Regional evidence of modulation of cardiac adiponectin level in dilated cardiomyopathy: pilot study in a porcine animal model

Adiponectin (ADN) is an adipocyte-derived protein that is abundantly present in plasma [1] and is also expressed in cardiomyocytes [2] and endothelial cells [3]. Two ADN receptors, AdipoR1 and AdipoR2, have been characterized: AdipoR1 is abundantly expressed in skeletal muscle, whereas AdipoR2 is predominantly expressed in the liver [4]. T-cadherin, an additional ADN receptor, is expressed in cardiomyocytes and is involved in ADN-mediated cardioprotection [5]. The remodeling of the failing heart has induced a number of investigators to test the hypothesis that modulation of myocardial metabolism might prove therapeutically advantageous [6].

Potential downstream effectors of ADN receptors include AMP-activated protein kinase (AMPK) and peroxisome proliferator-activated receptor-α (PPAR-α). AMPK is considered an energy-sensing enzyme that on stimulation enhances glucose use and fatty acid oxidation, whereas PPAR-α is a key nuclear transcription factor, regulating expression of genes involved in fatty acid uptake and use [7]. It was recently shown that cardiac ADN and its receptors are elicited by PPAR-γ activation, confirming the role of ADN in maintaining myocardial energy homeostasis [2].

In addition, cardioprotective effects of ADN have been well-demonstrated [8‐11]. Many clinical studies have shown that ADN is inversely correlated with an increased cardiovascular risk, and hypo-adiponectinemia has been established as an independent cardiovascular risk factor in coronary artery disease [12‐17]. Similarly, in patients with dilated cardiomyopathy (DCM) without overt heart failure (HF), ADN down-regulation is associated with severe endothelial/microvascular dysfunction [18]. Conversely, in patients with overt HF, ADN levels are increased [19] in presence of an increased mortality risk [20]. Further investigation of the autocrine/paracrine role of the cardiac ADN system in the failing heart at different stages could aid in understanding the “ADN paradox” [21]. This issue is still debated and little information is available.

The main aim of our study was to assess the regional modulation of ADN system in a clinically relevant model of non-ischemic HF, both at cardiac and systemic circulating levels. For this purpose, the myocardial expression of ADN, ADN receptors and modulators was determined in an experimental model of LV-pacing induced HF [22] and compared with healthy myocardium. ADN signaling was investigated to address myocardial remodeling, inflammation and metabolic process in different regions of the non-ischemic failing left ventricle (LV) regarding biochemical and functional patterns.

We investigated the myocardial ADN pathway in a clinically relevant animal model of non-ischemic dilated cardiomyopathy. The main findings of our study are: first, increased circulating ADN levels in HF minipigs did not affect myocardial AMPK expression and activity, nor PPARα expression; second, sustained high-rate LV pacing caused a significant reduction of ADN level and increased AdipoR1 level in PS, but not in OS; third, regional ADN down-regulation was related to severe contractile impairment, yet the key regulators of altered cardiac metabolism, inflammation and remodeling were unchanged compared to control hearts.

High circulating ADN levels have been reported repeatedly in HF [19‐21, 28] and several mechanisms have been utilized to explain this increase. Recent studies have reported that natriuretic peptides enhance ADN production via the cGMP pathway, by human adipocytes both in vitro and in vivo in patients with HF [29, 30]. We have previously described that BNP and CNP expression increased in adult minipigs after 21 days of LV pacing [25]. It is well-known that natriuretic peptides are secreted almost exclusively by cardiac tissue and their levels are increased in patients with HF [31]. It has been hypothesized that the ADN increase could be a compensatory effect in order to restore metabolic homeostasis, due to its anti-inflammatory and insulin-sensitizing effects [32, 33]. Abnormalities in glucose regulation, such as impaired glucose tolerance and insulin resistance, have been shown to correlate with HF severity [34]. In our animal model, we previously found that plasma insulin concentration was significantly higher in HF compared to control in presence of homogenous LV increase of glucose uptake [22]. In our study, no change was observed in AMPK activity and PPARα expression in presence of a marked reduction of global and regional LV function. The lack of cardiac activation of the ADN downstream molecules could suggest that their associated cardiac effects are not related to high ADN plasma levels. This may agree with previous studies reporting an association between high levels of ADN in HF and severity as well as increased mortality, indicating for ADN a role as prognostic biomarker of HF [20, 21]. Based on these findings, a “functional ADN resistance” in HF at the receptor level has been hypothesized. This resistance could determine an attenuated ADN response during the progression of HF [21], as recently observed at the level of skeletal muscle in chronic HF [35]. In this paper, a strong positive correlation between AdipoR1 and PPARα/AMPK gene expression was found. In our study, we found an increased expression of AdipoR1 in the HF pacing site in presence of myocardial ADN mRNA down-regulation, yet PPARα/AMPK gene expression and AMPK activity were not significantly different from healthy heart. As to T-cadherin, a physical association of ADN with T-cadherin is a prerequisite for ADN’s physiological activity in the heart [5]; thus, the reduction of T-cadherin mRNA expression in failing heart, although not significant, could support the reduced activity of ADN/AdipoR system in this condition.

The different behavior of systemic and cardiac ADN expression (Figure 1) as well as the presence of a regional variation in myocardial tissue after pacing are in tune with the existence of a local cardiac ADN regulation, independent of the circulating ADN, as previously described in human [36, 37] and murine hearts [38]. The cardiac protein expression of ADN in our samples was difficult to detect. We detected small amounts of ADN in the extracellular matrix surrounding the injured areas of the PS, but not in OS and controls. Consistent with these findings, previous studies have reported the presence of ADN in damaged regions of the heart, also due to its leakage from the vascular section [39‐41]. Moreover, down-regulation of ADN mRNA expression, compared to controls, and the myocardial over-expression of AdipoR1 in PS suggested the existence of a possible feedback loop, as previously observed in skeletal muscle [35].

The mechanisms underlying this down-regulation are yet unknown. It is well-recognized that inflammatory cytokine production, particularly TNF-α, plays a critical pathogenic role in cardiovascular complications, and a reciprocal action between TNF-α and ADN exists [42, 43]. However, in our model we found that cardiac ADN was down-regulated in absence of inflammatory response. In fact, TNF-α at the mRNA and protein level, was unchanged in each region of failing left ventricle and no inflammatory cells were detected. Accordingly, myocardial expression of PPAR-γ, a known upstream regulator of ADN [2], was also unchanged in each region of the failing heart. On the other hand, Sturk et al. provided evidence, directly in cardiomyocytes, of a cardiac regulatory feedback loop in ADN expression, without identifying any specific cardiomyocyte-derived secreted factors that exert this negative feedback [36]. Otherwise, a negative relationship between ADN and BNP mRNA expression was observed in this study (Figure 3C), suggesting that impaired contractile function in the failing heart could affect the ADN/AdipoR1 system. In a recent study in murine hearts with transverse aortic constriction, inverse correlations between myocardial ADN and BNP as well as heparin-binding epidermal growth factor (HB-EGF) were shown, indicating an important role for ADN in mediating the myocardial hypertrophic signaling pathway [38]. However, in our experimental animal model, cardiac tissue did not develop hypertrophic remodeling, as shown by the collagen immunohistochemistry (Figure 3E). The lack of collagen deposits in PS in absence of coronary stenosis and inflammation suggests the occurrence of non-ischemic ventricular remodeling due to mechanical stress, which could impair the cardiac ADN/AdipoR1 system. Sen et al.[44] demonstrated in vitro that continuous mechanical strain inhibits ADN expression at the transcriptional level.

As a matter of fact, it has been suggested that ADN may directly protect cardiomyocytes [37] largely via an AMPK-mediated signal pathway [9]. However, the lack of variation of AMPK expression and activity in our animal model supports the hypothesis that other pathways drive the cardioprotection afforded by ADN. In fact, treatment with ADN is still effective in improving the cardiac function in AMPK knockout mice [45]. The cardioprotective action of ADN may be performed by suppression of TNF-α signaling via the COX-2-prostaglandin E2-linked cascade [46]. In our conditions, TNF-α did not show any changes between HF and controls either at mRNA or protein level (Figure 4). In addition, it has been reported that ADN might prevent excess NO generation by inhibiting iNOS expression [45]. In our model, no significant variation of iNOS expression was found in HF compared to controls and no correlation with myocardial ADN was observed.

Our findings were observed in a reliable translational animal model of non-ischemic HF. The pacing-induced HF model is considered the gold standard in HF experimental research due to its relative similarities with many features of clinical dilated cardiomyopathy [51]. In conclusion, we observed that in spite of its high peripheral concentrations, plasma ADN did not show cardioprotective effects, confirming its role as a negative prognostic biomarker of HF. In particular, the effect of adiponectin in non-ischemic HF is AMPK-independent. These observations suggest the occurrence of novel mechanisms by which reduced cardiac ADN levels may regionally mediate the decline of cardiac function.