The development of early host response to Pseudomonas aeruginosa lung infection is critically dependent on myeloid differentiation factor 88 in mice.

Toll-like receptors (TLR) induce distinct patterns of host responses through myeloid differentiation factor 88 (MyD88)-dependent and/or -independent pathways, depending on the nature of the pathogen. Pseudomonas aeruginosa is a cause of serious lung infection in immunocompromised individuals and cystic fibrosis patients. The role of the TLR-MyD88 pathway in P. aeruginosa-induced lung infection in vivo was examined in this study. MyD88-/- mice demonstrated an impaired clearance of P. aeruginosa from the lung. Little or no neutrophil recruitment was observed in the airways of MyD88-/- mice following P. aeruginosa lung infection. This observation was associated with a reduced production of inflammatory mediators that affect neutrophil recruitment, including macrophage-inflammatory protein-2, tumor necrosis factor, and interleukin-1beta in the airways of MyD88-/- mice. Similarly, MyD88-/- mice showed inhibited NF-kappaB activation in the lung following P. aeruginosa infection. Interestingly, P. aeruginosa infection induced a 7.5-fold increase of TLR2 mRNA expression in the lungs of MyD88+/+ mice. Furthermore, host responses to P. aeruginosa lung infection in TLR2-/- and TLR4 mutant mice were partially inhibited compared with the responses of respective control mice. Taken together, our results indicate that the MyD88-dependent pathway is essential for the development of early host responses to P. aeruginosa infection, leading to the clearance of this bacterium, and that TLR2 and TLR4 are involved in this process.

Toll-like receptors (TLR) induce distinct patterns of host responses through myeloid differentiation factor 88 (MyD88)-dependent and/or -independent pathways, depending on the nature of the pathogen. Pseudomonas aeruginosa is a cause of serious lung infection in immunocompromised individuals and cystic fibrosis patients. The role of the TLR-MyD88 pathway in P. aeruginosainduced lung infection in vivo was examined in this study. MyD88 ؊/؊ mice demonstrated an impaired clearance of P. aeruginosa from the lung. Little or no neutrophil recruitment was observed in the airways of MyD88 ؊/؊ mice following P. aeruginosa lung infection. This observation was associated with a reduced production of inflammatory mediators that affect neutrophil recruitment, including macrophage-inflammatory protein-2, tumor necrosis factor, and interleukin-1␤ in the airways of MyD88 ؊/؊ mice. Similarly, MyD88 ؊/؊ mice showed inhibited NF-B activation in the lung following P. aeruginosa infection. Interestingly, P. aeruginosa infection induced a 7.5-fold increase of TLR2 mRNA expression in the lungs of MyD88 ؉/؉ mice. Furthermore, host responses to P. aeruginosa lung infection in TLR2 ؊/؊ and TLR4 mutant mice were partially inhibited compared with the responses of respective control mice. Taken together, our results indicate that the MyD88-dependent pathway is essential for the development of early host responses to P. aeruginosa infection, leading to the clearance of this bacterium, and that TLR2 and TLR4 are involved in this process.
The MyD88-dependent pathway is essential for the host defense against microbial infections in vivo from organisms such as Staphylococcus aureus (4) and Toxoplasma gondii (5). In contrast, resistance to Mycobacterium tuberculosis infection is effected largely through MyD88-independent pathways (6,7). In acute polymicrobial peritonitis, the effective antibacterial immune response occurs in the absence of MyD88 (8). In other cases, both MyD88-dependent and -independent mechanisms are involved. For example, Listeria monocytogenes activates an immune response through an ordered, sequential MyD88-independent and -dependent fashion (9). Specific microbes utilize individual TLR and adaptor pathways to induce immune responses that are tailored to the given microbial infection (10).
Pseudomonas aeruginosa, an opportunistic Gram-negative bacillus, is the major pathogen in cystic fibrosis patients (11) and a common cause of nosocomial pneumonia (12,13). Two major features of P. aeruginosa lung infection are the recruitment of neutrophils and the production of various cytokines and chemokines in the local tissue (14). Neutrophils play an essential role in the clearance of P. aeruginosa from the lung (15,16). Recruitment of neutrophils to the lung is largely induced by the production of inflammatory mediators in the airways (17). P. aeruginosa-induced mediator production in the airways is initiated by the interaction between P. aeruginosa and the host cell receptors (15). A major effort has been made to identify the molecules responsible for the initiation of P. aeruginosa-induced inflammation. Several cell surface molecules have been implicated in the direct interaction with P. aeruginosa, including the cystic fibrosis transmembrane conductance regulator (18,19), complement receptor 3 (20), gangliotetraosylceramide (21), CD91 (22), and syndecan-1 (23). More recently, TLR2, -4, and -5 have been associated with P. aeruginosa infection (24 -27). In vivo, decreased expression of TLR4 appears to correlate with impaired resistance to P. aeruginosa infection (25). In vitro, TLR2 and TLR5 are involved in P. aeruginosa flagella-induced activation of epithelial cells (24), and TLR2 and TLR4 are involved in monocyte and macrophage activation by a component of P. aeruginosa alginate (26). However, the contribution of the TLR-MyD88 pathway to the effective host response to P. aeruginosa lung infection remains a critical area of investigation.
Here we demonstrate that MyD88-deficient (MyD88 Ϫ/Ϫ ) mice showed little or no neutrophil recruitment or production of the neutrophil attractants MIP-2, TNF, and IL-1␤ and had an impaired ability to clear P. aeruginosa from the lung, which vastly contrasted with the findings of robust MIP-2, TNF, and IL-1␤ production, neutrophil infiltration, and bacterial clearance demonstrated in wild type animals. These findings suggest an essential role for the MyD88-dependent pathway in the host defense against P. aeruginosa lung infection in vivo. The different immune responses seen in MyD88 Ϫ/Ϫ , TLR2 Ϫ/Ϫ , and TLR4 mutant mice suggest that the host defense against P. aeruginosa lung infection may involve multiple members of the MyD88-dependent TLR/IL-1 superfamily.
Lung Infection with P. aeruginosa and Collection of Lung and Bronchoalveolar Lavage Fluid-P. aeruginosa strain 8821 (a gift from Dr. A. Chakrabarty, University of Illinois, Chicago) is a mucoid strain isolated from a cystic fibrosis patient (29). Mice were infected intranasally with 1 ϫ 10 7 or 1 ϫ 10 9 CFU of P. aeruginosa. After 4, 8, or 24 h, mice were sacrificed by cardiac puncture, and phosphate buffer solution was infused into the heart to remove blood from the lungs. Bronchoalveolar lavage fluid (BALF) was obtained by laving the lung three times with 1 ml of phosphate buffer solution containing soybean trypsin inhibitor (100 g/ml). Lung tissue was obtained for the detection of cytokines and myeloperoxidase (MPO) activity, for bacterial CFU counting (left lobes), and for histology, RT-PCR, and Western blot analysis (right lobes).
Lung tissue was homogenized in 50 mM HEPES buffer (4 l/mg lung tissue) containing soybean trypsin inhibitor (100 g/ml). For counting bacterial CFU, 10 l of the homogenate was plated on an agar dish and incubated for 24 -48 h at 37°C. The homogenate was centrifuged at 14,000 rpm for 30 min at 4°C. The supernatant was stored at Ϫ80°C for later analysis of cytokines. The pellet was resuspended and homogenized in 0.5% cetyltrimethylammonium chloride (4 l/mg lung tissue) and centrifuged as above. The clear extract was used for the MPO assay.
BALF (10 l) was plated on an agar dish and incubated for 24 -48 h, and then bacterial CFU were counted. For the detection of cytokines and MPO activity, BALF was centrifuged at 1500 rpm for 5 min at 4°C. Supernatants were used for cytokine analysis. The pellets were resuspended in 1 ml of NH 4 Cl (0.15 M) and spun as before to lyse red blood cells. The supernatants were discarded, and the pellets were resuspended in 0.5% cetyltrimethylammonium chloride (250 l/mouse) and centrifuged, and the clear extracts were used for MPO assay.
MPO Assay-The MPO assay was used to determine the infiltration of neutrophils into the lungs of the mice as described previously (30). Briefly, samples in duplicate (75 l) were mixed with equal volumes of the substrate (3 mM 3,3Ј,5,5Ј-tetramethyl-benzidine dihydrochloride, 120 M Resorcinol, and 2.2 mM H 2 O 2 ,) for 2 min. The reaction was stopped by adding 150 l of 2 M H 2 SO 4 . The optical density was measured at 450 nm.
Real-time Quantitative RT-PCR-Real-time quantitative RT-PCR was performed using a 7000 Sequence detector (PerkinElmer Life Sciences). Total RNA was isolated from the lung tissue and reverse tran-scribed using SuperScript II (Invitrogen). Specific quantitative assays for MIP-2, IL-12, TLR2, and TLR4 were performed using Assays-on-Demand TM reagents containing 6-FAM dye-labeled TaqMan® minor groove binder probes (PerkinElmer Life Sciences) according to the manufacturer's protocol (31). Glyceraldehyde-3-phosphate dehydrogenase was used as an endogenous reference. Data were analyzed using a relative standard curve method according to the manufacturer's protocol, and values from wild type untreated mice were used as the calibrator. Thus, data are expressed as the -fold increase relative to wild type untreated mice.
Electrophoretic Mobility Shift Assays-A consensus double-stranded NF-B oligonucleotide (Promega, Madison, WI) (5Ј-AGT TGA GGG GAC TTT CCC AGG C-3Ј) was used for the electrophoretic mobility assay. Probe labeling was accomplished by treatment with T4 kinase (Promega) in the presence of [ 32 P]adenosine triphosphate (Amersham Biosciences). Labeled oligonucleotides were purified on a Sephadex G-25M column (Amersham Biosciences). 10 g of nuclear protein was added to a 10-l volume of binding reaction with 1 g of poly(dI-dC) (Amersham Biosciences) and incubated at room temperature for 15 min. Labeled double-stranded NF-B oligonucleotide was added to each reaction mixture, which was incubated at room temperature for 30 min and separated by electrophoresis on a 6% polyacrylamide gel in 0.5ϫ Tris-boric acid-EDTA buffer. Gels were vacuum-dried and subjected to autoradiography. Cold competition was carried out by adding 1 l (1.75 M) of specific unlabeled double-stranded probe to the reaction mixture. Unlabeled double-stranded oligonucleotide (1 l, 1.75 M) that does not bind NF-B was used for nonspecific competition (data not shown). Polyclonal antibodies to p50 and p65 (1 g/10 l) (Santa Cruz Biotechnology) were used for supershift assays for NF-B proteins (data not shown).
Histology-Mice lungs were fixed in 10% formalin overnight and then in 100% ethanol for paraffin embedding and sectioning. Slides were deparaffinized with Citrisolv (Fisher) and rehydrated through decreasing concentrations of ethanol. Slides were stained with Harris hematoxylin-eosin to illustrate lung histology.
Statistics-Data are presented as mean Ϯ S.E. of the indicated number of experiments. Statistical significance was determined by assessing means with analysis of variance and the Tukey-Kramer multiple comparison test or by using an unpaired t test. Differences were considered significant at p Ͻ 0.05.
FIG. 1. Impaired clearance of P. aeruginosa from the lungs of MyD88 ؊/؊ mice. MyD88 ϩ/ϩ and MyD88 Ϫ/Ϫ mice were challenged intranasally with P. aeruginosa (mucoid strain 8821). After 24 h, the BALF (a) and the right lungs (b) were collected for bacterial colony counting. Data are the mean Ϯ S.E. of 6 mice/group.

Impaired Clearance of P. aeruginosa in MyD88
Ϫ/Ϫ Mice-We used MyD88-deficient mice to examine the role of MyD88 in the host defense against P. aeruginosa lung infection in vivo. MyD88 ϩ/ϩ and MyD88 Ϫ/Ϫ mice were inoculated intranasally with 1 ϫ 10 9 CFU of P. aeruginosa strain 8821. BALF and lung tissue were collected 24 h later for the detection of viable bacteria by counting CFU. MyD88 ϩ/ϩ mice are able to clear P. aeruginosa (Fig. 1). Significantly higher CFU counts can be seen in both the BALF and lung tissue from MyD88 Ϫ/Ϫ mice (Fig. 1), suggesting that MyD88 Ϫ/Ϫ mice demonstrate impaired bacterial clearance in comparison with MyD88 ϩ/ϩ mice.
Impaired Neutrophil Recruitment to the Lung in MyD88 Ϫ/Ϫ Mice-Given that neutrophils are essential for the clearance of P. aeruginosa during acute lung infection (15), we tested whether the impaired bacterial clearance in MyD88 Ϫ/Ϫ mice FIG. 2. Defective neutrophil recruitment into the airways of MyD88 ؊/؊ mice following P. aeruginosa lung infection. MyD88 ϩ/ϩ and MyD88 Ϫ/Ϫ mice were inoculated intranasally with P. aeruginosa (Psa) (mucoid strain 8821, 1 ϫ 10 7 CFU/mouse). Mice that were not treated with bacteria or those that received saline were used as controls (saline). a and b, after 4 or 8 h, BALF and lung tissue were collected for the determination of MPO activities. Data are the mean Ϯ S.E. of 5-10 mice/group (*, p Ͻ 0.05, compared with the saline group). c-f, 8 h after infection, mice were sacrificed, and the upper lobe of the left lung was collected for hematoxylin-eosin staining. c, MyD88 ϩ/ϩ mouse lung at ϫ40; d, MyD88 ϩ/ϩ mouse lung at ϫ100; e, MyD88 Ϫ/Ϫ mouse lung at ϫ40; f, MyD88 Ϫ/Ϫ mouse lung at ϫ100. was caused by defective neutrophil infiltration. MyD88 ϩ/ϩ and MyD88 Ϫ/Ϫ mice were inoculated intranasally with P. aeruginosa strain 8821. After 4 or 8 h, the BALF and lung tissue were collected for measuring MPO activities. In MyD88 ϩ/ϩ mice, infection with P. aeruginosa for 4 h induced an increased activity of MPO in the lung but not in the BALF, suggesting that at this time point neutrophils in circulation were recruited to the lung but had not yet reached the airways. After 8 h of infection, a significant increase of MPO activities was observed in both BALF and lung tissue in MyD88 ϩ/ϩ mice (Fig. 2, a and  b). Strikingly, little change in MPO activity was observed in either BALF or lung tissue after 4 or 8 h of P. aeruginosa lung infection in the MyD88 Ϫ/Ϫ mice (Fig. 2, a and b), suggesting an impaired neutrophil recruitment to the lung in MyD88 Ϫ/Ϫ mice. This finding is confirmed by histological examination of the lung. Lung tissue from MyD88 Ϫ/Ϫ mice showed little neutrophil infiltration after 8 h of P. aeruginosa infection, compared with the prominent neutrophil recruitment in MyD88 ϩ/ϩ mice (Fig. 2, c-f).
P. aeruginosa-induced Production of MIP-2, TNF, and IL-1␤ in MyD88 ϩ/ϩ but Not in MyD88 Ϫ/Ϫ Mice-MyD88 Ϫ/Ϫ neutrophils have normal migration ability in vivo because neutrophil recruitment to the septic focus has been shown to be normal in MyD88 Ϫ/Ϫ mice during polymicrobial infection (8). Thus, we reasoned that the impaired neutrophil recruitment in the lungs of MyD88-deficient mice might be caused by a defect in the production of neutrophil chemoattractants. Accordingly, several cytokines and chemokines that are important for neutrophil migration in vivo were examined. These include the direct neutrophil chemoattractant MIP-2 (33, 34) and indirect attractants TNF and IL-1␤ (17). To examine MIP-2 mRNA expression by real-time RT-PCR, lung tissue from MyD88 ϩ/ϩ and MyD88 Ϫ/Ϫ mice after 4 h of P. aeruginosa infection was used to isolate RNA. A significant increase of MIP-2 mRNA expression was seen in MyD88 ϩ/ϩ mice. In contrast, there was little MIP-2 response in MyD88 Ϫ/Ϫ mice (Fig. 3a). To determine MIP-2 production at the protein level, BALF and lung tissue homogenates from MyD88 ϩ/ϩ and MyD88 Ϫ/Ϫ mice after 4 or 8 h of P. aeruginosa lung infection were used to determine MIP-2 protein concentration by ELISA. P. aeruginosa infection induced a significant increase of MIP-2 in the BALF and lung tissue in MyD88 ϩ/ϩ mice. In contrast, there was little MIP-2 production in MyD88 Ϫ/Ϫ mice (Fig. 3, b and c).
Levels of TNF and IL-1␤ were also examined because of their essential roles in neutrophil recruitment to the lung in vivo (17). In MyD88 ϩ/ϩ mice, P. aeruginosa lung infection induced an increased production of IL-1␤ in the lung tissue (Fig. 4b) and TNF in the BALF (Fig. 4c). In contrast, there was little response of IL-1␤ or TNF in lung tissue or BALF from MyD88 Ϫ/Ϫ mice (Fig. 4, b and c). It is also noteworthy that in wild type mice, P. aeruginosa-induced IL-1␤ remained largely in the lung tissue (Fig. 4b), whereas the majority of TNF was secreted in the BALF (Fig. 4c).
Given that IFN-␥ and IL-12 have been implicated in P. aeruginosa lung infection (35,36), these two cytokines were examined in both MyD88 ϩ/ϩ and MyD88 Ϫ/Ϫ mice. Mice were challenged for 4 or 8 h with an intranasal inoculation of P. aeruginosa (strain 8821, 1 ϫ 10 7 CFU/mouse). The BALF and lung tissue homogenates were used to determine the levels of production of IFN-␥ and IL-12 by ELISA. In contrast to the robust response of neutrophil attractants MIP-2, IL-1␤, and TNF, there was little IFN-␥ and IL-12 response in both MyD88 ϩ/ϩ and MyD88 Ϫ/Ϫ mice after 4 or 8 h of P. aeruginosa lung infection (data not shown).
To examine the role of MyD88 in P. aeruginosa-induced NF-B activation, lung tissue homogenates from P. aeruginosainfected (4 h) or untreated mice were used to isolate nuclear extracts for the determination of NF-B activation by electrophoretic mobility assay. P. aeruginosa-induced NF-B activation was seen in wild type mice (Fig. 5). In contrast, P. aeruginosa-induced NF-B activation was markedly reduced in MyD88 Ϫ/Ϫ mice. These data are consistent with the defect of P. aeruginosa-induced mediator production in MyD88 Ϫ/Ϫ mice as shown in Figs. 3 and 4.
Increased TLR2, but Not TLR4, mRNA Expression in the Lung after P. aeruginosa Stimulation-Because MyD88 is important in TLR signaling, the lack of a P. aeruginosa-induced response in MyD88 Ϫ/Ϫ mice suggests that TLR may have a critical role in P. aeruginosa-induced host responses. To examine whether P. aeruginosa infection induces changes in TLR2 and TLR4 expression, lung tissue from P. aeruginosa-infected (4 h) MyD88 ϩ/ϩ and MyD88 Ϫ/Ϫ mice was used to determine mRNA levels by real-time RT-PCR. Results were expressed as the -fold increase relative to the level in wild type untreated mice by using the average value from these animals as a calibrator. As shown in Fig. 6a, the basal levels of TLR2 expression in MyD88 ϩ/ϩ and MyD88 Ϫ/Ϫ mice were similar. Interestingly, P. aeruginosa infection induced a greater increase in TLR2 expression in MyD88 ϩ/ϩ mice (7.5-fold) than in MyD88 Ϫ/Ϫ mice (1.7-fold) (Fig. 6a), supporting the concept that TLR2 is involved in P. aeruginosa lung infection. However, little change in TLR4 level in the lung was found in MyD88 ϩ/ϩ and MyD88 Ϫ/Ϫ mice (Fig. 6b).
To examine whether TLR4 plays a role in P. aeruginosainduced up-regulation of TLR2 expression, TLR4 mutant mice were used. Lung tissue from saline-treated or P. aeruginosainfected (4 h) mice was subjected to real-time RT-PCR analysis for TLR2 mRNA. Data are expressed as the -fold increase relative to the level in saline-treated mice. As shown in Fig. 6a, P. aeruginosa infection had little effect on TLR2 expression, suggesting an important role for TLR4 in P. aeruginosa-induced TLR2 expression.
P. aeruginosa-induced Differential Responses in TLR2 Ϫ/Ϫ and TLR4 Mutant Mice-To determine the specific contributions of TLR2 and TLR4 in the development of early immune responses in the lung following P. aeruginosa infection, TLR2 Ϫ/Ϫ and TLR4 mutant mice as well as their corresponding control mice were infected intranasally with P. aeruginosa strain 8821 for 4 h. Lung homogenates were used to determine neutrophil infiltration (MPO) and cytokine production. TLR2 Ϫ/Ϫ mice showed a partial decrease in neutrophil recruitment compared with wild type mice (Fig. 7a). However, a significant P. aeruginosa-induced MIP-2 and IL-1␤ production was observed in TLR2 Ϫ/Ϫ mice (Fig. 7, b-d), suggesting that TLR2 alone may not be a major component in mediating P. aeruginosa-induced responses. Similarly, TLR4 mutant mice showed P. aeruginosa-induced production of MIP-2, IL-1␤, and TNF when compared with saline-treated mice (Fig. 7, f-h). Compared with TLR4 ϩ/ϩ mice, TLR4 mutant mice demon-strated a significant decrease of neutrophil recruitment and MIP-2, IL-1␤, and TNF production (Fig. 7, e-h). DISCUSSION Depending on the nature of the microbial infection, the mechanism of the host defense varies, including variations in the receptor usage, signaling pathways, cellular participants, and the pattern of gene expression (10,15). The mechanisms of the host defense against P. aeruginosa lung infection in vivo remain incompletely defined. The lung has a unique relationship with the environment and has developed distinct strategies to defend itself from microbial invasion with innate immune mechanisms that are primarily responsible for the elimination of bacterial organisms (37,38). P. aeruginosa, which generates a mucoid phenotype after colonizing lung tissue, appears to have a specialized relationship with the lung (39). P. aeruginosa-induced lung damage is the major cause of death in cystic fibrosis patients (11) and accounts for 40% of the cause of deaths in people with ventilator-associated pneumonia (13). We attempted to determine the role of the TLR-MyD88 pathway in the host defense against P. aeruginosa lung infec- Lung tissue was collected from MyD88 ϩ/ϩ (C57BL/6), MyD88 Ϫ/Ϫ , or TLR4 mutant mice 4 h after intranasal administration of P. aeruginosa strain 8821 (1 ϫ 10 7 CFU/mouse). Mice that were not treated with bacteria or those that received saline were used as controls (saline). Total RNA isolated from the lungs was subjected to real-time RT-PCR analysis for TLR2 or TLR4 expression. For comparison between MyD88 ϩ/ϩ and MyD88 Ϫ/Ϫ mice, data are expressed as -fold increase relative to MyD88 ϩ/ϩ control mice (normal basal expression). In the study of TLR2 expression in TLR4 mutant mice, data are expressed as -fold increase relative to saline-treated TLR4 mutant mice. Data are the mean Ϯ S.E. of 3-6 mice/group (*, p Ͻ 0.01, compared with the saline group).
tion. Consistent with a recent finding (40), our results suggest that the MyD88-dependent pathway is a central component of the initiation of P. aeruginosa-induced early immune responses in the lung, leading to the clearance of this bacterium. Because neutrophils play a major role in the clearance of P. aeruginosa from the lung, the defective clearance of P. aeruginosa seen in MyD88 Ϫ/Ϫ mice is likely caused by a deficient recruitment of neutrophils into the airways. Given that MyD88 Ϫ/Ϫ neutrophils appear to have a normal migratory ability (8), the defective influx of neutrophils into the airways is most likely the result of insufficient production of neutrophil attractants in the lung of MyD88 Ϫ/Ϫ mice as observed in this study. Accordingly, our results support the model in which, during acute P. aeruginosa lung infection, MyD88 is absolutely required for the early FIG. 7. P. aeruginosa-induced differential immune responses in the lung in TLR2 ؊/؊ and TLR4 mutant mice. Lung tissue and BALF were collected from TLR2 Ϫ/Ϫ and TLR4 mutant mice and their respective control mice after intranasal infection with P. aeruginosa (Psa) (mucoid strain 8821, 1 ϫ 10 7 CFU/mice) for 4 h. Mice that were not treated with bacteria or those that did not receive saline served as controls (saline). Lung homogenates were used to determine MPO activity (a and e) and MIP-2 (b and f) and IL-1␤ (c and g) production. TNF (d and h) was measured using BALF samples. Data are the mean Ϯ S.E. of 12-16 mice/group. production of the cytokines and chemokines MIP-2, IL-1␤, and TNF, which are responsible for neutrophil recruitment and subsequent bacterial clearance. However, this does not exclude the possibility of the involvement of MyD88-independent pathways such as the TRIF/TRAM pathway in the P. aeruginosainduced host response, especially in the later phase of the infection. Several IFN-regulatory factor (IRF) 3-regulated cytokines and chemokines such as RANTES and IFN-␣ inducible protein 10 are induced by the TRIF/TRAM pathway in response to bacterial LPS stimulation (41,42). RANTES and IFN-␣ inducible protein 10 are up-regulated during P. aeruginosa infection (43). Thus, the role for the TRIF/TRAM pathway in P. aeruginosa-induced lung infection requires further study.
Given the importance of MyD88 in TLR signaling (1) and the circumstantial evidence of the association between P. aeruginosa and TLR2 or TLR4 (24 -26), immune responses to P. aeruginosa infection were examined in MyD88 Ϫ/Ϫ , TLR2 Ϫ/Ϫ , and TLR4 mutant mice, with the goal of assessing the relative contribution of these molecules as well as additional TLRs in P. aeruginosa infection. Apparently, the pattern of P. aeruginosa-induced immune responses in TLR2 Ϫ/Ϫ or TLR4 mutant mice is different from that seen in MyD88 Ϫ/Ϫ mice. These results suggest that neither TLR2 nor TLR4 functions individually as the only component responsible for P. aeruginosainduced immune response in the lung in vivo. This difference suggests that additional TLRs, such as TLR5 (24,44), or synergistic effects between different TLRs (45) may be involved. Nevertheless, the increase of TLR2 expression in the lung after P. aeruginosa infection and the partial inhibition of immune responses in TLR2 Ϫ/Ϫ mice support a role for TLR2 in P. aeruginosa-induced lung inflammation. This is consistent with recent in vitro studies (46) demonstrating that TLR2 is involved in the activation of macrophages and monocytes by mannuronic acid polymers, a component of P. aeruginosa alginate that is specifically related to the mucoid phenotype (26). Interestingly, TLR2 is mobilized into an apical lipid raft receptor complex in epithelial cells after P. aeruginosa stimulation. In addition, TLR2 together with TLR5 and gangliotetraosylceramide have been shown to be involved in P. aeruginosa flagella-induced epithelial cell activation in vitro (24).
Although the level of TLR4 mRNA was not changed, the inhibition of P. aeruginosa-induced neutrophil infiltration and production of MIP-2, IL-1␤, and TNF in TLR4 mutant mice supports a role for TLR4 in P. aeruginosa lung infection. However, unlike the nearly complete inhibition in MyD88 Ϫ/Ϫ mice, a significant P. aeruginosa-induced production of MIP-2, IL-1␤, and TNF was observed in TLR4 mutant mice. Other studies (47,48) have demonstrated P. aeruginosa-induced immune responses in C3H/HeJ mice. Thus, other TLRs in addition to TLR4 or synergistic effects between TLRs may be involved in host responses to P. aeruginosa lung infection.
The increase in TLR2 mRNA expression and the unchanged TLR4 mRNA level in P. aeruginosa-infected lungs suggest that TLR2 and TLR4 may have different roles during P. aeruginosa infection. It is likely that the contribution of TLR2 may increase when the infection progresses, considering the 7.5-fold increase of TLR2 mRNA in the lungs of P. aeruginosa-infected wild type mice. The P. aeruginosa-induced increase of TLR2 mRNA was not observed in TLR4 mutant mice. This suggests that P. aeruginosa up-regulates TLR2 through activation of TLR4. It is widely recognized that TLR4 is associated with LPS-induced tolerance before exposure to bacterial LPS, leading to a state of hyporesponsiveness to subsequent LPS stimulation (49). Thus, it is possible that TLR4 plays a major role during the early phase of P. aeruginosa infection, whereas TLR2 may play a greater role as the infection progresses.
It has been well established that MyD88 transduces cell surface signals to transcription factor NF-B, which regulates NF-B-dependent gene expression, including TNF and IL-1␤ (1). P. aeruginosa-induced NF-B activation has been reported previously (50) and was confirmed in our study in the infected lung tissue using an electrophoretic mobility shift assay. The defective production of TNF and IL-1␤ in MyD88 Ϫ/Ϫ mice is likely caused by the blockade of signaling from TLRs to NF-B in these animals, because P. aeruginosa-induced NF-B activation in the lung was markedly reduced in MyD88 Ϫ/Ϫ mice.
Interestingly, unlike RANTES and IFN-␣ inducible protein 10, which can be regulated in a MyD88-NF-B-independent manner (42), MyD88 Ϫ/Ϫ mice showed an almost complete deficiency in MIP-2 expression of both mRNA and protein levels. RANTES and IFN-␣ inducible protein 10 contain IRF binding sites in their promoters and are regulated through the TRIF-IRF3/7 pathway during TLR activation (42,51). In contrast, there is no report regarding the existence of an IRF-binding site in the MIP-2 promoter. Although the MIP-2 promoter contains a conserved NF-B consensus motif (52), TLR-mediated MIP-2 production appears to be NF-B-independent because the NF-B inhibitor pyrrolidinedithiocarbamate blocks LPS-induced TNF and IL-1␤ production but has no effect on MIP-2 expression (53). Thus, MyD88 may regulate MIP-2 production through additional transcription factors rather than through NF-B. Alternatively, MyD88-dependent MIP-2 production may be secondary to the production of TNF and IL-1␤ because IL-1␤ and TNF are able to regulate MIP-2 expression (54,55). Thus, the role for MyD88 in the host defense against P. aeruginosa lung infection is likely to rely on the interplay of multiple P. aeruginosa-induced mediators in the lung, such as MIP-2, TNF, and IL-1␤.
In summary, our results suggest that the MyD88-dependent pathway plays an essential role in P. aeruginosa-induced early immune responses, including MIP-2, IL-1␤, and TNF production and subsequent neutrophil recruitment and bacterial clearance. Multiple MyD88-dependent TLRs, including TLR2 and TLR4, may be involved in the host defense against P. aeruginosa lung infection.