Dysregulation of the WNK4-SPAK/OSR1 pathway has a minor effect on baseline NKCC2 phosphorylation

Abstract

The with-no-lysine kinase 4 (WNK4)-sterile 20/SPS-1-related proline/alanine-rich kinase (SPAK)/oxidative stress-responsive kinase 1 (OSR1) pathway mediates activating phosphorylation of the furosemide-sensitive Na⁺-K⁺-2Cl⁻ cotransporter (NKCC2) and the thiazide-sensitive NaCl cotransporter (NCC). The commonly used pT96/pT101-pNKCC2 antibody cross-reacts with pT53-NCC in mice on the C57BL/6 background due to a five amino acid deletion. We generated a new C57BL/6-specific pNKCC2 antibody (anti-pT96-NKCC2) and tested the hypothesis that the WNK4-SPAK/OSR1 pathway strongly regulates the phosphorylation of NCC but not NKCC2. In C57BL/6 mice, anti-pT96-NKCC2 detected pNKCC2 and did not cross-react with NCC. Abundances of pT96-NKCC2 and pT53-NCC were evaluated in Wnk4^−/−, Osr1^−/−, Spak^−/−, and Osr1^−/−/Spak^−/− mice and in several models of the disease familial hyperkalemic hypertension (FHHt) in which the CUL3-KLHL3 ubiquitin ligase complex that promotes WNK4 degradation is dysregulated (Cul3^+/−/Δ9, Klhl3^−/−, and Klhl3^R528H/R528H). All mice were on the C57BL/6 background. In Wnk4^−/− mice, pT53-NCC was almost absent but pT96-NKCC2 was only slightly lower. pT53-NCC was almost absent in Spak^−/− and Osr1^−/−/Spak^−/− mice, but pT96-NKCC2 abundance did not differ from controls. pT96-NKCC2/total NKCC2 was slightly lower in Osr1^−/− and Osr1^−/−/Spak^−/− mice. WNK4 expression colocalized not only with NCC but also with NKCC2 in Klhl3^−/− mice, but pT96-NKCC2 abundance was unchanged. Consistent with this, furosemide-induced urinary Na⁺ excretion following thiazide treatment was similar between Klhl3^−/− and controls. pT96-NKCC2 abundance was also unchanged in the other FHHt mouse models. Our data show that disruption of the WNK4-SPAK/OSR1 pathway only mildly affects NKCC2 phosphorylation, suggesting a role for other kinases in NKCC2 activation. In FHHt models NKCC2 phosphorylation is unchanged despite higher WNK4 abundance, explaining the thiazide sensitivity of FHHt.

NEW & NOTEWORTHY The renal cation cotransporters NCC and NKCC2 are activated following phosphorylation mediated by the WNK4-SPAK/OSR1 pathway. While disruption of this pathway strongly affects NCC activity, effects on NKCC2 activity are unclear since the commonly used phospho-NKCC2 antibody was recently reported to cross-react with phospho-NCC in mice on the C57BL/6 background. Using a new phospho-NKCC2 antibody specific for C57BL/6, we show that inhibition or activation of the WNK4-SPAK/OSR1 pathway in mice only mildly affects NKCC2 phosphorylation.

INTRODUCTION

The thick ascending limb (TAL) and distal convoluted tubule (DCT) of the mammalian kidney play essential roles in the control of extracellular fluid volume, urinary concentration, calcium and magnesium homeostasis, and pH balance (1). In the TAL, the major Na⁺ transport pathway is the Na⁺-K⁺-2Cl⁻ cotransporter (NKCC2), which is specifically inhibited by loop diuretics, and the major pathway along the DCT is the thiazide-sensitive Na⁺-Cl⁻ cotransporter (NCC). Mutations in SLC12A1 (encoding NKCC2) cause Bartter syndrome type 1, characterized by hypokalemic alkalosis, hypomagnesemia, hypercalciuria, and early presentation with severe volume depletion (2), while mutations in SLC12A3 (encoding NCC) cause Gitelman syndrome (NCC), characterized by hypokalemic alkalosis, hypocalciuria, hypomagnesemia, and milder clinical manifestations (3).

Both NKCC2 and NCC are phosphorylated at several serine and threonine residues along their amino-terminal intracellular tails by oxidative stress-responsive kinase 1 (OSR1) and sterile 20/SPS-1-related proline/alanine-rich kinase (SPAK) (4, 5), which are themselves activated via phosphorylation by upstream with-no-lysine (WNK) kinases. Although OSR1 and SPAK are both expressed along the TAL and DCT, the disruption of each kinase has different effects on cotransporter phosphorylation in vivo. Constitutive kidney-specific (KS-)Osr1^−/− mice display only modestly reduced phosphorylation of only NKCC2 (6), while constitutive Spak^−/− mice display a marked reduction in phosphorylation of only NCC (7–9). These data suggest that OSR1 and SPAK are the main activators of NKCC2 and NCC in the kidney, respectively. The strong effect of SPAK on NCC activity is supported by the similar phenotypes of Slc12a3^−/− mice (10) and Spak^−/− mice, with both showing hypocalciuria and hypomagnesemia (7–9), recapitulating Gitelman syndrome. KS-Osr1^−/− mice exhibited a Bartter-like phenotype with slightly higher urine volume and urinary Ca²⁺ excretion and normal renin concentration (6). However, their phenotype was much milder than that of Slc12a1^−/− mice that displayed hypokalemic metabolic alkalosis, more severe urinary Ca²⁺ wasting, higher plasma renin, and required treatment with indomethacin to prevent early death due to severe dehydration as a result of polyuria (11). These differences in phenotype severity suggest that NKCC2 phosphorylation is only mildly stimulated by the WNK4-SPAK/OSR1 pathway.

Dysregulation of the WNK-SPAK/OSR1 pathway causes the disease familial hyperkalemic hypertension (FHHt), which presents as the mirror image of Gitelman syndrome with hyperkalemia, hypertension, hyperchloremia, and metabolic acidosis. The pathophysiology of FHHt is primarily attributed to inappropriate activation of NCC. FHHt is caused by mutations in the genes encoding WNK1 and WNK4 as well as mutations in cullin 3 (CUL3) and Kelch-like 3 (KLHL3) (12, 13), components of an E3 ubiquitin ligase that mediates WNK degradation. CUL3 is a scaffold protein that can interact with different substrate adaptors and a RING ubiquitin ligase to form unique complexes that target specific substrates. Interaction with KLHL3 generates an E3 ubiquitin ligase that targets WNK kinases for proteasomal degradation. FHHt-causing mutations in WNKs, CUL3, or KLHL3 impair WNK4 degradation leading to excessive NCC phosphorylation. CUL3 mutations cause the most severe form of FHHt, and while the earlier onset of hypertension reported is likely to involve effects on the vasculature (14), hyperactivation of NKCC2 in the kidney could potentially contribute to the more severe electrolyte defects. CUL3 is expressed along all nephron segments, but KLHL3 is only expressed at high levels along DCT, with very low expression in TAL (15). However, KLHL2, a homolog of KLHL3, is expressed at significant levels in TAL (16) and mediates the degradation of WNK4 in vitro and in vivo (17, 18). These observations suggest that inappropriate NKCC2 phosphorylation arising from impaired CUL3/KLHL2-mediated WNK4 degradation along TAL may contribute to CUL3-mediated FHHt. Indeed, we have previously reported higher NKCC2 phosphorylation in a mouse model of CUL3-mediated FHHt (19).

A recent study from Moser et al. (20) demonstrated that C57BL/6J (BL/6) mice have a five amino acid deletion (Δ97–101) in NKCC2 that interferes with the binding of the most commonly used pT96/pT101 NKCC2 antibody. This results in cross-reaction of this antibody with pNCC (20), calling into question some previous findings with regard to WNK4-SPAK/OSR1 regulation of NKCC2 that used mice on the BL/6 background, including our previous report of increased NKCC2 phosphorylation in CUL3-mediated FHHt. Therefore, to clarify the effect of WNK4-SPAK/OSR1 signaling on NKCC2 activity in the kidney, we generated a new pT96-NKCC2 antibody and used several mouse models to determine more conclusively 1) the role of the WNK4-SPAK/OSR1 pathway in activation of NKCC2 and 2) whether WNK4-SPAK/OSR1-NKCC2 is activated along the TAL in CUL3-mediated but not KLHL3-mediated FHHt.

MATERIALS AND METHODS

Animals

Animal experiments were approved by the Oregon Health and Science University Institutional Animal Care and Use Committee (protocol IP00286). Generation and characterization of Slc12a3^−/− mice (10), Wnk4^−/− mice (21, 22), Osr1^−/− mice (23, 24), Spak^−/− mice (8, 25), Osr1^−/−/Spak^−/− mice (24), Klhl3^−/− mice (26), Cul3^−/− mice (15, 26–28), Cul3 heterozygotes also expressing CUL3-Δ9 (Cul3^+/−/Δ9) (19, 26, 28), and Klhl3 homozygous mutation (Klhl3^R528H/R528H) (29) have been published previously. For disruption of Osr1, Klhl3, and Cul3 and expression of CUL3-Δ9, the doxycycline-inducible tubule-specific Pax8-rtTA/TRE-LC1 system was used (26, 28, 30). CUL3-Δ9 was expressed from the Loxp-STOP-Loxp-Cul3-Δ9-IRES-tdTomato transgene (31). To induce recombination at floxed sites, 2 mg/mL doxycycline (Thermo Fisher Scientific, Waltham, MA) in 5% sucrose in drinking water was administered for 3 wk. After doxycycline treatment, mice were returned to regular drinking water for at least 2 wk before experiments were performed. Control mice given 5% sucrose drinking water were phenotypically equivalent to WT mice and were genetically identical littermates of those that received doxycycline. All mice were backcrossed onto the C57BL/6J background, and males or females were used as indicated. All samples were collected between 10 AM and 2 PM, and furosemide response tests ran from 9 AM to 3 PM.

PCR Genotyping

Genomic DNA extracts were prepared from tail snips by heating overnight at 55°C in 300 μL of digestion solution containing 5 mM EDTA, 200 mM NaCl, 100 mM Tris (pH 8.0), 0.2% SDS, and 0.4 mg/mL proteinase K, followed by ethanol precipitation. Primers used in this study are provided in Supplemental Table S1 (https://doi.org/10.6084/m9.figshare.22557724.v1).

Antibodies

Antibody sources, species, dilutions, and validation references are provided in Supplemental Table S2 (https://doi.org/10.6084/m9.figshare.22564036.v1). An antibody against pT96-NKCC2 was generated in the rabbit by Phosphosolutions (Aurora, CO) using the antigen LQ(pT)MDAVPKIE, which differs from the anti-pT96-NKCC2 antibody from the Loffing group (YYLQ(p) TMDAV) (20). This peptide was predicted to generate an antibody specific to C57BL/6J pT96-NKCC2 but with minimal affinity for NCC. Antibody specificity of this fraction was tested by immunohistochemistry and Western blot analysis.

Kidney Slices

Kidney slices (300-μm) were serially cut using a Leica Vibratome 1000 S in ice-cold buffer containing 108.5 mM NaCl, 25 mM NaHCO₃, 1.0 mM NaH₂PO₄·1H₂O, 2.5 mM CaCl₂, 1.8 mM MgCl₂, 3 mM K⁺, and 25.0 mM glucose. Evenly cut sections were selected, divided into two groups, and incubated at 35°C for 30 min in buffer with vehicle (DMSO) or 20 nM calyculin A (inhibits PP1 and PP4) (32). Carbogen gas (95% CO₂ and 5% O₂) was bubbled through buffers throughout the procedure. Kidney sections were then snap-frozen in liquid nitrogen for Western blot analysis.

Western Blot Analysis

Harvested kidneys were snap frozen in liquid nitrogen and then stored at −80°C until use. A half kidney was homogenized using a Potter homogenizer in 1 mL cold buffer containing 300 mM sucrose, 50 mM Tris·HCl (pH 7.4), 1 mM EDTA, 1 mM EGTA, 1 mM NaVO₄, 50 mM NaF, 1 mM ditiothreitol, 1 mM phenylmethane sulfonyl fluoride, 1 µg/mL aprotinin, 4 µg/mL leupeptin, and phosphatase inhibitor (PhosStop; Roche, Mannheim, Germany). Homogenates were centrifuged at 6,000 rpm for 15 min at 4°C, and supernatants were transferred to a new tube and then stored at −80°C. Protein loading was adjusted by densitometric quantitation of total protein after Coomassie staining (see Supplemental Fig. S1; all Supplemental Figures are available at https://doi.org/10.6084/m9.figshare.22557709.v1) (19, 26, 28, 33). Protein samples were separated by electrophoresis on 4–12% Criterion XT Bis-Tris gels or 4–15% Criterion TGX stain-free gels (Bio-Rad Laboratories, Hercules, CA) and transferred to LF PVDF membranes using the Trans-Blot Turbo transfer system (Bio-Rad Laboratories). Membranes were blocked with 5% nonfat milk in TBS-Tween (Thermo Fisher Scientific), followed by incubation with primary antibody overnight at 4°C. Anti-KLHL3 antibody was diluted in CanGetSignal (TOYOBO, Osaka, Japan). Appropriate horseradish peroxidase-conjugated secondary antibody in blocking buffer was added to membranes for 1 h at room temperature. Membranes were developed using enhanced chemiluminescence, Western Lightning Plus–ECL (Perkin Elmer, Waltham, MA), and visualized using the Pxi digital imaging system (Syngene, Frederick, MA), and densitometry was performed with ImageJ (https://imagej.net/nih-image/).

Immunofluorescence

Kidneys were perfusion-fixed by 3% paraformaldehyde in PBS (pH 7.4), removed, dissected, and cryopreserved in 800 mosmol/kgH₂O sucrose in PBS overnight before embedding in Tissue-Tek O.C.T. compound (Sakura Finetek, Torrance, CA). Slides were prepared by cutting 5-μm sections and were stored at −80°C until use. Sections were incubated with 0.5% Triton X-100 in PBS for 30 min and then blocked with 5% BSA in PBS for 30 min, followed by incubation with primary antibody, diluted in blocking buffer, overnight at 4°C. Sections were incubated with appropriate fluorescent dye-conjugated secondary antibody, diluted in blocking buffer, for 1 h at room temperature before being mounted with Prolong Diamond Antifade Mountant (Invitrogen, Carlsbad, CA). Fluorescence images were acquired using a Keyence BZ-X810 fluorescence microscope (Osaka, Japan). For immunolocalization at least 2 mice per group were analyzed to confirm findings; representative images are shown.

Blood Analysis

Blood was collected via cardiac puncture under isoflurane anesthesia and transferred into heparinized tubes and then centrifuged at 2,000 g for 5 min at room temperature. Plasma was removed and stored at −80°C until use. Plasma sodium and potassium were determined by a flame photometer.

Furosemide Response Test

The furosemide response test was performed as described previously (22). Five male control and 5 male Klhl3^−/− mice were injected intraperitoneally with vehicle (1.2% ethanolamine in 0.09% NaCl solution) and then placed in metabolic cages, and urine samples were collected for 4 h. On a different day, the same mice were injected with furosemide (25 mg/kg body wt) in vehicle, followed by 4-h urine collections. After a washout period of 7 days, hydrochlorothiazide (HCTZ) was injected daily at 25 mg/kg body wt for five consecutive days. On day 5, the furosemide response test was performed with either vehicle or furosemide (25 mg/kg body wt) injected 1 h following the HCTZ injection. Urine sodium was determined by a flame photometer.

Statistics

The null hypothesis was tested using two-tailed unpaired t tests and one-way ANOVA using GraphPad Prism 9 as indicated in the figures. Post hoc analysis was performed using Dunnett’s multiple-comparison tests. All data are plotted as means ± SE. Significant P values are described as *P < 0.05, **P < 0.01, and ***P < 0.001.

RESULTS

Generation of anti-pT96-NKCC2 C57BL/6J Antibody

A recent analysis of mouse genomic sequences for Slc12a1, encoding NKCC2, by Moser et al. (20), revealed that in contrast to the majority of 22 mouse strains in the Ensembl database (https://useast.ensembl.org/Mus_musculus/Info/Index) including 129S6/SvEVTac (129 Sv) mice, C57BL/6J (BL/6) mice have a five amino acid deletion in the region targeted by anti-pT96/pT101 (Fig. 1A). Alignment of the NCC sequence from BL/6 and 129 Sv mice shows significant homology between NCC and NKCC2 in the region surrounding pT96/pT101 (Fig. 1A). Therefore, this deletion in NKCC2 hinders reliable detection of NKCC2 phosphorylation in BL/6 mice when using the commonly used anti-pT96/pT101 antibody (20). We generated a new antibody specifically directed against the T96 phospho-site in BL/6 mice using the pT96-NKCC2 immunizing phosphopeptide, designed to minimize cross-reactivity with phosho-NCC (Fig. 1A). We validated the specificity of anti-pT96, comparing it with both anti-pT96/pT101 NKCC2 and anti-pT53-NCC antibodies, using Western blot analysis and immunofluorescence. In kidney lysates from BL/6 mice, the anti-pT96/pT101 NKCC2 antibody detected a strong diffuse signal that was similar to that observed with anti-pT53-NCC; a weaker more discrete signal was detected in kidney lysates from 129 Sv mice (Fig. 1B). Although total NKCC2 was detected in both BL/6 and 129 Sv lysates, the anti-pT96-NKCC2 antibody detected a strong discrete signal in BL/6 lysates that was absent in 129 Sv lysates (Fig. 1B). Protein phosphatase treatment (+PPs) decreased signal detected by the anti-pT96 antibody but not by anti-total NKCC2, but did not completely ablate it (Supplemental Fig. S2). This suggests anti-pT96 preferentially detects phosphorylated NKCC2 but also has some affinity for nonphosphorylated NKCC2. To eliminate the detection of nonphosphorylated NKCC2, we preblocked with the nonphosphopeptide (non-phospho). Blocking with the immunizing phosphopeptide (phospho) ablated almost all signal detected with the anti-pT96 antibody, while blocking with non-phospho only modestly reduced the signal (Fig. 1C). This blocking effect was confirmed by immunofluorescence (Fig. 1D). The signal detected by anti-pT96-NKCC2 did not overlap with parvalbumin, a marker of the DCT (Fig. 1E). Therefore, for subsequent Western blots and immunofluorescence with the anti-T96 antibody we used 100 ng/mL nonphospho peptide to block nonphosphorylated NKCC2. Finally, the anti-pT96-NKCC2 antibody detected significantly higher NKCC2 phosphorylation without significant changes in total NKCC2 in kidney slices incubated with the phosphatase inhibitor calyculin A, confirming that the antibody can detect higher pT96-NKCC2 abundance (Supplemental Fig. S3). All subsequent experiments were performed using samples from mice on the BL/6 background.

Figure 1.

Generation of anti-pNKCC2 (T96 C57BL/6J) antibody. A: analysis of genomic sequences using Ensembl (https://useast.ensembl.org/Mus_musculus/Info/Index) revealed that C57BL/6J mice, in contrast to most mouse strains including 129S6/SvEVTac mice, have a 5 amino acid deletion (Δ97–101) in the region targeted by anti-pT96/pT101 antibody. This was confirmed by PCR of genomic DNA and sequencing (20). Alignment of the NCC sequence from C57BL/6J and 129S6/SvEVTac mice shows significant homology between NCC and NKCC2 in the region surrounding pT96/pT101 (blue background). The amino acid sequences of immunizing peptides are shown for the antibodies used: anti-pT96/pT101 NKCC2 (34), anti-pT53-NCC (35), and anti-pT96-NKCC2 (validated here). The colored lines above the aligned sequences represent the pT96 (red) and pT96/pT101 (blue) immunizing peptide, which was designed to minimize cross-reactivity with NCC. B: anti-pT96/pT101 NKCC2 antibody produces a strong diffuse signal in kidney lysates from male C57BL/6J (BL/6) mice with a weaker, more discrete signal in kidney lysates from male 129S6/SvEVTac (129 Sv) mice. Anti-pT96-NKCC2 produces a strong, more discrete signal in BL/6 mice, which is absent in those from 129 Sv mice. Western blotting (C) and immunofluorescence (D) showed blocking with the immunizing phosphopeptide (phospho) ablates signal generated by anti-pT96 antibody in male BL/6 wild-type mice, while blocking with the non-phosphopeptide (non-phospho) only modestly reduced signal. E: immunofluorescence of kidney sections from male wild-type mice in the BL/6 background shows that anti-pT96 immunoreactivity (red) does not co-localize with parvalbumin immunoreactivity (green), a marker of the early DCT. Representative images of at least 2 mice total were analyzed with similar results. Scale bars: 50 µm. DCT, distal convoluted tubule; NKCC2, Na⁺-K⁺-2Cl⁻ cotransporter; NCC, Na⁺-Cl⁻ cotransporter.

To evaluate cross-reaction of anti-pT96/pT101 NKCC2 and anti-pT96-NKCC2 antibodies with phosphorylated NCC, we performed Western blots for pT53-NCC, pT96/pT101 NKCC2, pT96-NKCC2, and total NKCC2 using Slc12a3^−/− (NCC knockout) mice. Consistent with the results of Moser et al. (20), pT96/pT101 NKCC2 abundance was significantly lower in Slc12a3^−/− mice (37% vs. 100% WT, P = 0.0004), but unlike anti-pT53 abundance, was not completely absent (Fig. 2A). In contrast, pT96-NKCC2 and NKCC2 abundances were higher in Slc12a3^−/− mice (Fig. 2B and Supplemental Fig. S4). These suggest that this newly generated anti-pT96-NKCC2 antibody, unlike anti-pT96/pT101 NKCC2, does not cross-react with pT53-NCC.

Figure 2.

No cross-reactivity of anti-pT96-NKCC2 antibody with pNCC. A: Western blots of the whole kidney showed the strong diffuse band detected by the anti-pT53-NCC antibody was absent in Slc12a3^−/− mice. Similarly, anti-pT96/pT101 NKCC2 antibody detected a strong diffuse signal in WT mice, which was absent in Slc12a3^−/− mice, but a more discrete signal likely representing pT96/pT101 NKCC2 was still detected, suggesting cross-reactivity of the anti-pT96/pT101 NKCC2 antibody with pNCC. B: anti-pT96 signal was discrete and higher in Slc12a3^−/− mice and resembled NKCC2 signal, suggesting that anti-pT96 does not cross-react with pNCC. Individual values and means ± SE are shown. Statistical differences were examined by unpaired t tests. M, male; NKCC2, Na⁺-K⁺-2Cl⁻ cotransporter; NCC, Na⁺-Cl⁻ cotransporter; WT, wild-type.

Osr1 Deletion Modestly Decreases the pT96-NKCC2-to-NKCC2 Ratio

In vitro, NKCC2 can be phosphorylated at T96 by both SPAK and OSR1 (5). Constitutive kidney-specific (KS) Osr1 deletion in mice was reported to lower NKCC2 phosphorylation, leading to a Bartter-like phenotype (6). Global Spak deletion in mice dramatically lowers NCC phosphorylation at several sites including pT53 (7–9). The combination of constitutive Spak disruption and inducible disruption of Osr1 along renal epithelium further lowers pT53-NCC (24). To eliminate the potential developmental effects of Osr1 deletion and clarify the effect of inducible KS Osr1 deletion and/or Spak deletion on NKCC2 phosphorylation, we took advantage of our new pT96-NKCC2 antibody. We used several mouse models we previously generated with constitutive SPAK deletion and/or tubule-specific doxycycline-inducible OSR1 deletion: Osr1^−/−, Spak^−/−, and Osr1^−/−/Spak^−/− mice (8, 24). Western blot analysis showed that OSR1 was significantly lower in Osr1^−/− and Osr1^−/−/Spak^−/− mice (Supplemental Fig. S10). SPAK exists as three isoforms, full-length (FL) SPAK, SPAK2, and KS-SPAK, with SPAK2 and KS-SPAK antagonizing FL-SPAK activity (8, 36). Consistent with previous study (8), all three isoforms were not detected in Spak^−/− or Osr1^−/−/Spak^−/− mice (Supplemental Fig. S5). As would be expected for each particular model, phosphorylated forms of SPAK and OSR1 were largely absent. Compared with controls, pT96-NKCC2 abundance did not differ in Osr1^−/−, Spak^−/−, or Osr1^−/−/Spak^−/− mice, while NKCC2 abundance trended to be higher in Osr1^−/− mice and was significantly higher in Osr1^−/−/Spak^−/− mice (Fig. 3A). The pT96-NKCC2/NKCC2 ratio was significantly, but modestly, lower in Osr1^−/− and Osr1^−/−/Spak^−/− mice (Fig. 3A). In contrast, pT53-NCC abundance trended to be lower in Spak^−/− and Osr1^−/−/Spak^−/− mice, consistent with our previous study (Fig. 3A) (8, 24). pT96/pT101 NKCC2 abundance trended to lower in Spak^−/− mice and was decreased in Osr1^−/−/Spak^−/− mice (Supplemental Fig. S6). Effects on total and pT53-NCC were consistent with previous reports (7–9). Total NCC was lower in Spak^−/− and Osr1^−/−/Spak^−/− mice and the pT53-NCC/NCC ratio trended to lower in Spak^−/− and lowered in Osr1^−/−/Spak^−/− mice (Fig. 3B). pT53-NCC abundance trended towards higher and NCC abundance was higher in Osr1^−/− mice (Fig. 3B), which may reflect a response to lower plasma potassium levels (Supplemental Fig. S7), which promote NCC phosphorylation and activation (37, 38). These results suggest that Spak deletion strongly impairs NCC phosphorylation while inducible Osr1 deletion in adults only modestly impairs NKCC2 phosphorylation.

Figure 3.

Inducible Osr1 deletion modestly decreases pT96-NKCC2/NKCC2 ratio but not pT96-NKCC2 abundance. A: pT96-NKCC2 abundance did not differ in Osr1^−/−, Spak^−/−, or Osr1^−/−/Spak^−/− mice, while NKCC2 abundance trended to be higher in Osr1^−/− mice and was significantly higher in Osr1^−/−/Spak^−/− mice. B: pT53-NCC abundance trended to be lower in Spak^−/− and Osr1^−/−/Spak^−/− mice and higher in Osr1^−/− mice. Total NCC was lower in Spak^−/− and Osr1^−/−/Spak^−/− mice, and the pT53-NCC/NCC ratio trended to lower in Spak^−/− mice and was lower in Osr1^−/−/Spak^−/− mice. Individual values and means ± SE are shown. Statistical differences were examined by one-way ANOVA, followed by Dunnett’s multiple comparisons tests (vs. controls). M, male; NKCC2, Na⁺-K⁺-2Cl⁻ cotransporter; NCC, Na⁺-Cl⁻ cotransporter.

Wnk4 Deletion Does Not Significantly Affect pT96-NKCC2 Abundance

Wnk4 deletion strongly decreases NCC phosphorylation and its activity (21, 22), but NKCC2 activity was only mildly decreased in Wnk4^−/− mice despite a large reduction in pNKCC2 abundance, as assessed using the pT96/pT101 antibody (22). To clarify the effect of Wnk4 deletion on NKCC2 phosphorylation, we determined pT96-NKCC2 and total NKCC2 abundances in Wnk4^−/− mice. Western blot analysis showed that WNK4 was completely absent in Wnk4^−/− mice (Fig. 4A). Immunofluorescence showed that WNK4 is mainly expressed along medullary TAL and only weakly along cortical TAL in control mice, consistent with a previous report (22), but this WNK4 signal was completely absent in Wnk4^−/− mice (Fig. 4B). Also consistent with this previous study, pT53-NCC abundance was almost absent in Wnk4^−/− mice and pT96/pT101 NKCC2 abundance was reduced to 22% of control levels in Wnk4^−/− mice (Fig. 4C), similar to our data obtained for Slc12a3^−/− mice (Fig. 2A). However, pT96-NKCC2 abundance only trended to be slightly lower (78% vs. 100% WT, P = 0.06), and NKCC2 abundance did not differ in Wnk4^−/− mice compared with control mice (Fig. 4D). Our data with the pT96 antibody are more consistent with the relatively small difference in furosemide responsiveness previously reported (22). Immunofluorescence showed that the pT96-NKCC2 signal was less intense along cortical and medullary TAL, but was preserved in some cortical TALs (Fig. 4E). We therefore next evaluated whether the phosphorylated activated forms of SPAK and OSR1 (pSPAK/pOSR1, note that the antibodies used cannot distinguish) were preserved along cortical TAL in Wnk4^−/− mice. Western blot analysis showed that pSPAK/pOSR1 were 50% lower in Wnk4^−/− mice, but OSR1 did not differ, and SPAK and SPAK2 were higher in Wnk4^−/− mice (Fig. 5A). Immunofluorescence showed that pSPAK/pOSR1 was expressed at the apical membrane along medullary TAL through DCT in controls (Fig. 5B). In Wnk4^−/− mice, pSPAK/pOSR1 was only localized to cytoplasmic puncta along DCT, but was still localized to the apical membrane along TAL (Fig. 5B). Furthermore, pSPAK/pOSR1 signal was lower primarily along the medullary TAL, but was mainly preserved along cortical TALs, where pT96-NKCC2 signal was also preserved (Fig. 5B), suggesting another kinase regulates phosphorylation of SPAK/OSR1.

Figure 4.

Wnk4 deletion does not significantly lower pT96-NKCC2 abundance. A and B: Western blot analysis and immunofluorescence showed that WNK4 abundance and signal were completely absent in Wnk4^−/− mice. C: pT53-NCC abundance was almost absent in Wnk4^−/− mice, and pT96/pT101 NKCC2 abundance was clearly reduced in Wnk4^−/− mice. D: pT96-NKCC2 only trended to be mildly lower and NKCC2 abundance was not changed in Wnk4^−/− mice compared with control mice. E: immunofluorescence showed that pT96-NKCC2 abundance was reduced along both cortical and medullary TAL, but was preserved along a subset of cortical TALs (yellow arrowheads). Individual values and means ± SE are shown. Statistical differences were examined by unpaired t tests. Representative images of at least 2 mice total were analyzed with similar results. Scale bars: 50 µm. F, female; NKCC2, Na⁺-K⁺-2Cl⁻ cotransporter; NCC, Na⁺-Cl⁻ cotransporter; TAL, thick ascending limb.

Figure 5.

Apical localization of phosphorylated activated forms of SPAK and OSR1 (pSPAK/pOSR1) is mildly reduced along TAL in Wnk4^−/− mice. A: Western blot analysis showed that pSPAK/pOSR1 were 50% lower in Wnk4^−/− mice, but OSR1 did not differ, and SPAK and SPAK2 were higher in Wnk4^−/− mice. Statistical differences were examined by unpaired t tests. B: immunofluorescence showed that pSPAK/pOSR1 was expressed at the apical membrane along medullary and cortical TAL and DCT in controls. In Wnk4^−/− mice, pSPAK/pOSR1 was localized to cytoplasmic puncta along DCT but to the apical membrane along medullary and cortical TAL. pSPAK/pOSR1 expression was lower primarily along medullary TAL, but was mainly preserved along cortical TALs, where pT96-NKCC2 signal was also preserved (yellow arrowheads). Representative images of at least 2 mice total were analyzed with similar results. Scale bars: 50 µm. F, female. DCT, distal convoluted tubule; TAL, thick ascending limb.

The WNK1 gene generates two isoforms, the full-length kinase active form of WNK1 (L-WNK1) and a truncated kinase inactive form of WNK1 (KS-WNK1), through alternative promoters. L-WNK1 expression is higher along TAL than DCT (39), while KS-WNK1 is specifically expressed along DCT (39). KS-WNK1 is essential for the formation of cytoplasmic puncta termed WNK bodies, which are proposed to promote SPAK/OSR1 phosphorylation and activation along DCT (40). To clarify why pSPAK/pOSR1 is still localized to the apical membrane along the TAL following Wnk4 disruption, we determined whether WNK1 is expressed along the apical membrane of TAL in Wnk4^−/− mice. The anti-WNK1 antibody used detects both L-WNK1 and KS-WNK1. Immunofluorescence showed that WNK1 was localized to the apical membrane along TAL through DCT in controls, while WNK1 was localized to cytoplasmic puncta along the DCT but was still apical along TAL in Wnk4^−/− mice (Fig. 6, A and B). These data suggest that along the TAL, L-WNK1 may phosphorylate SPAK and OSR1 along the TAL following Wnk4 disruption. Thus, Wnk4 deletion strongly reduces NCC phosphorylation by impairing SPAK/OSR1 phosphorylation and apical localization, but the effect on NKCC2 phosphorylation is likely mitigated by L-WNK1 activity.

Figure 6.

WNK1 is apically localized along TAL in both control and Wnk4^−/− mice. A and B: immunofluorescence showed that WNK1 was localized to the apical membrane along TAL and DCT but its signal was relatively low. In Wnk4^−/− mice, WNK1 signal was more intense along DCT and localized to cytoplasmic puncta along DCT but the apical membrane along TAL. Representative images of at least 2 female mice total were analyzed with similar results. Scale bars: 50 µm. DCT, distal convoluted tubule; TAL, thick ascending limb.

WNK4 is Higher along the Cortical TAL in FHHt Models with Cul3 or Klhl3 Mutations

The CUL3-KLHL3 ubiquitin ligase complex promotes WNK4 degradation. Mutations in both CUL3 and KLHL3 cause FHHt by hyperactivating the WNK4-SPAK/OSR1-NCC pathway. Although Klhl3 is highly expressed along the DCT (15, 26, 41), KLHL2, another CUL3 ubiquitin ligase substrate adapter, is more highly expressed along the TAL (16). Since KLHL2 mediates the degradation of WNK4 in vitro and in vivo (17, 18), dysfunction of the CUL3-KLHL2 ubiquitin ligase might promote NKCC2 hyperactivation in CUL3-mediated FHHt. Indeed, we previously reported higher phosphorylation of NKCC2 in a CUL3 model of FHHt on the BL/6 background using the anti-pT96/pT101 NKCC2 antibody (19). However, since this antibody cross-reacts with pNCC on the BL/6 background, whether the WNK4-SPAK/OSR1-NKCC2 pathway is dysregulated in FHHt is now unclear. Here, we hypothesized that WNK4 abundance is increased along TAL in FHHt resulting from Cul3 mutation, which is expressed along the entire nephron, but not in Klhl3 FHHt models due to DCT-specific expression of KLHL3. To test this, we assessed WNK4 expression along TAL in Cul3^+/−/Δ9 mice and in Klhl3^−/− mice. Consistent with our previous results (19, 26), Western blot analysis showed that WNK4 abundance was increased in both Cul3^+/−/Δ9 mice and Klhl3^−/− mice (Fig. 7, A and B). Although baseline WNK4 abundance was clearly detected in the medulla but not in the cortex in controls, immunofluorescence showed a clear overlap of WNK4 and NKCC2 signal along a subset of cortical TALs in Cul3^+/−/Δ9 mice (Fig. 7C) and along cortical TAL in Cul3^−/− mice (Supplemental Fig. S8A). Consistent with single-cell RNA sequence data from the McMahon group showing low Wnk4 mRNA abundance in macula densa cells (42), the WNK4 signal was unchanged in cells immediately adjacent to the DCT in Cul3^+/−/Δ9 or in Cul3^−/− mice (Fig. 7C and Supplemental Fig. S8A). Unexpectedly, WNK4 and NKCC2 overlap was clearly observed along cortical TAL in Klhl3^−/− mice (Fig. 7D). Since Cul3 disruption increases KLHL3 protein abundance (26), we performed immunofluorescence for KLHL3 in Cul3^−/− mice to determine whether KLHL3 protein is also expressed along cortical TAL. KLHL3 was very weakly expressed along some cortical TAL tubules in controls, but its expression was higher in Cul3^−/− mice (Supplemental Fig. S8B), consistent with our previous findings that KLHL3 is itself a substrate for KLHL3. These data suggest the CUL3-KLHL3 complex may regulate WNK4 degradation along cortical TAL and that NKCC2 phosphorylation may be higher in FHHt.

Figure 7.

WNK4 is increased along a subset of cortical TALs in Cul3^+/−/Δ9 mice and Klhl3^−/− mice. A and B: Western blot analysis showed that WNK4 abundance was higher in Cul3^+/−/Δ9 mice and Klhl3^−/− mice than in controls, while KLHL3 abundance was almost absent in Klhl3^−/− mice. C and D: immunofluorescence showed that WNK4 was expressed strongly along medullary TAL but only weakly along cortical TAL in control mice. WNK4 signal was stronger along DCT and a subset of cortical TALs in Cul3^+/−/Δ9 mice and Klhl3^−/− mice. Circles indicate NKCC2+/WNK4- tubules and arrowheads indicate NKCC2+/WNK4+ tubules. Individual values and means ± SE are shown. Statistical differences were examined by unpaired t tests. Representative images of at least 2 mice total were analyzed with similar results. Scale bars: 50 µm. DCT, distal convoluted tubule; F, female; M, male; TAL, thick ascending limb.

pT96-NKCC2 Abundance and NKCC2 Activity Are Not Altered in CUL3 or KLHL3 FHHt Models

To clarify the effect of higher WNK4 on NKCC2 phosphorylation in both FHHt models, we next determined pT96-NKCC2 abundance in Cul3^+/−/Δ9 mice and Klhl3^−/− mice. Consistent with our previous results (19), abundances of pT53-NCC and pT96/pT101 NKCC2 were both higher in Cul3^+/−/Δ9 mice compared with controls (Fig. 8A). However, pT96-NKCC2 abundance did not differ between Cul3^+/−/Δ9 mice and controls (Fig. 8B). Similar results were observed in Klhl3^−/− mice (Fig. 8, C and D) and in mice with a constitutive homozygous FHHt-causing Klhl3 mutation (Klhl3^R528H/R528H mice) (Supplemental Fig. S9). To confirm no effect on NKCC2 activity in Klhl3^−/− mice, we performed a furosemide response test. Natriuresis induced by acute treatment with furosemide was slightly lower in Klhl3^−/− mice compared with control mice (Fig. 9A). Since pT96-NKCC2 was similar in the two groups, we reasoned that the furosemide effect may have been blunted in Klhl3^−/− mice by load-induced increases in NaCl reabsorption along the DCT, where NCC is activated by KLHL3 deletion (26, 41), compared with control mice. To determine the effects of furosemide in the absence of sodium reabsorption along the DCT, we repeated this experiment in mice pre-treated with HCTZ to block NCC activity. In this case, the furosemide response was similar between Klhl3^−/− mice and controls (Fig. 9B).

Figure 8.

pT96-NKCC2 abundance is unchanged in Cul3^+/−/Δ9 mice and Klhl3^−/− mice. A and B: Western blot analysis showed that abundances of pT53-NCC and pT96/pT101 NKCC2 were higher in Cul3^+/−/Δ9 mice compared with controls, consistent with our previous study (19), while pT96-NKCC2 abundance did not significantly differ between Cul3^+/−/Δ9 mice and controls. C and D: similarly, in Klhl3^−/− mice, pT96-NKCC2 and NKCC2 abundances were unchanged, although abundances of pT53-NCC and pT96/pT101 NKCC2 were increased. Individual values and means ± SE are shown. Statistical differences were examined by unpaired t tests. DCT, distal convoluted tubule; F, female; M, male; NKCC2, Na⁺-K⁺-2Cl⁻ cotransporter; NCC, Na⁺-Cl⁻ cotransporter; TAL, thick ascending limb.

Figure 9.

Effects of furosemide on urinary sodium excretion in control and Klhl3^−/− mice. A: 4 h of urine collection after a single furosemide injection (25 mg/kg body wt ip) showed that natriuresis induced by furosemide was mildly blunted in male Klhl3^−/− mice compared with male control mice. B: after treatment with hydrochlorothiazide for 5 days, furosemide response was similar between male Klhl3^−/− mice and male control mice, suggesting no increase in NKCC2 activity in Klhl3^−/− mice. Individual values and means ± SE are shown. Statistical differences were examined (left graphs) by two-way ANOVA, with post hoc analysis performed using unpaired t tests with Bonferroni correction or (right graphs) by two-tailed unpaired t tests.

To clarify the reason for no change in NKCC2 activity despite increased WNK4, we evaluated whether pSPAK/pOSR1 was increased along TAL in Klhl3^−/− mice. Western blot analysis showed abundances of pSPAK/pOSR1, OSR1, SPAK, and SPAK2 were higher in Klhl3^−/− mice (Fig. 10A). Immunofluorescence showed that pSPAK/pOSR1 was highly expressed in both medullary and cortical TAL in controls, although WNK4 signal was modest along cortical TAL. In Klhl3^−/− mice, pSPAK/pOSR1 signal was clearly higher along DCT but not along cortical TAL compared with controls (Fig. 10B). In Klhl3^−/− mice, Cul3^+/−/Δ9, and Cul3^−/− mice, WNK4 signal was also not higher along the segment of cortical TAL adjacent to DCT that likely corresponds to macula densa (Fig. 7, B and D, and Supplemental Fig. S8A). Notably, phosphorylation of SPAK and OSR1 were similar in all cortical TALs in Klhl3^−/− mice (Fig. 10B). As shown above, WNK1 was localized to the apical membrane along TAL and DCT in WT or control mice (Fig. 6B and Fig. 11B). In contrast, in Klhl3^−/− mice, WNK1 was mainly localized to cytoplasmic puncta along the DCT in Klhl3^−/− mice but remained apical along the TAL (Fig. 11, A and B). Since L-WNK1 is relatively resistant to ubiquitination by the CUL3-KLHL3 complex but KS-WNK1 is sensitive to it (43), these results suggest that apical WNK1 expression along cortical TAL is mainly L-WNK1, which contributes to SPAK and OSR1 phosphorylation under normal conditions. These results suggest that NKCC2 phosphorylation and/or activity is not further increased in FHHt caused by Cul3 and Klhl3 mutations despite higher WNK4 abundance since SPAK/OSR1 are already maximally activated by both L-WNK1 and WNK4.

Figure 10.

Phosphorylated forms of SPAK and OSR1 are not increased along cortical TAL in Klhl3^−/− mice. A: Western blot analysis showed that phosphorylated activated forms of SPAK and OSR1 (pSPAK/pOSR1) abundance was higher in Klhl3^−/− mice along with higher abundances of OSR1, FL-SPAK, and SPAK2. B: immunofluorescence showed that pSPAK/pOSR1 was highly expressed in both medullary and cortical TAL in controls. In Klhl3^−/− mice, pSPAK/pOSR1 expression was clearly increased along DCT but not along cortical TAL. Individual values and means ± SE are shown. Statistical differences were examined by two-tailed unpaired t tests. Representative images of at least 2 mice total were analyzed with similar results. Scale bars: 50 µm. DCT, distal convoluted tubule; M, male; TAL, thick ascending limb.

Figure 11.

WNK1 remains at the apical membrane along TAL in Klhl3^−/− mice. A and B: WNK1 was localized to apical along TAL and DCT in control mice. In Klhl3^−/− mice, WNK1 signal was higher and mainly localized to puncta along DCT but remained at the apical membrane along the TAL. Representative images of at least 2 male mice total were analyzed with similar results. Scale bars: 50 µm. DCT, distal convoluted tubule; TAL, thick ascending limb.

DISCUSSION

In this study, we generated a new anti-pT96-NKCC2 antibody that can specifically detect NKCC2 phosphorylation in BL/6 mice. This new antibody is similar to that generated by Moser and colleagues (20), since it does not cross-react with pNCC, but differs since it cannot detect pT96-NKCC2 in 129 Sv mice. Using this new antibody, we sought to clarify the effects of the WNK4-SPAK/OSR1 signaling pathway on NKCC2 phosphorylation in vivo (Fig. 12).

Figure 12.

Comparison of pT96-NKCC2 and pT53-NCC abundances with disruption or dysregulation of the WNK4-SPAK/OSR1 pathway. Top: abundance of phosphorylated NKCC2 (pT96-NKCC2) is slightly lower in with-no-lysine 4 kinase knockout (Wnk4^−/−) mice but not different in oxidative stress-responsive kinase 1 (OSR1) knockout, sterile 20/SPS-1-related proline/alanine-rich kinase (SPAK) knockout or OSR1/SPAK double knockout mice (Osr1^−/−, Spak^−/−, or Osr1^−/−/Spak^−/−, respectively), although the pT96-NKCC2/total NKCC2 ratio is lower in Osr1^−/− and Osr1^−/−/Spak^−/− mice. Cullin 3 (CUL3) and Kelch-like 3 (KLHL3) models of familial hyperkalemic hypertension (FHHt)(Cul3^+/−/Δ9 and Klhl3^−/− mice) display higher WNK4 abundance along a subset of cortical thick ascending limbs (TALs), but pT96-NKCC2 abundance is not different. Bottom: in contrast, abundance of phosphorylated NCC (pT53-NCC) is clearly lower in Wnk4^−/−, Spak^−/−, and Osr1^−/−/Spak^−/− mice and higher in FHHt models. These data suggest that dysregulation of the WNK4-SPAK/OSR1 pathway has a minor effect on NKCC2 activity. BS Type 3, Bartter type 3 syndrome; NKCC2, Na⁺-K⁺-2Cl⁻ cotransporter; NCC, Na⁺-Cl⁻ cotransporter.

Our data confirmed previous findings that Spak deletion strongly impairs NCC phosphorylation, but there are discrepant data regarding NKCC2 phosphorylation and activity in Spak^−/− mice. A major limitation has been that anti-pNKCC2 antibodies have not been validated in Nkcc2^−/− mice, which display perinatal lethality (11) and to our knowledge are no longer available. Our data using a new antibody specific to BL/6 p-T96 NKCC2 suggest only a minor effect on NKCC2 phosphorylation. In contrast, three groups including ours independently reported higher pNKCC2 abundance (7–9, 24). These studies used Spak^−/− mice on the BL/6 background, and the discrepancy is likely related to differences in antibody specificity. In all cases, the anti-pNKCC2 antibodies used were generated with antigens containing at least four of the amino acids deleted in BL/6 NKCC2 (YYLQTpFGHNTpMDAVP) (34), (TYYLQ(pT)FGHN) (7), and (YYLR(pT)FGHN(pT)MDAVPRK) (9), suggesting each can cross-react with pNCC. Although other experimental differences such as diet, water intake, and plasma electrolytes may differ, we speculate that these studies showed higher pNKCC2 due to antibody cross-reactivity. SPAK disruption dramatically lowers pNCC as confirmed using an antibody validated in NCC knockout mice on the BL/6 background (35). This would result in less competition for pNKCC2 with pNCC for binding to the nonspecific antibody, leading to the detection of a higher pNKCC2 signal and the conclusion that pNKCC2 abundance is therefore higher. Consistent with this, Lin et al. reported significantly lower NCC activity but not NKCC2 activity in Spak^−/− mice, as demonstrated by diuretic response tests (7); in our hands furosemide responsiveness trended higher but was also not significantly different (8). Contrasting these data, furosemide-responsive Na⁺ flux, i.e., NKCC2 activity, was lower in isolated microperfused TAL from Spak^−/− mice (44). Given that furosemide responsiveness was normal in intact Spak^−/− mice (7), this in vitro finding may reflect the loss of circulating hormonal factors required to stimulate alternative pathways that activate NKCC2 in the absence of SPAK, such as vasopressin. This in vitro study did not assess NKCC2 phosphorylation or another important determinant of NKCC2 activity, plasma membrane localization (45). Finally, in the present study we observed lower NKCC2 phosphorylation in Spak^−/− mice using the pT96/pT101 NKCC2 antibody (Supplemental Fig. S6), contradicting our prior findings (8). This may reflect lower competition between pNKCC2 and pNCC for antibody binding since Western blot conditions including protein loading, gel type, and antibody dilution differed (8, 24). We also found that induction of Osr1 deletion in adult mice had no significant on effect NKCC2 phosphorylation, contrasting the findings of Lin et al. who reported a significant reduction (∼30% of control abundance) of pT96-NKCC2 in constitutive renal tubule-specific Osr1^−/− mice, as well as reduced furosemide responsiveness (6). Developmental effects of constitutive versus inducible OSR1 deletion may contribute to the differential effects on NKCC2 phosphorylation. However, similar reductions of the pT96-NKCC2/NKCC2 ratio were observed, suggesting an effect of OSR1 deletion in both models. Note that Lin et al. reported higher pNCC abundance but normal thiazide responsiveness in Osr1^−/− mice (6). Similar to our speculation regarding target competition in SPAK^−/− mice, lower pNKCC2 abundance may permit the nonspecific antibody to generate a higher pNCC signal in this model. One way to resolve many of these issues would be to backcross these models to other backgrounds, but this would take significant effort. The BL/6 specific antibodies generated here and by the Loffing group (20) may help resolve these technical issues.

It is unclear why the combined deletion of SPAK and OSR1 increases total NKCC2 abundance compared with controls, but since p-T96-NKCC2 did not differ from controls, this may reflect a compensatory response. The mild phenotype of constitutive and inducible Osr1^−/− mice and the modest effects on pT96-NKCC2 abundance in inducible Osr1^−/− mice and Osr1^−/−/Spak^−/− mice suggests an alternative kinase compensates to phosphorylate NKCC2 in the absence of OSR1 and SPAK or that OSR1 is not involved in NKCC2 phosphorylation in vivo. One possible candidate for in vivo phosphorylation of NKCC2 is Traf2- and NCK-interacting kinase (TNIK), which was identified using a targeted proteomics approach (46). TNIK binds and phosphorylates NKCC2 at T96 and T101 in vitro. In vivo, pT96/pT101 NKCC2 abundance was lower in Dahl SS rats administered a TNIK inhibitor (KY-05009) (46) and in Tnik^−/− mice (47). To attempt to eliminate potential compensatory effects of other kinases, we determined pT96-NKCC2 abundance after short-term Osr1 disruption. After 5 days of treatment with doxycycline, we observed a similar decrease in the ratio of pT96-NKCC2/NKCC2 but no change in pT96-NKCC2 abundance in Osr1^−/− mice (Supplemental Fig. S10), suggesting another kinase rapidly compensates following Osr1 deletion. This may explain why the phenotype of KS-Osr1^−/− mice is much milder than that of Slc12a1^−/− mice. In contrast, no kinase can compensate along the DCT in Spak^−/− mice, resulting in a phenotype closely resembling that of Slc12a3^−/− mice.

We previously showed NKCC2 activity was only mildly lower in Wnk4^−/− mice, but a strong reduction of NKCC2 phosphorylation was detected using the anti-pT96/pT101 NKCC2 antibody (22). Here we observed a trend to but no significant reduction in NKCC2 phosphorylation using the new pT96-NKCC2 antibody, which is consistent with the mild reduction in acute natriuresis in response to furosemide administration (22). Furthermore, pT96-NKCC2 expression was preserved along a subset of cortical TALs, while it was mainly reduced along medullary TAL. Decreased NKCC2 activation in the renal medulla is likely to reduce urinary concentrating ability by reducing the cortico-medullary osmotic gradient, which along with reduced Na⁺ reabsorption along the DCT due to lower NCC activity may contribute to the elevated urine volume in Wnk4^−/− mice (21).

Consistent with relatively little effect on pT96-NKCC2, phosphorylation of SPAK and OSR1 were only partially reduced by Wnk4 disruption and pSPAK/pOSR1 remained localized to the apical membrane along TAL. The detection of apical WNK1 signal along TAL in Wnk4^−/− mice suggests WNK1 may compensate for a lack of WNK4. Interestingly, phosphorylation of SPAK and OSR1 was mainly lower along medullary TAL, although apical WNK1 signal was still detected in both medulla and cortical TAL. One possible explanation for this difference may be differences in extracellular chloride concentration between the cortex and medulla. Intracellular chloride inhibits WNK autophosphorylation, and thus activity, via direct binding to the catalytic domain, impairing downstream phosphorylation of SPAK and OSR1 (48). Recently, ex vivo experiments from Moser and colleagues showed that higher extracellular concentration (110 mM Cl⁻ vs. 5 mM Cl⁻) lowers pT96-NKCC2 phosphorylation (20). Since chloride concentration is quite high in the interstitium of the renal medulla (49), the high extracellular chloride may attenuate the phosphorylation of SPAK and OSR1 by WNK1.

We have previously reported higher phosphorylation of NKCC2 in Cul3^+/−/Δ9 mice using the anti-pT96/pT101 NKCC2 antibody and suggested hyperactivation of NKCC2 may contribute to the more severe disease caused by CUL3-Δ9. However, our data using the new pT96-NKCC2 antibody showed that NKCC2 phosphorylation is not different from controls in Cul3^+/−/Δ9 mice. NKCC2 phosphorylation was also unaltered in Klhl3^R528H/R528H or Klhl3^−/− mice that model recessive forms of FHHt (41). This is consistent with the mounting evidence that the effects of CUL3-Δ9 on the vasculature contribute to the more severe phenotype of Cul3 mutations rather than purely renal effects (14). In contrast to pT96-NKCC2, WNK4 abundance was higher along a subset of cortical TALs and along all DCT tubules in CUL3 and KLHL3 FHHt models, suggesting the CUL3-KLHL3 complex regulates WNK4 abundance along not only DCT but also cortical TAL. However, pSPAK/pOSR1 abundance was increased only along the DCT, but not along cortical TAL in Klhl3^−/− mice. This correlated with no change in pT96-NKCC2 abundance. Apical WNK1 signal was observed along the apical membrane of TAL in controls, which was unchanged by Klhl3 disruption. Since WNK4 expression is lower along cortical TAL compared with medullary TAL, WNK1 may play a more important role in regulating NKCC2 phosphorylation along cortical TAL, although this requires further testing using TAL WNK1 knockout, once appropriate animal models are available. However, though the expression of both WNK1 (mainly L-WNK1) and WNK4 was weak along cortical TAL under normal conditions, higher WNK4 abundance in the FHHt models could not enhance phosphorylation of SPAK/OSR1 and NKCC2. One possible explanation is that high WNK abundance may be not required to maximally promote NKCC2 phosphorylation along cortical TAL compared with medullary TAL because the extracellular chloride concentration is much lower in the cortex than in the medulla. This may be supported by the clear effect of Wnk4 disruption on NCC phosphorylation. WNK4 expression is much lower along DCT compared with medullary TAL under normal conditions, but the expression is enough to activate NCC phosphorylation. Furthermore, it is also likely that SPAK2 exerts a dominant-negative effect on FL-SPAK/OSR1 and hence NKKC2 phosphorylation along the TAL (8, 36), which may blunt the effects of increased WNK4 abundance because SPAK2 abundance was also increased in Klhl3^−/− mice. Thus, our data suggest that while both L-WNK1 and WNK4 promote phosphorylation of SPAK/OSR1 and NKCC2 along cortical TAL and WNK4 abundance is increased along a subset of cortical TALs in models of FHHt, this does not lead to downstream phosphorylation of SPAK/OSR1 and thus NKCC2. This may explain why NCC but not NKCC2 is activated in FHHt, and why FHHt patients are exquisitely sensitive to thiazide diuretics.

The effects of disruption or dysregulation of WNK4-SPAK/OSR1 pathway on NKCC2 phosphorylation are mild compared with effects on NCC phosphorylation (Fig. 12), suggesting the contribution of several pathways to NKCC2 phosphorylation. It has been reported that NKCC2 phosphorylation at T96 and T101 may result from the convergence of several pathways mediated by protein kinase A (PKA), NO synthase 3 (NOS3), and the phosphatase calcineurin, in addition to the WNK4-SPAK/OSR1 pathway (45). NKCC2 reabsorbs 20% of filtered Na⁺ along the TAL, which is much greater than the 5–10% of Na⁺ reabsorption via NCC. Thus, NKCC2 dysregulation causes a severe phenotype (11) and may be the reason why NKCC2 phosphorylation is more tightly regulated with multiple compensatory pathways compared with NCC.

There are limitations to our study that will need to be addressed with further experiments. First, sexual dimorphism has been reported for expression levels of several ion transport pathways including total NCC and pNCC (50, 51) and total NKCC2 (52, 53). While properly powered studies have not been performed to assess sex differences in the various models made by us or others, our data confirm previous findings that disruption of the CUL3-WNK-SPAK pathway has a large effect on pNCC abundance. Furthermore, FHHt does not present with any obvious sex differences with regard to disease severity (54). Similarly, despite differences in baseline expression, responses of total and pNCC to angiotensin II-induced hypertension (55) or furosemide administration (50) are similar. While we cannot exclude the existence of sex differences in each mouse model, we believe any that exist are likely to be quite small. Critically for the present study, in each model used, large effects on NCC phosphorylation were observed while effects on NKCC2 phosphorylation were minor in the same animals, regardless of sex. This supports our central conclusion that dysregulation of the CUL3/KLHL3-WNK-SPAK pathway plays a greater role in the regulation of NCC phosphorylation than it does NKCC2 phosphorylation. Another important limitation is that we only assessed differences in NKCC2 phosphorylation under baseline conditions. The regulatory pathway may become more important under certain conditions including low sodium or water deprivation. In this regard, our previous study assessing the effects of vasopressin on NKCC2 phosphorylation in Spak^−/− mice on the BL/6 background (56) was limited by the use of the anti-pT96/pT101 NKCC2 antibody now known to be nonspecific. Finally, while we assessed NKCC2 activity in vivo in Klhl3^−/− mice, for other models we only quantified pT96 NKCC2 phosphorylation. NKCC2 activity is also stimulated by SPAK/OSR1-independent phosphorylation at S126 (5) and by cAMP-mediated trafficking to the plasma membrane (45). Therefore, changes in pT96 phosphorylation may not be equivalent to NKCC2 activity. Indeed, compared with NKCC1, NKCC2 activity is relatively less stimulated by changes in SPAK/OSR1-mediated phosphorylation in vitro (57).

Perspectives and Significance

Our data suggest that disruption of the WNK4-SPAK/OSR1 pathway has a minor effect on NKCC2 phosphorylation in the kidney (Fig. 12), suggesting compensatory regulation of NKCC2 by other pathways maintains the sodium, divalent cation, and fluid homeostasis. This is in contrast to the strong regulation of the WNK4-SPAK pathway on NCC activity, which can be rapidly changed in accordance with plasma potassium concentration influencing intracellular chloride concentration (37, 38). Despite a modest effect on NKCC2 phosphorylation, the reduction in activity of both NKCC2 and NCC may explain why Wnk4^−/− mice recapitulate the phenotype of Bartter type 3 syndrome, which presents with hypokalemic metabolic alkalosis but normocalciuria. The combined dysfunction of calcium handling along TAL and DCT likely results in an overall neutral effect on urinary calcium excretion. Finally, WNK4 abundance is higher along several cortical TAL in CUL3 and KLHL3 FHHt model, but this does not increase phosphorylation of SPAK/OSR1 and NKCC2 and may explain why FHHt patients are highly responsive to administration of thiazide diuretics.

DATA AVAILABILITY

Data will be made available upon reasonable request.

SUPPLEMENTAL DATA

Supplemental Figs. S1–10: https://doi.org/10.6084/m9.figshare.22557709.v1.

Supplemental Table S1: https://doi.org/10.6084/m9.figshare.22557724.v1.

Supplemental Table S2: https://doi.org/10.6084/m9.figshare.22564036.v1.

GRANTS

Y.M. received a postdoctoral award from the Uehara Foundation, R.J.C. is funded by National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) Grants DK120790 and DK137034, M.C.-B. is funded by CONACyT (Mexico) Grant 101720, G.G. is funded by DGAPA-UNAM IN203422, and J.A.M. is funded by NIDDK Grant DK098141.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

Y.M. and J.A.M. conceived and designed research; Y.M., L.T.N., A.S., R.J.C., X.-T.S., M.R.G., H.C.-C., M.C.-B., and J.A.M. performed experiments; Y.M., A.S., and J.A.M. analyzed data; Y.M., and J.A.M. interpreted results of experiments; Y.M. prepared figures; Y.M., and J.A.M. drafted manuscript; Y.M., H.C.-C., M.C.-B., G.G., and J.A.M. edited and revised manuscript; Y.M., L.T.N., A.S., R.J.C., X.-T.S., M.R.G., H.C.-C., M.C.-B., G.G., and J.A.M. approved final version of manuscript.