DeepCure Selects First AI-Generated Candidate DC-9476 for Autoimmune Diseases

BOSTON -- DeepCure, a therapeutics company specializing in AI-driven drug discovery for inflammation and immune diseases, has unveiled its first development candidate, DC-9476. This third-generation BRD inhibitor is selective for the BD2 domain of Brd4 and is slated to enter clinical trials by 2025. The BD2 domain of Brd4 is integral in regulating a variety of cytokine-related pathways such as TNF-alpha, IL6, IL-17, and IL-1b, all of which play significant roles in autoimmune diseases.

Autoimmune diseases often involve epigenetic alterations, including histone modifications that change chromatin structure and gene expression. The BET protein family, which includes Brd4, interprets these histone modifications, leading to long-lasting gene expression patterns in both innate and adaptive immune cells. Additionally, Brd4 BD2 activity can directly influence transcription factors. By inhibiting Brd4 BD2 with small molecule drugs, it is possible to disrupt harmful cycles of gene expression across multiple pathways, thus restoring the self-resolving processes that typically follow inflammatory triggers.

Preclinical models for various autoimmune diseases, such as rheumatoid arthritis (RA) and Still’s disease, have demonstrated DC-9476's promising efficacy. In RA models, DC-9476 outperformed existing therapies such as TNF-alpha inhibitors, IL-6 inhibitors, and the JAK inhibitor tofacitinib. It also enhanced the effectiveness of TNF-alpha inhibitors when used in combination therapy. Unlike previous generations of non-selective inhibitors, DC-9476 has shown an excellent safety profile, notably avoiding issues like thrombocytopenia.

The discovery of DC-9476 was made possible through DeepCure’s proprietary AI platform, which leverages differences in the structure and function among the BET protein family and their bromodomains (BD1 and BD2). This AI platform integrates advanced machine learning (ML) and physics-based tools to identify interaction sites on the protein surface (PocketExpander™) and to design novel, synthetically feasible compounds (MolGen™).

Professor Francesco Del Galdo, M.D., Ph.D., a Professor of Experimental Medicine at the NIHR Biomedical Research Centre and the Institute of Rheumatic and Musculoskeletal Medicine at the University of Leeds, commented on the significance of this discovery. He emphasized that autoimmune diseases result from multiple disrupted pathways and noted that while targeted therapies against specific cytokines, kinases, or receptors have been beneficial, they often come with increased toxicity when combined. DC-9476, with its novel mechanism for addressing multiple pathways and excellent safety profile, holds great promise for the treatment of autoimmune diseases.

DeepCure’s CEO & Co-Founder, Kfir Schreiber, expressed enthusiasm about this milestone, stating, “We’re excited to announce our first AI-generated candidate, which we plan to advance into clinical trials next year. This validates our generative AI and physics-based drug discovery engine.” He further highlighted that DC-9476 is the first of many innovative development candidates from their platform, with the potential to revolutionize the treatment of inflammation and immune diseases.

DeepCure, founded by researchers at MIT, aims to accelerate the development of breakthrough therapies using artificial intelligence and related technologies to innovate small molecule discovery. Based in Boston, MA, the company’s team of engineers, chemists, and biologists work together to solve complex problems, ultimately aiming to significantly impact patient health.