Abstract

During experimental lung metastasis, tumor cells adhere to the pulmonary microvasculature and activate coagulation via surface-expressed tissue factor (TF), leading to local fibrin deposition and platelet activation. This thrombotic microangiopathy, which facilitates tumor cell survival and extravasation, results in intravascular hemolysis and consumption of platelets and clotting factors. While interventional studies have demonstrated convincing efficacy of anticoagulants and anti-platelet agents in inhibiting metastasis, no information is available on how tumor biology may be affected by congenital bleeding disorders such as hemophilia A. We therefore used a syngeneic model of mouse melanoma to provide mechanistic insight into this area of uncertainty. By conventional RT-PCR, indirect flow cytometry using a novel polyclonal rabbit anti-mouse TF antibody, and one-stage clotting assays we detected strong expression of TF mRNA, antigen, and procoagulant activity in the murine B16F10 melanoma cell line. For metastasis experiments, control (n=17) and hemophilic mice (n=18), both on a mixed B6/129 background, were injected via the lateral tail vein with 2 x 105 B16F10 cells. After 18 days, macroscopic tumor nodules were counted on the lungs of sacrificed animals. Compared to controls (median 194, IQR 112–241), lung tumor formation was significantly reduced in mice with hemophilia A (100, 63–125; P<0.01). This genetic protection was completely reversed by a single dose of human FVIII (hFVIII) at 100 U/kg body weight 15 min prior to tumor cell injection (190, 156–300; n=10, P<0.001) and further pronounced by lepirudin (5 μg), a direct thrombin inhibitor (64, 58–73; n=7, P<0.05), suggesting that thrombin generation contributed to hematogenous metastasis in the absence of FVIII. Infusion of hFVIII into control mice (n=10) did not significantly alter the efficacy of melanoma lung seeding (P=0.35), although 50% of animals in this group showed confluent colony formation (>300 tumor nodules) as compared to 0% in the group pretreated with PBS. As expected, lepirudin significantly reduced numbers of pulmonary foci in control mice (96, 82–114; n=7, P<0.01). In these experiments, mortality was associated with metastatic tumor burden rather than with hemostatic competence of study mice. To assess tumor cell-induced coagulation activation in vivo, mice were injected intravenously with PBS (n=5 per group) or 1 x 106 B16F10 cells (n=10 per group). After 15 min, blood was collected by puncture of the retro-orbital plexus and analyzed for platelet count and plasma levels of FXa and hemoglobin, a sensitive marker of intravascular hemolysis. Consistent with our previous findings, injection of B16F10 cells evoked laboratory changes of consumptive coagulopathy in both groups of mice, with less pronounced changes in FVIII-deficient animals. Three weeks after subcutaneous implantation of 1 x 106 B16F10 cells into flanks of control and hemophilic mice (n=10 per group), there was a trend towards smaller tumor volumes in the latter group (5.2±3.4 cm3 vs. 9.4±5.3 cm3, P=0.08). Spontaneous lung metastasis was observed in 25% of control as compared to 0% of hemophilia A mice. Although strong TF expression by B16F10 cells may promote thrombin-dependent metastasis in mice with hemophilia A, amplification of coagulation by host FVIII appears to be necessary for maximum melanoma lung seeding.

Author notes

Corresponding author