myxin B) and liposome-entrapped antibiotics (amikacin and polymyxin B) on bacterial attachment ... neomycin, bacitracin, polymyxin B, and amikacin (1, 2, 19).
ANTIMICROBIAL AGENTS AND CHEMOTHERAPY, Dec. 1995, p. 2645–2649 0066-4804/95/$04.0010 Copyright q 1995, American Society for Microbiology
Vol. 39, No. 12
Effects of Free and Liposome-Encapsulated Antibiotics on Adherence of Pseudomonas aeruginosa to Collagen Type I ˙ BIETA A. TRAFNY,* MAŁGORZATA STE ´ SKA, MAŁGORZATA ANTOS, ELZ ˛ PIN AND JACEK GRZYBOWSKI Microbiology Department, Military Institute of Hygiene and Epidemiology, Warsaw, Poland Received 23 May 1995/Returned for modification 8 August 1995/Accepted 28 September 1995
The adherence of 27 clinical Pseudomonas aeruginosa strains to collagen type I was investigated by using a solid-phase assay. The influence of free antibiotics (amikacin, gentamicin, piperacillin, bacitracin, and polymyxin B) and liposome-entrapped antibiotics (amikacin and polymyxin B) on bacterial attachment to collagen type I was examined. The greatest inhibitory effect was shown for free and liposomal amikacin, which decreased the attachment of 74 and 100% of tested strains, respectively. The mean percent attachment (6 standard deviation) in the presence of free amikacin was 65.7% (6 12.0%) as measured by solid-phase assay. In the presence of liposomal amikacin, the attachment ranged from 17.3% (6 6.0%) to 42.1% (6 9.4%), depending on the antibiotic solvent. In contrast, polymyxin B, even at a subinhibitory concentration, enhanced attachment of all P. aeruginosa isolates to collagen. Liposomal polymyxin B displayed a protective effect only when the encapsulated drug was of a low concentration. Application of liposome-encapsulated amikacin may be advantageous in injured tissues in which extracellular matrix structures become exposed. Pseudomonas aeruginosa remains one of the most important agents of bacteremia and is especially associated with pneumonia, urinary tract infections, and burns. Adherence of bacteria to host cells or to extracellular matrix (ECM) materials is considered the first event in the process of colonization, often followed by tissue invasion. Fibronectin, laminin, proteoglycans, and collagens are ECM components which may be involved in binding of some gram-positive and gram-negative bacteria to host tissues (9, 23, 24). With its wide distribution, collagen type I, a major component of submucosal underlying connective tissues, is a potential receptor for bacterial adhesion for P. aeruginosa, which usually infects patients with tissue damage. It has been shown that P. aeruginosa adheres significantly to a matrix rich in type I collagen (18). Several factors may affect the bacterial adhesion process. In general, antibiotics inhibit the adhesion of bacteria, even at sub-MICs (e.g., adherence of Staphylococcus aureus to collagen), but in certain cases they may enhance adhesion (3, 17). After tissue damage, the ECM structures may become exposed to microbial colonization. Hence, the question arises how a topically applied antibiotic can modulate bacterial adhesion. The most commonly used topical antibiotics are gentamicin, neomycin, bacitracin, polymyxin B, and amikacin (1, 2, 19). In recent years, use of liposomes with encapsulated antibiotics has been proposed as an effective delivery system for the treatment of bacterial and fungal infections. Encouraging results have been obtained with liposomal ampicillin in the treatment of Listeria monocytogenes infections, liposomal gentamicin against Mycobacterium avium complex in AIDS patients, and liposomal amphotericin B administered to patients with invasive or superficial fungal infections and other microbial diseases (5, 8). In this study, we investigated the effect of free and liposomal antibiotics on adherence of clinical P. aeruginosa isolates to collagen type I.
MATERIALS AND METHODS Bacterial strains and growth conditions. The 27 P. aeruginosa clinical isolates used in this study were obtained from the collection of the Department of Clinical Bacteriology, Children Memorial Hospital, Warsaw, Poland. Seventeen of them were isolated from urinary tract infections, eight strains were from respiratory tract infections, and two were from brain abscesses. For each experiment, bacteria were subcultured on Bacto-Agar plates (Difco Laboratories, Detroit, Mich.) and grown overnight at 378C. The organisms were then harvested and resuspended in phosphate-buffered saline (PBS) (0.02 M phosphate and 0.15 M NaCl, pH 7.4) and adjusted to an A600 of 0.6, corresponding to 1 3 109 to 2 3 109 CFU/ml. The bacterial concentration was confirmed by quantitative culture on agar. All strains were nonmucoid; the purity of cultures was determined by periodic observation of colony morphology. All strains were serologically typed with commercially available antisera (Institut Pasteur, Paris, France) and represented serotypes 1, 2, 6, 8, and 11. Seven strains were polyagglutinable. Materials. Purified bovine collagen type I, extracted from tendons by limited pepsin digestion, was obtained from Tissue Bank, Kielce, Poland. The purity of the collagen preparation (95%) was estimated by hydroxyproline content analysis. Egg phosphatidylcholine (60%) (Sigma Chemical Co., St. Louis, Mo.), type X-E, was purified by chromatography on an alumina column as described by New (15) and was stored as chloroform stock solutions at 2258C under nitrogen. Hyperimmune human polyvalent gamma globulin against P. aeruginosa was kindly donated by J. Emo ¨d, Institute HUMAN for Serobacteriological Production and Research, Budapest, Hungary. Affinity-purified goat anti-human immunoglobulin G (gamma chain specific)-peroxidase conjugate was purchased from Sigma Chemical Co. Antimicrobial agents. The following antimicrobial agents were purchased from the indicated manufacturers as standard powders or solutions as follows: polymyxin B sulfate from Pfizer GmbH, Karlsruhe, Germany; bacitracin and piperacillin sodium salt from Sigma Chemical Co.; amikacin sulfate (Biodacyna) from Institute of Biotechnology and Antibiotics, Warsaw, Poland; neomycin sulfate from Polfa, Warsaw, Poland; and gentamicin sulfate from Fluka Chemie AG, Buchs, Switzerland. Stock solutions of the antibiotics were prepared from standard powders and then were diluted in PBS to achieve the desired concentrations. MIC determinations. The MICs of antibiotics for P. aeruginosa strains were determined by the broth macrodilution method using doubling dilutions prepared in 1 ml of Oxoid nutrient broth (Unipath Ltd., Basingstoke, England). The inoculum (0.05 ml) was adjusted to yield a concentration of 5 3 105 CFU/ml. The tubes were incubated at 378C for 18 h, after which the MIC was defined as the lowest concentration of antibiotic which inhibited visible growth. Liposomes. The ‘‘hand-shaking’’ method (15) was used to prepare neutral liposomes. A 100-mg sample of phosphatidylcholine and 50 mg of cholesterol (Fluka Chemie AG) were suspended in 5 ml of chloroform in a 250-ml roundbottomed flask and rotary evaporated at 308C to a thin dry film. The residual solvent was removed by vacuum evaporation for 1 h, and then 5 ml of antibiotic was added. The concentrations of added polymyxin B dissolved in PBS ranged from 0.05 to 3 mg/ml. The concentration of added amikacin dissolved either in PBS or in MSS (manufacturer’s stabilizing solution, pH 4.5, containing sodium
* Corresponding author. Mailing address: Military Institute of Hygiene and Epidemiology, Kozielska 4, 01-167 Warsaw, Poland. Phone: (48) 2-685-32-06. Fax: (48) 2-238-10-69. 2645
2646
TRAFNY ET AL.
citrate, sodium hydrogen sulfite, and sulfuric acid) was 4 mg/ml. The flask was rotated in an evaporator at 60 rpm for 30 min at room temperature and allowed to stand for 2 h. Nonentrapped material was separated by dialysis overnight against PBS or MSS. The concentration of encapsulated antibiotic was determined as follows. Liposomes were ruptured by treatment with 0.1% sodium deoxycholate, and the antibiotic content in the supernatant was evaluated by radial diffusion in agar with a lawn of P. aeruginosa ATCC 27853. The antibiotic concentration was calculated by comparison with a standard curve. Solid-phase assay for P. aeruginosa binding to collagen. The wells of 96-well flat-bottomed polystyrene plates (Dynatech Laboratories Inc., Chantilly, Va.) were coated with 0.1 ml of collagen type I dissolved in 0.1 M acetic acid at a concentration of 500 mg/ml and allowed to dry at 378C overnight. After that, the plates were extensively washed with PBS-T (PBS with 0.1% Tween 20). Bacterial solutions of individual strains at a concentration of 107 CFU/ml were incubated in the wells for 1 h. Then the plates were again extensively washed with PBS-T, and bacteria adhering to collagen were fixed with 0.3% formalin for 10 min. Detection of bound bacteria was performed with human gamma globulin anti-P. aeruginosa and anti-human immunoglobulin G-peroxidase conjugate. After enzymatic reaction with o-phenylenediamine, the A490 was measured with a microELISA reader (Dynatech Laboratories). Competitive assay. To the microtiter wells coated with 500 mg of collagen type I per ml, a bacterial suspension of each strain at a concentration of 107 CFU/ml was added together with 25 mg of soluble collagen type I per ml. After 1 h of incubation at 378C, P. aeruginosa binding to collagen type I immobilized in the wells was evaluated as described above. The results were expressed as percent inhibition, calculated as follows: % inhibition 5 100 2 % attachment (see below). Effect of free antibiotics on P. aeruginosa attachment. Experiments were performed as follows. A 100-ml volume of bacterial suspension with a density of 107 CFU/ml and 5 ml of antibiotic solution were added to collagen-coated microtiter wells. The final concentration of antibiotic in the well was 16 mg/ml. Next, the plates were incubated for 1 h at 378C. Then a solid-phase assay was performed as described above. The percent attachment compared with that of the controls was calculated by the following formula: % attachment 5 [bacteria adhering in the presence of antibiotic/bacteria adhering in the absence of antibiotic (control)] 3 100. In another set of experiments, bacteria used for the assay were grown in the presence of one-half the MIC of polymyxin B for 24 h at 378C and sedimented at 10,000 3 g for 15 min. Then the density of bacteria was adjusted to 107 CFU/ml, and the adherence solid-phase assay was performed as described above. Effect of liposomal antibiotics on P. aeruginosa attachment. A 20-ml volume of liposomal amikacin (190 mg/ml) or liposomal polymyxin B (24, 48, and 1,500 mg/ml) prepared as described above was incubated for 1 h with 0.1 ml of bacterial cells (107 CFU/ml). Then a solid-phase assay was used to detect the adhered bacteria and to calculate the percent attachment. After 1 h of incubation of liposomal antibiotics with 107 CFU of bacterial cells per ml, the number of living cells, as determined by plating, did not decrease because the stability of liposome preparation during 1 h of incubation with bacterial cells, i.e., the concentration of released antibiotic, was too low to kill bacterial cells. To establish the avidity of bacterial cell binding to liposomes, small pieces of polystyrene (2.5 by 7.5 mm) were coated with collagen type I (500 mg/ml) and incubated in bacterial suspensions (107 CFU/ml) of individual strains for 1 h at 378C. After that, the polystyrene pieces were extensively washed in PBS-T and next were shaken for 0.5 h in a suspension of liposome-encapsulated amikacin (in PBS). Organisms adhered to the pieces of collagen-coated polystyrene were detected with anti-pseudomonal gamma globulin and anti-human immunoglobulin G-peroxidase conjugate. The percentage of bacteria adhered to the collagen-coated bars was calculated in comparison with control samples (bacteria of the same strains extensively washed with PBS-T instead of liposome suspension). Statistical methods. The two-tailed Student t test was used for determining the statistical significance (P values) of differences of the means for control and experimentally treated bacteria.
RESULTS Attachment of P. aeruginosa to collagen-coated microtiter wells. In the solid-phase adherence assay, the A490 increased linearly within the range of 106 to 108 CFU/ml with increasing concentrations of bacteria added to the collagen-coated microtiter wells. A density of 107 CFU of bacterial cells per ml was chosen as a medial concentration. The mean absorbance 6 standard deviation (SD) for 27 tested strains adhering to collagen was 0.93 6 0.32. These assay conditions allowed us to observe either a decrease or an increase in bacterial attachment caused by antimicrobial agents. Liquid-phase collagen type I inhibited attachment of individual strains to solid-phase collagen type I; the mean percent inhibition 6 SD was 68.3%
ANTIMICROB. AGENTS CHEMOTHER. TABLE 1. Influence of various antibiotics on the adherence of P. aeruginosa strains to collagen-coated microtiter wells Antibiotica
No. (%) of strains adheringb
Mean % attachment 6 SDc
Amikacin Gentamicin Neomycin Piperacillin Bacitracin Polymyxin B
20 (74) 3 (11) 9 (33) 6 (22) 11 (41) 26 (96)
65.7 6 12.0 78.7 6 5.8 70.4 6 8.2 78.7 6 10.5 81.9 6 4.0 174.4 6 45.2
The concentration of each antibiotic in the microtiter wells was 16 mg/ml. Twenty-seven strains were tested. Numbers indicate strains for which differences in A490 readings in the presence and in the absence of antibiotic (control) were statistically significant (P , 0.05). c Mean percent attachment (6 SD) of strains for which attachment was affected by antibiotics (compared with untreated controls, which had 100% attachment) in three separate experiments, each performed in duplicate. a b
6 9.1% (compared with attachment of control samples, in which bacteria were not blocked by soluble collagen). Modulation of P. aeruginosa attachment by antimicrobial agents. Six antibiotics were used in adherence experiments (Table 1). Gentamicin, neomycin, amikacin, bacitracin, and polymyxin B were included in the study as antibiotics which might be applied topically. Piperacillin, which is very effective against P. aeruginosa, was studied to compare the actions of different antibiotics. The greatest inhibitory effect on adherence of P. aeruginosa isolates was produced by amikacin (74% of sensitive strains). The lowest mean percent attachment was observed also for the strains treated with amikacin. Gentamicin and neomycin had inhibitory effects on three (11%) and nine (33%) of the tested P. aeruginosa strains, respectively. Sensitivity to the inhibitory action of aminoglycosides was not dependent on the serotypes of tested strains or on the sources of clinical strains (data not shown). Piperacillin affected adherence to collagen of 22% of P. aeruginosa strains. Bacitracin is applied topically as a drug effective against gram-positive bacteria. The P. aeruginosa strains were resistant to it. Eleven strains (41%) showed a decrease in attachment to collagen in the presence of this antibiotic; however, the reduction of adherence caused by bacitracin was the lowest among the tested antibiotics. In the presence of polymyxin B, the number of adhering bacteria increased significantly, and this phenomenon was observed for the majority of tested strains. All strains tested had different susceptibilities to the antibiotics examined, expressed as MICs. No correlation between the MIC and inhibition of adherence by appropriate antibiotics was observed. For instance (Fig. 1), amikacin even at a concentration of 4 mg/ml had no effect on the adherence of one isolate, but for another isolate an MIC of 1 mg of amikacin per ml was sufficient to block adherence (up to 50%). A similar phenomenon was noticed also for gentamicin and piperacillin (data not shown). Inhibition of adherence by aminoglycosides was to some extent a dose-dependent phenomenon (Fig. 2). Amikacin at a concentration of 1 mg/ml did not affect the adherence of any strains tested. The inhibitory effect was observed only when the antibiotic was used at concentrations of 16 mg/ml or higher. The MICs of polymyxin B for tested strains ranged from 0.12 to 0.50 mg/ml, so the concentration of polymyxin B used in experiments exceeded the MICs for some strains many times over. To elucidate if such an excess of antibiotic molecules could be responsible for the observed increase in adherence, the solid-phase assay was performed with organisms grown in the presence of one-half the MIC of polymyxin B. As displayed
VOL. 39, 1995
ANTIBIOTICS AND P. AERUGINOSA ADHERENCE TO COLLAGEN
FIG. 1. Effect of amikacin on the adherence of P. aeruginosa isolates in comparison with MICs. Bacteria were incubated with 16 mg of amikacin per ml in collagen-coated microtiter wells. h, adherence; s, MIC. Results shown are means of three separate experiments, each performed in duplicate.
in Fig. 3, a significant increase in the number of adhering organisms also was observed. One-half the MIC of antibiotic added to the wells during a 1-h adherence assay had no significant effect on attachment (except for one strain). Effect of liposomal amikacin. As shown in Table 2, ‘‘empty’’ liposomes (containing PBS) had no effect on adherence of the majority of tested strains. The encapsulation of amikacin altered the effect of liposomes on adherence of P. aeruginosa to collagen-coated microtiter wells. For all strains, diminished numbers of adhered organisms were seen in the presence of liposomal amikacin. Liposomes with encapsulated amikacin in MSS were able to protect collagen almost completely from adherence of bacterial cells of all strains examined. Also, an important influence on the protective capabilities of liposomes was the encapsulation of acidic solution (MSS) inside vesicles. The assay performed with collagen-coated polystyrene
2647
FIG. 3. Attachment of P. aeruginosa strains to collagen-coated microtiter plates: influence of sub-MIC of polymyxin B. Adherence assays were performed for 1 h with bacteria grown in the absence of antibiotic and assayed either in the presence of one-half the MIC of polymyxin B (z) or in the absence of antibiotic (h) and with bacteria grown in the presence of one-half the MIC of polymyxin B, washed extensively in PBS, and assayed in the absence of drug (s). Results shown are means 6 SDs for four samples. p, P , 0.01.
pieces, to which P. aeruginosa isolates had adhered for 1 h and which had then been extensively washed with liposomal amikacin (in PBS), demonstrated that about one-third of adhering organisms were detached in 0.5 h by liposomal amikacin (mean percent attachment 6 SD, 66.8% 6 17.3%) in comparison with control samples (attachment of bacteria washed with PBS was assumed as 100%). Effect of liposomal polymyxin B. We next investigated the capacity of liposomal polymyxin B at three different concentrations to modulate P. aeruginosa adherence (Table 3). An inhibitory effect was observed only when the concentration of liposomal polymyxin B was 4 mg/ml. Liposomal polymyxin B at a concentration of 8 mg/ml caused a significant increase in the number of adhering organisms and affected 50% of tested strains. All 27 strains examined were sensitive to the modulatory action of liposomal polymyxin B only when the drug concentration was 250 mg/ml.
TABLE 2. Effect of liposomal amikacin on the adherence of P. aeruginosa isolates to collagen-coated microtiter wells
FIG. 2. Attachment of P. aeruginosa strains in the presence of amikacin. Bacteria were incubated in collagen-coated microtiter wells with the indicated concentrations of amikacin for 1 h. Data are means 6 SDs for triplicate samples of triplicate experiments for each of 27 tested strains and are expressed as percent attachment relative to that of controls (bacteria incubated in the absence of antibiotic 5 100% attachment). p, P , 0.01 (differences were estimated in comparison with results obtained in the presence of 1 mg of amikacin per ml).
Content of liposomes
No. (%) of strains adheringa
Mean % attachment 6 SDb
PBS (control) Amikacin in PBS Stabilizing solutionc (control) Amikacin in stabilizing solutionc
2 (7) 27 (100) 27 (100) 27 (100)
83.1 6 1.9 42.1 6 9.4 30.7 6 14.6 17.3 6 6.0
a Twenty-seven strains were tested. Numbers indicate strains for which differences in A490 readings in the presence and in the absence of liposomal formulation (control) were statistically significant (P , 0.05). b Mean percent attachment (6 SD) of liposome-treated isolates compared with untreated isolates (100% attachment) in two separate experiments, each performed in triplicate. c MSS; see Materials and Methods.
2648
TRAFNY ET AL.
ANTIMICROB. AGENTS CHEMOTHER.
TABLE 3. Effect of liposomal polymyxin B on the adherence of P. aeruginosa isolates to collagen-coated microtiter wells Concn of entrapped polymyxin B (mg/ml)
No. (%) of strains adheringa
Mean % attachment 6 SDb
4 8 250
15 (56) 13 (48) 27 (100)
67.7 6 13.3 148.5 6 32.8 193.5 6 45.8
a
See Table 2, footnote a. Mean percent attachment (6 SD) of liposome-treated isolates compared with untreated isolates (100% attachment) in two separate experiments, each performed in triplicate. b
DISCUSSION Exposed ECM proteins may play a role as a target for bacterial attachment. Thus, it is reasonable to compare the affinities of P. aeruginosa clinical strains for binding to collagen type I (one of the most abundant matrix proteins) in the presence of antibiotics applied to protect injured tissues from bacterial infection. Various systems have been employed to investigate bacterial interactions and modulation of these interactions with ECM proteins (6). Among them, the microtiter plate adherence assay seems to be easy to perform and reproducible. In spite of its semiquantitative assessment of the binding data, this type of assay allowed demonstration of the inhibitory effects of different agents on bacterial adherence (20, 21, 25). A solid-phase adherence assay applied in our studies was performed in an excess of collagen molecules accessible on the plates. The results of the assay were dependent on the number of adhered P. aeruginosa cells. We assume that binding of organisms to collagen immobilized on the plates was specific because it was inhibited (;70%) by the liquid-phase collagen in the competitive assay. Our data illustrated that antimicrobial agents exert significant effects on adherence of P. aeruginosa isolates to collagencoated microtiter plates. For five antibiotics (amikacin, neomycin, gentamicin, piperacillin, and bacitracin), the ability to decrease P. aeruginosa adherence to immobilized collagen molecules was demonstrated, but their influence varied with the antibiotic concerned. Among aminoglycosides, amikacin displayed the most profound effect on adherence of P. aeruginosa; 74% of tested strains were sensitive to its action. Under the same conditions, gentamicin inhibited adherence of 11% of strains, whereas neomycin inhibited adherence of 33% of strains. Aminoglycosides, with their cationic nature, perturb the overall architecture of the bacterial outer membrane by competing with divalent cations. These cations are necessary to keep lipopolysaccharides (LPSs) and other phospholipid molecules properly packed in the membrane (10). Replacement of those cations results in distortion of the outer membrane and may change bacterial adhesive properties. Proper interpretation of the observed differences in aminoglycoside-induced inhibition of adherence requires further studies. We believe that differences in aminoglycoside structure and in the arrangement of the positive charges on the antibiotic molecules are responsible for different drug affinities for LPS molecules and the observed differences in adherence. Similar reasoning may be used to explain the variability among strains in their sensitivity to aminoglycoside-induced inhibition of adherence. LPSs on the intact bacteria are heterogeneous in the length of O antigen attached to the core and may differ in affinity of cation binding. Mai and colleagues (11) demonstrated that in the presence
of aminoglycosides such as amikacin, tobramycin, and gentamicin, the adherence of P. aeruginosa to Dacron fiber does not decrease. However, the investigated process of adhesion was unspecific, as Dacron fiber is a biologically inert substrate, and it is difficult to make the comparison with adherence to collagen. Mai et al. also did not observe the correlation between the MIC for the organism and the adherence-inhibitory effect of the drug tested. Our results confirm this finding; for amikacin, gentamicin, and piperacillin, no such correlation was seen. It is surprising to us that bacitracin inhibited the adherence of so many isolates although this antibiotic is not active as an antipseudomonal agent. The outer membrane of gram-negative bacteria is an effective permeability barrier for bacitracin. Therefore, bacitracin interference with P. aeruginosa adhesion to collagen may be related to its action on the bacterial surface. This explanation is intuitive and needs to be proved experimentally. Polymyxins also compete for divalent cation-binding sites within the outer membrane (4), and their overall effect resembles that of aminoglycosides. Therefore, one can assume that the two types of antibiotics possess similar adherenceinhibitory properties. However, in contrast to aminoglycosides, polymyxin B even at subinhibitory concentrations enhanced attachment of all tested P. aeruginosa isolates to collagen in our experiments. A possible interpretation of this finding is that adherence of bacteria to collagen in the presence of polymyxin B is unspecific and may be caused by increased bacterial hydrophobicity. Data for decreasing cell surface charge and increasing cell hydrophobicity after P. aeruginosa treatment with polymyxin B nonapeptide have been presented recently (12). Another mechanism of unspecific adherence may be deduced from the extremely high affinity of polymyxin B binding to bacterial LPS. Drug molecules form a stable complex with the lipid A region of bacterial LPS (13). Polymyxin B complexed with LPS of P. aeruginosa remains bound even throughout purification (4). Such a stable LPS-polymyxin B complex may exist on the surface of the adhering bacterial cell, increasing its affinity for binding to collagen molecules; polymyxin B is known to bind to host tissues (22). We have some preliminary observations to support this hypothesis, but they have yet to be carefully confirmed. We wondered if encapsulation of amikacin and polymyxin B inside liposome vesicles would affect their influence on bacterial attachment to collagen. Empty liposomes (containing PBS) proved to be ineffective in protection of collagen molecules from adherence of P. aeruginosa isolates. It has been shown previously (7) that phosphatidylglycerol liposomes decrease the total number of P. aeruginosa organisms adhering to respiratory epithelial cells. It was proposed that the observed protection was due to masking of epithelial receptors or blocking of bacterial adhesions by neutral phospholipids. Our liposomal preparations contained phosphatidylcholine and cholesterol but were ineffective in protection of collagen unless polycationic amikacin or acidic MSS was encapsulated. We believe that in addition to the vesicles containing amikacin or acidic molecules, liposomes were coated with some of these molecules, as has been previously shown for gentamicin (14). Those external molecules supply a positive charge to vesicles. Thus, interaction of bacterial cells with liposomes may take place as a result of physical attractive forces. Such a mechanism of bacterium-liposome interaction has been proposed for liposomes with encapsulated ciprofloxacin (16). Data from this study demonstrated that the interaction of liposomes with encapsulated amikacin (in PBS) with P. aeruginosa isolates is of high affinity. Liposomal amikacin was able to detach about one-third of adhering bacteria from immobilized collagen molecules.
ANTIBIOTICS AND P. AERUGINOSA ADHERENCE TO COLLAGEN
VOL. 39, 1995
Encapsulated liposomal polymyxin B at concentrations above 8 mg/ml caused visible enhancement of attachment of P. aeruginosa strains. Polymyxin B, like other peptides, is probably incorporated inside the liposome membrane (15). It may be speculated that the release of antibiotic and its action on the bacterial cell are related to disruption of liposomal membranes which abolished the protective capacities of positively charged liposomes. In conclusion, the antimicrobial agents tested in this study decreased the attachment of P. aeruginosa to collagen, except for polymyxin B. Liposomal amikacin significantly protected collagen molecules. The observations described above need to be investigated in more detail to elucidate the mechanism of adherence inhibition induced by free and liposomal antibiotics. ACKNOWLEDGMENT This work was supported by grant 0T00A 020 08 (to E.A.T.) from Committee of Scientific Research (KBN), Poland. REFERENCES 1. Baxter, C. R., and G. T. Rodeheaver. 1990. Interventions: hemostasis, cleansing, topical antibiotics, debridement and closure, p. 71–82. In New directions in wound healing. E. R. Squibb & Sons, Inc., Princeton, N.J. 2. Boyce, S. T., A. P. Supp, G. D. Warden, and I. A. Holder. 1993. Attachment of an aminoglycoside, amikacin, to implantable collagen for local delivery in wounds. Antimicrob. Agents Chemother. 37:1890–1895. 3. Butcher, W. G., J. Close, D. Krajewska-Pietrasik, and L. M. Switalski. 1994. Antibiotics alter interactions of Staphylococcus aureus with collagenous substrata. Chemotherapy 40:114–123. 4. Conrad, R. S., and C. Galanos. 1989. Fatty acid alterations and polymyxin B binding by lipopolysaccharides from Pseudomonas aeruginosa adapted to polymyxin B resistance. Antimicrob. Agents Chemother. 33:1724–1728. 5. Coune, A. 1988. Liposomes as drug delivery system in the treatment of infectious diseases. Potential applications and clinical experience. Infection 16:141–147. 6. Doig, P., and T. J. Trust. 1993. Methodological approaches of assessing microbial binding to extracellular matrix components. J. Microbiol. Methods 18:167–180. 7. Girod de Bentzmann, S., O. Bajolet-Laudinat, F. Dupuit, D. Pierrot, C. Fuchey, M. C. Plotkowski, and E. Puchelle. 1994. Protection of human respiratory epithelium from Pseudomonas aeruginosa adherence by phosphatidylglycerol liposomes. Infect. Immun. 62:704–708. 8. Gregoriadis, G., and A. T. Florence. 1993. Liposomes in drug delivery. Clinical, diagnostic and ophthalmic potential. Drugs 45:15–28.
2649
9. Isaaks, R. D. 1994. Borrelia burgdorferi bind to epithelial cell proteoglycans. J. Clin. Invest. 93:809–819. 10. Kadurugamuwa, J. L., A. J. Clarke, and T. J. Beveridge. 1993. Surface action of gentamicin on Pseudomonas aeruginosa. J. Bacteriol. 175:5798–5805. 11. Mai, G. T., J. G. McCormack, W. K. Seow, G. B. Pier, L. A. Jackson, and Y. H. Thong. 1993. Inhibition of adherence of mucoid Pseudomonas aeruginosa by alginase, specific monoclonal antibodies, and antibiotics. Infect. Immun. 61:4338–4343. 12. Morris, C. M., A. George, W. W. Wilson, and F. R. Champlin. 1995. Effect of polymyxin B nonapeptide on daptomycin permeability and cell surface properties in Pseudomonas aeruginosa, Escherichia coli, and Pasteurella multocida. J. Antibiot. 48:67–72. 13. Morrison, D. C., and D. M. Jacobs. 1976. Binding of polymyxin B to the lipid A portion of bacterial lipopolysaccharides. Immunochemistry 18:813–818. 14. Nacucchio, M. C., M. J. Gatto Belora, D. O. Sordelli, and M. d’Aquino. 1988. Enhanced liposome-mediated activity of piperacillin and gentamicin against gram-negative bacilli in vitro. J. Microencapsulation 5:303–309. 15. New, R. R. C. 1990. Preparation of liposomes, p. 33–103. In R. R. C. New (ed.), Liposomes, a practical approach. Oxford University Press, New York. 16. Nicholov, R., A. E. Khoury, A. W. Bruce, and F. DiCosmo. 1993. Interaction of ciprofloxacin loaded liposomes with Pseudomonas aeruginosa cells. Cells Mater. 3:321–326. 17. Piatti, G. 1994. Bacterial adhesion to respiratory mucosa and its modulation by antibiotics at sub-inhibitory concentrations. Pharm. Res. 30:289–299. 18. Plotkowski, M. C., M. Chevillard, D. Pierrot, D. Altemayer, J. M. Zahm, G. Colliot, and E. Puchelle. 1991. Differential adhesion of Pseudomonas aeruginosa to human respiratory epithelial cells in primary culture. J. Clin. Invest. 87:2018–2028. 19. Riegels-Nielsen, P., F. Espersen, L. R. Holmich, and N. Frimodt-Moller. 1995. Collagen with gentamicin for prophylaxis of postoperative infection. Staphylococcus osteomyelitis studied in rabbits. Acta Orthopaed. Scand. 66:69–72. 20. Schadow, K. H., W. A. Simpson, and G. D. Christensen. 1988. Characteristics of adherence to plastic tissue culture plates of coagulase-negative staphylococci exposed to subinhibitory concentrations of antimicrobial agents. J. Infect. Dis. 157:71–77. 21. Sexton, M., and D. J. Reen. 1992. Characterization of antibody-mediated inhibition of Pseudomonas aeruginosa adhesion to epithelial cells. Infect. Immun. 60:3332–3338. 22. Sixl, F., and H. J. Galla. 1981. Polymyxin interaction with negatively charged lipid bilayer membranes and the competitive effect of Ca21. Biochim. Biophys. Acta 643:626–635. 23. Umemoto, T., and I. Namikawa. 1994. Binding of host-associated treponeme proteins to collagens and laminin: a possible mechanism of spirochetal adherence to host tissues. Microbiol. Immunol. 38:655–663. 24. Westerlund, B., and T. K. Korhonen. 1993. Bacterial proteins binding to the mammalian extracellular matrix. Mol. Microbiol. 9:687–694. 25. Zhang, Q., T. F. Young, and R. F. Ross. 1994. Microtiter plate adherence assay and receptor analogs for Mycoplasma hyopneumoniae. Infect. Immun. 62:1616–1622.
