Abstract
We used a mouse model of acute respiratory infections to investigate the role of Toll-like receptor 2 (TLR2) and TLR4 in the host response to Haemophilus influenzae. Acute aerosol exposures to wild-type strains of H. influenzae showed that TLR4 function was essential for TNF-α induction, neutrophil influx, and bacterial clearance. To determine how lipooligosaccharide (LOS) modifications would affect the role of TLR4 in inducing the host response, we used acute infections with an H. influenzae strain expressing a mutation in the htrB gene. This mutant strain expresses an LOS subunit with decreased acylation. In response to H. influenzae htrB infection, tumor necrosis factor alpha (TNF-α) secretion remained TLR4 dependent. But the decrease in LOS acylation made the neutrophil influx and the bacterial clearance also dependent on TLR2, as shown by the decreased host response elicited in TLR2 knockout mice compared to C57BL/6 mice. A subsequent analysis of TLR2 and TLR4 gene expression by quantitative PCR indicated that TLR4 function induces TLR2 expression and vice versa. These results indicate that some changes in the LOS subunit of H. influenzae can favor signaling through non-TLR4 receptors, such as TLR2. The results also indicate a close interaction between TLR4 and TLR2 that tightly regulates the expression of both receptors.
Nontypeable Haemophilus influenzae (NTHI) is a commensal and opportunistic pathogen. It can cause serious complications, from otitis media in children (26) to exacerbation of chronic obstructive pulmonary disease in adults (21). Since NTHI expresses lipooligosaccharide (LOS), Toll-like receptor 4 (TLR4) was thought to be the major mammalian receptor involved in H. influenzae-dependent signaling. Previous evidence has implicated TLR4 function in H. influenzae clearance in the mouse lung (30). Additional in vitro evidence in human cell lines has also shown TLR2, another member of the toll-like receptor family, to mediate H. influenzae signaling (22, 23), indicating that multiple host receptors mediate the immune response to acute H. influenzae infections.
Modifications of the LOS and lipopolysaccharide (LPS) subunits have been described in a large number of bacteria (3, 8, 9, 11). Often, these modifications are thought to improve the survival of the bacteria in the host. To define how modifications in the LOS subunit of H. influenzae affect the induction of the host response, we used two H. influenzae strains to induce acute respiratory tract infections in a mouse model. The LOS subunits expressed by both H. influenzae isolates are well characterized. While strain H. influenzae 2019 expresses a wild-type LOS with hexaacylation (11), the H. influenzae htrB strain expresses LOS isoforms with pentaacylated and tetraacylated lipid A moieties (24, 27). In human airway xenograft studies, these changes in the acylation state did not affect adherence (27), suggesting that the effect of the htrB mutation was limited to the endotoxin moiety, specifically the acylation of lipid A. Since lipid A is known to be the major immunogenic component of bacterial endotoxins (18) and in vitro studies have shown that lipid A exerts its immunogenic effect through efficient activation of TLRs, specifically TLR4 (13, 20), it was not surprising that the H. influenzae htrB strain showed reduced virulence in human xenograft studies (27). TLR4 signaling has recently been shown to induce TLR2 expression in endothelial cells (4), providing evidence for cross talk between several TLRs. In addition to the regulation by TLR4, TLR2, the receptor for lipoproteins, can form complexes with TLR1 and TLR6 (17, 28, 29). The results indicating transcriptional regulation between TLR2 and TLR4 support a close interaction between TLR2 and TLR4 in the initiation of the host response. Since both TLRs were shown to signal in response to acute H. influenzae infections, we used a mouse model of acute H. influenzae lung infections to determine which aspects of the host response were regulated by TLR2 and TLR4.
We used a mouse model of acute respiratory infection and measured functional endpoints, such as neutrophil infiltration and cytokine secretion. We determined that TLR4 function was required for cytokine secretion and neutrophil chemotaxis, following an acute infection with wild-type H influenzae. By contrast, H. influenzae htrB strain infections also required TLR2 expression for clearance and neutrophil influx, while cytokine secretion was still largely dependent on TLR4 function. This comparison of the host response induced by wild-type and mutant H. influenzae strains indicated a significant increase in TLR2 expression following an acute infection with the H. influenzae htrB mutant. The change in TLR2 expression suggests that some mutations in LOS may lead to an increased involvement of receptors other than TLR4 as a means of clearing invading pathogens from the lung.
MATERIALS AND METHODS
Animals.
Male C57BL/6 mice, aged 6 to 8 weeks, were purchased from Jackson Laboratories (Bar Harbor, Maine). Age-matched, male TLR4 knockout (TLR4KO) (7) and TLR2 knockout (TLR2KO) (28) mice, which were backcrossed for a minimum of six generations into a C57BL/6 background, were obtained from Akira (Osaka University, Osaka, Japan). Mice were given food and water ad libitum. Prior approval of all experimental procedures involving animals was obtained from the institutional animal care and use committee.
Bacterial strains.
Nontypeable H. influenzae strain 2019 (2) and the isogenic htrB mutant B29 (11) (gifts of M. A. Apicella, University of Iowa) were plated on brain heart infusion agar (BHI; Difco) supplemented with hemin (ICN Biochemical) and NAD (Sigma). On the day of exposure, bacterial colonies were eluted from the plates, and the optical density was measured to determine the bacterial density (27). A stock solution of 1011 CFU/ml in 0.9% NaCl was diluted to 1010 CFU/ml in Hanks balanced salt solution for use in exposures. A total of 4 ml of the bacterial solution was used for the exposure of two mice at a time using the nose-only exposure setup described below.
Assay for bacterial clearance.
Bacterial density and pulmonary deposition were measured by plating aliquots of the lung lavage at both 4 and 24 h postexposure on supplemented BHI plates (27). Following overnight incubation, colonies were counted and numbers were averaged. Time points were chosen to correlate with maximum expression of cytokines and influx of neutrophils, respectively. Preliminary experiments indicated that despite some bacterial clearance, sufficient numbers of live bacteria remained in the lung to assay host-induced clearance.
Mouse model of acute respiratory infection.
The exposure setup uses a nose-only aerosol system that was developed for mice and described previously (5). Briefly, the exposure system, previously optimized for live exposure to Pseudomonas aeruginosa (14), was adapted for H. influenzae exposure and includes a jet nebulizer (Mallinckrodt Medical, St. Louis, Mo.) attached to a Lucite chamber. The chamber has holes to accommodate the intake and outlet valves for delivery of the aerosol. Additional holes are sized to accommodate tubes for holding the mice. The tips of the tubes are trimmed off to permit the rostrum of the mouse to protrude into the aerosol. A driving pressure of 25 lb/in2 is used to deliver the aerosol for 24 min. Exposed mice were euthanized at 4 or 24 h postchallenge. For strain comparisons, mice from each strain to be tested were exposed in parallel to minimize variations due to aerosol delivery. In addition, a swipe test of the chamber prior to each exposure set verified the lack of carryover bacterial contamination between exposures. To ensure delivery of equal numbers of wild-type and mutant bacteria, filters were inserted into the chamber during the exposure. Following exposure, the filters were washed in Luria-Bertani (LB) for 3 h and then plated on supplemented BHI plates to ensure delivery of comparable numbers of wild-type and mutant H. influenzae strains.
Whole-lung lavage.
At specific time points after the completion of the exposure, mice were sacrificed, the chest was opened, and lungs were lavaged as described previously (5). Following the lavage, the samples were processed as previously described (15). Briefly, the lavage volume was noted and centrifuged in 15-ml conical tubes for 5 min at 200 × g. The supernatant fluid was decanted and frozen at −70°C for subsequent use. The cell pellet was resuspended and washed twice in Hanks balanced salt solution (without Ca2+ or Mg2+). After the second wash, an aliquot of the sample was taken for cell count with a cytocentrifuge (Cytospin-2; Shandon Southern, Sewickley, Pa.). Duplicate slides were obtained for each lavage sample. Staining was done with a Diff Quick stain set (Harleco, Gibbstown, N.Y.) as previously described (15). Cell counts were assigned blinded to mouse strain and exposure conditions.
Cytokine assays.
All cytokines assays were done in duplicate, based on 50 μl of lavage fluid, as described previously (15), with a commercially available enzyme-linked immunosorbent assay kit (R&D Systems, Minneapolis, Minn.) according to the manufacturer's recommendations.
RNA isolation and real-time RT-PCR.
Total RNA was isolated from mouse lungs with RNA-STAT 60 (Tel Test, Inc., Friendswood, Tex.), following the manufacturer's instructions. RNA was quantified spectrophotometrically. Real-time reverse transcription-PCR (RT-PCR) was performed with SYBR Green detection using an ABI Prism 7000 Sequence Detection System (Applied Biosystems, Foster City, Calif.). Each 50-μl reaction mixture contained 10 to 80 ng of cDNA (to establish linearity of response), SYBR Green Universal PCR Mix with hot-start Amplitaq Gold enzyme (Applied Biosystems), and 100 nM gene-specific primers. GAPDH primers were used to establish an internal standard for the calculation of TLR2 and TLR4 gene expression. For each unknown sample, the relative amount of TLR2 and TLR4 was calculated by the comparative CT method as described in the user bulletin supplied by the manufacturer (ABI Prism 7700 Sequence Detection System Update 2001; Applied Biosystems). The ΔΔCT was calculated by the subtraction of the unstimulated control ΔCT from the stimulated ΔCT value. Results represent the average percent change relative to saline-treated control mice and were calculated with the equation 2(−ΔΔCT), with the range determined by ΔΔCT plus the standard deviation and ΔΔCT minus the standard deviation. Real-time RT-PCRs were done in duplicate and repeated twice.
Statistics.
A minimum of four mice were used per condition and time point. Readouts for inflammatory markers for each set were averaged, and the standard deviation was determined. The statistical significance of differences between sets was analyzed with the SPSS 9.0 software package (SPSS, Chicago, Ill.). A P level of 0.05 by paired t test was used to determine statistical significance with equal variances assumed.
RESULTS
Deletion of TLR4, but not TLR2, prevented clearance after an acute infection with wild-type H. influenzae.
C57BL/6, TLR2KO, and TLR4KO mice were exposed to an aerosol containing live NTHI 2019 as well as the H. influenzae htrB strain and euthanized at 4 or 24 h postexposure. As a measure of bacterial survival in the lung, aliquots of lung lavages from infected mice were plated. For infections with wild-type bacteria, C57BL/6 and TLR2KO mice showed a decrease in the number of live bacteria retrieved from the lung lavage by 24 h postexposure, while TLR4KO mice did not. Moreover, in TLR4KO mice, the bacterial counts increased as the infection progressed from 4 to 24 h postexposure (from 5.46 × 103± 5.1 ×103 CFU/ml at 4 h to 8.42 × 105± 2.18 ×105 CFU/ml at 24 h) (Fig. 1A).
FIG. 1.
TLR4 is required for bacterial clearance in response to an acute infection with NTHI 2019. C57BL/6, TLR4KO, and TLR2KO mice were exposed to H. influenzae 2019 (A) and the H. influenzae htrB strain (B) as described in Materials and Methods. TLR2KO and C57BL/6 mice, but not TLR4KO mice, were able to clear the wild-type bacteria by 24 h (P = 0.004). By contrast, the bacterial load in TLR4KO mice increased by 24 h compared to 4 h (P = 0.014). In mice infected with the H. influenzae htrB strain, C57BL/6 and TLR2KO mice had comparable levels of live bacteria at 4 h postexposure, but only C57BL/6 mice were able to clear them efficiently. In TLR2KO mice, the levels of mutant bacteria remained unchanged, while in TLR4KO mice the levels of bacteria increased from 4 to 24 h postexposure. All results are based on a minimum of four mice per set and are indicated as average counts and standard deviations.
Even though both TLR2KO and C57BL/6 mice showed a reduction in the number of live bacteria as the infection progressed from 4 to 24 h, the number of bacteria recovered in the lung lavage differed significantly between strains. While TLR2KO mice had initial bacterial loads similar to those seen in TLR4KO mice, C57BL/6 mice had significantly more bacteria in the lavage. The lower levels of live bacteria in the lavages of the knockout mice could be due to an accelerated clearance. Such an accelerated clearance has been reported for TLR4KO mice exposed to live bacteria (6). The results indicate that TLR4, but not TLR2, function is required for an efficient clearance of wild-type H. influenzae.
For acute infections with the H. influenzae htrB stain, a mutant strain whose LOS is not completely acylated, the analysis of the bacterial load in the lung lavages revealed an efficient clearance of the H. influenzae htrB strain in C57BL/6 mice (Fig. 1B). While a significant number of bacteria were present at 4 h postinfection, by 24 h postinfection no live bacteria could be retrieved from the lavage. Despite the altered LOS subunit, C57BL/6 mice were able to efficiently clear the H. influenzae htrB strain. By contrast, neither knockout strain was able to clear the bacteria. TLR4KO mice, similar to exposures with wild-type H. influenzae, showed an increase in the bacterial load from 4 to 24 h postinfection. TLR2KO mice, in contrast to the efficient clearance of wild-type H. influenzae, could not clear the mutant H. influenzae, whose numbers remained constant between 4 and 24 h postexposure (Fig. 1B). The expression of TLR4 and TLR2 is therefore required for an efficient clearance of the mutant H. influenzae htrB strain.
TLR4 expression was required for the neutrophil influx following an acute infection with H. influenzae 2019, while the H. influenzae htrB strain required TLR2 and TLR4 function.
To further characterize the host response during an acute infection with wild-type H. influenzae, we also measured lavage neutrophils. TLR4KO mice did not show any inflammatory infiltrates at 4 or 24 h postexposure, evidenced by an almost exclusive presence of macrophages in the lavage (data not shown). By contrast, both C57BL/6 and TLR2KO mice had a significant number of neutrophils at both times (Fig. 2A), with higher numbers present in TLR2KO mice than in wild-type mice at 4 and 24 h postexposure. The increased number of neutrophils present in the lavage specimens from TLR2KO mice suggests a hyperresponsive phenotype in TLR2KO mice exposed to wild-type H. influenzae.
FIG. 2.
TLR4 expression is required for the neutrophil influx in response to an acute infection with H. influenzae 2019. Neutrophil infiltration was measured in the lung lavages of C57BL/6, TLR2KO, and TLR4KO mice infected with wild-type (A) and mutant (B) H. influenzae strains as described in Materials and Methods. (A) In TLR2KO and C57BL/6 mice, the highest levels of neutrophils were seen at 4 h, with a decreased number present at 24 h postexposure (P = 0.007). In TLR4KO mice, no neutrophil increase was seen at either time point, with levels remaining below those of TLR2KO mice (P = 0.003) and C57BL/6 mice. (B) C57BL/6 mice had only minimal levels of neutrophils by 4 h postexposure but showed a strong influx of neutrophils by 24 h postexposure. TLR4KO mice had some neutrophil influx at 4 h postexposure, but the number of lavage neutrophils was reduced by 24 h postexposure. Compared to C57BL/6 and TLR4KO mice, TLR2KO mice induced the lowest levels of lavage neutrophils. Results are given as averages and standard deviations of a minimum of four mice per set.
Since TLR2 and TLR4 are required for an efficient clearance of the H. influenzae htrB strain, we anticipated that both TLR2KO and TLR4KO mice would induce lower levels of neutrophils in the lung lavage. Acute exposure to the H. influenzae htrB strain altered the neutrophil influx in C57BL/6 mice but did not weaken it. The presence of the mutant LOS subunit on H. influenzae induced a delayed neutrophil influx in C57BL/6 mice (Fig. 2B). In TLR4KO mice, neutrophil influx peaked early but at a significantly lower level than in C57BL/6 mice. This result suggests that although TLR4 function is required for an effective host response to the H. influenzae htrB strain, other factors besides TLR4 contribute to the migration of neutrophils during acute H. influenzae htrB strain infection. Based on previous experiments in cell lines, we anticipated one of these factors to be TLR2 (22, 23). We therefore tested the neutrophil influx in TLR2KO mice (Fig. 2B). TLR2KO mice showed an almost complete absence of lavage neutrophils at 4 and 24 h postexposure, confirming that TLR2 function is required for an effective host response to the H. influenzae htrB strain.
The results indicate that the mutation in the LOS subunits delays the induction of the host response in wild-type mice and that both TLR4 and TLR2 are required for neutrophil activation. TLR2 expression, in particular, is an important mediator of neutrophil influx during acute respiratory infection with the H. influenzae htrB strain, since TLR2KO mice show an almost complete lack of lavage neutrophils.
TLR4 expression was required for the cystokine secretion following an acute infection with H. influenzae 2019, while the H. influenzae htrB strain requires TLR2 and TLR4 function.
Measurement of tumor necrosis factor alpha (TNF-α) and macrophage inflammatory protein 2 levels (data not shown) in the lavage confirmed an effective host response in C57BL/6 and TLR2KO mice infected with wild-type H. influenzae, since both mouse strains showed high levels of TNF-α in the lavage at 4 h postexposure. TNF-α levels in TLR4KO mice did not differ significantly from saline baseline levels (Fig. 3A). By 24 h postexposure, the levels of TNF-α in C57BL/6 and TLR2KO mice had declined to a baseline level, which was identical in all three mouse strains following saline exposure (data not shown).
FIG. 3.
TLR4 expression is required for the cytokine secretion in response to an acute infection with H. influenzae 2019. TNF-α secretion in the lung lavage was measured as described in Materials and Methods in mice infected with wild-type H. influenzae (A) and mice infected with the H. influenzae htrB strain (B). (A) TLR2KO and C57BL/6 mice showed comparably high levels of TNF-α 4 h postexposure, with a return to baseline at 24 h as indicated by the saline-exposed C57BL/6 control. TNF-α levels in TLR4KO mice remained at baseline throughout. TNF-α levels are given in picograms per milliliter of lavage fluid. (B) TNF-α levels were measured by enzyme-linked immunosorbent assay. Both C57BL/6 and TLR2KO mice showed similar levels of TNF-α at 4 and 24 h postexposure to the H. influenzae htrB strain. By 24 h, cytokine levels in all mouse strains decreased to baseline. The levels of TNF-α in TLR4KO mice were similar to those seen in saline-exposed mice throughout the infections. All results are based on four mice per set; all assays were done in duplicate. Results are given as averages, and standard deviations are indicated.
The analysis of the TNF-α levels in the lavage indicates that TLR4 function is required for TNF-α secretion as well as neutrophil influx. In addition, the increased influx of neutrophils in TLR2KO mice suggests that TLR2 could act as a repressor of the immune response during an acute infection with H. influenzae 2019, potentially by limiting TLR4 function. Cross talk between TLR2 and TLR4 has been shown previously (4).
Analysis of the neutrophil influx indicated that both TLR2 and TLR4 function contributes to the host response to the H. influenzae htrB strain. We next assayed the levels of cytokine secretion in lung lavage. Results from the exposures with wild-type H. influenzae suggested a requirement for TLR4 to induce TNF-α secretion. This result was confirmed when TNF-α expression in the lung lavage was measured following an infection with the mutant H. influenzae. The levels of lavage cytokines were lower than those seen in mice infected with wild-type H. influenzae and suggest that some aspects of the host response induced by the H. influenzae htrB strain are weakened. Despite the role of TLR2 in inducing neutrophil influx and clearance, TLR2KO mice showed levels of TNF-α in the lung similar to those in C57BL/6 mice exposed to the H. influenzae htrB strain. TLR4KO mice, on the other hand, had only baseline levels of TNF-α in the lavage, similar to the levels observed in C57BL/6 control mice exposed to saline. By 24 h postexposure, the TNF-α levels in all mouse strains had receded to identical baseline levels (Fig. 3B). This result indicates that TLR4, but not TLR2, expression is required to induce TNF-α secretion during acute H. influenzae htrB strain infections.
The htrB mutation altered the induction of TLR2 expression during acute H. influenzae infections.
The preceding analysis suggested that TLR2 function was more important in the immune response to the H. influenzae htrB strain than during acute infections with H. influenzae 2019. We therefore used real time RT-PCR to determine whether TLR2 expression was differentially expressed in C57BL/6 mice (Fig. 4, top, and Table 1). A comparison of the TLR2 mRNA levels in C57BL/6 mice infected with H. influenzae 2019 and the H. influenzae htrB strain showed an increased expression of TLR2 in C57BL/6 mice infected with the mutant H. influenzae strain.
FIG. 4.
Acute infections with the H. influenzae htrB strain induce higher expression levels of TLR2 than wild-type bacteria. Total RNA was isolated from the lungs of infected C57BL/6 and TLR4KO mice as described. Reverse-transcribed cDNA was amplified with gene-specific primers for GAPDH and TLR2 for quantitative real-time RT-PCR. Amplifications revealed a higher induction of TLR2 expression in C57BL/6 mice (top) infected with the H. influenzae htrB strain mutant than infections with NTHI 2019 at 4 h postexposure (P = 0.016). In TLR4KO mice (bottom), TLR2 mRNA was present at lower levels following infection with wild-type bacteria than with mutant H. influenzae (P = 0.003, 4 h postexposure). Results are representative of two independent experiments with tissue samples from two different mice. Statistically significant differences in TLR2 mRNA induction at 4 h postexposure are indicated with asterisks.
TABLE 1.
TLR2 mRNA expression in C57BL/6 and TLR4KO mice exposed to H. influenzae
| Mouse strain, condition | Expression
|
Relative fold increase over control | |||
|---|---|---|---|---|---|
| TLR2 | GAPDH | ΔCT | ΔΔCT | ||
| C57BL/6, saline | 28.53 ± 0.69 | 22.44 ± 1.35 | 6.08 ± 0.65 | 0 ± 0.65 | |
| C57BL/6, NTHI 4 h | 26.23 ± 0.74 | 22.80 ± 0.85 | 3.43 ± 0.14 | −2.65 ± 0.14 | +5.31 ± 0.61 |
| C57BL/6, NTHI 24 h | 27.07 ± 0.66 | 22.51 ± 1.28 | 4.55 ± 0.61 | −1.53 ± 0.61 | +2.06 ± 1.26 |
| C57BL/6, htrB 4 h | 25.86 ± 0.89 | 22.93 ± 1.05 | 2.93 ± 0.15 | −3.15 ± 0.15 | +7.93 ± 0.97 |
| C57BL/7, htrB 24 h | 27.57 ± 0.84 | 22.67 ± 0.56 | 4.90 ± 0.78 | −1.18 ± 0.78 | −1.49 ± 1.30 |
| TLR4KO, saline | 29.19 ± 0.03 | 24.65 ± 0.78 | 4.53 ± 0.81 | 0 ± 0.81 | |
| TLR4KO, NTHI 4 h | 28.92 ± 0.92 | 26.38 ± 1.02 | 2.53 ± 0.09 | −2.0 ± 0.09 | +2.26 ± 0.25 |
| TLR4KO, NTHI 24 h | 28.72 ± 0.84 | 24.34 ± 0.83 | 4.38 ± 0.01 | −0.15 ± 0.01 | −0.63 ± 0.01 |
| TLR4KO, htrB 4 h | 26.44 ± 1.00 | 24.89 ± 0.59 | 1.55 ± 0.41 | −2.98 ± 0.41 | +6.18 ± 0.23 |
| TLR4KO, htrB 24 h | 28.23 ± 0.89 | 24.73 ± 0.44 | 3.49 ± 0.44 | −1.04 ± 0.45 | +0.38 ± 0.65 |
In contrast to the similar induction of TLR2 mRNA seen at 24 h postexposure with either bacterial strain, at 4 h postexposure the levels of TLR2 mRNA C57BL/6 mice infected with wild-type bacteria differed significantly from those mice infected with the mutant H. influenzae isolate. Infection with the H. influenzae htrB strain induced significantly higher levels of TLR2 mRNA at 4 h than infections with wild-type bacteria. C57BL/6 mice infected with the H. influenzae htrB strain showed a >8-fold increase in TLR2 mRNA than saline-exposed control mice. By contrast, wild-type H. influenzae induced <6-fold-more TLR2 mRNA by 4 h after the initial exposure. This difference in the induction of TLR2 mRNA was significant (P = 0.016). By 24 h postexposure, C57BL/6 mice infected with either bacterial strain showed similar levels of TLR2 induction. TLR2 mRNA was increased about twofold in mice from both experimental groups, and no significant difference was observed between mutant and wild-type bacteria (P = 0.62). The increased expression of TLR2 in mice infected with the H. influenzae htrB strain suggests that mutations in the LOS moiety of H. influenzae caused a differential induction of non-TLR4 receptors. Since cross talk between TLR4 and TLR2 was reported previously (4), we also analyzed TLR2 mRNA expression in TLR4KO mice exposed to mutant and wild-type H. influenzae to determine whether the lack of TLR4 function would alter the expression of TLR2 (Fig. 4, bottom, and Table 1). While the highest levels of TLR2 mRNA in TLR4KO mice were also observed at 4 h postexposure, the levels of TLR2 mRNA were reduced in TLR4KO mice compared to C57BL/6 mice. Fan et al. (4) reported that TLR4 signaling induced TLR2 expression in endothelial cells. It was therefore likely that the expression of TLR2 would be reduced in the lungs of TLR4KO mice. Acute infections with NTHI 2019 resulted in less than half the induction of TLR2 mRNA in the lungs of TLR4KO mice compared to the sixfold increase seen in C57BL/6 mice.
The difference in TLR2 mRNA induction between TLR4KO and C57BL/6 mice was less pronounced during infections with the H. influenzae htrB strain. Acute infections with the H. influenzae htrB strain in TLR4KO mice resulted in higher levels of TLR2 expression at 4 h postexposure than in infections with wild-type bacteria. These results indicate that TLR2 mRNA induction can occur independently of TLR4 function but that lack of TLR4 function decreases the levels of TLR2 mRNA in TLR4KO mice compared to C57BL/6 mice.
Deletion of TLR2 expression prevented the induction of TLR4 mRNA.
Analysis of the TLR2 mRNA expression indicated that the lack of TLR4 function reduced the induction of TLR2 expression during acute infection with H. influenzae. To determine whether the interaction between TLR4 and TLR2 also affected TLR4 induction in TLR2KO mice, we analyzed the expression of TLR4 in C57BL/6 and TLR2KO mice during acute infection with H. influenzae.
Analysis of TLR4 mRNA levels revealed that the infection with wild-type bacteria was associated with low levels of TLR4 mRNA (Fig. 5A and Table 2) in C57BL/6 mice. TLR4 mRNA expression reached its maximum by 24 h postinfection, with some induction observed at 4 h postexposure. Acute infections with the H. influenzae htrB strain were associated with significantly higher levels of TLR4 mRNA in C57BL/6 mice, despite the lack of a host response associated with the mutant bacteria. Consistent with the maximum influx of neutrophils at 24 h postexposure, the highest levels of TLR4 mRNA were observed at the later euthanasia point.
FIG. 5.
TLR4 mRNA induction in C57BL/6 and TLR2KO mice exposed to H. influenzae. (A) TLR4 mRNA induction in C57BL/6 mice infected with H. influenzae. TLR4 mRNA, as measured by real-time RT-PCR, was significantly stronger induced in C57BL/6 mice infected with the H. influenzae htrB strain compared to infections with NTHI 2019 (P < 0.001). (B) TLR4 mRNA induction in TLR2KO mice infected with H. influenzae. In TLR2KO mice infected with NTHI, maximum TLR4 mRNA induction was observed within 4 h of the acute infection, with decreased levels at 24 h postexposure. In TLR2KO mice infected with mutant H. influenzae, minimal levels of TLR4 mRNA were induced at either time point. The difference in TLR4 mRNA induction between the H. influenzae htrB strain and NTHI infections at 4 h postexposure was significant (P < 0.001). Statistically significant differences in TLR4 mRNA induction at 4 h postexposure are indicated with asterisks.
TABLE 2.
TLR4 mRNA expression in C57BL/6 and TLR2KO mice exposed to H. influenzae
| Mouse strain, condition | Expression
|
Relative fold increase over control | |||
|---|---|---|---|---|---|
| TLR4 | GAPDH | ΔCT | ΔΔCT | ||
| C57BL/6, saline | 34.3 ± 1.37 | 21.72 ± 0.89 | 12.57 ± 0.47 | 0 ± 0.47 | |
| C57BL/6, NTHI 4 h | 31.95 ± 0.84 | 21.73 ± 0.78 | 10.21 ± 0.06 | −2.36 ± 0.06 | +1.02 ± 0.01 |
| C57BL/6, NTHI 24 h | 31.53 ± 0.66 | 22.00 ± 0.93 | 9.53 ± 0.61 | −3.04 ± 0.61 | +2 ± 1.20 |
| C57BL/6, htrB 4 h | 32.75 ± 1.03 | 21.64 ± 0.96 | 11.11 ± 0.07 | −1.46 ± 0.07 | +3.40 ± 0.21 |
| C57BL/7, htrB 24 h | 32.63 ± 0.38 | 21.91 ± 0.84 | 10.72 ± 0.46 | −1.85 ± 0.47 | +6.98 ± 3.60 |
| TLR2KO, saline | 37.53 ± 1.12 | 22.61 ± 0.62 | 14.92 ± 0.49 | 0 ± 0.49 | |
| TLR2KO, NTHI 4 h | 36.31 ± 0.65 | 24.15 ± 0.72 | 12.16 ± 0.06 | −2.75 ± 0.06 | +5.75 ± 0.29 |
| TLR2KO, NTHI 24 h | 37.5 ± 0.77 | 23.10 ± 0.72 | 14.39 ± 0.05 | −0.52 ± 0.05 | +0.45 ± 0.05 |
| TLR2KO, htrB 4 h | 38.01 ± 0.74 | 23.46 ± 0.95 | 14.54 ± 0.20 | −0.37 ± 0.20 | +0.31 ± 0.19 |
| TLR2KO htrB 24 h | 37.92 ± 0.68 | 23.75 ± 0.88 | 14.17 ± 0.2 | −0.75 ± 0.2 | +0.69 ± 0.23 |
Analysis of TLR4 mRNA expression in TLR2KO mice yielded a different expression profile (Fig. 5B and Table 2). In TLR2KO mice infected with NTHI 2019, TLR4 mRNA reached a maximum induction by 4 h postexposure, followed by a rapid decline as the infection progressed to 24 h postexposure. This indicates that TLR2 expression is not required for the induction of TLR4 mRNA after an acute infection with H. influenzae expressing a wild-type LOS subunit. Since C57BL/6 mice showed significantly lower levels of TLR4 mRNA at 4 h after an exposure to NTHI 2019, lack of TLR2 expression is associated with an augmentation of TLR4 induction. This suggests that the presence of functional TLR2 may interfere with the induction of TLR4 mRNA early during the infection. This could indicate an active repressor function of TLR2 or competition between TLR2 and TLR4 for a common intermediate.
We next analyzed TLR4 mRNA expression in TLR2KO mice infected with the H. influenzae htrB strain. C57BL/6 mice had shown a significant induction of TLR4 mRNA, following an acute infection with mutant H. influenzae. By contrast, TLR2KO mice infected with the H. influenzae htrB strain showed significantly lower levels of induction of TLR4 mRNA, which is consistent with the lack of a host response observed in these mice. The differences in TLR4 mRNA induction between C57BL/6 and TLR2KO mice after acute infection with an H. influenzae strain expressing a mutant LOS subunit further support the conclusion that TLR2 expression and signaling are important in the induction of a host response, in part by induction of TLR4 expression.
DISCUSSION
Previous evidence in the literature has identified roles for both TLR2 and TLR4 in host defenses against NTHI (10, 23, 30). Since H. influenzae is also a common respiratory pathogen, we wanted to determine how TLR2 and TLR4 function contributed to host defenses in the lung. The acute exposures to wild-type H. influenzae indicated that TLR4 function is required for efficient clearance of the bacteria from the lung, as well as for neutrophil activation and TNF-α secretion in the lung (Fig. 1 and 2). TLR4KO mice showed significant defects in mounting a host response to wild-type H. influenzae, while TLR2KO mice were able to mount a more efficient host response than C57BL/6 mice.
To further define the role of TLR2 and its interaction with TLR4, we exposed mice to the H. influenzae htrB strain. In general, the htrB mutation did not affect the structure of the oligosaccharide, with the exception of phosphorylation; its major effect was on acylation of the lipid A unit. Infection with the mutant H. influenzae strain showed a reduced rate of survival (Fig. 1), which has already been noted in an in vitro model of airway infection with human epithelial cells (27). While TLR2KO mice were able to clear wild-type bacteria (Fig. 1), acute infections with the H. influenzae htrB strain in TLR2KO mice resulted in a lack of clearance (Fig. 1). Analysis of the lavage neutrophils also indicated a deficient host response of TLR2KO mice to H. influenzae htrB infections. These differences in the host response of TLR2KO mice suggest that TLR2 function is more important in the host response to the H. influenzae htrB strain than to NTHI 2019.
The shift towards a TLR2-driven host response in H. influenzae htrB infection was also evident at the molecular level. In C57BL/6 mice, TLR2 mRNA showed a higher rate of induction in infections with the H. influenzae htrB strain than with NTHI 2019. Some of the TLR2 mRNA induction was TLR4 dependent, since TLR4KO mice showed lower levels of TLR2 mRNA than C57BL/6 mice (Fig. 4). This is consistent with previous results (4) that suggest that TLR4 signaling induces TLR2 expression. Moreover, the difference in TLR2 mRNA expression between infections with NHTi and infections with the H. influenzae htrB strain was much more pronounced in TLR4KO mice than in C57BL/6 mice (Fig. 4). This result indicates that the absence of TLR4 expression augments the shift towards a TLR2-driven immune response.
We next analyzed the changes in TLR4 expression. Acute infections of C57BL/6 mice with either bacterial strain led to an increase in TLR4 expression as the infection progressed from 4 to 24 h postexposure (Fig. 5), with higher levels of TLR4 mRNA during infections with the H. influenzae htrB strain than with wild-type H. influenzae. By contrast, TLR4 mRNA levels in TLR2KO mice infected with NTHI 2019 reached a maximum level within 4 h of the initial infection. The early maximum in TLR4 expression may be the basis of the heightened immune response seen in TLR2KO mice (Fig. 1 and 2) and suggests that the presence of functional TLR2 in C57BL/6 mice can impede the induction of TLR4 mRNA. The high levels of TLR2 expression and activation early during the infection, as in C57BL/6 mice, could therefore limit the host response by reducing the levels of TLR4 induction (Fig. 5). The outer membrane protein complex of H. influenzae is known to activate TLR2 during acute infections in an myeloid differentiation factor 88-dependent manner, which could limit the ability of TLR4 to engage and signal through this shared TLR signaling intermediate (10).
Acute infections with the H. influenzae htrB strain showed an almost complete lack of TLR4 induction in TLR2KO mice compared to C57BL/6 mice. This difference indicates that expression of TLR4 is impaired during infections with the H. influenzae htrB strain, unless TLR2 is present and functional. This result therefore indicates a role for TLR2-mediated signaling in the induction of TLR4 expression, which is also supported by the timing during the infection at which both receptors reach maximum expression levels in C57BL/6 mice. TLR2 expression is reaching a maximum level by 4 h postinfection (Fig. 4), while TLR4 mRNA levels are induced to their highest levels only by 24 h post infection (Fig. 5). Lastly, a TLR2-dependent induction of TLR4 is supported by the lack of TLR4 induction in TLR2KO mice (Fig. 5). While TLR4 mRNA levels increase significantly in C57BL/6 mice exposed to the H. influenzae htrB strain, TLR2KO mice exposed to the mutant bacteria show only a minimal induction of TLR4 expression (Fig. 5). The lack of TLR4 mRNA induction in TLR2KO mice exposed to H. influenzae expressing a mutant LOS subunit therefore suggests two possible interpretations. Since TLR4 mRNA is efficiently induced in TLR2KO mice infected with NTHI 2019, but not during infections with the H. influenzae htrB strain, the LOS subunit in the htrB strain may inefficiently interact with TLR4 (Fig. 5). In addition, TLR2 expression may induce TLR4 expression, since TLR4 mRNA is induced in C57BL/6 mice infected with the H. influenzae htrB strain.
The analysis of the host response to the H. influenzae htrB strain and wild-type H. influenzae supports an important function for TLR2 in the host response to inhaled H. influenzae. TLR2 can both limit and induce TLR4 expression, depending on the bacterial strain used. TLR2 limits the induction of TLR4 during the early stages of the infection with wild-type bacteria in C57BL/6 mice compared to TLR2KO mice, but will induce TLR4 expression during the later stages of the infection with the H. influenzae htrB strain. Experiments involving oral administration of Salmonella into mice have reported a time-dependent involvement of TLR4 and TLR2 in inducing the host response, with TLR4 function preceding the TLR2-dependent induction of macrophage function (31). The experiments using wild-type H. influenzae support a role of TLR4 in inducing TLR2 function. Since the acute exposures with the H. influenzae htrB strain also show a TLR2-dependent induction of TLR4 expression, the timing of the TLR expression and function may be dependent on the respective bacterial strain.
The shift towards a TLR2-driven host response during acute infections with bacteria expressing specific LOS mutations has been observed previously in vitro. The Neisseria meningitidis lpxA mutant is viable in the absence of functional endotoxin expression and signals through TLR2 (19) rather than TLR4/MD-2 in vitro. Interestingly, the lpxA mutant shows an increased expression of the Opa protein (25), suggesting that lack of endotoxin expression may be compensated by increased expression of lipoproteins. The htrB mutant of H. influenzae is therefore the second example of a bacterial mutant that shifts from a TLR4-driven to a TLR2-driven immune response due to a mutation in the lipooligosaccharide component. The use of a mouse model to study the effect of the htrB mutation defines for the first time phenotypic effects of such a mutation in vivo and its effect on TLR function and expression in the lung. For the development of respiratory therapies against H. influenzae infections, both TLR2 and TLR4 have therefore to be targeted.
The results involving the acute infections with the H. influenzae htrB strain further indicate, on a more conceptual basis, that the expression of TLR4 can be induced by TLR2 function. This finding adds to previous reports by Fan et al. (4), which show that TLR4 signaling induces TLR2 expression in endothelial cells. Previous reports have already indicated both synergy (1) and cross-tolerance (12) between these two receptors, as well as extensive heterodimer formation between TLRs such as TLR4/TLR5 heterodimers (16). These results suggest therefore a complex interaction between TLRs that regulates the expression and function of these receptors during infections.
Acknowledgments
We thank W. M. Foster (Duke University) for help with the optimization of the nose-only aerosol model and Rebecca W. Todd (Wake Forest University Health Sciences) for excellent technical assistance. We also thank M. Seeds (Wake Forest University Health Sciences) for critical reading of the manuscript.
This work was funded through a research award from the American Lung Association, North Carolina Affiliate.
Editor: V. J. DiRita
REFERENCES
- 1.Beutler, E., T. B. Gelbart, and C. West. 2001. Synergy between TLR2 and TLR4: a safety mechanism. Blood Cells Mol. Dis. 27:728-730. [DOI] [PubMed] [Google Scholar]
- 2.Campagnari, A. A., M. R. Gupta, K. C. Dudas, T. F. Murphy, and M. A. Apicella. 1987. Antigenic diversity of lipooligosaccharides of nontypable Haemophilus influenzae. Infect. Immun. 55:882-887. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Ernst, R. K., E. C. Yi, L. Guo, K. B. Lim, J. L. Burns, M. Hacket, and S. I. Miller. 1999. Specific lipopolysaccharide found in cystic fibrosis airway Pseudomonas aeruginosa. Science 286:1561-1565. [DOI] [PubMed] [Google Scholar]
- 4.Fan, J., R. S. Frey, and A. B. Malik. 2003. TLR4 signaling induces TLR2 expression in endothelial cells via neutrophil NADPH oxidase. J. Clin. Investig. 112:1234-1243. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Foster, W. M., D. M. Walters, M. Longphre, K. Macri, and L. M. Miller. 2001. Methodology for the measurement of mucociliary function in the mouse by scintigraphy. J. Appl. Physiol. 90:1111-1118. [DOI] [PubMed] [Google Scholar]
- 6.Haziot, A., N. Hijiya, S. Gangloff, J. Silver, and S. Goyert. 2001. Induction of a novel mechanism of accelerated bacterial clearance by lipopolysaccharide in CD14-deficient and Toll-like receptor 4-deficient mice. J. Immunol. 166:1075-1078. [DOI] [PubMed] [Google Scholar]
- 7.Hoshino, K., O. Takeuchi, T. Kawai, H. Sanjo, T. Ogawa, Y. Takeda, K. Takeda, and S. Akira. 1999. Cutting edge: toll-like receptor 4 (TLR-4) deficient mice are hyporesponsive to lipopolysaccharide: evidence for TLR-4 as the Lps gene product. J. Immunol. 162:3749-3752. [PubMed] [Google Scholar]
- 8.Howard, M. D., A. D. Cox, J. N. Weiser, G. G. Schurig, and T. J. Inzana. 2000. Antigenic diversity of Haemophilus somnus lipooligosaccharide: phase-variable accessibility of the phosphorylcholine epitope. J. Clin. Microbiol. 38:4412-4419. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Inzana, T. J., R. P. Gogolewski, and L. B. Corbeil. 1992. Phenotypic phase variation in Haemophilus somnus lipooligosaccharide during bovine pneumonia and after in vitro passage. Infect. Immun. 60:2943-2951. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Latz, E., J. Franko, D. T. Golenbock, and J. R. Schreiber. 2004. Haemophilus influenzae type b-outer membrane protein complex glycoconjugate vaccine induces cytokine production by engaging human toll-like receptor 2 (TLR2) and requires the presence of TLR2 for optimal immunogenicity. J. Immunol. 172:2431-2438. [DOI] [PubMed] [Google Scholar]
- 11.Lee, F. K., D. S. Stephens, B. W. Gibson, J. J. Engstrom, D. Zhou, and M. A. Apicella. 1995. Microheterogeneity of Neisseria lipooligosaccharide: analysis of a UDP-glucose 4-epimerase mutant of Neisseria meningitidis NMB. Infect. Immun. 63:2508-2515. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Lehnert, M. D., S. Morath, K. S. Michelsen, R. R. Schumann, and T. Hartung. 2001. Induction of cross-tolerance by lipopolysaccharide and highly purified lipoteichoic acid via different Toll-like receptors independent of paracrine mediators. J. Immunol. 166:5161-5167. [DOI] [PubMed] [Google Scholar]
- 13.Lien, E., J. C. Chow, L. D. Hawkins, P. D. McGuinness, K. Miyake, T. Espevik, F. Gusovsky, and D. T. Golenbock. 2001. A novel synthetic acyclic lipid A-like agonist activates cells via the lipopolysaccharide/toll-like receptor 4 signaling pathway. J. Biol. Chem. 276:1873-1880. [DOI] [PubMed] [Google Scholar]
- 14.Lorenz, E., D. C. Chemotti, K. Vandal, and P. A. Tessier. 2004. Toll-like receptor 2 represses nonpilus adhesin-induced signaling in acute infections with the Pseudomonas aeruginosa pilA mutant. Infect. Immun. 72:4561-4569. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Lorenz, E., M. Jones, C. Wohlford-Lenane, N. C. Meyer, K. L. Frees, N. C. Arbour, and D. A. Schwartz. 2001. Genes, other than TLR4, are involved in the response to LPS. Am. J. Physiol. 281:L1106-L1114. [DOI] [PubMed] [Google Scholar]
- 16.Mizel, S., A. Honko, M. Moors, P. Smith, and A. West. 2003. Induction of macrophage nitric oxide production by gram-negative flagellin involves signaling via heteromeric Toll-like receptors 5/Toll-like receptor 4 complexes. J. Immunol. 170:6217-6223. [DOI] [PubMed] [Google Scholar]
- 17.Ozinsky, A., D. M. Underhill, J. D. Fontenot, A. M. Hajjar, K. D. Smith, C. B. Wilson, L. Schroeder, and A. Aderem. 2000. The repertoire for pattern recognition of pathogens by the innate immune system is defined by cooperation between Toll-like receptors. Proc. Natl. Acad. Sci. USA 97:13766-13771. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Persing, D. H., R. N. Coler, M. J. Lacy, D. A. Johnson, J. R. Baldridge, R. M. Hershberg, and S. G. Reed. 2002. Taking toll: lipid A mimetics as adjuvants and immunomodulators. Trends Microbiol. 10:S32-S37. [DOI] [PubMed] [Google Scholar]
- 19.Pridmore, A. C., D. H. Wyllie, F. Abdillahi, L. Steeghs, P. van der Ley, S. K. Dower, and R. C. Read. 2001. A lipopolysaccharide-deficient mutant of Neisseria meningitidis elicits attenuated cytokine release by human macrophages and signals via toll-like receptor (TLR) 2 but not via TLR4/MD2. J. Infect. Dis. 183:89-96. [DOI] [PubMed] [Google Scholar]
- 20.Raetz, C. R., and C. Whitfield. 2002. Lipopolysaccharide endotoxins. Annu. Rev. Biochem. 71:635-700. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Sethi, S., N. Evans, B. J. Grant, and T. F. Murphy. 2002. New strains of bacteria and exacerbations of chronic obstructive pulmonary disease. N. Engl. J. Med. 347:465-471. [DOI] [PubMed] [Google Scholar]
- 22.Shuto, T., A. Imasato, H. Jono, A. Sakai, H. Xu, T. Watanabe, D. R. Rixter, H. Kai, A. Andalibi, F. Linthicum, Y. L. Guan, J. Han, A. C. Cato, D. J. Lim, S. Akira, and J. D. Li. 2002. Glucocorticoids synergistically enhance nontypeable Haemophilus influenzae-induced Toll-like receptor 2 expression via a negative cross-talk with p38 MAP kinase. J. Biol. Chem. 277:17263-17270. [DOI] [PubMed] [Google Scholar]
- 23.Shuto, T., H. Xu, B. Wang, J. Han, H. Kai, X. X. Gu, T. F. Murphy, D. J. Lim, and J. D. Li. 2001. Activation of NF-κB by nontypeable Hemophilus influenzae is mediated by toll-like receptor 2-TAK1-dependent NIK-IKKα/β-IκBα and MKK3/6-p38 MAP kinase signaling pathways in epithelial cells. Proc. Natl. Acad. Sci. USA 98:8774-8779. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Starner, T. D., W. E. Swords, M. A. Apicella, and P. B. McCray, Jr. 2002. Susceptibility of nontypeable Haemophilus influenzae to human β-defensins is influenced by lipooligosaccharide acylation. Infect. Immun. 70:5287-5289. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Steeghs, L., R. den Hartog, A. den Boer, B. Zomer, P. Roholl, and P. van der Ley. 1998. Meningitis bacterium is viable without endotoxin. Nature 392:449-450. [DOI] [PubMed] [Google Scholar]
- 26.St. Geme, J. W., III. 2000. The pathogenesis of nontypable Haemophilus influenzae otitis media. Vaccine 19(Suppl. 1):S41-S50. [DOI] [PubMed] [Google Scholar]
- 27.Swords, W. E., D. L. Chance, L. A. Cohn, J. Shao, M. A. Apicella, and A. L. Smith. 2002. Acylation of the lipooligosaccharide of Haemophilus influenzae and colonization: an htrB mutation diminishes the colonization of human airway epithelial cells. Infect. Immun. 70:4661-4668. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Takeuchi, O., K. Hoshino, and S. Akira. 2000. Cutting edge: TLR2-deficient and MyD88-deficient mice are highly susceptible to Staphylococcus aureus infection. J. Immunol. 165:5392-5396. [DOI] [PubMed] [Google Scholar]
- 29.Takeuchi, O., K. Hoshino, T. Kawai, H. Sanjo, H. Takada, T. Ogawa, K. Takeda, and S. Akira. 1999. Diffferential roles of TLR-2 and TLR-4 in recognition of gram-negative and gram-positive bacterial cell wall components. Immunity 11:443-451. [DOI] [PubMed] [Google Scholar]
- 30.Wang, X., C. Moser, J.-P. Louboutin, E. Lysenko, D. Weiner, J. Weiser, and J. Wilson. 2002. Toll-like receptor 4 mediates innate immune responses to Haemophilus influenzae infection in mouse lung. J. Immunol. 168:810-815. [DOI] [PubMed] [Google Scholar]
- 31.Weiss, D. S., B. Raupach, K. Takeda, S. Akira, and A. Zychlinsky. 2004. Toll-like receptors are temporally involved in host defense. J. Immunol. 172:4463-4469. [DOI] [PubMed] [Google Scholar]