Saccharomyces boulardii decreases inflammation and intestinal colonization by Candida albicans in a mouse model of chemically-induced colitis

Abstract

The present study was designed to investigate the effects of Saccharomyces boulardii on inflammation and intestinal colonization by Candida albicans in a BALB/c mouse model of colitis that had been induced by dextran-sulfate-sodium (DSS). Colonization with C. albicans was established by oral gavage with a 200 µL suspension of 10⁷ yeast cells. A 1.5% solution of DSS was administered in drinking water 1 h after C. albicans oral challenge, while 10⁷ cells of S. boulardii was inoculated daily by oral gavage for 1 week. Faeces were collected daily for 2 weeks. Seven groups of mice consisting of those that were administered either C. albicans or S. boulardii or both were sacrificed after 14 days and samples of the colon were taken for histological scoring and real-time PCR (RT-PCR) analysis of inflammatory cytokines and toll-like receptors (TLRs). Compared to control animals that did not receive DSS, the number of C. albicans colonies recovered from faeces was significantly greater in mice receiving DSS. In contrast, the colony forming units (CFUs) of C. albicans were greatly reduced in mice receiving S. boulardii. The administration of this yeast decreased the severity of DSS-induced clinical scores and histological inflammation. At the mRNA expression level, an increase in TLR2 and TLR4 resulting from the presence of S. boulardii was associated with a reduction in the inflammatory cytokines TNFα and INFγ. In mice receiving DSS and C. albicans, TLR4 was over-expressed by stimulation with both yeasts, but TLR2 and TNFα, which were increased by the administration of C. albicans alone, were decreased in the presence of S. boulardii. These results indicate that S. boulardii decreased inflammation and C. albicans colonization in this BALB/c mouse model of colitis.

Candida albicans, Saccharomyces boulardii, dextran sulphate sodium, inflammation

Introduction

Candida albicans is a frequent cause of systemic infections and mortality in patients undergoing chemotherapy for cancer 1, 2, as well as immunosuppressive 3 and prolonged antibiotic therapy 4. The passage of viable C. albicans through the gastrointestinal (GI) mucosa into the host bloodstream is believed to be an important mechanism leading to systemic candidosis 5–7. It has been suggested that probiotics, which have the ability to reduce the growth of C. albicans within the gut, could prevent systemic invasion 8. Saccharomycesboulardii, a non-pathogenic yeast, is recognized to have a positive probiotic effect when given orally as a lyophilized preparation to treat acute infectious gastroenteritis and antibiotic associated diarrhoea 9, 10. The use of live S. boulardii cells is significant since previous studies have shown differential effects of live versus killed S. boulardii cells in several models (cholera toxin 11, E. coli thermostable toxin 12 or mortality induced by C. difficile13). Regarding C. albicans, oral administration of live S. boulardii cells in immunosuppressed mice, has been demonstrated to inhibit C. albicans translocation to the mesenteric lymph nodes, liver, and kidneys 14.

In the context of inflammatory bowel diseases, controlled clinical trials have indicated that S. boulardii cells were effective in the prevention and/or the treatment of Crohn's disease (CD) and ulcerative colitis (UC) 15, 16. In a recent study, a possible relationship was observed between C. albicans and human bowel inflammation leading to CD. It was found that C. albicans could act as an immunogen for anti-oligomannose antibodies (named ASCA), which are markers of CD 17. The present study was designed to investigate the effects of S. boulardii on inflammation and intestinal colonization by C. albicans in a BALB/c mouse model of colitis induced by dextran-sulfate-sodium (DSS).

Materials and methods

Yeast strains used

C. albicans wild type strain SC5314 18 and S. boulardii strain (Ultra-Levure; BIOCODEX, Gentilly, France) were used throughout the study and maintained at 4°C on Sabouraud dextrose agar (SDA). For the experimental studies, cells of both test isolates were grown in Sabouraud dextrose broth at 37°C in a shaking incubator for 18 h.

Animals

All animal experiments conformed to the Ministère de l'Agriculture et de la Forêt Resolution on the use of animals in research and were approved by the Subcommittee on Research Animal Care of the Faculty of Medicine, Lille (protocol 2003-35).

Six- to eight-week-old BALB/c male mice were used in these investigations. Mice were maintained for us by Charles River Laboratories (France). Animals were group housed and had free access to regular rodent chow and tap water. Mice were killed by cervical dislocation 14 days post inoculation of C. albicans.

Inoculum preparation and induction of colitis

Each animal was inoculated by oral gavage with 200 µl of phosphate-buffered saline (PBS) containing 10⁷ live C. albicans cells. One hour after C. albicans oral challenge, a single-daily dose of 10⁷ live S. boulardii cells was administrated by oral gavage for 1 week. Once the mice were administered the yeast they were given 1.5% DSS (MW 36–50 kDa; MP Biomedicals, LLC, Germany) in drinking water for 14 days to induce intestinal inflammation. The presence of yeasts in the intestinal tract was followed each day through plate counts of faeces (approximately 0.1 g/sample) collected from each animal. The faecal samples were suspended in 1 ml saline, ground in a glass tissue homogeniser and inoculated onto Candi-Select medium® (Bio-Rad laboratories, Marnes la Coquette, France). This chromogenic medium is designed for the isolation of yeasts from clinical specimens and the identification of medically important species by the colour of their colonies 19. C. albicans colonies were counted after 48 h incubation at 37°C and the results expressed as colony forming units (CFUs) per gram of faeces. In accord with manufacturer's instructions (Gentaur Molecular Products, Brussels, Belgium), faecal smears were stained using Fungiflora Y, a fluorescent dye that binds with fungal cell wall glucans. 20. In brief, 200µl of reagent A was added to faecal smears, the contents were mixed gently, and then 200 µl of reagent B was added to each slide and the contents were again mixed. The smears were stained for 10 min, rinsed with distilled water, the slides were dried, and coverslips were mounted over the stained areas with an aqueous mounting medium. The slides were then examined through the use of an immunofluorescent microscope (Leica Microsystems AG, France). At day 14, the animals were sacrificed and their colons were removed to be cut into sections for histological examination.

Experimental design

Mice were distributed into the following control and experimental groups; (a) healthly mice without any treatment as normal control (CTL), (b) mice gavaged orally with C. albicans but without any treatment (CaCTL), (c) mice gavaged orally with S. boulardii (SbCTL), (d) mice treated with DSS (DSS), (e) a group gavaged orally with C. albicans and treated by DSS (CaDSS),(f) mice gavaged orally with S. boulardii and treated with DSS (SbDSS) and (g) a group inoculated by oral gavage with both S. boulardii and C. albicans and treated with DSS (CaSbDSS).

Determination of clinical score and histological score

The body weight of each mouse was determined daily and examined for stool consistency and the presence of blood in the rectum. Clinical scores, as described previously, were assessed independently by two investigators blinded to the protocol 21. Briefly, no weight loss was scored as 0, a loss of 1–5% as 1 point, 5–10% as 2 points, 10–20% as 3 points, and >20% as 4 points. Stool consistency was recorded as; 0 = well-formed pellets, 2 = pasty and semi-formed stools that did not stick to the anus, and 4 = liquid stools that remained adhesive to the anus. Bleeding was reported as; 0 points for no blood in haemoccult, 2 points for positive haemoccult, and 4 points for gross bleeding from the rectum. These scores were added, divided by 3 and multiplied by 2, resulting in a total clinical score ranging from 0 (healthy) to 8 (maximal activity of colitis). The entire colon from the caecum to the anus was removed post mortem. Rings of the transverse part of the colon were fixed overnight in 4% paraformaldehyde-acid and embedded in paraffin for histological analysis. Sections (4 µm thick) were stained with May-Grunwald-Giemsa (Merck) and histological scoring was performed by two independent investigators blinded to the protocols. The infiltration of inflammatory cells was evaluated using the following scoring system; (a) rare inflammatory cells in the lamina propria were counted as 0, (b) increased numbers of inflammatory cells, including neutrophils in the lamina propria as 1, (c) confluence of inflammatory cells, extending into the submucosa as 2, and a score of 3 was given for transmural extension of the inflammatory cell's infiltrate. For epithelial damage, the absence of mucosal damage was counted as 0, the presence of discrete focal lymphoepithelial lesions were counted as 1, the occurrence of mucosal erosion/ulceration was assigned a score of 2, and 3 was given for extensive mucosal damage and extension through deeper structures of the bowel wall. The two subscores were added, with the combined histological scores ranging from 0 (no changes) to 6 (extensive cell infiltration and tissue damage).

Real-time mRNA quantification

Following the manufacturer's instructions, total RNA was isolated from colon samples using a NucleoSpin RNA II kit (Macherey-Nagel, France), with 20–50 units of DNase I (RNase-free) at 37°C for 30 min to avoid contamination with genomic DNA. RNA quantification was performed by spectrophotometry (Nanodrop; Nyxor Biotech, France). Reverse transcription of mRNA was carried out according to the manufacturer's instructions in a final volume of 26 µl from 1 µg total RNA using 300 U M-MLV reverse transcriptase (Invitrogen, France) with 500 ng oligo (dT) 12–18 and 50 U ribonuclease inhibitor (RNase-Out, Promega). PCR was performed using an ABI 7000 prism sequence detection system (Applied Biosystems, France) with SYBR green (Applied Biosystems, France). Amplification was carried out in a total volume of 25 µl containing 0.5 µl of each primer (Table 1), and 1 µl of cDNA prepared as described above. SYBR green dye intensity was analysed using Abiprism 7000 SDS software (Applera Corp.). All results were normalized to the housekeeping gene β-actin.

Table 1

Mouse oligonucleotide sequences.

Primer	Sequence (5′–3′)
β-actin S	TCA CCC A CA CTG TGC CCA TCT ACG A
β-actin AS	CAG GGG AAC CGC TCA TTG CCA ATG
TNF-α S	TGG GAG TAG ACA AGG TAC AAC CC
TNF-α AS	CAT CTT CTC AAA ATT CGA GTG ACA A
INFγ S	TTT CAC CCT GAA GTC GTT GGA
INFγ AS	ATG GAA AGG AGG GAT ACA GAC GT
TLR-2 S	ACG GGG CCA TCC TTG TT
TLR-2 AS	TTA TCT TGC GCA GTT TGC AGA A
TLR-4 S	GAC CAA GCC TTT CAG GGA ATT
TLR-4 AS	GGA CGT GTA AAC CAG CCA GGT

S, sense; AS, anti-sense.

Statistics

Data are expressed as the mean±SE of five mice in each group. All comparisons were analysed by the Mann-Whitney U test. Statistical analyses were performed using the StatView 4.5 statistical program (SAS Institute Inc., Meylan, France). Differences were considered significant when the P value was below 0.05.

Results

Effects of DSS on GI colonization by C. albicans

C. albicans colonization in mice was followed daily for two weeks by performing plate counts of faeces (Fig. 1). When control mice were inoculated with C. albicans, very few CFUs were recovered from stools throughout the 14 day experimental period. In contrast, in DSS-treated mice (group 5) significantly higher numbers of CFUs were recovered from stool samples beginning on day 5 (Fig. 1).

Fig. 1

Faecal Candida albicans counts in BALB/c mice. Each data set represents the mean values of Candida albicans counts for CaCTL mice (control group of mice inoculated with Candida albicans by oral gavage), CaDSS mice (group of mice inoculated with Candida albicans by oral gavage and treated with DSS) and CaSbDSS mice (group of mice inoculated by oral gavage both Saccharomycesboulardii and Candida albicans and treated with DSS). (^‡P<0.05 in CaDSS mice vs. CaCTL mice; *P<0.05 in CaDSS mice vs. CaSbDSS mice; and ^†P<0.05 in CaSbDSS mice vs. CaCTL mice.)

Effects of S. boulardii on GI colonization by C. albicans in the DSS model

When stool smears from mice treated by oral gavage with C. albicans and/or S. boulardii were examined under the microscope after staining with Fungiflora, the predominant morphology of C. albicans consisted of elongated or ovoid blastospores whereas S. boulardii had their spherical morphology (Fig. 2B).

Fig. 2

Microscopic examination of stools smears from mice administered orally by gavage with Candida albicans (a) or Saccharomycesboulardii (b) after staining with Fungiflora. Magnification×100.

The number of yeasts in the stools was determined by plating faecal suspensions on Candi-Select medium (Fig. 3). While the blue colonies C. albicans could be readily determined, the green color zones around S. boulardi colonies made it difficult to assess their recovery.

Fig. 3

Quantification of Candida albicans (blue) and Saccharomycesboulardii (light green) colony-forming units present in mice stools after dilution and seeding on Candi-Select medium.

In mice receiving C. albicans alone, the number of CFUs recovered from the stools was very low for the first 4 days post gavage and the yeast was quite undetectable from day 5 to the conclusion of the studies (Fig. 1). By contrast, in mice receiving DSS, the number of C. albicans CFUs dramatically increased on day 5 and continued to increase until day 14. In mice receiving DSS, C. albicans and S. boulardii, the number of C. albicans colonies in the faeces was significantly reduced from day 5 to day 7 during the one week treatment with S. boulardii. Reduction was maintained at very low levels up to 4 days after the end of S. bourlardii administration and began to slightly increase reaching a plateau on day 11 post inoculation. By contrast, DSS-treated mice who did not receive S. boulardii showed a steady increase between days 11 and 14 (Fig. 1).

Determination of clinical and histological score

DSS administration to BALB/c mice was associated with clinical changes, which included weight loss starting on day 7 and the appearance of occult faecal blood starting on day 9 (Fig. 5A). Administration of S. boulardii reduced the clinical signs associated with DSS. Similarly, S. boulardii reduced clinical scores of mice receiving DSS and inoculated with C. albicans, but this effect did not differ from that observed in mice treated with DSS alone followed by S. boulardii administration.

Histological analysis showed that DSS induced epithelial erosions, a massive inflammatory response and destruction of the normal crypt structure of the colon. These results were unaffected by the presence of C. albicans (Fig. 4). When S. boulardii was administrated, a reduction of epithelial damage and destruction of crypts was noted. Histological scores showed that this reduction of inflammation following S. boulardii administration was significant in both DSS and DSS-C. albicans groups although the reduction was less pronounced in the presence of C. albicans (Fig. 5B).

Fig. 4

Histological characteristics of colon samples taken from mice. (a) inflamed colon from Candida DSS-treated mouse, (B) colon from Candida plus Saccharomyces DSS-treated mouse. Histological sections (4 µm thick) stained with May-Grunwald-Giemsa revealed living cells (b), whereas the grey colour indicates direct cell damage and death resulting from DSS treatment and Candida colonization (A). Magnification×5 (A and B).

Fig. 5

Clinical and histological analysis of DSS-induced colitis in mice. (A) Clinical score was determined by assessing rectal bleeding and stool consistency. (B) Histological score was performed blindly by two independent examiners as described in Materials and methods (140 examinations). Data are expressed as the means±SE in each group. (^†P<0.05 in SbDSS mice vs. DSS mice; ^‡P<0.05 in CaSbDSS mice vs. CaDSS mice; and *P<0.05 in CaSbDSS mice vs. DSS mice.)

Inflammatory cytokine mRNAs expression in the colon

The contribution of inflammatory cytokines particularly TNFα and INFγ to the tissue-damaging process in mice that had received DSS and C. albicans and/or S. boulardii was determined by RT-PCR analysis (Fig. 6). TNFα mRNAs increased in the colon of BALB/c mice after DSS exposure and was significantly higher with the presence of C. albicans. In each of the groups of mice, administration of S. boulardii was associated with a reduction of TNFα expression. INFγ mRNAs increased in the DSS group but did not show an additional increase in DSS-induced colitis mice treated with C. albicans. S. boulardii administration to mice exposed to DSS or DSS with C. albicans was associated with decreased INFγ mRNAs expression levels in the colon (Fig. 6).

Fig. 6

Relative expression levels, determined by real-time quantitative PCR, of TNFα, INFγ, TLR-2 and TLR-4 mRNA in colons taken from BALB/c mice administered 1.5% DSS in drinking water for 14 days. Data are expressed as the means±SE in each group. (^†P<0.05 in SbDSS mice vs. DSS mice; ^‡P<0.05 in CaSbDSS mice vs. CaDSS mice; and *P<0.05 in CaSbDSS mice vs. DSS mice.)

Toll-like receptor (TLR) mRNAs expression in the colon

TLR2 and TLR4 were analysed as they are known to respond to yeast stimulation. In order to assess the way each yeast species differentially stimulates these TLRs, their mRNAs were analysed by real-time RT-PCR (Fig. 6). Although the histological scores were not modified by the inoculation of C. albicans and regular administration of S. boulardii in the absence of DSS (Fig. 4), an over-expression of TLR2 and TLR4 was still evident at day 14 and comparable to the one induced DSS alone. DSS however was found to augment the stimulation induced by single yeast species. TLR2 and TLR4 appeared to respond more to C. albicans than to S. boulardii. Administration of S. boulardii during the process of colonization by C. albicans reduced C. albicans stimulation of TLR2 but presence of both yeasts led to a significant cumulative effect on TLR4 over-expression.

Discussion

In humans, C. albicans colonizes the mucosal surfaces and is adapted to adhere to all parts of the gastrointestinal (GI) tract from the oral cavity to the rectum and peri-anal tissue 22. Genetic identity between colonizing strains and strains isolated from blood cultures suggests that the GI tract is a major reservoir of yeasts involved in systemic candidiasis 23–25, an important medical and economic problem in modern hospitals 26, 27. In this study, DSS treatment of experimental mice resulted in clinical effects that mimic several pathologic conditions in humans. The use of DSS resulted in a significant increase in C. albicans colonies recovered from mice stool samples and provided the model used to investigate the effects of S. boulardii on GI colonization, clinical scores of inflammation and expression of inflammatory cytokines. Oral administration of S. boulardii for 1 week to treated mice reduced the number of C. albicans colonies isolated from stool samples. However, when administration was stopped, the number of C. albicans colonies recoverable from stool began to increase indicating that S. boulardii does not permanently colonize the intestinal tract. Thus, S. boulardii exerts transient colonization, which may be sufficient to protect the intestinal mucosa against colonization by C. albicans. It has been previously demonstrated that presence of high numbers of S. boulardii cells prevented the population of C. albicans from exceeding a high level in faeces 28 and oral ingestion of S. boulardii inhibited the translocation of C. albicans from the GI tract 14. This antagonistic action of S. boulardii on C. albicans has also been shown in vitro29.

In this study, histological analysis revealed that oral administration of S. boulardii decreased the degree of epithelial damage and inflammation in both DSS and C. albicans DSS-treated groups compared to the group of mice treated only with DSS. It has been shown that the pathogenesis of inflammatory bowel disease, particularly CD, involves genetically influenced dysregulation of the mucosal immune response to antigens present in the normal bacterial flora 30. In susceptible individuals, tissue damage and intestinal inflammation result from immunological response to the indigenous flora or failure of normal regulatory constraints on mucosal immune responsiveness to enteric bacteria 31. However, maintenance of epithelial barrier function is essential for the preservation of mucosal integrity. Therefore reinforcement of barrier function, together with its immune modulatory and metabolic properties, may be central in the mechanism of action of S. boulardii. The latter exerts protective effects toward various enteric pathogens 32 and beneficial effects on the host intestinal mucosa by increasing secretory IgA production 33.

S. boulardii differs from S. cerevisiae by several phenotypic, metabolic, and genetic properties 34–36. Probiotic properties are probably not restricted to S. boulardii and could induced by S. cerevisiae and other members of the genus. However, Castagliuolo et al. have shown that the inhibition of C. difficile toxin A-mediated secretion and permeability in rat ileal loops by S. boulardii was not observed using a S. cerevisiae strain 37.

S. boulardii was found to have a beneficial effect on the human intestine, especially in the maintenance treatment of either UC or CD inflammatory bowel diseases 15, 16. By contrast, several reports suggest that C. albicans may have adverse effects on IBDs 17, 38. Anti-C. albicans antibodies are generated during UC 39 and the presence of C. albicans DNA in colonic mucosa was suggested to be an important marker of this disease 40. For CD, clinical and experimental molecular evidence has indicated that C. albicans is able to induce the production of ASCAs, which recognize the immunodominant antigen of CD 17. Interestingly, S. boulardii intake in patients with CD does not generate ASCA and these patients do not develop fungemia during 3 months of treatment 41.

The results from RT-PCR, which showed increased levels of the inflammatory cytokines TNFα and INFγ following DSS administration were in accord with previous studies 42–44. While TNFα levels were increased by C. albicans stimulation in BALB/c mice, the yeast did not exert any effect on the INFγ levels. In contrast, oral administration of S. boulardii decreased both TNFα and INFγ levels in C. albicans DSS-treated mice. This is in accordance with previous studies concerning the effect of S. boulardii on these cytokines. Two investigations have shown that S. boulardii produces a soluble anti-inflammatory factor (SAIF), that inhibits NF-κB activation and attenuates pro-inflammatory signaling in host cells 45, 46. As to IFNγ, a recent study has shown that S. boulardii inhibition of IFNγ production in the colon was correlated with the trapping of T cells in mesenteric lymph nodes 47.

It has been demonstrated that TLR signalling pathways control the homeostasis of the epithelium and appear critical for protection of the host against gut injury in controlling cytoprotective factors and epithelial cell proliferation 48. By incubating human mononuclear cells with C. albicans yeasts, an antibody against TLR2 led to an inhibition of TNFα and IL-1β production, while an antibody against TLR4 did not influence the production of proinflammatory cytokines 49, 50. The present study shows that the presence of C. albicans and S. boulardii up regulated TLR2 and TLR4 in the colon. Administration of DSS increased both these stimulations. Interestingly, a complex interplay was evidenced between both yeasts since S. boulardii reduced TLR2 stimulation induced by C. albicans whereas TLR4 expression was increased by the presence of both yeasts.

Numerous studies have suggested that TLR4 was more involved than TLR2 in protecting the mice against C. albicans. Prior investigations have noted that TLR2 deficiency confers increased resistance to infection but TLR4^-/- mice have increased susceptibility 49, 51. However, respective roles of TLRs in regulating a complex balance between Th1 and Th2 responses may depend on the strain of mice used in the studies 52–54. It has been demonstrated that the expression levels of TLR2 and TLR4 mRNAs were higher in dendritic cells (DC) isolated from spleens of naive BALB/c mice than in those from naive C57BL/6 mice 55. In this study, 1.5% DSS for 2 weeks was used to induce colitis in BALB/c mice, in which the TLR4 expression was different from C57BL/6 mice exposed to 5% DSS for 1 week (data not shown). These results suggest that the concentration of DSS and the genetic background are critical factors in this model.

Conclusion

In conclusion, the present study shows that C. albicans colonization in mice can be promoted by inflammation and confirms the ability of S. boulardii to reduce C. albicans colonization, as previously established in a model of antibiotic treated mice 56. S. boulardii exhibited opposing effects, i.e., a down regulation of TNFα and TLR2 and an up regulation of TLR4. These effects can be considered as beneficial for the host with respect to inflammation and resistance to C. albicans. Since this model involves two closely genetically related yeasts 57, it could facilitate further studies necessary to determine the nature of yeast molecules responsible for beneficial versus deleterious effects on the intestines of mammals.

Acknowledgements

The authors thank Emilie Gantier and Edmon Erdual for their excellent technical assistance.