption of filtered protein; this would be analogous to our findings in other models. Alternatively, other phosphatases may serve a redundant compensatory function in maintaining the protein permeability barrier when PTPro is absent throughout glomerular development. The increased expression of podocyte vimentin (30) is an example of changes in other glomerular proteins in PTPro (−/−) mice. This reasoning is supported by the finding that suppression of individual protein tyrosine phosphatase receptors does not necessarily lead to profound abnormalities in neuronal development (27). Several members of type III receptor protein tyrosine phosphatases may serve cooperative as well as competitive functions. Similarly, multiple knockouts may be required to fully define the role of PTPro in podocyte differentiation and in maintenance of the permeability barrier.
Recent studies suggest that the extracellular domain of PTPro is homophilic (3, 24). Its signaling functions are incompletely understood and may be complex. For example, PTPro controls the sensitivity of ephrin receptors in retinal cells and determines their topographical growth (22). In these studies, PTPro dephosphorylates a phosphotyrosine residue conserved in the juxtamembrane region of ephrin receptors, which is required for the activation and signal transmission by these receptors. PTPro is also expressed in hematopoietic stem cells and plays a role in stem cell adhesion and in homophilic cell-cell interactions (2). It controls B cell-receptor signaling and modulates tonic B cell receptor signaling (6).
A potential substrate for PTPro in neurons, neuronal pentraxin with chromo domain (NPCD), has been identified. This protein associates with PTPro in neurons and interacts with and may serve as a substrate for PTPro kinase activity (4, 5). NPCD is also expressed in glomeruli and colocalizes with PTPro. Its function in determining the outgrowth of cytoplasmic extensions from neurons may be analogous to its function in determining the normal morphology of podocyte foot processes. At present, the identity of the normal substrate for PTPro catalytic activity in glomeruli is not known, and the means by which this activity serves to maintain the filtration barrier in mature glomeruli remains to be determined.
Our results confirm that antibodies directed to PTPro cause increased glomerular protein permeability. In particular, antibodies to a specific region of the extracellular domain result in impairment of the permeability barrier. The possible role of phosphatase activity of PTPro in maintaining the glomerular protein permeability barrier is intriguing and worthy of further study. Results of the present studies support our hypothesis that tonic activity of PTPro is required to maintain the integrity of the filtration barrier. The study of isolated glomeruli provides the opportunity to define the effects of alterations in individual glomerular receptors and their signaling function on the filtration barrier. Our results emphasize the complexity of glomerular function and the potential role of altered tyrosine phosphorylation in its control.
This work was supported by National Institute of Diabetes and Digestive and Kidney Diseases Grants R01-DK-39255 and R01-DK-46073 (to J. E. Wiggins) and R01-DK-43752 (V. J. Savin).
The results were presented, in part, at the annual meeting of the American Society of Nephrology held on November 3–6, 1996, in New Orleans, LA.
- Copyright © 2009 the American Physiological Society