Magnesium Research (1995) 8, 1, 65-76
Summary: In order to assess total magnesium concentrations in human red blood cells (erythrocytes--ErMg), atomic absorption spectrometry (AAS) provides high accuracy and precise method rapid and amenable to automation. Taking care of eliminating the chronic marginal magnesium deficits, normal values of ErMg evaluated through a direct method and expressed as mmol/litre of packed cells are 2.3 ± 0.24. Inductively coupled plasma-mass spectrometry (ICP-MS) is a multielemental analytical technique. Particle induced x-ray emission (PIXE) also provides multielemental capability, but is time-consuming and costly. Microelectrodes are the gold standard for intracellular Mg2+ measurements. But microelectrodes and fluorescence probes measure the activity of magnesium ions, not the concentrations. Ionized magnesium content of human intact erythrocyte is mainly assessed with the NMR method and with the zero point titration. The concentration of ionized magnesium as estimated by NMR (31P NMR method) was found to be 0.20 ± 0.02 mmol/litre cell water and with the zero point titration 0.55 ± 0.12. The uncertainty concerning the two current used techniques for free magnesium determination is worsened by the fact that magnesium inside red cells continually oscillates in vivo. Free magnesium constitutes a small part of total magnesium. Further studies are necessary to assess the importance of its variations in clinical medicine. Efflux of ErMg is controlled through membranous sodium-dependent and sodium-independent pathways and through genetic and neurohormonal regulations. Variations in the total or ionized ErMg do not necessarily mean that similar changes should exist in the magnesium pool. But it remains the basic static cellular magnesium item. Its value will be subsequently enhanced when it takes place among the clinical and paraclinical data of dynamic magnesium investigations.
Key words: Atomic absorption spectrometry (AAS), erythrocyte, fluorescence, genetic and neurohormonal control, ion selective electrode (ISE), magnesium, nuclear magnetic resonance (NMR), null point, particle-induced X-ray emission (PIXE), red blood cell (RBC).
During acute or chronic experimental magnesium deficiency a decrease in erythrocyte magnesium (ErMg) is observed. It occurs more slowly and discreetly than in plasma magnesium. The measurement of this cellular magnesium item being easy, it has been conferred a primary role in routine static analysis of total or ionized cellular magnesium. Concentration of erythrocyte magnesium is related to their age and their quick renewal goes with an increase of erythrocyte magnesium unrelated to a magnesium overload.
The dogma of the lack of exchange between erythrocyte and plasma magnesium is far from being intangible. Red blood cells are distinct from the other cells of the organism as they have numerous specific characteristics namely the absence of nucleus and of mitochondria and therefore have a particularly low cellular magnesium concentration. But it remains none the less true that in clinical practice, erythrocyte magnesium concentration constitutes the easiest way of investigating cellular magnesium.
The aim of this mini-review is:
(1) To evaluate the current analytical techniques for measuring total magnesium and free magnesium in red blood cells;
(2) To sum up our present knowledge on membranous and systemic control of the red cell magnesium content.
And to conclude on the interest of erythrocyte magnesium evaluation as the best routine static cellular magnesium item for the investigation of magnesium status. Its value will subsequently be enhanced when it takes place among the clinical and paraclinical data of dynamic magnesium investigations.
The most widely used technique for measuring total magnesium is atomic absorption spectrophotometry (AAS)1. The sensitivity, defined as the concentration required for 1 per cent absorbance, is about 0.01 µ /ml for flame AAS at the 285.2 nm resonance line, while electrothermal (graphite furnace) AAS(ET-AAS) has a sensitivity of 0.17 pg2. The sensitivity of magnesium measurement by flame atomic absorption is sufficiently high that the graphite furnace instrument is infrequently required for biological determinations. The flame most often used is air-acetylene since it is sensitive, prohibits many interferences, and is readily available. While a fuel-rich air-acetylene flame has a greater sensitivity towards magnesium, an oxidizing (fuel-lean) flame is less susceptible to some interferences3. Thus, as a general rule, an oxidizing flame is recommended except when extreme sensitivity is required.
The worst interferences in magnesium AAS are observed with metals that form acid oxides that are stable at high temperature4. These elements, which include aluminium, silicon, titanium and zirconium, cause severe interferences only when present in relatively high concentrations4. Many other metals have been reported to interfere with magnesium AAS under various conditions including lithium, sodium, potassium, rubidium, chromium, selenium, beryllium, iron, vanadium, molybdenum, caesium, strontium, calcium, and barium,5,6. Chloride, oxalate, ethylenediaminetetraacetic acid (EDTA), and 8-hydroxyquinoline have been reported as anion interferences5,6. Although a large number of interferences exist, most are easily overcome. An air-acetylene flame prevents interferences due to sodium, potassium, calcium, phosphate and iron7. The presence of 0.1-1 per cent (w/v) lanthanum chloride or strontium chloride eliminates the remaining interferences except for those caused by chromium and titanium50. EDTA (0.4 per cent) overcomes the interference of chromium. Some commonly used biological buffers cause little or no interference of magnesium AAS. High (non-physiological) concentrations of phosphate have been reported to interfere5,7. However this interference is not seen with either an air-acetylene flame or a strontium or lanthanum diluent5,7. Metal-free solution of 50 mM Tris, Mes, or HEPES do not interfere with the determination of 0.25 µ/ml magnesium1. Protein has occasionally been reported to interfere8,9. In these cases, this effect is probably due to high sample viscosity which may alter the sample aspiration rate. Dilution of the sample is generally a convenient remedy to this problem. Factors that exacerbate interferences are a cool flame and the presence of nitrate or sulphate4,6.
Direct method: The assay of intracellular magnesium concentration in erythrocytes is performed in fresh erythrocytes which are washed three times with ice-cold solutions such as 140 mM choline chloride, 100 mM MgCl2, a Tris buffer pH 7.40 containing 146 mM choline chloride, 1mM MgCl2, 1 mM CaCl2, 5 mM ortho-phosphoric acid, a solution containing 75 mM MgCl2, 85 mM sucrose and 10 mM Tris-MOPS pH 7.40, a sodium phosphate-saline buffer pH 7.4010, or with 0.15 M saline 11.
After cell washing the recentrifugation of the cell suspension at higher g forces and longer times than 1200 g for 5 min does not result in further packing of the erythrocytes12.
Fluorescence polarization studies revealed a 15 per cent increase in the fluidity of membranes from magnesium-deficient rat erythrocytes and analysis of the membranes showed decreased amounts of magnesium13. Thus severe magnesium deficiency might change optimum centrifugation conditions.
The cells are then diluted and lysed with ice-cold bidistilled water. To speed up haemolysis, 0.2 per cent purified saponin is sometimes used 14.
After cell lysis the membranes are removed by centrifugation at 2000 g for 20 min. The haemolysate is diluted with an acidic solution, for example 0.1 M HCL)4 + 125 mM S r2 +10 or 0.3 M HCL containing 0.5 per cent (w/v) lanthanum chloride11 and the analysis is performed in an atomic absorption spectrophotometer. Usually erythrocyte magnesium concentrations are expressed as mmol/litre of packed cells (mean ± SEM) (2.08 ± 0.04), sometimes as mmol/kg wet weight (1.33 ± 0.06), as mmol/kg dry weight (4.03 ± 0.17)10, or as ng/10E6 cells (3.63-6.42) and µ/grams of haemoglobin15. But these red blood cell magnesium concentrations have been established without taking care of eliminating the marginal chronic magnesium deficit. If that is done, normal erythrocyte magnesium concentrations expressed as mmol/litre of packed cells are 2.30 ± 0.2414,40.
Indirect method: Although the direct method is considered to be more nearly accurate, some have speculated that part of the observed variability may be due to the distribution of cells in the erythrocyte pellet. Nucleated erythrocytes and reticulocytes, cells known to have higher concentrations of magnesium than mature erythrocytes, would be distributed at the top of the pellet. Thus the magnesium content of a specimen could be affected by the cell-size distribution of the final aliquot used for analysis.
Deuster et al. evaluated three methods (two indirect and one direct) for determining the magnesium content of red blood cells, to compare methodological differences and to establish a method suitable for use in field studies. For the indirect methods, erythrocytes in whole blood were lysed by adding either de-ionized water (I) or nitric acid, 2 mol/litre (II). For the direct method (III), erythrocytes were isolated by density centrifugation, washed, then digested in concentrated HNO3. Magnesium concentrations were measured by atomic absorption spectrophotometry in plasma and whole blood for the indirect method, and in the pellet for the direct method. Haematocrit and haemoglobin were measured, and erythrocytes were sized and counted on all samples. When values for the three methods were compared, that by method I was significantly lower than those by methods II and III. Values obtained by method II were 100.1 percent of that by the direct method. The indirect method with 2 mol/litre HNO3 lysing solution provides a reproducible, reliable, accurate, and simple technique for measuring magnesium in erythrocytes (micrograms per gram of haemoglobin): results (method II) in micrograms/gram of haemoglobin were (mean ± SD) 116.6 ± 14.7.
Bonnay et al.16 used a Kodak Ektachem 700 to measure red blood cell magnesium by using the neutralized supernatant of the perchloric acid-treated haemolysate. The supernate is used as if it were a plasma sample (overall dilution 1:6) in the Ektachem.
X-ray emission can be induced by particle beams (PIXE). During particle excitation of material, characteristic x-rays are emitted from target atoms. The PIXE method, used in conjunction with these microbeams, provides unique possibilities due to a combination of good analytical sensitivity, good spatial resolution and multielemental capability. The beam is reduced to microscopic size by an original designed microbeam line. The beam section can be adjusted from 1.5 µm to 10µm. Micro-PIXE is one of the few techniques able to quantify and locate an element within a cell. Micro-PIXE has enough qualities to become a powerful routine method in biology but it is still time-consuming and needs a particle accelerator which is only available in nuclear centres. The competitive methods SIMS (secondary ion mass spectrometry) and SEM (scanning electron microprobe) are either monoelemental or less sensitive but achieve best spatial resolutions.
Various techniques have been used to obtain erythrocyte populations of various ages. Most of the investigations of erythrocyte aging have been based on the assumption of an age-dependent difference in erythrocyte density. Given that the potassium concentration (mM/kg dry mass) of an individual erythrocyte can be used to mark the degree of senescence of the erythrocyte and given that electron-probe x-ray microanalysis technique allows the quantitative analysis of multiple elements in each erythrocyte then it becomes feasible to study systematic variations in the ionic concentration of the erythrocyte during the senescence process. Such a study circumvents the need for erythrocyte separation procedures. Specifically as potassium and Cl decreased in concentration, calcium increased in concentration whereas sodium and magnesium did not demonstrate a significant pattern of change. These findings are in keeping with the past observations that erythrocyte senescence is accompanied by: decreased water, decreased potassium and decreased volume, in addition to increased density and increased calcium12.
Schuette et al.18 recently described a general method for the accurate isotopic determination of magnesium (24Mg25Mg26Mg) in biological materials, which is based on inductively coupled plasma mass spectrometry (ICP-MS).
By itself, AAS is an analytical technique capable of providing high accuracy and precision of trace element quantitation. Generally, flame AAS is quite selective, rapid, amenable to automation, and therefore the technique of choice whenever adequately sensitive. Background absorption is not a particular problem and is completely compensated for by using a deuterium background corrector; however, it should always be checked for. Transport interferences are encountered with viscous sample solutions, and may call for viscosity-matched standards. inasmuch as high-salt solutions are often inevitable in flame AAS, it is essential to thoroughly optimize observation height and acetylene flow, so as to avoid time-consuming standard addition calibrations. Generally recognized assets of electrothermal AAS (ET-AAS) are its low detection limits, small size sample requirements, and the possibility of direct sampling. As such it is a method of choice. It has, however, some drawbacks and limitations: serious background problems, severe interferences, carbide formation, more complicated calibration, more critical optical alignment, problems with measuring fast transient signals, pipetting errors, pronounced sensitivity drift and memory effects, higher contamination risk, much longer instrumental time (e.g., 2-4 min per sample vs. 5-10 s in flame AAS), higher qualification of operator required, and higher cost.
Inductively coupled plasma-mass spectrometry (ICP-MS) is a multielement analytical technique. Its analytical domain overlaps both those of ICP emission and graphite furnace atomic absorption. For example, atomic spectroscopy detection limits (micrograms/litre) are 0.1 for flame AAS, 0.004 for ET-AAS, 0.08 for ICP emission and 0.007 for ICP-MS. In short, ICP-MS offers the analytical productivity of ICP emission with the detection limits of ET-AAS. ICP-MS is also being used for stable isotope tracer measurements. In this technique, chemicals enriched in one or more stable isotopes of an element are added to a dynamic system to trace the flow of the added element through it. Typical application is in the biomedical field for dietary studies. The enriched stable isotopes are relatively easy to obtain and are easier to work with than radioisotopes for tracer work.
Micro-PIXE is one of the few techniques able to quantify and locate an element within a cell. Besides spatial resolution Micro-PIXE provides multielemental capability. However this technique is time-consuming and costly.
While magnesium ions can modify numerous cell processes in vitro, their in vivo role remains unclear. The main reason for the uncertainty regarding the intracellular role of magnesium was, until recently, the technical difficulties of measuring the free or ionized concentration of the ion, for it is the free and not the total concentration that is the key physiological parameter 19.
Magnesium can bind to proteins or anions. Due to the binding, an equilibrium is established between the free magnesium and the bound magnesium. If the concentration of the total number of binding sites is (X)T and the total magnesium is (Mg)T, and K the equilibrium constant, free magnesium concentration (Mg2+) is given by the following equation19:
Thus the concentration of the free magnesium concentration in cells depends on the equilibrium constant K, as well as on the total concentrations of either magnesium or binding sites. The methods that have been used in the past to measure the ionized magnesium is measured directly with microelectrodes and should thus give the most accurate estimations of the free magnesium levels in the tissues. In the second group, the ionized magnesium is not directly determined, but is estimated from some reaction, depending on the ionized magnesium concentration.
Previously, the concentration of ionized magnesium in red cells was estimated using adenylate kinase equilibrium in intact cells and red cell lysates, since this enzyme requires magnesium as an activator20.
Then the molal concentration of ionized red blood cell magnesium has been measured with a magnesium electrode in a stroma-free, freeze-thaw haemolysate of fresh human oxygenated red blood cells with a divalent cation-specific electrode; a value of about 0.5 mM was found21.
From the numerous approaches three are now routinely used to study the intracellular free magnesium concentration and its regulation, namely: magnesium sensitive microelectrodes, fluorescence probes (Mag-Fura) and nuclear magnetic resonance spectroscopy (NMR).
Magnesium ion activities are evaluated by ion-selective electrodes. They are usually transformed into ion concentrations using a calibration procedure. This transformation, which is ultimately based on the calculation of activity coefficients, is assuming a constant ion background (constant ionic strength, I). Nevertheless it is the activity coefficient of the calibration solution which is strictly constant. This is not the case even with a biological sample. It would be more adequate to specify activities. Matrix interaction, degree of discrimination of background ions (selectivity), life time of the membrane, and dynamic response behaviour have to be taken into account to avoid systematic errors. Actually the ionophore ETH 7025 (N', N'', N'''-imino-di8, 1-octanediyl) tris (N-heptyl-N-methyl-malonamide) is the only one with a pronounced discrimination of calcium ions and a sufficiently high selectivity to monovalent ions, suitable for measurements in blood22.
In the absence of divalent cations, the fluorescence excitation spectrum of furaptra shows a maximum at 370 nm when the emission wavelength is set at 510 nm. Upon addition of magnesium or calcium, the fluorescence excitation maximum of furaptra is 335 nm23. In contrast, addition of magnesium or calcium causes a decrease in intensity but not a shift in the wavelength of the fluorescence emission maximum at 510 nm with the excitation wavelength set at 370 nm. Because of the shift in the excitation maximum upon magnesium binding, the free intracellular magnesium ((Mg2+)f) concentration measured from the fluorescence excitation spectrum of furaptra is obtained from the ratio method according to:
(Mg2+)f = KD Smin (R-Rmin)/maxXR(max-R) where R is the fluorescence intensity ratio at the wavelengths of 335 and 370 nm observed with the biological sample, Rmin and Rmax are the fluorescence intensity ratios in the absence and presence of saturating amounts of magnesium and Smin and Smax are the fluorescence intensities in the absence of magnesium and in the presence of saturating magnesium, respectively. The main advantage of the ratio method is that magnesium measurements are independent of the concentration of the fluorescence indicator used. The fluorescence intensity measured for the sample must be corrected for cell autofluorescence by subtraction of the measurements observed with unloaded cells at the wavelengths of 335 and 370 nm. The fluorescence emission spectrum is the mirror image of the absorption spectrum.
The normal cytosolic free calcium concentration in most cells is about 100-200 nM, a value 250-500 times lower than the apparent KD for calcium binding to furaptra. Thus under basal conditions, only approximately 0.2-0.4 per cent of the furaptra will be complexed with calcium whereas approximately 33-36 per cent of the furaptra will be magnesium complexed since basal Mgi is 0.5-1 mM. Furaptra in solutions which mimic the intracellular milieu has an apparent KD-Mg of 1.5 mM, an apparent KD-Ca in the range of 20-60 µM, and a PK of 5.0. Furaptra is now available from Molecular Probes (Eugene, OR, USA) under the designation Magfura-2. In spite of the limitations mentioned above, furaptra has been extensively used to measure Mgi in numerous cells: BC3HI cells, 3T3 fibroblasts, peripheral blood lymphocytes, MDCK cells, pancreatic β-cells, vascular smooth muscle line A7r5, isolated myocytes, kidney cells (MDCK), hepatocytes24.
An alternative approach is provided by treating the cells with the ionophore A23187 and Mg 2+25. Using a divalent cation ionophore A23187 the authors have investigated magnesium buffering in intact human red blood cells and have found that the fresh, oxygenated, inosine-fed cell has three main buffer systems which bind nearly 90 per cent of the total magnesium present inside the cell in physiological conditions. Net magnesium movements across the intact red cell membrane renders the cell highly permeable to magnesium, allowing rapid equilibration of magnesium ions across the membrane and with the intracellular buffers. At electrochemical equilibrium the distribution of magnesium ions is given by Mgi2, = r2 Mgo2+, where Mgi2+ and Mgo2+ are the molal concentrations of ionized magnesium in cell water and in the medium, respectively and r2 is the Donnan distribution ratio for divalent cations at equilibrium. The total magnesium content (Mgt) of the cells is the sum of the free and bound forms (Mgb) where Mgt and Mgb are the molal concentrations of total and bound magnesium in cell water. From these equations bound magnesium is equal to Mgb = Mgt - r2 Mgo2, . The total magnesium content of the cells was measured by AAS in the trichloroacetic supernatant of cell lysates. Cell water was measured as the difference between wet and dry weights of packed-cell sample corrected for extracellular fluid. The original magnesium content of the cells used in the present experiments was 3.46 ± 0.05 mmol/kg cell water. The method described here can be used to investigate the effect of metabolism and deoxygenation on the magnesium buffer curve in intact cells as well as the magnesium dependence of active sodium and calcium transport. The main disadvantages of the null-point method are the following: (i) it does not give a direct measurement of (Mg2+)i, (ii) fairly dense suspensions of cells are required, and (iii) an independent assessment of intracellular pH or, when appropriate, r2 is also required.
The concentration of ionized magnesium in the oxygenated cells was found to be 0.39 mM and was not greatly affected by changes in the composition of the medium. The concentration of ionized magnesium in deoxygenated cells showed more dependence on the composition of the medium. Values of 0.54 and 0.62 mM were found in cells incubated in TRIS- and HCO3 buffered media respectively. Only a small increase of 0.16-0.22 mM was found in the concentration of ionized magnesium when the cells were deoxygenated. This increase might be due to a change in the binding of the important magnesium chelators, 2,3-diphosphoglycerate (2,3-DPG) and ATP by haemoglobin26.
Nuclear magnetic resonance (NMR) is a spectroscopic technique that can be used to measure (Mg2+)i in intact cells. The main advantage of NMR is that it is noninvasive and therefore ionic and metabolic changes in the intracellular environment can be monitored within an essentially unperturbed living system.
The measurement of (Mg2+)i by NMR is based on the fact that the frequency separation between the α and β or β and γ-phosphates of ATP depend on the ratio of MgATP to unbound ATP. Therefore, the NMR spectrum of intracellular ATP may be used to measure the fraction of ATP that is completed to Mg2+ and, thereby, to estimate the level of (Mg2+)i in intact cells. The main disadvantage of the technique is that (Mg2+), is not measured directly but through its effects on the NMR spectra of 31P and the accuracy of the estimate of (Mg2+) crucially depends on precise knowledge of the dissociation constant for MgATP under physiological conditions 27.
The phosphorus NMR spectra of intracellular ATP in glycolysing human red blood cells maintained at 37° C in an atmosphere containing 5 per cent CO2 have been obtained and quantitated under completely aerobic and anaerobic conditions. The results showed that 84 ± 4 per cent and 78 ± 4 per cent of the total ATP are completed to magnesium in the aerobic, and anaerobic states of the cell, respectively. The intracellular concentration of free magnesium was determined to be 0.25 ± 0.07 mM in the aerobic and 0.67 ± 0.15 mM in the anaerobic state in a sample of normal erythrocytes. The magnesium concentrations calculated on a water basis were determined to be 1.02 vs 0.83 mM for ATPMg, 0.73 vs 0.80 mM for ATPMgHb, 0.44 vs 0.20 mM for glycerate-2,3-P2 9DPG) Mg complex (DPGMg), in aerobic and anaerobic conditions respectively. Since the magnesium in the red cell is largely complexed, the three-fold increase in free magnesium under fully anaerobic conditions would significantly affect the rates of enzymatic reactions28. The concentration of ionized magnesium inside red cells is not therefore constant, but continually oscillates as the cells circulate.
To avoid some of the problems with 31P-NMR, several workers have successfully used exogenous NMR magnesium indicators to determine magnesium29. The cells are suspended in a loading buffer containing acetoxymethyl ester of MG-APTRA. 19F NMR studies of 4-methyl-5fluoro-APTRA-loaded human erythrocytes indicate a basal free magnesium level of 0.25 mM. The lack of change in cytosolic free magnesium upon raising the extracellular magnesium to 6 nM is consistent with the data of others suggesting that intracellular magnesium is slow to exchange or equilibrate with extracellular magnesium.
To our knowledge, ionized magnesium content of human intact erythrocytes has been measured mainly by the NMR method or by zeropoint titration. Both microelectronics and Mag-Fura measure the activity of magnesium ions, not the concentration.
As with most of the fluorescent indicators, one needs to be concerned about changes in autofluorescence, and indicator leakage from the cell. The excitation maximum for magnesium complexed furaptra is 335 nm, thus overlapping with cell NADH fluorescence. Thus, controls should be done in cells without indicators to assure that changes in fluorescence attributed to furaptra are not significantly contributed to by changes in NADH. Also the presence of the chelator in the cytoplasm of the cell increases the cell buffering capacity for that ion. One also needs to be sure that the addition of the chelator to the cell does not alter cell function. While fluorescent probes offer the advantage of ease of use and the possibility of routine advantage of ease of use and the possibility of routine measurement, neither MagFura-2 nor Mag-Fura-5 are ideally suited for the study of intracellular magnesium regulation, because of calcium and pH sensitivity, respectively. Fluorescence is a highly sensitive technique that can detect concentrations as low as 10-8 M, as opposed to 10-4 M for NMR spectroscopy. Because of the high sensitivity of fluorescence, only 5 µM of furaptra is required for loading, as opposed to 50 µM of the 19 F NMR indicators; thus, toxicity effects are minimized.
Ionized magnesium in erythrocytes has been measured in 14 healthy controls with both 31P NMR and zero point titration 30. The concentration of ionized magnesium as estimated by NMR was significantly lower than measured by zeropoint titration, but no significant correlation could be detected between the erythrocyte ionized magnesium levels estimated with these two methods. (Mg2+)i (mean ± ISD) measured with the 31P NMR method (mmol./litre cell wt) was found to be 0.20 (0.03) and with the zeropoint titration (mmol/litre cell wt) 0.55 (0.12). The NMR and zeropoint titration methods gave different values for ionized magnesium concentration in human erythrocytes, which showed no correlation. Thus the uncertainty concerning the two currently used techniques for free magnesium determination in erythrocytes (i.e., NMR and zeropoint titration methods) is worsened by the fact that ionized magnesium inside red cells continually oscillates in vivo.
Intracellular free magnesium concentration falls to one-third of its original level after 11 days storage at 4° C in standard blood preservation media. The fall occurs in spite of little change in the total magnesium content of the cells, indicating a change in the buffering characteristics of the cytoplasm 31. Binding of magnesium to ligands other than ATP and 2,3-bisphosphoglycerate must increase during storage. Storage did not significantly affect basal or sodium-stimulated efflux of magnesium from magnesium-loaded red cells. During storage the degree of magnesium binding attributable to ATP and DPG declines. Calcium but not magnesium permeability of human red cells increases during cold storage 32.
In order to study magnesium homoeostasis, it is necessary to know the ionized as well as the total magnesium concentration since it is ionized magnesium that is regulated by transport systems. Plasma ionized magnesium concentration is about 0.5 mM and the membrane potential in erythrocytes is about 9 mv, inside negative. These values predict that intracellular ionized magnesium concentration should be about 1 mM at electrochemical equilibrium. In fact measured values in red cells are all well below this level. Analysis of the ionized intracellular magnesium concentration (Mg2+)i also shows whether or not magnesium is at equilibrium across the membrane. Red cell magnesium permeability appears to be very low. It has been suggested in human studies that any 28Mg entering the cells did so during early phases of erythropoeisis. The observation that magnesium can move across the human red cell membrane prompted careful exploration of the fluxes from cells with normal magnesium content into media containing sodium but nominally free of magnesium. Fluxes ranging from about 4-7 µmol/litre/cell/h were detected after correction for haemolysis. Although these fluxes are small, they are much higher than would be expected from passive diffusion through the lipid of the membrane33.
Intracellular free and bound magnesium represent a magnesium buffer system. If the capacity of the intracellular magnesium buffer is exhausted, constant pMgi is supported by net magnesium influx or net magnesium efflux across the plasma membrane. Moreover, magnesium uptake may occur without significant changes in free magnesium because of buffering of intracellular magnesium34.
When rat or chicken erythrocytes are loaded with magnesium by incubation of the cells at increased (Mg2+)o in the presence of the divalent cation ionophore A23187, the erythrocytes take up magnesium. Magnesium efflux took place only when (Mg2+)i was increased and stopped when the physiological magnesium content was reached. Magnesium efflux was energy-dependent and reduced by reduction of intracellular ATP in magnesium-loaded human erythrocytes. In vitro, magnesium efflux was specifically dependent on (Na+). Substitution of (Na+)o by K+, L+ or Choline+ could not support magnesium efflux. Magnesium efflux was stoichiometrically coupled with the uptake of extracellular sodium: with human erythrocytes a ratio of 3 sodium to 1 magnesium as reported, in analogy to the Na+/Ca2+ antiport at the cell membrane. Magnesium efflux was irreversible. When erythrocytes were magnesium depleted, loaded with sodium and incubated at elevated (Mg2+)o in the presence of ATP, there was no magnesium uptake. Irreversibility of Na+/Mg2+ antiport only holds for net magnesium efflux.
In chicken erythrocytes 28Mg2+ uptake occurs in exchange for non-radioactive intracellular 24Mg2+. The 28Mg2+ - 24Mg2+ exchange in loaded and non-loaded cells showed the same inhibition by amiloride and (Na+)o. Therefore it can be concluded that in magnesium-loaded and non-loaded cells the magnesium efflux system exchanges intracellular for extracellular magnesium. In this exchange system extracellular sodium competes with extracellular magnesium. After magnesium loading the magnesium exchanger has changed its properties, leading to Na+/Mg2+ antiport. This effect is an analogy to the function of Na+/H+ antiport which becomes active at increased (H+)i. It can be concluded that in the presence of high (Mg2+)i, a higher portion of Mg2+ -Mg2+ exchange occurs simultaneously with Na+/Mg2+. In rat erythrocytes (Na+)i competes with (Mg2+), for Na+/Mg2+ antiport. All these results were obtained from rat or chicken erythrocyte experiments. In chicken, rat, and human erythrocytes Na+/Mg2+ antiport was ATP-dependent.
In magnesium-loaded human erythrocytes a major fraction of magnesium efflux occurred in sodium-free potassium medium. Sodium-independent magnesium efflux from magnesium-loaded human, rat and chicken erythrocytes was the highest in sucrose medium and reduced to the same degree in KCl, LiCl and choline chloride. Sodium-independent magnesium efflux was reduced by high (Cl)o and by (4,4'-diisothiocyanatostilbene-2,2-disulphonicacid) DIDS.
From these results it was concluded that sodium-independent magnesium efflux functions in combination with net Cl- efflux for charge compensation. Magnesium efflux from human magnesium-loaded erythrocytes was found (mmol/litre cells*30 min) in sucrose (sodium-independent 0.89, in choline chloride (sodium-independent) 0.33 and in NaCl-choline chloride (sodium-dependent) 0.16). Sodium-dependent magnesium efflux from human, rat and chicken erythrocytes can be partly inhibited by serum albumin34. In experiments with magnesium-depleted erythrocytes, no significant reuptake of magnesium was found. Only when reticulocytes were magnesium-depleted could reuptake of magnesium be observed. In erythrocytes there is an additional magnesium efflux operating in combination with Cl efflux for charge compensation.
Major differences in erythrocyte (Er) magnesium content have been shown to exist between different ethnic groups35. Mean erythrocyte magnesium values were 20-30 per cent lower in black African men than in Amerindian Quechuas of the same sex and age. European Caucasians exhibited intermediate values. These differences appeared to be relatively stable and partly independent of the climate in which the subjects were residing. In a given population, large differences between subjects were also observed. Healthy adult male blood donors examined in Paris revealed a 65 per cent difference between the lowest value (1.55 mmol/litre of erythrocytes) and the highest one (2.89 mmol). Large and fairly stable interethnic and interindividual variations led Henrotte35 to propose the hypothesis of a genetic control of the erythrocyte magnesium level in humans. However, the relative contribution of genetic vs environmental factors was a priori difficult to assess. The existence of a genetic regulation of erythrocyte magnesium content has now been fully confirmed. By comparing monozygotic twins living together and apart, Darin et al.36 were able to discriminate the genetic component from the familial non-genetic component. Environmental factors that begin to differ after twins have been separated, had some influence on the plasma magnesium and none (or very little) on the erythrocyte magnesium values. There is a very significant association between erythrocyte magnesium and HLA-B antigens. The largest associated variations with HLA-B antigens (between B35+ Bw6+ and B38+ subjects) are equal to 10 per cent of the total population erythrocyte magnesium mean. The association between erythrocyte magnesium and HLA is much stronger in men than in women, suggesting the interaction of sex-linked factors; one of these factors could be the larger variability of erythrocyte magnesium in women with age, menstrual cycle and oral contraceptive intake.
Thus, erythrocyte magnesium level is controlled by genetic factors. In erythrocytes, other phenomena may hamper the interpretation of results. Erythrocyte magnesium levels decrease with cell age: high levels in reticulocytes decrease and reach a steady state in mature erythrocytes and decrease again in older cells. Genetically low erythrocyte magnesium values could, therefore, be ascribed to the older age of the erythrocyte population. The interpretation of the genetic regulation of cell magnesium is also hindered by the fact that the total intracellular magnesium level represents both bound and free magnesium contents. A statistically significant correlation has been found between free and total magnesium contents37.
A complex neuro-hormonal system is instrumental in the control of the stability of intracellular magnesium and of the consequences of magnesium status disturbances. Exchanges between extracellular compartments and soft tissues elicit secretion of hormones and neurohormones: adrenalin and insulin, taurine (and perhaps glutamyltaurine), and finally beta-stimulation of the adrenergic receptors.
These neuro-hormonal factors regulating cellular magnesium content are very important to appreciate the significance of the variations of erythrocyte magnesium concentrations in case of several pathological or iatrogenic disturbances such as diabetes mellitus, phaeochromocytoma or stress pathology38-40.
It should be emphasized that variations in the concentration of total or ionized magnesium in a given cell do not necessarily mean that similar changes should exist in other cellular concentrations or in the magnesium pool. Changes may occur in the course of disturbances in the distribution of the ion and constitute evidences of functional or organic disorders in the analysed cell.
It remains that the basic cellular magnesium item for evaluating the magnesium status is the static measurement of erythrocyte magnesium which once again has been experimentally and clinically validated in the investigation of primary chronic marginal magnesium deficiency. Its value will be subsequently enhanced when it takes its place among the clinical and paraclinical data of dynamic investigations41-44.
1. Martin M.T. & Shapiro, R. (1988): Atomic absorption spectrometry of magnesium. Methods Enzymol. 158, 365-370.
2. Sturgeon, R.E. & Berman, S.S. (1983): Determination of the efficiency of the graphite furnace for atomic absorption spectrometry. Anal. Chem. 55, 190-200.
3. Harrison, W.W. & Wadlin, W.H. (1969): Magnesium spinell interferences in air-acetylenevs nitrous oxide-acetylene flames in atomic absorption spectrometry.
4. Elwell, W.T. & Gidley, A.F. (1961): In: Atomic absorption spectrophotometry, pp. 67-75. New York: Macmillan.
5. Ramakrishna, T.V., Robinson, J.W. & West, P.W. (1966): The determination of calcium and magnesium by atomic absorption spectroscopy. Anal. Chim. Acta 36, 57-64.
6. Halls, D.J. & Townshend, A. (1966): A study of some interferences in the atomic absorption spectrophotometry of magnesium. Anal. Chim. Acta 36, 273-278.
7. Willis, J.B. (1960): The determination of metal in blood serum by atomic absorption spectroscopy. II. Spectrochim Acta 16, 273-278.
8. Price, W.J. (1972): In: Analytical atomic absorption spectrometry, pp. 163-166. New York: I Heyden.
9. Stendig-Lindberg, G., Penciner, J., Rudy, N. & Wacker, W.E.C. (1984): Comparison of diluents for serum magnesium. Magnesium 3, 50-56.
10. Lijnen, P., Hespel, P., Lommelen, G., Laermans, M., M'Buyamba-Kabangu, J.R. & Amery, A. (1986): Intracellular sodium, potassium and magnesium concentration, ouabain-sensitive ribidium-uptake and sodium-efflux and Na+,K+ cotransport activity in erythrocytes of normal male subjects studied on two occasions. Meth. Find. Exptl. Clin. Pharmacol. 8, 525-533.
11. Fisher, P.W.F., Belonje, B. & Giroux, A. (1993): Magnesium status and excretion in age-matched subjects with normal and elevated blood pressure. Clin. Biochem 26, 207-211.
12. Cameron, I.L., Hardman, W.E., Smith, N.K.R., Fullerton, G.D. & Miseta, A. (1993): Changes in the concentration of ions during senescence of the human erythrocyte. Cell. Biol. Int. 17, 93-98.
13. Tongyai, W., Rayssiguier, Y., Motta, C., Gueux, E., Maurois, P. & Heaton, F.W. (1989): Mechanism of increased erythrocyte membrane fluidity during magnesium deficiency in weanling rats. Am. J. Physiol. 257, C270-C276.
14. Rousselet, F. & Durlach, J. (1971): Méthodes analytiques et exploration pratique du metabolisms du magnesium en clinique
humaine. Symposium on magnesium deficit in human pathology, Vol. 1, p. 65-90. Vittel.
15. Duester, P.A., Trostman, U.H., Bernier, L.L. & Dolev, E. (1987): Indirect vs direct measurement of magnesium and zinc in erythrocytes. Clin. Chem. 33/34, 529-532.
16. Bonnay, M.M., Loegac, E.M., Brunier, A.P. & Bohuon, C.J. (1988): Intra-erythrocytic Mg assay in the Kodak Ektachem 700 analyser. Clin. Chem. 34, 2154-2155.
17. Moretto, P., Llabador, Y., Simonoff, M., Bara, M., Guiet-Bara, A. & Durlach, J. (1993): The nuclear microprobe: a powerful microanalysis technique for magnesium and trace element research. In: Health and disease (role of magnesium and trace minerals), eds. R. Nath & K.D. Gill, pp. 93-101, New Delhi: Ashish Publishing.
18. Schuette, S., Vereault, D., Ting B.T.G. & Janghorbani, M. (1988): Accurate measurement of stable isotopes of magnesium in biological materials with inductively coupled plasma mass spectroscopy. Analyst 113, 1837-1842.
19. McGuigan, J.A.S., Buri, A., Chen, S., Illner, H. & Lilthi, S. (1993): Some theoretical and practical aspects of the measurement of the intracellular free magnesium concentration in heart muscle: consideration of its regulation and modulation. Magnesium and the cell. ed. N.J. Birch, pp. 91-120. London: Academic Press,
20. Rose, I.A. (1968): The state of magnesium in cells as estimated from the adenylate kinase equilibrium. Proc. Natl. Acad. Set. USA 61, 1079-1086.
21. Berger, H., Jänig, G.R., Gerber, G., Ruckpaul, K. & Rapoport, S.M. (1973): Interaction of haemoglobin with ions, Eur. J. Biochem. 38, 553-562.
22. Spichiger, U.E., Eugster, R., Schaller, U., Haase, E. & Simon, W. (1991): Magnesium-selective electrodes: state of the art. Magnes. Bull. 13, 140-144.
23. Mota de Freitas, D. & Dorus, D. (1993): Techniques for measuring magnesium in tissues from hypertensive, psychiatric and neurological patients. In: Magnesium and the cell ed. N.J. Birch, pp. 51-79. London: Academic Press.
24. Murphy, E. (1993): Measurement of intracellular ionized magnesium. Miner. Electrolyte Metab. 250-258.
25. Flatman, P. & Lew, V.L. (1977): Use of ionophore A23187 to measure and control free and bound cytoplasmic Mg in intact red cells. Nature 267, 360-362,
26. Flatman, P.W. (1980): The effect of buffer composition and deoxygenation on the concentration of ionized magnesium inside human red blood cells. J. Physiol. 300, 19-30.
27. Alvarez-Leefmans, F.J., Giraldez, F. & Gamino, S.M. (1987): Intracellular free magnesium in excitable cells: its measurement and its biologic significance. Can. J. Physiol. Pharmacol. 65, 915-925.
28. Gupta, R.K., Benovic, J.L. & Rose, Z.B. (1978): The determination of the free magnesium level in the human red blood cell by 31P NMR. J. Biol. Chem. 253, 6172-6176.
29. Levy, L.A., Murphy, E., Raju, B. & London, R.E. (1988): Measurement of cytosolic free magnesium ion concentration by 19F NMR. Biochemistry 27, 4041-4048.
30. Geven, W.B., Vogels-Mentink, G.M., Willems, J.L., Hilbers, C.V., Os, C.V. & Monens, L.A.H. (1991): Ionized Mg content of human erythrocytes measured by the NMR method compared with values obtained by zeropoint titration. In: Magnesium- a relevant ion, eds. B. Lasserre & J. Durlach, pp. 291-293. London: John Libbey.
31. Flatman, P.W. (1988): The control of red cell magnesium. Magnes. Res. 1, 5-11.
32. Bock, J. L. & Yusuf, Y. (1988): Further studies on alterations in magnesium binding during cold storage of erythrocytes. Biochem. Biophys. Acta 941, 225-231.
33. Flatman, P.W. (1991): Mechanisms of magnesium transport. Annu. Rev. Physiol. 53, 259-271.
34. Vormann, J. & Günther, T. (1993): Magnesium transport mechanisms. In: Magnesium and the cell, ed. N.J. Birch, pp. 137-155. London: Academic Press.
35. Henrotte, J.G. (1993): Genetic regulation of cellular magnesium content. In: Magnesium and the cell, ed. N.J. Birch, pp. 137-155. London: Academic Press.
36. Darin, P., Michotte, Y., Defrise-Gussenhoven, F. & Henrotte, J.G. (1981): The inheritance of plasma and red blood cell magnesium and zinc levels studied from twin and family. Acta. Genet. Med. Gemell. 30, 67-50.
37. Santarromana, M., Delepierre, M., Ferray, J.C., Franck, G., Garay, R. & Henrotte, J.C. (1989): Correlation between total and free magnesium levels in human red blood cells. lnfluence of HLA antigens. Magnes. Res. 2, 281-283.
38. Durlach, J. (1988): Regulation of cellular magnesium. In: Magnesium in clinical practice, ed.. J. Durlach, pp, 31-40. London: John Libbey.
39. Durlach, J. (1988): Secondary magnesium deficits; magnesium overload. In: Magnesium in clinical practice, ed. J. Durlach, pp. 114-215. London: John Libbey.
40. Classen, H. G. (1990): Systemic stress and the role of magnesium. In: Metal ions in biological systems, vol. 26. Compendium on magnesium , eds. H. Sigel & A. Sigel, pp. 321-329. New York: Marcel Dekker.
41. Durlach, J. (1988): Methods of evaluating magnesium status. In: Magnesium in clinical practice, ed. J. Durlach, pp. 40-60. London: John Libbey.
42. Durlach, V., Millart, H., Meer, L., Grulet. H,, Gross, A. & Leutenegger, M. (1991): Magnesium status in a group of so-called 'spasmophilic patients'. Magnes. Res. 4, 233--234.
43. Nowitzki-Grimm, S., Grimm, P., Thoni, H. & Classen, H.G. (1991): Diagnosis of magnesium, status. Magnes. Bull. 13, 107-115.
44. Beretta, P., Ambrosini, G., Concari, M. & Bargellini, A. (1993): Is magnesium content in erythrocytes suitable for evaluating cation retention after oral physiological supplementation in marginally magnesium deficient subjects. Magnes. Res. 6, 149-153.
All articles by Dr. Durlach are copyrighted, and permission is granted to Web users only to make single hard copies for personal use. Additional reprints should be obtained from the originating journals. Excerpts may be used by the media with attribution to Dr. Durlach.
This page was first uploaded to The Magnesium Web Site on April 29, 1996