Correspondence Dr. K. M. O'Shaughnessy, Clinical Pharmacology Unit, Box 110, Addenbrooke's Hospital, Cambridge CB2 2QQ, UK. Tel.: +44 1223 762578; Fax: +44 1223 762576; E-mail: email@example.com
A high salt intake in industrialized countries is an important cardiovascular risk factor. It remains at typically twice the maximum recommended levels of 5–6 g/day, and halving this would have enormous public health benefit in preventing stroke and cardiovascular disease. Salt homeostasis can also be affected pharmacologically by diuretic drugs that target mechanisms within the distal kidney nephron to cause salt wasting. Indeed, thiazide-type diuretics are among the most widely used agents in the management of hypertension and work by blocking NCCT, the NaCl-transporter in the distal nephron. The biology of this membrane transporter was not previously well understood until the discovery of the molecular basis of a rare familial form of hypertension called Gordon syndrome (pseudohypoaldosteronism type 2, PHAII). This has established that the NCCT transporter is dynamically regulated in the kidney by WNK kinases and a signaling cascade using a second kinase called SPAK. Common polymorphisms in the SPAK gene have recently been shown to affect blood pressure in human cohorts and removing its function lowers blood pressure in mice. The SPAK-deficient mouse actually has a phenotype reminiscent of Gitelman syndrome. This suggests that specific inhibitors of SPAK kinase may provide a novel class of diuretic drugs to lower blood pressure through salt wasting. The expectation is that SPAK inhibitors would mimic the on-target effects of thiazides but not their adverse off-target effects. An antihypertensive drug that could lower blood pressure with the efficacy of a thiazide without producing metabolic side effects such as hyperuricaemia or impaired glucose tolerance is therapeutically very attractive. It also exemplifies how data coming from the rare monogenic hypertension syndromes can together with genome-wide association studies in hypertension deliver novel druggable targets.
Hypertension is a very common disorder that affects some 25% of the adult population. More importantly, its untreated consequences represent a large burden of morbidity and mortality for industrialized societies. Hypertension is a major risk factor for cardiovascular disease (CVD) such as stroke, heart attack, and kidney disease and the World Health Organization predicts that CVD will become the leading cause of death worldwide by 2010 . Both genetic and environmental factors are involved in determining blood pressure. The importance of salt intake as an environmental factor is simply put: it can be easily modified, at least in principle .
Our current high salt intake is a modern development. Distant ancestors on the open savannahs of Africa were adapted to a very low-salt environment consuming much less than 1 g/day . However, the development of agriculture and farming has lead to a steady increase in salt consumption over the last few millennia . The arrival of modern food processing and high-salt “fast” food in the last few decades has ensured that our daily intake of salt is now many times that needed to maintain the normal physiological function of sodium. Indeed, in industrialized society the typical consumption of salt (NaCl) often exceeds 12 g/day. Although expert groups such as the British Hypertension Society (BHS) suggest that is should not exceed 6 g/day, and public health benefits in terms of CVD can be expected if this was reduced further to 3 g/day ; which of course is still far in excess of our ancestral consumption.
Salt-Sensitive and Salt-Resistant Hypertension
Although epidemiological observations have clearly shown that hypertension is rarer in societies which consume a low salt diet (<5 g/day) versus those consuming a higher salt diet (>15–20 g/day) , controversy remains. Indeed, although the relationship between blood pressure and salt intake is obvious in populations at these extremes of <3–5 or >20 g/day, it has been difficult to establish this correlation for populations consuming between >5 and <20 g/day . Studies of chimpanzees, the species genetically closest to humans, have shown that in a colony to which half had salt added progressively to their diet, additional salt caused a significant rise in blood pressure . However, the individual effect of salt differed between chimpanzees: from a large to a small blood pressure rise, or no rise at all. The same inter-individual variation in the susceptibility to the pressor effect of salt was observed in epidemiological studies and clinical trials in humans . In fact, the term salt “sensitivity” (despite being poorly defined) implies that some people are responsive to a change in their salt intake, with a strong correlation between a high salt intake and hypertension, whereas some individuals are not (so-called “salt resistant individuals”) . This heterogeneity probably explains why it has been difficult to observe a clear relationship between dietary salt intake and blood pressure within a mixed population of individuals when the change in salt intake is not as extreme as <3–5 or >20 g/day. However, it is important to note that blood pressure reduction in response to dietary sodium restriction is more pronounced in hypertensive compared to normotensive subjects. Moreover, in populations such as the Yanomama of the Carajas rainforest in Brazil, and natives of New Guinea who consume a very low salt diet, BP does not rise as steeply with age as it does in industrialized populations illustrating the long-term protective benefit of such a diet .
A recent meta-analysis of 13 prospective studies of reduced salt diets emphasizes that even in industrialized societies a lowered salt intake is very worthwhile. At a population level, a 5 g/day reduction in salt intake could prevent 1.25 million deaths from stroke and some 3 million deaths per annum from cardiovascular disease worldwide . Therefore, the public health benefits of consuming a low salt diet should be obvious and are reflected in the advice offered in numerous international guidelines [12–14]. It is often claimed that a low salt diet makes food unpalatable, but a progressive reduction in salt intake is accompanied by a reciprocal change in our ability to detect the saltiness of food. So that eventually there is no detectable change in the taste of food on the lower salt diet .
Nevertheless, the public health dilemma facing us is how the general population can be persuaded to usefully lower their salt intake. Because most of our salt intact is ‘hidden’ in processed food, this requires the close cooperation of the food industry. To date, it has been reluctant to engage in this process especially in adopting ‘traffic lights’ and other ways of transparently labeling the salt content of food.
Sodium Transport along the Nephron: Lessons from Rare Monogenic Blood Pressure Syndromes
According to Guyton's hypothesis, the kidney plays an essential role in setting the long-term blood pressure . During the last decade, molecular genetics have identified mutations in a number of genes involved in the control of blood pressure, derived from the study of rare Mendelian forms of hypertension and hypotension. Remarkably, given the variety of physiological systems that affect arterial pressure, these mutated gene products all ultimately affect the control of NaCl transport in the distal nephron of the kidney .
It is worth contrasting here the success that unraveling the monogenic forms of hypertension has had on kidney physiology with the attempts to dissect the much commoner essential hypertension. Not only is this a complex genetic trait and hence driven by multiple genes, but isolating the genes themselves remains problematic. Hence, despite very large samples sizes (several tens of thousand), the recent Global BPgen  and CHARGE  meta-analyses only shared four genetic loci between them; the causative allele was not identified within any of these nor did they include any of the genes identified as causing monogenic forms of high blood pressure. As we discuss later, SPAK kinase may be the first gene product to break this mould.
The sodium cation itself (Na+) is a major determinant of extracellular volume and thus blood pressure, because it represents (together with its associated counter anions Cl− and HCO3−) the main osmolyte in the ECF. Following filtration at the glomerulus, the recovery of Na+ into the ECF requires movement across kidney epithelial cells, which involves sequential transport across apical and basolateral surfaces. The basolateral transport of Na+ generally involves the Na,K-ATPase pump, which is expressed throughout the length of the nephron. In contrast, movement across the apical membrane is carried out by a diversity of transporters and ion channels that are characteristic of each region of the nephron.
The greatest fraction of filtered Na+ (60–70%) is reabsorbed in the proximal tubule principally by the sodium/hydrogen exchanger isoform 3 (NHE3) . A further 20–30% of Na+ is reabsorbed in the thick ascending limb of Henle (TAL) mainly by the Na+/K+/2Cl− cotransporter, NKCC2. This part of the nephron is the target of the loop diuretics, furosemide and bumetamide, which act by blocking NKCC2 . Potassium (which becomes rate-limiting) is recycled across in the apical membrane by the renal outer medulla potassium channel, ROMK, whereas Cl− exits the cell via the basolateral membrane, through chloride channels and transporters including CLCNKB. Mutations within any of these genes in the loop (NKCC2, ROMK, and CLCNKB) leads to Bartter syndrome: respectively type I, II, and III [21–24]. Bartter syndrome is characterized by hypokalemia with metabolic alkalosis and hypercalciuria. The blood pressure can be normal or low (especially in the neonatal period) depending on the degree of salt wasting.
In the distal convoluted tubule (DCT), about 5–7% of the filtered sodium is reabsorbed through the thiazide-sensitive sodium/chloride cotransporter, NCCT. The DCT is the site of two other monogenic diseases. First, loss-of-function mutations within NCCT leads to Gitelman syndrome , which resembles Bartter syndrome but with hypomagnesaemia and hypocalciuria. Second, Gordon syndrome  also known as pseudohypoaldosteronism type 2 (PHAII) that is the clinical mirror of Gitelman syndrome and caused by increased NCCT membrane expression in the kidney. It is characterized by hypertension and hyperkalemic acidosis and caused by mutations within one of two members of the novel family of kinases called the “With No lysine (K)” kinases (WNK1 and WNK4) [27,28].
The final adjustment of Na+ reabsorption occurs in the last part of the DCT, the connecting tubule (CNT) and the collecting duct (CD). In these segments of the nephron, Na+ is transported through the amiloride-sensitive epithelial sodium channel, ENaC, located in the apical membrane [29,30]. It is made up of α, β, and γ subunits and gain-of-function mutations localized in the PY motif of the C-terminal tail of the β and γ subunits lead to another Mendelian disease, Liddle syndrome . Liddle syndrome is characterized by hypertension and hypokalemic alkalosis associated with hypoaldosteronemia. The clinical inverse of Liddle syndrome also exists in the form of pseudohypoaldosteronism type 1 (PHAI) , which exists in both autosomal recessive and dominant forms. The first is caused by homozygous loss-of-function mutations in ENaC  and is very severe, whereas the second is milder, and caused by loss-of function mutations in the mineralocorticoid receptor (MR) . These syndromes are summarized in Figure 1.
The Importance of the Na–Cl Cotransporter (NCCT)
Attention has classically focused on ENaC as the principal regulator of Na+ reabsorption in the distal nephron. However, the striking effect that genetic manipulation of NCCT produces on blood pressure has led to a reappraisal of the role of NCCT as a key blood pressure regulator in humans. Hence, patients with loss-of-function mutations of the NCCT transporter (Gitelman syndrome) show salt wasting and lowered blood pressure whereas those with loss-of-function mutations in WNK proteins, which negatively regulate NCCT retain salt and become hypertensive (Gordon syndrome).
The pharmacological sequences of NCCT inhibition by thiazide diuretics provide equally robust evidence for NCCT as a significant blood pressure regulator in the general population. Thiazide diuretics are one of the oldest and most widely used class of antihypertensive medication  with all cause mortality benefits equal to angiotensin-converting enzyme inhibitors and calcium channel antagonists . The low cost of thiazide diuretics also makes them a highly cost-effective treatment for hypertension . Thiazides however are no panacea; they have a significant side effect profile. Some adverse effects are directly related to NCCT inhibition such as hypokalemia, hypercalcemia and possibly hypomagnesemia, but most appear to be idiosyncratic off-target effects including gout, photosensitization, impotence, and impaired glucose tolerance. Hence, a better understanding of the molecular mechanisms that regulate NCCT could lead to novel diuretics with fewer adverse effects.
Regulation of NCCT
NCCT is a protein of 12 transmembrane spanning domains and the third member of the SLC12A family of sodium cotransporters. It exists as a homodimer  and is expressed mainly in the DCT where it transports sodium and chloride in a 1:1 stoichiometry. It is regulated by alteration both in its level of expression in the surface membrane (Type 1 regulation) and its phosphorylation state (Type 2 regulation), which directly affects transporter kinetics. This is illustrated in Figure 2 and reflects the two functional categories of mutation seen in Gitelman syndrome, termed Type 1 and Type 2, respectively [38,39].
Type 1 involves alteration in forward trafficking and expression of NCCT at the apical membrane through defective glycosylation (retarding NCCT in the ER) caused by some Gitelman syndrome mutations or reduced WNK4-mediated degradation through the lysosomal pathway caused by PHA II mutations. Type 2 regulation involves alteration of intrinsic transporter kinetics by phosphorylation of conserved N-terminal amino acids, especially T58. Phosphorylation is induced by hypotonic stress and Cl− depletion and is effected by the kinases SPAK/OSR1.
Regulation of NCCT Surface Expression (Type 1 Regulation)
This can occur either as a failure of NCCT to undergo its normal processing in the endoplasmic reticulum (ER) and Golgi apparatus or by enhanced forward trafficking from the Golgi apparatus to the plasma membrane. The former occurs in NCCT transporters harboring Type 1 Gitelman mutations that impair glycosylation. The latter is mediated by three of the four WNK kinases [40,41]. WNK4 is the most fully characterized of these, and normally reduces NCCT expression at the surface membrane by diverting the transporter for lysosomal degredation using a pathway dependent on the signaling molecule AP-3 . Hence, in PHAII patients with WNK4 mutations their hypertension is caused by the mutant WNK4 failing to divert NCCT for degradation, and the increased forward trafficking of NCCT to the apical membrane results in its over expression at the cell surface and excessive salt absorption through the transporter [43–45].
Other members of the WNK kinase family also affect NCCT surface membrane expression. Intronic WNK1 mutations lead to the expression of an unusually long WNK1 isoform in the kidney [46,47]. This long WNK1 (L-WNK1), unlike the short WNK1 form (S-WNK1), is thought to inhibit WNK4 thus causing increased expression of NCCT at the apical surface and so enhanced NaCl reabsorption [45,46]. A third WNK kinase, WNK3, also increases NCCT expression at the surface membrane [41,48], although the precise mechanism is not clear and appears to vary between the brain and renal isoforms. No disease-causing mutations in WNK3 have been reported so far.
Regulation of the Phosphorylation State of NCCT (Type 2 Regulation)
Gamba's group first observed that Cl− depletion of Xenopus oocytes expressing a fluorophore-tagged NCCT protein resulted in an increase in Na+ transport without any change in surface membrane expression of the transporter . This activation of NCCT activity was associated with the phosphorylation of residues in the N-terminus of the protein: specifically, two threonines (T53 and T58) and one serine (S71) residue. Site mutation of each residue to alanine prevented low Cl− conditions from activating NCCT, but the effect was most pronounced for T58 (equivalent to T60 in human NCCT) . We have also observed that even under chloride replete conditions, mutation of NCCT to mimic a nonphosphorylated state of the transporter (T58A) reduced NCCT activity whereas T58D mutation (mimicking constitutive phosphorylation) produced the opposite effect . The reason T58 mutation dramatically alters NCCT function is probably because it impairs the phosphorylation of the adjacent S/T residues .
The importance of these N-terminal residues in controlling NCCT activity has been verified in mammalian cell  culture and significantly the human T58 homologue, T60, has been identified as the mutant residue in some Gitelman pedigrees . These N-terminal T/S residues are in fact conserved in other members of the SLC12 superfamily (Figure 3) and N-terminal phosphorylation of NKCC1 is also observed to increase transporter activity without affecting surface membrane expression .
NCCT Regulation by SPAK and OSR1
WNK kinases generally require intact catalytic activity to regulate SLC12A cotransporters [40,41,44,52], although there is little evidence that they directly phosphorylate the transporters themselves . In fact, it is likely they form part of a kinase cascade with the phosphorylation targets of the WNKs being the STE20/SPS-1-related proline/alanine rich kinases SPAK and OSR1 . SPAK and OSR1 are members of the GCK (Germinal center kinase) subfamily of the STE20 family of MAPK (mitogen-activated protein kinase)-like protein kinases . They are widely expressed in many tissues including the kidney  and have been well characterized in their ability to phosphorylate and activate NCCT and NKCC1 in response to either osmotic stress, WNK1, or WNK4 [47,55,56]. SPAK and OSR1 both have an N-terminal serine/threonine kinase domain, a conserved region, the S-motif, and a unique 92 amino acid regulatory sequence known as the conserved C-terminal (CCT) . The CCT contains a binding site which recognizes RFx[I/V] motifs found in both activators (WNK1 and WNK4 ) and substrates (NKCC1 and NCCT) [47,56,59,60] and was role was recently modeled from x-ray crystallographic data .
In operational terms, SPAK is thought to bind to the N-terminal of NCCT via an RFx[I/V] motif where it phosphorylates the key conserved Ser/Thr residues causing activation of the transporter (T46, T55, and T60 in human NCCT) . Equivalent residues in NKCC1 and NKCC2 are also phosphorylated by both low Cl− conditions or SPAK/OSR [57,62–64]. Of note, expression of a kinase-dead SPAK protein in Xenopus oocytes reduces the activity of NCCT , NKCC1 [53,55], and NKCC2 . However, it remains unclear whether the mechanism by which WNK1 and WNK4 regulate NCCT expression at the apical membrane also depends upon SPAK/OSR1. In the case of NKCC1, co-expression of SPAK in Xenopus oocytes had no effect on surface expression of the transporter .
No monogenic syndrome has been attributed to SPAK mutations to date, although it seems likely they would present clinically as a Gitelman-like syndrome . Nevertheless, milder variation in SPAK function could contribute to essential hypertension. Hence, a recent Genome Wide Association Study in the Amish has identified striking association between SNPs in STE39 (the gene encoding SPAK) and hypertension . At least one of the intronic SNPs in STK39 appeared to increase SPAK expression, which could increase NCCT phosphorylation, and hence blood pressure by promoting salt retention. Finally, we have looked at the in vivo role of SPAK using a SPAK kinase-dead knock in mouse, which expresses a mutant form of SPAK that is unable to phosphorylate its usual target proteins including NCCT. This mouse demonstrates obvious hypotension that is paralleled by a substantial reduction in NCCT expression and phosphorylation at T58 and T55 . This provides the first evidence that NCCT is directly regulated by SPAK in vivo.
Summary and Future Directions
Modifying salt intake to lower blood pressure is an important strategy to help lower an individual's blood pressure on or off antihypertensive therapy. This therapy may itself make use of diuretics, especially thiazide diuretics, to promote salt wasting. The various targets for these diuretic drugs are now well understood from the study of the molecular basis of rare monogenic blood pressure syndromes such as PHAII. Subjects with PHAII have mutations in WNK kinases that regulate the expression of NCCT, the target for thiazide diuretics. However, WNK kinases do not direct phosphorylate NCCT, but act through a second downstream kinase called SPAK.
The discovery of this WNK–SPAK kinase signaling pathway has identified novel drug targets to modulate NCCT function. SPAK itself is a particularly attractive one, because SNP variants within the SPAK gene are associated with hypertension . These same variants also appear to affect SPAK gene expression, which suggests that human blood pressure may be relatively sensitive to small changes in the WNK–SPAK signaling pathway. Inhibiting SPAK within the kidney it likely to have a highly selective effect on tubular function since SPAK is only expressed within the DCT . A closely related kinase, OSR1, is more widely expressed but recent work in a SPAK-deficient mouse has shown that it cannot substitute for the loss of SPAK activity . These mice have a profound reduction in both NCCT and its phosphorylated form in the DCT and a phenotype that closely resembles Gitelman syndrome. This raises the intriguing possibility that SPAK inhibitors may be highly effective thiazide-like diuretics, but without the off-target metabolic side effects of thiazides. In fact, the SPAK-deficient mouse is normoglycaemic and its overall phenotype suggests that SPAK deficiency (and hence perhaps SPAK inhibitors) is well tolerated. Of course, producing selective kinase inhibitors that can distinguish between SPAK and the closely related OSR1 may be challenging, but not insurmountable from experience with other kinase inhibitors. Indeed, given the size of the blood pressure reduction in SPAK-deficient mice even modest inhibition of SPAK may be sufficient to produce a clinically useful and novel antihypertensive drug.
MG was funded by a British Heart Foundation Clinical PhD studentship and a Sackler studentship. AMZ was supported by the Swiss National Science Foundation.