CAREER: Origins of Cellular Heterogeneity by Dynamically-linked Multiplexed Imaging

About This Grant

T cells are a type of white blood cell that plays an important role in the functioning of the immune system. Therapies based on T cells (CAR T-cell therapy) are proving to be effective for certain cancers. T cells are isolated from a patient’s blood and modified to express a binding protein (CAR) that allows the T-cell to more effectively attach to tumor cells. Once modified, the CAR T-cells are grown to create millions of them, and then introduced into the patient, where they bind to cancer cells and kill them. The weak link in this process is the growth step. As the cells grow in number, they create inexact copies of the original cell, so ultimately there is a broad distribution of effectiveness in the resulting CAR T-cells. This reduces the effectiveness of the treatment. This project will attempt to track the development of these growing populations and determine what causes this variability. In parallel, the PI will be developing hands-on lessons and computational tutorials on light microscopy for high school and college-level communities. This should enhance STEM education and broaden the understanding of single-cell biological engineering. This project aims to decode how single-cell phenotypic heterogeneity arises and can be precisely controlled by linking dynamic cellular behaviors to deep molecular phenotyping. Currently, a major limitation is the inability to observe how a cell’s molecular phenotype relates to the multimodal signals and dynamic paths it takes through time and space. To address this challenge, the research leverages three cutting-edge technologies: (1) spatial proteomic multiplexed imaging; (2) in vitro cell tracking; and (3) engineered biomaterial artificial antigen-presenting cells (aAPCs) that deliver controlled stimulatory signals to T cells. The scope of the project is centered on three main objectives. First, it will image and computationally extract dynamic cell interactions to link these events directly to downstream T cell phenotypes. Second, it will establish a barcoded, dynamic, multi-input T cell stimulation platform and decode the complete stimulation histories of individual cells as they acquire their final phenotypes. Third, it will use these integrated datasets—encompassing dynamic behavior, molecular signatures, and multiplexed stimulation conditions—to learn and predict how T cell phenotypes emerge from complex, time-varying inputs. To accomplish these goals, the project employs existing and newly developed computational workflows, including segmentation, unsupervised clustering, machine learning, and data-driven multiscale modeling. By constructing a powerful, decodable perturbation screening platform, the research will uncover how distinct combinations and sequences of stimulation events shape T cell fate. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.