Internalized β2-Adrenergic Receptors Oppose PLC-Dependent Hypertrophic Signaling

Abstract

Background:

Chronically elevated neurohumoral drive, and particularly elevated adrenergic tone leading to β-adrenergic receptor (β-AR) overstimulation in cardiac myocytes, is a key mechanism involved in the progression of heart failure. β1-AR and β2-ARs are the two major subtypes of βARs present in the human heart, however, they elicit different or even opposite effects on cardiac function and hypertrophy. For example, chronic activation of β1-ARs drives detrimental cardiac remodeling while β2-AR signaling is protective. The underlying molecular mechanisms for cardiac protection through β2-ARs remain unclear.

Methods:

β2-AR signaling mechanisms were studied in isolated neonatal rat ventricular myocytes and adult mouse ventricular myocytes using live cell imaging and western blotting methods. Isolated myocytes and mice were used to examine the roles of these signaling methods in regulation of cardiac hypertrophy.

Results:

Here we show that β2-AR activation protects against hypertrophy through inhibition of PLCε signaling at the Golgi apparatus. The mechanism for β2-AR-mediated PLC inhibition requires internalization of β2-AR, activation of Gi and Gβγ subunit signaling at endosomes and ERK activation. This pathway inhibits both angiotensin II and Golgi-β1-AR-mediated stimulation of phosphoinositide hydrolysis at the Golgi apparatus ultimately resulting in decreased PKD and HDAC5 phosphorylation and protection against cardiac hypertrophy.

Conclusions:

This reveals a mechanism for β2-AR antagonism of the PLCε pathway that may contribute to the known protective effects of β2-AR signaling on the development of heart failure.

Graphical Abstract

INTRODUCTION

Catecholamines of the sympathetic nervous system (SNS) regulate heart rate, contractility and vascular resistance through activation of adrenergic receptors. Prolonged elevation of circulating catecholamines, including epinephrine and norepinephrine, in response to cardiac injury, vascular disease, or stress, is tied closely to the pathogenesis of heart failure ¹. These hormones act in part through β-adrenergic receptors (βARs) on cardiac myocytes, where chronic activation drives maladaptive cardiac hypertrophy, apoptosis, and fibrosis, ultimately resulting in heart failure ². βARs are G protein-coupled receptors (GPCRs) and consist of three subtypes, β1, β2 and β3. β1-adrenergic receptors (β1-ARs) comprise of 80% of βARs in the healthy human hearts with the β2-ARs comprising the remaining 20% ². Although β1-ARs and β2-ARs respond to the same physiologic agents, share high sequence homology ³, and core signaling pathways (both couple to Gα_s and stimulate cAMP production), they have distinct or even opposite physiological and pathological roles in the heart ⁴. Specifically, chronic pathological stimulation by endogenous catecholamines leads to decreased β1-AR levels and function in the heart, and mild overexpression of β1-ARs results in cardiac hypertrophy and heart failure in mice ^5,6. In contrast, the β2-AR population remains unchanged during heart failure and moderate levels of β2-AR overexpression leads to positive inotropic effects ^7–10. β2-AR deletion in mice leads to exaggerated cardiac hypertrophy in a pressure overload-induced model, and β1-AR deletion mice had a comparable responses to wild type, however, in β1-AR/β2-AR dual knockout mice, cardiac hypertrophy was abolished ^11,12. These observations suggest that coordinated signaling events downstream these receptors are critical in mediating cardiac hypertrophy and heart failure.

Phospholipase C (PLC) enzymes have important roles in the heart ^13–18. PLCs mediate hydrolysis of phosphatidylinositol 4,5 biphosphate (PIP₂) downstream of GPCR and receptor tyrosine kinase (RTKs) activation leading to production of inositol triphosphate (IP3) and diacylglycerol (DAG) ^19,20. Classically, PLCβ isoforms of the PLC family are stimulated downstream of GPCRs by Gq and Gβγ subunits from Gi ²¹. More recently, PLCε has also been shown to be downstream of multiple GPCR families and RTKs due to its ability to be directly activated by a diverse array of upstream regulators including multiple members of the family of small GTPases (Ras, Rho and Rap) and Gβγ subunits, ^22–25.

A pathway that has been extensively studied in cardiac myocytes in our laboratory is regulation of PLCε by Rap after cAMP-dependent stimulation of the Rap GEF, Exchange Protein Activated by cAMP (EPAC) ^16–18,26. We previously identified an essential role of PLCε in regulating cardiac hypertrophy in an animal model of pressure-overload induced cardiac hypertrophy ¹⁸. In cardiac myocytes, PLCε scaffolds to muscle-specific A kinase anchoring protein (mAKAP) along with PKA, EPAC, protein kinase D (PKD) and other hypertrophic regulatory proteins at nuclear envelope ^27,28. Activation of PLCε at this location induces the hydrolysis of phosphatidylinositol-4-phosphate (PI4P) in the closely associated Golgi apparatus, generating inert inositol bisphosphate (IP2) and local DAG to drive the activation of PKD at nucleus and phosphorylation of HDAC leading to the hypertrophic gene expression ¹⁸.

While exploring mechanisms for cAMP-dependent stimulation of the Golgi Epac/PLCε pathway we found that stimulation of cell surface β1-ARs with Iso was ineffective in cardiac myocytes ²⁹. Rather, adrenergic stimulation of the Golgi Epac/PLCε pathway required a pre-existing pool of Golgi localized β1-ARs ²⁶ which stimulate production of a pool of cAMP with privileged access to the Epac/PLCε module at the Golgi/NE interface. Access of Epinephrine (Epi) and norepinephrine (NE) to the intracellular β1-ARs requires an organic cation transporter subtype, OCT3 ^30,31 and blockade of OCT3 attenuated catecholamine-stimulated cardiomyocyte hypertrophy ²⁶ and contractile responses ³². These studies clearly demonstrate physiologically relevant roles of intracellular β1-ARs, however, how plasma membrane/sarcolemma β-AR signaling coordinates with intracellular β1-AR signaling remains poorly explored.

It is well established that intracellular cAMP pools generated at different subcellular compartments downstream of Gs coupled receptors lead to distinct phenotypic outcomes. In cardiac myocytes, β1-ARs but not β2-AR localize on the periphery of the nuclear envelope ^33,34 or Golgi ^26,32,35 in addition to cell surface and T-tubule locations of both β1 and β2-ARs. Activation or β2-ARs results in internalization into endosomes where they continue to signal, initiating a set of signaling and transcriptional control events distinct from those at the plasma membrane ³⁶. Internalization of activated β2-ARs in cardiac myocytes has been documented ^37–39 but functional roles for these internalized receptors have not been thoroughly examined in cardiomyocytes.

In this study, we demonstrate that activation of plasma membrane β2-ARs inhibits stimulation of the hypertrophic EPAC/PLCε pathway at the Golgi downstream of Golgi-β1-ARs and plasma membrane angiotensin II receptors in cardiac myocytes. We describe a mechanism where activated β2-ARs internalize from the plasma membrane, activate Gi, releasing Gβγ, subunits leading to ERK activation and inhibition of PI hydrolysis at Golgi apparatus. This prevents PLCε-mediated downstream hypertrophic signaling including nuclear PKD activation and histone deacetylase (HDAC) phosphorylation. These data reveal a new mechanism that could underly the anti-hypertrophic versus hypertrophic signaling balance between β1 and β2-ARs and give insights into novel strategies for treatment of heart failure by deliberate inhibition of internal β1-ARs and combined with selective β2-AR activation for heart failure therapy.

Methods

Data Availability statement.

The data that support the findings of this study are available from the corresponding author upon reasonable request. Detailed methods are provided in the Supplementary Material. Please see the Major Resources Table in the Supplemental Material.

Results

Activation of β2-ARs opposes Golgi-β1-AR-mediated PLCε activation at the Golgi apparatus.

We previously demonstrated that the cell permeant agonist dobutamine (Dob) stimulates PLCε-dependent PI4P hydrolysis through activation Golgi β1-ARs.²⁶. To assay Golgi PLCε activity we transduce cells with adenoviruses expressing GFP-FAPP (four phosphate adapter protein)-PH which binds selectively to PI4P and measure stimulus-dependent alterations in Golgi associated fluorescence using live-cell time lapse confocal fluorescence imaging in both neonatal rat ventricular myocytes (NRVMs) and adult ventricular myocytes (AVMs) ^18,26. Figure 1A shows co-localization of the PI4P-specific fluorescent probe with the Golgi marker GM130. Regions of interest corresponding to the Golgi apparatus in proximity to the nucleus are quantitated over time before and after agonist addition. As further evidence that Golgi PLCε activation by Dob requires Golgi localized β1-ARs in AVMs we expressed Golgi-targeted-mApple-NB80 (enos-mApple-NB80) in AVMs to block downstream engagement of Gs by Golgi βARs, and monitored PI4P hydrolysis. Localization of enos-mApple-NB80 at the Golgi apparatus was confirmed by co-localization AVMs with the Golgi specific fluorescence sensor GFP-FAPP-PH (Fig S1A). Expression of enos-mApple-NB80 blocked Dob-stimulated PLCε activation while the control enos-mApple had no effect (Fig S1B). This recapitulates previous results demonstrating that Golgi βARs mediate activation of Golgi PLCε by β1-AR agonists in AVMs.

Figure 1. A pathway downstream of β2-ARs suppresses β1-AR stimulation of PLCε at the Golgi.

A) AVMs were transduced with adenoviruses expressing GFP-FAPP-PH for 18 hours prior to fixation and staining for GM130, a cis-Golgi marker. Scale bars = 20μm B) NRVMs were transduced with FAPP-PH-GFP and imaged by a live cell confocal microscopy. NRVMs were stimulated with dobutamine at either 10 μM or 100 nM at the arrow and remaining fluorescence intensity at the Golgi apparatus was monitored over time. C) AVMs transduced with GFP-FAPP-PH were stimulated with dobutamine at either 10 μM or 100 nM at the arrow indicated and Golgi associated fluorescence was monitored over time. NRVMs (D and F) and AVMs (E) were pretreated with a selective β2-AR antagonist ICI-118,551 (50 nM) for 30 min before stimulation with dobutamine at the indicated concentrations and Golgi PI4P hydrolysis was measured. G) AVMs were pretreated with either Sal (100 nM) or vehicle control 30 min before imaging and dobutamine was added at the arrow. Data for B, D, and F were from at least n=8 cells from 3 separate preparations of NRVMs and analyzed with a two-way repeated measures ANOVA (sphericity corrected) with Tukey’s post-hoc test. Data for C, G and E were collected from n=3–5 cells from at least 3 independent preparations of AVMs. C and G were analyzed with a Wilcoxon test, and E was analyzed with a Friedman test and Dunn’s post-hoc test. (See Table S1 for exact cell numbers).

Unexpectedly, while analyzing the concentration-dependence for Dob-mediated Golgi PLCε activation, we found that 100 nM Dob was optimal, but 10 μM Dob did not stimulate PI4P hydrolysis in NRVMs or AVMs (Fig 1B, C, S1C). Dob is relatively selective β1-AR agonist but activates β2-ARs at higher concentrations (pKd 5.6 vs 4.8, β1-ARs vs β2-ARs)⁴⁴.

To assess whether the inhibitory effect on PI4P hydrolysis at higher concentrations of Dob was due to activation of β2-ARs, we pretreated NRVMs or AVMs with the selective β2-AR antagonist ICI-118,551 prior to Dob addition. Preincubation with ICI-118,551 uncovered stimulation of PI4P hydrolysis by 10 μM Dob comparable to that seen with 100 nM Dob in both NRVMs and AVMs (Fig. 1D, E). ICI-118,551 alone has no effects on 100 nM Dob-stimulated PI4P hydrolysis confirming that Dob-stimulated PI4P hydrolysis is β1-AR-dependent (Fig. 1F). To extend this observation, we determined if salmeterol (Sal), a selective β2-AR agonist, could inhibit PI4P hydrolysis stimulated by Dob. Preincubation with Sal (100 nM) inhibited PI4P hydrolysis stimulated by 100 nM Dob (Fig. 1G). Taken together, this data indicates that β2-AR stimulation opposes PLCε activation downstream of Golgi β1-ARs.

β2-AR-dependent inhibition of PI4P hydrolysis is at the level of PLCε signaling.

Previous studies from our laboratory demonstrated that Golgi β1-AR-induced PI4P hydrolysis requires the EPAC activation downstream of βARs and upstream of PLCε in NRVMs ²⁶. To determine if β2-ARs inhibit PI4P hydrolysis downstream of β1-AR-dependent cAMP accumulation, we stimulated PLCε-dependent PI4P hydrolysis using the EPAC-selective cAMP analog, 8-(4-chlorophenylthio)-2-O-methyl-cAMP-acetoxymethyl ester (cpTOME-AM). Pretreatment with Sal prevented cpTOME-AM-stimulated Golgi PI4P hydrolysis indicating that β2-ARs inhibit PLCε activation at the level of EPAC or its downstream effectors, not at the level of cAMP generation by Golgi β1-ARs (Fig 2A, B).

Figure 2. PLCε is the likely target for β2-AR-dependent inhibition of Golgi PI4P hydrolysis.

A) Diagram of β2-AR-dependent blockade of cpTOME-AM mediated activation of EPAC/PLCε. B) AVMs were pretreated with or without Sal (100 nM) before stimulation with cpTOME-AM (10 μM) and Golgi PI4P associated fluorescence was monitored with time. C) Diagram of β2-AR-dependent blockade of AngII mediated activation of PLCε at the Golgi apparatus. D) AVMs were pretreated with or without Sal (100 nM) before stimulation with Angiotensin II (1 μM) and Golgi PI4P associated fluorescence was monitored. Data are from n=3–11 cells from at least 3 independent preparations of AVMs (See Table S1 for exact cell numbers). Data in A were analyzed with a Friedman test and Dunn’s post hoc analysis; Data in B were analyzed with a repeated measures mixed effects model with Tukey’s post-hoc analysis.

We previously demonstrated that a Gq coupled receptor (endothelin 1 A receptor) activates mAKAP associated PLCε through a pathway independent from EPAC ¹⁸. We reasoned that if the inhibitory effect of β2-AR activation was at the level of PLCε beyond EPAC/Rap, activation of PLCε by an alternative pathway would also be inhibited (Fig 2C). To test this AVMs were treated with Angiotensin II (AngII), which activates ATII receptors that couple to Gq and Gi/o but not Gs and is highly relevant to cardiac pathophysiology. AngII strongly stimulated PI4P hydrolysis which was blocked by Sal pretreatment (Fig 2D). This indicates that β2-AR-dependent inhibition is at the level of PLCε/mAKAP complex because two independent signaling pathways involving either Rap or signals downstream of Gq converge at the level of PLCε activation at the Golgi.

β2-AR-dependent inhibition relies on Gi-Gβγ signaling.

Activation of β2-ARs stimulates cAMP production and PKA activation. We have reported that PKA activation counters the stimulation of Golgi PLCε by EPAC ²⁹. We tested whether PKA is involved in β2-AR-dependent PLCε inhibition (Fig. 3A). Preincubation AVMs with a PKA inhibitor, PKI, did not uncover PI4P hydrolysis upon stimulation with 10 μM Dob indicating that β2-AR-mediated inhibition is PKA-independent (Fig. 3B). β2-ARs preferentially activate Gs, but also couple to Gi/o ^37,45. To test the role of Gi we treated AVMs with the Gi/o inhibitor, pertussis toxin (PTX). PTX treatment blocked Sal-mediated inhibition of PI4P hydrolysis induced by Dob (Fig. 3C). PTX treatment alone did not affect Dob-stimulated PI4P hydrolysis (Fig. S2A). To further support a role for Gα_i in β2-AR-dependent inhibition of PLCε activity, NRVMs were treated with siRNA directed at Gα_i2, the prominent Gα_i isoform in cardiac myocytes (Fig. 3D, E). Gα_i2 siRNA treatment prevented salmeterol-dependent inhibition of Dob-stimulated PI4P hydrolysis. Gα_i2 siRNA treatment had no effect on Dob-stimulated PI4P hydrolysis in the absence of salmeterol, and in cells treated with random control siRNA salmeterol blocked Dob-stimulated PI4P hydrolysis. We then tested the role of Gβγ liberated from Gi/o activation downstream of β2-ARs in the signaling events. Treatment with gallein, a Gβγ inhibitor that selectively inhibits interactions with downstream effectors ^46,47, prevented Sal-dependent inhibition of Dob-stimulated PLCε activation (Fig. 3F). Gallein had no effect on either basal PI4P levels or Dob-stimulated PI4P depletion (Fig. S2B).

Figure 3. β2-AR-Gi-Gβγ signaling axis counters activation of PLCε.

A) Diagram of possible signaling pathways downstream of β2-ARs. B) AVMs were pretreated with or without the PKA inhibitor, myrPKI (1 μM) 30 min before stimulation with dobutamine at the indicated concentrations and Golgi PI4P hydrolysis was assessed. C) AVMs were pretreated with either PTX (100 ng/ml) or vehicle control overnight and followed by pretreatment with Sal 30 min before imaging and dobutamine was added at arrow. D) Representative images of NRVMs cotransfected with siGLO red/Gα_i2-siRNA mix (1:1 20 nmol each) and co-transduced with FAPP adenovirus. Scale bar = 40 μm E) NRVMs were transfected with Gα_i2 targeting siRNA or control scrambled siRNA. 48 hours after transfection cells were treated with the indicated agonists. F) AVMs were pretreated with the Gβγ inhibitor, gallein (10 μM) or vehicle control for 30 min prior to the pretreatment with Sal for 30 min and dobutamine was added at the arrow. Data for B, C and F are n=3–6 from 3 independent preparations of AVMs and (See Table S1 for exact cell numbers) and were analyzed Friedman’s test with Dunn’s post-hoc analysis. E; data are n=6–7 from 3 independent preparations of NRVMs and analyzed with a repeated measures mixed effects model (sphericity corrected) with Tukey’s post-hoc test.

β2-ARs inhibit PLCε activation via ERK signaling

Given that β2-ARs can signal through β-arrestin, a nexus for activation of protein kinases including ERK, we tested whether ERK activation downstream of β2-ARs is involved.

Pharmacological inhibition of ERK activity with the highly selective MEK inhibitor PD0325901 abolished Sal-mediated inhibition of Golgi PI4P hydrolysis (Fig. 4A). Similarly, ERK2 siRNA treatment blocked Sal-mediated inhibition of PI4P hydrolysis (Fig. 4B) but had no effect on cells stimulated with Dob alone. Previous studies have reported that β1and β2-ARs can stimulate epidermal growth factor (EGFR) transactivation, leading to ERK activation ^48–50. This pathway requires the release of Gβγ subunits and is PTX sensitive. Additionally, βAR-mediated EGFR transactivation confers protective effects in isolated NRVMs ⁵¹. To assess the involvement of β2-AR-EGFR transactivation, AVMs were pretreated with an EGFR inhibitor, gefitinib followed by preincubation with or without Sal followed by 100 nM Dob stimulation. EGFR inhibition did not alter the Sal-dependent inhibition of Dob-mediated PI4P hydrolysis (Fig. 4C). We also examined if β2-AR-EGFR transactivation contributes to ERK activation downstream of β2-ARs activation. AVMs were stimulated with Sal over a time course of 0–30 min with or without gefitinib preincubation and ERK phosphorylation was assessed by western blotting. Sal treatment leads to clear ERK activation at 30min and EGFR inhibition with Gefitinib had no statistically significant effect on this stimulation. (Fig. 4 D, E). Gefitinib is a potent inhibitor of EGFR and while it may have off target effects, the lack of effect on Sal-dependent inhibition of PI4P hydrolysis and Sal-dependent ERK activation at 30 min, strongly indicates that EGFR transactivation is not involved in this pathway. These observations favor a model where β2-ARs activate Gi and release Gβγ subunits, leading to EGFR-independent ERK activation that antagonizes Golgi PLCε activation.

Figure 4. β2-AR-ERK signaling axis counters activation of PLCε.

A) AVMs were pretreated with the MEK inhibitor, PD0325901 (100 nM) for 30 min prior to the pretreatment with Sal for 30 min and dobutamine was added at the arrow. B) NRVMs were transfected with ERK2 targeting siRNA or control scrambled siRNA. 48 hours after transfection cells were treated with the indicated agonists. C) AVMs were pretreated with or without the EGF receptor inhibitor, gefitinib (10 μM) for 2 hours before the pretreatment with Sal for 30 min and dobutamine was added at the arrow. D) Top panel: Acutely isolated adult ventricular myocytes were pretreated or without gefitinib for 2 hours before the treatment with Sal for indicated times followed by western blotting for p-ERK and total ERK. Shown are representative western blots from three independent preparations of AVMs. Bottom panel, NRVMs were treated with or without gefitinib for 2 h followed by 10 min treatment with EGF followed by western blotting. E) Quantitation of data in D from three AVM preparations (FOB=fold over basal) with the 30 min time analyzed with an unpaired Mann-Whitney U test. Data in A and C were analyzed with a Friedman test and Dunn’s post-hoc test. Data in B was analyzed with a two-way repeated measures ANOVA (sphericity corrected) with Tukey’s post-hoc test. E. See table S1 for exact cell numbers.

Internalized β2-ARs are required to activate ERK and inhibit PLCε signaling.

β2-ARs undergo activation-dependent internalization which has been implicated in ERK activation ⁵². Inhibition of receptor internalization using a dynamin inhibitor, dyngo-4a markedly reduced the β2-AR-mediated blockade of PLCε activation (Fig. 5A) suggesting β2-ARs mediate PLCε inhibition from endosomes rather than the plasma membrane. Salmeterol is a partial β2-AR agonist for both Gs and β-arrestin recruitment relative to Iso, yet does engage arrestin to some degree ^53,54. To determine if Sal causes β2-AR internalization in cardiac myocytes, NRVMs were transfected with flag-β2-ARs and internalization was monitored using fluorescent anti-flag M1 antibody and confocal microscopy. Stimulation of cells with Sal led to accumulation of intracellular β2-AR associated fluorescence confirming that Sal causes β2-AR internalization in myocytes (Fig 5B). To determine if endocytic blockade affects β2-AR stimulation of ERK phosphorylation, freshly isolated AVMs were pretreated with or without dyngo-4a and time-dependent ERK activation stimulated by Sal was measured by western blotting. Dyngo-4a significantly inhibited Sal-elicited ERK phosphorylation at 30 min consistent with a mechanism requiring receptor internalization (Fig. 5C, D). To explicitly examine the involvement of the endosomal Gβγ in the regulation of Golgi PLCε activity, we selectively perturbed endosomal Gβγ using the C-terminus of G protein-coupled receptor kinase (GRK2ct), a highly selective Gβγ inhibitor, fused to a 2XFYVE domain sequence, highly specific to for binding to PI3P enriched endosomes ⁵⁵, to generate an endosomal targeted GRK2ct (FYVE-mApple-GRK2ct) (Fig. 5E). Expression of FYVE-mApple-GRK2ct in AVMs led to a punctate expression pattern consistent with an endosomal location (Fig. 5F). In AVMs co-expressing FAPP-PH-GFP and FYVE-mApple-GRK2ct, Sal was unable block Dob-induced PI4P hydrolysis (Fig. 5G) indicating that the inhibitory signaling downstream of β2-ARs is indeed driven by endosomal Gβγ.

Figure 5. β2-AR-dependent blockade of PLCε activation relies on endosomal Gβγ.

A) AVMs were pretreated with the dynamin inhibitor Dyngo-4a (40 μM) for 30 min prior to pretreatment with Sal for 30 min and dobutamine was added at the arrow. Data was analyzed with a Friedman test with Dunn’s post-hoc test. 4–5 cells from N=3 cell preparations. B) Representative images showing Sal or Iso mediated β2-AR internalization. NRVMs were transfected with Flag-β2-ARs for 24 hours. Cells were then labeled with M1-Flag-488 for 10min at 37°C and then treated with negative control, Sal (100 nM), or Iso (10 μM) at 37°C. Images were acquired at the indicated times by confocal microscopy. Scale bars = 10 μm. C-D) Acutely isolated adult ventricular myocytes were pretreated in the presence or absence of Dyngo-4A before the treatment with Sal for indicated times followed by western blotting for p-ERK and total ERK. Shown is a representative western blot from four independent preparations of AVMs with quantitation shown in D. Data at the 30 min time was analyzed with a Mann Whitney U test N=4 independent cell preparations. E) Diagram of blockade of Gβγ signaling at endosomal membranes. F) Representative image of AVMs expressing FYVE-GRK2ct. Scale bar = 10 μm. G) AVMs were transduced with adenoviruses expressing FYVE-mApple-GRK2ct and FAPP-PH-GFP for 18 hours before imaging. AVMs were stimulated with Dobutamine alone or in the presence of Sal. Golgi associated PI4P fluorescence intensities at 0 and 30 min were measured. Data was analyzed with a Kruskal-Wallis test. N=4. See table S1 for exact cell numbers.

Activation of β2-ARs inhibits nuclear PKD activation.

A signaling event directly downstream of perinuclear PLCε is the activation of nuclear PKD ¹⁸. PKD is directly activated by DAG, the principal active product of PLCε activity acting on Golgi PI4P. If Golgi PLCε is inhibited downstream of Sal, we would predict that β2-AR activation would inhibit nuclear PKD activation. To determine if activation of β2-ARs reduces agonist-dependent PKD activation, NRVMs were pretreated with Sal for 30 min before addition of AngII for 30 min and analysis of PKD phosphorylation by western blotting. Sal treatment eliminated AngII mediated PKD activation (Fig. 6A, B). To more specifically examine nuclear PKD activation downstream of β2-AR and Ang II activation, AVMs were transduced with an adenovirus expressing a nuclear-localized fluorescence resonance energy transfer (FRET) PKD activation reporter, nDKAR ⁴² (Fig. 6C). AVMs were pretreated with either vehicle or Sal for 30 min and AngII-mediated changes in FRET were measured in the nucleus at 0 and 30 min. AngII caused a significant nuclear PKD activation which was suppressed by β2-AR activation (Fig. 6D).

Figure 6. Nuclear PKD activation downstream of PLCε is suppressed by β2-AR activation.

A-B) NRVMs were pretreated with either Sal or vehicle before addition of AngII (1 μM) for 30 min and followed by western blotting for p-PKD and total PKD. Shown is a representative western blot from four separate preparations of NRVMs. Quantitation is shown in B. C) Representative images of AVMs expressing nuclear-DKAR. D) The nuclear region of AVMs expressing nDKAR was selected and the CFP/YFP ratio was measured before and after addition of AngII (1 μM) for 30 min addition in the presence with or without pretreatment with Sal for 30 min. Scale bars = 20 μm. E) Heart lysates from WT mice infused with either saline (Veh) or salmeterol (Sal) (25 μg/kg/day) together with or without AngII (1.5 mg/kg/day) for 14 days. Western blotting was performed to determine the level of p-PKD, p-HDAC, total HDAC and GAPDH. Shown is a representative western blot from 4 mice each group. F) Quantitation of PKD phosphorylation from E. G) Quantitation of HDAC phosphorylation from E. H) Heart weight/body weight (HW/BW) ratios from AngII and AngII+Sal treated mice. N=4 animals each condition. All data were analyzed with a Kruskal-Wallis test.

We had previously shown that cardiac specific deletion of PLCε in mice eliminates PKD activation in a TAC model suggesting that PKD phosphorylation can be used as a proxy for PLCε activity in vivo. To extend our observations to mice, we examined PKD phosphorylation in response to the chronic AngII infusion together with Sal by implanting AngII +/− Sal subcutaneous osmotic minipumps into mice for 14 days. Chronic stimulation with Sal significantly attenuated AngII-mediated PKD activation in mouse hearts (Fig. 6E and F). Similarly, HDAC phosphorylation was blunted in Sal treated group (Fig. 6E and G). PKD phosphorylation of HDAC is partially responsible mediating MEF-dependent hypertrophic gene transcription ⁵⁶. Salmeterol did not statistically significantly inhibit AngII-driven cardiac hypertrophy as measured as a reduction in heart size (HW/BW) compared to animals treated with AngII alone, likely because of the low number of animals tested in this experiment (Fig. 6H).

In a separate cohort of mice, osmotic pumps containing AngII or AngII+Sal were implanted in mice for 14 days and changes in cardiac functional and morphometric parameters were analyzed. Heart mass increased in AngII treated animals but not in animals treated with AngII+Sal, assessed by direct visualization and mass determination (Fig 7A). Ventricular myocyte cross sectional area increased in AngII treated mice which was substantially reversed with coadministration of Sal (Fig 7B). Fractional shortening marginally decreased with Ang II and cotreatment with Sal may have improved function, but these parameters were not statistically significant (Fig 7C, left panel). LVIDs trended toward an increase in AngII treated mice which was reversed AngII+Sal treated animals (Fig 7C, middle panel). LVIDd did not change with AngII administration but Sal treatment led to a decrease in LVIDd (Fig 7C, right panel).

Figure 7. Animals were implanted with osmotic pumps containing Saline (NC), Salmeterol (Sal), AngII or AngII+Sal for 14 days followed by echocardiography, sacrificed and tissue harvesting.

A) Scale bar on hearts is 2 mm and 1 mm on HE stained sections. B) Cells were stained with WGA and myocyte area was quantitated using Image J. Scale bar is 60 μm. Data were analyzed from at least 300 cells from 3 stained sections each condition (See Table S1 for exact cell numbers) using a nested hierarchical unpaired One-Way ANOVA with Tukey’s post-hoc test. C) Echocardiography analysis of fractional shortening (FS), systolic left ventricular inside dimension (LVIDs), and diastolic left ventricular inside dimension (LVIDd). Data in A and C were analyzed using an ordinary one-way unpaired ANOVA with Tukey’s post-hoc test. See Table S1 for animal numbers.

These observations together suggest β2-ARs inhibit development of hypertrophy in mice by preventing detrimental AngII-mediated PLCε signaling in cardiomyocytes. However, we cannot exclude the possibility that Sal mediated protection in mice involves inhibition of AngII-mediated hypertension.

Dob-induced NRVM hypertrophy is inhibited by β2-AR activation and requires Gβγ signaling in endosomes.

In our previous studies, we showed that Golgi resident β1-AR signaling to PLCε is required for catecholamine-mediated cardiomyocyte hypertrophy in NRVMs ²⁶. Our data indicate that Gβγ signaling from endosomes is required for Sal to inhibit Dob stimulated PI4P hydrolysis, a key driver of cardiomyocyte hypertrophy. To determine whether Sal-protects against Dob-stimulated hypertrophy and if endosomal Gβγ signaling is required for this, NRVMs were transduced with fyve-mApple-GRK2ct expressing adenovirus or control β-Gal adenovirus followed by treatment with Dob (100 nM) with or without Sal for 42 h and cell area and expression of the hypertrophic marker, atrial natriuretic factor (ANF), were measured. Dob stimulated a significant increase in cell area (Fig. 8A) and ANF expression (Fig. 8B). Co-treatment with Sal blunted Dob-induced hypertrophy assessed by these two measurements in β-Gal transduced myocytes, (Fig. 8 A, B). In myocytes transduced with fyve-mApple-GRK2ct Sal was unable to inhibit Dob-dependent hypertrophy (Fig. 8 C, D). These effects are not likely due to fibroblast contamination since the NRVM preparations had very few contaminating fibroblasts (<1%) (Fig. S4). These data demonstrate that β2-AR signaling is protective against cardiomyocyte hypertrophy, consistent with the results from mice, and that Gβγ signaling from endosomes is required.

Figure 8. Activation of β2-ARs inhibits dobutamine induced cardiomyocyte hypertrophic growth and requires endosomal Gβγ signaling.

A and B, Neonatal rat ventricular myocytes (NRVMs) were transduced with β-galactosidase expressing adenovirus for 6 hours then stimulated with dobutamine for 42 hours in the presence of Sal or vehicle control followed by fixation. A, Cells were stained for α-actinin to identify cardiomyocytes and quantitated for cell area by image J. B, Cells were stained for ANF (atrial natriuretic factor) expression and with DAPI (4’,6-diamidino- 2-phenylindole) to identify nuclei. The fluorescence intensity of ANF rings was quantified by image J. C and D, Same as A except NRVMs were transduced with FYVE (Fab 1, YOTB, Vac 1, and EEA1 domain)-mApple-GRK2ct (G proteincoupled receptor kinase 2 c-terminus) expressing adenovirus before stimulation with dobutamine in the presence of Sal or vehicle control (NC), followed by fixation. C, Cell area. D, ANF expression. FOB is fold over basal. Cell size was quantified from n=1283–1383 cells from 6 separate wells each condition, repeated with 2 separate preparations of NRVMs. ANF was quantified from n=890–1348 cells from 6 wells each condition, repeated with 2 separate preparations of NRVMs. Scale bars is 40 μm. All data were analyzed with an ordinary one-way ANOVA with Tukey’s post-hoc test. NC indicates negative control.

Discussion

Our previous work led to development of a model where stimulation of Golgi β1-ARs by the membrane permeant agonist Dob produces a local pool of cAMP with privileged access to the EPAC/PLCε/mAKAP complex, generating DAG derived from PI4P depletion, activating PKD and mediating cardiac hypertrophy ^18,26. Unexpectedly, while low concentrations of Dob induced robust PI4P hydrolysis, saturating concentrations of Dob did not. Here we demonstrate that β2-AR activation opposes activation of Golgi PLCε-PKD pathway by two clinically relevant hypertrophic stimuli which could play an essential role in the ability of β2-ARs to limit cardiac hypertrophy. In cardiac myocytes we show that the mechanism involves agonist-driven internalization of β2-ARs where they couple to ERK activation via Gi-Gβγ at the endosome which in turn inhibits prohypertrophic PLCε signaling at the Golgi. Heart failure is often associated with up-regulation of Gi and enhanced β2-AR-Gi signaling ^57,58 which may be a compensatory mechanism to overcome the detrimental effects induced by circulating mediators and sympathetic neurotransmitters, many of which are likely to signal through PLCε.

Compartment specific ERK signaling in the heart.

Extensive studies suggest ERK is a key player in regulating cardiomyopathy. ERK signaling can produce both the beneficial and deleterious effects depending on the context. Transgenic mice overexpressing activated ERK are reported to show maladaptive ⁵⁹ or adaptive ⁶⁰ hypertrophic responses, while other studies utilizing mice with ERK deletion mice suggest ERK1/2 signaling may be dispensable in pathologic cardiac hypertrophy ^61,62. In addition, an in vitro study indicates that ERK1/2 activation is required for the ANF mediated antihypertrophic response ⁶³. It is now becoming clear that the signaling cascades and outcomes downstream of ERK largely depend on the subcellular locations of phosphorylated ERK1/2 ⁶⁴. Therefore, a more sophisticated dissection of ERK signaling mechanisms is required.

ERK is activated downstream of multiple GPCRs which are themselves spatially compartmentalized ⁶⁵. GPCRs inside the cell activate G proteins and produce local cAMP accumulation to initiate distinct subcellular signaling events ^66,67. G proteins including Gs, Gq, Gi and Gβγ subunit or β-arrestin-mediated signaling pathways activate ERK cascades. Different ligands cause different subcellular destinations of activated ERK via their preference to shift GPCR coupling to G-proteins or β-arrestin ^68,69. Recently, it has been shown that after ligand-stimulated β2-ARs endocytosis, endosomal ERK signaling is activated by an endosome-localized active Gα_s to subsequently stimulate nuclear ERK activity to control gene expression in HEK cells ⁵². Our studies support the idea that in AVMs, β2-AR-induced accumulated ERK signaling requires receptor endocytosis and endosomal Gβγ released from Gi. This endosome-receptor initiated ERK signaling serves as a repressor for PLCε which has been implicated in mediating cardiac hypertrophy. This study provides a functional role for this endosomal β2-AR-ERK signaling axis in preventing the development of cardiac hypertrophy.

PKD signaling in cardiac remodeling

Cardiac-specific deletion of PKD improves cardiac function in response to pressure overload or angiotensin II signaling ⁷⁰, making it a promising therapeutic target to treat heart failure resulting from maladaptive cardiac hypertrophic signaling. We previously demonstrated that DAG generated from PI hydrolysis by perinuclear scaffolded PLCε activates PKD in close proximity to the nucleus, where it phosphorylates HDAC to regulate hypertrophic gene expression ¹⁸. Our data suggest that acute activation of β2-ARs can protect against pathological cardiac hypertrophy through inhibition of phosphorylation of nuclear PKD activation and perhaps other kinases such as CamKII. Hence, currently clinically effective β2-AR agonists can serve as potent PKD repressors. In addition to the role of PKD in cardiac hypertrophy and fibrosis, it is also a critical signaling molecule in cancer-associated functions such as migration, cell proliferation, or survival ⁷¹. Thus, our observations could extend to other pathophysiological systems.

Molecular mechanisms underlying the cardioprotective effects of β2-AR signaling have been investigated previously. Gi-dependent-PI3K-AKT signaling downstream of β2-ARs contributes to its cardioprotective effects through prevention of apoptosis ⁷². In addition, β2-ARs can mediate EGFR and PDGFR transactivation, promoting cardiomyocyte survival ⁷³. These studies focused on cardiomyocyte apoptosis. We propose that inhibition of hypertrophic PLCε signaling is an additional mechanism for β2-AR dependent cardiac protection against detrimental cardiac remodeling.

Combined therapy with a β2-AR agonist and a selective β1-AR blocker.

β-blockers (carvedilol, bisoprolol, and metoprolol) are clinically used to treat patients with heart failure by blocking detrimental β1-AR-G protein signaling, although their complete mechanism of action is not fully understood. More recently, it has been suggested that effective β-blockers tend to be hydrophobic which allows them to access the internal pools of β1-ARs ^26,35. As a non-selective β-blocker, carvedilol has been to shown to produce superior clinical outcomes relative to other β-blockers ⁷⁴, though the mechanism is unclear. Carvedilol has been suggested to activate β2-ARs and ERK signaling ⁷⁵. It has been proposed that carvedilol might have dual therapeutic benefits in heart failure by preventing the deleterious signaling from β1-ARs and promoting the beneficial effects of β2-ARs ⁴. This therapeutic concept combining β1-AR blockade with a β2-AR agonist has been previously proposed and investigated in a coronary ligation model mice and rats where the combination treatment was more effective at preserving cardiac function that β1 blockade alone ^72,76,77. Our data indicate that in addition to effects on vascular tone selective β2-ARs ligands can act directly through cardiac myocytes to improve cardiac function by preventing the deleterious PLCε stimulated by chronically elevated catecholamines and angiotensin.

Isoproterenol paradox

Isoproterenol (Iso), a non-selective and relatively membrane impermeant β1-AR and β2-AR agonist, does not induce acute PLCε activation ²⁹, but with chronic exposure, it causes cardiac hypertrophy in vitro and vivo. One possible explanation for this discrepancy could be with acute Iso treatment, β2-AR activation counteracts signaling from β1-ARs or AngII. However, with chronic exposure to strong sustained stimulation by Iso (or other full agonists including norepinephrine and epinephrine), β2-ARs, while not downregulated, may desensitize leading to the loss of their protective signaling. Alternatively, Iso may depend on a different plasma membrane mediated hypertrophic pathway. These speculations need to be experimentally verified.

Limitations to this study.

While we have defined a pathway for β2-AR dependent inhibition of PLCε as a mechanism for inhibition of hypertrophy development, this does not directly show that this pathway is responsible for β2-AR dependent inhibition of hypertrophy in vivo. The study also does not show that salmeterol/β2-AR dependent protection in vivo is due to direct actions on cardiac myocytes. In our in vivo experiments it remains possible β2-AR actions in the vasculature, for example, could be responsible for the observed effects. Definitive demonstration of the in vivo significance of β2-AR-PLCε inhibition in cardiac hypertrophy protection will require more detailed in vivo investigation in knockout and transgenic mouse models.

Novelty and Significance.

What is Known

• β2-adrenergic receptor signaling is protective against cardiac hypertrophy.

• Activation of subcellular PLC (phospholipase C) at the Golgi apparatus stimulates cardiac hypertrophy through stimulation of local cAMP (cyclic adenosine monophosphate) production that has preferential access to PLCε located at the Golgi-nuclear envelope interface.

• β1 adrenergic receptors in the Golgi apparatus activate PLCε at the Golgi apparatus but cell surface β1 adrenergic receptors cannot.

What New Information Does this Article Contribute.

• β2 adrenergic receptors inhibit PLCε at the Golgi apparatus through a mechanism that requires β2 adrenergic receptor internalization and signaling to ERK (extracellular regulated kinase) from endosomes.

• Activation of β 2adrenergic receptors in endosomes prevents β1 adrenergic receptor-mediated stimulation of cardiomyocyte hypertrophy.

• Phospholipase Cε at the Golgi apparatus integrates both positive and negative inputs into hypertrophic signaling pathways.

The mechanisms for β2 adrenergic receptor dependent protection against development of cardiac hypertrophy have not been well defined. Internalization GPCRs (G protein-coupled receptors) including the β2 adrenergic receptor has emerged as a alternate signaling mechanism that produces signals distinct from those at the plasma membrane. In these studies, a mechanism was identified that likely contributes to the ability of β2 adrenergic receptors to protect against cardiac hypertrophy and heart failure. This supports that the use of β2 adrenergic receptor selective agonists for heart failure therapy deserves further investigation.

Non-standard abbreviations and Acronyms.

GPCR

G protein-coupled receptor

βAR

β-adrenergic receptor

PLC

phospholipase C

PIP2

phosphatidylinositol 4,5 bisphosphate

PI4P

phosphatidylinositol 4-phosphate

DAG

diacylglycerol

IP3

inositol 1,4,5-trisphosphate

RTK

receptor tyrosine kinase

Epac

Exchange Protein Activated by cAMP

mAKAP

muscle-specific A kinase anchoring protein

IP2

inositol 1,4-bisphosphate

PKD

protein kinase D

OCT

organic cation transporter

epi

epinephrine

NE

norepinephrine

FAPP

four phosphate adapter protein

PH

pleckstrin homology domain

HDAC

histone deacetylase

AVM

adult ventricular myocyte

NRVM

neonatal rat ventricular myocyte

WGA

wheatgerm agglutinin

ANF

atrial natriuretic factor

GRK2CT

C-terminus of G protein coupled receptor kinase 2

ERK

extracellular regulated kinase

TAC

transaortic constriction

Sal

salmeterol

Dob

dobutamine

Iso

isoproterenol

cpTOME-AM

8-(4-chlorophenylthio)-2-O-methyl-cAMP-acetoxymethyl ester

nDKAR

nuclear D kinase activity reporter

ATII

angiotensin II

PTX

pertussis toxin

EGFR

epidermal growth factor receptor

EGF

epidermal growth factor

CamKII

calcium-calmodulin dependent protein kinase II

PDGFR

platelet derived growth factor receptor.