Neutrophil-Derived IL-1β Impairs the Efficacy of NF-κB Inhibitors against Lung Cancer

Abstract

Although epithelial NF-κB signaling is important for lung carcinogenesis, NF-κB inhibitors are ineffective for cancer treatment. To explain this paradox, we studied mice with genetic deletion of IKKβ in myeloid cells and found enhanced tumorigenesis in Kras(G12D) and urethane models of lung cancer. Myeloid-specific inhibition of NF-κB augmented pro-IL-1β processing by cathepsin G in neutrophils, leading to increased IL-1β and enhanced epithelial cell proliferation. Combined treatment with bortezomib, a proteasome inhibitor that blocks NF-κB activation, and IL-1 receptor antagonist reduced tumor formation and growth in vivo. In lung cancer patients, plasma IL-1β levels correlated with poor prognosis, and IL-1β increased following bortezomib treatment. Together, our studies elucidate an important role for neutrophils and IL-1β in lung carcinogenesis and resistance to NF-κB inhibitors.

Figure 1

Inhibition of NF-κB signaling in myeloid cells increases lung tumorigenesis and epithelial cell proliferation. A) Representative photomicrographs and B) Number of lung tumors in WT and IKKβ^Δmye mice at 16 weeks after a single injection of urethane (n=16-22 mice per group). C) Representative photomicrographs showing an AAH lesion (red arrow) in the lung of WT mice or tumor in IKKβ^Δmye mice, and D) Number of AAH lesions counted per H&E-stained lung section (3 sections per mouse) from WT and IKKβ^Δmye mice harvested at week 6 after injection of urethane (n=9-10 mice per group). E) Immunostaining for PCNA⁺ cells and (F) Number of PCNA⁺ cells per lung section (averaged from 25 sequential fields taken at 40× magnification) from WT and IKKβ^Δmye mice harvested at week 6 after urethane injection (n=3-4 per group). G-H) Lethally-irradiated LSL-Kras^G12D mice received bone marrow from WT (WT→Kras^G12D) or IKKβ^Δmye (IKKβ^Δmye→Kras^G12D) mice. Lung tumors were induced by instillation of IT adeno-Cre (1.5×10⁷ PFU). G) Representative photomicrographs and H) Number of lung tumors in WT→Kras^G12D and IKKβ^Δmye→Kras^G12D mice at 8 weeks after adeno-Cre (n=4-9 mice per group) *p < 0.05. See also Figure S1.

Figure 2

Neutrophils are increased in the lungs of mice lacking myeloid NF-κB signaling. A) Number of total BAL cells in WT and IKKβ^Δmye mice at baseline (C) and at 6 weeks after urethane injection (U) (n=7-9 mice per group; *p < 0.05 compared with urethane-treated WT mice). B) Representative FACS plots and (C-D) Percentages of viable CD45⁺/CD11b⁺/Gr1^hi neutrophils (Gr1^hi), CD45⁺/CD11b⁺/Gr1^low monocytes (Gr1^low), and CD45⁺/CD11b⁺/Gr1^neg macrophages (Gr1^neg) in the lungs of WT and IKKβ^Δmye mice at (C) baseline and (D) 6 weeks after urethane injection (n=4-11 mice per group; *p < 0.05 compared with WT). E) Representative FACS plots and (F) total viable CD45⁺/CD11b⁺/Ly6G⁺ neutrophils in the lungs of WT→Kras^G12D and IKKβ^Δmye→Kras^G12D mice 8 weeks after adeno-Cre (n=4 mice per group; *p < 0.05 compared with WT→Kras^G12D). Ly6G identifies the granulocytic subgroup of the Gr1 marker. See also Figure S2.

Figure 3

Neutrophils promote lung tumorigenesis in the absence of myeloid NF-κB signaling. All mice were treated with isotype control IgG or anti-Ly6G depletion antibodies (100 g by IP injection) for the first 6 weeks following urethane injection. A) Representative FACS plots and (B) total viable CD45⁺/CD11b⁺/Ly6C⁺/Ly6G⁺ lung neutrophils demonstrating depletion efficiency in IKKβ^Δmye mice harvested 3 days after the last dose of antibody (n=4 mice per group). C) Number of AAH lesions per lung section from IgG- and anti-Ly6G-treated WT and IKKβ^Δmye mice at 6 weeks after urethane injection (n=6-9 mice per group). D) Lethally-irradiated WT mice received bone marrow from WT or IKKβ^Δmye mice. Lung tumors at 16 weeks after urethane injection in bone marrow chimera mice treated with IgG or anti-Ly6G antibodies for the first 6 weeks of tumorigenesis (n=6-8 mice per group). *p < 0.05.

Figure 4

Mature neutrophils are increased in the lungs during early tumorigenesis in the absence of myeloid NF-κB signaling. A) FACS sorting strategy and photomicrographs demonstrating cell morphology of lung monocytes (CD45⁺/CD11b⁺/Ly6C⁺/Ly6G⁻) and neutrophils (CD45⁺/CD11b⁺/Ly6C⁺/Ly6G⁺) isolated from lungs of WT and IKKβ^Δmye mice at 1 week after urethane injection. B) Numbers of CD11b⁺/Ly6G⁺ neutrophils (Ly6G⁺), CD11b⁺/Ly6C⁺ monocytes (Ly6C⁺), and CD11b⁺/Ly6G^neg/Ly6C^neg macrophages (Ly6G/C^neg) in the lungs of WT and IKKβ^Δmye mice at 1 week after urethane injection (n=3 mice per group, representative of 2 independent experiments; *p<0.05 compared to WT). C) Flow cytometry plot (including fluorescence minus one [FMO] control) and (D) mean fluorescence intensity (MFI) showing expression of MPO in viable CD45⁺/CD11b⁺/Ly6G⁺ cells from lungs of WT and IKKβ^Δmye mice at 1 week after urethane injection (n=4 mice per group). Expression of (E) N1 and (F) N2 markers in CD45⁺/CD11b⁺/Ly6G⁺ cells isolated from lungs of IKKβ^Δmye mice at 1 week after urethane injection (n=4-5 mice per group). G) CD45⁺/CD11b⁺/Ly6G⁺ cells isolated from lungs of IKKβ^Δmye mice at 1 week after urethane injection do not impair the ability of allogeneic dendritic cells (DC) to induce proliferation of CFSE-labeled responder CD4⁺/CD25⁻ T cells (Teff) (1:1, performed in duplicate).

Figure 5

Neutrophils from IKKβ^Δmye mice produce increased IL-1β following urethane injection. Expression of cytokines by A) mRNA and B) protein in the lungs of WT and IKKβ^Δmye mice harvested 1 week after urethane (n=10-11 mice per group; *p < 0.05 compared with WT). C) Concentration of IL-1β in the conditioned media following 12-hour culture of lung Ly6G⁺ neutrophils, Ly6C⁺ monocytes, or Ly6G/C^neg macrophages isolated from IKKβ^Δmye mice at 1 week after urethane injection (n=3; *p < 0.05 compared with Ly6C⁺ and Ly6G/C^neg). D) Concentration of IL-1β in the conditioned media following 12-hour culture of equal numbers of lung Ly6G⁺ neutrophils from WT and IKKβ^Δmye mice isolated at 1 week after urethane injection (n=8 mice per group). E) IL-1β protein levels in lung homogenates at 1 week after urethane in the lungs of IKKβ^Δmye mice treated with liposomal clodronate or PBS on day 5 following urethane injection (n=6 mice per group). F) IL-1β protein levels at 1 week after urethane in the lungs of IKKβ^Δmye mice treated with anti-Ly6G antibodies (100 μg) or control IgG antibodies by IP injection on days −1, 2, and 5 relative to the day of urethane injection (n=3-5 mice per group; *p < 0.05 compared with IKKβ^Δmye mice treated with control IgG antibodies). Lung Ly6G⁺ neutrophils were isolated from WT and IKKβ^Δmye mice at 1 week after urethane injection. G) IL-1β concentration in the conditioned media after culture with inhibitors (all 100 M) of caspase-1 (Ac-YVAD-CMK), neutrophil elastase and proteinase-3 (MeOSuc-APPV-CMK), or cathepsin G (Z-GLP-CMK) (n=3-8 replicates per group; #p>0.05 compared to WT Unt; *p<0.05 compared to either WT or IKKβ^Δmye Unt). H) mRNA expression of cathepsin G in lung Ly6G⁺ neutrophils isolated from WT and IKKβ^Δmye mice at 1 week after urethane injection.

Figure 6

Pharmacological inhibition of NF-κB increases IL-1β in mice and indicates worse survival in NSCLC patients. A) Schematic representation of NF-κB inhibition protocol using bortezomib (Bort). In addition to urethane, WT mice were treated with IP injections of Bort (1 mg/kg) or vehicle control (Veh). B) BAL cells in Bort- or Veh-treated WT mice at 1 week after urethane injection (n=4-5 mice per group; *p<0.05 compared to Veh). C) Serum and (D) lung IL-1β protein levels from Bort- or Veh-treated WT mice 1 week after urethane (n=6 mice per group). E) Schematic representation of the NF-κB inhibition protocol using BAY 11-7082 (BAY). In addition to urethane, WT mice were treated with IP injections of the specific NF-κB inhibitor BAY (10 mg/kg) or Veh. F) BAL cells in BAY- and Veh-treated WT mice at 1 week after urethane injection (n=8 mice per group). G) Number of Ly6G⁺ neutrophils and Ly6C⁺ monocytes in the lungs of BAY- or Veh-treated mice at 1 week after urethane injection (n=4-5 mice per group; *p<0.05 compared to Veh). H) IL-1β protein levels in the lungs of BAY- or Veh-treated mice at 1 week after urethane injection (n=8 mice per group). I) IL-8, (J) TNF, (K) IL-6, and (L) IL-1β protein levels in the plasma of NSCLC patients treated before (0hr) and 24hr after treatment with bortezomib (1 mg/m²) (n=28 patients; *p < 0.05 compared with 0 hr). M) Correlation analysis between progression-free survival and baseline plasma IL-1β protein levels in advanced NSCLC patients treated with bortezomib plus standard chemotherapy (p=0.026). See also Figures S3, S4 and Table S1.

Figure 7

IL-1β facilitates lung tumorigenesis by stimulating epithelial cell proliferation and supports resistance to bortezomib therapy. A) Schematic representation of IL-1 receptor antagonist (IL-1ra) treatment protocol. WT and IKKβ^Δmye mice were injected with a single dose of urethane and treated by osmotic pump delivery of 60 mg/kg/day of IL-1ra or PBS for the first 4 weeks. B) Number of AAH lesions per H&E-stained lung section harvested from IKKβ^Δmye mice at week 6 after injection of urethane (n=9 mice per group, *p < 0.05 compared with PBS). C) Lung tumors on H&E-stained lung sections from WT and IKKβ^Δmye mice cut at predetermined depths (5 sections per mouse, n=7 mice per group; *p<0.05 compared with PBS-treated IKKβ^Δmye mice). D) Number of PCNA⁺ cells per lung section (averaged from 25 sequential fields taken at 40× magnification) from IKKβ^Δmye mice harvested at week 6 after urethane injection (n=9 mice per group; *p<0.05 compared with PBS). E) Fold change of subcutaneous LLC tumor volume over 10 days of treatment with vehicle control, bortezomib, IL-1ra, or bortezomib plus IL-1ra (n=6-9 mice per group; *p<0.05 compared with control). F-H) Inducible Kras^G12D mice were treated with doxycycline (dox) for 4 weeks to develop lung tumors. (F) Percentage of Ly6G⁺ and Ly6C⁺ cells in the lungs of dox-inducible Kras^G12D mice treated for 1 additional week with Bort or Veh plus dox (*p<0.05 compared to Veh). (G) Representative photomicrographs and (H) Numbers of surface lung tumors in mice treated with dox alone for 4 weeks followed by 4 weeks of treatment with dox plus vehicle control, bortezomib, IL-1ra, or bortezomib plus IL-1ra (n=6-7 mice per group; *p<0.05 compared with control.