MimicMaker: an integrated platform for identifying relevant mimics of T cell antigens - SUMMARY Advances in immunology are revolutionizing medicine. The last 10-20 years have seen the rapid development of a range of transformative immunotherapies, with numerous others on the horizon, including new precision and personalized therapies for cancer and infectious disease. Advances in immunology are also powering new ways to address health conditions that emerge from immune dysfunction or dysregulation. Many immunotherapies center on the cellular arm of the adaptive immune system. In cellular immunity, T cells use their αβ T cell receptors (TCRs) to recognize small peptides (or “epitopes”) bound and presented by class I or class II major histocompatibility complex (MHC) proteins. While specificity is a hallmark of T cell recognition, TCRs are also broadly cross-reactive. The biological imperative for TCR cross-reactivity results from the limited size of an individual’s T cell repertoire compared to the vastly larger number of potential epitopes, as well as the need for T cells to recognize self for positive selection and homeostasis. However, TCR cross-reactivity also poses substantial risks for new and existing immunotherapies and leads to autoimmunity and cellular rejection of transplanted organs. TCR cross-reactivity can almost always be ascribed to the concept of molecular mimicry, where cross-recognized peptides share key structural and physical features. However, mimicry is frequently obscured behind structural and physicochemical complexities. These complexities have generally prevented the prospective identification of mimics of T cell epitopes, greatly complicating derisking and pre-clinical testing in immunotherapy and hindering our ability to address the underlying immunology of T cell driven autoimmunity and transplant rejection. Indeed, molecular mimics have traditionally only been identified and understood after cross-recognition – often in the form of a clinical presentation or complication – is observed. To address this major limitation in translational immunology and advance precision immunotherapy, we have begun developing a technology platform that, beginning with a known target epitope and its presenting MHC protein, uses data science, AI-driven 3D structural modeling, and structure-based scoring that incorporates advances in deep learning and structural analysis to prospectively identify molecular mimics within genetic databases. Our platform, termed MimicMaker, directly addresses the complexities of cross-reactivity in T cell recognition. It leverages our decades of experience in structural immunology, protein biophysics, and immunoinformatics. In two iterative Specific Aims, we will develop and refine MimicMaker, generating and using new data from our mouse model of virus-accelerated transplant rejection. Collaborations with industry and academia are in place to help test, optimize, and validate the platform and, with success, facilitate its adoption and eventual commercialization to improve outcomes in areas such as oncology, autoimmunity, and transplantation.
