Extracellular vesicle immunopeptidomics recovers immunogenic neoantigen peptides from cancer patient plasma

Personalized immunotherapy using patient-specific neoantigens is a critical strategy in immuno-oncology. DNA/RNA sequencing and MHC affinity modeling are used to select optimal neoantigen peptides for development. Tissue-based peptide-MHC (pMHC) mass spectrometry (MS) immunopeptidomics studies have been used to confirm MHC presentation of neoantigen peptides. Since these pMHC analyses are tissue intensive, novel ways to identify MHC presentation of patient-specific neoantigen peptides are needed. Extracellular vesicles (EV) are constitutively expressed by all tumor and immune cells and represent a potential plasma source of patient specific pMHC complexes. Using a proprietary EV purification method, we report detection and characterization of patient specific neoantigen pMHC complexes from very small volumes of patient plasma using targeted proteomics. Methods: Consented patient plasma was used in this study, and patient metadata (DNA/RNA sequencing, HLA haplotyping, and neoantigen peptide binding predictions) was used to define potential neoantigens. EV were purified from plasma (0.5 mL) using proprietary size exclusion chromatography (SEC) methods. Class I MHC-presented peptides were enriched from samples using a pan-MHC antibody column, and enriched peptides were analyzed by high-resolution targeted LC-MS/MS with heavy peptide standards . Single chain trimers (SCT) for pMHC complexes were used to evaluate neoantigen-specific T cell frequencies in the peripheral blood, sort T cells for TCR sequencing, and engineer TCR-T cells for further neoantigen characterization. Results: EVs purified from patient plasma using proprietary SEC methods surprisingly yielded ~20 greater amounts of total class I MHC protein based on ELISA than standard EV preps, illustrating a potential for pMHC detection. Plasma enriched EVs from two patients (advanced CRC and osteosarcoma (OST)) were used in subsequent experiments. Multiple patient-specific pMHC were recovered from each patient, 3 peptides from OST and 5 peptides from CRC. Expression of multiple pMHC were confirmed in tumor tissue by MS. Peptides represent candidate neoantigens based upon expression and MHC presentation. Immunogenicity was confirmed for multiple peptides (4 CRC/ 2 OST) by identification of neoantigen-specific CD8+ T cells in peripheral blood using SCT staining and flow cytometry. Subsequent studies will be reported with engineered TCR-T's. Conclusions: Identification of patient-specific tumor antigens is a prerequisite for personalized immunotherapy. We have developed a workflow for Extracellular Vesicle Immunopeptidomics (EV-IMPX) to validate neoantigen presentation using small-volume patient plasma samples. We expect that peptide selection based on EV-IMPX will enhance immunotherapy by narrowing neoantigen selection to MHC-presented peptides with the highest likelihood of tumor killing.