Phenotypic switch of CD8(+) T cells reactivated under hypoxia toward IL-10 secreting, poorly proliferative effector cells

Abstract

CD8(+) T cells controlling pathogens or tumors must function at sites where oxygen tension is frequently low, and never as high as under atmospheric culture conditions. However, T-cell function in vivo is generally analyzed indirectly, or is extrapolated from in vitro studies under nonphysiologic oxygen tensions. In this study, we delineate the role of physiologic and pathologic oxygen tension in vitro during reactivation and differentiation of tumor-specific CD8(+) T cells. Using CD8(+) T cells from pmel-1 mice, we observed that the generation of CTLs under 5% O2, which corresponds to physioxia in lymph nodes, gave rise to a higher effector signature than those generated under atmospheric oxygen fractions (21% O2). Hypoxia (1% O2) did not modify cytotoxicity, but decreasing O2 tensions during CTL and CD8(+) tumor-infiltrating lymphocyte reactivation dose-dependently decreased proliferation, induced secretion of the immunosuppressive cytokine IL-10, and upregulated the expression of CD137 (4-1BB) and CD25. Overall, our data indicate that oxygen tension is a key regulator of CD8(+) T-cell function and fate and suggest that IL-10 release may be an unanticipated component of CD8(+) T cell-mediated immune responses in most in vivo microenvironments.

Keywords: CD8+ T cell; Hypoxia; IL-10; Oxygen; T-cell reactivation.

Figure 1

CTLs generated under physioxia have a higher effector profile than those generated under atmospheric oxygen fraction. (A and B) CTLs from Pmel‐1 splenocytes were generated under 21% (black bars) or 5% O₂ (gray bars) and were analyzed for RNA expression. Results represent (A) the mean relative gene expression or (B) the RNA absolute count + SEM of five independent experiments (n = 5). The checkerboard represents the 42 genes analyzed (only genes modulated by more than 30% with a p < 0.05 are colored in red or blue). (C and D) CTLs from Pmel‐1 splenocytes were generated under O₂ fractions as indicated and were analyzed for marker expression by flow cytometry. (C) CTLs were stained for CD62L and CD44, gated on live CD8⁺ cells, at day 0, day 3, and day 6 postpriming. Dot plots show one representative experiment at day. Line graph shows the mean percentage ± SEM CD44⁺ CD62L^‐ cells among CTLs from at least four independent experiments (n ≥ 4). (D) CTLs generated under 21 and 5% were preconditioned for 3 days under 21 (“21 to 21%”), 5 (“5 to 5%”) or 1% (“21 to 1%” and “5 to 1%”) O₂. The resulting CTLs were assessed for their capacity to kill EL‐4 tumor target cells pulsed with hgp100 (Pmel‐1 CTLs; n ≥ 4) or Ova peptide (OT‐I CTLs; n = 3) for 4 h under the oxygen fraction used for preconditioning. Results show the mean ± SEM specific lysis of tumor cells from at least three independent experiments. ns: not statistically significant, *p < 0.05, **p < 0.01, ***p < 0.001 (A, B: Three‐way ANOVA; C, D: Student's t‐test).

Figure 2

Hypoxia modulates expansion and RNA profile of reactivated CTLs. CTLs generated under 21% (squares) or 5% (circles) O₂ from Pmel‐1 splenocytes were reactivated for indicated times under 21% O₂ (closed squares with solid line), 5% O₂ (closed circles with solid line) or 1% O₂ (open squares or open circles with dashed line). Results show (A) mean cell number, (B) mean cell division number, (C) mean cell viability, and (D) mean apoptotic cells ± SEM out of at least three independent experiments (n ≥ 3). CTLs generated under (E) 21% or (F) 5% O₂ from Pmel‐1 splenocytes were reactivated for two days under indicated oxygen fractions. (E) RNA profile from CTLs reactivated under 21% (black bars) or 1% O₂ (gray bars). (F) RNA profile from CTLs reactivated under 5% (black histograms) or 1% O₂ (gray histograms). Results represent the mean relative gene expression + SEM out of four independent experiments (n = 4). The checkerboard represents the 42 genes analyzed (only genes modulated by more than 30% with a p < 0.05 are colored in red or blue). To display common genes modulated under each condition, genes composing the checkerboard are organized identically (in an arbitrary fashion). ns: not statistically significant, *p < 0.05, **p < 0.01, *** p <0.001 (A–D: Student's t‐test; E, F: Three‐way ANOVA).

Figure 3

Decreasing oxygen fraction promotes IL‐10 production by reactivated CTLs. CTLs were generated under 21% O₂ and reactivated under 21% (21 to 21%) or 1% O₂ (21 to 1%), or were generated under 5% O₂ and reactivated under 5% (5 to 5%) or 1% (5 to 1%) O₂. Four days post‐reactivation cell surface expression of (A) CD137 or (B) CD25 was assessed by flow cytometry. (A and B) Open histograms: control isotype staining; closed histograms: staining for the molecule of interest. Line graphs: median fluorescence intensity from three independent experiments (n = 3). (C) Dot plots: IL‐10 and IFN‐γ intracellular staining from CTLs reactivated for two days. (D) Percentage of IL‐10⁺ or IFN‐γ⁺ CTLs from CTLs generated under 21% (squares) or 5% O₂ (circles) that were reactivated under 21% (closed squares with solid line), 5% (closed circles with solid line) or 1% O₂ (open squares and open circles with dashed line), mean ± SEM of 4–5 independent experiments (n ≥ 4). The table represents the p‐value obtained by two‐tailed Student's t‐test calculations. Comparison were made between CTLs generated at 21% and reactivated at 1% versus CTLs generated at 21% and reactivated at 21% O₂ (“21 to 1%” versus “21 to 21%”) or between CTL generated at 5% and reactivated at 1% versus CTL generated at 5% and reactivated at 5% O₂ (“5 to 1%” versus “5 to 5%”). Mean cytokine secretion of (E) IL‐10, (G) IFN‐γ, (H) IL‐2, and (I) TNF‐α ± SEM out of three independent experiments (n = 3).*p < 0.05. (F) Cell yield from CTLs reactivated for four days in the presence of IL‐10 blocking antibody (open bars) or isotype control (closed bars), mean + SD of two independent experiments (n = 2). *p < 0.05 (Student's t‐test).

Figure 4

CD25, CD137, and IL‐10 are expressed by CD8⁺ TILs ex vivo, and are positively regulated by hypoxia after reactivation. Spleen and tumor CD8⁺ T cells were purified from E.G7‐OVA‐bearing mice. (A) CD25 and CD137 expression was analyzed by flow cytometry. Dot plots show one representative experiment. Bar graphs represent the mean ± SEM percent CD25‐positive or CD137‐positive CD8⁺ T cells from six independent experiments (n = 17). (B) il10 and hif2a expression were quantified by qPCR. Results show mean + SEM expression from five independent experiments (n = 13). (C–F) CD8⁺ TILs were cultured for three days with or without αCD3/αCD28‐coated beads. (C) Results show mean ± SEM absolute cell number and viability from four independent experiments (n = 8). (D) CD25 and CD137 were analyzed by flow cytometry. Open histograms: control isotype staining; closed histograms: staining for the molecule of interest. Line graphs (dashed lines: unstimulated; solid lines: αCD3/αCD28‐stimulated): median fluorescence intensity ± SEM from four independent experiments (n = 8). (E) IL‐10 and (F) IFN‐γ secretion were quantified by ELISA. Results show mean ± SEM cytokine secretion from four independent experiments (n = 8). ns: not statistically significant, *p < 0.05, **p < 0.01, ****p < 0.0001 (Student's t‐test).