Editor: Artur Ulmer
Regulation of the IL-10/IL-12 axis in human dendritic cells with probiotic bacteria
Version of Record online: 26 JUL 2011
© 2011 Bioneer A/S. FEMS Immunology & Medical Microbiology © Federation of European Microbiological Societies. Published by Blackwell Publishing Ltd
FEMS Immunology & Medical Microbiology
Volume 63, Issue 1, pages 93–107, October 2011
How to Cite
Gad, M., Ravn, P., Søborg, D. A., Lund-Jensen, K., Ouwehand, A. C. and Jensen, S. S. (2011), Regulation of the IL-10/IL-12 axis in human dendritic cells with probiotic bacteria. FEMS Immunology & Medical Microbiology, 63: 93–107. doi: 10.1111/j.1574-695X.2011.00835.x
- Issue online: 8 SEP 2011
- Version of Record online: 26 JUL 2011
- Accepted manuscript online: 27 JUN 2011 12:15PM EST
- Received 4 May 2010; revised 6 June 2011; accepted 10 June 2011., Final version published online 26 July 2011.
- dendritic cells;
- commensal bacteria;
- immune modulation;
In this study, we have used monocyte-derived dendritic cells (DCs) to design a screening model for the selection of microorganisms with the ability to suppress DC-secreted IL-12p70, a critical cytokine for the induction of T-helper cell type 1 immune responses under inflammatory conditions. By the treatment of DCs with cocktails containing TLR agonists and proinflammatory cytokines, the cells increased the secretion of the Th1-promoting cytokine IL-12p70. Clinically used probiotics were tested for their IL-10- and IL-12p70-stimulating properties in immature DCs, and showed a dose-dependent change in the IL-10/IL-12p70 balance. Lactobacillus acidophilus NCFM™ and the probiotic mixture VSL#3 showed a strong induction of IL-12p70, whereas Lactobacillus salivarius Ls-33 and Bifidobacterium infantis 35624 preferentially induced IL-10. Escherichia coli Nissle 1917 induced both IL-10 and IL-12p70, whereas the probiotic yeast Saccharomyces boulardii induced low levels of cytokines. When combining these microorganisms with the Th1-promoting cocktails, E. coli Nissle 1917 and B. infantis 35624 were potent suppressors of IL-12p70 secretion in an IL-10-independent manner, indicating a suppressive effect on Th1-inducing antigen-presenting cells. The present model, using cocktail-stimulated DCs with potent IL-12p70-stimulating capacity, may be used as an efficient tool to assess the anti-inflammatory properties of microorganisms for potential clinical use.
Specific probiotics have shown to be efficient in symptom relief or reduction of surgery-induced remission in conditions like inflammatory bowel disease (IBD) and irritable bowel syndrome (IBS). The effect of these probiotics is diverse and includes modulation of the gut immune system through interaction with gut epithelial cells and immune cells, primarily the gut-associated dendritic cells (DCs), which have the capability to respond toward microbial signals through toll-like receptor (TLR) signalling (Pamer, 2007; Coombes & Powrie, 2008; Goriely et al., 2008). A number of in vitro models and animal models have been established in order to evaluate the immune effects of microorganisms and to understand and predict the immune modulation by these microorganisms in patients suffering from either Crohn's disease (CD), ulcerative colitis (UC), arthritis or atopic dermatitis (Vanderpool et al., 2008).
CD and UC are both complex immune disorders of the gastrointestinal tract, for which the disease pathology is still being explored. CD involves mainly the activation of T-helper 1 (Th1) and to some extent T-helper 17 (Th17) responses (Fuss et al., 2006), which are believed to be activated mainly by gut-associated DCs. DC-mediated activation of Th1 and Th17 cells involves induction of DC maturation markers like CD40, CD80, CD86 and MHC II molecules, together with DC-secreted cytokines IL-12p70, IL-6, IL-23 and TGFβ, whereas T-cell secreted cytokines include tumor necrosis factor (TNFα), the Th1 cytokine IFNγ and the Th17 cytokine IL-17 (Neurath, 2007). UC is usually believed to be a Th2-driven inflammation involving cytokines like IL-4 and IL-13 (Bouma & Strober, 2003), but recent evidence has also shown the involvement of Th17-type responses in UC (Kobayashi et al., 2008).
A range of animal models and in vitro models based on primary cells have been used to screen for immune modulating properties of commensal strains. In the chemically induced TNBS mouse colitis model, Lactobacillus acidophilus NCFM™ exacerbates symptoms of inflammation, whereas Lactobacillus salivarius Ls-33 is able to significantly protect against colitis development when administered orally for 5 days before TNBS challenge (Foligne et al., 2007a). Similar models have been used to analyze the bacterium Faecalibacterium prausnitzii. Low abundance of F. prausnitzii was found to be associated with a higher risk of postoperative recurrence of ileal CD, and F. prausnitzii-induced IL-10, but very low levels of the Th1-promoting cytokines IL-12p70 and IFNγ, when incubated with human peripheral blood mononuclear cells (PBMCs) (Sokol et al., 2008). An interesting ex vivo model involving the test of microorganisms for immune modulation showed that two Lactobacillus strains were able to suppress TNFα production when the strains were incubated with inflamed mucosa from CD patients in vitro (Borruel et al., 2002).
Immune regulatory effects of species of Lactobacillus and Bifidobacterium were determined in DCs, and shown to modulate both cytokine secretion and maturation in different ways. Interestingly, several strains are able to suppress IL-12p70 secretion from DCs treated with a potent IL-12p70-inducing strain, but not lipopolysaccharide-induced IL-12p70, indicating that probiotic strains have the ability to interact with and influence DC-mediated cytokine secretion (Christensen et al., 2002).
Whole-blood assays showed that the addition of a probiotic mixture of eight different microbial strains, VSL#3, caused the specific induction of IL-10 and reduced IL-12p70 in steady-state CD11+ DCs, and VSL#3 was able to suppress lipopolysaccharide-induced IL-12p70 production. The same IL-10/IL-12p70 balance was seen when lamina propria CD11+ DCs from human intestine were treated with VSL#3 (Hart et al., 2004).
With the complex immune pathological picture of IBD in mind, probiotic treatment of IBD should be carefully conducted, because strains beneficial in one condition might be harmful in another. The intestinal probiotics coexist in a fine balance with pathogens, and disturbances of this balance might be a contributing factor to disease development as proposed by Sokol and colleagues. Furthermore, an imbalanced intestinal microbiota might be altered by the addition of probiotics, in order to re-establish intestinal homeostasis. Examples of such probiotics are Escherichia coli Nissle 1917, which has shown efficiency in maintaining remission in UC (Rembacken et al., 1999; Kruis et al., 2004), and one study in CD showing a reduced risk of relapse (Malchow, 1997). Bifidobacterium infantis 35624 has not shown to be effective in UC, but causes the relief of symptoms in IBS (Whorwell et al., 2006). IBS is immunologically considered as a Th1-related inflammatory condition, involving elevated expression of proinflammatory cytokines like TNFα, IL-6 and IL-1β and a decreased IL-10/IL-12p70 ratio (O'Mahony et al., 2005; Öhman et al., 2009). VSL#3 efficiently maintains remission in UC patients (Bibiloni et al., 2005), and also in pouchitis (Mimura et al., 2004), and induces relief of symptoms in IBS (Kim et al., 2005). Finally, the yeast Saccharomyces boulardii has been shown to maintain clinical remission in both CD and UC (Guslandi et al., 2000, 2003) and can reduce symptoms in IBS (Maupas et al., 1983).
In the present study, we have analyzed the immune modulatory properties of five single microbial strains and the microbial mixture VSL#3 containing eight different microbial strains. For all six microbial samples, clinical data or experimental animal data are available for comparison and correlation with our proposed model. For VSL#3, B. infantis 35624 and E. coli Nissle 1917, the cytokine responses seen in our study are correlated with the clinical data available for these bacteria.
Saccharomyces boulardii was, in comparison with the other microorganisms tested, immunologically ‘inert,’ showing very subtle induction of cytokines and DC markers CD40, PDL1, PDL2 and ILT3, and was unable to suppress the proinflammatory cocktails. We suggest that the present DC model, and in particular the use of cocktails that promote a Th1-directing DC phenotype, can be valuable tools in the identification and screening of novel probiotic bacteria, with implications for use in IBD and IBS.
Materials and methods
DC development from monocytes
PBMC were purified from buffy coats of healthy donors by centrifugation over a Ficoll-hypaque (GE-healthcare) gradient. Monocytes were isolated from PBMC by the positive selection of CD14+ cells by magnetic beads (Dynal, Invitrogen, San Diego, CA) according to the manufacturer's instructions. The CD14+ monocytes were cultured in six-well plates at a concentration of 2 × 106 cells mL−1 (3 mL per well) in RPMI/5% FCS supplemented with recombinant human GM-CSF (20 ng mL−1) and IL-4 (20 ng mL−1) (PeproTech, Rocky Hill, NJ). The medium was changed after 2 and 3 days. After 6 days of culture, the immature DCs were recultured into 96-well plates at a concentration of 105 cells per well, and left untreated or treated with lipopolysaccharide (Sigma-Aldrich, Brøndby, Denmark) or cocktail for 24 h.
Preparation of microorganisms
Probiotic bacteria were grown anaerobically in MRS broth (Oxoid, Nepean, Canada) at 37 °C for 48 h to the stationary phase. For bifidobacteria, 2.8 mM l-cysteine and 20 mM dl-threonine were added to the growth medium. Saccharomyces boulardii were grown in YPG broth [2% (w/v) glucose (Merck, Darmstadt, Germany), 1% (w/v) Bacto peptone (Difco, Detroit, MI), 0.5% (w/v) yeast extract (Difco), pH 5.6] at 30 °C. Cells were harvested, washed twice in phosphate-buffered saline (PBS) and pelleted by centrifugation before freeze-drying. Before the addition of bacteria to DCs, the freeze-dried bacteria were resuspended in RPMI media supplemented with 5% FCS, and added directly to the DCs. To determine the amount of bacteria, tubes were weighed empty and with freeze-dried bacteria. Bacteria included in the experiments are shown in Table 1, with information on CFU per mg freeze-dried bacteria. VSL#3, which contains eight different bacteria, was used freeze-dried as supplied by the manufacturer (Sigma-Tau Pharmaceuticals Inc.). Initial experiments showed that DC cytokine responses did not differ for bacteria harvested in the early stationary (24 h) vs. the late stationary (48 h) phase, and neither did we observe a significant difference in DC maturation between the addition of live and freeze-dried bacteria.
|Microorganism||Description (CFU mg−1)||Source|
|Bifidobacterium infantis 35624||Commercial strain 4.4 ± 2.2 × 107||Procter & Gamble, OH|
|Lactobacillus acidophilus NCFM™||Commercial strain 1.4 ± 1.7 × 107||Danisco, Copenhagen, Denmark|
|Lactobacillus salivarius Ls-33||Commercial strain 8.5 ± 0.6 × 106||Danisco|
|Escherichia coli Nissle 1917||Commercial strain 3.0 ± 0.3 × 109||Ardeypharm, Herdecke, Germany|
|VSL#3: Lactobacillus acidophilus MB443 Lactobacillus delbrueckii subps bulgaricus MB453 Lactobacillus casei MB 451 Lactobacillus plantarum MB452 Bifidobacterium longum Y10 Bifidobacterium infantis Y1 Bifidobacterium breve Y8 Streptococcus salivarius subsp. thermophilus MB455||Mixture of eight different bacteria* 2 × 108||VSL Phamaceuticals, Gaithersburg, MD|
|Saccharomyces boulardii||Probiotic yeast 7.4 ± 0.2 × 106 (Saccharomyces cerevisiae strain)||Ardeypharm|
Measurements of cytokines
Cytokines were determined using enzyme-linked immunosorbent assay (ELISA) kits for human IL-12p70, TNFα (R&D Systems, Minneapolis, MN), IFNγ, IL-17 and IL-23 (eBioscience, San Diego, CA). DCs were set up in 96-well plates with 100 000 cells per well, and probiotic tests were performed in triplicates. After 24 h of incubation with cocktails, the cell supernatants were diluted to reach the linear range for each ELISA assay. ELISAs were prepared immediately after the removal of the media from the wells or from media that were stored at −80 °C until analysis.
The total amount of each of the cytokines, IL-10 and IL-12p70, secreted in the range of 1–200 μg mL−1 bacterium was calculated as the area under the curve drawn between the average cytokine secretion from the DCs at each concentration of bacterium. The method used was based on numerical integration – the rectangular rule. The bacterial concentrations used in the experiments were used to define the intervals for the calculations of the area of the rectangles. The calculation of the area of the rectangles was carried out as follows:
where x1 is the lower concentration of bacteria, x2 is the higher concentration, y1 is the level of cytokine induced by x1 concentration of bacteria and y2 is the level of cytokine induced by x2 concentration of bacteria.
DC cocktail screening and validation
Potent IL-12p70-stimulating cocktails were designed by a series of combinations of proinflammatory cytokines and TLR agonists. Cocktail 1 contains lipopolysaccharide (0.1 μg mL−1), IFNγ (20 ng mL−1) and TNFα (50 ng mL−1), cocktail 2 contains IFNγ (20 ng mL−1), TNFα (50 ng mL−1), polyinosinic polycytidylic acid (Poly I : C, 12.5 μg mL−1), IL-1β (10 ng mL−1) and IFNα (6 ng mL−1) and cocktail 3 contains IFNγ (20 ng mL−1), Poly I : C (10 μg mL−1) and IL-1β (10 ng mL−1). The cytokines TNFα, IFNγ, IFNα, IL-6 and IL-1β were from Peprotech (London, UK) and were all diluted in cell media to reach the indicated concentrations. Lipopolysaccharide and Poly I : C were from Sigma (St. Louis, MI). TLR agonists were of TLR grade. Cocktails were combined in a 5 × working stock and added to DCs 6 h after the addition of probiotics. Cell density was 105 cells per well in 96-well plates. For validation experiments, dexamethasone was added (0.1–1 μM) 4 h before the addition of cocktail.
IL-10 neutralization study
Recombinant human IL-10 Peprotech (London, UK) was added at 20 ng mL−1 either alone or combined with 200 μg mL−1B. infantis 35624 or E. coli Nissle 1917. After 4 h of incubation, cocktail was added in order stimulate the DCs toward an IL-12-producing phenotype. To investigate the effect of IL-10 in the system, the neutralizing monoclonal Ab (mAb) MAB217 (R&D Systems) was added dose-dependently at the time of addition of either the recombinant IL-10 protein or the probiotic strain. IL-12p70 was measured after 24 h.
IL-17 response in the PBMC model
PBMCs were purified using the hypaque-Ficoll method (Foligne et al., 2007a). 105 cells were plated per well in 96-well round-bottom plates, and test reagents were added. The system was optimized for probiotic concentration (1, 10 and 100 μg mL−1) and timing for IL-17 production. Similar results were seen for all the doses tested, and IL-17 was maximally produced after 3 days. Concanavalin A, (Sigma) was added at 10 μg mL−1 as a positive control for IL-17 production.
Flow cytometric analysis
Harvested DCs were washed twice with PBS supplemented with 1% FBS. Fc receptors were blocked with excess human immunoglobulin G (IgG; Sigma-Aldrich) on ice for 10 min. Immunofluorescence staining was performed by incubation of DCs for 30 min at 4 °C with each mAb diluted to the optimal concentration according to the manufacturer's instructions. A double staining was performed with anti-CD11C-APC combined with one of the following mAbs: anti-HLA-DR-Pe, anti-CD80-Pe, anti-CD86-Pe, anti-CD40-Pe, anti-PDL1-Pe, anti-PDL2-Pe and anti-ILT3-Pe (all antibodies are from Becton Dickinson, Pharmingen). Relevant isotype controls were always used. Samples were acquired on a FACSArray (Becton Dickinson). At least 5000 mononuclear cells were gated using a combination of forward-angle and side scatter to exclude dead cells and debris. Data were analyzed using facsdiva software.
Differential induction of IL-10 and IL-12p70 from DCs treated with microorganisms in different doses
The balance between the secretion of IL-10 and IL-12p70 in immune models has been an attractive parameter for the determination of the immune modulatory properties of microorganisms (Christensen et al., 2002; Hart et al., 2004; Foligne et al., 2007a). We wished to determine the IL-10/IL-12p70 balance for selected probiotics and microorganisms with a history of immune modulation in clinical trials or experimental models. Each microorganism was added at several doses ranging from 1 to 1000 μg mL−1 to human monocyte-derived DCs (Fig. 1). As a representative experiment, monocyte-derived DCs from a human donor showed that two strains induced the secretion of IL-10 with no detectable IL-12p70 secretion (L. salivarius Ls-33 and B. infantis 35624). Lactobacillus acidophilus NCFM™ and E. coli Nissle 1917 caused the induction of both IL-10 and IL-12p70, and the probiotic mixture VSL#3 caused the induction of primarily IL-12p70. Saccharomyces boulardii caused very low induction of both IL-10 and IL-12p70. For L. acidophilus NCFM™ and E. coli Nissle 1917, we noticed a dose-dependent change in the IL-10/IL-12p70 balance, because doses at 100 μg mL−1 or below yielded an IL-10/IL-12p70 ratio below 1, and at doses above 100–200 μg mL−1, the IL-10/IL-12p70 ratio was above 1.
Because there was a dose-dependent change in the IL-10/IL-12p70 balance for some strains, and a variation in the response using DCs from different donors, we performed dose–response experiments in multiple donors. In order to normalize responses from different donors, four control strains were included in each experiment at a dose of 100 μg mL−1 (L. acidophilus NCFM™, B. infantis 35624, L. salivarius Ls-33, VSL#3), representing bacteria that induce different IL-10 and IL-12p70 responses. We analyzed the five strains and VSL#3 in multiple donors (n=6–12) and normalized their responses using the four control samples for each donor. The total amount of cytokine, secreted over the range of 1–200 μg mL−1 of added microorganism, was integrated (the area under the curve was determined for each cytokine) and is shown in Fig. 2. The maximum concentration limit at 200 μg mL−1 was used instead of 1000 μg mL−1 because the highest dose for most strains induced preferentially IL-10 and not IL-12p70. Furthermore, a concentration of 1000 μg mL−1 correlates to a bacteria : DC ratio of approximately 200–2000 depending on the strain type, which can be considered a nonphysiological ratio based on current knowledge of DC–microbial interaction in the gut (Christensen et al., 2002; Vanderpool et al., 2008). Two strains caused the secretion of significantly more IL-10 than IL-12p70: B. infantis 35624 (P<0.005) and L. salivarius Ls-33 (P<0.005). Two probiotics caused significantly stronger secretion of IL-12p70 than IL-10: L. acidophilus NCFM™ (P<0.05) and VSL#3 (P<0.005). VSL#3 caused a sevenfold higher induction of IL-12p70 than IL-10. For E. coli Nissle 1917 and S. boulardii, there were no significant differences between the induction of IL-12p70 and IL-10, and S. boulardii showed a remarkable lack of cytokine induction when added to DCs. These results show that care has to be taken when interpreting results from microbial strains in immune models, and that immune modulating properties should optimally be determined in a range of concentrations and not at one single concentration, when selecting the most promising microorganisms for clinical trials.
Selection of bacteria with potent IL-12p70-suppressing abilities using proinflammatory cocktails as a screening tool
Because E. coli Nissle 1917 caused the induction of IL-10 and IL-12p70 to similar levels (Fig. 2), the previously assumed correlation between a high IL-10/IL-12p70 ratio and a potential anti-inflammatory effect (Hart et al., 2004; Foligne et al., 2007a) did not apply to this probiotic bacterium, which has shown beneficial effects in both CD and UC (Malchow, 1997; Rembacken et al., 1999; Kruis et al., 2004). Thus, we wished to examine whether E. coli Nissle 1917 and the other probiotics were able to suppress IL-12p70 secretion when added before a cocktail with potent IL-12p70-stimulating capacity, mimicking a DC-induced Th1-directed response (Coombes & Powrie, 2008). Based on previous work with the identification of IL-12p70-stimulating cocktails containing proinflammatory cytokines and TLR agonists, three cocktails were selected with potent ability to stimulate IL-12p70 secretion (Jensen & Gad, 2010). All three cocktails significantly induced IL-12p70 from DCs compared with untreated and lipopolysaccharide-treated cells (Fig. 3a). When microorganisms were added at a dose of 100 μg mL−1 to the DCs 6 h before the addition of cocktail 1, and incubated for another 24 h, several strains were able to prevent the cocktail-induced IL-12p70 secretion (Fig. 3b). The most potent strain was E. coli Nissle 1917, which was able to invert the IL-10/IL-12p70 balance completely, compared with cocktail treatment alone. Escherichia coli Nissle 1917 induced high levels of IL-10, while IL-12p70 secretion was strongly reduced. Bifidobacterium infantis 35624 showed a similar effect, although slightly less potent, with a reduction of IL-12p70 secretion and increased IL-10 secretion. Both strains were more potent in suppressing IL-12p70 secretion than a 4-h treatment of DCs with 1 μM of dexamethasone, and furthermore, increased the secretion of the anti-inflammatory cytokine IL-10. Lactobacillus salivarius Ls-33 was able to partly reduce IL-12p70 secretion, and induce IL-10 secretion, while the two most potent IL-12p70-stimulating strains L. acidophilus NCFM™ and VSL#3 further promoted IL-12p70 secretion.
The ability to suppress IL-12p70 secretion from cocktail-treated DCs was further explored using different doses of the four clinically used probiotic microorganisms: E. coli Nissle 1917, B. infantis 35624, S. boulardii and VSL#3, and the three cocktails shown in Fig. 3a (Fig. 4a–c). The probiotics were added 4 h before the addition of the cocktails in doses of 1–50–200 μg mL−1, and dexamethasone was added at 0.1 and 1.0 μM. For all three cocktails, E. coli Nissle 1917 was able to suppress IL-12p70 secretion at all concentrations tested, with a potency at 1 μg mL−1 being similar to the suppression seen for 1.0 μM dexamethasone. Bifidobacterium infantis 35624 was also dose-dependently able to suppress the IL-12p70 secretion from all three cocktails, although the inhibition was less potent compared with E. coli Nissle 1917. Saccharomyces boulardii and VSL#3 were not able to suppress IL-12p70 secretion, and rather seemed to further support the cocktail-induced IL-12p70 secretion, in particular in combination with cocktail 3.
Is IL-10 induction involved in the IL-12p70-suppressive activity seen for E. coli Nissle 1917 and B. infantis 35624
IL-10 is a potent anti-inflammatory cytokine, which may be involved in mediating the suppressive activity seen for E. coli Nissle 1917 and B. infantis 35624 in the suppression of cocktail-induced IL-12p70 secretion. We wished to test whether IL-10 secretion was important for the suppressive activity of these two strains using IL-10-neutralizing antibody. First, we tested whether IL-10 by itself is able to suppress cocktail-induced IL-12p70 secretion, by adding recombinant IL-10, 4 h before the addition of cocktail 1 (Fig. 5a). Indeed, 20 ng mL−1 of IL-10 could suppress cocktail-induced IL-12p70 secretion, and this suppression was prevented by the addition of the IL-10-neutralizing antibody MAB217 in a dose-dependent manner (Fig. 5a). In order to test whether the secretion of IL-10 induced by E. coli Nissle 1917 and B. infantis 35624 could be responsible for the IL-12p70-suppressive activity, we added 200 μg mL−1 bacteria together with increasing concentration of the neutralizing antibody, 4 h before the addition of cocktail (Fig. 5b). In this case, the antibody was not able to prevent the IL-12p70-suppressive activity of the two probiotics, showing that IL-10 secretion is not directly involved in the inhibitory activity of E. coli Nissle 1917 and B. infantis 35624.
Saccharomyces boulardii-treated DCs show reduced stimulation of proinflammatory cytokines TNFα, IL-6, IL-12p40, IL-23 and CXCL1
Saccharomyces boulardii induced very low levels of IL-10 and IL-12p70 (Figs 1 and 2) and did not suppress any of the three IL-12p70-inducing cocktails, which led us to speculate whether the clinical effects of S. boulardii are not associated with DC interaction and possibly independent of effects on immune cells. We speculated whether S. boulardii was able to stimulate other proinflammatory cytokines and chemokines when added to human DCs. A dose–response study was conducted at doses of 1–50–200 μg mL−1, and the levels of TNFα, IL-6, IL-23, IL-12p40 and CXCL1 were measured in the culture supernatant (Fig. 6a–e). The Th17-inducing DC cytokine IL-23, identified in inflammatory tissue from CD patients (Fuss et al., 2006; Kobayashi et al., 2008), was significantly higher in DCs from E. coli Nissle 1917, L. acidophilus NCFM™ and B. infantis 35624-treated DCs than S. boulardii-treated DCs (P<0.05–0.005), with basically nondetectable levels in S. boulardii-treated DCs (Fig. 6a).
IL-6 induction by S. boulardii-treated DCs was significantly lower than for DCs treated with all other strains (P<0.005) (Fig. 6b), whereas IL-12p40 secretion was the lowest for S. boulardii- and L. salivarius Ls-33-treated DCs (Fig. 6c). TNFα and CXCL1 were induced in a dose-dependent manner from all six types of probiotics (E. coli Nissle 1917, B. infantis 35624, L. salivarius Ls-33, L. acidophilus NCFM™, S. boulardii and VSL#3); however, S. boulardii induced lower levels than all the other strains (P<0.05–0.005) (Fig. 6d and e). Taken together, these results show that DCs respond weakly toward the presence of S. boulardii and indicate that S. boulardii is immunologically ‘silent’ due to the low ability to induce cytokines and chemokines. Furthermore, E. coli Nissle 1917 was the most potent inducer of the seven cytokines and chemokines tested, which is a surprise considering the ability to suppress cocktail-induced IL-12p70 secretion (Figs 3 and 4). In particular, the strong secretion of E. coli Nissle 1917-treated DCs of the Th17-promoting cytokines IL-6 and IL-23 was surprising, and made us speculate whether E. coli Nissle 1917 possessed Th17-inducing properties. In an autologous PBMC model, we added the six probiotic strains and measured the ability of the strains to induce IL-17 after 3 days of incubation. None of the strains were able to induce IL-17 to levels as Concanavalin A-treated cells, and even showed IL-17 levels below the level seen for untreated control cells (Fig. 6f), showing that although E. coli Nissle 1917 induces IL-6 and IL-23 from DCs, the strain is not able to induce Th17 cell differentiation in a PBMC model.
Probiotic-treated DCs show differential expression of stimulatory and inhibitory DC receptors
DCs activate the adaptive immune system by a combination of cytokine and chemokine secretion together with physical interaction with T-cells through membrane receptors (Coombes & Powrie, 2008). We wished to investigate whether DC-associated maturation markers were affected in a similar manner by probiotic treatment, in particular, for S. boulardii-treated DCs, because these DCs showed low secretion of proinflammatory cytokines. We analyzed classical markers associated with DC-stimulated T-cell activation HLA-DR, CD40, CD80 and CD86, but also receptors associated with inhibition of DC-mediated T-cell activation of PDL1, PDL2 and ILT3 (Fig. 7) (Cella et al., 1997; Zhang et al., 2006; Wang et al., 2008). The markers were analyzed on DCs treated with three concentrations of probiotic for 24 h (1–10–100 μg mL−1), and the mean fluorescence intensity (MFI) values for probiotic-treated DCs were adjusted to untreated cells for each donor to compensate for differences in basal expression between the donors, and expressed as fold increase compared with untreated cells normalized to 1. Because of large datasets, representative dot-blots for gating, isotype controls and CD11c and CD80 dot-blots are shown (Fig. 7a–c), and representative histograms are shown for one donor on the CD80 marker for untreated DCs, cocktail 1 and E. coli Nissle 1917-treated DCs (Fig. 7d–f). Datasets for three donors were prepared and expressed in fold change vs. untreated DCs. The most dramatic changes were seen for CD40, PDL1, PDL2 and ILT3 (Fig. 7g). CD40 was induced significantly for B. infantis 35624 (P<0.01), VSL#3 (P<0.05) and E. coli Nissle 1917 (P<0.05) compared with untreated cells, whereas E. coli Nissle 1917 and VSL#3 (P<0.05) induced significantly higher CD80 levels than S. boulardii (Fig. 7g). For the inhibitory molecules, S. boulardii and B. infantis 35624 showed significant induction of PDL1 compared with untreated cells (P<0.05), whereas B. infantis 35624, VSL#3 and E. coli Nissle 1917 induced significantly higher levels than S. boulardii (all P<0.05). Bifidobacterium infantis 35624, VSL#3 and E. coli Nissle 1917 all induced significantly higher levels of PDL2 than untreated cells (P<0.05), and finally, VSL#3 (P<0.05) and E. coli Nissle 1917 (P<0.05)-treated DCs induced ILT3 significantly compared with S. boulardii-treated DCs. Expressions of HLA-DR, CD80 and CD86 were only slightly induced, for example with B. infantis 35624 showing significantly increased CD80 expression vs. untreated cells. These data show that indeed S. boulardii is a weaker inducer of DC maturation markers like CD40, CD80 and PDL1, in particular compared with B. infantis 35624, VSL#3 and E. coli Nissle 1917, which is in line with what was seen for the secretion of inflammatory cytokines and chemokines (Fig. 6). Secondly, B. infantis 35624, VSL#3 and E. coli Nissle 1917 are potent inducers of the inhibitory DC receptors compared with both untreated and S. boulardii-treated DCs, indicating that these strains might be involved in induction of tolerance (Cella et al., 1997; Zhang et al., 2006; Wang et al., 2008).
In the present study, we have used monocyte-derived DCs to investigate the immune modulatory properties of microorganisms with beneficial clinical effects in either IBD or IBS, or with well-documented effects in animal models of TNBS-induced colitis. We have correlated these clinical and experimental data with the results obtained in this DC-based in vitro model, and shown that several immune regulatory effects are important when probiotic or commensal bacteria are evaluated in immune models.
One finding was that all tested microorganisms showed a dose-dependent change in their IL-10- and IL-12p70-stimulating capacity. For strains with the ability to stimulate IL-12p70 production (L. acidophilus NCFM™, E. coli Nissle 1917 and VSL#3), the stimulation of IL-12p70 was high at the lower doses of bacteria, while IL-10 secretion increased with bacterial load. The IL-10/IL-12p70 ratio thus increases dramatically with increasing dose. This is an important finding when designing models for screening of microorganisms with potential probiotic effects and prediction of their immune modulatory effects.
This is supported by other studies, where a few strains over a lower dose range have shown similar effects (Konstantinov et al., 2008). Consequently, potential probiotic strains should be tested at several doses in DCs from several human donors, with the inclusion of control strains at a fixed dose for normalization of the test results.
A second finding in our study was that the two clinically used probiotic microorganisms (S. boulardii and E. coli Nissle 1917) would possibly not have been selected for clinical trials, if the selection criteria were based on a high IL-10/IL-12p70 ratio (Foligne et al., 2007a), because E. coli Nissle 1917 stimulated the secretion of equal amounts of IL-10 and IL-12p70 and S. boulardii stimulated nearly undetectable levels of IL-10 and IL-12p70.
A third finding was that E. coli Nissle 1917 and B. infantis 35624, which have both shown beneficial effects in Th1-associated intestinal inflammation like CD and IBS (Malchow, 1997; Whorwell et al., 2006; Plassmann & Schulte-Witte, 2007), had the ability to suppress cocktail-induced secretion of IL-12p70 from DCs to levels comparable with 1.0 μM dexamethasone. This correlates with the ability of B. infantis 35624 to reverse the IL-10/IL-12p70 balance in IBS patients (O'Mahony et al., 2005) and the ability of E. coli Nissle 1917 to reduce the risk of relapse and minimize the need for glucocorticoid treatment in CD patients (Malchow, 1997).
The fact that two distinctly different bacteria can suppress IL-12p70 secretion from three different cocktails, which activates the innate response through TLR3, TLR4 and a panel of different cytokine receptor-activated pathways, shows that E. coli Nissle 1917 and B. infantis 35624 are potent suppressors of DC activation. Although exogenously added IL-10 was able to suppress cocktail-induced IL-12p70 secretion, IL-10 secretion from DCs treated with these two probiotics was not involved in the suppressive activity of these strains, shown by the addition of a neutralizing antibody toward IL-10.
A fourth finding was that the yeast S. boulardii was unable to stimulate significant amounts of cytokines like IL-6, IL-10, IL-12p70, IL-23 and TNFα, and was not able to induce DC maturation markers or inhibitory receptors to the extent seen for the other strains, in particular for CD40, PDL1, PDL2 and ILT3, and was unable to reduce cocktail-induced IL-12p70 secretion from three different cocktails. The positive clinical effect of S. boulardii as seen in UC, CD and IBS could likely be dependent on factors other than a direct suppressive effect on DCs. A likely mechanism was shown by Thomas et al. (2009), who showed that conditioned media from S. boulardii suppressed lipopolysaccharide-induced expression of DC activation, maturation markers CD40, CD80, and the migration receptor CCR7, and decreased the secretion of proinflammatory cytokines. Other studies have shown beneficial effects of S. boulardii in intestinal inflammation, involving the prevention of pathogen adherence to the intestinal cell wall, neutralization of bacterial endotoxins or improvements of the intestinal barrier (Pothoulakis et al., 1993; Tasteyre et al., 2002).
A fifth finding in our study is the induction of immune suppressive receptors on probiotic-treated DCs: PDL1, PDL2 and ILT3. These receptors are involved in regulating and modulating immune stimulatory responses, thereby suppressing effector T-cell responses, and differentiation of regulatory T-cells required for the induction of tolerance (Cella et al., 1997; Zhang et al., 2006; Wang et al., 2008). Whether these inhibitory receptors are central to understanding the effects of probiotics in terms of immune suppression, serve as possible feedback mechanisms or are involved in induction of immune tolerance may be explored in future studies.
The finding that E. coli Nissle 1917 was the most potent inducer of TNFα, IL-6, IL-12p40, IL-23 and CXCL1 from DCs was a surprise based on the promising clinical data on this probiotic bacterium in IBD, as well as the ability of this bacterium to suppress cocktail-induced IL-12p70 secretion. In particular, the high induction of IL-6 and IL-23, which are key initiators of Th17 type responses, is surprising based on the ability to suppress Th17-associated diseases (Malchow, 1997; Rembacken et al., 1999; Kruis et al., 2004). However, E. coli Nissle 1917 was not able to induce significantly higher levels of IL-17 in a PBMC model, despite the strong induction of IL-6 and IL-23 from DCs.
Two Lactobacillus strains, L. salivarius Ls-33 and L. acidophilus NCFM™, with well-known effects in experimental models of colitis, showed strong opposite effects in TNBS-induced colitis in mice when given in a preventive study 5 days before TNBS treatment. Lactobacillus salivarius Ls-33 showed a strong ability to suppress colitis development, whereas L. acidophilus NCFM™ caused a more severe colitis development (Foligne et al., 2007a). Furthermore, oral intake of L. acidophilus NCFM™ in a human clinical trial induced an increase in antigen-specific IgA and IgM titers in serum, which was not seen after the intake of L. salivarius Ls-33, suggesting a stronger immune response initiated by L. acidophilus NCFM™ (Paineau et al., 2008).
In line with these observations, L. salivarius Ls-33 and L. acidophilus NCFM™ showed, in our study, IL-10/IL-12p70 ratios that support the above findings. When the total amount of cytokine was calculated over the range of 1–200 μ mL−1 of added bacteria to the DCs, L. salivarius Ls-33 showed an IL-10/IL-12p70 ratio of approximately 10, whereas L. acidophilus NCFM™ showed an IL-10/IL-12p70 ratio of approximately 0.4. Furthermore, L. salivarius Ls-33 showed an ability to suppress cocktail-induced IL-12p70, although not as potent as E. coli Nissle 1917 and B. infantis 35624, while L. acidophilus NCFM™ further stimulated IL-12p70 secretion, an observation that we have seen reproducibly in both human and mouse DCs.
VSL#3 is efficient in the maintenance of remission in UC, and is successful in combination with standard therapy in patients with UC (Bibiloni et al., 2005). Furthermore, VSL#3 has shown promising effects in pouchitis and can reduce symptoms in IBS (Gionchetti et al., 2003; Kim et al., 2003; Mimura et al., 2004; Tursi, 2004). We showed that VSL#3 was as potent as L. acidophilus NCFM™ in inducing IL-12p70 in DCs, and was not able to block cocktail-induced IL-12p70 secretion from any of the cocktails. Our findings for VSL#3 are in contrast to the findings obtained by Hart et al. (2004), who incubated VSL#3 with whole human blood and subsequently analyzed the blood DCs for IL-10 and IL-12p70 content. These studies showed increased IL-10 with the baseline IL-12p70 content and showed that VSL#3 was able to suppress lipopolysaccharide-induced IL-12p70 production intracellularly. Other results support our finding, with preferential IL-12p70 production from VSL#3-treated murine bone marrow-derived DCs (Drakes et al., 2004).
Comparison of results from different types of immune models on the same strains is important in order to understand the effect of probiotics in clinical and experimental studies. Steady-state DCs from whole blood or from inflammatory tissue from CD or UC patients as used by Borruel et al. (2002) or Hart et al. (2004) are suitable for examination of a small number of probiotic samples, but due to limited cell numbers, patient heterogeneity, tissue accessibility and potential alteration of the DC phenotype due to purification procedures, such studies are less suitable for screening purposes of potential probiotic strains. In contrast, studies on human monocyte-derived or mouse bone marrow-derived DCs are suitable for screening purposes, but these DCs might diverge phenotypically from the DCs present in the intestine, depending on the differentiation procedure applied. Several DC differentiation protocols have been developed for different applications. A majority of models apply IL-4 and GM-CSF to mouse bone marrow or human monocytes during differentiation, resulting in immature DCs with a preinflammatory phenotype (Zeuthen et al., 2006; Shortman & Naik, 2007). Such cells are suitable for inflammatory screening, as applied in our present cocktail-induced IL-12p70-producing model, and may be similar to the DC phenotype seen in intestinal inflammation compared with steady-state DCs as seen in, for example, whole blood.
We have, in our model, focused on DCs and their production of IL-10 and IL-12p70 and to some extent on IL-23 produced from probiotic-treated DCs. The IL-10/IL-12p70 balance has, in both experimental models of colitis and in clinical studies, been shown to be important for disease development. First, DCs treated with the IL-10-inducing strain L. salivarius Ls-33 could reduce disease symptoms when transferred to TNBS-induced colitic mice (Foligne et al., 2007b). Second, an engineered recombinant Lactococcus lactis strain with the constitutive expression of murine IL-10 could reduce disease symptoms in a DSS-induced mouse colitis model (Steidler et al., 2000). Third, in human studies, neutralization of the IL-12/IL-23 p40 subunit in CD patients, but not in UC patients, improved the clinical score and reduced the production of IL-12p70, IL-23, IL-6 and IL-17 in lamina propria cells after treatment, emphasizing the importance of the p40 IL-12/IL-23 subunit in the development of CD (Fuss et al., 2006).
Screening models used for the selection of promising immune suppressive strains should optimally be able to indicate the overall immune effects of a given strain in itself in terms of the IL-10/IL-12 balance, and should potentially show the ability of a strain to suppress immune responses. The effect of the same strain in the complex microbiota of the intestinal content is far more complex (Round & Mazmanian, 2009). A model to screen for the probiotic effect in such a complex environment should optimally constitute a model with a functional epithelial barrier including M-cells and DCs able to sense the lumen content by dendrites protruding the epithelial barrier (Coombes & Powrie, 2008; Round & Mazmanian, 2009). In this regard, the model presented in our study, and a majority of other in vitro screening models, are limited due to the lack of complexity seen in the functional intestine. However, in vitro models are able to contribute with basic knowledge on specific strains and mechanisms of immune regulation, which should be taken into consideration when selecting microbial strains for experimental and clinical testing.
Our study suggests that the evaluation of potential probiotics to influence the human immune system should be considered carefully, when selecting strains for clinical trials. Our study supports the previously used IL-10/IL-12p70 balance as a good predictive parameter for the evaluation of some types of bacteria like Lactobacillus strains, as seen for L. salivarius Ls-33 and L. acidophilus NCFM™, but further suggests applying a dose–response evaluation in order to make a correct interpretation of the response. Furthermore, our study suggests an additional model, where the anti-inflammatory potential of bacteria can be evaluated using a strong proinflammatory stimuli like one of the identified IL-12p70-stimulating cocktails. The ability of a potential probiotic bacterium to suppress the cocktail-induced IL-12p70 secretion, as seen in our study for E. coli Nissle 1917 and B. infantis 35624, correlated well with the clinical observation for these two bacteria.
This work was supported by the Danish Agency for science, technology and innovation. We wish to thank Trine Møller and Bente M. Thorup for excellent technical assistance.
- 2005) VSL#3 probiotic-mixture induces remission in patients with active ulcerative colitis. Am J Gastroenterol100: 1539–1546. Direct Link:, , , , , , & (
- 2002) Increased mucosal tumour necrosis factor alpha production in Crohn's disease can be downregulated ex vivo by probiotic bacteria. Gut51: 659–664. , , , , , , , & (
- 2003) The immunological and genetic basis of inflammatory bowel disease. Nat Rev Immunol3: 521–533. & (
- 1997) A novel inhibitory receptor (ILT3) expressed on monocytes, macrophages, and dendritic cells involved in antigen processing. J Exp Med185: 1743–1751. , , , , , & (
- 2002) Lactobacilli differentially modulate expression of cytokines and maturation surface markers in murine dendritic cells. J Immunol168: 171–178. , & (
- 2008) Dendritic cells in intestinal immune regulation. Nat Rev Immunol8: 435–446. & (
- 2004) Bacterial probiotic modulation of dendritic cells. Infect Immun72: 3299–3309. , & (
- 2007a) Correlation between in vitro and in vivo immunomodulatory properties of lactic acid bacteria. World J Gastroentero13: 236–243. , , , , , , , , & (
- 2007b) A key role of dendritic cells in probiotic functionality. PLoS One2: e313. , & (
- 2006) Both IL-12p70 and IL-23 are synthesized during active Crohn's disease and are down-regulated by treatment with anti-IL-12 p40 monoclonal antibody. Inflamm Bowel Dis12: 9–15. , , , , , , , & (
- 2003) Prophylaxis of pouchitis onset with probiotic therapy: a double-blind, placebo-controlled trial. Gastroenterology 124: 1202–1209. , , , , , , , , & (
- 2008) How microorganisms tip the balance between interleukin-12 family members. Nat Rev Immunol8: 81–86. , & (
- 2000) Saccharomyces boulardii in maintenance treatment of Crohn's disease. Digest Dis Sci45: 1462–1464. , , & (
- 2003) A pilot trial of Saccharomyces boulardii in ulcerative colitis. Eur J Gastroen Hepat15: 697–698. , & (
- 2004) Modulation of human dendritic cell phenotype and function by probiotic bacteria. Gut53: 1602–1609. , , et al. (
- 2010) Differential induction of inflammatory cytokines by dendritic cells treated with novel TLR-agonist and cytokine based cocktail: targeting dendritic cells in autoimmunity. J Inflamm 7: 37. & (
- 2003) A randomized controlled trial of a probiotic, VSL#3, on gut transit and symptoms in diarrhoea-predominant irritable bowel syndrome. Aliment Pharm Therap17: 895–904. , , , , , & (
- 2005) A randomized controlled trial of a probiotic combination VSL# 3 and placebo in irritable bowel syndrome with bloating. Neurogastroent Motil17: 687–696. , , , , , , & (
- 2008) IL23 differentially regulates the Th1/Th17 balance in ulcerative colitis and Crohns disease. Gut57: 1682–1689. , , et al. (
- 2008) S layer protein A of Lactobacillus acidophilus NCFM regulates immature dendritic cell and T cell functions. P Natl Acad Sci USA105: 19474–19479. , , , , , , , , & (
- 2004) Maintaining remission of ulcerative colitis with the probiotic Escherichia coli Nissle 1917 is as effective as with standard mesalazine. Gut53: 1617–1623. , , et al. (
- 1997) Crohn's disease and Escherichia coli. A new approach in therapy to maintain remission of colonic Crohn's disease?J Clin Gastroenterol25: 653–658. (
- 1983) Treatment of irritable bowel syndrome. Double blind trial of Saccharomyces boulardii. Médecine Chirurgie Digestives2: 77–79. , & (
- 2004) Once daily high dose probiotic therapy (VSL#3) for maintaining remission in recurrent or refractory pouchitis. Gut53: 108–114. , , , , , , , , & (
- 2007) IL-23: a master regulator in Crohn disease. Nat Med 13: 26–28. (
- 2009) T-cell activation in patients with irritable bowel syndrome. Am J Gastroenterol104: 1205–1212. , , , , , , & (
- 2005) Lactobacillus and Bifidobacterium in irritable bowel syndrome: symptom responses and relationship to cytokine profiles. Gastroenterology 128: 541–551. , , , , , , , , & (
- 2008) Effects of seven potential probiotic strains on specific immune responses in healthy adults: a double-blind, randomized, controlled trial. FEMS Immunol Med Mic 53: 107–113. , , , , , , , , & (
- 2007) Immune responses to commensal and environmental microbes. Nat Immunol11: 1173–1178. (
- 2007) Treatment of irritable bowel syndrome with Escherichia coli strain Nissle 1917 (EcN): a retrospective survey. Med Klin 102: 888–892. & (
- 1993) Saccharomyces boulardii inhibits Clostridium difficile toxin A binding and enterotoxicity in rat ileum. Gastroenterology104: 1108–1115. , , , , , & (
- 1999) Non-pathogenic Escherichia coli versus mesalazine for the treatment of ulcerative colitis: a randomised trial. Lancet 354: 635–639. , , , & (
- 2009) The gut microbiota shapes intestinal immune responses during health and disease. Nat Rev Immunol9: 313–323. & (
- 2007) Steady-state and inflammatory dendritic-cell development. Nat Rev Immunol7: 19–30. & (
- 2008) Faecalibacterium prausnitzii is an anti-inflammatory commensal bacterium identified by gut microbiota analysis of Crohn disease patients. P Natl Acad Sci USA105: 16731–16736. , , et al. (
- 2000) Treatment of murine colitis by Lactococcus lactis secreting interleukin-10. Science289: 1352–1355. , , , , , , & (
- 2002) Inhibition of in vitro cell adherence of Clostridium difficile by Saccharomyces boulardii. Microb Pathogenesis5: 219–225. , , , & (
- 2009) Saccharomyces boulardii inhibits lipopolysaccharide-induced activation of human dendritic cells and T cell proliferation. Clin Exp Immunol156: 78–87. , , , , & (
- 2004) Acute diverticulitis of the colon – current medical therapeutic management. Expert Opin Pharmaco5: 55–59. (
- 2008) Mechanisms of probiotic action: implications for therapeutic applications in inflammatory bowel diseases. Inflamm Bowel Dis 14: 1585–1596. , & (
- 2008) Programmed death 1 ligand signaling regulates the generation of adaptive Foxp3+CD4+regulatory T cells. P Natl Acad Sci USA105: 9331–9336. , , , , & (
- 2006) Efficacy of an encapsulated probiotic Bifidobacterium infantis 35624 in women with irritable bowel syndrome. Am J Gastroenterol101: 1581–1590. Direct Link:, , , , , , , & (
- 2006) Lactic acid bacteria inducing a weak interleukin-12 and tumor necrosis factor alpha response in human dendritic cells inhibit strongly stimulating lactic acid bacteria but act synergistically with gram-negative bacteria. Clin Vaccine Immunol13: 365–375. , & (
- 2006) Regulation of T cell activation and tolerance by PDL2. P Natl Acad Sci USA103: 11695–11700. , , et al. (