In Developing Oral Treatments, Complexity is the New Normal
Kurt Nielsen, Ph.D., talks with Dan Smithey, Ph.D., CEO of Serán Bioscience, about oral drug innovation, large language models, complexity, and how flexibility can create new pathways to reach market quickly.
Twenty years ago, when small molecules ruled the world, they were relatively simple in that they were low molecular weight, relatively soluble, and discovery teams had a narrow focus on those molecules with a complexity that they were willing to progress. Do you remember Lipinski’s Rule of Five that we all tried to adhere to for years and years?
As biology has advanced, we have discovered new targets that we think will impact human disease, the requirements for molecular structures, and whether they're small molecules or other modalities. Biologics have increasingly become more and more complex, not just bigger molecules, but we’re seeing much more structural diversity, the ability to not only interact non covalently, but covalently and with allosteric interactions instead of specific interactions that are known within pockets of proteins to have certain types of function.
That complexity is forcing the whole industry to rethink how we develop drugs, whether it's the screening of biology in discovery teams to where Serán gets involved, which is trying to understand how to formulate, develop, and ultimately manufacture all the way through clinical trials. Since we started Serán about nine years ago, we've seen a dramatic increase in complexity and diversity.
We're finding that we don't have the perfect tools anymore, if they ever existed, but we certainly do not have the tools now necessary to deliver these molecules, whether it is formulation or even manufacturing. So, I think it is very important that as an industry we respond to this challenge because as CDMOs, if we're not providing the best possible solutions to our customers, then we’re not doing your job. It requires a much deeper scientific approach to CDMO services, so it's an exciting time to take on those new challenges.
Without going too far into the details of the chemistry, diversity is everything from molecular weight, the types of interactions that are needed to specifically bond to essentially other proteins or proteins in general to invoke some sort of response, but it's also in such as peptides. We're seeing peptides, obviously with GLP-1 focus becoming more prevalent, and oral delivery of peptides is a big deal. However, the molecular weight of a GLP-1 is obviously dramatically different than that of a traditional small molecule, and they are also very different structurally and compositionally.
It’s those sorts of challenges, albeit GLP-1’s are at the extreme end of the range, that are driving diversity. Both a peptide and a traditional small molecule however require delivery and formulation approaches that ultimately give us the appropriate pharmacokinetic (PK) and pharmacodynamic (PD) response.
Stability is a big deal for a peptide modality too, and the gastrointestinal (GI) tract poses all kinds of issues when you're trying to deliver something orally, as it must get across the lipid bilayer within the GI tract and then into the bloodstream etc. So, it is not only the complexities bought on by the diversity of molecular structure we have to tackle, but also the diversity of challenges we need to address to get these molecules into the body with the proper PK. It’s very different than it was years ago.
One of the things we're also seeing, which is becoming a bigger and bigger deal, is insoluble molecules. We thought we had solved that problem when it was ‘theme of the day’ over the last perhaps two decades. Industry has embraced that we can handle poorly soluble molecules, because we need them, we need their hydrophobic interactions to invoke a response, and so we came up with some tools to do that. Amorphous systems are one such solution, and there are other approaches. But as these molecules have grown in both size and complexity, now we have another problem to solve, which is permeability. We can raise the solubility to get more drug absorbed within the GI tract, but permeability is a barrier that's much more difficult to overcome and there are fewer adjustments that can be made, or ‘dials to turn’, as there's more risk to doing things that aren’t good within the GI tract.
That's a challenge that I think is still fundamentally unmet, although there are new tools that we and others are exploring. That's something that we'd probably need to spend more time understanding.
Artificial Intelligence (AI) in Drug Development
This is the very beginning of a revolution in how we manage large volumes of data. Large language models, which is what we're mostly talking about when we refer to ‘AI’, are great at handling massive repositories of data and trying to come up with a most probable result from that analysis based on some sort of input. There's no doubt that that is important in our industry as well, whether it's trying to look at clinical data or manufacturing data and trying to make better decisions. I am very interested in understanding how those tools can help us.
Also, there's all the fantastic protein folding models that are out there. That’s a different kind of AI and a different sort of machine learning, or understanding, but they're all based on the same sort of algorithmic approach.
But, given the complexity we’re facing, we need tools like that to overcome massively complex systems such as the human body and understand how to deal with the challenges that poses in delivering drugs. These tools will become increasingly important, there's no doubt about it.
As an industry, we need to continue to think and address those issues.
The Complexities of Molecules, Drug Delivery, and Humans!
Looking at drug development for a moment, we must think about the complexity of all the different steps, the stakeholders in that process, and the challenges of dealing with multiple interfaces between groups of people and companies. We know that systems often fail at the point of interface, so an integrated approach to the steps in the drug development process would save time and money. Outsourcing is here to stay, but how could we move to a more integrated approach to drug development today? How do you see that evolving and what do you think it takes to make it successful?
I think there's no doubt that a large amount of experience makes an amazing impact on efficiency by not repeating old mistakes, and by being able to apply proven ideas to new things.
Every organization needs a mix of ‘gray hairs’ and ‘smart young people’ to solve difficult problems, and it’s super important that we don't forget about the past when we're trying to tackle new problems. The integrated approach is so important because not only do we have more complexity, but we are also challenged to move more quickly as an industry. If we think about developing a drug that takes 10 to 12 years and a billion dollars, we know that's not sustainable, and sometimes I’m amazed it ever was.
At Serán, we really tried to think about ways to shorten that whole development paradigm. We don't do clinical trial work, but as a chemistry, manufacturing and controls (CMC) related company, we see real benefits in creating an approach to commercialization early. In phase one alone, having an approach that we can have a lot of confidence in, whether it's the analytical component, the formulation, or the manufacturing elements (that's very easy to scale and to move forward quickly). The old paradigm of fail fast is no longer as applicable as it used to be, we really want to be able to succeed first, and you can do that by using the more advanced tools that are available now, whether they're material science or other experience-based ideas that allow us to progress more quickly, and with a lot less risk.
Given the development costs, the only way we're going to make an ROI work better and provide a better outcome for patients, is to go faster to commercialization. That's where tools like AI can really make a big difference, by helping us to understand the vast amounts of data we've collected previously and apply learning to the next problem. But we also must be smart about how we develop things.
Adding scale and remaining nimble and flexible means being able to react to data in real time and make decisions quickly so you can move to the next step. In early phases, flexibility means that when a client’s API gets delayed in customs, an organization such as Serán doesn’t reallocate the clinical slot but instead, we do all we can to keep on schedule. We're aiming for a business design that allows clients to manage delays and when the API shows up, the customer can still receive their material for their first in human study.
Then, flexibility downstream means that as market conditions change, your manufacturing changes too, to smaller batches and shorter changeover times. That’s how we handle operational complexity.
When we speak about complexity, it's not just complexity of molecules, it's complexity of clinical trials and of the market dynamics, which all require flexibility.