Leukemia (2014), 1–5 © 2014 Macmillan Publishers Limited All rights reserved 0887-6924/14 www.nature.com/leu
ORIGINAL ARTICLE
Ibrutinib inhibits collagen-mediated but not ADP-mediated platelet aggregation S Kamel1, L Horton1, L Ysebaert2, M Levade3,4, K Burbury5, S Tan1, M Cole-Sinclair1, J Reynolds6, R Filshie1, S Schischka1, A Khot5, S Sandhu5, MJ Keating7, H Nandurkar1,8 and CS Tam1,5,8 The BTK (Bruton’s tyrosine kinase) inhibitor ibrutinib is associated with an increased risk of bleeding. A previous study reported defects in collagen- and adenosine diphosphate (ADP)-dependent platelet responses when ibrutinib was added ex vivo to patient samples. Whereas the collagen defect is expected given the central role of BTK in glycoprotein VI signaling, the ADP defect lacks a mechanistic explanation. In order to determine the real-life consequences of BTK platelet blockade, we performed light transmission aggregometry in 23 patients receiving ibrutinib treatment. All patients had reductions in collagen-mediated platelet aggregation, with a significant association between the degree of inhibition and the occurrence of clinical bleeding or bruising (P = 0.044). This collagen defect was reversible on drug cessation. In contrast to the previous ex vivo report, we found no in vivo ADP defects in subjects receiving standard doses of ibrutinib. These results establish platelet light transmission aggregometry as a method for gauging, at least qualitatively, the severity of platelet impairment in patients receiving ibrutinib treatment. Leukemia advance online publication, 2 September 2014; doi:10.1038/leu.2014.247
INTRODUCTION Ibrutinib is an irreversible inhibitor of Bruton’s tyrosine kinase (BTK) with significant activity in chronic lymphocytic leukemia (CLL), mantle cell lymphoma (MCL) and Waldenstrom’s macroglobulinemia. The drug has recently been approved by the United States Food and Drug Administration for the treatment of CLL and MCL.1–4 Ibrutinib is generally well tolerated but is associated with an increased risk of bleeding. For example, the phase I/II studies of ibrutinib reported increased bruising in 17% of subjects, and intracranial hemorrhages occurred in 2%, leading to current recommendations for avoidance of concomitant warfarin therapy, and for interruption of ibrutinib for 3–7 days before and after invasive procedures.1,3 In the phase III study comparing ibrutinib with ofatumumab in the treatment of CLL, bleeding-related adverse event rates were 44% for ibrutinib vs 12% for the control arm, although the rates of serious bleeding were low in both arms (1% vs 2%, respectively).2 There are compelling reasons to suggest that the bleeding associated with ibrutinib may be due to platelet dysfunction. First, the general pattern of bleeding is consistent with primary hemostatic failure, comprising mainly of bruising or subcutaneous bleeding following minimal trauma. Second, this bleeding has occurred despite improvement in platelet counts in most patients treated with ibrutinib, and no defects in classical coagulation parameters (for example, prothrombin time and activated partial thromboplastin time) had been identified in over 350 patients treated across multiple studies.1–3 Finally, BTK and TEC are critical and nonredundant mediators of platelet glycoprotein VI signaling following collagen binding,5,6 and both enzymes are irreversibly inhibited by ibrutinib at clinically relevant concentrations with the IC50 being 0.5 and 78 nM, respectively.7
Previous studies of platelet function following ibrutinib exposure have yielded variable results. Farooqui et al.8 tested samples from ibrutinib-treated patients using the PFA-100 instrument with adrenaline and adenosine diphosphate (ADP) as agonists; no functional abnormalities were identified using this method. Rushworth et al.9 added ibrutinib exogenously to patient samples and evaluated platelet function using the more sensitive method of light transmission aggregometry (LTA), and showed defects in both collagen and ADP responses. Whereas the collagen defect is expected and is due to on-target inhibition of BTK and TEC, the ADP impairment lacks a mechanistic explanation, and may have profound clinical implications given the potency of ADP-inhibiting drugs (for example, clopidogrel) as antithrombotic agents.10 To evaluate the in vivo impact of ibrutinib on platelet aggregation, we performed LTA testing in subjects receiving stable ibrutinib therapy. MATERIALS AND METHODS Human Ethics Research Committee approved platelet studies were performed in 23 subjects receiving ibrutinib either on clinical studies of relapsed CLL2 or through private purchase/compassionate access programs. Fourteen patients were enrolled from Melbourne, Australia and nine were from Toulouse, France. Clinical bleeding events included bruising, hematoma or hemorrhage from any site according to Common Toxicity Criteria (version 4.0). Platelet counts and the use of antiplatelet or anticoagulant drugs were recorded. All subjects had tests performed for coagulation assays including prothrombin time and activated partial thromboplastin time performed at baseline and while on ibrutinib (within reference range in all). Platelet aggregometry tests were performed when the platelet counts had improved sufficiently following ibrutinib therapy to permit reliable testing (generally 4100 × 109/l). Where possible, patients
1 Department of Hematology, St Vincent’s Hospital, Melbourne, Victoria, Australia; 2Département d'Hématologie, IUCT-Oncopole, Toulouse, France; 3Inserm, U1048 and Université Toulouse 3, Toulouse, France; 4Laboratoire d’Hématologie CHU de Toulouse, Toulouse, France; 5Department of Hematology, Peter MacCallum Cancer Centre, Melbourne, Victoria, Australia; 6Faculty of Health, Deakin University, Melbourne, Victoria, Australia; 7Leukemia Department, University of Texas MD Anderson Cancer Center, Houston, TX, USA and 8 University of Melbourne, Parkville, Victoria, Australia. Correspondence: Dr CS Tam, Peter MacCallum Cancer Centre, St Andrew's Place, East Melbourne, Victoria 3002, Australia. E-mail: [email protected] Received 23 June 2014; revised 7 August 2014; accepted 14 August 2014; accepted article preview online 20 August 2014
Ibrutinib and bleeding S Kamel et al
2 without baseline thrombocytopenia were also tested by platelet aggregometry before ibrutinib initiation. Platelet aggregometry by LTA is a relatively complex test requiring considerable experience and technical expertise to perform; thus, LTA was performed using the local protocol already established at the two testing centers.11,12 In Melbourne, 20 ml of citrated blood was collected from each subject using a 21-G needle and centrifuged for 15 min with a relative centrifugal force of 100 at 20 °C. This resulted in a layer of platelet-rich plasma, which was gently aspirated off. The remaining blood was then centrifuged for 10 min with a relative centrifugal force of 4000 leaving platelet-poor plasma, which was used as a measure of maximum possible transmission during LTA. The procedure was similar at Toulouse. Both centers used the key agonists of interest at identical concentrations (collagen 2 μg/ml and 10 μg/ml, and ADP 5 μM); in addition, ADP 2.5 μM, arachidonic acid (AA) 1.5 mM, adrenaline 25 μg/ml, and ristocetin 0.5 mg/ml and 1.2 mg/ml were used in the Melbourne protocol, whereas CRP (collagen-related peptide) 9 μg/ml and the thromboxane A2 analog U46619 5 μM were used in the Toulouse protocol. The source for collagen was equine tendon, purchased from Helena Laboratories (Beaumont, TX, USA) for the Melbourne site and Takeda Laboratories (Linz, Austria) for the Toulouse site. Continuous variables were compared using the Mann–Whitney U-test and categorical variables were compared using the Fisher exact or chi-square tests, as appropriate. The relationship between the binary bleeding outcome and peak aggregation was explored with a logistic regression analysis and a corresponding receiver-operating characteristic curve. All P-values were based on two-sided hypothesis tests.
RESULTS Patient characteristics Table 1 summarizes patient characteristics at the time of LTA testing. The median age was 69 years, and 74% of patients were aged 65 years or older. The majority (96%) received ibrutinib for the treatment of CLL (ibrutinib 420 mg daily), and one patient (4%) was treated for MCL (ibrutinib 560 mg daily). Eight patients (35%) had baseline thrombocytopenia ( o100 × 109/l, lowest value 65 × 109/l). At the time of testing, only two patients had thrombocytopenia (92 × 109/l and 95 × 109/l). Eight (35%) patients were on concomitant antiplatelet or anticoagulant medications: five patients were on aspirin, two patients were on aspirin and
Table 1.
Patient characteristics
Patient characteristics
Value
Tested at Melbourne, Australia 14 (61%) Tested at Toulouse, France 9 (39%) Median age (range), years 69 (48–82) Age ⩾ 65 years 17 (74%) Male 18 (78%) Chronic lymphocytic leukemia 22 (96%) Mantle cell lymphoma 1 (4%) 133 (65–239) Baseline platelet count, median (range), × 109/l 8 (35%) Platelet count at baseline o100 × 109/l Platelet count at aggregometry, median (range), × 109/l 142 (92–237) 9 2 (9%) Platelet count at aggregometry o 100 × 10 /l Median time from ibrutinib initiation to platelet 4 (2–24) aggregometry (range), weeks No antiplatelet or anticoagulant therapy 15 (65%) Aspirin only 5 (22%) Aspirin and clopidogrel 2 (9%) Rivaroxaban 1 (4%) Clinical bleeding Bruising (all grade 1) Surgical bleeding (grade 2) On aspirin at time of bleeding On aspirin+clopidogrel at time of bleeding On rivaroxaban at time of bleeding
Leukemia (2014) 1 – 5
7 (30%) 6 (26%) 1 (4%) 1 0 0
clopidogrel, and one patient was on rivaroxaban. Clinical bleeding while taking ibrutinib occurred in seven patients (30%): six patients with grade 1 bruising (one patient was on concomitant aspirin) and one patient with grade 2 surgical bleeding (wound hematoma following skin cancer excision, performed without cessation of ibrutinib). Platelet aggregometry (LTA) testing The results of LTA for subjects on stable ibrutinib therapy are shown in Figure 1. The results for agonists other than collagen (Figure 1a) were normal, including those for ADP at both high (5 μM) and low (2.5 μM) concentrations. As expected, patients on concomitant therapy with aspirin or clopidogrel had reduced responses to AA and ADP, respectively (data not shown). Reduced adrenaline response (n = 2) was commonly seen in normal donors, and was not considered abnormal.11 Subjects on ibrutinib showed marked impairment of responses to collagen (Figure 1b). At high concentrations of collagen (10 μg/ml), mean maximum aggregation values were 96%, 80% and 38% for controls, subjects before ibrutinib and subjects during ibrutinib therapy, respectively (P o 0.001 comparing pre- and post-ibrutinib). The differences in values between controls and subjects before ibrutinib are consistent with the differences in baseline platelet count (median of 245 × 109/l for controls, compared with a median of 133 × 109/l for study subjects, P = 0.10). In contrast, the differences in collagen response before and after taking ibrutinib occurred despite an improvement in platelet count on therapy (median of 142 × 109/l, P = 0.09 compared with pre-ibrutinib), and thus represent bona fide impairments of functional response. For collagen 2 μg/ml, the mean maximum aggregation values were 92%, 68% and 14% for controls, subjects pre-ibrutinib and subjects during ibrutinib therapy, respectively (Po0.001 comparing pre- and post-ibrutinib). Information on the effect of ibrutinib on platelet response to CRP (a specific agonist of glycoprotein VI, in contrast to collagen that has broader platelet interactions including binding to the integrin α2β1) was available in nine subjects tested at Toulouse. Similar to the results of collagen, subjects on stable ibrutinib displayed a marked reduction in peak aggregation response to CRP 9 μg/ml. Before starting ibrutinib, median peak CRP aggregation was 75% (range: 52–89%). This peak aggregation value decreased to a median of 13% (range: 1–51%, P o 0.001 compared with baseline) while on ibrutinib therapy. Correlation with clinical bleeding No patient in this study had grade 3–4 bleeding. Of the seven patients with less severe degrees of bleeding, only one was on concomitant antiplatelet or anticoagulant therapy (aspirin alone). In patients with bleeding, the events occurred early at a median of 26 days (range: 15–35 days) after starting the drug. The remaining patients were followed up for a median 91 days (range: 42–616 days) without bleeding events. Comparing the LTA response to high-dose collagen (10 μg/ml), mean peak aggregation values were significantly lower in subjects with bleeding (26%) compared to those without (44%, P = 0.039; Figure 1c). The probability of bleeding declined with increasing preservation of peak aggregation to 10 μg/ml collagen (P = 0.044; Supplementary Figure 1a). At 31% peak aggregation (corresponding to the 10th lowest observation indicated in the receiveroperating characteristic curve in Supplementary Figure 1b), the sensitivity and specificity for clinical bleeding were 86% and 75%, respectively. Almost all patients had marked impairment of platelet aggregation when tested at collagen 2 μg/ml; this concentration was therefore not discriminatory for correlation with clinical bleeding in our cohort. © 2014 Macmillan Publishers Limited
Ibrutinib and bleeding S Kamel et al
Platelet aggregation impairment is reversible We performed serial LTA on one patient who interrupted ibrutinib therapy owing to an intercurrent illness (Figure 2). In this patient, the collagen defect was fully reversed when tested 1 week after cessation of ibrutinib therapy and recurred 1 week after restarting therapy. Thus, we confirmed that the ibrutinib effect on collageninduced platelet aggregation was specific and fully reversible.
Percentage Aggregation
100
75
50
25
A
D
U
P
46
H
ig
61
9
h
Lo w
is R
A D P
A
A
dr
A
0
Agonist
Percentage Aggregation
100
75
50
25
Percentage Aggregation
75
50
25
bl
bl
ee
ee
di
di
ng
ng
0
no
dr
C
100
© 2014 Macmillan Publishers Limited
on l2 ol
on C
C
Agonist
Agonist
ug
e l2 C
ol
C tr
ol
on 0 l1 ol
pr
ol l2
ug dr
0 l1 ol C
C
on t
ro
lC
ol l1
pr
0
e
0
DISCUSSION This study reports the use of the sensitive method of LTA to assess platelet function in subjects treated with ibrutinib at approved doses. We showed a specific and reversible effect on collageninduced platelet aggregation, and some evidence (based on small patient numbers) that there may be a correlation between the degree of reduction in collagen response and occurrence of clinical bleeding. In contrast to the ex vivo work reported by Rushworth et al.,9 we found no evidence of impairment of ADP response. Of note, the study performed by Rushworth used a concentration of ibrutinib (0.5 μM) that was ~ 50% higher than the mean peak plasma concentration seen in subjects receiving ibrutinib at 560 mg daily,4 and it is possible that ex vivo manipulation of patient samples, with exogenous addition of drug, may have influenced the reliability of platelet testing for this purpose. What is the clinical significance of our observations? First, these results link the bleeding phenotype seen in ibrutinib-treated subjects (primary hemostatic failure) with a plausible mechanism (platelet dysfunction). This bleeding tendency is expected to be mild when considered in isolation, as excessive bruising or bleeding has not been reported in registry studies of subjects with congenital BTK mutations.5,13 Furthermore, in a second congenital disease—that of platelet glycoprotein VI deficiency— the bleeding phenotype tends to be variable, with the majority having no obvious hemostatic defects.14,15 Thus, reduction or absence in collagen-induced platelet response appears to be largely dispensable in normal life, presumably due to redundancy in other platelet activation pathways. The situation may change, however, with concomitant administration of other antiplatelet and anticoagulant agents and simultaneous inhibition of multiple pathways. Although the initial episodes of intracranial bleeding on ibrutinib therapy occurred in patients on concomitant warfarin (thus leading to the warning about not combining ibrutinib and warfarin), the platelet inhibition properties of ibrutinib as revealed in this paper suggest that concomitant use of multiple antiplatelet drugs (for example, ibrutinib plus aspirin and clopidogrel) may be similarly deleterious. Second, although we did not have the opportunity to examine in detail the kinetics of collagen response recovery following ibrutinib cessation, our study does provide some insight into the decay of ibrutinib effect in patients who require interruption of
Figure 1. LTA in samples taken from patients on stable ibrutinib therapy. The percentages of peak responses for each of the different platelet agonists are shown, with the horizontal bar indicating the mean. (a) Subjects on ibrutinib had normal responses to platelet agonists other than collagen. This figure excludes AA results for subjects on aspirin and ADP results for subjects on clopidogrel. AA = AA 1.5 mM; Adr = adrenaline 25 μg/ ml; Ris = ristocetin 1.2 mg/ml; ADP low = ADP 2.5 μM; ADP high = ADP 5 μM; U46619 = U46619 5 μM. (b) Subjects on stable ibrutinib therapy have reduced responses to collagen at low (2 μg/ml; ‘coll 2’) and high (10 μg/ml, ‘coll 10’) concentrations. For collagen 10 μg/ml, the mean maximum aggregation values were 96%, 80% and 38% for controls, subjects before ibrutinib (pre) and subjects during ibrutinib therapy, respectively (P o 0.001 comparing preand post-ibrutinib). For collagen 2 μg/ml, the mean maximum aggregation values were 92%, 68% and 14% for controls, subjects pre-ibrutinib and subjects during ibrutinib therapy, respectively (P o 0.001 comparing pre- and post-ibrutinib). The differences in maximum aggregation between controls and subjects preibrutinib are likely due to differences in baseline platelet counts. (c) Seven subjects had grade 1–2 bleeding while on ibrutinib therapy, these patients had a significantly lower mean maximal response to collagen 10 μg/ml when compared with subjects without clinical bleeding (26% for subjects with bleeding compared with 44% for subjects without bleeding, P = 0.039). Leukemia (2014) 1 – 5
3
Ibrutinib and bleeding S Kamel et al
4
Figure 2. Serial LTA results in one patient who stopped and restarted ibrutinib therapy. Note the typical collagen defect (green curve, with arrow) seen on ibrutinib therapy (a), which fully resolved after drug interruption for 1 week (b) and recurrence after drug reinitiation (c). Also note the lack of ADP effect with drug stoppage and reinitiation.
drug for invasive procedures. On the basis of the knowledge that ibrutinib irreversibly inhibits BTK but has a short half-life ( o 3 h),4 coupled with the fact that platelets are anucleate and are therefore incapable of synthesizing new BTK, we deduce that the decay of collagen inhibition should be directly proportional to the physiological half-life of platelets (3–4 days).11 We also infer from this knowledge that platelet transfusions outside the 3-h half-life window should reverse the hemostatic defect associated with ibrutinib, in the event of serious bleeding. In conclusion, our data support the safety of the current recommendations regarding interruption of ibrutinib therapy for invasive procedures, and provide the rationale for further mechanistic studies into the antiplatelet effects of ibrutinib and related BTK inhibitors. Clinical studies of ibrutinib and other BTK inhibitors should assess platelet function particularly in patients with clinical bleeding, and in those taking one or more antiplatelet medications.
CONFLICT OF INTEREST Dr Tam received honorarium from Janssen-Cilag and Pharmacyclics. The remaining authors declare no conflict of interest.
Leukemia (2014) 1 – 5
ACKNOWLEDGEMENTS We thank Bernard Payrastre and Cedric Garcia (Inserm, U1048 and Université Toulouse 3) for assistance in platelet testing, and acknowledge the generosity of the subjects who participated in this study. This work was funded by the authors’ institutions.
REFERENCES 1 Byrd JC, Furman RR, Coutre SE, Flinn IW, Burger JA, Blum K et al. Targeting BTK with ibrutinib in relapsed chronic lymphocytic leukemia. N Engl J Med 2013; 369: 32–42. 2 Byrd JC, Brown JR, O'Brien S, Barrientos JC, Kay NE, Reddy NM et al. Ibrutinib versus ofatumumab in previously treated chronic lymphoid leukemia. N Engl J Med 2014; 371: 213–223. 3 Wang ML, Rule S, Martin P, Goy A, Auer R, Kahl BS et al. Targeting BTK with ibrutinib in relapsed or refractory mantle-cell lymphoma. N Engl J Med 2013; 369: 507–516. 4 Advani RH, Buggy JJ, Sharman JP, Smith SM, Boyd TE, Grant B et al. Bruton tyrosine kinase inhibitor ibrutinib (PCI-32765) has significant activity in patients with relapsed/refractory B-cell malignancies. J Clin Oncol 2013; 31: 88–94. 5 Quek LS, Bolen J, Watson SP. A role for Bruton's tyrosine kinase (Btk) in platelet activation by collagen. Curr Biol 1998; 8: 1137–1140. 6 Atkinson BT, Ellmeier W, Watson SP. Tec regulates platelet activation by GPVI in the absence of Btk. Blood 2003; 102: 3592–3599. 7 Honigberg LA, Smith AM, Sirisawad M, Verner E, Loury D, Chang B et al. The Bruton tyrosine kinase inhibitor PCI-32765 blocks B-cell activation and is efficacious in models of autoimmune disease and B-cell malignancy. Proc Natl Acad Sci USA 2010; 107: 13075–13080.
© 2014 Macmillan Publishers Limited
Ibrutinib and bleeding S Kamel et al
5 8 Farooqui M, Lozier JN, Valdez J, Saba N, Wells A, Soto S et al. Ibrutinib (PCI 32765) rapidly improves platelet counts in chronic lymphocytic leukemia/small lymphocytic lymphoma (CLL/SLL) patients and has minimal effects on platelet aggregation. ASH Annual Meeting Abstracts 2012; 120: 1789. 9 Rushworth SA, MacEwan DJ, Bowles KM. Ibrutinib in relapsed chronic lymphocytic leukemia. N Engl J Med. 2013; 369: 1277–1278. 10 Squizzato A, Keller T, Romualdi E, Middeldorp S. Clopidogrel plus aspirin versus aspirin alone for preventing cardiovascular disease. Cochrane Database Syst Rev 2011; (1): CD005158. 11 Lewis BB, Bates I. Dacie and Lewis Practical Haematology. 10th edn. Churchill Livingstone, Elsevier, 2006.
12 Gratacap MP, Martin V, Valera MC, Allart S, Garcia C, Sie P et al. The new tyrosinekinase inhibitor and anticancer drug dasatinib reversibly affects platelet activation in vitro and in vivo. Blood 2009; 114: 1884–1892. 13 Winkelstein JA, Marino MC, Lederman HM, Jones SM, Sullivan K, Burks AW et al. X-linked agammaglobulinemia: report on a United States registry of 201 patients. Medicine (Baltimore) 2006; 85: 193–202. 14 Arthur JF, Dunkley S, Andrews RK. Platelet glycoprotein VI-related clinical defects. Br J Haematol 2007; 139: 363–372. 15 Zahid M, Mangin P, Loyau S, Hechler B, Billiald P, Gachet C et al. The future of glycoprotein VI as an antithrombotic target. J Thromb Haemost 2012; 10: 2418–2427.
Supplementary Information accompanies this paper on the Leukemia website (http://www.nature.com/leu)
© 2014 Macmillan Publishers Limited
Leukemia (2014) 1 – 5
ORIGINAL ARTICLE
Ibrutinib inhibits collagen-mediated but not ADP-mediated platelet aggregation S Kamel1, L Horton1, L Ysebaert2, M Levade3,4, K Burbury5, S Tan1, M Cole-Sinclair1, J Reynolds6, R Filshie1, S Schischka1, A Khot5, S Sandhu5, MJ Keating7, H Nandurkar1,8 and CS Tam1,5,8 The BTK (Bruton’s tyrosine kinase) inhibitor ibrutinib is associated with an increased risk of bleeding. A previous study reported defects in collagen- and adenosine diphosphate (ADP)-dependent platelet responses when ibrutinib was added ex vivo to patient samples. Whereas the collagen defect is expected given the central role of BTK in glycoprotein VI signaling, the ADP defect lacks a mechanistic explanation. In order to determine the real-life consequences of BTK platelet blockade, we performed light transmission aggregometry in 23 patients receiving ibrutinib treatment. All patients had reductions in collagen-mediated platelet aggregation, with a significant association between the degree of inhibition and the occurrence of clinical bleeding or bruising (P = 0.044). This collagen defect was reversible on drug cessation. In contrast to the previous ex vivo report, we found no in vivo ADP defects in subjects receiving standard doses of ibrutinib. These results establish platelet light transmission aggregometry as a method for gauging, at least qualitatively, the severity of platelet impairment in patients receiving ibrutinib treatment. Leukemia advance online publication, 2 September 2014; doi:10.1038/leu.2014.247
INTRODUCTION Ibrutinib is an irreversible inhibitor of Bruton’s tyrosine kinase (BTK) with significant activity in chronic lymphocytic leukemia (CLL), mantle cell lymphoma (MCL) and Waldenstrom’s macroglobulinemia. The drug has recently been approved by the United States Food and Drug Administration for the treatment of CLL and MCL.1–4 Ibrutinib is generally well tolerated but is associated with an increased risk of bleeding. For example, the phase I/II studies of ibrutinib reported increased bruising in 17% of subjects, and intracranial hemorrhages occurred in 2%, leading to current recommendations for avoidance of concomitant warfarin therapy, and for interruption of ibrutinib for 3–7 days before and after invasive procedures.1,3 In the phase III study comparing ibrutinib with ofatumumab in the treatment of CLL, bleeding-related adverse event rates were 44% for ibrutinib vs 12% for the control arm, although the rates of serious bleeding were low in both arms (1% vs 2%, respectively).2 There are compelling reasons to suggest that the bleeding associated with ibrutinib may be due to platelet dysfunction. First, the general pattern of bleeding is consistent with primary hemostatic failure, comprising mainly of bruising or subcutaneous bleeding following minimal trauma. Second, this bleeding has occurred despite improvement in platelet counts in most patients treated with ibrutinib, and no defects in classical coagulation parameters (for example, prothrombin time and activated partial thromboplastin time) had been identified in over 350 patients treated across multiple studies.1–3 Finally, BTK and TEC are critical and nonredundant mediators of platelet glycoprotein VI signaling following collagen binding,5,6 and both enzymes are irreversibly inhibited by ibrutinib at clinically relevant concentrations with the IC50 being 0.5 and 78 nM, respectively.7
Previous studies of platelet function following ibrutinib exposure have yielded variable results. Farooqui et al.8 tested samples from ibrutinib-treated patients using the PFA-100 instrument with adrenaline and adenosine diphosphate (ADP) as agonists; no functional abnormalities were identified using this method. Rushworth et al.9 added ibrutinib exogenously to patient samples and evaluated platelet function using the more sensitive method of light transmission aggregometry (LTA), and showed defects in both collagen and ADP responses. Whereas the collagen defect is expected and is due to on-target inhibition of BTK and TEC, the ADP impairment lacks a mechanistic explanation, and may have profound clinical implications given the potency of ADP-inhibiting drugs (for example, clopidogrel) as antithrombotic agents.10 To evaluate the in vivo impact of ibrutinib on platelet aggregation, we performed LTA testing in subjects receiving stable ibrutinib therapy. MATERIALS AND METHODS Human Ethics Research Committee approved platelet studies were performed in 23 subjects receiving ibrutinib either on clinical studies of relapsed CLL2 or through private purchase/compassionate access programs. Fourteen patients were enrolled from Melbourne, Australia and nine were from Toulouse, France. Clinical bleeding events included bruising, hematoma or hemorrhage from any site according to Common Toxicity Criteria (version 4.0). Platelet counts and the use of antiplatelet or anticoagulant drugs were recorded. All subjects had tests performed for coagulation assays including prothrombin time and activated partial thromboplastin time performed at baseline and while on ibrutinib (within reference range in all). Platelet aggregometry tests were performed when the platelet counts had improved sufficiently following ibrutinib therapy to permit reliable testing (generally 4100 × 109/l). Where possible, patients
1 Department of Hematology, St Vincent’s Hospital, Melbourne, Victoria, Australia; 2Département d'Hématologie, IUCT-Oncopole, Toulouse, France; 3Inserm, U1048 and Université Toulouse 3, Toulouse, France; 4Laboratoire d’Hématologie CHU de Toulouse, Toulouse, France; 5Department of Hematology, Peter MacCallum Cancer Centre, Melbourne, Victoria, Australia; 6Faculty of Health, Deakin University, Melbourne, Victoria, Australia; 7Leukemia Department, University of Texas MD Anderson Cancer Center, Houston, TX, USA and 8 University of Melbourne, Parkville, Victoria, Australia. Correspondence: Dr CS Tam, Peter MacCallum Cancer Centre, St Andrew's Place, East Melbourne, Victoria 3002, Australia. E-mail: [email protected] Received 23 June 2014; revised 7 August 2014; accepted 14 August 2014; accepted article preview online 20 August 2014
Ibrutinib and bleeding S Kamel et al
2 without baseline thrombocytopenia were also tested by platelet aggregometry before ibrutinib initiation. Platelet aggregometry by LTA is a relatively complex test requiring considerable experience and technical expertise to perform; thus, LTA was performed using the local protocol already established at the two testing centers.11,12 In Melbourne, 20 ml of citrated blood was collected from each subject using a 21-G needle and centrifuged for 15 min with a relative centrifugal force of 100 at 20 °C. This resulted in a layer of platelet-rich plasma, which was gently aspirated off. The remaining blood was then centrifuged for 10 min with a relative centrifugal force of 4000 leaving platelet-poor plasma, which was used as a measure of maximum possible transmission during LTA. The procedure was similar at Toulouse. Both centers used the key agonists of interest at identical concentrations (collagen 2 μg/ml and 10 μg/ml, and ADP 5 μM); in addition, ADP 2.5 μM, arachidonic acid (AA) 1.5 mM, adrenaline 25 μg/ml, and ristocetin 0.5 mg/ml and 1.2 mg/ml were used in the Melbourne protocol, whereas CRP (collagen-related peptide) 9 μg/ml and the thromboxane A2 analog U46619 5 μM were used in the Toulouse protocol. The source for collagen was equine tendon, purchased from Helena Laboratories (Beaumont, TX, USA) for the Melbourne site and Takeda Laboratories (Linz, Austria) for the Toulouse site. Continuous variables were compared using the Mann–Whitney U-test and categorical variables were compared using the Fisher exact or chi-square tests, as appropriate. The relationship between the binary bleeding outcome and peak aggregation was explored with a logistic regression analysis and a corresponding receiver-operating characteristic curve. All P-values were based on two-sided hypothesis tests.
RESULTS Patient characteristics Table 1 summarizes patient characteristics at the time of LTA testing. The median age was 69 years, and 74% of patients were aged 65 years or older. The majority (96%) received ibrutinib for the treatment of CLL (ibrutinib 420 mg daily), and one patient (4%) was treated for MCL (ibrutinib 560 mg daily). Eight patients (35%) had baseline thrombocytopenia ( o100 × 109/l, lowest value 65 × 109/l). At the time of testing, only two patients had thrombocytopenia (92 × 109/l and 95 × 109/l). Eight (35%) patients were on concomitant antiplatelet or anticoagulant medications: five patients were on aspirin, two patients were on aspirin and
Table 1.
Patient characteristics
Patient characteristics
Value
Tested at Melbourne, Australia 14 (61%) Tested at Toulouse, France 9 (39%) Median age (range), years 69 (48–82) Age ⩾ 65 years 17 (74%) Male 18 (78%) Chronic lymphocytic leukemia 22 (96%) Mantle cell lymphoma 1 (4%) 133 (65–239) Baseline platelet count, median (range), × 109/l 8 (35%) Platelet count at baseline o100 × 109/l Platelet count at aggregometry, median (range), × 109/l 142 (92–237) 9 2 (9%) Platelet count at aggregometry o 100 × 10 /l Median time from ibrutinib initiation to platelet 4 (2–24) aggregometry (range), weeks No antiplatelet or anticoagulant therapy 15 (65%) Aspirin only 5 (22%) Aspirin and clopidogrel 2 (9%) Rivaroxaban 1 (4%) Clinical bleeding Bruising (all grade 1) Surgical bleeding (grade 2) On aspirin at time of bleeding On aspirin+clopidogrel at time of bleeding On rivaroxaban at time of bleeding
Leukemia (2014) 1 – 5
7 (30%) 6 (26%) 1 (4%) 1 0 0
clopidogrel, and one patient was on rivaroxaban. Clinical bleeding while taking ibrutinib occurred in seven patients (30%): six patients with grade 1 bruising (one patient was on concomitant aspirin) and one patient with grade 2 surgical bleeding (wound hematoma following skin cancer excision, performed without cessation of ibrutinib). Platelet aggregometry (LTA) testing The results of LTA for subjects on stable ibrutinib therapy are shown in Figure 1. The results for agonists other than collagen (Figure 1a) were normal, including those for ADP at both high (5 μM) and low (2.5 μM) concentrations. As expected, patients on concomitant therapy with aspirin or clopidogrel had reduced responses to AA and ADP, respectively (data not shown). Reduced adrenaline response (n = 2) was commonly seen in normal donors, and was not considered abnormal.11 Subjects on ibrutinib showed marked impairment of responses to collagen (Figure 1b). At high concentrations of collagen (10 μg/ml), mean maximum aggregation values were 96%, 80% and 38% for controls, subjects before ibrutinib and subjects during ibrutinib therapy, respectively (P o 0.001 comparing pre- and post-ibrutinib). The differences in values between controls and subjects before ibrutinib are consistent with the differences in baseline platelet count (median of 245 × 109/l for controls, compared with a median of 133 × 109/l for study subjects, P = 0.10). In contrast, the differences in collagen response before and after taking ibrutinib occurred despite an improvement in platelet count on therapy (median of 142 × 109/l, P = 0.09 compared with pre-ibrutinib), and thus represent bona fide impairments of functional response. For collagen 2 μg/ml, the mean maximum aggregation values were 92%, 68% and 14% for controls, subjects pre-ibrutinib and subjects during ibrutinib therapy, respectively (Po0.001 comparing pre- and post-ibrutinib). Information on the effect of ibrutinib on platelet response to CRP (a specific agonist of glycoprotein VI, in contrast to collagen that has broader platelet interactions including binding to the integrin α2β1) was available in nine subjects tested at Toulouse. Similar to the results of collagen, subjects on stable ibrutinib displayed a marked reduction in peak aggregation response to CRP 9 μg/ml. Before starting ibrutinib, median peak CRP aggregation was 75% (range: 52–89%). This peak aggregation value decreased to a median of 13% (range: 1–51%, P o 0.001 compared with baseline) while on ibrutinib therapy. Correlation with clinical bleeding No patient in this study had grade 3–4 bleeding. Of the seven patients with less severe degrees of bleeding, only one was on concomitant antiplatelet or anticoagulant therapy (aspirin alone). In patients with bleeding, the events occurred early at a median of 26 days (range: 15–35 days) after starting the drug. The remaining patients were followed up for a median 91 days (range: 42–616 days) without bleeding events. Comparing the LTA response to high-dose collagen (10 μg/ml), mean peak aggregation values were significantly lower in subjects with bleeding (26%) compared to those without (44%, P = 0.039; Figure 1c). The probability of bleeding declined with increasing preservation of peak aggregation to 10 μg/ml collagen (P = 0.044; Supplementary Figure 1a). At 31% peak aggregation (corresponding to the 10th lowest observation indicated in the receiveroperating characteristic curve in Supplementary Figure 1b), the sensitivity and specificity for clinical bleeding were 86% and 75%, respectively. Almost all patients had marked impairment of platelet aggregation when tested at collagen 2 μg/ml; this concentration was therefore not discriminatory for correlation with clinical bleeding in our cohort. © 2014 Macmillan Publishers Limited
Ibrutinib and bleeding S Kamel et al
Platelet aggregation impairment is reversible We performed serial LTA on one patient who interrupted ibrutinib therapy owing to an intercurrent illness (Figure 2). In this patient, the collagen defect was fully reversed when tested 1 week after cessation of ibrutinib therapy and recurred 1 week after restarting therapy. Thus, we confirmed that the ibrutinib effect on collageninduced platelet aggregation was specific and fully reversible.
Percentage Aggregation
100
75
50
25
A
D
U
P
46
H
ig
61
9
h
Lo w
is R
A D P
A
A
dr
A
0
Agonist
Percentage Aggregation
100
75
50
25
Percentage Aggregation
75
50
25
bl
bl
ee
ee
di
di
ng
ng
0
no
dr
C
100
© 2014 Macmillan Publishers Limited
on l2 ol
on C
C
Agonist
Agonist
ug
e l2 C
ol
C tr
ol
on 0 l1 ol
pr
ol l2
ug dr
0 l1 ol C
C
on t
ro
lC
ol l1
pr
0
e
0
DISCUSSION This study reports the use of the sensitive method of LTA to assess platelet function in subjects treated with ibrutinib at approved doses. We showed a specific and reversible effect on collageninduced platelet aggregation, and some evidence (based on small patient numbers) that there may be a correlation between the degree of reduction in collagen response and occurrence of clinical bleeding. In contrast to the ex vivo work reported by Rushworth et al.,9 we found no evidence of impairment of ADP response. Of note, the study performed by Rushworth used a concentration of ibrutinib (0.5 μM) that was ~ 50% higher than the mean peak plasma concentration seen in subjects receiving ibrutinib at 560 mg daily,4 and it is possible that ex vivo manipulation of patient samples, with exogenous addition of drug, may have influenced the reliability of platelet testing for this purpose. What is the clinical significance of our observations? First, these results link the bleeding phenotype seen in ibrutinib-treated subjects (primary hemostatic failure) with a plausible mechanism (platelet dysfunction). This bleeding tendency is expected to be mild when considered in isolation, as excessive bruising or bleeding has not been reported in registry studies of subjects with congenital BTK mutations.5,13 Furthermore, in a second congenital disease—that of platelet glycoprotein VI deficiency— the bleeding phenotype tends to be variable, with the majority having no obvious hemostatic defects.14,15 Thus, reduction or absence in collagen-induced platelet response appears to be largely dispensable in normal life, presumably due to redundancy in other platelet activation pathways. The situation may change, however, with concomitant administration of other antiplatelet and anticoagulant agents and simultaneous inhibition of multiple pathways. Although the initial episodes of intracranial bleeding on ibrutinib therapy occurred in patients on concomitant warfarin (thus leading to the warning about not combining ibrutinib and warfarin), the platelet inhibition properties of ibrutinib as revealed in this paper suggest that concomitant use of multiple antiplatelet drugs (for example, ibrutinib plus aspirin and clopidogrel) may be similarly deleterious. Second, although we did not have the opportunity to examine in detail the kinetics of collagen response recovery following ibrutinib cessation, our study does provide some insight into the decay of ibrutinib effect in patients who require interruption of
Figure 1. LTA in samples taken from patients on stable ibrutinib therapy. The percentages of peak responses for each of the different platelet agonists are shown, with the horizontal bar indicating the mean. (a) Subjects on ibrutinib had normal responses to platelet agonists other than collagen. This figure excludes AA results for subjects on aspirin and ADP results for subjects on clopidogrel. AA = AA 1.5 mM; Adr = adrenaline 25 μg/ ml; Ris = ristocetin 1.2 mg/ml; ADP low = ADP 2.5 μM; ADP high = ADP 5 μM; U46619 = U46619 5 μM. (b) Subjects on stable ibrutinib therapy have reduced responses to collagen at low (2 μg/ml; ‘coll 2’) and high (10 μg/ml, ‘coll 10’) concentrations. For collagen 10 μg/ml, the mean maximum aggregation values were 96%, 80% and 38% for controls, subjects before ibrutinib (pre) and subjects during ibrutinib therapy, respectively (P o 0.001 comparing preand post-ibrutinib). For collagen 2 μg/ml, the mean maximum aggregation values were 92%, 68% and 14% for controls, subjects pre-ibrutinib and subjects during ibrutinib therapy, respectively (P o 0.001 comparing pre- and post-ibrutinib). The differences in maximum aggregation between controls and subjects preibrutinib are likely due to differences in baseline platelet counts. (c) Seven subjects had grade 1–2 bleeding while on ibrutinib therapy, these patients had a significantly lower mean maximal response to collagen 10 μg/ml when compared with subjects without clinical bleeding (26% for subjects with bleeding compared with 44% for subjects without bleeding, P = 0.039). Leukemia (2014) 1 – 5
3
Ibrutinib and bleeding S Kamel et al
4
Figure 2. Serial LTA results in one patient who stopped and restarted ibrutinib therapy. Note the typical collagen defect (green curve, with arrow) seen on ibrutinib therapy (a), which fully resolved after drug interruption for 1 week (b) and recurrence after drug reinitiation (c). Also note the lack of ADP effect with drug stoppage and reinitiation.
drug for invasive procedures. On the basis of the knowledge that ibrutinib irreversibly inhibits BTK but has a short half-life ( o 3 h),4 coupled with the fact that platelets are anucleate and are therefore incapable of synthesizing new BTK, we deduce that the decay of collagen inhibition should be directly proportional to the physiological half-life of platelets (3–4 days).11 We also infer from this knowledge that platelet transfusions outside the 3-h half-life window should reverse the hemostatic defect associated with ibrutinib, in the event of serious bleeding. In conclusion, our data support the safety of the current recommendations regarding interruption of ibrutinib therapy for invasive procedures, and provide the rationale for further mechanistic studies into the antiplatelet effects of ibrutinib and related BTK inhibitors. Clinical studies of ibrutinib and other BTK inhibitors should assess platelet function particularly in patients with clinical bleeding, and in those taking one or more antiplatelet medications.
CONFLICT OF INTEREST Dr Tam received honorarium from Janssen-Cilag and Pharmacyclics. The remaining authors declare no conflict of interest.
Leukemia (2014) 1 – 5
ACKNOWLEDGEMENTS We thank Bernard Payrastre and Cedric Garcia (Inserm, U1048 and Université Toulouse 3) for assistance in platelet testing, and acknowledge the generosity of the subjects who participated in this study. This work was funded by the authors’ institutions.
REFERENCES 1 Byrd JC, Furman RR, Coutre SE, Flinn IW, Burger JA, Blum K et al. Targeting BTK with ibrutinib in relapsed chronic lymphocytic leukemia. N Engl J Med 2013; 369: 32–42. 2 Byrd JC, Brown JR, O'Brien S, Barrientos JC, Kay NE, Reddy NM et al. Ibrutinib versus ofatumumab in previously treated chronic lymphoid leukemia. N Engl J Med 2014; 371: 213–223. 3 Wang ML, Rule S, Martin P, Goy A, Auer R, Kahl BS et al. Targeting BTK with ibrutinib in relapsed or refractory mantle-cell lymphoma. N Engl J Med 2013; 369: 507–516. 4 Advani RH, Buggy JJ, Sharman JP, Smith SM, Boyd TE, Grant B et al. Bruton tyrosine kinase inhibitor ibrutinib (PCI-32765) has significant activity in patients with relapsed/refractory B-cell malignancies. J Clin Oncol 2013; 31: 88–94. 5 Quek LS, Bolen J, Watson SP. A role for Bruton's tyrosine kinase (Btk) in platelet activation by collagen. Curr Biol 1998; 8: 1137–1140. 6 Atkinson BT, Ellmeier W, Watson SP. Tec regulates platelet activation by GPVI in the absence of Btk. Blood 2003; 102: 3592–3599. 7 Honigberg LA, Smith AM, Sirisawad M, Verner E, Loury D, Chang B et al. The Bruton tyrosine kinase inhibitor PCI-32765 blocks B-cell activation and is efficacious in models of autoimmune disease and B-cell malignancy. Proc Natl Acad Sci USA 2010; 107: 13075–13080.
© 2014 Macmillan Publishers Limited
Ibrutinib and bleeding S Kamel et al
5 8 Farooqui M, Lozier JN, Valdez J, Saba N, Wells A, Soto S et al. Ibrutinib (PCI 32765) rapidly improves platelet counts in chronic lymphocytic leukemia/small lymphocytic lymphoma (CLL/SLL) patients and has minimal effects on platelet aggregation. ASH Annual Meeting Abstracts 2012; 120: 1789. 9 Rushworth SA, MacEwan DJ, Bowles KM. Ibrutinib in relapsed chronic lymphocytic leukemia. N Engl J Med. 2013; 369: 1277–1278. 10 Squizzato A, Keller T, Romualdi E, Middeldorp S. Clopidogrel plus aspirin versus aspirin alone for preventing cardiovascular disease. Cochrane Database Syst Rev 2011; (1): CD005158. 11 Lewis BB, Bates I. Dacie and Lewis Practical Haematology. 10th edn. Churchill Livingstone, Elsevier, 2006.
12 Gratacap MP, Martin V, Valera MC, Allart S, Garcia C, Sie P et al. The new tyrosinekinase inhibitor and anticancer drug dasatinib reversibly affects platelet activation in vitro and in vivo. Blood 2009; 114: 1884–1892. 13 Winkelstein JA, Marino MC, Lederman HM, Jones SM, Sullivan K, Burks AW et al. X-linked agammaglobulinemia: report on a United States registry of 201 patients. Medicine (Baltimore) 2006; 85: 193–202. 14 Arthur JF, Dunkley S, Andrews RK. Platelet glycoprotein VI-related clinical defects. Br J Haematol 2007; 139: 363–372. 15 Zahid M, Mangin P, Loyau S, Hechler B, Billiald P, Gachet C et al. The future of glycoprotein VI as an antithrombotic target. J Thromb Haemost 2012; 10: 2418–2427.
Supplementary Information accompanies this paper on the Leukemia website (http://www.nature.com/leu)
© 2014 Macmillan Publishers Limited
Leukemia (2014) 1 – 5
