S153: mechanisms of trombogenesis in patients with chronic myeloid leukemia under treatment with ponatinib and other tyrosine kinase inhibitors

Background: Second- and third-generation tyrosine kinase inhibitors (TKIs) allow deeper and faster responses in CML patients but may increase the risk of thrombotic events. Ponatinib, useful in resistant cases, is associated with a high thrombotic risk. The underlying mechanisms remain unclear. Although it has been described in a mouse model that ponatinib may induce von Willebrand Factor (VWF) secretion mediating platelet adhesion and development of thrombotic microangiopathy, information is lacking in patients. Aims: To elucidate the thrombogenic mechanisms of ponatinib that promote a thromboinflammatory state to identify new therapeutic targets for thrombosis prevention. Methods: Peripheral blood was collected from chronic phase CML patients treated with imatinib 300-400 mg once daily (QD); nilotinib 400 mg QD/300 mg twice daily (BID); ponatinib 30-45 mg QD; or in TKI-free remission, recruited from GELMC hospitals (n≥20/group), as well as from sex- and age-matched healthy donors. Hemostasis and thromboinflammatory markers were then evaluated. Results: We found no differences in biological or molecular parameters between groups. As expected, ponatinib-treated patients were exposed to more lines of treatment, i.e. shorter ponatinib exposure time. Ponatinib-treated CML patients showed higher VWF:Ag plasma levels compared to controls and other TKIs (p<0.01 and p<0.05 respectively); higher VWF:CB levels (p<0.05) and a decrease in ADAMTS13 activity (p<0.05). For NETs markers, imatinib and ponatinib therapy increased cfDNA plasma levels (p<0.001 and p<0.01 respectively), but only ponatinib increased citH3-DNA, a specific NET marker (p<0.05). PFA-100 assay showed that ponatinib-treated patients had prolonged Col/Epi cartridge closure time (>300s, p<0.001), whereas imatinib and nilotinib did not change it. Compared to controls and other TKIs, platelets from ponatinib patients showed reduced (50%) fibrinogen binding and P-selectin exposure after TRAP6 and CRP stimulation, as well as reduced p-Syk, suggesting a defective GPVI activation pathway. However, plasma PF4 levels were similar in all CML patients, suggesting that platelet hyporeactivity under ponatinib treatment is not caused by in vivo platelet activation in the circulation. Compared to controls, the Luminex assay showed higher plasma sE-selectin levels in imatinib- and ponatinib-treated patients (p<0.01 and p<0.05 respectively) and a similar trend for thrombomodulin levels. Importantly, ponatinib-treated samples had twice the levels of tissue factor (TF), a potent procoagulant factor (p<0.001), while the other TKIs had no effect on TF levels. To investigate whether these higher TF levels were ponatinib dose-dependent, a subgroup of patients treated with ponatinib 15 mg QD after failure to discontinue (PonaZero study) was investigated (n=16) and no changes in TF levels were found, supporting that plasma TF increase occurs only at ponatinib doses associated with higher thrombotic risk. Summary/Conclusion: We describe for the first time that VWF, NETs and TF play a role in the pathogenesis of ponatinib-associated thrombosis and could be useful thromboinflammatory biomarkers to identify thrombotic risk. Indeed, plasma TF levels were remarkably elevated only in patients treated with high doses of ponatinib (30-45 mg QD) compared to low doses (15 mg QD), other TKIs or controls. As platelets from ponatinib-treated patients are hyporeactive due to a defective GPVI pathway, antiplatelet therapy may not help prevent thrombotic events in these patients. However, data from this study identify novel biomarkers and targets for thromboprophylaxis in this setting. Keywords: Thrombosis, Tissue factor, Chronic myeloid leukemia, Tyrosine kinase inhibitor