Author's School

School of Medicine

ORCID

https://orcid.org/0000-0002-3855-7950

Author's Department/Program

Biology and Biomedical Sciences

Concentration/Secondary Program

Molecular Genetics and Genomics

Language

English (en)

Date of Award

12-15-2026

Degree Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Chair and Committee

Malachi Griffith (Chair); Obi L. Griffith (Co-Chair)

Committee Members

Todd A. Fehnigher, Jennifer A. Foltz, William E. Gillanders, Robert D. Schreiber

Abstract

Precision oncology aims to match cancer treatment to the biology of an individual patient’s tumor. However, tumors that share a diagnosis can differ in their molecular features, cellular composition, clinical course, and response to therapy, making it difficult to determine which treatment will benefit an individual patient. This dissertation integrates genomic, transcriptomic, and single-cell multiomic computational analysis to characterize tumors across their somatic alterations, transcriptional cell states, microenvironment, and the immune response they elicit.

The first study describes a rare case of donor-derived follicular lymphoma following kidney transplantation. Genomic profiling showed that the donor and recipient tumors shared 31 somatic mutations, strongly supporting a common origin. However, while the donor was diagnosed only two years after donating the kidney, disease did not emerge in the recipient until 12 years after transplantation. This demonstrates how the same ancestral clone can follow markedly different evolutionary timelines in different hosts.

The next chapter uses single-nucleus profiling to resolve heterogeneity in large B-cell lymphoma. Analysis of 232 tumor and control biopsies identified three recurrent lymphoma microenvironment archetypes: FMAC, characterized by cancer-associated fibroblasts and tumor-associated macrophages; LN, characterized by a lymph node architecture with naïve and memory T cells; and TEX, characterized by exhausted CD8+ T cells and activated macrophages. These archetypes were associated with distinct clinical outcomes following CD19-directed chimeric antigen receptor T-cell therapy. Complementary analysis of malignant B cells in the subsequent chapter identifies five recurrent transcriptional topics. Of the five topics, one recapitulated activated B-cell biology, one recapitulated germinal-center B-cell biology, two resolved the dark-zone signature into states with distinct prognostic associations, and one identified a plastic precursor-like population present across most tumors. Together, these studies show how malignant-cell state and microenvironmental context can refine patient stratification beyond existing genomic classifications and reveal potential therapeutic vulnerabilities.

The dissertation next examines approaches for manipulating antitumor immunity. In a phase I trial involving patients with newly diagnosed high-grade glioma, the long-acting interleukin-7 agonist NT-I7 increased absolute lymphocyte counts for over 12 weeks. Single-cell transcriptomics and immune profiling showed that the treatment produced early increases in CD4+ T cells, CD8+ T cells, and natural killer cells, accompanied by immune activation and selective expansion of CD8+ T-cell clonotypes.

The next chapter describes the development of an end-to-end protocol for personalized neoantigen vaccine design that comprises two parts. The first is a computational component, ImmunoNX, which integrates tumor DNA and RNA with matched-normal DNA to perform quality control, somatic variant calling, HLA typing, expression analysis, and neoantigen prediction. The second part turns those predictions into a vaccine, in which the candidates are prioritized through expert-guided review to select the neoantigens that are ultimately included. This workflow has supported more than 184 patients across 11 clinical trials.

Finally, the dissertation turns to the question of which of the predicted neoantigens used in vaccines actually elicit a T cell response. This information is needed both to determine whether the vaccine generated an immune response, and also to refine the computational prediction pipelines based on neoantigen immunogenicity. This chapter uses complementary tetramer-based and bulk and single-cell T cell receptor sequencing-based approaches to identify neoantigen-specific T cells.

Together, this dissertation presents a framework for personalizing cancer treatment across multiple levels: stratifying patients according to tumor-intrinsic and microenvironment states and further identifying patient-specific targets for immunotherapy.

DOI

https://doi.org/10.48765/tvks-b829

Download

Share

COinS