Blood cell JAKtivation aggravates cerebral venous thrombosis

In this issue of Blood Advances, Bourrienne et al present the first preclinical data on how the gain-of-function JAK2 V617F mutation that underlies most myeloproliferative neoplasms (MPNs) may modify outcomes in cerebral venous sinus thrombosis (CVST).¹ Their study was spurred by observations that individuals harboring a hematopoietic JAK2 V617F mutation (irrespective of whether or not an overt MPN was present) have increased intracerebral hemorrhage (ICH)² and worse outcomes with CVST.³ 

The authors found that MPN mice harboring Jak2 V617F in hematopoietic cells indeed had decreased survival 1 day after experimental CVST compared with wild-type controls (76% vs 100%). ICH occurred twice as often and was more severe in MPN mice, likely explaining the increased mortality. Notably, increased ICH in MPN mice coincided with increased microvascular venular thrombosis, consistent with the reputed mechanism of increased thrombus burden leading to increased venous pressure, resulting in blood-brain-barrier disruption and ICH.

JAK2 V617F is a somatically acquired mutation in hematopoietic cells resulting in constitutive activation of the JAK-STAT signaling pathway. Cytokine-independent myeloid cell proliferation occurs, resulting in the characteristic phenotype of MPN, with increased numbers of mature granulocytes, erythrocytes, and platelets. Although quantitative increases of erythrocytes in MPN can alter blood rheology and increase thrombotic risk,⁴ it is also known that JAK2 V617F-mutated clonal hematopoiesis (with normal peripheral blood cell counts) increases the risk of venous thrombosis threefold,⁵ myocardial infarction 12-fold,⁶ and ischemic stroke 1.4-fold⁷ over the general population. This cell count–independent hypercoagulability may be because of the direct effects of JAK2 V617F on mutated myeloid cells, the indirect effects of JAK2 V617F on nonmutated cells, or a combination of both.

In this study, although MPN mice had a polycythemia vera phenotype with elevated hemoglobin and hematocrit, the severity of ICH after CVST did not correspond to peripheral blood hemoglobin levels, implicating the qualitative cellular changes elicited by JAK2 activation. Indeed, the authors found basal activation of both neutrophils (increased integrin CD11b) and platelets (increased platelet factor 4, decreased platelet P-selectin, and increased soluble P-selectin), in addition to increased platelet-neutrophil aggregates in MPN mice compared with wild type, consistent with other reports.⁸ Neutrophil and platelet activation was further increased after CVST. JAK2 V617F-mediated neutrophil integrin expression⁸ allows for enhanced neutrophil attachment to endothelium, which was indeed visualized by Bourrienne et al soon after CVST. Growing evidence implicates neutrophils in hypercoagulability, largely via neutrophil extrusion of extracellular traps.⁹ In this study, although netosis was detected by elevated plasma levels of citrullinated histone 3, there was no direct evidence of netosis in the brains of MPN mice after CVST. An alternative explanation for the enhanced venular thrombosis in MPN mice could be the increased expression of neutrophil tissue factor (TF). TF is the primary initiator of the extrinsic coagulation pathway and is normally expressed by subendothelial perivascular cells in which it serves as a hemostatic envelope to rapidly plug vessel wall damage to prevent hemorrhage. TF can also be aberrantly expressed by circulating hematopoietic cells, including neutrophils, under pathologic conditions such as antiphospholipid syndrome¹⁰ and MPN.^11,12 Thus, it is plausible that JAK2 V617F-mutated neutrophils more frequently interact with the endothelium owing to increased integrin expression and support thrombosis by surface TF exposure. In line with this, TF was found to significantly contribute to venular thrombosis in a different murine model of CVST.¹³ 

Platelets also carry the JAK2 V617F mutation in MPN, yet their role in MPN-associated hypercoagulability is controversial. MPN platelets appear hyperactive in vivo, yet ex vivo show impaired function.^14,15 Elevated platelet counts have not consistently been associated with increased thrombosis risk but are instead linked to increased bleeding risk. The Jak2 V617F mouse model used by Bourrienne et al has well-characterized platelet dysfunction in addition to reduced levels of high-molecular-weight von Willebrand factor multimers.¹⁶ In spite of this, no surgical bleeding was observed in the MPN mice during the craniotomy or CVST procedure in the present study. Platelet activation in CVST recently came into the spotlight during the COVID-19 pandemic after a small number of individuals developed vaccine-induced thrombotic thrombocytopenia with associated CVST. Platelet factor 4 (PF4) activating antibodies were identified in these individuals suggesting a causal role. Subsequent mechanistic studies by Buka et al¹⁷ revealed that PF4 activated the thrombopoietin receptor on platelets resulting in activation of JAK2. JAK2 activation is also a key feature in COVID-19 infection where both CVST incidence and severity are increased.¹⁸ Surprisingly, little is known about inciting mechanisms of CVST, but recent evidence implicates activated platelet receptor signaling in abnormal adhesion to cerebral venous sinuses.¹⁹ It is therefore possible that the constitutive activation of platelet JAK2 in MPN may be linked not only to increased mortality but also to enhanced initiation of CVST.

Increasing variant allele fraction (VAF) of JAK2 V617F correlates with thrombosis risk in MPN.²⁰ The MPN mouse model used by Bourrienne et al had a Jak2 V617F VAF of 50%, which is equivalent to every hematopoietic cell having 1 mutant copy of Jak2. Curiously, however, CVST has been reported in individuals with low JAK2 V617F VAF of <10%, wherein <20% of granulocytes are mutated and peripheral blood cell counts are often normal. Skewing toward a higher mutation burden in platelets has been reported,²¹ so it is plausible that those with low JAK2 V617F VAF (which is typically measured in granulocytes) have a higher burden of clonal platelets.

CVST is more common in females than males, and the use of estrogen-containing oral contraceptives and puerperium are risk factors for CVST development.^1,18 In contrast, CVST outcomes are worse when they occur in men and when they occur in women not taking oral contraceptives or who are not postpartum.¹⁸ Thus, the risk factors for developing CVST and for developing severe CVST may be different. Notably, JAK2 V617F clones are more likely to be homozygous in men and heterozygous in women, rationally resulting in higher hematopoietic cell JAK2 activation in men as compared to women.²² Bourrienne et al only used female mice for the present study, though it would be of interest to also evaluate male mice in the same model to further explore the effect of sex on severity of MPN-associated CVST.

In summary, the present study contributes to the growing body of evidence that “JAKtivation” of myeloid cells contributes to hypercoagulability in MPN. Further studies are necessary to parse whether Jak2-mutated neutrophils, platelets, or both cell types are contributing to the enhanced thrombus propagation and poor CVST outcomes in this murine model.

Conflict-of-interest disclosure: B.R. has received consulting fees from PharmaEssentia, Incyte, and MorphoSys.