A novel protective role for activating transcription factor 3 in the cardiac response to metabolic stress

This editorial refers to ‘ATF3 expression in cardiomyocytes preserves homeostasis in the heart and controls peripheral glucose tolerance’ by R. Kalfon et al., pp. 134–146. Type 2 diabetes mellitus (T2D) can promote the development of a specific form of cardiomyopathy independent of coronary artery disease and hypertension. This is the so-called diabetic cardiomyopathy, which represents a major cause of heart failure in Western countries.1 Unfortunately, the molecular mechanisms underlying the pathogenesis of diabetic cardiomyopathy still need to be elucidated. Similarly, the molecular and metabolic pathways mediating the cardiac response to diabetes-associated metabolic derangements are still elusive.1 The article by Kalfon et al.2 significantly extends our current knowledge regarding the signalling pathways that are modulated in the heart in response to diabetes. They identified activating transcription factor 3 (ATF3) as a critical sensor of the metabolic stress associated with T2D in cardiomyocytes. ATF3 was found to protect the heart against T2D-induced cardiac remodelling and dysfunction, also controlling peripheral insulin sensitivity. ATF3 is an immediate early transcription factor of the basic leucine zipper (bZIP) superfamily, whose expression is generally low in physiological conditions. However, it is up-regulated by multiple stressors, including those that are involved in T2D cardiomyopathy, such as metabolic, oxidative, and endoplasmic reticulum stress, as well as inflammation.3 ATF3 functions by binding to the cyclic AMP response element (CRE) and, in turn, modulates the expression of multiple target genes including pro-inflammatory cytokines and chemokines.3 Previous work demonstrated that ATF3 modulates cardiac adaptation to different kinds of stressors, exerting both adaptive and maladaptive functions. ATF3-deficient mice develop hypertrophy, fibrosis and cardiac dysfunction in response to aortic banding.4 ATF3 was shown to reduce cardiomyocyte death in response to doxorubicin.5 Myocardial ATF3 is rapidly activated by ischaemic preconditioning, reducing ischaemic injury, inflammation, and neutrophil infiltration.6 On the other hand, ATF3 gene deletion was found to limit cardiac hypertrophy in response to phenylephrine.7 A previous study showed that systemic ATF3 gene disruption reduced mouse survival in response to chronic myocardial infarction, although it also limited fibrosis and hypertrophy.8 In addition, studies conducted in transgenic mice with cardiomyocyte-specific ATF3 overexpression demonstrated that prolonged activation of ATF3 exerts detrimental effects. ATF3 overexpression was seen to promote massive hypertrophy, fibrosis, cardiac dysfunction, and conduction abnormalities.9 Several reasons may explain the dual effect of ATF3 activation in response to cardiac stress. ATF3-mediated responses appear to be adaptive or maladaptive depending on the stress condition. The level and duration of ATF3 activation may be critical for determining whether the transcription factor is beneficial or detrimental. The effects of ATF3 during stress may also be cell type specific. In addition, it is possible that the signalling mechanisms promoting the up-regulation of ATF3 in response to stress may determine its cellular effects. Finally, ATF3 binding patterns may have a crucial role in controlling ATF3 targets and functions. For example, previous work demonstrated that the ATF3 homodimer can bind to CRE elements whereas the ATF3 heterodimer with CHOP/GADD153 cannot bind to it.10 The study by Kalfon et al.2 significantly extended this previous evidence characterizing the cardiac phenotype of mice with cardiac-selective deficiency of ATF3 (ATF3 cKO) subjected to 15 weeks of high fat diet (HFD). Cardiomyocyte-restricted inactivation of ATF3 caused more severe cardiac fibrosis and more pronounced hypertrophy and inflammation in response to HFD, with respect to what was observed in wild type animals. ATF3 cKO mice also displayed more severe cardiac dysfunction than controls in response to T2D. At a molecular level, cytokines, such as TNF-α, IL-6 and TGF-β3, were significantly increased in ATF3 cKO mice with respect to wild type mice in response to HFD. Tissue inhibitor metalloproteinases 1 (TMP1) was also up-regulated in the hearts of diabetic ATF3 cKO, corresponding to increased cardiac fibrosis. In addition, knockout mice showed a marked up-regulation of foetal genes, a molecular signature of maladaptive remodelling. Several aspects regarding the role of ATF3 in the cardiac adaptation to diabetes remain to be clarified. First of all, it is not clear how ATF3 is up-regulated by HFD. Activation of p38 and PKA may be involved in these mechanisms.11 In addition, previous work demonstrated that ATF3 may be induced by lipid-dependent activation of Toll-like receptor 4 (TLR4).12 In this regard, Kalfon et al.2 found that palmitate is sufficient to induce ATF3 in cardiomyocytes. Finally, ATF3 was previously found to be activated by the unfolded protein response,6 which is usually activated in the heart in response to metabolic stress. The signalling mechanisms through which ATF3 limits the progression of diabetic cardiomyopathy also remain to be clarified. ATF3 was previously shown to repress p53 expression in response to doxorubicin,5 suggesting that this factor may directly preserve cardiomyocyte survival. It can also repress Egr1,13 a transcription factor involved in the promotion of cardiac hypertrophy and damage.14 In addition, ATF3 may limit hypertrophy also by restraining the ERK1/2 and JNK signalling. Finally, ATF3 interferes with TLR4 signaling15 and can repress NF-κB,16 which may explain its anti-inflammatory and cardioprotective effects. In regard to the latter, recent evidence demonstrated that genetic inactivation of TLR4 or NF-κB is able to prevent diabetic cardiomyopathy.17,18 Kalfon et al.2 found that cardiac ATF3, in addition to its beneficial effects in the heart, is also critically implicated in a heart-to-periphery communication, controlling insulin sensitivity of peripheral organs. Genetic inactivation of ATF3 in cardiomyocytes exacerbated systemic insulin resistance and blood glucose intolerance in response to HFD, further increasing metabolic stress. This suggests the existence of a dangerous vicious cycle in which myocardial dysfunction and metabolic stress mutually worsen in knockout mice. The precise mechanism underlying this phenomenon remains unclear. It is possible that an increased formation of cytokines in the heart of ATF3 cKO may aggravate systemic inflammation and in turn insulin sensitivity. In addition, Kalfon et al.2 found that ATF3 gene deletion significantly reduces myocardial insulin sensitivity through the down-regulation of IRS1 and 2. It is possible that the cardiac response to insulin may be an important determinant of systemic insulin sensitivity. Of note, this is not the only instance in which cardiomyocyte-specific mechanisms affect systemic insulin sensitivity. Olson’s group previously showed that cardiomyocyte-specific genetic disruption of MED13 significantly worsens HFD-induced obesity and insulin resistance.19 Schematic representation of the molecular mechanisms potentially involved in the cardioprotective effects of ATF3 in response to T2D, through which it restrains the development of diabetic cardiomyopathy. UPR, unfolded protein response; TLR4, toll-like receptor 4; IRS1/2, insulin receptor substrate; CM, cardiomyocyte. The authors wish to thank Dr Derek Yee for the critical revision of the manuscript. Conflict of interest: none declared.