Comprehensive genomic analysis of breast tumors revealed four main subclasses of breast cancer, each of which has extensive genetic heterogeneity, said an international consortium of scientists in The Cancer Genome Atlas Network that included those from Baylor College of Medicine in a report in the current issue of the journal Nature.

"In the last 10 years, new technologies have allowed us to look at these tumors at a variety of different levels," said Dr. Chad Creighton, an associate professor in the NCI-designated Dan L Duncan Cancer Center at BCM, who worked on the analysis of the various genomic datasets, which included, for all genes, the sequence, expression, copy, and any epigenetic silencing. Dr. Lawrence Donehower, professor of molecular virology and microbiology at BCM, concentrated on analysis involving a particular tumor suppressor gene called p53.

"We took all these data and used them together to identify these four basic subtypes of tumors that we've known about for some time," said Creighton. "Among other things, this study aimed to get at the fundamental question of just how many subtypes of breast cancer there are, because that can determine treatment."

One of largest effort to date

"This is one of the largest multi-institutional and multiplatform efforts of this kind to date," said Donehower. For this report, members of the consortium evaluated tumors and tissue samples from 825 patients with breast cancer. Of these, 510 tumors from 507 patients underwent sequencing of the whole exome (the protein-coding portion of the genome) and identified 30,626 somatic mutations. (Somatic refers to cells or tissues other than the egg or sperm that make up the germ cells of the body.) The 35 genes judged "significantly mutated" by one analysis included nearly all the genes known to be mutated in breast cancer and several novel ones.

They also performed a gene expression analysis, which included messenger RNA (which carries the blueprint for a protein) and microRNA (small single strands of genetic material that affect gene regulation), DNA methylation (the addition of a methyl molecule to a protein that usually turns a gene off), DNA copy number (alterations of the genome that can result in deletion of a gene or additional copies of gene) and a technique called reverse phase protein array that evaluates the activities of proteins.

Subtypes identified

When the results of these studies were taken together, the Cancer Genome Atlas consortium identified the subtypes as:

  • Luminal A breast cancers, which were the most heterogeneous or varied across all platforms.
  • Luminal B breast cancers that had a diversity of significantly mutated genes with TP53 and PIK3CA being the most frequent.
  • HER2-based tumors, which express the protein HER-2 that is targeted by the drug trastuzumab (Herceptin®).
  • Basal-like tumors or triple-negative breast cancers that lack most common drug target—estrogen receptor, progesterone receptor and HER-2.

Looking for new treatments

The study also noted that the different types do not seem to evolve from one another, said Creighton. For example, most luminal A breast cancers could not evolve into basal-like cancers, as luminal A cancers harbor a PIK3CA mutation that basal-like cancers lack.

Understanding the genes involved in the different subtypes can give researchers a target for developing new treatments.

"We wanted to understand what are the subtypes and what are the pathways that distinguish those subtypes," said Creighton.

A total of 91 institutions took part in the study. Funding for the research came from the National Institutes of Health and the National Cancer Institute, the Susan G. Komen Foundation for the Cure, the U.S. Department of Defense and the Breast Cancer Research Foundation.

For more information on basic research at Baylor College of Medicine, please go to