Severe platelet dysfunction in NHL patients receiving ibrutinib is absent in patients receiving acalabrutinib

Abstract

The Bruton tyrosine kinase (Btk) inhibitor ibrutinib induces platelet dysfunction and causes increased risk of bleeding. Off-target inhibition of Tec is believed to contribute to platelet dysfunction and other side effects of ibrutinib. The second-generation Btk inhibitor acalabrutinib was developed with improved specificity for Btk over Tec. We investigated platelet function in patients with non-Hodgkin lymphoma (NHL) receiving ibrutinib or acalabrutinib by aggregometry and by measuring thrombus formation on collagen under arterial shear. Both patient groups had similarly dysfunctional aggregation responses to collagen and collagen-related peptide, and comparison with mechanistic experiments in which platelets from healthy donors were treated with the Btk inhibitors suggested that both drugs inhibit platelet Btk and Tec at physiological concentrations. Only ibrutinib caused dysfunctional thrombus formation, whereas size and morphology of thrombi following acalabrutinib treatment were of normal size and morphology. We found that ibrutinib but not acalabrutinib inhibited Src family kinases, which have a critical role in platelet adhesion to collagen that is likely to underpin unstable thrombus formation observed in ibrutinib patients. We found that platelet function was enhanced by increasing levels of von Willebrand factor (VWF) and factor VIII (FVIII) ex vivo by addition of intermediate purity FVIII (Haemate P) to blood from patients, resulting in consistently larger thrombi. We conclude that acalabrutinib avoids major platelet dysfunction associated with ibrutinib therapy, and platelet function may be enhanced in patients with B-cell NHL by increasing plasma VWF and FVIII.

Graphical abstract

Figure 1.

Acalabrutinib and ibrutinib therapy cause dysfunctional GPVI-mediated platelet aggregation. PRP from patients receiving ibrutinib (Ibr; n = 6), acalabrutinib (Acal; n = 7), healthy donors (n = 9), or Btk inhibitor naïve CLL patients (n = 5) was loaded into 96-well microtiter plates containing lyophilized platelet agonists. Plates were shaken for 5 minutes at 37°C and aggregation was measure by light transmission to give an end point measurement. (A) Scatter plot of EC₅₀ values calculated from concentration response curves for ADP, U46619, epinephrine, TRAP-6, CRP-XL, and collagen. Each point represents the response of a patient receiving acalabrutinib (red triangle) or ibrutinib (green square); healthy donor (gray circle) or Btk inhibitor naïve CLL patient (white diamond). NR, patients that did not respond to an agonist at the highest concentration tested. Concentration response curves for GPVI agonists (B) CRP-XL and (C) collagen; each point represents the mean % aggregation values ± standard error of the mean (SEM). Significant differences relative to healthy donors were tested by 2-way ANOVA with the Turkey multiple comparisons test. Aggregation of washed platelets from healthy donors pre-treated with a range of (D) acalabrutinib or (E) ibrutinib concentrations and then stimulated with 3 to 0.01 µg/mL CRP-XL.

Figure 2.

SFK activation and adhesion to collagen is spared by acalabrutinib but not ibrutinib. Washed human platelets from healthy donors were preincubated with a range of (A) acalabrutinib or (B) ibrutinib and then stimulated with 1 µg/mL CRP-XL for 3 minutes at 37°C; tyrosine phosphorylation of Src (Y418) and Btk (Y223) were measured by western blot using site-specific antibodies and total levels of Tec tyrosine phosphorylation were measure by ELISA. Points represent mean levels of tyrosine phosphorylation relative to vehicle-treated controls ± SEM. (C) Representative images of phosphoblots. Washed platelets treated with a range of acalabrutinib, ibrutinib, or dasatinib concentrations and stimulated as previously and blotted for (D) tyrosine phosphorylation of Lyn (Y396) or (E) total phosphotyrosine using a site-specific antibody or 4G10, respectively. Points represent mean levels of tyrosine phosphorylation relative to vehicle-treated controls ± SEM. (F) Aggregation 1 µg/ml CRP-XL stimulated of washed platelets was performed in 96-well plates after treatment with a range of concentrations of acalabrutinib, ibrutinib, or dasatinib concentration response curves are mean platelet aggregation ± SEM relative to vehicle-treated control. (G) Acalabrutinib, ibrutinib, or dasatinib-treated washed platelets from healthy donors were allowed adhere to collagen-coated wells of a 96-well plate after 45 minutes at 37°C; graphs plot concentration response curves to mean numbers of adhered platelets ± SEM relative to vehicle treatment. (H) Representative images of phosphoblots for platelets treated with tyrosine kinase inhibitors.

Figure 3.

Acalabrutinib inhibits signal transduction downstream of GPVI with lower potency than ibrutinib. Washed human platelets from healthy donors were preincubated with a range of acalabrutinib or ibrutinib concentrations and then stimulated with 1 µg/mL CRP-XL for 3 minutes at 37°C and phosphorylation of (A) PLCγ2 (Y759) and (B) PKC substrates (S/T) was measured by western blotting. (C) Representative western blot images for PLCγ2 (Y759) and S/T phosphorylated PKC substrates. (D) Cytosolic Ca²⁺ following stimulation with 1 µg/mL CRP-XL was measured in fura-2 loaded platelets in real time for 5 minutes. Points represent the mean response relative to vehicle ± SEM. Representative [Ca²⁺]_i traces following incubation with (E) acalabrutinib or (F) ibrutinib.

Figure 4.

Acalabrutinib is a less potent inhibitor of integrin α_IIbβ₃activation and granule secretion than ibrutinib. Washed human platelets from healthy donors were preincubated with a range of acalabrutinib or ibrutinib concentrations and then stimulated with 1 µg/mL CRP-XL for 20 minutes. (A) P-selectin and (B) fibrinogen binding was measured by fluorescence-activated cell sorting. (C) Platelets were stimulated with 1 µg/mL CRP-XL for 3 minutes at 37°C. Phosphorylation of β3 (Y773) was measured by western blotting. Points represent mean responses relative to vehicle ± SEM. (D) Representative β3 (Y773) blots.

Figure 5.

Ibrutinib, but not acalabrutinib, causes thrombus instability. Whole blood from healthy donors was incubated with vehicle, 1 µM ibrutinib, and 1 µM acalabrutinib for 20 minutes and flowed through collagen-coated microfluidic flow chambers at a shear rate of 1000 s⁻¹ for 6 minutes. (A) Representative images of thrombi formed under each of the conditions tested at 2-minute intervals. (B) A summary of effects of 1 µM ibrutinib and 1 µM acalabrutinib on thrombus formation on collagen and other platelet signaling and functional measurements relative to vehicle-treated control. Bars represent the mean ± SD; P values were calculated using multiple tests with Holm-Sidak correction for multiple comparisons.

Figure 6.

Thrombus formation on collagen correlates with platelet count during treatment with acalabrutinib but not with ibrutinib. (A) Blood from healthy donors (n = 6), Btk-inhibitor naïve CLL patients (n = 5), patients receiving ibrutinib (n = 6) or acalabrutinib (n = 8) was perfused through collagen-coated microfluidic flow chambers at room temperature at a shear rate of 1000 s⁻¹ before being fixed with 10% formyl saline. Fixed samples were stained with DioC6 and z-stack images acquired to enable estimation of thrombus volume. (A) Representative images of thrombi in blood from a healthy donor and patients in the presence or absence of Haemate P (intermediate factor VIII) added ex vivo. Thrombus volume plotted against platelet count for (B) Btk inhibitor naïve CLL patients; patients treated with (C) acalabrutinib or (D) ibrutinib with or without added Haemate P. Correlation between platelet count and thrombus volume is significant in the presence or absence of Haemate P for acalabrutinib patients (significantly nonzero, P ≤ .05) but nonsignificant for ibrutinib patients (P > .05).

Figure 7.

Platelet function is improved by ex vivo addition of Haemate P. Comparison of the volume of thrombi formed in the presence or absence of Haemate P added ex vivo to blood from healthy donors (n = 6), Btk inhibitor naïve CLL patients (n = 5), or patients receiving ibrutinib (n = 6) or acalabrutinib (n = 8, as described in Figure 7). Two-way ANOVA with repeated measures and Sidak multiple comparisons test.