Welcome everybody. I know some people were running from the very relatable Plenary session to now so thank you all that's how we're starting a couple minutes late welcome to our symposium This is steroidogenesis and androgen action sexual development across mammals You know rather than jump right in there's going to be a couple minutes of talk because this is also a symposium That's really honoring a giant in endocrinology. Dr.. Gene Wilson. You know there are There aren't that many giants that you can say that I think everybody will agree is just somebody who's contributed so much to the field That it's almost immeasurable, and I think Gene Wilson is one of those. He's really a leader in studying androgen actions He made he's made seminal contributions related to five alpha reductase and really a better understanding of how androgen signaling in You know functions, and how it relates to sexual development, so my name is Steve Thomas I am at the University of Rochester. I was formerly at UT Southwestern and and Gene was my mentor And this is Hans guy who was also at UT Southwestern and obviously Gene had a great impact on him as well and And all of the speakers here have a connection to Gene and a connection to UT Southwestern as it turns out So so what I thought I'd do is I'm going to give you one sort of brief memory of Gene Then there's something I'm going to read from From Marilyn Renfroe who was supposed to be here, but unfortunately couldn't make it and then we'll start So I've got a lot of memories of Gene some of them probably I can't say in public So I'll stick to the ones that I can which is that really what? One of the reasons I came to UT Southwestern was because when I was looking for a place to go I was studying a kind of an unusual signaling mechanism And there was a lot of debate about whether it's real or not and and I was a little unsure of myself And I just remember when I interviewed there Gene Wilson basically saying to me You know this is really cool and different We want to support you to do something crazy and unusual and why would you not do that right? He was the only person that I ever spoke to during my interview trail who just said do this crazy thing Let's see how it works. We'll figure it out And so I knew I was gonna go to UT Southwestern after that conversation with Gene And he helped he helped me make sure that it did work out. So that's my memory of Gene I want to read a quick statement from Marilyn and and then we'll move on from there. So Marilyn, uh, I'm just gonna read it basically So Gene Wilson first visited us in 1986 when he became interested in the potential of marsupials for the understanding of virilization As it took place well after birth while the young were still in the pouch however Our first study showed that circulating levels of testosterone and DHT did not drive virilization which led to the discovery of an alternate pathway Involving the metabolite 5-alpha andrastein diol Which was done in collaboration with several of Gene's UT Southwestern colleagues including Fred George, Cedric Shackleton and Richard Aukus Who you'll hear from later today. The pathway occurs in many mammals including humans and is now recognized as an explanation for several cases of abnormal virilization So, um, he's he left an amazing legacy for us with publications that still form the basis for many of our marsupial experiments There was just a small part of his worldwide seminal research on the endocrinology of DHT and testosterone He will greatly be missed all over the world valet gene D Wilson and that's from Marilyn Renfroe Renfroe and Jeff Shaw So with that we can move on to the first presentation Do we can switch off this slides these slides and go on to Dr. Nima Sharifi who we all know very well up here and down there and Nima is coming from Cleveland and I can't see the title of his talk quite yet There you go roll of androgen metabolism and prostate cancer welcome Nima and we're really delighted to have you here great Thank you. Thank you, Steve, and it's a real incredible honor to be Participating in this session in honor of Gene Wilson and thank you so much for arranging this So as mentioned, I'll be talking about the role of androgen metabolism and prostate cancer For my presentation feel free to take photos share whatever You'd like to do Here are my disclosures So this is just one of many Seminal observations done by Gene and this is really the classic 1968 paper in the Journal of Biological Chemistry where Gene Wilson and Nicholas Brzezinski identified DHT as the primary active androgen in the prostate cell nucleus So what does this mean for prostate cancer? so in sort of typical male gonadal physiology You've got as you know gonadal testosterone going into circulation once it hits the prostatic tissue It's converted by five alpha reductase to the much more potent DHT Where the androgen receptor is not bound to? Ligand it's typically most of it is in the cytoplasm when it binds DHT It gets into the nucleus and induces expression of hundreds of androgen responsive genes And there's a characteristic for example translocation that oftentimes allows androgens to directly drive oncogene expression and tumor progression So this is one way of viewing the biologic states of the disease and prostate cancer And I should say I'm not really alone in viewing it this way so when someone comes in with advanced disease or metastatic disease Circulating PSA and tumor burden oftentimes go hand in hand Usually the disease is driven primarily by gonadal testosterone stimulating the androgen receptor So AR transcriptional activity is on it drives tumor progression and increase in tumor volume and the standard of care now for about eight decades In part is is basically treating with medical or surgical castration Depriving AR of its primary stimulus from gonadal testosterone Leading to a decrease in AR transcriptional activity, so that's your therapeutic response you decrease tumor volume and so that typically lasts one to two years and Then eventually you get the development of castration resistant prostate cancer so AR turns back on This is transcriptional activity going back up And tumor progression occurs here, so what's really interesting is if you biopsy these tumors that are progressing on men absent gonadal testosterone You find that the level of biologically active androgens in the tissues itself are now elevated So this is an observation first made by Jack Geller in the 1970s People didn't believe them later studies were done by mass spec, and it's it's real and true right so these tumors take non gonadal precursor Steroids and make their own potent androgens, so this is a metabolic issue or metabolic question So now the question is how do we know that this is an important disease driver is this entirely correlative? Or does it truly drive bad outcomes and so we have a series of practice changing phase 3 clinical trials including with this androgen synthesis inhibitor abiraterone that blocks CYP17 and Adrenal androgens, and if you use that drug compared with placebo in men with castration resistant prostate cancer Even men who have tumors that are resistant to docetaxel chemotherapy you prolong overall survival So I'll say that just one more time tumors that are resistant to docetaxel chemotherapy Treated with adrenal androgen ablation that prolongs overall survival, and if you use this drug up front along with Castration in the setting of castration sensitive prostate cancer you get an even more profound effect on overall survival by about a year And in other clinical trials with these potent AR antagonists enzalutamide apalutamide That displace DHT from the AR ligand binding domain you get a comparable effect on overall survival so This intertumoral DHT generation is important. Hopefully you're convinced of that with these studies now The question is how do these tumors make? DHT so in 2008 I arrived at UT Southwestern in Dallas as an assistant professor And so I before then I was primarily interested in somatic genetic alterations that directly affect the androgen receptor But we also recognize that we need to look at how this how some of those alterations affect androgen metabolism So, you know, I'm a medical oncologist But had meetings with Gene Wilson, Rich Aukus, Mike McFall, Steve Hamas So I probably better belonged in the division of endocrinology metabolism than hematology oncology I certainly spent more time with the folks in endocrinology, but Gene said listen take my TLC plates, right? I'm not using them any longer. Why don't you look at this directly and I said terrific and we never published the reason for doing those experiments and the reason is that the controls didn't work as Expected that totally pivoted the direction of what we ended up doing. So let me tell you Why that's the case. So this is what was generally believed prior to that That from adrenal precursor steroids the DHA goes to androstenedione Androstenedione goes to testosterone. Testosterone is 5-alpha reduced to DHT following the classic gonadal pathway And what we found this is the actual data on on the TLC Androstenedione treatment you can see that there is certainly conversion to DHT. So you do get from here to here On the other hand when you treat this with the supposed intermediate metabolite testosterone Here up to 24 hours you get no DHT And the reason is that androstenedione Predominantly goes through an alternative intermediate metabolite 5-alpha androstenedione on to DHT So the primary route to DHT in castration resistant disease from adrenal precursors Circumvents testosterone and so we clinically confirmed this in fresh tissues from patients, etc So this is thought to be the dominant pathway in the human adrenal reticularis From cholesterol all the way to DHEA and DHEA sulfate this stuff goes down to circulation And this is the predominant pathway in the prostatic tissue. So this first enzyme here 3-beta HSD Is a step that we we suspected may be gatekeeping in the development of castration resistant prostate cancer This is Charlie Dye who spent a couple years in my lab as a med students now of oncology fellow Dana Farber So let's just sort of pivot here and take a broader perspective for sex steroids from adrenal precursors So the question here we wanted to we thought was important to ask Really considering this over some time is Does HSD 3b1 which is the gene that encodes for 3-beta HSD 1 does it genetically regulate? Basically adrenal sex steroid dependent human phenotypes So you can see that 3-beta HSD 1 is required for conversion from DHEA to androstenedione Androstenedione and testosterone are basically reversibly interconvertible through over a dozen 17-beta HSD's 3-beta HSD is an irreversible reaction and These are both as I mentioned both substrates meaning androstenedione and testosterone substrates for 5-alpha reductase and required for potent androgen synthesis and AR stimulation from adrenal DHEA and these Both as you know are also substrates for aromatase and required for ER stimulation from adrenal DHEA so a Number of years ago. We identified completely dimorphic metabolic phenotypes and we came to call these adrenal permissive versus adrenal Restrictive phenotypes. So this is the experimental setup. Basically. We're looking at DHEA metabolism again The first step is 3-beta HSD all the way to DHT. This is the adrenal restrictive phenotype We put the DHEA on it basically just sits there. There's very little flux through 3-beta HSD very little DHT synthesis right, so basically the precursors are stuck upstream of 3-beta HSD or the gene is HSD 3b1 and the opposing phenotype metabolic phenotype is the adrenal permissive phenotype so you can see an Alternative models and tumors the DHEA is rapidly consumed 3-beta HSD is relatively fast So basically you've opened up the floodgates right adrenal permissive because it enables the conversion from adrenal precursors to downstream potent sex steroids so the mechanism is all in this paper that we published about a decade ago in cell and it basically has to do with a single nucleotide change that encodes for Basically a missense and this confers a difference in susceptibility to ubiquitination and degradation Right, so the adrenal permissive allele is resistant to ubiquitination You have high steady-state levels of the enzyme the adrenal restrictive allele is Is turned over more rapidly you have low steady-state levels of the enzyme in cells so when we're thinking about Phenotypes that might be conferred by the adrenal permissive allele We have to be mindful of genetic ancestry and the reason is that if you look for the adrenal permissive allele frequency Worldwide there's a huge variation by geography in ancestry the highest frequency of the adrenal permissive allele is Actually in people of European descent where typically one and two people have at least one copy of the adrenal permissive allele There's actually a north-south gradient. The highest frequency is actually in Italy and Spain It tends to be much lower in most East Asian countries. So for example in Korea It's a 5% allele frequency and in most African populations is comparably low So for example in the Yoruba Nigerian population, it's 6% so huge difference so our major hypothesis For our group right now is that HST 3b1 genetically regulates sex steroid driven phenotypes in the absence of gonadal sex steroids in other words in the adrenal only setting for sex steroids So, you know, what are the questions that we want to ask? Physiologically one obviously it's in the title of the talk prostate cancer and this is specifically in men who've undergone medical castration standard treatment for advanced prostate cancer in women, it's estrogen dependent post-menopausal Breast cancer right absent gonadal estradiol and for endometrial cancer. I think it's a fair question as well Median age of diagnosis is 60 years. Obviously typically post-menopausal So thinking sort of simplistically and physiologically you have two endogenous sources or two major endogenous sources of androgens So for prostate cancer when you treat with standard medical castration The adrenal precursors have to be converted over metabolically, right? So just like other metabolic dependencies in cancer or oncogene dependencies If you provide an alternative supply to the tumors, then maybe they can get along and progress, right? so the thinking here is that HST 3b1 inheritance may act as a spigot and if you open that up these tumors may progress more rapidly So does the adrenal permissive allele basically affect the duration of response to medical castration? So this is the first cohort that we looked at to answer this question and so this is basically progression-free survival in men with advanced disease time from castration to progression So men who do not inherit the adrenal permissive allele have the best outcomes Those who inherit one progress more rapidly and those who inherit two alleles Progress profoundly more rapidly. This is Jason Hearn who led a lot of this work He was a radiation oncology resident with us in Cleveland, and he's now an associate professor of radiation oncology at the University of Michigan So that was a first cohort and then obviously what we wanted to see if this holds true in other cohorts This is a collaborative Another study that we did with Manish Kohli and Don Tyndall again zero versus one versus two alleles at Mayo Clinic So there are three cohorts in the first study that we published in Lancet Oncology in 2016. This is the first independent validation done by that group at University of Utah Again, those who get two adrenal permissive alleles progress profoundly more rapidly and sometimes people say well It's not really true unless you see it in Boston So we did a study with the group at Dana-Farber and again, you can see a similar phenomenon zero versus one versus two Adrenal permissive alleles and that was in JAMA Oncology. So now there's 12 cohorts worldwide that I think quite clearly validates That inheritance of this adrenal permissive allele leads to poor clinical outcomes in the context of medical castration including cohorts from Japan, Spain, a phase 3 clinical trial that's practice changing, another study from Hopkins and ICR, Minnesota Hopkins, and there are others as well. So let's pivot the question a bit and move on to the next disease The question here now is does HST3B1 play a role in ER-driven postmenopausal breast cancer? So prior to this so far as we can tell there is no genetic link between estrogen metabolism, estrogen exposure, and breast cancer. There is a study from the group at USC that linked, for example, a non-coding SNP in aromatase to modest changes in circulating estrogens, but that didn't link all the way to breast cancer risk. And we thought HST3B1 might regulate estrogen-driven breast cancer again in the adrenal only setting. So our hypothesis was that homozygous adrenal permissive inheritance may be associated with ER-driven postmenopausal breast cancer. Why homozygous adrenal permissive? The reason is that's where that's the context where we saw the most severe phenotypes in prostate cancer with castration. So this is a study that we published just last year in JCI Insight. This is Megan Cruz who's one of our breast medical oncologists who led the accrual of this our single institution study where we looked at the frequency of the homozygous adrenal permissive genotype for ER-positive postmenopausal breast cancer. You can see it's about 17%. So that's elevated compared with what we anticipated and found in sort of the general population of 9 to 10%. For ER-negative postmenopausal disease, the numbers are low, but you can see it's not elevated. If anything, it may be suppressed. So obviously we wanted to validate. So we did this with Serena Zanell at Cambridge. Again, you can see ER-positive postmenopausal, the homozygous adrenal permissive genotype is enriched, and if anything, suppressed in ER-negative disease. That was actually surprising to us. And in TCGA, it's the same thing, right? So about between 14 and 15% and lower in ER-negative disease. So that's sort of enrichment of that genotype in postmenopausal estrogen-driven breast cancer. Does this mean anything for clinical outcomes? So this is a story that's now in press, and it's a collaborative study led by Megan Flanagan, who's a breast surgical oncologist at the University of Washington, looking at 640 women with postmenopausal ER-positive breast cancer, stage 1 to 3, all treated with local therapy, and looking at clinical outcomes, time to recurrence. And what you see here is actually quite remarkable, that if you look at distant metastatic recurrence, for those who are homozygous adrenal permissive, the hazard ratio for recurrence is 4.9 compared with those who don't inherit this allele. So the total number of events is relatively small, but you can see there appears to be a pretty strong signal there, and when you look at breast cancer mortality, again, this is homozygous adrenal permissive with hazard ratio of 3.5. So let's pivot again. Now we're also focusing, we're continuing the studies on prostate and breast cancer. The breast cancer work is just at its very beginning, but we felt compelled to look at endometrial cancer as well. And unlike breast cancer and prostate cancer, where hormonal therapy is sort of well-established, you've got gazillions of adjuvant trials in breast cancer and in prostate cancer. We're moving from, you know, using the new hormonal therapies in castration-resistant disease to castration-sensitive disease earlier and earlier. In endometrial cancer, it's very difficult to find high-quality specimens from great clinical trials. Oftentimes, the GYN oncologist will reach for other therapies depending on the context, right? So there's a number of different ways to look at endometrial cancer, and this is just taking you through that over the years. These tend to be more hormone therapy responsive, but it's not really well worked out, and these tend to be less hormone therapy responsive. So the old categorization is type 1 and type 2 endometrial cancer. Type 1 tends to more often express ER and PR. Endometrioid histology tends to be more hormone therapy responsive. Serous histology, not so much. And the more modern categorization I think that people are going to is from the TCGA, where you have categorization of copy number low versus copy number high, less well differentiated here, right? So we had to ask the question, can these phenotypes be driven in part by inherited HST3B1-driven endogenous estrogen biosynthesis? So this is Jeff McManus, a research associate in our group, and Roberto Vargas that we work with, a GYN oncologist at our institution. So is there an association with the type of endometrial cancer? And this is the odds ratio of having the copy number high, less well differentiated, less hormone therapy responsive subtype for each adrenal restrictive or less estrogen synthesizing allele. And you can see that it's certainly elevated, whether you do a univariate analysis adjusted for race, or just looking at the white subjects, which is the majority for TCGA. Now the next question that we wanted to answer is, is there an association with, so that's just an association with the type of endometrial cancer, is there an association with actual clinical outcomes? So we went to UK Biobank to answer this question. You can see there's about 2,500 women with endometrial cancer in this analysis. And what you hopefully can appreciate are that women who are homozygous adrenal restrictive, that's the AA, have worse endometrial cancer mortality compared with those who inherit two of the adrenal permissive allele as well as the heterozygotes, right? So there's, I think the power sort of falls apart here, but here it's quite clear and you can see the P value. So what's really remarkable about this is that the copy number high subtype that's less well differentiated, these are enriched for P53 mutations, et cetera, all these somatic alterations, but this precedes all of that, right? Because this is inherited and I think that's what makes it more powerful and easier to follow in terms of biomarker analyses, right? So in conclusion, for the adrenal permissive HST3B1 allele, this enables more rapid conversion from DHEA to sex steroids, that's a cellular phenotype that I showed you earlier on. For prostate cancer, there's more rapid progression and death after castration. This is now very well established. You also actually get worse outcomes with apiraterone or enzalutamide and those are studies published by others, European Group and Hopkins and I didn't show you those data just because there's lack of time here. So for breast cancer, it seems like there's a, the adrenal permissive allele enables a greater likelihood of ER positive postmenopausal breast cancer and it seems to have the opposing effect on ER negative disease. And it seems like there's a pretty strong signal that there's worse outcomes with ER positive postmenopausal disease and I should say in that paper, that story that's in press, 90% of those women got adjuvant hormonal therapy. So this is worse outcomes and worse breast cancer mortality in spite of adjuvant therapy and our thinking is that these women may require longer adjuvant therapy. We actually see a similar signal in prostate cancer in one of the phase three clinical trials of enzalutamide. I didn't show you those data but I'm happy to discuss it. And so for endometrial cancer with the adrenal permissive allele, this seems to be associated with the copy number low disease and low grade disease that tends to be more hormone therapy responsive and that may confer better overall and cancer specific survival. What we don't have for endometrial cancer is the association with hormone therapy responsiveness. Again, it's more challenging to get but we're working toward doing that. So pivoting from cancer to something else, this is something that wakes me up every day. The bigger questions in my mind beyond the cancer related questions. So why is HST3B1 associated with genetic ancestry? So I think there are at least two possibilities. One is that you get these different frequencies by geography because of just simply genetic drift. An alternative is that maybe genetic selection through human history. For something where you have a clear biochemical phenotype with other cancer associated clinical phenotypes, I suspect that there may be some level of selection and then that leads to the next natural question which is what are the other HST3B1 driven human phenotypes? So why did this evolve differently? I think that the cancer associated outcomes is probably, this is speculation, an accident of human evolution and I'd love to understand what else this is doing in human physiology. And I'll end by saying this is a picture of Gene and a quote from Gene where he says, really my entire life is based on the influence of others and I really want to echo that sentiment and that's certainly the case for me both with the interactions with Gene, the former division of endocrinology at UT Southwestern, everybody here continues interactions with Rich Aukus, other colleagues, people in the lab who I also consider colleagues as well. In many ways what's more pleasurable to me, even more pleasurable to me than the work is actually the interaction with people. It's quite meaningful and enriching to me. And these are, I try to show photos and show names of the folks in the lab who are actually involved in a lot of this work. This is our current group. These are folks who've left. We have a number of collaborators in Cleveland and outside of Cleveland as well, in particular Rich Aukus. I want to thank patients and their families for everything that we do. We really try to stay very close to human mechanisms. So if we identify a mechanism in the lab that's metabolic, we like to confirm quickly in the clinic. And if we can't do that, then we typically move on to other questions. But to do that, you really have to have buy-in from patients to donate tissues, to donate human biospecimens, and their family members as well who oftentimes have to buy in as well. So that's very important. And these are our funding sources. And this is just COVID-friendly outing for our group. So thank you for your attention. And I'd love to take any questions. Thank you. Well, thank you very much. People can line up at the microphone. And if there are any questions from people listening remotely, please go ahead and type them in as well. Go right ahead. Christy Brown, Weill Cornell Medicine. Beautiful talk. Very exciting work as well. So we do see an increase in 3-beta-HSD1 in the breast adipose tissue in relation to obesity. So this suggests a transcriptional regulation locally. And I wonder if you've considered the compounding effect of transcriptional regulation and this adrenal permissive allele. It might not be so easy to answer. But do you have any clues from gene expression in your cohorts in relation to these alleles? Yeah, absolutely. So that's it. And thank you for that question. We are certainly looking at a number of other mechanisms that may regulate 3-beta-HSD. We actually just published a paper in Cancer Research that hypoxia, intermittent hypoxia may actually transcriptionally upregulate HSD3B1. But it's the reoxygenation that may be required to supply the cofactor for 3-beta-HSD. And there are other aspects of function. I'd say half our group is obsessively focused on 3-beta-HSD. Trevor Penning, University of Pennsylvania. Hi, Trevor. Great talk. Thank you. Chrissy Brown stole my first question was about the transcriptional regulation of HSDB1. But my second comment was in the endometrial case with the permissive HSD3B1, do you think this might be related to increased flux to progesterone and affecting ratios of E and P? What was the very last thing you said? And change the ratios of E versus P. Yeah, I mean, as you know, 3-beta-HSD is absolutely necessary for conversion from pregnenolone to progesterone. And I mean, the other question that we actually, so again, I'm a medical oncologist, but there's gonna be a maternal fetal medicine fellow working with our group to look for pregnancy-associated outcomes. We have a little bit of data. Ratio of progesterone to estrogen, I don't really know. We haven't looked at that specifically. But the progesterone angle is something that we're also very, very interested in. Great, thank you. Thank you. Okay, if there are no further questions. Thank you. That was terrific. Thank you. I'm doing mine. Thank you. So our next speaker is Wayne Tilley. Pull up his. He's from University of Adelaide in Australia. He's gonna talk about modulating androgen receptor signaling in breast cancer. What I didn't realize when we invited you was that you were actually at UT Southwestern before any of us were. And so there's a connection that I didn't even know and it just turned out to be the case. There's a reason I have less hair. I don't know. Age. So yes, thank you very much for the opportunity to present. It really is a privilege. And as our chairperson, oh, there are my disclosures. As our chairperson said, I had the very good fortune to spend four years at UT Southwestern with Gene where I cloned the human androgen receptor with Marco Marcelli and Mike McFall. And it was an exciting time to be there because 5-alpha reductase was cloned, aromatase was cloned and the like. It really was a tour de force of molecular endocrinology. And I was fortunate to see Gene before COVID hit. And I was very fortunate with Ganesh Raj and a colleague at UT Southwestern to write this sort of, I guess, reflection of Gene Wilson's legacy. But I'd like to just highlight Gene's own memoir there because if you have any interest in knowing more about him, it's a terrific read. And you get a lot of insights and you can see the butterfly there. I mean, he just loved so many things, going to the Antarctic, going to different places and observing nature, I think. And so it was just an extraordinary time. And I put this up because towards my time at UT Southwestern, I did an endocrine round. And I had Brown Goldstein, Dan Foster, Gene all in the front row. And it was on estrogen receptor and androgen receptor and breast and prostate cancer. And people looked at you a little weird then when you talked about androgen receptor and breast cancer and converse ER and prostate. But Gene sort of really encouraged me to pursue the AR breast cancer avenue. And that's why I put this up. And it really took us a long while. It took us almost 40 years to sort the damn story out. But we got there last year. And that last slide I put up, or the last sort of image there, I mean, that came out in JCO in 1995. And it really did stimulate an interest in, because it actually showed that the androgen receptor was the sole predictor of response to medroxyprogesterone acetate, which was used as a second line hormone therapy following failure of tamoxifen. And so that actually told us the AR must be playing some role in mediating hormone responsiveness. But, of course, then you fast forward and now, you know, the current treatment paradigm for ER-positive breast cancer is basically to build a bigger hammer. And that's, of course, led to, you know, SARMs, rheumatase inhibits, and now a whole generation of SIRDS. But unfortunately, we still get therapy-resistant breast cancer. That's the main cause of breast cancer deaths. And, as you well know, it's largely because of, you know, in part reactivation of the estrogen receptor. And, of course, you know, 10 years of an AI, or a combination of tamoxifen and AI, has incredibly, you know, debilitating effects on some patients to such an extent that it often leads to treatment noncompliance. So we've really had an interest in, you know, can you actually take a different approach to this? And is it possible to stimulate homeostatic mechanisms that inhibit or modulate estrogen receptor activity in the normal breast? And this sort of started with Jason Carroll in Cambridge, and Hicham Mohamed, who may be here. He actually developed this RIME technique, which allowed us to show that, on a background of estrogen, I don't know if I can get this working. Maybe it's not working. Hmm. I think your finger's in front of you. Oh, it's just blocking it, sorry. Yeah, okay. Good, thank you. So in this panel here, this is on background of estrogen. And then we did this RIME technique. The only factor that was significantly gained in terms of an interact with the estrogen receptor, and we added progesterone, was indeed the progesterone receptor. So this was a striking finding, and suggested this was a critically important sort of interaction. And as you can see on this panel here, when you actually add progesterone, as you expect, you get a market increase in PR chromatin binding. But at the same time, you had a massive gain in ER chromatin binding as well, and then you had some conserved sites here. Now, I show this just to highlight later the difference in terms of how AR works, because one of our reviewers, when we were at Nature Medicine, said, well, where's the novelty? We know PR reprograms. Well, we actually just had to take them back and politely redirect them to the actual data in this paper to show that it's a completely different mechanism. But unfortunately, sometimes when people just read abstracts, they get things wrong. So this led quickly to two clinical trials of treatment in IVR positive breast cancer, one in the UK, one in Australia. These trials were both complete later this year, and so that will be really interesting to see whether this basic observation we made some five, six years ago will actually really have the potential for repurposing PR agonist as a potential therapy in endocrine sensitive and maybe metastatic breast cancer. So what we also noted back then is that in luminal B breast cancers, which have the poorer outcome, there's some 40% of luminal B tumors have a copy number loss in the TCGA data set. But if you actually look at AR expression in TCGA, what you notice is that despite the copy number loss, there's no difference in AR expression. And so this really made us think that maybe we should look more closely at what AR is doing, and that maybe in these poorer outcome positive breast cancers, the AR, exploiting AR as the therapeutic target could be worth pursuing. So the AR is expressed in a high percentage of breast cancers. This was a very large cohort study, and you can see here that in DCIS, invasive ductal cancer, or lymph node mets, the combination of the orange and the green is AR, and the orange is both AR and ER expression. So large, high percentage, almost 89% of these tumors express the AR, and some 60% express both receptors. And so if you look at the dual staining down here with ER and red, AR and green, you can see the majority of these tumor cells here are orange, indicating the co-location of the two. So this suggested that you had potential for genomic interaction of these two signaling pathways in an ER positive tumor. We knew that from previous studies that the androgen receptor is an independent predictor of survival in breast cancer. And historically, of course, androgens were used to target breast cancer prior to the advent of tamoxifen, and certainly Kennedy published in 1958 in the New England Journal using fluoximestrone as a therapeutic agent, and it actually had comparable efficacy to tamoxifen, and indeed Ingalls and colleagues at the Mayo Clinic, even in the 90s, were looking at combinatorial approaches with tamoxifen and an AR agonist. So it never really went away until ultimately the aromatase inhibitors came along, and that was really the demise of AR agonism as such then. And of course, back then, they didn't actually, you know, in the 50s, they didn't know a lot about the androgen receptor, they didn't have a biomarker for response, so the response rates were actually relatively impressive. So you fast forward to present day, and there's a lot of confusion about how best to exploit AR expression in ER positive disease. And it's in part because in triple negative, or subset of triple negative disease, there's a suggestion that it may be beneficial to target the AR with an AR antagonist, although I think now we're getting better models, even that is questionable. And I'm happy to discuss that later. But also, because the AR is, and of course, if you read a lot of reviews, and it just frustrates me, they lump triple negative with ER positive, and they say, well, okay, it's maybe bad in ER negative, therefore it's gonna be bad in ER positive. And then they say, well, the AR is an oncogene in prostate cancer, and we use an AR antagonist there. So surely if the AR is expressed in breast cancer, we should use an AR antagonist. Well, actually, women are not quite the same as men, and breasts are not quite the same as prostates. And it turns out that based on probably some limited preclinical data, what we ended up with is this extraordinary scenario where we had clinical trials being done with both AR agonists and antagonists at the same time. Pfizer have actually pulled two of their AR antagonist studies. But because of this clinical controversy, it really drove us to develop better preclinical models so we could actually understand how best to target the androgen receptor. So one approach we took was to use these patient-derived breast tissue explants. This is where we take tissue rapidly from surgery within one hour into the lab. You can actually coach them on a gelatin sponge, treat them then for a period of time with various hormones. And what you can see on the top here, this is normal breast, that with estradiol, you see a market increase in BRDU incorporation here, and this is suppressed by co-treatment with diastereotestosterone. And then if you look at invasive breast cancer, you see exactly the same. If one quantitates Q67, and a larger number of these tumors cultured in this way, you can see that the combined treatment suppresses the E2 induction of Q67. And if you then look in gene center enrichment analysis with RNA-seq from, so this is now using RNA-seq from these tumor samples cultured on the sponges, you can see that with E2, you're actually getting, sorry, no, what's this going, it's actually giving us a problem there. That's not over now, I think it actually may have, oh no, yeah, okay, good. So you can actually see with E2 here, you've got a positive correlation with these G2M checkpoint genes, and then this is reduced or a negative correlation when you co-treat with diastereotestosterone. This is true in the normal breast and in invasive breast cancer. Oh, sorry, what you can see more clearly in the middle, if we now just look at cell cycle genes and GSCA, you can see again, positive association with E2 treatment, and this negative correlation with co-treatment with E2 plus DHT. If you look at androgen response, now you see the converse negative correlation with E2 and a positive association with E2 plus DHT. This is not the most attractive-looking sort of plot, and it's partly because in GSCA, the androgen-stimulated genes are DHT-induced genes in own capsules, but what we find actually in the breast is that the overlap between DHT-induced in prostate and in breast human tissues is not that great, so it was actually probably surprising that we saw such a good association there. So then we moved on. We really wanted to develop sort of contemporary models of breast cancer, and so these are PDX models. The model on the left-hand side is developed by Carol Sartorius of Colorado, and the one on the right is one of Elena Welm's models that Elgene Lim had at the Garvin Institute, and because of the prior controversy about this, we were very keen to do these experiments across three different laboratories, so the left-hand panel was done in Denver, the right-hand was in Sydney, and we've now, in a collaboration with Elena Welm, acquired all of her models and characterized them for ERP and AR responsiveness, and we've been able to replicate these findings, but time doesn't really allow me to go into this, but what I want to just highlight today, and you see in both these models have ESR1 mutations, and on the left-hand side in orange, you can see the E2 stimulated growth, and this is clearly inhibited by both dihydrotestosterone in the blue and Asam, a selective anticeptor modulator, and Novasam in the purple. If you look at the HCI05 model, here's the very estrogen-sensitive, but what you see is this profound inhibition with both the Asam and dihydrotestosterone, and you can see on the left-hand side here, if we now look at tumors, so we usually set up parallel experiments, so we can take tumors at five or seven days, you can see estrogen stimulation, QE67 here, and then with the Novasam, you see induction of androgen receptor in red, and a marked reduction of QE67, so these early changes in QE67 and AR are reflective of what we're seeing long-term with tumor volume. This is now an in vivo ChIP-seq analysis, so we take the tumors also at five or seven days, and we can actually do ChIP-seq analysis for ER and AR. This is actually for ER, and again, as I just mentioned, I just wanted to highlight that as opposed to PR activation where we saw a marked gain, here now, the key finding is that we've got conservation, but here we have loss rather than gain of ER, so this is the main feature that's quite different to PR activation, this loss of ER binding. You can see this in the read density plots here, that this is vehicle, black is a Novasam, and then DHT's really suppressing this ER binding, and this is true in both models. If we actually, just to give you an example, and this is in the intron, one of the MIB gene, so you can see with estradiol, you've got stimulative or recruitment of ER chromatin here, and then when you co-treat with diuretic testosterone, you see the complete loss of this ER binding, and it's markedly suppressed with the SAM, and then if you look at the RNA-seq data from these tumors, you can see this correlates with a marked reduction and you can see this correlates with a marked reduction in MIB expression with the combined treatment compared to estradiol. So the next question we want to ask is, well, what about an ER antagonist, enzalutamide? So this is, again, the HCI005 model. Orange was the E2-treated tumors. You can see co-treatment with enzalutamide has no effect on tumor growth. This is the DHT treatment. If you actually combine E2 plus DHT plus enzalutamide in the black here, you see you do see some reversal of the DHT effects. This indicates the DHT is working in the system. It also suggests clinically that if you were to give an AR antagonist, you may well counter the natural endogenous benefit of AR signaling within a breast cancer. So that's something we're very interested to pursue. We also wanted to know, and people then have also argued, well, OK, this is final works in estrogen sensitive disease. What about estrogen independent disease? So this is a model that we developed with LG and Limit, the Garvin Institute, the GAR 1513. So this is now, the orange here is actually growth in the absence of estradiol. And you can see co-treatment with enzalutamide has no effect on growth. Here, DHT is not quite as effective as in the sensor, but it's still significantly inhibiting or retarding tumor growth. So what about the mechanism of this? These are a two-factor log ratio, or an M-plot. And to try and just keep it simple for you, so you've got ER, this is showing ER binding. And it's on the background of adding in dihydrotestosterone. And what you can see here is that in this panel, these are ER binding sites that are lost. And in this panel, these are now binding that's sort of gained both AR gain and ER gain. So here, you can see loss of PR. And I think I see nib here. I'm having trouble reading that from here. And then, and that doesn't really help. And if you go to the middle panel here, you can actually see here, this is ER being recruited. And it's still present with DHT. But you can see no AR actually present at these sites. And again, DHT here sort of suppressing PR binding here. So you get a suppression of this recruitment with both of these two genes, the nib and the PR. And if you look over here, you can see that the expression of both these genes has also concomitantly decreased. If you look at the genes, the binding that's gained, the ER binding at these sites, so a number of these are tumor suppressors. And so I just give you two examples here. So this is now quite different. So with these AR target genes, you can actually see with the DHT, you've got stimulation of AR binding. But if you look at the ER, you can see with the E2, there's no binding. But there is some recruitment of ER as well. So you're getting marked recruitment of the AR and some recruitment of ER in both cases. And this is associated with an increase of the expression of the two genes in the RNA-seq data. And how does this actually work? Well, we think it's critically dependent on these two ER coactivators, P300 and SRC3. And this is, again, another m-plot or two-factor plot. And what you see, again, these same genes that where ER binding is lost with co-treatment with diuretic testosterone, you also see loss of SRC3 and P300. And then, again, in the genes that were where you saw gain of AR and ER binding with co-treatment, you see recruitment of P300 and SRC3. So effectively, DHT treatment is tethering SRC3 and P300 away from these ER target genes. Now, if you look at the actual motifs associated with these, the genes where ER binding is lost, this is predominantly an ERE. And then if you look at the motifs associated where there is a gain in binding, here you see FOXA1 and GREs and AREs, as you would expect. So this is consistent with AR recruiting genes to its own response element, but pulling these genes away from sitting on EREs. So to summarize that, what we believe happens when AR is agonized, it can bind adjacent to an ERE to displace ER, and that leads to destabilization of the ER transcriptional complex. So effectively, it's a very good ER target therapy. But what we think is probably more important is this AR agonism and AR binding to its own canonical response elements, and now tethering P300 and SRC3 away from this complex. And it may or may not have ER in this complex. And it's actually, therefore, effectively a dual therapy, because you've got a very effective ER target therapy, but you've also got activation of these AR regulated pathways, which includes tumor suppressors and a number of other genes that are probably critical for the durable growth inhibition that we see in the PDX models compared to estrogen therapy. It requires AR, an AR agonist, and AR binding to DNA. And this came out last year in Nature Medicine. And Elgene Lim on the left, Teresa Hickey in the middle, and Jason Carroll were critical in this work. So a key question is, well, how does this compare with standard of care endocrine target therapies? And this is the HCI005 model again. Contains an ESR1 mutation. So here's the estrogen-stimulated growth. You can see inhibition with tamoxifen here. But it's nowhere near as potent as the dihydrotestosterone suppression. And if you combine it with tamoxifen, it's comparable to DHT. If you look at the key 67, you can just see in the right-hand side that you've got market induction of AR with either DHT or the combination. And you've got suppression of key 67. Now look in the GAR15, which is resistant to E2. So this is the vehicle growth. You can actually see that SARM actually has some efficacy here. In this model, you can see the pulvocycloid has some activity. So this is in a model that's partially resistant to CDK46. And when you actually inhibitors, and when you actually now combine pulvocycloid with the SARM, you can see this more potent inhibition here. So we're actually seeing also evidence, not only that it's more effective than tamoxifen, but in a range now of PDX models that you can even combine this with a CDK46 inhibitor to improve the durability of the response. And again, on the right here, you can see the inhibition of proliferation with the combination. And this is quantitated here. We now have actually developed with LG and LIM models that are truly resistant to CDK46 inhibitors. And so this is a completely new model. So again, you can see the control growth here. You can see that it's resistant to the CDK46 inhibitor. And this is actually resistant to pulvocycloid and abemocycloid. But you do get some activity with the SARM. And then actually, you can see in combination, it's almost like you've maybe slightly resensitized, and you've got this additive effect. So we're really interested in trying to dissect the mechanism whereby you can get this additive effect of AR activation with a CDK46 inhibitor. And that, again, is just the quantitation. And we now have more models that we're exploring. So taking this forward, there are a number of clinical translational studies underway. We're currently writing up a phase 2 study that was done with GTX and now in ANOVA-SARM. It's taken a while to get this data. And this is with Beth Ovenmeyer from Dana-Farber and Cullard-Palmiller in the UK. And in that study, they were able to show that AR greater than 40% positivity was a good biomarker. Based on our paper that came out last year, VIRU established the R-test trial, which you can see the nature of it. It's a phase 3 trial on metastatic breast cancer. This actually led to ANOVA-SARM getting a fast-track designation from the FDA in January. So incredibly quick sort of translation here. And they've developed a partnership with Roche to develop a companion diagnostic to take this forward in breast cancer. And there'll be another trial starting with a Bemaciclib later this year in combination to test this combinatorial observation that we've seen. We're actually working with other companies to develop new agonists and take them to the clinic. Because just like most of the other target drugs, it's unlikely we're going to see a long-term, durable response in all cases. So we need other agents. And also, we need the academic freedom to be able to do investigator-initiated trials, which unfortunately, VIRU have suppressed. And we're also working with Claris Therapeutics because they have a novel delivery system that they've been using for some time that circumvents uptake for other portal circulation. And so it circumvents liver toxicity. And so we're looking at how we might be able to take that through into clinical trials as well. In the interest of time, I'm going to skip over this bit. I've actually had trouble going faster with my eyes not focusing on that or not. So in conclusion, I think AR is a tumor suppressor role in AR-positive breast cancer. AR competes with the recruitment of key co-activators to mediate growth inhibition. An AR agonist strategy, not an antagonist strategy, advocated for treatment of AR-driven breast cancer. And that selective antireceptor modulators, or SARMs, confer durable growth inhibition of AR-positive breast cancer. But I did show you some evidence. I didn't really emphasize it, but we actually believe there are some instances where a pure AR agonist is actually has better growth inhibitory characteristics and may lead to a more durable response. Finally, because I get asked this all the time, what about AR-negative breast cancer? There's some work from Miles Brown, now to Dana Farber and colleagues, suggesting based on limited cell line models that there may be a subset of AR-negative breast cancer, subset of triple negatives, where I has expressed that AR antagonists may be effective. But if you look at the data, the DHT stimulation is very modest. So again, with LG and LIM, we've built new PDX models now, triple negative breast cancer. And this is just to show you one line where clearly DHT is, I mean, it's only a modest inhibition, but it's actually DHT is clearly not stimulating. And interestingly, you can see that the DHT treatment results in marked up-regulation of AR nuclear activity. And it leads to stimulation of key AR-regulated genes and, indeed, down-regulates some key proliferative genes. So we think the story is not complete in AR-negative breast cancer. And there may even, indeed, be context there where AR agonists will be beneficial. But that certainly needs more work to resolve. So in conclusion, just thank you for your attention. And there are many people here, as Nima showed, who have been absolutely essential to this work. And likewise, I'd really like to acknowledge the patients and their families who are invaluable as either providing tissue and being active patient advocates within our research program. And with that, thank you. And that's a view for those of you who want to get to Australia at some stage. That's the beach from the front of our house in Adelaide. Thank you. Thank you. Thank you very much. That was terrific. Carol? Hi, Wayne. Thank you so much. That is not the view in front of my house in Minneapolis. For the 70% or so of patients who did not respond in the clinical trial, did they look at, say, AIB1 amplification, which is very common? Is there some way to track who might respond yet? And then I just wanted to know if your TAM-resistant and your estrogen-independent models were ever performed in ovexed mice? So latter question, no, not in ovariectomized mice yet. That's something we will do. Gosh, we did 15 PDX models, I think, for NatureMed. And actually, we've probably put more emphasis on building the new CDK46-resistant models and trying to tease out exactly how you would best use an AR agonist in metastatic endocrine-resistant disease. That trial started over a decade ago with GTX. There were completely unselected patients. There was no biomarker. There was no nothing. But it did meet the statistical significant cut point. That's why we're keen to write it up. Unfortunately, the whole thing was a mess when GTX collapsed. And then it got bought up by another company. And we nearly bought the drug. And then it ended up with Vero again. And then they wouldn't let us get access to the data or anything for us. We've had to even struggle to get the data. So unfortunately, we can't really answer that question, other than say they were totally unselected. Yeah, that's impressive then. And I was just wondering, because your mechanism involves SRC3, if you upregulate that, you might be able to bypass that, right? But very cool, thank you. There's so many questions we'd love to ask if we could get that tissue. The normal breast makes PSA. And you can use high-sensitivity PSA to screen for polycystic ovarian syndrome. Can you use that as a biomarker of response to your treatment, since it's a secreted protein? Unfortunately, we see PSA in virtually where we've looked at in the serum of most of these women. So it probably is a marker of response. Whether it's a good biomarker of selection or not is a completely different point. What I didn't have time to show you, we started with 142 genes as candidate biomarkers. We've narrowed that down to about five now. And so we're really trying to come up with something independent of AR per se as a biomarker of patient response. We think that's absolutely critical if this is really going to get traction in the clinic. We've been talking with Laura Essman with iSpy2, and she's extremely keen to have molecular markers where she can take AR agonism and iSpy2. Very nice talk. Can you comment on the side effects associated with AR agonism versus aromatase inhibitors, for example? No, they're the polar opposite. Metabolically? You would expect there to be completely different metabolic effects based on what we know. We haven't looked so much at the metabolic effects yet, but in terms of the majority of the side effects, I mean, cognitive function, libido, bone, just general well-being of women, it's the complete opposite with AR agonism versus AIs or tamoxifen. It's so obvious that the women on the trials know whether they're on placebo or drug because they feel well, and then their girlfriends want the drug as well. So yeah, it's pretty black and white. All right, we got one more, and then we gotta move on. So sorry to everyone. Cutting you off. Wayne. So this is Jennifer Ricker from the University of Colorado. So we just completed a trial, and we're submitting the paper in women with metastatic breast cancer who have been relapsed on after aromatase inhibitor, fulvestrant in many cases, CDK4-6 in some cases, and enzalutamide, the anti-androgen, does show clinical benefit, and none of the patients progress or increase the progression of the disease, and about 25% had clinical benefit rate at 24 weeks. So I think it's gonna all depend on the context, and maybe early on it could be agonist or antagonist, just like high-dose estrogen and high-dose testosterone did historically work. So I think it's really very context-dependent. It may be, but that's why we've gone out of our way to still keep building models to try and understand exactly what's occurring in an endocrine-resistant metastatic setting. We're also looking at the whole role of estrogen deprivation versus AR agonism in the process of metastasis per se. I didn't have time to show you a little bit of preliminary data there, but we've put an enormous amount of energy into building labeled PDXOs that we can use in the MIND model and do lineage tracing to specifically look at what cells are going to what metastatic sites, and what's driving that, and how does AR agonism work. In terms of the ENSA effect that you're seeing, again, that's why we put a lot of effort into building those CDK4-6-resistant models, so we can actually model that effect and see, is ENSA actually working via an AR antagonism, or is it an off-target effect? So the dose of ENSA and how that's working in a breast tumor, we really don't know yet. We need Nehmer and whatever to help us look at all this, and what's going on with metabolism. Yeah, we're using the same dose that they use for prostate cancer in our clinical trial. And I think that might well be too high in a woman. You know, some of Alana's models, like the one you used, the HCI013, that have ESR1 mutations, they do metastasize, even in mice that don't have estrogen. And so that's what we're doing as well, to model the actual situation in women that acquire these mutations only after aromatase inhibitor treatment. So the estrogen deprivation, and we think they kind of switch over to using AR when there's no estrogen. So we don't see, well, we don't see that yet. So we now have, we have a collaboration with Alana, we have every one of her models, we have all the PDXOs, we've labeled them all. We're now looking at estrogen, progesterone, and androgen response in all of them. So give us 12 months, we should be able to tell you what's going on with metastasis as well in those models, I hope, because this needs to be sorted out. It's absolutely axiomatic to taking this into it as a new endocrine therapy. So. There's no doubt that agonists make women feel better, so. But it makes them feel better. Oh, yeah. There's no question about that. Quality of life goes up. I think that means a lot for breast cancer patients. Well, thank you very much. I mean, it just really emphasizes the importance of basic research and clinical trials together to try and understand what are obviously super complicated and probably changing over time, depending upon the stage of the cancer. That's the problem. I mean, this is an evolving disease. It's not static. And we absolutely need to make people collect tissue with clinical trials. Exactly. It's a real emphasis to collecting tissue, which is brought up by the first two speakers. Yeah, before I actually introduce the next speaker, I just wanted to mention that I actually had the good fortune of knowing Dr. Gene Wilson. Besides being a phenomenal physician scientist, I'll tell you this, he was a kind and warm-hearted human being. Honestly, a true Texan. Dr. Wilson really cared about his mentees. And I can tell you personally that one of the mentees that he was quite fond of actually happens to be our next speaker, Dr. Richard Aukas, from the University of Michigan. Dr. Aukas will be talking to us about the traditional and alternative pathways to human androgens. Please welcome Dr. Aukas. Thank you. Okay. Well, thank you, Hans. I'm delighted to be here to sub for Maryland. These are two tough acts to follow. So I'm gonna give you a whirlwind tour about what we used to know and what we know now. These are my disclosures. So yes, this is a tribute to Gene. And yes, what hasn't been mentioned is Gene was a former president of the Endocrine Society. And that Gene was very instrumental in kicking off some of the early successes I had in my career. And I'm totally indebted to his kindness and his support and his intellect over the time. There won't be another Gene Wilson, and that's for sure. So Gene and Jim Griffin and Mike McFaul and the others that you heard about came up with the classic paradigm. So when I was a resident and an endocrine fellow, it was that LH regulates the testis to make testosterone. Testosterone goes into the target cell. Most of its actions are mediated by testosterone itself. But a few, particularly virilization of the external genitalia require conversion in the target cell to dihydrotestosterone by 5-alpha reductase. And then on the right is probably the most famous pH curve in the history of enzymology where Jim Griffin showed that there were two peaks and that the patients with clinical 5-alpha reductase deficiency were missing the acidic peak which turned out to be the type 2 enzyme after David Russell cloned the two isoforms. And then of course there's the adrenal androgens. And then the adrenal androgens are responsible for the process of adrenarche. So during fetal life, the fetal adrenal makes a ton of DHEA sulfate which is actually, if you didn't know this, it's the source of estrogens during pregnancy, it's the precursor, it comes from the fetal adrenal. And then there's involution of the fetal adrenal and then in early childhood there is growth of the zona reticularis and the process of adrenarche that occurs independent of puberty that leads to the development of pubic and axillary hair in both boys and girls. DHEA has peaks at about 25 years of age and then declines throughout life. So that was the classic paradigm. So the conventional pathway, to dress it up in a little more detail, all steroids come from cholesterol, the sign chain cleavage enzyme which is P45011A1 generates the first committed precursor pregnenolone. And then the downstream enzymes. My lab has worked on P45017A1 which does both the 17-hydroxylase and 17-20-lyase activities to get you from 21 carbon steroids to 19 carbon steroids, which are the precursors to testosterone and thus ultimately dihydrotestosterone. So now, one thing is that as you purify this enzyme, you lose the lyase activity. And that's because there's an essential cofactor called cytochrome B5 that's required only for the 1720 lyase activity. And so what my lab has worked on for the last 20 years is how does that happen, how does cytochrome B5 selectively stimulate the 1720 lyase activity. And we know that patients who are deficient in cytochrome B5, the males, have ambiguous genitalia and failure to progress through puberty. So this is a clinically relevant action. This is not an in vitro artifact of cytochrome B5. So cytochrome B5 is a very small hemoprotein. And there's only a bunch of possible ways that it can interact with another protein because there's only a limited number of surface exposed residue. And one of the things I showed in Walt Miller's lab with Tim Lee was that apocytochrome B5 that's missing the heme also stimulates activity. So based on this, we felt it was not an electron transfer effect, but rather an allosteric effect. So we sought to understand the interaction of B5 with 17A1. So Jackie Nafin, all of us, who was a postdoc in Dallas with me, she rolled up her sleeves and she made a whole bunch of mutants on those surface exposed residues. And they came up into three groups. These mutations did nothing. These mutations were a little low at one equivalent, but then eventually got to full activity. And then the ones here never got full stimulation. And this double mutant E4849 really didn't have any ability to stimulate cytochrome B5. Now it could be misfolded, but no, that's not true because we could, it bound heme, it formed a different spectrum, and it transferred electrons from flavoproteins at the same rate as the wild type. So it is a functional protein, it's just not stimulating lyase activity. I took out a lot of data slides. I'm gonna show you one data slide. So we suspected that these are the key residues on B5. We predicted that they would be interacting with residues that were key for 1720 lyase activity. And we knew from the work that David Geller and had done with me in Walt Miller's lab on patients that Bernice Mendonca had identified that these two residues, these two arginines, R347 and R358 were neutralized in patients with isolated 1720 lyase deficiency. So bedside to bench, we took those two mutations and we tried to see if they interacted with those negatively charged residues that we had identified on cytochrome B5. So we tried to form a cross-link with this carbodiibin reagent, which doesn't form cross-links with arginines. It forms cross-links with lysines. So we had to bait the P450. We had to substitute lysines for arginines at those residues. We had to show that it had normal lyase activity that was stimulated by B5, which it did. And then we cross-linked the two proteins and were able to generate these complexes. And you see that it forms the complex in the presence of the carbodiimide. When you use the double mutant, you get much less, if any, complex formed. So then we cut out these bands and we chopped it up and sent it to the mass spec and we looked for the cross-links. And to make a long story short, these two residues which we were looking for were one of the cross-links. And then these two residues, glutamate 42 and the arginine to lysine 347 was the other cross-link. And when we lined those up, 48 and 49 are pointing right at R358. So we believe that this is actually the molecular structure of the interaction that is essential for 1720 lyase activity. But I didn't show you what that does. So we spent the next 10 years trying to figure out how that turns out to be a stimulant of 1720 lyase activity. And you gotta understand that cytochrome P450 enzymes go through a cycle. They have to bind substrate, accept two electrons from NADPH, bind oxygen. And if you're lucky enough to go all through this cycle, you get to compound one, which does the chemistry. Now, if you don't put those protons on that iron-oxygen species in the right order, then what you do is you lose partially reduced oxygen, either superoxide or hydrogen peroxide. And we call those uncoupling reactions. So basically, the shortstop couldn't turn the double play. And you just took electrons and you reduced oxygen. So we did the balance sheet. We looked at product formation and NADPH consumption. And when you divide the two, you get the coupling. So for the hydroxylase reaction, for progesterone, it's about 50% coupled. Cytochrome B5 doesn't do much. The pregnenolone reaction is similar, but a little more complicated because there's a side product form. When we looked at the lyase reaction, cytochrome B5 increased the coupling 10 times, which is the rate of stimulation in the steady state kinetics. So we can entirely account for the increase in rate by the increase in coupling. So in other words, the enzyme is not going through the cycle faster in the presence of cytochrome B5. It's just successful 10 times as often as it is in the absence of cytochrome B5. So now we go to the wallaby. So Gene wanted to know what the circulating androgens. Was the paradigm correct? Did the target tissues convert testosterone to dihydrotestosterone? And because eutherian mammals are placental mammals, you can't get to them in utero. He used the marsupial with Marilyn Renfroe and her colleagues. So the joey is born altricial, meaning that it's undeveloped from the chest on down. No legs, all right, just kind of a tail. And then it climbs up with its front paws into the pouch. And then it suckles. And during that time, the rest of physical differentiation and sexual differentiation occurs. So you can measure steroids, you can give injections and so on. So one day I come to work and Gene Wilson is sitting in the lab. And he's trying to do thin layer chromatography like Nima showed you on 10 metabolites that come out of these experiments he did in Australia. Okay, and I'm saying, Gene, don't you know we have machines that do that for you now? So we fired up our HPLC. And what we found is that the progesterone was going to 17-hydroxyprogesterone. But then it wasn't going on to androstenedione and testosterone. And Gene also showed with Cedric Shackleton that the final product was androstenediol in the circulation. Not testosterone, not dihydrotestosterone. And for reasons that I'm not sure Gene could even tell you if he was still here, he did the experiment in the presence of a 5-alpha reductase inhibitor. And it stopped at 17-hydroxyprogesterone. So the 17-hydroxyprogesterone was actually getting 5-alpha reduced in the testis. And it was being, and then those 5-alpha reduced stories are rapidly 3-alpha reduced by the enzymes that Trevor Penning studies. And those form this intermediate pregnenediol. And that was going through the 17-20 lyase reaction to androsterone, which is 17-keto, reduced to androstenediol. That was the circulating androgen, which was taken up in the target tissue and then was this enzyme, this 3-alpha HSD enzyme or rho-DH enzyme or 17-beta-HSD type 6. That was actually the target tissue enzyme that was generating dihydrotestosterone. And I know Vipka's in the audience. And when I showed this, she said, well, so this backdoor pathway that you showed to dihydrotestosterone, and so that name has stuck. Gene doesn't like that term, but we call it the alternate or the backdoor pathway because it goes through the backdoor without ever going through the conventional androgens. So Manisha Gupta in my lab did the experiment with the human enzyme. We said, well, can the human enzyme do this reaction? And so this is with P-diol. This is with 17-hydroxypregnenolone, the good substrate. And this is the poor substrate, 17-hydroxyprogesterone. And you can see that it's actually the best substrate that we've ever found for the lyase reaction. In fact, you get as much androsterone formed in the absence of B5 as you do in the presence of B5 with 17-hydroxypregnenolone pathway. So you get about threefold stimulation by B5. So we have to add now to the conventional pathway the alternative or backdoor pathway. And then, of course, as Nima showed you in prostate cancer, this abbreviated alternative pathway that goes through 5-alpha-androstenedione. Now, is this relevant to human beings? So one way to look at this is to look at androgen metabolites. So the normal fetal adrenal makes DHEAS. It's converted to androstenedione. And then that and testosterone come out in the urine as basically a one-to-one or one-to-two mixture of androsterone, which is 5-alpha, and edioclanolone, which is 5-beta. So one of the bizarre things that was being worked on around this time was why do girls virilize in P450 oxidoreductase deficiency? Well, in that case, they were accumulating 17-hydroxyprogesterone in the adrenal. And if it could be 5-alpha reduced, which has now been shown by VIPCA's group that the 5-alpha reductase is present in the fetal adrenal, that you would make pediole. And pediole would disproportionately come out in the urine as androsterone because it's already 5-alpha reduced. So we have to look at the ratio of these 5-alpha and 5-beta androgens in the urine. And Cedric Shackleton did that in mothers carrying POR fetuses. And so the hatched bars are normal pregnancies. And the gray bars are women who are carrying a fetus with POR deficiency. And the androsterone-edioclanolone ratio was increased about five-fold in those women. So during fetal life, the fetal adrenal is making androsterone in this disease. And also in 21-hydroxylase deficiency because you also accumulate 17-hydroxyprogesterone in that adrenal gland. And Clemens Kermath and Stefan Vudi showed that in the first year of life, the children with 21-hydroxylase deficiency have markedly elevated ratios of androsterone to edioclanolone, that's all these open circles. But then after about the first year of life, then you get more of a one-to-one ratio, more of the normal ratio. So at least in fetal life, at least in the first year of life, at least in POR deficiency and in 21-hydroxylase deficiency, this pathway is in fact active. And it probably actually explains the genital virilization in these kids and why kids with, for example, 3-beta-HSD deficiency do not have much genital virilization, right? They have a little clitoromegaly because they can't use this pathway in their adrenals with the 3-beta-HSD deficiency. So taking on 21-hydroxylase deficiency, what do they make after the first year of life? So this is the classical pathway with the block at 21-hydroxylase and the overflow to androgens. And you might think that this enzyme here, P45011B1, is sitting around twiddling its thumbs with nothing to do because there's no substrate getting past the block. But in fact, it reaches back and metabolizes 17-hydroxyprogesterone to 21-deoxycortisol. This enzyme is actually highly expressed in the 21-hydroxylase deficient adrenal. So what else is a substrate for 11-hydroxylase? To show that, I need my assistant here, Nemo the fish, to tell me what's in common with this girl with classic 21-hydroxylase deficiency. And the answer is that fish make 11-ketotestosterone. Now, if any of you were marine biology majors, you would know that the fish testis has 11-hydroxylase in it. And 19-carbon steroids are excellent substrates for 11-hydroxylase. I think these are sticklebacks. So this is like a good laboratory fish. And the territorial males, the breeding males, their 11-ketotestosterone is higher than the other males. It's higher than their testosterone by a long shot. So this is the breeding male hormone in the teleost fishes, the bony fishes. So they make it by taking testosterone, 11-hydroxylating, and the same 11-beta-HSD enzyme that converts cortisol to cortisone converts 11-hydroxytestosterone to 11-ketotestosterone. Well, if you have an adrenal that's accumulating 70-hydroxyprogesterone, you can force a little bit of androsterone past the block, and that androsterone is gonna get metabolized by 11-hydroxylase to 11-hydroxyandrostenedione. In fact, using adrenal vein sampling samples, we save everything at Michigan, and we see a lot of primary aldo, and we get a lot of adrenal vein blood. The third most abundant steroid in adrenal vein blood is 11-hydroxyandrostenedione, behind cortisol and DHEAS, okay? So you make a lot of this, even normal people. And then in peripheral tissues, it gets oxidized to 11-ketoandrostenedione, and then that's an excellent substrate for AKR1C3. It's actually a much better substrate than androstenedione itself, and that gets converted to 11-ketotestosterone, and that can be 5-alpha reduced to 11-ketodihydrotestosterone. So Adina Turku took a bunch of serum samples from ladies with, mostly ladies, and a few men with 21-hydroxylase deficiency and some matched controls. We developed mass spec assays for all these analytes, and you can see that their 11-ketotestosterone level is twice as high as their testosterone level. And their 11-hydroxyandrostenedione is twice as high as their androstenedione. So for the last 50 years, we've been measuring the wrong steroids in this disease because we didn't know about this, okay? Now people had looked at 11-hydroxyandrostenedione in the past, but they didn't look for their subsequent metabolites, and it didn't really distinguish PCOS from non-PCOS women, so it was sort of shelved in the 1980s. Now what's cool about this is in the women, there's a linear correlation of 11-keto-T with T because when they're in poor control, they're making both of these from their adrenals. But in the men, there's an inverse correlation because, for example, this man on the right, he's making all that testosterone from his testes, and he's not making 11-ketotestosterone from his adrenals. So he's in good control, and his genital axis is intact. This man here is in terrible control, and all that testosterone, which is close to the male range, is coming from his adrenal, okay? So that's one of the things you can use that for. I'm gonna go further here, okay? Now let's look at normal children, and if you look at girls with premature adrenarche, remember, we use DHEAS as our marker of adrenarche, and age-matched girls, yes, their DHEAS is much higher. Their testosterone is not that much different, but what is different, and what is three times or four times higher than their testosterone is their 11-ketotestosterone. So this is the androgen that causes adrenarche, okay? And it has about the same bioactivity in a reporter gene assay as testosterone, induces a very similar spectrum, not exact, but a very similar spectrum of activities as testosterone, especially those classic androgen-responsive genes. And if you take kids with premature puberty, okay, who do not have an elevated DHEAS, but they do have early puberty, yeah, the testosterones are similar, but the 11-ketotestosterone is, again, four times higher, much higher than the control children. So even in kids who aren't in official adrenarche, who have puberty, this is the androgen that is causing it. So now let's look at old people, okay? So these are pre- and post-menopausal women, and we all know that androgens go down with age. Androstenedione, testosterone, DHEA, DHEA sulfate, adial, adial sulfate, they all go down from pre-menopausal to post-menopausal women. But 11-oxyandrogens do not go down, okay? So in the post-menopausal women, the 11-ketotestosterone is, if anything, a little higher, and the ratio is definitely higher. So there is no such thing as androgen deficiency in women with adrenal glands as they age. So Adina Turku and Eldabio and our colleagues in Michigan, they looked at androstenedione, testosterone, in men and women across the spectrum. 11-hydroxyandrostenedione, 11-ketotestosterone, maybe a little decline of 11-ketot and men, but no decline in 11-hydroxyandrostenedione, and definitely no decline in 11-ketotestosterone in women. Nope, these androgens stay up. All right, I would be remiss if I didn't finish with something on 5-alpha reductase. So one of the things that Gene always lamented is that nobody's ever been able to purify the 5-alpha reductase enzyme because it's so hydrophobic. So we took this on, Wei-Ming Peng and one, Valentin Goyko in my lab, and we expressed it in E. coli with a histidine tag, ran it through the column. We were able to generate homogeneous type 1 and type 2 enzymes, and then we had to take out the detergent and reconstitute in liposomes. That was the key. So that purified protein is not active. We then had to put it back in the lipids, and we were able to get activity. And one of the interesting things we showed, Nima's done a lot of work about abiraterone metabolism, and that abiraterone can be 5-alpha reduced by the type 1 enzyme. It's not reduced by the type 2 enzyme. So now we've taken that a little bit further. We now actually use these streptactin tags rather than polyhistidine tags, and we're able to express and purify, for example, the type 2 enzyme, and you might say, what the heck is that? And it's not a dimer or a trimer, but it's actually a chaperone because if we wash the column with magnesium and ATP, it comes off, and then we get beautifully, squeaky clean enzyme. And now we're incorporating that into these phospholipid nanodiscs where we can pick the size of the nanodisc and the lipid composition. And so Sanctual-M has done this with a couple different liposomes and a couple of different nanodiscs, and we can get 5-alpha reductase activity reconstituted in this way. So tribute to Gene and his mentoring and his legacy of whom we've all benefited greatly and the society has benefited greatly. I've already mentioned a lot of the people that did the work on the 11-oxyandrogens, the enzymology and the backdoor pathway through the years, and thank you for your attention. Thank you. While people are thinking about their questions, Dr. August, I have a quick question in terms of the alternative pathway. Can you hear me now? Yeah, sure. Go ahead. In terms of the alternative pathway, we've been thinking about male sexual development and especially in terms of the classical paradigm that you showed, where does that alternative pathway kind of play in? So there's some evidence that it participates in parallel, but I think it actually is mostly involved in pathologic states in development. That's my guess right now, but those are really hard experiments to do. There's a lot of inference you have to make, but I think it's probably mostly in pathologic states. Yeah. I actually have a question while you're coming to the mic. This is a practical question. I've heard you talk about this before and I love it. The question is, for clinicians, being able to measure these keto steroids would really be a game changer for a lot of different situations. So where are we in the process? So one of my non-disclosures, I do not consult for LabCorp Esoteric, but they offer this assay commercially. So you can order on anybody in the United States through LabCorp Esoterics. You can get 11-hydroxyandrostenedione, 11-ketotestosterone, and one more in a package. There are other reference labs. I'm pretty sure that Mayo is working to get this on their menu too. We have used it. For one example was a woman with non-classic 21-hydroxylase who was having trouble getting pregnant and had very high androgens, very high testosterone despite good treatment, low progesterone, low 17-OHP. So we measured the 11-oxyandrogens. They were suppressed. So we knew it was coming from her ovaries. And so she had an ovarian tumor as a result. So that was one of the ways we've used this to determine what the source of the androgens is. Okay. Okay. Bernice. Thank you, Richard, for your beautiful presentation. I just want to join the tribute to Gene Wilson. I worked with him many years and he helps me a lot and gives me a lot of thoughts and everything good. I think, as you told, his mentoring is very clear here in this presentation with all the beautiful things we heard here. Thank you very much. Thank you, Bernice. And I think Gene also was instrumental in many international collaborations through the Endocrine Society and that's another legacy that he's left us. Just a quick question. Has anyone looked at 11-ketotestosterone in postmenopausal breast cancer cohorts? Not that I am aware of. No, I think we're just scratching the surface on this right now. And the question I get all the time is, does abiraterone lower 11-ketotestosterone? So Nima and I did a quick study. We did a study of abiraterone and 21-hydroxylase deficiency. We had some leftover samples and we were able to show that in those women that the 11-oxyandrogens went down with abiraterone. Nima had a handful of CRPC patients and on abiraterone and prednisone, it's pretty much zero. So that may actually be why abiraterone works, not because it's lowering that little bit of residual testosterone. Great. This concludes this fantastic session. Thank you to all the speakers. Thank you all for coming.

androgen receptor

breast cancer

targeting AR signaling

therapeutic strategy

AR expression

AR agonists

selective estrogen receptor modulators

tumor growth inhibition

genomic interaction

estrogen receptor signaling

alternative pathways

testosterone

cholesterol