The Cardiac Re-AKT-Ion to Chronic Volume Overload

This article refers to ‘Molecular and structural transition mechanisms in long-term volume overload’†, by BA Mohamed et al., published in this issue on pages 362–371. In response to pathological changes of haemodynamic load and neuroendocrine activation, the left ventricle (LV) undergoes structural remodelling, and a central feature of this process is cardiomyocyte hypertrophy. Increased afterload can be imposed by arterial hypertension, aortic stenosis and/or neuroendocrine activation with ensuing vasoconstriction, and elevated preload by aortic (AR) or mitral regurgitation (MR) as well as renal dysfunction with sodium and volume retention. While pathological LV hypertrophy is a major risk factor for the development of heart failure,1 the heart can also develop an adaptive form of hypertrophy in response to physiological stress, such as exercise2 or pregnancy.3 These different types of hypertrophy can be differentiated by microscopic, functional and molecular criteria, where adaptive hypertrophy is typically characterized by: (i) maintained systolic and diastolic function; (ii) improved myocardial microvascularization by enhanced angiogenesis; (iii) the absence of fibrosis; and (iv) on a molecular level, activation of, for example, vascular endothelial growth factor- (VEGF-) and/or insulin and insulin-like growth factor 1 (IGF1) receptors, signaling to phosphatidylinositide 3-kinase (PI3K) and protein kinase B/AKT.4, 5 In contrast, maladaptive cardiac remodelling is characterized by (among others): (i) diastolic and/or systolic dysfunction; (ii) decreased or unchanged myocardial vascularization; (iii) cell death (through apoptosis or necrosis) and fibrosis; and (iv) on a molecular level, activation of Ca2+/calmodulin-dependent protein kinase II (CaMKII), calcineurin and various mitogen-activated proteins (MAPs) kinases, and oxidative stress.6 The type of cardiac remodelling depends on the type of hemodynamic load (Figure 1). In a seminal study, Toischer et al.7 used transaortic constriction (TAC) to increase afterload and an aortocaval shunt operation to increase preload in mice and observed typical characteristics of maladaptive remodelling in response to afterload, but mostly features of adaptive remodelling in response to preload, despite similar total wall stress.7 On a molecular level, activation of CaMKII, inflammation, apoptosis, and fibrosis stood out in response to afterload, but were absent after preload elevation. In contrast, the protein kinase B/AKT pathway was activated by preload, but not afterload. These differences were apparent by 1 week after the respective intervention and, in the longer run (after 8 weeks), led to concentric hypertrophy with more severe systolic dysfunction in response to afterload, but rather eccentric hypertrophy with only mildly impaired systolic function after an elevation of preload.7 In the current issue of the European Journal of Heart Failure, Mohamed et al.8 now extend their previous findings7 by analysing in more detail the long-term effects of volume overload on the LV. They observe that 20 weeks after the shunt procedure, LV dilation and systolic dysfunction induce global cardiac decompensation with pulmonary and hepatic congestion. In contrast to the early changes after 1 week,7 CaMKII is now activated and AKT inactivated, resembling the maladaptive remodelling pattern induced by afterload elevation already after 1 week.7 They propose that the reverse transition of these two particular signaling pathways in the course of volume overload may be a critical step in the transition from adaptive to maladaptive remodelling, resulting in functional deterioration and decompensation (Figure 1). Previous studies have revealed the causal role of CaMKII in mediating maladaptive cardiac remodelling and its progression to heart failure in response to elevated afterload.7, 9 In contrast, activation of AKT is required for adaptive cardiac hypertrophy10 and ameliorates maladaptive cardiac remodelling in response to pressure overload.4 Here, Mohamed et al.8 convincingly demonstrate that the early activation of AKT that occurred after 1 week7 protects the heart for some time from volume overload-induced heart failure, as the relatively slow progression to LV dysfunction, decompensation, and death after the shunt operation was accelerated when AKT was genetically inactivated. The observation that after long-term volume overload (but not short-term volume overload) programmed cell death (apoptosis) was increased is in agreement with the pro-apoptotic role of CaMKII11 and anti-apoptotic effects of AKT.12 Finally, the late occurrence of apoptosis that parallels cardiac dilation and decompensation is in agreement with a possibly causal role of low levels of apoptosis for the development of dilated cardiomyopathy and death13 as well as the early observation that in human terminal heart failure, the rates of apoptosis are increased compared with non-failing hearts.14 So how should we put the data of this8 and the previous study of Toischer et al.7 into a larger clinical context (Figure 1)? To do so, let us consider again in some more detail the experimental intervention of the aortocaval shunt. While diastolic wall stress (preload) was increased by 277% by the shunt,7 this bypassing of the (high-resistance) peripheral circulation must reduce afterload. In fact, when estimating the effective arterial elastance (Ea) from the representative pressure–volume loops presented in the earlier study,7 vascular afterload is clearly reduced in the shunt model.15 Therefore, this model—at least qualitatively—resembles the haemodynamic changes in pregnancy (Figure 1), where the placenta is a functional shunt that (together with elevated VEGF and other vasodilating factors) decreases vascular resistance.16 Interestingly, pregnancy-induced hypertrophy is also mediated by AKT17 and is not associated with elevated levels of BNP,18 which is in agreement with adaptive AKT activation and lack of BNP upregulation in response to volume overload by the shunt at the earlier time-point after 1 week.7, 8 When pregnancy is complicated by hypertension and/or pre-eclampsia, by which afterload increases, then the combined increase of preload and afterload indeed increases BNP levels (Figure 1).18 Concerning valvular heart disease, acute MR creates volume overload with unchanged or slightly decreased afterload, which is explained by the unloading of the LV with ejection of blood into the left atrium instead of the higher resistance arterial vascular system.19 At early stages of (primary) MR, diastolic function can be even supernormal, and even at later stages is rather uncompromised,19 in agreement with a rather adaptive remodelling phenotype (Figure 1). With persistence and/or progression of MR, however, afterload returns to normal or even increases (Figure 1), related presumably to sympathetic activation.19, 20. In support of this, beta-blockade improved LV function after experimental MR in dogs, and this was associated with improvement of contractility of isolated cardiac myocytes from these hearts.20 In the present study,8 long-term maladaptive remodelling in response to volume overload was associated with downregulation of the sarcoplasmic reticulum Ca2+ ATPase (SERCA), a central defect of Ca2+ handling in heart failure that is reversed by beta-blocker treatment in human patients.21 Therefore, one may speculate that, in long-term volume overload, sympathetic activation, which, when sustained, signals to CaMKII rather than to protein kinase A22 through β-adrenergic receptors, may indeed contribute to the transition from the adaptive to maladaptive remodelling observed in the current study by Mohamed et al.8 (Figure 1). In contrast to MR, where afterload is mostly unchanged, AR is associated with an increase of both preload and afterload (Figure 1), where elevated afterload results from the increased aortic stroke volume as the sum of forward stroke volume plus the regurgitant volume. At early stages, the LV with preserved systolic function can use the Frank–Starling mechanism to compensate this afterload mismatch (‘preload reserve’). At later stages, however, this preload reserve is exhausted because of progressive LV dilatation and systolic dysfunction, resulting in an ‘afterload mismatch’.23 At such a late stage of maladaptive cardiac remodelling, improvement of LV ejection performance after aortic valve replacement is related mostly to the decrease in cardiac afterload rather than to an improvement of LV contractile function per se,24 suggesting that indeed, combined preload and afterload in AR can induce a heart failure phenotype comparably severe as that in response to aortic stenosis (Figure 1).25 Mohamed et al.8 have thus to be congratulated for providing additional insights into the mechanisms of structural, functional and molecular remodelling in response to volume overload. Their data suggest that sustaining AKT signalling and/or antagonizing CaMKII-directed signalling may be promising targets to slow or even prevent maladaptive cardiac remodelling in response to pathological levels of preload, but also afterload (Figure 1). In this context, small molecules to sustain AKT signaling are currently being developed26 and may be promising tools for the treatment of patients with heart failure. C.M. is supported by the Deutsche Forschungsgemeinschaft (DFG; Heisenberg Programm, SFB-894 and Ma2528/7–1) and by the Deutsche Herzstiftung (Margret Elisabeth Strauß-örderung). Conflicts of interest: none declared.