INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

NITRIC OXIDE (NO) can regulate oxygen supply to tissues by activating soluble guanylate cyclase in vascular smooth muscle, resulting in vascular relaxation and increased blood flow (16, 23). However, NO can also potentially regulate tissue oxygen consumption by binding to the oxygen-binding site of cytochrome oxidase, resulting in reversible inhibition of mitochondrial respiration (5, 7, 9, 31). Binding of NO to cytochrome oxidase is competitive with oxygen, and at physiological levels of oxygen in tissue (roughly 30 µM O₂) half-maximal inhibition of respiration occurs at 60 nM NO (7). This concentration is similar to that required for half-maximal activation of soluble guanylate cyclase (45 nM NO) (1). Higher concentrations of NO or its derivatives peroxynitrite and S-nitrosothiols can irreversibly inhibit respiration at multiple sites within mitochondria (4, 10, 14, 15, 20, 26). Reversible inhibition of oxygen consumption by NO has been found in isolated cytochrome oxidase (7, 13, 40), mitochondria (9, 25, 31), and cultured cells (6, 11, 19). In vivo it has been observed that inhibitors of NO synthase cause large increases in tissue and whole body oxygen consumption that are not attributable to any changes in vascular supply (17, 18, 21, 22, 33, 34), suggesting that basal release of NO tonically inhibits tissue respiration in vivo (41, 42). However, it is unclear whether any such NO inhibition of tissue respiration is mediated by a direct action of NO on mitochondrial respiration or indirectly, e.g., via cGMP.

Induction of the inducible isoform of NO synthase (iNOS) by endotoxin and cytokines results in a high, sustained concentration of NO (27), giving rise to reversible inhibition of cellular respiration rate in astrocytes (6) and cells coincubated with iNOS-expressing macrophages (8). iNOS expression also causes irreversible inhibition of mitochondrial respiratory components in astrocytes (3), hepatocytes (36), tumor cells (15, 37), and vascular smooth muscle cells (12, 38).

During local inflammation or the systemic inflammation of sepsis, iNOS is induced in a wide range of tissue cells (29, 30). Sepsis or endotoxemia can cause hypotension, vascular insufficiency, lactic acidosis, and multiple organ failure, and these symptoms have been associated with excessive NO production from iNOS (29, 30, 39). Mitochondrial dysfunction has also been implicated in sepsis, and in principle septic symptoms could be due to NO inhibition of mitochondrial respiration alone (28, 32). We set out to test 1) whether the induction of iNOS in the aorta and endothelial cells by endotoxin and interferon- would result in a significant inhibition of respiration, and 2) the mechanism of any such inhibition, its reversibility, and sensitivity to oxygen concentration.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Wistar rats (12-16 wk old) were euthanized with CO₂. Thoracic aortas were removed, cleaned of adhering fat and connective tissue in Hanks' balanced salt solution supplemented with 5 µM indomethacin, and cut into ~4-mm rings. Rings were incubated for 24 h at 37°C in DMEM (without serum) plus 500 IU/ml penicillin and streptomycin in an incubator gassed with 95% air-5% CO₂. Alternatively, to induce iNOS, aortic rings were incubated for 24 h in DMEM (plus 500 IU/ml penicillin and 500 IU/ml streptomycin) supplemented with 10 µg/ml lipopolysaccharide (LPS endotoxin from Salmonella typhimurium; Sigma) and 50 U/ml interferon-. For measurement of respiratory rate, aortic rings were washed with Krebs buffer, mounted on steel hooks, and put into 1 ml of Krebs-HEPES buffer (in mM: 118 NaCl, 4.8 KCl, 1.2 KH₂PO₄, 1.2 MgSO₄, 1 CaCl₂, 11 glucose, and 25 HEPES, pH 7.4) plus 0.5 mM L-arginine. Oxygen consumption by aortic rings was measured in a sealed and stirred vessel with a Clarke-type oxygen electrode built into the bottom of the vessel (Rank Brothers, Bottisham) maintained at 37°C.

Endothelial cells were isolated from rat aortas by digestion (7 min at 37°C) with 3 mg/ml collagenase (Sigma C-0130) in medium 199 and cultured in DMEM plus 15% fetal calf serum, 5 ng/ml basic fibroblast growth factor, 500 IU/ml penicillin, and 500 IU/ml streptomycin in an incubator gassed with 95% air-5% CO₂ at 37°C. Respiratory measurements were made on 7-9th passage cultures that were ~80% confluent. Cells were activated by adding 10 µg/ml LPS (endotoxin from S. typhimurium, Sigma) and 50 U/ml interferon- to the cultures for 24 h before respiratory measurements. To measure oxygen consumption, cells were removed from culture flasks by gentle scraping and centrifuged and resuspended at ~3 million cells per millimeter in the Krebs-HEPES buffer (content given above) plus 0.1 mM L-arginine. Of this cell suspension, 0.7 ml was added to a sealed, stirred vessel containing both oxygen and NO (World Precision Instruments) electrodes maintained at 37°C (7). The NO electrode was calibrated with NO-saturated, deoxygenated water, assuming this contains 2 mM NO (6).

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

We tested whether the induction of iNOS in rat aortic rings by endotoxin and interferon- would inhibit aortic oxygen consumption. Figures 1 and 2 show that the oxygen consumption of rings maintained in culture conditions with endotoxin and interferon- for 24 h was strongly inhibited relative to rings cultured in the absence of endotoxin and interferon-. Moreover, the oxygen consumption of endotoxin- and interferon--treated rings became strongly oxygen dependent, so that the rate of oxygen consumption was almost proportional to oxygen concentration over the physiological range up to 160 µM O₂ (Figs. 1 and 2). Reoxygenation of hypoxic rings returned the oxygen consumption rate to the previous high, normoxic rate (data not shown). In contrast, the oxygen consumption of untreated rings was relatively oxygen independent, remaining linear until low oxygen levels were reached (Figs. 1 and 2).

View larger version (8K): [in this window] [in a new window]   Fig. 1.   Representative traces showing oxygen consumption by aortic rings. Five aortic rings (13-19 mg wet wt) were incubated in 1 ml Krebs-HEPES buffer containing 0.5 mM L-arginine. Traces b and d, control aorta; trace c, aorta treated with lipopolysaccharide (LPS)-interferon-; trace a, LPS-interferon--treated aorta preincubated 1 h with 100 µM N-(3-(aminomethyl)benzyl)acetamide (1400W). Where indicated, aliquots of 1 µM nitric oxide (NO) were added to control aorta causing immediate but reversible inhibition of respiration. Horizontal dotted line, zero level of oxygen. Representative traces are of at least 3 experiments in each condition.

View larger version (23K): [in this window] [in a new window]   Fig. 2.   LPS-interferon--induced inhibition of respiration rate of aortic rings is oxygen dependent and is reversed by oxyhemoglobin (HbO2) or 1400W but not by ODQ. Experimental conditions were as in Fig. 1. Aortic rings were preincubated for 5 min with 10 µM ODQ or for 1 h with 100 µM 1400W; 10 µM HbO2 was added immediately before measuring oxygen consumption by aorta. Data are means ± SE of at least 3 experiments. *Statistically significant effect if compared with control at the same oxygen concentration.

To test whether the inhibition of respiration was rapidly reversible and due to NO, we added either a NO scavenger (oxyhemoglobin) or an iNOS inhibitor [N-(3-(aminomethyl)benzyl)acetamide · 2HCl, 1400W] to the endotoxin- and interferon- treated rings before measuring oxygen consumption. These additions resulted in reversal of the respiratory inhibition (Figs. 1 and 2), indicating that the inhibition of oxygen consumption was rapidly reversible and due to NO from iNOS.

Because most of the physiological effects of NO are mediated by stimulation of soluble guanylate cyclase, we tested whether a specific inhibitor of this enzyme (1H-[1,2,4]oxadiazolo[4,3-a]quinoxalin-1-one, ODQ) could reverse the inhibition of respiration in endotoxin- and interferon--treated rings. ODQ did not reverse the respiratory inhibition (Fig. 2), indicating that the NO inhibition of respiration was not mediated by cGMP.

To test whether NO alone could cause respiratory inhibition, we added authentic NO (1 µM) to control rings. This caused an immediate inhibition of aortic respiration, which completely reversed, however, during several minutes (Fig. 1), indicating that NO is a potent, reversible inhibitor of aortic respiration. Moreover, NO inhibition of aortic respiration was greater at lower oxygen concentrations (Fig. 1) and was insensitive to the soluble guanylate cyclase inhibitor ODQ (data not shown). All of these results in aortic rings are consistent with endotoxin and interferon- causing inhibition of aortic respiration via NO inhibition of cytochrome oxidase in competition with oxygen.

We further tested whether induction of iNOS in aortic endothelial cells in culture would result in significant inhibition of cellular respiration and whether any such inhibition would be reversible. Figures 3 (representative traces) and 4 (mean data) show that endothelial cells activated for 24 h with endotoxin and interferon- produced NO in the presence of L-arginine (steady-state concentration of 0.54 ± 0.11 µM NO) and that cellular respiration was substantially inhibited relative to untreated cells (92% inhibition). This inhibition was immediately reversed on addition of the NO scavenger oxyhemoglobin. This is again consistent with NO from iNOS causing potent but reversible inhibition of cellular respiration via NO binding to cytochrome oxidase.

View larger version (13K): [in this window] [in a new window]   Fig. 3.   LPS-interferon- activation of endothelial cells causes NO production and inhibition of cellular respiration, reversed by HbO2. Rat aortic endothelial cells were cultured without (control, A) or with (B) LPS-interferon- for 24 h. Then, 3.1 million cells (LPS-interferon- treated) or 3.8 million cells (control) were resuspended in 0.7 ml Krebs-HEPES plus 0.1 mM L-arginine and placed in a sealed, stirred vessel with NO and oxygen electrodes. HbO2 (10 µM) was subsequently added to remove the NO.

View larger version (18K): [in this window] [in a new window]   Fig. 4.   LPS-interferon- activation of endothelial cells causes inhibition of cellular respiration, reversed by HbO2: pooled results. Values are means ± SE of 3 separate experiments measuring the oxygen consumption rate at 25-35% of oxygen saturation for control and LPS-interferon--treated cells (immediately before and after addition of oxyhemoglobin). *Statistically significant difference from control; # statistically significant difference from LPS-interferon--treated aortas in the absence of HbO2.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

We have shown here that incubation with endotoxin and interferon- causes a substantial inhibition of oxygen consumption in both isolated aortas and aortic endothelial cells. This inhibition is largely or completely reversible by an NO scavenger or iNOS inhibitor, suggesting that NO from iNOS is reversibly inhibiting respiration.

Treatment of mice with endotoxin has previously been shown to reduce myocardial oxygen consumption ex vivo, an effect that was reversed by addition of an inhibitor of NO synthase (24). This inhibition of respiration did not occur in animals that were knockout for iNOS (24).

We have also shown that addition of 1 µM authentic NO causes rapid and reversible inhibition of aortic respiration, which is greater at lower oxygen concentrations. The NO-induced inhibition of respiration is likely to be due to NO binding to cytochrome oxidase, because we have previously shown that this causes rapid, reversible inhibition that is competitive with oxygen and occurs over the range of NO concentrations used here (5, 7). This is also consistent with the finding that endotoxin- and interferon--treated aortic rings have an oxygen consumption rate that is almost proportional to oxygen level, whereas untreated rings or treated rings supplemented with an NO scavenger or iNOS inhibitor have oxygen consumption rates that are relatively insensitive to oxygen level.

It has previously been found that endotoxin and interferon- treatment of cultured vascular smooth muscle cells for 48 h resulted in an irreversible inhibition of mitochondrial respiration, attributed to activation of poly-ADP ribosyltransferase (PARS or PARP) and subsequent NAD⁺ and ATP depletion (38). However, in the conditions of our experiments, we found no significant irreversible inhibition of cellular respiration at 24 h, although this does not exclude such an inhibition appearing later or in different conditions. In particular, in vivo neutrophil recruitment and activation during inflammation may provide a source of superoxide from which peroxynitrite may be derived and which might cause an irreversible inhibition of respiration.

A substantial inhibition of cellular respiration by iNOS-derived NO in competition with oxygen has important implications for inflammation and sepsis. It has been suggested that tissue dysoxia during septic shock, and the consequent organ failure, may be due to inhibition of respiration by NO (24, 28, 32). Furthermore, although no experimental evidence is available at present, it is likely that inhibition of respiration also contributes to the vasodilatation and hyporeactivity to vasoconstrictors.

Inflammation and sepsis are known to be accompanied by glycolysis, increased production of lactate, and inhibition of tissue function (32, 35). Our present and previous findings suggest that these changes might indeed be due to inhibition of respiration by NO and the ensuing oxidative stress (2, 5, 8, 24).