Pancreaticobiliary biofilm: is cholesterol the answer?

Bacteria in the biliary tree have been the bane of surgeons and gastroenterologists alike. On the one hand, up to 50% of patients with chronic cholecystitis have positive bile cultures, as do 75% of patients with acute cholecystitis, and virtually 100% of individuals with cholangitis.^1,2 On the other hand, gut flora are thought to play a key role in the formation of at least black pigment and mixed gall stones, by virtue of mucus production, and the elaboration of B-glucuronidase that deconjugates bilirubin leading to its precipitation with calcium and palmitate.^3–7 Moreover, in the form of a biofilm in which bacteria deposit themselves in complex patterns within a mucopolysaccharide or glycocalyx infrastructure, their presence has been associated with progressive occlusion of implanted medical devices which have a lumen and infections resistant to conventional courses of antibiotics.⁸ Plastic biliary stents, most often placed for malignant obstructive jaundice, have also been noted to occlude by virtue of biofilm development.^9,10 This biofilm appears to be multidimensional and consists of slime, immunoglobulins, and other proteins in conjunction with multiple bacteria species in the setting of incomplete stent occlusion, and calcium bicarbonate and palmitate interlaced within a bacterial matrix with complete stent occlusion.^4,9,11–13 In either setting, the cholesterol concentration within the stent is relatively low.

Attempts to improve prosthesis patency by use of prophylactic antibiotics, antimucin drugs, ursodeoxycholic acid, and changes in stent polymer have invariably been disappointing.^14–16 Clinical stent occlusion leads to jaundice and bacterial cholangitis, with polymicrobial infections in up to 90% of patients in several studies.^2,11,17 Recently, similar changes to include polymicrobial bacterial contamination and pancreatic sepsis have been noted in patients with indwelling pancreatic prostheses.^18,19

In this issue of Gut, Swidsinski and colleagues²⁰ present elegant insights into the presence of bacterial biofilm and the viability of those bacteria in duodenal, bile duct, and pancreatic duct mucosa as well as in gall stones, common bile duct stones, and biliary stents (see page 388). Using oligonucleotide probes, fluorescence in situ hybridisation (FISH) studies were used to characterise bacterial species within a biofilm and three different parameters used to define presumptive bacterial viability: (1) amenability of cells to FISH versus ratio of EUS 338 Cy3 positive and Gram positive cells (DNA staining); (2) hybridisation time required to achieve fluorescence signal; and (3) time associated with complete exhaustion of fluorescence. Decreased bacterial viability was assumed in the presence of decreased DNA staining, prolonged time to achieve autofluorescence, and rapid exhaustion of fluorescence staining.

Results included the presence of a bacterial biofilm in brown/mixed gall stones, findings supported by Stewart, et al, who noted that 73% of pigmented gall stones/bile duct stones contained bacteria,⁴ data comparable with those previously noted with black pigment stones. In contrast, no biofilm was noted within the mucosa of 20 gall bladder or five bile duct walls or in the elutes of 132 cholesterol gall stones that were tested.

The latter data can be construed as a different pathogenesis of cholesterol gall stone formation (mucin eliciting cholesterol crystal nucleation in supersaturated bile).²¹ Alternatively, as the authors suggest, it is possible that cholesterol has the potential to inhibit bacterial growth and may actually be a novel mechanism to inhibit bacterial growth within biofilms.

Perhaps the most intriguing findings in the study were the presence of bacterial attachment to pancreatic duct epithelium in seven of nine patients with chronic calcific pancreatitis as well as the dramatic diminution in bacterial viability of bacteria within a biofilm with progressive biliary stent occlusion. Our group has previously noted negative bacterial cultures in patients with chronic calcific pancreatitis at the time of initial manipulation but an average of 3.4 enteric bacteria in aspirated juice once a stent has been placed.¹⁸ Moreover, we have noted that pancreatic sepsis can occur uncommonly and that prosthesis occlusion is necessary but not a sufficient explanation for why the patient develops infectious complications of pancreatic endotherapy. The findings by Swidsinski et al, that individuals with chronic calcific pancreatitis who have undergone endoscopic therapy develop complex biofilms within the duct epithelium may mean nothing more than exposure of the pancreatic duct to duodenal flora by virtue of stent placement or pancreatic sphincterotomy.²⁰ If so, further work should demonstrate that distal bile duct walls have comparable bacterial biofilm in individuals who have undergone biliary sphincterotomy. The latter may or may not play a role in the subsequent development of common bile duct stones seen in a subset of patients, even those with widely patent sphincterotomies. Alternatively, the development of biofilm in patients with chronic calcific pancreatitis may be onerous and associated with increased stone formation, paralleling recent studies demonstrating pancreatic stone protein within biliary stents prior to complete occlusion.¹¹

Perhaps the most interesting aspect of the manuscript from a personal standpoint was the finding that bacteria were always above, and not below, the sludge matrix in patients with variably occluded biliary stents, and that bacterial concentrations in patent, but narrowed, stents were higher at the liver, as opposed to the duodenal, end of the prosthesis. In contrast, occluded stents, as well as brown pigment and mixed cholesterol gall stones, were associated with decreased viable bacteria as detectable by FISH, and less than 10% of the bacteria seen by Gram stain or autofluorescence were amenable to FISH. Had the authors simply cultured the bile proximal to the stent, however, our group as well as others would have shown polymicrobial bacterial contamination that may ultimately result in “stent flu” or frank cholangitis.^2,17,18

The implied conclusion that cholesterol stone formation may be a physiological response to limit bacterial biofilm development in bile and therefore useful to prevent biofilm development and subsequent stent occlusion in patients with biliary prostheses holds some merit. More likely, however, cholesterol coated stents or pharmacological manipulation of biliary excretion of cholesterol would not preclude mucin or protein deposition onto the inner stent surface, findings noted by most authors, prior to bacterial adherence.^10–13 Nevertheless, Swidsinski et al have added further fuel to the importance of bacterial biofilms and the need to delineate novel therapies to prevent their initiation or preclude their propagation.