The role of nonsense mediated decay in antigenic HLA peptide generation

Background: The nonsense mediated decay (NMD) pathway has recently been implicated as
a novel target for cancer immunotherapy [1,2,3], however the fundamental biological link
between NMD and HLA presentation of mutated epitopes is unclear. Two competing
hypotheses have been proposed, the first states that NMD must be bypassed, which hence
leads to increased mutated protein abundance/HLA presentation via canonical mRNA
translation. The second model proposes that the non-canonical "pioneer round of
translation", which occurs directly after nuclear export in coordination with NMD, generates
the major source of antigenic peptides [4,5]. Addressing this fundamental question is of
urgent biological/therapeutic relevance. In addition, the exon junction complex model of NMD
efficiency provides a poor predictive classification of which variants will and won't be trigger
degradation, creating a challenge in interpreting the germline pathogenic and immunogenic
potential of individual mutations.
Goals:
1. To establish whether frameshift mutated proteins generate HLA presented neoepitopes via
canonical mRNA translation, or the "pioneer round of translation". Dr Litchfield's laboratory
will utilise genomics, transcriptomics and immunopeptidomics data to identify a set of known
frameshift mutated proteins which trigger validated HLA peptides. Dr Carlton's laboratory will
use live cellular imaging to track these mutated proteins in vivo, to establish the route of
translation and HLA presentation, using an adapted version of the TRICK assay (PMID:
25792328).
2. To develop an improved predictive model to classify whether premature termination codon
mutations will or won't trigger NMD, by using a saturation genome editing approach. Utilising
the design generated from the rotation project below, premature termination codon
mutations will be introduced at every amino acid position across the length of a selected set
of reference genes using CRISPR-Cas9 editing. Cells will then be screened using a
fluorescence reporter to establish whether NMD was triggered at each mutated position. Dr
Litchfield's laboratory will then develop a novel machine learning classifier algorithm to
predict which mutations do and don't trigger NMD.