Expression of tissue factor pathway inhibitor by cultured endothelial cells in response to inflammatory mediators

We recently proposed that endothelium may represent the primary physiologic site of synthesis of the tissue factor pathway inhibitor (TFPI). In support of this conclusion, we have now found that the poly(A)⁺ RNAs obtained from rabbit and bovine lung tissues contain abundant amounts of TFPI messenger RNAs (mRNAs), whereas the poly(A)⁺ RNAs obtained from the liver of these animals contain less than 5% of that found in the lung tissues. Because inflammatory mediators are known to upregulate tissue factor (TF) expression by the endothelium, we have examined the effect of these agents on the TFPI expression by the cultured endothelial cells. When cultured human umbilical vein endothelial cells were stimulated (in 10% fetal bovine serum) with phorbol myristate acetate (PMA), endotoxin, interleukin-1, or tumor necrosis factor-α, the TF mRNA increased ≈7- to 10-fold within 2 to 4 hours. Unstimulated cells constitutively expressed TFPI mRNA and its levels either did not change or increased slightly (up to 1.5-fold) upon stimulation with these inflammatory agents. TF mRNA abruptly declined to a negligible level and the TFPI mRNA returned essentially to the basal level at ≈24 hours. The membrane-bound TF clotting activity of induced cells peaked between 4 and 8 hours, and finally declined. The cumulative TFPI activity secreted into the media was either unchanged or slightly higher in the induced cell cultures as compared with that present in the noninduced cultures. Endothelial cells were also cultured in 10% heat-inactivated human serum derived from plasma or whole blood. TFPI secreted into the media containing whole blood serum was consistently higher (≈1.5-fold at 8 hours) than that secreted into the media supplemented with serum obtained from plasma lacking the formed elements; these cells also expressed similarly increased levels of TFPI mRNA. Moreover, PMA-stimulated cells cultured in whole blood serum expressed modestly increased levels of TFPI mRNA (≈1.5-fold); supernatants from these cells also contained similarly increased TFPI activity. Cumulatively, our data indicate that, unlike thrombomodulin and fibrinolytic enzymes synthesized by the endothelial cells, TFPI synthesis is not downregulated and may be slightly upregulated during an inflammatory response. Inspection of the 5′ flanking region of the TFPI gene showed a conserved GATA-binding motif located ≈400 bp upstream of the proposed transcription initiation site(s). This motif by binding to the GATA-2 transcriptional factor may keep the endothelium in an 'on' state for constitutive expression of TFPI. A modest increase in TFPI expression by treatment of cells with whole blood serum or PMA may be due to the presence of two transforming growth factor-β-like responsive elements and of two phorbol ester-like responsive elements located in the 5′ end of the gene.