mTOR in Growth and Protection of Hypertrophying Myocardium

Summary

In response to an increased hemodynamic load, such as pressure or volume overload, cardiac hypertrophy ensues as an adaptive mechanism. Although hypertrophy initially maintains ventricular function, a yet undefined derailment in this process eventually leads to compromised function (decompensation) and eventually culminates in congestive heart failure (CHF). Therefore, determining the molecular signatures induced during compensatory growth is important to delineate specific mechanisms responsible for the transition into CHF. Compensatory growth involves multiple processes. At the cardiomyocyte level, one major event is increased protein turnover where enhanced protein synthesis is accompanied by increased removal of deleterious proteins. Many pathways that mediate protein turnover depend on a key molecule, mammalian target of rapamycin (mTOR). In pressure-overloaded myocardium, adrenergic receptors, growth factor receptors, and integrins are known to activate mTOR in a PI3K-dependent and/or independent manner with the involvement of specific PKC isoforms. mTOR, described as a sensor of a cell’s nutrition and energy status, is uniquely positioned to activate pathways that regulate translation, cell size, and the ubiquitin-proteasome system (UPS) through rapamycin-sensitive and -insensitive signaling modules. The rapamycin-sensitive complex, known as mTOR complex 1 (mTORC1), consists of mTOR, rapamycin-sensitive adaptor protein of mTOR (Raptor) and mLST8 and promotes protein translation and cell size via molecules such as S6K1. The rapamycin-insensitive complex (mTORC2) consists of mTOR, mLST8, rapamycin-insensitive companion of mTOR (Rictor), mSin1 and Protor. mTORC2 regulates actin cytoskeleton in addition to activating Akt (Protein kinase B) for the subsequent removal of proapoptotic factors via the UPS for cell survival. In this review, we discuss pathways and key targets of mTOR complexes that mediate growth and survival of hypertrophying cardiomyocytes and the therapeutic potential of mTOR inhibitor, rapamycin.

Introduction

Cardiac hypertrophy, observed for example, during aortic stenosis or chronic hypertension, is the basic adaptive response of the adult heart in response to an increased hemodynamic overload ¹. Although this compensatory mechanism preserves ventricular function in the short-term, this growth phenomenon often leads to pathological remodeling and precipitates a transition into congestive heart failure (CHF). Therefore, a major research goal is to develop new clinical interventions in order to delay or reverse this transition from compensated hypertrophy to cardiac failure by preferentially allowing only regulated physiological growth that would help normalize the wall-stress of the myocardium and improve the ability of cardiomyocytes to survive in the stressed environment. This review focuses on a key molecule, known as mammalian target of rapamycin (mTOR), which could be a common molecular branch point, regulating growth and survival of hypertrophying myocardium.

mTOR is a serine/threonine kinase that is a critical link between external stimuli and cell growth regulation. mTOR has been extensively studied in several cell systems and disease conditions due to its unique position in various cell signaling pathways, including those that regulate cell size, cell growth, protein synthesis, cytoskeleton, nutrition and energy sensing and protein degradation via the ubiquitin proteasome system (UPS). A major characteristic of mTOR is its specific inhibition by rapamycin-FKBP12 complex. Recent studies indicate that mTOR has both rapamycin-sensitive and -insensitive functions by forming two distinct multiprotein complexes. These two types of signaling complexes (for reviews, ^2–4) are formed as mTOR complexes with Raptor and Rictor to form mTORC1 and mTORC2, respectively. mTORC1 primarily governs growth regulation through p70 S6 kinase (S6K1), 4E-binding protein (4EBP) and protein phosphatase-2A (PP2A) in order to promote cell growth and size ⁵, whereas mTORC2 is known to regulate the actin cytoskeleton^6,7 and contribute to the activation of Akt ⁸ that triggers cell survival pathways, including those that are responsible for the removal of deleterious proteins via the UPS ⁹. Although it is conceivable that these two mTOR-mediated processes might be defective in CHF, the characterization of these two arms of mTOR signaling has not been achieved in detail. Initial studies including ours ^10–14 have characterized the activation of the rapamycin-sensitive mTOR (mTORC1) pathway in promoting translation in hypertrophying cardiomyocytes, although how this pathway might contribute to the overall increase in cardiomyocyte size remains unaddressed. Furthermore, the possible protective role of mTORC2, which is widely studied in cancer, is yet to be explored in hypertrophying heart. In this review, we focus on the current understanding of these two mTOR complexes and their possible role in mediating both growth and protection of hypertrophying cardiomyocytes, especially when they are under mechanical stress. We also discuss therapeutic values of rapamycin via its effects on mTOR complexes in hypertrophying heart.

mTOR Structure:

mTOR (TOR in yeast, dTOR in Drosophila), which belongs to the phosphoinositide-3-kinase-related kinase (PIKK) family, is an unconventional protein serine/threonine kinase. It is highly conserved from yeast to mammals and is often found as complexes of molecular mass ranging from 280 to 500 kDa. In mammalian cells, mTOR is a 280 kDa protein, which has been found to have six functional domains interspersed by short linker sequences (Fig. (1A); source ⁴). In the N-terminus, two identical HEAT (Huntington, Elongation factor 3A, A subunit of PP2A, and TOR1) domains reside in tandem, and they are believed to mediate the multimeric complex formation with other protein targets. C-terminus to the two HEAT domains is the FAT (FRAP, ATM and TRRAP) domain, and it is highly conserved among all members of the PIKK family. The FAT domain is followed by the FRB (FKBP12-rapamycin-binding) domain, which is unique to TOR family kinases, since the hydrophobic cleft of this domain offers a high affinity binding site for the FKBP12-rapamycin complex. The solution structure of the FRB domain in complex with rapamycin and phosphatidic acid has been published recently ¹⁵. This and a previous structural study ¹⁶ indicate that the FRB domain interacts with both entities of the FKBP-rapamycin complex. Thus, FRB domain has been considered as a potential regulatory domain of mTOR activity. The FRB domain is followed by the catalytic domain. The catalytic activity of mTOR is directed towards S/T residues in the hydrophobic motif of AGC family kinases, including S6 kinase 1 and Akt. The substrate specificity of mTOR is largely dictated by partners associated within the respective mTOR complexes. Thus, the binding of FKBP12-rapamycin complex to the FRB domain can suppress the catalytic function of mTORC1 but not mTORC2 (discussed below). Finally, the domain present at the C-terminus is the FAT-C domain, and it has a sequence homology to the FAT domain. Both FAT and FAT-C domains are thought to contribute to the active conformation of the kinase domain.

Fig. (1): Molecular organization of mTOR and its assembly with binding partners:

(A) Structural features of mTOR, consisting of the following six functional domains (from N to C-terminus): Two HEAT domains that are present in tandem at the N-terminus; FAT domain in the middle of the molecule; FRB domain; Kinase domain; FATC domain at the C-terminus. The importance of these structural domains is described in the text. (B) Cellular partners of mTOR, constituting the two mTOR complexes. mTORC1: Complexing of mTOR with mLST8 and Raptor; mTORC2: Complexing of mTOR with mLST8 and Rictor, mSin1 and Protor.

Distinct mTOR Complexes:

mTORC1 and mTORC2 have been characterized to associate with common proteins, such as mLST8, and distinct proteins that alter their biological function and rapamycin sensitivity (Fig. (1B)). Of these distinct proteins, Raptor constitutes mTORC1, whereas three other proteins, namely Rictor, Protor-1 (also known as PRR5) and mSin1, have been characterized to associate with mTOR exclusively in mTORC2. These are cytoplasmic protein complexes and have differential sensitivity to nonionic detergent extraction (Fig. (2A)). That is, during isolation from mammalian cells, mTORC1 and mTORC2 can exist in their native forms if cells are lysed in a buffer without detergent, or with 0.3% Tween-20 or 0.3% CHAPS. However, if cells are lysed in 0.3% N-acetyl glucoside, 1% NP-40 or 1% Triton X-100, only a partial mTOR complex, consisting of mTOR and mLST8 that are common to both complexes, can be isolated (Fig. (2B))^3,17,18. Raptor of complex-1 and Rictor, mSin1 and Protor components of complex-2 no longer associate with mTOR. Based on the detergent solubility, Rictor, mSin1 and Protor components of mTORC2 were found to coexist as an independent complex, suggesting their mTOR independent interaction ^6,7,18. Therefore, usage of appropriate buffer conditions is critical to study the intracellular organization of mTOR complexes.

Fig. (2): Detergent sensitivity of mTORC1 and mTORC2:

mTOR complexes are sensitive to detergent during their isolation. Extraction of cells or tissue samples with buffers containing Tween-20, CHAPS or no detergent preserves both mTORC1 (mTOR, mLST8 and Raptor) and mTORC2 (mTOR, mLST8, Rictor, mSin1 and Protor). However, if the buffer contains N-acetylglucoside, NP-40 or Triton X-100, both complexes undergo partial dissociation. Presence of these detergents causes dissociation of Raptor from mTORC1, although the binding between mTOR and mLST8 is preserved. Similarly, these detergents cause dissociation of mTORC2 into two parts: one compartment contains mTOR and mLST8 and the other compartment contains Rictor, mSin1 and Protor.

Activation of mTOR complexes

Multiple pathways that promote cell growth have been shown to converge at the activation of mTOR. Mechanical stimulation and several agonists that stimulate growth factor receptors, α- and β-adrenergic receptors and integrins have been shown to activate mTORC1 and mTORC2 in numerous cell types ⁴. As shown in (Fig. (3)), in resting cells, the activity of mTORC1 is controlled via its interaction with the GTP/GDP-bound Rheb, a member of the small GTPase proteins ¹⁹, and this in turn is under the control of the small GTPase-activating function of TSC1/TSC2 (tuberous sclerosis complex) ^20–23. During mechanical or hormonal stimulation, the TSC1/TSC2 complex is inactivated through phosphorylation by various upstream pathways, including Akt. Rheb is then converted to the active GTP-bound form, resulting in the activation of mTORC1. In pressure-overloaded myocardium, others ^14,24–29 and we ^{10–12,30,31} have shown activation of mTORC1 complex as evidenced by mTOR phosphorylation at S2448, sensitivity to rapamycin inhibition, and activation of its major downstream targets, such as p70S6K (S6K1), ribosomal protein S6 and 4EBP. Our recent studies ³⁰ demonstrate that both PI3K-dependent and independent mechanisms control mTORC1 activation. The binding and activation of class IA PI3K at an activated receptor, for example with growth factors and insulin initiates a cascade of signaling events that leads to the activation of both mTORC1 and mTORC2. Subsequent to PI3K activation, generation of phosphatidylinositol(3,4,5)-triphosphate (PIP3) at the plasma membrane results in the recruitment of the PDK1 (phosphoinositide-dependent kinase-1) and Akt. In this manner, Akt is phosphorylated at T308 by PDK1, and the activated Akt then mediates mTORC1 activation. In addition, class IB PI3K, which is known to undergo activation during adrenergic stimulation ³² and class III PI3K, which is activated during the availability of amino acids and nutrients ³³, have also been shown to activate mTORC1, although the precise intermediate steps have yet to be established. Our studies using adult cardiomyocytes ¹² indicate that ligand binding to β3 integrin and internalization of a soluble integrin-binding peptide, RGD, can result in mTOR and S6K1 activation. This activation, which is sensitive to rapamycin, was found to occur independent of either class IA PI3K or PKC, suggesting class IB and/or class III PI3K as possible intermediate kinases.

Fig. (3): Activation and role of mTORC1:

Multiple hypertrophic agonists trigger their cognate cell surface receptors, transmitting biochemical signals that converge at the activation of mTORC1. For this, the GTPase, Rheb, needs to be in its GTP-bound form, which is achieved by the inactivation of GTPase activation function of TSC1/TSC2. In this manner, Akt activation by growth factor receptors, adrenergic receptors and/or integrins results in the mTORC1 activation through suppression of TSC1/TSC2 complex. For this, stimulation of growth factor receptor (such as insulin receptor)-mediated Class IA PI3K and β-adrenergic receptor-mediated Class IB PI3K contribute to Akt activation that suppresses TSC1/TSC2. Furthermore, α- and β- adrenergic stimulation results in the activation of PKA and/or PKC that also leads to the loss of TSC1/TSC2 activity. Also, amino acid availability and integrin-mediated internalization of ECM products that might activate Class III PI3K, also contribute to mTORC1 activation. Finally, mTORC1 is also known to serve as an energy sensor, and this is primarily mediated via cellular energy (ATP/AMP ratio) that regulates AMP-activated protein kinase (AMPK). When the cellular energy level is low, AMPK activates TSC1/TSC2 resulting in the suppression of mTORC1 activation. When the energy level is high, the AMPK activity, and thus TSC1/TSC2 activity are low, resulting in the activation of mTOR.

Activated mTORC1 mediates phosphorylation of S6K1, 4EBP and PP2A. S6K1 activation contributes to both translational initiation and ribosomal biogenesis in addition to promoting cell size via SKAR phosphorylation. Similarly, phosphorylation of 4EBP causes its dissociation from eIF4E thereby relieving its negative regulation on eIF4E-mediated translational initiation. Furthermore, the increased phosphorylation of mTORC1 substrates is sustained through mTORC1-mediated phosphorylation and inactivation of PP2A. However, prolonged S6K1 activation has been shown to form a negative feedback loop where it phosphorylates and blocks IRS-1 function, thus resulting in the loss of insulin sensitivity.

In a recent study ³⁰, we carefully analyzed the importance of specific PKC isoforms mediating mTOR/S6K1 activation. In these studies, both insulin and endothelin-1A were found to stimulate class IA PI3K-dependent and -independent phosphorylation of both mTOR at S2448 and S6K1 at T389 in adult cardiomyocytes. While both pathways required the activity of PKCδ isoform, only endothelin stimulation however required PKCε, in addition to PKCδ. These studies therefore indicate that PKCδ might be a critical player both during class IA PI3K-dependent and -independent activation of mTORC1.

mTORC2 Formation and Activation:

Whereas both mTOR complexes appear to be present even in quiescent unstimulated cells ^3,17,34, the mechanism of activation has been well characterized for mTORC1, and factors that function upstream of mTORC2 have yet to be defined (Fig. (4)). Growth factors, nutrients and increased levels of cellular energy that activate mTORC1 via TSC1/TSC2 and Rheb were also found to regulate actin polymerization. Therefore, it is suggested that these anabolic factors may also activate mTORC2, since mTORC2 is linked to actin cytoskeleton dynamics ^3,17. Since our recent studies ³⁰ demonstrate that PKC isoforms contribute to mTORC1 activation, it is possible that they may also play an upstream role in the activation of mTORC2.

Fig. (4): Activation and role of mTORC2:

Stimulation of cells with growth factor and insulin results in the activation of class I PI3K that leads to the activation of mTORC2, although whether PI3K is directly involved in this process or is a mediator through activated Akt is not clear. Activated mTORC2 then mediates Akt phosphorylation at the S473 site. In this manner, S473 phosphorylated Akt both at the plasma membrane and in the nucleus phosphorylates its target proteins for cell survival and cytoskeletal organization. Finally, mTORC2 was also found to mediate PKCα phosphorylation and regulate both Rho family GTPases and paxillin for cytoskeletal organization. Overall, activated mTORC2 regulates both cell survival and actin-cytoskeletal organization via phosphorylation of its potential substrates, such as Akt and PKCα.

Concerning the downstream mediators of mTORC2, it is clear that this complex does not contribute to the activation of mTORC1 substrates, S6K1 and 4EBP. Several studies have shown that the phosphorylation of Akt at the S473 site is mediated via mTORC2 (reviewed in ³⁵). Whether mTORC2 functions as the sole kinase for this important phosphorylation of Akt for cell survival is not clear. A recent study ³⁶, performed in adult heart and isolated neonatal rat ventricular cardiomyocytes, demonstrates that p21-activated kinase (PAK) might also serve as a direct upstream kinase of Akt for S473 phosphorylation. Furthermore, Rictor, one of the binding partners of mTOR in mTORC2, has also been found to associate with integrin-linked kinase (ILK) independent of mTOR, resulting in the S473 phosphorylation of Akt ³⁷. Thus, in addition to mTORC2, at least two other independent kinases have been identified for Akt S473 phosphorylation. Therefore, further studies are required to prove that mTORC2 signaling is responsible for Akt activation in hypertrophying heart and that the activation of Akt by this mechanism mediates cardioprotective effects. In addition to its cell survival role, mTORC2 was found to modulate the phosphorylation of PKCα and play a role in regulating the actin cytoskeleton ⁷. Similarly, Jacinto, et al ⁶ demonstrated an upstream role for mTORC2 in Rho family GTPase activation and paxillin phosphorylation, again linking mTORC2 with actin cytoskeletal organization. Finally, several studies demonstrate nuclear localization of Akt in multiple cell types ³⁸, including cardiomyocytes ³⁹ where phosphorylated Akt plays a cell survival role. Whether mTORC2 contributes to Akt phosphorylation and nuclear localization in the myocardium has yet to be established.

Activation of mTOR Complexes During Cardiac Hypertrophy:

The two distinct complexes of mTOR have been characterized in several cell types but an equally thorough analysis in the heart and in the isolated cardiomyocyte is lacking. Clearly, more information on the mTORC1 complex is present in the literature as many laboratories, including ours, have used rapamycin for years to gain a better understanding of the role that mTOR plays during cardiac hypertrophy. As mentioned earlier, a read-out of mTORC1 activity is the phosphorylation of S6K1 at T389. Our in vivo studies using an RVPO model of pressure-overload ¹¹, as well as our in vitro studies using insulin and other hypertrophic agonists in adult feline cardiomyocytes ^11,12,30, demonstrate phosphorylation of S6K1 at T389. Furthermore, this agonist-stimulated phosphorylation was blocked by pretreatment of cardiomyocytes with rapamycin. Since T389 phosphorylation is primarily mediated by mTORC1, these studies demonstrate the activation of mTORC1 during hypertrophic stimulation of both adult heart in vivo and isolated cardiomyocytes in vitro. Similarly, our studies ¹² demonstrate that insulin and to a lesser degree RGD, but not TPA stimulation of adult cardiomyocytes, results in the phosphorylation of AKT at the S473 site. Since phosphorylation at this site is primarily mediated by mTORC2 ⁸, we also analyzed if it occurs in pressure-overloaded myocardium (Fig. (5)). Thus in pressure-overloaded right ventricle but not in normally loaded same animal control left ventricle, phosphorylation of AKT at both S473 and T308 sites occurs at the level of cardiomyocytes. These studies strongly demonstrate that both mechanical stimulation, such as in vivo pressure-overload of the heart, and agonist stimulation of adult cardiomyocytes in vitro can activate both mTORC1 and mTORC2. In future studies it will become more clear how the different downstream effectors of mTOR complexes mediate growth and survival of the myocardium.

Fig. (5): Phosphorylation of Akt in pressure-overloaded myocardium:

Confocal analysis (top panel) shows both S473 and T308 phosphorylated Akt is enriched in cardiomyocytes of 48 h pressure-overloaded myocardium. For this, specific antibodies were used to stain α-actinin (in red) and pAkt S473 or T308 (in green). Western blot analyses (bottom panel) demonstrate the presence of S473 phosphorylated Akt in both 24 and 48 h pressure-overloaded myocardium and in agonist-stimulated adult cardiocytes.

Role of mTORC1 in Hypertrophic Growth:

Increases in myocyte size and ventricular mass are distinguishing characteristics of cardiac hypertrophy that have been repeatedly shown to occur by means of increased protein synthesis. As a rapamycin-sensitive complex, mTORC1 appears to play a central role in transmitting extracellular signals of both mechanical and mitogenic origin, into signals that mediate several aspects of hypertrophic growth. Indeed, several studies have shown that mTORC1 activity is critical to the regulation of cell growth by modulation of several key translational control mechanisms that play a role in facilitating increases in protein synthesis ⁴⁰. These include the phosphorylation of S6K1 and 4EBP1. Although the full activation of S6K1 requires additional signaling inputs, mTORC1 is responsible for critical phosphorylation and activation of this kinase. S6K1 is known to play multiple roles in the regulation of translation initiation via its interactions with eIF3 ⁴¹, eIF4b ⁴² and PP2A ⁴³. S6K1 has also been shown to regulate translation elongation via the phosphorylation of eEF2 kinase ⁴⁴. In regards to cardiac hypertrophy, studies by several laboratories ^14,45, including ours ^10,11, demonstrate a robust activation of S6K1 in pressure-overloaded myocardium. Although the activity of 4EBP has not been fully characterized in the intact heart, levels of hyperphosphorylated 4EBP in isolated adult cardiomyocytes were shown to increase in response to TPA and insulin treatment as well as electrically stimulated pacing ⁴⁶. Our recent study ³¹ further demonstrated that a selective translation of messages possessing 5’-terminal oligopyrimidine (5’-TOP) sequence in the 5’-untranslated region (5’-UTR) occurs in pressure-overloaded myocardium and that such a preferential translation during hypertrophic stimulation of cardiomyocytes in vitro was found to be sensitive to rapamycin inhibition. Since translation of these messages results in the generation of ribosomal proteins, these studies show the importance of mTORC1 in enhancing the translational capacity of cardiomyocytes via ribosomal biogenesis.

Although correlative evidence suggested that S6K1 activation increased ribosomal biogenesis via S6 phosphorylation, studies using S6K1 knockout mice exhibited an overall reduction in animal and cell size ⁴⁷ with no appreciable change in S6 phosphorylation. On the other hand, S6K2 knockout mice exhibited loss of S6 phosphorylation without affecting the cell or animal size ⁴⁸. Furthermore, a recent study using S6K1/S6K2-double knockout mice ⁴⁹ showed that pressure-overload-induced hypertrophic growth could occur even in the absence of these anabolic kinases, at least during the initial period of 1 wk after aortic banding. However, as a cautionary note, a related kinase, p90rsk, not studied in this specific context, has been envisaged to play a possible compensatory role for the absence of S6Ks ^48,49. In favor of this, Boluyt et al ⁵⁰ demonstrated that the loss of pressure-overload-induced S6K1 activation in rapamycin-treated mice occurs initially, although S6K1 activation was observed later in 3-day pressure-overloaded mice even in the presence of rapamycin implicating existence of an alternative pathway. Hence, it is important to use the double knockout mice for further studies in pressure-overloaded myocardium to address whether: (i) ribosomal biogenesis, which is important for a sustained hypertrophic growth, can still occur, (ii) individual cell size increases, as reported during a sustained pressure-overload in wild type animals, can still take place, (iii) myocardial cell loss increases and (iv) there is hyperactivation of p90rsk that might compensate for S6Ks’ absence.

Rapamycin-sensitive mTOR activity has been shown to control overall cell size via S6K1-mediated phosphorylation of both S6 protein ⁵¹ and a newly identified S6K1 target, SKAR (S6K1 Aly/Ref like target protein) ⁵². Further studies on SKAR ⁵³ demonstrate that it contributes to pre-mRNA splicing downstream of mTOR/S6K1 pathway. Our studies clearly demonstrate S6 phosphorylation both in PO myocardium in vivo and agonist-stimulated adult cardiomyocytes in vitro. Similarly, our recent work on SKAR (Moschella and Kuppuswamy, unpublished observation) in pressure-overloaded myocardium exhibits both increased expression and posttranslational modification, including phosphorylation of SKAR at multiples sites. Therefore, phosphorylated S6 protein and SKAR could be important mediators in determining cardiomyocyte size during hypertrophic growth downstream of mTORC1.

Finally, mTOR has been linked to autophagy, lysosomal degradation of long-lived proteins, macromolecular aggregates and damaged intracellular organelles ^54,55. Activated mTOR (in this case mTORC1) acts as a negative regulator of autophagy (Schmelzle T, Cell, 2000). This process appears to be mediated through a translational mechanism involving eEF2 kinase inhibition, where mTOR-mediated phosphorylation of eEF2 kinase results in the loss of its function to promote autophagy ⁵⁶. Therefore, one potential mechanism for the mTOR-mediated hypertrophic growth could be via regulating autophagy-related protein degradation.

Role of mTORC2 in Cardioprotection of Hypertrophying Heart:

Akt, a well-characterized effector molecule of mTORC2 ⁵⁷, has been shown to function as a progrowth and cell survival kinase ⁵⁸. Akt was first identified as a cellular counterpart of murine AK T8 retrovirus, and the product was identified as a serine/threonine kinase related to protein kinase A and C and therefore named protein kinase B (PKB). Mammalian genomes contain three Akt genes (Akt1, Akt2 and Akt3) and Akt1 is widely expressed in several tissues including the heart ^48,59. The current model of its activation suggests that the activation of PI3K generates PIP3 in the plasma membrane, which subsequently recruits Akt via its pleckstrin homology domain ⁶⁰. At the membrane, Akt undergoes phosphorylation at two critical sites, T308 and S473. The T308 phosphorylation in the activation loop is mediated by PDK1. However, the S473 phosphorylation in the hydrophobic motif of the C-terminal tail of Akt is predominantly mediated by mTORC2 ⁸, although recent studies demonstrate that PAK1 ³⁶ and ILK ³⁷ also serve as upstream kinases for this phosphorylation.

Activated Akt is a critical regulator of cell survival (Fig. (4)), and several studies in multiple cell types have demonstrated that Akt is sufficient to block cell death induced by a variety of apoptotic stimuli ^61,62. Akt mediates these effects by modification of several apoptotic targets, including Bad, forkhead transcription factors (FOXO), and IκB kinase (IκKα). Many of these targets and other intermediary proteins of their pathways are degraded via ubiquitination subsequent to their phosphorylation by Akt. In the case of Akt’s role in the heart, transgenic mice overexpressing Akt exhibited increased cardiomyocyte size and contractility in vivo ^27,63,64, and several studies link Akt activation in the heart to physiological hypertrophy ^64–67. Akt1-deficient mice demonstrate lack of both swim training-induced cardiac hypertrophy and IGF-1-stimulated protein synthesis, suggesting the importance of Akt activation for physiological growth of the heart ⁶⁵. Importantly, pressure overloading of Akt1-deficient mice via TAC results in exaggerated pathological hypertrophy and fibrosis ^65,68, which also supports a role of mTORC2 signaling for only physiological hypertrophy. Other independent studies ^39,69 demonstrate that the nuclear localization of Akt antagonizes certain features of cardiac hypertrophy and improves systolic function in TAC-induced hypertrophy in mice. It is suggested that acute activation of Akt especially in the nucleus may be beneficial to the heart ⁷⁰. Overall, Akt, which is known to function both as an upstream activator of mTORC1 and downstream mediator of mTORC2, may promote cardiomyocyte growth and protection, which are critical events at least during early stages of PO-induced hypertrophic heart.

Ubiquitin Proteasome System (UPS) is a Downstream Effector of mTORC2:

The UPS offers the major non-lysosomal removal of proteins in the cell, in which protein degradation requires both modification of specific target proteins and activation of ubiquitin-26S proteasome system. Ubiquitination is a process where ubiquitin, an 8.5 kDa polypeptide, is covalently attached to the ε-amino group of lysine residues of target proteins and is mediated by three independent enzyme steps, involving E1, E2 and E3 ligases ⁷¹. The target proteins are either mono-, multi- or poly-ubiquitinated, and the nature of such ubiquitination determines their fate in terms of whether they should either have a modified function or be degraded ⁷². Although a basal level of UPS activity is required for cellular homeostasis by removing misfolded and short-lived proteins, its function is imperative under stress conditions that increase the release of deleterious proteins in the cell. Adult cardiomyocytes, having no proliferative potential, can be expected to have the ability to change their proteomic profile in response to stress stimuli. The importance of proteasome function in the stressed heart is now beginning to emerge ^73–76 (see below).

In response to mechanical load, the UPS may also regulate the adaptation of cardiac cell growth and survival in coordination with mTOR. In fact, inhibiting the proteasome attenuates hypertrophic growth in vivo ^77,78. Conversely, since in vitro inhibition of the UPS with various hypertrophic agonists decreases Akt level ⁷⁸, and since Akt or mTOR inhibition also inhibits hypertrophic growth ^14,79–81, it is likely that the UPS plays an integral role in regulating mTOR signaling in hypertrophic growth. It is conceivable that in the stressed heart, activation of several E3 ligases can be expected to mediate ubiquitination of these proteins for subsequent proteasomal degradation ⁸². Furthermore, amino acids released from such degradation might be used for the subsequent mTORC1-mediated synthesis of critical proteins ⁷⁴. Under stress conditions, such as mechanical overload to the heart, a sharp rise in damaged or denatured proteins can be expected. These deleterious proteins and activated proapoptotic signals have to be eliminated to preserve cardiac function. Indeed, accumulation of hyperubiquitinated proteins, resulting from the lack of their proteasomal degradation, has been observed in failing human hearts ^83,84. Since dysfunction of this system has been shown to preclude heart failure ⁸⁵, it may contribute to the pathological transition between compensatory and decompensatory hypertrophy. A compelling example of this phenomenon was observed in a decompensated hypertrophic model, where increased cell death correlating with decreased proteasome function could be rescued by siRNA for pro-apoptotic targets of Akt, Bax and p53, which had accumulated due to insufficient degradation ⁸⁵. In this context, pathways that are activated by mTORC2 and Akt result in the phosphorylation and subsequent ubiquitin-mediated elimination of several proapoptotic proteins. Additional target proteins include FOXO (Forkhead O) and IκB (inhibitor of nuclear factor kappa B). The mTOR/Akt pathway has been shown to regulate the cellular level of FOXO transcription factors ^74,86,87. The mTOR/Akt/FOXO axis can control both atrophy and autophagy. Akt inactivates FOXO by phosphorylation, causing its nuclear exclusion ⁸⁸ and ubiquitin-mediated destruction ^74,86,87. FOXO is known to induce atrophy by transcribing atrogenes MURF-1, which causes myofibrillar breakdown, and atrogin-1, which inhibits the hypertrophic agonist calcineurin ⁸⁹, both through ubiquitin-mediated degradation. Furthermore, FOXO transcription factor is necessary and sufficient to cause autophagy in skeletal muscle ^90,91. In a ubiquitin-independent manner, FOXO induces autophagy by transcribing autophagy-related genes, LC3/Atg8 and Bnip3 ⁹¹. Overall, loss of FOXO during IGF-1 stimulation and subsequent Akt induction is sufficient to preserve growth ⁸⁸. Since this process is rapamycin-insensitive ⁹¹, it is mTORC2 that is responsible for the ubiquitin-mediated inhibition of atrophy and autophagy gene induction by FOXO transcription factors. Therefore, mTORC2 can regulate hypertrophic growth independent of mTORC1 activation.

In the case of NFκB, it has been established that NFκB is required for hypertrophic growth in TAC mice, where inhibiting NFκB by cardiac transfection of a non-phosphorylatable mutant of its inhibitor IκB attenuated hypertrophy and NFκB activity ⁹². Since this IκB mutant lacks phosphorylation sites for its ubiquitin-mediated degradation and the subsequent activation of NFκB ^92,93, successful delivery of this virus into the heart inhibits NFκB activation induced by increased cardiac load. Further, it was shown that mTORC2 is responsible for this activation. Overexpressing a constitutively active form of Akt in a cardiac-specific manner is sufficient to promote a significant increase in NFκB activity and hypertrophy ⁸⁰. Inhibiting mTORC1 with rapamycin partially attenuates hypertrophic growth but does not affect NFκB activity, indicating the role of mTORC2 in mediating NFκB nuclear localization and activation. While the loss of NFκB activity results in a partial loss of hypertrophic growth, inhibition of both mTORC1 and NFκB exhibits maximum loss of cardiac hypertrophy ⁸⁰. These studies show that mTORC2 contributes to cardiomyocyte survival and thus contractile function via NFκB-mediated regulation of several pro-survival genes.

mTORC2 and the Actin Cytoskeleton:

Cardiac hypertrophy is associated with actin cytoskeletal rearrangement, a process that is regulated to a large extent by integrins. Integrin activation leads to specific cytoskeletal rearrangement via effectors such as vinculin, talin, FAK, PYK2, Src, p130cas, paxillin, Nck and Shc (reviewed in ^94–96). Integrins also are involved in signaling to small GTPases. Several studies show that perturbation of components downstream of integrin signaling lead to cytoskeletal defects ⁹⁷. Integrin activation, like in non-myocytes, activates a similar signaling paradigm in cardiomyocytes that are embedded in collagen along with an integrin-binding peptide (Arg-Gly-Asp, RGD) ^98,99. Cardiac-specific β1 integrin knockout resulted in myocardial disarray typical of a disorganized cytoskeleton ¹⁰⁰. We have shown the involvement of β3 integrin during hypertrophic signaling ^{12,99,101–103}.

Both mTORC1 and mTORC2 have been implicated in regulating the actin-cytoskeleton. Involvement of mTOR in cytoskeletal reorganization has been shown in Swiss 3T3 fibroblasts in which active mTOR, Akt, PDK1, PI3K and S6K1 are localized in actin arcs of the leading edge in a rapamycin-sensitive manner. However, this earlier study did not analyze the components of mTORC2 ¹⁰⁴. In astrocytes, mTOR has been shown to regulate stress fiber formation downstream of neurofibromin 1 (NF1−/−). In NF1−/− cells, mTOR has been found to be hyperactive in a Ras-dependent manner. Moreover, Rac1 has been shown to be hyperactive downstream of mTOR. These cells had reduced actin stress fibers and rapamycin treatment rescued the defective actin phenotype ¹⁰⁵. Recently, components of mTORC2 have attracted more attention in terms of their cytoskeletal effect. Loewith et al ¹⁷, while identifying the two complexes of TOR in yeast, showed that TORC2 controls actin cytoskeleton via the effector pathway consisting of ROM2, RHO1, PKC1 and the PKC1-controlled MAPK cascade. Subsequent studies further revealed a stronger role for TORC2 than TORC1 in actin cytoskeleton. In NIH 3T3 fibroblasts, addition of insulin, serum or lysophosphatidic acid induced actin cytoskeletal rearrangement including stress fiber and lamellipodia formation. Such a rearrangement was insensitive to rapamycin ⁶. Knockdown of mTOR, mLST8 or mAVO3 (Rictor) lead to a reduction in stress fiber formation. In addition, a key cytoskeletal protein, paxillin, was also found to be hypo-phosphorylated at Y118. These actin cytoskeletal defects were rescued by overexpressing active Rho (RhoA-L63) or active Rac (Rac1-L61), suggesting that mTORC2 is upstream of these small GTPases ⁶. It is worth mentioning here that Rac1 has been shown to be important for mediating cardiac hypertrophy. In isolated adult cardiomyocytes embedded in collagen along with RGD peptide for integrin activation, we have shown the cytoskeletal recruitment of Rac1 and phosphorylation of PAK (Willey CD, PhD Dissertation). Also, targeted deletion of Rac1 in the heart leads to a reduced hypertrophy upon angiotensin stimulation ¹⁰⁶. Sarbassov et al ⁷ showed that reduced Rictor or mTOR by siRNA-mediated knockdown leads to a defective actin stress fiber formation. While control cells exhibited actin staining at the cell cortex, cells without Rictor or mTOR showed prominent cytoplasmic actin fibers with reduced cortical actin. Paxillin distribution was also altered in these cells. This study also showed that mTORC2 regulates actin cytoskeleton via PKCα, since Rictor or mTOR knockdown reduced S657 phosphorylation of PKCα. A recent study in glioblastoma cell lines and in primary brain tumor cells identified high levels of Rictor accompanied by Akt S473 phosphorylation, indicating mTORC2 activation. Such an activated mTORC2 positively regulates integrin β1 and β3 expression and S657 phosphorylation of PKCα, indicating that integrin and mTOR signaling are intertwined ¹⁰⁷.

Overall, these studies demonstrate that mTORC2, in addition to its potential role in cell survival via Akt, have a significant role in actin-cytoskeleton dynamics, and therefore mTORC2 is envisaged to contribute to both cytoskeletal alteration and survival of cardiomyocytes during hypertrophic stimulation by converging various upstream signals. However, rapamycin-sensitive cytoskeletal regulations reported in the past need to be revisited more carefully in the light of recent advances in delineating the functions of mTOR complexes. In addition, the dose and time course of rapamycin treatment need to be critically reviewed before interpreting the cytoskeletal effects of rapamycin.

Rapamycin and its Cardioprotective Effects:

Rapamycin has been in clinical practice to prevent kidney and heart transplant rejection ^108,109, and this macrolide has been suggested to have antitumor effects in certain cancer patients, although it did not exert the same effect on all cancer cells (Reviewed in ¹¹⁰). In patients with vascular disease, it is also used to coat stents in order to prevent restenosis during coronary angioplasty ¹¹¹. However, the therapeutic effects of rapamycin in patients with heart disease remain unclear. While there are reports that blocking mTORC1 with rapamycin can abolish the protective effect of ischemic or pharmacological preconditioning ^112–114, other independent studies indicate that rapamycin treatment has a cardioprotective effect against ischemia/reperfusion injury ¹¹⁵. It is suggested that rapamycin concentrations and/or timing of its administration during the ischemia/reperfusion cycle may contribute to such opposing effects ¹¹⁶. It appears that in studies where the heart is pretreated with rapamycin prior to the ischemia/reperfusion injury, a cardioprotective effect was observed; however, if rapamycin is administered just prior to the onset of reperfusion, the preconditioning-mediated protective effect is lost. Additional studies are needed to understand the differential effects of rapamycin treatment.

In pressure-overloaded myocardium, activation of mTORC1 and its downstream effector molecules, S6K1 and 4EBP is known to promote cardiac hypertrophy. Whether the initial growth process that requires factors, such as mTOR/S6K1, also contributes to the pathological remodeling that eventually results in the transition of a compensated heart to decompensating failing heart is not clear. These questions were partly addressed by rapamycin administration in vivo using a mouse pressure-overload model. Shioi et al ¹⁴ demonstrated that rapamycin administration prior to the initiation of left ventricular pressure-overload in mice reduced hypertrophic cardiac growth significantly without compromising ventricular function. Although these studies were performed in a 1 wk pressure-overload period, they reveal that the blocking of the hypertrophic growth process does not affect cardiac performance. Whether the loss of the initial compensatory growth would affect the transition into decompensatory failing heart in long-term pressure-overloaded animals is not known. Further studies by this group ²⁸ demonstrate that in mice with established decompensated hypertrophy, rapamycin treatment significantly improves LV function, fractional shortening and ejection fraction, suggesting that rapamycin may control pathological remodeling.

How does treatment of rapamycin that blocks mTORC1 and its anabolic growth effects, improve ventricular function? Based on available literature, three possibilities come to mind: First, although rapamycin is a highly specific inhibitor of mTOR, certain studies demonstrate that it affects one or more ion channels. In the study conducted by Ghatta et al ¹¹⁷, rapamycin appears to exert its cardioprotective effect in ischemic heart through the ATP-dependent potassium channels, since blocking the channel activity abolishes the rapamycin-mediated protective effect. Similarly, both rapamycin and FK506 that are known to bind with the FK-506 binding protein (FKBP) have been shown to affect both ryanodine receptor and -Ca²⁺ release ¹¹⁸. Therefore, the improved ventricular function in the rapamycin treated animal studies could be partly mediated by the changes in ion channel activity. Whether the concentration of rapamycin for the improved ventricular function of hypertrophied heart was sufficient to alter these ion channels is not known. Second, it is possible that the activation of mTORC1 and its effectors, such as S6K1 and 4EBP, may result in the translation of certain deleterious proteins that contribute to pathological remodeling in addition to certain required proteins for the compensatory growth. In this scenario, it is easy to reconcile how the loss of mTORC1 activity with rapamycin can improve ventricular function. However, in this case, a broader proteomic approach needs to be taken to draw definitive conclusions. Third, another attractive possibility is that rapamycin promotes mTORC2 activation, thereby promoting Akt-mediated physiological growth and protection of hypertrophying myocardium. Previous studies indicate that rapamycin treatment enhances S473 phosphorylation of Akt through enhancing insulin sensitivity ¹¹⁹. In this context, studies demonstrate that cells achieve an increased response to insulin via blocking the desensitization process of the insulin receptor substrate-1 (IRS-1). Here, subsequent to mTORC1 activation, the associated S6K1 activation has been shown to phosphorylate IRS-1, resulting in the loss of an insulin effect. Therefore, rapamycin, by blocking S6K1 activation, has been shown to prevent IRS-1 phosphorylation and desensitization. Although rapamycin has been shown to cause dissociation of mTOR from mTORC1, this effect does not appear to affect the balance between mTORC1 and mTORC2. Our recent work (Johnston and Kuppuswamy, unpublished work) demonstrates an augmented level of insulin-stimulated Akt phosphorylation at S473 by acute low dose rapamycin treatment in adult cardiomyocytes. Thus, rapamycin treatment affects protein synthesis and hypertrophic growth, but it is quite possible that it may contribute to the cardioprotective effect via an Akt-mediated survival mechanism. Collectively, these studies suggest the following: 1) rapamycin reduces hypertrophic growth; 2) it ameliorates hypertrophy-induced heart failure; 3) rapamycin could affect the function of not only mTORC1 but also mTORC2 and other targets if the treatment is longer or at high doses; 4) the precise dose and duration for selective cardiac phenotypic outcome is yet to be determined to reap the full benefits of rapamycin’s protective effects; and 5) care must be taken while interpreting the pharmacological effects of rapamycin by considering dose and duration of rapamycin administration.

Conclusions and Perspectives:

Studies thus far indicate the existence of at least two mTOR complexes with distinct functions, and their role in cardiac biology are beginning to emerge. While mTORC1 and mTORC2 seem to have distinct downstream effectors, it remains to be understood how these two signaling arms are coordinated. For instance, the mechanism of coupling mTORC1-mediated translational activation with mTORC2-mediated cytoskeletal reorganization would answer some key questions in biology, like how protein synthesis and cell volume increase are coordinated. Furthermore, the fact that Akt serves as a prime substrate for mTORC2 implies how mTOR could contribute to cell survival by forming unique complexes with cellular proteins. This would be an important area of research to prevent cardiomyocyte loss in the stressed heart. Although mTOR could very well coordinate many of these processes, the evidence is sparse as to its role at this stage. In addition, the existence of other mTOR complex components outside these two well-established complexes, mTORC1 and mTORC2, creates an additional layer of signaling diversity and needs to be fully understood for therapeutic targeting.

Based on the importance of Akt activation for physiological hypertrophy, the activation of mTORC2 that functions upstream of Akt could be perceived to improve ventricular function of the stressed heart; however, mTORC1, which is activated initially as part of the compensatory growth via translational regulation, may also contribute to the transition of the decompensatory phase observed in pressure overloaded myocardium. Therefore, additional studies are needed to establish whether promoting the role of mTORC2, while simultaneously suppressing mTORC1 function would improve ventricular function of the failing heart. The data obtained to date with rapamycin as a prospective cardio-therapeutic drug is encouraging. However, possible long term and off-target effects need to be ascertained to enhance the clinical usability of this drug and its analogues. Future studies on mTOR signaling pathways can be expected to unveil potential new drug targets for patients with heart disease.