Bi fi dobacterium strains inhabiting the gastrointestinal tract of rat as potential probiotics for animals

The aim of the study was to evaluate in vitro the potential probiotic functionality in the gut of animal-originated bifi dobacteria strains. Eleven B. animalis and fi ve B. pseudolongum strains isolated from Wistar rats were characterized for their tolerance to low pH and bile, and seven B. animalis and four B. pseudolongum strains for their ability to adhere to HT-29 cells and mucus. The strains showed high resistance to three-hour exposure to pH 3 (34-100% of the population survived) and six-hour exposure to the presence of 3% bile (28-93%), as well as differentiated adhesive abilities. The number of bacterial cells attached to 100 HT-29 cells ranged from 9 to 33 and from 17 to 645 for B. pseudolongum and B. animalis strains, respectively, whereas corresponding levels of adherence per square centimetre of polystyrene-immobilized mucin ranged from 3 × 105 to 4 × 105 and from 5 × 105 to 889 × 105 bacterial cells. Three B. animalis strains (KSP4, PS46 and PS11) resistant to low pH and bile, well-adhering to epithelial cells and mucus were selected as potential probiotics for animals and will be evaluated in future in vivo studies.


INTRODUCTION
Probiotics are live microorganisms that, when administered through the digestive tract, positively affect the host's health.A variety of microbial species WASILEWSKA E. ET AL.
Bifi dobacteria are Gram-positive, rod-shaped, anaerobic bacteria that inhabit the gastrointestinal tract of humans and animals, where they constitute a part of the dominant anaerobic fl ora benefi cially affecting host health (Bezkorovainy and Miller-Catchpole, 1989).For years, generally recognized as safe (GRAS), bifi dobacteria have been used as active ingredients in functional food products.Studies on the role of bifi dobacteria in the maintenance of host health have revealed that probiotic strains not only counteracted proliferation of opportunistic and harmful bacteria and reduced diarrhoea incidence, but also positively infl uenced digestion, host metabolism and immunity (Tannock, 1999;Dunne, 2001).The use of bifi dobacteria in animal feeding could signifi cantly support animal breeding.
Much knowledge about the function of probiotic bifi dobacteria is based, however, upon work done in humans, whereas the results might not always be the same in animals.Furthermore, it is generally accepted that probiotic effectiveness is only achieved with a suffi cient dose of living cells.Surveys of probiotic products on the market have revealed defi ciencies in the viability of probiotic strains, especially in those containing bifi dobacteria (Fasoli et al., 2003;Masco et al., 2005).Since probiotic cultures may encounter acidic conditions both in technological processes and during gastric transit, tolerance to low pH is a critical factor infl uencing probiotic functionality in the gut.Natural bacterial habitats are settled by different species, thus, it is also of paramount importance that the candidates for probiotics should be selected from the host species, especially if colonization is essential to gain the ultimate goal of application of the probiotic (Fuller, 1992).Therefore, studies on the selection of new strains highly resistant to gastrointestinal and technological conditions able to colonize a host's intestine seem to be of great importance.
Most of the studies on probiotics have focused on the selection of probiotic strains designed for humans.There are only a few reports suggesting that animaloriginated bifi dobacteria such as B. animalis and B. pseudolongum strains may play a role in intestinal mucosal defence (Bezkorovainy and Miller-Catchpole, 1989;Mengheri et al., 1991;Meng et al., 1998).Likewise, to the best of our knowledge, there is no previous publication encompassing a larger number of animal-originated strains in terms of their sensitivity to gastrointestinal conditions or adhesive abilities.Some authors have described the resistance to unfavourable gastric conditions of B. animalis subsp.lactis strains, closely related to B. animalis subsp.animalis species inhabiting animal intestines (Masco et al., 2004(Masco et al., , 2005(Masco et al., , 2007)).Our previous investigations revealed that the rat's colon harbours bifi dobacteria from the species of B. animalis and B. pseudolongum, with the fi rst being predominant (Wasilewska et al., 2003).
The aim of the present study was to characterize strains of B. animalis and B. pseudolongum for their sensitivity to an acidic environment and presence of bile as well as their ability to adhere to HT-29 epithelial-like cells and mucin.Although these assays are not fully suffi cient to predict the functionality of strains in the gut, they are considered as the most important prerequisites in the selection of Bifi dobacterium strains to be used as probiotics.

Bacterial strains and culture conditions
All bifi dobacteria strains tested (see Table 1) were isolated in our laboratory from Wistar rats and identifi ed to species using phenotypic and molecular methods (Wasilewska et al., 2003).

Tolerance to low pH
Active cultures were thoroughly mixed and transferred quantitatively into two test tubes, in the fi rst, the pH was adjusted to 3.0 with sterile 1 N HCl, whereas into the second (control) an equivalent volume of sterile distilled water was added.Both tubes were incubated in a tissue culture incubator (Cytoperm 8088, Hereus Instruments GmbH, Austria) in an atmosphere with lowered oxygen content (5% O 2 and 22% CO 2 ) at 37 o C for 3 h.Numbers of live bacterial cells were counted as colony forming units (CFU) per ml on Garche's agar.The cultures were neutralized with 1 N NaOH before plating.The plates were incubated anaerobically (Gas Pak anaerobic system) at 37 o C for 48 h.Survival was calculated as the percentage of surviving population relative to the control population.

Resistance to bile salts
The procedure was similar as described above, however, the pH of the cultures was adjusted to 6 with 1 N NaOH before transferring into tubes and then a 15% solution of ox bile (Sigma) was added to a fi nal concentration of 3%.The exposure to bile lasted for 6 h.

Adhesion to HT-29 cells
Seven B. animalis and four B. pseudolongum strains listed in Table 3 were tested for their adhesion to epithelial cells.Epithelium-like HT-29 cell line was purchased from Deutsche Sammlung von Mikroorganismen und Zellkulturen (DSMZ).HT-29 cells were grown routinely in Dulbecco's modifi ed Eagle's minimal essential medium (DMEM) (Gibco BRL) supplemented with 2 mmol/l of L-glutamine, 1% (v/v) of non-essential amino acids and 10% (v/v) of inactivated (30 min, 56 o C) foetal calf serum (Gibco BRL).For all experiments, cells were cultured in a tissue culture incubator (Cytoperm 8088, Hereus Instruments GmbH, Austria) at 37 o C in a 5% CO 2 -95% air atmosphere and 80% humidity.
For the bacterial adhesion assay, HT-29 cells were inoculated at a concentration of 2×10 4 cells per cm 2 of glass coverslips placed in six-well tissue-culture plates (Nunc, Germany).The culture medium was changed every other day and well differentiated monolayers were used at a post-confl uence phase after 18-20 days of incubation.
Overnight bifi dobacterial cultures (24 h-old) were harvested by centrifugation at 5000 g for 10 min, washed in sterile phosphate buffer saline (PBS, pH 7.2) and resuspended in a mixture of modifi ed Garche's broth (specially prepared according to a tissue culture procedure) and Dulbecco's MEM medium (at a ratio of 1:1) to a fi nal concentration of 1×10 9 CFU per ml (OD 550 =1.2).The bacterial suspensions were added to the monolayers washed twice with PBS buffer in the amount of about 1×10 8 bacterial cells per cm 2 of a well, and incubated for 1 h under conditions described for HT-29 cells.After incubation, the monolayers were washed fi ve times with sterile PBS to remove non-adherent bacteria, preserved in graded methanol series: 25, 50, 75 and 96% (v/v) by immersing for 5 min in each solution, next Gram stained and analysed microscopically.Adherence was evaluated in 20 random microscopic fi elds and the number of bacterial cells attached per 100 epithelial cells was calculated.Each experiment was conducted in duplicate.

Binding to mucus
The strains tested towards adhesion to mucus are listed in Table 3. Adhesion to mucus was determined according to the method described by Cohen and Laux (1995) with some modifi cation.In brief, 1 mg/ml of porcine mucin (Sigma; M2378) was diluted in HEPES (N-2-hydroxy-ethyl-piperazine-N'-2-ethanesulfonic acid)-Hank's buffer (HH; 10 mM HEPES; pH 7.4) and immobilized with 250 µl on polystyrene microtitre 24-well plates (Nunc) by overnight incubation at 4 o C.
Freshly multiplied bacterial strains were harvested by centrifugation (5000 g for 10 min), washed twice with HH buffer and resuspended in HH.The optical density of bacterial suspensions at 600 nm was adjusted to 0.5 ± 0.02 in order to standardize the number of bacteria (1×10 8 -2.5×10 8 CFU per ml).
The wells with immobilized mucus were washed twice with 500 µl HH to remove excess mucus.Afterwards, a suspension of 250 µl of each strain was added to two parallel wells.After 1 h incubation at 37 o C, the wells were washed twice with 500 µl HH to remove unattached cells and thereafter the number of bacterial cells attached to mucus was determined.To this end, 1 ml of Triton X-100 (1% solution in HH buffer) was added to each well, next the wells were left for 5 min at room temperature and thoroughly scrapped with a sterile pipette tip in order to release all the attached bacterial cells to the suspension.Since it was impossible to scoop out the whole suspension inside the well, and also in order to correct for intra-assay variation, 0.5 ml of the bacterial suspension from two parallel wells were mixed together and analysed as a one trial.The number of bifi dobacteria attached per well was counted as CFU using the pour plate method.Incubation of the plates was carried out at 37 o C for 48 h under anaerobic conditions.Each adhesion assay was performed in duplicate.The fi nal results were expressed as the number of bacterial cells attached per cm 2 of the surface covered with mucin.

Scanning electron microscopy
Monolayers of the epithelial cells on glass coverslips with the attached bacterial cells were used for scanning electron microscopic (SEM) studies.After the adhesion assay the monolayers were fi xed with 2.5% glutaraldehyde for 1 h at room temperature, washed twice in PBS and postfi xed in 2% OsO 4 for 30 min.Afterwards, the samples were washed three times in PBS and dehydrated in a graded acetone series: 25, 50, 75, 90 and three times in 100% (v/v) by immersing for 5 min in each solution, and then dried at the critical point using liquid CO 2 as the transition medium.Dehydrated samples were coated with gold in a vacuum evaporator (JOEL, JEE-4X) and examined with a JEOL 1200EX TEMSCAN electron microscope equipped with an EM-ASID10 scanning device, at 80 kV.WASILEWSKA E. ET AL.

Statistical analysis
All results are presented as the average of at least two independent experiments with standard deviation.Each experiment was performed with two parallels to correct for an intra-assay variation.A simple correlation coeffi cient was calculated to measure the correlation between the features tested.

RESULTS
Eleven B. animalis and fi ve B. pseudolongum strains isolated from rat faeces were tested for their sensitivity to low pH or the presence of bile.The results showing the behaviour of the strains in the presence of stress factors are reported in Tables 1 and 2. After 3-h exposure to pH 3, the survival of all the Bifi dobacterium strains ranged from 34 to 102% of the population (Table 1).A comparison of inhibitory effects of bile salts on the strains yielded similar results.After 6-h exposure to the presence of 3% of ox bile in the medium, the number of live cells decreased about 1.1 to 3.6-times, which means 28-93% survivability of bacterial cells (Table 2).No signifi cant correlation was found between the two tested features (r=-0.17).A/jejunum 38.5 ± 1.9 39.4 ± 1.9 102 B. pseudolongum KSI-9 E/faeces 133.6 ± 10.9 47.0 ± 6.1 35 1 the strains were exposed to pH 3 for 3 h; 2 the capital letter denotes the source rat; 3 survival is shown as percentage of surviving population relative to population in control tube; 4 mean ± standard deviation 13.9 ± 5.5 3.9 ± 2.7 28 B. pseudolongum KSI-9 76.0 ± 7.4 35.6 ± 4.6 47 1 the strains were exposed to 3% ox bile for 6 h; 2 survival is shown as percentage of surviving population relative to population in control tube; 3 mean ± standard deviation Epithelial-like HT-29 cells were used to assess the ability of the Bifi dobacterium strains to adhere to epithelial cells.The adherence of the tested strains was differentiated and the number of bacterial cells attached to 100 HT-29 cells amounted from 9 to 645 (Table 3).9 ± 3 3 ± 1 1 the number of bacterial cells attached to 100 epithelial HT29 cells in monolayer; mean ± standard deviation; 2 the number of bacterial cells attached to 1 cm 2 of the surface covered with mucin The highest number of attached cells was observed for B. animalis PS46.Two other B. animalis strains, PS11 and KSP4, also adhered well (~150 cells), whereas the four remaining B. animalis strains and all B. pseudolongum strains tested adhered weakly (9-67 cells).Bacterial cells adhered to the monolayers of epithelial-like HT-29 cells in a diffuse way -as dispersed single cells, or in a cluster (Figures 1 and 2).The tested bifi dobacterial strains were also able to bind to mucus, and their adherence to mucin was as much differentiated as their adhesion to epithelial cells (Table 3).The highest binding rates were observed for B. animalis PS11, PS37, PS46 and KSP4 strains (360-889×10 5 bacterial cells attached per cm 2 of the surface covered with mucin).The other strains, except for well-adhering B. animalis KS29a3 (21×10 5 cells) showed weak binding abilities to mucin, below 5.5×10 5 cells per cm 2 (Table 3).An insignifi cant positive linear correlation (r=0.32) was observed between adhesion of the tested strains to epithelial cells and to mucin.In general, the strains adhering well to epithelial cells also adhered well to mucin.However, B. animalis PS37 cells adhered weakly to HT-29 cells and strongly to mucin.

DISCUSSION
Resistance to gastric acidity and bile salts as well as attachment to mucus or epithelial cells are the most important prerequisites for the effectiveness of probiotics in the host gut (Charteris et al., 1998).So far, most of the studies on probiotics have been focused on the selection of probiotic strains designed for humans and originating from humans, therefore animal-originated bifi dobacteria have been studied very rarely.Resistance to unfavourable gastric conditions is a rare property among human-originated bifi dobacterial species, such as B. bifi dum, B. infantis, B. longum, B. breve, B. adolescentis or B. catenulatum, a great majority of the strains belonging to those species appeared to be sensitive to low pH or bile (Clark and Martin, 1994;Lankaputhra and Shah, 1995;Zavaglia et al., 1998).The performed studies revealed that the tested animal-originated strains survived the 3-h exposure to pH 3 well.In all cases, 0 to merely 2.8-fold reduction of live cells was observed, whereas survival at the level of >10 6 CFU/ml is suffi cient for probiotic effectiveness in the gut (Charteris et al., 1998;Collado and Sanz, 2006).Resistance to low pH and bile of single animal-originated strains of B. animalis and B. pseudolongum (originating from type or industrial collections) was observed by Lankaputhra and Shah (1995) and Zavaglia et al. (1998).Among the other tested bifi dobacteria, strains of B. animalis subsp.lactis seem to show the highest tolerance to physiological and technological stress conditions and, as a consequence, have been preferentially included in probiotic products in the food industry, although many researchers stress the importance of host specifi city of probiotic strains (Jayamanne and Adams, 2006;Masco et al., 2007).The B. lactis strains were isolated from fermented milk and, as they are closely related to B. animalis species, are classifi ed at a subspecies level of B. animalis (Meile et al., 1997;Masco et al., 2004).The observed resistance of B. animalis strains isolated from rats may be a species-specifi c feature; however, more studies encompassing a greater number of B. animalis strains originating from different animal species should be performed.Due to high tolerance to low pH, B. animalis strains can also be more suitable for technological processing than other bifi dobacteria.Our previous research on the survival of B. animalis strains in commercial bioyoghurts with pH 4.5-4.7 during 4 weeks of storage under refrigeration revealed merely a 2-4-fold reduction of live bifi dobacterial cells (Bielecka et al., 2000).
Resistance to stress factors of the tested animal-originated strains was confi rmed during their exposition to bile.The applied 3% concentration of bile corresponds to the top physiological concentrations of bile in the human gut and makes up an equivalent of 1.5% of bile salts.Likewise, in an acidic environment, all tested B. animalis and B. pseudolongum strains survived the 6-h exposure to bile well, although in comparison with their survivability at pH 3, the numbers of viable cells of the strains tested were slightly reduced.Nevertheless, the applied high concentration of bile and the observed survival rate suggest that the tested strains may be good candidates for the elaboration of probiotic products.
Interaction of probiotic bacteria with the intestinal mucosa may enhance the possibility for colonization, immune system modulation, and defence against enteropathogens (Tannock, 1999;Dunne, 2001).Unlike sensitivity to unfavourable gastrointestinal conditions, the tested strains differed in their adhesion to both mucosal enterocytes and mucus.The experiments revealed a dissimilarity of the adhesive properties of both tested Bifi dobacterium species, and also a strong diversity of the adhesive abilities among the B. animalis strains.All well-adhering strains belonged to B. animalis species and four out of the eleven strains tested strongly adhered to mucus, whereas only one strain showed strong adherence to epithelial cells.Crociani et al. (1995) described good adhesion of two B. animals and one B. pseudolongum strain to Caco-2 cells.Good adherence of some B. animalis subsp.lactis strains to mucus has also been documented in other recent papers (Kirjavainen et al., 1998;He et al., 2001;Rinkinen et al., 2003).However, all those strains originated from type or industrial culture collections.
An important function of benefi cial microfl ora is to form a barrier against colonization of the gastrointestinal tract by pathogenic bacteria.Commensal bacteria colonizing the mucus layer adjacent to the epithelium block receptors and protect the intestine from pathogen invasion.Light and scanning electron microscopy studies of the adhesion to HT-29 cells showed that the tested bifi dobacterial cells adhered in clusters, probably to mucus secreted by goblet cells, or in diffuse wayto a brush border of enterocytes (Figures 1 and 2).This suggests that the receptors for bacterial adhesion on mucus may be different, or present in different numbers than on enterocytes, but also that some bifi dobacterial cells can be able to form biofi lm on the mucosal layer and nearby goblet cells.Such a hypothesis is in agreement with high adhesion of some strains to mucus, especially those isolated directly from the intestine.Our previous scanning electron microscopy studies of bacterial adhesion to rat colon revealed that bifi dobacteria adhered to intestinal epithelium by fi laments, also connecting bacterial cells and consequently forming characteristic microcolonies (Bielecka et al., 1998).Bernet et al. (1993) observed higher adhesion of some human-originated bifi dobacterial strains to HT29-MTX (a homogenous subpopulation of goblet cells) than to Caco-2 cells.The ability of probiotic bacteria to attach to host mucus seems to be of a great importance as it ensures mutual interaction.Enterocytes are continuously exfoliated from the tips of the villi and replaced with new cells migrating out of the crypt to reach the villous apex.The strains adhering to both mucus and enterocytes seem to be well predisposed to intestinal colonization.Our previous in vitro research revealed that the B. animalis PS46 strain, well adhering to HT-29 cells and mucin, also strongly adhered to the intestinal wall of rat colon (Bielecka et al., 1998).
In conclusion, among the tested strains three B. animalis strains (PS11, PS46 and KSP4) adhered well to mucin and enterocytes and can be considered to be potential probiotic candidates.Although the ecology of animal-type bifi dobacteria has been studied very rarely, it is known that both B. animalis and B. pseudolongum strains are natural inhabitants of a majority of animals (hogs, chickens, turkeys, cattle, calves, piglets, sheep, mice, rats, dogs) (Bezkorovainy and Miller-Catchpole, 1989).All of them can be potential probiotic recipients; nevertheless, in vivo studies on the evaluation of the effectiveness of selected strains in host gut are necessary.

Table 1 .
Viability of the tested Bifi dobacterium strains after exposure to low pH 1

Table 2 .
Viability of the tested Bifi dobacterium strains after exposure to bile salts 1

Table 3 .
Adhesion of the Bifi dobacterium strains to HT29 cells and to mucus