Sympatho-adrenergic mechanisms in heart failure: new insights into pathophysiology

Evidence for activation of the sympatho-βAR system in HF

Clinical studies on patients with HF have contributed to the consensus that the enhanced sympatho-βAR activity represents compensatory mechanism in response hemodynamic instability. By using a range of techniques [5], SNS overactivity in human HF patients can be ensured by: (1) increased circulating levels of NE; (2) enhanced spillover of cardiac sympathetic neurotransmitter, (3) augmented sympathetic neuronal firing rates [6], and (4) blunted β₁AR response due to sustained stimulation by neuronal or circulation derived catecholamines. In patients with HF, elevated circulating levels of NE and epinephrine bear prognostic implication [7]. Esler et al. [8] furthered this finding by measuring NE spillover through administration of a radiotracer (³H-NE) together with blood sampling using catheter positioned in the coronary sinus. Kinetics of regional sympathetic neurotransmission is able to determine, by using high-performance liquid chromatography, NE and a variety of its precursors or metabolites. Application of this technique to HF patients revealed enhanced NE release from cardiac sympathetic nerves together with faulty neuronal NE reuptake that further exaggerated rise in NE levels [4]. Patients with higher cardiac NE spillover rate had adverse long-term prognosis [4, 9, 10]. Direct recording of sympathetic nerve firing rates by using microneurography is able to determine muscle sympathetic nerve activity (MSNA). A recent meta-analysis of clinical studies on HF patients indicated that progressive increase in MSNA was observed in association with NYHF class [6], or with classification based on left ventricular ejection fraction [LVEF, i.e. heart failure with preserved ejection fraction (HFpEF), heart failure with mid-range ejection fraction (HFmrEF), and heart failure with reduced ejection fraction (HFrEF)] [11]. Moreover, MSNA was correlated with heart rate, LV end-diastolic dimension (EDD), LVEF, brain natriuretic peptide (BNP) or plasma NE levels [11]. Likely due to non-invasive in nature, heart rate variability (HRV) has been commonly used to estimate autonomic nervous activity. Whilst increased lower-frequency component of HRV is widely viewed as a parameter for enhanced SNS activity, studies using other more direct methodologies, as discussed above, suggest that the lower frequency of HRV is not directly related to cardiac sympathetic nervous activity [12]. Recently, nuclear imaging technique has been applied into the assessment of cardiac sympathetic nervous activity in patients with HF [13]. Current studies indicate that SNS imaging might be used for risk assessment [14], treatment monitoring, and matching treatment strategies with disease states. However, prospective and controlled trials are essential to confirm the usefulness of SNS imaging indices in guiding clinical management of HF with better outcomes.

In the setting of HF, activation of the sympatho-βAR system has been well documented to exert pro-arrhythmic action, especially in the presence of substantial myocardial remodeling and fibrosis [15, 16]. Enhanced sympatho-βAR activity is known to evoke electrical instability via mechanisms involving induction of diastolic Ca²⁺ leaky through RyR [17, 18], and reduction in slow-component of delayed rectifier K⁺ current (I _Ks) leading to prolongation of action potential duration [19]. Using cardiac NE spillover methodology to assess NE spillover and store, Brunner-La Rocca et al. [9] revealed in HF patients, that a lower NE store together with higher NE release predicted risk of fatality due to HF worsening, whilst the combination of large NE store and high NE spillover indicated risk of sudden death. Direct evidence for the pro-arrhythmic property of SNS activity derived from preclinical and clinical studies showing that sympathetic denervation is protective against ventricular tachyarrhythmias in diseased settings like acute MI, long QT syndrome, channelopathies, cardiomyopathy or HF [20], [21], [22], [23]. A key role of the autonomic nervous system in the pathogenesis of atrial fibrillation has also been documented [24]. In pacing-induced canine model of HF, recording of nerve activity of the sympathetic stellate ganglion revealed augmented nerve firing that was directly associated with paroxysmal atrial tachycardia in these dogs with HF [25]. Using the same pacing-induced canine HF model, atrial tachyarrhythmias were markedly inhibited by ablation of bilateral stellate and thoracic ganglia [24]. For a commonly used anti-arrhythmic drug amiodarone, experimental and clinical studies have revealed its reserpine-like “sympatholytic” action, by which neurotransmitter storage of cardiac sympathetic nerves is partially exhausted [26], [27], [28]. Interestingly, chronic treatment with amiodarone to a rat model of dilated cardiomyopathy (DCM) induced by autoimmune myocarditis partially restored cardiac sympatho-β₁AR activity [28], suggesting the potential of chronic drug therapy-induced neuromodulation of the failing heart.

Adverse biological consequences via βAR signaling

Numerous studies have established the notion that intense or sustained activity of the sympatho-βAR system results in myocardial remodelling at molecular, cellular and organ levels and therefore is maladaptive (Figure 1). Observational studies have documented that in patients with HF, elevated levels of physiological and biochemical parameters of SNS activity predict poor prognosis [4, 7, 29]. In preclinical studies, administration of β-agonists either as a toxic dosage or by continuous delivery, regimes simulating sympathetic storm or constant hyperactivity, results in cardiac damage and remodeling featured by cardiomyocyte loss, inflammatory infiltration, interstitial fibrosis and dysfunction [30, 31]. Supporting evidence also comes from studies using genetically modified mouse models with enhanced β-adrenergic signaling due to cardiac overexpression of βARs or downstream signaling molecules like AC and PKA [32], [33], [34]. Understanding the downstream βAR biology and cardiotoxic signalling has set up the foundation for the clinical use of β-blockers. Clinical use of β-blockers are beneficial in alleviating symptoms, limiting remodelling and improving cardiac function. Importantly, β-blocking therapy significantly improved long-term prognosis in patients from diagnosis of HF [35, 36]. Prolonged receptor exposure to catecholamines results in increased βAR-phosphorylation by PKA or GPCR kinases (GRKs). In addition to uncoupling βAR from G-proteins, there is increased association of βAR-β-arrestins, which recruits other signal proteins thereby promoting non-classical signaling (vide infra). There is evidence that diverse βAR signaling is more likely to occur in the failing heart contributing to adverse biological consequences. For instance, the cardiomyopathy phenotype seen in transgenic (TG) mice overexpression β₂AR (β₂-TG) was largely reversed by crossing with mice expressing dominant-negative mutant p38-mitogen-activated protein kinase (MAPK) [37]. In contrast, studies have also implicated detrimental or cardioprotective signaling via βAR–β-arrestin coupling [38, 39].

β₁AR downregulation and desensitization

Under conditions of cardiac sympathetic hyperactivity together with elevated circulating levels of catecholamines, previous studies have shown β₁AR downregulation, i.e. loss of surface membrane β₁AR together with attenuated β₁AR responsiveness, i.e. receptor desensitization [40, 41]. The causal relationship of sympathetic overdrive and β₁AR downregulation/desensitization in the failing heart has been well documented [41]. Clinically, augmented NE kinetics is correlated with the substantial loss of myocardial β₁AR in biopsies of HF patients [4, 8, 42]. Experimentally, in rapid-pacing induced canine HF model, isotope-trace based assay revealed a reverse correlation between interstitial NE levels and β₁AR density (r = −0.848, P < 0.001) [43]. The process of β₁AR downregulation is a consequence of sustained receptor stimulation that initiates receptor phosphorylation by GRKs and receptor binding by β-arrestins. β-arrestin-bound β₁AR then translocate from cell membrane to intracellular vesicles, where β₁AR either recycle back to membrane or undergo degradation [29, 44]. Another key factor contributing to the loss of β₁AR in the failing myocardium is downregulation of gene transcription of β₁AR (but not β₂AR), which is commonly observed by a scale of 50%–70% relative to controls by studies on numerous animal models of HF or myocardial biopsy of from human patients. Using quantitative autoradiography, Elnatan et al. [45] studied biopsies from various regions of the heart from HF patients, and observed that β₂AR accounted for 36%–72% of total βAR, indicating substantial loss of β₁AR. Desensitization of remaining β₁AR starts from receptor phosphorylation either by PKA or by GRKs, followed by β-arrestin binding but uncoupling from G-proteins [40, 46]. Importantly, distinct differences exist between β₁AR and β₂AR with β₂-subtype maintaining its density albeit receptor relocation occurs, which is accompanied with altered intracellular signaling (vide infra). However, earlier studies also showed an enhanced β₂AR-Gi signaling in the failing heart, likely due to increased Gi-protein abundance and β₂AR phosphorylation by PKA that increases β₂AR–Gi coupling whereas reduces Gs-coupling [47]. The β₂AR-Gi coupling is known to exert inhibitory regulation of β₁AR responsiveness [40, 46].

Establishment of the β-blocking therapy for HF patients

It has been nearly 60 years since the initial discovery made by James Black et al. [48] on the first β-antagonist propranolol. This work was followed by subsequent earlier clinical testing on patients with HF [49]. Swedish cardiologist Dr. Waagstein pioneered clinical testing during 1970s for treatment of patients with chronic HF using β-blockers [50]. Although the very first paper was published in 1975 [49], it took almost two decades for this class of drugs to be accepted as a front-line medicine for patients with HF [51, 52]. The current consensus is that β-blockers may protect the heart against adverse biological effects owing to sympathetic overdrives. Additionally, β-blockers might also suppress activity of the renin-angiotensin-aldosterone system by blocking renin release following renal sympatho-βAR activation, and inhibit inflammatory signalling [5, 44]. Earlier and recent meta-analysis of randomized clinical trials have confirmed the efficacy of β-blocking therapy on patients with the spectrum of HF syndrome, judged by hard clinical outcomes like re-hospitalization, sudden cardiac death, cardiovascular mortality or all-cause mortality [35, 36]. There is also evidence for β-blocking therapy to re-sensitize the suppressed β₁AR signaling in the failing heart [44].

Increased appreciation on the role of β₂AR signaling in heart failure

Recent studies on β₂AR have demonstrated (1) significant receptor redistribution on cell membrane, (2) adverse signaling, and (3) increased potential of pro-arrhythmic action. In the setting of ischemia or simulated ischemic conditions, activation of β₂AR is cardioprotective in large part through its Gi-signalling that opposes βAR-Gs signalling [53]. Alternations in β₂AR signaling in the failing myocardium are also revealed. In the failing heart, GRK2 is activated and mediates β₂AR phosphorylation by which β₂AR-Gi signaling is enhanced, which might facilitate development of HF under certain diseased conditions, such as chronic pressure-overload [54]. Importantly, β₂AR signaling pattern is altered in the failing myocardium. Upon ligand binding, activation of β₁AR of cardiomyocytes increases intracellular level of cAMP and PKA activity in a scale of virtually whole cell, while β₂AR-mediated intracellular changes are restricted to a limited compartment [55, 56]. Such differences are due to diverse receptor localization with β₁AR more evenly distributed on cellular membrane while β₂ARs are embedded within transverse tubules (t-tubules) where other signal proteins, like caveolin-3 and G-proteins, are richly represented forming multi-protein clusters (signalsomes) (Figure 2) [57]. In the failing heart, cardiomyocytes are characterized by overt disruption of t-tubules where synchronization of events like Ca²⁺-influx triggered SR Ca²⁺ release (CICR) occurs [58]. T-tubules also co-localize β₂AR-signalsomes critical for compartmentalized β₂AR signaling (Figure 2) [57]. With loss of t-tubules [56, 57] as well as caveolin-3 [59], β₂AR redistribute to cell surface and, upon ligand stimulation, alters the mode of signaling akin to that of β₁AR (Figure 2) [57, 59]. Surface scans of cardiomyocytes using scanning ion conductance microscopy (SICM) showed that structural alteration of t-tubules of failing cardiomyocytes were associated with regional change of functional βARs, in particular, compartmented β₂AR-cAMP signal was disturbed. Instead, β₂AR-cAMP signal was observed to propagate along the cardiomyocyte [56, 57]. Indeed, an earlier study by Kaumann et al. [60] showed that in failing myocardium of patients with DCM or ischemic cardiomyopathy, stimulation of β₂AR and β₁AR yielded similar extent of downstream signaling and functional responses.

Figure 2:

T-tubule localized β₂AR signalosomes and redistribution of β₂AR in the failing cardiomyocyte. A. Cardiomyocytes are rich in t-tubular structures, where sarcolemma membrane invaginate into the internal space of the cell body. T-tubules contain abundant lipid microdomains and caveolae and are rich in ion channels, e.g. L-type calcium channels, important for membrane potential and excitation-contraction (EC) coupling and mobilization of sarcoplasmic reticulum (SR) Ca²⁺ pool. T-tubules are the location of cluster of signaling proteins including G-proteins, adenylyl cyclase (AC), G-protein receptor kinase (GRK), A-kinase anchoring proteins (AKAP), protein kinase A (PKA), protein phosphatase 2A, and phosphodiesterases (PDE), forming highly efficient signalosomes. β₂AR are also highly enriched in the caveolae/lipid microdomain of t-tubules where it induces local cAMP-PKA and Ca²⁺ signals. In comparison, β₁AR are largely distributed throughout sarcolemma membrane and mediate a global cAMP-PKA and intracellular Ca²⁺ signals. B. In the failing myocardium, there is significant loss of t-tubular network and hence partial disintegration of signaling clusters. Hence, β₂ARs undergo redistribution from t-tubules to sarcolemma membrane, which renders β₂AR signaling upon ligand binding similar to that of β₁AR. AR: adrenergic receptor.

Ventricular dilatation and myocardial remodelling (including hypertrophy and fibrosis) form the pro-arrhythmic substrates [16]. In this setting, sympatho-βAR activation acts as a powerful trigger of arrhythmias. An earlier study by us provided the direct evidence for SNS activation as a trigger of ventricular arrhythmias [15]. Using in situ perfused heart from rats with chronic myocardial infarction (MI) and HF, electrical stimulation of cardiac sympathetic nerves was potent in inducing ventricular tachyarrhythmias with the severity of arrhythmias dependent on the intensity of nerve stimulation and the infarct size [15]. We observed that SNS stimulation-evoked arrhythmias could only be inhibited by the combined use of β₁- and β₂-selective antagonists. Other studies also highlighted the role of β₂AR in the onset of ventricular tachyarrhythmias in the failing heart [18, 61]. In a rabbit model of HF with severe cardiac remodeling and fibrosis, induced by combination of aortic valve regurgitation and pressure-overload, administration of the β₂-agonist zinterol evoked ventricular tachyarrhythmias in failing hearts while having no effect in normal hearts [18]. This pro-arrhythmic action of zinterol was abrogated by the concomitant use of the β₂-antagonist ICI-118,551 while blockade of β₁AR was ineffective. In ventricular myocytes isolated from failing rabbit hearts, stimulation of β₂AR induced increased extent of Ca²⁺ transient amplitudes, SR Ca²⁺ leaking and phosphorylation of phospholamban [18]. In this HF rabbit model, expression and function of β₁AR were markedly downregulated or attenuated whilst that of β₂AR was maintained, changes emulating that seen in human failing heart [18, 44]. Hence β₂-AR activation is arrhythmogenic in the failing heart by increasing the SR-Ca²⁺ loading and the spontaneous release of Ca²⁺ from the SR. In another study on LV myocardium from advanced HF patients, distinct pro-arrhythmic role of β₁AR and β₂AR signaling was observed [61]. Specifically, treatment of the failing human myocardium with selective β₂-, but not β₁-agonists enhanced conduction velocity, shortened action potential duration, and increased intracellular Ca²⁺ transients. β₁AR stimulation only induced automaticity in Purkinje fibers of the sub-endocardium, whereas β₂AR stimulation significantly induced not only automaticity of Purkinje fibers of the sub-endocardium, but also Ca²⁺-mediated delayed afterdepolarizations in the sub-epicardium and mid-myocardium. This heterogeneous ectopic activity by β₂AR stimulation suggests that in human failing myocardium, β₂AR signaling becomes more important than β₁AR in inducing electrophysiological abnormalities across the entire layer of the ventricular wall.

Using genetically modified mouse models with overexpressing in cardiomyocytes of genes encoding β₁AR, β₂AR Gsα, AC or PKA, studies have demonstrated similar cardiomyopathy phenotypes, notably age-dependent onset of HF, fibrosis, ventricular arrhythmias and premature death [32, 33, 62], [63], [64]. With aging, β₂AR transgenic (β₂-TG) mice exhibited fibrotic cardiomyopathy and premature deaths due either to HF or sudden death [32, 65, 66]. In a telemetry study on conscious β₂-TG mice [64], we found that administration of the β₂-antagonist ICI-118,551 significantly reduced the frequency of ventricular tachyarrhythmias, effect only seen in β₂-TG mice with advanced age and severe cardiac fibrosis [64]. This finding suggests that the presence of overt fibrosis as the substrate permits the pro-arrhythmic effect of β₂AR activation.

Using β₂-TG mice at young ages when overt cardiomyopathy is still absence, we studied influence by enhanced β₂AR activity on cardiac adaptation to disease challenge by transverse aorta constriction (TAC) or MI. β₂-TG mice tolerate poorly to chronic pressure-overload with facilitated onset of HF with exacerbated cardiac remodeling, and premature deaths [67, 68]. Under conditions of chronic MI, however, β₂-TG mice displayed better preserved LV function and less severe HF relative to control littermates [69]. Thus, the effect of enhanced activity of β₂AR on HF is aetiology-dependent.

Clinical and experimental studies have generated strong evidence for a vital role of SNS-βAR in the pathogenesis of Takatsubo cardiomyopathy (or stress cardiomyopathy). In patients with Takatsubo cardiomyopathy, circulating levels of catecholamines were markedly increased, even much higher compared with that of patients diagnosed with acute MI [70]. Experimentally, exposure to large doses of β-agonists (isoproterenol, ISO, or epinephrine up to 500 mg/kg) is able to induce temporal and reversible LV dysfunction, and that β₂AR/Giα signalling is critical in Takatsubo cardiomyopathy with acute HF. Paur et al. [71] reported in rats the presence of relatively higher β₂AR density at the apical myocardium relative to the basal region of the heart. Stimulation of myocardial β₂AR preferentially activates the Gi-mediated cardio-suppressant effect that was more prominent in the apical than the basal regions. Accordingly, use of β₂AR antagonist levosimendan was cardioprotective against acute of LV dysfunction and fatality [71].

βAR signaling through EPAC and GRK/β-arrestins

One of the breakthroughs in βAR signalling is the demonstration of the downstream signaling via exchange protein directly activated by cAMP (EPAC), which was initially discovered in 1998 [72]. cAMP is the most well-documented signaling molecule that is generated by membrane-bound AC upon βAR stimulation and subsequently activates PKA. Identified as novel cAMP-binging proteins, EPAC is now emerging as an alternative cAMP effector. Binding of cAMP to the regulatory subunit of EPAC leads to conformational change with exposure of its catalytic region. EPAC1 and EPAC2 are structurally similar multi-domain proteins. Whilst EPAC2 expression is restricted to neuroendocrine tissues, EPAC1 is ubiquitously expressed with a high expression level in the heart. EPAC1 is known to form signalosomes with partner proteins including β-arretins, Ca²⁺/calmodulin-dependent protein kinase (CaMK-II), a small G-protein Rap, within the subcellular compartments, rendering EPAC1 as a multifunctional signaling molecule. There is also evidence for mitochondria-localized EPAC1 that mediates cardiomyocyte death by promoting mitochondrial Ca²⁺ overload and opening of mitochondrial permeability transition pore (mPTP) [73]. Mitochondrial cAMP promotes mitochondrial Ca²⁺ influx from the ER via VDAC1/Mcu (voltage-dependent anion-selective channel-1/mitochondria calcium uniporter) through recruitment of EPAC1 [74]. Indeed, mitochondria Ca²⁺ level plays a role in the regulation of ATP biosynthesis [75]. It remains to be investigated whether such mitochondria-EPAC1 participates in the regulation of energy metabolism in cardiac physiology and, more importantly in failing cardiomyocytes. It has been shown that cAMP can also be generated by soluble AC (sAC), which is activated by Ca²⁺ signal. This extends our understanding on compartmented sAC/cAMP/EPAC signalling. It is likely that sAC/cAMP/EPAC signaling allows for subcellular fine turning of cAMP-mediated signalling that may be directly (via AC/cAMP) or indirectly (via Ca²⁺ signal) linked to the classical βAR/Gs/AC/cAMP/PKA signaling.

Cardiac expression of EPAC1 was markedly increased in various animal models of pathological cardiac remodelling such as chronic catecholamine infusion, pressure overload induced by TAC [72], or in hearts of human patients with end-stage HF due to DCM or ischemic cardiomyopathy [72, 73, 76]. By employing specific gene deficient models or inhibitor, several studies have investigated the role of EPAC1 under diseased conditions. A recent study on both EPAC1-deficient mouse model and EPAC1 inhibitor (AM-001) provided evidence for cardioprotective effects by inactivating EPAC1 in various forms of cardiac stress [77]. Activation of EPAC1 has been shown to disturb Ca²⁺ homeostasis, enhance reactive oxygen species (ROS) generation, stimulate gene expression and exacerbate myocardial remodelling [78]. EPAC1 or EPAC2 KO mice did not exhibit change in baseline cardiac function. Under diseased conditions such as chronic β-agonist stimulation or TAC, however, EPAC1-KO mice exhibited reduced severity of cardiac hypertrophy, cardiomyocyte apoptosis or fibrosis relative to wild-type controls [79, 80]. Pereira et al. [81] studied EPAC1 and EPAC2 deficient mice subjected to TAC for 3-weeks, and observed that stimulation of β₁AR, but not β₂AR, mediated a CAMK-II-dependent SR-Ca²⁺ leak and arrhythmias, which were abolished in EPAC2-KO, but not in EPAC1-KO mice, despite the fact that EPAC2 protein was undetectable in the heart tissues. Whether such action occurs in the setting of HF remains to be investigated.

Significant breakthroughs have also been made in βAR signaling through G-protein receptor kinases (GRKs)/β-arrestins. Among GRK family members, GRK2, GRK3 and GRK5 are highly expressed in the heart [82]. GRKs, particularly GRK2 and GRK5, mediate phosphorylation of βAR that promotes (a) receptor binding by GPCR adapter proteins β-arrestins with subsequent βAR desensitization/downregulation, and (b) β-arrestin-mediated signaling. In diseased and failing hearts, expression and activity of GRK2 and GRK5 are elevated. Therapeutic implication of GRK overactivity is at two folds: First, GRK-phosphorylates βAR thereby leading to receptor downregulation and desensitization. Causal role of GRK in mediating βAR downregulation/desensitization is indicated by studies using mouse models of GRK2 overexpression, or expression of C-terminal domain (βARKct) that prevents GRK2 translocation to the membrane thereby sparing βAR from phosphorylation [83]. By crossing βARKct-TG mice with genetic mice with cardiomyopathy or subjected to chronic disease stress, studies showed improvement in cardiac function and chamber dilatation, and reduction in premature death [83]. Second, GRK/β-arrestin signaling mediate a large part of intracellular events evoked by βAR stimulation in a G-protein independent fashion. GRK2 is known to inhibit insulin signaling pathway and therefore may interfere with cardiac and systemic metabolism [84, 85]. In patients with acute MI, elevated expression of lymphocyte GRK2 during the acute phase was correlated with LV dysfunction and dilatation determined acutely or during a 2-year follow-up period, implying that GRK2 level predicts adverse cardiac outcomes in MI [86]. In a cryogenic MI mouse model, Ciccarelli et al. [87] reported that inhibition of GRK2 using a peptide (C7) improved mitochondrial morphology and function together with improvement in contractile function. Lymperopoulos et al. [88] reported that GRK2 signaling is responsible for downregulation of presynaptic α₂AR, that promotes NE release. In addition, GRK2 activates NF-κB signaling thereby promoting inflammation and hypertrophy. Thus, it is important to further establish the role of GRK2 in HF and test targeted therapeutic interventions. Unlike previous view of functionally interchangeable, recent studies have revealed significant differences between β-arresins1 and β-arrestin2 [83, 89, 90]. Notably, β-arrestin2, but not β-arrestin1, mediates positive inotropy via activation of SERCA2a [91], protects heart against apoptosis through transactivation of epidermal growth factor receptor (EGFR) and extracellular signal-regulated kinase (ERK)1/2 activation, and antagonizes NF-κB inflammatory signaling [91], [92], [93], [94]. In contrast, there is ample evidence for detrimental cardiac effects mediated by β-arrestin1. Unlike β-arrestin2, β-arrestin1 is unable to transactivate EGFR which mediates cardioprotective signaling [83, 89]. When studied under stressed conditions like MI, β-arrestin1 KO mice were not only spared of β₁AR downregulation/desensitization, but also had reduced apoptotic cell death, ventricular remodelling and dysfunction and mortality [89]. Thus, specific pharmacological interventions that regulate the counteractive signaling mediated by β-arrestin1 and β-arrestin2 would be expected to be useful in HF therapy [83].

Emerging evidence for β₁AR-Giα coupling

Whilst traditional view of βAR–Giα coupling is restricted to β₂-subtype, there have been studies showing β₁AR-Giα interaction. Belevych et al. reported that in cultured cardiomyocytes, β₁AR mediated signaling, measured as Ca²⁺ influx via L-type Ca²⁺ channel, was blunted by simultaneous activation of protein kinase C (PKC), effect that was abolished by pertussis toxin, suggesting coupling of β₁AR and Gαi/o [95]. Martin et al. [96] also observed in transfected CHO cells, switching from Gs to Gi coupling following PKA-mediated phosphorylation of β₁AR. These in vitro findings have gained support by recent in vivo studies. In β₁-TG mouse model of cardiomyopathy, knockout of Gαi2 worsened the severity of cardiomyopathy and increased premature death [97], implying that at least under diseased conditions, there exists interaction between β₁AR and Giα. Carvedilol is a non-selective β-blocker with 13-fold more potent as a competitively β₂- than β₁-antagonist in human heart tissues [98]. Interestingly, carvedilol binding to β₁AR initiates β₁AR-Giα interaction that favours β-arrestin-mediated signaling whereas β₁AR-Gs signaling is inactivated [99]. In another study on cardiomyocytes prepared from diabetic mice, treatment with carvedilol promoted β₁AR-Giα coupling leading to activated signaling through phosphatidylinositol-3-kinase (PI3K)/Akt/nitric oxide synthase (NOS3)/-protein kinase G (PKG) [100]. This signaling pathway is cardioprotective measured by reduced myocardial apoptosis or hypertrophy, improved myocardial energy status, and better preservation of contractile function [100]. Thus, it is likely that under diseased settings, β₁AR-Giα interaction activates downstream signaling pathways that involve β-arrestin, PI3K and cGMP/PKG. Such interaction is facilitated by binding to β₁AR of carvedilol as a biased-ligand (but not by other β-blockers).

βAR activates Hippo signaling pathway

Recent studies have documented βAR-Hippo pathway signaling. The Hippo pathway is a highly conserved signalling that plays a vital role in the regulation of organ size during development [101]. The main signal output of Hippo pathway is through yes-associated protein (YAP), which is a transcription co-regulator (Figure 3). YAP orchestrates with transcription factors, like TEA domain transcription factors (TEADs) and regulates expression of numerous target genes. The nuclear localization of YAP is essential for its transcriptional activity. YAP activity is suppressed by upstream kinases, Mst1 (mammalian sterile-like kinase-1) and Lats (large tumour suppressor homolog), through Ser¹²⁷-phosphorylation of YAP [101]. Recent studies have implicated a key role of Hippo-YAP pathway in heart disease [101], [102], [103], [104]. Whereas enhanced YAP activity facilitates post-MI cardiac healing by promoting cardiomyocyte regeneration [101], activation of Hippo signalling with YAP-inactivation, like transgenic overexpression of Mst1 (Mst1-TG) or cardiomyocyte-restricted knockout of YAP (YAP-cKO) or TEAD1 (TEAD1-cKO), similarly results in DCM phenotype with severe interstitial fibrosis [104], [105], [106], [107], [108], implying a common mechanism involved. Studies have shown activated cardiac Hippo signalling with YAP inactivation in patients or animals with a variety of cardiomyopathies [31, 63, 102, 109, 110].

Figure 3:

Diagram depicting βAR signaling through the Hippo pathway in the heart. There is strong evidence for coupling of β-adrenergic and Hippo signaling pathways in the myocardium. βAR stimulation by genetic or pharmacological means effectively activates Hippo signaling featured by activation of upstream kinases Mst1 (mammalian strile-20 like kinase1) and Lats (large tumour suppressor homolog), and enhanced inhibitory Ser¹²⁷-phosphorylation of YAP (yes-associated protein) and its paralogue TAZ (transcription coactivator with PDZ-binding motif). The latter results in cytoplasmic retention of YAP with increased binding to scaffold protein 14-3-3 and subsequent degradation via the ubiqutine/proteasome system. As a consequence, the activity of YAP as transcription co-activator or co-repressor, is turned off leading to altered expression of numerous YAP-target genes. Studies in vitro also suggest βAR/PKA-mediated direct Lats activation bypassing Mst (broken line). MOB1: Mps one binder kinase activator like 1; Sa: salvador homologue; TEAD: TEA-domain transcription factor. ⊗ Indicates epigenetic factors.

Recent studies have demonstrated regulation of the Hippo pathway by GPCRs. βAR-Hippo pathway coupling was firstly implicated by Yu et al. [111] who discovered in mammalian cell lines that overexpression of Gs-coupled glucagon receptor, β₂AR or a constitutively active form of Gsα, increased Ser¹²⁷-YAP phosphorylation. In cell lines like mouse embryonic fibroblasts, Kim et al. also showed that activation of PKA using the AC activator forskolin induced Ser¹²⁷-YAP phosphorylation indicating YAP inactivation by cAMP/PKA signalling [112]. In human adenocarcinoma cell line, stimulation with catecholamines induced Ser¹²⁷-YAP phosphorylation, which was prevented by propranolol [113]. Several lines of recent findings, largely from in vivo studies, have implicated that activation of βAR-Hippo signaling with YAP inactivation contribute to pathogenesis of heart disease (Figure 3). First, pharmacological (i.e. ISO) or transgenic activation of βAR (i.e. β₁-TG, β₂-TG) leads to cardiac upregulation of Mst1 [31, 63]. ISO stimulation also induces a dose- and time-dependent Ser¹²⁷-YAP phosphorylation in the heart [31]. Second, activation of Hippo signaling by βAR stimulation involves PKA and Mst1. Cell experiments showed that ISO-mediated Mst1 upregulation/YAP phosphorylation were abolished by PKA inhibitors but mimicked with the use of forskolin, implying PKA-dependent signalling (Figure 3) [31]. In β₁-TG mice with cardiomyopathy, inactivation of Mst1 by crossing with TG mice harbouring dominant-negative mutant Mst1 gene (dnMst1-TG with inhibition of native Mst1 activity) largely abolished cardiomyopathy phenotypes (e.g. inhibited cardiomyocyte apoptosis and fibrosis, and improved LV function) [63], findings that underscore a pivotal role of Mst1 signaling in this β₁-TG model of cardiomyopathy. Similar protection against ISO-cardiotoxicity was observed in Mst1-KO mice [63]. Third, the operation of βAR-Hippo signalling pathway promotes cardiac expression of genes like galectin-3 (pro-inflammatory and pro-fibrotic protein), CTGF (connective tissue growth factor (CTGF) and Bcl-2 interacting mediator of cell death (BIM), an upstream pro-apoptotic protein) that are known to contribute to the development and worsening of heart disease (Figure 3) [31, 107, 114, 115]. In dnMst1-TG mice [108], ISO-induced upregulation of galectin-3 and BIM was significantly inhibited [31]. In contrast, in Mst1-TG model with activated Hippo signaling and a markedly high galectin-3 content, ISO-induced net increment in galectin-3 expression was 10-times greater relative to that in control animals [31]. Finally, in cardiomyoblast H9c2 cells, ISO-induced Ser¹²⁷-YAP phosphorylation and galectin-3 upregulation, effects that were simulated by siRNA-mediated YAP knock-down [31]. Similarly, in the Mst1-TG model with YAP inactivation [31], RNA sequencing revealed that 68% of YAP-target genes are upregulated [31]. These findings indicate that in addition to transcription co-activator, YAP also acts as a co-repressor for expression of certain genes (such as galectin-3), hence removal of YAP inhibitory signal as a result of activation of the βAR-Hippo signaling results in upregulation of these genes (Figure 3) [31]. With increasing amount of evidence for a role of Hippo pathway in heart disease, further research is warranted to investigate the βAR-Hippo signaling in development of HF and the therapeutic effect using β-blockers.

Autoantibodies against βAR

Autoantibodies against β₁AR (β₁AR-AAb) could be detectable in 30%–90% of patients with DCM or up to 55% of patients with ischemic cardiomyopathy [116], [117], [118], [119]. Studies have indicated that β₁AR-AAb exert diverse allosteric effects, ranging from inhibitory to agonist-promoting activities, and that presence of β₁AR-AAb negatively impacts LV function [116, 120], albeit there are contradictory reports. Among DCM patients, presence of activating β₁AR-AAb was associated with severe ventricular dysfunction, higher risk of ventricular tachyarrhythmias and sudden cardiac death, and hence might be involved in the progression of DCM [117, 121]. In a clinical study consisting of 2,062 patients diagnosed with chronic HF due to DCM or ischemic cardiomyopathy, 379 (21.6%) cases died during a 36-month follow-up period. Whereas patients with positive β₁AR-AAb had comparable fatality due to non-arrhythmic causes, their risk of mortality in the form of sudden cardiac death was markedly higher (hazard ratio = 4.5 for DCM, HR = 3.75 for ischemic cardiomyopathy) [122]. β₁AR-AAb may become undetectable following implantation of LV assistant device or anti-HF therapy.

Whereas the significance for the presence of βAR-AAb remains partially understood, studies have shown that a subset of βAR-AAb may exert β-agonist-like activity. Binding of activating-AAb to the second extracellular loop of β₁AR effectively and persistently activates downstream signaling pathways involving cAMP/PKA or p38MAPK [116, 120]. simulating prolonged activation of β₁AR by catecholamines. β₁AR-AAb may mediate apoptosis, pathological remodelling, and β₁AR desensitization and downregulation. In vitro studies have been undertaken to explore action of βAR-AAb. In inflammatory macrophage-like cells, treatment with β₁AR-AAb stimulated TNF-α expression and secretion, effect that was blocked by the selective β₁-antagonist metoprolol and partially by the PKA inhibitor H89 [123]. In cultured fibroblasts, use of β₁AR-AAb obtained from patients with heart disease stimulated cell proliferation [120]. Presence of β₂AR-AAb in patients with HF has also been reported [124]. A higher ratio of β₁AR-AAb/β₂AR-AAb was associated with greater risk of worsening HF implying a potential of cardioprotection of β₂AR-AAb. Similar to β₁AR-AAb, β₂AR-AAb could bind to and activate β₂AR. Cao et al. [124] furthered this finding by showing anti-apoptotic property of β₂AR-AAb in cultured cardiomyocytes challenged with β₁AR-AAb and also in mice treated with the anti-cancer drug doxorubicin.

Limited number of clinical studies have indicated beneficial effect from β-blocking therapy in patients with positive AAb [116]. Recently, β₁AR-AAb-targeted therapies have received increasing interest. Therapies for immune-absorption and neutralizing (e.g. using small peptides) or oligonucleotides of RNA- or DNA-like aptamers, that bind with high affinity to and neutralize AAb from β₁AR-binding and facilitating clearance [116]. These therapies have been tested in preclinical models of cardiomyopathy or human HF patients with results showing effectiveness in removing β₁AR-AAb. Recent studies have shown that β₁AR-AAb-targeted therapies improved LVEF, NYHA (New York Heart Association) HF class or 6MWD (6-min walking distance), and lowed BNP level [116]. However, long-term effects remain to be obtained from large-scale trials. Nonetheless, β₁AR-AAb have the potential to be novel therapeutic target for HF patients who are β₁AR-AAb-positive.

Sympatho-β-adrenergic signaling in immune cells

Cross-talk exists between the SNS and the immune system, forming an important limb of the SNS to integrate body function and maintain homeostasis. Under conditions of inflammation or organ injury, sympathetic nerve firing rates increase leading to release of catecholamines in tissues such as born marrow, the spleen and lymphatic notes, and activation of βAR in immune cells thereby modulating immune responses at a desired extent.

Recent studies have established that the SNS-βAR system regulates bone marrow-dwelled progenitors or immune cells to maintain homeostasis of immune system in physiology and to govern immune responses following pathological stimuli. Subtypes of βAR are expressed in different immune cells derived from hematopoietic cells. β₃AR are restricted to bone marrow stromal cells, whereas β₂ARs are highly expressed in hematopoietic and the stromal progenitor cells [125]. Sympathetic nerve fibers release NE that activates β₃AR of bone marrow niche cells to form upstream hematopoietic stem cells, followed by release of these cells into the blood circulation (Figure 4). With chronic stress, there is increased proliferation of these most primitive hematopoietic progenitors resulting in elevated levels of disease-promoting inflammatory leukocytes (Figure 4) [126], [127], [128]. In response to acute cardiovascular stress, like MI or stroke, numbers of neutrophils and monocytes harboured in the bone marrow increased within 2–3 days due to hyper-proliferation of specific myeloid progenitor cells, event associated with increased circulating counts of monocytes and neutrophils [129, 130]. Nahrendorf and colleagues investigated kinetics of myocardial macrophages that contributes to cardiac remodelling in mice with chronic MI (eight weeks) and HF, and observed monocytosis as well as a 3-fold increase in myocardial macrophages [131]. In addition to resident macrophages, an important source of myocardial macrophages comes from the bone marrow hematopoietic stem and progenitor cells that displayed hyper-proliferation in animals with HF relative to controls. Sager et al. [131] observed that relative to control mice, NE level in bone marrow was 2-fold higher in HF, which was accompanied with enhanced release of these cells into blood circulation, splenic retention and proliferation, and myocardial recruitment. Thus, in mice with HF, higher sympathetic tone to bone marrow facilitate proliferation of bone marrow hematopoietic progenitor cells through activation of β₃AR (Figure 4), which leads to monocytosis and higher tissue density of monocytes like the spleen. These monocytes from the bone marrow and spleen then infiltrate the myocardium contributing to cardiac macrophages pool [131]. This finding bears therapeutic implications given that inhibiting monocyte recruitment would limit ventricular remodeling.

Figure 4:

Sympathetic innervation of bone marrow and immune organs and βAR-mediated regulation of hematopoietic cells and immune cells. Sympathetic nerves, which innervate bones, can penetrate into the bone marrow compartment, where hematopoietic stem cells and progenitor cells (HSC) reside as specific niches. Catecholamines released by sympathetic varicosities activate hematopoiesis by stimulating β₃AR in HSCs thereby promoting bone marrow cell differentiation, proliferation and egression of cells into blood circulation (a). After departing bone marrow, circulating immune cells relocate in immune organs like the spleen or lymph notes (b), where cells proliferate and mature and, upon sympathetic drive, depart immune organs into circulation (c). Circulating immune cells including monocytes (mono), lymphocytes (lym) and neutrophils (neu) then infiltrate the diseased myocardium (d) and promote regional inflammation. Immune cells are equipped mainly with β₂ARs that regulate innate or acquired immunity, lymphocyte homing, immune cell maturation and expression of cytokines/chemokines. mϕ: macrophages; plate: platelet.

β₂AR are the major βAR subtype of bone marrow-derived immune cells and play a vital role in neuronal regulation of inflammatory responses. Many earlier studies, largely in the in vitro setting, reported immunosuppressive action of β₂AR stimulation. Lamkin et al. [132] studied gene patterns in human macrophages, and found that β₂AR stimulation induced a transcriptome typical of anti-inflammatory M2-pattern, together with suppression of pro-inflammatory M1-related genes. However, there is also report for β₂AR signaling synergistically interacts with inflammatory stimuli (i.e. Toll-like receptor activators like lipopolysaccharide or phorbol myristate acetate) leading to activation of immune cells. Studies that carried out in vivo have provided evidence for β₂AR-mediated immune-enhancement [126, 133]. In mice with acute MI, Grisanti et al. [126] demonstrated that β₂AR-mediated activation of immune cells is critical in regulating inflammatory-fibrotic healing. In mice depleted of β₂AR gene and subjected to MI, inflammatory infiltration into the infarcted myocardium was blunted together with retention of immune cells within the spleen, changes associated with a markedly increased mortality (100% vs. 20% in control mice) due largely to cardiac rupture as a consequence of impaired healing. In contrast, mice with bone marrow cell-specific depletion of β₁AR or β₃AR only showed insignificant change in the mortality. In mice with TAC-induced pressure overload, renal sympatho-β₂AR activation stimulated secretion of a cytokine CSF2 (colony stimulating factor 2), which is potent in activating monocytes or macrophages resided in the stressed myocardium [134]. Thus, the SNS-β₂AR involves in the heart–kidney interplay contributing to the adaptation to hemodynamic stress (Figure 4).

The spleen stores a large number of immune cells as well as platelets that could be mobilized upon acute stress or sustained sympathetic activation, contributing to inflammatory response [130, 135]. Following acute MI, circulating platelets undergo dynamic changes in size and activity [106, 136]. In mice subjected to acute MI, we observed release of splenic platelets into blood circulation contributing to the increased platelet size, platelet-monocyte conjugation, and myocardial accumulation of platelets, changes that are highly pro-inflammatory. Such cardio-splenic response can be abolished by the angiotensin-converting enzyme inhibitor perindopril and β-antagonist atenolol [135], implying that elevated levels of Angiotensin-II and NE promote splenic release of immune cells and platelets (Figure 4). Clinical studies have also suggested that conditions associated with sympathetic activation, such as intense exercise or administration of catecholamines, induce splenic release of platelets. Conversely, sympatho-βAR activation leads to shrink in splenic size and increase in circulating level of certain immune cell types. Thus, in the setting of HF, enhanced sympathetic drive might induce splenic release of immune cells and platelets that subsequently promote inflammatory responses in the diseased heart.

βAR signaling in fibroblasts

Fibroblasts account for approximately 20%–30% of cardiac cell population with this number increasing substantially under pathological conditions. Over many years, research investigating the β-adrenergic mechanism in the setting of HF had been focused on cardiomyocytes. It has now been appreciated that events involving fibroblasts and inflammatory cells contribute significantly to the onset and development of HF. Fibroblasts are equipped mainly with β₂AR as well as downstream signaling molecules including AC, PKA, signaling scaffolding proteins like A-kinase anchoring proteins (AKAP) β-arrestins, EPAC and phosphodiesterases (PDE). Using FRET-based sensors, Grisan et al. [137] reported that cAMP signaling in cultured cardiac fibroblasts is mediated predominantly by β₂AR. Whereas some earlier ex vivo studies reported that stimulation of β₂AR suppressed fibroblast proliferation and collagen synthesis, recent studies have revealed pro-fibrotic actions mediated by β₂AR that involve fundamental cellular processes i.e. myofibroblast differentiation, proliferation, migration, and synthesis and release of extracellular matrix proteins [138]. For instance, in cultured cardiac fibroblasts, stimulation of β₂ARs activates DNA synthesis, cell proliferation, expression of pro-inflammatory and fibrotic molecules like interleukin-6 (IL-6) [30, 138], [139], [140], [141], [142]. One possible explanation to these contradictory findings comes from a study by Li et al. [143] using cultured fibroblasts from human cardiac biopsy of control subjects or HF patients, they observed that expression of β-arrestins was elevated that diminished β-agonist-mediated inhibition of fibroblast activity and collagen synthesis. Indeed, in cells with knockdown of β-arrestins, treatment with β-agonists inhibited fibroblast activity and collagen expression [143]. Enhanced β-arrestin signaling mediated a pro-fibrotic phenotype by promoting signaling that involves TGF-β/Smad2/3 and ERK1/2 [143]. Traveras et al. [144] studied mice with cardiac ischemia-reperfusion (IR) or fibroblasts prepared from biopsies of human patients with end-stage HF, and found that Gβγ-GRK2 signaling plays a key role in myocardial fibrogenesis. In addition, several studies have well documented that β₂AR stimulation of fibroblasts promotes expression and release of IL-6 through signaling pathways involving Gsα/REK1/2 [145], or p38MAPK [146]. There is also strong evidence for fibroblast-derived IL-6 in promoting cardiomyocyte hypertrophy [147]. However, the use of fibroblasts in culture, although allowing for dissection of molecular mechanisms in relation to fibrogenesis, is a very simplified model considering the in vitro environment where fibroblasts direct in contact with cardiomyocytes and indirectly influenced by factors released from cardiomyocytes.

A pivotal role of βAR in cardiac fibrosis signaling is suggested by our previous study on mice disrupted of both β₁- and β₂-AR genes (β_1/2-KO). β_1/2-KO and control mice were subjected to TAC to induce chronic pressure-overload for up to 12 weeks. Significant myocardial interstitial fibrosis together with upregulated expression of fibrotic and inflammatory genes, and cardiac hypertrophy, whereas all being evident in control animals, were abolished in β_1/2-KO mice (Figure 5) [148]. Imaeda et al. recently generated mice with myofibroblast-restricted deletion of β₂AR and tested effects of treatment with ISO at a toxic dose (50 mg/kg/day for two weeks). ISO-induced LV hypertrophy, dysfunction and fibrosis, seen in wild-type control mice, were largely prevented in KO mice [31]. The role of fibroblast β-adrenergic signalling in contributing to myocardial hypertrophy was further supported by the observation that mice with fibroblast-restricted overexpression of catalytic subunit of PKA developed cardiac hypertrophy [31]. Thus, these findings suggest that βARs not only in cardiomyocytes, but also in fibroblasts, play a central role in mediating cardiac hypertrophy and fibrosis in diseased settings.

Figure 5:

Cardiac phenotype of mice with dual deletion of β₁AR and β₂AR (β_1/2-KO) under conditions of chronic pressure overload. Following a 12-week period of pressure overload, wild-type (WT) control mice displayed significant cardiac interstitial fibrosis (A), hypertrophy (B) and signs of HF (e.g. pulmonary congestion, LV dilatation and dysfunction, atrial thrombus formation). All these changes, together with upregulation of pro-fibrotic and inflammatory genes in the stressed myocardium, were abolished in dual β_1/2-KO mice. These findings underscore an essential role of βAR signaling in mediating fibrosis and hypertrophy induced by pressure overload [148]. SH, sham-operation; TAC, transverse aorta constriction. Figure adopted from Kiriazis et al. with permission [148].

Galectin-3 belongs to the lectin family and exerts pro-fibrotic and pro-inflammatory actions through binding to numerous intracellular and membrane glycoproteins. Currently, galectin-3 is regarded as a useful biomarker for HF as well as a disease mediator. We explored whether βAR activation might lead to elevation of circulating galectin-3. Several murine models were studied including IR, ISO cardiotoxicity and β₂AR-TG [64, 66]. In these models, circulating galectin-3 level was elevated in proportion to the increased galectin-3 content in the myocardium, and the cardiac source of circulating galectin-3 was further confirmed by the presence of trans-cardiac galectin-3 gradient [149], indicating spillover of cardiac galectin-3 into blood circulation [149]. Findings from these models differ from that in Mst1-TG mice with DCM [108], in which cardiac content of galectin-3 increased by 50-folds, the highest among the models studied (7–18 folds), and yet circulating galectin-3 level was unchanged [149]. Interestingly, ISO treatment in Mst1-TG mice for 2 days markedly increased circulating levels of galectin-3, indicating that βAR activation facilitates cardiac galectin-3 release into blood circulation. Importantly, βAR-mediated increase in both circulating and cardiac galectin-3 levels was reversed by the use of β-antagonists [31]. Significance of βAR-stimulated galectin-3 expression has been implicated by preclinical studies demonstrating that inhibition of galectin-3 by genetic or pharmacological means suppressed ISO-induced cardiac fibrosis and inflammation [114]. In a recent study, galectin-3 KO and wild-type control mice were subjected to chronic ISO administration. In control animals, ISO-induced increase in cardiac galectin-3 content (by 8 folds), LV hypertrophy, fibrosis and suppressed LV function, together with upregulation of a number of fibrotic and inflammatory genes [31]. All these phenotypes at molecular and global levels were abolished in galectin-3 KO mice, except for a similar degree of LV hypertrophy [31]. These findings call for clinical research to explore the influence of β-blockers on dynamic changes in circulating level of galectin-3 in relation to the overall therapeutic efficacy by β-blocking therapy.

Ample evidence has been accumulated for the interactions between cardiomyocytes and fibroblasts that are critical in fibrogenesis and pathological consequences [16]. We recently studied the role of βAR in mediating such hetero-cellular interactions leading to fibrogenesis via galectin-3, K_Ca3.1 (calcium-activated K⁺-channel), and junctional protein connexin-43 (Cx43) (Figure 6) [150, 151]. Specifically, stimulation of cardiomyocyte βAR enhances expression of galectin-3, which induces K_Ca3.1 expression in non-cardiomyocyte cells. In addition to further inducing galectin-3 expression, K_Ca3.1 enhances Ca²⁺ influx via transient receptor potential (TRP) channels leading to activation of inflammatory cells and fibroblasts (Figure 6A) [150]. In mouse models of ISO-cardiotoxicity or β₂-TG, treatment with a K_Ca3.1 inhibitor suppressed inflammation and fibrosis [150]. Cx are membrane proteins that are expressed in cardiomyocytes and fibroblasts and form hemichannels and gap junctions between cells. Zhang et al. [142] recently reported cell-type distinct regulation by βAR of expression and localization of Cx43. Whilst β₂AR stimulation upregulates expression of Cx43 in fibroblasts together with enhanced fibroblast activity, βAR activation in cardiomyocyte downregulates Cx43 expression together with redistribution of Cx43 from interdisc to laterals (Figure 6B). Moreover, βAR stimulation leads to formation of inflammasomes that cleavage inactive pre-IL-18 to active IL-18, which further augments Cx43 expression in fibroblasts (Figure 6B) [142]. These changes are expected to increase the probability of formation of Gap-junctions between heterogeneous cells (i.e. cardiomyocytes and fibroblasts), changes in favour of electrophysiological disturbance and arrhythmogenesis [16, 152]. Gap-junction mediated coupling of adjacent fibroblasts has also been shown to promote fibroblast differentiation, proliferation and migration, as well as facilitate inflammatory-fibrotic process [153, 154].

Figure 6:

Schematic diagrams depicting βAR-mediated heterogeneous cell interactions promoting cardiac fibrosis. (A) Stimulation of βAR in cardiomyocytes induces expression and release of galectin-3 (Gal-3), which in turn activates expression of intermediate-conductance Ca²⁺-activated K⁺ (K_Ca3.1) channels in fibroblasts or immune cells. Activity of K_Ca3.1 channels leads to membrane hyperpolarization and subsequent Ca²⁺ influx via transient receptor potential (TRP) cation channels by which fibroblasts or inflammatory cells are activated [150]. (B) βAR activation in cardiomyocytes suppressed Cx43 expression and shifted Cx43 localization from interdisc to lateral sides of cells. Cx43 expression in fibroblasts is upregulated via direct activation of β₂AR/PKA signaling cascade [142]. Besides, interleukin-18 (IL-18) released via βAR-mediated formation of inflammasomes in cardiomyocytes further upregulates Cx43 expression in fibroblasts [142]. These changes would favour the probability of cardiomyocyte-fibroblast coupling via Gap-junctions with consequent electrophysiological instability [16, 152, 154], or activation of coupled fibroblasts [153].

Potential role of cardiac α₁AR in HF

It is estimated that α₁ARs, mainly α_1AAR and α_1BAR subtypes, account for approximately 10% of total AR in the heart [155]. However, relative to βAR, cardiac α₁AR has been much less investigated. Studies using a diversity of approaches have shown that activation of α_1BAR in the heart subjected to disease insults leads to adverse consequences [155], [156], [157]. Under conditions of chronic pressure overload, we observed in α_1B-TG mice more severe hypertrophy, fibrosis and HF relative to non-TG littermates, indicating adverse effect of α_1BAR in this pathological setting [158]. In contrast, accumulating evidence indicates that stimulation of α_1AAR holds promise to be developed for treatment of HF. Mice with cardiomyocyte specific knockout of both α_1AAR and α_1BAR, showed poor tolerance to TAC challenge with significant apoptosis, LV dilatation and premature deaths due to HF [159]. Further investigation indicated that it was loss of α_1AAR, but not α_1BAR, that was responsible for cardiac maladaptation. Huang et al. [160] studied cardiomyocytes with selective depletion of α_1A-AR, and observed an increased cell death due to apoptosis or necrosis induced by stimuli of NE, doxorubicin or H₂O₂.

Several studies have addressed the downstream signaling mechanisms responsible for the cardioprotective action of α_1AAR. Whilst treatment with α_1AAR agonist (A61603) provides cardiac protection in wile-type mice subjected to TAC, such effect of α_1A-agonist was lost in mice with cardiac expression of inactive mutant α_1AAR that prevents Gq-protein coupling [161]. Cardiac protection conferred by α_1AAR is mediated by diverse Gq-dependent signaling that involves PKCδ/glucose transporter-1/4 to promote metabolism [162], Gq/ERK to antagonize apoptosis [161], or pro-angiogenesis occurring post-MI [163]. Cardioprotection by enhanced α_1AAR signaling is further supported by studies using a mouse model with cardiac overexpression of α_1AARs (α_1A-TG). α_1A-TG mice displayed cardiac hyper-contractile function without adverse myocardial hypertrophy or fibrosis [164]. Using α_1A-TG and non-TG mice subjected to chronic MI or pressure overload (TAC), studies have shown that genetic α_1A-AR activation preserved LV function and reduced HF death under both disease settings [156, 163, 165]. Similar cardioprotection by treatment with a selective α_1A-agonist was observed in animals treated with anti-cancer drug doxorubicin [166, 167]. In cardiac biopsy samples from patients without or with cardiomyopathy and HF, all three subtypes of α₁AR were expressed at mRNA level [168, 169], only α_1AAR and α_1BARs were detectable at the protein level (60% α_1BARs) which was maintained in the failing myocardium, implying that α₁ARs do not undergo downregulation [168, 169]. We also showed, in mice with chronic pressure-overload that expression of α_1AAR in hypertrophic myocardium was unchanged [158]. These findings are in keeping with another report showing that in myocardial samples from HF patients, α_1AAR mediated positive inotropy remains intact [170]. The mechanism of the inotropic action, seen in α_1A-TG mice or use of selective agonist, is only partially understood but involves increased Ca²⁺ sensitivity of contractile apparatus following phosphorylation of contractile regulatory proteins by RhoA/ROCK kinases, and elevated intracellular Ca²⁺ signaling [171]. The increased Ca²⁺ entry by α_1AAR activation is achieved through transient receptor potential channels (TRPC6) that relocate from plasma to sarcolemma upon activation of α_1AAR/Gq/PLCβ/diacylglycerol signaling [172].

Changes in presynaptic regulation of neurotransmission by α₂AR

Sympathetic neuronal release of NE is controlled by diverse presynaptic receptors that mediate either stimulatory (e.g. β₂AR, angiotensin AT₁R) or inhibitory regulation (e.g. α₂AR, dopaminergic DA2, muscarinic M2, purinergic A1, NPY-receptor). Of them, α₂AR mediates the most potent suppression of NE exocytosis, by which α₂AR determines the extent of stimulation of postsynaptic βAR or αAR [173, 174]. Among three subtypes (α_2A, α_2B and α_2C), α_2AAR and α_2CAR are functionally important. In mice deficient of α_2AAR (α_2A-KO), loss of presynaptic suppression of NE release results in higher baseline heart rate (by 47%), lower NE content (by 57%), and lower βAR density (by 26%) [175], changes similar to that seen in the human failing heart. Under chronic pressure overload by TAC, Hein’s group documented, in α_2A-KO and α_2C-KO mice, increased mortality due to HF, more severe LV hypertrophy and fibrosis, poorer LV function, and elevated levels of circulating catecholamines [176]. In human studies, facilitatory regulation of NE release by presynaptic β₂AR is not evident, but inhibitory regulation by α_2CAR is potent [177, 178]. The α_2CAR deletion variant with loss of four amino acids in the third intracellular domain 322–325, is commonly in the population [179, 180]. Such loss-of-function polymorphism significantly impairs the signal transduction of α_2CAR and presynaptic inhibition of NE release. Importantly, Brede et al. [176] showed, for the first time, that in chronic HF patients who had the α_2CAR deletion variant exhibited more severe HF relative to patients with intact α_2CAR. However, down-regulation of presynaptic α_2CAR in cardiac sympathetic nerves and the adrenal gland has been noticed in HF patients [88, 177], which might explain the lack of significant change in cardiac NE kinetics in patients with the α_2CAR deletion variant [178].

Several studies have implicated the prognostic significance of α_2CAR polymorphism in patients with HF. Small et al. [180] reported that the α_2CAR deletion variant is more common in blacks than whites (61% vs. 41%), which might contribute to a poorer prognosis in blacks than whites with HF. However, in a study on 345 patients with DCM and a mean follow-up period of 4.9 years, patients carrying the α_2CAR deletion variant had lower risk for total adverse event rates (−83%) or death rate (−87%) [179]. This seemingly contradictory finding might be due to the common use in this patient cohort of β₁-antagonist, that spares potentially protective α₁AR or β₂AR. Potential drug intervention of presynaptic receptors has recently been explored. Using the rat model of MI, treatment with CHF-1024 as an agonist for α₂AR and DA2, dose-dependently reduced cardiac release of NE and expression of TNFα, effects accompanied by alleviated myocardial damage and improved LV function [181], indicating cardioprotection via blunted sympathetic surge post-MI. In a randomised clinical trial on 1,000 patients with severe HF, treatment with nolomirole, an agonist of α₂AR and DA2, had no benefits in terms of HF-hospitalization or survival, nor to alter heart rate and blood pressure [182]. In another study on HF patients and controls, significance of the α_2CAR deletion variant and β₁AR within Gs-binding domain high-function-variant (Arg398) relative wild-type (Gly398) was investigated [180]. In black subjects, risk of HF increased by 5.6-fold with the α_2CAR only, but by 10.1-fold in subjects carrying both α_2CAR deletion variant and β₁AR high-function-variant [180]. In a clinical trial on 1,040 HF patients testing the β-clocker bucindolol, which is also a sympatholytic agent, Aleong et al. [183] explored the impact of the α_2CAR deletion variant on effect of bucindolol, and found that α_2CAR wild-type patients respond to this drug with a pronounced sympatholytic effect, lower incidence of malignant ventricular tachyarrhythmias, and a better survival. In contrast, such effects of bucindolol were attenuated or completely lost in patients carrying the α_2CAR deletion variant. Collectively, the α_2CAR deletion variant augments NE release, which is potentially detrimental to the diseased and failing heart. However, such influence is also subjected to modulation by β₁AR high- or low-function variants.

Remodeling of cardiac sympathetic nerves

In addition to the well-documented cardiac sympathetic overdrive and enhanced NE release, studies have revealed morphological and functional abnormalities of sympathetic nerves in the failing heart, likely as a combined consequence of sustained enhancement of sympathetic neuronal tone and altered retrograde signals from the myocardium [8, 10]. A range of abnormalities of cardiac sympathetic nerves have been identified in the failing heart (Figure 7). First, numerous studies have reported lower sympathetic nerve density in the failing myocardium, suggesting loss of sympathetic neurons or axons [5, 42, 184], albeit there are reports of increased nerve density in hypertrophic or failing myocardium.

Figure 7:

Remodeling of cardiac sympathetic nerves in the failing myocardium. Most likely due to chronic sympathetic hyperactivity, cardiac sympathetic nerves exhibit functional and phenotypic changes. Changes in sympathetic neurons consist of reduced expression of tyrosine hydroxylase (TH, a limiting enzyme of NE synthesis) and NE transporter (NET), and yet expression of choline acetyltransferase (ChAT) is induced in certain neurons. These changes, defined as neurotransmitter switching, are characteristic of neuronal “rejuvenation”. The mechanism involves alterations of retrograde signaling of cardiomyocytes on neurons, notably upregulation of interleukin-6 (IL-6) family cytokines (e.g. leukaemia inhibitory factor, LIF) and down-regulation of nerve growth factor (NGF). LIF and NGF bind to respective receptors on the presynaptic membrane, undergo reverse-transportation to reach neuronal bodies, where they induce phenotypic changes. At varicosities, there are overt reduction in NE reuptake but increased NE breakdown by monoamine oxidase (MAO), which, together with excessive release and reduced biosynthesis, result in partial NE depletion. ACh: acetylcholine; DOPEG: 3,4-dihydroxyphenylethyl glycol; LIF-R: LIF receptor; TrkA: NGF-activated tyrosine kinase receptor; VMAT: vesicular monoamine transporter.

Second, earlier studies showed a partial NE depletion in the failing myocardium, which is attributable to long-term excessive NE release together with attenuation of both NE biosynthesis and NE reuptake (Figure 7) [10, 42, 185]. Backs et al. [186] observed in the rat with TAC-induced HF and significant cardiac remodelling, that whereas neuronal NE transporter (NET) mRNA level was unchanged in neurons localized at the left stellate ganglion, there was a 40% reduction in NET binding sites, change independent of nerve density. In varicosities, NE is transported through vesicular monoamine transporter (VMAT) back into vesicles for storage, or alternatively metabolised by monoamine oxidase (MAO) to form dihydroxyphenyl glycol (DOPEG) that can diffuse freely through sarcolemma membrane into circulation (Figure 7). Patients with HF showed greater cardiac spillover of DOPEG [8, 42], suggesting increased NE breakdown. It remains unknown whether there is concomitant attenuation in the activity of VMAT in the failing heart that would exacerbate NE metabolism and depletion.

Third, sympathetic nerves undergo neurotransmitter-switching with phenotype changes (Figure 7) [187]. Expression of interleukin (IL)-6 cytokines is elevated in the failing myocardium or in cardiac cells upon βAR stimulation [141, 188, 189]. Leukemia inhibitory factor (LIF) is a member of IL-6 cytokine family and could regulate sympathetic neuronal plasticity through binding to its dual-molecule receptor LIF-receptor and gp130. Earlier studies in cultured sympathetic neurons revealed that LIF induced switch of neurotransmitters from NE to acetylcholine (ACh). Using the failing heart of salt-sensitive Dahl rats, Kanazawa et al. [187] reported an increased abundance of cholinergic neurons, including peripheral axons as well as neuronal bodies, that express markers like choline acetyl-transferase (ChAT) and vesicular ACh transporter (VAChT). This finding was further confirmed by another study on samples of LV and stellate ganglia from autopsied HF patients showing neuronal marker switching from tyrosine hydroxylase (TH⁺, i.e. noradrenergic) to choline-acetyltransferase (ChAT⁺, i.e. cholinergic) [187]. Whereas the significance of such cholinergic trans-differentiation of cardiac sympathetic nerves remains to be fully illustrated, this likely represents an adaptive change to protect against cardiotoxicity by excessive sympathetic drive, akin to that of β₁AR-downregulation and desensitization or alternatively by the use of β-blockers. Together with reasons like excessive NE release, reduced NE synthesis and faulty NE reuptake [186, 190], such neurotransmitter switching adds to an alternative mechanism responsible for the partial depletion of NE content in the failing heart (Figure 7). Changes in neurotransmitter identity is a common feature during neuronal development. Kimura et al. [185] proposed that neurotransmitter switching as well as other loss-of-function changes (e.g. reduction in NE uptake, lower NE content, suppressed NE synthesis) are “sympathetic rejuvenation”, which is characterized by expression of markers of immature neurons and likely driven by LIF and cardiotrophin-1. Thus, the augmented expression of IL-6 family cytokines in the failing heart or under β-adrenergic stimulation [30, 139, 189, 191], induces functional loss of sympathetic nerves together with neurotransmitter plasticity, that might impact disease compensation and development. Recent studies have shown that after MI, cardiac sympathetic nerves undergo transient neurotransmitter switching, evidenced by expression of ChAT in adrenergic neurons and, upon stimulation, simultaneous release of NE and ACh within the first two weeks post-MI [192, 193]. There is evidence for ACh co-released with NE counteracting the pro-arrhythmic electrophysiological property induced by NE, such as heterogeneity of action potential duration crossing the infarcted heart [192, 193]. By testing effects of different Ca²⁺-channel blockers in HF patients, a recent clinical study revealed switching of neuronal Ca²⁺ channels from N-type (restricted to peripheral nerves) to L-type (central nerves) [194].

One important mechanism for the abnormal cardiac sympathetic nerves is altered expression in the failing myocardium of neurotrophic factors, most notably nerve growth factor (NGF). Sympathetic neuronal function and innervation density are regulated by target organ-derived neurotrophic factors, which mediate retrograde signals essential for the development and differentiation of neurons and acquisition of neuronal properties [185]. These neurotrophic factors bind to respective receptors localized at presynaptic membrane (e.g. tyrosine kinase receptor, TrKA, for NGF), and then undergo retrograde axonal transportation to reach the neuronal body where they regulate plasticity, function and phenotype of sympathetic neurons (Figure 7). We studied cross-cardiac gradient of NGF concentration by sampling arterial and coronary sinus blood from patients with HF due to DCM or ischemic cardiomyopathy (average LVEF 17%, NYHA class 3.1), as well as NGF expression in heart tissues from rats with chronic MI-induced HF or in cultured cardiomyocytes stimulated with NE [195]. In HF patients, blood concentration of NGF and trans-cardiac NGF gradient were dramatically lower relative to healthy controls, implying reduced NGF production in the failing myocardium. This was confirmed by reduction in NGF expression at mRNA (−40%) and protein (−24%) levels in rat failing hearts. Furthermore, in cultured cardiomyocytes, expression of NGF was inhibited by treatment with NE [195]. Similar findings have been reported by subsequent studies on different animal models of HF or NE stimulation. Kimura et al. [196] reported that in rats received chronic NE infusion, animals showed cardiac downregulation of NGF expression and loss of sympathetic nerve fibers. In dogs with an 8-week period of rapid-pacing and HF, expression level of NGF in the myocardium was lower at mRNA and protein levels, together with downregulated expression by neurons of NGF-receptor TrKA [197]. These changes were in parallel with loss of sympathetic nerve densities and reduced activity of TH, the rate-limiting enzyme for NE biosynthesis [197]. In rats with TAC and HF, animals showed reduction in both NGF expression and NE reuptake by NET, and a partial NE depletion [198]. Importantly, there was a report showing that in rats with HF, NGF injection into the left stellate ganglion restored NET expression and NE content, changes accompanied by improvement of LVEF [198]. This finding is in keeping with our previous study in mice with cardiac overexpression of NGF, in which we observed cardiac hyper-innervation, higher tissue NE content and enhanced NE reuptake, but with preservation of βAR-response and LV geometry and function, likely due to concomitant enhancement of NE reuptake [199]. However, there is discrepancy regarding NGF expression levels in hypertrophic and failing hearts. In myocardial tissues from human subjects with ischemic heart disease, there was 2.5-fold increase in NGF level that was associated with increased sympathetic axon diameters. Similar morphological changes were also observed in spontaneously hypertensive rats or dogs with ischemic cardiomyopathy [200]. The reason(s) for such discrepancy regarding NGF level in diseased myocardium remain unclear, but influences such as aetiologies, diversity of cardiac sampling regions or disease stage, cannot be excluded. Indeed, NGF expression is known to be differentially regulated by factors relevant to heart disease, for instance, calcineurin-NFAT downregulates whilst endothelin-1 (ET-1) upregulated NGF expression [185]. Collectively, these data indicate a dynamic and constant NGF expression in the heart and that dysregulation of neurotrophic action of NGF in the heart contribute to neuronal dysfunction irrespective of reduced or increased sympathetic densities in the hypertrophic or failing heart.

NPY is a neuropeptide of 36 amino acids, present in central nervous tissue and peripheral sympathetic nerves and, upon sympathetic activation, co-released with NE. Whilst earlier studies observed higher cardiac and circulating levels of NPY under conditions of cardiovascular stress (hypertension, MI, hypertrophy or HF) [201], a few recent studies have enriched our understanding on the role of NPY. In rats with acute MI, Huang et al. [202] observed a 3–4-fold increase in cardiac and plasma levels of NPY, albeit the expression level of NPY at the sympathetic ganglion was not determined. Interestingly, they showed that NPY-KO rats subjected to MI had smaller infarct size and lower level of cardiomyocyte apoptosis, evidence of cardioprotion [202]. In a recent study on patients with acute MI and by blood sampling from the coronary sinus, Kalla et al. [203] showed that in patients exhibited sustained VT or VF, NPY co-released with NE from the infarcted heart is pro-arrhythmic, even in the presence of β-antagonists. Kalla et al. [203] furthered this finding by studying perfused rat hearts subjected to the sympathetic stellate ganglion stimulation, which increased VF vulnerability even in the presence of the β₁-antagonist metoprolol. Elevated VF vulnerability was prevented with the use of Y1 receptor antagonist (BIBO3304) but induced by NPY infusion. Over several decades, the role of NPY in chronic HF remains unclear. Olujini et al. recently investigated the implication of coronary sinus NPY levels in relation to clinical outcomes of chronic HF patients [204]. In 105 patients with HF (mean LVEF 26%), higher NPY levels in coronary sinus blood was associated with overall adverse cardiac events (death, HF hospitalization, requirement of heart transplantation or placement of cardiac assistant device, adjusted HR 9.3, P < 0.001) [204]. Furthermore, Ajijola et al. [204] observed in stellate ganglion biopsy from chronic HF patients, lower NPY protein content but unchanged NPY mRNA level, indicating increase NPY release. Medzikovic et al. [205] observed that a nuclear receptor Nur77 was potent in inducing NPY expression, and they further showed that Nur77 gene deletion (Nur77-KO) or treatment with Y1-antagonist (BIBO3304) were beneficial in a model of drug-induced cardiomyopathy.

Neuromodulation of the sympathetic nervous system by device-based interventions

It is well established that central imbalance of the autonomic nervous system is a key component of HF pathophysiology [4, 206]. In this regard, sustained activation of the SNS together with suppression of the parasympathetic nervous system characterize clinical features of patients with HF. Advanced from the previous viewpoint of peripheral SNS abnormalities in HF [7], increasing number of studies have suggested central nervous abnormalities in HF subjects [4, 207]. Using several independent methodologies for clinical assessment of SNS activity, studies have revealed augmented sympathetic efferent firing rates, increased circulating level of catecholamines, and blunted baroreflex sensitivity or heart rate variability [11, 12]. Thus, HF could be a clinical condition in which abnormalities in the SNS is evident not only peripherally, but also centrally. The autonomic nervous activity is centrally regulated by processing numerous afferent inputs that are both inhibitory or excitatory in nature. In this regard, the pioneer studies by Esler and colleagues on renal sympathetic denervation (RSD) for treatment of resistant hypertension have re-directed research attention into the area of neuromodulation by device-based interventions to achieve centrally resetting of the SNS activity in HF (Figure 8).

Figure 8:

Neuromodulation with suppression of cardiovascular sympathetic nervous activity by device-based interventions. Increasing number of preclinical studies have shown that SNS activity could be suppressed through non-pharmacological device-based procedures, including thoracic spinal cord stimulation (TSCS), vagus nerve stimulation (VNS), carotid body denervation (CBD) or renal sympathetic denervation (RSD). Currently preclinical studies indicate that these interventions similarly induce central resetting, reduce cardiac sympathetic outflow and plasma NE levels, leading to cardioprotective actions. However, clinical studies on HF patients thus far could only replicate parts of these findings from large animal models with reasons to be explored.

Studies have shown that neuromodulation is achievable through device-based interventions or pharmacology means (Figure 8). Potential non-pharmacological therapeutic approaches include ganglionated plexus ablation, thoracic spinal cord stimulation (TSCS), carotid body denervation (CBD), renal sympathetic denervation (RSD) and cervical vagal nerve stimulation (VNS).

Less explored areas of cardiac βAR actions

Mitochondrial dysfunction and metabolic remodeling are regarded as the common mechanism leading to the development and worsening of HF. Whether sympatho-βAR signaling affects cardiac mitochondrial function remains much less investigated. With the recent demonstration of βAR signaling via Hippo pathway [114], we are investigating myocardial mitochondrial function and metabolism. Using two cardiomyopathy mouse models due to transgenic activation of βAR (β₂-TG) or Mst1 (Mst1-TG) as upper-stream kinase of Hippo signaling, our studies thus far have demonstrated significant abnormalities in cardiotranscriptome with over 40% genes dysregulated in both models. Importantly, bioinformatic analyses of RNAseq data revealed that similar in both models, mitochondrion-related gene-sets are among those most significantly downregulated, and in contrary, pro-fibrotic gene-sets are among the most significantly upregulated gene-sets (Figure 9, and our unpublished finding) [107]. The similarity in both models of these changes in transcriptome strongly indicates a pivotal role of the βAR-Hippo signaling pathway in mediating mitochondrial abnormalities seen in the failing heart. Other signaling mechanisms also likely contribute to metabolic abnormalities, including GRK2 mediated insulin-resistance [229], and enhanced ROS generation upon βAR overactivation. Currently there has been lack of effective drugs that exert mitochondria protective action. Progress in this filed would widen our understanding on adverse β-adrenergic signaling in the heart and allow for development of novel therapeutic intervention targeting mitochondrial dysfunction and abnormal energy metabolism (Figure 9). Another intriguing area is pro-angiogenesis action mediated by β₂AR, as reported by numerous studies on cancer tissues [230]. The similar angiogenesis due to β₂AR activation was also observed in the heart [231]. Considering the significance of angiogenesis in the heart, it is worthwhile to explore the potential role of such apparently compensatory action mediated by β₂AR in HF.

Figure 9:

Genetic activation of either βAR or Hippo signaling pathway induces similar alterations in cardiotranscriptome by RNA sequencing. Data were from adult transgenic (TG) mouse models with cardiac overexpression of β₂AR (β₂-TG) or mammalian sterile-20-like kinase 1 (Mst1-TG). In both models, RNA sequencing of the left ventricular myocardium revealed profound downregulation of numerous gene-sets of mitochondrial dynamics or metabolism, together with upregulation of fibrotic gene-sets. Data are based on published study [107] and our unpublished findings. BCAA: branched chain amino acids; ECM: extracellular matrix; NADH: nicotinamide adenine dinucleotide.

Unmet needs in drug development targeting β-adrenergic signaling

There are apparent limitations of current β-blocking therapy, given that the sympatho-βAR activation bears compensatory significance. Certain cohorts of HF patients show β-blocker intolerance for reasons including worsening HF symptoms, bradycardia, orthostatic hypotension, fatigue or disease-like asthma. In addition, clinical studies have revealed that in HF patients with atrial fibrillation, β-blocker therapy failed in achieving benefits assessed by reduction in risk of cardiovascular events like HF re-hospitalization and survival [232]. Anaemia is one of commonly comorbidities of HF and a recent J-CHF trail revealed that in patients with HFrEF, the therapeutic effects of carvedilol, estimated by improvement of LVEF and normalization of elevated BNP, were blunted in patients complicated with anaemia [233]. Furthermore, β-blocking therapy may have limited effect in patients who are positive in β₁AR-AAb [116, 118]. Although reasons remain unclear for the blunted efficacy of β-blockers in these clinical scenarios, a very likely explanation is the non-selective blockade of βAR signaling by current β-blockers. It would be desirable to develop drugs for specific inhibition of adverse signaling pathways sparing the compensatory signaling. This requires further in-depth understanding on β-adrenergic signaling networks, and identify key and drugable signal molecules. An alternative approach is the development of biased ligands. The biased ligand refers to a specific ligand that could stimulate a fraction of signal molecules of a given receptor thereby activating only desired downstream effects while avoiding potential adverse effects. This concept has stimulated research on development of biased ligands of GPCRs. Carvedilol is classically defined as a non-selective β-antagonist and widely used for HF therapy. This drug has been shown to act as a biased β₂AR ligand stimulating β-arrestin-dependent, but not G-protein-dependent signaling, such as activation of ERK signaling [234]. A recent study showed that use of carvedilol with such selective signaling is useful for treatment of skeletal muscle wasting, which is commonly seen in patients with severe HF [235]. Cardioprotection via β₂AR-β-arrestin biased signaling has also been achieved both in vitro and in vivo with the use of a small peptide (ICL1-9) [39, 236]. Development of ligands that selectively activate β-arrectin2 and/or suppress β-arrestin1 also hold promise for future HF therapy by promoting cardiac beneficial while limiting detrimental signaling.

βAR signaling in hematopoietic/immune cells and fibroblasts

Last two decades have witnessed major progress in our understanding regulation by SNS-βAR of inflammatory cells (hematopoietic cells, immune cells) and fibroblasts, cell types critical in fibrogenesis. By regulating tissue damage as well as healing processes, inflammation plays a key role in heart disease and HF. However, clinical trials of drugs targeting inflammatory molecules for treatment of heart disease have yielded disappointing outcomes and new strategy is yet to be established. With increasing body of knowledge in this field, it is likely to establish new therapeutic strategy for HF by using selective βAR agonists and antagonists to regulate inflammatory homeostasis. To achieve precise fine-turning of immune responses without unwanted adverse effect, such pharmacological regulation of the βAR-immune axis would be aetiology-dependent, and key factors to be considered include the timing and duration of therapeutic targeting of βARs in immune cell. Furthermore, there is strong evidence for interactions, likely through the paracrine mechanism, between heterogeneous cells, i.e. cardiomyocytes, fibroblasts, endothelial cells or inflammatory cells, contributing to myocardial remodeling in HF. The role of βAR in mediating communication between different cell types deserves further investigation.

Role of sympatho-adrenergic mechanism in subtypes of HF

Current understanding on changes in the SNS in HF has largely come from research on patients with reduced LVEF (HFrEF). It has now been well recognized that approximately half of all patients with HF have preserved LVEF (HFpEF) vs. the rest exhibiting moderately-reduced or reduced LVEF (HFmrEF, HFrEF). However, studies on the SNS in HFpEF are sparse. It is critical to conduct comparative research to explore the sympathetic nervous activity on patients or animals with HFpEF vs. HFrEF, given that establishment of effective therapy for HFpEF would be difficult without better understanding of HFpEF pathophysiology, including the sympathy-adrenergic mechanism and the use of β-blocking therapy. In this regards, two very recent studies on patients with chronic HF have revealed that circulating NE level was higher in patients with HFrEF than those with HFpEF, and that patients with HFmrEF had NE level comparable to HFpEF subgroup [237, 238]. Interestingly, elevated NE level remains to be the independent predictor for long-term prognosis in all subtypes of HF patients albeit the predictive power is the strongest in HFrEF vs. other subtypes.

Better understanding sympathetic nerve remodelling in HF

Driven by sustained afferent reflexes, cardiac sympathetic efferent activity is enhanced in the setting of HF. This, together with alterations in a group of neuroregulatory molecules like NGF and LIF, result in profound remodelling of sympathetic nerves and ganglion-localized sympathetic neurons. Whereas current information remains somewhat fragmented, the features of such remodelling likely depend on the stage and aetiology of HF, forming a vital part of HF pathophysiology. Progress in this area would be expected to shade light on neuronal mechanisms in development of HF and arrhythmogenesis.

Neuromodulation for centrally mediated suppression of sympathetic activity

The management of symptoms and the strategy for improving prognosis of HF patients have largely been based on pharmacological treatments. Modern cardiology leads the progress of device-based therapies for conditions that traditionally could only be treated with medications. To achieve autonomic neuromodulation for treatment of HF, centrally acting agents (e.g. imidazole compounds) or device-based interventions hold promise to be further established as effective neuromodulatory therapies for HF (Figure 8). In-depth investigation, including clinical trial, is urgently needed on how device-based neuronal interventions, such as RSD, reduces the severity of HF and arrhythmic burden. It is anticipated that this filed would add new dimension to future HF therapy.