Antibody-like therapeutics to redirect effector T cells to CMV-positive glioblastomas
Introduction
To address the short-comings of current therapeutics targeting glioblastoma (GBM), we aimed to develop high affinity antibody-like proteins that specifically bind the pp65 protein found on the surface of the >90% of all GBM tumors. The pp65 protein is a validated GBM target that is unique among human tissues since it is produced by human cytomegalovirus, a feature which is expected to reduce “on-target, off-tumor” side effects common to many antibodies targeting tumor associated antigens. Once developed, this targeting protein can be incorporated into many different therapeutic modalities, including engineered T cells and bi-specific T-cell engagers, and combined with complementary strategies to reduce the immunosuppressive GBM tumor microenvironment and reverse T-cell exhaustion.
Methods
We designed four sequential structure-guided libraries of the human T-cell receptor (TCR) RA14 and selected for variants with increased affinities using a mammalian cell display system. We then generated a bispecific T-cell engager with one arm comprising the TCR and the other an antibody binding the T-cell antigen CD3. This protein was characterized for peptide specificity, in vitro T-cell activation against peptide-pulsed and CMV-infected target cells, with a syngeneic mouse tumor model in progress.
Results
We engineered an antibody-like TCR that binds the pp65-A2 complex with a binding affinity of 0.6 nM that detected as few as ~15 pMHC complexes on the target cell surface. This protein exhibited an attractive thermostability profile and expressed at levels similar to an antibody. Incorporation into a bispecific T-cell engager format revealed a strong correlation between TCR affinity and T-cell activation in the presence of peptide-pulsed target cells; it was also able to activate T cells in the presence of CMV-infected fibroblasts. Screening against a panel of peptides with single point mutations and similar peptides culled from the human repertoire observed no significant off-target activation at low micromolar peptide concentrations.
Conclusion
Personalized therapeutic strategies such as autologous transfer of pp65-specific T cells and ex vivo pulsing of dendritic cells with pp65 have shown promise to increase GBM patient survival. To leverage these results into a protein therapeutic with predictable pharmacokinetics and dosing, low toxicity risks and straightforward manufacturing processes, we developed antibody-like proteins to redirect T cells to eliminate pp65-positive cells. In vitro data support further development and evaluation of this protein in a relevant mouse model.