Functional Regulation of TFPI in Membrane Lipid Rafts.

Abstract The assembly of tissue factor-factor VIIa (TF-FVIIa) complex results in the proteolytic activation of factors X (FX) and IX and ultimate thrombin generation. The inhibition of this pathway by the Kunitz-type inhibitor, tissue factor pathway inhibitor (TFPI), involves the formation of a stable TF-FVIIa-FXa-TFPI complex. TFPI in endothelial cells (EC) locates primarily in rafts and caveolae, which are membrane microdomains enriched in cholesterol, glycosphingolipids (GSL) and caveolins, and which regulate the function of TFPI. Since caveolin-1 supports the TFPI-dependent inhibition of TF-FVIIa, we aimed to decipher the role played by the individual components of rafts in the anticoagulant function of cell surface TFPI. To this end, we studied the distribution of TFPI, TF and caveolin-1 by immunofluorescence microscopy, and we assayed the functional activity of TFPI after cholesterol-complexation on EC (EA. hy926 and HUVEC) and HEK293 expressing TFPI or TFPI+caveolin-1, or we used GSL-deficient CHO mutant cell lines. In EC, cholesterol complexation with filipin led to patching of TFPI over the cell surface and reduced inhibition of TF-FVIIa. Extraction of cholesterol from the external leaflet of the membrane with methyl-β-cyclodextrin (M-β-CD) shifted the partition of TFPI from predominantly raft-associated to the non-raft cellular fractions isolated through temperature-induced phase separation of Triton X-114 lysates. Although activation of FX by TF-FVIIa was significantly enhanced by M-β-CD and reversed after cholesterol replenishment, the effect was only modestly affected by the TFPI activity reduction. By immunofluorescence we observed that M-β-CD produced redistribution of both TFPI and TF over the EC and 293 cell surface with apparent segregation into separate domains and complete lack of co-localization. Such accumulations of TF will likely promote strong procoagulant activity when not inhibited by TFPI. Since M-β-CD selectively disrupts the glycerophospholipid-rich regions of the membrane while leaving the caveolar cholesterol virtually intact, we also tested progesterone, which extracts cholesterol specifically from caveolae. Treatment of HEK293 cells with progesterone for 2 hrs reduced significantly the inhibition of TF-FVIIa-dependent activation of FX by TFPI for TFPI+Cav+ cells but not for TFPI+ cells, suggesting that the process was specific for cells that have caveolae. To study the role of GSL for the activity of TFPI, we used Ly-B cells, a GSL-deficient mutant derived from CHO-K1, which have a defect in the LCB1 subunit of serine palmitoyltransferase. Characterization of endogenous TFPI in CHO-K1, Ly-B and its genetically corrected revertant Ly-B/cLCB1 (cLCB) revealed strong similarities between CHOs and EC with regard to the expression and function of TFPI. Whereas not affecting cLCB cells, incubation of Ly-B for 2 days in sphingolipid-deficient medium shifted the partition of cellular TFPI from the detergent-soluble (rafts) to the water-soluble (non-raft) fraction, which suggests that GSL play a major role in the distribution and function of the membrane TFPI. The fundamental knowledge developed by these studies will improve our understanding of the mechanisms by which TFPI functions against TF-FVIIa procoagulant activity on cell surfaces. In the long term, they may guide novel therapeutic approaches to prevent inflammation and thrombosis.