NaCl cotransporter activity and Mg²⁺ handling by the distal convoluted tubule

Abstract

The genetic disease Gitelman syndrome, knockout mice, and pharmacological blockade with thiazide diuretics have revealed that reduced activity of the NaCl cotransporter (NCC) promotes renal Mg²⁺ wasting. NCC is expressed along the distal convoluted tubule (DCT), and its activity determines Mg²⁺ entry into DCT cells through transient receptor potential channel subfamily M member 6 (TRPM6). Several other genetic forms of hypomagnesemia lower the drive for Mg²⁺ entry by inhibiting activity of basolateral Na⁺-K⁺-ATPase, and reduced NCC activity may do the same. Lower intracellular Mg²⁺ may promote further Mg²⁺ loss by directly decreasing activity of Na⁺-K⁺-ATPase. Lower intracellular Mg²⁺ may also lower Na⁺-K⁺-ATPase indirectly by downregulating NCC. Lower NCC activity also induces atrophy of DCT cells, decreasing the available number of TRPM6 channels. Conversely, a mouse model with increased NCC activity was recently shown to display normal Mg²⁺ handling. Moreover, recent studies have identified calcineurin and uromodulin (UMOD) as regulators of both NCC and Mg²⁺ handling by the DCT. Calcineurin inhibitors paradoxically cause hypomagnesemia in a state of NCC activation, but this may be related to direct effects on TRPM6 gene expression. In Umod^−/− mice, the cause of hypomagnesemia may be partly due to both decreased NCC expression and lower TRPM6 expression on the cell surface. This mini-review discusses these new findings and the possible role of altered Na⁺ flux through NCC and ultimately Na⁺-K⁺-ATPase in Mg²⁺ reabsorption by the DCT.

INTRODUCTION

Mg²⁺ is the second most abundant intracellular divalent cation and plays an essential role in many fundamental cellular processes (15). It is a cofactor for Na⁺-K⁺-ATPase and many kinases, regulates stabilization of nucleotides, and is critical for mitochondrial function, ion channels, and neuromuscular excitability (15, 29). Therefore, Mg²⁺ deficiency is associated with systemic defects but often presents with symptoms of neuromuscular hyperexcitability such as tremor, muscle cramps, tetany, and generalized seizures (15). Hypomagnesemia is frequently associated with hypokalemia and/or hypocalcemia, which can result in cardiac arrhythmia and sudden death (15). Mg²⁺ deficiency is also linked to various diseases including type 2 diabetes mellitus, hypertension, atherosclerosis, dyslipidemia, myocardial infarction, and psychiatric disorders (15).

Plasma Mg²⁺ concentration is maintained by the interplay of renal reabsorption, intestinal absorption, and bone exchange (29). In the kidney, ∼70% of total plasma Mg²⁺ is filtered by the glomeruli, of which ∼96% is reabsorbed along the nephron. Most reabsorption occurs along the thick ascending limb of the loop of Henle (TAL; 50–70%), with less along the proximal convoluted tubule (PCT; 10–25%) (54). Only 3–7% of the filtered Mg²⁺ is reabsorbed in the distal convoluted tubule (DCT), but this segment determines urinary excretion since it is the major site of regulated reabsorption (54). The critical role of the TAL in Mg²⁺ handling is demonstrated by the effects of claudin disruption of Mg²⁺ handling by this segment. However, the DCT plays a critical role in fine tuning Mg²⁺ homeostasis, as demonstrated by the presentation of hypomagnesemia as a central feature of Gitelman syndrome, caused by inactivating mutations in the DCT-specific NaCl cotransporter (NCC) (74), and it can also occur as a side effect of thiazide diuretics, which inhibit NCC. Moreover, several other congenital syndromes that lead to DCT dysfunction can present with renal Mg²⁺ wasting, hypomagnesemia, and hypocalciuria, resembling Gitelman syndrome (84), and thus confirming the critical role of the DCT in Mg²⁺ homeostasis.

Epidermal growth factor (EGF) is highly expressed along the DCT, and proEGF is cleaved to generate active EGF that binds to EGF receptor (EGFR) at the basolateral surface (31). EGF signaling has been shown to play an important role in Mg²⁺ homeostasis. Recently, uromodulin (UMOD; also known as Tamm-Horsfall protein) and calcineurin have been identified as novel factors modulating Mg²⁺ handling by the DCT, providing new insights into the relationship between NCC activity and Mg²⁺ homeostasis. Other recent studies have explored the link between Mg²⁺ and NCC activity more directly, using dietary Mg²⁺ restriction and a genetic mouse model that results in hyperactive NCC. We also briefly discuss the possible role of changes in intracellular Cl⁻ concentration in Mg²⁺ handling. This mini-review discusses these new findings, focusing on how altered Na⁺ flux through NCC and ultimately Na⁺-K⁺-ATPase may be a central regulator of Mg²⁺ reabsorption.

Mg²⁺ REABSORPTION ALONG THE DCT

Along the DCT, Mg²⁺ is transported via active transcellular pathways, which has been functionally divided into early (DCT1) and late (DCT2) subsegments (50, 54). Apical Mg²⁺ entry along DCT1 and DCT2 is mediated by transient receptor potential channel subfamily M member 6 (TRPM6) (69, 86), likely in complex with transient receptor potential channel subfamily M member 7, which is required for maximal expression of TRPM6 (68). Mg²⁺ entry through the channel is driven by the electrical potential across the apical membrane (−70 mV), which may be established by the voltage-gated K⁺ channel (Kv1.1) (Fig. 1) (24, 52). TRPM6 mutations result in hypomagnesemia and renal Mg²⁺ wasting (86), whereas patients with Kv1.1 mutations fail to lower fractional renal Mg²⁺ excretion in the presence of hypomagnesemia (24). Intestine-specific Trpm6^−/− mice suffer from hypomagnesemia, whereas kidney-specific Trpm6^−/− mice show normal Mg²⁺ homeostasis (11). It remains unclear what causes this discrepancy between mice and humans; however, it is well established that inhibition of EGFR that increases TRPM6 transcription and activation (37, 78) induces hypomagnesemia with renal Mg²⁺ wasting in both mice and humans (16, 65). In addition, mutations in the gene encoding EGF result in isolated, recessive renal hypomagnesemia (31).

Fig. 1.

Apical and basolateral membrane proteins that participate in Mg²⁺ handling in the early distal convoluted tubule (DCT1). NaCl cotransporter (NCC), voltage-gated K⁺ channel Kv1.1, and transient receptor potential channel subfamily M member 6 (TRPM6) are expressed in the apical membrane, whereas Na⁺-K⁺-ATPase, Kir4.1/Kir5.1, voltage-gated Cl⁻ channel ClC-K2, SLC41A1, and cyclin and CBS domain divalent metal cation transport mediator-2 (CNNM2) are expressed in the basolateral membrane. Mg²⁺ entry through TRPM6 is driven by the electrical potential across the apical membrane (−70 mV), which may be established by Kv1.1. Parvalbumin (PV) has been proposed to act as the intracellular Mg²⁺ shuttle. SLC41A1 and CNNM2 may play a key role in basolateral Mg²⁺ extrusion. Yellow circles represent Mg²⁺. FXYD2, Na⁺-K⁺-ATPase γ-subunit.

The intracellular protein that binds and transports Mg²⁺ to maintain normal intracellular Mg²⁺ levels during transcellular reabsorption has not been identified (54), but the cytoplasmic Ca²⁺/Mg²⁺-binding proteins parvalbumin (PV) and calbindin D28K (CaBP28) are relatively highly expressed along the DCT and may serve this role. PV is specifically expressed in DCT1 (7, 50), whereas CaBP28 is expressed along with apical Ca²⁺ transporter and basolateral Ca²⁺-extruding proteins in DCT2 and the connecting segment. This different distribution suggests that PV may act as the intracellular Mg²⁺ shuttle and that the major site for transcellular Mg²⁺ reabsorption is the DCT1 (Fig. 1). In support of a dominant role for PV, Pv^−/− mice may display mild urinary Mg²⁺ wasting (63), whereas Cabp28^−/− mice do not (47). However, the question remains of whether there is even the need for intracellular proteins to buffer intracellular Mg²⁺. The presence of CaBP28 keeps intracellular Ca²⁺ very low (0.12 µmol/L) despite ongoing transepithelial flux (54), whereas the intracellular concentration of Mg²⁺ is much higher (∼800 µmol/L) and is similar to tubular and plasma Mg²⁺ concentrations (52, 54).

The basolateral Mg²⁺ extrusion mechanism remains unclear, but SLC41A1, which acts as a Na⁺/Mg²⁺ exchanger, has been implicated (Fig. 1) (43). A SLC41A1 mutation that reduces its expression in patients with nephronophthisis-related disorders by an in-frame deletion of a transmembrane helix shows a defect in Mg²⁺ transport activity in human embryonic kidney (HEK)-293 cells (35). Another Na⁺/Mg²⁺ exchanger, SLC41A3, is also highly expressed in the DCT (14), but its role is less clear. Slc41a3^−/− mice showed hypomagnesemia but no change in urinary Mg²⁺ excretion (14). Furthermore, it has been previously reported that SLC41A3 is a mitochondrial Mg²⁺ transporter rather than a plasma membrane protein (51). Cyclin and CBS domain divalent metal cation transport mediator-2 (CNNM2) might also play a key role in the Mg²⁺ reabsorptive mechanism (Fig. 1), although its true function has not yet been clarified (23). CNNM2 localizes at the basolateral membrane of the DCT in the human kidney, and mutations in patients cause renal Mg²⁺ wasting and hypomagnesemia (75). Recently, Cnnm2^+/− mice and kidney-specific Cnnm2^−/− mice have been reported to display lower Mg²⁺ levels in serum but no increase in urinary Mg²⁺ excretion, suggesting no defect in renal Mg²⁺ reabsorption (21). Of the three, SLC41A3 and CNNM2 deficiency do not cause renal Mg²⁺ wasting, so SLC41A1 is considered to be the prime candidate to facilitate Mg²⁺ extrusion; generation of DCT-specific Slc41a1^−/− mice might resolve the issue of basolateral Mg²⁺ extrusion.

NCC AND Na⁺-K⁺-ATPASE IN DCT Mg²⁺ TRANSPORT

DCT Mg²⁺ transport depends upon NCC function. Loss-of-function mutations in SLC12A3, which encodes NCC, cause Gitelman syndrome, characterized by hypokalemia, hypocalciuria, metabolic alkalosis, and renal Mg²⁺ wasting and hypomagnesemia (74). Ncc^−/− mice and mice treated with thiazide diuretics, which specifically inhibit NCC, phenotypically resemble Gitelman syndrome and display marked atrophy of DCT1 (50, 60). Although hydrochlorothiazide treatment also inhibits the Na⁺-driven Cl^–/${\text{HCO}}_3^{-}$ exchanger and pendrin in β-intercalated cells of the collecting duct (48), this is probably not relevant to DCT Mg²⁺ handling. Furthermore, mice lacking STE20/SPS1-related proline-alanine-rich protein kinase (SPAK), a kinase that phosphorylates NCC, display almost complete ablation of NCC activity and decreased DCT1 mass (28). These findings strongly suggest that NCC activity influences DCT1 mass (Fig. 2, A and B). Trpm6 mRNA and protein expression are significantly lower in Ncc^−/− mice (60). DCT1 atrophy might therefore simply result in lower TRPM6 expression as a consequence of lower DCT1 mass (Fig. 2A). However, TRPM6 expression levels were reduced in mice with 6 days of thiazide treatment despite strong NCC expression and no evidence of DCT atrophy (60). More recently, experiments in 1-day-old Ncc^−/− mice showed lower TRPM6 protein expression before impaired DCT1 outgrowth was observed (70). Thus, NCC may indirectly influence Trpm6 mRNA and protein expression before DCT1 atrophy is manifested, through unidentified pathways.

Fig. 2.

Schematic model of the effect of NaCl cotransporter (NCC) activity on distal convoluted tubule (DCT) Mg²⁺ handling. A: NCC inactivation reduces early DCT (DCT1) mass and Na⁺-K⁺-ATPase activity, leading to hypomagnesemia. This is caused by decreased Mg²⁺ entry through transient receptor potential channel subfamily M member 6 (TRPM6). B: NCC activation increases DCT mass, but TRPM6 per unit DCT1 mass is downregulated, resulting in a minimal effect on Mg²⁺ entry. C: Na⁺-K⁺-ATPase activity is a central determinant of DCT Mg²⁺ reabsorption because it provides the driving force for Mg²⁺ entry via TRPM6. Hepatocyte nuclear factor-1β (HNF1B) plays a key role in Mg²⁺ handling, since it promotes transcription of Na⁺-K⁺-ATPase subunit-γ (FXYD2) and Kir5.1, which enhance Na⁺-K⁺-ATPase activity. Kir4.1/Kir5.1 activity is closely related to NCC activity. D: Mg²⁺ deficiency may promote NCC degradation via ubiquitin ligase neuronal precursor cell developmentally downregulated 4-2 (NEDD4-2). Increased Na⁺ delivery to downstream segments can promote K⁺ secretion. Mg²⁺ deficiency upregulates TRPM6 expression, which enhances Mg²⁺ entry, but NCC downregulation can cause DCT1 atrophy. Combined with reduced levels of the intracellular Mg²⁺ concentration ([Mg²⁺]_i) for Na⁺-K⁺-ATPase activity, this would be expected to reduce Mg²⁺ entry via TRPM6 channels.

Although loss of NCC activity causes hypomagnesemia, increased NCC activity does not cause hypermagnesemia. NCC is activated by phosphorylation of several serine and threonine residues at its NH₂ terminal. The pathways involved in this process are complex (72), but in simple terms, with no lysine kinase 4 (WNK4) phosphorylates and activates SPAK, which then directly phosphorylates NCC. This NCC-activating pathway is regulated by inhibition of WNK4, which can occur by two mechanisms. First, direct binding of Cl⁻ prevents WNK4 autophosphorylation (5). Second, abundance of WNK4 is decreased by proteasomal degradation following ubiquitination by an E3 ubiquitin ligase consisting of the scaffold cullin 3, the substrate adaptor Kelch-like 3 (KLHL3), and the ubiquitin conjugating ligase RING (12). The disease familial hyperkalemic hypertension (FHHt) is caused by mutations in the genes encoding WNK4, cullin 3, and KLHL3 as well as WNK1 (which results in its ectopic expression) (53). The net effect of these mutations is to inappropriately activate NCC, as reflected by the fact that all features of FHHt (which also include metabolic acidosis and hypercalciuria) are reversed by thiazides. FHHt is the mirror image of Gitelman syndrome, with the notable exception that it is not associated with Mg²⁺ imbalance (53). Consistent with this, in mice with DCT1-specific expression of constitutively active mutant SPAK (CA-SPAK), which display NCC hyperactivation, serum Mg²⁺, urinary Mg²⁺ excretion, and Trpm6 mRNA levels were similar to those of control mice (83). Since CA-SPAK mice display increased DCT1 mass (27), this suggests compensatory downregulation of TRPM6 (lower TRPM6 per unit DCT1 mass) to maintain Mg²⁺ homeostasis. One caveat is that mRNA abundances may not be directly match protein levels, so TRPM6 protein abundances and activities need to be evaluated. Another recent study confirmed that NCC activation is not associated with altered Mg²⁺ homeostasis (Fig. 2B). Low dietary K⁺ intake strongly activates NCC via WNK/SPAK-mediated phosphorylation via activation of the Kir4.1/Kir5.1 heterotetramer in the basolateral membrane of the DCT (88). van der Wijst et al. (82) demonstrated that although low-K⁺ diet activated NCC, it did not alter Trpm6 mRNA expression, serum Mg²⁺ concentration, or Mg²⁺ excretion. In both studies, TRPM6 currents were not measured, and compensatory mechanisms in other segments may also contribute to maintenance of Mg²⁺ homeostasis. Another explanation proposed for the different effects of NCC inactivation and activation is that membrane potential does not change much with NCC activation (52). However, the actual membrane potential of the DCT has not been measured while NCC activity was manipulated. Although activation or inactivation of NCC activity would affect Kir4.1/Kir5.1 activity through Na⁺-K⁺-ATPase coupling, this may not change the basolateral membrane potential because Cl⁻ efflux through voltage-gated Cl⁻ channel ClC-K2 should also change in the same direction (see potential role of intracellular cl⁻ in mg²⁺ handling by the dct for more details). Another explanation is that the effect of NCC activation on DCT1 mass may be smaller than that of NCC inactivation (Fig. 2, A and B). Indeed, the DCT1 mass of Ncc^−/− mice is reduced to <0.1-fold (50), whereas that of CA-SPAK mice is increased to only ∼1.4-fold (27), and TRPM6 mRNA abundance is decreased in Ncc^−/− mice but not in CA-SPAK mice. Thus, reduced TRPM6 due to DCT atrophy may be central in causing Mg²⁺ wasting with NCC inactivation, but with NCC activation DCT mass and TRPM6 are disconnected.

Genetic disorders that influence Na⁺-K⁺-ATPase function and cause urinary Mg²⁺ wasting confirm the importance of ATPase activity for DCT Mg²⁺ reabsorption (84). These include loss-of-function mutations of other genes highly expressed in DCT, including Na⁺-K⁺-ATPase subunit-γ (FXYD2) (55), inward rectifying K⁺ channel subfamily J member 10 (KCNJ10) (66), hepatocyte nuclear factor-1β (HNF1B) (1), and pterin-4-α-carbinolamine dehydratase 1 (PCBD1) (18). FXYD2 interacts with Na⁺-K⁺-ATPase, modulating its activity (39, 52). Fxyd2^−/− mice show destabilized and depressed Na⁺-K⁺-ATPase activity (39) and a nonsignificant trend to increased urinary Mg²⁺ excretion (39, 52). In addition, Na⁺-K⁺-ATPase activity depends on K⁺ recycling via Kir4.1/Kir5.1 (encoded by KCNJ10/KCNJ16) expressed in the basolateral membrane of the DCT (Fig. 2C). Kir4.1/Kir5.1 activity is closely related to NCC activity (Fig. 2C), as demonstrated by dramatically lower reduced NCC expression and phosphorylation in Kir4.1 knockout mice (13); thus, kidney-specific deletion of Kir4.1 in mice causes DCT atrophy (76), which may also contribute to hypomagnesemia. HNF1β binds the FXYD2 promoter and activates its transcription (Fig. 2C), whereas PCBD1 regulates FXYD2 transcription as a dimerization cofactor for HNF1B (18). A recent study has demonstrated that HNF1β also activates the Kcnj16 promoter, and kidney-specific Hnf1b disruption downregulated mRNA expression of not only Kir5.1 but also Kir4.1 (44). Thus, HFN1B promotes Mg²⁺ reabsorption by regulating Na⁺-K⁺-ATPase activity both directly and indirectly (Fig. 2C).

Altered NCC activity might also be expected to affect Mg²⁺ homeostasis through effects on Na⁺-K⁺-ATPase activity. NCC activation and the following elevation of intracellular Na⁺ concentration might be expected to increase Na⁺-K⁺-ATPase activity, but, as discussed above, NCC activation (e.g., in FHHt) is not associated with altered Mg²⁺ handling in part due to effects on TRPM6 expression (Fig. 2B). Whether NCC activation significantly alters Na⁺-K⁺-ATPase activity has not been determined. In contrast, NCC inactivation and the resulting decrease in NaCl entry may have a significant inhibitory effect on Na⁺-K⁺-ATPase, thus reducing the driving force for Mg²⁺ entry via TRPM6 (Fig. 2A) (52). Indeed, Na⁺-K⁺-ATPase activity of rat DCT segments is ∼30–50% lower after thiazide treatment (22, 56). Although the link between NCC activity and basolateral Mg²⁺ extrusion is unclear, Mg²⁺ extrusion via SLC41A1 is considered to be driven by the inward Na⁺ electrochemical gradient generated by Na⁺-K⁺-ATPase (52). Thus, NCC inactivation may also affect the driving forces for basolateral Mg²⁺ extrusion.

The DCT is clearly important for Mg²⁺ reabsorption, so Mg²⁺ depletion exerts significant effects on DCT function. Dietary Mg²⁺ depletion stimulates expression of Trpm6 and PV (30). The highest Na⁺-K⁺-ATPase activity along the renal tubule is found in the DCT (40). Mg²⁺ can directly influence Na⁺-K⁺-ATPase activity (2) and is required for mitochondrial production of ATP to power Na⁺-K⁺-ATPase (29). Thus, there is reason to suspect that Mg²⁺ depletion may impair DCT Na⁺ transport by reducing Na⁺-K⁺-ATPase function (Fig. 2D), and this segment may be particularly susceptible. Mg²⁺ depletion may also impair DCT Na⁺ transport by reducing NCC activity. We recently found that dietary Mg²⁺ restriction as short as 3 days decreased abundances of total and phosphorylated NCC (17). NCC is ubiquitinated and targeted for degradation by neural precursor cell expressed, developmentally downregulated 4-2 (NEDD4-2) (Fig. 2D) (3). In mice lacking tubular NEDD4-2, Mg²⁺ restriction no longer reduced total or phosphorylated NCC, suggesting that Mg²⁺ depletion increases NEDD4-2 activity (17). Although the mechanism by which Mg²⁺ depletion activates NEDD4-2 has not been determined, one possibility is that the decrease in intracellular Mg²⁺ induced by dietary Mg²⁺ restriction elevates the intracellular free ionized Ca²⁺ concentration (90). NEDD4-2 can be activated by either directly binding Ca²⁺ via its Calb/C2-binding domain, which releases its autoinhibition, or after binding to substrate (87). Reduced Na⁺ entry would also lower Na⁺ available to drive Na⁺-K⁺-ATPase, exacerbating reductions in Na⁺ and Mg²⁺ entry; DCT1 atrophy would have an additional detrimental effect. However, Mg²⁺ restriction upregulates protein abundance of the Na⁺-K⁺-ATPase activator HNF1B (81). To resolve the relative contributions of these opposing forces on Na⁺-K⁺-ATPase activity, further studies are needed to evaluate DCT1 mass and Na⁺-K⁺-ATPase activity during Mg²⁺ restriction. For an extensive recent discussion of the mechanisms linking Na⁺ and Mg²⁺ reabsorption along the DCT, the reader is referred to Franken et al. (19).

Mg²⁺ depletion has long been known to promote urinary K⁺ excretion, sustaining hypokalemia (34). Classically, this phenomenon is explained by Mg²⁺-mediated suppression of outward K⁺ currents through the renal outer medullary K⁺ channel (ROMK) in the connecting segment and collecting duct, such that Mg²⁺ depletion allows unbridled K⁺ secretion through ROMK (34). We recently found in mice that the effect of dietary Mg²⁺ restriction to downregulate NCC overrode the effect of K⁺ restriction to activate it (17). This observation provides a plausible mechanism for increased Na⁺ delivery and flow in the setting of combined hypomagnesemia and hypokalemia.

POTENTIAL ROLE OF INTRACELLULAR Cl⁻ IN Mg²⁺ HANDLING BY THE DCT

Changes in intracellular Cl⁻ concentration play an important role in regulating NCC. Cl⁻ prevents WNK4 autophosphorylation and activation by directly binding to its catalytic domain (5), so decreased intracellular Cl⁻ concentration phosphorylates and activates the WNK4-SPAK-NCC pathway (Fig. 3). ClC-K2 is mainly expressed in the TAL and DCT (32) and is known to be the main pathway for basolateral Cl⁻ exit in DCT cells (32, 42). The driving force for Cl⁻ extrusion via ClC-K2 is provided by the inside negativity of the membrane potential, which is determined by Kir4.1/Kir5.1 activity (32, 88). Recently, Su et al. (77) generated a transgenic mouse model expressing an optogenetic kidney-specific Cl⁻ sensor and found that chemical inhibition of Kir4.1/Kir5.1 significantly increased intracellular Cl⁻ concentration. Thus, intracellular Cl⁻ levels are in a large part determined by Kir4.1/Kir5.1 activity and regulate NCC activation (Fig. 3). Cl⁻ efflux through ClC-K2 may also be modulated independently of Kir4.1/Kir5.1, but this has not yet been demonstrated.

Fig. 3.

Schematic model of the effect of intracellular Cl⁻ on distal convoluted tubule (DCT) Mg²⁺ handling. Left: decreased intracellular Cl⁻ levels ([Cl⁻]_i) activate the with no lysine kinase 4 (WNK4)-STE20/SPS1-related proline/alanine-rich kinase (SPAK)-NaCl cotransporter (NCC) pathway. The driving force for Cl⁻ extrusion via the voltage-gated chloride channel ClC-K2 is provided by the inside negativity of the membrane potential, which is determined by Kir4.1/Kir5.1 activity. Serum K⁺ levels regulate Kir4.1/5.1 activity. Right: ClC-K2 deficiency is predicted to increased [Cl⁻]_i, leading to inhibition of the WNK4-SPAK-NCC pathway. This induces DCT1 atrophy, decreasing transient receptor potential channel subfamily M member 6 (TRPM6) expression and activity, with the consequence of Mg²⁺ wasting. FXYD2, Na⁺-K⁺-ATPase subunit-γ.

Loss-of-function mutations in the CLCNK2 gene that encodes ClC-K2 cause Bartter syndrome type 3, but some patients display renal Mg²⁺ wasting and phenotypes similar to Gitelman syndrome (20, 61). Therefore, CLC-K2 may affect not only Mg²⁺ reabsorption along the TAL but also along the DCT1 by modulating the WNK4-SPAK-NCC cascade. Recently, Hennings et al. (32) generated Clcnk2^−/− mice that displayed a phenotype resembling this syndrome and a profound decrease in total and phosphorylated NCC. Consistent with this, these mice exhibited no response to furosemide but also a severely blunted response to thiazide (32). Interestingly, Clcnk2^−/− mice showed a marked atrophy of the DCT1 and renal Mg²⁺ wasting (32). These effects of Clcnk2^−/− mice are similar to those of renal epithelium-specific kcnj10^−/− mice (13, 76, 91) that phenocopy the human disease epilepsy, ataxia, sensorineural deafness, and tubulopathy (EAST) syndrome (also called SeSAME) caused by mutations in KCNJ10 (8, 71). ClC-K2 deficiency would be predicted to increased intracellular Cl⁻ concentration, leading to inhibition of WNK4. This would lower NCC activity and hence induce DCT1 atrophy, decreasing TRPM6 expression and activity, with the consequence of Mg²⁺ wasting (Fig. 3).

DCT Mg²⁺ HANDLING AND CALCINEURIN INHIBITION

Calcineurin inhibitors (CNIs) such as tacrolimus (previously referred to as FK506) and cyclosporin A are immunosuppressants widely used to prevent rejection after solid-organ transplantation and to treat autoimmune disease. CNI use is frequently accompanied by hypomagnesemia and hypertension through effects on the kidney (57). Calcineurin is a Ca²⁺/calmodulin-dependent serine/threonine phosphatase that influences expression of downstream genes by dephosphorylating transcription factors, such as nuclear factor of activated T cells, but can also act through nongenomic mechanisms (49). The DCT expresses both the calcineurin-α and -β isoforms (45), suggesting that effects on this segment may contribute to CNI side effects; studies in rodents support this.

CNI-induced hypertension is mediated in part by increased NCC activity due to reduced NCC dephosphorylation (Fig. 4) (33). To inhibit calcineurin and exert immunosuppressive effects, tacrolimus must bind to the 12-kDa FK506-binding protein (FKBP12), but it can also have calcineurin-independent effects. To address whether the effect of tacrolimus is calcineurin dependent, Lazelle et al. (45) generated inducible renal epithelium-specific FKBP12 knockout (KS-Fkbp12^−/−) mice and found that tacrolimus-induced NCC activation and hypertension were absent in KS-Fkbp12^−/− mice. These results strongly suggest that tacrolimus influences the calcineurin pathway (Fig. 4).

Fig. 4.

Role of 12-kDa FK506-binding protein (FKBP12) in distal convoluted tubule (DCT) Mg²⁺ handling. NaCl cotransporter (NCC) activity and transient receptor potential channel subfamily M member 6 (TRPM6) expression under normal conditions (left) and after tacrolimus treatment (right) are shown. TRPM6 is regulated by epidermal growth factor (EGF) that is generated by cleavage from proEGF and stimulates EGF receptor (EGFR). Stimulation of EGFR promotes TRPM6 transcription via the EGF-ERK1/2- activator protein-1 (AP-1) pathway and increases cell surface protein abundance and activation. The tacrolimus-FKBP12 complex inhibits calcineurin, leading to hypertension and hypomagnesemia as a result of NCC activation and TRPM6 downregulation. Tacrolimus treatment can decrease Egf mRNA expression and transcriptional activity of AP-1, leading to TRPM6 downregulation.

Dietary K⁺ alters NCC phosphorylation via modulation of the WNK-SPAK signaling pathway (88). High-K⁺ diet may also activate calcineurin by inducing cell depolarization, which opens voltage-dependent Ca²⁺ channels, providing another possible pathway for NCC regulation (73). Furthermore, CNIs increase abundance of WNK kinases and SPAK (33). Recent findings demonstrate that tacrolimus increases WNK4 abundance by preventing calcineurin-mediated dephosphorylation of the ubiquitin-ligase complex component KLHL3 (38). Phosphorylation of KLHL3 impairs degradation of WNK4. Thus, tacrolimus prevents WNK4 degradation, resulting in increased SPAK activity, with more NCC phosphorylation, increasing DCT Na⁺ reabsorption and blood pressure (Fig. 4).

The effects of tacrolimus raise a paradox since tacrolimus causes NCC activation associated with lower Mg²⁺ reabsorption and hypomagnesemia. A recent study by Gratreak et al. (26) exploring Mg²⁺ homeostasis in KS-Fkbp12^−/− mice may provide a possible explanation. Administration of tacrolimus to rats profoundly downregulates Trpm6 mRNA and protein levels, leading to renal Mg²⁺ wasting (Fig. 4) (59). In KS-Fkbp12^−/− mice, tacrolimus-induced hypomagnesemia and reduction of Trpm6 mRNA were completely abolished (26), suggesting that calcineurin downregulates TRPM6 expression through a pathway that differs from its NCC regulatory pathway. CNI-treated rats have shown downregulation of Egf mRNA paralleling effects on Trpm6 (Fig. 4) (46) that may hint at mechanism. In HEK-293 cells, EGFR activation upregulated TRPM6 transcription via the ERK1/2 activator protein-1 (AP-1) pathway and increased TRPM6 cell surface protein abundance and activation (Fig. 4) (37). In rats, EGF administration did not rescue CNI-induced TRPM6 downregulation, despite the fact that its administration to control rats upregulated TRPM6 expression (46). In addition, in renal tubular epithelial cell lines, CNI inhibited the ERK1/2-AP-1 pathway that upregulates Trpm6 mRNA (36). Therefore, calcineurin may affect Mg²⁺ transport by regulating TRPM6 expression via a EGF-ERK1/2-AP-1 pathway (Fig. 4). However, conflicting data from Kiely et al. (41) showed that CNI activated ERK1/2 in Madin-Darby canine kidney cells. Whether this difference is a consequence of the cell type examined or other technical differences remains unclear. Decreased EGF-EGFR-ERK1/2 pathway activation may also affect NCC activity. A recent study by Cheng et al. (10) reported that EGFR tyrosine kinase inhibition resulted in lower ERK1/2 phosphorylation, lower NCC activity, and increased Na⁺ excretion, suggesting a direct effect of EGFR on NCC. This NCC inactivation should contribute to decreased Mg²⁺ entry via TRPM6 by decreased Na⁺-K⁺-ATPase activity. However, the effect of EGF-EGFR-ERK1/2 pathway by CNI on NCC is likely to be smaller than that of WNK4-SPAK pathway because CNI activates NCC as discussed above.

Several questions remain. How is it that CNIs stimulate NCC activity but suppress TRPM6 expression? Is this because they influence DCT1 and DCT2 differently? Do they alter cell morphology or number in the DCT? Since calcineurin can act independently of nuclear factor of activated T cells (49), another possibility is differential interactions with other signaling molecules. Further studies will be required to evaluate these hypotheses.

UMOD AND DCT Mg²⁺ HANDLING

UMOD is a glycosylphosphatidyl-inositol-anchored protein that is targeted to the apical membrane of tubular cells, where it is subsequently cleaved and released into the lumen by the serine protease hepsin (9). UMOD is highly expressed in the TAL and after cleavage is excreted in the urine, where it is the most abundant protein in healthy subjects (89). Urinary UMOD protects against urinary tract infections and renal stone formation and stimulates renal salt reabsorption (89).

Umod^−/− mice exhibit renal Mg²⁺ wasting with compensatory upregulation of mRNA expression of Trpm6, Hnf1b, Egf, Fxyd2, and Pv (58). Since UMOD enhances the surface expression of the Na⁺-K⁺-2Cl⁻ cotransporter (NKCC2) and ROMK (89), its deletion is likely to indirectly affect paracellular Mg²⁺ along the TAL. However, although Umod^−/− mice display increased urinary Mg²⁺ wasting, administration of a loop diuretic to these mice increases urinary Mg²⁺ to a degree similar to that seen in control mice (58). This suggests effects of UMOD deletion on additional segments. Indeed, in the DCT, despite increasing Trpm6 mRNA, UMOD deletion decreased apical TRPM6 protein abundance (58). Cotransfection experiments in HEK-293 cells showed that UMOD enhanced TRPM6 cell surface abundance and current density. Whole cell patch-clamp recording of HEK-293 cells treated with purified UMOD enhanced TRPM6 cell surface abundance and current density by impairing TRPM6 endocytosis (58). UMOD could exert similar effects on TRMP6 in native DCT1 cells to influence Mg²⁺ handling by this segment (Fig. 5A).

Fig. 5.

Schematic model of the effect of uromodulin (UMOD) on distal convoluted tubule (DCT) Mg²⁺ handling. A: UMOD levels and transient receptor potential channel subfamily M member 6 (TRPM6) cell surface abundance differ between normal conditions (left) and low-Mg²⁺ conditions (right). UMOD can enhance TRPM6 cell surface abundance and current density by impairing TRPM6 endocytosis. Furthermore, luminal UMOD levels increase under low-Mg²⁺ conditions. Thus, luminal UMOD may play a role in Mg²⁺ reabsorption along the DCT by regulating cell surface expression of TRPM6 and by altering NaCl cotransporter (NCC) activity. B: effects of intracellular UMOD on NCC. Left, UMOD can facilitate NCC phosphorylation and activation. Right, UMOD deficiency is associated with decreased phosphorylated NCC in the early DCT (DCT1), leading to increased NCC phosphorylation in the late DCT (DCT2). Decreased NCC activity in DCT1 may cause inhibition of Na⁺-K⁺-ATPase activity, further contributing to decreased Mg²⁺ entry via TRPM6.

Although UMOD is classically thought of as a TAL marker, Tokonami et al. (80) found significant UMOD expression in mouse and human DCT1, at ∼10% of TAL expression levels. The Ca²⁺-sensing receptor (CaSR) plays a facilitating role, since its activation decreased UMOD excretion by reducing intracellular levels of cAMP, which is involved in the trafficking of many proteins in tubular cells (79). Another study showed that genetic and pharmacological inhibition of ROMK lowered urinary UMOD excretion, suggesting that a membrane potential-dependent trafficking mechanism might influence UMOD excretion (67). Since hepsin, CaSR, and ROMK are all expressed in the DCT (9, 25, 85), they could regulate UMOD trafficking and release not only in the TAL but also along the DCT. Therefore, UMOD synthesized in the DCT1 itself may also modulate Mg²⁺ reabsorption by this segment via direct effects on TRPM6 as discussed above.

Global disruption of UMOD was associated with decreased phosphorylated NCC in DCT1, consistent with a coexpression study of UMOD and NCC in HEK-293 cells, suggesting a facilitating role for UMOD in NCC activation (Fig. 5B) (80). UMOD excretion was significantly increased in Mg²⁺-restricted wild-type mice (58). Therefore, UMOD might be expected to stimulate Mg²⁺ reabsorption along the DCT by activating NCC, but we found lower phosphorylated NCC in Mg²⁺-restricted mice (17), suggesting that other pathways (e.g., NEDD4-2 activation) override the ability of UMOD to stimulate NCC during Mg²⁺ deficiency. However, there is evidence that UMOD may contribute to Mg²⁺ handling by the DCT. Tokonami et al. (80) reported that NCC activity was lower in DCT1 but higher in DCT2 in Umod^−/− mice at baseline. Chronic furosemide treatment increased NCC phosphorylation along DCT2 but not along DCT1 in Umod^−/− mice, despite increased NCC phosphorylation along DCT1 in control mice (80). Importantly, acute addition of hydrochlorothiazide to chronic furosemide treatment decreased Mg²⁺ excretion in wild-type mice but not in Umod^−/− mice. NCC inactivation along DCT1 in Umod^−/− mice should inhibit Na⁺-K⁺-ATPase activity, further contributing to decreased Mg²⁺ entry via TRPM6 (Fig. 5B). Thus, regulation of NCC activity by UMOD along DCT1 might also contribute to normal Mg²⁺ handling.

Further complexity is added when UMOD regulatory proteins expressed along both the TAL and DCT are considered. CaSR inhibits UMOD but activates NCC. CaSR also binds and responds to Mg²⁺, and DCT cells modulate intracellular signals in response to changes in extracellular Mg²⁺ as well as in extracellular Ca²⁺ (4). Since CaSR is expressed in the apical membrane of DCT (25), reduced luminal Mg²⁺ might lower CaSR activation. Lower CaSR activity would be predicted to increase UMOD release from the DCT1 by lowering CaSR-mediated cAMP inhibition of UMOD trafficking, stimulating TRPM6 to prevent Mg²⁺ loss. In contrast, CaSR stimulation activates NCC through the KLHL3-WNK4-SPAK pathway in vitro and in vivo (6). Therefore, lower CaSR activation during Mg²⁺ deficiency might result in NCC inhibition.

Hepsin, which cleaves and releases UMOD, is expressed at similar levels in the DCT and TAL (9). However, Olinger et al. (62) showed that in mice carrying a hepsin missense mutation cellular UMOD accumulation affected NKCC2 activity but not total and phosphorylated NCC. This raises the possibility that UMOD is differentially regulated in the two segments. However, given that UMOD deletion differentially influences NCC activation in DCT1 and DCT2, and since NCC was only evaluated in whole kidney, the two segments should be evaluated separately. On the other hand, DCT1-made UMOD might be still released because another protease, prostasin, which directly cleaves UMOD in vitro (9), is strongly expressed in the DCT (64). Although its deletion did not change urinary UMOD levels (9), the effect of prostasin on UMOD release from DCT1 remains unknown. Whether hepsin activity changes in response to changing dietary Mg²⁺ intake is unknown. Finally, how ROMK stimulation of urinary UMOD excretion contributes to Mg²⁺ handling has not been explored.

There are other unanswered questions regarding the contribution of UMOD synthesized by the DCT1 to Mg²⁺ handling. Can it increase TRPM6 membrane abundance before its hepsin-mediated cleavage? Although urinary UMOD excretion increases with Mg²⁺ restriction, is this additional UMOD derived from both the TAL and DCT1? What are the relative contributions of UMOD effects along the TAL and DCT in renal Mg²⁺ handling? To fully understand the roles of UMOD and its regulators in Mg²⁺ handling by the DCT, TAL- and DCT-specific knockout models will be needed.

SUMMARY

Recent studies have identified calcineurin and UMOD as regulators of renal Mg²⁺ handling through effects on the DCT. Several lines of evidence suggest that inhibition of NCC might contribute to hypomagnesemia by altering the drive for apical Mg²⁺ entry by inhibiting Na⁺-K⁺-ATPase. This is supported by the etiology of other monogenic diseases that resemble Gitelman syndrome. Mg²⁺ deficiency may itself promote Mg²⁺ loss by directly lowering Na⁺-K⁺-ATPase activity since it serves as a cofactor, and downregulation of NCC by decreased intracellular Mg²⁺ may exacerbate this. In contrast, inappropriate activation of NCC, as seen in FHHt or CA-SPAK mice, does not lead to altered Mg²⁺ handling. This may be due to the fact that Na⁺-K⁺-ATPase is high along the DCT compared with other segments, rendering it more susceptible to conditions that lower Na⁺-K⁺-ATPase activity. CNIs paradoxically cause hypomagnesemia in a state of NCC activation, but this may be related to direct effects on TRPM6 gene expression or selective effects on DCT2. Further studies are needed to work out precisely how these regulating factors interact with dietary Mg²⁺ and the mechanisms underlying altered DCT Mg²⁺ entry.

GRANTS

Y.M. received a postdoctoral fellowship from Uehara Memorial Foundation, and J.A.M. is funded by National Institute of Diabetes and Digestive and Kidney Diseases Grant DK098141.

DISCLOSURES

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

AUTHOR CONTRIBUTIONS

Y.M. drafted manuscript and prepared figures; J.A.M. edited and revised manuscript; Y.M. and J.A.M. approved final version of manuscript.