Microbiota influence the development of the brain and behaviors in C57BL/6J mice

Abstract

We investigated the contributions of commensal bacteria to brain structural maturation by magnetic resonance imaging and behavioral tests in four and 12 weeks old C57BL/6J specific pathogen free (SPF) and germ free (GF) mice. SPF mice had increased volumes and fractional anisotropy in major gray and white matter areas and higher levels of myelination in total brain, major white and grey matter structures at either four or 12 weeks of age, demonstrating better brain maturation and organization. In open field test, SPF mice had better mobility and were less anxious than GF at four weeks. In Morris water maze, SPF mice demonstrated better spatial and learning memory than GF mice at 12 weeks. In fear conditioning, SPF mice had better contextual memory than GF mice at 12 weeks. In three chamber social test, SPF mice demonstrated better social novelty than GF mice at 12 weeks. Our data demonstrate numerous significant differences in morphological brain organization and behaviors between SPF and GF mice. This suggests that commensal bacteria are necessary for normal morphological development and maturation in the grey and white matter of the brain regions with implications for behavioral outcomes such as locomotion and cognitive functions.

Fig 1. MRI methods.

(A). Macromolecular proton fraction map of four weeks old mouse. Color bar indicates pseudo-color mapping of MPF in percent units. (B). High resolution T1-weighed volumes of ex-vivo mouse brain images with skulls in situ. A group of four brains were imaged at the same time. (C). Automatic structural parcellation of a mouse brain obtained from MT image using multi-atlas label fusion method.

Fig 2. Regional volume differences between SPF and GF Mice.

Regional volume differences in SPF and GF mice at four (n = 15 for both female and male SPF mice; n = 6 for female and n = 9 for male GF mice) and 12 (n = 13 for both female and male SPF and n = 7 for both female and male GF mice) weeks of age. Asterisks indicate significant differences when p-value is at least <0.05 between the treatment groups or between the individual groups in post-hoc comparisons where treatment group factor of two-way ANOVA for each testing age was significant.

Fig 3. Tensor based morphometry comparison between SPF and GF mice.

Animal numbers: SPF (n = 5 for both females and males) and GF (n = 5 for both females and males) mice at 12 weeks of age. No significant difference was detected at four weeks of age with the same number of animals in each group. The pseudo-colored statistical parametric map is overlaid on a gray scale template. Statistically significant regional volume expansion in olfactory bulbs and prefrontal cortex in GF group is indicated by pseudo-colored voxels in red-yellow scale. Color bar indicates corrected p-values range and p value of <0.05 was considered statistically significant.

Fig 4. Statistical parametric maps of comparisons between SPF and GF mice using FA and radial diffusivity metrics.

Animal numbers: SPF (n = 5 for both females and males) and GF (n = 5 for both females and males) mice at four weeks. No significant difference was detected at 12 weeks of age with the same number of animals in each group. Green skeleton lines were overlaid on gray scale FA, shown on sagittal, coronal and horizontal sections. Statistically significant voxels, where the parameters were less in the GF group, are depicted in red-yellow scale for FA. Color bar indicates corrected-p-values range and p value of <0.05 was considered statistically significant. Abbreviations of anatomic structures: ac-anterior commissure, cc-corpus callosum, ic-internal capsule, fi-fimbria, ot-optic tract, pwm-periventricular white matter.

Fig 5. Macromolecule proton fraction (in percentage) in SPF and GF mice.

Animal numbers at four (n = 15 for both female and male SPF mice; n = 6 for female and n = 9 for male GF mice) and 12 (n = 13 for both female and male SPF and n = 7 for both female and male GF mice) weeks of age. Asterisks indicate significant differences when p-value is at least <0.05 in between the groups in post-hoc comparisons where treatment group factor of two-way ANOVA for each testing age was significant.

Fig 6. Myelin content determination by Luxol fast blue stain.

Animal numbers at four (n = 6 for both SPF and GF mice) (A) and 12 (n = 6 for both SPF and GF mice) weeks of age (B). Asterisks indicate significant differences when p-value is at least <0.05 between SPF and GF group.

Fig 7. Open field and elevated plus maze behavioral tests.

SPF mice were more mobile (A) and traveled with faster speed (B) at four weeks (n = 15 for both female and male SPF mice; n = 6 for female and n = 9 for male GF mice), but not at 12 weeks (n = 13 for both female and male SPF and n = 7 for both female and male GF mice). SPF mice spent more time in the center quadrant in the open field test at four weeks, but not at 12 weeks (C). D. GF mice were not different from SPF mice at four weeks and at 12 weeks in time spent in closed arm in elevated plus maze. Asterisks indicate significant differences when p-value is at least <0.05.

Fig 8. Memory and learning behavioral tests using Morris water maze and fear conditioning.

No significant difference during training was found between SPF (n = 15 for both females and males) and GF (n = 6 for females and n = 9 for males) mice at 4 weeks (A), but at 12 weeks (B) SPF (n = 13 for both females and males) mice had significantly higher learning curve slopes than GF (n = 7 for both females and males) mice by RM ANOVA. (C). SPF mice spent significantly more time in the platform quadrant during the probe trial of the Morris water maze test at 12 weeks, but were not different from SPF mice at 4 weeks. (D). SPF males swim speed was significantly higher than GF males at 12 weeks. (E). Freezing time was longer in contextual fear conditioning test in male SPF mice at 12 weeks. (F). Freezing time was shorter for SPF mice in the second part of cued fear conditioning test (different chamber and sound cue presented) at 12 weeks. Asterisks indicate significant differences when p-value is at least <0.05.

Fig 9. Social interaction tests using three-room chamber.

In a sociability test SPF and GF mice were not different in time spent in chamber with a strange mouse (A) or with an empty cage (B) at either four (n = 15 for both female and male SPF mice; n = 6 for female and n = 9 for male GF mice) or 12 (n = 13 for both female and male SPF and n = 7 for both female and male GF mice) weeks of age. In the test for social novelty, no difference between SPF and GF groups was found in time spent with a familiar mouse (C), but there was a significant decrease in time spent with a novel mouse in GF mice at 12 weeks (D). Asterisks indicate significant differences when p-value is at least <0.05.

Fig 10. Stereological estimation of cell numbers between SPF and GF mice.

(A). Estimation of neuron number in hippocampal regions at 12 weeks (n = 6 for both SPF and GF mice) of age. (B). Neuronal density in motor cortex at four weeks. (C). Oligodendrocyte density in white and gray matter regions at 12 weeks of age. Asterisks indicate significant differences when p-value is at least <0.05.

Fig 11. Summary of differences in structures and behaviors between SPF and GF Mice.

Figure links the morphologic changes of regions in the brain with correlated behaviors.