The inotropic peptide βARKct improves βAR responsiveness in normal and failing cardiomyocytes through G_ßγ-mediated L-type calcium current disinhibition

Abstract

Rationale

The Gβγ-sequestering peptide βARKct derived from the G-protein coupled receptor kinase 2 (GRK2) carboxy-terminus has emerged as a promising target for gene-based heart failure (HF) therapy. Enhanced downstream cAMP signaling has been proposed as the underlying mechanism for increased β-adrenergic receptor (βAR) responsiveness. However, molecular targets mediating improved cardiac contractile performance by βARKct and its impact on Gβγ-mediated signaling have yet to be fully elucidated.

Objective

We sought to identify Gβγ-regulated targets and signaling mechanisms conveying βARKct-mediated enhanced βAR responsiveness in normal (NC) and failing (FC) adult rat ventricular cardiomyocytes.

Methods and Results

Assessing viral-based βARKct gene delivery with electrophysiological techniques, analysis of contractile performance, subcellular Ca²⁺ handling and site-specific protein phosphorylation, we demonstrate that βARKct enhances the cardiac L-type Ca²⁺ channel (LCC) current (Ica) both in NCs and FCs upon βAR stimulation. Mechanistically, βARKct augments Ica by preventing enhanced inhibitory interaction between the α1-LCC subunit (Cav1.2α) and liberated Gβγ subunits downstream of activated βARs. Despite improved βAR contractile responsiveness, βARKct neither increased nor restored cAMP-dependent protein kinase A (PKA) and calmodulin-dependent kinase II (CaMKII) signaling including unchanged protein kinase Cε (PKCε), ERK1/2, Akt, ERK5 and p38 activation both in NCs and FCs. Accordingly, though βARKct significantly increases Ica and Ca²⁺ transients being susceptible to suppression by recombinant Gβγ protein and use-dependent LCC blocker, βARKct-expressing cardiomyocytes exhibit equal basal and βAR-stimulated sarcoplasmic reticulum Ca²⁺ load, spontaneous diastolic Ca²⁺ leakage and survival rates and were less susceptible to field-stimulated Ca²⁺ waves compared with controls.

Conclusion

Our study identifies a Gβγ-dependent signaling pathway attenuating cardiomyocyte Ica upon βAR as molecular target for the Gβγ-sequestering peptide βARKct. Targeted interruption of this inhibitory signaling pathway by βARKct confers improved βAR contractile responsiveness through increased Ica without enhancing regular or restoring abnormal cAMP-signaling. βARKct-mediated improvement of Ica rendered cardiomyocytes neither susceptible to βAR-induced damage nor arrhythmogenic SR Ca²⁺ leakage.

Introduction

Intensive research efforts have been devoted to understanding the pathophysiological role of G-protein coupled receptor kinase 2 (GRK2/βARK1) in diseased myocardium.^1-3 Development of the Gβγ protein sequestering peptide βARKct, which is derived from the GRK2 carboxy-terminus, has emerged as a major consequence of this work, potentially enabling the generation of novel clinical treatments for heart failure (HF). Numerous studies using genetic mouse models as well as viral-based cardiac gene delivery strategies demonstrate that βARKct expression in acutely and chronically failing myocardium rescues performance and improves survival in several experimental HF animal models including post-ischemic, dilated and hypertrophic cardiomyopathy.^4-14

Mechanistically, βARKct improves and restores blunted β-adrenergic receptor (βAR) mediated contractile responsiveness of normal and failing myocardium in vitro and in vivo.^4-14 At least in failing myocardium, resensitized βAR signaling has been proposed as a major underlying molecular mechanism for βARKct's inotropic effects although compartmentalized cAMP-dependent signaling has not yet been determined (reviewed in^1-2). However, given the negative consequences of chronically enhanced cAMP-dependent signaling in HF patients^{15, 16} experimentally reflected by the detrimental outcome of mouse models with cardiomyocyte-specific overexpression of βAR,¹⁷ Gs,¹⁸ PKA¹⁹ and phosphatase inhibitor 1 (I-1)²⁰, there is legitimate concern about using a gene-based strategy potentially facilitating βAR signaling in cardiomyocytes.

Interestingly, it appears βARKct may not be exerting all of its effects directly through GRK2 inhibition and this is likely due to its ability to sequester Gβγ. Studies suggest a direct Gβγ protein-dependent modulation of voltage-dependent Ca²⁺ channels of the N-(α1_B) and P/Q-(α1 ) type.^{21, 22} More recently, an inhibitory interaction between Gβγ proteins and the L-(α1_A) type Ca²⁺ channel (LCC), a critical regulator of basal and βAR-stimulated cardiomyocyte performance, has been reported in oocytes²³. Based on this finding and by virtue of the ability of βARKct to sequester Gβγ proteins, the current study focused on cardiac LCC as a molecular target for inhibitory Gβγ protein and explored signaling pathways conveying βARKct-mediated enhanced βAR responsiveness in normal (NC) and failing (FC) adult rat ventricular cardiomyocytes.

Materials and Methods

An expanded Materials and Methods section is available online at http://www.circaha.org. All animal procedures and experiments were performed in accordance with corresponding institutional guidelines of Thomas Jefferson University and University of Heidelberg.

Experimental rat heart failure model, rat cardiomyocyte isolation and in vitro adenoviral gene transfer protocol

Left ventricular adult cardiomyocytes (FCs) were enzymatically isolated from failing rat hearts employing an experimental post-cryoinfarction heart failure model and sham-operated animals were used to obtain normal cardiomyocytes (NCs).²⁴ To achieve cardiomyocyte βARKct expression in vitro, NCs and FCs were subjected to a second generation replication-deficient serotype 2 βARKct adenovirus (AdβARKct). To control adenoviral transfection efficiency by fluorescence microscopy, AdβARKct expressed βARKct and the green fluorescent protein (GFP) reporter gene under control of two independent cytomegalovirus promoter. Control cells in each group were infected with a corresponding adenovirus carrying the GFP cDNA alone (AdGFP). Our protocol²⁴ resulted in GFP expression in more than 90% of cells after 24 hours in culture without visible signs of toxicity. For detailed protocols, please refer to the expanded methods section.

Cardiomyocyte contractility, intracellular Car²⁺ transients and SR Ca²⁺ load

24 hours after plating, contractile parameters and intracellular Ca²⁺ transients in AdGFP and AdβARKct transfected NCs and FCs were obtained by video edge detection (VED) and epifluorescent assessment of Fura2-AM signals under basal conditions and isoproterenol stimulation.²⁵ Recordings were taken under 2Hz continuous electrical stimulation at steady-state levels both at 5 min and 30 min after isoproterenol stimulation. Sarcoplasmic reticulum (SR) Ca²⁺ content was immediately assessed after termination of Ca²⁺ transient measurements by abrupt exposure to Na⁺/Ca²⁺ free solution supplemented with caffeine (20 mM). The peak of the caffeine-induced cytosolic Ca²⁺ rise was used as semiquantitative index of the SR Ca²⁺ load.²⁶ For further details on chemical treatments, kinase inhibitors and procedures, see the expanded methods section.

Ca²⁺ spark measurements

Ca²⁺ sparks in intact quiescent adult rat NCs and FCs were monitored using a Leica SP2, (Mannheim, Germany) laser scanning confocal microscope (LSCM) under basal conditions and isoproterenol stimulation as described in detail in the expanded methods section. ²⁷ Recordings were started 15 min after isoproterenol stimulation.

Diastolic SR Ca²⁺ wave measurements and assessment of cardiomyocyte cell death

Chronic diastolic Ca²⁺ waves in field-stimulated (2 Hz) and FURA2-AM loaded AdGFP and AdβARKct transfected NCs and FCs were evoked by exposure to combined isoproterenol and caffeine treatment in HEPES-modified medium 199. Cell death was determined in HEPES-modified medium 199 cultured, quiescent AdGFP and AdβARKct transfected cardiomyocytes subjected either to isoproterenol (10^-7 M) or caffeine (1 mM) treatment for 24 h by assessment of ball-shaped (contracted) cardiomyocytes as described previously.²⁸ Concurrent LDH release was measured with a commercially available kit. For detailed description of the procedures, refer to the expanded methods section.

L-Type calcium current recordings

L-type Ca²⁺ currents (I_Ca) were recorded in single cardiomyocytes with the whole cell-patch clamp configuration of the MultiClamp 700 amplifier (Molecular Devices, Sunnyvale, CA, USA).^29-30 To test the effects of Gb/Gy proteins on Ica, pipettes were filled with an intracellular solution (20 mM KCL, 100 mM K-Aspartate (DL), 20 mM TEA-Cl, 10 mM HEPES, 4.5 mMMgCl₂, 4 mM Na₂ATP, 1mM Na₂CrP, 2.5 mM EGTA, pH 7.4 KOH) and recombinant human Gb₃/Gy₃ (for expression and purification see protocol below) was introduced into the cell via the patch pipette. Procedural details are given in the expanded methods section.

Expression and purification of recombinant human Gβ₃ and Gy₃ proteins

Human myc-DDK-tagged Gβ₃ and Gγ₃ were expressed in HEK293 cells and purified by immunoabsorption. Purity and specificity was confirmed by silver staining and anti-DDK/FLAG staining with an anti-DDK monoclonal antibody as described in details in the expanded methods section.

Western blotting, cellular PKA enzyme activity and assessment of cellular cAMP content

Protein expression and site-specific phosphorylation analyses was carried our for GFP, CSQ, βARKct, GRK2, PLB, Ser16-PLB, Thr17-PLB, CaMKII, Thr286-CamKII, TnI, Ser23/24-TnI, S100A4, PKC, p-PKCα, ERK1/2, p-ERK1/2, Akt, p-Akt, p38, p-38, ERK5, p-ERK5 and GAPDH, and assessment of cellular PKA enzyme activity and cellular cAMP content was performed with a commercially available non-radioactive PKA kinase and cAMP enzyme immunoassay kit, respectively, from cell (in vitro isoproterenol stimulation for 30 min) and tissue lysates as described in details in the expanded methods section.

RyR and Ca_v1.2 Immunoprecipitation and Ca_v 1.2 back phosphorylation with PKA

Immunoprecipitation of RyR2 and Ca_v1.2α²⁶ and PKA-dependent back phosphorylation of immunoprecipitated Ca_v1.2³¹ was carried out as described in details in the expanded methods section to assess RyR2 serine-2808 and serine 2815 phosphorylation levels and interaction between Gβγ and Ca_v1.2α and PKA-dependent phosphorylation of the Ca_v1.2α and Ca_v1.2α subunit. For details see expanded methods.

Adeno-associated virus serotype 9 generation and systemic in vivo AAV9 cardiac-targeted gene transfer protocol

In vivo cardiac-targeted βARKct expression in normal mouse hearts was obtained by using retro-orbital injection of an AAV9 harboring the βARKct gene driven by a cardiomyocyte-specific CMV-MLC0.26 promoter. AAV9 harboring the GFP gene under the same promoter served as control. Adult male C57/B6 mice (age 12 weeks) were anesthetized with isoflurane (2%) and 200 μl of 37° heated PBS containing either 1×10¹¹ total viral particles of either AAV9-GFP or AAV9-βARKct were injected into the circulation accessing the retro-orbital venous plexus via the medial corner of the right eye with a 28G1/2 insulin syringe. For detailed description on AAV9 generation and systemic gene transfer refer to the expanded methods section.

Isolation of mouse heart cardiomyocyte and non-cardiomyocyte fraction from AAV9-transfected hearts and echocardiographic analysis of cardiac function

βARKct and GFP expression, enzymatic isolation of cardiomyocyte and non-cardiomyocyte fractions and functional assessment in normal mouse hearts³ were performed 6 weeks after systemic AAV9 application by Western blotting and echocardiography as described in detail in expanded methods.

Statistics

Data are generally expressed as mean±SEM. An unpaired two-tail student's t-test and twoway repeated measures ANOVA were performed for statistical comparisons. For all tests, a P value of <0.05 was considered as significant.

Results

βARKct-mediated inotropic effects in cardiomyocytes require direct β-AR stimulation

We first analyzed Ca²⁺ handling in NCs and FCs subjected either to βARKct (AdβARKct) or control adenovirus (AdGFP). There was a greater increase in Ca²⁺ transient amplitudes in field-stimulated (2Hz) βARKct-expressing NCs over control when exposed to the βAR agonist isoproterenol (Figure 1A). In contrast, both groups responded similarly to the adenylyl cyclase (AC) activator forskolin (FSK). Similar to NCs, direct βAR but not AC stimulation triggered a persistent gain-in function in βARKct-expressing FCs over AdGFP (Figure 1B). Basally, βARKct did not result in an inotropic effect in NCs or FCs (Figure 1A and B). Results for fractional shortening mirrored this pattern (Table 1). In line with our in vitro data, βARKct-expressing mouse hearts, using cardiomyocyte-targeted adeno-associated serotype 9 virus (AAV9) mediated transgene expression, exhibited increased left ventricular fractional shortening compared to controls when stimulated with isoproterenol but not with the AC activator NKH477. Statistical analysis and cardiomyocyte-targeted βARKct expression in NCs and mouse hearts is shown in Supplemental Figure I.

Figure 1. βARKct enhances βAR-dependent contractile reponsiveness in NCs and FCs independent of SR Ca²⁺ handling.

A and B, Representative tracings of Ca²⁺ transients in field-stimulated (2Hz) control (AdGFP) and βARKct-expressing (AdβARKct) NCs (A) and FCs (B) 5 min after stimulation with isoproterenol (Iso) and forskolin (Fsk). Similar results were obtained at 30 min and for Iso 10-7 _M both in NCs and FCs at both time points (data not shown). The dotted line in B indicates the elevated diastolic FURA2 ratio in FCs which is unaltered by βARKct. C, Representative tracings of caffeine-evoked Ca²⁺ transients in NCs and FCs reflecting equally increased SR Ca²⁺ content upon βAR stimulation. Statistical analysis and data for unchanged basal SR Ca²⁺ load are given in Table 1. Black bars indicate addition of caffeine. Dashed lines in A, B and C indicate the inotropic gain by βARKct upon isoproterenol stimulation over corresponding controls. Similar results were obtained for Iso 10-7 M both in NCs and FCs (data not shown). D, Representative confocal line-scans of elementary Ca²⁺ events in control and βARKct-expressing NCs and FCs upon βAR-stimulation. Statistical analysis showing similar Ca²⁺ spark frequency between groups is given in Table 1. FURA2 emission ratios shown in A, B and C are given as arbitrary units. For cell numbers see Table 1.

Table 1.

Contractile and Ca²⁺ handling properties of control (AdGFP) and βARKct-expressing NCs and FCs

	NC-AdGFP	NC-AdβARKct	FC-AdGFP	FC-AdβARKct	number of cells
FS (%)					80-120
Basal	14.1±0.3	14.3±0.3	7.82±0.6‡	8.91±0.5‡
Iso	18.6±0.4†	22.3±0.5*,†	10.1±0.4†‡	20.8±0.8*,‡
Fsk	17.6±0.6†	16.9±0.4†	11.3±0.5†	10.0±0.5†
SR Ca2+ load/Caffeine-CaT (Fura-2 ratio)					50-60
Basal	0.19±0.2	0.18±0.2	0.13±0.1‡	0.14±0.2‡
Iso	0.27±0.3†	0.26±0.2†	0.17±0.2†	0.18±0.2†
Fsk	0.21±0.2†	0.21±0.6†	0.17±0.3†	0.16±0.6†
Ca2+ sparks frequency (events/100μm/sec)					50-60
Basal	3.31±0.3	3.05±0.4	5.62±0.3‡	5.41±0.3
Iso	5.15±0.2†	4.95±0.3†	8.42±0.2†	8.58±0.2†
Peak Ica current (pA/pF)					12-14
Basal	12.1±1.0	14.7±1.6	10.1±1.9	14.1±1.6‡
Iso	18.6±2.4†	27.5±3.1*,†	11.6±2.2‡	28.6±3.1*,†
Fsk	19.6±1.8†	20.1±1.2†	16.3±1.4‡	15.5±2.4‡

Iso, isoproterenol (10⁻⁹ M); Fsk, forskolin (10⁻⁷ M); Caffeine-induced Ca²⁺ transient (Caffeine-CaT) is used as an index of the SR Ca²⁺ load.

^*P<0.05 AdβARKct vs. corresponding AdGFP

^†P<0.05 Iso and Fsk vs. corresponding basal

^‡P<0.05 FCs vs. corresponding NCs.

Similar findings in control and βARKct-expressing NCs and FCs were obtained for 10⁻⁷M isoproterenol (data not shown).

βARKct enhances Ca²⁺ cycling independent of SR Ca²⁺ load and diastolic RyR2 function

Given that SR Ca²⁺ load critically grades the Ca²⁺ transient amplitude upon βAR stimulation,³² we assessed the impact of βARKct on SR Ca²⁺ content in NCs and FCs. Despite greater Ca²⁺ transient amplitudes, neither βAR-stimulated βARKct-expressing NCs nor FCs showed increased SR Ca²⁺ load over control (Figure 1C and Table 1). FSK treatment resulted in the same SR Ca²⁺ content in corresponding GFP and βARKct-expressing cardiomyocytes (Table 1). This observation prompted analysis of diastolic RyR2 function to determine whether βARKct might enhance the diastolic SR Ca²⁺ leak. However, βARKct neither altered basal (Table 1) nor βAR-mediated increases in frequency (Figure 1D and Table 1) or amplitude (data not shown) of elementary Ca²⁺ events in NCs and FCs.

βARKct improves cardiomyocyte Ica upon βAR stimulation

Analysis of steady state whole cell LCC current (Ica) revealed a greater peak Ica amplitude in βAR-stimulated (Figure 2 A and B) but not FSK (Table 1) treated βARKct-expressing NCs and FCs compared with GFP cells. A much smaller but still significant difference in basal peak Ica was also evident between control and βARKct-expressing cardiomyocytes (Figure 2 C and D). Additional measurements enabled us to determine a significant difference in voltage-dependent steady-state availability (f_∞) that further increased upon βAR stimulation while Ica activation (d_∞), Ica time-to-peak and decay constants were not different between groups (Supplemental Figure II). Subsequent calculation of the maximal steady state window Ica (f_∞×d_∞) revealed significantly higher values for βARKct-expressing NCs (control ~5% vs. ~9% βARKct, P<0.05) which is consistent with sustained Ca²⁺ influx at more depolarized membrane potentials. Upon βAR stimulation, f_∞×d_∞ increased similarly (~23%) in βARKct-expressing NCs and FCs (data not shown). In keeping with these data, βARKct rendered field-stimulated Ca²⁺ transients in NCs more susceptible to the use-dependent LCC antagonist nifedipine as reflected by the leftward shift of relative IC50 compared with control (Supplemental Figure II).

Figure 2. βARKct enhances Ica current in NC and FCs upon βAR stimulation.

A to D, Voltage-dependence of peak Ica in control (AdGFP) and βARKct-expressing (AdβARKct) NCs and FCs under βAR-stimulation (A and B) and basal conditions (C and D). Data for equally increased peak Ica in control and βARKct-expressing cells upon forskolin stimulation are shown in Table 1. *; P<0.05 AdβARKct basal and Iso-stimulated NCs and FCs vs. corresponding AdGFP NC and FC controls, †; P<0.05 Iso-stimulated control FCs vs. Iso-stimulated control NCs, ‡; P<0.05 Iso-stimulated control NCs vs. basal control NCs. Cells were derived from three different isolations, n=12-14 cells in each group. Data are shown as mean±SEM.

βARKct does not alter cAMP-dependent kinase signaling downstream of activated βARs

Previous studies described facilitated βAR signaling conferring inotropic actions of βARKct (reviewed in^1-2). We therefore investigated both compartmentalized and total PKA activity. There was equal basal and βAR-triggered PKA-mediated increases of phospholamban (PLB) phosphorylation at serine 16 (PLB-ser16, shown for 10^-9 M isoproterenol) both in control and βARKct-expressing NCs and FCs (Figure 3A). Similar results were obtained for the sarcomeric PKA target troponin I (TnI) at serine 23/34 (TnI-ser23/24) and indistinguishable EC50 for isoproterenol-mediated PLB and TnI phosphorylation matched by equal PKA activity in control and corresponding βARKct-expressing cardiomyocytes (Supplemental Figure III). We also assayed PKA-dependent changes in phosphorylation of the α1C-LCC (Cav1.2α) (Figure 3B), the β2C-LCC subunit (data not shown) as well as RyR2 (RyR-ser2808, Figure 3C). Basal and PKA-dependent backphosphorylation of Ica and site-specific phosphorylation of RyR2 were not different between βARKct-expressing and control NCs and FCs (Figure 3B and C). In further support of a PKA-independent inotropic effect, myr-PKI as well as alternative chemical PKA inhibitors such as Rp-8-Br-cAMPS and H89 (data not shown) blocked the isoproterenol-mediated increase in PLB-ser16 phosphorylation both in control and βARKct-expressing cardiomyocytes but did not abrogate the relative βARKct-mediated gain-in function (Supplemental Figure IV). Assessment of cellular cAMP concentrations unveiled a similar increase in their levels both in isoproterenol-stimulated control and βARKct-expressing NCs and FCs with equally blunted βAR-induced AC activity in the latter group (Supplemental Figure IV). We next determined compartmentalized CamKII activity and activation as downstream target of exchange protein activated by cyclic AMP (Epac).³³ CaMKII-dependent PLB (PLB-thr17) and RyR2 (RyR2-ser2815) phosphorylation as well as activated CaMKII (CaMKII-thr268, Figure 4A-C) was similar in isoproterenol-stimulated control and βARKct-expressing cardiomyocytes. We also found that isoproterenol had indistinguishable effects on PKCα and ERK1/2 phosphorylation levels in βARKct-expressing and control cells (Supplemental Figure VI) as well as on p38, ERK5 and Akt activity (data not shown). Further corroborating a CaMKII- and PKC-independent effect of βARKct, the chemical selective CamKII and broad-spectrum PKC inhibitor myr-AIP and bisindolylmaleimide-1 (BIM), respectively, equally diminished the isoproterenol-induced increase in the Ca²⁺ transient amplitude in both groups without abolishing the βARKct-mediated gain-in function (data not shown). The CaMKII inhibitor KN93 and its inactive analog KN92 were not employed given that both compounds were reported to unspecifically inhibit Cav1.2α activity in a CaMKII-independent manner. ³⁴

Figure 3. βARKct neither alters βAR downstream cAMP-dependent kinase (PKA) signaling in NCs nor FCs.

A, Representative immunoblots (left) reflecting similar basal and isoproterenol-induced PKA-dependent changes in phospholamban (PLB) serine 16 phosphorylation both in control and βARKct-expressing NCs and FCs expressed as PLB-ser16/PLB ratio (right). B, representative Cav1.2α immunoprecipitation and corresponding autoradiography signals for PKA-dependent Cav1.2α back-phosphorylation (left). Statistical analysis expressed as Cav1.2α autoradiography intensity/Cav1.2α protein shows unaltered PKA-dependent back-phosphorylation in isoproterenol-stimulated control NCs and FCs compared with corresponding βARKct-expressing groups. C, Representative immunoblots of immunoprecipitated RyR2 and corresponding PKA-dependent RyR2 serine 2808 phosphorylation (left). Statistical analysis (right) shows RyR2-ser2808 phosphorylation both in control and βARKct-expressing NCs and FCs compared with basal conditions depicted as RyR2-ser2808/RyR2 ratio. *; P<0.05 Iso-stimulated NCs and FCs vs. corresponding basal AdGFP and AdβARKct NCs and FCs, †; P<0.05 Iso-stimulated FCs vs. corresponding Iso-stimulated NCs. Experiments were performed from five different cell isolations. Data are shown as mean±SEM. Data are shown for 10-9_M isoproterenol and similar results were obtained for 10-7M (data not shown).

Figure 4. βARKct neither alters βAR downstream calmodulin-dependent kinase II (CaMKII) signaling in NCs nor FCs.

A, Representative immunoblots (left) reflecting similar basal and CamKII-dependent increases in phospholamban (PLB) threonine 17 phosphorylation both in control and βARKct-expressing NCs and FCs expressed as PLB-thr17/PLB ratio (right). B, Representative immunoblots of immunoprecipitated RyR2 and RyR2 serine 2815 phosphorylation (left). Statistical analysis (right) shows significant but similar increases in RyR2-ser2815 phosphorylation both in control and βARKct-expressing NCs and FCs compared with corresponding basal conditions depicted as RyR2-ser2815/RyR2 ratio. C, representative immunoblots for CaMKII and activated CaMKII (CaMKII-thr286) (left) showing unchanged CaMKII-thr268/CamKII ratio basally and in response to isoproterenol between groups. *; P<0.05 Iso-stimulated NCs and FCs vs. corresponding basal AdGFP and AdβARKct NCs and FCs, †; P<0.05 basal and Iso-stimulated FCs vs. corresponding NCs. Experiments were performed from five different cell isolations. Data are shown as mean±SEM. Data are shown for 10-9_M isoproterenol and similar results were obtained for 10-7_M (data not shown).

βARKct inhibits Gβγ-mediated suppression of Ica in cardiomyocytes upon βAR stimulation

A previous study reported on a direct inhibitory effect of Gβγ proteins on cardiac LCC activity through interaction with the cytosolic amino- and carboxy-terminus of the pore-forming Cav1.2α subunit.²³ Gβγ-sequestering peptides derived either from Gα_i1 or the GIRK1 potassium channel prevented this effect and subsequently enhanced Ica. Activation of heterotrimeric G-proteins by βAR agonists in cardiomyocytes is known to release Gβγ from activated Gα subunits.¹ To assess whether Gβγ can suppress Ica in cardiomyocytes, we assayed the effect of recombinant Gβ₃ and Gγ₃ isoforms on basal and βAR-stimulated Ica in voltage-clamped NCs and FCs. Gβ₃ and Gγ₃ were chosen because studies suggest they are the most abundant Gβγ protein isoforms in rat adult cardiomyocytes.³⁵ Application of 30 nM recombinant human Gβ₃/Gγ₃ suppressed βAR-stimulated Ica both in control NCs and FCs (Figure 5A and 5B). Gβ₃/Gγ₃-mediated suppression of Ica was prevented both in NCs and FCs expressing βARKct (Figure 5C and D). In line with the notion of a potentially rate-limiting concentration of βARKct, increasing Gβ₃/Gγ₃ concentrations by 10-fold to 300 nM eventually abrogated enhanced Ica both in βARKct-expressing NCs and FCs (Figure 5C and D). 3 nM Gβ₃/Gγ₃ did not have an effect in GFP or βARKct-expressing NCs and FCs (data not shown).

Figure 5. βARKct inhibits Gγβ-mediated suppression of Ica in βAR-stimulated NCs and FCs.

A to D, Voltage-dependence of peak Ica in control (AdGFP) and βARKct-expressing (AdβARKct) NCs (upper row) and FCs (lower row) in response to βAR-stimulation and installation of recombinant myc-DDK-Gα₃α₃. Myc-DDK installation served as control and showed no effect on Ica compared to untreated cells (data not shown). Note that 30 nM Gα₃α₃ (dashed red line) supresses Ica only in control (A and B) but not βARKct-expressing (C and D) NCs and FCs. However 300 nM Gα₃α₃ (solid red line) diminished Ica in βARKct-expressing NCs and FCs (C and D). Ica measurements were derived from four different cell isolations. n=8-11 cells in each group. Data are shown as mean±SEM.

βARKct attenuates Gβγ binding to LCC in cardiomyocytes upon βAR stimulation

Gβ-binding to Cav1.2α is enhanced by βAR stimulation in control NCs and FCs compared to basal conditions (Figure 6A). βARKct attenuated Gβ/Cav1.2α interaction resulting in a decrease of Gβ/Cav1.2α binding ratios in NCs and FCs (Figure 6A). We also tested whether recombinantly expressed Gβ₃/Gγ₃ proteins used for Ica measurements bound to Cav1.2α in extracts of control and βARKct-expressing NCs. 30 nM myc-DDK-tagged Gβ₃/Gγ₃ proteins were added to cellular extracts of isoproterenol-stimulated control and βARKct-expressing NCs. Cav1.2α/myc-DDK-tagged Gβ₃ co-precipitated in control but not βARKct-expressing NCs (Figure 6B). Addition of 300 nM recombinant Gβ₃/Gγ₃ protein - a concentration that suppressed Ica in βARKct-expressing cardiomyocytes -resulted in Gβ₃/α1C-LCC co-precipitation (Figure 6B).

Figure 6. βARKct inhibits βAR-triggered interaction between Gγβ and Cav1.2α in NCs and FCs.

A, Representative immunoblots showing immunoprecipitated Cav1.2α (upper panel) demonstrate significantly enhanced co-precipitation of Gα (lower panel) in control NCs and FCs upon β-AR stimulation which is significantly attenuated in by βARKct. Similar results were obtained for Gα (data not shown). Data were normalized to isoproterenol-stimulated control NCs (AdGFP). B, Representative immunoblots for immunoprecipitated Cav1.2α from cellular extracts of isoproterenol-stimulated NCs (10^-7M) with addition of 30 nM myc-DDK-Gβ₃ demonstrate co-precipitation of myc-DDK-Gα₃ and endogenous Gβ in control NCs that is prevented in extracts of βARKct-expressing NCs (left panel). However, a ten-fold increase in myc-DDK-Gα₃ (300 nM) results in Cav1.2α/myc-DDK-Gα₃ co-precipitation. Similar results were obtained for myc-DDK-Gβ₃ and endogenous Gγ (data not shown). Cells for immunoprecipation (IP) were derived from four different isolations. Each IP (n=4) was carried out in duplicates. Data are shown as mean±SEM. Isoproterenol was used at 10^-7M.

βARKct-mediated improvement of Ica renders cardiomyocytes susceptible neither to cell death nor arrhythmogenic diastolic Ca²⁺ waves

Previous studies demonstrated detrimental effects of chronically enhanced Ica in cardiomyocytes both with constitutive α1C-LCC and β2C-LCC subunit overexpression under basal conditions and βAR stimulation in vitro ^{28, 36}. However, cardiomyocytes with βARKct-mediated Ica enhancement exhibited a trend towards improved survival under basal conditions and isoproterenol treatment for 24 h compared with corresponding controls (Supplemental Figure VI). In support of this notion, there was a trend towards greater LDH levels in the supernatant of control cells (Supplemental Figure VI). Indistinguishable cellular viability was also evident between groups after 24 h caffeine (1 mM) treatment, which sensitizes RyR2s to cytosolic Ca²⁺ (Supplemental Figure VI). Of note, when field-stimulated βARKct-expressing and control NCs and FCs were subjected to a combination of caffeine (0.5 mM) and isoproterenol (10^-7 M) that has been shown to trigger diastolic Ca²⁺ waves,³⁷ βARKct apparently conveyed a protective effect by attenuating the onset of arrhythmogenic Ca²⁺ waves (Supplemental Figure VII).

Discussion

Among other targets,² βARKct has emerged as a promising candidate for gene-based therapy targeting abnormal βAR responsiveness in failing myocardium.^4-14 Our study characterizes a Gβγ-based molecular mechanism that attenuates cardiac LCC/Ica in normal and diseased cardiomyocytes upon βAR stimulation. Targeted interruption of this inhibitory pathway by βARKct results in Ica disinhibition and improved βAR contractile responsiveness. Mechanistically, this action is based on the ability of βARKct to sequester Gβγ proteins. Apparently, this mechanism seems neither to require nor rely on cAMP-dependent kinase signaling. These results might have both physiological and clinical implications as they advance the molecular understanding of βARKct towards a cAMP-independent inotrope and, at the same time, provide novel insight into a Gβγ-based mechanism in cardiomyocytes negatively shaping Ica upon βAR stimulation. Of note, βARKct-mediated improvement of Ica rendered cardiomyocytes neither susceptible to SR Ca²⁺ overload nor diastolic Ca²⁺ leakage or enhanced βAR-induced cell damage.

Enhanced Ica as potential mechanism for inotropic effects of βARKct in cardiomyocytes

Given that Ica constitutes one of the dominant factors in controlling cardiac inotropy,^32,38 enhancing Ica upon βAR stimulation appears as a potent means to grade cardiomyocyte performance. This is reflected by augmented βAR-mediated contractile responsiveness in βARKct-expressing NCs exhibiting increased Ica in response to isoproterenol treatment. Our data suggest that a similar mechanism is in effect in FCs yielding blunted Ica increase upon βAR stimulation. The latter has been considered as common pathophysiological feature in FCs across several experimental HF models and human failing cardiomyocytes³⁹. Reversal of this defect seems to be key for βARKct to restore cytosolic Ca²⁺ handling and βAR contractile responsiveness in our experimental setting. The rather modestly enhanced basal Ica on the other hand did not translate into heightened contractile performance and Ca²⁺ cycling in βARKct-expressing cardiomyocytes. Although counterbalancing outward currents carried by potassium (I_K) remain inactivated in the voltage-clamped mode,³⁹ they are active under field-stimulation and likely candidates to effectively antagonize βARKct-mediated moderate enhancement of basal inward Ica. Hence, at least in this study, biologically relevant amplification of cardiac Ica by βARKct seems to be limited to and requires direct βAR stimulation both in normal and diseased rat cardiomyocytes.

Our data reveal that the inotropic effect of βARKct is due to a larger steady state window Ica based on greater LCC availability. Previous studies suggest that sustained Ica can result in progressive Ca²⁺ load in cardiomyocytes potentially compromising contractile performance and cellular viability.²⁸ In addition, progressive SR Ca²⁺ load can also render cardiomyocytes susceptible to arrhythmogenic Ca²⁺ abnormalities.⁴⁰ Our data however reveal that βARKct-mediated Ica enhancement does not lead to SR Ca²⁺ overload with subsequent diastolic Ca²⁺ leakage and exaggerated cell death. This suggests that βARKct-expressing cardiomyocytes can avoid these potentially fatal consequences by balancing increased systolic Ica through augmented diastolic Ca²⁺ efflux under steady state conditions. This can potentially be accomplished in part by altered sodium-calcium exchanger (NCX) activity or alternative mechanisms, which will be a subject for future studies. Conclusively, our results provide novel mechanistic insight and identify Ica as significant mechanism conferring inotropic actions of βARKct upon βAR stimulation – potentially through a Gβγ-based mechanism as further discussed below.

Role of cAMP-dependent signaling for inotropic effects of βARKct

The differential effect of βARKct on Ica and SR function prompted further investigations of Ca²⁺ handling proteins with respect to their ultimate role as molecular effectors of βAR downstream signaling.³² Due to recent reports highlighting the necessity to determine site-specific phosphorylation of cAMP-dependent kinase targets,^41-42 we assessed both PKA- and CaMKII-dependent activity at distinct sarcoplasmic, sarcomeric and sarcolemmal targets which more accurately reflect transient and compartmentalized cAMP gradients than global cellular cAMP measurements. In line with previously reviewed data,^43-44 control experiments revealed abnormal βAR downstream signaling in FCs compared with NCs. Our results unveiled indistinguishable PKA-and CaMKII-dependent phosphorylation patterns at different subcellular sites in βARKct-expressing and control normal and diseased cardiomyocytes closely reflecting similar changes in SR Ca²⁺ load and diastolic RyR2 function upon βAR stimulation. Unchanged EC50 values for distinct PKA-dependent targets as well as total PKA activity and cellular cAMP levels further corroborate the conclusion that βARKct-mediated enhanced βAR contractile responsivenes and Ca²⁺ cycling in our setting is neither based on nor requires enhanced cAMP-dependent downstream signaling at least under in vitro conditions. Importantly, unchanged PKA-dependent phosphorylation of LCC subunits as well as unaltered CaMKII activity render both cAMP-dependent kinases less likely candidates conveying βARKct-mediated changes in Ica.⁴⁵

Extending our signaling analysis into alternative cAMP-independent signaling pathways downstream of activated βARs unmasked an identical pattern. βARKct did not alter PKCα and ERK1/2 or Akt, ERK5 or p38 activation. Although PKCα and ERK1/2 have been implicated in cAMP-independent CamKII activation³³ and α-arrestin/metalloproteinase transactivated EGPR-mediated cardioprotection⁴⁶ while the GRK2 carboxy-terminus including the βARKct sequence has been shown to inhibit Akt⁴⁷, our data do not support a role for those kinases in βARKct inotropic actions in our experimental setting. Though Akt activity is unchanged in our experimental setting, inhibition of PI3Kα can exert similar effects on Ica upon βAR stimulation than βARKct ⁴⁸ and future studies need to determine whether these effects might be mechanistically linked. Collectively, our in vitro data support the conclusion that short-term inotropic effects of βARKct in normal and diseased cardiomyocytes are independent of “classical” cAMP-related and alternative kinase-dependent signaling pathways downstream of activated βARs.

Role of Gβγ protein related signaling for inotropic effects of βARKct

Studies have pointed towards a role of Gβγ proteins in regulating ion channel activity²¹ including voltage-dependent Ca²⁺ channels of the N- (α1_B) and P/Q- (α1_A) type.²² Potentially relevant for cardiac Ica regulation, Dascal and co-workers recently provided evidence that Gβγ can bind to the cytosolic Cav1.2α amino- and carboxy-terminus of the LCC as expressed in full subunit composition in an oocyte system.²³ This interaction attenuated Ica and co-expression of Gβγ sequestering peptides prevented the inhibitory Gβγ/Cav1.2α interaction thus disinhibiting Ica. Herein, we describe for the first time an interaction between Gβγ and Cav1.2α both in NCs and FCs. It is tempting to speculate that agonist-occupied βARs either coupled to Gs or Gi,¹ are a likely source for “inhibitory” Gβγ release potentially reflecting an intrinsic negative feedback mechanism that could limit the increase in Ica upon βAR stimulation in normal and failing myocardium although the G protein source remains to be determined.

Targeted interruption of this signaling by βARKct apparently decrease the inhibitory interaction between liberated Gβγ and Cav1.2α leading to enhanced and restored blunted βAR-stimulated Ica in NCs and FCs by virtue of its ability to sequester Gβγ proteins. Our data might also rejuvenate interest into previously reported inotropic effects of other Gβγ sequestering peptides in cardiomyocytes such as the amino-terminal truncated phosducin.⁴⁹ In line with our molecular data, βARKct-expressing cardiomyocytes exhibited greater sensitivity to the use-dependent LCC blocker verapamil indicating a higher dependency on LCC activity and PKA inhibitors only blunted the PKA-dependent but not Gβγ-dependent βARKct-mediated increase in Ca²⁺ transient amplitudes upon βAR stimulation. Similarly, CamKII and PKC inhibitors were equally ineffective in abolishing the βARKct-mediated gain-in-function. Together with our signaling data, these results strongly suggest that βARKct conveys its inotropic actions both in NCs and FCs, at least in vitro, by sequestering inhibitory Gβγ proteins, which does not translate into augmented cAMP-dependent downstream signaling.

Implications and limitations of the study

Our results provide evidence for a mechanism effective both in normal and diseased cardiomyocytes linking βARKct-mediated enhancement of βAR responsiveness to a Gβγ-dependent pathway that operates independent of cAMP-downstream signaling. Consistent with our data showing unchanged cellular cAMP levels between controls and βARKct-expressing cells basally and in response to short-term βAR stimulation, first studies describing superior βAR-mediated contractile responsiveness in mice with cardiomyocyte-restricted βARKct expression already demonstrated similar increases in AC activity and cAMP formation in vivo as control mice.^{8, 50} These studies therefore pointed towards a potential cAMP-independent mechanism underlying βARKct-mediated enhanced βAR contractile responsiveness. Our current study, which systematically assessed βAR downstream signaling beyond cAMP therefore advances our understanding by demonstrating that neither enhanced PKA nor CaMKII signaling accounts for improved βAR-mediated contractile responsiveness in βARKct-expressing normal and failing rodent cardiomyocytes. Although most reports,^4-14 including ours, use selective βAR agonists, a recent study reported on rather subtle inotropic changes in βARKct-expressing isolated murine cardiomyocytes when exposed to norepinephrine.⁵¹ As the latter stimulates both αARs and βARs which exerts inhibitory and stimulatory effects on Ica, respectively,⁴⁵ it is possible that selective βAR stimulation might provide a “biased” picture of βARKct’s inotropic potency. It might be less effective in the context of endogenous catecholamines and future studies on βARKct are needed employing epinephrine and norepinephrine in experimental HF models approximating human cardiovascular physiology and structure to address this clinically relevant implication.

It is currently unclear why the previously demonstrated ability of βARKct to inhibit Gβγ-dependent GRK2 activity in molecular assays^4-14 is not translated into faciliated downstream AC signaling both in normal and diseased adult ventricular cardiomyocytes in our short-term experimental setting. However, previous studies have used assessment of AC-mediated cAMP formation in myocardial homogenates and membrane preparations and found enhanced cAMP formation.^{4, 52-53} However, this methodological approach neither adequately reflects cellular cAMP homeostasis nor downstream signaling in subcellular domains in intact cardiomyocytes. Our study is the first to systematically determine this in βARKct-expressing adult normal and diseased cardiomyocytes. Notably, although our results are obtained in a short-term in vitro setting, they indicate that Gβγ-dependent mechanisms might co-exist with GRK2 inhibition through which βARKct can improve βAR contractile responsiveness independent of cAMP-dependent kinase signaling. This concept may be consistent with the idea that inhibition of GRK2 resensitizes βARs on the membrane thus allowing for more signaling through heterotrimeric G proteins upon βAR stimulation. Sequestration of liberated Gβγ through βARKct may in turn disinhibit Ica through the βAR-dependent mechanism described in this study potentially leading to enhanced βAR contractile responsiveness without increasing cAMP formation.

Precise understanding of how this potentially cAMP-independent mechanism translates into therapeutic mechanisms that eventually result in adaptive events with secondarily improved AC activity and cAMP homeostsais is currently limited by the in vitro design of our study. Future in vivo studies are needed to delineate the relative contribution and time line of βARKct effects on Gβγ-mediated regulation of Ica and βAR sensitivity by dissecting early and late molecular effects in βARKct-expressing failing hearts and isolated cells. In addition, our study is further limited by the fact that we are currently unaware of other potential Gβγ-dependent cardiac targets of βARKct. Potential interference of βARKct i.e. with nucleoside disphosphate kinase (NDPK) mediated regulation of Gβγ-dependent signaling in cardiomyocytes^54-55 is just one potential candidate. This highlights the need for a more comprehensive approach to identify βARKct targets which are involved in its inotropic actions and eventually translate into previously reported therapeutic effects.

Given comprehensive evidence for the therapeutic benefit of βAR blockers such as metoprolol or carvedilol in clinical practice, it is clear that βARKct-based gene therapy can only be considered as an adjuvant approach to βAR blockers. Previous studies in small experimental HF models already suggested potential synergistic and/or additive therapeutic benefits employing both metoprolol and βARKct-based gene therapy in a rat HF model. Given that βAR blocker therapy protects from sympathetic overdrive but does not result in complete βAR blockade, βAR-dependent βARKct effects are most likely partially preserved in HF upon βAR blocker therapy. However, these data have to be viewed with caution as long as such benefit needs to be shown in large-animal HF models which more closely approximate human cardio-vascular physiology.

Conclusion

In summary, our study identifies a Gβγ-dependent signaling pathway attenuating cardiomyocyte Ica upon βAR as molecular target for the Gβγ-sequestering peptide βARKct. Targeted interruption of this inhibitory pathway by βARKct potentially confers improved βAR contractile responsiveness through increased Ica rather than increased or restored cAMP downstream signaling (summarized in Figure 7). This potentially unveils a new perspective on βARKct towards cAMP-independent effects on cardiac performance and shed new light on therapeutic strategies that improve βAR responsiveness in failing myocardium.

Figure 7. Simplified molecular model for βARKct-mediated Ica enhancement in cardiomyocytes elicited by ßAR stimulation.

(A) ßAR activation results in heterotrimeric G-protein stimulation and release of Gβγ (1). Our results indicate that Gβγ can bind to the LCC and attenuate ßAR-triggered systolic Ca²⁺ influx potentially functioning as a negative feedback mechanism (1). (B) The Gβγ-sequestering βARKct peptide can attenuate the inhibitory interaction between Gβγ and LCC (2) thereby preventing Gβγ-mediated Ica suppression upon ßAR stimulation. Our data suggest that Ica disinhibition (3) could significantly contribute to βARKct-mediated enhancement of ßAR contractile responsiveness. Our study further indicates that inotropic actions of βARKct neither rely on enhanced PKA nor CaMKII (not shown) signaling and occur independent of SR Ca²⁺ load. Under steady-state conditions when Ca²⁺ influx matches Ca²⁺ efflux, it is tempting to speculate that βARKct-mediated Ica enhancement during systole is balanced by augmented diastolic Ca²⁺ extrusion reflected by unchanged net SR Ca²⁺ load. Further studies need to clarify this issue.

Novelty and Significance.

What is known?

• The Gβγ-sequestering peptide βARKct improves α-adrenergic receptor (βAR) responsiveness in experimental heart failure (HF).

• Improved βAR responsiveness by βARKct has been attributed to enhanced cAMP-dependent signaling.

• Gβα proteins can inhibit voltage-dependent Ca²⁺ channels of the N-(α1_B), P/Q-(α1_A) and L-(α1_A) type.

What New information Does this article Contribute

• Gβα subunits liberated from activated βAR/heterotrimeric G-protein complexes inhibit the L-type calcium channel (LCC) current (Ica) increase in cardiomyocytes upon αAR stimulation.

• βARKct improves and restores Ica in normal and failing cardiomyocytes by preventing Gβα-dependent LCC suppression upon αAR stimulation.

• βARKct-mediated LCC disinhibition conveys improved βAR responsiveness independent of cAMP-dependent kinase signaling

Summary

HF is characterized by alterations in βAR signaling due to, at least in part, increased GRK2 levels/activity. Improved βAR responsiveness by the inotropic peptide and GRK2 inhibitor αARKct in normal and failing hearts has been attributed to enhanced cAMP signaling. However, due to detrimental effects of clinical drug regimens that chronically enhance cAMP levels (i.e. phosphodiesterases) legitimate concern exists regarding molecular strategies facilitating αAR signaling in HF. Here, we describe for the first time a cAMP-independent mechanism conveying αARKct-mediated improvement of βAR responsiveness both in normal and failing cardiomyocytes. Binding of liberated Gαα subunits to the LCC seems to limit Ica increases upon βAR stimulation in cardiomyocytes. Targeted interruption of this inhibitory feed-back loop by αARKct through its ability to sequester Gαα proteins appears key to improved βAR contractile responsiveness. Of note, this occurs independent of downstream cAMP signaling, at least in isolated cardiomyocytes, and αARKct-mediated Ica improvement caused neither Ca²⁺-induced cell death nor enhanced susceptibility to Ca²⁺-triggered arrhythmias, which implicates further, hitherto unknown mechanisms. Overall, our study unveils a new perspective of αARKct's molecular actions on cardiac performance and sheds new light on molecular therapeutic strategies that can improve βAR responsiveness in failing myocardium through Gαα-sequestration rather than facilitating cAMP signaling.

Non-standard Abbreviations and Acronyms

AC

adenylyl cyclase

AAV9

adeno-associated virus 9 Ad

Ad

adenovirus βAR β-adrenergic receptor

βARK

β-adrenergic receptor kinase

CaMKII

Calmodulin dependent kinase II

ERK1/2

Extracellular regulated kinase 1/2

FSK

forskolin

FC

failing cardiomyocyte

GRK2

G-protein coupled receptor kinase 2

GFP

green fluorescent protein

HF

heart failure Ica L-type Ca²⁺ channel current

LCC

L-type Ca²⁺ channel

NC

non-failing cardiomyocyte

MOI

multiplicity of infection

PKA

cAMP-dependent protein kinase

PKC

protein kinase C

PLB

phospholamban

RyR2

ryanodine receptor 2

SR

sarcoplasmic reticulum

TnI

troponin I

LV

left ventricle