AI unlocks new pathways in pain relief discovery

An unexpected discovery in an automated lab in Shanghai sparked a breakthrough in pain management. A neuroscience team at Insilico Medicine used their AI platform, PandaOmics, to analyze vast datasets and identified a protein target, Z, which was previously linked to a different disease. The AI's insight led to the development of a new molecule, ISM9528, which outperformed morphine in preclinical trials. This non-opioid candidate represents a significant step toward addressing chronic pain, a global issue affecting nearly 27% of the population, with over 24% of U.S. adults impacted. Traditional pain treatments like opioids and NSAIDs come with severe side effects and limited efficacy, prompting the team to explore novel biological pathways. By integrating AI-driven target discovery, experimental validation, and generative chemistry, the team designed ISM9528, which can cross the blood-brain barrier and selectively bind to Target Z. The success of this approach highlights AI's potential to transform drug discovery by enabling systematic scientific serendipity. Insilico Medicine's MMAI Gym, an open platform for training AI models in drug discovery, has shown improvements in model performance by up to 10-fold. The company's lead AI-discovered asset, rentosertib, entered Phase III trials in mid-2026, marking the first AI-discovered target and molecule to reach late-stage evaluation. Despite preclinical success, human trials will determine ISM9528's potential as a transformative non-opioid treatment. Alex Zhavoronkov, PhD, founder and CEO of Insilico Medicine, emphasized the synergy between AI and human expertise in advancing medical innovation. The company has partnered with 13 of the world’s 20 largest pharmaceutical firms, underscoring growing industry interest in AI-driven drug discovery. As the field moves toward clinical trials in 2027, the case of ISM9528 illustrates how AI can expand the scope of biological inquiry and enable new therapeutic possibilities.