New plugs for CCM bleeds

In this issue of Blood, Lopez-Ramirez et al link hemorrhage associated with cerebral cavernous malformations (CCMs) to anticoagulant properties of lesion endothelial cells, suggesting novel ways to attenuate accompanying seizures and stroke.¹ 

CCMs are central nervous system vascular lesions that affect ∼1 in 200 humans. They are characterized by abnormal blood flow and hemorrhage that can lead to neurologic impairment, seizures, and stroke. Most CCM research has focused on the underlying genetic mutations and the cellular mechanisms leading to endothelial cell dysfunction, which results in the dilated and tortuous vessels that characterize the lesions.²  It has been assumed that hemorrhage results from defective endothelial cell junctions (ie, a leaky bag) and/or rupture as a result of fragile vessels and disrupted blood flow. An intriguing new study by Lopez-Ramirez et al shifts focus to the coagulant properties of endothelial cells and shows that CCM lesions have increased anticoagulant proteins that contribute to the bleeding phenotype and may provide novel therapeutic targets.

CCM disease results from mutations in 1 of 3 genes (CCM1/Krit1, CCM2, CCM3/PDCD10) that are inherited as a single mutant allele in the familial form, with a second hit leading to somatic mutation of the remaining normal allele. The more common sporadic form of the disease also results from somatic mutation of CCM genes.³ CCM mutations affect endothelial cell junctions, and evidence supports some form of transition from stable endothelium to a more active and migratory state called endothelial-to-mesenchymal transition (EndoMT) in CCM lesions.⁴  A previous genome-wide transcriptomic analysis by the group⁵  revealed that 2 components of a pathway that prevents blood clotting were increased in endothelial cells from mice carrying a CCM1 mutation, Thbd, and endothelial protein C receptor (EPCR). Thbd encodes thrombomodulin (TM), a protein that binds to thrombin, prevents conversion of fibrinogen into insoluble fibrin, and enhances production of activated protein C (APC); EPCR also enhances generation of APC. APC is a potent natural anticoagulant, which led the authors to hypothesize that CCM-induced increase in endothelial cell anticoagulant status contributes to the hemorrhage associated with morbidity and mortality.

The authors first examined patient samples and found increased plasma levels of soluble TM in patients with CCM and increased TM and EPCR RNA and protein in human CCM lesions. They then analyzed 2 mouse models that recapitulate aspects of CCM1 (Krit1^iECKO [inducible endothelial cell-knockout]) or CCM3 (Pdcd10^iECKO) and found changes in the CCM lesions similar to those seen in humans. Moreover, the degree of hemorrhage correlated with TM levels in murine lesions, which suggested a causal relationship in vivo. Because the transcription factors KLF2 and KLF4 are known to regulate endothelial cell anticoagulant gene expression^6,7  and are upregulated in CCM,^8,9  the authors next asked whether these factors are upstream of anticoagulant expression using human umbilical vein and brain endothelial cells. They found that overexpression of both KLF factors upregulated TM but not EPCR, whereas reduced KLF levels prevented increased TM expression induced by reducing KRIT1 protein levels. To bring the study back into the mouse, the authors showed that reducing endothelial cell TM levels via gene inactivation in a CCM background (Pdcd10^iECKO; Thbd^iECKO) partially rescued the bleeding phenotype of the mice, and neutralizing antibodies against either TM or EPCR had a similar effect.

These exciting results suggest that, although the etiology of CCM resides in protein complexes that affect the stability of endothelial cell junctions, the effects of the mutations on other endothelial cell functions such as coagulation are also important and may be worth considering as novel therapeutic avenues. In this light, it is especially intriguing that systemic administration of blocking antibodies for TM or EPCR to mice partially rescued the bleeding phenotype that is thought to be causative of neurologic symptoms and morbidity in human patients. This work leads to several new questions that are important for a better understanding of the disease and for considering therapeutic interventions. Identification of upstream EPCR regulators will help us better understand how anticoagulant activity is regulated in CCM lesions. Although APC was increased in cultured endothelial cells with reduced KRIT1 protein levels, this study did not investigate the effect of protein C neutralization on CCM in vivo.Because APC is downstream of increased TM and EPCR in CCM, it may be a more effective therapeutic target. It will be important to determine whether systemic blockade of the anticoagulant pathway leads to abnormal coagulation in other parts of the body, which could complicate the use of anticoagulant blockers in managing CCM. The results from this study suggest that soluble TM may be a useful biomarker for CCM in human patients, especially given that imaging is the current modality for screening. Finally, taking a page from cancer therapeutics, a combination of both anticoagulant blockers and agents that restore endothelial junction integrity may be more potent at preventing hemorrhage than single-modality therapeutics. For example, the RhoA pathway is activated downstream of CCM mutations to disrupt endothelial cell junction integrity,¹⁰  so combinatorial treatment with RhoA pathway inhibitors and anticoagulant blockers may enhance treatment efficacy. Regardless, this new study suggests potential novel avenues for managing CCM, because hemorrhage is the proximal event leading to the most devastating sequelae of CCM disease.