Roles of phosphodiesterases in the regulation of the cardiac cyclic nucleotide cross-talk signaling network

Abstract

The balanced signaling between the two cyclic nucleotides (cNs) cAMP and cGMP plays a critical role in regulating cardiac contractility. Their degradation is controlled by distinctly regulated phosphodiesterase isoenzymes (PDEs), which in turn are also regulated by these cNs. As a result, PDEs facilitate communication between the β-adrenergic and Nitric Oxide (NO)/cGMP/Protein Kinase G (PKG) signaling pathways, which regulate the synthesis of cAMP and cGMP respectively. The phenomena in which the cAMP and cGMP pathways influence the dynamics of each other are collectively referred to as cN cross-talk. However, the cross-talk response and the individual roles of each PDE isoenzyme in shaping this response remain to be fully characterized. We have developed a computational model of the cN cross-talk network that mechanistically integrates the β-adrenergic and NO/cGMP/PKG pathways via regulation of PDEs by both cNs. The individual model components and the integrated network model replicate experimentally observed activation-response relationships and temporal dynamics. The model predicts that, due to compensatory interactions between PDEs, NO stimulation in the presence of sub-maximal β-adrenergic stimulation results in an increase in cytosolic cAMP accumulation and corresponding increases in PKA-I and PKA-II activation; however, the potentiation is small in magnitude compared to that of NO activation of the NO/cGMP/PKG pathway. In a reciprocal manner, β-adrenergic stimulation in the presence of sub-maximal NO stimulation results in modest cGMP elevation and corresponding increase in PKG activation. In addition, we demonstrate that PDE2 hydrolyzes increasing amounts of cAMP with increasing levels of β-adrenergic stimulation, and hydrolyzes increasing amounts of cGMP with decreasing levels of NO stimulation. Finally, we show that PDE2 compensates for inhibition of PDE5 both in terms of cGMP and cAMP dynamics, leading to cGMP elevation and increased PKG activation, while maintaining whole-cell β-adrenergic responses similar to that prior to PDE5 inhibition. By defining and quantifying reactions comprising cN cross-talk, the model characterizes the cross-talk response and reveals the underlying mechanisms of PDEs in this non-linear, tightly-coupled reaction system.

Keywords: Cardiac myocytes; Computational model; Cyclic nucleotide cross-talk signaling network; NO/cGMP/PKG pathway; Phosphodiesterases; β-adrenergic pathway.

Fig. 1. Cyclic nucleotide cross-talk signaling network

(A) The cN cross-talk signaling network model is composed of the β-adrenergic pathway (red background), the NO/cGMP/PKG signaling pathway (blue background), and cross-talk between them (yellow background). Cross-talk is mediated by PDEs 1–5. In the regulation of cAMP- (B) and cGMP- (C) hydrolysis, cNs exert positive (green arrows) or negative (red arrows) regulation of PDE activities. In particular, PDE2 hydrolysis rate of either cN is stimulated (green arrow) by low concentrations of the other cN but is suppressed (red arrow) if the concentrations are sufficiently high. To avoid crowding the figure, the hydrolysis reactions of cNs are omitted in (B) and (C), which would have been drawn as red arrows originating from each PDE to cAMP in (B) and cGMP in (C). Instead, hydrolysis of cAMP and cGMP are respectively represented by ovals of faded red in (B) and faded blue in (C).

Fig. 2. Experimental validation of PDE models

Experimental data predicted by the PDE models. Dotted symbols are experimental data (obtained from purified PDEs); lines are simulation results. (A) PDE1 cAMP hydrolysis rate at 6.25 μM [cGMP] (normalized to the maximum rate) versus data of Yan et al. [57]. (B) Normalized PDE1 cGMP hydrolysis rate at 6.25 μM [cAMP] versus data of Yan et al. [57]. (C) Effects of [cGMP] on PDE2 cAMP hydrolysis rates (1 μM [cAMP]) compared to data of Prigent et al. [58]. Reported rates are normalized against rate without cGMP. (D) PDE2 cGMP hydrolysis rates (10 μM [cGMP]) as a function of [cAMP] compared to data of Russell et al. [59], normalized to that without cAMP. (E) PDE3 cAMP hydrolysis (0.2 μM [cAMP]) as a function of [cGMP] versus data of He et al. [60], normalized to that without cGMP. (F) Normalized PDE3 cGMP hydrolysis (0.2 μM [cGMP]) versus data of He et al. [61].

Fig. 3. Experimental validation of cross-talk signaling network model

Model predicts experimental data from rat cardiac ventricular myocytes. Dotted symbols are experimental data; lines are simulation results. (A) Dose-response relationship of [cAMP] in response to ISO, with (blue) and without (black) the PDE inhibitor IBMX (100 μM), versus data of Vila-Petroff et al. [47] (blue dots) and Kuznetsov et al. [62] (black dots). (B) Simulated [cAMP] time course in response to 10 nM [ISO] compared to data of Vila-Petroff et al. [63] (filled black dots) and Zaccolo et al. [48] (hollow blue dots). (C) Simulated [cGMP] with (gray bar) and without 100 μM [IBMX] (white bar) in the absence and presence of SNAP (100 μM) compared to data of Castro et al., 2006 [49] (black dots) and 2010 [40] (blue dots). (D) Simulated [cGMP] time course under 100 μM [SNAP] perfusion followed by 100 μM [IBMX] application compared to data of Castro et al. [49]. (E) Simulated [cGMP] time course in presence of SNAP (100 μM) with intermittant application of the specific PDE inhibitors EHNA (PDE2 inhbitor, 10 μM) and Sildenafil (Sil, PDE5 inhibitor, 0.1 μM), and the non-specific PDE inhibitor (100 μM [IBMX]) compared to data of Castro et al. [49]. (C)–(D) Details on simulations and data processing are provided in Suppl. Sect. VIIIA.

Fig. 4. Cross-talk responses as compared to direct pathway responses

Direct and cross-talk responses are shown in black and gray respectively. (A)–(C) NO-elicited direct response in the NO/cGMP/PKG pathway (black) is compared to its cross-talk response in the β-adrenergic pathway (gray) at 1 nM [ISO] and with varying [SNAP]. (D)–(F) ISO-elicited direct response of the β-adrenergic pathway (black) is compared to its cross-talk response in the NO/cGMP/PKG pathway (gray) at 5 nM [SNAP] and with varying [ISO]. (A) [cGMP] versus [cAMP]. (B) PKG versus PKA-I activation as a percentage of total [PKG] and [PKA-I] respectively. (C) PKG versus PKA-II activation as a percentage of total [PKG] and [PKA-II] respectively. (D) [cAMP] versus [cGMP]. (E) PKA-I versus PKG activation as a percentage of total [PKA-I] and [PKG] respectively. (F) PKA-II versus PKG activation as a percentage of total [PKA-II] and [PKG] respectively.

Fig. 5. PDE interactions underlying cross-talk response

Shadings for PDEs 1–5 are respectively black, red, blue, green, and cyan, with gray indicating net increases across all PDEs. (A) and (B) Percent increases in PDE cAMP hydrolysis rates, under low (1 nM) and high (1 μM) [ISO] respectively. The percent increases are calculated against PDE2 cAMP hydrolysis rates without SNAP but with the same indicated [ISO]. (C) and (D) Percent increases in PDE cGMP hydrolysis rates under low (5 nM) and high (500 nM) [SNAP] and indicated [ISO], normalized to PDE2 cGMP hydrolysis rates before ISO application but under the same indicated [SNAP].

Fig. 6. Role of PDE2 in cross-talk signaling network

(A) Percent cAMP hydrolyzed by PDE2 relative to total cAMP hydrolyzed at each [ISO], without SNAP and with simultaneous applications of 5 nM, 50 nM, and 500 nM [SNAP]. (B) Percent cGMP hydrolyzed by PDE2 against total cGMP hydrolyzed at each [SNAP], without ISO and with simultaneous applications of 10 nM, 100 nM, and 1 μM [ISO].

Fig. 7. NO/cGMP/PKG pathway responses under PDE5 inhibition

All simulation data shown are percent increases upon PDE5 inhibition relative to their respective values prior to inhibition under a medium level of [SNAP] (50 nM) without ISO stimulation. Varying degrees of PDE5 inhibition (20%, 50%, and 90%) is simulated with simultaneous SNAP application under the indicated [ISO] for 30 min. (A) Percent increases in cGMP hydrolysis rates of PDEs 1–3 and 5 upon 90% PDE5 inhibition. (B)–(D) respectively, percent increases in PDE2 cGMP rates, [cGMP], and PKG activation upon 20%, 50%, and 90% PDE5 inhibition.

Fig. 8. β-adrenergic pathway activity under PDE5 inhibition

All simulation data shown are percent increases upon PDE5 inhibition relative to their respective values prior to inhibition under the same [SNAP] (50 nM) without ISO stimulation (same normalization as in Fig. 7). (A) Percent increases in cAMP hydrolysis rates of PDEs 1–4 upon 90% PDE5 inhibition. (B)–(D) respectively, percent increases in [cAMP], PDE2 cAMP rates, and PDE3 cAMP rates upon 20%, 50%, and 90% PDE5 inhibition.