Toll-Like Receptor 2 Mediates Fatal Immunopathology in Mice During Treatment of Secondary Pneumococcal Pneumonia Following Influenza

Infectious Agents

Influenza virus A/Puerto Rico/8/34 (PR8) was grown in Madin-Darby canine kidney cells. S. pneumoniae strain A66.1 (type 3) was transformed with the lux operon (Xenogen Corp) [23]. Inactivation was accomplished by a 5-minute suspension in 90% ethanol. An azithromycin-resistant variant of the luciferase-containing A66.1 strain was generated by serial passage in increasing concentrations of the antibiotic under a protocol approved by the Institutional Biosafety Committee of St Jude. The minimum inhibitory concentrations (MICs) for the susceptible and resistant strains were 0.023 and 0.023 μg/mL for ampicillin, and 0.064 and 5 μg/mL for azithromycin, respectively. The azithromycin-resistant, ampicillin-sensitive phenotype was confirmed by culture in Todd Hewitt broth in the presence of ampicillin (100 μg/mL), azithromycin (3 or 100 μg/mL), or combinations followed by plating of aliquots at intervals to determine colony counts. An isogenic, pneumolysin-deficient mutant of A66.1 (pln-) was created by complete gene replacement with an erythromycin resistance cassette via the SOEing method (primer sequences available on request) [24]. Complete deletion was confirmed by sequencing.

Mice

Seven- to 8-week-old female BALB/c, C3H/OuJ, and C3H/HeJ mice, as well as male and female C57Bl/6 mice (wild type and lacking Toll-like receptor [TLR] 4 or TLR2), were obtained from Jackson Laboratories and used in biosafety level 2 facilities in these studies in a manner in accordance with the guidelines of the Committee on Care and Use of Laboratory Animals. All experimental procedures were done under general anesthesia with inhaled isoflurane 2.5% (Baxter Healthcare Corp) and were approved following review by the St Jude Institutional Animal Care and Use Committee.

Infectious Model

Infectious agents were administered intranasally in a volume of 100 μL (50 μL per nostril) to anesthetized mice as described previously [16]. Influenza virus was given at a dose of 30 (BALB/c, C3H-derived strains) or 60 (C57Bl/6-derived strains) median tissue culture infective doses (TCID₅₀). Except where otherwise noted, S. pneumoniae was given 7 days later at a dose of 200 or 700 colony-forming units (CFU). Following secondary bacterial infection, bioluminescent imaging was used to follow development and progression of pneumonia using defined parameters to ensure experimental groups were balanced with all mice at an early stage of pneumonia as described [16, 20, 23].

Treatment of SBP

Treatment was started upon identification of pneumonia by bioluminescent imaging as described [22]. Mice were given antibiotics (Sigma-Aldrich) intraperitoneally twice daily in divided doses (ampicillin 200 mg/kg/day; clindamycin 120 mg/kg/day; vancomycin 60 mg/kg/day) or once daily (azithromycin 10 mg/kg/day first dose, 5 mg/kg/day thereafter) for a total of 7 days or were mock-treated with the diluent, phosphate-buffered saline (PBS). This dose of azithromycin is predicted to result in lung tissue levels above the MIC for the sensitive strain but below the MIC for the resistant strain [25]

Histopathology

Microscopic evaluation of lungs was performed by an experienced veterinary pathologist (K. L. B.) blinded to study purpose and design and to group composition. A semiquantitative grading scheme was utilized to score 2 parameters, the overall character of the pneumonic process and the pathology specifically observed in the interstitum and terminal airways as previously described [20, 21].

Statistical Analyses

Comparison of survival between groups of mice was done with the log-rank χ² test on the Kaplan-Meier survival data. Comparisons of histopathology scoring between groups were done using analysis of variance (ANOVA). A P value of <.05 was considered significant for these comparisons. SigmaStat for Windows software, version 3.11 (SysStat Software) was utilized for all statistical analyses.

Outcomes Based on Mode of Antibiotic-Mediated Killing

A model of treatment of SBP based on earlier work [16, 22, 26] was developed for this study. Female BALB/c mice infected with a sublethal dose of influenza virus (30 TCID₅₀) followed 7 days later with PBS, or mock-infected with PBS followed 7 days later with a sublethal dose of S. pneumoniae (200 CFU), had 100% survival (Figure 1). When pneumococcal challenge followed influenza virus infection, however, pneumonia detectable through bioluminescent imaging resulted, and all superinfected mice died within 4 days of exposure to the second agent. Similar results were obtained in CH3/OuJ and C57Bl/6 mice, although slightly larger doses of both infectious agents were needed for C57Bl/6 mice (60 TCID₅₀ of virus and 700 CFU of S. pneumoniae). Treatment with the cell wall–active agents ampicillin or vancomycin resulted in rapid elimination of bacteria from the lungs of mice by bioluminescent imaging (data not shown), but only 30%–50% of mice survived. Treatment with the protein synthesis inhibitors clindamycin or azithromycin resulted in significantly better outcomes, with 100% survival seen in the azithromycin group.

The lungs of mice from treatment groups were examined microscopically. Mice infected only with a sublethal dose of S. pneumoniae had mild perivascular infiltration of lymphocytes but no pathologic changes suggestive of pneumonia (Figure 2). Virus-infected mice exhibited hypertrophy and hyperplasia of epithelial cells lining the airways, accompanied by a moderate lymphocytic infiltrate and edema. Lungs from mice coinfected with influenza virus and bacteria followed by treatment with ampicillin showed a severe fibrinopurulent broncho-interstitial pneumonia. The inflammatory response was characterized by a copious neutrophilic infiltrate accompanied by edema, fibrin deposition, and necrosis of the alveolar walls and bronchiolar epithelium at the alveolar bronchiolar junction. In mice treated with azithromycin, however, alveolar wall necrosis and fibrin deposition were not seen. The broncho-interstitial pneumonia was less severe and was characterized by an inflammatory response composed predominately of foamy macrophages within the alveolar spaces and only minimal numbers of neutrophils (Figure 2). Overall, the character and severity of the pneumonic process resembled that engendered by virus alone and was significantly less in azithromycin-treated animals than in ampicillin-treated animals.

We hypothesized that the superior outcomes seen with azithromycin derived from its purported anti-inflammatory activity [27]. We reasoned that if this anti-inflammatory property was independent of the mechanism of antibacterial killing, then it should still be able to contribute to rescue of superinfected mice with severe lung inflammation in the setting of resistance. Therefore, we developed an azithromycin-resistant version of the pneumococcal strain being used in our model of SBP and repeated our antibiotic treatment studies (Figure 3). As before, untreated mice succumbed to infection within a few days of secondary challenge, while ampicillin therapy only modestly improved survival (9 of 18 [50%] survived), despite complete elimination of bacteria from the lungs of mice by bioluminescent imaging (data not shown). Azithromycin was unable to rescue mice superinfected with the resistant strain due to overgrowth of bacteria in the lungs (0 of 5 [0%] survived). However, the addition of azithromycin to ampicillin resulted in survival of 16 of 18 mice (89%), an outcome significantly better than ampicillin alone, despite lack of additive effects on killing in vitro. We conclude that the anti-inflammatory activity of azithromycin is independent of its antibacterial effect and can contribute to improved outcomes from SBP, even when it is caused by resistant strains of S. pneumoniae.

Induction of TLR4 Through Release of Pneumolysin From Lysed Bacteria

Expression of the pneumococcal cytotoxin pneumolysin has been associated with activation of TLR4 [28]. We hypothesized that release of pneumolysin from lysed bacteria following exposure to β-lactams was driving the inflammatory response that resulted in mortality during treatment of SBP. If our hypothesis was correct, we reasoned that either knock-out of pneumolysin from the bacteria or lack of TLR4 expression in the mice would prevent the fatal inflammatory burst during ampicillin treatment of SBP. We therefore generated a pneumolysin-deficient mutant (pln-) of the A66.1 strain used in the model and infected TLR4-sufficient (C3H/OuJ) and TLR4-deficient (C3H/HeJ) mice with and without antecedent influenza virus infection. The pln- mutant was attenuated in mice compared with the wild type (WT) (Figure 4 and data not shown). In the presence of virus, however, all mice developed pneumonia and both the pln- mutant and the WT were lethal in 100% of mice when 1 × 10⁴ CFU of bacteria was used. Expression of TLR4 provides protection against pneumococcus in this mouse model, since the WT was attenuated in TLR4-sufficient C3H/OuJ mice, but caused 80% mortality in TLR4-deficient C3H/HeJ mice. However, this attenuation was not based on TLR4 recognition of pneumolysin, as the pln- mutant showed a similar pattern of disease with increased virulence in the absence of TLR4. We conclude from these experiments that pneumolysin is important for pathogenesis in the lungs of mice, but the relatively greater virulence of S. pneumoniae seen in mice deficient for TLR4 is not due to a lack of TLR4 interactions with pneumolysin.

Having established the parameters of the model using the pln- mutant and the TLR4-deficient C3H/HeJ mouse strain, we next addressed our main hypothesis. TLR4-sufficient and -deficient influenza-infected mice were challenged with WT or the pln- mutant and then treated with ampicillin following development of pneumonia. Lack of TLR4 did not alter disease in mice preinfected with influenza virus; no mice survived coinfection at these doses (Figure 4). Treatment with ampicillin rescued between 29% and 50% of superinfected mice, with no statistically significant differences evident between treated groups regardless of whether the WT or pln- mutant was utilized. To confirm this finding, we repeated this experiment with WT bacteria in C57Bl/6 mice and TLR4^−/− mice. In both the parental strain C57Bl/6 and TLR4^−/− mice, only 4 of 13 (31%) animals survived treatment of pneumococcal SBP following influenza (Figure 5). We conclude from these data that neither release of pneumolysin from lysed bacteria nor activation of TLR4 is responsible for the inflammatory burst and accompanying immunopathology observed during treatment of SBP with β-lactam agents.

Importance of TLR2 for Mortality

We next examined the importance of a second Toll-like receptor, TLR2, during lysis of pneumococci in the setting of influenza virus infection. TLR2 recognizes cell wall breakdown products of S. pneumoniae and thereby participates in innate immune responses through induction of inflammation [29]. C57Bl/6 wild-type or TLR2^−/− mice were infected with influenza, then challenged with pneumococci. No differences in survival were seen in untreated WT or TLR2^−/− mice. Similarly to previous experiments, 9 of 18 (50%) WT C57Bl/6 mice survived infection and treatment with ampicillin (Figure 5). However, lack of TLR2 provided significant protection as 16 of 17 (94%) mice survived. We conclude that TLR2 contributes to the fatal inflammatory response observed in mice during treatment of SBP with β-lactams.

The mechanism underlying worse outcomes in TLR2-sufficient mice is of obvious interest. We used inactivated S. pneumoniae to probe the relationship between provision of cell wall–associated PAMPs and disease in the context of influenza. Administration of 1 × 10⁸ CFU equivalents of inactivated bacteria was assessed in mice with and without antecedent influenza virus infection. Inactivated bacteria alone did not cause significant disease, and 100% of mice survived (Figure 6). Preinfection with influenza virus, however, primed mice for significant mortality (2 of 8 mice [25%] survived). This experiment was repeated in WT and TLR2^−/− mice. Only 4 of 9 (44%) WT C57Bl/6 mice survived exposure to inactivated bacteria in the setting of influenza, while 7 of 9 (77%) TLR2^−/− mice survived. Thus, live bacteria are not required either for disease in the secondary pneumonia model following pneumonia or to observe TLR2-mediated mortality during treatment. These data suggest that TLR2 recognition of PAMPs contained in inactivated bacteria, likely cell wall breakdown products, is responsible for immunopathology and mortality during β-lactam–based treatment regimens for SBP following influenza.