The Ectopyrase CD73 Is a Critical Checkpoint in Thromboinflammation

Introduction: Purinergic signaling through extracellular nucleotides such as adenosine is at the intersection of inflammation and coagulation pathways. CD73 is the key vascular enzyme involved generating AMP-derived extracellular adenosine to reduce inflammation, but its impact on thromboinflammation remains unexplored.

Hypothesis: We hypothesized that CD73 suppresses innate immune activation and protects against thromboinflammation.

Methods: To test this hypothesis, we recruited ten patients with ultrarare, loss of function mutations in the gene encoding CD73 which can result in an autosomal recessive, arterial calcification disease. Medical history and blood samples were obtained from patients for experiments. Preclinical experiments to study thromboinflammation were performed in mice with genetic and post-natal CD73 inhibition with appropriate controls.

Results: We first confirmed loss of functional CD73 enzymatic activity in these patients using a novel, live cell phosphohydrolysis assay. Four of the 10 patients had a history of thrombosis (2 arterial, 2 venous). Neutrophils from patients lacking CD73 formed more spontaneous neutrophil extracellular traps (NETs) than healthy controls measured using multiple assays, which was recapitulated by bone marrow-derived neutrophils from mice genetically deficient for CD73 (CD73KO). Using single-cell RNA-seq of human neutrophils, we further characterized circulating neutrophil phenotypes in affected patients and healthy controls. We observed in these patients three distinct, overexpressed neutrophil subsets corresponding to mature and activated; immature; and IFN-responsive phenotypes. The differential gene expression of neutrophils from CD73-deficient patients was largely driven by genes involved in IFN response and granule genes. In contrast, CD14+ monocytes purified from patients lacking CD73 had similar cellular tissue factor expression as healthy controls, and similar tissue factor activity on extracellular vesicles released following monocyte stimulation with LPS.

To mechanistically explore the role of CD73 in venous thrombosis, we used two murine models of deep vein thrombosis (electrolytic injury and stenosis). Genetic CD73 deletion and inhibition of CD73 in WT mice using a blocking mAb resulted in increased murine venous thrombogenesis, which was abrogated by exogenous apyrase administration. The increased thrombosis in CD73KO mice was attributed, in part to excess NETs formation characterized by increases in plasma cell-free DNA early after thrombosis induction, citrullinated histone H3 in the thrombosis, and attenuation of in vivo thrombosis by administration of DNase. Immunohistochemical analysis revealed increased myeloperoxidase in thrombi from CD73-deficient mice. Mechanistically, the severe thromboinflammation phenotype induced by genetic or post-natal CD73 inhibition in mice could be rescued by stimulation of the GPCR adenosine 2A receptor with CGS21680, suggesting CD73’s enzymatic activity to generate adenosine and trigger downstream cAMP signaling as a key mechanism in mitigating venous thrombosis.

Conclusions: These translational studies identify a previously unknown thromboinflammatory phenotype in patients with a rare, monogenic disease and establishes CD73 as a critical checkpoint in thromboinflammation using mechanistic investigations in human studies and animal models. CD73 inhibitors are also being tested in cancer clinical trials as an immune checkpoint inhibitor. Our ongoing studies will test whether a clinically used CD73 inhibitor is capable of triggering a thromboinflammatory phenotype in mice. These studies are especially relevant in the context of the large number of patients with prothrombotic cancers who may potentially receive a CD73 inhibitor for cancer treatment.