Enterococci, normally commensal bacteria of the human oral cavity and gastrointestinal tract, have become important opportunistic pathogens in hospital settings. While not historically regarded as an important uropathogen in uncomplicated community-acquired cystitis, E. faecalis
is now considered a major agent of hospital-acquired UTIs where catheterization and indwelling medical devices can negatively impact patients' defenses against pathogens (20
). However, the lack of a robust animal model of enterococcal UTI has impeded the study of this pathogen within the urinary tract. In previously established models of ascending UTIs, E. faecalis
fails to establish persistent bladder colonization and primarily displays tropism to the kidneys (21
). The absence of persistent infections in these murine models reflects the profile of E. faecalis
in community-acquired UTIs where this bacterium represents less than 5% of clinical isolates (55
). As an opportunistic pathogen, E. faecalis
requires changes in bladder homeostasis following urinary catheterization in order to achieve successful infection of the urinary tract.
In this study, we investigated the pathophysiology of E. faecalis
-mediated CAUTIs using a murine model of foreign body-associated UTI that was modified from previous studies (28
). Our results indicate that the presence of silicone tubing in the murine bladder elicits histopathological and immunological changes that are similar to the effects seen in patients with spinal cord injuries and postsurgical patients with indwelling urinary catheters (8
). Furthermore, we show that E. faecalis
forms biofilms on the silicone implants and is able to persist at high titers in the bladders and kidneys of animals with implants.
The striking difference in bladder colonization and virulence from previous models is a direct consequence of implantation, which provides an abiotic surface for biofilm production. Biofilms have been implicated in the establishment of chronic infections, including CAUTIs (6
). Our data indicate that biofilm production on the surface of the implants is associated with persistent enterococcal cystitis. Bacteria within biofilms may continuously seed the bladder, preventing effective clearance by the host and leading to persistent cystitis.
) and GelE (52
) have been shown to be important in E. faecalis
biofilm formation. Atn has been shown to be a mediator of autolysis and DNA release, facilitating attachment to abiotic surfaces, thus promoting biofilm formation in vitro
). The extracellular protease GelE (1
) and the serine protease (SprE) mediate Atn-dependent DNA release during enterococcal biofilm formation (71
). GelE has been associated with virulence in several models of infection (19
). When introduced into mice with silicone implants to assess their contribution to CAUTI, both atn
- and gelE
-deficient mutants were as virulent as the wild-type strain. Thus, under these conditions, urinary tract colonization and biofilm formation on the implant did not require Atn or GelE. While these factors are important determinants of DNA-dependent biofilm production in vitro
, the ability of E. faecalis
to colonize the implants and promote infection in the absence of both Atn and GelE argues that extracellular DNA may not be an important contributor of the extracellular matrix of enterococcal biofilms in vivo
in the urinary tract or that its release may occur in an Atn/GelE-independent manner. Consistent with this finding was the observation that gelatinase was not expressed in 71/163 clinical isolates (42
) of E. faecalis
, and the presence of gelE
did not correlate with gelatinase expression (7
) or their ability to form biofilms in vitro
). However, a recent study by Arciola et al. (1
) found a correlation with a high level of expression of GelE and the ability to form biofilms in enterococci from orthopedic implant infections, indicating that the factors utilized by enterococci in forming biofilms may depend on the substrate and/or site of infection. Identification of the major components of the extracellular matrix of in vivo
enterococcal biofilms will lead to a better understanding of CAUTIs and will be of valuable interest for designing therapeutics that promote the prevention and disruption of these structures.
Sortases have also been implicated in enterococcal biofilm production (17
). Recent studies have shown that SrtA, a transpeptidase that anchors LPXTG-containing surface proteins to the cell walls of Gram-positive bacteria (40
), plays a major role during in vitro
biofilm development (17
). Deletion of srtA
in E. faecalis
prevented adherence and subsequent biofilm growth on abiotic surfaces (17
). Although Kemp et al. (30
) reported that disruption of srtA
did not significantly affect virulence in a murine model of ascending UTI, our data reveal that SrtA is critical in the establishment of CAUTI. The srtA
-deficient mutant is unable to colonize the implants in vivo
and is significantly attenuated in CAUTI. These findings further underscore the importance of biofilm production during E. faecalis
pathogenesis and strongly suggest that SrtA-dependent substrates may be required for attachment to abiotic surfaces and/or to damaged uroepithelium in vivo
during the establishment of infection. It will be of interest to identify SrtA-dependent substrates that are critical in CAUTI. For example, surface proteins known as MSCRAMMs (m
olecules) have been identified in E. faecium
and E. faecalis
). MSCRAMMs, some of which are putative SrtA-dependent substrates, can bind to host cells via interactions with host extracellular matrix proteins such as collagen and fibrinogen. The disruption of urothelial surfaces, as the result of catheterization, may expose factors that SrtA-dependent substrates, such as MSCRAMMs, can recognize to promote attachment and colonization.
Additionally, catheter-mediated abrasions to bladder epithelial cells have been previously shown to induce the production of lactate dehydrogenase and small amounts of IL-6 and IL-8 (2
). In this report, we also show that the presence of these implants elicits the expression of specific inflammatory markers in the bladders, including granulocyte colony-stimulating factor (G-CSF), keratinocyte-derived cytokine (KC) or CXCL1 (chemokine [C-X-C motif] ligand 1), and IL-6, whereas GM-CSF and MIP-1α (chemokine [C-C motif] ligand 3 [CCL3]) expression was decreased. All the above cytokines are produced by macrophages, neutrophils, and epithelial cells and are important in the recruitment of PMNs and other immune cells to sites of infection (3
). The presence of E. faecalis
in animals with implants specifically elicits the production of IL-1β and MIP-1α during CAUTI. These cytokines, produced by neutrophils and other immune cells, are important mediators of the inflammatory response to microbial infections (5
). In particular, IL-1β was shown to be released during UPEC-mediated acute cystitis (24
). The CAUTI model will be useful in unraveling the balance and interplay between inflammation and bacterial colonization and persistence.
The efficient bacterial clearance and the significantly lower titers recovered from the urinary tracts of mice without implants and/or with the loss of implants further corroborate the importance of a foreign body in bacterial persistence. Furthermore, our results indicate that the presence of a silicone implant in the bladder allows E. faecalis
to successfully ascend to the kidneys and establish persistent infection in animals with implants. The onset of acute pyelonephritis has also been reported in postmortem studies of the elderly with indwelling catheters at the time of death (79
). In the clinical setting, the removal of infected indwelling urinary catheters is one of the most effective methods used to resolve bacteriuria and CAUTIs (74
). However, removing the indwelling medical device, even combined with long course of antibiotic treatment (74
), may not be sufficient for complete resolution of the infection, especially with the rise in antibiotic resistance observed in nosocomial settings (11
). In addition, implant removal upon infection in other device-associated infections like prosthetic valve endocarditis is not in itself efficacious (10
) and thus may not be a suitable therapeutic approach in all instances. Understanding the pathogenesis of CAUTIs may lead to new and better ways to treat and prevent these diseases.
The optimized murine model of foreign body-associated UTI presented here is highly relevant for the investigation of the mechanisms underlying Enterococcus-mediated CAUTIs, since it couples biofilm production to enterococcal virulence during E. faecalis uropathogenesis. This murine model is a valuable and robust tool in the identification and characterization of novel biofilm determinants and bacterial virulence factors and host responses pertinent to the pathogenesis of E. faecalis in the urinary tract and the mechanisms underlying enterococcal ascension to the kidneys. Importantly, our optimized model will serve as an ideal platform for testing potential therapeutics against enterococcal infections and antibiofilm compounds.