Alterations of human erythrocyte membrane fluidity by oxygen-derived free radicals and calcium

Free RadicalBiology & Medicine, Vol. 9, pp. 507-514, 1990 Printed in the USA. All rights reserved.

0891-5849/90 $3.00+ .00 © 1990 PergamonPress plc

Original Contribution ALTERATIONS OF HUMAN OXYGEN-DERIVED

ERYTHROCYTE MEMBRANE FLUIDITY FREE RADICALS AND CALCIUM

BY

HIROSHI WATANABE,* AKIRA KOBAYASHI, TAKAHISA YAMAMOTO, SHINGO SUZUKI, HIDEHARU HAYASHI, and NOBORU YAMAZAKI Third Department of Internal Medicine, Hamamatsu University School of Medicine, 3600 Handa-cho, Hamamatsu, 431-31, Japan (Received 22 November 1989; Revised and Accepted 14 March 1990)

Abstract--Two possible reasons for the structural alterations of cell membranes caused by free radicals are lipid peroxidation and an increase in the intracellular calcium ion concentration. To characterize the alterations in membrane molecular dynamics caused by oxygen-derived free radicals and calcium, human erythrocytes were spin-labeled with 5-doxyl stearic acid, and alterations in membrane fluidity were quantified by electron spin resonance spectrometry. The in vitro generation of oxygen free radicals, using hypoxanthine (0.43 mM) plus xanthine oxidase (0.07 U/mL) decreased membrane fluidity, and the addition of superoxide dismutase and catalase inhibited the effect on membrane fluidity of the hypoxanthine-xanthine oxidase system. Hydrogen peroxide (0.1 and 1 mM) also decreased membrane fluidity and caused alterations to erythrocyte morphology. In addition, a decrease in membrane fluidity was observed in erythrocytes incubated with 2.8 mM CaC12. On the other hand, incubation of erythrocytes with calcium-free solution decreased the changes in membrane fluidity caused by hydrogen peroxide. These results suggest that changes in membrane fluidity are directly due to lipid peroxidation and are indirectly the result of increased intracellular calcium concentration. We support the hypothesis that alterations of the biophysical properties of membranes caused by free radicals play an important role in cell injury, and that the accumulation of calcium amplifies the damage to membranes weakened by free radicals. Keywords--Membrane fluidity; Free radicals; Calcium ion; Human erythrocyte; Radical scavenger; Spin-label; Electron spin resonance

INTRODUCTION

Oxygen-derived free radicals and hydrogen peroxide have been suggested to contribute to cell damage. The plasma membrane is a critical site of free radical reactions for several reasons. Extracellularly generated free radicals may initiate toxic reactions at the plasma membrane with the unsaturated fatty acids (phospholipids) present in membranes and the transmembrane proteins containing oxidizable amino acids both being susceptible to free radical damage. Also, it has been suggested that increased membrane permeability caused by lipid peroxidation or the oxidation of important structural proteins can cause a breakdown of transmembrane ion gradients and inhibition of integrated cellular metabolic processes. In fact, decreased membrane fluidity has been found *Author to whom correspondence should be addressed. 507

in studies employing the fluorescent probe diphenylhexatriene following peroxidation of phospholipid vesicles, erythrocytes, and microsomes. 1'2"3Bruch et al. 4 examined the effect of lipid peroxidation on membrane fluidity in sonicated phospholipid vesicle (liposomes) by using a series of doxyl stearate spin probes, and reported that membrane fluidity decreased following lipid peroxidation, In contrast, Grzelinska et al. 5 have reported that the membrane fluidity of erythrocytes increased following peroxidation. Thus, contradictory experimental results have been obtained with fatty acid spin probes. Two possible reasons for structural alterations of phospholipids are free radical lipid peroxidation and an increase in the intracellular calcium concentration. This study investigated alterations of membrane fluidity and morphological changes in erythrocytes caused by oxygen-derived free radicals and calcium. Human erythrocytes were used as a membrane model after

508

H. WATANABEet al.

spin-labeling with 5-doxyl stearic acid, and alterations of membrane fluidity were quantified by electron spin resonance spectrometry. The morphological study was performed by scanning electron microscopy. MATERIALS AND METHODS

Spin labeling of erythrocytes We used human erythrocytes as our membrane model system, since the effects on oxidation of erythrocyte membranes can serve as a general model for the oxidative damage of biomembranes and since Grzelinska et al. 5 reported that membrane fluidity increased following the peroxidation of erythrocytes. Fresh blood samples were collected from healthy adult donors into heparinized tubes, and centrifuged at 3,000 × g for 5 min. The plasma and buffy coat were then removed and erythrocytes were washed three times with cold isoosmotic NaCI. They were then suspended in modified Hank's balanced salt solution with the following composition (in mM): NaCI, 137; KC1, 5.4; CaC12, 0.5; glucose, 5.6; KH2PO4, 0.44; Na2HPO4, 0.33 (pH 7.2). A fatty acid spin label, 5-oxyl stearic acid (5-DSA), which has a stable nitroxide radical ring at the 5th carbon position from the carboxyl group of the acyl chain, was used in this study. This probe becomes oriented in the membrane such that the long axis of the probe is parallel to the phospholipid fatty acids. Erythrocytes were labeled by incubating 500 pL of a 50% (v/v) suspension of washed erythrocytes with 1.5 mL of Hank's solution containing 10 pg of 5-DSA for 2 hours at 4°C. Spin-labeled erythrocytes were washed twice with Hank's solution to remove unincorporated spin labels.

Free radical generation system Superoxide anion radicals were generated by the xanthine oxidase reaction using hypoxanthine as a substrate. Xanthine oxidase was pretreated with 0.4 mM phenylmethylsulfonyl fluoride to inhibit trypsin-like activity, which is present in the commercial product as a contaminant. Spin-labeled erythrocytes (final hematocrit, 12.5%) were incubated with hypoxanthine (0.043 mM and 0.43 mM) and xanthine oxidase (0.007 U/mL and 0.07 U/mL) for various durations (5, 10, 20, 30, 45, 60, and 90 min) at 37°C in a shaking thermostat bath. The amounts of hypoxanthine and xanthine oxidase used were decided on the basis of previous findings. 6 Generation of superoxide anion radicals was ascertained by chemiluminescence techniques using 2-methyl-6-(p-methoxyphenyl)-3,7dihydro-imidazol[ 1,2-a]pyrazin-3-one (MCLA).7

MCLA-dependent luminescence is originated by superoxide anion radicals reacting with MCLA. Thus, the extent of radical generation could be determined by monitoring the chemiluminescence intensity with a luminescence reader (Aloka Co., Japan). Addition of MCLA caused a strong emission of light with a wavelength around 465 nm in the presence of superoxide anion radicals. Experiments with hydrogen peroxide and erythrocytes were performed in the presence of sodium azide (1.2 mM) in order to inhibit endogenous catalase activity. Spin-labeled erythrocytes (final hematocrit, 12.5%) were incubated with various concentrations of hydrogen peroxide (10 5, 10-4, and 10 3 M) for various durations (5, 10, 20, 30, 45, 60, and 90 min) at 37°C in a shaking thermostat bath. It has been found previously that hydrogen peroxide is maintained under normal conditions at concentrations of 10 7 to 10 -9 M by intracellular catalase and peroxidases. 8 Since a dramatic increase in free radical generation occurs during reperfusion, we used three increasing concentrations of hydrogen peroxide (10 -5, 10 -4, and 10 -3 M) in this study. After incubation, the erythrocytes were suspended in 10 volumes of Hank's solution at 4°C, centrifugated at 3,000 x g for 3 min, and the pellet was then transferred to a Pyrex capillary tube.

Measurement of membrane fluidity Electron spin resonance (ESR) spectra were obtained with a JES-FE2XG spin resonance spectrometer (JEOL, Japan) operating at a center field strength of 3,280 Gauss with an 8-min scan time to scan 100 Gauss, a 0. l-s time constant, a modulation amplitude of 2.0 Gauss, and microwave power of 8 milliwatts. Spin-labeled erythrocytes, aspirated into a Pyrex capillary tube, were placed in a quartz holder which was maintained at a constant temperature of 37°C. The ESR spectrum of human erythrocytes labeled with 5-DSA revealed the rapid anisotropic motion typical of fatty acids residing within a phospholipid membrane bilayer (Fig. 1). The fluidity of the membrane-incorporated label was quantified by measuring the order parameter S, as described by Gaffney. 9 The parallel (2T //) and perpendicular (2T A_) components of the hyperfine splittings were measured graphically, and the order parameter S was calculated from these values. This order parameter gives a measure of the degree of structural order in the membrane, with the order parameter being 1 for a spin-label moving rapidly about only one axis and 0 for rapid isotropic motion. An increase in the value of S is interpreted as a decrease in membrane fluidity.

Free radicals, Ca2÷, and membranefluidity

509

Erythrocytes were then sprayed with a thin gold layer, and observed by scanning electron microscopy using a JSM-35 apparatus (JEOL, Japan).

Reagents

i

,:~--2TJ _~:

-2Tr---

lOGauss

Fig. 1. Electron spin resonance spectrum of human erythrocytes labeled with 5-DSA. The parallel (2T //) and perpendicular(2T .1_) components of the hyperfine splittings were measured graphically and the order parameterS was calculated from these values.

Effects of free radical scavengers on membrane fluidity Superoxide dismutase (SOD; 100 U) or SOD (100 U) plus catalase (0.03 mg/mL) were added to Hank's solution which contained hypoxanthine (0.43 mM) and spin-labeled erythrocytes, after which xanthine oxidase (0.07 U/mL) was added to the system. The concentrations of these chemicals used in this study were determined from information obtained in previous studies. 6"Is Mixtures were incubated for various durations (5, 10, 20, 30, 45, 60, and 90 min) at 37°C in a shaking thermostat bath. After incubation, the erythrocytes were suspended in 10 vols of Hank's solution at 4°C and centrifuged at 3,000 × g for 3 min, and then the pellet was transferred to a Pyrex capillary tube.

The 5-DSA spin-label, hypoxanthine, xanthine oxidase, hydrogen peroxide, SOD, and catalase were purchased from Sigma Chemical Co. MCLA was kindly supplied by Prof. Nakano of the College of Medical Care and Technology at Gunma University in Japan.

Statistics All order parameter S values are reported as the mean --- SD. Significance of difference was determined using the paired or nonpaired Student's t-test as was appropriate. A P value of <0.05 was considered significant. RESULTS

Effects of hypoxanthine-xanthine oxidase on erythrocyte membrane fluidity Superoxide anions were generated by the xanthine oxidase reaction using hypoxanthine as a substrate. The generation of superoxide anions was confirmed by chemiluminescence techniques using MCLA-dependent luminescence (Fig. 2), and it was shown that emis-

E

Hypoxanthine 430pM |Xanthlne Oxidase 0.07 unlts/ml /,.~//MCLA 4,AJM ffer pHT.4 lOmM

40. Effects of Ca2÷ on membrane fluidity Spin-labeled erythrocytes were incubated with various concentrations of CaCI2 (0, 1,0, and 2.8 mM) for 5, 10, 20, 30, 45, 60, and 90 min at 37°C in a shaking thermostat bath. In addition, spin-labeled erythrocytes were incubated simultaneously with hydrogen peroxide (10 -3 M) and either Ca2+-free Hank's solution and EGTA (0.5 mM), to exclude the influence of contaminating Ca 2÷, or normal Hank's solution containing CaCI2 (0.5 mM) for 90 min at 37°C in a shaking thermostat bath. After incubation, the erythrocytes were suspended in 10 vols of Hank's solution at 4°C, centrifuged at 3,000 × g for 3 min, and then the pellet was transferred to a Pyrex capillary tube.

Scanning electron microscopy Erythrocytes incubated with hydrogen peroxide (10 -3 M) for various durations (0, 5, and 30 min) were fixed in 2% glutaraldehyde, dehydrated with graded alcohols, and dried by the critical point drying method.

0 0

'- 30 X >O9

Z

20

LU

_z -r-

1C

..J

/+SOD 0

0

1

2

3

4

INCUBATION

5

6

?

~min

TIME

Fig. 2. MCLA-dependentluminescencein a hypoxanthine-xanthine oxidase system. The incubation mixture contained 4 pM MCLA, 430/zM hypoxanthine,0.07 U/mL of xanthineoxidase, 10 mM (pH 7.4) Tris HCI buffer, and Hank's balanced salt solution. Supcroxide dismutase (100 U) completely abolished the enzyme-induced luminescence.

H. WATANABEet al.

510

sion was completely inhibited by administration of SOD. The effects of hypoxanthine and xanthine oxidase on the value of the order parameter S are shown in Fig. 3A. Low amounts of hypoxanthine (0.043 mM) plus xanthine oxidase (0.007 U/mL) induced no significant changes in the S values of erythrocytes. However, high concentrations of hypoxanthine (0.43 mM) plus xanthine oxidase (0.07 U/mL) caused the value of S to increase significantly after 10 min of incubation. The value became maximal at 20 min, and then shared a plateau thereafter. However, the value of S remained unchanged, when erythrocyte membranes were incubated with hypoxanthine (0.43 mM) or xanthine oxidase (0.07 U/mL) alone (Fig. 3B). This indicated that superoxide anions had decreased the membrane fluidity of human erythrocytes.

Effect of hydrogen peroxide on erythrocyte membrane fluidity Spin-labeled erythrocyte membranes were incubated with hydrogen peroxide at concentrations rang-

o3

0.670

~ 0.660

0.

tit

'

,

1

l

ing from 10 -5 to 10 -3 M. Incubation of erythrocytes with hydrogen peroxide (10 -5 M) did not produce a significant alteration of the order parameter S. With 10 -4 M hydrogen peroxide, however, the value of S increased significantly after 10 min of incubation, and then reached a plateau level at 20 min. These alterations of the value of S were similar to those seen in the hypoxanthine-xanthine oxidase system. In the presence of hydrogen peroxide (10 -3 M), the value of S increased after 5 min, was maximal at 10 min, and decreased thereafter (Fig. 4). Thus, the membrane fluidity of human erythrocytes was shown to be reduced by hydrogen peroxide in a concentration-dependent manner.

Effects of free radical scavengers on membrane fluidity MCLA-dependent chemiluminescence was completely inhibited by SOD in this study (Fig. 2). This indicated that the amount of SOD added was sufficient to scavenge the superoxide anion radicals generated in the hypoxanthine-xanthine oxidase system. However, SOD alone had no effect on the value of S in the hypoxanthine-xanthine oxidase system (Fig. 5A). On the other hand, combined treatment with SOD and catalase significantly inhibited the alteration of erythrocyte membrane fluidity by a 10-min incubation in the hypoxanthine-xanthine oxidase system (Fig. 5B).

0

0.660"

Effects of calcium on membrane fluidity 0 5 10 20

30

45

60

g'Omin

0.670b3

Incubation of erythrocytes with CaC12 (1.0 mM) did not produce significant alteration of the value of S (Fig. 6). On the other hand, incubation with CaCI: (2.8 mM) resulted in a significant increase in the value of S after 10 min. Incubation of erythrocyte membranes with Ca 2+-free solution partially inhibited the increase of S

0.560 EL "(3 0

0.690-

*~m,M

O. 650.

0.680-

6 ~ 1'o 2b ~b

4'~

6b

9bm~o

E

¢o

Fig. 3. Panel A is the effect of hypoxanthine plus xanthine oxidase on the order parameter S. The concentrations of hypoxanthine (HX) and xanthine oxidase (XO) were as follows:e~---.--~,control; o------o, HX (0.043 mM) + XO (0.007 U/mL); m--no, HX (0.43 mM) + XO (0.07 U/mL). Values are the mean - SD (n = 5). * = p < 0.05 compared to control at each incubation time. Panel B is the effect of hypoxanthine or xanthine oxidase alone on the order parameter S. The concentrationof hypoxanthine (HX) and xanthine oxidase (XO) were as follows: l----i, HX (0.43 raM) + XO (0.07 U/mL); ~-------~,HX (0.43 mM); o.------o, XO (0.07 U/mL). Values are the mean +- SD (n = 5). * = p < 0.05 compared to single administration of XO or HX at each incubation time.

0.670

el

o

T

0.660

°i

0.65( 0 ~ 10

2"0 ~ 0

4:5

6"0

gOmin

Fig. 4. Effects of hydrogen peroxide on the order parameter S of human erythrocytes. The concentrations of hydrogenperoxide were as follows: e o, 10-s M; o o, 10-4 M; "-', 10-3 M. Values are the mean -+ SD (n = 5). * = p < 0.05 compared to before administration of hydrogen peroxide.

Free radicals, Ca =+, and membrane fluidity

A

511

0.670 o)

m 0.670 .............

i 0.660

E ~. 0.660

• • Control o----o lmM CaCI= ........ ZSmld CaC~

~5 0.650

2

0

0 5 10

OS'~ 20 :~

4s

eo

9~min

B

20

30

45

60

90mln

Fig. 6. Effects of CaCI2 on the order parameter S of the human erythrocytes. The concentrations of CaCI2 were as follows: e.-.-.---.-e, control; o---------o, 1 mM and B----a, 2.8 mM CaC12". Values are the mean --- SD (n = 5). * = p < 0.05 compared to control at each incubation time.

0.670. E

O.o.ee 2

0

0S~

20

30

4~5

6(3

90min

Fig. 5. Effects of radical scavenger on the order parameter S in the hypoxanthine (HX, 0.45 raM) plus xanthine oxidase (XO, 0.07 U/ mL) system. Panel A is the effect of superoxide dismutase (SOD, 100 U). =---------=, SOD alone; e..-.--.--e, HX + XO; o . . . . o, SOD with HX + XO. Values are the mean - SD (n = 5). There was no significant difference between with and without SOD in the hypoxanthine-xanthine oxidase system. Panel B is the effect of combined treatment with SOD (100 U) and catalase (0.03 m g / m L ) on the order parameter S. -= ~, HX + XO; o - - - o , a combination of SOD and catalase. * = p < 0.05 compared to I-IX + XO at each incubation time. Values are the mean -+ SD (n = 5).

caused by 10 -3 M hydrogen peroxide with 0.5 mM CaCI2 (Fig. 7).

pholipids is free radical lipid peroxidation. It has been found that free radicals generated in an aqueous phase can attack erythrocyte membranes to induce the chain oxidation of lipids and proteins, and eventually cause hemolysis. ~1This study showed that membrane fluidity, as reflected by changes in the electron spin resonance spectrum of a fatty acid nitroxide probe, decreased following administration of oxygen derived free radicals. Moreover, this study showed that hydrogen peroxide caused significant morphological changes of erythrocytes which transformed from discocytes to echinocytes during the 5 min of incubation and finally all erythrocytes changed to sphero-echinocytes. A previous spin-labeled study of membrane fluidity in erythrocytes peroxidized by gamma-irradiation has been reported by Grzelinska et al.5 who used the methyl esters of 5-doxyl palmitate and 12-doxyl stearate to monitor membrane fluidity. However, they found no change in fluidity with 5-doxyl palmitate and that flu-

Effects of hydrogen peroxide on erythrocyte morphology Hydrogen peroxide (10 -3 M) caused significant morphological changes (Fig. 8). A transformation of the normal discoid shape to a spiculated shape was induced during 5 min of incubation and after 30 min of incubation the erythrocytes had been transformed to sphero-echinocytes.

0.690 CO

0..

O.68O 0.670

~/~l/I I

I

IO. 51aM BG~&) --

~ O.B60 0.650

DISCUSSION

Peroxidation of the unsaturated fatty acids of cell membrane phospholipids is accompanied by alterations of membrane structural and functional characteristics. Lipid peroxidation also affects the physical properties of membranes, including membrane fluidity. ~° One possible cause of such structural alterations of phos-

5 10

20

30

45

60

90rnin

Fig. 7. Effects of 10 -3 M hydrogen peroxide with calcium-free solution on the order parameter S of human erythrocytes. The erythrocytes were incubated with 10 -3 M hydrogen peroxide in the presence ( ~ - - - e ) of 0.5 m M CaCI=, or in the absence (o . . . . o) of calcium and with 0.5 m M EGTA. EGTA was added to exclude the influence of contaminating Ca =+. Values are the mean -+ SD (n = 5). * = p < 0.05 compared to with CaCI= at each incubation time.

512

H. WATANABEet al.

A

B

12

Fig. 8. Morphological changes of human erythrocytes induced by application of 10 3 M hydrogen peroxide. Panel A is the control. Panels B and C show the morphological changes after 5 and 30 min of incubation.

idity actually increased slightly with 12-doxyl stearate. On the other hand, Vladimirov et al. ~o,~zhave reported increased membrane rigidity (decreased fluidity) following peroxidation of model and biological membranes in experiments using the fluorescent probes pyrene, perylene, and 4-dimethylchalcone. Bruch et al. 4 have also reported decreased membrane fluidity following the peroxidation of sonicated soybean phospholipid vesicles, which they detected using electron spin resonance probes. The results observed by Vladimirov et al. and Bruch et al. are consistent with the results in our study, but the results observed by Grzelinska et al. are contrast with our results. Thus, there are contradictory experimental results regarding the changes in membrane fluidity following peroxidation. Grzelinska et al. obtained an increase in fluidity in the

presence of 0.5% (about 0.1 M) ethanol, a concentration that is sufficient to increase the fluidity of a variety of membranes to a significant extent. 4 Therefore, the results observed by Grzelinska et al. may be attributable to the combined effects of ethanol and lipid peroxidation. In this study, we found that high concentrations of hypoxanthine and xanthine oxidase induced significant changes in the order parameter S of human erythrocytes, showing that superoxide anions may decrease membrane fluidity. It has been found that the phospholipids in membranes are oxidized by a free radical chain mechanism and that free radical scavengers suppress this peroxidation. However, SOD alone had no effect on the alterations of membrane fluidity caused by the hypoxanthine-xanthine oxidase system, although the concentration of SOD that we used was sufficient to scavenge all the superoxide anions generated. On the other hand, combined treatment of erythrocytes with SOD and catalase prevented the alterations of membrane fluidity caused by this system. SOD catalyzes the dismutation of superoxide to hydrogen peroxide, and catalase mediates the conversion of hydrogen peroxide to water and molecular oxygen. The effect of the addition of catalase suggests that hydrogen peroxide or the highly reactive hydroxyl radicals may be more important contributors to radicalmediated alterations of membrane fluidity than are superoxide anions. Another possible explanation for the lack of effect of added SOD may be due to the fact that this enzyme promotes the formation of hydrogen peroxide and that, in the presence of exogenous SOD, the endogenous catalase and peroxidase activities may be insufficient. 13This problem is most likely to express itself in experimental systems such as the buffer-perfused heart, in which the endogenous blood catalase activity is eliminated. 14 The important role of the calcium ion in cellular functions involving membranes has long been recognized. There is considerable evidence that the calcium ions can interact directly with the phospholipids which are arranged in monolayers or model bilayers and thereby restrict the freedom of motion or lipid fluidity of such arrays. A prior investigation has demonstrated that calcium ions can decrease the lipid fluidity of hepatocyte plasma membranes in vitro by influencing membrane-bound enzymes to alter the lipid composition.~5 Moreover, it has been reported that free radicals enhanced calcium release from the sarcoplasmic reticulum, ~6 and also that they inhibited sarcolemmal Na +, K+-ATPase, possibly causing the activation of the Na+-Ca 2÷ exchange mechanism in the myocardium. '7 Kaneko et al. 18 have demonstrated that free radicals depressed the Ca2+-pump activity in the car-

Free radicals, Ca 2+, and membranefluidity diac sarcolemmal membrane. Na+-Ca 2+ exchange mechanism may play a role in regulating the cellular calcium level in several tissues, for example, dog erythrocytes, squid axon, and cardiac muscle. However, Mcnamara et al. 19 and Sarkadi 2° have demonstrated that a significant Na+-Ca 2+ exchange was absent from human erythrocytes with physiological internal sodium concentration. Another possibility of calcium influx is an enhancement of passive calcium transport due to membrane disorganization. Extracellular calcium level modulates the passive calcium transport as well as the Na+-Ca 2÷ exchange mechanism. In this study, an incubation of erythrocyte membrane with a calcium-free solution partially inhibited the changes in membrane fluidity caused by hydrogen peroxide. We have already reported that hydrogen peroxide (10-3 M) increases the intracellular calcium concentration of isolated guinea pig ventricular myocytes using fura-2. 2j These data suggest that the decrease in membrane fluidity caused by hydrogen peroxide might be partially related to an increase in intracellular calcium concentration through passive calcium transport. We found that a decrease in membrane fluidity was observed in the erythrocytes incubated with CaCl2 (2.8 mM). This result is consistent with other reports. 22,23 Ohnishi and Ito 23 showed the decrease in membrane fluidity at higher calcium concentrations (2-10 mM, CaCI2) using phosphatidylcholine spin labels. 23 Livingstone and SchachteP 5 have reported that calcium ions modulate the lipid dynamics of rat hepatocyte plasma membranes in a complex manner involving both "direct" and "indirect" mechanisms for altering lipid fluidity. The former mode of action of the calcium ion on fluidity appears to involve the binding of ions t o anionic sites in the lipid bilayer. In contrast, the latter effect of calcium ions on fluidity appears to involve the stimulation of certain membrane-bound enzymes which alter membrane lipid composition. The direct action can be reversed by the addition of excess EDTA, whereas the indirect action cannot be reversed by the chelation of calcium ions using EDTA. We found that incubation of erythrocyte membranes with a calcium-free solution partially inhibited the changes in membrane fluidity caused by hydrogen peroxide (10 -3 M). Thus, it seems likely that the decrease in membrane fluidity caused by hydrogen peroxide may be partially related to an increase in intracellular calcium concentration. Another possibility of the alteration in membrane fluidity is the modulation of magnesium ion concentration, as the magnesium ion is also known to affect membrane integrity. Ohnishi and I t o 23 investigated the effects of various divalent cations (Mg 2+, Ba 2÷, and C a 2÷) o n the phosphatidylserinephosphatidylcholine membranes. It is clearly indicated

513

that Ba 2+ and Sr 2+ caused decreases in membrane fluidity; however, the changes were smaller than that induced by Ca 2+ , and Mg 2+ did not cause a decrease in membrane fluidity. These results suggest that free radicals may directly change membrane fluidity by lipid peroxidation and also may indirectly change fluidity as a result of increasing the intracellular calcium concentration. Alternatively, since membrane fluidity could modulate membrane functions such as ion permeability and enzyme activities, the alterations in membrane fluidity might themselves lead to an increase in the intracellular calcium ion concentration. In addition, the increase of intracellular calcium concentration may alter the configuration of the spectrin-actin network attached to the internal surface of the erythrocyte membrane, and may contribute to the irreversible stiffening of the membrane. 24 Recently, it has been shown that an increase in the cytoplasmic calcium concentration of the erythrocytes induced marked alterations in the transbilayer organization of the membrane phospholipids accompanied by changes in cell shape, membrane-protein composition, intracellular ATP levels, and transglutaminase activity, and that these changes might affect the decrease in membrane fluidity. 25 This study demonstrated that oxygen-derived free radicals induced alterations in membrane fluidity and morphology in human erythrocytes, and that combined treatment with SOD and catalase protected erythrocytes from alterations of membrane fluidity. In addition, we showed that calcium was related to the alterations of membrane fluidity caused by free radicals. Since it has been shown that both free radicals and calcium accumulate in the myocardium during ischemia and reperfusion, these results support the hypothesis that alterations of the biophysical properties of membranes caused by oxygen-derived free radicals play an important role in cell injury, and that the accumulation of calcium amplifies the damage to membranes which have been weakened by free radicals. The use of free radical scavengers may thus lead to improved function of ischemic organs following reperfused after transplantation, coronary angioplasty, and other such procedures. REFERENCES

1. Barrow, D. A.; Lentz, B. R. A model for the effect of lipid oxidation on diphenylhexatrienefluorescencein phospholipid vesicles. Biochim. Biophys. Acta 645:17-23; 1981. 2. Rice-Evans,C.; Hochstein,P. Alterationsin erythrocytemembrane fluidityby phenylhydrazine-inducodperoxidationof iipids. Biochem. Biophys. Res. Comm. 100:1537-1542; 1981. 3. Eichenberger,K.; B6hni, P.; Winterhalter,K. H.; Kawato,S.; Richter, C. Microsomallipid peroxidationcauses an increase in the order of the membranelipid domain.FEBS Len. 142:5962; 1982.

514

H. WATANABEet al.

4. Bruch, R. C.; Thayer, W. S. Differential effect of lipid peroxidation on membrane fluidity as determined by electron spin resonance probes. Biochim. Biophys. Acta 733:216-222; 1983. 5. Grzelinska, E.; Bartasz, G.; Gwozdzinski, K.; Leyko, W. A spin-label study of the effect of gamma radiation on erythrocyte membrane. Influence of lipid peroxidation on membrane structure. Int. J. Radiat. Biol. 36:325-334; 1979. 6. Battelli, M. G.; Corte, E. D.; Stirpe, F. Xanthine oxidase type D (dehydrogenase) in the intestine and other organs of the rat. Biochem. J. 126:747-749; 1972. 7. Nakano, M.; Sugioka, K.; Ushijima, Y.; Goto, T. Chemiluminescence probe with cypridina luciferin analog, 2-methyt-6phenyl-3,7-dihydro-imidazo[1,2-a]pyrazin-3-one, for estimating the ability of human granulocytes to generate 02-. Analy. Biochem. 159:363-369; 1986. 8. Freeman, B. A.; Crapo, J. D. Biology of disease. Free radicals and tissue injury. Labo. Invest. 47:412-426; 1982. 9. Gaffney, B. J.; McConnel, H. M. The paramagnetic resonance spectra of spin labels in phospholipid membranes. J. Magn. Res. 16:1-28; 1974. 10. Vladimirov, V. A.; Olenev, V. I.; Suslova, T. B.; Cheremisina, Z. P. Lipid peroxidation in mitochondrial membrane. In: Paoletti, R.; Kritchevsky, D., ed. Advances in lipid research. New York: Academic Press; 17:173-249; 1980. 1!. Niki, E.; Komuro, E.; Takahashi, M.; Urano, S.; Ito, E.; Terao, K. Oxidative hemolysis of erythrocytes and its inhibition by free radical scavengers. J. Biol. Chem. 263:19809-19814; 1988. 12. Dobretsov, G. E.; Borschevskaya, T. A.; Petrov, V. A.; Vladimirov, Y. A. The increase of phospholipid bilayer rigidity after lipid peroxidation. FEBS Lett. 84:125-128; 1977. 13. Bermier, M.; Manning, A. S.; Hearse, D. J. Reperfusion arrhythmias: dose-related protection by anti-free radical interventions. Am. J. Physiol. 256:HI344-H1352; 1989. 14. Shlafer, M.; Kane, P. F.; Kirsh, M. M. Superoxide dismutase plus catalase enhances the efficacy of hypothermic cardioplegia

15. 16. 17.

18. 19. 20. 21.

22. 23. 24.

25.

to protect the globally ischemic, reperfused heart. J. Thorac. Cardiovasc. Surg. 83:830-839; 1982. Livingstone, C. J.; Schachter, D. Calcium modulates the lipid dynamics of rat hepatocyte plasma membranes by direct and indirect mechanisms. Biochem. 19:4823-4827; 1980. Okabe, E.; Hess, M. L.; Oyama, M.; Ito, H. Characterization of free radical-mediated damage of canine cardiac sarcoplasmic reticulum. Arch. Biochem. Biophys. 225:164-177; 1983. Kako, K.; Kato, M.; Matsuoka, T.; Mustapha, A. Depression of membrane-bound Na+-K+-ATPase activity induced by free radicals and by ischemia of kidney. Am. J. Physiol. 254:C330C337; 1988. Kaneko, M.; Beamish, R. E.; Dhalla, N. S. Depression of heart sarcolemmal Ca2+-pump activity by oxygen free radicals. Am. J. Physiol. 256:H368-H374; 1989. Mcnamara, M. K.; Wiley, J. S. Passive permeability of human red blood cells to calcium. Am. J. Physiol. 250:C26-C31; 1986. Sarkadi, B. Active calcium transport in human red cells. Biochim. Biophys. Acta. 604:159-190; 1980. Hayashi, H.; Miyata, H.; Watanabe, H.; Kobayashi, A.; Yamazaki, N. Effects of hydrogen peroxide on action potentials and intracellular Ca2÷ concentration of guinea pig heart. Cardiovas. Res. 23:767-773; 1989. Brasitus, T. A.; Dudeja, P. K. Modulation of lipid fluidity of small- and large-intestinal antipodal membranes by Ca 2.. Biochem. J. 239:625-631; 1986. Ohnishi, S.; Ito, T. Calcium-induced phase separations in phosphatidylserine-phosphatidylcholine membranes. Biochem. 13: 881-887; 1974. Lorand, L.; Weissmann, L. B.; Epel, D. L.; Bruner-Lorand, J. Role of the intrinsic transglutaminase in the Ca2÷-mediated crosslinking of erythrocyte proteins. Proc. Natl. Acad. Sci. USA 73:4479-4481; 1976. Chandra, R.; Joshi, P. C.; Bajpai, V. K.; Gupta, C. M. Membrane phospholipid organization in calcium-loaded human erythrocytes. Biochim. Biophys. Acta. 902:253-262; 1987.