TNF- Contributes to Endothelial Dysfunction in Ischemia/Reperfusion Injury

【摘要】  Background- Despite the importance of endothelial function for coronary regulation, there is little information and virtually no consensus about the causal mechanisms of endothelial dysfunction in myocardial ischemia/reperfusion (I/R) injury. Because tumor necrosis factor- (TNF- ) is reportedly expressed during ischemia and can induce vascular inflammation leading to endothelial dysfunction, we hypothesized that this inflammatory cytokine may play a pivotal role in I/R injury-induced coronary endothelial dysfunction.

Methods and Results- To test this hypothesis, we used a murine model of I/R (30 minutes/90 minutes) in conjunction with neutralizing antibodies to block the actions of TNF-. TNF- 4-fold after I/R. To determine whether TNF- abrogates endothelial function after I/R, we assessed endothelial-dependent (ACh) and endothelial-independent (SNP) vasodilation. In sham controls, ACh induced dose-dependent vasodilation that was blocked by the nitric oxide synthase (NOS) inhibitor L-NMMA (10 µmol/L), suggesting a key role for NO. In the I/R group, dilation to ACh was blunted, but SNP-induced dilation was preserved. Subsequent incubation of vessels with the superoxide (O 2 ·- ) scavenger (TEMPOL), or with the inhibitors of xanthine oxidase (allopurinol, oxypurinol), or previous administration of anti-TNF- restored endothelium-dependent dilation in the I/R group and reduced I/R-stimulated O 2 ·- production in arteriolar endothelial cells. Activation of xanthine oxidase with I/R was prevented by allopurinol or anti-TNF-.

Conclusions- These results suggest that myocardial I/R initiates expression of TNF-, which induces activation of xanthine oxidase and production of O 2 ·-, leading to coronary endothelial dysfunction.

We examined the role of TNF- in endothelial dysfunction after myocardial ischemia-reperfusion (I/R) injury. I/R injury compromised endothelium-dependent dilation to acetylcholine (ACh). TNF- 4-fold after I/R. Anti-TNF- restored vasodilation to ACh and reduced xanthine oxidase activity, the production of O 2 ·-.

【关键词】  coronary artery disease endothelial function nitric oxide microcirculation reactive oxygen species

Introduction

Tumor necrosis factor- (TNF- ) is an inflammatory cytokine 1 that is expressed by macrophages and cardiac tissue early during the myocardial ischemia-reperfusion (I/R) injury. 2,3 Elevations of TNF- expression also appear to cause cardiomyopathy. 4,5 Interestingly, both cardiomyopathy and I/R injury are characterized by endothelial dysfunction, but, a putative role for TNF- in the abnormal vasodilatory responses during I/R has not been elucidated.

Studies from our laboratory 6 and others 7,8 have shown that I/R produces vascular endothelial dysfunction as defined by abrogated endothelium-dependent dilation. This adverse effect on endothelial function can occur acutely 6 or chronically (I/R injury in the pig can produce endothelial dysfunction for weeks). 9 However, endothelial dysfunction may be amplified by neutrophil-generated factors including oxygen-derived free radicals, cytokines, proteases, and lipid mediators. 10 I/R generates high levels of free radicals 11 composed of both reactive oxygen intermediates and nitric oxide (NO) via a complex sequence of events. When generated in sufficient concentrations, free radicals directly injure the myocardium and may even cause cell death. 12 The production of O 2 ·- is generally accepted as a contributing factor to I/R-induced microvascular injury. In addition, a decrease in endothelium-dependent dilation occurs soon after the generation of O 2 ·- radicals during reperfusion, 11,13-15 thus suggesting that endothelial generation of O 2 ·- radicals acts as a triggering mechanism for endothelial dysfunction.

A role for TNF- in myocardial ischemic injury was directly implicated by the work of Heusch and Schulz. 16 These investigators observed that in a model of coronary embolization, expression of TNF- produced progressive coronary constriction that exacerbated the initial magnitude of ischemia. Current evidence suggests that TNF- participates in myocardial I/R injury and cardiac allograft rejection. 17,18 A mechanism by which TNF- may induce injury pertains to the production of O 2 ·-, but linking TNF- expression to O 2 ·- production and endothelial dysfunction in endothelial injury during I/R has not been previously proposed. The purpose of our study was to determine the role of TNF- in endothelial dysfunction during I/R injury by studying NO-mediated vasodilation to ACh in isolated and pressurized coronary arterioles, O 2 ·- production, activation of xanthine oxidase, and expression of TNF- with and without neutralizing antibodies to TNF-.

Methods

Please see online Material at http://atvb.ahajournals.org.

Data Analysis

At the end of each experiment, the vessel was relaxed with 100 µmol/L SNP to obtain its maximal diameter at 60 cm H 2 O intraluminal pressure. 19,20 All diameter changes in response to agonists were normalized to the vasodilation in response to 100 µmol/L SNP and expressed as a percentage of maximal dilation. All data are presented as mean±SEM. Statistical comparisons of vasomotor responses under various treatments were performed with 2-way ANOVA and intergroup differences were tested with Bonferonni inequality. Significance was accepted at P <0.05.

Results

I/R Increased TNF- mRNA Expression and Plasma Concentration of TNF- in Murine Coronary Arterioles

Figure 1 A shows the expression of TNF- mRNA (204 bp) in left ventricular coronary arterioles of sham and I/R groups; expression of TNF- 4-fold after I/R. Figure 1 B shows the inflammatory cytokine TNF- is significantly increased after I/R (n=4), but anti-TNF- attenuate the concentration of free TNF- in I/R because the bulk of the circulating cytokine is bound to the antibody. However, plasma concentration of IL-6, IL-8, and IFN- are not increased after I/R (data not shown). Our studies indicated that after I/R, levels of TNF- were the only cytokines to increase ( P <0.05).

Figure 1. Real-time polymerase chain reaction (PCR) of TNF- mRNA from left ventricular coronary arterioles in sham and I/R mice. A, Bar graph showing fold change of TNF- expression. TNF- expression is significantly higher in I/R mice than those in sham mice. B, Plasma concentration of inflammatory cytokine TNF- is significantly increased after I/R, but anti-TNF- attenuates the concentration of free TNF- in I/R. Data represent means±SD, n=4. * P <0.05 vs sham mice, # P <0.05 vs wild-type (WT)-I/R.

Effects of I/R on NO-Mediated Vasodilation to ACh

Isolated murine coronary arterioles from control and sham animals dilated dose-dependently to the endothelium-dependent agonist, ACh ( Figure 2 A). Administration of the nitric oxide synthase (NOS) inhibitor L-NMMA (10 µmol/L) reduced the vasodilatory responses to ACh compared with controls ( Figure 2 A). In a similar manner as L-NMMA, ACh-induced vasodilation was impaired after I/R (30 minutes/90 minutes). Function of smooth muscle was preserved, because SNP-induced vasodilation was equivalent in both groups ( Figure 2 B).

Figure 2. Isolated mice coronary arterioles from control mice (without I/R) dilated in a concentration-dependent manner to ACh (n=12, A). L-NMMA (10 µmol/L, 30 minutes, n=6) inhibited vasodilation to ACh vs control. ACh-induced vasodilation was impaired after I/R, but that to the endothelial-independent vasodilator, SNP, was normal (n=5, B). ACh-induced vasodilation was not impaired in sham group compared with the control (n=5). n=number of vessels. * P <0.05 vs control.

Role of TNF- in I/R-induced Vascular Dysfunction

Neutralizing antibodies to TNF- (I.P. 0.1 mg/mouse containing 16 mg protein/mL, administered 3 hours before initiating I/R) maintained NO-mediated coronary arteriolar dilation after I/R, but administration of nonimmune IgG did not (intraperitoneal 0.1 mg/mouse containing 16 mg protein/mL, administered 3 hours before initiating I/R) ( Figure 3 A). Anti-TNF- did not affect dilation in the sham group. We administered TNF- (1 ng/mL, 60 minutes) and assessed responses of microvessels to ACh to show the acute administration of TNF- could mimic the responses of I/R. ACh-induced dilation was significantly blunted by TNF- or L-NMMA, but combined treatment (TNF- plus L-NMMA) did not induce further abrogation of endothelial dilation ( Figure 3 B).

Figure 3. Neutralizing antibodies to TNF- (anti-TNF-, 16 mg protein/mL; 0.1 mL/mouse) restored NO-mediated coronary arteriolar dilation in murine I/R, but nonimmune IgG (0.1 mL/mouse containing 16 mg protein/mL) did not (n=4, A). * P <0.05 vs sham. B, Effect of acute TNF- stimulation (1 ng/mL, 60 minutes) on the ACh-induced dilation of coronary arterioles before and after treatment with NO synthesis inhibitor L-NMMA. ACh dilated the control vessels in a concentration-dependent manner (n=4, B). These dilations were inhibited by L-NMMA (10 µmol/L, n=5, B) or pretreating the vessels with TNF- (1 ng/mL, n=4, B). The combination of TNF- and L-NMMA did not further inhibit the ACh-induced vasodilation (n=5, B). * P <0.05 vs sham or control.

Roles of Superoxide, Xanthine Oxidase, Mitochondrial Respiratory Chain, and NAD(P)H Oxidase in I/R-Induced Vascular Dysfunction

Administration of an O 2 ·- scavenger, TEMPOL ( Figure 4 A; n=7) or xanthine oxidase inhibitor, allopurinol ( Figure 4 A; n=7), or oxypurinol ( Figure 4 B; n=4) maintained vasodilation to ACh in I/R, but rotenone or apocynin did not. Furthermore, TEMPOL (data not shown; n=3), allopurinol ( Figure 4 A; n=4), oxypurinol ( Figure 4 B; n=4), rotenone ( Figure 4 C; n=5), or apocynin ( Figure 4 C; n=4) did not affect ACh-induced vasodilation in sham groups.

Figure 4. Administration of a O 2 ·- scavenger, TEMPOL (1 mmol/L, n=7, A), or xanthine oxidase inhibitor, allopurinol (10 µmol/L, n=7, A), or oxypurinol (100 µmol/L, n=4, B), restored vasodilation after I/R, but NADPH oxidase inhibitor apocynin (100 µmol/L, n=4, C) or mitochondrial respiratory chain inhibitor rotenone (1 µmol/L, n=4, C) did not. Furthermore, incubation with either TEMPOL (data not shown), allopurinol (n=4, A), oxypurinol (n=4, B), apocynin (n=4, C) or rotenone (n=4, C) did not alter the vasodilatory responses to ACh in sham group. * P <0.05 vs sham.

I/R-Induced Superoxide Production in Murine Coronary Arterioles

In control conditions, DHE fluorescence revealed sparse levels of O 2 ·- throughout the vessel wall ( Figure 5 A). Incubation of arterioles isolated from coronary arterioles of control mice hearts with TNF- (1 ng/mL, 60 minutes) caused a pronounced increase in fluorescent signals in both endothelial and smooth muscle layers, indicating augmented levels of O 2 ·- production in these cell types. I/R also significantly elevated production of O 2 ·- in both endothelial and smooth muscle layers during I/R injury compared with sham control (vehicle). Administration of O 2 ·- scavenger TEMPOL, or allopurinol, or anti-TNF- abolished O 2 ·- production, but apocynin or nonimmune IgG did not (DHE fluorescence; Figure 5 A). Figure 5 B shows the results from EPR spectroscopy to quantify the production of O 2 ·-. O 2 ·- production increased after I/R compared with the sham group ( P <0.05) and, importantly, administration of anti-TNF- reduced the expression of O 2 ·- to the level observed in the sham group.

Figure 5. DHE-fluorescence imaging of O 2 ·- in coronary arterioles (A). DHE fluorescence was markedly elevated in both endothelial (arrow head) and vascular smooth muscle (arrow) cells of arteriolar sections after TNF- treatment (1 ng/mL, 60 minutes, n=4) and I/R (n=4) compared with vehicle (sham). Apocynin and nonimmune IgG did not, but TEMPOL, allopurinol, and anti-TNF-, markedly reduced the fluorescent signals (n=4). A, Data representative from 4 separate experiments. B, EPR analysis of O 2 ·- production in isolated coronary arterioles in WT (sham), I/R, and I/R plus anti-TNF- mice. a, Representative EPR tracing of tissue lysates obtained from WT, I/R, and I/R plus anti-TNF- mice. b, The actual concentrations of O 2 ·- normalized by protein concentration. n=4 (B). * P <0.05 vs sham; P <0.05 vs I/R.

I/R Increased Xanthine Oxidase Activity in Murine Coronary Arterioles

Xanthine oxidase activity was measured in isolated coronary arterioles from 6 groups (sham, I/R, anti-TNF- plus sham, anti-TNF- plus I/R, allopurinol plus sham, and allopurinol plus I/R) to investigate the connection between xanthine oxidase and TNF- in I/R-induced endothelial dysfunction ( Figure 6 ). Importantly, this increase in activity was prevented by anti-TNF- or allopurinol to levels similar in the sham group with or without treatments.

Figure 6. Xanthine oxidase activity in mouse left ventricular coronary arterioles. I/R significantly increased xanthine oxidase activity compared with sham mice. After neutralization with anti-TNF-, xanthine oxidase activity was decreased but is still higher than those in sham mice. Allopurinol abolished I/R-induced increase of xanthine oxidase activity. However, anti-TNF- and allopurinol did not affect the xanthine oxidase activity in sham mice. Data represent mean±SEM, n=9. * P <0.05 vs sham. # P <0.05 vs I/R.

Discussion

Our major finding is that antibody neutralization of TNF- prevented coronary endothelial dysfunction during myocardial I/R injury. We also found that neutralization of TNF- reduced O 2 ·- generation and xanthine oxidase activity during I/R, and blockade of xanthine oxidase mimicked the actions of anti-TNF- on O 2 ·- production and endothelial function. Molecular evidence indicated that the expression of TNF- (mRNA) was significantly increased in I/R injury. These results suggest that myocardial I/R initiates expression of TNF-, which induces activation of xanthine oxidase and production of O 2 ·-, leading to endothelial dysfunction. Our findings support the concept that TNF- plays a pivotal role in endothelial dysfunction during myocardial I/R injury. We believe that our findings further understanding of the mechanism(s) underlying endothelial dysfunction after myocardial I/R injury.

I/R and Endothelium-Dependent Dilation in Murine Coronary Microvessels

The mechanisms of I/R injury are multifactorial and incompletely understood. Previous studies 6-8 have shown that I/R produces impaired vascular endothelial function as defined by abrogated endothelium-dependent dilation. This impairment has been attributed to arginase, 6 depletion of tetrahydrobiopterin, 21 O 2 ·-, cytokines, proteases, and lipid mediators. 10 Endothelium-derived NO is an important endogenous vasodilator that regulates microvascular tone. The endothelial release of NO underlies the mechanism of coronary arteriolar dilation to ACh. 22,23 The endothelial NO component was evident in the present study because L-NMMA attenuated the ACh-induced vasodilation. Consistent with our previous studies, we show here that ACh evokes dilation of coronary arterioles (probably by activation of endothelial NOS). In the present study, endothelium-dependent ACh-induced vasodilation in mouse coronary artery was impaired after I/R, whereas endothelium-independent SNP-induced vasodilation was normal, and these diminished responses were not further reduced by treatment with a NOS inhibitor, suggesting that I/R produces endothelial dysfunction in a murine model of reperfusion injury. I/R injury is known to induce endothelial dysfunction, 21 and our results speak to a causative role for TNF- because anti-TNF- prevented endothelial dysfunction.

The Role of TNF- in Coronary Vascular Function in I/R Injury

Endothelial dysfunction is an important early recurring phenomenon in virtually all forms of I/R injury, including a variety of circulatory shock states. 10 The role of TNF- in ischemic injury in the intact heart is controversial with some groups reporting beneficial effects, 24,25 whereas others find a detrimental role. 26,27 TNF- clearly initiates expression of an entire spectrum of inflammatory cytokines. 23,36 Accordingly, we hypothesized that TNF- expressed in the endothelium might play a pivotal role in the regulation of NO-mediated vasodilation by increasing O 2 ·- production, thereby decreasing NO bioavailability. Although we did not unequivocally eliminate a role for other cytokines in I/R injury, our results show that IL-6, IL-8, and INF- were not elevated. Nevertheless, we are compelled in confine our conclusions to the effect of TNF- because we did not systematically study these other cytokines.

A critical issue about our interpretations pertains to the specificity of the anti-TNF-. Within this context, we have made several observations that demonstrate specificity of the antibody. 31 Although the antibody is polyclonal, it shows no cross-reactivity with other cytokines. 31 The ability of the antibody to bind to a few other recombinant cytokines was tested to a limited degree and we found binding only to murine TNF-. 31 The antigen was recombinant murine TNF- and therefore it is very unlikely that this IgG neutralizes other cytokines. 31 We used nonimmune control IgG. 31 Moreover, in an animal model of lipopolysaccharide (LPS)-induced shock, anti-TNF- completely blocked increases in plasma TNF-, because the antibody bound the cytokine. 31 In previous published and unpublished studies, anti-TNF- has not universally attenuated LPS-induced effects on pulmonary parameters. 32-35 This means TNF- is not solely responsible for LPS-induced effects, and our antibody does not block whatever it is that causes these effects. Thus, we believe with conviction that the antibody is specific for TNF-, and thus our interpretations and conclusions about the pivotal role this cytokine plays in endothelial I/R injury are correct.

There is no previous report on the endogenous role of TNF- in this respect to our knowledge, and it has been unclear whether TNF- plays a direct role in endothelial dysfunction during I/R injury. In the present study, we documented that TNF- is critical for the development of endothelial reperfusion injury. TNF- expression was significantly increased in murine coronary arterioles after I/R injury; neutralizing antibodies to TNF- restored NO-mediated coronary arteriolar dilation in the I/R group, but did not affect the endothelium-dependent vasodilation in sham controls. To further establish this deleterious role of TNF-, we attempted to mimic the endothelial injury by incubating isolated arterioles with TNF-. A 60-minute incubation of endothelium-intact coronary arterioles with TNF-, which did not alter resting diameter, blunted the ACh-induced vasodilatory response. Thus, TNF- under the conditions of our experimental protocols appears to have a deleterious effect on vascular function by inducing endothelial dysfunction.

One interesting observation was an increase in TNF- expression (4-fold) and an increase in plasma concentration of TNF- (6- fold) after a total of 120 minutes of I/R. Although, we are not sure of the cell type that is expressing TNF-, a component of the signal may be related to infiltrating cells such a macrophages. If TNF- is, as we and others suspect, one of the initiators of a cytokine cascade, then it may also be one of the genes that is induced very quickly after perturbation.

Roles of ROS and Xanthine Oxidase in Coronary Vascular Function in I/R Injury

Endothelial dysfunction appears to occur very early after reperfusion, signaled by the endothelial generation of a large burst of superoxide radicals. 14,28 The decrease in endothelium-dependent dilation has been shown to occur soon after the generation of O 2 ·- by the reperfused coronary endothelium, 14 suggesting that endothelial generation of O 2 ·- radicals acts as a trigger mechanism for endothelial dysfunction. O 2 ·- also reduces the bioavailability of NO, which would compromise dilation to endothelium-dependent stimuli. Thus, it appears that impaired NO-dependent functions, eg, vasodilation, are the direct result of the overproduction of O 2 ·- during the development of I/R injury. Although many studies document oxygen radical formation during I/R, there is a paucity of knowledge regarding mechanisms responsible for stimulating the production of O 2 ·-. Our results extend this knowledge by demonstrating the role of TNF- in I/R-induced endothelial injury via stimulation of xanthine oxidase to produce O 2 ·- as shown by our observation that anti-TNF- prevented I/R-induced xanthine oxidase activation and O 2 ·- production.

Although there are multiple intracellular sources for formation of oxygen free radicals [eg, mitochondria, NAD(P)H oxidase, etc], our results support the idea that the major enzyme activated by TNF- during I/R is xanthine oxidase. We are not sure that allopurinol, in vitro, is acting as a free radical scavenger; however, we can state with conviction, based on our new results, that allopurinol is inhibiting xanthine oxidase. Our results indicated that the pathway for TNF- -induced endothelial dysfunction is mediated by activation of xanthine oxidase and the subsequent production of O 2 ·-. This idea is supported by the finding of Pereda et al, 29 which shows that simultaneous inhibition of TNF- production and xanthine oxidase activity greatly reduced local systemic inflammatory response in acute pancreatitis. 29 Our study shows that I/R injury increased the xanthine oxidase activity, and antibody to TNF- and xanthine oxidase inhibitor, allopurinol, separately attenuated the xanthine oxidase activity in I/R injury. In the majority of studies cited pertaining to reperfusion injury, elevations in O 2 ·- production are thought to be a critical factor in this process. Our results demonstrate the production of TNF- is essential in eliciting this oxidative stress. The present study indicates that I/R increases TNF-, which stimulates endothelial generation of O 2 ·- through activation of xanthine oxidase in the endothelium and contributes to the endothelial dysfunction. To our knowledge, this is the first functional study to link the mechanism(s) of I/R injury-in terms of endothelial dysfunction-to TNF-, the subsequent activation of xanthine oxidase, and thus the production of O 2 ·- in coronary arteriolar endothelium. One caveat: we studied the role of TNF- in endothelial dysfunction only in an acute setting. The chronic endothelial injury that occurs for weeks after I/R 9 may be the result of a different mechanism. The study of this effect deserves further investigation.

In conclusion, our results demonstrate that the endothelial dysfunction occurring subsequent to I/R injury is caused by TNF-, which induces activation of xanthine oxidase and production of O 2 ·-. We speculate that many conditions with a moderate elevation of TNF-, eg, septicemia, acute myocardial infarction, chronic heart failure, atherosclerosis, viral myocarditis, and cardiac allograft rejection 18,30 may potentially mitigate the role of NO in coronary flow regulation and thus contribute to further deterioration of cardiac function. Selective modulation of xanthine oxidase signaling may provide a novel therapeutic target to prevent or alleviate coronary diseases associated with TNF- activation.

Acknowledgments

This study was supported by grants from American Heart Association grant-in-aid (0455435B), Atorvastatin Research Award (2004-37), American Heart Association SDG (110350047A) to Dr Zhang, National Institutes of Health grants COBRE (P20 RR18766) to Dr Lanier, and National Institutes of Health grants (HL32788, HL65203) to Dr Chilian. We thank Drs Arthur M. Feldman, Steven Lanier, Michael Davis, Jay D. Humphrey, and Mark Entman for their expert assistance.

【参考文献】
  Sack MN. Tumor necrosis factor- in cardiovascular biology and the potential role for anti-tumor necrosis factor- therapy in heart disease. Pharmacol Ther. 2002; 94: 123-135.

Squadrito F, Domenica A, Zingarelli B, et al. Tumor necrosis factor involvement in myocardial ischaemia-reperfusion injury. Eur J Pharmacol. 1993; 237: 223-230.

Meldrum DR. Tumor necrosis factor in the heart. Am J Physiol. 1998; 274 (3Pt 2): R577-R595.

Habib FM, Springall DR, Davies GJ, et al. Tumour necrosis factor and inducible nitric oxide synthase in dilated cardiomyopathy. Lancet. 1996; 347: 1151-1155.

Engel DPR, Armstong RC, Sivasubramanian N, Mann DL. Cardiac myocyte apoptosis provokes adverse cardiac remodeling in transgenic mice with targeted TNF overexpression. Am J Physiol Heart Circ Physiol. 2004; 287: H1303-H1311.

Hein TW, Zhang C, Wei Wang M, et al. Ischemia-reperfusion selectively impairs nitric oxide-mediated dilation in coronary arterioles: counteracting role of arginase. FASEB J. 2003; 17: 2328-2330.

Tsao PS, Ma X, Lefer AM. Activated neutrophils aggravate endothelial dysfunction after reperfusion of the ischemic feline myocardium. Am Heart J. 1992; 123: 1464-1471.

Yamamoto S, Matsui K, Itoh N, et al. The effect of an Na + /H + exchange inhibitor, SM-20550, on ischemia/reperfusion-induced endothelial dysfunction in isolated perfused rat hearts. Int J Tissue React. 2001; 23: 1-7.

Lee JJ. Olmos L. Vanhoutte PM. Recovery of endothelium-dependent relaxations four weeks after ischemia and progressive reperfusion in canine coronary arteries. Proc Assoc Am Phys. 1996; 108: 362-367.

Lefer AM, Lefer DJ. Pharmacology of the endothelium in ischemia-reperfusion and circulatory shock. Annu Rev Pharmacol Toxicol. 1993; 33: 71-90.

Downey JM, Omar B, Ooiwa H, et al. Superoxide dismutase therapy for myocardial ischemia. Free Rad Res Comms. 1991; 12-13:703-720.

Chandrasekar B, Smith JB, Freeman GL. Ischemia-reperfusion of rat myocardium activates nuclear factor- B and induces neutrophil infiltration via lipopolysaccharide-induced CXC chemokine. Circulation. 2001; 103: 2296-2302.

Salvemini D, Cuzzocrea S. Superoxide, superoxide dismutase and ischemic injury. Current opinion in investigational drugs. 2002; 3: 886-895.

Lefer AM, Ma X Cytokines and growth factors in endothelial dysfunction. Crit.Care Med. 1993; 21 (Suppl): S9-S14.

Hsueh WA, Quinones MJ Role of endothelial dysfunction in insulin resistance. Am J Cardiol. 2003; 92 (4A): 10J-17J.

Thielmann M, Dorge H, Martin C, et al. Myocardial dysfunction with coronary microembolization: signal transduction through a sequence of nitric oxide, tumor necrosis factor-alpha, and sphingosine. Circ Res. 2002; 90: 807-813.

Oral H, Dorn GW, Mann DL. Sphingosine mediates the immediate negative inotropic effects of tumor necrosis factor-a in the adult mammalian cardiac myocyte. J Biol Chem. 1997; 272: 4836-4842.

Meldrum DM, Cleveland JC, Cain BS, et al. Increased myocardial tumor necrosis factor- in a crystalloid-perfused model of cardiac ischemia-reperfusion injury. Ann Thorac Surg. 1998; 65: 439-443.

Mohanty JG JJ, Schulman ES, Raible DG. highly sensitive fluorescent micro-assay of H2O2 release from activated human leukocytes using a dihydroxyphenoxazine derivative. J Immunol Method. 1997; 202: 133-141.

Hein TW, Kuo L. LDLs impair vasomotor function of the coronary microcirculation: role of superoxide anions. Circ Res. 1998; 83: 404-414.

Tiefenbacher CP, Chilian WM, Mitchell M, et al. Restoration of endothelium-dependent vasodilation after reperfusion injury by tetrahydrobiopterin. Circulation. 1999; 94: 1423-1429.

Ammar RF Jr, Gutterman DD, Brooks LA, et al. Free radicals mediate endothelial dysfunction of coronary arterioles in diabetes. Cardiovasc Res. 2000; 47: 595-601.

Bagi Z, Koller. A., Kaley G. PPAR-g activation, by reducing oxidative stress, increases NO bioavailability in coronary arterioles of mice with type 2 diabetes. Am J Physiol Heart Circ Physiol. 2004; 286: H742-H748.

Cai D, Xaymardan M, Holm JM, et al. Age-associated impairment in TNF-alpha cardioprotection from myocardial infarction. Am J Physiol Heart Circ Physiol. 2003; 285: H463-H469.

Eddy LJ, Goeddel DV, Wong GH. Tumor necrosis factor-alpha pretreatment is protective in a rat model of myocardial ischemia-reperfusion injury. Biochem Biophys Res Commun. 1992; 184: 1056-1059.

Donnahoo KK, Meng X, Ao L, et al. Differential cellular immunolocalization of renal tumour necrosis factor-alpha production during ischaemia versus endotoxaemia. Immunology. 2001; 102: 53-58.

Shames BD, Barton HH, Reznikov LL, et al. Ischemia alone is sufficient to induce TNF-alpha mRNA and peptide in the myocardium. Shock. 2002; 17: 114-119.

Tsao PS, Aoki N, Lefer DJ, et al. Time course of endothelial dysfunction and myocardial injury during myocardial ischemia and reperfusion in the cat. Circulaiton. 1990; 82: 1402-1412.

Pereda J, Sabater L, Cassinello N, et al. Effect of simultaneous inhibition of TNF-alpha production and xanthine oxidase in experimental acute pancreatitis: the role of mitogen activated protein kinases. Ann Surg. 2004; 240: 108-116.

Lefer AM, Tsao P, Aoki N, et al. Mediation of cardioprotection by transforming growth factor-beta. Science. 1990; 249 (4964): 61-64.

Bagby GJ, Plessala KJ, Wilson LA, Thompson JJ, and Nelson S. Divergent efficacy of anti-TNFa antibody in intravascular and peritonitis models of sepsis. J Infect Dis. 1991; 163: 83-88.

Lang CH, Obih J-CA, Bagby GJ, Bagwell JN, and Spitzer JJ. Endotoxin-induced increases in regional glucose utilization by small intestine: A TNF-independent effect. Am J Physiol Gastrointest Liver Physiol. 1991; 260: G548-G555.

Bagby GJ, Lang CH, Skrepnik N, Golightly G, and Spitzer JJ. Regulation of glucose metabolism after endotoxin and during infection is largely independent of endogenous tumor necrosis factor. Circ Shock. 1993; 39: 211-219.

Xie J, Joseph KO, Bagby GJ, Giles TD, and Greenberg SS. Dissociation of TNF-alpha from endotoxin-induced nitric oxide and acute-phase hypotension. Am J Physiol. 1997; 273: H164-H174.

Xie J, Kolls J, Bagby GJ, and Greenberg SS. Independent suppression of nitric oxide and TNFa in the lung of conscious rats by ethanol. FASEB J. 1995; 9: 253-261.

Moller DE. New drug targets for type 2 diabetes and the metabolic syndrome. Nature. 2001; 414: 821-827.

作者单位：Departments of Anethesiology, Surgery, and Physiology (C.Z., X.X.), LSU Health Sciences Center, New Orleans, La; the Department of Medical Physiology (W.W., L.K.), College of Medicine, Texas A&M University System Health Science Center, College Station, Tex; the Section of Cardiovascular Sciences