Sustained desensitization to bacterial Toll-like receptor ligands after resolution of respiratory influenza infection

Abstract

The World Health Organization estimates that lower respiratory tract infections (excluding tuberculosis) account for approximately 35% of all deaths caused by infectious diseases. In many cases, the cause of death may be caused by multiple pathogens, e.g., the life-threatening bacterial pneumonia observed in patients infected with influenza virus. The ability to evolve more efficient immunity on each successive encounter with antigen is the hallmark of the adaptive immune response. However, in the absence of cross-reactive T and B cell epitopes, one lung infection can modify immunity and pathology to the next for extended periods of time. We now report for the first time that this phenomenon is mediated by a sustained desensitization of lung sentinel cells to Toll-like receptor (TLR) ligands; this is an effect that lasts for several months after resolution of influenza or respiratory syncytial virus infection and is associated with reduced chemokine production and NF-kappaB activation in alveolar macrophages. Although such desensitization may be beneficial in alleviating overall immunopathology, the reduced neutrophil recruitment correlates with heightened bacterial load during secondary respiratory infection. Our data therefore suggests that post-viral desensitization to TLR signals may be one possible contributor to the common secondary bacterial pneumonia associated with pandemic and seasonal influenza infection.

Figure 1.

Long-term impairment of TLR5-dependent neutrophil recruitment after resolution of an influenza infection. BALB/c (A and C) or C57BL/6 (A and B) mice were infected i.n. with 50 HA of influenza X31 (post-flu, filled bars) or PBS as control (ctrl, open bars), left to recover for 4–6 wk, and challenged with 1 μg flagellin FliC. (A) The percentage and total number of macrophages and neutrophils in the lungs and airways was monitored by flow cytometry 6 h after i.n. instillation of flagellin or PBS as indicated. n = 5 mice/group. Data are representative of 6 experiments. (B) The number of neutrophils recruited to the airways is shown at different time points after flagellin instillation (n = 5 mice/time point). (C) Airway neutrophil recruitment 6 h after FliC treatment 4–6 wk (n = 13–15 mice/group) or 3 or 6 mo (n = 4–5 mice/group) after the initial Influenza infection. Error bars represent the mean ± the SEM.

Figure 2.

Reduced airway cell recruitment occurs after secondary bacterial challenge. Control (PBS-treated) or post-influenza mice (6 or 2 wk after influenza, as indicated) were inoculated with 1 μg LPS or 25 μg LTA (A) or were infected with 5 × 10⁶ CFU of P. aeruginosa PAK strain (B; n = 5–8 mice/group), 5 × 10⁶ CFU of group B Streptococcus (C; n = 6 mice/group), or 10⁵ CFU of S. pneumoniae (D; n = 6–15 mice/group). The cell number in the airways was monitored at the time points indicated. The number of bacteria was also evaluated in the lungs of infected mice (B–D). *, 3 d after infection, one control mouse out of 6 did not survive the infection (17%), whereas 3 out of 5 mice died in post-influenza mice (60% death). Error bars represent the mean ± the SEM.

Figure 3.

The recruitment defect in post-influenza mice is associated with an impaired TLR-induced cytokine production. (A) Airway and lung neutrophil recruitment in post-influenza mice (2 mo after influenza) or control mice (2 mo after PBS treatment) were compared after i.n. or i.v. administration of flagellin, 1 μg FliC, or after co-instillation of 1 μg of recombinant KC and MIP-2α. n = 4. (B) MIP-2α and KC levels were measured in the BAL fluid and lung homogenates of post-influenza or control mice 6 and 24 h after FliC administration. (C) Levels of other proinflammatory cytokines induced by TLR5 signaling in the airways of post-influenza compared with control mice. For B and C, n = 3–5 mice/time point; this is representative of 4 experiments. (D) Relative fold increase of KC and MIP-2α mRNA transcripts purified from the lungs of control and post-influenza mice at different time points after flagellin (1 μg) i.n. challenge. Data were normalized to levels of β-actin mRNA and are represented relative to the mRNA levels in PBS-treated mice. n = 3. Error bars represent the mean ± the SEM.

Figure 4.

Involvement of epithelial cells and AMs in post-influenza TLR desensitization in vivo. (A) AMs were isolated from post-influenza mice (6 wk after influenza) and control mice, and the level of TLR mRNA transcripts was assessed by real-time PCR. (B) Post-influenza mice (6 wk after influenza) and control mice were challenged i.n. with 1 μg FliC and BAL, and lungs were collected 1 h later. AMs were purified, and the level of different RNA transcripts, as indicated, was assessed by real-time PCR. Results are expressed as the ratio between the gene of interest and a housekeeping gene (18S). n = 3–5 mice/condition. Error bars represent the mean ± the SEM.

Figure 5.

Inhibition of TLR-induced NF-κB in AMs is responsible for reduced cell recruitment in post-influenza mice. (A) AMs isolated from control and post-influenza mice were stimulated for the indicated time with 1 μg/ml FliC and further stained for p65 (green) or a nuclear marker (propidium iodide, red). Colocalization of p65 and propidium iodide is visible in control AM after 60 min of flagellin activation (yellow). The percentage of cells with p65 translocation is shown. n = 6 mice/condition. Bar, 10 μm. (B) AMs were stimulated overnight with Flic and LPS at indicated concentrations and KC and MIP-2α were measured in the supernatant. Results are expressed as fold increase compared with medium level to normalize the data obtained from different cultures obtained from individual (n = 3) or pooled mice (n = 10). (C) Irradiated wild type C57BL/5 mice were reconstituted with bone marrow from mice expressing the DT receptor under the control of the CD11c promoter. These mice were then infected with influenza or received PBS as control. 2 wk later, half of the mice in each group received an intratracheal injection of DT leading to the depletion of AM and further left 3 wk for naive AM to reconstitute the airways. Nondepleted mice received PBS as control (no depletion). Neutrophil airway recruitment in control and after influenza mice was then evaluated 6 h after i.n. flagellin challenge in DT-treated and nontreated mice (right). The number of neutrophils before flagellin treatment was comparable between all groups and similar to naive mice (not depicted). Error bars represent the mean ± the SEM.