Klin Wochenschr (1985) 63:578-582 Klinische

Wochen- schrift

© Springer-Verlag 1985

The Fatty Acid Composition of Red Cells Deficient in Glucose-6-Phosphate Dehydrogenase and Their Susceptibility to Lipid Peroxidation

M.R. Clemens, H. Einsele, and H.D. Waller Eberhard-Karls-Universit~t Tiibingen, Medizinische Klinik, Abteilung Innere Medizin II

Summary. Oxidant damage to red cell membranes could play a part in the pathogenesis of acute and chronic haemolysis in glucose-6-phosphate dehy- drogenase deficiency. Therefore, we studied the substrate for red cell membrane lipid peroxidation, i.e. the content of various polyunsaturated fatty acids in ghosts, and the susceptibility of red cells to lipid peroxidation in normal subjects and in sub- jects deficient in glucose-6-phosphate dehydroge- nase. The fatty acid composition of red cell mem- branes and plasma was analysed by capillary column gas chromatography. The sensitivity of red cells to lipid peroxidation was evaluated after hy- drogen-peroxide-induced oxidant stress. The de- gree of lipid peroxidation was monitored by mea- suring the release of pentane and ethane formed during the breakdown of n-6 and n-3 fatty acids. The red cell sensitivity to lipid peroxidation was found to be higher in subjects with glucose-6-phos- phate dehydrogenase deficiency than in normal subjects. In the former, saturated fatty acids, in particular palmitic and stearic acid, were found to be decreased, whereas the proportion of arachi- donic acid showed a clear increase. Fatty acid anal- ysis of plasma did not reveal significant abnormali- ties in enzyme-deficient patients, which could ex- plain the alteration of membrane fatty acids. Our results suggest that the increased content of sub- strate for lipid peroxidation, particularly arachi- donic acid, in red cell membranes of subjects defi- cient in glucose-6-phosphate dehydrogenase, should be considered in an evaluation of an en- hanced sensitivity to red cell lipid peroxidation.

Abbreviations: G-6-PD =Glucose-6-phosphate dehydrogenase; GSH=Reduced glutathione; GSSG=oxidized glutathione; MCV=Mean corpuscular volume; MDA=Matondialdehyde; RBC = Red blood cells; TBA = Thiobarbituric acid

Key words: Anaemia Fatty acids - Glucose-6- phosphate-dehydrogenase deficiency - Lipid hyd- roperoxides - Lipids

Deficiency of G-6-PD is the most common inborn metabolic disorder of red blood cells. The clinical expression of G-6-PD variants encompasses a con- tinuous spectrum of haemolytic syndromes. The deficiency state is unrecognized in most affected individuals. With the most prevalent G-6-PD vari- ants - the Mediterranean type - haemolytic crises are induced by certain drugs, bacterial or viral in- fection, stress and following ingestion of fava beans. Under most of these conditions, red cel l glutathione (GSH) oxidation to the disulfide (GSSG) is increased due to the failure of NADPH regeneration in the deficient cells (Ben6hr and Waller 1974; Kosower et al. 1982). The biochemi- cal basis of the onset of haemolytic crisis in G-6- PD deficient subjects is still unknown. Since oxi- dant damage to red cell membranes could play a part in the pathogenesis of acute and chronic hae- molysis in G-6-PD deficiency, we studied the sub- strate for red cell membrane lipid peroxidation, i.e. the content of polyunsaturated fatty acids in ghosts, and the susceptibility of red cells to lipid peroxidation in normal and G-6-PD deficient sub- jects.

When red cells are incubated with high concen- trations of hydrogen peroxide in the presence of a catalase inhibitor, membrane lipids are slowly and partly oxidized (Stocks and Dormandy 1971). The currently used hydrogen peroxide stress test, based on the measurement of TBA reactive materi- al, usually MDA, has provided a simple and sensi- tive method for clinical studies. However, several

M.R. Clemens et al. : G-6-PD Deficiency of RBC and Lipid Peroxidation

Table 1. Haematological data of G-6-PD deficient patients

579

Patient Haemolysis Haemoglobin MCV (fl) Reticulocytes Oxidized GSH G-6-PD activity no. (g/dl) (o/o) (mmol/l RBC) (IU/10 I1 RBC)

1 No 13.2 91 5 2.92 5 2 No 12.9 89 3 2.18 1,2 3 No 14.1 101 8 1.32 0.5 4 Compensated 12.5 79 21 1.82 0.4

Haemolytic crisis 7.2 ND 160 2.07 0.1 5 Compensated 14.0 93 64 2.02 0.4 6 Compensated 13.0 88 72 2.18 0.7 7 No 12.5 106 t0 2.14 2.0

Normal ranges 12.~16.0 82-92 3-I 0 2.45 +_ 0.30 17.1±2.3

aspects of the test and its interpretation remain unclear. In particular, the measurement of TBA reactive material is not specific for oxidation prod- ucts of lipids. Furthermore, the application of this test in experiments with red cells is complicated by a potentially interfering colour reaction (Stocks and Dormandy 1970) and by the reaction of MDA itself with red cell membranes (Pfafferott et al. 1982; Jain and Hochstein 1980).

The new method used in this study circumvents these problems which are known to be serious in evaluating the degree of lipid peroxidation in red cells. The measurement of pentane and ethane, formed during the peroxidative breakdown of po- lyunsaturated fatty acids, is known to serve as a sensitive and reliable index for lipid peroxidation in red cells (Clemens et al. 1983; Clemens et al. 1984).

Materials and Methods

All chemicals used were analytical grade. Cellu- loses were purchased from Sigma (Munich, FRG) and all other chemicals from Merck (Darmstadt, FRG). Blood was taken from normal healthy adults and from non-icteric patients with G-6-PD deficiency (Mediterranean type) and mixed with sodium citrate as an anticoagulant (1 : 5, v/v). Red cells were separated from plasma by the method of Beutler et al. (1976), with microcristalline and alpha-cellulose (1 : 1, w/w).

The clinical characteristics of G-6-PD deficient patients are shown in Table 1. Of the seven enzy- mopenic patients, four had a compensated haemo- lysis by the time of the study. Haemoglobin and MCV was estimated using a Coulter counter (Model S-plus). Reticulocyte count was performed with a blood smear stained with fresh methylene blue. GSH analysis was carried out using the meth-

od of Beutler et al. (1963) and G-6-PD activity was measured as described by Beutler et al. (1977).

The susceptibility to lipid peroxidation was es- timated by the incubation of red cells (haematocrit 2.5%) with 10mmol/1 hydrogen peroxide and 0.25 mmol/1 sodium azide (final concentrations) in a phosphate-buffered solution of pH 7.4 for 2 h at 37 ° C. The analysis of hydrocarbons in the head space vials was performed as previously described (Clemens and Remmer 1982; Clemens et al. i983). After 2 h of red cell incubation with hydrogen per- oxide, haemolysis was determined photometrically by measuring the haemoglobin concentration in the supernatant of red cell suspensions after con- version to cyanmethaemoglobin at 540 nm. Isola- tion of ghosts, lipid extraction, and fatty acid anal- yses were performed as previously described (Cle- mens et al. 1985). All values of normal and G-6-PD deficient subjects were normally distributed. Therefore, statistical differences were calculated using Student's t-test.

Results

The incubation of normal red cells with hydrogen peroxide resulted in the release of small amounts of pentane and ethane. The formation of these al- kanes in red cells obtained from patients with G-6- PD deficiency was found to be higher than in con- trols in almost all measurements (Fig. 1). Patient 4 was studied in haemolytic crisis and in chronic compensated haemolytic status. The susceptibility of red cells to lipid peroxidation was not found to be enhanced during haemolytic crisis when com- pared with values obtained during chronic haemo- lytic status.

Haemolysis occurring after incubation of red cells with hydrogen peroxide for 2 h was found to be in the upper range of the normal distribution in patients with G-6-PD deficiency (Fig. 2). Hae-

580 M.R. Clemens et al. : G-6-PD Deficiency of RBC and Lipid Peroxidation

16 t -

1

Ethane Pentane Fig. 1. Measurement of hydrocarbons formed in red cells as an index for lipid peroxidation. Red cells were incubated with 10 mmol/l hydrogen peroxide in the presence of 0.25 mmol/ sodium azide in PBS pH 7.4 for 2 h at 37 ° C. Shaded areas represent the normal range (mean±SD of 35 control experi- ments). • Patients with G-6-PD deficiency. Arrows indicate the value obtained from patient 4 during the haemolytic crisis. A1- kane measurement was not performed with red cells from pa- tient 7

SO

u~

O.J . - p

60

4O

30 i

20-

10

CoNmls G6 PU- Fig. 2. Measurement of" haemolysis of red ceils after hydrogen peroxide incubation. For details see Fig. 1. The a r r o w indicates the value obtained from patient 4 during the haemolytic crisis. Haemolysis was not determined for patient 7

molysis after red cell incubation with hydrogen peroxide was not found to be higher in patient during haemolytic crisis than during chronic hae- molytic status.

Since polyunsaturated fatty acids in the red cell membrane, particularly arachidonic acid, are the major site of peroxidative damage (Einsele et al. 1985), we studied the fatty acid composition of red cell membranes obtained from 13 normal sub- jects and 7 patients with G-6-PD deficiency. Signif- icant changes in fatty acid composition were obvi- ous: saturated fatty acids, particularly palm±tic and stearic acid, were found to be decreased in patients with enzyme deficiency, whereas the proportion of arachidonic acid was markedly increased (Table 2). The content of arachidonic acid in membranes of G-6-PD deficient red cells was generally above nor- mal. There was no correlation between the content of arachidonic acid and the number of reticulo- cytes. Fat ty acid analysis of red cell membranes obtained from patient 4 was performed during chronic haemolytic status and not during haemo- lyric crisis.

Fatty acid analysis of plasma did not reveal significant changes between normal and G-6-PD deficient subjects, which could have explained the alteration of membrane fatty acid composition of red cells (Table 3). In particular, arachidonic acid was not found to be increased in plasma of enzy- mopenic patients. Values of fatty acid analysis re- ported here for plasma and red cells in normal subjects are in good agreement with the findings of other authors (Ways and Hanahan 1964; Schu- botz et al. 1976; Owen et al. 1982; Blaton et al. 1984).

Table 2. Fatty acid composition of the red cell membrane of patients with G-6-PD deficiency and normal subjects

Patient no. Palm±tic acid Stearic acid Oleic acid Linoleic acid Arachidonic Docosahexaenoic acid acid

1 23.2 16.2 17.8 13.1 21.5 8.2 2 20.8 17.0 18.2 16.0 2•.2 6.9 3 25.4 •5.2 17.4 13.6 21.9 6.5 4 22.0 16.7 16.1 13.3 23.2 8.8 5 22,2 18.1 •6.6 13.3 21.9 7.8 6 21.0 •6.6 16.0 11.9 24.8 9.8 7 22.3 16.7 16.1 13.3 23.2 8.8

Mean_+ SD 22.9 ± 1.9 16.8 + 1.0 16.7 + 1.2 13.6_+ 1.3 22,3 _ 1.3 7.8 ± 1.2

Normal 28.5 ± 3.5 19.4 ± 1.4 14.0 ± 2.0 t2.6 + 1.3 17.9 ± 1.9 7.6 ± 1.9 subjects

P < 0.01 0.001 0.01 NS 0.001 NS

Fatty acid analysis was performed three times for each ghost extraction. Values are given as mean in % fatty acids. P was calculated by using Student 's t-test. Normal subjects n = 13. Patient 4 was analysed while haemolysis was being compensated for

M.R. Clemens et al. : G-6-PD Deficiency of RBC and Lipid Peroxidation 58J

Table 3. Plasma fatty acid composition of patients with G-6-PD deficiency and normal subjects

Fatty acid Control G-6-PD- P < subjects patients (8) (7)

Pahnitic acid 24.1 ±2.5 22.9.+ 1,5 NS Stearic acid 8.5 .+ 1.0 8.7 _+_ 0.4 NS Oleic acid 21.0 ± 1.9 25.4 .+ 3.0 0.01 Linoleic acid 32.3 ± 3.6 29.9_+ 2.7 NS Arachidonic acid 10.9_+ 1.3 9.7 ± 1.2 NS Docosahexaenoic acid 3.4 _+ 1.0 3.4 -+ 0.5 NS

Fatty acid amounts are expressed in % fatty acids_+SD. P was calculated by using Student's t-test

Discussion

Peroxidation of membrane lipids is currently be- lieved to represent a crucial step in the process of intravascular haemolysis in G-6-PD deficiency (Chiu et al. 1982). The link between drug or fava bean sensitivity and G-6-PD deficiency is generally considered as an inability to remove the excess of hydrogen peroxide that is generated (Mavelli et al. 1984). However, the drugs that are haemolytic in G-6-PD deficiency produce free radicals on reac- tion with haemoglobin. These radicals accelerate haemichrome and choleglobin formation, and can form damaging hydroperoxides with membrane phospholipids (Carrell et al. 1977). However, there is evidence that membrane damage in G-6-PD defi- cient red cells can and does occur independently of intracellular alteration of haemoglobin (John- son et al. 1979; Kosower et al. 1982; Alhanaty et al. 1984).

Loss of deformability and membrane changes are presumed to be responsible for the enhanced in vivo destruction of G-6-PD deficient red cells, with easier entrapment of the cells and enhanced recognition of the altered membranes by the se- questration system. Intrachain disulfide formation may also result in altered membrane properties, e.g. changes in ion transport, complement binding, and agglutinability. These protein alterations which result in an increased cell membrane rigidity, may partly explain the finding that the hydrogen peroxide-induced haemolysis of G-6-PD deficient red cells in our test system was within the normal range and not, as expected, a consequence of the enhanced susceptibility to lipid peroxidation above the normal range.

On the other hand, Tillmann et al. (1977a, b) have reported observations that G-6-PD deficient red cells have favourable rheological properties by an increased flexibility of these cells. This finding

may be explained partly by the increased content of arachidonic acid in those cells, a fact known to cause an enhanced membrane fluidity.

Changes in antioxidant enzymes, such as super- oxide dismutase, catalase and glutathione peroxi- dase, may influence the susceptibility of enzymo- penic red cells to lipid peroxidation. Conflicting data exist on the alterations of these enzymes in G-6-PD deficient cells.In the most extensive of sev- eral studies, normal levels of all three enzyme activ- ities have been observed in G-6-PD deficient sub- jects with either positive or negative histories of haemolytic crisis (Gerli et al. 1982). With respect to these findings, the increased sensitivity of G-6- PD deficient red cells found in this study was un- likely to be caused by abnormalities in antioxidant enzymes. Since an increased content of substrate for lipid peroxidation, i.e. arachidonic acid, is a consistent finding in G-6-PD deficient red cells, unrelated to the clinical state of the patient (acute or no haemolysis), we suggest that this may partly explain the enhanced sensitivity of these cells to lipid peroxidation.

Our results fail to confirm observations of De Flora et al. (1981) who found arachidonic acid in significantly higher amounts in favic but not in asymptomatic G-6-PD deficient subjects. How- ever, our results showing that the amount of po- lyunsaturated fatty acids affects the sensitivity of membranes to lipid peroxidation is consistent with recent observations of Szebeni et al. (1984), which showed that rapid haemoglobin oxidation and lip- id peroxidation occur in experiments with unsatur- ated liposomes but not with saturated phospholip- ids. Mowri et al. (1984) studied the effect of lipid composition of liposomes on their sensitivity to peroxidation and showed that the content of po- lyunsaturated fatty acids is responsible for the de- gree of lipid peroxidation.

No evidence was found that the reported red cell membrane abnormalities in G-6-PD deficient cells were due to alterations of the plasma fatty acid composition. Szeinberg et al. (1965) have pub- lished data indication that there is an increased proportion of phosphatidytserine in G-6-PD defi- cient cells. Since this phospholipid contains about 30% aracidonic acid (W Kegler, MR Clemens, H Remmer, unpublished data), the increased amount of arachidonic acid in enzymopenic cells reported here may be an effect of alteration in phospholipid composition.

Acknowledgements. The technical assistance of U. Wegner is gratefully acknowledged. Sincere thanks are due to M. Blau- rock for statistical calculations, H. Remmer for helpful discus- sions and G. Pawelec for correcting the language.

582 M.R. Clemens et al.: G-6-PD Deficiency of RBC and Lipid Peroxidation

R e f e r e n c e s

Alhanaty E, Snyder M, Sheetz MP (1984) Glucose-6-phosphate dehydrogenase deficient erythrocytes have an impaired shape recovery mechanism. Blood 63 : 1198-i 202

Ben6hr HC, Waller HD (1974) Hematological manifestations in enzymatic deficiencies of glutathione reduction. In: Flohe L, Benthr HC, Sies H, Waller HD, Wendel A (eds) Glutath- ione. Georg Thieme, Stuttgart, pp 184-191

Beutler E, Blume KG, Kaplan JC, Lthr GW, Ramot B, Valen- tine WN (1977) International commitee for standardization in haematology: recommended methods for red-cell enzyme analysis. Br J Haematol 35:331-340

Beutler E, Duron O, Kelty BM (1963) Improved method for the determination of blood glutathione. J Lab Clin Med 61 : 882-888

Beutler E, West C, Blume KG (1976) The removal of leucocytes and platelets from whole blood. J Lab Clin Med 88:328-333

Blaton V, De Buyzere M, Declercq B, Pracetyo A, Vanderkelen G, Delanghe J, Spincemaille J (1984) Effect of polyunsatur- ated isocaloric fat diets on plasma lipids, apolipoproteins and fatty acids. Atherosclerosis 53 : 9-20

Carrell RW, Winterbourn CC, French JK (1977) Haemoglobin - a frustrated oxidase? Implications for red cell metabolism. Hemoglobin 1 : 815-827

Chiu D, Lubin B, Shohet SB (1982) Peroxidative reactions in red cell biolog. In : Pryor WA (ed) Free Radicals in Biology, Voi. V, Academic Press, New York, pp 115-160

Clemens MR, Einsele H, Frank H, Remmer H, Waller HD (1983) Volatile hydrocarbons from hydrogen peroxide-in- duced lipid peroxidation of erythrocytes and their cell com- ponents. Biochem Pharmacol 32:3877-3878

Clemens MR, Einsele H, Remmer H, Waller HD (1985) De- creased susceptibility of red blood cells to lipid peroxidation in patients with alcoholic liver cirrhosis. Clin Chim Acta 145 : 283-288

Clemens MR, Returner H (1982) Volatile alkanes produced by erythrocytes : an assay for in vitro studies on lipid peroxida- tion. Blut 45 : 329-335

Clemens MR, Remmer H, Waller HD (1984) Phenylhydrazine- induced lipid peroxidation of red blood cells in vitro and in vivo : monitoring by the production of volatile hydrocar- bons. Biochem Pharmacol 33:1715-1718

De Flora A, Morelli A, Benatti U, Pontremoli S, Melloni E, Salamino F, Sparatore B, Michetti M, Melloni T (1981) Membrane lipid components of normal and gtucose-6-phos- hate dehydrogenase-deficient erythrocytes of asymptomatic and favic subjects. Acta Biol Germ 40: 563 570

Einsele H, Clemens MR, Remmer H (1985) Effect of ascorbate on red blood celt lipid peroxidation. Free Radical Res Comm (in press)

Gerli GC, Beratta L, Bianchi M, Agostoni A, Gualandri V, Orsini GB (1982) Erythrocyte superoxide dismutase, cata- lase, and glutathione peroxidase in glucose-6-phosphate de- hydrogenase deficiency. Scand J Haematol 29:135-140

Jain SK, Hochstein P (1980) Polymerization of membrane com- ponents in aging red blood cells. Biochem Biophys Res Commun 92: 247-254

Johnson G J, Allen DW, Cadman S, Fairbanks VF, White JG,

Lampkin BC, Kaplan MC (1979) Red-cell-membrane poly- peptide aggregates in glucose-6-phosphate dehydrogenase mutants with chronic hemolytic disease. N Engl J Med 301 : 522-527

Kosower NS, Zipser Y, Faltin Z (1982) Membrane thioI-disul- fide status in gtucose-6-phosphate dehydrogenase deficient red cells. Biochim Biophys Acta 691:345-352

Mavelli I, Ciriolo MR, Rossi L, Meloni T, Forteleoni G, de Flora A, Benatti U, Morelli A, Rotilio G (1984) Favism: a hemolytic disease associated with increased superoxide dismutase and decreased glutathione peroxidase activities in red blood cells. Eur J Biochem 139:13-18

Mowri H,Nojima S, Inoue K (1984) Effect of lipid composition of liposomes on their sensitivity to peroxidation. J Biochem 95:551-558

Owen JS, Bruckdorfer KR, Day RC, McIntyre N (1982) De- creased erythrocyte membrane fluidity and altered lipid composition in human liver disease. J Lipid Res 23 : 124-132

Pfafferott C, Meiselman If J, Hochstein P (1982) The effect of malondialdehyde on erythrocyte deformability. Blood 59:12-15

Schubotz R, Goebel KM, Kaffarnik H (1976) Ver/inderungen in den Membranlipiden der Erythrocyten bei/ithanolindu- zierter Hyperlipid/imie (Zieve-Syndrom). Klin Wochenschr 54:827-833

Stocks J, Dormandy TL (1970) A direct thiobarbituric-acid reacting chromogen in human red blood cells. Clin Chim Acta 27:117-120

Stocks J, Dormandy TL (19Tt) The autoxidation of human red cell lipids induced by hydrogen peroxide. Br J Haematol 20:95-111

Szebeni J, Winterbourn CC, Carrell RW (1984) Oxidative inter- actions between haemoglobin and membrane lipid. A lipo- some model. Biochem J 220:685-692

Szeinberg A, Zaidman J, Clejan L (1965) Investigation of the lipid content of n o d a l and glucose-6-phosphate dehy- drogenase deficient red cells. Biochim Biophys Acta 98 : 598-606

Tillmann W, Gahr M, Labitzke N, Schr6ter W (1977) Men> brahe deformability of erythrocytes with glucose-6-phos- phate dehydrogenase Hamburg. Acta haemat 57 : 162-167

Tillmann W, Labitzke N, Schr6ter W (1977) Gfinstige rheolo- gische Eigenschaften der Erythrozyten beim Glucose-6- phosphatdehydrogenase-Mangel. Klin Wochenschr 55: 385-391

Ways P, Hanahan DJ (1964) Characterization and quantifica- tion of red celt lipids in normal man. J Lipid Res 5:318

Received December 13, 1984 Revised March 13, 1985 Accepted April 9, 1985

Dr. M.R. Clemens Medizinische Klinik Ot fried-Miiller-Stral3e 10 D-7400 Tfibingen 1 Bundesrepublik Deutschland