Mucosal Administration of Flagellin Protects Mice from Streptococcus pneumoniae Lung Infection

Abstract

Streptococcus pneumoniae is a major cause of pneumonia in infants and the elderly. Innate defenses are essential to the control of pneumococcal infections, and deficient responses can trigger disease in susceptible individuals. Here we showed that flagellin can locally activate innate immunity and thereby increase the resistance to acute pneumonia. Flagellin mucosal treatment improved S. pneumoniae clearance in the lungs and promoted increased survival of infection. In addition, lung architecture was fully restored after the treatment of infected mice, indicating that flagellin allows the reestablishment of steady-state conditions. Using a flagellin mutant that is unable to signal through Toll-like receptor 5 (TLR5), we established that TLR5 signaling is essential for protection. In the respiratory tract, flagellin induced neutrophil infiltration into airways and upregulated the expression of genes coding for interleukin 6 (IL-6), tumor necrosis factor alpha (TNF-α), CXCL1, CXCL2, and CCL20. Using depleting antibodies, we demonstrated that neutrophils are major effectors of protection. Further, we found that B- and T-cell-deficient SCID mice clear S. pneumoniae challenge to the same extent as immunocompetent animals, suggesting that these cell populations are not required for flagellin-induced protection. In conclusion, this study emphasizes that mucosal stimulation of innate immunity by a TLR not naturally engaged by S. pneumoniae can increase the potential to cure pneumococcal pneumonia.

Streptococcus pneumoniae (pneumococcus) causes respiratory infections among infants and the elderly worldwide (40, 44). Capsular polysaccharide is the main virulence factor, and its composition defines 91 serotypes of pneumococcus (42). Certain serotypes colonize the human nasopharynx asymptomatically, representing a reservoir for interindividual transmission of the bacteria. In some individuals, colonization may progress to pneumococcal pneumonia and invasive disease (19, 36). In contrast, other serotypes, such as serotype 1, are rarely associated with colonization but cause invasive infections (28).

Activation of innate defenses is essential for the control of pneumococcal infection (1, 20). Toll-like receptor 2 (TLR2), TLR4, and TLR9, as well as the adaptor MyD88, participate in the early detection and clearance of pneumococcus in the lungs (reviewed in reference 42). The cytosolic receptors Nod1 (nucleotide-binding oligomerization domain 1) and Nod2 are also involved in the recognition of pneumococci (29). TLR signaling activates mucosal innate responses that culminate with the recruitment of phagocytes, such as polymorphonuclear neutrophils (PMN) and macrophages, and the production of microbicidal agents (for a review, see reference 8). This process triggers rapid eradication of the pathogen by phagocytosis as well as by extracellular killing. In MyD88-deficient animals, S. pneumoniae is unable to intrinsically trigger any PMN recruitment into the airways, and animals thus have increased susceptibility to pneumonia (1). The contribution of TLR signaling in humans has been highlighted by a recent study showing that some MyD88 polymorphisms are associated with increased susceptibility to pneumococcal infection (43).

The modulation of immunity by the activity of innate receptors to elicit protective responses against infections is an emerging concept (6, 35). The rationale is to promote innate responses that greatly exceed in magnitude, quality, and dynamics the innate response triggered by the pathogen itself. The effectiveness of TLR agonists for therapeutic treatment of infectious diseases has been demonstrated in several animal models, including models of respiratory infections (6, 21, 35). TLR5 senses bacterial flagellins, which are the main constituents of flagella. Various cells of the pulmonary tract, including the epithelial cells (14, 33), express TLR5, and mucosal administration of flagellin induces MyD88-dependent signaling, characterized by the swift production of various proinflammatory cytokines and chemokines (3, 10, 15, 16, 27, 33), as well as by rapid and heavy neutrophil infiltration into the airways (2, 10, 16). Although S. pneumoniae does not have flagella, we hypothesized that activation of TLR5 signaling may promote new and appropriate protective innate defenses against ongoing acute pneumococcal infections. Here we report that local stimulation of innate immunity by flagellin from Salmonella enterica serovar Typhimurium blocks the progression of pneumococcal pneumonia in mice.

MATERIALS AND METHODS

Bacterial preparation.

Streptococcus pneumoniae serotype 1 (clinical isolate E1586) was obtained from the National Reference Laboratory, Ministry of Health, Montevideo, Uruguay (39). Working stocks were prepared as follows. Todd-Hewitt yeast broth (THYB) (Sigma-Aldrich, St. Louis, MO) was inoculated with fresh colonies of S. pneumoniae grown on blood agar plates, and the culture was incubated at 37°C until it reached an optical density at 600 nm (OD₆₀₀) of 0.7 to 0.9. Cultures were stored at −80°C in THYB plus 12% (vol/vol) glycerol for as long as 3 months. For mouse infection, working stocks were thawed, washed with sterile physiological saline solution (saline), and diluted to the appropriate concentration. The number of bacteria in stocks was confirmed by plating serial dilutions onto blood agar plates.

Proteins.

Native flagellin (FliC) from Salmonella enterica serovar Typhimurium SIN22 and recombinant flagellin mutants (FliC_Δ174-400 and FliC_{Δ174-400/89-96*}) were prepared as described previously (27). Both mutants lack the hypervariable domain of flagellin, and FliC_{Δ174-400/89-96*} also carries a further amino acid substitution (amino acids [aa] 89 to 96) that prevents TLR5 signaling (27). All proteins contained low levels of lipopolysaccharide (LPS) (less than 30 pg LPS per μg, as determined with the Limulus assay). In some experiments, trypsin-hydrolyzed FliC (FliC/T) was used as a control. Native FliC was heated for 5 min at 65°C before use to ensure that the proteins were mostly monomers. The dose of flagellin used was 1 μg for all the experiments, since we and others have previously shown that administration of this dose by the intranasal (i.n.) route is optimal for eliciting strong local immune activation (15, 16). Further, unless otherwise specified, FliC, FliC/T, FliC_Δ174-400, or FliC_{Δ174-400/89-96*} was always coadministered i.n. with the S. pneumoniae suspension. To exclude any direct effect of flagellin on bacterial viability, viable counts were determined prior to and after the incubation of S. pneumoniae with the same concentration of flagellin used for the in vivo assay. There were no significant differences between the number of bacteria recovered after incubation with flagellin and that recovered under the control condition.

Animal infections.

Female mice of the BALB/c, C57BL/6J, and outbred NMRI strains (6 to 8 weeks old) were obtained from the National Division of Veterinary Laboratories (Uruguay) or the Janvier Laboratories (France). Female SCID mice (C.B-17 SCID) were obtained from the Institut Pasteur de Lille breeding facilities. These mice are characterized by the lack of B and T lymphocytes and by agammaglobulinemia. Animals were maintained in individually ventilated cages and were handled in a vertical laminar flow cabinet (class II A2; Esco, Hatboro, PA) for infection. All experiments complied with current national and institutional regulations and ethical guidelines (CHEA [Universidad de la República, Montevideo, Uruguay] and A59107 [Institut Pasteur de Lille]). Mice were anesthetized by intraperitoneal (i.p.) injection of 2.2 mg ketamine (Richmond Vet Pharma, Grand Bourg, Buenos Aires, Argentina) plus 0.11 mg xylazine (Portinco, Montevideo, Uruguay) in a total volume of 200 μl or by inhalation of isoflurane (Belamont, SAS, France) using a non-rebreathing anesthesia system (DRE Compact 150; DRE Veterinary, Louisville, KY). Bacteria and flagellins were administered into the nostrils of mice in 20 to 50 μl of saline. The survival of mice was recorded daily.

For depletion of granulocytes, 100 μg of anti-Gr-1 (RB6-8C5) or an isotype control (HB152) was administered i.p. 24 h before i.n. challenge with S. pneumoniae (24). The anti-Gr1 injection was found to deplete 96.8% ± 1.2% of PMN in bronchoalveolar lavage (BAL) fluid after intranasal flagellin treatment.

Determination of bacterial loads in lungs and spleens.

Lungs and spleens were collected at selected time points after intranasal challenge and were homogenized with an Ultra Turrax homogenizer (IKA-Werke, Staufen, Germany). Viable counts were determined by plating serial dilutions onto blood agar plates.

Quantitative reverse transcription-PCR (qRT-PCR).

Lungs were homogenized in Trizol reagent (Invitrogen, CA) with an Ultra Turrax homogenizer and were stored at −80°C. RNA was extracted according to the manufacturer's instructions. Prior to cDNA synthesis, 1 μg total RNA was treated with DNase I (Invitrogen), and first-strand cDNA synthesis was carried out using random primers (Invitrogen) and Moloney murine leukemia virus (M-MLV) reverse transcriptase (Invitrogen). Real-time PCR was performed using a QuantiTect SYBR green PCR kit (Qiagen, Hilden, Germany) in a Rotor-Gene 6000 real-time rotary analyzer (Corbett, Sydney, Australia) according to the following protocol: 15 min at 95°C, followed by 40 cycles at 95°C for 15 s and 60°C for 1 min. The primers are listed in Table 1 and were used at a final concentration of 0.9 μM. The expression of the gene of interest was normalized by using β-actin as a housekeeping gene. Results are presented as fold increases in mRNA levels over those for the saline-treated group.

TABLE 1.

Sequences of primers used for qRT-PCR

Gene name	Forward primer (5′-3′)	Reverse primer (5′-3′)	Product length (bp)
Actb	GCTTCTTTGCAGCTCCTTCGT	CGTCATCCATGGCGAACTG	68
Cxcl1	CTTGGTTCAGAAAATTGTCCAAAA	ACGGTGCCATCAGAGCAGTCT	84
Cxcl2	CCCTCAACGGAAGAACCAAA	CACATCAGGTACGATCCAGGC	72
Ccl20	TTTTGGGATGGAATTGGACAC	TGCAGGTGAAGCCTTCAACC	69
Il4	ACAGGAGAAGGGACGCCAT	GAAGCCCTACAGACGAGCTCA	95
Il6	GTTCTCTGGGAAATCGTGGAAA	AAGTGCATCATCGTTGTTCATACA	78
Il17a	CTCCAGAAGGCCCTCAGACTAC	GGGTCTTCATTGCGGTGG	69
Il17f	CCCATGGGATTACAACATCACTC	CACTGGGCCTCAGCGATC	66
Il23a	GCCCCGTATCCAGTGTGAAG	CGGATCCTTTGCAAGCAGAA	78
Tgfb1	GCTGAACCAAGGAGACGGAAT	GAGTTTGTTATCTTTGCTGTCACAAGA	76
Tnf	CATCTTCTCAAAATTCGAGTGACAA	CCTCCACTTGGTGGTTTGCT	63

Determination of PMN infiltration into the airways and lungs.

For BAL sampling, the trachea was cannulated, and 1 ml of phosphate-buffered saline (PBS) plus 1 mM EDTA was instilled six times and recovered by gentle aspiration; this process was repeated twice. Cells were suspended in FACS-EDTA buffer (PBS, 0.1% azide, and 1% bovine serum albumin [Sigma-Aldrich] plus 5 mM EDTA). Lung cells were isolated after collagenase and DNase treatment as previously described (34) and were filtered through a 40-μm-pore-size cell strainer. Immune cells were separated in a two-layer Percoll (Sigma-Aldrich) gradient. Briefly, cells were suspended in 35% isotonic Percoll solution, carefully placed on top of a 70% isotonic Percoll solution, and centrifuged for 30 min at 2,600 × g and room temperature without a brake. The top ring of cells, corresponding mostly to epithelial cells, was discarded, and immune cells were recovered from the ring of cells closest to the 70% Percoll layer. Cells were filtered using a 100-μm-pore-size cell strainer, washed, and stained for fluorescence-activated cell sorter (FACS) analysis. Neutrophils were identified by forward scatter-side scatter (FSC-SSC); by positive staining either for anti-Ly6G (clone 1A8) and anti-Ly6C (clone HK1.4) or for anti Gr-1 (clone R6B8c5), as well as for anti-CD11b (clone M1/70); and by negative staining for anti-CD11c (clone HL3). Antibodies were obtained either from BD Biosciences, from BioLegend (San Diego, CA), or from AbD Serotec (United Kingdom). After fixation with 4% paraformaldehyde (PFA), flow cytometry data were acquired on a FACSCalibur cytometer with CellQuest software (version 3.3; BD Biosciences).

Histological analysis.

Lungs were fixed in 4% formalin (Sigma-Aldrich) for 24 h and were then embedded in paraffin. Lung blocks were sectioned at a thickness of 5 μm using a Leica microtome (Leica Microsystems, Wetzlar, Germany) and were adhered to silanized slides. Three mice per group were analyzed. Six sections from each mouse were stained with hematoxylin and eosin and were analyzed in a double-blind manner by using a Nikon Eclipse 80i microscope and a Nikon DS-Ri1 digital camera. Images were processed using NIS-Elements BR (version 3.0) software (Laboratory Imaging).

Statistical analysis.

A log rank (Mantel-Cox) test was performed for analysis of survival curves. For comparison between two groups, the Mann-Whitney test was performed. P values of <0.05 were considered significant in all cases. Statistical analysis was carried out using GraphPad Prism (GraphPad Software, San Diego, CA).

RESULTS

Intranasal delivery of flagellin protects mice against a lethal challenge with S. pneumoniae.

We first determined the minimal dose of S. pneumoniae that causes 100% mortality in BALB/c mice upon intranasal (i.n.) administration. Animals were infected with increasing doses of a clinical isolate of S. pneumoniae serotype 1, and survival was assessed daily (data not shown). We defined 4 × 10⁵ CFU as the minimal lethal dose (MLD) that kills all animals within 72 to 120 h.

The capacity of flagellin to control pneumococcal pneumonia was then assessed by comparing the survival of mice challenged intranasally with S. pneumoniae to that of mice instilled with flagellin (FliC) and S. pneumoniae. As a control, mice were also challenged with S. pneumoniae and flagellin that had been hydrolyzed with trypsin previously (FliC/T). As shown in Fig. 1A, FliC-treated mice had a survival rate of 75%, while untreated or FliC/T-treated animals died within 3 to 4 days after challenge. The protection induced by flagellin ranged from 75 to 100% in different independent experiments.

FIG. 1.

Flagellin protects BALB/c mice against a lethal respiratory challenge with S. pneumoniae. BALB/c mice (n = 8) were infected i.n. with 4 × 10⁵ CFU of S. pneumoniae (Sp) serotype 1 in saline alone (filled squares) or in saline supplemented either with 1 μg flagellin (FliC) (open circles) or with 1 μg trypsin-digested flagellin (FliC/T) (filled circles). (A) The survival of mice was monitored daily. The survival of the FliC-treated group was statistically significantly different from that of the saline group and the FliC/T-treated group. The results are representative of 1 out of 3 experiments. (B and C) CFU counts were determined in the lungs (B) or spleens (C) of mice (n = 6) at days 1, 2, and 7 after challenge with S. pneumoniae alone (filled squares) or S. pneumoniae plus flagellin (open circles). The solid lines represent means, and the dashed lines represent the CFU detection limit. The daggers indicate that all animals in the group were dead by day 7. Significant differences between groups (P < 0.05) are indicated by asterisks. Results are representative of 1 out of 2 experiments.

Coadministration of flagellin with S. pneumoniae challenge resulted in significant reductions of bacterial counts in the lungs within the first 24 h after challenge compared with those for animals that received S. pneumoniae alone. Moreover, bacterial loads in the lungs continued to increase until death in all mice that received no treatment, while in flagellin-treated mice, bacteria were cleared from the lungs by day 2 after infection and could no longer be detected at the local level (Fig. 1B). The capacity of flagellin to protect against systemic infection was also evaluated by assessing bacterial loads in the spleens of treated and untreated mice at different time points after challenge. As shown in Fig. 1C, high numbers of S. pneumoniae organisms were detected in the spleens of all untreated animals as early as 48 h after infection. In contrast, bacteria could not be detected in the spleens of flagellin-treated mice at any time point (Fig. 1C).

We also evaluated whether flagellin could exert a protective response against pneumococcal infection when administered before or after the infection. All animals receiving flagellin intranasally 12 to 24 h before pneumococcal challenge survived, while all control mice died by day 4. To assess the therapeutic value of flagellin administration, BALB/c mice were infected with S. pneumoniae, and 24 h later, they either were treated with 1 μg of flagellin by the i.n. route or remained untreated. The level of protection among animals receiving flagellin ranged from 60 to 100%, while all untreated mice died of infection. As shown in Fig. 2, therapeutic administration of flagellin induced significant reductions of bacterial loads in the lungs (Fig. 2A) as well as in the spleen (Fig. 2B) as detected 48 h after infection (24 h after flagellin administration). By day 7, S. pneumoniae was no longer detectable in the lungs or spleens of surviving animals. Therefore, flagellin shows prophylactic as well as therapeutic effects in pneumococcal pneumonia.

FIG. 2.

Therapeutic effect of FliC after S. pneumoniae challenge. BALB/c mice were infected intranasally with 4 × 10⁵ CFU of S. pneumoniae (filled squares); 24 h after challenge, 1 μg of FliC was administered i.n. to one group of animals (open circles). CFU counts were determined in the lungs (A) or spleens (B) at day 2 and day 7 after challenge with S. pneumoniae. Solid and dashed lines represent means and the CFU detection limit, respectively. The dagger indicates that all animals in the group were dead by day 7. Significant differences between groups (P < 0.05) are indicated by asterisks. Results are representative of 1 out of 2 experiments.

The capacity of flagellin to induce protection was also assessed in C57BL/6 mice and in mice of the outbred strain NMRI. The MLD of S. pneumoniae serotype 1 was found to be 2 × 10⁶ CFU for both strains, and flagellin-mediated protection was evaluated with 5× MLD. Administration of flagellin 12 h before bacterial challenge induced 80% protection in C57BL/6 mice; similarly, 100% protection was achieved in NMRI animals when flagellin was administered 32 h to 6 h before challenge. Flagellin was also protective when coadministered with S. pneumoniae to mice of the C57BL/6 and NMRI strains, although to a lower extent (40%). Taken together, these results show that flagellin treatment is protective in different mouse strains.

We next addressed whether TLR5 signaling is necessary for the protection elicited by flagellin treatment. For this purpose, we used the recombinant flagellins FliC_Δ174-400 and FliC_{Δ174-400/89-96*} (27). FliC_Δ174-400 has the same capacity to promote mucosal TLR5 signaling as native flagellin, whereas FliC_{Δ174-400/89-96*} carries mutations that prevent TLR5 signaling. While all mice that received FliC_Δ174-400 and S. pneumoniae survived the challenge, none of those receiving the mutant FliC_{Δ174-400/89-96*} did so (Fig. 3). These results strongly suggest that TLR5 signaling is required for protection.

FIG. 3.

TLR5 signaling is required for flagellin-mediated protection against S. pneumoniae infection. BALB/c mice (n = 8) were infected i.n. with 4 × 10⁵ CFU of S. pneumoniae (Sp) serotype 1 in saline either alone (filled square) or supplemented either with 1 μg FliC_Δ174-400 (open circles) or with 1 μg FliC_{Δ174-400/89-96*}, lacking the TLR5-signaling motif (filled circle). Survival of mice was recorded daily. The survival of the FliC_Δ174-400-treated group was statistically significantly different from that of the untreated and FliC_{Δ174-400/89-96*}-treated groups (P < 0.05). Results are representative of 1 out of 2 experiments.

Flagellin treatment promotes proinflammatory gene expression and exacerbates transient cellular infiltration into the lungs during pneumococcal pneumonia.

We then analyzed whether flagellin treatment modifies the lungs' transcriptional response to pneumococcal infection. Mice were challenged with S. pneumoniae alone or with S. pneumoniae plus flagellin as before. Another group received flagellin alone as a control. Twenty-four hours after treatment and infection, lungs were harvested in order to analyze the expression of selected genes by qRT-PCR. Administration of flagellin alone or in combination with S. pneumoniae provoked a dramatic increase in Cxcl1, Cxcl2, and Ccl20 mRNA levels over those observed with pneumococcal challenge alone (Fig. 4). Flagellin treatment also increased the expression of Tnf, although the difference was consistent, it was not statistically significant. Expression of Il6 was increased in animals that were challenged and treated with flagellin but not in those that received flagellin or S. pneumoniae alone, suggesting a synergistic effect of flagellin and pneumococcal infection on Il6 expression (Fig. 4). Among all groups, mRNA levels of the Tgfb1, Il17a, Il17f, Il23, and Il4 genes remained unchanged from those for mock-treated animals (data not shown).

FIG. 4.

Flagellin upregulates chemokine and cytokine gene expression in the lungs of infected animals. BALB/c mice (n = 5) were instilled i.n. either with saline, with 1 μg flagellin (FliC), or with 4 × 10⁵ CFU of S. pneumoniae (Sp) serotype 1 either in saline alone or in saline supplemented with 1 μg FliC. Mice were sacrificed 24 h after challenge, and the expression of the Cxcl1, Cxcl2, Ccl20, Il6, and Tnf genes was assessed by qRT-PCR on total lung RNA. Values were normalized to that of β-actin mRNA, and the results are expressed relative to mRNA levels in saline-treated mice. Data are means; error bars represent standard errors of the means. Significant differences between groups (P < 0.05) are indicated by asterisks;. Results are representative of 1 out of 3 experiments.

To assess if the expression of proinflammatory genes correlated with inflammation and cellular infiltration into the airways, we performed histological analysis of lung tissue obtained 24 h after flagellin treatment and infection. As shown in Fig. 5, S. pneumoniae induced moderate cellular infiltration, restricted to certain bronchioles and some perivascular areas close to these bronchioles (Fig. 5D). In contrast, flagellin treatment, alone or together with pneumococci, induced edema, as well as heavy infiltration of cells affecting not only perivascular and peribronchial regions but also some areas of the surrounding lung parenchyma (Fig. 5B and E). Remarkably, at day 7, lungs from mice that had received flagellin and pneumococci showed complete resolution of the inflammation, with no apparent cellular infiltration or edema (Fig. 5C). These results suggest that flagellin induces a strong but transient inflammatory response that promotes the clearance of bacteria without causing any permanent alteration of lung morphology or function.

FIG. 5.

Flagellin promotes heavy cell infiltration and lung remodeling and induces rapid restoration of tissue architecture. Mouse lungs were obtained at day 1 or day 7 after intranasal instillation of saline (A and F), 4 × 10⁵ CFU of S. pneumoniae serotype 1 (D), S. pneumoniae plus 1 μg flagellin (B and C), or flagellin alone (E) and were stained with hematoxylin and eosin. B, bronchiole; BV, blood vessel. Thin arrows indicate edema; arrowheads indicate cellular infiltration. On day 1, edema and recruitment of inflammatory cells were evident in peribronchial and perivascular regions in flagellin-treated animals. Infected but untreated animals showed less inflammatory-cell infiltration and edema at 24 h than flagellin-treated and infected mice. Flagellin-treated animals showed resolution of inflammation by day 7. Total magnification, ×200 for panels and ×1,000 for insets. The images shown are representative of the analysis of 6 sections for each individual and 3 mice per experimental group.

Flagellin-induced protection requires Gr-1-expressing cells but is independent of B and T lymphocytes.

Recruitment of neutrophils into the airways is a hallmark of both pneumococcal infection and nasal flagellin treatment (15, 18), and here we showed that flagellin treatment and infection activated the expression of genes involved in neutrophil recruitment (Fig. 4). Thus, we aimed at comparing the kinetics of neutrophil infiltration in animals challenged with S. pneumoniae and either left untreated or treated with flagellin. BAL specimens were collected at different time points after challenge and were stained with anti-Gr-1, anti-CD11b, and anti-CD11c antibodies. As shown in Fig. 6A, pneumococcal challenge induced the recruitment of PMN in all animals. However, mice treated with flagellin at the time of the challenge showed a more rapid and pronounced infiltration of PMN into the airways than mice challenged with S. pneumoniae alone. At 24 h, PMN infiltration peaked in both groups, and the difference between groups was maximal. However, by 48 h, the numbers of PMN were no longer significantly different between groups. Thus, coadministration of flagellin with pneumococci promoted a rapid and transient recruitment of a high number of neutrophils into the airways.

FIG. 6.

Flagellin-mediated neutrophil recruitment is required for protection against pneumonia. (A) Kinetics of PMN infiltration. BALB/c mice (n = 3) were infected i.n. with 4 × 10⁵ CFU of S. pneumoniae (Sp) serotype 1 in saline alone (filled squares) or in saline supplemented with 1 μg flagellin (FliC) (open circles). Animals were sacrificed at the indicated time points after challenge. The total number of cells in bronchoalveolar lavage fluid was determined, and the number of neutrophils (defined as Ly6C⁺ Ly6G⁺ CD11b⁺ cells) was determined by flow cytometry. (B) Depletion of PMN abrogates FliC-mediated protection. BALB/c mice (n = 6) were injected with 100 μg of an anti-Gr-1 monoclonal antibody or a control antibody. Animals were infected i.n. 24 h later with 4 × 10⁵ CFU of S. pneumoniae serotype 1 in saline alone (squares) or in saline supplemented with 1 μg flagellin (FliC) (circles). Survival was recorded daily. The survival of the nondepleted FliC-treated group was statistically significantly different from that of depleted FliC-treated animals or nondepleted untreated animals (P < 0.05). (C) BALB/c mice were infected intranasally with 4 × 10⁵ CFU of S. pneumoniae (filled square); 24 h after challenge, 1 μg of FliC was administered i.n. to one group of animals (open circles). At selected time points after infection (days 1, 2, and 7), BAL was carried out for 3 mice per time point in order to perform FACS analysis. The total number of cells in BAL fluid was determined, and the number of neutrophils was determined by flow cytometry. The dagger marks the time point at which all animals in the group were dead. For panels A and C, data are means; error bars represent the standard errors of the means; and significant differences between groups (P < 0.05) are indicated by asterisks. The results are representative of 1 out of 2 experiments.

Subsequently, we determined whether neutrophils were critical for flagellin-mediated protection in the coadministration setting of flagellin treatment. For this purpose, animals were injected i.p. with a monoclonal antibody specific for granulocyte receptor-1 (Gr-1, or Ly6G/Ly6C) or an isotype control antibody 24 h before challenge alone or challenge plus treatment. Animals that received the isotype control and were treated with FliC survived the challenge (Fig. 6B). In contrast, anti-Gr-1 treatment depleted >95% of neutrophils from the airways and abrogated flagellin-mediated protection against S. pneumoniae. These results showed that Gr-1-expressing cells, likely PMN, are critical effectors of flagellin-induced protection in pneumococcal infection.

Because the protection induced by coadministration of flagellin relied on a rapid influx of PMN into the lungs, we also analyzed the kinetics of PMN recruitment when flagellin was administered after the establishment of the infection. Animals were infected with S. pneumoniae and 24 h later either were treated with flagellin or remained untreated. BAL samples were collected at days 2, 3, and 7 after infection and were used for FACS analysis. Like coadministration of flagellin with the challenge, therapeutic administration of flagellin also induced a peak of PMN infiltration at 24 h after flagellin instillation, which decreased rapidly afterwards. By day 7, no neutrophils were detected in the lungs of animals treated with flagellin (Fig. 6C). Overall, these results suggest that the mechanism behind flagellin-induced protection in the therapeutic setting is similar to that observed when flagellin was coadministered with the bacterial challenge.

Since B and T lymphocytes are involved in the early phase of pneumococcal infection (12, 17, 18), we evaluated their roles in flagellin-induced protection. SCID mice (deficient in antibodies, B cells, and T cells) as well as immunocompetent BALB/c mice were challenged with 2 × 10⁷ CFU of S. pneumoniae either alone or with flagellin. Lungs and spleens were collected 36 h after infection for the determination of bacterial counts. As shown in Fig. 7A, flagellin coadministration promoted the clearance of bacteria in the lungs of SCID mice to an extent similar to that in BALB/c mice. Both SCID and BALB/c mice also showed lower bacterial counts in the spleen upon flagellin treatment (Fig. 7B), demonstrating that they were able to control not only local but also systemic infection. SCID mice recruited similar numbers of PMN into the lungs and alveolar spaces as BALB/c mice 16 h after the instillation of flagellin (Fig. 7C). In summary, these results show that Gr-1-expressing cells require neither B cells nor T lymphocytes in order to trigger flagellin-induced protection.

FIG. 7.

Flagellin-mediated protection against pneumonia does not depend on B and T lymphocytes. (A and B) BALB/c (solid symbols) or SCID (open symbols) mice (n = 6) were infected i.n. with 2 × 10⁷ CFU of S. pneumoniae serotype 1 in saline alone (squares) or in saline supplemented with 1 μg flagellin (FliC) (circles). Bacterial counts were determined in lung (A) and spleen (B) homogenates 36 h after challenge. Significant differences between groups (P < 0.05) are indicated by asterisks. Results are representative of 2 experiments. (C) Recruitment of PMN (defined as Ly6G⁺ Ly6C⁺ CD11b⁺ CD11c⁻ cells) into BAL fluid and lungs after i.n. instillation of 1 μg flagellin for SCID (open bars) or BALB/c (filled bars) mice (n = 5). Note that the numbers of PMN recruited in the two strains were not statistically significantly different. The results are presented as medians and standard errors of the means.

DISCUSSION

Innate immunity is essential for controlling pneumococcal infection, as shown by the requirements for TLR and MyD88 to prevent early colonization of the respiratory tract by S. pneumoniae (1, 20). The immune response to pneumococcal airway infection is characterized by a pronounced, brisk recruitment of neutrophils into the lungs (18), and phagocytic killing of pneumococci by PMN is considered a major defense mechanism. Nevertheless, S. pneumoniae can evade the host's innate defenses by inhibiting or delaying complement deposition and the respiratory burst of phagocytes (31). Hence, the neutrophil influx is often ineffective at clearing the infection until serotype-specific antibodies are produced and bypass the inhibition of complement deposition, enhancing opsonophagocytosis. In this study we assessed whether exogenous administration of an agonist of a TLR not naturally engaged by S. pneumoniae, namely, the TLR5 agonist flagellin, could strength innate immunity, enabling it to control an acute respiratory pneumococcal infection. We found that local administration of flagellin promoted the survival of mice challenged with a lethal dose of S. pneumoniae serotype 1 by enhancing local and systemic bacterial clearance. Flagellin treatment was effective when performed before, during, or after infection establishment in BALB/c, C57BL/6, and NMRI mice.

It has been demonstrated that in vivo administration of flagellin upregulates the expression of proinflammatory cytokines (10, 15, 16, 37, 38). Here we show that coadministration of flagellin at the time of S. pneumoniae challenge also upregulates the expression of the PMN-specific chemokine/activator genes Cxcl1 and Cxcl2, as well as Tnf and Ccl20, in the lungs, whereas S. pneumoniae alone induces these genes poorly. In agreement with the chemokine and cytokine expression profile, analysis of lung tissue sections showed a heavy infiltration of cells into the peribronchial and perivascular regions that was more pronounced in the lungs of flagellin-treated animals than in those of infected and untreated animals. Of note, despite the pronounced inflammatory response induced by flagellin, lung tissue recovered fully by day 7, while untreated animals died of infection. The analysis of BAL samples suggested that administration of the TLR5 agonist at the time of pneumococcal challenge as well as after the establishment of infection induced accelerated and more-pronounced PMN recruitment. PMN infiltration was transient and peaked at 24 h after flagellin treatment, after which time it started to decrease. Depletion of Gr-1-expressing cells, most likely PMN, abolished the protection, demonstrating that these cells are necessary for controlling pneumococcal lung infections. The self-limiting nature of the flagellin-mediated inflammatory response is a very relevant finding, since exacerbated inflammation could be associated with failure of the lung barrier and function. The molecular mechanisms controlling TLR5 response not only upregulate proinflammatory genes but also trigger response termination. Therefore, mucosal treatment with flagellin could be considered as a therapy against pneumococcal pneumonia, enhancing neutrophil infiltration and concurrently limiting inflammation, that merits further evaluation in clinical trials.

Besides PMN, several studies have also reported that T and B lymphocytes, as well as natural antibodies, may play important roles in the early control of pneumococcal pneumonia. Kadioglu et al. showed that T lymphocytes accumulate in zones of peribronchial inflammation at early stages of the immune response and are involved in the defense against pneumococci, since major histocompatibility complex (MHC) class II-deficient mice lacking CD4⁺ T cells are more susceptible to infection than their wild-type counterparts (17, 18). Haas et al. showed that CD19-deficient mice, which have impaired development of B1a cells and impaired natural antibody production, have increased susceptibility to pneumococcal infection (12). However, as indicated by the results obtained in SCID mice, neither T nor B cells are required for flagellin-induced local and systemic clearance of bacteria. Taken together, our results strongly suggest that changing the PMN dynamic results in effective killing of pneumococci, even in the absence of B and T lymphocytes.

Our results also showed that TLR5 signaling is required for the protection induced by flagellin. In the airways, TLR5 is expressed by alveolar macrophages (30) and epithelial cells (16, 32, 33), suggesting that these resident cells may be key players in the induction of protective innate defenses against S. pneumoniae upon flagellin treatment. In line with this, recent studies have suggested that the airway epithelium is the TLR5-activated tissue involved in chemokine production and PMN recruitment in response to flagellated bacteria (10, 16). On the other hand, murine neutrophils express TLR5 (13, 45); thus, TLR5 signaling may also activate PMN directly and enhance their S. pneumoniae killing capacity. Similarly, Lysenko et al. established that heat-killed Haemophilus influenzae can specifically increase the capacity of PMN to kill pneumococci in a Nod1-dependent manner (22, 23). Based on this information and our own results, we can postulate that the mechanism behind the early PMN recruitment and protection induced by flagellin most likely depends on soluble mediators produced by nonhematopoietic cells, likely chemokines produced by epithelial cells, and is independent of antibodies as well as of T and B cells.

Besides the TLR5 signaling activity, it has been shown recently that flagellin can also engage the cytoplasmic sensor IPAF/NLRC4 (11, 25, 26). However, since the mutant flagellin used here is unable to signal through TLR5 but still harbors the motif involved in IPAF/NLRC4 signaling (26), we can speculate that this pathway does not make a major contribution to the protective effect induced by flagellin treatment.

Current therapies for the prophylaxis and treatment of S. pneumoniae infection have limitations in preventing or curing pneumococcal disease (reviewed in references 4, 19, and 41); thus, new strategies of immune intervention are still required. Several reports (5, 7, 9) have shown that the administration of bacterial lysates and whole heat-killed bacteria stimulates protective responses against infection. However, the undefined nature of these preparations is usually a problem when designing drugs for human use. Our results showed that local stimulation with a single and well-characterized molecule, specifically flagellin, is sufficient for augmenting lung innate immune defenses and controlling pneumococcal pneumonia, highlighting the benefits of using microbe-associated molecular patterns as the basis for developing antimicrobial therapies.