Theses Doctoral

Human T-cell Negative Selection in Health and Disease

Madley, Rachel Caroline

Thymic negative selection has been identified as a crucial checkpoint in thymocyte development that purges the T-cell repertoire of autoreactive T cells through apoptosis of the cells after strong T cell receptor (TCR) stimulation. It has been well established that efficient thymic negative selection is required to prevent severe monogenic autoimmune diseases, such as Autoimmune Polyendocrinopathy-Candidiasis-Ectodermal Dystrophy. The involvement of negative selection in other T cell-mediated autoimmune diseases remains unclear. This is largely due to the lack of fully-humanized physiological models for the study of human thymic negative selection. In the work presented here, I aim to study human thymic negative selection in healthy control (HC) immune systems and to determine whether negative selection is impaired in immune systems from individuals with Type 1 diabetes (T1D), a T cell-mediated autoimmune disease.

To facilitate these studies, I developed novel humanized mouse and organ culture models. These models built on the previously described Personalized Immune (PI) mouse model [1]. The PI mouse model allows for the rederivation of a fully humanized immune system from a human donor by transplanting hematopoietic stem cells (HSC) and progenitors from an individual adult donor and a human fetal thymus fragment to immunodeficient mice. The HSCs then reconstitute all immune cell lineages, including T cells which develop in the human thymus fragment. This model is extremely powerful because it allows for the study of a human’s immune system in a model that is conducive to experimental replicates and interventions, unlike studies done directly on human patients. To optimize this model for the study of thymic negative selection, I developed two other PI models. The first is the TCR-transgenic PI thymic organ culture (TOC) model. This model allows for the study of the selection of a specific TCR in a culture system combining human HSCs and thymus fragments. The second model is the TCR-transgenic PI mouse model. This model allows for the study of the thymic selection of a specific TCR in a fully humanized in vivo model. The work presented here utilized these three powerful PI models to interrogate the thymic negative selection process in human health and disease at a depth not previously possible.

Using these models, we demonstrated the first evidence for thymic negative selection of an insulin-reactive TCR that recognizes a naturally expressed antigen in healthy human immune systems. These studies also demonstrated that robust negative selection requires HSCs expressing the HLA-restriction element of the TCR, and without the expression of that HLA on HSCs, negative selection is reduced and performed in later stages of thymic development. When comparing the phenotypic and functional characteristics of thymocytes undergoing negative selection in HC and T1D immune systems, T1D thymocytes in some immune systems had differential expression of TCR-signaling and negative selection markers and resistance to apoptosis and cell death after strong TCR stimulation. Studies on the negative selection of a specific insulin-reactive TCR in healthy and T1D immune systems demonstrated that in healthy immune systems central tolerance to this TCR involved a combination of negative selection and T regulatory cell conversion. This is the first demonstration of combined tolerogenic induction in the human immune system. In contrast, some T1D immune systems demonstrated impaired negative selection of this insulin-reactive TCR and impaired conversion of these autoreactive T cells to T regulatory cells. Further, when comparing the gene expression profile of HC and T1D thymocytes undergoing negative selection, there are multiple genes important in thymic selection and apoptosis that are differentially regulated.

Overall, this data provides unique insights into the process of thymic negative selection in healthy immune systems. It also provides the first evidence that thymic negative selection is impaired in some T1D immune systems and this impairment is possibly driven by differential gene expression. The models developed will allow for further study into human thymic selection in health and disease. These findings answer the important question of whether thymic negative selection is impaired in autoimmune disease, which has been debated in the field of T1D research and the wider immunology field. More importantly, it opens the door to targeting of the thymic negative selection pathway with therapeutics for T1D and other autoimmune diseases.


  • thumnail for Madley_columbia_0054D_16434.pdf Madley_columbia_0054D_16434.pdf application/pdf 3.91 MB Download File

More About This Work

Academic Units
Microbiology, Immunology, and Infection
Thesis Advisors
Sykes, Megan
Ph.D., Columbia University
Published Here
April 20, 2021