Genetic profiling of tumors shows high sccess in predicting how breast cancer patients will respond to chemotherapy drug

By Lorraine Cichowski and Peter O'Connel

RICHMOND, Va. – Nearly 90 percent of the time, gene expression patterns in tumor cells appear to accurately predict whether a breast cancer patient will respond well or show resistance to treatment with the anti-cancer drug Taxotere, according to a study in the current issue of The Lancet. 

Genetic profiling of a patient's tumor could help clinicians decide which chemotherapy drug or drug combination is best for that patient's breast cancer, either before or after removing the cancerous breast tumor, says Dr. Peter O'Connell, professor and chair of the Department of Human Genetics at Virginia Commonwealth University and a faculty member at VCU's Massey Cancer Center. O'Connell was senior author of the study, which appears in the Aug. 2 issue of The Lancet.

"The dilemma is that no currently available clinical test says, "this drug will help this breast cancer patient but this drug will not"," said O'Connell. "It once was thought that surgery alone cured most women of breast cancer and that unpleasant and toxic chemotherapy after breast cancer surgery was only necessary for the minority of women whose tumors had obviously spread. But it turned out that one third of the women who didn't have chemotherapy had recurrences of cancer. So now most patients with operable breast cancer undergo different forms of chemotherapy to kill any cancer cells that might have spread. For example, they might be treated with a relatively new agent such as Taxotere and a combination of the conventional chemotherapy drugs doxorubin (Adriamycin) and cyclophosamide (Cytoxan).

"There is new evidence that many women can benefit equally from single drug treatments. Clearly, if these same women get two drugs, their benefit erodes because of potential drug toxicity from over treatment. We need to find a better way to determine who should get one therapy or the other – or both."

O'Connell, who joined VCU in July 2002, led the team of scientists and clinicians at the Breast Care Center at Houston's Baylor College of Medicine and its affiliated Methodist Hospital when this study was done. The team used DNA microarray technology to study tumor response to the drug docetaxel, which is sold under the brand name Taxotere by the French drug company Aventis. The microarrays used in the study measure usage patterns of nearly every human gene and genetically "bar code" tumors based on tumor-to-tumor differences in gene expression, which is the process cells use to select which subset of their genes is made into proteins. Differences in gene expression can build a molecular profile of an individual cancer.

For this study, researchers prepared the bar codes by using breast tumor needle biopsies obtained from 24 patients before they began preoperative chemotherapy with docetaxel to shrink their tumors. Preoperative chemotherapy is an increasingly common practice in breast cancer treatment because patients whose tumors shrink substantially after pre-surgery chemotherapy generally can be treated with breast-conserving lumpectomies instead of mastectomies, and their chances for long-term survival are enhanced.

To measure the response of the tumor to docetaxel, the researchers took additional biopsies after each chemotherapy treatment – every three weeks for 12 weeks. The patients then had surgery to remove their tumors, and all received Adriamycin-Cytoxan chemotherapy after surgery to meet the current standard for breast cancer treatment.

O'Connell and colleagues found that consistent differences in expression of just 92 of the 12,625 genes tested predicted how tumors ultimately responded to Taxotere therapy. Sensitive tumors, for example, had higher expression of genes involved in stress response and programmed cell death, while resistant tumors showed reduced expression of those same genes but increased expression of protein metabolism and DNA synthesis genes.

The analysis correctly classified tumors as sensitive or resistant to Taxotere in about 90 percent of the cancers in the 24 patients in the study. It also correctly predicted Taxotere response for all patients in an independent group of six women with breast cancer who were studied to validate the predictive method.

"What we determined is that tumors have a 92-line bar code that predicts Taxotere response even before therapy starts and that this bar code identifies patients most likely to benefit from single agent Taxotere therapy," O'Connell said. "The genetic bar code is hard wired into the tumor before treatment, so doctors can rely on the original breast tumor for the answers about chemotherapy and other questions about optimal treatment strategies, regardless of whether or not the tumor has spread."

Taxotere is part of a new class of chemotherapy agents that interferes with tumor cell division rather than DNA synthesis. Taxotere has been found to be more effective at shrinking tumors and killing cancer cells, with fewer side effects, than the older Adriamycin-Cytoxan drug combination. But, because every person's cancer is slightly different, Taxotere doesn't work for everyone. Although the study looked at the effect of Taxotere on shrinking tumor size before surgery, O�Connell said that molecular profiling of the tumor is an equally useful guide if clinicians use more conventional post-surgery chemotherapy treatment.

Since his arrival at VCU, O'Connell has continued this study with surgeons, radiologists, pathologists and clinical oncologists affiliated with VCU�s Cancer Genomics Project and Massey Cancer Center. He hopes to expand the study to examine genetic predictors of response to both Taxotere and Adriamycin-Cytoxan chemotherapy in a larger group of patients.