INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Enterotoxigenic Escherichia coli (ETEC) strains are frequent causes of piglet diarrhea during the preweaning and immediate postweaning periods. Among the different ETEC strains (K88-, K99-, or 987P-expressing strains), those expressing K88 fimbrial antigen are the most prevalent (1, 13). These fimbriae mediate the adhesion of E. coli K88-expressing strains to the intestinal epithelial mucosa and also to the mucus layer lining the small intestine (7), and thereafter the organism elaborates one or two enterotoxins, heat-stable toxin and heat-labile toxin, which induce massive fluid and electrolyte secretion into the gut lumen (1, 13). Among ETEC variants expressing the ab, ac, or ad K88 fimbriae, those possessing the K88ac fimbrial antigen are the most common variant found in pathogenic E. coli isolates in the United States (26). Antibiotics are routinely used in an attempt to control pathogens, but the organisms are becoming resistant to the more commonly used treatments, making antibiotic therapy unreliable (10). Furthermore, the use of antimicrobial growth promoters may cause the development of resistance in a number of important pathogenic bacterial species (28). Recently, the European Union decided to ban the use of four widely used antibiotics, i.e., tylosin, virginiamycin, spiramycin, and zinc bacitracin, as growth promoters from July 1999. As a consequence, there is an urgent need to seek an alternative to antibiotics for the purpose of enhancing the health status and production performance of domestic animals.

Probiotics have been used to reduce the colonization of the intestines of animals by pathogens. This, in turn, reduces the prophylactic use of antibiotics as feed additives in animal production (12, 15, 23). Most probiotic bacteria are of intestinal origin and belong to the lactic acid-producing bacteria (LAB) such as bifidobacteria, lactobacilli, and enterococci. Many studies have focused on lactobacilli and bifidobacteria and have been carried out to elucidate the mechanism of bacterial adhesion and the ability of these bacteria to inhibit the adhesion of pathogens to the intestinal mucosa. However, there are very few studies, if any, on the adhesion of enterococci and their inhibition of the adhesion of pathogens to the intestine, although more than 60% of probiotic preparations in the market contain strains of enterococci (9, 11, 25). Recently, Netherwood et al. (21) reported that probiotics containing both genetically modified (GM) and non-GM Enterococcus faecium changed the bacterial flora of the chicken gastrointestinal tract, but with conflicting results: the relative amount of E. faecalis in the total eubacterial population increased in the presence of the non-GM strain and decreased in the presence of the GM probiotic was used compared with the results obtained with an untreated control group (21).

It has been established that receptors for the K88 fimbriae are detectable in the mucus overlying the epithelial cells of the piglet small intestine (17). The presence of receptors in mucus may play a role in the pathogenesis of K88-bearing strains (7). The objective of the present study was to investigate if E. faecium 18C23 has the ability to inhibit the adhesion of E. coli K88ac and K88MB to the small intestinal mucus of piglets.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Bacteria and culture conditions. Three strains of LAB, namely, E. faecium 18C23, Lactobacillus acidophilus 80H10, and Lactobacillus casei subsp. rhamnosus 47G19, were used. These bacteria were originally isolated from cottage cheese and obtained from StarLab Inc. (St. Joseph, Mo.). E. faecium 18C23 was grown in KF Streptococcus broth (Difco, Detroit, Mich.), and the Lactobacillus strains were cultured in MRS broth (Difco). ETEC K88ac was obtained from the Pennsylvania State University E. coli Reference Center (University Park, Pa.). A strain of the hemolytic ETEC K88+ bacterium (K88+MB) was obtained from the Animal Health Center, Veterinary Services Branch, Manitoba Agriculture, Winnipeg, Canada. The isolate, which was obtained from the feces of a piglet suffering from diarrhea, was shown to give a positive agglutination test with anti-K88 ETEC antibody but not with anti-K99 ETEC antibody. E. coli K88+MB was tested as the K88ac phenotype (16). Primary cultures of ETEC strains were grown overnight in tryptic soy broth (Difco) at 37°C using a 1% innoculum from stocks stored at 20°C. All strains of bacteria were stored at 20°C in the appropriate medium containing 30% glycerol.

Preparation of labeled bacteria. Both LAB and E. coli K88 were labeled as described by Laux et al. (18). E. coli K88 was grown overnight and LAB were incubated for 10, 15, 25, and 40 h at 37°C in their respective media containing 10 µCi of [methyl-1,2-³H]thymidine (118 Ci mmol¹; Amersham International, Little Chalfont, United Kingdom). Cells were harvested by centrifugation at 2,000 × g for 15 min, washed twice in phosphate-buffered saline (PBS; pH 7.2), and resuspended in PBS. The suspension of ETEC K88ac and K88MB was adjusted to an optical density of 1.0 at 600 nm (approximately 10⁹ CFU ml¹) and was used in the following tests.

Preparation of small intestine mucus from piglets. Four 14 ± 2-day-old healthy Cotswold piglets were obtained from the Glenlea Swine Research Unit, University of Manitoba, Winnipeg, Canada. The piglets used in this study were of the K88-susceptible phenotype, since K88-bearing E. coli MB was able to bind to the intestinal epithelial cells when tested microscopically using the procedure described by Wilson and Hohmann (27). The small intestine mucus used in the present study was isolated from piglets by the procedure of Jin et al. (16). Briefly, a microscope slide was used to gently scrape the mucosal surface into HEPES-Hanks' buffer (HH buffer; pH 7.4). The mucosal scrapings were then centrifuged twice at 27,000 × g at 4°C for 15 min to remove solids. The supernatant, containing crude intestinal mucus, was then analyzed for protein content by the method of Lowry et al. (20). Bovine serum albumin (BSA) was used as a standard.

Adhesion assay. The adhesion of radioactively labeled bacteria to the intestinal mucus receptor was studied by using a modification of the methods of Blomberg et al. (3). All adhesion assays were performed in triplicate in multiwell polythene tissue culture plates (Linbro, flat bottom, 1.6 cm in diameter; Nalge Nunc International, Roskilde, Denmark). Briefly, the intestinal mucus was immobilized overnight at 4°C in wells at 0.5 mg of protein per ml. Control wells were immobilized with bovine serum albumen (BSA, Sigma, St. Louis, Mo.) at 1 mg of protein per ml or with 0.2 ml of PBS as a polythene control. After immobilization, the wells were washed twice with 0.5 ml of HH buffer (pH 7.4) containing 0.5% mannose (Sigma), 0.2 ml of [methyl-1,2-³H]thymidine-labeled LAB or E. coli was added to each well, and the plates were incubated for 1 h at 37°C. The wells were washed three times with 1.0 ml of HH buffer. Adherent bacteria were recovered by adding 0.5 ml of 0.5% sodium dodecyl sulfate (SDS). The samples, after incubation for 3 h at 37°C, were collected and mixed thoroughly with 10 ml of scintillation cocktail (ICN Biomedicals Inc., Aurora, Ohio). The radioactivity of the SDS-extracted sample was enumerated using a liquid scintillation spectrophotometer (LKB-Wallac RackBeta).

Characterization of the intestinal mucus receptor for E. faecium adhesion. To examine the nature of the intestinal mucus receptor for E. faecium adhesion, three proteolytic enzymes (Sigma), i.e., pronase, proteinase K, and trypsin (0.2 ml of solutions containing 800 µg of enzyme per ml), were added individually to wells containing immobilized mucus and the plates were incubated for 2 h at 37°C and then overnight at 4°C. The plates were then washed, and the adhesion assay was performed with E. faecium 18C23 culture incubated for 25 h. Controls were the immobilized mucus that was treated with BSA (800 µg ml¹) rather than enzymes. Immobilized mucus on tissue culture wells was also treated with 0.2 ml of 0.01 M sodium metaperiodate or 0.2 ml of 0.01 M sodium iodate in 0.2 M sodium acetate buffer (pH 4.5), incubated in the dark for 3 h at 4°C, washed twice with HH buffer, and assayed for bacterial adhesion.

Assay of inhibition of adhesion. The assay of inhibition of adhesion was similar to that carried out previously (16). The E. faecium culture was prepared by incubation in KF Streptococcus broth for 25 h at 37°C. The final concentration of E. faecium was approximately 5 × 10⁸ CFU ml¹. The intestinal mucus was immobilized overnight at 4°C in wells at a concentration of 0.5 mg of protein per ml. After immobilization, the wells were washed twice with 0.5 ml of HH buffer (pH 7.4) containing 0.5% mannose (Sigma). For the competition test, suspensions of the broth cultures of E. faecium (0.1 ml) and E. coli K88ac (0.1 ml) or K88MB (0.1 ml) were added simultaneously to the immobilized mucus and the mixture was incubated for 60 min at 37°C. For the displacement test, E. coli was added to immobilized mucus and incubated for 60 min. After the mucus was washed three times with 0.5 ml of HH buffer, 0.2 ml of the suspension of E. faecium was added and the mixture was incubated for another 60 min at 37°C. After the reaction, the plates were washed three times with HH buffer (1 ml). Adherent bacteria were recovered by adding 0.5 ml of 0.5% SDS. Attachment of labeled E. coli to mucus in wells treated for the same period with sterile KF broth (control for culture and supernatant inhibition) and PBS buffer (control for inhibition by bacterial cells) was taken as 100%. The inhibition of adhesion was assessed as the percentage of radioactivity in the test culture relative to the respective control. All assays were performed three times each in duplicate.

1

Statistical analysis. The data were analyzed using the SAS program (24). Tukey's test was used to identify differences among groups after analysis of variance. In the dose-dependent inhibition test, data were analyzed using regression with dummy coding.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Bacterial adhesion to small intestine mucus receptors. The three strains of LAB, i.e., E. faecium 18C23, L. acidophilus 80H10, and L. casei 47G19, were tested for their ability to adhere to the immobilized intestinal mucus of piglets, and the results obtained using the LAB culture incubated for 25 h are presented in Fig. 1. E. coli K88ac, which showed a strong ability to attach to the intestinal mucus in our previous study (16), was used as a positive control, while L. acidophilus I26, which was isolated from the chicken intestine (14), was used as a negative control. The result showed that the adhesion of ³H-labeled E. faecium 18C23 to the small intestine mucus was much higher than the adhesion of L. acidophilus 80H10 and L. casei 47G19 (Fig. 1). Similar patterns were obtained using LAB cultures incubated for 10, 15, and 40 h (data not shown). Approximately 9% of E. faecium 18C23 adhered to the small intestine mucus, while less than 1.5% of the other two strains did so. The adhesion of E. faecium 18C23 to the small intestine mucus was 8.5-fold higher than that to BSA, while the adhesion of L. acidophilus 80H10 and L. casei 47G19 was similar to (0.5-fold to 1.5-fold higher) their adhesion to BSA. A similar attaching pattern of these strains to polythene was also observed. The ratio of E. faecium 18C23 attachment to mucus and to polythene was much higher than the ratios for L. acidophilus 80H10 and L. casei 47G19 (Fig. 1). These results indicate that the binding of E. faecium 18C23 to the piglet intestinal mucus receptor was specific. Since incubation for 25 h yielded the highest attachment of E. faecium 18C23 to the mucus receptor (1.5 to 1.8 times higher than the values for cultures incubated for shorter or longer periods), this incubation time was used in the following experiments unless stated otherwise.

View larger version (17K): [in this window] [in a new window]   FIG. 1.   Relative adhesion of bacteria to the intestinal epithelial mucus. The result is expressed as the ratio between the number of bacteria attached to mucus and the number attached to BSA albumin (open bars) or polythene (solid bars). LA, L. acidophilus 80H10; LC, L. casei 47G19; EF, E. faecium 18C23; I26, L. acidophilus I26 (a chicken isolate, as a negative control); K88ac, E. coli K88ac (as a positive control).

In vitro competitive exclusion study. Living E. faecium 18C23 culture efficiently inhibited the adhesion of E. coli K88ac and K88MB to the piglet intestine mucus receptor. Inhibition of adhesion of E. coli K88ac to the small intestine mucus receptor was found to be dose dependent (Fig. 2). Inhibition of >90% was observed with 10⁹ CFU or more of living E. faecium 18C23 culture when E. faecium 18C23 and E. coli were simultaneously added to immobilized mucus. The adhesion-inhibiting effects of the E. faecium 18C23 culture declined dramatically when the E. faecium 18C23 culture was used at 10⁷ CFU ml¹ or less. In addition, the E. faecium 18C23 culture was more effective in inhibition of adhesion of K88MB than of K88ac (P = 0.02 [Fig. 2]). On the other hand, in the displacement test, there was no significant reduction in the adhesion of E. coli K88ac and K88MB to mucus with culture, washed cells, or supernatant (data not shown).

View larger version (12K): [in this window] [in a new window]   FIG. 2.   Inhibition of adhesion of E. coli K88ac (open circles) and K88MB (open squares) to piglet intestinal mucus by E. faecium 18C23. The results are expressed as a percentage of E. coli K88 binding to mucus relative to the control (treated with PBS buffer [pH 7.0]).

Characterization of inhibitory substances. The bacterial cells, supernatant, or original culture of E. faecium 18C23 were tested separately to identify which part was involved in the inhibition. The original EF18C23 culture, washed bacterial cells, or culture supernatant remarkably reduced the attachment of E. coli K88ac and K88MB to the intestinal mucus receptor (Table 2). Compared to the original EF18C23 culture, washed cells or culture supernatant inhibited the adhesion of E. coli to a lesser degree. However, this inhibiting activity was recovered after the washed cells and the supernatant were recombined, with the recombined mixture having a similar inhibitory ability to that of the original E. faecium 18C23 culture.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

The genus Enterococcus is a member of the normal microflora of the alimentary tracts of pigs. E. faecium, E. faecalis, E. hirae, and E. cecorum are the most frequently found species in the swine intestine (8). Together with the other LAB, i.e., lactobacilli and bifidobacteria, enterococci are widely used in probiotic products (9, 11). The results of the present study showed that the culture of E. faecium 18C23, washed bacterial cells or its supernatant, inhibited the adhesion of E. coli K88ac or K88MB to the small intestine mucus of piglets. Somewhat surprisingly, this is, to our knowledge, the first report of its kind. With other LAB such as lactobacilli or bifidobacteria, similar results have been reported, that LAB inhibits the attachment of pathogenic bacteria in both in vitro and in vivo (2-4, 12). Conway (6) found that the adhesion of pathogenic E. coli to piglet ileal epithelial cells can be inhibited by pretreating the ileal cells with whole Lactobacillus cells. Blomberg et al. (3) also reported that three Lactobacillus strains of porcine origin reduced the adhesion of E. coli K88ab and K88ac by approximately 50% in an in vitro study.

The inhibitory effects of LAB on pathogens have been attributed to steric hindrance of binding sites (22), pH values (19), or certain components of the lysed cell wall (19, 22). The results of the present study support the idea that E. faecium 18C23 cells or the substances released in its culture might occupy the binding sites, although these binding sites are not necessarily the same epitope as that for E. coli K88. The mucus receptor for E. coli K88 is involved with glycoprotein and had been characterized as an 80-kDa protein in our previous study (17). Apparently the nature of the mucus receptor for E. faecium 18C23 seems to be a glycoprotein, at least in part based on the results that (i) treatment with trypsin reduced the adhesion of this organism, indicating the protein nature of the receptor, and (ii) treatment with metaperiodate decreased the adhesion of E. faecium 18C23 to the mucus receptor, also indicating the carbohydrate is involved with the attachment to the receptor. However, the mucus receptor for E. faecium 18C23 may not be same as that for E. coli K88ac, because treatment of intestinal mucus with pronase and proteinase reduced the adhesion of E. coli K88ac (17) but increased the adhesion of E. faecium 18C23 (see above). This result may imply that E. coli and E. faecium 18C23 might not attach to the same domain of receptor. Therefore, the inhibition of adhesion of E. coli K88ac by E. faecium 18C23 or its supernatant might be through steric hindrance. This type of inhibition has been previously proposed by Ouwehand and Conway (22) for the inhibition of E. coli K88 by Lactobacillus spp. or the compounds released into their cultures, by Chauviere et al. (5) for the inhibition of human ETEC adhesion by heat-killed L. acidophilus cells, by Bernet et al. (2) for the inhibition of cell attachment and invasion of enterovirulent bacteria by L. acidophilus and its spent culture supernatant fluid, and by Chan et al. (4) for the inhibition of adhesion of uropathogenic E. coli by Lactobacillus whole cells or their cell wall fragments.

The culture pH has been proposed to be an important factor in the inhibition of adhesion of pathogens to the mucosa (19). In the present study, the inhibiting effect was not solely a pH effect since considerable inhibitory action was demonstrated after neutralizing the mixture or spent culture supernatant to pH 7.0. In addition, citrate buffer at pH 4.0 or acidified fresh KF broth at pH 4.0 did not reduce the adhesion of E. coli K88ac or K88MB to the intestinal mucus receptor. The results disagree with the results of Lehto and Salminen (19), who reported that inhibition of Salmonella enterica serovar Typhimurium adhesion to Caco-2 cell cultures by Lactobacillus strain GG spent culture supernatant was only a pH effect. It should be noted that the two studies used different intestinal receptors (human Caco-2 cell line versus porcine mucus) and different LAB (Lactobacillus versus E. faecium). A synergic inhibition effect by the E. faecium 18C23 cells and the spent culture supernatant was obtained in the present study. The inhibition effect became weaker when using the E. faecium 18C23 cells or the spent culture supernatant individually compared with the original E. faecium 18C23 culture. However, a similar inhibitory ability with the original E. faecium 18C23 culture was observed when the washed cells and the spent culture supernatant were recombined. This result may imply that the substances from both E. faecium 18C23 cells and the spent culture supernatant contribute to the inhibition of adhesion of E. coli K88 to the small intestine mucus receptors.

In conclusion, the present study found that (i) living E. faecium 18C23 efficiently inhibited the adhesion of E. coli K88ac and K88MB to the piglet intestinal mucus and that the inhibition of adhesion of E. coli K88ac to the mucus was dose dependent; (ii) the substances from both the intact E. faecium 18C23 cells and the spent culture supernatant contribute to the inhibition of adhesion of E. coli K88 to the small intestine mucus; (iii) the inhibition of adhesion of E. coli K88ac by E. faecium 18C23 or its supernatant might occur by steric hindrance and altered pH; and (iv) the results of this in vitro study warrant further testing to determine if E. faecium 18C23 can prevent diarrhea caused by E. coli K88, possibly by inhibiting the colonization of the host intestine by these pathogens.