Antagonism by paraquat of the relaxing effects of acetylcholine and A23187 in rat thoracic aorta

European Journal of Pharmacology Environmental Toxicology and Pharmacology Section 292 (1995) 315-320

ELSEVIER

environmentaltoxicotogy and pharmacology

Antagonism by paraquat of the relaxing effects of acetylcholine and A23187 in rat thoracic aorta Kuei-Sen Hsu 1, Shoei-Yn Lin-Shiau * Institute of Pharmacology, College of Medicine, National Taiwan Unit~ersity, No. 1, Jen-Ai Rd., Sec. 1, Taipei, Taiwan Received 28 March 1994; revised 27 October 1994; accepted 1 November 1994

Abstract

Paraquat, a widely used herbicide, has been reported to be capable of producing superoxide. In the present paper, therefore, the possibility of paraquat inhibiting endothelium-dependent relaxation, mediated by the production of nitric oxide, was tested. The relaxing effects of acetylcholine and A23187, but not that of sodium nitroprusside, in the rat thoracic aorta were found to be inhibited by paraquat in a concentration-dependent manner. In contrast, paraquat was totally inactive with regard to the aortic contractions induced by either norepinephrine or prostaglandin F2~. The inhibitory action of paraquat could be antagonized by superoxide dismutase but not by catalase and indomethacin. All of these findings indicate that superoxide anions produced by paraquat in the endothelium contribute to a decrease in the relaxation response to acetylcholine and A23187 by interfering with endothelium-derived nitric oxide.

Keywords: Paraquat; Superoxide anion; Nitric oxide (NO); Acetylcholine; A23187

I. Introduction

The vascular endothelium is widely recognized as a tissue with a variety of functions, including relaxation of vascular tone. The response of isolated blood vessels to several vasoactive agonists is modulated by the presence of endothelial cells. Indeed, these cells can release both dilator and constrictor substances (Luscher, 1990a; Vane et al., 1990). Imbalances between the release of these factors is observed in aging and diseases such as hypertension, atherosclerosis and ischaemia (Koga et al., 1989; Luscher, 1990b). The major vasodilator substances produced by endothelium are endothelium-derived relaxing factor (EDRF), identified as nitric oxide derived from L-arginine, and prostacyclin, a metabolite of the arachidonic acid pathway. Endothelin, produced by the endothelium, is one of the most potent vasoconstrictors. Paraquat is a quaternary ammonium bipyridyl compound (1,Y-dimethyl-4,4'-bipyridium dichloride), widely

* Corresponding author. Fax: 886-2-3915297. I Present address: Institute of Pharmacology, College of Medicine, National Cheng Kung University, Tainan, Taiwan. 0926-6917/95/$09.50 © 1995 Elsevier Science B.V. All rights reserved

SSDI 0 9 2 6 - 6 9 1 7 ( 9 4 ) 0 0 0 5 9 - X

used as a broad spectrum herbicide. It exerts its herbicidal activity by interfering with the electron transfer system, and thereby inhibits reduction of NADP to N A D P H during photosynthesis. Paraquat produces degenerative lesions in the lung following systemic administration to man and animals. Respiratory failure is a common cause of death in paraquat intoxication (Rose et aI., 1974). Although it has been claimed that progressive pulmonary fibrosis is the major cause of respiratory failure (Rose et al., 1974; Smith and Heath, 1976), toxic myopathy has also been reported to be a late manifestation of paraquat poisoning (Smith and Heath, 1976; Smith and Rose, 1977). Several aspects of the pulmonary toxicity of paraquat resemble the toxicity of various other lung toxins, including oxygen, nitrofurantoin and bleomycin (Bus and Gibson, 1984). Although a definitive mechanism of paraquat toxicity has not yet been delineated, it is apparent that a cyclic single electron reduction/oxidation of the parent molecule is a critical mechanistic event. The redox cycling of paraquat in biological systems has two potentially important consequences relevant to the development of toxicity: generation of a reactive oxygen species (e.g., superoxide anion, hydrogen peroxide, hydroxyl radical) that is highly toxic to the tissue; a n d / o r deple-

316

K.-S. Hsu, S.-Y. Lin-Shiau / Eur. J. Pharmacol. Enriron. Toxicol. Pharmacol. Section 292 (1995) 315-320

tion of cellular reducing equivalents (e.g., N A D P H ) necessary for normal function (Bus and Gibson, 1984; Heath and Smith, 1977; Smith and Heath, 1976; Smith and Rose, 1977). The same result has been reported for endothelial cell exposure to paraquat, which induces extracellular superoxide anion formation (Britigan et al., 1992). Furthermore, superoxide anion breaks down endothelium-derived nitric oxide and has also been reported to contract vascular smooth muscle (Auck-Schwellk et al., 1989; Gryglewski et al., 1986). Hydrogen peroxide relaxes vascular smooth muscle via the regulation of cyclic G M P (Burke and Wolin, 1987). Endothelial prostacyclin release has been reported to be stimulated (Harlan and Callahan, 1984) or depressed (Whorton et al., 1985) by oxygen-derived free radicals, depending on the concentration of radical present (Taylor et al., 1983). In addition, oxygen-derived radicals and proteolytic enzymes are cytotoxic, and both can cause damage to the endothelium (Ward and Varani, 1990). In order to investigate the effect of free radicals produced by endothelium on the regulation of vascular tone, in this paper, we use paraquat as a generator of activated oxygen (e.g., superoxide anion, hydrogen peroxide, hydroxyl radical) to determine the effect of free radicals on the endothelium-dependent relaxation of rat thoracic aorta to acetylcholine and ionophore A23187.

The aorta was equilibrated in the medium for 90 rain with three changes of Krebs' solution and maintained under an optimal tension of 1 g before specific experimental protocols were initiated. Contractions were recorded isometrically via a force-displacement transducer connected to a Grass Model 7 polygraph. The rings were contracted by addition of norepinephrine or prostaglandin F2, in the organ bath. At maximal contraction the aorta was exposed to acetylcholine or A23187 to 'challenge' the tissue and to test the functional integrity of endothelium. Some experiments were carried out in the presence of cyclooxygenase inhibitor indomethacin or the oxygen-radical scavengers superoxide dismutase (SOD) or catalase, which respectively scavenges superoxide anion or hydrogen peroxide. These compounds were added to the organ baths 10 rain prior to the addition of paraquat. 2.3. Statistical analysis Results are given as means _+ S.E.M. The number of experiments are indicated by n. The significance of difference was evaluated by Student's t test. When more than one group was compared with one control, significance was evaluated according to ANOVA. Probability values ( P ) of less than 0.05 were considered to be significant. A

2. Materials and methods

i~e--o-

1o o ,

i

2.1. Materials

•

C 8o ~

Paraquat (1,1'-dimethyl-4,4'-bipyridylium dichloride), norepinephrine, prostaglandin F2~, acetylcholine, A23187, superoxide dismutase (SOD), catalase, indomethacin and sodium nitroprusside were purchased from Sigma Chemical Company, St. Louis, MO, USA. These compounds were dissolved in water. Prostaglandin F2~ (PGF2~) at a stock concentration of 1 mM in ethanol was stored at -20°C.

•

,~S/' :

~

60-

. //

J

Z 40j o

i / 8

9

. ,////

<7

i

/c

6

-7

-5

4

3

Norepinephrine (log [M])

tO0 1

i 80-

2.2. Mechanical response Wistar rats of either sex weighing 250-300 g were killed by a blow to the head. The thoracic aorta was isolated and excessive fat and connective tissue were removed. The vessels were cut into rings about 5 mm in length and mounted in organ baths containing 10 ml of Krebs' solution of the following composition (mM): NaCI 118.2, KCI 4.7, CaC12 1.9, MgSO 4 1.2, K H 2 P O 4 1.2, N a H C O 3 25 and glucose 11.7. The tissue bath solution was maintained at 37.0_+ 0.5°C and oxygenated with 95% O 2 + 5% CO 2. Two stainless steel hooks were inserted into the aortic lumen; one was fixed while the other was connected to a transducer.

~°~

-8

-7

6

-5

Prostaglandin F2~ (log [M]) Fig. 1. Effects of paraquat on the tonic contraction caused by norepinephrine and prostaglandin F2c, in the rat thoracic aorta. Paraquat (e, 100 ~zM) was incubated with aorta for 20 rain, then cumulative concentrations of either norepinephrine (A) or prostaglandin F2~ (B) were used to trigger the contraction. Control (O). Data are presented as means _+S.E.M. (n = 5).

K.-S. Hsu, S.-Y. Lin-Shiau / Eur. Z Pharmacol. Enciron. Toxicol. Pharmacol. Section 292 (1995) 315-320

3. Results

3.1. Effect of paraquat on norepinephrine- and prostaglandin F2,-induced tonic contraction In the rat thoracic aorta, norepinephrine and prostaglandin F2a caused a tonic contraction which

lasted for at least 30 min. A relaxation (more than 90%) induced by acetylcholine (1 izM) indicated that the endothelium was intact in this preparation. Pretreatment with paraquat (100 tzM) for 20 min had no effect on either norepinephrine (10-8-3 × 10 -4 M)- or prostaglandin F2,~ (10 8-10-5 M)-induced tonic contraction in rat thoracic aorta (Fig. 1).

3.2. Effect of paraquat on arterial relaxation Addition of paraquat (100 ~M) to rat thoracic aorta rings had no significant effect on the basal tension. The relaxation response of rat aorta to acetylcholine (10 - 9 10 -4 M), was, however, found to be impaired. The maximal relaxations were decreased from 96.8 + 2.2% to 60.3 _+4.7% and 97.2 + 1.9% to 62.7 + 2.9% of the

precontraction either by norepinephrine (3 /zM) (P < 0.001) or prostaglandin F z a (1 / z M ) ( P < 0 . 0 0 1 ) (Figs. 2A and 2B). Similar results were also found for ionophore A23187 ( 1 0 - 9 - 3 × 10 -5 M)-induced relaxation of rat aorta. The maximal relaxations were decreased from 99.8 ___0.7% to 63.8 _+3.9% and 98.6 + 0.7% to 61.2_+ 4.8% of the precontraction either by norepinephrine (3 /zM) (P < 0.001) or prostaglandin F 2 , (1 / z M ) ( P < 0 . 0 0 1 ) (Figs. 2C and 2D). In addition, the inhibitory effects of paraquat on the relaxation of rat aorta to either acetylcholine or A23187 were reversible after washout of paraquat. Paraquat (1-3000 IzM) inhibited the relaxation response of rat aorta to acetylcholine (3/~M) and A23187 (3 IzM) in a concentration-dependent manner. The ICs0 was calculated to be about 168.8 and 124.7 IzM, respectively, of the precontraction with norepinephrine (3 /zM) and 118.4 and 121.3 p~M, respectively, of the precontraction with prostaglandin F2, (1 /~M) (Fig. 3). If sodium nitroprusside was used to relax the rat thoracic aorta independently from the endothelium, the responsiveness was not altered significantly by addition of paraquat (100 p.M) to the organ bath (Fig. 4).

C

A

1O0 -

T~ 0 - - 0 - - 0 - - 0

100 ©

60

./

40-

c~

"X"

0 / l

o3

r/0--0--0--0

80-

r/~ o

317

~

x

~ T I l l

~ T e / i

T

l i 1

/'

1 ~t

o

60-

5/°

X-

0

T

X-

II "----~ ~i

.

r

1

O3

;~/<

40-

6t .

20-

20

&

I

oi t

I

-9

I

#

"l

t

I

I

I

d

-fl

-5

-4

-9

-8

-7

-6

-5

I

-7 Acetylcholine

-8

13

(log

[Mj)

~

,/,

8o

v 0

.

.

*

Z

eo

8/?

60

rio

. r~ /Y

S

O3 O3

A 2 3 1 8 7 (log [M])

D 100

I00 -

40

/

cJ

-/

2O

~/l

*l

-8

-X* T T~ , ~I-- i i

.

'< •

Ti 0 ~

i/~

01

#

I

o-/.

. at.

z0/:<:~i;/'

~/

/.0_ I/ ---di__l-/-

-9

40.

I

-4

I

-7

Acetylcholine

I

-6

( l o g [M])

I

-5

I

-4

I

-9

I

-B

--

i

I

-7

-6

--t

I

-5

-4

A 2 3 1 8 7 (log [M])

Fig. 2. Effects of paraquat ( 1 0 0 / i M ) on the relaxation response to acetylcholine and ionophore A23187 in rat thoracic aorta. Concentration-effect curves for acetylcholine ( A , B) and A23187 (C, D) in arterial rings, after precontraction with either norepinephrine (3 ~ M ) ( A , C) or prostaglandin F2,~ (1 ~ M ) (B, D), were obtained in the absence ( © ) or presence of paraquat (e, 100 /~M). The results are expressed as a percentage of the precontraction. Data are presented as m e a n s + S.E.M. (n = 6). * : P < 0.05 as compared with the control.

318

I~ - S. Hsu, S. - Y. Lin- Shiau / Eur. J. Pharmacol. Environ. Toxicol. Pharmacol. Section 292 (1995) 3 1 5 - 3 2 0

A

,oO. o.

A

iS"

i0

I00

1olD0 BOO0

B

lOOT

×

•

n

O

~d

B I00

L

7

i' i

i

iI

.oi zo f

1

@4

fill

lo

ioo

I oo

sooo

P a r a q u a t (/.zM)

:c:¢.r

5uperoxme

2at~i~

:rv:or:l,~: :a :::

=:~mutas,

Fig. 3. Concentration-dependent inhibition of acetylcholine- and A23187-induced relaxation in rat thoracic aorta after precontraction with norepinephrine ( A ) or prostaglandin F2, (B) induced by paraquat. Various concentrations of paraquat were applied to rat aorta for 20 rain. The maximal relaxation responses to either acetylcholine ( O , 3 /xM) or A23187 (o, 3 p~M) were significantly diminished by the paraquat in a concentration-dependent manner. T h e results are expressed as a percentage of the precontraction. Data are presented as m e a n s + S.E.M. (n = 5).

3.3. Effect of oxygen-radical domethacin

scavengers

and in-

To elucidate the mechanism of action of paraquat on the acetylcholine- and A23187-induced relaxation, the experiments were repeated in the presence of superoxide dismutase, catalase and indomethacin (Fig. 5). Treatment of the arterial rings with indomethacin

fo--o

~oo~

Fig. 5. Influence of treatment with indomethacin or oxygen radical scavengers on the paraquat-induced decrease in maximal relaxation response to acetylcholine and A23187 in rat thoracic aorta. Maximal relaxation response to acetylcholine ( A ) or A23187 (B) in the absence (open columns) and presence of paraquat (100 txM) after a 20 rain incubation (closed column) is expressed as the percentage of the precontraction with norepinephrine (3/xM). Superoxide dismutase (100 U / m l ) , catalase (1000 U / m l ) or indomethacin (10 IxM) were added to the organ baths 10 min prior to the application of paraquat. Data are presented as m e a n s + S.E.M. (n = 5). *: P < 0.05 as compared with the control.

and paraquat resulted in a further significant reduction of the maximal relaxation caused by either acetylcholine or A23187, while neither superoxide dismutase (I00 U / m l ) nor catalase (1000 U / m l ) had an effect on the maximal relaxation response to acetylcholine and A23187. Superoxide dismutase, simultaneously incubated with paraquat prevented the decrease in relaxation. However, in the presence of catalase, paraquat still diminished the responsiveness of the arterial rings (Fig. 5).

4. Discussion

-8

-7

-6

-5

N i t r o p r u s s i d e (log [M]) Fig. 4. Concentration-effect curve of nitroprusside in rat thoracic aorta in the absence ( O ) and presence (e) of paraquat. The relaxation is expressed as a percentage of precontraction with norepinephrine (3 /xM). Data are presented as m e a n s _+S.E.M. (n = 4).

The present study demonstrated that the herbicide, paraquat, when added to organ baths containing segments of rat thoracic aorta, decreased the relaxation response to acetylcholine and A23187. The diminished response is restored by treatment with superoxide dismutase, which is indicative of involvement of oxygenderived free radicals. The importance of the endothelium is stressed by the observation that the relaxation to sodium nitroprusside is not affected by the addition of paraquat.

K.-S. Hsu, S.-Y. Lin-Shiau /Eur. J. Pharmacol. Environ. Toxicol. Pharmacol. Section 292 (1995) 315-320

Acetylcholine and ionophore A23187 are capable of evoking the release of vasodilator substances from the endothelium (Vane et al., 1990). Two potent dilators released from the endothelium are nitric oxide and prostacyclin (Ignarro et al., 1981). The relaxation responses to acetylcholine and A23187 are completely dependent on an intact endothelium; they are partially inhibited by the nitric oxide synthesis inhibitor, N Gmonomethyl-L-arginine, or the cyclo-oxygenase inhibitor, indomethacin. Simultaneous use of these inhibitors blocks the relaxation totally. Therefore, it is likely that the endothelium of rat thoracic aorta releases two substances, nitric oxide and prostacyclin, which mediate relaxation (Brown et al., 1992). In contrast to acetylcholine and A23187, sodium nitroprusside caused relaxation by releasing nitric oxide at or in the vascular smooth muscle cells. The generated nitric oxide activates soluble guanylate cyclase, catalyzing cyclic GMP accumulation and relaxation (Ignarro, 1989). The mechanism of nitric oxide liberation from sodium nitroprusside remains undefined, but is known to differ from that of the organic nitrovasodilators (e.g., S-nitroso-N-acetyl-penicillamine (SNAP) and 3morpholino-sydnonimine (SIN-l)) (Gryglewski et al., 1992). It is often assumed that spontaneous chemical degradation of sodium nitroprusside to nitric oxide mediates sodium nitroprusside vasorelaxation (Ignarro et al., 1981). An alternative hypothesis is that nitric oxide generation results from a metabolic interaction of sodium nitroprusside with the vascular smooth muscle cells (Kowaluk et al., 1992). Furthermore, Gruetter et al. (1979) showed that relaxation by sodium nitroprusside is not inhibited by impermeable heme proteins which bind nitric oxide, and suggested that sodium nitroprusside releases nitric oxide intracellularly. In this way, we used sodium nitroprusside as a generater of nitric oxide in the vascular smooth muscle, mimicking the vasodilator action of endothelium-derived nitric oxide. In the present study, we found that paraquat decreased relaxation response to either acetylcholine or A23187, but not to sodium nitroprusside. It is concluded that the effect of paraquat is not related directly to the action of nitric oxide in vascular smooth muscle ceils. The second instillation of extracellular interaction between nitric oxide and paraquat-induced superoxide anion was supported by the finding that the decrease in relaxation response to acetylcholine and A23187 caused by paraquat is completely restored by superoxide dismutase, which can hardly enter endothelial cells (Ignarro et al., 1981). Therefore, it is likely that paraquat induces endothelium-dependent extracellular superoxide anion formation, which directly interferes with endothelium-derived nitric oxide in the extracellular space of vascular smooth muscle cells. The fact that paraquat can induce cytotoxic oxygeny free radical production in endothelial ceils has been

319

reported (Britigan et al., 1992). These reactive metabolites are associated with cell injury in various pathological states, such as acute hypertension, ischaemia and shock (Wang et al., 1990; Wei et al., 1985). Damage to endothelial cells may lead to an altered responsiveness of the vasculature. However, in the present study, the response to either the acetylcholine or A23187 is only influenced in the presence of paraquat, and not when the paraquat is washed out of the organ bath. This finding indicates that no irreversible functional damage has occurred. Some evidence has been obtained that oxygen radicals may influence the endothelium-derived vasodilators, nitric oxide (Gryglewski et al., 1986) and prostacyclin (Taylor et al., 1983; Whorton et al., 1985). The paraquat-dependent decrease in responsiveness to acetylcholine and A23187 is similar, both in the presence and absence of the cyclo-oxygenase inhibitor, indomethacin. This result indicates that paraquat affects the nitric oxide pathway, because the prostacyclin release is inhibited by indomethacin. One of the pharmacological criteria for identifying endothelium-derived relaxing factor (EDRF) as nitric oxide, is that its action is augmented by superoxide dismutase (SOD) in a superfusion cascade. Nitric oxide is a labile substance and its stability is decreased by superoxide anion, but not by other oxygen-derived reactive metabolites, such as hydrogen peroxide (Gryglewski et al., 1986). Indeed, we have found that the decrease in relaxation response to acetylcholine and A23187 caused by paraquat is completely restored by superoxide dismutase, but not by the hydrogen peroxide scavenger catalase. In summary, the presence of paraquat in the organ bath causes a superoxide dismutase-inhibitable suppression of the acetylcholine-induced relaxation of rat thoracic aorta. It is concluded that superoxide anions, produced by paraquat in the endothelium, impair endothelium-dependent relaxation by direct interference with endothelium-derived nitric oxide. Paraquat-induced endothelial toxicity is, therefore, a potentially useful model for evalution of oxidant mechanisms of toxicity.

Acknowledgements This work was supported by a research grant from the National Science Council, Republic of China (NSC 82-0412-B002-093).

References Auck-Schwellk, W., Z.S. Katusic and P.M. Vanhoutte, 1989, Contractions to oxygen-derivedfree radicals are augmentedin aorta of the spontaneouslyhypertensiverats, Hypertension13, 859.

320

K.-S. Hsu, S.-Y. Lin-Shiau / Eur. J. Pharmacol. Emqron. Toxicol. Pharmacol. Section 292 (1995) 315-320

Britigan, B.E., T.L. Roeder and D. Michael-Shasby, 1992, Insight into the nature and site of oxygen-centered free radical generation by endothelial cell monolayers using a novel spin trapping technique, Blood 79, 699. Brown, M.L., C.A. Clark, R. Vaillancourt and D. Deykin, 1992, Elevated glucose alters A23187-induced release of arachidonic acid from porcine aortic endothelial cells by enhancing reacylation, Biochem. Biophys. Acta 1165, 239. Burke, T.M. and M.S. Wolin, 1987, Hydrogen peroxide elicits pulmonary arterial relaxation and guanylate cyclase activation, Am. J. Physiol. 252, H721. Bus, J.S. and J.E. Gibson, 1984, Paraquat: model for oxidant-initiated toxicity, Environm. Health Perspect. 55, 37. Gruetter, C.A., B.K. Barry, D.B. Mcnamara, D.Y. Gruetter, P.J. Kadowitz and L.J. Ignarro, 1979, Relaxation of bovine coronary artery and activation of coronary arterial guanylate cyclase by nitric oxide, nitroprusside and a carcinogenic nitrosamine, J. Cyclic Nucleotide Res. 5, 211. Gryglewski, R.J., R.M.J. Palmer and S. Moncada, 1986, Superoxide anion is involved in the breakdown of endothelium-derived vascular relaxing factor, Nature 320, 454. Gryglewski, R.J., A. Zembowicz, D. Salvemini, G.W. Taylor and J.R. Vane, 1992, Modulation of the pharmacological actions of nitrovasodilators by methylene blue and pyocyanin, Br. J. Pharmacol. 106, 838. Harlan, J.M. and K.S. Callahan, 1984, Role of hydrogen peroxide in the neutrophil-mediated release of prostacyclin from cultured endothelial cells, J. Clin. Invest. 74, 442. Heath, D. and P. Smith, 1977, The pathology of the lung in paraquat poisoning, in: Biochemical Mechanisms of Paraquat Toxicology, ed. A.P. Autor (Academic Press, New York) p. 39. Ignarro, L.J., 1989, Biological actions and properties of endothelium derived nitric oxide formed and released from artery and vein, Circ. Res. 65, 1. Ignarro, L.J., H. Lippton, J.C. Edwards, W.H. Baricos, A.L. Hyman, P.J. Kadowitz and C.A. Gruetter, 1981, Mechanism of vascular smooth muscle relaxation by organic nitrates, nitrites, nitroprusside and nitric oxide: Evidence for the involvement of S-nitrosothiols as active intermediates, J. Pharmacol. Exp. Ther. 218, 739.

Koga, T., Y. Takata, K. Kobayashi, S. Takishita, Y. Yama-Shita and M. Fujishima, 1989, Age and hypertension promote endothelium-dependent contractions to acetylcholine in the aorta of the rat, Hypertension 14, 542. Kowaluk, E.A., P. Seth and H.L. Fung, 1992, Metabolic activation of sodium nitroprusside to nitric oxide in vascular smooth muscle, J. Pharmacol. Exp. Ther. 262, 916. Luscher, T.F. (ed.), 1990a, Endothelial vasoactive substances and cardiovascular diseases (Karger, Basel). Luscher, T.F., 1990b, The endothelium: target and promoter of hypertension ? Hypertension 15, 482. Rose, M.S., E.L. Smith and I. Wyatt, 1974, Evidence for energy-dependent accumulation of paraquat into rat lung. Nature 252, 314. Smith, P. and D. Heath, 1976, Paraquat CRC Crit. Rev Toxicol 4, 411. Smith, L.L. and M.S. Rose, 1977, Biochemical changes in lungs exposed to paraquat: In: Biochemical Mechanisms of Paraquat Toxicology (A.P. Autor Ed.), Academic Press, New York, pp. 187. Taylor, L., M.J. Menconi and P. Polgar, 1983. The participation of hydrogen peroxides and oxygen radicals in the control of prostaglandin synthesis. J. Biol. Chem. 258, 6855. Vane, J.R., E.E. Anggard and R.M. Botting, 1990, Regulatory functions of the vascular endothelium. New Engl. J. Med. 323, 27. Wang, J., H. Chen, T. Wang, Y. Diao and K. Tian, 1990, Oxygen-derived free radicals induced cellular injury in superior mesenteric artery occlusion shock: protective effect of superoxide dismutase. Circ. Shock 32, 31. Ward, P.A. and J. Varani, 1990, Mechanisms of neutrophil-mediated killing of endothelial cells. J. Leukoc. Biol. 48, 97. Wei, E.P., H.A. Kontos, C.W. Christman, D.S. Dewitt and J.T. Povlishock, 1985 Superoxide generation and reversal of acetylcholine-induced cerebral arteriolar dilation after acute hypertension. Circ. Res. 57, 781. Whorton, A.R., M.E. Montgomery and R.S. Kent, 1985, Effect of hydrogen peroxide on prostaglandin production cellular integrity in cultured porcine aortic endothelial cells. J. Clin. Invest. 76, 295.