Flow Cytometry Gating Strategy and t-SNE analysis
Phenotypes were defined using a combination of expression of surface markers with intracellular cytokine expression after stimulation with PMA/ionomycin (versus relevant negative and isotype controls). Representative flow cytometric data from second generation bank EBV-specific T cell final product with initial gating of lymphocytes/ singlets/ live cells is outlined in Supplementary Figure 1. Lineage was identified through CD4 versus CD8 labelling, and differentiation status through surface expression of CD45RO/ CD45RA and CD62L or CCR7 (Supplementary Fig 1B). Analysis of cytokine expression (Supplementary Fig 1C) was based on gating CD8+/CD45RO+ viable lymphocytes, dividing into IL-2low and IL-2highpopulations, and then quadrant gating each IL-2 subpopulation for IFN-γ and TNF-α co-expression. A contrasting profile of T cell cytokine expression of a heterogeneous pan T cell compartment in PBMC freshly isolated from buffy coat donors is shown in Supplementary Figure 2.
Stochastic Neighbour Embedding (t-SNE) analysis was applied to multi-parameter flow cytometric data. Briefly, analyses were gated on lymphocyte/singlet/live as above using FlowJo, and reduced to a representative 10,000 events using Downsample. The t-SNE maps were generated using the FlowJo plugin for 800 iterations at perplexity 20. Events from each Downsample population were spatially correlated in terms of likeness for all fluorescent parameters outlined previously. Manual gating overlays were used to subdivide CD8+ and CD4+ cells into T cell memory populations based on surface marker expression: naive (CCR7+/CD45RO-), central memory TCM (CCR7+/CD45RO+), effector memory TEM (CCR7-/CD45RO+) and terminally-differentiated effectors expressing CD45RA TEMRA (CCR7-/CD45RO-).