ORCID
https://orcid.org/0000-0002-5491-2537
Date of Award
8-14-2023
Degree Name
Doctor of Philosophy (PhD)
Degree Type
Dissertation
Abstract
The adaptive immune system has inherent antitumor properties that are capable of inducing tumor-specific cell death. CD8+ and CD4+ T cells, two immune cell types critical to this process, recognize antigens bound by class I and II major histocompatibility complexes on the cell surface, respectively. After antigen recognition, T cells can signal growth arrest and cell death to tumor cells displaying the antigen. Neoantigens, or tumor-specific peptides, are targeted by cancer immunotherapies to induce T cell-mediated degradation. Treatments such as personalized vaccines and immune checkpoint blockade thus provide a unique avenue to selectively recognize and destroy tumor cells while keeping normal cells intact. Clinical trials for personalized vaccines have demonstrated that neoantigens can induce patient-specific T cell responses; however, computationally predicting neoantigens from somatic missense mutations and indels usually render only a small subset of immunogenic candidates. Additional strategies are needed to improve neoantigen prediction accuracy to increase suboptimal immunotherapy patient responses in multiple cancer types. Chapter 2 describes whole exome sequencing (WES) and RNA sequencing (RNAseq) of multi-sector tumors to determine the clonality and immunogenicity of neoantigens in glioblastoma multiforme (GBM). This tumor type historically contains a high level of molecular and cellular heterogeneity along with low tumor mutational ix burden (TMB), which has been linked to lower immunotherapy response. We explore the clonality of variants, neoantigens, copy number variation (CNV), and overall immune environment by performing multi-sector sequencing of primary gliomas compared to secondary brain metastases from primary malignant breast, lung, and skin tissue. As expected, the variants called in each glioma tumor section were significantly less clonal, whereas the metastases were more homogeneous. Further evidence for molecular heterogeneity was found when the same pattern emerged in class I/II neoantigens and CNV events. Finally, we concluded that multi- sector sequencing could increase evidence for additional neoantigens in brain tumors. Chapter 3 addresses a novel variant type for neoantigen prediction: cis-splicing regulatory variants that cause tumor-specific alternative splicing (AS) events. Alternative transcript isoforms are implicated in many aspects of cancer progression, including proliferation, motility, and invasion. These variants cause changes in mature mRNA, which then become translated into proteins. Then, proteasomal cleavage creates peptides that can bind to human leukocyte antigen (HLA) alleles and form major histocompatibility complexes (MHC) on the cell surface. To increase a patient’s candidate neoantigen count for personal vaccines, we present pVACsplice, a tool to predict neoantigens based on novel tumor exon-exon junctions. The pipeline begins with a splicing regulatory variant associated with novel junction coordinates, which are then used to assemble an alternative isoform. By comparing the wild type and alternative proteins, we select all unique kmers from the alternative sequences caused by inframe indels or frameshift events. We are collecting preliminary evidence for this pipeline by validating the accuracy of assembled isoforms in long and short-read RNA sequencing data in Integrated Genomics Viewer. Overall, this research is tailored to expand the neopeptide landscape per tumor and help prioritize candidates for consideration in personalized immunotherapies.
Language
English (en)
Chair and Committee
Daniel Link
Recommended Citation
Richters, Megan, "Investigating Novel Neoantigen Sources for Personalized Cancer Vaccines" (2023). Arts & Sciences Electronic Theses and Dissertations. 3136.
https://openscholarship.wustl.edu/art_sci_etds/3136