Enterococcus faecalis Tropism for the Kidneys in the Urinary Tract of C57BL/6J Mice
http://www.100md.com
感染与免疫杂志 2005年第4期
Department of Molecular Microbiology, Washington University, St. Louis, Missouri
ABSTRACT
Enterococcus faecalis is a gram-positive bacterium that can cause a variety of nosocomial infections of which urinary tract infections are the most common. These infections can be exceptionally difficult to treat because of drug resistance of many E. faecalis isolates. Despite their troublesome nature, little is known about the host or bacterial factors necessary for E. faecalis to cause disease in the urinary tract. Using a mouse model of urinary tract infection, we have shown that E. faecalis is capable of persisting in the kidneys of mice for at least 2 weeks. In contrast, bacterial titers from the bladders of the same mice were inconsistent and tended to be much lower than those recovered from the kidney. This preference for the kidney over the bladder is also observed in other clinical E. faecalis strains. Histologic examination of bladder and kidney tissues demonstrated that E. faecalis induced an inflammatory response in the kidney but not in the bladder. This inflammatory response was TLR2 independent and did not induce inflammatory markers typically associated with uropathogenic Escherichia coli. Using a competition assay, we demonstrated that a pyelonephritis clinical isolate had a growth advantage over a laboratory strain of E. faecalis in the kidneys but not in the bladders of mice. Taken together, these results demonstrate that E. faecalis has tropism for the kidneys in the urinary tracts of mice and that this system can be used to study factors involved in the pathogenesis of urinary tract infections.
INTRODUCTION
Enterococcus faecalis, while normally a gut commensal, is a frequent cause of many serious human infections, including urinary tract infections, endocarditis, bacteremia, and wound infections. Among the diseases that E. faecalis causes, urinary tract infections are the most common, responsible for approximately 110,000 cases yearly, many of which are nosocomial. Infections with E. faecalis can be especially troublesome to treat because of their frequent resistance to multiple antibiotics, including vancomycin, a drug of last resort for many gram-positive infections (for a review, see reference 8).
Both rat and mouse model systems have been used to study factors involved in the pathogenesis of E. faecalis in the urinary tract. Studies by Guze and colleagues showed that E. faecalis has a growth advantage over other enterococcal species in rat kidneys in a hematogenous inoculation pyelonephritis model (18). Another model system was used to show that pyelonephritis caused by Pseudomonas aeruginosa was aggravated by coinfection with E. faecalis, as determined by histological changes in the kidney (30). In this model, ligation of the urethra was used to induce retrograde reflux of bacteria into the kidney, increasing the susceptibility of the mouse to infection. In a bladder catheterization model of urinary tract infection, the Esp (enterococcal surface protein) adhesin was found to increase persistence in the urinary bladder of mice, although no histological changes were observed (26). Another study was unable to demonstrate a critical function for aggregation substance, a well-characterized enterococcal adhesin, in a murine model of cystitis (12). Thus, a robust murine cystitis model of E. faecalis infection has yet to be established, most likely due to complicating host and bacterial factors reflecting the adaptations enterococcus has evolved to coexist with its host. In order to more effectively devise strategies and therapies to prevent and treat enterococcal urinary tract infections, an infection model that can give a sensitive readout of virulence factors must be developed.
In contrast, many of the host-pathogen interactions critical for bladder infections caused by uropathogenic Escherichia coli, the most common etiologic agent of urinary tract infections, have been well-defined in murine models. Uropathogenic E. coli initially interacts with bladder epithelium through the expression of type 1 pili, which mediate adhesion and invasion into the uroplakin-expressing umbrella cells on the surface of the bladder (19). The invasion of E. coli into the urinary epithelium results in expression of a variety of cytokines, induction of apoptosis in infected cells, and a massive neutrophil inflammatory response. This response is mediated primarily by Toll-like receptor 4, which has been demonstrated to respond to lipopolysaccharide (22, 23). The interaction of E. coli with the mouse kidney epithelium during pyelonephritis is less well characterized, but the importance of P pili is well established as an initial mediator of adherence in other animal models (21). Similar to bladder infections, acute E. coli-mediated pyelonephritis in mice is characterized by a strong neutrophil immune response (4).
In this study, we describe the tropism of E. faecalis for the kidneys in C57BL/6J mice. The consequences of this tropism were investigated by characterizing the host response to E. faecalis and contrasting these differences to the response induced by uropathogenic E. coli.
MATERIALS AND METHODS
Strains and growth conditions. A summary of all the strains used in this study, their origins, and relevant drug resistances are shown in Table 1. For infection of mice, E. faecalis strains were grown statically overnight (typically 12 to 15 h) at 37°C in brain heart infusion (BHI) medium (Difco) without antibiotics. Clinical strains other than 0852 were contributed by Thomas Hooten and Walter Stamm (University of Washington) and categorized as either cystitis or pyelonephritis strains using clinical criteria.
Animals. Female wild-type C57BL/6 mice were obtained from Jackson Laboratories and C57BL/6 TLR2–/– mice were a gift of S. Akira (29). TLR2–/– mice were 36 to 48 weeks in age at the time of inoculation. Wild-type 12-week-old and 52-week-old mice were used to control for the various ages of the TLR2–/– mice. These experiments demonstrated that there was no substantial difference in the recoveries of E. faecalis due to the ages of the mice. In all other experiments, mice were 8 to 12 weeks of age.
Inoculation and CFU enumeration. Cultures were collected by centrifugation at 6,000 x g for 10 min and resuspended in phosphate-buffered saline (PBS) to an approximate density of 107 to 108 CFU/ml. Female C57BL/6J mice were anesthetized with inhaled isoflurane and then inoculated transurethrally with 200 μl of bacterial suspension similar to methods described in Schilling et al. (23). Although the inoculation volume of 200 μl resulted in leakage during and after inoculation, we found that this volume resulted in the most consistent infections.
At the appropriate time points, mice were sacrificed by cervical dislocation after inhalation anesthesia and the bladders and kidneys were harvested. For histologic evaluation, tissue was fixed in neutral buffered formalin and embedded in paraffin. To determine the number of bacteria present in these tissues, bladders or pairs of kidneys were homogenized in 0.025% Triton X-100 in PBS and plated at different dilutions on BHI agar (BactoAgar from BD) supplemented with antibiotics where appropriate. CFU were enumerated after 24 h of incubation at 37°C. Statistical analysis of the kinetic experiment was performed using Student's one-tailed t test for correlated samples using the logarithmic values at each time point. When the recovered titers were below the limit of detection, the recovered CFU value was set to 1 for statistical analysis.
Immunohistochemistry. For immunohistochemical analysis of bladder tissue, 5-μm-thick sections were prepared. Sections were deparaffinized using Hemo De (Fisher) (twice for 10 min), rinsed with isopropanol (three times for 3 min) and washed with PBS (three times for 5 min). Tissue sections were blocked with 1% bovine serum albumin, 0.2% nonfat dry milk, and 0.3% Triton X-100 in PBS (PBS-BB) for 15 min at room temperature. Primary antibody raised in rabbit against Streptococcus group D antigen (Lee Laboratories) was added in PBS-BB and incubated overnight at 4°C. After PBS washings (three times for 5 min), tissue was incubated with Alexafluor 555-labeled donkey anti-rabbit antibody (Molecular Probes) in PBS-BB and incubated overnight at 4°C. After PBS washes (three times for 5 min), tissue was counterstained with bis-benzimide (Sigma) to reveal nuclear morphology.
Quantitative real-time PCR. Methods were as described in Mysorekar et al. (20) with the following modifications. Female C57BL/6J mice were inoculated transurethrally with 200 μl of PBS, E. faecalis 0852, or uropathogenic E. coli isolate NU14 grown statically in Luria-Bertani (LB) for 48 h to induce type 1 pilus expression. At the time of sacrifice, kidneys were divided lengthwise, and one half was harvested for RNA isolation and the other half was titered to confirm the presence of bacteria. RNA from infected bladders and kidneys was collected either 6 h or 24 h after infection individually from each mouse using a commercially available affinity matrix-based kit (RNeasy kits, Qiagen). cDNA was generated using random hexamers and quantitative real-time PCR was performed as described (20) using the Bio-Rad iCycler. Primers for amplification of glyceraldehhyde-3-phosphate dehydrogenase, Mip-2, and Socs-3 are as previously described (20).
Competitive infection. Cultures of OG1X and BP78 were resuspended in PBS and mixed in equal volumes to make the mixed inoculum. The BP78 to OG1X ratio was determined by plating the mixed inoculum on medium selective for each strain; 48 h after infection, mice were sacrificed and their bladders and kidneys were collected and homogenized as described above. The homogenate was then plated onto either BHI agar supplemented with streptomycin, to select for OG1X, or tetracycline, to select for BP78. The competition index was calculated similarly to Freter et al. using OG1X as the reference strain (3). Briefly, the competition index = [(CFU of BP78/CFU of OG1X recovered from mice)/(CFU of BP78/CFU of OG1X present in initial inoculum)].
RESULTS
E. faecalis can cause a reproducible infection in C57BL/6J mouse kidneys. In order to study E. faecalis cystitis, we inoculated 50 μl of 108 CFU of an E. faecalis strain, 0852, from a diagnosed urinary tract infection transurethrally into female C57BL/6J mice. This protocol led to inconsistent recovery of bacteria from the bladder but noticeably higher titers in the kidney. Based on this observation, we used a 200-μl inoculum volume to intentionally induce retrograde reflux of the inoculum from the bladder into the kidney (11).
The ability of E. faecalis 0852 to persist in bladders and kidneys over a 2-week period is shown in Fig. 1. Fifteen minutes after inoculation, bacterial titers were high in both the bladder and the kidney. E. faecalis did not persist in the bladder, as bacterial titers decreased dramatically after 15 min, and many bladders were sterile at later time points. In contrast, recovery of bacteria from the kidneys remained steady over the first 12 h; 24 h after inoculation, bacterial levels in the kidney decreased but nevertheless persisted over a 2-week period. The conclusion from these studies was that E. faecalis persisted at higher titers in the kidney than in the bladder over a 2-week time frame (Fig. 1, P < 0.05 for five of nine time points, P < 0.1 for eight of nine time points).
The large dropoff in bladder titers between 15 min and 6 h was most likely due to the clearance of nonadherent bacteria from the bladder by mechanical forces of urine flow and other innate defenses. The persistence of E. faecalis in the kidneys over this time period indicated that the bacteria in the kidney were able to establish residence capable of evading innate defenses. Consistent with this hypothesis was the finding that 16% of mice had recoverable titers of 0852 in the kidneys despite having sterile bladders; conversely, no mice had recoverable bacterial titers in the bladder if their kidneys were sterile (Fig. 1B, squares). Given the variability of CFU per bladder, we hypothesized that bacteria recovered from the bladder at later time points may represent bacteria shed from the kidney. Taken together, the persistence of 0852 in the kidney and its clearance from the bladder suggest that E. faecalis has tropism for the kidneys in the urinary tract of C57BL/6J mice.
Multiple E. faecalis clinical isolates display tropism for the kidney in the urinary tract of C57BL/6J mice. To determine if the tropism of E. faecalis for the kidney in C57BL/6J mice is specific to strain 0852 or whether it represents a general feature of this species, we inoculated mice with five different strains of E. faecalis and harvested their bladders and kidneys for titers 48 h after infection (Fig. 2). Two of these strains, B1223 and B1384, were cystitis isolates, and three others, P1503, BP78, and BP250, were pyelonephritis isolates. Consistent with what had been observed with 0852, all of the strains persisted in the kidney to higher levels than in the bladder. Only three mice of the 29 in the group infected with cystitis-derived isolates had more bacteria in the bladder than in the kidney. While this finding may represent an adaptation of the cystitis isolates to have increased adherence to bladder epithelium, the kidney still appeared to have the larger bacterial burden in the majority of mice.
Transurethral inoculation of 0852 results in inflammation in the kidneys but not the bladder. Kidneys and bladders were subjected to histological analysis at various time points following inoculation with E. faecalis 0852. Hematoxylin and eosin staining revealed that there was little difference between infected and uninfected bladders at any of the time points examined (6 h, 24 h, 2 days, and 4 days; 24-h time point shown in Fig. 3). In contrast, E. coli induces cystitis, as measured by the disruption of the bladder epithelium, marked edema, and recruitment of numerous neutrophils at the same time point (4, 14, 19).
Kidney sections from E. faecalis-infected mice, however, showed an inflammatory infiltrate in the renal pelvis. The inflammation was most consistently evident at 24 h after inoculation but was also seen at other time points as small, isolated collections of inflammatory cells. The level of inflammation at 24 h in the kidney was variable, sometimes appearing quite extensive, with inflammatory cells lining the entire pelvis (Fig. 3b and 3e), but small, localized patches of inflammation were also observed along the pelvis (as in Fig. 3c and 3f). The cellular infiltrate in the kidneys is primarily monocytic, as determined by histologic features. This is in contrast to the mostly neutrophilic infiltrate seen in E. coli pyelonephritis (4). The results of the histologic analysis of the bladder and kidney demonstrate that E. faecalis can consistently cause pathology in the kidney but not the bladder, further confirming the tropism of E. faecalis for the mouse kidney.
Immunohistochemistry was also utilized to demonstrate the presence of enterococcal antigen within the infected kidney. Using the rabbit Lancefield group D antibody, we showed staining in association with areas of inflammation at 24 h postinfection, demonstrating that the inflammatory cells were recruited in response to the presence of E. faecalis rather than damage to the kidney parenchyma by the inoculation procedure (Fig. 4).
Inflammatory markers induced by uropathogenic E. coli are not upregulated in the bladder in response to E. faecalis. The interaction of uropathogenic E. coli with the bladder epithelium results in the expression of a variety of proteins involved in epithelial renewal and immune function (20). Among these, the proinflammatory marker Mip-2, the mouse orthologue to human interleukin-8, and Socs-3, a modulator of cytokine induction, were highly upregulated. We investigated whether E. faecalis induced a similar response in the urinary tracts of mice.
E. faecalis 0852- or sham-infected mouse bladders and kidneys were harvested 6 h or 24 h after inoculation, and RNA was collected in order to quantify the relative levels of Mip-2 and Socs-3 by quantitative reverse transcription-PCR. The relative induction of Mip-2 and Socs-3 after infection with NU14 (a clinical uropathogenic E. coli isolate) or E. faecalis 0852 was investigated (Fig. 5); 24 h after infection with either E. coli NU14 or E. faecalis 0852, NU14 induced Mip-2 and Socs-3 176-fold and 15-fold, respectively, over E. faecalis in the bladder. In the kidney, Mip-2 and Socs-3 were also more strongly induced by E. coli NU14 but not to the same magnitude (24-fold greater induction of Mip-2 and 4-fold greater induction of Socs-3 at 6 h, when the differences were most pronounced). Detection of focal inflammation in the kidney caused by E. faecalis may be diluted in this assay, which surveys gene expression levels of the whole kidney. Regardless, the observation that the uropathogenic E. coli strain NU14 was a more potent inducer of inflammatory mediators in both the bladder and kidneys implies a distinct pathogenic mechanism for E. faecalis in the urinary tract.
Innate response to E. faecalis in the kidneys is TLR2 independent. TLR4 has been demonstrated to be an important mediator of the innate host response to E. coli urinary tract infections (22, 23). TLR2 has been shown to mediate the host innate response to some gram-positive constituents such as lipoteichoic acid and peptidoglycan (24, 31). To determine if TLR2 plays a role in the initial response to E. faecalis, TLR2-deficient C57BL/6 were inoculated with 0852, and kidney titers were compared to wild-type mice 24 h after inoculation. There was no observable difference between the TLR2-deficient mice and the wild-type mice 24 h after infection, indicating that TLR2 did not play a critical role in the innate response (Fig. 6). Histologic examination of TLR2 and wild-type C57BL/6J mice revealed no observable difference in the levels of inflammation between wild-type and mutant mice (data not shown).
Clinically derived E. faecalis has a survival advantage over OG1X in the kidneys of mice. To assess the utility of mouse kidney infections as a model system to study pathogenesis of E. faecalis, the clinical strain BP78 and the laboratory strain OG1X were inoculated either separately or in a mixed suspension. As shown in Fig. 7a, OG1X and BP78 colonized the bladders and kidneys of mice equally well when infected separately, as determined 48 h after infection. However, in the mixed-infection model, BP78 showed a competitive advantage over OG1X of approximately 10-fold in the kidney, but there was no discernible difference in the bladder (Fig. 7b).
DISCUSSION
Enterococci are usually perceived as commensal bacteria that coexist with their host under most circumstances as part of the normal flora. While not regarded as particularly virulent organisms, enterococci can be significant agents of urinary tract infections in the hospital setting, where patients' defenses can be compromised by catheterization, immune deficiencies, or both. Given that E. faecalis is an opportunistic pathogen, animal model systems to study ascending urinary tract infections have been difficult to develop but will be essential for understanding the molecular basis of enterococcal disease.
Two recent studies used an ascending model of urinary tract infections similar to that used for uropathogenic E. coli to investigate the function of the Esp and aggregation substance adhesins in the pathogenesis of E. faecalis urinary tract infection. A role for aggregation substance has not yet been elucidated, whereas mutations in esp reduced colonization of the bladder. An interesting caveat in these studies was that approximately 30% of the mice had sterile bladders even when challenged with wild-type E. faecalis (12, 26). These results highlight the need to elucidate the host-pathogen interactions required to maintain robust urinary tract infections with enterococci. Interestingly, both studies noted significant colonization of the kidney tissue despite using a volume of inoculum designed to deliver bacteria only to the bladder.
In the present study, we discovered that E. faecalis has tropism for the kidneys in female C57BL/6J mice. This finding was made using a model where the inoculation volume of E. faecalis was increased in order to facilitate direct delivery of the bacteria to both the bladder and the kidneys simultaneously. Using this technique, we were able to recover bacteria from >90% of kidneys in mice infected with E. faecalis. In the same mice, the bladders were frequently sterile even though the kidneys were colonized. The molecular basis of this tropism for the kidney is unknown. Preferential adherence and/or invasion of E. faecalis to kidney versus bladder epithelium is a potential explanation, but host factors involved in innate defenses may also play a role.
The use of cultured bladder cells to study enterococcal pathogenesis in the urinary tract has limitations. Superficial umbrella cells lining the lumen of the bladder have specialized properties, including an ability to assemble uroplakins on their surface. Uroplakins form impermeable plaques that coat the luminal surface of the bladder (15). Cultured epithelial cells are not terminally differentiated, do not assemble uroplakins into plaques, and lack other distinguishing features of superficial umbrella cells, such as their large surface area and pentagonal shape (27). Thus, tropism is best studied in an animal model such as the one described here.
E. faecalis-mediated urinary tract infection appears to occur via a strikingly different mechanism than E. coli-mediated urinary tract infection. In a typical E. coli urinary tract infection, the bacteria invade the superficial umbrella cells and replicate to high levels, forming intracellular biofilms, a process that induces a TLR4-mediated cytokine response that recruits neutrophils to the site of infection (1, 22). In enterococcal urinary tract infection, there is little to no inflammation in the bladder. Furthermore, neutrophils represented only a minority of the inflammatory cells targeted to the site of infection, which typically is in the kidney. We found that TLR2, which is hypothesized to play a similar role in the innate response to gram-positives as TLR4 does for gram-negative pathogens (28, 29), does not play a substantial role in the innate response to E. faecalis in the urinary tract. Several observations reinforce this hypothesis. The presence or absence of TLR2 had no significant impact on the recovery of E. faecalis 24 h after infection. A luciferase reporter cell line expressing TLR2 was not stimulated by enterococci (32). Finally, there is substantial evidence that the host uses TLR2-independent pathways to stimulate inflammatory cascades in response to gram-positive infection (2, 5, 17). The recent discovery of TLR11, a Toll-like receptor that specifically recognizes uropathogens, suggested an additional innate mechanism used by the host to respond to enterococci (32).
All six E. faecalis strains tested in this study demonstrated the same tropism for the kidney, suggesting that most E. faecalis strains contain the core set of virulence factors necessary to cause disease in the kidney. When inoculated separately, OG1X, a commonly used laboratory strain of E. faecalis that lacks the recently described enterococcal pathogenicity island (25) and plasmid elements, and BP78, a clinical pyelonephritis isolate, persisted at similar levels over a 48-h period. This implies that OG1X has the essential molecular elements necessary to persist in the kidney. However, in a mixed-competition experiment, BP78 had a survival advantage of approximately 10-fold, suggesting that uropathogenic E. faecalis isolates have additional factors that lead to increased fitness in the urinary tract. Thus, mixed-infection experiments proved to be the most sensitive readout of additional virulence factors that enhance persistence in the kidney.
This work raises numerous questions related to human disease. The ability of enterococci to cause disease in the kidney is well established. The diagnosis of upper versus lower urinary tract infection caused by enterococci, as with gram-negative bacilli, is generally based on signs and symptoms. Thus, the presence of fever and flank pain, with or without lower tract symptoms of dysuria, frequency, and urgency suggest upper tract infection (13). Lower tract symptoms without upper tract symptoms are assumed to represent bladder infection, although it has been reported that about 30% of women with symptoms of bladder infection have silent upper tract infection (13). It will be important to determine whether the findings described in this work can be extended to humans. If enterococci have a strong tropism for the kidney, it is possible that many or most episodes of enterococcal cystitis are due to seeding from an upper tract infection. It is also possible that asymptomatic bacteriuria with enterococci is often localized to the upper tract rather than the bladder. The observation that enterococci fail to elicit a strong induction of inflammatory cytokines in our murine model is consistent with this hypothesis. If enterococcal urinary tract infections are often associated with upper tract infection, there may be treatment implications, since upper tract infection is usually treated for a longer duration than bladder infections (6).
ACKNOWLEDGMENTS
We thank members of the Hultgren laboratory for reading the manuscript, T. Hooten and W. Stamm for providing clinical strains, and S. Akira for TLR2-deficient mice.
This work was supported by National Institutes of Health and Office of Research on Women's Health Specialized Center of Research P50DK6454002 with the Food and Drug Administration (to S.J.H.), NIH grant AI-38273 (M.G.C.), Individual National Research Service Award F32DK06388702 (to S.M.), Pediatric Physician-Scientist Training Grant 5T32HD4301003 (E.H.), and Medical Scientist Training Program grant T32 GM07200 (A.K.).
REFERENCES
1. Anderson, G. G., J. J. Palermo, J. D. Schilling, R. Roth, J. Heuser, and S. J. Hultgren. 2003. Intracellular bacterial biofilm-like pods in urinary tract infections. Science 301:105-107.
2. Edelson, B. T., and E. R. Unanue. 2002. MyD88-dependent but Toll-like receptor 2-independent innate immunity to Listeria: no role for either in macrophage listericidal activity. J. Immunol. 169:3869-3875.
3. Freter, R., P. C. O'Brien, and M. S. Macsai. 1981. Role of chemotaxis in the association of motile bacteria with intestinal mucosa: in vivo studies. Infect. Immun. 34:234-240.
4. Haraoka, M., L. Hang, B. Frendeus, G. Godaly, M. Burdick, R. Strieter, and C. Svanborg. 1999. Neutrophil recruitment and resistance to urinary tract infection. J. Infect. Dis. 180:1220-1229.
5. Henneke, P., O. Takeuchi, J. A. van Strijp, H. K. Guttormsen, J. A. Smith, A. B. Schromm, T. A. Espevik, S. Akira, V. Nizet, D. L. Kasper, and D. T. Golenbock. 2001. Novel engagement of CD14 and multiple toll-like receptors by group B streptococci. J. Immunol. 167:7069-7076.
6. Hooton, T. M., and W. E. Stamm. 1997. Diagnosis and treatment of uncomplicated urinary tract infection. Infect. Dis. Clin. North Am. 11:551-581.
7. Hultgren, S. J., W. R. Schwan, A. J. Schaeffer, and J. L. Duncan. 1986. Regulation of production of type 1 pili among urinary tract isolates of Escherichia coli. Infect. Immun. 54:613-620.
8. Huycke, M. M., D. F. Sahm, and M. S. Gilmore. 1998. Multiple-drug resistant enterococci: the nature of the problem and an agenda for the future. Emerg. Infect. Dis. 4:239-249.
9. Ike, Y., D. B. Clewell, R. A. Segarra, and M. S. Gilmore. 1990. Genetic analysis of the pAD1 hemolysin/bacteriocin determinant in Enterococcus faecalis: Tn917 insertional mutagenesis and cloning. J. Bacteriol. 172:155-163.
10. Ike, Y., R. A. Craig, B. A. White, Y. Yagi, and D. B. Clewell. 1983. Modification of Streptococcus faecalis sex pheromones after acquisition of plasmid DNA. Proc. Natl. Acad. Sci. USA 80:5369-5373.
11. Johnson, J. R., and J. J. Brown. 1996. Defining inoculation conditions for the mouse model of ascending urinary tract infection that avoid immediate vesicoureteral reflux yet produce renal and bladder infection. J. Infect. Dis. 173:746-749.
12. Johnson, J. R., C. Clabots, H. Hirt, C. Waters, and G. Dunny. 2004. Enterococcal aggregation substance and binding substance are not major contributors to urinary tract colonization by Enterococcus faecalis in a mouse model of ascending unobstructed urinary tract infection. Infect. Immun. 72:2445-2448.
13. Johnson, J. R., and W. E. Stamm. 1989. Urinary tract infections in women: diagnosis and treatment. Ann. Intern. Med. 111:906-917.
14. Justice, S. S., C. Hung, J. A. Theriot, D. A. Fletcher, G. G. Anderson, M. J. Footer, and S. J. Hultgren. 2004. Differentiation and developmental pathways of uropathogenic Escherichia coli in urinary tract pathogenesis. Proc. Natl. Acad. Sci. USA 101:1333-1338.
15. Kachar, B., F. Liang, U. Lins, M. Ding, X. R. Wu, D. Stoffler, U. Aebi, and T. T. Sun. 1999. Three-dimensional analysis of the 16 nm urothelial plaque particle: luminal surface exposure, preferential head-to-head interaction, and hinge formation. J. Mol. Biol. 285:595-608.
16. Langermann, S., S. Palaszynski, M. Barnhart, G. Auguste, J. S. Pinkner, J. Burlein, P. Barren, S. Koenig, S. Leath, C. H. Jones, and S. J. Hultgren. 1997. Prevention of mucosal Escherichia coli infection by FimH-adhesin-based systemic vaccination. Science 276:607-611.
17. Malley, R., P. Henneke, S. C. Morse, M. J. Cieslewicz, M. Lipsitch, C. M. Thompson, E. Kurt-Jones, J. C. Paton, M. R. Wessels, and D. T. Golenbock. 2003. Recognition of pneumolysin by Toll-like receptor 4 confers resistance to pneumococcal infection. Proc. Natl. Acad. Sci. USA 100:1966-1971.
18. Montgomerie, J. Z., G. M. Kalmanson, and L. B. Guze. 1977. Virulence of enterococci in experimental pyelonephritis. Urol. Res. 5:99-102.
19. Mulvey, M. A., Y. S. Lopez-Boado, C. L. Wilson, R. Roth, W. C. Parks, J. Heuser, and S. J. Hultgren. 1998. Induction and evasion of host defenses by type 1-piliated uropathogenic Escherichia coli. Science 282:1494-1497.
20. Mysorekar, I. U., M. A. Mulvey, S. J. Hultgren, and J. I. Gordon. 2002. Molecular regulation of urothelial renewal and host defenses during infection with uropathogenic Escherichia coli. J. Biol. Chem. 277:7412-7419.
21. Roberts, J. A., B. I. Marklund, D. Ilver, D. Haslam, M. B. Kaack, G. Baskin, M. Louis, R. Mollby, J. Winberg, and S. Normark. 1994. The Gal(alpha 1-4)Gal-specific tip adhesin of Escherichia coli P-fimbriae is needed for pyelonephritis to occur in the normal urinary tract. Proc. Natl. Acad. Sci. USA 91:11889-11893.
22. Schilling, J. D., S. M. Martin, C. S. Hung, R. G. Lorenz, and S. J. Hultgren. 2003. Toll-like receptor 4 on stromal and hematopoietic cells mediates innate resistance to uropathogenic Escherichia coli. Proc. Natl. Acad. Sci. USA 100:4203-4208.
23. Schilling, J. D., M. A. Mulvey, C. D. Vincent, R. G. Lorenz, and S. J. Hultgren. 2001. Bacterial invasion augments epithelial cytokine responses to Escherichia coli through a lipopolysaccharide-dependent mechanism. J. Immunol. 166:1148-1155.
24. Schwandner, R., R. Dziarski, H. Wesche, M. Rothe, and C. J. Kirschning. 1999. Peptidoglycan- and lipoteichoic acid-induced cell activation is mediated by Toll-like receptor 2. J. Biol. Chem. 274:17406-17409.
25. Shankar, N., A. S. Baghdayan, and M. S. Gilmore. 2002. Modulation of virulence within a pathogenicity island in vancomycin-resistant Enterococcus faecalis. Nature 417:746-750.
26. Shankar, N., C. V. Lockatell, A. S. Baghdayan, C. Drachenberg, M. S. Gilmore, and D. E. Johnson. 2001. Role of Enterococcus faecalis surface protein Esp in the pathogenesis of ascending urinary tract infection. Infect. Immun. 69:4366-4372.
27. Surya, B., J. Yu, M. Manabe, and T. T. Sun. 1990. Assessing the differentiation state of cultured bovine urothelial cells: elevated synthesis of stratification-related K5 and K6 keratins and persistent expression of uroplakin I. J. Cell Sci. 97:419-432.
28. Takeuchi, O., K. Hoshino, and S. Akira. 2000. Cutting edge: TLR2-deficient and MyD88-deficient mice are highly susceptible to Staphylococcus aureus infection. J. Immunol. 165:5392-5396.
29. Takeuchi, O., K. Hoshino, T. Kawai, H. Sanjo, H. Takada, T. Ogawa, K. Takeda, and S. Akira. 1999. Differential roles of TLR2 and TLR4 in recognition of gram-negative and gram-positive bacterial cell wall components. Immunity 11:443-451.
30. Tsuchimori, N., R. Hayashi, A. Shino, T. Yamazaki, and K. Okonogi. 1994. Enterococcus faecalis aggravates pyelonephritis caused by Pseudomonas aeruginosa in experimental ascending mixed urinary tract infection in mice. Infect. Immun. 62:4534-4541.
31. Yoshimura, A., E. Lien, R. R. Ingalls, E. Tuomanen, R. Dziarski, and D. Golenbock. 1999. Cutting edge: recognition of gram-positive bacterial cell wall components by the innate immune system occurs via Toll-like receptor 2. J. Immunol. 163:1-5.
32. Zhang, D., G. Zhang, M. S. Hayden, M. B. Greenblatt, C. Bussey, R. A. Flavell, and S. Ghosh. 2004. A Toll-like receptor that prevents infection by uropathogenic bacteria. Science 303:1522-1526.(Andrew L. Kau, Steven M. )
ABSTRACT
Enterococcus faecalis is a gram-positive bacterium that can cause a variety of nosocomial infections of which urinary tract infections are the most common. These infections can be exceptionally difficult to treat because of drug resistance of many E. faecalis isolates. Despite their troublesome nature, little is known about the host or bacterial factors necessary for E. faecalis to cause disease in the urinary tract. Using a mouse model of urinary tract infection, we have shown that E. faecalis is capable of persisting in the kidneys of mice for at least 2 weeks. In contrast, bacterial titers from the bladders of the same mice were inconsistent and tended to be much lower than those recovered from the kidney. This preference for the kidney over the bladder is also observed in other clinical E. faecalis strains. Histologic examination of bladder and kidney tissues demonstrated that E. faecalis induced an inflammatory response in the kidney but not in the bladder. This inflammatory response was TLR2 independent and did not induce inflammatory markers typically associated with uropathogenic Escherichia coli. Using a competition assay, we demonstrated that a pyelonephritis clinical isolate had a growth advantage over a laboratory strain of E. faecalis in the kidneys but not in the bladders of mice. Taken together, these results demonstrate that E. faecalis has tropism for the kidneys in the urinary tracts of mice and that this system can be used to study factors involved in the pathogenesis of urinary tract infections.
INTRODUCTION
Enterococcus faecalis, while normally a gut commensal, is a frequent cause of many serious human infections, including urinary tract infections, endocarditis, bacteremia, and wound infections. Among the diseases that E. faecalis causes, urinary tract infections are the most common, responsible for approximately 110,000 cases yearly, many of which are nosocomial. Infections with E. faecalis can be especially troublesome to treat because of their frequent resistance to multiple antibiotics, including vancomycin, a drug of last resort for many gram-positive infections (for a review, see reference 8).
Both rat and mouse model systems have been used to study factors involved in the pathogenesis of E. faecalis in the urinary tract. Studies by Guze and colleagues showed that E. faecalis has a growth advantage over other enterococcal species in rat kidneys in a hematogenous inoculation pyelonephritis model (18). Another model system was used to show that pyelonephritis caused by Pseudomonas aeruginosa was aggravated by coinfection with E. faecalis, as determined by histological changes in the kidney (30). In this model, ligation of the urethra was used to induce retrograde reflux of bacteria into the kidney, increasing the susceptibility of the mouse to infection. In a bladder catheterization model of urinary tract infection, the Esp (enterococcal surface protein) adhesin was found to increase persistence in the urinary bladder of mice, although no histological changes were observed (26). Another study was unable to demonstrate a critical function for aggregation substance, a well-characterized enterococcal adhesin, in a murine model of cystitis (12). Thus, a robust murine cystitis model of E. faecalis infection has yet to be established, most likely due to complicating host and bacterial factors reflecting the adaptations enterococcus has evolved to coexist with its host. In order to more effectively devise strategies and therapies to prevent and treat enterococcal urinary tract infections, an infection model that can give a sensitive readout of virulence factors must be developed.
In contrast, many of the host-pathogen interactions critical for bladder infections caused by uropathogenic Escherichia coli, the most common etiologic agent of urinary tract infections, have been well-defined in murine models. Uropathogenic E. coli initially interacts with bladder epithelium through the expression of type 1 pili, which mediate adhesion and invasion into the uroplakin-expressing umbrella cells on the surface of the bladder (19). The invasion of E. coli into the urinary epithelium results in expression of a variety of cytokines, induction of apoptosis in infected cells, and a massive neutrophil inflammatory response. This response is mediated primarily by Toll-like receptor 4, which has been demonstrated to respond to lipopolysaccharide (22, 23). The interaction of E. coli with the mouse kidney epithelium during pyelonephritis is less well characterized, but the importance of P pili is well established as an initial mediator of adherence in other animal models (21). Similar to bladder infections, acute E. coli-mediated pyelonephritis in mice is characterized by a strong neutrophil immune response (4).
In this study, we describe the tropism of E. faecalis for the kidneys in C57BL/6J mice. The consequences of this tropism were investigated by characterizing the host response to E. faecalis and contrasting these differences to the response induced by uropathogenic E. coli.
MATERIALS AND METHODS
Strains and growth conditions. A summary of all the strains used in this study, their origins, and relevant drug resistances are shown in Table 1. For infection of mice, E. faecalis strains were grown statically overnight (typically 12 to 15 h) at 37°C in brain heart infusion (BHI) medium (Difco) without antibiotics. Clinical strains other than 0852 were contributed by Thomas Hooten and Walter Stamm (University of Washington) and categorized as either cystitis or pyelonephritis strains using clinical criteria.
Animals. Female wild-type C57BL/6 mice were obtained from Jackson Laboratories and C57BL/6 TLR2–/– mice were a gift of S. Akira (29). TLR2–/– mice were 36 to 48 weeks in age at the time of inoculation. Wild-type 12-week-old and 52-week-old mice were used to control for the various ages of the TLR2–/– mice. These experiments demonstrated that there was no substantial difference in the recoveries of E. faecalis due to the ages of the mice. In all other experiments, mice were 8 to 12 weeks of age.
Inoculation and CFU enumeration. Cultures were collected by centrifugation at 6,000 x g for 10 min and resuspended in phosphate-buffered saline (PBS) to an approximate density of 107 to 108 CFU/ml. Female C57BL/6J mice were anesthetized with inhaled isoflurane and then inoculated transurethrally with 200 μl of bacterial suspension similar to methods described in Schilling et al. (23). Although the inoculation volume of 200 μl resulted in leakage during and after inoculation, we found that this volume resulted in the most consistent infections.
At the appropriate time points, mice were sacrificed by cervical dislocation after inhalation anesthesia and the bladders and kidneys were harvested. For histologic evaluation, tissue was fixed in neutral buffered formalin and embedded in paraffin. To determine the number of bacteria present in these tissues, bladders or pairs of kidneys were homogenized in 0.025% Triton X-100 in PBS and plated at different dilutions on BHI agar (BactoAgar from BD) supplemented with antibiotics where appropriate. CFU were enumerated after 24 h of incubation at 37°C. Statistical analysis of the kinetic experiment was performed using Student's one-tailed t test for correlated samples using the logarithmic values at each time point. When the recovered titers were below the limit of detection, the recovered CFU value was set to 1 for statistical analysis.
Immunohistochemistry. For immunohistochemical analysis of bladder tissue, 5-μm-thick sections were prepared. Sections were deparaffinized using Hemo De (Fisher) (twice for 10 min), rinsed with isopropanol (three times for 3 min) and washed with PBS (three times for 5 min). Tissue sections were blocked with 1% bovine serum albumin, 0.2% nonfat dry milk, and 0.3% Triton X-100 in PBS (PBS-BB) for 15 min at room temperature. Primary antibody raised in rabbit against Streptococcus group D antigen (Lee Laboratories) was added in PBS-BB and incubated overnight at 4°C. After PBS washings (three times for 5 min), tissue was incubated with Alexafluor 555-labeled donkey anti-rabbit antibody (Molecular Probes) in PBS-BB and incubated overnight at 4°C. After PBS washes (three times for 5 min), tissue was counterstained with bis-benzimide (Sigma) to reveal nuclear morphology.
Quantitative real-time PCR. Methods were as described in Mysorekar et al. (20) with the following modifications. Female C57BL/6J mice were inoculated transurethrally with 200 μl of PBS, E. faecalis 0852, or uropathogenic E. coli isolate NU14 grown statically in Luria-Bertani (LB) for 48 h to induce type 1 pilus expression. At the time of sacrifice, kidneys were divided lengthwise, and one half was harvested for RNA isolation and the other half was titered to confirm the presence of bacteria. RNA from infected bladders and kidneys was collected either 6 h or 24 h after infection individually from each mouse using a commercially available affinity matrix-based kit (RNeasy kits, Qiagen). cDNA was generated using random hexamers and quantitative real-time PCR was performed as described (20) using the Bio-Rad iCycler. Primers for amplification of glyceraldehhyde-3-phosphate dehydrogenase, Mip-2, and Socs-3 are as previously described (20).
Competitive infection. Cultures of OG1X and BP78 were resuspended in PBS and mixed in equal volumes to make the mixed inoculum. The BP78 to OG1X ratio was determined by plating the mixed inoculum on medium selective for each strain; 48 h after infection, mice were sacrificed and their bladders and kidneys were collected and homogenized as described above. The homogenate was then plated onto either BHI agar supplemented with streptomycin, to select for OG1X, or tetracycline, to select for BP78. The competition index was calculated similarly to Freter et al. using OG1X as the reference strain (3). Briefly, the competition index = [(CFU of BP78/CFU of OG1X recovered from mice)/(CFU of BP78/CFU of OG1X present in initial inoculum)].
RESULTS
E. faecalis can cause a reproducible infection in C57BL/6J mouse kidneys. In order to study E. faecalis cystitis, we inoculated 50 μl of 108 CFU of an E. faecalis strain, 0852, from a diagnosed urinary tract infection transurethrally into female C57BL/6J mice. This protocol led to inconsistent recovery of bacteria from the bladder but noticeably higher titers in the kidney. Based on this observation, we used a 200-μl inoculum volume to intentionally induce retrograde reflux of the inoculum from the bladder into the kidney (11).
The ability of E. faecalis 0852 to persist in bladders and kidneys over a 2-week period is shown in Fig. 1. Fifteen minutes after inoculation, bacterial titers were high in both the bladder and the kidney. E. faecalis did not persist in the bladder, as bacterial titers decreased dramatically after 15 min, and many bladders were sterile at later time points. In contrast, recovery of bacteria from the kidneys remained steady over the first 12 h; 24 h after inoculation, bacterial levels in the kidney decreased but nevertheless persisted over a 2-week period. The conclusion from these studies was that E. faecalis persisted at higher titers in the kidney than in the bladder over a 2-week time frame (Fig. 1, P < 0.05 for five of nine time points, P < 0.1 for eight of nine time points).
The large dropoff in bladder titers between 15 min and 6 h was most likely due to the clearance of nonadherent bacteria from the bladder by mechanical forces of urine flow and other innate defenses. The persistence of E. faecalis in the kidneys over this time period indicated that the bacteria in the kidney were able to establish residence capable of evading innate defenses. Consistent with this hypothesis was the finding that 16% of mice had recoverable titers of 0852 in the kidneys despite having sterile bladders; conversely, no mice had recoverable bacterial titers in the bladder if their kidneys were sterile (Fig. 1B, squares). Given the variability of CFU per bladder, we hypothesized that bacteria recovered from the bladder at later time points may represent bacteria shed from the kidney. Taken together, the persistence of 0852 in the kidney and its clearance from the bladder suggest that E. faecalis has tropism for the kidneys in the urinary tract of C57BL/6J mice.
Multiple E. faecalis clinical isolates display tropism for the kidney in the urinary tract of C57BL/6J mice. To determine if the tropism of E. faecalis for the kidney in C57BL/6J mice is specific to strain 0852 or whether it represents a general feature of this species, we inoculated mice with five different strains of E. faecalis and harvested their bladders and kidneys for titers 48 h after infection (Fig. 2). Two of these strains, B1223 and B1384, were cystitis isolates, and three others, P1503, BP78, and BP250, were pyelonephritis isolates. Consistent with what had been observed with 0852, all of the strains persisted in the kidney to higher levels than in the bladder. Only three mice of the 29 in the group infected with cystitis-derived isolates had more bacteria in the bladder than in the kidney. While this finding may represent an adaptation of the cystitis isolates to have increased adherence to bladder epithelium, the kidney still appeared to have the larger bacterial burden in the majority of mice.
Transurethral inoculation of 0852 results in inflammation in the kidneys but not the bladder. Kidneys and bladders were subjected to histological analysis at various time points following inoculation with E. faecalis 0852. Hematoxylin and eosin staining revealed that there was little difference between infected and uninfected bladders at any of the time points examined (6 h, 24 h, 2 days, and 4 days; 24-h time point shown in Fig. 3). In contrast, E. coli induces cystitis, as measured by the disruption of the bladder epithelium, marked edema, and recruitment of numerous neutrophils at the same time point (4, 14, 19).
Kidney sections from E. faecalis-infected mice, however, showed an inflammatory infiltrate in the renal pelvis. The inflammation was most consistently evident at 24 h after inoculation but was also seen at other time points as small, isolated collections of inflammatory cells. The level of inflammation at 24 h in the kidney was variable, sometimes appearing quite extensive, with inflammatory cells lining the entire pelvis (Fig. 3b and 3e), but small, localized patches of inflammation were also observed along the pelvis (as in Fig. 3c and 3f). The cellular infiltrate in the kidneys is primarily monocytic, as determined by histologic features. This is in contrast to the mostly neutrophilic infiltrate seen in E. coli pyelonephritis (4). The results of the histologic analysis of the bladder and kidney demonstrate that E. faecalis can consistently cause pathology in the kidney but not the bladder, further confirming the tropism of E. faecalis for the mouse kidney.
Immunohistochemistry was also utilized to demonstrate the presence of enterococcal antigen within the infected kidney. Using the rabbit Lancefield group D antibody, we showed staining in association with areas of inflammation at 24 h postinfection, demonstrating that the inflammatory cells were recruited in response to the presence of E. faecalis rather than damage to the kidney parenchyma by the inoculation procedure (Fig. 4).
Inflammatory markers induced by uropathogenic E. coli are not upregulated in the bladder in response to E. faecalis. The interaction of uropathogenic E. coli with the bladder epithelium results in the expression of a variety of proteins involved in epithelial renewal and immune function (20). Among these, the proinflammatory marker Mip-2, the mouse orthologue to human interleukin-8, and Socs-3, a modulator of cytokine induction, were highly upregulated. We investigated whether E. faecalis induced a similar response in the urinary tracts of mice.
E. faecalis 0852- or sham-infected mouse bladders and kidneys were harvested 6 h or 24 h after inoculation, and RNA was collected in order to quantify the relative levels of Mip-2 and Socs-3 by quantitative reverse transcription-PCR. The relative induction of Mip-2 and Socs-3 after infection with NU14 (a clinical uropathogenic E. coli isolate) or E. faecalis 0852 was investigated (Fig. 5); 24 h after infection with either E. coli NU14 or E. faecalis 0852, NU14 induced Mip-2 and Socs-3 176-fold and 15-fold, respectively, over E. faecalis in the bladder. In the kidney, Mip-2 and Socs-3 were also more strongly induced by E. coli NU14 but not to the same magnitude (24-fold greater induction of Mip-2 and 4-fold greater induction of Socs-3 at 6 h, when the differences were most pronounced). Detection of focal inflammation in the kidney caused by E. faecalis may be diluted in this assay, which surveys gene expression levels of the whole kidney. Regardless, the observation that the uropathogenic E. coli strain NU14 was a more potent inducer of inflammatory mediators in both the bladder and kidneys implies a distinct pathogenic mechanism for E. faecalis in the urinary tract.
Innate response to E. faecalis in the kidneys is TLR2 independent. TLR4 has been demonstrated to be an important mediator of the innate host response to E. coli urinary tract infections (22, 23). TLR2 has been shown to mediate the host innate response to some gram-positive constituents such as lipoteichoic acid and peptidoglycan (24, 31). To determine if TLR2 plays a role in the initial response to E. faecalis, TLR2-deficient C57BL/6 were inoculated with 0852, and kidney titers were compared to wild-type mice 24 h after inoculation. There was no observable difference between the TLR2-deficient mice and the wild-type mice 24 h after infection, indicating that TLR2 did not play a critical role in the innate response (Fig. 6). Histologic examination of TLR2 and wild-type C57BL/6J mice revealed no observable difference in the levels of inflammation between wild-type and mutant mice (data not shown).
Clinically derived E. faecalis has a survival advantage over OG1X in the kidneys of mice. To assess the utility of mouse kidney infections as a model system to study pathogenesis of E. faecalis, the clinical strain BP78 and the laboratory strain OG1X were inoculated either separately or in a mixed suspension. As shown in Fig. 7a, OG1X and BP78 colonized the bladders and kidneys of mice equally well when infected separately, as determined 48 h after infection. However, in the mixed-infection model, BP78 showed a competitive advantage over OG1X of approximately 10-fold in the kidney, but there was no discernible difference in the bladder (Fig. 7b).
DISCUSSION
Enterococci are usually perceived as commensal bacteria that coexist with their host under most circumstances as part of the normal flora. While not regarded as particularly virulent organisms, enterococci can be significant agents of urinary tract infections in the hospital setting, where patients' defenses can be compromised by catheterization, immune deficiencies, or both. Given that E. faecalis is an opportunistic pathogen, animal model systems to study ascending urinary tract infections have been difficult to develop but will be essential for understanding the molecular basis of enterococcal disease.
Two recent studies used an ascending model of urinary tract infections similar to that used for uropathogenic E. coli to investigate the function of the Esp and aggregation substance adhesins in the pathogenesis of E. faecalis urinary tract infection. A role for aggregation substance has not yet been elucidated, whereas mutations in esp reduced colonization of the bladder. An interesting caveat in these studies was that approximately 30% of the mice had sterile bladders even when challenged with wild-type E. faecalis (12, 26). These results highlight the need to elucidate the host-pathogen interactions required to maintain robust urinary tract infections with enterococci. Interestingly, both studies noted significant colonization of the kidney tissue despite using a volume of inoculum designed to deliver bacteria only to the bladder.
In the present study, we discovered that E. faecalis has tropism for the kidneys in female C57BL/6J mice. This finding was made using a model where the inoculation volume of E. faecalis was increased in order to facilitate direct delivery of the bacteria to both the bladder and the kidneys simultaneously. Using this technique, we were able to recover bacteria from >90% of kidneys in mice infected with E. faecalis. In the same mice, the bladders were frequently sterile even though the kidneys were colonized. The molecular basis of this tropism for the kidney is unknown. Preferential adherence and/or invasion of E. faecalis to kidney versus bladder epithelium is a potential explanation, but host factors involved in innate defenses may also play a role.
The use of cultured bladder cells to study enterococcal pathogenesis in the urinary tract has limitations. Superficial umbrella cells lining the lumen of the bladder have specialized properties, including an ability to assemble uroplakins on their surface. Uroplakins form impermeable plaques that coat the luminal surface of the bladder (15). Cultured epithelial cells are not terminally differentiated, do not assemble uroplakins into plaques, and lack other distinguishing features of superficial umbrella cells, such as their large surface area and pentagonal shape (27). Thus, tropism is best studied in an animal model such as the one described here.
E. faecalis-mediated urinary tract infection appears to occur via a strikingly different mechanism than E. coli-mediated urinary tract infection. In a typical E. coli urinary tract infection, the bacteria invade the superficial umbrella cells and replicate to high levels, forming intracellular biofilms, a process that induces a TLR4-mediated cytokine response that recruits neutrophils to the site of infection (1, 22). In enterococcal urinary tract infection, there is little to no inflammation in the bladder. Furthermore, neutrophils represented only a minority of the inflammatory cells targeted to the site of infection, which typically is in the kidney. We found that TLR2, which is hypothesized to play a similar role in the innate response to gram-positives as TLR4 does for gram-negative pathogens (28, 29), does not play a substantial role in the innate response to E. faecalis in the urinary tract. Several observations reinforce this hypothesis. The presence or absence of TLR2 had no significant impact on the recovery of E. faecalis 24 h after infection. A luciferase reporter cell line expressing TLR2 was not stimulated by enterococci (32). Finally, there is substantial evidence that the host uses TLR2-independent pathways to stimulate inflammatory cascades in response to gram-positive infection (2, 5, 17). The recent discovery of TLR11, a Toll-like receptor that specifically recognizes uropathogens, suggested an additional innate mechanism used by the host to respond to enterococci (32).
All six E. faecalis strains tested in this study demonstrated the same tropism for the kidney, suggesting that most E. faecalis strains contain the core set of virulence factors necessary to cause disease in the kidney. When inoculated separately, OG1X, a commonly used laboratory strain of E. faecalis that lacks the recently described enterococcal pathogenicity island (25) and plasmid elements, and BP78, a clinical pyelonephritis isolate, persisted at similar levels over a 48-h period. This implies that OG1X has the essential molecular elements necessary to persist in the kidney. However, in a mixed-competition experiment, BP78 had a survival advantage of approximately 10-fold, suggesting that uropathogenic E. faecalis isolates have additional factors that lead to increased fitness in the urinary tract. Thus, mixed-infection experiments proved to be the most sensitive readout of additional virulence factors that enhance persistence in the kidney.
This work raises numerous questions related to human disease. The ability of enterococci to cause disease in the kidney is well established. The diagnosis of upper versus lower urinary tract infection caused by enterococci, as with gram-negative bacilli, is generally based on signs and symptoms. Thus, the presence of fever and flank pain, with or without lower tract symptoms of dysuria, frequency, and urgency suggest upper tract infection (13). Lower tract symptoms without upper tract symptoms are assumed to represent bladder infection, although it has been reported that about 30% of women with symptoms of bladder infection have silent upper tract infection (13). It will be important to determine whether the findings described in this work can be extended to humans. If enterococci have a strong tropism for the kidney, it is possible that many or most episodes of enterococcal cystitis are due to seeding from an upper tract infection. It is also possible that asymptomatic bacteriuria with enterococci is often localized to the upper tract rather than the bladder. The observation that enterococci fail to elicit a strong induction of inflammatory cytokines in our murine model is consistent with this hypothesis. If enterococcal urinary tract infections are often associated with upper tract infection, there may be treatment implications, since upper tract infection is usually treated for a longer duration than bladder infections (6).
ACKNOWLEDGMENTS
We thank members of the Hultgren laboratory for reading the manuscript, T. Hooten and W. Stamm for providing clinical strains, and S. Akira for TLR2-deficient mice.
This work was supported by National Institutes of Health and Office of Research on Women's Health Specialized Center of Research P50DK6454002 with the Food and Drug Administration (to S.J.H.), NIH grant AI-38273 (M.G.C.), Individual National Research Service Award F32DK06388702 (to S.M.), Pediatric Physician-Scientist Training Grant 5T32HD4301003 (E.H.), and Medical Scientist Training Program grant T32 GM07200 (A.K.).
REFERENCES
1. Anderson, G. G., J. J. Palermo, J. D. Schilling, R. Roth, J. Heuser, and S. J. Hultgren. 2003. Intracellular bacterial biofilm-like pods in urinary tract infections. Science 301:105-107.
2. Edelson, B. T., and E. R. Unanue. 2002. MyD88-dependent but Toll-like receptor 2-independent innate immunity to Listeria: no role for either in macrophage listericidal activity. J. Immunol. 169:3869-3875.
3. Freter, R., P. C. O'Brien, and M. S. Macsai. 1981. Role of chemotaxis in the association of motile bacteria with intestinal mucosa: in vivo studies. Infect. Immun. 34:234-240.
4. Haraoka, M., L. Hang, B. Frendeus, G. Godaly, M. Burdick, R. Strieter, and C. Svanborg. 1999. Neutrophil recruitment and resistance to urinary tract infection. J. Infect. Dis. 180:1220-1229.
5. Henneke, P., O. Takeuchi, J. A. van Strijp, H. K. Guttormsen, J. A. Smith, A. B. Schromm, T. A. Espevik, S. Akira, V. Nizet, D. L. Kasper, and D. T. Golenbock. 2001. Novel engagement of CD14 and multiple toll-like receptors by group B streptococci. J. Immunol. 167:7069-7076.
6. Hooton, T. M., and W. E. Stamm. 1997. Diagnosis and treatment of uncomplicated urinary tract infection. Infect. Dis. Clin. North Am. 11:551-581.
7. Hultgren, S. J., W. R. Schwan, A. J. Schaeffer, and J. L. Duncan. 1986. Regulation of production of type 1 pili among urinary tract isolates of Escherichia coli. Infect. Immun. 54:613-620.
8. Huycke, M. M., D. F. Sahm, and M. S. Gilmore. 1998. Multiple-drug resistant enterococci: the nature of the problem and an agenda for the future. Emerg. Infect. Dis. 4:239-249.
9. Ike, Y., D. B. Clewell, R. A. Segarra, and M. S. Gilmore. 1990. Genetic analysis of the pAD1 hemolysin/bacteriocin determinant in Enterococcus faecalis: Tn917 insertional mutagenesis and cloning. J. Bacteriol. 172:155-163.
10. Ike, Y., R. A. Craig, B. A. White, Y. Yagi, and D. B. Clewell. 1983. Modification of Streptococcus faecalis sex pheromones after acquisition of plasmid DNA. Proc. Natl. Acad. Sci. USA 80:5369-5373.
11. Johnson, J. R., and J. J. Brown. 1996. Defining inoculation conditions for the mouse model of ascending urinary tract infection that avoid immediate vesicoureteral reflux yet produce renal and bladder infection. J. Infect. Dis. 173:746-749.
12. Johnson, J. R., C. Clabots, H. Hirt, C. Waters, and G. Dunny. 2004. Enterococcal aggregation substance and binding substance are not major contributors to urinary tract colonization by Enterococcus faecalis in a mouse model of ascending unobstructed urinary tract infection. Infect. Immun. 72:2445-2448.
13. Johnson, J. R., and W. E. Stamm. 1989. Urinary tract infections in women: diagnosis and treatment. Ann. Intern. Med. 111:906-917.
14. Justice, S. S., C. Hung, J. A. Theriot, D. A. Fletcher, G. G. Anderson, M. J. Footer, and S. J. Hultgren. 2004. Differentiation and developmental pathways of uropathogenic Escherichia coli in urinary tract pathogenesis. Proc. Natl. Acad. Sci. USA 101:1333-1338.
15. Kachar, B., F. Liang, U. Lins, M. Ding, X. R. Wu, D. Stoffler, U. Aebi, and T. T. Sun. 1999. Three-dimensional analysis of the 16 nm urothelial plaque particle: luminal surface exposure, preferential head-to-head interaction, and hinge formation. J. Mol. Biol. 285:595-608.
16. Langermann, S., S. Palaszynski, M. Barnhart, G. Auguste, J. S. Pinkner, J. Burlein, P. Barren, S. Koenig, S. Leath, C. H. Jones, and S. J. Hultgren. 1997. Prevention of mucosal Escherichia coli infection by FimH-adhesin-based systemic vaccination. Science 276:607-611.
17. Malley, R., P. Henneke, S. C. Morse, M. J. Cieslewicz, M. Lipsitch, C. M. Thompson, E. Kurt-Jones, J. C. Paton, M. R. Wessels, and D. T. Golenbock. 2003. Recognition of pneumolysin by Toll-like receptor 4 confers resistance to pneumococcal infection. Proc. Natl. Acad. Sci. USA 100:1966-1971.
18. Montgomerie, J. Z., G. M. Kalmanson, and L. B. Guze. 1977. Virulence of enterococci in experimental pyelonephritis. Urol. Res. 5:99-102.
19. Mulvey, M. A., Y. S. Lopez-Boado, C. L. Wilson, R. Roth, W. C. Parks, J. Heuser, and S. J. Hultgren. 1998. Induction and evasion of host defenses by type 1-piliated uropathogenic Escherichia coli. Science 282:1494-1497.
20. Mysorekar, I. U., M. A. Mulvey, S. J. Hultgren, and J. I. Gordon. 2002. Molecular regulation of urothelial renewal and host defenses during infection with uropathogenic Escherichia coli. J. Biol. Chem. 277:7412-7419.
21. Roberts, J. A., B. I. Marklund, D. Ilver, D. Haslam, M. B. Kaack, G. Baskin, M. Louis, R. Mollby, J. Winberg, and S. Normark. 1994. The Gal(alpha 1-4)Gal-specific tip adhesin of Escherichia coli P-fimbriae is needed for pyelonephritis to occur in the normal urinary tract. Proc. Natl. Acad. Sci. USA 91:11889-11893.
22. Schilling, J. D., S. M. Martin, C. S. Hung, R. G. Lorenz, and S. J. Hultgren. 2003. Toll-like receptor 4 on stromal and hematopoietic cells mediates innate resistance to uropathogenic Escherichia coli. Proc. Natl. Acad. Sci. USA 100:4203-4208.
23. Schilling, J. D., M. A. Mulvey, C. D. Vincent, R. G. Lorenz, and S. J. Hultgren. 2001. Bacterial invasion augments epithelial cytokine responses to Escherichia coli through a lipopolysaccharide-dependent mechanism. J. Immunol. 166:1148-1155.
24. Schwandner, R., R. Dziarski, H. Wesche, M. Rothe, and C. J. Kirschning. 1999. Peptidoglycan- and lipoteichoic acid-induced cell activation is mediated by Toll-like receptor 2. J. Biol. Chem. 274:17406-17409.
25. Shankar, N., A. S. Baghdayan, and M. S. Gilmore. 2002. Modulation of virulence within a pathogenicity island in vancomycin-resistant Enterococcus faecalis. Nature 417:746-750.
26. Shankar, N., C. V. Lockatell, A. S. Baghdayan, C. Drachenberg, M. S. Gilmore, and D. E. Johnson. 2001. Role of Enterococcus faecalis surface protein Esp in the pathogenesis of ascending urinary tract infection. Infect. Immun. 69:4366-4372.
27. Surya, B., J. Yu, M. Manabe, and T. T. Sun. 1990. Assessing the differentiation state of cultured bovine urothelial cells: elevated synthesis of stratification-related K5 and K6 keratins and persistent expression of uroplakin I. J. Cell Sci. 97:419-432.
28. Takeuchi, O., K. Hoshino, and S. Akira. 2000. Cutting edge: TLR2-deficient and MyD88-deficient mice are highly susceptible to Staphylococcus aureus infection. J. Immunol. 165:5392-5396.
29. Takeuchi, O., K. Hoshino, T. Kawai, H. Sanjo, H. Takada, T. Ogawa, K. Takeda, and S. Akira. 1999. Differential roles of TLR2 and TLR4 in recognition of gram-negative and gram-positive bacterial cell wall components. Immunity 11:443-451.
30. Tsuchimori, N., R. Hayashi, A. Shino, T. Yamazaki, and K. Okonogi. 1994. Enterococcus faecalis aggravates pyelonephritis caused by Pseudomonas aeruginosa in experimental ascending mixed urinary tract infection in mice. Infect. Immun. 62:4534-4541.
31. Yoshimura, A., E. Lien, R. R. Ingalls, E. Tuomanen, R. Dziarski, and D. Golenbock. 1999. Cutting edge: recognition of gram-positive bacterial cell wall components by the innate immune system occurs via Toll-like receptor 2. J. Immunol. 163:1-5.
32. Zhang, D., G. Zhang, M. S. Hayden, M. B. Greenblatt, C. Bussey, R. A. Flavell, and S. Ghosh. 2004. A Toll-like receptor that prevents infection by uropathogenic bacteria. Science 303:1522-1526.(Andrew L. Kau, Steven M. )