Podcast Transcript

David Brühlmann [00:00:41]:
Every drug substance is supposed to be one molecule, one peak, and one clean story. Matthew O’Connor’s therapeutic isn’t. As CEO of Scientific Affairs at Cyclarity Therapeutics, he has built a cyclodextrin dimer that goes after 7-ketocholesterol, a molecule linked to artery plaque, increasing the risk of heart attacks and strokes. And that molecule doesn’t fit neatly as a small molecule nor a biologic. Today, he walks us through the science, the engineering, and the CMC puzzle of developing something that refuses to sit in an existing box. Matthew, welcome. It’s good to have you on today.

Matthew O’Connor [00:02:41]:
Thanks so much, David. It’s wonderful to be here.

David Brühlmann: [00:02:51]:
To start us off, Matthew, share something that you believe about drug development that most people disagree with.

Matthew O’Connor: [00:03:00]:
Well, something that most people disagree with is that the development of the drug side itself is boring. I think that’s what most people think, and that’s something that I disagree with. So I think the development side is exciting and important, and it doesn’t get enough attention from people both inside and outside of the industry.

David Brühlmann [00:03:13]:
Very good point. I like that. Obviously, I’m biased, but I definitely agree with you that it is very interesting. Before we dive deeper into this very topic, into drug development, I would like to start out with your story, Matthew. Tell us what drew you first into science and what were some interesting pit stops along your career?

Matthew O’Connor [00:03:39]:
When I was 16, I started thinking and started reading about the biology of aging and thinking about what are the key challenges facing humanity in the world? And I thought that those were, in terms of human health, revolved around most of what happens to us as we age. As I was sort of coming out of the era of people dying from things like infectious disease, childbirth, and violence. So that was a former era, and now we’ve moved into the era of chronic disease. And when you get down into the details, I think you find that what drives chronic disease is the aging process itself. And so I just started going further and further down that rabbit hole and realizing that this was one of those fundamental limitations for humanity that needed to be addressed. And I decided to dedicate my life to it.

David Brühlmann [00:04:34]:
Excellent. And what is the problem now you are solving with your company?

Matthew O’Connor [00:04:41]:
So dovetailing on what I was just talking about, about fundamental mechanisms of aging, I eventually, after some decades of study, came to the conclusion that the best way to think about it, not necessarily the only way to think about it, but that I think the most practical way to think about how to fix our systems as we age is to divide and conquer and look at what is breaking, what aspects of our biological processes as we age are breaking. And so this divide-and-conquer approach is one that I espoused when I was working at SENS Research Foundation, an aging research foundation.

I was managing many projects, but one project that fascinated me is one in which you have accumulation inside of your cells of various kinds of molecular junk. One of those fundamental products is a byproduct of oxidized cholesterol, and that accumulates in our cells and tissues as we age. And so it turns out that that is a key molecule that accumulates in our blood vessels in the process of plaque development, of atherosclerosis. So that one target is our drug target, and we invented our drug, which we’re currently just dubbing UDP-003, to scavenge it.

David Brühlmann [00:06:01]:
So if I may characterize, you have now this new drug that’s going to reduce the deposit of that oxidized cholesterol in our blood vessels. And so how does that work? What is the drug? What are the mechanisms of action? And how does it finally solve this very problem?

Matthew O’Connor [00:06:22]:
Yes, so we investigated various approaches to try to target or get rid of or degrade the oxidized cholesterol, and I settled on a technology called a cyclodextrin, which is a macrocycle. It’s a cyclic carbohydrate, and it binds to the oxidized cholesterol, and it seems to be able to pull it out of cells and tissues and then float away with it and be excreted through your urine. So it’s a quite simple mechanism of action. There’s no complicated metabolic steps. There’s nothing in this process in which either the drug itself or the drug target is metabolized or changed in any way before it’s excreted from the body. So it’s a very simple sequestration and excretion mechanism of action.

David Brühlmann [00:07:14]:
Oh, okay. And because of that, I imagine you don’t expect a lot of side effects, right?

Matthew O’Connor [00:07:20]:
That’s exactly right. And that’s how we designed the drug. And I think the beauty of it is its simplicity. And the fact that if you imagine a perfect version of this drug, if it was perfectly specific for its target. Now, one thing that I didn’t exactly mention is that this oxidized form of cholesterol is not created intentionally by your body, by your metabolism, by your biochemistry, and so it’s not useful in any way. You don’t use it for energy, you don’t use it like you use regular cholesterol to help make up your cell walls or anything like that. And so if I could wave a magic wand and make all of the particular isoform of cholesterol that we’re targeting is 7-ketocholesterol, so you don’t have any enzymes that create it, and so if I could wave a magic wand and just make all of it disappear from your body, you’d just be better off.

And so if our drug was infinitely specific for 7-ketocholesterol, then you would indeed have zero side effects with our drug because it would only bind the target, it would pull it out, it would float away, you would urinate all of it out, and you’d just be better off. So that’s kind of our starting point, our ideal creation. And so we set about creating something that had as much affinity and specificity for this type of cholesterol as we could make. And we came up with something that was pretty good in this regard and has a lot of affinity and specificity for its target. And so it does its job very well and indeed has very few side effects or toxicities.

David Brühlmann [00:08:57]:
What kind of work went into designing this, or identifying this very cyclodextrin that had this high specificity? I imagine that’s not just a simple screening exercise. I guess a lot more thought went into that.

Matthew O’Connor [00:09:13]:
Yeah, kind of three steps. The first being just the ideation of figuring out that cyclodextrins were a good potential material for this, which on one hand is kind of weird because cyclodextrins have been in medicine for about 25 years, depending on how you count it. But there’s only one cyclodextrin-based drug that’s approved yet. While not an unheard-of type of material, it is unusual.

And so when I was digging through the literature trying to think of alternative approaches to this, I came across cyclodextrins as sequestration molecules. And in fact, in the 1980s and ’90s, all of the existing sort of catalog cyclodextrins had been studied pretty extensively for their cholesterol-binding properties, and even to some extent their cholesterol derivative binding properties, like the oxidized cholesterol versions. I was able to learn a lot about what makes a cyclodextrin safe, and alternatively, how can you make a cyclodextrin dangerous, and also how to make a cyclodextrin bind cholesterol or a different form of cholesterol.

And so we had some ideas about how to do that before we went into in silico and started modeling these. And that harkens back to 2017, back at SENS Foundation, when I had an intern by the name of Amelia Anderson, and she had the idea to start doing computational chemistry and modeling these. And that wasn’t at all inside of my experience. And so I said, what do you need? And she said, I need $100 for a software license. And I said, go for it. And so she spent her summer playing around with modeling these and showing me different models and videos that she was putting together of these.

That led us to the idea, together with insights from the literature, that what we really wanted wasn’t just a single monomer of cyclodextrin, that we actually wanted 2 molecules of cyclodextrin to wrap around and encapsulate a single molecule of the target, oxidized cholesterol. And that’s what gave birth to this idea that what would happen if we dimerized these? And of course, the simulations made them look really good. And that gave birth to the 3rd phase of this project, which was doing some synthetic chemistry and making the first prototypes, cyclodextrin dimers, in which we dimerized them in a particular configuration.

So cyclodextrins are asymmetric molecules, and so we dimerized them in a head-to-head configuration, which is what all of our pre-chemistry data was proposing, and we made these head-to-head beta-cyclodextrin dimers, which turned out to have phenomenal affinity. So we didn’t just double by putting 2 molecules together, we didn’t just double the affinity for the target, we— took us years to actually measure this, but we ended up increasing it by something like 1,000-fold.

David Brühlmann [00:12:02]:
What is the impact on the drug itself since it’s a dimer? Does it behave like a traditional small molecule? Is it behaving more like a large molecule? Is it something else?

Matthew O’Connor [00:12:15]:
It’s kind of something in between. And I almost think that we need some different language to describe molecules like this because it’s technically a large molecule because it’s more than 1,000 daltons. That’s the definition of a small molecule, is less than 1,000 daltons. Well, our molecule is more than 2,000 daltons, but less than 3,000 daltons. So like that’s bigger than a traditional small molecule and bigger than the definition of a small molecule, but a lot smaller than what we tend to think of as a large molecule, like an antibody or a therapeutic enzyme, for example. And so it sort of falls in this zone of what I sometimes call a medium-sized molecule.

We are being regulated by FDA as a small molecule. So there’s that. Which was a big question for me when we were starting this process. Are they going to send us to CDER or CBER? I didn’t know for certain. We guessed that we would get sent to CDER, and we did, but we didn’t know for certain going into it. But the behavior, to your point, is, you know, something a little in between. We’re kind of in between a small molecule drug and a therapeutic antibody. We’re capturing something, which is kind of what therapeutic antibodies can do.

You could imagine a therapeutic antibody doing something like this. It turns out that we have investigated the idea and the potential of designing antibodies that are specific for the oxidized cholesterol isoform that we’re targeting, but it turns out to be very difficult. We’ve been completely unsuccessful at making something that had good specificity for this, but it’s capturing a molecule the way that a therapeutic antibody can scavenge a target. But we’re building it and engineering it kind of more like you would do a traditional small molecule starting material and doing chemical reactions one by one to add the different chemical groups to it that you need to dimerize and do the chemical substitutions to create the final product that has the high specificity and affinity for the target.

David Brühlmann [00:14:11]:
Is this the same pathway the small peptides undergo? Also, the small molecule pathways is the same? I believe it’s below 40 amino acids that there’s also a cutoff, even if it’s produced biologically. There’s also some strange legacy thinking of the pathway that even if you produce a molecule recombinantly, if it’s under a certain number of amino acids, it’s also considered a small molecule.

Matthew O’Connor [00:14:35]:
I’m not familiar with the regulatory path on those, so I’m not certain if there’s a specific set of definitions that you have to fulfill that is similar to that. I can see the logic and the point, so it may be. And there is a logical similarity to what you’re talking about because the starting material, as you’re alluding to, is in fact a biological product. So the basic cyclodextrin starting materials are produced by enzymes which were originally produced by microbes. And so our starting material is just a completely— every beta-cyclodextrin molecule is identical to every other beta-cyclodextrin in the world, but they’re originally produced by microbes and now produced in giant vats by enzymes that produce these and then can be purified. And that’s the starting material for our drug. So yes, it’s similar to what you’re talking about.

David Brühlmann [00:15:30]:
Speaking of the enzymes, there’s also a lot of work that goes into finding the right enzymes, right, that enable you to produce those cyclodextrins.

Matthew O’Connor [00:15:37]:
You know, that’s really not our problem. This is a problem that was solved, I don’t know how many decades ago, way more than the sort of three decades of like the modern cyclodextrin era. Cyclodextrins have been mass-produced and purified and used for many, many purposes, for something like 80 years, and these processes have been solved even at an industrial level many, many decades ago.

People are surprised that this kind of chemistry has like a long, long, and deep history. Cyclodextrins have been used in various types of solvents. They’re used in commercial products. For example, Febreze has cyclodextrins in it. Cyclodextrins are used in food products primarily as fiber additive, and that’s been around for a long time and predates their uses in medicine. So the industrial and food uses of cyclodextrins predate their uses in medicine by a long time. And so that means that these basic building blocks that you’re talking about can be made extremely efficiently and extremely cheaply and are basically just not our problem. You can buy very high quality, high purity, and high quantity cyclodextrin starting material for very, very cheap, like dollars per ton at scale. So that’s not a significant part of the problem that we needed to solve.

David Brühlmann [00:16:55]:
And how did this inform your manufacturing strategy? Do you do this in-house? Are there a lot of CDMOs that can do that very easily?

Matthew O’Connor [00:17:03]:
Well, none of it is done in-house, but there are not a lot of CDMOs that can do it very easily, despite the number of emails that I receive every single day from— I don’t even know how so many CDMOs exist in the world, but they email me every day saying, we can make your product and we’d be really excited to do so. And they really don’t know what they’d be getting themselves into. It’s not trivial and it’s not easy. And it’s kind of a whole sub-area of chemistry that you have to learn. And so we’ve worked with different groups around the world at different times over the last decade, groups that specialize in cyclodextrins, ones that specialize in carbohydrate chemistry, and then we’ve taught different CDMOs, taken our process and carefully taught them.

And so we have an internal head of chemistry named Keivan Sadrerafi, who is a specialist cyclodextrin chemist. And anytime he’s talking to somebody new, you know, they come out and they’re like, oh, we’re just going to make all these changes to your process and it’s going to make it more efficient, stuff like that. And he’s like, okay, I’m really excited to hear about what kinds of things we’re going to do, but let’s take a breath here and make sure you understand what we’re trying to do and why some of your ideas might not just work on the first try because these things don’t always work exactly the same way that a simple small molecule is going to work for you.

David Brühlmann [00:18:25]:
Speaking of challenges, I’d like to talk about your drug substance, because if I’m correct, it’s an isomeric mixture rather than a single defined molecule. So how do you demonstrate identity or the batch-to-batch consistency with such a complex mixture?

Matthew O’Connor [00:18:42]:
The original starting material that we were talking about— so there’s three flavors of cyclodextrin, 3 sizes: alpha, beta, and gamma. We’re working with beta, which is the medium-sized ones. Those are all single isomers. You could make a single isomer of a cyclodextrin. So a beta-cyclodextrin molecule has 21 hydroxyl groups around the outside of it, which are perfect places to do substitution chemistry on. And that gives you a lot of variety, but also a lot of complexity, which is both a feature and a bug. And there’s aspects of it that are perfectly symmetrical and aspects that are asymmetrical. And so you can generate an almost infinite amount of complexity or probably in fact an infinite amount of complexity if you include the infinite amount of complexity of potential chemistry that you could just keep adding on to these things.

But if you just want to start with a single chemical group and you react it with one of these hydroxyl groups, well, that’s a partially random reaction or a stochastic reaction. And so you take one molecule, let’s say a methyl group, the simplest group that you might want to add or substitute onto one of these hydroxyl groups, where it reacts is going to be stochastic and semi-random. And then you add a second one, and now you’re going to be putting that on in a semi-random place. So you put it on in one site, we’ll call it site 1 out of 21, and then the next site is probably not going to be right next to it, right? And it’s probably going to be a little bit far away from it, but it’s going to be somewhat predictable. It’s probably going to be on the opposite side of a molecule in a semi-predictable place.

And then the third one and fourth one, and you can go on from there. So if you fully substitute it, now you have a single isomer that has 21 methyl groups that have wrapped around your molecule. But if you partially substitute it, you get something that is going to be somewhat predictable in terms of the number of reactions and the placement of the reactions, but not perfectly. So now you have something that you can do different types of measurements on. So you can do different types of mass spec, you can measure the molecular weight. And so you’ll say, okay, this one has 10 methyl groups on it.

And then you start saying, well, where are those methyl groups? And that’s where it starts getting a little tricky. And so you can do NMR, you can do 2D NMR, and you can start mapping those out. And then using different types of mass spec and NMR and 2D NMR, you can start getting a sense for where are these. And once you get to a certain level of complexity, you’re not going to define every single isomer there, but you’re going to see a pattern on your spectra. And then you make another batch and you see the same pattern, and you say, it looks like we’ve got a very similar distribution of isomers here. And you can create your documentation and show it to regulators and say, we believe that we’ve produced 2 batches, they’re very similar to each other. And of course, you pile on top of that all the data that you’ve done in vitro, in cell culture, in animals, and you say, this is safe and it has the same activity as the previous batches. And eventually you say, you should let us use this in people.

David Brühlmann [00:21:37]:
And how will this be administered to people? Is that a liquid formulation? Is that lyophilized? How does that work?

Matthew O’Connor [00:21:45]:
We envision this being administered to people as a liquid formulation and as an injectable. Cyclodextrins are notoriously poorly orally absorbed. I mentioned cyclodextrins being fiber supplements before, and it turns out that just about every cyclodextrin that you invent, pretty much no matter what you do to it, is not going to be orally absorbed into the bloodstream. It’s just going to be passed through. So you do need to find another way to get it in, and so it’ll be an injectable. We envision it being a subcutaneous injectable like a GLP-1 drug, in that it would be delivered at home through one of these self-administered pens, but we could also do it with like a needle patch or something like that, an osmotic pump, all stuff that we’re interested in and looking at, but it does have to be an injectable.

The first cyclodextrin drug was called sugammadex and is now marketed as Bridion. The first version of that was a lyophilized powder that was mixed up by the pharmacist or by the nurse or the anesthesiologist because that’s an anesthesiology product, and that was powder in a vial, but it got quickly changed. When Merck acquired it, they created a liquid formulation, and we envisioned the same thing happening with our drug.

David Brühlmann [00:22:55]:
Matthew O’Connor’s story is a reminder that the toughest CMC problems often start with a simple question: what box does this actually belong in? We’re just getting into the plant-floor realities and the regulatory conversations that followed. If you’re enjoying this conversation with Matthew O’Connor, please leave a review on Apple Podcasts or your favorite platform. Thank you so much for tuning in today, and I’ll see you in part 2.