All-In Summit: Gene therapy and a new era of medicine with Dr. Nicole Paulk

Paulk’s central bet is to turn gene therapy from a bespoke rare-disease product into a reusable cancer platform. She spun UCSF work into Siren Biotechnology to combine AAV delivery with immunotherapy, using viral payloads that alert the immune system to tumors rather than building a separate vector for every disease.

The economic case for that platform is stark: Paulk said conventional viral gene therapies typically take 10–15 years and $2–3 billion to develop for one indication. Many target only 100–200 patients, so a reusable therapy across cancers would attack both the cost structure and tiny-market problem.

Siren’s strongest evidence in the talk was preclinical, not human: its virus eliminated implanted human brain tumors in treated mice while controls developed massive tumors and died. Paulk said survival improved enough to regard the mice as effectively cured, with a first clinical trial then being prepared for 2025.

The clearest proof that viral gene therapy can transform a human life came from an inherited-blindness patient, not Siren’s cancer program. Paulk described a 10-year-old who needed 214 seconds to navigate a low-light maze before treatment, later had “perfect vision,” and eventually obtained a driver’s license.

For near-term gene therapy, regulation may be a bigger delay than laboratory science. Paulk said Siren was ready to begin its brain-cancer trial but faced about 18 months of paperwork, while a new viral-gene-therapy program could still take 10–15 years largely because of regulatory review between stages.

Paulk identified capital—not technology—as biotechnology’s biggest current bottleneck, with manufacturing a persistent second constraint. She put average viral-gene-therapy development at $2–3 billion and said even a major CDMO such as Catalent might not touch a 5,000-liter virus run for two and a half years.

AAV’s small payload is less limiting than it sounds: Paulk said it carries about 4.75 kilobases, yet roughly 80% of protein-coding genes fit for gene transfer. Gene editing expands the reach further because the vector can deliver cutting or editing machinery rather than an entire replacement gene.

The most provocative conclusion was that technical capability is beginning to outrun the medical-regulatory framework. Paulk said shorter-sleep genetics and night vision are technically achievable, but FDA trials are built around diseases—not enhancements such as needing less sleep or aging—leaving permission and ethics as the unresolved barrier.