The Uromigos Episode 153: ASCO GU 2022 Overview of PARP Inhibitors in Prostate Cancer

Pete Nelson, MD, describes mechanism of action and current clinical application in prostate cancer

Episode Transcript

Tom:

Hey everybody. I’m joined here with Pete Nelson. Pete it’s an honor for you to be here with us today. We’re going to talk a little bit about PARP inhibition. They are these 2 big, randomized phase 3s coming up at GU ASCO. We just wanted to get a feel, Brian and I wanted to learn a bit more certainly and we wanted to ask you a couple of questions. If you could introduce yourself to start with and then we’ll fire away from there.

Pete Nelson:

Sure. You know, it’s a great pleasure to be here with you. My name’s Pete Nelson. I’m from the Fred Hutchinson Cancer Research Center in the University of Washington in Seattle. I’m a medical oncologist, but also have a research lab focused on DNA damage repair, androgen receptor, essentially everything prostate cancer.

Tom:

Pete, can we start by just saying, let’s just start with PARP inhibition and the mechanics of action. There’s this double stand break. There’s single strand break. There’s synthetic lethality. What’s going on with all of these terms?

Pete Nelson:

Right, so the essence of it really starts fundamentally with the requirement for cells to continually have mechanisms to repair DNA. So, DNA is constantly under attack from both exogenous and internal DNA damaging components and so, also as cells replicate and divide their DNA and as transcription occurs, DNA breaks occur as well. So, we have these remarkable systems, sort of failsafe systems, redundancy to some level, that are involved in repairing different types of DNA damage.

Probably the most lethal or the worst type of damage is when both strands of the DNA are broken. And there are at least 2 mechanisms that are responsible for repairing these double strand breaks. One of them, which has more errors in it, is called non-homologous end joining, or NHEJ. The jargon gets pretty complex pretty quickly, but NHEJ is one pathway and then the second pathway is called homology directed repair or homologous recombination.

And so, these are somewhat redundant, although it’s important to recognize that the homology directed repair, has much more fidelity. When you have repair by that mechanism, you preserve the sequence of the DNA, whereas with the non-homologous end joining, or NHEJ, there are errors incorporated into the repair DNA. So that’s, I think a bit of a background, but happy to go deeper, or yeah.

Brian:

So, Pete, the idea of course, is that cancer cells can’t repair this DNA damage and therefore are killed when they’re treated with PARP inhibitors, but can you tie that into, because there’s so many different mutations. So, what gets confusing is that, I think when you have an EGFR and ALK mutation, I can understand that it’s a single mutation and a cancer cell, but this is different, and I think it’s this litany of BRCA and related that I think gets confusing to some of us.

Pete Nelson:

Yep. So, think about it this way, that, so cancer cells also have usually the ability to repair DNA. In fact, they repair it usually fairly efficiently, although damage does accumulate. The key aspect for the BRCA genes and when we think about synthetic lethality and PARP inhibitors, so these tumor cells, there’s a subset of cancers where there’s a defect in homology directed repair or homologous recombination. And these are the BRCAness tumors.

So, often you may inherit a defect, a mutation in one of the genes, such as BRCA2, B-R-C-A-2, that’s probably the most common and that predisposes individuals to breast cancer, ovarian cancer, prostate cancer, pancreas cancer, these epithelial cancers, which is still pretty remarkable that it’s that subset of cancers that someone may be at risk for, but not lung cancer or colorectal cancer. So, don’t want to digress too far.

But when you have an inherited mutation in, for example BRCA2, the tumors that develop in those individuals, have lost the second copy. So, we’re born with 2 copies. One of them is defective. The tumor acquires a loss of the second ileal. So, those tumors are then exquisitely dependent on NHEJ, the other repair pay pathway. So, PARP inhibitors work by blocking than the other pathway. So, your normal cells in the body, the PARP inhibitor will block NHEJ in those cells, but they still have an intact homology directed repair pathway.

So, that’s where you get this synthetic lethality, you get a differential response in those tumors that have a defective homology directed repair, because then they’re vulnerable to knocking out that other arm, that fail safe, that redundancy pathway.

Now you can, in addition to the hereditary inherited forms, cancers can also arise by biallelic or losing both copies somatically and we see both of those occur in prostate cancer at roughly the same frequency. So, that’s why you get this synthetic lethality. That’s why PARP inhibitors are thought to be precision medicine in terms of their effectiveness in those tumors that have a defective homology directed repair pathway and that was the entry point for the 2 studies. The PROfound and TRITON study. They were biomarker selected trials. The patients coming in had to have been tested to confirm that they had an alteration, a loss in one of several genes involved in homology directed repair or homologous recombination.

Tom:

So, essentially what’s happening is, we got a DNA abnormality to BRCA1 or BRCA2 or a series of other genes. If that’s pre-existing and you put PARP inhibition on top of that, that results in synthetic lethality and cell death. And that’s in broad terms, how this is working. One of the things which I’m confused about, is in the ovarian cancer studies or some of them, they’ve been reliant on the presence of [inaudible 00:07:11] of germline mutations, where people have looked in the blood to look for mutations, whereas in prostate cancer, we’re looking in the tissue for mutation. Is that correct and why are we using different approaches?

Pete Nelson:

Well, it mostly advances in technology. So, you can also use circulating tumor DNA or cell-free DNA as a diagnostic tool in prostate cancer. So, foundation, for example, has a circulating tumor DNA test or a lot of home brew tests that different labs have developed. So, you can use either one.

Tom:

Okay.

Pete Nelson:

[crosstalk 00:07:44] brain cancer. Yeah. You can use a tissue-based assay in ovarian or breast.

Tom:

So, Pete, I’m bought into BRCA1, I’m bought into BRCA2. I’m not so bought into ATM and then there’re a string of other hangers on. What are they doing there? Why are they included? How important are they?

Pete Nelson:

Right. So, when the studies were originally done, if you go back and look in preclinical models, it’s been well understood that this repair pathway is extremely complex. There are multiple proteins involved in complexes that help repair the DNA and many of these are involved in the Fanconi’s anemia syndrome or heritable Fanconi’s, which we can get into if you like, but there’s complexes of proteins and these are involved both in signaling DNA damage to tell the cell, “Yes, we’ve got damage. You need to now assemble these complexes.” And then particular proteins in the complexes that actually carry out the repair work.

So, in preclinical studies, there were a spectrum of genes thought to be that if lost or mutated would confer this vulnerability. And so, those were selected as entry criteria. What has panned out though, is that they’re not all created equal in terms of, so although ATM, which you mentioned, that’s one of these sensor proteins or pathways. It’s not directly involved in the repair process generally. And it’s turned out that patients with ATM mutations, don’t seem to really respond to PARP inhibition. Now there’s some nuance there, because it may be the type of ATM mutation that may play a role. So, it does look like occasionally, there are ATM patients that respond, but they may have other defects that just weren’t recognized either.

So, it is very complex, but I would just agree with you that the clinical data, do not support ATM as being a susceptibility gene to PARP inhibition, although it is still on the label for Olaparib and that was one of the, I think confusing things to many. When the trials were designed, so the large ones, which now the PARP inhibitors have been approved are Olaparib and rucaparib.

So, the PROfound study, which was Olaparib, included a large panel of these genes, I think at least 15. And whereas rucaparib, the TRITON 2 study, only included really 2 genes, primarily BRCA1 and BRCA2. And when the FDA looked at the data, they approved this whole panel of 15 markers for Olaparib, whereas only 2 are really approved technically for rucaparib. So, it is a bit confusing because these 2 PARP inhibitors are fundamentally identical. There is some nuance with other PARP inhibitors, in terms of how effectively they block the PARP enzyme function, versus trap PARP on DNA, which is also one of their mechanisms of effectiveness.

Tom:

[crosstalk 00:11:14] Could you just expound on that for us, because that’s an important issue. So, there are different PARP inhibitors and this trap concept, which people feel is more important, which of the drugs are effective at which of those two? And how important do you think they are? And is there a better PARP inhibitor?

Pete Nelson:

The answer to your last part, we don’t know yet whether there’s a better PARP inhibitor. They’ve never really… I don’t know if they ever will be tested head-to-head clinically. In preclinical models, there’s a PARP inhibitor called Talazoparib, which seems to have the most potent trapping effect, but they’ll all trap to some degree. And so, I would just say for the two approved at this point, at least for prostate, the Olaparib and rucaparib, they would be fundamentally identical, such that if somebody failed, for example, Olaparib, it’s very unlikely they’re going to respond to rucaparib [crosstalk 00:12:14]

Tom:

[crosstalk 00:12:14] Pete, in the magnitude trial, niraparib’s being investigated. Is niraparib as good as the other two in trapping or is it less potent?

Pete Nelson:

I honestly don’t know exactly how much more potent it would be at trapping. It does in preclinical models roughly have the same effect in a, for example, a BRCA2 loss tumor. So, I don’t think we’re going to see major differences, in terms of response rates or outcomes.

Brian:

So, Pete, you talked about some resistance a little bit or how biomarker patients don’t respond, because they don’t all respond and these drugs are, what I would consider, typical targeted therapy. You get response rates in the 40%, you get PFS in the 6-, 8-month range, sort of with many targeted therapies. So, how does that work? How do biomarker patients not respond, or do we know anything about resistance mechanisms?

Pete Nelson:

Yeah. So, that’s a couple really important questions. First, we need to be careful of what we call biomarker positive. So, unfortunately the FDA label doesn’t explicitly state, you need biallelic loss. So, it doesn’t explicitly state, you need both copies lost. And when you look at the reports, when you send off a test to commercial labs such as Foundation, or even in internal, sometimes they don’t actually report whether there is or isn’t biallelic loss and it is harder to determine that from a circulating DNA assay than a tissue assay, but the general thinking is, you need to have both copies of a given gene lost. That monoallelic or single copy loss, will not confer susceptibility.

So, there is a concern when you look at the trials, they didn’t mandate biallelic loss. Okay, some of these non-responders, I should say, might have been classified as biomarker positive because one of the copies might have been [inaudible 00:14:20], but they didn’t absolutely confirm that the second copy was lost. So, that’s the first, I think, important take-home point from a clinician. If you really are going to allocate the therapy, it’d be optimal to know that both copies of a given susceptibility to gene or loss.

The second thing is that the mutations don’t actually tell you whether the tumor does have a functional loss of DNA repair. In this case, homology directed repair. So, there are some other assays that can tell you functionally, whether the tumor has these defects. One that’s used more in ovarian cancer, is simply a term, called loss of heterozygosity. So, you see very large changes in the genome of these tumors and there is an assay that produces, what’s called LOH score. So, that’s a signature left by that damage mechanism that you can read out in a tumor.

And then there are some assays trying to be developed, that will also give you a functional readout. So, in a tumor, if you induce damage, you can see, what are called foci of RAD51. That’s another one of these DNA repair genes. This gets a little bit complex, but if you don’t see RAD51 in the tumor cell, it indicates that, that tumor cell is defective in homology directed repair and there are some labs trying to develop an immunohistochemical assay for these RAD51 foci.

So, a lot more still to be done, to be more precise about, first, the biomarker allocation and then your next question really was about resistance.

So, it was recognized years ago in ovarian cancer, that one of the mechanisms of resistance to either platinum or PARP inhibition, in patients with homology directed repaired deficiency or BRCAness, is it turns out that within the tumor, a second mutation can occur. Let’s talk just about BRCA2. There may be a single alteration in one of the bases that disrupts the protein, but a second mutation further down, for example, can correct that mutation and produce then a functional BRCA2 gene. And this is well now recognized to happen, and we’ve seen it happen in prostate cancer as well. It’s called a reversion mutation.

So, that’s why you could check again in a patient at resistance, to see based on the tumor sequence, has it actually reverted back to a wild type and then competent tumor cell. So, that is a pretty common mechanism of resistance.

Tom:

It looks like treatment doesn’t work forever and it looks like resistance does occur in the majority of patients. Other mechanisms such as efflux pumps have been thought to be important. Can you see dynamic changes in LOH associated with therapy and is LOH a better marker of response and resistance than a panel of genes that may or may not have double breaks?

Pete Nelson:

Right. So, that’s a great question. I haven’t seen LOH dynamically change. It’s hard to reverse LOH, for example. You’ve got these very large alterations in the genome of that tumor cell with big copy losses and copy gain. So, it’s pretty difficult to reverse that and that’s why you really need some future functional assay of ongoing competence, which is where these RAD51 based assays may be relevant.

You’re right, there is some work trying to look at other resistance mechanisms and these efflux pumps may certainly be one. There’s some other kind of acquired alterations. There’s another gene called schlafen-11. Again, this gets pretty complicated, but schlafen-11 has also been looked at as a response or resistance metric to PARP inhibitors as it modulates how PARP activity is mediated on the DNA.

So, stay tuned, but you’re exactly right there. Most patients do fail. Either due to reversion mutations, for example, or other mechanisms that we still don’t yet fully understand. There are some exceptional responders though, which is obviously very gratifying for this type of precision medicine.

Brian:

Pete, can you talk about misclassification with clonal hematopoiesis with CHIP and how that, especially in prostate, with an aging population. How does that work?

Pete Nelson:

Yeah, so that’s another really important question. Particularly if somebody’s using a blood-based assay of circulating tumor DNA assay. So, as we get older, we accumulate what’s called clonal hematopoiesis and that is in the bone marrow, you get clonal production of cells that may carry mutations and you can detect them because these are clonal. That means they’re coming from one original cell and a number of these turn out to be involved in DNA repair. ATM is a very common one. You can see mutations in ATM in lymphocytes, for example, or in the circulating DNA that’s not coming from of your tumor, but actually coming from these …

Brian:

I see.

Pete Nelson:

… premalignant or more benign cells. So, it’s a really cautionary note that they’re getting better. The answer to that is, you do a simultaneous germline assay or a lymphocyte assay and unfortunately there are a number of companies that don’t do that routinely, but that would readily identify whether that mutation is from clonal hematopoiesis or from your tumor.

Tom:

Brian, it sounds like you know what you’re talking about here. I’m very impressed.

Brian:

I think, I don’t know if it’s that it’s a new year, but our questions are definitely better on this podcast.

Tom:

[crosstalk 00:20:47] They are.

Brian:

I’m not sure.

Tom:

[crosstalk 00:20:48] Listen, I’m going to bring it down a little bit. I’m going to bring it down a bit if I may, Pete. This weekend, I went to see Lord Nelson’s flagship at Portsmouth Harbor, and he sailed his fleet into the Spanish and French fleet in 1805. And he went in very slowly and essentially what happened, is he did something called, cross the T and he split the Spanish fleet into two. I can see that part working in kind of a similar way bizarrely, but the point I’m trying to make, I think, is that this has crept up very slowly. PARP inhibition has been around for a long period of time and now we have 2 randomized phase 3 s coming out on the same day.

What has changed to make this thrust into prostate cancer? Because the ovarian cancer doctors have been using the drugs for some period of time. Could you give us a background of the data we’ve had to date and why things have been moving so slowly?

Pete Nelson:

Well, one could argue slow, one could argue fast. We learn a lot from those cancers that have a predominance of these alterations, which ovarian cancer, it was one of the malignancies extremely responsive to platinum. So, it was pretty natural for the evolution of therapeutics to follow that platinum sensitivity and then it was recognized, this platinum sensitivity is because, often of homology directed repair, whereas in prostate cancer, platinums written large have not proven to be very effective across the board, yet there are subsets of patients that we do treat with platinum therapy.

So, I think the ovarian field was primed for that, same thing in triple negative breast cancer. They were moving very quickly having therapy. Since they couldn’t use endocrine therapy in those cancers, they were quickly looking for other alternatives. Whereas in prostate, we’ve been read to the androgen receptor focus for decades.

So, we came late to the field in that area, and it wasn’t recognized until, within the last 10 years, that there was also this heritability around BRCAness in prostate as well. So, I think once that was recognized, then the field moved pretty quickly.

From the first, there were a couple prostate cancers in the Olaparib phase one that Dr. Debona led, that showed exceptional responses. So, there was a lot of excitement. The question was, what the best biomarkers were, because if you restricted it early on, for example, just to BRCA2, it’s going to be hard to accrue those trials and so it moved into these larger biomarker panels and things.

So, I’ve been excited by the pace of it more recently with at least 2 inhibitors approved now and probably more to follow. I think the major question is, which these trials that you alluded to are going to start addressing, which is both problematic and then opportunity, is broadening the use of PARP inhibitors.

So, right now it’s pretty selective and arguably not selective enough with these panels of markers that we just talked about. These 15 markers for Olaparib are probably overestimating that group of patients likely to respond. So, the question in the field is, is there something else that could be used to together with a PARP inhibitor to broaden their effectiveness and broaden their use?

Brian:

So, Pete, my last question and thank you, this has been amazing. If you were the king of developing PARP inhibitors in prostate cancer, you have unlimited resources, what would you prioritize? Is it better biomarkers? Is it resistance mechanisms? Is it combinations? Where would you put the resource for the most clinical benefit?

Pete Nelson:

I think for the biggest patient benefit, it would be the biomarker. You don’t want to treat patients needlessly that aren’t destined to respond. It is expensive. There are side effects. So, I would definitely put resources there, but you’ve hit the other 2 areas as well that are critical. I do think there are a lot of opportunities that PARP inhibitors open up for combination therapy and then since you do see these fantastic responses in some patients, but they’re not durable. Is there some way to extend their utility? So, I would put, just as you’ve said, I’d put resources into all 3 of those camps, because I think any of them will make a big impact in the field for patients.

Tom:

Pete, the previous study to Olaparib and rucaparib with single agent trials. One was randomized, one was a single agent study. What sort of values and these trials propel and magnitude are combining with abiraterone? Is abiraterone as good a partner as any other out there? And what PFS results, what benchmark results did we see with PROfound and Tyson and what can we expect, without obviously talking about the results of these upcoming trials? What results are we expecting? Are we expecting overall survival of 25 months? Put us in a ballpark figure of what the single agents do and what you’d expect for the combination?

Pete Nelson:

Yeah. So, one of the challenges with PROfound, it was a very advanced patient population. So, I think what we’re probably going to be seeing in the future, both with these trials that will be reported, but others, is using them much earlier. So, there’s no real reason to use these, for example, after docetaxel. After an ARSI after, docetaxel, they should be very effective and potentially even more effective early in the disease course. So, I would be using them early.

So, your question about abiraterone. There is some evidence that, and this is the rationale I think, that led to the trial, but I honestly don’t know. I wasn’t involved in the designs of the trials and the rationales for abiraterone.

There is evidence that the androgen receptor is involved in DNA repair in a couple ways. One, it does look like the androgen receptor directly regulates, turns on or represses a number of the genes involved in homology directed repair. And there is some evidence with androgen deprivation, just ADT, where you could imagine further inhibiting AR signaling with abiraterone or enzalutamide or others, that you then repress DNA repair.

So, if you now have something that, either in combination represses DNA repair, i.e., PARP inhibition, plus an AR pathway inhibitor, or induces damage in the context of repressing repair, that could improve outcomes. So, I would guess that the trial design, the abiraterone component was considered an approach to repress DNA repair by repressing AR signaling and then you get the added benefit of the PARP inhibition.

Now, the counter is also true, which is really fascinating. There’s a trial that Mike Schweizer led here at the University of Washington, of using super physiologic androgen. So, super physiologic testosterone, a group at Hopkins and others and there’s a lot of preclinical work, also actually induces DNA strand breaks. So, the androgen receptor and the estrogen receptor, when they initiate transcription, they need to cut DNA. We need to uncoil it to allow the nuclear hormone receptor to transcribe.

So, super physiologic androgen has been shown to induce DNA damage and there was a small 30 patient or so, trial of using high dose androgen plus a PARP inhibitor, Olaparib, that showed pretty remarkable outcomes. It’s yet to be moved forward into any sort of phase 3. But I think what you’re seeing probably here, is some way to broaden the use of the PARP inhibitor.

I’m not aware that these trials that are going to be reported, were biomarkers selected. I think they were not and so that’s the idea behind, can you take advantage of the PARP inhibitor, by combining it with another pathway and the logical sense here, is repressing DNA repair by repressing an AR signal. If that makes sense? [crosstalk 00:30:06]

Brian:

[crosstalk 00:30:06] Yeah. I think we probably should wrap up, but this has been amazing. Thanks for the education around current use, but also where we’re going, and we look forward to the ASCO GU results. Tom, anything else?

Tom:

Yeah Pete, it’s been magic. Thank you so much. I’ve really enjoyed it. You definitely sound like you know what you’re talking about.

Pete Nelson:

You led me on. Those were great questions.

Brian:

Thanks a lot, Pete. Appreciate your time. Take care.

Tom:

All right. Thanks very much. See you soon, I hope.