I believe that an in-depth examination of outlier responses will become a standard component of the way we develop drugs. We will continue to be disappointed if a drug fails to achieve its primary endpoint and proves ineffective in most patients. We should, however, make sure we do not miss opportunities to continue to develop drugs that have profound activity in a subset of patients.
—David B. Solit, MD
In 2012, David B. Solit, MD, Geoffrey Beene Chair and Director of the Center for Molecular Oncology at Memorial Sloan Kettering Cancer Center (MSKCC) in New York, and his colleagues published the results of a phase II study1 of 45 patients with advanced bladder cancer. The purpose of the clinical trial was to evaluate the effectiveness of everolimus (Afinitor) in delaying disease progression.
All patients in the study received the drug, but so few responded, the drug was deemed not worthy of further development in this disease. However, one patient—a 73-year-old woman whose metastatic bladder cancer had become resistant to standard treatment—had such an impressive response to everolimus (in just 3 months her cancer had disappeared) that Dr. Solit and his research team initiated an in-depth investigation to determine the genetic basis for her extreme, outlier sensitivity to the drug.
They performed whole-genome sequencing on a sample of the patient’s tumor and germline DNA derived from blood. This analysis revealed approximately17,000 mutations in the tumor that were not found in her normal cells. Upon analyzing these 17,000 potentially pathogenic mutations, aberrations in two genes, TSC1 and NF2, stood out as the likely basis for her exceptional response to everolimus. Four years later, the patient is doing well, without evidence of disease recurrence and remains on everolimus therapy.
The ASCO Post talked with Dr. Solit about how advances in whole-genome sequencing methodology are being used to define the molecular basis of other exceptional responses to cancer therapy and how this knowledge is leading to changes in clinical trial design and biomarker development.
Finding Key Mutations
Please talk about the results of your study of everolimus in patients with bladder cancer.
One thing we’ve always recognized in oncology is the heterogeneity of outcomes among our patients. Some patients do well with surgery or chemotherapy or other systemic agents, whereas some patients do poorly. For years, we’ve tried to figure out why there is that variability in outcomes. This is important to know because for many patients we may already have an effective drug available for them, but it may be a drug that only works in a minority of patients with a specific tumor type. If we do not have a predictive biomarker to tell us who is going to respond to a particular drug, the drug may not get developed to its full potential.
We hypothesized that if we could figure out why this one patient in our everolimus study was so unique in her response to the drug, then maybe we could identify other patients for whom everolimus would also be of utility.
We initially sequenced some of the genes we thought might have been the basis for her response based upon their known role in regulating mTOR signaling and prior reports showing that they were mutated in some bladder cancers. But using this candidate approach, we were unable to identify the basis for our patient’s extraordinary response. Then, taking advantage of recent advances in sequencing technology that weren’t available in years past, we were able to sequence the patient’s entire tumor genome.
It turned out that she had two mutations that jumped out as the likely basis for her response. One was a two-base pair deletion in TSC1, which results in a truncation of the protein and loss of expression. And the other was a nonsense mutation in NF2, which also results in early truncation of the protein and loss of expression. These two genes are interesting because the loss of both of the proteins leads to activation of mTORC1, which is the direct target of everolimus.
We were then able to study the tumors of other patients with bladder cancer and ask, do these mutations occur commonly in this disease? We found that the answer is no. Rather, they are only found in a minority of patients. In the case of TSC1, somewhere between 6% and 8% of patients with bladder cancer have mutations in this gene. NF2 is very rare; it probably occurs in just 1 in 100 or so patients with bladder cancer.
How will you now use this information to treat your patients with bladder cancer?
We have proposed a second study where we look for TSC1 and NF2 mutations prospectively and then select those patients for enrollment in a trial of everolimus. We don’t necessarily assume that all the patients are going to respond as well as our index patient did in the first trial. That patient had two mutations that likely cooperated to promote the formation of her tumor conferring extreme sensitivity to the drug. Other patients may have, for example, a TSC1 mutation and a second mutation that makes them less sensitive to everolimus.
Will the whole-genome sequencing of your patients’ tumor become standard of care before enrollment into a clinical trial?
We are starting a program in which we will broadly begin profiling patients with bladder cancer for TSC1 and other alterations. The assay that we are going to use initially tests for about 340 genes that have a strong link to cancer formation. We are planning on profiling not only all patients with bladder cancer, but all patients who have advanced cancer and need a new treatment option. We currently have several trials open at MSKCC that target this pathway for which they may be eligible.
Are you also studying patients with other cancer types who have had exceptional responses to a drug?
Yes, we have completed retrospective studies on several additional cases. For example, we studied a patient with ovarian cancer who has had a nearly 4-year response to a MEK inhibitor. We found that the patient had a mutation in MAP2K1, which encodes MEK1, which was very difficult to detect using older methodologies. The incredible outlier response in this patient was thus due to a mutation in the target of the drug that sensitized her tumor to the MEK inhibitor. We just didn’t know that information when the trial was performed. We can now, however, start looking for patients prospectively that have similar mutations.
Will you eventually conduct prospective studies on all outlier patients?
Yes. We have now conducted such studies in a number of cancers, and we are starting to see other groups begin to study these exceptional responders. I think there is interest in repeating this paradigm very broadly to ensure that we do not miss opportunities to determine why an individual patient is responding to a specific drug.
In the past, we had to look at one gene at a time, but with newer technologies, we can sequence the whole genome, or at least the whole exome. Using these techniques, I expect that we can determine in almost all such patients why they are unique in their response to a particular therapy. Of course, some complexity remains, in that some patients may have so many mutations and some drugs may hit multiple different targets, making it difficult even with whole-genome analysis to pinpoint the genomic basis for an individual patient’s drug sensitivity. Despite these challenges, our success rate has been very high with this paradigm.
Because each patient’s tumor is molecularly specific, aren’t all patients considered “outliers”?
What you are saying is not completely untrue, but I believe that there are definitely patterns. For example, one of the things that came out of our study was that TSC1 is probably most commonly somatically mutated in bladder cancer (of the major tumor types), and that really was not appreciated previously. This highlighted that we should at least be considering the TSC1 mutant group of tumors as a subset of bladder cancer, for which we could develop effective combination therapies for most patients.
While everolimus as a single agent may not be an effective therapy for many patients, by defining the pattern of mutations that co-occur with TSCI mutation, we can, hopefully, develop a combination of therapies for those in whom a single-agent approach is ineffective.
How might the study of exceptional responders to cancer therapy change drug development in the future?
I believe that most pharmaceutical companies are still going to focus on broadly active drugs that work in large populations. I don’t think that’s going to change dramatically, but if we can make it easier to identify smaller populations that exhibit exceptional responses, there will be an interest in developing drugs for this narrower group of patients.
Furthermore, if the larger pharmaceutical companies are not interested in these rare cancer subtypes, I can see smaller companies or the National Institutes of Health (NIH) filling that void. [National Cancer Institute Director] Harold Varmus, MD, has proposed a national N-of-1 initiative to help advance this concept.
Specifically, the NIH has observed that small numbers of patients on many of the older studies that were deemed negative had profound and durable responses to agents whose development did not proceed, as no biomarker of drug sensitivity had yet been identified. On the basis of our work, there is now great interest in going back and studying such tumors to determine why those patients were unique.
I believe that an in-depth examination of outlier responses will become a standard component of the way we develop drugs. We will continue to be disappointed if a drug fails to achieve its primary endpoint and proves ineffective in most patients. We should, however, make sure we do not miss opportunities to continue to develop drugs that have profound activity in a subset of patients. ■
Disclosure: Dr. Solit reported no potential conflicts of interest.
1. Iyer G, Hanrahan AJ, Milowsku MI, et al: Genome sequencing identifies a basis for everlimus sensitivity. Science. August 23, 2012 (early release online).