ABSTRACT:We used 7Li N M R spin-lattice (TI) and spin-spin (Tz) relaxation time measurements to investigate the binding of Li+ in human red blood cell (RBC) suspensions. In RBCs containing 1.4 mM Li+, the intracellular 7Li N M R T2 relaxation value (0.30 f 0.03 s) was much smaller than the corresponding TI value (6.0 f 0.1 s), yielding a ratio of TI to T 2 of 20. For 1.5 mM LiCl solutions whose viscosities were adjusted to 5 CP with glycerol, the values of the T1/T2 ratios were as follows: 49 for unsealed RBC membrane (2.0 mg of protein/mL); 4.4 for spectrin (1.9 mg/mL); 1.5 for 5.4 mM 2,3-bisphosphoglycerate (BPG); 2.2 for 2.7 mM carbonmonoxyhemoglobin (COHb); 1.6 for 2.0 mM ATP; and 1.2 for a 50/50% (v/v) glycerol-water mixture. Intracellular viscosity and the electric field gradients experienced by Li+ when traversing the spectrin-actin network therefore are not responsible for the large values of the Tl/ T2 ratios observed in Li+-loaded RBCs. We conclude that the RBC membrane is the major Li+ binding site in Li+-loaded RBCs (KI, = 215 f 36 M-l) and that the binding of Li+ to COHb, BPG, spectrin-actin, or ATP is weak. Partially relaxed 7Li N M R spectra of Li+-containing RBC membrane suspensions indicated the presence of two relaxation components, one broad and one narrow. At the same extravesicular Li+ and protein concentrations, the TI values for right-side-out RBC vesicle suspensions were at least 2-fold larger than those for inside-out RBC vesicle suspensions; the inner layer of the RBC membrane, which has a larger percentage of anionic phospholipids than the outer layer, contributes mostly to Li+ binding.

We investigated the mechanism of competition between Li+ and Mg2+ in Li(+)-loaded human red blood cells (RBCs) by making 7Li and 31P NMR and fluorescence measurements. We used 7Li NMR relaxation times to probe Li+ binding to the human RBC membrane and ATP; an increase in Mg2+ concentration caused an increase in both 7Li T1 and T2 values in packed Li(+)-loaded RBCs, in suspensions of Li(+)-loaded RBC ghosts, in suspensions of Li(+)-containing RBC membrane, and in aqueous solutions of ATP, indicating competition between Li+ and Mg2+ for binding sites in the membrane and ATP. We found that increasing concentrations of either Li+ or Mg2+ in the presence of human RBC membrane caused an increase in the 31P NMR chemical shift anisotropy parameter, which describes the observed axially symmetric powder pattern, indicating metal ion binding to the phosphate groups in the membrane. Competition between Li+ and Mg2+ for phosphate groups in ATP and in the RBC membrane was also observed by both fluorescence measurements and 31P NMR spectroscopy at low temperature. The ratio of the stoichiometric binding constants of Mg2+ to Li+ to the RBC membrane was approximately 20; the ratio of the conditional binding constants in the presence of a free intracellular ATP concentration of 0.2 mM was approximately 4, indicating that Li+ competes for approximately 20% of the Mg(2+)-binding sites in the RBC membrane. Our results indicate that, regardless of the spectroscopic method used, Li+ competes with Mg2+ for phosphate groups in both ATP and the RBC membrane; the extent of metal ion competition for the phosphate head groups of the phospholipids in the RBC membrane is enhanced by the presence of ATP. Competition between Li+ and Mg2+ for anionic phospholipids or Mg(2+)-activated proteins present in cell membranes may constitute the basis of a general molecular mechanism for Li+ action in human tissues.

Because Mg2+ and Li+ ions have similar chemical properties, we have hypothesized that Li+/Mg2+ competition for Mg2+ binding sites is the molecular basis for the therapeutic action of lithium in manic-depressive illness. By fluorescence spectroscopy with furaptra-loaded cells, the free intracellular Mg2+ concentration within the intact neuroblastoma cells was found to increase from 0. 39 +/- 0.04 mM to 0.60 +/- 0.04 mM during a 40-min Li+ incubation in which the total intracellular Li+ concentration increased from 0 to 5.5 mM. Our fluorescence microscopy observations of Li+-free and Li+-loaded cells also indicate an increase in free Mg2+ concentration upon Li+ incubation. By 31P NMR, the free intracellular Mg2+ concentrations for Li+-free cells was 0.35 +/- 0. 03 mM and 0.80 +/- 0.04 mM for Li+-loaded cells (final total intracellular Li+ concentration of 16 mM). If a Li+/Mg2+ competition mechanism is present in neuroblastoma cells, an increase in the total intracellular Li+ concentration is expected to result in an increase in the free intracellular Mg2+ concentration, because Li+ displaces Mg2+ from its binding sites within the nerve cell. The fluorescence spectroscopy, fluorescence microscopy, and 31P NMR spectroscopy studies presented here have shown this to be the case.

Lithium has been used clinically in the treatment of manic depression. However, its pharmacologic mode of action remains unclear. Characteristics of Li ± interactions in red blood cells (RBCs) have been identified. We investigated Li + interactions on human neuroblastoma SH-SY5Y cells by developing a novel 7Li NMR method that provided a clear estimation of the intra-and extracellular amounts of LV in the presence of the shift reagent thulium-1,4,7,10tetrazacyclododecane-N , N', N", N"-tetramethylene phosphonate (HTmDOTP4). The first-order rate constants of Li + influx and efflux for perfused, agarose-embedded SH-SY5Y cells in the presence of 3 mM HTmDOTP4 were 0.055 ±0.006 (n = 4) and -0.025 ±0.006 min~1(n = 3), respectively. Significant increases in the rate constants of Li + influx and efflux in the presence of 0.05 mM veratridine indicated the presence of Nachannel-mediated Litransport in SH-SY5Y cells. 7Li NMR relaxation measurements showed that Li~is immobilized more in human neuroblastoma SH-SY5Y cells than in human RBCs. A 0, starting area of the intracellular resonance of Li * -loaded cells in Li + efflux measurements; A~, intracellular resonance area at time t; AA, atomic absorption; DIDS, 4,4 '-diisothiocyanatostilbene-2,2 '-disulfonic acid; DMEM, Dulbecco's modified Eagle's medium; HTmDOTP 4~, thulium-1,4,7,1 0-tetrazacyclododecane-N, N', N", N"-tetramethylenephosphonate; PCr, phosphocreatine; P,, inorganic phosphorus; ppm, parts per million; RBC, red blood cell; T,, spin-lattice relaxation value; T 2, spin-spin relaxation value; v0, initial rate of Li * influx.

Li þ /Mg 2þ competition has been implicated in the therapeutic action of Li þ treatment in bipolar illness. We hypothesized that this competition depended on cell-specific properties. To test this hypothesis, we determined the degree of Li þ transport, immobilization, and Li þ /Mg 2þ competition in lymphoblastomas, neuroblastomas, and erythrocytes. During a 50 mmol/L Li þ -loading incubation, Li þ accumulation at 30 min (mmoles Li þ /L cells) was the greatest in lymphoblastomas (11:1 AE 0:3), followed by neuroblastomas (9:3 AE 0:5), and then erythrocytes (4:0 AE 0:5). Li þ binding affinities to the plasma membrane in all three cell types were of the same order of magnitude; however, Li þ immobilization in intact cells was greatest in neuroblastomas and least in erythrocytes. When cells were loaded for 30 min in a 50 mmol/L Li þ -containing medium, the percentage increase in free intracellular [Mg 2þ ] in neuroblastoma and lymphoblastoma cells ($55 and $52%, respectively) was similar, but erythrocytes did not exhibit any substantial increase ($6%). With the intracellular [Li þ ] at 15 mmol/L, the free intracellular [Mg 2þ ] increased by the greatest amount in neuroblastomas ($158%), followed by lymphoblastomas ($75%), and then erythrocytes ($50%). We conclude that Li þ immobilization and transport are related to free intracellular [Mg 2þ ] and to the extent of Li þ /Mg 2þ competition in a cell-specific manner. #