Authors contributed equally to this work
Colitogenic and non-colitogenic commensal bacteria differentially trigger DC maturation and Th cell polarization: An important role for IL-6
Article first published online: 18 MAY 2006
Copyright © 2006 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
European Journal of Immunology
Volume 36, Issue 6, pages 1537–1547, June 2006
How to Cite
Frick, Julia S., Zahir, N., Müller, M., Kahl, F., Bechtold, O., Lutz, Manfred B., Kirschning, Carsten J., Reimann, J., Jilge, B., Bohn, E. and Autenrieth, Ingo B. (2006), Colitogenic and non-colitogenic commensal bacteria differentially trigger DC maturation and Th cell polarization: An important role for IL-6. Eur. J. Immunol., 36: 1537–1547. doi: 10.1002/eji.200635840
- Issue published online: 24 MAY 2006
- Article first published online: 18 MAY 2006
- Manuscript Accepted: 10 APR 2006
- Manuscript Revised: 14 MAR 2006
- Manuscript Received: 4 JAN 2006
- Dendritic cells;
- E. coli;
- B. vulgatus;
- interleukin 6;
- T cells
We investigated whether commensal bacteria modulate activation and maturation of bone marrow-derived DC and their ability to prime CD4+ T cells. We used Escherichia coli mpk, which induces colitis in gnotobiotic IL-2-deficient (IL-2–/–) mice, and Bacteroides vulgatus mpk, which prevents E. coli-induced colitis. Stimulation of DC with E. coli induced TNF-α, IL-12 and IL-6 secretion and expression of activation markers. Moreover, stimulation of DC with E. coli increased T cell activation and led to Th1 polarization. Stimulation with B. vulgatus led only to secretion of IL-6, and DC were driven into a semi-mature state with low expression of activation markers and did not promote Th1 responses. B. vulgatus-induced semi-mature DC were non-responsive to stimulation by E. coli in terms of maturation, T cell priming and TNF-α but not IL-6 production. The non-responsiveness of B. vulgatus-stimulated DC was abrogated by addition of anti-IL-6 mAb or mimicked with rIL-6. These data suggest that B. vulgatus-induced IL-6 drives DC into a semi-mature state in which they are non-responsive to proinflammatory activation by E. coli. This in vitro mechanism might contribute to the prevention of E. coli-triggered colitis development by B. vulgatusin vivo; high IL-6 mRNA expression was consistently found in B. vulgatus-colonized or B. vulgatus/E. coli co-colonized IL-2–/– mice and was associated with absence of colitis.
Dendritic cells (DC) are antigen-presenting cells (APC) with an important function in the immune system. As initiators of T cell responses, they have the capacity to stimulate naive T cells 1. It is assumed that DC, as migratory cells, transport microbial antigens to the lymph nodes and the spleen for induction of systemic immune responses 2. DC effectively induce primary immune responses against microbial infections and other stimuli 3, 4. The activation of DC can be induced by infectious agents and inflammatory products 4–10. These stimuli induce DC maturation, characterized by up-regulation of costimulatory molecules and MHC class II, cytokine production and the ability to activate T cells 4, 6, 9, 11. DC recognize microbes possibly by Toll-like receptors interacting with pathogen-associated microbial patterns such as lipopolysaccharide, peptidoglycane lipoarabinomanan or DNA containing unmethylated CpG motifs 4, 12–15.
In addition to bacterial stimuli, cytokines regulate DC maturation. It is described that type I interferons, produced by virally infected cells, can activate DC 15. Furthermore, MIP-1α and MIP-1β produced by epithelial cells in response to Toxoplasma gondii antigens can trigger IL-12 production by some murine DC 16. In contrast, recent studies describe that IL-6 may suppress DC activation and maturation 17. In bone marrow-derived DC, the MHC class II, CD86, CD80 and CD40 expression levels as well as IL-12 secretion were strongly enhanced by LPS treatment but suppressed by pretreatment with IL-6. Furthermore, the capability of DC to present antigen was reduced after treatment with IL-6 in vitro. Compared to wild-type mice, splenic DC from IL-6-deficient (IL-6–/–) mice treated with LPS in vivo showed enhanced expression of CD86 17, suggesting that IL-6, at least under certain conditions, may act as an immunosuppressive cytokine for DC differentiation 17.
IL-2–/– mice are a model for microflora-triggered, chronic immune-mediated colitis in hosts with dysregulated T cell function 18–20. Recently it was shown that IL-2–/– but not IL-2+/+ gnotobiotic mice mono-colonized with Escherichia coli mpk develop colitis, whereas IL-2–/– mice colonized with Bacteroides vulgatus mpk do not 21. Furthermore, B. vulgatus mpk was able to prevent the E. coli mpk-induced colitis 21. The genus Bacteroides comprises 30% of bacteria in the feces and the mucus overlying the intestinal epithelium and represents one of the dominant bacterial divisions 22. DNA array experiments and cellular studies revealed that E. coli mpk is a non-pathogenic commensal strain.
In this study we addressed whether and how colitogenic and non-colitogenic commensal bacteria modulate DC activation, maturation and the ability to prime CD4+ T cells. For this purpose we used the commensal E. coli mpk strain, which induces colitis, and the commensal B. vulgatus strain, which prevents E. coli-induced colitis in IL-2–/– mice.
B. vulgatus and E. coli induce different TNF-α, IL-12(p70), IL-6 and IL-10 production by DC
To assess cytokine production by DC upon stimulation with colitogenic (E. coli) and non-colitogenic (B. vulgatus) bacteria, murine DC were stimulated with viable bacteria at different MOI (1, 0.1, 0.01). After 1 h gentamicin was added to kill bacteria. To exclude that E. coli or B. vulgatus overgrew the DC cultures, the CFU of viable bacteria after 20 h incubation was determined; no viable bacteria were detectable (data not shown). Furthermore, bacterial invasion assays revealed that neither E. coli nor B. vulgatus invaded and survived within DC (data not shown). Culture supernatants were collected after 20 h incubation, and the levels of TNF-α, IL-12(p70), IL-6 and IL-10 were determined by ELISA (Fig. 1). E. coli activated DC and induced TNF-α, IL-12(p70) and IL-10 secretion in a dose-dependent manner. In contrast, B. vulgatus induced no IL-12(p70), TNF-α or IL-10 secretion by DC. E. coli induced high amounts of IL-6 already at an MOI of 0.01, while DC stimulation with B. vulgatus resulted in a dose-dependent increase in IL-6 secretion. Stimulation of DC with B. vulgatus at an MOI of 1 induced comparable IL-6 levels as stimulation with the E. coli strain. Stimulation of DC with heat-killed B. vulgatus or heat-killed E. coli resulted in comparable cytokine secretion patterns (data not shown). Thus B. vulgatus, unlike E. coli, does not induce IL-12(p70) or TNF-α production but induces IL-6 secretion by DC. These data suggest that B. vulgatus, in contrast to E. coli, lacks the ability to trigger inflammatory responses by DC rather than exerting an active inhibition of cytokine production.
Maturation of DC is differentially modulated byE. coli and B. vulgatus
In order to investigate whether and how E. coli and B. vulgatus modulate DC maturation, DC were stimulated with E. coli or B. vulgatus and analyzed for CD80, CD86, MHC class II and CD40 surface expression by flow cytometry. E. coli stimulated DC to express high levels of MHC class II, CD80, CD86 and CD40 and led to highly activated and matured DC. In contrast, expression of these surface molecules was nearly unaffected in DC exposed to B. vulgatus, corresponding rather to their expression on unstimulated DC, suggesting a reduced ability of B. vulgatus to activate and maturate DC (Fig. 2a).
B. vulgatus prevents E. coli-induced TNF-α secretion by DC
In gnotobiotic IL-2–/– mice, E. coli-induced colitis could be prevented by co-colonization with B. vulgatus, which was associated with decreased TNF-α and IFN-γ mRNA expression in colon tissue 21. To examine whether B. vulgatus prevents the E. coli-induced secretion of proinflammatory cytokines such as TNF-α and IL-12(p70) in vitro, DC were exposed to B. vulgatus prior to stimulation with E. coli. The E. coli-induced TNF-α secretion was reduced by pretreatment with B. vulgatus in a dose-dependent manner to almost 20% (pretreatment of DC with B. vulgatus at an MOI of 0.01, 0.1 or 1; Fig. 3a). Likewise, the E. coli-induced IL-12 secretion was inhibited by prestimulation with B. vulgatus (data not shown). Moreover, in contrast to reduction of TNF-α and IL-12, IL-6 levels increased in a dose-dependent manner in E. coli-stimulated DC that were pretreated with B. vulgatus, suggesting that semi-mature DC are selectively non-responsive to E. coli-induced TNF-α and IL-12 production but not to E. coli-induced IL-6 production (Fig. 3a). Notably, we observed that upon pretreatment with B. vulgatus, DC stimulated with E. coli expressed surface markers at lower levels as compared to DC stimulated with E. coli only (Fig. 2b). Therefore B. vulgatus seems to suppress the responsiveness of DC to E. coli concerning TNF-α and IL-12 production as well as maturation, including expression of DC surface markers. In order to exclude that the inhibitory effect of B. vulgatus is due to down-regulation of TLR2 or TLR4, we analysed TLR2 and TLR4 mRNA expression by RT-PCR. Stimulation of DC with B. vulgatus had no effect on TLR2 expression but increased TLR4 expression. Therefore we can exclude that down-regulation of TLR accounts for the reduced responsiveness of DC treated with B. vulgatus (data not shown). This observation is in line with recent studies demonstrating that IL-6 may drive DC into a semi-mature phenotype, with reduced responsiveness towards a subsequent LPS stimulus 23.
When DC were pretreated with E. coli, the DC produced TNF-α, IL-12 and IL-6 and were mature as indicated by high MHC class II expression. In keeping with previous data, these DC did not respond to a second E. coli stimulus with regard to TNF-α production (Fig. 3b). However, while this state of non-responsiveness is an inherent feature of mature DC, the non-responsiveness in B. vulgatus-treated semi-mature DC is probably based on a different mechanism.
Supernatant of DC stimulated with B. vulgatus or rIL-6 blocks E. coli-induced DC maturation and TNF-α production
As described above, stimulation of DC with E. coli induces TNF-α, IL-12(p70) and IL-6, whereas stimulation of DC with B. vulgatus leads only to secretion of IL-6 and induces a semi-mature DC phenotype less responsive to subsequent E. coli stimulation. This might be an IL-6-dependent process since IL-6, in the absence of TNF-α and IL-12(p70), is supposed to suppress DC activation and maturation 17.
Therefore we hypothesized that B. vulgatus-induced IL-6 might promote the development of semi-mature DC. To test this assumption, we investigated whether supernatants of B. vulgatus-treated DC can inhibit E. coli-induced TNF-α secretion. For this purpose, immature DC were incubated with the supernatant of DC stimulated with B. vulgatus at an MOI of 1 (B. vulgatus-conditioned supernatant) prior to stimulation with E. coli. The DC incubated with B. vulgatus-conditioned DC supernatant produced significantly less TNF-α (10% supernatant conferred ∼70% reduction in TNF-α secretion) upon E. coli stimulation than control DC (Fig. 4a). Furthermore, incubation of DC with B. vulgatus-conditioned supernatant resulted in reduced CD40 expression upon E. coli stimulation compared to the expression on DC stimulated with E. coli only (Fig. 4b).
Since it was described that IL-6 affects DC maturation 17, we hypothesized that B. vulgatus-induced IL-6 accounts for the inhibitory effects. To evaluate this hypothesis, DC were incubated with recombinant IL-6 for 24 h, followed by stimulation with E. coli (Fig. 4c). We observed a significant decrease in E. coli-induced TNF-α secretion by the IL-6-pretreated DC compared to DC stimulated with E. coli alone. Likewise, expression of CD40 upon E. coli stimulation was decreased in IL-6-pretreated DC as compared to DC stimulated with E. coli only (Fig. 4b). These results demonstrate that IL-6 mimics the effect of B. vulgatus-conditioned DC supernatant.
In a second experiment, immature DC were treated with B. vulgatus-conditioned supernatant plus neutralizing anti-IL-6 mAb (Fig. 4d) prior to stimulation with E. coli. The supernatant mixed with anti-IL-6 mAb did not mediate suppression of TNF-α secretion induced by E. coli; the TNF-α levels observed were equal to TNF-α levels released from untreated E. coli-stimulated DC.
From these results we can conclude that (i) B. vulgatus induces secretion of IL-6 but not TNF-α and IL-12(p70) by DC, (ii) maturation of DC upon stimulation with B. vulgatus is impaired and (iii) B. ;vulgatus-treated semi-mature DC are less responsive to subsequent activation with E. coli. Supernatants from B. vulgatus-treated DC also induce non-responsiveness of DC to E. coli, and this inhibitory effect seems to be mediated by IL-6. Therefore, we suggest that B. vulgatus-induced IL-6, in the absence of TNF-α and IL-12(p70), may act as a modulatory cytokine that triggers differentiation of DC into a semi-mature, E. coli-non-responsive state.
Differential T helper cell polarization by DC stimulated with colitogenic and non-colitogenic bacteria
To investigate whether DC differentially matured by B. vulgatus and E. coli differentially prime T cells and T helper cell polarization, DC were stimulated with viable bacteria at an MOI of 1 and then added to naive ovalbumin (OVA)-specific T cells from OT-II mice and OVA323–339 peptide. T cells co-incubated with E. coli-stimulated DC secreted significantly (∼90%) more IFN-γ than T cells co-cultured with B. vulgatus-stimulated DC. Furthermore, DC exposed to B. vulgatus followed by E. coli stimulated 50% lower IFN-γ secretion from T cells compared to DC stimulated with E. coli only (Fig. 5). These data indicate that B. vulgatus inhibits E. coli-induced DC activation and maturation and thus T cell priming and Th1 polarization.
Next, the supernatants of DC stimulated with B. vulgatus (MOI 1) or E. coli (MOI 1) were collected and added to naive T cells from DO11.10 mice, irradiated APC and OVA323–339 peptide. As evident from Fig. 6, supernatants from DC stimulated with E. coli promoted IFN-γ production by OVA-specific CD4+ T cells, whereas B. vulgatus-conditioned supernatants did not. Furthermore, we found a total suppression of IL-4 secretion by T cells incubated with E. coli-conditioned supernatant, whereas the secretion of IL-4 was only slightly diminished by incubating T cells with B. vulgatus-conditioned supernatant compared to T cells stimulated only with medium (Fig. 6). However, the supernatant of B. vulgatus-stimulated DC did not promote IL-4 secretion by T cells. Supernatants of E. coli-stimulated DC induced high IL-10 production by T cells, while supernatants from B. vulgatus-stimulated DC induced only low IL-10 production (Fig. 6). Together these results indicate that E. coli but not B. vulgatus activates DC to induce IFN-γ production by naive T cells and thus drives naive T cells into the Th1 direction. Finally, supernatants from DC stimulated with B. vulgatus (MOI 1) followed by E. coli (MOI 1) induced less IFN-γ secretion than supernatants from E. coli-stimulated DC (∼40% reduction). This leads to the conclusion that supernatant from B. vulgatus-stimulated DC is able to inhibit the IFN-γ secretion by CD4+ T cells induced by supernatants from E. coli-stimulated DC (Fig. 7).
B. vulgatus mpk-colonized mice show increased colonic IL-6 mRNA expression
As shown previously 21, E. coli-colonized gnotobiotic IL-2–/– mice show increased levels of TNF-α, IFN-γ and CD14 mRNA as compared to B. vulgatus mono-colonized or E. coli and B. vulgatus co-colonized mice. In order to analyze whether E. coli and B. vulgatus promote different IL-6 production in gnotobiotic IL-2–/– mice in vivo, we investigated the IL-6 mRNA levels in colonic tissue of gnotobiotic IL-2–/– and IL-2+/+ mice at 20 and 33 wks of age prior to and after onset of colitis. The IL-6 mRNA levels were normalized to the housekeeping gene GAPDH. We observed increased expression of IL-6 mRNA in the colonic tissue of B. vulgatus-colonized mice compared to E. coli-colonized mice at 20 wks of age (Fig. 8a) and increased expression of IL-6 mRNA in the colon of 33-wk-old (Fig. 8b) E. coli and B. vulgatus co-colonized mice.
The main findings of this study are: (i) B. vulgatus induces IL-6 but not TNF-α and IL-12(p70) secretion by DC, whereas E. coli induces secretion of IL-6, TNF-α as well as IL-12(p70) by DC; (ii) B. vulgatus inhibits maturation of DC, which might be explained by a paracrine effect of IL-6; (iii) E. coli-stimulated but not B. vulgatus-stimulated DC polarized naive T cells into the Th1 direction, an effect that was both IL-12- and LPS-dependent, as the addition of anti-IL-12 antibodies or polymyxin B reduced IFN-γ secretion by T cells (data not shown); (iv) B. vulgatus-treated semi-mature DC are non-responsive to stimulation with E. coli in terms of maturation, T cell priming and TNF-α and IL-12 (but not IL-6) production. In conclusion, these data suggest that B. vulgatus-induced IL-6 accounts, at least partially, for a decreased capacity for T cell priming and decreased Th1 polarization.
B. vulgatus is an anaerobic Gram-negative rod closely related phylogenetically to Porphyromonas gingivalis. LPS of P. gingivalis has an attenuated potential to stimulate TNF-α and IL-12(p70) secretion compared to E. coli LPS 24. Moreover, P. gingivalis LPS binds to LPS-binding protein (LBP) with 100-fold less affinity than does E. coli LPS 25. It is known that E. coli and P. gingivalis induce different classes of immune responses in vivo26. E. coli LPS induces a Th1-like response with abundant IFN-γ but little or no IL-4, IL-13 and IL-5, whereas P. gingivalis LPS induces T cell responses characterized by high levels of IL-4, IL-13 and IL-5. E. coli LPS consistently induces IL-12(p70) in CD8+ DC subsets, while P. gingivalis LPS does not 26. This data is in line with our finding that B. vulgatus induces no IL-12(p70) secretion by DC. Thus, we assume that the attenuated activation of DC upon exposure to B. vulgatus and the impairment of TNF-α and IL-12(p70) secretion as compared to secretion from E. coli-challenged DC is due to a lower activation of the cells. Additionally, our data indicate that B. vulgatus does not promote Th1 responses. On the other hand, a recent study reported that Gram-negative bacteria, including B. vulgatus and E. coli, may have an immunomodulatory effect on human monocyte-derived DC by IL-12-, IL-27- and IL-23-dependent imprinting of a strong Th1-polarizing capacity 27. Whether this discrepancy is due to differences in bacterial strains or differences between murine bone marrow-derived DC and human monocyte-derived DC remains to be shown.
In animal models, different strains of the same bacterial species may act variably in terms of induction of inflammatory bowel disease. In HLA-B27 transgenic rats, B. vulgatus but not E. coli leads to colitis 28. In contrast, in IL-2–/– mice E. coli mpk but not B. vulagtus mpk induces colitis associated with up-regulation of colonic TNF-α, IFN-γ and mCD14 mRNA expression 21. Moreover, in IL-2–/– mice B. vulgatus prevents E. coli-induced colitis 21. Corresponding to this phenomenon, we found that in vitro, stimulation of DC with B. vulgatus leads to a non-responsiveness of DC to E. coli, indicated by reduced TNF-α secretion and reduced Th1 polarization, including reduced IFN-γ secretion. This can be explained by the fact that stimulation of DC with B. vulgatus drives DC into a semi-mature phenotype characterized by reduced responsiveness towards a second microbial stimulus with E. coli and a reduced capacity for T cell activation. As stimulation of DC with B. vulgatus had no effect on TLR2 mRNA expression and rather led to up-regulation of TLR4 mRNA expression, we can exclude that down-regulation of TLR accounts for the reduced responsiveness of B. vulgatus-conditioned DC. In contrast, stimulation of immature DC with E. coli induces highly activated and mature DC, a proinflammatory DC response and promotion of T cell activation as well as polarization toward a Th1 response. However, simultaneous addition of B. vulgatus and E. coli had no inhibitory effect on the E. coli-triggered TNF-α production (data not shown), implicating that an active regulatory process is involved in the B. vulgatus-induced non-responsiveness of DC. However, as treatment of DC with B. vulgatus (MOI 0.01) induced no significant IL-6 secretion but led nevertheless to reduced E. coli responsiveness, we cannot exclude that mechanisms other than IL-6 might contribute to tolerance induction in B. vulgatus-treated DC.
The differentiation of DC into a semi-mature phenotype, reflected by a lack of TNF-α and IL-12 production, inhibition of maturation and LPS non-responsiveness, was IL-6-dependent. Consequently, neutralization of IL-6 with anti-IL-6 mAb abolished the inhibitory effect of B. vulgatus-conditioned supernatant. Interestingly, these DC were selectively non-responsive to E. coli-induced production of the proinflammatory cytokines TNF-α and IL-12 but not E. coli-induced production of IL-6. Therefore, it might well be that via a paracrine loop, IL-6 secreted by semi-mature DC induces differentiation of other immature DC into a semi-mature state.
IL-6 is known to influence cell growth, differentiation and migration during immune responses, haematopoiesis and inflammation 29. IL-6 also affects the differentiation of myeloid lineages, including macrophages and DC 30, 31. In the presence of IL-6, the number of resting/immature DC in lymph nodes and spleen is increased, whereas the number of activated mature DC is decreased upon stimulation with LPS in vivo 17. IL-6 plays an important role in T cell differentiation through two independent molecular mechanisms. First, stimulation of T cells by IL-6 leads to up-regulation of nuclear factor of activated T cells (NFAT), a transcription factor regulating IL-4 transcription 31, resulting in IL-4 expression and promotion of Th2-polarized T cell differentiation 32. Second, IL-6 up-regulates the expression of silencer of cytokine signalling (SOCS)1 in CD4+ cells, which inhibits IFN-γ signalling and thus Th1 differentiation 33. The presence of IL-6 may shift the Th1/Th2 balance towards Th2 32. In keeping with these findings, we hypothesize that in our model B. vulgatus-triggered IL-6 induces semi-mature DC, which inhibit the induction of a proinflammatory Th1 responses. It is therefore tempting to speculate that IL-6, in the absence of TNF-α and IL-12(p70), can be considered a modulatory rather than simply a proinflammatory cytokine.
On the other hand, IL-6 secreted by splenic DC upon TLR stimulation is an important factor in T cell activation, enabling Treg-mediated suppression of T cell proliferation to be overcome 34. Thus, IL-6–/– mice were severely compromised in their ability to induce T cell proliferation upon immunization with OVA and LPS (as an adjuvant) 34. Production of IL-6 by DC in response to TLR ligation during stimulation appears to be critical for T cell activation, because it allows pathogen-specific T cells to overcome the suppressive effect of CD4+CD25+ T cells 34. However, as splenic DC include different subsets of DC, we cannot exclude the possibility that our findings are limited to a system working with only myeloid DC.
IL-6 has been reported to act as a proinflammatory cytokine in several models of acute inflammatory bowel disease. A neutralizing antibody to IL-6R prevents colitis in a model of TNBS colitis and in IL-10–/– mice via induction of apoptosis in lamina propria T cells 35. In a T cell transfer colitis model, administration of anti-IL-6R mAb prevented colitis 36. Furthermore, IL-6–/– mice are less susceptible to DSS colitis compared to wild-type mice. However, in IL-2–/– mice, IL-6 mRNA expression is increased if B. vulgatus is applied for colonization and co-colonization compared with E. coli mono-colonized mice. Conversely, expression of TNF-α, IFN-γ and CD14 mRNA is significantly increased in E. coli-colonized mice as compared to B. vulgatus-colonized and E. coli and B. vulgatus co-colonized mice 21. Therefore, we speculate that B. vulgatus stimulates IL-6 expression in colonic tissue but not expression of proinflammatory cytokines such as TNF-α and IFN-γ in vivo. IL-6, in the absence of TNF-α or IL-12, may promote semi-mature DC that are less responsive to E. coli stimulation. Pretreatment of DC with IL-6 down-regulates CCR7 expression upon stimulation with LPS, which leads to a reduced migratory capacity 23. Furthermore, these DC are less responsive towards stimulation with LPS 23. We speculate that colonization with B. vulgatus might, via IL-6 production, induce semi-mature DC with reduced expression of CCR7 and reduced migration, which could cause accumulation of IL-6-secreting DC in the intestine, reduced migration to mesenteric lymph nodes, reduced T cell activation and thus prevention of colitis. In contrast, in E. coli-colonized mice, activated and matured DC would migrate to the mesenteric lymph nodes, activate T cells and trigger induction of colitis. This scenario might contribute to the prevention of E. coli-induced colitis by B. vulgatus.
Taken together, these data provide evidence that IL-6 might play an ambiguous role in inflammatory diseases. Depending on the host status (acute inflammatory bowel disease with presence of proinflammatory cytokines such as TNF-α or disease remission with absence of proinflammatory cytokines), IL-6 may act as a proinflammatory or an anti-inflammatory cytokine, promoting or inhibiting T cell priming and Th1 responses. Future studies will have to elucidate the mechanisms of B. vulgatus-mediated prevention of colitis in vivo and investigate the molecular mechanisms that enable bacteria to trigger semi-mature DC, the signalling pathways involved in this process and the differential non-responsiveness of bacteria-triggered TNF-α and IL-12 production compared with bacteria-triggered IL-6 production.
Materials and methods
Mice and bacteria
C57BL/6 × 129Sv mice, OT-II mice and DO11.10 mice were obtained from our own breeding. BALB/c mice were obtained from Harlan Winkelmann. All mice were kept under specific pathogen-free conditions. Male and female mice were killed at 6–10 wks of age. Animal experiments were reviewed and approved by an appropriate institutional review committee. The bacteria used for stimulation of the murine DC were E. coli mpk 21 and B. vulgatus mpk 21. The E. coli strain was grown in Luria-Bertani (LB) media under aerobic conditions at 37°C. B. vulgatus was grown in Brain-Heart-Infusion (BHI) medium under anaerobic conditions at 37°C.
Bone marrow cell culture and DC
Bone marrow cells were isolated and cultured as described by Lutz et al. 37 with minor modifications. Cells were harvested at day 8 and used to evaluate the effects of challenge with E. coli and B. vulgatus on cytokine release and expression of surface markers as described below.
Induction of cytokine release and surface molecule expression by DC
Cells were stimulated at day 8 with viable bacteria at MOI of 0.01, 0.1 and 1 at 37°C, 5% CO2. Gentamicin was added 1 h after stimulation, and cells were incubated for 20 h; to exclude bacterial overgrowth, the CFU of viable bacteria was determined after 20 h. LPS (1 μg/mL) was used as positive control, PBS was used as a negative control and gentamicin was added to exclude effects of gentamicin on cytokine induction. In case of costimulation experiments, DC were treated with B. vulgatus (MOI 0.1–100) at day 7, and at day 8 E. coli (MOI 1) was added. For prestimulation with IL-6 (Sigma), IL-6 was added 24 h prior to E. coli at different concentrations (2.5 ng/mL–10 ng/mL). Cell culture supernatant was collected, filter sterilized and stored at –80°C until used for ELISA, DC prestimulation and T cell stimulation. Inhibition of E. coli growth by addition of 10% B. vulgatus-treated DC supernatant was not observed. Cells were subjected to surface marker staining for flow cytometry analysis as described below. For invasion assays, cells were stimulated and washed after 1 h, followed by addition of fresh medium containing 1% gentamicin. After another hour incubation, cells were harvested for CFU determination of invasive E. coli or B. vulgatus.
Stimulation of T cells with supernatants from stimulated DC
DO11.10 mice transgenic for an OVA323–339-specific TCR 38 obtained from our own breeding were used as a source of antigen-specific T cells. Splenocytes from BALB/c mice were irradiated and used as APC. Naive CD4+ T cells were prepared from DO11.10 mice using a magnetic cell sorting kit (Miltenyi Biotec) according to the manufacturer's instructions. Naive T cells (2.5 × 105) from spleens of DO11.10 mice were cultured in the presence of 5 × 106 splenic APC from BALB/c mice, 0.6 μM OVA323–339 peptide and 10% cell culture supernatant from stimulated DC. DC supernatants were maintained for the entire period of stimulation with or without addition of IL-6 and IL-10 neutralizing mAb. The cultures were split, and cell numbers were equalized in each well at day 3 and/or day 5. IL-2 (20 U/mL) was added for the last 4 days. Cell cultures were stimulated with PMA (50 ng/mL) and ionomycin (1 μg/mL; Sigma) and cultured for an additional 24 h. Supernatants were collected for assessment of IFN-γ 39, IL-4 and IL-10 (BD Pharmingen).
Stimulation of T cells with stimulated DC
CD4+ T cells were isolated from spleens of OT-II mice as described above. Bone marrow-derived DC were stimulated with viable bacteria (MOI 1). Gentamicin was added after 1 h to kill extracellular bacteria, and the cells were cultured for an additional 24 h with OVA (100 μg/mL). Stimulated DC (2 × 105) and naive T cells (8 × 105) were co-cultured for 4 days. T cell polarization was determined by analyzing the amount of IFN-γ in cell culture supernatants.
Cytokine quantification in cell culture supernatants
For analysis of IL-4, IL-6, IL-10, IL-12(p70) and TNF-α concentrations, commercially available ELISA kits (BD PharMingen) were used according to the manufacturer's instructions. Determination of IFN-γ concentrations was performed as described previously 39.
Immunostaining and flow cytometry
Cells (4 × 105) were incubated in 50 μL PBS containing fluorochrome-conjugated antibodies at a concentration of 10 μg/mL. A total of 1 × 104 cells were analyzed. The following antibodies were used for staining: PE-conjugated anti-mouse CD11c, clone HL3 (Armenian Hamster IgG1, λ); FITC-conjugated anti-mouse CD80, clone 16–10A1 (Armenian Hamster IgG2, κ); FITC-conjugated anti-mouse CD86, clone GL1 (Rat IgG2a, κ); FITC-conjugated anti-mouse I-A/I-E, clone 2G9 (Rat IgG2a, κ); and FITC-conjugated anti-mouse CD40, clone 3/23 (Armenian hamster IgM κ).
Determination of mRNA expression in intestinal tissue by semiquantitative real-time RT-PCR
The colon including the caecum was cut into pieces, transferred into 3 mL cold TRIZOL reagent (Invitrogen Life Technologies, Karlsruhe, Germany), homogenized and finally snap-frozen in liquid nitrogen. RNA isolation was performed according to the manufacturer's instructions. Extracted RNA was dissolved in water containing 0.1% diethyl-pyrocarbonate. For reverse transcription 4 μg RNA was mixed with 0.5 μg oligo(dT)12–18 primers (Invitrogen Life Technologies), and diethyl-pyrocarbonate-treated water was added to a final volume of 10 μL, followed by incubation at 65°C for 10 min. After adding 10 μL of a solution containing 5 × first strand buffer, 20 nmol/L dithiothreitol, 200 U Superscript II (Invitrogen Life Technologies), 40 U RNase Out (Invitrogen Life Technologies) and 2 mmol/L desoxynucleoside triphosphate (Roth), the mixture was incubated at 37°C for 60 min. Finally, the samples were heated at 90°C for 5 min, diluted with diethyl-pyrocarbonate-treated water and stored at –20°C until further use. Real-time RT-PCR was carried out in duplicates in 96-well format on a GeneAmp 5700 Sequence Detection System (Applied Biosystems/Applera, Darmstadt, Germany). Each 20 μL reaction contained 10 μL TaqMan Universal PCR MasterMix (No AmpErase UNG, Applied Biosystems), 1 μL target gene-specific Assay-on-Demand Gene Expression Assay Mix (Applied Biosystems), 4 μL PCR grade water and 5 μL cDNA. Thermal cycling conditions for all reactions were as follows: 2 min at 50°C, 10 min at 95°C and then 40 cycles of 15 s at 95°C and 1 min at 60°C. The standard curve method was used for semiquantitative data analysis, whereas 10-fold dilutions of pooled cDNA from all samples were used as standards. Data were normalized by dividing the values of the target gene by the values of the housekeeping gene GAPDH.
Statistical analysis was performed using the paired Student's t-test, with p>0.05 considered significant. Error bars represent ± SEM.
This work was supported by IZKF Tübingen (BMBF) and the DFG.