Diurnal Expression of PD-1 on Tumor-Associated Macrophages Underlies the Dosing Time-Dependent Antitumor Effects of the PD-1/PD-L1 Inhibitor BMS-1 in B16/BL6 Melanoma-Bearing Mice

Abstract

Cancer cells have acquired several pathways to escape from host immunity in the tumor microenvironment. Programmed death 1 (PD-1) receptor and its ligand PD-L1 are involved in the key pathway of tumor immune escape, and immune checkpoint therapy targeting PD-1 and PD-L1 has been approved for the treatment of patients with certain types of malignancies. Although PD-1 is a well-characterized receptor on T cells, the immune checkpoint receptor is also expressed on tumor-associated macrophages (TAM), a major immune component of the tumor microenvironment. In this study, we found significant diurnal oscillation in the number of PD-1-expressing TAMs collected from B16/BL6 melanoma-bearing mice. The levels of Pdcd1 mRNA, encoding PD-1, in TAMs also fluctuated in a diurnal manner. Luciferase reporter and bioluminescence imaging analyses revealed that a NF-κB response element in the upstream region of the Pdcd1 gene is responsible for its diurnal expression. A circadian regulatory component, DEC2, whose expression in TAMs exhibited diurnal oscillation, periodically suppressed NF-κB-induced transactivation of the Pdcd1 gene, resulting in diurnal expression of PD-1 in TAMs. Furthermore, the antitumor efficacy of BMS-1, a small molecule inhibitor of PD-1/PD-L1, was enhanced by administering it at the time of day when PD-1 expression increased on TAMs. These findings suggest that identification of the diurnal expression of PD-1 on TAMs is useful for selecting the most appropriate time of day to administer PD-1/PD-L1 inhibitors.

Implications: Selecting the most appropriate dosing time of PD-1/PD-L1 inhibitors may aid in developing cancer immunotherapy with higher efficacy.

Figure 1.

Diurnal variations in the population rate of PD-1–expressing TAMs in mouse B16/BL6 melanoma-forming tumor masses. A, Schematic depicting the isolation method of TAMs from mouse B16/BL6 melanoma-forming tumor masses. B, Temporal mRNA expression profiles of Bmal1, Clock, Per1, Per2, Cry1, Dec1, Dec2, Dbp, Nfil3, Rorα, and Rev-erbα in TAMs and circulating monocytes. Data were normalized by 18S rRNA levels. Values are the mean with SD (n = 3). There were significant time-dependent variations in the mRNA levels of all circadian clock gene in both TAMs and circulating monocytes (P < 0.01, respectively; one-way ANOVA). C, The left diagrams show the representative proportion of PD-1⁺ TAMs collected at ZT6 and ZT18. Right panel shows temporal profiles of the population of PD-1–expressing F4/80⁺ CD11b⁺ CD206⁺ TAMs. Values are the mean with SD (n = 3). There was a significant time-dependent variation in the population of PD-1–expressing TAMs. (F_5,12 = 63.251, P < 0.001; one-way ANOVA). D, Temporal expression profile of Pdcd1 mRNA in TAMs. Data were normalized by 18S rRNA levels. Values are the mean with SD (n = 3). There was a significant time-dependent variation in Pdcd1 mRNA levels (F_5,12 = 4.167, P = 0.020; one-way ANOVA). The horizontal bar at the bottom of each panel indicates light and dark cycles.

Figure 2.

DEC2 regulates the circadian expression of Pdcd1 mRNA in macrophages. A, Temporal mRNA expression profiles of Per2, Bmal1, and Pdcd1 in RAW264.7 cells, whose circadian clocks were synchronized by treatment with 100 nmol/L DEX for 2 hours. Nontreatment cells were set as a nonsynchronized control. Data were normalized by the levels of 18S rRNA and the mean of each group was set at 1.0. Values are the mean with SD (n = 3). There were significant time-dependent variations in Per2, Bmal1, and Pdcd1 in DEX treatment group (F_{12, 26} = 53.225, P < 0.001 for Per2; F_{12, 26} = 12.609, P < 0.001 for Bmal1; F_{12, 26} = 18.874, P < 0.001 for Pdcd1; one-way ANOVA). B, DEC2 negatively regulates the transcription of the Pdcd1 gene. RAW264.7 cells were cotransfected with Pdcd1(-2050/+63)::Luc, and expression vectors for PER1, PER2, CRY1, DEC1, DEC2, CLOCK/BMAL1, NFIL3, DBP, RORα, and REV-ERBα. The values are the mean with SD (n = 3). The value of empty vector (pcDNA3.1)-transfected RAW264.7 cells was set at 1.0. *, P < 0.05; **, P < 0.01; ***, P < 0.001; significant difference from empty vector (pcDNA3.1)-transfected groups (F_10,33 = 18.109, P < 0.001; ANOVA with the Tukey Kramer post hoc test). C, Temporal expression profiles of DEC2 protein in mock-transduced and Dec2 KD RAW264.7 cells, whose circadian clocks were synchronized by treatment with 100 nmol/L DEX. Data were normalized by the β-ACTIN levels. Values are the mean with SD (n = 3). (F_{1, 28} = 216.110, P < 0.01 for group; F_6,28 = 17.930, P < 0.01 for time point; F_6,28 = 12.966, P < 0.01 for time point×group; two-way ANOVA). D, Temporal expression profiles of Pdcd1 mRNA in mock-transduced and Dec2 KD RAW264.7 cells, whose circadian clocks were synchronized by treatment with 100 nmol/L DEX. Data were normalized by the levels of 18S rRNA. Values are the mean with SD (n = 3). (F_1,28 = 323.663, P < 0.01 for group; F_6,28 = 8.459, P = 0.016 for time point; F_6,28 = 4.890, P < 0.01 for time point×group; two-way ANOVA).

Figure 3.

Repression of NF-κB–mediated transactivation by DEC2 underlies the circadian expression of Pdcd1. A, Bioluminescence profiles driven by Pdcd1(-2050/+63)::Luc-, Pdcd1(-1540/+63)::Luc-, and Pdcd1(-913/+63)::Luc-transfected NIH3T3 cells after treatment with 100 nmol/L DEX for 2 hours. The upper schematic diagrams show luciferase reporter constructs containing different lengths of the upstream region of the mouse Pdcd1 gene. Closed boxes indicate the sites homologous with clock gene response elements and the numbers of nucleotide residues indicate the distance from the transcription start site (+1). Values are the mean with SD (n = 6–8). B, Bioluminescence profiles driven by Pdcd1(-1540/+63)::Luc in circadian clock-synchronized NIH3T3 cells transfected with Dec2-expressing vectors or control (pcDNA) vectors. C, Location of the NRE in the upstream region of the mouse Pdcd1 gene. D, Suppression of p65-mediated transactivation of the NRE::Luc by DEC2. RAW264.7 cells were cotransfected with NRE::Luc, and expression vectors for p65 and DEC2. Values are the mean with SD (n = 3). The value of empty vector (pcDNA3.1)-transfected RAW264.7 cells was set at 1.0. **, P < 0.01; significant difference between the two groups (F_3,8 = 237.051, P < 0.001; one-way ANOVA with Tukey Kramer post hoc test). E, Suppression of p65-mediated transactivation of the Pdcd1 (–1540/+63)::Luc by DEC2. RAW264.7 cells were cotransfected with Pdcd1(–1540)::Luc, and expression vectors for p65 and DEC2. Values are the mean with SD (n = 3). The value of empty vector (pcDNA3.1)-transfected RAW264.7 cells was set at 1.0. **, P < 0.01; significant difference between the two groups (F_3,8 = 29.215, P < 0.001; one-way ANOVA with Tukey Kramer post hoc test). F, Suppression of LPS-induced nuclear translocation of p65 by DEC2. RAW264.7 cells were transfected with empty vector (pcDNA3.1) or Dec2-expressing vector and then treated with 1 µg/mL of LPS for 30 minutes. **, P < 0.01; significant difference between the two groups (F_3,8 = 96.447, P < 0.001; one-way ANOVA with Tukey Kramer post hoc test).

Figure 4.

Regulation of antitumor immunity of RAW264.7 macrophages by DEC2. (A and B) The viability of B16/BL6 melanoma (A) and mock-transduced or Dec2 KD RAW264.7 cells (B) after treatment with 10 µmol/L BMS-1 for 24 hours. C, BMS-1 increases the antitumor immunity of RAW264.7 cells under coculture conditions. B16/BL6 melanoma was cocultured with mock-transduced or Dec2 KD RAW264.7 cells, and cells were treated with 10 µmol/L BMS-1 or vehicle (0.2% DMSO) for 24 hours. All experiments were conducted without synchronization of the circadian clock. Values are the mean with SD (n = 6). **, P < 0.01; significant difference between the two groups (F_3,20 = 419.160, P < 0.001; one-way ANOVA with Tukey Kramer post hoc test).

Figure 5.

Dosing time-dependent change in the ability of BMS-1 to inhibit the growth of B16/BL6 melanoma implanted in mice. A, Difference in the number of phagocytic macrophages in B16/BL6 tumor masses at 7 days after the initiation of BMS-1 administration at ZT6 and ZT18. Cflar KD B16/BL6 cells were implanted into subcutaneous tissue of the back region in C57BL/6J mice. From 9 days after implantation, mice were intratumorally (i.t.) injected with a single daily dose of BMS-1 (50 µg) or vehicle (10% DMSO in PBS) at ZT6 or ZT18. The left panel shows immunofluorescence labelling of F4/80 (red) in the GFP-expressing B16/BL6 (green) tumor masses. The scale bar indicates 50 µm. The phagocytic macrophages were manually counted in a blinded manner. Values are the mean with SD (n = 7–8). **, P < 0.01; significant difference between the two groups (F_2,19 = 17.746, P < 0.001; one-way ANOVA with Tukey Kramer post hoc test). B, Difference in the volume of B16/BL6 melanoma-forming tumor cells at 7 days after the initiation of BMS-1 administration at ZT6 and ZT18. Values are the mean with SD (n = 7–8). *, P < 0.05; significant difference between the two groups (F_2,20 = 5.271, P = 0.0145; one-way ANOVA with Tukey Kramer post hoc test). Upper photograph shows dissected tumor from B16/BL6 melanoma-implanted mice at 7 days after the initiation of BMS-1 administration at ZT6 and ZT18.

Figure 6.

Schematic diagram underlying mechanism of the dosing time-dependent changes in the antitumor effects of PD-1/PD-L1 inhibitor BMS-1 in B16BL6 melanoma-implanted mice. The time-dependent repression of p65-mediated transactivation of Pdcd1 by DEC2 induces the circadian expression of PD-1 in TAMs. The antitumor efficacy of BMS-1 is enhanced by administering at the time of day when PD-1 expression is increased on TAMs.