Phosphatase regulatory subunits in beta-adrenergic signalling: a delicate balancing act

This editorial refers to ‘Role of type 2A phosphatase regulatory subunit B56α in regulating cardiac responses to β-adrenergic stimulation in vivo’, by S.-L. Puhl et al., pp. 519–529. Catecholamines enhance all aspects of cardiac function mainly through activation of beta1-adrenergic receptors (AR) among the three isotypes1 expressed on cardiac myocytes (CM). This is well known to be mediated through downstream increased phosphorylation of proteins involved in excitation–contraction (EC) coupling. While a lot of attention has been dedicated to the identification and regulation of kinases2 mediating these events, the other side of the coin, de-phosphorylation by phosphatases has been less explored. Puhl et al.3 build upon a previous pioneering study by their group4 that used an unbiased proteomic approach to analyse the CM myofilament sub-proteome, including phosphoproteins, before and after beta-AR stimulation. In that study, they had identified the B56alpha subunit of protein phosphatase 2A (PP2A) as dynamically translocated from the Z-line to a more diffuse cytosolic location after isoproterenol (Iso) treatment. PP2A is a serine/threonine phosphatase composed of three subunits, A (scaffolding), B (regulatory), and C (catalytic), with B regulatory subunits belonging to a large family encoded by alternatively spliced genes. This allows for a diversity of associations of heterotrimeric subunits, probably to subserve target specificity of the phosphatase in particular cell types and locations.5 This and other previous studies6 suggested a potential role of PP2A-B56alpha in the regulation of EC coupling by targeting the phosphatase to specific regulatory proteins in response to beta-AR stimulation, just as AKAPs do for A-kinase.7 The independent observation that B56alpha is downregulated in human failing hearts8 further hinted at pathophysiologic relevance. In the present article, Puhl et al.3 used a mouse model with systemic genetic deletion of B56alpha to test this hypothesis. In their detailed analysis, they characterized the phenotype of wild-type (WT), hypomorphic (HET) and homozygous (HOM) knockout mice (of both genders) ex vivo and in vivo at 2 and 6 months, at baseline, under acute dobutamine (DOB) stimulation or after ‘chronic’ (2 weeks) treatment with Iso to produce remodelling. They further isolated adult mouse CM to analyse specific phosphoproteins known to regulate EC coupling and cardiac contractility. HET or HOM genetic deletion produced a neutral cardiac phenotype, i.e. not different from WT, in the basal state, at 2 months in either sex; HOM male hearts showed a slight decrease in weight normalized to tibia length by gravidometry at 6 months only. Upon DOB stimulation, HOM mice developed lower ejection fraction (EF) and fractional shortening (FS) compared with WT, but similar increases in heart rate. However, acute stimulation of isolated CM with Iso did not reveal any appreciable difference in the phosphorylation of phospholamban (PLB), troponin I (TnI), or myosin binding protein-c (MyBP-c) between genotypes; phosphorylation of the ryanodine receptor (RYR2) on Ser2814 was lower at baseline in HOM hearts, but unchanged by Iso in CM of both genotypes. A 2-week stimulation with Iso (30 µg/kg/min, by osmotic pump) produced a mild hypertrophy in WT, without echocardiographic, histologic, or molecular signs of pathologic remodelling (concentric-not excentric or dilated-hypertrophy, absence of fibrosis or enhanced expression of the ‘foetal gene program’); the hypertrophic effect of Iso was blunted in HOM mice. Surprisingly, their echocardiographic indexes of left ventricular (LV) function (FS, EF) were now conserved. While the genetic deletion of B56alpha does alter the beta-AR effect, both on acute LV function and hypertrophy upon prolonged infusion, the exact mechanism is difficult to identify. Most of the putative target phosphoproteins involved in beta-AR signalling remained unchanged in HOM mice and although HOM mice exhibit a reduced RyR2 phosphorylation on Ser2814 at baseline, this is not reproduced upon acute activation of their isolated CM with Iso. Also, Pulh et al. carefully verified the impact of B56alpha deletion on PP2A activity, which they found to be decreased, so that the reduced RyR2 phosphorylation would have to result from an indirect (yet unknown) mechanism. (Hyper)phosphorylation of RyR2 on Ser2814 has been mostly implicated in sarcoplasmic reticulum (SR) calcium ‘leak’ upon activation of calcium-calmodulin kinase II (CamKII) in pathologic hypertrophic remodelling and heart failure,9 so that reduced phosphorylation on this residue (as observed here) may have been protective. Also, mice expressing a recombinant RyR2 non-phosphorylatable on Ser2814 are resistant to hypertrophy in response to transaortic constriction10 or to fibrosis after a 4-week infusion of Iso,11 which are rather harsh stimuli. However, this CamKII pathway is unlikely to have been recruited by the mild hypertrophic stimulus used here. Perhaps more insight could be obtained from (future) experiments analysing EC coupling (calcium transients, calcium sparks), as well as the phenotype of these B56alpha HOM mice under harsher stress. Others have generated similar mouse models with genetically altered B56alpha expression, which unfortunately provide little help in clarifying the concept. A two-fold overexpression of B56alpha resulted in increased basal contractility and myofilament calcium sensitivity in the basal state, correlated with decreased phosphorylation of TnI and MyBP-c; but the response to Iso was inversely reduced, with decreased phosphorylation of PLB on Ser16.12 Another model of B56alpha genetic deletion13 has produced a phenotype at odds with the results of the present study, as acknowledged by the authors; indeed, although this other study has also found reduced RyR2 phosphorylation (and calcium waves/sparks), this was ascribed to increased (not decreased) PP2A activity, together with functional consequences mostly on heart rate (unlike the present study). While Puhl et al. propose a technical explanation for the divergent results in PP2A activity (readout with a chromogenic substrate non-specifically targeted by various phosphatases), this other study tried to discriminate the PP2A-specific activity from total phosphatase activity using a (relatively) selective PP2A inhibitor.13 Differences in the stoichiometry of B56alpha vs. other subunits between genetic models more likely account for the discrepancy. The fact that a residual PP2A activity was still measurable in the present study despite complete loss of B56alpha, but in presence of upregulated alpha4 subunit, illustrates the importance of such compensatory adaptations in chronic genetic deletion models. What, then, of the observed decrease in B56alpha in human heart failure?8 It is important to remember that this was measured at a relatively advanced stage of pathology, unlike the observations in the mouse models above, and not as the result of a genetic manipulation with complete abrogation, which probably represents an extreme in the stoichiometric relationship, as discussed earlier. Of interest, adenoviral expression (in isolated rat CM) of the microRNA, miR-1 (implicated in cardiac pathology) decreased B56alpha (one of its validated targets), which was paralleled with hyperphosphorylation of RyR2 on Ser2814 (the CamKII site), calcium sparks and arrhythmia.14 Although the result on RyR2 phosphorylation is again at odds with the present study, one finding consistent with the earlier observation of Avkiran’s group is the change in subcellular localization of PP2A (catalytic subunit), from the Z-line to a more diffuse distribution in the cytosol upon downregulation of B56alpha. How this may be driven by (some amount of) adrenergic stimulation in the context of cardiac pathology remains to be clarified, if to provide some insight for targeted therapy. Altogether, divergent phenotypes and effects on phosphorylation perhaps are not unexpected from perturbations of PP2A, a complex effector phosphatase, the regulation of which is notoriously different between human and animal models, across cardiac chambers or even between cardiac cell types.15 Perhaps this also shows the methodological limitation of classic genetic deletion/overexpression models, and the need for more comprehensive approaches taking into account the disease stage and combined effects on transcription, post-transcriptional modulation not only of the effector of interest, but also signalling partners in a ‘network’ approach.16 Conflict of interest: none declared.