Phosphatidylinositol (PtdIns) is a relatively rare phospholipid in cell membranes. However, because of its unique ability to be transiently phosphorylated, it is critical for platelet signaling and vascular plug formation. Phosphatidylinositol transfer proteins (PITPs) facilitate the transfer of aqueous insoluble phosphatidylinositol in vitro from one cell membrane to another. Although murine platelets contain multiple PITP variants, PITPα is by far the most abundant isoform. Platelets are best known for their contribution to hemostasis, however, several lines of evidence indicate that they also contribute to tumor metastasis formation. To test the hypothesis that platelet phosphoinositide signaling contributes to tumor dissemination, we asked whether deletion of platelet PITPα impacts tumor metastasis formation by engineering PITPαfl/fl PF4Cre+ mice that lack PITPα only within their platelets and megakaryocytes. We found that loss of PITPα in platelets decreased PtdIns(4)P and PtdIns(4,5)P2 production by 30-40%. Even more striking was the 80% reduction in IP3 formation following thrombin-stimulation. However, we found no significant defect in the platelet ex vivo aggregation in response to typical doses of most of the platelet agonists, and only a small defect in response to low doses of thrombin. We investigated whether platelets lacking PITPα normally formed thrombi in vivo by using three well-established murine models. First, we observed an extremely minor (but statistically real) prolongation of the tail bleeding time in the PITPαfl/fl PF4Cre+ mice. Second, we found that mice lacking PITPα in their platelets had no defect in forming intravascular clots in response to a chemical induced carotid injury. Finally, we observed that thrombosis and platelet a-granule secretion in response to a laser-induced injury were completely normal in the knockout mice. Therefore, despite the biochemical defect in phosphoinositide signaling induced by the loss of PITPα, there was essentially no hemostastic defect. To determine whether PITPα-mediated phosphoinositide metabolism in platelets is required for tumor dissemination, we utilized a well-characterized B16F10 melanoma model of tumor metastasis. We observed that lung metastasis formation was reduced by 47%±18% in mice lacking PITPα in their platelets. We also found that during the first 3 hours after tumor injection, the control mice, but not the PITPαfl/fl PF4Cre+ mice, developed a rapid and transient thrombocytopenia. Histology analysis of the lung tissue at this time point revealed the presence of 30% more clots in the lung tissue of control mice. Further analysis showed that these thrombi were actually heterogenous complexes composed of tumor cells surrounded by platelets and fibrin. To understand why platelet PITPα influences tumor-induced fibrin formation, we investigated the ability of tumor cells to cause thrombin generation in platelet rich plasma derived from PITPαfl/fl PF4Cre+ and PITPαfl/fl PF4Cre- mice. We observed that the loss of PITPα in platelets resulted in an 88% reduction of thrombin generation compared to the controls. Furthermore, we found that PITPαfl/fl PF4Cre+ platelets also have impaired Annexin V binding suggesting that the defective fibrin formation seen in PITPαfl/fl PF4Cre+ mice is likely due to a role of PITPα in the exposure of phosphatidylserine on the platelet surface that is required for thrombin generation. Finally, we observed a mucosal immune response composed of NK-cells, T-cells, and neutrophils that was strikingly hyperplastic in mice lacking platelet PITPα at 48 hours after tumor injection, but was essentially absent in the control mice. Together, these findings demonstrate that PITPα-mediated phosphoinositide metabolism within platelets is not essential for platelet plug formation in vivo, but is required for the dissemination of tumors in vivo. Our work further demonstrates that platelet PITPα is required for tumor-induced phosphatidylserine exposure and thrombin generation. This process induces a shroud of platelets and fibrin that surround the surface of tumor cells, and thereby protects the tumors from elimination by the mucosal immune system. These results demonstrate that it is possible to clearly distinguish the platelet signaling processes required for platelet plug formation from those processes that augment metastasis formation.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.