Dr. Robert Weinberg of the Massachusetts Institute of Technology pointed out in a recent campus talk that about 90 percent of cancer deaths are caused not by the primary tumor itself, but by its metastases — the secondary, distant growths arising from the primary tumor. Weinberg's lab and others have been making significant progress toward understanding how tumors acquire the ability to invade and metastasize.

Each cancer has its own development pathway, but many types of human cells, Weinberg explained, can be transformed experimentally into cancer cells by altering five regulatory pathways: the replicative lifespan pathway (involving telomeres), apoptosis (cell death pathway involving p53), proliferative stimuli from outside the cell (like growth factors), control of the cell cycle (e.g., pRb) and the mobilization of resources that give the cell specific advantages (like the enzymes involved in nutrient metabolism).

Yet the story of cancer extends beyond these five pathways. Carcinomas, which constitute about 4 out of every 5 tumors, are made of both cancerous epithelial cells and their stroma — the connective tissue framework that supports their growth. Transformed human mammary epithelial cells form tumors much faster in mice when supplied with stromal cells like fibroblasts; Weinberg's lab has found that stromal cells from most breast cancers are far better at supporting tumor growth than those from normal breast tissue.

While many tumor-associated stromal cells probably originate in the normal, adjacent host stroma, others come from circulating cells. Weinberg's lab has shown that tumor-associated stroma are superior at recruiting the cells that form the vasculature to support a tumor's growth. His lab is now working to identify the signals that primary tumors use to recruit stromal cells and to identify the precursors to these stromal cells that support tumor growth.

Weinberg went on to discuss how metastatic tumor cells arise. The invasion-metastasis process involves several complex steps: local invasiveness, invasion into blood and lymphatic vessels, transport through the circulation, escape from blood vessels into the surrounding tissue, formation of a micrometastasis, and finally colonization, the growth of a micrometastasis into a full metastasis.

Through gene expression array screening in highly metastatic mouse breast cancer cells, Weinberg's lab identified two transcription factors that are involved in invasion and metastasis, Twist and Mesenchyme Forkhead. They identified two more, Goosecoid and Slug, to make a quartet of transcription factors that are normally active during early embryogenesis, in steps of development that require a change in cell behavior known as the epithelial-mesenchymal transition. These transcription factors allow epithelial cells in the embryo to acquire many characteristics also seen in motile, invasive cells. Metastatic cancers are somehow able to upregulate these embryonic transcription factors, thereby acquiring the ability to invade and migrate. The signals for these changes originate in nearby stroma, but their exact nature is still unclear.

What is clear is that genetic mutations don't tell the whole story of a cancer. The tumor's stroma affects both its formation and its ability to metastasize. This understanding is broadening the strategies researchers are using to combat cancer.