The effect of cultured endothelial cells on factor VIII procoagulant activity.

Cultured human umbilical vein endothelial cells produce a protein that has von Willebrand factor activity and forms immunoprecipitates with rabbit antibody to purified plasma factor VIII/von Willebrand factor (FVIII/vWF) protein, but it has no FVIII procoagulant activity. Of the three characteristics of plasma FVIII/vWF protein, only FVIII procoagulant activity is readily destroyed by trace proteases. A previous report from this laboratory demonstrated protease activity in culture medium under conditions that had been used by others to show that endothelial cells do not synthesize protein with FVIII procoagulant activity. However, even if cultured endothelial cells are placed in protease-free culture medium, no FVIII procoagulant activity can be detected, despite an increase in the level of protein with vWF activity from 0 to 0.57 microgram/ml by 48 hr. This observation and the lack of protease activity in medium left in contact with the cells for 48 hr led to the hypothesis that proteases exist on the surface of cultured umbilical vein endothelial cells. Protease activity was quantitated by the hydrolysis of p-nitroaniline from the substrate, N-benzoyl-phenylalanyl-valyl-arginyl-p-nitroanilide and by degradation of the procoagulant activity of added purified plasma FVIII/vWF protein. In the absence of endothelial cells, no protease activity was present in protease-free culture medium whether or not it had previously overlaid cultured cells. This medium did not cause cleavage of p-nitroaniline from the tripeptide substrate, and 83% of added FVIII procoagulant activity remained after 48 hr. When the synthetic tripeptide was incubated in contact with cultured endothelial cells, 7.3 +/- 0.8 X 10(-10) moles of p-nitroaniline/hr was released; moreover, only 47% of the added FVIII procoagulant activity remained after 48 hr. Given this rate of destruction, it can be calculated that sufficient protease activity exists on the surface of cultured endothelial cells to degrade the procoagulant activity of approximately 1.6 microgram FVIII/vWF protein/hr. This degradation rate is 45 times the rate of release of FVIII/vWF protein from cultured endothelial cells when assessed by the generation of protein with vWF activity. Hence, the detection of FVIII procoagulant activity, if in fact synthesized by cultured endothelial cells, will be most difficult.