Read the Full Video Transcript

Andrea Miyahira: Hi, everyone. I'm Andrea Miyahira at the Prostate Cancer Foundation. Joining me today is Dr. Johann De Bono of the Institute of Cancer Research and the Royal Marsden NHS Foundation Trust. He is sharing a recent paper from his group elucidating acquired PARP inhibitor resistance in advanced prostate cancer that was published in Cancer Cell. Thanks, Dr. De Bono, for joining us.

Johann De Bono: Pleasure, Andrea. I would like to present to you today the data that my team acquired over a decade of work studying in advanced prostate cancer sufferers, mechanisms of PARP inhibitor resistance, specifically focused on the olaparib investigational trials that my team ran in the TOPARP-A and TOPARP-B trials.

In these trials, from the outset, we pursued the pharmacological trail and collected serial plasma samples, as well as circulating tumor cell samples through EpCAM-based capture pre-, during, and at the end of therapy. These allowed us to pursue studies of that plasma DNA by deep whole-genome sequencing as a baseline at progression, as well as serially by targeted sequencing and by low-pass whole-genome analysis.

We also were able to develop assays to FACS sort circulating tumor cells to single-cell purity, amplify the DNA from these single circulating tumor cells, and evaluate these cells by low-pass whole-genome sequencing serially during therapy. And this is particularly important for our studies on the BRCA2 HomDel tumors.

Our main message overall from this paper is that in metastatic prostate cancer with BRCA2 and PALB2 alterations, where you see the most activity with these drugs (with PARP inhibitors), the majority of tumors, almost all of them, show evidence for return of function of these key homologous recombination genes.

In the BRCA2 PALB2-mutated tumors, many of which are germline but not all—some are purely somatic. We see reversions that restore function of these genes and restore homologous recombination repair, with very robust evidence of POL theta microhomology-mediated repair. Some would call this POL theta–mediated repair, suggesting that therapeutic strategies targeting and blocking POL theta may delay emergence of this resistance in tumors with BRCA2 and PALB2 mutations, both somatic and germline.

I should say that the BRCA2 HomDel tumors respond for a lot longer, as we published in Cancer Discovery some years ago, and take longer to acquire resistance. And to our surprise, these tumors progressed with a return of BRCA2 function. The BRCA2 gene was restored in these tumors, having been BRCA2 HomDel at baseline.

And this seemed to be largely, in most subjects, due to clonal evolution of a rare subclone that had not deleted both BRCA2 alleles. Although the question still remains whether in some tumors these cancer cells are acquiring the BRCA2 gene by fusion with normal cells. Over really this last decade, we have, through work from my group, proven that PARP inhibitors improve outcomes from prostate cancer.

Our first paper was published in 2009 in The New England Journal of Medicine for the phase I study of olaparib, where we showed responses in BRCA2 germline patients, including mCRPC subjects with BRCA2 loss. This led to that phase II study called TOPARP—TOPARP-A and TOPARP-B—and then to this randomized phase III trial called PROfound in mCRPC that was either treated with one line of chemotherapy or no chemotherapy in patients who had at least one next-generation AR-targeting drug or RP. And this changes standard of care.

But our question here today is, what causes resistance to PARP inhibitor therapy? And this is particularly important. Why is it important? Well, it's important because if the tumors progress with continued DNA repair defects, then arguably a better PARP inhibitor or a platinum-based therapy should be administered. And this is a really key question.

And it has, for example, been mooted that resistance to olaparib can be driven by, for example, P-glycoprotein increased expression, which can pump these drugs out of these cells and cause resistance, although in our studies we rarely see overexpression of Pgp in these tumors in CRPC biopsies.

So in this trial, we collected biopsies serially pre- and post-therapy and at progression when possible. We collected serial ctDNA from all the subjects—baseline, during therapy, and at progression. And as I said, we analyzed more than 3,000 single circulating tumor cells purified to single-cell purity, with low-pass whole-genome sequencing done on over 3,000 single cells.

We demonstrated, to our surprise, that in BRCA2 HomDel tumors—which at baseline were almost exclusively, by deep whole-genome sequencing, homozygously deleted—that actually at progression, in all but one subject, we saw restoration of at least one allele of the BRCA2 genes, which was a great surprise to us.

You'll see that in one subject, the first subject on the left, 1,115 days on trial, that this subject does not have, at progression biopsy, restoration of that BRCA2 allele. And that's an important point that we're still trying to understand. But in all the other subjects, we see restoration of that BRCA2 gene.

And here you see, in one subject, subject 23, each row representing one CTC. We have overall 52 single circulating tumor cells, low-pass whole-genome sequencing here. And you see that at baseline, if you look at chromosome 13, you see that the cells—chromosome 13 at baseline—most of which have deep deletion of the BRCA2 allele (amplified down here at the bottom of chromosome 13), BRCA2 being here. Most (but not all) the cells of BRCA2 did show homozygous deletion and deep blue. And at progression, most of the cells that are coming out at progression clearly have restoration of at least one allele of BRCA2, confirming that return of BRCA2. And this was confirmed by RNAish and FISH. And you see that at baseline in the biopsies, you see very little red probe for BRCA2, and at progression, the red probe is emerging in the tumor cells, in keeping with restoration of the BRCA2 gene.

And actually, if you look at the RNAish, you see minimal RNAish BRCA2 staining at baseline, but at progression, you see many tumor cells restoring BRCA2 expression. And really, this was a surprise to us, although perhaps not all the cells restore BRCA2. And this is something that still needs to be evaluated in more detail.

In the BRCA2 PALB2-mutated tumors, which progress much more rapidly than the BRCA2 HomDels, PARP inhibitor resistance is really related here to reversion mutations. And what was really important in this study is that for the first time, we show that in some subjects, you have these, surprisingly, these BRCA2 reversions before the patients ever get a PARP inhibitor.

And this is likely because these patients have had previous DNA-damaging agents like radiation therapy that clearly may cause BRCA2 reversions as a mechanism of resistance for the tumor cell. But what you see in this cartoon in this graph is that the patients get different BRCA2 reversions. Some get them very rapidly, some don't get them at all, as in this green or dark green subject here who is on trial still at 30 months without progression.

But you see that in this patient here, you get early increases in reversions, and this subject comes off trial very, very quickly. And in many subjects, by six months or so, you're starting to see reversions. And these generally are essentially restoring BRCA2 expression and putting BRCA2 back in frame.

And what's particularly interesting is that the number of reversions, and the first identification of reversion of BRCA2, really define the rate of progression—so that the more reversions we see, the more quickly the patient's going to progress. And the quicker we see reversions, the shorter the time to progression in these subjects. And that's really the first demonstration of that prospectively.

And importantly, for future therapeutic studies, the majority of these reversions, we are convinced, are being driven by POL theta–mediated microhomology end joining. And this is very important because it identifies an enzyme that, if we block this by emerging compounds that target POLQ—either at its DNase or its RNase domain—these POLQ inhibitors, when combined with PARP inhibition, may delay the emergence of these resistant mutations or even prevent them and maximize duration of benefits on PARP inhibition with a POL theta inhibitor.

And trials are now ongoing with multiple POL theta inhibitors with PARP inhibition to try and confirm this hypothesis being real, which has been shown preclinically in a number of studies. So in conclusion, we have shown that PARP inhibitor resistance almost always involves restoration of homologous recombination repair. And this is important because it validates that the primary mechanism of action of PARP inhibitors in prostate cancer is actually killing tumor cells by synthetic lethality, not through AR blockade, et cetera.

But clearly, more data are required to elucidate if subclones remain that don't restore HR (homologous recombination repair) but retain the HR defects. And maybe these subclones get resistance through other mechanisms, and that needs to be elucidated. But we have looked hard for other mechanisms of resistance in deep whole-genome sequencing and plasma and in the single-cell analysis, but we have not found any other mechanisms to date.

Although we are still studying that, actually, and we have some exciting data that we'll hopefully share in the near future that may shed more light on this space. We've also shown that restoration of HR in these patients is likely to cause resistance to further PARP inhibition or platinum. So if the patient is progressing on, say, olaparib, there probably is no or limited scope for benefit with, say, talazoparib.

Although we'll have to see in studies—we'll have to really show that. And we have previously published for ovarian cancer that actually, when the patients progress on olaparib, the response rate to platinum is much, much reduced, especially if they're refractory to platinum at platinum progression and they're progressing on—sorry, on PARP inhibitor progression.

So if they're progressing while on the PARP inhibitor, they're really quite unlikely in ovarian cancer to respond to platinum. So they're truly PARP inhibitor–refractory; you're unlikely to get platinum sensitivity. And similarly, if you're platinum-refractory, you're not going to respond to PARP inhibition. So really, it's quite important. If you're really progressing on PARP inhibition, you are unlikely to benefit on platinum as a next therapy, although some patients may do, and that needs to be elucidated further.

And finally, combined PARP inhibition and POL theta inhibition may block this microhomology-mediated end joining, prevent reversion mutation generation, and delay resistance emergence to increase benefit from PARP inhibition, which is what we hope to achieve. And finally, I would say that the emerging reversion mutations may actually result in new therapeutic strategy sensitization. And work is ongoing to really evaluate that further. So thank you for your attention.

Andrea Miyahira: Thank you so much, Dr. De Bono, for sharing this study. Have you looked at recurrence patterns by imaging in patients with homozygous BRCA2 deletions? Or do you know if recurrence only occurs in some lesions or in all lesions?

Johann De Bono: So Andrea, thank you for that very important question. So the first thing to say is we have published that PARP inhibition can, in some patients, induce mixed responses—i.e. some lesions improving and some not improving—in the same patient. And this is actually seen with really almost all the drugs we have in the clinic. You see that with RPs, lutetium PSMA, taxanes, in keeping with the concept of heterogeneous disease in the patient.

I think with PARP inhibition, every patient is different. There are some patients who progress in only a small number of lesions, and some patients who progress more broadly. And I think efforts need to be pursued to really evaluate this further. But the beauty of the blood biomarker studies is that you're really studying, we hope, what's happening across all the metastases, and that is the advantage of blood analysis over biopsy studies.

Andrea Miyahira: Thank you. And does PARP inhibition or other treatments like radiation therapy, as you mentioned, force the BRCA2-alternate DNA repair pathways like POLQ to act abnormally in the BRCA-deleted tumors?

Johann De Bono: Correct. So that's the hypothesis driven by a number of groups, and amazing work by many groups—Alan D'Andrea and Dana-Farber and others—that is really driving the study of combinations of POL theta inhibitors or PARP inhibitors. And I think this really does merit evaluation, although obviously we do worry about increased toxicity with these combinations. But certainly, these studies are ongoing. So watch this space.

Andrea Miyahira: OK, thanks. And you saw that there were some patients who became resistant to PARP but did not have the reversion mutations. So do you know what other factors might be contributing to resistance or allow continued survival and growth of tumor cells that have unstable or unrepairable DNA?

Johann De Bono: So what has surprised me is the convergence of resistance mechanisms. I expected to see many other mechanisms of resistance emerging. But in almost all the patients, bar one, we have seen essentially restoration of HR as a mechanism of resistance.

Now, that surprised me—that actually it's really a restoration of the BRCA2 or PALB2 genes. Preclinically, particularly in mouse models, many other mechanisms of resistance have been described. And we've been looking at these really hard with colleagues like Chris Lord, who is really an expert in DNA repair here at my institution, and first described that synthetic interaction between PARP inhibition and BRCA loss with Alan Ashworth 20 years ago. But what I can say to you is that, to date, in prostate cancer patients, we are not so far identifying other mechanisms of resistance, which was a surprise to me.

Andrea Miyahira: And you mentioned that using a POLQ inhibitor in combination with a PARP inhibitor may be a possibility. So how would you use this strategy, and how would you determine if a reversion mutant subclone is already pre-existing?

Johann De Bono: So if the subclone is pre-existing, then it is highly likely that the patient is less likely to benefit from a POL theta inhibitor. Although please note that actually the number of pre-existing subclones with a reversion was very, very small and uncommon. So I don't think that's going to be a major issue.

I should also point out that a POL theta inhibitor is not going to benefit the BRCA HomDel subjects. It's only going to benefit the patients with a BRCA2 or PALB2 mutation, or in other cancers BRCA1 in ovarian and breast mutation. So it is our hope that by blocking POL theta, if we can safely deliver the two drugs together, PARP inhibition and POL theta inhibition, that we will delay time to progression and show that there's a much longer duration of progression-free survival.

Now, I'll be honest, these studies are going to be challenging because the two drugs in combination may be more toxic. As you know, PARP inhibition has a limited therapeutic window due to particularly hematological toxicity, anemia, et cetera. And the other worry I have is that we are starting to see myelodysplasia and acute myeloid leukemia emerging in these trials.

And I think particularly in subjects who have been on these drugs for two years or more, this is going to become a major concern for me. We need to be keeping our eyes on this. And please note that the ongoing PARP inhibitor phase III trials are not pursuing, once the patient comes off study, the capture of the AML and myelodysplasia toxicity late on.

So we may be underreporting these toxicities. But certainly, in ovarian cancer, there is concern that 5% or more of subjects with ovarian cancer are getting myelodysplasia and AML post-PARP inhibition. But nonetheless, this is a very important question: Can we delay progression with this combination? I think if we do, I'm very confident we can actually demonstrate patient benefit by the combination of a PARP inhibitor alone, not only by impacting PFS, but also by impacting OS through randomized trials comparing PARP inhibitor alone and the combination.

And what this paper does that we've just published is it tells us generally when we'd expect these reversions to emerge. And if we can show in a non-randomized setting that the combination shifts the emergence of reversion or prevents them from coming out in that first six months, then I think we can be fairly confident that we're going to impart benefit.

Andrea Miyahira: OK, thank you. And what are your next steps for this?

Johann De Bono: Well, we are running these combination studies—POL theta with PARP inhibition. So that's going on. We are trying to figure out how these tumors that restore BRCA function develop other new therapeutic vulnerabilities, which I think they do. And that work is ongoing in my lab.

And we are trying to really innovate in pursuing new therapeutic avenues with regards to other vulnerabilities—for example, B7-H3 targeting, HER3 targeting, PSMA targeting—with regards to the uncommon switch, the neuroendocrine disease, which in my experience, having studied more than 1,000 CRPC biopsies, is not that common (less than 10% neuroendocrine disease). We’re looking at ways to try and push neuroendocrine disease back to luminal and AR-dependent disease. And I hope that strategies will emerge that will allow us to do that. We're also very excited about trying to target inflammation to reverse resistance. So we'll watch that space.

Andrea Miyahira: Well, thank you so much. I look forward to the results from those studies. And thanks for sharing this with us today.

Johann De Bono: Pleasure, thank you so much.