Daniel Frigo: Oh, thank you, Andrea. Thanks for giving me an opportunity to talk about some of our recent work focused on, really, prostate cancer metabolism and signaling. And so, before I begin, I just want to highlight that this work was really led by two now former graduate students in the lab. It was first started by Chenchu Lin, and then it was really taken across the finish line by Thomas Pulliam, shown here in the lower right.
So what I was going to do today is give a quick background, on one slide, of basically why we even got into this in the first place, talk about some of the key pieces of data, and then give an overview of our working model of what we think is going on. And then provide a few couple take home points, at the end.
Our main motivation actually started probably 10 or 15 years ago. I think it was at the time when many were realizing that the transition from hormone-sensitive prostate cancer to castration-resistant prostate cancer was actually driven by this reawakening or this reactivation of, in most cases, of the androgen receptor, AR. And so, this actually stimulated quite a number of groups and companies then to come up with second generation ways to block the androgen receptor. As we know, these are effective but they're often not curative, and it's been difficult to have sustained AR inhibition in many patients.
So we took a slightly different approach than the fact at the time that we were looking for, knowing that an androgen receptor is, at its core, is a transcription factor, we were looking for downstream elements that it regulates, that could be driving the disease. And so, many of the downstream effectors are also leftover from basic development, so we were hoping that by identifying things that could be downstream, we could parse out those that were leftover from basic development to cancer, with a thought that this would then bypass the androgen receptor, if we could target these. And then also, we think that there's probably an additional bonus where, perhaps some of these things are also regulated by other pro-cancer signaling pathways as well.
To cut a long story short, we identified one particular gene called CAMKK2, which encodes for a protein called calcium/calmodulin-dependent protein kinase kinase 2. And so what this does is that, then this is a serine/threonine kinase then, that phosphorylates a restricted number of downstream targets that can then elicit its biological effect. So what we initially showed and then, soon after a couple other groups basically verified and also went on to show as well, is that this is normally low in normal prostate, high in prostate cancer. You give hormone therapy, it drops, and then it comes back up in castration-resistant prostate cancer. So we went on and showed a number of different functional studies, such as the one shown here, where you can knock out CAMKK2 in models of castration-resistant prostate cancer, and it severely impairs tumor growth. Importantly for us that, this is work that was done by Ian Mills's group, that as a kinase, it's druggable. So this is a small molecule STO-609 that was shown to have antitumor efficacy in diverse models. This has then gone on to be shown in a number of different genetically engineered mouse models, as well.
While we were interested in trying to figure out how this kinase functions too, a lot of the early work, including work from our own group, had linked it to AMPK, which is this master regulator of metabolism. And so we were very interested in that, but it was clear that that probably wasn't the only downstream effector. And so while we were looking at different transcriptomics data, we found that many of the genes that were regulated had these cyclic AMP response elements, suggesting that it was being responsive to CREB signaling. And so we were interested in this, because activated CREB had been shown to be increased in bone metastatic prostate cancer, so through a variety of different preclinical models we showed that there was a tight correlation between CAMKK2, as well as this CREB phosphorylation, which is needed for its activation. And we showed that this to be functionally true in a variety of different models I'm showing you here, in the lower left. This is genetically engineered mouse models in which we conditionally deleted the CAMKK2 in the prostate epithelium, and showed that impaired the phosphorylation of and activation of CREB.
As I mentioned, we initially thought, based on some prior work, that this was largely due to its phosphorylation of AMPK, which has also been linked to CREB. But through a series of well-controlled experiments, we found out that we were totally wrong. Turns out that actually, after eliminating essentially a number of other candidates, this is mediated through CAMK1, and particularly an isoform called CAMK1a. And so, this phosphorylation event by CAMK1 was critical for the activation of CREB, and its downstream effects in prostate cancer. And so, shown here in cells, and you can also do this in animal models as well, if you knock out CAMKK2, shown here in the red, you impair prostate cancer cell growth. This can be restored if you over-express the wild type, but not a version of CREB in which you've mutated that single phosphorylation site, Serine 133.
We went on then to demonstrate that, within the CREB family, there's actually high homology between, particularly two of the family members, CREB1 and ATF1. So as shown here, there's redundancy between these two isoforms, two family members. And so you need to, if you really want get the full antitumor effect, you have to inhibit both. And so this was shown genetically, and then also, we were interested in the fact that there was a recent small molecule inhibitor of CREB that was developed called 666-15, that targets actually both of these isoforms. And so, shown in a variety of different animal models of castration-resistant prostate cancer, as well as PDX models of enzalutamide resistance, we had sustained, basically anti-cancer benefits, in some of these pre-clinical models.
Finally, while we were looking downstream to see what was it about CREB, because it's known to regulate a wide variety of genes, what was it that was really driving it? We were interested in the fact that there's repeated signatures of cholesterol metabolism that were regulated, either when we did, when we genetically, molecularly, or pharmacologically perturbed CREB signaling. And so functionally, similar to statins shown here in the blue, and this is a positive control that depletes intracellular cholesterol, this is NBCD. Inhibitors of CREB such as 666-15, or inhibitors of CAMKK2 such as STO-609, could impair the intracellular cholesterol levels, and you could confirm this with genetic approaches that were shown in the paper. From a functional standpoint, actually these impaired, then, cell growth. And this could be rescued when you added back cholesterol, particularly in the form of HDL. We actually, there's less rescue when you give LDL, and we think that that's because this is impairing the uptake of LDL particles. At least, that's part of the mechanism that's ongoing.
And so we have this working model now, where AR, one of the things that aberrant AR signal, I should say, can up regulate the expression of CAMKK2. It can be activated by other mechanisms as well, such as things that would increase calcium influx. There are multiple different pathways, one that we identified and talked about here is CAMK1, which can then phosphorylate and activate CREB which can, probably amongst other things, to be perfectly honest, increase cholesterol metabolism. Possibly through de novo, but we also think that there's probably a role for increasing cholesterol uptake as well, and this can contribute to disease progression.
So for a few take-home points, is that this particular kinase is CAMKK2. There is an AMPK-dependent component, but this I think is more of an adaptive response. But then here, there's clearly a second pathway, this AMPK-independent pathway, that involves CAMK1 and CREB. There is high redundancy between these, so if you're going to target CREB, you have to basically take into consideration for that. But importantly, if you could target CREB, you can actually overcome resistance to existing AR-targeted agents. And so, this pathway of CAMKK2 and CREB can then promote prostate cancer progression through increasing cholesterol metabolism. And so because of that, we're interested in further pursuing this as a potential therapeutic target.
So I'll stop there, and thank you for your attention.
Andrea Miyahira: Well, thank you so much Dr. Frigo, for sharing this with us. So have you evaluated the various CRPC molecular subsets, such as AR positive, NEPC Double-Negative, or apocrine, et cetera, for a role for CREB and CREB targeting? And in what stages of molecular subsets is this pathway acting to drive prostate cancer?
Daniel Frigo: So those are important questions. So we focus largely on AR-driven, more classic adenocarcinoma, because of the link between androgen receptor direct regulation of CAMKK2 expression. And then when we did the associations in diverse preclinical models, you could see a strong association between AR, adenocarcinoma, CAMKK2, as well as the CREB.
Now that being said, there were completely AR-independent negative neuroendocrine prostate tumors that also would express CREB. And interestingly, there's work done by one of my colleagues here in Houston, Wenliang Li, who's over at UT Health, had done work showing that CREB is important in response to anti-androgens, and then activation in neuroendocrine, basically treatment-induced neuroendocrine prostate cancer differentiation. So as we also basically touched on within the paper, too, it's interesting that you have both androgens that could strongly promote CREB activation, but so can anti-androgens, which seems like it's a paradox. But actually, this is through two different pathways. The level of activation is much stronger, certainly in our hands, in terms of androgens. And we have a very direct regulation. It's a little bit more amorphous, but it kind of speaks to the fact that, no matter how you go about it, the prostate cancer cells will rely, need CREB signaling. And so that's why, to some degree it's kind of an attractive target, because you see that sort of all roads come back to CREB in this regard.
Andrea Miyahira: Thanks. And what are the normal functions of the CREB pathway, and what side effects would you anticipate by blocking CREB and ATF1?
Daniel Frigo: Yeah, so I don't doubt that targeting CREB will undoubtedly, if you can target CREB in the cancer cells, it will have an effect. I think it is going to come down to the therapeutic window question. And so CREB does have broad effects throughout the body, which will have to be considered. So normally two of its biggest roles are, for example, in the brain. In terms of memory and learning. One would think that if you were going to target CREB, you could avoid this pharmacologically by avoiding things that cross the blood-brain barrier.
So the other big thing though, is that you'll see CREB plays quite a significant role in metabolic tissue. So things like liver, pancreas, muscle, where it's known to regulate glucose and lipid metabolism. We did actually, we were worried about this, and actually we did some studies that were reported part of the paper. And we did not see, with the inhibitor, substantial effects on, for example, various serum blood markers of overall metabolic homeostasis, glucose metabolism. But truth be told though, we never did it in the context of fasted, which I think fasted mice, et cetera, which could be important too.
There are certainly other roles for CREB that are important in development, in this regard though, we think it actually might be less of an issue because obviously the patient population is much older. One would imagine then, if you were actually going to look for side effects, I think I'd want to look for cognitive functions. But then, in particular, this what happens with the systemic metabolism as well.
Andrea Miyahira: Okay, thanks. And can the impact of this pathway on tumor cell metabolism be leveraged therapeutically, or possibly even with lifestyle modifications?
Daniel Frigo: Yeah, I think one of the things I'm interested in, in terms of more near-term clinical benefit is, there's a lot of interest from a variety of different sort of real world studies and retrospective analysis of use of metabolic targeting things, for a variety of conditions. And so a prime example is the use of statins. So there's been a lot of interest in possibly, does statin use benefit men with prostate cancer? And could actually, do statins themselves have direct antitumor activity? In fact, there's a large Phase 3, randomized Phase 3 trial that's now ongoing in Scandinavia that's directly testing whether or not statins can promote, or basically improve, have anti-cancer effects in men following the advent of hormone therapy. I would love to know whether or not, could you enrich for responders, if you find tumors that have high, for example, CAMKK2-CREB signaling within those tumors? I would predict that you would. And so maybe it's a way to guide some existing therapies that are already quite widespread, that a lot of our patients are taking.
Andrea Miyahira: Okay, thanks. And what are your next steps?
Daniel Frigo: Yeah. So one, I'm still not a hundred percent clear what aspects of cholesterol metabolism are really, truly regulated here, so we want to do a deeper dive in that regard. And then there's also the therapeutic angle. We talked about some of the potential side effects. I'm a little bit leery about whether or not, because of potential widespread functions of CREB, if that's really a viable therapeutic target. Plus it's oftentimes difficult, as you may know, to target transcription factors that have some of its own physical challenges and chemical challenges. But that's not to say that, one could imagine a scenario where you can selectively deliver a CREB inhibitor to that.
We've actually been quite a bit interested in developing CAMKK2 inhibitors. And we think that there's benefits there, because one, it has more of a restricted expression profile. And so we think that there'd be fewer side effects if you come up with truly effective inhibitors. We know that it's also highly druggable. And then, there's some additional benefits for CAMKK2 functions in terms of CAMKK2 inhibition actually has benefits for countering metabolic syndrome and bone complications, which are two of the common comorbidities in prostate cancer. And there might also be some positive effects with regards to reactivated immune system, which we're pursuing now, as well. So there's probably more to say on that in the next six months with the development of next generation CAMKK2 inhibitors, and marching those towards the clinic.
Andrea Miyahira: Okay. Well, I look forward to the next study. So thanks so much, Dr. Frigo, for sharing this with us today.
Daniel Frigo: Certainly.