Intestinal tissues induce an SNP mutation in Pseudomonas aeruginosa that enhances its virulence: possible role in anastomotic leak

Abstract

The most feared complication following intestinal resection is anastomotic leakage. In high risk areas (esophagus/rectum) where neoadjuvant chemoradiation is used, the incidence of anastomotic leaks remains unacceptably high (≈ 10%) even when performed by specialist surgeons in high volume centers. The aims of this study were to test the hypothesis that anastomotic leakage develops when pathogens colonizing anastomotic sites become in vivo transformed to express a tissue destroying phenotype. We developed a novel model of anastomotic leak in which rats were exposed to pre-operative radiation as in cancer surgery, underwent distal colon resection and then were intestinally inoculated with Pseudomonas aeruginosa, a common colonizer of the radiated intestine. Results demonstrated that intestinal tissues exposed to preoperative radiation developed a significant incidence of anastomotic leak (>60%; p<0.01) when colonized by P. aeruginosa compared to radiated tissues alone (0%). Phenotype analysis comparing the original inoculating strain (MPAO1- termed P1) and the strain retrieved from leaking anastomotic tissues (termed P2) demonstrated that P2 was altered in pyocyanin production and displayed enhanced collagenase activity, high swarming motility, and a destructive phenotype against cultured intestinal epithelial cells (i.e. apoptosis, barrier function, cytolysis). Comparative genotype analysis between P1 and P2 revealed a single nucleotide polymorphism (SNP) mutation in the mexT gene that led to a stop codon resulting in a non-functional truncated protein. Replacement of the mutated mexT gene in P2 with mexT from the original parental strain P1 led to reversion of P2 to the P1 phenotype. No spontaneous transformation was detected during 20 passages in TSB media. Use of a novel virulence suppressing compound PEG/Pi prevented P. aeruginosa transformation to the tissue destructive phenotype and prevented anastomotic leak in rats. This work demonstrates that in vivo transformation of microbial pathogens to a tissue destroying phenotype may have important implications in the pathogenesis of anastomotic leak.

Conflict of interest statement

Competing Interests: JCA is co-founder of Midway Pharmaceuticals which develops high molecular weight polyethylene glycol polymers as epithelial cytoprotectants. JG is a PLoS ONE Editorial Board member. This does not alter the authors’ adherence to all the PLoS ONE policies on sharing data and materials.

Figures

Figure 1. Anastomotic leak in rats exposed to pre-operative radiation and intestinal P. aeruginosa.

(A) Sketch of anastomosis model and treatment groups. Treatment groups: I, anastomosis only; II, anastomosis + cecal injection of P. aeruginosa MPAO1, 107 CFU; III, radiation + anastomosis; IV, radiation +anastomosis + cecal injection of P. aeruginosa MPAO1 (107 CFU). Black arrows indicate the anastomotic site. (B) Excised and exposed suture lines of anastomotic sites. All suture lines are grossly intact except for group IV where ulceration/dehiscence is noted by the black arrow. (C) H&E staining of anastomotic tissues. Arrows and brackets indicate width of tissue apposition at suture line. (D, D’, E) Methylene blue assessment of anastomotic integrity demonstrating rare to no leaks in groups I–III (D) and gross extravasation in group IV (D’). Arrows indicate the site of anastomosis. (E) Incidence of anastomotic leak between groups. n = 12 (group I), n = 16 (group II), n = 9 (group III), n = 18 (group IV), *p<0.01.

Figure 2. Scanning electron microscopy (SEM) images of anastomosis tissues.

Blue arrows indicate healed anastomosis (30x), intact intestinal epithelium (300x), and macrophages on epithelial surface (3,500x) in group II. Orange arrows indicate discontinuity near or at the anastomosis (30x), disrupted intestinal epithelium (300x), and a high degree of bacterial colonization/adherence at the edge of non-healed anastomoses (5,000x). 50 images from each group of 5 mice were obtained, and representative images are displayed.

Figure 3. P1 and P2 phenotypes of P. aeruginosa MPAO1.

(A) Pyocyanin production seen as green color pigmentation on solid PIA and liquid TSB media. (B) Kaplan-Meyer survival curves of C. elegans N2 feeding on P1 and P2. Cumulative survival is represented of 2 experiments, n = 7/dish, 5 dishes/experiment, p<0.01. (C) Swarming motility. (D) Wound healing assay. Wound width was calibrated and measured using the MicroSuite software for imaging applications (Olympus SZX16). Wound healing of −100% indicates a 2 fold increase in the wound width compared to the baseline width. n = 12, *p<0.01. (E,F) Collagenase activity of P1 and P2 measured by degradation of fluorescent labeled collagen I (E) and collagen IV (F) as substrates. n = 6, *p<0.01. Fluorescence values were normalized to cell density measured by absorbance at 600 nm. Results are representative of 3 independent experiments. (G) RAPD fingerprint analysis demonstrating a similar genetic background of the P1 and P2 phenotype strains.

Figure 4. P2 induces significant apoptosis and structural changes in the tight junction protein ZO-1 in in IEC-18 monolayers.

(A) IEC-18 cells infected with P1 and P2 for 3 hrs were analyzed for apoptosis and necrosis with Apoptic&Necrotic&Healthy Cells Quantification kit (Biotium, Inc.) using fluorescence microscope Axiovert 35 (Zeiss,Germany). FITC-Annexin V (apoptotic cells, green), EtD-III (necrotic cells, red). (B) Staining of nuclei with Hoechst 33342. (C) IEC-18 monolayers treated with antibody to ZO-1.

Figure 5. Wound healing assay.

The P2 strain recovered from anastomotic tissues (in vivo) and the P2 strain recovered after co-incubation of MPAO1 with anastomotic tissues (ex vivo) similarly destroy wounded epithelial IEC-18 monolayers.

Figure 6. SNP mutation in MexT is responsible for P2 phenotype.

(A) Genome DNA sequence comparative map of P. aeruginosa MPAO1-P1 and MPAO1-P2 at the DSM-1707 backbone annotated with the MexT locus. Grey and teal bands: annotated coding regions; red tick: location of MexT locus; green ticks: tRNAs; black ticks: rRNAs; inner circle GC content. (B–D) swarming motility in (B) MPAO1-P1 (P1), (C) MPAO1-P2 (P2), and (D) MPAO1-P2 in which mexT was replaced by mexT gene amplified from MPAO1-P1 (P2/mexTP1). (E) Growth curves at 100 µg/ml chloramphenicol demonstrating acquisition of chloramphenicol resistance in P2/mexTP1. (F) Collagenase activity measured by fluorescence of fluorescent labeled gelatin as a substrate. n = 6, *p<0.01. Results are representative of 3 independent experiments.

Figure 7. Protective effect of PEG/Pi.

(A) Inhibition of wound apposition (healing) by P. aeruginosa MPAO1-P2. n = 5, *p<0.01. (B) Prevention of the cytotoxic effect of MPAO1-P2. n = 6, *p<0.01. (C) Frequency of transformation MPAO1 to P2 phenotype. n = 3, *p<0.01. (D) Frequency of anastomotic leak in rats. n = 10, *p<0.01. (E) SEM images of anastomotic tissues treated with 5% PEG/Pi demonstrating intact intestinal epithelium covered with a mucus-like film structure.