Panelists: Joe OSullivan, MD, FRCPI, FFRRCSI, FRCR, The Northern Ireland Cancer Centre, Belfast City Hospital; Johann de Bono, PhD, MB, ChB, Institute of Cancer Research, Royal Marsden Hospital; Chris Parker, MD, FRCR, MRCP, Institute of Cancer Research, Royal Marsden Hospital; Bertrand Tombal, MD, PhD, Cliniques Universitaires Saint-Luc
Joe O’Sullivan, MD, FRCPI, FFRRCSI, FRCR: The other interesting part of biology, Johann, is the androgen receptor biology. It’s been an important part of prostate cancer therapy for as long as we’ve been treating it. Do you think the earlier use of AR-targeted therapy will drive early resistance? What are your thoughts on that topic?
Johann de Bono, PhD, MB, ChB: I think, first of all, there are clearly many different subtypes of prostate cancer with different risks of that resistance emerging. There is incontrovertible evidence that there are resistance mechanisms emerging with AR copy number gain, the gene copy number gain, the PCR expression, AR mutations, and AR rearrangements. So, the gene is fused, rearranged. It’s changed. You only express splice variants. I suspect there are going to be other aberrations, too. These are clearly going to emerge, particularly after giving abiraterone or enzalutamide, next-generation hormonal agents. By giving these drugs earlier, we are likely to get these resistance mechanisms earlier, particularly in cancers with that genomic instability, that intercellular intra-patient heterogeneity. It’s a concern for me that as we give these drugs earlier, I’m seeing a lot of patients with modest disease volume having all the available treatments very early on. It concerns me that that is not necessarily a good thing for our patients. I don’t know if, Bertrand, you agree with that.
Bertrand Tombal, MD, PhD: No, I’m even worried. I recently did a clinical trial using AR, and people even use that in very, very early onset disease, even prevention trials. But we should keep in mind that there are extremely potent drugs in terms of inducing AR rearrangement in general.
Chris Parker, MD, FRCR, MRCP: Johann, do you have the same concern with conventional, traditional androgen deprivation, or is it just in relation to the new drugs, the abiraterones and enzalutamides?
Johann de Bono, PhD, MB, ChB: The traditional LHRH analogs were not blocking AR signaling that well, and it’s really quite a shame that we took so long to figure that out. I think the concern in some patients, even within weeks of castration, is that androgen receptor splice variants are clearly emerging. Work from Martin Gleave and others has shown that within weeks of castration, you have AR splice variants. The emerging data would indicate, and we’ve just published a paper on this, that AR splice variants upregulate DNA repair, particularly nonhomologous end joining, and that causes radiation resistance. So, our current strategy of giving these drugs before radiation may be indeed the worst thing possible to generate splice variants that cause resistance and is seriously a major concern until we generate other methods to block non-homologous end joining: for example, by inhibiting DNA-PKc. We’ve shown in our paper that if you give DNA-PKc inhibitors, you actually block that repair and you photosensitize the radiation despite the presence of AR splice variants.
Chris Parker, MD, FRCR, MRCP: On the other hand, we do have large randomized trials comparing radiation alone versus radiation plus ADT, showing significant survival benefits for the combination.
Johann de Bono, PhD, MB, ChB: The survival benefits are there, but I can assure you that they’ll be even higher if we can block those splice variants and that DNA repair process, based on the data emerging.
Joe O’Sullivan, MD, FRCPI, FFRRCSI, FRCR: Do you think the issue is the neoadjuvant part of the hormone that comes before the radiation?
Johann de Bono, PhD, MB, ChB: That’s certainly a concern. If you give neoadjuvant hormone therapy, and many groups have shown this, very early on, in 15% or so, you are getting those splice variants expressed very early. Then, radiation is less likely to work in those cancers that are AR splice variant expressing and are really driving that repair process that causes radiation resistance.
Chris Parker, MD, FRCR, MRCP: There was an RTOG phase III trial that looked at radiation plus short-course hormones, and they were randomized between the timing of the hormones. So, half got neoadjuvant and half got subsequent ADT. Actually, the neoadjuvant strategy seems to be better empirically. That was the evidence from RTOG-9413.
Joe O’Sullivan, MD, FRCPI, FFRRCSI, FRCR: That’s right.
Johann de Bono, PhD, MB, ChB: That makes sense because, actually, resistance mechanisms depend on population statistics, and the less cancer cells you have left to treat by radiation, the less likely you’re going to get resistance. But that does not denigrate the argument that splice variants are actually causing resistance. In fact, this is not surprising. We’ve already shown that blocking AR increases survival, but we’re not blocking AR enough. We know that and that AR splice variants can emerge very early on, and this is a serious concern for our current paradigm.
Joe O’Sullivan, MD, FRCPI, FFRRCSI, FRCR: What do we know about neuroendocrine differentiation? Is there a biological switch that happens with neuroendocrine differentiation?
Johann de Bono, PhD, MB, ChB: First of all, the word neuroendocrine makes me somewhat nervous, and the reason I say this is that we have looked at over 300 metastatic castration-resistant prostate cancer biopsies. We’ve done a number of neuroendocrine markers, CD56, synaptophysin, chromogranin, we’ve looked at AR downregulation. We’ve looked at cadherin E and N. We don’t see much evidence of true neuroendocrine differentiation, although it does depend on how you define your endocrine. I think it would be fair to say there is still significant confusion as to what the neuroendocrine is. However, it is quite evident now based on 2 important science papers from 2 different groups earlier this year that there actually is evidence that you develop RB1 loss. It can be monoallelic or bi-allelic loss of RB1, and it usually follows the earlier p53 mutation and loss of function of p53. That double loss of p53 and RB1 is generating this luminal to basal transition with decreased AR expression and a more basal phenotype. I’m not sure we should call this neuroendocrine. I think it’s probably a switch in cell phenotype, but sometimes, you are correct, it does become neuroendocrine.
Now, what’s particularly interesting here is that RB1, which is this late subclonal hit, is a gene that sits next to BRCA2. These are 2 closely opposing genes on chromosome 13. There are gene co-sets. In our data, which was published in Clinical Cancer Research recently, we’ve shown that many of these cancers, in fact the majority, lose BRCA2 and RB1 at the same time. It’s a subclonal late loss. And therefore, that explains why when you get this “neuroendocrine,” if I can call it that, phenotype, you are seeing some platinum sensitivity. But it is subclonal, and when you do RB1 and immunohistochemistry—we’ve done a lot of this—you see islands of RB1 loss and a lot of cells that still have RB1, so this is truly subclonal.
Bertrand Tombal, MD, PhD: Many centers have tried to look systematically at these panels of genes in what we call Gleason 9, Gleason 10, or very aggressive disease. Actually, if you look at these high-risk localized diseases with very, very high Gleason scores, what we call neuroendocrine is many times a subset where if you simply do AR staining, they’re not AR, and they don’t simply express staining. That’s where you find all these aberrations. Is it neuroendocrine? Or, are they simply not AR-dependent, which is not like lung cancer? But if you look at these patients in various subsets, and for surgeons who look at Gleason 9, you’re going to find these patients on a regular basis.
Johann de Bono, PhD, MB, ChB: What’s quite important for us to understand is that it’s really like playing snooker. You can pot all the red balls and you’re left with the colored balls, but actually it’s more complicated because if you pot the colored balls, the red balls can come back. Chris and I are treating a patient who is getting therapy that is probably clearing his neuroendocrine BRCA2-type disease with platinum. What’s fascinating is that he had a very low PSA before he got therapy, and when you gave him platinum, actually, his cancer recurred with a very high PSA. Before, he always had a low PSA, indicating that we’ve now cleared the “neuroendocrine,” if you want to call it that. I don’t think we should call it neuroendocrine. And more AR different clones come back. That really drives the message that we have to be thinking about combination therapy.