TLRs 2 and 4 Are Not Involved in Hypersusceptibility to Acute aeruginosa Lung Infections

This information is current as Reuben Ramphal, Viviane Balloy, Michel Huerre, Mustapha of September 29, 2021. Si-Tahar and Michel Chignard J Immunol 2005; 175:3927-3934; ; doi: 10.4049/jimmunol.175.6.3927 http://www.jimmunol.org/content/175/6/3927 Downloaded from

References This article cites 51 articles, 24 of which you can access for free at: http://www.jimmunol.org/content/175/6/3927.full#ref-list-1 http://www.jimmunol.org/ Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication

by guest on September 29, 2021 *average

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2005 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

TLRs 2 and 4 Are Not Involved in Hypersusceptibility to Acute Lung Infections1

Reuben Ramphal,* Viviane Balloy,† Michel Huerre,‡ Mustapha Si-Tahar,† and Michel Chignard2†

TLRs are implicated in defense against microorganisms. Animal models have demonstrated that the susceptibility to a number of Gram-negative pathogens is linked to TLR4, and thus LPS of many Gram-negative bacteria have been implicated as virulence factors. To assess the role of this pathogen-associated molecular pattern as it is exposed on intact Pseudomonas aeruginosa, the susceptibility of mice lacking TLR4 or both TLR2 and TLR4 was examined in a model of acute Pseudomonas . These mutant mice were not hypersusceptible to the Pseudomonas challenge and mounted an effective innate response that cleared the organism despite low levels of TNF-␣ and KC in the airways. Bacterial and neutrophil counts in the lung were similar in control ؊ ؊ and TLR-deficient mice at 6 and 24 h after infection. MyD88 / mice were, however, hypersusceptible, with 100% of mice dying Downloaded from within 48 h with a lower dose of P. aeruginosa. Of note there were normal levels of IL-6 and G-CSF in the airways of TLR mutant mice that were absent from the MyD88؊/؊ mice. Thus, the susceptibility of mice to P. aeruginosa acute lung infection does not go through TLR2 or TLR4, implying that Pseudomonas LPS is not the most important virulence factor in acute pneumonia caused ,by this organism. Furthermore, G-CSF treatment of infected MyD88؊/؊ mice results in improved clearance and survival. Thus the resistance to infection in TLR2/TLR4؊/؊ mice may be linked to G-CSF and possibly IL-6 production. The Journal of Immunology, 2005, 175: 3927–3934. http://www.jimmunol.org/

seudomonas aeruginosa is the major cause of morbidity The role of LPS as a virulence factor in lung disease has been in cystic fibrosis (CF),3 where there is a state of chronic under considerable investigation for a number of organisms (8, 12, P colonization, punctuated by recurrent exacerbations. This 13). Although the inhalation of P. aeruginosa LPS has been re- organism is also a major cause of acute nosocomial in ported to result in severe lung inflammation (9), a critical role for patients who are ventilated (1). Besides the very high frequency of LPS expressed by the whole organism has only been occasionally infection in this patient population, P. aeruginosa is notable for the demonstrated to be responsible for susceptibility. Indeed, there are very high rate of attributable mortality in acute pulmonary infec- studies with C3H/HeJ mice, which have a loss-of-function muta- by guest on September 29, 2021 tions (1) that occurs even in the face of appropriate effective an- tion in the TLR4 gene, that do not support the idea that the LPS of tibiotics (2). Given our current state of knowledge concerning bac- this organism is the primary virulence factor (14, 15). It is believed terial virulence, inflammation, and disease, it may be deduced that that the proinflammatory action of LPS is mediated through TLR4 this susceptibility is due either to intoxication by a virulence factor in cooperation with CD14 and MD-2 molecules making up the or to the innate immune response, with severe inflammation re- TLR4 receptor complex (16), although studies by Pier et al. in sulting from the response to a virulence factor(s). In humans, there 1981 (17) indicated that C3H/HeJ mice do respond to Pseudomo- is an association between the production of the type III secreted nas LPS. However, the LPS of P. aeruginosa has been reported to toxins (TTSTs) of P. aeruginosa and the severity of the lung dis- be recognized by both TLR4 and TLR2, depending on the origin ease and mortality caused by this organism (3). Animal studies of the strain of P. aeruginosa and the cell types used to examine with certain, but not all, strains of this organism also suggest the the TLR4-LPS interaction (18). Thus, there may be other pathways involvement of these toxins (4–6), but other virulence factors, by which cells respond to LPS. LPS from a strain adapted to the such as LPS (7–10), phospholipase C (10), and flagellin (11), have CF respiratory tract was recognized by human TLR4, whereas LPS also been suggested to play roles. from an environmental strain was not recognized (18). This dif- ferential recognition occurred with human cells, but not with *Department of Medicine, University of Florida, Gainesville, FL 32610; †Institut mouse cells, i.e., mouse TLR4 recognizes LPS from the environ- Pasteur, De´fense Inne´e et Inflammation, Paris, France; Institut National de la Sante´et de la Recherche Me´dicale, E336, Paris, France; and ‡Institut Pasteur, Recherche et mental strain (18). In contrast, LPS from a nonadapted or envi- Expertise Histotechnologie et Pathologie, Paris, France ronmental strain of P. aeruginosa is recognized by human TLR2 Received for publication December 9, 2004. Accepted for publication July 8, 2005. (19). Such strains, coming from the environment, are likely to be The costs of publication of this article were defrayed in part by the payment of page found in acute lung disease. Thus, any analysis of the role of charges. This article must therefore be hereby marked advertisement in accordance Pseudomonas LPS in animal models of disease needs to consider with 18 U.S.C. Section 1734 solely to indicate this fact. these variations in recognition of LPS by TLRs. To dissect the 1 This work was supported by National Institutes of Health Grants AI45014 and AI47852 (to R.R.). relevant virulence factors in acute lung disease due to P. aerugi- nosa, we have begun by examining the responses of mice that have 2 Address correspondence and reprint requests to Dr. Michel Chignard, Institut Pas- teur, De´fense Inne´e et Inflammation, Institut National de la Sante´et de la Recherche deletions in the genes encoding either TLR4 or both TLR2 and Me´dicale E336, 25 rue du Dr. Roux, Paris 75015, France. E-mail address: TLR4. We find that neither mutation confers increased suscepti- [email protected] bility to acute lung infections produced by the intratracheal injec- 3 Abbreviations used in this paper: CF, cystic fibrosis; BAL, bronchoalveolar lavage; LB, Luria Bertoni; PAMP, pathogen-associated molecular pattern; PMN, polymor- tion of a reasonably well-characterized laboratory strain of P. phonuclear neutrophil; TTST, type III secreted toxin; WT, wild type. aeruginosa. However, we note that LPS-independent susceptibility

Copyright © 2005 by The American Association of Immunologists, Inc. 0022-1767/05/$02.00 3928 TLRs AND P. aeruginosa

to the lung infection in this animal model goes through the MyD88 LD50 doses for the normal control mice. Similar survival experiments were Ϫ Ϫ pathway, thus implicating another TLR or another receptor sharing performed with the MyD88 / mice using one-tenth of the wild-type (WT) LD to ascertain that susceptibility was mediated through a MyD88-depen- a similar signaling pathway. We observed that this MyD88 path- 50 dent pathway and to verify that the clone of strain PAK used in these exper- way is essential for IL-6 and G-CSF production, and that G-CSF iments was capable of causing rapid death in a susceptible mouse strain. treatment was partially protective of P. aeruginosa-infected A last series of animal experiments was performed to ascertain the ef- MyD88Ϫ/Ϫ mice. fects of the TLR and MyD88 mutations in defense against this strain of P. aeruginosa. Groups of eight mice were infected by the intratracheal route

using approximately one-tenth of the LD50 for the control mice. Bron- Materials and Methods choalveolar lavages (BAL) were performed on these mice 6 and 24 h after Bacterial strain and growth conditions used infection, after pentobarbital euthanasia. The lavage fluids were diluted or used undiluted and plated on LB agar plates to obtain viable bacterial P. aeruginosa P. aeruginosa The strain PAK, a widely studied strain of counts in the lavage fluid. Total and differential cell counts were performed originally obtained from S. Lory (Harvard University, Boston, MA) was on the lavage fluid. Total cell counts were measured in the BAL fluids with grown overnight in Luria Bertoni (LB) broth, then transferred to fresh a Coulter counter (Coulter Electronics), and cell differential counts were medium and grown for 4–5 h to midlog phase. The culture was centrifuged determined after cytospin centrifugation and staining with Diff-Quick prod- ϫ g at 3000 for 15 min, and the cell pellet was washed twice with cold ucts. Murine cytokine concentrations in BAL fluid were determined using PBS. For LD determinations, the pellet was suspended in one-tenth the 50 DuoSet ELISA kits obtained from R&D Systems and the Multiplex assay original volume. The bacterial count was ascertained by plating serial di- was performed with a Bio-Plex cytokine assay kit (Bio-Rad). Finally, in a lutions on LB agar plates. For other determinations, the bacterial pellet was Ϫ Ϫ last set of experiments, MyD88 / mice were treated s.c. with 3 ␮gof diluted in its original volume, and the OD was adjusted to give the ap- G-CSF/mouse (Neupogen; Amgen Europe) given 1 h before and at the time proximate desired inocula. The inocula were verified by serial 10-fold di- of infection, followed by the same treatment at 24 and 48 h after infection.

lutions of the bacterial suspensions and plating on LB agar. This strain of Downloaded from P. aeruginosa is known to contain and express a full complement of vir- Histology ulence factors, including pili, flagella, and type III secreted exoenzymes S, T, and Y and has a smooth LPS belonging to serotype 6. Mutant and WT mice infected with one-tenth of the LD50 for the control mice were killed with pentobarbital 24 h after infection. The lungs were Mouse strains fixed in formalin, sectioned and stained with H&E. Males from several mouse strains were used for the challenge experiments. Calculations Ϫ/Ϫ Ϫ/Ϫ Ϫ/Ϫ

TLR2 , TLR4 , and MyD88 mice were obtained from S. Akira http://www.jimmunol.org/ (Osaka University, Osaka, Japan) and were backcrossed eight times with The LD50 was calculated by the Reed-Muench method (22) using a com- C57BL/6 to ensure similar genetic backgrounds. Double-knockout TLR2/ puter program available at ͗http://ntri.tamuk.edu/cgi-bin/ld50/ld50͘ and TLR4Ϫ/Ϫ mice were generated by breeding TLR2Ϫ/Ϫ mice and TLR4Ϫ/Ϫ verified by probit analysis to obtain exact confidence intervals. Survival of mice. C57/BL6 mice from which these mice were derived were used as the the different mouse strains was compared using Kaplan-Meier analysis control mice. These latter mice were supplied by Centre d’Elevage R. log-rank test. Cytokine levels, polymorphonuclear neutrophil (PMN) Janvier and were used at ϳ8 wk of age. Mice were fed normal mouse chow counts, and bacterial counts were expressed as the mean Ϯ SEM. Differ- and water ad libitum and were reared and housed under standard conditions ences between groups were assessed for statistical significance using the with air filtration. Mice were cared for in accordance with Pasteur Institute Kruskal-Wallis ANOVA test, followed by the Mann-Whitney U test. A guidelines in compliance with European animal welfare regulations. value of p Ͻ 0.05 was considered statistically significant.

Animal infections Results by guest on September 29, 2021

Several different types of challenge experiments were performed. In the LD50 determinations in control and TLR mutant mice

first series of experiments, LD50 determinations were made on the control Ϫ/Ϫ Ϫ/Ϫ Ϫ/Ϫ LD50 determinations for the control C57BL/6 and TLR4 mice mice, TLR4 mice, and TLR2/TLR4 mice. LD50 determinations were not performed on MyD88Ϫ/Ϫ mice, because these mice were known were performed twice using a total of 45 mice for each mouse to be very susceptible to P. aeruginosa strain PAK (20). Mice were anes- strain in two experiments, because the doses given were slightly thetized by i.m. injection of a mixture of ketamine-xylazine and were Ϫ/Ϫ different. The LD50 determination for the TLR2/TLR4 mutant placed supine. A plastic catheter (diameter, 0.86 mm) was inserted into the mice was performed only once. The results are shown in Table I. trachea via the oropharynx. The proper insertion was verified by checking Ϫ/Ϫ ϳ the formation of mist due to expiration on a mirror placed in front of the The LD50 of the TLR4 mice was 3–4 times that of the con- Ϫ/Ϫ external end. A 50-␮l bacterial suspension was laid down at the internal trol mice, indicating that TLR4 mice were not hypersusceptible end of the catheter with a micropipette using a sterile disposable tip for gel to P. aeruginosa given directly into the lungs despite reports of loading that was introduced into the catheter. Mice were then immediately hypersusceptibility of such mice to Gram-negative bacteria. In held upright to facilitate bacterial inhalation and until normal breathing fact, the mice appeared to be slightly more resistant, but these resumed. This protocol allows highly reproducible infection of the whole lung and has been used in studies of invasive pulmonary Aspergillosis by values were not statistically significant using probit analysis. The Ϫ/Ϫ ϳ us (21). It may also mimic the main mechanism, aspiration, by which P. LD50 for the TLR2/TLR4 mice was 3 times that of the con- aeruginosa is acquired in acute infections, as opposed to aerosolization of trol mice, indicating that these mice were also not hypersusceptible the organisms, where much of the inoculum may bypass the airways and to P. aeruginosa. reach the alveoli depending on the particle size used. For the LD50 deter- minations, groups of five mice were infected by direct intratracheal injec- Mouse survival experiments tion of the bacterial strain, whose numbers were adjusted to deliver Ͼ108 CFU into the lungs at the highest concentration. Initial LD50 determina- To examine whether there were any differences in the time to death tions used five groups of mice receiving 10-fold dilutions; however, later that may be reflective of how the different mutant mice handle the determinations required only four groups to obtain both 100 and 0% mortality. Pseudomonas infection, survival experiments were performed us- ing between one and two LD doses of the microorganism for the After ascertaining the LD50 of the mouse strains, the survival times of 50 the different mouse strains were examined using between one and two control C57BL/6 mice. The survival curves are shown in Fig. 1.

a Table I. LD50 determinations for WT and TLR mutant mice

WT TLR4Ϫ/Ϫ TLR2, 4Ϫ/Ϫ

1.5 ϫ 107 (1.8 ϫ 106–1.25 ϫ 108) 5.0 ϫ 107 (2.1 ϫ 107–1.4 ϫ 108) 4.2 ϫ 107 (2.0 ϫ 107–2.5 ϫ 108)

a LD50 determination and confidence intervals. Differences are not statistically significant by probit analysis. The Journal of Immunology 3929

FIGURE 1. Survival curves of WT C57BL6 (WT), TLR4Ϫ/Ϫ, and TLR2/TLR4Ϫ/Ϫ mice. TLR mutant mice were given the same dose (1–2 FIGURE 3. PMN count from the BAL fluid of WT and TLR-deficient LD50) used for the WT mice. Ten mice per group were used for WT and TLR mutant mice. mice 6 and 24 h after intratracheal injection of a sublethal dose of P. aeruginosa. Data are the mean Ϯ SEM from four or five mice per time point. Differences are not statistically significant. Analysis of these data using a Kaplan-Meier test indicated that there was no significant difference in time to death when the TLR wide disparity of the values, the difference was only important for knockout mice were compared with the control C57BL/6 mice and the double mutant compared with the WT mice. Notably, however, confirmed that there were no significant differences in the suscep- IL-6 and G-CSF production was not impaired 6 h after infection, Downloaded from tibility of the TLR knockout mice vs the control mice, because the confirming the observations on IL-6 made in the previous assays. survival rates were similar using a fixed dose of bacteria. Ϫ/Ϫ Ϫ Ϫ Bacterial counts and inflammatory response of MyD88 mice Bacterial counts and inflammatory response of TLR4 / and TLR2/TLR4Ϫ/Ϫ mice MyD88 is an adaptor molecule for the signaling pathway of TLR2 and TLR4 as well as for most of the other TLR. The susceptibility C57BL/6 control mice and TLR knockout mice were infected in- Ϫ/Ϫ of MyD88 mice reported by Power et al. (23) was confirmed in http://www.jimmunol.org/ tratracheally with one-tenth of the LD50 of P. aeruginosa used for our hands. Thus, using ϳ1⁄10 to 1⁄20 of the bacterial dose used for control mice. Animals were killed 6 and 24 h after infection. BAL control mice, MyD88Ϫ/Ϫ mice died within 72 h (Fig. 5A). To gain fluids were cultured to obtain bacterial numbers in the airspaces insight into this increased susceptibility, BAL fluids were collected and were examined for total cell count, differential cell counts, and at 6 and 24 h. There was no appreciable PMN response 6 h after cytokine/chemokine content. infection in MyD88Ϫ/Ϫ mice, but there was a response at 24 h, Bacterial counts obtained from the BAL fluid of the different which was ϳ5 times lower than that in WT mice (Fig. 5B). By mouse strains are shown in Fig. 2. There were no significant dif- ferences in the numbers of P. aeruginosa obtained at the time

intervals indicated. The total PMN counts performed on the same by guest on September 29, 2021 BAL fluids were similar in the TLR4 mutant and WT mice and were slightly reduced in the TLR double-mutant animals (Fig. 3). Cytokine determinations on the BAL fluids indicated an impaired production of TNF-␣ and KC in response to the P. aeruginosa challenge in the TLR4Ϫ/Ϫ mice and the TLR2/TLR4Ϫ/Ϫ mice, but there was no impairment of the IL-6 response 6 h after infection in these mice (Fig. 4). BAL fluids were then further analyzed by a multiplex cytokine assay to determine whether other cytokines, not measured in the initial assays were similarly unaffected by the TLR mutations. A total of 18 cytokines/chemokines were measured (Ta- ble II). A number of these were present in small quantities and were not different in the three mouse lines. The production of several cytokines/chemokines was impaired in the mutant mice, as noted in the ELISA, e.g., TNF-␣ and KC production, but addi- tionally RANTES and MIP-1␣ production, were also impaired to a large extent. IL-1␤ production was also reduced, but due to the

FIGURE 4. Cytokine/chemokine concentrations in the BAL fluid of WT and TLR-deficient mice 6 and 24 h after intratracheal injection of a FIGURE 2. CFU of P. aeruginosa isolated from the BAL fluid of WT sublethal dose of P. aeruginosa. Six hours after infection, TLR mutant and TLR-deficient mice 6 and 24 h after intratracheal injection of a sub- mice show a dramatic impairment of TNF-␣ (A) and KC (B) production, lethal dose of bacteria. Data are the mean Ϯ SEM of four or five mice per but show an IL-6 response (C) similar to that of WT mice. Data are the time point. Differences are not statistically significant. mean Ϯ SEM from four or five mice per time point. 3930 TLRs AND P. aeruginosa

Table II. BAL fluid cytokine/chemokine responses to P. aeruginosa infection of the lungsa

WT TLR4Ϫ/Ϫ TLR2, 4Ϫ/Ϫ

Very low levels IL-2 7.4 Ϯ 0.2 5.3 Ϯ 0.5 5.2 Ϯ 0.3 IL-3 1.4 Ϯ 0.2 0.4 Ϯ 0.1 0.3 Ϯ 0.1 IL-4 0.5 0.4 0.4 Low levels–importantly reduced IL-1␣ 44.4 Ϯ 7.9 25.5 Ϯ 1.9 25.7 Ϯ 2.3 IL-10 11.4 Ϯ 0.6 5.8 Ϯ 0.6 5.4 Ϯ 0.3 IL-12 (p40) 23.3 Ϯ 1.3 11.6 Ϯ 4.1 18.3 Ϯ 4.0 IL-12 (p70) 13.4 Ϯ 1.4 7.7 Ϯ 0.8 7.3 Ϯ 0.8 IL-17 19.5 Ϯ 0.5 11.3 Ϯ 1.7 15.5 Ϯ 1.1 IFN-␥ 9.0 Ϯ 0.7 5.5 Ϯ 1.1 6.4 Ϯ 0.8 High levels–very importantly reduced (Ͼ80%) KC 1597 Ϯ 187 416 Ϯ 177 328 Ϯ 64 MIP-1␣ 1592 Ϯ 501 103 Ϯ 25 96.4 Ϯ 5.7 RANTES 115 Ϯ 16.9 17.5 Ϯ 4.4 17.7 Ϯ 2.0 TNF-␣ 4814 Ϯ 883 234 Ϯ 80 112 Ϯ 5.0 High levels–reduced IL-1␤ 383 Ϯ 116 246 Ϯ 57 173 Ϯ 36 Downloaded from Low levels–not importantly modified IL-5 12.8 Ϯ 0.4 14.6 Ϯ 6.3 18.7 Ϯ 2.6 GM-CSF 45 Ϯ 2.0 53 Ϯ 24 43 Ϯ 9.0 High levels–not importantly modified G-CSF 2386 Ϯ 419 1582 Ϯ 439 2082 Ϯ 400 IL-6 2212 Ϯ 124 2098 Ϯ 400 2424 Ϯ 96 http://www.jimmunol.org/ a BAL fluid cytokine/chemokine responses to P. aeruginosa infection of the lungs measured by a multiplex assay 6 h after infection (picograms per milliliter of BALF). Ϫ/Ϫ Significance refers to differences between the mutant mice (TLR4 and TLR2, FIGURE 5. Survival (A), PMN count (B), and bacterial count (C) from 4Ϫ/Ϫ) and the WT. BAL fluid of WT and MyD88Ϫ/Ϫ mice infected with the same sublethal dose of P. aeruginosa. Data are the mean Ϯ SEM from four or five mice per time point. contrast, the bacterial counts in MyD88Ϫ/Ϫ mice increased dra- matically 24 h after infection (Fig. 5C). To determine whether this Ϫ Ϫ susceptibility of MyD88 / mice was connected to a lack of IL-6 their survival. The log-rank test for comparisons of Kaplan-Meier and G-CSF, we also examined whether these cytokines were survival curves indicates a significant increase in survival of G- by guest on September 29, 2021 present in these mice. As illustrated in Fig. 6, both IL-6 and G-CSF CSF-treated mice ( p ϭ 0.009). As shown in Fig. 8, 48 h after Ϫ Ϫ productions were impaired in MyD88 / mice. As anticipated, infection all control mice had died, whereas 60% of the treated synthesis of TNF-␣ and KC was totally abolished under the same mice were alive. Another group of animals was infected and experimental conditions (Fig. 6). treated in the same way, but BAL were performed 6 h (four ani- mals per group) and 24 h (three animals per group) after infection Histological examination to ascertain whether the administration of G-CSF was a protective Histological examination was performed on the lungs collected factor against the pulmonary challenge. Consistent with the sur-

24 h after infection with one-tenth of the LD50 for WT mice. As vival data, G-CSF-treated mice were able to mount an inflamma- shown in Fig. 7, a diffuse and acute alveolitis was observed in WT tory PMN response and had ϳ1000-fold fewer bacteria in the mice (A1). Numerous PMN were present in the alveoli and around lungs (Fig. 8). the vessels (A2). For TLR4Ϫ/Ϫ (B1 and B2) and TLR2/TLR4Ϫ/Ϫ (C1 and C2) mice, the picture was not very different; there was a Discussion diffuse recruitment of PMN within the alveoli without necrosis or LPS or endotoxins have been implicated in the pathogenesis of hemorrhages. The inflammatory response, as best as can be judged, acute lung disease because of the demonstration that the instilla- Ϫ Ϫ was thus similar in all mouse groups except MyD88 / mice. tion of LPS from and Pasteurella hemolytica into Indeed, multiple focal alveolitis with inflammatory cells, espe- the lungs of animals produced lesions similar to those seen in cially aggregates of PMN and macrophages, were observed within animals with experimentally induced, acute Gram-negative pneu- these foci of alveolitis (D1), whereas few PMN were observed in monia (24). The subsequent observations that intratracheal admin- the unaffected alveoli (D2). istration of LPS induced IL-1␤ and TNF-␣ expression as well as an inflammatory infiltrate (25) suggested that LPS was an impor- Survival, bacterial counts, and inflammatory response of Ϫ/Ϫ tant virulence factor in the pathogenesis of Gram-negative bacte- G-CSF-treated MyD88 mice rial pneumonia. However, most experiments involving lung in- Due to the normal G-CSF production of TLR mutant mice and the flammation and LPS have been performed with isolated E. coli lack of such a response in MyD88Ϫ/Ϫ mice, which succumbed LPS (24, 26–28). Nevertheless, the LPS of P. aeruginosa has also rapidly to the pulmonary infection, MyD88Ϫ/Ϫ mice were infected been implicated as a pathogenetic factor in acute lung disease, with similar doses of P. aeruginosa (3 ϫ 105/mouse) and were because Pseudomonas LPS given to mice by inhalation, intrana- given 3 ␮g of G-CSF/mouse 1 h before and at the time of infection, sally or intratracheally, results in lung inflammation (7–10). followed by the same treatment 24 and 48 h after infection. Saline- Although these experimental approaches demonstrate that treated mice served as controls. G-CSF-treated and saline-treated Pseudomonas LPS is capable of eliciting an innate immune re- animals (eight per group) were observed for 4 days to measure sponse, proof that an LPS-mediated innate immune response via The Journal of Immunology 3931

FIGURE 6. Cytokine/chemokine concentrations from BAL fluid of WT and MyD88Ϫ/Ϫ mice infected with the same sublethal dose of P. aeruginosa. Data for WT mice are the same as those in Fig. 4 and were repeated for direct comparison, except for G-CSF mea- surement. MyD88Ϫ/Ϫ mice are unable to mount TNF-␣ and KC responses as seen in TLR mutant mice, but in contrast to these mice, IL-6 and G-CSF are barely de- tectable in MyD88Ϫ/Ϫ mice. Data are the mean Ϯ SEM from four or five mice per time point. Downloaded from

TLR2 or TLR4, is essential for clearance of whole organisms from been conclusively demonstrated. This kind of information would the lungs or that LPS-mediated inflammation via these TLRs is a be best obtained with the use of whole organisms carrying non- critical response for virulence during an actual infection has not signaling mutations in the lipid A portion of the Pseudomonas LPS molecule, but Pseudomonas lipid A mutants have been generally http://www.jimmunol.org/ nonviable. Another approach to elucidating the role of Pseudomo- nas LPS in lung infections is the use of mice that are unresponsive to LPS. Such mice, mainly C3H/HeJ that have a loss-of-function mutation of the tlr4 gene, have been described to be hypersuscep- tible to a variety of systemic Gram-negative infections (29–33). Pseudomonas challenge studies with such mice have not been con- clusive. George et al. (14) reported that intranasal challenge of by guest on September 29, 2021

FIGURE 7. Histological examination of the lungs of WT and TLR- and MyD88-deficient mice 24 h after infection of mice with sublethal doses of P. aeruginosa. Lung sections stained with H&E. The presented sections are FIGURE 8. Survival (A), PMN count (B), and bacterial count (C) from from WT (A1 and A2), TLR4Ϫ/Ϫ (B1 and B2), TLR2/TLR4Ϫ/Ϫ (C1 and BAL fluid of MyD88Ϫ/Ϫ mice treated or not with G-CSF and infected with C2), and MyD88Ϫ/Ϫ (D1 and D2) mice. Magnification: A1–D1, ϫ40 (left P. aeruginosa. Data are the mean Ϯ SEM from three or four mice per time column); A2–D2, ϫ400 (right column). point. 3932 TLRs AND P. aeruginosa

C3H/HeJ with several P. aeruginosa strains did not demonstrate Based on the data obtained with MyD88Ϫ/Ϫ mice, the present

differences in the LD50 between these mice and LPS-sensitive study suggests that susceptibility to or defense against acute mice. Similarly, Yu et al. (15) showed that for the P. aeruginosa Pseudomonas lung infection goes through a MyD88-dependent strain PAO1 given i.p., there was no difference in susceptibility pathway, but does not involve the two TLRs examined. A similar between C3H/HeJ mice and C57BL/6 mice. However, in a study to observation has been made with Legionella pneumophilia, another examine the role of TLR4 in the defense against P. aeruginosa, Gram-negative organism, where TLR4-deficient mice were also using the cytotoxic Pseudomonas strain PA103 to challenge C3H/ shown to respond normally to a lung challenge (44). These results HeJ mice via the intratracheal route (34), the authors concluded are in contrast with studies of other acute Gram-negative bacterial that TLR4 signaling was essential for survival. Another recent pneumonia models, such as H. influenzae (32), P. hemolytica (45), study (23) has also examined this question concerning the role of and (33), where innate immunity mediated LPS in susceptibility using a mucoid P. aeruginosa strain and has through TLR4 has been demonstrated to be necessary for survival. concluded that susceptibility does indeed go through MyD88, and The factors that determine this differential susceptibility among these Gram-negative bacteria, however, are unknown. It is possible that both TLR2 and TLR4 are involved. However, no LD50 or survival data were provided. It is of note that in both studies, P. that the agonistic activity of Pseudomonas LPS is lower than that aeruginosa strains were defective in many virulence factors (35). of other pathogen-associated molecular patterns (PAMPs) ex- Finally, a recent abstract report by Hajjar et al. (Toll 2004 Meeting, pressed by P. aeruginosa. Conversely, in H. influenzae, P. hemo- ͗www.umassmed.edu/toll2004/abstracts.cfm͘) appears to support lytica, and K. pneumoniae, the agonistic activity of LPS may be our findings that the control of P. aeruginosa replication is inde- higher than that of other PAMPs. Other PAMPs expressed by P. pendent of TLR2 and TLR4. It should, however, be pointed out aeruginosa that may account for the MyD88-dependent response Downloaded from that our conclusions are based on the study of a single invasive of the lung to infection are flagellin and CpG DNA through the strain of P. aeruginosa that produces exoenzyme S, and it is pos- activation of TLR5 and TLR9, respectively. However, LPS may sible that other strains may behave differently with regard to the still be a possible agonist; Schroeder et al. (46) reported that ␬ response mediated by TLR2 or TLR4. Undoubtedly, this study NF- B may be activated by a pathway that involves CF transmem- contrasts with those that have used purified LPS, where the inflam- brane conductance regulator-mediated uptake of LPS. Such a path- matory response is quite florid (9). The most likely explanation may way would allow these mutant mice to respond to LPS as an ag- http://www.jimmunol.org/ onist, as reported by Pier et al. (17) for C3H/HeJ mice. be that the actual amounts of LPS released during an acute pulmonary The only clues concerning what may mediate resistance to this infection may be much less than what is delivered by the inhalation of infection in TLR mutant mice are IL-6 and G-CSF production, pure LPS, partly accounting for differences seen. because these responses were elicited in TLR mutant mice and Besides LPS, the TTSTs are now being implicated as major were absent in MyD88 mutant mice. Both IL-6 and G-CSF have virulence factors in acute lung disease due to P. aeruginosa (3). been implicated in defense against P. aeruginosa. G-CSF stimu- Thus, an exoenzyme U mutant does not cause lung injury or death lates the proliferation of PMN, enhances their phagocytic and mi- in mice (34). The actions of these toxins may involve TLR4, be- crobicidal activities, and has recently been shown to increase the

cause exoenzyme S, an important TTST, has been reported to ac- by guest on September 29, 2021 survival of mice during the course of P. aeruginosa pneumonia tivate both TLR2 and TLR4 in vitro (36). The studies we describe, (47). The current study shows that the defect in MyD88 mutant however, do not support an important role for exoenzyme S acti- mice can be partially corrected by the administration of G-CSF, vation of TLR2 or TLR4 in determining resistance to this organism Ϫ Ϫ suggesting one major defect in MyD88 / mice is failure of re- P. aeruginosa or susceptibility, because the strain of used is known cruitment of PMN to the lungs, and that this response may have to produce large quantities of exoenzyme S (37). Porins of P. been protective in TLR mutant mice. IL-6 has also been implicated aeruginosa were also shown to be as effective as LPS in inducing in P. aeruginosa corneal infections (48). IL-6 knockout mice are leukocyte activation (38). Although this has not been demon- more susceptible to such infections, and exogenous IL-6 limits strated, Pseudomonas porins may be active through TLR2 as re- bacterial proliferation in these mice (48). It is also possible to link ported for neisserial and influenzae porins (39, 40). a well-recognized agonist, Pseudomonas flagellin (49), to IL-6, The findings of our study suggest that the failure to recognize LPS, because this protein induces IL-6 expression and secretion by cor- exoenzyme S, and porins, three known potent agonists of TLR2 neal cells (50) as well as after a lung challenge (51). However, and/or TLR4, does not result in hypersusceptibility. flagellin should have evoked TNF-␣, KC, and other responses via The conclusions achieved in these studies are consistent with TLR5, even in TLR4Ϫ/Ϫ mice (51) and TLR2/TLR4Ϫ/Ϫ mice, several other observations of the role of Pseudomonas LPS in trig- unless it uses a novel pathway to IL-6 synthesis in airway cells. gering an innate immune response. Hybiske et al. (41) recently Whatever the agonist is, it may use IL-6, G-CSF, and/or other reported that the application of strain PAK to the apical surface of unmeasured mediators to initiate the cellular response and goes polarized CF epithelial cells evoked a very low cytokine response, through the MyD88 pathway. and fewer genes were up-regulated than when the organism was In summary, the studies we report suggest that the recognition of placed basally. This observation suggests that LPS on bacterial LPS by TLR2 or the TLR4 complex is not central for susceptibility surfaces is not efficiently recognized by the apical surface of epi- to or defense against acute P. aeruginosa lung infections. This thelial cells, which may be explained by either TLR expression at suggests that the major PAMP may be another molecule that goes this location or even the poor agonist activity of Pseudomonas through a MyD88-dependent pathway. The studies also suggest LPS. Along the same lines of evidence, it has been reported that that there are novel aspects of the signaling pathways triggered by polarized airway epithelial cells, both normal and CF, do not ex- this microorganism that leads to IL-6, G-CSF, and possibly other press TLR4 on the apical surfaces, but in an intracellular compart- cytokine expressions that do not seem to parallel the expression of ment (42) or on the basolateral aspects (43). The latter report also other major inflammatory cytokines and that may be protective. showed that challenging the apical surfaces with Pseudomonas These studies, however, do not implicate this unknown pathway as LPS did not result in activation of NF-␬B, even after 24 h, but that the sole pathway, because there may be redundant innate immune TLR2 is activated in response to a whole organism challenge to responses to P. aeruginosa, such that removal of the TLR2 and these cell lines. TLR4 arms may still leave a protective response in place, and it is The Journal of Immunology 3933 conceivable that if this unknown pathway is blocked, another path- 19. Erridge, C., A. Pridmore, A. Eley, J. Stewart, and I. R. Poxton. 2004. Lipopoly- way, such as the TLR4 pathway, may respond to protect the host. saccharides of Bacteroides fragilis, Chlamydia trachomatis and Pseudomonas aeruginosa signal via Toll-like receptor 2. J. Med. Microbiol. 53: 735–740. However, our findings do suggest that the lipid A part of the LPS 20. Skerrett, S. J., H. D. Liggitt, A. M. Hajjar, and C. B. Wilson. 2004. Cutting edge: molecule is not the principal susceptibility-determining PAMP. myeloid differentiation factor 88 is essential for pulmonary host defense against Identification of this pathway, the agonist, and the cell type that Pseudomonas aeruginosa but not Staphylococcus aureus. J. Immunol. 172: 3377–3381. triggers the protective innate immune response in these mutant 21. Balloy, V., M. Huerre, J.-P. Latge, and M. Chignard. 2005. Differences in pat- mice should provide more fundamental information on under- terns of infections for corticosteroid treatment and chemotherapy in experimental standing acute lung infections caused by P. aeruginosa. invasive pulmonary aspergillosis. Infect. Immun. 73: 494–503. 22. Reed, L. J., and H. A. Muench. 1938. A simple method of estimating fifty percent endpoints. Am. J. Hygiene 27: 493–503. 23. Power, M. R., Y. Peng, E. Maydanski, J. S. Marshall, and T. Lin. 2004. The Disclosures development of early host response to Pseudomonas aeruginosa lung infection is The authors have no financial conflict of interest. critically dependent on MyD88 in mice. J. Biol. Chem. 279: 49315–49322. 24. Brogden, K. A., R. C. Cutlip, and H. D. Lehmkuhl. 1984. Response of sheep after localized deposition of lipopolysaccharide in the lung. Exp. Lung Res. 7: 123–132. References 25. Ulich, T. R., L. R. Watson, S. M. Yin, K. Z. Guo, P. Wang, H. Thang, and J. del Castillo. 1991. The intratracheal administration of endotoxin and cytokines. 1. Chastre, J., and J. Y. Fagon. 2002. Ventilator-associated pneumonia. I. Characterization of LPS-induced IL-1 and TNF mRNA expression and the Am. J. Respir. Crit. Care Med. 165: 867–903. LPS-, IL-1-, and TNF-induced inflammatory infiltrate. Am. J. Pathol. 138: 2. Osmon, S., S. Ward, V. J. Fraser, and M. H. Kollef. 2004. Hospital mortality for 1485–1496.

patients with bacteremia due to Staphylococcus aureus or Pseudomonas aerugi- Downloaded from 26. Chignard M., and V. Balloy. 2000. Neutrophil recruitment and increased perme- nosa. Chest 125: 607–616. ability during acute lung injury induced by lipopolysaccharide. Am. J. Physiol. 3. Hauser, A. R., E. Cobb, M. Bodi, D. Mariscal, J. Valles, J. N. Engel, and J. Rello. 279: L1083–L1090. 2002. Type III protein secretion is associated with poor clinical outcomes in 27. Faffe, D. S., V. R. Seidl, P. S. Chagas, V. L. Goncalves de Moraes, patients with ventilator-associated pneumonia caused by Pseudomonas aerugi- V. L. Capelozzi, P. R. Rocco, and W. A. Zin. 2000. Respiratory effects of li- nosa. Crit. Care Med. 30: 521–528. popolysaccharide-induced inflammatory lung injury in mice. Eur. Respir. J. 15: 4. Finck-Barbancon, V., J. Goranson, L. Zhu, T. Sawa, J. P. Wiener-Kronish, 85–91. S. M. Fleiszig, C. Wu, L. Mende-Mueller, and D. W. Frank. 1997. ExoU ex- 28. Mizgerd, J. P., M. L. Scott, M. R. Spieker, and C. M. Doerschuk. 2002. Functions Pseudomonas aeruginosa http://www.jimmunol.org/ pression by correlates with acute cytotoxicity and ep- of I␬B proteins in inflammatory responses to Escherichia coli LPS in mouse ithelial injury. Mol. Microbiol. 25: 547–557. lungs. Am. J. Respir. Cell Mol. Biol. 27: 575–582. 5. Smith, R. S., M. C. Wolfgang, and S. Lory. 2004. An adenylate cyclase-con- 29. Cross, A. S., J. C. Sadoff, N. Kelly, E. Bernton, and P. Gemski. 1989. Pretreat- trolled signaling network regulates Pseudomonas aeruginosa virulence in a ment with recombinant murine tumor necrosis factor ␣/cachectin and murine mouse model of acute pneumonia. Infect. Immun. 72: 1677–1684. interleukin 1␣ protects mice from lethal bacterial infection. J. Exp. Med. 169: 6. Tang, H. B., E. DiMango, R. Bryan, M. Gambello, B. H. Iglewski, 2021–2027. J. B. Goldberg, and A. Prince. 1996. Contribution of specific Pseudomonas 30. Wang, M., K. C. Jeng, and P. I. Ping. 1999. Exogenous cytokine modulation aeruginosa virulence factors to pathogenesis of pneumonia in a neonatal mouse or neutralization of interleukin-10 enhances survival in lipopolysaccharide- model of infection. Infect. Immun. 64: 37–43. hyporesponsive C3H/HeJ mice with Klebsiella infection. Immunology 98: 7. Hirakata, Y., N. Furuya, K. Tateda, M. Kaku, and K. Yamaguchi. 1993. In vivo 90–97. production of exotoxin A and its role in endogenous Pseudomonas aeruginosa 31. Bernheiden, M., J. M. Heinrich, G. Minigo, C. Schutt, F. Stelter, M. Freeman, septicemia in mice. Infect. Immun. 61: 2468–2473. D. Golenbock, and R. S. Jack. 2001. LBP, CD14, TLR4 and the murine innate by guest on September 29, 2021 8. Szarka, R. J., N. Wang, L. Gordon, P. N. Nation, and R. H. Smith. 1997. A immune response to a peritoneal Salmonella infection. J. Endotoxin Res. 7: murine model of pulmonary damage induced by lipopolysaccharide via intranasal 447–450. instillation. J. Immunol. Methods 202: 49–57. 32. Wang, X., C. Moser, J. P. Louboutin, E. S. Lysenko, D. J. Weiner, 9. Skerrett, S. J., T. R. Martin, E. Y. Chi, J. J. Peschon, K. M. Mohler, and J. N. Weiser, and J. M. Wilson. 2002. Toll-like receptor 4 mediates innate C. B. Wilson. 1999. Role of the type 1 TNF receptor in lung inflammation after immune responses to Haemophilus influenzae infection in mouse lung. J. Im- inhalation of endotoxin of Pseudomonas aeruginosa. Am. J. Physiol. 276: munol. 168: 810–815. L715–L727. 33. Branger, J., S. Knapp, S. Weijer, J. C. Leemans, J. M. Pater, P. Speelman, 10. Wieland, C. W., B. Siegmund, G. Senaldi, M. L. Vasil, C. A. Dinarello, and S. Florquin, and T. van der Poll. 2004. Role of Toll-like receptor 4 in Gram- G. Fantuzzi. 2002. Pulmonary inflammation induced by Pseudomonas aerugi- positive and Gram-negative pneumonia in mice. Infect. Immun. 72: 788–794. nosa lipopolysaccharide, phospholipase C, and exotoxin A: role of interferon 34. Faure, K., T. Sawa, T. Ajayi, J. Fujimoto, K. Moriyama, N. Shime, and regulatory factor 1. Infect. Immun. 70: 1352–1358. J. P. Wiener-Kronish. 2004. TLR4 signaling is essential for survival in acute lung 11. Liaudet, L., C. Szabo, O. V. Evgenov, K. G. Murthy, P. Pacher, L. Virag, injury induced by virulent Pseudomonas aeruginosa secreting type III secretory J. G. Mabley, A. Marton, F. G. Soriano, M. Y. Kirov, et al. 2003. Flagellin from toxins. Respir. Res. 5: 1. Gram-negative bacteria is a potent mediator of acute pulmonary inflammation in 35. Mahenthiralingam, E., M. E. Campbell, and D. P. Speert. 1994. Nonmotility and . Shock 19: 131–137. phagocytic resistance of Pseudomonas aeruginosa isolates from chronically col- 12. Brigham, K. L., and B. Meyrick. 1986. Endotoxin and lung injury. Am. Rev. onized patients with cystic fibrosis. Infect. Immun. 62: 596–605. Respir. Dis. 133: 913–927. 36. Epelman, S., D. Stack, C. Bell, E. Wong, G. G. Neely, S. Krutzik, K. Miyake, 13. Gatti, S., R. Faggioni, B. Echtenacher, and P. Ghezzi. 1993. Role of tumor ne- P. Kubes, L. D. Zbytnuik, L. L. Ma, et al. 2004. Different domains of Pseudo- crosis factor and reactive oxygen intermediates in lipopolysaccharide-induced monas aeruginosa exoenzyme S activate distinct TLRs. J. Immunol. 173: pulmonary oedema and lethality. Clin. Exp. Immunol. 91: 456–461. 2031–2041. 14. George, S. E., M. J. Kohan, M. I. Gilmour, M. S. Taylor, H. G. Brooks, 37. Sundin, C., M. C. Wolfgang, S. Lory, A. Forsberg, and E. Frithz-Lindsten. 2002. J. P. Creason, and L. D. Claxton. 1993. Pulmonary clearance and inflammatory Type IV pili are not specifically required for contact dependent translocation of response in C3H/Hej mice after intranasal exposure to Pseudomonas spp. Infect. exoenzymes by Pseudomonas aeruginosa. Microb. Pathog. 33: 265–267. Immun. 59: 3585–3591. 38. Cusumano, V., M. A. Tufano, G. Mancuso., M. Carbone, F. Rossano, M. T. Fera, 15. Yu, H., J. C. Boucher, N. S. Hibler, and V. Deretic. 1996. Virulence properties F. A. Ciliberti, E. Ruocco, R. A. Merendino, and G. Teti. 1997. Porins of Pseudo- of Pseudomonas aeruginosa lacking the extreme-stress sigma factor AlgU. Infect. monas aeruginosa induce release of tumor necrosis factor ␣ and interleukin-6 by Immun. 64: 2774–2781. human leukocytes. Infect. Immun. 65: 1683–1687. 16. da Silva Correia, J., K. Soldau, U. Christen, P. S. Tobias, and R. J. Ulevitch. 39. Massari, P., P. Henneke, Y. Ho, E. Latz, D. T. Golenbock, and L. M. Wetzler. 2001. Lipopolysaccharide is in close proximity to each of the proteins in its 2002. Cutting edge: immune stimulation by neisserial porins is Toll-like receptor membrane receptor complex transfer from CD14 to TLR4 and MD-2. J. Biol. 2 and MyD88 dependent. J. Immunol. 168: 1533–1537. Chem. 276: 21129–21135. 40. Galdiero, M., M. Galdiero, E. Finamore, F. Rossano, M. Gambuzza, 17. Pier, G. B., R. B. Markham, and D. Eardley. 1981. Correlation of the biologic M. R. Catania, G. Teti, A. Midiri, and G. Mancuso. 2004. Haemophilus influen- responses of C3H/HEJ mice to endotoxin with the chemical and structural prop- zae porin induces Toll-like receptor 2-mediated cytokine production in human erties of the lipopolysaccharides from Pseudomonas aeruginosa and Escherichia monocytes and mouse macrophages. Infect. Immun. 72: 1204–1209. coli. J. Immunol. 127: 184–191. 41. Hybiske, K., J. K. Ichikawa, V. Huang, S. J. Lory, and T. E. Machen. 2004. 18. Hajjar, A. M., R. K. Ernst, J. H. Tsai, C. B. Wilson, and S. I. Miller. 2002. Human Cystic fibrosis airway epithelial cell polarity and bacterial flagellin determine host Toll-like receptor 4 recognizes host-specific LPS modifications. Nat. Immunol. 3: response to Pseudomonas aeruginosa. Cell. Microbiol. 6: 49–63. 354–359. 42. Guillot, L., S. Medjane, K. Le-Barillec, V. Balloy, C. Danel, M. Chignard, and 3934 TLRs AND P. aeruginosa

M. Si-Tahar. 2004. Response of human pulmonary epithelial cells to lipopoly- cells and activates NF-␬B translocation. Proc. Natl. Acad. Sci. USA 99: saccharide involves Toll-like receptor 4 (TLR4)-dependent signaling pathways: 6907–6912. evidence for an intracellular compartmentalization of TLR4. J. Biol. Chem. 279: 47. Babalola, C. P., C. H. Nightingale, and D. P. Nicolau. 2004. Adjunctive efficacy 2712–2718. of granulocyte colony-stimulating factor on treatment of Pseudomonas aerugi- 43. Muir, A,. G. Soong, S. Sokol, B. Reddy, M. I. Gomez, A. Van Heeckeren, and nosa pneumonia in neutropenic and non-neutropenic hosts. J. Antimicrob. Che- A. Prince. 2004. Toll-like receptors in normal and cystic fibrosis airway epithelial mother. 53: 1098–1100. cells. Am. J. Respir. Cell Mol. Biol. 30: 777–783. 48. Cole, N., S. Bao, F. Stapleton, A. Thakur, A. J. Husband, K. W. Beagley, and 44. Lettinga, K. D., S. Florquin, P. Speelman, R. van Ketel, T. van der Poll, and M. D. Willcox. 2003. Pseudomonas aeruginosa keratitis in IL-6-deficient mice. A. Verbon. 2002. Toll-like receptor 4 is not involved in host defense against Int. Arch. Allergy Immunol. 130: 165–172. pulmonary infection in a mouse model. J. Infect. Dis. 49. McDermott, P. F., F. Ciacci-Woolwine, J. A. Snipes, and S. B. Mizel. 2000. 186: 570–573. High-affinity interaction between Gram-negative flagellin and a cell surface 45. Chapes, S. K., D. A. Mosier, A. D. Wright, and M. L. Hart. 2001. MHCII, Tlr4 polypeptide results in human monocyte activation. Infect. Immun. 68: 5525–5529. and Nramp1 genes control host pulmonary resistance against the opportunistic 50. Zhang, J., K. Xu, B. Ambati, and F. S. Yu. 2003. Toll-like receptor 5-mediated bacterium Pasteurella pneumotropica. J. Leukocyte Biol. 69: 381–386. corneal epithelial inflammatory responses to Pseudomonas aeruginosa flagellin. 46. Schroeder, T. H., M. M. Lee, P. W. Yacono, C. L. Canon, A. A. Gerecker, Invest. Ophthalmol. Vis. Sci. 44: 4247–4254. D. E. Golan, and G. B. Pier. 2002. CFTR is a pattern recognition molecule that 51. Honko, A. N., and S. B. Mizel. 2004. Mucosal administration of flagellin induces extracts Pseudomonas aeruginosa LPS from the outer membrane into epithelial innate immunity in the mouse lung. Infect. Immun. 72: 6676–6679. Downloaded from http://www.jimmunol.org/ by guest on September 29, 2021