Urine electrolyte, mineral, and protein excretion in NHERF-2 and NHERF-1 null mice

Rochelle Cunningham, Ali Esmaili, Eric Brown, Rajat S. Biswas, Rakhilya Murtazina, Mark Donowitz, Henry B. Dijkman, Johan van der Vlag, Boris M. Hogema, Hugo R. De Jonge, Shirish Shenolikar, James B. Wade, Edward J. Weinman

Abstract

The adaptor proteins sodium/hydrogen exchanger regulatory factor (NHERF)-1 and NHERF-2 have overlapping tissue distribution in renal cells and overlapping specificity in their binding to renal transporters and other proteins. To compare the kidney-specific differences in the function of these adaptor proteins, NHERF-1 and NHERF-2 null mice were compared with wild-type control mice. In NHERF-2 null mice, the renal proximal tubule abundance and distribution of NHERF-1 and NHERF-3 were not different from those in wild-type animals. The glomerular expression of podocalyxin and ZO-1 also did not differ. NHERF-1 null mice had increased urinary excretion of phosphate, calcium, and uric acid compared with wild-type control and NHERF-2 null mice. Because of the association between NHERF-2 and podocalyxin in glomeruli and ClC-5 in the renal proximal tubule, the urinary excretion of protein was determined. There were no differences in the urinary excretion of protein or low-molecular-weight proteins between wild-type control, NHERF-1−/−, and NHERF-2−/− mice. These studies indicate that the increased urinary excretion of phosphate and uric acid are specific to NHERF-1 null mice and highlight the fact that predictions about the role of adaptor proteins such as the NHERF proteins obtained from studies of model cell systems must be confirmed in whole animals.

  • mouse kidney
  • PDZ proteins
  • sodium/hydrogen exchanger regulatory factor

the proteins of the sodium/hydrogen exchanger regulatory factor (NHERF) family of PDZ adaptors (NHERF-1, -2, -3, and -4) are highly expressed in epithelial tissues, where they regulate their target proteins such as transporters and ion channels, receptors, signaling proteins, and cytoskeletal elements (8, 15, 16). The NHERF proteins are capable of binding to one another, and there also is considerable overlap in the binding of the individual NHERF proteins to target proteins including renal brush-border membrane sodium/hydrogen exchanger 3(NHE3) and sodium-dependent phosphate transporter 2a(Npt2a) (15). Study of the physiological role of NHERF proteins in model cell culture systems has not always reproduced the more complex biology seen in native tissues. An alternate approach to studying the physiological role of the NHERF proteins has been the development of genetically altered mice in which the individual NHERF genes have been inactivated (1, 5, 8). The first to be developed was the NHERF-1−/− mouse line. These animals demonstrated defective cAMP-mediated regulation of NHE3 and defective targeting of Npt2a resulting in increased urinary excretion of phosphate (8, 13). In addition, we reported (2, 17) that NHERF-1−/− mice have increased excretion of uric acid associated with decreased apical membrane abundance of URAT1. More recently, the NHERF-2 null mouse line was developed, but the phenotype of these animals has not yet been compared with NHERF-1 null mice (1).

In the present experiments, we have studied the cellular location and expression of NHERF-1 and NHERF-3 in NHERF-2−/− mice. In addition, we have compared the urinary excretion of electrolytes and minerals in NHERF-2 null mice to that in wild-type control and NHERF-1 null mice. Finally, by virtue of the interaction between NHERF-2 and podocalyxin in podocytes of the renal glomerulus and with ClC-5 in the renal proximal tubule, there are theoretical reasons to predict that NHERF-2 null mice would have proteinuria (4, 10). Accordingly, we have compared the urinary excretion of protein in wild-type mice with that in both NHERF-1 and NHERF-2 null mice.

MATERIALS AND METHODS

NHERF-1−/− mice bred into a C57BL/6 background for six generations and parental wild-type inbred control C57BL/6 mice were used in the present experiments (8). NHERF-2−/− mice in a FVB/n genetic background developed by Hogema, de Jonge, and coworkers (1) were bred into a C57BL/6 background for six generations to permit comparison to the NHERF-1 null animals. The Institutional Animal Care and Use Committee at the University of Maryland School of Medicine approved all animal protocols and procedures.

Throughout the study period, the animals were fed a normal rodent chow diet and had free access to water. Confocal microscopy was performed with paraffin-fixed renal tissue as previously described (14). For immunocytochemistry, appropriate species-specific secondary antibodies coupled to Alexa 488 or 568 dyes (Molecular Probes) were used at 1:100 dilution. Electron microscopic studies were performed with small pieces of renal cortex fixed in 2.5% glutaraldehyde dissolved in 0.1 M sodium cacodylate buffer, pH 7.4, for 7 days at 4°C. After being washed in cacodylate buffer, the tissue was postfixed in Palade-buffered 2% OsO4 for 1 h, dehydrated, and embedded in Epon 812, Luft's procedure (Merck, Darmstadt, Germany). Ultrathin sections (90 nm) were stained with 4% uranyl acetate for 45 min and subsequently with lead citrate for 5 min at room temperature. Brush-border membranes were isolated from mouse kidneys by previously described methods (13). The proteins were resolved with 10% SDS-PAGE and transferred to nitrocellulose, and Western immunoblotting was used to determine the abundance of NHERF-1, NHERF-2, NHERF-3, NHE3, Npt2a, and ezrin (11, 13). The NHERF-1 and NHERF-2 antibodies were polyclonal antibodies previously characterized by our laboratories (11). NHERF-3 antibodies were kindly provided by Biber and coworkers (3). Antibodies to NHE3 (Chemicon International) and ezrin (Abcam) were obtained from commercial sources. Immunocomplexes were detected with enhanced chemiluminescence (ECL) (Amersham, Arlington Heights, IL). Densitometry measurements of the Ponceau S-stained transfers and the intensity of ezrin immunoblots were used to ensure equal loading of the lanes.

For metabolic studies, 8- to 10-wk-old animals were individually housed in metabolic cages for 48 h. After a 24-h period of acclimatization, 24-h urine collections were obtained. A separate group of animals were killed to obtain blood samples. Concentrations of calcium and phosphate in serum and urine concentrations of creatinine, calcium, uric acid, phosphate, sodium, potassium, and chloride were determined with an autoanalyzer. Mouse parathyroid hormone (PTH) concentrations were assayed by ELISA.

Urine proteins were determined by the method of Lowry et al. (6). Data are expressed as means ± SE and compared with Student's t-test. For the longer-term studies of protein excretion in the urine of NHERF-1−/− mice, 8- to 54-wk-old animals were housed in metabolic cages as above. In some cases, individual animals were studied at more than one age. Eighteen percent SDS-PAGE polyacrylamide Tris-tricine-urea gel electrophoresis was used to resolve low-molecular-weight proteins in the urine.

RESULTS

Figure 1 illustrates the glomerular expression of NHERF-2 in wild-type mice and the total absence of staining in NHERF-2−/− mice. In contrast, the abundance and distribution of podocalyxin were normal in NHERF-2−/− mice (Fig. 1). To confirm the structural integrity of the glomeruli, we also stained the tissue with an anti-ZO-1 antibody (Fig. 2). No differences in ZO-1 staining were seen between wild-type control mice and NHERF-2 null mice. Finally, we examined the glomeruli by electron microscopy but did not detect significant abnormalities in NHERF-2−/− mice (Fig. 3). When considered together, these findings indicate that the overall structure of the glomerulus is maintained in the absence of NHERF-2. In prior studies in NHERF-1 null mice, we demonstrated (11) that the abundance and cellular distribution of NHERF-2 and NHERF-3 did not differ from those in wild-type controls. As shown in Fig. 4, there were no differences in NHERF-1 and NHERF-3 in NHERF-2−/− mice compared with control mice. In addition, the cellular localization of NHE3 and Npt2a did not differ between wild-type and NHERF-2 null mice (not shown). Figure 5 is a representative composite of two near-identical immunoblots of renal brush-border membrane from wild-type and NHERF-2−/− animals. There were no differences in the relative abundance of NHERF-1, NHERF-3, NHE3, Npt2a, and ezrin between wild-type and NHERF-2 null mice.

Fig. 1.

Representative confocal microscopy images of renal glomeruli stained with an antibody to sodium/hydrogen exchanger regulatory factor-2(NHERF-2) (A and B) and podocalyxin (C and D) in wild-type (WT; A and C) and NHERF-2−/− (B and D) mice.

Fig. 2.

Higher-resolution representative confocal microscopy images of renal glomeruli stained with an antibody to NHERF-2 (A and B) and ZO-1 (C and D) in WT (A and C) and NHERF-2−/− (B and D) mice. Scale bars, 7.4 μm.

Fig. 3.

Representative transmission electron microscopic images of the ultrastructure of WT (A and C) and NHERF-2−/− (B and D) mice. A and B: overall glomerular ultrastructure. C and D: podocyte foot processes and glomerular endothelium.

Fig. 4.

Representative confocal microscopy images of renal proximal convoluted tubules stained with an antibody to NHERF-1 (A and B) and NHERF-3 (C and D) in WT (A and C) and NHERF-2−/− (B and D) mice. Scale bars, 9 μm.

Fig. 5.

Composite immunoblot of renal brush-border membranes from WT and NHERF-2−/− mice stained for NHERF-1, NHERF-2, NHERF-3, sodium/hydrogen exchanger 3(NHE3), sodium-dependent phosphate transporter 2a(Npt2a), and ezrin.

To determine whether there are functional differences between NHERF-1−/− and NHERF-2−/− mice, selected blood and urine electrolytes and minerals were measured (Table 1). Serum concentrations of calcium averaged 9.3 ± 0.3 and 9.1 ± 0.1 mg/dl [P = not significant (NS)] in wild-type control (n = 6) and NHERF-2−/− (n = 6) animals, respectively. Serum concentrations of phosphate averaged 9.5 ± 0.6 mg/dl in wild-type control animals (n = 5) and 9.1 ± 0.6 mg/dl (P = NS) in NHERF-2−/− animals (n = 5). Because of the limited amount of sample available from a single animal and the volume required for the PTH assay, it was necessary to pool samples. Serum PTH concentrations averaged 73 ± 4 pg/ml in wild-type mice (pooled from 6 animals) and 81 ± 6 pg/ml in NHERF-2−/− mice (pooled from 5 animals). Compared with wild-type control mice, 8-wk-old NHERF-1 null mice had increased urine-to-creatinine ratios of phosphate, calcium, and uric acid, results that confirm prior observations (17). The urinary excretion of sodium, potassium, and chloride did not differ. In contrast, the urinary excretion of sodium, potassium, chloride, phosphate, calcium, and uric acid in NHERF-2 null mice did not differ from that in wild-type control mice.

View this table:
Table 1.

Urinary excretion of sodium, potassium, chloride, calcium, phosphate, and uric acid in male and female wild-type, NHERF-1−/−, and NHERF-2−/− mice

Given the possibility that NHERF-2 null mice might have proteinuria as a consequence of NHERF-2 binding podocalyxin in the glomerulus and/or the ClC-5 chloride channel in the proximal tubule, we examined the urine protein-to-creatinine ratios in 8- to 10-wk-old male and female wild-type, NHERF-1 null, and NHERF-2 null mice (4, 10). As summarized in Table 2, there were no differences between the groups. Figure 6 indicates that there were also no differences in the excretion of low-molecular-weight proteins among these three groups of animals.

Fig. 6.

Representative urine protein electrophoresis from WT, NHERF-1−/−, and NHERF-2−/− mice.

View this table:
Table 2.

Urine protein-to-creatinine ratios in wild-type, NHERF-1 null, and NHERF-2 null mice

NHERF-1 null mice develop time-dependent deposition of calcium in the interstitium of kidney, particularly in the papilla (17). To determine the long-term consequences of the absence of NHERF-1, we measured the urine protein-to-creatinine ratio in male and female mice at ages 8–54 wk (Fig. 7). There was a time-dependent decrease in the urine protein-to-creatinine ratio in both male and female mice but no differences between wild-type and NHERF-1 null mice.

Fig. 7.

Urine protein-to-creatinine ratios plotted as a function of age in wild-type mice (○) and NHERF-1−/− (□) male (A) and female (B) mice. Trend lines for WT (solid lines) and NHERF-1−/− (dashed lines) animals are shown.

DISCUSSION

The NHERF proteins contain multiple protein interactive PDZ domains and in the case of NHERF-1 and NHERF-2, a COOH-terminal domain that binds the cytoskeletal proteins ezrin, radixin, moesin, and merlin. To date, >40 proteins have been identified that bind to these adaptor proteins (9). Moreover, the NHERF proteins (NHERF-1, NHERF-2, NHERF-3, and NHERF-4) homodimerize and heterodimerize and are proposed to array an apical membrane/subapical mesh that binds and organizes targets into protein complexes (11). Since individual target proteins bind to more than one member of the NHERF family, a major focus of investigations is to determine the in vivo biological role of each NHERF protein. Study of model cell systems has provided valuable information on the interactions between the NHERF proteins and their targets, but predictions from such studies have not always been borne out in intact animals. For example, both NHERF-1 and NHERF-2 bind to NHE3 and are equivalent in mediating cAMP inhibition of this transporter when expressed in fibroblasts (18). In intact renal tissue, however, only NHERF-1 but not NHERF-2 mediates cAMP-associated inhibition of NHE3, results that highlight the complexity of the system and the need to confirm findings from cellular studies in model systems in intact animals (13). In the present experiments, we used the newly developed NHERF-2 null mouse line and compared selected aspects of renal structure and physiology to wild-type and NHERF-1 null mice (1).

As determined by confocal microscopy and by immunoblots of isolated brush-border membranes, the relative abundance and cellular localization of NHERF-1 and NHERF-3 were not altered in NHERF-2 null mice. The abundance and distribution of NHERF-2 and NHERF-3 were previously found to be unaffected by the absence of NHERF-1 (11). The above studies, however, do not rule out subtle differences in the subcellular distribution of these proteins. Nonetheless, it is reasonable to conclude that the differences between the animals were not the consequence of large reciprocating changes in the expression of one or more NHERF isoforms when one is developmentally silenced.

NHERF-1, NHERF-2, and NHERF-3 interact with Npt2a, the major sodium-dependent phosphate transporter in the proximal convoluted tubule of the kidney (3). NHERF-1 null mice manifest hypophosphatemia, an increase in the urinary excretion of phosphate, and a decrease in the renal proximal tubule apical membrane abundance of Npt2a (8). In contrast, the serum concentration of phosphate, the urinary excretion of phosphate, and the renal proximal tubule abundance of Npt2a in NHERF-2 null mice were not different from those in wild-type control mice. NHERF-1 null mice have increased urinary excretion of calcium and uric acid, whereas NHERF-2 null mice have rates of excretion comparable to those in wild-type control mice. Thus the abnormal urine profile for excretion of phosphate, calcium, and uric acid is distinct between NHERF-1 and NHERF-2 null animals and likely reflects the specificity of the interaction between NHERF-1 and Npt2a and selected transporters in the kidney. It is worth emphasizing that the urine samples were obtained while the mice were ingesting a normal rodent chow diet, and it remains possible that differences may become evident when the mice are challenged or stressed. Nonetheless, our present results are consistent with other experiments that indicate that while NHERF-1 and NHERF-2 have a number of binding proteins in common, they have distinctive effects on specific transporters in the kidney (16).

In glomerular podocytes, NHERF-2 binds to the sialoprotein podocalyxin, forming a protein complex that includes ezrin and an interaction with the actin cytoskeleton (10). Perfusion of kidneys with protamine sulfate or sialidase or induction of puromicin aminonucleoside nephrosis disrupts this complex and is associated with loss of foot processes, structural reorganization of the podocyte, and the prediction that the filtration barrier for protein would be compromised. In proximal tubule cells, the endocytic complex that regulates albumin reabsorption in renal proximal tubule cells includes NHE3, which interacts with both NHERF-1 and NHERF-2, and ClC-5, which binds to NHERF-2 (4, 12). Recent experiments in OK cells, a proximal tubule-like cell line, indicates that silencing of the NHERF-2 gene decreases cell surface expression of ClC-5 and the endocytosis of albumin while silencing of NHERF-1 increases cell surface expression of ClC-5 and albumin uptake (4). In the present experiments, we found no differences in the glomerular ultrastructure and no differences in the abundance and distribution of podocalyxin and ZO-1, a glomerular marker, between wild-type and NHERF-2 null animals, indicating that the structural integrity of the glomeruli was maintained in the absence of NHERF-2. Moreover, the urine protein-to-creatinine ratio in 8- to 10-wk-old male and female NHERF-2 null mice was not significantly different from that in wild-type control or NHERF-1 null mice. The urinary pattern of excretion of low-molecular-weight proteins also did not differ from that in control or NHERF-1 animals. Our failure to detect abnormalities in the urinary excretion of albumin in NHERF-2−/− mice may be explained, at least in part, by the recent studies by Russo and coworkers (7). These investigators provided evidence for very high rates of albumin filtration across the glomerulus associated with postglomerular retrieval pathways that included a low-capacity endocytic pathway that results in degradation of reabsorbed albumin and a high-capacity transcytotic pathway that reclaims filtered albumin intact. Thus it remains possible that our failure to detect abnormalities in protein excretion is the result of the presence of an intact NHERF-2-independent transcytotic pathway that compensates for impairments in the NHERF-2-dependent endocytic/degradation pathway.

Since NHERF-1 null mice have multiple defects in proximal tubule transport and develop calcium deposition in the interstitium of the kidney, we also examined the urine protein-to-creatinine ratio over a more extended time period from 8 to 54 wk of age (17). In wild-type and NHERF-1−/− mice, there is a slight age-related decline in the urine protein-to-creatinine ratio in both male and female animals but no differences between NHERF-1−/− mice and wild-type control mice of either sex. The normal rates of protein excretion and the lack of low-molecular-weight proteinuria in the NHERF-1−/− mice also suggest that the changes in the excretion of phosphate and uric acid are specific and are not the result of global alterations of proximal tubule function.

In summary, these experiments indicate that the renal cellular distribution of NHERF-1 and NHERF-3 were not altered in NHERF-2 null mice. The pattern of urinary excretion of electrolytes and minerals in NHERF-2 null mice is similar to that in wild-type control mice, and, in contrast to NHERF-1−/− mice, NHERF-2−/− mice do not excrete increased amounts of phosphate, calcium, or uric acid. Moreover, despite indications that NHERF-2 might be critical in the formation of the protein filtration barrier in glomeruli and the reabsorption of protein in the renal proximal tubule, NHERF-2 null mice do not exhibit increases in the urinary excretion of protein or low-molecular-weight proteinuria. These experiments extend the phenotypic expression of these interesting models and indicate the presence of significant differences in the biological roles of these two adaptor proteins.

GRANTS

These studies were supported by grants from the National Institutes of Health (NIH) (DK-55881 to E. J. Weinman and S. Shenolikar) and the Research Service of the Department of Veterans Affairs (E. J. Weinman). R. Cunningham is a recipient of a Minority Career Development Award from NIH and a Harold Amos Faculty Development Award from the Robert Wood Johnson Foundation.

Acknowledgments

The authors acknowledge the expert technical assistance of Fengying Wang and Deborah Steplock.

Footnotes

  • 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.

REFERENCES

View Abstract