The spontaneous formation of blood clots can be an indicator of neoplastic disease and was first described by Trousseau in 1865. Boccaccio and colleagues have now identified the genetic link between blood hypercoagulation and cancer.

The proto-ongogene MET , which encodes the hepatocyte growth factor/scatter factor receptor, is known to be important for invasive growth during tissue morphogenesis and repair, and is also mutated in a number of human cancers. Boccaccio and colleagues expressed a constitutively active form of human MET under the control of the albumin promotor to limit its expression to mouse hepatocytes. Cancer most often arises from the transformation of somatic cells, so the authors used a self-inactivating lentiviral vector to target these cells in adult mice. A GFP-expressing lentivirus was used as a control.

MET was successfully expressed in the somatic hepatocytes and nodules of dysplastic cells were evident after 3 months. By this time, however, many of the MET-expressing mice had already developed blood clots, evident in their tails, and by 6 to 7 months, as a result of the exhaustion of the blood-coagulation system, many had haemorrhages in all major organs — a condition known as disseminated intravascular coagulation (DIC). Only 5 of 25 mice survived this and, of these, only a few went on to develop overt liver neoplasia.

Hepatocytes are important for the synthesis of a number of blood-clotting factors, so to rule out a hepatocyte-specific effect the authors expressed MET in mammary epithelial cells and transplanted these into the spleens of mice. These mice also developed DIC.

Microarray analysis has indicated that MET upregulates the expression of a number of genes associated with blood homeostasis. Indeed, the mRNAs encoding two factors, plasminogen activator inhibitor type 1 (PAI1) and cyclooxygenase-2 (COX2) are among the most upregulated of the 12,000 genes assessed. PAI1 inhibits fibrin degradation and COX2 regulates platelet function, so the authors investigated whether the increased expression on these two proteins could explain the pro-thrombotic state. They found that PAI1 and COX2 were highly expressed in the MET-transformed mammary cells and were also expressed in the dysplastic liver nodules. Treatment of the mice with inhibitors of PAI1 or COX2 suppressed the onset of DIC, and, after prolonged treatment, the dysplastic liver foci also started to regress.

These data indicate for the first time that a single gene can both promote tumour formation in a small subset of cells and systemically affect blood coagulation by virtue of its effects on the expression of other genes. The therapeutic implications of these findings need to be carefully considered.