|Ph.D Student||Oved Kfir|
|Subject||T-Cells Regulation: From a single cell to a Systems View|
of Heterogeneous Populations
|Department||Department of Biology||Supervisor||Professor Yoram Reiter|
|Full Thesis text|
Cytotoxic T lymphocytes (CTLs) are central effector cells in host immunity. In addition to their role in pathogen clearance, CTLs are also capable of locating and eliminating tumor cells, yet, they frequently fail to do so. Numerous molecular mechanisms enable tumors to induce immune ignorance. The first part of this work describes a novel approach that direct CTLs against tumor cells, regardless of proper antigen presentation. We designed a chimerical protein that is composed of an MHC-peptide complex fused into an antibody fragment (pep-sc-HLA-scFv). The modular nature of these molecules enables one to use any antibody, HLA or peptide of interest. These molecules were shown to mediate a specific and effective tumor lysis. Applying this approach by our laboratory in animal models demonstrated an effective anti-tumor activity in-vivo. This approach may open the way for the development of new immunotherapeutic strategies.
In the second part of our work, we characterize the antigen-density dependent activity (ADDA) of CTLs. ADDA of CTLs is usually considered as an increasing monotonic function. We demonstrate that the ADDA of CTLs is a non-monotonic function with optimal activity at~100 MHC-peptide complexes on the surface of the target cell. Higher antigen densities, termed "supra optimal", led to a long lasting suppression of CTLs activity including profound alterations of effector functions, cytokine secretion, and gene expression.
The third and last part of our work deals with the immune system complexity. The immune system is built up of multiple cell types and soluble factors that form a network and can respond fast, accurately and effectively against pathogens or tissue damage. A major question is whether simple rules can be applied in order to predict and control this complex system. As a model system we used a poorly understood heterogeneous population termed tumor infiltrating T-lymphocytes (TILs). We measured the frequencies of multiple subpopulations and examined the relations between subpopulation composition and TIL function. We show that no individual subpopulation can accurately predict TIL response to cancer cells. However, by using a simple computational model we could generated a set of rules that accurately (>90%) predict TIL reactivity based on the subpopulation composition. Guided by these rules we were able to turn non-reactive TILs into reactive ones. This novel approach paves the way for predicting and controlling the activity of other heterogeneous populations such as bacteria, stem cells, engineered tissues etc.