Quantitative Approach to Supramolecular Assembly Engineering for Isolating and Activating Antigen-Specific T Cells

Abstract

T cell immunotherapy is a novel therapeutic strategy that aims to leverage the antigen-specific nature of a T cell immune response to treat a variety of immunological conditions. Over the past twenty years, T cell immunotherapy has been applied to treat several types of cancer, autoimmune conditions, and chronic infections, culminating in the FDA approval of two highly effective chimeric antigen receptor (CAR) T cell therapies targeting hematological cancers in 2017. While the initial success of T cell immunotherapy has been encouraging, identifying appropriate antigenic targets and optimizing T cell activation to promote effective responses in vivo remain significant challenges. In this dissertation, we discuss the development and application of new molecular tools for identifying, isolating, and activating antigen-specific T cells, which are directly relevant to the current challenges facing T cell immunotherapy. One of the greatest obstacles to developing a successful T cell immunotherapy is the selection of appropriate antigenic targets. T cells naturally recognize antigen-derived peptides presented on polymorphic major histocompatibility complex (MHC) proteins, and different MHC alleles exhibit different peptide binding specificities. Therefore, peptides that promiscuously bind multiple MHC alleles representing a diverse population have significant potential in the development of broadly protective peptide-based therapeutics and vaccines. A number of high-throughput in silico strategies have been developed to predict peptide-MHC binding; however, the accuracy of these approaches is generally inadequate for the reliable prediction of class II peptide-MHC (MHCII) interactions. In contrast, most experimental systems designed to measure peptide-MHCII binding emphasize quantitative detail over throughput. In this dissertation, we develop and validate a high-throughput screening strategy to evaluate peptide binding to four common MHCII alleles. Using this strategy, which we have termed microsphere-assisted peptide screening (MAPs), we screened overlapping peptide libraries of antigenic viral proteins and identified 12 promiscuously MHCII-binding peptides. Subsequent structural analysis indicated that nearly half of these peptides overlapped with antibody neutralization sites on the respective viral protein. Together, these results indicate that the MAPS strategy can be used to rapidly identify promiscuously binding and immunodominant peptides that have therapeutic relevance. Another significant challenge limiting the successful application of T cell immunotherapy is expanding a clinically relevant number of therapeutically effective T cells. The effectiveness of a T cell response is largely determined by the spatial and stoichiometric organization of signals delivered to the T cell during T cell activation. One strategy for promoting an effective T cell response is to tune the presentation of stimulatory and costimulatory signals through artificial antigen presentation. However, existing technologies have a limited ability to control the spatial and stoichiometric organization of T cell ligands on 3D surfaces. In this dissertation, we introduce a novel strategy for presenting highly organized clusters of stimulatory and costimulatory ligands to T cells using protein-scaffold directed assembly. Using this approach, we systematically investigated how the global surface density, local valency, and stoichiometric ratio of T cell ligands on a 3D cellular (yeast) surface can be manipulated to tune T cell activation. After validating this approach, we further develop more complex scaffold-assembly schemes to enhance the controllability of isolating and activating antigen-specific T cells. We believe that MAPS and artificial antigen presentation using protein-scaffold directed assembly provide a robust toolset for identifying, isolating, and activating antigen-specific T cells for T cell immunotherapy.