Perou Shares How Deeper Sequencing of Genome Helps Define, Treat Breast Cancer Subtypes

Charles M. Perou, PhD

For patients with early-stage invasive breast cancer and known estrogen receptor (ER), progesterone receptor (PR), and HER2 status, ASCO’s February 2016 release of their evidence-based recommendations¹ helped define appropriate use of biomarker assay results in guiding decisions on adjuvant treatment.

The recommendations state that beyond ER/PR and HER2 status, results from Oncotype DX, EndoPredict, PAM50, Breast Cancer Index, and urokinase plasminogen activator and plasminogen activator inhibitor type 1 have clinical utility in guiding treatment decisions on adjuvant therapy in subgroups of patients with breast cancer.

However, no biomarker test—except for ER/PR and HER2 status—was recommended for use to guide choices of specific drugs or treatment regimens.

The identification of intrinsic subtypes should be utilized to guide breast cancer treatment choices; however, more work needs to be done on sequencing the genome to determine what other biomarkers could carry significance in these patients, explains Charles M. Perou, PhD.

OncLive: How are lobular breast cancer subtypes being integrated into clinical arena?

In an interview with OncLive, Perou, May Goldman Shaw Distinguished Professor of Molecular Oncology, professor of Genetics, and Pathology and Laboratory Medicine, Lineberger Comprehensive Cancer Center, University of North Carolina School of Medicine, discusses how lobular breast cancer subtypes are used in clinical practice, what researchers have learned from sequencing the genome, and how determining intrinsic subtypes can help guide breast cancer therapy.Perou: There was a recent paper published from The Cancer Genome Atlas (TCGA) that was focused on studies in lobular breast cancers. From those studies, the main finding was a strong reinforcement that, on a molecular level, lobular breast cancer is really a distinct disease. It has a unique set of mutations, prominently including E-cadherin. However, another interesting finding included enrichment mutations in a gene called FOXA1.

Through gene expression profiling, we found 3 potential subtypes of lobular breast cancer, one of which showed a more proliferative phenotype and had a slightly worse prognosis than the other 2.

Can you go into detail with these expressive subtypes?

What are the immediate next steps for this project?

The other 2 are defined by other features in the microenvironment. In 1 case, it looked like there were immune-cell infiltrates. In another case, it looked like perhaps there were fibroblasts or other types of stromal-cell infiltrates—suggesting that there are important interactions between the tumor and the microenvironment. Perhaps we can target the immune infiltrate type with immunotherapy.We gave them relatively simple names. We called one the “proliferative” type, as it showed higher expression of cell cycle—regulated genes and a worse prognosis. We called the second the “immune” type to signify the presence of immune cells—it actually had macrophages, which was somewhat different than what we see in ductal cancers. Finally, the “stromal” type has fibroblasts and other types of mesenchymal cells.If possible, the next steps are to look at existing studies with the new lens and see if it has any therapeutic differences. We now have some tools to go in and look at earlier studies and link to some of these new findings.

The Breast International Group (BIG) 1-98 study, for example, was a very large aromatase inhibitor versus tamoxifen study. There were enough lobular breast cancers where they could look at the 2 arms between the lobular cancers and see a difference. That would be a great study to go back and look at these expression subtypes—to look at the FOXA1 mutations and see if the mutants behave differently than the wild-type cases.

What are some unanswered questions with lobular breast cancers?

FOXA1 is a cofactor with ER and it is required for, in some way, full ER activity. Mutations and a cofactor of ER certainly hint that there are going to be some differences in potential endocrine responses.Is there a difference in responsiveness of aromatase inhibitors versus tamoxifen? A number of lobular cancers and their chemotherapy responsiveness is another place to look. These are maybe not as chemotherapy responsive in general as ER-positive cancers, which are also maybe not that chemotherapy responsive.

What has sequencing the genome taught us?

Additionally, are immune checkpoint inhibitors going to benefit some or a large portion of lobular cancers? That is a very important question we need to address soon.Sequencing the genome has taught us, in one respect, a huge amount. From a clinical impact, I would say, “Not as much as we would have hoped.”

By sequencing the genome to invoke TCGA’s project, we have done 1100 breast cancers. We really know all of the genes that are mutated at a relatively high frequency of 3% or more, and we know all of the copy changes that occur.

We also know a great deal about the methylation profile, so we now have the parts list. This is the landscape that we have to deal with. It may change in the metastatic setting, and we have additional studies ongoing and in the future about how metastatic settings may be different than earlier settings.

We now can find a number of mutations that may be biomarkers for responsiveness for a number of drugs. Our immediate task is to run clinical trials to answer, on the gene-by-gene basis, whether a given gene is actually a biomarker, as opposed to saying, “I’m going to put 100 genes together in a panel, and I’m just going to apply the panel.”

We may not learn my hypothesis against the panel. There are probably 10 to 20 genes that are of clinical value and 80 that are not. Let’s figure out which of the 10 to 20 are valuable, and we can ignore the 80 that are not. In the biology world, we want to look at them all.

How can community oncologists use this logic in clinical practice?

We really need to figure out which ones are clinically helpful. We are figuring that out, and the list of helpful ones is growing. PI3K mutations may be a biomarker of sensitivity to drugs targeting PI3K. If that holds up, it will be great because 40% of patients with breast cancer have this mutation.I hope they learn more about when it is appropriate to use gene panel testing for somatic mutations to guide treatment choices, and maybe when it’s not appropriate. When they actually do use it and get the results, they should understand that it is still a very hard interpretation of what the results are saying.

Should intrinsic subtypes guide cancer therapy?

For most of these genes for breast cancer, many are still speculative and not validated biomarkers. The only DNA-based approved assay in breast cancer is HER2 gene amplification. All of these other assays are still in the “figuring it out” stage. We still need to be cautious.This really relates to the new ASCO biomarker guidelines. These guidelines are very helpful because they look at a large number of gene expression tests and determine how they should be used, and what the level of evidence was about each test. It has really provided some help for the field because there are 8 or 9 of these tests with varying levels of evidence for particular uses. The expert panel did a good job of trying to mesh this all together in a single document.

It seemed that none of the tests were approved for use in HER2-positive patients. Why?

I am a developer of the PAM50 assay, which had good evidence to be used to make adjuvant systemic therapy decisions in the node-negative setting.None of the tests were approved for use in HER2-positive patients or patients with triple-negative breast cancer (TNBC). The ones that were, in essence, approved for use were in HR-positive/HER2-negative patients, which is still the majority of patients.

Is it worth having another test developed for HER2 or TNBC?

However, it is important to point out that, aside from HER2 testing, there really is no other approved biomarker. For TNBC, these tests don’t really add to predicting chemotherapy response or prognosis for these patients. The guidelines did a good job of saying, “Here is where they can be used and for what, and here is where we don’t recommend them to be used.”It would be helpful to have genomic or genetic tests that may further stratify HER2 or TNBC patients. My laboratory is working to develop such tests, as are others. However, if you look at TNBC by the PAM50 risk of recurrence score, they are all pretty much called high risk. There is very little prognostication within TNBC for the existing tests.

If someone says to a patient, “You’re triple-negative,” you really know right away that they are highly likely to be high risk. If you take it a step further by extrapolating to chemotherapy, then they are going to be recommended to get chemotherapy, which is already the indication for TNBC.

You can see that, on one hand, the test is saying the right thing; however, on the other hand, they are also recapitulating what you already know and do. If we can find a test that found really good outcome triple-negative patients who had such a good prognosis that they didn’t need chemotherapy, that would be valuable; however, none of the tests do that right now. In the HER2-positive world, it’s similar; that is why the guidelines are as they are.

1. Harris LN, Ismaila N, McShane LM, et al. Use of biomarkers to guide decisions on adjuvant systemic therapy for women with early-stage invasive breast cancer: American Society of Clinical Oncology Clinical Practice Guideline. J Clin Oncol. 2016;34(10):1134-1150.