Functional gastrointestinal disorders, represented mainly by irritable bowel syndrome (IBS), are among the most prevalent gastrointestinal alterations in the western population. Alterations in bowel habits, abdominal pain, and discomfort, believed to reflect increased visceral sensitivity, are hallmarks of IBS. Symptoms in IBS fluctuate over time in intensity and character, but the mechanisms underlying these cycles remain unclear. Several factors, including stress, intestinal infection, drugs, and diet have been reported to exacerbate symptomatology, and might be key components of the pathophysiology of the disease.[2, 3] A growing body of evidence suggests that IBS pathogenesis is likely dependent on the interaction between local immune reactions within the intestinal wall and environmental factors in genetically susceptible individuals. In particular, stress and perturbations of the gut commensal microbiota have been recognized as two potential factors contributing to the onset, maintenance, and exacerbation of both functional and inflammatory gastrointestinal disorders.[4, 5] Indeed, stressful life events or depression are risk factors for the onset or relapse of intestinal inflammation and for symptoms presentation in IBS patients. Similarly, growing evidences suggest that IBS patients have a dysbiotic intestinal microbiota.[4, 6] Despite these evidences, the exact role of gut microbiota and stress, individually or as interactive factors, in the pathophysiology of IBS remains largely unknown.
In this study, we characterized the interaction between stress and microbiota and their potential role modulating functional colonic responses to stress and the induction of inflammatory-like changes in mice. First, we assessed the effects of repetitive psychological stress (water avoidance stress, WAS) and antibiotic treatment, individually or in combination, on the composition of ceco-colonic commensal microbiota and the induction of inflammatory-like changes in the colon. In the same animals, endocrine and colonic motor responses to stress were assessed simultaneously. To characterize the ceco-colonic microbiota, we determined changes in both luminal and wall (epithelium)-adhered microbiota. The assessment of inflammatory responses was based on inflammatory markers, histological evaluation of the colon, and quantification of luminal secretory-IgA (s-IgA). s-IgA is considered the main anti-inflammatory immunoglobulin of the mucosal intestinal immune system regulating the number, composition, and functions of luminal bacteria.[7, 8] Moreover, we also determined changes in relevant systems that have been involved in sensory responses within the colon, with particular relevance to IBS, namely the endocannabinoid and the serotonergic systems. For this, colonic expression of cannabinoid receptors type 1 and 2 (CB1 and CB2) and activity of the serotonergic system [density of enterochromaffin cells (EC) and expression of the tryptophan hydroxylase isoform 1 and 2 (TPH1 and TPH2)] were characterized in the same animals. Finally, to determine if these alterations translate into functional changes in visceral sensitivity, we tested visceral pain-related responses in animals treated with antibiotics, with or without the addition of stress. For this, we assessed the presence of visceral pain-related behaviors associated with the intracolonic administration of capsaicin, as previously described.[9, 10]
- Top of page
- Materials and Methods
- Author Contribution
In this study, we show that the colonic functional (motor) and endocrine responses to stress are essentially not affected by relatively large alterations of the ceco-colonic microbiota, either luminal or attached to the colonic wall, during an antibiotic treatment. Moreover, we show that microbiological changes, due to antibiotics and stress, are able to modulate the immune and sensory systems, namely the endocannabinoid and the serotonergic systems, within the colon, without the induction of a manifest state of intestinal inflammation. While antibiotics, per se, did not affect visceral pain-related responses, they prevented stress-induced hypersensitivity. This suggests that antibiotics-mediated effects on sensory systems might have functional consequences, leading to the modulation of visceral sensitivity.
Our results confirm the validity of chronic WAS as a valid, mild stressor in mice, as previously published.[14, 15, 26] Mice did not habituate to the stress protocol, as shown by the persistent colonic response along the 7-day period of WAS. Moreover, the efficacy of the stress paradigm is further demonstrated by the raise in plasma corticosterone and the increase in weight of the adrenal glands at the end of the last stress session.
Total bacterial counts were not affected by stress. However, repetitive WAS significantly increased the counts of Clostridium spp. and favored the appearance of Lactobacillus spp. These changes agree with those described in mice subjected to social stress, where the main change in the microbiota was an increase in the Clostridia group. Interestingly, the Verrucobacteria group, present in a relatively high proportion in non-stressed mice, was undetectable in stressed animals. This group of microorganisms, which degrade mucus within the gastrointestinal tract,[17, 27] might have relevance in gastrointestinal diseases. For instance, an enhancement of the mucin-degrading microbiota in dysbiotic patients predispose to Crohn's disease. During stress, the thickness of the mucus layer was reduced, in agreement with O'Malley et al.[28, 29] A reduction in mucus abundance might be a factor reducing also the relative abundance of Verrucobacteria. Alternatively, we cannot discard that these changes are secondary to the combined enhancing effects of stress on colonic motility and mucus secretion,[30-32] leading to an increased discharge of mucus and therefore to a net reduction in mucus content and associated bacteria. Moreover, although goblet cell density remained stable, stress increased the proportion of mature goblet cells, indicative of an increase in mucus production and secretion. Despite these changes in mucus content, wall-adhered microbiota was not affected by stress.
As expected, treatment with wide-spectrum, non-absorbable antibiotics significantly reduced total bacterial counts. The reduction in bacterial counts was coupled to a specific dysbiosis which implied a proliferation of Lactobacillus spp. and Enterobacteria; whereas the Clostridium spp. and the Verrucobacteria groups were reduced. Interestingly, only antibiotic-induced changes in luminal microbiota were associated with an increase in bacterial wall adherence. This is important because adhered microbiota has been suggested to be the one directly interacting with the host's bacterial recognition systems, thus eliciting either beneficial or harmful responses within the gut.[34, 35] The relationship between luminal counts and epithelial attachment seems to be strain dependent. Overall, changes in bacterial wall adherence correlated positively with changes in luminal counts. However, the Clostridia group was reduced during antibiotic treatment, but presented an increased rate of adherence. This negative relationship might reflect the heterogeneity of Clostridium coccoides cluster XIVa. From the present data, we cannot rule out the possibility that antibiotics are affecting only a part of this cluster, leading to a relative selection of bacteria with high wall adherence capacities. In fact, it is well reported that most antibiotics can increase the risk of developing Clostridium difficile colitis[36, 37] and that the relapse of colitis in patients with recurrent C. difficile infections is associated with reduced intestinal microbial diversity. Nevertheless, the role of gut commensal microbiota in intestinal inflammation remains controversial, and beneficial effects of wide spectrum antibiotics has been shown in DSS-induced colitis in rats. The mucous layer represents also a protective barrier preventing bacterial wall adherence. Therefore, a loss of mucus should be regarded as a factor favoring bacterial–host interactions.[40, 41] Antibiotics had only a marginal effect reducing the mucous layer, thus suggesting that the mucus, per se, might play a minor role affecting bacterial wall adherence in the present conditions. Ceco-colonic dysbiosis was further enhanced when antibiotic-treated mice were subjected to stress. This was associated with a significant increase in the incidence of wall adherence, observed for all bacterial groups assessed, and a clear reduction in the thickness of the mucous layer.
Commensal microbiota is necessary for the development of spontaneous colitis, as suggested by observations in mice deficient in interleukin 10; however, gut commensal microbiota could also have a protective role, as seen in germfree mice with DSS-induced colitis.[42-44] These apparent discrepancies might be associated with the composition of the microbiota, the immaturity of the immune system, the environmental conditions of housing, and the type of treatment applied (duration and antibiotics used). In any case, the potential pathophysiological implications of these observations warrant further investigations. In humans, increased bacterial wall adherence has been suggested as a pathogenic factor leading to local immune responses that favor the appearance and maintenance of intestinal inflammation.[21, 45] Interestingly, antibiotic-induced dysbiosis had no impact on the gut-to-brain modulation of endocrine responses to psychological stress. This agrees with recent data suggesting that the gut-to-brain signaling is established during the early post-natal phase and that commensal microbiota is important during that imprinting period.[4, 46] Once the gut is colonized and the commensal microbiota established, changes in microbiota composition seem to have a minor impact in gut-to-brain signaling, at least as stress-related endocrine responses relates. Despite this, intestinal microbiota has been related as a putative factor affecting gut sensory systems leading to altered behavioral[47, 48] and local visceral responses, such as visceral pain.[12, 49] For instance, gut commensal microbiota is fundamental for the development of inflammatory pain in mice.[12, 49, 50] Here, we assessed changes in the endocannabinoid and the serotonergic systems, two of the main sensory systems within the gut, with a demonstrated involvement in secretomotor- and visceral pain-related responses.[49, 51-54] In the present conditions, antibiotics selectively upregulated the expression of CB2; an effect further enhanced by the addition of stress. This agrees with data suggesting that gut microbiota is able to upregulate the endocannabinoid system within the gut. Modification in the commensal microbiota by addition of specific bacterial strains (namely L. acidophilus) has been shown to upregulate CB2 expression in rats and mice, leading to the induction of visceral analgesia. In agreement with this, changes in CB2 expression correlated positively with luminal counts of Lactobacillus spp., which increased with antibiotic treatment and were further enhanced in stressed antibiotic-treated mice. Overall, these observations further support the view that bacteria of the Lactobacillus spp. group should be regarded as a beneficial component of the microbiota, which might be implicated in the modulation of visceral pain responses through the modulation of the intestinal endocannabinoid system. On the other hand, counts of Clostridium spp. correlated in a negative manner with the CB2 expression reinforcing the potential role assigned to this bacterial group as a pathogenic component of the microbiota.
Expression of TPH1 and TPH2 and density of EC cells served to assess the activity of the serotonergic system. As expected, expression of TPH2, the isoform responsible for the synthesis of neuronal serotonin, was very low in whole colonic homogenates. On the other hand, expression TPH1, responsible for serotonin synthesis in EC cells, was detected at relatively high levels. Interestingly, TPH1 was upregulated in stressed animals, independently of the antibiotic treatment. These observations might suggest that, although not directly assessed, serotonin synthesis and availability is increased during stress, with commensal microbiota playing a minor role per se. Overall, this agrees with studies showing that serotonin availability might be increased within the colon during stress. However, density of EC cells was not affected by stress, thus suggesting a cellular hyperactivity, rather than a hyperplasia. This contrasts with inflammatory models of gut dysfunction, such as the experimental infection with Trichinella spiralis, in which increased availability of serotonin has been associated with a hyperplasia of EC cells.[57, 58] The functional consequences of these changes in the cannabinoid and serotonergic systems warrant further studies, outside the original scope of the present work.
The changes observed in the expression of sensory-related systems are likely to have a functional significance. This is demonstrated by the changes in visceral pain-related responses observed in antibiotic-treated vs non-treated animals. In agreement with previous reports, we show that intracolonic capsaicin evokes behavioral responses consistent with the induction of visceral pain.[9, 10] Moreover, an increase in pain-related events was observed in stressed animals, thus confirming data indicating that repeated psychological stress induces visceral hypersensitivity in rodents.[26, 59] Interestingly, stress-induced hyperalgesic responses were completely prevented by the treatment with antibiotics. However, in non-stressed animals, antibiotics had no significant effects on visceral pain-related behaviors. This might suggest that the modulatory effects exerted by antibiotics are able to compensate states of altered (increased) sensitivity, without affecting basal responses. Therefore, it is feasible to assume that the changes observed in CB2 expression and serotonin availability might lead to functional effects modulating states of altered visceral sensitivity. Similarly, other sensory mediators not directly assessed here and involved in visceral pain responses, such as vanilloids, might be involved in the responses observed. Overall, these observations further support an involvement of gut microbiota as a modulatory component of gut sensory functions.
As mentioned, none of the treatments applied resulted in evident intestinal inflammation. Although enlargement of the cecum was observed in antibiotic-treated animals, this was not associated with consistent histopathological alterations. It is interesting to point out that despite the increased host–bacterial interaction observed in dysbiotic mice, no signs of colonic inflammation (either macroscopical, microscopical or biochemical) were observed following the treatment with antibiotics. This contrasts with previous reports that observed signs of intestinal inflammation during both antibiotic treatment and stress.[12, 41, 61, 62] In particular, the appearance of stress-induced intestinal inflammation has been related with a mast cell infiltrate and the facilitation of bacterial wall adherence in rats.[12, 41, 61, 62] However, in our conditions, the density of mast cells was not increased by stress. Although inflammatory markers were unaltered, luminal s-IgA levels were increased during dysbiosis. Luminal s-IgA contributes to the suppression of immune reactions generated by commensal bacteria[63, 64] and, when binding to bacteria, prevent bacterial translocation. Increased s-IgA levels might represent a mucosal response, likely triggered by the increased rate of bacterial attachment during dysbiosis, aiming the prevention of local and systemic inflammation and bacterial translocation. Multiple factors ranging from the species/strain used to the intensity of the stressors applied or the microbial environment might contribute to the final immune response to a dysbiotic state. Systematic studies addressing these aspects will be necessary to determine the relative contribution of these factors to the final responses observed within the gut.
In summary, the current study shows that gut commensal microbiota and stress are likely to act as interactive components in the maintenance of gut homeostasis and in the development of gut pathophysiology. Changes observed here suggest that microbiota and stress are able to selectively modulate gut sensory mechanisms, in the absence of obvious structural or biochemical alterations compatible with the presence of intestinal inflammation. Nevertheless, a mucosal immune response, characterized by increased s-IgA production, could be observed. Moreover, the treatment with antibiotics was associated with a reduction in stress-induced visceral hypersensitivity, thus suggesting that microbiota, influencing sensory-related systems within the gut, is able to modulate visceral pain arising from the intestine. Overall, these data support the potential involvement of stress and gut microbiota in the alterations observed in patients with functional gastrointestinal disorders, characterized by secretomotor and sensory alterations in the absence of structural changes. These observations warrant further studies dissecting the pathways altered by stress and gut microbes and the associated functional changes. Our observations support the view that the beneficial effect of certain bacterial strains, used as probiotics, might be associated with the modulation of the activity of endogenous sensory-related systems, such as the endocannabinoid system.