Article Figures and data Abstract eLife digest Introduction Results Discussion Materials and methods Data availability References Decision letter Author response Article and author information Metrics Abstract Gene-environment interactions impact the development of neuropsychiatric disorders, but the relative contributions are unclear. Here, we identify gut microbiota as sufficient to induce depressive-like behaviors in genetically distinct mouse strains. Daily gavage of vehicle (dH2O) in nonobese diabetic (NOD) mice induced a social avoidance behavior that was not observed in C57BL/6 mice. This was not observed in NOD animals with depleted microbiota via oral administration of antibiotics. Transfer of intestinal microbiota, including members of the Clostridiales, Lachnospiraceae and Ruminococcaceae, from vehicle-gavaged NOD donors to microbiota-depleted C57BL/6 recipients was sufficient to induce social avoidance and change gene expression and myelination in the prefrontal cortex. Metabolomic analysis identified increased cresol levels in these mice, and exposure of cultured oligodendrocytes to this metabolite prevented myelin gene expression and differentiation. Our results thus demonstrate that the gut microbiota modifies the synthesis of key metabolites affecting gene expression in the prefrontal cortex, thereby modulating social behavior. https://doi.org/10.7554/eLife.13442.001 eLife digest A combination of genes and environmental factors underlie an individual’s risk of developing a mental illness. Among the environmental factors, it is becoming clear that communication between the gut and the brain is involved, but we do not understand how these two organs communicate. Our gut contains a variety of bacteria that help us to digest food and there is some evidence that changes in these bacterial communities can influence our mental health. Transplanting feces from one individual to the gut of another is one way to alter the communities of bacteria in the gut. Here, Gacias et al. investigated whether fecal transplants are sufficient to induce social avoidance behavior – a symptom of depression – in mice. The experiments show that introducing specific combinations of bacteria into the gut is indeed able to cause healthy adult mice to avoid social interactions. This effect was caused by changes in the “myelin” sheath that surrounds many nerve fibers and could be prevented by giving the mice antibiotics, which decreased the number of bacteria in the gut. Further experiments revealed that the mice that became depressed after fecal transplants had higher levels of a molecule called cresol, which is produced by certain gut bacteria. Gacias et al. found that cresol is able to reduce the amount of myelin produced by brain cells. Therefore, these findings show that changing the communities of bacteria in the gut can result in the accumulation of molecules that influence social behavior. Future work will aim to identify bacteria that can reduce the amount of cresol produced in the gut, which may have the potential to treat depression. https://doi.org/10.7554/eLife.13442.002 Introduction Despite the diffuse prevalence of mental illness and the large efforts spent in identifying genetic elements of susceptibility, there is a need to define the role of environment—gene interactions. In addition to genetic predisposition, there is extensive epidemiologic literature emphasizing the role of environmental exposure in the development of mild to severe mood disorders. The aftermath of traumatic life events, for instance, is often characterized by the onset of severe depression or post-traumatic stress disorder (Shalev et al., 1998). The interplay between genes and environmental variables has gained recent attention, and several immunologic and lifestyle contributors have been proposed to modulate depressive symptoms. The detection of high levels of serum cytokines and the higher incidence of depression in patients with autoimmune disorders (Postal and Appenzeller, 2015; Walker et al., 2011; Moll et al., 2011; van Hees et al., 2015; Feinstein et al., 2014) has suggested a role for neuroinflammation (Godbout et al., 2008; Menard et al., 2016; Audet et al., 2014). Deficiency of specific nutrients such as omega-3 fatty acids has been reported in subsets of patients with mental illnesses (Ohara, 2005; Patrick and Ames, 2015; Poudel-Tandukar et al., 2009; Panagiotakos et al., 2010), highlighting the link between mood disorders and the bioavailability of metabolites. There is evidence that bioactive metabolites act as mediators of gut—brain communication, as shifts in gut microbial composition impact brain neurochemistry (Cryan and Dinan, 2012; Collins et al., 2012; Desbonnet et al., 2014; Bercik et al., 2010; 2011). Indeed, psychiatric comorbidities often accompany conditions characterized by an aberrant gut microbiota composition, such as irritable bowel syndrome, functional gastrointestinal disorder, and inflammatory bowel disease (Gevers et al., 2014; Morgan et al., 2012; Haberman et al., 2014; Carroll et al., 2011; Addolorato et al., 1997). Conversely, altered gut microbiota composition and function have been reported in patients with major depressive disorders and children with autism (Jiang et al., 2015; De Angelis et al., 2015; De Angelis et al., 2013; Parracho et al., 2005). The gut microbiota is a complex microbial ecosystem that rapidly responds to environmental changes and can modulate brain development, function, and behavior (Cryan and Dinan, 2012; Collins et al., 2012; Desbonnet et al., 2014; Bercik et al., 2010; 2011; Wu et al., 2011; Daniel et al., 2014; Lax et al., 2014). These studies suggest that social behavior may be affected by abnormal interactions between gut microbiota and the brain, though the underlying mechanisms remain only partially understood. One hypothesis for the pathogenesis of depressive-like behaviors has been suggested through studies on social isolation in mice (Liu et al., 2012; 2016; Makinodan et al., 2012), which revealed a reduction of myelinated fibers in the prefrontal cortex (PFC), associated with changes in the oligodendrocyte transcriptome (Liu et al., 2012; 2016). Myelination is a dynamic process that continues into adulthood and contributes to physiologic brain function (Liu et al., 2012; 2016; Makinodan et al., 2012; Sánchez et al., 1998; Gibson et al., 2014; McKenzie et al., 2014). Oligodendrocytes are the myelinating cells of the central nervous system (CNS), and neuropathologic and transcriptomic studies have reported downregulated oligodendroglial transcripts and reduced myelin thickness in the brains of patients with schizophrenia, major depression, and bipolar disorder (Tkachev et al., 2003; Aston et al., 2005; Katsel et al., 2005). These data highlight the role of myelin in mental illness and depressive-like behaviors, though it remains to be established whether myelination in the adult PFC and social behavior are affected by alterations in gut microbiota composition. This study characterizes the gut microbiota in mice with social avoidance behavior and demonstrates that transfer of specific bacterial taxa is sufficient to alter adult PFC myelination and results in behavioral changes consistent with a depressive-like phenotype. Results Non-obese diabetic (NOD) and C57BL/6 mice display differential susceptibility to develop depressive-like symptoms in response to daily gavage Although gastric gavage and subcutaneous injections are routine, daily procedures used to administer drugs or special diet to rodents, the potential behavioral effects they may induce in mice have not been investigated. Daily gastric gavage with vehicle for two weeks (Figure 1A) was sufficient to induce social avoidance behavior in NOD mice (Figure 1B), without affecting their overall locomotor activity (Figure 1C). This depressive-like behavior induced by daily gavage was dependent on the specific mouse strain, as C57BL/6 mice were not affected (Figure 1D,E) (Moy et al., 2008). Subcutaneous injection of vehicle did not elicit any behavioral effect in either strain (Figure 1—figure supplement 1). Daily gastric gavage with an antibiotic cocktail proven to deplete the gut microbiota (Reikvam et al., 2011) failed to induce the social avoidance behavior in NOD mice (Figure 1B), and similarly had no effect on the C57BL/6 mice (Figure 1D). The antibiotic regimen was well tolerated by both NOD and C57BL/6 mice, did not impact body weight or glucose levels, and did not result in any gastric hemorrhage or visible stomach damage (Figure 1—figure supplement 2 and Figure 1—figure supplement 3). Consistent with previous reports, only chronic oral antibiotic treatment (but not subcutaneous delivery) induced enlargement of the large intestine (Figure 1—figure supplement 2), a macroscopic sign associated with microbiota depletion (Reikvam et al., 2011). Interestingly, daily gavage also induced an anxiety-like behavior in both NOD and C57BL/6 mice, as revealed by the elevated plus maze (EPM) (Figure 2B,D). However, the anxiety-like behavioral change displayed in response to daily gavage was not affected by oral antibiotic treatment (Figure 2B,D), suggesting that only the depressive-like behavior is mediated by alterations in gut microbiota. To further validate this hypothesis, we conducted the forced swim test (FST), which is considered a measure of despair-like behavior, in NOD and C57BL/6 mice after daily gavage with either vehicle or antibiotics. The despair-like behavior was induced by vehicle gavage in the NOD strain, and was prevented by oral antibiotic treatment (Figure 2C), but was not detected in the C57BL/6 mice (Figure 2E). Together, these results indicate that daily gavage of vehicle induces social avoidance and despair-like behaviors in NOD mice, but not in C57BL/6 mice, and that this effect is not observed when gavaging antibiotics orally and not subcutaneously. Figure 1 with 3 supplements see all Download asset Open asset The strain-specific social avoidance behavioral response to daily gavage is affected by oral antibiotic treatment. (A) Experimental timeline: vehicle or antibiotic mix were administered daily by gastric gavage (g.g.) for 14 days. Behavioral testing was performed before (baseline) and after treatment. (B–D) Results of the Social Interaction (SI) test for NOD (B) and C57BL/6 (D) mice. Oral antibiotic treatment did not affect locomotor activity measured during the social interaction test (C,E) (3 independent experiments with 8 mice per group/experiment for a total of n=23–24 mice per condition). Data are mean ± S.E.M; *p<0.05, **p<0.01 based on one-way ANOVA with Bonferroni’s post hoc test; n.s. indicates not significant. https://doi.org/10.7554/eLife.13442.003 Figure 2 with 1 supplement see all Download asset Open asset The strain-specific anxiety- and despair-like behavioral responses to daily gavage are differentially affected by oral antibiotic treatment. (A) Experimental timeline: vehicle or antibiotic mix were administered daily by gastric gavage (g.g.) for 14 days. Behavioral testing was performed before (baseline) and after treatment. Figure shows the results for the Elevated Plus Maze (EPM) and Forced Swim Test (FST) for NOD (B, C) and C57BL/6 (D, E) mice after oral treatment (g.g.). Baseline measurements for FST were performed in a separate cohort of mice (n=10) to avoid carryover effects (3 independent experiments with 8 mice per group/experiment for a total of n=24 mice per condition). Data are mean ± S.E.M; *p<0.05, **p<0.01, ***p<0.001 based on one-way ANOVA followed by Bonferroni’s post hoc test; n.s. indicates not significant. https://doi.org/10.7554/eLife.13442.007 Behavioral differences in genetically different mice are associated with specific gut microbiota composition To further characterize the effect of vehicle and antibiotic treatment on gut microbiota composition, we conducted 16S rRNA sequencing analysis of cecal and fecal samples collected after behavioral testing and after 14 days of treatment (Figure 3A). Unweighted UniFrac distances (Lozupone and Knight, 2005) were calculated between all pairs of fecal samples based on their microbiota composition. Based on these distances, Principal coordinate analysis (PCoA), an ordination method conceptually similar to principal component analysis, revealed a clear separation between vehicle-gavaged and baseline NOD (Figure 3) and between vehicle and antibiotic treated NOD and C57BL/6 mice (Figure 3—figure supplement 1A). PCoA analysis revealed clear differences between NOD mice before (“baseline”) and after oral treatment with antibiotics (Figure 3B), with differences also observed between samples before and after treatment with vehicle (Figure 3C). Since the depressive-like behavior was only observed in oral vehicle-treated NOD mice, we focused on identifying the specific microbiota that differ in these animals before and after treatment. Analysis of Operational Taxonomic Units (OTUs, defined as groups of 16S rRNA gene sequences with high similarity and that broadly correspond to a bacterial species) identified several taxonomic groups that were exclusively found in the vehicle-treated mice (Figure 3D and Gacias et al., 2016). These taxa represent potential candidates associated with the depressive-like phenotype observed in NOD mice. Linear discriminant analysis effect size (LEfSe) (Segata et al., 2011), a biomarker discovery method based on the Kruskal–Wallis and Wilcoxon tests, was used to identify key bacterial taxa enriched in vehicle-treated versus antibiotic-treated animals in each strain (Figure 3—figure supplement 1B–E). As expected, Proteobacteria were enriched in antibiotic-treated animals, while vehicle-treated mice had enrichment in Bacteroidetes and Firmicutes (Figure 3—figure supplement 1C,E). Figure 3 with 1 supplement see all Download asset Open asset Enrichment of bacterial OTUs induced by gastric gavage (g.g.) in NOD mice. (A) Experimental timeline indicating time points of fecal collection (arrows) relative to behavioral testing and treatment. (B,C) Principal coordinate analysis plots of unweighted UniFrac distances of microbiota in fecal samples at baseline and after 14 days of daily g.g. of antibiotics or vehicle in NOD mice. Each dot represents the microbiota of a sample (1 sample = feces pooled from 3–5 mice), color-coded by treatment (vehicle or antibiotic) and time-point. The percentage of variation explained by each principal coordinate (PC) is shown in parentheses. All samples were rarefied at 5000 sequences. (D) Analysis of unique Operational Taxonomic Units (OTUs) present in NOD vehicle-treated mice compared to their fecal microbiota at baseline. Figure shows representative taxa enriched in fecal samples of NOD vehicle-treated mice compared to their baseline samples. Each bar represents the microbiota of an individual sample (1 sample = 3–5 mice per cage). See Gacias et al. (2016). https://doi.org/10.7554/eLife.13442.009 Analysis of tissue samples revealed similar differences between vehicle- and antibiotic-treated mice in both strains (p<0.01, adonis with 999 permutations). No significant changes in gut microbiota composition were detected when antibiotics were administered subcutaneously. Modification of the gut microbiota following oral antibiotic administration induces unique changes in the medial prefrontal cortex adult myelination of NOD mice To identify possible CNS transcriptional signatures associated with the behavioral outcomes described in the vehicle-gavaged, but not antibiotic-gavaged, NOD mice, we performed an unbiased transcriptomic analysis of the medial prefrontal cortex (mPFC) using RNA sequencing. This analysis revealed decreased expression of genes related to myelination (Figure 4 and Figure 4—figure supplements 1,2 and Gacias et al., 2016) in vehicle-gavaged NOD mice - characterized by social avoidance behavior - compared to antibiotic-treated mice, whose behavior was comparable to baseline controls (Figure 1). The differences in myelin gene transcripts in the mPFC of vehicle-gavaged NOD compared to antibiotic-treated mice were validated by quantitative real-time qPCR (Figure 4B) and immunohistochemistry (Figure 4C). These differences were detected only in NOD mice, and not in C57BL/6 mice that showed no change in social behavior with oral gavage (Figure 4D,E). The differences in myelin gene expression in the mPFC could not be attributed to a nonspecific effect of antibiotic treatment, as there were no differences observed after subcutaneous delivery (Figure 4—figure supplement 1). The regional specificity of the transcriptional changes was also assessed in NOD mice by evaluating samples from a distinct brain region, the nucleus accumbens (NAc), revealing no difference in the two treatment groups (Figure 4—figure supplement 1). These data provide further support for the relationship between defective mPFC adult myelination and depressive-like behavior, as indicated by the lower levels of myelin transcripts and reduced area of MBP immunostaining in vehicle-gavaged NOD mice exhibiting social avoidance. The results also demonstrate that the transcriptional and behavioral effects were prevented by oral antibiotic treatment. Figure 4 with 2 supplements see all Download asset Open asset Myelin transcripts and myelinated fibers in the medial prefrontal cortex (mPFC) of adult NOD mice with social avoidance behavior. (A) Experimental timeline: vehicle or antibiotic mix were administered daily by gastric gavage (g.g.) for 14 days and mPFC was harvested for RNA extraction and quantitative real-time qPCR or immunohistochemsitry (B,D) qPCR of myelin transcripts after vehicle or antibiotic treatment of NOD (B) and C57BL/6 (D) mice. Values were normalized to 36b4 mRNA levels and are referred as fold change relative to vehicle-treated values (n=6 mice per group). (C,E) Representative confocal images and quantification of MBP+ fibers (red) in mPFC of NOD (C) and C57BL/6 (E) mice after vehicle or antibiotic treatment. DAPI (blue) was used as nuclear counterstain. Scale bar, 50 μm. Graph represents quantification of MBP+ fibers per surface area (n=3 for NOD; n=4 for C57BL/6). Data are mean ± S.E.M; **p<0.01, ***p<0.001 based on unpaired t test. n.s. indicates not significant. https://doi.org/10.7554/eLife.13442.011 Transplantation of fecal microbiota from vehicle-treated NOD mice to depleted C57BL/6 recipients is sufficient to recolonize the gut and transfer transcriptional, and behavioral traits To determine whether the social avoidance behavior and mPFC transcriptional changes induced by daily gavage of vehicle in NOD mice were caused by the enrichment of specific gut bacteria, we transferred the cecal content of vehicle-treated or antibiotic-treated NOD mice into C57BL/6 recipients, whose endogenous flora had been depleted by antibiotic treatment (Figure 5A). Social behavior in C57BL/6 depleted recipients was assessed before and after transplantation with microbiota from either vehicle-gavaged (Group I) or antibiotic-gavaged (Group II) NOD donors. The behavior of the C57BL/6 recipients resembled that of the donors: Social avoidance behavior was detected in Group I recipients, and was not observed in Group II recipients (Figure 5B,C). Intriguingly, transplantation of vehicle-gavaged NOD microbiota also transferred the transcriptional changes in the mPFC, but not in the NAc, as shown by the lower levels of myelin gene transcripts (Mag, Mog, Plp1, Mobp) in Group I mice compared to Group II recipients (Figure 5D). The functional consequences of the transcriptional changes in myelin genes were further validated by electron microscopy, and ultrastructural analysis revealed decreased myelin thickness in Group I recipients displaying a social avoidance behavior (Figure 5E). Quantification of myelin thickness relative to axonal diameter (g ratio) revealed that Group I recipients transplanted with vehicle-gavaged NOD microbiota, presented thinner myelin than Group II, recipients of antibiotic-treated NOD donors. No significant differences between the two groups were observed in the NAc (Figure 5E). The transfer of depressive-like behavior from donor to recipient was further validated by the detection of increased immobility at the FST in Group I mice compared to Group II (Figure 5—figure supplement 1). Figure 5 with 1 supplement see all Download asset Open asset Social avoidance behavior transfer from NOD donors to microbiota depleted C57BL/6 by fecal transplantation. (A) Experimental timeline for donor (NOD) and transplant-recipient (C57BL/6) mice. (B,C) Results from Social Interaction (SI) tests conducted in C57BL/6 recipients before and after transplantation with either microbiota from vehicle-treated (Group I; B) or antibiotic-treated (Group II, C) NOD mice. Graphs represent the amount of time spent (seconds) in the interaction zone when a target is present. Red dashed bar represents the interaction time of the NOD donors. Data are mean ± S.E.M; *p<0.05, **p<0.01 based on a two-way ANOVA (n=12 mice per experiment, 2 replicates of 12 for a total of 24 mice per condition). (D) Graphs indicate the relative levels of myelin gene transcripts in mPFC and NAc of C57BL/6 recipients displaying (Group I) or not displaying (Group II) social avoidance behavior after transplantation with NOD microbiota (n=6–8 mice per group; *p<0.05, **p<0.01, ***p<0.001 based on unpaired t test). (E) Electron micrographs and quantified g-ratios of myelinated axons in mPFC and NAc in Group I and Group II C57BL/6 recipients after transplantation with the NOD microbiota. Scale bar, 1 μm. (n=3 per treatment and condition; statistical differences between groups were determined using two-tailed t-test; n.s. indicates not significant). https://doi.org/10.7554/eLife.13442.014 Collectively, these findings suggest that the gut microbiota of vehicle-gavaged NOD donors was sufficient to transfer the depressive-like behavior, modulate transcript levels in the mPFC, and impact region-specific adult myelination in microbiota-depleted C57BL/6 recipients. The genomic DNA content was measured in fecal pellets of C57BL/6 recipients to validate the depletion of the gut microbiota with 14 days of antibiotic treatment, and to evaluate the effectiveness of recolonization after transplantation (Figure 6B,D). Analysis of alpha diversity (the number of bacterial taxa present in a sample or group of samples) further confirmed the microbiota depletion (Figure 6C,E). In both groups, diversity was significantly reduced from baseline after antibiotic treatment (Figure 6C,E; p<0.01 ANOVA with Tukey’s honest significant difference (HSD) post-hoc analysis). As expected, after transplantation Group II mice still exhibited a significantly depleted diversity compared to baseline (Figure 6E; p<0.01 ANOVA with Tukey’s HSD), while bacterial diversity in Group I had recovered to levels similar to baseline and was not significantly different (Figure 6C; p=0.09, ANOVA with Tukey’s HSD). These results suggested that transfer of behavioral traits was associated with restoration of bacterial diversity to baseline levels. In order to determine the differences in microbiota compositions associated with the behavioral phenotype, we conducted PCoA analysis based on unweighted UniFrac analysis (Figure 6F). Although all pooled fecal samples from NOD donors and C57BL/6 recipients clustered together at baseline (samples on the right side of the plot), treatment with antibiotics resulted in a drastic reshaping of the bacterial communities of both NOD (middle of the plot) and C57BL/6 (bottom-left side) mice. The microbiota composition of Group II mice after transplant (which did not display social avoidance behavior) was distinct from baseline, similar to antibiotic-treated animals pre-transplant (top-left side). However, Group I recipients which displayed social avoidance behavior (#19 and #18 on the plot), had compositions that were close to those of their vehicle-treated NOD donors. In contrast, Group I recipients which did not display social avoidance behavior (#17 on the plot), clustered with Group II recipients. This result suggests that the transplant procedure was not equally effective in all Group I mice. The distance in microbiota composition between vehicle-gavaged donors and recipients was significantly correlated with the social avoidance behavior, as measured by social interaction time (Figure 6G; p=0.01). This result suggests that the ability to successfully transfer the gut microbiota from vehicle-gavaged NOD donors was significantly correlated with the transmission of the depressive-like behavior. LEfSe analysis revealed a number of taxa that were significantly different between Group I and Group II C57BL/6 recipients (Figure 6—figure supplement 1 and Gacias et al., 2016). We further refined this analysis, by identifying the specific OTUs transferred from vehicle-gavaged NOD donors to Group I recipients (Figure 6—figure supplement 2 and Gacias et al., 2016). Members of the Clostridiales order, including Lachnospiraceae and Ruminococcaceae, were among those present in equal proportions both in the donors and the recipients in Group I recipients displaying a depressive-like behavior (i.e. samples #18 and #19), while absent in Group I recipients that did not exhibit such behavior (i.e. sample#17; Figure 6F and Gacias et al., 2016). We further confirmed these taxa as potentially responsible for this phenotype by qPCR using primers specific to these bacterial groups (Figure 6—figure supplement 2C). In order to identify differences undetectable at the OTU level, we performed oligotype analysis in those OTUs established as potentially responsible for the depressive-like behavior (Segata et al., 2011; Eren et al., 2014). Oligotype analysis is an entropy-based method to identify single nucleotide differences in sequences from closely related organisms. We found that most OTUs were composed of a single high-abundance oligotype (Figure 6—figure supplement 3A–C,E,G–O) and therefore support the conclusions from the OTU-level analysis. However, we identified three OTUs that had two oligotypes with similar abundances and distribution across samples: OTU 183849 (Blautia producta, a member of the Lachnospiraceae, Figure 6—figure supplement 3D), 188840 (unidentified member within Lachnospiraceae, Figure 6—figure supplement 3F), and 4418586 (unidentified member within Clostridiales, Figure 6—figure supplement 3P). Additional inspection of these sequences revealed the oligotypes GTT and TTT from the Blautia producta OTU, as well as the TG and TT oligotypes from the Lachnospiraceae OTU, had B. producta JCM 1471 as the closest reference sequence in NCBI; the oligotypes from Clostridiales had no close reference sequence. Overall, these results show that either a single oligotype or a combination of two oligotypes with similar abundance distributions were dominant within the analyzed OTUs, which suggested they might drive the observed social phenotypes. Figure 6 with 3 supplements see all Download asset Open asset Effect of fecal transplantation on bacterial mass and biodiversity in microbiota depleted C57BL/6 recipients. (A) Experimental timeline for donors (NOD) and transplanted recipients (C57BL/6). (B,C) Graphs represent fecal biomass (µg of gDNA relative to total fecal weight) of C57BL/6 recipients prior to transplantation (#1 before and #2 after 14 days of antibiotic treatment) and at end point after-transplantation (#3) with donor microbiota (n=3 pooled samples per time-point, each sample represents 1 sample = pooled feces from 3–5 mice. Data are mean ± S.E.M; *p<0.05, **p<0.01 based on one-way ANOVA with Bonferroni’s post hoc test). (C,E) Rarefaction curves comparing alpha diversity of fecal microbiota samples from C57BL/6 recipients at different experimental time-points (#1, #2, and #3). (F) Principal coordinate analysis plot of unweighted UniFrac distances of fecal samples from NOD donors and C57BL/6 mice at different time-points. (#1, #2, and #3). Each dot represents the microbiota of a sample, colored by group, treatment, and time-point (n=3 pooled samples per time-point; each sample corresponds to pooled feces from 3–5 mice). The percentage of variation explained by each principal coordinate (PC) is shown in parentheses. (E) Relationship between social interaction time and unweighted UniFrac distance to NOD donor mice (n=3) for all C57BL/6 recipients (n=10). Each point represents a single C57BL/6 animal, colored by group (light blue: Group_I, transplanted with NOD-vehicle microbiota; pink: Group_II, transplanted with NOD-antibiotic microbiota). Linear regression analysis indicates a significant correlation (p=0.0103) between the variables. https://doi.org/10.7554/eLife.13442.016 The gut metabolome is altered in microbiota-transplanted C57BL/6 mice displaying altered social and despair-like behaviors. Several studies have demonstrated that gut metabolites can impact the homeostatic host-microbiota interactions and affect behavior (Daniel et al., 2014; Hsiao et al., 2013). To determine whether altered taxa in the gut microbiota could impact the levels of metabolites, which in turn drive behavioral and transcriptional changes observed in the mPFC, we performed an unbiased metabolomic analysis of gut tissue from C57BL/6 recipients with (Group I) and without (Group II) social av
Abstract The N‐myc downstream regulated gene family member 1 ( NDRG1 ) is a gene whose mutation results in peripheral neuropathy with central manifestations. While most of previous studies characterized NDRG1 role in Schwann cells, the detection of central nervous system symptoms and the identification of NDRG1 as a gene silenced in the white matter of multiple sclerosis brains raise the question regarding its role in oligodendrocytes. Here, we show that NDRG1 is enriched in oligodendrocytes and myelin preparations, and we characterize its expression using a novel reporter mouse ( TgNdrg1‐EGFP ). We report NDRG1 expression during developmental myelination and during remyelination after cuprizone‐induced demyelination of the adult corpus callosum. The transcriptome of Ndrg1‐EGFP+ cells further supports the identification of late myelinating oligodendrocytes, characterized by expression of genes regulating lipid metabolism and bioenergetics. We also generate a lineage specific conditional knockout ( Olig1 cre/+ ;Ndrg1 fl/fl ) line to study its function. Null mice develop normally, and despite similar numbers of progenitor cells as wild type, they have fewer mature oligodendrocytes and lower levels of myelin proteins than controls, thereby suggesting NDRG1 as important for the maintenance of late myelinating oligodendrocytes. In addition, when control and Ndrg1 null mice are subject to cuprizone‐induced demyelination, we observe a higher degree of demyelination in the mutants. Together these data identify NDRG1 as an important molecule for adult myelinating oligodendrocytes, whose decreased levels in the normal appearing white matter of human MS brains may result in greater susceptibility of myelin to damage.
The interplay between genetic factors and cell-specific epigenetic changes may be highly relevant in development of multiple sclerosis (MS). Environmental risk factors for MS are able to modify the epigenome, constituting a link between environment exposure and changes in gene expression. In this review we discuss the most relevant epigenetic findings described in different tissues from MS patients and the future application of epigenetic advances in MS field.
Significance We have experimentally investigated the immunoregulatory effects of human gut microbiota in multiple sclerosis (MS). We have identified specific bacteria that are associated with MS and demonstrated that these bacteria regulate T lymphocyte-mediated adaptive immune responses and contribute to the proinflammatory environment in vitro and in vivo. Thus, our results expand the knowledge of the microbial regulation of immunity and may provide a basis for the development of microbiome-based therapeutics in autoimmune diseases.
Ceramide C16 is a sphingolipid detected at high levels in several neurodegenerative disorders, including multiple sclerosis (MS). It can be generated de novo or from the hydrolysis of other sphingolipids, such as sphingomyelin or through the recycling of sphingosine, in what is known as the salvage pathway. While the myelin damage occurring in MS suggests the importance of the hydrolytic and salvage pathways, the growing interest on the importance of diet in demyelinating disorders, prompted us to investigate the involvement of de novo ceramide C16 synthesis on disease severity. A diet rich in saturated fats such as palmitic acid, as found in many highly processed foods, provides substrates for the ceramide C16 synthetic enzymes ceramide synthase 6 (CERS6) and 5 (CERS5), which are expressed in the central nervous system. Using the experimental autoimmune encephalomyelitis (EAE) model of inflammatory demyelination, we show here that mice with CamK2a+ neuronal specific deletion of both CerS6 and CerS5 show a milder course of EAE than wild type mice, even when fed a diet enriched in palmitic acid. At a cellular level, neurons lacking both CerS6 and CerS5 are protected from the mitochondrial dysfunction arising from exposure to oxidative stress and palmitic acid in the medium. These data underscore the importance of a healthy diet avoiding processed foods for demyelinating disorders and identifies endogenous neuronal synthesis of ceramide C16 as an important determinant of disease severity.
Altered myelin structure and oligodendrocyte function have been shown to correlate with cognitive and motor dysfunction and deficits in social behavior. We and others have previously demonstrated that social isolation in mice induced behavioral, transcriptional, and ultrastructural changes in oligodendrocytes of the prefrontal cortex (PFC). However, whether enhancing myelination and oligodendrocyte differentiation could be beneficial in reversing such changes remains unexplored. To test this hypothesis, we orally administered clemastine, an antimuscarinic compound that has been shown to enhance oligodendrocyte differentiation and myelination in vitro , for 2 weeks in adult mice following social isolation. Clemastine successfully reversed social avoidance behavior in mice undergoing prolonged social isolation. Impaired myelination was rescued by oral clemastine treatment, and was associated with enhanced oligodendrocyte progenitor differentiation and epigenetic changes. Clemastine induced higher levels of repressive histone methylation (H3K9me3), a marker for heterochromatin, in oligodendrocytes, but not neurons, of the PFC. This was consistent with the capability of clemastine in elevating H3K9 histone methyltransferases activity in cultured primary mouse oligodendrocytes, an effect that could be antagonized by cotreatment with muscarine. Our data suggest that promoting adult myelination is a potential strategy for reversing depressive-like social behavior. SIGNIFICANCE STATEMENT Oligodendrocyte development and myelination are highly dynamic processes influenced by experience and neuronal activity. However, whether enhancing myelination and oligodendrocyte differentiation is beneficial to treat depressive-like behavior has been unexplored. Mice undergoing prolonged social isolation display impaired myelination in the prefrontal cortex. Clemastine, a Food and Drug Administration-approved antimuscarinic compound that has been shown to enhance myelination under demyelinating conditions, successfully reversed social avoidance behavior in adult socially isolated mice. This was associated with enhanced myelination and oligodendrocyte differentiation in the prefrontal cortex through epigenetic regulation. Thus, enhancing myelination may be a potential means of reversing depressive-like social behavior.
