Read the Full Video Transcript

Neeraj Agarwal: Hi. We have today, on UroToday, our esteemed guest, Dr. Emmanuel Antonarakis, Professor of Medicine, a medical oncologist par excellence, one of the legends in the field from the University of Minnesota. Welcome.

Emmanuel Antonarakis: Thank you, Neeraj.

Neeraj Agarwal: So we are going to discuss very exciting, very interesting data on the correlation of PARP1 mRNA expression and response to PARP inhibitors in metastatic hormone-sensitive prostate cancer. I, personally, thought it was very interesting because, as you said in our earlier discussion, we have a drug. We have a target. I could not find any literature looking at the target expression. So congratulations.

Emmanuel Antonarakis: Thank you. Yeah.

Neeraj Agarwal: Tell us about it.

Emmanuel Antonarakis: There's a lot to unpack in there, but let's try to break it down for the audience. So we've known about PARP inhibitors. We've heard about them now for almost a decade, even in prostate cancer, right? And, of course, PARP inhibitors target the PARP1 enzyme. That's how they work.

And we all know that these HRR deficiency mutations are the biomarkers of sensitivity. But we don't think that much about the target itself. So we began to wonder, a few months ago, whether the expression of the PARP1 transcript or protein could be prognostic for sensitivity to PARP inhibitors. And also we wanted to know whether there's other just prognostic information about disease aggressiveness, outcomes, et cetera.

Now, we have a problem. And the problem is we don't have a good way to measure or detect the PARP1 protein. For example, immunohistochemistry for PARP1 is not reliable. There are not many labs that can do it reliably. But the transcript—we have available data from whole transcriptome sequencing platforms.

So what we decided to do was we collaborated with our friends at Veracyte, and they have the Decipher platform, as you may know. It's a gene expression array platform, where they analyze all of the transcripts in the genome.

And we did something very interesting. We looked at about 50,000 cases of transcriptome analysis from patients that had had radical prostatectomies. And we decided to look comprehensively at the PARP enzymes. You know how many there are? 16.

Neeraj Agarwal: Wow.

Emmanuel Antonarakis: So PARP1, PARP2, PARP3, all the way up to PARP16. And each one of those is a different gene with a different transcript.

So first question we wanted to ask is, if you look at the highest quartile of expression of those 16 PARP transcripts versus lowest quartile of expression, does that correlate with aggressive features—Gleason Grade Group 5, extraprostatic extension, seminal vesicle invasion?

And out of all those PARP enzymes, PARP1 and PARP2, if the levels were higher, the highest quartile relative to the lowest quartile, were associated with all those bad prognostic features. Interestingly, we saw the exact opposite effect and statistically significant with, for example, PARP6 and PARP7. So if PARP6 or PARP7 are overexpressed, that's actually a good feature.

So that was the first part. The second part was, is this prognostic for metastasis-free survival? So of those 50,000 patients, there were three subcohorts for which we had the metastasis data, time to first metastasis. And once again, we looked at the 16 PARP enzymes. And we looked at which ones are favorably prognostic and poorly prognostic for metastasis. In PARP1 and PARP2, PARP1 was stronger—higher level portends faster metastasis.

Now, the third part was the one that I liked the most and is also the most clinically relevant and fits with our theme today of hormone-sensitive prostate cancer. As you remember, Neeraj, a few years ago, we conducted the first study in a biochemically recurrent prostate cancer, rising PSA after a previous radical prostatectomy.

We've published that paper already. It was 51 gentlemen. And of the 51 patients, about a third of them had BRCA2 mutations, and then the other patients had other HRR mutations.

You might recall from the JAMA Oncology publication, every single one of the BRCA2 patients had a biochemical response—every single one, including some complete responses. And then there were some minor degrees of responses with the other genes.

So using the prostatectomy samples from those patients, we ran the Veracyte transcriptome assays, microarray assays, and two interesting things from that prospective study that came out. The first is that the patients that had the BRCA2 mutations or deletions, they had an upregulation of PARP1 enzyme, but not PARP2 and not the other ones.

So as a secondary phenomenon of BRCA2 deletion, there was a negative feedback loop that upregulates PARP1 enzyme. But the most clinically relevant part is that the expression of PARP1 enzyme at the transcript level was prognostic in that trial, the BCR trial, of response to olaparib. That was the therapy that we were using.

And the important thing, again, keeping with our theme of hormone-sensitive prostate cancer, that was the only trial that evaluated olaparib without any form of androgen deprivation or ARPI. It was an entirely hormone-sparing approach—PARP inhibitor without any hormonal observations.

So PARP1 enzyme, yes, it seems to be important. It's associated with the worse prognosis, worse pathology, extracapsular extension, seminal vesicle invasion, faster time to metastasis. But, in the context of a PARP inhibitor, then became the most sensitive subtype for response to PARP.

Neeraj Agarwal: Fascinating, indeed. Whole story is fascinating. So what are the key takeaways—you just said them—for our viewers who are watching us today, regarding the practical implication of PARP1 expression? Could it be a new biomarker for those patients, where we are not sure if they have HRR gene mutations or not? Or maybe some other way they can help us in choosing PARP inhibitor therapy?

Emmanuel Antonarakis: I love the fact how you're always thinking ahead, Neeraj. Yes, I think this could be a biomarker. I think there's two things, though. First, we need to validate this. It would be great if we had other prospective studies which are much larger than 51 patients, especially if they're randomized studies, where we had an NGS assay that analyzes DNA and RNA.

Because if we had such an assay, for example, as you well know, there are many commercial companies out there—Tempus is one of them, Caris Life Sciences—there are a few others—that when you order that test or when you include that test in your prospective trial, they will sequence the exome and the transcriptome.

With the transcriptome piece, we could evaluate PARP1 transcript as a biomarker. But the second part—so it needs to be validated. The second part is even more interesting but requires more confirmation also, which is that we all know that there are small subsets of patients that don't have a detectable HRR mutation. It might be there, and we can't detect it, or it might not be there. Or it might be a different gene that we are not currently considering as an HRR gene.

But what if some of those patients have super high PARP1 transcript level, so they are showing the evidence of an HRR deficiency by the fact that their PARP1 expression at the transcript level is high, but you can't find the gene mutation?

Neeraj Agarwal: Or we don't know of that gene mutation, but we can see the downstream effect of PARP upregulation.

Emmanuel Antonarakis: Yeah. So this could be a patient that you may consider treating with a PARP inhibitor. So I think that's the future. There's a lot more steps that need to be taken before we get there. But we are always looking into the future, not just the present. And I think this will be one of the exciting things that I hope will come out from this research.

Neeraj Agarwal: Fantastic. Again, thank you for sharing those findings. And hopefully, PARP1 expression in the prostate cancer patients emerge as a biomarker of response and a guide for therapy selection for PARP inhibitors down the line in our patients. Thank you.

Emmanuel Antonarakis: Thank you very much. Thank you.