SEARCH

SEARCH BY CITATION

Keywords:

  • behavior;
  • germfree;
  • microbiota;
  • stress

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Funding
  8. Disclosures
  9. Author Contributions
  10. References

Background

There is increasing evidence suggesting the existence of an interaction between commensal microbiota, the gut and the brain. The aim of this study was to examine the influence of commensal microbiota on the host behaviors in a contamination-free environment, which was verified by culture-based methods.

Methods

Open-field and marble-burying tests were used to analyze anxiety-like behaviors and locomotor activity in gnotobiotic BALB/c mice with a common genetic background in a sterile isolator. The monoamine levels in several regions of the brain were measured in germfree (GF) mice and commensal fecal microbiota-associated mice (EX-GF).

Key Results

A 24-h exposure to the environment outside the sterile isolators rendered GF mice less anxious than those not contaminated, while there was no change in the locomotion. EX-GF mice, the gnotobiotic mice with normal specific pathogen-free microbiota, were less anxious and active than GF mice using open-field and marble-burying tests. The norepinephrine, dopamine, and serotonin turnover rates were higher in the EX-GF mice than in the GF mice in most regions of the brain, suggesting that monoaminergic neurotransmission might increase in the EX-GF mice comparing the GF mice. Monoassociation with Brautia coccoides reduced the anxiety level, but it did not affect the locomotor activity. In contrast, colonization with Bifidobacterium infantis decreased the locomotor activity, while having little effect on the anxiety level.

Conclusions & Inferences

These results strongly support the current view that gut microorganisms modulate brain development and behavior.


Abbreviations
GF

germfree

SPF

specific pathogen free

Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Funding
  8. Disclosures
  9. Author Contributions
  10. References

There is an increasing and intense current interest in the role that gut bacteria play in maintaining the health of the host. Gut microbiota have now been estimated to include at least 1800 genera and up to 40 000 species of bacteria based on the analysis of 16S ribosomal RNA.[1] They have an estimated mass of 1–2 kg, number 100 trillion, and together possess 100 times the number of genes present in the human genome.[2] These bacteria not only play a principal role in the postnatal maturation of the mammalian immune system[3, 4] but also aid in the digestion and absorption of macromolecules and act as a barrier to gut pathogens by blocking attachment to gut binding sites, which is the first step of bacterial pathogenicity.[5] Moreover, it is also rapidly becoming apparent that the gut microbiota plays a major role in the development and regulation of neuroendocrine systems such as the hypothalamic–pituitary–adrenal axis, a central integrative system crucial for the successful physiological adaptation of the organism to stress. In fact, our previous study on germ free (GF) and gnotobiotic mice demonstrated that exposure to gut microbes is a critical environmental determinant that regulates the development of the HPA stress response and also the set point for this axis.[6] These findings provided evidence of a novel link between indigenous microorganisms and the nervous system, indicating a new aspect of the ‘gut-brain axis’. Further progress in this field has since been achieved by several research groups,[7-10] and this concept is now integrated into an elaborate interaction among the microbiota, the gut and the brain, which together are known as the ‘microbiota-gut-brain axis’.[8, 11]

Recently, animal studies done by several independent groups show the commensal microbiota to be a crucial factor for modulating the host behavioral profile.[12-15] These exciting studies certainly open up a new avenue for this research field; however, no behavioral analyses have so far been performed in a strictly controlled, contamination-free environment.

In this study, we developed a novel method for evaluating mouse behavior inside a contamination-free environment using a closed population of an inbred strain of mice with a common genetic background. Based on this method, behavioral analyses were conducted in gnotobiotic mice that were kept contamination free even at the time of behavioral testing.

Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Funding
  8. Disclosures
  9. Author Contributions
  10. References

Animals

Germfree BALB/c mice were originally obtained from Central Institute for Experimental Animals (Kawasaki, Japan). These mice have been bred more than 10 generations in isolators without bacterial contamination in our laboratory. Some pairs of male and female GF mice were picked up as grandparents of mice for experiments. Their first offspring were used as parents of gnotobiotic mice. To make gnotobiotic mice associated with gut microbiota from specific pathogen-free (SPF) mice, the parents mice were orally given stools of SPF mice, and their offsprings were used as EX-GF mice.[6, 16] To produce such gnotobiotic mice associated with Bifidobacterium infantis (Bi) or Blautia (Clostridium) coccoides (Bc), 109 CFU of live Bi or Bc suspended in 0.5 mL of phosphate buffered saline were intragastrically inoculated into the parents, and their offspring were used as Bi- or Bc-associated mice in the experiments. The Bi is a predominant bacterium present in the infant gut.[17] The Bc is the most dominant inhabitant of the human gut,[18] and is also reported to be the dominant subgroup of clostridia, comprising a mean of 43% of the total bacteria in the fecal flora of healthy adults.[19] All the mice were maintained in Trexler-type flexible-film plastic isolators with sterile food (CL-2; CLEA Japan Inc., Tokyo, Japan) and water. Only male mice were used for the experiments. Surveillance for bacterial contamination was thoroughly performed by a periodic bacteriologic examination of feces.

All experiments were approved by the Ethics Committee for Animal Experiments of Tokai University.

Determination of fecal bacteria

The bacterial composition of intestinal contents of mice was enumerated according to a method described previously.[3, 6] Briefly, an approximately 0.1 g (wet weight) fecal sample was suspended in a serially diluted tenfold from 10−1 to 10−8. A 0.05 mL aliquot of each dilution was then spread on three non-selective and 11 selective agar plates. The plates were incubated for 3 days, and each bacterial colony was identified and enumerated. In the experiments using monoassociated mice, the mice were confirmed to be monoassociated according to results showing that only Bi or Bc was detected in their stool. The mean counts of Bi and Bc were 10 and 9.5 log10 CFU g−1 of feces, respectively.

Behavioral analyses

Open-field (OF) and marble-burying (MB) tests were selected as behavioral analyses of gnotobiotic mice because the validity and reliability of these tests are confirmed by a large number of previous studies,[20] and the instruments used for the tests are not too large in size to be put into a plastic isolator. Boxes for OF and MB tests were already placed in a sterile isolator, before moving mice there from the other isolator in which they had been bred. The test was performed 7 days after the sterilization of stock refilling to avoid any influence of peracetic acid, which was used for sterilization of the isolator, on the behavior of the mice. Behavioral tests of the mice were started at 7 weeks of age, and performed between 9:00 a.m. and 5:00 p.m. under low illumination (<5 lux). To minimize the variation with time, the tests of the two groups were equally distributed and performed alternately. The test room was maintained at 21 ± 1 °C.

OF test

The OF test was carried out to evaluate the levels of anxiety and locomotor activity, according to the method described previously with some modifications.[21, 22] In brief, mice were placed individually in the center of an OF box (L × B × H, 20.3 cm × 20.3 cm × 26.0 cm) in a sterile isolator. The square bottom of the box was equally divided into 16 subsquares (4 x 4). Their behaviors during the experiments were recorded and calculated by a computer system (Limelight TM version 2.57, Actimetrics, Inc., Wilmette, IL, USA). Total distance travelled for 30 min (DT30) was automatically calculated as spontaneous locomotor activities. The time spent in the 12 peripheral subsquares for 30 min (TS30) was also automatically calculated as a parameter of anxiety-like behavior.

MB test

The MB test was performed in a sterile isolator using a transparent polycarbonate box (L × B × H, 20.3 cm × 30.0 cm × 26.0 cm) with a 5-cm sawdust layer covering the floor (Fig. 1). Although the MB test has been established as a widely used model of anxiety due largely to the observation that ‘traditional’ anxiolytics, such as benzodiazepines, are effective at reducing MB behavior,[23] it is also regarded as a model for obsessive–compulsive disorder, due to the pharmacological efficacy of clinically applied drugs used for the treatment of obsessive–compulsive disorder.[24, 25]

image

Figure 1. Behavior analysis in a sterile isolator. Panels (A) and (B) show representative photo images of the system to evaluate marble-burying behavior of the mice in an isolator.

Download figure to PowerPoint

Twenty clear green glass marbles (diameter 1.5 cm) were evenly spaced over the sawdust layer (four rows of five marbles per row). The mice were individually left for 30 min in the box placed in an aseptic isolator. Thereafter, they were moved away from the box and then the number of buried marbles (NBM), defined as at least two thirds covered, was counted as a parameter of anxiety-like behavior.

Measurement of monoamine levels in the brain

Germfree and EX-GF mice derived from the same grandparents were killed by cervical dislocation in an isolator at 16 weeks of age. The brains were removed and dissected to the medial prefrontal cortex, striatum, hippocampus, and brainstem. The experiments to obtain brain samples were performed between 9:00 and 12:00 hours. Samples were weighed and kept at −80 °C until the assay. The levels of monoamines and their metabolites (contents/wet tissue) were analyzed using a previously described method with some modifications.[26] Briefly, the samples were homogenized and deproteinized in 0.2 mol L−1 perchloric acid containing 100 mmol L−1 EDTA 2Na. The homogenate was left for 30 min for deproteinization. The homogenate was centrifuged at 20 000 × g for 15 min at 0 °C. The pH of supernatant was adjusted to approximately 3.0 by adding 1 mol L−1 sodium acetate and the resultant supernatant was filtered through 0.2 μm filter (Millipore, Bedford, MA, USA). A 30 μL portion of filtrate was applied to a high-performance liquid chromatography (HPLC) system (Eicom, Kyoto, Japan) with a 150 mm × 3.0 mm octadecyl silane (ODS) column (SC-5ODS; Eicom) and an electrochemical detector (ECD-300, Eicom) at an applied potential of +0.75 V vs Ag/Ag Cl reference analytical electrode. Changes in the electric current (nA) were recorded in a computer using an interface system (Power Chrom ver 2.3.2.j, AD Instruments, Tokyo, Japan). The mobile phase contained 0.1 mol L−1 aceto-citric acid buffer (pH 3.5), methanol, 0.46 mol L−1 sodium 1-octane sulfonate, and 0.015 mmol L−1 disodium ethylenediaminetetraacetic acid (830 : 170 : 1.9 : 1) at a flow rate of 0.5 mL min−1. The concentrations of monoamines and metabolites including dopamine (DA), norepinephrine (NE), serotonin (5-HT), the DA metabolite dihydroxyphenylacetic acid (DOPAC) and homovanillic acid (HVA), NE metabolite 3-methoxy-4-hydroxyphenylglycol (MHPG), and 5-HT metabolite 5-hydroxyindoleacetic acid (5-HIAA) were determined, and their levels in the brain were calculated. Turnover rates (DOPAC/DA, HVA/DA, MHPG/NE, and 5-HIAA/5-HT) were also calculated. The limit of detection of the system for all monoamines was 0.1 pg/sample. The tryptophan levels were also done by the above HPLC system according to the previous method with some modifications.[27]

In another set of experiments, the brain monoamine levels were measured using the GF mice, which were kept in a sterile isolator, and the ‘GF mice’, which were subjected to a 24-h exposure to the external environment.

Statistical analysis

All the continuous variables are presented as the means ± standard error (SE). Comparison of changes between the two groups was performed using paired or unpaired t-test. In some experiments, it was also evaluated using linear mixed models for repeated observations. The comparison of the changes among the three groups was performed by Dunnett's post hoc test after the analysis of variance. All analyses were performed using IBM SPSS Statistics version 19.0.0 for Windows (SPSS Inc., Chicago, IL, USA). A P value of less than 0.05 was considered significant.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Funding
  8. Disclosures
  9. Author Contributions
  10. References

Exposure to a conventional environment renders GF mice less anxious

The GF mice were subjected to both OF and MB tests before and 24 h after exposure to an external environment outside the sterile isolator. In another set of experiments, the GF mice of the same age were given both OF and MB tests twice at 24-h intervals under a contamination-free environment, which was confirmed by culture-based methods. As shown in Fig. 2, there was no significant difference in the DT30 between before (1st test) and after exposure to the external environment (2nd test); however, such exposure significantly decreased the TS30. The NBM was lower in ‘GF mice’ exposed to conventional environment than in those kept under sterile conditions. In contrast, the GF mice failed to show any significant variation in the TS30 and NBM between the first and the second behavioral analyses. The bacteriological analysis using culture method showed such ‘GF mice’ to be colonized with a significant number of commensal bacteria such as Enterococci and Staphylococci in the fecal contents (Table 1). However, the 24-h exposure of the GF mice to the external environment did not affect either the monoamines or their metabolites levels in the medial prefrontal cortex, striatum, hippocampus, or brainstem of the brain (Table 2).

Table 1. Number of bacteria in the intestinal contents
ID no.Number of bacteria (log10 CFU/g contents)
EnterococciStaphylococciBacilliEnterobacteriaAnaerobes
  1. The number of bacteria in intestinal contents of ‘GF mice’ (n = 10) was analyzed at 24 h after the mice were moved to semiclean room. CFU, colony-forming unit; ND, not detected.

14.65.9NDNDND
24.35.3NDNDND
34.3NDNDNDND
44.13.5NDNDND
54.1NDNDNDND
62.85.4NDNDND
72.65.8NDNDND
82.62.3NDNDND
92.32.3NDNDND
102.3ND3.1NDND
Mean (SE)3.4 (0.9)3.0 (0.7)~~~
Table 2. Effects of exposure to external environment on monoamine levels in various regions of the brain
 GF24 h-GF
  1. The GF mice at 8–10 weeks of age were divided into the following two groups: one group was kept in a sterile isolator (GF, n = 6), while the other was subjected to a 24-h exposure to the external environment (24 h-GF, n = 8). Twenty-four hours later, their brains were removed and processed for the monoamine analysis. All data represent the means ± SE (ng g−1). The concentration of 5-HT, DA, NA, and their major metabolites in each brain region are expressed as nanograms per gram of tissue wet weight.

Prefrontal cortex
5-HT284.9 (24.8)268.8 (9.1)
5-HIAA275.9 (29.9)316.0 (44.6)
5-HIAA/5-HT0.97 (0.09)1.17 (0.14)
DA361.9 (27.5)448.0 (66.0)
DOPAC84.8 (6.4)114.0 (11.2)
HVA104.4 (6.4)127.0 (11.8)
DOPAC/DA0.24 (0.01)0.27 (0.03)
HVA/DA0.29 (0.02)0.30 (0.02)
NE226.0 (44.5)262.0 (31.0)
MHPG68.13 (6.1)84.3 (7.9)
MHPG/NE0.53 (0.27)0.39 (0.09)
Hippocampus
5-HT420.6 (38.6)406.2 (25.2)
5-HIAA402.8 (49.4)312.8 (26.4)
5-HIAA/5-HT0.95 (0.05)0.78 (0.06)
DA1921 (382)1963 (184)
DOPAC399.4 (69.2)316.6 (53.4)
HVA332.7 (57.3)341.1 (39.6)
DOPAC/DA0.22 (0.02)0.16 (0.02)
HVA/DA0.18 (0.02)0.17 (0.01)
NE487.7 (61.2)386.5 (45.1)
MHPG84.0 (12.6)82.8 (6.6)
MHPG/NE0.18 (0.04)0.23 (0.03)
Striatum
5-HT512.0 (56.8)564.8 (25.2)
5-HIAA342.7 (36.9)440.6 (41.2)
5-HIAA/5-HT0.69 (0.06)0.78 (0.07)
DA3065 (631)2929 (550.8)
DOPAC501.8 (99.6)574.9 (195)
HVA464.7 (98.8)545.5 (143)
DOPAC/DA0.17 (0.01)0.18 (0.03)
HVA/DA0.16 (0.01)0.09 (0.01)
NE639.1 (90.1)750.3 (89.4)
MHPG73.8 (11.0)81.4 (13.9)
MHPG/NE0.13 (0.03)0.12 (0.02)
Brainstem
5-HT517.6 (27.8)443.1 (43.8)
5-HIAA416.7 (50.7)393.0 (18.0)
5-HIAA/5-HT0.82 (0.12)0.95 (0.12)
DA602.6 (104)674.8 (70.7)
DOPAC49.6 (9.5)51.3 (5.6)
HVA55.1 (10.4)103.2 (16.0)
DOPAC/DA0.08 (0.01)0.08 (0.01)
HVA/DA0.09 (0.01)0.16 (0.03)
NE735.8 (57.7)709.9 (59.0)
MHPG67.4 (6.7)62.3 (8.0)
MHPG/NE0.09 (0.01)0.09 (0.01)
image

Figure 2. Effects of exposure to the external environment on behaviors and fecal contents. GF mice at 8–10 weeks of age underwent both the OF and MB tests (1st test) in a sterile isolator. Then, they were divided into the following two groups: the one group of mice (closed column) was taken out of the isolator and moved to a semiclean room generally used for SPF mice, whereas the other (open column) was moved to another isolator kept under strictly germfree conditions. Both groups of mice were again taken back into the original isolator 24 h after being moved, and underwent the second test of their behaviors. All data are expressed as the mean ± SE (n = 10 per group). *P < 0.05 and ***P < 0.001 were significantly different from the corresponding basal value.

Download figure to PowerPoint

These results thus indicated that even a short-time exposure to the environment outside the sterile isolators makes ‘GF mice’ less anxious, which may be related to colonization with the bacteria of environmental origin.

Commensal microbiota decreases both spontaneous locomotor activity and anxiety-like behavior of the host

We therefore performed all subsequent behavioral tests in a sterile isolator using GF, EX-GF, Bc, and Bi mice at 7, 10, and 16 weeks of age.

Figure 3 shows that both the DT30 and TS30 were significantly lower in EX-GF than in GF mice at all the time points including 7, 10, and 16 weeks old. The TS30 was significantly lower in the Bc mice than in the GF mice both at 7 and 10 weeks old, whereas there was no significant difference in the DT30 between the Bc and GF mice at all ages tested. On the contrary, the Bi mice showed a significant decrease in DT30, but a comparable level of TS30 in comparison with GF mice at all ages tested. The NBM was significantly lower in EX-GF than in GF mice at 10 and 16 weeks old. The Bc mice also exhibited a significant decrease in the NBM at all time points tested in comparison with the GF mice. Repeated measures anova showed that there were no significant differences in the NBM between the GF and Bi mice. The linear mixed models after adjusting for locomotor activity also demonstrated that both the TS30 and NBM were still significantly lower in EX-GF than in GF mice even after adjusting for DT30 (TS30, P < 0.001; NBM, P < 0.05), which thus indicates that the anxiety levels evaluated by the TS30 and NBM were indeed lower in the EX-GF mice than in the GF mice across different time scales regardless of the degree of locomotor activity.

image

Figure 3. Effects of commensal microbiota on behaviors of the host. GF, EX-GF, Bc, and Bi mice at the age of 7, 10, and 16 weeks were subjected to OP and MB tests, and DT30, TS30, and NBM were evaluated as described in the methods. All data are expressed as the mean ± SE (n = 14–19 per group). *P < 0.05, **P < 0.01 and ***P < 0.001 were significantly different from the corresponding GF values.

Download figure to PowerPoint

These results thus indicate that either a complete indigenous microbiota or some strains of anaerobic bacteria residing in the gut are capable of changing some aspects of the host behavioral phenotype.

Commensal microbiota influence monoamine and metabolites levels in the host brain

Finally, we investigated the influence of commensal microbiota on brain monoamine levels of their host. As shown in Table 3, 5-HIAA levels showed a significant decrease in the brainstem of EX-GF mice compared with GF mice. 5-HIAA/5-HT turnover rates were significantly higher in the striatum of the EX-GF mice than in that of the GF mice. No significant difference in tryptophan level, a precursor amino acid of 5-HT, was found in any regions of the brain between the GF and EX-GF mice.

Table 3. Monoamine, tryptophan, and major metabolite levels in various regions of the brain
 GFEX-GF
  1. All data represent the means ± SE (ng g−1). The concentration of NA, DA, 5-HT, and their major metabolites in each brain region are expressed as nanograms per gram of tissue wet weight. Tryptophan level is milligram per gram of tissue wet weight. *P < 0.05 was significantly different from the corresponding GF values (n = 7–10 per group).

Prefrontal cortex
Tryptophan5.2 (1.17)5.6 (0.35)
5-HT317.4 (53.9)343.6 (25.8)
5-HIAA194.2 (28.2)217.6 (33.5)
5-HIAA/5-HT0.637 (0.03)0.633 (0.03)
DA340.2 (64.1)171.0 (18.9)*
DOPAC44.2 (5.8)37.5 (2.5)
HVA129.1 (32.8)97.1 (8.7)*
DOPAC/DA0.152 (0.03)0.232 (0.02)*
HVA/DA0.301 (0.02)0.593 (0.05)*
NE261.9 (49.8)328.3 (18.9)
MHPG63.8 (11.2)94.7 (10.0)
MHPG/NE0.271 (0.04)0.316 (0.03)
Hippocampus
Tryptophan4.7 (0.91)4.0 (0.40)
5-HT454.4 (53.7)432.2 (30.2)
5-HIAA279.4 (24.3)305.8 (27.7)
5-HIAA/5-HT0.717 (0.12)0.729 (0.07)
DA1223 (358)1049 (328)
DOPAC151.0 (40.9)92.1 (27.1)
HVA218.4 (32.7)190.5 (45.8)
DOPAC/DA0.146 (0.03)0.180 (0.03)
HVA/DA0.242 (0.06)0.328 (0.07)
NE332.1 (48.5)307.9 (19.7)
MHPG76.3 (16.8)108.8 (6.8)
MHPG/NE0.232 (0.04)0.372 (0.04)*
Striatum
Tryptophan3.4 (0.33)3.6 (0.21)
5-HT718.6 (74.1)681.9 (40.9)
5-HIAA466.7 (94.2)461.7 (33.5)
5-HIAA/5-HT0.541 (0.04)0.701 (0.04)*
DA2018 (498)941.0 (311)
DOPAC244.7 (46.2)135.2 (23.1)*
HVA374.7 (73.6)238.7 (46.6)
DOPAC/DA0.141 (0.02)0.237 (0.04)
HVA/DA0.219 (0.03)0.396 (0.06)*
NE732.3 (114.6)527.4 (27.4)
MHPG85.3 (9.5)112.0 (7.4)*
MHPG/NE0.123 (0.01)0.214 (0.02)*
Brainstem
Tryptophan2.9 (0.13)3.2 (0.31)
5-HT506.6 (28.2)461.8 (22.0)
5-HIAA397.7 (30.1)304.1 (13.7)*
5-HIAA/5-HT0.785 (0.04)0.667 (0.04)
DA233.2 (26.0)79.0 (7.3)*
DOPAC18.1 (1.6)19.6 (1.5)
HVA34.8 (2.2)34.8 (2.2)
DOPAC/DA0.08 (0.01)0.246 (0.03)*
HVA/DA0.171 (0.03)0.474 (0.05)*
NE572.2 (35.4)472.5 (26.1)*
MHPG62.5 (6.4)110.7 (8.8)
MHPG/NE0.110 (0.01)0.225 (0.01)*

Dopamine levels exhibited a significant lower level in the prefrontal cortex and brainstem of the EX-GF mice than in those of the GF mice. Lower striatum DOPAC and cortex HVA levels were also found in the EX-GF mice in comparison with the GF mice. The EX-GF mice showed significantly higher turnover rates of not only DOPAC/DA in the brainstem and medial prefrontal cortex but also HVA/DA in the brainstem, striatum, and medial prefrontal cortex in comparison with the GF mice. MHPG levels were higher in the striatum and brainstem of the EX-GF mice than in those of the GF mice. MHPG/NA rates were also significantly higher in the brainstem, striatum, and hippocampus of EX-GF than in those of GF mice, thus suggesting that catecholaminergic neurotransmission might be increased in EX-GF mice in comparison with GF mice.

Taken together, these results indicate that commensal microbiota affect the brain monoamine metabolism, which may be involved in the modulation of host behavioral phenotypes.

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Funding
  8. Disclosures
  9. Author Contributions
  10. References

The current study demonstrated the EX-GF mice to exhibit a significant reduction in both the spontaneous motor activity and anxiety-like behavior in comparison with the GF mice under contamination-free experimental conditions, which was verified using culture-based methods. Monoassociation with Bc, which is a predominant bacterium in human gut,[18] also decreased the anxiety level, but failed to affect the locomotor activity. On the contrary, monocolonization with Bi, which is a major species of Bifidobacterium in the infant gut,[17] reduced the locomotor activity but had little effect on the anxiety level. The NE and DA turnover rates were higher in the EX-GF mice than in the GF mice in several regions of the brain, whereas the 5-HT turnover rates were significantly higher in the striatum, but lower in the brainstem of the EX-GF mice than in that of the GF mice. These results thus provide strong evidence that commensal microbiota influence the development of the brain and thus alter the behavior in their hosts. A change in brain monoamines may be involved in the mechanism for this microbiota-induced alteration in the host behaviors.

Two research groups reported[13, 14] that GF mice are less anxious than SPF mice when tested by the light-box test or the elevated plus maze test. In contrast, the current study found that commensal microbiota reduced both the spontaneous motor activity and anxiety-like behavior of their host. This reduced anxiety-like behavior was also verified by a multivariate statistical analysis after adjusting for motor activity. It should be noted that the current study was completed using male BALB/c mice rather than female Swiss Webster or NMRI mice, as in the previous works. Hence, both the strain and sex of the mice may have contributed to the observed differences.

In the current study, bacterial examination was done on only a limited number of culturable bacteria because the early bacterial communities formed during colonization in mammals are composed primarily of the facultative anaerobes such as Escherichia and Enterococcus.[28-31] Nonetheless, it should be noted that precise information on the bacteria obtained using molecular-based methods is required to accurately evaluate the effects of detected bacteria on the resulting behavioral changes.

In a recent series of studies, Bercik et al. demonstrated that anxiety-like behavior observed in Trichuris muris-infected mice is improved by administration of Bifidobacterium longum but not by Lactobacillus rhamnosus.[8, 32] In addition, Bravo and coworkers[33] also reported that JB-1 strain of L. rhamnosus reduces stress-induced anxiety- and depression-related behavior, accompanied by altered GABAAα2 receptor mRNA expression in the brain. Interestingly, surgical vagotomy failed to inhibit the B. longum-induced anxiolytic effects in the former system, while preventing the anxiolytic and antidepressant effects of L. rhamnosus (JB-1) in the latter system. The present results, in which Bc and Bi affected the anxiety and locomotor levels of the host in a reverse manner, suggest that the predominant pathways and the associated behavioral responses may be dependent on the nature of the bacterial species employed.

In the current study, there were no significant differences in DT30 between before and after exposure to the external environment; however, exposure significantly decreased TS30. NBM was lower in the ‘GF mice’ exposed to a conventional environment than in those kept under sterile conditions. The precise mechanisms whereby a short-time exposure to the external environment affects anxiety-like behaviors, but not locomotor activity, remain unclear; however, these results suggest that anxiety, rather than locomotion, may be more susceptible to a change in environment or exposure to conventional bacteria. In fact, Lyte and associates demonstrated that anxiety-like behaviors increase 7–8 h after a per-oral challenge with Citrobacter rodentium.[22] The alternative possibility that brain monoamine may be involved in these exposure-induced behavioral changes seems unlikely because a 24-h exposure of GF mice to a conventional environment did not affect the monoamines or their metabolites levels in the medial prefrontal cortex, striatum, hippocampus, or brainstem. Further studies are needed to clarify the molecules and pathways involved in this phenomenon.

In the OF test, the Bc mice exhibited decreases in TS 30, an indicator of anxiety-like behavior at 7 and 10 weeks, but not at 16 weeks, in comparison with that observed in the GF mice. On the other hand, the Bc mice exhibited significant decreases in NBM at all time points examined. The reason for the discrepant results between the OF and NBM at 16 weeks remains so far unclear. The discrepancy may be related to the fact that TS30 and NBM reflect different aspects of anxiety behavior.

The mice used in the current study that had not undergone behavioral analyses were killed in the isolator and subjected to sampling for the measurement of monoamines. Therefore, the results were free from the influences of stressors induced by both the tests and contamination by environmental microbes. This procedure is critically important to quantify a reliable basal level of monoamines because exposure to acute stress activates the monoaminergic system and elevates monoamine turnover rates.[34] As a result, NE, DA, and 5-HT turnover rates in EX-GF mice were lower in those in GF mice in most areas of the brain. These results are in a sharp contrast to those reported by Heijtza et al.,[13] in which these monoamine turnovers were higher in the striatum of GF mice than in that of SPF mice. Although the precise reason for this discrepancy remains unclear, it may be related to the different procedures used for sample collection or the different strains of animals that were used.

Accumulating evidence obtained by animal studies has shown an extensive role for the NE and DA in the modulation of executive function.[35] In addition, there is also an established dysfunction in the brain DA and NA systems in attention-deficit/hyperactivity disorder.[36] The present finding that GF mice showed increased locomotor activity concomitant with lower levels of DA and NA turnover rates in the brainstem and striatum in comparison with EX-GF mice, suggests that the absence or modulation of commensal microbiota might be involved in the pathogenesis of psychiatric disorder such as attention-deficit/hyperactivity disorder and autism.[37-40]

In this study, the sterility testing was carried out using culture-based methods. GF animals did not show any signs or symptoms suggesting contamination throughout the experiments; however, it should be noted that culture-based approaches are susceptible to false-negative results, as unculturable contaminants would not be detected. Therefore, further studies using not only culture-based methods but also PCR-based methods, should be performed in future studies to ensure the GF status.

In conclusion, the present results strongly support the current view that gut microorganisms modulate brain development and behavior. Given our previous work demonstrating gut microbes to exert protective effects against impaired HPA stress response, commensal bacteria may thus play a crucial role against developing stress-related disorders such as anxiety and depression, through providing the host with ‘stress resilience’[41, 42] necessary for adapting to a rapidly changing external environment. Clearly, further studies are called for, but these data will offer a firm foundation for this rapidly developing field of research and help to clarify the complex interactions and pathways involved in the ‘microbiota-gut-brain axis’.

Funding

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Funding
  8. Disclosures
  9. Author Contributions
  10. References

This work was partially supported by Research Fund of Tokai University School of Medicine (Y. Koga), Grants-in-Aid for General Scientific Research from the Ministry of Education, Culture, Sports, Science, and Technology in Japan (N. Sudo, No. 19390192, No. 22659144 and No. 24659350), and a research grant from the Yakult Bioscience Research Foundation (N. Sudo).

Disclosures

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Funding
  8. Disclosures
  9. Author Contributions
  10. References

All authors have no financial relationship with a commercial entity that has an interest in the subject of this manuscript.

Author Contributions

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Funding
  8. Disclosures
  9. Author Contributions
  10. References

RN had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis; RN, HT, KM, YK, and NS contributed to study concept and design; RN, HT, ST, MF, YA, and YK took care of the acquisition of data; RN, HT, KM, YK, and NS interpreted data and analysis; RN, KM, YK, and NS drafted the manuscript; HT, KM, ST, TH, and YA did critical revision of the manuscript for important intellectual content; RN, YK, and NS performed statistical analysis; YK and NS obtained funding; RN, HT, TH, YA, and YK provided administrative, technical, or material support. MF, YK, NS supervised the study.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Funding
  8. Disclosures
  9. Author Contributions
  10. References