INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

PROINFLAMMATORY CYTOKINES are key mediators of cellular damage in immune and inflammatory responses. Only recently have proinflammatory cytokines been found to play a role in the myocardium. Tumor necrosis factor (TNF)- was the first cytokine to be associated with the development of congestive heart failure (8, 11, 13, 21). More recently, clinical and basic studies have suggested that interleukin (IL)-6 and IL-1 may also be involved in the development of congestive heart failure. The interest in IL-6 has been prompted by the fact that IL-6, like TNF-, is cardiodepressive and elevated in patients with congestive heart failure (3, 9, 15, 19, 20, 22). Moreover, there is evidence from a subgroup analysis of a multicenter trial (15) suggesting that a mortality benefit was associated with reduced IL-6. Similarly to TNF- and IL-6, IL-1 is cardiodepressive and elevated in patients with congestive heart failure (3, 6, 19). In addition, like TNF-, IL-1 has been found to be expressed in the myocardium of patients with dilated cardiomyopathy (4).

Little is known regarding the molecular mechanisms that regulate myocardial expression of proinflammatory cytokines. It is known that adenosine, a breakdown product of ATP, inhibits the release of TNF- and IL-6 by human monocytes through activation of the A₂ receptor (1). We have recently shown that the A₂ receptor also mediates the suppressing effect of adenosine on TNF- release by isolated rat cardiomyocytes and muscle strips obtained from patients with end-stage heart failure (23, 24). It is not known whether adenosine affects other proinflammatory cytokines in the heart. Furthermore, it is unclear whether adenosine receptors other than the A₂ receptor are involved in the regulation of proinflammatory cytokines in the heart. Therefore, we investigated the effect of adenosine and its three cardiac receptors (10) on the expression of IL-6 and IL-1 in isolated cardiomyocytes, rat papillary muscles, and trabecular muscles obtained from human patients with advanced congestive heart failure.

The results of the present study suggest that, in isolated cardiomyocytes, adenosine inhibits the expression of TNF- through activation of the A₂ receptor but induces robust expression of IL-6 through activation of the A₃ and possibly the A₁ receptor. In the failing human heart, however, adenosine inhibits the expression of TNF- without affecting the expression of IL-6. Immunohistochemistry showed that the differential effect between myocytes and the failing human heart was primarily due to the fact that inflammatory cells, which respond differently to adenosine, were the major source of IL-6 in the failing human heart.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Neonatal rat cardiomyocytes. Cardiomyocytes were prepared from ventricles of 1-day-old Sprague-Dawley rats as previously described (23) using a commercially available cardiomyocyte isolation kit (Worthington Biochemical). Cells recovered after trypsin and collagenase digestion were preplated on untreated plastic flasks for 1 h to reduce nonmyocyte cell numbers. Nonadherent cells enriched for cardiomyocytes were cultured in DMEM/F-12 medium containing 5% horse serum, 1 mM glutamine, 10 mM HEPES, 0.1 mM 5-bromo-2'-deoxyuridine, 5 µg/ml insulin, 5 ng/ml selenium, 5 µg/ml transferrin, and 10 µg/ml gentamicin. Horse serum was treated with AG1X-10 (Pharmacia) resin as previously described (18) to reduce serum 3,3',5'-triiodo-L-thyronine to undetectable levels as determined by radioimmunoassay. Cells were plated on pronectin (Promega)-coated tissue culture plates at a density of 1 × 10⁵ cells/cm² and grown at 37°C in 95% O₂-5% CO₂. Some experiments were performed under hypoxia (addition of 100% N₂) to simulate ischemia. Cells were cultured for 48 h before experiments were started. We have previously shown (23) that these preparations contain 93-95% cardiomyocytes and <0.1% white blood cells.

Rat papillary muscles. Rat papillary muscles were isolated from female rats (~250 g) and immediately cut into ~2 × 1 × 1-mm strips as previously described (23). The muscle strips were incubated in DMEM/F-12 medium containing 5% horse serum, 1 mM glutamine, 10 mM HEPES, 5 µg/ml insulin, 5 ng/ml selenium, 5 µg/ml transferrin, and 10 µg/ml gentamicin. Muscle strips were equilibrated for 1 h at 37°C in 95% O₂-5% CO₂ before experiments were started. During the experiment, muscle strips were kept in commercially available 12-well plates at 37°C in 95% O₂-5% CO₂.

Human trabecular muscles. Human heart tissue was isolated from six patients with end-stage cardiomyopathy at the time of transplantation or during insertion of a left ventricular assist device as previously described (24). Clinical characteristics of the six patients have been reported elsewhere (24). 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/F-12 medium containing 5% horse serum, 1 mM glutamine, 10 mM HEPES, 5 µg/ml insulin, 5 ng/ml selenium, 5 µg/ml transferrin, and 10 µg/ml gentamicin. Muscle strips were equilibrated for 1 h at 37°C in 95% O₂-5% CO₂ before experiments were started. During the experiment, muscle strips were kept in commercially available 12-well plates at 37°C in 95% O₂-5% CO₂. Five muscle strips were placed in each well.

Stimulation with lipopolysaccharide. As previously described (23, 24), exposure to lipopolysaccharide (LPS) Escherichia coli 0127 (Sigma) was used to induce production of proinflammatory cytokines in neonatal rat cardiomyocytes, rat papillary muscles, and human muscle strips. Neonatal rat cardiomyocytes were exposed to LPS (10 ng/ml, 4-h incubation) after 48 h in culture. A higher dose of LPS (10 µg/ml, 4-h incubation) was used to induce cytokine production in rat papillary muscles and human heart muscle strips.

Treatment with adenosine. To assess the effect of adenosine, we used adenosine (1-10 µM; Sigma) as well as the selective A₁-receptor agonist N⁶-cyclopentyladenosine (CPA; 10 nM-10 µM), the selective A₂-receptor agonist N⁶-[2-(3,5-dimethoxyphenyl)-2-(2-methylphenyl)-ethyl]adenosine (DPMA; 10 nM-10 µM), the selective A₃-receptor agonist N⁶-benzyl-5'-(N-ethylcarboxamido)adenosine (N⁶-benzyl-NECA; 10 nM-10 µM), the selective A₁-receptor antagonist 8-cyclopentyl-1,3-dipropylxanthine (DPCPX; 10 µM), the selective A₂-receptor antagonist 8-(3-chlorostyryl)caffeine (CSC; 10 µM), or the selective A₃-receptor antagonist 3-ethyl-5-benzyl-2-methyl-4-phenylethynyl-6-phenyl-1,4-(±)-dihydropyridine-3,5-dicarboxylate (MRS-1191; 0.1 nM-10 µM) (all from Research Biochemical International), as previously described (23, 24). All compounds were added at the same time as the LPS.

Release of IL-1 and IL-6. At time 0 and 4 h after the addition of LPS E. coli 0127, the supernatants were collected, immediately frozen in liquid N₂, and stored at 70°C until analysis. The levels of IL-1 and IL-6 in the supernatants were measured with enzyme-linked immunosorbent assay (ELISA) kits (R&D Systems). To detect low levels of IL-1 and IL-6, all samples were concentrated through Centricon 10 concentrators (Amicon) as previously described (23, 24). Recovery was equal for all measured samples.

RNase protection assay. Total RNA was isolated from neonatal cardiomyocytes and human tissue samples by an acid guanidinium thiocyanate-phenol-chloroform extraction method (2) and quantified by spectrophotometry. The human tissue samples were homogenized using a Polytron for 30-60 s before phenol-chloroform extraction. A multiprobe RNase protection assay system (RiboQuant, Pharmingen) was used to detect and quantify transcript levels of selected cytokines. Briefly, RNA (10 µg) was hybridized with multiple radiolabeled RNA probes overnight, treated with RNase, and electrophoresed on a 5% polyacrylamide gel. Radioactive images from hybridization were obtained with a PhosphorImager (Molecular Dynamics) and quantified with Image-Quant software (Molecular Dynamics). A value for each transcript was normalized to that for glyceraldehyde-3-phosphate dehydrogenase to correct for differences in RNA mass. Data were in turn normalized to the mean of the control samples, arbitrarily set at 100%.

Immunohistochemistry. Muscle sections obtained from patients with cardiomyopathy were frozen in optimal cutting temperature medium. Blocks were cut on a cryostat at 10 µm, and sections were mounted on Superfrost Plus slides (Fisher Scientific). Sections were immersion fixed in 95% ethanol, rinsed in phosphate-buffered saline (PBS), and treated for 30 min with 5% normal goat serum. Primary antibody, polyclonal rabbit anti-human IL-6 (Genzyme), was diluted 1:100 to achieve a final concentration of 10 µg/ml. Samples were treated with primary antibody for 24 h at 4°C. Sections were rinsed briefly with PBS and then treated with a 1:200 dilution of biotinylated goat anti-rabbit secondary antibody (Jackson Labs) for 2 h, followed by PBS rinse and avidin-biotin complex (Vector Laboratories) for 45 min. Visualization of the reaction was achieved by treatment with 3-amino-9-ethyl-carbazole in 0.1 M acetate buffer, pH 5.2. Sections were weakly counterstained with Mayer's hematoxylin. Double-immunofluorescent detection was performed using anti-IL-6 antibody and the lectin RCA-1 (Ricinus communis agglutin), which labels macrophages and endothelial cells. Staining was performed as described above, except that fluorescent-labeled secondary antibodies (Jackson ImmunoResearch) were used at a 1:100 dilution. The immunofluorescent samples were analyzed with an argon/krypton laser scanning confocal microscope (Molecular Dynamics). Images were collected and analyzed by using both Image-1 (Molecular Dynamics) and operating system 5.3 software (SiliconGraphics).

Statistical analysis. Results are expressed as means ± SE. Data were subjected to analysis of variance (one-way ANOVA, Fisher test), and a value of P < 0.05 was considered to be statistically significant.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Expression of IL-6 in rat cardiomyocytes and rat papillary muscles. Neonatal rat cardiomyocytes released large amounts of IL-6 into the medium under control conditions. The IL-6 concentration reached 53.8 ± 7.3 pg/ml at 4 h (n = 4) (Fig. 1A). The addition of LPS induced a significant increase in IL-6 (P < 0.05). Unexpectedly, adenosine did not inhibit IL-6 release but, on the contrary, induced the release of IL-6. The effect of adenosine on IL-6 was additive to the effect of LPS (P < 0.005). Adenosine also induced the release of IL-6 in the absence of LPS (P < 0.001). The effect of adenosine was not dose dependent, with 1 µM inducing a 1.9-fold and 10 µM inducing a 2.2-fold increase in IL-6 (P = not significant). Under hypoxic conditions (addition of N₂), myocyte release of IL-6 was increased by 50% (P < 0.05) compared with that under control conditions. Cells remained morphologically intact but ceased beating during hypoxia.

View larger version (43K): [in this window] [in a new window]   Fig. 1.   Release of interleukin (IL)-6 by neonatal rat cardiomyocytes (NCM). A: effect of adenosine. NCM were incubated for 4 h with diluent (Con), 1 or 10 µM adenosine (Ado), 10 ng/ml lipopolysaccharide (LPS), or LPS + Ado. Supernatant was collected, and IL-6 was measured with ELISA (n = 4 experiments). * P < 0.001 vs. Con; ** P < 0.005 vs. LPS. B: effect of adenosine-receptor agonists. NCM were incubated with selective A1 agonist N6-cyclopentyladenosine (CPA; 10 nM) (A1), selective A2 agonist N6-[2-(3,5-dimethoxyphenyl)-2-(2-methylphenyl)-ethyl]adenosine (DPMA; 10 nM) (A2), or selective A3 agonist N6-benzyl-5'-(N-ethylcarboxamido)adenosine (10 nM) (A3) (n = 4 experiments). * P < 0.005 vs. Con. C: effect of adenosine-receptor antagonists. NCM were incubated with 10 µM Ado or Ado plus selective A1 antagonist 8-cyclopentyl-1,3-dipropylxanthine (DPCPX; 10 µM) (A1 antag.), selective A2 antagonist 8-(3-chlorostyryl)caffeine (CSC; 10 µM) (A2 antag.), or selective A3 antagonist 3-ethyl-5-benzyl-2-methyl-4-phenylethynyl-6-phenyl-1,4-(±)-dihydropyridine-3,5-dicarboxylate (MRS-1191; 10 µM) (A3 antag.) (n = 4). * P < 0.01; ** P < 0.001 vs. Con.

View larger version (93K): [in this window] [in a new window]   Fig. 2.   Effect of Ado on IL-6 transcript levels in NCM. A: representative RNase protection assay. TNF-, tumor necrosis factor-; GAPDH, glyceraldehyde-3-phosphate dehydrogenase. B: quantitative analysis. Transcript levels were measured with RNase protection and corrected for GAPDH. Control was taken as 100%; dosages were 10 ng/ml LPS and 10 µM Ado (n = 5). * P < 0.05 vs. Con.

View larger version (11K): [in this window] [in a new window]   Fig. 3.   Effect of Ado on release of IL-6 by rat papillary muscles (RPM). RPM were isolated and incubated for 4 h with 10 µM Ado or 10 ng/ml LPS. Supernatant was collected, and IL-6 was measured with ELISA (n = 6). * P < 0.05; ** P < 0.005 vs. Con.

Expression of IL-6 in the failing human heart. The human trabecular muscles released high amounts of IL-6 into the medium under control conditions (Fig. 4A). The level of IL-6 reached 1,565 ± 311 pg/ml after 4 h (n = 6). The addition of LPS had no significant effect on the release of IL-6 by the human trabecular muscles. In contrast with isolated rat cardiomyocytes and rat papillary muscles, adenosine had no effect on IL-6 in the failing human heart.

View larger version (53K): [in this window] [in a new window]   Fig. 4.   Effect of Ado on IL-6 in failing human heart. A: release of IL-6 (left) and IL-6 transcript levels (right). Heart muscle strips obtained from patients with end-stage congestive heart failure at time of transplantation or insertion of a left ventricular assist device (LVAD) were incubated in presence of diluent (Con), 10 µg/ml LPS, or LPS + 10 µM Ado. Left: supernatant was collected after 4 h, and IL-6 was measured with ELISA (n = 6). Right: transcript levels were measured with RNase protection and corrected for GAPDH. Control was taken as 100%. * P < 0.05 vs. Con. B: representative RNase protection assay.

View larger version (131K): [in this window] [in a new window]   Fig. 5.   Muscle sections of a cardiomyopathic left ventricle stained with anti-IL-6 (red). Note positive staining for IL-6 in perivascular aggregates: positive staining for IL-6 (A) and negative control (B) are shown at magnification ×200; positive staining for IL-6 is also shown at higher magnification ×400 (C and D).

View larger version (48K): [in this window] [in a new window]   Fig. 6.   Confocal micrograph showing double immunolabeling of cardiomyopathic left ventricle with anti-IL-6 (green) and lectin RCA-1 (Ricinus communis agglutin; red), which identifies macrophages and endothelial cells. Note prominent staining for IL-6 in perivascular aggregates and faint positive staining in interstitium (A and B).

IL-1 expression in rat cardiomyocytes, rat papillary muscles, and failing human heart. Neonatal myocytes did not release significant amounts of IL-1 under control conditions (Table 1; n = 5), and treatment with LPS did not increase the release of IL-1. Adenosine (1-10 µM), the selective A₁ agonist CPA, the selective A₂ agonist DPMA, or the selective A₃ agonist N⁶-benzyl-NECA had no significant effect on IL-1 in the presence or absence of LPS. At the transcriptional level, the expression of IL-1 was comparable to the expression of IL-6 (Fig. 2B). However, in contrast to their effects on IL-6, both LPS and adenosine had no significant effect on IL-1 expression (Fig. 4B and Table 1; n = 5). Similarly, LPS and adenosine had no effect on the expression of IL-1 in rat papillary muscles (data not shown).

View this table: [in this window] [in a new window]   Table 1.   Effect of adenosine on IL-1 release and transcript levels in neonatal cardiomyocytes and the failing human heart

TNF- expression in rat cardiomyocytes and the failing human heart. To confirm that LPS exposure induced TNF- expression in the present experiments and that this reexpression was comparable to that seen in earlier experiments (23, 24), we assessed the effects of LPS on TNF- mRNA levels. In neonatal rat cardiomyocytes, baseline transcript levels of TNF- were undetectable. These results were consistent with the previous observation that neonatal cardiomyocytes do not release measurable amounts of TNF- in the absence of LPS (23). However, transcript levels of TNF- increased 16-fold after the addition of LPS (P < 0.05) (Fig. 7). Adenosine (10 µM), added at the same time as LPS, inhibited the expression of TNF- in neonatal myocytes by 40% (P < 0.05). Similarly, the selective adenosine A₂-receptor agonist DPMA (10 nM) inhibited the expression of TNF- by 37% (P < 0.05) (Fig. 7; n = 5).

View larger version (13K): [in this window] [in a new window]   Fig. 7.   Effect of Ado A2-receptor agonist DPMA on TNF- transcript levels in NCM (A) and failing human heart (B). A: NCM were incubated with diluent (Con), 10 ng/ml LPS, or LPS + DPMA (LPS + A2). Transcript levels were measured with RNase protection and corrected for GAPDH. Control was taken as 100% (n = 5). * P < 0.05 vs. Con. B: heart muscle strips were obtained from patients with end-stage congestive heart failure at time of transplantation or insertion of an LVAD and incubated with diluent (Con), 10 µg/ml LPS, or LPS + A2. Transcript levels were measured with RNase protection and corrected for GAPDH. Control was taken as 100% (n = 6). * P < 0.05 vs. Con.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

The results of the present study suggest that adenosine receptors have a differential effect on the expression of cytokines in the rodent heart: the adenosine A₂ receptor suppresses the expression of TNF-, and the adenosine A₃ receptor induces the expression of IL-6. The differential effect of adenosine on myocardial cytokine expression was demonstrated in isolated cardiomyocytes and rodent papillary muscle preparations. However, experiments with trabecular muscles from patients with end-stage cardiomyopathy demonstrated that significant levels of cardiac IL-6 were due to inflammatory cells in the interstitial space and that the IL-6 expression was not responding to adenosine.

We have previously shown (23) that neonatal cardiomyocytes do not release significant amounts of TNF- in the absence of LPS. Accordingly, the present study did not detect TNF- mRNA in myocytes under control conditions with the use of an RNase protection assay. However, myocytes expressed IL-1 and IL-6 under control conditions and continuously released significant levels of IL-6 into the medium. This was consistent with reports by others (25). By contrast, the myocytes did not release significant amounts of IL-1.

As expected, the addition of LPS induced a 16-fold increase in TNF- transcript levels in neonatal cardiomyocytes (P < 0.05). Adenosine and the adenosine A₂-receptor agonist DPMA inhibited the expression of TNF- by 40% (P < 0.05). The effect of adenosine and DPMA on TNF- transcript levels was comparable with the effect on TNF- protein levels reported previously (23). This confirmed the importance of the A₂ receptor in mediating the TNF- suppressing effect of adenosine in cardiomyocytes.

However, while decreasing TNF-, adenosine significantly increased the expression of IL-6 in neonatal myocytes. Adenosine induced a 2.2-fold increase in IL-6 release (P < 0.001) and a 4.7-fold increase in IL-6 transcript levels (P < 0.05). The effect of adenosine was not concentration dependent at micromolar concentrations but was additive to the effect of LPS (P < 0.01). The effect of adenosine on IL-6 release could be replicated by the selective adenosine A₃ agonist N⁶-benzyl-NECA, but not by the A₂ agonist DPMA or the A₁ agonist CPA, indicating that the effect was likely mediated through the A₃ receptor. The involvement of the A₃ receptor was confirmed using selective adenosine antagonists. At nanomolar concentrations, only the A₃-receptor antagonist MRS-1191 was able to completely inhibit the effect of adenosine on IL-6. The A₂ antagonist CSC had no effect on the adenosine-induced release of IL-6, excluding a significant participation of this receptor. The A₁ antagonist DPCPX had a significant inhibiting effect on the adenosine-mediated release of IL-6 when micromolar concentrations were used. Therefore, some additional participation of the A₁ receptor cannot be excluded.

Using inhibitors of adenosine transport and adenosine kinase, we have previously shown that locally produced adenosine has the same effect on TNF- expression in cardiomyocytes as exogenously added adenosine (23, 24). In the present study, local production of adenosine was induced by hypoxia (simulated ischemia). As expected, hypoxia had an effect on IL-6 similar to that of added adenosine.

The demonstration that adenosine induces the cardiac expression of IL-6 may be of clinical relevance, because adenosine is continuously released during myocardial ischemia and in congestive heart failure (5, 16) and because IL-6 appears to play a pathophysiological role in the development of congestive heart failure (7, 12, 15, 19, 20, 22). IL-6 has been found to exert negative inotropic effects on the hamster papillary muscle and chick ventricular myocytes (3, 9). Hypoxia, the calcium ionophore ionomycin, IL-1, and catecholamines have all been shown (25) to induce the release of IL-6 by neonatal cardiomyocytes in culture. Interestingly, adenosine has been shown to limit IL-6 release in human monocytes (1). However, adenosine has also been shown to induce the release of IL-6 from rat adrenal zona glomerulosa cells, ovarian cells, and anterior pituitary cells through stimulation of the adenosine A₂ receptor (17). Thus there is a high degree of tissue specificity in the cellular response to selective adenosine agonists.

The present study identifies the A₃ receptor as an important mediator of cardiac IL-6 expression. Until the recent work of Liang and Jacobson (14), the physiological role of the cardiac A₃ receptor was poorly defined. Using the selective A₃ antagonist MRS-1191, they showed that the A₃ receptor is implicated in mediating sustained cardioprotection in a myocyte model of ischemia. It remains to be determined how the A₃ receptor can mediate cardioprotection and, at the same time, induce the expression of a cardiodepressive cytokine. The effects of adenosine on IL-6 expression were cytokine specific because neither adenosine nor specific adenosine analogs were able to influence IL-1 expression.

Compared with neonatal myocytes, the heart muscle strips obtained from patients with end-stage heart failure expressed ~20-fold higher levels of TNF- and IL-6. The high baseline expression of TNF- and IL-6 in the failing human heart confirms reports by others (7, 21) and further substantiates the observation that proinflammatory cytokines are expressed in the failing human heart. However, the difference in the observed responses obtained with rat myocytes and human myocardium may be due not only to differences in pathophysiological state (failing vs. nonfailing) but also to a species-related difference. Similarly to its effect in cardiomyocytes, the A₂ agonist DPMA reduced the expression of TNF- in the human heart by >50% (P < 0.05). This confirms the important role of the A₂ receptor in suppressing TNF- in the failing human heart.

In contrast to the effect seen in cardiomyocytes, adenosine did not induce the expression of IL-6 in human muscle strips. However, experiments using rat papillary muscles showed that adenosine induces IL-6 in the intact myocardium. This disparity could be explained by a species-specific difference between rodent and human heart or, alternatively, by heart failure-related changes in human myocardial A₃ receptors. However, immunohistochemical studies localized the IL-6 expression in the failing human heart to perivascular aggregates consisting of macrophages, with less abundant expression in endothelial cells and myocytes. Because adenosine limits the release of IL-6 by human monocytes (1), a stimulatory effect of adenosine on IL-6 in human myocytes might be balanced by an inhibitory effect of adenosine on IL-6 in extracellular matrix macrophages. Thus the presence, or absence, of macrophages and other inflammatory cells appears critical for the overall modulation of cytokines by adenosine in the intact human myocardium. Whether other molecules have a similar differential effect on IL-6 in both cell types remains to be determined. It appears intuitive that most factors would stimulate IL-6 production in macrophages and, therefore, it is difficult to speculate that cellular infiltrates may have a beneficial effect in the heart.

In summary, the present results suggest that adenosine receptors differentially regulate the expression of proinflammatory cytokines in the heart. Cardiac expression of TNF- and IL-6 are inhibited and induced by the A₂ and A₃ receptor, respectively. Furthermore, the present data identify a critical role for inflammatory cells in the expression of cytokines in the human failing myocardium. These results may be relevant in the consideration of a therapeutic role for adenosine and adenosine analogs for cardioprotection in patients with congestive heart failure.