| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
* Institut National de la Santé et de la Recherche Médicale U773, Centre de Recherche Biomédicale Bichat-Beaujon (CRB3), Paris France; Université Paris 7, Denis Diderot, Paris, France;
Service d’Anatomie et Cytologie Pathologiques, Centre Hospitalier Universitaire de Poitiers, Poitiers, France;
Unité de Réponses Précoces aux Parasites et Immunopathologie, Institut National de la Recherche Agronomique, Paris, France; and
Unité de Pathogénie Bactérienne des Muqueuses, Institut Pasteur, Paris, France
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
still remained greater in UPEC-infected than in naive C3H/HeJ (Lpsd) mice. Using primary cultures of microdissected Lpsn MCDs that expressed TLR4 and its accessory molecules MD2, MyD88, and CD14, we also show that UPECs stimulated both a TLR4-mediated, MyD88-dependent, TIR domain-containing adaptor-inducing IFN-
-independent pathway and a TLR4-independent pathway, leading to bipolarized secretion of MIP-2. Stimulation by UPECs of the TLR4-mediated pathway in Lpsn MCDs leads to the activation of NF-
B, and MAPK p38, ERK1/2, and JNK. In addition, UPECs stimulated TLR4-independent signaling by activating a TNF receptor-associated factor 2-apoptosis signal-regulatory kinase 1-JNK pathway. These findings demonstrate that epithelial collecting duct cells are actively involved in the initiation of an immune response via several distinct signaling pathways and suggest that intercalated cells play an active role in the recognition of UPECs colonizing the kidneys. |
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
Recognition of UPEC isolates by the mucosal cells lining the urinary tract elicits potent inflammatory responses. This process involves TLR4, a pattern-recognition receptor that recognizes LPS, the major cell wall constituent of all Gram-negative bacteria (9). LPS recognition by TLR4 results in the recruitment of multiple cytoplasmic signaling molecules, including MyD88 and the TNFR-associated factor (TRAF) 6, which, in turn, recruit and assemble additional molecules to activate downstream signaling components involving the transcription factors NF-B, MAPK p38, ERK1/2, and JNK, leading to the production of proinflammatory cytokines and chemokines (10, 11). C3H/HeJ mice (Lpsd), which have a loss-of-function mutation in the tlr4 gene (12), are unresponsive to LPS and fail to clear Gram-negative bacteria colonizing the lower urinary tract and kidneys (13).
Type 1 piliated E. coli has been shown to enhance the inflammatory response of the bladder epithelium by mediating bacterial invasion via a LPS recognition, TLR4-mediated pathway (14). Fisher et al. (15) also showed that the binding of type 1 or P fimbriae to their respective uroepithelial cell receptors activates a TLR4-mediated mucosal response, as assessed by analyses of neutrophil recruitment to the urinary tract of infected mice, through the requirement of distinct adaptor proteins. These findings suggested that the recognition of fimbrial adhesins by cell surface mucosal receptors may directly activate, independently of LPS, the TLR4 signaling pathway.
The mechanisms of interaction between UPECs and bladder epithelial cells have been extensively investigated (14, 16, 17). UPECs can also interact with renal tubule cells. UPECs may form microcolonies at the surface of renal proximal tubule cells from the kidneys of young rats and provoke Ca2+ oscillations induced by the secreted toxin -hemolysin, which, in addition, stimulates the production of IL-6 and IL-8 in cultured human renal epithelial A498 cells (18). However, the exact contribution of renal epithelial cells, as primary or additional sources of chemokine and cytokine production, to increasing the immune response still remains to be clarified. Recent studies using the hemopoietic chimeric Lpsn and Lpsd mice demonstrated that bladder epithelial cells (19), as well as intrinsic renal epithelial cells (20), contribute together with bone marrow-derived cells to the initiation of antibacterial immunity. Murine and human renal tubule cells express some of the TLR expressed in hemopoietic cells and secrete chemokines upon LPS stimulation (21, 22). In addition, TLR2 and TLR4 have been shown to be up-regulated in the distal nephron of inflamed, postischemic, reperfused kidneys (23). These data indirectly suggest that the most distal renal tubule segment, i.e., the medullary collecting duct (MCD), which is the first to come into contact with ascending UPEC strains, might be a site of bacterial adherence and, thus, is involved in triggering early inflammatory response.
In this study we analyzed the role of tubule epithelial cells during kidney colonization by two pyelonephritic UPEC strains lacking -hemolysin, and we characterized the resulting innate immune response. The role of TLR4 in the initiation of immune response was evaluated in whole kidneys and confluent cultures of MCDs microdissected from adult LPS-sensitive C3H/HeOuJ (Lpsn) and LPS-nonresponsive C3H/HeJ (Lpsd) mice. Experiments were also conducted with MyD88-deficient (MyD88–/–) mice (24), which exhibit no inflammatory response to the TLR4 ligand LPS (25), and Lps2 mutant (Lps2–/–) mice (26), which are deficient for the TIR domain-containing adaptor inducing IFN- (TRIF), an adaptor protein that regulates an MyD88-independent signaling pathway upon stimulation by TLR3 and TLR4 ligands (26, 27). We show that UPEC isolates colonized kidneys after transurethral inoculation and preferentially and constantly adhered to apical membranes of MCD intercalated cells expressing TLR4. UPECs induced the expression of proinflammatory mediators in MCD cells and stimulated bipolarized secretion of the chemokine MIP-2 (CXCL2/MIP-2) via both TLR4-mediated MyD88-dependent and TLR4-independent pathways requiring distinct MAPK-related signaling molecules. These findings indicate that MCD cells are involved in UPEC recognition and actively participate to the initiation of immune innate response by triggering rapid stimulation of proinflammatory mediators in injured kidneys.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
Two uropathogenic strains of E. coli (AL511 and AL10) isolated from the urine of women presenting pyelonephritis were used. They belong to two different O serogroups (O9 for AL511, and O101 for AL10), carry pap and afa-8 adhesin-encoding genes, and lack the hly -hemolysin-encoding gene (28). In addition, the AL511 isolate carries f17Ac adhesin genes and produces factors contributing to serum resistance. These UPEC isolates were selected on the basis of their in vitro interactions with and ability to stimulate cytokine/chemokine mRNA expression in immortalized mouse collecting duct mpkCCDcl4 cells (not shown). The nonpathogenic E. coli strain MG1655 was used as the control (29). The strains were grown in static Luria-Bertani broth at 37°C for 24 h.
Animals and infections
Eight- to 10-wk-old female C3H/HeOuJ (Lpsn), C3H/HeJ (Lpsd), and wild-type C57BL/6J mice were obtained from The Jackson Laboratory. TLR2 (tlr2–/–) and MyD88 (MyD88–/–) knockout mice (24, 30), originally obtained from S. Akira (Osaka University, Osaka, Japan), have been further backcrossed eight times into C57BL/6 mice to ensure a similar genetic background. C57BL/6 Lps2–/– mice (26) were provided by D. J. Philpott (Institut Pasteur, Paris, France). Mice submitted to sanitary control tests to ensure proper pathogen-free status were housed in the same animal facility at the Institut Pasteur for 1 wk before the experiments began. Lpsn and Lpsd mice, subjected to water restriction for 12 h, were anesthetized and then infected with 50 µl of bacterial suspension (108 bacteria) in sterile PBS introduced via the transurethral route into the bladder. A soft polyethylene catheter (Insyte Autoguard soft catheter, 0.7-mm external diameter; Vygon) was used as described (5, 8). Two days after bacterial inoculation the mice were sacrificed, and the kidneys and bladder were removed aseptically. The two halves of one kidney were fixed or quick frozen in liquid nitrogen. The contralateral kidney was homogenized, diluted in sterile PBS, and plated on Luria-Bertani agar plates to enumerate the number of CFUs. For in vivo TNF- neutralization experiments, Lpsn and Lpsd mice were i.p. injected 1 h before bacterial inoculation with 1 ml of PBS containing 1 mg of purified anti-mouse TNF- from the MP6-XT3 mAb or with 1 mg of the rat IgG1 isotype control (eBioscience) (31). All experiments were performed in accordance with the guidelines of the French Agricultural Office and in compliance with the legislation governing animal studies.
Microdissection and culture of MCDs
Kidneys from naive Lpsn and Lpsd mice or wild-type C57BL/6 (tlr2+/+), tlr2–/–, or MyD88–/– mice were rapidly removed under sterile conditions and incubated in modified defined medium supplemented with 0.1% collagenase (Roche Diagnostics) for 45 min at 37°C. MCDs were then microdissected under sterile conditions as described (32, 33). Pools of isolated MCDs (8–12 fragments) were seeded onto collagen-coated 48-well trays or onto Transwell filters (0.4- or 3-µm pore size, 0.33-cm2 insert growth area; Corning Costar). Isolated MCDs were grown to confluence for 15 days at 37°C in a 5% CO2-95% air atmosphere in the same modified defined medium (33). E. coli strains AL10 or AL511 (5 x 105 bacteria/filter) were added to the apical medium bathing confluent cell layers. Transepithelial resistance and potential were measured as described (33).
Histological and immunohistochemical studies
Kidneys were fixed in Dubosc-Brazil solution, rinsed in PBS, embedded in paraffin, and stained with H&E or periodic acid-Schiff. Immunohistochemical studies using Abs raised against E. coli (1/200; Interchim), murine TLR4 (1/200; provided by M. W. Hornef, University of Freiburg, Freiburg, Germany) (34), aquaporin-2 (AQP-2), and the chloride channel 5 (ClC-5) (1/200) (35) were performed using avidin-biotin blocking kit and alkaline phosphatase and peroxidase substrate kits (Vector Laboratories) according to the manufacturer’s instructions. Alkaline phosphatase and peroxidase activities were revealed with diaminobenzidine (brown reaction), nitroblue tetrazolium (purple reaction) (Lab Vision), or Vector Novared (red reaction) (Vector Laboratories) substrate solutions. Indirect immunofluorescence studies were also conducted on cultured MCDs using Abs against NF-B (Santa Cruz Biotechnology), E. coli, cytokeratins K8–K18, and ClC-5 and species-specific Alexa 488- and Cy3-conjugated IgG as secondary Abs (Jackson ImmunoResearch Laboratories). Cells were stained with phalloidin to visualize F-actin. Specimens were examined using a confocal laser-scanning microscope (CLSM-510-META; Zeiss) and photographed.
Scanning microscopy
E. coli strains (5 x 105 bacteria/filter) were added to the apical side of confluent cultures of MCDs grown on permeable filters for 3 h at 37°C. Cell layers were then rinsed in PBS, fixed in 4% glutaraldehyde in 0.1 M phosphate buffer for 1 h at 4°C, dehydrated in ascending concentrations of acetone, critical point dried, coated with gold, and then examined under a JEOL-JSM840A scanning electron microscope.
Real-time and RT-PCR
Total RNA was extracted from whole kidneys or cultured MCDs using the RNeasy mini kit (Qiagen), and reverse transcribed using Moloney murine leukemia virus reverse transcriptase (Invitrogen Life Technologies). cDNA was subjected to real-time PCR by using a ABI 7000 sequence detector (Applied Biosystems) (36). The mouse primers (GenBank accession nos. in parentheses) and TaqMan probes used were as follows: IL-1 (NM_008361), nt 527–551, nt 678–658, and probe, nt 629–656; IL-6 (NM_031168), nt 526–546, nt 608–587, and probe, nt 555–583; MIP-2 (NM_009140), nt 181–204, nt 262–237, and probe, nt 210–228; MCP-1 (NM_011333), nt 107–125, nt 233–213, and probe, nt 160–180; RANTES (NM_013653), nt 224–245, nt 294–268, and probe, nt 247–266; inducible NO synthase (iNOS) (NM_010927), nt 2176–2196, nt 2270–2250, and probe, nt 2198–2220; TNF- (NM_013693), nt 401–425, nt 575–553, and probe, nt 436–461; and -actin (NM_007393), nt 694–713, nt 831–811, and probe, nt 764–786. PCR data were reported as the relative increase in mRNA transcripts vs that found in kidneys of naive mice or untreated cultured MCDs and corrected by the respective levels of -actin mRNA used as the internal standard (36). For RT-PCR, cDNA and non-reverse transcribed RNA (400 ng) from cultured Lpsn MCDs were amplified for 30–35 cycles in 40 µl of total PCR buffer (50 mM KCl and 20 mM Tris-HCl (pH 8.4)) containing 100 µM dNTP, 1 or 1.5 mM MgCl2, 1 U of Taq polymerase, 10 pmol of TLR4 (BC029856, nt 1999–2020 and nt 2309–2288), 10 pmol of MD-2 (NM_016923, nt 12–34, and nt 327–306), 5 pmol of MyD88, 10 pmol of CD14 (NM_009841, nt 223–243 and nt 926–902), 5 pmol of TLR2 (NM_011905, nt 2335–2354 and nt 2816–2835), 30 pmol of CD45 (NM_011210, nt 3127–3146 and nt 3531–3551), 30 pmol of the -subunit of the epithelial sodium channel (ENaC), and 35 pmol of the cystic fibrosis transmembrane conductance regulator (CFTR) primers. The primers used for MyD88, -ENaC, and CFTR were the same as those previously described (33, 37). The thermal cycling program was 94°C for 30 s, 61°C (TLR2), 60°C (TLR4 and CFTR), 55°C (MD-2, CD14, and CD45), 53°C (MyD88), or 54°C (-ENaC) for 30 s and then 72°C for 1 min. Amplification products were run on a 2% agarose gel and then stained with ethidium bromide and autoradiographed.
Immunoblot analysis
For the detection of TLR4, the kidneys from a naive Lpsn mouse were removed and frozen in liquid nitrogen. An aliquot of homogenized proteins (50 µg of total protein) in Laemmli buffer was then processed for Western blotting using an anti-TLR4 Ab (34). Confluent MCD cells grown in 48-well plates precoated with rat tail collagen were lysed in 50 µl of 62.5 mM Tris-HCl (pH 6.8), 2% SDS, 10% glycerol, and 50 mM DTT and sonicated for 15 s at 4°C. For each condition tested, cell lysates from six separate wells were pooled. Samples (15 µg/lane) were electrophoresed using 10% SDS-PAGE and transferred to a nitrocellulose membrane. Abs against IkB-, p38, TRAF2, MAPK-activated protein kinase 2 (MAPKAPK-2), c-Jun (Cell Signaling Technology), TRIF (Imgenex), JNK, TRAF6, apoptosis signal regulatory kinase 1 (ASK1), ERK1/2 (Santa Cruz Biotechnology), and -actin (Sigma-Aldrich) were used to detect the corresponding Ags. Protein phosphorylation was analyzed using Abs against phosphorylated p38, JNK, ERK1/2, MAPKAPK-2, c-Jun, and ASK1 (Cell Signaling Technology). Protein bands were revealed using peroxidase-conjugated goat anti-rabbit or anti-mouse IgG (Jackson ImmunoResearch Laboratories) and detected using the ECL Plus Western blotting detection system (Amersham Biosciences).
ELISA
Cells were incubated for 3 h with UPECs alone and with anti-mouse TNF- mAb or rat IgG1 isotype control (eBioscience) or with or without specific cell-permeable MAPK inhibitors (Calbiochem). MIP-2 and TNF- secreted in cell supernatants were then determined using ELISA kits (R&D Systems) according to the manufacturer’s instructions. For in vivo TNF- measurements, kidneys from Lpsn and Lpsd mice infected with AL511 isolates were homogenized in 1 ml of PBS and kept at –80°C until use. Tissue samples were then thawed and assayed to measure the levels of TNF- production. Results were standardized to the amount of protein detected for each sample using the Bio-Rad protein assay with BSA as standard.
Statistical analysis
Results are expressed as means ± SEM. The significance of the differences was analyzed using Student’s t test (p < 0.05 was considered to be significant).
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
The UPEC strains AL10 and AL511 and the nonpathogenic strain MG1655 used as control were inoculated into the bladder of healthy Lpsn mice via the transurethral route. Bacteria in bladder and kidney tissues were counted 48 h postinfection. The AL10 and AL511 isolates were found to colonize the bladder and kidneys more effectively (p < 0.05) than MG1655 (Fig. 1A). All of the kidneys from mice inoculated with the two UPEC isolates were colonized with bacteria, whereas no (50% of inoculated mice) or only a few bacteria were detected in the kidneys following the transurethral inoculation of the nonpathogenic E. coli MG1655. Differences between the nonpathogenic strain and the UPEC isolates were also observed when kidney sections were examined by immunohistochemistry using an anti-E. coli polyclonal Ab. No positive E. coli immunostaining was detected in kidneys of naive mice (not shown), and almost no staining was detected in kidneys from mice inoculated with MG1655 bacteria (Fig. 1B). Positive E. coli immunostaining was detected in the kidneys colonized with the AL10 and AL511 isolates; the two UPEC isolates were observed adhering to the luminal surface of some, but not all, cells from the collecting duct sections (Fig. 1, C–F). Positive cytoplasmic staining was also observed in collecting duct cells from the kidneys of mice infected with AL511 isolates (Fig. 1F). To further characterize the bacterial attachment to collecting duct cells composed of principal and intercalated cells, sections were double stained with Abs raised against E. coli and the water channel AQP-2 or the chloride channel ClC-5 expressed in the principal and intercalated cells, respectively (35, 38). AL10 (Fig. 1G) and AL511 (not shown) isolates were observed adhering to ClC-5-expressing intercalated cells. In contrast, few or no bacteria adhered to AQP-2-expressing principal cells (Fig. 1H).
|
96 kDa in the mouse kidney (Fig. 2A). Immunohistochemical studies using this TLR4 Ab showed that most if not all cells from tubule sections, including collecting ducts, expressed TLR4, mainly in the cytoplasm (Fig. 2B). In addition, some of the collecting duct cells appeared intensely stained (Fig. 2B). As control, the staining was not detected when tissue sections were incubated with the anti-TLR4 Ab plus an excess of the peptide used for immunization (Fig. 2C). Examination of kidney sections from mice infected with the AL511 isolates also revealed that the collecting duct cells exhibiting intense positive E. coli staining (i.e., intercalated cells) corresponded to the more intensely TLR4-stained cells (Fig. 2, D and E). The preferential binding of UPECs to intercalated cells was also observed in ex vivo experiments using isolated MCDs microdissected from the kidneys of Lpsn mice. The MCDs formed confluent cell layers expressing K8–K18 cytokeratins and the tight junction-associated protein ZO-1 (Fig. 3A). Indirect immunofluorescence studies revealed that 50% of the MCD cell layers expressed the ClC-5 Cl– channel (Fig. 3A). Cell layers were not contaminated by hemopoietic cells, as they did not express any mRNA transcripts of the bone marrow-derived cell marker CD45 (Fig. 3B). Cultured MCDs expressed mRNAs for TLR4 and its accessory protein MD-2 and for the TLR4 adaptor proteins MyD88 and CD14 (Fig. 3C). They also expressed mRNAs for TLR2 and for the ENaC -subunit and CFTR Cl– channel, both of which were expressed in principal and intercalated cells (39, 40) (Fig. 3C). MCDs grown to confluence on filters developed high transepithelial electrical resistance (1941 ± 32 
· cm2, n = 28), and a negative potential (–32 ± 8 mV, n = 28), both of which are features of collecting duct cells (33). Overall, these results indicated that the confluent cultures of microdissected MCDs formed highly purified tight epithelial cell layers composed of 50% ENaC-expressing principal cells and 50% ClC-5-expressing intercalated cells. Three hours after the apical addition of E. coli AL511 isolates to confluent MCDs grown on filters, bacteria were found to be preferentially associated with the apical surface of the cells (Fig. 3, D-G) corresponding to ClC-5-positive stained cells (Fig. 3F, inset).
|
|
Lpsd and Lpsn mice were then inoculated with the E. coli strains AL10 and AL511 to determine the consequence of kidney colonization on the induction of proinflammatory mediators and to what extent the resulting inflammatory response was dependent upon TLR4 expression. Lpsd mice were more susceptible to infection by both of these pathogenic strains of E. coli, because the renal bacterial burden at day 2 after challenge was significantly x63-fold (AL10) to x108-fold (AL511) greater (p < 0.05) than in kidneys from Lpsn mice (Fig. 4A). As in Lpsn kidneys, the AL511 (and AL10; data not shown) isolates were found to be concentrated at the cell surface of collecting ducts cells (Fig. 4B, upper panels). Infiltrating cells surrounding tubules, including MCDs, were detected in Lpsn kidneys and also to a lower extent in Lpsd kidneys colonized with the AL511 (and AL10; data not shown) isolates (Fig. 4B, middle and lower panels).
|
in kidneys of infected Lpsn mice; AL511 isolates caused an increase in a broader range of proinflammatory mediators, including IL-1, MCP-1, and iNOS, than the AL10 isolates (Fig. 4C). In contrast, mRNA expression of all the proinflammatory mediators was dramatically reduced in infected Lpsd mouse kidneys, but this was less pronounced after infection with AL511 than with AL10 isolates (Fig. 4C). As a result, the expression of MIP-2, IL-1, and TNF- remained greater in AL511-infected than in AL10-infected Lpsd mouse kidneys (Fig. 4C), suggesting that some UPEC strains such as AL511 may induce significant TLR4-independent stimulation of proinflammatory mediators. MCD cells are involved in the induction of the inflammatory response to UPEC isolates
Because UPECs specifically interact with the apical surface of MCD cells expressing TLR4, we addressed the question of whether these epithelial cells directly contribute to the inflammatory response and chemoattraction of immune cells. Since in vivo experiments did not allow us to determine the exact role of parenchymatous renal cells in these processes, our experiments were conducted using primary cultures of MCDs microdissected from kidneys of naive Lpsn and Lpsd mice. Indirect immunofluorescence studies using the anti-ClC-5 Ab revealed that Lpsd MCDs exhibited the same percentage (50%) of intercalated cells as in Lpsn MCDs (data not shown). Apical addition of AL10 isolates to confluent cultures of Lpsn MCDs grown on filters caused dramatic and preferential increases in the expression of MIP-2 and TNF- compared with untreated Lpsn MCDs (Fig. 5A). Apical addition of AL511 isolates induced an even wider increase in the expression of proinflammatory mediators, including IL-1, MIP-2, MCP-1, iNOS, and TNF- (Fig. 5A). AL511 isolates, and to a lesser extent AL10 isolates, rapidly stimulated the expression of MCP-1 in Lpsn MCDs; but, in contrast to that observed in the day 2 postinfected kidneys, UPECs did not significantly stimulate the expression of RANTES in cultured Lpsn MCDs (Fig. 5A). Whichever strain of E. coli was used, the expression of proinflammatory mediators was considerably lower in Lpsd MCDs than in Lpsn MCDs (Fig. 5A). However, both UPEC isolates still induced significantly greater expression of MIP-2 (x13- to x15-fold) and TNF- (x6.4- to x7.2-fold) mRNA in Lpsd MCDs as compared with untreated Lpsd MCDs. The induction profiles of proinflammatory mediators triggered by UPECs closely resemble those observed in whole kidneys colonized by UPECs.
|
MIP-2 is required for neutrophil passage across the epithelial barrier of the infected urinary tract in experimental models of UTI (41, 42). The question arises of whether collecting duct cells may attract leukocytes by secreting chemokines toward their basal (i.e., interstitium) and/or apical (i.e., tubule lumen) sides. Adding UPECs to the apical side of confluent Lpsn MCDs grown on filters resulted in apical and basal secretion of MIP-2 (Fig. 5B, left panel). AL10 and AL511 isolates also stimulated apical and basal secretion of MIP-2 by confluent cultures of Lpsd MCDs, although to a lesser extent than in Lpsn MCDs (Fig. 5B, left panel). In contrast, the inactivation of the tlr2 gene did not affect the secretion of MIP-2. Apical and basal secretion of MIP-2 stimulated by UPECs remained identical in cultured MCDs dissected from wild-type and tlr2–/– mouse kidneys (Fig. 5B, right panel).
UPEC isolates activate both NF-B-dependent and -independent signaling pathways in MCD cells
TLR-mediated recognition of bacterial components by immune cells induces a cascade of signaling events that finally leads to the activation of the transcriptional factor NF-B, which is critical in the regulation of genes involved in inflammation (43). Signals that induce NF-B activity will cause the phosphorylation of the inhibitor proteins IBs by IB kinases as well as their dissociation and subsequent degradation, which then allows NF-B to translocate into the nucleus (44, 45). In Lpsn MCDs, AL511 bacteria induced time-dependent degradation of IB- and concomitant NF-B nuclear translocation that could be prevented by preincubating the cells with the specific NF-B inhibitor SN50 (Fig. 6A). In contrast, AL511 (and AL10; data not shown) isolates did not cause nuclear translocation of NF-B and did not alter the amount of IB- in Lpsd MCDs (Fig. 6A). The cellular activation caused by UPECs resulted in a subsequent transcriptional up-regulation of MIP-2 in Lpsn MCDs (Fig. 6A). The two UPEC strains also stimulated the secretion of MIP-2 independently of NF-B activation in Lpsd MCDs (Fig. 6B). The specific NF-B inhibitor SN50 reduced the production of MIP-2 elicited by the two UPEC strains in Lpsn MCDs by 50% but had no significant inhibitory action on MIP-2 produced by Lpsd MCDs (Fig. 6B).
|
(26, 27, 46, 47, 48). To address the role of MyD88 as well TRIF in the TLR4-mediated pathway of MCDs, the production of MIP-2 elicited by the two UPEC strains was measured in primary cultured MCDs dissected from the kidneys of MyD88–/– and TRIF-deficient Lps2–/– mice. The production of MIP-2 stimulated by UPECs was significantly reduced (p < 0.01) in primary cultured MyD88–/– MCDs as compared with that measured in wild-type C57BL/6 counterparts (Fig. 6C). Similarly to what is observed in Lpsd MCDs, SN50 had no significant inhibitory action on MIP-2 produced by cultured MyD88–/– MCDs (Fig. 6C). In contrast, the production of MIP-2 stimulated by UPECs remained identical in wild-type C57BL/6 and Lps2–/– MCDs (Fig. 6C). SN50 reduced to the same extent (50%) the production of MIP-2 elicited by the two UPEC strains in wild-type C57BL/6 and Lps2–/– MCDs (Fig. 6C). These results indicate that the TLR4-mediated stimulation of MIP-2 occurs through a MyD88-dependent pathway but not through a TLR4-mediated MyD88-independent pathway. In addition, 50% of secreted MIP-2 occurs through a TLR4-independent activated pathway. E. coli AL511 isolate activates MAPK pathways in MCD cells
The detection of NF-B-independent secretion of MIP-2 by Lpsn and Lpsd MCDs raised the question of which other signaling pathways are activated by UPECs. All subsequent experiments were conducted using the AL511 strain of E. coli. AL511 did not stimulate the expression of TRIF in both Lpsn and Lpsd MCDs (Fig. 7, A and C). In contrast, AL511 caused time-dependent activation of both phosphorylated p38 and ERK1/2 MAPKs in Lpsn MCDs but not in Lpsd MCDs (Fig. 7, A and C). In any case, AL511 did not affect the total amounts of ERK1/2 and p38. It is important to note that this pathogenic isolate had different effects on the expression of TRAF6 and TRAF2, two signaling adapter molecules shared by the IL-1R/TLR family and the TNFR superfamily that control NF-B, p38, and JNK signaling cascades (49). AL511 induced a time-dependent increase in TRAF6 when compared with the levels of -actin, used as internal standard, in Lpsn MCDs only (Fig. 7, A and C). In contrast, AL511 induced a time-dependent increase in TRAF2 in both Lpsn and Lpsd MCDs (Fig. 7, B and C). The fact that the adapter molecule TRAF2, which is involved in the TNF-induced activation of JNK independently of NF-B (50, 51, 52), is stimulated by AL511 Lpsd MCDs suggests that UPECs may directly activate JNK independently of TLR4 and TRAF6. AL511 produced a time-dependent increase in the amount of phosphorylated JNK without altering the amount of total JNK, not only in Lpsn MCDs but also in Lpsd MCDs (Fig. 7, B and C). Further evidence that this UPEC strain may activate TRAF2 independently of TLR4 was provided by the fact that AL511 also stimulated the phosphorylated form of ASK1, which is known to associate rapidly with TRAF2 in a TNF-dependent manner (53) in both Lpsn and Lpsd MCDs, (Fig. 7, B and C).
|
50% when Lpsn MCDs were preincubated with the JNK inhibitor SP600125 and the p38 inhibitor SB203580 as well as the MEK1/2 kinase inhibitor PD98059 (Fig. 8A). As controls, the phosphorylation of c-Jun and MAPKAPK-2, which are direct substrates of JNK and p38, respectively, and the phosphorylation of ERK1/2 were tested to assess the selectivity of the MAPK inhibitors at the concentration used in this study. In all cases, each of the inhibitors tested selectively inhibited their corresponding MAPK substrate without altering the phosphorylation of the other MAPKs (Fig. 8B). In contrast to what was observed in Lpsn MCDs, the secretion of MIP-2 caused by AL511, which was not significantly affected by SB203580 and PD98059, was almost completely inhibited when Lpsd MCDs were preincubated with the JNK inhibitor SP600125 (Fig. 8A). The MEK1/2, p38, and JNK inhibitors had very similar inhibitory effects on the secretion of MIP-2 elicited by AL511 in cultured wild-type C57BL/6 (MyD88+/+) MCDs and MyD88–/– MCDs, respectively (not shown). As in Lpsd MCDs, the JNK inhibitor SP600125 inhibited by 84% the production of MIP-2 elicited by AL511 in MyD88–/– MCDs (AL511, 1103 ± 107; AL511 plus SP600125, 173 ± 28 pg/ml, n = 6; p < 0.01).
|
was not totally abolished in Lpsd MCDs incubated with UPECs (see Fig. 4), the persistent production of TNF- should be sufficient to promote MIP-2 secretion via the TRAF2-ASK1-JNK pathway. The stimulated production of TNF-, although lower in AL511-treated Lpsd MCDs than in AL511-treated Lpsn MCDs, was still significantly greater (p < 0.05) than in untreated Lpsd MCDs (Fig. 8C). Preincubating the cells with an anti-mouse TNF- Ab significantly reduced (p < 0.05) the secretion of MIP-2 by 35 and 63% in Lpsn and Lpsd MCDs, respectively (Fig. 8C). Pretreatment of MCDs with the anti-mouse TNF- Ab also almost totally abolished the increase in phosphorylated ASK1 caused by AL511 in both Lpsn and Lpsd MCDs (Fig. 8D). These findings demonstrate that the TNF- produced by activated MCD cells may directly stimulate the secretion of MIP-2 by activating a TRAF2-ASK1-JNK signaling pathway and therefore account for the recruitment of leukocytes in UPEC-infected Lpsd mouse kidneys. The production of TNF- was dramatically increased in the kidneys of AL511-infected Lpsn mice and to a lesser extent in those of infected Lpsd mice as compared with that measured in the kidneys of naive mice (Fig. 8E). To better assess the participation of TNF- in renal bacterial clearance, bacterial counts and histological examinations were then performed on kidneys from Lpsn and Lpsd mice pretreated with the neutralizing anti-TNF- Ab 1 h before the intravesical inoculation of AL511 isolates. The kidneys from Lpsd mice treated with the anti-TNF- Ab were much more susceptible to AL511 infection than those from untreated Lpsd mice, because the renal bacterial burden at day 2 after challenge was significantly x104-fold greater than that for kidneys of untreated Lpsd mice infected with AL511 (Fig. 8F). Numerous intrarenal abscesses were more frequently detected in the infected kidneys from Lpsd mice treated with the neutralizing anti-TNF- Ab (seven of eight infected mice) than in those from untreated Lpsd mice (one of eight infected mice; data not shown) or Lpsd mice pretreated with the rat IgG1 isotype control (none of eight infected mice) (Fig. 8G). Neutralization of TNF- also led to significant greater increase (x127-fold) in the number of CFU in AL511-infected Lpsn kidneys than in the kidneys from untreated or IgG1 isotype control-pretreated Lpsn mice (not shown). However, no significant difference was observed in renal inflammatory reaction (data not shown). These results strongly suggest that, in addition to the activation of the main TLR4-dependent pathway, the activation of the TLR4-independent signaling pathway mediated by TNF- also participates in the renal clearance of UPECs. Fig. 9 summarizes the TLR4-dependent and -independent pathways and corresponding signaling molecules leading to stimulated secretion of MIP-2 by UPEC AL511 bacteria in mouse renal collecting duct cells.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
Previous studies have shown that renal epithelial cells are sensitive to purified LPS or intact UPECs. For example, the pyelonephritogenic E. coli ARD6 strain interacts with proximal tubule cells from young rats to induce large Ca2+ oscillations (19). Tsuboi et al. (21) also reported that LPS mediates the production of MCP-1 and RANTES in primary cultures of mouse renal tubule epithelial cells exhibiting a proximal phenotype. We provide here evidence that collecting duct epithelial cells may directly participate in the recognition and rapid induction of proinflammatory mediators in response to UPEC strains. We show that two pyelonephritis-associated E. coli isolates, AL10 and AL511, both of which colonize kidneys in the mouse model of the upper UTI and lack the virulent toxin -hemolysin, interact with the apical membrane of collecting duct cells and, more specifically, with intercalated cells, suggesting that this particular cell type may be a preferential site for bacterial attachment. The profile of stimulated proinflammatory mediator expression caused by UPECs is quite similar in the day 2 postinfected kidneys and in cultured Lpsn MCDs. However, UPECs that stimulate the expression of RANTES in whole kidneys did not activate its expression in cultured MCDs. LPS has been shown to stimulate the production of both MCP-1 and RANTES in cultured mouse primary renal proximal tubule cells (21). Although the NF-B activation seems to be essential for LPS-mediated up-regulation of these two chemokines, the inhibition of JNK and p38 MAPK impairs the expression of RANTES but not of MCP-1 (21), suggesting that RANTES and MCP-1 are differently regulated. Such dissimilarity in the regulation of these two mononuclear-directed chemokines may also account for the differences observed in day 2 postinfected kidneys exhibiting predominant expression of RANTES in interstitial cells on one hand and a preferential expression of MCP-1 in the 3-h infected MCDs on the other hand. Alternatively, the disparity observed may also reflect differences in the stimulated rates of induction and/or degradation of the two chemoattractant chemokines.
The molecular mechanism(s) by which E. coli strains activate renal epithelial cells still remain not fully elucidated. In C3H/HeJ Lpsd mice the renal inflammatory response to P fimbriated E. coli is TLR4 dependent (55, 56), and this activation signaling in epithelial A498 cells lacking CD14 was independent of LPS and lipid A myristoylation (56). More recently, Fischer et al. (15), using murine models of ascending UTI, have shown that P and type 1 fimbriated E. coli may use different adaptor molecules to influence neutrophil activation and bacterial clearance, but that in both cases MyD88 is required for efficient bacterial clearance. These results suggested that P and type 1 fimbriae, through their recognition by their respective epithelial cell receptors, may engage specific TLR4-associated proteins for the induction of innate immune response in the urinary tract (15). In this study, we show that murine wild-type MCDs in primary culture constitutionally express a functional TLR4 and its accessory molecules, including CD14, and are an important source of chemokines and cytokines following the adhesion of UPECs at their cell surface. The interaction of the two UPEC isolates with collecting duct intercalated cells induces an early and strong inflammatory response characterized by the up-regulation of proinflammatory mediators, including proinflammatory cytokines (IL-1, IL-6, and TNF-), chemokines (MIP-2, RANTES, and MCP-1), and iNOS in the infected kidneys from Lpsn mice. Consistent with what is observed in UPECs-infected bladder epithelial cells (14, 19), the inflammatory response in kidneys colonized by the AL10 and AL511 UPEC isolates appears to be mediated mainly by TLR4. However, the two UPEC strains induced markedly different stimulation profiles of proinflammatory mediators in the kidneys of Lpsn and Lpsd mice, suggesting that different bacterial factors produced by the UPECs may be implicated in the activation of distinct TLR4-dependent and -independent signaling pathways. MIP-2 and its human counterpart IL-8 play key roles in the migration of neutrophils to infected mucosal sites to protect them against invading pathogens (41, 42, 57). The two UPECs stimulated MIP-2 mRNA expression and protein secretion in both Lpsn kidney and cultured Lpsn MCDs through a main TLR4-dependent pathway. The results obtained with cultured MCDs dissected from the kidneys of MyD88–/– and Lps2–/– mice also strongly suggest that UPECs activate TLR4 pathway through a predominant, MyD88-dependent, mediated pathway. The persistent activation of proinflammatory mediators (mainly MIP-2 and TNF-) in Lpsd mouse kidneys, together with the persistent stimulation of basal and apical secretions of MIP-2 in Lpsd MCDs caused by both UPEC strains, further suggested that UPECs may also activate MCD cells via nonmediated TLR4 cellular signaling pathways.
In mammalian cells, the recognition of LPS by TLR4 results in the recruitment of multiple signaling molecules including TRAF6, which activates downstream components involving NF-B, p38, and JNK (58). UPEC AL511 isolates activate the TLR4-mediated ERK1/2 signaling pathway and increase the abundance of TRAF6 and down-stream activation of NF-B, p38, and JNK in Lpsn MCDs only (Fig. 9). Conversely, UPEC AL511 isolates had no stimulatory effect on these signaling molecules in TLR4-defective Lpsd MCDs. As a result of the TLR4-mediated, TRAF6-activated pathway, the stimulated secretion of MIP-2 in Lpsn MCDs is partially or almost entirely abolished by the inhibitors of NF-B, p38, or MEK1/2. We also demonstrate that the NF-B-independent stimulation of MIP-2 caused by AL511 depends mainly, if not exclusively, on the activation of JNK in Lpsd MCDs (Fig. 9). JNK is activated by TNF- and IL-1 and by exposing cells to a variety of environmental stress conditions (58). TRAF2 transduces the signals required for TNF-mediated activation of NF-B (59), p38, and JNK (50, 60). However, distinct activated cellular mechanisms may lead to diversification of TRAF2 signaling. Analysis of embryonic fibroblasts from TRAF2-deficient mice that reveals a severe reduction in TNF-mediated JNK activation and only a mild effect on NF-B activation suggests that a TRAF2-independent pathway of NF-B activation must exist (51). Habelhah et al. (61) have shown that TNF-induced ubiquitination of TRAF2, resulting from its translocation into insoluble lipid rafts, is required for the activation of JNK, but not for that of p38 and IB kinases. UPEC strain AL511 also stimulates TRAF2 and activates JNK, but not NF-B or p38, in TLR4-defective Lpsd MCDs. The fact that the selective inhibition of JNK by SP600125 almost completely prevented the stimulation of the secretion of MIP-2 caused by UPECs in Lpsd MCDs provides further evidence that the TRAF2-JNK pathway corresponds to the TLR4-independent pathway activated by UPECs in Lpsd MCDs.
The interaction between UPECs and both Lpsn and Lpsd MCDs caused increased expression of ASK1. This evolutionarily conserved MAPK kinase kinase is activated in response to various cytotoxic stresses such as TNF- and reactive oxygen species (ROS) (62). ASK1 is selectively required for the TLR4-mediated, LPS-induced activation of both JNK and p38 (63). The stimulation of TRAF6 and ASK1 caused by strain AL511 therefore must account for the activation of p38 in Lpsn MCDs. In contrast, the UPEC isolate does not affect TRAF6 but does stimulate TRAF2 and activates ASK1 in Lpsd MCDs. These findings strongly suggest that the formation of a TRAF2-ASK1 complex leads to the subsequent activation of JNK and, consequently, to stimulated production of MIP-2 in Lpsd MCDs. Recently, the antioxidant pyrrolidine dithiocarbamate has been shown to induce NF-B-independent MIP-2 promoter activation mediated via the JNK pathway and the subsequent activation of the AP-1 transcription factor in mouse macrophage RAW 264.7 cells (64). These findings suggest that the TLR4-independent secretion of MIP-2 stimulated by UPECs results from the activation of AP-1 mediated via the TRAF2-ASK1-JNK pathway. We also provide some evidence suggesting that the UPEC isolates stimulate the production of TNF-, particularly in Lpsn MCDs, but also, although to a much lesser extent, in Lpsd MCDs. Interestingly, the neutralization of renal TNF- production significantly altered the production of MIP-2 and blunted the induction of phosphorylated ASK1 caused by the AL511 isolate in both Lpsn and Lpsd MCDs. These data therefore provide convincing evidence that the stimulation of TNF- production by epithelial collecting duct cells activated by UPECs may directly activate the TLR4-independent production of MIP-2. Matsuzawa et al. (65) have demonstrated that ASK1 specifically mediates LPS-induced TLR4 signaling via a ROS-dependent activation of the TRAF6-ASK1-p38 pathway. This means that we cannot rule out the possibility that ROS produced locally by polymorphonuclear neutrophils or collecting duct cells may activate the TLR4-dependent TRAF6-ASK1-p38 pathway in Lpsn MCDs and, together with TNF-, activate the TLR4-independent, ASK1-JNK-activated pathway that we have identified in Lpsd MCDs. Alternatively, UPECs might activate TLR4-dependent and/or TLR4-independent pathway through their recognition by other pattern recognition receptors. TLR11, which is expressed in murine tubule epithelial cells but not in humans, has also been shown to play a key role in the renal bacterial clearance from mouse kidneys through the activation of identical TLR4-dependent signaling pathways (66). Nevertheless, our results strongly suggest that the TLR4-independent pathway activated by UPECs also participates in the protection of the kidneys against ascending pathogens; after blocking TNF-, TLR4-deficient Lpsd mice as well as Lpsn mice are much more susceptible to renal bacterial invasion than their respective untreated Lpsn and Lpsd mice counterparts and develop more rapid and extended renal inflammatory lesions. In human pathology, uncomplicated UTIs and acute pyelonephritis caused by UPECs are the most common forms of bacterial infection in renal transplant recipients. UTIs are thought to be directly attributable to exposure to pathogens during the early postoperative period and to immunosuppressive therapy (67). A recent major retrospective study has revealed that UTIs occurring after renal transplantation are often associated with an increased risk of septic shock and subsequent death (68). Therefore, it cannot be excluded that the frequent occurrence of ascending pyelonephritis caused by UPECs in renal transplanted patients may be favored, at least in part, by alterations in the activation of cellular signaling pathways due to prolonged immunosuppressive therapy. Further clinical studies will be needed to test this hypothesis.
To summarize, the present study contributes to elucidation of the role played by collecting duct intercalated cells in recognizing pyelonephritic E. coli isolates and provides evidence that collecting duct cells actively participate to the renal clearance of bacteria by mediating a reliable immune response via TLR4-mediated NF-B-, ERK1/2-, p38-, and JNK-activated signaling pathways or via an alternative TLR4-independent, TNF-mediated, TRAF2-ASK1-JNK-activated pathway.
|
Acknowledgments |
|---|
|
Disclosures |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
|
Footnotes |
|---|
1 This work was funded by Institut National de la Santé et de la Recherche Médicale (INSERM) and by Institut Pasteur Grant PTR 165 (to C.L.B. and D.B.-G.). A.V. was in receipt of an Interface INSERM-AP-HP fellowship. 
2 C.L.B, D.B.-G, and A.V. made equal contributions to this work.
3 Address correspondence and reprint requests to Dr. Alain Vandewalle, Institut National de la Santé et de la Recherche Médicale U773, Centre de Recherche Biomédicale Bichat-Beaujon CRB3, Unité de Formation et de Recherche de Médecine Xavier Bichat, BP 416, 16 Rue Henri Huchard, F-75870 Paris Cedex 18, France. E-mail address: vandewal{at}bichat.inserm.fr
4 Abbreviations used in this paper: UTI, urinary tract infection; AQP-2, aquaporin-2; ASK1, apoptosis signal-regulatory kinase 1; CFTR, cystic fibrosis transmembrane conductance regulator; ClC-5, chloride channel 5; ENaC, epithelial sodium channel; iNOS, inducible NO synthase; MAPKAPK-2, MAPK-activated protein kinase 2; MCD, medullary collecting duct; ROS, reactive oxygen species; TRAF, TNF receptor-associated factor; TRIF, TIR domain-containing adaptor inducing IFN-; UPEC, uropathogenic Escherichia coli.
Received for publication April 5, 2006. Accepted for publication July 12, 2006.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
