Ribavirin-induced externalization of phosphatidylserine in erythrocytes is predominantly caused by inhibition of aminophospholipid translocase activity

European Journal of Pharmacology 693 (2012) 1–6

Contents lists available at SciVerse ScienceDirect

European Journal of Pharmacology journal homepage: www.elsevier.com/locate/ejphar

Molecular and cellular pharmacology

Ribavirin-induced externalization of phosphatidylserine in erythrocytes is predominantly caused by inhibition of aminophospholipid translocase activity Marie-Claire Kleinegris a, Ger H. Koek b, Kelly Mast a, Eveline H.C. Mestrom a, Jef L.N. Wolfs a, Edouard M. Bevers a,n a b

Department of Biochemistry, Cardiovascular Research Institute Maastricht (CARIM), Maastricht University, The Netherlands Section of Gastroenterology/Hepatology, Department of Internal Medicine, University Hospital Maastricht, The Netherlands

a r t i c l e i n f o

abstract

Article history: Received 17 April 2012 Received in revised form 5 July 2012 Accepted 11 July 2012 Available online 21 August 2012

Ribavirin in combination with interferon-a is the standard treatment for chronic hepatitis C, but often induces severe anemia forcing discontinuation of the therapy. Whereas suppression of bone marrow by interferon may impact on the production of erythrocytes, it has been suggested that accumulation of ribavirin in erythrocytes induces alterations causing an early removal of these cells by the mononuclear phagocytic system. Externalization of phosphatidylserine, which is exclusively present in the cytoplasmic leaflet of the plasma membrane, is a recognition signal for phagocytosis in particular of apoptotic cells. Here, we demonstrate that surface exposure of phosphatidylserine upon prolonged treatment of erythrocytes with ribavirin results mainly from inactivation of the aminophospholipid translocase, an ATP-dependent lipid pump, which specifically transports phosphatidylserine from the outer to the inner leaflet of the plasma membrane. Inactivation is due to severe ATP depletion, although competitive inhibition by ribavirin or its phosphorylated derivatives cannot be excluded. Phospholipid scramblase, responsible for collapse of lipid asymmetry, appears to be of minor importance as erythrocytes of patients with the Scott syndrome, lacking Ca2 þ -induced lipid scrambling, are equally sensitive to ribavirin treatment. Neither the antioxidant N-acetylcysteine nor the pan-caspase inhibitor Q-VD-OPH did affect ribavirin-induced phosphatidylserine exposure, suggesting that oxidative stress or apoptotic-related mechanisms are not involved in this process. In conclusion, we propose that spontaneous loss of lipid asymmetry, not corrected by aminophospholipid translocase activity, is the mechanism for ribavirin-induced phosphatidylserine exposure that may contribute to ribavirininduced anemia. & 2012 Elsevier B.V. All rights reserved.

Keywords: Hepatitis C treatment Red blood cell Phosphatidylserine exposure Lipid asymmetry Scramblase Anemia

1. Introduction Worldwide, over 170 million people are infected with hepatitis C virus (HCV) and each year, this number increases by 3–4 million (Alter, 2007; Szabo et al., 2003). The infection progresses to a chronic form in more than 80% of the affected individuals. The standard treatment for the chronic form of HCV infection is a combination therapy with pegylated-interferon-a and ribavirin, which was found to successfully eradicate the virus in 40–90% of the patients (Webster et al., 2009). Several mechanisms of action of each of these drugs have been proposed, but are not subject of this study. The most common and significant adverse effect of ribavirin treatment is severe hemolytic anemia requiring discontinuation of n Correspondence to: Department of Biochemistry (CARIM), Maastricht University, PO Box 616, 6200 MD Maastricht, The Netherlands. Tel.: þ31 43 3881 687; fax: þ 31 43 3884 159. E-mail address: [email protected] (E.M. Bevers).

0014-2999/$ - see front matter & 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.ejphar.2012.07.041

the treatment in about 10% of the patients (Russmann et al., 2006; van Soest et al., 2009). The mechanism by which ribavirin causes anemia is largely unknown. Ribavirin is taken up by the erythrocyte via an es-type nucleoside transporter (Glue, 1999). Inside the cell, ribavirin is rapidly phosphorylated to ribavirin-triphosphate (Page and Connor, 1990), which accumulates in the cell due to lack of dephosphorylating enzymes. It has been suggested that the accumulating ribavirin-triphosphate leads to energy depletion and consequently a higher susceptibility of the cells to oxidative stress (De Franceschi et al., 2000), resulting in changes in membrane fluidity causing extravascular hemolysis and removal by the mononuclear phagocytic system. Alternatively, as was recently demonstrated by Homma et al. (2009), uptake and phosphorylation of ribavirin causes phosphatidylserine to appear in the outer leaflet of the erythrocyte membrane. An increased surface exposure of phosphatidylserine forms a signal for erythrophagocytosis leading to a reduced number of circulating cells (Fadok et al., 2000; McEvoy et al., 1986; Schroit et al., 1985).

2

M.-C. Kleinegris et al. / European Journal of Pharmacology 693 (2012) 1–6

The erythrocyte membrane consists of a bilayer in which the various phospholipids are non-randomly distributed over both membrane leaflets (for recent reviews, consult Daleke, 2008; Zwaal et al., 2005): the outer leaflet consists mainly of cholinecontaining phospholipids, phosphatidylcholine and sphingomyelin, whereas the aminophospholipids, phosphatidylserine and phosphatidylethanolamine, are virtually exclusively localized in the inner leaflet. This asymmetric distribution results from the concerted action of an aminophospholipid translocase, a P-type ATPase, which actively pumps phosphatidylserine and phosphatidylethanolamine from the outer leaflet to the inner leaflet at the expense of ATP and a floppase, which pumps all phospholipid classes from the inner to the outer leaflet, though at a much smaller rate. The latter protein in erythrocytes was identified as a multidrug resistance associated protein ABC C1 (MRP1), (reviewed by Folmer et al., 2009). Loss of lipid asymmetry may occur upon activation of phospholipid scramblase, which shuttles phospholipids between both monolayers in a non-specific and energy-independent way (reviewed by Bevers and Williamson, 2010). This may occur in pathological red blood cells or upon artificially increasing the intracellular Ca2 þ concentration by means of selective Ca2 þ -ionophores (Zwaal et al., 2005). The aim of the present study was to investigate the mechanism(s) of ribavirin-induced phosphatidylserine exposure by studying the effect of this drug on the above mentioned lipid transporters.

2. Materials and methods 2.1. Materials Ribavirin, ionomycin and bovine serum albumin were obtained from Sigma (St. Louis, MO) and Luciferin–Luciferase reagent from Chrono-Log Corporation (Havertown). Fluorescent labeled annexin A5 (Annexin A5-FITC) was from Invitrogen (Leiden). Fluorescent labeled phosphatidylserine, 1-palmitoyl-2-{12-[(7-nitro-2-1,3-benzoxadiazol-4-yl)amino]lauroyl)-sn-glycero-3-phosphoserine (NBD-phosphatidylserine), was from Avanti Polar-lipids, Inc. Caspase inhibitor Q-VD-OPH was obtained from Enzo Life Sciences. 2.2. Isolation of erythrocytes and ribavirin treatment Human blood was obtained from healthy volunteers, after informed consent, by venipuncture and collected in citrate 3.2% (ratio 9:1). After removal of the plasma and buffy coat, the erythrocytes were washed three times with HEPES buffer at pH 7.4 (10 mMHEPES, 136 mMNaCl, 2.7 mMKCL, 2 mMMgCl2,) at room temperature and subsequently centrifuged at 1000g for 5 min. Washed cells were stored in HEPES-buffer at a concentration of 1  108 ml  1, in the presence of 5 mM glucose and 50 units/ml penicillin-G and 50 mg/ml streptomycin. Erythrocytes were incubated for 48 h at 37 1C with ribavirin at final concentrations as indicated in the text in the presence of 1 mM CaCl2 or 1 mM EGTA. 2.3. Ionomycin treatment In the presence of extracellular Ca2 þ ions, Ionomycin causes maximal phosphatidylserine-exposure in erythrocytes. To be sure of an optimal distribution of Ionomycin in the erythrocyte suspension, 5 mM Ionomycin was diluted in HEPES buffer to a concentration of 0.1 mM. Every first day of the experiment, 0.5 ml of untreated cells, suspensions without ribavirin and in the presence of 1 mM CaCl2, were extracted and incubated for

20 min with 15 ml HEPES/Ionomycin (final concentration 3 mM) at 37 1C. Phosphatidylserine-exposure was determined at different time-points; these were chosen in order to measure the initial rates of phosphatidylserine exposure. 2.4. Measurement of phosphatidylserine exposure Surface-exposed phosphatidylserine was measured using annexin A5-FITC: 4 ml samples were added to 200 ml HEPESbuffer containing 3 mM CaCl2 and incubated with 2 ml A5 on ice. After 5 min, phosphatidylserine-exposure was measured by flow cytometry using an Accuri C6 Flow Cytometer. The annexin A5 fluorescence intensity was measured in fluorescence channel FL1-H (excitation 488 nm, emission 533/30 nm). Data was analyzed for 10,000 events per aliquot using the CFlows software. We determined annexin A5 positive cells as the percentage of cells that exceed the marker set to include 99% of the untreated erythrocytes at day 0. 2.5. Measurement of aminophospholipid translocase activity To measure translocase activity, 0.5 ml samples of the suspensions were incubated at 37 1C with 0.5 mM NBD-phosphatidylserine, added from a 1 mmol/l stock solution in HEPES-buffer. At consecutive moments in time, 5 ml of the sample was added to 250 ml HEPES buffer containing 0.1 mM EGTA and 1% bovine serum albumin and another 5 ml sample was added to 250 ml HEPES-buffer, with 0.1 mM EGTA without 1% bovine serum albumin. Bovine serum albumin rapidly extracts lipid probe from the outer plasma membrane, thus the amount of NBDphosphatidylserine localized on the inner leaflet and the total amount of NBD-phosphatidylserine in the plasma membrane can be determined. The samples were analyzed by flow cytometry. 2.6. Measurement of intracellular ATP Cells were lysed by adding 150 ml cells (109 ml  1) to 850 ml ice-cold distilled water for 5 min. To measure ATP, 200 ml of the lysate was added to a cuvette containing 300 ml HEPES buffer at 37 1C in a lumi-aggregometer. 50 ml reconstituted luciferine– luciferase reagent (Chronolume) was added and the fluorescence intensity was measured using a Lumi aggregometer (Chronolog Corporation). The ATP-content of freshly obtained erythrocytes was set at 100%. 2.7. Measurement of hemolysis Samples (100 ml) from the incubations at days 1 and 2 were centrifuged at 1000g for 10 s and 25 ml of the supernatant was added to 1 ml distilled water. The Hb-concentration was measured spectrophotometrically at 415 nm. The absorption of 25 ml of the erythrocyte suspension lysed in distilled water was defined as 100% hemolysis.

3. Results 3.1. Ribavirin causes a time dependent increase in surface exposed phosphatidylserine Plasma steady state concentrations of ribavirin in patients under treatment have been estimated around 10 mM (Inoue et al., 2006). Incubation of washed erythrocytes at this concentration ribavirin for more than a week at ambient temperature did not induce significant phosphatidylserine exposure (data not shown), likely because this period is too short to accumulate sufficient

M.-C. Kleinegris et al. / European Journal of Pharmacology 693 (2012) 1–6

80

3

100

70 Annexin A5 positive cells (%)

Annexin A5 positive cells (%)

80 60 50 40 30 20

60

40

20 10 0 0h

24 h

48 h

Fig. 1. Ribavirin-induced phosphatidylserine exposure in human erythrocytes. Washed cells at a concentration of 108 ml  1 in the presence of 1 mM CaCl2 and 5 mM glucose were incubated with ribavirin and phosphatidylserine exposure was measured using fluorescent-labeled annexin A5. Percentage annexin positive cells in the absence (white bars) or the presence of 0.1 mM (hatched bars), 1 mM (grey bars) and 5 mM ribavirin (black bars) after 24 and 48 h at 37 1C. Data are mean values 7S.D. (A, n ¼15 and B, n¼ 5).

ribavirin and ribavirin-phosphorylated compounds. In order to observe effects of ribavirin within a few days, experiments were performed at 37 1C at increased ribavirin concentrations. As shown in Fig. 1, at a concentration of 100 mM ribavirin, 371% annexin A5 positive cells were found after 24 h, not significantly different from the controls without ribavirin; after 48 h, however, the number of annexin A5 positive cells increased to 2477% compared to 874% for the control cells. Incubation at 1 mM ribavirin for 24 and 48 h at 37 1C increased the percentage of annexin A5 positive erythrocytes from 11 74% to 54711%, respectively. Raising the ribavirin concentration to 5 mM did not significantly increase the number of annexin positive cells, indicating that the rate of inward transport is virtually saturated at 1 mM ribavirin, in agreement with an estimated Km of 0.44 mM (Jarvis et al., 1998). The relatively large standard deviations in these experiments are due to the variability in response between different donors. The increased number of positive cells did not result from hemolysis, which would allow annexin A5 to gain access to the cytoplasmic leaflet of the membrane: cell lysis increased from 3% to 4% in the absence and from 4% to 6% in the presence of ribavirin after 24 and 48 h, respectively. These results confirm and extent data by Homma et al. (2009), who found 2.15% annexin positive cells after 18 h incubation with 1 mM ribavirin. 3.2. Effect of ribavirin on phosphatidylserine exposure is related to the intracellular ATP content As shown by De Franceschi et al. (2000), after 12 h incubation with 1 mM ribavirin, cellular ATP levels are markedly reduced, which may compromise mechanisms required for maintenance of membrane lipid asymmetry. To find support for this hypothesis, we measured the effect of ribavirin on phosphatidylserine exposure in the presence and absence of glucose. As shown in Fig. 2, prolonged glucose deprivation by itself already causes a significant increase in the number of phosphatidylserine exposing cells varying from 974% after 24 h (not shown) to 54716% at 48 h. Under these conditions, the presence of 1 mM ribavirin has a small additional effect on cells,

0 -glucose

+ glucose

Fig. 2. Effect of glucose on ribavirin-induced phosphatidylserine exposure in human erythrocytes.Washed cells were incubated at 37 1C with (black bars) and without (white bars) 1 mM ribavirin in the presence and the absence of 5 mM glucose. After 48 h the percentage of annexin positive cells was determined by flow cytometry. Data are mean values7 S.D. (n ¼6).

leading to 1376% phosphatidylserine positive cells at 24 h (not shown) and 60718% after 48 h incubation, respectively. In glucose containing medium, only 674% of the erythrocytes expose phosphatidylserine after 48 h at 37 1C, indicating that lipid asymmetry is maintained in the majority of cells. Despite the presence of 5 mM glucose, however, 43722% of the cells became annexin A5 positive after 48 h incubation at 1 mM ribavirin. This strongly suggests that the major effect of ribavirin relates to a drop in intracellular ATP, in agreement with the polyphosphorylation that occurs after ribavirin uptake by the cells. Indeed, incubation with 1 mM ribavirin for 24 h decreased the ATP content to less than 6% compared to 79% in the absence of ribavirin (Fig. 3). Prolonged incubation to 48 h reduced these values to 2% and 5% for the ribavirin-treated cells and the controls, respectively. Obviously, between 24 and 48 h, even in the presence of glucose, cells have lost most of their ATP contents. The same results were obtained when cells were incubated with ribavirin in the presence of 20 mM glucose (data not shown), suggesting that the uptake and phosphorylation of ribavirin not only consumes ATP, but also hampers its production. 3.3. Ribavirin treatment causes a decrease in aminophospholipid translocase activity. Aminophospholipid translocase activity is responsible for maintenance of membrane lipid asymmetry by rapidly transporting phosphatidylserine from the outer to the inner leaflet of the plasma membrane. This activity can be monitored by means of fluorescent labeled phosphatidylserine (NBD-phosphatidylserine), which is readily incorporated in the outer leaflet of the membrane, and when transported to the inner leaflet, is no longer extractable by bovine serum albumin. Thus, an increased residual fluorescence of the cells in time after extraction with bovine serum albumin is interpreted as probe being present in the inner leaflet of the plasma membrane. As shown in Fig. 4 (white bars), 60 min after addition of NBD-phosphatidylserine to fresh erythrocytes 80 77% of the probe has been transported to the inner leaflet. Repeating the same experiment after 24 h incubation in the presence of glucose, 74 78% NBD-phosphatidylserine

4

M.-C. Kleinegris et al. / European Journal of Pharmacology 693 (2012) 1–6

3.4. Effect of ribavirin treatment on scramblase activity

100

residual ATP (%)

80

60

40

20

0 0

10

20

30 time (h)

40

50

Fig. 3. Effect of ribavirin and glucose on the ATP content of human erythrocytes during prolonged incubation. Washed erythrocytes (108 ml  1) were incubated for 48 h at 37 1C and at different time intervals, samples were taken to determine the ATP content as described in the section materials and methods. Conditions used were as follows: 5 mM glucose, no ribavirin (J); no glucose, no ribavirin (K); 5 mM glucose and 1 mM ribavirin (&); no glucose and 1 mM ribavirin (’). ATP content of freshly isolated erythrocytes was set at 100%. Data are mean values 7S.D. (n ¼3).

100

80

60

40

20

0 0h

24 h

48 h

100

80 Annexin A5 positive cells (%)

non-exchangeable NBD-PS at 60 min (%)

non-exchangeable NBD-PS at 5 min (%)

100

Surface exposure of phosphatidylserine may result from activation of phospholipid scramblase, which can be induced by increasing intracellular Ca2 þ as for instance by means of a Ca2 þ -ionophore. We hypothesized that ribavirin-induced ATP depletion causes a loss of plasma membrane Ca2 þ -ATPase activity resulting in a gradual increase of the intracellular Ca2 þ concentration leading to activation of the scramblase. To investigate whether ribavirin-induced phosphatidylserine exposure results from activation of the scramblase, we compared the effect of 48 h ribavirin treatment in the presence and the absence of extracellular Ca2 þ . As shown in Fig. 5A, in the presence of extracellular Ca2 þ , 49714% of the cells were found annexin positive versus 3177% in the absence of Ca2 þ (p ¼0.026). Since the binding of annexin requires Ca2 þ , it cannot be ruled out that phosphatidylserine exposure is induced during the binding assay and that the fraction annexin positive cells found in the absence of extracellular Ca2 þ is overestimated. We therefore used lactadherin, which was shown to bind phosphatidylserine in the absence of Ca2 þ (Shi et al., 2004), but similar results were obtained (data not shown). To exclude the possibility that Ca2 þ -induced lipid scrambling is diminished or lost after 48 h incubation or that accumulated phosphorylated ribavirin metabolites interfere with intracellular Ca2 þ , cells were treated with ionomycin for an additional 60 min (Fig. 5A, gray bar). The result shows that the cells have retained full Ca2 þ -dependent scramblase activity. Together, these data suggest that Ca2 þ -induced lipid scrambling may contribute to some extent, but is not the major cause for loss of lipid asymmetry.

60

40

20

0

0h

24 h

80 * 60

40

20

48 h

Fig. 4. Effect of prolonged ribavirin treatment on aminophospholipid translocase activity in human erythrocytes. Inward transport of phosphatidylserine was measured after 0, 24 and 48 h treatment with ribavirin. To measure time dependent inward transport, 5 and 60 min after addition of NBD-phosphatidylserine (NBD-PS in the figure), samples were taken and extracted with bovine serum albumin to remove non-internalized probe and residual fluorescence intensity was measured by flow cytometry. Left, percentage non-exchangeable (internalized) NBD-phosphatidylserine 5 min after addition of the probe; right, percentage non-exchangeable NBD-phosphatidylserine after 60 min. White bars, control cells without ribavirin; black bars, cells treated with 1 mM ribavirin.

is present in the inner leaflet after 60 min; at 48 h incubation, this value decreased to 55713%, i.e. 68% of the original aminophospholipid translocase activity at the start of the experiment. In the presence of 1 mM ribavirin, however, the amount of NBDphosphatidylserine that can be transported to the inner leaflet within 60 min after 24 and 48 h incubation with ribavirin is reduced to 1879% and 17 76%, respectively. This clearly demonstrates that prolonged incubation in the presence of ribavirin results in loss of aminophospholipid translocase activity.

0 Ca2+ ribavirin

+ +

Ionomycin

-

control + -

+ +

+ +

patient +

+ -

+

-

-

+

Fig. 5. Role of Ca2 þ -activated scramblase in ribavirin-induced phosphatidylserine exposure in human erythrocytes. (A). Effect of extracellular Ca2 þ on ribavirininduced phosphatidylserine exposure. Erythrocytes are incubated with 1 mM ribavirin for 48 h in the presence of 1 mM CaCl2 (white bars) or 1 mM EGTA (black bars) and phosphatidylserine exposure is monitored by flow cytometry using fluorescent-labeled annexin A5. The gray bar shows the percentage of cells that become annexin positive after treatment with ionomycin, demonstrating that scramblase can still be activated in cells treated with ribavirin for 48 h. Data are mean values 7S.D (n¼ 6). *P¼ 0.026. (B). Phosphatidylserine exposure induced by ribavirin in erythrocytes from a patient with the Scott syndrome. Cells were incubated with ribavirin for 48 h in the presence (white bars) and the absence (black bars) of extracellular Ca2 þ and phosphatidylserine exposure was measured by annexin A5 binding. The gray bar in B represents the number of annexin positive cells obtained after ionomycin treatment of cells not incubated with ribavirin, to illustrate the impaired Ca2 þ -induced scrambling in this syndrome. (Values are mean 7 S.D. of one experiment in triplicate.) It should be noted that cell lysis under all conditions remained below 5%.

M.-C. Kleinegris et al. / European Journal of Pharmacology 693 (2012) 1–6

To further explore the contribution of Ca2 þ -induced lipid scrambling in phosphatidylserine exposure caused by ribavirin treatment, we have investigated the effect of ribavirin on erythrocytes from a patient with the Scott syndrome. The Scott syndrome is a rare bleeding disorder, characterized by an impaired Ca2 þ -induced lipid scrambling in all cells of the hematological lineage (Zwaal et al., 2004). Incubation of the patient’s red blood cells with 1 mM ribavirin for 48 h in the presence and the absence of extracellular Ca2 þ resulted in 53 75% and 40 76% annexin positive cells, respectively, similar to the control (Fig. 5B). Treatment with ionomycin in the absence of ribavirin does not result in phosphatidylserine exposure (less than 3% annexin positive cells) (Fig. 5B), indicative of impaired Ca2 þ -induced scrambling. The additional effect of extracellular Ca2 þ during ribavirin treatment on the fraction of phosphatidylserine exposing cells therefore seems unrelated to the scramblase. Increased lipid oxidation has been suggested to result in loss of lipid asymmetry (Kagan et al., 2000; Lang et al., 2003). As shown above, ribavirin treatment results in severe ATP depletion, which may lead to reduced glutathione levels compromising oxidative repair mechanisms. The presence of N-acetylcysteine (1 mM), however, did not reduce the number of phosphatidylserine exposing cells induced after 48 h incubation with ribavirin. Alternatively, a possible role for caspases in the mechanism of ribavirin-induced phosphatidylserine exposure was investigated. Ribavirin did not induce caspase activity; moreover, the presence of 10 mM Q-VD-Oph, an irreversible inhibitor of caspases 3, 8 and 9, during the ribavirin treatment did not significantly change the fraction of phosphatidylserine exposing cells (data not shown).

4. Discussion The present study demonstrates that prolonged incubation of erythrocytes with ribavirin causes inactivation of aminophospholipid translocase, which results in accumulation of phosphatidylserine in the external leaflet of the erythrocyte membrane. Since surface exposed phosphatidylserine is known to provide a signal for the mononuclear phagocyte system to remove cells (Fadok et al., 2000), this effect of ribavirin is likely to contribute to the various degrees of anemia that accompany treatment with this drug. The inhibitory effect of ribavirin on aminophospholipid translocase is likely due to a lack of adequate energy supply, since this protein was demonstrated to require hydrolysable ATP for transport. It may be argued that in vivo the continuous presence of glucose prevents ribavirin-induced ATP depletion. This seems unlikely because no difference in ribavirin-induced phosphatidylserine exposure was observed when 5 mM or 20 mM glucose was present during the experiment. Indeed, lack of ATP caused by ribavirin is the result of impaired glycolysis rather than consumption of ATP due to polyphosphorylation of ribavirin once inside the cell (Hitomi et al., 2011). Alternatively, being a synthetic nucleoside analog, ribavirin or its phosphorylated form could compete with ATP as substrate for the aminophospholipid translocase. This possibility, however, could not be investigated since attempts to restore the cellular ATP content by replenishing glucose after ribavirin treatment did not succeed, likely due to a compromised glycolytic pathway. Previous studies have shown that mere inhibition of aminophospholipid translocase is not sufficient to cause loss of lipid asymmetry in erythrocytes (Bevers et al., 1995; Connor et al., 1992; de Jong et al., 1997). Although this seems in disagreement with the present results, it should be emphasized that in the earlier mentioned studies, either the time was insufficient or the temperature was too low to allow a significant transbilayer lipid migration. Spontaneous transbilayer lipid diffusion is a slow process; in protein-free lipid vesicles half times of phospholipid

5

flip-flop are in the order of days (Kornberg and McConnell, 1971), but the presence of transmembrane proteins such as glycophorin accelerates the rate of lipid diffusion (de Kruijff et al., 1978). Passive transmembrane lipid diffusion in erythrocytes varies on the type of lipid and t1/2 was estimated to vary from 5 to 30 h (Zachowski, 1993). We conclude that a slow spontaneous transbilayer movement of lipids, combined with an inactive aminophospholipid translocase is the most likely explanation for the increased surface exposure of phosphatidylserine that is observed upon prolonged treatment with ribavirin. In the current view, the normal asymmetric lipid distribution is the resultant of a rapid and specific inward transport of aminophospholipids by the translocase, counterbalanced by a slow and non-specific transport of all phospholipids by the so-called floppase, which was identified in red blood cells as multidrug resistance related protein 1 (ABC C1) (Dekkers et al., 2000; Sohnius et al., 2003). However, since the outward transport of lipids by ABC C1 is also ATP dependent, it is unlikely that the activity of this protein contributes to ribavirin-induced phosphatidylserine exposure. Drug-induced phosphatidylserine exposure in erythrocytes has been studied extensively by Lang and coworkers (for a review, see (Foller et al., 2008). Although being anucleated cells, erythrocytes may undergo a process of programmed cell death with many characteristics similar to those of apoptosis in nucleated cells, for which reason the term eryptosis was coined (Lang et al., 2005). Central to this process appears to be a rise in intracellular Ca2 þ which causes activation of a Ca2þ -sensitive K þ channel (Gardos channel) resulting in efflux of K þ and subsequent Cl  leakage, loss of water and cell shrinkage. Increased intracellular Ca2 þ is also thought to activate phospholipid scramblase, which is responsible for surface exposed phosphatidylserine, providing a signal for removal of senescent red blood cells. The present results indicate that ribavirin-induced phosphatidylserine exposure occurs in the absence of extracellular Ca2 þ , raising the question whether the phospholipid scramblase is involved. Here, we demonstrate that erythrocytes from a patient with the Scott syndrome, characterized by the virtually complete absence of Ca2 þ -dependent lipid scrambling, show normal ribavirin-induced phosphatidylserine exposure. Recently, a mutation in TMEM 16F (Ano 6), a member of the anoctamin family of Ca2þ -activatable Cl  channels, was found to be responsible for the defective Ca2 þ -induced lipid scrambling in Scott syndrome patients (Castoldi et al., 2011; Suzuki et al., 2010). We therefore conclude that TMEM16F is not required for ribavirin-induced phosphatidylserine exposure in erythrocytes. Schoenwaelder et al. (2009) reported that phosphatidylserine exposure in blood platelets can be induced via two distinct pathways, one caused by platelet activation and the other through induction of programmed cell death via the pro-apoptotic proteins Bax/Bak. Apoptosis-induced phosphatidylserine exposure in platelets was inhibited by the pan-caspase inhibitor Q-VD-OPh, indicating the involvement of activated caspases. We could recently demonstrate that this pathway of phosphatidylserine exposure is normal in platelets from a patient with the Scott syndrome (Van Kruchten et al., 2011). Although it is tempting to speculate that ribavirin induced phosphatidylserine exposure in erythrocytes is related to eryptosis, we could not find indications for caspase activity being involved. The ATP depletion associated with ribavirin treatment may also compromise the cells reductive capacity leading to increased oxidative stress. Selective oxidation of phosphatidylserine has been proposed to be one of the mechanisms for externalization (Kagan et al., 2000). The lack of effect of the strong antioxidant N-acetylcysteine on ribavirin-induced phosphatidylserine exposure makes this mechanism less likely. It should be emphasized here, that increased production of malondialdehyde (a marker of lipid peroxidation) resulting from increased oxidative stress as previously observed (De Franceschi et al., 2000), could lead to a false positive

6

M.-C. Kleinegris et al. / European Journal of Pharmacology 693 (2012) 1–6

detection of phosphatidylserine by annexin A5 as demonstrated by Balasubramanian et al. (2001). This possibility can be ruled out as similar results were obtained with annexin A5 and lactadherin, the latter protein being selective for phosphatidylserine (Shi et al., 2004). We have previously shown that the rate of Ca2 þ -induced phospholipid scrambling varies among different individuals. Since the present results do not support a significant contribution of the phospholipid scramblase in the loss of lipid asymmetry resulting from prolonged incubation with ribavirin, measuring Ca2 þ -dependent lipid scrambling in red blood cells cannot be used as a biomarker to predict the risk of ribavirin-induced anemia. Indeed, a preliminary study failed to show a correlation between Ca2 þ -induced- and ribavirin-induced phosphatidylserine exposure in erythrocytes from 10 healthy individuals (data not shown). The ribavirin concentrations used in this in vitro study are far exceeding the therapeutic plasma concentration. van Soest et al., (2009) found no phosphatidylserine exposure after 24 h treatment with 10 mM ribavirin and in the present study, 10 mM ribavirin for 1 week at ambient temperature did not result in annexin A5 positive cells either. However, it may be assumed that at an effective plasma concentration of 10 mM, the same levels of ribavirin will finally accumulate as reached here within 48 h using ribavirin at 1 mM. Intracellular ribavirin levels in HCV patients receiving ribavirin have been estimated to range from 0.8 to 1.6 mM (Inoue et al., 2006), and this intracellular concentration was reached in vitro within 18 h upon incubation with 1 mM ribavirin (Homma et al., 2009). It should be emphasized that ribavirin-induced ATP depletion also induces morphological changes in erythrocytes as described by Homma et al. (2009) which may lead to early removal by the reticuloendothelial system. Whether these morphological changes are directly related to the loss of lipid asymmetry remains to be investigated. It is appropriate to mention that phosphatidylserine is not detectable on the surface of red cells from patients treated with ribavirin. In the present study, it is shown that ribavirin-induced phosphatidylserine exposure is a rather slow process with a considerable lag time. Lack of phosphatidylserine-exposed red cells in these patients is presumably due to their rapid removal from the circulation. We conclude that a slow spontaneous transbilayer movement of lipids, combined with an inactive aminophospholipid translocase is the most likely explanation for the increased surface exposure of phosphatidylserine that is observed upon prolonged treatment with ribavirin, thus contributing to the development of anemia. References Alter, M.J., 2007. Epidemiology of hepatitis C virus infection. World J. Gastroenterol. 13, 2436–2441. Balasubramanian, K., Bevers, E.M., Willems, G.M., Schroit, A.J., 2001. Binding of annexin V to membrane products of lipid peroxidation. Biochem 40, 8672–8676. Bevers, E.M., Wiedmer, T., Comfurius, P., Zhao, J., Smeets, E.F., Schlegel, R.A., Schroit, A.J., Weiss, H.J., Williamson, P., Zwaal, R.F., Sims, P.J., 1995. The complex of phosphatidylinositol 4,5-bisphosphate and calcium ions is not responsible for Ca2 þ -induced loss of phospholipid asymmetry in the human erythrocyte: a study in Scott syndrome, a disorder of calcium-induced phospholipid scrambling. Blood 86, 1983–1991. Bevers, E.M., Williamson, P.L., 2010. Phospholipid scramblase: an update. FEBS Lett. 584, 2724–2730. Castoldi, E., Collins, P.W., Williamson, P.L., Bevers, E.M., 2011. Compound heterozygosity for 2 novel TMEM16F mutations in a patient with Scott syndrome. Blood 117, 4399–4400. Connor, J., Pak, C.H., Zwaal, R.F., Schroit, A.J., 1992. Bidirectional transbilayer movement of phospholipid analogs in human red blood cells. Evidence for an ATP-dependent and protein-mediated process. J. Biol. Chem. 267, 19412–19417. Daleke, D.L., 2008. Regulation of phospholipid asymmetry in the erythrocyte membrane. Curr. Opin. Hematol. 15, 191–195.

De Franceschi, L., Fattovich, G., Turrini, F., Ayi, K., Brugnara, C., Manzato, F., Noventa, F., Stanzial, A.M., Solero, P., Corrocher, R., 2000. Hemolytic anemia induced by ribavirin therapy in patients with chronic hepatitis C virus infection: role of membrane oxidative damage. Hepatology 31, 997–1004. de Jong, K., Geldwerth, D., Kuypers, F.A., 1997. Oxidative damage does not alter membrane phospholipid asymmetry in human erythrocytes. Biochemistry 36, 6768–6776. de Kruijff, B., van Zoelen, E.J., van Deenen, L.L., 1978. Glycophorin facilitates the transbilayer movement of phosphatidylcholine in vesicles. Biochim. Biophys. Acta 509, 537–542. Dekkers, D.W., Comfurius, P., van Gool, R.G., Bevers, E.M., Zwaal, R.F., 2000. Multidrug resistance protein 1 regulates lipid asymmetry in erythrocyte membranes. Biochem. J. 350 (Part 2), 531–535. Fadok, V.A., Bratton, D.L., Rose, D.M., Pearson, A., Ezekewitz, R.A., Henson, P.M., 2000. A receptor for phosphatidylserine-specific clearance of apoptotic cells. Nature 405, 85–90. Foller, M., Huber, S.M., Lang, F., 2008. Erythrocyte programmed cell death. IUBMB Life 60, 661–668. Folmer, D.E., Elferink, R.P., Paulusma, C.C., 2009. P4 ATPases–lipid flippases and their role in disease. Biochim. Biophys. Acta 1791, 628–635. Glue, P., 1999. The clinical pharmacology of ribavirin. Semin. Liver Dis. 19 (Suppl. 1), 17–24. Hitomi, Y., Cirulli, E.T., Fellay, J., McHutchison, J.G., Thompson, A.J., Gumbs, C.E., Shianna, K.V., Urban, T.J., Goldstein, D.B., 2011. Inosine triphosphate protects against ribavirin-induced adenosine triphosphate loss by adenylosuccinate synthase function. Gastroenterology 140, 1314–1321. Homma, M., Hosono, H., Hasegawa, Y., Kohda, Y., 2009. Morphological transformation and phosphatidylserine exposure in erythrocytes treated with ribavirin. Biol. Pharm. Bull. 32, 1940–1942. Inoue, Y., Homma, M., Matsuzaki, Y., Shibata, M., Matsumura, T., Ito, T., Kohda, Y., 2006. Erythrocyte ribavirin concentration for assessing hemoglobin reduction in interferon and ribavirin combination therapy. Hepatol. Res. 34, 23–27. Jarvis, S.M., Thorn, J.A., Glue, P., 1998. Ribavirin uptake by human erythrocytes and the involvement of nitrobenzylthioinosine-sensitive (es)-nucleoside transporters. Br. J. Pharmacol. 123, 1587–1592. Kagan, V.E., Fabisiak, J.P., Shvedova, A.A., Tyurina, Y.Y., Tyurin, V.A., Schor, N.F., Kawai, K., 2000. Oxidative signaling pathway for externalization of plasma membrane phosphatidylserine during apoptosis. FEBS Lett. 477, 1–7. Kornberg, R.D., McConnell, H.M., 1971. Inside–outside transitions of phospholipids in vesicle membranes. Biochemistry 10, 1111–1120. Lang, K.S., Duranton, C., Poehlmann, H., Myssina, S., Bauer, C., Lang, F., Wieder, T., Huber, S.M., 2003. Cation channels trigger apoptotic death of erythrocytes. Cell Death Differ. 10, 249–256. Lang, K.S., Lang, P.A., Bauer, C., Duranton, C., Wieder, T., Huber, S.M., Lang, F., 2005. Mechanisms of suicidal erythrocyte death. Cell Physiol. Biochem. 15, 195–202. McEvoy, L., Williamson, P., Schlegel, R.A., 1986. Membrane phospholipid asymmetry as a determinant of erythrocyte recognition by macrophages. Proc. Natl. Acad. Sci. USA 83, 3311–3315. Page, T., Connor, J.D., 1990. The metabolism of ribavirin in erythrocytes and nucleated cells. Int. J. Biochem. 22, 379–383. Russmann, S., Grattagliano, I., Portincasa, P., Palmieri, V.O., Palasciano, G., 2006. Ribavirin-induced anemia: mechanisms, risk factors and related targets for future research. Curr. Med. Chem. 13, 3351–3357. Schoenwaelder, S.M., Yuan, Y., Josefsson, E.C., White, M.J., Yao, Y., Mason, K.D., O’Reilly, L.A., Henley, K.J., Ono, A., Hsiao, S., Willcox, A., Roberts, A.W., Huang, D.C.S., Salem, H.H., Kile, B.T., Jackson, S.P., 2009. Two distinct pathways regulate platelet phosphatidylserine exposure and procoagulant function. Blood 114, 663–666. Schroit, A.J., Madsen, J.W., Tanaka, Y., 1985. In vivo recognition and clearance of red blood cells containing phosphatidylserine in their plasma membranes. J. Biol. Chem. 260, 5131–5138. Shi, J., Heegaard, C.W., Rasmussen, J.T., Gilbert, G.E., 2004. Lactadherin binds selectively to membranes containing phosphatidyl-L-serine and increased curvature. Biochim. Biophys. Acta 1667, 82–90. Sohnius, A., Kamp, D., Haest, C.W., 2003. ATP and GSH dependence of MRP1mediated outward translocation of phospholipid analogs in the human erythrocyte membrane. Mol. Membr. Biol. 20, 299–305. Suzuki, J., Umeda, M., Sims, P.J., Nagata, S., 2010. Calcium-dependent phospholipid scrambling by TMEM16F. Nature 468, 834–838. Szabo, E., Lotz, G., Paska, C., Kiss, A., Schaff, Z., 2003. Viral hepatitis: new data on hepatitis C infection. Pathol. Oncol. Res. 9, 215–221. Van Kruchten, R., Wolfs, J.L.N., Feijge, M.A.H., Mattheij, N.J.A., Collins, P.W., Heemskerk, J.W.H., Bevers, E.M., 2011. Role of TMEM16F in agonist-induced but not apoptosis-induced phosphatidylserine exposure in Scott platelets. J. Thromb. Haemost. 9, 723. van Soest, H., Renooij, W., van Erpecum, K.J., 2009. Clinical and basal aspects of anemia during antiviral therapy for hepatitis C. Ann. Hepatol. 8, 316–324. Webster, D.P., Klenerman, P., Collier, J., Jeffery, K.J., 2009. Development of novel treatments for hepatitis C. Lancet Infect. Dis. 9, 108–117. Zachowski, A., 1993. Phospholipids in animal eukaryotic membranes: transverse asymmetry and movement. Biochem. J. 294, 1–14. Zwaal, R.F., Comfurius, P., Bevers, E.M., 2004. Scott syndrome, a bleeding disorder caused by defective scrambling of membrane phospholipids. Biochim. Biophys. Acta 1636, 119–128. Zwaal, R.F., Comfurius, P., Bevers, E.M., 2005. Surface exposure of phosphatidylserine in pathological cells. Cell. Mol. Life Sci. 62, 971–988.