Abstract 2589: Novel workflow to generate natural killer cell sensitivity prediction model based on the epithelial mesenchymal transition related reprogramming at a single cell resolution

Abstract

Abstract Single cell technology allows us to examine cell heterogeneity and identify rare cell populations; however, the high cost of single cell RNAseq limits its adoption into clinical and research environments. As an inexpensive alternative, immunofluorescent assays analyzing cell morphologies have been used in high content screening to examine cell phenotype and response to external stresses. Natural Killer cells (NK) are the bodies first line of defense and gained traction as a potential target in treating cancer. A study showed that cancer cells undergoing the process epithelial mesenchymal transition (EMT) increase their vulnerability to NK mediated cytotoxicity. We hypothesize that the EM morphology change can be correlated with NK sensitivity. Our goal is to develop an image-based, cost-effective, and efficient workflow capable of extracting cancer cell morphological features to predict its NK sensitivity at the single-cell level. To optimize this workflow, we use A549 cell line cells treated with transforming growth factor- β (TGF-β) as a representative models of cells with various EM phenotypes as NK sensitivities. We utilized a Cell Painting inspired novel image based live cell cytotoxicity assay. To capture cell morphological features, we optimized an organelle specific and live cell compatible staining panel that targets the nucleus, nucleic acid, cell membrane, F-actin, and mitochondria. The cell viability is assessed by Propidium Iodide, which stains dead cell nuclei. Our image analysis pipeline can extract the morphological features and track the cells of interest during the experiment and assess their viability. Subsequently, the morphology and viability data will be processed by a classification algorithm that could discern morphological features of cancer cells that indicate high NK vulnerability. We started with treating A549 cells with 5ng/mL TGF-β for 0, 24, 48, and 72 hours, and successfully applied the organelle staining panel. We observed distinct cell morphologies at each timepoint of TGF-β treatment, with cells shifting from a cobblestone (no TGF-β) to a spindle-like (72 hours TGF-β) morphology. Notably, the F-actin signal was significantly overexpressed in the 72 hours TGF-β treated A549 cells. A CellProfiler based preliminary image analysis pipeline was established for cell segmentation, cell features extraction, and cell viability tracking with quality control filters. Further optimization for cell segmentation, viability tracking, and classification algorithms will be needed to enhance the robustness of the workflow. Our established workflow demonstrates the feasibility of live cell functional assay with organelle staining. This workflow could be adopted to correlate and predict single tumor cell susceptibility to cytotoxic immune cells, with the goal of developing a therapeutic index. Citation Format: Yin Zou, Harrison Ball, Cade Harris, Nithya Ramnath, Venkateshwar Keshamouni, Sunitha Nagrath. Novel workflow to generate natural killer cell sensitivity prediction model based on the epithelial mesenchymal transition related reprogramming at a single cell resolution [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 2589.