The collagenolytic activity of fecal Enterococcus faecalis predicts anastomotic leak: development of a novel assay to clinically track the collagenolytic phenotype
Author(s):
Kristina Guyton, University of Chicago; Baddr Shakhsheer, University of Chicago; Robin Klabbers, Radboud University Nijimegen Medical Centre; James Luo, University of Chicago; Monika Krezalek, University of Chicago; Lynn Hancock, University of Kansas; Olga Zaborina, University of Chicago; John Alverdy, University of Chicago
Background: Anastomotic leakage is a feared and poorly understood complication of intestinal surgery. We have recently demonstrated that collagenolytic properties of the Enterococcus faecalis gelatinase (GelE) play a key and causative role in anastomotic leak in rats.
Hypothesis: Here we hypothesize that the predominance of a collagenolytic phenotype in intestinal E. faecalis predicts anastomotic leak.
Methods: Adult rats (n=17) underwent distal colon resection, local devascularization and recto-sigmoid anastomosis and were sacrificed six days later and evaluated for evidence of anastomotic leakage (41% leak rate). Pre- and post-operative luminal contents were harvested for analysis. To count individual collagenolytic E. faecalis colonies, we developed a high-throughput method using esculin bile plates complemented with skim milk. Collagenolytic Colony Forming Unit (CCFU) was measured by enumerating enterococcal colonies surrounded by a zone of milk clearing normalized to dry stool weight. Stool CCFU/g was determined at the day of operation (POD0) and the day of sacrifice (POD6) and the difference reflects the in vivo change during anastomotic healing. The novel assay was validated by comparing with the degradation of fluorescent gelatin using wild-type rat intestinal strains, a clinical VRE isolate (V583), and its mutants deficient of, or complemented with, the gelE/sprE operon that expresses quorum-sensing dependent gelatinase activity.
Results: Enterococci hydrolyzed esculin producing a black pigment surrounding the colony and many also displayed areas of milk clearing reflecting casein hydrolysis. The area of milk clearing correlated with the level of gelatinase activity in E. faecalis rat intestinal isolates as measured by the fluorescent-labeled gelatin method (n=24, R2= 0.80, P<0.0001). Compared to the lab strain V583, no milk clearing was seen with the gelE deletion mutant. In contrast, over-expression of gelE from a multi-copy plasmid enhanced the zone of clearance. In stool, the lowest change in CCFU/g was observed in rats with minimal leak (∆log= 0.01) and the highest CCFU/g change was observed in rats with severe leaks (∆log= 0.92) (P<0.05). The change in concentration of collagenolytic enterococci in stool correlated with higher clinical leak rate (n=12, R2= 0.61, P<0.005).
Conclusions: Microbial collagenolytic activity in stool may be a useful marker to predict anastomotic leak. In addition to E. faecalis, this phenotype association may exist in other intestinal bacteria and is under investigation.