Probing the immunopeptidome: enhanced epitope discovery through sHLA technology and bioinformatics

© 2018 Dr Katherine Elise ScullHuman leukocyte antigen (HLA) molecules are cell-surface glycoproteins that present peptides, derived from diverse protein antigens, for surveillance by T lymphocytes. This immunosurveillance seeks signs of disease or abnormality, facilitating the eradication of infected or malignant cells. Collectively, the vast array of HLA-bound peptides (including immunogenic epitopes) is termed the immunopeptidome. Many research groups seek to identify the peptides which comprise the immunopeptidome of particular HLA allomorphs or cell types using tandem mass spectrometry. However, the nature of the immunopeptidome presents particular challenges for such discovery studies; the peptides are not only hugely diverse both qualitatively and quantitatively, but their complex biological origins render standard proteomic methodologies problematic. For example, conventional analysis software typically identify peptides by matching tandem mass spectra with sequences from genome-based protein databases, but HLA-bound peptides can have sequences not found in such databases, causing the software to ignore or misidentify such peptides. My thesis aims to facilitate epitope discovery in two ways. Firstly, I investigated an existing experimental technique, secreted HLA (sHLA) technology, in which a secreted form of a chosen HLA allele is transfected into cells, allowing straightforward and specific purification of the HLA allotype of choice. I validated the use of sHLA technology by showing that sHLA resembles natural membrane-bound HLA in terms of the molecules’ maturation kinetics and peptide repertoires. Secondly, I used computational methods and novel bioinformatics to aid immunopeptidomics studies. This involved the development and implementation of several computer programs. The major program is ‘Mmers’, which aids identification of peptides with non-standard sequences, in addition to those with sequences found in standard genome-based protein databases. Mmers generates comprehensive ‘artificial databases’ which include all of the possible permutations of amino acids for peptides of a given length, then searches spectra using Mascot-based scoring. To test Mmers, I obtained tandem mass spectrometric data from complex SW480 and SW620 colon cancer immunopeptidome samples, and searched for peptides of 8-11 amino acids in length using Mmers, alongside conventional software. I showed that Mmers can identify many sequences in agreement with Mascot, ProteinPilot and PEAKS DB. Furthermore, despite statistical challenges necessitating more manual inspection than desired, Mmers allowed me to identify four novel peptides with non-standard sequences, which I validated in comparison to synthetic peptides by MRM. I wrote two minor programs, Shifty and Spliceprot, to start investigating the possible biological origins of the four peptides. I found that two of the peptides could derive from unusual transcription or translation from the WAPL oncogene and the PTEN tumour suppressor gene, respectively. In conclusion, my thesis confronts the challenging complexity of the immunopeptidome, and seeks to provide novel tools to help researchers understand it more fully, with potential benefits for developing cancer immunotherapies and vaccines, and for determining the causes of autoimmune disease