Date of Award


Document Type


Degree Name

Doctor of Philosophy (PhD)


Biomedical Sciences


Microbiology, Immunology, and Biochemistry

Research Advisor

Paul G. Thomas, PhD


Hongbo Chi, PhD Andrew M. Davidoff, MD Tanja A. Gruber, MD, PhD Maureen A. McGargill, PhD Mark A. Miller, PhD


CAR T cells, Lineages, Regulatory T cells, TCR, TCR Repertoire, Tregs


T lymphocytes with atypical antigen targets or mechanisms or antigen recognition have been implemented into therapeutic approaches for the treatment of disease with auspicious results. Two such populations are regulatory T cells (Tregs) and CAR (chimeric antigen receptor-modified) T cells. The functional role of Tregs in the immune system is non-canonical, in that they prevent hyper-inflammation and resolve immunopathology by constraining conventional immune cell effector functions and producing mediators that promote tissue repair. Similarly, the targets of CAR T cells are also self-derived but are recognized through an engineered receptor designed to bypass the conventional mechanism of antigen-recognition. Despite their success in the clinic, additional studies are needed to identify cellular characteristics associated with optimal treatment success. For Tregs, the role of T cell receptor (TCR) specificity against disease-associated antigens is not clear due, in part, to a lack of comparative studies of TCR repertoires across diverse disease settings, which could inform which diseases would benefit from disease antigen-specific or disease antigen non-specific Treg-based cellular therapies. If disease antigen-specific, these studies can emphasize the need to identify additional Treg-associated antigens for \textit{ex vivo} expansion. For CAR T cells, the phenotypes linked to robust effector responses have not been thoroughly assessed due to limited access to patient samples throughout the entire course of treatment. To address these gaps, we have acquired TCR sequencing data of Tregs from varied disease contexts including acute infection, allergy, cancer, and autoimmunity, followed by integrated analysis of TCR repertoire and single-cell gene expression to comprehensively identify TCR features associated with disease-responsive Tregs and assess the potential relationship between the TCR and transcriptional fate. Secondly, we acquired both pre- and post-infusion CAR T cell samples from patients participating in the SJCAR19 clinical trial to track CAR T cell clonotypes upon infusion to identify signatures associated with optimal cytotoxicity and persistence. Treg TCRs sequenced from murine models of influenza (Flu), acute asthma (AS), non-small cell lung carcinoma (NSCLC), and type 1 diabetes (T1D) showed disease-specific repertoire features. Distinct repertoire features were identified in measures of clonality, gene segment usage, and amino amino acid similarity, and physicochemical properties. Surprisingly, despite the distinct repertoire characteristics, a high degree of repertoire sharing was displayed, especially across lung conditions, though generation probabilities suggest there is an increased likelihood of utilizing similar gene segments simply by chance. Treg TCR features may influence the heterogeneity of Treg phenotypes. Focusing on Tregs isolated from the lung of influenza-infected mice at 4, 7, 10, and 23 dpi and from the lungs of NSCLC mice, single-cell gene expression suggested converging functional gene expression profiles as influenza Tregs matured compared to NSCLC Tregs, wherein both 23dpi and NSCLC Tregs showed evidence of tissue-resident and tissue-reparative transcriptional profiles. To assess whether certain TCR features were associated with Treg transcriptional heterogeneity, we applied Clonotype Neighbor Graph Analysis (CoNGA) to identify clusters of cells based on simultaneous similarity in gene expression landscape and TCR amino acid sequence. CoNGA analysis identified 3 distinct Treg-associated gene expression clusters with high TCR similarity, suggesting that transcriptional heterogeneity in the Treg population is not driven by differences in TCR features, and likely, TCR specificity. Lastly, single-cell gene expression of pre- and post-infusion CAR T cells identified distinct transcriptional landscapes with pre-infusion CAR T cells displaying signatures of proliferation and post-infusion CAR T cells having largely acquired either a cytotoxic effector signature associated with expression granzymes and effector cytokines or a dysfunctional signature associated with cellular death and T cell exhaustion. Having identified CAR T cells with optimal effector responses, single-cell TCR data was utilized to identify pre-infusion cells clonally related to functional effector post-infusion CAR T cells, deemed effector precursors. Comparison of effector precursor gene expression to all other pre-infusion CAR T cells demonstrated higher expression of granzymes, \textit{GNLY}, and \textit{IFNG}, suggesting effector precursors were primed for optimal cytotoxic responses upon infusion. Using differential expression of genes encoding surface proteins, we demonstrated that effector precursors can be enriched in the pre-infusion sample to increase the proportion of ideal CAR T cells patients receive upon infusion. Taken together, these findings provide insight into TCR repertoires and subsequent functional association of these unconventional T cell populations, which will assist in the improvement of treatment strategies using Tregs or CAR T cells. Moreover, these data can be applied in the development of other T cell-based therapies, such as CAR-Tregs, by informing the role of the endogenous TCR in conditioning CAR-Treg phenotypes in distinct tissues and providing a novel method of leveraging the TCR to elucidate transcriptional signatures and surface phenotypes leading to optimal CAR-Treg responses.

Declaration of Authorship

Declaration of Authorship is included in the supplemental files.




2022-014-Wilson-DOA.pdf (152 kB)
Declaration of Authorship

Available for download on Saturday, July 13, 2024