Whole-Genome Study of ER+ Breast Tumors: Treatment Clues and Complexity

A team from the Washington University School of Medicine and the Siteman Cancer Center, St. Louis, has sequenced whole genomes of tumors from 50 women with luminal (ER and/or PR positive) HER2 negative breast cancer. The researchers compared the sequences to matched DNA from the same patients’ healthy cells, to identify mutations present only in the cancer cells. Because endocrine therapy–resistant HER2 negative breast cancer is clinically aggressive and is the most common cause of breast cancer–related death, the researchers also sought to identify specific mutations linked to treatment outcomes.

The study represents one of the largest cancer genomics investigations reported, with more than 10 trillion chemical bases of DNA sequenced. While its results underscore the complexity of breast cancer biology, the mutations uncovered may provide further clues to inform personalized therapy of this common breast cancer subtype.

Lead investigator Matthew J. Ellis, MD, PhD, presented the findings (abstract LB-87) at the AACR 102 Annual Meeting, which is being held in Orlando, Florida, from April 2 through 6. The tumor samples analyzed were from two trials assessing neoadjuvant endocrine therapy (POL and ACOSOG Z1031). All of the patients had ER-positive tumors. The cell-proliferation marker Ki67 was used to classify the surgical samples; 24 samples were endocrine therapy–resistant by Ki67 testing (Ki67 ≥ 10%), and 26 were sensitive (Ki67 ≤ 10%).

Using a supercomputer to conduct massively parallel DNA sequencing, DNA samples from the tumors were sequenced and analyzed in collaboration with Washington University’s Genome Institute. To ensure accuracy of the data, sequencing of tumor DNA and healthy DNA was repeated about 30 times for each patient.

There were more than 1,700 mutations in the tumors altogether. Most of the mutations were unique to individual patients, and involved single-nucleotide variations, frame shifts, translocations, and deletions. Similar to other published clinical studies, Dr. Ellis and his colleagues found two common mutations in many of the patients’ cancers: in extension analysis of 121 samples from the POL/Z1031 studies, PIK3CA mutations were present in 43% and TP53 mutations were found in about 15%. In addition, mutation of tumor-suppressor gene MAP3K1 was found in about 9% of the luminal breast cancer samples (compared with only one MAP3K1 mutation found in 60 samples of basal-like [ER/PR/HER2 negative] breast cancer). The genes ATR and MYST3 were mutated at about the same frequency as MAP3K1 in the samples, a statistically significant level. Other significantly mutated genes included RUNX1, PRSS8, and ZNHIT2.

Mutations in MAP3K1, especially in combination with PIK3CA mutations, was preliminarily associated with tumors that responded to aromatase-inhibitor therapy.

A total of 21 significantly mutated genes were found in the samples, most at frequencies well below 10%. Given that breast cancer is a common disease occurring in thousands of women, however, even mutation frequencies of 5% may be clinically important in terms of identifying potentially appropriate therapies.  An understanding of the clinical significance of these rarer events will require the analysis of thousands of specimens with clinical annotation and analysis of further samples is underway.

The investigators categorized the identified mutations into a variety of cancer pathways, including phosphoinositol-3-kinase signaling, cell cycle regulation, metabolism, mitotic spindle regulation, and DNA mismatch repair. In conclusion, they said their analyses “suggest a classification of the disease that is not based on individual gene abnormalities but on common cancer pathway events that ultimately determine outcome for patients with luminal-type disease.”

In discussing the study, Dr. Ellis commented that “while analyses of whole genome expression allow us to further categorize breast tumors into subtypes of clinical importance, we still do not have a good understanding of the cancer biology behind many of the identified gene sets.” Also, he said, a variety of gene expression profiles can be associated with the same treatment outcome, for example, response to adjuvant endocrine therapy. “The availability of massive parallel sequencing adds another layer of complexity to identifying the genes of greatest relevance to outcome,” he said. “Yet as the information available to us expands exponentially, the ultimate goal remains the same: to identify driver gene aberrations that we can target for therapy so that outcomes for our patients with breast cancer can be improved.”