The Clotting Quest is Over: Vascular Endothelium is the Source of Both Factor VIII and von Willebrand Factor

As a medical student in 1974, I read in the 7th Edition of Wintrobe’s Clinical Hematology that the sites of production of factor VIII (FVIII) had yet to be defined.¹  Roles for the liver, spleen, and reticuloendothelial system had been proposed. Transplantation of a normal liver into a hemophilic dog resulted in a 50 percent increase in FVIII, while transplanting a hemophiliac liver into a normal dog decreased the FVIII by 50 percent.^2,3  These experiments suggested that extrahepatic cells also produce FVIII. In the 1980s, immunohistochemical assays showed FVIII antigen in guinea pig liver and spleen. FVIII antigen could be found in normal human liver sinusoidal endothelial cells, as well as in extracts of human lymph nodes, lung, liver, and spleen.⁴  The localization of antigen in tissues did not, however, distinguish sites of FVIII synthesis from those of storage, and such experiments were subject to misinterpretation due to entrapment of plasma FVIII in tissues. FVIII mRNA was detected in isolated human hepatocyte-rich extracts, as well as in the spleen, the lymph nodes, and the kidney, but not in white blood cells or cultured endothelial cells. More recently, cell sorting studies of microvascular endothelial cells have documented FVIII production.⁵  Von Willebrand Factor, which circulates with FVIII, is known to be synthesized in endothelial cells and megakaryocytes, but definitive proof of the endothelial origin of FVIII in vivo was lacking. Now, however, the results of two elegant studies published in Blood, one from David Ginsburg’s laboratory at the University of Michigan and the other from Robert Montgomery’s laboratory at the Blood Center of Wisconsin, demonstrate that endothelial cells are indeed the primary source of plasma FVIII.

The Michigan group took advantage of the special function of LMAN1, a cargo receptor that is responsible for efficient secretion of factors V and VIII into plasma. Lman1 mutations in mice reduced plasma concentrations of both factors to 10 to 15 percent of normal. Using Lman1 conditional murine knockouts, the investigators demonstrated that endothelial cells were the primary site of FVIII biosynthesis, while hepatocytes made no significant contributions to the FVIII plasma pool. Confirmatory studies using endothelial-specific mRNA isolation and PCR showed high level expression of FVIII by endothelial cells from multiple tissues.

The Wisconsin group utilized a Cre/lox-dependent conditional knockout mouse model in which exons 17 and 18 of FVIII are flanked by loxP sites, and when Cre-recombinase is expressed, the FVIII gene is inactivated. By breeding these mice with various tissue-specific Cre strains, hepatocyte-specific FVIII-knockout animals had similar FVIII levels as control mice, while endothelial-knockout mice had a severe hemophiliac phenotype with no detectable FVIII. And bone marrow transplant studies showed that hematopoietic cells do not contribute to FVIII production.