Activator protein 1 suppresses antitumor T-cell function via the induction of programmed death 1

Abstract

T cells play a critical role in tumor immunosurveillance by eliminating newly transformed somatic cells. However, tumor cell variants can escape from immunological control after immunoediting, leading to tumor progression. Whether and how T cells respond to tumor growth remain unclear. Here, we found that tumor-infiltrating T cells exhibited persistently up-regulated expression of the activator protein 1 (AP-1) subunit c-Fos during tumor progression. The ectopic expression of c-Fos in T cells exacerbated tumor growth, whereas the T-cell-specific deletion of c-Fos reduced tumor malignancy. This unexpected immunosuppressive effect of c-Fos was mediated through the induced expression of immune inhibitory receptor programmed death 1 (PD-1) via the direct binding of c-Fos to the AP-1-binding site in the Pdcd1 (gene encoding PD-1) promoter. A knock-in mutation of this binding site abrogated PD-1 induction, augmented antitumor T-cell function and repressed tumor growth. Taken together, these findings indicate that T-cell c-Fos subsequently induces PD-1 expression in response to tumor progression and that disrupting such induction is essential for repression of tumor growth.

Fig. 1.

Tumor-infiltrating T cells exhibit increased c-Fos expression during tumor progression. (A) Lung-infiltrating T cells were FACS-purified from LLC-inoculated or noninoculated C57BL/6 mice at day 21 postinoculation for mRNA quantification (mean ± SD; n = 3; **P < 0.005 by Student t test). (B) EGFP expression in lung-infiltrating T cells from treated Cd4Cre⁺c-Fos^+/fl mice (mean ± SEM; n = 5; *P < 0.05 and **P < 0.005 by Student t test). (C) EGFP⁺ T cells from LLC-inoculated reporter mice at day 21 postinoculation were purified with their EGFP⁻ counterparts for c-Fos mRNA quantification (mean ± SD; n = 3; ***P < 0.001 and **P < 0.005 by Student t test). (D) Flow-cytometric analysis of CD44 and EGFP expression in T cells from the lung, MLN, and SPL of reporter mice at day 21 post-LLC inoculation. Statistical analysis is shown in Table S1.

Fig. 2.

T-cell c-Fos overexpression promotes tumor growth. (A) Survival of LLC-inoculated FosTg mice compared with littermate controls (FosTg, n = 11; control, n = 10; P < 0.005 by log-rank test). (B) Number of lung tumors at day 28 post-LLC inoculation (FosTg, n = 8; control, n = 8; ***P < 0.001 by Student t test). (C) Lungs at days 21 and 28 post-LLC inoculation. (Scale bars, 5 mm.) The results are representative of four independent experiments. (D) Hematoxylin and eosin (HE)-stained lung sections at day 28 post-LLC inoculation (magnification, 40× and 400×). (Scale bars, 200 μm.) (E) Ki-67 staining of lung sections. Proliferating tumor cells of each tumor were quantified as described in Materials and Methods; more than 50 HPFs were counted from each group (magnification, 400×; mean ± SEM; FosTg, n = 15 from 3 mice; control, n = 8 from 3 mice). (Scale bars, 200 μm.) (F) Cleaved caspase-3 staining and apoptotic tumor cell quantification (magnification, 400×; mean ± SEM; FosTg, n = 14 from 3 mice; control, n = 7 from 3 mice; *P < 0.05 by Student t test). (Scale bars, 200 μm.)

Fig. 3.

T-cell c-Fos depletion reduces tumor growth. (A) Survival of LLC-inoculated Foscko mice compared with littermate controls (Foscko, n = 10; control, n = 18; P < 0.001 by log-rank test). (B) Number of lung tumors at day 28 post-LLC inoculation (Foscko, n = 7; control, n = 12; **P < 0.005 by Student t test). (C) Lungs at days 21 and 28 post-LLC inoculation. (Scale bars, 5 mm.) The results are representative of nine independent experiments. (D) HE-stained lung sections at day 28 post-LLC inoculation (magnification, 40× and 400×.) (Scale bars, 200 μm.) (E) Ki-67 staining of lung sections. Proliferating tumor cells were quantified (magnification, 400×; mean ± SEM; Foscko, n = 9 from 3 mice; control, n = 10 from 3 mice; **P < 0.005 by Student t test). (Scale bars, 200 μm.) (F) Cleaved caspase-3 staining and apoptotic tumor cell quantification (magnification, 400×; mean ± SEM; Foscko, n = 14 from 3 mice; control, n = 13 from 3 mice; ***P < 0.001 by Student t test). (Scale bars, 200 μm.)

Fig. 4.

PD-1 is a potential downstream target of c-Fos in tumor-infiltrating T cells. (A) CD44^hi tumor-infiltrating T cells were purified from FosTg and control mice at day 21 post-LLC inoculation for mRNA quantification (mean ± SD; n = 3; **P < 0.005 by Student t test). (B) EGFP⁺ tumor-infiltrating T cells were purified from Foscko and control mice at day 21 post-LLC inoculation for mRNA quantification (mean ± SD; n = 3; **P < 0.005 by Student t test). (C and D) Flow-cytometric analysis of PD-1 expression in CD44^hi tumor-infiltrating T cells from FosTg mice (C), EGFP⁺ tumor-infiltrating T cells from Foscko mice (D), and controls at day 28 post-LLC inoculation (solid line, anti-PD-1; shaded in gray, isotype control). The results are representative of three independent experiments.

Fig. 5.

c-Fos (AP-1) transactivates PD-1 expression by directly binding to the AP-1–binding site D in the Pdcd1 promoter. (A) Jurkat cells were electroporated with the pGL3 vector or Pdcd1-pGL3 (as shown in Fig. S4B) for reporter assay (mean ± SD; n = 3; **P < 0.005 by Student t test). (B) c-Fos plasmids combined with Jun plasmids were cotransfected with the pGL3 vector or the Pdcd1-pGL3 into Jurkat cells for reporter assay (mean ± SD; n = 3; **P < 0.005 and *P < 0.05 by Student t test). (C) Increasing concentrations (μg) of the AP-1 (c-Fos and JunB) constructs were cotransfected with the pGL3 vector or the Pdcd1-pGL3 into Jurkat cells for reporter assay (mean ± SD; n = 3; ***P < 0.001 and **P < 0.005 by Student t test). (D) Tumor-infiltrating T cells were purified from C57BL/6 mice at day 21 post-LLC inoculation and lysed for ChIP assay (mean ± SD; n = 3; *P < 0.05 by Student t test). (E) The reporter activity of the mutant form DM-pGL3 and Pdcd1-pGL3 was measured (mean ± SD; n = 3; ***P < 0.001 by Student t test).

Fig. 6.

Knock-in mutation of the AP-1–binding site D leads to repressed tumor growth and enhanced T-cell antitumor functions. (A) Survival of LLC-inoculated Pdcd1^KI/KI mice compared with littermate controls (Pdcd1^KI/KI, n = 17; control, n = 16; P < 0.001 by log-rank test). (B) Lung tumor numbers were counted at day 35 post-LLC inoculation (Pdcd1^KI/KI, n = 9; control, n = 7; **P < 0.005 by Student t test). (C) Images of lungs at days 28 and 35 post-LLC inoculation. (Scale bars, 5 mm.) The results are representative of nine independent experiments. (D) Lung sections with HE staining at day 28 post-LLC inoculation (magnification, 40× and 400×). Ki-67 (E) and cleaved caspase-3 (F) staining of lung sections were performed. (Scale bars, 200 μm.) Quantification was not performed because of the small tumor sizes observed in Pdcd1^KI/KI mice. (G) CD44 expression in tumor-infiltrating T cells at day 35 post-LLC inoculation (statistically analyzed in Table S1). (H) Flow-cytometric analysis of PD-1 expression in CD44^hi tumor-infiltrating T cells at day 35 post-LLC inoculation (solid line, anti-PD-1; shaded in gray, isotype control). The results are representative of three independent experiments. (I) Tumor-infiltrating lymphocytes were harvested at day 35 post-LLC inoculation. Cytokine expression was assessed. (J) Surface CD25 and intracellular Foxp3 expression in tumor-infiltrating CD4 T cells. Production of IL-17a (K), IL-4 (L), and IFN-γ (M) from tumor-infiltrating CD4 T cells was analyzed. Granzyme B (N) and perforin (O) expression in tumor-infiltrating CD8 T cells assessed. Results from (I) to (O) were statistically analyzed as shown in Table S1.

Fig. 7.

Mutation of the AP-1–binding site relieves spontaneous tumor growth and enhanced T-cell antitumor function in Pdcd1^KI/KI PyMT mice. (A) Flow-cytometric analysis of PD-1 expression in CD44^hi tumor-infiltrating T cells from mice at day 130 after birth (solid line, anti-PD-1; shaded in gray, isotype control). (B) Tumor onset was monitored for tumor-free survival analysis (Pdcd1^KI/KI PyMT, n = 9; control, n = 20; P < 0.001 by log-rank test). (C) Mouse mammary tumor growth was monitored starting from 6 wk after birth, up to 15 wk (mean ± SEM, Pdcd1^KI/KI PyMT, n = 9; control, n = 18; *P < 0.05 by Student t test with Welch’s correction). (D) Mouse tumor sections with HE staining at day 130 after birth (magnification, 40× and 400×). (Scale bars, 200 μm.) Ki-67 (E) and cleaved caspase-3 (F) staining of tumor sections from (D) were performed, more than 150 HPFs were counted for each group (magnification, 400×; mean ± SEM; Pdcd1^KI/KI PyMT, n = 10 from 3 mice for Ki-67; n = 11 from 3 mice for cleaved caspase-3; control, n = 11 from 3 mice for Ki-67; n = 10 from 3 mice for cleaved caspase-3; *P < 0.05 by Student t test). (Scale bars, 200 μm.) (G) Cytokine expression was assessed. (H) Surface CD25 and intracellular Foxp3 expression in tumor-infiltrating CD4 T cells. Production of IL-17a (I), IL-4 (J), and IFN-γ (K) from tumor-infiltrating CD4 T cells was analyzed as in G. Granzyme B (L) and perforin (M) expression in tumor-infiltrating CD8 T cells from mice in A was assessed. Results from G–M were statistically analyzed as shown in Table S1.