Angiotensin II has a positive inotropic effect in cardiomyocytes both by myofilament Ca²⁺ sensitization and intracellular Ca²⁺ rise (Mattiazzi, 1997; Watanabe and Endoh, 1998). Experiments in isolated rabbit myocytes demonstrated that Ang II induced a positive inotropic effect without associated increases in either Ca²⁺ inward current or Ca²⁺ transients, but was accompanied by intracellular alkalosis that could potentially increase myofilament sensitivity to Ca²⁺ (Ikenouchi et al., 1994). It has also been shown that Ang II increases phosphorylation of myosin light chain 2 (MLC2) by the MLC–PKC pathway, thereby increasing myofilament Ca²⁺ responsiveness (Morano et al., 1988; Clement et al., 1992). This Ang II–AT₁R activation of myofilaments could, therefore, form the basis for length-dependent activation (Frank–Starling mechanism). Indeed, Rockman’s group recently revealed that gene deletion or selective inhibition of AT₁R in mouse hearts abrogates the Frank–Starling relationship (Abraham et al., 2016).

Using β-arrestin1/2 deficient mice, they also found that the loss of β-arrestin proteins abrogates Frank–Starling relationship without activating PKC. It has been demonstrated that β-arrestin-biased AT₁R activation enhances myocyte contractility without increasing intracellular Ca²⁺ concentration (Rajagopal et al., 2006), and myofilament Ca²⁺ sensitivity associated with reduced TnI and MyBPC phosphorylation and enhanced tropomyosin phosphorylation (Monasky et al., 2013; Ryba et al., 2017). However, protein phosphorylation in the β-arrestin1/2 deficient mice does not differ compared to wildtype controls. This may be due to distinct β-arrestin signaling pathways downstream of stretch–AT₁R compared to ligand–AT₁R stimulation (Wang et al., 2018). Further detailed investigation of posttranslational modifications of myofilament proteins by proteomic analyses will identify novel proteins critical to the AT₁R-dependent modulation of the Frank–Starling relationship.

In addition to the positive inotropic effect caused by myofilament Ca²⁺ sensitization, Ang II enhances contractility by a Ca²⁺ dependent mechanism. It has been proposed that Ang II–AT₁R binding triggers endothelin-1 (ET-1) production/release, which in turn activates endothelin type A receptor (ET_AR) to induce transactivation of epidermal growth factor receptor (EGFR) through ROS-induced ROS release (Cingolani et al., 2006; Yeves et al., 2015; Zhang et al., 2015). This results in NHE-1 activation to induce Na⁺ influx that in turn triggers Ca²⁺ entry via reverse mode NCX, thereby enhancing contractility (Petroff et al., 2000; Pérez et al., 2003) (Figure 3, red arrows). It has been suggested that these mechanisms could be the basis for the SFR (the in vitro equivalent of the Anrep effect). Cingolani’s group has described a similar complex signaling pathway upon stretch-induced Ang II production, involving autocrine/paracrine activation of AT₁R and ET_AR (Cingolani et al., 2011), ROS-induced ROS release (Caldiz et al., 2007), the transactivation of EGFR (Brea et al., 2016), ERK1/2 activation (Pérez et al., 2011), and NHE-1 and NCX activation (Pérez et al., 2001), thus increasing myocyte contraction (Figure 3, red arrows).

Recently, we demonstrated that transient receptor potential canonical (TRPC) 6 channel is an additional important upstream determinant of SFR. Our study demonstrated that gene deletion and selective blockade of TRPC6 channel abrogates SFR and the Anrep effect in physiologically loaded myocytes, muscle strips, and intact hearts (Seo et al., 2014a). More recently, it was found that TRPC3 is an equivalent contributor to this process (Yamaguchi et al., 2018) as supported by evidence that TRPC3 and 6 can form functional heterotetramers (Hofmann et al., 2002). The mechanosensitivity of TRPC6 had been demonstrated in smooth muscle cells (Spassova et al., 2006) and adult cardiomyocytes (Dyachenko et al., 2009) exposed to membrane stretch or shear stress. Some, however, have questioned the mechanosensing capacity of TRPC6 and attributed it to its upstream stretch-activated GPCRs and/or artifacts from TRPC6 overexpression in heterologous systems (Gottlieb et al., 2008). Importantly, TRPC3/6 is a downstream component of AT₁R, and Ang II stimulation is known to activate these channels via G_αq protein, PLC and diacylglycerol (DAG) signaling pathway (Onohara et al., 2006). Indeed, stretch-induced slow increase in Ca²⁺ was suppressed by either TRPC3/6 or AT₁R inhibition (Yamaguchi et al., 2018). Ca²⁺ influx through TRPC3/6 may directly stimulate SFR, while it is also conceivable that TRPC3/6 activates ERK1/2 (Yao et al., 2009; Chiluiza et al., 2013) upstream of NHE-1, thus inducing an inward Na⁺ current that in turn triggers Ca²⁺ entry via NCX (Poburko et al., 2007, 2008; Louhivuori et al., 2010) (Figure 3, blue arrows). While many of these studies have suggested the involvement of AT₁R as an upstream component of SFR signaling pathways, there are several opposing reports describing that Ang II and AT₁R are not involved in the process because SFR was not suppressed by Ang II receptor blockers (ARBs) (Calaghan and White, 2001; von Lewinski et al., 2003; Shen et al., 2013). Despite the inverse agonistic effect of ARBs (Sato et al., 2016), mechanical stretch may differentially activate AT₁R even in the presence of ARBs which should normally suppress the ligand-stimulated signaling pathway. More direct evidence that links AT₁R to SFR, using AT₁R-deficient models, will be needed to clarify this mechanism.

It is well established that Ang II can induce cardiomyocyte hypertrophy through the activation of multiple intracellular signaling pathways such as the mitogen-activated protein kinase (MAPK) signaling cascade (Thorburn et al., 1994; Zechner et al., 1997; Wang et al., 1998a), c-Jun N-terminal kinase (Sadoshima and Izumo, 1993; Ramirez et al., 1997; Wang et al., 1998b), Akt–mammalian target of rapamycin (mTOR) (Gorin et al., 2001; Diniz et al., 2009) and calcium–calmodulin-activated phosphatase calcineurin (Molkentin et al., 1998; Wilkins and Molkentin, 2004). In particular, calcineurin is a pivotal regulator of pathological cardiac hypertrophy preferentially activated by mechanical stress on GPCRs (Molkentin, 2004; Heineke and Molkentin, 2006). Once activated by increases in Ca²⁺, calcineurin mediates the hypertrophic response through its downstream transcriptional effector NFAT (Crabtree and Olson, 2002). It has been accepted that a major source of Ca²⁺ for activation of calcineurin is Ca²⁺ influx through TPRC3 and 6 channels. Indeed, Ang II–AT₁R activation promotes calcineurin–NFAT signaling that requires the DAG-induced Ca²⁺ signaling pathway through TRPC3 and 6 (Onohara et al., 2006). Importantly, this signaling pathway can be activated by mechanical stress. TRPC3 overexpressed transgenic mice exhibit an increase in calcineurin–NFAT activation in vivo, and increased hypertrophy after Ang II/Phenylephrine and pressure-overload stimulation (Nakayama et al., 2006). In addition, TRPC6 transgenic mice also resulted in enhanced sensitivity to mechanical stress, with an increase in calcineurin–NFAT signaling, and severe cardiac hypertrophy and failure (Kuwahara et al., 2006). Although TRPC3/6 channels are linked to other G_αqPCRs such as ET_AR and α-adrenergic receptors, AT₁R may be the putative central actor in stress-induced hypertrophy, considering the mechanosensing capacity of the receptor.

AT₁R activation by either Ang II or mechanical stress not only induces pathological signaling but also promotes physiological hypertrophy and prosurvival signaling. In embryonic, neonatal, and adult cardiomyocytes, Ang II–AT₁R activation promotes transactivation of EGFR, which in turn activates MAPK and Akt–mTOR pathways (Thomas et al., 2002; Diniz et al., 2009). Rockman’s group recently demonstrated that mechanical stress in cells and the hearts activates AT₁R-induced prosurvival signaling in a β-arrestin-dependent manner that does not require Ang II release (Rakesh et al., 2010). The formation of an AT₁R–β-arrestin complex by mechanical stress induces EGFR transactivation and subsequent ERK and Akt signaling pathways, which suppresses cardiomyocyte injury (Rakesh et al., 2010). They later found that membrane stretch uniquely promotes the coupling of the inhibitory G protein (G_αi) that is required for the recruitment of β-arrestin2 and activation of downstream signaling pathways, such as EGFR transactivation and ERK phosphorylation (Wang et al., 2018). G_αi proteins primarily inhibit the cAMP-dependent pathway by inhibiting adenylyl cyclase (AC) activity. Although previous studies have shown that Ang II may promote AT₁R–G_αi coupling to inhibit AC and to regulate Ca²⁺ channels in certain tissues or cell types (Hescheler et al., 1988; Maturana et al., 1999), it is not yet clear if the G_αi–β-arrestin complex has the same function.

Taken together, chronic mechanical stress can induce both G_αq–TRPC3/6 dependent and G_αi–β-arrestin dependent signaling pathways to differentially promote pathological and prosurvival signaling (Figure 4). Although selective G_αq–TRPC3/6 signaling initially works as an adaptive response to mechanical stress by enhancing myocardial contractility through SFR, prolonged signaling eventually worsens cardiac function. On the other hand, β-arrestin-dependent AT₁R signaling is proposed to enhance cardiac contractility by a Ca²⁺-independent mechanism, and chronically activates prosurvival signaling, making it a pathway of high clinical potential to ameliorate acute and chronic heart failure.

Pathological hypertrophy induced by the overstimulation of AT₁R by Ang II or mechanical stress can eventually lead to heart failure and sudden death associated with arrhythmia. One of the current therapeutics for these conditions is AT₁R blocking drugs, known as ARBs. Several ARBs are known to have inverse agonistic action which can inactivate the GPCR state, and thereby suppress the constitutive activity of receptors. Such drugs can suppress mechanical stretch-induced signals through AT₁R and may exhibit enhanced therapeutic effects for these disease states (Zou et al., 2004; Wei et al., 2011).

While endogenous Ang II activates both G protein and β-arrestin signaling pathways, several AT₁R ligands, such as [Sar1, Ile4, Ile8]-Ang II (SII), TRV120023, TRV120027, and TRV120067 have been shown to selectively activate β-arrestin-mediated pathways and therefore termed as β-arrestin-biased agonists (Rajagopal et al., 2010; Strachan et al., 2014). These β-arrestin-biased AT₁R agonists have been demonstrated to have beneficial effects on the disease condition caused by mechanical stress. For example, TRV120027 causes cardiac unloading action while preserving renal function in a canine model of acute heart failure (Boerrigter et al., 2011). However, a Phase 2B trial of TRV120027 in acute heart failure (BLAST-AHF) resulted in no benefit over placebo (De Vecchis et al., 2017). On the other hand, TRV120023 diminishes myocyte apoptosis caused by mechanical stress by selectively activating ERK1/2 cardioprotective signaling pathways in isolated mouse hearts (Kim et al., 2012). Long-term treatment with TRV120067 in the mouse model of dilated cardiomyopathy not only prevented maladaptive signaling but also improved cardiac function by altering the myofilament response to Ca²⁺ via β-arrestin signaling pathways (Ryba et al., 2017). Notably, TRV120023 and TRV120067 have shown better efficacy in cardiac function and cardioprotection compared to ARBs, suggesting their possibility to become novel therapeutic drugs for heart failure.

Targeting AT₁R downstream signals, such as protein kinase G (PKG) and TRPC3/6, could be another therapeutic strategy. Our study previously revealed that adverse myocardial responses induced by mechanical stimulation to these channels are suppressed by post-transcriptional modification of TRPC6 channel by activation of the cGMP–PKG pathway (Seo et al., 2014a). Indeed, there is growing evidence that stimulation of the cGMP–PKG pathway within cardiac myocytes dampens cardiac stress responses, and its activation can attenuate pathological hypertrophy, protect against ischemic injury and enhance cell survival (Zhang and Kass, 2011). One means to activate the cGMP–PKG pathway is to inhibit the degradation of cGMP by phosphodiesterase-5 (PDE5). PDE5 inhibitors [e.g., sildenafil (Viagra)] have proven their efficacy in treating pressure-overload cardiac hypertrophy and failure in animal models (Takimoto et al., 2005; Nagayama et al., 2009). Although the RELAX trial, a large clinical trial of PDE5 inhibition in heart failure with preserved ejection fraction (HFpEF), failed to show robust beneficial effects (Redfield et al., 2013), single-center trials in patients with heart failure with reduced ejection fraction (HFrEF) reported improved exercise capacity and quality of life (Lewis et al., 2007). Indeed, a meta-analysis of randomized controlled trials in heart failure shows statistically significant improvement of clinical outcomes in patients with HFrEF rather than HFpEF (De Vecchis et al., 2017). Our study demonstrated that sildenafil attenuates pathological hypertrophy by promoting TRPC6 phosphorylation by PKG in the mouse model of muscular dystrophy in which the heart is susceptible to mechanical load (Seo et al., 2014a). Direct antagonism of TRPC3/6 channels is also proven effective for preventing pathological hypertrophy in experiment levels (Seo et al., 2014b). Recently, the orally bioavailable selective TRPC6 inhibitor (BI 749327) was tested in mice, providing in vivo evidence of therapeutic efficacy for cardiac and renal stress-induced disease with fibrosis (Lin et al., 2019). Thus, direct inhibition of TRPC6 could be an alternative strategy to effectively suppress pathological cardiac hypertrophy and failure induced by adverse mechanical stress.

A potent inotropic effect of apelin has been demonstrated in cardiomyocytes (Farkasfalvi et al., 2007; Wang et al., 2008; Peyronnet et al., 2017), muscle strips (Dai et al., 2006), isolated hearts (Szokodi et al., 2002; Perjés et al., 2014) and in vivo heart disease models (Berry et al., 2004; Charo et al., 2009). It has been proposed that the increase in myocardial contractility is attributed to both Ca²⁺-dependent (Dai et al., 2006; Wang et al., 2008) and Ca²⁺-independent mechanisms (Farkasfalvi et al., 2007; Charo et al., 2009; Parikh et al., 2018). The latter is considered to rely on apelin’s action on myofilament sensitivity to Ca²⁺. It was first demonstrated by Farkasfalvi et al. (2007) in studies with normal and failing cardiomyocytes that displayed increased sarcomere shortening in the absence of increased Ca²⁺ transient amplitude after apelin administration. They suggested one of the mechanisms involves increased myofilament sensitivity to Ca²⁺ as apelin activates NHE-1 with consequent intracellular alkalinization. Subsequently, a study using isolated single left ventricular myocytes from apelin deficient and APJ deficient mice revealed that loss of apelin or APJ causes impaired contraction with no difference in intracellular Ca²⁺ kinetics, suggesting apelin and APJ affect either myofilament Ca²⁺ sensitivity or cross-bridge cycling kinetics (Charo et al., 2009). This may be attributable to the activation of MLC kinase through parallel and independent activation of PKCε and ERK1/2 signaling stimulated by apelin (Perjés et al., 2014).

The phosphorylation level of myofilament proteins can affect length-dependent activation, which positively regulates the Frank–Starling relationship. Recently, the direct effect of apelin on length-dependent activation was examined in mechanically preloaded cardiomyocytes; apelin increased compliance of the myocytes as indicated by the negative regulation of end-diastolic force-length relationship, which in turn enhanced contractility as indicated by increased Frank–Starling gain (dimensionless index for contractility) (Peyronnet et al., 2017). Increased cardiomyocyte compliance is presumably related to titin phosphorylation. However, the observed positive regulation of the Frank–Starling relationship by apelin must be dependent on other contractile proteins such as cTnI and MLC or alkalosis, because decreased titin-based stiffness is associated with reduced length-dependent activation of myocardium (Methawasin et al., 2014; Ait-Mou et al., 2016; Beqqali et al., 2016). Recently, we demonstrated that APJ-deficient cardiomyocytes showed negative regulation of the Frank–Starling relationship with no increase in Ca²⁺ transients in response to stretch (Parikh et al., 2018). Our study also provided mechanistic insights for apelin’s positive inotropic and Frank–Starling effects. We demonstrated reduced protein kinase A (PKA) phosphorylation of cTnI at Ser²²/Ser²³ in response to apelin. This is known to increase myofilament Ca²⁺ sensitivity, and is consistent with apelin-dependent AC–cAMP–PKA inhibition through G_αi activation (Szokodi et al., 2002; Scimia et al., 2012) (Figure 5).

In addition to this Ca²⁺-independent mechanism, apelin is thought to exert its inotropic action by increasing the availability of intracellular Ca²⁺. Apelin’s inotropic effect in isolated hearts is dependent on PLC, PKC, NHE-1, and NCX activation (Szokodi et al., 2002). Notably, this inotropic response develops slowly to reach a plateau within 10–30 min and is then sustained, which is different from classical β-adrenergic receptor activation that develops rapidly over a matter of seconds. Although Ca²⁺ dynamics were not examined in this study, a later study proposed that apelin-induced increased contractility is the result of increased Ca²⁺ transients rather than changes in myofilament Ca²⁺ responsiveness (Dai et al., 2006). Most recently, these observations were consolidated in a report showing that apelin has positive inotropic and lusitropic actions on isolated myocytes with enhanced calcium-induced calcium release. This enhanced Ca²⁺ release is achieved through increased Ca²⁺ influx through NCX and increased rate of Ca²⁺ uptake to Ca²⁺ storage by sarcoplasmic reticulum Ca²⁺-ATPase (SERCA), as controlled by PKC-directed phosphorylation (Wang et al., 2008). It is intriguing that the time course and the signaling pathways underlying the effect of apelin show some similarities to the mechanism of SFR (Figure 6 compared to Figure 3), in which NHE-1 and NCX are the primary downstream actors. Although the role of APJ in SFR has yet to be fully explored, it is conceivable that this signaling pathway modulates this physiological phenomenon both through apelin–APJ binding and APJ’s mechanosensing ability.

We previously demonstrated a beneficial effect of chronic apelin supplementation on cardiac performance with reduced left ventricular preload and afterload, and increased contractile function without evidence of hypertrophy (Ashley et al., 2005). Chronic stimulation of APJ by apelin not only increases cardiac performance but also attenuates the development of pressure-overload heart failure through the inhibition of TGF-β-driven profibrotic activity (Pchejetski et al., 2012) and NOX2-derived ROS production (Koguchi et al., 2012). In apelin-deficient mice, hypertrophic response to pressure overload was unchanged, but the progressive impairment of systolic function was observed (Kuba et al., 2007). Myocardial infarcted hearts in apelin deficient mice exhibit exacerbated postinfarction remodeling and impaired functional recovery with a significant reduction of prosurvival phospho-Akt and ERK1/2 signals in the infarct and peri-infarct regions (Wang et al., 2003). These studies clearly demonstrate the simultaneous inotropic and antihypertrophic effects of apelin in ischemia and pressure overload.

Conversely, cardiomyocyte-specific overexpression of APJ causes cardiac hypertrophy and contractile dysfunction in mice (Murata et al., 2016). This indicates that APJ has a capacity to activate multiple downstream signaling pathways such as G_αi- and G_αq-dependent pathways (Chapman et al., 2014), some of which may be independent of apelin’s protective effects. It has also been shown that APJ can integrate chemical (apelin) and mechanical (stretch) stimuli and translates these into opposite cardiac outcomes by differentially activating downstream pathways (Scimia et al., 2012). Specifically, apelin activates APJ through G_αi protein to exert its cardioprotective effect, while stretch stimulates APJ to recruit β-arrestins, which promote pathological hypertrophy (Figure 7). At the cellular level, the mechanosensing capacity of APJ has been confirmed in H9c2 cardiomyocytes in which the increase in diameter, volume and protein content of cardiomyocytes under static pressure was ameliorated by APJ shRNA (Xie et al., 2014). Recently, our study provided supporting evidence that myocyte-specific deletion of APJ is protective against pressure-overload heart failure, showing the abrogation of mechanosensing capacity, reduced Ca²⁺ transient, and remarkable suppression of cellular hypertrophy and fibrosis (Parikh et al., 2018). These studies suggest that APJ integrates apelin and stretch stimuli, biasing the level of G protein versus β-arrestin signaling to attenuate or stimulate hypertrophy. While the downstream mechanisms of stretch-induced β-arrestin signaling remain undefined, the function of β-arrestins has also been observed to block the interaction of APJ with G_αi proteins (i.e., desensitization), which may contribute to the pro-hypertrophic program (Scimia et al., 2012). These studies have implications for the consideration of APJ as a drug target, because the greatest benefit may be obtained not simply by apelin stimulation, but rather by selectively activating the G_αi-dependent signaling pathway or by inhibiting the ability of APJ to respond to mechanical stretch. For this purpose, further efforts are critical to clarify the intricacies of downstream integration of ligand and mechanosensitive signaling by APJ.

Apelin’s positive inotropy and anti-hypertrophic effects support its therapeutic potential in preventing and treating cardiovascular disease. The majority of current heart failure therapies are targeted at the inhibition of deleterious neurohormonal axes that are upregulated in the later stages of the disease. In this aspect, the apelin–APJ system is attractive because it appears to be downregulated in heart failure (Japp and Newby, 2008), and stimulation of this pathway may have additive or even synergistic efficacy to current therapy by targeting complementary but separate pathways. Indeed, intravenous administration of [Pyr¹] apelin-13, an active fragment of apelin, in heart failure patients showed efficacy with peripheral and coronary vasodilatation and increases in cardiac output (Japp et al., 2010).

Nonetheless, the therapeutic application of apelin is limited because of its extremely short biological half-life and parenteral administration. This is attributed to the degradation by endogenous proteases circulating in the blood (Japp et al., 2008; Zhen et al., 2013). Furthermore, apelin is hydrolyzed and partially deactivated by angiotensin I converting enzyme 2 (ACE2) (Vickers et al., 2002; Wang et al., 2016). Thus, many efforts have been directed to the development of apelin analogs or novel agonists of APJ (Yamaleyeva et al., 2016). In particular, combinatorial screening of the NIH small molecule library identified a full non-peptide APJ agonist, ML233, at activating both G_αi- and β-arrestin-dependent pathways (Khan et al., 2011). In addition, using molecular dynamics simulations, Brame et al. (2015) designed cyclic analogs and identified a biased agonist, MM07, which activates APJ by preferentially stimulating G_αi-dependent pathways but not β-arrestin. Because stimulation of β-arrestin pathway could be deleterious, G protein-biased APJ ligand MM07 represents a potential new therapeutic for heart failure.