There is an unsettled debate about the role of magnesium as a ‘chronic regulator’ of biological functions, as opposed to the well-known role for calcium as an ‘acute regulator’. New and old findings appear to delineate an increasingly complex and important role for magnesium in many cellular functions. This review summarizes the available evidence for a link between the regulation of intracellular magnesium availability and the control of cell growth, energy metabolism and death, both in healthy and diseased conditions. A comprehensive view is precluded by technical difficulties in tracing magnesium within a multicompartment and dynamic environment like the cell; nevertheless, the last few years has witnessed encouraging progress towards a better characterization of magnesium transport and its storage or mobilization inside the cell. The latest findings pave the road towards a new and deeper appreciation of magnesium homoeostasis and its role in the regulation of essential cell functions.
Mg (magnesium; referring to both bound and free forms) is referred to as the intracellular divalent cation par excellence. Its biological role is extremely versatile as it can serve structural functions (e.g. fluidity and stability of phospholipid bilayers, protein tertiary or quaternary structures and DNA double helices) as well as dynamic functions (e.g. cofactor or allosteric modulator of enzyme activities). Thanks to its double-positive charge, Mg can bridge negatively charged molecules; thanks to its small atomic radius it can occupy most narrow sites, thus competing with other divalent cations for specific binding sites [1,2]. Within this flexible scenario, however, the best recognized function of Mg is its association with ATP4− and the consequent facilitation of transphosphorylation reactions that are crucial to cell activation/deactivation as, for example, in signal transduction pathways. Given these premises and in view of its supra-millimolar concentration in most tissues, Mg should be considered the most aspecific ligand within the cell. Nevertheless, several genetic and acquired diseases, such as hypomagnesaemia associated with hypocalcaemia, cardiovascular diseases, hypertension, diabetes and metabolic syndrome, to mention the most important ones, suggest that a derangement of Mg availability and metabolism may be considered the primary cause and/or consequence of pathophysiological conditions [3–6].
The discussion of the multiple functions or pathways that are affected by intra-/extra-cellular Mg availability in specialized tissues (such as, for example, cation reabsorption in kidney tubules, regulation of glutamate receptor in neurons and contraction in muscles) is beyond the scope of this review and would require an entire volume. To provide a general view of the importance of Mg in cellular pathophysiology, we will review and discuss Mg-related cell functions that are shared by most tissues, namely proliferation, metabolism and apoptosis.
Mg AND CELL PROLIFERATION
In the 1970s and 1980s, several investigators described the essential role of Mg in the proliferation of yeast and mammalian cells [7,8]. Harry Rubin postulated the theory of ‘the coordinated control of cell proliferation’, and proposed that Mg was the key factor that regulated the different steps of this complex process . At the onset of the process, Mg is involved in receptor-mediated mitotic signals mediated by, for example, growth factors; then, Mg is essential to sustain protein synthesis occurring prior to cell division. Subsequently, Mg is involved in DNA duplication, as polymerases and ligases require Mg-ATP2−. And finally, Mg is involved in the cytoskeleton re-arrangement leading to the formation of the mitotic spindle and cytokinesis. Rubin  was able to describe the relationship between the sequence of these events (growth rate, thymidine and uridine incorporation etc.) and the availability of Mg. These observations provided the conceptual foundations for probing the involvement of Mg in the different phases of the cell cycle. It has been shown that incubation of cells in low Mg concentrations causes growth arrest, resulting in an increased percentage of cells in G0/G1-phase, a decreased percentage of cells in S-phase and, occasionally, a slightly increased percentage of cells in G2/M-phase. All these results are consistent with the multiple roles of Mg in each phase of the cell cycle, as illustrated in Figure 1.
Pathway of cell proliferation and its regulation by Mg
In the attempt to investigate how Mg could regulate cell proliferation at the molecular level, we have shown that incubation in low-Mg medium triggers the up-regulation of the cell cycle inhibitor p27 , whereas others have found up-regulation of p21 . Further findings supporting the capability of Mg to influence the cell cycle were obtained by DNA expression profiling (RNA arrays). It was found (, and F. I. Wolf, V. Trapani, M. Simonacci, A. Boninsegna, A. Mazur, and J. A. M. Maier, unpublished work) that a low Mg content in the growth medium causes up-regulation of p53, as well as of Jumonji and numblike, two newly identified negative modulators of cell proliferation [12,13]. In parallel, cyclins D and F and the transcription factor E2F, which promote cell cycle progression by activating CDKs (cyclin-dependent kinases) and transcription of S-phase-specific genes respectively, were down-modulated.
High Mg availability, as obtained by increasing Mg concentration in the extracellular medium, accelerates cell proliferation to a variable extent [14,15]. When cells are Mg-depleted, the re-addition of Mg causes a rapid and substantial rise in the proliferation rate, suggesting that cell cycle arrest is reversible . Nevertheless, a chronic adaptation to an increased extracellular Mg neither causes a substantial increase in the proliferation rate nor leads to a significant increase in cell Mg content . Under such defined conditions, however, RNA arrays of cells grown in high Mg uncovered the up-regulation of several cell-cycle-related genes, specifically cyclin F and ETS-related transcription factor (F. I. Wolf, V. Trapani, M. Simonacci, A. Boninsegna, A. Mazur, and J. A. M. Maier, unpublished work).
Different cell types have a different dependence of their proliferation rate on extracellular Mg availability. Endothelial cells proved to be highly sensitive to changes of Mg availability, as shown by the fact that their growth rate decreased to a considerable extent when the cells were maintained in 0.1 mmol/l extracellular Mg; HC11 mammary epithelial cells, usually grown in 0.8 mmol/l extracellular Mg, were less sensitive and had a 50% growth inhibition only upon exposure to 0.05 mmol/l Mg [9,18]. Tumour cells proved to be the most resistant cell type; for example, HL-60 leukaemia cells, MCF7 mammary carcinoma and Ehrlich ascites tumour cells had an unaltered growth rate at 0.05 mmol/l extracellular Mg [9,19]. However, every cell strain can adapt to non-physiological concentrations of extracellular Mg, as shown by the case of chronic adaptation to very high or very low Mg-containing medium [14,20].
A common feature in the Mg-dependent control of cell growth pertained to the modulation of cell-cycle-inhibitory proteins such as p27 and p21 [9,10,21]. We described in detail the time-dependent and concentration-dependent up-regulation of p27 in HC11 mammary epithelial cells grown in Mg-deficient medium, as well as the inhibition of p27 expression under the opposite conditions (i.e. 45 mmol/l extracellular Mg) [9,14]. These findings led us to suggest that the extracellular Mg availability can ‘directly’ affect the cell cycle by influencing the transcription of related genes. At this time, however, we do not know the complete molecular pathway whereby extracellular Mg can affect gene expression. Two mechanisms can be envisaged. The first one implies that a drastic decrease in the extracellular Mg concentration may determine sizeable modifications of Mg pools inside the cells. This, in turn, affects pathways leading to increased p53 levels and to a consequent transcription of cell-cycle-inhibitory proteins; one such mechanism probably occurred in HC11 mammary epithelial cells. The second mechanism hypothesizes that extracellular Mg may act via a mechanism similar to that occurring for Ca-sensing receptors . The first mechanism sounds more plausible, but some linking steps are nonetheless missing. We know that many cell types respond to incubation in Mg-free medium by activating a measurable Na-dependent Mg efflux, as shown, for example, in , which should perturb intracellular Mg pools and alter Mg-dependent enzymatic activities. On the basis of such a mechanism, Rubin  proposed that mTOR (mammalian target of rapamycin), a PI3K (phosphoinositide 3-kinase)-related kinase which initiates protein synthesis, might be the key activity modulating G1-phase protein synthesis. This kinase is characterized by a Km for ATP of 1.0 mmol/l, which is 50–100 times greater than that of most protein kinases. A concentration of 1 mmol/l would be close to the concentration of ATP within the cell, but one should also consider that Mg-ATP2− is the only active form of ATP. It follows that the limiting step of this specific kinase reaction would be the full availability of Mg to form Mg-ATP2− at a concentration close to 1 mmol/l .
There is limited experimental evidence to explain how the intracellular distribution of Mg is affected by removing extracellular Mg. Interestingly, modifications of Mg availability may lead to sizeable changes in the contents of other cell ions, namely K+ loss, and Ca2+ and Na+ accumulation . Consequently, the resulting cellular effects, that may at first appear ascribable to Mg deprivation, are in fact secondary to other ion fluctuations.
In some cell types, such as cardiomyocytes, vascular smooth muscle cells and hepatocytes, intracellular Mg can be modulated by receptor-mediated stimuli such as hormones or growth factors [26–28]. The difficulties in obtaining relevant results are due, in part, to the lack of specific and sensitive methods to trace intracellular Mg. Studies on rat thymocytes grown under Mg deficiency suggest that Mg availability might affect the expression level of stress proteins, potential triggers of the inhibition of the cell cycle .
The relationship between Mg and oxidative stress has been a matter of investigation in the last few decades. Many studies are available in the literature, but most of them have been obtained by inducing hypomagnesiaemia in vivo, a condition that unavoidably involves the contribution of many other factors, such as inflammation, cytokine production and activation of phagocyte oxidative burst [30,31]. With the most novel experimental approaches, primarily arrays, the mRNA levels of some scavenger enzymes were in fact shown to be affected by Mg availability. In particular, glutathione transferase was up-regulated in some tissues [29,32]; however, an up-regulation of scavenger enzymes does not necessarily reflect the actual detoxifying ability of the cell, as demonstrated by several studies in which the enzyme activities of tissues or cells kept under Mg-deficient conditions proved to be unaffected, stimulated or even inhibited [33,34]. In addition, direct evidence that the depletion of extracellular Mg affected the levels of intracellular ROS (reactive oxygen species) was obtained under limited experimental conditions [29,35,36]. Finally, it must be pointed out that endogenous ROS can trigger different kinds of cellular responses, varying from a stimulation of cell proliferation (as demonstrated by activation of Akt pathway) to growth arrest or apoptosis [as in the case of NF-κB (nuclear factor κB)-driven signals] [37,38].
Growth arrest can be triggered by up-regulation of p53 in response to DNA damage. In principle, low extracellular Mg could induce DNA oxidative damage by two different mechanisms: first, by increasing intracellular ROS, and secondly, by influencing DNA repair mechanisms [39,40]. In principle, the two mechanisms could act synergistically, but a proof of evidence is lacking. We have recently shown that LLC (Lewis lung carcinoma) tumours grown in Mg-deficient mice have higher levels of 8-OHdG (8-hydroxydeoxyguanosine), the most important guanine oxidation product, than do the same tumours grown in Mg-repleted mice .
Whether extracellular Mg acted through a mechanism similar to that of Ca-sensing receptors remains a matter of speculation. The recently identified inward transporters of Mg, namely TRPM6 and TRPM7 (where TRPM is transient receptor potential melastatin) ion channels, possess unique structural characteristics . The cytosolic C-terminus of TRPM6 and TRPM7 has an α-kinase activity that has inspired the definition of this class of channels as ‘chanzymes’ . The possibility that these proteins acted as channels and kinases suggested that they might play a role in the regulation of Mg homoeostasis and related cell functions . The activity of TRPM6/TRPM7 appears to be modulated by Mg, whether in the form of an Mg-ATP complex or in the form of an Mg2+ ion that becomes available inside or outside of the cell . However, the contribution of the channel-associated kinase activity to the transport of Mg and the regulation of cell signals and functions awaits further clarification [45,46].
In conclusion, although the experimental evidence supports the hypothesis that Mg is essential to sustain cell growth and that Mg depletion directly induces growth arrest by up-regulating p27 and p21, a general molecular mechanism leading to G0/G1 arrest is still lacking and further work is required to ultimately associate the intracellular consequences of decreased Mg availability (i.e. modifications of intracellular Mg pools) with the modulation of key proteins responsible for the control of cell proliferation.
Mg AND ENERGY METABOLISM
The large majority of the enzymes involved in glycolysis, the Krebs cycle and the respiratory chain, which represent the core of energy metabolism, are Mg-dependent by two main mechanisms: Mg can act either as an allosteric modulator, as in the case of enolase, or as a cofactor in the form of Mg-ATP2− . A complete list of the enzymes involved in energy metabolism and modulated by Mg would be too long to be exhaustively discussed in the present review. In principle, it can be stated that, in most cases, Mg deficiency depresses metabolism. On the basis of previous investigations by our group, we focus here on the role of Mg in regulating glucose uptake and glycolysis, two functions that are key points in diseases such as diabetes and cancer.
A number of years ago, we observed in tumour cells that the addition of excess Mg into the extracellular medium led to an impressive ATP hydrolysis occurring in less than 15 min . We showed that this phenomenon was due to a Mg-dependent stimulation of phosphofructokinase, as fructose 1,6-diphosphate was accumulated. This effect only occurred in intact cells, as permeabilization did not reproduce the same phenomenon . We hypothesized that extracellular Mg either stimulated glucose uptake or favoured the interaction between glycolytic enzymes and the cytoskeleton to form an ideal channel of ordered glycolytic enzymes, as suggested by Srivastava and Bernhard . At that time, we were unable to demonstrate that Mg caused a direct stimulation of the glucose transport mechanisms . Of note, these experiments were carried out on tumour cells, which are known to exhibit very high aerobic glycolysis. This raises the possibility that the glycolytic activity of tumour cells was high enough to preclude any further stimulation by Mg.
Following our pioneering work on the energy metabolism of tumour cells, the relationship between Mg and glucose utilization emerged from the observations that pathologies such as diabetes and metabolic syndrome are associated with a defined disregulation of Mg homoeostasis, leading to hypomagnesaemia and decreased intracellular Mg content . Intracellular Mg regulates insulin action, insulin-mediated glucose uptake and vascular tone. Mg deficit has been proposed as a possible underlying common mechanism of the ‘insulin resistance’ of different metabolic conditions . The association between Mg availability and glucose utilization by an insulin-mediated mechanism supports our early hypothesis that Mg could stimulate glucose transport into cells. Several studies have attempted to demonstrate such a mechanism. Romani and co-workers have shown in perfused hearts and isolated cardiomyocytes that insulin stimulates Mg uptake in the presence of extracellular glucose . These results indicate that extracellular and/or intracellular Mg may modulate glucose transport and/or utilization . In hepatocytes, the same authors showed that glucose mobilization was essential for Mg extrusion induced by adrenergic agonists and that Mg2+ transport through the hepatocyte plasma membrane was altered under diabetic conditions as a result of a decrease in intracellular glucose [51,52].
However, molecular details on how extracellular or intracellular Mg could directly or indirectly regulate glucose uptake and vice versa are still lacking. It is also possible that, in these circumstances, Mg modulates some phosphorylation/dephosphorylation events which are crucial for regulating the membrane transport of glucose and Mg, and that insulin resistance consists of a drastic modification of these regulatory reactions, as reviewed recently [6,53].
The remaining energetic pathways, namely the Krebs cycle and the respiratory chain, take place in mitochondria which are considered intracellular Mg stores. The concept of intracellular Mg stores derives from qualitative, rather than quantitative, observations; this is due to the peculiar characteristics of intracellular Mg homoeostasis, which does not have chemical gradients similar to those described for Ca. In an attempt to quantify Mg in intracellular compartments by electron probe microanalysis, we found that the amount of Mg detected in nuclei, mitochondria or cytosol/microsome regions had variations hardly approaching 10% (118±3, 101±3 and 116±4 mmol/kg of dry weight respectively) . From a qualitative point of view, however, two lines of evidence support a role for mitochondria in the storage of Mg. First, by live-cell imaging and using novel Mg fluorescent probes, it has been shown that uncouplers induce an Mg efflux which is distinct from ATP hydrolysis and Ca efflux . Secondly, specific Mg transporters located in the inner mitochondrial membrane have been genetically characterized and distinguished from those found in other membranes. Among these are yMrs2 in yeast  and hsaMrs2p in humans ; their expression proved to be crucial to the survival of mitochondria.
As mentioned in the Introduction, the chemical characteristics of Mg favour its binding to many species, such as proteins, phospholipids or nucleic acids . Among all intracellular species, ATP is the one that displays the highest affinity for Mg (Kd=1×10−5 mol/l); ATP is therefore considered the major intracellular chelator or the ‘buffer’ of intracellular Mg. As ATP is synthesized mainly in mitochondria, it is not surprising that these organelles are rich in Mg. Indeed, the release of Mg from mitochondria has been associated with the release of adenine nucleotides, namely through ANT (adenine nucleotide translocase), by a cAMP-dependent mechanism [58,59]. Interestingly, this enzyme has been associated with the mitochondrial permeability transition, as discussed below.
The regulation of Mg within mitochondria has different pathophysiological roles. In addition to ATP synthesis and transport, Mg can affect the rate of respiration by regulating the function of some mitochondrial dehydrogenases, namely succinate and glutamate dehydrogenases  and cytochrome c oxidase . Mg can also affect the mitochondrial membrane potential, and hence the mitochondrial volume, by regulating the K+/H+ antiporter . However, the ultimate role of Mg in the pathophysiology of mitochondria has yet to be clarified by integrated studies, including the molecular biology of its transporters, its movements within subcellular compartments and the biological consequences for the organelles. In this context, experimental models consisting of cells characterized by overexpression or genetic deletion of Mg transporters should be of great help in understanding Mg regulation and function.
It must be pointed out, however, that mitochondria should not be viewed as the only intracellular Mg stores; the endoplasmic reticulum is also very rich in Mg, a feature that may well be consistent with its role in promoting protein synthesis [1,24]. On the other hand, information on Mg transport and distribution in the endoplasmic reticulum network is scarce and inconclusive. Recently, efforts to map intracellular Mg distribution using newly synthesized Mg-selective fluorochromes by live-cell imaging have provided a topology of intracellular Mg distribution which suggests that the cation is concentrated in membrane-rich perinuclear areas probably occupied by mitochondria and endoplasmic reticulum . More experimental work is needed to better understand the role of Mg in the regulation of energy metabolism and intracellular trafficking.
Mg AND APOPTOSIS
Several important aspects of Mg biochemistry and physiology point to a possible role for this cation in the apoptotic process. As discussed above, Mg is a key modulator of cell proliferation and metabolism, and Mg availability appears to affect the occurrence of oxidative stress, although the precise mechanisms are still unclear and debated. It is well known that the ultimate effect of several stimuli related to growth and differentiation may greatly vary, from proliferation to death, depending on the cell type, on the intensity or the duration of the stimulus, and on other concomitant factors. ROS are a typical example of such versatile biochemical players [37,38]. In the case of Mg, the issue is complicated further by the technical difficulty in detecting transient Mg movements at the intracellular level. Therefore it is not surprising that the available literature on the link between Mg and apoptosis is scarce and often contradictory.
Some have described that Mg deprivation induces cell death by apoptosis, for example in primary cultures of rat hepatocytes . In these experimental conditions, characterized by prolonged incubation in serum-free medium, the authors found a decrease in glutathione concentration and an increase in lipid peroxidation. They concluded that Mg deficiency provoked apoptotic death by an increased susceptibility to oxidative stress. Dietary Mg restriction in rats also accelerated thymus involution, a typical example of apoptosis , although in this case the inflammatory process could be a confounding factor. On the other hand, several studies have indicated a promoting role for Mg in apoptosis. We have shown that, under Mg deprivation, HL-60 leukaemia cells had growth inhibition and some features of neutrophil-like differentiation, and eventually developed resistance to apoptosis induced by etoposide and teniposide . Consistently, others have shown that both intrinsic  and extrinsic  apoptosis were accompanied by an early increase in intracellular Mg. Patel and co-workers  showed that the bile salt glycodeoxycholate induced apoptosis in rat hepatocytes, as detected by DNA fragmentation and in vitro endonuclease activity; at the same time, intracellular Mg levels, as measured by ratio imaging of Mg/Fura-2, had a 2-fold increase within the first 2 h of treatment. Therefore it was hypothesized that apoptosis is promoted by an influx of Mg which stimulates the activity of nuclear endonucleases. Chien and co-workers  also found an increase in intracellular Mg as an early event in apoptosis, but this was attributed to the mobilization of internal pools. Fas treatment of B-cells resulted in an increase in the cytosolic levels of free Mg and the number of cells that mobilized Mg; this occurred before phosphatidylserine externalization and DNA fragmentation. Interestingly, the intracellular source of Mg was hypothesized to be in the mitochondria, as extended treatment with the mitochondrial uncoupler FCCP (carbonyl cyanide p-trifluoromethoxyphenylhydrazone) appeared to impair Mg mobilization and the consequent apoptotic process.
The latter suggestion is particularly appealing, as it is well established that mitochondria play a central role in the onset of the apoptotic programme. Several lines of evidence also support this view. First of all, although the opening of the permeability transition pore appears to be dispensable for the release of cytochrome c from mitochondria, this is not the case for the presence of Mg, which instead appears to be an absolute requirement [69,70]. Another study  showed that an apoptotic compound, gliotoxin, might specifically activate an Mg efflux system from rat brain mitochondria in conditions of preserved mitochondrial integrity (i.e. high membrane potential, no swelling and retention of other ions). Finally, and most importantly, it has been shown that mitochondria are in fact intracellular Mg stores and that Mg release can occur following Ca release and prior to ATP hydrolysis . The presence of a specific mitochondrial channel for Mg, Mrs2, corroborates these findings ; as the activity of this channel is membrane-potential-dependent, it can be speculated that stored Mg might be released through the channel by depolarization. This situation would not be in contradiction with the findings by Chien et al. , where the increase in Mg preceding the onset of apoptosis is abolished by extensive depolarization: in such a case, Mg mitochondrial stores would be completely depleted by the prolonged FCCP pretreatment before the onset of Fas-mediated apoptotic signal.
Others have proposed that Mg release from mitochondria might be driven by ANT. In fact, Romani et al.  found that Mg efflux from mitochondria induced by cAMP and ADP was blocked by specific ANT inhibitors, such as atractyloside and bongkrekic acid. As ANT is one of the putative components of the mitochondrial permeability transition pore, these findings provide another interesting link between Mg efflux and the release of pro-apoptotic molecules, such as of cytochrome c, AIF (apoptosis-inducing factor), endonucleases G etc., as shown in Figure 2.
Pathways of apoptosis and the possible involvement of Mg
Overall, the role of Mg in apoptosis is far from being defined. If on one hand an anti-apoptotic effect of Mg can be ascribed to the improvement in the antioxidant defences, it is also true that the experimental design and the biochemical characteristics of oxidative agents, as well as the extent of the exposure to these agents, may all contribute to generate variability in the final outcome. In addition, as extensively discussed in the present review, Mg has pleiotropic effects that extend beyond the modulation of oxidative stress. For example, Mg is an essential cofactor for topoisomerase II, an enzyme involved in DNA replication and transcription or repair; this may help to explain our results found previously . Mg is also necessary for the activity of Ca/Mg-dependent endonucleases, which provides a straightforward reason for the increase of Mg prior to the apoptotic event par excellence, i.e. DNA fragmentation. The most recent findings [55,67–71] about the existence of mitochondrial Mg stores and specific channels suggest a finer involvement of Mg in the apoptotic cascade, and raise the possibility that this ion might be an important control element in life compared with death decisions. A change in intracellular Mg levels might result in altered Mg compartmentalization and mobilization and might eventually modify the progress of apoptosis. On the other hand, a change in mitochondrial function might also alter, reduce or abolish mobilization of Mg from intracellular stores in response to appropriate stimuli, which would certainly produce vast cellular effects and potentially impair subsequent apoptotic events. At present, these are only attractive speculations. It is most important to clarify whether the increase in intracellular Mg that occurs following an apoptotic stimulus was a coincidental event (due to, for example, mitochondrial depolarization) or a causative determinant of the apoptotic cascade. As such, Mg should be considered a ‘second messenger’ that triggers downstream events, as postulated by Grubbs and Maguire  some 20 years ago and reappraised by others more recently .
In conclusion, our understanding of the role of Mg in programmed cell death is still incomplete, because of technical limitations and of the inherent complexity of the apoptotic process. Nevertheless, there are indications that the role of Mg might be much more important than believed previously. The next few years should witness a deeper insight into this scenario.
In this brief review, we have attempted to provide an overview of the pathophysiological role of Mg in vital functions, such as cell proliferation, energy metabolism and apoptosis. Owing to its abundance and biological versatility, the role of Mg is only partially understood. With regard to the regulation of cell proliferation, there is convincing evidence that Mg deficiency induces growth arrest by affecting the expression levels of cell-cycle-regulatory proteins, including p27, p21, cyclins and CDKs. At the same time, Mg can regulate other steps of cell proliferation such as protein synthesis, DNA duplication and mitosis. As for energy metabolism, Mg regulates several crucial enzymes, and thus participates in insulin-dependent glucose uptake and utilization. Interestingly, mitochondria have all the characteristics to be considered as Mg stores, as they possess specific channels to take up Mg from the cytosol. In this regard, there is a challenging possibility that Mg participates not only in ATP synthesis, but also in the apoptotic pathway. As mentioned, we hope that the next few years will be devoted to improving our appraisal of Mg biochemistry and pathophysiology, so as to shed light on its unique roles in normal and diseased conditions.
We are gratefully indebted to Professor Achille Cittadini, who inspired the early research on Mg and kept supporting and encouraging our later work.