Introduction

Top Abstract Introduction Use of Transgenic Models... Use of Transgenic Mice... Gene Knockout Strategies in... Inhibition of GRK2 in... Consequences of Specific... Conclusions and Future... References

Regulation of the cardiovascular system is under tight control in order to establish and maintain homeostasis. G protein-coupled receptor (GPCR) signaling is an essential component of this regulation. Adrenergic receptors (ARs) are a member of the superfamily of GPCRs that bind epinephrine and norepinephrine, thereby mediating their intracellular effects. Within the AR family, there are three main subfamilies, the ₁, ₂, and ARs. Each of these subfamilies comprises three subtypes. The ₁AR subfamily is made up of _1A, _1B, and _1C; the ₂ARs comprise _2A, _2B, and _2C; and the AR subtypes include ₁, ₂, and ₃. The signaling complexity does not reside solely in the large number of transmembrane-spanning receptor subtypes, as each receptor is coupled to distinct heterotrimeric G proteins that activate a multitude of second messengers upon agonist occupancy (Koch et al., 2000). In addition, regulation of signaling through the ARs is desensitized by a family of GPCR kinases (GRKs) that phosphorylate the receptor when it is in the agonist-bound conformation (Lefkowitz, 1998). It is apparent that the study of cardiovascular AR regulation is complex, and for this reason, many of the initial studies to delineate the pathways were completed in vitro using either cell lines or primary cell culture. However, more recent data suggest that in vitro situations do not necessarily duplicate in vivo interactions; therefore, in vivo studies are critical to understand the physiological and pathological significance of adrenergic receptor regulation (Eckhart et al., 2000; Gainetdinov et al., 2000).

	Use of Transgenic Models to Delineate in Vivo GRK Selectivity

Top Abstract Introduction Use of Transgenic Models... Use of Transgenic Mice... Gene Knockout Strategies in... Inhibition of GRK2 in... Consequences of Specific... Conclusions and Future... References

The obviousness for the need to study GPCR specificity in vivo is evident when examining the properties of GRKs. To date, seven members of the GRK family, GRK1-7, have been identified (Gainetdinov et al., 2000). Of the seven members of the GRK family, GRK2, GRK3, GRK5, and GRK6 have been found in the cardiovascular system, although GRK2, otherwise known as the -adrenergic receptor kinase 1 (ARK1), GRK3, and GRK5 are the most abundant forms (Feldman et al., 1995; Koch et al., 2000). The GRKs have a tri-domain structure. The central catalytic domain has homology to other kinases such as the cAMP-dependent protein kinase A. This central domain is flanked by amino-terminal and carboxyl-terminal domains, both of which contain elements involved in the regulation of GRK activity (Inglese et al., 1993; Palczewski, 1997). Whereas GRK5 binds phospholipids through a poorly defined carboxyl terminal polybasic domain and is normally associated with the membrane, GRK2 and GRK3 reside in the cytoplasm. Upon agonist binding to the appropriate GPCR, GRK2 and/or GRK3 translocate to the membrane via association of the carboxyl-terminal portion that contains a pleckstrin homology domain and which can interact specifically with the membrane-bound -subunits of activated G proteins (Inglese et al., 1995). Interestingly, although GRK2 and GRK3 share 85% identity across their entire sequence, they share a more limited 52% identity within this carboxyl-terminal G binding domain (Koch et al., 1993; Daaka et al., 1997). Therefore, although these kinases share a common endpoint of GPCR phosphorylation, the disparity between them suggests that there are differences in their selectivity, which can hypothetically occur in their membrane targeting in response to different G pools.

In vitro studies have hinted at the need to study selectivity not only in a cell-specific way but also in tissue- and even species-specific manners, and this is particularly apparent when studying GRK phosphorylation of the _1BAR. Initial examination of GRK selectivity against activated _1BARs was performed in vitro using fairly generic fibroblast-like cell lines, including COS-7, HEK293, and DDT1-MF2 (Diviani et al., 1996). The _1BAR is coupled to the heterotrimeric G protein, Gq. Activation of the _1BAR results in phospholipase C (PLC) activation by Gq. PLC produces diacylglycerol (DAG) that activates protein kinase C and inositol 1,4,5-trisphosphate that results in Ca²⁺ accumulation. Overexpression of either the GRK2 or GRK3 in COS-7 and HEK293 cells increased agonist-induced phosphorylation of transiently transfected _1BARs and promoted desensitization of the receptor-mediated stimulation of PLC activation (Diviani et al., 1996). Therefore, there did not appear to be any GRK specificity between these homologous GRKs. Interestingly, purified bovine GRK2 was not capable of phosphorylating endogenous _1BARs purified from DDT1-MF2 cells (Diviani et al., 1996). Moreover, it was also found that stable overexpression of GRK2 in rat FRTL-5 cells was not able to desensitize inositol 1,4,5-trisphosphate accumulation due to stimulation of endogenous _1BARs (Iacovelli et al., 1999). In addition, it was also seen in the COS-7 and HEK293 cells that there were species differences between using rat and bovine GRK3 to transfect into the cells. Therefore, although both GRK2 and GRK3 are equally capable of phosphorylating _1BARs in vitro, even within the context of cell culture, some GRK properties, perhaps post-translational modification, compartmentalization, and/or conformation of the protein, are apparently important in determining protein-protein interaction and kinase activity. Thus, this illustrates clearly that to understand GPCR signaling within the cardiovascular system, it was essential to study GRK specificity in vivo in the heart.

To study GRK specificity of the _1BAR in vivo we took advantage of transgenic mouse models (Eckhart et al., 2000). Transgenic mice have been created with cardiac overexpression of either a constitutively active mutant of the _1BAR (Milano et al., 1994b) or the wild-type _1BAR using the cardiac-specific -myosin heavy chain (MHC) promoter (Akhter et al., 1997). In addition, mice with cardiac-targeted overexpression of GRK2 (Koch et al., 1995), GRK3 (Iaccarino et al., 1998), and GRK5 (Rockman et al., 1996) were all available. We bred the various transgenic mice to create double transgenic, hybrid mice. Thus, we had overexpression of both the _1BAR and the various GRKs, which provided unique and powerful models to examine cardiac-specific in vivo selectivity.

Cardiac GRK2 overexpression, at both low and high levels, was unable to inhibit _1BAR-mediated hypertrophy, DAG content, ventricular ANF re-expression, and mitogen-activated protein (MAP) kinase activation (Eckhart et al., 2000). These in vivo results were similar to the findings in cell culture that noted that GRK2 overexpression did not affect endogenous _1BAR signaling, but they contrasted with the findings that GRK2 overexpression inhibited overexpressed _1BAR-mediated signaling. Importantly, GRK3 completely attenuated all of the in vivo _1BAR signaling parameters (Eckhart et al., 2000). GRK5 had variable effects in vivo and was capable of inhibiting _1BAR-mediated hypertrophy, ventricular ANF re-expression, and MAP kinase signaling but not DAG content (Eckhart et al., 2000). Therefore, these in vivo studies illustrate that there is specificity among GRK2, GRK3, and GRK5 in desensitizing _1BAR signaling in the heart and that the GRKs play distinct roles in the normal regulation of myocardial signaling and function. Moreover, these results indicate that the relevant GRK for _1BAR desensitization in the heart appears to be GRK3.

Other in vivo studies have also shown selective GRK selectivity in the heart. Although GRK2 overexpression was capable of inhibiting both AR-mediated and angiotensin II receptor-mediated stimulation of cardiac function (Koch et al., 1995), GRK5 overexpression chronically uncoupled myocardial ARs but not angiotensin II receptors (Rockman et al., 1996). In vivo cardiac GRK3 overexpression had no effect on AR or angiotensin II signaling but inhibited myocardial thrombin-mediated MAP kinase stimulation (Iaccarino et al., 1998). The lack of AR desensitization in GRK3-overexpressing mice was surprising; however, it illustrates that GRK2 and GRK3 are not isozymes but distinct GRKs that act on independent sets of GPCR substrates in vivo. Therefore, in vivo in the heart, GRK2 phosphorylates ARs and angiotensin II receptors, GRK3 is selective for the _1BAR and thrombin receptors, and GRK5 phosphorylates ARs. Importantly, these in vivo studies have allowed us to discern selective preferences for the GRKs in the absence of specific chemical antagonists.

	Use of Transgenic Mice to Delineate Differences in Cardiac AR Signaling

Top Abstract Introduction Use of Transgenic Models... Use of Transgenic Mice... Gene Knockout Strategies in... Inhibition of GRK2 in... Consequences of Specific... Conclusions and Future... References

Because of the similarity in sequence within the adrenergic receptor subfamilies, it also has been challenging to distinguish distinct signaling pathways and physiological consequences of the cardiac ₁AR and ₂ARs using pharmacological methods. In addition, acquisition of the data for the in vivo cardiac ₁AR and ₂AR pathways is essential because it provides us with the ability to better address treatments for cardiovascular diseases in which there is altered GPCR signaling. For example, in heart failure, there is an impairment of the myocardial AR system such that there is a decrease in total ARs, a parallel decrease in agonist-stimulated adenylyl cyclase activity, and an even greater decrease in agonist-mediated inotropy (Bristow et al., 1982; Brodde, 1993; Ungerer et al., 1993). Both ₁AR and ₂ARs are expressed in the heart of most mammalian species with ₁ARs being the most abundant subtype expressed at levels close to 75% of the total AR population and the receptor subtype that decreases during heart failure (Brodde, 1993).

Both ₁AR and ₂ARs couple to adenylyl cyclase in the heart resulting in positive inotropy mediated by cAMP and increased intracellular calcium (Bristow et al., 1989), although recently cell studies have revealed that ₂ARs can also couple to G_i, an inhibitor of adenylyl cyclase that is important for AR-mediated MAP kinase activation (Koch et al., 1993, 1994; Daaka et al., 1997; Xiao et al., 1999; Zou et al., 1999). In addition, ₁AR stimulation has been linked to induction of apoptosis, or programmed cell death, whereas ₂AR stimulation has no effect or may be protective (Lefkowitz et al., 2000; Zhu et al., 2001). Therefore, although both receptors are present in the cardiac myocyte and mediate inotropic effects, it appears that ₁AR and ₂AR signaling is fundamentally different, and studies on transgenic mice have allowed us to begin to discern the differences in signaling between the AR subtypes.

Recent studies in transgenic mice with cardiac overexpression of ARs have provided a more thorough appreciation of the signaling complexity. Mice with cardiac-specific overexpression of the ₂ARs at lower levels of overexpression (30-50-fold) to levels greater than 100-fold over endogenous myocardial AR expression have biochemically and physiologically enhanced cardiac function which at nonstimulated baseline levels is equal to or greater than in vivo function of nontransgenic littermate control mice with maximum doses of the -agonist isoproterenol (Milano et al., 1994a; Turki et al., 1996; Liggett et al., 2000). Interestingly, there is no additional contractility response as measured by left ventricular contractility in response to isoproterenol in the ₂AR-overexpressing mice (Milano et al., 1994a), and this is probably due to a marked increase in spontaneously isomerized receptors present in the active conformation (Koch et al., 2000). Importantly, mice with 60- and 100-fold overexpression of ₂ARs have enhanced heart rates and contractility from birth (Koch et al., 2000). The 60-fold overexpressing mice show minimal pathology with age, although the 100-fold overexpressing mice had significant pathology by 1 year (Koch et al., 2000). Mice with greater than 350-fold ₂AR overexpression displayed aggressive cardiomyopathy (Liggett et al., 2000). Therefore, even within the in vivo setting, it is essential to not overwhelm the signaling elements within the organ, thus decreasing the selectivity of interaction. It is difficult to interpret this ₂AR-induced cardiomyopathy, with 350-fold overexpression, because there may be potential for any protein, including benign proteins, overexpressed using transgenic technology to produce nonphysiologic effects (Huang et al., 2000).

As an interesting sideline to AR overexpression, transgenic mice with ventricular-targeted adenylyl cyclase V overexpression displayed increased heart rates and fractional shortening as assessed by echocardiography, although there was no response of in vivo hemodynamics (heart rate or contractility) to isoproterenol infusion, suggesting that the levels of adenylyl cyclase do not constrain AR signaling in the cardiomyocyte (Tepe et al., 1999; Ostrom et al., 2000). Additionally, several downstream targets of AR-mediated phosphorylation have also been manipulated and shown to be important to AR signaling. Phospholamban (PLB) is one such critical AR substrate, in which ablation of PLB is associated with attenuation of the contractile responses to AR stimulation in the mouse heart. This attenuation of isoproterenol-mediated increases in contractility of the PLB-knockout hearts is not due to alterations in the AR signal transduction pathway or the degree of phosphorylation of other cardiac regulatory phosphoproteins in myofibrils and the sarcolemma (Kiss et al., 1997). Therefore, a multitude of components act in concert to increase the complexity of interpretation of AR signaling.

To add further to the complexity, differences in ₁AR versus ₂AR signaling in vivo have recently been documented in transgenic mouse models. Interestingly, whereas an approximately 100-fold increase in ₂ARs appears to be well tolerated, cardiac-targeted ₁AR overexpression at only 5 to 15 times endogenous AR levels causes dilated cardiomyopathy and premature death (Engelhardt et al., 1999; Bisognano et al., 2000). These results mimic what others have found in cell culture (Communal et al., 1999; Zaugg et al., 2000). With this very modest overexpression, these mice present significant early ventricular remodeling, including fibrosis (Engelhardt et al., 1999). Transgenic mice with myocardial-targeted Gs overexpression also develop cardiac pathology (Iwase et al., 1996), indicating a similarity to ₁AR but not ₂AR overexpression (Koch et al., 2000). Thus, ₂AR overexpression offers increased inotropy without the apparent deleterious effects on the heart seen with overexpression of ₁AR or Gs, and these models may provide insight into novel therapeutic strategies for enhancing ₂AR signaling but not ₁AR signaling during compromised heart function. As noted above, these specific differences in pathophysiology seen in these different mice may be due to the apparent ₁AR versus ₂AR differences in stimulation of apoptosis and/or specific G protein coupling.

	Gene Knockout Strategies in the Mouse to Study AR Function

Top Abstract Introduction Use of Transgenic Models... Use of Transgenic Mice... Gene Knockout Strategies in... Inhibition of GRK2 in... Consequences of Specific... Conclusions and Future... References

Intricacies in the distinction between ₁AR and ₂AR signaling have also been examined in mice with gene-targeted deletion of ₁AR or ₂ARs (Rohrer, 1998; Chruscinski et al., 1999). When isolated cardiac muscle from ₁AR knockout mice is stimulated with isoproterenol, inotropic and chronotropic responses are not observed (Rohrer, 1998). Disruption of the ₂AR gene does not have severe implications for cardiac physiology. Consequences are only observed during exercise. This is probably because ₂ARs are the major subtype mediating vasodilation in the vasculature, and loss of these receptors would disrupt vasoregulation and energy metabolism during increased catecholamine release (i.e., exercise) (Chruscinski et al., 1999). Cardiac function is certainly dependent on vascular resistance, and because the ₂ARs are the major subtype in the vascular system, drawing conclusions regarding cardiac ₂AR disruption difficult is difficult.

Recently, a Cre-LoxP technique has been devised that allows for site-specific gene deletion. Mice are generated with targeted insertion of loxP (locus of crossing over) recognition sites for the cre (causes recombination) recombinase enzyme that recognizes these sites and mediates excision of the sequence between the recognition sites (Sauer, 1998). The targeted gene with flanking loxP sites is referred to as the floxed locus. The floxed gene is present in all cells of the body throughout development. The loxP sites are placed such that expression of the floxed gene is unaffected. Tissue-specific targeted deletion is achieved by expression of cre in a tissue-specific manner, such as using the -MHC promoter. Cre expression will mediate excision of the floxed sequence only in those cells in which cre is expressed and the gene of interest will remain and be functional in all non-cre-expressing cells. This cre/lox technique will be particularly powerful for examining the cardiac implications of ₂AR disruption in absence of the vascular effects, and further conclusions will be able to be made as to the distinction in the cardiac signaling of these two AR subtypes. This also will be a powerful strategy to further investigate the role of GRKs in cardiovascular regulation (see below).

	Inhibition of GRK2 in the Heart Has Therapeutic Potential

Top Abstract Introduction Use of Transgenic Models... Use of Transgenic Mice... Gene Knockout Strategies in... Inhibition of GRK2 in... Consequences of Specific... Conclusions and Future... References

As noted above, gene disruption is not always a straightforward or viable technique to study signaling implications and this has been demonstrated in GRK2 knockout mice (Jaber et al., 1996). GRK2 gene ablation leads to embryonic death with severe cardiac malformations (Jaber et al., 1996). Therefore, to study the inhibition of GRK2 activity in the heart, alternative strategies had to be used. The cre/lox strategy described above is certainly a worthwhile strategy; however, another interesting approach that has been used successfully is the technique of signaling interruption via the use of peptide inhibitors expressed in a tissue-specific manner in transgenic mice. This technique has been used to inhibit the activity of GRK2. Since GRK2 activity is elevated in heart failure and is responsible for the lack of AR responsiveness in a number of animal models and human heart failure (Koch et al., 2000), it is of interest to lower GRK2 activity in the heart to observe the physiologic consequences through expression of the last 194-amino acid, pleckstrin-homology domain-containing portion of the GRK2 (ARKct) in the heart. This GRK2 (or ARK1) activity can be inhibited via competition for binding to G released following GPCR stimulation and blockade of GRK2 membrane translocation (Koch et al., 1994, 1995). Mice expressing the ARKct in their myocardium have enhanced left ventricular contractility at baseline and an augmented response to isoproterenol administration (Koch et al., 1995). In addition, breeding this ARKct mouse with a mouse model of heart failure that has disruption of the muscle lim protein "rescued" this mouse model as demonstrated by restored fractional shortening, normalized systolic, and diastolic hemodynamic function, as well as normalized responsiveness to -agonism (Rockman et al., 1998). The use of this peptide inhibitor strategy in vivo illustrates the devastating impact of elevated GRK2 levels in the setting of heart failure and indicates that GRK2 inhibition is a novel therapeutic strategy for heart failure (Koch et al., 2000). In addition, in the absence of chemical inhibitors of GRK2 and the lethality of complete GRK2 ablation, it provides a broader understanding of the importance of minimizing the activity of this enzyme to restore AR signaling and normalize cardiac function.

	Consequences of Specific Inhibition of Myocardial Gq-Mediated Signaling

Top Abstract Introduction Use of Transgenic Models... Use of Transgenic Mice... Gene Knockout Strategies in... Inhibition of GRK2 in... Consequences of Specific... Conclusions and Future... References

Like the ARKct, a peptide inhibitor strategy has been used to study the role of class-specific G protein signaling in the heart (Akhter et al., 1998). It has been well documented in vitro that activation of Gq signaling, through receptors including the ₁ARs, angiotensin II, and endothelin receptors, leads to hypertrophy via a number of mechanisms. In vitro, these mechanisms include MAP kinase activation and calcium signaling; however, it was not clear whether this was also the case in vivo. Studies in vivo using Gq-coupled receptor agonists did result in cardiac hypertrophy; however, because these agents exert potent vasoconstrictive effects, it is impossible to definitively assess the role of cardiac Gq-coupled signaling using this approach. To circumvent this, a peptide inhibitor specific for inhibiting Gq signaling was targeted to the hearts of transgenic mice with the MHC promoter (Akhter et al., 1998). This peptide consisted of the last 54 amino acids of Gq (GqI) and represents the carboxyl-terminal portion of the Gq that interacts with the activated Gq-coupled receptor. This peptide can selectively block Gq-coupled signaling, and when it was introduced into the hearts of transgenic mice, it significantly attenuated the ventricular hypertrophy response induced by transverse aortic constriction (TAC) and pressure overload (Akhter et al., 1998). When TAC was applied, the GqI transgenic mice developed significantly less left ventricular hypertrophy at any given pressure load as compared with that of nontransgenic littermate control mice, demonstrating that Gq signaling is crucial for TAC-induced in vivo ventricular hypertrophy (Akhter et al., 1998). Thus, this strategy demonstrated the importance of cardiac Gq-signaling in vivo in isolation of any vascular effects. These data also suggest that class-specific G protein inhibition offers potentially significant advantages over single receptor antagonists, especially in conditions where multiple hormones or neurotransmitters may be involved, such as in hypertension and myocardial hypertrophy (Koch et al., 2000).

	Conclusions and Future Directions

Top Abstract Introduction Use of Transgenic Models... Use of Transgenic Mice... Gene Knockout Strategies in... Inhibition of GRK2 in... Consequences of Specific... Conclusions and Future... References

Studies using cardiac-specific gene targeting of molecules have provided us with a greater understanding of GPCR signaling in vivo. Transgenic mice are powerful models because they allow for tissue-specific modification of various aspects of the signaling cascade. They have helped us appreciate that although there is much sequence similarity among the various AR subfamilies, specific subtypes play very distinct roles. This includes the specific activity of the GPCR-desensitizing GRKs. GRKs phosphorylate and desensitize distinct subtypes and are undoubtedly regulated differently with respect to expression and distribution. The ₁AR and ₂AR possess unique functional consequences, although they both can signal through adenylyl cyclase. Finally, as reviewed above, peptide inhibitor expression makes it possible to discern the tissue-specific impact of cardiac-specific signaling inhibition in the setting of genetic or induced cardiac pathology.

The cardiovascular system is complicated and multifactorial with common elements in both the cardiac and vascular component. The function and regulation of both cardiac and vascular systems are dependent on each other; therefore, transgenic animals provide us with excellent models with which to discern in vivo GPCR signaling distinctions between these two tissues. Currently, studies in transgenic mice have primarily used cardiac targeting utilizing the MHC promoter, but vascular smooth muscle promoters can also be utilized to study the role of AR signaling and GRK selectivity in the vasculature. Importantly, studies in cardiac transgenic mouse models have demonstrated the potential for GRK2 inhibition as a novel therapeutic strategy for improving the function of the failing heart, and GRK2 inhibitors may be a future new class of pharmacologic agonists (Koch et al., 2000). Finally, class-specific Gq inhibition may also represent a novel strategy limiting pathological cardiac hypertrophy and may replace multiple Gq-coupled receptor antagonist therapy.