Originally published Nov. 18.
Clovis Oncology has designed a clinical trial program for its PARP inhibitor rucaparib, through which the company hopes to define a DNA repair gene signature associated with drug response, use that signature to garner US Food and Drug Administration approval for the drug in a molecularly defined patient subset, and launch the first next-generation sequencing-based companion test.
At the World CDx conference in Boston last week, Clovis Senior Director Mitch Raponi said that the company had launched a Phase II biomarker program for rucaparib involving patients with high-grade serous ovarian cancer, “to focus on what is the most appropriate companion diagnostic approach.” In parallel with the Phase II program, in which the firm is planning to “lock down” a DNA repair gene signature associated with drug response, Clovis will also initiate a Phase III trial in which it will use that genetic test to prospectively stratify patients and establish rucaparib's safety and efficacy in molecularly defined patient subsets.
A number of studies have now shown that patients with different types of cancer carrying germline BRCA mutations – which are inherited and occur in every cell in the body – benefit from treatment with PARP inhibitors. However, Raponi noted that “clearly there are patients who do not have germline BRCA mutations that are also deriving clinical benefit, particularly those that have somatic BRCA mutations in their tumor tissue, as well as other non-BRCA DNA repair genes.”
As cells divide and replicate in our bodies, the DNA inside those cells is damaged thousands of times. Most of these DNA lesions are patched up by proteins charged with this task. However, mutations in key repair genes can throw a wrench in this corrective mechanism, which in turn can cause diseases.
BRCA1 and BRCA2 are examples of genes involved in homologous recombination, the process by which cells repair breaks in both strands of DNA by exchanging nucleotide sequences between two like DNA molecules. Women who carry mutations in BRCA1 and BRCA2 genes have an increased risk of hereditary breast and ovarian cancer, but those same mutations that hinder their cells from repairing DNA damage make them more likely to respond to PARP inhibitors, such as rucaparib.
These inhibitors block PARP proteins – also involved in repairing DNA breaks in cells – from doing their job. Since cancer patients with BRCA mutations already have an impaired ability to repair DNA, the idea is that when given a PARP inhibitor, their rapidly multiplying tumor cells are inundated with faulty DNA and die. By this logic, researchers believe that cancer patients with mutations in other genes involved in DNA repair may also benefit from PARP inhibition. And a companion test that detects these mutations would help doctors identify the patients who would respond particularly well to this class of drugs.
In Phase I studies, Clovis has seen encouraging responses to rucaparib from metastatic breast, ovarian, and pancreatic cancer patients with BRCA1/2 mutations. In the Phase II trial, Clovis will enroll all high-grade serous ovarian cancer patients who have relapsed on platinum therapy and give them rucaparib. Clovis is hoping to refine its biomarker signature, including BRCA1/2 mutations and a number of other DNA repair markers, in the all-comer population.
In Phase III, Clovis will compare rucaparib against placebo in high-grade serous ovarian cancer patients who have responded to platinum therapy. In this part of the study, researchers will validate the gene signature from Phase II and identify best responders to the drug using a step-wise approach to treat BRCA-mutated patients with rucaparib first and then patients with other DNA repair-associated gene mutations.
“Our goal is to develop a homologous recombination deficiency signature, also known as a BRCA-ness-type signature, to try to identify all patients who are likely to benefit from rucaparib,” Raponi said.
Citing data published by The Cancer Genome Atlas in Nature in 2011, Raponi estimated that approximately 50 percent of women with high-grade serous ovarian cancer had defects in homologous recombination genes, a significant proportion of which are not BRCA mutations. Using whole-genome sequencing and other types of analyses, TCGA researchers have reported that 96 percent of the tumors harbored mutations affecting the tumor suppressor gene TP53; 3 percent have BRCA1/2 somatic mutations; 9 percent harbor germline BRCA1 mutations; and 8 percent have BRCA2 germline mutations.
Several drug developers are working with Myriad Genetics, to investigate the impact of germline BRCA mutations on how breast and ovarian cancer patients respond to PARP inhibitors, but Raponi noted that this may be short-sighted. “A number of our competitors in this space are focusing on just germline BRCA mutations but we think at Clovis, they're missing a whole other population of patients, if you don't look at the tumor,” he said.
A race to market
Market analysts predict that PARP inhibitors could sell for $12,500 to $15,000 for a month's supply, bringing the cost of a year of treatment to as much as $180,000. According to the National Cancer Institute, BRCA mutations account for up to 10 percent of all breast cancers and approximately 15 percent of ovarian cancers overall. By targeting these drugs for the germline BRCA-mutated cancer population, developers of PARP inhibitors are racing to snap up a relatively small subset of the breast and ovarian cancer market.
In an attempt to establish a broader indication and differentiate its drug from other PARP inhibitors that will likely enter the market before rucaparib, Clovis is exploring DNA repair response markers beyond germline BRCA mutations. At the same time, Clovis is working with Foundation Medicine to try to carve out a regulatory path to market for its NGS-based test as a companion diagnostic, a challenge few drug developers are willing to take on at the moment.
However, drug/diagnostic co-development has its risks, and in order to minimize uncertainties, competing drug companies are working with Myriad, the leading provider of BRCA mutation testing. Myriad has nearly two decades of experience with BRCA testing, more than any other company on the market. However, Myriad's flagship BRACAnalysis is a blood test that interrogates germline BRCA mutations. Therefore, BRACAnalysis will not identify those cancer patients with somatic BRCA alterations, which they acquire only in tumor cells.
Earlier this year, Foundation launched its NGS-based test, called FoundationOne, which analyzes formalin-fixed paraffin-embedded tissue and can gauge more than 200 cancer-related genes. The company currently performs the test in a CLIA-certified lab.
Clovis is hoping to file an NDA for rucaparib by 2017, and in the ideal scenario, around the same time, Foundation would also file the pre-market approval application for the NGS-based companion test. The FDA has expressed a willingness to work with sponsors to figure out the regulatory pathway for NGS-based diagnostics, but hasn't recommended any concrete guidelines.
Another reason Clovis' competitors have chosen to partner with Myriad and use BRACAnalysis as the companion test to their drugs is because doing so will likely allow them to bring their products to market more quickly than if they decided to use a newer technology platform. Myriad has already initiated Phase III trials with three drug developers – AstraZeneca, BioMarin, and Tesaro – which are using BRACAnalysis to identify best responders to their investigational PARP inhibiting agents.
Of these partners, Myriad has said that AstraZeneca's olaparib with BRACanalysis as a companion test will likely be the first to market in Europe in 2015. The drugmaker is using the BRCA mutation test in two Phase III trials: one investigating olaparib maintenance therapy in the first-line setting in ovarian cancer patients with BRCA mutations, and another assessing the drug in platinum-sensitive, relapsed ovarian cancer patients who harbor these mutations.
Tesaro is using Myriad's BRACAnalysis test in two Phase III trials, one involving platinum-sensitive ovarian cancer patients and another involving germline BRCA-positive breast cancer patients.
“We are evaluating both germline BRCA mutation status and platinum sensitivity as potential predictors of clinical response in our Phase III study in ovarian cancer,” said a spokesperson for the company. “Our trial is enrolling both germline BRCA-positive and -negative patients, all of whom are platinum sensitive, in two separate cohorts. Regardless of BRCA status, given the preponderance of homologous recombination deficiencies among platinum sensitive patients, both germline BRCA positive and negative patients could be sensitive to PARP inhibitors, and this study design will enable us to analyze responses in each of these groups.”
BioMarin has similarly said it will study its PARP inhibitor BMN-673 in metastatic breast cancer patients with germline BRCA mutations. However, the company is also betting that its drug will be more potent than olaparib or rucaparib. According to data presented in October by NCI researchers at a major medical conference, BMN-673 was 100-fold more effective than the other two agents at trapping PARP proteins at DNA damage sites. These accumulating trapped proteins, in turn, become toxic to cells.
With regard to the greater PARP trapping ability seen with BMN-673, Clovis's Chief Medical Officer Andrew Allen told PGx Reporter that while BioMarin's drug may be more potent in this context, it also results in more toxicities, giving it a “chemotherapy-like profile” where patients experienced multi-lineage myelosuppression and alpoecia. Researchers have not seen these types of adverse events with rucaparib, Allen said.
AstraZeneca and BioMarin didn't respond to questions ahead of press time.
Capturing all responders
While all three drug developers are using BRACAnalysis to investigate how patients with germline mutations are responding to their drugs in pivotal trials, at least one of these firms, AstraZeneca, has publicly discussed exploratory work with NGS. AstraZeneca used Foundation's NGS test to retrospectively investigate how patients with somatic BRCA mutations responded to its PARP inhibitor olaparib. Data from this Phase II trial showed that patients with somatic and germline BRCA mutations had very similar median progression-free survival while on olaparib.
In a presentation at the Next Generation Dx Summit in Washington, DC, this summer, J. Carl Barrett, VP of translational science oncology at AstraZeneca, reported “strong concordance” between BRCA1/2 mutation calls by Foundation's test and by Sanger sequencing, in this same Phase II study. There were some discordant results, which Barrett said highlighted the bioinformatics challenges of using NGS technologies. However, noting that NGS testing identified the subset of sporadic BRCA patients, he said that AstraZeneca is continuing to investigate other homologous recombination genes.
In the retrospective Phase II study, when comparing the response of patients with normal BRCA genes treated with olaparib and placebo, researchers reported a hazard ratio of 0.53. This, Clovis' Raponi said, indicates that some patients without BRCA mutations are still responding to olaparib. “There are probably patients in this non-BRCA [mutated] group who are responding very well, may be similar to the BRCA [mutated] group, and another group that's not seeing any clinical benefit. The question is, who are those patients?” he posited.
Meanwhile, the fact that AstraZeneca partnered with Foundation to study how patients with somatic BRCA-mutations responded to its PARP inhibitor suggests that the company may decide to expand the indication of its drug in the future after initially launching it in germline BRCA-mutated patients. If it decides to do so, AstraZeneca can choose from a number of available platforms, including FoundationOne and Myriad's NGS-based homologous recombination deficiency test that gauges BRCA1/2 gene mutations, as well as a number of other markers involved in DNA repair.
Drug developer PharmaMar has already partnered with Myriad to use its HRD test in a Phase II study of PM1183, an agent that causes double-stranded DNA breaks to kill cells. [Editorial Note: After this article was published, BioMarin announced it was also using Myriad's HRD test to investigate how different tumor types respond to BMN-673.] Myriad is planning to launch its HRD test for platinum therapies in 2015 and for PARP inhibitors in 2017.
Myriad researchers and collaborators published data in the British Journal of Cancer last year showing that the HRD test can gauge whether ovarian tumor cells have impaired homologous recombination ability. The researchers noted in the paper that "a substantial fraction" of tumors with intact BRCA1, BRCA2, and RAD51C have elevated homologous recombination deficiency scores. "Two possible explanations are that there is a substantial rate of defects in other genes in the HR pathway in many of these samples, or that the HRD score is non-specific," the researchers wrote. The company has also shown in a study that its HRD score predicts whether patients with triple-negative and BRCA1/2 mutated breast cancer respond to neoadjuvant platinum-based therapy.
Complex program
Raponi acknowledged at the meeting in Boston that the drug/test development pathway Clovis has decided to take on is more challenging than the pharmacogenetic biomarker strategy employed by its competitors. “This is a complex clinical program. Clearly we're all familiar with approved companion diagnostic tests, which when you compare to this type of program, is relatively straightforward,” he reflected.
Foundation recently published in Nature Biotechnology an analytical validation study of the assay in over 2,000 patients. “This is a very important step forward,” Raponi said of Foundation's publication, noting that “it will require some refinement with regard to how we do this particularly in the context of companion diagnostics.” He noted that both Foundation and Clovis are having ongoing communications with the FDA to figure out the requirements for regulatory approval of a rucaparib and an NGS-based companion test.
But before tackling regulatory requirements, Clovis will have to finalize its biomarker signature for validation in Phase III studies, a difficult task made even more complex by NGS. “Here we're looking at a whole array of genes, a signature of genes that represent a pathway. And not only are they multiple genes, but they're not necessarily oncogenes,” Raponi explained. “Many of these genes are tumor suppressor genes. And the issue there is that we're not looking at hot spot mutations, we're looking at genes that have hundreds or thousands of mutations throughout the coding region of the gene.”
Clovis researchers used the published literature and past experiments from groups like TCGA and preclinical models to generate their hypothesis about the genes of interest related to rucaparib response. Based on that work, the company is currently focused on a list of 30 genes that they plan to further refine in the Phase II trial. “We … plan to optimize that signature in our Phase II biomarker study, ARIEL 2, and while all of this is going on, in parallel we're running this ARIEL 3 program,” Raponi said. “Prior to this final analysis of … this pivotal trial, we will lock down the final companion diagnostic test and prospectively [assess] that in ARIEL 3.”
Using the Foundation platform, Clovis also plans to look beyond gene mutations into other molecular features involved in DNA repair, such as gene amplifications, translocations, and large deletions. “Something we're also looking at … is how can you leverage this type of genomic information to look at a phenotype of DNA repair deficiency,” Raponi said. “So, you can think of how when a cell has deficiency in DNA repair, the DNA gets screwed up and becomes scarred. And by looking at these [molecular markers] we could come out with a phenotypical read out of what a homologous deficient cell would look like.”
Running parallel Phase II and Phase III programs with a companion test will require even greater coordination and strategic alignment between the research teams at Clovis and at Foundation than if the companies were doing these studies sequentially. “From the clinical side, time is of the essence,” Raponi said. “And while five years ago, developing a molecule in the clinic would take 10 years, today we're really looking at five to six years. That has huge implications for developing companion diagnostics.”
Clovis is also exploring a pharmacogenetic strategy for a number of other drugs in its pipeline, including a non-small cell lung cancer treatment, CO-1686, for patients with EGFR mutations who are resistant to EGFR inhibiting drugs. Clovis is working with Qiagen to advance a companion test for CO-1686.
Clovis was also developing the pancreatic cancer drug CO-101 as an alternative for poor responders to gemcitabine, but the company had to stop developing the agent after a clinical trial failed to show any improvement in survival time in a molecularly defined subset of patients. Clovis pursued this PGx strategy after multiple retrospective studies had indicated there was a link between hENT1 expression and survival for patients on gemcitabine.