Integration of Phospho-Signaling and Transcriptomics in Single Cells Reveals Distinct Th17 Cell Fates

Single-cell RNA sequencing (scRNA-seq) provides the resolution and scale necessary to identify transcriptional programs but fails to capture post-transcriptional information critical to decipher signaling networks and cellular states. We present Vivo-seq, a novel platform that integrates single-cell RNA-seq and intracellular CITE-seq following cellular fixation with a deep eutectic solvent, which preserves multiple domains of biological information beyond RNA transcripts alone. Vivo-seq enables simultaneous capture of both transcriptional and phospho-signaling states in single cells. Applying this platform to developing T-helper 17 (Th17) cells, we find that simultaneous phosphorylation of ERK1/2 and c-FOS leads to maximal IL-2 and IL-17 production. Furthermore, we show that early IL-2 production imprints developing Th17 cells for enhanced maintenance or cytokine-dependent transdifferentiation during subsequent antigenic stimulation. By integrating transcriptional and phospho-signaling information at single-cell resolution, we identify a hyperactivated Th17 cellular state associated with early IL-2 production that has downstream consequences on functional plasticity. Murine naive T cells were isolated from the spleen, split into 3 separate cultures, and polarized under Th17 conditions for 3 days. On day 3, the cultures were split and half were left in basal culture medium (unstimulated) and the other half were treated with PMA/ionomycin for 3 hours (stimulated). The cultures were then fixed in a deep eutectic solvent (DES; VivoPhix) and stored at -80C. On the day of single-cell RNA-seq (scRNA-seq), cultures were transitioned to aqueous buffer, as described in the manuscript, blocked, split in half, and then stained with hashtag antibodies and one of two different intracellular CITE-seq panels. 12 total unique samples were generated (n=3 unstimulated cultures + n=3 stimulated cultures x 2 panels of inCITE-seq antibodies). All samples were hashed into the same well for 10X Genomics 3’ V3 scRNA-seq.