Bo Ren

Topic

Proteomic Remodeling in Cancer-Associated Systems: Responses to Microbial and Therapeutic Stress

School of Molecular Life Sciences

Date & location

• Tuesday, November 3, 2026
• 9:00 A.M.
• Clearihue Building, Room B007

Examining Committee

Supervisory Committee

• Dr. David Goodlett, School of Molecular Life Sciences, University of Victoria (Supervisor)
• Dr. Javier Alfaro, School of Molecular Life Sciences, UVic (Co-Supervisor)
• Dr. Julian Lum, School of Molecular Life Sciences, UVic (Member)
• Dr. Lisa Reynolds, School of Molecular Life Sciences, UVic (Member)
• Dr. Marie-Ève Tremblay, School of Medical Sciences, UVic (Outside Member)

External Examiner

• Dr. Bingyun Sun, Department of Chemistry, Simon Fraser University

Chair of Oral Examination

• Dr. Jon Husson, School of Earth and Ocean Sciences, UVic

Abstract

Mass spectrometry-based proteomics can characterize how cancer cells and tumor-associated tissues respond to bacterial exposure and treatment-associated stress and can also directly identify bacterial proteins. However, the low abundance of bacteria in human-derived samples, along with sequence similarities, can complicate their direct identification and preclude reliable measurement of quantitative differences.

This thesis first reviews the evidence for microbes in tumor-associated samples, followed by three experimental studies: 1) proteomic responses of cancer cells to controlled bacterial infection, 2) benchmarking of low-abundance bacterial peptide detection in defined bacterial-human protein mixtures, and 3) comparison of paired primary and recurrent ascites samples from patients with high-grade serous ovarian carcinoma (HGSOC).

The review considered DNA, RNA, spatial, proteomic, immunopeptidomic, and functional evidence for bacteria and bacteria-derived material in human tumors. These forms of evidence support different claims about bacterial presence, activity, and biological effects on the host, and therefore require different forms of validation.
In the first study, host responses were examined in melanoma and ovarian cancer cells exposed to bacterial infection. Staphylococcus aureus produced extensive proteomic changes, with changes involving mitochondrial metabolism, cellular stress, and proliferation observed across the experiments. Despite confirmed bacterial exposure and a reproducible host proteomic response, bacterial proteins were not confidently detected by MS-based proteomics.

As a result, a second study benchmarked bacterial peptide identification using defined mixtures of bacterial and human proteins. In silico spectral library searching and conventional database searching differed in sensitivity and background behavior, but neither separated bacterial signal from human-only background at 1% bacterial protein. Reproducible and confident bacterial protein identification above the human-only background emerged only when bacterial protein was present at 2% relative to the cell-line proteome. Consistent with this detection limit, bacterial identifications in infected cells did not exceed those observed in the human-only background despite the reproducible host response.

The third study compared proteomes from paired primary and recurrent HGSOC ascites from three patients using unsorted, CD45⁻ tumor-enriched, and CD45⁺ immune-enriched cell populations. Few individual proteins differed significantly between primary and recurrent samples. Gene Set Enrichment Analysis (GSEA) identified suppression of interferon-associated programs across all three cell populations, metabolic and DNA-repair-associated changes in the CD45⁻ tumor-enriched cell population, and altered immune programs in the CD45⁺ immune-enriched cell population. Because the recurrent samples were collected after treatment, these differences were interpreted as recurrence-associated changes reflecting both treatment history and disease progression.

Overall, this thesis shows that detecting a change in the host proteome, directly identifying bacterial proteins, and interpreting differences between primary and recurrent patient samples are distinct analytical problems. A reproducible host proteomic response did not guarantee confident detection of bacterial proteins, and differences between primary and recurrent HGSOC samples could not be attributed to treatment alone. Each question therefore requires controls and validation appropriate to the conclusion.