1. Cell Biology
Download icon

β3-Adrenoceptor redistribution impairs NO/cGMP/PDE2 signalling in failing cardiomyocytes

  1. Sophie Schobesberger
  2. Peter T Wright
  3. Claire Poulet
  4. Jose L Sanchez Alonso Mardones
  5. Catherine Mansfield
  6. Andreas Friebe
  7. Sian E Harding
  8. Jean-Luc Balligand
  9. Viacheslav O Nikolaev  Is a corresponding author
  10. Julia Gorelik  Is a corresponding author
  1. Myocardial Function, National Heart and Lung Institute, Imperial College London, ICTEM, Hammersmith Hospital, United Kingdom
  2. Institute of Experimental Cardiovascular Research, University Medical Center Hamburg-Eppendorf, German Center for Cardiovascular Research (DZHK) partner site Hamburg/Kiel/Lübeck, Germany
  3. Physiologisches Institut, University of Würzburg, Germany
  4. Pole of Pharmacology and Therapeutics (FATH), Institut de Recherche Expérimentale et Clinique (IREC), Université Catholique de Louvain (UCLouvain), Belgium
Research Article
  • Cited 0
  • Views 397
  • Annotations
Cite this article as: eLife 2020;9:e52221 doi: 10.7554/eLife.52221

Abstract

Cardiomyocyte β3-adrenoceptors (β3-ARs) coupled to soluble guanylyl cyclase (sGC)-dependent production of the second messenger 3’,5’-cyclic guanosine monophosphate (cGMP) have been shown to protect from heart failure. However, the exact localization of these receptors to fine membrane structures and subcellular compartmentation of β3-AR/cGMP signals underpinning this protection in health and disease remain elusive. Here, we used a Förster Resonance Energy Transfer (FRET)-based cGMP biosensor combined with scanning ion conductance microscopy (SICM) to show that functional β3-ARs are mostly confined to the T-tubules of healthy rat cardiomyocytes. Heart failure, induced via myocardial infarction, causes a decrease of the cGMP levels generated by these receptors and a change of subcellular cGMP compartmentation. Furthermore, attenuated cGMP signals led to impaired phosphodiesterase two dependent negative cGMP-to-cAMP cross-talk. In conclusion, topographic and functional reorganization of the β3-AR/cGMP signalosome happens in heart failure and should be considered when designing new therapies acting via this receptor.

Introduction

Over the last two decades, functional β3-adrenergic receptors (β3-ARs) have been found and studied in cardiomyocytes isolated from various species including humans and rodents (Gauthier et al., 1998; Mongillo et al., 2006; Hammond and Balligand, 2012). Depending on the cell type (cardiomyocytes vs adipocytes or atrial vs ventricular myocytes), β3-ARs have been reported to couple to both stimulatory (Gs) and inhibitory (Gi) proteins and to regulate cardiac contractility. In human and rodent ventricular myocardium, catecholamine binding to β3-ARs elicits negative inotropic and positive lusitropic effects by signalling via Gi and the second messenger 3’,5’-cyclic guanosine monophosphate (cGMP) (Gauthier et al., 1998; Mongillo et al., 2006). Unlike β1- and β2-AR, the β3-AR is resistant to agonist-induced desensitization, (Liggett et al., 1993; Nantel et al., 1993) and its expression is increased in heart failure as well as in sepsis and diabetic cardiomyopathy (Amour et al., 2007; Moniotte et al., 2007; Moniotte et al., 2001). It was hypothesised that β3-AR/cGMP pools can attenuate excessive cardiotoxic β1-AR/cAMP signalling, as well as pathological cardiac hypertrophy and remodelling which takes place in cardiomyocytes during the progression towards heart failure (Mongillo et al., 2006; Hammond and Balligand, 2012; Takimoto et al., 2005). Endothelial nitric oxide synthase (eNOS), has been detected in close proximity to β3-ARs in cardiomyocyte caveolae structures. The caveolae are believed to provide discrete signalling domains, necessary for the autonomic regulation of the heart (Feron et al., 1998). It has been shown indirectly that β3-AR/cGMP is most likely degraded by the phosphodiesterases 2 and 5 (Mongillo et al., 2006; Takimoto et al., 2005). Recently, overexpression of β3-AR in transgenic mice has been shown to protect the heart from catecholamine-induced hypertrophy and remodelling via an eNOS/soluble guanylyl cyclase (sGC)/cGMP-dependent signalling pathway. The same study showed localization of β3-ARs together with eNOS in caveolae-enriched membrane fractions, which had been separated via ultracentrifugation (Belge et al., 2014). Another mouse study identified the sGC subunit α1 as the facilitator of the NO dependent but Ca2+ independent effects of β3-AR using sGC α1 KO mice (Cawley et al., 2011). Despite its name the ‘soluble’ sGC has been shown to act in close association with β3-ARs and membrane located signalosomes (Mongillo et al., 2006; Feron et al., 1998). However, the exact localization of functional β3-ARs in adult cardiomyocytes and the spatio-temporal regulation of their cGMP signals as well as their potential interaction with cAMP signalling pathways have not been studied before.

In this study, we employ a highly sensitive Förster Resonance Energy Transfer (FRET)-based biosensor, red cGES-DE5, in combination with scanning ion conductance microscopy (SICM). We demonstrate that in healthy rat cardiomyocytes, functional β3-ARs are localized exclusively within the transverse (T)-tubules and stimulate a cGMP pool which is predominantly regulated by phosphodiesterases (PDEs) 2 and 5. Furthermore, by using the cAMP specific FRET-based biosensor Epac1-camps we show that β3-AR stimulation can decrease overall adenylate cyclase dependent cAMP levels in healthy cardiomyocytes by a PDE2-mediated cGMP-to-cAMP cross-talk. This cross-talk appears to be disrupted in heart failure, where β3-AR stimulation no longer has a significant effect on overall cAMP levels. In failing cells, β3-AR/cGMP signals decrease in the T-tubules. Heart failure leads to altered co-localization of sGC and caveolin-3, as shown via immunocytochemical staining. Together, these alterations result in the impairment of the β3-AR-dependent cGMP signalling pathway and of a PDE2-mediated β3-AR induced decrease of local cAMP.

Results

Echocardiography and biometric data show heart failure phenotype

To study β3-AR-dependent cGMP dynamics, we used ventricular cardiomyocytes isolated from healthy and failing rat hearts transduced with an adenovirus to express a highly sensitive cytosolic FRET biosensor red cGES-DE5. As a model of heart failure, we used rats which underwent left coronary artery ligation for 16 weeks (Lyon et al., 2009). Echocardiographic and biometric data from these animals are summarized in Figure 1. Data analysis showed typical clinical signs of heart failure, including a loss of pump function, left ventricular dilation and left ventricular wall thickening.

Histograms of echocardiography and biometric data in rat age matched control (AMC) hearts and hearts with myocardial infarction (MI).

(A) Ejection Fraction, (B) Heart weight (HW) corrected to tibia length (TL), (C) left ventricular diastolic internal dimension (LViDd), (D) left ventricular systolic internal dimension (LViDs), (E) end-diastolic volume, (F) end-systolic volume, (G) end-diastolic left ventricular posterior wall thickness (LVPWd), (H) end-systolic left ventricular posterior wall thickness (LVPWs). Statistical significance was analysed via two-sided T-test. ***p<0.001.

Isoproterenol (ISO) induces a β3-AR-dependent cGMP increase in adult rat cardiomyocytes

β-adrenergic stimulation (ISO, 100 nmol/L) of healthy control rat ventricular cardiomyocytes expressing the cGMP biosensor red cGES-DE5, led to the production of substantial amounts of cGMP (Figure 2A) in about 2/3 of all tested cells. In failing cardiomyocytes isolated from rats 16 weeks post-myocardial infarction, administration of the same saturating concentrations of ISO resulted in a significant two-fold reduction in the amount of detectable cGMP (Figure 2B, p=0.0465). Blocking β1- and β2-ARs (with 100 nmol/L CGP20712A and 50 nmol/L ICI118551, respectively) in control cells did not abolish this cGMP production (Figure 2C). The signal was however strongly and significantly inhibited in control cells by the application of the β3-AR antagonist SR59230A (Figure 2D, p=0.0316) or by the nitric oxide synthase (NOS) blocker, nitro-L-arginine methyl ester (L-NAME, Figure 2E, p=0.0217).

Measurements of β3-AR-dependent cGMP responses in adult cardiomyocytes.

Representative FRET tracings of a control (A) or a failing cardiomyocyte (B) treated with isoproterenol (100 nmol/L). FRET responses of control cardiomyocytes pre-treated for 5 min with either with the β1-AR and β2-AR inhibitors CGP20712A (100 nmol/L) and ICI118,551 (50 nmol/L) (C), β3-AR inhibitor SR59230A (100 nmol/L) (D) or for 10 min with the nitric oxide synthase blocker L-NAME (300 µmol/L) (E), before the application of isoproterenol (100 nmol/L). (F) Quantification of whole cell cGMP-FRET responses from protocols described in A-E). Error bars represent standard error of the mean. Numbers of cells/hearts are shown above the bars. Statistical significance was calculated via Mann Whitney U-test for independent treatments versus control followed by Bonferroni correction: *p<0.05; **p<0.01.

β3-AR/cGMP is preferentially controlled by PDE2 and PDE5

Next, we stimulated cells with ISO and then applied selective inhibitors of the various cGMP-degrading PDEs to investigate the regulation of β3-AR/cGMP dynamics. Following the application of selective PDE blockers we applied the non-selective PDE inhibitor IBMX. We found that β3-AR/cGMP levels are under the control of multiple PDEs. PDE1 inhibition has a minimal effect on β3-AR/cGMP production, whereas PDE2 and PDE5 represent the most prominent β3-AR/cGMP degrading families (Figure 3B,D). Both PDE2 and PDE5 contribute to more than a half of the overall PDE inhibitory response as determined by IBMX treatment (Figure 3B,D and E). Furthermore, we observed that in failing cells, the PDE2, PDE3 and PDE5 inhibitor effects, while not statistically significant, showed a tendency towards increasing (Figure 3B,D and E).

Investigation of phosphodiesterase regulation of β3-AR/cGMP in adult cardiomyocytes.

Representative FRET response curves of control (grey line) and failing (black line) cardiomyocytes following whole cell treatment with isoproterenol (100 nmol/L) followed by the PDE1 blocker vinpocetine (VINPO, 10 µmol/L) (A), the PDE2 inhibitor EHNA (10 µmol/L) (B), the PDE3 inhibitor cilostamide (CILO, 10 µmol/L) (C) and the PDE5 inhibitor tadalafil (TAD, 100 nmol /L) (D) followed by the non-specific PDE blocker IBMX (100 µmol/L). The scatter plot/histograms present whole cell cGMP-FRET responses evoked by PDE inhibition further to the isoproterenol responses in % from (A–D) (E) Error bars represent standard error of the mean. Numbers of cells/hearts are shown below the bars. Statistical significance was calculated via mixed ANOVA followed by χ2-test: No statistically significant differences between control and failing conditions for any PDE could be detected, only tendencies to increased responses for PDE2, PDE3 and PDE5 inhibitors.

Functional β3-ARs are localized in the T-tubules of healthy cells and migrate to the non-tubular sarcolemma in heart failure

Using SICM/FRET we were able to localize functional β3-ARs by measuring cGMP-FRET signals following local ligand application from the SICM nanopipette. This approach stimulates cardiomyocytes specifically within T-tubules or on the non-tubular sarcolemma. In healthy cardiomyocytes, we observed that functional β3-ARs reside mainly in the T-tubules with very few responses being detectable outside of T-tubules (Figure 4A–E), whereas in failing cells, the β3-ARs responses after localized stimulation can be detected in both tubulated and non-tubulated areas across the sarcolemma (Figure 4F–J).

Figure 4 with 1 supplement see all
Identification of β3-AR/cGMP signal localization using scanning ion conductance microscopy (SICM) combined with Förster Resonance Energy Transfer (FRET).

Representative SICM surface scan of a 10 × 10 µm area of a healthy (A) and a failing cardiomyocyte (F). White arrows indicate T-tubule or crest structures and a dotted white line indicates the areas selected for the topographical profiles presented in (B) and (G). Representative topographical profiles of healthy (B) and failing cardiomyocytes (G). Images present schematics of local β3-AR stimulation with Isoproterenol (50 µmol/L) either inside a T-tubule opening or on the cell crest via the SICM nanopipette. Representative FRET response curves during perfusion with the β1-AR blocker CGP20712A (100 nmol/L) and the β2-AR blocker ICI118551 (50 nmol/L) and local stimulation inside the T-tubule (C) and the crest of a control cardiomyocyte (D) or the T-tubule (H) and crest (I) of a failing cardiomyocyte. Scatter plots presenting the localised cGMP-FRET responses of control (E) and failing cardiomyocytes (J). Error bars represent standard error of the mean. Numbers of cells/hearts are shown above the bars. *p<0.05, n.s. – not significant by Mann-Whitney U test.

The increased activity of β3-ARs in non-tubulated surface areas in failing cells might be linked to a disrupted association of β3-AR with caveolar signalosomes. We investigated this hypothesis by using the cell-permeable peptide disruptor of caveolar signalling TAT-C3SD. The addition of this peptide leads to the dissociation of caveolar signalosomes by inhibiting signalling which is dependent upon the binding to the caveolin-3 specific scaffolding domain (C3SD) (Macdougall et al., 2012). In control cells, β3-AR-cGMP responses in the T-tubules are higher than in the non-tubulated cell surface areas, as is seen in Figure 4E. However, the T-tubular localization of the receptor can be abolished by treating cells with the TAT-C3SD peptide (Figure 4—figure supplement 1), so that the response level in the crest areas equals the response level of the T-tubules.

Heart failure disrupts β3-AR associated sGC localization in caveolin-rich membrane domains

To precisely investigate the localization of the components of the β3-AR signalosome in control cells and failing cardiomyocytes, we performed immunocytochemical staining of sGC subunits together with caveolin-3 or α-actinin. We detected partial co-localization of sGCα1 with caveolin-3 (Figure 5A,C,E) which was significantly decreased by about 20% in failing cardiomyocytes (Figure 5E, p=0.0014). Concomitantly, sGCα1 co-localization with the microfilament protein α-actinin (Figure 5B,D,F) had a tendency to increase in failing cardiomyocytes (Figure 5F, p=0.0244), whereas sGCβ1 subunit co-localization with caveolin-3 also decreased by about 16% (Figure 5—figure supplement 1, p=0.0044), indicating a redistribution of sGC away from caveolin-rich microdomains.

Figure 5 with 1 supplement see all
Investigation of sGC and caveolin-3 localization in control and failing cardiomyocytes.

Representative, confocal images of sGCα1 (A) and caveolin-3(B) in control and failing cardiomyocytes. Magnified representations of double staining of sGCα1(C) with caveolin-3 and of sGCα1 (D) with α-actinin. Quantification of sGC (E) and caveolin-3 and of sGC (F) and α-actinin co-localization in control and failing cells. Error bars represent standard error of the mean. Numbers of cells/hearts are shown above the bars. Statistical significance was analyzed via mixed ANOVA followed by χ2-test; **p<0.01, n.s. – not significant.

Heart failure impairs PDE2 mediated cGMP-to-cAMP cross-talk after β3-AR stimulation

To investigate whether the β3-AR signals we detected in the experiments above were able to influence cAMP signalling in cardiomyocytes, we expressed the cAMP biosensor Epac1-camps in healthy and failing cardiomyocytes. To analyse the cGMP-to-cAMP cross-talk, we treated the cell with the adenylyl cyclase activator forskolin with and without β3-AR agonist CL316,243 (Figure 6A,B). In healthy cells, stimulation of β3-AR led to a significant reduction of approximately 10.3% of the forskolin stimulated cAMP production (Figure 6C, p=0.040). The PDE2 inhibitor BAY60-7550 used in this setting was able to abolish the observed β3-AR agonist effects on cAMP levels (Figure 6A,C, p=0.0458). We used forskolin to directly activate adenylate cyclases downstream of β-ARs to allow us a direct comparison between control and MI cells, which might have been complicated otherwise due to β1-AR receptor desensitisation. Nonetheless, our result suggests that β3-ARs acts via increased PDE2 activation to attenuate cAMP responses. In failing myocytes, the effect of CL316,243 on the cAMP response was no longer significant (Figure 6B,C), suggesting that the PDE2 mediated cGMP-to-cAMP crosstalk is disrupted by disease.

β3-AR signalling can affect cAMP levels via PDE2.

Representative FRET responses for control cardiomyocytes (A) or failing cardiomyocytes from myocardial infarction (MI) hearts (B) stimulated using forskolin (FOR, 10 µmol/L) applied with or without the β3-AR agonist CL316243 (1 µmol/L) and with or without pre-treatment with the PDE2 inhibitor BAY60-7550 (100 nmol/L). (C) Scatter plot/histogram presenting whole cell cAMP-FRET responses depicted as the percentage of the maximal possible cAMP FRET response (=Forskolin followed by IBMX). The measured Forskolin or IBMX responses were the respective maximal responses, equalling the lowest FRET ratio value, achieved after each stimulus. The effect of the β3-AR agonist on forskolin induced cAMP levels is no longer discernible after inhibition of PDE2 in control cardiomyocytes. This effect is no longer significant in failing cardiomyocytes. Error bars represent standard error mean. Statistical significance was calculated via mixed ANOVA followed by χ2-test* p<0.05.

Discussion

The pharmacological modulation of β3-AR for the treatment of cardiac hypertrophy and heart failure has recently emerged as a promising therapeutic route in translational research (Belge et al., 2014). Previous studies have suggested that the β3-AR are localized in caveolae, in close proximity to its signalling partners eNOS and sGC (Mongillo et al., 2006; Belge et al., 2014). However, the exact sub-membrane localization of β3-AR and the compartmentation of its signalling to cGMP were not well understood. Alterations in β3-AR signalling in disease states have been difficult to study due to the lack of appropriate imaging techniques and specific antibodies which work in situations of relatively low endogenous expression. In this work, we studied the exact submembrane localization of β3-AR and alterations to β3-AR/cGMP signalling in failing cells using new cutting-edge biophysical approaches such as SICM and FRET. We were able to directly visualize compartmentalized β3-AR/cGMP production in the cellular context of adult rat cardiomyocytes and its disruption in heart failure. Using FRET imaging in the presence of either β3-AR agonists or antagonists, we show that these cells can produce cGMP following direct agonist stimulation of β3-ARs (see Figure 2). In some, but not all cells, residual cGMP production was still detectable despite β3-AR and NOS inhibition. This is an observation which has previouslybeen reported in neonatal rat cardiomyocytes (Mongillo et al., 2006). β3-AR dependent cGMP pools are mainly formed in the T-tubules (see Figure 4C) due to the localization of β3-ARs in close proximity to the caveolar signalosome, which among other molecules is comprised of eNOS and sGC. To further substantiate the association of functional β3-AR to caveolae, we dissociated the signalosome of these lipid structures in healthy cardiomyocytes using a previously published peptide, which targets the caveolin-3 scaffolding domain (Macdougall et al., 2012). As a result, β3-AR-cGMP pools were induced outside of T-tubular domains (see Figure 4—figure supplement 1) which corroborates the importance of caveolar signalosomes for proper β3-AR regulation.

Using various family selective PDE inhibitors, we uncovered that β3-AR/cGMP signals are predominantly regulated by local pools of PDE2 and PDE5 (see Figure 3B,D and E). These findings are similar to what has been described for atrial natriuretic peptide-stimulated pools of submembrane cGMP in rat cardiomyocytes (Castro et al., 2006) and eNOS/sGC-dependent pools of cGMP in mouse cells (Takimoto et al., 2005). Further regulation via the cGMP specific and IBMX insensitive PDE9 cannot be completely excluded, although a recent study in mice suggests that PDE9 degrades cGMP pools generated downstream of natriuretic peptide receptors acting independently of NO (Lee et al., 2015). Interestingly, the overall β3-AR/cGMP response was significantly reduced in failing cells (Figure 2B,F) despite increased β3-AR expression reported in cardiac disease (Hammond and Balligand, 2012; Moniotte et al., 2001). This reduction in the signal could be in part be due to higher PDE activity on cGMP (Figure 3B,E) and an altered subcellular arrangement of sGC (Figure 5). sGC was found to redistribute away from caveolin-3 in heart failure, as demonstrated in this work by immunocytochemical double staining. We have observed a trend to an increased overlap between our sGCα and α-actinin staining in our confocal imaging, which could potentially represent an increased redistribution of sGCα to the areas of the Z-disc not directly associated with the T-tubules or caveolar signalosomes. Though the immunocytochemical method is limited in its spatial resolution and can therefore not resolve the caveolae structures themselves, it allows us to detect an alteration in sGC localization in heart failure, which could potentially be indicative of dysregulated caveolar signalosomes as reported previously in a pressure overload induced heart failure model using mice (Tsai et al., 2012). Decreased expression of eNOS in the caveolae, alongside increased expression of neuronal NOS (nNOS) at the sarcoplasmic reticulum has been reported in studies of tissue from humans with heart failure (Damy et al., 2004; Drexler et al., 1998). As eNOS function in cardiomyocytes has been shown to attenuate beta-adrenergic contractile responses (Farah et al., 2018), potential changes in local NOS activities for example due to decreased eNOS-caveolin-3 association (Feron et al., 1998) could further contribute to altered β3-AR/cGMP signal disruption. The slightly increased contributions of PDE2,3 and 5 to cGMP regulation which we observed in heart failure, are probably responsible for the overall reduction in cGMP levels, as can be seen when comparing total ISO plus IBMX responses in control and failing cells (Figures 2F and 3E). Increased expression and activity of the dual-specificity PDE2 has been shown in human heart failure (Mehel et al., 2013), as well as in a rodent model of aortic banding, where an increase in both cGMP and cAMP hydrolysis took place (Yanaka et al., 2003). At the same time as being degraded by PDE2, cGMP can also increase the effect of PDE2 on cAMP by binding to the PDE’s GAF-B domain and leading to a conformational change, thereby triggering the so-called cGMP-to-cAMP crosstalk. This crosstalk between the second messengers of which one (cAMP) has increasingly been associated with detrimental signalling pathways in the context of heart failure, which the other (cGMP) could potentially attenuate (Mongillo et al., 2006; Moniotte et al., 2001; Yanaka et al., 2003). Interestingly, we have also detected a decreased β3-AR induced attenuation of the forskolin induced cAMP response, due to PDE2 activity, which it was possible to abolished by blocking the PDE2 using a specific inhibitor (see Figures 6 and 7). This observed impairment of the cGMP/cAMP crosstalk further supports the hypothesis of dramatic spatial rearrangements in the β3-AR/sGC/PDE2 signalosome which impair cGMP dynamics and could potentially lead to depressed β3-AR effects on cardiac contractility demonstrated in similar animal models (Mongillo et al., 2006) and in human heart tissue samples (Moniotte et al., 2001). The observed response of β3-AR stimulation of about 10 percent on overall cAMP levels, as measured via the cytosolic FRET sensor Epac1-camps, could be of physiological relevance when brought into the context of lowering pathological cAMP signalling levels on a whole or in confined signalling compartments. We believe that altered compartmentation of subcellular cGMP may play a role in disease-associated changes in β-AR signalling. However, heart failure does not completely ablate β3-AR/cGMP responses, leaving room for a residual cardioprotective action of β3-ARs in the failing heart. This therapeutic potential is currently being addressed pharmacologically using the β3-AR agonist mirabegron in clinical studies to treat heart failure with preserved and reduced ejection fraction (Pouleur et al., 2018).

Schematic of β3-AR/cGMP signalling in healthy (left side) and failing (right side) cardiomyocytes.

In healthy cardiomyocytes, functional β3-ARs are associated with caveolar signalosomes and localized mostly in T-tubules. Via Gi/NOS/NO/sGC/cGMP signalling they can suppress strong cAMP responses by stimulating increased PDE2 dependent cAMP degradation through cGMP binding to GAF-B domain of PDE2. In heart failure, increased presence of β3-AR activity at non-tubular plasma membrane (Crest) and away from caveolin-3 associated membrane domains might disrupt receptor-associated cGMP signalosomes and lead to disrupted cGMP/cAMP-crosstalk.

In summary, our study reveals mechanisms of submembrane localization of cardiomyocyte β3-AR, which regulates the compartmentation of receptor coupling to cGMP production and disease-driven alterations in β3-AR/cGMP signalling. These data add insights to the growing body of data regarding the therapeutic implications for the potential treatment of heart failure by β3-AR agonists.

Materials and methods

Experimental reagents

Request a detailed protocol

M199 medium (Invitrogen UK, 11150), taurine (Biochemica, A1141), creatine monohydrate (Sigma Aldrich, C3630), penicillin/streptomycin (Merck, A2212), carnitine hydrochloride (Sigma Aldrich, C9500) BSA (Sigma Aldrich, A6003), laminin (Sigma Aldrich L2020), isoproterenol hydrochloride (Sigma Aldrich, I6504), ICI118551 (Tocris UK, 0821), CGP 20712A (Tocris UK, 1024), SR 59230A hydrochloride (Tocris UK, 1511), L-NAME hydrochloride (Tocris UK, 0665), Vinpocetine (Sigma Aldrich, V6383), EHNA hydrochloride (Sigma Aldrich, E114), Cilostamide (Tocris UK, 0915), Tadalafil (Santa Cruz USA, sc-208412), IBMX (Santa Cruz sc-201188), self-made rabbit sGCα and β subunit antibodies (specificity tested in KO animals; Friebe et al., 2018), mouse α-actinin (Sigma Aldrich, A7732), mouse Caveolin-3 (BD Transduction Laboratories, 610421, specificity tested in KO animals; Woodman et al., 2002), secondary Alexa Fluor antibodies 488 nm, 514 nm, 568 nm and 633 nm (Life Technologies), BSA (Fisher Scientific UK, BPE9704), fluorescence mounting medium (Vectashield Germany, H-1000), MaTek glass-bottom dishes (MaTek USA, P35G-1.5–10 C), TAT-scram and TAT-Cav3 peptides (a gift from Dr. Sarah Calaghan from Leeds, England).

Myocardial infarction (MI) model

Request a detailed protocol

All animal experiments were performed in the United Kingdom (UK) according to the standards for the care and use of animal subjects determined by the UK Home Office (ASPA1986 Amendments Regulations 2012) incorporating the EU directive 2010/63/EU. The Animal Welfare and Ethical Review Body Committee of Imperial College London approved all protocols.

The parts of the investigation (on isolated cardiomyocytes from healthy rats) which were performed in Germany, conformed to the guide for the care and use of laboratory animals published by the National Institutes of Health (Bethesda, Maryland; Publication No. 85–23, revised 2011, published by National Research Council, Washington, DC). The experimental procedures were in accordance with the German Law for the Protection of Animals and with the guidelines of the European Community.

The following procedure was exclusively performed at Imperial College London in the UK: Left descending coronary artery ligation was performed as described (Nikolaev et al., 2010). Rats were monitored by echocardiography in M-mode under anaesthesia (2% isoflurane). Animals with induced MI were sacrificed 16 weeks after MI for ventricular cardiomyocyte isolation via enzyme digestion of the Langendorff perfused heart as described (Lyon et al., 2009). Age matched animals served as control. Echocardiographic and biometric data are summarized in Figure 1. All animals were male.

Peptide dissociation of caveolae signalosome

Request a detailed protocol

For disruption of the caveolar signalosome, cardiomyocytes were treated with a cell-membrane penetrating TAT peptide targeting the Caveolin-3 scaffolding domain (C3SD) (sequence: YGRKKRRQRRRGGGGDGVWRVSYTTFTVSKYWCYR) or with a scrambled control peptide without any cellular targets in cardiomyocytes (sequence: YGRKKRRQRRRGGGGYWTVYTKVDFCGSRYVRTSW) as described previously (Macdougall et al., 2012). Cells were incubated with peptides by directly putting the peptides into the cell medium for 30 min at 37°C.

Drug concentrations

Request a detailed protocol

The concentrations for pharmacological agonists and antagonists for the β-AR subtypes and the PDE subtypes investigated were taken from previously published work, which established the drug affinity through competitive radioligand binding assays, FRET sensor dose-response levels and PDE activity assays (Nikolaev et al., 2010; Nikolaev et al., 2006; Hoffmann et al., 2004; Johnson et al., 2012).

Whole cell and SICM/FRET cGMP measurements

Request a detailed protocol

After isolation, adult rat ventricular myocytes were plated onto laminin-coated cover glasses or MatTek dishes and cultured in M199 media supplemented with creatine 5 mM, taurine 5 mM, carnitine 5 mM, bovine serum albumin 1%, ascorbate 1 mM and penicillin/streptomycin 1%, before being subjected to FRET and SICM/FRET measurements at room temperature as described (Nikolaev et al., 2010) 44–52 hr after transduction with adenovirus expressing the cGMP-FRET biosensor red cGES-DE5 (Götz et al., 2014) at a multiplicity of infection equal 300. The respective buffer, at pH = 7.4, in which the cells were imaged contained 144 mmol/L NaCl, 5.4 mmol/L KCl, 1 mmol/L MgCl2, 1 mmol/L CaCl2 and 10 mmol/L HEPES. Whole cell β3-AR/cGMP levels were measured by treating control and failing cardiomyocytes with 100 nmol/L isoproterenol. To determine the source of isoproterenol induced cGMP whole cell FRET measurements were performed by pre-blocking control cardiomyocytes either with 100 nmol/L of the β3-AR inhibitor SR 59230A for 5 min or with 100 nmol/L of the β1-AR blocker CGP 20712A and 50 nmol/L of the β2-AR blocker ICI 118,551 for 5 min, or with 300 µM Nω-Nitro-L-arginine methyl ester hydrochloride for 10 min, before applying 100 nmol/L of isoproterenol. For whole cell FRET measurements of phosphodiesterase dependent regulation of β3-AR/cGMP signals, cells were treated with 100 nmol/L isoproterenol followed by either 10 µmol/L of the specific phosphodiesterase blocker vinpocetine for PDE1, 10 µmol/L erythro-9-amino-β-hexylα-methyl-9H-purine-9-ethanol (EHNA) for PDE2, 10 µmol/L cilostamide for PDE3 or 100 nmol/L Tadalafil for PDE5, and 100 µmol/L of the unspecific PDE inhibitor 3-Isobutyl-1-methylxanthine (IBMX). To study β3-AR/cGMP localization on cardiomyocyte surface structures, cells were continuously superfused with FRET buffer containing 100 nmol/L of the β1-AR blocker CGP 20712A and 50 nmol/L of the β2-AR blocker ICI 118,551. SICM was used to scan and visualize cardiomyocyte T-tubule openings and crest surface structures. Next, the scanning nanopipette was lowered onto either T-tubule openings or crest structures and receptor ligand was locally applied from the nanopipette filled with 50 µmol/L Isoproterenol (ISO) and 50 µmol/L CGP 20712A and 25 µmol/L ICI 118,551 by switching SICM pipette holding potential from −200 to +500 mV as described (Schobesberger et al., 2016).

FRET-based cAMP measurements

Request a detailed protocol

Freshly isolated cardiomyocytes were transduced with Epac1-camps (Nikolaev et al., 2004) adenovirus virus for 44–52 hr and exposed to whole cell FRET measurements for β3-AR effects on cAMP levels. Cells were kept in FRET buffer containing 50 nmol/L β2AR inhibitor ICI 118,551 before treating them with 100 nmol/L of the β3-AR agonist CL 316,243 plus 50 nmol/L ICI 118,551 or just adding FRET buffer with 50 nmol/L ICI 118,551 followed by addition of 10 µmol/L Forskolin and consequently 100 µmol/L IBMX into the solution. Additional cells were treated using the same protocol but with or without 100 nmol/L of the PDE2 inhibitor BAY 60–7550 in the cell bath.

Confocal imaging

Request a detailed protocol

Freshly isolated cardiomyocytes were plated onto laminin coated coverslips and fixed in ice cold methanol for 10 min. Next, they were blocked with PBS based buffer containing 0.3% Triton X-100% and 5% fetal calf serum. Cells were then double stained overnight, at 4°C with primary antibodies, diluted in PBS containing 0.3% Triton and 1% BSA, for example against sGCα1 together with α-actinin or Caveolin-3. Cells were then washed three times for 5 min with PBS before they were exposed to secondary antibodies diluted in PBS containing 0.3% Triton and 1% BSA for 1 hr. Finally, cells were washed again in PBS three times for 5 min before being mounted onto cover slides with mounting medium. Imaging was performed using an inverted Zeiss LSM-800 laser scanning microscope equipped with a 40x oil immersion objective and controlled by ZEN imaging software which was used for the colocalization analysis and detection of the respective Pearson’s correlation coefficients.

Statistics

Statistical differences were analysed using OriginPro 8.6 and GraphPad Prism 7 software. Normal data distribution was determined by Kolmogorov-Smirnov test. For the comparison of two independent groups with skewed distribution Mann-Whitney U test was used. For the comparison of normally distributed data, a two-tailed T-test (for comparing morphometric and echocardiographic data from independent animals, Figure 1) or a mixed ANOVA followed by χ2-test (when data from several cells from multiple individual animals were compared) were applied. Differences were considered significant at p-values below 0.05. All data are presented as means ± s.e.m.

References

  1. 1
  2. 2
  3. 3
  4. 4
  5. 5
  6. 6
  7. 7
  8. 8
  9. 9
  10. 10
  11. 11
  12. 12
  13. 13
  14. 14
  15. 15
  16. 16
  17. 17
  18. 18
  19. 19
  20. 20
  21. 21
  22. 22
  23. 23
    The human beta 3-adrenergic receptor is resistant to short term agonist-promoted desensitization
    1. F Nantel
    2. H Bonin
    3. LJ Emorine
    4. V Zilberfarb
    5. AD Strosberg
    6. M Bouvier
    7. S Marullo
    (1993)
    Molecular Pharmacology 43:548–555.
  24. 24
  25. 25
  26. 26
  27. 27
  28. 28
  29. 29
  30. 30
  31. 31
  32. 32

Decision letter

  1. Nir Ben-Tal
    Reviewing Editor; Tel Aviv University, Israel
  2. Anna Akhmanova
    Senior Editor; Utrecht University, Netherlands
  3. Marc Freichel
    Reviewer; University of Heidelberg, Germany

In the interests of transparency, eLife publishes the most substantive revision requests and the accompanying author responses.

Acceptance summary:

Several complementary biophysical methods are used to explore β3-Adrenoceptor localization in normal and impaired cells. The results show that redistribution of the receptor impairs NO/cGMP/PDE2 signaling in failing cardiomyocytes. This is interesting and timely because of therapeutic implications when targeting the receptor.

Decision letter after peer review:

Thank you for submitting your article "β3-Adrenoceptor redistribution impairs NO/cGMP/PDE2 signaling in failing cardiomyocytes" for consideration by eLife. Your article has been reviewed by three peer reviewers, and the evaluation has been overseen by a Reviewing Editor and Anna Akhmanova as the Senior Editor. The following individual involved in review of your submission has agreed to reveal their identity: Marc Freichel (Reviewer #3).

The reviewers have discussed the reviews with one another and the Reviewing Editor has drafted this decision to help you prepare a revised submission.

Summary:

Schobesberger et al. analyze the localization of β3-adrenergic receptor and associated signaling molecules in signaling domains of adult rat cardiomyocytes and its redistribution in a model of myocardial ischemia. They use FRET-based cGMP and cAMP sensors and scanning ion conductance (SICM) and confocal microscopy. The authors show Isoproterenol-evoked cGMP generation in control cardiomyocytes, its contribution by β3-adrenoceptors using β1 and β2 antagonists and its dependence on NO generation. The authors can demonstrate that ISO-evoked GMP production is reduced in failing cardiomyocytes. Using a large set of phosphodiesterase inhibitors, they elaborate the contribution of PDE2 and PDE5 isoforms for ISO-evoked cGMP generation. Combined FRET and SICM surface imaging reveals that β3-adrenoceptor-mediated cGMP generation occurs predominantly in the T tubular cleft in healthy myocytes, and that in failing myocytes the cGMP response is similar in T tubules and at the crest. Modulation of caveolin binding similarly equalizes T-tubule and crest signalling in healthy myocytes. Using confocal microscopy, the authors then show data suggesting that the β3-adrenergic receptor-associated signaling molecules sGCα1 and sGβ1 dissociate from caveolin 3-containing domains in failing cardiomyocytes. In a final set of experiments the authors demonstrate that β3-adrenoceptor activation attenuates cAMP generation (also measured by a FRET biosensor) in myocytes from healthy animals but that this control is lost in failing cells.

Opinion:

The study provides some interesting new data and is relevant because unlike the β1- and β2-ARs, β3-ARs are not desensitized in HF, and their activator is currently undergoing a Phase IIb clinical trial for treating heart failure. The combined SICM and FRET methods are unique to this investigative group and are valuable for the question at hand. There are a number of intriguing results presented, and the presentation is generally quite clear. However, the conclusions of the study are somewhat overstated, based on the data presented (and the way these data are analysed). This must be addressed.

Essential revisions:

1) First and foremost, recent work of PC Simpson showed clearly that β3-AR does not exist in murine cardiac myocytes. The authors should examine their data in view of these observations or argue against.

2) Evidence that the myocytes are from a state of heart failure should be included in the Results (echo data).

3) The choice of drug concentrations (β1-AR, β2-AR antagonists, PDE inhibitors) must be justified – selectivity of drugs for the targets is key to the interpretation of most data in this manuscript.

4) There are issues with use of statistical tests. MW or t-tests are used to compare multiple groups (Figure 1, Figure 4—figure supplement 1, Figure 5). This should be re-done with ANOVA or non-parametric equivalent. Please notice also other comments about statistical analysis of the data.

5) Interpretation of t-tubule/crest responses (Figure 3 and Figure 4—figure supplement 1) does not appear justified by the data. In Figure 3, comparison of E and J suggest that T-tubule responses are lost but not that they are redistributed to other areas of the sarcolemma. By contrast, comparison of the impact of the C3SD peptide and its scrambled control in healthy cells suggests a gain of crest signalling, rather than a loss of t-tubular signals.

6) Care must be taken with interpretation of work using the C3SD peptide. The sequence replicates the so-called caveolin scaffolding domain (CSD) which for many years was considered to mediate the regulatory role of caveolins through interaction with a complementary caveolin binding motif (CBM) in target proteins. This central tenet has been questioned within the last decade by work from the Parton and Dart laboratories which show that most CBM in caveolin's 'binding partners' are inaccessible for interaction with the CSD. This does not necessarily weaken the evidence for a role for CSD, which may exert effects via alternative mechanisms. However, we are not aware of any evidence that the C3SD peptide leads to 'deletion of intact caveolae'. The cartoons (Figure 4—figure supplement 1C, D) imply that Cav3 (presumably the red line, although not defined in legend) is a caveolar neck protein, which is not accurate. Furthermore, these cartoons suggest that with the C3SD peptide, the t-tubular cGMP signal is reduced which is not borne out by the data in part A of this figure. The legend states 'in cardiomyocytes treated with..C3SD..the cGMP signal in…crest areas is even stronger than when stimulated in t-tubules' but the comparison between these 2 datasets is not made (also see 3 above).

7) It is important to acknowledge the limitation of confocal imaging to provide robust support for the location of proteins in caveolae (lateral resolution ≈200 nm, caveolae 50-100 nm). Could more meaningful information be gained from confocal images (Figure 4) by a separation of interior (t-tubular) and surface (crest) staining?

8) The increased co-localisation of sGC and α-actinin is not explained (Figure 4F) – this seems to suggest increased sCG at z discs, i.e., t-tubules in failing cells which is not consistent with the overall interpretation of data.

9) Figure 5 – the magnitude of the impact of β3-AR stimulation on cAMP production (indexed with the Epac-based sensor) is small even in control cells. If this significant difference is retained when the data are analysed using ANOVA, the authors should comment on the relevance of such a small change. Perhaps measurement of cAMP signals in different subcellular compartments might reveal profound differences.

10) No direct evidence is presented here that β3-AR are caveolae-based (presumably because of lack of an appropriate antibody) yet some unsubstantiated statements are made e.g. 'β3 -AR dependent cAMP pools are formed in caveolae'. Some evidence (referenced in the manuscript) supports the location of some sGC subunits in buoyant caveolae-containing fractions, and the evidence for caveolar location of eNOS is robust. However, comprehensive evidence to support statements regarding the location of the β3 signalosome in caveolae is lacking.

11) Figure 6. There seems to be some lack of accuracy in depicting current (and past) data from this group. Why are β1-AR only in crest in healthy myocytes (vs. Nikolaev et al. 2010 Science). Are the PDE2/5 symbols misplaced at the bottom of panel A (surely these degrade cGMP in control cells as shown in Figure 2). The data in Figure 5C suggest that the normal β3-AR dependent reduction in cAMP seen in control cells is absent in failing cells (i.e. cAMP lower in control cells in the presence of β3-AR stimulation). Yet comparison of panel A and B in Figure 6 suggests that cAMP is lower in failing cells in the presence of β3-stimulation. Even given an increase in PDE2 expression/activity in failing cells (not shown explicitly) which could increase degradation of cAMP, how can the data in Figure 5 be reconciled with the cartoons in Figure 6?

12) It appears that the differently targeted β3-AR stimuli applied to T-tubules vs. Crest membrane regions were performed in different cells, with comparisons between them analyzed with non-paired statistics. This is not explicitly stated but is apparent by the statistical test used (Mann Whitney U). This seems problematic given that the level of FRET response varies from cell to cell, and even accepting IBMX-based normalization between cells, the statistical power to identify quantitative changes would be challenging. If getting both stimuli in the same cell is impossible and unpaired analysis is required, then the ability to properly normalize between cells is critical. The concern with it is raise by the key data shown in Figure 3. While the extent to which β3-AR-mediates cGMP at the Crest appears similar between healthy and failing cardiomyocytes, the FRET response at T-tubule declines in the latter group. If this was due to a redistribution of β3-AR to the Crest region, would you not expect a rise in the% FRET response there? Is it possible the results can be explained by a decline in β3-R levels or signal-coupling in the T-tubules after MI? Data in Figure 3E and 3J should be combined into a 2WANOVA, and a statistical test made for the interaction between heart failure (or not) and location where the agonist was applied.

13) This study is not the first to examine β3 signaling in plasma membrane compartments or its modulation by heart failure (see Trappanese et al., Basic Res Cardiol, 2015), however there are differences in their findings to the current ones that deserve comment. They too found β3R expressed in both caveolin enriched and heavy (non-lipid raft) regions shown by direct Western Blot. Their results differ however, in that the distribution in each region did not appear to change in the canine HF model (mitral regurgitation), and signaling seemed only observed if the β3R was stimulated in the non-lipid raft compartment. The methods were clearly different, but results should be discussed, particularly in light of Comment 12. Ideally, using a membrane fraction, then Cav3 pull down, and β3-AR protein analysis would help confirm if there is indeed a redistribution in this model that was not observed in the canine model reported previously. If there is a real difference, the question then becomes what goes on in human – is it more like rat or dog?

14) There are a number of figures shown where the FRET tracings themselves do not appear to represent the summary data. Examples are in Figure 2 and 5. There are also concerns that for some analysis, very few animals were used, (e.g. Figure 1F, Figure 4—figure supplement 1A, B, Figure 5—figure supplement 1, Figure 5C), and this raises concerns about reproducibility.

15) Figure 5 is another key figure in the study, looking at potential crosstalk between β3-AR-stimulated cGMP and cAMP via PDE2 regulation. The modest decline from β3-AR agonist in controls seems to be reproduced in the post MI cells too, though the statistical variance is slightly different so one is borderline significant and the other not. The proper test would be a 2WANOVA, where the impact of heart condition is one variable, and the β3-AR application the other. This could be done for the β3-AR response, and then for +/- PDE2-I response where β3-AR stimulation is present. To support the authors' conclusions, there would need to be a significant interaction.

The authors choose forskolin to stimulate cAMP and PDE2-dependent hydrolysis, but this should be contrasted to a β-AR agonist to better support their conclusion (e.g. as depicted in Figure 6).

16) The novelty of the immunofluorescent microscopy data is questioned by prior reports showing similar co-distribution of sGC and Caveolin 3 in normal and failing hearts (e.g. Tsai et al., 2012). There is also a concern that the microscopy is based on freshly isolated cells, whereas the functional data were obtained 48 hours post isolation (time needed to introduce the FRET probes). What is the immunohistology distribution after this 48 hours? Myocytes are known to change the localization of ion channels and various GPCRs during this time period.

17) The MI model is not described, so it is difficult to know the severity of dysfunction generated. Equally important, the location in the heart from which myocytes cells were isolated post MI – is not noted. This is important; e.g. was this peri-infarct, or from the remote territory? How was that determined? How reduced was the EF generated and what other evidence was there supporting this as heart failure – e.g. elevated filling pressures, pulmonary edema, etc.. What was the sex of the animals?

18) The scheme shown in Figure 6 is interesting but not really confirmed by the data shown and so remains speculative. This is because Figure 5 does not test the relative signaling involved in the two different compartments (TT vs. Crest), and with cAMP being stimulated coupled to a Cav3/localizing β-AR, versus not.

19) In Figure 1B and f the authors show a reduced cGMP generation after Isoproterenol stimulation. However, it is not clear whether this is entirely mediated via β3 receptors. The authors should do the similar experiments in the presence of CGP plus ICI (β1 and β2 adrenoceptor blockers).

20) In Figure 2E it is not entirely clear whether the FRET response by IBMX (about 4%) represents the cGMP generation on top of Isoproterenol evoked cGMP generation. If not the IBMX effect together with Isoproterenol is not larger than with ISO alone (Figure 1F). This should be clarified.

21) In the first paragraph of the subsection “Functional β3-ARs are localized in the T-tubules of healthy cells and redistribute to the non166 tubular sarcolemma in heart failure”, the authors state that in failing hearts cGMP generation can be detected across the sarcolemma. The authors should more specifically state that in healthy cardiomyocytes there is no crest response which becomes now obvious in failing hearts. At least this is demonstrated in Figure 3D and I.

22) The immunostainings shown in Figure 4 and Figure 5—figure supplement 1 entirely depend on the specificity of the antibodies directed against sGCα1, caveolin3 and sGCβ1. The authors should comment in the manuscript about what is known about the specificity of these antibodies. Have they been used previously in knockout control cells or what other efforts have been made to demonstrate their specificity?

23) Figure 4—figure supplement 1: In the model in figure B and C, it is suggested that cGMP formation is reduced in the T tubules after induction of heart failure. However, if there is no statistical significance Figure 4—figure supplement 1A, B between the first bar (T tubule scramble) and the third bar (T tubular C3SD). Thus, this model is not supported entirely by the data.

24) In Figure 5 the authors should somehow indicate more clearly at which time points the changes in cAMP generation are quantified in C. Is this at the time point just before IBMX application? The authors should clearly state in the manuscript the percentage of inhibition in cAMP generation by the β3-agonists. It seems to be in the range of 10-15%?

25) The model in Figure 6 in healthy cardiomyocytes indicates that cGMP act on PDE2 to limit cAMP generation. The authors should elaborate in a little bit more detail how this is achieved for readers not entirely familiar with the previous literature.

https://doi.org/10.7554/eLife.52221.sa1

Author response

Essential revisions:

1) First and foremost, recent work of PC Simpson showed clearly that β3-AR does not exist in murine cardiac myocytes. The authors should examine their data in view of these observations or argue against.

We are aware of the recent publication from the PC Simpson group describing single cell PCR analysis of all adrenergic receptor types in wildtype mice (Circ Res. 2017; 120(7): 1103-1115). However, in this manuscript we deal with rat myocytes only. Therefore, we can’t comment on the mouse cells.

Different publications have clearly shown that there are species differences in the expression profiles of adrenergic receptors and thatβ3-AR are indeed expressed in rat cardiomyocytes. Also, it is proven that they are overexpressed in various disease states in the rats (i.e. The effect of diabetes on expression of β1-, β2-, and β3-adrenoreceptors in rat hearts. Dinçer UD, Bidasee KR, Güner S, Tay A, Ozçelikay AT, Altan VM. Diabetes. 2001 Feb;50(2):455-61; The Trend of β3-Adrenergic Receptor in the Development of Septic Myocardial Depression: A Lipopolysaccharide-Induced Rat Septic Shock Model. Yang, N., Shi, X. L., Zhang, B. L., Rong, J., Zhang, T. N., Xu, W., and Liu, C. F., Cardiology, (2018), 139(4), 234–244; Cardiac effects of long-term active immunization with the second extracellular loop of human β1- and/or β3-adrenoceptors in Lewis rats. Montaudon E, Dubreil L, Lalanne V, Vermot Des Roches M, Toumaniantz G, Fusellier M, Desfontis JC, Martignat L, Mallem MY. Pharmacol Res. (2015), 100:210-9).

Additionally, in many publications the effects of β3-AR specific agonists and antagonists have been reported on the rat cardiomyocytes (i.e. Anti-hypertrophic and antioxidant effect of β3-adrenergic stimulation in myocytes requires differential neuronal NOS phosphorylation. Watts VL, Sepulveda FM, Cingolani OH, Ho AS, Niu X, Kim R, Miller KL, Vandegaer K, Bedja D, Gabrielson KL, Rameau G, O'Rourke B, Kass DA, Barouch LA. J Mol Cell Cardiol. (2013), 62:8-17.; Nebivolol: a multifaceted antioxidant and cardioprotectant in hypertensive heart disease. Khan MU, Zhao W, Zhao T, Al Darazi F, Ahokas RA, Sun Y, Bhattacharya SK, Gerling IC, Weber KT. J Cardiovasc Pharmacol. (2013), 62(5):445-51).

Moreover, we have already expressed our criticism of the PC Simpson group publication in an open letter, stating that their methodological approach may not represent the presence of β3-AR correctly, particularly not in disease states, where they are upregulated: https://doi.org/10.1161/CIRCRESAHA.117.310942).

Nevertheless, in our experiments we also noticed that not every cell responded to Isoproterenol treatment with a distinct cGMP FRET signal, therefore we added “β-adrenergic stimulation (ISO, 100 nmol/L) of healthy control rat ventricular cardiomyocytes expressing the cGMP biosensor red cGES-DE5, led to the production of substantial amounts of cGMP (Figure 1A) in about 2/3 of all tested cells” into our Results. Additionally, we changed the graphs containing Isoproterenol stimulation of cGMP into scatter plots, to make the non-responders visually discernible from the responding cells (see Figures 2, 4 and Figure 4—figure supplement 1).

2) Evidence that the myocytes are from a state of heart failure should be included in the Results (echo data).

We agree with the reviewer and have added the respective echo data as Figure 1.

Additionally, we added the following passage into the Results section:

“Echocardiographic and biometric data show heart failure phenotype. Echocardiographic and biometric data were collected and are summarised in Figure 1.”

And the following sentence into the Materials and methods section: “Echocardiography and biometric data are summarised in Figure 1.”

3) The choice of drug concentrations (β1-AR, β2-AR antagonists, PDE inhibitors) must be justified – selectivity of drugs for the targets is key to the interpretation of most data in this manuscript.

The choice of drug concentrations in this manuscript is based on our previous work/publications: Nikolaev et al., 2006; Nikolaev et al., 2010; Hoffmann et al., 2004.

The above publications are providing the drug concentrations for the β-Adrenergic subtypes.

Johnson et al., 2012.

We added the following paragraph into the Materials and methods section about the choice of concentrations “Drug concentrations. The concentrations for pharmacological agonists and antagonists for the β-AR subtypes and the PDE subtypes investigated were taken from previously published work, which established the drug affinity through competitive radioligand binding assays, FRET sensor dose-response levels and PDE activity assays.” with references to the above papers.

4) There are issues with use of statistical tests. MW or t-tests are used to compare multiple groups (Figure 1, Figure 4—figure supplement 1, Figure 5). This should be re-done with ANOVA or non-parametric equivalent. Please notice also other comments about statistical analysis of the data.

We thank the reviewer for the observation and have now performed renewed statistical tests in most figures using either non-parametric Mann Whitney U or a mixed ANOVA followed by χ2-test with cells assigned to the respective rats they were isolated from. As a result, Figure 2 and Figure 5E lost their statistical significance, but are retained in the manuscript as Figure 3 shows the main PDEs degrading β3-AR-dependent cGMP and both figures show a potential trend of altered cGMP regulation.

5) Interpretation of t-tubule/crest responses (Figure 3 and Figure 4—figure supplement 1) does not appear justified by the data. In Figure 3, comparison of E and J suggest that T-tubule responses are lost but not that they are redistributed to other areas of the sarcolemma. By contrast, comparison of the impact of the C3SD peptide and its scrambled control in healthy cells suggests a gain of crest signalling, rather than a loss of t-tubular signals.

We thank the reviewer for the suggestion and have reworded the text accordingly as follows:

“The increased activity of β3-ARs in non-tubulated surface areas in failing cells might be linked to a disrupted association of β3-AR with caveolar signalosomes. […] However, this difference can be abolished by treating cells with the TAT-C3SD peptide, but not with a scrambled peptide, with the level the response on the crest areas increasing to the level of T-tubules (Figure. 1, p=0,0458).”

In addition, to better visualise the cell responses to β3-AR stimulation we changed the presentation of Figures 1,2 and 4.

6) Care must be taken with interpretation of work using the C3SD peptide. The sequence replicates the so-called caveolin scaffolding domain (CSD) which for many years was considered to mediate the regulatory role of caveolins through interaction with a complementary caveolin binding motif (CBM) in target proteins. This central tenet has been questioned within the last decade by work from the Parton and Dart laboratories which show that most CBM in caveolin's 'binding partners' are inaccessible for interaction with the CSD. This does not necessarily weaken the evidence for a role for CSD, which may exert effects via alternative mechanisms. However, we are not aware of any evidence that the C3SD peptide leads to 'deletion of intact caveolae'.

We thank the reviewer for this correction and have reworded this part in the manuscript to “We investigated this hypothesis by using the cell-permeable peptide disruptor of caveolar signalling TAT-C3SD. The addition of this peptide leads to the dissociation of caveolar signalosomes by inhibiting signalling which is dependent upon the binding to the caveolin-3 specific scaffolding domain (C3SD)(MacDougall et al., 2012)”, while citing the MacDougall et al., 2012 publication from the Calaghan S. group.

The cartoons (Figure 4—figure supplement 1C, D) imply that Cav3 (presumably the red line, although not defined in legend) is a caveolar neck protein, which is not accurate. Furthermore, these cartoons suggest that with the C3SD peptide, the t-tubular cGMP signal is reduced which is not borne out by the data in part A of this figure. The legend states 'in cardiomyocytes treated with..C3SD..the cGMP signal in…crest areas is even stronger than when stimulated in t-tubules' but the comparison between these 2 datasets is not made (also see 3 above).

We thank the reviewer for this observation and have changed the cartoon to portray Cav 3 more accurately. We also performed additional C3SD peptide measurements for the Figure 5—figure supplement 1.

7) It is important to acknowledge the limitation of confocal imaging to provide robust support for the location of proteins in caveolae (lateral resolution ≈200 nm, caveolae 50-100 nm). Could more meaningful information be gained from confocal images (Figure 4) by a separation of interior (t-tubular) and surface (crest) staining?

We thank the reviewer for this observation and the suggestion to look at different cell areas in our confocal images. We have performed further Pearson Correlation Coefficient measurements by drawing the regions of interest so they either measured the membranes or the cell interior. We could however detect no significant differences in the cell areas. We added a remark about the limitation of confocal microscopy to detect structures underneath the diffraction limit in the Discussion, as follows: “Though the immunocytochemical method is limited in its spatial resolution and can therefore not resolve the caveolae structures themselves, it allows us to detect an alteration in sGC localisation in heart failure and, which could potentially be indicative of dysregulated caveolar signalosomes as reported previously in a pressure overload induced heart failure model using mice (Tsai et al., 2012)”.

8) The increased co-localisation of sGC and α-actinin is not explained (Figure 4F) – this seems to suggest increased sCG at z discs, i.e., t-tubules in failing cells which is not consistent with the overall interpretation of data.

We understand the point the reviewer is making, however after more stringent statistical testing Figure 5F has lost its significance. Nonetheless, we opted to leave the images and graph in the manuscript as it might give a potential trend of where the sGCα subunit might be relocating to in heart failure we and have added the following sentence in the Discussion section: “We have observed a trend to an increased overlap between our sGCα and α-actinin staining in our confocal imaging, which could potentially represent an increased redistribution of sGCα to the areas of the Z-disc not directly associated with the T-tubules or caveolar signalosomes.”

9) Figure 5 – the magnitude of the impact of β3-AR stimulation on cAMP production (indexed with the Epac-based sensor) is small even in control cells. If this significant difference is retained when the data are analysed using ANOVA, the authors should comment on the relevance of such a small change. Perhaps measurement of cAMP signals in different subcellular compartments might reveal profound differences.

We have acquired new data that and added to the figure, this effect is statistically significant (see Figure 6).

However, we appreciate the reviewer’s comment and have added the following paragraph into the Discussion:

“The observed response of β3-AR stimulation of about 10 percent on overall cAMP levels, as measured via the cytosolic FRET sensor Epac1-camps, could be of physiological relevance when brought into the context of lowering pathological cAMP signalling levels on a whole or in confined signalling compartments.”

10) No direct evidence is presented here that β3-AR are caveolae-based (presumably because of lack of an appropriate antibody) yet some unsubstantiated statements are made e.g. 'β3 -AR dependent cAMP pools are formed in caveolae'. Some evidence (referenced in the manuscript) supports the location of some sGC subunits in buoyant caveolae-containing fractions, and the evidence for caveolar location of eNOS is robust. However, comprehensive evidence to support statements regarding the location of the β3 signalosome in caveolae is lacking.

We thank the reviewer for this comment. In the past we have shown β3-AR association with caveolae in myocytes transgenically expressing the human β3-AR, which allowed us to perform membrane fractionation protein assays as well as a proximity ligation assay using a human specific β3-AR and Caveolin 3 antibody. We also currently have a paper in press (Dubois-Deruy E ESCHF 2020, in press) containing further evidence of this association via a PLA colocalization signal between the endogenous rat β3-AR and AMPK (itself in caveolae) from rat neonatal cardiac myocytes. In any case we added a remark in the Introduction to mention our previous work on the caveolae and β3-AR association: “Recently, overexpression of β3-AR in transgenic mice has been shown to protect the heart from catecholamine-induced hypertrophy and remodelling via an eNOS)/soluble guanylyl cyclase (sGC)/cGMP-dependent signalling pathway. The same study showed localization of β3-ARs together with eNOS in caveolae-enriched membrane fractions, which had been separated via ultracentrifugation (Belge et al., 2014)”.

11) Figure 6. There seems to be some lack of accuracy in depicting current (and past) data from this group. Why are β1-AR only in crest in healthy myocytes (vs. Nikolaev et al. 2010 Science). Are the PDE2/5 symbols misplaced at the bottom of panel A (surely these degrade cGMP in control cells as shown in Figure 2). The data in Figure 5C suggest that the normal β3-AR dependent reduction in cAMP seen in control cells is absent in failing cells (i.e. cAMP lower in control cells in the presence of β3-AR stimulation). Yet comparison of panel A and B in Figure 6 suggests that cAMP is lower in failing cells in the presence of β3-stimulation. Even given an increase in PDE2 expression/activity in failing cells (not shown explicitly) which could increase degradation of cAMP, how can the data in Figure 5 be reconciled with the cartoons in Figure 6?

We thank the reviewer for spotting the misplaced PDE symbols and have changed the overall schematic to better reflect current and old data (see Figure 7 and its legend).

12) It appears that the differently targeted β3-AR stimuli applied to T-tubules vs. Crest membrane regions were performed in different cells, with comparisons between them analyzed with non-paired statistics. This is not explicitly stated but is apparent by the statistical test used (Mann Whitney U). This seems problematic given that the level of FRET response varies from cell to cell, and even accepting IBMX-based normalization between cells, the statistical power to identify quantitative changes would be challenging. If getting both stimuli in the same cell is impossible and unpaired analysis is required, then the ability to properly normalize between cells is critical. The concern with it is raise by the key data shown in Figure 3. While the extent to which β3-AR-mediates cGMP at the Crest appears similar between healthy and failing cardiomyocytes, the FRET response at T-tubule declines in the latter group. If this was due to a redistribution of β3-AR to the Crest region, would you not expect a rise in the% FRET response there? Is it possible the results can be explained by a decline in β3-R levels or signal-coupling in the T-tubules after MI? Data in Figure 3E and 3J should be combined into a 2WANOVA, and a statistical test made for the interaction between heart failure (or not) and location where the agonist was applied.

Unfortunately, a proper normalization with the red-DE5 FRET sensor is not yet established and performing both measurements of crest and T-tubule in one cell could lead to potential desensitization of receptors and alterations in the FRET sensor. We appreciate the reviewer’s suggestion that we use a more stringent statistical test and have now used a mixed ANOVA followed by χ2-test. We also updated the text to better reflect the data of an increased crest signal in heart failure.

13) This study is not the first to examine β3 signaling in plasma membrane compartments or its modulation by heart failure (see Trappanese et al., Basic Res Cardiol, 2015), however there are differences in their findings to the current ones that deserve comment. They too found β3R expressed in both caveolin enriched and heavy (non-lipid raft) regions shown by direct Western Blot. Their results differ however, in that the distribution in each region did not appear to change in the canine HF model (mitral regurgitation), and signaling seemed only observed if the β3R was stimulated in the non-lipid raft compartment. The methods were clearly different, but results should be discussed, particularly in light of Comment 12. Ideally, using a membrane fraction, then Cav3 pull down, and β3-AR protein analysis would help confirm if there is indeed a redistribution in this model that was not observed in the canine model reported previously. If there is a real difference, the question then becomes what goes on in human – is it more like rat or dog?

We thank the reviewer and believe the suggested experiment could be of potential value if it were feasible. However, due to the relatively low expression of β3-ARs and the current lack of a suitable antibody, which could detect rat β3-ARs with appropriate specificity, we have refrained from performing this experiment. We agree that the data from Trappanese et al., 2015, does not match our data exactly, however we assume that this can at least in part be explained by the different pathophysiological courses of heart failure in the two different models. We have shown in the past, that in human a residual cGMP increase was detectable after β3-AR agonist application on biopsies from patients with different types of heart disease. (see: Gauthier et al., 1998).

14) There are a number of figures shown where the FRET tracings themselves do not appear to represent the summary data. Examples are in Figure 2 and 5. There are also concerns that for some analysis, very few animals were used, (e.g. Figure 1F, Figure 4—figure supplement 1A, B, Figure 5—figure supplement 1, Figure 5C), and this raises concerns about reproducibility.

The reviewers concern about low animal numbers was noted and additional experiments were performed for figures in which only two animals were used, with the exception of Figure 4—figure supplement 1A, B) in order to increase n numbers and alleviate the concerns about reproducibility. We kept the current traces of FRET experiments, since we believe that the traces are representative of the summary data.

15) Figure 5 is another key figure in the study, looking at potential crosstalk between β3-AR-stimulated cGMP and cAMP via PDE2 regulation. The modest decline from β3-AR agonist in controls seems to be reproduced in the post MI cells too, though the statistical variance is slightly different so one is borderline significant and the other not. The proper test would be a 2WANOVA, where the impact of heart condition is one variable, and the β3-AR application the other. This could be done for the β3-AR response, and then for +/- PDE2-I response where β3-AR stimulation is present. To support the authors' conclusions, there would need to be a significant interaction.

The authors choose forskolin to stimulate cAMP and PDE2-dependent hydrolysis, but this should be contrasted to a β-AR agonist to better support their conclusion (e.g. as depicted in Figure 6).

We thank the reviewer for the suggestions and have performed further experiments to increase n-number and for statistics we used a mixed ANOVA followed by χ2-test to incorporate the variance introduced by the different animals. The differences are still significant. We chose forskolin as our cAMP stimulant as we wished to have a direct comparison between control and heart failure cells. Due to β1-AR desensitization in heart failure such a direct comparison would have been complicated, if we had opted to use a direct β-AR agonist.

16) The novelty of the immunofluorescent microscopy data is questioned by prior reports showing similar co-distribution of sGC and Caveolin 3 in normal and failing hearts (e.g. Tsai et al., 2012). There is also a concern that the microscopy is based on freshly isolated cells, whereas the functional data were obtained 48 hours post isolation (time needed to introduce the FRET probes). What is the immunohistology distribution after this 48 hours? Myocytes are known to change the localization of ion channels and various GPCRs during this time period.

We thank the Editor and the reviewers for raising this important point. Cardiomyocytes do indeed remodel but the remodelling of rat cells is not so extensive as to preclude useful experiments. Papers from our group (Gorelik et al. Cardiovascular Research, 2006) and others (Pavlovic et al. Exp Physiol, 2010) have demonstrated that factors such as the organization of the surface of the cardiomyocytes (z-groove ratio) and t-tubule density (rat cells lose only 25% of their t-tubule area over the first 48 hrs). Functional factors are not markedly altered, neither is the capacitance of the membrane or the staining of tubular structures after 48 hours despite the cells looking morphologically quite different due to the rounding of the myocytes ends. Rat cells compare favourably with those isolated from mouse which completely lose their form and function over 48 hrs (Pavlovic et al. Exp Physiol, 2010).

17) The MI model is not described, so it is difficult to know the severity of dysfunction generated. Equally important, the location in the heart from which myocytes cells were isolated post MI – is not noted. This is important; e.g. was this peri-infarct, or from the remote territory? How was that determined? How reduced was the EF generated and what other evidence was there supporting this as heart failure – e.g. elevated filling pressures, pulmonary edema, etc.. What was the sex of the animals?

We thank the reviewer for their observation and have added the respective echocardiography and biometric data as Figure 1.

Additionally, we added the following passage into the Results section:

“Echocardiographic and biometric data were collected and are summarised in Figure 1.”

And the following into the Materials and methods section: “Echocardiographic and biometric data are summarised in Figure 1.”

Animals were all male, we have added this information in the Materials and methods section as follows: “All animals were male.”

18) The scheme shown in Figure 6 is interesting but not really confirmed by the data shown and so remains speculative. This is because Figure 5 does not test the relative signaling involved in the two different compartments (TT vs. Crest), and with cAMP being stimulated coupled to a Cav3/localizing β-AR, versus not.

We appreciate the reviewer’s point, but with the experiment shown in Figure 5 (now Figure 6) we wished to detect if there was an overall potential of β3-AR activation to lower and thereby attenuate cAMP levels in control and heart failure. We have slightly changed Figure 6 (now Figure 7) to be less speculative. See Figure 7.

19) In Figure 1B and f the authors show a reduced cGMP generation after Isoproterenol stimulation. However, it is not clear whether this is entirely mediated via β3 receptors. The authors should do the similar experiments in the presence of CGP plus ICI (β1 and β2 adrenoceptor blockers).

We agree with the reviewer that it would benefit our work if we could show the β3-AR response only in the MI cells (with inhibiting β1 and β2 receptors); however, we were unable to perform the suggested experiment as we did not have any more MI cells (in the 2 months of time that we were given for the revision).

20) In Figure 2E it is not entirely clear whether the FRET response by IBMX (about 4%) represents the cGMP generation on top of Isoproterenol evoked cGMP generation. If not the IBMX effect together with Isoproterenol is not larger than with ISO alone (Figure 1F). This should be clarified.

We thank the reviewer for this point and have tried to make the data description clearer by adding the following passage into the figure legend: “The scatter plot/histograms present the average whole cell cGMP-FRET responses evoked by PDE inhibition further to the isoproterenol responses in% from (A-D) (E).”, that means i.e. if the ratio after ISO application was at 0.98 and the following ratio after single PDE inhibition was at 0.97 and the ratio after IBMX application at 0.96, then the values would be 1 for the single PDE and 2 for the total PDE inhibition.

21) In the first paragraph of the subsection “Functional β3-ARs are localized in the T-tubules of healthy cells and redistribute to the non166 tubular sarcolemma in heart failure”, the authors state that in failing hearts cGMP generation can be detected across the sarcolemma. The authors should more specifically state that in healthy cardiomyocytes there is no crest response which becomes now obvious in failing hearts. At least this is demonstrated in Figure 3D and I.

We thank the reviewer for this comment and have changed the manuscript text accordingly to “In healthy cardiomyocytes, we observed that functional β3-ARs reside mainly in the T-tubules with very few responses being detectable outside of T-tubules (Figure 4A-E), whereas in failing cells, they β3-ARs responses after localised stimulation can be detected in both tubulated and non-tubulated areas across the sarcolemma (Figure 4F-J).”

22) The immunostainings shown in Figure 4 and Figure 5—figure supplement 1 entirely depend on the specificity of the antibodies directed against sGCα1, caveolin3 and sGCβ1. The authors should comment in the manuscript about what is known about the specificity of these antibodies. Have they been used previously in knockout control cells or what other efforts have been made to demonstrate their specificity?

We agree with the reviewer that the specificity of antibodies is essential for the experiments shown in our manuscript. Therefore, we used the best available antibodies and added the specificity info into the text. We have commented on the specificity of the antibodies we employed in our manuscript in the Materials and methods section as follows: “rabbit sGCα and β subunit antibodies (Prof. Andreas Friebe from Würzburg, Germany, specificity tested in KO animals Friebe, Voußen and Greoneberg, 2018), mouse α-actinin (Σ Aldrich, A7732), mouse Caveolin 3 (BD Transduction Laboratories, 610421, specificity tested in KO animals (Woodman et al., 2002)”.

Additionally, we tested the specificity of sGCα. Freshly isolated adult mouse cardiomyocytes (either from wild type of from global sGC knockout mice) were fixed with ice-cold methanol and co-stained with sGCb1 and Cav3 antibodies as described in the Materials and methods and Figure 5—figure supplement 1) Confocal images were using the small laser intensities in all conditions similar to those used in this manuscript.

23) Figure 4—figure supplement 1: In the model in figure B and C, it is suggested that cGMP formation is reduced in the T tubules after induction of heart failure. However, if there is no statistical significance in Figure 4—figure supplement 1A, B between the first bar (T tubule scramble) and the third bar (T tubular C3SD). Thus, this model is not supported entirely by the data.

We thank the reviewer for the observation and have altered the cartoon accordingly, so that the cGMP formation in HF tubules does not appear smaller. See figure 4—figure supplement 1.

24) In Figure 5 the authors should somehow indicate more clearly at which time points the changes in cAMP generation are quantified in C. Is this at the time point just before IBMX application? The authors should clearly state in the manuscript the percentage of inhibition in cAMP generation by the β3-agonists. It seems to be in the range of 10-15%?

We thank the reviewer for the suggestion and have attempted to clarify the measurement steps in the manuscript with the following explanation in the Figure 5 figure legend: “Scatter plot/histogram presenting whole cell cAMP-FRET responses depicted as the percentage of the maximal possible cAMP FRET response (= Forskolin followed by IBMX). The measured Forskolin or IBMX responses were the respective maximal responses, equalling the lowest FRET ratio value, achieved after each stimulus.” Additionally, we added the average response rate of 10.3% into the manuscript text as follows: “In healthy cells, stimulation of β3-AR led to a significant reduction of approximately 10.3% of the forskolin stimulated cAMP production (Figure 5C)”.

25) The model in Figure 6 in healthy cardiomyocytes indicates that cGMP act on PDE2 to limit cAMP generation. The authors should elaborate in a little bit more detail how this is achieved for readers not entirely familiar with the previous literature.

We thank the reviewer for the suggestion and have elaborated more on this point in the Discussion section as follows: “At the same time as being degraded by PDE2, cGMP can also increase PDE2 activity on cAMP more strongly by binding to it, thereby triggering the so-called cGMP-to-cAMP crosstalk.This crosstalk between the second messengers of which one (cAMP) has increasingly been associated with detrimental signalling pathways in the context of heart failure, which the other (cGMP) could potentially attenuate (Mongillo et al., 2006; Moniotte et al., 2001; Yanaka et al., 2003)”, as well as in the figure legend (see Figure 7).

https://doi.org/10.7554/eLife.52221.sa2

Article and author information

Author details

  1. Sophie Schobesberger

    1. Myocardial Function, National Heart and Lung Institute, Imperial College London, ICTEM, Hammersmith Hospital, London, United Kingdom
    2. Institute of Experimental Cardiovascular Research, University Medical Center Hamburg-Eppendorf, German Center for Cardiovascular Research (DZHK) partner site Hamburg/Kiel/Lübeck, Hamburg, Germany
    Contribution
    Formal analysis, Investigation
    Contributed equally with
    Peter T Wright
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0001-8268-0019
  2. Peter T Wright

    Myocardial Function, National Heart and Lung Institute, Imperial College London, ICTEM, Hammersmith Hospital, London, United Kingdom
    Contribution
    Data curation, Formal analysis, Investigation, Methodology
    Contributed equally with
    Sophie Schobesberger
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-6504-590X
  3. Claire Poulet

    Myocardial Function, National Heart and Lung Institute, Imperial College London, ICTEM, Hammersmith Hospital, London, United Kingdom
    Contribution
    Data curation, Investigation, Visualization, Methodology
    Competing interests
    No competing interests declared
  4. Jose L Sanchez Alonso Mardones

    Myocardial Function, National Heart and Lung Institute, Imperial College London, ICTEM, Hammersmith Hospital, London, United Kingdom
    Contribution
    Formal analysis, Investigation, Methodology
    Competing interests
    No competing interests declared
  5. Catherine Mansfield

    Myocardial Function, National Heart and Lung Institute, Imperial College London, ICTEM, Hammersmith Hospital, London, United Kingdom
    Contribution
    Resources, Data curation, Methodology
    Competing interests
    No competing interests declared
  6. Andreas Friebe

    Physiologisches Institut, University of Würzburg, Würzburg, Germany
    Contribution
    Conceptualization, Resources, Funding acquisition
    Competing interests
    No competing interests declared
  7. Sian E Harding

    Myocardial Function, National Heart and Lung Institute, Imperial College London, ICTEM, Hammersmith Hospital, London, United Kingdom
    Contribution
    Conceptualization, Resources, Supervision
    Competing interests
    No competing interests declared
  8. Jean-Luc Balligand

    Pole of Pharmacology and Therapeutics (FATH), Institut de Recherche Expérimentale et Clinique (IREC), Université Catholique de Louvain (UCLouvain), Brussels, Belgium
    Contribution
    Conceptualization
    Competing interests
    No competing interests declared
  9. Viacheslav O Nikolaev

    Institute of Experimental Cardiovascular Research, University Medical Center Hamburg-Eppendorf, German Center for Cardiovascular Research (DZHK) partner site Hamburg/Kiel/Lübeck, Hamburg, Germany
    Contribution
    Conceptualization, Resources, Software, Supervision, Funding acquisition, Methodology
    Contributed equally with
    Julia Gorelik
    For correspondence
    v.nikolaev@uke.de
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-7529-5179
  10. Julia Gorelik

    Myocardial Function, National Heart and Lung Institute, Imperial College London, ICTEM, Hammersmith Hospital, London, United Kingdom
    Contribution
    Conceptualization, Resources, Software, Supervision, Funding acquisition, Methodology
    Contributed equally with
    Viacheslav O Nikolaev
    For correspondence
    j.gorelik@imperial.ac.uk
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0003-1148-9158

Funding

British Heart Foundation (12/18/30088)

  • Julia Gorelik

Wellcome (WT090594)

  • Julia Gorelik

Deutsche Forschungsgemeinschaft (Fr 1725/3-2)

  • Andreas Friebe
  • Viacheslav O Nikolaev

National Institutes of Health (ROI-HL grant 126802)

  • Julia Gorelik

Deutsche Forschungsgemeinschaft (NI 1301/3-2)

  • Andreas Friebe
  • Viacheslav O Nikolaev

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

Acknowledgements

We thank Peter O’Gara for cardiomyocyte isolation. Confocal microscopy was performed at the FILM imaging facility of Imperial College London and at the Department of Pharmacology and Toxicology, University Medical Centre Hamburg-Eppendorf.

Ethics

Animal experimentation: All procedures performed in the UK were carried out according to the standards for the care and use of animal subjects determined by the UK Home Office (ASPA1986 Amendments Regulations 2012) incorporating the EU directive 2010/63/EU. The Animal Welfare and Ethical Review Body Committee of Imperial College London approved all protocols. The parts of the investigation, which were performed in Germany, conformed to the guide for the care and use of laboratory animals published by the National Institutes of Health (Bethesda, Maryland; Publication No. 85-23, revised 2011, published by National Research Council, Washington, DC). The experimental procedures were in accordance with the German Law for the Protection of Animals and with the guidelines of the European Community.

Senior Editor

  1. Anna Akhmanova, Utrecht University, Netherlands

Reviewing Editor

  1. Nir Ben-Tal, Tel Aviv University, Israel

Reviewer

  1. Marc Freichel, University of Heidelberg, Germany

Publication history

  1. Received: October 4, 2019
  2. Accepted: March 25, 2020
  3. Accepted Manuscript published: March 31, 2020 (version 1)
  4. Version of Record published: April 7, 2020 (version 2)

Copyright

© 2020, Schobesberger et al.

This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.

Metrics

  • 397
    Page views
  • 97
    Downloads
  • 0
    Citations

Article citation count generated by polling the highest count across the following sources: Crossref, PubMed Central, Scopus.

Download links

A two-part list of links to download the article, or parts of the article, in various formats.

Downloads (link to download the article as PDF)

Download citations (links to download the citations from this article in formats compatible with various reference manager tools)

Open citations (links to open the citations from this article in various online reference manager services)

Further reading

    1. Cell Biology
    2. Structural Biology and Molecular Biophysics
    Jeffrey T Lock, Ian Parker
    Research Article Updated
    1. Cell Biology
    2. Human Biology and Medicine
    Adela Constantinescu-Bercu et al.
    Research Article Updated