Although TLR5 regulates the innate immune response to bacterial flagellin, it is unclear whether its function is essential during in vivo murine infections. To examine this question, we challenged Tlr5−/− mice transurethrally with Escherichia coli. At 2 days postinfection, wild-type mice exhibited increased inflammation of the bladder in comparison to Tlr5−/− mice. By day 5 postinfection, Tlr5−/− mice had significantly more bacteria in the bladders and kidneys in comparison to wild-type mice and showed increased inflammation in both organs. In addition, flagellin induced high levels of cytokine and chemokine expression in the bladder that was dependent on TLR5. Together, these data represent the first evidence that TLR5 regulates the innate immune response in the urinary tract and is essential for an effective murine in vivo immune response to an extracellular pathogen.
Toll-like receptors are a family of germline-encoded innate immune receptors that recognize pathogen-associated molecular patterns, such as bacterial flagellin (TLR5), LPS (TLR4), and lipopeptides (TLR1/2/6) (1). Expression of TLRs varies among cells and tissues, suggesting that individual TLRs may regulate distinct pathogen and organ-specific roles in host defense to different pathogens (1). We previously discovered that the ligand for TLR5 is bacterial flagellin, the most abundant protein in the whip-like tails of flagellated bacteria (2). We defined the TLR5 recognition site on flagellin and found that it is conserved among a wide variety of flagellated bacteria (3), although select bacterial species possess unique flagellin molecules that evade TLR5 recognition (4). TLR5 is expressed in epithelial cells of the airways, intestine, and urogenital tract, as well as on hemopoietic cells of the innate and adaptive immune system and has recently been shown to be involved in the transport of flagellated Salmonella typhimurium from the intestinal tract to the mesenteric lymph nodes (5).
Many important pathogenic bacteria, both Gram-positive and Gram-negative, are flagellated. Flagellated uropathogenic E. coli (UPEC)4 cause 70–90% of all urinary tract infections (UTI), and their pathogenesis involves contact between bacteria and the epithelial cell surface of the urogenital tract, a site of TLR5 expression in humans (6). UPEC colonize the urethra and ascend to the bladder, where they can persist at high levels (7). In addition to cystitis in the bladder, UPEC may ascend to the kidney and cause serious complications, including pyelonephritis and bacteremia (8). E. coli is recognized by several TLRs, including TLRs 2, 4, 5, and 11, and likely also TLR9. Previous studies indicate that TLR4 and TLR11 regulate susceptibility to UTIs (9, 10, 11). However, it is not currently known whether TLR5 is critical for host defense to UTIs or whether there is sufficient TLR redundancy to obviate its requirement.
Materials and Methods
Mice, bacteria, and TLR agonists
Tlr5−/− mice (strain designation B6.129P2-Tlr5tm1Aki) were derived and backcrossed to a C57BL/6 background for eight generations as previously described (5). Wild-type (WT) control mice were from a C57BL/6 background (The Jackson Laboratory). E. coli strain CFT073, from a patient with acute pyelonephritis (American Type Culture Collection), was grown in Luria-Bertani (LB) medium in static culture at 37°C for 48 h. Expression of type 1 pili was confirmed for each experiment by testing for yeast agglutination (12). Flagellin was purified from S. typhimurium as described in Ref. 3 and was heated to 70°C for 15 min to monomerize it. Contaminating endotoxin was removed by passage through a 100-kDa molecular mass cutoff filter (Millipore) followed by endotoxin removal on a polymixin B column (Pierce). The resulting flagellin did not show detectable endotoxin by Limulus assay (Cambrex). Ultrapure LPS was purchased from List Biologicals.
RNA was extracted from organs with TRIzol (Invitrogen Life Technologies), DNase treated with TURBO DNA-free (Ambion), and cDNA produced with Superscript II (Invitrogen Life Technologies). Real-time PCR was performed with TaqMan Fast (Applied Biosystems) on an Applied Biosystems Prism 7900 HT. Primer/probe sets for elongation factor 1α (EF1α) were designed with Primer Express 1.0 (PerkinElmer) with a 5′-FAM and 3′-TAMRA modification (Biosearch Technologies). Primer/probe sets for mouse TLR5 and cytokines/chemokines were purchased from Applied Biosystems. Threshold cycle (Ct) values were transformed by 1/2Ct, and then normalized to EF1α for each organ.
NF-κB luciferase reporter assay
Chinese hamster ovary (CHO) K1 cells stably expressing mouse TLR5 and NF-κB luciferase reporter constructs (3) were stimulated with heat-killed bacteria for 4 h and assayed for luciferase activity. Assays were done in duplicate, and the experiment was repeated three times. Percent fold induction was calculated by dividing the luciferase values for each bacterial dose by the maximal luciferase value for the bacteria in each experiment.
UTI model of infection
The Institute for Systems Biology and Osaka University Institutional Animal Care and Use Committees approved all animal protocols. Forty-eight-hour static cultures of E. coli CFT073 were resuspended in cold PBS at 1 × 109 CFU/ml. Anesthetized mice were inoculated transurethrally with 5 × 107 E. coli in 50 μl, and urethras were coated with collodion (Sigma-Aldrich) (13). Six hours postinfection, collodion was removed by blotting with acetone. At each time point, organs were homogenized in 1 ml of 0.025% Triton X-100/PBS and plated on LB-agar to enumerate CFUs.
Comparisons were made with a two-tailed Mann-Whitney U test or a Student’s t test. A p ≤ 0.05 was considered to be significant. Statistics were calculated with PRISM4 (GraphPad).
Bladders and kidneys were fixed in 10% formalin-buffered saline and embedded in paraffin. Four-micrometer sections were cut, stained with H&E, and examined by a pathologist blinded to mouse genotype.
Results and Discussion
Tlr5−/− mice are more susceptible to E. coli urinary tract infection
To test our hypothesis that TLR5 is critical for host defense against E. coli UTIs, we first examined bladder and kidney tissue for TLR5 expression. We extracted RNA from tissues of C57BL/6 mice and evaluated expression levels by real-time PCR (Fig. 1,A). TLR5 was expressed in both bladder and kidney, which suggested that it may regulate critical aspects of the immune response during UTI. We next examined whether Tlr5−/− mice were more susceptible to urinary tract infections. When grown in static culture, uropathogenic E. coli forms type 1 pili (12) that enhance adherence to bladder epithelia and increase bladder colonization. We first determined whether growth in static culture resulted in flagellin expression. CHO cells stably expressing mouse TLR5 and a NF-κB-dependent luciferase reporter construct responded to statically grown heat-killed bacteria in a dose-dependent manner, detecting fewer than 8000 bacteria, a multiplicity of infection of ∼0.1 (Fig. 1 B). Control CHO cells expressing the pEF6 vector alone did not respond to bacteria (data not shown).
To test the role of TLR5 during infection in vivo, we inoculated WT and Tlr5−/− mice transurethrally with 5 × 107 CFU of statically cultured E. coli. Bladder and kidneys were harvested 4 h, 24 h, 2 days, and 5 days after infection and the number of CFU in each organ was determined. At early time points, no difference was seen in bladder CFU between WT and Tlr5−/− mice (Fig. 1,C). In contrast, although all mice remained infected at day 5, the number of CFU per bladder was reduced in WT mice but rose dramatically in Tlr5−/− mice. WT mice had a median of 475 CFU/bladder (interquartile range (IQR): 212.5–685) in comparison to Tlr5−/− mice with a median of 7.7 × 105 CFU/bladder (IQR: 9.0 × 104–4.1 × 106, p < 0.0001 by Mann-Whitney U test) (Fig. 1 C).
We next examined whether the E. coli disseminated to the kidney. Bacteria were present in the kidneys of both WT and TLR5-deficient animals, and no significant CFU differences were observed between WT and Tlr5−/− mice in the kidney at early time points (Fig. 1,D). By day 5 after infection, however, no bacteria were detected in WT kidneys, but Tlr5−/− mice had a median of 6.5 × 102 CFU/kidney (IQR: 0- 6.5 × 104, p = 0.0274 by Mann-Whitney U test) with six of nine infected mice showing kidney counts (Fig. 1 D). Together, these data suggest that Tlr5−/− mice are unable to control bacterial replication and cannot clear the infection from the kidneys by day 5.
Tlr5−/− mice exhibit decreased inflammation at 2 days postinfection
We next examined histologic sections from WT and knockout mice by light microscopy to determine the pathologic consequences of TLR5 deficiency. Bladders and kidneys from WT and Tlr5−/− mice exhibited similar levels of inflammation at 4 and 24 h after infection (data not shown). In contrast, at 2 days postinfection, Tlr5−/− mice showed decreased inflammation in the bladder relative to WT mice (Fig. 2 A). WT mice exhibited prominent submucosal edema and infiltration of the submucosa and epithelium by leukocytes. There was no significant inflammation in the kidneys from WT or Tlr5−/− mice at the 2-day time point (data not shown).
By day 5 postinfection when TLR5-deficient mice showed increased bacterial counts, the situation was reversed, with prominent inflammation in the TLR5-deficient animals. WT mice showed minimal to no inflammation in the bladder and kidney, and bacteria were not visible in the lumen of these organs (Fig. 2, B and C). In contrast, bladders from Tlr5−/− mice showed prominent submucosal edema with leukocyte infiltration into the submucosa and invasion of the epithelial layer, as well as focal microabscess formation and accumulation of leukocyte-rich exudates on the bladder surface (Fig. 2,B). Bacteria were readily visible and present predominantly on the surface of the Tlr5−/− urothelium without prominent evidence of enclosure within the umbrella cells of the bladder, such as the intracellular bacterial communities described by others (14). A similar but less pronounced pattern of inflammation was also present in the Tlr5−/− kidneys at day 5, and primarily involved the urothelium in the renal pelvis with focal extension into the interstitium of the renal medulla (Fig. 2 C and data not shown). Together, these data indicate a turning point at 2 days postinfection: Tlr5−/− mice manifest decreased inflammation that leads to overwhelming bacterial growth and more severe inflammation by day 5 after infection.
Flagellin induces early expression of proinflammatory genes in the bladder
To identify proinflammatory molecules up-regulated by TLR5 in the bladder that might lead to increased early inflammation in WT mice, we examined the in vivo response to transurethral administration of flagellin and compared this with LPS, another prominent TLR agonist present in E. coli. Mice were inoculated transurethrally with 30 μg of flagellin or 10 μg of ultrapure LPS in PBS, and bladders were harvested at 4 h. Real-time PCR was performed on bladder tissue for several proinflammatory cytokines and chemokines. Transurethral inoculation of flagellin up-regulated expression of KC (CXCL1), MIP2 (CXCL2), MCP-1 (CCL2), IL-6, and TNF-α mRNA, but not β-defensin 1 mRNA in WT mice (Fig. 3). As expected, Tlr5−/− mice did not respond to flagellin in the bladder. In contrast to flagellin, LPS delivered into the bladders of WT mice did not induce transcription of these proinflammatory genes. These results demonstrate that flagellin induces a robust TLR5-dependent innate immune response in the murine bladder that may account for its critical role in UTI pathogenesis.
Taken together, these data demonstrate that TLR5 plays a crucial role in host defense to UPEC infection by mediating flagellin-induced inflammatory responses in the bladder that limit bacterial replication in both the bladder and kidney. Two additional TLRs, TLR4 and TLR11, have also been shown to play a role in E. coli-induced UTI. TLR4-deficient C3H/HeJ mice exhibit a reduced inflammatory response to UPEC and exhibit significantly higher bacterial counts in the bladder and kidneys (9, 15, 16). In contrast to the inflammatory response to flagellin, we found that the bladder was relatively unresponsive to LPS. There is conflicting evidence about whether the urinary epithelium responds to LPS (12, 17, 18, 19). Our experiments comparing in vivo delivery of highly purified flagellin and LPS suggest that TLR5 and TLR4 regulate distinct bladder innate immune responses with TLR5 regulating a relatively dominant role initially. TLR11, which is a pseudogene in humans, is expressed in both kidney and bladder epithelial cells of mice (11). Infection of TLR11-deficient mice with the UPEC strain 8NU resulted in approximately equal colonization of the bladders of TLR11-deficient and WT mice, but significantly more bacteria ascended to the kidneys of TLR11-deficient mice (11). Thus, it is possible that TLR5 and TLR11 play complementary roles in the mouse urinary tract, with TLR5 limiting bladder replication and TLR11 primarily controlling bacterial invasion of the kidney.
Bacterial motility is important for virulence in some models of UTI, which suggests that TLR5-flagellin interactions may be important for bacterial uropathogenesis. Flagellum-negative mutants of Proteus mirabilis are significantly less successful at bladder colonization and do not progress to the kidneys as readily (20). Furthermore, two recent studies suggest that flagellar mutants of E. coli are less able to colonize the mouse urinary tract (8, 14). Flagellar-based motility may be beneficial in early colonization of the urinary tract, but may not be required for maintenance of infection (8). This is in agreement with a study that demonstrated down-regulation of flagellin genes by E. coli CFT073 several days after in vivo infection (21). These studies combined with our results suggest that TLR5 recognition of flagellin is an important component of the innate immune response to E. coli during the early stages of UTI when flagellin expression and motility contribute to colonization of the urinary tract.
In addition to TLR5 recognition of extracellular bacterial flagellin, two novel intracellular flagellin receptors have recently been described that are both members of the nucleotide-binding oligomerization domain leucine-rich repeat family (22, 23, 24). Naip5 detects flagellin from Legionella pneumophila that reaches the macrophage cytosol via the bacteria’s type IV secretion system (24, 25). Ipaf detects cytoplasmic flagellin injected into macrophages by the type III secretion system of S. typhimurium (22). The roles of these additional flagellin receptors in UTI are not known, but the UPEC strain used in this study, like many other UPEC (26), does not encode a type III secretion system (27). Thus, intracellular flagellin receptors may not be able to compensate for the lack of flagellin recognition by TLR5 in this infection model.
Although TLR5 has been implicated in the innate immune response to mucosal infection (5, 28), there has been limited in vivo data to substantiate this claim. A recent study of Tlr5−/− mice did not find a unique role for TLR5 in defense to Salmonella or Pseudomonas (29). Our study provides the first evidence that TLR5 regulates a critical and nonredundant role in the innate immune response to a murine infection with extracellular flagellated bacteria. We have previously identified a TLR5 polymorphism present in the general population that results in a stop codon that abrogates TLR5 signaling and is associated with increased susceptibility to Legionnaire’s disease (30). These murine studies support a hypothesis that individuals who possess this TLR5 variant will also be more susceptible to UTI.
We thank April Clark and David Rodriguez for expert assistance with animal care, Cynthia Lorang for real-time PCR assistance, and Evgeni Sokurenko and Scott Weissman for advice.
The authors have no financial conflict of interest.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
T.R.H. was supported by National Institutes of Health Grant R01 AI061464-02, K.D.S. was supported by National Institutes of Health Grant R01 AI062859, and A.A. was supported by National Institutes of Health Grant R01 AI052286.
Abbreviations used in this paper: UPEC, uropathogenic E. coli; UTI, urinary tract infection; EF1α, elongation factor 1α; CHO, Chinese hamster ovary; LB, Luria-Bertani; IQR, interquartile range; WT, wild type.