Importance of the Kupffer Cell Network

Toll-Like Receptor 4 Dependence of Innate and Adaptive Immunity to Salmonella: Importance of the Kupffer Cell Network

This information is current as Andrés Vazquez-Torres, Bruce A. Vallance, Molly A. of October 2, 2021. Bergman, B. Brett Finlay, Brad T. Cookson, Jessica Jones-Carson and Ferric C. Fang J Immunol 2004; 172:6202-6208; ; doi: 10.4049/jimmunol.172.10.6202 http://www.jimmunol.org/content/172/10/6202 Downloaded from

References This article cites 38 articles, 20 of which you can access for free at: http://www.jimmunol.org/content/172/10/6202.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 October 2, 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 © 2004 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

Toll-Like Receptor 4 Dependence of Innate and Adaptive Immunity to Salmonella: Importance of the Kupffer Cell Network1

Andre´s Vazquez-Torres,* Bruce A. Vallance,† Molly A. Bergman,‡ B. Brett Finlay,† Brad T. Cookson,‡§ Jessica Jones-Carson,* and Ferric C. Fang2‡§

Mammalian cells recognize LPS from Gram-negative bacteria via the Toll-like receptor 4 (TLR4) complex. During experimental Salmonella infection, C3H/HeJ mice carrying a dominant-negative mutation in TLR4 exhibited delayed chemokine production, impaired NO generation, and attenuated cellular immune responses. However, dramatically enhanced bacterial growth within the Kupffer cell network before the recruitment of inﬂammatory cells appeared to be primarily responsible for the early demise of Salmonella-infected TLR4-deﬁcient mice. LPS-TLR4 signaling plays an essential role in the generation of both innate and adaptive immune responses throughout the course of infection with Gram-negative bacteria. Alternative pattern-recognition receptors Downloaded from cannot completely compensate for the loss of TLR4, and compensation occurs at the expense of an increased microbial burden. The Journal of Immunology, 2004, 172: 6202–6208.

ulticellular organisms use pattern-recognition recep- mice to S. Typhimurium in vivo. Our findings indicate that TLR4 tors to detect and respond to pathogenic microorgan- signaling controls diverse functions contributing to host resistance M isms. Toll-like receptors (TLRs)3 are an important to Salmonella. Recognition of Gram-negative bacteria by TLR4 http://www.jimmunol.org/ family of pattern-recognition receptors expressed at cell surfaces, stimulates early expression of chemokines, recruitment, and acti- where they recognize and trigger cellular responses to a wide va- vation of phagocytic cells, and the eventual establishment of cell- riety of pathogen-associated molecular patterns (1). In association mediated adaptive immunity. In addition, we have demonstrated a with the LPS-binding protein and CD14 and MD-2 proteins, the critically important role for TLR4 in the activation of the Kupffer pattern-recognition receptor TLR4 specifically responds to LPS cell network to limit Salmonella replication in the liver before the derived from Gram-negative bacteria (2). Accordingly, C3H/HeJ recruitment of inflammatory cells. mice carrying a dominant-negative mutation in the cytoplasmic domain of TLR4 (3) are resistant to LPS-induced shock and death by guest on October 2, 2021 (4). However, despite their endotoxin resistance, TLR4-deficient Materials and Methods C3H/HeJ mice have increased mortality following infection with Measurement of bacterial virulence live Gram-negative bacteria such as Salmonella enterica serovar Typhimurium (S. Typhimurium) (5) or Escherichia coli (6), sug- Viability of wild-type C3H/HeN or congenic TLR4-deficient C3H/HeJ (The Jackson Laboratory, Bar Harbor, ME) mice was recorded daily after gesting that LPS/TLR4 signaling is important for the stimulation i.p. challenge with ϳ1000 CFUs of virulent S. Typhimurium strain 14028s of protective immune responses. (American Type Culture Collection, Manassas, VA) or an isogenic chlor- TLRs have been implicated in such diverse processes as che- amphenicol resistance-cassette insertion in invA invasion gene (invA::cm) mokine/cytokine production (7), phagocytic cell recruitment and derivative (11). Selected groups of mice were euthanized by CO2 inhalation, and the livers were aseptically removed and homogenized in sterile function (8, 9), and the generation of adaptive immunity (10). The PBS (12). Viable counts were determined after overnight culture on Luria- present study was undertaken to identify specific TLR4-dependent Bertani (LB) agar plates. functions responsible for the enhanced susceptibility of C3H/HeJ

Macrophage killing assays *Department of Microbiology, University of Colorado Health Sciences Center, Den- † Peritoneal exudate cells from wild-type C3H/HeN mice or congenic TLR4- ver, CO 80262; Biotechnology Laboratory, University of British Columbia, Van- deﬁcient C3H/HeJ mice were harvested 4 days after i.p. inoculation of 1 couver, British Columbia, Canada; and Departments of ‡Microbiology and ¤Labora- tory Medicine, University of Washington School of Medicine, Seattle, WA 98195 mg/ml sodium periodate, as described (13). The peritoneal exudate cells were resuspended in RPMI 1640 supplemented with 10% heat-inactivated Received for publication November 13, 2003. Accepted for publication March FCS (Gemini Bio-Products, Calabassas, CA), 1 mM sodium pyruvate, 10 3, 2004. mM HEPES, and 2 mM L-glutamine (all from Sigma-Aldrich, St. Louis, The costs of publication of this article were defrayed in part by the payment of page MO). The macrophages were selected by adherence to a 96-well plate and charges. This article must therefore be hereby marked advertisement in accordance cultured for 48 h at 37¡C in a 5% CO2 incubator. Selected groups of with 18 U.S.C. Section 1734 solely to indicate this fact. macrophages were incubated with 20 U/ml murine IFN-␥ (Life Technol- 1 Support for this work was provided by the National Institutes of Health (AI10181, ogies, Grand Island, NY) 24 h before infection. The phagocytes were chal- AI39557, RR16082, AI47242, AI54959) and the Schweppe Foundation (Chicago, IL). lenged with S. Typhimurium opsonized with 10% normal mouse serum at 2 Address correspondence and reprint requests to Dr. Ferric C. Fang, Department of a multiplicity of infection of 10:1, allowed to internalize the bacteria for 15 Microbiology, University of Washington School of Medicine, Box 357242, Seattle, min, and washed with prewarmed medium containing 6 ␮g/ml gentamicin. WA 98195-7242. E-mail address: [email protected] To estimate the contribution of NO to the anti-Salmonella activity of the ␮ 3 Abbreviations used in this paper: TLR, Toll-like receptor; S. Typhimurium, Salmo- macrophages, 250 M N-monomethyl L-arginine was added to the phago- nella enterica serovar Typhimurium; FliC, ﬂagellin; iNOS, inducible NO synthase; cytes at the time of the infection. At different time points after challenge, LB, Luria-Bertani; MPO, myeloperoxidase; invAϻcm, chloramphenicol resistance- the macrophages were lysed with 0.5% sodium deoxycholate and the sur- cassette insertion in invA invasion gene. viving bacteria enumerated on LB agar plates.

Measurement of inducible NO synthase (iNOS) expression secondary Abs, including Alexa Fluor 488 goat anti-rabbit (Alexis, San Diego, CA) and either Alexa Fluor 594 streptavidin or PE-Cy5 goat anti-rat Total RNA was isolated from sodium periodate-elicited macrophages using (Cedarlane Laboratories, Hornby, Ontario, Canada) Abs (all at 1/200) for TRIzol (Invitrogen, Carlsbad, CA). ssDNA was synthesized from purified 4 h. Coverslips were mounted in Mowiol mounting medium (Sigma-Al- RNA by oligo(dT) priming and Moloney murine leukemia virus reverse drich), and the tissues were viewed using a Zeiss (Oberkochen, Germany) transcriptase (Promega, Madison, WI). Expression of iNOS was monitored Axiovert s100 TV microscope attached to a Bio-Rad Radiance Plus con- Ј by PCR analysis of synthesized cDNA using primers 5 -CCTTC focal microscope with a ϫ63 oil objective. Images of 512 ϫ 512 pixels Ј CGAAGTTT CTGGCAGCAGC and 5 -GGCTGTCAGAGCCTCGTG (102 ϫ 102 ␮m) were acquired using Bio-Rad Lasersharp software (Her- GCTTTGG with TaqDNA polymerase. Amplification of GAPDH was used cules, CA). Sections of 0.2 ␮m thickness were assembled into flat projec- as an internal control. Production of reactive nitrogen species by Salmo- tions using NIH Image and imported into Adobe Photoshop (San Jose, nella-infected macrophages was determined by spectrophotometric detec- CA). Liver tissues from six infected mice were studied, and Ͼ20 fields of tion of nitrite at 550 nm using the Griess reagent (0.5% sulfanilimide and view were assessed per mouse. 0.05% N-1-naphthylethylenediamide hydrochloride in 2.5% acetic acid) (13). ELISPOT measurement of IFN-␥-producing T lymphocytes Measurement of chemokine expression The frequency of Salmonella- or FliC (flagellin)-specific IFN-␥-producing Total RNA was purified from 50Ð100 mg of liver tissue from Salmonella- T lymphocytes in immunized wild-type or TLR4-deficient mice was de- infected C3H/HeN and C3H/HeJ mice using the TRIzol reagent. The RNA termined by an ELISPOT assay essentially as previously described (15). was treated with DNase I and purified further with the Qiagen RNeasy kit Wild-type C3H/HeN or TLR4-deficient C3H/HeJ mice were injected with (Qiagen, Hilden, Germany). Single-stranded cDNA was synthesized from 5 ϫ 107 heat-killed FliC-expressing S. Typhimurium via an i.p. route. purified RNA, as above. Differential chemokine gene expression in the Seven days after infection, spleen cells were harvested for ELISPOT assay liver of C3H/HeN and C3H/HeJ mice was assessed at 6 and 24 h after to detect CD4ϩ T cells. Splenic CD4ϩ T cells were isolated by incubation Salmonella infection using the mouse chemokine GEArray kit, according on ice with FITC-conjugated anti-CD4 (BD PharMingen, San Diego, CA) Downloaded from to the instructions recommended by the manufacturer (Superarray, before selection with a magnetic cell sorter, and stimulated for 48 h at 37¡C Frederick, MD). with 5 ϫ 106 irradiated syngeneic splenocytes as APCs plus 5 ϫ 107/ml heat-killed S. Typhimurium, 100 ␮g/ml purified FliC (flagellar) protein, or Measurement of myeloperoxidase (MPO) activity no Ag. Cells were harvested with Histopaque (density 1.083; Sigma-Al- ␥ MPO was extracted from homogenized liver tissue using hexadecyltrim- drich), seeded to ELISPOT plates coated with anti-IFN- (BD PharMin- ␥ ethylammonium bromide (Sigma-Aldrich). Enzymatic activity was re- gen), incubated for 18 h, and assayed for IFN- secretion using a biotin- ␥ corded spectrophotometrically as the increase in A in a reaction con- ylated anti-IFN- secondary Ab and an avidin-HRP detection system 460 (Sigma-Aldrich). IFN-␥-producing spots were enumerated with a dissect- http://www.jimmunol.org/ taining 0.3% H2O2 and 0-dianisidine (Sigma-Aldrich). Units of MPO/g tissue/min were calculated using an extinction coefficient of ⑀ ϭ 13.5. ing microscope. Histopathology and immunofluorescence microscopy Results Hepatic tissue from Salmonella-infected wild-type C3H/HeN or TLR4- deficient C3H/HeJ mice was collected 0Ð24 h after i.p. infection, fixed in Susceptibility of TLR4-deficient mice to live Salmonella 10% formalin, stained in H&E, and examined by light microscopy for the TLR4-deficient C3H/HeJ mice succumbed within 4 days after i.p. presence of microabscesses and granulomas. For immunofluorescence studies (14), wild-type C3H/HeN or TLR4- infection with 1000Ð1500 CFU wild-type S. enterica strain ATCC deficient C3H/HeJ mice were i.p. infected with 1000 CFU of S. Typhi- 14028s (Fig. 1A), a sublethal challenge for the congenic wild-type by guest on October 2, 2021 murium 14028s and euthanized by CO2 inhalation 16 h after infection. C3H/HeN mice. The increased susceptibility of TLR4-deficient Livers were aseptically dissected, rinsed in ice-cold TBS, and flash frozen mice to salmonellosis was paralleled by the enhanced proliferation in OCT compound (Sakura Finetech, Torrance, CA), using a combination Ϫ of bacteria in the livers of these mice (Fig. 1B). During the first6h of isopentane and liquid N2, and stored at 70¡C. Serial sections were cut at a thickness of 15 ␮m and fixed in ice-cold acetone for 20 min. Tissues after infection, the growth rate of Salmonella was comparable in were blocked with 1% BSA in TBS with 0.1% Triton X-100. This was wild-type and TLR4-deficient mice. However, C3H/HeJ mice followed by addition of the primary Abs: rabbit anti-Salmonella (Difco, were found to harbor 100- to 1000-fold more bacteria in their Detroit, MI) at a 1/500 dilution, and biotinylated mouse anti-cytokeratin 18 livers than wild-type controls by 24 h after the onset of infection. (Progen Biotechnik, Heidelberg, Germany) and rat anti-F4/80 (Serotec, Raleigh, NC) at a 1/100 dilution. Incubation with primary Abs was per- These observations indicate that TLR4 is essential for the formed overnight (18 h) and followed by extensive washing with 0.1% development of an effective early innate immune response to Triton X-100 in TBS. The tissues were subsequently incubated with the Salmonella.

FIGURE 1. TLR4-deﬁcient mice have increased susceptibility to Salmo- nella infection. Wild-type C3H/HeN and TLR4-deﬁcient C3H/HeJ mice were inoculated i.p. with 1000Ð1500 CFU of S. typhimurium strain 14028s. Survival of mice (A) and bacterial burden in the liver (B) were monitored for 5 days after infection. The data were ,ء .obtained from six mice per group Denotes p Ͻ 0.05 by Student’s t test. 6204 TLR4 DEPENDENCE OF INNATE AND ADAPTIVE IMMUNITY

Antibacterial actions of peritoneal macrophages from TLR4- that TLR4 is not required for early NADPH phagocyte oxidase- deficient mice mediated Salmonella killing. In fact, elicited macrophages from C3H/ Salmonella virulence in mice closely corresponds to the ability of HeN and C3H/HeJ mice produced similar quantities of hydrogen per- the bacteria to survive in murine macrophages (16). Periodate- oxide and superoxide 1 h after infection with Salmonella as measured elicited peritoneal macrophages from TLR4-deficient mice were by HRP-dependent oxidation of phenol red and cytochrome c reduc- found to be as effective at killing Salmonella during the initial 6 h tion (data not shown). However, TLR4-deficient macrophages were after phagocytosis as were wild-type cells (Fig. 2A). This suggests substantially impaired in their ability to restrict Salmonella growth at Downloaded from http://www.jimmunol.org/ by guest on October 2, 2021

FIGURE 2. Peritoneal macrophages from TLR4-deficient mice exhibit impaired NO-dependent antimicrobial activity for Salmonella. Periodate-elicited peritoneal macrophages from wild-type C3H/HeN or TLR4-deficient C3H/HeJ mice were infected with S. Typhimurium or Staphylococcus aureus bacteria. Selected groups of macrophages were treated with 200 U/ml murine rIFN-␥ 24 h before infection. Surviving intracellular bacteria were enumerated on LB plates, and the results were expressed as percentage of survival (A), which correlates inversely with macrophage antimicrobial activity. Mean CFU Ϯ SEM Denotes p Ͻ 0.01 by Student’s t test. Nitrite accumulation in culture ,ء .from six independent experiments obtained on 2 separate days are shown Denotes p Ͻ 0.001 by ,ء .(supernatants of infected macrophages was measured spectrophotometrically using the Griess reagent 20 h after infection (B Student’s t test. iNOS and control GAPDH mRNA were visualized in an agarose gel after RT-PCR amplification (C). Selected groups of periodate-elicited G C3H/HeN or C3H/HeJ peritoneal macrophages were treated with 250 ␮M N -monomethyl-L-arginine (MMLA) to inhibit NO synthase at the time of G Salmonella infection. Controls received 250 ␮M N -monomethyl-D-arginine, which does not inhibit NO synthesis. Percentage of Salmonella survival was Denotes ,ء .determined at 6 or 20 h after infection (D). Mean CFU Ϯ SEM from three to six independent experiments obtained on 2 separate days are shown p Ͻ 0.001 by Student’s t test. The Journal of Immunology 6205

20 h after infection, an action previously shown to be dependent on sidered whether defective recruitment of inflammatory cells might iNOS (13). Stimulation of TLR4-deficient macrophages with IFN-␥ account for the early immune defect of TLR4-deficient animals 20 h before Salmonella ingestion was able to completely restore the observed despite preservation of early Salmonella killing mecha- antibacterial activity of these cells at 20 h (Fig. 2A). The antibac- nisms in peritoneal macrophages. terial defect of TLR4-deficient macrophages was specific for Binding of LPS by its TLR4 cognate receptor triggers a signal- Gram-negative bacteria such as Salmonella, because these macro- ing cascade that stimulates the expression of multiple CXC che- phages were able to kill Gram-positive Staphylococcus aureus as mokines, which leads to the recruitment of acute inflammatory efficiently as wild type (Fig. 2A). cells to sites of infection. At 6 h after Salmonella inoculation, The impaired anti-Salmonella activity of TLR4-deficient mac- when the bacterial burden was similar in TLR4-deficient and wild- rophages at 20 h correlated with a marked reduction in the pro- type animals (Fig. 1B), Gro1, IFN-␥-inducible protein 10, and duction of nitrite, an oxidation product of NO (Fig. 2B). IFN-␥ monokine induced by IFN-␥ chemokines were preferentially ex- treatment significantly increased the ability of both wild-type and pressed in wild-type C3H/HeN mice (Fig. 3A). However, robust TLR4-deficient macrophages to synthesize nitrite (Fig. 2B). Pro- levels of Gro1, IFN-␥-inducible protein 10, monokine induced by duction of nitrite by Salmonella-infected macrophages from C3H/ IFN-␥, and macrophage-inflammatory protein 2 chemokine HeJ mice paralleled the presence of iNOS mRNA (Fig. 2C). Col- mRNA were observed in the livers of TLR4-deficient animals by lectively, these observations demonstrate that the reduced anti- 24 h after infection. The chemokine response observed at the later Salmonella activity of TLR4-deficient macrophages results from a time point in C3H/HeJ mice is likely to reflect the induction of failure of iNOS transcription and iNOS-dependent NO production. TLR4-independent signaling pathways by the greater bacterial This is further substantiated by the ability of the NOS inhibitor burden present in the immunodeficient mice. Of interest, hepatic G Downloaded from N -monomethyl-L-arginine to reduce the anti-Salmonella activity expression of the chemokine macrophage-inflammatory protein 2 of wild-type macrophages at 20 h to the level of TLR4-deficient associated with neutrophil recruitment was detected at6hinlivers cells (Fig. 2D). of Salmonella-infected wild-type mice, but not of their TLR4-deficient counterparts. As neutrophils are the initial inflammatory Chemokine responses to Salmonella infection in cells recruited to the site of Salmonella infection (14) and are TLR4-deficient mice highly efficient at killing Salmonella (17), it was further considered The preservation of early bacterial killing by macrophages with whether defective recruitment of neutrophils might contribute to http://www.jimmunol.org/ defective late NO-dependent antimicrobial actions was somewhat the hypersusceptibility of TLR4-deficient mice to infection with unexpected given the impaired ability of TLR4-deficient mice to Gram-negative bacteria. limit bacterial replication during the initial 24 h of infection (Fig. 1A). Earlier work has suggested that iNOS only plays a significant Neutrophilic infiltration in TLR4-deficient mice role in host defense against Salmonella at significantly later time Histopathological examination of H&E-stained liver sections at points (12). A further indication that absent iNOS activation could 24 h after infection demonstrated the presence of small microab- not account for the defective innate immunity of C3H/HeJ mice scesses packed with neutrophils throughout the hepatic paren-

came from the measurement of iNOS mRNA in the livers of in- chyma of TLR4-deficient C3H/HeJ mice (Fig. 3B), in marked con- by guest on October 2, 2021 fected mice, which revealed almost no detectable iNOS expression trast to published studies of invasive E. coli infection in C3H/HeJ in either wild-type or TLR4-deficient animals at the 24-h time mice (6), in which no neutrophilic infiltration was observed. Pres- point (data not shown). Moreover, administration of the iNOS in- ence of microabscesses was not evident in the livers of wild-type 6 hibitor L-N -(1-iminoethyl)lysine failed to increase 24-h organism C3H/HeN mice until 3 days after infection (data not shown). In- burdens in C3H/HeN mice (data not shown). It was therefore con- filtration of neutrophils in the hepatic parenchyma of wild-type and

FIGURE 3. TLR4-deficient mice exhibit neutrophilic infiltration of the liver despite delayed CXC chemokine expression. TLR4-deficient C3H/HeJ (J) or wild-type C3H/HeN control (N) mice were infected with Salmonella i.p., as described in Fig. 1. The abundance of mRNA corresponding to selected chemokine genes was quantitated in liver extracts of Salmonella-infected mice using a DNA microarray (A). H&E sections were prepared from livers harvested at the indicated times after infection (B). The micrographs shown (original magnification ϫ400) are representative of data collected from two separate experiments. Recruitment of numerous neutrophils (arrow) into liver tissue was independently corroborated by measuring MPO activity at the ,ء .indicated times after infection (C). Mean values Ϯ SEM from six to nine independent experiments performed on 2 or more separate days are shown Denotes p Ͻ 0.005 by Student’s t test. 6206 TLR4 DEPENDENCE OF INNATE AND ADAPTIVE IMMUNITY

TLR4-deficient mice was independently assessed by monitoring burdens were present, a few bacteria could also be observed within MPO activity. In agreement with the histopathological evidence of hepatocytes (data not shown). a neutrophilic response in C3H/HeJ mice, abundant MPO activity was detected within the livers of TLR4-deficient mice (Fig. 3C) following perfusion. Thus, an initial delay in chemokine produc- Susceptibility of TLR4-deficient mice to invA mutant Salmonella tion is rapidly compensated for in TLR4-deficient mice, with a To distinguish the importance of pathogen-directed and phago- resulting exuberant recruitment of polymorphonuclear leukocytes cyte-directed mechanisms of bacterial internalization, infection to the site of Salmonella infection exceeding the magnitude of studies were repeated using noninvasive invA::cm and congenic response seen in wild-type animals. Despite the delayed CXC che- wild-type Salmonella in wild-type and TLR4-deficient mice. The mokine response, the hypersusceptibility of C3H/HeJ mice to inability to invade nonphagocytic cells did not impair the ability of acute Salmonella infection could not be attributed to a poor neu- S. Typhimurium to proliferate in the livers of TLR4-deficient C3H/ trophilic response. HeJ mice (Fig. 5), providing additional support for the microscopic Localization of Salmonella in hepatic parenchyma of evidence that hepatic bacteria were principally replicating within TLR4-deficient mice the Kupffer cell network. A modest trend toward reduced survival of invA::cm mutant bacteria relative to wild-type S. typhimurium Somewhat paradoxically then, TLR4-deficient mice are deficient was observed in TLR4-proficient C3H/HeN mice. This may indi- in early innate immunity to Salmonella despite a robust acute in- cate that invA-dependent cytotoxicity for macrophages (11) can flammatory cell response and preserved early antibacterial actions partially counter the TLR4-dependent host response. of inflammatory phagocytes. Studies were therefore conducted to Downloaded from determine the localization of Salmonella within the liver during the 6- to 24-h period of enhanced bacterial proliferation and before the Adaptive immune responses to Salmonella in recruitment of acute inflammatory cells. TLR4-deficient mice Salmonella were directly visualized by fluorescence and confocal microscopy in relation to F4/80-expressing Kupffer cells and An important function of TLRs, in addition to the control of innate cytokeratin 18-expressing hepatocytes 16 h after i.p. infection. immunity, is to provide a linkage to adaptive immune responses Confocal microscopy confirmed intracellular localization of the (18, 19). The principal mechanism by which TLRs are believed to http://www.jimmunol.org/ organisms. No bacteria could be visualized in the livers of wild- promote adaptive immune responses is through the induction of type mice at this early time point. However, numerous Salmonella costimulatory ligand and cytokine responses in APCs. The gener- were found to localize within Kupffer cells in the livers of TLR4- ation of adaptive cellular immunity to Salmonella Ags was there- deficient mice (Fig. 4), demonstrating an important role of LPS- fore compared in wild-type and TLR4-deficient mice. To negate TLR4 signaling pathways in triggering innate antibacterial actions the influence of impaired innate immunity on Ag load, heat-killed of Kupffer cells before the recruitment of inflammatory cells. Sal- Salmonella was used as an immunogen. monella organisms within Kupffer cells were clustered in discrete Seven days after i.p. administration of FliC (flagellar Ag)-expressing heat-killed Salmonella, a significantly greater number of loci throughout the liver, suggesting proliferation and localized by guest on October 2, 2021 ϩ spread within the Kupffer cell network. When higher organism IFN-␥-producing CD4 T cells responding to Salmonella or purified FliC was observed in the spleens of wild-type mice in com- parison with TLR4-deficient mice (Fig. 6), indicating that LPS- TLR4 signaling modulates the generation of adaptive immune responses to Salmonella.

FIGURE 4. Salmonella proliferates in Kupffer cells of TLR4-deficient mice before the recruitment of inflammatory cells. TLR4-deficient C3H/ HeJ or wild-type C3H/HeN control mice were infected with Salmonella i.p., as described in Fig. 1, and sacrificed 16 h after infection. A three-color merged view obtained by immunofluorescence confocal microscopy demonstrates Salmonella (green) within Kupffer cells (blue, labeled with ␣-F4/ 80) in the livers of TLR4-deficient mice (A). Hepatocytes were labeled with ␣-cytokeratin 18 (red). To better visualize the Kupffer cell network, the FIGURE 5. TLR4-deficient mice have increased susceptibility to infec- image was cropped, the hepatocyte signal was removed, and the blue F4/80 tion with invA mutant Salmonella. Wild-type C3H/HeN and TLR4-defi- signal (Kupffer cells) was changed to red (B). A Z section along the plane cient C3H/HeJ mice were inoculated i.p. with 1000Ð1500 CFU of wild- indicated by the dashed line confirms the presence of the Salmonella inside type or noninvasive invA::cm S. typhimurium. Bacterial burden in the liver the Kupffer cells. Each figure was originally obtained at a magnification of was determined 24 h after infection. The data were obtained from six mice Denotes p Ͻ 0.05 by Student’s t test. For HeN WT vs HeN ,ء .ϫ630. No Salmonella could be visualized within the livers of wild-type per group C3H/HeN at the same time point. invA, p ϭ 0.08. The Journal of Immunology 6207

sues in addition to promoting the recruitment of inflammatory cells. Both Kupffer cells and hepatocytes have been shown to ex- press TLR4 (23, 24), and TLR4-deficient Kupffer cells have been found to be deficient in their ability to kill ingested E. coli (6), but the mechanisms of TLR4-dependent antimicrobial actions in Kupffer cells and hepatocytes remain to be established. In TLR4-deficient mice infected with S. Typhimurium 14028s, the critical inadequacy appears to be in the timely generation of immune responses. Although compensatory pathways ultimately stimulate chemokine production, inflammatory cell recruitment, or phagocyte activation, in each case the response occurs later than would have been the case in the presence of TLR4, with the ultimately lethal consequence of a higher microbial burden. TLR4- FIGURE 6. Adaptive cellular immune responses to Salmonella are at- deficient mice exhibit a potent, if somewhat belated chemokine tenuated in TLR4-deficient mice. Reduced numbers of CD4ϩ T cells re- response (Fig. 3A), possibly reflecting compensation by other sponding to heat-killed Salmonella or purified flagellar (FliC) Ag were TLRs, but the ensuing influx of inflammatory cells (Fig. 3B) was observed 7 days after immunization with heat-killed Salmonella in TLR4- ineffective in controlling the infection. Although the lack of TLR4- ء deficient C3H/HeJ mice as compared with wild-type C3H/HeN controls. , dependent iNOS stimulation (25, 26) can be compensated by the Denotes p Ͻ 0.001 by Student’s t test. addition of IFN-␥ in vitro (Fig. 2A), this necessitates a costly delay Downloaded from in macrophage activation until the arrival of cytokine-producing Discussion cells in vivo. Similarly, TLR4-deficient mice can generate protec- Several of the 11 known mammalian TLRs have been demon- tive adaptive immune responses to vaccination (27), but the kinet- strated to recognize various bacterial components, including li- ics of the adaptive responses are protracted (Fig. 6). ␥ popeptides (TLR2), LPS (TLR4), flagellin (TLR5), and unmeth- Compensation for TLR4 deficiency by IFN- can be rational- ylated CpG DNA (TLR9) (1). Because typical bacteria produce ized by the importance of both NF ␬B and transcription factor http://www.jimmunol.org/ more than one of these ligands, some functional redundancy STAT1 activation for iNOS expression. TLR4 agonists by them- among individual TLRs might be anticipated. However, point mu- selves are believed to be sufficient for iNOS stimulation because tations in human TLR4 resulting in reduced function are associated TLR4-dependent signaling leads to both NF-␬B activation and en- with an increased risk of Gram-negative infections (20), and dogenous production of IFN-␤, leading to STAT1 phosphorylation TLR4-deficient mice have enhanced susceptibility to Gram-nega- (28) and iNOS transcription via IFN regulatory factor-1 (29, 30). tive bacteria (5, 6), indicating that the presence of other TLRs TLR2 stimulation, in contrast, does not result in iNOS stimulation cannot necessarily compensate for the loss of another. Similarly, unless IFN-␤ or -␥ to activate STAT1 (28, 31) is provided the mouse strains used in this study both possess a functional exogenously. by guest on October 2, 2021 Nramp1/S1c11a1 locus (21), but this important determinant of host Although C3H/HeJ mice are unable to respond to LPS via resistance to Salmonella is not sufficient to compensate for the loss TLR4, their ability to recognize bacterial ligands via other TLRs of TLR4 in the C3H/HeJ mice. such as TLR2, TLR5, and TLR9 remains intact. Mice infected with The present study analyzed TLR4-dependent elements of innate Salmonella begin to produce IFN-␥ ϳ3Ð5 days after the onset of and adaptive immunity in the well-characterized Salmonella mu- infection (32), and TLR4-deficient mice retain the ability to pro- rine infection model. A single Pro712His mutation in TLR4 (3) has duce IFN-␥ in response to Salmonella (33). Sources of IFN-␥ in- profound and pleiotropic effects on the resistance of C3H/HeJ mice clude NK cells (34), which can be activated by stimulation of to Salmonella infection. TLR4 plays an important role in both receptors such as TLR2 (35). Although the production of IFN-␥ early and late immune responses in vivo, ranging from chemokine can eventually overcome the absence of LPS-TLR4 signaling for production to phagocyte activation and the elicitation of adaptive the stimulation of iNOS expression, the delayed synthesis of NO is immunity. Most significantly, our studies demonstrate that LPS- likely to be detrimental to the infected host. TLR4 interactions are required for the initial limitation of bacterial Although the various TLRs can provide a host with versatility replication within host tissues before the arrival of inflammatory through the recognition of a wide range of pathogenic microor- cells. This innate antimicrobial function appears to be principally ganisms, each TLR may require a different level of organism burden mediated by tissue histiocytes (e.g., Kupffer cells of the liver), with for a given pathogen to reach the critical threshold for activation. The a possible secondary contribution from parenchymal cells (e.g., subtle or absent phenotypes of mice lacking single TLRs in some hepatocytes). This contrasts markedly with earlier observations in immunocompetent mice, in which hepatic Salmonella were found infection models (36) have led to the suggestion that these receptors within infiltrating phagocytic cells (14) rather than resident cells, and may be functionally redundant. However, the ability of TLR4 to sense only at later time points. Enhanced replication of Salmonella in the minute quantities of LPS provides the host with a highly sensitive and livers of C3H/HeJ animals was not abrogated by an invA mutation essential system to detect the presence of Salmonella or other Gram- abolishing Salmonella invasiveness for nonphagocytic cells, further negative bacterial pathogens that cannot be fully supplanted by the substantiating that bacteria in the liver are principally replicating recognition of alternative ligands. within the Kupffer cell network of TLR4-deficient mice. The role of LPS-TLR4 signaling in the development of lethal Along with a recent study using an experimental urinary tract endotoxic shock (37) has led to the suggestion that TLR blockade infection model (22), the present observations reinforce that resi- may be useful in the treatment of sepsis (38). However, the mul- dent cells in peripheral tissues contribute more to TLR-dependent titude of essential immune functions dependent on TLRs as dem- host resistance than is generally appreciated. However, in contrast onstrated in this study suggests that targeting TLRs in the setting to the urinary tract model, the present study demonstrates that of acute bacterial infection will have to be undertaken with great TLR4 can regulate direct antimicrobial actions in peripheral tis- caution. 6208 TLR4 DEPENDENCE OF INNATE AND ADAPTIVE IMMUNITY

Acknowledgments 19. Rescigno, M., M. Urbano, M. Rimoldi, B. Valzasina, G. Rotta, F. Granucci, and P. Ricciardi-Castagnoli. 2002. Toll-like receptor 4 is not required for the full We are grateful to Leigh Knodler and Danika Goosney for assistance with maturation of dendritic cells or for the degradation of Gram-negative bacteria. confocal microscopy, to Casey Calkins for assistance with MPO assays, Eur. J. Immunol. 32:2800. and to Yisheng Xu for assistance with the measurement of chemokine gene 20. Smirnova, I., N. Mann, A. Dols, H. H. Derkx, M. L. Hibberd, M. Levin, and expression. B. Beutler. 2003. Assay of locus-speciﬁc genetic load implicates rare Toll-like receptor 4 mutations in meningococcal susceptibility. Proc. Natl. Acad. Sci. USA 100:6075. References 21. Forbes, J. R., and P. Gros. 2001. Divalent-metal transport by NRAMP proteins at 1. Takeda, K., and S. Akira. 2003. Toll receptors and pathogen resistance. Cell. the interface of host-pathogen interactions. Trends Microbiol. 9:397. Microbiol. 5:143. 22. Schilling, J. D., S. M. Martin, C. S. Hung, R. G. Lorenz, and S. J. Hultgren. 2003. 2. Beutler, B., X. Du, and A. Poltorak. 2001. Identiﬁcation of Toll-like receptor 4 Toll-like receptor 4 on stromal and hematopoietic cells mediates innate resistance (Tlr4) as the sole conduit for LPS signal transduction: genetic and evolutionary to uropathogenic Escherichia coli. Proc. Natl. Acad. Sci. USA 100:4203. studies. J. Endotoxin Res. 7:277. 23. Dory, D., H. Echchannaoui, M. Letiembre, F. Ferracin, J. Pieters, Y. Adachi, 3. Poltorak, A., X. He, I. Smirnova, M. Y. Liu, C. Van Huffel, X. Du, D. Birdwell, S. Akashi, W. Zimmerli, and R. Landmann. 2003. Generation and functional E. Alejos, M. Silva, C. Galanos, et al. 1998. Defective LPS signaling in C3H/HeJ characterization of a clonal murine periportal Kupffer cell line from H-2K(b) and C57BL/10ScCr mice: mutations in Tlr4 gene. Science 282:2085. -tsA58 mice. J. Leukocyte Biol. 74:49. 4. Sultzer, B. M. 1968. Genetic control of leukocyte responses to endotoxin. Nature 24. Liu, S., D. J. Gallo, A. M. Green, D. L. Williams, X. Gong, R. A. Shapiro, 219:1253. A. A. Gambotto, E. L. Humphris, Y. Vodovotz, and T. R. Billiar. 2002. Role of 5. O’Brien, A. D., D. L. Rosenstreich, I. Scher, G. H. Campbell, R. P. MacDermott, Toll-like receptors in changes in gene expression and NF-␬B activation in mouse and S. B. Formal. 1980. Genetic control of susceptibility to Salmonella typhi- hepatocytes stimulated with lipopolysaccharide. Infect. Immun. 70:3433. murium in mice: role of the LPS gene. J. Immunol. 124:20. 25. Royle, M. C., S. Totemeyer, L. C. Alldridge, D. J. Maskell, and C. E. Bryant. 6. Cross, A., L. Asher, M. Seguin, L. Yuan, N. Kelly, C. Hammack, J. Sadoff, and 2003. Stimulation of Toll-like receptor 4 by lipopolysaccharide during cellular P. J. Gemski. 1995. The importance of a lipopolysaccharide-initiated, cytokine- invasion by live Salmonella typhimurium is a critical but not exclusive event mediated host defense mechanism in mice against extraintestinally invasive Esch- leading to macrophage responses. J. Immunol. 170:5445. 26. Ding, A., S. Hwang, H. M. Lander, and Q. W. Xie. 1995. Macrophages derived

erichia coli. J. Clin. Invest. 96:676. Downloaded from 7. Netea, M. G., C. A. Van Der Graaf, A. G. Vonk, I. Verschueren, from C3H/HeJ (Lpsd) mice respond to bacterial lipopolysaccharide by activating J. W. Van Der Meer, and B. J. Kullberg. 2002. The role of Toll-like receptor NF-␬B. J. Leukocyte Biol. 57:174 (TLR) 2 and TLR4 in the host defense against disseminated candidiasis. J. Infect. 27. McSorley, S. J., B. T. Cookson, and M. K. Jenkins. 2000. Characterization of ϩ Dis. 185:1483. CD4 T cell responses during natural infection with Salmonella typhimurium. 8. Weinstein, D. L., C. R. Lissner, R. N. Swanson, and A. D. O’Brien. 1986. Mac- J. Immunol. 164:986. rophage defect and inflammatory cell recruitment dysfunction in Salmonella sus- 28. Schilling, D., K. Thomas, K. Nixdorff, S. N. Vogel, and M. J. Fenton. 2002. ceptible C3H/HeJ mice. Cell. Immunol. 102:68. Toll-like receptor 4 and Toll-IL-1 receptor domain-containing adapter protein 9. Mattsby-Baltzer, I., B. Ahlstrom, L. Edebo, and P. de Man. 1996. Susceptibility (TIRAP)/myeloid differentiation protein 88 adapter-like (Mal) contribute to max- of lipopolysaccharide-responsive and -hyporesponsive ItyS mice to infection with imal IL-6 expression in macrophages. J. Immunol. 169:5874. http://www.jimmunol.org/ rough mutants of Salmonella typhimurium. Infect. Immun. 64:1321. 29. Kamijo, R., H. Harada, T. Matsuyama, M. Bosland, J. Gerecitano, D. Shapiro, 10. Medzhitov, R. 2001. Toll-like receptors and innate immunity. Nat. Rev. Immunol. J. Le, S. I. Koh, T. Kimura, S. J. Green, et al. 1994. Requirement for transcription 1:135. factor IRF-1 in NO synthase induction in macrophages. Science 263:1612. 11. Chen, L. M., K. Kaniga, and J. E. Gala«n. 1996. Salmonella spp. are cytotoxic for 30. Martin, E., C. Nathan, and Q. W. Xie. 1994. Role of interferon regulatory factor cultured macrophages. Mol. Microbiol. 21:1101. 1 in induction of nitric oxide synthase. J. Exp. Med. 180:977. 12. Mastroeni, P., A. Vazquez-Torres, F. C. Fang, Y. Xu, S. Khan, C. E. Hormaeche, 31. Decker, T., S. Stockinger, M. Karaghiosoff, M. Muller, and P. Kovarik. 2002. and G. Dougan. 2000. Antimicrobial actions of the NADPH phagocyte oxidase IFNs and STATs in innate immunity to microorganisms. J. Clin. Invest. and inducible nitric oxide synthase in experimental salmonellosis. II. Effects on 109:1271. microbial proliferation and host survival in vivo. J. Exp. Med. 192:237. 32. Pie, S., P. Matsiota-Bernard, P. Truffa-Bachi, and C. Nauciel. 1996. ␥ Interferon 13. Vazquez-Torres, A., J. Jones-Carson, P. Mastroeni, H. Ischiropoulos, and and interleukin-10 gene expression in innately susceptible and resistant mice F. C. Fang. 2000. Antimicrobial actions of the NADPH phagocyte oxidase and during the early phase of Salmonella typhimurium infection. Infect. Immun. inducible nitric oxide synthase in experimental salmonellosis. I. Effects on mi- 64:849. by guest on October 2, 2021 crobial killing by activated peritoneal macrophages in vitro. J. Exp. Med. 33. Freudenberg, M. A., Y. Kumazawa, S. Meding, J. Langhorne, and C. Galanos. 192:227. 1991. ␥ Interferon production in endotoxin-responder and -nonresponder mice 14. Richter-Dahlfors, A., A. M. Buchan, and B. B. Finlay. 1997. Murine salmonel- during infection. Infect. Immun. 59:3484. losis studied by confocal microscopy: Salmonella typhimurium resides intracel- 34. Kirby, A. C., U. Yrlid, and M. J. Wick. 2002. The innate immune response differs lularly inside macrophages and exerts a cytotoxic effect on phagocytes in vivo. in primary and secondary Salmonella infection. J. Immunol. 169:4450. J. Exp. Med. 186:569. 35. Becker, I., N. Salaiza, M. Aguirre, J. Delgado, N. Carrillo-Carrasco, L. G. Kobeh, 15. Cookson, B. T., and M. J. Bevan. 1997. Identification of a natural T cell epitope A. Ruiz, R. Cervantes, A. P. Torres, N. Cabrera, et al. 2003. Leishmania li- presented by Salmonella-infected macrophages and recognized by T cells from pophosphoglycan (LPG) activates NK cells through Toll-like receptor-2. Mol. orally immunized mice. J. Immunol. 158:4310. Biochem. Parasitol. 130:65. 16. Fields, P. I., R. V. Swanson, C. G. Haidaris, and F. Heffron. 1986. Mutants of 36. Reiling, N., C. Holscher, A. Fehrenbach, S. Kroger, C. J. Kirschning, S. Goyert, Salmonella typhimurium that cannot survive within the macrophage are avirulent. and S. Ehlers. 2002. Toll-like receptor (TLR)2- and TLR4-mediated pathogen Proc. Natl. Acad. Sci. USA 83:5189. recognition in resistance to airborne infection with Mycobacterium tuberculosis. 17. Shiloh, M. U., J. Ruan, and C. Nathan. 1997. Evaluation of bacterial survival and J. Immunol. 169:3480. phagocyte function with a fluorescence-based microplate assay. Infect. Immun. 37. Heumann, D., and T. Roger. 2002. Initial responses to endotoxins and Gram- 65:3193. negative bacteria. Clin. Chim. Acta 323:59. 18. Janeway, C. A., Jr., and R. Medzhitov. 2002. Innate immune recognition. Annu. 38. Lakhani, S. A., and C. W. Bogue. 2003. Toll-like receptor signaling in sepsis. Rev. Immunol. 20:197. Curr. Opin. Pediatr. 15:278.