scFv(CD19)-CD28-CD3zeta CAR-T Lentivirus

Characterizing intercellular communication and tracking its changes over time is critical to understanding the coordination of biological processes that mediate normal development, disease progression, and responses to perturbations such as therapeutics. Existing tools are unable to capture time-dependent cell-to-cell interactions and rely primarily on databases compiled from limited contexts. Here, researchers introduce DIISCO, a Bayesian framework designed to characterize the temporal dynamics of cellular interactions using single-cell RNA sequencing data from multiple time points. The approach leverages structured Gaussian process regression to reveal time-resolved interactions between different cell types based on their co-evolution, incorporating prior knowledge of receptor-ligand complexes. Researchers demonstrate the interpretability of DIISCO in both simulated data and new data collected from T cells co-cultured with lymphoma cells, demonstrating its potential to reveal dynamic cell-cell cross talk.

To demonstrate the application of DIISCO to biological data, the researchers generated single-cell data from a controlled in vitro experimental setting. Specifically, green fluorescent protein (GFP)-transduced anti-CD19 (CAR-T cells) were co-cultured with MEC1 cells, a CD19-expressing B cell (chronic lymphocytic leukemia [CLL]) leukemia cell line, as CAR-T cell targets (Figure 1A). CD19 CAR-T cells were generated by transducing healthy donor T cells with a third-generation lentiviral vector encoding a bicistronic construct containing FMC63 CD19 scFv-CD28-CD3ZETA (currently known as CD247) and either GFP or FMC63 CD19 scFv-41BB-CD3ZETA. Four biological replicates were analyzed using scRNA-seq at 10 time points over 24 hours, resulting in high-quality data for 49,283 total cells. The data showed changes in composition over time, prompting the use of this dataset as a test case for studying temporal interactions with DIISCO (Figure 1B). By clustering and examining the expression of a curated set of genes, the researchers identified four major cell types (meta-clusters): cancer cells, exhausted CD8+ T cells, activated CD8+ T cells, and other CD8+ T cells that had neither activation nor exhaustion markers. MEC1 cancer cells were annotated based on the expression of CD19 and CD79A. T cells were annotated based on the expression of CD3E, CD3D, and CD8A. Activated T cells were defined by the expression of CD69, CD27, and CD28, while exhausted cells were defined by the expression of TIGIT, PDCD1, TGFB1, and LAG3 (Figure 1C-E). Clusters that were positive for both T cell and MEC1 gene markers were removed as doublets.

Figure 1. CAR T experimental setup and data preprocessing. (Park C, et al., 2024)