The vagus nerve is an important pathway signaling immune activation of the gastrointestinal tract to the brain. Probiotics are live organisms that may engage signaling pathways of the brain-gut axis to modulate inflammation. The protective effects of Lactobacillus reuteri (LR) and Bifidobacterium infantis (BI) during intestinal inflammation were studied after subdiaphragmatic vagotomy in acute dextran sulfate sodium (DSS) colitis in BALB/c mice and chronic colitis induced by transfer of CD4+ CD62L+ T lymphocytes from BALB/c into SCID mice. LR and BI (1 × 109) were given daily. Clinical score, myeloperoxidase (MPO) levels, and in vivo and in vitro secreted inflammatory cytokine levels were found to be more severe in mice that were vagotomized compared with sham-operated animals. LR in the acute DSS model was effective in decreasing the MPO and cytokine levels in the tissue in sham and vagotomized mice. BI had a strong downregulatory effect on secreted in vitro cytokine levels and had a greater anti-inflammatory effect in vagotomized- compared with sham-operated mice. Both LR and BI retained anti-inflammatory effects in vagotomized mice. In SCID mice, vagotomy did not enhance inflammation, but BI was more effective in vagotomized mice than shams. Taken together, the intact vagus has a protective role in acute DSS-induced colitis in mice but not in the chronic T cell transfer model of colitis. Furthermore, LR and BI do not seem to engage their protective effects via this cholinergic anti-inflammatory pathway, but the results interestingly show that, in the T cell, transfer model vagotomy had a biological effect, since it increased the effectiveness of the BI in downregulation of colonic inflammation.
the term inflammatory bowel disease (IBD) comprises primarily ulcerative colitis and Crohn's disease (CD). IBD is a chronic, uncontrolled inflammation of the intestinal mucosa that can affect any part of the gastrointestinal tract (26). The mechanisms responsible for IBD remain poorly understood, but it is widely accepted that normal intestinal bacterial flora contribute significantly to the pathogenesis. This is demonstrated by the lack of intestinal inflammation in all transgenic models if the mice are raised under germ-free conditions, while rodents of the same genetic background, raised in a conventional setting, develop severe colitis (28).
Extensive research is focused on modifying the intestinal flora with probiotic bacteria to attenuate inflammatory activity (2, 29, 36). Probiotics are live organisms that, when ingested, promote beneficial health effects and well being to the host. Probiotic organisms have been used to treat a variety of human intestinal conditions, including diarrhea (22, 32), and aspects of inflammatory bowel disease (IBD), such as pouchitis (11). Organisms used as probiotics are most frequently of the Lactobacilli or Bifidobacterium species. Probiotics and commensal bacteria represent promising sources for novel therapeutic strategies in IBD. Although there is clear evidence for the efficacy of probiotics in the treatment of experimental colitis, the underlying mechanisms through which probiotics exert their effects are unknown. A variety of probiotic preparations have been tested in gut flora-associated animal models of colitis (3, 21, 30).
The described cholinergic anti-inflammatory pathway is a physiological mechanism that modulates host inflammatory responses via cholinergic mediators or electrical stimulation of the vagus nerve (33, 34). Our group among others has demonstrated the protective effect of vagal signaling in experimental colitis (8–10, 35). Because the vagus nerve plays a prominent role in the pathways of communication between the gastrointestinal tract and the brain, it is of interest to investigate if the anti-inflammatory effects of different probiotics are (partly) mediated through the vagus. However, no murine model of colitis has been employed so far to study the anti-inflammatory effects of probiotic organisms after subdiaphragmatic vagotomy.
Distinct mechanisms are involved in the acute Dextran sulfate sodium (DSS) and the chronic T cell transfer model of colitis. It has been shown that neither T nor B cells are required for the induction of acute DSS-induced colitis, since this model still induces inflammation in SCID mice (4). Therefore, the effect of vagotomy and the impact of feeding probiotic organisms was evaluated in both the acute DSS and the chronic T cell transfer model of colitis. The beneficial downregulatory anti-inflammatory effects of Lactobacillus reuteri (LR) and Bifidobacterium infantis (BI) were equally effective in vagotomized and nonvagotomized animals. We show that different probiotics ameliorated intestinal inflammation in mice but that their working mechanisms of action do not involve the vagus nerve in the studied models.
MATERIALS AND METHODS
Male CB-17/IcrHanHsd SCID mice and CB6F1/OlaHsd donor mice (BALB/c background), 6–10 wk old were obtained from Harlan. Before the study, all mice were housed under specific pathogen-free conditions. Male BALB/c mice were obtained from Harlan and were 6–7 wk of age upon delivery. Following initiation of the trials, all mice consumed a standard nonsterile diet and were housed in a conventional environment. The animals had free access to chow food and tap water. All animal experiments were approved by the local institutional review board. The different experiments described in this paper were conducted at several places, with the acute colitis LR study being done at McMaster University and the acute colitis BI study at the Brain-Body Institute, both in Hamilton, Canada. The SCID study was performed at the Alimentary Pharmabiotic Centre in Cork, Ireland.
From frozen stocks (−80°C), LR (ATCC no. 23272) cells were suspended in fresh Man-Rogosa-Sharpe (MRS) broth and plated in MRS agar, cultured anaerobically at 37°C for 24 h. Bacteria were resuspended in MRS broth and grown in 50-ml tubes for another 48 h at 37°C under anaerobic conditions. After 2 days, the bacteria were checked by Gram stain. Tubes were centrifuged at 2,000 revolutions/min (rpm) for 15 min at room temp, resuspended at the designated concentration of 5 × 109 bacteria/ml in MRS broth using a Vitek colorimeter, and frozen until use. BI (UCC no. 35624) was obtained from Alimentary Health (Cork, Ireland) and was originally isolated from the terminal ileum of a patient undergoing urinary tract reconstructive surgery, and its selection was based on its possession of a number of desirable probiotic properties. The BI strain was propagated at 37°C in MRS supplemented with 0.05% (wt/vol) cysteine hydrochloride (Sigma) under anaerobic conditions. Anaerobic growth conditions were achieved using Anaerocult A (Merck) sachets and a BBL GasPak system. Cells that were used for freeze-drying were harvested from early stationary phase cultures by centrifugation for 15 min at 10,000 rpm in a Beckman JA-14 rotor. The pelleted cells were resuspended in 20 ml of a reconstituted skimmed milk (RSM) cryoprotectant [18% RSM (Kerry ingredients) and 2% sucrose (Irish Sugar)] and removed immediately to a −80°C freezer. Frozen cells were transferred to a STERIS GT2 freeze drier and dried for 24–48 h. Upon completion of freeze-drying, the material was aseptically transferred to a sterile plastic bag and crushed using a pestle before being stored at −80°C.
Mice were anesthetized using gaseous anesthesia with isoflurane using a rodent anesthetics machine. The stomach and lower esophagus were visualized after an upper midline laparotomy. The skin and abdominal wall were incised along the ventral midline, and the intestine was retracted to allow access to the left lateral lobe of the liver (LLL) and the stomach. The LLL was retracted, and a ligature was placed around the esophagus at its entrance to the stomach to allow gentle retraction to clearly expose both vagal trunks. These were dissected, and all neural and connective tissue surrounding the esophagus below the diaphragm was removed to transect all small vagal branches. At least a 2-wk recovery period was allowed.
Assessment of vagotomy.
To assess the completeness of the vagotomy, a food intake analysis was performed based on the satiety effect of cholecystokinin-octapeptide (CCK-8) (Sigma Aldrich, St. Louis, MO), which is known to be mediated by the afferent vagus nerve (7, 14). All animals were food deprived for 20 h and then received an injection (ip) of 8 μg CCK·kg−1·mouse−1, and their 2-h food intake was measured. Food intake was decreased significantly by 46% in CCK-8-injected sham mice compared with animals that received saline (data not shown). Subdiaphragmatic vagotomy abolished the satiety effect of CCK-8. Any vagotomized animals that still decreased their food intake significantly were excluded from the study.
Induction of acute colitis.
Acute colitis was induced by exposing mice to DSS (mol wt 36,000–50,000) (MP Biomedicals) in their drinking water to a final concentration of 4% (wt/vol). Mice were allowed free access to this solution for 5 days and received tap water for two more days before death on day 8. Control mice received tap water throughout the experiment.
Induction of chronic colitis.
Chronic colitis was induced by transfer of CD4+ CD62L+ T lymphocytes from BALB/c mice into SCID mice. Splenic CD4+ CD62L+ T lymphocytes from BALB/c mice were isolated. Splenocytes from donor mice were used as a source of CD4+ T cells for reconstitution of SCID recipient mice. A single cell suspension was prepared in Hanks' balanced salt solution (Sigma-Aldrich) with 10% FCS (Sigma-Aldrich) from the donor spleen and filtered through a sterile 70-μm nylon cell strainer (BD Falcon). Erythrocytes were lysed using the Mouse Erythrocyte Lysing kit (R&D Systems). The CD4+CD62L+ T cell population was isolated using column separation (Miltenyi Biotech), first depleting non-CD4+ T cells and second positively selecting for the CD4+CD62L+ T cell subpopulations. These procedures resulted in a population ∼96% pure as assessed on the flow cytometer (BD Biosciences). Isolated donor CD4+CD62L+ T cells were resuspended in 200 μl of 1:1 PBS-DMEM (GIBCO) at a concentration of 1 × 106 cells and injected interperitoneally in each SCID mouse.
Administration of LR and BI.
Studying the effects of probiotics in the acute model, mice were given: 1) 1 × 109 LR in 200 μl/mouse a day for a total of 9 days (daily gavaging started 2 days before DSS treatment) or 2) BI were given in the drinking water starting 2 wk before DSS treatment based on previous experience. The concentration was adjusted so that, on average, mice consumed 1 × 109 colony-forming units (cfu)/ml daily. During DSS exposure, mice were given 1 × 109 cfu of BI by daily gavage. In the chronic model, the recipient sham and vagotomized SCID mice consumed 1 × 109 cfu/ml of BI in drinking water daily from the beginning of T cell reconstitution over a period of 10 wk. LR was not tested in the chronic SCID model.
Clinical assessment of colitis.
Throughout the experiments, animals were clinically assessed for signs of intestinal inflammation, including measurement of body weight, inspection of stools for diarrhea and blood, and appearance. The disease activity index, modified from Okayasu et al. (24), was determined (Table 1).
Determination of myeloperoxidase activity.
Myeloperoxidase (MPO) activity was measured as previously described (17). Briefly, full-thickness segments of the middle and distal colon were sonicated for 20 s in 50 mM KH2PO4 buffer (pH 6.0) containing 0.5% hexadecyltrimethylammonium bromide (ratio 1:10 wt/vol). Homogenates were then centrifuged at 10,000 g (10 min at 4°C). The supernatant was shock-frozen using liquid nitrogen, thawed three times, and centrifuged again for 10 min at 10,000 g at 4°C. Supernatants were assayed spectrophotometrically for MPO activity and protein measurements. MPO activity was expressed as units per milligram of protein, where 1 unit corresponds to the activity required to degrade 1 mmol of hydrogen peroxide in 1 min at room temperature. The protein concentration of the supernatant was determined using a Bradford assay kit (Bio-Rad Laboratories, Hercules, CA), and results were expressed as units of MPO per milligram of tissue protein.
Six cytokines [interleukin (IL)-6, IL-10, IL-12, tumor necrosis factor (TNF), monocyte chemoattractant protein (MCP)-1, and interferon (IFN)-γ] were analyzed. When studying the effects of LR on colitis after vagotomy, a FACScan was used (BD Biosciences, San Diego, CA). Because of the availability of new technology, the samples of the BI experiment were measured using a six-bead array flow cytometry system (BD Biosciences). The SCID experiment was conducted in the laboratory in Ireland where a LSRII (BD Biosciences) was used. The six-bead array system is less sensitive and results in lower readings. Because samples were measured in different machines in different experiments, while data were comparable, different ranges resulted. However, this does not affect the differences between groups.
Full-thickness segments of the middle and distal colon were pooled and homogenized in PBS containing protease inhibitors (1 tablet/25 ml PBS). Protease inhibitor cocktail tablets were purchased from Roche and used according to the manufacturer's instructions. Homogenates were centrifuged at 10,000 rpm (10 min at 4°C), and the supernatant was collected and stored at −20°C until measured. Results were expressed as cytokine levels per milligram of total protein.
Spleens were individually collected in cold cell culture medium (Hanks' salt) or DMEM with 10% FCS, mechanically disrupted, and filtered through a cell strainer (70-μm pore size). Red blood cells were lysed, and isolated lymphocytes were counted and resuspended. A total of 1 × 106 cells/ml were incubated in anti-CD3-coated plates (10 μg/ml of anti-CD3 at 4°C overnight) in 200 μl of culture medium for 48 h. The spontaneous production of cytokines and the production of cytokines following stimulation with anti-CD28 monoclonal antibodies (2 mg/ml) were measured. Four wells per condition were used, and results are given in picograms per milliliter (mean ± SE).
Results on cytokine levels were analyzed by a one-way ANOVA followed by a post hoc test (Bonferroni). Probability values of P < 0.05 were considered to reflect a statistically significant difference. Graphs were constructed with Prism software (Prism; GraphPad Software, San Diego, CA).
In the acute model, the body weight of vagotomized mice did not differ significantly from that of sham vagotomized animals (data not shown). DSS exposure significantly affected the weight of all groups of mice treated with DSS and lost weight compared with their weight on day 1. Control mice gained weight compared with their starting point. LR and BI did not have a significant effect on weight in either sham or vagotomized control or DSS-exposed animals. In the T cell transfer model of colitis, there was no significant effect of vagotomy on weight. BI did affect weights in the SCID mice, since treated mice lost less weight compared with placebo treatment in sham and vagotomized groups (data not shown).
Disease activity score.
Important clinical parameters, including consistency and blood in stool, fur color, and texture (Table 1), were changed significantly after the induction of acute and chronic colitis. In the acute DSS model, vagotomy further increased severity of the clinical score (Table 2). The onset of sickness parameters was earlier in vagotomized animals compared with sham. All animals exposed to DSS receiving LR or BI clinically improved, including vagotomized mice. In the chronic model, SCID mice started showing signs of sickness ∼8 wk after CD4+CD62L injections. Vagotomized mice had an increased clinical score throughout the experiment and at the final day. At the end of the 10-wk treatment period, mice receiving BI appeared healthier than did untreated mice. This was reflected by a significantly lower clinical score (Table 2).
In DSS-treated groups, the MPO activity was significantly higher in vagotomized compared with sham animals. In sham animals and vagotomized mice, treatment with BI and LR significantly decreased the MPO activity in both sham and vagotomized animals (Fig. 1A). In SCID mice, MPO levels were also increased after vagotomy, but this increase was not significant. In this model, BI did decrease levels of MPO in vagotomized animals, whereas the levels in sham-operated mice were not affected by BI (Fig. 1B).
In DSS-treated groups, the level of TNF was significantly higher in vagotomized compared with sham animals (Fig. 2). LR and BI decreased the level of tissue TNF in sham and vagotomized mice (Fig. 2). The level of IL-6 partly shows conflicting results. As explained in materials and methods, protective effects of LR and BI were studied at different time points and in different animal facilities, which could explain why IL-6 levels were similar between sham and vagotomized mice (Fig. 3A) while the level of IL-6 was significantly greater after vagotomy (Fig. 3B). However, LR did greatly and consistently decrease the levels of IL-6 in sham and vagotomized mice; BI only decreased tissue IL-6 in vagotomized animals. MCP-1 levels were not affected by BI, whereas LR very strongly reduced MCP-1 levels both in sham and vagotomized animals (sham: control 25 ± 6; DSS/control 549 ± 71; DSS/LR 53 ± 11; vagotomy: control 39 ± 5; DSS/control 521 ± 129; DSS/LR 32 ± 5). In the chronic T cell transfer model of colitis, the tissue cytokine levels were very low, and no differences between groups were found (data not shown).
Anti-CD3/CD28-stimulated splenocytes secreted more TNF and IL-6 (Fig. 4, A and B) after vagotomy. In vitro BI appeared to be more potent compared with LR in downregulating the secretion of cytokines. In addition, BI seems to work better after vagotomy. Lymphocytes from spleen from untreated CD4+ CD62L+ T cell-injected SCID mice secreted high levels of proinflammatory cytokines TNF and IL-6. There was no difference in secretion between sham and vagotomized groups. However, administration of BI led to a significant reduction of TNF, IL-6, and IFN-γ (data not shown) in vagotomized animals, whereas the level of secretion in sham-operated mice remained unchanged (Figs. 4C and 5C).
Administration of certain commensal organisms (e.g., Lactobacillus and Bifidobacterium species) to murine models attenuates the development of colitis (3, 21, 23, 25). However, the mechanisms whereby this is achieved are poorly understood. We tested the efficacy of LR and BI in a model of acute and chronic experimental intestinal inflammation after subdiaphragmatic vagotomy, since this has been shown to remove the tonic cholinergic anti-inflammatory efferent effect of the vagus (33). We confirm the anti-inflammatory involvement of the vagus nerve in acute intestinal inflammation in mice (35). In contrast to our expectations, however, the working mechanism of both probiotics LR and BI do not involve the vagus nerve in either of the studied models, and their anti-inflammatory effectiveness was retained in both.
The vagus nerve is known to be a major communicator between the gastrointestinal tract and the brain. It has been demonstrated that all regions of the colon, except the rectum, are innervated by the vagus nerve (1). Work done in our laboratory suggests that the anti-nociceptive effects of LR may involve nerves in the intestine (15). We hypothesized that probiotics engaged the vagus nerve for their protective effects in acute and chronic colitis. Our results show that LR strongly attenuated inflammation in both sham and vagotomized animals and BI had a stronger effect after vagotomy in the acute and chronic T cell transfer model of colitis.
From previous studies, we know that vagotomy increases inflammation in the chronic DSS model (35), but no proinflammatory effect was seen in this model in SCID mice lacking T and B lymphocytes. The overall level of inflammation in the SCID is higher compared with the acute and chronic DSS models. Various explanations could account for this difference. The injection of a population of foreign T cells may have led to a reaction that is so drastic that cutting the vagus did not lead to additional inflammation. The vagus carries efferent effector anti-inflammatory signals via a cholinergic α7-nicotinic pathway (6), and this may have an unknown effect on the T cell system. Because regulatory T cells are not present in SCID mice, cutting the vagus may result in a similar inflammatory state in sham and vagotomized animals. Publications from Ghia et al. (8–10) have shown, albeit in C57Bl/6 mice, that vagal protection weakens in time and is replaced with other anti-inflammatory mechanisms. However, in our experiments in Balb/c mice, we did not see any increase in inflammation in the chronic T cell transfer model of colitis in the first place, but the vagotomy had a biological effect, since it increased the effectiveness of the BI in downregulation of colonic inflammation. These unexpected findings are the subject of intense ongoing investigation.
Multiple mechanisms have been proposed to account for probiotic action in different clinical conditions. The probiotic cocktail VSL#3 has been demonstrated to induce dendritic cell secretion of IL-10 while attenuating T cell production of IFN-γ (12). Probiotic effects on epithelial cell function have been demonstrated in vitro and in vivo (13, 19). LR has been shown to be capable of inhibiting the onset of experimental colitis in transgenic IL-10-deficient mice (21) and has further been shown to reduce intestinal permeability to macromolecules (20). Probiotics are capable of inhibiting TNF production in activated macrophages (16, 27). A recent publication by Lin et al. (18) showed that LR suppressed the production of TNF from lipopolysaccharide-activated human monocytes/macrophages. Transcriptional factor studies demonstrated that the suppression was achieved by inhibiting activation of c-jun and the activation of mitogen-activated protein kinase-regulated transcription factor AP-1. Because only TNF production was suppressed in activated cells, macrophages may respond to probiotics based on their relative state of activation. Because overproduction of TNF is implicated in pathogenesis of intestinal inflammation and even further elevated by subdiaphragmatic vagotomy, enteric LR- or BI-mediated inhibition of TNF and alteration of cytokine profiles may highlight an important immunomodulatory role in the present study.
In our laboratory, LR has been shown to be effective in an ovalbumin model for murine allergic asthma (5). Furthermore, LR possesses anti-nociceptive properties since it was shown to be effective in a visceral pain model in rats (15). In the present study, LR was able to strongly decrease the DSS-induced colitis. Surprisingly, downregulation of inflammation by LR did not seem to require the presence of an intact vagus. The use of BI in SCID mice following transfer of CD4+ CD62L+ T cells ameliorated the development of intestinal inflammation. Furthermore, BI seems to work more effectively after vagotomy. Because this is especially the case in SCID mice, this effect cannot be due solely to an increased inflammatory state, since vagotomy did not increase overall inflammation in SCID mice.
In the context of IBD, anti-inflammatory bacteria may signal to the gastrointestinal epithelium and perhaps mucosal regulatory T cells or dendritic cells (31). Although it is known that T cells are not required to induce acute DSS colitis (4), the findings that BI works better after vagotomy could imply a role for T cells. Possibly the population of CD4+ T cells after vagotomy induced inflammation, whereas CD4+ T cells from sham-operated mice could be less pathogenic. Unpublished pilot observations have shown that inflammation increased in mice after receiving CD4+ cells from vagotomized otherwise healthy animals, supporting the possibility that there are changes in T cell populations after vagotomy. However, considerably more experiments are needed to further study these possibilities.
Perspectives and Significance
The vagus nerve has a tonic efferent cholinergic anti-inflammatory effect via a specific nicotinic α7 pathway that appears to be active in some models of colitis and other forms of subdiaphragmatic inflammation. Because probiotic organisms have systemic and local effects, it was a reasonable assumption that the vagus nerve might be involved in their working mechanism. The two different strains that we tested in the present study did not support this assumption. LR and BI did not require the presence of an intact vagus for their protective effects. Thus the mechanisms of action of at least two probiotic organisms do not involve the vagus nerve. However, this does not necessarily mean that the vagus could not be involved in the working mechanisms of other probiotic strains. Interestingly, vagotomy produced different results in the acute and chronic models of colitis that were used, having little proinflammatory effects in the T cell transfer model and dramatic effects in the DSS model. These results clearly point to effects of an unknown nature of the vagus on T cells and offer additional avenues of investigation on the effects of both probiotic organisms and the vagus nerve in inflammation.
This work was supported in part by the Broad Medical Research Program. C. O'Mahony, F. Shanahan, and L. O'Mahony were supported in part by Science Foundation Ireland by the Health Research Board of Ireland and the Higher Education Authority of Ireland.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2008 the American Physiological Society