Read the Full Video Transcript

Charles Ryan: Hello and welcome. We are going to have another conversation today in our series on, "The Big Provocative Questions In Prostate Cancer." And I am delighted to be joined by Dr. Nima Sharifi, who is a friend, and a colleague, and Director of the Center for GU Malignancies Research at the Lerner Research Institute in Cleveland, Ohio, based at the Cleveland Clinic. And Nima is going to talk about targeting the androgen-androgen receptor interaction and why it remains more important than ever. So, we're going to listen to his talk and then have a little question and answer after that.

Nima Sharifi: Great. Thank you, Chuck. Thanks for having me on. Okay, so I will talk about whether or not we can make more progress on targeting the androgen-androgen receptor interaction. You could break down the mechanisms of resistance in a variety of different ways. This is one way to do it. Obviously, you have the potent androgens, testosterone, and DHT that interact with the androgen receptor here on the left.

The first category of tumors that are progressing after castration is those that are androgen-dependent. And if you are androgen-dependent, then you must be androgen receptor-dependent because that is how androgens work. And the types of things that fall in this category are drivers that originate from adrenal androgens or adrenal precursor steroids, as well as other extragonadal androgens. What do I mean by that? For example, de novo androgen synthesis from cholesterol; these are mechanisms that really have to utilize enzymes to convert these precursor steroids to potent androgens like T and DHT.

The second category are tumors that become castration-resistant through androgen agnostic pathways. They don't care about the presence or absence of androgens yet they are driven by the androgen receptor. So, what do we mean by these specifically? Well, there are these androgen receptor variants that are truncated and are on all the time, regardless of the presence or absence of androgens. There are also mutations in the AR ligand-binding domain that are stimulated for example, by glucocorticoids. So they don't care about the presence of androgens. And there are other ligand-independent mechanisms, phosphorylation to be one of them.

The third separate category are those tumors that become resistant by way of mechanisms that are androgen agnostic. So they don't care about the presence of androgens but are also androgen receptor agnostic. These are things like neuroendocrine tumors and other tumors that are AR expressing but they are driven by another stimulus, meeting molecular drivers from alpha genes.

Now, I think it's important to mention this gene, HSD3B1 that encodes for this enzyme, 3-beta-HSD1. And it's important because it is required for all pathways to potent androgen synthesis, from adrenal precursors and de novo steroidogenesis. There is a very common germline variant with the protein-coding change that differs in frequency by race. And this falls into different categories of physiology. The first one is the adrenal restrictive allele. And the second is the adrenal permissive allele. I'll explain what these are. The adrenal restrictive allele is this 1245A sequence that encodes for a rapidly degraded protein. So you have low levels of protein and the conversion from precursors to potent androgens is very slow as a consequence. So you get low levels of DHT synthesis from non-gonadal or non-testicular precursor steroids. And so that is why it is adrenal restrictive because it limits conversion of that.

The adrenal permissive allele is 1245C. So the functional consequence is that you have a stable protein and you enable conversion from non-gonadal precursors to DHT. So this leads to high levels of androgen synthesis. So graphically, how do we show this? Well, adrenal restrictive is something like this where you basically pulled all the precursors back behind the dam and very little dose forward to potent androgens. And adrenal permissive is when the dam basically opened up the flood gates and you get the more rapid conversion to DHT synthesis, which could lead to more rapid progression to castrate-resistant prostate cancer. So you might see these are nice graphics, but really what are the data? What are the clinical data?

And this is just a summary of the clinical data from at least eight cohorts of patients with castrate-sensitive prostate cancer. And the basic nutshell summary is, that when you inherit two copies of the adrenal permissive allele, this is about 8-10% of the general population. You progress more quickly on castration and in some of these studies that are powerful overall survival, this does decrease overall survival. When you do not inherit any copies of the more active or adrenal permissive allele, progression-free survival is much longer. And then if you inherit one copy of your heterozygous, you're typically somewhere in between these two. And that's fairly consistent across all these studies.

So how does that fit into these categories of mechanisms of resistance and how should we think about that for the overall topic at hand for today with Dr. Ryan? Well, HSD3B1 is certainly an enzyme and it falls under this first category because this would drive castration-resistant prostate cancer through mechanisms that are both androgens as well as androgen receptor-dependent. And then the second category, I think a lot of us recognize that these variants may be important for driving castrate-resistant prostate cancer as well as AR amplification.

What happens to these castration-resistant prostate cancer tumors in terms of clinical outcomes when these patients are treated with drugs like enzalutamide, which displaces androgens from the AR ligand-binding domain as well as abiraterone which blocks further synthesis of extragonadal androgens. Well, one might expect that in this first category driven by extragonadal androgen synthesis, that these tumors should respond well, at least as well as other tumor types, to drugs like enzalutamide and abiraterone, because these things block the action or the synthesis of the ligands that drive CRPC here. Now I'll tell you about the data in a little bit.

For the second category, there are data that already exists showing that when you have an increase in the copy number of androgen receptors genomically, you don't respond as well for some of these agents, in particular abiraterone. There is not a lot of clinical data out there, but there is enough of it out there. So these tumors do not respond as well. I think there is fairly clear evidence along those lines from cell-free DNA data.

So now let's go to the data for that first category, for HSD3B1 driven tumors. There is a couple of stories that came out last year in the Annals of Oncology. The first is, a collaborative study between the group at Hopkins and the ICR in London. And what it showed is that for patients who have two copies of the more activity of the adrenal permissive allele, they actually have, sorry, shorter overall survival compared with the rest of the cohort by about seven months. And overall survival from first-line ADT is about two years shorter, again, for these homozygous adrenal permissive patients. This is actually surprising because these tumors drive CRPC by a ligand-dependent or androgen-dependent process. And you are treating with that enzalutamide or abiraterone yet they are not doing as well. And there is a second study that accompanied this first study that was done with the group in Spain, as well as the Vancouver Prostate Center.

Now, the population is a bit different here and has a bit of a mixture of first and second-line abiraterone or enzalutamide treatments that were previously treated with docetaxel. And what they observed is that there was a decrease in PSA response rate but not overall survival. But all in all, these two studies together with about 800 patients suggest poor outcomes for patients who are homozygous adrenal permissive, once they are CRPC and then when they are treated with abiraterone or enzalutamide. So what is my interpretation here? One is that, interestingly enough, there is a plausible role for sustained steroidogenesis in a subset of patients who become resistant and have poor outcomes with abiraterone or enzalutamide. I think this could have value for the development of new therapeutic agents that targets androgen access in a way in which we have to select these patients very carefully. Because obviously if you're developing a new therapeutic that targets the androgen receptor access, you don't want to target, for example, neuroendocrine tumors.

And the second is that these homozygous adrenal permissive patients effectively don't respond well to multiple lines of hormonal therapies, not just medical castration. So the bottom line here is that HSD3B1 genetically regulates extragonadal androgen synthesis and this, in turn, dictates clinical outcomes after the amount of testosterone deprivation, multiple cohorts have shown that. And there are surprisingly worse outcomes for those who are homozygous adrenal permissive. And this is about 8-10% of the caucasian population. The major question here is, in my mind, are these poor outcomes reversed when you use these agents enzalutamide, abiraterone, or apalutamide upfront along with medical castration? Can we somehow target this particular mechanism? So I'll stop there.

Charles Ryan: Great. Well, thank you for that overview. And there is a lot of questions that come from that. And what I want to do is sort of focus on what, we as clinicians should think about these data. And my first take-home is that there is heterogeneity in terms of androgen production. And that may explain heterogeneity outcomes that we see. You've done some work that has looked at health outcomes outside of the world of prostate cancer with the adrenal permissive genotype, correct?

Nima Sharifi: Yes, yes we have.

Charles Ryan: And did you want to talk about that at all? I just find that interesting because it sort of speaks to the fact that this may be not just specifically looking at prostate cancer mechanism but may be related to something else as well, just general health too.

Nima Sharifi: Yeah. No, absolutely. So basically, and at this point, I could say probably it's more of a belief. We are accumulating the data that... because this basically regulates conversion from... So, if you asked the endocrinologist, "What does DHEA do in normal physiology?" We know that these adrenal precursors are involved in a variety of different pathophysiologic processes. What does it do in normal physiology? That's still not very well worked out. And actually, we do have another study published in asthma suggesting that these adrenal androgens play a role in inflammatory processes and actually can regulate anti-inflammatory response to oral glucocorticoids.

Charles Ryan: Exactly.

Nima Sharifi: So it seems a bit counterintuitive and probably requires a separate discussion. But I think it probably pinches on a variety of different physiologic processes, not just prostate cancer. And so that work is ongoing and I'll hope to have more for you probably.

Charles Ryan: I'm fascinated by that aspect of things because part of what I think about is your data has sort of this bi-directionality to it. In some ways when you say the flood gates are open and there's a lot of androgen being made, I'm thinking, oh, that's bad, because that means the cancer is going to grow. On the other hand, I think, well, that might create or foster a form of the disease that is very androgen-dependent. And that's something we can target.

The third bucket of prostate cancer is the one that we really want to avoid, which is the AR-independent and androgen-independent sort of form of the disease. And so the question is if we had a very restrictive androgen environment, is that which grows out going to be that third bucket, that very resistant disease? And so we've got really good tools, abiraterone, enzalutamide, et cetera, standard androgen deprivation therapy, but yet they are not good enough. And I want you to speak a little bit to that sort of almost paradox, which is to say, shouldn't those drugs work better in the 1245C patients, for example?

Nima Sharifi: Yeah, absolutely. And actually, that was my expectation. These are clinical data that open up entirely new possible mechanisms of resistance. So one of the questions is, okay, when you're treating someone with abiraterone, if they are homozygous 1245C, their responses aren't as good, is it because abiraterone gives you incomplete blockade of the adrenal androgens and extragonadal androgens? Is there a mechanism of androgen synthesis that is completely independent of CYP17? Or is this something else entirely? So I think the clinical data that I showed you with those two recent papers, I am hoping that we will get more of that in the coming months that can better define exactly what this is. And the other thing that I really want to look forward to seeing is not only data on HSD3B1, but we need to get a more comprehensive metabolic and molecular profile of individual patients.

What happens to these patients? Do you get ARG amplification in these patients or do you not? Are they mutually exclusive? Do you get PTEN, p53 mutations, or is that a completely different category? You might suspect that that could be a separate category but I think so far in the genetic profiles that we have, they are not well powered. In order to look at all these things on top of one another, you need extraordinarily well-powered studies, and then you want good clinical annotation on top of that. And now you're asking for a lot with clinical trial quality data, right?.

Charles Ryan: Yes.

Nima Sharifi: But I'm hoping that we can gather that as a community in the coming years.

Charles Ryan: Yeah. Some of the data are out there, whether it's sufficient to demonstrate the validity of the interaction of all of those variables is another question, but it is a fun question. I enjoy thinking through those questions with you. One of the things I have to say also is as I sit in my world of designing clinical trials, there is a whole army of people designing clinical trials for the very resistant disease that you just mentioned, the TP53s and the RBs. And that is certainly worthwhile.

At the other end of the spectrum, we have a number of people, and a number of trials, and a number of observations now about exceptional responders and patients. I've got people now who had metastatic disease on ADT and abiraterone, and then one year, two years, three years pass and they are doing really really well with a PSA of zero. And you wonder, were these, the patients where the flood gates were open and they had this entirely androgen-dependent tumor, and by cutting it off so well with the next generation drugs that we have, are we going to, get a nice return on that investment in time? And so it's just kind of interesting having these research questions on both ends of the spectrum.

Nima Sharifi: Yeah. No, absolutely. Thinking back to my sort of earlier years in terms of trying to figure out, okay, we see some patients who have, for example, a long response on castration and subsequent let's say short response or a long response on subsequent therapies prior to abiraterone or enzalutamide. And so there have been a number of sort of older studies. If you have a long response to castration, does it mean you're going to have a long response to the next agent, the secondary hormonal therapy, or does it mean the opposite? And so I think the answer to questions like that really depends on the mechanism that drives that castrate-resistant disease. You can have mechanisms that will lead to, for example, rapid CRPC and perhaps rapid resistance to the next agent.

And conversely, you can think up mechanisms. So this all depends on ligand dependence versus independence in part, and I'm sure there are other factors that we're not thinking about. So, as we look at things now, I think even now with the molecular data that we have from Stand Up to Cancer and other things, including metabolic profiling, I think our views are probably overly simplistic. And as we dig into this, it's going to become increasingly complex in ways that we can't even foresee. 

Charles Ryan: And your assumption through all of this is that all of this rubber hits the androgen receptor road. And then the question is, could other steroid receptors play a role in other things? So it's extraordinarily complex. It's extraordinarily interesting, isn't it? I mean, to be thinking through this and to be discovering the nuance of this mechanism that was first targeted, 75, 80 years ago is extraordinary. And you are really leading that conversation. So I thank you for your time and always a great opportunity to catch up with you about what you're thinking. And it's a great addition to our provocative question series. Thank you, Dr. Nima Sharifi.

Nima Sharifi: Thank you, Dr. Ryan. Thanks for having me.