CHICAGO (GenomeWeb) – At the intersection of genomics and proteomics, investigators are doing basic research bent on improving existing immunotherapies and better understanding response.
Attendees at the American Association for Cancer Research heard about just a few of these potential applications during a major symposium on cancer immunotherapy analyses here yesterday.
Michal Bassani-Sternberg, an oncology researcher and director of immunopeptidomics at the UNIL/CHUV Ludwig Cancer Research Center's experimental therapeutics center, highlighted a series of studies using mass spectrometry-based approaches for prioritizing promising neoantigen targets in tumors — insights that may be applied to everything from personalized cancer vaccine development to CAR-T treatments.
Bassani-Sternberg noted that personalized neoantigen development is increasingly of interest in clinical trials. Generally speaking, investigators have been identifying potential neoantigen-producing mutations in tumors using exome or whole-genome sequencing on tumor biopsy samples.
In 2016, for example, a Washington University McDonnell Genome Institute-led team outlined a tumor DNA and RNA sequencing-based pipeline for finding tumor-specific antigens.
It has been more difficult to prioritize potential neoantigens, though a wide range of strategies have been attempted, from expression level data to human leukocyte antigen (HLA) binding site estimates.
For their part, Bassani-Sternberg and her colleagues have been focusing on in vivo processing and an antigen presentation-based prioritization method, including efforts to identify mutated tumor peptides interacting with the immune system with mass spectrometry-based approaches.
In a proof-of-concept paper published in Nature Communications in 2016, Bassani-Sternberg and co-authors from centers in Germany outlined a mass spec-centered search for promising antigen targets in five individuals with melanoma. That work — one of the studies Bassani-Sternberg described at the AACR meeting — suggested it is possible to directly identify plausible neoantigen candidates, including peptides that prompted a response from one of the patients own T cells.
For a more recent preprint article in BioRxiv, Bassani-Sternberg's team has been continuing to characterize potential HLA ligand contributors, including proteasome-spliced peptides, which may inform future neoantigen search and prioritization efforts.
Bassani-Sternberg cautioned that mass spec-based immunopeptidomics does not appear to be feasible for all cancer patients. With that in mind, investigators are also working on strategies for making better neoantigen predictions.
For example, Bassani-Sternberg was first author on a PLOS Computational Biology study last year that used insights gleaned from dozens of new and previously reported HLA peptidomic datasets to predict class I HLA binding motifs and neoantigens in melanoma and lung cancer collections.
"These high-throughput and unbiased data enable us to refine models of HLA-I binding specificity for many alleles, including some that had no ligand until this study," she and her co-authors wrote in that paper, "and improve predictions of neoantigens from exome sequencing data in melanoma and lung cancer samples."
In a related presentation during the same symposium, Washington University pathology and immunology researcher Robert Schreiber shared details from experiments done in mice with chemically induced sarcomas that were treated with immune checkpoint inhibitor drugs, alone or in combination.
Using 10x Genomics' single-cell RNA sequencing and Fluidigm's mass cytometry (CyTOF) platforms, he noted that he and his team have identified distinct macrophage immune cell populations in control mice relative to mice with induced sarcomas treated with anti-PD1 antibodies, anti-CTLA4, or both.
By following the dynamics of immune cell population expansions or contractions, Schreiber noted, the investigators are getting a clearer look at the immune processes at play in checkpoint blockade-treated mouse models — analyses that might ultimately lead to candidate markers for treatment response.