This article requires a subscription to view the full text. If you have a subscription you may use the login form below to view the article. Access to this article can also be purchased.
Recent studies indicate that pendrin, an apical Cl−/HCO3− exchanger, mediates chloride reabsorption in the connecting tubule and the cortical collecting duct and therefore is involved in extracellular fluid volume regulation. The purpose of this study was to test whether pendrin is regulated in vivo primarily by factors that are associated with changes in renal chloride transport, by aldosterone, or by the combination of both determinants. For achievement of this goal, pendrin protein abundance was studied by semiquantitative immunoblotting in different mouse models with altered aldosterone secretion or tubular chloride transport, including NaCl loading, hydrochlorothiazide administration, NaCl co-transporter knockout mice, and mice with Liddle’s mutation. The parallel regulation of the aldosterone-regulated epithelial sodium channel (ENaC) was examined as a control for biologic effects of aldosterone. Major changes in pendrin protein expression were found in experimental models that are associated with altered renal chloride transport, whereas no significant changes were detected in pendrin protein abundance in models with altered aldosterone secretion. Moreover, in response to hydrochlorothiazide administration, pendrin was downregulated despite a marked secondary hyperaldosteronism. In contrast, α-ENaC was markedly upregulated, and the molecular weight of a large fraction of γ-ENaC subunits was shifted from 85 to 70 kD, consistent with previous results from rat models with elevated plasma aldosterone levels. These results suggest that factors that are associated with changes in distal chloride delivery govern pendrin expression in the connecting tubule and cortical collecting duct.
Long-term BP regulation is remarkably linked to renal salt excretory function. Normally, the kidney adapts urinary NaCl excretion to match exactly daily dietary NaCl intake. Among the different nephron segments, the critical importance in vascular volume and BP regulation of the aldosterone-sensitive distal nephron (ASDN), which consists of part of the distal convoluted tubule (DCT), the connecting tubule (CNT), and the collecting duct (CD), now is well established (1). In fact, during the past decade, mutations in genes that cause syndromes that are characterized by hypertension or hypotension have been identified, and most of these genes have turned out to be involved in the control of Na+ absorption in the ASDN (for review, see reference ).
It is widely assumed that the pressor effect of dietary salt depends primarily on Na+. However, several studies indicate that Cl− is necessary for Na+ to exert its pressor effect (3–7), and primary alteration in distal Cl− transport may have consequences on BP (8–10). Remarkably, in the CNT and the CD, Na+ reabsorption is not linked molecularly to Cl− transport directly. Na+ reabsorption is activated by the hormone aldosterone and is achieved through the apical epithelial sodium channel (ENaC), which is expressed by CNT cells and principal cells of CD. Cl− transport does not occur through principal cells but rather through the paracellular pathway or through the intercalated cells (11). In the B and non-A non-B type intercalated cells, pendrin, an apical anion exchanger (12–14), is believed to play a key role in BP regulation, most likely by controlling distal nephron chloride reabsorption (15,16). Pendrin is upregulated by the injection of deoxycorticosterone pivalate, a pharmacologic analogue of aldosterone (16). However, we have demonstrated that pendrin also can be regulated in response to changes in Cl− intake by an aldosterone-independent mechanism (17). In vivo relevance of both mechanisms remains unsettled, but this question is of particular importance. Therefore, the goal of our study was to examine whether pendrin is regulated in vivo primarily by a chloride-dependent mechanism, by aldosterone, or by the combination of both determinants.
To achieve this goal, we examined the protein expression of pendrin in different mouse models with disturbed NaCl balance that allow separate analysis of the effects of changes in distal chloride transport or in aldosterone secretion. We systematically examined the parallel regulation of ENaC as a control. Our data demonstrate that Cl−-dependent pendrin regulation is able to override the effects of aldosterone. These data also suggest that pendrin might be an independent modulator of NaCl transport by the ASDN.
Materials and Methods
Animal Models and Treatments
All animals used in this study were treated in full compliance with the French government animal welfare policy (agreement no. RA024647151FR).
A treated group and a control group were handled in parallel during a 6-d period. Each group consisted of seven adult C57BL/6 male mice. Treated mice were given 0.28 M NaCl in the drinking water and were fed ad libitum with standard laboratory mouse chow (AO3; Scientific Food and Engineering, Augy, France). Control mice were fed ad libitum with standard laboratory mouse chow and had free access to distilled water for the same period of time.
Mouse Model of Liddle Syndrome with Constitutive Activation of the Apical Sodium Channel ENaC.
Transgenic mice were generated by insertion of a stop codon (corresponding to residue R566 in human) into the mouse βENaC gene locus as described previously (18). Seven homozygous transgenic mice (ENaCL/L) and seven wild-type littermate controls (ENaCWT) were used for these series. All of the animals were outbred on a mixed 129/Ola-C57BL/6 genetic background. Animals were 5 to 6 mo of age and weighed 28 to 30 g.
NaCl Co-Transporter (NCC) Knockout Mice.
Mice were generated by a standard gene-targeting technique as described previously (19). Eight adult homozygous NCC-deficient male mice (NCC−/−) and eight littermate wild-type mice (NCC+/+) were used for this series. All animals were inbred on the C57BL/6 genetic background (>10 generations of backcrossing). The animals were 7 to 8 mo of age and weighed 32 to 34 g.
Acute Hydrochlorothiazide Treatment
Hydrochlorothiazide (HCTZ; Sigma-Aldrich, St. Louis, MO), a specific inhibitor of the NCC co-transporter of the DCT, was given to seven adult C57BL/6 male mice for 48 h, at a dose of 130 mg/kg body wt per d mixed with powdered standard laboratory mouse chow. The control group was composed of seven adult C57BL/6 male mice and was fed with the same powdered food without HCTZ. The dose was determined by preliminary experiments in which various doses were tested (65, 130, and 260 mg/kg body wt per d); 130 mg/kg body wt per d was found to be the minimal dose required to detect a significant increase in renal Na+ excretion in metabolic studies (data not shown).
Animals were housed in metabolic cages. They first were allowed to adapt for 5 d to the cages before the experimental period. Urine then was collected daily under light mineral oil for urinary flow and electrolyte measurements. At the end of the experimental period, animals were killed after anesthesia was induced by peritoneal injection of ketamine and xylazine (0.1 and 0.01 mg/g body wt, respectively). Plasma was collected on heparin, and kidneys were removed. Urinary pH and Pco2 were measured with a pH/blood-gas analyzer (Radiometer ABL555; Copenhagen, Denmark). Serum and urine electrolytes and urine creatinine were measured with a Konelab 20i autoanalyzer (Thermo Electron Corp., Eragny Parc, France). Blood pH, Pco2 and Po2 were measured with an AVL Compact 1 pH/blood-gas analyzer (AVL Instruments Médicaux, Eragny-sur-Oise, France). Plasma and urinary aldosterone were measured by RIA (DPC Dade Behring, La Défense, France). Plasma renin activity was determined by RIA of angiotensin I generated by incubation of the plasma at pH 8.5 in the presence of an excess of angiotensinogen as described previously (20).
Rabbit polyclonal antibody directed to C-terminal amino acids 630 to 643 of mouse pendrin has been characterized (21,22). Affinity-purified rabbit polyclonal antibody against the B1 subunit of the H+-ATPase was generated by S. Breton (Massachusetts General Hospital, Charlestown, MA). This antibody was raised against the last 13 amino acids of the C-terminal tail of the rat ATP6V1B1 subunit isoform (C-QGAQQDPASDTAL). The peptide was coupled to KLH via a cysteine positioned to its N-terminal end. Western blotting showed one single band in kidney extracts and complete inhibition of the signal after preincubation of the antibody with the B1 peptide (data not shown). Rabbit polyclonal antibodies directed to α-ENaC (amino acids 46 to 68), β-ENaC (amino acids 617 to 638), and γ-ENaC (amino acids 629 to 650) were described by Masilamani et al. (23). The ENaC antibodies were tested in preliminary experiments to verify that the specificity was identical to that of the antibodies from the original batch (data not shown).
Membrane fraction preparation and immunoblotting procedures for comparing two sets of samples of renal cortical membranes with regard to relative abundance of specific proteins were described in detail previously (17,24). Coomassie blue–stained polyacrylamide gels were used to control equality of protein loading for each series, as described in detail previously (17,24). The dilutions of primary antibodies used in this study were as follows: Anti-pendrin, 1:30,000; anti–H+-ATPase, 1:30,000; anti–α-ENaC, 1:3000; anti–β-ENaC, 1:20,000; and anti–γ-ENaC, 1:2000. Densitometric values were normalized to the mean for the control group in a given experiment, which was defined as 100%, and results were expressed as means ± SE.
Kidneys of vehicle and HCTZ-treated mice were fixed by vascular perfusion with 3% paraformaldehyde and 0.05% picric acid in PBS, frozen in liquid propane, and stored at −80°C until use. Kidneys of NCC−/− and NCC+/+ mice were taken from a previous experiment (25). Tissue was processed for immunolabeling as described previously (25). Briefly, cryosections (5 μm thick) were incubated overnight at 4°C with anti-pendrin antibody (dilution 1:400). Binding sites of the primary antibodies were detected with a Cy3-conjugated goat anti-rabbit IgG (Jackson ImmunoResearch Laboratories, West Grove, PA) diluted 1:1000.
Kidney sections were immunostained for pendrin and were analyzed by an experienced investigator who was blinded to the treatment and genotype of the mice. Using a Zeiss epifluorescence microscope, CNT and CCD profiles were randomly selected and the number of intercalated cells with apical pendrin immunostaining was quantified and expressed as percentage of the total number of pendrin-positive cells analyzed. At least 200 pendrin-positive cells were evaluated per animal.
All data are presented as means ± SEM. Comparisons among groups were assessed by ANOVA or unpaired t test. Differences were considered significant at P < 0.05.
Mouse Pendrin Protein Abundance Was Decreased in Response to Long-Term NaCl Loading
We previously described that an increased oral Cl− load reduces pendrin expression in kidneys of rats (17). Because it is possible that mice and rats differ with respect to mechanisms of NaCl transport regulation, we first tested whether we could reproduce this effect in mice. In response to high NaCl intake, extracellular fluid volume (ECFV) was only moderately expanded in the treated mice, because there was no detectable change in plasma aldosterone concentration when compared with control group values (361 ± 57 pM in the NaCl-loaded group versus 467 ± 64 pM in the control group; P = 0.24). In the same way, plasma renin activity tended to be lower in the treated group, but this difference did not reach statistical significance (882 ± 286 ng angiotensin I/h per ml in treated mice versus 1452 ± 494 ng angiotensin I/h per ml in controls; P = 0.3). As previously found in rat, pendrin was significantly downregulated (71 ± 3% in NaCl-loaded mice versus 100 ± 8% in controls; P = 0.006; Figure 1, A and B). α-ENaC protein abundance tended to be lower, but a statistically significant difference was not demonstrated (74 ± 7% in NaCl-loaded mice versus 100 ± 11% in control group; P = 0.07; Figure 1, A and C). In contrast, as shown in Figure 1, A and D, the profile of γ-ENaC was changed with a shift from the 70-kD “active” molecular form, which was decreased (23 ± 5% in NaCl-loaded mice versus 100 ± 13% in control mice; P = 0.0001), toward the 85-kD “inactive” form (128 ± 2% in NaCl-loaded mice versus 100 ± 3% in control mice; P < 0.0001). β-ENaC protein abundance was markedly upregulated in treated mice (184 ± 10% in NaCl-loaded mice versus 100 ± 8% in control group; P < 0.0001; Figure 1, A and E). Importantly, this corresponded exactly to the opposite changes observed previously by Masilamani et al. (26) in response to dietary NaCl restriction in the rat.
No Changes in Pendrin Protein Abundance Were Detectable in ENaCL/L Mice Despite a Marked Secondary Hypoaldosteronism
We next assessed the effect of ECFV expansion/low aldosterone concentration in the presence of unaltered chloride intake by analyzing pendrin protein abundance in mice with Liddle’s mutation. The phenotype of ENaCL/L mice was exactly as described previously by Pradervand et al. (18,27). Despite their activated Na+ transport in the CD, Liddle mice did not differ from their controls with respect to their body weight, acid–base status, or plasma electrolyte concentrations (Table 1). However, mutated mice exhibited an increase in ECFV as attested by a dramatic decrease in urinary and plasma aldosterone concentrations (Table 1). Despite this secondary hypoaldosteronism, no changes in pendrin protein abundance were detectable in Liddle mice (95 ± 13% in ENaCL/L mice versus 100 ± 12% in ENaCWT mice; P = 0.8; Ballerinas Mephisto Schwarze Schwarze Dora Ballerinas Dora Schwarze Mephisto Mephisto xvIxqHO, A and B). However, we cannot exclude the possibility that hypoaldosteronism secondary to the Liddle mutation led to a small decrease in pendrin protein abundance that was below the detection limit of our immunoblotting technique (<20% change). In contrast, both α- and γ-ENaC subunits clearly were downregulated in Liddle mice (66 ± 3% in ENaCL/L mice versus 100 ± 5% in ENaCWT mice [P < 0.0001; Figure 2, A and C] and 38 ± 4% in ENaCL/L mice versus 100 ± 8% in ENaCWT mice [P < 0.0001; Figure 2, A and D], for α- and γ-ENaC, respectively). As expected, the β-ENaC subunit was not detectable in ENaCL/L mice because the mutation deleted the C-terminal end of the protein that contains the epitopes that are recognized by the β-ENaC–specific antibody (Figure 2A).
Pendrin Abundance Is Downregulated whereas ENaC Expression Is Stimulated in Response to Acute HCTZ Administration
To examine the effect of a secondary hyperaldosteronism as a result of a renal loss of NaCl (i.e., without NaCl restriction) on pendrin protein expression, we next tested in wild-type mice the acute effect of HCTZ, a pharmacologic inhibitor of NCC, the NaCl transporter of the DCT. Physiologic data are shown in Tables 2 and 3. As expected, treated mice exhibited a marked increase in urinary excretion of Na+ and Cl− after the first day of drug administration when compared with control values. After 48 h of HCTZ treatment, the urinary excretion of Na+ and Cl− returned toward normal values, indicating that the mice compensated (Table 3). HCTZ-treated mice developed a marked decrease in ECFV, attested by a marked decrease in body weight, and an increase in plasma protein, urea, and aldosterone concentrations (see Table 2). Unexpected, pendrin protein abundance was not up- but downregulated in treated mice when compared with controls, as shown in Figure 3, A and B (67 ± 9% in HCTZ group versus 100 ± 11% in vehicle group; P = 0.04). In contrast, α-ENaC expression was increased (155 ± 11% in HCTZ group versus 100 ± 5% in vehicle group; P < 0.001; Figure 3, A and C), γ-ENaC expression showed a shift from the 85- to the 70-kD band (70 kD: 351 ± 23% in HCTZ group versus 100 ± 20% in vehicle group [P < 0.0001]; 85 kD: 53 ± 4% in HCTZ group versus 100 ± 7% in vehicle group [P < 0.0001]; Figure 3, A and D), and β-ENaC expression was decreased (47 ± 5% in HCTZ group versus 100 ± 6% in vehicle group; P < 0.0001; Figure 3, A and E), consistent with previous results from rat models with elevated plasma aldosterone levels (23,26).
Because recent data indicate that subcellular redistribution of pendrin is an important mechanism of pendrin regulation (15,16,21), the possibility exists that total protein abundance as assessed by Western blot on renal membrane fractions may not represent the physiologically active pool of pendrin. Therefore, we next assessed whether HCTZ treatment induced a translocation of pendrin protein to the apical membrane while total protein abundance was decreased. However, Figure 4 shows that after 2 d of HCTZ administration, pendrin apical expression was markedly decreased, further supporting the conclusion that the functional pool of pendrin is decreased in this situation.
Several Membrane Proteins from the Distal Nephron, Including Pendrin, Were Upregulated in NCC−/− Mice
The effect of HCTZ on pendrin protein that was observed in the preceding series of experiments is paradoxic with respect to the need for NaCl reclamation in response to blockade of the NaCl co-transporter of the DCT, NCC. Therefore, we next studied the regulation of pendrin in NCC−/− mice because these mice have the same defect in NaCl transport chronically. Moreover, in contrast to acute HCTZ treatment, this defect generally is compensated in mice that ingest a NaCl-replete diet, presumably because of increased NaCl reabsorption in the downstream segments CNT and CD, where pendrin and ENaC are expressed (Brautschuhe Krause Blockabsatz Marineblau Elegantpark Hochzeit Riemchen HC1808 Damen Pumps Satin 36 Gr wqE48On6). Physiologic data from the NCC−/− and NCC+/+ mice that were used in these studies are summarized in Table 4. As described previously (19), when fed a Na+-replete diet, NCC−/− mice were normal with respect to their weight, plasma electrolyte concentrations, and acid–base balance. Figure 5, A and B, shows that pendrin protein expression was markedly increased in the NCC−/− mice when compared with the NCC+/+ mice (260 ± 24% in NCC−/− versus 100 ± 8% in NCC+/+; P < 0.0001). ENaC protein abundance also was increased in the NCC−/− mice (Figure 5), with an increase in the protein abundance of all of the three ENaC subunits: α-ENaC (162 ± 11% in NCC−/− versus 100 ± 9% in NCC+/+; P < 0.001; Figure 5, A and C), β-ENaC (183 ± 11% in NCC−/− versus 100 ± 12% in NCC+/+; P < 0.001; Figure 5, A and E), and γ-ENaC (Figure 5, A and D). However, for γ-ENaC, the increase consisted of an increased expression of the two different 70- and 85-kD molecular forms of the protein (348 ± 34 and 120 ± 5% in NCC−/− versus 100 ± 9 and 100 ± 6% in NCC+/+; P < 0.0001 and P = 0.03 for the 70- and 85-kD bands, respectively).
Loffing et al. (25) reported that the fractional cortical tubular volume of CNT is approximately twice as high in NCC−/− than in NCC+/+ mice. This likely reflects a hypertrophy/hyperplasia of the CNT in NCC knockout mice that contributes to compensation of the Na+ transport defect in the preceding DCT. To test whether the upregulation of the Cl− transporter pendrin that was observed in our experiments also was explained by the hypertrophy/hyperplasia of the CNT and therefore to an increase in the number of pendrin-positive intercalated cells, we assessed pendrin protein expression by immunofluorescence studies on kidney sections from NCC−/− and NCC+/+ mice. Figure 6 shows that the number of pendrin-positive cells was markedly increased in NCC−/− mice when compared with NCC+/+ mice, whereas the intensity of pendrin labeling did not seem to be different. These results indicated that the upregulation of NaCl transporters from the CNT and CD was at least partly the consequence of tubular hypertrophy. This assumption was supported further by the fact that the protein abundance of the intercalated cell–specific B1 subunit of the H+-ATPase was increased in magnitude similar to that of pendrin in NCC−/− mice (317% ± 33 in NCC−/− versus 100% ± 12 in NCC+/+; P < 0.0001; Figure 7). However, although no obvious difference in subcellular localization of pendrin was noted in this model (Figure 8A), careful morphometric quantification showed that the percentage of pendrin-positive cells with apical pendrin expression was increased slightly (Figure 8B).
The CNT and the CD play a critical role in the fine-tuning of NaCl and acid–base balance. In the CNT and the CD, unlike most nephron segments, transepithelial reabsorption of Na+ and Cl− occurs through two separate cellular pathways (11,29,30). The other unique feature of these nephron segments is that Cl− reabsorption is linked molecularly to bicarbonate secretion. Therefore, these distinct Na+ and Cl− pathways allow independent adaptation of NaCl and NaHCO3 reabsorption. However, Na+ reabsorption, which is under the control of the steroid hormone aldosterone, needs to be coordinated to Cl− reabsorption to affect ECFV. Therefore, the aim of our study was to assess whether aldosterone plays a major role in regulation of pendrin or whether pendrin is regulated primarily by a Cl−-dependent mechanism that is aldosterone independent. Our data confirm our previous observation made in the rat that pendrin is strikingly dependent on factors that are associated with changes in distal Cl− delivery (17). In contrast, pendrin expression did not always correlate with circulating aldosterone level. Indeed, pendrin was downregulated by an increase in NaCl intake even in the absence of a detectable decrease of aldosterone secretion; conversely, in experiments with Liddle mice, we could not detect any significant changes in pendrin protein abundance despite a marked decrease in aldosterone concentration. This could reflect that pendrin is sensitive mainly to an increase and not to a decrease in aldosterone levels, but the experiments shown in Figures 3 and 4 clearly demonstrate that when distal chloride delivery was increased as a result of diminished NaCl reabsorption in the DCT, pendrin was downregulated despite a marked secondary hyperaldosteronism. This contrasts with the previous notion that aldosterone is a strong stimulatory factor of pendrin expression. Verlander et al. (16) found that prolonged administration of the synthetic aldosterone analogue DOCP increases pendrin mRNA expression and cell surface abundance of pendrin in intercalated cells. However, in the latter study, aldosterone was administered exogenously at high concentration (i.e., able to raise BP significantly). Therefore, it is possible that aldosterone when administrated in a supraphysiologic dose has a much more pronounced effect on pendrin. One other possibility, as noted by the authors themselves, is that the effect of DOCP was indirect. For example, chronic DOCP treatment leads to metabolic alkalosis and hypokalemia (16), two situations that have been shown to induce the translocation of pendrin from the cytosolic compartment to the apical membrane of the cells (21). Nevertheless, the data in Figures 3 and 4 clearly suggest that in the physiologic range of aldosterone concentration, factors that are associated with changes in distal chloride delivery are the more important determinants regulating pendrin.
The mechanism of Cl−-dependent regulation of pendrin remains unsettled, but our data raise the possibility that it involves the capability to sense changes in the rate of urinary Cl− delivery or in luminal fluid Cl− concentration. In fact, we document adaptive changes in pendrin protein abundance in response to maneuvers that are expected to alter dramatically distal chloride delivery while plasma [Cl−] was unchanged. It is interesting that pendrin has been detected not only in mammals but also in the gill epithelium of a euryhaline elasmobranch, Dasyatis sabina (Atlantic stingray), in cells with a phenotype similar to that of B intercalated cells (31). In Dasyatis sabina, pendrin is believed to be an important mechanism for chloride reclamation required for the fish to adapt to fresh water (i.e., to low [Cl−] conditions in the surrounding environment) (32). In contrast, increasing the salinity of the water markedly decreased pendrin protein abundance (31). This observation supports the hypothesis that renal pendrin regulation in mice might be due to changes in intratubular chloride concentration. However, because we did not measure luminal chloride delivery in the CNT and the CD in our various experimental conditions, it is not possible from the data of this study to conclude firmly that pendrin was regulated by changes in distal chloride delivery or concentration per se. Moreover, several endocrine or paracrine factors are expected to have been altered in parallel or by changes in distal chloride delivery and might be variably involved in pendrin regulation (e.g., vasopressin, bradykinin). Evidence for additional complexity in the regulation of pendrin is the disparity between the acute effect (48 h) of HCTZ to decrease greatly pendrin expression and the upregulation of pendrin in NCC null mice. It should be emphasized that the latter model involves a lifelong adaptation to the absence of NCC function, in contrast to the acute effect of HCTZ. Long-term changes in electrolyte transport conditions may induce secondary morphologic changes in the kidney. In fact, in the mouse model of NCC disruption, we found that the upregulation of pendrin was largely attributable to an increased prevalence of pendrin-positive intercalated cells as a result of hypertrophy/hyperplasia of the CNT (see Figures 6 and 8). Similar morphologic changes occur also in other models with increased ion transport activity in the ASDN, such as prolonged furosemide treatment (33–37), and may explain the reported upregulation of pendrin in response to this diuretic (17). The signals underlying these chronic adaptations are not known. Clearly, the disparity in pendrin regulation between the acute HCTZ model and the chronic NCC null model indicates that factors other than acute changes in distal chloride delivery must be involved in pendrin regulation.
Our results indicate that pendrin is regulated at the protein level by factors that are associated with changes in distal chloride delivery or transport, whereas aldosterone in the physiologic range plays a minor role. However, in response to situations that cause distal nephron hypertrophy and hyperplasia, such as long-term inactivation of NCC, pendrin expression is upregulated by several mechanisms, including hyperplasia of intercalated cells. In this setting, the B and non-A non-B intercalated cells may participate in the renal compensatory mechanisms that limit NaCl loss by increasing pendrin-mediated Cl− reabsorption.
Support was provided to P.S.A. by National Institutes of Health grants DK33793 and DK17433, to S.B. by National Institutes of Health grant DK38452, to J.L. by Swiss National Science Foundation grant 3200B0-105769/1, and to D.L.-C. by a Marie-Heim-Vögtlin Fellowship of the Swiss National Science Foundation.
We thank Dr. Edith Hummler and Prof. Bernard Rossier for providing the Liddle mice and Dr. Gary E. Shull for providing the NCC knockout mice and for helpful discussions.
Published online ahead of print. Publication date available at www.jasn.org.
- © 2006 American Society of Nephrology