Myocardial overexpression of the C-terminus of β-adrenergic receptor kinase (βARKct) has been shown to result in a positive inotropic effect or an improvement of survival in heart failure. However, it is not clear whether this beneficial effect is mainly because of dominant-negative inhibition of βARK1, and a consecutive resensitization of β-adrenergic receptors (βAR), or rather due to inhibition of other Gβγ-mediated effects. In this study, we tested whether overexpression of N-terminally truncated phosducin (nt-del-phosducin), another Gβγ-binding protein that does not resensitize βARs owing to simultaneous inhibition of GDP release from Gα subunits, shows the same effects as βARKct. Adenoviral gene transfer was used to express nt-del-phosducin and βARKct in isolated ventricular cardiomyocytes and in myocardium of rabbits, which suffered from heart failure because of rapid ventricular pacing. βAR-stimulated cAMP formation was increased by βARKct, but not by nt-del-phosducin, whereas both proteins inhibited Gβγ-mediated effects. Both transgenes also increased contractility of normal and failing isolated cardiomyocytes and improved contractility in rabbits with heart failure after gene transfer in vivo. In conclusion, overexpression of nt-del-phosducin enhances the contractility of cardiomyocytes to the same extent as βARKct, suggesting that the effects of βARKct might be owing to inhibition of Gβγ rather than to βAR resensitization.
Heart failure is characterized by a blunted responsiveness of the β-adrenergic receptor (βAR) system, which is both because of a downregulation of βAR and an increased activity of βAR kinase (βARK1; synonymous: G protein-coupled receptor kinase 2).1 βARK1 phosphorylates agonist-occupied βAR and triggers the binding of β-arrestins and the process of homologous desensitization and internalization.2 During this process, the enzyme is specifically targeted to the membrane by a direct interaction between dissociated βγ subunits of activated G proteins (Gβγ) and residues within the carboxyl terminus of βARK1.3,4 This C-terminal part of the protein has been termed ‘βARKct’. Overexpression of βARKct in several models of heart failure has been shown to result in improved hemodynamic conditions, for example, after somatic gene transfer into rabbits with heart failure because of rapid pacing5 or myocardial infarction.6,7 Also, crossbreeding of cardiac-specific βARKct-transgenic mice with mouse models of heart failure due to ablation of the muscle LIM protein (MLP (−/−) mice),8 due to a myosin heavy chain mutant9 or due to overexpression of calsequestrin10 prevented the development of heart failure or improved survival. This effect was additive to the treatment of heart failure with β-blockers.10
However, it is not clear whether this beneficial effect of βARKct is mainly because of dominant-negative inhibition of βARK1, and a consecutive resensitization of βAR, or rather due to inhibition of other Gβγ-mediated effects. Gβγ proteins exert a modulatory effect on phospholipase C, potassium channels, and other second messenger systems11 and Gβγ-mediated effects are inhibited by βARKct.12 Therefore, we compared overexpression of βARKct with that of phosducin, another Gβγ-binding protein, which also translocates upon receptor activation from the cytosol to the cell membrane, binds Gβγ subunits with high affinity13 and inhibits Gβγ-mediated signaling.14 Similar to βARKct, phosducin can compete with βARK for Gβγ subunits and thereby impair βARK-mediated phosphorylation of receptors.15 The Gβγ-binding affinities of βARKct (∼300 nmol/l), phosducin and an N-terminally truncated variant (‘nt-del-phosducin’ nt-del-phd) (15 and 33 nmol/l) are comparable.15,16,17 In contrast to βARKct, however, phosducin and nt-del-phosducin do not cause overall resensitization of the β-adrenergic receptor system owing to simultaneous inhibition of GDP release from Gα subunits.11,14,18 The structure of phosducin and nt-del-phd are compared schematically in Figure 1. Preliminary experiments using various variants of phosducin had shown that truncation of N-terminal sequences resulted in a phosducin with more reproducible Gβγ-binding effects in intact cells (Humrich et al., manuscript in preparation). As we had shown before that an N-terminally truncated phosducin (amino acids 64–246) has a Gβγ-binding affinity not different from full-length phosducin,17 we used an N-terminally truncated variant of phosducin starting at the second ATG codon at 156 base pairs (‘nt-del-phosducin’) for the present study. Overexpression of nt-del-phosducin and βARKct were compared for their effects on contractility and the development of heart failure.
In vivo adenoviral delivery of transgene to failing hearts
Overexpression of all transgenes was investigated by studying the coexpression of GFP in the hearts after in vivo gene transfer, since all transgenes were expressed together with GFP. For this purpose, fresh sections were cut from all hearts. Figure 2 shows an example of a macroscopic slice of a rabbit heart infected with Ad-nt-del-phosducin-GFP, in which GFP coexpression occurred throughout the left ventricle. In a series of reproducible experiments, the amount of green fluorescent cells after gene transfer in vivo was determined by isolating cardiomyocytes and determining GFP fluorescence by FACS. Similar to the data shown before,19 32±4% of cardiomyocytes showed green fluorescence 1 week after gene transfer.
Transgene expression assessed by Western blotting
Western blotting documented that expression of nt-del-phosducin was detectable with a specific antibody, the transgene being 6 kDa smaller than full-length phosducin (Figure 3a). Also βARKct was expressed at the expected size of 27 kDa in cardiomyocytes (Figure 3b).
Intracellular cAMP formation in cardiomyocytes
In order to investigate the effects of nt-del-phosducin or βARKct on G-protein-mediated signalling, we measured cAMP accumulation in isolated cardiomyocytes with the alpha screen assay after ex vivo gene transfer. The β-adrenergic receptor agonist isoproterenol increased intracellular cAMP content in all groups in both healthy (Figure 4a) and failing cardiomyocytes (Figure 4b). In βARKct-expressing healthy cardiomyocytes, the increase in intracellular cAMP formation in response to 1 μmol/l isoproterenol was more than two-fold increased compared to Ad-GFP-infected control cells or to Ad-nt-del-phosducin-infected cells. The experiments for healthy cardiomyocytes were also reproduced by using the ELISA system to detect cAMP, leading to similar results (not shown). Also in failing cardiomyocytes, βARKct resulted in significantly increased cAMP formation (Figure 4b).
Inhibition of Gβγ-mediated effects
To investigate the functional consequences of Gβγ inhibition, we studied the effects of both transgenes on Gβγ signalling stimulated by two different agonists. Figure 5a shows that intracellular IP3 formation in response to both bradykinin and acetylcholine was equally reduced in the presence of βARKct and nt-del-phosducin, indicating that Gβγ-mediated effects were effectively inhibited by both transgenes. Figure 5b demonstrates that similar effects were observed in failing cells.
Effect on β-adrenergic receptor density
In βARKct-expressing or nt-del-phosducin-expressing cardiomyocytes, no significant changes of β-adrenergic receptor density occurred compared to Ad-GFP-infected control cells (14±4, 13.5±3 and 13.8±2 fmol/mg protein, respectively).
In order to investigate the effects of nt-del-phosducin or βARKct on cardiomyocyte contractility, we measured fractional shortening and velocity of shortening in single, isolated cardiomyocytes from both failing and normal hearts after ex vivo gene transfer. Compared to Ad-GFP-infected control cells, basal and maximal contractility in response to isoproterenol were markedly increased in nt-del-phosducin- and in βARKct-expressing cardiomyocytes (Figure 6a). Overexpression of nt-del-phosducin and βARKct also enhanced the maximal contraction amplitude of failing cardiomyocytes in response to isoproterenol (Figure 6b). Similar results were obtained for shortening velocity (not shown). The concentration–response curves of both nt-del-phosducin- or βARKct-expressing normal and failing cells were significantly shifted to the left. In all batches of virus-infected cells, GFP-negative cardiomyocytes (which did not express the transgenes) were also tested for contractility. These cells did not show any difference compared to noninfected cells, thus demonstrating the comparability of preparation quality (not shown).
Rapid pacing was used to induce heart failure in all animals. The average +dp/dtmax-value in failing hearts was 2200±320 mmHg/s (versus 3200±390 mmHg/s in healthy controls; P<0.05), and left ventricular end-diastolic pressure (LVEDP) increased from 3.6±0.4 to 13±3.4 mmHg during rapid pacing (P<0.05).
Improvement of LV dysfunction in pacing-induced heart failure
Heart failure was induced in rabbits by 1 week of rapid pacing at 320 bpm and its extent was quantified by echocardiography prior to direct gene delivery of Ad-GFP, Ad-βARKct-GFP or Ad-nt-del-phosducin-GFP. The rabbits were then paced one additional week at 360 bpm before final echocardiographic and hemodynamic measurements were performed. Figure 7 shows that in the βARKct- and nt-del-phosducin-expressing groups, the first derivatives of LV pressure (dp/dt max) in response to isoproterenol were significantly higher than in the Ad-GFP-infected control group. This was also true for the increases in systolic LV pressure. Also, LV filling pressures at basal conditions tended to be lower in animals whose hearts expressed βARKct (8.3±3.7 mmHg) or nt-del-phosducin (10.5±3.7 mmHg) than in the group expressing GFP only (13±3.4 mmHg), although this trend did not reach statistical significance.
LV fractional shortening (FS) was assessed by serial echocardiography, and the ratio of FS at the end of the experiment versus before gene transfer was determined. In the βARKct- and nt-del-phosducin-expressing groups, FS did not change during the second week of rapid pacing, whereas in the Ad-GFP-infected group, a clear decrease in FS occurred (Figure 8a). Figure 8b shows that also the left ventricular end-diastolic dimensions were further dilated in the Ad-GFP-infected group, whereas the LV-EDD ratios at the end of the experiment versus before gene transfer were only little changed in the nt-del-phd- or βARKct-expressing groups, as determined by echocardiographic M mode measurements.
The present study shows that overexpression of an N-terminally truncated phosducin (‘nt-del-phosducin’) results in a clear positive inotropic effect in both normal and failing cardiomyocytes after gene transfer ex vivo. The effects are comparable to overexpression of βARKct, although nt-del-phosducin does not resensitize β-receptor-dependent cAMP formation. Moreover, cardiac function was clearly improved in rabbits with heart failure after in vivo gene transfer of both transgenes. These results suggest that nt-del-phosducin and possibly also βARKct exert their positive effects by the inhibition of those Gβγ-mediated pathways, that are not linked to a resensitization of β-adrenergic receptors.
To determine the direct functional significance of nt-del-phosducin and βARKct overexpression on myocardial performance in the absence of tonic sympathoadrenal neural activation and mechanical loading, we measured the contractility of left ventricular myocytes isolated from normal or failing hearts after ex vivo gene transfer. We detected a clear increase in isoproterenol-dependent contractility of isolated cardiomyocytes in the presence of βARKct, corroborating the findings in isolated cardiomyocytes of βARKct-transgenic mice20 and the increased cAMP formation measured after gene transfer into failing cardiomyocytes.5
Also the finding of improved contractility after gene transfer of Ad-βARKct-GFP in vivo corresponds well with the improvement of LV function and outcome measured in similar disease models5 or in cardiac-specific βARKct-transgenic mice.5,9,8 To a similar extent as βARKct, also overexpression of nt-del-phosducin enhanced basal contraction and maximal contractility of both normal or failing cardiomyocytes. Moreover, a clear leftward shift of the concentration–contractility curve occurred (Figure 6). We also found that overexpression of both, βARKct and nt-del-phosducin, increased contractility and prevented further deterioration of heart failure after in vivo gene transfer (Figure 7 and Figure 8).
In contrast to their concomitant effects on contractility, the two Gβγ-binding proteins showed a fundamentally different effect on intracellular cAMP accumulation in both, healthy and failing cardiomyocytes. Isoproterenol-dependent cAMP formation was more than two-fold increased in βARKct-expressing healthy cardiomyocytes, whereas nt-del-phosducin-expressing cells, except for the stimulation with 100 nmol/l isoproterenol, did not differ from controls (Figure 4a). In failing cardiomyocytes, βARKct overexpression resulted in clearly increased basal and isoproterenol-dependent cAMP formation, whereas nt-del-phosducin overexpression only led to a trend towards increased cAMP formation at low agonist concentrations, which did not, however, reach statistical significance.
The different effects on cAMP formation may be because of the capacity of phosducin, but not of βARKct, to bind to Gsα13 and inhibit GDP release from Gsα, the rate-limiting step of G protein activation, or to its higher binding affinity for Gβγ compared to βARKct.15,17 An alternative explanation would be a specific βARK-directed dominant-negative effect of βARKct, but not of phosducin. Structurally different domains of phosducin seem to mediate binding to Gβγ and Gsα. The augmentation in contractility induced by nt-del-phosducin is apparently independent of an increase in intracellular cAMP accumulation and therefore most probably unrelated to the resensitization of the receptors.
Most probably, the beneficial effects of nt-del-phosducin on cardiac contractility in heart failure depend on its capacity to sequester Gβγ and, consequently, to inhibit Gβγ-dependent pathways such as phospholipase C-β and phosphatidyltidylinositol (IP3)21 or mitogen-activated protein (MAP) kinase. MAP and PI3-kinase activities have recently been shown to be inhibited by activated Gβγ-BARK1.22,23
As phosphorylation of phosducin has been shown to abolish the effects of phosducin on G protein function,14,15,24,25 the deletion of the phosphorylation sites at amino acids 6, 36 and 54 by N-terminal truncation was introduced in the mutant used in the present study. This modification resulted in a phosducin mutant producing consistent effects in intact cells.
In conclusion, overexpression of nt-del-phosducin enhances the contractility of cardiomyocytes to the same extent as βARKct, suggesting that the effects of βARKct are probably owing to inhibition of Gβγ rather than to βAR resensitization.
Construction and purification of recombinant adenovirus
An N-terminally truncated phosducin (nt-del-phosducin) was cloned by using a PCR-based strategy with a forward primer encompassing the second ATG codon (pos. 157–159) and a reverse primer encompassing the last 20 bases of the coding sequence of bovine phosducin. Recombinant (E1/E3-deficient) flag-tagged adenoviruses for βARKct (Ad-βARKct-GFP) and for nt-del-phosducin (Ad-nt-del-phosducin-GFP) were generated, expressing the transgenes and green fluorescence protein (GFP) under control of two independent CMV promotors.26 As a control, Ad-GFP without further transgenes was used. Large virus stocks were prepared as described previously.27 Adenoviral titers were determined using plaque titration and GFP expression titration in non-E1-expressing cells.
Model of heart failure
Pacemakers from Vitatron, Düsseldorf, Germany, were implanted into New Zealand White rabbits (weight 3.6±0.3 kg; from Harlan, Munich, Germany). After 2 days, rapid pacing was initiated at 320 bpm. Under this protocol, a tachycardia-induced heart failure (HF) develops reproducibly over 1 week. Pacing was then continued at 360 bpm, which predictably led to a further deterioration of heart failure. The project was approved by the institutional ethics review board.
Adenoviral gene transfer to rabbit myocardium
After the first week of rapid pacing, all rabbits received catheter-based adenoviral gene transfer (5 × 109 PFU) to the myocardium as described before.27 For the intervention, the rabbits were anesthetized with fentanyl and propofol. The efficacy of gene transfer was assessed in all hearts after the end of the experiments by investigating transverse freeze-cut sections for expression of GFP by fluorescence microscopy under UV long-wave illumination. Morphological changes were assessed after fixation with 4% paraformaldehyde. Gene transfer led to reproducible transgene expression in ∼30% of cardiomyocytes as had been shown before by isolating cardiomyocytes after in vivo gene transfer.19
Measurement of cardiac function and LV hemodynamics
Left ventricular contractility and dimensions were measured by echocardiography in all rabbits at baseline, immediately prior to gene transfer after 1 week of rapid pacing, and 1 week postgene transfer after a total of 2 weeks of pacing. Echocardiography was performed using a 7.5 MHz probe (Hewlett Packard Sonos 1000) after the induction of anesthesia as previously described. For echocardiography, a 7.5 MHz probe was fixed on a tripod. Standard sections were recorded, which were well reproducible. Both M mode and B mode measurements were carried out. ECG was monitored continuously.
For final LV catheterization, a 3F micromanometer-tipped catheter (Millar) connected to a differentiating device was placed in the left ventricle via a sheath placed in the carotid artery. After definition of basal contractility and left ventricular pressure, 200 μl of NaCl (0.9%) was injected as a negative control. Isoproterenol was infused intravenously at increasing doses. After a 20 min equilibration period, tip catheter measurements were carried out.
Preparation and culture of adult ventricular cardiomyocytes and adenovirus infection
Single calcium-tolerant ventricular cardiomyocytes were isolated from healthy or failing White New Zealand rabbit hearts as described before.28 Briefly, the hearts were perfused and digested with collagenase. The isolated cardiomyocytes were cultured in modified M199 on laminin-precoated dishes (5–10 μg/cm2) at a density of 1.5 × 105 cells per cm2 (at 5% CO2 and 37°C). For contraction experiments, the cells were infected with adenovirus (multiplicity of infection (moi) 1 PFU/cell) 5 h after plating. In total, 50–60% of the infected cardiomyocytes expressed the transgene at this titer. For cAMP and IP3 assays, an moi of 10 PFU/cell was used, resulting in >95% expression efficacy.
Western blot of infected cardiomyocytes
Cardiomyocytes were harvested 48 h after adenoviral infection. The cells were homogenized and cytosolic extracts were then used for Western blotting by using an antibody raised against βARKct (GRK-2 antibody, Santa Cruz Biotech, Santa Cruz, CA, USA), or a polyclonal rabbit antiserum raised against recombinant bovine phosducin purified from E. coli. Horseradish peroxidase-coupled goat anti-rabbit antibodies by Dianova, Germany, were used as second antibodies.
Shortening measurements in isolated cardiomyocytes
Contractility of infected cardiomyocytes was measured by an electro-optical monitoring system connected to on-line digitalized assessment of amplitude and velocity of shortening and of relaxation as described before.28 After the contraction amplitude reached stability, increasing concentrations of isoproterenol were applied.
Determination of intracellular cAMP concentrations
Healthy or failing cardiomyocytes were investigated 48 h after adenoviral infection. The cells were harvested and stimulated with increasing concentrations of isoproterenol for 20 min. The reaction was stopped by adding 100 μl of a 20 mmol/l phosphate-EDTA buffer (pH 7.0) in the presence of IBMX (1 mmol/l) to inhibit cAMP degradation, followed by boiling at 100°C for 7 min. This suspension was centrifuged, and the supernatant was used for ELISA assays with cAMP-specific antibodies (Stratagene, Heidelberg, Germany, cat. no. 200020), following the manufacturer's instructions. Alternatively, cells were lysed with a Tween 20-containing buffer, and a PerkinElmer alpha screen cAMP kit using anti-cAMP-antibody-coupled microbeads (cat. no. 6760600) was used according to the manufacturer's instructions to determine cAMP accumulation directly.
Density of β-adrenergic receptors was determined in membranes from single cardiomyocytes, which were centrifuged and resuspended with 5 mmol/l Tris-HCl (pH 7.4) and 2 mmol/l EDTA. Radioligand binding was done with 3H-CGP12177 in concentrations ranging between 0.05 and 5 nmol/l in triplicates, with or without propranolol to determine nonspecific binding. The reaction was terminated by rapid filtration through Whatman GF/C filters, and bound radioactivity was measured in a β-counter.
Determination of phospholipase C activity
For IP3 assays, adenovirus-infected cardiomyocytes were stimulated with the respective agonists for 1 min, and the reaction was stopped by adding perchloric acid (4%) and scratching the cells off. They were centrifuged at 2000 g, then 10 μl of KOH (10 mol/l) was added. The pellet was resuspended and centrifuged again, and the protein content of each sample was determined by the method of Bradford.29 The supernatant was used for an assay kit with 3H-inositol-(1,4,5)-trisphosphate and a binding protein (Amersham, cat. no. TRK 1000) to measure IP3 concentration, following the manufacturer's instructions. In failing cardiomyocytes, a PerkinElmer alpha screen IP3 measurement kit using IP3-binding protein coupled to microbeads (cat. No. 6760621) and a GST detection kit (cat. No. 6760603x) was used to determine IP3 accumulation directly.
All data are expressed as means±standard error of the means (s.e.m.). For statistical analysis, analysis of variance (ANOVA) for repeated measurements followed by Scheffe's testing, or, where appropriate, Student's t-test with two-tailed distribution, was used. For EC50-values, simple one-way ANOVA followed by LSD testing was used. For all analyses, a value of P<0.05 was considered to be statistically significant.
About this article
Nature Reviews Cardiology (2009)