Deep-Learning-Driven\nDiscovery of SN3–1, a\nPotent NLRP3 Inhibitor with Therapeutic Potential for Inflammatory\nDiseases

The\nNLRP3 inflammasome plays a central role in the pathogenesis\nof various intractable human diseases, making it an urgent target\nfor therapeutic intervention. Here, we report the development of SN3–1,\na novel orally potent NLRP3 inhibitor, designed through a lead compound\nstrategy centered on deep-learning-based molecular generative models.\nOur strategy enables rapid fragment enumeration and takes into account\nthe synthetic accessibility of the compounds, thereby significantly\nenhancing the optimization of lead compounds and facilitating the\ndiscovery of potent inhibitors. X-ray crystallography provided insights\ninto the SN3–1 inhibitory mechanism. SN3–1 has shown\na favorable safety profile in both acute and chronic toxicity assessments\nand exhibits robust pharmacokinetic properties. Furthermore, SN3–1\ndemonstrated significant therapeutic efficacy in various disease models\ncharacterized by NLRP3 activation. This study introduces a potent\ncandidate for developing NLRP3 inhibitors and significantly expands\nthe repertoire of tools available for the discovery of novel inhibitors.