Optimization of an MHC-I immunoprecipitation protocol for mass spectrometry analyses of persister-specific antigens in acute myeloid leukemia

Acute myeloid leukemia (AML) is the most frequent and lethal leukemia among adults. Chemotherapy (cytarabine (Ara-C)) is the first-line treatment option and leads to high remission rates (75%). However, most patients eventually relapse. Relapse is mediated by persisters (i.e., cells surviving the treatment) that present a vast transcriptomic reprogramming characterized by the expression of stress responses, such as senescence and diapause. Since we previously showed that alterations in the transcriptome are reflected in the immunopeptidome (set of MHC-I-associated peptides (MAPs)), we hypothesize that persisters present specific and immunogenic MAPs deriving from stress responses. Therefore, we aim to identify such MAPs to design a peptide vaccine targeting these cells and preventing relapse. Mass spectrometry (MS) on immunoprecipitated MHC-I molecules and MAPs is the only method to explore the immunopeptidome. Preliminary analyses showed that few MHC-I molecules could be isolated from AML cell lines. Accordingly, few MAPs (31) were identified by MS. Therefore, we sought to improve the yield of our isolation method and found that performing eight MHC-I elutions instead of two dramatically increased the amount of collected MHC-I. We also found that using next-generation trapped ion mobility time-of-flight (timsTOF) MS dramatically (~300-fold) increased the number of identified MAPs. After bioinformatic filtering of peptides based on their size distribution and predicted binding to MHC-I molecules, a clear enrichment of MHC-I epitopes was observed. Identified MAPs were also radically different (~20% overlap) and had greater GRAVY indexes than others identified previously in the same cell line with Orbitrap MS, suggesting that Tims-TOF enables the discovery of previously unreported MAPs, due to its capacity to capture more hydrophobic MAPs. In conclusion, our MS protocol is operational and will allow us to identify enough MAPs to find good candidates for designing the vaccine.