Mutual Regulation of Epicardial Adipose Tissue and Myocardial Redox State by PPAR-γ/Adiponectin Signalling

Abstract

Rationale: Adiponectin has anti-inflammatory effects in experimental models, but its role in the regulation of myocardial redox state in humans is unknown. Although adiponectin is released from epicardial adipose tissue (EpAT), it is unclear whether it exerts any paracrine effects on the human myocardium.

Objective: To explore the cross talk between EpAT-derived adiponectin and myocardial redox state in the human heart.

Methods and results: EpAT and atrial myocardium were obtained from 306 patients undergoing coronary artery bypass grafting. Functional genetic polymorphisms that increase ADIPOQ expression (encoding adiponectin) led to reduced myocardial nicotinamide adenine dinucleotide phosphate oxidase-derived O2 (-), whereas circulating adiponectin and ADIPOQ expression in EpAT were associated with elevated myocardial O2 (-). In human atrial tissue, we demonstrated that adiponectin suppresses myocardial nicotinamide adenine dinucleotide phosphate oxidase activity, by preventing AMP kinase-mediated translocation of Rac1 and p47(phox) from the cytosol to the membranes. Induction of O2 (-) production in H9C2 cardiac myocytes led to the release of a transferable factor able to induce peroxisome proliferator-activated receptor-γ-mediated upregulation of ADIPOQ expression in cocultured EpAT. Using a NOX2 transgenic mouse and a pig model of rapid atrial pacing, we found that oxidation products (such as 4-hydroxynonenal) released from the heart trigger peroxisome proliferator-activated receptor-γ-mediated upregulation of ADIPOQ in EpAT.

Conclusions: We demonstrate for the first time in humans that adiponectin directly decreases myocardial nicotinamide adenine dinucleotide phosphate oxidase activity via endocrine or paracrine effects. Adiponectin expression in EpAT is controlled by paracrine effects of oxidation products released from the heart. These effects constitute a novel defense mechanism of the heart against myocardial oxidative stress.

Keywords: adiponectin; adipose tissue; myocardium; obesity; oxidative stress.

Figure 1.

Circulating adiponectin (AdN) is positively related with myocardial redox state in patients with ischemic heart disease. In the Clinical Associations Studies, high circulating AdN levels were paradoxically related with high myocardial nicotinamide adenine dinucleotide phosphate (NADPH)–stimulated superoxide (O₂⁻) (A) and high plasma malonyldialdehyde (MDA; a marker of systemic oxidative stress; B). There was no association of circulating adiponectin with plasma interleukin-6 (IL-6; C) or high-sensitivity C-reactive protein (hsCRP; D). High circulating adiponectin was also positively related with high plasma brain natriuretic peptide (BNP; E). Values are expressed as median (25th–75th percentile).

Figure 2.

Genetically conferred increases in adiponectin (AdN) bioavailability are causally associated with lower nicotinamide adenine dinucleotide phosphate (NADPH) oxidase activity in the human myocardium. The total number of rs17366568G alleles (polymorphism in ADIPOQ gene) and rs266717T alleles (polymorphism in ADIPOQ gene promoter) had an additive effect on circulating AdN levels (A) and was associated with reduced NADPH–stimulated superoxide (O₂⁻) in human myocardium (B). The number of rs17366568G/rs266717T alleles was positively associated with higher ADIPOQ gene expression in thoracic adipose tissue (ThAT, C), but not associated in epicardial adipose tissue (EpAT; D). Patients with higher ADIPOQ (E) or peroxisome proliferator-activated receptor (PPAR)-γ (F) gene expression in EpAT also had higher NADPH-stimulated O₂⁻ production in their myocardium. Higher myocardial NADPH-stimulated O₂⁻ was not associated with endogenous ADIPOQ gene expression in the heart (G), but was associated with higher gene expression of adiponectin receptor-1 (AdipoR1), but not of AdipoR2 or T-cadherin (CDH13) in human myocardial tissue (H). Values are expressed as median (25th–75th percentile). RLU indicates relative light units.

Figure 3.

Effects of recombinant adiponectin (AdN) on myocardial redox state in humans.Ex vivo incubation of human myocardium with AdN (10 μg/mL) for 2 h resulted in increased phosphorylation of AMP-kinase (AMPK)-α at Thr172 (p-AMPK; A) leading to AMPK activation as assessed by the phosphorylation status of its downstream target acetyl-CoA carboxylase (ACC) at Ser79 (p-ACC; B), an effect reversed by compound C (CC; 10 μmol/L; A and B). AdN reduced superoxide (O₂⁻) production in human myocardium (C) and specifically nicotinamide adenine dinucleotide phosphate (NADPH) oxidase activity, as assessed by measuring the Vas2870-inhibitable (40 μmol/L) O₂⁻ (D); both effects were reversed by CC (C and D). These effects of AdN on myocardial O₂⁻ were also confirmed by dihydroethidium (DHE) staining; AdN reduced both total and Vas2870-inhibitable DHE fluorescence, and these effects were reversed by CC (E–G). Importantly, AdN prevented Rac1 activation (assessed by measuring the ratio of GTP-Rac1:total Rac1 [t-Rac1; H]) and reduced the membrane-bound fraction of Rac1 (m-Rac1; I); both effects were reversed by CC. Similarly, AdN prevented p47^phox phosphorylation at its activatory site Ser359 (p-p47^phox, J) and reduced the membrane-bound fraction of nicotinamide adenine dinucleotide phosphate (NADPH) oxidase subunit p47^phox (m-p47^phox; K) in an AMPK-dependent manner (as both effects were reversed by CC; J and K). A and B, AdN: n=7 to 13 per group; CC: 4 to 7 per group; (C, D, H, J, K) AdN, n=7 to 12; CC group, n=4 to 6; (E–I) n=3 to 4 per group. Values are expressed as fold change vs control group and shown as mean±SEM; *P<0.05, **P<0.01 vs control.

Figure 4.

Activation of nicotinamide adenine dinucleotide phosphate (NADPH) oxidase in H9C2 cardiomyocytes triggers peroxisome proliferator-activated receptor gamma (PPAR-γ) signaling in rat pericardial fat: identifying a novel inside-to-outside signal. To examine whether under conditions of increased endogenous oxidative stress cardiac myocytes release a transferable factor able to affect the activation of PPAR-γ/adiponectin signaling in rat epicardial adipose tissue (EpAT), we exposed H9c2 cells (differentiated to cardiac myocytes) to NADPH 100 μmol/L for 2 h, whereas rat EpAT was conditioned ex vivo (A). After 2 h, the rat EpAT was transferred into the H9c2 wells and cocultured for an additional 16 h (A). At the end of the incubation period, gene expression was studied in the rat EpAT. Addition of NADPH to intact H9c2 cells grown on coverslips led to a striking increase of NADPH oxidase–derived superoxide (O₂⁻) that was partly inhibitable by either Vas2870 (a pan-Nox inhibitor) or gp91-dstat (a specific inhibitor of NOX2; B), as demonstrated by real-time monitoring using lucigenin-enhanced chemiluminescence. Coincubation of rat EpAT with H9c2 cardiac myocytes stimulated with NADPH resulted in an upregulation of ADIPOQ (C), PPAR-γ (D), and CD36 (E) in EpAT at 16 h. All these effects were prevented by polyethylene glycol (PEG)-SOD (300 U/mL) or vas2870 (10 nmol/L). The presence of unstimulated H9C2 cells or NADPH alone had no effect on the expression of ADIPOQ, PPAR-γ, or CD36 genes in the rat EpAT (C–E). Concentration of gp91-dstat was 50 μmol/L. Values are presented as mean±SEM. B, n=7; (C–E), n=7; *P<0.05 vs control group; ***P<0.0001 vs resting; †P<0.05 vs NADPH alone. DMEM indicates Dulbecco’s Modified Eagle Medium; RLU, relative light units; and SOD, superoxide dismutase.

Figure 5.

Myocardial oxidation product 4-hydroxynonenal (4HNE) as a mediator of the inside-to-outside signal from the human myocardium to epicardial adipose tissue (EpAT). In human myocardium, increased nicotinamide adenine dinucleotide phosphate (NADPH) oxidases activity (upper tertile of NADPH-stimulated O₂⁻ in the Clinical Associations Studies) were associated with significantly greater 4HNE (A) and malonyldialdehyde (MDA; B) production. Incubation of EpAT and thoracic (ThAT) adipose tissue with MDA (1 mmol/L) for 16 h had no significant impact on ADIPOQ gene expression (C). On the contrary, incubation of EpAT with 4HNE (30 µmol/L) for 16 h induced a striking increase in ADIPOQ (D), PPAR-γ (E) and CD36 (F). The effects of 4HNE on ADIPOQ and CD36 gene expression were reversed by the inhibitor of PPAR-γ activity, T0070907 (10 μmol/L; D and F). Importantly, 4HNE had no significant impact on the expression of ADIPOQ (G), PPAR-γ (H), or CD36 (I) genes in ThAT. Values are represented as fold change compared to control group (mean±SEM). A and B, n=4 per group; (C) n=5 per group; (D–I), n=5 to 8 per group, *P<0.05; **P<0.01 vs low myocardial NADPH oxidase activity (A and B) or control (C to I).

Figure 6.

Testing the inside-to-outside paracrine effects of the heart on pericardial adipose tissue using a cardiomyocyte-specific Nox2-tg mouse model. In the cardiac myocyte–specific NOX2-transgenic mouse, myocardial nicotinamide adenine dinucleotide phosphate (NADPH) oxidases were activated, as assessed by both the NADPH-stimulated (A) and Vas2870-inhibitable superoxide (O₂⁻) signal (B), and by increased formation of 4-hydroxynonenal (4HNE) protein adducts when compared with wild-type (wt) animals (C). There was no difference in the myocardial protein levels of malonyldialdehyde (MDA) adducts (D). Increased myocardial oxidative stress and 4HNE adducts formation in mNOX2-tg mice led to increased ADIPOQ gene expression in the fat attached to the heart (pericardial adipose tissue [PerAT]), but not in remote AT depots, eg, subcutaneous AT (ScAT; E). mNOX2-tg mice also had increased endogenous levels of ADIPOQ gene expression in myocardial tissue (F), but there was no difference in the myocardial gene expression levels of any of adiponectin receptors, T-cadherin (CDH13), AdipoR1 and AdipoR2 (G); (A–E), n=5 to 6 per group; (F and G) n=9 to 10 per group, *P<0.05, **P<0.01 vs wt group. RLU indicates relative light units.

Figure 7.

Testing the inside-to-outside paracrine effects of the heart on adipose tissue using a pig model of rapid atrial pacing (RAP). In a pig model of RAP, myocardial nicotinamide adenine dinucleotide phosphate (NADPH) oxidases activity was significantly higher in the paced animals than in sham, as shown by both the NADPH-stimulated and Vas2870-inhibitable superoxide (O₂⁻) signal (A and B). RAP also increased formation of 4-hydroxynonenal (4HNE; but no malonyldialdehyde [MDA]) protein adducts (C and D) and upregulated ADIPOQ gene expression in epicardial AT (EpAT) but not in remote AT depots, eg, ScAT (E). There was no difference in endogenous ADIPOQ gene expression levels in the heart (F) or myocardial expression of adiponectin receptors (G) between sham-operated and RAP animals; all panels, n=5 per group, *P<0.05, **P<0.01 vs sham group. RLU indicates relative light units.