A perennial puzzle for cancer researchers is why some tumors spread to other parts of the body, or metastasize, while others don't. Because most cancer deaths result from metastases, answering this question could help doctors anticipate metastatic disease and help choose the optimal treatment. Now researchers have taken a small step in that direction: They've deciphered a genetic "signature" that, in several tumor samples, appears to predict whether a cancer will spread.
Tumor cells metastasize once they acquire certain genetic mutations. These might make a cell hardier in a foreign environment--say, when a breast cell relocates to bone--or more able to escape the primary tumor and survive while rushed through the bloodstream. But distinguishing between garden-variety cancer cells and the more cunning metastatic type has been tough.
In a recent attempt, oncologist Todd Golub of the Dana-Farber Cancer Institute in Boston, Massachusetts, and his colleagues collected gene expression profiles, taken from public databases, from 76 tumors. Sixty-four were from a primary site (although the team members didn't know whether those cancers had later metastasized), and 12 were from metastases. Because the metastatic and primary samples weren't from the same individuals, and because the samples were drawn from a hodgepodge of cancers (lung, breast, prostate, colorectal, uterus, and ovary), Golub didn't really expect to see anything interesting.
To the researchers' surprise, they found distinct patterns of expression among 17 genes. They were particularly intrigued to see that some primary tumors shared an expression pattern with the metastatic ones. Could these tumors be the ones that would spread? The group turned to more profiles: 78 breast cancer, 62 lung cancer, and 21 prostate cancer samples, all from primary tumors, and with follow-up information about whether they had metastasized. Although a bad outcome was possible without the gene expression pattern, patients with the genetic signature were found to survive for less time than those who didn't have it, they wrote in the 9 December online edition of Nature Genetics.
"This study has provided strong evidence demonstrating that the molecular signature of metastasis exists prominently in the primary tumor," says Mary Hendrix, a cancer researcher at the University of Iowa in Iowa City. But Hendrix stresses that the study needs to be repeated with a larger sample. Furthermore, she says, it would be nice to know the function of the 17 genes--a few of which had been previously linked to cancer--because it could help refine tumor treatments.