Exogenous nitric oxide prevents endotoxin-induced glomerular thrombosis in rats

Kidney International, Vol. 46 (1994), pp. 711—716

Exogenous nitric oxide prevents endotoxin-induced glomerular thrombosis in rats GUNNAR WESTBERG, PAMELA J. SHULTZ, and LEOPOLDO RAIJ Department of Medicine, Veterans Administration Medical Center, and the University of Minnesota School of Medicine, Minneapolis, Minnesota, USA

Exogenous nitric oxide prevents endotoxin-induced glomerular throm-

bosis in rats. Nitric oxide (NO) synthesized from L-arginine is an endogenous vasodilator and inhibitor of platelet adhesion and aggregation. Gram-negative lipopolysaccharide (LPS) can induce NO synthesis,

which may mediate the pathophysiologic effects of endotoxemia. In addition, our previous studies suggested that LPS-induced NO may protect against thrombosis in rats. In the present study, male SpragueDawley rats given LPS (0.1 mg/kg) i.p. increased their urinary excretion of

NO2 + NO3 (stable end-products of NO) by 4.3-fold. Rats given 10 rg/kg/hr i.v. of nitroglycerin (GTN), an exogenous NO donor, showed a similar increase. L-NAME, an inhibitor of NO synthesis, abrogated the increase in urinary NO2 + NO3 in LPS-treated rats but not in rats given GTN. Glomerular thrombosis developed in rats given LPS + L-NAME (thrombosis score = 3.02 0.4), while those given LPS + L-NAME + GTN were largely protected (thrombosis score = 1.37 0.5, P < 0.05). Atrial natriuretic peptide (ANP), an NO-independent vasodilator, neither increased urinary NO2 + NO3 nor prevented glomerular thrombosis (thrombosis score = 2.68 0.5, NS). Hydralazine, another vasodilator without effects on NO or platelets, also failed to prevent glomerular thrombosis in rats given LPS + L-NAME. We conclude that in endotoxemia, the antithrombogenic properties of endogenously synthesized NO are important in preventing alomerular thrombosis. The exogenously NO

in the urinary excretion rate of NO2 + NO3, the stable end products of NO, and do not develop glomerular thrombosis. However, if LPS-induced synthesis of NO is prevented by prior or

concomitant treatment with an L-arginine analog, such as NG nitro-L-arginine methyl ester ( L-NAME), generalized glomerular thrombosis occurs [7]. In human septic shock, severe vasodilation, hypotension, vascular damage and disseminated intravascular coagulation can be observed. The persistent hypotension and low systemic vascular resistance are usually resistant to vasoconstrictors such as norepi-

nephrine. Even in critical care units, the mortality rate of these patients remains high [8]. Recent reports have demonstrated that inhibition of NOS will quickly restore blood pressure to normal in both animals and humans with endotoxemia [9—11]. However, in rats and in dogs with endotoxemia the administration of an NOS inhibitor resulted in increased mortality [7, 12]. In a recent study,

Wright, Rees and Moncada [13] reported that in rabbits that

donor, GTN, can substitute for the antithrombogenic effect of endogenous NO. Clinically, administration of NO synthesis inhibitors to treat endotoxic shock may need to be combined with concomitant administration of exogenous NO donors to prevent microvascular thrombosis.

received LPS, blocking of NOS with an arginine analog, results in irreversible shock and increased mortality. The administration of an NO donor, S-nitroso-N-acetyl-penicillamine, improved perfusion, prevented shock and led to a decreased mortality. Unfortunately, intravascular coagulation or thrombosis was not investigated in that study.

Endotoxin, a lipopolysacharide (LPS) from gram negative

effects such as glomerular thrombosis will have to be prevented.

If NOS inhibitors are to be useful in septic shock, adverse We therefore undertook the present study to explore whether induce endothelial injuly and intravascular coagulation, often intravenous administration of an exogenous NO donor, nitroglycleading to glomerular thrombosis [1—3]. LPS and/or cytokines, erin (glyceryl trinitrate, GTN), could substitute for endogenous released during endotoxemia, increase expression of the inducible NO and prevent glomerular thrombosis induced by LPS in nitric oxide synthase (NOS) [4, 5], leading to synthesis of large L-NAME-treated rats, To evaluate whether the vasodilation amounts of nitric oxide (NO) from its precursor L-arginine. NO induced by NO was important in the prevention of glomerular induces vascular relaxation and inhibits platelet aggregation via thrombosis, we compared the effect of GTN with that of two other increased synthesis of cGMP due to activation of soluble guany- vasodilators, atrial natriuretic peptide (ANP) and hydralazine. late cyclase [6]. Thus, NO may play an important role in the ANP has been shown to be effective in preventing acute renal hypotension observed during endotoxemia or sepsis [3]. On the failure after ischemia [14, 15], and is known to increase cGMP via other hand, in the kidney increased NO formation may be particulate guanylate cyclase. Hydralazine, on the other hand, important in maintaining renal perfusion during shock and pre- causes direct vascular smooth muscle relaxation, probably via bacteria (LPS), and certain cytokines, have been demonstrated to

venting platelet aggregation and thrombosis in glomeruli [7]. We have previously shown that rats given LPS have a marked increase

effects on calcium.

Methods Experimental design

Received for publication September 20, 1993 and in revised form March 28, 1994 Accepted for publication March 28, 1994

© 1994 by the International Society of Nephrology

The experimental design used in these studies is described in

Figure 1. Male Sprague-Dawley rats (275 to 350 g, Harlan Laboratories, Indianapolis, Illinois, USA) were used in all studies. Rats were fasted for six hours and then placed in metabolic cages

711

Westbeig et a!: Nitric oxide and glomerular thrombosis

712

=

Anesthesia Intravenous infusion, GTN,ANP or saline LPS

L-NAME4

Sacrifice 1

1

23 4567

16-hr baseline —1 0 1 6 hr urine fasting Urine collection, hours collection

Fig. 1. Schematic representation of the experimental design, including the timing of i.p. injections (arrows) and intravenous infusions (hatched bar).

for overnight (16 hr) urine collections to determine baseline Group 5, 3 rats), a PE-50 catheter was placed in the femoral artery urinary NO2 + NO3 excretion. Water was provided ad libitum, for continuous blood pressure monitoring (TNF transducer, and rats were fasted for the duration of the experiment. The rats were then briefly anesthetized (Brevital 100 mg/kg i.p.) and a PE-50 catheter was placed in the femoral vein, tunneled subcuta-

Gould, Oxnard, California, USA). As small doses of heparin were required to maintain line patency in these rats, kidney tissue from these rats was not used for the morphologic studies.

neously and extruded in the posterior cervical region for the

Seven hours (8 hr for group 5) after the injection of LPS, rats

constant infusion of GTN, ANP, hydralazine or saline vehicle. No were anesthetized with mactin (5-butyl-5-ethyl-thiobarbituric heparin was administered. A baseline blood sample was obtained acid, 100 mg/kg i.p.). The left kidney was removed through a from the femoral vein. Once awake, the rats were kept in midline incision and processed for morphologic analysis. A final individual metabolic cages for the seven hours of the experiment blood sample was collected from the aorta and the animals killed by exsanguination. (Fig. 1). The following five experimental groups were studied: Group 1. LPS + L-N4JvIE (N = 19). These rats were given LPS Moi'phologic analysis (0.1 mg/kg i.p.) at time 0 and L-NAME (75 mg/kg i.p.) at —30 Kidney tissue was fixed in buffered formalin and stained with minutes and 0, 2, 4 and 6 hours after the LPS injection. In five of

these rats, a catheter was introduced in the femoral vein as

the Periodic acid Schiff (PAS) technique. The stained sections described above. These rats received normal saline intravenously were coded and read without knowledge of which group the tissue (1.2 ml/hr) beginning one hour before the LPS injection. These came from. The extent of the deposition of fibrin-like PAS

rats were used as controls to evaluate the effects of volume positive material in the capillaries was evaluated. We have substitution and instrumentation as given in rats in the other previously shown by immunofluorescence microscopy that these groups. Group 2. LPS + L-NAME + GTN (N = 15). L-NAME and LPS

deposits do in fact contain fibrin [7]. The percentage of capillaries in each glomeruli which contained fibrin thrombosis was the basis for a semi-quantitative score: Score 0: No thrombosis seen; Score

were given as in Group 1. In addition, GTN was dissolved in sterile saline in a glass vial, aspirated into a glass syringe and 1: some, but not more than 10%, of the glomerular capillaries infused via the femoral vein catheter at a rate of 1.2 mI/hr. GTN showed thrombosis; Score 2: 10 to 25%; Score 3: 26 to 50%; Score infusion was started at —60 minutes, at 2 g/kg/min and increased 4: 51 to 100%. Fifty glomeruli were evaluated from each rat and stepwise to 10 jxg/kg/min over the next hour. This dose was the mean score for each group of rats was calculated. continued for the rest of the experiment. Group 3, LPS + L-NAME + ANP (N = 12). LPS and L-NAME were given as in Group 1. In addition, ANP, dissolved in normal saline, was infused at a rate of 1.2 ml/hr, beginning at —60 minutes

Biochemical measurements

Urinary excretion rate of NO2 + NO3 was determined as previously described [7]. Urinary excretion of cyclic 3', 5'at 2 jxg/kg/hr with a stepwise increase to a maximum of 16 guanosine monophosphate (cGMP) was determined by either pg/kg/hr [14], and kept at this rate for the duration of the enzyme-linked immunoassay or radioimmunoassay (Amersham, UK) [7, 14]. Both assays yielded similar results. experiment. The following chemicals were used: N°-nitro-L-arginine methyl Group 4. GTN only (N 4). LPS and L-NAME were not given in this group. These rats were given only an infusion of GTN in ester (L-NAME; Sigma Chemicals, St. Louis, Missouri, USA);

normal saline in a concentration and at a rate identical to that lipopolysaccharide of E. col4 serotype 0127 (Difco, Detroit, given in Group 2. Michigan, USA); nitroglycerin (glyceryl trinitrate, GTN; NitroGroup 5. LPS only (N 15). These rats were injected i.p at time stat, Parke-Davis, Morris Plains, New Jersey). Atrial natriuretic 0 with LPS (0.1 mg/kg). No other injections or infusions were peptide, human (ANP; Atriopeptin III, Sigma Chemicals); Brevital (Methohexital sodium, Eli Lilly Comp., Indianapolis, Indiana, An additional group of rats (N = 7) was given LPS + L-NAME USA); hydralazine HCI (Sigma Chemicals). and an intravenous infusion of a different vasodilator, hydralazine. Statistical analysis The protocol for these experiments was exactly as for the groups given.

given LPS + L-NAME + GTN or ANP. Based on previous

Paired, two-tailed Student's t-test was used for comparison

studies [16], we gave hydrazaline at a dose of 0.4 mg/kg/hr in a between the baseline and experimental values. The blood pressure volume of 1.2 ml/hr, which resulted in blood pressures comparable recordings and intergroup comparisons were analyzed by analysis of variance (ANOVA) and post-hoc Scheffe's test, with differto that in the LPS + L-NAME + GTN group. To measure intrarterial blood pressure, subgroups of rats in ences considered significant for P < 0.05. Values are presented as each group (Group 1, 5 rats; Group 2, 4 rats; Group 3, 5 rats, mean standard error of the mean (sEM).

713

Westbe,g et a!: Nitric oxide and glomerular thrombosis

Table 2. Urinary excretion of cGMP

E 180 E

Group Baseline

160

1

140

2 3 4

120

5

t

100 80

a)

—1

0

1

2

3

4

5

6

7

Experimental protocol — LPS + L-NAME >N &ANP GTN only LPS only

N 17

12 8 10 4 6

UV cGMP nmol/hr/100 g body wt 0.99 1.00 1.13

0.10 0.13 0.27

2.63 0.59

0.89 0.043 1.71

0.28a

Urine was collected for 7 hr, except in Group 3, where most of the rats died before the end of the experiment and the collection time was on the average 5 hr. a P < 0.05 vs. all others

Time, hours Fig. 2. Mean intraarterial blood pressure during the 7 hours of the experiment in rats given LPS + L-NAME (•, N = 5), LPS + L-NAME + GTN (A, N = 4), LPS + L-NAME + ANP (, N = 5 at all time points, except N = 2 at 7 hr). Blood pressures were significantly lower in the LPS only group compared to all other groups at 1, 3, 5 and 7 hr (P < 0.05). Data are

shown as mean SEM for each point. Table 1. Urinary excretion of NO2 + NO3

Group Baseline 1

2 3 4 5

Experimental protocol — LPS + L-NAME >N &ANP GTN only LPS only

N 17 19 11 10 4 7

Urine flow rated mi/hr 0.6 1.4

0.09

o.la

2.7 016ab 1.4 0.18 0.7 1.0

0.09 0.13a

UV NO2 + NO3 nmol/hr/100 g body wt

75 6.4

49 8.2a

201 225ab 48

9.2a

248 251ab 321

36.5

rate. Urinary NO2 + NO3 excretion rates (UVNO2 + No3) are also shown in Table 1. Injection of LPS resulted in a 4.3-fold increase in the urinary excretion rate of these endproducts of NO metabolism, compared to baseline. A similar increase in urinary excretion of NO2 + NO3 required a dose of GTN of 10 1.gJkg/min (Group 4, GTh only). L-NAME given with LPS (Group 1) completely inhibited the increase in the urinary NO2 + NO3 excretion. In fact there was a significant decrease in urinary excretion of the NO metabolites in both the LPS + L-NAME and the LPS + L-NAME + ANP groups,

compared to baseline. GTN administration in LPS + L-NAME treated rats induced a significantly higher urinary excretion of NO2 +

NO3 (P < 0.05 vs. LPS + L-NAME) which was similar to that obtained in rats given GTN alone. Thus L-NAME did not block the increase in NO2 + NO3 excretion resulting from GTN infusion.

Urinary cGMP excretion rates for the experimental groups are shown in Table 2. LPS caused an increase in the urinary excretion

bp < 0.01 vs. Group 1 and Group 3

P < 0.01 vs. baseline

of cGMP (UVCGMP) of 1.7-fold compared to baseline. When L-NAME was given with LPS (Group 1) no increase occurred.

cp < 0.01 vs. Group 2

GTN alone (Group 4) caused no increase in urinary cGMP

d Urine was collected for 7 hr, except in Group 3 in which the collection time averaged 5 hr.

excretion. The infusion of GTN in LPS + L-NAME treated rats also had no effect on cGMP excretion, while ANP infusion caused

a substantial increase (P < 0.05 vs. Groups 1 and 2).

Results All rats that received LPS + L-NAME (Groups 1, 2 and 3) appeared lethargic and ill at the end of the experimental period. In Group 3 (LPS + L-NAME + ANP), all but 2 of the 12 rats became moribund and had to be killed early, between 3 and 6-1/2 hours after the injection of LPS. Mean arterial blood pressures for rats in Groups 1, 2, 3 and 5 are shown in Figure 2. Administration of LPS only was followed by a slight decrease in blood pressure from 137 5.6 before injection to 125 5.3 mm Hg at the end of the experiment. As

While rats injected with LPS only showed very little evidence of

glomerular thrombosis, the animals that had been injected with LPS + L-NAME demonstrated extensive thrombosis, often involving the entire glomerular capillary bed (Fig. 3A). The extent of thrombosis was similar in the subgroup of five rats who along with LPS and L-NAME also received an intravenous infusion of normal saline at a rate of 1.2 mi/hr (data not shown). For this reason these rats were included in the analysis of the LPS + L-NAME group. In the LPS + L-NAME + GTN group (Group 2), only scattered fibrin thrombin were seen in glomerular capillaries (Fig. 3B). The thrombosis score was 1.37

expected, the blood pressure increased in all rats receiving compared to 3.02 L-NAME. The mean blood pressure values were not significantly different between all the groups that received L-NAME, but were significantly higher (P < 0.05) in these three groups compared to the "LPS only" Group at time 1, 3, 5 and 7 hours.

Urinary flow rate increased compared to baseline in all five experimental Groups 1,2, 3 and 5 (Table 1). The increase was most marked for the LPS + L-NAME + GTN group, where it was more than four times the baseline level. No diuretic effect of ANP was noted in addition to that observed with LPS + L-NAME alone. Rats given GTN only (Group 4) did not show an increase in urinary flow

0.5 in this group,

0.4 (P < 0.025) for the LPS + L-NAME

group, and not significantly different than in the LPS only group (0.59 0.4, P = 0.06). The administration of large doses of ANP

did not protect the kidneys from thrombosis. The thrombosis score in rats treated with LPS + L-NAME + ANP was 2.68

0.5,

not significantly different from that obtained in the LPS + L-NAME group (Fig. 4).

In separate studies we also tested another NO-independent

vasodilator, hydralazine. These rats developed glomerular thrombosis to a similar degree as rats treated with LPS + L-NAME or with LPS + L-NAME + ANP (data not shown). Like the LPS +

714

Westberg et al: Nitric oxide and gloinerular thrombosis

Westberg et al: Nitric oxide and glomerular thrombosis

715

Fig. 3. (A). Representative photomicrograph of glomendi from a rat treated with LPS + L-NAME, as described in Methods and Figure 1. Extensive thrombi are seen within most capillary loops (glomerular thrombosis score = 4). (Masson Trichrome stain, original magnification X 460). (B). Photomicrograph

of a glomerulus from a rat treated with LPS + L-NAME + GTN, as described in Methods and Figure 1. As shown, most glomeruli from the rats in this group appeared normal with no evidence of thrombi (glomerular thrombosis score = 0. However, some glomeruli from this group had thrombi in a minority of the glomerular capillaries (see Fig. 4). (Masson Trichrome stain, original magnification X 460).

rats. Our findings argue against this contention for several reasons: (1) L-NAME alone does not cause glomerular thrombosis [7]; (2) although organic nitrates are vasodilators, infusion of GTN in amounts similar to those used in the current study was found not to affect renal vascular resistance in rats [20]; (3) ANP has been shown to ameliorate the development of acute renal

4 0 0 0)

3

(1) Cl)

0

failure caused by vasoconstriction and ischemia [14, 15]. However,

.0

E

0 2

*

(Cl

E

0

1

0 0 LPS÷L-NAME

LPS÷L-NAME÷GTN LPS+L-NAME+ANP

Fig. 4. Glomendar thrombosis score in rats treated with LPS only (N = 15),

LPS + L-NAME (N = 19), LPS + L-NAME + GTN (N = 15), orLPS + L-NAME + ANP (N = 12). The experimental design is given in Figure 1 and in the Methods. The glomerular thrombosis score is based on the percentage of capillaries within a glomerulus which have thrombi. Data shown represent mean SEM from at least fifty glomeruli from each

group. () denotes P < 0.05 versus both LPS + L-NAME and LPS +

we found that ANP administered in doses large enough to cause a substantial increase in urinary cGMP levels failed to prevent heavy fibrin deposition in the glomeruli; and (4) another unrelated vasodilator, hydralazine, also failed to prevent the glomerular thrombosis observed with LPS + L-NAME. Thus, we conclude that renal vasoconstriction, leading to a decrease in renal blood flow, does not account for the thrombosis observed in these rats. Although there are many studies showing that GTN generates NO in vitro [21—23], our study seems to be the first to show that the

administration of an organic nitrate in vivo is followed by an increase in urinary excretion of NO metabolites. To our initial surprise, no increase in urinary cGMP occurred during GTN infusion, either when GTN was infused alone or together with LPS + L-NAME. Previous studies in our laboratory have shown an increase in urinary cGMP excretion in rats where NO synthesis

L-NAME + ANP groups.

was stimulated with acetylcholine [19] or LPS [7]. It is well

L-NAME + ANP, rats receiving LPS + L-NAME + hydralazine also showed reduced survival time in this protocol.

established that organic nitrates, such as GTN, vasodilate through the activation of soluble guanylate cyclase resulting in an increase in intracellular cGMP [24, 25]. However, infusion of GTN was found not to increase the urinary excretion of cGMP in patients

[26]. Furthermore, in patients with congestive heart failure,

Discussion nitroglycerin ointment in doses sufficient to have a clear hemodyWe have previously demonstrated that increased production of namic effect did not cause a measurable increase in plasma cGMP NO during endotoxemia was crucial in protecting against glomer- [27]. Thus, it appears that in vivo administration of organic

ular capillary thrombosis [7]. Rats given LPS increased their nitrates does not result in increases in urinary cGMP, despite urinary excretion of NO2 + NO3 and did not show substantial increases in NO metabolites in the urine. The reason for this fibrin deposition in the glomeruli. When the NO formation was disparity is not known. This is the first study that shows that an organic nitrate can inhibited with L-NAME, extensive glomerular thrombosis developed. L-NAME alone did not cause fibrin deposition [7]. In the prevent the formation of glomerular thrombosis. However, an current study, we have demonstrated that when endogenous NO antithrombotic effect of nitrates has previously been suggested synthesis is inhibited, exogenous nitric oxide can protect the [28]. Platelet deposition on damaged endothelium in the carotid glomeruli from endotoxin-induced thrombosis. As an exogenous NO donor we chose GTN [17, 18], in a dose that gave a threefold

artery of pigs can be inhibited by organic nitrates [29]. Similarly, platelet aggregates in the retina of humans can be disaggregated using GTN [30]. Furthermore, patients treated with intravenous

increase in the urinary excretion of NO2 + NO3, to mimic the urinary excretion of NO metabolites seen in LPS-treated rats GTN have prolonged bleeding times and decreased platelet (Table 1). In contrast ANP, a cGMP dependent but NO-indepen- adhesion to endothelial cells [28, 31]. ANP, on the other hand, has dent vasodilator, caused no increase in the excretion of NO not been found to affect platelets, presumably because ANP acts metabolites, and glomerular thrombosis was not prevented. More- only on particulate guanylate cyclase, which is not present in over, hydralazine, another powerful vasodilator with a different platelets [32]. To further elucidate the action of GTN on platelets, mechanism of action from either GTN or ANP, also failed to we attempted to measure the cGMP content of platelets from the rats in our study [6, 33]. We found that in rats given LPS + arrest glomerular thrombosis. Since L-NAME in the dose used in this study acts as a L-NAME, the platelets had aggregated and could not be effecvasoconstrictor [7, 19], we thought it important to consider the tively separated from the erythrocytes. In rats given LPS + possibility that L-NAME induced vasoconstriction and ischemia L-NAME + GTN, the platelets were not aggregated, but the lack may contribute to the glomerular thrombosis seen in LPS-treated of comparison groups made it impossible to assess the effect of

716

Westberg et

al: Nitric oxide and glomerular thrombosis

Inhibition of NO synthesis in septic shock. (letter) Lancet 339:435, GTN on platelet cGMP. However, the observation that platelets 1992 from LPS + L-NAME rats are highly aggregatory is consistent with the concept that platelet aggregation may contribute to the 12. COBB JP, NATAN5ON C, HOFFMAN WI), LODATO F, Bn'ncs 5, 1(0EV CA, SOLOMON MA, ELIN RJ, Hossnmn JM, DANNER R: W-Aminoglomerular thrombosis we observe in these animals. Furthermore, L-Arginine, an inhibitor of nitric oxide synthase, raises vascular the apparent correction of this enhanced aggregation in the LPS resistance but increases mortality rates in awake canines challenged + L-NAME + GTN group, supports our hypothesis that exogewith endotoxin. J Exp Med 176:1175—1182, 1992 nous NO protects against glomerular thrombosis by decreasing 13. WRIGHT CE, REE5 DD, MONCADA 5: Protective and pathological roles of nitric oxide in endotoxin shock, Cardiovasc Res 26:48—57, 1992 platelet aggregation. 14. Sia&w 5, WEIDMAJ4 P, ZIMMERMAN A: Urodilatin, not nitroprusside, Based on our studies, we conclude that either nitric oxide, or a combined with dopamine, reverses ischemic acute renal failure. substance akin to it, is released from GTN in quantities sufficient Kidney Int 42:1153—1159, 1992 to increase NO levels in vivo, as evidenced by the increased 15. ScHAwiuw's K, HEIDBRENNER E, GRIMM D, HEIDLAND A: Norepinephrine-induced acute renal failure: Beneficial effects of atrial urinary NO2 + NO3 excretion. Furthermore, the NO produced natriuretic factor. Nephron 44:240—244, 1986 from GTN prevents glomerular thrombosis in rats given LPS and inhibitors of endogenous NOS. We speculate that the prevention 16. RAn L, CHIOU X-C, OWENS R, WRIGLEY B: Therapeutic implications

of glomerular capillary thrombosis by GTN is primarily due to the effect of NO on platelet function. If the use of inhibitors of NOS

for the treatment of septic shock becomes clinically feasible, simultaneous treatment with nitroglycerin or similar exogenous NO donors may prevent intravascular coagulation and organ thrombosis. Acknowledgments Gunnar Westberg received support from Astra-Hassle Inc., Sweden. We thank Jonathon P. Tolins, M.D., for helpful suggestions and review of

the manuscript. The technical assistance of Ms. Karen Coffee, Ms. Kimberly Freude and Dawn Holmes is gratefully acknowledged. Reprint requests to Leopoldo Rae); M.D., Renal Section 111 J, VA Medical Center, One Veteran's Drive, Minneapolis, Minnesota 55417, USA.

References 1. RAn L, KEANE WF, Micaa AF: Unilateral Shwartzman reaction; Cortical necrosis in one kidney following in vivo perfusion with endotoxin. Kidney mt 12:91—95, 1977 2. GAYNOR E, DEB0uvIER CA, SPART TH: Vascular lesions: Possible

pathogenetic basis for the generalized Shwartzman reaction. Science 170:986—988, 1970 3. BONE RC: The pathogenesis of sepsis. Ann mt Med 115:457—469, 1991

4. STUEHR DJ, MARLErrA MA: Mammalian biosynthesis: Mouse mac-

rophages produce nitrite and nitrate in response to Escherichia coli

of hypertension-induced glomerular injury. Am J Med 79(3C):37—41, 1985 17. FUNG H-L, CHUNG S-J, BAUER JA, CHONG 5, KOWALUK EA: Biochemical mechanism of organic nitrate action. Am J Cardiol 70:4B— lOB, 1992 18. VANN AF, MORDVINTSEV PJ, KLE5HCHEV AL: [Nitric oxide appear-

ance in animal tissues in vivo] (Russian). Stud Biophys 102:135—143, 1984 19. TouNs JP, PALMER RMJ, MONCADA 5, RAIJ L: Role of endotheliumderived relaxing factor in regulation of renal hemodynamic responses. Am J Physiol 258:H665—H662, 1990 20. FLAIM F, WEITZEL RL, ZEus R: Mechanism of action of nitroglycerin during exercise in a rat model of heart failure. Improvement of blood flow to the renal, splanchnic and cutaneous beds, Cisc Res 49:458—468, 1981 21. KUHN M, OnEN A, FROHLICH IC, FOR5TERMANN U: Glyceryl trini-

trate increases platelet cyclic GMP after metabolism in fibroblasts. EurJPharm 200:175—178, 1991 22. FEELISCH M, KELM M: Biotransformation of organic nitrates to nitric

oxide by vascular smooth muscle and endothelial cells. Biochem Biophys Res Corn 180:286—293, 1991

23. NOACK E, FEELI5CH M: Molecular mechanisms of nitrovasodilator bioactivation, Basic Res Cardiol 86 (Suppl 2):37—50, 1991 24. BOHME E, GRAF H, SCHULTZ 0: Effects of sodium nitroprusside and other smooth muscle relaxants on cyclic GMP-formation in smooth muscle and platelets. Adv Cycl NucI Res 9:131—143, 1978 25. AXELS5ON KL, ANDER55ON RGG: Tolerance towards nitroglycerin,

induced in vivo, is correlated to a reduced cGMP response and an alteration in cGMP turnover. Eur J Pharm 88:71—79, 1983

lipopolysaccharide. Proc Nail Acad Sci USA 82:7738—7742, 1985

26. VORDERwINKLER KP, ARTNER-DWORZAIC E, JAKOB 0, MAIR J, DIEN-

5. RnoMsti MW, VALLANCE P, WHITLEY 0, FOXwELL N, MONCADA 5:

5171 F, PICHLER M, PUSCHENDORF B: Release of cyclic guanosine monophosphate evaluated as a diagnostic tool in cardiac diseases. Clin Chem 37:186—190, 1991 27. OGA5AWARA B, OGAWA K, SASSA H: Effect of nitroglycerin ointment

Platelet adhesion to human vascular endothelium is modulated by constitutive and cytokine induced nitric oxide. Cardiovasc Res 27: 1380—1382, 1993

6. MELLION BT, IGNARRO 12, Om.smlN EU, PoNmcoRvo EU, Hnw' AL, Kuowrrz PJ: Evidence for the inhibitory role of guanosine 3',S'monophosphate in ADP-induced human platelet aggregation in the presence of nitric oxice and related vasodilators. Blood 57:946—955, 1981 7. SHULTZ, PJ, R.&u L: Endogenously synthesized nitric oxide prevents endotoxin-induced glomerular thrombosis. J Clin Invest 90:1718—1725, 1992 8. PARRILLO JE, PARKER MM, NATA.N5ON C, SUFFREDINJ RL, DANNER

on plasma norepinephrine and cyclic nucleotides in congestive heart failure. J Cardiovasc Pharm 3:867—875, 1981

28. LoscALzo J: Antiplatelet and antithrombotic effect of organic nitrates. Am J Cardiol 70:18B—22B, 1992 29. LAid JY, CHESEBRO JH, FESTER V: Platelets, vasoconstriction, and nitroglycerin during arterial wall injury. A new antithrombotic role for an old drug. Circulation 78:712—716, 1988 30. KURITZKY 5: Nitroglycerin to treat acute loss of vision. N EngI J Med 323:1438, 1990

RE, CuNNI0N RE, OGNIBENE FP: Septic shock in humans: Advances in the understanding of pathogenesis, cardiovascular dysfunction, and therapy. Ann Intern Med 113:227—236, 1990 9. KILBOURN RU, JuuIw4 A, GRoss SS, GRIFFITh OW, LEVI R, ADAMs J, LODATO RF: Reversal of endotoxin-mediated shock by N°-mcthylL-arginine, an inhibitor of nitric oxide synthesis. Biochem Biophys Res Commun 172:1132—1138, 1990 10, PEmos, A, BENNET D, VALLANCE P: Effect of nitric oxide synthase

31. AISSA J, FEELI5CH M: Generation of nitric oxide account for the anti-platelet effects of organic nitrates—The role of plasma compo-

inhibitors on hypotension in patients with septic shock. Lancet 338:

33. WATANABE H, KAKJHANA M, OHTSUKA 5, ENOMOTO T, YA5UI K,

1557—1558, 1991

11. GRouia'eos 5, Scttiamici J, CAKMAKCI H, JUNG Fl, L&RGIADER F:

nents and vascular cells, In: Nitric Oxide from L-Arginine. A Bioregulatory System, edited by MONCADA 5, HIGGS EA, New York, Excerpta Medica 1990, pp 142—144 32. LANG CC, Lw CS, BELCH JJ, STRUTHER5 AD: Effect of natriuretic factor on platelet function in whole blood ex-vivo in man. Eur I Clin Pharm 39:589—591, 1990 SUGISHITA Y: A potentially useful indicator to evaluate the effects of nitroglycerin and nitrate tolerance. Circulation 88:29—36, 1993