An in vivo model to study the pathobiology of infectious biofilms on biomaterial surfaces A. Buret,* K. H. Ward, M. E. Olson, and J.W. Costerton Departments of Biological Sciences, Microbiology and Infectious Diseases, University of Calgary, Calgary, Alberta, Canada This study examines the morphology, ultrastructure, and microbiology of the intact biofilm developing on an implant surface. Silastic subdermal implant material was colonized with I? aeruginosa and surgically inserted into the peritoneal cavity of adult rabbits. After 4, 8, 28, and 42 days implants were recovered and the intact biofilms examined. P aeruginosa colonized the implant throughout the entire experimental time. Microcolonies of glycocalyx-coated bacteria were observed within the biofilm. However, the bulk of the biofilm was host-generated and typically contained phagocytes trapped within a thick mesh of fibrin. Polymorphonuclear neutrophils were the predominant cell type. Isolated erythrocytes,
macrophages, and fibroblasts were also observed. By day 28, the biofilm was enclosed in a fibrous capsule of vascularized connective tissue. The low numbers of neutrophils seen in biofilms from sterile Silastic sheets implanted into control animals suggested that neutrophilia may represent a specific cellular response to the bacterial colonization. The results indicate that the cell-mediated immune response provides for most of the biofilm mass on colonized implant surfaces. Inactivated phagocytes trapped in fibrin may ”wall-off” the embedded bacterial microcolonies and thus shield them from live phagocytic leucocytes. Such a mechanism may play an important role in the pathogenesis of prosthetic device infections.
IN TRODUCTION
Since Elek and Conen demonstrated the ability of implanted foreign bodies to potentiate infection,’ the pathobiology of infection in prosthetic devices has become the focal point of an increasing number of Although most implant-related infections are caused by gram-positive bacteria, infections due to gram-negative bacteria are potentially more serious? It is generally accepted that generation of a biofilm, resulting from the combined effects of bacterial adhesion to the surface of the implant, production of bacterial glycocalyx, and accumulation of host inflammatory products, plays an important role in the pathogenesis of foreign body Yet, several factors pertinent to the pathogenesis of foreign body infections remain incompletely understood, and little is known about the physical and chrono*To whom correspondence should be addressed at Dept. of Biological Sciences, University of Calgary, Biosciences 025, 2500 University Dr. N.W. Calgary, Alberta, Canada T2N 1N4. Journal of Biomedical Materials Research, Vol. 25, 865-874 (1991) 0 1991 John Wiley & Sons, Inc. CCC 0021-9304/91/070865-10$4.00
BURET ET AL.
866
logical arrangements in sifu of bacterial and tissue components within the biofilm over the course of infection. The aims of this study were (a) to develop an animal model to investigate the evolution of a bacterial biofilm in vivo and (b) to examine the ultrastructural organization within the biofilm by light and electron microscopy.
MATERIALS A N D METHODS
Bacterial strain and culture methods The organism used in this study was a laboratory strain of Pseudomonas aeruginosa (I’AO-1). Bacteria were maintained on nutrient agar slopes at -70°C. Samples for experimentation were cultivated overnight at 37°C in a chemically defined medium lacking iron5 following standard procedures. Implant material Non-reinforced Silastic subdermal implant material (Dow Corning Corp., Medical Products, Midland, MI, USA) was cut into 4.5/2 cm sheets, sterilized in ethylene oxide for 4 h, and degassed for 48 h. Implants were assigned to two groups: (a) An experimental group, in which Silastic sheets were colonized with I! aeruginosa prior to implantation. Bacteria used for colonization were harvested by centrifugation at 1600g for 10 mins, washed once in sterile phosphate buffered saline (PBS), and resuspended to a concentration of lo7 CFU/mL of PBS. Colonization was achieved by 3-h incubation of the implants at 37°C on a rocking platform in a solution containing suspended bacteria. All implants were rinsed in sterile I‘BS prior to insertion in order to remove non-adherent bacteria. (b) Results were compared to a control group of sterile Silastic sheets. Experimental design Adult female New Zealand white rabbits (n = 8, Vandermeer C., Sherwood Park, AB, CDN T8C 1H5) weighing approximately 3 kg were used in all experiments. Under halothane anesthesia, the rabbit’s abdomen was shaved and disinfected with undiluted Hibitane (chlorhexidine acetate, Ayerst Laboratories, Montreal, Canada) solution. A midline incision was performed through all layers and a Silastic sheet was inserted into each side of the peritoneal cavity and sutured by all four corners to the internal peritoneal fascia with 3-0 propylene monofilament suture (Prolene, Ethicon Ltd., Peterborough, Ontario, Canada). The internal peritoneal fascia and the subcutaneous layer were closed with interrupted or continuous 3-0 absorbable polyglycolic acid sutures (Vicryl, Ethicon Ltd.) respectively and the skin was closed with 3-0 interrupted propylene monofilament sutures. The experimental protocol in-
PATHOBIOLOGY OF INFECTIOUS BIOFILMS
867
volved two groups: In group 1,Silastic implants were colonized with P aeruginosa. Four, 8, 28, and 42 days postsurgery, the implant was surgically removed and processed as described below. In group 2, the control group, sham-treated sterile Silastic implants were studied 4 and 28 days postsurgery. On every sampling day, three-four implants from two-three different animals were studied in each group.
Processing of recovered implants Immediately after recovery (i.e., at day 4,8,28, or 42), implants were cut into two equal pieces. One piece was placed into 1mL of sterile PBS and used to assess the progression of the I? aeruginosa infection using standard quantitative recovery procedures? Briefly, the surface of the implant was scraped into 5mL sterile PBS with a sterile scalpel blade. Scrapings and Silastic in PBS were vortexed for 1 min and ultrasonicated for 10 min at very low output to detach cells from the Silastic. The suspension was then homogenized 3 x 30 s to break up bacterial clumps. Duplicate 1/10 dilution series were obtained and plated onto nutrient agar. Bacterial colonization of the implant was calculated in CFU/cm2 of Silastic. The second half of the recovered implant was placed into 5% gluteraldehyde and 0.15% ruthenium red in 0.1 M cacodylic buffer, pH 7.2, for morphological assessment by light and transmission electron microscopy (TEM).
Morphology Implants were stored overnight in the original gluteraldehyde fixative at 4°C. Specimens were then washed in 0.05% ruthenium red cacodylic buffer (0.1 M at pH 7.2), post-fixed in Os04,dehydrated in distilled acetone, cleared with propylene oxide and infiltrated in Spurr's low viscosity medium (J.B. EM Services Inc., Dorval, Quebec, Canada). Following infiltration, intact biofilms were lifted off the implant and placed into the embedding molds so that the transverse plane of the biofilm would be exposed to the block face for sectioning. Sections (0.25 pm) were obtained and stained with basic toluidine blue for light microscopic evaluation and for selecting areas for thin sectioning. Ultra-thin sections (60 nm) for TEM were stained with saturated uranyl acetate in 50% ethanol, followed by 0.4% lead it rate.^ Specimens were examined on a Hitachi 600 Transmission Electron Microscope at 50 kV
RESULTS
Viable bacterial counts
At the time of implantation, 1.43 k 0.08 x lo5 (mean k standard error) viable P aeruginosa per cm2 colonized the biomaterial. The Silastic remained
BURET ET AL.
868
colonized with sessile I? aeruginosa for the entire experimental time and, following an initial decrease in bacterial counts on day 4 (1.07 k 0.6 x lo4 CFU/cm2),increased to 2.09 1.2 x lo5 CFU/cm2by day 42 (Fig. 1).None of the sites of suture showed any evidence of infection throughout the time of study. In contrast, on postmortem examination, sites of implantation from both groups exhibited signs of inflammation, although implants recovered from control animals remained sterile throughout the time of study. +_
Biofilm composition Glycocalyx-coated bacteria were observed in isolated microcolonies through the entire thickness of the colonized biofilm (Fig. 2). An increase in biofilm mass was observed as time progressed (Fig. 3). Four days after implanting the colonized Silastic, viable and nonviable (as assessed from morphological examination under transmission electron microscopy) neutrophils were the predominant cell type found in the biofilm. Isolated erythrocytes and macrophages were also present (Fig. 4). Later in the infection (day 28, 42), the biofilm was enclosed by a fibrous capsule of vascularized connective tissue (Fig. 5). On day 42, in 1 animal out of the 3 studied, the biofilm had reached macroscopic proportions and contained f h i d pus. Formation of a host-generated biofilm was also observed on noncolonized Silastic. At both
0
n
0 7
(3
0 -I
4-
W
(v
E
0
i
-
-
3 LL 0
2-
I
1
1
1
PAT HOBIOLOGY OF INFECTIOUS BIOFILMS
Figure 2. Transmission electron micrograph illustrating a microcolony of glycocalyx-coated (GLY) P aeruginosa in a biofilm recovered from a colonized implant 8 days postsurgery.
Figure 3. Light micrographs of biofilms form colonized implants 2 (A), 4 (B), and 28 (C) days postsurgery showing a significant increase in biofilm thickness over time. The biofilm recovered from a control animal on day 4 (D) is noticeably thinner and is covered with a layer of flattened macrophages (inset). Arrows indicate the bottom of the biofilm. Basic fuchsin and methylene blue.
869
BURET ET AL.
870
Figure 4. Transmission electron micrograph of a biofilm recovered 4 days after surgery. Erythrocyte (E), neutrophil ( N ) and macrophage (M) were trapped within a mesh of fibrin (F).
sampling times, biofilm recovered from control animals was noticeably thinner than biofilm containing bacteria (Fig. 3D). In control biofilms, most macrophages were localized in an almost confluent layer covering the biofilm, and the neutrophil to macrophage ratio was lower than that observed on colonized implants (Fig. 6). Fibrin deposition was observed in biofilms from both experimental groups, but it appeared organized in layers on colonized implants (Fig. 6B). DISCUSSION
The ability of bacteria to adhere and to colonize foreign body surfaces has become a major concern in the use of cardiovascular urinary catheters,’”’’ suture materials and tissue adhesive^,'^,'^ or any prosthetic im-
PATHOBIOLOGY OF INFECTIOUS BIOFILMS
871
Figure 5. Micrographs illustrating that by day 28, the biofilm was composed of a vascularized capsule of connective tissue (A) surrounding high numbers of viable and nonviable inflammatory cells (B). (A) methylene blue and basic fuchsin; (b) transmission electron micrograph.
plants.23Minimal biofilm recovery from implanted devices would be in contradiction with the prevalence of such infections. However, this may occur if precautions were not taken during removal of the implant. As reported previo~sly,'~ the biofilm layer may slide off the device during removal and remain within the host tissue. In vim experimentation is necessary when studying physiologic activities of these biofilm b a ~ t e r i aScanning .~~ electron microscopy is one of the tools most commonly used to visualize intact biofilms in s i h 6 Difficulties inherent to TEM studies involve disorganization of biofilm structure during sample preparation. In previous studies of foreign-body-related biofilms, TEM was performed on scrapings obtained from the biomaterial s~rface.~,"~ No TEM methodology allowing direct processing of intact biofilms recovered from hard substrata was available. This report describes a new in v i m model allowing recovery and observation of cross-sections from intact biofilms developing on biomaterial surfaces. Although the recovery procedures used in this experiment result in an underestimate of the true bacterial colonization,6 I? aeruginosa proliferation was demonstrated on precolonized implants. Light microscopy and TEM yielded data on biofilm thickness, morphology, and ultrastructural organization. Our results suggest that the bulk of a biornaterialrelated biofilm is host-generated. This illustrates that tissue response to implanted devices is a major component in biofilm formation and that the
872
BURET ET AL.
Figure 6. Toluidine blue transverse sections of control (A) and colonized (B) biofilms recovered on day 4. Note the higher proportion of neutrophils (N) in colonized compared to control biofilms, and the presence of a mesh of fibrin (F) organized in layers within the colonized biofilm. M: macrophages.
mere presence of a non-phagocytosable implant in the host will trigger an inf lammatory response, thus confirming the findings of other authors." The role of mononuclear and polymorphonuclear cells in peritoneal defense has been discussed in previous reports2,17and there is evidence to suggest that peritoneal macrophages can phagocytize and kill bacteria as effectively as polymorphonuclear neutrophils." Findings from this study confirm the involvement of a chronic cell-mediated immune response (predominantly polymorphonuclear neutrophilic) to an intraperitoneal bacterial infection, and further demonstrate the importance of peritoneal macrophages in the defense against such infections. Trapping of phagocytes and microorganisms within the infectious biofilm was clearly apparent throughout the present experiment. In one instance, the walling-off and necrosis of host and pathogen debris resulted in abscess formation. In this case, bacterial counts were significantly higher than when no abscess was formed. This provides further evidence that such a phenomenon may shield bacteria from an early inflammatory response and hence account for the late infections commonly observed in association with prosthetic devices. As suggested previously,2' this shielding factor may combine with PMN inactivation and promote the onset of infection. Other reports, analyzing the role of local host factors in foreignbody Staphylococcus aureus infections, also described neutrophilia during the
PATHOBIOLOGY OF INFECTIOUS BIOFILMS
873
early stages following implantation and formation of a vascularized capsular tissue several weeks later.20In agreement with previous ~ t u d i e s , ~our ~ ”data ~ show that foreign-body-related infections promote the accumulation of a heavy mesh of fibrin on the biomaterial surface. Deposition of fibrin on inserted catheters appears to promote staphylococcal adherence?l Fibrin may play a similar role in the pathogenesis of peritoneal implant infections. In conclusion, the model described in this paper can be used to study the in v i m development of a biofilm on infected peritoneal implants. Bacteria proliferated in microcolonies within the biofilm despite the massive host inf lammatory response thus suggesting, in accordance with other studies, that the inactivated leukocytes and the mesh of fibrin accumulating in the biofilm may shield bacteria from live phagocytic cells. The similarities between the pathobiology of biofilms observed in the present study and previous data obtained from other experimental’’ and human trial^^,^,'"'^,^^ make this model a useful development for the study of the pathogenesis of foreign-body-related infections and for the future experimentation of potential solutions to the growing problem of such infections. This work was supported by the Natural Sciences and Engineering Research Council of Canada and the Alberta Heritage Foundation for Medical Research.
References 1.
2. 3. 4.
5. 6.
7. 8. 9.
10. 11.
S. D. Elek and P. E. Conen, “The virulence of Staphylococcus pyogenes for man: a study of the problems of wound infection,“ Br. I. Exp. Pathol., 38, 573-586 (1957). S. H. Dougherty, ”Pathobiology of infection in prosthetic devices,” Rev. Infect. Dis., 10(6), 1102-1117 (1988). M. Jacques, T. J. Marrie, and J.W. Costerton, “Review: Microbial colonization of prosthetic devices,“ Microb. Ecol., 13, 173-191 (1987). T. Khoury and J.W. Costerton, “Biofilms in nature and medicine,” in 3rd ECCM Symposium Foreign Body-Related Infections, Elsevier, Amsterdam, 1988, pp. 2-15. M. R.W. Brown, H. Anwar, and P. A. Lambert, “Evidence that mucoid Pseudomonas aeruginosa in the Cystic Fibrosis lung grows under ironrestricted conditions,” FEMS Microbiol. Lett., 21, 113-117 (1984). J.W. Costerton, J.C. Nickel, and T. I. Ladd, “Suitable methods for the comparative study of free-living and surface-associated bacterial populations,“ in Bacteria in Nature, Vol. 2, J. s. Poindexter and E. R. Leadbetter (eds.), Plenum, New York, 1986, pp. 49-84. J.H. Venable and R. Coggeshall, ‘A simplified lead citrate strain for use i n electron microscopy,” I. Cell. Biol., 25, 407-408 (1965). D.G. Maki, C. E. Weise, and H.W. Sarafin, ‘A semi-quantitative method of identifying intravenous catheter-related infection,“ N. Engl. I. Med., 296, 1305-1309 (1977). G. L. Archer, “Experimental endocarditis,” in infections of Prosthetic Heart Valves and Vascular Grafts, R. J. Duma (ed.), University Park Press, Baltimore, 1977, pp. 43-59. W. E. Stam, ”Guidelines for prevention of catheter-associated urinary tract infection,” Ann. Intern. Med., 82, 386-390 (1975). T. I. Ladd, J.C. Nickel, and J.W. Costerton, ”Rapid method for detection of adherent bacteria on Foley urinary catheters,” I. Clin. Microbiol., 21 (6), 1004-1006 (1985).
BURET ET AL.
874 12. 13.
14.
15. 16.
17. 18.
19. 20.
21.
22.
S. Katz, M. Izhar, and D. Mirelman, ”Bacterial adhesion to surgical sutures. A possible factor in suture-induced infection,“ Ann. S i q . , 194, 35-41 (1981). M. E. Olson, I. Ruseska, and J.W. Costerton, ”Colonization of n-butyl2-cyanoacrylate tissue adhesive by Staphylococcus epidermidis, “ 1. Biomed. Mater. Res., 22, 485-495 (1988). M. K. Dasgupta, R. A. Ulan, K. B. Bettcher, V. Burns, K. Lam, J. B. Dossetor, and J.W. Costerton, ”Effect of exit site infection and peritonitis on the distribution of biofilm encased adherent bacterial microcolonies (BABM) on Tenckhoff (T) catheters in patients undergoing continuous ambulatory peritoneal dialysis (CAPD),” Ada Peritoneal Dialysis, 102-109 (1986). T. J. Marrie and J.W. Costerton, ”Scanning and transmission electron microscopy of in situ bacterial colonization of intravenous and intraarterial catheters,” 1. Clin. Microbiol., 19(5), 687-693 (1984). G.D. Christensen, L.M. Baddour, D.L. Hasty, J.L. Lowrance, and W. A. Simpson, “Microbial and foreign body factors in the pathogenesis of medical device infections,” in Infections Associated w i f h Indwelling Medical Devices, A . L. Bisno and F. A. Waldvogel (eds.), Am. SOC.Microbiol., Washington, DC 20006, 1989, pp. 27-59. R. M. Hurley, D. Muogabo, G.W. Wilson, and M.A.M. Ali, “Cellular composition of peritoneal effluent: Response to bacterial peritonitis,” Can. Med. Assoc. J., 117, 1061-1062 (1977). H.A Verbrugh, W.F. Keane, J . R . Hoidal, M.F. Freiberg, G.R. Elliott, and P. K. Peterson, “Peritoneal macrophages and opsonins: Antibacterial defense in patients undergoing chronic peritoneal dialysis,” I. Infect. Dis., 147(6), 1018-1029 (1983). W. M. Scheld, J. A. Valone, and M. A. Sande, ”Bacterial adherence in the pathogenesis of endocarditis. Interaction of bacterial dextran, platelets, and fibrin,” J. Clin. Invest., 61, 1394-1404 (1978). P. E. Vaudaux, D. P. Lew, and F. A. Waldvogel, “Host factors predisposing to foreign body infections,” in Infections Associated with lndwelling Medical Devices, A. L. Bisno and F. A. Waldvogel (eds.), Am. SOC.Microbiol., Washington, DC 20006, pp. 3-26. P. Vaudaux, D. Pittet, A. Haeberli, E. Huggler, U.E. Nydegger, D. P. Lew, and F. A. Waldvogel, ”Host factors selectively increase staphylococcal adherence on inserted catheters: A role of fibronectin and fibrinogen or fibrin,” J. Inf. Dis., 160(5), 865-875 (1989). M. K. Dasgupta, K. B. Bettcher, R.A. Ulan, V. A. Burns, K . Lam, J. B. Dossetor, and J.W. Costerton, ”Relationship of adherent bacterial biofilms to peritonitis in chronic ambulatory peritoneal dialysis,” Perit. Dial. Bull., 7, 168-173 (1987).
Received February 19, 1990 Accepted February 12, 1991
macrophages, and fibroblasts were also observed. By day 28, the biofilm was enclosed in a fibrous capsule of vascularized connective tissue. The low numbers of neutrophils seen in biofilms from sterile Silastic sheets implanted into control animals suggested that neutrophilia may represent a specific cellular response to the bacterial colonization. The results indicate that the cell-mediated immune response provides for most of the biofilm mass on colonized implant surfaces. Inactivated phagocytes trapped in fibrin may ”wall-off” the embedded bacterial microcolonies and thus shield them from live phagocytic leucocytes. Such a mechanism may play an important role in the pathogenesis of prosthetic device infections.
IN TRODUCTION
Since Elek and Conen demonstrated the ability of implanted foreign bodies to potentiate infection,’ the pathobiology of infection in prosthetic devices has become the focal point of an increasing number of Although most implant-related infections are caused by gram-positive bacteria, infections due to gram-negative bacteria are potentially more serious? It is generally accepted that generation of a biofilm, resulting from the combined effects of bacterial adhesion to the surface of the implant, production of bacterial glycocalyx, and accumulation of host inflammatory products, plays an important role in the pathogenesis of foreign body Yet, several factors pertinent to the pathogenesis of foreign body infections remain incompletely understood, and little is known about the physical and chrono*To whom correspondence should be addressed at Dept. of Biological Sciences, University of Calgary, Biosciences 025, 2500 University Dr. N.W. Calgary, Alberta, Canada T2N 1N4. Journal of Biomedical Materials Research, Vol. 25, 865-874 (1991) 0 1991 John Wiley & Sons, Inc. CCC 0021-9304/91/070865-10$4.00
BURET ET AL.
866
logical arrangements in sifu of bacterial and tissue components within the biofilm over the course of infection. The aims of this study were (a) to develop an animal model to investigate the evolution of a bacterial biofilm in vivo and (b) to examine the ultrastructural organization within the biofilm by light and electron microscopy.
MATERIALS A N D METHODS
Bacterial strain and culture methods The organism used in this study was a laboratory strain of Pseudomonas aeruginosa (I’AO-1). Bacteria were maintained on nutrient agar slopes at -70°C. Samples for experimentation were cultivated overnight at 37°C in a chemically defined medium lacking iron5 following standard procedures. Implant material Non-reinforced Silastic subdermal implant material (Dow Corning Corp., Medical Products, Midland, MI, USA) was cut into 4.5/2 cm sheets, sterilized in ethylene oxide for 4 h, and degassed for 48 h. Implants were assigned to two groups: (a) An experimental group, in which Silastic sheets were colonized with I! aeruginosa prior to implantation. Bacteria used for colonization were harvested by centrifugation at 1600g for 10 mins, washed once in sterile phosphate buffered saline (PBS), and resuspended to a concentration of lo7 CFU/mL of PBS. Colonization was achieved by 3-h incubation of the implants at 37°C on a rocking platform in a solution containing suspended bacteria. All implants were rinsed in sterile I‘BS prior to insertion in order to remove non-adherent bacteria. (b) Results were compared to a control group of sterile Silastic sheets. Experimental design Adult female New Zealand white rabbits (n = 8, Vandermeer C., Sherwood Park, AB, CDN T8C 1H5) weighing approximately 3 kg were used in all experiments. Under halothane anesthesia, the rabbit’s abdomen was shaved and disinfected with undiluted Hibitane (chlorhexidine acetate, Ayerst Laboratories, Montreal, Canada) solution. A midline incision was performed through all layers and a Silastic sheet was inserted into each side of the peritoneal cavity and sutured by all four corners to the internal peritoneal fascia with 3-0 propylene monofilament suture (Prolene, Ethicon Ltd., Peterborough, Ontario, Canada). The internal peritoneal fascia and the subcutaneous layer were closed with interrupted or continuous 3-0 absorbable polyglycolic acid sutures (Vicryl, Ethicon Ltd.) respectively and the skin was closed with 3-0 interrupted propylene monofilament sutures. The experimental protocol in-
PATHOBIOLOGY OF INFECTIOUS BIOFILMS
867
volved two groups: In group 1,Silastic implants were colonized with P aeruginosa. Four, 8, 28, and 42 days postsurgery, the implant was surgically removed and processed as described below. In group 2, the control group, sham-treated sterile Silastic implants were studied 4 and 28 days postsurgery. On every sampling day, three-four implants from two-three different animals were studied in each group.
Processing of recovered implants Immediately after recovery (i.e., at day 4,8,28, or 42), implants were cut into two equal pieces. One piece was placed into 1mL of sterile PBS and used to assess the progression of the I? aeruginosa infection using standard quantitative recovery procedures? Briefly, the surface of the implant was scraped into 5mL sterile PBS with a sterile scalpel blade. Scrapings and Silastic in PBS were vortexed for 1 min and ultrasonicated for 10 min at very low output to detach cells from the Silastic. The suspension was then homogenized 3 x 30 s to break up bacterial clumps. Duplicate 1/10 dilution series were obtained and plated onto nutrient agar. Bacterial colonization of the implant was calculated in CFU/cm2 of Silastic. The second half of the recovered implant was placed into 5% gluteraldehyde and 0.15% ruthenium red in 0.1 M cacodylic buffer, pH 7.2, for morphological assessment by light and transmission electron microscopy (TEM).
Morphology Implants were stored overnight in the original gluteraldehyde fixative at 4°C. Specimens were then washed in 0.05% ruthenium red cacodylic buffer (0.1 M at pH 7.2), post-fixed in Os04,dehydrated in distilled acetone, cleared with propylene oxide and infiltrated in Spurr's low viscosity medium (J.B. EM Services Inc., Dorval, Quebec, Canada). Following infiltration, intact biofilms were lifted off the implant and placed into the embedding molds so that the transverse plane of the biofilm would be exposed to the block face for sectioning. Sections (0.25 pm) were obtained and stained with basic toluidine blue for light microscopic evaluation and for selecting areas for thin sectioning. Ultra-thin sections (60 nm) for TEM were stained with saturated uranyl acetate in 50% ethanol, followed by 0.4% lead it rate.^ Specimens were examined on a Hitachi 600 Transmission Electron Microscope at 50 kV
RESULTS
Viable bacterial counts
At the time of implantation, 1.43 k 0.08 x lo5 (mean k standard error) viable P aeruginosa per cm2 colonized the biomaterial. The Silastic remained
BURET ET AL.
868
colonized with sessile I? aeruginosa for the entire experimental time and, following an initial decrease in bacterial counts on day 4 (1.07 k 0.6 x lo4 CFU/cm2),increased to 2.09 1.2 x lo5 CFU/cm2by day 42 (Fig. 1).None of the sites of suture showed any evidence of infection throughout the time of study. In contrast, on postmortem examination, sites of implantation from both groups exhibited signs of inflammation, although implants recovered from control animals remained sterile throughout the time of study. +_
Biofilm composition Glycocalyx-coated bacteria were observed in isolated microcolonies through the entire thickness of the colonized biofilm (Fig. 2). An increase in biofilm mass was observed as time progressed (Fig. 3). Four days after implanting the colonized Silastic, viable and nonviable (as assessed from morphological examination under transmission electron microscopy) neutrophils were the predominant cell type found in the biofilm. Isolated erythrocytes and macrophages were also present (Fig. 4). Later in the infection (day 28, 42), the biofilm was enclosed by a fibrous capsule of vascularized connective tissue (Fig. 5). On day 42, in 1 animal out of the 3 studied, the biofilm had reached macroscopic proportions and contained f h i d pus. Formation of a host-generated biofilm was also observed on noncolonized Silastic. At both
0
n
0 7
(3
0 -I
4-
W
(v
E
0
i
-
-
3 LL 0
2-
I
1
1
1
PAT HOBIOLOGY OF INFECTIOUS BIOFILMS
Figure 2. Transmission electron micrograph illustrating a microcolony of glycocalyx-coated (GLY) P aeruginosa in a biofilm recovered from a colonized implant 8 days postsurgery.
Figure 3. Light micrographs of biofilms form colonized implants 2 (A), 4 (B), and 28 (C) days postsurgery showing a significant increase in biofilm thickness over time. The biofilm recovered from a control animal on day 4 (D) is noticeably thinner and is covered with a layer of flattened macrophages (inset). Arrows indicate the bottom of the biofilm. Basic fuchsin and methylene blue.
869
BURET ET AL.
870
Figure 4. Transmission electron micrograph of a biofilm recovered 4 days after surgery. Erythrocyte (E), neutrophil ( N ) and macrophage (M) were trapped within a mesh of fibrin (F).
sampling times, biofilm recovered from control animals was noticeably thinner than biofilm containing bacteria (Fig. 3D). In control biofilms, most macrophages were localized in an almost confluent layer covering the biofilm, and the neutrophil to macrophage ratio was lower than that observed on colonized implants (Fig. 6). Fibrin deposition was observed in biofilms from both experimental groups, but it appeared organized in layers on colonized implants (Fig. 6B). DISCUSSION
The ability of bacteria to adhere and to colonize foreign body surfaces has become a major concern in the use of cardiovascular urinary catheters,’”’’ suture materials and tissue adhesive^,'^,'^ or any prosthetic im-
PATHOBIOLOGY OF INFECTIOUS BIOFILMS
871
Figure 5. Micrographs illustrating that by day 28, the biofilm was composed of a vascularized capsule of connective tissue (A) surrounding high numbers of viable and nonviable inflammatory cells (B). (A) methylene blue and basic fuchsin; (b) transmission electron micrograph.
plants.23Minimal biofilm recovery from implanted devices would be in contradiction with the prevalence of such infections. However, this may occur if precautions were not taken during removal of the implant. As reported previo~sly,'~ the biofilm layer may slide off the device during removal and remain within the host tissue. In vim experimentation is necessary when studying physiologic activities of these biofilm b a ~ t e r i aScanning .~~ electron microscopy is one of the tools most commonly used to visualize intact biofilms in s i h 6 Difficulties inherent to TEM studies involve disorganization of biofilm structure during sample preparation. In previous studies of foreign-body-related biofilms, TEM was performed on scrapings obtained from the biomaterial s~rface.~,"~ No TEM methodology allowing direct processing of intact biofilms recovered from hard substrata was available. This report describes a new in v i m model allowing recovery and observation of cross-sections from intact biofilms developing on biomaterial surfaces. Although the recovery procedures used in this experiment result in an underestimate of the true bacterial colonization,6 I? aeruginosa proliferation was demonstrated on precolonized implants. Light microscopy and TEM yielded data on biofilm thickness, morphology, and ultrastructural organization. Our results suggest that the bulk of a biornaterialrelated biofilm is host-generated. This illustrates that tissue response to implanted devices is a major component in biofilm formation and that the
872
BURET ET AL.
Figure 6. Toluidine blue transverse sections of control (A) and colonized (B) biofilms recovered on day 4. Note the higher proportion of neutrophils (N) in colonized compared to control biofilms, and the presence of a mesh of fibrin (F) organized in layers within the colonized biofilm. M: macrophages.
mere presence of a non-phagocytosable implant in the host will trigger an inf lammatory response, thus confirming the findings of other authors." The role of mononuclear and polymorphonuclear cells in peritoneal defense has been discussed in previous reports2,17and there is evidence to suggest that peritoneal macrophages can phagocytize and kill bacteria as effectively as polymorphonuclear neutrophils." Findings from this study confirm the involvement of a chronic cell-mediated immune response (predominantly polymorphonuclear neutrophilic) to an intraperitoneal bacterial infection, and further demonstrate the importance of peritoneal macrophages in the defense against such infections. Trapping of phagocytes and microorganisms within the infectious biofilm was clearly apparent throughout the present experiment. In one instance, the walling-off and necrosis of host and pathogen debris resulted in abscess formation. In this case, bacterial counts were significantly higher than when no abscess was formed. This provides further evidence that such a phenomenon may shield bacteria from an early inflammatory response and hence account for the late infections commonly observed in association with prosthetic devices. As suggested previously,2' this shielding factor may combine with PMN inactivation and promote the onset of infection. Other reports, analyzing the role of local host factors in foreignbody Staphylococcus aureus infections, also described neutrophilia during the
PATHOBIOLOGY OF INFECTIOUS BIOFILMS
873
early stages following implantation and formation of a vascularized capsular tissue several weeks later.20In agreement with previous ~ t u d i e s , ~our ~ ”data ~ show that foreign-body-related infections promote the accumulation of a heavy mesh of fibrin on the biomaterial surface. Deposition of fibrin on inserted catheters appears to promote staphylococcal adherence?l Fibrin may play a similar role in the pathogenesis of peritoneal implant infections. In conclusion, the model described in this paper can be used to study the in v i m development of a biofilm on infected peritoneal implants. Bacteria proliferated in microcolonies within the biofilm despite the massive host inf lammatory response thus suggesting, in accordance with other studies, that the inactivated leukocytes and the mesh of fibrin accumulating in the biofilm may shield bacteria from live phagocytic cells. The similarities between the pathobiology of biofilms observed in the present study and previous data obtained from other experimental’’ and human trial^^,^,'"'^,^^ make this model a useful development for the study of the pathogenesis of foreign-body-related infections and for the future experimentation of potential solutions to the growing problem of such infections. This work was supported by the Natural Sciences and Engineering Research Council of Canada and the Alberta Heritage Foundation for Medical Research.
References 1.
2. 3. 4.
5. 6.
7. 8. 9.
10. 11.
S. D. Elek and P. E. Conen, “The virulence of Staphylococcus pyogenes for man: a study of the problems of wound infection,“ Br. I. Exp. Pathol., 38, 573-586 (1957). S. H. Dougherty, ”Pathobiology of infection in prosthetic devices,” Rev. Infect. Dis., 10(6), 1102-1117 (1988). M. Jacques, T. J. Marrie, and J.W. Costerton, “Review: Microbial colonization of prosthetic devices,“ Microb. Ecol., 13, 173-191 (1987). T. Khoury and J.W. Costerton, “Biofilms in nature and medicine,” in 3rd ECCM Symposium Foreign Body-Related Infections, Elsevier, Amsterdam, 1988, pp. 2-15. M. R.W. Brown, H. Anwar, and P. A. Lambert, “Evidence that mucoid Pseudomonas aeruginosa in the Cystic Fibrosis lung grows under ironrestricted conditions,” FEMS Microbiol. Lett., 21, 113-117 (1984). J.W. Costerton, J.C. Nickel, and T. I. Ladd, “Suitable methods for the comparative study of free-living and surface-associated bacterial populations,“ in Bacteria in Nature, Vol. 2, J. s. Poindexter and E. R. Leadbetter (eds.), Plenum, New York, 1986, pp. 49-84. J.H. Venable and R. Coggeshall, ‘A simplified lead citrate strain for use i n electron microscopy,” I. Cell. Biol., 25, 407-408 (1965). D.G. Maki, C. E. Weise, and H.W. Sarafin, ‘A semi-quantitative method of identifying intravenous catheter-related infection,“ N. Engl. I. Med., 296, 1305-1309 (1977). G. L. Archer, “Experimental endocarditis,” in infections of Prosthetic Heart Valves and Vascular Grafts, R. J. Duma (ed.), University Park Press, Baltimore, 1977, pp. 43-59. W. E. Stam, ”Guidelines for prevention of catheter-associated urinary tract infection,” Ann. Intern. Med., 82, 386-390 (1975). T. I. Ladd, J.C. Nickel, and J.W. Costerton, ”Rapid method for detection of adherent bacteria on Foley urinary catheters,” I. Clin. Microbiol., 21 (6), 1004-1006 (1985).
BURET ET AL.
874 12. 13.
14.
15. 16.
17. 18.
19. 20.
21.
22.
S. Katz, M. Izhar, and D. Mirelman, ”Bacterial adhesion to surgical sutures. A possible factor in suture-induced infection,“ Ann. S i q . , 194, 35-41 (1981). M. E. Olson, I. Ruseska, and J.W. Costerton, ”Colonization of n-butyl2-cyanoacrylate tissue adhesive by Staphylococcus epidermidis, “ 1. Biomed. Mater. Res., 22, 485-495 (1988). M. K. Dasgupta, R. A. Ulan, K. B. Bettcher, V. Burns, K. Lam, J. B. Dossetor, and J.W. Costerton, ”Effect of exit site infection and peritonitis on the distribution of biofilm encased adherent bacterial microcolonies (BABM) on Tenckhoff (T) catheters in patients undergoing continuous ambulatory peritoneal dialysis (CAPD),” Ada Peritoneal Dialysis, 102-109 (1986). T. J. Marrie and J.W. Costerton, ”Scanning and transmission electron microscopy of in situ bacterial colonization of intravenous and intraarterial catheters,” 1. Clin. Microbiol., 19(5), 687-693 (1984). G.D. Christensen, L.M. Baddour, D.L. Hasty, J.L. Lowrance, and W. A. Simpson, “Microbial and foreign body factors in the pathogenesis of medical device infections,” in Infections Associated w i f h Indwelling Medical Devices, A . L. Bisno and F. A. Waldvogel (eds.), Am. SOC.Microbiol., Washington, DC 20006, 1989, pp. 27-59. R. M. Hurley, D. Muogabo, G.W. Wilson, and M.A.M. Ali, “Cellular composition of peritoneal effluent: Response to bacterial peritonitis,” Can. Med. Assoc. J., 117, 1061-1062 (1977). H.A Verbrugh, W.F. Keane, J . R . Hoidal, M.F. Freiberg, G.R. Elliott, and P. K. Peterson, “Peritoneal macrophages and opsonins: Antibacterial defense in patients undergoing chronic peritoneal dialysis,” I. Infect. Dis., 147(6), 1018-1029 (1983). W. M. Scheld, J. A. Valone, and M. A. Sande, ”Bacterial adherence in the pathogenesis of endocarditis. Interaction of bacterial dextran, platelets, and fibrin,” J. Clin. Invest., 61, 1394-1404 (1978). P. E. Vaudaux, D. P. Lew, and F. A. Waldvogel, “Host factors predisposing to foreign body infections,” in Infections Associated with lndwelling Medical Devices, A. L. Bisno and F. A. Waldvogel (eds.), Am. SOC.Microbiol., Washington, DC 20006, pp. 3-26. P. Vaudaux, D. Pittet, A. Haeberli, E. Huggler, U.E. Nydegger, D. P. Lew, and F. A. Waldvogel, ”Host factors selectively increase staphylococcal adherence on inserted catheters: A role of fibronectin and fibrinogen or fibrin,” J. Inf. Dis., 160(5), 865-875 (1989). M. K. Dasgupta, K. B. Bettcher, R.A. Ulan, V. A. Burns, K . Lam, J. B. Dossetor, and J.W. Costerton, ”Relationship of adherent bacterial biofilms to peritonitis in chronic ambulatory peritoneal dialysis,” Perit. Dial. Bull., 7, 168-173 (1987).
Received February 19, 1990 Accepted February 12, 1991
