Diabetologia (1990) 33:511-514 Diabetologia Springer-Verlag 1990 Review Magnesium and glucose homeostasis G. Paolisso 1, A. Scheen 3, E D'Onofrio 2 and R Lef6bvre 3 Istituto di Gerontologia e Geriatria, 2 Istituto di Medicina Generale, Terapia Medica e Malattie del Metabolismo, University of Naples, Naples, Italy and 3 Division of Diabetes, Nutrition and Metabolic Disorders, Department of Medicine, University of Libge, Li6ge, Belgium Summary. Magnesium is an important ion in all living cells losses and insulin resistance. The extent to which such a low being a cofactor of many enzymes, especially those utilising intracellular magnesium content contributes to the develop- high energy phosphate bounds. The relationship between in- ment of macro- and microangiopathy remains to be estab- sulin and magnesium has been recently studied. In particular lished. A reduced intracellular magnesium content might it has been shown that magnesium plays the role of a second contribute to the impaired insulin response and action which messenger for insulin action; on the other hand, insulin itself occurs in Type 2 (non-insulin-dependent) diabetes mellitus. has been demonstrated to be an important regulatory factor Chronic magnesium supplementation can contribute to an of intracellular magnesium accumulation. Conditions associ- improvement in both islet Beta-cell response and insulin ac- ated with insulin resistance, such as hypertension or aging, tion in non-insulin-dependent diabetic subjects. are also associated with low intracellular magnesium con- tents. In diabetes mellitus, it is suggested that low intracellu- Key words: Magnesium, insulin, glucose homeostasis, lar magnesium levels result from both increased urinary diabetic complications, dietary magnesium supplements. Magnesium homeostasis In normal man, daily magnesium intake should be be- tween 240 and 480 mg to maintain an adequate magne- Magnesium is one of the most abundant ions present in liv- sium balance. No single factor appears to play a leading ing organisms. It is distributed in three major compart- role in the regulation of magnesium metabolism as does, ments of the body: about 65 % in the mineral phase of skele - for instance, vitamin D for calcium homeostasis [2]. Data ton, some 34% in the intracellular space and only 1% in the collected from measurements performed in a large series extracellular fluid [1]. The levels of magnesium in the plas- of animal species have shown that the small intestine is the ma of healthy people are remarkably constant, being on main site of magnesium absorption, but that the pattern of average 0.85 mmol/1 and varying less than 15% from this absorption varies with the species studied [5]. Most likely value [2]. The distribution of normal values for serum or there is a common mechanism for the transport of calcium plasma magnesiumis similarin men and women and almost and magnesium across the small intestinal wall [5], a the- one third is bound to plasma proteins. The remaining two- ory, however, that has been challenged [6]. Magnesium ex- thirds, which is diffusible or ionized, appears to be the bio- cretion is performed through renal pathways since an logically active component [2]. As reviewed by Flatman [3] amount equivalent to one third of the daily magnesium in- many cells keep their magnesium content well below elec- take is excreted through the kidney [2]. troch emical equilibrium, indicating that they possess an ac- Magnesium balance appears to be regulated by differ- tive magnesium transport system. This is also true for mito- ent hormones known to affect magnesium transport. chondria. The source of energy for magnesium transport Among them, calcitonin [7] and parathormone [8] have may be the coupling of magnesium exit to the obligatory long been thought to play a major role. Noradrenaline entry of either sodium (as in the nerve or muscle cells), and adrenaline appear to have different effects depend- protons (as in mitochondria) or potassium (a s in synapto- ing upon the tissue considered, since they stimulate mag- somes or pancreatic Beta-cells), which travel down their nesium uptake by fat cells while they reduce magnesium electrochemical gradients. In fact there is some evidence uptake by cardiac muscle cells [2]. Insulin has also been that a separate magnesium extrusion pump, driven by me- suggested as a regulatory hormone of the magnesium tabolic energy directly, does exist [4]. balance. In fact, Lostroh and Krahl [9, 10] were the first 512 G. Paolisso et al.: Magnesium and glucose homeostasis to demonstrate that insulin added in vitro promptly pro- Magnesium and insulin action motes a net increase in the accumulation of magnesium and potassium in uterine smooth muscle cells. These au- Numerous in vitro studies have pointed out the major role thors [9, 10] suggested that insulin, after interacting with of magnesium in insulin action [9, 10, 19, 20]. Lostroh and its own receptor on the plasma membrane, can affect an Krahl [9, 10] suggested magnesium as a second messenger ATPase pump increasing magnesium and potassium for insulin action. In fact, cellular magnesium deficiency is cellular entry. Recently reported data support an effect of correlated to an impaired function of many enzymes uti- insulin on magnesium transport. Indeed, during the lising high energy phosphate bonds, which are involved in course of an oral glucose tolerance test, a significant glucose metabolism, and require magnesium as a cofactor. decline in plasma magnesium with a contemporary signi- Furthermore, Tonyai et al. [20] have demonstrated that a ficant increase in erythrocyte magnesium levels does low erythrocyte magnesium content per se can increase occur [11]. Such opposite changes in plasma and erythro- membrane microviscosity and have suggested that this cyte magnesium levels are also seen during the course of mechanism may impair the interaction of insulin with its a euglycaemic hyperinsulinaemic glucose clamp [11]. Fi- receptor on the plasma membrane. nally, in vitro investigations have shown that erythrocytes In vivo, Moles and McMullen [21] have suggested that accumulate magnesium in the presence of glucose hypomagnesaemia may contribute to the insulin resis- (5 mmol/1) and insulin (100 mU/1), an effect entirely tance observed during the treatment of diabetic ketoaci- abolished by ouabain, while glucose alone had no signifi- dosis, while Durlach and Rayssiguer [22] reported that cant effect [11]. These in vivo and in vitro results thus chronic magnesium deficiency Contributes to reduce in- suggest that insulin is an important modulator of intracel- sulin sensitivity. lular magnesium content; furthermore, there are indi- Recent studies have also shown that aging and essen- cations that, as in other energy producing system, an tial hypertension, two classic conditions associated with ATPase-dependent pump is involved in the mechanisms insulin resistance [23-25], are also associated with an im- by which insulin regulates the erythrocyte magnesium paired insulin-mediated accumulation of magnesium into content [11]. erythrocytes. In essential hypertension [26], a significant reduction in plasma and erythrocyte magnesium levels and a reduced erythrocyte magnesium uptake response to incubation in the presence of insulin and high extracellu- Magnesium deficiency in man lar magnesium levels have been reported; these abnor- malities were associated with an increase in erythrocyte The existence of a state of magnesium deficiency has been membrane microviscosity. It was suggested that changes doubted for many years. However, severe hypomagne- in the physical state of the plasma membrane as well as in saemia is a well recognized clinical syndrome charac- insulin sensitivity were co-responsible for the lower ery- terized by: a)muscular symptoms (spasmophilia, gross throcyte magnesium level found in these patients. In aging muscular tremour, ataxia, tetany); b)psychic disorders [27], the reduced erythrocyte magnesium content was ex- (agitation, confusion and hallucinations); and, c) cardio- plained on the grounds of the well known insulin-resistant logical signs (low-voltage T-wave at the ECG). Labora- state frequently observed in this condition [23, 24]. The tory data (low serum magnesium levels associated with a extent to which changes in plasma membrane liquid com- normal serum calcium concentration and a normal blood position, which frequently occur and impair the interac- pH) confirm the diagnosis. As recently reviewed by Rein- tion of insulin with its receptor in the elderly [28], also con- hart [12], measurement of magnesium levels in the plasma tribute remains an open question. or serum is the usual method for determining magnesium homeostasis. However, it is well known that there may be a dissociation between serum and intracellular levels of Magnesium and diabetes mellitus magnesium [13, 14] and that intracellular levels may in- deed better reflect homeostasis. In this respect, erythro- In 1971, Londono and Rosenbloom [29] were the first to cytes or lymphocytes are frequently used while muscle demonstrate, in diabetic children, that a glucagon injec- biopsies or magnesium balance studies, although more tion induced a significant decline in plasma magnesium sensitive indices of magnesium deficiency, are rarely per- and calcium levels. Subsequently, Rosenbloom [30] ob- formed [12]. served that the decline in plasma magnesium and calcium Among all clinical conditions associated with a deple- levels observed during the course of an oral glucose toler-