Interference with ERK^Thr188 phosphorylation impairs pathological but not physiological cardiac hypertrophy

Abstract

Extracellular signal-regulated kinases 1 and 2 (ERK1/2) are central mediators of cardiac hypertrophy and are discussed as potential therapeutic targets. However, direct inhibition of ERK1/2 leads to exacerbated cardiomyocyte death and impaired heart function. We have previously identified ERK^Thr188 autophosphorylation as a regulatory phosphorylation of ERK1/2 that is a key factor in cardiac hypertrophy. Here, we investigated whether interference with ERK^Thr188 phosphorylation permits the impairment of ERK1/2-mediated cardiac hypertrophy without increasing cardiomyocyte death. The impact of ERK^Thr188 phosphorylation on cardiomyocyte hypertrophy and cell survival was analyzed in isolated cells and in mice using the mutant ERK2^T188A, which is dominant-negative for ERK^Thr188 signaling. ERK2^T188A efficiently attenuated cardiomyocyte hypertrophic responses to phenylephrine and to chronic pressure overload, but it affected neither antiapoptotic ERK1/2 signaling nor overall physiological cardiac function. In contrast to its inhibition of pathological hypertrophy, ERK2^T188A did not interfere with physiological cardiac growth occurring with age or upon voluntary exercise. A preferential role of ERK^Thr188 phosphorylation in pathological types of hypertrophy was also seen in patients with aortic valve stenosis: ERK^Thr188 phosphorylation was increased 8.5 ± 1.3-fold in high-gradient, rapidly progressing cases (≥40 mmHg gradient), whereas in low-gradient, slowly progressing cases, the increase was not significant. Because interference with ERK^Thr188 phosphorylation (i) inhibits pathological hypertrophy and (ii) does not impair antiapoptotic ERK1/2 signaling and because ERK^Thr188 phosphorylation shows strong prevalence for aortic stenosis patients with rapidly progressing course, we conclude that interference with ERK^Thr188 phosphorylation offers the possibility to selectively address pathological types of cardiac hypertrophy.

Cardiac hypertrophy has been identified as an independent risk factor of diastolic dysfunction, heart failure, arrhythmias, and sudden death (1). Various triggers initiate cardiac hypertrophy, but the type of trigger has been identified as decisive for an adaptive vs. a maladaptive outcome of cardiac hypertrophy. Whereas cardiovascular diseases such as hypertension, aortic stenosis, or myocardial infarction generally induce a pathological type of hypertrophy, chronic physical exercise or postnatal cardiac growth entail a compensatory and physiological type of hypertrophy. In contrast to physiological hypertrophy, pathological hypertrophy is characterized by accumulation of interstitial collagen and cell death, which both have been shown to contribute to increased cardiovascular risk (2–5). Therefore, it would be of great therapeutic interest to prevent pathological hypertrophy; however, such prevention requires a fine balance between inhibition of maladaptive and preservation of compensatory adaptive hypertrophy.

Hypertrophic stimuli are mediated via several intracellular signaling cascades that ultimately affect cardiac gene transcription. A central signaling cascade, which has been shown to promote the development of cardiac hypertrophy, is the mitogen-activated protein (MAP) kinase cascade consisting of the kinases rapid activation of fibrosarcoma (Raf), MAP/ERK kinase (MEK1/2), and ERK1/2. However, inhibition of ERK1/2-mediated hypertrophy using dominant-negative mutants or MEK inhibitors in isolated cardiomyocytes and in mice was accompanied by increased cardiomyocyte death (6–13). Therefore, to date, it has not been possible to exploit the potential of the ERK1/2 cascade as a therapeutic target in cardiac hypertrophy.

Here, we describe a possibility to selectively impair hypertrophic ERK1/2-mediated functions while leaving the antiapoptotic effects of ERK1/2 intact. We have previously discovered an ERK2^Thr188 autophosphorylation (ERK2^Thr188 corresponds to Thr208 in ERK1; subsequently referred to as ERK^Thr188 phosphorylation) that appears to be essential for the induction of ERK1/2-mediated cardiac hypertrophy in response to various stimuli (14, 15). Because inhibition of ERK1/2 is an effective strategy to prevent cardiac hypertrophy (6, 8, 12), we aimed to investigate whether specific interference with ERK^Thr188 phosphorylation had an effect on hypertrophy and on apoptosis in vivo. We further investigated the occurrence of ERK^Thr188 phosphorylation in patients with different forms of aortic stenosis. Our data suggest a clinical relevance of this autophosphorylation and further indicate that interference with ERK^Thr188 phosphorylation may be a selective therapeutic strategy in pathological ERK1/2-mediated cardiac hypertrophy.

Results

Interference with ERK^Thr188 Phosphorylation Attenuates ERK-Mediated Cardiomyocyte Hypertrophy but Not Cell Survival.

First, we assessed whether interference with ERK^Thr188 phosphorylation may be a strategy to attenuate cardiomyocyte hypertrophy without a simultaneous exaggeration of apoptotic cell death. For this study, we used an ERK2 mutant that cannot be phosphorylated at threonine 188, ERK2^T188A (T188A). Interestingly, overexpression of ERK2^T188A did not affect catalytic ERK activity, as monitored by MEK1/2-mediated phosphorylation of ERK1/2 at the so-called TEY motif [pERK1/2(TEY); Fig. S1 A–C] (6, 7). To exclude an overexpression artifact, we compared ERK2^T188A with similar expression levels of ERK2 wild-type (WT) [ERK2^T188T (T188T)] (Fig. S1D).

ERK2^T188A overexpression in neonatal rat cardiomyocytes (NRCMs) significantly attenuated phenylephrine-induced cardiomyocyte hypertrophy, as assessed by incorporation of [³H]isoleucine (Fig. 1A) and by cell size analysis (representative cells in Fig. 1B and quantification of cell size in Fig. S1E) (16, 17). Strikingly, ERK2^T188A did not enhance H₂O₂-induced apoptotic cell death, as determined by terminal deoxynucleotidyl transferase dUTP nick-end labeling (TUNEL) (Fig. 1C) and procaspase-3 levels (Fig. S1F) (18–20). In contrast, overexpression of ERK2^T188T affected neither the hypertrophic (Fig. S1G) nor apoptotic response of NRCMs (Fig. S1H). Thus, expression of ERK2^T188A causes an efficient inhibition of cardiomyocyte hypertrophy but has no effect on apoptosis.

Fig. 1.

Interference with ERK^Thr188 phosphorylation does not abrogate antiapoptotic effects of ERK. (A–C) NRCMs were transduced with Flag-tagged ERK2^T188T (T188T) or ERK2^T188A (T188A) and stimulated with phenylephrine (PHE) (4 µM; 24 h) (A and B) or with hydrogen peroxide (H₂O₂) (100 µM; 1 h) (C) as indicated. (A) [³H]Isoleucine incorporation assay (n = 9). *P < 0.001 vs. all other conditions. (B) Phalloidin staining. (Scale bar: 10 μm.) (C) TUNEL assay (n = 6–8; at least 150 cells per experiment and group). *P < 0.01 vs. unstimulated controls. (D–F) NRCMs were pretreated with PD98059 (30 µM; 60 min) and stimulated with phenylephrine (4 µM; 24 h) (D and E) or with H₂O₂ (100 µM; 1 h) (F) as indicated. (D) [³H]Isoleucine incorporation assay (n = 6). *P < 0.001 vs. all other conditions; not significant (n.s.) vs. unstimulated controls. (E) Phalloidin staining. (Scale bar: 10 μm.) (F) TUNEL assay (n = 7; at least 150 cells per experiment and group). *P < 0.01 vs. control; ^#P < 0.01 vs. all other conditions. (G) Analyses of procaspase-3 levels in NRCMs with (siERK1/2) or without (CON and CON siRNA) knock down of endogenous ERK1 and ERK2 expression. NRCMs were transfected with siERK1/2 or CON siRNA and adenovirally transduced with Flag-tagged ERK2^T188T (T188T), Flag-tagged ERK2^T188A (T188A), or GFP (CON). (n = 5–6). *P < 0.05. Shown are representative immunoblots.

These results are particularly interesting because, as reported in the Introduction, inhibition of the catalytic ERK1/2 activity (e.g., by using the MEK1/2 inhibitor PD98059) attenuates cardiomyocyte hypertrophy in analogous assays (Fig. 1 D and E and Fig. S1I) (21). In contrast to ERK2^T188A, PD98059 leads to enhanced H₂O₂-induced apoptotic cell death (Fig. 1F and Fig. S1J).

To ensure that ERK2^T188A overexpression does indeed not affect apoptotic ERK1/2 signaling, we next analyzed its effect on apoptosis at endogenous expression levels. For these experiments, we used siRNAs to reduce endogenous ERK1/2 expression (Fig. 1G, middle blot). Like ERK1/2 inhibition with PD98059 (Fig. S1K), knock down of ERK1/2 activated the programmed cell death machinery, as evidenced by decreased procaspase-3 levels (Fig. 1G, upper blot and graph). Accordingly, subsequent reexpression of WT ERK2 (T188T) reversed the apoptotic effect of the ERK1/2 knockdown (i.e., normalized the procaspase-3 expression levels; Fig. 1G), confirming the antiapoptotic role of endogenously expressed ERK1/2. Interestingly, reexpression of ERK^T188A normalized procaspase-3 expression levels as efficiently as did reexpression of ERK^T188T (Fig. 1G).

Taken together, our experiments confirmed that ERK1/2, indeed, mediate both cardiomyocyte hypertrophy and survival in NRCMs. Our experiments further suggest that interference with ERK^Thr188 phosphorylation—in contrast to catalytic inhibition of ERK1/2—might be a strategy to specifically target ERK-mediated hypertrophy without impairing ERK-mediated cell survival.

Inhibition of Hypertrophy by ERK2^T188A Does Not Exacerbate Apoptosis in Vivo.

We then sought to confirm these findings obtained in isolated NRCMs in intact animals. For this study, we used FVB/N mice with transgenic cardiac expression of ERK2^T188A (T188A), and as controls, we used nontransgenic WT mice, as well as mice transgenic for ERK2 WT (T188T). ERK2 constructs were transgenically expressed at similar levels as the overall expression level of endogenous ERK1 and ERK2 (Fig. S2A). These mice were subjected to chronic pressure overload stress induced by transverse aortic constriction (TAC) to induce hypertrophy and apoptosis (22). Aortic pressure gradients were the same in the three groups: 70.7 ± 4.3 for WT, 72.5 ± 6.0 for ERK2^T188T, and 71.9 ± 2.6 for ERK2^T188A.

Transgenic cardiac expression of ERK2^T188A significantly reduced the hypertrophic response to TAC compared with control mice (WT or T188T), as measured by heart weight-to-tibia length ratios (Fig. 2A), LV wall thickness (Table S1), and cardiomyocyte size (Fig. 2B and Fig. S2B). Furthermore, echocardiographic analyses showed normal cardiac function in ERK2^T188A-expressing mice (Table S1). Sirius red staining of histological sections revealed that ERK2^T188A-overexpressing mice developed even less interstitial fibrosis in response to LV pressure overload (Fig. 2 C and D). Interestingly, whereas ERK-mediated cardiac hypertrophy was efficiently inhibited in ERK^T188A transgenic mice, the extent of apoptosis induced by TAC was comparable in all three genotypes (Fig. 2 E and F).

Fig. 2.

Inhibition of hypertrophy by phosphorylation-deficient ERK2^T188A does not exacerbate apoptosis in vivo. (A–F) WT mice and transgenic mice expressing ERK2^T188T (T188T) or ERK2^T188A (T188A) without (CON) or with surgically induced TAC (6 wk). Analysis of heart weight-to-tibia length ratios (A), histological H&E-stained sections (scale bar: 50 μm) (B), quantification (C) and representative images (D) of interstitial fibrosis in Sirius red-stained sections of ventricular myocardium (scale bar: 200μm), caspase-3 activity (E), and TUNEL-positive cell nuclei (F) (n = 5–14 mice per group). *P < 0.01; ^#P < 0.01 vs. unstimulated controls.

Thus, interference with ERK^Thr188 phosphorylation permits efficient inhibition of TAC-induced cardiac hypertrophy and cardiac remodeling, with no adverse effects on cell survival and cardiac function.

ERK2^T188A Prevents Activation of Nuclear, Hypertrophic but Not Cytosolic Antiapoptotic ERK1/2 Signaling.

To understand the mechanism that enables ERK2^T188A to distinguish between the different ERK functions (i.e., hypertrophy vs. apoptosis), we compared its effects on ERK1/2 target phosphorylation, looking at both cytosolic and nuclear targets. As ERK1/2 targets, which mediate the antiapoptotic effects of ERK1/2 and are, thus, primarily located within the cytosol, we investigated p90 ribosomal S6 kinase (p90RSK) and B-cell lymphoma 2-interacting mediator of cell death (BIM) (23). As a nuclear target contributing to the transcriptional hypertrophic effects of ERK1/2, we analyzed ERK1/2-mediated phosphorylation of the nuclear ETS (E twenty-six)-like 1 transcription factor (Elk1) (24). Both PD98059 and ERK2^T188A expression inhibited phenylephrine-induced phosphorylation of nuclear Elk1, but only PD98059 interfered with phenylephrine- or H₂O₂-induced phosphorylation of cytosolic p90RSK and BIM (representative blots in Fig. 3 A and B; quantification in Fig. S3 A–F and G–L).

Fig. 3.

Interference with ERK^Thr188 phosphorylation specifically attenuates nuclear ERK-target activation. (A and B) Western blot analyses of NRCMs adenovirally transduced with Flag-tagged WT ERK2^T188T (T188T) or Flag-tagged phosphorylation-deficient ERK2^T188A (T188A) or pretreated with PD98059 (30 µM; 60 min) as indicated. Cells were stimulated without (CON) or with phenylephrine (PHE) (4 µM; 10 min) or H₂O₂ (100 µM or 400 µM; 1 h). (A and B) Analysis of nuclear (Elk1) and cytosolic (p90RSK and BIM) ERK1/2 targets using phospho-specific antibodies directed against ERK1/2 phosphorylation sites of these target proteins. (C) Representative confocal images of NRCMs transduced with YFP-tagged T188T or YFP-T188A stimulated without or with phenylephrine (4 µM; 10 min) or H₂O₂ (100 µM; 1 h). Nuclei were stained with DAPI. (Scale bar: 20 μm.) (D) Immunoblot analysis of ERK^Thr188 phosphorylation of mock-transfected NRCMs stimulated with phenylephrine (4 µM; 10 min) or H₂O₂ (100 µM; 60 min). All experiments were reproduced at least four times.

These experiments suggest that interference with ERK^Thr188 phosphorylation (in contrast to inhibition of ERK1/2 activity by PD98059) selectively inhibits nuclear ERK targets, which may be explained by a prominent role of ERK^Thr188 phosphorylation in nuclear accumulation of ERK1/2: (i) the ERK2^T188A mutant did not—unlike WT ERK2—accumulate within the nucleus in response to a hypertrophic stimulus (Fig. 3C); and (ii) WT ERK2 did not accumulate in the nucleus in response to H₂O₂ (Fig. 3C), which activated ERK1/2 [i.e., pERK1/2(TEY)] but did not induce ERK^Thr188 phosphorylation (representative immunoblots in Fig. 3D and quantification in Fig. S3 M–P).

ERK2^T188A Does Not Affect Physiological Growth of the Heart.

Even though cardiac hypertrophy is a risk factor for heart failure, therapeutic prevention of cardiac hypertrophy in patients should ideally impair neither the heart’s normal postnatal growth nor its adequate growth in response to exercise (3, 4). Therefore, we next investigated the influence of ERK2^T188A expression on physiological growth of the heart. For this analysis, we assessed cardiac growth (i) in mice until the age of 9 mo and (ii) in mice that exercised voluntarily in a running wheel for a period of 3 wk.

Postnatal cardiac growth up to 9 mo, as well as exercise-driven cardiac growth, was indistinguishable between ERK2^T188A-expressing and control mice (WT and T188T), as assessed by ventricular wall thicknesses (Fig. 4A), cardiomyocyte size (representative sections in Fig. 4B and quantification in Fig. 4C), and heart weight-to-tibia length or heart weight-to-body weight ratios, respectively (Fig. S4 A and B). Furthermore, echocardiographic analysis of heart function (Table S2) and histological analysis (representative sections in Fig. 4D and quantification in Fig. 4E) did not reveal any pathological changes in ERK2^T188A-expressing mice at the age of 9 mo or after 3 wk of running-wheel exercise. Thus, interference with ERK^Thr188 phosphorylation does not appear to interfere with physiological and healthy growth of the heart. In line with these findings, ERK^Thr188 phosphorylation, which is necessary for ERK1/2-mediated cardiac hypertrophy (14), was elevated only in TAC-operated mice but not in exercised mouse hearts (Fig. 4F).

Fig. 4.

ERK^Thr188 phosphorylation is not involved in physiological hypertrophy. (A–E) Echocardiographically determined interventricular wall thickness (IVS) (A), representative histological H&E-stained sections (scale bar: 50 μm) (B), quantification of cross-sectional cardiomyocyte areas (C), corresponding Sirius red-stained sections (scale bar: 200 μm) (D), and quantification of interstitial fibrosis of ventricular myocardium (E) of WT mice and transgenic mice with cardiac expression of wild-type ERK2^T188T (T188T) or phosphorylation-deficient ERK2^T188A (T188A) at the age of 8 wk (CON), at the age of 9 mo (9 MON), and after 3 wk of voluntary running-wheel exercise (RUN) (n = 7–15). *P < 0.05 vs. CON. (F) Representative immunoblots of ERK1/2 phosphorylations in heart lysates of 8-wk-old WT mice, heart lysates of mice after 3 wk of voluntary running-wheel exercise (RUN), or after 6 wk of pressure overload (TAC) of n = 9 experiments. Bar graph shows the quantification of pERK(Thr188) from n = 9 experiments. *P < 0.001 vs. all other conditions.

Differential ERK^Thr188 Phosphorylation in Patients with High- and Low-Gradient Aortic Stenosis.

To further characterize the involvement of ERK^Thr188 phosphorylation in cardiac hypertrophy, we analyzed septum samples of patients with aortic valve stenosis. These patients were subgrouped into high (≥40 mmHg) or low (<40 mmHg) transvalvular pressure gradients, which showed low or high degrees of interstitial fibrosis, respectively (Fig. 5A) (25). Despite different transvalvular pressure gradients, both groups had a similar extent of hypertrophy at the time of surgery, as determined by echocardiographic analysis of heart weight (Table S3), diastolic wall thickness (Table S3), and histological analysis of cardiomyocyte cell area (Fig. 5B). At the time of surgery, ejection fractions and aortic valve areas were also undistinguishable in both patient groups (Table S3). Furthermore, canonical ERK1/2 activation (pERK[TEY]; Fig. 5C) was similar in both patient groups. Interestingly, however, ERK^Thr188 phosphorylation was >eightfold increased in patients with high transvalvular pressure gradients but not significantly (<twofold) enhanced in patients with low gradients (Fig. 5C). Of these patients, 10/12 with high pressure gradients, but only 1/8 with low gradients, had a rapid course of disease (Fig. 5D and Table S3). This indicates a close correlation between pressure gradients, rapid progression, and ERK^Thr188 phosphorylation, even though the extent of ventricular hypertrophy at the time of surgery was similar in all patients (Table S3, Fig. 5B, and Fig. S5). Thus, in contrast to similar canonical ERK1/2 activation in all samples, increased ERK^Thr188 phosphorylation was seen specifically in the cases with rapid progression and high gradients (Fig. 5C). This indicates that ERK^Thr188 phosphorylation depends on the type of hypertrophic trigger and seems to be correlated with the rapidity of disease development (Fig. 5D), rather than with its final extent of hypertrophy (Fig. S5).

Fig. 5.

Differential ERK^Thr188 phosphorylation in patients with high- and low-gradient aortic stenosis. Analysis of septum samples of patients with aortic valve stenosis subgrouped in patients with high (≥40 mmHg; high-gradient) or low (<40 mmHg; low-gradient) transvalvular pressure gradients. (A) Masson–Goldner staining of interstitial fibrosis. (Scale bar: 200 μm.) (B) Analysis of cardiomyocyte cell area. (C) Immunoblot analysis of ERK^Thr188 phosphorylation (pERK[Thr188]) and ERK1/2 activation in patients with low and high gradients compared with nonfailing control myocardium, as monitored by canonical ERK1/2 phosphorylation (pERK[TEY]) (n = 5–12 samples per group). *P < 0.05. (D) Speed of disease progression in patients with high- or low-gradient aortic stenosis. Rapid progress is defined by an increase in peak aortic flow velocity ≥0.3 m⋅s⁻¹y⁻¹.

Taken together, the analyses of patients with aortic valve stenosis further substantiate the hypothesis that ERK^Thr188 phosphorylation is not increased in all types of cardiac hypertrophy but is dependent on the type of hypertrophic trigger and is, in particular, associated with high pressure load in mice and with rapidly progressing forms of aortic stenosis in patients.

Discussion

Pathological hypertrophy is a key risk factor for myocardial infarction, arrhythmia, and sudden death, whereas physiological growth of the heart in response to exercise is generally accepted to be protective, i.e., to preserve or even improve heart function (1, 4, 26). Although maladaptive hypertrophy involves cell death, collagen deposits, fibrosis, and ventricular stiffening, these hallmarks of pathological remodeling are absent in the “athlete’s” heart (1, 27). Although there is an ongoing debate whether heart failure-associated ventricular hypertrophy has a compensatory role in heart failure patients, evidence is increasing that inhibition of pathological hypertrophy may be beneficial (3, 14, 15, 28–30).

On the molecular level, pathological cardiac hypertrophy is mediated by various proteins including, Gα_q, extracellular signal-regulated kinases, protein kinase C, Ca²⁺/calmodulin, and nuclear factor of activated T cells. Even though these proteins are associated with pathological hypertrophy, many of them also play vital roles in normal cardiomyocyte functions (4–7, 31, 32). Therefore, it is a challenging but critical task to identify features within pathological hypertrophic signaling pathways that allow to specifically target maladaptive functions of candidate proteins.

The ERK1/2 cascade has been reported to be activated and to be involved in diverse instances of cardiac hypertrophy and heart failure including cardiomyopathies, both in rodents and humans, and has, thus, been considered as a putative therapeutic target in pathological hypertrophy (6, 33–35). However, many mouse models point toward a pivotal role of ERK1/2 in the prevention of apoptotic cell death: cell death is significantly increased in mice with cardiac overexpression of dominant-negative Raf1, in mice treated with MEK1/2-inhibitors, and in ERK1^−/−ERK2^+/− mice (11, 12, 36), whereas overexpression of constitutively active MEK1 increases cell survival (37, 38). Therefore, inhibition of ERK1/2 activity is detrimental for the heart and, thus, cannot be used to target cardiac hypertrophy.

Our present study identifies ERK^Thr188 phosphorylation as a tool to differentiate between these distinct functions of ERK1/2 in the heart: increased cell death as a maladaptive and reduced hypertrophy as a potentially beneficial consequence of ERK1/2 inhibition. With the help of ERK2^T188A, the impact of ERK^Thr188 phosphorylation on cardiomyocyte hypertrophy in vitro and in vivo was demonstrated. Expression of ERK2^T188A attenuated the hypertrophic responses of pathological stimuli such as phenylephrine, angiotensin II, isoprenaline, and chronic pressure overload (14, 15). However, protective, antiapoptotic ERK1/2 signaling in vitro and in vivo (Fig. 1C and Fig. 2 E and F), as well as cardiac function (Tables S1 and S2), were not disturbed by ERK2^T188A expression. Therefore, ERK2^T188A appears to display its dominant-negative effect on nuclear ERK1/2 signaling via dimerization with endogenous WT ERK1/2 and by retaining them within the cytosol. This hypothesis is supported by the following: (i) the ability of ERK2^T188A to dimerize with ERK2 (ref. 14 and Fig. S6 A and B); (ii) the ability of ERK2 to heterodimerize with ERK1 (ref. 15 and Fig. S6 A and B); (iii) the cytosolic retention of YFP-tagged WT ERK2 in the presence of ERK2^T188A (Fig. S6C); and (iv) earlier observations that both ERK dimerization and ERK^Thr188 phosphorylation are required for nuclear accumulation (15, 39).

Interestingly, ERK^Thr188 phosphorylation may be controlled by β-arrestins, interaction partners of ERK1/2. Overexpression of β-arrestin2 prevented the interaction of ERK2 with Gβγ subunits, an interaction that is necessary to induce ERK2^Thr188 phosphorylation (Fig. S6 F–I) (40, 41). β-Arrestin2 may, thus, be an endogenous competitor of ERK^Thr188 signaling.

ERK^Thr188 phosphorylation seems to be a prerequisite for nuclear accumulation of ERK1/2, and, therefore, ERK^Thr188 phosphorylation opens up a possibility to interfere with ERK-mediated cardiac hypertrophy without inhibiting catalytic ERK1/2 activity. Targeting of ERK^Thr188 phosphorylation may, therefore, allow to address the Raf/MEK/ERK1/2 cascade as a promising therapeutic target in cardiac hypertrophy. Of interest in this context, interference with ERK^Thr188 phosphorylation by expression of ERK2^T188A did not interfere with physiological (postnatal or exercise-driven) growth of the heart, and even long-term interference with ERK^Thr188 phosphorylation did not reveal any cardiac adverse effects (Fig. 4 and Table S2). However, because this mechanism appears to be of a general nature, it will be important to delineate it in more detail in both cardiac and noncardiac cells and tissues.

The analysis of septal samples of patients with aortic valve stenosis further confirms that ERK^Thr188 phosphorylation is a marker for pathological hypertrophy: ERK^Thr188 phosphorylation was significantly increased in patients with high pressure gradients and with a rapid course of the disease (Fig. 5 C and D).

Interestingly, in patients with aortic valve stenosis, we found a negative correlation for fibrosis and ERK^Thr188 phosphorylation. Low-gradient patients had a significantly higher degree of interstitial fibrosis than those with high-gradient stenosis (Fig. 5A), but ERK^Thr188 phosphorylation was only significantly up-regulated in high-gradient patients (Fig. 5C). This observation essentially excludes the possibility that up-regulation of ERK^Thr188 phosphorylation in myocardial biopsies may originate from increased numbers or activation of cardiac fibroblasts attributable to increased interstitial fibrosis (42). Further clinical studies should aim to delineate whether these differences in ERK^Thr188 phosphorylation are simply attributable to the mechanical differences (gradient) or whether the two forms are different clinical entities.

Future studies will have to establish how these effects of ERK^T188A may be therapeutically exploited. Although gene therapy with ERK^T188A may be a possibility, it would seem most attractive to search for small molecule inhibitors that specifically interfere with ERK^Thr188 phosphorylation. Because the antiapoptotic effects of ERK1/2 depend on their enzymatic/catalytic activity (Fig. 1F), mechanistic prerequisites for ERK^Thr188 phosphorylation may represent interesting strategies for specific interference with ERK^Thr188 phosphorylation. ERK^Thr188 phosphorylation is dependent on (i) dimerization of activated ERK1/2 and (ii) subsequent binding of Gβγ subunits to dimeric ERK1/2 (14, 15). Thus, inhibition of ERK1/2 dimerization or inhibition of the ERK/Gβγ interaction might represent possibilities to specifically interfere with ERK^Thr188 phosphorylation and, thus, with ERK1/2-mediated hypertrophy without affecting the antiapoptotic ERK1/2 signaling in the heart. To date, these approaches have been positively tested with overexpressed peptides and with mutant proteins in cardiomyocytes in vitro and/or in vivo and await molecular modeling of small molecule inhibitors (14, 15). Such experiments will have to consider that ERK1/2 mediate a panoply of functions within the body and that such targeting strategies may interfere with pivotal ERK1/2 functions other than pathological hypertrophy.

Taken together, our results show that specific interference with ERK^Thr188 phosphorylation permits differential targeting of distinct ERK1/2 effects (hypertrophy vs. cell survival), as well as of distinct types of hypertrophy (pathological vs. physiological). Selective occurrence of ERK^Thr188 phosphorylation in patients with rapidly progressing, high-gradient aortic valve stenosis further suggests ERK^Thr188 phosphorylation as a favorable target in pathological cardiac hypertrophy.

Materials and Methods

An expanded section on materials and methods is available in SI Materials and Methods.

Mice, Rats, and Cardiomyocytes.

The generation of transgenic mice overexpressing ERK2^T188A (T188A) and ERK2^T188T (T188T) under the control of the mouse α-myosin heavy chain promoter was described previously (14). For all experiments, male mice with an FVB/N background were used and as control isogenic WT FVB/N mice. Pregnant Sprague–Dawley rats were purchased from Janvier. NRCMs were prepared from left ventricles of 1- to 2-d-old Sprague–Dawley rats as described previously (14, 15, 43). Care of the animals was performed in accordance with the Committee on Animal Research of the regional government (Regierung von Unterfranken, Würzburg, Germany), who reviewed and approved all experimental protocols (Az.54-2531.01-62/06 and Az.55.2-2531.01-20/10) according to the corresponding national legislation.

Analysis of Patients with Aortic Valve Stenosis.

Twenty patients were regularly followed from moderate (aortic valve area, 1.0–1.5 cm²) to severe symptomatic aortic valve stenosis (aortic valve area, <1.0 cm²) and underwent transthoracic echocardiography at least 1 y before and within 3 wk before aortic valve replacement. Patients were grouped in high-gradient (≥40 mmHg; n = 12) and low transvalvular pressure gradient (<40 mmHg; n = 8) (25). Rapid progress of the stenosis was defined as an increase in peak aortic flow velocity ≥0.3 m⋅s⁻¹⋅y⁻¹ (44). Echocardiography was performed using the Vivid7 System (GE Vingmed Ultrasound) with a 3.5-NHz transducer. Biopsy samples of the basal septum were taken during aortic valve replacement. Informed consent was obtained from all patients. The study was reviewed and approved by the Ethics Committee of the Medical Faculty of the University of Würzburg and was conducted according to the principles outlined in the Declaration of Helsinki.

Statistical Analysis.

Statistical significance was evaluated by one-way ANOVA using GraphPad Prism software. Bonferroni’s test was used as post hoc test, with P < 0.05 regarded as significant. All data are shown as means ± SEM or as indicated. Individual experiments were reproduced at least three times. All histological and echocardiographic evaluations were performed in a blinded fashion.