Inactivation of Adhesion and Invasion of Food-Borne Listeria monocytogenes by Bacteriocin-Producing Bifidobacterium Strains of Human Origin

Abstract

Three bacteriocin-producing bifidobacterial isolates from newborns were identified as Bifidobacterium thermacidophilum (two strains) and B. thermophilum (one strain). This study was undertaken to evaluate the ability of these strains to compete with food-borne Listeria monocytogenes for adhesion and invasion sites on Caco-2 and HT-29 cells. The bifidobacteria adhered at levels ranging from 4% to 10% of the CFU added, but none of the bifidobacteria were able to invade cells. The abilities of Listeria to adhere to and to invade cells varied widely depending on the strain tested. Three groups of Listeria were identified based on invasiveness: weakly invasive, moderately invasive, and highly invasive strains. One strain from each group was tested in competition with bifidobacteria. B. thermacidophilum RBL70 was the most effective in blocking invasion of Listeria, and the decreases in invasion ranged from 38% to 90%. For all three bifidobacterial strains, contact between the cell monolayer and the bifidobacteria for 1 h before exposure to Listeria increased the degree of inhibition. Finally, visualization of competition for adhesion sites on cells by fluorescent in situ hybridization suggested that the two bacteria tended to adhere in close proximity.

Listeria monocytogenes is a facultatively intracellular, gram-positive bacterium responsible for severe food-borne infections leading to meningitis, encephalitis, and gastroenteritis (10, 41). Despite its low incidence in Western countries, L. monocytogenes is ranked first among the food-borne pathogenic bacteria in terms of death rate (40%), far ahead of Campylobacter, Salmonella, and Escherichia coli O157:H7. This high mortality rate has made L. monocytogenes and its spread the subject of intensive research. L. monocytogenes has to cross the intestinal barrier in order to produce central nervous system or feto-placental unit infections (41). The passage through the intestinal barrier is probably the most critical step. Recently, Lecuit et al. (24) showed that adhesion of L. monocytogenes and entry into epithelial cells of the small intestine involve several bacterial cell surface proteins. One of these proteins, internalin A, promotes entry into human cells by interacting with the host receptor E-cadherin (8, 19, 24). Internalin A is covalently anchored to the cell wall and belongs to a large family of leucine-rich repeat proteins (6). Recently, it was shown that an L. monocytogenes mutant with a mutation in the internalin A gene was completely deficient in attachment to the epithelial surface and was unable to invade (30). Animal studies have shown that the absence of either the bacterial surface protein or the intestinal cell receptor leads to a significant decrease in Listeria infectivity.

As it is for most enteric bacterial pathogens, the small intestine, more specifically the ileum, is the preferred site for adhesion and invasion of infectious L. monocytogenes (23). In this environment, the resident flora is less abundant than it is in the colon (10⁷ CFU/g in the ileum compared to 10¹⁰ CFU/g in the colon), and this probably increases the capacity for adhesion to epithelial cells. However, the concentration of bifidobacteria, a normal component of the “autochthonous microbiota,” is relatively high in the human ileum, up to 10⁷ CFU/g (23, 27), suggesting that the genus Bifidobacterium may contain bacteria that are suitable probiotics for limiting Listeria infection.

The genus Bifidobacterium has been extensively studied because of its beneficial effects on health, especially protection of the intestinal tract from microbial infection (1, 16, 21, 26, 28). Several mechanisms have been proposed to explain the efficacy of bifidobacteria in preventing infection. These mechanisms include production of antimicrobial agents, such as organic acids, blocking the adhesion of pathogens and toxins to epithelial cells, and modulating the immune response of the host (32). Inhibition of enteric pathogens by bifidobacteria is well documented, especially for gram-negative bacteria. Gagnon et al. (12) showed that bifidobacteria isolated from feces of newborns were able to inhibit in vitro adhesion of E. coli O157:H7 to the Caco-2 and HT-29 cell lines. The level of inhibition varied from 0.1 to 12% depending on the bifidobacterial strain tested and its initial concentration. Similar results have been obtained for inhibition of Salmonella enterica serovar Typhimurium in both in vitro and in vivo studies (2, 25).

Recently, our group has isolated several new strains of Bifidobacterium from infant feces. Three of these strains, strains RBL67, RBL68, and RBL70, have been shown to produce bacteriocin-like peptides with strong inhibitory activity against L. monocytogenes (38). In the present study, we identified the species to which the three bifidobacterial isolates belong and evaluated the potential of these three strains to inhibit in vitro adhesion and invasion of L. monocytogenes in human epithelial cell lines. To our knowledge, this is the first study that demonstrated that there is competition between L. monocytogenes and bifidobacteria for adhesion sites on enteric cells and that there is inhibition of pathogen invasion.

MATERIALS AND METHODS

Bacterial strains and growth media.

The sources of food-borne Listeria strains and Bifidobacterium sp. strains RBL67, RBL68, and RBL70 are listed in Table 1. All strains were maintained as 20% glycerol stock preparations at −80°C. The bifidobacteria used in this study were previously isolated from fecal samples from 1-week-old breast-fed infants by Touré et al. (38). Isolates RBL57, RBL68, and RBL70 have been shown to produce antibacterial substances similar to a bacteriocin that have high antilisterial activity (38).

TABLE 1.

Bacterial strains used in this study

Strain	Referencea	Source	Serotype
B. thermophilum subsp. infantis RBL67	STELA	Infant feces	Unknown
B. thermacidophilum subsp. suis RBL68	STELA	Infant feces	Unknown
B. thermacidophilum subsp. suis RBL70	STELA	Infant feces	Unknown
L. ivanovii LSD5	CFIA	Unknown	Unknown
L. seeligeri LSD11	CFIA	Unknown	Unknown
L. welshimeriLSD12	CFIA	Unknown	Unknown
L. murrayiLSD14	CFIA	Unknown	Unknown
L. monocytogenes LSD15	CFIA	Frozen whole egg	Unknown
L. monocytogenes LSD338	CFIA	Cheese	Unknown
L. monocytogenes LSD339	CFIA	Egg	Unknown
L. monocytogenes LSD340	CFIA	Ice cream	Unknown
L. monocytogenes LSD341	CFIA	Cheese	Unknown
L. monocytogenes LSD346	CFIA	Milk	Unknown
L. monocytogenes LSD348	CFIA	Frozen yolk	Unknown
L. monocytogenes LSD523	CFIA	Unknown	1c
L. monocytogenes LSD524	CFIA	Unknown	1/2c
L. monocytogenes LSD526	CFIA	Unknown	4e
L. monocytogenes LSD529	CFIA	Unknown	3a
L. monocytogenes LSD530	CFIA	Unknown	3b
L. monocytogenes LSD531	CFIA	Unknown	3c
L. monocytogenes LSD535	CFIA	Unknown	1/2a

^aSTELA, STELA Dairy Research Group, Pavillon Comtois, Université Laval, Quebec, Canada; CFIA, Laboratory Services Division, Canadian Food Inspection Agency, Ottawa, Ontario, Canada.

All Listeria strains were grown at 37°C for 16 h in tryptic soy broth (Difco Laboratories, Detroit, MI) supplemented with 0.6% (wt/vol) yeast extract (Difco). Bifidobacteria were cultured in De Man-Rogosa-Sharpe (MRS) broth (9) obtained from Rosell Institute Inc. (Montreal, Québec, Canada) supplemented with 0.05% (wt/vol) l-cysteine hydrochloride (Sigma Chemical Co., St. Louis, MO) and incubated anaerobically at 37°C for 16 h in Oxoid jars using an atmosphere generation system (Oxoid AnaeroGen; Oxoid Ltd., Basingstoke, Hampshire, England).

Identification of bifidobacteria.

The identities of Bifidobacterium sp. isolates RBL67, RBL68, and RBL70 were confirmed by 16S rRNA gene sequence analysis, using the PCR amplification protocol proposed previously (35). The genomic 16S rRNA gene was amplified using Bifidobacterium genus-specific 16S rRNA gene primers lm26 (5′-GATTCTGGCTCAGGATGAACG-3′) and lm3 (5′-CGGGTGCTI*CCCACTTTCATG-3′) (20). Bifidobacterial purity was verified by using universal bacterial primers Bak4 (5′-AGGAGGTGATCCARCCGCA-3′) and BakLLW (5′-AGTTTGATCMTGGCTCAG-3′) developed by Schürch (35). PCR amplification was carried out with a GeneAmp 2400 PCR system (Perkin-Elmer, Norwalk, CT). Amplified products were separated by electrophoresis (pH 8.0, 100 V for 45 min) using an i-Mupid mini-electrophoresis unit (Cosmo Bio Co. Ltd., Tokyo, Japan) in a 1% agarose gel (Invitrogen, Burlington, Ontario, Canada) and were stained with ethidium bromide. PCR-amplified products that were approximately 1.4 kb long were excised from gels and purified using QIAGEN columns (QIAGEN, Mississauga, Ontario, Canada). Automated DNA sequence analysis was carried out for both strands with a DNA sequencer (ABI Prism 3100; Applied Biosystems, Foster City, CA). Sequence data were assembled and analyzed using GCG Wisconsin package 10 (Accelrys, San Diego, CA). Database homology searches were performed using the FASTA network service at the National Center for Biotechnology Information (www.ncbi.nlm.nih.gov). The bootstrap method was employed to determine the statistical confidence of phylogenetic relationships (11).

Epithelial cell lines.

Enterocyte-like Caco-2 (ATCC HTB-37) and HT-29 (ATCC HTB-38) cells were obtained from the American Type Culture Collection (Rockville, MD). Caco-2 cells were cultured as monolayers in 75-cm² flasks (Falcon, Becton and Dickinson Company, Franklin Lakes, NJ) with Dulbecco's modified Eagle's medium (DMEM) (Gibco, Invitrogen, Burlington, Ontario, Canada) containing 20% (vol/vol) fetal bovine serum (Wisent Inc., St. Bruno, Québec, Canada), 1% (vol/vol) nonessential amino acids (Gibco), and 1% (vol/vol) antibiotic solution (100 U/ml penicillin and 100 μg/ml streptomycin; Gibco). HT-29 cells were cultured in RPMI 1640 medium (Gibco) containing 10% (vol/vol) fetal bovine serum, 2 mM l-glutamine (Gibco), and 1% (vol/vol) antibiotic solution. For adhesion and invasion assays, Caco-2 or HT-29 cells were seeded in 24-well plates (2 cm²/well; Falcon) to obtain semiconfluent monolayers (10⁶ cells/well) in 3 days. Cell cultures were replenished with DMEM and RPMI 1640 without antibiotics 24 h before the assay was performed. All cell cultures were incubated at 37°C in a 5% (vol/vol) CO₂ atmosphere.

Adhesion assays.

The adhesion of L. monocytogenes and the adhesion of bifidobacteria to epithelial cells were evaluated separately and in a competition assay. Briefly, fresh overnight cultures of L. monocytogenes and bifidobacteria were harvested by centrifugation at 10,000 × g for 10 min and washed twice with sterile phosphate-buffered saline (PBS). The washed pellets were suspended in DMEM for assays with Caco-2 cells and in RPMI 1640 medium for assays with HT-29 cells at final concentrations of 5 × 10⁸ CFU/ml for L. monocytogenes and 5 × 10⁶ to 5 × 10⁸ CFU/ml for bifidobacteria. Cell monolayers in 24-well plates were then inoculated with 250 μl of a single strain for individual adhesion assays or with a combination of Listeria and Bifidobacterium for competition assays. For the latter assays, bifidobacteria were added either at the same time that Listeria was added or 1 h before or 1 h after Listeria was added. Plates were then incubated under anaerobic conditions at 37°C for 30 min, after which free bacteria were eliminated by washing the cell layers twice with PBS. Cells with adherent bacteria were harvested with trypsin, centrifuged at 10,000 × g for 5 min, and suspended in PBS. Adherent L. monocytogenes cells were enumerated on Listeria selective media (Oxoid) and incubated for 24 to 48 h at 37°C in aerobic conditions. Adherent bifidobacteria were enumerated on MRS agar media and incubated for 24 to 48 h at 37°C in anaerobic conditions. The adhesion capacity was expressed as the number of adherent bacteria divided by total number of bacteria added, multiplied by 100. The inhibition of adhesion of L. monocytogenes in a competition assay was expressed as a percentage, using the following formula: inhibition of adhesion = 100(1 − T₁/T₂), where T₁ and T₂ are the numbers of adherent Listeria cells (CFU/well) in the presence and absence of bifidobacteria, respectively.

Invasion assays.

The invasion of epithelial cells by Listeria or bifidobacteria was studied using the gentamicin-based assay described by Bunduki et al. (3) and Gaillard et al. (13). A bacterial suspension (250 μl, like the suspension used for the adhesion assay) was spread on cell monolayers and incubated at 37°C for 1.5 h before it was washed two times with 250 μl of PBS. For competition assays, L. monocytogenes and bifidobacteria were added either together or 1 h apart as described above for the adhesion assays. The monolayers were washed twice with PBS before addition of 250 μl of 1 mg/ml gentamicin (Sigma) and incubation for an additional 1.5 h under the same conditions to kill noninvading bacteria. The monolayers were then washed with PBS and lysed with 250 μl of ice-cold 0.1% (vol/vol) Triton X-100 (Sigma). Appropriate dilutions were prepared in PBS and plated on Listeria selective agar (aerobic conditions, 37°C, 48 h) and MRS agar (anaerobic conditions, 37°C, 48 h) for enumeration of L. monocytogenes and bifidobacteria, respectively. The invasion capacity was determined by multiplying the number of invading bacterial cells/total number of adherent bacterial cells by 100. The inhibition of invasion was expressed as a percentage, using the relative decrease in invasion by Listeria in the presence of bifidobacteria, as determined by the following formula: inhibition of invasion = 100(1 − T_B/T_L), where T_B and T_L are the numbers of invading Listeria cells (CFU/well) in the presence and absence of bifidobacteria, respectively.

FISH.

All fluorescent in situ hybridization (FISH) experiments were performed by using cell monolayers with adherent and invading Listeria and bifidobacteria (from competition and noncompetition assays) and an adaptation of a protocol described by Wagner et al. (43). Monolayers were harvested by addition of 100 μl of trypsin and washed two times with PBS. Cells were then fixed for 3 h at 4°C in PBS containing 4% (wt/vol) paraformaldehyde. These cells were kept frozen in a 50% ethanol-PBS solution before hybridization. Fifteen microliters of diluted paraformaldehyde-fixed cells was transferred to eight-well microscope slides and dried for 10 min at 60°C. The slides had previously been washed with a 70% (vol/vol) ethanol solution containing 0.1% (vol/vol) hydrochloride acid (HCl) and treated with 0.01% poly-l-lysine. The fixed cells were dehydrated successively in 50%, 80%, and 100% ethanol baths (3 min each) at room temperature and then treated with 20 μl of a 10-mg/ml lysozyme (Sigma) solution for 20 min at room temperature, followed by washing with distilled water and repetition of the ethanol dehydration procedure. Fixed samples were then treated at room temperature with a proteinase K solution (0.05 mg/ml; Sigma) for 5 min. After one wash with distilled water and a final dehydration procedure, the slides were dried at room temperature for 20 min and hybridized for 16 h at 46°C in a hybridization solution (0.09 M NaCl, 20 mM Tris-HCl, 0.1% sodium dodecyl sulfate, 35% formamide; pH 7.2) containing 50 ng/ml of Listeria genus-specific probe Lis1255 (5′-fluoro-ACCTCGCGGCTTCGCGAC-3′) (34, 39) and 50 ng/ml of Bifidobacterium genus-specific probe Bif164 (5′-Cy3-CATCCGGCATTACCACCC-3′) (22, 33). Each slide was then washed with hybridization buffer and incubated in washing buffer (0.07 M NaCl, 20 mM Tris-HCl, 0.1% sodium dodecyl sulfate; pH 7.2) for 20 min at 48°C. After a final wash with distilled water, the slides were allowed to dry completely at room temperature before they were treated with a Slowfade antifade kit (Molecular Probes, Invitrogen) used according to the manufacturer's instructions and were visualized with an Olympus fluorescent microscope (Olympus, Melville, NY).

Statistical analysis.

Adhesion and invasion assays were repeated three times, and strain differences were tested by analysis of variance. The significance of differences between strains was determined by Fisher's least-significant-difference test at a level of certainty of 5% with the StatGraphics Centurion 15.0 software (Statpoint Inc., Herndon, VA).

RESULTS

Identification of bifidobacterial species.

The sequence data for the 16S rRNA gene (approximately 1.4 kb) of bifidobacterial isolates RBL67, RBL68, and RBL70 were compared with the available sequences in the GenBank database. The 16S rRNA gene sequences of isolates RBL68 and RBL70 exhibited the highest levels of similarity to the sequences of B. thermacidophilum subsp. suis (GenBank accession no. AY148470.1) (99%) and bacterial isolate mpn (GenBank accession no. AF357558) (98%). The sequence of the 16S rRNA gene of strain RBL67 was also related to the B. thermacidophilum subsp. suis sequence (99%). Using DNA-DNA hybridization, von Ah et al. (42) reported levels of genome homology of 77.2, 77.9, and 86.25% between strain RBL67 and B. thermacidophilum subsp. suis LMG21689^T, B. thermacidophilum LMG21395^T, and B. thermophilum DSM20210^T, respectively. Strain RBL67 was found to differ from B. thermophilum DSM20210^T in its tolerance to low pH and carbohydrate fermentation characteristics. Based on these results, a novel subspecies of B. thermophilum, Bifidobacterium thermophilum subsp. infantis, was created for strain RBL67 (GenBank accession no. DQ340558).

Bacterial adhesion on epithelial cells.

The efficiencies of adhesion of the bifidobacterial and Listeria strains to Caco-2 and HT-29 cells are shown in Table 2. The levels of adhesion of bifidobacteria ranged from 3.24% to 7.68% for Caco-2 cells and from 3.45% to 8.37% for HT-29 cells. In general, there was no significant difference between adhesion of bifidobacteria to Caco-2 cells and adhesion of bifidobacteria to HT-29 cells. B. thermacidophilum subsp. suis RBL70 was the most efficient strain in terms of adhesion to epithelial cells, and the values were 7.68% and 8.37% with Caco-2 and HT-29 cells, respectively. The least efficient of the bifidobacteria was strain RBL68, while the values for strain RBL67 were intermediate.

TABLE 2.

Adhesion and invasion of bifidobacterial and listerial strains with Caco-2 and HT-29 epithelial cells^a

Bacterial strain	Adhesion (%)		Invasion (%)
	Caco-2 cells	HT-29 cells	Caco-2 cells	HT-29 cells
B. thermophilumsubsp. infantis RBL67	5.43 ± 0.58 DEF	6.23 ± 0.55 CDE
B. thermacidophilum subsp. suis RBL68	3.24 ± 1.02 FGHI	3.45 ± 0.91 FGH
B. thermacidophilum subsp. suis RBL70	7.68 ± 1.09 BCD	8.37 ± 1.54 C
L. ivanovii LSD5	4.53 ± 0.65 EFG	7.92 ± 0.10 BCD	1.84 ± 0.45 M	2.66 ± 0.70 KL
L. seeligeri LSD11	1.18 ± 0.07 HIJ	1.43 ± 0.15 EFG	4.69 ± 0.33 L	5.02 ± 1.23 JK
L. welshimeriLSD12	1.03 ± 0.5 HIJ	0.55 ± 0.09 G	18.33 ± 1.25 F	27.94 ± 0.91 C
L. murrayiLSD14	34.6 ± 1.2 A	65.7 ± 9.7 A	7.46 ± 1.02 K	8.96 ± 0.28 HI
L. monocytogenes LSD15	1.67 ± 0.66 HIJ	1.41 ± 0.33 EFG	13.48 ± 1.17 HI	13.18 ± 1.41 FG
L. monocytogenes LSD338	1.02 ± 0.38 HIJ	0.81 ± 0.17 EFG	34.89 ± 0.32 C	42.64 ± 2.13 B
L. monocytogenes LSD339	0.32 ± 0.24 J	0.46 ± 0.10 G	11.24 ± 1.12 IJ	10.57 ± 0.31 GH
L. monocytogenes LSD340	1.14 ± 0.51 HIJ	1.38 ± 0.31 EFG	22.57 ± 2.06 E	26.62 ± 0.47 CD
L. monocytogenes LSD341	0.66 ± 0.21 IJ	0.44 ± 0.11 G	38.72 ± 3.26 B	42.97 ± 7.59 AB
L. monocytogenes LSD346	0.012 ± 0.017 J	0.0037 ± 0.0014 G	15.26 ± 1.47 GH	16.47 ± 1.09 E
L. monocytogenes LSD348	1.04 ± 0.03 HIJ	1.27 ± 0.19 EFG	44.65 ± 0.85 A	45.49 ± 0.82 A
L. monocytogenes LSD523	1.89 ± 0.19 HIJ	2.29 ± 0.48 DEFG	0.64 ± 0.14 M	0.74 ± 0.33 L
L. monocytogenes LSD524	0.12 ± 0.01 J	0.22 ± 0.06 G	0.9 ± 0.29 M	0.83 ± 0.06 L
L. monocytogenes LSD526	0.42 ± 0.03 J	0.4 ± 0.13 G	0.78 ± 0.17 M	0.99 ± 0.34 L
L. monocytogenes LSD529	6.88 ± 1.15 BCDE	7.44 ± 0.36 BCD	7.41 ± 1.09 K	9.57 ± 0.61 HI
L. monocytogenes LSD530	6.42 ± 0.36 CDE	9.45 ± 1.13 B	0.0046 ± 0.0011 M	0.0056 ± 0.007 L
L. monocytogenes LSD531	0.31 ± 0.03 J	0.33 ± 0.11 G	9.14 ± 1.15 JK	10.04 ± 1.21 HI
L. monocytogenes LSD535	0.56 ± 0.30 IJ	0.65 ± 0.16 FG	25.18 ± 1.53 DE	26.31 ± 0.82 CD

^aThe data are means ± standard variations values followed by different letters in the same column are significantly different as determined by a least-significant-difference test at a P value of <0.05.

The adhesion of Listeria strains was more variable, and the levels ranged from 0.012% and 0.0037% with Caco-2 and HT-29 cells, respectively, for L. monocytogenes LSD346 to 34.6% and 65.71% with Caco-2 and HT-29 cells, respectively, for L. murrayi LSD14. Three strains exhibited levels of adhesion ranging from 4.53% to 6.88% with Caco-2 cells and from 7.44% to 9.45% with HT-29 cells. For the remaining strains, the levels of adhesion ranged from 0.1% to 2.0% with both Caco-2 and HT-29 cells. As observed with bifidobacteria, no significant difference between adhesion on Caco-2 cells and adhesion on HT-29 cells was noted.

With the exception of L. murrayi LSD14, which adhered more than any other strain, the adhesion of non-L. monocytogenes Listeria strains, including L. ivanovii LSD5, L. seeligeri LSD11, and L. welshimeri LSD12, did not differ from the adhesion of L. monocytogenes strains.

Bacterial invasion of epithelial cells lines.

The abilities of bifidobacteria and Listeria to invade the Caco-2 and HT-29 epithelial cells lines are also summarized in Table 2. First, no invasion of either cell line was observed with the three bifidobacterial strains.

As observed for adhesion, invasion of both epithelial cell lines by Listeria strains was highly variable. The highest levels of invasion were obtained with L. monocytogenes LSD348 (44.65 and 45.49% for Caco-2 and HT-29 cells, respectively). The lowest levels, 0.0046 and 0.0056%, were observed with L. monocytogenes LSD530. Again, there was no significant difference between invasion of Caco-2 cells and invasion of HT-29 cells, except for strains LSD12, LSD338, LSD340, and LSD341, for which the level of invasion was always higher with HT-29 cells. In addition, no correlation between adhesion and invasion was observed for any Listeria strain. For example, the highly adherent strain LSD530 showed a very low level of invasion, while the weakly adherent strain LSD346 was highly invasive (15.26 and 16.47% for Caco-2 and HT-29 cells, respectively).

Competition for adhesion.

Three L. monocytogenes strains (LSD348, LSD339, and LSD530) were selected and tested in adhesion and invasion competition assays. The invasion percentages for these three strains were very different.

Three concentrations (10⁶, 10⁷, and 10⁸ CFU/ml) were tested for each bifidobacterial strain. When Listeria and bifidobacteria were added simultaneously, no inhibition of the adhesion of Listeria was observed (data not shown). Moreover, Listeria adhesion increased slightly when bifidobacteria were added at a concentration of 10⁸ CFU/ml. Similar results were obtained when L. monocytogenes was added 1 h before bifidobacteria were added.

When bifidobacteria were added at a concentration of 10⁸ CFU/ml 1 h before Listeria was added, strong inhibition of adhesion was observed, notably in the case of L. monocytogenes LSD348, the most invasive strain tested (Fig. 1). Decreases of 21, 16, and 31% in adhesion to Caco-2 cells were observed with Bifidobacterium strains RBL67, RBL68, and RBL70, respectively. Slightly greater decreases were observed for RBL67 and RBL70 with HT-29 cells (29 and 40%, respectively), while RBL68 decreased L. monocytogenes adhesion three times as much (47%) on HT-29 cells as on Caco-2 cells. Finally, no protective inhibition of pathogen adhesion was observed when bifidobacteria were added at concentrations of 10⁶ and 10⁷ CFU/ml (data not shown).

FIG. 1.

Effect of 1 h of preincubation with bifidobacteria on adhesion of L. monocytogenes strain LSD348 to epithelial cells. Different letters above columns indicate that the values are significantly different (P < 0.05).

Figure 2 shows adhesion of bifidobacteria and Listeria to HT-29 cells, as visualized by phase-contrast microscopy (Fig. 2A) and FISH (Fig. 2B to D). Some of the bifidobacteria present on the cell monolayer are near Listeria cells.

FIG. 2.

Visualization of bacterial adhesion on HT-29 cells. (A) Phase-contrast image. Magnification, ×1,000. (B to D) FISH images of B. thermacidophilum subsp. suis RBL70 (red), L. monocytogenes LSD348 (green), and the two strains in competition, respectively.

Inhibition of invasion.

Figure 3 shows the inhibition of invasion of enteric cells by L. monocytogenes strains LSD339, LSD348 and LSD530 in the presence of Bifidobacterium strains RBL67, RBL68, and RBL70 added at the same time. The degree of inhibition depended on both the bifidobacterial strain and the Listeria strain. With B. thermophilum subsp. infantis RBL67, the level of invasion inhibition varied from 35 to 87%; the latter value was obtained for L. monocytogenes LSD530 on HT-29 cells. With B. thermacidophilum subsp. suis RBL68, 70% inhibition was observed for LSD530 on HT-29 cells, while LSD339 invasion was not inhibited at all for either cell line, nor was invasion of HT-29 cells by LSD348. With B. thermacidophilum subsp. suis RBL70, the level of inhibition of L. monocytogenes LSD339 on HT-29 cells was 91%, which was the highest level of inhibition observed, while the level of inhibition of LSD348 on Caco-2 cells was only 38%. The levels of invasion inhibition were higher on HT-29 cells with all strain combinations except RBL67 and RBL68 with LSD348. RBL70 inhibited invasion of HT-29 cells by LSD339, LSD348, and LSD530 by 91, 78, and 88%, respectively. L. monocytogenes LSD348, the most invasive strain tested, was more invasive than the other two strains except for Caco-2 cells in the presence of RBL68.

FIG. 3.

Inhibition of L. monocytogenes LSD339, LSD348, and LSD530 invasion of epithelial cells by B. thermophilum subsp. infantis RBL67 (top) and B. thermacidophilum subsp. suis RBL68 (middle) and RBL70 (bottom). Different letters above columns indicate that the values are significantly different (P < 0.05).

Figure 4 shows the effect of 1 h of prior contact between epithelial cells and bifidobacteria on invasion of L. monocytogenes LSD348. The efficacy of the three bifidobacteria for inhibiting invasion by Listeria is clear, with values ranging from 73 to 99.8%, compared to 0 to 78% when both bacteria were added simultaneously. Strain RBL68 was the least effective strain, while the two other bifidobacterial strains showed similar activities.

FIG. 4.

Effect of 1 h of preincubation with bifidobacteria on invasion of epithelial cells by L. monocytogenes strain LSD348. Different letters above columns indicate that the values are significantly different (P < 0.05).

Finally, when the bifidobacteria were not added until 1 h after L. monocytogenes was added, strain RBL67 was still able to limit the invasion of the most invasive Listeria strain tested (Fig. 5) by as much as 67% on HT-29 cells. This was in stark contrast with the results obtained for the adhesion assay, which showed that there was no inhibition of L. monocytogenes adhesion when bifidobacteria were added 1 h after Listeria was added.

FIG. 5.

Effect of contact with bifidobacteria 1 h after exposure to L. monocytogenes strain LSD348 on invasion of epithelial cells by the pathogen. The same letter above columns indicates that the values are not significantly different (P > 0.05).

DISCUSSION

In this study, we used two in vitro models to test the abilities of three bifidobacterial strains, isolated from infant feces, to adhere to intestinal cells using two epithelial cell models. The first model is based on the human colon adenocarcinoma cell line Caco-2, whose characteristics simulate structural and functional characteristics of mature enterocytes in vitro (31). The second model is based on the HT-29 cell line, which is also derived from a human colon adenocarcinoma. This cell line also has structural and functional features of mature human enterocytes, but when cultured in the presence of glucose, it maintains a constant proliferation rate with practically no further differentiation (44). Our results indicate that high percentages of B. thermophilum subsp. infantis RBL67 and B. thermacidophilum subsp. suis strain RBL68 and RBL70 cells adhered to both cell lines. The adhesion of strains RBL67 and RBL70 was found to be greater than that reported for B. bifidum RBL71, which we previously found to be highly adhesive (12), and B. pseudolongum ATCC 25526 (7), which is also considered highly adhesive. The adhesion by B. thermacidophilum subsp. suis strain RBL68 is comparable to the adhesion by B. bifidum RBL71. The three bifidobacterial strains tested in the present study did not invade the intestinal epithelial cells, confirming that the genus Bifidobacterium is generally noninvasive. This criterion is an important safety consideration.

The ability of L. monocytogenes to invade intestinal cells is a good indicator of its virulence (18, 29, 32, 40). On this basis, we tested the abilities of 14 strains of L. monocytogenes and four other Listeria species isolated from food commodities in Canada to invade intestinal cell lines. Our results revealed widely different levels of adhesion and invasion among the strains tested. Such strain variability is very common and may explain the difference in virulence among L. monocytogenes strains. We also noted that adhesion and invasion were greater with HT-29 cells than with Caco-2 cells. This may have been due to facilitation of L. monocytogenes invasion by factors related to epithelial cell proliferation (44). Velge et al. (41) reported that entry of L. monocytogenes into Caco-2 and HT-29 cells was closely related to the proliferation process. These authors proposed that entry of L. monocytogenes into Caco-2 and HT-29 cells is mediated by interaction between E-cadherin on the surface of the proliferative cell and the bacterial cell surface protein “internalin.” Since the proliferation rate in HT-29 cells is constant and is higher than that in Caco-2 cells (44), the interaction between Listeria and HT-29 cells is expected to be greater than the interaction between Listeria and Caco-2 cells.

As recently reviewed by Servin (37) and Servin and Coconnier (36), one of the better-documented effects of probiotic bifidobacteria on human health is the prevention of infection. As observed for many invasive enteropathogens, L. monocytogenes pathogenicity involves adhesion of the bacteria to intestinal epithelial cells and invasion in two distinct steps (14, 18, 36). In this study, we examined the abilities of three bifidobacterial strains to block the adhesion and invasion of three strains of L. monocytogenes with different invasive potentials. All three bifidobacteria failed to inhibit the adhesion of Listeria when the two organisms were added simultaneously. We even observed a slight increase in Listeria adhesion at the highest concentration of bifidobacteria used. These results differ from those of Coconnier et al. (4), who showed that there was dose-dependent inhibition of adhesion by L. monocytogenes in competition with Lactobacillus acidophilus LB on Caco-2 cells. These authors tested live and heat-killed Lactobacillus and observed 75% and 80% inhibition, respectively, with 1 × 10⁹ lactobacilli/ml. Considering that our incubation time was 30 min, compared to the 60 min used by Coconnier et al. (4), the levels of adhesion of our Bifidobacterium strains on epithelial cells should have been lower. Coconnier et al. did not detect any differences when heat-killed lactobacilli were added before the Listeria strain was added, while we observed significantly enhanced inhibition of the pathogen when live bifidobacteria were used.

Inhibition of adhesion of other enteric pathogen by bifidobacteria or Lactobacillus has also been reported. Gagnon et al. (12) showed that a B. bifidum strain isolated from infant feces reduced adhesion of E. coli O157:H7 to Caco-2 cells in a dose-dependent manner. Other studies have also reported inhibition of adhesion of enteropathogens, such as E. coli, S. enterica serovar Typhimurium, and Staphylococcus aureus, in a dose-dependent manner (2, 5, 15, 25). In a very limited number of studies workers have examined interactions between bifidobacteria and Listeria. Our study showed the potential probiotic activities of the B. thermophilum subsp. infantis RBL67 and B. thermacidophilum subsp. suis RBL68 and RBL70 strains from newborns to reduce the adherence of Listeria spp., particularly L. monocytogenes, to both epithelial cells lines and to reduce invasion of these cells lines.

While the bifidobacterial strains tested in this study were not very effective in inhibition of the adhesion of L. monocytogenes when the two bacteria were added simultaneously, they were quite effective in blocking the invasion of epithelial cells. In fact, two of the three strains tested inhibited invasion by 35 to 90% depending on the mode of addition and the cell line. Interestingly, when bifidobacteria were added 1 h before Listeria was added, the level of inhibition increased to nearly 100% for the majority of the strain combinations. One possible explanation for the differences between the adhesion and invasion inhibition results in this study is the different contact times (30 min for adhesion assays and 90 min for invasion assays). Since Listeria is able to invade epithelial cells as soon as it comes into contact with them (18), we limited the incubation time to 30 min in order to minimize invasion and obtain a better estimate of inhibition of adhesion. We also tested the inhibition of adhesion with a prolonged incubation time, but no significant differences were noted. Adhesion and invasion of L. monocytogenes occur by separate mechanisms involving different proteins located on the bacterial cell wall (14, 17, 18). The first step in the invasion process is adhesion of the bacteria on the epithelial cell, which is modulated by fibronectins. This step is believed to be nonspecific and may occur with a wide variety of cell types and other molecules (14). The second step involves a more specific interaction of other bacterial proteins called internalins with other receptors, which results in high-affinity host-cell-specific interactions and therefore in the internalization of the bacteria in the host cells (14, 18). Considering this mechanism, the greater ability of bifidobacteria to inhibit L. monocytogenes invasion may suggest that there is an interaction with specific receptors on epithelial cells that are essential for Listeria internalins to promote entry of the pathogen into the host cells.

Fluorescent in situ hybridization performed with a cell monolayer after a competition assay for invasion showed that some of the bifidobacteria present were physically near Listeria cells, suggesting that competition between the two bacteria for adhesion and invasion is possible. However, further experiments are needed to determine possible interactions between the two bacteria and the impact that such interactions could have on the inhibition of listerial invasion.

In conclusion, the bifidobacterial strains used in this study poorly blocked adhesion of L. monocytogenes strains isolated from food samples but were able to block invasion by L. monocytogenes. Our results also suggest that competition between probiotic and pathogenic bacteria should be analyzed by considering adhesion and invasion two distinct events, in order to characterize the mode of action by which the probiotic is able to decrease infection. Finally, the results obtained in this study reinforce previous data that suggest that probiotics are able to prevent infections by pathogens when sufficient numbers are present in the intestinal flora. The three bifidobacterial strains used in this study should be considered probiotic strains with interesting potential for preventing enteric infections in humans. In vivo studies using animal models for Listeria infection are currently under way to confirm the results obtained in vitro.