BenevolentAI poaches drug discovery executive from GlaxoSmithKline 

The supercomputer used at BenevolentAI
The DGX-1 supercomputer that powers BenevolentAI's research.

BenevolentAI has hired Ian Churcher, D.Phil., from GlaxoSmithKline. The appointment of Churcher as VP of drug discovery furthers the secretive AI-enabled startup’s track record of poaching executives from Big Pharma.

Churcher joins BenevolentAI after heading up a GSK Discovery Performance Unit that researched fields including protein degradation. Prior to joining GSK, Churcher worked on Alzheimer’s disease at Merck, giving him a history that overlaps with BenevolentAI’s interest in neurodegeneration and deal with an unnamed U.S. drugmaker targeting the hard-to-treat disease. 

BenevolentAI lured Churcher away from GSK, ironically at a time when the Big Pharma is dialling up its own AI activities, on the strength of the potential for its data-crunching platform to improve the discovery and development of drugs. That potential also persuaded Jackie Hunter, Ph.D., D.Sc., formerly of GSK, to join BenevolentAI, where she has been joined by émigrés from Pfizer and other Big Pharmas. 

Churcher joins BenevolentAI as it is starting to come out of its shell following a period flying about as far under the radar as is possible for a company that is hiring executives from Big Pharma, raising upward of $100 million and striking deals worth up to $800 million.

The first evidence of BenevolentAI’s clinical development programs has now emerged in filings with regulators on both sides of the Atlantic. Led by the former Glenmark Pharmaceuticals CMO with assists from clindev staffers poached from Takeda, BenevolentAI is running a phase 2b trial of bavisant in Parkinson’s disease patients who suffer from excessive daytime sleepiness.

Johnson & Johnson put bavisant through a clutch of phase 1 and 2 trials starting about 10 years ago to assess the H3 receptor antagonist in healthy volunteers and patients with attention deficit hyperactivity disorder (ADHD). But the Big Pharma’s interest in the compound waned after it failed to move the needle in a 430-person ADHD trial.

However, for BenevolentAI’s purposes, the side-effect profile of bavisant is more relevant than the efficacy miss. The most common adverse event was dose-dependent insomnia. That finding added to evidence of bavisant’s effect on wakefulness and sleeping.

BenevolentAI, after crunching the numbers in its AI platform, identified bavisant as a potential treatment for Parkinson’s disease patients who suffer from excessive daytime sleepiness. And it followed up that finding by licensing the drug and other clinical-stage assets from J&J about one year ago. 

London-based BenevolentAI has since filed to run a 230-patient dose-finding trial. The trial will look at doses ranging from 0.5 to 3 mg/d, notably lower than the 1 to 10 mg/d administered by J&J.