Predicting which drugs will work on cancer patients
Doctors play a guessing game about which drugs to use in combatting cancer. One problem is human individualitywhat helps one person may fail to help another. In the search for the proper medicine, doctors must often subject a patient to a sequence of powerful drugs, many of which turn out to be ineffective against the malignant cells. Now a simple technique promises a means of testing the effectiveness of drugs in a specific case of cancerwithout having to administer them to the patient.
A team of researchers at the University of Arizona Cancer Center in Tucson reported last week in the New England Journal of Medicine that cells from a patient's cancer can be grown, or cultured, in the laboratory and tested there to determine which drugs work.
Researchers have long been frustrated by their inability to get cancer cells from patients' tumors to grow rapidly in culture. But the Arizona team, led by Dr. Sydney Salmon and Cell Biologist Anne Hamburger, discovered three years ago that by "conditioning" culture medium with spleen cells taken from mice prone to cancer, they can grow tumor cells from people with common forms of cancer. (The mouse cells apparently produce some yet unidentified factor that supports the growth of certain human cancer cells.) According to Salmon, the cancer cells that thrive and form colonies in the laboratory's plastic petri dishes appear to be the tumor's "clonogenic," or "stem," cells. Though they account for less than 1% of all the cells in a tumor, these cells are thought to be the cancer's key replicating units; they divide and migrate, "seeding" new cancers in the body in a process called metastasis.
The Arizona team began applying anticancer drugs to cells taken from tumors and then culturing the cells in order to, in their words, "determine whether there are correlations between what is observed in the petri dish and in the patient." Tumor cells taken from nine people with myeloma, a bone marrow malignancy, and nine with ovarian cancer were exposed to varying concentrations of several anticancer drugs, then cultured in petri dishes. The researchers compared the effects of the drugs on the cultured cells with the patients' responses to the same drugs. In all but one case, the effects matched. If the drug prevented cancer cells from growing in the culture, it also killed them in the patient. If the drug had no effect in the petri dish, it did not help the patient. The team also found that while patients might have the same type of cancer, their cells in culture showed markedly different responses to the same drug. Sometimes sensitivity to a drug varied by a factor of 20 or 30.
The main value of the laboratory test, says Salmon, is that it can help the physician plan individual courses of treatment. For example, only 20% of people with cancer of the colon or rectum respond to the drug fluorouracil; the other 80% suffer needlessly from the drug's toxic effects. The new technique may have another benefit: it could be used to evaluate new anticancer drugs without endangering cancer patients.