HOUSTON – (Aug. 9, 2016) – Breast cancer tumors are complex and dynamic. They comprise a population of continuously dividing cells that carry different genetic mutations. On a paper published today in Nature Communications, researchers from Baylor College of Medicine, Washington University School of Medicine, MD Anderson Cancer Center and the Mayo Clinic reveal, for the first time, that treating human estrogen-receptor positive (ER-positive) breast cancer tumors with estrogen-deprivation therapy results in changes in the spectrum of mutations in the tumor population, and point towards the possibility of using this information to improve cancer treatment.
"The majority of ER-positive breast cancers are not a single tumor but more like a family of related tumors referred to as 'sub-clones,'" said senior author Dr. Matthew Ellis, professor and director of the Lester and Sue Smith Breast Center at Baylor. "The tumors are like a large family. A family has the same genetic origin – the same parents – but each family member has distinct genetic characteristics. The brothers and sisters are clearly individuals but they are also related. When we treat the tumor with aromatase inhibitors, an estrogen-deprivation therapy that lowers the levels of estrogen the tumor needs to grow, we are creating a situation where certain members in the tumor family are able to persist and grow while others perish. The surviving members of the tumor family are likely the ones that will cause future problems with recurrence."
Although researchers have extensively studied the genetic heterogeneity in breast cancer in untreated samples, they knew little about how aromatase inhibitors, such as letrozole, anastrozole and exemestane, affect the genetic diversity of the tumor.
"In this study, we present a first answer to this question by studying 22 human breast cancer tumors scheduled for surgery," said Ellis. "To reduce tumor size before surgery, we treated the tumors with estrogen-deprivation therapy for four months. The tumors were then surgically removed. We analyzed in great detail the effect of estrogen-deprivation therapy on the gene mutation patterns of the tumors by studying the entire genomic structure – the whole genome – of each of the tumors on biopsies taken before and after estrogen-deprivation therapy."
"In the post-treatment samples, we found many new mutations or enrichment of mutations present at low levels in the pre-treatment samples," said Ellis. "This means that under the environmental stress of the treatment, the tumors are spawning new sub-clones which subsequently can survive and grow despite therapy, and that is why we are having difficulty treating ER-positive breast cancer. We found this result for a majority of ER-positive breast cancers we studied."
A majority of the breast cancer tumors the researchers studied comprised a number of sub-clones with a common origin – they were all members of the same tumor family. But, the researchers also discovered that some patients had more than one tumor of different origin.
"Even though each patient in this study was diagnosed only with a single tumor, looking at the cancer genome allowed us to see that in some cases the patient actually had two separate tumors growing closely together. We call these "collision tumors," said first author Dr. Christopher Miller, research faculty at the McDonnell Genome Institute at Washington University in St Louis.
In these cases, the two separate tumors were like "two unrelated families growing so close together they were originally incorrectly identified as a single family," said Ellis.
Undiagnosed collision tumors could explain why sometimes tumors with an initial good prognosis have an unexpected relapse after surgical treatment.
In one of the 22 tumors, the researchers discovered ER-negative tumor cells that were hiding inside a mostly ER-positive tumor. "By the end of four months of therapy – because the treatment had shrunk the ER-positive tumor – we could detect this second ER-negative tumor, and treat it accordingly to its nature before it grew larger," said Ellis. "Without this approach, that ER-negative tumor would have never been diagnosed early and treated."
"If a patient with breast cancer has the tumor surgically removed, it won't be possible to detect the cells with the genetic makeup most likely to be driving relapse," said Ellis. "But, if, on the other hand, we start by treating the tumor with aromatase inhibitors before surgery for a few months, so we can track the behavior of that tumor, we would get a more complete picture of the cancer. We can potentially detect sub-clones that can cause relapse in the future."
"Our results suggest that studying the genetic makeup of a tumor at diagnosis is not enough – periodically scanning the genome in several biopsy samples to understand how it is changing may help us evolve treatment strategies to match," said Miller.
"The results emphasize the importance of proper trial design coupled with sample banking and annotation within clinical trials toward ultimately arriving at a better understanding of the disease and its treatment," said senior author Dr. Elaine R. Mardis, Robert E. and Louise F. Dunn distinguished professor of medicine and co-director of the McDonnell Genome Institute at Washington University School of Medicine.
Treatment with aromatase inhibitors before surgery can have advantages for patients. "Because the treatment shrinks the tumor, patients are more likely to have breast-conserving surgery," said Ellis.
Other contributors to this work include Yevgeniy Gindin, Charles Lu, Obi L. Griffith, Malachi Griffith, Dong Shen, Jeremy Hoog, Tiandao Li, David E. Larson, Mark Watson, Sherri R. Davies, Kelly Hunt, Vera Suman, Jacqueline Snider, Thomas Walsh, Graham A. Colditz, Katherine DeSchryver and Richard K. Wilson from Baylor College of Medicine, Washington University School of Medicine, MD Anderson Cancer Center and Mayo Clinic.
This work was supported by the National Cancer Institute of the National Institutes of Health under award numbers U10CA180821 and U10CA180882, and the following grants: 5U10CA180833 and 5U10CA180858. This research was also supported in part by grants U54HG003079 from the National Human Genome Research Institute, R01-CA095614, U24-CA114736, U10-CA076001, and U01-CA114722 from the National Cancer Institute; by the Breast Cancer Research Foundation; Komen Promise Grant PG12220321, a Komen St Louis Affiliate Clinical Trials Grant; and support for Z1031 from Pfizer and Novartis; the McNair Medical Foundation Scholar program and the Cancer Prevention Research Institute of Texas established investigator award.