Cardiovascular toxicity of Bruton tyrosine kinase inhibitors: forget about selectivity but watch the clock

TO THE EDITOR:

Bruton tyrosine kinase (BTK) is an important molecular driver of B-cell malignancies such as chronic lymphocytic leukemia (CLL), Waldenström macroglobulinemia, and others. Ibrutinib, a first-generation BTK inhibitor (BTKi), and acalabrutinib and zanubrutinib, second-generation BTKis, represent valuable opportunities for managing B-cell malignancies; however, all of them also cause cardiovascular (CV) toxicity in the form of hypertension, atrial flutter/fibrillation, or less frequently, ventricular arrhythmias.¹ In this article, Moslehi et al² describe the results of a pooled analysis of 10 clinical trials in which zanubrutinib was used as a single agent, primarily in the setting of CLL/small lymphocytic leukemia (SLL), and focus on head-to-head comparisons between zanubrutinib and ibrutinib in the ALPINE trial of patients with relapsed/refractory CLL/SLL and in cohort 1 of the ASPEN trial of patients with Waldenström macroglobulinemia.^3,4 The analysis illuminates zanubrutinib as a well-tolerated BTKi that causes significantly lower exposure-adjusted incidence rates of atrial flutter/fibrillation, hypertension, and ventricular arrhythmias than ibrutinib. The improved CV safety of zanubrutinib emerges even when spurious data from the ALPINE trial, which indicated similar rates of hypertension with zanubrutinib and ibrutinib, are taken into account.² The thoughtful analysis by Moslehi et al² is well in agreement with the data reported by other investigators who scrutinized the same database, albeit with different aims and approaches.⁵ 

CV toxicity of BTKis does not come as a surprise. BTK is highly homologous to TEC and Src kinases that relay homeostatic signals in the CV system⁶; moreover, ibrutinib, acalabrutinib, and zanubrutinib inhibit BTK by covalent bond to Cys481 in the ATP-binding pocket of the kinase,⁷ but Cys481 and other druggable cysteines are found in many other kinases, including PI3K and the kinase domains of Erbb2 and Erbb4 receptors, all being important CV regulators.^8-12 BTKis can thus inhibit many more kinases than just BTK, paving the road to off-target CV effects.

Why does zanubrutinib cause fewer CV events than ibrutinib? One intuitive explanation is that zanubrutinib is more selective than ibrutinib at inhibiting BTK,^2,5,13 but logical as it may sound, this assumption does not come without limitations. Colorful renditions of the human kinome exposed to BTKis suggest that both zanubrutinib and acalabrutinib show improved selectivity toward BTK^13,14 but data were generated under conditions that call for caution. The concept of selectivity embraces parameters such as the affinity of the inhibitor for the intended target (measured as Ki in binding experiments) and the potency with which the target is inhibited (measured as 50% inhibitory concentration [IC₅₀] in enzymatic assays). For covalent inhibitors, as the BTKis are, selectivity should be measured by setting the inhibitor at concentrations not too different to Ki or IC₅₀, so as to avoid artifactual inhibition of kinases that BTKis would bind at higher concentrations¹⁵; yet, 1 popular kinome assay, yielding a selectivity rank order of acalabrutinib ≥ ACP-5862 (acalabrutinib active metabolite) > zanubrutinib > ibrutinib, was generated with BTKis at 1 μM, which was orders of magnitude higher than IC₅₀ or Ki values.¹⁴ In a different kinome assay, yielding a selectivity rank order of zanubrutinib > acalabrutinib > ibrutinib > ACP-5862, BTKis were used at 100× IC₅₀¹³; this was partially appropriate for ibrutinib and zanubrutinib, whose concentrations remained lower than Ki, but it was not so for acalabrutinib and ACP-5862, whose concentrations were overly higher than Ki (Table 1). The bright side of this story is that zanubrutinib seems to be more selective than ibrutinib in both kinome assays, which in principle would be consistent with its improved CV liability. The dark side of the story is that the rank order of selectivity of the 3 BTKis changes from 1 experimental setting to another, making a comprehensive picture difficult to obtain at this point in time.

Having considered all the caveats and uncertainties around selectivity, I suggest that looking at the characteristics of patient candidates for BTKis may help scrutinize these drugs from different viewpoints. Advanced age and comorbidities fingerprint most of these patients. Moslehi et al report that ∼6% of them presented at treatment with a history of prior atrial fibrillation,² a figure consistent with a Mayo Clinic retrospective analysis of 2444 patients diagnosed with CLL/SLL.¹⁶ Perhaps more importantly, the Mayo Clinic analysis showed that patients without prior atrial fibrillation eventually developed it during follow-up at an apparent rate of 1% per year, in spite of that no more than one-third received active treatment and no more than 3% of them received ibrutinib as part of treatment.¹⁶ These facts illuminate atrial fibrillation (and hypertension) as inevitable pathophysiologic events in a population of older comorbid patients and suggest that BTKis might be viewed as “triggers” rather than etiological agents of CV events.

Possible differences between ibrutinib and second-generation BTKis as triggers of CV events were characterized for the first time in a phase 3 trial of ibrutinib vs acalabrutinib in patients with previously treated CLL.¹⁷ In that study, the median time to any grade atrial fibrillation was significantly longer for acalabrutinib than for ibrutinib; however, the curves of cumulative incidence eventually aligned after ∼50 months of follow-up. Pointing on head-to-head comparisons in ASPEN cohort 1 or in the ALPINE trial, Moslehi et al show that atrial fibrillation occurs early during the first 6 months of treatment with ibrutinib but develops more slowly during treatment with zanubrutinib.² Notably, and different from acalabrutinib, the curve of zanubrutinib-triggered atrial fibrillation did not align with ibrutinib at >40 or 50 months follow-up in ALPINE³ or ASPEN cohort 1,⁴ respectively. Pending the availability of new and more extensive data on hypertension in head-to-head comparisons of zanubrutinib vs ibrutinib, this focus on time to atrial fibrillation may help identify different means for differentiating BTKis, with both acalabrutinib and zanubrutinib proving less potent than ibrutinib at accelerating the development of events that otherwise would occur at a later time point; moreover, zanubrutinib holds promise as a BTKi with both longer time to, and reduced overall incidence of, CV events. The clinical management of patients treated with BTKis should perhaps be tailored to such differentiation so as to intensify surveillance over the time window when the risk of triggered events increases. Watching the clock may thus become important (Figure 1).

Shifting the focus from selectivity toward time to event may also help refine research directions; whereas better characterization of selectivity remains a must by definition, understanding the kinetics of off-target effects over time emerges as an equally important goal. How do BTKis approach unintended kinases in peripheral compartments and cell types that express such kinases? How does the irreversible inhibition of off-target kinases compare with target resynthesis and steady-state concentrations of BTKis at that target? Questions such as these should be answered soon; the clock is ticking.

Contribution: G.M. designed and wrote the manuscript.

Conflict-of-interest disclosure: G.M. reports consultancy and honoraria from Johnson & Johnson Innovative Medicine, BeiGene, Incyte, and Servier.

Correspondence: Giorgio Minotti, Department of Medicine, University and Fondazione Policlinico Universitario Campus Bio-Medico, Via Alvaro del Portillo 21, Rome 00128 Italy; email: g.minotti@policlinicocampus.it.