Deoxygenation-induced and Ca2+ dependent phosphatidylserine externalisation in red blood cells from normal individuals and sickle cell patients

Cell Calcium 51 (2012) 51–56

Contents lists available at SciVerse ScienceDirect

Cell Calcium journal homepage: www.elsevier.com/locate/ceca

Deoxygenation-induced and Ca2+ dependent phosphatidylserine externalisation in red blood cells from normal individuals and sickle cell patients Erwin Weiss a , Urszula M. Cytlak a , David C. Rees b , Anna Osei b , John S. Gibson a,∗ a b

Department of Veterinary Medicine, University of Cambridge, Madingley Road, Cambridge CB3 0ES, UK Department of Pediatric Haematology, King’s College Hospital, Denmark Hill, London SE5 9RS, UK

a r t i c l e

i n f o

Article history: Received 16 June 2011 Received in revised form 29 September 2011 Accepted 23 October 2011 Available online 22 December 2011 Keywords: Sickle cell Phosphatidylserine Calcium affinity Magnesium Deoxygenation

a b s t r a c t Phosphatidylserine (PS) is usually confined to the inner leaflet of the red blood cell (RBC) membrane. It may become externalised in various conditions, however, notably in RBCs from patients with sickle cell disease (SCD) where exposed PS may contribute to anaemic and ischaemic complications. PS externalisation requires both inhibition of the aminophospholipid translocase (or flippase) and activation of the scramblase. Both may follow from elevation of intracellular Ca2+ . Flippase inhibition occurs at low [Ca2+ ]i , about 1 ␮M, but [Ca2+ ]i required for scrambling is reported to be much higher (around 100 ␮M). In this work, FITC-labelled lactadherin and FACS were used to measure externalised PS, with [Ca2+ ]i altered using bromo-A23187 and EGTA/Ca2+ mixtures. Two components of Ca2+ -induced scrambling were apparent, of high (EC50 1.8 ± 0.3 ␮M) and low (306 ± 123 ␮M) affinity, in RBCs from normal individuals and the commonest SCD genotypes, HbSS and HbSC. The high affinity component was lost in the presence of unphysiologically high [Mg2+ ] but was unaffected by high K+ (90 mM) or vanadate (1 mM). The high affinity component accounted for PS scrambling in ≥2/3rd RBCs. It is likely to be most significant in vivo and may be involved in the pathophysiology of SCD or other conditions involving eryptosis. © 2011 Elsevier Ltd. All rights reserved.

1. Introduction In red blood cells (RBCs), as in other cells, the aminophospholipid phosphatidylserine (PS) is usually confined to the inner leaflet of the plasma membrane [1]. Externalised PS may have pathological significance; it is prothrombotic, encourages phagocytosis by macrophages, and can bind to receptors on activated endothelial cells. This has led to the concept of “eryptosis”, an RBC process sharing some of the features of apoptosis in nucleated cells, and in particular lipid scrambling [2]. PS exposure is seen in a number of disease states and notably in sickle cell disease (SCD) in which a high, but variable, percentage of RBCs (about 2–10%) show exposed PS [1,3]. Externalised PS in SCD patients may therefore contribute to the anaemic and ischaemic complications of the disease. Notwithstanding its potential importance, however, the precipitating causes of PS exposure remain uncertain. Distribution of PS is the consequence of the activity of two main transport systems: the aminophospholipid translocase (APLT or flippase) is an ATP-driven transporter moving PS from outer to inner bilayer, whilst the scramblase–whose identity remains uncertain [4] – is a Ca2+ -dependent system which is able to

∗ Corresponding author. Tel.: +44 1223 337638; fax: +44 1223 337610. E-mail address: [email protected] (J.S. Gibson). 0143-4160/$ – see front matter © 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.ceca.2011.10.005

translocate PS rapidly in either direction [5]. Following APLT inhibition alone, PS redistribution and exposure will occur – but only slowly. For rapid PS exposure both APLT inhibition and also scramblase activation are required. The ratio of activities of APLT/scramblase is considered crucial for PS localisation [6]: at high ratio PS is predominantly internalised, as the ratio declines, PS exposure is favoured. In RBCs from SCD patients, APLT inhibition and scramblase activation may result from a number of factors including the oxidative stress which follows from instability of HbS leading to accumulation of heme and free iron [7,8]. Ca2+ is also thought to be involved as RBCs from these patients have perturbed Ca2+ homeostasis especially when deoxygenated [9]. A deoxygenation-induced conductance (termed Psickle ) allows entry of Ca2+ and this together with a concurrent reduction in activity of the plasma membrane Ca2+ pump can lead to elevation of intracellular Ca2+ [9–11]. Although APLT inhibition occurs with fairly high Ca2+ affinity (c. 1 ␮M) [12], it is generally accepted that scramblase activation requires a [Ca2+ ]i about two orders of magnitude higher (c. 100 ␮M) [4,13]. These relatively high [Ca2+ ]i s are unlikely to be achieved in vivo in a viable RBCs in which the high capacity Ca2+ pump minimises increases in free intracellular [Ca2+ ], notwithstanding increased Ca2+ permeability such as occurs with Psickle activation [11,14]. The Ca2+ pump-leak characteristics may therefore mitigate against the involvement of Ca2+ in PS scrambling in viable RBCs. In addition, Gardos channel activation, and ensuing rapid K+ and

E. Weiss et al. / Cell Calcium 51 (2012) 51–56

Cl− loss and RBC shrinkage would be markedly stimulated at [Ca2+ ]i s of a few hundred nanomolar [15], considerably lower than those thought to activate the scramblase. Shrinkage itself would lead to PS exposure [16,17]. These features question whether Ca2+ is directly involved in PS scrambling. In this report, we reassess the role of Ca2+ in PS exposure over a range of [Ca2+ ]i s of 6 orders of magnitude in RBCs from both main SCD genotypes, HbSS and HbSC and from normal individuals (HbAA). Results show that Ca2+ -induced scrambing can in fact occur at much higher Ca2+ affinity than hitherto proposed. Intracellular Ca2+ levels may therefore co-ordinate several eryptotic events (flippase inhibition, lipid scrambing, Gardos channel activation). This report also provides the first determination of the Ca2+ levels required for PS exposure in RBCs from SCD patients.

Percentage of RBCs with externalised phosphatidylserine

52

15

(a) HbSS

0mM Ca2+ * 1.1mM Ca2+

Deoxygenated conditions

10

* *

5

0 0

20

40

60

80

2. Materials and methods 2.1. Blood samples and salines Anonymised, discarded, routine blood samples (taken into the anticoagulant EDTA) were collected from individuals of HbSS, HbSC or HbAA genotypes. RBCs were washed into saline, comprising (in mM) NaCl 55, KCl 90, MgCl2 0.15, inosine 10 and HEPES 10, pH 7.4 at 37 ◦ C (HK saline), or the same except with NaCl 140 and KCl 4 (LK saline). Osmolality was 290 ± 5 mOsm kg−1 . They were then given a final wash in saline with nominally 0 Ca2+ and 2 mM EGTA to remove contaminant Ca2+ . In most experiments, RBCs (0.8% haematocrit) were incubated at 37 ◦ C for 30 min in the presence of vanadate (1 mM – except where indicated otherwise, e.g. Fig. 3). Apart from the experiment shown in Fig. 1, experiments were carried out in oxygenated RBCs. To deoxygenate RBCs, they were placed in gently rotating Eshweiler tonometers flushed with humidified N2 . 2.2. Modulation of intraecellular Ca2+ and Mg2+ To clamp intracellular [Ca2+ ] ([Ca2+ ]i ), RBCs were treated with the ionophore bromo-A23187 (nominally 2–6 ␮M – although the activity of different batches varied and each was titrated to establish the minimal concentration required to clamp [Ca2+ ]i reliably) and incubated at different free [Ca2+ ]o s along with EGTA (2 mM). [Mg2+ ]o in these experiments was usually 0.15 mM which keeps [Mg2+ ]i at physiological levels [18]. In all experiments, especially those involving bromo-A23187, pH was carefully controlled and always checked, as any pH change will alter free [Ca2+ ] especially over the lower concentration ranges. Controls using RBCs loaded with the intracellular Ca2+ indicator fluo-4 (fluo-4-AM, 5 ␮M for 30 min at 37 ◦ C then washed once) showed that the increase in Ca2+ permeability was homogeneous and >95% RBCs became Ca2+ loaded in <5 min. At the concentrations used, bromo-A23187, in the absence of added Ca2+ , had no effect on RBC size and shape. In some experiments, extracellular Mg2+ was also altered (Fig. 5). In the presence of bromo-A23187, free intracellular [Ca2+ ] (or [Mg2+ ]) is clamped at a value equal to their free extracellular concentration multiplied by the square of the Donnan ratio, r2 = ([H+ ]i /[H+ ]o )2 which is about 2 [18,19].

Percentage of RBCs with externalised phosphatidylserine

Time (min) 15

0mM Ca 2+ 1.1mM Ca 2+

(b) HbSC Deoxygenated conditions

10

5

*

*

*

0 0

20

40

60

80

Time (min) Fig. 1. Percentage of red blood cells (RBCs) from sickle cell disease patients with externalised phosphatidylserine (PS). RBCs from (a) homozygous HbSS or (b) heterozygous HbSC patients were suspended in low potassium (LK)-containing saline (4 mM K+ ) without bromo-A23187 for 80 min in the presence of 0 or 1.1 mM [Ca2+ ]o . Samples were deoxygenated in gently rotating tonometers, flushed continually with N2 . Serial aliquots were taken and PS labelled with FITC-lactadherin. Percentage of positive RBCs was determined by FACS. Data represent means ± S.E.M. for RBCs from 10 (HbSS) or 5 (HbSC) patients; *p < 0.05.

binding and subsequent washing was carried out in the presence of vanadate (1 mM) to prevent any possibility of APLT-mediated resequestration of PS (notwithstanding its likely inhibition by low temperature). Control experiments showed that vanadate had no effect on lactadherin binding. Fluorescence was then measured by flow cytometry (FACSCalibur, BD) about 45 min after incubation at 37 ◦ C. The size gate for RBCs was established in control experiments using FITC-labelled anti-glycophorin A antibody (Cambridge Bioscience). The fluorescence gate for positive PS-exposing cells was set using RBCs unlabelled with FITC-lactadherin. 2.4. Statistics Unless stated otherwise, data are presented as individual results representative of at least 5 other experiments. Means are given ±S.E.M. in RBC samples from n different individuals. They were compared for statistical significance using Student’s t test.

2.3. Measurement of phosphatidylserine exposure 3. Results After 30 min incubation (except for Fig. 1), RBC aliquots were harvested and placed in cold (4 ◦ C) lactadherin-binding solution (HK or LK saline, as appropriate) with 16 nM lactadherin-FITC (Cambridge Bioscience), RBC density 4 × 105 ml−1 . Lactadherin-binding was carried out for 15 min in the dark at room temperature after which RBCs were washed once in cold binding buffer lacking FITC-lactadherin and kept on ice until FACS analysis. Lactadherin

3.1. Phosphatidylserine exposure in RBCs from SCD patients and effect of deoxygenation In freshly washed but otherwise untreated RBCs, PS exposure in RBCs from HbSS homozygotes was 1.2–13.2% with a mean ± S.E.M. of 4.5 ± 0.4% (n = 72), in agreement with previous reports [1,3].

3.2. The effect of modulation of intracellular Ca2+ The effect of clamping Ca2+ in bromo-A23187-treated RBCs was then investigated in detail. In most experiments, RBCs were also exposed to vanadate (to inhibit the P-type ATPases including the Ca2+ pump and APLT) during Ca2+ loading. Ca2+ entry and stimulation of the Ca2+ pump would otherwise rapidly deplete ATP and could lead to secondary effects. HK saline was also used, to prevent secondary shrinkage of RBCs following Gardos channel activation subsequent to Ca2+ loading. Elevation of [Ca2+ ]i led to rapid PS exposure with levels reaching a plateau after 30 min and remaining at this level for up to 2 h. For example at 100 ␮M, the percentage of PS positive RBCs was 59, 76, 81 and 80% after 10, 30, 60 and 120 min, respectively. In subsequent experiments therefore, 30 min incubation was chosen. The effect of clamping [Ca2+ ]i at different concentrations between 0 ␮M and 10 mM was determined in RBCs from the two most common SCD genotypes (homozygous HbSS and heterozygous HbSC). The relationship of PS exposure to [Ca2+ ] was similar for both (Fig. 2). RBCs from normal individuals (HbAA genotype) were also tested and again showed a similar Ca2+ dependency (Fig. 2). In these experiments, PS exposure was greatest in RBCs from normal (HbAA) individuals (p < 0.01), perhaps because of greater damage from the presence of HbS in RBCs from SCD patients which, amongst other effects, may affect scrambling activity. At the highest [Ca2+ ] (>1 mM), PS exposure was sometimes reduced, as observed previously in lymphocytes and RBCs [23,24]. In all these experiments, there were two apparent components of PS scrambling, one requiring submicromolar [Ca2+ ]s and complete by about 10 ␮M, the other requiring 10–100-fold greater concentrations. In HbSS cells, the component with higher Ca2+ affinity had an EC50 of 1.8 ± 0.3 ␮M; the low affinity component had an EC50 of 306 ± 123 ␮M (both means ± S.E.M., n = 5). These different components may reflect the likely existence of more than one RBC scramblase enzyme [4] with different Ca2+ dependences, although other explanations are possible (see Section 4). The low affinity component, although always present, was sometimes small (e.g. for normal RBCs in Fig. 2 or for sickle cells in Fig. 4). The greater proportion of RBCs (≥2/3rds) showed PS scrambling at the lower Ca2+ concentrations.

53

100

(a)

80

60

40

HbSS HbAA HbSC

20

0 10 0

10 1

10 2

10 3

10 4

2+

Extracellular Ca (µM)

Percentage of RBCs with externalised phosphatidylserine

In RBCs from HbSC heterozygotes, PS exposure was also higher than for RBCs from normal HbAA individuals, but at about half the level observed for HbSS patients, ranging from 0.2 to 5.4%, mean 2.2 ± 0.3% (n = 29), p < 0.02. Deoxygenation has been previously shown to induce PS exposure, although experiments are rarely carried out in salines with physiological [Ca2+ ]o s (e.g. see [20,21]). We therefore investigated the effect of 80 min deoxygenation on RBCs from both major SCD genotypes incubated in LK saline either in the absence of Ca2+ (with addition of 2 mM EGTA) or in saline containing Ca2+ at normal free plasma levels (1.1 mM; Fig. 1). When RBCs were deoxygenated in the absence of extracellular Ca2+ , there was little change in PS exposure. Similar findings occurred in oxygenated RBCs at 1.1 mM Ca2+ . Deoxygenation in 1.1 mM Ca2+ , however, stimulated a progressive externalisation of PS. At 5 mM [Ca2+ ]o , PS exposure in RBCs from HbSS patients was consistently reduced compared to values obtained with 1.1 mM Ca2+ . For example, after 1 h, PS exposure was 8.3 ± 1.5% falling to 7.3 ± 0.8% in RBCs incubated at 1.1 mM [Ca2+ ]o or 5 mM, respectively. In this context, it is interesting to note that deoxygenation-induced cation fluxes have been previously reported to be inhibited by extracellular [Ca2+ ] over 2 mM [22]. These experiments showed that PS exposure was Ca2+ -dependent and deoxygenation-induced, but that deoxygenation alone, with the concomitant sickling shape change, was insufficient to elicit scrambling.

Percentage of RBCs with externalised phosphatidylserine

E. Weiss et al. / Cell Calcium 51 (2012) 51–56

(b)

100

*

80

*

*

*

60

40

*

HbSS (n=10) HbAA (n=4) HbSC (n=5)

20

0

* 10 0

10 1

10 2

10 3

10 4

Extracellular Ca2+ (µM) Fig. 2. Ca2+ dependence of phosphatidylserine (PS) exposure. RBCs were incubated in high potassium (HK)-containing saline (90 mM K+ ) with bromo-A23187 and vanadate (1 mM) and different free [Ca2+ ]o s for 30 min. HK saline prevents K+ loss following activation of the Gardos channel, thus removing secondary effects due to cell shrinkage. In all cases, extracellular [Mg2+ ] was 0.15 mM. RBCs from three genotypes were examined, those from normal HbAA individuals and those from SCD patients of both HbSS and HbSC genotype. RBC aliquots were then taken, external PS labelled with FITC-lactadherin and percentage of positive RBCs determined by FACS. (a) Single representative experiment; (b) means ± S.E.M. *Significant difference (p < 0.01) between percentage PS exposure in RBCs from normal individuals (HbAA) and homozygous (HbSS) SCD patients.

3.3. The effect of vanadate on phosphatidylserine exposure These findings appear at odds with most reports in the literature which generally hold that Ca2+ -induced scrambling occurs at low affinity (20–100 ␮M are often quoted e.g. [4,13]). In the above experiments, vanadate (1 mM) was present during Ca2+ loading. It was possible that the presence of vanadate may have altered the apparent Ca2+ affinity [25]. The Ca2+ dependence of PS exposure was therefore determined in RBCs which were not exposed to vanadate until addition of lactadherin. The pattern of Ca2+ -induced PS scrambling in the absence of vanadate was very similar to that observed in its presence (Fig. 3), although perhaps occurring at slightly lower Ca2+ affinity, which may result from continued activity of the flippase or the plasma membrane Ca2+ pump.

100

75

50

25

+ Vanadate - Vanadate

0 10

0

10

1

10

2

10

3

Percentage of RBCs with externalised phosphatidylserine

E. Weiss et al. / Cell Calcium 51 (2012) 51–56

Percentage of RBCs with externalised phosphatidylserine

54

100

80

60

[Mg2+]o 0.15mM

40

0.3mM 20

1.0mM 2.0mM

0 10 0

10 1

10 2

10 3 2+

Extracellular Ca

Extracellular Ca2+ (µM)

10 4

(µM)

Fig. 3. The effect of vanadate on Ca2+ dependence of phosphatidylserine exposure. Experiments were carried out as for Fig. 2, but in the presence or absence of vanadate (1 mM) during incubation with bromo-A23187. RBCs were from patients homozygous for HbS (HbSS). Data represent means ± S.E.M. for RBCs from 5 patients.

Fig. 5. The effect of magnesium on Ca2+ dependence of phosphatidylserine exposure. Experiments were carried out as for Fig. 2, except that [Mg2+ ]o was either 0.15, 0.3, 1 or 2 mM. RBCs were from patients homozygous for HbS (HbSS). Single experiment representative of 4 others.

3.4. The effect of potassium on phosphatidylserine exposure Aside from Fig. 1, experiments were also carried out at high [K+ ]o , to prevent secondary RBC shrinkage following Ca2+ -induced activation of the Gardos channel. It has been suggested that K+ interferes with scrambling [26], although high K+ is reported to reduce PS exposure and is therefore less likely to stimulate artefactually a higher Ca2+ affinity for scrambling. To test whether high K+ had an effect on PS scrambling, the experiments were repeated at physiological K+ and Na+ (Fig. 4). If anything, PS exposure was slightly increased in the LK saline, perhaps following RBC shrinkage [27,28] – but the high affinity component was still apparent.

levels was investigated (Fig. 5). At [Mg2+ ]i s present in intact normal RBCs (given by [Mg2+ ]o of 0.15 and 0.3 mM which clamp [Mg2+ ]i at the physiological levels of 0.3 mM in oxygenated RBCs and 0.6 mM in deoxygenated ones [18]), addition of Mg2+ had no effect on PS exposure. As [Mg2+ ] was raised to 1 mM, however, the high Ca2+ affinity component of PS scrambling became obscured such that the overall Ca2+ affinity fell to levels close to the value of around 100 ␮M given in the literature. For example, at [Ca2+ ] of 10 ␮M, PS exposure fell from 67 ± 8% at 0.15 mM Mg2+ to 40 ± 14% at 2 mM (both n = 5; p < 0.05). Higher concentrations of Mg2+ did not alter this pattern. Nevertheless it appears from these results that unphysiologically high [Mg2+ ]i inhibits Ca2+ -induced PS scrambling.

3.5. Effect of Mg2+ on Ca2+ -induced phosphatidylserine exposure

4. Discussion

Percentage of RBCs with externalised phosphatidylserine

As Mg2+ often competes with Ca2+ in cellular reactions and has been shown to slow lipid reorientation [29], in the final series of experiments, the effect of altering RBC Ca2+ at different [Mg2+ ]

100

75

50

25

HK LK

0 10 0

10 1

10 2

10 3 2+

Extracellular Ca

10 4

(µM)

Fig. 4. The effect of potassium on Ca2+ dependence of phosphatidylserine exposure. Experiments were carried out as for Fig. 2, but in either high-potassium (HK)containing saline (90 mM) or low-potassium (HK)-containing saline (4 mM). RBCs were from patients homozygous for HbS (HbSS). Data represent means ± S.E.M. for RBCs from 5 patients.

The present findings confirm that PS scrambling in intact RBCs from SCD patients upon deoxygenation required extracellular Ca2+ . On clamping intracellular Ca2+ (using the A23187 method), there appeared to be two components of Ca2+ -induced scrambling, one apparent at a few micromolar [Ca2+ ], the other requiring some two orders of magnitude higher levels. The higher affinity component predominated and probably accounts for most scrambling in vivo. The presence of unphysiologically high levels of Mg2+ (1 mM outside) resulted in loss of this high affinity component and may account for it being overlooked hitherto. RBCs from SCD patients are known to show high levels of PS exposure. Our results confirmed this for both main SCD genotypes, HbSS and HbSC. Deoxygenation is well recognised as a trigger resulting in higher levels of exposure (e.g. see [20,21]) but the mechanism is unclear. We show that deoxygenation per se in the absence of extracellular Ca2+ had no effect on PS scrambling. Thus although deoxygenation-induced HbS polymerisation and the sickling shape change would disrupt any stabilisation of lipid asymmetry mediated by the RBC cytoskeleton [30–32], this in itself was insufficient to cause PS scrambling. Rather the presence of Ca2+ was also required. Previous reports suggest that intracellular [Ca2+ ] of around 100 ␮M are required for scramblase activation. Deoxygenation of RBCs from SCD patients perturbs Ca2+ homeostasis. It results in activation of a deoxygenation-induced cation conductance, Psickle , permeable to Ca2+ [33]. There is also a modest inhibition of the plasma membrane Ca2+ pump (PMCA) [11]. Notwithstanding, it is

E. Weiss et al. / Cell Calcium 51 (2012) 51–56

unlikely that these effects are able to raise intracellular [Ca2+ ] to the high levels required to stimulate scrambling. Mean values of about 30 nM have been reported [11] and even though it is likely that subpopulations achieve higher levels of Ca2+ , values of some 100 ␮M are likely too high. How then can Ca2+ participate in the scrambling process? We used A23187 to clamp intracellular [Ca2+ ] across 6 orders of magnitude to investigate the role of Ca2+ . Findings were consistent with the presence of two components of PS scrambling, occurring with high (around a few micromolar) and low (several hundred micromolar) affinity scrambing. This was evident in RBCs from all three genotypes examined (HbAA, HbSS and HbSC). The component of scrambling seen at low Ca2+ levels predominated. These results appear at variance with previous findings which indicate the Ca2+ -induced PS scrambling requires some 20–100 ␮M [4,13]. These reported values, however, rely on a relatively limited number of publications. Some of these concern purified PLSCR1 in liposomes [34,35] – the RBC scramblase and PLSCR1 are now considered to be different entities [4,36]. In one study on a leukaemic cell line damaged through radiation with ultraviolet light, the EC50 for intracellular Ca2+ (as opposed to the extracellular concentration in which cells were suspended) was also very low, and probably submicromolar (see Figures 5 and 6 in Ref. [37]), more in line with our findings. Others reports involve ghosts rather than intact RBCs [34,38]. In several studies, furthermore, experiments were carried out in the presence of high concentrations of extracellular Mg2+ (1 mM) [38–40]. In RBCs from normal individuals (HbAA), free [Mg2+ ]i oscillates between about 0.3 mM when oxygenated, rising to about 0.6 mM when deoxygenated [18]. Mg2+ levels in RBCs from SCD patients are likely to be lower [2,41,42]. An extracellular Mg2+ of 1 mM in the presence of the ionophore A23187 would set intracellular Mg2+ at over 2 mM, considerably in excess of these normal intracellular Mg2+ levels. Results of Fig. 5 show that these high [Mg2+ ]s dampen the extent of PS scrambling in response to Ca2+ . It appears therefore that high [Mg2+ ]i inhibits Ca2+ -induced PS scrambling and its presence may explain why higher affinities have not been observed previously. The mechanism responsible for the two apparent components of Ca2+ -induced PS scrambling requires further consideration. We speculated that there may be more than one enzyme involved in PS transport, for example the different ABC transporters reported to be active in the membrane of RBCs [43]. Other possibilities are actions of different second messengers, perhaps themselves with different Ca2+ affinities which impact on a single scrambling process. Amongst these, PKC, caspases and ceramide are obvious candidates [2,44,45]. This aspect is currently under investigation.

5. Conclusion Results shown here reveal an hitherto overlooked high affinity component of Ca2+ -induced RBC PS scrambling at physiologically relevant levels. Experimental conditions, in particular the presence of high Mg2+ , may explain why this component has not been identified previously. Findings apply equally to both main SCD genotypes (HbSS and HbSC) and also to normal (HbAA) individuals. It seems more appropriate that Ca2+ levels required for APLT inhibition and PS scrambling are similar, thus facilitating the co-ordination of a number of eryptotic events [2]. The scrambling activity of RBCs persists for several hours following Ca2+ elevation [40]. Our findings suggest that perturbed Ca2+ homeostasis in RBCs from SCD patients, even if short-lived following transient deoxygenation, may contribute to PS exposure and play a role in pathogenesis of the condition. Notwithstanding, whether Ca2+ directly stimulates

55

scramblase activity, or is mediated through a second-messenger pathway, remains to be established [4]. Conflict of interest The authors declare no competing financial interests. Acknowledgements We thank the British Heart Foundation, the Medical Research Council and the BBSRC for financial support. References [1] F.A. Kuypers, Red cell membrane lipids in hemoglobinopathies, Current Molecular Medicine 8 (2008) 633–638. [2] F. Lang, et al., Mechanisms and significance of eryptosis, Antioxidants and Redox Signalling 8 (2006) 1183–1192. [3] K. de Jong, et al., Characterization of the phosphatidylserine-exposing subpopulations of sickle cells, Blood 98 (2001) 860–867. [4] E.M. Bevers, P.L. Williamson, Phospholipid scrambling: an update, FEBS Letters 584 (2010) 2724–2730. [5] C.W.M. Haest, Distribution and movement of membrane lipids, in: I. Bernhardt, J.C. Ellory (Eds.), Red Cell Membrane Transport in Health and Disease, Springer Verlag, Berlin, 2003, pp. 1–25. [6] L.A. Barber, et al., Aminophospholipid translocase and phospholipid scramblase activities in sickle erythrocyte subpopulations, British Journal of Haematology 146 (2009) 447–455. [7] R.P. Hebbel, et al., Spontaneous oxygen radical generation by sickle erythrocytes, Journal of Clinical Investigation 70 (1982) 1253–1259. [8] R.P. Hebbel, et al., Accelerated autoxidation and heme loss due to instability of sickle hemoglobin, Proceedings of the National Academy of Sciences of the United States of America 85 (1988) 237–241. [9] V.L. Lew, R.M. Bookchin, Ion transport pathology in the mechanism of sickle cell dehydration, Physiological Reviews 85 (2005) 179–200. [10] C.H. Joiner, Cation transport and volume regulation in sickle red blood cells, American Journal of Physiology 264 (1993) C251–C270. [11] Z. Etzion, et al., Effects of deoxygenation on active and passive Ca2+ transport and on the cytoplasmic Ca2+ levels of sickle cell anemia red cells, Journal of Clinical Investigation 92 (1993) 2489–2498. [12] M. Bitbol, et al., Ion regulation of phosphatidylserine and phosphatidylethanolamine outside–inside translocation in human erythrocytes, Biochimica et Biophysica Acta 904 (1987) 268–282. [13] D. Kamp, T. Sieberg, C.W.M. Haest, Inhibition and stimulation of phospholipid scrambling activity. Consequences for lipid asymmetry, echinocytosis, and microvesiculation of erythrocytes, Biochemistry 40 (2001) 9438–9446. [14] T. Tiffert, et al., Effects of deoxygenation on active and passive Ca2+ transport and cytoplasmic Ca2+ buffering in normal human red cells, Journal of Physiology 464 (1993) 529–544. [15] T. Tiffert, J.L. Spivak, V.L. Lew, Magnitude of calcium influx required to induce dehydration of normal human red cells, Biochimica et Biophysica Acta 943 (1988) 157–165. [16] F. Lang, et al., Cation channels, cell volume and the death of an erythrocyte, European Journal of Physiology 447 (2003) 121–125. [17] K.S. Lang, et al., Involvement of ceramide in hyperosmotic shock-induced death of erythrocytes, Cell Death and Differentiation 11 (2004) 231–243. [18] P.W. Flatman, The effect of buffer composition and deoxygenation on the concentration of ionized magnesium inside human red blood cells, Journal of Physiology 300 (1980) 19–30. [19] M.C. Muzyamba, E.H. Campbell, J.S. Gibson, Effect of intracellular magnesium and oxygen tension on K+ –Cl− cotransport in normal and sickle human red cells, Cellular Physiology and Biochemistry 17 (2006) 121–128. [20] B. Lubin, et al., Abnormalities in membrane phospholipid organization in sickled erythrocytes, Journal of Clinical Investigation 67 (1981) 1643–1649. [21] N. Blumenfeld, et al., Transmembrane mobility of phospholipids in sickle erythrcoytes: effect of deoxygenation on diffusion and asymmetry, Blood 77 (1991) 849–854. [22] V.L. Lew, Z. Etzion, R.M. Bookchin, Dehydration response of sickle cells to sickling-induced Ca2+ permeabilization, Blood 99 (2002) 2578–2585. [23] P. Williamson, et al., Phospholipid scramblase activation pathways in lymphocytes, Biochemistry 40 (2001) 8065–8072. [24] K. de Jong, F.A. Kuypers, Sulphydryl modifications alter scramblase activity in murine sickle cell disease, British Journal of Haematology 133 (2006) 427–432. [25] M. Foller, et al., Vanadate-induced suicidal erythrocyte death, Kidney and Blood Pressure Research 31 (2008) 87–93. [26] J.L.N Wolfs, et al., Direct inhibition of phospholipid scrambling activity in erythrocytes by potassium ions, Cellular and Molecular Life Sciences 66 (2009) 314–323. [27] K.S. Lang, et al., Enhanced apoptosis in sickle cell anemia, thalassemia and glucose-6-phosphate dehydrogenase deficiency, Cellular Physiology and Biochemistry 12 (2002) 365–372.

56

E. Weiss et al. / Cell Calcium 51 (2012) 51–56

[28] J. Schneider, et al., Suicidal erythrocyte death following cellular K+ loss, Cellular Physiology and Biochemistry 20 (2007) 35–44. [29] U. Henseleit, G. Plasa, C.W.M. Haest, Effects of divalent cations on lipid flipflop in the human erythrocyte membrane, Biochimica et Biophysica Acta 1029 (1990) 127–135. [30] C.W.M. Haest, et al., Spectrin as a stabilizer of the phospholipid asymmetry in the human erythrocyte membrane, Biochimica et Biophysica Acta 509 (1978) 21–32. [31] P.F.H. Franck, et al., Uncoupling of the membrane skeleton from the lipid bilayer, Journal of Clinical Investigation 75 (1985) 183–190. [32] E. Middelkoop, et al., Studies on sickle erythrocytes provide evidence that the asymmetric distribution of phosphatidylserine in the red cell membrane is maintained by both ATP-dependent translocation and interaction with membrane skeletal proteins, Biochimica et Biophysica Acta 937 (1988) 281–288. [33] M.D. Rhoda, et al., Ca2+ permeability in deoxygenated sickle cells, Blood 75 (1990) 2453–2458. [34] F. Basse, et al., Isolation of an erythrocyte membrane protein that mediates Ca2+ -dependent transbilayer movement of phospholipid, Journal of Biological Chemistry 271 (1996) 17205–17210. [35] J.G. Stout, et al., Change in conformation of plasma membrane phospholipid scramblase induced by occupancy of its Ca2+ binding site, Biochemistry 37 (1998) 14860–14866. [36] Q. Zhou, et al., Normal hemostasis but defective hematopoietic response to growth factors in mice deficient in phospholipid scramblase, Blood 99 (2002) 4030–4038. [37] D.L. Bratton, et al., Appearance of phosphatidylserine on apoptotic cells requires calcium-mediated nonspecific flip-flop and is enhanced by loss of

[38] [39]

[40]

[41]

[42]

[43]

[44]

[45]

the aminophospholipid translocase, Journal of Biological Chemistry 272 (1997) 26159–26165. L.A. Woon, et al., Ca2+ sensitivity of phospholipid scrambling in human red cell ghosts, Cell Calcium 25 (1999) 313–320. B. Verhoven, R.A. Schlegel, P. Williamson, Rapid loss and restoration of lipid asymmetry by different pathways in resealed erythrocyte ghosts, Biochimica et Biophysica Acta 1104 (1992) 15–23. P. Williamson, et al., Ca2+ induces transbilayer redistribution of all major phospholipids in human erythrocytes, Biochemistry 31 (1992) 6355–6360. O.E. Ortiz, V.L. Lew, R.M. Bookchin, Deoxygenation permeabilizes sickle cell anaemia red cells to magnesium and reverses its gradient in the dense cells, Journal of Physiology 427 (1990) 211–226. J.P. Willcocks, et al., Simultaneous determination of low free Mg2+ and pH in human sickle cells using 31P NMR spectroscopy, Journal of Biological Chemistry 277 (2002) 49911–49920. K. Köck, M. Grube, G. Jedlitschky, L. Ovevermann, W. Siegmund, C.A. Ritter, H.K. Kroemer, Expression of adenosine triphosphate-binding cassette (ABC) drug transporters in peripheral blood cells: relevance for physiology and pathology, Clinical Pharmacokinetics 46 (2007) 449–470. D. Mandal, A. Mazumder, P. Das, M. Kindu, J. Basu, Fas-, caspase 8-, and caspase 3-dependent signalling regulates the activity of the aminophospholipid translocase and phosphatidylserine externalization in human erythrocytes, Journal of Biological Chemistry 280 (2005) 39460–39467. B.A Klarl, P.A. Lang, D.S. Kempe, O.M. Niemoeller, A. Akel, M. Sobiesiak, K. Eisele, M. Podolski, S.M. Huber, T. Wieder, F. Lang, Protein kinase C mediates erythrocyte “programmed cell death” following glucose depletion, American Journal of Physiology 290 (2006) C244–C253.