Utilising molecular dynamics simulations to understand and engineer T-cell receptors

Approximately 90% of all therapeutic targets in the human proteome operate solely inside the cell, making them unavailable for recognition by antibodies which instead bind antigens presented on the exterior of the cell surface. To target the 90%, the human immune system utilises a class of binding proteins known as T-cell receptors (TCRs). These TCRs recognise peptide fragments sourced from proteins produced inside the cell that have subsequently been degraded and transported to the cell surface by the human leukocyte antigen (HLA, pHLA with peptide bound). TCRs are membrane bound and attached to T-cells and use their six complementarity-determining region (CDR) loops to bind antigenic pHLA molecules (i.e. peptides that come from protein antigens). TCR binding to pHLA molecules induces an immune response from the T-cell, which ultimately leads to the antigen presenting cells death. The capability of TCRs to identify antigens which are not naturally expressed on the cell surface (unlike antibodies) has helped drive the development of a new class of therapeutics that consist of a soluble, bispecific TCR engineered to bind a specific antigenic pHLA for the treatment of various diseases (such as cancers and viral infections). Natural TCRs bind with characteristically poor affinities (~μM) and half-lives (~seconds), which are undesirable properties for a therapeutic. The CDR loops on TCRs are therefore normally subjected to affinity maturation to produce TCRs with affinities in ~pM range for their target pHLA. This does however carry a significant risk in terms of safety, as the very large majority of peptides presented by HLA molecules are sourced from endogenous (i.e. healthy) proteins and must not be bound by the TCR in order to avoid the production of an autoimmune response on the healthy cells. These requirements for a highly specific TCR that binds with high affinity to its target pHLA is the primary motivation for this thesis, as herein, fundamental engineering principles for generating TCRs with these properties are determined and protocols to evaluate these properties are developed and demonstrated. This insight is obtained through combinations of structural analysis, molecular dynamics simulations and free energy calculations, providing an atomistic description of how this has occurred for several TCRs. Furthermore, we characterise how different peptide cargo can tune the molecular flexibility of the entire pHLA molecule, including regions distal from the HLA binding site. These findings suggest peptide dependant tuning of the HLA molecule may play a role in regulating the functional outcome of an immune response. Ultimately, this work and the principles identified herein will aid in the rational design of high affinity and high specificity TCRs as therapeutics for various diseases.