Adrenergic Receptor Signaling in Heart Failure

Role of Increased Adrenergic Drive in the Natural History of Heart Failure

In patients with CHF, the adrenergic nervous system is critical to the support of myocardial function. However, chronic activation contributes to progressive myocardial dys function and left ventricular pathological remodeling. Regardless of etiology, a decrease in myocardial function results in activation of afferent signaling from baroreceptors and chemoreceptors to the central nervous system, such that efferent cardiac adrenergic nerves release more NE, activating β-ARs, increasing heart rate and, depending on the degree of functional reserve, increasing myocardial contractility. The increases in heart rate and contractility favorably affect both systolic and diastolic function, thereby increasing cardiac output. As heart failure worsens, adrenergic drive continues to increase in an attempt to compensate for the progressive loss of cardiac function ( Fig. 6.1 ). However, long-term exposure to high concentrations of NE has marked adverse effects on myocardial and cardiac myocyte biology ( Fig. 6.2 ). Conversely, inhibitors of adrenergic signaling that have demonstrated favorable effects on heart failure natural history include β-blocking agents and inhibitors of the renin–angiotensin–aldosterone system, which indirectly lower adrenergic drive. There is extensive clinical trials literature supporting the use of these classes of agents as well as definitive guidelines (American Heart Association/American College of Cardiology [AHA/ACC]) for their optimal administration (see Chapter 37 ).

Although the “cardiotoxic” effects of catecholamines, in particular NE, have been recognized for more than a century, we have only recently begun to understand the detailed molecular mechanisms by which sustained increases in cardiac adrenergic drive adversely affect myocardial biology and structure/function phenotype. Treatment of isolated cardiac myocytes with concentrations of NE equivalent to those in the failing human heart cause dramatic changes in cell morphology and, within a matter of days, up to a 60% loss of myocyte viability. This effect is mimicked by exposing myocytes to the nonselective β-agonist isoproterenol, an effect that is inhibited pharmacologically by the β-blocker, propranolol, suggesting that the acute toxic effects of NE are mediated almost exclusively through β-ARs rather than through α-ARs. As described below, signaling via the β ₁ -AR is generally considered to be more biologically harmful than via the β ₂ -AR. This finding is supported by clinical trial data, where differences in β ₁ -AR blocking doses and enrollment criteria are considered ; further, β ₁ -AR selective blockade appears equally efficacious as treatment with nonselective β-blockers, meaning that β ₁ -AR signaling is the key general mechanism mediating reversible pathobiologic effects in the failing heart.

As stated above, elevated plasma concentrations of NE reflecting overflow from adrenergically activated organs including the heart are a well-recognized biomarker of CHF. However, as adrenergic activation persists, cardiac NE stores become depleted. Ultimately, other tissue sources of catecholamines (i.e., the adrenal medulla) are stimulated, leading to generalized adrenergic activation and spillover of cardiac derived NE into the systemic circulation. As systemic or cardiac NE levels increase, so does mortality in patients with heart failure. These observations have been validated further by evidence from numerous clinical trials of β-blocker therapy, demonstrating reverse remodeling (improved systolic function and decreased left ventricular volumes, typically measured as the ejection fraction) and improved survival in patients with heart failure who receive β-blocker therapy, long term.

Interestingly, in marked contrast to adults with HFrEF, β-blocker therapy does not appear to be particularly efficacious in pediatric heart failure patients. Although the findings are somewhat anecdotal, pediatric patients with systolic heart failure appear to respond more favorably than adults to type 3 phosphodiesterase (PDE) inhibitors. Underlying this difference are markedly different gene expression patterns and adrenergic receptor biology in adult versus pediatric populations.

A recent realization is that insufficient adrenergic drive in individuals with CHF can also increase mortality. In contrast to β-blocking agents, sympatholytic agents (agents that lower adrenergic activity) can actually increase mortality, presumably via excessive withdrawal of adrenergic support to the failing heart. In contrast, β-blocking agents are pharmacologically reversible mass-action agents, and in their presence adrenergic support to the failing heart can be accessed if needed simply by increasing cardiac adrenergic drive. One β-blocker in clinical development, bucindolol, is both a nonselective β-blocker and a mild sympatholytic agent. When NE lowering from bucindolol is mild, the result is an enhancement of efficacy. However, in some patients, the degree of sympatholysis produced by bucindolol can be excessive, obviating any benefit realized from the β-blockade. It is now understood that the sympatholytic effects of bucindolol are under control of prejunctional α _2C adrenergic receptors, the implication being that exaggerated/adverse NE lowering can be avoided by not treating patients who have a loss of function α _2C polymorphism in combination with a decreased function β ₁ -AR genetic variant (see below).

Historical studies suggested that one underlying mechanism of NE toxicity involves cyclic adenosine monophosphate (cAMP)-mediated increases in intracellular calcium. Increased formation of the second messenger cAMP leads to activation of protein kinase A (PKA), which in turn phosphorylates a variety of target proteins involved in regulating intracellular calcium, including key targets such as phospholamban (PLB), L-type calcium channels (LTCCs, CaV _1.2 ), and ryanodine receptors (RyR2). Enhanced phosphorylation of calcium channels leads to an increased flux of extracellular calcium into the cell, activating downstream effectors including proteases, kinases, and phosphatases. Activation of the β ₁ -AR subtype also appears to promote increased calcium influx via cAMP-independent mechanisms. Importantly, β ₁ -ARs can be cAMP-independent activators of CaMKII, a signaling pathway known to have particularly adverse effects on cardiac myocytes. Regardless of the exact mechanism(s) invoked, excessive β-AR stimulation results in oxidative stress and calcium overload, conditions promoting cell death through a number of mechanisms, including necrosis and apoptosis. Long-term adrenergic activation has also been shown to cause changes in the gene expression profile of cardiac myocytes, the general response being activation of the so-called fetal gene program that is under β ₁ -adrenergic control via membership in a larger gene signaling network. As discussed in Chapter 1 , the fetal gene program, which is a molecular surrogate for hypertrophy/heart failure, includes downregulation in the gene expression of the fast-contracting α-isoform of myosin heavy chain (MHC) (MYH6) and SR Ca ²⁺ ATPase (ATP2A2) , upregulation of the slow contracting β-MHC isoform (MYH7) , and upregulation of ANP (NPPA) and BNP (NPPB) .

In the failing heart, β-ARs and certain downstream PKA-dependent effectors in the β-adrenergic signaling cascade undergo agonist-mediated, time-dependent downregulation and desensitization, presumably in an attempt to withdraw the heart from an excess of deleterious signaling. These changes result in a marked attenuation of the failing heart’s ability to respond to either endogenous or exogenous catecholamines. Consequently, the amount of cAMP generated in response to a given amount of adrenergic stimulation can be significantly decreased. Evidence has been provided through experiments using isolated, denervated preparations of human heart tissue revealing that failing hearts have a reduced cAMP content compared to nonfailing controls. Extending this observation to the clinical standpoint, reduced responsiveness to catecholamines translates into a decrease in myocardial reserve, which is most commonly manifested as an impaired ability to exercise and means less ability to respond to any form of circulatory stress. In contrast, PKA-independent signaling pathways, such as CaMKII, may undergo upregulation.

In summary, the role of increased adrenergic drive in the natural history of heart failure is a seemingly a paradoxical one. On the one hand, adrenergic activation is essential to meeting the demands of increased physiological and pathological stress in the failing heart; on the other hand, chronic adrenergic stimulation is a major contributor to progressive myocardial dysfunction and remodeling that characterizes heart failure. However, from an evolutionary biology standpoint, this situation is entirely plausible; adrenergic mechanisms have evolved to allow reproductive-age individuals to cope with and escape from highly stressful short-term situations; they are not particularly well suited to chronic support of the failing heart that typically occurs beyond the reproductive years.

Adrenergic Receptor Pharmacology

In the 1940s, Ahlquist recognized that adrenergic receptors could be subdivided into two major classes, α- and β-ARs ( Fig. 6.3 ), both of which serve as binding sites for the endogenous catecholamines, epinephrine and NE. Almost 20 years later, Lands and colleagues, using rank orders of agonists, proposed that β-ARs could be divided into β ₁ - and β ₂ -AR subtypes. A few years later, use of selective β ₁ - and other β-AR antagonists formalized this classification and expanded it to include an additional β-AR subtype, later named the β ₃ -AR. Building on this basic pharmacology was work in the 1970s and 1980s centered on direct identification of receptors by radioligand binding and the biochemistry of signaling pathways. Based on the groundbreaking peptide biochemistry work of Khorana and colleagues on rhodopsin, subsequent work by Lefkowitz and colleagues described the molecular cloning of a number of G-protein coupled receptors (GPCRs) including α ₁ -, α ₂ -ARs, and the β ₁ - and β ₂ -ARs. Perhaps more than any other previous advance in the field, this knowledge permitted the analysis of adrenergic receptor biology in exquisite detail and allowed for an appreciation of the true complexity of adrenergic receptor signaling, including the definition of a multitude of gene regulatory motifs and the recognition of highly organized multi-protein signaling complexes. Ultimately, the combined work of Lefkowitz and Kobilka on GPCR pharmacology, biochemistry, and molecular biology, as well as Kobilka’s subsequent work on GPCR structural biology, resulted in their being awarded the Nobel Prize in Chemistry in 2012.

Relevant to the human heart, there are two major subclasses of α-ARs, α ₁ - and α ₂ -ARs, of which there are several subtypes (see Fig. 6.3 ). These α ₁ -AR subtypes are coupled to the Gq/G11-family of G-proteins, linked to stimulation of phospholipase C (PLC), which in turn promotes phosphoinositide (PIP ₂ ) turnover. The two primary products of this reaction are inositol triphosphate (IP ₃ ), which stimulates the release of intracellular calcium, and diacylglycerol (DAG), which stimulates protein kinase-C (PKC) activity. Increases in intracellular calcium and PKC-mediated phosphorylation of target proteins both activate signaling pathways and gene expression patterns which, depending on the PKC subtype, result in a hypertrophic myocardial phenotype as well as vasoconstriction. Examples of other Gq-coupled receptors whose stimulation also results in a hypertrophic phenotype are the angiotensin II AT1 and endothelin-1 receptors. Each of these receptor pathways represents pharmacological targets central to the management of heart failure and/or its precursors, such as systemic or pulmonary hypertension.

In contrast to α ₁ -, α ₂ -ARs are generally described as inhibitory presynaptic or prejunctional neuronal receptors functioning to suppress the release of NE from the synapse. The potential importance of this function in the setting of heart failure has been emphasized in a study by Small and colleagues, wherein a naturally occurring “loss of function” polymorphism of the α2C-AR, when present in combination with a “gain of function” polymorphism of the β ₁ -AR, produced a significantly increased risk of heart failure in certain populations. These and other β-AR polymorphisms, and their relevance to CHF, are discussed in greater detail below.

β-ARs are also subdivided, with three distinct receptor subtypes encoded by separate genes β ₁ - (ADRB1) , β ₂ - (ADRB2) , and β ₃ - (ADRB3) . β ₁ - and β ₂ -ARs were first described by agonist or antagonist pharmacologic responses, and then by molecular cloning. In contrast, β ₃ -ARs were first definitively identified by molecular cloning after being classified as beyond the β ₁ , β ₂ paradigm or “atypical” in various organs, as determined by pharmacologic response. β ₃ -ARs are generally associated with regulation of metabolic rate, modulation of nitric oxide (NO) formation, gut and urinary tract smooth muscle relaxation and, to a lesser degree, with changes in cardiac contractility. β ₃ receptors are expressed at low abundance but are coupled via a nitric oxide synthase (NOS)/NO/soluble guanylate cyclase/cGMP mechanism to a negative inotropic response in human ventricular myocardium, where they may play a cardioprotective role. In contrast, β ₁ - and β ₂ -ARs have the clear functional role of increasing myocardial performance, and both are expressed at a substantial abundance in human ventricular myocardium with β ₁ -AR density greater than β ₂ .

β ₁ - and β ₂ -ARs, respond equally well to the full agonist isoproterenol, and are distinguished primarily by their affinity for NE. β ₁ -ARs have 10- to 30-fold greater affinity for NE than β ₂ -ARs. An even greater degree of distinction is achieved through the relative affinities of selective antagonists. In the context of heart failure, the most clinically relevant compounds differentiating between β ₁ - and β ₂ -ARs are the β ₁ -AR selective antagonists metoprolol, bisoprolol, and nebivolol. In contrast, β ₃ -ARs are characterized by their relatively low affinity for propranolol, by the selective β ₁ -antagonist CGP 20712a, or by their response to β ₃ selective agonists such as BRL 37344 ⁴² or CL316243. Two β-blocking agents in use or development, nebivolol and bucindolol, have β ₃ -agonist activity that may contribute to their vasodilator properties or possibly to their myocardial effects.

Altered β -Adrenergic Receptor Signal Transduction in the Failing Heart

As manifest in heart failure, chronic heightened adrenergic activation ultimately results in diminished β-AR-mediated signaling as a consequence of decreased in β ₁ -AR gene expression (mRNA and protein expression), uncoupling of both β ₁ - and β ₂ -ARs secondary to receptor phosphorylation from βARK1 (GRK2) or other kinases, and to increases in the inhibitory G protein(Gαi) ( Fig. 6.4 ). In nonfailing, adult human ventricles, β ₁ -ARs constitute approximately 70% to 80% of all β-ARs. However, in individuals with heart failure the β ₁ -AR is selectively downregulated, resulting in an ∼60:40 ratio of β ₁ - to β ₂ -ARs.

Phosphorylation of β-ARs occurs by the G-protein receptor kinase (GRK) family, specifically GRK2, which phosphorylates both β ₁ - and β ₂ receptor subtypes in an agonist occupancy-dependent manner, and GRK5, a kinase of considerable importance linking β ₁ -ARs to epidermal growth factor receptor (EGFR) transactivation and cardioprotective effects, but also to CaMKII signaling, and β ₂ -ARs to Src and MAP kinase (extracellular signal-regulated kinase [ERK]) signaling. As initially described, β-ARs also can be phosphorylated and desensitized by other kinases including PKA and PKC. Therefore in failing ventricular myocardium, both β ₁ - and β ₂ -receptor subtypes undergo desensitization, with β ₁ -ARs exhibiting downregulation, and both β ₁ - and β ₂ -ARs uncoupling from PKA signaling. Similar results have been obtained in the in vivo setting with intravenous administration of dobutamine, demonstrating marked blunting of contractility in the failing human heart compared to no decrement in function with calcium infusion.

The major consequence of β-AR downregulation and desensitization is that for a given level of adrenergic activation, less of the second messenger cAMP is produced. In turn, decreased formation of cAMP leads to diminished activation of PKA. However, decreased PKA activity does not necessarily translate into decreased phosphorylation of all putative PKA substrates. Related to the above, although PKA activity and PLB phosphorylation are decreased in heart failure, cardiac RyR2s have been shown to be hyperphosphorylated. The net phosphorylation “state” of the RyR and PLB, and thus their activities, is due, at least in part, to their being in separate subcellular A-kinase anchoring protein-scaffolded microdomains. In these discreet microdomains different PDEs control cAMP levels and PKA activation; phosphatase species and activities may also differ. Specifically, in the failing heart there is evidence for the downregulation of phosphatases PP1 and PP2A that are associated with RyR2, as well as to changes in PKA activity. Alternatively, recent studies have shown that RyR2 can be phosphorylated by CaMKII, inducing SR calcium leakiness and engendering arrhythmias, an effect that appears to be mediated by cardiac β ₁ -ARs and the “exchange protein directly activated by cAMP-2,” EPAC2. Interestingly, a recent paper by Pereira and colleagues, showed that the EPAC/CaMKII-mediated phosphorylation of RyR2 occurs through a pathway that includes AKT (also known as protein kinase B) and NOS1. The authors showed that increased cAMP levels activate EPAC, which in turn activates AKT resulting in NOS1 activation. NOS1 promotes S-nitrosylation of CaMKII resulting in phosphorylation of RyR2 and increased leakiness.

In contrast to the hyperphosphorylation of RyR2, PKA-mediated phosphorylation of PLB appears to be decreased in heart failure, which is presumably due to reduced cAMP levels in the SERCA II-PLB microdomain and to the activation of the phosphatase PP1. PP1 is regulated by protein phosphatase inhibitor I, PPI-1, a protein whose abundance is itself downregulated in heart failure. The reduced PLB phosphorylation results in a greater inhibition of SR Ca ²⁺ ATPase (SERCA II), resulting in an impaired ability of the heart to both relax and contract. In contrast, hyperphosphorylation of RyR2 results in defective channel function during excitation-contraction coupling. Together, these changes result in decreased contractility in response to adrenergic stimulation and a propensity for developing arrhythmias. Unfortunately, however, the relationship between RyR2 phosphorylation state and its biological relevance is not as clear as might be desired, as articulated in a recent set of “controversies” articles.

Substantial evidence from model systems points to significant differences between β ₁ - and β ₂ -ARs and their ability to stimulate apoptosis in cardiac myocytes with stimulation of the β ₁ - but not the β ₂ -AR pathway resulting in an increased rate of cardiomyocyte apoptosis. Disruption of β ₂ -AR coupling to Gi by treatment with pertussis toxin can abrogate the potential protective, anti-apoptotic signaling pathways rendering the β ₂ -AR pro-apoptotic. Chesley and colleagues, have provided evidence to support the supposition that the anti-apoptotic effects of β ₂ -AR are mediated in part by the stimulation of the PI3K and the Akt-PK-D pathways.

In addition to the mechanisms described above, the pro-apoptotic effect of the β ₁ -AR pathway also may be due to increased CaMKII activity. Zhu and colleagues, have shown that β ₁ -AR stimulation results in myocyte apoptosis, which is independent of PKA activation but is dependent on CaMKII, suggesting that β ₁ -AR stimulation may selectively result in CaMKII activation independently. These studies demonstrate that inhibition of CaMKII, LTCC phosphorylation, or buffering of intracellular Ca ²⁺ results in an attenuation of β ₁ -AR mediated apoptosis; conversely, overexpression of CaMKIIδ _c induces apoptosis. In addition, Yoo and colleagues, using transgenic mice with knockouts of β-AR subtypes, have demonstrated that post-infarct, isoproterenol-mediated increases in CaMKII signaling and apoptosis are dependent exclusively on the β ₁ -AR. However, recent data from Dewenter and colleagues indicate that CaMKII signaling persists despite the use of β-blockers in both experimental and human heart failure, arguing for the development and use of selective CaMKII inhibitors in heart failure. This conclusion is supported by the findings of Kreusser and colleagues, indicating that in a mouse transverse aortic constriction (TAC) model, double knockout of CaMKIIδ/γ progression of cardiac dysfunction and deposition of fibrosis is attenuated.

Excessive β ₁ -AR signaling also activates the fetal gene program ( see also Chapter 1 ). In this context, a change in MHC isoform expression favoring the low(er) ATPase activity β isoform would be expected to decrease contractile function, as would the downregulation in SERCA II, another adult gene that is considered part of the adult-fetal program. Although the signaling cascades responsible for this activation are complex, CaMKII appears to play a prominent role, and the altered expression of so-called fetal and adult genes appears mediated by β ₁ - and not β ₂ -AR stimulation. CaMKII activity is upregulated in the failing human heart, suggesting that despite the desensitization that occurs in more proximal β ₁ -receptor signaling steps, adverse signaling through CaMKII may be maintained or even enhanced ( see also Chapter 1 ).

As described above, CaMKII is involved in various pathologic processes in heart failure including apoptosis, arrhythmias and changes in gene expression. To gain a better understanding of the role of CaMKII in pathologic cardiac disease, knockout mouse models of CaMKII and a transgenic model overexpressing a CaMKII peptide inhibitor have been generated. Mice expressing the inhibitory peptide are resistant to pathological changes related to β-AR stimulation, and do not exhibit decreases in fractional shortening, increases in heart/body weight or decreases in left ventricular internal diameter. Different results are obtained in knockout mouse model studies. In response to TAC in 6-week-old animals, knockout of CaMKIIδ blunted fibrosis and reduced heart/body weight ratio, myocyte area, and pathologic changes in gene expression. In an 8-week TAC model, CaMKIIδ downregulation did not prevent the development of pathologic hypertrophy, but blunted the transition to heart failure, including chamber dilation, ventricular dysfunction, lung edema, cardiac fibrosis, and apoptosis. However, a 6-week-old CaMKII knockout mouse model of severe thoracic aortic banding showed improved contractility, increased myofilament sensitivity to Ca ²⁺ , decreased apoptosis but increased fibrosis. However, no improvement in ejection fraction/fractional shortening was observed. The authors also showed that knockout of CaMKII in the setting of severe heart failure may be detrimental due to reduced efficiency of β-AR regulation of contraction and exacerbation of diastolic dysfunction. These studies suggest that although CaMKII is likely to be primarily detrimental, it may also be important for β-AR regulation of contractile function in the failing heart.

In contrast to the general desensitization phenomena exhibited by β ₁ - and β ₂ -ARs in failing human ventricular myocardium, β ₃ -ARs exhibit upregulation. Since the β ₃ -AR knockout mouse undergoes excessive remodeling in response to pressure overload, increased expression of β ₃ -ARs in the failing human heart may be a salutary adaptive mechanism.

Molecular Basis of β -Adrenergic Receptor Signaling

Canonically, β ₁ - and the β ₂ -ARs couple to stimulatory G-proteins, Gs, which are composed of three subunits, an α-subunit, and a βγ-heterodimeric subunit, Fig. 6.5 . Both the α and βγ subunits are competent to signal independently Gαs is a member of the much larger class of G-protein α-subunits, of which there are ∼20 subtypes. Their heterogeneity is a central basis of specificity for GPCR signaling, a subject reviewed in detail elsewhere. Additional layers of signaling diversity are afforded by combinatorial permutations of various βγ-heterodimeric subunits (5 β-subtypes and 11 γ subtypes) with the various α-subunits.

As a result of agonist stimulation, G-protein α and βγ subunits dissociate with the free αs-subunit acting to stimulate adenylate cyclases (AC), for which there are multiple isoforms. In cardiac tissues, the most abundant isoforms are AC V and VI with the activities of each being feedback inhibited by increases in cytosolic calcium. In turn, adenylate cyclase stimulation increases production of cAMP that, in turn, binds to the regulatory subunits of PKA. Binding causes the dissociation of the regulatory and catalytic subunits of PKA, with the catalytic subunits proceeding to phosphorylate consensus serine (S) and threonine (T) residues on a broad spectrum of intracellular target proteins. In a number of tissues, including cardiac, PKA and its targets are in close approximation because of A-kinase anchoring proteins (AKAPs). An association between the β ₂ -AR and AKAP79 has been described. Additionally, there is recent clarity as to the scaffolding proteins that interact with β ₁ -ARs. Thus, unique interactions between β ₁ - and β ₂ -ARs and specific AKAPs and/or other anchoring proteins support the “micro-domain” concept of receptor subtype-specific signaling, discussed in greater detail below.

Via the “canonical” Gαs/adenylyl cyclase pathway, β-AR stimulation results in the phosphorylation by PKA of a number of proteins including (LTCC, CaV _1.2 ) and PLB, the primary modulator of the sarcoplasmic reticulum (SR)-associated ATP-dependent calcium pump, SERCA II, as well as modulatory proteins associated with regulation of the contractile apparatus, for example, troponin I (TnI). More directly, however, PKA phosphorylates β-ARs, as well as other GPCRs, resulting in partial uncoupling and desensitization of receptors to further agonist stimulation. Changes in the phosphorylation state of calcium-handling proteins results in increased [Ca ⁺⁺ ]i and enhanced rates of myocardial contraction and relaxation. However, the increase in intracellular calcium has a number of other effects beyond enhanced contractility, including the activation of a number of calcium-sensitive enzymes, for example, Ca ⁺⁺ /calmodulin and calcineurin (phosphatase PP2B). On a more global level, changes in [Ca ⁺⁺ ]i can result in profound changes in gene expression patterns.

It has been known for some time that β ₁ - and β ₂ -ARs differ in their efficiency of coupling to adenylate cyclase and to cAMP production, with the β ₂ -AR subtype being more efficiently coupled to cAMP production than is the β ₁ -AR. In trabeculae isolated from the human heart, selective stimulation of contraction by either receptor pathway generally correlates well with receptor abundance. However, the fact that the β ₂ -AR produces more cAMP per unit of stimulus does not necessarily translate into a greater inotropic potential for β ₂ -ARs. Rather, these data favor an argument that there is distinct intracellular compartmentalization of cAMP pools produced by β-ARs, with the existence of discrete microdomains of high cAMP concentration in cardiomyocytes. Specifically, in neonatal cardiomyocytes, high concentrations of cAMP are found in the proximity of T tubules and the junctional SR. An argument is also made that diffusion of cAMP outside of these microdomains is quickly curtailed by the action of PDEs. Support for this notion is the finding that there appears to be close coordination between agonist-stimulated recruitment of PDEs to the β ₂ -AR, a process facilitated by interaction with β-ARs, and the attenuation of signaling. Recent data indicate that the differential signaling and architecture of β ₁ - and β ₂ -ARs appears to be defined, in part, by a differential association with PDEs, with PDE4 activity preferentially targeting cAMP produced by the β ₁ -AR, whereas cAMP produced by the β ₂ -AR being targeted by multiple subtypes of PDEs. Perhaps counterintuitive to the highly localized responses of β-ARs being dictated by scaffolding proteins is the recent study by Nikolaev and colleagues, demonstrated in transgenic mice that isoproterenol-mediated increases in cAMP in cardiac tissues leads to “far-reaching” cAMP signals for the β ₁ -AR, whereas β ₂ -AR generated cAMP appears to remain locally confined. An excellent recent study detailing the compartmentalization of β-AR signaling, the interplay of PDEs, phosphatases, and subtype specific spaciotemporal patterning has recently been published by Xiang.

A very recent study delineates further cAMP and β-AR compartmentalization. Interestingly, in murine models, β ₂ -ARs and cAMP demonstrate anatomical variation such that at the apex of the heart β ₂ -ARs uniquely modulate increased contractility. This appears to be primarily due to variations in the size and organization of T-tubules and caveolae impacting microdomains.

It is clear that both the β ₁ -AR and the β ₂ -AR pathways couple to signaling pathways beyond the “traditional” Gαs/adenylyl cyclase pathway. Perhaps the best example of this diversity of signaling is the finding that the β ₂ -AR can interact with the inhibitory G-protein, Gαi. This is of interest in the setting of CHF, for two reasons. As described earlier, increased circulating NE concentrations result in chronic activation of β-ARs. This, in turn, results in increased βARK activity and an increased phosphorylation state of β-ARs resulting ultimately in the uncoupling of the receptor from the stimulatory Gαs signaling pathway. It is this phosphorylated form of the β ₂ -AR that has an increased propensity to couple to the inhibitory Gαi pathway and its downstream effectors, including inhibition of AC and activation of MAPKs. Amplifying this effect is the finding that the abundance of Gαi proteins is significantly elevated in the failing human heart. Thus, the increased phosphorylation state of the β ₂ -AR and the increased abundance of Gαi protein both facilitate increased trafficking of β ₂ -AR signaling via Gαi-associated pathways. The ramifications of this are potentially significant, particularly in regard to protective, anti-apoptotic effects as well as to myocardial growth regulatory effects.

The β ₁ -AR, although thought to be less “promiscuous” than the β ₂ -AR in its ability to couple to pathways other than to Gαs, nonetheless does exhibit signaling via additional pathways, as noted above, to the CaMKII pathway. However, recent data from Wang and colleagues support the conclusion that the β ₁ -AR can indeed couple to Gαi, an effect uniquely promoted by the β-AR antagonist, carvedilol. In fact, Gαi appears to be required for β ₁ -AR β-arrestin-biased signaling for both EGFR transactivation and for MAPK signaling. However, unlike β ₂ -AR Gαi-mediated ERK activation, carvedilol-induced β ₁ -AR/Gαi MAPK signaling does not appear to require Gβγ nor GRK/PKA receptor phosphorylation. As noted above, current published evidence supports the finding that carvedilol’s induction of β ₁ -AR/Gαi interaction is unique; it will be of future interest to determine whether or not nebivolol, also a β-arrestin-biased agonist, has any requirement for Gαi.

At least in part, β ₁ -AR extra-Gαs coupling, much like that of the β ₂ -AR, appears to be facilitated by the specificity of the carboxy-terminus PDZ (PSD-95/Dlg/ZO-1) domain of the receptor peptide, a domain that drives the interaction with proteins containing specific PDZ recognition motifs. An excellent review on the general topic of regulation of β-AR signaling by scaffolding proteins has been published by Hall and Lefkowitz. Examples of β ₁ -AR interacting proteins are PSD-95, MAGI-2, the endophilins (SH3p4/p8/p13), and CNrasGEF, all of which appear to have unique downstream roles. The association of the β ₁ -AR with PSD-95, which is a neuronal-associated protein, distinctly places the β ₁ -AR in the proximity of the synapse, a location perhaps driven by adrenergic innervation. Here, PSD-95 appears to inhibit the internalization of the β ₁ -AR, an effect that is modulated by the kinase GRK5. Conversely, interaction of the β ₁ -AR with endophilins appears to negatively modulate Γαs coupling and to promote β ₁ -AR internalization. A more detailed picture of the β ₁ -AR SAP97-AKAP79/150 scaffold at its PSD-95/DLG/ZO1 motif has been demonstrated to be necessary for both PKA mediated phosphorylation of the β ₁ -AR as well as for receptor trafficking. This is in contrast to the findings describing the association of the β ₁ -AR with CNrasGEF (PDZ-GEF1), an effector molecule that activates Ras, and thus growth pathways. In this case, the mechanism of Ras activation is rather unique given that it appears to be via Gαs rather than the well-documented Gβγ signaling pathway.

Additional evidence detailing the importance of PDZ-mediated interactions has been described by Xiang and colleagues. Specifically, mutation of the PDZ motif by disruption of the carboxy-terminus renders the β ₁ -AR significantly more sensitive to agonist-mediated cell surface downregulation. Further, deletion of C-terminal amino acids renders the β ₁ -AR more β ₂ -AR-like in that it becomes capable of interacting with Gi. Thus, even though the β ₁ -AR and β ₂ -AR both contain PDZ motifs, the interacting proteins that each couples to are unique and serve as an important basis of different regulatory and signaling properties. To expand this concept, the β ₂ -AR, has been shown to modulate signaling independent of its interactions with G-proteins. For example, the β ₂ -AR appears to couple to at least three other pathways: (1) NHERF, which in turn regulates the function of the Na ⁺ /H ⁺ exchanger ; (2) to a non-PKA-dependent interaction with L-type calcium channels ; and (3) coupling to the phosphatidylinositol 3′-kinase pathway (PI3K), a pathway associated with the inhibition of apoptosis, and to the PI3K-PTEN pathway, which is relevant to the regulation of both myocardial contractility and cell size and shape.

As alluded to above, β-AR stimulation results in the dissociation of the stimulatory G-protein, Gs, into α and βγ subunits, each of which are signaling entities. A major function of the βγ heterodimer is to recruit βARKs in proximity to the receptor. By virtue of the lipid-modified (isoprenylation) of the γ subunit, which promotes localization of the βγ subunit to the sarcolemmal membrane, the heterodimer is in position to orchestrate the co-localization of the β-AR with βARK. Protein/protein interaction of the βγ subunit with βARK occurs via a pleckstrin homology (PH) domain. In this way agonist-dependent, βARK-mediated phosphorylation of β-ARs serves as an important mode of β-AR desensitization. In addition to playing a scaffolding role, βγ subunits are also signaling molecules in their own right. Downstream targets of signaling include subsets of adenylyl cyclases, PI3K, K ⁺ channels, and Ras/Raf/MEK/MAPK pathway(s) (see Fig. 6.5 ).

Recently, significant attention has been paid to the specifics of β-AR localization and receptor trafficking. Differential targeting of β-AR subtypes to specific intracellular compartments has been described. Of note is the apparent caveolar co-localization of the β ₂ -ARs with the inhibitory protein, Gαi, as well as with PKA subunits and with GRK2. What is now becoming clear is that the localization of the β ₂ -AR to the caveolar compartment is an essential component of its signaling capabilities. In contrast, β ₁ -ARs, as well as a number of other proteins including Gαs, and ACs V and VI, are not specifically associated with caveolae; instead, these proteins have a more generalized distribution.

An important and newly emergent concept of β-AR signaling is that of the connection between β ₁ -AR signaling with the cardioprotective effects of the EGFR pathway. Work from Rockman and colleagues has shown that GRK5/6 mediated phosphorylation of the β ₁ -AR targets the association of β-arrestin to the receptor, which, in addition to activating MAP kinases, facilitates the activation of matrix metalloproteinases causing the release of heparin-bound EGF, which then activates the EGFR. Activation of the EGFR leads to activation of Ras/Raf/ERK/Akt and subsequent anti-apoptotic, cardioprotective effects (see Fig. 6.5 ).

Recent evidence indicated that the differentiation of β ₁ - versus β ₂ -AR signaling also extends to the cardiac progenitor cells (CPCs) and their differentiation and survival. Interestingly, β-AR stimulation appears to promote the differentiation of CPCs via β ₂ -ARs. However, once these cells become committed to a myocyte lineage, they begin to express β ₁ -ARs. Counteracting that progression, continued β-AR stimulation via β ₁ -AR leads to a loss of CPCs.