A mouse model of pseudohypoaldosteronism type II reveals a novel mechanism of renal tubular acidosis

Pseudohypoaldosteronism type II (PHAII) is a genetic disease characterized by association of hyperkalemia, hyperchloremic metabolic acidosis, hypertension, low renin


P
seudohypoaldosteronism type II (PHAII), also known as Gordon syndrome or familial hyperkalemic hypertension, is a rare genetic disease characterized by the association of hypertension, hyperchloremic metabolic acidosis, and hyperkalemia in the absence of renal failure. 1 The pathogenesis of this syndrome has remained puzzling for decades because hyperkalemia along with renal metabolic acidosis are 2 features of hypoaldosteronism, whereas volume-dependent hypertension is observed in states caused by an excess of mineralocorticoids.A study by Schambelan  et al. suggests that the disease is caused by excessive electroneutral absorption of chloride by the distal nephron. 2 The observation that most symptoms of PHAII can be corrected by thiazide diuretics supports the possibility that a gain of function of the thiazide-sensitive sodium chloride (NaCl) cotransporter NCC of the distal convoluted tubule (DCT) plays a central role in PHAII.A breakthrough in our understanding of this disease came in 2001 from the discovery that mutations in genes encoding 2 serine-threonine kinases of the WNK family, WNK1 and WNK4, account for PHAII in some patients. 3WNK1 and WNK4 are elements of a complex signaling network involved in the regulation of ion transport in the distal nephron. 4PHAII mutations in the WNK1 gene are large intronic deletions that cause an increase in WNK1 expression, whereas WNK4 mutations are missense mutations clustered in a short acidic segment distal to the kinase domain. 3These missense mutations also result in increased expression of WNK4 due to a decrease in degradation of the protein. 5The ubiquitin-ligase complex formed by the proteins KLHL3 and CUL3 has been identified as responsible for this degradation, 4 and mutations in both proteins have been recently identified in patients with PHAII. 6,7ogically, the studies conducted over the past 15 years have focused on the regulation of NCC by WNK1 and WNK4.Transgenic mice expressing a WNK4 transgene carrying 1 of the missense mutations identified in PHAII patients (WNK4-Q562E; TgWnk4 PHAII mice) displayed all clinical features of the syndrome. 8The same phenotypes were seen in a Wnk4(D561A/þ) knock-in mouse model 9 and in Wnk1 þ/FHHt mice, another PHAII model in which the first intron of Wnk1 is deleted. 10In these models, increased NCC phosphorylation and expression at the apical cell membrane result in enhanced NaCl absorption by the DCT, and vascular volume expansion, favoring the onset of hypertension.Conversely, Wnk4 knockout mice exhibited a mild Gitelman-like syndrome, with normal blood pressure, increased plasma renin activity, and reduced NCC expression and phosphorylation. 11he aforementioned studies confirm the importance of NCC in the pathogenesis of PHAII syndrome.However, other studies have suggested that increased NCC activity or abundance is not sufficient to cause PHAII.In transgenic mice overexpressing NCC, NCC abundance was increased but the mice displayed a phenotype similar to wild-type mice. 12dditionally, mice inactivated for KS-WNK1 or Nedd4-2 did not exhibit hyperkalemic hypertension despite increased NCC expression and phosphorylation. 13,14n the kidney, WNK4 is expressed not only in the DCT, but has also been found in the cortical collecting duct (CCD). 3,15,16Alongside the amiloride-sensitive epithelial sodium channel ENaC, which is expressed by principal cells, we have demonstrated the presence of a novel thiazide-sensitive electroneutral NaCl uptake mechanism in renal intercalated cells (ICs) of the CCD. 17NaCl is taken up from urine by the coordinated action of the Cl -/HCO 3 -exchanger pendrin/ SLC26A4 and the sodium ion (Na þ )-driven Cl -/2HCO 3 exchanger NDCBE/SLC4A8.Overexpression of pendrin in ICs can favor the onset of chloride-dependent hypertension, indicating that this system likely plays a role in the regulation of blood pressure in vivo. 18Conversely, pendrin-deficient mice are protected against mineralocorticoid-induced hypertension 19 and develop a lower blood pressure during NaCl restriction. 20Importantly, acute inactivation of pendrin results in a lower blood pressure. 21Finally, we demonstrated that NCC compensated for NDCBE inactivation and that double deletion of these proteins in mouse induces hypokalemia. 22Therefore, we hypothesized that this novel system might be involved in the pathogenesis of PHAII.
In the present study, we demonstrate that PHAII-causing mutation of WNK4 increases pendrin activity and pendrin/ NDCBE-dependent NaCl absorption.We also demonstrate that pendrin hyperactivity is necessary to drive hyperchloremic metabolic acidosis and hyperkalemia in PHAII because both symptoms are corrected by pendrin genetic ablation.

TgWnk4 PHAII mice exhibit hyperkalemic metabolic acidosis
As shown in Figure 1a-e, TgWnk4 PHAII mice exhibit the phenotypic abnormalities characteristic of PHAII: metabolic hyperchloremic acidosis and hyperkalemia.Urinary aldosterone excretion was significantly increased in TgWnk4 PHAII mice (Figure 1f), whereas renin mRNA abundance was lower in the PHAII mice (Figure 1g) as observed in most patients experiencing PHAII.As shown by others, 8,9 NCC and phosphor-T53 NCC protein abundance was dramatically increased in TgWnk4 PHAII mice (Figure 1h).
Electroneutral thiazide-sensitive NaCl transport is activated in the CCD of TgWnk4 PHAII mice Transepithelial fluxes of Na þ (J Na ), potassium ion (K þ ; J K ), chloride ion (Cl -; J Cl ) and the transepithelial voltage (V te ) were measured in microperfused CCDs isolated from TgWnk4 PHAII mice fed a standard diet (Figure 2).No transport activity was detectable in CCDs isolated from control mice, as previously described. 17However, CCDs isolated from TgWnk4 PHAII mice exhibited net NaCl absorption but did not develop a significant lumen negative V te and did not secrete K þ , suggesting the absence of detectable ENaC activity.Furthermore, NaCl absorption was fully inhibited by luminal addition of 100 mM hydrochlorothiazide (HCTZ).
Taken together, these experiments demonstrate that electroneutral NaCl transport mediated by pendrin/NDCBE is stimulated by PHAII-causing mutation of WNK4.
ENaC-dependent K D transport mechanisms are impaired in TgWnk4 PHAII mice despite unchanged renal ENaC activity ENaC activity is generally detectable not in the CCD but only in the connecting tubule (CNT), a nephron segment, which is not suitable for in vitro microperfusion.Thus, to estimate the total ENaC activity, we next measured the effect of an acute amiloride injection in TgWnk4 PHAII and control mice on urinary Na þ and K þ excretion.The natriuretic response to amiloride was identical for both groups (Figure 3a, left panel), suggesting that Na þ reabsorption through ENaC is not altered in TgWnk4 PHAII mice.However, amiloride had significantly less effect on urinary K þ excretion in TgWnk4 PHAII mice than in controls (Figure 3a, right panel).These experiments demonstrate that ENaC-dependent K þ transport mechanisms per se are inhibited in TgWnk4 PHAII mice.
We next examined the relative protein abundance of the a and g subunits of ENaC in TgWnk4 PHAII and control mice by immunoblot analyses of plasma membrane-enriched preparations isolated from renal cortex (Figure 3b).The abundance of the full-length 90 kDa form and the cleaved N-terminal 30 kDa fragment of a-ENaC was higher in TgWnk4 PHAII mice.The abundance of the full-length form of g-ENaC (85 kDa) was decreased in TgWnk4 PHAII mice, whereas the abundance of the cleaved 70 kDa form of g-ENaC was increased.These alterations are indicative of an activation of ENaC possibly by aldosterone.
Normalization of plasma K concentration does not correct metabolic acidosis in TgWnk4 PHAII mice Hyperkalemia can cause renal tubular acidosis. 23To check the importance of high plasma potassium in the development of acidosis in PHAII, we tested whether normalization of plasma potassium concentration can alleviate the acidosis of TgWnk4 PHAII mice.TgWnk4 PHAII mice were fed with a synthetic low-K þ diet for 4 days (Figure 4).As expected, plasma K þ was normalized following this treatment.However, the decrease in plasma K þ did not modify either plasma HCO 3 or pH, indicating that acidosis in PHAII is not caused primarily by hyperkalemia.
Renal acidosis in TgWnk4 PHAII mice is not caused by impaired proton secretion We next evaluated whether metabolic acidosis in TgWnk4 PHAII mice is due to impaired proton secretion by the distal nephron.Acid was administrated as NH 4 Cl 0.28 M in the drinking water for 15 days.Figure 5 shows that before acid loading, only plasma HCO 3 -(panel b) was significantly lower  b a s i c r e s e a r c h in TgWnk4 PHAII mice, whereas urine pH (panel c), urinary titrable acid excretion (panel d), and ammonium excretion (panel e) were identical in TgWnk4 PHAII and control mice.After 2 days of acid loading, blood pH and HCO 3decreased in both groups.However, TgWnk4 PHAII mice developed a much stronger metabolic acidosis.When acid loading was continued, control mice were able to cope with the acid load and exhibited normal blood pH and HCO 3at day 15 of the treatment.By contrast, severe metabolic acidosis persisted in TgWnk4 PHAII mice.Interestingly, following acid loading, urine pH decreased and ammonium excretion increased maximally in both groups, indicating that proton secretion, and ammonium transport or metabolism, are normal in these mice.
The effect of acid loading was also tested in WNK1 þ/FHHt mice, another PHAII model in which the first intron of WNK1 is deleted. 10When submitted to an acid load (Supplementary Figure S1), these mice responded exactly like TgWnk4 PHAII mice, indicating that the same mechanism is causing acidosis in both PHAII models independently of the mutation.
Finally, Figure 5f shows a marked shift in the relationship between urinary ammonium excretion and blood bicarbonate concentration (reflecting the severity of metabolic acidosis) in TgWnk4 PHAII versus control mice.This demonstrated that, while proton secretion or ammonium excretion capabilities are preserved in TgWnk4 PHAII mice, these mice clearly have a defect in renal net acid excretion.This led us to hypothesize that metabolic acidosis in PHAII is not caused by impaired acid excretion but rather is due to increased renal loss of base.

Pendrin activity is increased in PHAII-mutant WNK4 mice
We next examined pendrin activity by measuring Cl -dependent alkalinization in ICs in CCDs isolated from TgWnk4 PHAII and control mice (Figure 5g, left panel).The rate of intracellular alkalinization in response to luminal Cl - removal, an estimate of apical Cl -/HCO 3 -exchange activity, was much higher in ICs from TgWnk4 PHAII mice than in control mice (Figure 5g, right panel).These results indicate that pendrin activity per cell is increased in this model.
Further, immunofluorescence staining revealed an increase in pendrin labeling in the collecting ducts located in renal medullary rays of TgWnk4 PHAII mice (Figure 5h).ICs exist in 3 different forms: (i) a-ICs characterized by apical expression of the V-ATPase and basolateral expression of an anion exchanger, which is an alternately spliced product of the erythrocyte AE1 gene, (ii) b-ICs, which harbor apical pendrin and basolateral V-ATPase, and (iii) non-a-, non-b-ICs with apical pendrin and apical V-ATPase.The total number of ICs remains generally constant; however, the relative number of a-ICs versus b-ICs is tightly controlled and adapted depending on the acid-base status. 24We therefore evaluated the number of the different versions of ICs in the CCDs of control and TgWnk4 PHAII mice using triple immunofluorescence labeling for pendrin, for the E subunit of the V-ATPase, and for AE1 (Figure 5h).In both transgenic and control mice, CCDs located in medullary rays exhibited aand b-ICs but no non-a-, non-b-ICs.We counted 1867 ICs (defined as those that stained for the E subunit of the V-ATPase) from 4 independent control mice, and the results are displayed as white bars in Figure 5i.Of these ICs, 37% stained for pendrin, whereas 63% had AE1 staining.In 4 independent TgWnk4 PHAII mice, we counted 1955 ICs (black bars in Figure 5i).The fraction of b-ICs was significantly increased in CCDs of TgWnk4 PHAII mice as 50% of ICs were pendrin-positive.Conversely, the number of a-ICs was reduced and the total number of ICs was not different Amiloride elicits significant increase in Na þ excretion and decrease in potassium ion (K þ ) excretion in both wild-type and TgWnk4 PHAII mice 6 hours after injection.Amiloride-induced natriuresis was not different between groups.K þ excretion after amiloride was significantly higher in TgWnk4 PHAII mice, indicating that ENaC-dependent K þ secretion is decreased in these mice.Values are the mean AE SEM for 8 mice.Statistical significance was assessed by 1-way analysis of variance.***P < 0.001 and ****P < 0.0001, vehicle versus amiloride.*P < 0.05, wild-type versus TgWnk4 PHAII mice following amiloride injection.(b) a-ENaC and g-ENaC protein abundance was assessed by Western blot of plasma membrane-enriched preparations obtained from the renal cortex of wild-type (n ¼ 4) and TgWnk4 PHAII (n ¼ 6) mice.Bar graph shows a summary of densitometric analyses.Densitometric values were normalized to the mean for the control group that was defined as 100%, and results were expressed as mean AE SEM.Statistical significance was assessed by unpaired t-test.*P < 0.05, **P < 0.01.To optimize viewing of this image, please see the online version of this article at www.kidney-international.org.between control and TgWnk4 PHAII mice.Because pendrin activity per cell is increased, we can conclude that total pendrin activity is increased in CCDs of mutant mice.As shown in Figure 5j, this increased pendrin transport activity was associated with a significant increase in CCD fractional volume.

b a s i c r e s e a r c h
To verify whether hyperactivity of pendrin participates to PHAII phenotype, we next tested the effects of pendrin disruption in mice harboring the mutation of WNK4 causing PHAII.Double-mutant mice (TgWnk4 PHAII ;Pds -/-) no longer exhibited hyperchloremic metabolic acidosis and hyperkalemia (Table 1).These results confirmed that activation of pendrin participates to hyperchloremic acidosis and hyperkalemia in PHAII.
Pendrin disruption in TgWnk4 PHAII mice led to a significant increase in hematocrit (Table 1).Renin mRNA was also higher in double transgenic TgWnk4 PHAII ;Pds -/-mice (Figure 6).No significant difference, however, was detected in blood pressure measured by telemetry (Figure 6) or in aldosterone levels (7.26 AE 0.57 nM/mM creatinine in TgWnk4 PHAII ;Pds -/-mice, n ¼ 10 vs. 7.93 AE 0.63 nM/mM creatinine in TgWnk4 PHAII ;Pds þ/þ mice, n ¼ 12, NS).The effects of pendrin disruption on blood pressure remained limited, presumably because NCC is still markedly stimulated.Accordingly, no change in protein abundance of NCC and phosphor-T53 NCC was observed between TgWnk4 PHAII ;Pds þ/þ and TgWnk4 PHAII ;Pds -/-mice by immunoblot analyses of plasma membrane-enriched preparations isolated from renal cortex (Figure 7a).Protein abundance of ENaC subunits was also assessed.Both a and g subunits decreased in TgWnk4 PHAII ;Pds -/-mice compared with TgWnk4 PHAII ;Pds þ/þ mice, consistent with the effect of isolated deletion of pendrin on ENaC expression described previously. 25e next measured total ENaC-dependent Na þ absorption and K þ secretion in double-mutant mice by assessing the effects of an acute injection of amiloride on urinary excretion of Na þ and K þ .The natriuretic response to amiloride (Figure 7b, left panel) was not different between TgWnk4 PHAII ;Pds þ/þ and TgWnk4 PHAII ;Pds -/-mice, indicating that ENaC activity is not altered by pendrin disruption.However, the decrease in urinary K þ excretion following amiloride was higher in TgWnk4 PHAII ;Pds -/-mice than in TgWnk4 PHAII ;Pds þ/þ mice (Figure 7b, right panel), suggesting that ENaC-dependent K þ secretion is increased after pendrin disruption in TgWnk4 PHAII mice.

DISCUSSION
Studies performed either in humans or in animal models have confirmed that PHAII-causing mutations lead to increased phosphorylation, membrane expression, and activity of the cotransporter NCC of the DCT cells.

b-ICs exchange HCO 3
-for Cl -through the apical Cl -/HCO 3 -exchanger pendrin/SLC26A4, fulfilling their primary function of HCO 3 -excretion.When pendrin is functionally coupled with the Na þ -driven Cl -/2HCO 3 -exchanger NDCBE/SLC4A8, thiazide-sensitive, electroneutral NaCl absorption then occurs in these cells. 17We therefore performed this study to determine to what extent pendrin might participate to the PHAII phenotype.
Here we show that pendrin activity and thiazide-sensitive electroneutral NaCl absorption through the pendrin/NDCBE transport system are activated in isolated CCDs of TgWnk4 PHAII mice.Elucidating the mechanism of pendrin activation was beyond the scope of this study.However, several studies in the literature support the idea that WNK4 is involved in pendrin activation.For instance, the pendrin/NDCBE transport system is the dominant mechanism accounting for Na þ absorption in the CCD of Ncc knockout mice, 17 whereas in Wnk4 knockout mice, which exhibit dramatically low NCC expression, Na þ absorption in CCDs does not occur through the pendrin/NDCBE system but through ENaC. 11Shibata et al. identified a phosphorylation site at S843 in the mineralocorticoid receptor (MR S843-P ) exclusively in ICs, which prevents ligand binding. 16MR S843-P dephosphorylation restores MR activation by aldosterone and increases pendrin expression.TgWnk4 PHAII mice (gain of function of WNK4) have lower levels of MR S843-P , whereas Wnk4 knockout mice showed increased levels of MR S843-P .Altogether, these studies strongly support the role of WNK4 in pendrin regulation through the modulation of MR phosphorylation.
Pendrin deletion in TgWnk4 PHAII mice decreases signs of hypervolemia.Renin expression and hematocrit were mice ( ## P < 0.01 and #### P < 0.0001).No significant difference was observed between control and TgWnk4;Pds -/-mice.significantly increased in pendrin-deleted TgWnk4 PHAII mice compared with TgWnk4 PHAII mice, although not to the levels seen in control mice.There was a trend of decreasing systolic blood pressure in pendrin deleted TgWnk4 PHAII mice compared with TgWnk4 PHAII mice, but the difference (5 mm Hg) did not reach statistical significance, possibly because of the small number of mice per group.It is worth noting that, in the study by Lalioti et al., the increase in systolic blood pressure in TgWnk4 PHAII mice compared with wild-type mice was about the same magnitude at 7 mm Hg. 8 Nevertheless, these results indicate that NCC hyperactivation is sufficient to drive hypertension in PHAII.As with NCC, 8 genetic deletion of pendrin in PHAII mice normalized blood K þ concentration.Interestingly, overexpression of either NCC or pendrin has no effect on blood K þ concentration. 12,18Thus, concomitant activation of NCC and pendrin is necessary to drive hyperkalemia in this model.Conversely, combined deletion of NCC and NDCBE, the partner of pendrin in ICs, induces hypokalemia, 22 while isolated deletion of either NCC or pendrin in the mouse has no or slight effect on blood K þ concentration under standard conditions. 19,20,26,27These findings indicate that NCC and pendrin act in concert to control plasma K þ concentration.9][30] Decreased plasma K þ concentration was also shown to promote pendrin induction by aldosterone, a mechanism that has been proposed to counteract the progression of hypokalemia. 31Our study suggests that these 2 mechanisms of protection against hypokalemia become aberrantly activated in PHAII mice and thereby participate in the generation of hyperkalemia.Increased activity of both NCC and the pendrin/NDCBE system is expected to favor the electroneutral NaCl reabsorption in the distal nephron at the expense of the electrogenic Na þ /K þ exchange promoted by ENaC, therefore leading to K þ retention.
We found that ENaC-dependent K þ secretion was decreased in TgWnk4 PHAII mice and restored when pendrin was deleted.However, we were not able to detect any change in amiloride-induced urinary Na þ excretion (reflecting total ENaC activity) between control, TgWnk4 PHAII , and pendrindeleted TgWnk4 PHAII mice.Expected changes in ENaC protein expression were nevertheless observed between these mice.This suggests that ENaC activity is modulated by other factors such as HCO 3 -or ATP as previously demonstrated. 25,32Altogether, these findings indicate that a decrease in K þ channel activity per se is involved in the pathogenesis of hyperkalemia in this model.A decrease in renal outer medullary potassium channel (ROMK) expression in the DCT2/ CNT was previously reported in a PHAII mouse model caused by a WNK1 mutation. 10 ) and TgWnk4;Pds -/-(N ¼ 5) mice.Amiloride elicits significant increase in Na þ excretion and decrease in K þ excretion in both TgWnk4;Pds þ/þ and TgWnk4;Pds -/-mice 6 hours after injection.Amiloride induced natriuresis was not different between groups.K þ excretion after amiloride was significantly lower in TgWnk4;Pds -/-mice, indicating that epithelial sodium channel (ENaC)dependent K þ secretion is increased in these mice.Values are the mean AE SEM.Statistical significance was assessed by 1-way analysis of variance.**P < 0.01 and ****P < 0.0001, vehicle versus amiloride.****P < 0.0001, TgWnk4;Pds þ/þ versus TgWnk4;Pds -/-mice following amiloride injection.To optimize viewing of this image, please see the online version of this article at www.kidney-international.org.b a s i c r e s e a r c h A recent study by Grimm et al. proposed that tubular remodeling in response to NCC activation explains renal potassium retention in PHAII. 33In this latter study, the authors showed that constitutive activation of STE20/SPS1-related proline/alanine-rich kinase in the DCT (CA-SPAK), which leads to PHAII phenotype, induces a marked increase in DCT1 mass, with a commensurate reduction in the CNT mass resulting in a dramatic decrease in both ENaC and ROMK expression.This latter model significantly differs from the one studied here.We observed an increase in CCD mass that is paralleled by an increase in pendrin activity with only a modest decrease in CNT mass (CNT fractional volumes were 5.16% AE 0.54% vs. 3.41% AE 0.36% of kidney cortex in control and TgWnk4 PHAII mice, respectively (n ¼ 4 for each group, P ¼ 0.036), which likely explains the decrease in ENaC and ROMK currents seen in the DCT2/early CNT of TgWnk4 PHAII mice. 34However, the observation that pendrin deletion normalized blood K þ concentration in TgWnk4 PHAII mice indicates that remodeling of the CNT is not the primary cause of K þ retention in our model.In summary, both models (CA-SPAK and WNK4-PHAII) are characterized by the combination of NCC activation along with a second mechanism that differs from a model to another but that results ultimately in a decrease in K þ secretion.
In vitro studies have shown direct inhibitory effect of WNK4 on ROMK or BK. 5,35However, because decreased urinary K þ secretion in TgWnk4 PHAII mice is reversed by pendrin deletion, it is unlikely that hyperkalemia in PHAII is the consequence of direct inhibition of ROMK and/or BK by WNK4.Changes in acid-base status affect potassium secretion in the distal nephron; specifically, metabolic acidosis inhibits whereas metabolic alkalosis stimulates K þ secretion. 36onsistent with this observation, it has been demonstrated that intracellular pH is an important modulator of the native low-conductance K þ channel in CCD, 37,38 such that acidification within the physiological range results in a reduction of single channel activity.Thus, correction of the acidosis following pendrin deletion would remove the acidic pH inhibition of ROMK activity.The reversion of hyperkalemia by genetic ablation of pendrin in TgWnk4 PHAII mice could also be due to enhanced BK activity in ICs.The BKb4 subunit is present with the BKa subunit in ICs as well as L-WNK1, which activates BK. 39 Chloride binding to WNK1 precludes its autophosphorylation and thereby inhibits kinase activity. 40t is thus possible that inactivation of pendrin by decreasing intracellular Cl -concentration activates L-WNK1 selectively in ICs, which in turns activates BK.
Finally, one of the most exciting findings is the identification of pendrin hyperactivity as a novel mechanism for renal acidosis.Indeed, our experiments shown in Figures 4 and 5 clearly demonstrated that the metabolic acidosis is not caused primarily by hyperkalemia or by defective V-ATPase activity.Indeed, (i) metabolic acidosis was not significantly alleviated by the normalization of blood K þ concentration; and (ii) ammonium excretion and proton secretion were not impaired, and the mice were still able to achieve maximal urine acidification and ammonium excretion under stimulated condition.By contrast, our results showing that pendrinmediated Cl -/HCO 3 -exchange per cell as well as the total number of pendrin-expressing cells are markedly increased while pendrin genetic ablation completely corrects metabolic acidosis in TgWnk4 PHAII mice indicate that metabolic acidosis is caused by excessive bicarbonate secretion by the distal nephron.
Activation of pendrin in TgWnk4 PHAII mice is accompanied by an alteration of the structure of the CCD.Indeed, we show an increase in b-ICs at the expense of a-ICs in the CCDs of TgWnk4 PHAII mice.Several studies report that acidosis provokes the differentiation of a-ICs from b-ICs, but only one other study, by Grimm et al., observes that the IC conversion works in the opposite direction in a Spak knockout mouse model, 41 which has very low level of NCC. 42n increased number of pendrin-positive IC cells was also observed in Ncc knockout mice as the result of hyperplasia of the CNT seen in these mice. 43,44t might be puzzling, if metabolic acidosis is caused by a distal renal leak of bicarbonate, that TgWnk4 PHAII did not exhibit overt bicarbonaturia.6][47][48] If not, these patients would not be in steady state and would die from uncontrolled acidosis.Generally, the renal leak of bicarbonate can be unmasked when blood bicarbonate concentration is normalized after massive infusion of bicarbonate.However, we were not able to perform this experiment in our model as the mice tolerated bicarbonate infusion very poorly and died during the bicarbonate infusion.
Interestingly, our observation might not be limited to PHAII patients.Indeed, it has been recently demonstrated that calcineurin inhibitors, which are extensively used as immunosuppressive agents in organ transplantation, alter the WNK signaling pathway, and thereby can activate NCC causing hypertension. 49These patients do not only exhibit hypertension but are also frequently affected by hyperkalemia and metabolic acidosis. 50,51It is tempting to speculate that pendrin hyperactivity will be also detected in these patients and can account for these symptoms.
In conclusion, our study demonstrates a role of pendrin in acid-base and potassium homeostasis, and identifies a renal leak of bicarbonate by the distal nephron as a novel type of renal tubular acidosis.We propose to define this novel mechanism of renal acidosis as type 5 RTA.[51]

Physiological studies
All experiments were conducted using male mice (3-5 months old) and performed in accordance with the relevant guidelines of the French Ministry of Agriculture (Authorization Executive Order B751532) for scientific experimentation on animals, the European Communities Council Directive, and international ethical standards.
For urine collection, individual mice were studied in metabolic cages (Tecniplast France, Lyon, France).Mice were pair-fed with standard laboratory powdered chow containing 0.3% sodium (INRA, Jouy-en-Josas, France).After 3 to 5 days of adjustment, 24-hour urine was collected under mineral oil.
Blood was sampled from the retro-orbital sinus of isofluoraneanesthetized mice.

Blood parameters and urinary analyses
Blood analyses were performed on an ABL 77 pH/Blood-Gas Analyzer (Radiometer, Copenhagen, Denmark).Blood bicarbonate concentration was calculated from the measured values using the Henderson-Hasselbach equation.Urinary Na þ and K þ electrolytes were measured with a flame photometer (IL943; Instruments Laboratory, Lexington, MA).Urinary pH was measured with a pH/ blood-gas analyzer (ABL 555, Radiometer).Urinary NH 4 þ and titratable acid were measured by titration with a DL 55 titrator (Mettler Toledo, Viroflay, France).Urine aldosterone was measured by RIA (DPC Dade Behring/Siemens Healthcare, Erlangen, Germany).All urinary values were adjusted by urinary creatinine previously quantified by Jaffé colorimetric method.

Measurements of blood pressure by radiotelemetry
The catheter of the telemeter was inserted into the left femoral artery.The transmitter probe was positioned subcutaneously on the flank.After a 1-week recovery period in individual cages, mice were placed on a receiver and blood pressure (BP) and locomotor activity were recorded continuously in freely moving mice, in a light/darkcycled recording room (7 a.m. to 7 p.m.), for 3 consecutive days on a normal diet as previously described. 21 vitro microperfusion of CCDs CCDs were isolated and microperfused in vitro as previously reported. 17,18,32,53munofluorescence studies and immunoblot analyses Kidneys were fixed in situ by retrograde perfusion of the aorta with a solution of 4% PFA in phosphate buffer.Harvested kidneys were washed in ice-cold phosphate buffer for 30 minutes before freezing in cold isopentane.Then 4-mm cryosections were stained with primary antibodies (a rabbit anti-pendrin diluted 1:200, a guinea pig anti-AE1 diluted 1:5,000, and a chicken anti-Atp6v1e1, which detects the V-ATPase, diluted 1:500), and secondary antibodies (goat antirabbit Alexa 555 diluted 1/800 (Invitrogen, Carlsbad, CA), donkey anti-guinea pig Cy5 diluted 1:2,000 (Jackson ImmunoResearch Laboratories, West Grove, PA), goat anti-chicken Alexa 488 diluted 1/2000 (Invitrogen).Images were acquired with a Zeiss LSM 710 laser scanning microscope (Carl Zeiss France, Marly-le-Roi, France).Fractional volume of CCDs was measured as described previously. 54r Western blot analyses, 15 mg of the plasma membrane-enriched protein fraction isolated from renal cortex was separated on reducing 7.5% or 10% SDS-polyacrylamide gels.Protein loading was assessed on gels run in parallel and stained with Coomassie blue. 55Blots were probed with anti-NCC, 1:10,000; anti-NCC phospho-Thr53; 1:10,000; anti-a-ENaC, 1:5,000; anti-g-ENaC, 1:10,000; and antipendrin, 1:3,000.For a complete description of all the study methods, see the Supplementary Methods.

DISCLOSURE
All the authors declared no competing interests.

Figure 1 |
Figure 1 | TgWnk4 PHAII mice exhibit phenotypic abnormalities characteristic of PHAII.(a-e) Blood gas and electrolytes were measured in wild-type (white bars) and TgWnk4 PHAII (black bars) mice.Values are represented as mean AE SEM, n ¼ 8 mice per group.Statistical significance was assessed by unpaired Student t-test.*P < 0.05, **P < 0.01, and ***P < 0.001.(f) Urinary aldosterone (U aldo.) excretion was measured in wild-type (white bars) and TgWnk4 PHAII (black bars) mice.Data are presented as mean AE SEM, n ¼ 8 for control mice and n ¼ 7 for TgWnk4 PHAII mice.Statistical significance was assessed by unpaired Student t-test.*P < 0.05.(g) Renin gene expression was analyzed by quantitative reverse-transcription polymerase chain reaction in whole kidney of wild-type (white bars) and TgWnk4 PHAII (black bars) mice.Results are expressed in arbitrary units relative to the expression of a geometrical mean of 4 housekeeping genes: ubiquitin, HPRT, 18s rRNA, and GAPDH.The control group has been arbitrarily set to 1. Values are represented as mean AE SEM, n ¼ 6 mice per group.Statistical significance was assessed by unpaired Student t-test.*P < 0.05.(h) Immunoblots for sodium chloride cotransporter (NCC) and its T53 phosphorylated residue (pNCC) on plasma membrane-enriched fractions from renal cortex of TgWnk4 PHAII (n ¼ 6) and wild-type (n ¼ 4) mice.Bar graph shows a summary of densitometric analyses.Densitometric values were normalized to the mean for the control group that was defined as 100% and results were expressed as mean AE SEM.Statistical significance was assessed by unpaired Student t-test.**P < 0.001.To optimize viewing of this image, please see the online version of this article at www.kidney-international.org.

Figure 2 |
Figure 2 | Electroneutral sodium chloride transport system is activated in the cortical collecting duct (CCD) of TgWnk4 PHAII mice and abolished by luminal HCTZ.Na þ (JNa), Cl -(JCl), and K þ (JK) transepithelial fluxes and transepithelial voltage (Vte) were measured in microperfused CCDs isolated from wild-type mice (white bars, N ¼ 6) and TgWnk4 PHAII mice (black bars, N ¼ 5) fed a standard 0.3% sodium diet.Transepithelial ion fluxes and voltage were also measured in the presence of 100 mM hydrochlorothiazide in the tubular lumen of CCDs isolated from TgWnk4 PHAII mice (striped bars, N ¼ 6).Statistical significance was assessed by an unpaired Student t-test.*P < 0.05.

Figure 3 |
Figure 3 | Epithelial sodium channel (ENaC) expression and activity in TgWnk4 PHAII mice.(a) Effect of amiloride injection on urinary Na þ and K þ excretion in wild-type and TgWnk4 PHAII mice.Amiloride elicits significant increase in Na þ excretion and decrease in potassium ion (K þ ) excretion in both wild-type and TgWnk4 PHAII mice 6 hours after injection.Amiloride-induced natriuresis was not different between groups.K þ excretion after amiloride was significantly higher in TgWnk4 PHAII mice, indicating that ENaC-dependent K þ secretion is decreased in these mice.Values are the mean AE SEM for 8 mice.Statistical significance was assessed by 1-way analysis of variance.***P < 0.001 and ****P < 0.0001, vehicle versus amiloride.*P < 0.05, wild-type versus TgWnk4 PHAII mice following amiloride in-

Figure 4 |
Figure 4 | Low-potassium ion (KD) diet decreases blood plasma potassium concentration in TgWnk4 PHAII mice but does not modify blood pH or blood HCO 3-.Blood gas parameters were measured in 7 TgWnk4 PHAII mice fed a normal diet and after 4 days on a low-K þ diet.Statistical analysis was assessed by paired t-student.NS, not significant; P > 0.05.

Figure 5 |
Figure 5 | TgWnk4 PHAII mice display renal tubular acidosis due to increased apical Cl/HCO 3 exchange activity in b-intercalated cells.
(a-e) Time course of blood gas analysis and urinary acid-base parameters from wild-type (WT; white bar and open circle) and TgWnk4 PHAII mice (black bar and black square) during 15 days of acid loading (280 mM of NH 4 Cl in drinking water).Values are represented as the mean AE SEM for 8 mice.Statistical significance was assessed by unpaired Student t-test.*P < 0.05, **P < 0.01).(f) Relationship between urinary NH 4 þ and plasma [HCO 3 -] before and after the acid loading.(g) Representative profile of intracellular pH changes following luminal chloride removal in intercalated cells in isolated cortical collecting duct (CCDs) from WT (black curve) and TgWnk4 PHAII (red curve) mice.Average initial rates of pH increase after chloride removal in intercalated cell (ICs) from WT (open bars, N ¼ 4) and TgWnk4 PHAII (black bars, N ¼ 7) mice.Values are represented as the mean AE SEM, *P < 0.05, unpaired Student t-test.(h) Immunofluorescence images of collecting ducts in medullary rays from WT and TgWnk4 PHAII mice.Kidney sections were labeled for pendrin, for the E subunit of the V-ATPase, and for AE1.b-ICs were identified as apical pendrin (red) and basolateral V-ATPase (green) positive cells.a-ICs were identified as apical V-ATPase (green) and basolateral AE1 (blue) positive cells.Bar ¼ 50 mm.(i) Fraction of a-IC cells and b-IC cells in relation of total ICs in CCD.WT mice (open bars) and TgWnk4 PHAII (black bars), N ¼ 4 for each group.*P < 0.05, unpaired t-test.(j) Fractional volume of CCDs in WT (open bars) and TgWnk4 PHAII mice (black bars), N ¼ 4 mice per group.*P < 0.05, unpaired Student t-test.To optimize viewing of this image, please see the online version of this article at www. kidney-international.org.

Figure 6 |
Figure 6 | Effect of pendrin deletion in TgWnk4 PHAII mice on renin expression and systolic blood pressure.Left panel: renin gene expression was analyzed by quantitative reverse-transcription polymerase chain reaction in whole kidney of 7 TgWnk4;Pds þ/þ (black bar) and 6 TgWnk4;Pds -/-(gray bar) mice.Results are expressed in arbitrary units relative to the expression of a geometrical mean of 4 housekeeping genes: ubiquitin, HPRT, 18s rRNA, and GAPDH.The TgWnk4;Pds þ/þ group has been arbitrarily set to 1. Values are represented as mean AE SEM.Statistical significance was assessed by unpaired Student t-test.**P < 0.01.Right panel: Systolic blood pressure was measured by telemetry in 6 TgWnk4;Pds þ/þ (black bar) and 5 TgWnk4;Pds -/-(gray bar) mice.Data correspond to the 12-hour night period mean AE SEM.

Figure 7 |
Figure 7 | Effect of pendrin deletion in TgWnk4 PHAII mice on sodium ion (Na D ) and chloride ion (Cl -) transporters.(a) Immunoblots performed on plasma membrane-enriched fraction of renal cortex from TgWnk4;Pds þ/þ (N ¼ 6) and TgWnk4;Pds -/-(N ¼ 6) mice.Bar graph shows a summary of densitometric analyses.Densitometric values were normalized to the mean for the TgWnk4;Pds þ/þ mice that was defined as 100%, and results were expressed as mean AE SEM.Statistical significance was assessed by unpaired Student t-test.(b) Effect of amiloride injection on urinary Na þ and potassium ion(K þ ) excretion in TgWnk4;Pds þ/þ (N ¼ 7) and TgWnk4;Pds -/-(N ¼ 5) mice.Amiloride elicits significant increase in Na þ excretion and decrease in K þ excretion in both TgWnk4;Pds þ/þ and TgWnk4;Pds -/-mice 6 hours after injection.Amiloride induced natriuresis was not different between groups.K þ excretion after amiloride was significantly lower in TgWnk4;Pds -/-mice, indicating that epithelial sodium channel (ENaC)dependent K þ secretion is increased in these mice.Values are the mean AE SEM.Statistical significance was assessed by 1-way analysis of variance.**P < 0.01 and ****P < 0.0001, vehicle versus amiloride.****P < 0.0001, TgWnk4;Pds þ/þ versus TgWnk4;Pds -/-mice following amiloride injection.To optimize viewing of this image, please see the online version of this article at www.kidney-international.org.