This study shows that circulating plasma activin A levels in T2D men without cardiovascular complications associate with impaired myocardial glucose metabolism, independent of M-value. Furthermore, we observed a positive association with the LVMV-ratio. Metformin treatment decreased activin A levels, while pioglitazone had no effect. Furthermore, in patients treated with pioglitazone, changes in plasma activin A levels were borderline significantly correlated with changes observed in LVMV-ratio. There was no association between changes in plasma activin A and myocardial glucose metabolism after either pioglitazone or metformin treatment. These results suggest an involvement of activin A in the pathogenesis of early cardiac derangements in T2D.
Importantly, the pathogenesis of diabetic cardiomyopathy is not yet completely unraveled. It has been proposed that myocardial dysfunction in patients with T2D is related to depletion of endothelial progenitor cells and increased oxidative stress . Others demonstrated that early T2D is associated with an increase in the sarco/endoplasmic reticulum Ca2+-ATPase (SERCA) / phospholamban-ratio and that insulin directly stimulates SERCA expression and relaxation velocity . Finally, circulating levels of osteoprotegerin and adiponectin were found to associate with cardiac abnormalities in men with uncomplicated T2D . The present study identifies activin A as additional potential factor contributing to the pathogenesis of myocardial dysfunction in diabetic cardiomyopathy.
Levels of circulating activin A were in the same range as other clinical studies [17–19]. Although one study reported that high activin A levels associated with abnormal glucose regulation in patients with myocardial infarction though without known T2D , we as well as others did not find altered activin A levels between patients with (uncomplicated) T2D and controls [19, 21, 22]. Nevertheless, activin A levels tended to be higher in patients with cardiovascular disease . In patients with stable and unstable angina, levels of activin A were elevated as compared to healthy controls . Also in T2D patients with coronary artery disease higher levels of activin A were found as compared to T2D patients without coronary artery disease . Finally, in heart failure patients, increased activin A levels were demonstrated as compared to healthy controls [18, 19].
Some studies have reported beneficial effects of activin A. Activin A is a dimeric protein consisting of two inhibin βA monomers linked by a single disulfide bond. In mice, maternal diabetes induces cardiac malformations in the embryos by downregulation of inhibin βA, and in association decreased activation of the post-receptor signaling components Smad2 and/or Smad3 [23, 24]. Furthermore, Zhao et al. demonstrated that activin A was able to rescue myocardial cell proliferation and endocardial cell migration in mouse embryonic hearts which was suppressed by maternal diabetes . Other investigators have proposed that activin A could have beneficial effects on inflammation and atherogenesis, and that high activin A levels reflect a counteracting anti-inflammatory and anti-oxidative response [19, 20]. The present findings do not support this, but are corroborated by our earlier studies in which we showed that activin A derived from EAT from T2D patients impairs cardiomyocyte insulin sensitivity by inhibiting the insulin-mediated phosphorylation of Akt through induction of microRNA-143 [6, 26].
The association of activin A with LVMV-ratio indicates that this factor is involved in early myocardial remodeling in T2D as well, as LVMV-ratio is one of the features of isolated LV diastolic dysfunction [7, 13, 27]. Importantly, this is supported by our results showing that changes in activin A levels after only pioglitazone were borderline positively correlated with changes in LVMV-ratio. Although an in vitro study on monocytes reported no effect of pioglitazone on activin A release , it should be noted that the key effector for pioglitazone, PPARγ, is predominantly expressed in adipose tissue. Furthermore, expression is activin A is ubiquitous [6, 7, 28, 29], and high expression levels have been reported in human adipocyte progenitor cells . During differentiation into adipocytes activin A expression is downregulated, and inhibition of the activin A signaling has been found to promote human adipogenesis . Based on these observations, one may speculate that pioglitazone improves adipocyte differentiation and adipose tissue quality, thereby lowering activin A production by mature adipocytes. Interestingly, a recent report directly linked activin A released from EAT to the development of cardiac fibrosis . This raises the possibility that the reduction of the LVMV-ratio in pioglitazone-treated patients may result from local effects of pioglitazone on epicardial adipocytes rather than a systemic effect on circulating activin A levels. Accordingly, we could recently demonstrate that activin A release is increased in EAT from patients with T2D [6, 26]. This fat depot is not separated by a fascia from the underlying tissue and shares the blood supply with the myocardium. Consequently, activin A released from EAT can directly affect cardiac function in a paracrine fashion.
The current findings do not completely exclude the possibility that the effects of pioglitazone on the LVMV-ratio result from an improved left ventricular diastolic function in this treatment group. In support of this notion is that metformin treatment did not affect cardiac function, including the LVMV-ratio , despite lowering circulating activin A levels. An in vitro study on monocytes also reported a downregulation of activin A release following metformin exposure . However, the absence of a beneficial effect of metformin on determinants of left ventricular diastolic function may prevent beneficial effects on left ventricular hypertrophy despite lower circulating activin A levels.
Another issue that should be considered is that the expression and release of activin A is not confined to EAT. Rather activin A is expressed by a large number of cells and tissues, including (epicardial) adipose tissue [6, 7, 28, 29]. Because the circulating levels of activin A were not different between controls and patients with T2D, it seems likely that the amount of activin A found in the circulation is also derived from these other sources. Nevertheless, both our previous in vitro studies [6, 26] and the present clinical study support the notion that activin A participates in the development of cardiac abnormalities in patients with T2D.
This study has several limitations. First, the number of control subjects examined in this study was relatively small. Consequently, additional studies remain required to confirm the lack of association between activin A and cardiometabolic parameters in subjects without T2D. Second, our study was conducted in males. Although this may limit generalization of our findings, it should be noted that multiple studies showed that gender is not a confounding factor for levels of activin A levels in the circulation [18, 19, 22, 30, 31].