Nontypeable Haemophilus influenzae (NTHi) is considered a major pathogen underlying middle ear infection. This study characterized the role of Toll-like receptor 4 in the innate immune responses to acute otitis media induced by NTHi in mice. We used C3H/HeJ mice, which have nonfunctional Toll-like receptor 4, and normal wild-type C3H/HeN mice. NTHi were injected into the tympanic bulla, and middle ear effusions and tissues were collected. In C3H/HeN mice, the severity of acute otitis media decreased promptly with a significant reduction in bacterial recovery from middle ear effusions 48 h after injection. In contrast, all C3H/HeJ mice had otitis media at all time points examined, and increasing bacterial counts from middle ear effusions were detected in C3H/HeJ mice 72 h after injection. Expression of intracellular adhesion molecule-1 by the middle ear mucosa paralleled the number of polymorphonuclear cells in the middle ear in both strains. The findings of transmission electron microscopy revealed that phagocytosis and phagosome maturation of polymorphonuclear cells was impaired in C3H/HeJ mice. Our findings indicate that Toll-like receptor 4 plays a part in the early accumulation and functional promotion of polymorphonuclear cells in the middle ear for eradicating NTHi infection.
Acute otitis media is one of the most common infectious diseases in children, and the peak incidence of this disease occurs in early childhood. Acute otitis media is associated with hearing loss, delayed speech development, permanent middle ear damage and mucosal changes (Giebink, 1989). Nontypeable Haemophilus influenzae (NTHi) is considered a major pathogen responsible for acute otitis media (Berman, 1995) In recent years, the number of antibiotic-resistant strains of NTHi has increased, and it is vital to gain insight into the pathogenesis of acute otitis media (Commisso et al., 2000).
Innate immunity represents the first line of defense against invading pathogens, and it is important for the elimination of bacteria from the respiratory tract. Recently, Toll-like receptors (TLRs) have emerged as key regulators of innate immune responses to infection in mammals. To date, 11 different members of the TLR family have been identified, and they recognize molecular patterns derived from pathogens, such as bacteria and fungi (Brightbill & Modlin, 2000; Takeda et al., 2003; Yarovinsky et al., 2005). The interactions between TLRs and molecular patterns lead to activation of intracellular signaling pathways, such as the nuclear factor-kappa B (NFκB) pathway, and to expression of genes involved in production of cytokines and chemokines and activation of the adaptive immune system (Takeda et al., 2003).
In these TLR families, TLR2 is reported to be a regulator of the pathogenesis of acute otitis media with NTHi. TLR2 detects components of NTHi and regulates the host inflammatory responses, and NTHi lipoprotein P6 activates NFκB in human epithelial cells via TLR2 signaling (Shuto et al., 2001). TLR4 mediates lipopolysaccharide and lipooligosaccharide responsiveness and recognizes gram-negative bacteria via the lipopolysaccharide/lipooligosaccharide moiety present on the surfaces of these microorganisms (Chow et al., 1999; Hoshino et al., 1999). Recent studies have examined the role of TLR4 in host defenses against gram-negative bacterial infection in the lung, and have shown that this receptor contributes to protective innate immune responses (Wang et al., 2002; Branger et al., 2004; Wieland et al., 2005). Moreover, specific polymorphisms in the human TLR genes may be associated with increased susceptibility to infection. Alterations of the human TLR4 gene have been shown to be more prevalent in intensive care unit patients than in healthy volunteers, and a significantly higher incidence of gram-negative infection has been documented in patients with these mutations (Agnese et al., 2002). These findings indicate that TLR4 is important for protective immunity against bacterial infections. However, the details of the TLR4-mediated innate immune responses in the middle ear are not known and little is known about how TLR4 participates in protective immunity in the middle ear. Clarifying the process of innate immune responses in the middle ear is important for establishing a strategy for treating acute otitis media.
In this study, we produced a mouse model of acute otitis media by inoculating NTHi into the middle ear of C3H/HeJ (TLR4-deficient) mice, which have a nonfunctional TLR4 protein, and normal wild-type (WT) C3H/HeN mice to investigate the mechanism of protective innate immunity via TLR4 in the middle ear.
Materials and methods
C3H/HeJ and C3H/HeN mice were purchased from Charles River Laboratories (Atsugi, Japan). All mice were maintained in a pathogen-free facility until they were 6 weeks old, at which time they were used for experiments. All experiments were approved by the Committee on Animal Experiments of Oita University (Oita, Japan).
Middle ear challenge with live NTHi and induction of experimental otitis media
Strain 76 of NTHi, which was isolated from the nasopharynx of a patient with otitis media with effusion at Oita University, was used for the middle ear challenge. NTHi was grown on chocolate agar at 37°C under 5% CO2 for 16 h, and three to five clones were then transferred to another plate and incubated for 4 h. The bacterial concentration was determined by optimal density at wavelength of 600 nm and a bacterial suspension of 106 CFU mL−1 in phosphate-buffered saline (PBS) was prepared and stored on ice. We confirmed the bacterial concentration by counting the colonies after overnight incubation. The bacterial suspension was injected into the right tympanic cavity, followed by the method described previously (Sabirov et al., 2001). In brief, after an incision was made in the submandibular skin, the right inferior bulla was exposed, and two microholes were made in the bulla with a 27-gauge needle. A micropipette was inserted into one of the holes, and 10 μL (104 CFU) of live NTHi suspension was injected slowly. Fifty mice from each strain were used in the experiments. The mice were monitored otomicroscopically to confirm the presence of middle ear effusion (MEE) and tympanic membrane changes. The tympanic membrane color and opacity were recorded and graded in increasing order of inflammation and increasing correlation with the presence of MEEs according to the severity of acute otitis media as follows: 0, gray and translucent tympanic membrane without MEE (normal); 1, gray and opaque tympanic membrane with serous or mucoid MEE; 2, yellow and opaque tympanic membrane with purulent MEE. These grading scales are referred to Giebink & Wright (1983). At 6, 12, 24, 48 and 72 h after injection, 10 mice from each group were sacrificed under deep anesthesia by intraperitoneal administration of pentobarbital solution. Samples of MEEs were obtained by myringotomy at the time of decapitation, and the middle ears were washed with 250 μL physiologic saline. MEEs were diluted serially with PBS, and 10 μL of the diluted samples was plated on chocolate agar. Bacterial colonies were counted after overnight incubation to measure the rates of NTHi clearance from the middle ear. NTHi was identified by standard bacteriologic techniques, including Gram staining and determination of V and × growth factor requirements. Inflammatory cells in MEE samples were counted with a hemocytometer, and differential cell counting was performed by smear with Wright's staining (Diff Quick, American Scientific Products, McGaw Park, IL).
After the mice were sacrificed and samples of MEEs were collected, the mice were fixed by intracardiac perfusion with each fixative; 10% neutral-buffered formalin, periodate lysine paraformaldehyde and Karnovsky's fixative (2.5% glutaraldehyde and 2% paraformaldehyde in 0.2 M cacodylate buffer, pH 7.4) for histologic evaluation. The heads of mice were immersed in the same fixative for 6 h and decalcified in 0.12 M EDTA (pH 7.5) for 2 weeks at 4°C. For hematoxylin-eosin (H&E) staining, tissues were dehydrated through a graded series of ethanols, cleared in xylene and embedded in paraffin. For immunohistochemistry, the tissues were embedded in OCT compound (Sakura Finetechnical Co., Tokyo, Japan) and stored at −80°C until use. For transmission electron microscopy, the tissues were then postfixed in 1% osmium tetroxide in 0.1 M cacodylate buffer (pH 7.4) at 4°C for 2 h. Following postfixation, the specimen was dehydrated through a graded series of ethanols and embedded in EPON 812. Untreated mice were used as controls and were evaluated with the same procedure for histologic evaluation.
Serial paraffin sections (6 μm) containing the tympanic bulla were prepared. Sections of the middle ear mucosa around the tympanic orifice were stained with H&E, and histopathologic changes were observed under light microscopy.
Immunohistochemistry for Intracellular adhesion molecule-1 (ICAM-1)
Frozen sections 10 μ thick of the middle ear mucosa around the tympanic orifice were made using a cryostat and washed in PBS. Specimens were treated with 3% hydrogen peroxide in absolute methanol for 20 min at room temperature. Sections were exposed to 5% normal goat serum in PBS for 30 min and then incubated with biotinylated goat anti-mouse ICAM-1 antibody (GT, Minneapolis, MN) diluted 1 : 100 in 1% bovine serum albumin (BSA)-PBS for 18 h at room temperature. After a rinse with PBS, sections were incubated with ABC reagent (Vector Laboratories, Burlingame, CA) for 1 h and developed in 0.05% 3,3′-diaminobenzidine-0.01% H2O2 substrate medium in 0.1 M phosphate buffer (DAB-H2O2) for 8 min.
Confocal laser scanning microscopy for TLR2 and TLR4
Inflammatory cell infiltration into the middle ear was the most intense 24 h after injection in both strains, and sections of middle ear 24 h after injection were used for confocal microscopic examination. To detect TLR2 and TLR4 in the middle ear, confocal laser scanning microscopy was performed. Cryostat sections (10 μm) were incubated with FITC-labeled anti-mouse TLR2 antibody (eBioscience, San Diego, CA) diluted 1 : 100 in 1% BSA-PBS for 6 h at room temperature, followed by PE-labeled anti-mouse TLR4 antibody (eBioscience) diluted 1 : 100 in 1% BSA-PBS for 6 h at room temperature. Sections were rinsed in PBS, processed for immunofluorescence and examined with an LSM 5 PASCAL (Zeiss, Germany). PE-labeled rat IgG2a antibody and FITC-labeled Rat IgG2b antibody (eBioscience) were used for isotype-matched control staining.
Transmission electron microscopy (TEM)
Sections of the middle ear 24 h after injection were used for the TEM study. Ultra-thin sections that contained the middle ear mucosa and polymorphonuclear cells (PMNs) in the middle ear cavity were prepared. The samples were attached to a carbon-coated copper grid by air fixation, stained with saturated uranyl acetate in 50% methanol and lead acetate, washed in 0.02 N sodium hydroxide and distilled water and dried for 1 min under a lamp. Samples were viewed with a JEOL 1200XII at 80 kV.
Unpaired Student's t-test was used to evaluate the significance of differences in the severity score of acute otitis media, bacterial counts and the number of inflammatory cells in MEEs between WT and TLR4-deficient mice. A P value of less than 0.05 was considered significant.
Severity of acute otitis media
The tympanic membrane color and opacity were recorded and graded in increasing order of inflammation and increasing correlation with the presence of MEEs according to the severity of acute otitis media. The course of acute otitis media after injection from 6 to 72 h in both strains is shown in Fig. 1. In the WT mice, 100% of mice (10/10) had MEE 6–24 h after injection; however, the frequency of detection of MEE (score of >0) decreased to 60% of mice (6/10) by 48 h after injection. The presence ratio of MEE was 30% of mice (3/10) at 72 h after injection. In contrast, in TLR4-deficient mice, all mice had acute otitis media at each time point in the experiment. Microscopic examination of the eardrum showed deterioration with purulent MEEs in all mice 48 and 72 h after injection. The severity score of acute otitis media according to tympanic membrane findings is shown in Table 1. The severity score of acute otitis media in WT mice differed significantly from that in TLR4-deficient mice 48 and 72 h after injection (mean severity score±SD in WT vs. TLR4-deficient mice; 48 h, 0.8±0.8 vs. 2.0±0; 72 h, 0.3±0.5 vs. 2.0±0; P<0.01).
Table 1. The severity score of acute otitis media (AOM)
Ten mice from each group were monitored otomicroscopically at 6, 12, 24, 48 and 72 h after injection of NTHi.
Severity score of AOM was measured according to the tympanic membrane (TM) color and opacity, and the presence of MEEs as follows: 0, gray and translucent TM without MEE (normal); 1, gray and opaque TM with serous or mucoid MEE; 2, yellow and opaque TM with purulent MEE. The severity score expressed as the mean score ± standard deviation (SD).
Bacterial clearance and inflammatory cell infiltration
Bacterial counts of NTHi from the middle ear are shown in Fig. 2. WT mice showed significantly lower bacterial recovery after 24 h than did TLR4-deficient mice (P<0.01), and the bacterial counts were almost negligible 72 h after injection. In TLR4-deficient mice, bacterial counts continued to increase significantly even at 72 h after injection. Cytologic examination revealed marked infiltration by PMNs at all time points in both WT and TLR4-deficient mice, and the infiltrating cells in MEEs from WT mice comprised 93–100% PMNs, 0–5% macrophages/monocytes and 0–2% lymphocytes. In TLR4-deficient mice, the infiltrating cells in MEEs comprised 93–100% PMNs, 0–4% macrophages/monocytes, and 0–2% lymphocytes. There was no difference in the percentages of inflammatory cell types. The number of inflammatory cells in MEEs of WT mice was significantly higher 6 and 12 h after injection than in TLR4-deficient mice (the mean in WT vs. TLR4-deficient; 6 h, 2.5 × 106 vs. 0.5 × 106; 12 h, 5.0 × 106 vs. 1.0 × 106 cells mL−1; P<0.05) but had decreased by 48 and 72 h after injection (Fig. 2). In contrast, in TLR4-deficient mice, the number of inflammatory cells continued to increase at 72 h after injection. The elevated number of PMNs in TLR4-deficient mice was not reflected in bacterial clearance from the middle ear.
Before injection, no inflammatory cells were seen in the middle ear in either WT or TLR4-deficient mice (Fig. 3a and e). In WT mice, middle ear mucosa showed inflammation containing PMNs 6 h after injection (Fig. 3b) and these inflammatory responses were strong 12 and 24 h after injection (Fig. 3c). After 24 h, these responses became weak and had almost disappeared by 72 h after injection (Fig. 3d). In contrast, in TLR4-deficient mice, the subepithelial space appeared slightly thickened with only slight infiltration by inflammatory cells 6 h after injection (Fig. 3f); however, the subepithelial space became severely thickened due to edema, and PMNs had strongly infiltrated the middle ear mucosa and cavity between 12 and 72 h after injection (Fig. 3g and h). Control mice, which were not treated with the middle ear procedure, had no immune response to either strain.
Immunohistochemistry for ICAM-1
We also examined the expression of ICAM-1 in the middle ear mucosa after injection of NTHi. In WT mice, epithelial cells and subepithelial vessels of the middle ear were stained strongly with anti-ICAM-1 antibodies 6–24 h after injection, but ICAM-1 expression was decreased 48 and 72 h after injection. In TLR4-deficient mice, expression of ICAM-1 on the middle ear surface was weak 6 h after injection, but epithelial cells and subepithelial vessels of the middle ear showed strong ICAM-1 immunostaining 12–72 h after injection (Fig. 4). Control mice, which were not treated with the middle ear procedure, showed weak immunostaining in both strains. The intensity of ICAM-1 expression in the middle ear mucosa correlates closely with PMN infiltration into the middle ear in both strains.
In WT mice, PMNs that had infiltrated the middle ear mucosa showed strong immunostaining of both TLR2 and TLR4 24 h after injection (Fig. 5c). In contrast, in TLR4-deficient mice, PMNs showed hardly any staining of TLR2 and TLR4 (Fig. 5g). The intensity of TLR2 and TLR4 immunostaining on the epithelial surface of the middle ear mucosa was weak in both strains, but the intensity of staining in WT mice was slightly stronger than that in TLR4-deficient mice. No detectable staining was observed with isotype-matched control antibody.
TEM was used to assess the difference in PMN function between WT mice and TLR4-deficient mice. Activated PMNs in the middle ear cavity internalized NTHi and contained phagosomes with NTHi. Dying PMNs with internal bacteria were also seen in WT mice. Phagocytosis and phagosome maturation were impaired in TLR4-deficient mice (Fig. 6)
This study was designed to clarify the role of TLR4 in the response to acute otitis media caused by injection of NTHi. NTHi is a gram-negative bacterium that contains lipooligosaccharide. Although there are structural differences between lipooligosaccharide and lipopolysaccharide, lipid A, which is present in both, is thought to be responsible for TLR4 signaling, and lipooligosaccharide also induces TLR4 signaling (Zughaier et al., 2004). Our present results indicate that TLR4 plays a critical role in sensing and affecting the initial innate immune responses to middle ear challenge with NTHi. The severity of acute otitis media was reduced in WT mice but not in TLR4-deficient mice. The eardrum findings correlated closely with the number of bacteria in MEEs. When we examined the number of inflammatory cells present in the MEE of WT and TLR4-deficient mice, we found that there were 4–10 times more PMNs in WT mice than in TLR4-deficient mice 6 and 12 h after injection. The number of PMNs in WT mice then dropped sharply until 48 h after challenge. In contrast, the number of PMNs in TLR4-deficient mice continued to increase at 72 h after injection.
PMNs recognize ICAM-1, which is constitutively expressed, and it is known that ICAM-1 expression in the airway is increased by stimulation with lipopolysaccharide and gram-negative bacteria (Beck-Schimmer et al., 1997; Frick et al., 2000). In this study, the number of PMNs in MEEs paralleled ICAM-1 expression in the middle ear mucosa, and these results indicate that induction of ICAM-1 expression on middle ear epithelial cells by NTHi increased PMN adherence to the epithelial cell surface. It has recently been reported that ICAM-1 expression is affected by TLR4 signaling. Lipopolysaccharide-TLR4 signaling and PMN NADPH oxidase activate NFκB signaling and upregulate TLR2 expression in endothelial cells, and this increased TLR2 expression via NFκB signaling results in enhanced ICAM-1 expression and augmented PMN migration (Fan et al., 2003). Andonegui et al. (2003) reported that endothelium-derived TLR4 is the key molecule in lipopolysaccharide-induced neutrophil sequestration into lungs. These findings suggest that induction of ICAM-1 expression on the middle ear mucosa in the early phase of acute otitis media may be involved in TLR2 and TLR4 in WT mice, although the intensity of TLR2 and TLR4 immunostaining on the epithelial surface of the middle ear mucosa was weak in WT mice. However, infiltration by PMNs was augmented even in TLR4-deficient mice beginning 12 h after injection, in conjunction with the increase in ICAM-1 staining. It has been reported that the expression of TLR2 mRNA in response to lipid A is impaired in TLR4-deficient mice, and that the expression of TLR2 mRNA is increased in response to high levels of lipopolysaccharide (Matsuguchi et al., 2000; Liu et al., 2001). TLR2 signaling pathways mediated activation of NFκB by NTHi in epithelial cells (Shuto et al., 2001). In our preliminary study, the concentration of endotoxin, which contains lipooligosaccharide, in MEEs in our mouse model of acute otitis media increased in parallel with the bacterial counts in MEEs (data not shown). For this reason, ICAM-1 expression on the middle ear mucosa in TLR4-deficient mice may be involved in signaling via TLR2 in the later phases of acute otitis media.
PMNs have a large variety of antimicrobial functions as well as the ability to produce cytokines to initiate inflammatory responses and chemokines to induce trafficking of immune cells. These functions are stimulated through TLRs (Hayashi et al., 2003; Parker et al., 2005). One of the most important functions of PMNs is phagocytosis and a bactericidal effect against microorganisms. In this study, bacterial clearance from the middle ear cavity was enhanced in WT mice but not in TLR4-deficient mice 24–72 h after injection of NTHi despite the accumulation of large numbers of PMNs. Phagocytosis and phagosome maturation in PMNs were observed in WT but not TLR4-deficient mice in our TEM study. It is known that alveolar macrophages in the lung play a crucial role in eradicating bacterial infection, and TLR4 is especially important for a successful host defense in pulmonary infection with NTHi (Wieland et al., 2005). However, the bactericidal function of PMNs in relation to TLRs is not well understood, and the contribution of PMNs to innate immune responses has not been assessed fully. This is the first report in which impaired phagocytosis by PMNs has been shown in TLR4-deficient mice by TEM. Recently, it was reported that bacteria-induced activation of the TLR signaling pathway regulates phagocytosis at multiple steps including internalization and phagosome maturation in macrophages (Honstettre et al., 2004). Phagocytosis and phagosome maturation of macrophages were impaired in the absence of TLR signaling with TLR2 and TLR4 (Blander & Medzhitov, 2004). In this study, we also showed that PMNs that had infiltrated into the middle ear showed strong immunostaining with both TLR2 and TLR4 24 h after injection of NTHi in WT mice. In contrast, PMNs in TLR4-deficient mice showed little or no staining of TLR2 and TLR4. These results indicate that TLR2 and TLR4 may be involved in promoting phagocytosis and phagosome maturation in PMNs. Our present data suggest that TLR4 plays an important role in the recognition of NTHi and the bactericidal function of PMNs.
Our present results clarify the role of TLR4 in the innate immune response to NTHi-induced acute otitis media in mice. The severity of acute otitis media and the number of bacteria from MEE were increased in TLR4-deficient mice, and our findings indicate that TLR4 is critical for the early accumulation of PMNs via ICAM-1 expression in the middle ear after NTHi challenge. TLR4 is also essential for promotion of phagocytosis of NTHi by PMNs. In short, TLR4 plays an essential and vital role in eradicating NTHi infection in the middle ear.
This work was supported by a Grant-in-Aid for Scientific Research (B) from the Ministry of Education, Science, Sports and Culture of Japan (15790944). We thank I. Ichimiya and Y. Takenaka (in our department) for invaluable assistance and G. Rodgers for correcting this manuscript.