Biofilm formation by clinical isolates and the implications in chronic infections

Abstract

Background: Biofilm formation is a major virulence factor contributing to the chronicity of infections. To date few studies have evaluated biofilm formation in infecting isolates of patients including both Gram-positive and Gram-negative multidrug-resistant (MDR) species in the context of numerous types of infectious syndromes. Herein, we investigated the biofilm forming capacity in a large collection of single patient infecting isolates and compared the relationship between biofilm formation to various strain characteristics.

Methods: The biofilm-forming capacity of 205 randomly sampled clinical isolates from patients, collected from various anatomical sites, admitted for treatment at Brooke Army Medical Center (BAMC) from 2004-2011, including methicillin-resistant/methicillin susceptible Staphylococcus aureus (MRSA/MSSA) (n=23), Acinetobacter baumannii (n=53), Pseudomonas aeruginosa (n=36), Klebsiella pneumoniae (n=54), and Escherichia coli (n=39), were evaluated for biofilm formation using the high-throughput microtiter plate assay and scanning electron microscopy (SEM). Relationships between biofilm formation to clonal type, site of isolate collection, and MDR phenotype were evaluated. Furthermore, in patients with relapsing infections, serial strains were assessed for their ability to form biofilms in vitro.

Results: Of the 205 clinical isolates tested, 126 strains (61.4%) were observed to form biofilms in vitro at levels greater than or equal to the Staphylococcus epidermidis, positive biofilm producing strain, with P. aeruginosa and S. aureus having the greatest number of biofilm producing strains. Biofilm formation was significantly associated with specific clonal types, the site of isolate collection, and strains positive for biofilm formation were more frequently observed to be MDR. In patients with relapsing infections, the majority of serial isolates recovered from these individuals were observed to be strong biofilm producers in vitro.

Conclusions: This study is the first to evaluate biofilm formation in a large collection of infecting clinical isolates representing diverse types of infections. Our results demonstrate: (1) biofilm formation is a heterogeneous property amongst clinical strains which is associated with certain clonal types, (2) biofilm forming strains are more frequently isolated from non-fluid tissues, in particular bone and soft tissues, (3) MDR pathogens are more often biofilm formers, and (4) strains from patients with persistent infections are positive for biofilm formation.

Figure 1

Biofilm formation by clinical bacterial isolates. Biofilm formation by individual clinical isolates of E. coli (n=39), P. aeruginosa (n=36), S. aureus (n=23), K. pneumoniae (n=54), and A. baumannii (n=53) using the 96-well microtiter plate assay. Biofilm formation was assessed by staining the attached bacteria with 0.1% CV and measuring the OD values at 570nm (CV₅₇₀) after 48 h growth at 37°C. Bars are representative of the average biofilm biomass from three independent experiments for each clinical isolate tested. Error bars indicate the standard error. Dashed line (−−-) indicates average biofilm biomass value (OD₅₇₀ = 0.122) for S. epidermidis ATCC 12228, the positive control for biofilm formation. Bars representing individuals strains are color coded to indicate site of isolation. Clinical isolates were ordered randomly with the exception of S. aureus which was separated by methicillin resistant (MRSA) and sensitive (MSSA) strains.

Figure 2

Scanning Electron Microscopy (SEM) images of biofilms. Representative SEM images of biofilms established on polystyrene pegs following 48 h incubation at 37°C from a selected biofilm producing strain of each bacterial species; including A) S. epidermidis ATCC 12228 (positive control), B) E. coli, C). P. aeruginosa,D) S. aureus (MRSA), E) S. aureus (MSSA), F) K. pneumoniae, and G) A. baumannii. SEM pictures were taken at 2000X magnification; inset white bar is representative of 10 microns.

Figure 3

Association between biofilm formation and pulsed-field type. Relationship of the biofilm-forming capacity of individual strains and pulsed-field type (PFT). Data points represent the mean biofilm biomass of individual isolates tested from each of the unique PFTs, as determined by the microtiter plate assay. Line bar represents the median biofilm biomass for each PFT group. Only pulsed-field types with >3 individual strains were used for the comparison. Other denotes those groups with <3 clinical isolates. Error bars represent the standard deviation among the results for different isolates. One-Way ANOVA analysis with Holm-Sidak comparison test was used to determine statistical differences between groups. Asterisks indicate those groups that were statistically significant to the majority of PFT; *p<0.05.

Figure 4

Biofilm formation is associated with the site of isolation. A) Distribution of biofilm-forming strains among fluid or non-fluid culture source. Bars are representative of % biofilm positive strains. Statistical analysis was performed using the X² test. B) Comparison of biofilm biomass (CV₅₇₀) from clinical isolates collected from various anatomical sites, including urine (n=27), blood (n=45), tissue deep (n=82), tissue superficial (n=31), bone (n=8) and respiratory sites (n=12), as determined by the microtiter plate assay. Data points represent the average biofilm biomass of individual isolates tested from each anatomical site and line bar represents the median biofilm biomass of the group. One-Way ANOVA analysis with Holm-Sidak comparison test was used to determine statistical differences between groups. Asterisks indicate those groups that were statistically significant; *p<0.05.