Antibody-Based Therapy for Enterococcal Catheter-Associated Urinary Tract Infections

Abstract

Gram-positive bacteria in the genus Enterococcus are a frequent cause of catheter-associated urinary tract infection (CAUTI), a disease whose treatment is increasingly challenged by multiantibiotic-resistant strains. We have recently shown that E. faecalis uses the Ebp pilus, a heteropolymeric surface fiber, to bind the host protein fibrinogen as a critical step in CAUTI pathogenesis. Fibrinogen is deposited on catheters due to catheter-induced inflammation and is recognized by the N-terminal domain of EbpA (EbpA^NTD), the Ebp pilus's adhesin. In a murine model, vaccination with EbpA^NTD confers significant protection against CAUTI. Here, we explored the mechanism of protection using passive transfer of immune sera to show that antisera blocking EbpA^NTD-fibrinogen interactions not only is prophylactic but also can act therapeutically to reduce bacterial titers of an existing infection. Analysis of 55 clinical CAUTI, bloodstream, and gastrointestinal isolates, including E. faecalis, E. faecium, and vancomycin-resistant enterococci (VRE), revealed a diversity of levels of EbpA expression and fibrinogen-binding efficiency in vitro Strikingly, analysis of 10 strains representative of fibrinogen-binding diversity demonstrated that, irrespective of EbpA levels, EbpA^NTD antibodies were universally protective. The results indicate that, despite diversity in levels of fibrinogen binding, strategies that target the disruption of EbpA^NTD-fibrinogen interactions have considerable promise for treatment of CAUTI.

Importance: Urinary catheterization is a routine medical procedure, and it has been estimated that 30 million Foley catheters are used annually in the United States. Importantly, placement of a urinary catheter renders the patient susceptible to developing a catheter-associated urinary tract infection, accounting for 1 million cases per year. Additionally, these infections can lead to serious complications, including bloodstream infection and death. Enterococcus strains are a common cause of these infections, and management of enterococcal infections has been more difficult in recent years due to the development of antibiotic resistance and the ability of strains to disseminate, resulting in a major threat in hospital settings. In this study, we developed an antibiotic-sparing treatment that is effective against diverse enterococcal isolates, including vancomycin-resistant enterococci, during catheter-associated urinary tract infections.

FIG 1

E. faecalis colocalized with Fg on human urinary catheters. Urinary catheters with an indwelling time of 18 h (A), 24 h (B and C), 8 days (D), or 9 days (E) were recovered from patients with an enterococcal UTI. The presence and distribution of bacteria and fibrinogen were assessed by immunofluorescence using antibody staining to detect fibrinogen (anti-Fg; green) and E. faecalis (anti-group D; red). As a negative control, a piece of the catheter was incubated with the secondary antibody only to assess background fluorescence.

FIG 2

Passive immunization with anti-EbpA^Full and anti-EbpA^NTD antibodies prevents E. faecalis CAUTI. Mice (n = 10) were given a dose of 100 µl of PBS sera or 100 µl of anti-EbpA^Full or anti-EbpA^NTD with a titer of 1 × 10⁷. (A) Experimental timeline. Sac, time of mouse sacrifice. (B and C) Detection of EbpA^Full and EbpA^NTD antibodies in bladder (B) and in urine (C) analyzed by diluting each sample 1:100 before serial dilution. (D and E) Prophylaxis treatment performed using 1 dose at 4 h prior infection (D) or 2 doses at 4 h prior infection and 12 h postinfection (hpi) (E). Values represent means ± SEM. The Mann-Whitney U test was used for mouse experiments; P < 0.05 was considered statistically significant. *, P < 0.05; **, P < 0.005; ***, P < 0.0005; ****, P < 0.00005; ns, values were not statistically significantly different. The horizontal bar represents the median value. The horizontal broken line represents the limit of detection (LOD) of viable bacteria. Animals that lost the catheter were not included in this work.

FIG 3

Release of anti-EbpA^NTD antibodies into the bladder and urine is mediated by bladder inflammation upon catheterization, and passive transfer of anti-EbpA^NTD antibodies reduced bacterial titers of an existing E. faecalis infection. (A) Experimental timeline. (B) Detection of anti-EbpA^NTD antibodies in urine samples, bladder homogenates, and serum samples was analyzed by diluting each sample 1:100 before serial dilution. (C) Experimental timeline. Mice (n = 10) were implanted with catheters and challenged with 1 × 10⁷ CFU of E. faecalis OG1RF. A dose of PBS serum, anti-EbpA^Full, anti-EbpA^NTD, anti-EbpA^CTD, or anti-group D antibody was administered intraperitoneally (i.p.) at 12 h postinfection (hpi) (anti-EbpA^Full, anti-EbpA^NTD, anti-EbpA^CTD, and anti-group D antibody titers of 1 × 10⁷). (D and E) Following 24 h of infection, bacterial burdens in bladder tissue (D) or recovered from catheters (E) were quantitated as the number of CFU recovered. Values represent means ± SEM. The Mann-Whitney U test was used; P < 0.05 was considered statistically significant. *, P < 0.05; **, P < 0.005; ***, P < 0.0005; ns, values were not statistically significantly different. The horizontal bar represents the median value. The horizontal broken line represents the limit of detection of viable bacteria. Animals that lost the catheter were not included in this work.

FIG 4

A prior enterococcal infection does not confer protection against a subsequent infection. (A and B) Mice were mock infected (n = 5) or infected with E. faecalis OG1RF (n = 15) (A); infection was followed by measuring the bacterial burden in urine (B). (C) After antibiotic treatment, mice were infected and bacterial burdens in bladder tissue or recovered catheters were quantitated as the number of CFU recovered. (D to G) Mice were immunized with mock serum (n = 5), serum from infected mice (n = 15 [D and E] or n = 10 [F and G]), anti-EbpA^NTD antibodies (n = 15 [D and E] or n = 5 [F and G]), or diluted anti-EbpA^NTD antibodies (n = 5). Following 24 h of infection, bladders (D and F) and catheters (E and G) were harvested and bacterial burdens were quantified. Values represent means ± SEM. The Mann-Whitney U test was used; P < 0.05 was considered statistically significant. *, P < 0.05; **, P < 0.005; ***, P < 0.0005; ns, values were not statistically significantly different. The horizontal bar represents the median value. The horizontal broken line represents the limit of detection of viable bacteria. Animals that lost the catheter were not included in this work.

FIG 5

Ebp pilus operon presence and analysis of EbpA conservation and expression and of its function in binding to Fg in clinical enterococcal strains. (A) Ebp operon scheme. (B) Alignment of 137 unique EbpA peptide and nucleotide sequences by using a hidden Markov model. vWA, von Willebrand factor A. (C) Pearson correlation statistical analysis was performed to measure the correlation between EbpA expression and Fg binding of each tested enterococcal strain (n = 3). Each dot represents the average of results from two independent experiments, each consisting of 3 biological replicates. As a negative control for the expression of EbpA, ΔEbp pilus and ΔEbpA strains were used, and as a negative control for Fg binding, the EbpA MIDAS mutant strain (AWAGA) was used. The star symbols indicate the strains that were selected for additional in vivo analyses.

FIG 6

EbpA^NTD-based vaccine and anti-EbpA^NTD antibody treatment prevented and reduced bacterial titers of different enterococcal isolates. (A) Mice were immunized and received two booster immunizations with doses of 100 µg EbpA^NTD. Four weeks following the final immunization, mice were implanted with catheters and challenged with 2 × 10⁷ CFU of E. faecalis OG1RF (n = 5) or with the indicated clinical enterococcal strains (n = 5). Following 24 h of infection, bacterial burdens in bladder tissue or recovered catheters were quantitated as the number of CFU recovered. org, organ. (B to E) To assess the therapeutic effect of anti-EbpA^NTD antibodies against enterococcus-infected mice, antibodies were administered intraperitoneally 12 hpi (anti-EbpA^NTD antibody titer of 1 × 10⁷). Enterococcal strains were divided into two groups: (i) those that highly expressed EbpA and bound to Fg (B and C; n = 5) and (ii) those that did not express EbpA (n = 10 [EC#15 and EC#43] and n = 15 [EC#49]) or bind to Fg (D and E; n = 15) under in vitro conditions. Bladder (B) and catheter (C) bacterial burdens were quantified as described below. Values represent means ± SEM. The Mann-Whitney U test was used; P < 0.05 was considered statistically significant. *, P < 0.05; **, P < 0.005; ***, P < 0.0005; ns, values were not statistically significantly different. The horizontal bar represents the median value. The horizontal broken line represents the limit of detection of viable bacteria. Animals that lost the catheter were not included in this work.