Browsing by Author "Uerlich, Martin Theis"

Glioblastoma, the primary type of glioma, remains a major threat with very few treatment options. Currently, only surgery, radiotherapy and limited medications are available to treat patients. Therefore, there is a great need for novel treatments, and immunotherapies could prove to be a viable option. Today, several immunotherapy approaches are under investigation, and for many of them, the discovery of antigens that distinguish tumor cells from healthy cells is essential. Neoantigen discovery involves various methods to identify or predict antigens that are characteristic and unique to tumor cells. A promising approach is the identification of antigens presented on human leucocyte antigen (HLA) complexes. To find these antigens (or peptides) experimentally, the tumor cells need to be lysed and the elute peptides separated from the HLA class complexes. These peptides can then be evaluated for their binding affinity and immunogenicity to screen out candidates for various antigen-based immunotherapy. The focus of this thesis project was to develop and evaluate an experimental pipeline to characterize the immunopeptidome of glioblastoma cells with special focus on peptides presented on either HLA- class I or HLA class II complexes. For this purpose, different elutions were tested in an immunoaffinity purification protocol. Furthermore, two liquid chromatography coupled tandem mass spectrometry (LC-MS/MS) protocols were compared based on peptide yield and binding affinity prediction of the identified peptides. Finally, two computational models for grouping the peptides based on their binding motifs were compared. To optimize the immunoaffinity purification method, three peptide elution conditions with acetonitrile concentration ranging from 30% to 80% were tested and 40% was found to provide the highest peptide yield and quality. The peptides were then identified by two LC-MS/MS protocols and the column properties and peptide yields were compared. The larger column diameter, longer gradient times and overall, more refined HLA-II peptide protocol of the second method provided better results. Finally, two separate motif deconvolution tools were evaluated based on their success in assigning the identified peptides to cell line specific HLA alleles. Semi-supervised clustering proved to be more feasible for HLA class II peptide grouping, especially at lower peptide numbers. For future application, these results may provide direction on how to modify current immunoaffinity purification and mass spectrometry based neoantigen discovery pipelines for HLA class II peptides. This is particularly valuable because the HLA class II immunopeptidome has not been studied to the same extent as the HLA class I immunopeptidome. However, known HLA class II target antigens could expand potential immunotherapy strategies to include more arms of the immune system, particularly CD4+ cells, which have been shown to enhance CD8+ and immunotherapy efficacy.