Activated protein C up-regulates procoagulant tissue factor activity on endothelial cells by shedding the TFPI Kunitz 1 domain

Abstract

Thrombin and activated protein C (APC) signaling can mediate opposite biologic responses in endothelial cells. Given that thrombin induces procoagulant tissue factor (TF), we examined how TF activity is affected by APC. Exogenous or endogenously generated APC led to increased TF-dependent factor Xa activity. Induction required APC's proteolytic activity and binding to endothelial cell protein C receptor but not protease activated receptors. APC did not affect total TF antigen expression or the availability of anionic phospholipids on the apical cell membrane. Western blotting and cell surface immunoassays demonstrated that APC sheds the Kunitz 1 domain from tissue factor pathway inhibitor (TFPI). A TFPI Lys86Ala mutation between the Kunitz 1 and 2 domains eliminated both cleavage and the enhanced TF activity in response to APC in overexpression studies, indicating that APC up-regulates TF activity by endothelial cell protein C receptor-dependent shedding of the Kunitz 1 domain from membrane-associated TFPI. Our results demonstrate an unexpected procoagulant role of the protein C pathway that may have important implications for the regulation of TF- and TFPI-dependent biologic responses and for fine tuning of the hemostatic balance in the vascular system.

Figure 1

APC enhances TF procoagulant activity on endothelial cells. EAhy926 cells were treated in all experiments with TNFα (5nM) for 5 hours to induce TF expression, unless otherwise noted. (A) Different concentrations of APC or active site–blocked APC (APC blocked with dansyl-glutamyl-glycyl-arginyl-chloromethylketone [DEGR-APC]) were added for the final 3 hours of TNFα incubation. Procoagulant TF activity was analyzed as described in “Xa generation assay.” Activity of generated factor Xa relative to only TNFα-stimulated controls is shown. (B) APC (20nM) was added for the indicated period at the end of the TNFα incubation. (C) Cells were incubated for 5 hours in the presence or absence of TNFα, and different concentrations of thrombin and APC (10nM) were added for the final 3 hours as indicated. (D) APC (20nM) and/or the PI3K inhibitor wortmannin (100nM) was added for the final 3 hours of TNFα incubation. Means ± SEM are shown; n = 5 (A, B), 6 (C), or 3 (D). *P < .05; **P < .005, compared with corresponding sample without APC.

Figure 2

Increased TF procoagulant activity in response to exogenous and endogenously generated APC requires EPCR binding but not PAR1 cleavage. TNFα-induced (5 hours) cells were treated for the final 3 hours with the indicated agonists. Blocking anti-EPCR (RCR-252, 25 μg/mL), nonblocking anti-EPCR (RCR-92, 25 μg/mL), cleavage blocking anti-PAR1 (ATAP2 and WEDE15, 10 and 25 μg/mL, respectively), and anti–protein C (C1, 25 μg/mL) were added 30 minutes before the agonists. Factor Xa generation is shown in panels A and B. For the experiments shown in panel B, APC activity in the cell medium at the end of the incubation time was also analyzed. No amidolytic activity was detected in the absence of protein C and only the results in the presence of protein C are shown. Means ± SEM; n = 5 (A) and 9 (B). **P < .005 compared with the corresponding sample without APC or protein C.

Figure 3

APC does not lead to increased surface expression of TF or availability of anionic phospholipids. TNFα- or control-treated cells were incubated for 3 hours with the indicated agonists. (A) Surface expression of TF was determined by immunoassay using monoclonal anti-TF. Prothrombin activation (B) and annexin V binding (C) were analyzed on the cell surface as measures of negatively charged phospholipids as described in “Reagents, antibodies, and assays.” HgCl₂ was added only for 30 minutes. Agonists were removed by washing steps before the assays were performed. Means ± SEM with n = 6 are shown.

Figure 4

APC leads to a loss of the K1 domain from cell-associated TFPI. (A) Cells were treated with siRNA targeting TFPI or control for 24 or 48 hours. TFPI was detected by Western blotting in cell lysates using a K2 domain–specific antibody. (B) Cells were treated for 3 hours with the indicated agonist concentrations followed by detection of TFPI with anti-K2 or anti-K1 domain–specific antibodies. (C) Cells were treated with 20nM APC or factor Xa for the indicated time period followed by detection of TFPI with anti-K2. (D) Cells were preincubated for 15 minutes with 25 μg/mL blocking (RCR-252) or nonblocking (RCR-92) anti-EPCR followed by a 3-hour incubation with 20nM APC or factor Xa and Western blotting using anti-K2. (E) Cells and supernatant were analyzed after incubation with TNFα (5 hours) and 20nM APC (3 hours) as indicated. Growth medium was replaced with TNF-containing serum-free DMEM as in the Xa generation experiments. (F) Cells and supernatant were analyzed without replacing the growth medium for the agonist incubation. Where indicated, 25nM recombinant soluble TFPI (tTFPI) were added before APC. (G) Cells were treated for 3 hours with the indicated concentrations of thrombin in the absence or presence of protein C. (H) Cells were agonist treated as indicated in the presence of 80nM protein C. Polyclonal anti-TFPI was used for detection in panels E through H. Typical results from 2-4 experiments are shown in all panels.

Figure 5

The TFPI K2 domain is detected by immunoassay on APC-treated but not factor Xa-treated cells. Cells were treated with various concentrations of APC and factor Xa for 3 hours followed by cell surface immunoassay using the indicated antibodies. Thrombin (5nM) was present during the 3-hour incubation in panels B and C where indicated. Means ± SEM with n = 6 are shown, *P < .05; **P < .005, compared with no APC/Xa.

Figure 6

TFPI K1 domain shedding is required for the up-regulation of procoagulant TF by APC. (A) TNFα-induced (5 hours) cells were treated for the final 3 hours with APC (20nM). Factor Xa generation was analyzed after a 30-minute pretreatment with different concentrations of blocking polyclonal anti-TFPI. (B) HEK293t cells were transfected with expression constructs for EPCR alone or in the presence of TFPIα or TFPIβ. The following day cells were incubated for 3 hours with control or APC (60nM) and TFPI in the supernatant (S) or cells (C) was analyzed by Western blotting using polyclonal anti-TFPI. (C) HEK293t cells were transfected with EPCR alone or in the presence of wild-type TFPIβ or TFPIβ K86A. The following day cells were incubated for 3 hours with control or APC (60nM) and TFPI expression in the cells was analyzed by Western blotting using polyclonal anti-TFPI. (D) HEK293t cells were triple transfected as indicated with EPCR, TFPI, and different amounts of TF expression construct. Factor Xa generation was analyzed the next day after a 3-hour incubation with control or APC (60nM). (E) HEK293t cells transfected with EPCR, TFPIβ, and TF (20 ng/well) were treated for 3 hours with APC (60nM). Xa generation was analyzed after a 30-minute pretreatment with the indicated concentrations of blocking polyclonal anti-TFPI. (F) HEK293t cells transfected with EPCR, TFPIβ, and different amounts of TF or TNFα-induced EAhy926 cells were treated for 3 hours with APC (60nM) followed by analysis of Xa generation. Means ± SEM with n = 3 are shown in panels A, D, E, and F. *P < .05; **P < .005. Typical results from 2-3 experiments are shown in panels B and C.