Schematic a shows B16-gp33 tumor bearing B6.Cas9 mice receiving P14 control and P14 IRF knockout cells at day 7 with FACS at day 14. Scatter plot b shows CD90.1 positive output to input ratios declining significantly below 1 for IRF2 knockout, IRF4 knockout, and IRF8 knockout compared to their respective controls, all with significant differences marked. Paired line graphs c show TPEX and TEX percentages among PD-1 positive cells across IRF2, IRF4, and IRF8 knockouts, where TPEX shows significant decreases for IRF2 and IRF8 knockouts and TEX shows a significant increase for IRF8 knockout. Paired line graphs d show TOX geometric mean fluorescence intensity times 10 to the power of 3 declining significantly across all three knockouts and percentage PD-1 positive cells remaining largely unchanged. Paired line graphs e show percentage IFN gamma positive rising significantly for IRF2 and IRF8 knockouts, TNF alpha positive rising significantly for IRF8 knockout only, and Granzyme B positive rising significantly across all three knockouts with representative data shown. Schematic f shows the same experimental setup leading to single cell RNA sequencing at day 14, with violin plots showing expression levels of key genes including Lef1, Tcf7, Tox, Nfatc1, Gzmk, Ifng, and others across four clusters defined as C1 Tex, C2 Tex Cycling, C3 Tpex, and C4 ToxLow EffLike. UMAP plots g show integrated and batch corrected clustering for IRF2, IRF4, and IRF8 control and knockout pairs with sample sizes of 4, 5, and 7 respectively, alongside bar graphs showing cluster frequency distributions where C1 shows a significant difference in IRF2 knockout and cluster proportions shift variably across knockout conditions. Gene set enrichment analysis plots h show effector versus exhausted CD8 T cell upregulation scores for IRF8 knockout, IRF4 knockout, and IRF2 knockout versus control.
Deletion of IRF2, IRF4, and IRF8 attenuates exhaustion in tumor-specific CD8 + T cells. (a) Schematic representation of experimental procedure used to assess effects of IRF2, IRF4, and IRF8 KO in P14 CD8+ TILs from B16-gp33–bearing B6-Cas9 mice. (b) Recovered (output) versus injected (input) CD90.1+ control (c) or KO (ko) CD8+ T cells from co-transfer experiments of IRF2, 4, or 8 KO with control-transduced P14+ CD8 T cells represented as ratio output/input 7 days after co-transfer. Data were pooled from three independent experiments (n = 13 for IRF2 KO, n = 12 for IRF4 KO, and n = 14 for IRF8 KO). (c) Percentages of TPEX and TEX populations in IRF KO and control-transduced cells. (d and e) TOX gMFI and PD-1+ percentage (d) and percentage of IFNγ+, TNFα+, and GrzB+ (e) after restimulation with gp33 peptide and of P14+CD90.1+ control and IRF KO CD8 T cells from co-transfer experiment illustrated in a. Representative data from three independent experiments (n = 6 for IRF2 KO and n = 5 for IRF4 and IRF8 KO). (f) Schematic representation of experimental procedure used to assess transcriptional changes of IRF2, IRF4, and IRF8 KO in P14 CD8+ TILs from B16-gp33–bearing B6-Cas9 mice (top) and violin plots displaying expression levels of markers spanning progenitor, exhaustion, effector, and proliferative programs in the four clusters defined in g. (g) UMAP plots defining clustering after integration and batch correction for each P14 IRF KO and their respective co-transferred control CD8 TILs derived from scRNA-seq data and clustered using Seurat, with quantification for each of the four clusters in control versus IRF KO. n = 4 for IRF2 KO, n = 5 for IRF4 KO, and n = 7 for IRF8 KO. (h) GSEA of pseudobulk from scRNA-seq for the indicated IRF KO versus control using gene set effector versus exhausted CD8 T cells UP. GEO accession: GSE9650. Adjusted P value (adj. P val) and normalized enrichment score (NES) are shown for each comparison. In b, statistical analysis was done by two-tailed Student’s t test. In c–e and g, statistical analysis was done with paired two-tailed Student’s t test. *P < 0.05, **P < 0.01, and ***P < 0.001.