Article InfoFigure 1
ATM treatment alters the composition of intestinal microbiota. (A) Representative DGGE gel of fecal microbiota from control and ATM-treated mice, before and during the treatment. Samples were pooled from each cage (5 mice per cage). (B) Mean DGGE profiles of cecal microbiota from control (n = 17) and ATM-treated (n = 20) mice. (C) Mean DGGE profiles of controls (n = 15) and mice at 2 weeks after ATM treatment (n = 19). (D) Detailed analysis of the microbiota by sequencing single excised DNA bands from DGGE gel from control and ATM-treated mice. Shades within the inner ring represent specific bacterial species. (E) Total number of cultivable bacteria using blood agar media. All data are mean ± standard error of the mean.
Figure 2
Oral ATM treatment alters mouse behavior promoting exploration. Results of step-down and light/dark preference tests in orally ATM-treated mice (n = 39), mice 2 weeks after ATM treatment (n = 19), and control mice (n = 47). BDNF protein levels measured by ELISA in hippocampus and amygdala of control (n = 17) and ATM-treated (n = 20) mice.
Figure 3
ATM effect on behavior is mediated by intestinal microbiota. (A) Effect of IP administration of ATMs (1% of daily oral dose for 7 days, n = 15) on mouse behavior compared with controls (n = 15). (B) Behavior in germ-free BALB/c (n = 7) before and after ATM administration, and after colonization with SPF BALB/c microbiota.
Figure 4
Oral ATMs do not induce gut inflammation or changes in enteric neurotransmitters, and do not act through the autonomic nervous system. (A) Microscopic scores and myeloperoxidase activity of small intestine and colon from control (n = 15) and ATM-treated (n = 19) mice. (B) Levels of serotonin, dopamine, and noradrenalin as assessed by ELISA in control (n = 15) and ATM-treated (n = 19) mice, both in the small intestine and colon. (C) Behavior in controls (n = 15), ATM-treated mice (n = 15), ATM-treated mice with pyloroplasty alone (n = 15), ATM-treated mice with vagotomy and pyloroplasty (n = 15), mice with pyloroplasty only (n = 9), and vagotomy only (n = 12). (D) Behavior in controls (n = 15), ATM-treated mice (n = 15), ATM-treated mice with sympathectomy (n = 15), and mice with sympathectomy only (n = 15). Data are mean ± standard error of the mean, statistics by ANOVA and the Tukey test.
Figure 5
Intestinal microbiota transfer differentially affects behavior of recipient mice. (A) Representative DGGE gel of fecal microbiota from SPF BALB/c and NIH Swiss mice (pooled samples, 5 mice per cage). (B) Mean DGGE profiles of cecal microbiota from SPF BALB/c and NIH Swiss mice (n = 8 per group). (C) Step-down test in NIH Swiss and BALB/c mice at 3 weeks after colonization. SPF NIH Swiss (n = 22), and germ-free (GF) NIH Swiss mice colonized with either NIH Swiss (n = 43) or BALB/c (n = 41) microbiota (left panel). SPF BALB/c (n = 15), or GF BALB/c mice colonized with either BALB/c (n = 15) or NIH Swiss (n = 15) microbiota (right panel). (D) Step-down test in NIH Swiss mice colonized for 1 week with either BALB/c (n = 12) or NIH Swiss (n = 11) microbiota. BDNF levels in hippocampus and amygdala in NIH Swiss mice colonized with BALB/c (n = 7) or NIH Swiss (n = 7) microbiota for 1 week. Data are mean ± standard error of the mean, statistics by ANOVA and the Tukey test.
Figure 6
Intestinal microbiota does not affect circulating cytokines or gut neurotransmitters. (A) Levels of serum tumor necrosis factor-α, interferon-γ, and IL-1β in germ-free recipients who received either NIH Swiss (n = 21) or BALB/c (n = 20) microbiota. Data are mean ± standard error of the mean, statistics by t test. (B) Serotonin and dopamine levels in both small intestine and colon in mice colonized with NIH Swiss or BALB/c microbiota (n = 6 per group). Data are mean ± standard error of the mean, statistics by t test. *P < .05 vs BALB/c mice.
Supplementary Figure 1
IP administered ATMs do not affect intestinal microbiota. (A) Representative DGGE gels of cecal microbiota from individual mice treated with saline or ATMs IP. (B) Mean DGGE profiles of cecal microbiota from IP saline (n = 15) and IP ATM-treated (n = 15) mice.
Supplementary Figure 2
Intestinal microbiota does not affect brain BDNF at 3 weeks after colonization. BDNF levels in amygdala and hippocampus after 3-week colonization with NIH Swiss or BALB/c microbiota (n = 6 per group).
Background & Aims
Alterations in the microbial composition of the gastrointestinal tract (dysbiosis) are believed to contribute to inflammatory and functional bowel disorders and psychiatric comorbidities. We examined whether the intestinal microbiota affects behavior and brain biochemistry in mice.
Methods
Specific pathogen–free (SPF) BALB/c mice, with or without subdiaphragmatic vagotomy or chemical sympathectomy, or germ-free BALB/c mice received a mixture of nonabsorbable antimicrobials (neomycin, bacitracin, and pimaricin) in their drinking water for 7 days. Germ-free BALB/c and NIH Swiss mice were colonized with microbiota from SPF NIH Swiss or BALB/c mice. Behavior was evaluated using step-down and light preference tests. Gastrointestinal microbiota were assessed using denaturing gradient gel electrophoresis and sequencing. Gut samples were analyzed by histologic, myeloperoxidase, and cytokine analyses; levels of serotonin, noradrenaline, dopamine, and brain-derived neurotropic factor (BDNF) were assessed by enzyme-linked immunosorbent assay.
Results
Administration of oral antimicrobials to SPF mice transiently altered the composition of the microbiota and increased exploratory behavior and hippocampal expression of BDNF. These changes were independent of inflammatory activity, changes in levels of gastrointestinal neurotransmitters, and vagal or sympathetic integrity. Intraperitoneal administration of antimicrobials to SPF mice or oral administration to germ-free mice did not affect behavior. Colonization of germ-free BALB/c mice with microbiota from NIH Swiss mice increased exploratory behavior and hippocampal levels of BDNF, whereas colonization of germ-free NIH Swiss mice with BALB/c microbiota reduced exploratory behavior.
Conclusions
The intestinal microbiota influences brain chemistry and behavior independently of the autonomic nervous system, gastrointestinal-specific neurotransmitters, or inflammation. Intestinal dysbiosis might contribute to psychiatric disorders in patients with bowel disorders.
Keywords:
Host–Bacterial Interactions, Gut–Brain Axis, Commensal Bacteria, Inflammatory Bowel DiseaseAbbreviations used in this paper:
ATM (antimicrobial), BDNF (brain-derived neurotropic factor), DGGE (denaturing gradient gel electrophoresis), ELISA (enzyme-linked immunosorbent assay), IBS (irritable bowel syndrome), IL (interleukin), SPF (specific pathogen-free)Conflicts of interest These authors disclose the following: S. M. Collins, P. Bercik, and E. F. Verdu received grant support from Nestle Switzerland. The remaining authors disclose no conflicts.
Funding Supported by Canadian Institutes for Health Research and Crohn's and Colitis Foundation of Canada grants (S.M.C., P.B.); E. F. Verdu and Premysl Bercik hold Internal Career Research Awards from the Department of Medicine at McMaster University.
