Robust proteomic strategies to detect viral infection in tumors

Brian Searle, PhD - Translational Research Fellow, Institute for Systems Biology

Host: Matt Bush
 

Infections cause 15% of cancers worldwide, with HPV, Hepatitis B, and Hepatitis C being the leading viral causes. Two human herpes viruses, EBV and HHV-8, are known to cause cancer, and a third, human cytomegalovirus (HCMV), has been reported in multiple brain tumor types, including glioblastomas. However, despite confirmatory studies by many different labs, the relationship between HCMV and cancer has remained controversial, in part due to the absence of any HCMV DNA or RNA sequences in large cancer sequencing databases such as the TCGA. In this talk we show two proteomics applications of a new mass spectrometry strategy using data independent acquisition (DIA) to detect low levels of infection in human tissue. First, coupling DIA with gas-phase fractionation, we create high-resolution “digital maps” of tumors that can be iteratively mined as new hypotheses are derived. Second, with matched-matrix dilution experiments, we demonstrate an approach to assign analytical figures of merit, such as lower limits of detection and quantification, to every measured peptide. We then use these parameters to build rapid, fit-for-purpose infection assays targeting the most well-characterized peptides. With these approaches, we solve a two decades-old controversy in the immuno-oncology literature by demonstrating the widespread absence of HCMV in glioblastomas and other brain metastases, despite positive immunohistochemistry staining in the same tumors with multiple antibodies and antigens. New hypotheses emerge to reconcile our disparate datasets, and our findings open doors to future brain cancer immunotherapies.

Career Q&A with Dr. Brian Searle: Students and postdocs are invited to stay after Dr. Searle’s seminar on Monday afternoon to ask Dr. Searle ask scientific questions and questions about his career.