[CANCER RESEARCH 45, 5521-5525, November 1985] Tumor Cell Generation of Thrombin via Functional Prothrombinase Assembly1 Livingston VanDeWater, Paula B. Tracy,2 Denise Aronson, Kenneth G. Mann,2 and Harold F. Dvorak3 Departments of Pathology, Beth Israel Hospital and Harvard Medical School and the Charles A. Dana Research Institute, Beth Israel Hospital, Boston, Massachusetts 02215 (L. V. D. W., D. A., H. F. D.¡,and Hematology Research Section, Mayo Clinic, Rochester, Minnesota 55095 [P. B. T., K. G. M.¡ ABSTRACT These classical pathways are now recognized as a network of activation and feedback reactions that converge on a final com Prothrombinase affects the proteolytic conversion of pro- mon pathway of prothrombinase and thrombin generation. Pro- thrombin to thrombin and is the penultimate enzyme in the coagulants acting at steps proximal to this common pathway will common coagulation pathway. Prothrombinase is a complex in be ineffective in generating thrombin and depositing fibrin if which the proteinase, Factor Xa, a cofactor, Factor Va, and prothrombinase is not active. calcium are bound to a membrane surface to generate the active Prothrombinase complex formation is thus a critical step in enzyme. Guinea pig line 1 and line 10 tumor cells, grown as coagulation and one that has been largely neglected in studies primary cultures from ascites tumors or as cell lines in culture, of tumor-associated clotting. Prothrombinase is a complex of an provide a surface that interacts with coagulation Factor Va and active protease (Factor Xa) with a non-enzymatic cofactor (Fac Xa and with calcium ions to form this enzyme complex. Cultured tor Va) bound to an appropriate membrane surface in the pres human colorectal carcinoma cells (Colo 205) also participate in ence of calcium ions. Platelets (18-22), some inflammatory cells Prothrombinase complex assembly and function. Prothrom (23), endothelial cells (24), and certain phospholipid vesicles of binase generation was measured by following the kinetics of defined composition (25) all provide an appropriate surface for prothrombin conversion to thrombin. Thrombin generation was prothrombinase complex assembly; I.e., they afford a surface monitored continuously using the reversible thrombin inhibitor, that binds Factor Va and allows it to serve as a receptor for dansylarginine A/-(3-ethyl-1,5-pentanediyl)amide, which displays Factor Xa. In the absence of such a surface, the capacity of enhanced fluorescence upon binding to thrombin. Analyses of prothrombinase to cleave prothrombin to thrombin falls nearly kinetic data indicate that the apparent dissociation constants (1- three orders of magnitude (26); thus, provision of a suitable 4x10~10 mol/liter) and the number of Factor Va-Xa binding sites surface for prothrombinase activation, in addition to plasma per tumor cell are comparable to values reported for human and clotting factors, is required for fibrin formation under any circum bovine platelets, human lymphocytes, and monocytes. Guinea stance, including that occurring about tumors. pig lymphocytes were also active, while erythrocytes were inac We considered the possibility that tumor cells themselves tive, in the prothrombinase assay. Membrane vesicles, shed by might provide an appropriate surface for prothrombinase gener guinea pig and human tumor cells into conditioned medium, also ation and thus contribute directly to local fibrin deposition. We supported functional prothrombinase activity. Although earlier report here that several types of carcinoma cells, as well as studies indicated that tumor cells may initiate coagulation, this is membranes shed by such cells, provide an effective surface for the first demonstration that tumor cells are competent to bring activation of the prothrombinase complex. clotting to fruition by generating thrombin, a step essential to fibrin generation. These data suggest that tumor cells, in the MATERIALS AND METHODS presence of clotting initiators and appropriate coagulation fac tors, are sufficient to generate the fibrin deposited in solid tumors. Materials. Dulbecco's modified Eagle's medium, HEPES,4 calf serum, and fetal calf serum were obtained from Grand Island Biological Co. (Grand Island, NY), and fat-free bovine serum albumin was obtained from INTRODUCTION Miles (Elkhardt, IN). DAPA was prepared as described (25, 27). All other reagents were of analytical grade. Fibrin is a prominent component of the stroma of autochtho Coagulation proteins were purified from bovine plasma using published nous and transplanted tumors (1-5). Diverse functions such as procedures (28, 29). Factor V was activated by the addition of thrombin angiogenesis, desmoplasia, facilitation of fibroblast and endothe- (0.02 units/ml) for 3 min at 37°C (30). Factor X was activated using lial cell adhesion and motility, mast cell degranulation, and mod immobilized Factor X activator purified from Russell Viper Venom (29). ulation of the immune response have been attributed to fibrin Cell Culture. Guinea pig line 10 and line 1 bile duct carcinoma cells, and its catabolites (5-9). The fibrin deposited in tumors derives ascites variants, were passaged in the peritoneal cavities of inbred strain from plasma fibrinogen that has extravasated from local, hyper- 2 Sewall-Wright guinea pigs (11). The line 10 cells grow in culture as non-adherent cells. The ascites form of line 1 carcinoma cells and the permeable blood vessels. However, the mechanisms by which human colorectal carcinoma cell line (Colo 205) grow in culture as thrombin is generated in the extravascular space to convert adherent (tissue culture plastic) or non-adherent (bacteriological plastic) fibrinogen to fibrin are poorly understood. Several tumor cell cells. Cells (0.5-1 x 106 per ml) were maintained in Dulbecco's modified procoagulant activities have been described that act at various Eagle's medium supplemented with penicillin (50 ¿¿g/ml),streptomycin steps in the intrinsic and extrinsic clotting pathways (10-17). (50 units/ml), glucose (4.5 mg/ml), and 5% calf serum (line 1 and line 10) 1This work was supported by USPHSgrants CA 28471 and HL 17430. or 10% fetal calf serum (Colo 205, HT1376) (31, 32). Cells were prepared 2Current address: Departments of Medicine (P. B. T.) and Biochemistry (K. G. for prothrombinase assay by washing two times in phosphate-buffered M.), Universityof Vermont, Burlington, VT 05405. saline and once in HEPES-glucose buffer (150 mw NaCI:11 ÕTIMglu- 3To whom requests for reprints should be addressed, at Departmentof Pathol ogy, Beth Israel Hospital, 330 Brookline Ave., Boston, MA 02215. 4The abbreviations used are: HEPES, 4-(2-hydroxyethyl)-1-piperazineethane- Received 3/26/85; revised 8/9/85; accepted 8/12/85. sulfonic acid; DAPA,dansylarginineN-<3-ethyl-1,5-pentanediyl)amide. CANCER RESEARCH VOL. 45 NOVEMBER 1985 5521 Downloaded from cancerres.aacrjournals.org on September 27, 2021. © 1985 American Association for Cancer Research. TUMOR CELL PROTHROMBINASE ACTIVITY cose 10 mu HEPES, pH 7.2). from 0.01-5 nM. Changes in prothrombinase activity were inter Conditioned medium was obtained from guinea pig line 1 or 10 preted to reflect binding of Factor Va to a cellular site (21, 23, carcinoma cells or from human HT1376 by a modification of published procedures (10,11 ). Cells were washed and cultured in Eagle's Minimal 25). We found that the guinea pig line 1 and line 10 bile duct essential medium without phenol red (Flow Laboratories) and without serum at 2.5 x 10* cells/ml for 4 h at 37°C with 5% CO2. Cells (95% carcinomas and the human colorectal line, Colo 205, all provided viable by try pan blue exclusion) were removed by centrifugation at 160 efficient surfaces for prothrombinase generation. Using the anal x g for 10 min, and the medium was centrifuged at 10,000 x g for 20 ysis of kinetics of prothrombin activation to infer parameters of min. In some experiments, the resulting supernatant was centrifuged at Factor Va binding to the cell surface, we found that when Factor 100,000 x g for 90 min. Va concentrations were decreased below saturating levels, the Prothrombin Activation (Monitoring with DAPA. The formation of a reaction rate decreased according to a hyperbolic function (Chart functionally active prothrombinase complex on the surface of tumor cells 1) and represented a binding isotherm characteristic of a func was analyzed by assessing the influence of Factor Va and Xa on the tional Factor Va-cell interaction (21,23). Reaction velocities were rate of prothrombin activation to thrombin (21). This approach is based on the observation that deletion of the catalytic "membrane surface' assumed to be directly proportional to the number of occupied results in a 1000-fold decrease in reaction rate and, therefore, an Factor Va sites on the cells. The ratio of observed veloc- ineffective enzymatic complex (26). No thrombin is generated unless an ity:maximum velocity allowed calculation of the fraction of oc equivalent "membrane surface" is provided by tumor cells. Prothrom- cupied sites (21,23,25). A double reciprocal plot (Chart 1, inset) bmase activity was monitored continuously using DAPA, which emits depicts the reciprocal of free Factor Va cellular binding sites increased fluorescence when it binds to newly generated thrombin (21, plotted as a function of the ratio of nominal Factor Va concentra 23, 25, 27). Tumor cells (1 x 10* cells/ml) were preincubated for 2 min tion to the fraction of filled sites. The apparent dissociation at ambient temperature in a mixture containing prothrombin (1.39 ^ M) constant calculated from the slope is Ka »1.7 x 10~10mol/liter; and Ca2* (2 mw) and varying concentrations of Factor Va, in a final volume of 1 ml in HEPES-glucose buffer, pH 7.2. Addition of Factor Xa the number of functional cellular receptor sites calculated from the y-intercept is 45,000 sites per cell in this experiment. This (5 nM) initiated prothrombinase generation, and the time course of thrombin formation (OAPA fluorescence) was continuously recorded (26).