Ep203: Sri Kosuri on Small Molecule Correctors of Disease

Episode Summary

Executive Summary: Sri Kasuri, co-founder/CEO of Octant Bio, discusses building a platform for oral small-molecule correctors of protein misfolding and mistrafficking, with lead program OCT-980 for rhodopsin-associated autosomal dominant retinitis pigmentosa now in clinical testing. The conversation also covers his personal journey through his daughter’s cancer treatment, shaping his view of medicine, teamwork, and urgency.

Main Topics: Personal family cancer experience and its impact (Priority: 5/5): Kasuri recounts his daughter’s rare myeloid sarcoma diagnosis, the rapid St. Jude response, and her eventual cancer-free recovery, emphasizing how the experience deepened his appreciation for oncology and drug development. From computational biology to synthetic biology (Priority: 5/5): He traces his path from childhood computer fascination to Berkeley and MIT, where he moved from computational modeling into experimental synthetic biology, early gene-editing, and measurement technologies. Why start Octant Bio (Priority: 5/5): Kasuri explains leaving academia and co-founding Octant because many diseases would not be solved by gene therapy alone, and because emerging technologies plus new classes of small molecules suggested a better platform for drug discovery. Octant’s platform for correctors and multiplexed biology (Priority: 5/5): The company uses high-throughput cellular assays, deep mutational scanning, and iterative chemistry to identify and optimize small molecules that rescue protein function, especially for misfolding/mistrafficking diseases. Lead program: rhodopsin-associated autosomal dominant retinitis pigmentosa (Priority: 5/5): OCT-980 targets a large subset of rhodopsin mutations causing retinal protein misfolding. The goal is to improve low-light vision and ultimately halt degeneration, with clinical readouts based on imaging and functional vision tests. Expansion into other hard targets and AI/ML use (Priority: 4/5): Kasuri says the same platform can support programs in Fabry disease, P53-mutant cancer, and neurodegeneration, while AI/ML helps prioritize compounds, automate workflows, and scale program execution.

Key Arguments: Kasuri argues that many diseases are better addressed by small-molecule correctors than by gene therapy, especially where restoring native protein function can be disease-modifying. He contends that platform value comes from repeated experimental iteration: build thousands of analogs weekly, learn from cellular readouts, and converge on drug-like properties. He says rhodopsin ADRP is a strong first indication because patients are genetically identifiable, natural-history decline is measurable, and improvement—not just slowing decline—could shorten trials. He emphasizes that Octant’s assays are designed in human cells and anchored to human genetics, making them more predictive than relying on animal models alone. He believes AI/ML is most useful when paired with proprietary data, helping scientists choose the next best compound and eventually run more of the workflow agentically. He argues that undruggable targets like p53 may become tractable through new chemistry classes such as glue correctors, but only if the platform can continually optimize local chemical space.

Data Points: Daughter’s diagnosis timing: Thanksgiving 2024 - Kasuri says swelling above her eye began around Thanksgiving and led to a prolonged diagnostic workup. Time to St. Jude treatment: About 1 week from first contact to being in Memphis - He describes a rapid sequence: contact on Saturday, evaluation Sunday/Monday, tickets Wednesday, Memphis Friday. Treatment duration: About 9 months - His daughter participated in an AML23-style regimen over roughly nine months. Patient population for rhodopsin ADRP: About 20,000 patients in the US and EU - Kasuri cites a sizable but rare population that is genetically well identified. Mutation coverage of OCT-980: About 75% - He says the program works on roughly three-quarters of mutations causing rhodopsin misfolding/mistrafficking. Compounds generated in rhodopsin hit-to-lead: About 250,000 analogs - He describes weekly iterative chemistry over the course of a year. Weekly chemistry batch size: About 5,000 compounds per week - Octant builds and tests about 5,000 compounds weekly in the optimization cycle. P53 program compounds tested: About 750,000 compounds to date - He says two people ran the p53 program over the last couple of years on the platform. Natural history study size: 200 rhodopsin ADRP patients - A Moorfields Eye Hospital study tracked patients longitudinally over eight years. Natural history duration: 8 years - The study followed patients over a long timeline to quantify decline. First clinical readout timing: Q3 2027 - Kasuri expects the multi-dose patient study to yield meaningful data then. Human rod outer segment turnover: 2 to 3 weeks - He uses retinal biology turnover to estimate how quickly a functional effect might appear.

Pivotal Quotes: "Every personal family's emergency is their emergency." — Sri Kasuri: He describes the mindset and dedication of pediatric oncology teams during his daughter’s treatment. "If we can't simulate that, what can we possibly hope to simulate?" — Sri Kasuri: He explains why he pursued synthetic biology around the relatively simple T7 virus during graduate school. "You don't have a CRO partnership, you have a lottery ticket." — Sri Kasuri: Used during a sponsor segment-style discussion about the importance of standardized bioanalysis; reflects his preference for reproducible workflows.

Implications: Octant is betting that data-rich, cell-based corrector discovery can produce broadly accessible oral medicines for misfolding diseases. If OCT-980 works, it could validate a repeatable platform for ophthalmology, rare disease, and even oncology.