New Frontiers in Therapeutic Peptides - Dr. Karsten Eastman

Hi everybody, and welcome to this week's Talking Biotech podcast. Now, peptides are strings of amino acids. They're small strings, but they can be very important in metabolism and in physiology in general. And nature makes a lot of peptides with important roles, from things that are constituents of bee venom to a really potent one that's found in mushrooms that shuts down transcription, right? At least from polymerase II. So these peptides interact with signaling systems and other aspects of cellular function. In recent times, We've heard a lot about peptides like GLP agonists, and they've revolutionized weight loss and likely will change everything from healthcare costs to the price of a plane ticket to how much the junk food companies make in the stock market. There's a lot of things happening around peptides, but there are inherent limitations to peptides, and at least their use in a therapeutic context, everything from their stability to localization to bioavailability, all these goodies are affected by different aspects of peptide biology. So what if that could be controlled? And so today's guest will tell us about that. We're speaking with Dr. Karsten Eastman. He's the CEO and co-founder of Sothera Therapeutics. Wow, that's almost hard to say. Welcome to the podcast, Dr. Eastman.

Thank you so much for having me, Kevin. It's great to be here, and I'm really excited to share a little bit about peptides and what we're working on to make them a better therapeutic modality.

No, very good. It's actually pretty exciting stuff because in plant biology, we've seen such an emergence of peptides in so many different places where we never knew where they functioned before. So I'm not as up to it in the animal literature. I certainly know the big players, but maybe a good place to start would be what are the different roles for peptides in biology and animal biology or in disease states, normal physiology?

That's a great place to start. And I think the most interesting place is just think of it in terms of insulin. Insulin is one of the most important peptides and has been critical for treating diabetes for a very long time. And I think everyone has an immediate recognition to why that particular peptide would be so important. If you are diabetic, you need insulin to survive and live. So that's A really potent piece to realize is peptides are responsible for keeping us alive at the baseline minimum, but they're used in many other ways too. They're signaling molecules, neuro— they can be used for neural transmission. They activate different receptors, and they themselves can be functional units that can be appended onto different cells or different proteins as an additional tailoring piece. Peptides are essentially everywhere in biology and physiology in human health.

And so the problem I think that comes to mind right away when you bring up those important contexts is that, isn't it just important where they're expressed and how long they stick around? I mean, you don't want insulin coursing through your veins all the time, right?

That's absolutely right. If you have insulin sticking around all of the time, your blood sugar levels will completely deplete, and then you also end up in a really bit of a pickle that can also be quite bad for you. Peptides are great in terms of biological signaling and control. They typically have very short half-lives, so they get in, they have the effect that they're supposed to have, and then they're very quickly degraded back to individual amino acids and recycled for other uses, which, as I mentioned, is great for biology, but typically very bad for therapeutics. If you have a therapeutic, you want it to stick around for a long time so you get that continued biological response. If you don't have that, well, then you don't have a very good therapeutic.

And along that line, what are the usual mechanisms of degradation? Are they going through ubiquitination and turnover in that kind of proteolysis cycle?

Typically, no. Well, that is one way of it being degraded, but that's more in line for larger proteins. Peptides themselves, because they're usually unfolded, undergo proteolysis. So they can be chewed from either the N or C terminus. There are specific residues that will get cleaved, and they'll just be broken down into smaller parts. But ubiquitination is not a typical method for peptides to be recycled and degraded.

And have efforts been made to kind of, let's say, fortify the ends of peptides or even inhibit the specific proteases? Are those parts of any kind of therapeutics?

Yes, actually, that's a good way— that's a really good point you bring up. One thing that's been done somewhat recently is I believe Eli Lilly appended a polyproline tail to their GLP-1 therapeutic. And those prolines, since they're so rigid, if you have multiple prolines in a row, it can actually act as a way of stopping that C-terminal processing. So there are ways of disguising peptides from proteolysis. Another way that's commonly used is the introduction of unnatural amino acids. If you have something something that nature has never seen before, the likelihood of it being processed or processed quickly is much smaller.

Yeah. Are there even possibilities of using the— now, see, you have D and L-form amino acids, and I always get it backwards which one is which, but are they using like L-form instead of the naturally occurring forms to make mimetics that maybe can do the same thing but are less likely to be degraded?

So yes, if you use the D-form of the amino acid, which is the enantiomer, that is a great way of having something that has similar chemical properties but is less likely to be recognized by the biological machinery that would degrade it. One other option is to use a peptoid, and peptoids have their amino acid side chains coming off of the backbone nitrogen rather than the alpha carbon.

Hmm.

So there's lots of ways to I guess, hide or disguise peptides to make them better therapeutics.

Yeah. So just to kind of dial it back for folks who maybe aren't protein chemists is that we're making things up of amino acids. They're following a very predictable backbone sequence of NCC in each amino acid. So nitrogen, 2 carbons, where the R group or the part that makes it a specific amino acid, the chemical moiety that makes it an amino acid comes off that central carbon, and then these things are strung together. So the idea here would be to have it come off the nitrogen rather than the carbon, which would give the overall structure a similar chemical property in the way it would fold and interact with its immediate environment, but not necessarily be recognized by the proteolytic— or I'm sorry, by the proteolytic machinery that would degrade it. Do I have that right?

That's absolutely right. That's a great way of summarizing it and reframing it. That's good. Thank you.

Yeah, no, it's important. I hate to lose a listener because we start geeking out too much.

It's really easy to geek out. So reframing is always excellent.

Yeah, it is. But we would have an audience of 3. There's like you, me, and a couple of people who are way into this. But I guess maybe a good place to start Heading towards the technologies that you propose are to look at an important class of peptides that we've heard a lot about recently, and that's the GLP-1 therapeutics or GLP-1-like compounds. And can you give us a little bit of background exactly what these are and why they have been so important?

So it's glucagon-like peptide, which is what they're all modeled after essentially, which plays an important role in regulating appetite and satiety mechanisms. What these peptides do is they essentially make you feel more full and suppress that hunger feeling that you get, while some of them also have shown signs of increasing metabolism. So you, you'll burn through your energy stores quicker. The combination of those 3 pieces makes it very easy to see how these peptides are so important for weight loss.

Yeah, that's really interesting that they're showing roles for weight loss and appetite suppression, but also a lot of other interesting things that happen because there are GLP receptors in the brain. And I guess that they've had other roles with things like gambling addiction or nail biting or everything from alcoholism to— so what are some of the other potential applications they may see here?

Well, then we start getting into areas of, you know, the limbic system and, you know, reward and consequence type interactions. And I think that once you start pulling things towards brain chemistry, it becomes a lot more difficult to study. For example, I know that Novo was looking at using them potentially for Alzheimer's, but some of the clinical readouts there didn't necessarily have as strong of a correlation as they would have hoped. So I think that There are options here. And if you think about it, your reward system is often tied to food. I mean, food tastes good. If you've had a really stressful week or a stressful day, sometimes going and getting your favorite snack is a great way to get that burst of brain chemicals to help you feel better. But if you sort of have a dampening on just the baseline feeling of, oh, I need to feel good, or, oh, I'm hungry, I need to satiate this signal, I think that some of these second-order effects start to make a little bit more sense, but they're a lot harder to operationalize and quantify simply because then we're entering brain chemistry. And that's, that's a very different space than, say, modulating things related to, you know, hunger and appetite.

Yeah, actually, I never really thought about that before. And I've heard a lot about these compounds and been really interested in them. And the idea of feeling hungry maybe is something that is coming from You know, one, uh, one aspect of this, but the decisions you make when hungry are probably coming from the brain and what you start to feel satiety and, um, other aspects of, you know, I don't want to say food addiction, but, uh, the choices that we make that are influenced through, let's say, executive function tying into that. And so that's where this gets to be a little bit more messy, but still it makes perfect sense when you talk about the other behaviors that are kind of controlled through GLP-1 agonists.

Yeah, I think, I think there's a lot of places that this can be compounded. So using these GLP-1s as an additional piece where maybe you just need one other supporting medication to really help some of these other brain-regulated behaviors. So I think it makes sense, but right now there's— that needs a little bit more time before we can start making really well-educated conclusions about what they're doing and how they affect the brain chemistry.

Yeah, so, so far we've really framed the problem. We've talked about the idea of peptides and what they are, and at least one common place where they're really playing a major role in society and in public health. And so we'll take a break here, and when we come back on the other side of the break, we'll talk about the ways that Sathera is changing the way peptides are regulated and how that improves their therapeutic This is the Talking Biotech Podcast. We'll be back in just a moment. And now we're back on the Talking Biotech Podcast. We're speaking with Dr. Karsten Eastman. He's the CEO and founder of Sathera Therapeutics. We started out by talking about peptides and some of their current applications and some of the inherent problems with peptide stability and why peptides are not always the best choice for— well, best choice for therapeutics, or some of the, I should say, The limitations to using peptides as therapeutics. There we go. But you mentioned this, um, the presence of the peptides throughout biology and how important they are in certain contexts, uh, in animals and certainly in plants we know they are. But what are some of the ways that nature pushes back against the degradation?

There's 2 main ways that nature has really pushed back on this. The first and the simplest is using cysteine to form a disulfide. So cysteine is one of the naturally occurring amino acids, and the side chain contains a thiol, so a sulfur and a hydrogen. When you have 2 disulfides next to each other in space, they will actually link together, and this forms what's known in the peptide space as a disulfide macrocycle. And that's the simplest way. A more complex way that nature does it is it will actually build large machines of multiple proteins that will build a peptide through non-conventional manners. So rather than it coming off of a ribosome, you have something called an NRPS, so a non-ribosomal polyketide synthase, which will build very diverse and complex peptides that often have one or more unnatural amino acids. But that's, that's a very difficult task to do in nature.

Well, when you talk about the disulfide bond and the formation of the disulfide bond. What are some of the— so just for the listener, these are typically used in stabilizing protein structure, right? So these are built to kind of give it a little bit of shape so that the business ends of the molecule are in their correct catalytic conformation or protein binding conformation, whatever. But what are some of the inherent problems with a disulfide bond? I mean, doesn't it require a certain environment in order to be able to form and maintain itself?

It does, yes, and that requires an oxidizing environment. So wherever there's oxygen, you usually get disulfides. However, in the body, oxygen is, you know, one of those problem molecules actually, and that's why we hear so much about, oh, you want to take antioxidants to ensure that you're healthy. Most cells are known as Reducing. And what that basically does to a disulfide is it will take this link and then pop it back off, which means that the peptide can then be proteolyzed like we were talking about earlier. So disulfides themselves are fairly weak bonds that are usually broken in the context of the body, especially in certain locations.

And so how does pharma get around this? I mean, if you want to use peptides therapeutically, you have to somehow trick this cellular environment, but also the natural mechanisms that break it down. So how does pharma address that?

Well, bluntly, for a long time they didn't. They would form the disulfide and just accept that the inherent half-life of these peptides would be extremely short. And that is really one of the main reasons why peptides as therapeutics haven't really taken off. However, chemistry has come a long way, and we do have a few more solutions that have been proposed and been accepted. One of the most common is head-to-tail cyclization. So on, say, the initial point of a peptide, you have an amine, and then on the terminal side, you have a carboxylic acid, and you can link those 2 together to actually form a head-to-tail additional peptide bond. That's one way they've solved it, and that's, you know, gives it a little bit more half-life. But this brings us to the idea of, okay, that's one way to do it, and you get this macrocycle that's essentially linked, and on that macrocycle, you have all of this chemical diversity on the outside, But one of the biggest lessons that biology teaches us is structure equals function. And if the only architecture that your screens or you're able to produce through traditional chemistry is a circle, that's really good for certain contexts, but it's inherently limiting in terms of higher-order architectures. And I think one of the most clear ways to think about this is There are now antibodies, which are biologics, not peptides, but they're still made of amino acids, that can go after more than one target simultaneously. So these are bispecific antibodies, and those are showing great promise in the context of many difficult-to-treat cancers.

Well, let's go back to that idea of that N-C linkage. Is that something that is synthesized in the laboratory, or is that something that can be catalyzed enzymatically inside a cell?

So that's typically synthesized in the laboratory. One of the most common ways to do it is an on-resin synthesis, or to just push that particular catalytic piece forward through an activator. But I mean, anything that happens catalytically could happen in the context of a cell, but usually you have to have an activation in order to install the next amino acid. And similarly here, there isn't an activating event that would lower that energy barrier such that it would happen spontaneously, at least not in many cases.

Yeah, it's interesting. I've never heard about that before. And when you mentioned it, I go, you know, that's true. You could do that. And it's surprising you don't see more proteins linked to each other because, you know, you have everything there to make that peptide bond. But let's— all right. So you come up with this idea that there are ways to limit the breakdown and then increase the persistence and maybe the therapeutic value of different peptides. How has Sathera got around that to solve that same question?

So what we've done, and the foundation of our platform, is we are interested in accessing diverse chemical space. So rather than us being limited to a single macrocycle, we've tried to find ways of, say, accessing architectures that are like insulin, or accessing architectures that are similarly to how a bispecific antibody is constructed. constructed. But in order to do that, you need to have control of the chemistry. If you try to make a very complex molecule that has multiple macrocycles layered on top of each other, if you do this through traditional synthetic chemistry, you have to go through what I lovingly like to call orthogonal protecting group gymnastics, which is just a chemist's fancy way of saying It's really slow, it takes a lot of time, and you have to have multiple steps of purification in order to get to your final product. We use enzymes to do a one-pot reaction to access that higher-order diversity, which means that rather than needing to really slowly do one peptide at a time, we can make diverse peptides, have them All controlled, all basically processed in a controllable way and get the same products every single time. So that's what really sets us apart is we are going at it in an enzymatic means rather than traditional synthetic chemistry.

Okay. So these are all in vitro ways of generating peptides that are essentially configured together. Or are these single peptides that are locked in different states where they're either active or inactive? Maybe there's a domain that would be protected before it would be subject to catalysis by the body or binding a receptor. Is that the idea that you're protecting different active domains of peptides, or is it all putting different ones together?

I think it's both at the same time. So, we can do both. Of course, so when you have a therapeutic candidate, you need to ensure that you can produce a lot of that therapeutic candidate and its domains are locked. So yes, that's where you would say synthesize a single peptide and a lot of that single peptide, use the enzyme to process it, and then have your new active, say, bispecific peptide moiety. However, we think that the real value here is there are many architectures that have never undergone high-throughput screening because those architectures cannot be accessed in a high-throughput Meaning if you tried to do the screen, you wouldn't be able to deconvolute what comes out of it. So the way that we're operating right now, and it's alluding more to your— the first point you were talking about, is we make, say, 1 trillion unique peptides, and all of those unique peptides have a barcode on them, and that barcode is what is the encoding sequence.

Okay.

Right. We can amplify nucleic acids through PCR. And I think that everyone learned way more about PCR than they wanted to during COVID But we can amplify those barcodes. So when we do this trillion molecule screen on a target of interest, and wash everything that doesn't bind away, we just amplify that barcode that's left over and can sequence it and get great results. Now, that's what I just explained is traditional mRNA display screening. This is something that's been done. This was kind of an evolution from phage display, which is where, you know, you use— you hijack viral machinery to display lots of peptides, and then you can amplify the genome of the virus to see what you've got. But when you do traditional mRNA display screening, you're still left with linear peptides. And linear peptides are floppy, and they don't spend a lot of time on target, which means that when you are washing your target away, there are a lot of peptides that in theory could be a potential therapeutic but aren't. So our spin on it is, after we have these huge libraries, we use our enzyme to tailor all of the library at once to make it more drug-like.

I see. So it's a question of taking— which, which was a Nobel Prize-winning Yeah. Idea, you know, this idea of phage display. It's things that we kind of stumbled into too, a little bit here in the lab years ago. But they take this basic idea of mRNA display where you're generating some sort of a peptide that could be potentially therapeutic, yet now you're able to stabilize its shape so that it is More likely to be a therapeutic, or will have therapeutic value, right? Help me out there a little bit.

Yeah, so it's not that the peptides that we pull from the screens themselves are going to be the therapeutic entity, but it's that it gives it a much stronger starting point. So, for example, some of the initial screens that we've done, these polymacrocyclic peptides have had single-digit nanomolar Kds. That's with no med cheming whatsoever, just what we pull from the screen. And, getting values like that on a first pass with zero med chem is absolutely wild.

Yeah, that's pretty impressive. So, what kind of molecules have you been able to identify so far? What kind of peptides have been found to be helpful with this kind of cyclization where now they have more therapeutic value?

So I think I can give 2 broad examples of some of the libraries that we work with. And I think that the simple examples will make the most sense. One type of peptide that we've been screening looks a bit like a barbell, where you have, say, a macrocycle here and a macrocycle here, and then they're connected to each other by an alanine chain. So this is the simplest form of how you could imagine a bispecific peptide, where one macrocycle will engage engage with one target, another macrocycle will engage with another. And this could be applied to a variety of different areas. But as I mentioned, having something like a bispecific peptide rather than a bispecific antibody, you could target 2 things simultaneously in cancer and have your payload that kills the cancer conjugated to that peptide. So this would give a lot more specificity towards making sure that the cancer treatment is As only deleterious for the cancers. That's part of it. And then what we're also trying to do is mimic natural protein folds that typically require hydrophobic collapse. And what I mean by that is when you have a, you know, your linear chain of amino acids, eventually in order to get the shape of your protein, that linear chain has to collapse, fold, and then enter into its final conformation where it's active. Trying to mimic, say, catalytic domains or specific architectures in the context of a small peptide is traditionally impossible because you can't get the right types of folding. But if you have a way of enzymatically stitching and editing the peptide together, then all of a sudden you can get these more complex nested motifs that look very similar to protein architectures, and we're also using those in our screens.

I see. Yeah, because that makes a lot of sense because many of the— well, I shouldn't say many of them. I can't really say in the universe of peptide shapes, but there are at least a number of domains that are relatively small and made up of a few amino acids, which articulate with other receptors or other surfaces that— or other proteins that are real significant.

Yeah, absolutely.

Yeah. So what are some of the other potential targets for this? You mentioned cancer as kind of what seems to be almost like an immunotherapy context in that you're exploiting the presence of specific surface antigens that are present on cancer cells that maybe are not present on normal cells. And are there other examples like this that are particularly attractive for your company?

Right now, if we cast the widest possible net, And maybe this is me thinking too much towards the future, but this really does seem to be target agnostic and indication agnostic, meaning that because peptides and proteins are found everywhere in nature and the body, that means that everywhere there is a peptide or a protein could be impacted by our platform. But some of the areas that we are thinking of right now are in the anti-inflammatory space. So there are a lot of targets that peptides could disrupt in order to tone down an inflammatory response. We're also considering anti-infective peptides. One of the main ways that microbes evolve resistance is just by tweaking one amino acid in whatever, say, method of resistance is necessary, and then all of a sudden all of your small molecule therapeutics no longer are effective. If you have a peptide that has a larger surface area and can interact on a bigger piece of that microbial machinery, then you can't just have one event to give resistance. It requires multiple events. And from an evolutionary standpoint, that's very difficult to do.

Yeah, you might actually even make it better.

Yes. Yes.

Yeah. So that was it. That was— sorry to interrupt you on that. It was just kind of picturing this in my head. You're exactly right, because now you're, you're not able to evade a mechanism by having a single mutation. Now it takes It would take a number of mutations and you could actually increase binding affinity. So that seemed rather interesting. Can we talk about business model a little bit?

Sure.

Is this something that you would team up with other companies that are currently designing, say, something like a GLP-1 agonist that would say, okay, this is good, but pharmacologically has a short half-life. Is there a way that we can tweak it using a Sathera approach? Or how does the business model work?

I think that there's 2 main ways that we're working. And the first is we've really looked at PeptiDream and Bicycle Therapeutics and how they built their peptide business. And really what they've done is they used the power of their screening platform to go after targets de novo. So someone at a large pharma company would say, this is a target of interest that we want to try to develop some new chemical material around. Can you screen this target for us using your platform. And both of those companies have been very successful in that space. And I mean, bluntly, I think that our platform is better because we can access more architectural space. And we've shown that our enzymes work with unnatural amino acids as well. So it's— there's a lot of space that we can go. So we do see ourselves as a screening partner and can help co-develop that particular chemical entity towards new targets. I mean, anything in the protein-protein interaction space, we believe that our peptides have great value in. And then the other areas, we are interested in our own targets internally. The reason why I went to the cancer space is because we want to go after these rare orphan indications, these rare cancers, and find ways to use what is a very powerful platform to get to results for those patients sooner. So I think those are the two main ways that we No, that's really good because the rare orphan cancer space is—

I hate to say it that way— is such an issue because so many people are still affected. Even if it's rare, it's still there. And folks just don't have the options that others do. And so something gives them a little bit of an optimistic look forward. Well, all of this is really optimistic and really adds a new level of understanding for me, at least in terms of how peptides peptides could be more effective therapeutics and used in novel ways, which is super cool. So, Dr. Karsten Eastman, thank you very much for joining me today. And let me know when there's a big breakthrough and let's talk again.

Excellent. Well, thank you so much, Kevin. It was great being here. And I really appreciate the opportunity to discuss the platform and Sathera with you. So thank you again.

Oh, very good. And for listeners, thank you very much for listening to another week of the Talking Biotech podcast. I hope you keep in mind the idea that peptides are biologically useful. They're all around us and within us and serve many important functions, and that their role as a therapeutic has been really held back by just the natural processes that destroy them. If we look at the diversity of amino acid-based compounds that we can make, it's almost infinite, especially when you incorporate non-natural amino acids. So the possibility to devise new therapeutics and as well as new other molecules that are useful in biology, antibiotics, everything is virtually limitless once we have a better command on ways to be able to manipulate peptides. This is a Talking Biotech podcast, and we'll talk to you again next week.