Fig 2: Association of CDK9 expression with tumor-infiltrating CD8+ T cell exhaustion in colorectal cancer. The frequencies of PD-1+CD8+ T cells, Tim-3+CD8+ T cells and CD39+CD8+ T cells were determined by flow cytometry. Each dot presents data generated from one patient (n=35). *P<0.05; **P<0.001; ***P<0.0001. CDK9, cyclin-dependent kinase 9; PD-1, programmed cell death 1; Tim-3, T-cell immunoglobulin mucin family member 3; N, paired normal tissue; CDK9Low, low expression of cyclin-dependent kinase 9; CDK9High, high expression of cyclin-dependent kinase 9; ns, not significant; T, tumor tissue.

Fig 3: Overexpression of Pla2g4a in CT26-induced CD39+?d Tregs in vitro.CT26-Pla2g4a and CT26-Vec cells were cocultured with intraepithelial lymphocytes derived from the ileum (IL-IELs), colon (CO-IELs), or spleen of BALB/c mice. (A and C) Representative pictures showed the number of CT26 cells in the coculture system after 12 (A) and 36 hours (C). Scale bars: 200 µm. (B and D) Bar diagrams summarizes the cell numbers of CT26-Pla2g4a and CT26-Vec cells cocultured with IL-IELs after 12 (B) and 36 (D) hours. Data are shown as mean ± SEM, using 2-way ANOVA followed by Sidak’s multiple-comparisons test. (E–G) Representative histogram plot and bar diagram summarizes the expression of CD39 on ?d T cells in IL-IELs (E), CO-IELs (F) and splenocytes (G) group. (H) Representative flow cytometric analysis of CD39+ Tregs in the total CD3+ T cells from splenocytes. Bar diagram summarizes the percentages of CD39+?d Tregs. Unpaired 2-tailed t test was used to compare variables. (I) Silencing Pla2g4a in CT26-Pla2g4a rescued the antitumor response mediated by ?d T cells. CT26-Pla2g4a were transfected with Pla2g4a siRNA or control siRNA. After 24 hours, CT26-Vec, CT26-Pla2g4a, and CT26-Pla2g4a–knock-down cells were cocultured with ?d T cells sorted from BALB/c mice with or without anti-PLA2G4A antibody (5 mg/mL). Scale bars: 200 µm. (J) Bar diagram summarizes the cell number of CT26. (K) Representative histogram plot of CD39 expression on ?d T cells in each group by flow cytometry. Bar diagram summarizes the frequency of CD39+ ?d Tregs/?d T cells. One-way ANOVA followed by Tukey’s multiple-comparison test was used for multiple-group analysis. All of the experiments were repeated 3 times; n = 3/group. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

Fig 4: Differentiation phenotype of CD39+?d Tregs in RSCRC/LSCRC.(A) Representative histogram plot of the expression of IL-17A/IFN-?/PD-1/Tim-3 in/on CD39+ Tregs derived from LSCRC/RSCRC. (B–E) The percentages of IL-17A (B), IFN-? (C), PD-1 (D), and Tim-3 (E) in CD39+?d T cells was calculated. Data are shown as mean ± SEM, using 2-way ANOVA followed by Sidak’s multiple-comparisons test. RSCRC, n = 5; LSCRC, n = 6. Each dot represents 1 patient. N, normal; T, tumor; RSCRC, right-sided colorectal cancer; LSCRC, left-sided colorectal cancer. *P < 0.05, **P < 0.01, ****P < 0.0001.

Fig 5: The infiltration of CD39+?d Tregs was significantly increased in RSCRC compared with LSCRC.(A) Representative flow cytometric analysis of the distribution of CD39+ Tregs in CRC and paired normal tissues. (B–D) The percentages of CD39+ ?d Tregs (B), CD4+ T cells (C), and CD8+ T cells (D) in the total tumor-infiltrating CD3+ T cells was calculated. Data are shown as mean ± SEM, using 2-way ANOVA followed by Sidak’s multiple-comparisons test. RSCRC, n = 15; LSCRC, n = 20. Each dot represents 1 patient. N, normal; T, tumor; RSCRC, right-sided colorectal cancer; LSCRC, left-sided colorectal cancer. **P < 0.01, ***P < 0.001.