NEW YORK (GenomeWeb) – Although enthusiasm for one specific biomarker, AR-V7, is driving the first major advances in blood-based testing to guide prostate cancer treatment, results of a new study have provided the strongest evidence so far that taking a broader approach can identify other genomic or molecular information that could also have clinical value.
The study, led by a group from Duke University and published in Clinical Cancer Research, used array-based comparative genomic hybridization to search for copy number changes in circulating tumor cells isolated from men with advanced prostate cancer who were receiving treatment for their disease using one of two hormonal therapies: enzalutamide or abiraterone acetate.
While some previous research efforts have shown that isolation and genomic or trancriptomic analysis of CTCs can uncover alterations with at least some clinical importance, they have been limited.
Andrew Armstrong, the study's senior author, told GenomeWeb this week that the goal of the recently published work was mainly to add incrementally to what had been seen in earlier studies, and also to establish the feasibility of a broader profiling approach in order to support a much larger study evaluating array-based and sequencing-based CTC assessment, analysis of circulating tumor DNA, and comparing three competing individual AR-V7 transcripts assays.
"Ar-V7 is measured at the RNA level. It's a splice variant and it seems to be strongly associated with resistance to these therapies, but it doesn't explain fully the resistance story. For example, there are many patients that are AR-V7 negative but still progress on these drugs," Armstrong said. "We also don't know if it's the [real] mechanism for resistance. Some patients may just have altered splicing and very complex tumors, so they have more V7, but that's not actually what's causing the resistance.
"We wanted to take a much more holistic approach to look at what could be reproducibly identified and what might be targetable, and what might change over time," he said.
In the study, Armstrong and his coauthors recruited a total of 29 men with metastatic castration-resistant prostate cancer and attempted to isolate CTCs from their blood using the EpCAM-based CellSearch assay. Seven of these men had to be excluded due to a lack of isolatable CTCs. Six others had samples that didn't pass quality control screening of their genomic DNA.
In all, the group was able to perform array CGH on DNA from CTCs in 12 men, in some cases from samples before and after experiencing disease progression on either enzalutamide or abiraterone.
All the men in the study treated with abiraterone eventually progressed. Three of these men also became enzalutamide resistant, while two men had a short-term response to enzalutamide after abiraterone, and one had a temporary response to docetaxel chemotherapy.
According to Armstrong and his team, the array CGH revealed a number of biologically relevant alterations. Overall, the results showed that there is both high between-patient and individual heterogeneity in the landscape of CTCs in these men, which suggests that tissue biopsies may not always capture all the relevant alterations present in a tumor.
The data also showed that there was clear clonally divergent evolution of CTCs during treatment with abiraterone or enzalutamide. If further validated, some of these changes could prove useful as biomarkers for tracking or predicting resistance, or potentially as new drug targets, the authors wrote.
"We saw some interesting novel things," Armstrong said, "including gain of the ERG oncogene, or transformations to neuroendocrine genes like MYCN that became gained in a patient that progressed and who lost AR amplification."
"Given recent findings linking ERG expression to taxane resistance, this finding has potential clinical implications for treatment," the study authors wrote.
CTCs from men in the study also showed alterations in BRD4, a target for which inhibitors are already being developed.
This first effort only included a handful of patients and was limited to copy number analyses via array CGH. But moving forward, Armstrong said the team is already well into a follow-up prospective observational study that could help clarify whether AR-V7 is indeed the most useful biomarker associated with response to particular therapies, or if there are other molecular features in either CTCs or ctDNA that might be more predictive, or add to the value of the single splice variant.
In this study, supported by the Prostate Cancer Foundation, Duke is leading a five-center effort to analyze CTCs from men with advanced prostate cancer before and after treatment with abiraterone, enzalutamide, and taxane chemotherapy using arrays, exome- and RNA-sequencing of CTCs, and sequencing of ctDNA.
In parallel, the team — which also includes Cornell University, Johns Hopkins University, Memorial Sloan-Kettering cancer Center, and the University of Chicago — is also comparing three different AR-V7 assays. One, which measures AR-V7 transcripts in the blood, was initially developed by researchers at Johns Hopkins and was also being investigated by Qiagen and Tokai Pharmaceuticals. Another, developed by Epic Sciences and academic collaborators, uses Epic's proprietary CTC isolation platform to measure AR-V7 in CTCs. The third is an assay currently in development that uses multiplex PCR, Armstrong said.
Armstrong said it's not clear yet how all the different arms may shake out in terms of clinical utility. "We are hoping to look at how DNA changes may reflect RNA changes, whether there is something beyond AR-V7, or whether these DNA elements we identified, like extra copies of AR, for example, are associated with AR-V7, and what really is the most predictive," he said.
An important limitation to CTC-based analyses is that a substantial proportion of men with these cancers do not appear to have isolatable CTCs, so the group's study may also reveal ctDNA changes that correlate with therapy response that could be picked up in these men and close that clinical gap.
Broad profiling of CTC DNA, RNA, and ctDNA in all patients with metastatic prostate cancer isn't exactly economically feasible or commercially viable, Armstrong stressed.
"You obviously want a much more predictive single biomarker," he said, "so that's where this study comes in. Let's say we find one thing, like AR amplification, that is highly predictive — you could have a simple probe for that using [any number of] different CTC isolation methods."
"It's also possible you could refine what we find in this study to a hotspot panel," he added. In that case, the group will also have laid important groundwork to show that such a panel can be useful applied to circulating nucleic acids, without the need for a tissue biopsy.
At the very least, the results will likely yield some needed clarity about how different methods of assessing AR-V7 directly compare to one another.