Key role of ERK1/2 molecular scaffolds in heart pathology

Abstract

The ability of cardiomyocytes to detect mechanical and humoral stimuli is critical for adaptation of the myocardium in response to new conditions and for sustaining the increased workload during stress. While certain stimuli mediate a beneficial adaptation to stress conditions, others result in maladaptive remodelling, ultimately leading to heart failure. Specific signalling pathways activating either adaptive or maladaptive cardiac remodelling have been identified. Paradoxically, however, in a number of cases, the transduction pathways involved in such opposing responses engage the same signalling proteins. A notable example is the Raf–MEK1/2–ERK1/2 signalling pathway that can control both adaptive and maladaptive remodelling. ERK1/2 signalling requires a signalosome complex where a scaffold protein drives the assembly of these three kinases into a linear pathway to facilitate their sequential phosphorylation, ultimately targeting specific effector molecules. Interestingly, a number of different Raf–MEK1/2–ERK1/2 scaffold proteins have been identified, and their role in determining the adaptive or maladaptive cardiac remodelling is a promising field of investigation for the development of therapeutic strategies capable of selectively potentiating the adaptive response.

Introduction

Heart failure is a complex disease due to inherited gene mutations or to acquired abnormalities in heart function mainly due to chronic exposure to work overload. This condition is usually detected by the heart, with the subsequent activation of a process called adaptive heart remodelling, a complex and coordinated reaction of the entire organ that includes cardiomyocyte hypertrophy, metabolic changes, and modifications in calcium handling and contractile function. The vascular bed and stromal tissue also undergo important remodelling to support the increased metabolic need and increased mechanical strain of the organ. This initial adaptive response can, however, progress toward contractile dysfunction and heart dilation due to cardiomyocyte death, inflammatory reactions and heart fibrosis. Heart remodelling is triggered by the activation of mechanical signals, cytokine growth factors and neurohumoral stimuli. While some of these stimuli mainly mediate an adaptive heart remodelling response, others result in maladaptive changes and the actual myocardial function is determined by the balance between these different stimuli [1]. This is due to the fact that different stimuli activate specific intracellular signalling cascades, leading to unique responses in term of cardiac cell metabolism, hypertrophy and survival. Numerous intracellular signalling pathways have been characterized as being involved in heart remodelling. Analyses of genetically modified mice overexpressing or lacking specific signalling components have greatly contributed to the concept that heart remodelling can be beneficial or maladaptive depending on the specific pathways involved [1–3]. It is well known that neurohumoral stimuli, such as angiotensin and endothelin-1, via Gq-coupled seven transmembrane receptor signalling, trigger a maladaptive cardiac hypertrophic response, leading to heart failure [4, 5]. Conversely, the PI3 K/AKT/GSK3β signalling cascade, activated by growth factors, such as IGF-1, via tyrosine kinase receptors, is known to induce a compensatory hypertrophic response in the absence of fibrosis and systolic dysfunction, and protect cardiomyocytes from apoptotic death [6–9].

Current understanding of the biochemical pathways triggered by different stress stimuli on the heart indicates that, in some instances, the same signalling molecules are activated both during adaptive and maladaptive remodelling generating confusion in the comprehension of the molecular mechanisms involved in cardiac remodelling and casting doubts on the possible uses of these signalling molecules as drug targets [10, 11]. The Ras-Raf–MEK1/2–ERK1/2 MAPK signalling pathway is a paradigmatic example of this dual role as it is involved both in maladaptive and beneficial remodelling.

In the ERK1/2 pathway, Ras GTPase stimulates the activity of Raf kinase, which phosphorylates MEK1/2 that sequentially phosphorylate and activate the downstream kinases ERK1/2. Once activated, ERK1/2 phosphorylate multiple intracellular targets, both in the cytoplasm and in the nucleus. Cytoplasmic substrates include approximately 70 proteins, among them EGFr, SOS, RSK1, cPLA2, p70 S6 kinase and phosphodiesterase 4D (PDE4) [12]. In the nucleus, ERK1/2 phosphorylates numerous transcription factors that induce reprogramming of cardiac gene expression.

Many reports in the literature support the beneficial role of ERK1/2 signalling in the heart. For instance, raf-1 null mice show left ventricular systolic dysfunction and heart dilatation without cardiac hypertrophy, and increased cardiomyocyte apoptosis [13]. Accordingly, transgenic mice with cardiac-specific expression of a dominant negative form of Raf-1 show blunted hypertrophic remodelling, high levels of cardiomyocyte apoptosis and increased mortality in response to pressure overload [14]. Furthermore, transgenic mice expressing activated MEK1 in cardiomyocytes develop stable concentric hypertrophy devoid of interstitial cell fibrosis [15]. In another study, Erk1 null and Erk2^−/+ heterozygous mice display heart decompensation and failure after long-term pressure overload due to an increase in cardiomyocyte apoptosis [16].

On the other hand, other studies point to ERK1/2 as a maladaptive signalling pathway in heart remodelling. Expression of a constitutively active Ras in the mouse heart promotes hypertrophic cardiomyopathy characterized by myofilament disarray and interstitial fibrosis [17, 18]. However, given that Ras activates several signalling cascades in addition to the MEK–ERK1/2 pathway, the involvement of other pathways cannot be excluded [19]. A well-known negative effect of ERK1/2 signalling in human heart function is highlighted by the fact that different mutations able to increase ERK1/2 pathway activationlead to cardiac pathologies in patients with Noonan and related syndromes, such as Costello, LEOPARD and CFC (cardio-facio-cutaneous) syndromes [15, 16, 20–25]. Knock-in mice expressing the Noonan syndrome-associated Raf1 ^L613V mutation exhibit enhanced ERK1/2 signalling, eccentric cardiac hypertrophy in basal conditions and an accelerated transition toward heart failure in response to pressure overload. Interestingly, postnatal treatment with MEK inhibition normalizes cardiac defects [26]. A further example is represented by the mouse model of Emery–Dreifuss muscular dystrophy, carrying a mutation in the A-type lamin gene and causing dilated cardiomyopathy. These mice show abnormal activation of ERK1/2 signalling in the heart [27] and treatment with MEK inhibitors prevents the development of left ventricular dilatation [28] and improves heart performance if administered after the onset of cardiac disease [29]. Furthermore, Lorenz et al. demonstrated that G protein-coupled receptor-induced ERK1/2 autophosphorylation of Thr188 causes localization of ERK1/2 to the nucleus, thus enhancing phosphorylation of nuclear targets of ERK1/2 [30]. Interestingly, Thr188 phosphorylation of ERK1/2 does not affect its kinase activity, but does mediate ERK1/2 nuclear accumulation. Transgenic mice expressing a Thr188 phosphomimetic Erk2 mutant (Erk2^T188D) are normal under basal conditions, but show enhanced hypertrophic remodelling in response to pressure overload associated with reduced fractional shortening and cardiac dysfunction [30].

The results discussed above clearly demonstrate that activation of the ERK1/2 pathway can result in either beneficial or maladaptive cardiac remodelling, which can be ascribed to different modes of activation.

ERK1/2 MAPK scaffold proteins in the heart

ERK signalling requires the spatial and temporal organization of three different kinases, Raf, MEK1/2 and ERK1/2, which act in sequential cascade, an event controlled by scaffold molecules. A number of scaffold molecules regulating the Raf–MEK1/2–ERK1/2 pathway have been identified [31]. These include β-arrestin, Gβγ subunits, the four-and-a-half LIM domain- protein 1(FHL1), Ras GTPase-activating-like protein 1 (IQGAP-1), kinase suppressor of Ras 1 (KSR1), mitogen-activated protein kinase organizer 1 (MORG1) and MEK Partner 1 (MP1), paxillin and G protein-coupled receptor (GPCR)-kinase interacting protein-1 (GIT1) [32] and dual-specificity tyrosine-phosphorylated and regulated kinase 1A (DYRK1A) [33].

A major function of molecular scaffolds is to bind different components of the signal transduction pathway keeping these molecules in the location and orientation needed for their reciprocal interactions and activation/deactivation. A second important function of scaffold molecules is to localise the signalling complexes to a given cellular sub-compartment in order to place them in close proximity to specific effectors and, therefore, to achieve signalling specificity. Moreover, scaffold molecules can often bind proteins of multiple signalling pathways, thus, contributing to the crosstalk between pathways and the organisation of the signalling network typical of a specific stimulus in a given cellular context [34].

Some of the Raf–MEK1/2–ERK1/2 scaffold molecules mentioned above have been studied for their role in cardiac tissue, while others, despite being expressed in the heart, have mainly been characterised in other organs.

Here, we discuss the involvement of scaffold proteins in mediating the assembly and spatiotemporal activation of specific Raf–MEK1/2–ERK1/2 signalling complexes and their impact on beneficial or maladaptive cardiac remodelling.

β-arrestins and Gβγ subunits

The seven transmembrane receptors, such as angiotensin and adrenergic receptors, activate the ERK1/2 signalling by different mechanisms involving at least two different scaffold proteins: a G protein-independent mechanism involving the β-arrestins [35, 36] and a G protein-dependent pathway involving Gβγ subunits [30, 37]. An additional G protein-dependent mechanism involves the Gα subunit which can activate the ERK1/2 pathway via classical GPCR downstream effectors, such as PKA and PKC [38, 39], possibly through the FHL1 scaffold protein [41], as will be discussed below.

β-arrestin 1 and 2 are multifunctional adaptor proteins, classically known for their role in G protein-coupled receptor (GPCR) desensitisation. In particular, recruitment of β-arrestins to the GPCR prevents further G protein coupling of the receptor, thereby leading to desensitisation [36]. β-arrestins also function as a scaffold, linking GPCR activation to several downstream effectors such as Src, the ubiquitin ligase, Mdm2 and MAPK cascades (ERK and JNK) [41]. In particular, β-arrestins bind Raf–MEK1/2–ERK1/2, aligning them in the appropriate orientation to activate each other [35]. β-arrestin–mediated scaffolding directs ERK1/2 activity to their cytoplasmic substrates, protecting cells from apoptosis [42]. Based on this property, it has been suggested that β-arrestins are cardioprotective in two different ways, namely by activating a G protein independent ERK1/2 pathway and by blocking the deleterious G protein pathway [43].

Gβγ complex downstream of the seven transmembrane receptors can also act as a Raf–MEK1/2–ERK1/2 scaffold and mediates the G protein-dependent ERK1/2 maladaptive signalling [30]. G scaffolding mediates ERK1/2 autophosphorylation at a specific site, Thr188, which is ultimately linked to detrimental remodelling, characterised by excessive hypertrophy and ventricle dilation in response to pressure overload [30, 37]. Thr188 phosphorylation strongly increases ERK1/2 nuclear localisation, raising the possibility that the ERK1/2 scaffold-driven nuclear localisation is at least partially responsible for maladaptive remodelling.

The four-and-a-half LIM domains protein 1 (FHL1)

The FHL family, composed of FHL 1, 2, 3, 4 and ACT, is a group of LIM-only proteins characterised by an N-terminal half LIM domain followed by four complete LIM domains [44, 45]. FHL1 is highly expressed in striated muscles and moderately expressed in cardiac muscle and other tissues [44]. FHL1 mutations have been found in hypertrophic cardiomyopathies [46] and are responsible for four distinct skeletal muscle diseases [44]. Moreover, FHL1 is upregulated in mouse hearts in response to pressure overload [47, 48] as well as in hearts of human patients with hypertrophic cardiomyopathy [49–51]. From a molecular point of view, FHL1 plays an important role in biomechanical stress responses involved in cardiac hypertrophy. FHL1 deficiency in mice prevents both pressure overload-induced and Gαq-mediated cardiomyopathy, suggesting that FHL1 coordinates maladaptive signalling downstream of Gαq [40]. FHL1 functions as a scaffold for the Raf–MEK1/2–ERK1/2 pathway in the heart by regulating its subcellular localisation at the I band and by regulating intensity and/or duration of ERK1/2 activation in response to pressure overload [40], negatively impacting on muscle compliance [52]. In summary, FHL1-mediated Gαq-dependent ERK1/2 activation leads to detrimental hypertrophic cardiac remodelling.

Ras GTPase-activating-like protein 1 (IQGAP-1)

IQGAP-1 is a ubiquitously expressed multi-domain protein originally identified as a putative GAP for CDC42. More recently, it has been shown that IQGAP-1 binds to Raf, MEK1/2 and ERK1/2, functioning as a scaffoldand activating the ERK1/2 pathway in response to EGF in fibroblasts and epithelial cells [53, 54]. Indeed, in the heart, IQGAP-1 is expressed both in cardiomyocytes and in fibroblasts and acts as a Raf, MEK1/2 and ERK1/2 scaffold molecule required for activating a temporally delayed wave of ERK1/2 signalling in response to pressure overload [55]. Interestingly, ERK1/2 activation at early time points after pressure overload is independent from IQGAP-1 [55], underlying the importance of scaffold proteins in determining the temporal specificity of signalling pathway activation. Moreover, IQGAP-1 is also able to bind Akt [56], and is required for Akt activation in response to pressure overload [55], highlighting its ability to orchestrate different signalling pathways. Overall, the absence of IQGAP-1 causes an accelerated transition towards left ventricle dilation and impaired contractility, indicating a role for IQGAP-1 in sustaining functional remodelling upon chronic pressure overload [55].

Interestingly, IQGAP-1 not only binds c-Raf, MEK1/2 and ERK1/2 in the heart but also interacts with the heat shock protein, Hsp90, and melusin, a muscle-specific small molecular chaperone regulating signal transduction [57, 58]. Chaperone molecules play an important role in assisting conformational changes required for activation and stabilisation of signalling proteins [59], suggesting that a cooperation of scaffolds and chaperones is required to organise a fully active signalosome complex.

Additional Raf-MEK-ERK scaffolds

The ERK scaffold proteins briefly discussed below have been characterised in several cellular systems, but their role in cardiac remodelling has not yet been addressed, despite the fact that they are expressed in the heart.

Kinase suppressor of Ras 1 (KSR1)

KSR1 is one of the best characterised scaffold proteins of the ERK pathway in yeast as well as in mammalian cells [31, 60]. In resting cells, KSR1 is retained in the cytoplasm and interacts with 14-3-3 proteins [61] and with the E3 ubiquitin ligase IMP (impedes mitogenic signal propagation) [62]. Cell stimulation with growth factors induces disassembly of the complexes, resulting in localisation of KSR1 to the plasma membrane where it facilitates signalling from c-Raf to MEK1/2 and to ERK1/2 [63]. Recent studies using cells obtained from KSR1-null mice revealed that KSR1 modulates the intensity and duration of Raf–MEK1/2–ERK1/2 signalling, promoting cell differentiation and oncogene-induced senescence [64, 65]. KSR1 can also function as a scaffold for iNOS in association with the heat shock protein, Hsp90 [66], thus suggesting that KSR1 can act as an integrator of multiple signalling pathways.

Mitogen-activated protein kinase organiser 1 (MORG1) and MEK Partner 1 (MP1)

MP1 can associate with both MEK1/2 and ERK1/2 and target this complex to late endosomes, resulting in localisation of ERK1/2 signalling to this intracellular compartment. MP1 also binds to MORG1 that in turn binds c-Raf and co-localises with vesicles in cells. The MORG1-bound Raf–MEK1/2–ERK1/2 complex is selectively stimulated by lysophosphatidic acid through G protein-coupled receptors, but is not activated through tyrosine kinase receptors in response to epidermal growth factor (EGF) or platelet-derived growth factor (PDGF). By contrast, Raf–MEK1/2–ERK1/2 complexes coordinated by KSR1 scaffold protein at the cell membrane respond to both G protein-coupled and tyrosine kinase receptors [67], further highlighting the key role of scaffolds in determining specific signalling modality.

Paxillin

Paxillin has been reported to function as a scaffold protein for c-Raf, MEK1/2, and ERK1/2, downstream of the hepatocyte growth factor receptor Met. ERK1/2 binding to paxillin is thought to control cell motility by phosphorylation of paxillin on S83, and promoting its binding to focal adhesion kinase, FAK. This results in localisation of ERK1/2 signalling to focal adhesions, activation of the Rho family GTPase Rac, and extension of lamellipodia [68].

G protein-coupled receptor (GPCR)-kinase interacting protein-1 (GIT1)

GIT1 is a scaffold protein for the MEK1–ERK1/2 pathway [69] known to bind paxillin and GRK2 (G protein-coupled receptor kinase 2). GIT1 is required for sustained activation of MEK1-ERK1/2 after stimulation with angiotensin II or epidermal growth factor [69], being involved both in G protein-coupled and tyrosine kinase receptor signalling [69]. Interestingly, GIT1 null mice show enhanced cardiomyocyte apoptosis and cardiac hypertrophy that progressed to heart failure due to impaired mitochondrial biogenesis and function [70]. However, the involvement of GIT1-dependent ERK1/2 signalling in the cardiac phenotype remains to be determined.

Dual-specificity tyrosine-phosphorylated and regulated kinase 1A (DYRK1A)

DYRK protein kinases phosphorylate serine and threonine residues on many substrates, as well as possessing autophosphorylation activity on tyrosine residues [71]. Mammalian DYRK1A is ubiquitously expressed in adult and foetal tissues with high expression in brain and heart during development. DYRK1A has been reported to function as a scaffold protein for Ras, B-Raf and MEK1, facilitating the formation of Ras–B-Raf–MEK1 multi-protein complexes, and prolonging the kinetics of ERK activation in PC12 neuronal cells in a cell-specific manner [33]. In addition, DYRK1A interacts with several transcription factors and its ability to phosphorylate NFAT has been reported [72, 73]. In cardiomyocytes DYRK1A acts as a negative regulator of the calcineurin–NFAT pathway by mediating NFAT nuclear export which leads to reduced pro-hypertrophic gene transcription [74]. Although DYRK1A has an important function in heart remodelling via the calcineurin–NFAT pathway [74–76], its potential role as an ERK1/2 scaffold molecule in cardiomyocytes has not yet been demonstrated.

Conclusions

As discussed above, several scaffold proteins can organise ERK1/2 signalling in the heart in response to different stimuli leading to differential impacts on heart remodelling. This is likely due to scaffold protein-mediated spatial regulation of ERK1/2 activation that in turn determines substrate specificity [12]. Adaptive remodelling is mainly regulated by β-arrestins and IQGAP-1, while Gβγ and FHL1 determine a maladaptive heart response. However, in the stressed heart, multiple stimuli act in combination, resulting in the simultaneous activation of multiple ERK1/2 pathways via different scaffold proteins (Fig. 1). In addition, the same receptor can activate multiple ERK1/2 pathways via different scaffolds (Fig. 2). Thus, the balance among different pathways defines the ultimate outcome on cardiac remodelling. In this scenario, the general inhibition of ERK1/2 pathway as a potential therapeutic approach is undesirable, as it will impair both detrimental and beneficial signalling. Therefore, understanding how signalling specificity is achieved is an urgent challenge of molecular cardiology, for the development of therapeutic strategies capable of selectively interfering with the correct target. It is clear that the scaffold component involved in the activation of the signalling pathway is a crucial element for selectively determining either adaptive or maladaptive signalling. In principle, the downregulation of a specific scaffold can be helpful in reducing a particular maladaptive signal, as exemplified by FHL1 knockout mice [40]. However, scaffold molecules are often multifunctional proteins involved in different signalling pathways. Detailed knowledge of the scaffold/kinase signalosome complexes will enable the use of peptide sequences or other compounds to displace specific kinases from a specific scaffold molecule resulting in interruption of a particular signalling cascade. This strategy has been adopted in a number of cases for selective attenuation of a signalling pathway [77, 78], but needs to be translated in vivo in order to evaluate the impact on heart remodelling.

Fig. 1.

Multiple stimuli can act in combination via different receptors, resulting in the simultaneous activation of multiple ERK1/2 pathways via different scaffold proteins

Fig. 2.

The same receptor can activate multiple ERK1/2 pathways via different scaffold proteins