R. A. Jacobson1,2, A. Williamson2, S. Gaines2, K. Wienholts3, B. D. Shogan2, O. Zaborina2, J. C. Alverdy2 1Rush University Medical Center,Surgery,Chicago, IL, USA 2University Of Chicago,Surgery,Chicago, IL, USA 3Radboud University,Nijmegen, Netherlands
Introduction: The protease plasminogen (PLG) and its activator urokinase play a critical role in regulating collagen remodeling during gastrointestinal healing. We have previously shown that colonization of anastomotic tissue by collagenolytic Enterococcus faecalis can cause anastomotic leak (AL). We have recently discovered that E. faecalis can cause overactivation of PLG. The aims of this study were to: 1) determine the mechanism of PLG activation in bacterial-mediated AL pathogenesis, and 2) analyze the ability of tranexamic acid (TXA), a suppressor of PLG activation, to prevent AL caused by bacterial pathogens.
Methods: Activation of PLG and pro-urokinase (puPA) by collagenoltyic E. faecalis V583 and mutant strains lacking the collagenase genes gelE and sprE was measured using fluorogenic assays. In an established model of colorectal AL caused by collagenolytic bacteria, mice received either TXA or vehicle control on postoperative days 1, 2 and 3 via enema. Anastomoses were evaluated on postoperative day 8 using anastomotic healing score (AHS – 0: perfectly healed; 1 – flimsy adhesions; 2 – dense adhesions; 3 – abscess; 4 gross disruption). Fluorescence microscopy was used to demonstrate the temporospatial dynamics of PLG deposition and the impact of TXA on this process.
Results: Wild type E. faecalis activated puPA to a greater extent than mutants deficient in GelE, SprE, or both; activation was partially rescued when mutant strains were complemented with their deficient genes (Fig 1A). In the presence of puPA, the parent strain activated more PLG than mutants (Fig 1B). The presence of TXA inhibited the ability of E. faecalis to bind (flow cytometric data not shown) and activate PLG (497.8±20.1 RFU/s 0 TXA vs 272.1±7.3 10mM TXA, p<.05). When applied to our mouse model of E. faecalis-induced AL, TXA reduced the incidence of leak compared to vehicle (AHS 2.7±0.7 vehicle vs 1.4±0.5 TXA). Rescue of AL with TXA was redemonstrated in our model of P. aeruginosa-induced leak (AHS 1.6±0.5 vs 3.8±0.4, p<.05). Fluorescence microscopy demonstrated qualitatively less PLG at the anastomotic site when mice received TXA compared to vehicle control (Fig 1C).
Conclusion: To our knowledge this is the first mechanistic description of PLG activation by E. faecalis and its involvement in AL. Our data demonstrate microbial PLG activation is a novel, plausible and generalizable mechanism for pathogen-mediated AL. TXA limits the ability of pathogens to activate PLG and disrupt the anastomosis. TXA is inexpensive and FDA approved for use in elective surgery; its local delivery to the anastomotic site may be a promising approach to prevent AL in high risk anastomoses.
