Ivan Kairatov is a titan in the biopharmaceutical sector, recognized for his pioneer work in integrating generative artificial intelligence into the traditional drug discovery pipeline. With a career spanning decades in research and development, Kairatov has been a vocal advocate for moving beyond the “trial and error” methods of the past. As the industry watches Insilico Medicine advance its AI-designed drug for Parkinson’s disease into human trials, Kairatov provides a seasoned perspective on how technology is finally cracking the code of the blood-brain barrier and the chronic neuroinflammation that has long stymied CNS research.
The following discussion explores the revolutionary shift from multi-year discovery timelines to rapid preclinical nominations, the technical hurdles of treating the central nervous system, and the strategic importance of global collaborations. We delve into the specifics of the ISM8969 clinical trial design, the biological significance of the NLRP3 inflammasome, and the emerging concept of “pharma superintelligence” through advanced benchmarking platforms like MMAI Gym.
Traditional drug discovery often spans several years and requires testing thousands of compounds. How does reducing this timeline through AI-driven platforms like Chemistry42 fundamentally change the landscape for neurodegenerative research?
The shift we are seeing is nothing short of a total transformation in how we approach human health. Historically, the early-stage discovery process is a grueling marathon that takes anywhere from 2.5 to 4 years, often leading into dead ends after millions of dollars are spent. Insilico has effectively shattered that paradigm by reaching the preclinical candidate nomination stage in an average of just 12 to 18 months. Instead of the scattergun approach of testing thousands of molecules, the Chemistry42 platform allows researchers to synthesize and test only 60 to 200 molecules per program. This level of precision is a massive relief for scientists who have spent years chasing “druggable” targets; it feels like finally having a high-resolution map in a territory where we used to be wandering in the dark. For neurodegenerative diseases like Parkinson’s, where time is literally brain tissue, cutting years off the development cycle means we can reach patients while they still have a significant quality of life to preserve.
The blood-brain barrier has historically been a graveyard for many promising neurological treatments. What makes the design of ISM8969 a potential “best-in-class” candidate for reaching the central nervous system?
For years, the industry has struggled with the fact that most effective NLRP3 inhibitors are peripherally restricted, meaning they can treat inflammation in the body but are blocked by the brain’s protective barrier. This is a tragedy because the NLRP3 inflammasome is a vital component of the innate immune response, but when it becomes chronically activated, it triggers an overproduction of pro-inflammatory cytokines that drive neurodegeneration. ISM8969 was specifically engineered to be brain-penetrant and orally available, a dual-threat profile that is incredibly difficult to achieve through traditional chemistry. In our preclinical models, we saw a balanced druggability profile that demonstrated favorable permeability and safety across multiple mouse disease models. By precisely optimizing this molecule through AI, we have created a tool that can finally access the target compartment in the brain to modulate pathological inflammation and support neuronal survival directly.
Can you elaborate on the strategic design of the Phase I trial in Australia and why the inclusion of obese participants is a critical component of this study?
The Phase I trial is a meticulously structured, randomized, double-blind, placebo-controlled study that will provide the first hard evidence of how this molecule behaves in the human body. We are conducting both single ascending dose and multiple ascending dose cohorts, involving 80 healthy participants and a very specific subset of 20 obese adult participants at risk of cardiovascular disease. This is a brilliant strategic move because obesity and cardiovascular risk are often linked to systemic inflammation, providing a more complex physiological environment to test the drug’s safety and pharmacokinetics. We will be collecting cerebrospinal fluid samples to verify that the drug is indeed crossing into the central nervous system at therapeutic levels. Seeing those first sets of data will be an emotional milestone for the team, as it moves the project from a digital concept to a tangible medical reality that could help millions.
With 31 preclinical candidates nominated and 13 receiving IND approval since 2021, how is the industry responding to the emergence of “pharma superintelligence” and platforms like MMAI Gym?
The industry is reaching a tipping point where the results are becoming too significant to ignore, as evidenced by the 13 programs that have already cleared the path for human trials. We are moving toward a state of “pharma superintelligence,” where AI doesn’t just assist researchers but serves as a “trainer and benchmark” for scientific reasoning through platforms like MMAI Gym. This system integrates thousands of benchmarks and real-world datasets, allowing organizations to train models that can handle domain-specific tasks with a level of rigor that humans simply cannot match alone. Collaborations with pioneers like Liquid AI and Human Longevity show that there is a growing ecosystem of companies eager to utilize these datasets to refine their own discovery engines. This isn’t just about speed anymore; it’s about a fundamental increase in the quality and success rate of the molecules we choose to put into people.
How does the co-development partnership with Hygtia Therapeutics reflect the changing financial and strategic landscape of biotechnology?
The collaboration between Insilico and Hygtia is a model for the future, where risk and reward are shared equally through a 50% stake in global rights. This strategic alliance, supported by the Shenzhen Pengfu Fund and Fosun Pharma, allows for an accelerated global development path that wouldn’t be possible in a silo. Insilico leads the IND submission and the initial Phase I trials, but they remain eligible for up to $66 million in upfront and milestone payments as the program matures. This structure provides the financial stability needed to push multiple candidates forward simultaneously while leveraging the clinical expertise of established pharmaceutical giants. It’s a sophisticated dance of innovation and infrastructure that ensures the most promising AI-generated candidates don’t just sit on a shelf but actually make it to the clinic.
What is your forecast for the future of AI-designed neurology treatments?
I forecast that within the next five to seven years, the use of generative AI in designing CNS-penetrant molecules will shift from being an innovative experimental approach to the industry’s primary standard. We will see a massive influx of successful Phase II and III trials for neurodegenerative conditions that were once considered “undruggable” because we can now optimize for multiple complex parameters—like blood-brain barrier penetration and toxicity—simultaneously. The successful first-in-human dosing of ISM8969 is the first domino in a long line of breakthroughs that will eventually make Parkinson’s a manageable chronic condition rather than a death sentence. As these AI platforms continue to learn from the data generated in trials like the one in Australia, the accuracy of our preclinical nominations will only improve, leading to a golden age of neuroscience where the time from target identification to patient treatment is measured in months, not decades.
