How Does Androgen Biosynthesis Occur Independent of CYP17A1 and CYP11A1?
SPECIAL LECTURE: How Does Androgen Biosynthesis Occur Independent of CYP17A1 and CYP11A1?
Nima Sharifi (University of Miami)
Introduction by Howard Soule (Prostate Cancer Foundation)
View the Transcript Below:
How Does Androgen Biosynthesis Occur Independent of CYP17A1 and CYP11A1
Howard Soule, PhD [00:00:11] Our next speaker is Nima Sharifi. There’s Nima. Nima’s been a part of the PCF Enterprise and our family for a long time now and is a very productive investigator now at the University of Miami. And I would have titled this Just When You Think You Knew Everything About Androgen Synthesis Here Comes Dr. Sharifi. So, thank you.
Nima Sharifi, MD [00:00:48] Thank you so much, Howard. I want to start by giving immense gratitude to you, to Andrea, to PCF, that really the support really started from the inception of our research program back in 2008 with the Young Investigator Award. So, I want to share with you a new story on androgen biosynthesis. This may be the first chapter of this. And what I’d like to do is basically start with the conclusions of the talk. So, you know exactly what I would like for you to take away from the data that I present to you. So, the background here, as you all know, is that in all endogenous androgens vertebrae are thought to require a CYP17 for biosynthesis. Of course, we have a variety of different treatments blocking gonadal and adrenal androgens that extend overall survival. But we also recognize that inhibition of all known androgen sources ends up not being complete. And we think we’ve identified a CYP17 independent pathway of making androgens that appears to require oxysterols, and it also appears to require a different p450 enzymes, CYP51. So, these are rain lilies, and the distinguishing feature of rain lilies as opposed to other types of flowers that bloom in response to sun is that rain lilies bloom in response to long periods of heavy rain. So, this is a seven-year-long story that’s I think now at least part of it is coming to fruition, led by this person, Ziqi Zhu in our group, but also enabled by many people in our research program. So, the canonical pathway starting from cholesterol going all the way to androgens is right there. And you can see that the very first enzymatic step that’s necessary is catalyzed by CYP11A1 or sidechain cleavage enzyme. And its job is to take off those six carbons off the cholesterol side chain. The second enzyme is a target for abiraterone or CYP17, catalyzes two reactions that together takes off two more carbons. So, this results in the generation of 19 carbon androgens. All androgens have 19 carbons. Here, this is DHEA, but the other androgens have a similar structure. But there’s a different pathway as well that actually doesn’t require those first two enzymes and instead requires CYP51 that appears to be capable of taking off those eight carbons off the side chain at once. So about 15 years ago, we had the first FDA approval for abiraterone, a drug that was pioneered by many people in this room. That was first for metastatic CRPC that was a taxane treated, and then of course, subsequent approvals, including for upfront treatment with initiation of systemic therapy. So, this is a really important part of our therapeutic armamentarium, obviously. But here’s a clinical problem, and these are data from Mary-Ellen Taplin and the groups in Boston and Seattle from a neoadjuvant clinical trial that they conducted beautifully. And this is looking at residual androgens in the prostatic tissues after either medical castration alone, that’s in the yellow bars, or medical castration plus abiraterone. So, the top bar graph there is testosterone, and you can see that it declines when you add CYP17 inhibition, but it’s not, you know, it’s not the magnitude of the effect isn’t dramatic. And you might say, well, DHT is more important, and so that’s on the bottom, and you get a better proportional decline, but it’s a subtotal decline, right? So given what’s left in terms of the androgens and the blue bars, it’s not hard for me to imagine that some of this might be responsible for resistance, at least in some patients. So, the first question I think we tackled as a field is can we do better with CYP17 inhibitions? So, there were, as you know, a variety of different CYP17 inhibitors that were tested, and none were clinically shown to be better than abiraterone. And then the question was should we go upstream to CYP11? So, Orion Pharma developed a CYP11 inhibitor, generated the clinical data, and of course. This is now with Merck, and we have two worldwide randomized phase three clinical trials. But let’s take a look at the process here a bit more carefully in the C4-2 human prostate cancer CRPC line. And here we’re culturing the cells, looking at the androgens inside the cells in vitro, either with charcoal strip serum or serum-free media, and the androgens you know appear to persist in spite of complete deprivation of serum. If we do the same thing and we incubate the cells with pharmacologic inhibitors of CYP17, we don’t see that it really puts a dent in the intracellular T and DHT. And so, at this point we were a little bit paranoid that the androgens that are there might be carryover from when they were first cultured in androgen replete conditions and the ARs binding to the androgens tightly and just not letting go. So here, so this is three beta HSD1, an enzyme that’s required to convert the precursor conformation of DHE and cholesterol, for example, to T. If we genetically get rid of that enzyme, those androgens go away. So, what this tells us, or at least tells me, is that an active biosynthetic process is necessary and that the androgens that we measure in this context, it’s not just a carryover. So, we wanted to look at this more carefully with stable isotope tracing studies, and we used C13 labeled cholesterol that’s labeled, the C13 is on the A ring there. If that’s converted over metabolically in the cells, then we should be able to detect C13 labeled T. And you know exactly where that’s coming from because we gave it the cholesterol. And we can detect this by mass spec, same retention time, three Dalton difference as you expect. And when we do the experiment, you can see clearly in that middle bar that when you add, when you culture in C13 labeled cholesterol, you do get the generation of C13 labeled testosterone. So, how are androgens synthesized from cholesterol independent of CYP17? Is that really possible? So, I think to approach that question, we have to basically step back and take a fresh perspective about you know, how do we know about CYP17? This really predates the abiraterone. So, before we, you know, approach this in the clinic, this is the first biochemical description of germline loss of function mutations of CYP17 described in 1966 at UCSF, right? So, for people who are born with loss of function mutations in CYP17, they can’t make androgens or estrogens. And so, they described, Biglieri described a subtotal loss of androgens in their physiology, right? And you get endocrine anomalies and other things. So, at this point, we had to basically reimagine how androgen synthesis could occur. And for, you know, this is a difficult process. So, for anybody who does mass spec or metabolism work, you know that if you’re doing mass spec for metabolites that are known, that are polar, that are abundant, that ionize really well, it’s cake. It’s pretty easy, right? So, TCA cycle intermediates, it’s not very difficult to do. In contrast, if you’re looking for things that are lipophilic, have molecular weights that are very close or identical to one another, don’t ionize well when you zap them to look for daughter ions, and are unknown, it’s incredibly challenging. One might say, you know, a treacherous swamp that you’d have to traverse with some pitfalls there, you can see that there’s an alligator hiding back there. Regardless, we decided to proceed, and our guess or hypothesis was that oxidized forms of cholesterol or oxysterols could be substrates for androgen biosynthesis. So, oxysterols, there’s a variety of them that are known. You know, the functions are sometimes worked out, sometimes not. But I just want to draw your attention to the one that’s right next to cholesterol, which is 1720 dihydroxy cholesterol, or DHC for short. And I want to fast forward through a lot of headaches that we went through just to tell you that DHC is a substrate for androgen biosynthesis. You put it onto cells, you get the generation of DHEA, testosterone, DHT, and others. And with the time course experiment, you can see that within two hours you see the appearance of DHEA, and with some delay, the generation of testosterone and DHT as well. So just very briefly for those interested in breast cancer, we know that in order to make estrogens, you have to make androgens first. So, we just wanted, we were curious if you also express aromatase, can this go all the way to estrogens? And the answer is yes. This is not a prostate cancer model, but I just want to show that that this can go all the way to estrogen biosynthesis, and I won’t talk about estrogens any further. We wanted to look at this even more carefully. So, we made a deuterated form of DHC. And we asked, well, can we follow this down to deuterated forms of androgens? And the answer is obviously yes. And when we treat with abiraterone, it doesn’t block the conversion to deuterated DHEA, right? So CYP17 appears to be not required. And then we treat it with a more generic P450 inhibitor, ketoconazole, and that appears to suppress a variety of different androgens, not completely, but it does suppress them. So, this suggests that there’s a different P450 enzyme that’s involved. And we don’t want to stop there. So, TG expressed all 57 human P450 enzymes. Asking which one, if any, can convert that oxysterol to androgens. And obviously there’s a singular very clear outlier, and that’s CYP51 that’s capable of making the androgens. CYP17 doesn’t do it, and CYP11 doesn’t do it, and the other P450s don’t do it. So, when he overexpresses CYP51, you get increased generation of deuterated androgens from deuterated oxysterol or DHC. And if you’re an enzymologist, you might say, you know, I don’t care about all the stuff that happens in the cell. If you really want to attribute this activity to that enzyme, you have to purify it and show me that that enzyme alone is responsible for that activity. So here on the left, you can see that you need DHC, the substrate, CYP51, and the cofactor, NADPH, and then you get the generation of DHEA from that oxysterol. And on the right, you can see that the reaction is saturable, which is consistent with enzymatic activity. So CYP51 is actually the oldest P450 evolutionarily, or at least that’s what most people think. And we wanted to find out whether this unusual androgen synthesis activity is specific to humans or if it occurs across multiple organisms. So, we tested a variety of different CYP51 orthologs. And the only place, and this is not comprehensive, obviously, but the only place we really appear to see androgen biosynthesis activity by CYP51 is in vertebrae, right? And we think it’s probably not an accident that AR and ER evolved, came in through evolution in vertebrae and not the other organisms. Not sure about that. We need to investigate it further, but that’s our suspicion. So, the last few slides that I shared with you are starting with the oxysterol. Now we wanted to figure out what if we start all the way back from cholesterol and give C13 labeled cholesterol. Do we still see that CYP51 is required to convert cholesterol all the way down to label testosterone? So, when we genetically get rid of CIP51, the generation of C13 labeled testosterone goes away. So, the answer appears to be yes, even when you start with cholesterol. And then, you know, obviously we want to know does this play any role in in resistance to hormonal therapies? These are Subcue Xenograft models in C4-2 and castrated mice. It delays the development of CRPC. And when we look at the tumors that come out, you can see that there’s a reduction in expression of PSA transcript and protein. And then finally, I called Alex Wyatt and I told him about the story, and I said, we don’t know anything about the human genetics, can you help us out? So, he and Ruby Liao generated these plots. From 730 men at cell-free DNA from 730 men with metastatic CRPC who have at least 15% tumor DNA in circulation. And what appears to be the case is that about 17% of these cases have an increase in copy number of CYP51. So again, there’s a lot more to be worked out, but that’s sort of a beginning here. So, I started with a conclusion, so I’ll be fast with the conclusions here. On the right, you have what was known about androgen biosynthesis from cholesterol requiring CYP11 and CYP17, but there’s a different pathway entirely that doesn’t require those enzymes and may require CYP51. And really the most important slide of the talk, which is to thank the people who did the work. I already mentioned Ziqi, who’s sort of in the back there, but also many of our group members contributed to this work. Want to thank our also our collaborators, Rich Auchus at University of Michigan, collaborators at Vanderbilt, Fred Guengerich, Galina Lepesheva, and others, Alex Wyatt and Ruby Liao, the leadership at our place, Dipen Parekh for Desai Sethi Urology Institute, and Stephen Nimer for Sylvester Comprehensive Cancer Center, and most importantly, Prostate Cancer Foundation, Howard, Andrea, thank you for your immeasurable support. I think we’ve often been accused of being myopically focused on this area. So, you know, whatever you think about this work, we wouldn’t have gotten to this place unless you had supported it. So, thank you for almost two decades of support. I’d love to take any questions.
Unknown [00:15:58] That was a really informative, excellent talk. So, there are some patients who are on ADT and even ADT and ARPi that still have, you know, low measure testosterones, 20, 30 versus others. It’s extremely low. So, I’m wondering if you ever you looked at you know circulating DNA in in those patients. And the other question is, have you identified any polymorphisms in these genes that may, you know, give you a clue as to why you see differences in testosterone synthesis?
Nima Sharifi, MD [00:16:27] So great questions. We haven’t looked at self-re DNA. I mean it’s not it’s not sort of our forte. I want to just be clear that the data that I shared with you is primarily basically cell autonomous synthesis of androgens, not coming from endocrine organs. I mean there may be a contribution from endocrine organs. I don’t know. We have looked for polymorphisms, and we don’t see anything that’s frequent that we suspect may be a contributor, but it’s possible we’ve missed something. Great questions.
Unknown [00:16:57] Hey Nima, Question for you. The conversion when you showed the C13 label testosterone, that was in cells. Have you tried that in vivo?
Nima Sharifi, MD [00:17:07] Not yet. Not yet.
Unknown [00:17:08] ‘Cause that would actually be really cool to see number one where it’s made in vivo, and also if you knock out those genes, even in maybe in organoids or something, that you don’t have to have a full knockout. That would be very interesting.
Nima Sharifi, MD [00:17:22] Absolutely. Yes. Thank you. Hi.
Unknown [00:17:27] So my question was in your C4-2 experiment with the xenograft where you knocked down the CYP51, why did CYP17A1 not compensate for the loss of CYP51 to make some androgen?
Nima Sharifi, MD [00:17:43] Why does CYP17 not compensate for CYP51? I don’t know. I’m not sure.
Unknown [00:17:52] Thank you
Kara Maxwell, MD, PhD [00:17:54] Hi, Kara Maxwell from Penn. That was really beautiful. Do you have, and you might not know yet because you just mentioned you haven’t looked in vivo yet, but do you have any idea of what lipoproteins might be delivering the cholesterol? Do you think it is, you know, exogenously delivered or synthesis in the prostate cancer?
Nima Sharifi, MD [00:18:10] That so, you know, even when you go back prior to prostate cancer, just looking at traditional endocrine tissues, right? So Leydig cells of the testes or the adrenal, it’s not clear how much of a contribution to steroidogenesis comes from basically uptake from HDL versus de novo biosynthesis. So, we don’t really know. That’s a wonderful question.
Unknown [00:18:34] Beautiful talk. I just wanna ask about the other oxysterol 27 and 25 hydroxycholesterol increased by adiposity and obesity in patient. Do the 7 dehydrocholesterol increase?
Nima Sharifi, MD [00:18:49] We tested those, those we don’t see that those are converted to androgens. It’s a great question.
Unknown [00:18:55] Do you have an increase in oxysterols in cancer? Did you look at those whether you know prostate cancer tissues have an increase in oxysterols?
Nima Sharifi, MD [00:19:05] So we actually don’t detect any DHC. And it’s not clear if it’s because of our assay. You know, sometimes when things are rapidly consumed, you don’t get accumulation or detection of intermediate metabolites. So, there’s a lot that we really need to figure out. But I think it’s pretty clear both the xenograft models and then in the cell lines, even in the absence of adding the oxysterol, you require CYP17 to get all the way down to the androgens. It’s a great question.
Unknown [00:19:35] Of course. So, building off that, this was amazing. Like, you know, we always talk about mechanisms of resistance, and we put up intertumoral synthesis and then we don’t really say much about it and except for you’re saying more about it. But I really want to know about what’s the regulating CYP51, what’s turning it on? You just said the oxysterols aren’t really around, so I would normally think LXR or something like that would be regulating this pathway.
Nima Sharifi, MD [00:20:05] Yeah. So first of all, I wouldn’t say that the oxysterols aren’t around. I think it’s difficult to characterize them from an analytic standpoint. And we have to do a better job at that and we’re working on that. What upregulates CYP51, you know, according to some of Alex’s data, maybe it’s an increase in copy number. There may be there certainly may be a factor from sterol or SREBP that, you know, a lot of the literature is actually quite old in terms of CYP51 regulation. So, I don’t know, but we’re looking at that right now. Yeah.
Unknown [00:20:38] Nima, beautiful story as always. Thank you. Where do you think DHC is made? Is it just within the tumor? Are there other organs that make it? And is there a normal function?
Nima Sharifi, MD [00:20:50] It’s not well characterized. And it may be that maybe DHC is not the correct intermediate. Maybe there’s another intermediate that serves the same function. Again, there’s a lot that we need to work out here. It’s a great question.
Unknown [00:21:05] Thanks. A fantastic talk. Thanks. Is there any cell line dependency, at least in the models you tested? Because I only saw the data for C4-2. Have you tested different cell lines and is this somehow related or like what I’m saying is it a universal phenomenon or like is it, can be, tumor specific?
Nima Sharifi, MD [00:21:30] You know, it’s expressed variably throughout various different cell line models. So, to what extent does each model depend on CYP51, I don’t really know. Thank you.