WNK kinases and renal sodium transport in health and disease: an integrated view

Abstract

The with no lysine (WNK) kinases comprise a novel branch of the human kinome that plays a central role in regulating renal sodium, potassium, and chloride transport, and, therefore, blood pressure. Mutations of two WNK kinases, WNK1 and WNK4, cause familial hyperkalemic hypertension (Gordon’s syndrome or Type II pseudohypoaldosteronism), a rare monogenic disease. Many aspects of WNK action have been elucidated during the past seven years. WNKs are all expressed along a short segment of renal distal tubule, where they modulate the activity of a wide variety of transport proteins. These diverse effects, however, make it difficult to describe an integrated model of WNK function within the kidney. Recently, work in vivo and in vitro has begun to clarify this picture. The present review emphasizes recent insights into mechanism by which WNK kinases interact to modulate sodium and potassium transport along the aldosterone-sensitive distal nephron. We describe a potential mechanism by which WNK4 mutations convert the action of WNK4 from inhibiting renal sodium chloride retention to stimulating it, thereby affecting both blood pressure and potassium balance. An explanation for how WNK kinases can alter the effects of aldosterone from primarily kaliuretic to primarily sodium chloride retentive, according to physiological need, is also described.

Hypertension affects 25% of the adult population in the developed world and is a major independent risk factor for stroke, myocardial infarction, and heart and kidney failure. Although many genetic and environmental contributors are involved, the kidney plays a dominant role, both in animal models¹, and in human essential hypertension². Most monogenic hypertensive syndromes result from increased Na+ transport along the aldosterone-sensitive distal nephron (ASDN)³. The majority of these, however, are associated with hypokalemia, indicating that activation of the epithelial Na+ channel, ENaC, is a primary pathophysiologic process. In contrast, familial hyperkalemic hypertension (FHHt, also known as Gordon’s Syndrome or Type II Pseudohypoaldosteronism) is characterized by hypertension with hyperkalemia indicating that stimulated ENaC cannot be the primary event. FHHt was first described in 1964⁴, and later shown to be inherited in an autosomal dominant manner^{5, 6}. Patients with FHHt all exhibit hyperkalemia, which appears to be the most consistent feature of the disease. Hypertension, while commonly present and sometimes severe, often appears later in the natural history. Other characteristic features include mild metabolic acidosis, suppressed plasma renin activity, and aldosterone levels that are lower than would be expected, considering the hyperkalemia. Infusing the chloride salt of Na+ (NaCl) does not increase urinary potassium excretion in patients with FHHt, as it does in normal individual, whereas infusing non-chloride salts of Na+ does increase K+ excretion in FHHt patients to normal levels^{7, 8}. Patients are often remarkably sensitive to thiazide diuretics, which can correct both the hyperkalemia and hypertension, in many cases⁹.

WNK kinases and FHHt

In 2001, some cases of FHHt were shown to result from mutations in WNK1 and WNK4¹⁰, identifying WNK kinases as previously undiscovered components of a novel electrolyte homeostasis pathway. Since that time, information about the physiological role of WNK kinases, their substrates and mechanisms of action, and their role in FHHt has accumulated rapidly. The identification of a previously unrecognized signal transduction pathway that plays a central role in renal electrolyte balance and blood pressure control has generated considerable interest and has been reviewed widely^11–15. Despite this, several fundamental questions regarding physiological and pathophysiological actions of WNK kinases, and the pathogenesis of FHHt, are only now being addressed. The focus of this review will be several of these remaining questions.

WNK1 was first identified by Cobb and colleagues during a search for novel members of the MAP/extracellular signal-regulated protein kinase family¹⁶. They named the new family of kinases With No Lysine (K) because the lysine critical for ATP binding in the catalytic site is in subdomain I rather than subdomain II, where it is located in other serine/threonine kinases¹⁶. When first identified, the function, upstream regulators and downstream targets of the WNK kinases were not clear. Cobb and colleagues demonstrated that WNKs are catalytically active and widely expressed, at least at the transcript level¹⁶. Four WNKs have now been identified in humans (WNK1–WNK4), and all exhibit a similar domain structure (see Figure 1), including a short amino terminal domain, a highly conserved kinase domain, and a longer carboxyl terminal domain, with at least two coiled coil domains. Like many other kinases, WNKs contain an autoinhibitory domain that inhibits kinase activity^{17, 18}; this domain, which contains two crucial phenylalanine residues, lies just beyond the kinase domain (Figure 1). In addition, there are several PXXP motifs that may interact with SH3 domains of other proteins. Two serine residues located within the activation loop (S382 and S378) appear to modulate kinase activity¹⁷.

Figure 1. Structures of WNK Kinases.

WNKs 1–4 are shown. All contain a homologous kinase domain (pink), an autoinhibitory domain (green) with two essential phenylalanine residues (F), and coiled coil domains (yellow). Two phosphorylated (P) serine residues that are essential for WNK1 activation are shown (S382 is more important). The alternative first exon in KS-WNK1 is shown in dark red. Regions of WNK4 mutations that cause FHHt are shown in orange. Approximate domain locations and specific residues are provided based on mouse WNKs.

WNK kinases appear to be involved in many physiological processes (Table 1). WNK1 is activated by both hyper- and hypotonicity, and is sensitive to osmotic stressors, such as NaCl, KCl, and carbohydrates^{16, 19}. This suggests that WNK kinases participate in cell volume regulation²⁰, because they also interact with, phosphorylate, and activate the tonicity-responsive kinases SPAK/PASK and OSR1^{21, 22}, thereby affecting cation chloride cotransporters. WNK kinases also bind to and phosphorylate synaptotagmin-2²³, a protein that regulates membrane fusion events and participates in neurotransmission, and they bind to the syntaxin-inhibitory protein, Munc18c²⁴. WNK kinases affect cell growth and apoptosis, through several pathways^25–27. WNK1 is expressed ubiquitously, and its expression in the cardiovascular system in particular (ref delaloy), may account for the in utero lethality of WNK1 deletion²⁸. WNK3 is highly expressed in the brain, and may play an important role in the regulation of volume and intracellular Cl− in GABA-ergic neurons (ref). Finally, WNK kinases modulate ion transport across epithelia (see below).

Table 1.

Physiological Effects of WNK Kinases

Function	Mediator
Cell Volume regulation	OSR1/SPAK21, 22, 81–83
Neurotransmission	Synaptotagmin, Syntaxin84, 85
Cell proliferation	caspase, Akt/PKB25, 27
Development	?28
Paracellular permeability	Claudins74–76
Transepithelial ion transport	Intersectin, clathrin,53, 62, 71, SGK165, 66, OSR1/SPAK21, 22, 81–83

Despite these diverse and essential functions, the WNK kinase mutations identified in humans cause a phenotype that results predominantly from kidney dysfunction, although minor effects in other organ systems may be present²⁹. Disease-causing WNK1 mutations are large deletions in the first intron that do not change the coding sequence¹⁰. These intronic mutations are believed to increase WNK1 protein expression¹⁰, although this observation has not been corroborated in renal tissue. FHHt-causing WNK4 mutations are missense mutations located within two discrete regions (see Figure 1), the first within an acidic motif adjacent to the first coiled coil domain, and the second adjacent to the second coiled coil domain^{10, 30}. The FHHt-causing WNK4 mutations lie outside the kinase domain.

Expression and regulation of WNK kinases in the kidney

To understand how the WNK kinases regulate electrolyte homeostasis, and how this relates to FHHt and normal physiology, it is important to understand the expression pattern of ion transport proteins in the distal nephron. The distal tubule (See Figure 2) can be defined as the region of the nephron between the macula densa and the confluence with another tubule to form the colleting duct. This region comprises a short segment of thick ascending limb, the true distal convoluted tubule (DCT), the connecting tubule (CNT) and the initial segment of cortical collecting tubule (ICT) (for a review, see31). The thiazide-sensitive Na-Cl cotransporter, NCC, has been localized exclusively to the DCT at the mRNA level using in situ hybridization^32–34 and single nephron PCR³⁵. At the protein level, NCC expression is also limited to DCT cells^{36, 37}; NCC expression therefore “defines” the DCT. Further expression analysis has revealed that the DCT can be subdivided into an “early” DCT (DCT1) and a “late” DCT (DCT2)³³. Both DCT1 and DCT2 express the NCC, but the DCT1 does not express the sodium-calcium exchanger (Na/Ca)³³ or ENaC³⁸, both of which are expressed along the DCT2³⁹. In contrast, the collecting duct, while expressing ENaC, does not express Na/Ca³³. The K channel, ROMK, and the Na-K-ATPase are expressed all along the distal tubule⁴⁰.

Figure 2. Distribution patterns of sodium and potassium transport and aldosterone signaling machinery along the distal nephron.

The figure shows the thick ascending limb (TAL), the distal convoluted tubule (DCT, with ‘early’ and ‘late’ segments), the connecting tubule (CNT), and the cortical (C), and outer medullary (OM) collecting ducts (CD). Locations of transport proteins and regulatory kinases are shown, including the the sodium chloride contransporter (NCC), the epithelial Na+ channel, ENaC, the potassium channel ROMK, WNK kinases, including kidney specific WNK1 (KS-WNK1), full length WNK1 (WNK1) and WNK4, serum and glucocorticoid induced kinase (SGK1), and 11-b hydroxysteroid dehydrogenase (11HSD2).

The WNK1 gene (WNK1 or PRKWNK1) produces at least two major products^41–43 (and several minor products⁴²), a full-length kinase-active WNK1 (WNK1), which is widely expressed, and a second truncated product (Figure 1). The truncated product is produced from a separate promoter and lacks the majority of the kinase domain; this product is, therefore, kinase-inactive. It appears to be expressed only by kidney tubule epithelial cells, predominantly along the distal convoluted tubule (DCT) and connecting tubule (see Figure 2). For this reason, the kinase-inactive isoform has been termed ‘kidney-specific WNK1’ (KS-WNK1), to differentiate it from the full length WNK1.

WNK4 is expressed by epithelial cells throughout the body, including cells of the distal nephron, where it localizes, at least in part, adjacent to tight junctions^{10, 44}. It is highly expressed by cells of the distal convoluted tubule and connecting tubule, but expression extends distally into the collecting duct and, at lower levels, into thick ascending limb⁴⁵.

WNK3 is a third member of the family that is also expressed by kidney epithelial cells and elsewhere in the body^46–48. In contrast to the renal expression patterns of KS-WNK1 and WNK4, WNK3 expression is not predominantly along the ASDN and instead is expressed throughout the nephron, from the proximal tubule to the collecting duct⁴⁷. Although mutations in WNK3 have not been reported to be associated with FHHt, WNK3 has recently been shown to regulate the same classes of ion transport proteins that are targets of WNK1 and WNK4 (see below).

WNK kinases regulate NCC in vitro

The clinical features of FHHt identify it as a disease of renal electrolyte transport, so investigation of the effects of WNK kinases was first directed at their role in modulating renal ion transport proteins. It is now clear that WNK kinases modulate the trafficking of many transport proteins to or from the plasma membrane, at least in vitro. In view of the fact that most of the defects in FHHt can be corrected by treatment with thiazide diuretics⁹, and FHHt presents as a “mirror-image” of Gitelman’s Syndrome, a disease that results from inactivating mutations of NCC⁴⁹, it is not surprising that WNK4 was soon shown to regulate NCC activity, in vitro^{30, 50, 51}.

WNK4 does not affect total cellular NCC protein abundance, but instead reduces NCC abundance at the plasma membrane (see Figure 3,^{10, 30, 51–53}). Immunoprecipitation studies have shown that WNK4 and NCC associate in a protein complex involving the carboxyl termini of both proteins^{50, 53, 54}. The role of kinase activity in modulating NCC activity is controversial. Two groups reported that the effects of WNK4 on NCC are dependent on its kinase activity^{30, 50} whereas another group found evidence of a kinase-independent action⁵⁴. Indeed, our group showed that a truncated form of WNK4 lacking the entire kinase domain inhibits NCC activity^{54, 55}. Further analysis identified a region near the carboxyl terminus of WNK4 that is required for NCC inhibition.

Figure 3. WNK effects in the DCT2 and CNT.

Model shows a late distal convoluted tubule cell (DCT2) cell and a connecting tubule cell (CNT, similar to a collecting duct cell) The DCT2 cell shows the WNK kinase signaling complex regulating NCC. Effects of WNK kinases are either stimulatory (red arrows) or inhibitory (green arrows). In addition to modulating trafficking of NCC, it appears likely that WNK kinases affect the phosphorylation state of the NCC⁵⁸, ⁸⁶, which has been shown to alter its activity independent of trafficking to the plasma membrane⁸⁶. It is not clear which WNK kinase is involved. Model also shows a CNT cell with WNK kinase effects on ENaC, ROMK, and paracellular permeability shown. Note that these WNK effects probably occur in the same cells and are shown separately for clarity.

Studies to determine the mechanism by which WNK4 reduces surface NCC expression suggest that WNK4 inhibits the insertion of NCC into the plasma membrane, rather than endocytosis. Studies in both Xenopus oocytes⁵² and mammalian cells⁵³ showed that the ability of WNK4 to reduce NCC surface expression is not affected by expression of a dominant-negative dynamin, suggesting that clathrin-dependent processes are not involved. Furthermore, the inhibitory effect of WNK4 on NCC was sensitive to inhibition of lysosomal proton pumps, suggesting that WNK4 reduces trafficking of NCC to the plasma membrane, ultimately leading to enhanced lysosomal degradation⁵³.

Unlike WNK4, WNK1 does not affect NCC activity directly (see Figure 3). Instead it suppresses the effects of WNK4 on NCC^{51, 52}. Functional studies of WNK1 have revealed that its actions to modulate WNK4 require physical association with WNK4 and intact kinase activity⁵⁴. The predominant renal isoform, KS-WNK1, which lacks intrinsic kinase activity, inhibits WNK1 kinase activity and inhibits its effects on NCC, presumably through a dominant-negative mechanism⁵⁵. KS-WNK1 interacts physically with WNK1 which, as noted, associates with NCC.

Surprisingly, WNK3 strongly stimulates NCC activity (see Figure 3⁴⁷). This effect is associated with an increase in NCC protein abundance at the plasma membrane and with an increase in NCC phosphorylation⁴⁷. A kinase-inactive form of WNK3 exerts inhibitory effects on NCC that resemble the effects of WNK4^{47, 56}.

Mouse models of WNK action

To date, four strains of genetically engineered mice have been generated to analyze the functions of the WNK kinases in vivo. A gene trap approach that disrupted the first WNK1 intron (and presumably left the KS-WNK1 promoter region intact) led to the production of WNK1 knockout mice²⁸. Mice homozygous for the disrupted allele die before embryonic day 13, possibly from cardiovascular defects. Heterozygotes were viable but had blood pressure that was significantly lower than wild type mice, lending support to the idea that L-WNK1 is a stimulator of sodium reabsorption.

Lalioti et al. generated two lines of mice transgenic for wild type WNK4 and for FHHt-causing Q562E WNK4⁵⁷. Animals overexpressing wild type WNK4 had lower blood pressure than wild type mice, whereas animals transgenic for WNK4 Q562E had higher blood pressure than wild type mice. With regard to electrolyte balance, mice overexpressing the FHHt-causing mutant WNK4 displayed hyperkalemia, to an extent similar to that seen in patients with FHHt. When challenged with a high K+ diet, Q526E WNK4 mice became more hyperkalemic, with clearly impaired K+ excretion. On a low K+ diet, wild type WNK4 transgenic mice became hypokalemic relative to wild type mice. Taken together, these data indicate that WNK4 plays a key role in regulating K+ balance, but the effects of WNK4 and mutant WNK4 on K+ balance are opposite.

Histological analysis of the kidneys of these transgenic mice revealed that overexpression of wild type WNK4 reduces the luminal surface area of the DCT, whereas overexpression of Q526E WNK4 increased the luminal surface area. Immunostaining indicated that expression levels of NCC were increased in Q526E WNK4 mice, but co-localization studies with segment markers indicated that NCC expression was still confined to the DCT. Little or no differences were observed in ROMK expression between wild type and transgenic mice on normal or high K+ diet, or in levels of ENaC expression⁵⁷. Importantly, interbreeding of Q562E WNK4 mice with NCC knockout mice resulted in complete amelioration of all the defects observed in the Q526E WNK mice, suggesting that dysregulation of NCC is the key mechanism underlying FHHt.

One limitation regarding the studies of Lalioti is that they were based on overexpression of wild type or expression of mutant WNK4 on a background of two wild type WNK4 alleles. A more recent report described generation of a WNK4 D561A knock-in mouse⁵⁸. Unlike the mice generated by Lalioti et al., which express mutant WNK4 on a wild type WNK4 background, these mice express one mutant WNK4 allele and one wild type allele, closely mimicking the human disease. The phenotype of these mice was remarkably similar to the phenotype of the WNK4 Q562E transgenic mice; the investigators also noted that the phenotypic effects were completely corrected by thiazide diuretics, like in humans⁵⁸. In addition, however, these investigators reported increased phosphorylation of the NCC, which may enhance NCC activity independent of effects on trafficking (see Figure 3). Taken together, these results strongly suggest that mutations in WNK4 cause FHHt primarily by increasing NCC activity, abundance, and phosphorylation in the DCT.

One result of these studies has led to some confusion, however. As noted, wild type WNK4 has been shown to inhibit NCC activity both in vitro and in animals. Some^{30, 50, 53}, though not all^{51, 59}, investigators have reported that mutations of WNK4 abrogate its inhibitory activity. This has led some investigators to suggest that FHHt results from loss of WNK4-mediated NCC inhibition^{30, 50, 53, 58, 60}. Yet, the WNK4 Q562E transgenic animals express mutant WNK4 on a background that includes two wild type WNK4 alleles⁵⁷; the knock-in animals express mutant WNK4 on a background of a single wild type allele⁵⁸. In both situations, wild type WNK4 is present and expressed within the kidney, suggesting that WNK4 mutations act as ‘gain-of-function’ rather than ‘loss-of-function’ mutations. Surprisingly then, animals made transgenic for additional copies of wild type WNK4 exhibit a phenotype that is opposite to FHHt and is strikingly similar to Gitelman’s syndrome⁵⁷. They have reduced NCC abundance and activity, and hypokalemia; thus, as demonstrated using oocytes experiments, wild type WNK4 reduces NCC activity and abundance at the plasma membrane. This indicates that wild type and mutant WNK4 proteins exert opposite effects on NCC. Taken together the results clearly show that FHHt results from disordered regulation of NCC by WNK4, but a unifying explanation of the functional effects of wild type and mutant WNK4 seems difficult to provide.

We recently reported evidence of WNK kinase effects that may help to resolve this conundrum. We found that WNK3 also participates in a WNK kinase signaling complex with WNK4⁵⁶. Lifton and colleagues showed that WNK3 stimulates NCC activity⁴⁷, in sharp contrast to the effects of WNK4. As noted above, WNK3 is expressed all along the nephron, including the DCT⁴⁷. We confirmed those results and showed further that WNK4 binds to and inhibits WNK3 actions on NCC; conversely, WNK3 blocks WNK4 inhibition of NCC⁵⁶. Intriguingly, when WNK3 and WNK4 are expressed in varying ratios, together with NCC, the range of NCC activity is much greater than when NCC is expressed with either WNK4 or WNK3 alone (see Figure 4). Furthermore, the net effect of expressing NCC with different molar ratios of WNK3/WNK4 is a finely graded regulation of NCC activity from negligible (when WNK4 abundance greatly exceeded WNK3) to very high (when WNK3 abundance greatly exceeded WNK4, see Figure 4). Thus, the expression of WNK3 with WNK4 acts, not so much as a molecular switch, but as a molecular rheostat, tightly controlling NCC activity across a wide range.

Figure 4. Effects of WNK3 and WNK4 on NCC activity.

schematic representation of effects of varying molar ratios of WNK3 and WNK4 on NCC activity. Circle indicates NCC activity in the absence of WNK kinases. Note that when WNK4 exceeds WNK3, NCC activity is low; when WNK3 exceeds WNK4, NCC activity is high.

We also found that FHHt-mutant WNK4 Q562E loses its ability to inhibit WNK3⁵⁶; yet, as noted above, FHHt does not appear to result from loss-of-function. This raised the possibility that interactions between WNK4 and WNK3 might help to explain the FHHt phenotype. In support of this hypothesis, we found that WNK4 Q562E binds to and inhibits the effect of wild type WNK4 on WNK3. Thus, WNK4 Q562E appears to act as a dominant-negative WNK4 modulator, leaving WNK3 activity unopposed and NCC activity strongly stimulated. Although this model is derived from observations made in vitro, and must be tested in more physiological systems, it is consistent with the opposing effects of wild type and mutant WNK4 on NCC activity. Wild type WNK4 suppresses NCC activity both directly, and by inhibiting WNK3. FHHt mutant WNK4 not only loses the ability to inhibit WNK3, but also blocks the effects of the wild type gene product, thereby leaving NCC activity enhanced. Figure 5 shows a simplified model of WNK3 and WNK4 interactions that could account for the phenotypes of the transgenic and knock-in mice.

Figure 5. Effects of wild type and mutant WNK4 on ion transport along the distal tubule.

schematic figure shows early DCT (DCT1) in green, late DCT (DCT2) in yellow, and connecting tubule/collecting duct (CNT/CCD) in red. A typical transepithelial voltage is also shown. Normally, the DCT1 reabsorbs NaCl, the CNT exchanges Na+ for K+ and the DCT does some of both; balanced WNK4/WNK3 effects maintain NaCl transport at baseline. When animals are made transgenic for wild type WNK4, WNK4 suppresses WNK3 effects and suppresses NCC activity, leading to reduced NaCl transport. When animals are made transgenic for FHHt mutant WNK4, or when the mutant is knocked in, the mutant allele inhibits the WNK4 effect on WNK3, leaving WNK3 to stimulate NaCl transport.

The WNK kinase signaling complex must play a role in normal electrolyte homeostasis, too. Although physiological regulators of WNK kinase activity and abundance are only beginning to be evaluated, it appears that dietary K+ loading increases WNK4 abundance⁴⁵ and increases the ratio of KS-WNK1/ WNK1^{45, 61, 62}. Figure 3 shows how both increased WNK4 abundance and an increased KS-WNK1/WNK1 ratio favor K+ secretion by shifting the DCT2 from an epithelium that transports Na+ primarily with Cl− (neutral), to one in which Na+ is transported largely in exchange for K+ (electrogenic). First, the increase in WNK4 inhibits NCC activity directly, thereby enhancing electrogenic Na+ absorption relative to electroneutral Na+ reabsorption; second, the increased KS-WNK1 inhibits WNK1’s ability to block WNK4, thereby inhibiting NCC secondarily; third, the increased WNK4 inhibits WNK3; since WNK3 is a potent NCC stimulator, this effect will further suppress NCC activity. As discussed below, these effects on NCC do not exclude important and physiologically relevant effects of WNK kinases on other ion transport pathways, but would tend to hyperpolarize the epithelium, favoring K+ secretion.

Lifton and colleagues proposed that an unidentified physiological ligand switches WNK4 activity from inhibitory (to NCC) to stimulatory (to NCC), mimicking the human disease, FHHt⁵⁰. According to the proposed WNK signaling complex model, WNK3 is a WNK4 ligand that fulfills this prediction (see Figure 3). Conversely, the same investigators suggested that an unidentified physiological ligand might switch WNK3 from stimulatory (to NCC) to inhibitory (to NCC)⁴⁷. The WNK signaling complex model (Figure 3) suggests that WNK4 is a WNK3 ligand that fulfills this prediction.

Mechanisms of Hyperkalemia in FHHt

Hyperkalemia is a universal feature of FHHt, and frequently develops prior to the onset of hypertension. Renal K+ excretion results primarily from K+ secretion along the distal tubule DCT2 and CNT^{31, 63} as well as the cortical collecting duct. Potassium secretion is driven by the electrical gradient generated by Na+ reabsorption (via ENaC) across an electrically tight epithelium. Paracellular permeability characteristics of the distal tubule are determined largely by expression of claudins, proteins that act as selective barriers to ion movement and therefore maintain the transepithelial voltage⁶⁴. Potassium is secreted largely, although not exclusively, via ROMK (Kir1.1) channels. In addition to affecting NCC, WNK kinases might, therefore, modulate K+ excretion by interacting with ENaC, ROMK, or claudins, among other proteins. The effects of WNK kinases on each of these protein classes will, therefore, be discussed (see Figure 3, CNT cell).

WNK kinases have been shown to modulate ENaC activity in cells, in Xenopus oocytes, and in vivo. WNK1 increases ENaC activity by activating phosphotidylinositol 3-kinase, stimulating glucocorticoid-induced kinase 1 (SGK1)^{65, 66}, a well-established ENaC-regulatory factor⁶⁷. This effect is dependent on an intact WNK1 kinase domain but also requires an intact amino terminal domain (N-terminal of the kinase domain). Interestingly, KS-WNK1, which lacks both the kinase domain and the amino terminal domain, has also been reported to stimulate ENaC, implying a different mechanism⁶⁸. In contrast to WNK1, WNK4 inhibits ENaC activity, an effect that is suppressed by SGK1⁶⁹. FHHt-mutant WNK4 (WNK4 Q562E) does not inhibit ENaC activity when expressed in Xenopus oocytes⁶⁹, and mice expressing an FHHt-mutant WNK4 exhibit increased ENaC activity in the kidney and colon^{58, 69}. Based on these results, and results discussed below, Lifton and colleagues suggest that the FHHt phenotype may result, in part, from the coordinated up-regulation of NCC and ENaC and down-regulation of ROMK (see below)⁶⁹. Three observations, however, argue that activation of ENaC, while probably present in FHHt²⁹, is not the primary factor in its pathogenesis. First, as noted above, thiazide diuretics correct both hypertension and hyperkalemia in FHHt, at least for patients who have mutations in WNK4⁹ and for animals carrying mutant WNK4^{57, 58}. Thiazides do not affect ENaC activity directly and instead increase Na delivery to distal sites of high ENaC expression⁷⁰. Second, as originally suggested by Lifton and colleagues¹⁰, stimulation of ENaC should generate hypokalemia, rather than the hyperkalemia that is consistently observed. Finally, Yang and colleagues showed that the increased ENaC activity in WNK4 D561A knock-in animals could be reversed by treatment with thiazide diuretics⁵⁸, suggesting that effects on ENaC are secondary to physiological changes.

WNK4 strongly inhibits ROMK activity in vitro, through a kinase-independent mechanism^{62, 71}. WNK4 reduces ROMK abundance at the plasma membrane, as it does with NCC. Yet the effect of WNK4 on ROMK is dynamin-dependent and involves clathrin-mediated endocytosis⁷¹, whereas the effect of WNK4 on NCC does not. Studies in Xenopus oocytes and HEK-293 cells have shown that WNK1 also inhibits ROMK activity. Two groups reported that the effect is dependent on intact kinase activity^{61, 62}, whereas another found that a kinase-dead WNK1 mutant was still effective⁷². Time course studies of ROMK plasma membrane expression suggest that WNK1 increases endocytosis of ROMK in a dynamin-dependent manner⁷², an effect involving interactions with the scaffolding protein intersectin (Huang). KS-WNK1 has no direct effect on ROMK activity, but blocks the effects of WNK1on ROMK^{61, 62}. KS-WNK1 therefore indirectly activates ROMK and inhibits NCC.

Lifton and colleagues⁷¹ reported that FHHt-causing mutant WNK4 inhibits ROMK more actively than does wild type WNK4 and suggested that hyperkalemia in FHHt could result from a reduced abundance of K channels at the plasma membrane⁷¹. As with the observations regarding ENaC, however, observations from patients and animal models argue that reduced K conductance is not a primary cause of hyperkalemia in FHHt, at least when the disease results from WNK4 mutations. First, patients with FHHt can excrete normal amounts of potassium when non-chloride salts of Na+, are infused^{7, 8}. This indicates that the potassium secretory apparatus is intact and suggests that reduced K+ secretion results from a reduced transepithelial voltage in the ASDN. Second, animals transgenic for mutant WNK4 do not exhibit any apparent changes in ROMK abundance^{57, 58}. Third, treatment with thiazides, which corrects the K+ secretory abnormalities^{57, 58} would not be expected to correct a defect in ROMK. It should be noted, however, that inhibition of NaCl reabsorption by thiazides would be expected to increase distal flow and could stimulate K secretion that occurs via the structurally unique maxi-K channels⁷³.

Based on the observations concerning the effect of non-chloride sodium salts on K+ secretion in FHHt, Schambelan and colleagues proposed that FHHt results from a “chloride shunt” in the distal nephron⁸. According to this model, increased chloride permeability in the distal nephron depolarizes the transepithelial voltage, which is oriented with the lumen negative relative to blood, thereby increasing Na+ reabsorption and decreasing K+ secretion; the former would contribute to hypertension, and the latter to hyperkalemia. The observation that WNK4 co-localizes with tight junctions¹⁰ suggested that WNK kinases might regulate paracellular chloride permeability and seemed consistent with the “chloride shunt” hypothesis. In support of this hypothesis, one group showed that WNK4 affects paracellular chloride permeability when overexpressed in cultured cells⁷⁴. Two groups also reported that FHHt-mutant WNK4 increased chloride permeability relative to Na permeability^{74, 75}. This effect of WNK4 was reported to require WNK4 catalytic activity, since a kinase-inactive WNK4 did not affect paracellular chloride permeability⁷⁵ and because WNK4 was reported to phosphorylate claudins 1–4 ⁷⁴. Recently, overexpression of L-WNK1 was reported to exert similar effects on chloride permeability⁷⁶. A relative increase in paracellular chloride permeability in response to mutant WNK4 would seem consistent with the “chloride shunt” model. Yet, as with effects of ENaC and ROMK, the ability of thiazide diuretics to correct the hyperkalemia in FHHt⁹ does not seem consistent with a predominant effect on paracellular processes. Thiazide diuretics do affect Cl transport, but have never been reported to alter paracellular processes. Further, direct measurement of paracellular permeability from collecting ducts of mice with WNK4 D561A knocked-in show no difference from wild type collecting ducts⁵⁸.

Although the data discussed do not support a primary role for altered ENaC, ROMK or claudin activity in the pathogenesis of FHHt induced by WNK4 mutations, this does not mean that the effects of WNK kinases on transport proteins other than NCC do not contribute importantly to electrolyte homeostasis normally. Most of the described effects of WNK kinases on these transport pathways are entirely consistent with physiologically regulated Na, Cl and K transport and likely contribute, under normal conditions. As described above, the unique effects of WNK4 mutations on NCC may therefore result from alterations in the WNK signaling complex that do not modulate ion transport mediated by other pathways. A full discussion of such possibilities is beyond the scope of this review.

The Aldosterone Paradox

The FHHt phenotype indicates that WNK kinases lie at a critical physiological control point that regulates the balance between NaCl absorption and K+ secretion, thereby modulating aldosterone’s actions to regulate blood pressure, on the one hand, and K+ balance, on the other. This suggests that consideration of WNK kinases might help answer a persistent question concerning the actions of aldosterone. Aldosterone is secreted in response to two, sometimes independent, stimuli; extracellular (ECF) fluid volume depletion and hyperkalemia. In the case of ECF volume depletion, aldosterone stimulates Na+ reabsorption (largely with chloride), while in hyperkalemia, aldosterone promotes K+ secretion⁷⁹. The ability of a single hormone to exert different effects has been documented by Kamel and Halperin^{79, 80} and termed, the Aldosterone Paradox, but a mechanistic explanation has not been apparent. Recently, it has been suggested that KS-WNK1 plays a key role^{45, 55, 61, 62}. Evidence for this comes from studies examining the regulation of WNK kinase expression by dietary electrolyte manipulation^{45, 55, 61, 62} and by aldosterone⁴⁵.

The balance between electroneutral Na+ reabsorption and electrogenic Na+ reabsorption plays a pivotal role in this model (see Figure 5). Under normal conditions, NCC mediates electroneutral NaCl reabsorption in the DCT1 and, to a lesser extent, the DCT2; in later segments, including the DCT2, ENaC activity generates a lumen-negative voltage, which drives K+ secretion. When aldosterone secretion is induced by dietary K+ loading, KS-WNK1, and probably WNK4, are induced^{45, 55, 61, 62}. As discussed above, both KS-WNK1 and WNK4 inhibit NCC trafficking and favor electrogenic Na+ reabsorption (Figure 3). Along with direct effects of aldosterone on ENaC, the balance of Na reabsorption along the DCT2 shifts from electroneutral to electrogenic, increasing the transepithelial voltage, and favoring K+ secretion. Effects of WNK kinases on ENaC, ROMK, and even claudins are likely to contribute additionally to these effects (see Figure 3). In contrast, during hypovolemia (low NaCl intake), aldosterone secretion is increased, but KS-WNK1 and WNK4 expression are not increased and may even be reduced⁴⁵. These changes activate the NCC, favoring electroneutral over electrogenic Na+ transport, thereby depolarizing the transmembrane voltage. In more distal segments that lack NCC, aldosterone still stimulates ENaC activity, which drives some K+ secretion, but the effect favors balanced NaCl reabsorption with Na+/K+ exchange.

Summary

Information about the roles of WNK kinases in normal physiology and in the pathogenesis of FHHt has accumulated rapidly. The pathogenesis of FHHt that results from mutations in WNK4 is currently being solved, although important questions remain. The disease involves activation of NCC along the DCT, probably because mutant WNK4 exerts a dominant-negative effect on wild type WNK4. This permits unrelieved activation of WNK3. In contrast, the mechanisms that underlie FHHt caused by WNK1 mutation are not as clear, but may also involve a dominant-negative effect, involving KS-WNK1. A decrease in the KS-WNK1/WNK1 ratio would activate NCC and inhibit ROMK. Whether these changes account, by themselves or with other processes, for the disease phenotype awaits experimental confirmation. Clearly, WNK kinases stand at the crossroads of renal Na+, K+ and Cl− transport; as such, they are attractive targets for drug development.