Regulation of magnesium balance

INTRODUCTION — Magnesium balance, like that of other ions, is a function of intake and excretion. The average daily magnesium intake is 360 mg (15 mmol). Approximately one-third of this magnesium is absorbed, principally in the small bowel through both a saturable transport system (presumably mediated by a channel encoded by the TRPM6 gene) and passive diffusion. (See "Hypomagnesemia: Causes of hypomagnesemia".)

Two other processes occur in the gut: the secretion of approximately 40 mg (1.7 mmol) in intestinal secretions, and the absorption of another 20 mg (0.8 mmol) in the large bowel. In the healthy adult, there is no net gain or loss of magnesium from bone so that balance is achieved by the urinary excretion of the approximately 100 mg (4.1 mmol) that is absorbed. Changes in intake are balanced by changes in urinary magnesium reabsorption, principally in the loop of Henle and the distal tubule.

UNITS OF MEASUREMENT — Before discussing the factors responsible for the regulation of magnesium balance, it is useful to review the different units that may be used to measure the plasma (or serum) magnesium concentration. Laboratories in the United States usually report the results in units of mEq/L or mg/dL, while other countries primarily use mmol/L. The relationship among these units can be expressed by the following equations:

Since the molecular weight of magnesium is 24.3 and the valence is +2, 1 mEq/L is equivalent to 0.50 mmol/L and to 1.2 mg/dL. Thus, a normal range of the plasma magnesium concentration of 1.4 to 1.7 mEq/L is equivalent to 0.70 to 0.85 mmol/L and 1.7 to 2.1 mg/dL.

RENAL HANDLING OF MAGNESIUM — Magnesium transport differs from that of most other ions in that the proximal tubule is not the major site of reabsorption. Approximately 80 percent of the total plasma magnesium is filtered at the glomerulus. Only 15 to 25 percent of the ultrafiltrable magnesium is reabsorbed passively in the proximal tubule (down concentration gradients generated by the reabsorption of sodium and water) and 5 to 10 percent in the distal tubule. Micropuncture studies (in which small pipettes are inserted into different nephron segments) indicate that the major site of magnesium transport is the thick ascending limb of the loop of Henle, where 60 to 70 percent of the ultrafiltrable magnesium is reabsorbed [1,2].

Proximal tubule — There is no active transport of magnesium in the proximal tubule, so reabsorption in this segment is thought to be passive and likely paracellular, driven by the concentration gradient generated in the late proximal tubule by luminal water reabsorption. Interest in this segment has been stimulated by the observation that inhibitors of the sodium-glucose cotransporter 2 (SGLT2) increase renal magnesium reabsorption. This phenomenon, however, is more likely due to increased sodium delivery to the loop of Henle. Interestingly, SGLT2 knockout mice have increased rather than decreased urine magnesium excretion [3].

Loop of Henle — The bulk of magnesium transport in the thick ascending limb, like that of calcium, is passive, occurring by paracellular diffusion between the cells and driven by the favorable electrical (lumen-positive) gradient resulting from the reabsorption of sodium chloride (figure 1) [1,4-6]. This lumen-positive voltage is generated by two distinct processes:

●Reabsorption of sodium, potassium, and chloride via the Na-K-2Cl cotransporter, with apical recycling of potassium via ROMK channels back into the lumen, is thought to be the major contributor to the electrical gradient (figure 2).

●Net reabsorption of sodium chloride dilutes the tubular fluid and generates a transepithelial concentration gradient toward the end of the thick ascending limb that favors backflux of sodium chloride from the peritubular to the tubular space. It has been hypothesized that the presence of a paracellular pathway in the cortical thick ascending limb that is selective for sodium over chloride ions generates an electrical diffusion potential that augments the lumen-positive voltage, favoring additional reabsorption of magnesium [7].

Paracellular magnesium reabsorption appears to be facilitated by a complex of two tight junction proteins, specifically, claudin-16 (also known as paracellin-1) [8] and claudin-19 [9], that interact to form paracellular divalent cation pores and directly mediate magnesium reabsorption [10,11].

It has been suggested that they may also act as a paracellular pore for Na in the cortical thick ascending limb, thereby facilitating the generation of the NaCl diffusion potential that is hypothesized to drive additional Mg reabsorption [12,13].

Conversely, claudin-10b is a claudin that mediates paracellular Na reabsorption in the thick ascending limb but seems to inhibit magnesium reabsorption. Claudin-10b is predominantly expressed in the medullary thick ascending limb in a population of cells that is distinct from those expressing claudin-16 and -19 [14]. Passive reabsorption of Na would be expected to dissipate the transepithelial voltage generated by transcellular Na, K, and Cl transport with apical K recycling and thereby reduce paracellular magnesium reabsorption.

Factors controlling magnesium transport act through changes in the voltage and/or permeability of the paracellular pathway. Thus, decreased reabsorption and magnesium wasting can be induced by the administration of a loop diuretic, which inhibits sodium and chloride reabsorption, or by mutations in claudin-16 or -19, which cause familial hypomagnesemia with hypercalciuria and nephrocalcinosis [8,9]. Conversely, inactivating mutations in claudin-10b cause a syndrome of renal sodium wasting that is associated with hypermagnesemia and urinary magnesium retention [15,16]. (See "Effect of diuretics on magnesium handling by the kidney" and "Hypomagnesemia: Causes of hypomagnesemia", section on 'Familial hypomagnesemia with hypercalciuria and nephrocalcinosis (FHHNC)' and "Hypomagnesemia: Causes of hypomagnesemia".)

Several other factors alter magnesium transport in the loop of Henle:

●Plasma magnesium concentration – Loop magnesium reabsorption (particularly in the cortical aspect of the thick ascending limb) varies with changes in the plasma magnesium concentration, which is the main physiologic regulator of urinary magnesium excretion [1,2]. Hypermagnesemia inhibits loop magnesium (and calcium) transport, while hypomagnesemia stimulates magnesium transport in both normal and magnesium-depleted animals, thereby decreasing further magnesium losses.

●Plasma calcium concentration – Hypercalcemia also inhibits magnesium (and calcium) reabsorption, leading to hypermagnesuria and hypercalciuria. This seems to occur by activation of the calcium-sensing receptor on the basolateral membrane of the thick ascending limb. Downstream of the calcium-sensing receptor is a complex signaling sequence whereby calcineurin and NFATc1 stimulate the transcription of two microRNAs, miR-9-1 and miR-374, that normally inhibit the expression of the tight junction protein, claudin-14 [17,18]. Claudin-14, once expressed, physically binds to and inhibits the claudin-16 and -19 complex, thereby inhibiting Mg reabsorption. Further discussion of the role of this receptor in the modulation of calcium and sodium transport is discussed separately. (See "Disorders of the calcium-sensing receptor: Familial hypocalciuric hypercalcemia and autosomal dominant hypocalcemia".)

●Hormones – A number of hormones have been shown to alter magnesium transport in the thick ascending limb in experimental studies. These include parathyroid hormone (PTH), calcitonin, glucagon, arginine vasopressin (AVP), and the beta-adrenergic agonists, all of which are coupled to adenylate cyclase in the thick ascending limb. Postulated mechanisms include an increase in luminal-positive voltage (via activation of basolateral membrane chloride conductance and apical Na-K-2Cl cotransport) and an increase in paracellular permeability (possibly by phosphorylation of paracellular pathway proteins) [19]. It is not known if these effects have any important role in normal magnesium homeostasis.

●Ras-related GTP binding D (RRAGD) – RagD is a subunit of the mTORC1 complex, and mutations in its gene RRAGD that hyperactivate mTORC1 cause autosomal dominant hypomagnesemia in association with nephrocalcinosis, suggesting that mTORC1 may regulate paracellular magnesium and calcium reabsorption in the thick ascending limb [20].

●Other factors – Metabolic alkalosis stimulates and metabolic acidosis, hypokalemia, and phosphate depletion inhibit magnesium reabsorption in the loop of Henle, contributing to changes in urinary magnesium excretion [1]. The mechanisms for these changes are unknown.

Distal reabsorption — Distal reabsorption of magnesium occurs in the distal convoluted tubule (DCT) via transcellular transport (figure 3). It is important for the fine tuning of renal magnesium excretion and hence is potentially a key step for physiological regulation.

Monogenic diseases that produce hypomagnesemia have helped elucidate important components of distal magnesium transport (see "Hypomagnesemia: Causes of hypomagnesemia"):

●Magnesium enters the DCT cells from the tubular lumen via a magnesium channel complex composed of TRPM6 and TRPM7; this occurs passively, driven by a favorable electrical gradient produced by potassium flux from the cell into the tubular lumen [21,22]. Passive influx of magnesium is believed to be the rate-limiting step in distal magnesium reabsorption.

●Mutations in TRPM6 cause autosomal recessive hypomagnesemia with secondary hypocalcemia, which is due both to impaired intestinal absorption of magnesium and a renal magnesium leak [23,24]. Several factors regulate TRPM6, including dietary magnesium [25], estrogens [25], and acid-base balance [26]. In addition, epidermal growth factor (EGF) regulates TRPM6 by stimulating protein trafficking to the plasma membrane [27]. This is why mutations in pro-EGF cause isolated renal hypomagnesemia [28], as does treatment with EGF inhibitors such as cetuximab. Insulin activates TRPM6 via binding of the insulin receptor substrate (IRS4) and phosphatidylinositol 3-kinase signaling, and IRS4 knockout mice show hypomagnesemia and urinary magnesium wasting [29,30]. The transcription factor prospero homeobox 1 (Prox-1), which is highly enriched in the DCT, also appears to regulate TRPM6. Knockout of Prox-1 in mice leads to hypomagnesemia accompanied by downregulation of TRPM6 expression, as well as that of the thiazide-sensitive cotransporter (NCC) [31].

●Mutations in TRPM7 have been described in patients with hypomagnesemia associated with seizures [32] or hemiplegic migraine [33].

●As magnesium entry is driven by the electrical gradient across the apical membrane, it is also regulated by the setpoint of the membrane potential. Mutations in the apical voltage-gated potassium channel, Kv1.1, cause autosomal dominant hypomagnesemia, presumably by altering the apical membrane potential [34].

●The mechanism for basolateral exit of magnesium in the DCT is incompletely understood. CNNM2 encodes a transmembrane protein that is localized to the basolateral membrane of the thick ascending limb and DCT and has been proposed to function either as a magnesium channel [35] or magnesium-sensitive sodium channel [36]. CNNM2 is also upregulated under conditions of magnesium deficiency [35] and is mutated in dominant isolated renal magnesium wasting [36]. Thus, CNNM2 could be a candidate for the basolateral exit pathway, although the driving force for magnesium exit is unclear.

●Distal magnesium reabsorption is in some way dependent upon the Na-K-ATPase. Mutations in the Na-K-ATPase gamma subunit, FXYD2 [37], or in hepatocyte nuclear factor 1B (which regulates transcription of FXYD2) [38], cause dominant isolated renal hypomagnesemia and hypomagnesemia with kidney malformations. In addition, infants with heterozygous mutations in the Na-K-ATPase alpha-1 subunit have been reported to have severe hypomagnesemia with renal magnesium wasting [39]. Exactly how the Na-K-ATPase regulates magnesium transport is unclear; it may determine the basolateral membrane sodium gradient or ensure hyperpolarization of the membrane potential and hence drive apical magnesium entry.

●At the basolateral membrane, a potassium channel constituted from a heteromultimer of Kir4.1 (encoded by KCNJ10) and Kir5.1 (encoded by KCNJ16) are hypothesized to recycle potassium across the basolateral membrane, thus facilitating Na-K-ATPase activity. Thus, mutations in KCNJ10 cause renal magnesium wasting as part of a syndrome called SeSAME or EAST, together with seizures, sensorineural deafness, ataxia, developmental delay, and a Gitelman-like tubulopathy [40,41]. Mutations in KCNJ16 cause a similar tubulopathy [42].

In addition, population-based, genome-wide association studies have identified a number of genetic variants that are associated with serum magnesium concentration. These include variants in MUC1, TRPM6, DCDC5, ATP2B1, and PRMT7 [43]. With the exception of TRPM6, the possible mechanisms for these associations are unknown. Urine magnesium excretion is also associated with genetic variants in TRPM6 and another gene, ARL15, which encodes a GTP-binding protein [44].

Distally acting diuretics can also affect magnesium handling. Chronic use of thiazide diuretics is associated with hypomagnesemia and renal magnesium wasting; in addition, Gitelman syndrome, which is due to an abnormality in the gene coding for NCC, is characterized by hypomagnesemia due to urinary magnesium wasting [45]. (See "Inherited hypokalemic salt-losing tubulopathies: Pathophysiology and overview of clinical manifestations" and "Hypomagnesemia: Causes of hypomagnesemia".)

Both thiazide diuretics and murine knockout of NCC downregulate TRPM6 expression; this likely explains the mechanism of the magnesuria caused by thiazides and Gitelman syndrome [46]. In addition, the potassium depletion associated with these conditions may contribute to the hypomagnesemia [1,47]. (See "Effect of diuretics on magnesium handling by the kidney".)

In contrast to thiazides, amiloride is a magnesium-sparing diuretic, possibly acting by hyperpolarizing of the membrane potential, thereby increasing the driving force for magnesium entry [48].

REGULATION OF PLASMA MAGNESIUM CONCENTRATION — In contrast to other ions, magnesium is treated by the body as an orphan: there are no hormones that have a substantial role in regulating urinary magnesium excretion, and bone, the principal reservoir of magnesium, does not readily exchange with circulating magnesium. This is in marked contrast to the roles of aldosterone and atrial natriuretic peptide in sodium excretion; antidiuretic hormone in water excretion; aldosterone and cell potassium stores in the regulation of potassium balance; and parathyroid hormone (PTH), vitamin D, calcitonin, and bone stores in the maintenance of calcium balance.

The inability to readily mobilize magnesium stores means that, with negative magnesium balance, the initial losses come primarily from the extracellular fluid; equilibration with bone stores does not begin for several weeks [49]. Thus, the plasma magnesium concentration falls rapidly with negative magnesium balance, leading to a marked reduction in magnesium excretion, unless urinary magnesium wasting is present (see "Hypomagnesemia: Causes of hypomagnesemia"). It is the enhanced reabsorption in the loop of Henle and distal tubule and fall in urinary magnesium excretion that protects the plasma magnesium concentration against further losses. The fractional excretion of magnesium, which is between 3 and 5 percent in subjects with normal kidney function ingesting a typical diet, can fall to below 0.5 percent with magnesium depletion due to extrarenal losses [2]. (See "Hypomagnesemia: Evaluation and treatment".)

Conversely, there is no protection against hypermagnesemia with loss of kidney function. In this setting, continued intake leads to magnesium retention that is predominantly in the extracellular fluid. (See "Hypermagnesemia: Causes, symptoms, and treatment".)

CELLULAR MAGNESIUM STORES AND CYTOSOLIC MAGNESIUM ACTIVITY — The total intracellular magnesium content approximates 8 to 10 mmol/L (10 to 20 mEq/L). However, most of the cell magnesium is bound to ATP and other intracellular nucleotides and enzyme complexes. Some studies, using magnesium-sensitive dyes based upon the FURA compound, indicate that the free cytosolic concentration is in the range of 0.6 to 0.8 mmol/L (1.2 to 1.6 mEq/L) [50]. Changes in the plasma magnesium levels alter cellular magnesium content slowly. It is not yet known how magnesium enters cells, but energy is not required because of the favorable electrochemical gradient. As the intracellular free magnesium is significantly below electrochemical equilibrium, however, there must be some type of energy-dependent extrusion mechanism [51]. An ATP-dependent sodium-magnesium exchanger has been described in several tissues, but its physiologic importance is not known.

SUMMARY

●Magnesium balance – Magnesium balance is a function of intake and excretion. The average daily magnesium intake is 360 mg (15 mmol). Changes in intake are balanced by changes in renal magnesium reabsorption. (See 'Introduction' above.)

●Renal handling of magnesium

•Eighty percent of the total plasma magnesium is filtered at the glomerulus. Fifteen to 25 percent is reabsorbed in the proximal tubule; 60 to 70 percent is reabsorbed in the thick ascending limb of the loop of Henle; and 5 to 10 percent in the distal tubule. (See 'Renal handling of magnesium' above.)

•The majority of magnesium transport in the thick ascending limb occurs by paracellular diffusion between the cells and is driven by the favorable electrical gradient resulting from the reabsorption of sodium chloride. Diffusion is facilitated by tight junction proteins called claudin-16 and claudin-19. (See 'Loop of Henle' above.)

•The plasma magnesium concentration is the major physiologic regulator of urinary magnesium excretion. Hypermagnesemia inhibits loop magnesium transport (thus increasing urinary excretion of magnesium), while hypomagnesemia stimulates magnesium transport (inhibiting its excretion). Other factors that decrease magnesium transport in the loop of Henle include hypercalcemia, metabolic acidosis, hypokalemia, phosphate depletion, the administration of loop diuretics, and mutations in paracellin-1 and related proteins. (See 'Loop of Henle' above.)

•Factors that alter distal tubule magnesium transport include diuretics and genetic mutations. Whereas amiloride, a magnesium-sparing diuretic, stimulates magnesium transport, thiazide diuretics are occasionally associated with magnesium wasting and hypomagnesemia. (See 'Distal reabsorption' above.)

●Regulation of plasma magnesium concentration

•Bone magnesium is the principal reservoir of magnesium. Because bone magnesium does not readily exchange with circulating magnesium, the plasma magnesium concentration falls rapidly with negative magnesium balance. This leads to a marked reduction in magnesium excretion, provided urinary magnesium wasting is not present. (See 'Regulation of plasma magnesium concentration' above.)

•There is no protection against hypermagnesemia with loss of kidney function. In this setting, continued intake leads to magnesium retention that is predominantly in the extracellular fluid. (See 'Regulation of plasma magnesium concentration' above.)

ACKNOWLEDGMENT — The UpToDate editorial staff acknowledges Zalman S Agus, MD, who contributed to an earlier version of this topic review.