Putting The Brakes On Hypertensive Remodeling: An ATF3 Mechanism of Myofibroblast Restraint

Heart disease remains the leading cause of death in this country, and for the first time in 30 years the Center for Disease Control and Prevention reported that cardiac-related deaths are on the rise. A primary contributor to these numbers is hypertensive heart disease, which is characterized by myocyte hypertrophy and widespread fibrosis. Fibrotic remodeling increases both cardiac arrhythmogenicity and myocardial stiffness that cause a rapid progression toward failure. Thus, decades of research have been directed towards identifying the basis for cardiac remodeling with an eye towards short circuiting this maladaptive process. In doing so the field has largely focused on cardiac myocytes, but the recent emergence of targeted genetic approaches for manipulating cardiac cells has revealed vital roles for non-myocytes in mediating cardiac plasticity. Fibroblasts have gained traction in this respect, as demonstrated by the surge of evidence shifting its role from a passive player responding to cardiac myocyte dysfunction to a primary component of the disease process^{1, 2}.

Building on this theme, findings from Li et al.³, which are featured in this issue of Circulation, underscore the importance of revisiting previously defined remodeling signals in a more targeted cellular context. This study was focused on uncovering new transcriptional mechanisms for cardiac remodeling, and with RNAseq analysis they identified 22 transcription factors that were differentially expressed in whole heart lysates from 2 independent mouse models of hypertensive heart disease- chronic angiotensin II (AngII) infusion and pressure overload (TAC). Within this overlapping subgroup activating transcription factor 3 (ATF3) had the greatest increase in expression and was the most consistently up-regulated of the 22 candidate factors. ATF3 has long been identified as a global stress-response mechanism because it's rarely expressed in healthy tissues, but is induced by a wide range of toxic or injury stimuli⁴, leaving open whether its function is protective or maladaptive. Certainly, ATF3's role in the heart has been hard to ascertain because of conflicting phenotypes from gain and loss of function genetic studies^5-8. For instance, cardiac myocyte-specific transgenic overexpression of ATF3 promoted cardiac hypertrophy in mice^{6, 8}; whereas, global loss of ATF3 function both promoted and protected against hypertrophic remodeling in response to pressure overload^{6, 7}. Indeed, ATF3's function in regulating myocyte growth was equally variable in ischemic and diabetic heart disease models, yet enhanced fibrosis^5-8 was consistent among studies of global ATF3 knockout mice. So when Li et al.³ uncovered that hypertension-induced ATF3 expression was substantially up-regulated in fibroblasts by a factor of 4 versus cardiac myocytes and macrophages, it suggested that ATF3 stress signaling in the fibroblast may have greater biological impact on cardiac plasticity than its previously implicated role in myocytes thereby providing an impetus to study ATF3 function in cardiac fibroblasts³.

Using chronic AngII exposure as a model of hypertension this study found that global ATF3 deficiency caused greater myocyte hypertrophy and fibrosis relative to littermates replete with ATF3. Both myocyte hypertrophy and cardiac fibrosis were significantly reduced by re-expressing ATF3 solely in fibroblasts of ATF3 knockout mice suggesting that ATF3's function in fibroblasts was the primary driver of remodeling in hypertension. In healthy hearts fibroblasts are responsible for extracellular matrix (ECM) homeostasis, which includes regulating its quantity, quality and organization⁹. During disease the ECM becomes extremely stiff due to fibroblasts that have differentiated into myofibroblasts, cells specialized for fibrotic matrix secretion (Figure 1). While not rigorously tested in vivo, Li et al.³ identified that ATF3 negatively regulates fibroblast to myofibroblast differentiation as loss of ATF3 increased myofibroblast gene expression (i.e. smooth muscle α-actin and collagen-1) in response to AngII infusion. AngII can have both systemic and localized effects on the heart. Systemically, AngII increases blood pressure providing additional afterload which triggers maladaptive hypertrophy, but it also directly induces fibroblast to myofibroblasts differentiation^{10, 11}. Discriminating between local and systemic effects was not feasible in this in vivo model, leaving unanswered whether cardiac cells remodel because of systemic increases in mechanical load or localized chemical signaling. To address this Li et al. performed in vitro co-culture experiments with wild-type myocytes and ATF3 knockout fibroblasts. Here under AngII stimulation wild-type myocytes hypertrophied when co-cultured with ATF3 knockout fibroblasts. These findings are particularly intriguing as they suggest a synergistic relationship between fibroblast and myocytes in the context of cardiac plasticity; whereby, fibroblast function can modulate myocyte growth. What fibroblast factor is promoting myocyte hypertrophy is still unclear as several factors including remodeled matrix, myofibroblast contractility and electrical coupling, secreted factors or all of the above could impact myocyte behavior.

Figure 1.

Hypothetical signaling network whereby ATF3 negatively regulates hypertension-induced fibrotic remodeling by restraining MKK3/MKK6-p38 activity. Blue text represents activated signaling pathways initiating fibroblast to myofibroblast formation and cardiac fibrosis, while red text signifies negative regulation that prevents fibrosis. The image representing cardiac fibrosis is a Masson's trichrome stained transverse section of the left ventricle following 2 weeks of chronic AngII infusion where blue staining represents matrix deposition.

Efforts directed towards understanding how ATF3 regulates hypertension-induced fibrosis uncovered that in AngII-activated fibroblasts ATF3 represses the Map2k3 locus through an association with histone deacetylase-1 (HDAC1)³. This correlated with a down regulation of the Map2k3 gene, Mitogen-Activated Protein Kinase Kinase 3 (MKK3, Figure 1). The relationship between Mitogen-Activated Protein Kinase (MAPK) signaling, ATF3, and HDAC1 interactions is not novel as these interactions were previously identified in mouse embryonic fibroblasts (MEFs)¹² but in conjunction with other inflammatory genetic loci⁴. However, in reference to matrix remodeling this finding represents a new branch in the cardiac fibrosis regulatory network¹. Given that many chemical and mechanical stimuli including AngII require MKK3-MKK6 activation of p38 signaling to induce fibroblast to myofibroblast differentiation¹⁰, these data fit with current regulatory paradigms for fibrosis. Left open is where ATF3 actually fits within the MKK3/MKK6-p38 signaling pathway in cardiac fibroblasts, as previous data suggest that p38 is required for ATF3 expression¹². Confirming this downstream placement would involve identifying whether p38 is required for disease-mediated up-regulation of ATF3 in cardiac fibroblasts. While this study failed to directly examine the requirement of ATF3-HDAC1 complex occupancy of the Map2k3 locus for blocking fibrosis and myofibroblast differentiation, the requirement of p38 as an activator of fibrotic gene expression in the context of ATF3 deficiency was functionally confirmed in vitro with pharmacologic inhibition and shRNA-mediated knockdown of p38. These results demonstrated a powerful effect of p38 inhibition to block the expression of fibrotic genes and myocyte hypertrophy in vitro. These outcomes, however, could still be independent of ATF3 and solely due to inhibition of p38 signaling itself. If ATF3 is indeed downstream of p38, then coupled with results from Li et al., it suggests that ATF3 functions as a brake on p38 signaling by down-regulating MKK3 expression (Figure 1). The observation that in the absence of a stimulus neither gain nor loss of ATF3 function in the fibroblast by itself causes the heart to remodel supports this hypothesis. Here if the ATF3 brake is removed, such as in the global ATF3 knockouts, p38 signaling which is sufficient for triggering fibrosis¹⁰ is fully active and maintained thereby enhancing fibrosis. Thus, ATF3 appears to function within a feedback loop that negatively regulates cardiac fibrosis by restraining p38 activity (Figure 1).

Although the magnitude of ATF3 gene expression was highest in fibroblasts it was nonetheless up-regulated in myocytes and macrophages. Whether ATF3 overexpression elicits the same activation and repressive epigenetic and transcriptional patterns in other cardiac cell populations wasn't tested beyond the fibroblast. So, could ATF3-restraint of MKK3/MKK6-p38 signaling apply to the myocyte and hypertrophic remodeling? Looking at p38's role in myocyte remodeling doesn't provide instant clarity due to the complexity of this signaling pathway. For example myocyte-specific activation of p38 signaling had no effect on cardiac remodeling, but dominant negative p38 expression promoted the hypertrophic response to TAC, suggesting p38 is antihypertrophic¹³. Yet myocyte-specific deletion of p38 revealed that it's not required for pressure overload-induced hypertrophy¹³. So, if ATF3 restrains p38 signaling in myocytes then one would predict that ATF3 knockout mice should behave similarly to those with myocyte activation of p38 signaling, which saw no alteration in hypertrophy. Since this is the opposite response from results presented here in which ATF3 knockout mice were prohypertrophic, it suggests that alternative transcriptional regulatory mechanisms underlie ATF3-dependent changes with respect to myocyte plasticity.

In summary, Li et al. established a new cardiac fibrosis regulatory mechanism operating through ATF3's restraint of p38-MAPK signaling, which substantially inhibits the heart's fibrotic response to hypertension. Many questions still remain as to its regulatory function in other cardiac cell populations including the myocyte that also experienced a rise in ATF3 expression with disease. Whether myocyte remodeling in this case was mediated by ATF3 itself or a synergistic response to altered fibroblast function is unknown. Further analysis to address these issues is needed in order to define ATF3's role in cardiac remodeling. This could be achieved with comparative and transcriptional analysis of conditional ATF3 knockouts created via cell-targeted deletions of ATF3 in the various cardiac cell populations. For now, this study took a step forward by identifying a way of dampening the fibrotic response to hypertension, which may yield insights into new therapeutic strategies for preventing maladaptive cardiac remodeling.