The Jackson Laboratory for Genomic Medicine Farmington, Connecticut, United States
Objective: Osteosarcoma (OS) remains a debilitating disease within the pediatric population, with few treatment developments over the past four decades, and very limited options for metastatic disease. In anticipation of the need for effective immunotherapy strategies for OS, we have developed a custom workflow to identify and select for OS tumor-specific protein variants which contain novel peptide sequences that could be exploited as novel targets for immunotherapy.
Methods: We built a custom computational proteogenomic workflow to integrate a wide variety of transcriptomic, proteomic, and genomic analysis tools for this study. With three pretreatment human OS clinical samples, we carried out Pacific Biosciences long-read RNA sequencing and Illumina short-read RNA sequencing to generate a high-quality OS transcriptome containing over 10K novel full length coding transcripts per tumor sample. Differential transcript usage analysis was performed for the full-length transcripts using 86 OS tumor samples from the TARGET (Therapeutically Applicable Research to Generate Effective Therapies) study and 126 healthy tissue samples from GTEx (Genotype-Tissue Expression). Predicted coding sequences for novel transcripts were validated by identifying peptide spectra matches (PSM) between the predicted protein sequence and publicly available mass spectrometry spectra files which included OS and other cancer cell lines. A select set of candidate prospective protein variants were further evaluated based on mass spec spectra of variant specific synthetic peptide sequences with OS cell line lysate.
Results: We identified over 30 novel transcripts that are enriched in OS, supported by at least one PSM, and contain a unique novel peptide sequence as compared to all Uniprot protein variants. Additional evaluation revealed 27 candidate protein isoforms that were predicted to contain immunogenic peptide sequences in regions of novel amino acid sequences. Lastly, using variant specific synthetic peptide sequences we further confirmed the unique regions of 5 OS enriched novel protein variants.
Conclusion: Our custom proteogenomic workflow identifies novel protein variants that could be promising candidates for novel immuno-oncology therapy development. Although the majority of candidates do not contain extracellular topology, our targets are predicted to bind MHC class I and can be used as targets for developing T cell receptor mimic (TCRm) antibodies.
