Upregulation of PD-L1 Expression by Prostaglandin E2 and the Enhancement of IFN-γ by Anti-PD-L1 Antibody Combined With a COX-2 Inhibitor in Mycoplasma bovis Infection

Abstract

Bovine mycoplasmosis caused by Mycoplasma bovis results in pneumonia and mastitis in cattle. We previously demonstrated that the programmed death 1 (PD-1)/PD-ligand 1 (PD-L1) pathway is involved in immune dysfunction during M. bovis infection and that prostaglandin E₂ (PGE₂) suppressed immune responses and upregulated PD-L1 expression in Johne's disease, a bacterial infection in cattle. In this study, we investigated the role of PGE₂ in immune dysfunction and the relationship between PGE₂ and the PD-1/PD-L1 pathway in M. bovis infection. In vitro stimulation with M. bovis upregulated the expressions of PGE₂ and PD-L1 presumably via Toll-like receptor 2 in bovine peripheral blood mononuclear cells (PBMCs). PGE₂ levels of peripheral blood in infected cattle were significantly increased compared with those in uninfected cattle. Remarkably, plasma PGE₂ levels were positively correlated with the proportions of PD-L1⁺ monocytes in M. bovis-infected cattle. Additionally, plasma PGE₂ production in infected cattle was negatively correlated with M. bovis-specific interferon (IFN)-γ production from PBMCs. These results suggest that PGE₂ could be one of the inducers of PD-L1 expression and could be involved in immunosuppression during M. bovis infection. In vitro blockade assays using anti-bovine PD-L1 antibody and a cyclooxygenase 2 inhibitor significantly upregulated the M. bovis-specific IFN-γ response. Our study findings might contribute to the development of novel therapeutic strategies for bovine mycoplasmosis that target PGE₂ and the PD-1/PD-L1 pathway.

Keywords: Mycoplasma bovis; PD-1; PD-L1; T-cell exhaustion; cattle; immune dysfunction; immunoinhibitory molecules; prostaglandin E2.

Figure 1

M. bovis upregulates PD-L1 and PGE₂ expression. (A–F) PBMCs from M. bovis-uninfected cattle were incubated with (A–C) live M. bovis (MOI of 0.1, 1, or 10) or (D–F) heat-killed M. bovis (1.5 ng/ml) for 24 h. (A) PD-L1 expression on CD11b⁺CD14⁺ monocytes (untreated: n = 14, MOI 0.1: n = 14, MOI 1: n = 14, MOI 10: n = 14) was determined using flow cytometry. (B) PGE₂ levels in culture supernatants (untreated: n = 9, MOI 0.1: n = 9, MOI 1: n = 9, MOI 10: n = 9) were determined using ELISA. (C) Positive correlation is noted between PD-L1 expression on CD11b⁺CD14⁺ monocytes and PGE₂ levels under live M. bovis stimulation (MOI of 10, n = 11). (D) PD-L1 expression on CD11b⁺CD14⁺ monocytes (untreated: n = 13, heat-killed M. bovis: n = 13) were determined using flow cytometry. (E) PGE₂ levels in culture supernatants (untreated: n = 13, heat-killed M. bovis: n = 13) were determined using ELISA. (F) Positive correlation is noted between PD-L1 expression on CD11b⁺CD14⁺ monocytes and PGE₂ levels under heat-killed M. bovis stimulation (n = 13). (G) PD-L1 expression on CD11b⁺CD14⁺ monocytes cultured in 24-h culture supernatant of PBMCs with or without live M. bovis (untreated: n = 9, MOI 0.1: n = 9, MOI 1: n = 9, MOI 10: n = 9) were analyzed by flow cytometry. (H) PBMCs from uninfected cattle were incubated with PGE₂, and PD-L1 expression on CD11b⁺CD14⁺ monocytes (DMSO: n = 6, PGE₂: n = 6) was determined using flow cytometry. The bars are the median of the values (A,B,D,E,G,H). Statistical significance was determined by the Steel–Dwass test (A,B,G) or Mann–Whitney U-test (D,E,H). Correlation statistics were analyzed using Spearman's correlation analysis (C,F). *P < 0.05.

Figure 2

PGE₂ production from CD14⁺ cells induced by M. bovis. CD14⁺ cells or CD14⁻ cells from uninfected cattle were incubated with live M. bovis (MOI of 10) for 24 h. (A) PGE₂ production was determined by ELISA (CD14⁻ cells untreated: n = 6, CD14⁻ cells Live M. bovis: n = 6, CD14⁺ cells untreated: n = 6, CD14⁺ cells M. bovis: n = 6). (B) PD-L1 expression on CD14⁺ cells (untreated: n = 6, live M. bovis: n = 6) was determined by flow cytometry. The bars are the median of the values (A,B). Statistical significance was determined by the Steel–Dwass test (A) or Mann–Whitney U-test (B). **P < 0.01.

Figure 3

TLR2 signaling upregulates PD-L1 and PGE₂ expression. PBMCs from uninfected cattle were incubated with FSL-1 (100 ng/ml). (A) PGE₂ production was determined by ELISA (PBS: n = 11, FSL-1: n = 11). (B) PD-L1 expression on CD14⁺ cells (PBS: n = 11, FSL-1: n = 11) was determined by flow cytometry. (C) Positive correlation is noted between PD-L1 expression on CD11b⁺CD14⁺ monocytes and PGE₂ levels under FSL-1 stimulation (n = 13). (D) PBMCs from uninfected cattle were incubated with FSL-1 (100 ng/ml) or heat-killed M. bovis (1.5 ng/ml) under inhibition of TLR signaling by SsnB and PGE₂ production was determined by ELISA (DMSO: n = 9, SsnB: n = 9, DMSO and FSL-1: n = 9, SsnB and FSL-1: n = 9, DMSO and M. bovis: n = 9, SsnB and M. bovis: n = 9). The bars are the median of the values (A,B,D). Statistical significance was determined by the Mann–Whitney U-test (A,B) or the Steel–Dwass test (D). Correlation statistics were analyzed using Spearman's correlation analysis (C). *P < 0.05.

Figure 4

Analysis of PGE₂ in cattle infected with M. bovis. (A,B) Serum PGE₂ levels in M. bovis-infected cattle (n = 89; Mastitis: n = 62, Pneumonia: n = 21, Pneumonia with arthritis: n = 6) and uninfected cattle (n = 18) was determined by ELISA. (C–E) Correlation between plasma PGE₂ levels in M. bovis-infected cattle with arthritis (n = 13), otitis (n = 10), pneumonia (n = 5), and other factors. (C) Correlation between plasma PGE₂ levels and plasma IFN-γ levels (n = 29). (D) Correlation between the plasma levels of PGE₂ and the proportions of PD-L1⁺ monocytes (n = 25). (E) Correlation between plasma PGE₂ levels and IFN-γ production from PBMCs against heat-killed M. bovis (n = 25). The bars are the median of the values (A,B). Statistical significance was determined by the Mann–Whitney U-test (A) or the Steel–Dwass test (B). Correlation statistics were analyzed using Spearman's correlation analysis (C–E). *P < 0.05, **P < 0.01.

Figure 5

PD-L1 and PGE₂ expression in the lung of M. bovis-infected cattle with pneumonia. Immunohistochemical staining of PD-L1 (A,B) and PGE₂ (C,D) in lung tissues of cattle with or without M. bovis infection was performed using anti-bovine PD-L1 mAb (6C11-3A11) and human PGE₂ polyclonal Ab (rabbit polyclonal). Black arrowheads, PD-L1 positive macrophages; black arrows, PD-L1 positive fibroblasts; white arrowheads, PGE₂-positive macrophages; white arrows, PGE₂-positive epithelial cells.

Figure 6

Activation of IFN-γ responses by the combination of meloxicam and anti-PD-L1 Ab. PBMCs from cattle infected with M. bovis (n = 7) were incubated with meloxicam and anti-PD-L1 mAb in the presence of anti-bovine CD3 and CD28 mAbs (A) or heat-killed M. bovis (B). IFN-γ production in culture supernatants was determined by ELISA. The bars are the median of the values (A,B). Statistical significance was determined by the Steel–Dwass test (A,B). *P < 0.05.