Cardiac Beta3-Adrenergic Receptors in the Clinical Arena: The End of the Beginning

This article refers to ‘The first-in-man randomized trial of a beta3 adrenoceptor agonist in chronic heart failure: the BEAT-HF trial’, by H. Bundgaard et al., published in this issue on pages 566–575. Despite significant advances in treatment options, heart failure remains a devastating disease with significant morbidity/mortality and economic burden. In this context, new therapeutic approaches are eagerly needed. The paper by Bundgaard in this issue of the journal1 proposes a treatment target, the third isotype of beta-adrenergic receptors (β3 AR) that was identified in the human myocardium >15 years ago,2 but only recently tested in the clinical arena thanks to the availability of a specific agonist, mirabegron. The molecule is available for clinical use in urology for the treatment of overactive bladder disease, a painful condition of frequent micturition, based on its well-established relaxing effect on the bladder smooth muscle that expresses β3 AR.3 Bundgaard et al. now present the results of a first pilot trial of mirabegron in patients suffering from heart failure with reduced ejection fraction (HFrEF), the rationale of which was based on pre-clinical data pointing to beneficial effects of β3 AR activation in the failing myocardium. Although the trial did not meet its primary endpoint (increase in LVEF over placebo), a non-pre-specified, subgroup analysis in patients with the lowest EF suggests potential benefits in subjects with more severe LV dysfunction. β3 ARs are clearly different from the two other well-known isotypes of β AR. In contrast to β1 and β2 ARs, their expression increases in the diseased myocardium2 and they are resistant to desensitization.4 i.e. they remain coupled to intracellular downstream signalling and maintain their effect in the presence of high concentrations of catecholamines. This already makes them rather attractive targets in the setting of heart failure. Their coupling and effects on rhythm and contractility also contrast with classical β1 and β2 adrenergic responses. In human ventricular biopsies in organ baths,2, 5 but also in non-failing ventricular tissue from mice with cardiac-specific expression of the human β3 AR,6 high concentrations of agonists induce a decrease of unloaded contractile shortening. A similar slight negative inotropic effect was also observed by Bundgaard and colleagues after administration of β3 AR agonists in non-failing sheep in vivo.7 However, the effect was exactly the opposite, i.e. an increase of LV contractility by activating β3 ARs in sheep with heart failure.7 Admittedly, others have found discordant results with β3 AR modulation in other animal models8 that would contradict the proposition of Bundgaard and collaborators. Their merit was to test the hypothesis in a pilot human trial. They also built upon solid pre-clinical data supporting protective effects of β3 ARs on the myocardium. Mechanistically, β3 ARs couple to activation of nitric oxide synthases (NOSs) in rodent and human myocardium;5, 9 the resulting increase in cGMP and cGMP-dependent protein kinase I (cGKI) activation in part explains the attenuation of excitation–contraction coupling and contractility in the normal cardiac cell, but also the protection in failing myocytes; cGKI acts on at least three different targets to attenuate hypertrophic remodelling,10 but also activates protein phosphatase 2A to decrease NADPH oxidase activation, thereby decreasing reactive oxygen species production. In the authors' hypothesis, this antioxidant effect attenuates the oxidative inactivation of the Na+–K+-ATPase pump, thereby restoring the pump activity.11 This, in turn, reduces intracellular Na in cardiac myocytes. The authors reasoned that, in non-failing myocytes, this decrease in Na would attenuate Ca transients and reduce contractility. However, in failing myocytes with Na overload, restoration of the pump activity would reduce the excess intracellular Na and restore excitation-contraction coupling and contractility.7 This was the rationale to test mirabegron (a β3 AR agonist) in HFrEF. In the present pilot human trial, they randomized patients with HFrEF under standard therapy to either mirabegron or placebo for 6 months; although the current recommended dose for the urological indication is 50 mg/day, the protocol included a dose escalation up to 300 mg/day, which was attained in >90% of the patients. Notably, all patients had to be treated with beta-blockers per protocol before inclusion (the majority received carvedilol or metoprolol, at >75% of recommended target doses in both arms), which probably protected those in the active treatment arm from off-target effects of mirabegron (i.e. agonism at β1 and β2 ARs) at such high doses. Also, none of the patients was treated with digoxin (which would have obliterated the expected beneficial effect of mirabegron through Na+–K+-ATPase activation). Patients were screened by echocardiography and included on the basis of LVEF <40%. Although this underestimated LVEF, as measured by cardiac computed tomography (CT) at enrolment, both sample size calculation and measurement of the primary endpoint (LVEF before/after treatment) were done by cardiac CT. Nevertheless, most of the patients had moderate HFrEF (EF 38 ± 17% by CT) with >95% at NYHA class II. The aetiology of heart failure was ischaemia in ∼50% in both arms, and all were in sinus rhythm. The trial was powered to detect a change of 4% LVEF by CT. The primary endpoint (increase in LVEF by CT after 6 months) was not reached. Perhaps this is not unexpected, given the relatively mild starting LV dysfunction, consistent with other functional and biological parameters (6-min walk test, VO2 max, and NT-proBNP), making it a challenge for any additional drug to improve LV function further over standard therapy. In an exploratory subgroup analysis of patients with baseline CT LVEF <40%, eight patients under mirabegron improved LVEF from baseline to follow-up [mean difference: 5.9% (0.4–11.4%); P = 0.025], while 17 patients under placebo showed no significant change [mean difference: 0.4% (−2.2 to 2.9%); P = 0.88]. Although the numbers are too small for a definitive mechanistic explanation, this increase in LVEF seemed to be due mostly to a decrease in systolic volumes, while there was no change (i.e. further decrease) in heart rate (as expected under full background β1–2 blockade). Based on their mechanistic hypothesis and these preliminary results, the authors expect a higher benefit of the β3 AR agonist in patients with more severe LV dysfunction. However, one may question such a simple conclusion. First, the endpoint chosen here, LVEF, may not be the best predictor of outcome in HF patients, and it remains to be demonstrated that the (modest) increase in LVEF translates into clinical (i.e. morbidity, other functional parameters, mortality) benefit. Interventional trials with positive inotropic drugs have mostly failed to demonstrate improved survival.12 Nevertheless, β3 AR agonists may confer benefits through other effects on LV remodelling that may be revealed in other trials with more mechanistic endpoints. Indeed, β3 ARs are expressed not only in cardiac myocytes, but also in endothelial cells of the human coronary vasculature, where they mediate both nitric oxide-dependent and independent relaxations that persist in coronary resistance vessels from diseased patients.13 In the remodelling myocardium, this would contribute to maintain coronary perfusion and avoid further ischaemic damage. In animal models, activation of β3 ARs on cardiac myocytes also attenuates hypertrophic remodelling and fibrosis in response to neurohormonal stress or haemodynamic overload.9 In combination, these effects may prevent adverse myocardial remodelling and improve not only systolic but also diastolic function. Such parameters were not evaluated in the BEAT-HF trial, but are being studied in a multicentre European trial on the effect of mirabegron in patients with structural heart disease (stage B) with, or at risk of developing heart failure with preserved EF (ISRCTN 65055502). Of note, a reduction of intracellular Na overload, as proposed through β3 AR activation, would also be beneficial in the context of diastolic dysfunction.14 Also, the combination of β3 AR agonists with β1 AR blockers, which seemed to be well tolerated in the present pilot trial, may be particularly suited given previous evidence for a mutual potentiation on downstream NOS/cGMP signaling.15 Finally, activation of β3 AR may confer indirect cardiovascular benefits through activation of extracardiac targets, e.g. on beige/brown fat, and subsequent improvements in peripheral metabolism and insulin sensitivity.16 Whether any or all of these effects will translate into clinically measurable variables and improved outcome will need to be confirmed from the results of trials ongoing or yet to come. In this respect, it is good news that the authors of the BEAT-HF trial decided to launch a larger, more mechanistic trial in HFrEF, while the Beta3lvh trial is exploring more long-term effects on remodelling, endothelial function, and metabolism. At the very least, the BEAT-HF trial demonstrated very few or no side effects (under full β1–β2 AR blockade) in HFrEF patients, which is also reassuring for those patients currently taking the drug for urological indications. This work was supported by grants from the Fonds National de la Recherche Scientifique (FNRS; PDR T.0144.13) and the European Union (UE Horizon2020 GA634559) to J.L.B. Conflict of interest: none declared.