Redox Regulation of Cardiac ASK1 (Apoptosis Signal-Regulating Kinase 1) Controls p38-MAPK (Mitogen-Activated Protein Kinase) and Orchestrates Cardiac Remodeling to Hypertension

Supplemental Digital Content is available in the text.

implicated in the development of heart failure, 7 being activated in cardiomyocytes and the heart by pathophysiological stresses such as redox stress and proinflammatory cytokines, including IL (interleukin)-1β. 8 p38-MAPKs and JNKs are generally activated in parallel, the only well-documented exception being with ischemia where there is selective activation of p38-MAPKs. 9 Hypertension is associated with increased redox stress and some degree of inflammation, 10 so modulating p38-MAPK or JNK activities may be therapeutically beneficial in hypertensive heart failure. p38-MAPKs and JNKs are, of themselves, potential therapeutic targets but, because of their role in inflammation, 11 it may be more desirable to target specific inputs into the cascade to modulate activity rather than block the signal completely. p38-MAPKs and JNKs are phosphorylated/activated by upstream MKKs (MAPK kinases), MKK3/6, and MKK4/7, respectively. These are phosphorylated/activated by upstream MAP3Ks (MKK kinases) which are more numerous, potentially responding selectively to specific stresses. 12 The greater diversity at the MAP3K level provides increased potential for selective targeting. Two such kinases are MAP3K5 (ASK1 [apoptosis signal-regulating kinase 1]) and MAP3K7 (TAK1 [transforming growth factor-activated kinase 1]), each of which has been placed upstream of both p38-MAPKs and JNKs in noncardiac cells where they regulate cell death responses. 13 ASK1 is activated by myriad cues that alter the cellular redox balance. 14,15 ASK1 is inhibited by association with thioredoxin, oxidation of which (by elevated ROS [reactive oxygen species], such as H 2 O 2 ) induces complex dissociation and ASK1 autophosphorylation of the activation loop (Thr838 in humans; Thr845 in mice/rats). ASK1 is associated with development of fibrosis in various tissues and is a therapeutic target for fibrotic diseases, including pulmonary arterial hypertension, chronic kidney disease, and nonalcoholic steatohepatitis. Indeed, ASK1 inhibitors have been developed and passed into phase-III clinical trials. 16,17 In the heart, ASK1 is activated in mouse models of pressure-overload, 14 ischemia/ reperfusion, 18 myocardial infarction, 14 and hypertension induced by Ang II (angiotensin II), 19 all of which are associated with increased ROS. Furthermore, studies in global ASK1 knockout mice demonstrate reduced cardiac cell death and remodeling in models of myocardial infarction, 20 indicating that it plays a detrimental role. How ASK1 might be involved in cardiac hypertrophy and remodeling is still far from clear. Nevertheless, therapies targeting ASK1 are in development, 21 and cardiac ASK1 is an attractive target for heart failure. 22 Here, we addressed the hypothesis that (since hypertension is associated with hypoxia and ROS) ASK1 is likely to be a prominent cardiac MAP3K in hypertension, and (because ASK1 promotes fibrosis in other tissues) its inhibition potentially reduces cardiac fibrosis. With reports that ASK1 is activated by various stimuli and signals nonselectively to p38-MAPKs and JNKs, we first clarified and delineated the ASK1 signaling pathway in the heart, establishing that ASK1 was specifically and selectively activated by moderate levels of redox stress, signaling selectively to p38-MAPK (not JNKs). ASK1, therefore, has an appropriate profile for cardiac activation in hypertension where, given its profibrotic effects in other tissues, it may promote cardiac fibrosis. Consistent with this, selonsertib (GS-4997), an ASK1 inhibitor developed as an antifibrotic agent for nonalcoholic steatohepatitis, 16 reduced cardiac fibrosis, and remodeling in mice treated with Ang II. Thus, ASK1 inhibitors represent a viable therapeutic modality for fibrosis in hypertensive heart disease.

Methodology
See the Data Supplement for a full description of materials and methods. Descriptions of cell/animal experiments are provided below. Data from this study are available from the corresponding authors upon reasonable request.
Neonatal rat ventricular myocytes were prepared and adult rat hearts perfused as described previously. 8,23,24 Cells were exposed to H 2 O 2 or IL1β at the concentrations/times indicated. In some experiments, cells were preincubated with selonsertib before treatment with H 2 O 2 or IL-1β. Hearts were equilibrated (15 minutes) and then perfused with H 2 O 2 or IL1β, or subjected to global ischemia with/ without reperfusion. In some experiments, hearts were perfused with/ without selonsertib or N-acetyl cysteine during the equilibration phase. Control hearts were perfused for the same total duration as the experimental conditions. An in vivo model of Ang II-induced hypertension (0.8 mg/[kg·d], 7d) was used to assess the effects of selonsertib (4 mg/[kg·d]) on cardiac remodeling in male wild-type C57BL/6J (10-12 week) mice. In vivo echocardiography was performed using a Vevo2100 system (Visualsonics). At the end of the experiment, mouse hearts were either perfusion fixed in situ using 10% buffered formalin and embedded for sectioning and histology, or rapidly excised and pulverized in liquid nitrogen for biochemical analyses. The use of all animals in our studies was performed in accordance with UK Animals (Scientific Procedures) Act of 1986 and the European Community Directive 86/609/EEC for animal experiments.

Statistics
Data are presented as detailed in the figure legends. Data were collected and curated in Microsoft Excel. Statistical analysis used GraphPad Prism 8.0 with t tests, 1-way or 2-way ANOVA with Holm-Sidak multiple comparison post hoc test as indicated. P<0.05 was considered statistically significant.

Activation of Cardiac ASK1 Is ROS-Dependent
Our hypothesis is that pathophysiological stresses signal through specific MAP3Ks for differential activation of p38-MAPKs versus JNKs. ASK1 is expressed in all cell types in the heart and an established ROS-sensing kinase, whereas IL-1β signals through a receptor complex, so these stimuli likely use different MAP3Ks. Activation of ASK1 compared with an alternative MAP3K (TAK1) was assessed by immunoblotting with antibodies to the activation loop phosphorylation sites. In neonatal rat cardiomyocytes, IL-1β activated TAK1, but not ASK1, whereas H 2 O 2 activated ASK1, but not TAK1 ( Figure 1A). Thus, different stimuli signal through different MAP3Ks. In perfused hearts, H 2 O 2 activated ASK1, not TAK1, consistent with the cardiomyocyte response, but IL-1β activated both kinases ( Figure 1B). We hypothesized that IL-1β may activate ASK1 via increased ROS in this context. Indeed, IL-1β significantly increased H 2 O 2 generation in perfused hearts ( Figure 1C), and perfusion of hearts with the ROS/H 2 O 2 scavenger, N-acetyl cysteine, appeared to reduce IL-1β-induced ASK1 activation ( Figure 1D) suggesting this was, indeed, the case. We conclude that ASK1 is selectively activated by ROS (which does not activate TAK1), whereas IL-1β signals more specifically through TAK1.

Hypertension
October 2020 Cardiac ASK1 Signals Selectively to p38-MAPKs, not JNKs ASK1 is reported to promote activation of both p38-MAPKs and JNKs. 14, 15 We used a selective ASK1 inhibitor, selonsertib (GS-4997), 25 to dissect the signaling pathway in cardiomyocytes ( Figure 2) and human cardiac fibroblasts ( Figure S1). Selonsertib is highly selective for ASK1 (K d 4.1 nmol/L) with just 2 off-target kinases (DYRK1A, K d 220 nmol/L; ribosomal protein S6 kinase 4, K d 430 nmol/L) that may be inhibited at higher drug concentrations, 16 neither of which is expressed at significant levels in adult hearts. 26 H 2 O 2 (1 mmol/L) activated p38-MAPK in rat neonatal cardiomyocytes within 5 minutes, whereas activation of JNKs was more pronounced after 15 minutes. Pretreatment with selonsertib inhibited activation of p38-MAPK(s) but not JNKs ( Figure 2A). Consistent with the IL-1β signal being mediated through an alternative pathway ( Figure  S2), selonsertib did not affect activation of p38-MAPKs or JNKs by IL-1β ( Figure 2B). In fibroblasts, H 2 O 2 (maximal at 100 µmol/L; Figure S1A) and IL-1β (maximal at 100 ng/ml; Figure  S1B) activated p38-MAPKs, but selonsertib only inhibited activation by H 2 O 2 ( Figure S1C). Thus, cardiac ASK1 signals selectively to p38-MAPKs in the context of redox signaling. The concentration-dependence of ASK1 activation by H 2 O 2 in cardiomyocytes was bell-shaped, with activation across a range of concentrations (0.1-10 mmol/L), maximal activation at 1 mmol/L (5 minutes), and reduced activation at higher concentrations ( Figure 3A). H 2 O 2 concentrations >1 mmol/L cause cell death, 27 but the lack of activation in this study (over 5 minutes) is not due to loss of cells because total ASK1 levels remain the same. Moreover, not all kinases become nonresponsive at higher H 2 O 2 concentrations. For example, Akt phosphorylation becomes maximal at concentrations >3 mmol/L H 2 O 2 in cardiomyocytes over the same duration. 28 Thus, the bell-shaped curve for ASK1 activation ( Figure 3A) suggests selective activation only at moderate (proapoptotic) concentrations of H 2 O 2 . Myocardial ischemia is associated with a moderate increase in ROS and activation of p38-MAPKs but not JNKs, whereas redox stress increases substantially on reperfusion and this leads to activation of JNKs. 9,29 Consistent with this, we detected ASK1 phosphorylation in perfused rat hearts subjected to ischemia and this phosphorylation, together with activation of p38-MAPKs, was suppressed by the ASK1 inhibitor, selonsertib ( Figure 3B). In contrast, with ischemia and reperfusion, we did not detect phosphorylation of ASK1 (not shown) and selonsertib did not inhibit phosphorylation of either p38-MAPKs or JNKs ( Figure 3C). Thus, ASK1 is a redox sensor in cardiomyocytes that responds only to moderate changes in ROS levels to mediate ROS-dependent redox signaling to p38-MAPKs.

ASK1 Inhibition Reduces Hypertension-Induced Cardiac Hypertrophy and Fibrosis
Studies in global knockout mice implicate ASK1 in cardiac cell apoptosis in heart failure associated with high myocardial wall stresses (eg, myocardial infarction or pressure-overload). 20 Knowledge of the role of cardiac ASK1 in models associated with more moderate myocardial wall stresses, such as hormone-induced hypertension, is limited. We used a model of hypertension induced by Ang II in C57Bl/6J male mice to assess this. We selected a moderate concentration that induces hypertension (0.8 mg/[kg·d] Ang II), 30 rather than a subpressor dose (eg, 0.288 mg/[kg·d]) 31 or one more associated with sudden cardiac death (eg, >2 mg/[kg·d]). Ang II significantly increased mRNA expression of the hypertrophic markers Nppa, Nppb, and Myh7. Selonsertib (4 mg/[kg·d]) significantly inhibited the increase in expression of Nppa and Nppb, but not Myh7 ( Figure 4A). We used echocardiography to assess cardiac dimensions and function at 3 days and then at 7 days ( Figure 4B and 4C). Selonsertib alone did not affect any variables measured ( Figure S3, Table S2). Ang II increased ejection fraction at 3 days, and this was not affected by selonsertib. At 3 and 7 days, the increase in left ventricular posterior wall thickness (but not the decrease in internal diameter) induced by Ang II in both systole ( Figure 4C) and diastole were inhibited by selonsertib suggestive of a selective effect on the structure of the myocardium. Consistent with this, Ang II significantly increased left ventricular myocardial cross-sectional area compared with vehicle controls, and this was significantly inhibited by selonsertib ( Figure 4D). Overall, these results demonstrate that ASK1 signaling plays a role in the cardiac hypertrophy resulting from hypertension.
Ang II-induced hypertension is associated with increased interstitial and perivascular fibrosis, 31 increased expression of remodeling proteins and enzymes, and elevated levels of pathological collagens. 32 Cardiac fibrosis was assessed in the left ventricular using Masson trichrome (for global collagen  deposition; Figure 5A) and picrosirius red (for fibrillar collagens; Figure 5B). Selonsertib significantly reduced cardiac fibrosis induced by Ang II, particularly of interstitial rather than perivascular fibrosis ( Figure 5A, lower). Ang II increased mRNA expression of Ddr2 (a fibrosis marker) and Tgfb (that drives fibroblast activation), extracellular remodeling-related enzymes (Timp1, Lox), a profibrotic signaling molecule (Post1), and an extracellular matrix protein (Thbs1). These were all reduced in hearts from mice treated with selonsertib, although a significant inhibition was only detected for Ddr2 and Timp1 ( Figure 5C). Ang II increased abundance of the fibrillar pathological collagens Col1a1 and Col3a1 and increased the basement membrane-linked Col4a1, with no effect on Col2a1 ( Figure 5D). As with other mRNAs assessed, selonsertib partially decreased the response to Ang II. However, selonsertib significantly reduced collagen 1 protein expression induced by Ang II (Figure 5E). These molecular findings are in accord with the reduction in cardiac fibrosis resulting from selonsertib treatment of Ang II-induced hypertension detected by histology ( Figure 5A and 5B).
Ang II-induced hypertension is associated with increased myocardial inflammation, 33 although whether this precedes or is consequential to cardiac fibrosis remains to be defined. Ang II increased mRNA expression of proinflammatory cytokines ( Figure S4) and inflammatory cell markers ( Figure  S5). Selonsertib did not affect upregulation of inflammatory cytokines but significantly reduced expression of myocarditis markers, reflective of protected myocardium. Therefore, our data strongly indicate that ASK1 operates as a profibrotic signal in the context of hypertension and is a potential therapeutic target to reduce pathological cardiac remodeling associated with this ( Figure 6).

Discussion
In recent years, ASK1 has emerged as a therapeutic target for fibrotic diseases, with ASK1 inhibitors in phase-III clinical trials for chronic kidney disease and nonalcoholic steatohepatitis. 16,17 Here, we show that ASK1 inhibitors, such as selonsertib, have potential therapeutic value in hypertensive heart disease, reducing cardiac fibrosis together with cardiomyocyte hypertrophy ( Figure 6). Thus, targeting the ASK1→p38-MAPK nexus has potential therapeutic viability as a treatment for hypertensive heart disease.
Our data with selonsertib are consistent with previous studies of mice with global ASK1 deletion subjected to a low dose of Ang II (0.288 mg/[kg·d]) in which cardiac hypertrophy and fibrosis were reduced. 31 Our study used a higher hypertensive concentration of Ang II (0.8 mg/[kg·d]), 30 resulting in greater fibrosis and hypertrophy and, still, selonsertib was effective at reducing both (Figures 4 and 5). Several studies indicate that ASK1 (as its name suggests) promotes cardiomyocyte death. Thus, global ASK1 knockout mice subjected to myocardial infarction or pressure-overload (models with much higher wall stresses than with Ang II-induced hypertension that is particularly associated with cardiomyocyte death) exhibit reduced cardiac hypertrophy and remodeling. 20 Furthermore, cardiomyocyte overexpression of ASK1 increases cardiomyocyte apoptosis in response to stressors, exacerbating progression to heart failure in myocardial infarction and pressure-overload mouse models. 14 ASK1 overexpression in cardiomyocytes does not promote hypertrophy in vivo per se, 14 yet selonsertib inhibited cardiomyocyte hypertrophy in Ang II-induced hypertension ( Figure 4). Here, it is important to consider the in vivo context taking into account the established profibrotic effects of ASK1. ASK1 inhibitors do not appear to affect blood pressure in humans (Selonsertib in adults with pulmonary arterial hypertension and Safety and Efficacy of Selonsertib in Adults With Nonalcoholic Steatohapetitis [NASH] and Bridging [F3] Fibrosis trials 16,17,34,35 ) or in a mouse model of kidney fibrosis associated with hypertension, 36 so it is unlikely that the reduction in cardiac fibrosis and cardiomyocyte hypertrophy with selonsertib seen in our studies with Ang II are due to an effect on blood pressure. It is more that the inhibition of fibrosis by selonsertib, associated with reduced paracrine signaling and inflammation, reduces cardiomyocyte workload, improving cardiomyocyte survival. The lower overall workload means that cardiomyocyte hypertrophy is also reduced.
As a ROS-sensing kinase, ASK1 is sensitive to an array of physiological and pathological processes and stimuli. Our data clearly support this because ASK1 is activated in cells in a time-and dose-dependent manner by direct redox challenge with H 2 O 2 (Figures 1 and 2). However, ASK1 was activated by moderate levels of ROS ( Figure 3) classically associated with apoptosis. 27 In contrast, high levels of oxidative stress more associated with necrotic forms of cell death did not activate ASK1 which has a bell-shaped dose-response curve. In accord with this, ASK1 was activated only by ischemia (associated with moderate levels of ROS) in perfused hearts, but not on reperfusion (with much higher levels of ROS). 29 Our data for ASK1 are consistent with previous studies of myocardial infarction in ASK1 knockout mice. 20 Here, ASK1 deletion reduced the percentage of apoptotic cardiomyocytes particularly in the peri-infarct zone, a region likely to be subjected to lower ROS levels than cells in the infarct zone itself. It is difficult to correlate the levels of exogenously applied H 2 O 2 with endogenous levels. However, IL-1β treated hearts exhibited an increase of ≈60 µmol/L H 2 O 2 per mg of protein in perfused rat hearts ( Figure 1C), so allowing for degradation of exogenous H 2 O 2 as it enters cells, together with loss of endogenous H 2 O 2 on production from the heart, the concentrations of H 2 O 2 required to activate ASK1 (Figures 1 and 2) are in an appropriate range.
Other stimuli, although not activating ASK1 directly, may still activate ASK1 via secondary production of ROS. This appears to be the case with IL-1β which activated ASK1 in an intact perfused heart in which it increased oxidative stress (Figure 1). We could not detect activation of ASK1 in cardiomyocytes at any concentration of IL-1β that we studied ( Figure S2). The reason is not clear, but the environment of cultured cardiomyocytes which have a good nutrient supply and efficient gas exchange contrasts with that of an intact beating heart in an ex vivo perfusion setting. The latter is under greater stress and subject to greater constraints with respect to oxygen supply and demand. This may deplete antioxidant reserves more rapidly and any small effect of IL-1β to increase ROS may be exaggerated to a sufficient degree that ASK1 may be activated. Other proinflammatory cytokines also activate ROS and stimulate ASK1 indirectly. For example, TNFα (tumor necrosis factor-α) signals through the ROS-generating NADPH oxidase, NADPH oxidase-1, to activate ASK1 in vascular smooth muscle cells. 37 Proinflammatory cytokines alone do not promote apoptosis in primary neonatal cardiomyocytes, 38 but the question remains of whether proinflammatory cytokines promote apoptosis via a ROS-ASK1 sensing system in compromised cells. The rheostat-like response of ASK1 to redox stress is in accord with our previous work showing that different degrees of oxidative stress differentially regulate signaling pathways and gene/protein expression. 24 For ASK1, activation by moderate levels of ROS is likely to favor regulated (less damaging) forms of cell death such as apoptosis over unregulated necrotic cell death associated with a high level of inflammation.
ASK1 is implicated in activation of p38-MAPKs or JNKs in a plethora of studies in many different cells, 39 raising a question of signal specificity. In cardiomyocytes and the heart overall, our data demonstrated that the signal from ASK1 is highly specific for p38-MAPK activation with little/no input into JNKs (Figures 2 and 3). In contrast, TAK1, as activated by IL-1β, activates both MAPK pathways (Figure 2; Figure S2), with cross-talk to ASK1 only in the intact heart consequential to an altered redox balance. One factor may be the hard-wiring of signaling networks in primary cells, particularly terminally differentiated cells, such as cardiomyocytes. However, although our data are consistent with previous studies in perfused hearts showing that ASK1 is activated in concert with p38-MAPKs, 15 they contrast with other studies of the heart suggesting that ASK1 activates JNKs. 14,40 It should be borne in mind that JNKs and p38-MAPKs are both stress-responsive, and interfering with one pathway via genetic modification may result in an increased stress that influences the other, making it difficult to dissect the signaling. For example, although cardiomyocyte ASK1 overexpression is associated with increased activation of JNKs (rather than p38-MAPKs) with pressureoverload over 1 to 8 weeks, 14 this may be secondary to the remodeling response which includes enhanced fibrosis and inflammation. In our case, by using a small molecule inhibitor, we were able to assess the acute (ie, direct) signaling pathway.
Although there is much speculation and suggestion that ASK1 inhibitors may be useful for cardiac diseases, 16,17,21 such drugs are not yet being assessed in clinical trials for this purpose. They have been tested in relation to nonalcoholic steatohepatitis, chronic kidney disease, and pulmonary hypertension in phase-II or phase-III trials, 16 so they are clearly considered safe. Our data with selonsertib support the concept that reducing ASK1 activity in the heart is a desirable strategy for reducing fibrosis and thus the workload of the heart in disease states such as hypertension and is independent of a direct effect on blood pressure. 36 Other studies with genetically modified mice indicate that ASK1 is also a strong therapeutic target in myocardial infarction in which inhibition of ASK1 may facilitate cardiomyocyte survival. A caveat is that because ASK1 is involved in fibrosis, inhibiting ASK1 may Figure 6. Schematic representation of ASK1 (apoptosis signal-regulating kinase 1) signaling in the heart. Cardiac cell stresses, such as ischemia and Ang II (angiotensin II), along with IL (interleukin)-1β to some extent, increase reactive oxygen species (ROS) signaling in the heart. These stressors or direct application of H 2 O 2 lead to ASK1 activation. In contrast, redox signaling does not activate TAK1 (transforming growth factor-activated kinase 1), which is activated directly by IL-1β. In cardiomyocytes, ASK1 signals selectively to p38-MAPKs (mitogen-activated protein kinases), which, in the context of Ang II-induced hypertension in vivo, results in cardiac fibrosis and inflammation. This, in turn, increases cardiomyocyte workload, resulting in cardiomyocyte hypertrophy. Continued stress leads to cardiomyocyte apoptosis exacerbating cardiac fibrosis/inflammation. This destructive cycle of cardiac remodeling is broken by selonsertib, a selective ASK1 inhibitor. MKK indicates MAPK kinase.
have a detrimental effect on establishing a stable fibrotic scar. It may be important, therefore, to tailor the therapy towards hypertension-associated cardiac dysfunction or other cardiac diseases with pronounced interstitial fibrosis.

Perspective
Hypertensive heart disease is a poorly controlled and progressive condition associated with myocardial fibrosis, along with cardiomyocyte dysfunction and death, cumulating in cardiac stiffening and heart failure. This study demonstrates that the ROS-sensing enzyme ASK1 represents an ideal therapeutic target for hypertensive heart disease. ASK1 inhibitors are developed as antifibrotic therapies in other diseases and may be useful for hypertensive heart disease, reducing fibrosis and, therefore, improving cardiac function. At a mechanistic level, our data define a very specific ROS→ASK1→p38-MAPK signaling modality, suggesting that p38-MAPK(s) play a significant role in development of cardiac fibrosis. Clearly, other components of the pathway constitute alternative therapeutic targets. Thus, targeting the ASK1→p38-MAPK nexus in the heart has potential therapeutic viability as a treatment for hypertensive heart disease.