American Journal of Hematology 41 :170-177 (1992)

Influence of Calcium Permeabilization and Membrane-Attached Hemoglobin on Erythrocyte Deformability E. Friederichs, R.A. Farley, and H.J. Meiselman Department of Physiology and Biophysics, University of Southern California, School of Medicine, Los Angeles, California

The present study was designed to evaluate the influence of intracellular calcium [Ca], regulated membrane attached hemoglobin (Hb,) on the deformability of human RBC and ghosts. [Ca], of RBC was elevated via the ionophore A23187 (10 kM); the deformability of RBC and resealed ghosts was determined via measuring RBC and ghost transit times through 5 km diameter pores with the Cell Transit Analyzer (CTA). Salient results included: (1) significantly increased RBC levels of Hb, following ionophore treatment; (2) elevated Hb, with increasing lysing medium calcium concentration (0-5 rnM); (3) decreased deformability of both intact RBC and ghosts with increasing Hb, and significant (P < 0.02 or better) linear relationships between Hb, and RBC or ghost transit times; and (4) an increased sensitivity to ionophore treatmenthnembraneattached hemoglobin for the higher percentiles of the CTA transit time distribution (i.e., for more rigid subpopulations). Our results thus indicate that calcium-induced interaction of hemoglobin with the RBC membrane produces cellular rheological changes; in addition, they demonstrate the usefulness of the CTA system in measuring both average RBC rheologic behavior and the distribution of cellular rheologic properties within an erythrocyte population. 0 1992 Wiley-Liss, h e .

Key words: calcium, hemoglobin, ionophore, RBC deformability, RBC membrane

INTRODUCTION

In the microcirculation, rigid cells have difficulty when passing through small vessels, splenic slits or capillaries [ 1,2]. Such cells, which may be only 2 or 3% of the total RBC population, may thus have large effects on blood flow and RBC distribution and may also decisively influence the results obtained during filtration experiments using narrow pore filters [l]. While the study of such subpopulations is possible using micropipette methods on individual cells (e.g., 3), this technique is extremely time consuming and yields relatively few observations. The concept of subpopulations of less deformable RBC is thus the rationale for the development of a new micropore filtration system, the Cell Transit Analyzer (CTA), which measures the rheological characteristics of a large number of individual erythrocytes and provides a frequency distribution of these properties [4,5]. Note that for RBC populations having a relatively small subpopulation of rigid cells, use of the pcrcentiles of the micropore transit time distribution, rather than only 0 1992 Wiley-Liss, Inc.

the overall mean transit time, provides a sensitive index to the presence of rigid erythrocytes [4]. Cell deformability (i.e., the ability of the entire cell to assume a new shape in response to deforming forces) has been ascribed to four factors (1,2,6): (1) surface area-tovolume ratio; (2) shape; ( 3 ) internal viscosity; and (4) membrane mechanical properties. While the effects of surface area-to-volume ratio, RBC shape, and internal viscosity have been extensively studied, the relative contribution of membrane properties to cellular deformability is less clear. In addition, although detailed characterization of RBC rheological properties has been accomplished (e.g ., 2), the biochemical mechanisms leading to premature destruction of red blood cells are less well known. However, several biochemical alterReceived for publication October 11, 1991; accepted March 10, 1992. Address reprint requests to Dr. H.J. Meiselman, Department of Physiology and Biophysics, University of Southern California, Medical School, 2025 Zonal Avenue, Los Angeles, CA 90033.

Calcium and Membrane-Attached Hemoglobin

ations have been suggested as possible mechanisms for membrane changes that may lead to altered cellular rheology (e.g., crosslinking of membrane proteins, accumulated oxidant damage, increased binding of hemoglobin) [3,7-131. 1. It is well known that RBC cytoskeletal proteins play an important role in determining membrane material properties (e.g., 2). Chasis and Mohandas [ 101 have proposed that the membrane skeleton can be viewed as a network of folding and unfolding spectrin molecules held together by protein-protein associations at both ends of the spectrin heterodimer. Increased interaction of either inter- or intramolecular proteins may thus have a profound effect on membrane deformability . 2. Interaction of hemoglobin with the erythrocyte membrane has been described by Shaklai and co-workers [12,13] and by Eisinger et al. [14] and Claster et al. [15], who conclude that there is a reversible, weak interaction with band 3, which depends on pH and ionic strength, and that oxidants enhance hemoglobin binding [ 131. 3. The role of various intracellular constituents (e.g., intracellular ions) in regulating membrane properties by influencing membrane attachment of hemoglobin has not been fully analyzed. In this context, the role of calcium in affecting erythrocyte deformability has been a subject of some controversy [16]; reports conflict as to whether membrane viscoelastic properties are altered at elevated RBC calcium concentrations [ 17-19]. Although there are reports of calcium- and/or magnesium-induced membrane association of hemoglobin [20,21], it is not yet clear whether Ca ions can alter membrane mechanical properties through a direct interaction with skeletal proteins or through a calcium-induced hemoglobin interaction with parts of the protein network. The present study was designed to evaluate the influence of intracellular calcium regulated membraneattached hemoglobin on the deformability of human RBC and derived ghosts; of particular interest were the effects of calcium permeabilization on the rheological behavior of various RBC subpopulations. Our results indicate significant correlations between membrane attached hemoglobin and both RBC and ghost deformability and suggest the importance of hemoglobin-membrane interactions as a determinant of erythrocyte deformability . MATERIALS AND METHODS Sample Preparation

171

KCI, pH = 7.4, 290 mOsm/kg), then recombined with autologous plasma to a hematocrit of -50%. Subsequently, 0.026 ml of a l-mg/ml solution of the calcium ionophore A23 187 (Sigma) dissolved in DMSO (dimethy1 sulfoxide) was added to 5 ml of the 50% hematocrit suspension, and the RBC suspension incubated for 30 min at 37°C (10 p M final A23187 concentration, 0.5% final DMSO concentration). Control, ionophore-free suspensions containing 0.5% DMSO were also incubated for 30 min at 37°C. Note that this incubation procedure with A23 187 has been previously shown to result in a twofold increase of intracellular calcium [22] and that the short incubation period in plasma maintains physiological levels of ATP, hence preventing echinocyte shape changes and volume loss due to potassium efflux [ 11,23,24]. Following incubation, the RBC were washed three times in HEPES buffer containing 0.4 g/dl human albumin (American Red Cross) to remove the ionophore [22]. The cells were then diluted in HEPES buffer to 5 X lo6/ ml for CTA analysis and to a hematocrit of 2 0 4 0 % for measurement of hematologic parameters (see below). In addition, RBC shape following incubation was examined by phase-contrast microscopy; evidence of shape transformation (i.e., echinocyte formation) was judged sufficient to reject the preparation and thus only RBC with normal, biconcave morphology were tested. Ghost Preparation

Resealed ghosts were prepared from control and ionophore-treated RBC via a modification of the method of Nash and Meiselman [25]. Whole blood was centrifuged (2,00Og, 5 min), the plasma and buffy coat removed, and the RBC suspended at 12.5% hematocrit in the isotonic HEPES buffer. Lysis was induced by adding 1 ml of ice-cold cell suspension to 37 ml of o"C, dilute HEPES buffer (20 mM HEPES, 5 mM KC1, pH 7.4, =40 mOsm/ kg). After a delay time of 2 min at O"C, the suspension was returned to isotonicity by addition of 7 ml of 5 times concentrated HEPES buffer; in one series of experiments, the delay time prior to restoring isotonicity was varied in order to produce ghosts with various volumes [25]. Resealing of ghosts was facilitated by incubation of the isotonic suspension at 37°C for 1 hr. Following resealing, the ghosts were centrifuged at 20,OOOg for 10 min, the supernatant removed, and the pellet washed twice in a large excess of isotonic HEPES buffer. The ghosts were finally resuspended in isotonic HEPES buffer for measurement of the membrane attached hemoglobin, ghost volume and ghost deformability; ghost morphology was also examined via phase microscopy.

Blood was obtained from healthy, adult laboratory personal via venipuncture into heparin (14 U/ml) and used within 4-6 hr. RBC were separated by centrifugation (2,00Og, 5 rnin), the buffy coat discarded, and the plasma saved. The RBC were washed twice in an isotonic Attached Hemoglobin Analysis HEPES (N-2-hydroxyethylpiperazine-N'-2-ethansulfo- Membrane-attached hemoglobin per cell was calcunic acid) buffer (20 mM HEPES, 140 mM NaCl, 5 mM lated from the absorbance of resealed ghost suspensions

172

Friederichs et al. TABLE 1. Effects of lonophore Treatment on RBC Volume, Membrane-AttachedHemoglobin (Hb,), and Deformability* MCV

(fl) Ionophore Control P
1 mM. Note also that at the nominal plasma calcium concentration (i.e., 2 mM), the amount of attached hemoglobin in these ghosts is slightly higher than that observed for ionophore-treated RBC in plasma (Table I); this finding may possibly be due to lower intracellular calcium when using the ionophore/plasma incubation procedure. Calcium and Membrane Cytoskeleton

Evaluation of membrane protein changes consequent to ionophore treatment or lysis in calcium-containing media was carried out via SDS-PAGE gel electrophoresis [26]. While very minor alterations of several bands appeared to exist (data not shown), the most obvious and

0

I

0

I

I

I

I

I

1 2 3 4 CALCIUM IN LYSING MEDIUM

5

8

(mM)

Fig. 7. Effect of lysing medium calcium concentration on membrane-attached hemoglobin (Hb,); resealed ghosts were prepared from washed, control (i.e., no ionophore) RBC with various concentrationsof calcium (0-5 mM) in the hypotonic lysing medium. Note that Hb, increases from about 0.8 pgicell in the absence of calcium to ~ 2 . pglcell 2 at 5 mM calcium, with essentially constant levels of Hb, above 1 mM calcium.

consistent change observed following either treatment with 10 pM A23 187 or lysis in 1-5 mM calcium was a marked increase above control of membrane-attached hemoglobin; these qualitative gel electrophoresis results are thus in agreement with our quantitative measures of Hb, (i.e., Table I, Fig. 7). DISCUSSION

It is generally agreed [ 1,2,6] that the ability of RBC to deform depends on the cellular membrane surface areato-volume ratio, cell shape, intracellular viscosity, and membrane mechanical properties. However, evaluation of the results presented herein suggests that the first three of these factors are not responsible for the observed effects of membrane attached hemoglobin (Hb,) on cellular deformability: (1) only a minimal decrease of MCV was found following ionophore treatment (Table I), with the resulting slight increase of the surface area to volume ratio expected to yield increased rather than decreased cell deformability [1,6]; (2) only RBC or ghosts with normal biconcave shapes were tested for deformability; and (3) data obtained for RBC ghosts, which have markedly reduced intracellular hemoglobin concentration and hence very low cytoplasmic viscosity, also showed significant effects of Hb, on cellular deformability (Fig. 6). In particular, it is of interest to note that the slopes of RBC and ghost mean transit time versus Hb, are similar (Figs. 1, 6). Thus, the present findings indicate that the

176

Friederichs et al.

effects of calcium-mediated hemoglobin binding on RBC and ghost membrane mechanical properties must be considered. Inasmuch as the CTA system provides average rheologic data as well as transit time histograms for each cell suspension [4,5], analyses of these frequency distributions yields information regarding different RBC subpopulations. The higher percentiles (i.e., P-90, P-95, P-99) show larger increases of transit time consequent to ionophore treatment (Fig. 3), thus suggesting that these subpopulations may have increased amounts of attached hemoglobin. This possibility is supported by the different slopes obtained when the various percentiles of the RBC transit time data are regressed versus Hb, (Fig. 4); the increased slopes for the higher percentiles may correspond to increased amounts of membrane-attached hemoglobin for the more rigid subpopulations, presumably the older cells. Garcia-Sanchez and Lew [27,28] have shown that ionophore-induced calcium distribution is markedly heterogeneous; these differences most likely reflect variations in active calcium transport among RBC. Intracellular calcium is contained within compartments, with these compartments having properties of endocytic inside-out vesicles capable of ATP-dependent calcium accumulation. Nevertheless, more than 99% of intracellular calcium is rapidly mobilized by A23 187 [ 10,l I] and therefore is not irreversibly bound or precipitated. Note that while calcium is present in both young and old RBC, the amount is many times greater in older cells [27,28]. Furthermore, it is notable that for non-ionophore treated RBC, older cells exhibit a higher membrane viscosity and a greater sensitivity of this membrane mechanical property to increases of MCHC [3]; these membrane rheologic effects were suggested to be due to concentration-dependent hemoglobin-membrane interactions. Thus, differing initial levels of intracellular calcium and of hemoglobin concentration may play a role in the observed responses of RBC subpopulations to calcium permeabilization (Figs. 3, 4). Clark and co-workers [ 171 have reported that increased internal viscosity is responsible for the decreased deformability of calcium-loaded RBC, with echinocyte formation also important at high calcium levels (i.e., >500 pM). Conversely, Dreher et al. [29] and Kuettner et al. [30] suggest that decreased deformability is due to increased membrane rigidity, but that changes occur only if calcium accumulation is accompanied by decreased cell volume, loss of potassium, and ATP hydrolysis. However, our results question the absolute need for these biochemicaUgeometrica1 changes: ( I ) Ionophore treatment was performed using RBCiplasma suspensions, thus yielding very small increases of intracellular calcium (i.e., from 0.2 pM to 0.4 pM, see Friederichs et al. [22]); (2) Incubation in plasma served to maintain MgATP at physiologic levels and to prevent loss of potas-

sium secondary to severe magnesium and ATP depletion [ 11,241; and ( 3 ) Echinocyte formation was not observed in the present study. While the CTA system is sensitive to MCHC and internal viscosity changes (i.e., Figs. 1,6), it seems obvious that there is a calcium-dependent mechanism that affects cellular deformability independently from internal viscosity. Data from O’Rear and co-workers [18] support the suggestion that calcium may be related to changes in RBC membrane viscoelastic properties: by subjecting RBC to fluid shear stresses, they induced increased intracellular calcium and decreased deformability without changes in ATP, magnesium, cell volume, or morphology, and without large sodium or potassium fluxes. Shiga et al. [ 191 also report decreased deformability consequent to calcium accumulation, and that only a partial reversal of the deformability changes was possible via exposure of RBC to hypotonic media. Several calcium-dependent events at the RBC membrane cytoskeleton have been described, including membrane binding of cytoplasmic proteins, proteolysis of membrane proteins, and nonreversible protein aggregation [7,8,9,3I]; aggregate formation only becomes important at about 0.1 mM calcium [31] and thus at much greater levels than achieved herein [22]. Furthermore, it is notable that our SDS-PAGE findings consistently indicated marked increases of membrane-attached hemoglobin following ionophore treatment or lysis in 1-5 mM calcium, with very minor alterations of other bands. While we cannot totally exclude specific calcium effects unrelated to elevated membrane-bound hemoglobin, both prior studies [3,18,19,32] and our observations lead to the conclusion that increased Hb, has an adverse effect on erythrocyte membrane rheologic behavior (i .e., higher membrane rigidity and/or viscosity) and hence on cellular deformability . The mechanisms responsible for calcium-induced loss of deformability are of considerable interest, since similar changes of RBC mechanical behavior occur in diseases such as sickle cell anemia, hereditary spherocytosis and thalassemia, as well as during the normal erythrocyte aging process [33]. It is of interest that Shrier et al. [34] could demonstrate that for thalassemic erythrocytes, the dynamic rigidity of the membrane is influenced not only by cell hemoglobin concentration but also by the extent of hemoglobin interaction with the membrane. However, it should be noted that for some of the abovementioned diseases, the relative importance of cytoplasmic viscosity versus membrane mechanical properties remain unclear [35] ; further studies, including those involving densityseparated RBC, are thus clearly warranted. ACKNOWLEDGMENTS

This work was supported by NIH research grants HL 15722, HL 41341, by an award from the American Heart

Calcium and Membrane-Attached Hemoglobin

Association-Greater Los Angeles Affiliate (537IG), and by a grant from Deutsche Forschungsgemeinschaft (Fr 752 1-1 and 1-2). The technical assistance of Ms. R.B. Wenby is gratefully acknowledged.

REFERENCES I . Dormandy J: “Red Cell Deformability and Filterability.” Boston: Martinus Nijhoff, 1983. 2. Cokelet GR, Meiselman HJ, Brooks DE: “Erythrocyte Mechanics and Blood Flow.” New York: Alan R. Liss, 1980. 3. Nash GB, Meiselman HJ: Red cell and ghost viscoelasticity: Effects of hemoglobin concentration and in vivo aging. Biophys J 43:63, 1983. 4. Koutsouris D, Guillet R, Wenby RB, Meiselman HJ: Determination of erythrocyte transit times through micropores. ii. Influence of experimental and physicochemical factors. Biorheology 2688 I, 1989. 5 . Koutsouris D, Guillet R, Lelievre JC, Guillemin MT, Bertholom P, Beuzard Y, Boynard M: Determination of erythrocyte transit times through micropores. i. Basic operational principles. Biorheology 25:763, 1988. 6. Donnandy I: “Blood Filtration and Blood Deformability.” Boston: Martinus Nijhoff, 1985. 7. Carraway KL, Tripleet RB, Anderson DR: Calcium-promoted aggregation of erythrocyte membrane proteins. Biochim Biophys Acta 379371, 1975. 8. King LE, Morrison M: Calcium effects on human erythrocyte membrane proteins. Biochim Bipohys Acta 471:162, 1977. 9. Allen DW, Cadman S: Calcium-induced erythrocyte membrane changes. The role of adsorption of cytosol proteins and proteases. Biochim Biophys Acta 55 1 : 1, 1979. 10. Chasis JA, Mohandas N: Erythrocyte membrane deformability and stability: Two distinct membrane properties that are independently regulated by skeletal protein associations. J Cell Biol 103:343, 1986. I I . Sheetz MP, Singer SJ: On the mechanism of ATP-induced shape changes in human erythrocyte membranes. I. The role of the spectrin complex. J Cell Biol73:638, 1977. 12. Shaklai N, Yguerabide J, Ranney HM: Interaction of hemoglobin with red cell membranes as shown by fluorescence chromophore. Biochemistry 16:5585, 1977. 13. Shaklai N, Frayman B, Fortier N, Snyder M: Crosslinking of isolated cytoskeletal proteins with hemoglobin: A possible damage inflicted to the red cell membrane. Biochim Biophys Acta 915:406, 1987. 14. Eisinger J, Flores J, Bookchin RM: The cytosol membrane interface of normal and sickle erythrocytes. J Biol Chem 259:7169, 1984. 15. Claster S, White E, Woolworth V, Quintanilha A: Degradation of erythrocyte glycophorin results in increased membrane bound hemoglobin. Arch Biochem Biophys 285: 147, 1991. 16. Engelman B: Calcium homeostasis of human erythrocytes and its pathophysiological implications. Klin Wochenschr 69: 137, 1991. 17. Clark MR, Mohandas N, Feo C, Jacobs MS, Shohet SB: Separate mechanisms of deformability loss in ATP-depleted and Ca-loaded cells. J Clin Invest 6 7 5 3 I , 1981.

177

18. O’Rear EA, Udden MM, McIntire LV, Lynch EC: Reduced erythrocyte deformability associated with calcium accumulation. Biochim Biophys Acta 691:274, 1982. 19. ShigaT, SekiyaM, Maeda N, Kon K, Okazaki M: Cell age-dependent changes in deformability and calcium accumulation of human erythrocytes. Biochim Biophys Acta 814:289, 1985. 20. Tasso M, Morelli A, de Flora A: Calcium-induced alterations in the levels and subcellular distribution of proteolytic enzymes in human red blood cells. Biochem Biophys Res Comniun 138:87, 1986. 21. Eaton JW, Branda RF, Hadland C, Dreher K: Anion channel blockade: Effects upon erythrocyte membrane calcium response. Am J Hematol 9:391, 1980. 22. Friederichs E, Radisch T, Winkler H: Calcium content of the erythrocytes: A sensitive and easy handling method for measuring free calcium ions, and modulation of the Cazi ion concentration by the calcium antagonists nifedipine and pentoxifylline. Clin Exp Pharmacol Physiol 16:387, 1989. 23. Reed PW, Lardy HA: A23187: A divalent cation ionophore. J Biol Chem 247:6970, 1972. 24. Reed PW: Calcium dependent potassium efflux from rat erythrocytes incubated with antibiotic A23187. Fed Proc 72:635, 1973. 25. Nash GB, Meiselman HJ: Effects of preparative procedures on the volume and content of resealed red cell ghosts. Biochim Biophys Acta 815:477, 1985. 26. Laemmli UK: Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227:660, 1970. 27. Garcia-Sanchez J, Lew VL: Heterogeneous calcium and adenosine triphosphate distribution in calcium-permeabilized human red cells. J Physiol (Lond) 407:523, 1988. 28. Garcia-Sanchez J, Lew VL: Detection and separation of human red cells with different calcium contents following uniform calcium permeabilization. J Physiol (Lond) 407505, 1988. 29. Dreher KL, Eaton JW, Kuettner JF, Breslawac BS, Blackshear PL, White JG: Retention of water and potassium by erythrocytes prevents calcium-induced membrane rigidity. Am J Pathol92:215, 1978. 30. Kuettner JF, Dreher JL, Rao GH, Eaton JW, Blackshear PL, White JG: Influence of the ionophore A23187 on the plastic behavior of normal erythrocytes. Am J Pathol88:81, 1977. 3 I . Anderson DR, Davis JL, Carraway KL: Calcium-promoted changes of the human erythrocyte membrane. J Biol Chem 252:6617, 1977. 32. Nash GB, Meiselman HJ: Effect of dehydration on the viscoelastic behavior of red cells. Blood Cells 17517, 1991. 33. Friederichs E, Lakomek M, Winkler H, Tillmann W: Effects of calcium ions and 2,3-diphosphoglycerate on the solubility of deoxygenated human hemoglobin. Clin Hemorheol 8:707, 1988. 34. Schrier SL, Rachmilewitz E, Mohandas N: Cellular and membrane properties of Alpha and Beta Thalassemic erythrocytes are different: Implication for differences in clinical manifestations. Blood 74:2194, 1989. 35. Mohandas N, Shrier SL: Mechanisms of red blood cell destruction in hemolytic anemias. In Mentzer WC, Wagner GM (eds): “The Hereditary Hemolytic Anemias.” New York: Churchill Livingstone, 1989.