β₁-Blockade Prevents Post-Ischemic Myocardial Decompensation Via Sphingosine-1 Phosphate Signaling

Abstract

Background

Although β-blockers increase survival in patients with heart failure (HF), the mechanisms behind this protection are not fully understood, and not all patients with HF respond favorably to them. We recently demonstrated that, in cardiomyocytes, a reciprocal down-regulation occurs between β₁-adrenergic receptors (ARs) and the cardioprotective sphingosine-1-phosphate (S1P) receptor-₁ (S1PR₁).

Objectives

The authors hypothesized that, in addition to salutary actions due to direct β₁AR-blockade, agents such as metoprolol (Meto) may improve post-myocardial infarction (MI) structural and functional outcomes via restored S1PR₁ signaling, and sought to determine mechanisms accounting for this effect.

Methods

We tested the in vitro effects of Meto in HEK293 cells and in ventricular cardiomyocytes isolated from neonatal rats. In vivo, we assessed the effects of Meto in MI wild-type and β₃AR knockout mice.

Results

Here we report that, in vitro, Meto prevents catecholamine-induced down-regulation of S1PR₁, a major cardiac protective signaling pathway. In vivo, we show that Meto arrests post-MI HF progression in mice as much as chronic S1P treatment. Importantly, human HF subjects receiving β₁AR-blockers display elevated circulating S1P levels, confirming that Meto promotes S1P secretion/signaling. Mechanistically, we found that Meto-induced S1P secretion is β₃AR-dependent because Meto infusion in β₃AR knockout mice does not elevate circulating S1P levels, nor does it ameliorate post-MI dysfunction, as in wild-type mice.

Conclusions

Our study uncovers a previously unrecognized mechanism by which β₁-blockers prevent HF progression in patients with ischemia, suggesting that β₃AR dysfunction may account for limited/null efficacy in β₁AR-blocker–insensitive HF subjects.

INTRODUCTION

Signaling through myocyte G-protein-coupled receptors (GPCRs), particularly β-adrenergic receptors (β₁, β₂, and β₃ARs), regulates the rate and force of contraction in response to increased workload (1). However, after myocardial infarction (MI), this signaling is profoundly altered, and cardiac βAR dysregulation and desensitization largely account for chronic post-MI decompensation (2). Accordingly, βAR-blockers attenuate the noxious effects of sympathetic catecholamines on the heart and prevent further βAR down-regulation. These compounds remain crucial to improving survival and preventing further left ventricular (LV) decompensation in patients who have experienced an ischemic event (1,3). Clinical β-blocker use is unequivocally proven to reduce oxygen consumption, prevent cardiac adverse remodeling, blunt cardiomyocyte apoptosis, inhibit βAR down-regulation, and reduce the risk of fatal arrhythmias (4). Yet, whether correcting βAR molecular perturbations justifies all mechanisms of β-blocker-induced cardioprotection in HF is a relevant question, as additional mechanisms appear to participate. Moreover, because not all patients respond to β-blockers, the overall clinical experience raises the question of how exactly β-blockers attenuate progression of HF after MI. Furthermore, delineating whether distinct, clinically-useful β-blockers have different mechanisms may explain specific patient responses and lead to fine-tuning of personalized HF therapy.

β-Blockers, such as alprenolol and carvedilol, act as classical receptor antagonists. However, they can also stimulate signaling pathways in a G protein–independent fashion (5). Moreover, selective β₁AR-blockers, such as metoprolol (Meto) or nebivolol, can also promote β₃AR up-regulation and/or activity in the heart, thus enhancing the cardioprotective signaling of this βAR subtype (6–8). Thus, answering these mechanistic questions is important, and may help tailor β-blockade therapies for patients with HF. Herein, we evaluated the role of sphingosine-1-phosphate receptor 1 (S1PR₁), a key GPCR expressed in the heart, which mediates the cardioprotective effect of sphingosine-1-phosphate (S1P), a natural lysophospholipid that appears tightly linked to β₁AR-signaling (9,10). In fact, β₁AR down-regulation can occur after S1P stimulation, whereas S1PR₁ down-regulation can be triggered by isoproterenol (ISO) treatment (9). This direct receptor cross-talk is also physiologically relevant, as these GPCRs interact and show reciprocal down-regulation in a rat model of post-MI HF (9). For these reasons, we hypothesize that, aside from the protective effects of β-blockade against catecholamine damage or βAR regulation in the myocardium (11), there are additional beneficial effects involving S1P and S1PR₁ signaling.

Methods

AGONISTS AND INHIBITORS

S1P was purchased from Cayman Chemicals (26993-30-6, Ann Arbor, Michigan); ISO (I6504), BRL 37344 (B169), and Meto (M5391) were all purchased from Sigma (Sigma-Aldrich, St. Louis, Missouri).

CELL CULTURE AND TRANSFECTION

Neonatal rat ventricular cardiomyocytes (NRVMs) were isolated from the hearts of 1- to 2-day-old rats, as previously described (12). HEK293 cells obtained from the American Type Culture Collection (Manassas, Virginia) were transfected as previously described (9).

CELL FRACTIONATION AND WESTERN BLOT ANALYSIS

Immunoblotting was performed as previously described (12). Protein levels of S1PR₁ (Y080010, ABM Inc, Richmond, Canada; 1: 1,000), SphK1 (sc-48825, Santa Cruz Biotechnology, Dallas, Texas; 1: 1,000), SphK2 (ab37977, Abcam, Cambridge, Massachusetts, 1: 1,000), β₃AR (PAB8502, Abnova, Walnut, California, 1: 1,000); glyceraldehyde-3-phosphate dehydrogenase (sc-32233, 6C5, Santa Cruz Biotechnology, Dallas, Texas, 1: 2,000) were assessed. Plasma membrane proteins were isolated from NRVMs, as previously described (12).

ANIMAL MODELS AND EXPERIMENTAL PROCEDURES

All animal procedures were performed in accordance with the guidelines of the Institutional Animal Care and Use Committee of Temple University School of Medicine. For in vivo experiments, we used wild-type C57BL/6 mice and global β₃AR knockout (KO) mice (13). All animals (females and males, 9 to 12 weeks of age) were bred and maintained on a C57BL/6 background.

The β₃AR KO genotype was assessed using specific primers: wild-type (WT) forward 5′ GTTGCGAACTGTGGACGTCAGTGG 3′; KO forward 5′ CGCATCGCCTTCTATCGCCTTCTTG 3′; and KO/WT reverse (common) 5′ AATGCCGTTGGCGCTTAGCCAC 3′

Surgically-induced MI was performed as previously described (14). Seven days post-MI, mice were randomly assigned to 1 of the following groups: MI; MI + Meto; MI + S1P; and MI + Meto/S1P. Meto was administered in drinking water (250 mg/kg/day) (15), whereas S1P (10 μM) was dissolved in phosphate-buffered saline (PBS) and was continuously infused subcutaneously into mice via an osmotic mini-pump (ALZET, DURECT Co., Cupertino, California) (16). Transthoracic echocardiography was performed to assess cardiac structure and function at baseline, 7 days, and 4 weeks post-MI, using a VisualSonics VeVo 2100 system (VisualSonics, Toronto, Canada), as previously described (12).

HISTOLOGY

Cardiac-specimens were fixed in 4% formaldehyde and embedded in paraffin. After deparaffinization and rehydration, 5-μm-thick sections were prepared, mounted on glass slides, and stained with 1% Sirius Red in picric acid (Sigma-Aldrich, St. Louis, Missouri) to detect interstitial fibrosis, as previously described (12). Terminal deoxynucleotidyl transferase-mediated dUTP nick-end labeling (TUNEL) was performed as previously described (12).

ANALYSIS AND STATISTICS

All procedures in mice were performed in a blinded fashion. Moreover, all LV tissue and blood samples derived from them were also coded, and the operators were blind to these codes until data quantification was complete. Data are expressed as means ± SE. Statistical significance was determined by a Student t-test or Mann-Whitney U test when sample size was <10. For multiple comparisons, 1-way analysis of variance, followed by Bonferroni post hoc correction, was performed. All data were analyzed using GraphPad Prism software version 6 (GraphPad Software, La Jolla, California).

RESULTS

METO PREVENTS ISO-INDUCED IN VITRO S1PR1 MEMBRANE DOWN-REGULATION

The βAR agonist ISO induces robust internalization and/or desensitization of the S1P receptor (S1PR₁), leading to inhibition of its protective signaling, which is dependent on down-regulation of β₁AR (9,10). Here, we tested whether selective β₁AR-blockade via Meto affects this reciprocal β₁AR-S1PR₁ down-regulation. To this end, we first treated neonatal rat ventricular myocytes (NRVMs) with ISO (1 μM) and S1P (250 nM) (each for 30 min), which resulted in a robust S1PR₁ plasma membrane down-regulation (Figure 1A). Next, we pre-treated NRVMs with Meto (10 μM) for 30 min, and then stimulated them with ISO (1 μM) or S1P (250 nM) for an additional 30 min. Pre-treatment of cells with Meto prevented ISO-, but not S1P-induced S1PR₁ down-regulation (Figure 1A).

Figure 1. Meto Abolishes β₁AR-Dependent Sarcolemmal S1PR₁ Internalization.

(A) Representative immunoblots (upper panels) and densitometric quantitative analysis (lower panel) of multiple (n = 3) independent experiments to evaluate S1PR₁ in crude plasma membrane preparations from NRVMs. Cells were nonstimulated (Ns) or stimulated with ISO (1 μM) or S1P (250 nM) for 30 min. Before ISO or S1P stimulation, a group of cells was pre-treated with Meto (10 μM) for 30 min. Ponceau Red staining was used to assess total protein loading. *p < 0.05 versus Ns. (B) Representative immunofluorescence images (scale bar: 10 μm) of HEK293 cells overexpressing β₁AR-Flag and S1PR₁-GFP (green). S1PR₁-GFP internalization over a 30-min time course (0 to 30 min) in cells treated with ISO (1 μM) or S1P (250 nM) is shown. Before ISO stimulation, a group of cells was pre-treated with Meto (10 μM for 30 min). β₁AR = beta-1 adrenergic receptor; GFP = green fluorescent protein; ISO = isoproterenol; Meto = metoprolol; NRVM = neonatal rat ventricular myocytes; S1P = sphingosine-1-phosphate; S1PR₁ = sphingosine-1-phosphate receptor-1;

To confirm these observations, we used HEK293 cells over-expressing both FLAG-tagged β₁AR (β₁AR-FLAG) and green fluorescent protein tagged S1PR₁ (S1PR₁-GFP) (Online Figure 1A) and checked for S1PR₁ internalization. Using confocal microscopy, we performed a time course evaluation (0 to 30 min), assessing the impact of Meto on ISO- and S1P-induced S1PR₁ internalization. In line with the NRVM data, we observed that ISO and S1P induced massive receptor internalization, and that Meto prevented ISO-induced effects (Figure 1B). Together, these data indicate that β₁AR-blockade prevents β₁-dependent S1PR₁ down-regulation in vitro.

BOTH METO AND S1P IMPROVE CONTRACTILITY AND PREVENT LV REMODELING AFTER MI

For in vivo support of the findings described in the preceding section, we used a mouse MI model because both β₁AR and S1PR₁ are down-regulated in infarcted myocardium (9). Five groups of randomized mice were studied: 1) sham-operated; 2) untreated MI; 3) MI treated with Meto (250 mg·kg⁻¹·day⁻¹); 4) MI treated with S1P (10 μM); and 5) MI treated with both Meto and S1P (Meto + S1P). LV function and dimensions were evaluated by echocardiography before MI, and then 1 (Online Figures 2A and 2B) and 4 weeks after MI (study termination) (Figure 2A). Four weeks after MI, echocardiographic analysis revealed that LV ejection fraction (EF) was markedly decreased, and LV dimensions and dilation were substantially enlarged in untreated post-MI mice, as opposed to sham-operated animals (Figures 2B–2D). However, Meto administration resulted in improved LV contractility (Figure 2B) and reduced LV dilation (Figure 2C–2D), as compared with untreated post-MI mice. S1P treatment also significantly increased EF and reduced the LV diameter versus control MI mice (Figure 2B–2D). Although Meto and S1P were compatible for coadministration, we did not observe any synergistic effects of these therapies (Figures 2B–2D). Interestingly, Meto and S1P, either alone or in combination, decreased heart rate (Figure 2E). Of note, the percentage of infarct size (measured at the end of the study) was similar among all study groups (Figure 2F).

Figure 2. Meto Ameliorates MI-Induced Cardiac Dysfunction and S1PR₁ Down-Regulation.

(A) Overall design of 4-week study of MI in mice. Echocardiography was performed at day 0 and then MI were surgically induced. Sham-operated mice were used as controls. After 1 week, echocardiography was performed in order to assess that all mice had similar cardiac dysfunction. Mice (total n = 12 to 22 per group) were randomized for treatments: MI (controls), MI + Meto (250 mg/kg/day), MI + S1P (10 μM) and MI + Meto/S1P. Four weeks post-MI (3 weeks post-treatment), echocardiography was performed and animals euthanized for tissue harvesting. (B–C) Dot plots showing the echocardiographic analysis performed at the end of the study period (4 weeks post-MI) of (B) left ventricular ejection fraction (EF, %) and (C) left ventricular internal diameter at diastole (LVIDd) of individual mice from each of the groups. (D) Representative images of whole hearts from mice. (E) Dot plots showing echocardiographic analysis of heart rate of individual mice from the each of the groups. (F) Dot plots (upper panel) and representative panels of picrosirius red staining (lower panels; scale bar: 500 μm) showing the percentage of infarct size in the MI, MI + Meto, MI + S1P, and MI + Meto/S1P mice (n = 6). For (B, C, and E) *p < 0.05 versus sham; #p < 0.05 versus MI. Echo = echocardiogram; MI = myocardial infarction. Other abbreviations as in Figure 1.

BOTH METO AND S1P REDUCE POST-MI CARDIAC APOPTOSIS AND FIBROSIS

Cardiac apoptosis and fibrosis are 2 major hallmarks of post-MI cardiac remodeling (12,17). Therefore, we compared the impact of Meto and/or S1P on these endpoints. As expected, cardiomyocyte cell death and collagen deposition were significantly increased in post-MI untreated mice versus sham animals (Figures 3A and 3B). These changes were accompanied by the up-regulation of profibrotic genes encoding connective tissue growth factor (CTGF), and type 1 (Col1) and type 3 collagen (Col3) (Online Figures 2A–2C). Importantly, Meto treatment induced a marked reduction of both cardiac fibrosis and apoptosis (Figures 3A and 3B and Online Figures 3A–3C). Conversely, S1P treatment significantly prevented MI-dependent apoptosis (Figure 3A), but had no effect on cardiac fibrosis (Figure 3B and Online Figures 3A–3C). Moreover, combined Meto + S1P treatment did not improve upon Meto alone in preventing the MI-induced increase in apoptotic rate and fibrosis (Figures 3A and 3B and Online Figures 3A–C).

Figure 3. Effects of Meto and S1P on Post-Ischemic Cardiac Fibrosis and Apoptosis.

(A) Dot plots showing myocyte cell death via TUNEL staining from the following groups (n = 5 to 9): sham; MI; MI + Meto (250 mg/kg/d); MI + S1P (10 μM); and MI + Meto/S1P. (B) Representative images (scale bar: 200 μm) and (C) quantitative data showing the percentage of cardiac fibrosis via picrosirius red staining from each of the groups (n = 5 to 8). (D) S1P serum levels measured by ELISA assay in each of the groups (n = 7 to 11). (E) Representative immunoblots (upper panels) and densitometric quantitative analysis (lower panel) of multiple independent (n = 3 to 4) experiments to evaluate SphK1 protein levels in total cardiac lysates from each of the groups. GAPDH levels were used for protein loading controls. For (A, C–E) *p < 0.05 versus sham; #p < 0.05 versus MI. ELISA = enzyme-linked immunosorbent assay; GAPDH = glyceraldehyde-3-phosphate dehydrogenase; SphK1 = sphingosine kinase-1; TUNEL = terminal deoxynucleotidyl transferase dUTP nick end labeling. Other abbreviations as in Figures 1 and 2.

The lack of any additional effect of Meto + S1P compared to Meto alone prompted us to further investigate the effects of Meto and S1P. In particular, we measured circulating S1P serum levels in all of our study groups. We found that although S1P in serum was significantly reduced in MI untreated mice compared with sham, Meto treatment was able to restore circulating S1P to sham levels (Figure 3D). As expected, S1P administration normalized circulating S1P levels, both when administered alone or with Meto (Figure 3D). Importantly, the impact of Meto on circulating S1P levels was paralleled by the rescued cardiac sphingosine kinase 1 (SphK1) expression (Figure 3E) and activity (Online Figure 4A). SphK1 is the enzyme involved in S1P production and consequent S1PR₁ activation (18). Of relevance, Meto obviated the S1P secretion deficiency found in post-MI WT mice (Online Figure 4B).

METO RESCUES B₃AR-DEPENDENT SPHK1 EXPRESSION IN INFARCTED HEARTS

To date, there is no molecular mechanism explaining how β-blockers lead to S1P production. Normally, β₃ARs, which are expressed at low levels in the myocardium, can be up-regulated and activated by Meto (6,19). Moreover, due to metabolic effects (1), the β₃AR is implicated in regulation of SphK1 expression and S1P release, without affecting SphK2 (20). Thus, we tested whether β₃AR signaling accounts for Meto-induced post-MI cardioprotection using β₃AR KO and WT mice. At 1 week after MI, all animal groups had similar LV dysfunction (Online Figures 5A and 5B). We then randomized mice to placebo or Meto treatment (Figures 4A and 4B). At the end of the study period, we found significant down-regulation of cardiac β₃AR protein levels in post-MI WT mice versus sham-operated animals (Figure 4C). Notably, Meto obviated this change (Figure 4C). As expected, β₃AR protein levels were not detectable in β₃AR KO mice (Figure 4C). Consistent with our previous studies (9), cardiac SphK1, SphK2 protein levels, and circulating S1P were all reduced after MI in WT mice with respect to sham (Figures 4D and 4E and Online Figure 6). Treatment with Meto restored SphK1 expression (Figures 4D and 4E and Online Figure 6). Intriguingly, SphK1 expression and S1P secretion were very low in β₃AR KO mice, regardless of MI and/or Meto status. Of note, SphK2 expression was similar in WT and in β₃AR KO mice. In aggregate, this evidence strongly supports the contention that β₃AR per se plays a crucial role in basal expression levels of SphK1 in the heart (Figures 4D and 4E).

Figure 4. β₃AR-Dependent Expression/Activation of Cardiac SphK1 and S1P Secretion in MI Mice Following Meto Treatment.

(A) Representative genotype analyses of mice (in duplicate) with WT (400 bp), or deleted (600 bp) β₃AR alleles. (B) Overall design of 4-week study in WT and β₃AR KO mice. Echocardiography was performed at day 0, followed by surgically-induced MI. Sham-operated mice were used as controls. After 1 week, echocardiography was performed in order to assess that all mice had similar cardiac dysfunction, and they were then randomized for treatments: MI (controls) and MI + Meto. (C) Representative immunoblots (upper panels) and densitometric quantitative analysis (lower panel) of multiple independent experiments (n = 10 to 13) to evaluate β₃AR protein levels in total cardiac lysates from the following groups of WT mice: sham; MI; and MI + Meto. Total cardiac lysates from β₃AR KO mice were used as a negative control. GAPDH levels were used for protein loading. *p < 0.05 versus sham. (D) Representative immunoblots (upper panels) and densitometric quantitative analysis (lower panel) of multiple independent experiments (n = 6) to evaluate SphK1 protein levels in total cardiac lysates from the following groups: WT (sham, MI, and MI + Meto) and β₃AR KO (sham, MI, and MI + Meto). GAPDH levels were used for protein loading controls. *p < 0.05 versus WT sham. (E) S1P blood serum levels measured by ELISA in the following groups (n = 7 to 10): WT (sham) and β₃AR KO (sham, MI, and MI + Meto). *p <0.05 versus WT sham. (F) Representative immunoblots (upper panels) and densitometric quantitative analysis (lower panel) of multiple (n = 4) independent experiments to evaluate SphK1 protein levels in NRVMs Ns or stimulated with ISO (1 μM) or BRL (1 μM) for 12 h. Before ISO or BRL stimulation a group of cells was pre-treated with Meto (10 μM for 30 min). *p < 0.05 versus Ns; #p < 0.05 versus ISO. KO = knockout; WT = wild-type. Other abbreviations as in Figures 1–3.

To confirm the role of β₃AR on SphK1 expression, we treated NRVMs with ISO (10 μM) or with the β₃AR agonist BRL 37344 (BRL, 1 μM) for 12 h. Before ISO or BRL stimulation, some groups of cells were also pre-treated with Meto (10 μM) for 30 min. Both ISO and BRL were able to induce SphK1 up-regulation, as compared with control, nonstimulated (Ns) cells (Figure 4F). More importantly, Meto pre-treatment robustly enhanced the ability of ISO and BRL to increase SphK1 expression. Meto alone enhanced SphK1 at levels comparable to those found in Ns cells (Figure 4F). Thus, Meto rescues β₃AR-dependent SphK1 expression through direct β₁AR blockade, making more catecholamines available for β₃AR activation and inhibitory G protein (G_i) coupling. In fact, pre-treatment of NRVMs with the G_i inhibitor pertussis toxin (PTX, 100 ng/ml for 18 h) prevented SphK1 up-regulation due to ISO or BRL (Online Figure 7).

METO-INDUCED IMPROVEMENT IN POST-MI HF DEPENDS ON BOTH β₃AR AND S1PR₁

The data presented so far suggests that S1PR₁ resensitization and β₃AR activation are important components of Meto-induced beneficial effects post-MI. To solidify this view, we compared the in vivo cardiac effects of Meto in post-MI WT and β₃AR KO mice. Meto improved cardiac function in post-MI WT mice, but not β₃AR KO mice (Figures 2B and 5A), despite a similar heart rate reduction in both groups (Figures 2E and 5B). Infarct sizes in β₃AR KO mice were similar under any experimental conditions (Online Figure 8), and no differences in cardiac function were observed between WT and β₃AR KO mice following MI (Figure 5A). Consistently, apoptosis and fibrosis were also similar in WT and β₃AR KO mice after MI (Figures 3A–3E). In contrast, whereas Meto reduced both of these endpoints in post-MI WT mice, it did not in β₃AR KO mice (Figures 3A–3C and 5C–5E). Finally, to confirm the role of S1PR1 in Meto-mediated cardioprotection, infarcted WT mice were treated with the S1PR1 antagonist W146. MI mice were randomized to 1 of the following groups: MI + vehicle (PBS + ethanol); MI + Meto (250 mg/kg/d); MI + W146 (8 μg/ml) (21); and MI + Meto/W146. W146 was administered via mini-osmotic pumps, as done with S1P. As expected, Meto prevented LV dysfunction and adverse remodeling (Online Figure 9), and these beneficial actions were prevented by W146 (Online Figure 9). Either vehicle or W146, taken alone had no sizable impact on post-MI decompensation. Thus, activation of the β3AR/S1PR1 axis accounts for the protection against MI afforded by Meto.

Figure 5. Lack of Meto-Dependent Post-MI Therapeutic Response in β₃AR KO Mice.

(A,B) Dot plots showing echocardiographic analysis of (A) left ventricular ejection fraction (EF, %) and (B) heart rates of individual mice from the following groups (n = 11 to 22): β₃AR KO (sham, MI, and MI + Meto). (C) Dot plots showing myocyte cell death via TUNEL staining from the following groups of mice (n = 5): β₃AR KO (sham, MI, and MI + Meto); (D,E). Representative images (scale bar: 200 μm) and quantitative data showing percentage of cardiac fibrosis via picrosirius red staining from the β₃AR KO (sham, MI, and MI + Meto) mouse groups (n = 5). For (A–C, E) *p < 0.05 versus sham. Abbreviations as in Figures 1–4.

CHRONIC TREATMENT WITH METO INCREASES CIRCULATING S1P LEVELS IN HUMAN HF PATIENTS

To add clinical relevance to our findings, we investigated whether β₁-blocker treatment can affect circulating S1P levels in human patients with HF. Patients treated with β₁AR antagonists and a similar HF group not treated with β-blockers (see Online Table 1 for specific patient characteristics) were used to asses S1P levels via an enzyme-linked immunosorbent assay (ELISA). Interestingly, clinical β₁-blockade was characterized by significantly higher blood S1P levels compared with the β-blocker-naïve control population (Online Figure 10).

DISCUSSION

Our study shows that the selective β₁-blocker Meto rescues S1PR₁- dependent signaling by preventing β₁AR-induced S1PR₁ down-regulation, which attenuates HF progression after MI in mice. This effect requires up-regulation of β₃AR signaling, which leads to increased circulating S1P levels and cardiac S1PR₁ stimulation, ultimately inducing prosurvival signals in the ischemic myocardium. Accordingly, the salutary actions of Meto are completely lost in β₃AR KO mice. Hence, the present study documents a previously unrecognized mechanism by which β₁-blockers, such as Meto, effectively improve survival and mitigate LV decompensation in mice with ischemic HF.

β₁-Blockers remain a forefront approach for treating acute or chronic HF of ischemic and nonischemic origin. They prevent direct cardiac catecholamine toxicity and promote βAR down-regulation in the failing heart, resulting in markedly reduced HF-related mortality (1). However, β₁-blockers are a heterogeneous class of drugs, and some appear to attenuate HF progression through mechanisms additive to βAR resensitization and prevention of catecholamine-induced damage. Uncovering these unknown modalities of protection may help tailor βAR-blocker therapies for each HF patient. In the present study, we focused on β₁ARs and S1PR₁s, 2 distinct, highly-expressed cardiac GPCRs, whose cross-talk appears to have important pathophysiological repercussions (9). In a rat model of post-MI HF, where S1P is significantly reduced and catecholamine secretion is increased, we previously found that β₁AR-hyperstimulation leads to marked S1PR₁ signaling dysregulation (9). Therefore, in this study, we tested whether, among other salutary actions, β₁AR-blockers are also able to tip the balance to favor S1PR₁ protective signaling. Accordingly, we show that Meto pre-treatment of cardiomyocytes in vitro abolishes ISO-induced S1PR₁ down-regulation, whereas no effect is detectable in presence of S1P stimulation, confirming that β₁AR activation could lead to S1PR₁ transactivation. We validated these in vitro findings in a relevant in vivo HF model in mice after MI. We found that Meto treatment rescued LV function and prevented further adverse LV remodeling, restoring, at least in part, S1PR₁ cardiac plasma membrane levels. S1PR₁ mediates the effect of S1P, which is a natural cardioprotective sphingolipid present at high concentrations in blood (9). Data showing that increasing cardiac S1PR₁ levels contributed to significantly reduce myocyte apoptosis and fibrosis in post-MI hearts treated with Meto is confirmed by the fact that combining S1P and Meto had beneficial effects similar to those exerted by Meto alone. In keeping with this, we found that chronic administration of this selective β₁-blocker was sufficient to restore circulating S1P levels, which were markedly reduced after MI. In turn, this increased bioavailability activated cardiac SphK1, which was ultimately responsible for the rise in S1P levels in the serum of Meto-treated infarcted mice. Of note, this rise was superimposable to that achieved after direct infusion of S1P in post-MI mice. Taken together, this evidence explains why combining S1P and Meto had no additional or synergistic effects. This conclusion not only points toward new and important mechanistic insights into the beneficial effects of β₁-blockade in ischemic HF, but also stresses the functional repercussions of a loss of cardiac S1P-S1PR₁ signals due to β₁AR-hyperstimulation after ischemic injury. In essence, our study represents a significant departure from previous in vitro and in vivo data showing that only direct S1PR₁ activation confers cardioprotection after an ischemic insult (22–26).

Another conceptual advancement offered by the present findings is the involvement of β₃ARs in Meto-induced protection involving the S1P-S1PR₁ axis. β₃ARs were recently proposed to be cardioprotective receptors, with nodal roles in cardiac metabolism involving regulation of lipolysis via enhanced Sphk1 activity (20). Our present findings significantly expand this view by showing that following MI, β₃AR expression is substantially down-regulated, along with markedly reduced SphK1 expression, and consequent impaired S1P secretion and worse LV remodeling. The dependence of SphK1/S1P signaling downstream of β₃AR activation is supported by the severe down-regulation of both SphK1 and S1P levels in β₃AR KO mice, and that these levels do not differ significantly from those found in WT mice after MI. Congruent with this, we observed similar LV dysfunction in post-MI WT and β₃AR KO mice. Nevertheless, the most striking finding is that Meto not only restores cardiac S1PR₁, but also prevents β₃AR down-regulation after MI, resulting in enhanced SphK1 activity and consequent S1P secretion. These biological effects were not seen in β₃AR KO mice. Meto did not consistently prevent HF progression in infarcted β₃AR KO mice. Thus, our data reveal that activating the β₃AR signal is a crucial and required step for β₁AR blockers (and likely other pharmacological interventions impinging on the β₁AR signal) because Meto does not improve the HF phenotype in post-MI β₃AR KO mice.

Although appearing somewhat modest, Meto or S1P treatment of mice with severe post-MI cardiac dysfunction resulted in a significant protective effect in terms of post-MI viable myocardium or LV function. Importantly, the treatments were started at 1 week post-MI and drugs were administered for just 3 weeks. Therefore, instituting Meto or S1P within the first week post-MI and/or prolonging the treatment up to 8 weeks or more might result in more pronounced effects, as previously demonstrated for S1P (27). These pending questions, along with possible dosing issues, should be addressed in large animals to more firmly establish the therapeutic potential of the present findings obtained in mice. Likewise, future studies to more finely dissect changes in S1PR₁ downstream signaling, particularly in β₃AR KO mice, are warranted.

All of the evidence reported here appears to be relevant to human HF. In fact, we report for the first time that patients with ischemic cardiomyopathy receiving treated with β₁-blockers (Online Table 1 and Online Figure 10) had significantly higher S1P serum levels. At this time, possible β₃AR up-regulation in these patients with ischemia receiving β₁-blockers can only be inferred from evidence obtained in post-MI WT mice. However, a previous report showed up-regulation of β₃AR protein levels in heart specimens from human subjects with either ischemic or dilated cardiomyopathy who received chronic treatment with β₁AR-blockers, digitalis, angiotensin-converting enzyme inhibitors, and diuretic agents (28). Overall, these data corroborate our view that it is not the HF condition per se or its etiology, but rather the pharmacological treatment used (i.e., β₁-blockers [or others]) that effectively increases cardiac β₃AR in these human HF subjects. Additional in-depth studies must be designed to directly address this important point. However, our study suggests another important translational consideration that may help in personalizing β₁AR-blocker therapy: the cohort of patients that do not respond or responds minimally to β₁-blockers may harbor abnormalities in β₃AR signaling.

In conclusion, our study documents a novel mechanism by which β₁AR-blockers induce a therapeutic response after MI: β₁AR blockade by Meto leads to up-regulation of β₃ARs, followed by S1P/S1PR₁-triggered protective and beneficial signaling. The latter event is followed by β₃AR-dependent S1P release outside the cardiomyocytes, which, in turn, activates the rescued S1PR₁ aligned on the sarcolemmal membrane of neighboring myocytes to ultimately trigger intracellular protective signaling (Central Illustration; Figure 6). Therefore, Meto acting on myocardial β₁ARs actually influences the signaling of 3 distinct receptors, both directly (S1PR1 down-regulation) and indirectly (up-regulation of β₃ARs and increased S1P secretion) (Central Illustration; Figure 6). The discovery of this new protective loop triggered by β₁AR blockers such as Meto advances our understanding of ischemic HF pathophysiology.

Central Illustration. β₁AR-Blocker–Dependent Activation of β₃AR/S1P Protective Signaling in Cardiomyocytes.

Following binding with the β₁AR, β₁AR-blockers prevent direct catecholamine toxicity on the cardiac myocytes. Circulating catecholamines then bind to β3ARs, leading to up-regulation and activation of SphK1, which phosphorylates sphingosine (Sph). The S1P generated is then secreted outside the myocytes, and directly activates S1PR1 in an autocrine and paracrine fashion, which, in turn, induces a therapeutic response in heart failure. β₁AR = beta-1 adrenergic receptor; β₁AR = beta-3 adrenergic receptor; CA = catecholamine; HF = heart failure; S1P = sphingosine-1-phosphate; S1PR = sphingosine-1-phosphate receptor; SphK1 = sphingosine kinase-1.

Figure 6. Scheme for Metoprolol-Dependent Induction of β₃AR and S1PR₁ Activity in Cardiomyocytes.

Following binding to the β₁AR, Meto prevents direct catecholamine toxicity on the myocardium and further β₁AR and S1PR₁ down-regulation (1). Circulating catecholamines then bind to β₃ARs (2), leading to up-regulation and activation of SphK1 (3), which phosphorylates the sphingolipid sphingosine (Sph) (4). The S1P generated is then secreted outside the myocytes and directly activates S1PR₁ in an autocrine and paracrine fashion (5), which, in turn, induces a therapeutic response in HF (6). Thus, blocking β₁AR positively affects all 3 receptor systems in failing cardiomyocytes. CA = catecholamine. Other abbreviations as in Figures 1 and 3.

Online Methods

Agonists and inhibitors

W146 was purchased by Cayman Chemicals (909725-62-8, Michigan, USA), Pertrussis Toxin (PTX) was purchased by Sigma (P7208, Sigma-Aldrich, St. Louis, MO, USA)

Confocal Microscopy

HEK293 expressing the S1PR₁-GFP and the WT β₁AR-Flag were visualized as previously described¹. The fluorescent data sets were visualized with a Zeiss 510 confocal laser scanning microscope and analyzed by LSM 510 software.

RNA extraction and Real-time PCR

Total RNA was isolated from NRVMs and from LV specimens with TRizol (Life Technologies, USA) accordingly to the company’s instructions. Real-time PCR was performed in duplicate on a CFX96 Real-Time PCR detection system (BioRad Lab. USA) using the SYBR green mix (BioRad Lab. USA) and specific primers for: CTGF (Fwd-5′-GGAAGACACATTTGGCCCAG-3′; Rev-5′-TAGGTGTCCGGATGCACTTT-3′); Col1a2 (Fwd-5′-TAGAAAGAACCCTGCTCGCA-3′; Rev-5′-CGGCTGTATGAGTTCTTCGC-3′) and Col3 (Fwd-5′-ATGAGGAGCCACTAGACTGC-3′; Rev-5′GGTCACCATTTCTCCCAGGA-3′). The expression levels of CTGF, Col1a2 and Col3 were normalized to the rRNA 18S. Specificity of PCR products was confirmed by melting curve and gel electrophoresis.

ELISA assay

Circulating and LV S1P levels were measured using a commercial kit (Echelon), according to manufacturer instructions and as previously described¹. About 1 mL and 5 mL of blood were collected from mice and humans respectively. Then the blood was centrifuged to 2,000 rpm a 15°C for 15 minutes and 25 μL of serum were used for the ELISA assay. About 1–2 mg of LV were lysed according to the kit instruction. The lysate was used for the assay.

Sphingosine Kinase activity assay

Cardiac SphK activity was measured using a commercial kit (Echelon), according to manufacturer instructions. About 1–2 mg of LV was lysed according to the kit instruction. The lysate was used for the assay.

In vivo S1PR₁ antagonism

Surgical induced MI was performed as previously described²³. Seven days post-MI, mice were randomly assigned to one of the following groups: MI, MI+W146 (8μg/mL) MI+Meto, MI+S1P and MI+Meto/S1P. Meto was administered in drinking water (250 mg/Kg/day)²⁴ while S1P (10 μM) was dissolved in PBS and was continuously infused subcutaneously into mice via an osmotic mini-pump (ALZET, DURECT Co., Cupertino, USA)²⁵.

Supplementary Material

1

Table 1. Characteristic of patients in the overall study population and in β₁-blockers–treated and non-treated HF patients

Supplementary Figure 1. Shown is co-expression of β₁AR-Flag (Red) and S1PR₁-GFP (green) in HEK293 cells (Scale bar: 50 μm).

10. Supplementary Figure 2.

Bar graphs showing the echocardiographic analysis performed after 1 week post-MI of (A) Left Ventricular (LV) Ejection fraction (EF, %); (B) heart rate (HR) of individual mice (n=7–12 mice per group) from the following group of treatment: Sham, MI, MI+Meto, MI+S1P and MI+Meto/S1P.

2. Supplementary Figure 3.

Bar graph showing quantitative data of quantitative RT-PCR experiments to evaluate myocardial A) CTGF, B) Col1a2, C) Col3 mRNA levels from the following groups (n=6–8 mice per group) of treatment: Sham, MI, MI+Meto, MI+S1P and MI+Meto/S1P mice. *, p<0.05 vs Sham;

3. Supplementary Figure 4.

Dot plots showing analysis of A) SphK activity and B) S1P levels measured by ELISA assay in LV lysates from the following groups: WT (Sham, MI and MI+Meto). n=10–13 mice per group for SphK activity and n=6 mice per group for S1P assay: *, p<0.05 vs Sham.

4. Supplementary Figure 5.

Bar graphs showing echocardiographic analysis of (A) Left ventricular (LV) Ejection fraction (EF, %) and (B) heart rate (HR) of individual mice after 1 week of MI from the following groups (n=17–19 mice per group): β₃AR KO (Sham, MI and MI+Meto).

5. Supplementary Figure 6.

Representative immunoblots (upper panels) and densitometric quantitative analysis (lower panel) of multiple independent experiments (n=4 mice per group) to evaluate SphK2 protein levels in total cardiac lysates from the following group: WT (Sham, MI and MI+Meto) and β₃AR KO (Sham, MI and MI+Meto). GAPDH protein levels were used to check protein loading. *, p<0.05 vs WT Sham.

6. Supplementary Figure 7.

Representative immunoblots (upper panels) and densitometric quantitative analysis (lower panel) of multiple (n=3) independent experiments to evaluate SphK1 protein levels in neonatal rat ventricular myocytes (NRVMs) either unstimulated (Ns) or stimulated with Isoproterenol (ISO, 1 μM) or BRL37344 (BRL, 1 μM) for 12 hours. Prior to ISO or BRL stimulation a group of cells was pre-treated with Pertussis Toxin (PTX, 100 ng/mL) for 18 hours. *, p<0.05 vs Ns.

7. Supplementary Figure 8.

Dot plots showing percentage of infarct size in post-MI β₃AR KO mice (n=5 mice per group) treated or not with Meto (250 mg/Kg/d).

8. Supplementary Figure 9.

A–B) Dot plots showing echocardiographic analysis of (A) Left ventricular (LV) Ejection fraction (EF, %) and (B) heart rate (HR) of individual mice from the following groups (n=7–8 mice per group): WT (MI, MI+W146, MI+Meto and MI+Meto/W146). *, p<0.05 vs MI.

9. Supplementary Figure 10.

Dot plots showing S1P blood serum levels measured by ELISA assay in human HF-patients treated (n=19) or not (n=20) with β₁-blockers. *, p<0.05 vs HF.

PERSPECTIVES.

COMPETENCY IN MEDICAL KNOWLEDGE

βAR-blockers are a key component of therapy for patients with HF, but some patients do not respond to this line of treatment. β₃-dependent S1P activation is crucial for cardiac protection, and β₃ signaling may underlie the differential responsiveness to this class of drugs.

TRANSLATIONAL OUTLOOK

Patients with HF who are resistant to β-blockade may harbor structural or functional β₃ variants, and therapies that stimulate these selectively might extend the benefit of to a broader segment of the patient population.

ABBREVIATIONS AND ACRONYMS

βAR

β-adrenergic receptor

GPCR

G-protein coupled receptor

HF

heart failure

ISO

isoproterenol

Meto

metoprolol

MI

myocardial infarction

S1P

sphingosine-1-phosphate

S1PR₁

sphingosine-1-phosphate receptor 1

SphK1

sphingosine kinase 1