Vol. 181, No. 3, 1991 December 31, 1991
AND BIOPHYSICAL RESEARCH COMMUNICATIONS Pages 1392-1397
CYTOPROTECTlVE FUNCTION OF NlTBlC OXIDE: INACTIVATION OF SUPEROXIDE RADICALS PRODUCED BY HUMAN LEUKOCYTES Gabor M. Rubanyi’, Elena H. Ho, Elinor H. Cantor, William C. Lumma and Lynne H. Parker Botelho ‘Schering AG, Berlin, Germany Merck Sharp & Dohme, West Point, PA 19486 Berlex Laboratories, Inc., Cedar Knolls, NJ 07927 Received
The oxygenderived free radical superoxide anion (.Oz’)plays an important role in the pathogenesis of various diseases. Recent demonstrations that .Ozsinactivates the potent vasodilator endotheliumderived relaxing factor (EDRF) and that EDRF is probably nitric oxide (NO) suggest that EDRF(N0) may act as an endogenous free radical scavenger. This hypothesiswas tested in an in vitro system by analyzing the effect of authentic NO (dilutions of a saturated aqueous solution) on .Oz’production (detectedspectrophotometrically as reduction of cytochrome @ by fMet-Leu-Phe-activated human leukocytes (PMN). NO depressed the rate of reduction of cytochrome g by ,Oz’ released from PMN’s or generated from the oxidation of hypoxanthine by xanthine oxkfase. This effectwasconcentration-dependent andoccurred at dilutions of the saturated NO solution (1:250 to 1 :lO) which inhibited platelet aggregation. NO had no direct effecton cytochrome g or on xanthine oxidase. These observations indicate that NO(EDRF) can be regarded as a scavenger of superoxide anion and they suggest that EDRF(N0) may provide a chemical barrier to cytotoxic free radicals 0 1991 Academic Press, Inc. u&-l.
Endothelium-derived relaxing factor (EDRF) is a labile non-prostanoid substance produced by the vascular endothelium (1). Its known biological functions are vasodilation (2,3) and inhibition of platelet aggregation (4) suggesting that EDRF contributes to the physiological regulation of blood vessel caliber and the fluidity of the blood. Pharmacologic and chemical evidence indicates that EDRF may be identical with nitric oxide (NO) (5-7)and one of its characterisiticfeatures is that the oxygen derived free radical superoxide anion (.O,-) inactivates it (8,9). Superoxide anion is a cytotoxic free radical which, if released in large quantities [for example, from activated polymorphonuclear leukocytes (PMN)], can cause tissue injury (10). Having an unpaired electron, NO can accept electrons, and thereby inactivate/scavenge .O; (11). It has been postulated that EDRF(N0) could be a natural extracellular scavenger of .O;, and one of its biological functions may be to act as a chemical barrier to cytotoxic free radicals (12,13). The present study was designed to test this hypothesis by analyzing the effectsof NO on .O; generated by activated PMN’s or by the xanthine oxkfase and hypoxanthine enzyme reaction. The data show that NO (EDRF) inactivates .O; produced by PMN’s which may represent a cytoprotectivefunction for this ubiquitous mediator. 0006-291X/91 $1.50 Copyright 0 1991 by Academic Press, Inc. All rights of reproduction in any form reserved.
181, No. 3, 1991
METHODS AND MATERIALS MaWrials: Cytochalasin B, fMet-Leu-Phe, cytochrome c, superoxide diimutase, adenosine diphosphate, xanthine oxidase, hypoxanthine, and xanthine were purchased from Sigma Chemical Co., St. Louis, MO. Preparation of NO: Deionized water(150 ml) wassparged with argon gas for 30 min. NO gas (99.0%) was introduced while maintaining the argon flow, and NO was bubbled through the water for 20 min. The saturated NO solution was transferredto argon-purged vials (50 ml/vial). The vials werecapped with rubber septa, sealed, and stored in the refrigerator. Studies with Polymorphonuclear Leukocytes (PMN’s): Human PMN’s were prepared from heparinized blood as follows: plasma was removed and replaced with an equal volume of 3% dextran in phosphate bufferedsaline (PBS). After sedimentation (20 min, 3PC) to remove red blood cells, the leukocyte-rich layer was applied to a discontinuous percolt gradient (550/a/O%), and centrifuged (400 g, 30 min, 25°C). PMN’s were harvested from the 55%-70% interface, washed with PBS, and resuspended to give the desired final cell count. PMN’s were activated by preincubation with cytochalasin B (5 pg/ml) for 10 min at 37°C. NO or test compounds were added at the indicated final concentrations or dilutions, followed immediately by the stimulator fMet-Leu-Phe (lc’ M). In the experiments designed to determine .O; production, cytochrome c (8~10~ M) was added as an indicator and production of superoxide radii1 by the activated PMN’s wai followed by measuring the reduction of q&chrome c (AS50 nm) for 10 min. Superoxide dismutase (SOD, 30 U/ml) was used to determine the extent of cytochrome c reduction caused by superoxide anion (SODsensitive reduction). In experiments designed to determine PMN aggregation, the reaction wascarried out in an aggregometer. Studies wfth Xanthine Oxidase: Superoxide wasgenerated by the reaction of xanthine oxidase (10.’ M) with hypoxanthine (IO4 M) at 25°C and the reduction of cytochrome c was followed as a change in 4%. Diluted aliquots of the saturated NO solution wereadded simultaneously with the addition of the substrate. SOD (30 U/ml) was included as a positive control as in studies with PMN’s (see above). The effect of NO on reduced cytochromes wasdetermined by adding an NO solution (1:lO final dilution) followingthe maximal reduction of cytochrome c (approximately 5 min) and the reaction was followed for an additional 5 min. The formation of urate from xanthine (1O4 M) by xanthine oxidase (IO-’ M) was monitored at 295 nm for 5 min at 25OC. The effects of the addition of 100 pl of control PBS (pH 7.4), acidified PBS (pH 2.4), or a saturated NO solution (pH 2.4) was assessed by determining the change in 4%. Studies of Platelet Aggregation: Human platelets wereprepared from whole blood mixed with 3.2% sodium citrate (911v/v) and centrifuged at 150 xg for 20 min. The upper layer consisting of platelet rich plasma (PRP) wasremoved and centrifuged at 1,000 xg for 15 min. The lowerlayer wasthe source of PRP and the upper layer or platelet poor plasma (PPP) was used as a blank. Platelet aggregation was measured with an aggregometer following the addition of 1o-5M ADP in the presence or absence of NO. RESULTS Stimulation of human PMN’s by f-Met-Leu-Phe (fMLP; 10.’ M) caused a time-dependent reduction of cytochrome c (Fig. IA) as determined by an increase in the absorbance of 550 nm (A&. The change in 4% was prevented by the presence of superoxide dismutase (SOD; 30 U/ml) (Fig. 1A) indicating that the reduction reaction was caused by .O,- released from PMN’s. Nitric oxide decreased the rate and extent of cytochrome c reduction in a concentration-dependent manner (Fig. 1Aj. Approximately 50% inhibition was observed at a final dilution of 120 of the saturated NO solution, with little or no effectseen at 1:lOO dilution. A saturated NO solution has a pH of approximately 2.4, therefore an acidified (pH 2.4) control solution 1393
Vol. 181, No. 3, 1991
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
Time (min) Figure 1. Effect of nitric oxide (NO) on the rate of cytochrome g reduction in the presence of human [email protected]
leukocytes (PMN’s) or xanthlne oxidase. The effect of varfous concentrations of a saturated nitric oxide (NO) solution on the reduction of @chrome E was determinedas an increase in & in the presence of fMLP-activated PMN’s (PanelA) or in the presence of xanthineoxidese plus hypoxanthine(PanelB). In
each experiment,the changein & of a cytochrome2 solutionwasrecordedfor 10 min (PanelA) or 5 min (Panel5) in the presenceof: PBS (a);a 1:I00 dilution of NO (b);a 1:20
dilution of NO (c); a 1:lO dilution of NO (d); or 30 Uml” of SOD (e). These are representativespectrophotometriitracesof experimentsthat wererepeatedfivetimes.
(without NO) was tested to rule out a non-specific effect due to acidification of the reaction medium. The acidified control had no effect on the reduction of cytochrome c, and neither the acidified control nor the saturated NO solutions (up to 1:lO dilution) had any significant effect on the the pH of the buffer medium. Incubation of xanthine oxidase with the substrate hypoxanthine resulted in the reduction of cytochrome c (Fig. 1B). The presence of 30 U/ml of superoxide dismutase in the reaction mixture (Fig. I B) completely inhibited the reduction of cytochrome c implying that .O, being produced by the xanthine oxidase reaction was responsible for the reduction of cytochrome c. Nitric oxide also inhibited the xanthine oxidaseinduced reduction of cytochrome c in a concentration-dependent manner (Fig. IB) with 50% inhibition occuring at a dilution of 120. Inhibition of xanthine oxidase itself could be ruled out since neither NO (1:I0 dilution) nor an acidified (pH 2.4) control solution had any effecton xanthine oxidase-catalyzedformation of urate from the substrate xanthine (data not shown). Direct interaction between NO and cytochrome c (i.e., reoxidation or reduction) wasalso ruled out, since NO at a 1:lO final dilution had no effecton the absorbance of cytochrome c already reduced by a 5 min exposure to xanthine oxidase in the presence of the substrate, hypoxanthine. Addition of NO to oxidized cytochrome c in the absence of xanthine oxidase and hypoxanthine had no effect on & (data not shown). Figure 2 showsa comparison between the inhibitory effect of NO on the initial rate of reduction of cytcchrome c (expressedas the change in bO) in the presence of fMLP-activated PM’s (Panel A) or in the presence of xanthine oxidase and the substrate hypoxanthine (Panel B). The slopes of the inhibition curves are not significantly different,and approximately 1:20 dilutions of asaturated NO solution evoked half-maximal inhibition in both cases. 1394
Vol. 181, No. 3, 1991
Figure2. Effectof NOonthe initialrateof reductionof cytochromec.in thepresence of PMN’s or xanthineoxidase. Thetwo figuresshowthe decrease m the initialcytochrome c reductionrateexpressed as the changein absorbance (&) min-’ as a functionof increasing concentrations of NOinthepresence of PMN’s(PanelA) or xanthineoxidase plushypoxanthine (Panel6). The slopesof the inhibitioncurves,-0.12and -0.15, respectively, werenotsignificantly different(p= 0.54,ANOVAtest).Thisfigurerepresents datefromfive experiments.
Figure3 showsa comparison of theconcentration-dependent effectsof NOintwo differentbiological assays:inhibitionof humanplateletaggregation induced with IO” M adenosine diphosphate and inhibition
of humanPMNactivationinducedwith lo” M fMLP (Fig. 3). Half-maximal inhibitionin thesetwo different assays was obtained at approximate NO dilutions of I:25 and 1:lO, respectively.
0.0 1 No
Figure3. Comparison of the Inhibitoryeffect of variousconcentrations of NO In different Mobgicalassays.Dilutions of a saturated NOsolution from1:lOOO (0.001)to 190 (0.1) weretestedfor inhibitory effectsonadenosine diphosphate (10’ M) -induced aggregation of humanplatelets (O), orfMLP-induced actkationof humanPM’s (A). Thedataforthe inhibition of ADP-induced aggregation of human platelets andfortheinhibition of activation of human PM’s arerepresentative experiments whereeachpointwasobtained intriplicate andquadruplicate, respectively. Eachexperiment wasrepeated at least3 times. 1395
181, No. 3, 1991
AND BIOPHYSICAL RESEARCH COMMUNICATIONS
DISCUSSION The purpose of the present studies was to evaluate the possibility that NO (EDRF) has a cytoprotective role as scavenger of the cytotoxic oxygen-derived free radical .O,-. A biological assay was designed to test this hypothesis using human PMN’s which are known to generate .O; when activated. The generation of .O; by activated PMN’s was spectrophotometrically determined as the rate of reduction of cytochrome c. Complete inhibition of cytochrome c reduction in the presence of SOD confirmed that the spectrophotometric change observed in the presence of activated PMN’s wasdue to the generation of .02-. The study showsthat NO attenuates the reduction of cytochrome c in the presence of fMtP-activated PMN’s. The data imply that this is the result of NO-induced inactivation of .02-, but does not rule out other mechanisms including inhibition of the production of .O; by PMNs. Although this study has not directly ruled out interference of NO with PMN function as a mechanism of action, the ability of NO to suppress the reduction of cytochrome c in the presence of an enzyme system that generates .O; (xanthine oxidase plus hypoxanthine) indicates that the proposed scavenging pathway is chemically feasible. The fact that the concentration-dependent inhibition of the initial rate of reduction of cytochrome c is the same whether superoxide is generated from the more complex PMN-system or from the xanthine oxidasesystem strengthens the argument that the most likely explanation for the observed inhibition of cytochrome c reduction is inactivation of .O,- by NO. The dilutions of the saturated NO solution necessary to achieve suppression of cytochrome c reduction were lowerthan reported earlier on other biological preparations (5-7). Although no attempts were made to measure it, this may be due to the fact that under the conditions used in this study, most of the NO wasconverted to nitrite anion. However, sodium nitrite had no effecton reduction of cytochrome c induced by fMLP-stimulated PMNs at concentrationsof up to 0.5 mM (unpublished observation). Another explanation for the “low” sensitivity is that detection of inactivation of .O; by the technique used requires more NO than other biological actions of the radical. This question was addressed by testing the platelet aggregation inhibitory actions of the same NO solution that was used in the PMN studies. Half-maximal inhibition of platelet aggregation occurred at similar dilutions as these required for suppression of cytochrome c reduction implying that NO(EDRF) may exert its cytoprotective action (i.e. inactivation of ,02’) at concentrations which inhibit platelet aggregation. The present findings suggest that one of the important biological functions of NO(EDRF) may be the inactivation of cytotoxic oxygen free radicals (.O;) from oxygenated blood or activated blood cells (e.g., PMN’s). The recent demonstration that NO (generated from acidified nitrite) reduces infarct size and myeloperoxidase activity in cat myocardium during ischemia-reperfusion supports this hypothesis (14). Several earlier observations suggest a hypothesis that the EDRF(N0) system may have developed in response to exposure of cells and organisms to high oxygen environment to serve as a natural protective mechanism against oxygen stress. EDRF(N0) production by the vascular endothelium is suppressed in the absence of oxygen (1). EDRF(N0) production is more pronounced in systemic arteries that contain oxygenated blood than in veins (1516). Chronic exposure of veins to arteriil blood facilitates the production 1396
Vol. 181, No. 3, 1991
of EDRF (17). EDRF production by the vascular endothelium appears during the phylogenesis in species who moved from water (lowoxygen environment) to air (high oxygen environment) (18). Although it remains to be determined whether endotheliumderived NO behaves like authentic NO, the present observations suggest that an important biological function of EDRF(N0) is to act as a chemical barrier to protect the endothelium (and other cells) from injuries caused by superoxide and perhaps other oxygen free radicals. In contrast to this hypothesis, lgnarro et al, reported that inhibition of NO synthetase, the enzyme responsible for the cellular production of NO, has a protective effect on the myocardium during reperfusion following &hernia (personal communication). Additionally, the interaction of NO with .Oz- has been shownto produce the stable peroxynitrite anion (ONO,-) and recent data by Beckman et al (19) suggest that decomposition of the protonated species produces a strong oxidant with reactivity similar to hydroxyl radical. However, the extremely short biological half-life of the hydroxyl radical implies that if generation occurs at a site distant from biologically important structures,such as membranes or DNA, the active species may never reach them. Controversyas to whether EDRF is NO or whetherthe savengening of super oxide radical by NO is beneficial suggests that more research is needed to elucidate the physiological relevance of the various chemical intermediates generated in these processes. ACKNOWLEDGMENTS The authors would like to thank Ms. Susan Packie and Ms. Patricia Rodgers for typing the manuscript. REFERENCES ::
3. 4. 6”. 7: 8. 9. 10. 11. :32: 14. 15. 16. 17. 18. 19.
Furchgott, R.F., and Zawadski, J.V. (1980) Nature 288, 373-376. Furchgott, R.F. (1983) Circ. Res. 53,557-573. Vanhoutte, P.M., Rubanyi, GM, Miller, V.M., and Houston, D.S. (1986) Ann. Rev. Physiol. 48,307320. Furlong, B., Henderson, A.H., Sweis, M.J., and Smith, J.A. (1987) Br. J. Pharmacol. 90, 687692. Furchgott, R.F., Khan, M.T., and Jothianandan, D. (1987) Thr0m.Re.s. 7, S5. Ignarro, L.F. Byrns, R.E., Buga, G.M., and Wood, K.S. (1987) Circ. Res. 61,866879. Palmer, R.F., Ferrige, D.A., and Moncada, S. (1987) Nature 327,542-526. Rubanyi, G.M., and Vanhoutte, P.M. (1986) Am. J. Physiol. 250, H822-H827. Gryglewski, R.J., Palmer, R.M.J., and Moncada, S. (1986) Nature 320,454-456. Rubanyi, G.M. (1988) Free Radical Biol. Med. 4,107-120. Blough, N.V., and Zafiriou, O.C. (1985) Inorg. Chem. 24,3502-35&I. Rubanyi, G.M. (1988) J. Mol. Cell. Cardiol. 2O(Suppl. V), S56. Feigl, E.O. (1988) Nature 331,490-491, Mulloy, D. E., Johnson, G., Tsao, P.S., and Lefer, A.M. (1989) Fed. Proc. 3, A1175. DeMey, JG., and Vanhoutte, P.M. (1982) Circ. Res. 51,439-447. Rubanyi, G.M., and Vanhoutte, P.M. (1988) Blood Vessels 25, 7581. Miller, V.M., Aarhus, L.L., and Vanhoutte, P.M. (1986) Am. J. Physiol. 251, H526H527. Miller, V.M., and Vanhoutte, P.M. (1986) Blood Vessels 23, 225-235. Beckman, J.S., Beckman, T.W., Chen, J., Marshall, P.A., and Freeman, B.A. (1990) Proc. Nat. Acad. Sci USA 87, 1620-1624.