BenevolentAI, cruel R&D: AI-enabled drug flunks midphase eczema trial to dent deal plans

BenevolentAI’s hopes of out-licensing its atopic dermatitis candidate this year have taken a hit. In a phase 2a trial, the topical pan-Trk inhibitor failed to improve eczema symptoms or itch, denting the prospects of the program and raising questions about the AI-enabled model that proposed the approach. 

While tropomyosin receptor kinases (Trk) are best known as the targets of cancer drugs, BenevolentAI’s platform identified the receptors as mediators of itch and inflammation in atopic dermatitis, the most common form of eczema. The potential to simultaneously address the symptoms in mild to moderate atopic dermatitis with fewer side effects than existing drugs led the biotech to design Trk inhibitors. 

BEN-2293, an inhibitor of TrkA, TrkB and TrkC, emerged as the lead candidate, and BenevolentAI stuck by it after a phase 1b clinical trial “demonstrated a trend towards a clinically meaningful effect,” betting that a larger, powered midphase study would validate the approach.

The bet failed. The phase 2a enrolled 91 adults to use either BEN-2293 or placebo on their affected skin for 28 days. BEN-2293 proved to be safe and well tolerated, the primary endpoint of the trial, but slipped up on the secondary efficacy endpoints. The drug candidate was no better than placebo at improving the area and severity of eczema or itching.

BenevolentAI had planned (PDF) to out-license the asset in the second half of the year to a pharma company focused on dermatology. The deal was one of two agreements on BenevolentAI’s road map for 2023, with the company also looking to enter into a collaboration along similar lines to its alliance with AstraZeneca. 

The failure is a blow to both BEN-2293 and, by extension, the platform that identified the approach to atopic dermatitis. But BenevolentAI is yet to give up on the drug candidate, pointing to evidence that the treatment performed better in patients with a greater extent of disease at baseline to make the case for further investigation. The biotech is reviewing the data before deciding on its next steps.  

While BenevolentAI’s interest in drugging Trk to treat atopic dermatitis stemmed from its AI platform, it is not the first company to try the approach. The now-defunct Creabilis hit upon the idea more than a decade ago and ran a phase 2 trial of its TrkA inhibitor in atopic dermatitis.