Going to cAMP just got more complicated
D. Bers
DOI: https://doi.org/10.1113/jphysiol.2007.140764
2007-09-01
Journal of Physiology
Abstract:In this issue of The Journal of Physiology, Pereira et al. (2007) demonstrate that some acute effects of cyclic AMP (cAMP)-dependent signalling on Ca2+ transport in adult cardiac myocytes may be independent of cAMP-dependent protein kinase (PKA). They show that activation of an alternative cAMP target (Epac, exchange protein activated by cAMP) leads to activation of Ca–calmodulin-dependent protein kinase II (CaMKII), with important consequent effects on sarcoplasmic reticulum (SR) Ca2+ release channels (activating potentially arrhythmogenic diastolic SR Ca2+ release and reducing SR Ca2+ content and Ca2+ transients).
Acute cAMP signalling in heart is classically attributed to direct activation of either PKA or cyclic nucleotide-gated channels (Bers, 2001), as physiological mediators of sympathetic regulation of the heart, via chronotropic, inotropic and lusitropic effects. In the classic pathway, β-adrenergic receptor (β-AR) activation causes G-protein (Gs)-dependent cAMP production (by adenylate cyclase), and cAMP binds to the regulatory subunit of PKA, activating the catalytic subunit (Fig. 1). PKA can then phosphorylate L-type Ca2+ channels, enhancing Ca2+ current (ICa) amplitude, which enhances cellular Ca2+ loading, triggers larger SR Ca2+ release (inotropic), and shifts activation of ICa gating to more negative membrane potential (allowing earlier activation during pacemaker activity, chronotropic). PKA also phosphorylates phospholamban, relieving its tonic inhibition of SR Ca2+-ATPase, enhancing the rate of Ca2+ uptake and relaxation (lusitropic) and enhancing SR Ca2+ content (inotropic). PKA can also phosphorylate the ryanodine receptor (RyR), which may sensitize the release process to a given ICa trigger (potentially inotropic, although details are controversial). Hyperpolarization and cyclic nucleotide-gated channels (HCN) that carry the pacemaker current, If, are activated directly by cAMP (which shifts activation to more negative membrane potential) contributing to the chronotropic effect of sympathetic stimulation.
Figure 1
Activation of Epac by cAMP and activation of CaMKII-dependent phosphorylation of the RyR Broken lines and question marks indicate aspects where the possible pathway requires further tests. Abbreviations are as in text except: INa is Na+ current, α ...
Epac1 and Epac2 are guanine nucleotide exchange factors discovered in database searches for proteins which might explain cAMP-dependent activation of the small GTPase Rap1, which was insensitive to PKA inhibition (Bos, 2006). In heart, Epac1 appears to be part of a PKA anchoring protein (mAKAP) complex at the nuclear envelope (Fig. 1; McConnachie et al. 2006). Epac1 binds to mAKAP via phosphodiesterase 4D3 (PDE4D3) and cAMP can activate two parallel pathways: (1) the canonical PKA pathway in which PKA phosphorylates and activates PDE4D3 to break down cAMP and limit cAMP signals, and (2) Epac1 activation of Rap1, which can inhibit ERK5-dependent phosphorylation and inhibition of PDE4D3, thereby also activating PDE4D3, but also intersecting with the mitogen-activated protein kinase pathways. In this vein, cAMP-dependent Epac activation has also been implicated in hypertrophic signalling via another small GTPase, Rac, and activation of calcineurin (Morel et al. 2005).
Pereira et al. (2007) show the first compelling evidence that Epac activation by a cAMP analogue (8-CPT) that is highly selective for Epac over PKA can acutely alter cardiac myocyte Ca2+ handling and that these effects are mediated by CaMKII. The dominant effect was a CaMKII-dependent (and PKA-independent) activation of Ca2+ sparks and reduction in SR Ca2+ content and twitch Ca2+ transients. Curran et al. (2007) also found that β-AR stimulation enhanced SR Ca2+ leak and lowered SR Ca2+ content in ventricular myocytes in a CaMKII-dependent (and PKA-independent) manner. However, in that case forskolin (which directly activates cAMP production by adenylate cyclase) did not mimic the β-AR effect on SR Ca2+ leak, despite evidence of potent PKA activation. Thus cAMP (and its ability to activate Epac) would not seem sufficient to activate SR Ca2+ leak in that study. They suggested something upstream of cAMP (e.g. arrestin) might mediate the non-PKA-mediated effect of β-AR on SR Ca2+ leak. More work will be needed to see if the pathways in these two studies converge.
A limitation in the study of Pereira et al. (2007) is that they relied exclusively on 8-CPT to activate Epac. While this may activate Epac very potently, 8-CPT activation may differ from that by β-AR in a couple of ways. The affinity of cAMP for PKA is much higher than for Epac, leaving open the question of whether physiological activation of β-AR might not achieve comparable Epac activation. It would be valuable to know if combined β-AR activation and PKA inhibition produces comparable Ca2+ handling effects to the direct activation of Epac by 8-CPT.
Further information about subcellular localization of Epac signalling in myocytes would be valuable. If Epac is mainly in mAKAP local signalling complexes at the nuclear envelope (Dodge-kafka & Kapiloff 2006), how is the Epac–CaMKII signal transmitted to the RyR at the sarcolemmal-SR junctions? Are other CaMKII targets influenced by this Epac-dependent pathway (ICa, phospholamban, transcription)? How does Epac activate CaMKII (is Ca2+ release from stores involved)?
Why have parallel cAMP signalling pathways such as Epac and PKA? Is the PKA pathway geared to acute transient fight-or-flight responses, whereas Epac (and its putative CaMKII, calcineurin and ERK5 effects) is more involved in longer term adaptive changes in the myocyte? Moreover, this new awareness of PKA-independent cAMP signalling may mean that some cAMP-dependent effects previously attributed to PKA should be re-examined to discern whether Epac might be involved (e.g. more careful studies with PKA inhibition, coupled with selective Epac activation).