Insilico Medicine nominates AI-designed candidate for ocular diseases

Insilico Medicine has nominated an artificial intelligence-designed drug candidate for the potential treatment of several inflammatory and degenerative eye diseases including dry age-related macular degeneration (dry AMD), uveitis and dry eye disease.

The company said ISM9077 had been selected as a preclinical candidate following studies showing activity across several species, favourable drug exposure and permeability and a safety profile that could support further development.

The candidate targets a protein referred to by the company as Target Y, which it says plays a role in pathological inflammation. Insilico believes the mechanism could potentially be applied beyond eye diseases to conditions including neurodegenerative diseases, inflammatory bowel disease, cardiovascular disease, metabolic dysfunction-associated steatohepatitis (MASH) and obesity.

“ISM9077 marks our 32nd preclinical candidate nominated since 2021, so this is an example of how generative AI delivers at scale,” said Alex Zhavoronkov, Found, CEO and CBO of Insilico Medicine. “This candidate features a moderately novel target, a novel molecule designed from scratch and a novel formulation possibility of eye drop. Beyond offering a non-invasive, highly effective option for millions of patients suffering from blinding ocular diseases, ISM9077 further enhances Insilico’s dual-purpose strategy to treat specific diseases while targeting ageing and extending healthspan, as Target Y is strongly implicated in ageing, longevity and multiple age-related conditions.”

AI used to design compound

Insilico said its researchers used its Chemistry42 generative chemistry platform alongside a structure-based drug design approach to develop ISM9077.

The team first established an understanding of the target’s binding pocket using co-crystal structures before using generative models to design new molecules. The compounds were assessed for factors including pharmacophore fit, drug-likeness, three-dimensional shape and structural novelty.

An AI-based model specific to the target was then used to predict molecular activity and prioritise compounds for further testing. The company said repeated rounds of evaluation and optimisation ultimately produced ISM9077.

Strong preclinical results

In preclinical studies, ISM9077 showed activity across several animal models of ocular disease following both oral administration and topical eye-drop delivery.

The company reported retinal exposure between two and 5.5 times plasma levels, alongside favourable oral bioavailability and low-to-moderate clearance. It also reported a wide safety margin and properties that could support development as an eye-drop treatment.

In models of dry AMD, ISM9077 improved retinal structure and visual function. Insilico said the candidate demonstrated approximately three times the efficacy of currently available therapy on key endpoints while also producing greater improvements in histopathology.

The compound also reduced inflammation and inflammatory cytokine release in uveitis models and improved retinal function. In dry-eye models, it increased tear production and reduced corneal inflammation with a rapid onset of action. The company said it outperformed cyclosporine A, an established treatment for the condition.

“Ocular diseases treatment calls for novel options, as current solutions haven’t reached a balance between toxicity management, efficacy results and dosing adherence,” said Feng Ren, Co-CEO and CSO of Insilico Medicine. “ISM9077 overcomes these limitations through AI-driven structural design, achieving superior target exposure and potential for eye drop formulation. Dry AMD is a leading cause of irreversible central vision loss in older adults and the positive efficacy results in therapeutic studies, as well as Cynomolgus monkey models, first validated the therapeutic value of Target Y for dry AMD, in a way much closer to human mechanism. We look forward to future clinical translation based on such AI-driven outcomes.”

Next steps towards clinical development

ISM9077 is currently at the preclinical candidate stage and will now need development and regulatory testing before any potential clinical trials can begin.

Insilico said the candidate is part of a broader AI-driven drug discovery programme which has nominated 32 preclinical candidates since 2021, with 13 having subsequently received IND approval or clearance.

The company said its approach typically takes 12 to 18 months to reach preclinical candidate nomination, compared with its estimate of 2.5 to four years for traditional early-stage drug discovery programmes.