Adenosine Inhibits Lipopolysaccharide-Induced Secretion of Tumor Necrosis Factor-α in the Failing Human Heart
Background—The proinflammatory cytokine tumor necrosis factor-α (TNF-α) has been implicated in the pathogenesis of congestive heart failure. Recent studies have shown that adenosine inhibits lipopolysaccharide (LPS)-induced expression of TNF-α in macrophages and rat cardiomyocytes. The aim of this study was to determine whether adenosine has a similar effect in the failing human heart.
Methods and Results—Left ventricular muscle strips were obtained from seven patients with end-stage congestive heart failure undergoing heart transplantation or insertion of a left ventricular assist device. The muscle strips were incubated at 37°C in 95% O2/5% CO2 and stimulated with LPS (10 μg/mL). TNF-α release in the supernatant was measured with ELISA, and muscle sections were stained for TNF-α. Muscle strips released TNF-α in the absence of LPS (0.22±0.05 pg · mL−1 · mg wet wt−1). TNF-α was immunolocalized to the cardiac myocyte, suggesting that the myocyte is a source for TNF-α production. Adenosine (10 μmol/L) decreased TNF-α by 40% (P<.05). The selective adenosine A2 receptor agonist DPMA (10 μmol/L) decreased TNF-α release by 87% (P<.001), whereas ITu (10 μmol/L), an adenosine-regulating agent that increases endogenous adenosine concentration, inhibited TNF-α release by 93% (P<.001).
Conclusions—Adenosine can significantly diminish TNF levels in the failing human heart and may represent a new pharmacological intervention in congestive heart failure.
Clinical and basic research studies support the hypothesis that the proinflammatory cytokine TNF-α plays an important role in the development of dilated cardiomyopathy and congestive heart failure: (1) patients with end-stage heart failure have elevated levels of TNF-α1 ; (2) TNF-α has direct negative inotropic effects on the myocardium2 ; (3) myocytes from failing human hearts re-express TNF-α3 ; (4) elevations in left ventricular filling pressures can stimulate TNF-α expression4 ; and (5) transgenic mice overexpressing TNF-α demonstrate interstitial fibrosis, interstitial infiltrates, adrenergic desensitization, and ventricular dilatation.5 However, the biochemical mechanisms responsible for modulating myocardial TNF-α levels remain undefined.
Recently, studies have demonstrated that the phosphodiesterase inhibitor vesnarinone,6 7 the antiarrhythmic agent amiodarone,8 and the cardiac glycoside ouabain9 inhibit the production of TNF-α by human mononuclear cells after challenge with LPS. A similar cytokine inhibitory effect can be found with the naturally produced nucleoside adenosine.10 Importantly, adenosine has also been shown to inhibit LPS-stimulated cytokine production by rodent myocardium, an effect that appeared to be mediated through the adenosine A2 receptor.11 However, rodent hearts are often not representative of human myocardium. Therefore, the present study was undertaken to assess the ability of adenosine to regulate cytokine production by failing human heart.
Human Trabecular Muscles
Human heart tissue was isolated from seven patients with end-stage cardiomyopathy at the time of transplantation or during insertion of an LVAD. The heart tissue was transported at 4°C in St Thomas cardioplegic solution and immediately cut into ≈2×1×1-mm strips. The muscle strips were incubated in DMEM/F12 medium containing 5% horse serum and equilibrated for 1 hour at 37°C before experiments were begun. The experiments were performed in 12-well plates at 37°C in 95% O2/5% CO2. The wet weight of muscle strips was determined at the end of the experiments.
Measurement of TNF-α
LPS Escherichia coli 0127 (Sigma Chemical Co) (10 μg/mL) was used to induce production of TNF-α in muscle strips as previously described.11 At times 0 and 4 hours after addition of LPS, the supernatants were collected, frozen in liquid nitrogen, and stored at −70°C until analysis. TNF-α in the supernatants was measured with a human TNF-α ELISA kit (R&D Systems). The accuracy of this ELISA kit was verified by repeat measurements with a kit from a different company (Genzyme). Both kits use a mouse monoclonal anti–TNF-α, and they provided comparable measurements. To lower the limit of detection to 1 pg/mL, all samples were concentrated through centricon 10 concentrators (Amicon) as previously described.11 A standard curve was generated with each set of samples assayed. Linear regression analysis of the standard curves yielded a correlation coefficient of >.99.
Immunohistochemical staining of trabecular muscles was performed as previously described.11 Tissue sections were treated with chicken anti-human TNF-α (Promega) and rabbit anti-chicken secondary antibody (Jackson Laboratories). Avidin-biotin complex (Vector Laboratories) was added, and visualization of the reaction was achieved by treatment with 3-amino-9-ethyl cabazole and 0.03% H2O2 in 0.1 mol/L acetate buffer, pH 5.2. Sections were weakly counterstained with hematoxylin.
Results are expressed as mean±SEM of determinations in muscle strips of seven patients. All experiments were performed in triplicate. Data were subjected to one-way ANOVA (Fisher test), and P<.05 was considered statistically significant.
All seven patients were male, and their mean age was 50±7 years (14 to 67 years). Three patients had ischemic cardiomyopathy and four, dilated cardiomyopathy. All patients were in NYHA functional class IV and had a similar medical regimen consisting of diuretics and intravenous inotropic agents. Six patients were receiving dobutamine and two, dobutamine and milrinone. One patient had an LVAD. Four patients underwent orthotopic heart transplantation. Two patients received a biventricular assist device and one, an LVAD.
Effect of Adenosine on TNF-α Release
Trabecular muscles from failing human hearts released TNF-α in the absence of LPS (0.22±0.05 pg · mL−1 · mg wet wt−1) (n=7). Furthermore, TNF-α was immunolocalized to the cardiac myocyte, suggesting that the myocyte was a source for TNF-α production (Fig 1⇓). As seen in Fig 2⇓, adenosine 10 μmol/L decreased TNF-α by 40% (P<.05). Similarly, the selective adenosine A2 receptor agonist DPMA 10 μmol/L decreased TNF-α release by 87% (P<.001), whereas ITu, an adenosine-regulating agent that increases endogenous adenosine concentration, inhibited TNF-α by 93% (P<.001).
Interestingly, the response to LPS challenge was more pronounced for muscle strips obtained from patients with ischemic cardiomyopathy (5.09±0.72 pg · mL−1 · mg wt−1, n=3) than for muscle strips from patients with dilated cardiomyopathy (1.2±0.29 pg · mL−1 · mg wt−1, n=4) (P<.001). There was no significant difference in age (52 versus 49 years) or in medical regimen between the two groups. In addition, adenosine had a significant effect (P<.005) on TNF-α release in the ischemic cardiomyopathy group but did not alter TNF-α release in the dilated cardiomyopathy group. By contrast, DPMA and ITu had a significant effect on TNF-α in both the ischemic and nonischemic groups.
In the present study, thin slices of myocardium from failing human heart expressed measurable levels of the proinflammatory cytokine TNF-α. This was not surprising, because earlier studies by Torre-Amione et al3 demonstrated that failing but not nonfailing human heart re-expressed abundant quantities of TNF-α. However, the relatively modest level of TNF-α secretion by the tissue slices could be enhanced 10-fold by the addition of LPS to the incubation medium. The LPS-stimulated TNF-α expression could be attenuated by adenosine and nearly abrogated by the addition of the A2 receptor agonist DPMA or the adenosine kinase inhibitor ITu. The ability of adenosine to inhibit LPS-stimulated TNF-α secretion is consistent with earlier studies in the rat in which adenosine inhibited myocardial TNF-α production, with maximal effect being observed with adenosine A2-selective agonists.11 However, in contrast to failing human heart, normal rodent myocardium does not express TNF-α in the absence of LPS or other stimuli.
Because adenosine was given simultaneously with LPS, it is undetermined whether adenosine can inhibit TNF-α production once it is stimulated. In fact, we observed in neonatal rat cardiomyocytes that adenosine is effective only when given during the time period from 1 hour before to 1 hour after LPS challenge.11 Adenosine was no longer effective when given 3 hours after LPS. This may limit the clinical usefulness of adenosine. It is known that milrinone inhibits TNF-α secretion from human mononuclear cells.12 In our study, there was no significant effect of milrinone therapy on the secretion of TNF-α.
Immunolocalization of TNF-α to the cardiac myocyte shows that the myocyte was a source of TNF-α production. These results in the human heart are also consistent with studies by Kapadia and colleagues4 demonstrating that stretch-induced TNF-α production is largely localized to the myocytes. However, it should be noted that all seven patients had NYHA class IV symptoms and were receiving inotropic support before cardiac transplantation or LVAD insertion. Therefore, we cannot exclude the possibility that myocardium isolated from patients with mild to moderate heart failure might not express TNF-α either at baseline or in response to LPS. Nevertheless, our results suggest that elevated serum levels of TNF-α that have been observed in patients with end-stage heart failure1 may be due in part to myocardial secretion of TNF-α.
It was interesting to note that muscle strips from patients with ischemic heart disease produced substantially more TNF-α in response to LPS than did strips from patients with dilated cardiomyopathy. By contrast, baseline levels were similar in both groups. The difference between the two groups is difficult to explain. It is possible that in patients with ischemic cardiomyopathy, some muscle samples were obtained from ischemic myocardium. This may have contributed to an enhanced secretion of TNF-α, because TNF-α mRNA is increased in ischemic myocardium.13 Greater fibroblast proliferation in the ischemic heart may also have contributed to the observed difference between the two groups. Adenosine had a significant effect on TNF-α release in the ischemic cardiomyopathy group but did not alter TNF-α release in the dilated cardiomyopathy group. By contrast, DPMA and ITu had a significant effect on TNF-α in both groups. It could be speculated that the diminished effect of adenosine in dilated cardiomyopathy was due to desensitization or downregulation of the adenosine A2 receptor. However, because of the small sample size, the significance of these findings is uncertain.
In conclusion, these in vitro experiments suggest that adenosine can significantly diminish TNF-α levels in myocardial tissue obtained from failing human heart. If confirmed in vivo, adenosine and/or adenosine-regulating agents may provide novel pharmacological strategies for the treatment of patients with end-stage congestive failure.
Selected Abbreviations and Acronyms
|LVAD||=||left ventricular assist device|
|TNF||=||tumor necrosis factor|
Part of this work was done during the tenure of a fellowship from the American Heart Association, Nation’s Capital Affiliate Inc (Dr Wagner). We are thankful to the members of the Division of Cardiothoracic Surgery of the University of Pittsburgh for providing the heart muscle samples.
- Received October 3, 1997.
- Revision received December 12, 1997.
- Accepted December 12, 1997.
- Copyright © 1998 by American Heart Association
Eichenholz P, Eichacker PQ, Hoffman WD, Banks SM, Parrilo JE, Danner RT, Natanson C. Tumor necrosis factor challenges in canines: patterns of cardiovascular dysfunction. Am J Physiol. 1992;263:H668–H675.
Torre-Amione G, Kapadia S, Lee J, Durand JB, Bies RD, Young JB, Mann DL. Tumor necrosis factor-α and tumor necrosis factor receptors in the failing human heart. Circulation. 1996;93:704–711.
Kapadia SR, Oral H, Lee J, Nakano M, Taffet GE, Mann DL. Hemodynamic regulation of tumor necrosis factor-α gene and protein expression in adult feline myocardium. Circ Res. 1997;81:187–195.
Kubota T, McTiernan CF, Frye CS, Slawson SE, Koretsky AP, Demetris AJ, Feldman AM. Dilated cardiomyopathy in transgenic mice with cardiac-specific overexpression of tumor necrosis factor-alpha. Circ Res. 1997;81:627–635.
Matsui S, Matsumori A, Matoba Y, Uchida A, Sasayama S. Treatment of virus-induced myocardial injury with a novel immunomodulating agent, vesnarinone: suppression of natural killer cell activity and tumor necrosis factor-α production. J Clin Invest. 1994;94:1212–1217.
Matsumori A, Shioi T, Yamada T, Matsui S, Sasayama S. Vesnarinone, a new inotropic agent, inhibits cytokine production by stimulated human blood from patients with heart failure. Circulation. 1994;89:955–958.
Matsumori A, Ono K, Nishio R, Nose Y, Sasayama S. Amiodarone inhibits production of tumor necrosis factor-α by human mononuclear cells: a possible mechanism for its effect in heart failure. Circulation. 1997;96:1386–1389.
Matsumori A, Ono K, Nishio R, Igata H, Shioi T, Matsui S, Furukawa Y, Iwasaki A, Nose Y, Sasayama S. Modulation of cytokine production and protection against lethal endotoxemia by the cardiac glycoside ouabain. Circulation. 1997;96:1501–1506.
Parmely MJ, Zhou WW, Edwards CK III, Borcherding DK, Silverstein R, Morrison DC. Adenosine and a related carbocyclic nucleoside analogue selectively inhibit tumor necrosis factor-α production and protect mice against endotoxin challenge. J Immunol. 1993;151:389–396.
Wagner DR, Combes A, McTiernan C, Sanders VJ, Lemster B, Feldman AM. Adenosine inhibits lipopolysaccharide-induced cardiac expression of tumor necrosis factor-α. Circ Res. 1998;82:47–56.