Mapping how small differences of tumor antigenicity translate into divergent functional outcomes for T cell activation

Speaker Details:

Gregoire Altan-Bonnet
National Institute of Health / Center for Cancer Research

Lecture Details:

April 4, 2019
12:30 p.m.
Investigator, PhD
Foege N130A, Wallace H. Coulter Seminar Room

Back to List


Recent progress in systems immunology have ushered a quantitative understanding of T cells’ ability to discriminate antigens (e.g. self vs non-self) on a short timescale (~1 hr). Yet understanding how T cell differentiation, proliferation and overall response unfold over long timescales (~1 week) is a challenge with fundamental and clinical applications. We analyzed tumor-infiltrating lymphocytes (TILs) using mass cytometry (CyTOF) and discovered that a large diversity of T cell phenotypes remains, even after weeks of IL-2-driven expansion ex vivo. We developed a new methodology (hierarchical agglomerative learning or HAL) to automatically identify clusters of differentiation within such single-cell measurements of leukocytes, and to quantify their phenotypic diversity. In turns, we found that TILs’ phenotypic diversity could be analyzed to define an immunoscore that correlates with clinical outcomes for immunotherapies based on the adoptive transfer of TILs. To model how such large phenotypic diversity emerges, we developed an ex vivo robotic platform that establishes and monitors T cell activation cultures, over long timescales. We used this system to map out the dynamics of T cell response to antigens of varied quality and quantity, across multiple readouts. We then quantitatively modeled how such high-dimensional measurements can be reduced to yield rigorous quantification of the quality and quantity of antigens T cell respond to. To conclude, we will discuss how such quantitative approaches can be applied to help clinicians monitor patients’ response and to usher new approaches in cancer immunotherapies.

Speaker Bio:

Grégoire Altan-Bonnet trained in Statistical Physics and nonlinear dynamics (PhD at the Rockefeller University) and in Immunology (post-doctoral Studies – NIAID, NIH). The ImmunoDynamics group he has been heading at Memorial Sloan-Kettering (2005-2015) and at the National Cancer Institute (2016-present) has been developing experimentally validated quantitative models of different aspects of the immune response. In particular, the group addressed the interplay between the robustness and variability of self/non-self discrimination, as well as the bridging of local and global cytokine regulations in the immune system. Altan-Bonnet & coworkers are currently focused on developing quantitative models of leukocyte-leukocyte communications within the cytokine network. Clinical applications in the field of cancer immunotherapy are being explored as well.