Runx transcription factors regulate early T-developmental progression by inducing T-lineage identity and innate lymphoid cell program genes via gene network regulation

Runx transcription factors are essential for generating functional T cells starting from the earliest stages in thymic T-development. To evaluate how Runx factors instruct T cell development, we perturbed Runx activities using 1) CRISPR-Cas9, simultaneously deleting two predominant, redundant Runx factors, Runx1 and Runx3, and 2) Runx1-overepxression at a stage before pro-T cells commit to a T cell fate. Then single-cell transcriptome analysis was performed. Runx1/Runx3 knockout (KO) vs. Runx1 overexpression (OE) resulted in contrasting deviations from the normal developmental trajectory. The separation of both perturbation conditions from the controls implied that Runx factors regulate gene networks in individual cells to control separate pathways instead of changing distributions of cells among different stages along the control pathway. Notably, a common core set of genes responded to both KO and OE, in opposite directions. Runx activation target genes were selectively enriched for T-identity program and innate lymphoid cell program genes, whereas inhibition targets were significantly associated with stem/progenitor program and myeloid lineage pathways. Pseudotime inference analysis suggested that Runx perturbations also substantially regulate developmental speed, as Runx KO significantly slowed down developmental progression, whereas Runx1 OE resulted in pronounced acceleration. This interpretation was largely supported by Runx activities regulating key transcription factors involved in establishing distinct gene network modules during T cell development. Indeed, further analysis showed that Runx target genes get combinatorial inputs from these Runx-regulated transcription factors as well as from Runx factors themselves. Together, our data suggest that Runx transcription factors drive early T developmental progression by launching gene expression modules associated with T-identity and innate lymphoid cells through mediating other transcription factor cooperativities. To recapitulate early T cell development in a perturbation-accessible way, the OP9-Dll1 in vitro culture system was exploited. For Runx1 overexpression conditions, freshly obtained bone-marrow-derived progenitor cells from 8-10 week old B6.Bcl2 transgenic mice were cultured on OP9-Dll1 cells with 10ng/ml of IL-7 and Flt3-ligand. On day 2, MSCV-Runx1-mCherry or MSCV-empty-mCherry vectors were delivered using spinfection. Two days after delivering retrovirus (total 4 days of culture) or 4 days post-infection (total 6 days of culture), each of the experimental replicates was stained with a distinct hashtag oligo and live Lineage marker-negative, mCherry+, CD45+ cells were FACS sorted for transcriptome analysis. For Runx1/Runx3 double knockout, slightly more time was allowed for the disruption of the targeted genes to reduce protein levels before gene expression was measured. Bone-marrow-derived progenitor cells from B6.Cas9;Bcl2 transgenic mice were co-cultured with OP9-Dll1 cells in medium supplemented with 10ng/ml each of IL-7 and Flt3-ligand. On day 2, E42-gRunx1-humanNGFR (hNGFR) and E42-gRunx3-mTurquoise vectors were retrovirally introduced. Three gRNA pairs were utilized for each Runx1 and Runx3 targeting. For control conditions, E42-gControl-humanNGFR (hNGFR) and E42-gControl-mTurquoise vectors were utilized, where control gRNA targets luciferase sequences. At 3 days post-infection (total 5 days of culture) or 6 days post-infection (total 8 days of culture), each of the experimental replicates was stained with a different hashtag oligo and live Lineage-, mTurquoise+, hNGFR+, CD45+ cells were FACS sorted. Equal numbers of FACS sorted cells from each of the experimental replicates from Runx1 OE and Runx1/Runx3 double KOs and controls were pooled and subjected to single-cell RNA-seq using 10X Chromium v3.