Diuretics are the most commonly prescribed drugs for hypertension in the United States [Article:22350108]. Diuretics modulate the reabsorption of sodium in the kidney and effect the homeostasis of water and sodium and thus the volume of blood and blood pressure. From a pharmacogenomics perspective, the most well studied drug in this class is the thiazide diuretic hydrochlorothiazide (HCT). HCT is excreted primarily by renal elimination with greater than 95% of the drug recovered unchanged in the urine [Article:1277708]. Therefore pharmacodynamic genes have been the focus for studies of the variation in response to thiazide diuretics. Thiazide diuretics act by inhibiting the sodium chloride co-transporter, NCC, coded for by SLC12A3 [Article:12590198]. Also acting on sodium regulation in the kidney, loop diuretics such as furosemide, target the sodium potassium chloride co-transporter (NKCC2, SLC12A1). In addition another sub-class of diuretics, potassium sparing diuretics like amiloride, act on the sodium channel ENac, coded for by SCNN1 family of genes.
This brief summary introduces the candidate genes in the pharmacodynamic pathway of diuretics and discusses the pharmacogenomic studies.
The sodium chloride co-transporter SLC12A3 is also known as the thiazide-sensitive transporter. It controls uptake of sodium accompanied by a chloride ion from the lumen into the kidney epithelial cell in the distal convoluted tubule, DCT [Article:10653443]. Sodium is then returned to the bloodstream via the ATP-dependent sodium-potassium pump, Na,K-ATPase [Article:1328175] (see figure 1). The Na,K-ATPase is coded for by several genes of the ATPase family, which show tissue specific expression [Article:1328175].
The uptake of sodium by SLC12A3 can be blocked by thiazide diuretics, or by rare, hereditary loss of function mutations in the SLC12A3 gene that result in Gitelman's syndrome [Articles:10653443, 12590198]. Thiazides inhibit the binding of chloride to SLC12A3, which then prevents the co-transport of sodium [Article:9167652]. In order to have its pharmacological effect the thiazide must first be excreted into the lumen by a transporter.
Both the trafficking of SLC12A3 to the membrane and phosphorylation and activation of SLC12A3 is controlled by the With No lysine kinases, WNKs [Article:22405999]. As seen in figure 1, the WNKs are central to a complex regulatory network that balances the relative roles of the different transporters and ion channels in the different parts of the kidney (for a full review of the detailed signaling see [Article:22405999]). In the DCT, WNK3 activates SLC12A3 both directly and via STK39 (also known as SPAK, or Ste20/SPS1-related kinase) [Articles:21907141, 21610494]. WNK1 may also activate STK39 [Article:22405999]. STK39 phosphorylates SLC12A3 in the DCT leading to its activation. STK39 can be differently spliced, with the short form or kinase defective form (KD-SPAK), which does not phosphorylate SLC12A3 or OXSR1, being expressed in the tall ascending limb of the kidney [Article:21907141]. SLC12A3 is inhibited by WNK4 by promotion of its degradation via the lysosomal pathway [Article:22405999], however angiotensin II (AngII) may act to counteract this by switching WNK4 to stimulate STK39 and so stimulate SLC12A3 [Article:22820370]. Also a kidney specific splice variant of WNK1 inhibits SLC12A3 by inhibition of WNK1, which in turn inhibits WNK3 [Article:22405999]. WNK1 can also inhibit the WNK4 effects on SLC12A3 [Article:22405999]. Mutations in the WNK1 or WNK4 genes cause FHHt, familial hyperkalemic hypertension or Gordon syndrome or pseudohypoaldosteronism type II, whereby SLC12A3 is overactive and results in hypertension [Article:17975670]. Patients with FHHt can have six to seven times the average response to HCT [Article:12107233].
There is also regulation of SLC12A3 by aldosterone [Article:21852580]. Aldosterone increases SLC12A3 protein expression by activation of SGK1, which phosphorylates NEDD4L and prevents NEDD4L binding to and inhibiting SLC12A3 [Article:21852580].
Other hormones may also influence the activity of SLC12A3. Male patients with mutations in SLC12A3 tend to have more severe symptoms of Gitelman syndrome than female patients [Article:17329572]. Estrogens have been shown to increase SLC12A3 density at the membrane in rats [Article:9541496].
The Na,K-ATPase is formed from three subunits alpha, beta and gamma. Several genes encode the catalytic alpha subunit with different tissue expression patterns. ATP1A1 is the predominant form expressed in the kidney [Article:1328175]. There are also several genes for the beta subunit (ATP1B1, ATP1B2, ATP1B3)[Article:9457675]. The gamma subunit is coded for by FXYD2 [Article:19865785]. Mutations in FXYD2 result in hypomagnesemia and hypocalciuria [Article:19865785]. In vitro studies have shown that the Na,K-ATPase is regulated by phosphorylation on the alpha subunit by protein kinase A (PKA) and protein kinase C (PKC)[Articles:22242112, 22162761, 22433860]. Phosphorylation can also affect the transporters trafficking to the membrane with G-protein coupled receptor kinases (GRKs) and arrestins promoting phosphorylation and phosphatase PP2AC reversing this process [Article:22242112]. The FXYD2 subunit expression is alternatively spliced and the ratios of these may be effected by the transcription factor HIF1B [Article:21130072]. Variants in the ARNT gene that codes for HIF1B may therefore result in hypomagnesemia [Article:21130072].
The distal convoluted tubule is responsible for only around five percent of the sodium reabsorbed within the nephron therefore thiazide diuretics have only a limited capacity to change sodium balance [Article:10653443] However, these drugs are often sufficient in patients with uncomplicated hypertension and no concurrent renal disease [Article:10653443]. One of the side effects of thiazides is hypokalemia, reduced serum potassium, since potassium is lost to the urine to counterbalance the electrochemical gradient when sodium resorption is inhibited [Article:20095916]. Thiazide-related potassium loss is dose dependent and at high doses can lead to arrhythmia and sudden cardiac death [Article:20095916]. Therefore, rather than increase the dose, if low doses of thiazides are insufficient to reduce blood pressure a second antihypertensive drug is added, such as an ACE inhibitor or AngII receptor blocker which often reduces the renal potassium excretion as well as synergistically reducing blood pressure [Article:20095916]. The mechanism for this counter-balancing is likely via AngII on WNK4 (as discussed above) [Article:22820370].
The main target of loop diuretic is the sodium potassium chloride co-transporter, NKCC2, SLC12A1, which is found predominantly in the thick ascending limb of the loop of Henle (i.e. earlier in the resorption process than the distal convoluted tubule where thiazides act). Thirty percent of the filtered sodium chloride is reabsorbed in the thick ascending limb [Article:15044637]. SLC12A1 transports sodium, potassium and two chloride ions across the apical membrane. Potassium is recycled out across the apical membrane by the potassium channel ROMK, coded for by KCNJ1 [Article:15044637]. At the basolateral membrane, sodium is exported from the cell via the Na,K-ATPase whereas chloride diffuses via the channels ClC-Ka and ClC-Kb, coded for by CLCNKA and CLCKNB [Article:21503667]. The membrane location of the chloride channels is regulated by barttin, BSND. Inactivating mutations in any of these genes (SLC12A1, KCNJ1, CLCNKA, CLCNKB and BSND) can result in Bartter syndrome, where sodium, water, potassium, chloride and calcium homeostasis is disrupted [Article:21503667].
As with the thiazide sensitive transporter, the phosphorylation and location of SLC12A1 at the membrane can be modulated by WNK signaling [Article:22405999]. An undefined WNK phosphorylates OXSR1 which in turn phosphorylates SLC12A1 and increases function [Article:21972418] OXSR1 is inhibited by the kinase-deficient splice variant of STK39 [Article:21907141].
Around 1-2% of sodium resorption occurs in the collecting duct [Article:19940300]. This region is the main site of action of the potassium-sparing diuretics, amiloride and triamterene, which target the epithelial sodium channel, ENaC. ENaC is coded for by the SCNN1 family of genes, SCNN1A, SCNN1B, SCNN1G, SCNN1D whose products combine to form a heterotrimeric protein, although SCNN1D may not be a relevant candidate gene for efficacy since it is more prominently expressed in tissues other than kidney [Article:22573384]. Mutations SCNN1B or SCNN1G can cause increased channel activity and Liddle syndrome, characterized by and early onset hypertension and hypokalemia [Article:9637708].
The WNKs are also involved in regulation of SCNN1 at the membrane [Article:22405999]. WNK1 and WNK4 can increase the amount of SCNN1 at the membrane [Article:22405999]. Aldosterone, acting via SGK1 and NEDD4L can also lead to increased SCNN1 at the membrane [Article:21852580]. Increased intracellular sodium leads to down regulation of SCNN1, and this mechanism is defective in Liddle syndrome [Article:9637708].
There have been a considerable number of PGx studies of thiazide diuretics (see table 2). A few studies have also looked at other diuretics or grouped drugs by class [Articles:20877298, 21692745, 17460608](see table 3). The majority of these have taken candidate gene approaches, confirming genes already known to be involved in hypertension and blood pressure control, with two GWAS studies involving diuretics published to date [Articles:18591461, 22566498].
The results of these associations can be loosely grouped into six intersecting sets, (the first three are depicted on figure 1): 1, variants that act at the SLC12A3 transporter (in SLC12A3, NEDD4L, WNK1), 2. variants that act at the Na,K-ATPase (in ADD1, NEDD4L), 3. variants that act at the SCNN1 transporter (in SCNN1G, NEDD4L, WNK1), 4. variants that act in the RAAS pathway (in ACE, AGT, AGTR1, CYP11B, REN, KNG1, NOS3, ), 5. variants that act in the beta-adrenergic pathway (in ADRB2, ADRB3, GNB3), 6. variants with other or unknown pathway effects (in SLC22A6 and SLC22A8, which are genes involved in the PK of thiazides and loop diuretics; YEATS4, a transcription factor; NT5C2; GRK5).
To date there have been few attempts to replicate associations identified in either candidate studies or GWAS. Some candidates have been replicated but further replication has failed in other cohorts eg ACE, ADD1. This appears to be a common problem with pharmacodynamic PGx, given that the clearest examples where PGx has been sufficiently replicated to be applied to practice in dosing recommendations (warfarin, codeine, mercaptopurines) are all PK. Many believe that will have to use some kind of complex network prediction for PD PGx [Article:22923055]. Some of the difficulty may be due to the different phenotypes measured (ambulatory vs office blood pressure), treatment regimen (treatment naïve vs long term treatment, or polytherapy) and heterogeneity of the patients (gender, race, age, dietary salt intake). While some suggest that pharmacogenetic effects of common variants may be larger than the common variants influencing blood pressure and hypertension, and therefore may be detectable in GWAS, if the effect sizes are as small, both the candidate gene and GWA studies so far are likely severely underpowered. This suggests that many of the reported associations may be false positives, and underscores the importance of replication studies.
Hypertension is a complex disease. If looked at from a molecular perspective it is a phenotype (elevated blood pressure) that results from alterations in the multiple pathways of blood pressure regulation. As such, determining the likely efficacy of different therapies for hypertension is challenging. However a large number of candidate genes and SNPs for both disease etiology and response to therapies have already been identified. There is a great need for these to be replicated in studies with sufficient size and power. Using current knowledge to better define molecular subtypes of hypertension in order to select more homogeneous patient populations during study design may enable the pharmacogenomics to be predicted.
Thorn Caroline F, Ellison David H, Turner Stephen T, Altman Russ B, Klein Teri E. "PharmGKB summary: Diuretics pathway, pharmacodynamics" Pharmacogenetics and genomics (2013).
If you would like to reproduce this PharmGKB pathway diagram:
Entities in the Pathway
Drugs/Drug Classes (5)
Relationships in the Pathway
Download data in TSV format. Other formats are available on the Downloads/LinkOuts tab.