Summary. Background: Tissue factor (TF)-bearing microparticles (MP) from different origins are thought to be involved in the pathogenesis of cancer-associated thrombosis. However, the role of circulating tumor cell-derived TF is not well understood. Methods: TF antigen and activity were measured in MP generated in vitro from human TF-expressing cancer cells by ELISA and clotting or thrombin generation assays, respectively. TF antigen and activity were also measured in vivo in cell-free plasmas from mice previously injected with in vitrogenerated MP or in cell-free plasmas from nude mice bearing orthotopically injected human cancer cells. Results: Tumor cellderived MP (TMP) exhibited strong TF-dependent procoagulant activity (PCA) in vitro and in vivo. Injection of TMP into mice was associated with acute thrombocytopenia and signs of shock, which were prevented by prior heparinization. Human TF antigen and activity could be detected in mouse cell-free plasmas up to 30 min after TMP injections. Human TF was detected in the spleen of injected mice and its clearance from circulation was delayed in splenectomized mice, suggesting the involvement of the spleen in the rapid clearance of circulating MP in vivo. Detectable levels of TF-dependent PCA and thrombin-antithrombin complex were found in cell-free plasmas from mice growing pancreatic human tumors, suggesting that circulating tumor-derived TF causes coagulation activation in vivo. Conclusions: MP derived from certain cancer cells exhibit TF-dependent PCA both in vitro and in vivo. These results provide new information about the specific contribution of tumor-derived MP to the hypercoagulable state observed in cancer.

Thromboembolic disease is a frequent complication in cancer. Tissue factor (TF), shown to be involved in tumor growth and metastasis, is also considered to play a central role in the pathogenesis of cancer-associated thrombosis. Circulating TF-bearing microparticles (TF+ MPs) have been found in the plasma of patients with different malignancies and are thought to contribute to their hypercoagulable state. Although numerous studies have focused on TF+ MPs derived from blood cells, there is no information available on the pathological relevance of MPs originating from tumor cells. We conducted a study to detect, enumerate and characterize the procoagulant activity (PCA) of MPs released specifically from tumor cells. MPs from high (MDA-231) and low (MCF-7) TF-expressing human breast carcinoma cells were generated ex vivo in whole blood or in buffer under stirring conditions for 45 minutes. The numbers (MPs/ml) of total and TF-expressing tumor-derived particles (TMPs) in cell-free plasmas were measured by flow cytometry using FITC-labeled annexin V and a PE-labeled monoclonal anti-TF antibody respectively. The PCA of TMPs was measured by a one stage clotting assay and a highly sensitive fluorogenic thrombin generation assay. In order to evaluate the PCA of circulating TMPs, we injected 2x106 TF+ MPs derived from MDA-231 cells into mice via the tail vein. Human TF antigen and activity were measured in cell-free mouse plasmas at various intervals (5–420 min) after injections by ELISA and clotting assay, respectively. MPs less than 1μm in diameter were released from tumor cells in both buffer and whole blood by stirring. TMPs positive for TF consisted of approximately 40% of the annexin V+ MPs, and such particles derived from as low as 1x105 MDA-231 cells could be enumerated reliably (2.5x104 MPs/105 cells). By ultracentrifugation of cell-free plasmas, we confirmed that TF antigen was associated entirely with the MP fraction. TMPs derived from as few as 450 MDA-231 cells shortened plasma recalcification times from 525 ± 114 to 265 ± 148 (P<0.01), and significantly accelerated thrombin generation as evidenced by a 3 fold shortening in lag time, and a 2 fold increase in the rate of thrombin formation and thrombin concentration. No PCA was detected with MCF-7-derived TMPs. The PCA of TMPs was inhibited completely by a blocking anti-TF monoclonal antibody, but not by saturating concentrations of annexin V (an inhibitor of phospholipid PCA) or corn trypsin inhibitor (an inhibitor of the intrinsic pathway). Five minutes following the injection of TMPs into mice, appreciable levels of human TF antigen and activity were detected in cell-free plasmas. Both TF activity and antigen decreased over time and were detectable no longer than 30 minutes after injection, indicating a rapid clearance of circulating TMPs in vivo. In contrast, when TMPs were incubated in human whole blood ex vivo for various intervals, TF activity remained unchanged in cell-free plasmas for at least 5 hrs and TF antigen was not detected by flow cytometry on any blood corpuscles, including platelets, at all intervals. However, when whole blood containing TMPs was clotted by recalcification, no TF activity could be detected in the serum, indicating the incorporation of TMPs in formed clots. In summary, MPs bearing active TF are released readily from tumor cells and can significantly activate coagulation even at very low concentrations. These results provide new insights towards the potential contribution of TMPs to the pathogenesis of cancer-associated thrombosis.

Preoperative evaluation of patients presenting with ovarian masses is challenging, partly due to shortcomings with the commonly used marker, CA-125. Ovarian cancer is associated with systemic coagulation activation. Measurement of D-dimer, serum tissue factor (TF), and the coagulation process as a whole are considered candidates for improving discrimination between benign and malignant ovarian masses. We therefore sought to identify possible benefits by analyzing preoperative coagulation status in conjunction with CA-125 in patients with ovarian masses. Preoperative blood from 95 patients with ovarian masses (75 benign, 20 malignant) and 30 controls was analyzed, prospectively. Thromboelastography served for global hemostatic assessment. Plasma TF antigen and D-dimer were measured by ELISA and microparticle-associated TF activity by thrombin generation assay. TF microparticles were enumerated by flow cytometry. Time to clot formation by thromboelastography was similar between patients having either benign or malignant ovarian tumors. Clot formation rate, clot strength, and coagulation index were significantly increased in patients having malignant versus benign tumors, indicating that thromboelastography differentiated malignant from benign tumors. D-dimer alone differentiated malignant from benign ovarian tumors and also improved differentiation when combined with CA-125. Circulating TF antigen, activity, and TF microparticle numbers, however, failed to differentiate benign from malignant tumors. Significant coagulation activation occurs in women with ovarian malignancies. Plasma D-dimer may help discriminate between patients with benign and malignant tumors. Thromboelastography may also contribute meaningfully when combined with CA-125 in the preoperative evaluation of ovarian masses. Larger studies are needed to assess these possibilities.

No abstract

3394 The generation of thrombin is the pivotal event in the process of blood coagulation. In vivo, thrombin generation is regulated by cooperation between the vascular endothelium and the pro- and anti-coagulant systems in blood, such as the thrombin/endothelial cell-dependent activation of the protein C anticoagulant pathway that ultimately leads to inactivation of factors Va and VIIIa. In vitro, thrombin generation is assessed most commonly in platelet-poor plasma by a fluorogenic substrate-based thrombin generation assay (TGA) in microtiter plates. While this assay can accurately measure the kinetics of thrombin generation in plasma (including Lag time [LT], Peak thrombin [PT], and endogenous thrombin potential [ETP]), it does not assess the influence of the endothelial cell-dependent protein C pathway on thrombin generation. As a result, the assay has limited value in the assessment of the hypercoagulable patient. In the present study we introduced a surrogate endothelium to the TGA, thus including activated protein C-induced inactivation of factors Va and VIIIa in the assay system. Wells of flat-bottomed microtiter plates were coated with 3 × 104quiescent EA.hy926 endothelial-like cells which consistently express thrombomodulin (TM) as well as the endothelial protein C receptor (EPCR). The concentration of active TM associated with EA.hy926 in the assay well was determined by a chromogenic assay and found to be ∼0.5 nM. Tissue factor (TF)-initiated thrombin generation was evaluated in normal pooled plasma (NP), and in protein C-deficient (PCd), protein S-deficient (PSd), and heterozygous factor V Leiden (fVL) plasmas, in the presence or absence of endothelial cells. Thrombin generation in NP was reduced in the presence of endothelial cells as evidenced by a 55% reduction in PT, and a 40% reduction in ETP. LT was prolonged by 83% in the presence of endothelial cells. However, in PCd, PSd and fVL plasmas, endothelial-induced suppression of thrombin generation was blunted. Specifically, only relatively small reductions were observed in PT (26% in PCd, 28% in PSd, 20% in fVL) and ETP (12% in PCd, 6% in PSd, and 14% in fVL). Furthermore, only in the presence of endothelial cells, addition of protein C to PCd plasma dose-dependently reversed the effect of protein C deficiency on PT and ETP. In separate experiments in the absence of endothelial cells, we noted that while similar results were obtained in the presence of the soluble form of thrombomodulin, the concentration required (30 nM) was 60 times higher than that expressed on the endothelial cell monolayer. This suggests that protein C activation occurs more efficiently in the presence of EA.hy926 cells. In summary, by introducing an endothelial cell monolayer to the TGA, and measuring thrombin generation kinetics in the presence or absence of these cells, we have adapted the assay to assess the contribution of the protein C anticoagulant system to thrombin generation in a physiologically relevant manner. This novel approach not only enables the functions of the endothelial-dependent PC pathway by expressing TM and EPCR, but also may provide other endothelial components relevant to thrombin generation (such as tissue factor pathway inhibitor [TFPI]). This approach to thrombin generation assessment may therefore have both research and clinical applicability. Disclosures: No relevant conflicts of interest to declare.

Coagulation and platelet activation occurs in patients with ovarian cancer, and several factors underlying such hypercoagulability have been implicated in tumor growth and metastasis. The incidence of venous thromboembolism throughout the course of ovarian cancer is high, and identification of hypercoagulable patients may be important not only for effective thromboprophylaxis, but also for future anti-hemostatic treatment strategies targeting disease progression. The purpose of this prospective study was therefore, first, to characterize in more detail, the hypercoagulable state in patients with ovarian tumors and, second, to define markers of coagulation or platelet activation that specifically identify patients with malignant disease. Thirty nine patients presenting to our institution with an ovarian mass of unknown type (41–69 yrs) and 32 age-matched controls were included in the study. Pre- and postoperative (7–14 days) blood samples were collected, and all whole blood (WB) analyses were performed and plasmas prepared and frozen within 2 hours thereafter. WB thrombelastography (TEG) served for global coagulation assessment. D-Dimer and prothrombin fragment F1+2 were used as markers of coagulation activation. Soluble CD40 ligand (sCD40L), soluble P-selectin, platelet-monocyte conjugates, and platelet-derived microparticles (PMP) were measured to assess in vivo platelet activation. Tissue factor (TF) antigen and TF+ microparticles were quantified in plasma by ELISA and flow cytometry, respectively. Compared to controls, mean values for TEG-R, a measure of clotting time, and TEG-MA, a measure of clot strength, were significantly decreased (10.1 vs. 7.3 min, P<0.05) and increased (63 vs. 68 mm, P<0.001), respectively, in preoperative samples. Median levels of TF antigen (88 vs. 28 pg/ml, P<0.01), D-Dimer (255 vs. 133 ng/ml, P<0.01), and sCD40L (223 vs. 162 pg/ml, P<0.01) as well as platelet-monocyte conjugates (66 vs. 45%, P<0.001) and plasma PMP (10.2 vs. 5.7x106/ml, P=0.01) were significantly higher in preoperative samples than in controls. Interestingly, no correlation was found between plasma TF antigen levels and absolute numbers of TF+ microparticles. Moreover, while we found significant correlations between the various markers of platelet activation, there was no obvious association between the two TF indices and any of the hemostatic parameters analyzed. D-Dimer values (1310 vs. 255 ng/ml, P<0.001) and TEG-MA (73 vs. 68 mm, P<0.05) were significantly elevated in postoperative as compared to preoperative samples, indicating further and prolonged coagulation activation by the surgical trauma. Furthermore, when compared to patients with benign tumors (n=31), values for D-Dimer (215 vs. 2329 ng/ml, P=0.009) and TEG-MA (66.9 vs. 72.6 mm, P=0.007) were significantly elevated in those with histologically proven ovarian cancer (n=8). Importantly, two of three patients with metastatic disease had the highest D-Dimer and TEG-MA values of the entire patient cohort. In summary, our results demonstrate significant coagulation and platelet activation in women presenting with an ovarian mass. In addition, plasma D-Dimer and WB TEG may be helpful in identifying individuals at high risk for perioperative thromboembolic complications and metastatic malignant disease.