Magnesium Homeostasis

Abstract

Mg²⁺, the fourth most abundant cation in the body, serves as a cofactor for about 600 cellular enzymes. One third of ingested Mg²⁺ is absorbed from the gut through a saturable transcellular process and a concentration-dependent paracellular process. Absorbed Mg²⁺ is excreted by the kidney and maintains serum Mg²⁺ within a narrow range of 0.7–1.25 mmol/L. The reabsorption of Mg²⁺ by the nephron is characterized by paracellular transport in the proximal tubule and thick ascending limb. The nature of the transport pathways in the gut epithelia and thick ascending limb has emerged from an understanding of the molecular mechanisms responsible for rare monogenetic disorders presenting with clinical hypomagnesemia. These human disorders due to loss-of-function mutations, in concert with mouse models, have led to a deeper understanding of Mg²⁺ transport in the gut and renal tubule. This review focuses on the nature of the transporters and channels revealed by human and mouse genetics and how they are integrated into an understanding of human Mg²⁺ physiology.

Magnesium, the fourth most abundant cation in the body and the second most abundant cation in intracellular fluid,¹ serves as a cofactor for about 600 cellular enzymes,² including GTPase, Na⁺-K⁺-ATPase, adenylate cyclase, phosphoinositide kinases, and phosphofructokinase.³ Mg²⁺ plays critical roles in protein and nucleic acid synthesis and ligand receptor interactions⁴ where it exerts its major biologic influences and is important for parathyroid hormone release from the gland. In humans, Mg²⁺ is distributed such that a little more than 50% is in the bone and a little less than 50% is in the muscle and other soft tissues. Serum contains about 0.3% of total body Mg²⁺ and red blood cells about 0.5%. Serum Mg²⁺ concentration is kept within a very narrow range, 0.7–1.25 mmol/L (1.4–2.5 mg/dl), and these levels are influenced by absorption from the gastrointestinal tract⁵ and excretion by the kidney.^6,7 In mammalian cells, total intracellular Mg²⁺ varies from 10 to 30 mmol/L, and most are protein or nucleotide bound.⁸ Approximately 0.5–1.2 mmol/L is free in the cytosol, comparable with concentrations in extracellular fluid. Mg²⁺ balance is tightly regulated by channels and transporters,⁹ and intracellular Mg²⁺ concentration is determined by the orchestration of Mg²⁺ influx and efflux across plasma membranes and reciprocal exchange from organelles such as mitochondria. A Mg²⁺ sensor has not been unequivocally identified in mammals. MgtE, a Mg²⁺ transporter in prokaryotes, is regulated by a Mg²⁺-specific riboswitch and is capable of Mg²⁺ sensing.¹⁰ The SLC41 family of transporters¹¹ comprises three family members: SLC41A1, SLC41A2, and SLC41A3; is expressed in all eukaryotes; and displays distant homology to the prokaryotic MgtE family of Mg²⁺ transporters.^12,13 Expression of SLC41A1 and SLC41A3 in Xenopus laevis (X. laevis) oocytes demonstrated Mg²⁺ currents at physiological Mg²⁺ concentrations, suggesting a channel-like function.^14,15 Whether this family of eukaryotic proteins may be the putative Mg²⁺ sensor is yet to be determined; however, the observation that the Slc41a3^−/− mouse is hypomagnesemic¹⁶ is intriguing.

Intestinal Mg²⁺ Transport

The principal site of magnesium absorption in the gut is the small intestine. The daily recommended intake of magnesium for adults is 300–400 mg/d.¹⁷ About one third of that is absorbed, and roughly the same amount is excreted in the urine to maintain balance. Absorption occurs by two distinct pathways, a saturable active transcellular and a nonsaturable passive concentration dependent paracellular pathway (Figure 1). Hypomagnesemia can result from structural alterations in intestine that decrease surface area for absorption, such as bariatric surgery, short bowel syndromes, and Whipple procedure, or decrease in transit time, which does not allow for equilibration, such as steatorrhea, Crohn disease, and chronic diarrhea syndromes.

Figure 1.

Nature of transport processes in the gut. In the gastrointestinal tract, predominantly the small intestine reabsorbs Mg²⁺ by two distinct processes. There is a saturable transcellular process that is engaged at low to normal Mg²⁺ intake and a concentration-dependent paracellular absorptive process at which high oral Mg²⁺ intake becomes the major route of intestinal absorption. These combined processes result in absorption of about one third of oral ingestion of Mg²⁺.

Transcellular Transport

The molecular basis for the apical magnesium entry step was initiated by the identification and cloning of TRPM6. TRPM6 is one of the eight members of the transient receptor potential melastatin (TRPM) cation channel subfamily members and is composed of 2022 amino acids, encoded by a large gene containing 39 exons located on chromosome 9q. TRPM6 and TRPM7 are members of a subgroup of Mg²⁺ transporters that belong to the TRPM superfamily of cationic channels.^18,19 Intestinal-specific Trmp6^−/− mice suffer from hypomagnesemia,²⁰ and the phenotype can be rescued by dietary supplementation of Mg²⁺. TRPM proteins possess six predicted transmembrane domains, and a hydrophobic region between transmembrane domains 5 and 6 forms the cationic pore when four channel subunits segregate to form the functional transporter.²¹ TRPM6 is found along the full length of the intestine, in the kidney, in the distal convoluted tubule, and in the lung and testis.^19,22 TRPM6 protein is localized to the apical membranes of the gastrointestinal tract and distal convoluted tubule, forming the gateway for entry and Mg²⁺ influx into the epithelial cells.¹⁵ Some groups have shown that functional expression of TRPM6 in the cell membrane requires coexpression with TRPM7, thus suggesting the formation of a heteromeric ion channel.^22,23 Loss-of-function mutations in the TRPM6 gene leads to the syndrome of hypomagnesemia with secondary hypocalcemia (HSH).^24,25 HSH (OMIM 607009) is an autosomal recessive disorder characterized by low serum magnesium levels due to diminished intestinal absorption of Mg²⁺ and renal Mg²⁺ wasting. Inherited mutations in the TRPM6 gene include premature stop codons, exon deletions, frameshift mutations, and inserted splice sites, which lead to endoplasmic retention.

Although the molecular nature of the magnesium entry step is understood in intestinal cells, the exit step is still a little murky. Potential candidates are the SLC41 family of proteins. All three members of the family are expressed in the human intestine.¹² Mg²⁺ absorption is similar in Slc41a3^+/+ and Slc41a3^−/− mice,¹⁶ suggesting that SLC41A3 does not play a critical role in the exit step. SLC41A1 does not seem to be a significant player,²⁶ but we need more information on the potential role of SLC41A2 in this process. Another group of potential transporters involved in the magnesium exit step are the cyclin M family of proteins, of which there are four (CNNM1–4). The name results from the weak sequence homology shared with the cyclin family, although a cyclin-like function has never been shown for CNNM1–4. CNNM2 (previously known as ACDP2) is the most conserved between man and mouse. Cnnm2^+/− mice display higher fecal Mg²⁺ content than Cnnm2^+/+ mice,²⁷ and Cnnm2 transcripts were significantly reduced in intestine of Cnnm2^+/− compared with Cnnm2^+/+ mice, and Cnnm2^+/− mice show mild hypomagnesemia with increased fecal Mg²⁺ content. Another mouse model has suggested that Cnnm4 is the basolateral located Mg²⁺ exit mechanism.²⁸ CNNM4 is strongly expressed in the intestine in a basolateral location. Cnnm4^−/− mice show hypomagnesemia because of intestinal malabsorption of magnesium. Although we cannot, on the basis of current information, unequivocally state the molecular nature of the Mg²⁺ extrusion step in humans, CNNM2 and CNNM4 emerge as strong candidates. Intestinal influx and exit steps are illustrated in Figure 2.

Figure 2.

Detailed mechanisms of gastrointestinal transcellular absorption. In the gut lumen, the apical entry step for transcellular transport is mediated by TRPM6/7, which is mutated in hypomagnesemia with secondary hypocalcemia (HSH) resulting in loss of function and contributing to the hypomagnesemia phenotype in this syndrome. For the basolateral exit step, several candidates have emerged. Experimental studies in Cnnm4^−/− mice show that these animals express hypomagnesemia and excrete high levels of Mg²⁺ in feces. In addition, Cnnm2+/− mice have higher Mg²⁺ in stools when compared with wild-type Cnnm2+/+ mice. In mice, Slc41a1 and Slc41a3 do not appear to play a critical role in intestinal Mg²⁺ absorption; however, we have limited information on the potential role of Slc41a2. All three variants are expressed in the intestine. The molecular nature of the players involved in paracellular transport is unclear.

Paracellular Transport

Increasing oral intake of magnesium is associated with increasing absorption through the paracellular pathway.⁵ This process is not infinite, however, as magnesium is a cathartic. With increasing oral ingestion, intestinal motility increases and transit times are shortened, leading to frequent bowel motions and frank diarrhea. Increasing oral intake is clinically effective, however, in correcting hypomagnesemia in patients and can be increased to the tolerability of the individual patient. Tight junctions between cells may comprise three sets of macromolecules: claudins, occludins, and tricellulins. The specific nature of the Mg²⁺ pore has not been delineated; however, the claudin family of proteins has attracted attention because all these channel formers exhibit at least one of the three types of selectivity: for cations (claudin-2, -10b, and -15), for anions (claudin-10a and -17), and for water (claudin-2).²⁹ Chronic use of proton pump inhibitors (PPIs) appears to exert their effects in part by raising intestinal pH and also inhibiting Mg²⁺ absorption through inhibition of paracellular transport.^30–32 In clinical hypomagnesemia due to PPIs, the kidney appears to appropriately lower its fractional excretion of magnesium (FE_Mg2+) and, therefore, can be significantly worsened by coadministration of drugs that promote renal magnesium wasting, e.g., diuretics.

Renal Mg²⁺ Transport

Seventy percent to 80% of serum Mg²⁺ is ultrafilterable, and the kidney adapts to wide ranges of magnesium intake by lowering the FE_Mg2+ to 0.5% or raising it to 80% as needed to maintain balance.³³ The kidney cannot reduce Mg²⁺ excretion to zero, and, therefore, patients in intensive care units without Mg²⁺ supplementation have a high risk for developing hypomagnesemia.³⁴ Patients administered magnesium sulfate in the management of eclampsia can excrete 90% of the load within 24 hours. Because the kidney is the sole route of Mg²⁺ elimination, clinical hypermagnesemia, although uncommon, is most likely observed in patients with kidney failure after the oral administration of Mg²⁺ containing cathartics or Mg²⁺ containing enemas. Normally, more than 95% of the filtered load of Mg²⁺ is reabsorbed by the mammalian renal tubule. The proximal tubule reabsorbs 20%–26%, and the thick ascending limb of Henle (TALH) accounts for 65%–70% of the filtered load. At both sites, the reabsorption is paracellular and passive. The distal convoluted tubule, particularly the early distal convoluted tubule, reabsorbs about 10%, and transport is transcellular, regulated, and, under normal physiologic states, determines the urinary Mg²⁺ content. It is useful to calculate the FE_Mg2+ to assess whether a low serum magnesium is due to decreased intake and/or intestinal absorption or renal Mg²⁺ wasting. Renal Mg²⁺ wasting is considered when the FE_Mg2+ is 2% or greater in a patient with hypomagnesemia.

Proximal Tubule

The adult proximal tubule reabsorbs about 20%–25% of the filtered load of Mg²⁺ but has a limited role in the conservation of magnesium. Intraluminal concentrations of Mg²⁺ rise as water is reabsorbed along the proximal nephron, thus delivering higher concentrations to the loop.^35–37 In the neonatal kidney, magnesium concentration remains similar to the ultrafiltrate along the rat proximal nephron such that the fractional magnesium along the nephron is in the order of 60%–70%.³⁸ The physiological basis for the difference between neonatal and adult proximal nephron may be a less mature paracellular pathway, which would allow more magnesium to move across with sodium, calcium, and water. Although the nature of the paracellular barrier is unknown, claudins 2 and 10a are expressed in the proximal tubule and remain possible candidates for this role.

Loop of Henle

The principal site of magnesium reabsorption is the TALH.^39,40 de Rouffignac and Quamme have observed magnesium transport in the cortical TALH. This transport is passive, and net Mg²⁺ absorption is predominantly dependent on the transepithelial voltage.^41,42 The molecular nature of this channel has evolved from a clearer understanding of the defect in familial hypomagnesemia with hypercalciuria and nephrocalcinosis (FHHNC). This is a rare autosomal renal tubulopathy, characterized by impaired renal reabsorption of magnesium and calcium in the TALH. Patients usually present with recurrent urinary tract infections, polyuria, nephrolithiasis, and nephrocalcinosis. Additional symptoms include rickets, hematuria, muscle tetany, seizures, failure to thrive, vomiting, and abdominal pain.^43,44 Twelve kindreds with recessive FHHNC (OMIM 248250) demonstrated linkage to a segment of chromosome 3q.⁴⁵ The open reading frame of the identified gene was called paracellin-1 (PCN-1), and sequence analysis showed it encoded a protein of 305 amino acids with four transmembrane domains and intracellular NH₂- and COOH-termini. The PCN-1 protein showed sequence and structural similarity to the members of the claudin family, of which there are currently 27 members.⁴⁶ Since the identification of PCN-1 (later renamed claudin-16), there have been more than 30 additional families with FHHNC that express various CLDN16 mutations.^47–50 RNAi knockdown of claudin-16 in transgenic mice is associated with Mg²⁺ and Ca²⁺ wasting in urine, nephrocalcinosis, and a reduction of the lumen-positive transepithelial potential,⁵¹ which is a determinant of the passive paracellular flux of Mg²⁺ in the TALH.

Konrad and colleagues⁵² have characterized a Swiss and eight Spanish/Hispanic families whose renal symptoms phenocopy the families with CLDN16 mutations but had additional ocular abnormalities such as macular coloboma, nystagmus, and myopia. In these families, no CLDN16 mutations could be demonstrated.^53,54 Genome-wide analysis demonstrated linkage to a region of chromosome 1q34.2 (OMIM 248190). The mutated gene emerged as CLDN19, which is found in a few organs but whose highest expression is in the kidney and eye. Similar renal expression patterns for CLDN16 and CLDN19 and protein expression studies suggest a colocalization of both claudins.⁵² Furthermore, coimmunoprecipitation studies and freeze fracture images confirm an interaction at the tight junctions of TALH calls. This interaction forms a heterodimer, Figure 3, with nonselective cation properties, and allows lumen to basolateral flux of Mg²⁺ and Ca²⁺ and basolateral to lumen flux of Na⁺ to establish the transepithelial voltage.^55,56 Because the defects involve a reduction in channel insertion at the tight junction, as well as a reduction of the transepithelial voltage, these patients tend to have severe hypomagnesemia.

Figure 3.

Essential features of paracellular Mg²⁺ reabsorption in TALH. The cortical TALH cell tight junctions are constituted by a heterodimer of claudin-16 and claudin-19, which form the nonselective cation channel. The transmembrane potential is generated by the movement of Na⁺ through the channel driven by the concentration gradient (interstitium>tubular lumen). Impairment of transmembrane potential can occur with loss of function of NKCC2, decreased ROMK, and CLC-Kb activity, all of which increase luminal Na⁺ concentration and decrease transmembrane potential. The CaSR inhibits ROMK through formation of 20-HETE acid and increases claudin-14 expression. 20-HETE, 20-hydroxyeicatetraenoic; CaSR, calcium-sensing receptor; ROMK, renal outer medullary K1; TALH, thick ascending limb of Henle.

Impaired magnesium absorption in the TALH may occur in Bartter syndrome. The syndrome is characterized by sodium and chloride wasting and hypokalemic metabolic alkalosis. The syndrome results from mutations in at least five genes in the TALH-encoding proteins, which directly or indirectly regulate sodium and chloride reabsorption in the TALH. They include autosomal recessive mutations in SLC12A1 encoding NKCC2 (OMIM 601678),⁵⁷ the apical inward rectifier KCNJ1 encoding the apical inward rectifier renal outer medullary K⁺ (ROMK) (OMIM 241200),⁵⁸ a basolateral chloride channel CLCNK encoding the basolateral chloride channel CLC-Kb (OMIM 602023),⁵⁹ and BSND encoding Barttin (OMIM 602522), an essential activator $\beta$-subunit for CLC-Kb.⁶⁰ Autosomal dominant gain-of-function mutant of the basolateral calcium-sensing receptor (CaSR) (OMIM 601198)⁶¹ can cause a variant of Bartter syndrome by inhibition of ROMK. It should be noted that the CaSR can have additional effects on Ca²⁺ reabsorption through effects on claudin-14.⁶² This variant results in a decrease of K⁺ flux through ROMK and a reduction of the Na⁺ gradient, thus reducing the flux of sodium through the claudin tight junction cation channel resulting in a decrease in transepithelial voltage favoring Mg²⁺ reabsorption. Other rare genetic diseases worth mentioning here are the HELIX syndrome, which results from mutations in CLDN10 and mutations in MAGELD2, which presents with a transient antenatal Bartter syndrome phenotype.⁶³ Because the cation channel is intact, the Mg²⁺ losses tend to be mild, and only 20% of patients with Bartter syndrome exhibit hypomagnesemia.

Certain classes of drugs produce hypomagnesemia by affecting Mg²⁺ reabsorption in the TALH. Loop diuretics produce their clinical effect by inhibiting NKCC2 on the apical side of the TALH,⁶⁴ and thus may effect a pathophysiology similar to that seen in antenatal Bartter syndrome. In experimental animals, sirolimus (rapamycin) downregulates the expression of NKCC2⁶⁵ and thus can initiate a similar pathophysiology as loop diuretics. Aminoglycosides stimulate the CaSR on the basolateral side of the TALH, thus producing pathophysiology similar to one of the Bartter syndrome variants.⁶⁴

Distal Convoluted Tubule

Early understanding of Mg²⁺ transport in the distal convoluted tubule has been gleaned from micropuncture and microperfusion studies.^66,67 A more complete understanding of Mg²⁺ transport, the molecular nature of the transporters, and their regulation in the distal convoluted tubule has evolved by characterization of clinical syndromes and the genetic mutations that are responsible for the hypomagnesemia that result. It is useful to discuss transcellular Mg²⁺ transport in the distal convoluted tubule in two parts: the entry step at the apical membrane and the exit step at the basolateral membrane.

The thiazide-sensitive NaCl electroneutral transporter (NCC), distributed along the apical membrane of the distal convoluted tubule, is energized by the Na⁺ gradient generated by the basolateral Na⁺-K⁺-ATPase.^68,69 A favorable membrane potential generated by the potassium channel Kv1.1 facilitates the entry of Mg²⁺ through the apical Mg²⁺ channel, TRPM6/7,⁷⁰ which is the rate-limiting step. Identification of human disorders that are associated with hypomagnesemia and the apical membrane proteins mutated in these disorders has aided a clearer understanding of factors regulating the apical entry step in the distal convoluted tubule. Pharmacological inhibition of NCC with thiazides⁷¹ or in mouse models of Gitelman syndrome,⁷² loss of NCC is associated with tubular atrophy and can lead to loss of apical TRPM6. Whether this observation has implications for the 47 million patients on thiazides in the United States is unknown.

Gitelman syndrome (OMIM 263800) is an autosomal recessive uncommon salt-losing tubulopathy characterized by hypokalemic metabolic alkalosis, hypomagnesemia, and hypocalciuria. Loss-of-function mutations in the SLC12A3 gene encoding NCC are the cause of most of these cases.^73,74 Eighteen percent to 40% of patients with clinical Gitelman syndrome are usually found to carry only one mutant allele after SLC12A3 screening.^75,76 In a large cohort of patients with Gitelman syndrome, missense mutations (approximately 59%) and a predisposition to large rearrangements (approximately 63%) caused by repeated sequences within the SLC12A3 gene⁷⁷ were observed. Recently it has been shown that intronic variants in the SLC12A3 gene were associated with a Gitelman-like phenotype.⁷⁸ Hypomagnesemia is a defining feature of Gitelman syndrome and appears to be a consequence of downregulation of TRPM6 in the distal convoluted tubule at the mRNA and protein level.⁷⁹ A Gitelman-like phenotype syndrome has also been described in patients with mitochondrial DNA variants in MT-TF and MT-TI.⁸⁰ By decreasing ATP generation in distal convoluted tubule, these variants will inhibit NCC function, thus decreasing TRPM6 expression through distal convoluted tubule atrophy.

HSH (OMIM 602014) is a result of autosomal recessive mutations in TRPM6,²¹ which has a restricted expression pattern in resorptive epithelia, the kidney, exclusively in the distal convoluted tubule and gastrointestinal tract.^24,25,70 Loss-of-function mutations cause a severe phenotype because there is impaired absorption of Mg²⁺ in gastrointestinal tract and renal Mg²⁺ wasting. Extracellular Mg²⁺ can affect TRPM6 function since Mg²⁺ restriction significantly upregulates the levels of mRNA encoding TRPM6. Renal tubule–specific Trmp6^−/− mice have produced conflicting results. The Chubanov group using a Sox2-Cre could show no effects on Mg²⁺ handling²⁰; however, using a Six2-Cre, Funato et al. showed hypomagnesemia and urinary Mg²⁺ wasting.⁸¹

In 2009, Glaudemans and colleagues described a Brazilian family with isolated autosomal dominant hypomagnesemia (OMIM 176260) and identified a missense mutation in KCNA1, resulting in loss of function of the encoded voltage-gated potassium channel Kv1.1.⁸² This mutation resulted in a N255D amino acid substitution in a highly conserved asparagine in S3 of Kv1.1, and immunohistochemistry demonstrated colocalization with TRPM6 in the apical membrane of the distal convoluted tubule. Later, a patient with tetany and hypomagnesemia was described who had a L328V amino acid substitution in Kv1.1, which was in TM loop 5 of the channel.⁸³ Both mutations were associated with loss of function in the potassium channel and suggests that Kv1.1 contributes to the positive transmembrane potential of the distal convoluted tubule necessary to facilitate Mg²⁺ entry through TRPM6.

Isolated renal hypomagnesemia (OMIM 131530) was characterized in a Dutch family with affected individuals presenting with hypomagnesemia and mental retardation.⁸⁴ The mutation mapped to chromosome 4q25 demonstrated autosomal recessive transmission and resulted in a P1070L amino acid substitution in pro-EGF. Normally, pro-EGF is sorted to both apical and basolateral membranes⁸⁵; however, the P1070L mutation disrupts the highly conserved P1070 required for the basolateral PXXP sorting motif.⁸⁶ The renal epidermal growth factor receptor (EGFR) is on the basolateral membrane of the distal convoluted tubule and thus inaccessible to the mutant pro-EGF because of impaired basolateral sorting. EGF stimulates TRPM6 specifically through the EGFR, and receptor engagement stimulates the Src family of tyrosine kinases and the small Rho-GTPase, Rac-1, resulting in a redistribution of vesicular TRPM6 to the apical plasma membrane.⁸⁷ Sixty-five percent of patients treated with cetuximab or panitumumab (EGFR-blocking antibodies) develop hypomagnesemia within 8 weeks of therapy. About 43% of patients treated with cisplatin develop hypomagnesemia.⁸⁸ Although the mechanisms for this observation are unclear, the evidence for downregulation of the TRPM6/EGF pathway in a rat model of cisplatin nephrotoxicity⁸⁹ suggest that this may explain the mechanism for hypomagnesemia in humans undergoing cisplatin-based therapy. The features of the apical entry step are illustrated in Figure 4.

Figure 4.

Regulation of apical Mg²⁺ entry in the distal convoluted tubule. In the distal convoluted tubule, TRPM6/7 is the rate-limiting pore for Mg²⁺ entry. Loss-of-function mutations in TRPN6 cause HSH, a severe form of hypomagnesemia presenting in early life. The apical expression of TRPM6 is regulated by EGF/EGFR, which promotes apical TRPM6 expression by modulating vesicular trafficking, and loss of EGF results in downregulation of apical TRPM6. Loss-of-function mutations or inhibition in NCC on the apical membrane cause a downregulation of TRPM6 through distal convoluted tubule atrophy. Finally, the driving force for Mg²⁺ absorption is the transmembrane potential generated by K⁺ extrusion through Kv1.1. Loss-of-function mutations in KCNA1, gene encoding Kv1.1, causes isolated autosomal dominant hypomagnesemia. DCT, distal convoluted tubule; EGFR, epidermal growth factor receptor.

In 1985, a group of Dutch investigators identified two unrelated families with renal Mg²⁺ wasting and autosomal dominant inheritance⁹⁰ (OMIM 6018114). More detailed analysis showed that the disease mapped to chromosome 11q23.⁹¹ The mutation associated with dominant isolated renal magnesium was due to a G41R amino acid substitution in the transmembrane region of the γ-subunit of Na⁺-K⁺-ATPase (FXYD2).⁹² FXYD2 is a type I transmembrane protein and is a regulator of pump activity.⁹³ The exact mechanism by which FXYD2 controls Mg²⁺ handling and the exit step at the basolateral membrane remains elusive. Expression of wild-type FXYD2 and the G41R mutant in X. laevis oocytes demonstrated that the G41R mutant generated whole-cell ion currents with a novel Mg²⁺-dependent gating on inward rectification.⁹⁴ Whether this is the exit step for Mg²⁺ at the basolateral membrane or whether FXYD2 interacts with other channel proteins on the basolateral membrane to affect this step is still unresolved. One view is that renal Mg²⁺ wasting associated with Na⁺-K⁺-ATPase/FXYD2 interaction is explained by destabilization and inactivation of Na⁺-K⁺-ATPase.⁹⁵

Additional support for Na⁺-K⁺-ATPase/FXYD2 in the renal reabsorption of Mg²⁺ was obtained in studies in a teenage patient with mutation in hepatocyte nuclear factor 1B (HNF1B), who presented with tetany and hypomagnesemia.⁹⁶ HNF1B is a transcription factor, mutations in which have been associated with MODY5 (OMIM 137920), a rare monogenetic form of diabetes mellitus,⁹⁷ in which about 50% of affected patients present with hypomagnesemia This disorder maps to chromosome 17q12 and exhibits autosomal dominant transmission. It is likely that HNF1B binds to a region near the splice site for FXYD and preferentially drives transcription of FXYD2b in the distal convoluted tubule.⁹⁶ Two highly conserved HNF1B binding sites have been recognized in the FXYD2 promoter region, and mutations in HNF1B disrupt this binding event. Of note, PCBD1 is a coactivator of HNF1B-mediated transcription necessary for fine-tuning FXYD2 transcription in the distal convoluted tubule, and mutations in the PCBD1 gene cause hypomagnesemia and renal Mg²⁺ wasting⁹⁸ (OMIM 126090).

The α1 subunit of Na⁺-K⁺-ATPase is encoded by ATP1A1 and is highly expressed in the kidney and central nervous system. De novo variants in ATP1A1 have been described, and some patients have hypomagnesemia⁹⁹ (OMIM 618314). The mechanisms may be due to reduced activity of the pump, which leads to downregulation of NCC and distal convoluted tubule atrophy as described earlier.

Two independent groups identified mutations in KCNJ10, which is mapped to chromosome 1q23 and encodes the inward rectifying K⁺ channel Kir4.1.^100,101 These investigators described the phenotypes as seizures, sensorineural deafness, ataxia, mental retardation, and electrolyte imbalance (SeSAME) syndrome and epilepsy, ataxia, sensorineural deafness, and tubulopathy (EAST) syndrome. The mutations were transmitted in an autosomal recessive pattern, and consanguinity was detected in several of the cohorts. Kir4.1 is expressed in epithelial cells of the kidney and inner ear as well as glial cells of the central nervous system.¹⁰² Loss of function caused by the various mutations in each case are very different; however, they provide genetic insights into the structural basis of pH sensing, gating, surface expression, and channel stability¹⁰² and probably accounts for the variability in clinical severity observed. In the kidney, Kir4.1 is expressed along the basolateral membrane of the distal convoluted tubule where it colocalizes with Na⁺-K⁺-ATPase. Kir4.1 is thought to recycle K⁺ into the interstitial space to provide sufficient K⁺ for optimum pump activity. Kcnj10 knockout mice display a dramatic reduction in NCC expression and phosphorylation, distal convoluted tubular atrophy, and urinary Mg²⁺ wasting.¹⁰³ There is also a growing body of evidence to suggest that Kir4.1/5.1 is a component of a chloride-dependent NCC regulatory pathway.¹⁰⁴

Schlingmann and colleagues performed whole-exome sequencing and whole-genome sequencing on a patient cohort with novel inherited salt-losing tubulopathy, hypomagnesemia, and dilated cardiomyopathy.¹⁰⁵ These studies uncovered heterozygous missense variants in a gene that encodes a small Rag guanosine triphosphatase (RRAGD), which results in mammalian target of rapamycin-activating mutations leading to autosomal dominant kidney hypomagnesemia-RRAGD (OMIM 608268), probably a result of downregulation of TRPM6.

In a transcriptome analysis, CNNM2 was upregulated in mouse kidney from mice on a low Mg²⁺ diet and in mouse distal convoluted tubular cells grown in low Mg²⁺ media.¹⁰⁶ When expressed in X. laevis oocytes, CNNM2 supported the transport of divalent cations, including Mg²⁺ but not Ca²⁺. In 2011, a group from The Netherlands identified loss-of-function mutations in CNNM2 in patients of two unrelated families with unexplained autosomal dominant hypomagnesemia.¹⁰⁷ The phenotype included hypomagnesemia, seizures, and intellectual disability and obesity syndrome (OMIM 616418). Phenotypic analysis of a larger cohort demonstrated hypomagnesemia in all patients, generalized seizures in 77% of patients who exhibited mild-to-moderate intellectual disability, and speech delay and severe obesity was observed in 89% of patients.¹⁰⁸ The finding that the Cnnm2^+/− mouse has similar phenotypes and impaired Mg²⁺ reabsorption in the gut and the observation that the Cnnm2^−/−, although embryonically lethal, was severely hypomagnesemic at birth confirm that CNNM2 is a major player in basolateral Mg²⁺ extrusion mechanism.²⁷ Recently it has been shown that the ADP-ribosylation factor-like protein 15 (ARL 15) is a novel negative regulator of Mg²⁺ transport by catalyzing N-glycosylation of CNNMs.¹⁰⁹

Goytain and Quamme¹⁴ showed that the SLC41A1 transcript was upregulated in response to low Mg²⁺. Using cDNA for mouse SLC41A1, they prepared cRNA and injected into X. laevis oocytes, demonstrating that heterologous expression of mouse SLC41A1 in X. laevis oocytes results in Mg²⁺ translocation. An exon-skipping mutation in SLC41A1 resulted in a nephronophthisis-like phenotype in a patient who eventually required renal transplantation.¹¹⁰ Abnormal serum or urinary electrolytes were, however, not detected. In differential transcriptomic studies in mouse distal convoluted tubule, another group¹¹¹ identified SLC41A3 as a potential player in Mg²⁺ homeostasis. The experimental evidence thus supported a role for SLC41A3 as a potential basolateral efflux mechanism for distal convoluted tubular Mg²⁺ transport. de Baaij and colleagues have characterized the functional role of SLC41A3 in the mouse by studying the Slc41a3^−/− mouse.¹⁶ They determined by quantitative PCR analysis that while Slc41a1 and Slc41a2 are expressed, Slc41a3 is the only SLC41 isoform with enhanced expression in the mouse distal convoluted tubule. Importantly the Slc41a3^−/− mice suffer from hypomagnesemia. Of some interest in their study, in contrast to Slc41a3, Mg²⁺-sensitive regulation of Slc41a1 expression was not observed in the distal convoluted tubule transcriptomic study.¹¹¹ These studies suggest that SLC41A3 is a major factor in mammalian renal magnesium homeostasis. Main features of the distal convoluted tubule exit step for Mg²⁺ are illustrated in Figure 5.

Figure 5.

Potential players involved in the distal convoluted tubule exit step for Mg²⁺. Loss of functions in the gene for the γ-subunit of Na⁺-K⁺-ATPase (FXYD2) is associated with autosomal dominant hypomagnesemia. HNF1B, a transcription factor, positively stimulates FXYD2 expression in concert with PCBD1, and mutations in the genes for HNF1B and PCBD1 are associated with hypomagnesemia. It is possible that autosomal dominant mutations in FXYD2 and HNF1B affect Na⁺-K⁺-ATPase function, causing dysregulation of ionic concentrations favoring the exit of Mg²⁺ through the basolateral membrane. ARL 15, the negative regulator of CNNM2, is shown. Loss-of-function mutations in Kir4.1/5.1 lead to downregulation of NCC and loss of TRPM6 and tubular atrophy. The nature of the exit pore has not been determined with certainty, but several potential candidates have emerged from identified human gene mutations and evidence from knockout mouse models. These include SLC41A3 and CNNM2.

In the past two decades, astute phenotyping and advances in whole-exome and whole-genome sequencing, coupled with informatics and in concert with knockout and transgenic mouse models, have resulted in the identification of rare hereditary disorders of renal magnesium handling and salt-losing tubulopathies. Scientists worldwide continue to identify key transporters and their regulatory binding partners, which contribute to a richer understanding of the renal biochemistry and physiology. However, there are still mountains to climb because it has been estimated that approximately 20% of patients with tubulopathies lack a genetic diagnosis.

Disclosures

A.R. Morrison reports consultancy agreements with Robert Wood Johnson Foundation–Harold Amos minority Faculty Development Program; honoraria from Robert Wood Johnson Foundation; and an advisory or leadership role for National Advisory Committee (NAC) of the Harold Amos Program.

Funding

None.

Author Contributions

A.R. Morrison conceptualized the study, wrote the original draft, and reviewed and edited the manuscript.