Prostate Cancer: Epic Sciences AR-V7 Test

Transcript:

Emmanuel S. Antonarakis, MBBCh: Another interesting assay to determine and detect AR-V7 has been developed by a company called Epic Sciences, and they have been working with my colleague, Dr. Howard Scher, at the Memorial Sloan Kettering Cancer Center. The Epic Sciences assay is somewhat different and unique from the Johns Hopkins assay. So, just to recap: in the Johns Hopkins assay, we were looking for CTC-specific mRNA for AR-V7, so we were looking for the transcript of AR-V7. Epic Sciences, on the other hand, is focusing on the AR-V7 protein product. They’re not looking at the RNA, they’re looking at the protein. And in addition, they’re not only looking for the protein, but they have the ability, on their platform, to look at the localization of the protein. In other words, where is the AR-V7 predominantly found? Is it predominantly in the cytoplasm of the cell outside the nucleus or is it predominantly intranuclear, inside the nucleus?

So, the initial study that was published by Dr. Howard Scher in JAMA Oncology 2015 showed that using their platform, which is again different from the Johns Hopkins platform, showed some very similar results, namely that patients who had AR-V7 protein detected on the circulating tumor cells using the Epic Sciences assay had worse outcomes to novel hormone therapy, enzalutamide and abiraterone. And, interestingly, they had a second part to their study where they looked at the prognostic value of AR-V7 protein on taxane response. Taxanes are the most common chemotherapies used for prostate cancer; the most common one being docetaxel and the second one being cabazitaxel. And one question that arises is, if patients with AR-V7—positive detection are resistant or have inferior outcomes to abiraterone and enzalutamide, does that also hold true for chemotherapy, specifically the taxane agents? What Dr. Scher showed using the Epic Sciences platform was that the AR-V7, a positivity presence did not appear to be associated with a primary resistance to chemotherapy. In other words, despite the AR-V7 protein being present, some patients, not all, could still respond favorably to taxane therapy.

The next step was to look for the differential impact of the biomarker in the setting of AR-directed therapy and the setting of chemotherapy. In other words, does the biomarker help to select therapy A over therapy B? Is it a treatment specific biomarker or not? When Dr. Scher’s group combined the data from their enzalutamide and abiraterone cohort and the data from their separate taxane cohort, they showed that the AR-V7 presence favored a response to chemotherapy over AR-directed therapy. But AR-V7 absence—in other words, a negative AR-V7 test—produced relatively equal outcomes with both AR-directed therapy and with chemotherapy.

Our group at Johns Hopkins also published a second paper using our AR-V7 mRNA-based assay, which also, in fact, showed very similar results; namely that the Johns Hopkins assay in the context of AR-directed therapy—in other words, enzalutamide or abiraterone—had a negative prognostic effect. But in the setting of taxane therapy, it did not appear to be a biomarker of absolute resistance, again supporting the notion that in the AR-V7—positive situation, a taxane therapy, may perhaps be more superior to an AR-directed therapy. But in the AR-V7–negative situation, both an AR-directed therapy and a taxane therapy may be equivalent.

Howard I. Scher, MD: In the early 90s, it was clear that assays were becoming available to look to see if cancer cells were spreading in the blood. The first test we used was essentially to look for PSA message. The messenger RNA that made PSA is a commonly used protein for diagnosis in patient management in prostate cancer. And we postulated that if a patient had cancer cells in their blood, they would have a worse prognosis than a patient who didn’t. And what was interesting to us was we were looking at patients who were treated with hormonal therapy whose PSAs were undetectable.

We were very happy. They seemed to be responding, treatments were working, and the patient felt fine. But we were still able to identify messenger RNA for PSA despite the PSA being 0. We know that PSA is regulated by androgens, which means that the amount that is produced will be lower if testosterone levels were lowered in a patient. But importantly, it signaled that there was still activity of the cancer that could be deleterious to the patient. In other words, the cancer cells were still spreading.

For me, this represented unique information, and we spent the next several years looking at different ways that we can identify whether or not those cells were present as a first test, if you will. Ultimately, we were able to identify them, to start working on ways to characterize them at a biological level.

Right now, there is one circulating tumor cell test that has reached the level of an FDA clearance. It’s a test called CELLSEARCH. It was marketed by Veridex, and what it essentially does is it looks for the EpCAM-positive cells that are circulating in the blood. It was a test that was designed to be reproducible, accurate, and was first approved in 2004 as an aid to monitoring breast cancer. And what the investigators were able to show is when they looked at a group of patients who had cancer cells in their blood, their prognosis was worse at the time of first treatment. And if a treatment was given and those circulating tumor cells were shown to decrease below a certain level, they live longer and became a measure of response.

So, with the Epic platform, there’s no capture of specific cell types. If you’re using an EpCAM capture, you’re not capturing or identifying all the cells that are present in the blood. What Epic does is essentially take a blood center. It will lysis the red cells and puts in all of what’s called the mononuclear fraction, which includes lymphocytes, white cells, as well as the rare tumor cells.

And it functions a little bit like a Xerox machine. It will scan the slide, and the software will identify cells that have certain characteristics based on staining. So, you could stain for the androgen receptor, you could stain for cytokeratin, and you could stain for a protein called CD45, which identifies the white cells. But the software will find the cell, localize the point, and count the number of events that occur.

This is different than the capture methods from blood, which basically take a pool of cells. So with the Epic platform, you can look on a cell-by-cell basis. And what you start to see, there are a range of cell types. There are some that are dying, there’s others that don’t express cytokeratin. You can see big cells, small cells, and you can see different patterns in the nucleus. Some look dotted, some look like there’s a bullet in the middle.

And what has happened over time is that we’ve started to characterize these different cell types and are starting to look at them, what they mean biologically. There is a predicate for this, and really the first application was in identifying cervical cancer in PAP smears. So, it’s the same concept of, if you will, facial recognition software, and the initial software was able to distinguish the difference between normal cells, inflammatory cells, and cancer cells. The beauty of it, scanning, you get an answer. And there were several assays that were actually approved to do that.

Think of the Epic platform as an extension of this, where not only are you looking at some basic features, but as the software gets more and more sophisticated, it’s almost like racial profiling. So, you can look at someone who has big eyes, blue eyes, green eyes, keep going through that at multiple levels, and that’s, in essence, what you can start doing and what we have been doing with the Epic platform.

Transcript Edited for Clarity