The switching role of β-adrenergic receptor signalling in cell survival or death decision of cardiomyocytes

Abstract

How cell fate (survival or death) is determined and whether such determination depends on the strength of stimulation has remained unclear. In this study, we discover that the cell fate of cardiomyocytes switches from survival to death with the increase of β-adrenergic receptor (β-AR) stimulation. Mathematical simulations combined with biochemical experimentation of β-AR signalling pathways show that the gradual increment of isoproterenol (a non-selective β1/β2-AR agonist) induces the switching response of Bcl-2 expression from the initial increase followed by a decrease below its basal level. The ERK1/2 and ICER-mediated feed-forward loop is the hidden design principle underlying such cell fate switching characteristics. Moreover, we find that β1-blocker treatment increases the survival effect of β-AR stimuli through the regulation of Bcl-2 expression leading to the resistance to cell death, providing new insight into the mechanism of therapeutic effects. Our systems analysis further suggests a novel potential therapeutic strategy for heart disease.

Figure 1. A schematic diagram for the β-AR signalling network.

The β-AR signalling network comprises four modules: the cAMP-PKA signalling module (thick grey arrow), the central feedback regulatory module (thick red arrow and orange square), the ERK1/2 signalling module (thick blue arrow) and the Ca²⁺ regulatory module (thick purple arrow and purple square). β₁-AR, β₁-adrenergic receptor; β₂-AR, β₂-adrenergic receptor; G_s, stimulatory G protein; G_i, inhibitory G protein; AC, adenylyl cyclase; CN, calcineurin; ICER, inducible cAMP early repressor; RSK, stress-activated protein kinase; MSK, mitogen-activated protein kinase.

Figure 2. Estimation of kinetic parameter values for the β-AR signalling network model.

The temporal profiles for signalling molecules in response to the indicated ISO concentrations were shown for the experimental results and the simulation results. The time courses of p-ERK1/2 (a,b), p-CREB (d,e) and ICER (f,g) were obtained from our experiments with primary adult rat cardiomyocytes. The time courses of the complex formation of SOS/Grb2 (ref. 67) (c), ICER protein (h), PDE3 protein (i), active CaMKII (j), cAMP (k), and PKA activity (l) were reproduced from the previous experimental data, where the cAMP levels and PKA activity were measured by real-time Förster resonance energy transfer-based live-cell imaging in the absence of PDE inhibitor. The kinetic parameters of the mathematical model were estimated from those experimental data. The experimental data represent mean±s.e.m. for at least three independent experiments; representative blot images and the time course of qRT–PCR results can be found in Supplementary Fig. 1. The simulation data represent mean±s.e.m. (or mean value) for the repetitive simulations (n=30) over up to 30% random variation of parameter values.

Figure 3. ISO induces the switching response of Bcl-2.

Time-dependent PKA activation (a), CREB phosphorylation (b), ICER expression (c), PDE3 expression (d), ERK1/2 phosphorylation (e) and Bcl-2 expression (f) in response to three different concentrations of ISO: 10 pM (grey line), 10 nM (blue line) and 10 μM (red line). The time courses of these signalling molecules were quite distinct depending on the ISO concentration. The dose–response profile of PKA activation (g), CREB phosphorylation (h), ICER expression (i), PDE3 expression (j), ERK1/2 phosphorylation (k) and Bcl-2 expression (l) for the indicated ISO concentration range are also shown. Each dose response (g–l) was observed at 24 h after the ISO stimulation. Active PKA, p-CREB, ICER, PDE3 and p-ERK1/2 showed sigmoidal dose–response profiles, and Bcl-2 expression increased at the nanomolar concentration range; however, it then decreased to further below its basal level at a micromolar concentration range. The data represent means±s.e.m. for the repetitive simulations (n=100) over up to 20% random variation of parameter values. *P<0.05; **P<0.01; ***P<0.001 compared with control group; Student’s t-test.

Figure 4. Switching response of Bcl-2 to β-AR stimulation in cardiomyocytes.

(a–d) Switching response profiles of Bcl-2 were observed in cardiomyocytes when we incubated them with the indicated concentrations of ISO for 12 or 24 h (a) or preincubated them with the indicated concentrations of a PKA inhibitor (RP-cAMPs) for 2 h before incubation with 1 μM ISO for 12 h (b). Representative Bcl-2 and α-tubulin (loading control) immunoblots (a,b) and the semi-quantified data (c,d) are shown. Note that the concentrations of RP-cAMP were rearranged in the order of increased signalling flux through the cAMP-PKA signalling module (d). The data represent means±s.e.m., n=3 biological and technical replicates (independent culture preparations). (e–j) Cardiomyocytes preincubated with the indicated concentrations of ISO were subjected to H₂O₂ (100 μM) or ionomycin (3.5 μM) treatment. Cell death was measured by ELISA (e,f). The survival rate was assessed by live-cell imaging (g,h; Supplementary Movies 3–9), and the data were summarized (i,j), where the insets are magnifications of the dash-outlined area (g,h). Scale bar, 100 μm. The data represent means±s.d. pooled from three biological and technical replicates (independent culture preparations) (e,f) or means±s.e.m. pooled from three biological and technical replicates (independent culture preparations) (i,j). *P<0.05; **P<0.01; ***P<0.001; NS, not significant compared with their control groups; Student’s t-test. Uncropped western blots are shown in Supplementary Fig. 14.

Figure 5. Identification of the core regulatory circuit that robustly generates the switching response profile of Bcl-2.

(a) A detailed β-AR signalling network that implements the switching response of the Bcl-2 expression. The grey-shaped circles indicate clustered signalling components. (b) A simplified β-AR signalling network. The eight links are essential for the generation of the switching response of Bcl-2. The circled numbers indicate the indexes of the regulatory links that are same as in a. (c) Identification of models that can robustly generate the switching response profile of Bcl-2 for each combination of circuit design (m^th) and parameter set (n^th). Inset: the identification is constrained to the switching response of Bcl-2 to the gradual increase of ISO. (d) The four most robust models that can produce the switching response of Bcl-2, in which ERK and ICER-mediated incoherent feed-forward loop is commonly included.

Figure 6. ISO concentration-dependent Bcl-2 expression for a different perturbation level of each essential regulatory link.

(a) Blockade of ICER-mediated link significantly increased Bcl-2 expression only at or above the micromolar concentration range of ISO. (b) Blockade of the ERK1/2-mediated link decreased Bcl-2 expression at 10⁻¹⁰–10⁻⁷ M ISO, where the dashed line denotes control and the solid line denotes the perturbed condition. The simulation of the dose–response profiles was observed at 12 h after ISO stimulation. (c–e) Control siRNA (siControl) or ICER siRNA (siICER) transfected cardiomyocytes were stimulated with the indicated concentrations of ISO. (c) Representative immunoblots showing Bcl-2 expression at the indicated ISO concentration determined in the presence of siControl or siICER. (d) Plots of Bcl-2 protein expression shown in c versus ISO dose. Data represent means±s.e.m., n=3 biological and technical replicates (independent culture preparations). (e) The survival rate was assessed by live-cell imaging. Data represent means±s.e.m., n=3 biological and technical replicates (independent culture preparations). (f–h) Cardiomyocytes were stimulated with the indicated concentrations of ISO in the presence or absence of PD98059 (f) Representative immunoblots showing Bcl-2 expression at the indicated ISO concentration determined in the presence or absence of PD98059. (g) Plots of Bcl-2 protein expression shown in f versus ISO dose. Data represent means±s.e.m., n=3 biological and technical replicates (independent culture preparations). (h) The survival rate was assessed by live-cell imaging. Data represent means±s.e.m., n=6 biological and technical replicates (independent culture preparations). ^##P<0.01; ^###P<0.001 compared to non-treated control group; *P<0.05 compared with no drug-treated and the same concentration of ISO-treated group; Student’s t-test. Uncropped western blots are shown in Supplementary Fig. 14.

Figure 7. β₁-blockers increase the tolerance of cardiomyocytes to cell death by expanding the survival range of the switching response profile of Bcl-2.

(a,b) Heat maps of Bcl-2 expression in response to treatment with ISO and β₁-blocker (a) or ISO and β₂-blocker (b). The survival range was increased by the β₁-blocker, whereas it was decreased by the β₂-blocker. (c,d) Simulation curves for Bcl-2 expression versus ISO dose in the presence of different concentrations of β₁-blocker (c) or β₂-blocker (d). Note that the bell-shaped curves were expanded toward a high ISO concentration with an increased β₁-blocker concentration, whereas the opposite pattern was observed for the β₂-blocker. The simulation of the dose–response profiles was observed at 24 h after ISO stimulation. The data represent mean±s.e.m. for the repetitive simulations (n=20) over up to 20% random variation of parameter values. (e,f) Representative immunoblots showing the ISO concentration-response effects for Bcl-2 expression in the presence of metoprolol or ICI 118,551. (g,h) The line graphs depict the semi-quantification of the immunoblots shown in e,f. The dashed lines represent the concentration-response profile of Bcl-2 expression with ISO alone (that is, the control data taken from Fig. 6i). (i,j) The survival rate was assessed by live-cell imaging in the presence of metoprolol or ICI 118,551. Cell death induced by higher concentration (10⁻⁷–10⁻⁶ M) of ISO is significantly reduced by metoprolol. Data represent means±s.e.m. pooled from more than three biological and technical replicates (independent culture preparations). *P<0.05; **P<0.01 with Student’s t-test. Uncropped western blots are shown in Supplementary Fig. 14.

Figure 8. The core regulatory circuit is pivotal for the concentration-dependent cell fate determination.

The eight essential regulatory links of the β-AR signalling network are primarily responsible for the switching response of Bcl-2 (thin solid lines). Among these, the ERK1/2 and ICER-mediated incoherent feed-forward loop is the core regulatory circuit that robustly generates the Bcl-2 switching response of cardiomyocytes, in which the ERK1/2 pathway at a low concentration range of ISO positively regulates Bcl-2 induction and thus cell survival ensues, whereas the PKA-ICER-CREB pathway at a high concentration range negatively regulates Bcl-2 induction and therefore promotes cell death.