Model of Chronic Equine Endometritis Involving a Pseudomonas aeruginosa Biofilm

Abstract

Bacteria in a biofilm community have increased tolerance to antimicrobial therapy. To characterize the role of biofilms in equine endometritis, six mares were inoculated with lux-engineered Pseudomonas aeruginosa strains isolated from equine uterine infections. Following establishment of infection, the horses were euthanized and the endometrial surfaces were imaged for luminescence to localize adherent lux-labeled bacteria. Samples from the endometrium were collected for cytology, histopathology, carbohydrate analysis, and expression of inflammatory cytokine genes. Tissue-adherent bacteria were present in focal areas between endometrial folds (6/6 mares). The Pel exopolysaccharide (biofilm matrix component) and cyclic di-GMP (biofilm-regulatory molecule) were detected in 6/6 mares and 5/6 mares, respectively, from endometrial samples with tissue-adherent bacteria (P < 0.05). A greater incidence (P < 0.05) of Pel exopolysaccharide was present in samples fixed with Bouin's solution (18/18) than in buffered formalin (0/18), indicating that Bouin's solution is more appropriate for detecting bacteria adherent to the endometrium. There were no differences (P > 0.05) in the number of inflammatory cells in the endometrium between areas with and without tissue-adherent bacteria. Neutrophils were decreased (P < 0.05) in areas surrounding tissue-adherent bacteria compared to those in areas free of adherent bacteria. Gene expression of interleukin-10, an immune-modulatory cytokine, was significantly (P < 0.05) increased in areas of tissue-adherent bacteria compared to that in endometrium absent of biofilm. These findings indicate that P. aeruginosa produces a biofilm in the uterus and that the host immune response is modulated focally around areas with biofilm, but inflammation within the tissue is similar in areas with and without biofilm matrix. Future studies will focus on therapeutic options for elimination of bacterial biofilm in the equine uterus.

Keywords: bacteria; biofilm; cyclic di-GMP; endometritis; equine; exopolysaccharide.

FIG 1

(A) Gross pathology of the equine endometrial surface of a representative mare 5 days postinoculation with lux-labeled P. aeruginosa. The uterine lumen was filled with 50 to 100 ml of purulent fluid. (B) Bioluminescent imaging of the equine uterus at 5 days postinoculation with lux-labeled P. aeruginosa. The highly luminescent areas are correlated with a high bacterial load of lux-expressing bacteria.

FIG 2

Bioluminescent imaging of the equine uterus at 5 days postinoculation with lux-labeled P. aeruginosa. Luminescence of a strongly adherent matrix on the endometrial surface was detected following repeated washing of the endometrium. The luminescence was present in the base of the uterus and extending into the uterine horns, indicating the presence of tissue-adherent lux-labeled P. aeruginosa.

FIG 3

LC-MS/MS quantitative analysis of the bacterial secondary messenger molecule, cyclic di-GMP, from uterine infections to detect P. aeruginosa biofilms. Elevated cyclic di-GMP levels were detected in a majority of tissue-adherent bacterial samples, except for those from samples obtained from horse 5, which were not elevated compared to those of control samples. Four samples of intraluminal fluid and tissue-adherent bacteria were collected at random locations from each infected uterus (n = 6). Four control samples were collected by uterine biopsy procedure from two uninfected mares. Amounts are represented as picomoles of cyclic di-GMP per gram of sample. The calculated limit of detection (LOD) is 0.64 pmol, and the calculated limit of quantification (LOQ) is 2.14 pmol. Experimental samples range from 3.1 pmol/g to 2,033 pmol/g cyclic di-GMP. Samples from which no cyclic di-GMP was detected are represented under the LOD line.

FIG 4

Detection of tissue-adherent P. aeruginosa from the equine endometrium was dependent on the fixative. H&E-stained endometrial sections from a representative mare with tissue-adherent P. aeruginosa are shown. The inflammatory response (brackets) in the endometrium was a severe lymphocytic infiltrate. (A) Tissue was fixed in Bouin's solution, and the tissue-adherent P. aeruginosa was clearly evident (black arrow). (B) Tissue was collected from the same mare and same location as those described for panel A but was fixed in 10% formalin. Note the lack of tissue-adherent P. aeruginosa in panel B. The black arrow points to where the tissue-adherent P. aeruginosa should be located.

FIG 5

Detection of tissue-adherent P. aeruginosa in endometrium samples. H&E image of endometrium with tissue-adherent P. aeruginosa on the luminal surface (black arrow) (A) and deep in the endometrial glands (black arrow) (B). (C) Differential interference contrast image of an endometrial gland below the luminal surface of the uterus; this is similar to the area represented in panel B by the black arrow. Immunofluorescent staining of tissue-adherent P. aeruginosa with an anti-Pseudomonas antibody (Alexa Fluor 405) (D) and anti-Pel lectin (Texas red) (E) and merged image detecting the Pel exopolysaccharide colocalized with P. aeruginosa (F). Immunofluorescent images are projected images of Z-stacks as processed by Volocity image analysis software in which 0.5-μm scanning increments were performed through approximately 10 μm of tissue. The scale bar is 4 μm.

FIG 6

Fold change in gene expression of inflammatory cytokines in the endometrium preinoculation, postinoculation with tissue-adherent P. aeruginosa, and postinoculation free of tissue-adherent P. aeruginosa. A proinflammatory response was noted with upregulation of IL-6 and IL-1β. Endometrium with tissue-adherent P. aeruginosa had significantly greater change in gene expression of IL-10, an immune-modulatory cytokine, than endometrium free of bacteria. A difference in lowercase letter indicates a significant difference in gene expression (P < 0.05).