It has been widely acknowledged that deformability and aggregability of erythrocytes are very important determinants of microvascular perfusion and that they are of a considerable clinical significance in many circulatory disorders including atherosclerosis, essential hypertension (1, 2) and diabetes mellitus (3). Such approach to red blood cell (RBC) biology raises a crucial question whether rheological properties of RBC are affected by secretory function of endothelium known to play a key role in maintaining of vascular homeostasis. This function of endothelium is achieved through the release of a variety of autocrine and paracrine substances (4), with endotheliumŒdependent prostacyclin (PGI2) and nitric oxide (NO), the best characterised and probably the most important among them (5, 6).

Essentially, PGI2 and NO have platelet suppressant, fibrinolytic, thrombolytic, vasodilator and cytoprotective properties mediated by cyclic nucleotides, i.e. c- AMP and c-GMP, respectively. We have recently described that PGI2 and NO, when released from leukocytes affect rheological functions of RBC (7). In addition, iloprost, stable prostacyclin analogue, as well as SIN-1, an active metabolite of molsidomine that spontaneously releases NO, were very potent modulators of RBC deformability in rats when given exogenously (7). On the other hand, erythrocytes were claimed to influence the secretory function of vascular endothelium and other blood cells such as platelets and/or polymorphonuclear leukocytes (8, 9).

In the present study we aimed to investigate the hypothesis that RBC deformability is regulated endogenously by balanced generation of two endothelial secretogogues: PGI2 and NO. Accordingly we assessed the deformability of erythrocytes in rats under the influence of non-selective inhibitors of cyclooxygenase (indomethacin, aspirin) or NO-synthase (NG -nitro- L-arginine methyl ester; L-NAME). Experiments were performed ex vivo and with whole blood or isolated erythrocytes in vitro to judge whether rheological properties of RBS could be also affected by non-endothelial sources of these two enzymes.


MATERIALS AND METHODS
Ex vivo experiments

Male Wistar rats were anaesthetised with thiopental (50 mg/kg i.p., Biochemic GmbH) and heparinized (800 U/kg i.v., Polfa). Blood samples (0.5 ml) were taken from the right carotid artery before (control sample) and after administration of following compounds into the left femoral vein: 1) SIN-1 (Sigma) infused at a dose of 2 mg/kg/min for 10 min and samples taken after 5 and 10 min (in some experiments 10 min infusion of SIN-1 was preceded with bolus injection of L-NAME (Sigma) at a dose of 10 mg/kg), 2) iloprost (Schering) in bolus injection at a dose of 10 µg/kg and sample taken after 15 min, 3) NG nitro-L-arginine methyl ester (L-NAME)in bolus injection at a dose of 10 mg/kg and samples taken after 15, 30 and 45 min, 4) aspirin (1 or 50 mg/kg, Sigma) or indomethacin (20 mg/kg, Sigma) in bolus injection and samples taken after 15, 30 and 45 min.

In vitro experiments

Male Wistar rats were anaesthetized with thiopental (50 mg/kg i.p.). Blood (8 ml) was collected from the right carotid artery into heparin (25 U/ml) solution. Erythrocytes were isolated using the method described by Jubelin and Giennan (10). Red blood cells were separated from white blood cells on a Ficoll packed column, then washed three times in PBS solution (Imperia Lab. UK), enriched in albumin (Sigma) and glucose (POCH S.A.) Cells were suspended in solution of PBS with 3% dextran for 30 min and again rinsed to remove dextran. Finally, cells were resuspended in PBS solution. Haematocrit was adjusted to the value of whole blood haematocrit. Microscopic examination of prepared suspension revealed neither white blood cells nor platelets. Samples of whole blood or suspension of erythrocytes in PBS were incubated with SIN-1 (3 µM), iloprost (1 µM), aspirin (0.05 mM or 3.0 mM), indomethacin (3.0 mM), L-NAME (10 µM) or placebo (saline) at 22°C for 15 min.

Red blood cell deformability

Erythrocyte deformability was measured by a laser shear stress diffractometer (Rheodyn SSD). This instrument measures ellipsoidal elongation of red blood cells in response to defined, physiologically relevant shear stress conditions, forced by rotation rate. For all experiments deformability of RBC was assessed at the shear stress of 60 Pa. Samples of blood (30 µl) were suspended in 2 ml of dextran solution (MW 60 000, osmolarity 300 mOsm, pH=7.4, viscosity 24 mPa). The instrument projects actual level of deformability of red blood cells as a deformability index (DI%) calculated automatically according to equation: DI=100(L-W)/(L+W) where L and W are the means of length and width of elongated red cells, respectively. In our data every change in erythrocyte deformability resulting from the activity of investigated compound was expressed as DI(%) - the result of a subtraction between reading of control DI and reading of DI in the presence of the compound or placebo. Thereby the values of DI above zero (plus sign) or less than zero (minus sign) indicate improvement or worsening of RBC deformability by particular compound, respectively.

Statistical analysis

Results were expressed as arithmetical means ± SD of n numbers of experiments and analyzed by Student`s T test for paired means to determine the significance of the response; ieplo values of less than 0.05 were considered as statistically significant.

The Institutional Review Committee approved the protocol for animal experiments.


RESULTS
NO and PGI2 administered exogenously

In experiments ex vivo 10 min intravenous infusion of SIN-1 (NO-donor) at a dose of 2 mg/kg/min induced significant improvement of deformability of erythrocytes. The change of deformability index ( DI) as compared with control reading before drug infusion was 1.41% ± 0.23, p<0.001 (Fig. 1a). Also iloprost - stable analogue of PGI2, efficiently increased deformability index of RBC 15 min after its bolus intravenous injection at a dose of 10 µg/kg (DI=2.05% ± 3.0, p<0.05) (Fig. 2a).

Fig. 1. The effect (DI% mean ± SD) of NO-donor - SIN-1 on RBC deformability in rats ex vivo (2 mg/kg/min i.v. infusions - A) and in vitro (3µM - B). ** p<0.01, * p<0.05, n=6

Improvements of RBC deformability by these two compounds were also evident in vitro when they were applied and incubated for 15 min in whole blood or in suspension of isolated erythrocytes. For SIN-1 at a concentration of 3 µM in whole blood and in isolated erythrocytes DI was 2.41% ± 1.4, p<0.05 and 4.08% ± 2.6, p<0.01, respectively (Fig. 1b). For iloprost at a concentration of 1 µM the corresponding data were 2.09% ± 1.06, p<0.01 and 2.95% ± 1.93, p<0.01 (Fig. 2b).

Fig. 2. The effect (DI% mean ± SD) of iloprost on RBC deformability in rats ex vivo 15 minutes after its bolus injection at a dose of 10 µg/kg i.v. (A) and in vitro (1 µM - B). ** p<0.01, * p<0.05, n=6

Endogenous generation of NO and PGI2

In ex vivo experiments an inhibitor of NO-synthase Œ L-NAME (10 mg/kg, i.v.) caused significant decrease of RBC deformability, reaching the maximal effect 45 min after its injection (DI = -2,21% ± 0.55, p<0.01) (Fig. 3a). Contrary to the above, L-NAME did not significantly affect RBC deformability in vitro (Fig. 3b).

Fig. 3. The effect (DI% mean ± SD) of L-NAME on RBC deformability in rats ex vivo (10 mg/kg i.v. - A) and in vitro (10 µM - B). ** p<0.01, * p<0.05, n=6

Worsening of deformability by NO-synthase inhibition ex vivo was spectacularly reversed by infusion of a NO-donor - SIN-1 at a dose of 2 mg/kg/min i.v. (DI = -0.14% ± 0.89, after 5 min infusion). As compared with the change in deformability induced by L-NAME (DI = -1.31% ± 0.83) statistical significance was at the level less than 0.001 (Fig. 4). Five minutes after infusion of SIN-1 deformability returned to the initial value.

Fig. 4. The effect (DI% mean ± SD) of SIN-1 infusion (2 mg/kg/min i.v.) on L- NAME-induced (10 mg/kg i.v.) worsening of RBC deformability in rats ex vivo. ** p<0.01, * p<0.05, n=6

Inhibition of cyclooxygenase in experiments ex vivo by intravenous injection of aspirin significantly worsened RBC deformability in a way similar to observed with L-NAME. The effect of aspirin appeared both at a low (1 mg/kg) and high (50 mg/kg) dose of the drug. DI 45 min after bolus injection was - 2.32% ± 1.32, p<0.01 and - 1.39% ± 0.72, p<0.01, respectively (Fig. 5). Surprisingly, aspirin at a concentration of 0.05 mM and 3.0 mM worsened RBC deformability also in suspension of isolated erythrocytes: DI = -1.37% ± 0.98, p<0.001 and -1.24% ± 1.0, p<0.001, respectively (Fig. 6). Other cyclooxygenase inhibitor - indomethacin (20 mg/kg i.v.) aggravated RBC deformability ex vivo similarly to aspirin (DI = -1.34% ± 0.9, p<0.05 15 min after injection), but contrary to aspirin it did not affect deformability in vitro up to 3 mM concentrations (Fig. 7).

Fig. 5. The effect (DI% mean ± SD) of aspirin on RBC deformability in rats ex vivo at a dose of 1 mg/kg i.v. (A) or 50 mg/kg i.v. (B). ** p<0.01, * p<0.05, n=6

Fig. 6. The effect (DI% mean ± SD) of aspirin (0.05 mM or 3 mM) on deformability of isolated red blood cells of rats. ** p<0.01, n=6

Fig. 7. The effect (DI% mean ± SD) of indomethacin on RBC deformability in rats ex vivo (20 mg/kg i.v. - A) and in whole blood (3 µM - B). ** p<0.01, * p<0.05, n=6

DISSCUSSION
So far distempered deformability of red blood cells in hypercholesterolaemia (11), arterial hypertension (1, 2), diabetes mellitus (3), coronary artery disease (12) and ageing (13) has been suggested to result from the injury of vascular endothelium with subsequent deficit of endothelium- derived relaxing and antiplatelet secretogogues - nitric oxide (NO) and prostacyclin (PGI2). Here we provide the evidence that RBC deformability may be also aggravated when endogenous generation of these compounds is pharmacologically suppressed by cyclooxygenase (indomethacin or aspirin) and NO-synthase (L-NAME) inhibitors. Moreover, we demonstrate that prostacyclin and nitric oxide directly improve deformability of red blood cells both in vivo and in isolated cells, and that they are capable to reverse damaging effects of their inhibitors. Thereby we propose the hypothesis on the balanced endothelial generation of these two secretogogues being responsible for the endogenous regulation of RBC deformability in physiological conditions.

What is the mechanism by which NO and PGI2 regulate deformability and which endogenous sources of NO and PGI2 are involved in the regulation of RBC deformability? Up to now numerous clinical studies (10, 14, 15, 16, 17) as well as our own exeprimental results have not enabled us to answer this question thoroughly. In our previous work on RBC deformability rabbit erythrocytes and other blood cells could interact freely and that was why we suggested that the deformability of red blood cells might be modulated by the presence of polymorphonuclear leukocytes (7). However, in experiments in vitro reported in the present study and elsewhere (14) NO and PGI2 did affect erythrocytes even if they were completely isolated from the blood and other blood cells. This is the reason for us to speculate that the regulation of deformability by NO and PGI2 seems likely to be related to the direct effect of these compounds on erythrocytes and/or on their membranes.

In most cases biological effects of NO are strictly related to the induction of c-GMP that subsequently, via protein kinase C, phosphorylates cell proteins including proteins of the cell membrane. Could it be also true for erythrocytes? Cyclic GMP-dependent mechanism by which NO may improve RBC deformability has been mainly claimed by advocates of the hypothesis that erythrocytes contain guanylate cyclase and two isoforms of NO-synthase (NOS- 2 and NOS-3) thanks to which they are capable to self regulate the deformability by their own generation of NO (10). However, in our hands the NO-synthase inhibitor ΠL-NAME, while worsening RBC deformability in vivo, did not influence rheological properties of erythrocytes either in whole blood or in suspension of isolated cells. This clearly indicates that in physiological conditions RBC deformability is regulated by NO originating rather from vascular endothelium than from blood cells, including erythrocytes. The involvement of c-GMP in mediating the NO-induced changes of RBC deformability seems to us quite possible. Increase of c-GMP content was found to correlate with changes of RBC deformability after administration of various c-GMP-activators such as sodium nitroprusside, sodium nitrite, atrial natriuretic peptide (ANP) or phosphodiesterase inhibitors (18, 19). On the other hand, among mechanisms by which NO may affect RBC deformability, there are also suggestions indicating that the effects of NO are independent on c-GMP. For instance, there is a possibility that NO directly improves deformability by activation of Ca +2 -dependent K + channels known to be induced by nitric oxide in vascular smooth muscle cells and claimed to regulate also rheological properties of RBC (20, 21, 22).

In our experiments, similarly to the effects of nitric oxide, RBC deformability was significantly improved by the stable analogue of PGI2 - iloprost. So far the mechanism of this phenomenon is unknown and to our knowledge PGI2 receptors have not been found on erythrocyte membrane. However, since beneficial effects on RBC deformability were observed also for PGE1 and for ß-adrenergic agonists (all of them very potent stimulators of adenylate cyclase), possible role of c-AMP in mediation of biological functions of PGI2 in erythrocytes appears acceptable (11, 23). In fact, an increased level of c-AMP induced by unspecific inhibition of phosphodiesterase is a commonly accepted mechanism for the activity of pentoxifylline - the only drug with clinically recognised RBC deformability improving properties successfully used for the treatment of peripheral arterial occlusive disease (24).

Interestingly, also NO has been found to inhibit phosphodiesterase E4 (PDE4). So, it could well be that both PGI2 and NO improve RBC deformability by identical c-AMP-dependent mechanism. At least for the moment we are unable to agree with this opinion and certainly further basic experimental work is required to solve the problem.

As prostacyclin plays a protective role for RBC, the inhibition of its endogenous generation by cyclooxygenase inhibitors was supposed to disturb rheological properties of these cells. Indeed, indomethacin and aspirin when injected into the animal significantly aggravated deformability of erythrocytes. In addition indomethacin did not affect deformability in vitro, indicating that the most meaningful portion of prostacyclin involved in the regulation of RBC deformability originated from vascular endothelium. However, aspirin surprisingly decreased deformability ex vivo even when used at an antiplatelet dose of 1 mg/kg which was expected to be too low to inhibit cyclooxygenase. Moreover, it had identical effects in isolated erythrocytes. Is there any way to explain such unexpected effects? Along with our observations detrimental effects of aspirin on RBC deformability was also revealed by clinical studies (17, 25) and in a few laboratory experiments (14, 26, 27), but not many comments of this phenomenon were proposed. The most plausible explanation is that aspirin induces acetylation of integral proteins of the red cell membrane that leads to the rigidizing effect on membrane lipid fluidity (26). It should be mentioned as well 662.that aspirin, like e.g. methyl acetylphosphate, is able to acetylate human haemoglobin which in its acetylated form is known to modify proteins and/or lipids of the red cell membrane changing their rheological properties (28, 29). Thus, in our opinion worsening of deformability by aspirin may result not only from inhibition of cyclooxygenase in vascular endothelium and from depletion of prostacyclin but, above all, it can be due to unspecific effects of aspirin on red cell membrane.

In conclusion, two endothelial secretogogues: prostacyclin and nitric oxide significantly improve deformability of red blood cells and thereby they may play an important role in regulation of rheological functions of red blood cells in flowing blood. The mechanism of this phenomenon is unclear but it seems to be related to the direct effects of these compounds on erythrocytes. Aspirin has a unique property, possibly independent of inhibition of cyclooxygenase, that consists in worsening of the deformability by a direct acetylating effect of red cell membrane. Further exploration of the mechanisms by which NO-donors and prostacyclin analogues improve RBC deformability may contribute to the extension of their therapeutic applications in circulatory disorders.


REFERENCES

1. Chabanel A, Schachter D, Chien S. Increased rigidity of red blood cell membrane in young spontaneously hypertensive rats. Hypertension 1987; 10: 603-607.
2. Gomi T, Ikeda T, Ikegami F. Beneficial effect of alpha-blocker on hemorheology in patients with essential hypertension. Am J Hypertension 1997; 10: 886.
3. MacRury SM, Lennie SE, McColl P, Balendra R, MacCuish AC, Lowe GD. Increased red cell aggregation in diabetes mellitus: association with cardiovascular risk factors. Diabetic Med 1993; 10: 21-26.
4. Gryglewski RJ. Interactions between endothelial mediators. Pharmacol Toxicol 1995; 77:1-9.
5. Moncada S, Higgs A. The L-arginine-NO pathway. N Engl J Med 1993; 329: 2002-2012.
6. Gryglewski RJ, Botting RM, Vane JR. Mediators produced by the endothelial cell. Hypertension 1988; 12: 530-548.
7. Korbut R, Gryglewski RJ. Nitric oxide from polymorphonuclear leukocytes modulates red blood cell deformability in vitro. Eur J Pharmacol 1993; 234: 17-22.
8. Sprague RS, Ellsworth ML, Stephenson AH, Lonigro AJ. ATP the red blood cell link to NO and local control of the pulmonary circulation. Am J Physiol 1996; 271: H2717-H2722.
9. Turitto VT, Weiss HJ. Red blood cells, their dual role in thrombus formation. Science 1980; 207: 541.
10. Jubelin BC, Gierman JL. Erythrocytes may synthetize their own nitric oxide. Am J Hypertension 1996; 9: 1214-1219.
11. Hayashi J, Ishida N, Hata Y, Saito T. Effect of beraprost, a stable prostacyclin analogue, on red blood cell deformability impairment in the presence of hypercholesterolaemia in rabbits. J Cardiovasc Pharmacol 1996; 27: 527-531.
12. Leschke M, Motz W, Strauer BE. Hemorheologic therapy applications in coronary heart disease. Wiener Med Wochenschr 1996; 136: 17-24.
13. Knapowski J, Wieczorowska-Tobis K, Witowski J. Pathophysiology of ageing. J Physiol Pharmacol 2002; 53, 2, 135-146
14. Starzyk D, Korbut R, Gryglewski RJ. Effects of nitric oxide and prostacyclin on deformability and aggregability of red blood cells of rats ex vivo and in vitro. J Physiol Pharmacol 1999; 50: 629-637.
15. Garnier M, Attali JR, Valensi P, Delatour-Hanss E, Gaudey F, Koutsouris D. Erythrocyte deformability in diabetes and erythrocyte membrane lipid composition. Metabolism: Clin Exper 1990; 39: 794-798.
16. Hadengue AL, Del-Pino M, Simon A, Levenson J. Erythrocyte disaggregation shear stress, sialic acid, and cell aging in humans. Hypertension 1998; 32: 324-330.
17. Korbut RA, Adamek-Guzik T. The effect of aspirin on rheological properties of erythrocytes in essential hypertension. Przegl Lek 2002; 59: 71-75.
18. Petrov V, Fagard R, Lijnen P. Human erythrocytes contain Ca2+, calmodulin-dependent cyclic nucleotide phosphodiesterase which is involved in the hydrolysis of cGMP. Meth Find Exp Clin Pharmacol 1998; 20: 387-393.
19. Zamir N, Tuvia S, Riven-Kreitman R, Levin S, Korenstein R. Atrial natriuretic peptide: direct effects on human red blood cell dynamics. Biochem Biophys Res Commun 1992; 188; 1003- 1009.
20. Bolotina VM, Najibi S, Palacino JJ, Pagano PJ, Cohen RA. Nitric oxide directly activates calcium-dependent potassium channels in vascular smooth muscle. Nature 1994; 368: 850-853.
21. Li Q, Jungmann V, Kiyatkin A, Low PS. Prostaglandin E2 stimulates a Ca2+ -dependent K+ channel in human erythrocytes and alters cell volume and filterability. J Biol Chem 1996; 271: 18651-18656.
22. Shields M, La Celle P, Waugh RE, Scholz M, Peters R, Passow H. Effects of intracellular Ca2+ and proteolytic digestion of the membrane skeleton on the mechanical properties of the red blood cell membrane. Biochim Biophys Acta 1987; 905: 181-194.
23. Tuvia S, Moses A, Gulayev N, Levin S, Korenstein R. Beta-adrenergic agonists regulate cell membrane fluctuations of human erythrocytes. J Physiol. 1999; 516: 781-792.
24. Scheffler P, de la Hamette D, Gross J, Mueller H, Scjieffer H. Intensive vascular training in stage IIb of peripheral arterial occlusive disease. The additive effects of intravenous prostaglandin E1 or intravenous pentoxifylline during training [see comments]. Circulation 1994; 90: 818-822.
25. SaniabadiAR, Fischer TC, McLaren M, Belch JF, Forbes CD. Effect of dipyridamole alone and in combination with aspirin on whole blood platelet aggregation, PGI2 generation, and red cell deformability ex vivo in man. Cardiovasc Res. 1991; 25: 177-183.
26. Watala C, Gwozdzinski K. Effect of aspirin on conformation and dynamics of membrane proteins in platelets and erythrocytes. Biochem Pharmacol 1993; 45: 1343-1349.
27. Slomiany BL, Slomiany A. Delay in oral mucosal ulcer healing by aspirin is linked to the disturbances in P38 mitogen-activated protein kinase activation. J Physiol Pharmacol 2001; 52, 185-194.
28. Xu AS, Labotka RJ, London RE. Acetylation of human hemoglobin by methyl acetylphosphate. Evidence of broad regio-selectivity revealed by NMR studies. J Biol Chem 1999; 274: 26629- 26632.
29. Xu AS, Macdonald JM, Labotka RJ, London RE. NMR study of the sites of human hemoglobin acetylated by aspirin. Biochim Biophys Acta 1999; 1432: 333-349.