Sail Biomedicines’ Kerry Benenato and Rajesh Ramaswamy discuss the roles artificial intelligence (AI) can play in informing programmability, using the example of Sail’s tailored Endless RNATM (eRNATM) medicines and addressable nanoparticle delivery vehicles.
Maks_lab Shutterstock
Human health and disease are complex, and the limited success of one-size-fits-all medicines has led to increased interest in personalised medicine. Tailoring medicines for small populations of patients is expensive, however, and yet, recent experience with COVID-19 vaccines has shown that adapting to changing circumstances is commercially viable with medicines that are readily programmable.
How did witnessing the rapid development of COVID-19 vaccines shape the vision for programmable medicine?
KB: Where the COVID-19 vaccines served as a particularly successful proof-of-concept for the application of messenger RNA (mRNA) as a therapeutic molecule, perhaps the aspect of these vaccines that is less well understood publicly and yet is equally as important is the proof that the mRNA molecules were programmable. As the coronavirus mutated over time and new strains became more prevalent in the population, we’ve been able to rapidly evolve the sequences of the vaccine mRNAs to meet those new challenges, something that is much more difficult and expensive with protein-based therapies or even small molecules.
Can you explain how Sail's eRNA differs from traditional mRNA-based therapies?
KB: Because mRNA is rapidly degraded when it is introduced to the body, the therapeutic window for mRNA-based medicines is typically brief. In the target cells, there is a rapid burst of protein expression followed by a steep drop-off. Thus, sustained therapy requires frequent dosing to maintain efficacy, which can lead to tolerance issues and immune responses in patients.
As a fully closed system, circular RNA like Sail’s Endless RNA or eRNA has a longer half-life in the body than linear mRNA and so it can be dosed less frequently than the equivalent mRNA. Natural circular RNAs don’t express proteins, however, so the first thing Sail needed to do was produce a circular RNA molecule that can express the proteins of interest. Experiments performed to date in cell culture and in animal models show that protein expression from eRNA is significantly more durable than from mRNA. In one experiment, for example, levels of a growth hormone were 1000-fold higher after 72 hours in mice when expressed by eRNA versus mRNA.
How does Sail leverage natural nanoparticle chemistry to improve the precision and efficacy of programmable medicines?
KB: Another challenge with RNA-based medicines is delivery of the therapeutic molecule to target cells and tissues. Although great advances have been made using lipid nanoparticles (LNPs) and many organisations are adept at producing them, LNPs typically migrate to the liver. While that is beneficial for some medical conditions, such as liver disease, it is of little use for conditions that impact extra-hepatic systems.
RR: At Sail, we view delivery not just as a logistics problem, but as an engineering challenge that demands precision. Nature has already evolved countless solutions for moving materials across biological barriers safely and efficiently. We leverage the chemical matter used by nature and use AI to train machines to learn the rules that govern tissue-specific delivery. By doing so, we’re building a programmable delivery platform—one that evolves beyond generic lipid nanoparticles toward targeted, tolerable systems optimised for each RNA medicine. It's an approach that fuses nature’s intelligence with machine intelligence to unlock delivery as a core enabler of programmable therapeutics.
Are there limitations to the types of proteins that can be programmed using eRNA, and how are these challenges addressed?
KB: To date, we have not yet run into any limitations in the type or size of proteins that can be produced by eRNA. We have introduced intracellular, secreted, and transmembrane proteins. As an example, in one of our partnered programs, we are able to express the very large cystic fibrosis transmembrane conductance receptor (CFTR), which is completely missing in a small percentage of people with cystic fibrosis.
How could eRNA medicines address unmet needs in rare diseases or complex chronic conditions?
RR: A big challenge in the rare disease space is the small patient population, which makes efforts to develop medicines to treat the disease less economically viable. One way to deal with that is to reduce the cost of drug development, and we think our AI engine gives us an advantage here. The better we learn and leverage the language of life, the higher probability that an eRNA medicine we develop will work and the faster we can get them into clinical trials and, eventually, to market.
KB: Sail’s approach allows us to fine-tune the candidate eRNA to the problem at hand and to the patient we want to treat. And our ability to rapidly produce new candidates allows us to go from initial idea to dosing an animal model within weeks to test our hypotheses.
What role do you see programmable medicine playing in the future of personalised medicine?
KB: For personalised medicine to be fully realised, programmable medicines are key. Human health and disease are complex, and as we have seen in many cases, a single therapy to treat them all has its limitations. Programmable medicines give us the ability to tailor a treatment’s efficacy to a given patient or patient subpopulation.
RR: When you change the cost aspects of the equation, more things are possible on the outcome front because now you can cater to a smaller pool of patients. Can we realise a world with RNA medicines where we are talking about the extreme of personalised medicine: the individual? I am not sure whether at that level, it would be economically viable. If you really master the rules of life, however, with the help of AI, you should be able to engineer every element of programmable medicines to suit narrower pools of patients.