Proteomic approach to discover human cancer viruses from formalin-fixed tissues

The challenge of discovering a completely new human tumor virus of unknown phylogeny or sequence depends on detecting viral molecules and differentiating them from host molecules in the virus-associated neoplasm. We developed differential peptide subtraction (DPS) using differential mass-spectrometry (dMS) followed by targeted analysis to facilitate this discovery. We validated this approach by analyzing Merkel cell carcinoma (MCC), an aggressive human neoplasm, in which ~80% of cases are caused by the human Merkel cell polyomavirus (MCV). Approximately 20% of MCC have a high mutational burden and are negative for MCV, but are microscopically indistinguishable from virus positive cases. Using 23 (12 MCV positive, 11 MCV negative) formalin-fixed MCC, DPS identified both viral and human biomarkers (MCV Large T antigen, CDKN2AIP, SERPINB5 and TRIM29) that discriminates MCV positive and negative MCC. Statistical analysis of 498,131 dMS features not matching the human proteome by DPS revealed 562 (0.11%) to be up-regulated in virus-infected samples. Remarkably, four (20%) of the top 20 candidate MS spectra originated from MCV T oncoprotein peptides and confirmed by reverse translation degenerate oligonucleotide sequencing. DPS is a robust proteomic approach to identify novel viral sequences in infectious tumors when nucleic acid-based methods are not feasible.