Platelet Pitpa Is Not Essential for Hemostasis, but Is Required for Tumor Metastasis Formation

It has been observed that cancer patients who have thrombocytosis frequently have metastases and a poor prognosis. Tumor metastasis formation is a complex process that involves tumor cells undergoing an intricate series of interactions with target tissues, vascular endothelium, lymphocytes, and even platelets. To test the hypothesis that platelet signaling contributes to the dissemination of tumors, we analyzed whether the disruption of phosphatidylinositol metabolism impaired metastasis formation. Phosphatidylinositol and its phosphorylated derivatives, phosphoinositides(PtdIns), are minor constituents of phospholipids at the cellular membrane level, yet they are essential components of intracellular signaling that regulate diverse cellular processes, including platelet plug formation. PtdIns are removed from cell membranes and transported through the cytoplasm by two distinct phosphatidylinositol transfer proteins (PITPs) called PITPα and PITPβ. In murine platelets, PITPα is the dominant isoform and it is present 7-fold more abundantly than PITPβ. We generated mice with a conditional deletion of PITPα specifically in platelets and in megakaryocytes, and we investigated the impact of this deletion on platelet-mediated hemostasis and tumor metastasis formation. Deletion of PITPα caused only a 15% reduction in the platelet count. However, the loss of PITPα impaired the synthesis of PtdIns(4)P and impacted the PtdIns(4,5)P₂ level by 40-50% in either resting or thrombin stimulated platelets. Furthermore, deletion of PITPα resulted in an 80% reduction in platelet InsP₃ production. These biochemical defects also correlated with ex vivo platelet function defects. Platelets lacking PITPα had an approximately 70~80% reduction in ex vivo platelet secretion following stimulation by low or high concentrations of thrombin, collagen, and U46619, (Thromboxane A2 analog) and ~40% reduction in platelet aggregation by low concentrations of thrombin. However, platelet accumulation area and secretion (as measured by P-selectin expression) at sites of laser-induced injuries in vivo were completely normal in mice lacking PITPα in their platelets. Therefore, despite the importance of PITPα in platelet function ex vivo, we could not demonstrate that this protein contributed to in vivo hemostasis. To determine whether PITPα-mediated phosphoinositide metabolism was critical for tumor dissemination in vivo, we utilized a well-described mouse model of lung metastasis formation following the intravenous injection of B16F10 melanoma tumor cells. We observed that lung metastasis formation was reduced by 47%±18% in mice lacking PITPα in their platelets. We also found that during the initial hours after tumor injection, control mice developed a rapid and transient thrombocytopenia. Histologic analysis revealed that tumor cell injection induced activation of wild-type platelets causing adhesion of the platelets to the melanoma cells within the vasculature. In contrast, mice with PITPα-null platelets did not become thrombocytopenic after injection with tumor cells, and also failed to form large heterogeneous complexes of platelets and melanoma cells. We also observed that in wild type mice, the tumors extravasated through the vascular wall as soon as 48 hours after injection of the melanoma cells, while this was rarely observed in mice lacking platelet PITPα. However, bronchus-associated lymphoid tissue (BALT) was strikingly hyperplastic in mice lacking platelet PITPα, but it was essentially absent in the control mice. Together, these findings demonstrate that PITPα within platelets is essential for normal phosphoinositide signaling, yet it is completely dispensable for in vivo hemostasis. In contrast, PITPα is essential for the dissemination of tumors in vivo through a complex process that involves platelet-tumor interaction and an inflammatory response that is mediated by lymphoid cells.