Collection Mémoires et thèses électroniques
AccueilÀ proposNous joindre

Chapitre 5. Effet inhibiteur des bifidobactéries humaines sur l’attachement des virus entériques aux lignées cellulaires intestinales

Table des matières

Inhibitory effect of human bifidobacteria on the attachment of enteric viruses to intestinal cell lines

L’objectif de ce chapitre était d’évaluer de façon in vitro l’effet des bifidobactéries isolées de fèces de nouveaux-nés pour inhiber l’attachement des virus entériques aux cellules intestinales Caco-2 et HT-29. Pour ce faire, les bifidobactéries ont été ajoutées avant, en même temps et après l’ajout des virus sur les cellules entériques afin d’évaluer leur impact sur le pourcentage d’attachement des rotavirus, du substitut de norovirus et du virus de l’hépatite A. Les trois souches de bifidobactéries humaines utilisées pour cette étude ont été sélectionnées principalement sur le critère d’adhésion très élevé aux Caco-2 et HT-29. Ce travail a été réalisé au pavillon Marchand (Université Laval, Québec) dans des laboratoires avec les équipements nécessaires pour la culture cellulaire, la virologie et la bactériologie.

Les résultats de ce travail seront soumis dans la revue Applied Environmental Microbiology. Les auteurs de cet article sont Mélanie Gagnon, Julie Jean, André Darveau, et Ismaïl Fliss. Ce chapitre présente le contenu intégral de ce manuscrit rédigé en anglais.

La capacité des souches probiotiques à adhérer à la muqueuse intestinale, à exclure, à entrer en compétition et à déplacer les pathogènes entériques viraux représente un aspect important pour traiter des infections entériques virales. La capacité d’adhésion de trois souches de bifidobactéries humaines aux cellules Caco-2 et HT-29 a été évaluée et comparée avec deux souches références de bifidobactéries de l’ATCC. L’adhésion des bifidobactéries a varié de 0.8% à 8.6% selon la concentration ajoutée (5×107 ou 5×109 CFU/ml). Le pourcentage d’attachement de rotavirus, d’un substitut-norovirus (FCV) et de l’hépatite A (HAV) à une concentration de 104 PFU/puits aux lignées cellulaires en absence de bifidobactéries a varié de 11.5% (rotavirus sur les cellules HT-29) à 20.8% (FCV sur les cellules Caco-2). Le contact des cellules avec les suspensions de bifidobactéries provenant de nouveaux-nés (5×109 CFU/ml) pendant 30 min avant l’addition des virus réduit l’attachement des rotavirus de près de 98%. Pour HAV, une réduction légèrement inférieure de l’attachement viral (près de 85%) sur les deux lignées cellulaires intestinales a été obtenue avec les mêmes souches probiotiques ajoutées avant les virus. Au contraire, aucune diminution substantielle de l’attachement des FCV a été observée dans l’essai d’exclusion. La présence simultanée des bactéries et virus sur les cellules intestinales produit des réductions variables de l’attachement viral, suggérant que différents mécanismes sont impliqués dans les essais d’exclusion et compétition alors que la capacité des bifidobactéries à déplacer les virus attachés est plus limitée. Les souches de bifidobactéries humaines testées démontrent un potentiel probiotique pour un usage prophylactique contre des entéropathogènes viraux tel que rotavirus.

The ability of probiotic bacteria to adhere to the intestinal mucosa and exclude, compete with and displace viruses is of great interest for preventing and treating viral enteric infections. In this study, adhesion of bifidobacteria isolated from human infants and two ATCC reference strains to Caco-2 and HT-29 cell growth ranged from 0.8% to 8.6% of applied loads of 5×107 or 5×109 CFU/ml. Percent adhesion increased with increasing bacterial concentration. Attachment of rotavirus, norovirus-surrogate (FCV) and hepatitis A at 104 PFU/well to cell growth in the absence of bifidobacteria ranged from 11.5% (rotavirus on HT-29 cells) to 20.8% (FCV on Caco-2 cells). Contacting cells with suspensions of bifidobacterial isolates from infants (5×109 CFU/ml) for 30 min prior to exposure to virus reduced rotavirus attachment by about 98%, while HAV attachment was reduced by 85% on both intestinal cell lines. No substantial decrease in FCV attachment was observed. Contacting cells with bacteria and virus at the same time produced variable reductions in viral attachment, suggesting different mechanisms for exclusion and competition, while the ability of the bifidobacteria to displace attached virus was more limited. These bifidobacteria thus appear to have potential as probiotics for prophylactic use against viral enteropathogens such as rotavirus.

The rotaviruses, noroviruses and hepatitis A virus (HAV) are major foodborne and waterborne pathogens and a worldwide public health concern (Koopmans et al., 2002; Clark and McKendrick, 2004). The infections they cause often require hospitalization particularly in the case of sensitive populations such as young children, elderly, pregnant women or immunocompromised persons (Gerba et al., 1996). These viruses are transmitted through the fecal-oral route by the ingestion of contaminated foods, particularly those that are raw or minimally processed (Seymour and Appleton, 2001). The first event in the process of infection is viral attachment to the cells of the intestinal lumen. Rotaviruses and noroviruses infect the mature epithelial cells lining the villi of the small intestine and replicates locally leading to diarrhea (Michelangeli and Ruiz, 2003). Recently, Blank et al. (2000) showed that prior to infecting the liver, HAV can infect intestinal cells and produce new virions therein, which proceed to their target cells in the liver to cause hepatitis. The nucleocapsids of these three nonenveloped viruses interact with polarized cells through attachment to a cell glycoprotein receptor (Levy et al., 1994). Rotavirus and norovirus infections cause structural and functional damages in cells of the small intestine, observable as shortened and abnormal microvilli (Koopmans et al., 2002). In contrast, HAV release from intestinal cells occurs without a cytopathic effect (Blank et al., 2000).

In numerous recent studies, treatments using probiotics have been proposed to prevent and control enteric infections (Gill, 2003; Sullivan and Nord, 2005; Huebner and Surawicz, 2006; De Vrese and Marteau, 2007). There is increasing evidence that lactobacilli and bifidobacteria, which are natural components of the gastrointestinal microbiota, possess antimicrobial activities that contribute to the host gastrointestinal defense system (Servin, 2004). Clinical trials have shown that selected strains of bifidobacteria have the ability to interfere with pathogens in the intestinal tract (Saavedra et al., 1994; Guarino et al., 1997).

The mechanisms by which bifidobacteria may protect the host from pathogens include production of inhibitory substances, blockade of adhesion sites and stimulation of the immune response (Picard et al., 2005) However, no in vitro study has examined interactions between bifidobacteria and enteric viruses. To investigate the interactions between enteric viruses with bifidobacterial strains on the surface of intestinal cells, we have used cellular models previously developed (Gagnon et al., 2007). These have proven useful for initial evaluation of intestinal host-microbial interactions and are inexpensive.

Among the numerous strains of Bifidobacterium that we have previously isolated from newborn infants, three (coded RBL67, RBL69 and RBL70) have shown strong resistance to gastrointestinal conditions and high probiotic potential (Touré et al., 2003; Moroni et al., 2006). In this study, we evaluated the ability of these strains to adhere to the intestinal Caco-2 and HT-29 cell surfaces and thereby to exclude, compete with and displace a norovirus-surrogate (feline calicivirus), human rotavirus and hepatitis A virus.

Bifidobacteriumsp. RBL67 (Bifidobacterium thermophilum), RBL69 (Bifidobacterium thermacidophilum) and RBL70 (also Bifidobacterium thermacidophilum) were previously isolated by Touré et al. (2003) from fecal samples of 1-week-old breast-fed infants. Their identification was done by 16S rDNA sequence analysis using a PCR amplification protocol described by Schürch (2002) with minor modifications (Moroni et al., 2006) and subsequently by DNA-DNA hybridization (von Ah et al., 2007). All other organisms were obtained from the American Type Culture Collection (ATCC; Rockville, MD). B. longum ATCC 15707 and B. pseudolongum ATCC 25526 were used respectively as low and high adhesive reference bacteria (Crociani et al., 1995). All strains were stocked in 20% glycerol at -80ºC. Bifidobacteria were cultured in De Man-Rogosa-Sharpe (MRS) broth (Rosell Institute Inc., Montréal, Québec, Canada) supplemented with 0.05% (wt/vol) L-cysteine-hydrochloride (Sigma Chemical Co., St. Louis, MO) and incubated anaerobically at 37°C for 18 h in jars using an atmosphere generation system (AnaeroGen, Oxoid Ltd., Basingstoke, Hampshire, England).

Feline calicivirus strain F9 (ATCC VR-782), a norovirus-surrogate, was propagated in Crandell’s feline kidney cells (CrFK, ATCC CCL-94) as described by Bidawid et al. (2003). Human rotavirus strain Wa (ATCC VR-2018) and hepatitis A virus (HAV) strain HM-175 (ATCC VR-1402) were propagated respectively in monkey kidney MA-104 cells (ATCC CRL-2378.1) and FRhK-4 cells (ATCC CRL-1688) using the methods of Ansari et al. (1988) and Mbithi et al. (1991, 1992). The viral stock solution consisted of cell culture supernatant obtained after an infection period followed by three freeze (-80ºC) and thaw (37ºC) cycles and low speed centrifugation (1000 rpm, 10 min). The supernatant was stored in 1-ml fractions at -80°C until use.

All cell lines were obtained from the American Type Culture Collection. All cell culture media and supplements were Gibco brand, obtained from Invitrogen Inc., Burlington, Ontario, Canada. All cell cultures were incubated at 37ºC in a 5% CO2 atmosphere. Caco-2 cells (ATCC HTB-37) were cultured as monolayers in 75-cm2 flasks (Falcon, Becton Dickinson and Company, Franklin Lakes, NJ) with Dulbecco’s Modified Eagle medium (DMEM) containing 4,500 mg of glucose/liter, supplemented with 20% (vol/vol) fetal bovine serum (FBS), 1% (vol/vol) nonessential amino acids, and 1% (vol/vol) antibiotic solution (100 U/ml penicillin and 100 μg/ml streptomycin). HT-29 cells (ATCC HTB-38) were grown in RPMI 1640 medium supplemented with 10% (vol/vol) fetal bovine serum and 2 mM L-glutamine and 1% (vol/vol) antibiotic solution. For bacterial adhesion and viral infection assays, Caco-2 or HT-29 cells were seeded in 24-well tissue culture plates (Falcon) at 2.8×104 and 4×104 cells per well, respectively. The culture medium was replaced at two-day intervals, and the cells were used at postconfluence (106 cells/well) after 15 or 21 days of culture for Caco-2 and HT-29 respectively. Intestinal cell cultures were replenished with DMEM and RPMI 1640 without antibiotics 18 h before the assays were performed.

FRhK-4 and MA-104 cells were grown as monolayer in Eagle’s minimal essential medium with Earle’s salts (MEM) supplemented with 10% (vol/vol) fetal bovine serum, 2 mM L-glutamine, 1% (vol/vol) nonessential amino acids, 1% (vol/vol) HEPES buffer, 0.1125% (wt/vol) sodium bicarbonate, 1% (vol/vol) antibiotic solution. They were seeded at concentrations of 4.2×104 and 4.7×104 cells per cm2, respectively. CrFK cells were seeded at a concentration of 4.2×104 cells per cm2 and grown in DMEM supplemented with 10% (vol/vol) newborn calf serum (NCS), 2 mM L-glutamine, 1% (vol/vol) nonessential amino acids and 1% (vol/vol) antibiotic solution.

Caco-2 and HT-29 cells represent two complementary in vitro intestinal epithelial models that differ by the formation of a monolayer of polarized differentiated cells in the case of Caco-2 cells and a multilayer of unpolarized and undifferentiated growth in the case of the HT-29 line (Rousset, 1986). These cell lines are considered accurate representations of the in vivo situation, in which both structures are found among proliferating intestinal cells (Zweibaum et al., 1991). The adhesion of B. thermophilum RBL67 to both cell lines was comparable to that obtained for B. pseudolongum ATCC 25526, a strain recognized as highly adhesive (Crociani et al., 1995). This isolate (RBL67) had also a greater capacity to adhere to Caco-2 cells than an isolate that we previously found to have high adhesiveness and identified as B. thermacidophilum RBL71 (Gagnon et al., 2006). Moreover, the adhesion of RBL70 was greater than that reported for RBL71 whereas that of RBL69 was less, even though both of these isolates have been identified as B. thermacidophilum. Differences in adhesion efficiency are strain-dependent and probably determined by physiology and factors such as proteinaceous component, polysaccharide, ionic charge and lipoteichoic acid (Crociani et al., 1995). This capacity of bifidobacteria to adhere to intestinal cells is considered as an important criterion for the in vitro selection of probiotics that may colonize (if only transiently) the human intestinal tract (Alander et al., 1999; FAO/WHO, 2002).

It has been suggested recently that bifidobacteria play a role in acting as a barrier against infection of the gastrointestinal tract by viral pathogens (Gill, 2003; Servin and Coconnier, 2003; Servin, 2004). As observed for many invasive enteropathogens, the ability of enteric viruses to attach to the surface of intestinal epithelial cells is a critical step in infection(Carter, 2005). On this basis, we examined the abilities of three human bifidobacterial strains to exclude, compete with and displace three enteric viruses. Exclusion and competition by bifidobacteria had profiles quite different from those of displacement. Viral attachment inhibition by exclusion and competition was generally much greater than the inhibition achieved by displacement. In fact, when added 30 min before or at the same time as enteric viruses, the three bifidobacterial strains inhibited viral attachment by 51% to 98%, depending on the concentration added and the cell line. This result differs from those of Botic et al. (2007), who obtained low to moderate inhibition (30% to 60% cell survival) of vesicular stomatitis virus (VSV) infection of IPEC-J2 (intestinal pig epithelial cell jejunum) cell monolayer using probiotic L. paracasei A14 at a concentration of 108 lactobacilli/ml. These authors demonstrated that there was a bacterial dose and contact time dependence for antiviral effect of probiotic bacteria. Considering that our probiotics concentration and contact time were different, this could explain the greater inhibition observed in our study. Botic et al. (2007) did not observed a significant change in the cell survival rate due to virus dilution, while we observed a virus dose effect on the inhibition of viral attachment.

While the bifidobacteria tested in this study were quite effective in excluding rotavirus and HAV, all three failed to inhibit FCV attachment even when added 30 min before the virus. In fact, FCV exhibited a greater capacity to adhere to HT-29 cells in the presence of probiotic bacteria. This increased attachment (up to 140%) may be explained by a mechanism that involves a low pH-dependent step. The bifidobacteria acidify the cell culture medium and as Kreutz et al. (1994) has demonstrated, acidic environment promote FCV attachment, with maximum binding at pH 6.5. In the case of rotavirus and HAV, this pH-effect was not observed. The human bifidobacteria in our study were also not very effective at displacing virus. When they were added 30 min after the viruses, the reduction in viral attachment ranged from 9% to 55%. Lee et al. (2003) reported similar interactions for two probiotic Lactobacillus strains, L. rhamnosus GG and L. casei Shirota, against enteropathogenic bacteria such as E. coli and Salmonella on Caco-2 cells. These authors demonstrated that displacement of pathogenic bacteria is a very slow process and once the pathogens are attached to their cell receptors, probiotics have little effect on displacing them. They propose that bacterial adhesins present on probiotics could bind to the mucosal surface in order to be effective in exclusion and competition interactions.

The greatest reductions in attachment (86% to 98%) in the exclusion and competition assays were obtained for rotavirus with all three bifidobacterial isolates. Immunofluorescent microscopy showed that bifidobacteria diminished rotavirus infection. It is noteworthy that we observed the greatest inhibition of rotavirus attachment in exclusion and competition assays with B. thermophilum RBL67, which has previously shown the greatest adhesiveness to Caco-2 and HT-29 cells among the three bifidobacteria isolates. However, further experiments at the molecular level are needed to determine the interactions between this Bifidobacterium and rotavirus on intestinal cells and to evaluate the impact of such interactions on rotavirus infectivity.

The antiviral effect of probiotic bacteria may involve any of several possible mechanisms. The pre- and co-incubation inhibition of viral attachment to human cell lines in vitro may be due to the formation of a bacterial biofilm that slows down the infectious process by preventing the access of enterovirulent organisms to the cell surface. This steric hindrance mechanism has been suggested by Botic et al. (2007) to explain inhibition of VSV infection by pre-incubation with probiotic bacteria. This non-specific mechanism occurs independently of the enteric virus tested. Another possible mechanism is based on the ability of probiotic bacteria to compete with viral pathogens for specific adhesion sites on the same intestinal mucosal receptors. In the case of the three nonenveloped enteric viruses in the present study, binding to intestinal glycoprotein receptors could be involved (Levy et al., 1994). However, Lee and Puong (2002) have suggested that probiotics have specific adhesines that bind to cellular receptors and that competition between probiotics and pathogens takes place for attachment to these receptors. The specificity of both the adhesines and the receptors still needs to be proven by future molecular studies. Inhibition of the attachment of other invasive enteropathogens to enterocytes by bifidobacteria has also been reported. Bernet et al. (1993) showed that the human strain B. longum 4 inhibited Salmonella typhimurium association with Caco-2 cells by 88% in competition assays. Moroni et al. (2006) have shown similar inhibition of Listeria monocytogenes invasion of Caco-2 and HT-29 cells using bifidobacteria isolated from newborn infants.

Our study has shown the probiotic potential of three bifidobacterial isolates from newborns, B. thermophilum (RBL67) and two variants of B. thermacidophilum (RBL69 and RBL70), to reduce the attachment of enteric viruses to Caco-2 and HT-29 cell lines. This could have significant impact on human health with regard to prevention and treatment of enteric viral infections. The bifidobacteria used in this study were not able to inhibit infection of cultured cell lines by rotavirus, FCV and HAV completely, but significant reductions of viral attachment were obtained when the bifidobacteria were incubated with the cells prior to or simultaneous with exposure to virus. The different degrees of inhibition obtained with the tested bifidobacteria indicate the need for case-by-case characterization of probiotic strains but suggest potential for adhesive strains as inhibitors of enteropathogenic viral attachment to enterocytes.

Finally, the cellular models described in this study offer appropriate tools for studying host-enteric virus interactions in the presence of probiotic bacteria. Better understanding of these interactions will provide a rational basis for using probiotics as dietary supplements. This should allow the selection of strains with potential application to the prevention or treatment of diarrhea caused by enteric viruses. Studies are in progress to confirm and characterize the antiviral properties of the selected bifidobacteria in a rodent model infected with rotavirus.

This work was supported by the Natural Sciences and Engineering Research Council of Canada (NSERC). M.G. was supported by a doctoral fellowship from the Fonds Québécois de la Recherche sur la Nature et les Technologies (FQRNT). We thank Ueli Von Ah for their technical assistance.

© Mélanie Gagnon, 2007