The importance of T cells in recognizing and destroying cancer cells has been shown by many different
approaches, most recently through the clinical effectiveness of checkpoint inhibitor therapies. Over the past few
decades, a completely independent approach has revealed the potency of T cells in an effect called graft vs
leukemia (GVL). The clinical benefit of GVL derives in large part from the long-appreciated phenomenon in which
allogeneic MHC products induce very strong T cell responses. In the case of GVL, the effect involves the
presentation of peptides by class I MHC on leukemia cells to alloreactive T cells derived from the donor bone
marrow. We recently showed that it was possible to use a single mutation in the Kb molecule to generate an ‘allo-
like’ molecule that could overcome tolerance in the induction of CD8+ T cells; once activated, the T cells could in
turn recognize antigens expressed by the wild type Kb expressed on a tumor. The full potential of this approach
will be realized with MHC mutations that can be identified experimentally in order to induce optimal anti-tumor T
cell responses. We will use our expertise in deep mutational scanning and protein engineering of class I MHC to
develop allo-like MHC molecules (‘mut-MHC’) that induce effective T cell responses against cancer antigens.
Our hypothesis is that MHC mutants selected for improvements in peptide binding and/or TCR affinity will
enhance T cell responses against multiple cancer antigens. A key advantage of this approach, compared to
efforts to identify cancer neoantigens for vaccine purposes, is that it does not require identification of the most
effective neoantigens and that it allows aberrant self-antigens to serve as more effective antigens.
By engineering the Kb system, we will be able to test the hypothesis in syngeneic mouse tumor models using
the mouse glioblastoma lines GL261, SMA-560, and CT2A. Importantly, our experience with engineering the
human HLA-A2 system will allow us to translate the findings to human cancers. The specific aims of the project
are: Aim 1. To use deep mutational scans of the Kb molecule to characterize enhanced functional
mutants. We have recently completed a deep mutational scan of Kb expressed on the yeast surface, examining
every substitution of the alpha1 and alpha2 domains (3,420 mutations), using antibody and TCR probes. This yielded 50
to 100 Kb mutations with candidate potential for improved peptide stability or TCR binding. This Aim will further
evaluate this collection of mutants. Aim 2. To examine mutated Kb molecules for their ability to overcome
tolerance against cancer peptide antigens. We will examine the top Kb mutants for their ability to present
cancer antigens, as whole tumor cell or DC vaccines, that overcome tolerance in syngeneic mouse glioblastoma
models (GL261, SMA-560, and CT2A). The results will provide a blueprint for the effective use of rationally
engineered MHC molecules to overcome T cell tolerance against cancer antigens. The approach would provide
an alternative, or adjunct, to other T cell directed therapies such as checkpoint inhibitors.