Intestinal inflammation in inflammatory bowel disease (IBD) and experimental models of colitis is characterized by a dysregulated intestinal immune response with elevated levels of Th1 cytokines. The luminal flora has been implicated as a major factor contributing to the initiation and perpetuation of inflammation in experimental colitis by mechanisms not known. Bacterial DNA contains unmethylated cytosin–guanosin dinucleotides (CpG) which strongly activate Th1-mediated immune responses. To test whether these CpG-motifs modulate intestinal inflammation we treated mice with dextran sulphate sodium (DSS)-induced colitis with CpG-containing oligodeoxynucleotides (CpG-ODN). CpG-ODN given after the onset of DSS colitis aggravated the disease, as indicated by a significantly increased loss of body weight and a 30% increase of the histological score. Further, we found a severe increase of proinflammatory cytokines (interleukin (IL)-6: 40-fold; interferon (IFN)-γ : 11-fold). In a pretreatment setting CpG-ODN reduced weight loss significantly and reduced intestinal inflammation by 45%. Colonic IFN-γ and IL-6 mRNA levels were reduced by 75%, and IL-10 was elevated by 400% compared to controls. The prophylactic CpG-effect was not imitated by IL-12 because IL-12 pretreatment was not protective. In time-course experiments, CpG-ODN pretreatment over 5 days resulted in a tolerance effect concerning its IFN-γ-inducing quality, and during the following days of colitis induction IL-10 secretion from mesenterial lymph node cells was elevated compared to controls. Therefore, the prophylactic effect of CpG-ODN might be explained by its tolerizing effect and/or the increased ability for IL-10 production during the consecutive intestinal inflammation.
The pathogenesis of chronic inflammatory bowel disease (IBD) is still unknown. Its aetiology is complex and multi-factorial, including genetic and environmental factors [1,2]. There is abundant evidence that the resident intestinal flora plays a critical role in the initiation and perpetuation of chronic intestinal inflammation, as demonstrated in numerous genetic mouse and rat models of spontaneous colitis who fail to develop disease under germ-free conditions [3–9]. On the other hand, there are several observations suggesting that an ‘over-clean’ environment, especially in early childhood, increases the risk of IBD [10,11]. Recently, mutations in the non-obese diabetic (NOD)2 gene were found to be associated with Crohn's disease . The NOD2 gene product is thought to operate as intracellular receptor for microbial products (e.g. lipopolysaccharides) resulting in NFκB activation in monocytes . Hence, there is cumulative convincing evidence for a multi-faceted role of bacteria and their products in the development of chronic intestinal inflammation.
Recently, the importance of bacterial DNA as activating product of the vertebrate immune system was recognized . As shown, cytosin–guanosin dinucleotide (CpG) sequence motifs composed of unmethylated CpG dinucleotides are the immunostimulatory component of bacterial DNA . Oligodeoxynucleotide (ODN)-containing CpG motifs are able to activate murine macrophages, lymphocytes, natural killer (NK) cells and dendritic cells. In response to this stimulus the cells process and present antigens and start to secrete a variety of proinflammatory cytokines, including interleukin (IL)-6, IL12, tumour necrosis factor (TNF) and interferon (IFN)-γ, leading eventually to a strong induction of a Th1-skewed immune response . CpG motifs occur stochastically in bacterial DNA (5%) and are rare in vertebrate DNA (1·5%) . Additionally, the cytosines in vertebrate DNA are mainly methylated in contrast to bacterial DNA. Based on these differences the vertebrate immune system has developed mechanisms to specifically recognize bacterial DNA. Recent findings indicate that the toll-like receptor (TLR) 9 is critical for the recognition of CpG motifs of bacterial DNA .
In vitro-generated unmethylated CpG-ODN are able to mimic the immune activating properties of bacterial DNA and can be used successfully as adjuvants in immunization studies [19,20] as well as in models of allergic diseases . Overall, the specific immune stimulatory properties of bacterial DNA lead to a strong Th1 response with elevated production of IFN-γ, IL-12, TNF and IL-6 that is similar to the immune reaction described for IBD. Therefore, and with regard to the possible contribution of bacteria in the pathomechanism of colitis, we hypothesize that the interaction between bacterial DNA and the intestinal immune system is a critical factor within the pathogenesis of IBD.
Previously, we were able to demonstrate a proinflammatory effect of CpG-ODN treatment during dextran sulphate sodium (DSS)-induced colitis, most evident in the chronic phase of the disease . In this study, we investigate the differential effects of ‘prophylactic’versus‘therapeutic’ CpG-ODN application on the outcome of colitis and try to explain the underlying pathomechanism.
MATERIALS AND METHODS
Female Balb/c mice weighing 20–22 g (Charles River, Germany) were used for the experiments and housed during the experiments in a conventional facility. The animal studies were approved by the local Institutional Review Board.
Reagents and cell culture conditions
DSS was purchased from ICN (MW 40,000, Eschwege, Germany) and Salmonella typhimurium-derived lipopolysaccharide (LPS) was purchased from Sigma (Deisenhofen, Germany). Murine IL-12 was a gift from M. Gately (Hoffmann-La Roche, New Jersey, USA), Phosphothioate-stabilized ODNs were obtained from Metabion (Martinsried, Germany). The sequence of the ODN is as follows:
For induction of colitis mice received 5% DSS in drinking water for 7 days, as described previously . Animals were treated with CpG- and control GpG-ODN from either days 3–7 during DSS application or over a time period of 5 days before induction of colitis with DSS. ODN were given i.p. daily for 5 days in doses of 10 µg and 2 µg in 100 µl sterile PBS per injection, as indicated below. IL-12 and LPS were administered i.p. in a dose of 50 ng and 1 µg in 100 µl sterile phosphate buffered saline (PBS), respectively, according to the above described schedule. It has been shown previously that in vivo LPS treatment using doses between 0·6 and 2 µg per mouse induces LPS hyporesponsiveness [25–27].
Assessment of histological score
From the distal third of the colon 1 cm of colonic tissue was removed and used for histological analysis as described previously [24,28]. Three sections were evaluated, each obtained at 100 µm distance. Mice were scored individually, each score representing the mean of three sections. Histological examination was performed by two independent investigators blinded to the source of treatment. Histology was scored as follows:
Epithelium (E). 0: normal morphology; 1: loss of globlet cells, 2: loss of globlet cells in large areas; 3: loss of crypts; 4: loss of crypts in large areas.
Infiltration (I). 0: no infiltrate; 1: infiltrate around crypt basis; 2: infiltrate reaching to lamina muscularis mucosae; 3: extensive infiltration reaching the L. muscularis mucosae and thickening of the mucosa with abundant oedema; 4: infiltration of the L. submucosa.
The total histological score represents the sum of the epithelium and infiltration score and ranges from 0 to 8.
Isolation and incubation of mesenterial lymph node cells
Mesenterial lymph nodes (pooled from each group of mice) were collected under sterile conditions in cold cell culture medium (RPMI-1640, 10% fetal calf serum (FCS), 100 U/ml penicillin and 100 µg/ml streptomycin from Gibco-BRL (Eggenstein, Germany) and β-mercaptoethanol 3 × 10−5m (Sigma, Deisenhofen, Germany). Lymph nodes were disrupted mechanically and filtered through a cell strainer (70 µm). 2 × 105 cells/well were incubated in 200 µl culture medium over 24 h and spontaneous cytokine secretion was measured in the supernatants by ELISA (all from Endogene, Woburne, MA, USA), using four wells per condition.
An additional colonic tissue specimen (1 cm) was harvested for quantitative RT-PCR and transferred to ice-cold RNAlater solution (Ambion, Austin, TX, USA). RNA was extracted using the RNeasy-kit (Quiagen, Hilden, Germany) and the Quiagen Shredder-kit following the manufacturer's recommendations. RNA was transcribed using the Promega (Mannheim, Germany) reverse transcription system as recommended by the manufacturer. Quantification of cytokine mRNA was performed using a light cycler (Roche, Molecular Systems, Mannheim, Germany). For standardization β-actin was amplified (Primer pair: 5′-TGG AAT CCT GTG GCA TCC ATG AAA C-3′ and 5′-TAA AAC GCA GCT CAG TAA CAG TCC G-3′).
Primers and conditions
IFN-γ. 5′-TGG AGG AAC TGG CAA AAG GAT GGT-3′ and 5′-TTG GGA CAA TCT CTT CCC CAC-3′, annealing temperature at 62°C, 3 mm MgCl2.
IL-1β . 5′-TGA AGG GCT GCT TCC AAA CCT TTG ACC-3′ and 5′-TGT CCA TTG AGG TGG AGA GCT TTC AGC-3′, annealing temperature at 60°C, 3 mm MgCl2.
IL-6. 5′-TGG AGT CAC AGA AGG AGT GGC TAA G-3′ and 5′-TCT GAC CAC AGT GAG GAA TGT CCA C-3′, annealing temperature at 62°C, 4 mm MgCl2.
TNF. 5′-GCG ACG TGG AAC TGG CAG AAG-3′ and 5′-GGT ACA ACC CAT CGG CTG GCA-3′, annealing temperature at 62°C, 3 mm MgCl2.
IL-10. 5′-TCC TTA ATG CAG GAC TTT AAG GGT TAC TTG-3′ and 5′-GAC ACC TTG GTC TTG GAG CTT ATT AAA ATC-3′, annealing temperature at 62°C, 3 mm MgCl2.
All primers were purchased from MWG-Biotech AG, Ebersberg, Germany. Cytokine mRNA was quantified for each individual mouse.
Statistical analysis was performed using Student's t-test (cytokine levels), the Mann–Whitney rank sum test (histological score) or the general linear model (daily weight loss). Error bars represent the standard error of the mean. P < 0·05 was considered to be statistically significant.
The time-point of CpG-ODN treatment before or after induction of colitis influences disease severity
To investigate possible divergent immune modulating effects of bacterial DNA in vivo in a murine model of colitis different time-points of ODN application were chosen to influence the outcome of intestinal inflammation. Initially, mice were treated with ODN1668 (CpG-ODN) or control ODN1668G (GpG-ODN) after the onset of colitis over a time-period of 5 days from days 3–7 of DSS feeding. As shown in Fig. 1, additional application of CpG-ODN during induction of colitis significantly enhanced loss of body weight (P < 0·0001) (Fig. 1a). Intestinal inflammation was more severe in animals treated with CpG-ODN than in mice receiving control GpG-ODN, as demonstrated by an increased histological score (CpG-ODN: 7·0 ± 0·3 versus 5·4 ± 0·5 control) (Fig. 1b). Notably, CpG-ODN treatment of healthy control animals did not influence body weight and did not result in intestinal inflammation (Fig. 1).
Subsequently, we tested the effect of CpG-ODN priming prior to the induction of colitis. Therefore, CpG-ODN or GpG-ODN (control) was administered daily for 5 days before DSS application. Surprisingly, animals pretreated with CpG-ODN (10 µg) were significantly less affected compared to GpG-ODN pretreated controls, illustrated by a significantly reduced weight loss (P < 0·0001) (Fig. 2a). The protective effect of CpG-ODN pretreatment was confirmed further by a marked reduction of the histological score (CpG-ODN 10 µg: 3·6 ± 0·22 and CpG-ODN 2 µg: 4·3 ± 0·56 versus control 6·6 ± 0·24) reflecting a less severe inflammatory infiltrate and reduced epithelial damage (Fig. 2b,c,d).
Pretreatment with the proinflammatory cytokine IL-12 or the bacterial component LPS does not simulate the protective effect of CpG-ODN motifs
It is known that in vitro stimulation of antigen presenting cells with CpG-ODN induces the secretion of IL-12 . To test whether CpG-induced production of this proinflammatory cytokine was involved in the protective effect seen in our experiment, we examined directly the influence of IL-12 pretreatment on the course of colitis. As shown in Fig. 2a, IL-12 pretreatment revealed a trend towards aggravation of the weight loss (one animal of six died on day 5) and the inflammatory damage within the colonic mucosa was increased slightly compared to controls (Fig. 2b). Furthermore, application of LPS, another bacterial component that is also known to induce activation of antigen-presenting cells in vitro and in vivo, before induction of colitis with DSS did not influence weight loss (data not shown) and the histological score was not significantly different (Fig. 2b).
Effects of CpG-ODN treatment on the colonic cytokine expression pattern before and during colitis induction
Intestinal inflammation is accompanied usually by expression of proinflammatory cytokines within the intestinal mucosa, reflecting the activation of APC and the induction of a Th1-dominated immune response. To investigate whether certain cytokine patterns reflect the differences between treatment strategies with CpG-ODN before and after the onset of inflammation we evaluated the effects of CpG-ODN administration on cytokine expression in the colonic tissue using real-time PCR. As anticipated, an almost 40-fold increase of colonic IL-6 and an 11-fold increase of IFN-γ mRNA transcripts compared to controls were detected when CpG-ODN treatment was started after the onset of colitis, whereas levels of TNF, IL-1β and IL-10 mRNA were not altered (Fig. 3a). In contrast, pretreatment with CpG-ODN resulted in a 85% decrease of IFN-γ and IL-6 mRNA levels and a 55% reduction of TNF mRNA expression in the colonic tissue at the end of DSS application, whereas mRNA expression levels of the anti-inflammatory cytokine IL-10 were found to be elevated fourfold compared to controls (Fig. 3b). IL-1β mRNA expression was not affected.
Divergent effects of CpG-ODN treatment before and during induction of colitis on cytokine secretion from mesenterial lymph node cells
To test whether the differences seen in cytokine expression on the mRNA level within the colonic tissue were accompanied by differential effects within draining lymph nodes, we measured cytokine production of mesenterial lymph node cells. As shown, isolated mesenterial lymph node cells from animals treated with CpG-ODN after onset of colitis secreted significantly higher amounts of IFN-γ (15-fold) compared to the GpG-ODN treated controls. IL-10 secretion was low overall and was not influenced significantly by CpG-ODN application (Fig. 4a).
In contrast, pretreatment with CpG-ODN led to a significant decrease of IFN-γ secretion to 22% and an increase in IL-10 secretion (2·5-fold), compared to GpG-ODN-treated controls (Fig. 4b).
Effects on the time-course of cytokine secretion from mesenteric lymph node cells during CpG-ODN treatment and colitis induction
It is known that during the course of colitis induction expression of certain cytokines changes over time, reflecting the development of a T cell response. To illuminate further the beneficial effects of CpG-ODN pretreatment we monitored the time-dependent alteration of mesenteric lymph node cell cytokine secretion from the moment of CpG-ODN injection until the end of DSS application. CpG-but not GpG-ODN pretreatment (starting at day −4) initially augmented IFN-γ production in normal mice (day − 3), whereas after 5 days of pretreatment, at the time of DSS introduction, no difference was noted between CpG-ODN-treated and GpG-ODN-treated control animals (day 0; Fig. 5b, left panel). During colitis, IFN-γ secretion in the CpG-ODN pretreated group remained stable over time. In contrast, the GpG-ODN-treated control group showed a marked increase in IFN-γ production of mesenterial lymph node cells during progression of colitis (Fig. 5b, left panel, dotted line). Additionally, we investigated the time-course of cytokine induction within the group of animals receiving CpG-ODN after onset of colitis. A strong increase in IFN-γ secretion (days 3–7) was observed immediately after the first injection of CpG-ODN compared to the GpG-ODN-treated control group (Fig. 5a, left panel). Almost no IL-10 production was seen in both groups when treated during colitis induction (Fig. 5a, right panel). In contrast, CpG-ODN pretreatment induced high levels of the anti-inflammatory cytokine IL-10 in healthy animals 2 days after the start of CpG-ODN injection, which dropped almost back to baseline later. After induction of colitis with DSS IL-10, production again increased significantly compared to GpG-ODN controls (Fig. 5b, right panel).
This report provides evidence of divergent effects of CpG motifs in an experimental model of intestinal inflammation depending on the time-point of application. A characteristic feature of CpG-ODN is the strong Th1-inducing capacity [30,31]. Systemic administration of CpG-ODN after the onset of acute colitis resulted in an aggravation of disease accompanied by increased IFN-γ, IL-6 and IL-12 (data not shown) production. All these cytokines have been identified to be relevant in numerous experimental models of colitis. In previous studies we have shown that IL-12 induced IFN-γ-aggravated acute DSS-induced colitis . Thus, the colitis-accelerating effects of CpG-ODN administration during acute colitis can be explained by their Th1-favouring, proinflammatory properties. This is also in line with our previous results, which showed proinflammatory effects of CpG-ODN on established chronic DSS-induced colitis . Taken together, these data indicate that DNA might be an important bacterial constituent contributing to the well-established pacemaker function of the intestinal flora in development and perpetuation of chronic colitis .
In contrast, treatment with CpG-ODN over 5 days before colitis induction was protective, as indicated by a reduced weight loss and a reduced microscopic inflammatory infiltrate and epithelial damage. While this corroborates a recent report by Rachmilewitz et al.  showing the protective effects of CpG-ODNs when administered before or at the beginning of acute DSS colitis induction, we extended this finding by characterizing alteration of cytokine profiles on mRNA and protein level. We could demonstrate that colonic IL-6 and TNF levels as well as IFN-γ and IL-12 secretion from mesenterial lymph node cells were reduced in acute DSS colitis after CpG-ODN pretreatment. This indicated a reduced proinflammatory cytokine expression at least in the late phase of colitis induction.
Based on our data we hypothesized first that Th1-priming by CpG-ODN enables the intestinal immune system to cope better with the bacterial challenge to which it is exposed after the DSS-induced breakdown of the intestinal barrier. However, when mice were pretreated with IL-12, no protective effect was found. Because in immunohistochemistry there was no obvious difference in CD3+ cells in colonic sections (data not shown), the decreased IFN-γ secretion cannot be explained by a decreased number of infiltrating T cells at the site of injury. The cytokine production in time-course experiments during CpG-ODN pretreatment indicated ‘desensitizing’ as a possible mechanism. We found that the high initial IFN-γ and IL-12 (data not shown) production ceased after 5 days of CpG-ODN administration and further led to a suppression of IFN-γ production, normally observed in DSS-treated mice as shown in Fig. 5. Therefore, we assumed desensitization to be a probable mechanism for the protective effects of CpG-ODN. Tolerance in the TLR family has been shown previously and was found to be independent of the degree of receptor expression and paracrine mechanisms [34,35]. Interestingly, Yeo et al. demonstrated a differential modulation of LPS-induced cytokine secretion by CpG-ODN preincubation in mouse macrophages . When RAW264·7 cells were preincubated with CpG-ODN, a tolerance specific for CpG-ODN- and LPS-induced IL-12- and TNF production was found, whereas LPS-induced IL-10 production increased significantly about 3–4-fold. This immune modulating process induced by CpG-ODN prestimulation might become important when the intestinal immune system is again confronted with CpG motifs of bacterial DNA or LPS derived from luminal bacteria during DSS feeding and the concomitant breakdown of the epithelial barrier. Thus, overstimulation and excessive IFN-γ production is avoided, IL-10 secretion increases and the severity of disease is reduced. The distinct role of bacterial DNA in the development of colitis is underlined further by our finding that LPS pre-exposure, known to induce tolerance towards LPS [25–27,34], did not lead to a significant attenuation of colitis.
Tolerance induction towards bacterial DNA and a modified immune response to bacterial constituents might be a possible mechanism explaining the positive effects of prophylactic CpG-ODN administration in intestinal inflammation. However, in experimental models of autoimmune diseases, which are assumed to be independent of bacterial components, CpG-ODN also proved to be efficient in preventing disease. Quintana et al. demonstrated a strong protective effect of CpG-ODN in NOD mice by modulating spontaneous autoimmunity . The effect was explained by a strong increase of IL-10 secretion shown in spleen cells of these mice. This observation corresponds with our data, which show an increase of IL-10 secretion during acute colitis in CpG-ODN compared to control GpG-ODN pretreated mice. The increase of IL-10 secretion, demonstrated from day 5 onwards, might therefore contribute partly to a down-regulation of the inflammatory process. As a consequence of the altered IFN-γ /IL-10 ratio the differentiation of T-cells towards a regulatory phenotype might be supported. These cells − which then exert their regulatory role only in part via IL-10 secretion − were shown to have strong protective and therapeutic effects in experimental models of colitis [38,39]. Interestingly, a recent publication indicated that regulatory T-cells express TLRs and ligation directly enhances their regulatory capacity . Further studies using the SCID-transfer model are currently being undertaken in our laboratory to clarify further the mechanism underlying the prophylactic effect of CpG-ODNs.
Rachmilewitz et al. proposed that the protective effect of CpG-ODN in experimental models of colitis was not mediated by CD4+ T-cells because protective effects were observed in both models of Th1- and Th2-driven inflammation . However, an increase of IL-10 production and an eventually more regulatory type of immune response were both able to reduce Th1- and Th2-mediated pathologies.
In summary, our results indicate that exposure to CpG-ODN during acute inflammation exacerbates the disease, whereas pre-exposition proved to be protective. Due to the proinflammatory effects of CpG-ODN after the onset of colitis, as shown by us, a possible therapy with CpG-ODN during an acute flare-up, as suggested by others [33,41], might have deleterious consequences. The colitis-exacerbating potential of CpG-ODN was overlooked by these authors because CpG-ODNs were applied not later than 2 days after the start of DSS feeding . However, using these compounds for the maintenance of remission might be promising, eventually preventing flares via the induction of a more regulatory type of immune response and the modulation of the immune response towards bacterial constituents. The fact that pre-exposition to bacterial DNA was protective is of special interest in regard to results from epidemiological studies which suggest that the incidence of IBD correlates positively with the increase of hygienic standards in developed countries [10,11]. Therefore, it is tempting to speculate that the reduced exposure to microbial components (such as bacterial DNA) contributes to an increased risk to develop a deregulated intestinal immune response − leading eventually to inflammatory bowel diseases in genetically susceptible hosts.
This work was supported by a grant from the European community QLG1-CT-1999-00549 to W. F. and a research grant from the University of Regensburg, Germany as part of the ReForM programme to F. O.