208. Single-Cell Gene Expression Profiles of Genetically Modified T Cells for Adoptive Immunotherapy

Abstract

System-level characterization of cell states requires detailed portrayal of cellular components such as gene transcripts or proteins. Here, we exploit a robust single-cell platform technology to characterize T cells expanded ex vivo intended for human application. Coupled with advanced custom analytics, we identify “drivers” (relatively few, but vital) genes which orchestrate T-cell activation and clonal expansion programming, versus “passengers” (relatively many, less critical) or resultant/downstream transcripts.Detailed analyses reveal that there is considerable variability between individual activated T cells, implying that cell state may undergo multiple stepwise transitions in a stochastic manner. Compared to interrogation of bulk T-cell populations (e.g., analyzed using RNA-Seq, microarrays, and Northern blotting), where heterogeneity in expression signals is attenuated by temporal and cell-cycle averaging, the genetic heterogeneity between cells in isogenic populations may be a reflection of fundamental phenomena such as transcriptional bursting. Our model describes cell state oscillation in tandem with transcriptional repressive (closed chromatin) and permissive conformations (open chromatin), versus simpler probabilistic models of transcription. T-cell variability can significantly impact upon individual cell behavior within seemly isogenic populations, and may be essential in molding survival and cytotoxicity within certain contexts, such as the rapidly changing and stressful tumor microenvironments. We find that lactate dehydrogenase A (LDHA) is a leading indicator for the metabolic adaptation during T-cell activation and together with other energetics related genes such pyruvate kinase isoform 2 (PKM2), can be harnessed to adjust our culture conditions during ex vivo expansion. These cellular fingerprints suggest useful ways for tuning the balance between oxidative and non-oxidative glucose metabolism to enhance the killing and thus therapeutic potential of our T cells. Our framework to harness catalogs of gene expression data into clinically-applicable insights can be adapted for meaningful characterization of other therapeutically-relevant source populations such as hematopoietic stem cells and reprogrammed NK cells. This is being reduced to practice by understanding the heterogeneity within and between genetically modified T-cell products generated for clinical use.