Characterization of the phosphatidylserine-exposing subpopulation of sickle cells.

Phosphatidylserine (PS), exclusively present in the inner monolayer of the normal red blood cell (RBC) membrane, is exposed in subpopulations of sickle cells. PS-exposing RBCs were found predominantly among the densest and the very light sickle cells. Within the light RBC fraction, PS exposure was found on reticulocytes, transferrin receptor-expressing reticulocytes, and mature RBCs. The last subset contained low-density valinomycin-resistant RBCs, previously shown to have high Na(+) and low K(+) content. This subpopulation contained the highest percentage of PS-exposing cells. The PS-exposing sickle cells did not show the sustained high cytosolic Ca(++) levels that have been shown to activate scramblase activity. Data from this study indicate that PS exposure can occur at different stages in the life of the sickle RBC and that it correlates with the loss of aminophospholipid translocase activity, the only common denominator of the PS-exposing cells. The additional requirement of scramblase activation may occur during transient increases in cytosolic Ca(++). (Blood. 2001;98:860-867)

Normal Ca2+ extrusion by the Ca2+ pump of intact red blood cells exposed to high glucose concentrations.

The ATPase activity of the plasma membrane Ca2+ pump (PMCA) has been reported to be inhibited by exposure of red blood cell (RBC) PMCA preparations to high glucose concentrations. It has been claimed that this effect could have potential pathophysiological relevance in diabetes. To ascertain whether high glucose levels also affect PMCA transport function in intact RBCs, Ca2+ extrusion by the Ca2+-saturated pump [PMCA maximal velocity (V(max))] was measured in human and rat RBCs exposed to high glucose in vivo or in vitro. Preincubation of normal human RBCs in 30-100 mM glucose for up to 6 h had no effect on PMCA V(max). The mean V(max) of RBCs from 15 diabetic subjects of 12.9 +/- 0.7 mmol. 340 g Hb(-1). h(-1) was not significantly different from that of controls (14.3 +/- 0.5 mmol. 340 g Hb(-1). h(-1)). Similarly, the PMCA V(max) of RBCs from 11 streptozotocin-diabetic rats was not affected by plasma glucose levels more than three times normal for 6-8 wk. Thus exposure to high glucose concentrations does not affect the ability of intact RBCs to extrude Ca2+.

Identification and characterization of a newly recognized population of high-Na+, low-K+, low-density sickle and normal red cells.

We describe a population of sickle cell anemia red cells (SS RBCs) ( approximately 4%) and a smaller fraction of normal RBCs (<0.03%) that fail to dehydrate when permeabilized to K(+) with either valinomycin or elevated internal Ca(2+). The nonshrinking, valinomycin-resistant (val-res) fractions, first detected by flow cytometry of density-fractionated SS RBCs, constituted up to 60% of the lightest, reticulocyte-rich (R1) cell fraction, and progressively smaller portions of the slightly denser R2 cells and discocytes. R1 val-res RBCs had a mean cell hemoglobin concentration of approximately 21 g of Hb per dl, and many had an elongated shape like "irreversibly sickled cells," suggesting a dense SS cell origin. Of three possible explanations for val-res cells, failure of valinomycin to K(+)-permeabilize the cells, low co-ion permeability, or reduced driving K(+) gradient, the latter proved responsible: Both SS and normal val-res RBCs were consistently high-Na(+) and low-K(+), even when processed entirely in Na-free media. Ca(2+) + A23187-induced K(+)-permeabilization of SS R1 fractions revealed a similar fraction of cal-res cells, whose (86)Rb uptake showed both high Na/K pump and leak fluxes. val-res/cal-res RBCs might represent either a distinct erythroid genealogy, or an "end-stage" of normal and SS RBCs. This paper focuses on the discovery, basic characterization, and exclusion of artifactual origin of this RBC fraction. Many future studies will be needed to clarify their mechanism of generation and full pathophysiological significance.

Polymer structure and solubility of deoxyhemoglobin S in the presence of high concentrations of volume-excluding 70-kDa dextran. Effects of non-s hemoglobins and inhibitors.

Earlier observations indicated that volume exclusion by admixed non-hemoglobin macromolecules lowered the polymer solubility ("Csat") of deoxyhemoglobin (Hb) S, presumably by increasing its activity. In view of the potential usefulness of these observations for in vitro studies of sickling-related polymerization, we examined the ultrastructure, solubility behavior, and phase distributions of deoxygenated mixtures of Hb S with 70-kDa dextran, a relatively inert, low ionic strength space-filling macromolecule. Increasing admixture of dextran progressively lowered the Csat of deoxyHb S. With 12 g/dl dextran, a 5-fold decrease in apparent Csat ("dextran-Csat") was obtained together with acceptable sensitivity and proportionality with the standard Csat when assessing the effects of non-S Hb admixtures (A, C, and F) or polymerization inhibitors (alkylureas or phenylalanine). The volume fraction of dextran excluding Hb was 70-75% of total deoxyHb-dextran (12 g/dl) volumes. Electron microscopy showed polymer fibers and fiber-to-crystal transitions indistinguishable from those formed without dextran. Thus when Hb quantities are limited, as with genetically engineered recombinant Hbs or transgenic sickle mice, the dextran-Csat provides convenient and reliable screening of effects of Hb S modifications on polymerization under near-physiological conditions, avoiding problems of high ionic strength.

Stochastic nature and red cell population distribution of the sickling-induced Ca2+ permeability.

To explore basic properties of the sickling-induced cation permeability pathway, the Ca2+ component (Psickle-Ca) was studied in density-fractionated sickle cell anemia (SS) discocytes through its effects on the activity of the cells' Ca2+sensitive K+-channels (KCa). The instant state of KCa channel activation was monitored during continuous or cyclic deoxygenation of the cells using a novel thiocyanate-densecell formation method. Each deoxy pulse caused a reversible, sustained Psickle-Ca, which activated KCa channels in only 10-45% of cells at physiological [Ca2+]o ("activated cells"). After removal of cells activated by each previous deoxy pulse, subsequent pulses generated similar activated cell fractions, indicating a random determination rather than the response of a specific vulnerable subpopulation. The fraction of activated cells rose monotonically with [Ca2+]o along a curve reflecting the cells' distribution of Psickle-Ca, with values high enough in a small cell fraction to trigger near-maximal KCa channels. Consistent with the stochastic nature of Psickle-Ca, repeated deoxygenated-oxygenated pulsing led to progressive dense cell formation, whereas single long pulses caused one early density shift. Thus deoxygenation-induced Ca2+-permeabilization in SS cells is a probabilistic event with large cumulative dehydrating potential. The possible molecular nature of Psickle-Ca is discussed.

HbC compound heterozygotes [HbC/Hb Riyadh and HbC/Hb N-Baltimore] with opposing effects upon HbC crystallization.

Compound heterozygotes of variant haemoglobins (Hbs) with HbC, with or without novel phenotypic changes, have provided insight into the molecular basis of the interacting haemoglobins and information concerning the role of specific residues in the crystallization of oxy HbC. A high phosphate buffer system has proved useful for studying the effects of variant haemoglobins (naturally co-existing with HbC in the red cell) on the oxy HbC crystallization process and has led us to conclude that beta87 and beta73 are contact sites of the oxy HbC crystal. We now present investigations from two HbC compound heterozygotes which exhibit opposing effects upon HbC crystallization: HbC/Hb N-Baltimore (beta95 Lys-->Glu) and HbC/Hb Riyadh (beta120 Lys-->Asn). The latter inhibits the in vitro crystallization of HbC, explaining the lack of erythrocyte abnormalities (with the exception of microcytosis) in the doubly heterozygous infant. In contrast, Hb N-Baltimore accelerates the crystallization of HbC, contributing to multiple abnormalities in red cell morphology, albeit in the absence of morbidity. We conclude that (1) beta120 and beta95 are additional contact sites in the crystal, and (2) the HbC/Hb Riyadh haemoglobinopathy demonstrates that crystallization may not be required for the generation of the observed microcytosis and increased red cell density in HbC-containing red cells.

Measurement of the distribution of anion exchange function in normal human red cells.

1. The aim of the present work was to investigate cell-to-cell variation in anion exchange turnover in normal human red cells. Red cells permeabilized to protons and K+ dehydrate extremely rapidly by processes that are rate-limited by the induced K+ permeability or by anion exchange turnover. Conditions were designed to render dehydration rate-limited by anion exchange turnover. Cell-to-cell variation in anion exchange function could then be measured from the distribution of delay times required for dehydrating cells to attain resistance to haemolysis in a selected hypotonic medium. 2. Red cells were suspended at 10% haematocrit in a low-K+ solution and, after a brief preincubation with 20 microM SITS at 4 degrees C, were warmed to 24 degrees C, and the protonophore CCCP was added (20 microM) followed 2 min later by valinomycin (60 microM). Delay times for cells to become resistant to lysis were measured from the instant of valinomycin addition by sampling suspension aliquots into thirty volumes of 35 mM NaCl. After centrifugation the per cent lysis was estimated by measuring the haemoglobin concentration in the supernatant. Typical median delay times with this standardized method were 4-5 min. 3. The statistical parameters of the delay time distributions report the population spread in the transport function that was limiting to dehydration. In the absence of SITS and CCCP, dehydration was limited by the diffusional Cl- permeability (PCl). Delay time distributions for PCl- and anion exchange-limited dehydration were measured in red cells from three normal donors. For both distributions, the coefficients of variation ranged between 13.0 and 15.2%, indicating a high degree of uniformity in PCl and anion exchange function among individual red cells.

Pathophysiology of sickle cell anemia.

The anemia results from the markedly shortened circulatory survival of SS cells, together with a limited erythropoietic response. Both independent properties of Hb S-polymerization of the deoxy-Hb and instability of the oxy-Hb-contribute to early red cell destruction by effects on the Hb and on the red cell membranes. The erythroid response is limited mainly by the low oxygen affinity of SS cells, caused by the polymer and the increased 2,3-DPG. But the worst culprits in these processes are the dense, dehydrated SS cells (including the ISCs), most of which are formed rapidly from non-Hb F-reticulocytes by cation transport mechanisms triggered by polymerization. Since the clinical consequences of microvascular occlusion far exceed those of anemia per se, measures to lessen the anemia must also inhibit polymerization and sickling.

A recombinant sickle hemoglobin triple mutant with independent inhibitory effects on polymerization.

As part of a comprehensive effort to map the most important regions of sickle hemoglobin that are involved in polymerization, we have determined whether two sites previously shown to be involved, Leu-88(beta) and Lys-95(beta), had additive effects when substituted. The former site is part of the hydrophobic pocket that binds Val-6(beta), the natural mutation of HbS, and the latter site is a prominent part of the hemoglobin exterior. A sickle hemoglobin triple mutant with three amino acid substitutions on the beta-chain, E6V/L88A/K95I, has been expressed in yeast and characterized extensively. Its oxygen binding curve, cooperativity, response to allosteric effectors, and the alkaline Bohr effect showed that it was completely functional. The polymer solubility of the deoxy triple mutant, measured by a new micromethod requiring reduced amounts of hemoglobin, was identical to that of the E6V(beta)/K95I(beta) mutant, i.e. when the K95I(beta) substitution was present on the same tetramer together with the naturally occurring E6V(beta) substitution, the L88A(beta) replacement had no additive effect on polymer inhibition. The results suggest that Lys-95(beta) on the surface of the tetramer and its complementary binding region on the adjoining tetramer are potential targets for the design of an effective antisickling agent.

K(86Rb) transport heterogeneity in the low-density fraction of sickle cell anemia red blood cells.

Previous studies have suggested ion transport heterogeneity among sickle cell anemia (SS) reticulocytes that could influence their dehydration susceptibility. We examined Ca2(+)-independent K transport in the lowest density (F1), reticulocyte-rich SS cells, measuring the effects of acidification, ouabain, and bumetanide on their unidirectional K(86Rb) fluxes. Unlike those of normal red blood cells and SS discocytes, the SS-F1 K(86Rb) fluxes were highly nonlinear, with large 5-min flux components (previously unobserved) and a more gradual decline over 60 min. Analysis revealed two distinct K pools: a rapid-turnover pool in a small fraction of cells, whose major ouabain-resistant K(86Rb) transport path showed distinctive properties including inhibition by high concentrations of bumetanide (> or = 1 mM) and stimulation at pH 7.0, and another heterogeneous, relatively slow-turnover pool, in most of the F1 cells, whose main ouabain-resistant K(86Rb) path was insensitive to bumetanide but was stimulated at pH 7.0, which is consistent with heterogeneous expression of the acid-sensitive K-Cl cotransport and with both rapid and slower generation of dehydrated SS cells.

Distribution of chloride permeabilities in normal human red cells.

1. The rate of dehydration of K+ permeabilized red cells is influenced by their Cl- permeability (PCl). In instances of pathological K+ permeabilization, cell-to-cell differences in PCl may determine which red cells dehydrate most. The present study was designed to investigate whether PCl differed significantly among red cells from a single blood sample. 2. Previously available methods measure only the mean PCl of red cell populations. We describe a 'profile migration' method in which dilute red cell suspensions in low-K+ media were permeabilized to K+ with a high concentration of valinomycin, rendering PCl the main rate-limiting factor for cell dehydration. As the cells dehydrated, samples were processed to obtain full haemolysis curves at precise times. Variations in PCl among cells would have appeared as progressive changes in the profile of their haemolysis curves, as the curves migrated towards lower tonicities. 3. Red cells from five normal volunteers showed no change in profile of the migrating haemolysis curves, suggesting that their PCl distributions were fairly uniform. Quantitative analysis demonstrated that intercell variation in PCl was less than 7.5%. 4. Results obtained with this technique were analysed using the Lew-Bookchin red cell model. The calculated PCl was within the normal range described in earlier studies.

Generation of normal human red cell volume, hemoglobin content, and membrane area distributions by "birth" or regulation?

Using flow cytometry and osmotic lysis measurements, we document here the means and coefficients of variation of the following red cell (RBC) properties: hemoglobin (Hb) content, volume, Hb concentration, and relative lytic tonicity distributions in populations of normal human RBCs, before and after density fractionation. The distributions showed a pattern characterized by much larger coefficients of variation of the Hb content and volume distributions than of the Hb concentration and relative lytic tonicity distributions. From analysis of the factors that determine those RBC properties, the patterns were interpreted as reflecting previously unrecognized statistical proportionalities between cell osmolyte content, Hb content, and membrane area. The possible origin of these statistical links was analyzed by considering alternative models with and without the participation of regulatory processes during cell maturation. A model was shown to be feasible in which mature RBC variability with proportional volume, area, and Hb content arises solely from cell size variability at the last erythroid cell division.

Measurement of the hemoglobin concentration in deoxyhemoglobin S polymers and characterization of the polymer water compartment.

Biological polymers contain freely exchangeable water within intermolecular crevices with restricted access to large extrapolymer solutes. Our recent studies highlighted large osmotic effects of such polymer water compartments (PWCs), and their substantial physiological and pathophysiological relevance. The size and accessibility of the PWC are critical parameters determining the polymers' osmotic properties. We report here a new experimental approach to investigate these parameters in deoxyhemoglobin S polymers. The size of the PWC is inversely related to the deoxyhemoglobin S concentration in the polymer (CP). Only an approximation of CP (approximately 69 g/dl) was previously available. By analyzing the distributions of soluble hemoglobin and a large molecular weight (MW) marker (14C-dextran, MW approximately 70kDa) in the supernatant and pellet of centrifuged gels, we obtained a reproducible value of CP, 54.7 (+/- 0.7)g/dl. This indicates that 60% of the polymer is composed of a water compartment inaccessible to soluble Hb and other non-interactive macromolecules. The accessibility properties of this PWC to smaller molecules were explored with markers of different MW. Non-interactive markers with MW < 200 kDa diffused freely in the PWC, whereas those with 300 kDa < MW < 1000 kDa showed partial exclusion. Higher MW markers were generally excluded, except molecules with elongated (rather than spherical) shapes or possible interactivity with hemoglobin. These results predict that dense sickle cells would significantly dehydrate on deoxygenation, generating a PWC of up to 60% to 80% of the cell water. Soluble enzymes would concentrate in the residual cytosol. For osmotic equilibrium, most of the ions and low MW substrates would concentrate in the PWC. Oxygenation-deoxygenation would thus cause dynamic oscillations in cell hydration and between states of single and double cytoplasmic water phases, the latter with a substantially altered internal environment. The relevance of such oscillations to the membrane and metabolic abnormalities of dense sickle cells requires further investigation.

Effects of deoxygenation on active and passive Ca2+ transport and on the cytoplasmic Ca2+ levels of sickle cell anemia red cells.

Elevated [Ca2+]i in deoxygenated sickle cell anemia (SS) red cells (RBCs) could trigger a major dehydration pathway via the Ca(2+)-sensitive K+ channel. But apart from an increase in calcium permeability, the effects of deoxygenation on the Ca2+ metabolism of sickle cells have not been previously documented. With the application of 45Ca(2+)-tracer flux methods and the combined use of the ionophore A23187, Co2+ ions, and intracellular incorporation of the Ca2+ chelator benz-2, in density-fractionated SS RBCs, we show here for the first time that upon deoxygenation, the mean [Ca2+]i level of SS discocytes was significantly increased, two- to threefold, from a normal range of 9.4 to 11.4 nM in the oxygenated cells, to a range of 21.8 to 31.7 nM in the deoxygenated cells, closer to K+ channel activatory levels. Unlike normal RBCs, deoxygenated SS RBCs showed a two- to fourfold increase in pump-leak Ca2+ turnover. Deoxygenation of the SS RBCs reduced their Ca2+ pump Vmax, more so in reticulocyte- and discocyte-rich than in dense cell fractions, and decreased their cytoplasmic Ca2+ buffering. Analysis of these results suggests that both increased Ca2+ influx and reduced Ca2+ pump extrusion contribute to the [Ca2+]i elevation.

The distribution of intracellular calcium chelator (fura-2) in a population of intact human red cells.

Using quantitative fluorescence microscopy of red cells loaded non-disruptively with 1-2.5 mmol/l cells of fura-2, we examined the distribution of the incorporated free chelator among and within individual cells. Cytoplasmic hemoglobin quenched the effective fluorescence yield of fura-2 by a factor of about 100. All red cells were found to fluoresce upon excitation at 380 nm, and the fluorescence intensities they emitted at 510 nm were approximately +/- 20% about the mean intensity, indicating a fairly uniform distribution of incorporated chelator among the cells. Red cells loaded with these high levels of fura-2 retained their biconcave shape, and a comparison between their transmission images at 415 nm and their fura-2 fluorescence images suggests that the concentration of fura-2 was also uniform throughout the cytosol. These results validate assumptions made in earlier experiments with non-fluorescent incorporated Ca2+ chelators, and demonstrate the feasibility of fura-2 and Ca2+ imaging of intact red cells, despite considerable quenching of probe fluorescence by hemoglobin.

Effects of deoxygenation on active and passive Ca2+ transport and cytoplasmic Ca2+ buffering in normal human red cells.

1. The effects of deoxygenation on cytoplasmic Ca2+ buffering, saturated Ca2+ extrusion rate through the Ca2+ pump (Vmax), passive Ca2+ influx and physiological [Ca2+]i level were investigated in human red cells to assess whether or not their Ca2+ metabolism might be altered by deoxygenation in capillaries and venous circulation. 2. The study was performed in fresh human red cells maintained in a tonometer either fully oxygenated or deoxygenated. Cytoplasmic Ca2+ buffering was estimated from the equilibrium distribution of 45Ca2+ induced by the divalent cation ionophore A23187 and the Vmax of the Ca2+ pump was measured either by the Co(2+)-exposure method or following ionophore wash-out. The passive Ca2+ influx and physiological [Ca2+]i were determined in cells preloaded with the Ca2+ chelator benz-2 and resuspended in autologous plasma. 3. Deoxygenation increased the fraction of ionized Ca2+ in cell water by 34-74% and reduced the Vmax of the Ca2+ pump by 18-32%. 4. To elucidate whether or not these effects were secondary to deoxygenation-induced pH shifts, the effects of deoxygenation on cell and medium pH, and of pH on cytoplasmic Ca2+ binding and Ca2+ pump Vmax in oxygenated cells were examined in detail. 5. Deoxygenation generated large alkaline pH shifts that could be explained if the apparent isoelectric point (pI) of haemoglobin increased by 0.2-0.4 pH units in intact cells, consistently higher than the value of 0.15 reported for pure haemoglobin solutions. 6. In oxygenated cells, the fraction of ionized cell calcium, alpha, was little affected by pH within the 7.0-7.7 range. Ca2+ pump Vmax was maximal at a medium pH of about 7.55. Comparison between pH effects elicited by HCl-NaOH additions and by replacing Cl- with gluconate suggested that Vmax was inhibited by both internal acidification and external alkalinization. Since deoxygenation alkalinized cells and medium within a range stimulatory for Vmax, the inhibition observed was not due to pH. 7. There was no significant effect of deoxygenation on passive Ca2+ uptake, or steady-state physiological [Ca2+]i level. 8. The deoxygenation-induced reduction in Ca2+ binding capacity may result from the increased protonation of haemoglobin on deoxygenation and from binding of 2,3-diphosphoglyceric acid (2,3-DPG) and ATP to deoxyhaemoglobin; inhibition of the Ca2+ pump may result from shifts in the [Mg2+]i/[ATP]i ratio away from a near optimal stimulatory value in the oxygenated state.

Osmotic effects of protein polymerization: analysis of volume changes in sickle cell anemia red cells following deoxy-hemoglobin S polymerization.

Polymerization-depolymerization of proteins within cells and subcellular organelles may have powerful osmotic effects. As a model to study these we analyzed the predicted volume changes following hemoglobin (Hb) S polymerization in sickle cell anemia (SS) red cells with different initial volumes. The theoretical analysis predicted that dehydrated SS red cells may sustain large polymerization-induced volume shifts whose direction would depend on whether or not small solutes were excluded from polymer-associated water. Experiments with SS cells from promptly fractionated venous blood showed oxygenation-induced swelling, maximal in the densest cells, in support of nonexclusion models. The predicted extent of cell dehydration on polymerization was strongly influenced by factors such as the dilution of residual soluble Hb and the increased osmotic contribution of Hb in cells dehydrated by salt loss, largely overlooked in the past. The osmotic effects of polymer formation may thus play an important part in microcirculatory infarction by dense SS cells, as they become even denser and stiffer during deoxygenation in the capillaries.

A mathematical model of the volume, pH, and ion content regulation in reticulocytes. Application to the pathophysiology of sickle cell dehydration.

We developed a mathematical model of the reticulocyte, seeking to explain how a cell with similar volume but much higher ionic traffic than the mature red cell (RBC) regulates its volume, pH, and ion content in physiological and abnormal conditions. Analysis of the fluxbalance required by reticulocytes to conserve volume and composition predicted the existence of previously unsuspected Na(+)-dependent Cl- entry mechanisms. Unlike mature RBCs, reticulocytes did not tend to return to their original state after brief perturbations. The model predicted hysteresis and drift in cell pH, volume, and ion contents after transient alterations in membrane permeability or medium composition; irreversible cell dehydration could thus occur by brief K+ permeabilization, transient medium acidification, or the replacement of external Na+ with an impermeant cation. Both the hysteresis and drift after perturbations were shown to depend on the pHi dependence of the K:Cl cotransport, a major reticulocyte transporter. This behavior suggested a novel mechanism for the generation of irreversibly sickled cells directly from reticulocytes, rather than in a stepwise, progressive manner from discocytes. Experimental tests of the model's predictions and the hypothesis are described in the following paper.