SAN ANTONIO (GenomeWeb) – Researchers at the Dana-Farber Cancer Institute and the Broad Institute are combining insights from exome sequencing and transcriptome profiles of metastatic breast cancer patients to gain a better understanding of how the disease progresses and the mechanisms cancer cells use to evade treatment.
Ofir Cohen, a computational biologist at the Broad, presented the first data from this effort at the San Antonio Breast Cancer Symposium yesterday, showing that metastatic, treatment-resistant, estrogen receptor-positive breast cancers have mutational and transcription characteristics not present in primary tumors. By mapping these changes, researchers are hoping to inform treatment choice, particularly combination therapy, guide clinical trial enrollment, and identify new drug targets in metastatic breast cancer — an understudied setting compared to primary tumors due to the challenge of acquiring biopsies.
"Even if you know the alterations and mutations in the primary setting, that may not be sufficient to guide treatment at a later time point in the metastatic setting," Cohen said. Moreover, just looking at the genomic aspects of the metastatic tumor may not tell the entire story, Nikhil Wagle, deputy director of DFCI's Center for Cancer Precision Medicine and another lead researcher, noted at the conference.
The study comes at a time when major cancer institutes around the country are profiling cancer patients so they can receive precision treatment or drugs that target the molecular drivers of their disease. But as some experts at the conference pointed out, within these programs, as few as 5 percent of patients actually match to a therapy, because even when the tumor drivers are known, there may not be drugs to target them.
The aim of their study wasn't to promote precision medicine per se, Wagle told GenomeWeb, but to improve understanding of the complex biology of metastatic breast cancer and resistance — the kinds of discovery work without which precision medicine cannot advance.
Understanding metastases and resistance
The most common type of breast cancers are those that have a large number of estrogen receptors on cancer cells. Women with these types of tumors can receive hormonal therapy and have better five-year survival rates than patients with little or no estrogen receptors on their cancer cells. However, earlier this year, longitudinal analysis of the International Breast Cancer Study Group clinical trials data showed that estrogen receptor-positive breast cancer patients were at higher risk of recurrence than estrogen receptor-negative patients beyond the first five years after treatment of the primary tumor and through 24 years.
Although the most common cause of death for metastatic breast cancer patients is treatment resistance, the biological mechanisms are not well understood, and there is a need to develop new drugs for these patients. However, recent studies suggest that the genomic features of advanced breast cancer can be substantially different from primary tumors, Cohen said at the meeting, highlighting that one study has shown that ESR1 mutations occurred in 25 percent of metastatic tumors versus 1 percent of primary tumors.
Hoping to elucidate other molecular features of this kind, Cohen and his team collected 149 tumor samples from treatment-resistant, estrogen receptor-positive, metastatic breast cancer patients at DFCI, as well as primary tumor samples from 44 of these patients. More than 95 percent of patients received at least one estrogen receptor-directed treatment before biopsy of the metastatic lesion.
Cohen's team conducted exome sequencing on all 149 samples using Illumina's HiSeq at 185x mean target coverage and 120x minimum coverage. They conducted transcriptome analysis using Illumina's TruSeq paired-end protocol followed by HiSeq on 128 metastatic biopsies at a mean depth of more than 35 million mapped reads.
Exome sequencing revealed several genes that are similarly mutated in primary and metastatic disease, such as TP53, GATA3, and PIK3CA. However, there were a number of genes that acquired single nucleotide variations and indels in the metastatic setting that researchers believe could be driving metastasis, such as CDH1, SMARCAD1, and ARID1A, while alterations in ESR1, ERBB2, AKT1, KRAS, and RB1 may be mechanisms of resistance to endocrine treatment. Researchers also found a number of amplified and deleted genomic segments that might be clinically significant in the metastatic setting.
Comparing primary and metastatic tumor samples when available, Cohen's group tried to capture how tumors evolved from the earlier to the advanced setting. At diagnosis, 77 percent of patients were HER2 negative, 9 percent were HER2 positive and HER2 status for another 9 percent were unknown. In the metastatic setting, six patients had become estrogen receptor-negative, seven had become HER2 negative, and five gained HER2 expression.
Researchers identified ESR1 mutations in 15 patients, and 14 of them acquired these mutations from the metastatic tumor. This data emphasizes, Cohen said, the important role that ESR1 mutations play "under the selection pressure of endocrine therapy in the metastatic setting" and may suggest the use of estrogen degrading treatment over aromatase inhibitors.
Similarly, five out of six patients with ERBB2 mutations acquired them in the metastatic setting, "suggesting resistance to ER-directed therapy," Cohen added, saying these patients may benefit from enrollment in trials of HER2-targeted treatment.
Additionally, three out of five patients had acquired RB1 mutations in the metastatic setting. "The mutational status of RB1 is very important in therapeutic choice since inactivation may predict resistance to CDK4/6 inhibitors," he said.
Wagle highlighted the case of one patient who, when first diagnosed with metastatic breast cancer in 2011, had estrogen hormone receptor-positive, HER2-negative disease. In 2015, when her gluteal mass was biopsied and tested as part of the present study, she had triple-negative breast cancer and alterations she previously didn't have in ESR1, RB1, and PTEN.
Cohen and his team are also integrating patients' transcriptome data with mutational information. As an example, he showed how his team had mapped transcriptional signatures of cell cycle type and estrogen on top of the landscape of ESR1 and RB1 alterations.
"While the phenotypic and functional consequences of mutations may be hard to assess in the clinical setting, additionally sequencing the transcriptome may provide a useful handle," he said. "The association between the mutational status and the transcriptional signatures suggests the functional consequences of these mutations in altering the tumor cell state."
Since cancer cells may also acquire resistance though epigenetic and regulatory mechanisms, these mechanisms may "leave their footprint on the transcriptome," he noted.
One challenge in this analysis is distinguishing which molecular events predispose a tumor to metastases and which lead to treatment resistance, since most of the samples used in the study were from metastatic patients who were also treatment resistant. Cohen's group is also performing functional assays to pinpoint the precise role of the identified events.
Growing focus on metastatic breast cancer
The study at DFCI and the Broad is part of a growing effort in the breast cancer research community to improve the understanding of metastatic disease using sequencing approaches.
At this same meeting, another group led by researchers from the University of Pittsburgh evaluated the genomic profiles of 20 breast cancer patients who developed brain metastasis, and found that 20 percent of those with HER2-negative primary tumors became HER2-positive when the disease spread to the brain.
Although there were transcriptional similarities between the primary and matched metastatic samples, researchers led by the University of Pittsburgh's Adrian Lee reported in the Journal of the American Medical Association that 100 genes were recurrently altered and 17 patients had differential expression of clinically actionable genes in their brain lesions compared to their primary tumors.
Based on this study, it's not yet known the degree of benefit metastatic patients who switched to HER2-positive status might derive from HER2-targeted treatment. "At the least, further studies are immediately warranted as we may be missing treatment opportunities in advanced cancer settings," Nolan Priedigkeit, the study's lead author, told GenomeWeb.
Brain metastasis is a common site of progression in breast cancer. However, obtaining a biopsy is risky, making it challenging to track the evolution of the metastatic disease, and so, clinical decisions are made based on molecular features of the primary tumor, explained Priedigkeit, an MD/PhD candidate at the University of Pittsburgh and Carnegie Mellon University Medical Scientist Training Program. However, one option in the future, may be to do blood-based circulating tumor DNA testing as a surrogate marker for what's happening within the tumor, he added.
In Cohen and Wagle's study, they are tracking patients' cell-free DNA with blood biopsies every two to three months, in addition to collecting more than 250 biopsy samples. In that study, researchers simultaneously perform a targeted panel in a CLIA lab, so that certain molecular information about patients' tumors can be returned to them via their doctors.
"There are definitely many examples of doctors using these results for clinical decision making, like enrolling patients on trials," Wagle told GenomeWeb. "We cannot report the specific responses, since those are part of clinical trials." He added though that patients are responding to clinical decisions that have been made, and there are efforts to try to quantify this.
One criticism of precision oncology approaches has been that sequencing identifies drivers of disease but often those drivers are not targetable by existing treatments. At the conference, Philippe Bedard from the Princess Margaret Cancer Center highlighted recent editorials that argue that despite the enthusiasm for precision oncology, the approach hasn't been proven in randomized-controlled trials and therefore shouldn't be broadly applied.
Wagle, a proponent of precision medicine, countered that there is also no evidence that the concept is invalid, and critics are ignoring the successful examples that are now considered standard treatment.
"The criticism of whatever's left is that nothing else works, so why are you testing all these other genes? The reason is that we don't yet have enough information," he said. "The reason we don't have more targets in precision medicine is we haven't learned enough and we haven't developed the right therapeutic strategies to target those alterations."
That's precisely why his group is doing the exome and transcriptome sequencing study. "I think precision medicine will ultimately yield good responses," he said. "But we have to make the discoveries and understand how to apply it, before we can say it's not going to work."