Selective Coupling of the S1P₃ Receptor Subtype to S1P-mediated RhoA Activation and Cardioprotection

Abstract

Sphingosine-1-phosphate (S1P), a bioactive lysophospholipid, is generated and released at sites of tissue injury in the heart and can act on S1P₁, S1P₂, and S1P₃ receptor subtypes to affect cardiovascular responses. We established that S1P causes little phosphoinositide hydrolysis and does not induce hypertrophy indicating that it does not cause receptor coupling to G_q. We previously demonstrated that S1P confers cardioprotection against ischemia/reperfusion by activating RhoA and its downstream effector PKD. The S1P receptor subtypes and G proteins that regulate RhoA activation and downstream responses in the heart have not been determined. Using siRNA or pertussis toxin to inhibit different G proteins in NRVMs we established that S1P regulates RhoA activation through Gα₁₃ but not Gα₁₂, Gα_q, or Gα_i. Knockdown of the three major S1P receptors using siRNA demonstrated a requirement for S1P₃ in RhoA activation and subsequent phosphorylation of PKD, and this was confirmed in studies using isolated hearts from S1P₃ knockout (KO) mice. S1P treatment reduced infarct size induced by ischemia/reperfusion in Langendorff perfused wild-type (WT) hearts and this protection was abolished in the S1P₃ KO mouse heart. CYM-51736, an S1P₃-specific agonist, also decreased infarct size after ischemia/reperfusion to a degree similar to that achieved by S1P. The finding that S1P₃ receptor- and Gα₁₃-mediated RhoA activation is responsible for protection against ischemia/reperfusion suggests that selective targeting of S1P₃ receptors could provide therapeutic benefits in ischemic heart disease.

Graphical abstract

1. Introduction

Restoration of blood flow after an ischemic episode (e.g. myocardial infarct) is necessary to prevent catastrophic heart failure but reperfusion can itself increase cardiomyocyte death, a process referred to as reperfusion injury [1]. Previous studies have shown that the circulating bioactive lysophospholipid sphingosine-1-phosphate (S1P) is endogenously released in response to cardiac injury [2, 3] and that S1P helps to protect the heart from the oxidative damage that leads to reperfusion injury [4–7]. S1P is a high affinity ligand for five G protein-coupled receptor (GPCR) subtypes denoted S1P_1–5 [8, 9]. The S1P_1–3 receptor subtypes are expressed in cardiomyocytes. S1P₁ is the predominant subtype expressed in the heart as well as in cardiomyocytes and it exclusively couples to Gα_i [10, 11]. Coupling to Gα_i leads to inhibition of cyclic AMP formation and accounts for the ability of S1P to decrease cardiac contractility [9, 10, 12, 13]. The S1P₂ and S1P₃ receptors can couple to Gα_i signaling in the heart [10] but these subtypes are, in addition, able to couple to Gα_q, Gα₁₂, and Gα₁₃ [9, 14, 15].

Activation of Gα_q stimulates phospholipase C-beta (PLCβ) and has been demonstrated to play a major role in the development of cardiac hypertrophy [16–20]. While S1P₂ and S1P₃ have been shown to couple to Gα_q in other systems [9, 14], it has not been determined whether stimulation of these receptors in cardiomyocytes activates Gα_q and PLCβ to elicit signals that lead to cardiomyocyte hypertrophy. The Gα_12/13 family of G proteins regulates the low molecular weight G protein Ras homolog gene member A (RhoA) through direct stimulation of guanine exchange factors [21–25]. Recently we demonstrated that cardiac expression of RhoA protects the heart against oxidative stress and ischemia/reperfusion (I/R) injury whereas gene deletion of RhoA decreases tolerance to ischemic damage [26]. We also showed that S1P could confer cardioprotection through RhoA and its downstream effectors [6].

In this study, we examined whether S1P regulates PLC activation and hypertrophy through S1P₂ or S1P₃ receptors and/or whether these receptor subtypes regulate activation of RhoA and in turn S1P-mediated cardioprotection. The data presented here demonstrate that S1P primarily signals through coupling to Gα₁₃ and activation of RhoA, that this pathway does not strongly activate PLCβ or contribute to development of cardiac hypertrophy, and that it is the S1P₃ receptor that regulates RhoA activation to mediate cardioprotection.

2. Materials and Methods

2.1. Animals

All animal procedures were performed in accordance with NIH Guide for the Care and Use of Laboratory Animals and approved by the Institutional Animal Care and Use Committee at the University of California – San Diego. Generation of global homozygous C57BL/6 S1P₂ KO and S1P₃ KO mice has been previously described [27]. All experiments were performed on age-matched male WT and KO littermates.

2.2. NRVM cell culture and reagents

Neonatal rat ventricular myocytes (NRVMs) were isolated from cardiac ventricles of 1- to 2-day-old Sprague-Dawley rat pups as described previously [28]. NRVMs were plated at a density of 3.0 x 10⁴/cm² and maintained overnight in Dulbecco-modified Eagle’s medium (DMEM) containing 15% fetal bovine serum overnight. Cells were either serum-starved with DMEM for 24 hours or transfected with siRNA for further analysis. Predesigned rat siRNA and scrambled control siRNA were purchased from Qiagen and used at 3 μg per 1 x 10⁶ cells. Cardiomyocytes were transfected with siRNA using DharmaFECT-1 transfection reagent from Thermo Fisher Scientific based on the manufacturer’s instruction. The Rho inhibitor C3 exoenzyme was obtained from Cytoskeleton (CT04). Pertussis toxin (PTX) was purchased from Calbiochem. Phenylephrine (PE) was obtained from Sigma Life Science. S1P was obtained from Avanti Polar. CYM-51736, is an allosteric agonist of S1PR₃ of the structure N,N-dicyclohexyl-5-(furan-3-yl)isoxazole-3-carboxamide [29, 30], was provided by Dr. Hugh Rosen (The Scripps Research Institute, La Jolla, CA).

2.3. Immunofluorescence

NRVMs were fixed in 3.5% paraformaldehyde solution, permeabilized in 0.2% NP-40 alternative, blocked in 2% bovine serum albumin (BSA) plus 10% goat serum, and then incubated in primary antibody against α-actinin (Sigma) or atrial natriuretic factor (Peninsula Laboratories) overnight at 4°C. Secondary antibodies conjugated to Alexa Fluor 488 and Alexa Fluor 555 (Invitrogen) were applied for 2 hours at room temperature. Cells were mounted with VECTASHIELD Hardset containing DAPI (Vector Labs). Images were acquired using confocal microscopy. Cell area was measured using a scale bar of 20 μm.

2.4. cDNA synthesis and qPCR analysis

RNA was isolated from NRVMs using Trizol. cDNA synthesis was carried out with the Verso cDNA synthesis kit (Thermo Scientific) and qRT–PCR was carried out using standard TaqMan primers and TaqMan Universal Mastermix II (Applied Biosystems) on a 7500 Fast Real–Time PCR system (Applied Biosystems). The data acquired was analyzed using the comparative C_T Method (i.e. the 2^−ΔΔC_T method) [31]. GAPDH levels were used as the internal control.

2.5. Phosphatidylinositol (PI) hydrolysis

After overnight culturing, NRVMs were labeled with tritium-labeled ([³H]) inositol (2.5 μCi/mL) for 24 hours. Cells were treated with agonists at various times in the presence of 25 mM lithium chloride, washed with cold PBS and incubated in cold 50 mM trichloroacetic acid (Sigma) for 40 minutes at 4°C. Samples were centrifuged and trichloroacetic acid was extracted with water-saturated ether. [3H]InsPs were isolated by ion exchange chromatography, and radioactivity was then measured by liquid scintillation counting.

2.6. Transverse aortic constriction

Transverse aortic constriction (TAC) was used on 8- to 10-week-old WT, S1P₂ KO, or S1P₃ KO mice to induce pressure overload hypertrophy as previously described [32–35]. The transverse aortic arch was visualized by a median sternotomy and a 7-0 silk ligature was tied around the aorta (27-gauge constriction) between the right brachiocephalic and the left common carotid arteries for one week.

2.7. GTP-RhoA pull-down assay

RhoA activation was determined as described previously [36]. Briefly, cell lysate was incubated with Rho binding domain of Rhotekin and then subjected to series of washes and centrifugations. 4 × Laemmli buffer was added and boiled for 5 minutes prior to SDS-PAGE analysis. Activated GTP-bound RhoA was detected by Western blotting for RhoA and normalized to total RhoA in lysate. For GTP-RhoA pulldown on isolated perfused hearts, tissue was flash-frozen, homogenized in RhoA lysis, debris pelleted via centrifugation, and the supernatant used for RhoA pull-down assay as described above.

2.8. Western blotting

Western blot analysis was performed according to protocols previously described [28]. The antibodies used for immunoblotting were the following: RhoA, Gα_q, Gα₁₂, and Gα₁₃ from Santa Cruz Biotechnology, glyceraldehyde-3-phosphate dehydrogenase (GAPDH), phospho-PKD (Ser^744/748), PKD, and α-actinin from Cell Signaling Technology. Peroxidase-conjugated secondary antibodies were used at a dilution of 1:2000 (Sigma) and the enhanced chemiluminescent substrate was from Thermo Fisher Scientific.

2.9. Isolated perfused heart (Langendorff) ischemia/reperfusion

Hearts from age-matched 8- to 12-week-old male WT, S1P₂ KO, or S1P₃ KO mice were removed quickly and perfused with modified Krebs-Henseleit buffer (118 mM NaCl, 4.7 mM KCl, 1.2 mM KH₂PO₄, 25 mM NaHCO₃, 0.5 mM EDTA, 1.2 mM MgSO₄, 11 mM glucose, 1.5 mM sodium pyruvate, and 2 mM CaCl₂) in a Langendorff apparatus (Radnoti) at a constant pressure of 80 mmHg. Hearts were stabilized for 10 minutes and then subjected to a period of global ischemia for 22 minutes followed by reperfusion for 60 minutes. To measure infarct size, triphenyl tetrazolium chloride (TTC) was used as described previously [28].

2.10 Dual-Luciferase reporter assay

NRVMs were plated onto 6-well plates. The following day, cells were transfected via Dharmafect1 (Dharmacon) for 8 hours with control siRNA or siRNA for Gα₁₂ or Gα₁₃. The following day, cells were transfected via Lipo2000 (Invitrogen) for 8 hours with an SRE.L reporter (1 μg/well) as well as a Renilla plasmid (100 ng/well) to normalize the luminescence signal. After transfections, cells were serum starved for 24 hours before stimulating with 1 μM S1P for 8 hours. Reporter activity was then measured using the Dual-Luciferase Reporter Assay (Promega) based on the manufacturer’s protocol.

2.11. Statistical analysis

All results are reported as means ± standard error of the mean (SEM). Comparison of two groups with one variable was accomplished using an unpaired Student’s t test, or if the control group was normalized to 1 with no variance, a paired t test. Data with two groups with multiple variables were compared with two-way analysis of variance (ANOVA) followed by Tukey’s multiple comparison test. Data from experiments with more than two groups with one variable were compared by one-way ANOVA followed by Tukey’s multiple comparison test, or repeated measures ANOVA followed by Tukey’s multiple comparison test if the control group was normalized to 1 with no variance. Probabilities ≤ 0.05 were considered significant.

3. Results

3.1. S1P does not induce cardiac hypertrophy

Stimulation of Gα_q-coupled receptors with ligands such as phenylephrine (PE) or endothelin regulates PLC to increase PI hydrolysis and downstream signals implicated in development of cardiac hypertrophy [37]. S1P can activate S1P receptor subtypes known to couple to Gα_q, but the role of S1P in control of cardiac hypertrophy appears to be minimal [38]. To further demonstrate and explore the basis for the limited efficacy of S1P as a hypertrophic agonist we first assessed the ability of S1P receptor stimulation to activate PLC in neonatal rat ventricular myocytes (NRVMs). NRVMs were serum-starved overnight, labeled with ³H-inositol, and treated with 0.3 μM S1P for 0, 1, 5, 10, 30, and 60 minutes. Responses were compared with those seen with 50 μM PE, an established hypertrophic agonist that activates α-adrenergic receptors. We observed a robust increase in the accumulation of inositol phosphates (InsPs) in response to PE whereas S1P elicited a much smaller response (Figure 1A). S1P receptor-mediated PLC activation could occur through the novel PLC isoform PLCε and its activation by RhoA [39, 40]. To determine if this was the mechanism by which S1P stimulated PI hydrolysis in NRVMs we inhibited RhoA function with C3 exoenzyme (2 μg/ml). The response to S1P was blocked whereas the response to PE was not (Figure 1B). In addition, knockdown of PLCε with siRNA prevented the PI response to S1P but not that to PE (Figure 1C). These data suggest that whereas PE works through the canonical Gα_q/PLCβ signaling pathway to elicit robust PI hydrolysis, the less robust PI response to S1P is mediated through RhoA signaling to PLCε.

Fig. 1. S1P receptor signaling stimulates modest increases in PI hydrolysis and does not mediate in vitro or in vivo cardiac hypertrophy.

(A) Time course of S1P- and PE-induced phosphatidylinositol hydrolysis. NRVMs were serum-starved overnight in the presence of [³H] inositol. Cells were then treated with agonists for 1, 5, 10, 30, and 60 minutes in the presence of LiCl before isolation of [³H] inositol phosphates (InsPs). The data displayed are the mean ± SEM. **P < 0.01 vs. vehicle (Veh), ##P < 0.01 vs. S1P (n = 5). (B) NRVMs were pretreated with 2.0 μg/mL C3 exoenzyme (Rho inhibitor) for 6 hours or (C) transfected with Control siRNA (siCtrl) or siRNA against PLCε for 48 hours before challenge with agonists for 60 minutes and assessed for InsPs production. **P < 0.01 vs. vehicle + Control or siCtrl, ##P < 0.01 vs. S1P + Control or siCtrl (n = 5). (D) Representative immunofluorescent images depicting NRVMs stained for α-actinin, atrial natriuretic factor (ANF), and nuclei with DAPI after treatment with either vehicle, 0.3 μM S1P, or 50 μM PE for 24 hours. Scale bar: 20 μm. (E) Quantified relative cell area and (F) quantified ANF positive cells (n = 400 cells). (G) mRNA expression of brain natriuretic peptide (BNP) and (H) skeletal muscle α-actin (ACTA1). **P < 0.01 vs. vehicle (n = 4). (I) heart weight (HW) to body weight (BW) ratio of WT, S1P₂ KO, and S1P₃ KO mice following transverse aortic constriction (TAC) to induce pressure overload hypertrophy for one week. **P < 0.01 vs. Sham (n ≥ 5).

To determine whether S1P treatment can elicit cardiomyocyte hypertrophy, NRVMs were treated with S1P or PE for 24 hours. Cells were subjected to immunofluorescence analysis for α-actinin, a cardiomyocyte cytoskeletal protein, and atrial natriuretic factor (ANF), a hypertrophic marker (Figure 1D). Immunofluorescence analysis revealed a large increase in ANF positive cardiomyocytes following PE treatment, but not following S1P treatment (Figure 1D, F). We also observed that PE, but not S1P, significantly increased cell surface area assessed by α-actinin staining (Figure 1D, E). qPCR analysis was used to measure the expression of brain natriuretic peptide (BNP) and skeletal muscle α-actin (ACTA1), which are both up-regulated in NRVMs treated with hypertrophic stimuli. PE treatment resulted in a 3.6-fold increase in BNP and 2.5-fold increase in ACTA1 mRNA expression, but S1P did not significantly increase expression of either hypertrophic marker (Figure 1G, H). Thus, S1P receptor stimulation is unable to trigger the conventional hypertrophic signaling pathways by which activation of receptors that couple to Gα_q/PLCβ elicits hypertrophy in NRVMs.

3.2. S1P₂ and S1P₃ receptors do not play a role in TAC induced hypertrophy in mouse hearts

Inhibition or genetic deletion of Gα_q prevents pressure overload hypertrophy induced by transverse aortic constriction (TAC) [19, 20], suggesting that this pathway is stimulated through upstream GPCR agonists. To determine whether either of the S1P receptor subtypes known to couple to Gα_q (S1P₂ and S1P₃) mediate development of hypertrophy in vivo, we subjected WT, S1P₂ KO, and S1P₃ KO mice to TAC for one week. We observed a significant increase in heart-weight relative to body-weight ratio, indicative of hypertrophy development, in WT mice and an equivalent increase in the S1P₂ and S1P₃ KO mice following TAC (Figure 1I). The finding that cardiac hypertrophy is still induced in the absence of these receptors suggests that S1P action on these receptors does not participate in the development of cardiac hypertrophy induced by pressure overload.

3.3. S1P induced RhoA activation is Gα₁₃ dependent

We postulated that the differential PI hydrolysis pathways used by PE and S1P reflects differences in G protein and RhoA activation by these agonists. RhoA activation was assessed by immunoprecipitating the GTP-bound activated RhoA using the Rho binding domain of Rhotekin, a RhoA effector. S1P robustly activated RhoA whereas PE did not (Figure 2A). S1P receptors have been shown to couple to Gα_i (for S1P₁) or to Gα_i, Gα_q, and Gα_12/13 (for S1P₂ and S1P₃). The G protein subunits that are most dedicated in coupling GPCRs to Rho activation through RhoGEFs are Gα₁₂ and Gα₁₃. In some systems, however, RhoA can also be activated through Gα_q and Gα_i. To test involvement of Gα_i in Rho activation induced by S1P cells were pretreated with PTX overnight followed by 5-minute S1P stimulation. S1P elicited a significant increase in activated RhoA which was unaffected by pretreatment with PTX (Figure 2B). To determine whether Gα_q, Gα₁₂ or Gα₁₃ were involved in S1P-mediated RhoA activation we used small interfering RNA (siRNA)-mediated knockdown. Treatment with siRNA for 72 hours reduced expression of each Gα subunit by greater than 50% (Figure 2C, D, E). Subsequent S1P treatment of cells transfected with control siRNA induced RhoA activation, which was not inhibited by knockdown of Gα_q or Gα₁₂ (Figure 2F). Knockdown of Gα₁₃, however, significantly attenuated RhoA activation by S1P. Using an SRE.L-luciferase readout for RhoA activation, we further demonstrated that S1P signaling in NRVMs was through Gα₁₃ but not Gα₁₂ (Figure 2G). These combined results implicate Gα₁₃ in S1P induced activation of RhoA in cardiomyocytes.

Fig. 2. S1P activates RhoA through Gα₁₃ in NRVMs.

(A) NRVMs were treated with 0.3 μM S1P or 50 μM PE for 5 minutes and RhoA activation was measured by a GTP-RhoA pull-down assay. The data displayed are the mean ± SEM (n = 5), **P < 0.01 vs. vehicle (Veh). (B) NRVMs were pretreated with 0.1 μg/mL pertussis toxin (PTX) overnight and then stimulated with 0.3 μM S1P for 5 minutes. RhoA activation was assessed by GTP-RhoA pull-down assay. NS indicates not significant (n = 4). (C) NRVMs were transfected with either Control siRNA (siCtrl) or siRNA against Gα_q, (D) Gα₁₂, or (E) Gα₁₃. G protein expression levels were assessed by Western blotting after 72 hour- knockdown, **P < 0.01 (n = 3). Cytoskeletal protein α-actinin was blotted as a loading control. (F) NRVMs were transfected with either siCtrl or siRNA against Gα_q, Gα₁₂, or Gα₁₃ for 72 hours and then stimulated with 0.3 μM S1P for 5 minutes. RhoA activation was assessed by GTP-RhoA pull-down assay. *, **P < 0.05, 0.01 vs. siCtrl vehicle (Veh), #P < 0.05 vs. siCtrl + S1P (n = 7). (G) Dual-Luciferase Reporter Assay was performed as described in Materials and Methods to assess RhoA activation. **P < 0.01 vs. siCtrl, ##P < 0.01 vs. siCtrl + S1P (n = 6).

3.4. S1P activates RhoA through the S1P₃ receptor in NRVMs and in isolated perfused hearts

To determine which S1P receptor subtypes are required for the activation of RhoA in response to S1P, siRNA for each of the S1P receptor subtypes was transfected into NRVMs. Treatment with siRNA resulted in a greater than 70% reduction in mRNA expression for each respective S1P receptor subtypes (Figure 3A, B, C) and did not affect the expression of other subtypes (data not shown). The siRNA-treated cells were stimulated with S1P for 5 minutes and RhoA activity was assessed by the pulldown of GTP-bound active RhoA. Knockdown of S1P₃ receptors markedly reduced RhoA activation by S1P. In contrast, RhoA was activated to an equivalent extent in control siRNA-treated cells, and S1P₁ or S1P₂ knockdown cells (Figure 3D). CYM-51736, an S1P₃ selective receptor agonist derived from CYM-5541 [29], also increased active RhoA (Figure 3E), albeit to a lesser extent (approximately 1.5 fold) than S1P (approximately 2.5-fold). As further support, we measured active RhoA in the presence of an S1P₂ receptor selective antagonist, JTE-013, which has been used at 1 μM to show S1P₂ selective signaling [41]. We observed that JTE-013 did not block RhoA activation by S1P (Figure 3F), further indicating S1P₃ but not S1P₂ involvement.

Fig. 3. S1P activates RhoA through S1P₃ in NRVMs and in isolated perfused hearts.

NRVMs were transfected with either siCtrl or siRNA against (A) S1P₁, (B) S1P₂, (C) or S1P₃. S1P receptor levels were assessed by qPCR analysis after 48-hour siRNA transfection. Data shown represent the mean ± SEM, **P < 0.01 vs. siCtrl (n = 3) (D) NRVMs were transfected with siRNA to knockdown S1P receptor subtypes for 48 hours and then stimulated with 0.3 μM S1P for 5 minutes. RhoA activation was assessed by GTP-RhoA pull-down assay. *, **P < 0.05, 0.01 vs. siCtrl, ##P < 0.01 vs. siCtrl + S1P, (n = 4). (E) NRVMs were transfected with siRNA to knock down S1P₃ receptor and then stimulated with 0.3μM S1P or 10μM CYM-51736 for 5 minutes. RhoA activation was assessed by the GTP-RhoA pulldown assay. **P < 0.01 vs. Veh (n = 4). (F) NRVMs were pretreated for 30 minutes with 1 μM of JTE-013, an S1P₂ receptor antagonist, and then stimulated with 0.3 μM S1P for 5 minutes. RhoA activation was assessed by the GTP-RhoA pulldown assay. *P < 0.05 vs. Veh (n = 3). (G) Isolated wild-type (WT), S1P₂ KO, and S1P₃ KO mouse hearts were perfused with either vehicle or 0.3 μM S1P in Krebs-Henseleit buffer for 5 minutes in Langendorff mode and RhoA activation was assessed by GTP-RhoA pull-down assay. **P < 0.01 vs. vehicle, ##P < 0.01 vs. WT + S1P (n = 3 to 5).

To extend our finding of S1P₃ receptor-mediated RhoA activation to the intact adult heart, WT, S1P₂ KO, or S1P₃ KO mouse hearts were isolated and perfused in the Langendorff mode. S1P was perfused for 5 minutes and RhoA activation was assessed using the GTP-RhoA pull-down assay. S1P significantly increased the amount of active RhoA in WT hearts and a similar increase was observed in hearts isolated from the S1P₂ KO (Figure 3G). In contrast, S1P treatment failed to activate RhoA in S1P₃ KO mouse hearts, indicating that the S1P₃ receptor plays a critical role in S1P induced RhoA activation in the adult mouse heart as it does in neonatal rat cardiomyocytes.

3.5. S1P₃ receptor mediates cardioprotection in Langendorff perfused mouse hearts against ischemia/reperfusion injury

Our lab has previously reported that S1P confers strong cardioprotection in isolated perfused mouse hearts through signaling to RhoA and PKD [6]. The findings above suggested that the cardioprotection induced by S1P would be mediated through activation of the S1P₃ receptor. To test this hypothesis, hearts from WT and S1P₃ KO mice were perfused with 0.3 μM S1P for 10 minutes, subjected to global ischemia for 22 minutes followed by reperfusion for 60 minutes, and cardiac damage assessed by staining with 1% TTC. S1P treatment significantly reduced infarct size in the WT mouse heart, supporting previous observations [6]. The protective effect of S1P was completely abolished in the S1P₃ KO hearts (Figure 4A). To further confirm the significant role of S1P₃ in cardioprotection we examined the effect of CYM-51736, a derivative of the previously reported S1P₃ receptor-specific agonist CYM-5541 [29]. Hearts isolated from WT mice were perfused with 10 μM CYM-51736 for 10 minutes followed by global I/R. CYM-51736 pretreatment resulted in a significant reduction in infarct size (Figure 4B), comparable to that observed in hearts perfused with S1P (Figure 4A) supporting the critical role of the S1P₃ receptor in S1P-mediated cardioprotection against I/R injury.

Fig. 4. S1P protects against ex vivo I/R through S1P₃ receptor.

Representative images of TTC-stained cross sections of isolated perfused mouse hearts after I/R injury (top) and quantification of infarct size (bottom). White areas are infarcted tissue and red areas are viable tissue. (A) Isolated WT and S1P₃ KO hearts were perfused with either Veh or 0.3 μM S1P for 10 minutes and subjected to 22 minute global ischemia followed by 60 minute reperfusion. *P < 0.05 vs. WT + Veh (n ≥ 5). The data shown are the mean ± SEM of each group’s animals. (B) WT hearts were perfused with either vehicle or 10 μM CYM-51736 (S1P₃ specific agonist) for 10 minutes and subjected to I/R. Infarct size was assessed by TTC staining. **P < 0.01 vs Vehicle (n = 5).

3.6. S1P₃ mediates the activation of PKD in NRVMs

We previously demonstrated that PKD is downstream of RhoA activation and contributes to S1P/RhoA-mediated cardioprotection [6]. Knockdown of S1P₃ in NRVMs significantly attenuated PKD phosphorylation in response to S1P whereas knockdown of S1P₁ or S1P₂ did not (Figure 5A). In addition, the S1P₃ selective agonist CYM-51736 increased PKD phosphorylation, and failed to do so following S1P₃ receptor knockdown (Figure 5B). PKD activation by CYM-51736 was attenuated by siRNA knockdown of Gα₁₃, or by inhibiting RhoA by treatment with C3 exoenzyme (Figure 5C). The S1P₂ antagonist JTE-013, shown above to have no effect on S1P-stimulated RhoA activation, also failed to block S1P-mediated phosphorylation of PKD (Figure 5D). Together, these results indicate that the cardiac S1P₃ receptor plays a critical role in initiating the RhoA/PKD signaling cascade that contributes to cardioprotection [6].

Fig. 5. The S1P₃ receptor mediates the activation of PKD by S1P in NRVMs.

Representative Western blots of NRVMs that were transfected with siRNA against S1P receptor subtypes for 48 hours followed by treatment with either (A) 0.3 μM S1P for 5 minutes or (B) 10 μM CYM-51736 for 20 minutes. The data displayed are the mean ± SEM. *P < 0.05 vs. siCtrl vehicle, #P < 0.05 vs. siCtrl + S1P or CYM-51736 (n = 4 to 5). (C) NRVMs were transfected with siRNA against Gα₁₃ for 72 hours or treated with 2 μg/ml C3 Rho inhibitor 12 hours, followed by treatment with 10 μM CYM-51736 for 15 minutes. Data displayed are the experimental means ± SEM. **P < 0.01 vs. siCtrl + Veh, #P < 0.05 vs. siCtrl + CYM-51736, ##P < 0.01 vs. siCtrl + CYM-51736 (n = 5). (D) NRVMs were pretreated for 30 minutes with 1 μM of S1P₂ receptor antagonist JTE-013, then stimulated with 0.3 μM S1P. Data displayed are the experimental means +/− SEM. **P < 0.01 vs. vehicle, #P < 0.05 vs. S1P stimulation (n = 3).

4. Discussion

S1P is a pleiotropic bioactive lysophospholipid that couples to a variety of GPCR subtypes to regulate biological functions, including cell proliferation, inflammation, and cardiac function [24, 42–44]. In the heart, S1P has been shown to confer cardioprotection against ischemic stress [3, 4, 45], and we recently reported that RhoA plays a crucial role in S1P-mediated cardioprotection against I/R injury [6, 10, 26, 46]. Identifying the S1P receptor subtype through which cardioprotection is mediated would inform further consideration of potential therapeutic targets for ischemic heart diseases. We demonstrate here that S1P₃ is the receptor subtype responsible for S1P-mediated RhoA activation, PKD activation and protection against I/R injury in cardiomyocytes and in the isolated perfused heart. Our results also suggest that Gα₁₃, but not Gα₁₂, couples S1P receptor stimulation to RhoA activation in the heart. We also provide extensive evidence that S1P signaling through this pathway causes limited PLC activation and is not sufficient to induce hypertrophic responses.

4.1. S1P and cardiac hypertrophy

Gα_q has been established to be essential in development of hypertrophy induced by GPCR agonists and by pressure-overload [19, 20], presumably through PLCβ and its downstream mediators [17, 47]. Our results demonstrate that S1P, in contrast to many other GPCR agonists, does not induce a hypertrophic response in NRVMs (Figure 1D, E, F, G, and H) nor is S1P receptor stimulation involved in pressure-overload induced hypertrophy in vivo (Figure 1I). Compared to the adrenergic agonist PE, which activates Gα_q-coupled α₁-adrenergic receptors and PLCβ, S1P-induced PI responses are extremely modest and appear to be RhoA/PLCε dependent (Figure 1A, B, and C). These results suggest that although S1P₂ and S1P₃ have the ability to activate Gα_q signaling [9, 14, 15], recruitment of Gα_q signaling through these receptors in cardiomyocytes is limited and is not sufficient to drive hypertrophic responses in the heart. The observation that the hypertrophic response induced by pressure-overload was unaffected in S1P₂ or S1P₃ KO mice further indicates that activation of these S1P receptors does not contribute to hypertrophy in vivo (Figure 1).

4.2. S1P receptor and Gα subtype coupling in RhoA signaling

S1P₁ exclusively couples to Gα_i, while S1P₂ and S1P₃ activate multiple G protein subtypes: Gα_i, Gα_q, Gα_12, and Gα₁₃ [12, 14, 15, 46]. In the context of RhoA activation, Gα₁₂ and Gα₁₃, members of the Gα_12/13 subfamily of G proteins, are recognized to transduce G protein-coupled receptor stimulation to the activation of RhoA through direct regulation of RhoGEFs [21, 24, 48–50]. Interestingly, the data presented here reveal that Gα₁₃ but not Gα₁₂ knockdown significantly attenuates S1P-induced RhoA activation, suggesting that Gα₁₂ cannot compensate for the function of Gα₁₃ in cardiomyocytes. We cannot rule out the possibility that the Gα₁₂ knockdown achieved by siRNA is insufficient to observe the contribution of Gα₁₂ to S1P-mediated RhoA activation in NRVMs. It is clear, however, that Gα₁₂ and Gα₁₃ can signal through distinct and nonredundant mechanisms, as evidenced by the finding that Gα₁₃ KO mice die during embryonic development, while Gα₁₂ KO mice are viable, fertile, and without obvious phenotype [51–53].

There is considerable evidence for preferential activation of RhoA through Gα₁₃ in response to upstream GPCR activation in some systems. A previous study using G protein knockdown or knockout in cardiomyocytes demonstrated that Gα₁₃ was specifically required for endothelin-1 and AngII to activate RhoA [54]. Activation of RhoA by a TXA₂ agonist has been observed in both wild-type and Gα₁₂ KO platelets, whereas Gα₁₃ KO platelets lacked RhoA activation [55]. Studies using Gα₁₂ and Gα₁₃ antibody microinjection in Swiss 3T3 cells demonstrated that lysophosphatidic acid (LPA) stimulated stress fiber formation, a RhoA-mediated response, through Gα₁₃ but not Gα₁₂ [56]. Using the SRE.L reporter assay, as demonstrated for S1P (Figure 2G), we also observed preferential involvement of Gα₁₃ but not Gα₁₂ in RhoA activation by LPA (data not shown).

Selective utilization of Gα₁₃ versus Gα₁₂ may reflect the specific nature of the RhoA GTP exchange factors (GEFs) that predominate in the tissue under study. The regulator of G protein signaling homology (RH) family of Rho GEFs (p115RhoGEF, leukemia-associated RhoGEF a.k.a. LARG, and PDZ-RhoGEF) are among the best characterized downstream effectors of Gα₁₂ and Gα₁₃. Previous seminal in vitro reconstitution experiments revealed that Gα₁₃, but not Gα₁₂, stimulates the activity of p115 RhoGEF activity and PDZ-RhoGEF [21, 23]. Leukemia-associated RhoGEF (LARG) is also selectively activated by Gα₁₃ and can only be activated by Gα₁₂ when this GEF is phosphorylated by tyrosine-kinase [57].

4.3. S1P receptor subtypes and Rho signaling

The expression of the five S1P receptor subtypes varies depending on cell type. Previous studies, including work from our laboratory and those of others showed that S1P₁, S1P₂, and S1P₃ are the major subtypes expressed in the heart [8, 27, 46, 58]. S1P₁ is the most abundant subtype in mouse cardiomyocytes, followed by S1P₃, while S1P₂ is expressed at relatively lower levels [5, 11, 27, 58]. The lack of involvement of the S1P₁ receptor in RhoA activation is consistent with our finding that blocking Gα_i, the only G protein to which S1P₁ couples, does not inhibit RhoA activation.

The relative roles of S1P₃ versus S1P₂ in RhoA activation appear to be tissue specific since studies in non-cardiomyocytes implicate S1P₂ in RhoA activation, or the combined effects of S1P₂ and S1P₃ on activation of RhoA [27, 59, 60]. Which subtype predominates could depend on subcellular localization of the receptor or its level of expression. The higher level of mRNA expression for S1P₃ versus S1P₂ in cardiomyocytes may explain the dominance of S1P₃ signaling to RhoA in the heart [5, 61]. There may also be as yet undefined differences in S1P₂ and S1P₃ coupling to Gα₁₂ versus Gα₁₃, for example differences in interaction with endogenous RhoGEFs in a given cell type as discussed above. Regardless of the mechanistic basis, our findings using siRNA-mediated S1P receptor knockdown and S1P₃ KO hearts are the first to reveal that S1P₃ is the receptor subtype that mediates RhoA activation by S1P in cardiomyocytes (Figure 3D, 3G).

4.4. S1P receptor subtypes and cardioprotection

A wide range of molecules downstream of S1P receptor activation have the potential to mediate cardioprotection. Our laboratory previously demonstrated that S1P treatment of NRVMs leads, through RhoA and its effect on PLCε activation, to activation of PKD [6]. Furthermore, the protective effect of S1P was lost in hearts from PKD KO or PLCε KO mice [6]. In line with the aforementioned significant contribution of S1P₃ to RhoA activation, we show here that PKD activation induced by S1P treatment also required the S1P₃ receptor (Figure 5A). These results place PKD downstream in the S1P₃/RhoA signaling axis. We demonstrate that S1P-mediated cardioprotection against I/R injury is abolished in S1P₃ KO hearts (Figure 4A), and stimulated by the S1P₃ agonist CYM-51736 (Figure 4B). These results are consistent with previously published work demonstrating that S1P associated with high density lipoprotein protects the heart against I/R injury in vivo through effects on S1P₃ [4]. We also demonstrate that selective activation of S1P₃ by CYM-51736 leads to robust activation of PKD (Figure 5C) and confers protection comparable to that observed with S1P (Figure 4B), whereas inhibition of S1P₂ by JTE-013 does not block PKD activation by S1P (Figure 5D). Thus, this study defines early players in the S1P signaling pathway and suggests that coupling of the S1P₃ receptor to Gα₁₃ and RhoA plays a major role in S1P mediated cardioprotection against I/R injury.

The involvement of S1P₃ receptors in RhoA cardioprotective signaling is of additional interest since fingolimod (FTY720), a widely used drug for clinical treatment of multiple sclerosis, is an agonist for both S1P₁ and S1P₃ receptors. Fingolimod has been reported to induce cardioprotection in a porcine model of I/R injury [62], as well as to protect against reperfusion-induced cardiac arrhythmias [58, 63, 64]. Notably the S1P₁ receptor which is predominant in the myocardium is downregulated by fingolimod treatment where S1P₃ receptors may remain [63]. In light of the findings presented here, signaling through the S1P₃ receptor could contribute to the protective effects of fingolimod.

All three S1P receptors expressed in the heart have been shown to couple to Gα_i, Our earlier studies using an in vivo I/R model concluded that both S1P₂ and S1P₃ activate Akt through Gα_i to confer cardioprotection [5]. Another study also demonstrated that S1P₂ and S1P₃ antagonists block S1P mediated cardioprotection [65]. The studies reported here show that signaling via the S1P₃ receptor to RhoA in cardiomyocytes of the isolated, perfused heart is both sufficient and required for cardioprotection in ex vivo I/R. Akt activation by S1P₂ and S1P₃ could contribute, along with the S1P₃-mediated RhoA activation, to the protective effect of S1P on cardiomyocytes observed in vivo, either by direct actions on cardiomyocytes or S1P-mediated effects on non-cardiac cells [41, 63].

4.5 Conclusion

We show in this study that S1P receptor stimulation does not induce hypertrophic responses in cardiomyocytes and is not necessary for the development of hypertrophy in response to pressure-overload. Using siRNA-mediated knockdown and knockout mouse models, we demonstrate for the first time that the actions of S1P on the S1P₃ receptor and its coupling to Gα₁₃ activate RhoA and PKD. Using gene silencing and deletion, we also demonstrate that S1P₃ is responsible for S1P-mediated cardioprotection against ex vivo I/R injury. These findings are supported by experiments using a recently developed S1P₃ selective agonist, CYM-51736. We suggest that specific drug targeting of S1P₃ receptors could provide a therapeutic benefit in ischemic heart disease without the undesirable effects of global activation of other cardiac S1P receptors.

Highlights.

• S1P signaling does not contribute to in vitro or in vivo cardiomyocyte hypertrophy

• S1P mediates RhoA activation in murine cardiomyocytes through the S1P₃ receptor

• S1P₃ couples to Gα₁₃ to regulate RhoA activation in cardiomyocytes

• S1P signaling through S1P₃ and RhoA elicits protection during ischemia/reperfusion