Clostridium butyricum relieve the visceral hypersensitivity in mice induced by Citrobacter rodentium infection with chronic stress

View article
Biochemistry, Biophysics and Molecular Biology

Introduction

Irritable bowel syndrome (IBS) is a common functional gastrointestinal disorder which is defined by chronic abdominal pain or discomfort and altered bowel habits, without histopathologic findings (Longstreth et al., 2006). It has a obviously high morbidity (15–23%) across the industrial world, causing a reduced quality of life in patients and a considerable socioeconomic burden (Guglielmetti et al., 2011; Saha, 2014). Acute infection in the gastrointestinal tract has six-fold risk factor to the development of IBS (Halvorson, Schlett & Riddle, 2006). There is a sub-type named post-infectious IBS (PI-IBS) characterized by new occurrence and frequent abdominal pain that exhibits visceral hypersensitivity, with altered bowel habits (mostly diarrhea), following the experience of an acute gastroenteritis episode (Barbara et al., 2019). PI-IBS represents about 4% to 36% of all IBS cases (Spiller & Campbell, 2006). Patients typically complain of the abdominal pain symptoms which persist for years and the increasing of visceral hypersensitivity is commonly found among them (Marshall et al., 2010; Lee, Annamalai & Rao, 2017).

Although the pathogenesis of PI-IBS is indistinct and complicated, there is plenty of evidences showing that the intestinal barrier function dysbiosis (Cotton, Beatty & Buret, 2011), low grade inflammation of intestinal mucosa (Spiller et al., 2000; Barbara et al., 2004), neuroendocrine crosstalk (Ng et al., 2018), and the changes of the gastrointestinal microbiota (Lupp et al., 2007; Krogius-Kurikka, Lyra et al., 2009; Malinen et al., 2005) are related to it. Remarkably, gut microbial dysbiosis or inflammation plays a key role in the pathogenesis of abdominal hypersensitivity of PI-IBS (Su et al., 2018). There are increases in the numbers of Firmicutes and Proteobacteria, while decrease in numbers of Bacteroidetes (Krogius-Kurikka, Lyra et al., 2009; Malinen et al., 2005). However, there is a relative paucity of research assessing the mechanism between visceral hypersensitivity and the gut microbial remodeling of PI-IBS.

The guidelines of therapeutic strategy for PI-IBS are insufficient, there are still several clinical therapies for the gastroenteritis episode, which can be divided into general treatment (antidiarrheal and fluid supplement) and modifications of the gut microbiome (probiotics and antibiotics) (Chey, Eswaran & Kurlander, 2015). The former treatment strategy has little influence on abdominal pain and uncomfortable symptoms, and the treatment of various antibiotic drugs can induce dramatic and irretrievable changes in the gut microbiota (Rodino-Janeiro et al., 2018). On the other hand, probiotics have been reported to be the active component used in this field. With colonized into the intestinal tract and remodeling of the gut microbiota in childhood gastroenteritis occasions (Freedman et al., 2018), probiotics have gained a great attention in recent years. The C. butyricum is a type of Gram-positive anaerobic bacterium in gut microbiota, and it could regulate intestinal immune function and prevent colitis in mice (Hayashi et al., 2013). It was reported that C. butyricum can improves symptoms, quality of life and stool frequency in IBS-D patients (Sun et al., 2018). Other kind of probiotics such as Lactobacillus and Bifidobacteria preperations also have potential in reducing abdominal symptoms and regulating bowel habits of IBS-D (Wang et al., 2014). Although there is phenomenal evidence for the C. butyricum treatment on IBS, whether it could influence the visceral hypersensitivity and gut microbiota of PI-IBS is is still unknown.

It is well known that Citrobacter rodentium, which resembles the pathogenic factor pathogens enteropathogenic E. coli (EPEC) and enterohaemorrhagic E. coli (EHEC) in humans (Mullineaux-Sanders et al., 2019), can produce a transient and self-limiting colitis in mice (Ibeakanma et al., 2009). It is interesting that the infectious mice treated with the chronic water avoidance test will have induced visceral hypersensitivity.

In this study, we prepared the protocol that we could use mice subjected to the C. rodentium infection plus chronic water stress as the rodent model of PI-IBS. According to the above-mentioned research results, our idea is that intervention of C. butyricum may actually involve in or influence the etiology of visceral hypersensitivity in PI-IBS through microbiotia modification.

Materials & Methods

Mice model

Thirty C57BL/6 male mice (5–6 weeks old, 15–20 g) were purchased from SPF Biotechnology Co., Ltd. (Beijing, China), and twenty-four of them were randomly assigned to four groups: the C. rodentium gavaged group (Crodentium group), the C. rodentium + Cbutyricum intervention group (Cbutyricum group, or probiotic group), the C. rodentium + antibiotics group (antibiotic group) and the control group with PBS gavaged group (n = 6 per group). The C. rodentium infected groups were gavaged with C. rodentium (2 ×109 colony forming units (CFU), 200 µl of PBS) on day 0. Mice in the probiotic and antibiotic groups were gavaged with C. rodentium together on day 0, while C. butyricum or antibiotics were joined in 3 days later. Mice in the probiotic group were gavaged C. butyricum (1 ×109 CFU, 200 µl of PBS) daily for one week. Mice in the antibiotic group were given gentamycin (Sigma, St. Louis, MO, USA) at 20 mg/kg/d and ampicillin (Sigma) at 500 mg/kg/d in drinking water for one week (Fig. 1). The rest of the six mice were given Crodentium and deeply euthanizing with 5-minutes CO2 asphyxiation at 30% Chamber Replacement Rates at day 0 (n = 2), day 5 (n = 2) and day 21 (n = 2) after infection, and their distal colonal tissues were collected for HE Staining. Mice that survived the study were bred for our other experiments. All mice were maintained on a 12-h light/dark cycle, with room temperature of 22 °C–24 °C and relative humidity of 50–60%. Mice were allowed aseptic food and water in a level 2 animal feeding facility at Shandong University Laboratory Animal Center, and we have mixed the bedding for all mouse cages randomly at the beginning of the mice experiment. All procedures in this experiment were in accordance with the ARRIVE guidelines, and were approved by the Shandong University Laboratory Animal Center and the Ethics Committee on Animal Experiment of Shandong University Qilu Hospital (animal experiment proof certificate number: Dull-2020-11).

Timeline for C. rodentium infection and C. butyricum/antibiotics intervention for mice.

Figure 1: Timeline for C. rodentium infection and C. butyricum/antibiotics intervention for mice.

C. rodentium gavage began at 0 days to 5–6 weeks old mice except the control group. The C. rodentium + C. butyricum intervention group and the C. rodentium + antibiotics group were given the probiotics or antibiotics, respectively. And the intervention last for 1 week (orange stripes). Three C. rodentium infectious groups were followed by water avoidance stress (WAS) for a period of 9 days (blue stripes). Mice feces samples were taken and stored weekly (yellow triangles) until day 30.

Culture of C. rodentium and C. butyricum

A single C. rodentium (ATCC51459) colony was inoculated into Luria broth (LB) medium and grown overnight (at 37 °C and 150 rpm for 12 h). The C. butyricum was seperated from C. butyricum capsules (ATaiNing, Qingdao Eastsea Pharmaceutical Co., Ltd., China, 420 mg per capsule, 1.5 × 107 CFU/g, CGMCC0313.1) and cultured anaerobically at 37 °C for 14 h in De Man, Rogosa and Sharpe (MRS) medium.

Water-avoidance stress (WAS)

The C. rodentium infected mice group, the probiotic group, and the antibiotic group were all given a 9-days chronic WAS paradigm (days 21- 30 after infection). For 1 h each day of the chronic stress period, mice were placed on a dry platform (70 mm diameter) in an acrylic bucket of water (280 mm diameter), standing 20 mm above water level and 10 mm below the brim of the bucket. Mice were stressed within a single period of chronic stress (9 days of WAS concluding on post-gavage day 21) as shown in Fig. 1.

Behavioral testing

Behavioral responses to CRD were assessed in four groups starting 1 days after the water-avoidance stress by measuring the abdominal withdrawal reflex (AWR) using a semiquantitative score. The testing of AWR indicated an express contraction in the abdominal wall musculature. Distention water balloons (6-Fr, two mm external diameter) were placed in the descending colons of mildly sedated mice (4% isoflurane, temporary inhaled anesthesia) and secured by taping the mouse’s tail, and then they were put into a transparent acrylic box. The mice were then housed in small boxes as a recovery room and allowed to wake up and adapt (1 h). Measurement of the AWR consisted of visual observation of the animal response to graded CRD (0.1, 0.15, 0.2, 0.25 and 0.3 ml of water injected into balloons) by double-blinded observers and assignment of the AWR score. Each mouse was tested three times at each graded CRD and averaging for analysis.

HE staining

The HE staining was conducted in the colonic histology specimens of the six C. rodentium infected mice. The paraffin-embedded distal colonic tissues were cut into 4 µm longitudinal sections, after proper specimen processing which involves dehydration, clearing, and paraffin infiltration. After deparaffinization and rehydration, the sections were stained with hematoxylin solution for 2 min followed by 3–4 drips of 1% acid ethanol (1% HCl in 70% ethanol) for 2 s. After the rinse by water, they were stained with eosin solution for 2–3 min and followed by dehydration with graded alcohol and xylene. The slides were observed and photographed using a fluorescence microscope (Nikon TI-FLC-E; Nikon,Tokyo, Japan).

16S gene pyrosequencing analysis of mice fecal samples

At the point-in-time of orange triangles in Fig. 1, feces samples were collected at day 0 and in each week after C. rodentium infection and they were weighed and homogenized in sterile PBS for microbiota analysis. Fecal samples from the four group mice were collected and DNA was extracted, and the 16S rRNA gene was amplified and sequenced. The 16S rRNA pyrosequencing was processed at Majorbio (Shanghai, China) by using the Illumina Miseq system. The Chao index and Shannon index were calculated to assess the alpha-diversity in each sample. Cluster analysis based on the Euclidean distance was conducted based on the relative abundances of all operational taxnomic units (OTUs). The principal co-ordinates analysis (PCoA) based on the Bray-Curtis distance was done in order to performed to assess the beta-diversity.

Data analysis

Data are expressed as mean ± SD, and error bars in figures represent SD. The statistical analysis of AWR scores was performed using paired Student’s t-tests, and significance was assigned when *P < 0.05, **P < 0.01, ***P < 0.001, ##P < 0.01. The raw data of 16S gene pyrosequencing were clustered into OTUs based on a 97% similarity and analyzed through RDP classifier 2.11. The community composition at each taxonomic level was calculated by Bayes algorithm. The OTUs were then imported into Mothur v 1.38.1 for analyzing the rarefaction curves and alptha diversity. The specific taxa were identified by the Linear discriminant analysis Effect Size (LEfSe) analysis with the value of Kruskal -Wallis sum-rank test set to 0.05. The taxonomy of each 16S rRNA gene sequence was analyzed by RDP Classifier algorithm (http://rdp.cme.msu.edu/) against the Silva (SSU132) 16S rRNA database using confidence threshold of 70%. SPSS was used to analyze the data, a 95% confidence level, P < 0.05 was used.

Results

C. rodentium infection plus repeated water-avoid stress caused the visceral hypersensitivity in mice

All mice exposed to C. rodentium alone showed signs of colitis early in the infection (Day 1 to Day 4), such as soft stools, rough body hair, languid movement and temporary body weight loss (Day 2 to Day 5). Two-week post infection, these symptoms gradually faded away, and all mice finally recovered from infectious colitis and brought into next experiments. On the 30th days after C. rodentium infection and WAS treatment, the AWR score showed that the infectious group had a lower pain tolerance for colorectal distension compared with the control group (n = 6, P < 0.01) as shown in Fig. 2A. Also the colonic inflammation pathological changes of the infectious group had recovered to normal within the assessment of the histopathological score (Fig. 2B). Based on these phenomena, we can draw the conclusion that C. rodentium infection and AWR stress could produce a PI-IBS-like mice model.

The AWR score and H&E-stained distal colon sections of mice.

Figure 2: The AWR score and H&E-stained distal colon sections of mice.

(A) Summary data (n = 6 mice per group) illustrating the AWR scores which were the responses to graded CRD with water injection in mice subjected to the PBS-only control group (blue circles), C. rodentium-only (orange squares), C. butyricum co-treatment with C. rodentium (pink triangles), and gentamicin-ampicillin co-treatment with C. rodentium (green triangles). The AWR scores were higher in the WAS plus C. rodentium infectious group compare with control group (** P < 0.01), and the co-treatment of C. butyricum could reverse this tendency (* P < 0.05) while co-treatment of antibiotics had no such effect on it (P > 0.05). When the water injected into balloons was 0.15 ml, the AWR scores of C. rodentium group was apparently higher than control group (## P < 0.01). Data are reported as mean ± SD; paired Student T-test. (B–D) H&E-stained distal colon sections from C. rodentium infectious mice at 0 day (B), 5 days (C), and 21 days (D) after infection. No inflammation is evident in sections of untreated mice and 21 days after infection mice, while there was inflammatory cell infiltration and hyperplasia in the mucosa of C. rodentium infectious mice at day 5.

C. butyricum could relieve the visceral hypersensitivity of PI-IBS-like mice

We evaluated whether the visceral hypersensitivity caused by C. rodentium infection and WAS could be influenced with probiotics or antibiotics treatment. It was observed that mice treated with the C. butyricum presented lower AWR scores in the behavioral testing than C. rodentium infected mice group (n = 6, P < 0.05). However, the gentamycin-ampicillin treatment group did not show this characteristic (n = 6, P > 0.05) (Fig. 2A).

The OTUs of gut microbes tremendously decreased after antibiotics treatment,which was not observed in C. butyricum treatment mice

The 16S rRNA sequence was performed to investigate the change of the gut microbiota. It was found that the number of OTUs were among 661–673 at the beginning in different groups. The C. rodentium infection and C. butyricum treatment did not affect the number of OTUs. However, the number of OTUs obviously decreased to 580 at day 7 and further tremendously decrease to 122 at day 14 in antibiotic treatment group (Fig. 3B). After WAS, the number of OTUs in antibiotic group decreased to 111 at day 30 (Fig. 3B). At day 30, there were only 69 OTUs which were shared in the four groups. In contrast, there were 456 OTUs which were shared in the three groups except for the antibiotics group (Fig. 3A). These results demonstrated the antibiotic indeed destroy the abundance of microbiota in PI-IBS-like mice, while C. butyricum did not influence it.

Comparison of the OTU of the microbiota in different groups.

Figure 3: Comparison of the OTU of the microbiota in different groups.

(A)The 16S rRNA sequence for the number of OTUs plotted to the Venn diagram of the four experimental groups at day 30 after WAS. There were only 69 OTUs which were shared in the four groups. In contrast, there were 456 OTUs which were shared in the three groups except for the antibiotics group. (B) OTU numbers were counted in the four groups at different times (day 0, day 7, day 14 and day 30). The antibiotic treatment after C. rodentium infection demonstrated a low bacterial species diversity compared with control and other infectious groups.

The α-diversity of gut microbiota obviously changed after antibiotics treatment, and the C. butyricum could maintain them within normal range

The α-diversity of gut microbiota was analyzed and the results are shown in Fig. 4. It was found that there was no difference in the Shannon and Chao indexes at day 0 (data not shown). After C. rodentium infection, the Shannon and Chao indexes were not affected at day 7, day 14 and day 30. The treatment of C. butyricum also did not affect the Shannon and Chao indexes. However, the Chao index of the antibiotics group significantly decreased compared with the control group (519.45 ± 52.71 VS. 436.72 ± 36.68, P < 0.05) at day 7. Also, both the Shannon and Chao indexes of the antibiotics group were significantly lower than those of the C. butyricum-treatment group (3.45 ± 0.41VS4.26 ± 0.41, P < 0.01; 436.72 ± 36.68VS.524.12 ± 91.84, P < 0.05). At day 14, the Shannon and Chao indexes of the antibiotics group decreased and were significantly lower than those of the other three groups (P < 0.001). At day 30, the Shannon and Chao indexes of the antibiotics group were still significantly lower than those of the C. rodentium infection group and the C. butyricum-treatment group (P < 0.001).

The α-diversity of gut microbiota in the four groups at different times.

Figure 4: The α-diversity of gut microbiota in the four groups at different times.

The Chao index of the antibiotic group decreased compared with the control group (519.45 ± 52.71 VS. 436.72 ± 36.68, P < 0.05) at day 7; both the Shannon and Chao indexes of the antibiotic group were also lower than the C. butyricum-treatment group (3.45 ± 0.41VS4.26 ± 0.41, P < 0.01; 436.72 ± 36.68VS.524.12 ± 91.84, P < 0.05) at day 14; at day 30, the Shannon and Chao indexes of the antibiotic group were still lower in antibiotic group compared with the other three groups (P < 0.001). The C. butyricum-treatment group did not affect the Shannon and Chao indexes at day 7, day 14, and day 30.

The β-diversity of the gut microbiota changed in different groups

The β-diversity of the gut microbiota was analyzed based on PCoA and the results are shown in Fig. 5. The structure of the gut microbiota was nearly the same at day 0 (R = 0.01, P = 0.41). It was found that C. rodentium infection did not affect the structure of the gut microbiota compared with the control group (R = 0.16, P = 0.09). However, antibiotic treatment significantly affected the structure of the gut microbiota at day 7 (R = 0.32, P = 0.001) and day 14 (R = 0.55, P = 0.001). At day 14, the structure of the microbiota was similar between the control and the C. butyricum groups. However, there were significant differences between the four groups (P < 0.05) (Fig. S1). After the WAS, the difference between the antibiotic treated group and the other three groups was further increased (R = 0.58, P = 0.001) (Fig. S2).

The β-diversity of the gut microbiota in the four groups was analyzed based on PCoA at different times.

Figure 5: The β-diversity of the gut microbiota in the four groups was analyzed based on PCoA at different times.

The structure of the gut microbiota was nearly the same at day 0 (R = 0.01, P = 0.41). C. rodentium infection did not affect the structure of the gut microbiota compared with the control group (R = 0.16, P = 0.09). However, antibiotics treatment significantly affected the structure of the gut microbiota at day 7 (R = 0.32, P = 0.001) and day 14 (R = 0.55, P = 0.001). After the WAS, the difference between the antibiotic treated group and the other three groups was further increased (R = 0.58, P = 0.001).

The microbiota composition changed in different groups

The main genera of the gut micriobiota (the percentages were above 1%) included more than 38 genera in all groups including Muribaculaceae, Lactobacillus, Bacteroides, Lachnospiraceae, Ileibacterium, Alistipes, and Helicobacter, ect. (Fig. 6A). The specific taxa that most likely contributed to the differences between the four groups at day 30 were determined by linear discriminant analysis effect size (Fig. 6B). The genera Lactobacillus, Turicibacter and Flavobacteriaceae were enriched in the control group while the genera Muribaculaceae, Lachnospiraceae, Helicobacter, Ileibacterium, Alistipes, and Ruminococcus were enriched in the C. rodentium infection group. After treatment with antibiotics, the genera Bacteroides, Blautia, Parasutterella, Parabacteroides, and Akkermansia were enriched. In contrast, the genera Oscillospirales, Mucispirillum, Dubosiella, and Odoribacter were enriched in the C. butyricum-treatment group. The amount of genera Lactobacillus and Odoribacter in the four groups are shown in Figs. 6C and 6D. It was found that the amount of the genus Lactobacillus in the control group was higher than those of other groups (P < 0.01). It increased in C. butyricum-treatment group without significant difference. Furthermore, the genus Odoribacter in the C. butyricum-treated group was higher than those of the other three groups (P < 0.05).

The microbiota composition changed in different groups.

Figure 6: The microbiota composition changed in different groups.

(A) The main genera of the gut microbiota (the percentages were above 1%) included more than 38 genera in all groups. (B) The specific taxa that most likely contributed to the differences between the four groups at day 30 were determined by linear discriminant analysis effect size. (C) The amount of genera Lactobacillus and Odoribacter in the four groups were all increased in the C. butyricum-treatment group compared with other two C. rodentium infectious groups. The genus Lactobacillus in the control group was higher than those of other groups. (* P < 0.05, ** P < 0.01, *** P < 0.001, **** P < 0.0001).

Discussion

In this study, after treatment with C. butyricum, the PI-IBS mouse model presented a lower AWR scores which means mitigatory peritoneal hypersensitivity. In contrast, antibiotics treatment had no effects on hypersensitivity. We also found that C. butyricum treatment could remodeling of the gut microbiota. However, antibiotics treatment could induce a drastic change in the intestinal microbiota.

In recent years, it was reported that lots of probiotic supplements have favorable effects on IBS as revealed through in clinical cases and animal studies (Ford et al., 2014; Didari et al., 2015). As the most commonly used probiotics, Bifidobacterium is of great benefit to on abdominal pain and bloating in IBS patients (Sun et al., 2018). Lactobacillus acidophilu s improve an improvement of barrier function and reduce cytokines secretion, which contribute to resulting in the benefit of relieving visceral hypersensitivity of IBS (Zeng et al., 2008). There was also a PI-IBS rodent model study to induce colitis through the Trichinella, and the intervention of Lactobacillus and yeast showed a decrease of visceral hypersensitivity in mice (Hong et al., 2019). The Trichinella causes persistent and histoclastie infection in gut tissue (Luo et al., 2019), which may not reflect the pathogenic condition of PI-IBS. In this study, the infection of C. rodentium is a self-limited colitis in mice which have a histological recovery in 10 days after infection (Collins et al., 2014). This rodent model could mimic a similar pathological and immunological state of PI-IBS cases. The C. butyricum is a well-known probiotic which was used for therapeutic strategy for IBS patients (Sun et al., 2018; Zhao et al., 2019). It revealed an important role in a hypersensitivity-relieving effect in the C. rodentium infection plus WAS mice model (Fig. 2A). Antibiotics are commonly used in the treatment of infectious diarrhea. In this study, an antibiotic mixture of gentamycin-ampicillin was also used to treat the C. rodentium infection. Though the fecal character greatly changed, it demonstrated no significant differences in hypersensitivity of infected mice plus WAS (Fig. 2A). Thus, the C. butyricum might be a better choice for infectious diarrhea than antibiotics.

There have been many studies on the pathogenesis of IBS, and one of the most important explains is gut microbiota remodeling. Gastroenteritis episodes induce variation in the gut microbiota, which usually return to normal in healthy people after recovery. However, those developing PI-IBS may have a relative inability to restore their microbiota (Neal, Barker & Spiller, 2002). It was reported that the C57BL/6 mice used in our study is one of C. rodentium-resistant mouse line which could only get a self-limiting colitis and more mild symptoms of diarrhea and weight loss compare to the sensitive mouse lines (Vallance et al., 2003). The amount of C. rodentium were only about 102–104/g feces 3 days after infection in the resistant mice (Osbelt et al., 2020). And the α-diversity remained stable and non-significant decrease during the infection until day 14 (Cannon et al., 2020). However, the intervention of C. rodentium did affect the structure of gut microbiota same as the published reports (Osbelt et al., 2020; Hoffmann et al., 2009). In this study, the structure of the microbiota was still not restored at day 14 after infection without treatment or treatment with antibiotic (Fig. 5 and Fig. S1). In contrast, the structure of the microbiota of the mice in the C. butyricum treatment group was similar to those of the control group (Fig. S1), which indicated its effect on maintaining the balance of gut microbiota. Also, chronic persistent stress, such as WAS, plays an important role in remodeling the gut microbiota of mice (Han et al., 2020).

In this study, we found that WAS treatment remodeled the microbiota and the Shannon and Chao indexes even decreased in control group (Fig. 3). However, the Shannon and Chao indexes did not change after WAS treatment in the C. butyricum treatment group. It was interesting that only the structure of the microbiota of the C. butyricum treatment group did not change after WAS (Fig. S2). This suggested that C. butyricum could protect the microbiota from chronic persistent stress. Furthermore, the genera Odoribacter and Lactobacillus increased more in the C. butyricum treatment group at day 30 than in the C. rodentium infections group and in the antibiotic group. It was reported that higher baseline proportions of Odoribacter were related to beneficial inflammatory-marker changes (Hod et al., 2018). In addition, probiotics of the genus Lactobacillus are also have potentially reduced abdominal symptoms of IBS (Bonfrate et al., 2020; Preston et al., 2018). These may be the reason why C. butyricum alleviates visceral hypersensitivity.

On the other hand, the antibiotics can affect the gut microbiota of PI-IBS (Becattini, Taur & Pamer, 2016). In this study, we found that the Shannon and Chao indexes as well as the structure of the microbiota significantly changed after the treatment of antibiotics (Figs. 4 and 5) and the tendencies could last for 30 days. It was also found that the genus Bacteroides increased at day 14 and 30 (Fig. 6A) in antibiotic group. The increase of the Bacteroidetes phylum composition may increase the susceptibility to infection and it is also abundant in PI-IBS patients (Downs et al., 2017). Therefore, probiotics are more suitable than antibiotics in remodeling the microbiota of patients after infection.

There may be some possible limitations in this study. First, whether or not probiotics include other strains of Clostridium have the same effect on the visceral hypersensitivity induced by the intestinal infection was not confirmed. Second, the PI-IBS-like rodent model can not identically imitate the pathological conditions of PI-IBS patients, it need to continually improve. Thirdly, we observed the C. butyricum could relive the visceral hypersensitivity and the gut microbiota was closely related to it, the specific molecular or immune mechanism was not explored and need further research.

Conclusions

In conclusion, it was found that the C. butyricum could relieve the visceral hypersensitivity in mice induced by C. rodentium infection plus chronic stress. It was also superior to antibiotics in remodeling the microbiota of the visceral hypersensitivity model mice. Thus, C. butyricum may be a candidate for the treatment of PI-IBS.

Supplemental Information

The β-diversity of the gut microbiota in the three C. rodentium infectious groups compared with control group, based on PCoA at day 14

The structure of microbiota was similar between the C. rodentium infectious group and control group, which was also happened in the C. butyricum group. However, there were significant differences between the antibiotic-treatment group and control group (P < 0.05).

DOI: 10.7717/peerj.11585/supp-1

Effect of WAS on the structure of the microbiota of four groups

After the WAS at day 30, the difference between the antibiotic antibiotic-treatment group and the other three groups was further increased. However, the structure of the microbiota of the C. butyricum - treatment group did not change after WAS.

DOI: 10.7717/peerj.11585/supp-2

The raw data and analysis of AWR scores in Fig. 2A

DOI: 10.7717/peerj.11585/supp-3

Sample name and group information of 16s gene data

DOI: 10.7717/peerj.11585/supp-4

ARRIVE 2.0 Checklist

-

DOI: 10.7717/peerj.11585/supp-5
1 Citation   Views   Downloads