Platelet granule secretion continuously prevents intratumor hemorrhage

Abstract

Cancer is associated with a prothrombogenic state capable of platelet activation. Platelets, on the other hand, can support angiogenesis, a process involved in the progression of tumor growth and metastasis. However, it is unclear whether platelet/tumor interactions substantially contribute to tumor physiology. We investigated whether platelets stabilize tumor vessels and studied the underlying mechanisms. We induced severe acute thrombocytopenia in mice bearing s.c. Lewis lung carcinoma or B16F10 melanoma. Intravital microscopy revealed that platelet depletion led to a rapid destabilization of tumor vessels with intratumor hemorrhage starting as soon as 30 min after induction of thrombocytopenia. Using an inhibitor of glycoprotein Ibalpha (GPIbalpha) and genetically engineered mice with platelet adhesion defects, we investigated the role of platelet adhesion receptors in stabilizing tumor vessels. We found that a single defect in either GPIbalpha, von Willebrand factor, P-selectin, or platelet integrin activation did not lead to intratumor hemorrhage. We then compared the ability of transfused resting and degranulated platelets to prevent intratumor hemorrhage. Whereas resting platelets prevented thrombocytopenia-induced tumor bleeding, circulating degranulated platelets did not. This suggests that the prevention of intratumor hemorrhage by platelets relies on the secretion of the content of platelet granules. Supporting this hypothesis, we further found that thrombocytopenia dramatically impairs the balance between propermeability and antipermeability factors in tumor-bearing animals, in particular depleting blood of angiopoietin-1 and serotonin. Our results show a crucial contribution of platelets to tumor homeostasis through continuous prevention of severe intratumor hemorrhage and consequent cell death. The study also suggests platelet function as a reasonable target for specific destabilization of tumor vessels.

Figure 1. Acute thrombocytopenia induces tumor bleeding independently of the tumor type, age, and location

A. At day 8 after subcutaneous implantation of LLC tumor cells, mice were injected with either the control IgG (control) or the platelet-depleting IgG (depleted) and tumors were photographed 18 hours later. Bar = 5 mm. B. H&E staining of the LLC tumors showed massive accumulation of red blood cells in the tumor stroma only in platelet-depleted mice (arrows). Bars = 100 μm. C. Thrombocytopenia was induced at day 4, 8, or 12 following subcutaneous LLC implantation. 18 hours later, intratumor hemoglobin content was determined and compared to control IgG-treated tumors (n = 4). D. 10 days following either subcutaneous (upper panel) or intravenous (lower panel) injection of B16F10 melanoma cells, mice were injected with either the control IgG (control) or the platelet-depleting IgG (depleted). 18 hours later, photographs of skin and lungs were taken. Hemorrhage was observed only in tumors from platelet-depleted mice (arrows). Bars = 5 mm.

Figure 2. Kinetics and localization of tumor bleeding in platelet-depleted mice

A. Mouse carrying a dorsal skinfold chamber. B. Mice bearing 5 day old LLC tumors were injected with Evans blue and either the control (upper panel) or the platelet-depleting antibody (lower panel) at time 0. Tumors were observed through the dorsal skinfold window for 3 hours. Times post-infusion are indicated. Bar = 500 μM. C. LLC tumor viewed 3 hours after induction of thrombocytopenia. White arrow: intratumor hemorrhage, black arrow: hemorrhage occurring from vessels surrounding the tumor. Bar = 500 μM.

Figure 3. Thrombocytopenia reduces cancer cell proliferation and locally affects tumor viability

At day 8 after subcutaneous implantation of LLC tumors, thrombocytopenia was induced and 48 hours later, tumors were harvested and sectioned. A. Mice were injected with BrdUrd 3 hours before sacrifice and the proliferative index was calculated as the percentage of BrdUrd positive nuclei relative to DAPI-stained nuclei (n = 10 microscopic fields out of 4 tumors for each). B. The apoptotic index was calculated as the percentage of TUNEL positive nuclei relative to DAPI-stained nuclei (n = 10 microscopic fields out of 5 tumors for each). C. H&E, DAPI and TUNEL staining of non-hemorrhagic and hemorrhagic areas of the LLC tumors. Arrows indicate fragmented and condensed nuclei. Bars = 20 μm.

Figure 4. Prevention of tumor hemorrhage by platelets is independent of platelet GPIbα

PBS or the GPIbα chimera (GPG-290) was injected intravenously at day 8 following subcutaneous LLC tumors implantation. A. Tail bleeding time was assessed 18 hours after GPG-290 injection. B. Comparison of hemoglobin content of control and GPG-290-treated tumors (n = 5). No difference in hemoglobin content was found between the two groups.

Figure 5. Platelet depletion effects on serum concentrations of VEGF, angiopoietin-1 and serotonin

A. Comparison of VEGF levels in LLC tumor-bearing control and platelet-depleted mice (n = 5). B. Comparison of angiopoietin-1 levels between control and platelet-depleted mice (n = 5). Inset. Western blot detection of angiopoietin-1. Lane 1: platelet poor plasma from control mouse, lane 2: plasma from platelet-depleted mouse, lane 3: serum from control mouse, lane 4: serum from platelet-depleted mouse. Arrow indicates angiopoietin-1. C. Comparison of serotonin levels between control and platelet-depleted mice (n = 4–6). Platelet depletion led to disappearance of serotonin and angiopietin-1 from serum without affecting VEGF levels.

Figure 6. Degranulated platelets are unable to prevent thrombocytopenia-induced tumor bleeding

A. Degranulation of thrombin-stimulated platelets was assessed by FACS analysis of P-selectin surface expression and by B. Quantitation of angiopoietin-1 in platelet supernatants by ELISA. C. At day 8 after tumor cells implantation, mice were injected with either the control IgG (control) or the platelet-depleting IgG (depleted). A subset of mice was transfused 30 minutes prior to the induction of thrombocytopenia with tyrode buffer (no transfusion) or with 7 × 10⁸ of either resting (resting platelets) or activated platelets (activated platelets) and subcutaneous LLC were photographed 18 hours later. Bar = 5 mm. D. Comparison of the hemoglobin content of control tumors and platelet-depleted tumors from mice transfused with either tyrode buffer, resting platelets, or activated platelets (n = 17–20).