The work represents a major step toward truly individualized breast cancer prevention in women at high risk for the disease based on their age, family history of breast cancer, and personal medical history.

"Our study reveals the first known genetic factors that can help predict which high-risk women should be offered breast cancer prevention treatment and which women should be spared any unnecessary expense and risk from taking these medications," said the study's lead scientist, James N. Ingle, M.D., professor of oncology at the Mayo Clinic in Rochester, Minn. "We also discovered new information about how the drugs tamoxifen and raloxifene work to prevent breast cancer."

Ingle and Mayo-based colleagues in the NIH Pharmacogenomics Research Network (PGRN) conducted the study in collaboration with PGRN-affiliated researchers at the RIKEN Center for Genomic Medicine in Tokyo. Data and patient DNA came from the long-running National Surgical Adjuvant Breast and Bowel Project (NSABP), supported by the National Cancer Institute.

"This innovative, PGRN-enabled international research partnership has produced the first gene-based method to identify which women are likely to benefit from a readily available preventive therapy," said PGRN director Rochelle Long, Ph.D., of the NIH's National Institute of General Medical Sciences. "Because the disease affects so many women worldwide, this work will have a significant impact."

The research, which shows nearly a six-fold difference in disease risk depending on a woman's genetic makeup, appears in the June 13, 2013, issue of Cancer Discovery.

Women undergoing breast cancer preventive treatment take tamoxifen or raloxifene for five years. In rare cases, the drugs can cause dangerous side effects, including blood clots, strokes and endometrial cancer.

Many women never try the therapy because the chance of success seems small (about 50 women in the NSABP trials needed to be treated to prevent one case of breast cancer) compared to the perceived risk of side effects. More women might benefit from the potentially life-saving strategy if doctors could better predict whether the therapy was highly likely to work. That's what the current study begins to do.

The investigators leveraged data from past NSABP breast cancer prevention trials that involved a total of more than 33,000 high-risk women -- the largest sets of such data in the world. Women in the trials gave scientists permission to use their genomic and other information for research purposes.

The scientists analyzed the genomic data by focusing on more than 500,000 genetic markers called single nucleotide polymorphisms (SNPs). Each SNP represents a single variation in the DNA sequence at a particular location within the genome.

To determine whether any SNPs were associated with breast cancer risk, the researchers computationally searched for SNPs that occurred more commonly in women who developed breast cancer during the trial than in women who remained free of the disease. The analysis identified two such SNPs -- one in a gene called ZNF423 and the other near a gene called CTSO.

Neither ZNF423 nor CTSO -- nor any SNPs related to these genes -- had previously been associated with breast cancer or response to the preventive drugs. The scientists' work revealed that women with the beneficial version of the two SNPs were 5.71 times less likely to develop breast cancer while taking preventive drugs than were women with neither advantageous SNP.

Using a variety of biochemical studies, the scientists learned that ZNF423 and CTSO act by affecting the activity of BRCA1, a known breast cancer risk gene. Healthy versions of BRCA1 reduce disease by repairing a serious form of genetic damage. Harmful versions of BRCA1 dramatically increase a woman's chance of developing breast cancer.

"The results of our collaborative research bring us a major step toward the goal of truly individualized prevention of breast cancer," said Ingle. "Our findings also underscore the value of studying the influence of gene variations on drug responses."

NIH's National Institute of General Medical Sciences and National Cancer Institute funded this research through grants U19GM61388, P50CA116201, U10CA37377, U10CA69974, U24CA114732 and U01GM63173.