Violin plots a show IRF family member expression levels ranging from approximately 0.75 to 2 across eight subsets including Tex, Tpex, SLEC, Eff Interm, Mem Prec, Eff Early, and Eff Cycling, where Irf8 shows notably wider violins peaking around 2 in Tex and Tpex populations, while other family members show narrower distributions below 1.5. Schematic b illustrates three experimental timelines where LCMV Cl13 and B16-gp33 models receive naive P14 cells at day negative 1 with infection or engraftment at day 0, while the B16-gp33 preactivated model receives tumor at day negative 7 and preactivated P14 cells at day 0, with all groups analyzed by FACS at day 21. Grouped bar graphs c show PD-1 on a gMFI times 10 to the power of 4 y-axis with LCMV Cl13 values around 2, B16-gp33 values reaching approximately 4, and B16-gp33 preactivated values around 2 to 3 with a significant difference marked between LCMV Cl13 and B16-gp33, while TOX values on a gMFI times 10 to the power of 3 y-axis range between 5 and 15 across all three conditions with no significant differences. Contour plots d and f show PD-1 versus TOX and IRF8 versus TOX distributions respectively, where LCMV Cl13 shows compact low-value populations, B16-gp33 shows broader distributions extending along both axes, and B16-gp33 preactivated shows populations shifting toward intermediate to higher values. Scatter plot e on a gMFI times 10 to the power of 3 y-axis up to 25 shows spleen data points across all conditions clustering below 5 while tumor data points for B16-gp33 range from 10 to 18 and B16-gp33 preactivated reach 13 to 20, with significant differences marked for both tumor comparisons. Contour plots g show SLAMF6 versus TIM3 with TPEX and TEX gates, where LCMV Cl13 shows a prominent TPEX cluster, B16-gp33 shows larger TEX accumulation, and B16-gp33 preactivated shows a condensed TPEX alongside a prominent TEX population. Paired line graph h on a gMFI times 10 to the power of 3 y-axis up to 30 shows LCMV Cl13 TPEX and TEX lines remaining flat below 5, B16-gp33 lines rising from 2 to 5 up to 10 to 20, and B16-gp33 preactivated lines showing the steepest rise reaching up to 25, with accompanying histograms confirming rightward shifted IRF8 distributions in TEX compared to TPEX populations.
Exhausted CD8 + T cells specifically express IRF8 in cancer. (a) Expression of IRF TF family members in virus-specific CD8 T cells based on publicly available scRNA-seq datasets from different LCMV infection studies (Material and methods). (b) Schematic representation of experimental procedure used to compare IRF8 levels in tumor- versus virus-specific exhausted CD8+ T cells. In the LCMV Cl13 chronic infection (top), 103 naive P14 CD8+ T cells have been transferred into Vβ5 mice 1 day prior to infection with LCMV Cl13. In the B16-gp33 condition (middle), 2 × 106 naive P14 CD8+ T cells have been transferred 1 day prior to B16-gp33 tumor engraftment (0.5 × 105 cells s.c.) in B6 mice. In the B16-gp33 (preact.) condition (bottom), B16-gp33 tumors where engrafted in B6 mice 7 days prior to adoptive transfer of 106 P14 CD8+ T cells preactivated in vitro during 48 h. Readout was done 21 days after in vivo antigen encounter. (c) Flow cytometric quantification of PD-1 (left) and TOX (right) in the indicated conditions. (d) Representative FACS plots for c showing TOX and PD-1 expression in the indicated population. (e) IRF8 protein quantification in transferred P14 CD8+CD44+PD-1+ CD8+ T cells in spleens from the three conditions in b and tumors from B16-gp33 and B16-gp33 (preact.). (f) Representative FACS plots for c showing TOX and IRF8 expression in the indicated population. (g) Representative FACS plots showing gating strategy for SLAMF6+TIM3− TPEX and SLAMF6−TIM3+ TEX among P14 CD8+CD44highPD-1+ cells in indicated populations. (h) IRF8 quantification in TPEX and TEX in the indicated condition (left) and representative FACS histograms (right). In b–h, pooled data are from two independent experiments with n = 6 for LCMV Cl13, n = 7 for B16-gp33, and n = 6 for B16-gp33 (preact.). In c, statistical analysis was done with ordinary two-way ANOVA. In e, statistical analysis was done with ordinary two-way ANOVA comparing each column with the LCMV Cl13 condition. In h, statistical analysis was done by multiple paired two-tailed Student’s t tests. Data in c and e are represented as mean with SD. *P < 0.05, **P < 0.01, and ***P < 0.001.