A Closer Look
New Tools Enhance Molecular Portraits of Breast Cancers
The HER 2 protein (red) can trigger cell growth responses in breast cancer cells (nuclei stained blue).
Using a combination of analytical tools, investigators with The Cancer Genome Atlas (TCGA) Research Network have completed a molecular study of breast tumors from 825 women. The results, recently reported in Nature, confirm the existence of four major subtypes of breast cancer and add new details about the biological changes underlying these diseases.
The researchers used up to six different technologies to characterize subsets of the tumors. In addition to sequencing DNA and RNA, the investigators profiled patterns of DNA methylation and counted the number of copies of genes in tumors. This was also the first TCGA study to report protein expression patterns in tumor samples.
The integration of these results has given researchers a catalog of the genetic and epigenetic abnormalities in each subtype of breast cancer, underscoring the idea that these tumors are, in many respects, distinct diseases.
"This paper and five others [describing breast cancer genomes] published this year in Nature provide a new roadmap for translational and basic research on breast cancer," said co-lead investigator Dr. Matthew Ellis of the Washington University School of Medicine in St. Louis. Researchers could spend a decade following up on these results, he added. (See the sidebar for links to the study abstracts.)
Previous studies had hinted that one of the subtypes, basal-like breast cancer, was genetically similar to a form of ovarian cancer. The TCGA study confirmed this idea and suggested that treatments currently being tested for some ovarian cancers could be tested against these breast cancers.
"This finding really stood out," said Dr. Ellis. "And it led to discussions [among the study authors] about the most appropriate types of chemotherapy for patients with breast cancer." The other subtypes are known as luminal A, luminal B, and HER2-enriched breast cancers.
Making Use of Multiple Technologies
Speaking at a press briefing on cancer research last week, NCI Director Dr. Harold Varmus acknowledged that the four breast cancer subtypes have been known for years. What's new, he explained, is that, for each subtype, TCGA investigators used multiple technologies to describe the "landscape of genetic abnormalities" in greater detail than in the past.
"We haven't had a storehouse of so much valuable information about each of these categories of cancer, with the same tumors analyzed for a wide variety of properties," he said. "It's the repository that is so important."
Because the study included hundreds of tumors, the researchers were able to detect uncommon but recurring mutations. Some of these mutations indicated that the tumors might respond to existing drugs. "Repurposing drugs will be important for treating this disease," said Dr. Ellis.
Even if a particular mutation occurs in only 2 percent of patients, Dr. Ellis continued, breast cancer is common enough that researchers should be able to enroll enough women in clinical trials to test existing drugs that target these mutations.
About 20 percent of the patients with basal-like tumors might be candidates for drugs known as PARP inhibitors based on analyses of the genes BRCA1 and BRCA2 in their tumors, the researchers said. The group of basal-like tumors includes triple-negative breast cancers, which are difficult to treat and disproportionately affect younger women and African Americans.
The Translation Phase
Only three genes—TP53, PIK3CA, and GATA3—were mutated in more than 10 percent of the patients' tumors. Drugs that target changes resulting from defects in PIK3CA are in development and could be tested in selected patients with breast cancer. However, designing and implementing large clinical trials can take years, the researchers cautioned.
"People always want to know when this kind of research is going to affect clinical care," said Dr. Charles Perou of the Lineberger Comprehensive Cancer Center at the University of North Carolina, another study leader. "Now that we've made these discoveries, we're in the translation phase."
Many of the new discoveries can now be tested in the context of clinical trials. For instance, the study suggested there may be at least two groups of patients with HER2-positive tumors, and these groups may have different responses to treatment.
"We had a hint of this from past gene-expression studies," said Dr. Perou. But the integrated results of the TCGA analysis, which included proteomics, are "far more convincing and suggestive than results based on any one technology alone."
Dr. Perou co-authored one of the first studies to use genomics to distinguish subtypes of cancer. The study, published in 2000, used what was then a new tool—DNA microarrays—to profile the expression of 8,000 genes in breast tumors from 42 women.
More than a decade later, the technological advances in genomics have been "astonishing," noted Dr. Ellis. The missing component right now is information about proteins and the biochemistry of cancer cells, he observed.
"Over the next 10 years, we need to study proteins in the same way that we have just studied DNA and RNA over the last decade," said Dr. Ellis. Only then, he added, "will we develop a complete picture of the biochemistry of cancer cells."
—Edward R. Winstead