French edition (Le magnesium en pratique cliniques. Published
in 1985 by J. B. Baillière, Editions Médicales
Internationales, 62 rue des Mathurins, 75008 Paris, France
First English language edition published in 1988 by
John Libbey & Company Ltd, 80/84 Bondway, London SW8 1SF, England (01) 582 5266
John Libbey Eurotest Ltd, 6 rue Blanche, 92120 Montrouge, France (1) 47 35 85 52
Excerpt. Pages 1-39.
Basically, the role of magnesium in clinical practice may only be considered in two conditions occurring with very different frequency. On the one hand there is the major and somewhat commonplace problem of magnesium deficits and on the other, the much rarer condition of magnesium overload which is almost always of iatrogenic origin.
The beginning of modern studies of the consequences of magnesium deficit in pathology is rightly attributed only to the demonstration of its physiologic importance in animals with the seminal experiment of Jehan Leroy which proved in 1926 the essential character of the ion for mice (750). One can cite as a predecessor only J. Gaube du Gers (492, 493, 494) who, more than a century ago, noticed in the same animal that a diet deficient in magnesium, composed of bread devoid of magnesium and distilled water, but also no doubt deficient in other nutrients, caused progressive sterility. Already magnesium appeared to him to be 'the metal of vital activity for what is most precious and noble in life: reproduction and sensation.' Subsequently it was the remarkable studies of the groups of E. MacCollum and D.M. Greenberg (281, 345) that in the Thirties revealed many of the physiological properties of magnesium, examining, in the rat initially the multiple effects of a lack of magnesium in the diet on development, reproduction, neuromuscular apparatus, and humoral balance. They then studied the specific reversibility of such defects by oral loading of magnesium. These studies provide as well the first experimental basis in animals for the diagnostic test for magnesium deficiency by oral loading with physiological doses of magnesium. Contrary to intakes of magnesium which have non-specific pharmacologic effects, oral physiological doses of magnesium are devoid of any activity. Their effectiveness in a clinical context constitutes the best proof of the existence of a developing magnesium deficiency (332).
On the other hand, one must refrain from attributing any compelling physiologic significance to earlier works that, extrapolating from magnesium's various pharmacodynamic properties (neurosedative, cardio-regulatory, vasodilating, antihemolytic, antitoxic...), whether studied in vitro or by massive parenteral doses, claimed to have discovered symmetrical harmful effects in the case of magnesium deficit. Nothing could be less conclusive: one can control a motor-stimulatory crisis or a vascular spasm by the pharmacodynamic effects of a parenteral injection of magnesium without in doing so providing the least proof of a deficit of this ion as cause (332). It is pleasant however to be able to note the predictive character of hypotheses stated more than a century ago, and subsequently verified, concerning the etiologic or physiopathologic role of magnesium in certain tetanies, epilepsies, cardiovascular complaints and hemolytic or toxic problems.
We have cited here, however, only carefully chosen examples where modern studies have proved past extrapolations to be well founded, and not any others. The latter have, on the contrary, been the source of a number of unwarranted pathologic attributions of paternity to magnesium deficit. For this reason few ions have generated as much enthusiasm and as much disdain.
Heated zealots like P. Delbet have seen in magnesium a sort of panacea, the lack of which plays a major role in the development of cancer, the spread of epidemics and even in the frequency of suicides (340).
For the skeptics, on the other hand, magnesium has appeared as a trace element of unclear biologic importance and physiologic significance, existing in the diet at levels sufficient to supply any needs, but remaining almost impossible to measure.
Between these two extremes it is today possible to find a balance.
Magnesium, the second most abundant intracellular cation, is a catalytic and structural element of major significance in the physiology of the human organism. Necessary for the anatomical and functional integrity of various subcellular organelles, it participates in all the major metabolic pathways, i.e., those involving carbohydrates, proteins and lipids, as well as in redox reactions. It is involved in the regulation of ion levels, maintaining the potassium level in the cell and exercising on the metabolism of calcium and phosphorus vitamin D-like effects.
Integral to processes of defense, magnesium exhibits a variety of effects: antistress, anti-allergic, anti-anaphylactic, anti-inflammatory, antiradiation. Magnesium plays a role in thermoregulation; it stimulates phagocytosis and the formation of antibodies, complement and elements of the properdin system.
Present in many tissues, it is active in the physiology of many systems, not only neuromuscular, osteo-articulatory, and dental, but also respiratory, endocrine, reproductive, ocular, digestive, hepatic, pancreatic, renal, cardiovascular and hematologic.
Dietary magnesium in many regions and particularly in France appears to be insufficient to satisfy any daily need that is markedly high with respect to body stores. Thanks in particular to atomic absorption spectrophotometry, magnesium can now be measured analytically with ease and accuracy. If its essentially intracellular distribution still makes it difficult to evaluate, the "deterring factors" which have previously prevented the systematic study of its pathology have been permanently removed. It has become quite easy to identify the various clinical manifestations of magnesium deficits. But one must not approach both primary and secondary deficits in the same fashion. A primary magnesium deficit only allows one to observe a more or less specific, strictly magnesium-dependent symptomatology while the signs of a secondary magnesium deficit must be distinguished from among the consequences of the causative illness. Moreover, a primary magnesium deficit clearly constitutes the only indication where corrective magnesium therapy represents the principal and specific form of therapeutic intervention. Such a possibility occurs with great frequency since it corresponds to the majority of cases previously described without their magnesium dependency having been seen at the time, for example normocalcemic latent tetany, hyperventilation syndrome, or idiopathic Barlow's disease (332, 345, 363, 371, 372, 375). A secondary magnesium deficit on the contrary presents as but one of the elements of a complaint. One must determine the etiologic factors that cause such a secondary deficit and then evaluate its relative importance according to the physiopathology of the illness that causes it. The correcting of a secondary magnesium deficit must always depend initially on treatment of the causative illness. It must invariably take into account the whole picture, ionic deficit included. Specific therapeutic measures that seek to correct a secondary magnesium deficit are therefore justified only if etiologic treatment of the responsible illness is either impossible or ineffective. Such measures then are no more than adjunct treatments, justified only by the ultimate importance of secondary magnesium deficits in the physiopathology of the original illness. They are valid only in as much as it is possible to correct the problems causing the deficit.
We intend successively to:
• At first, as a necessary preliminary to the study of the role of magnesium in pathology, recall some of the basic elements of the biology of magnesium:
— its properties
— its metabolism
— methods of analysis
• Then devote the largest part of this clinical study to the question of primary magnesium deficits and especially to their best known forms:
— Neuromuscular hyperexcitability: the typical example will be latent tetany in the adult. The practical approach to this strictly magnesium pathology will be described: clinical and paraclinical symptomatology, pathogenic etiology, physiopathology, development, clinical variability and diagnostic problems.
— Other primary magnesium deficits: endocrine-humoral, osteo-articulatory, allergic, cardiovascular.
• Secondary magnesium deficits will be categorized as deficits secondary to a "spontaneous" pathology or as deficits of "iatrogenic" origin, classifying them in both cases according to the principal mechanisms that cause them.
— Problems of intake or absorption
— Neuroendocrine and metabolic dysfunctions
— Renal hyperexcretion
• We will examine magnesium excess more briefly, taking as the typical example massive parenteral overload.
• Finally, we will consider the problems of therapy, stressing in particular:
— The treatment of magnesium deficits, contrasting treatment of simple "deficiency" forms, where it is sufficient to increase intake, with the treatment of severe "depletion" forms where only an ever increasing knowledge of the control mechanisms of magnesium metabolism will allow correction of the problem.
— Treatment of magnesium overload.
1. The properties of magnesium
Magnesium has an atomic number of 12 and an electronic structure with complete K and L shells that have 2 and 8 electrons respectively, while the outer M shell has only two electrons (Fig. 1) . This outer shell determines the reactive capacity of magnesium.
Its atomic weight is 24.3. It has an atomic radius of 0.66 Å in non-aqueous medium and 5.9 Å in aqueous medium. The three naturally occurring stable isotopes of magnesium are 24Mg (79%), 25Mg (10%) and 26Mg (11%). The two principal radioisotopes are 27Mg and 28Mg with half lives respectively of 9.5 minutes and 21.3 hours.
Magnesium belongs to group II of the third period of the Periodic Table of elements, a group that includes only two other physiologically important elements: calcium and zinc.
We will outline briefly the main biologic properties of magnesium, going from the biochemical level to the cellular and finally to the level of physiologic function.
1.1 The biochemical functions of magnesium
The biochemical properties of magnesium can be deduced from three principal modes of action (12, 13, 14, 191, 328, 337, 361, 364, 398, 409, 546, 690, 1147, 1327, 1337):
— the synthesis and utilization of energy-rich compounds
— the synthesis of electron and proton transporters
— the synthesis and activity of numerous enzymes
1.11 Synthesis and utilization of energy-rich compounds
The energy necessary for the functioning of the machinery of the cell must be stored in the form of potential chemical energy or energy-rich bonds.
Magnesium is necessary for the synthesis of various compounds that have energy-rich bonds of any type: the phosphoric anhydride bond that is found mainly in ATP or adenosine triphosphate, "the main fuel of life" (13), but also in GTP (guanosine triphosphate) as well as in other nucleoside triphosphates such as UTP (uridine triphosphate), CTP (cytosine triphosphate) and ITP (inosine triphosphate). It is also found in the phosphoamide bond of phosphocreatine, the phosphoenol bond of phosphoenolpyruvic acid, the mixed anhydride bond of 1,3-diphosphoglyceric acid and in the bond between an acid and a thiol group as in acyl coenzyme A or succinyl coenzyme A.
The permanent reconstitution of these compounds from the products of their own degradation is accomplished by phosphorylation coupled to redox reactions and magnesium is essential for the coupling of the phosphorylating and redox reactions (12, 13, 14). It thus allows cellular respiration to be used to store energy in the form of high energy phosphates, that is to say, in the form of phosphorylated compounds that contain energy-rich bonds.
Moreover, magnesium has also been established as necessary for the utilization of energy-rich bonds whether this occurs by transfer or by hydrolysis. Magnesium activates in effect all reactions that transfer the phosphorylated radical (1327): reactions involving phosphoryltransferases whether reversible (adenylate kinase or myokinase, creatine kinase, carbamate kinase, phosphoglycerate kinase, pyruvate kinase) or non-reversible (glucokinase, hexokinase, phosphofructokinase, adenylate cyclases that catalyze the conversion of ATP to cAMP, adenosine triphosphatases, i.e., normally membrane-bound polyphophatases that degrade ATP).
The formation and subsequent utilization of energy-rich bonds, processes that both require Mg2+, constitute the necessary basis for all synthetic reactions and for all cellular activities. This alone would be sufficient to establish the critical biologic importance of magnesium.
1.12 Synthesis of proton and electron transporters
Proton and electron transporters represent other basic compounds necessary for the redox coupled reactions. Magnesium is essential for their synthesis. It is involved in the formation of phosphopyridine nucleotides, DPN (diphosphopyridine nucleotide or NADH2 and TPN (triphosphopyridine nucleotide or NADPH2; in the formation of flavin nucleotides, FMN (flavin mononucleotide) and FAD (flavin adenine dinucleotide); and finally in that of coenzyme A.
1.13 Synthesis and activity of enzymes
Magnesium activates a very large number of enzymes (approximately 300). In some cases Mg2+ is linked to the substrate through chelation. The Mg-ATP complex represents a typical linking of compounds that becomes the true substrate of the enzyme. This kind of activation occurs for example with hexokinase and phosphoglycerate kinase. In other cases, Mg2+ is first bound to the enzyme, causing it thus to assume its active conformation which is able to act on the substrate. This is the case, for example, with enolase, pyruvate kinase and pyrophosphatase. These two mechanisms are not moreover incompatible. A single Mg2+ may initially form a Mg-ATP complex that represents the real substrate of an ATPase which will only act on this substrate when activated by a second Mg2+ (546).
All ATP-dependent enzyme reactions exhibit an absolute need for Mg2+ (12).
We will list some of the principal enzymes that are involved in the major metabolic pathways:
• Enzymes of carbohydrate metabolism (336, 396): glucokinase, hexokinase, galactokinase, phosphorylase phosphatase, phosphorylase kinase, phosphoglucomutase, 6-phosphofructokinase aldolase, triokinase, fructose-1,6-bisphosphatase, glucose-6-phosphatase, glucose-6-phosphate dehydrogenase, transketolase, phosphoglycerate kinase, phosphoryl glycerylmutase, enolase, pyruvate kinase, thiamine-pyrophosphate kinase, pyruvate decarboxylase, glycerokinase, glycerophosphatase, various pentoside kinases that activate B vitamins....
• Enzymes of nucleic acid and protein metabolism (188, 328, 684, 690, 732, 1149, 1327): RNA polymerase which allows the synthesis of RNA and especially that of messenger RNA which, associated with post-ribosomal factors of initiation and elongation and with polyamines, codes for amino acids to produce specific proteins; DNA polymerase which allows the reconstitution and recombination of DNA, ornithine carbamyl transferase, glutamine synthetase, carbamate kinase, argininosuccinate synthetase, creatine kinase, insulinase, leucine aminopeptidase which appears to be similar to hypertensinase....
• Enzymes of lipid metabolism (328, 732, 1065, 1067): acetylcoenzyme A synthetase, acylco A synthetase, beta-ketothiolase, diglyceride kinase, phosphatidate phosphatase, mevalonate kinase, phosphomevalonate kinase, lecithin-cholesterol-acyl transferase (LCAT).
• Finally, various other enzymes: alkaline phosphatase, cholinesterase, anti-hyaluronidase.
Enzyme activation is not however the only action of magnesium on these proteins.
• It can sometimes on the contrary inhibit them: this is the case for example with the ATPase activity of myosin and with most of the guanylate cyclases.
• It participates above all in their synthesis, not only as a constituent of a magnesium metalloenzyme such as alkaline phosphatase, but especially by its role in protein synthesis in general and therefore in that of enzymes in particular. The reduced levels of serum alkaline phosphatase and of erythrocyte glucose-6-phosphate dehydrogenase (G6PDH) in the case of magnesium deficiency in the calf (729, 730) or those of lymphocyte elongation factors 1 and 2 during magnesium deficit in the rat (473) are two examples.
Thus, Mg2+ which is necessary for the synthesis of energy-rich compounds, electron transporters and enzymes and controls their effects appears to be a major cellular component. One can understand that it constitutes the fundamental regulator of the cell cycle (1333). Compartmentalization of Mg2+ in the cell permits it to play a key role in the coordination of metabolic pathways where the limiting steps are represented by phosphorylation reactions (550, 550a, 1105, 1106, 1107, 1108, 1129, 1317).
1.2 The cellular properties of magnesium (12, 13, 14, 328, 337, 546, 549, 550, 550a, 617, 847, 1286, 1336, 1337)
The cytologic properties of magnesium deserve the very special attention due an ion that is almost totally intracellular. They are determined by the extreme heterogeneity of both its physicochemical and its anatomical distribution.
Even though its different concentrations clearly vary according to the cell type being studied as a function of the physiologic activities of various tissues, these concentrations present certain common characteristics.
1.21 Physicochemical heterogeneity
The cellular distribution of magnesium corresponds to a broad compartmentalization that is essentially a function of the tendency of this metal to form chelates with different cellular constituents, very often in competition with calcium. (12, 13, 14, 546, 549, 550).
• Phospholipids form complexes with magnesium as well as with calcium, which explains why magnesium is an integral part of the structure of many membranes, cellular or subcellular, plasma or mitochondrial, reticular, and sarcoplasmic.
• Nucleic acids form complexes with magnesium that have a much greater flexibility than those formed with calcium. They add to the high levels in ribosomes as well as in the nucleus.
• By virtue of the law of mass action, these different bound magnesium compounds (corresponding to about 70% of intracellular magnesium) are in equilibrium with a pool of free magnesium (i.e., about 30% of cellular magnesium).
• Heterogeneity may be observed not only among different substructures of the cell — one can, for example, contrast the richness of mitochondria and microsomes in magnesium with the more modest magnesium levels of the cytosol — but also within a given subcellular organelle. In a mitochondrion the intermembrane compartment is much richer than the inner or outer membranes. Similarly, in the cytosol, magnesium is not uniformly distributed. Its highest levels are found near negatively charged membranes or on the surface of polyanions. Buffered at a level of pMg 3 (pMg = -log [Mg2+]), intracellular Mg2+, whatever its site, is found in equilibrium with bound magnesium and in an optimal concentration for the exercise of its metabolic effects (546).
Thus, magnesium, distributed in the cell in a very heterogeneous manner, is found at high levels in all cellular and subcellular membranes, in mitochondria, microsomes, and the nucleus, always remaining in an equilibrium between bound and free forms that is optimal for the exercise of its effects, the most notable of which is its stabilizing capacity.
1.22 The cellular and subcellular stabilizing effects of magnesium
Magnesium appears to have a stabilizing effect not only for the cell membrane but also for various subcellular organelles.
• One of the major properties of magnesium is that of stabilizing membranes. This depends mainly on its structural role. The complexing of magnesium with phospholipids reduces membrane fluidity and lowers membrane permeability with parallel polarizing electrostatic effects.
Thus in magnesium deficit the permeability of plasma membranes increases. Cells become loaded with Ca2+ and Na+ and lose K+ and phosphorus. At the same time, the cell membrane depolarizes (70). The direct structural effects of magnesium on the membrane are of fundamental importance. They may be re-inforced by comparable effects on the ATP-dependent pumps that regulate the active transport of Na+, K+, and Ca2+. It is particularly necessary to stress the potentiating effect of magnesium on Na+/K+-adenosine triphosphatase (Na+/K+-ATPase) that controls the "active sodium pump." When activated, it removes sodium from the cell and thus allows potassium to remain. In this fashion, a lack of magnesium induces changes in cellular ion levels which in turn cause various ionic interactions. Increasing cytosolic Na+, for example, alters the distribution of the calcium load. The level of cytosolic Ca2+ which has already been increased by the hyperpermeability of the plasma membrane is increased even more by a flux of calcium from the mitochondrion into the cytosol that is caused by changes in the mitochondrial membrane (546). Calcium remains even more tightly bound in the cytosol because magnesium deficit increases the affinity of its binding sites for it (293).
Magnesium's stabilizing role is not limited to the plasma membrane: Mg2+ is necessary for the anatomical and functional integrity of a variety of subcellular organelles.
• Magnesium maintains a "strongly coupled" state in the mitochondrion, that is, a state where oxidative phosphorylation, the fundamental element of energy reactions, is highly efficient. The primary function of all mitochondria is to keep phosphorylation coupled to oxidation and to produce ATP, the basic fuel of life. With magnesium deficit, cellular respiration accelerates while phosphorylation slows down and one observes a reduction of the P/O ratio. In conjunction with this uncoupling, the mitochondria swell because of the increase in the permeability of their membranes. This membrane hyperpermeability during magnesium deficiency depends again less on the changes in ATPase activity than on the direct effect of the increase in membrane permeability. The uncoupling of oxidation and phosphorylation seems to be a function of an increase in the permeability of the inner mitochondrial membrane to the influx of protons, while ATPase activity remains normal (571).
• Ribosomes require magnesium in order to maintain their physical stability: their aggregation into polysomes which can form peptide chains is possible only in the presence of Mg2+. During magnesium deficiency they dissociate into small particles. The almost exclusive function of ribosomes is the biosynthesis of proteins. The stabilizing activity of magnesium on polysomes maintains the structure of RNA complexes, allowing, with the help of elongation factors (188, 473) and polyamines (1149), the formation of polypeptides and ultimately the most stable conformation of the protein (12, 13, 14, 323, 546, 549, 1327).
• Magnesium also appears to protect lysosomes. Magnesium deficit in the rat, for example, is accompanied by a degranulation of mast cells that may be compared to the degranulation of polynuclear basophils in man. This is not the result of a degranulating factor in serum, but rather of a direct cellular effect of magnesium deficit, perhaps a reduction in the level of cAMP (343). Thus a diet deficient in magnesium, like any other degranulating factor, can be classified among the conditioning agents that increase susceptibility to necrosis (918). Above all, magnesium deficit in the rat along with excessive blood levels of histamine and serotonin due to mast cell degranulation offers an experimental model approximating the clinical expression of an immediate type hypersensitivity. The binding of reagins to mast cells or their circulating equivalents, polynuclear basophils, allows these homocytotropic antibodies to bring about the same events (1038).
• In the nucleus more than half the magnesium is closely associated with nucleic acids and mononucleotides. Magnesium is necessary for the physical integrity of the double helix of DNA which carries genetic information and the code for specific proteins (1340). Its excess favors, at least in vitro, the formation of the Z form of DNA which is antigenic (367, 719, 1261). It likewise facilitates the formation of messenger RNA which transmits to ribosomes information about the sequence of specific amino acids. Finally, nuclear magnesium appears necessary for the structural integrity of chromosomes (115, 620, 621). Magnesium deficiency in female rats causes chromosomal anomalies in both mother and fetus.
Magnesium thus acts as a major cellular and subcellular stabilizing agent which is necessary for the stability of plasma membranes, for the integrity of mitochondria, lysosomes, polysomes, and chromosomes as well as for the integrity of the helix of DNA and of messenger RNA and of RNA complexes. One can understand, therefore, its use as a cell protecting agent for organ transplants (1349) or during open heart surgery (1151, 1330, 1389) as well as its indispensable character for major biochemical functions: activation of the major metabolic pathways for carbohydrates, proteins and lipids, redox reactions, acid-base equilibrium, ion metabolism.... Necessary for cellular integrity and distributed throughout the organism, magnesium clearly must participate in the principal physiologic functions.
1.3 The physiologic properties of magnesium
Our enumeration of these characteristics will constitute a logical introduction to clinical practice, but it will be brief enough to allow further discussion when we reach the subject of physiopathologic analysis.
It is necessary to avoid any confusion between the pharmacodynamic properties of magnesium and its physiologic properties. Pharmacodynamic properties are established either in vitro or in isolated organs or sometimes in cells of species quite removed from humans and for variations in magnesium intake that are incompatible with life. Physiologic properties correspond to those observed in vivo preferably in man or the whole animal, thus maintaining the possibility of general or local reactions. The best proof of a physiologic action of magnesium is to reveal its role in a deficiency by administering a dose to correct the deficiency. Similarly, its role may be revealed when an overload is corrected by ceasing magnesium intake.
1.31 Neuromuscular physiology (191, 222, 223, 224, 268, 332, 337, 345, 346, 348, 359, 363, 364, 370, 380, 466, 728, 749, 860, 1012, 1014, 1015, 1016, 1091)
The magnesium ion has a sedative effect on the nervous system. Thus, magnesium deficit provokes a diffuse neuromuscular hyperexcitability that has effects on both the central and the peripheral nervous systems. These effects act on cortex and subcortex, on the extra-dural neurons and on the neuromuscular junction (curare effect) and on both the voluntary and involuntary nervous systems. In regard to effects on the autonomic nervous system, it is of interest that magnesium accomplishes its sedative functions by stimulating their inhibitory effects and depressing their stimulatory activity (860). It acts especially as a ganglioplegic. Experimental deficiency may sometimes induce organic nervous lesions.
Magnesium's effects on muscles occur most often in a manner similar to those affecting nerve excitability: it achieves its neuro-sedative effects by lowering the excitability of the muscular fiber, either striated (muscle relaxing effect) or smooth (musculotropic effect). Experimental magnesium deficiency occasionally causes the lesions of myopathy.
On the other hand, magnesium increases muscular efficiency when storing actin molecules, using them in energy requiring reactions or synthesizing them.
1.32 The cardiovascular system and the formed elements of the blood
Magnesium acts on the containing system, the heart and the vessels, as well as on what it contains, the circulating formed elements and humoral risk factors.
1.321 Magnesium, the heart and the vascular system (28, 29, 32, 33, 457, 1165)
With the myocardium, magnesium acts as a calcium antagonist. It reduces conductivity and irritability. It exercises an energy enhancing effect. It exhibits cardio-protective effects, both anti-hypoxic and anti-ischemic. Magnesium deficiency may, conversely, induce cardiomyopathy.
Magnesium protects the vessel wall by reducing the possibility of calcium overload and of connective tissue disturbances and by direct vasodilating effects and indirect antispasmodic effects on the muscles. Conversely again, experimental magnesium deficiency may cause vascular lesions.
1.322 Magnesium and vasopathogenic blood factors
Magnesium plays a physiologic antithrombotic role. (323, 324, 325, 346, 353, 359, 1239, 1240). This effect depends on its platelet-stabilizing capacity. Magnesium deficiency in man and in the rat causes a magnesium curable hyperaggregation of platelets. This effect is due to a reduction in the plasma factor that degrades ADP, thus slowing the breakdown of serum aggregating activity. (323, 325, 1240). A direct action on membranes is also possible (421). This may also be linked to the erythrocyte stabilizing capacity of magnesium. Magnesium deficit reduces the deformability of red cells with a parallel decrease in erythrocyte ATP and 2,3-DPG (962) as well as structural alterations in membranes (421, 422, 423, 757).
The effects of magnesium on leucocytes — enhancing phagocytosis and the production of lymphocytes and their transformation into blast cells while moderating the inflammatory reaction (342, 349) — are less important in the analysis of the effects of magnesium deficit on the formed elements of the blood as a vascular risk factor than are its stabilizing effects on platelets and erythrocytes.
Magnesium deficit, on the other hand, favors vascular disease by its effects on the metabolism of phosphorus and calcium in calcinosis, on the metabolism of proteins, carbohydrates and glycoproteins in alterations analogous to connective tissue changes due to aging and finally on the metabolism of lipids and lipoproteins by enhancing atherogenic dyslipidemias (32, 33, 57, 146, 343, 396, 727, 1057, 1061, 1064, 1065, 1067, 1161).
1.323 Magnesium and other systems
Magnesium is necessary for the growth and mineralization of bone (338, 570, 902). It is involved in the synthesis and activation of alkaline phosphatases, pyrophosphatases and ATPase and in the formation of collagen and sulfated mucopolysaccharides (with induction of the synthesis of magnesium- and ATP-dependent phosphoadenosine phosphosulfate or PAPS, which is the key to sulfur activation). In this manner, magnesium encourages ossification, acting on the protein matrix of bone as well as on its mineralization and activating both osteoblasts and osteoclasts. On the other hand, magnesium may behave like a "poison" in the formation of crystals of apatite (323). Magnesium deficiency produces a true "aging" of bone with a slowed turnover of bone cells and a reduced receptivity to the hormone-like D vitamins as well as to parathyroid hormone. Magnesium deficiency affects the teeth before affecting skeletal bones (338, 694, 902, 1292).
Magnesium is involved in renal physiology. It has common reabsorption sites with calcium and is apparently necessary for the conservation of phosphorus and potassium as well as for that of amino acids, especially taurine. It also appears to be necessary for responsiveness to ADH and for the secretion of acid by the kidneys through its action on the carbonic anhydrase system and on ammonia production (1032, 1281).
Magnesium is involved as well in reproductive functions. Magnesium plays a role in the gametes' transmission of genetic information and activates the motility of the spermatozoa. It shows a "tocolytic" action on uterine muscle, i.e. decreases gravid myometrial contractility (389, 1228) and assures proper fetal trophicity (199, 389, 619, 621, 622, 624, 715a, 715b, 1326, 1411, 1412).
Magnesium participates in the synthesis of enzymes and of the mucins in digestive juices as well as in the synthesis of the digestive polypeptides e.g. CCK-PZ (cholecystokinin-pancreomyzin). It stimulates various hepatic functions: glycogenic, lipotropic and detoxifying.... Magnesium deficiency may bring on changes in gastric, intestinal or hepatic function. Also, while the exocrine pancreas may not be affected (383, 1265), the functioning of the endocrine pancreas may be altered by a severe and prolonged magnesium deficit (396).
In the lungs, magnesium participates in the synthesis of surfactant (1100).
Magnesium exhibits a number of interactions with hormones (732, 889, 1334). It may be necessary for their synthesis (insulin), their storage (catecholamines) or their release (PTH). It may act on their peripheral receptors, most often in a favorable way (PTH, oxytocin, ADH, vasopressin, insulin, beta adrenoreceptor), but sometimes as an antagonist (alpha-adrenoreceptor, thyroxin). Several mechanisms may explain the action of magnesium on the endocrine glands: effects on cyclic nucleotides. permeability of the plasma membrane, activation of genes and the synthesis of messenger RNA, the pentose phosphate pathway, mitochondrial coupling of oxidative phosphorylation (141, 328, 361, 398, 490, 732, 1383).
Magnesium is involved in active membrane mechanisms that prevent hyperhydration of the crystalline substance of the lens and its becoming opaque (339, 706, 799, 1226). It is active in the physiology of the retina in association with high levels of phosphorus, ATPase, carbohydrates and taurine (336, 397). It is also involved in olfaction and hearing (481, 658).
Finally, magnesium is involved in a number of defense processes. It exhibits a variety of "antistress" functions: antihypoxic, anti-allergic, anti-anaphylactic, anti-inflammatory. These effects also include stimulation of phagocytosis, a role in activating the complement system (fixing of C2 on C4 in the classic pathway, but also activation of the alternate pathway), regulation of acid-base equilibrium, redox reactions and homeothermia. These antistress functions also include antitoxic effects as for example in the cases of cardiotonic glycosides (234, 260, 905, 1165, 1262), vitamin D (1165), lead (540) and acetaldehyde (390).
Found in many tissues, magnesium is thus involved in almost all of the physiologic functions of the organism. We must therefore now discuss its metabolism in detail.
2. The metabolism of magnesium
Magnesium is absorbed, stored and excreted. In the case of marginal deficit or load, there is no symptomatology. In the case of severe deficit or overload, the symptoms become apparent. It is necessary thus, with a disturbance of magnesium metabolism, to analyze the mechanisms which permit the latency of marginal disturbances. To this end, one must contrast the direct cellular effects of the metabolic disturbance with the general responses that it evokes. Then it is necessary to determine from among the latter those that are useful, serving to maintain the constancy of the internal environment, i.e., distinguish the homeostatic responses from the harmful ones.
Magnesium deficiency produces low magnesium levels in the extra-cellular compartment and a reduction of levels in the cell along with hyperpermeability of the cell membrane. This depolarisation finally causes a lowered level of cellular potassium and a calcium overload (increase of intracellular Ca), in conjunction with a lowering of phosphorus levels and an increase in intracellular Na+. The increased influx of calcium into the cell produces lower blood levels of calcium and the release of potassium from the cell raises blood levels of potassium. Moreover, if the deficiency is prolonged, the cellular calcium overload may cause calcinosis, due to mixed apatite crystals which combine Ca, P and Mg. These salts that have no physiologic value increase the cellular levels of P and, paradoxically, those of Mg during prolonged severe MD (146).
Severe magnesium deficit may thus be accompanied in tissues by calcinosis and low cellular potassium levels (with sodium retention) while one can observe in blood low calcium, phosphorus and potassium levels and in urine low calcium and high phosphorus and potassium levels (361). It is therefore necessary to elucidate the regulatory mechanisms that cause the least severe magnesium deficits, both in experimental animals and in man, to remain latent and notably "normohumoral," that is to say, magnesium levels appear normal as do levels of calcium, phosphorus, potassium and sodium (Table 1).
Circulating levels of phosphorus-calcium and potassium-sodium are subject to well-defined endocrine regulation. Parathyroid hormone (PTH, calcitonin (CT) and 1,25-dihydroxycholecalciferol (1,25-(OH)2-D) are the major regulatory elements for phosphorus and calcium metabolism, renin, aldosterone and insulin for potassium and sodium metabolism. It is thus logical to assume that the compensatory mechanisms for disturbances of calcium and potassium metabolism in magnesium deficit can bring into play the parathyroid and medullary-thyroid glands and hormone-like vitamin D in the kidney to control the calcium disturbance and renin, aldosterone and insulin to control the potassium problem.
Thus it is the whole of the endocrine system that one must consider as entering into the regulation of magnesium homeostasis (732, 1334). It will be necessary in the study of hormonal modifications of magnesium homeostasis to distinguish between harmful and useful manifestations, between harmful endocrine responses that must be treated and homeostatic endocrine responses that must be respected. Vitamin D is hardly involved at all since it is not generally modified in vivo in magnesium deficit (16, 902, 903, 1062, 1371) while the renin-aldosterone system can exercise only harmful effects (732). With magnesium deficit there exists, in effect, a hyperfunction of the juxtaglomerular apparatus and an excessive production of aldosterone (189, 190) which can only aggravate any potassium problems (360). Therefore only the positive effects of other endocrine changes will be analyzed.
Magnesium overload is only roughly a mirror image of magnesium deficit. It is important to note that, as with magnesium deficit, the moderate forms remain latent and are clearly without consequence for calcium and potassium metabolism. Large overloads generally occur in conjunction with calcium problems but very rarely with disturbances of potassium metabolism (916).
The direct membrane effects of magnesium excess normally cause a release of calcium (and sodium) from the cell with subsequent hypercalcemia and an increased influx of potassium (and magnesium) with the resulting cellular excess of potassium and hypokalemia.
In fact, the compensatory mechanisms operate so vigorously that large magnesium overloads are, on the contrary, accompanied by hypocalcemia (with hypercalciuria) and problems with blood levels of potassium are unusual: there are rare cases of hypokalemia with hyperkaluria.
Therefore, in order to define the essentials of magnesium metabolism and of its control mechanisms, we must,
review how absorption, storage, and secretion of magnesium occur.
then analyze in order the regulatory mechanisms of this metabolism, studying them from four different points of view:
the regulation of magnesemia, and its remarkable stability which needs to be explained (14, 361, 380, 398, 583, 770, 1314, 1327),
the regulation of the principal ionic consequences (calcium and potassium) of magnesium disturbances (361, 398).
the regulation of cellular magnesium, since it is necessary to understand the relatively constant levels of intracellular Mg (75, 146,168, 591, 1313).
the regulation of the cellular consequences of magnesium problems, mainly those that concern cyclic nucleotide second messengers, but also the principal cellular ionic effects, stressing at each step the failure of these controls and contrasting their homeostatic value in the latent forms with their ineffectiveness in symptomatic decompensated forms (361, 398).
2.1 Absorption, storage and excretion of magnesium
Absorption of magnesium in man occurs mostly in the intestine, more in the small than in the large intestine and more in the distal small intestine than in the proximal (994). This absorption represents the sum of two mechanisms. One is a simple passive diffusion process and can thus in no way be part of a regulatory mechanism. The other involves the transport process of facilitated diffusion: passive diffusion from the intestinal lumen into the parietal cells, followed by passage from the cells into the blood by an energy dependent mechanism. The second process is saturable, which may play a role in resisting variations in magnesium intake (409).
The absorption of magnesium is linked to that of water (111). Absorption improves with increased transit time (1336) and depends as much on the physico-chemical nature of magnesium, especially its ionization (13), as on the equilibrium among the various constituents of the diet and among various hormonal secretions. Conditions favoring Mg absorption include a slight acidity and a diet rich in protein (especially animal protein), in unsaturated fatty acids, medium chain triglycerides and volatile fatty acids, in vitamin B, sodium, lactose and vitamin D. Similarly diets that enhance the secretion of insulin, PTH and perhaps certain polypeptide digestive hormones (vasoactive polypeptide hormone or VIP, calcium elevating peptide or CEP for example) also improve magnesium absorption while an alkaline environment, some vegetable proteins, ammonia compounds, saturated fats, dietary fiber, phytic acid, and an excess of P, Ca or alcohol tend to inhibit magnesium absorption (25, 162, 334, 349, 361, 363, 390, 391, 396, 398, 432, 601, 605, 742, 803, 821, 927, 1121, 1145, 1220). Overall, magnesium is an ion that is poorly absorbed. On the average only 30% of intake is absorbed, of which 10% is by passive diffusion (1101).
Magnesium stores are almost entirely intracellular.
More than half is found in skeletal bone, a quarter in skeletal muscle and the rest is found throughout the body, predominantly in the nervous system and in organs of high metabolic activity: myocardium, liver, digestive tract, kidneys, exocrine and endocrine glands and the hemo-lymphatic system (334).
Reserves of magnesium are found mainly in bone. It is in effect very difficult to lower magnesium levels in soft tissues during a magnesium deficiency (13, 14, 1327).
We have at present only partial knowledge about the mechanisms of magnesium transport in the cell and the maintenance of the gradient that exists between intra-and extracellular Mg. Uptake of magnesium by the cell seems to depend on a mechanism of facilitated diffusion while release of magnesium from the cell depends on active transport, both energy-using processes that require glucose (52, 396, 546, 550). Vitamins B6and D, insulin (either directly as an ionophoric hormone or indirectly as a hormone that stimulates hepatic and renal hydroxylations that activate vitamin D) and perhaps taurine are capable of increasing the cellular level of magnesium. These hormones and vitamins moreover represent at present the only "magnesium fixing" agents that are available to us. In the opposite manner, adrenalin, perhaps by beta stimulation, reduces the level of magnesium in the cell (332, 345, 358, 359, 363, 364, 380, 396, 397, 398).
The major excretion pathway for magnesium is through the kidneys (1281). Intestinal secretions and sweat are generally of secondary importance. Diffusible magnesium (ionized and complexed) in the plasma (68% of plasma magnesium) is filtered by the glomerulus and then reabsorbed at a level of 96.5, more in the ascending branch of the loop of Henley than in the proximal tubule where Na, K and Ca are absorbed at twice the rate. Secondarily, secretion may occur in the descending branch of the loop of Henley.
It is very important to note that the absorptive mechanism functions at or very near saturation (409, 1039, 1338).
A number of factors have been found to affect urinary output of magnesium. Expansion of extracellular fluids, especially by perfusion with Na, any agent causing hypercalcemia (Ca, vitamin D, lactose), proteins, all sugars metabolized in the kidney (like glucose), alcohol, acidifiers, ADH, aldosterone, T3 and T4, all increase magnesuria, while P (in moderate doses), CT, PTH, glucagon, insulin and the D vitamins (in doses that do not cause hypercalcemia) reduce magnesuria (72, 361, 390, 391,396, 398, 754, 976, 1025, 1039, 1040, 1041, 1187, 1281).
2.2 Regulation of magnesium metabolism
2.21 Regulation of magnesemia
We will first examine regulation of blood levels of magnesium, stressing their remarkable stability. This is the result of the direct intervention of the organs that govern magnesium metabolism, the kidneys and the intestines, as well as of medullary-adrenal hormonal regulation.
2.211 1 The role of renal excretion
Traditionally, renal excretion constitutes the major regulatory mechanism for magnesemia. Urinary magnesium excretion increases in the case of magnesium overload and drops with magnesium deficit (13, 14,1281,1327).
It is certain that, since magnesium reabsorption is accomplished at levels close to saturation for physiologic levels of blood magnesium, any hypermagnesemia will automatically induce a compensating hypermagnesuria. The importance of this control is undeniable; it represents an effective passive regulatory mechanism for magnesium overload (916).
However, this passive mechanism cannot explain completely the homeostatic reduction of magnesuria during magnesium deficiency. Certainly, non-saturated magnesium reabsorption can function at a maximum during deficiency and thus cause a lowering of blood magnesium levels (1041). But the latter conforms to criteria that are incompatible with a "passive" phenomenon: the persistence of an obligatory loss and the carbohydrate dependence of this phenomenon (336, 396, 581). In fact, fasting therapy with the obese shows that homeostatic urinary reduction of magnesuria is of short duration. It begins again only with the resumption of carbohydrate intake (1065). The reduction of magnesuria to combat magnesium deficiency represents an active phenomenon whose control signals must be complex.
2.211 2 Intervention of storage mechanisms
It is not impossible that the ratio of extracellular magnesium to intracellular reserves of magnesium remains constant by means of an adaptive release of certain labile forms of magnesium (993). This has been suggested for hepatic magnesium (825). The liver would in this case regulate plasma magnesium by acting directly on the active process that maintains the gradient between intracellular and circulating magnesium.
2.211 3 Intervention of absorptive mechanisms
In man, absorption ranges from 75.8% for a diet very low in magnesium (23 mg/day) to a level of 23% for a diet rich in magnesium (564 mg/day) (523). By virtue of its capacity for facilitated diffusion, intestinal absorption is a linearly saturable function and can thus account directly for this adaptation to variations in intake. It is also possible that indirect mechanisms of absorption may be involved in order to counteract a magnesium deficit that is not linked to an insufficient intake. An increase of magnesium absorption is in fact observed in the rat during magnesium deficit due to urinary hyperexcretion of magnesium caused either by a loop diuretic such as furosemide (950) or by increasing renal clearance with triiodothyronine (995). We do not know the steps in such a regulatory system with the exception of the slowing of intestinal transit, the only mechanism observed during magnesium deficiency that is capable of inducing a compensatory increase in the intestinal absorption of magnesium (176, 724).
One can propose the hypothesis of a local hormonal regulation of magnesium absorption through the intermediary of digestive polypeptide hormones. Magnesium is in fact capable of modifying their secretion and several of these peptide hormones are capable of affecting magnesium absorption (21, 396,1121).
Increasing magnesium absorption, during a deficit of this ion, is not in any case a constant regulatory mechanism. It is missing for example in the case of an experimental deficiency in humans (315).
In this way it is established that homeostasis of blood levels of magnesium depends directly on good renal function and secondarily on the efficiency of intestinal magnesium absorption. But these two processes at not sufficient to explain the constancy of the levels of magnesium in blood. Feedback mechanisms that regulate these two main factors at a distance as well as distribution of the ion must be involved.
2.212 1 Medullary-adrenal feedback control of magnesemia
Magnesemia can be controlled by a medullary-adrenal feedback mechanism which was first described in the rat (731, 732, 735, 1059, 1060). Intraperitoneal injection of adrenalin causes hypermagnesemia that is a function of the administrated dose, an effect confirmed with isoprenaline (1284, 1285). Noradrenalin remains without effect. These data have been confirmed in the rabbit where adrenalin increases serum and erythrocyte magnesium (890). Moreover, the medullary-adrenal system is involved in the genesis of various kinds of hypermagnesemia such as those due to stress, fasting, etc. Inversely, magnesium loading strongly inhibits medullary-adrenal secretion of adrenalin. Magnesium deficiency appears to be accompanied by excessive levels of adrenalin in the blood, perhaps due to stimulation of the pentose phosphate pathway in the medullary-adrenal gland (490). Magnesemia in the deficient rat is in fact negatively correlated with phosphorylase A levels in muscle and positively with the reduced glycogenesis that results (336, 1303). In the course of primary human magnesium deficit, one can clearly observe high levels of adrenalin in the urine which are moreover less constant and more moderate than any concomitant increase in urine noradrenalin (303,305, 1290). Finally, removal of the adrenal medulla produces an aggravation of experimental magnesium deficiency (735).
Thus all the elements of a medullary-adrenal regulatory feedback mechanism for magnesemia are found grouped together. With magnesium deficiency there would occur a hyperadrenalinemia that corrects the lowered blood levels of magnesium. In the case of magnesium overload there occurs a lowering of the physiologic levels of adrenalin that subsequently reduces the degree of hypermagnesemia (361, 398).
Adrenalin acts neither by increasing intestinal absorption of magnesium nor by reducing urinary excretion. It can act only by modifying exchanges between intracellular reserves of magnesium and circulating magnesium (361, 398), probably by stimulation of beta adrenal receptors, perhaps atypical ones (731), thus inhibiting influx of magnesium into the cell (546).
This feedback mechanism is all the more credible because excessive levels of plasma and erythrocyte magnesium have been observed in human patients with pheochromocytomas. In every one of the five cases observed, there has been a return to normal levels of magnesium after surgical correction of the medullary-adrenal tumor (890). Finally, it is important to stress that the action of adrenalin on magnesemia has been verified in humans for physiologic levels of the hormone (142, 230, 264).
2.212 2 Failure of medullary-adrenal feedback control
This feedback mechanism may be ineffective when the response is excessive or insufficient
2.212 21 During magnesium deficit, it can in the initial stages overshoot its goal and give rise to one of the rare forms with hypermagnesemia (345, 346, 363. 364). At a more severe level hyperadrenalinism exhibits lipolytic effects, however, lipolysis reduces magnesium levels in the blood (46 464, 1065), both in animals and in humans (stress, delirium tremens, infarction), probably by complexing with ionized magnesium that may be chelated with fatty acids or fixed in the adipocytes (427, 1325). This occurrence is even more harmful because magnesium deficit increases the toxicity of adrenalin (1323, 1331) (Fig. 2) . Finally, medullary-adrenal feedback control may on the contrary be missing. This is a frequent occurrence since in clinical practice elevated urinary excretion is observed in only half of the cases of chronic magnesium deficit. This is doubtless the effect of the habitual exhausting of a reactive hypersecretion that has been called on too many times (303, 1290).
2.212 22 With severe magnesium overload, shock secondary to the drastic reduction in blood pressure that is due to arterial vasodilation substitutes high blood levels of adrenalin due to stress for reduced medullary-adrenal effects. The consequences are not unfavorable for the reduction of hypermagnesemia. In moderate magnesium overload, blood levels of magnesium are reduced by a homeostatic hypoadrenalinism, while during severe overload, magnesium levels in the blood are reduced by a lipolysis that is due to a strong hyperadrenalinemia caused by stress. On the other hand, hyperadrenalinemia has a role in the hyperglycemia and the hypophosphoremia that are customary with severe magnesium overload (142, 396, 397).
2.221 Regulation of the calcium and potassium problems caused by magnesium deficit
2.221 1 Endocrine feedback control of hypocalcemia and low cellular potassium
It is logical that the hypocalcemia induced by the direct cellular consequences of magnesium deficit should involve a response by the control system for calcium homeostasis, whether it be hyperparathyroidism (hypersecretion of PTH), renal hypersecretion of 1,25-(OH)2D, or finally hyposecretion of calcitonin (CT Ø)
In fact, hypersecretion of PTH is actually observed, both in animals and in humans (570, 732, 950, 1055, 1187, 1352). Hyposecretion of CT is very difficult to discover because of the limited sensitivity of techniques for measuring this hormone. It is consistent that it has been described recently in the spasmophilic form of magnesium deficit (893, 1264). On the other hand, an increase in 1,25-(OH)2D does not seem to be involved; it is not normally observed in experimental animals or in humans during magnesium deficit (361, 398, 783, 902, 903, 1062, 1371). These normal levels of l,25-(OH)2D, seemingly inappropriate in hypocalcemia, may be caused by the effect of antagonists on vitamin synthesis. The direct cellular effects of magnesium deficit on the hepatic and renal tissues in which the magnesium-dependent steps of the hydroxylations that activate vitamin D occur should tend to reduce the synthesis of 1,25-(OH)2D as has actually been observed in vitro (903, 1062). On the contrary, various general and humoral effects may combine to stimulate the synthesis of vitamin D, whether it be insulin (361, 396, 398), hyperparathyroidism or hypophosphatemia (641). The digestive hormones may also participate in remedying calcium problems secondary to magnesium disturbances. VIP (vasoactive intestinal polypeptide) and CEP (calcium elevating peptide) are two hormones capable of increasing blood calcium levels and it is possible that magnesium is involved in the secretion of digestive hormonal peptides (25, 742, 821, 1121). An increase in thyroxin (T4) has been considered as a protective clement against nephrocalcinosis (1281), but its many harmful effects, especially an increased urinary clearance of magnesium, prevent its being considered as a regulatory factor (732). One should note that we are not referring here to studies of magnesium deficiencies in rats, the only animal to develop hypercalcemia in such cases (669, 732). In this species the intervention of the PTH-CT system works in the opposite direction; one observes a lowering of PTH and an increase in CT (732, 1224).
Low cellular potassium is not corrected by a relaxing of the renin-aldosterone system. Quite to the contrary, there occurs a hypertrophy of the renal juxtaglomerular apparatus with excessive secretion of aldosterone (189, 190). It is thus a question here of an endocrine reaction, apparently a harmful one, that makes the loss of potassium worse (361, 398, 732).
The low cellular potassium is probably improved by hypersecretion of insulin. This has never been the subject of direct proof, but seems likely. In effect insulin increases tolerance for carbohydrates as well as influx of potassium into the cell. Conversely, magnesium deficiency diminishes tolerance for carbohydrates and reduces effector sensitivity to insulin (336, 396, 489, 776). The response should be a reactive hyperinsulinism. Its signs have been described since a degranulation of the beta cells of the islets of Langerhans has been observed during experimental magnesium deficit (634). Moreover, hypomagnesemia produces hypersecretion from the perfused pancreas in the rat (118, 336, 361, 364, 396, 398, 651, 741, 822). Induction of this hyperinsulinism due to magnesium deficit might simply come from a lessening of the inhibitory effects of insulin secretion on the beta cells of the islets of Langerhans in the pancreas that are normal with Mg (396). This induction may be linked to stimulation of the pentose phosphate pathway (490). It might also depend on the stimulatory effects of insulin secretion due to hypersecretion of glucagon (741) or even on the much more hypothetical action of digestive peptide hormones (396, 398).
Thus, the indirect calcium and potassium homeostatic regulatory mechanisms could be summed up as a hypersecretion of PTH and insulin and a hyposecretion of CT.
2.221 2 Failure of calcium and potassium feedback controls
These feedback mechanisms may be ineffective either through a response that is insufficient or through one that is excessive.
2.221 21 Failure by lack of response
• PTH The mechanism of reactive hyperparathyroidism may be defective, due either to a lowering of receptor affinity (in bone rather than in the kidneys) (937, 1017), or above all to a reduction in parathyroid secretion (60, 897, 902, 1017, 1187). In severe deficits, one observes hyperparathyroidism. This is due not to insufficient synthesis but rather to a blocking of parathyroid hormone release (879, 902, 1187).
• CT. We do not know how long the eventual slowing of CT secretion in the course of prolonged primary magnesium deficit will last. However, in the particular case of magnesium deficit secondary to streptozotocin-induced experimental diabetes in the rat, it is possible to find an acute phase with hyperparathyroidism and reactive hypocalcitoninemia followed by a chronic phase with failure of these reactions. With the transition from hyper- to hyposecretion of parathyroid hormone, there occurs a parallel substitution of hypercalcitoninemia for slowed medullary-thyroid secretion (396, 611, 1189).
• Insulin. Insulin may be secreted over a long period in the course of severe magnesium deficit in man (274). However, this reaction ultimately disappears since severe and prolonged experimental deficits can produce hypoinsulinism (396). The unexpected hyposecretion of insulin during such magnesium deficits may derive either directly from suppression of the "permissive" role of magnesium in beta secretion in the islets of Langerhans, for which a minimum level of magnesium thus appears necessary, or indirectly from the hypocalcemia and low cellular potassium due to the magnesium deficit. Ca2+ and K+ are in fact the two major ions that stimulate insulin secretion (119, 266, 396, 475, 578, 1058, 1068, 1398).
• 1,25-(OH)2D. If, in the equilibrium between the direct cellular elements in magnesium deficiency that inhibit production of 1,25-(OH)2D and the general reactions that stimulate this synthesis (hyperparathyroidism, hyperinsulinism, hypophosphoremia), the latter become reversed, a reduced and harmful blood level of hormone-like vitamin D may be found. Magnesium therapy then corrects the lowered vitamin and calcium levels less rapidly than it does the magnesium-dependent hypoparathyroidism (259, 479).
2.221 22 Failure due to excessive response
• PTH and CT. Too sharp a response of the regulatory mechanisms for phosphorus and calcium explains the rare hypercalcemic (26, 543, 1076) or hypophosphoremic (543, 902) forms due to excessive hyper parathyroidism (1264), without one's being able at present to account for an excessive slowing of calcitonin secretion.
Hyposecretion of calcitonin which is exaggerated and harmful may however be observed in a situation that involves a urinary wasting of magnesium, for example, mineralocorticoid hypertension in the rat. Expansion of the extracellular compartment produces in this case urinary hyperexcretion of magnesium (and calcium) which is opposed by a hyposecretion of calcitonin in order to avoid a syndrome of magnesium (and calcium) deficit. This correction surpasses its goal, making magnesium (and calcium) balance clearly positive (124). The magnesium overload which thus occurs appears to play a key role in the physiopathology of this kind of hypertension, since a diet low in magnesium is sufficient to prevent it (1092). Magnesium excess must, in this model of endocrine and nutritional dysregulation, especially affect the vascular walls (146).
• Insulin. An excessive reactive hyperinsulinism accounts for the hypoglycemic forms of magnesium deficit (336,345,359,396, 1227) (Fig. 3) .
• 1,25-(OH)2D. If, among the elements that maintain equilibrium in the production of 1,25-(OH)2D, there occurs an excess of all or some of the three factors that produce it (hyperparathyroidism. hyperinsulinism hypophosphatemia), one can observe high blood levels of vitamin D. In this manner the only case where vitamin D is implicated in a homeostatic phenomenon involving magnesium is produced. This occurrence appears exceptional both in human and in veterinary medicine (647, 767). By exceeding its goal it can set the stage for hypercalcemia with an excess of circulating vitamin D (955).
2.222 Regulation of calcium and potassium disturbances during magnesium overload
2.222 1 Endocrine feedback control
Uncompensated magnesium overload normally produces a release of Ca from the cell (inducing hypercalcemia) and an excess of cellular K.
• The hypercalcemia is opposed by the usual response of the couple CT-PTH. i.e., a hypersecretion of calcitonin linked to hypoparathyroidism. It is interesting that the quickness with which these mechanisms operate — (probably because of the extreme danger created by acute hypercalcemia) — derives from the intervention of the medullary thyroid system which is even more rapid than that of the parathyroid (902).
• Excessive cellular potassium does not seem to be compensated by an activation of the renin-aldosterone system (916). If magnesium perfusion can increase the renin level in the isolated kidney (1367), magnesium perfusion in the human deficient in Na does not change the level of plasma renin (689). On the other hand, it would be tempting to attribute a role to a reduction in the physiologic secretion of insulin. Magnesium overload induces, in fact, a reduced peripheral cellular uptake of glucose, suggesting the possibility of hyposecretion of insulin (336, 396). In addition, hyposecretion of insulin is produced by magnesium overload both in the isolated rat pancreas (118,266) and in vivo, but in the latter case only with large doses of magnesium administered during glucose-stimulated insulin secretion (822).
Thus, the indirect homeostatic regulatory mechanisms for calcium and potassium during magnesium overload can be summed up as a hypersecretion of calcitonin coupled with a hyposecretion of PTH and insulin (Fig. 3) .
2.222 2 Failure of calcium and potassium feedback controls
In severe magnesium overload as, for example, in parenteral magnesium therapy with large doses, one finds an excessive response of the regulatory system opposing the resulting hypercalcemia. It is normal in this case to find hypocalcemia with hypophosphoremia. An analysis of the mechanisms of this response demonstrates the dominant role of calcitonin hypersecretion compared to that of a slowing of parathyroid secretion. There is a rapid delaying of the onset of hypocalcemia: the hypocalcemia is associated with hypophosphoremia (732, 902); and finally the hypocalcemia induced by the magnesium load may sometimes be accompanied by a hypersecretion of parathyroid hormone, doubtless to balance the hypocalcemia (263).
However hypophosphoremia may also be due to the release of adrenalin during the "stress" of low blood pressure. This activity of adrenalin in causing low blood phosphorus appears to depend directly on exchange effects between the intra- and extracellular milieux and not on its effects on CT-PTH secretions (142, 230, 264).
It might he, on the contrary, the ineffectiveness of the low insulin levels that explains the rare forms with hypokalemia (916) (Fig. 3) .
One can thus affirm that during magnesium overload four types of endocrine feedback control can intervene in a favorable manner (Fig. 4) .
During magnesium deficit one can find hypersecretion of medullary-adrenal adrenalin, hypersecretion of parathyroid PTH, hyposecretion of medullary thyroid CT and hypersecretion of insulin from the beta cells of the pancreas. And, conversely, during magnesium overload, low blood levels of adrenalin, hypersecretion of calcitonin, hypoparathyroidism and hypoinsulinism are evident.
It is very interesting that the two hormonal feedback controls for calcium, parathyroid and medullary-thyroid, exercise absolutely parallel favorable effects on the metabolism of magnesium and calcium (950, 1055). PTH and CT can simultaneously control not only the levels of circulating Ca but also those of Mg. It is however evident that the normal quantitative relationship between Ca and Mg is such that this method of regulation can only come into play with serious variations in magnesium levels (361, 398, 865, 1187). It is possible that the endocrine membrane receptors for the two ions are different. In fact, the inhibitory effect on PTH secretion of one of the cations is increased by even a minimal concentration of the other (173, 409). The role of calcitonin in this scheme of magnesium homeostasis accounts for the effects of thyroidectomy, independent of the iodinated hormones (732). These are null in the case of magnesium deficit where there is a slowing of CT secretion (732); they become evident with magnesium overload where, by preventing the release of calcitonin, they slow the return to normal of magnesium levels (732).
Insulin feedback control can also in parallel play a useful role in regard to both Mg and K. Insulin acts in a parallel manner on Mg2+ and K+ with, however, a much more intense effect on the monovalent ion than on the divalent. Insulin increases the absorption of magnesium (895) and its cellular uptake by its action as an ionophore in magnesium translocation (336, 361, 396, 398, 776, 1356, 1357) and diminishes urinary loss of magnesium, probably by decreasing glycosuria (336, 409). Insulin appears through these diverse effects to be a true "magnesium-sparing" hormone (361, 396, 398). Moreover, hyperinsulinism is capable as well of an antagonistic effect in regard to the lipolysis that lowers blood levels of magnesium and that is induced by large releases of adrenalin (396).
In reciprocal fashion, medullary-adrenal feedback control of magnesium reinforces calcium feedback control. In fact adrenalin stimulates secretion of PTH. Thus certain excessive reactive forms of hyperparathyroidism in magnesium deficit can be controlled by the use of beta blockers, the effects of which are particularly correlated with the rise of blood phosphorus levels (1290), But this effect may also constitute evidence for the direct control of adrenalin's ability to lower phosphorus levels (142).
Endocrine control of blood magnesium levels and that of the ionic consequences of the deficit or of the overload can therefore be summarized in the two following diagrams, referring particularly to tour types of glands:
medullary-adrenal, parathyroid, thyroid-medullary and the beta pancreas.
In magnesium deficit:
ä adrenalin, ä PTH, Ø CT, ä insulin
In magnesium overload:
Ø adrenalin, Ø PTH, ä CT, Ø insulin
In marginal or moderate deficit or overload, the efficiency of these controls finally results in latency. In severe forms of the deficit, an insufficient or excessive response by the regulatory systems ends in decompensation (Fig 4) .
It seems of interest to divide these four hormones into two groups:
1. The first is formed by the PTH-CT couple of which the effects are particularly important for exchanges between the extracellular compartment and body stores in the hard tissues (bones and teeth).
2. The second brings adrenalin and insulin together, presiding more particularly over exchanges between the extracellular compartment and body stores in the soft tissues.
2.23 Regulation of cellular magnesium
We know only a few of the details about the elements that control the cellular levels of magnesium and even less about its heterogeneous distribution in the cell.
2.231 Control of cellular magnesium levels
Parallel variations of adrenalin and insulin levels represent the only mechanism known in disorders of magnesium metabolism that tends to maintain a constant level of cellular magnesium in the soft tissues. These two hormones represent one of the two groups that assure regulation of magnesium levels in the blood, essentially regulating exchanges between the extracellular compartment and magnesium in the soft tissues. So, while these two hormones combine in a synergistic modification of magnesium levels in blood, they act in opposite ways in regard to cellular magnesium levels (Fig. 7) . Insulin, as an ionophoric hormone, facilitates translocation of magnesium in the cell where it increases the influx of magnesium (396, 776, 1356, 1357). In the opposite direction, adrenalin facilitates the loss of magnesium from the cell, perhaps by acting on the beta adreno-receptors (361, 398, 546). Thus the parallel and synergistic regulatory variations of blood magnesium levels do not exercise this homeostatic effect at the expense of cellular magnesium, but do so rather by assuring its stability. Since the gradient between intra- and extracellular magnesium is glucose-dependent (336, 396), it is not surprising to see that two of the major hormones in the regulation of the metabolism of carbohydrates assume so essential a role in the regulation of magnesium.
At present we do not know how to evaluate the ultimate physiologic role of three other "magnesium-fixing" agents also involved: the D vitamins, vitamin B6, and taurine. These three compounds are moreover also closely linked to carbohydrate metabolism (396).
2.232 Control of intracellular magnesium distribution
No control mechanism for the distribution of magnesium during disturbances of its metabolism has yet been described.
However an initial contribution to this absorbing problem can be found in the description of problems of intracellular distribution of taurine during magnesium deficit. This sulfonic amino acid which is capable of modifying the physicochemical state of intracellular Mg and Ca exhibits an intracellular distribution that is profoundly modified in experimental magnesium deficiency in the rat (401, 1341, 1342, 1343, 1344).
Moreover, if one sees that problems of calcium distribution can be related to changes in calmodulin levels (396, 398, 546), one can picture by analogy alterations in a hypothetical "magmodulin" that are linked to the intracellular distribution of magnesium (1333)
There also exist tissue modifications of insulin-dependent calmodulin (1002) and insulin secretion is involved in magnesium homeostasis (361, 364, 396, 398).
2.24 Regulation of the cellular effects of magnesium disturbances, especially those involving cyclic nucleotide second messengers, and also of major ionic effects
Regulation of the consequences in tissues of metabolic variations of an ion that is almost completely intracellular appears to be of major importance.
2.241 Regulation of cyclic nucleotide second messengers
It seems to us of particular interest to study the cellular effects of magnesium problems on cyclic AMP and GMP. These two nucleotides represent in fact cellular second messengers, constituting targets that are affected as much by local modifications of tissue ions as they are by the general reactions they produce. At the outset, we can consider changes in cyclic nucleotides induced by the involuntary neuro-receptors during modifications in magnesium metabolism.
2.241 1 Cellular effects
In vitro studies have allowed us to describe well the effects of a magnesium load as well as those of magnesium deficit on the involuntary nervous system.
A magnesium load directly favors the action of the autonomic nerves when they have a paralysing effect and depresses this action when the effect is stimulatory (860). This must correspond to the schema of beta-adrenergic stimulation (229, 697, 838, 971) and alpha-adrenergic and muscarinic inhibition (222, 223, 224, 860) (Fig. 5) .
Conversely, magnesium deficit directly induces beta-adrenergic inhibition and alpha-adrenergic and cholinergic stimulation. This corresponds to production by the corresponding involuntary receptors of diminished levels of cAMP which is beta-adrenergic dependent (136, 290, 595) and increased levels of cGMP which is alpha-adrenergic and acetylcholine dependent (509). It is tempting to attribute these modifications of cyclic nucleotides during magnesium deficit (where magnesium tends to decrease in the cell while calcium increases) to the classic Mg2+-dependence of adenylate cyclase (136, 290, 595, 813, 1111) and the Ca2+-dependence, at least in vivo, of guanylate cyclase (483, 1222, 1270).
2.241 2 General effects
2.241 21 Cyclic AMP
In latent forms of magnesium deficit, reduction in cAMP levels may be countered by in vivo beta stimulation induced by an increased production of catecholamines (both medullary-adrenal and sympathetic), and, secondarily, by the action of histamine on H2 receptors. These phenomena, observed both in experimental animals and in clinical practice with humans, offer an example of effective regulation of a cellular disturbance in magnesium deficit. This regulation may, however, exceed its goals. In the animal deficient in magnesium, there occur increases in cellular cAMP (546, 548, 549, 550, 1061) (Fig. 5) and in man one can find sympathicotonic forms of magnesium deficit, both responding to beta blockers (303, 363, 364, 399, 579, 635, 1290). On the other hand, hyperinsulinism and, although much more hypothetically, certain polypeptide digestive hormones tend to maintain the reduction of cAMP levels (141, 396) (Fig. 5)
2.241 22 Cyclic GMP
An increase in cGMP, on the contrary, does not, as far as we know, activate compensating mechanisms. Quite the opposite is true. The reactions of the organism during magnesium deficit appear to make the situation worse. Alpha-adrenergic stimulation, vagotonia, over-production of insulin, H1 receptor stimulation by histamine, hyperinsulinism and perhaps even a hypothetical production of peroxides and free radicals constitute factors that are capable of increasing the level of cGMP (509, 931, 1350). According to the present state of our knowledge, it appears that an increase in cGMP would represent a basic cellular sign of magnesium deficit. It is in agreement with the frequency of the amphotonic forms of magnesium deficit, responding both to dihydroergotamine and to belladonna (345, 358, 361, 363, 364, 398). It also is consistent with the aggravating effects of trinitrin (415,416, 657, 931).
In addition this increase in cGMP should counteract the stimulating effects of the neurohormones that increase the synthesis of cyclic AMP. In fact, the increased synthesis of the cyclic guanylic nucleotide is evidently accomplished at the expense of its basic substrate, GTP. Thus, neurohormones only effectively activate adenylate cyclases in the presence of sufficiently high levels of GTP (106, 290, 595).
However, since there are forms of magnesium deficit that do not exhibit these involuntary nervous system signs that are cyclic (IMP-dependent, there must be corresponding regulatory mechanisms.
The taurine hypothesis of cGMP regulation (Fig. 6) .
• Mobilization of taurine represents at present the only known mechanism capable of reducing the levels of cGMP during magnesium deficit (361, 364, 398). During experimental magnesium deficiency in the rat, high blood levels of taurine are found (613, 614, 1087, 1088, 1341, 1342, 1343) with a concomitant increase of taurine levels in skeletal muscle (1087, 1089), the intestines and the liver (1341. 1342, 1344), as well as in urine (614, 1088). Magnesium deficit also provokes in various tissues changes in the distribution of cellular taurine (401, 1341, 1342, 1343, 1344). Taurine, at least in nervous tissue, is capable of lowering an increased level of cGMP (539). An increase in cellular taurine could thus be the first mechanism described for the homeostatic reduction of increased cGMP during magnesium deficit.
• The involvement of this regulation by taurine depends on mechanisms that are still poorly understood. Increased cellular uptake of taurine during magnesium deficit might depend on beta adrenoreceptor stimulation which actually increases influx of taurine into the cell (66, 219, 220, 1073). More hypothetically, it might depend on a hyperinsulinism since insulin increases the influx of amino acids into the cell (818). This use of extracellular taurine would cause an increased synthesis of endogenous taurine which occurs in the "endocrine" kidney in man (488) or would facilitate by an unknown mechanism an increase in the intestinal absorption of the sulfonic amino acid and/or its precursors. This increase in taurine synthesis could be enhanced during magnesium deficit, as has been observed during zinc deficit (612,531), by a simultaneous reduction in protein synthesis. Common to both these deficiencies, low protein production increases the availability of the constitutive amino acids of protein. In particular, this leaves an increased level of sulfur amino acid precursors of taurine free for taurine synthesis (402b).
One can imagine similar effects in magnesium overload, but they have not yet been directly studied. The taurine feedback system of regulating the cellular level of cyclic GMP during problems of magnesium metabolism remains therefore still largely hypothetical at present (361, 364, 398).
Defects in the taurine regulatory system
This taurine regulatory system is also quite fragile. (402b)
• Beta-adrenal stimulation can be excessive. In this case it reverses its effects on the cellular uptake of taurine, reducing instead of increasing it (775).
• It may, in contrast, as with hyperinsulinism, exhaust itself. If both these situations occur, the two mechanisms that increase cellular taurine are removed
• Urinary excretion of taurine must not exceed absorption of exogenous taurine or endogenous renal synthesis of taurine (361, 397, 398).
2.242 Regulation of the cellular ionic effects of magnesium disorders
Thus three "hormones", insulin, adrenalin and taurine, combine to exercise synergistic effects by participating in cellular magnesium homeostasis and in the effects of magnesium disorders on cyclic nucleotide second messengers.
They do not, however, all have the same favorable effects on other consequences of magnesium disturbances (Fig. 7) .
If, in order to conserve cellular K+, insulin acts in a positive way (1356, 1357), adrenalin remains, at least in physiologic doses, without effects (142).
In addition, insulin and adrenalin both have harmful effects on cell membranes as well as on cell phosphorus and calcium metabolism.
On the one hand, the two hormones have additive effects in depolarizing membranes (332, 345, 363, 380, 397, 398), which is undesirable given the same cellular effects of magnesium deficit. On the other hand, insulin and adrenalin promote calcium and phosphorus overload. Insulin stimulates entry of phosphorus into the cell (818) and owes much of its effect to its power as an ionophore in translocating calcium (396). Through beta stimulation, adrenalin stimulates influx of calcium into the cell (35, 546, 549, 1213) and above all has at physiologic levels a powerful reducing effect on blood phosphorus levels through the transport of phosphorus from the extracellular compartment to the interior of the cell (142).
Thus adrenalin-insulin regulation brings along with its beneficial effects others which favor tissue calcinosis. One can probably find here an explanation for the general mechanisms that substitute calcinosis for the initial loss of phosphorus during cellular magnesium deficit, with precipitation of insoluble phosphates, thus fixing Ca and Mg in deposit is without biologic value.
Finally, adrenalin (by alpha stimulation) and insulin have effects that tend synergistically to increase the priority of the levels of cyclic GMP in relation to those of cyclic AMP (Fig. 7) .
These different negative effects of the adrenalin-insulin couple in regulation of the cellular effects of magnesium disturbances make the role of taurine all the more important. Moreover, it is doubtless not by chance that adrenalin and insulin favor cellular influx of taurine. This would appear rather to be a regulatory mechanism. Taurine seems in fact able to oppose the undesirable cellular effects of the adrenalin-insulin couple (Fig. 7) . Taurine appears to be a powerful membrane stabilizer (81, 82, 84, 88, 361, 398, 401, 626, 627, 1138, 1139) whose effects at physiologic doses are comparable to those of magnesium. They appear to be achieved more by a para-cellular action on the physico-chemical structure of the membrane than by a cellular route, that is to say, essentially by means of membrane ATPases (84). This action on membranes must facilitate the maintenance of potassium levels in the cell (Fig 7) .
It is therefore possible that the organism deficient in magnesium can try non-specifically to balance the membrane changes secondary to Mg2+ deficit by mobilizing the stabilizing capacity of taurine. But it is also possible that the action of taurine on the cell membrane seeks to keep Mg2+ an essentially intracellular cation, in the cell. Taurine would then be acting like a true "magnesium sparing" hormone (361, 398).
Taurine also appears to be able to chelate calcium. It can thus oppose the effects of calcium overload in the cell, the major mediator of the biochemical consequences of magnesium deficit, an effect enhanced by changes in its subcellular distribution during magnesium deficiency (401). This action must also contribute to the reduction in levels of cyclic GMP since guanylate cyclase in vivo is often Ca2+ activated (Fig. 7) (402b).
Regulation by taurine can be seen thus at the conclusion of this analysis of the regulatory mechanisms of magnesium homeostasis as having great importance by its effects whether specific in sparing magnesium or favorable non-specific ones on membranes, ions or cyclic nucleotides (Fig. 7) (402b).
It is not presently possible to know what role to assign in humans to L-gamma-glutamyl taurine, a peptide hormone secreted by the parathyroid glands. This hormone is thought to possess biologic properties analogous to those of taurine (451). Perhaps it represents that second hormone of the parathyroids which is distinct from parathyroid hormone and is active in magnesium metabolism, the presence of which we have suspected during experiments — now long since past — that we conducted on the relationship between the parathyroids and magnesium in the dog (373, 374).
Such is the state of our present knowledge of the metabolism of magnesium. The factors that control the stability of blood levels of magnesium and of intracellular Mg2+ as well as the different consequences of compensated magnesium disturbances include not only the effectiveness of renal excretion but also activation of a complex neuro-endocrine system. Interactions with the hard tissues bring into play the PTH-CT couple (and perhaps digestive polypeptide hormones). Exchanges with soft tissues elicit secretion of adrenalin and insulin, of taurine (and perhaps of L-glutamyl taurine), and finally beta stimulation of the adrenergic receptors. (Fig. 8) .
This page was first uploaded to The Magnesium Web Site on July 13, 2002