GUT MICROBES 2020, VOL. 12, NO. 1, e1813532 (15 pages) https://doi.org/10.1080/19490976.2020.1813532 RESEARCH PAPER How oral probiotics affect the severity of an experimental model of progressive multiple sclerosis? Bringing commensal bacteria into the neurodegenerative process Leyre Mestrea,b*, Francisco J. Carrillo-Salinas a,c*, Ana Feliú a,b, Miriam Mechaa,b, Graciela Alonsoa,b, Carmen Espejob,d,e, Laura Calvo-Barreirod,e, José L. Luque-Garcíaf, Héctor Estevezf, Luisa María Villar b,g, and Carmen Guazaa,b a Neuroimmunology Group, Functional and Systems Neurobiology Department, Instituto Cajal, CSIC, Madrid, Spain; bRed Española de Esclerosis Múltiple (REEM); cTufts University School of Medicine, Boston, MA, USA; dServei de Neurología-Neuroimmunología, Centre d’Esclerosi Múltiple de Catalunya, Vall d’Hebron Institut de Recerca, Hospital Universitari Vall d’Hebron, Barcelona, Spain; eUniversitat Autònoma de Barcelona, Bellaterra, Spain; fAnalytical Chemistry Department, Complutense University of Madrid, Madrid, Spain; gImmunology Department, Hospital Universitario Ramón y Cajal de Investigación Sanitaria (IRYCIS), Madrid, Spain ABSTRACT ARTICLE HISTORY A growing number of studies support that the bidirectional interactions between the gut micro­ Received 24 February 2020 biota, the immune system and the CNS are relevant for the pathophysiology of MS. Several studies Revised 15 July 2020 have reported alterations in the gut microbiome of MS patients. In addition, a variety of studies in Accepted 12 August 2020 animal models of MS have suggested that specific members of the gut commensal microbiota can KEYWORDS exacerbate or ameliorate neuroinflammation. Probiotics represent oral nontoxic immunomodula­ Theiler’s virus model; tory agents that would exert benefits when using in combination with current MS therapy. Here we probiotics; gut microbiota; investigate the effect of Vivomixx on the gut microbiome and central and peripheral immune neuroinflammation; multiple responses in a murine model of primary progressive MS. Vivomixx administration was associated sclerosis; treg cells; breg cells with increased abundance of many taxa such as Bacteroidetes, Actinobacteria, Tenericutes and TM7. This was accompanied by a clear improvement of the motor disability of Theiler’s virus infected mice; in the CNS Vivomixx reduced microgliosis, astrogliosis and leukocyte infiltration. Notably, the presence of Breg cells (CD19+CD5+CD1dhigh) in the CNS was enhanced by Vivomixx, and while spinal cord gene expression of IL-1β and IL-6 was diminished, the probiotic promoted IL-10 gene expression. One of the most significant findings was the increased plasma levels of butyrate and acetate levels in TMEV-mice that received Vivomixx. Peripheral immunological changes were subtle but interestingly, the probiotic restricted IL-17 production by Th17-polarized CD4+ T-cells purified from the mesenteric lymph nodes of Theiler’s virus infected mice. Our data reinforce the beneficial effects of oral probiotics that would be coadjuvant treatments to current MS therapies. Introduction Studies using the encephalomyelitis autoimmune Although the etiology of MS remains unknown, envir­ experimental (EAE) model of MS described lower onmental and genetic risk factors are involved in the incidence in germ free (GF) or antibiotic treated pathogenesis of the disease.1 Accumulating evidence mice than in specific pathogen free (SPF) mice,6–8 indicate that microbiome gut dysbiosis may be impli­ while recolonization restored the susceptibility to cated in the increased incidence of MS in developed develop EAE.3 However, other studies reported countries.2 Specifically, bi-directional interactions a therapeutic role of probiotic administration in between the gut microbiota, the intestinal barrier a variety of EAE models. On this matter, orally admin­ and the immune system appear to be relevant to the istration of Bacteroides fragilis,9,10 Prevotella pathophysiology of MS.3–5 The different stages and histicola,11 different species of Lactobacillus subtypes of MS are influenced by environmental (L. paracasei plus L. plantarum;12 L. Reuteri;13 factors,1 particularly in the relapsing remitting MS L. helveticus14), or multi-strain probiotics15 have forms (RRMS) with current successful therapies. demonstrated to improve EAE. In addition, probiotics CONTACT Leyre Mestre leyre@cajal.csic.es; Carmen Guaza cgjb@cajal.csic.es Neuroimmunology Group, Functional and Systems Neurobiology Department Instituto Cajal, CSIC; ; Madrid 28002, Spain *These authors contributed equally to this work © 2020 The Author(s). Published with license by Taylor & Francis Group, LLC. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. e1813532-2 L. MESTRE ET AL. supplementation to RRMS patients has been shown to Results modulate immunity by exerting anti-inflammatory Vivomixx treatment on the symptomatic chronic effects.16,17 phase of TMEV-IDD improves motor function Right now, there is little information regard­ ing the association between gut microbes and We explored the effect of manipulating the gut CNS inflammation/neurodegeneration in the microbiota by oral administration of Vivomixx dur­ primary progressive MS. Notably, MRI studies ing the chronic phase of disease (Figure 1(a)). in progressive MS show less brain permeability, Consumption of this probiotic improved motor probably reflecting CNS compartmentalized function (Figure 1(b)), witnessed through the inflammation. As a result, we have studied the increase in the horizontal and vertical activity dis­ effect of administering the oral multi-strain played by these mice (HACTV p < .01 vs TMEV- probiotic, Vivomixx, on the gut microbiome mice), as opposed to the more limited activity of the and plasma short chain fatty acids (SCFAs) in vehicle treated TMEV-mice (TMEV p < .01 vs sham the late chronic phase of a murine model of mice). The reduced latency to fall observed in PPMS, the Theiler’s murine encephalomyelitis TMEV-mice on the rotarod performance confirmed virus induced demyelinating disease (TMEV- the motor improvement induced by Vivomixx sup­ IDD). plementation (p < .01 vs TMEV-mice). Figure 1. Vivomixx-treatment improves the motor disability of TMEV-mice. (a) Scheme of the experimental groups. Mice were administered 3 × 108 cfu Vivomixx by oral gavage three times a week from day 70 pi to 85 dpi. Control mice were administered autoclaved water alone. b) Horizontal (HACTV) and vertical activity (VACTV), as well as motor coordination were evaluated. The data are presented as the means ± SEM: ***p < .001 vs. S; ##p < .01 vs T; n = 10 mice/group studied in two independent experiments. GUT MICROBES e1813532-3 Effects of oral Vivomixx on gut microbiota and in the CNS of TMEV-infected mice (Figure 3(e-g)). blood content of SCFAs Conversely, the presence of B cells (Figure 3(h)) was strongly influenced by oral Vivomixx administration, Miseq Illumina analysis of the bacterial population which significantly reduced the proportion of B cells revealed that TMEV-mice showed a decreased rela­ (Figure 3(i); p < .001 TMEV vs sham; p < .001, tive abundance of Bacteroides, Parabacteroides, TMEV plus Vivomixx vs TMEV). More interestingly, Prevotella, Sutterella or Desulfovibrio while an Vivomixx enhanced the presence of the Breg subpo­ increase of Anaerostipes, Dorea, pulation B220+CD19+CD5+CD1dhigh cells (Figure 3 Erysipelotrichaceae or Coprobacillus were observed (j); p < .05 vs TMEV). (Figure 2(a)). Sham mice subjected to oral Vivomixx treatment experiment subtle changes in gut microbiota composition. In the case of Oral Vivomixx modifies microglial morphology and Vivomixx-TMEV mice the relative abundance reduces astrogliosis promoting an Anaerostipes, Dorea, Oscillospira, anti-inflammatory profile in the spinal cord Enterobacteraceae or Ruminococcus were decreased Microglial cells maintain the CNS homeostasis, being while Bacteroides, Odoribacter, Lactobacillus, or activated in MS and its experimental models. As Sutterella were increased (Figure 2(a)). To give expected, the area of Iba-1 staining is larger in robustness to our data we compare the gut micro­ TMEV-mice (p < .01 vs sham mice) and Vivomixx biota composition of TMEV-mice before and after enhanced this effect, further increasing the area of Vivomixx treatment, we found an increased relative Iba-1 staining (p < .05 vs TMEV: Figure 4(a)). abundance of Bacteroidetes and a significant Nevertheless, a detailed examination of these cells decrease in Firmicutes. In addition, the percentage indicated subtle morphological changes in the spinal of bacteria belonging to Actinobacteria, Tenericutes cord of mice that received Vivomixx reflecting that and TM7 phyla tend to increase, but not signifi­ microglia exhibited anti-inflammatory activation or cantly (Figure 2(b)). At genus level (Figure 2(c)), attenuated proinflammatory activity (Figure 4(a)). In Vivomixx treatment tend to rebalance microbiota addition, probiotic supplementation was associated changes increasing the relative abundance of with weaker Vimentin-GFAP co-localization in the Bacteroides, Odoribacter or Lactobacillus, while spinal cord of TMEV-infected mice, whereas TMEV- Oscillospira, Ruminococcus or Bilophila were signif­ mice had a higher MoC than sham mice (Figure 4(b), icantly reduced. Gut microbiota changes would be p < .05 vs sham mice and p < .05 vs TMEV-mice). related to the increased levels of acetate and buty­ Indeed, when we analyzed the expression of pro- rate observed in the plasma of TMEV-mice after inflammatory cytokine genes (Figure 4(b)), we 15 days of probiotic treatment (Figure 2(d)). found that the elevated expression of IL-1β and IL- 6 in TMEV-mice (p < .001 and p < .05 vs sham, Oral Vivomixx alters CNS infiltrates in the TMEV-IDD respectively) was significantly reduced by Vivomixx augmenting the frequency of the regulatory (p < .01 and p < .05 vs TMEV-mice, respectively). B220+CD19+CD5+CD1high B lymphocytes Conversely, expression of the IL-10 gene was pro­ moted by Vivomixx (p < .01 vs TMEV-mice), while We assessed whether administering Vivomixx altered IL-4 expression displayed a tendency toward the T cell subpopulation in the CNS of TMEV-mice enhancement. Intriguingly, the elevated TNF-α gene (Figure 3(a)). A tendency toward a reduction in the expression observed in TMEV-mice remained proportion of CD4+ T cells was observed (Figure 3 unchanged by the probiotic treatment (Figure 4(c)). (b)), with no change in the frequency of CD8+ T cells (Figure 3(c)). However, the increase in the number of Peripheral immunological changes in the leukocytes in the CNS of TMEV-infected mice was symptomatic chronic phase of TMEV-IDD associated significantly dampened by probiotic supplementation with Vivomixx supplementation (Figure 3(d)). In addition, probiotic gavage did not significantly modify the proportions of The proportions of CD4+ and CD8+ T cells in the Foxp3+CD39+CD4+ or Foxp3−CD39+CD4+ T cells spleen of TMEV-mice were not significantly e1813532-4 L. MESTRE ET AL. Figure 2. Changes in gut microbiota population after Vivomixx treatment. Fecal pellets were collected freshly from each mouse before and after probiotic treatment. A) Stacked bar charts of the relative abundance of main bacterial communities, at genus level, present in the different experimental groups. Extended error bar plots, at phylum (b) and genus level (c), shows the differences in the mean proportions of bacterial taxa in TMEV mice before (black) and after treatment (green). Corrected p-values are shown on the right. D) Quantification of the plasma levels of acetate and butyrate. The data are presented as the means ± SEM: ***p < .001 vs. S; ##p < .01 vs T; n = 5 mice/group. GUT MICROBES e1813532-5 Figure 3. Effect of Vivomixx treatment on the T and B cell populations in the CNS. A single cell leukocyte suspension was obtained from the brain and spinal cord. A, H) Representative flow cytometry plots of CD45+CD3+CD4+ T cells, CD45+CD3+CD8+ T cells (a), and B220+CD19+ B cells and CD5+CD1dhigh cells gated on B220+CD19+ B cells (h). B-D) Quantification of the change in the number of CD45+CD3+CD4+, CD45+CD3+CD8+ T cells and leukocytes relative to S mice E) Representative flow cytometry plots of CD39Foxp3 cells gated on CD45+CD3+CD4+ T cells. F-G) Percentage of Foxp3+CD39+ and Foxp3−CD39+ T cells gated on CD45+CD3+CD4+ T cells, respectively. The data represent the change relative to the S mice. I) Quantification of the change in the percentage of B220+CD19+ B cells relative to the S mice. J) Percentage of CD5+CD1dhigh cells gated on B220+CD19+ B cells in the CNS: Groups S, T (n = 10); groups SVx, TVx (n = 5); **p < .01 vs S; ***p < .001 vs. S; #p < .05 vs. T; ###p < .001 vs T. affected by Vivomixx administration (Figure 5(a)), evident in T cells from TMEV-mice and this was although more intense CD4+ fluorescence was dampened when they received the probiotic (Figure e1813532-6 L. MESTRE ET AL. Figure 4. Vivomixx administration in the chronic phase of TMEV-IDD promotes anti-inflammatory responses in the spinal cord. A) Representative images of transversal spinal cord sections immunostained with Iba-1 to label microglia. Scale bar 100 μm. Below are higher magnification images (63x) of Iba-1 stained cells that indicate the changes in morphology. Scale bar 10 μm. Quantification of the area occupied by Iba-1 staining is also shown on the right. B) Representative images of GFAP (red) and Vimentin (green) staining in the spinal cord together the quantification of Vimentin-GFAP overlap through the Manders coefficient measure. C) qRT-PCR analysis of inflammatory markers (mIL-1β, mIL-6, mTNF-α, mIL-4 or mIL-10) in the spinal cord as assessed using the 2−ΔΔCt method. *p < .05 vs. S; **p < .01 vs. S; ***p < .001 vs. S; #p < .05 vs. T; ##p < .01 vs. T. 5(b)). Although the frequency of Treg cells was received Vivomixx (p < .05 and p < .001 vs TMEV, similar in TMEV-mice irrespective of their treat­ respectively) (Figure 5(c)). By contrast, there were ment with Vivomixx, the percentage of no significant changes in the proportion of B cells Foxp3+CD39+ and Foxp3−CD39+ T cells was dra­ (Figure 5(d)) or in the abundance of Breg cells in matically reduced in the spleen of TMEV-mice that the spleen of TMEV-mice receiving Vivomixx. GUT MICROBES e1813532-7 Figure 5. Subtle changes in the peripheral immunological response after Vivomixx administration to TMEV-mice. A spleen cell suspension was stained to analyze the CD4, CD8, CD39+ Treg and Breg cell populations. a) Representative flow cytometry plots of CD45+CD3+CD4+ T cells and CD45+CD3+CD8+ T cells, along with their corresponding quantification. b) Histogram of fluorescence emitted by CD3+CD4+ spleen cells from representative mice in each experimental group. c) Flow cytometry plots of Foxp3+CD39+ Treg cells gated on CD45+CD3+CD4+ T cells, together with a quantification of the percentage of Foxp3+ T cells (gated on CD4+ T cells), Foxp3−CD39+ and Foxp3+CD39+ Treg cells. d) Representative flow cytometry images and quantification of the B220+CD19+ cells in the spleen, together with a normalized histogram and quantification of CD1d expression in B220+CD19+CD5+ gated cells. Groups. S, T (n = 10); groups SVx, TVx (n = 5); **p < .01 vs S; ***p < .001 vs. S; #p < .05 vs. T; ###p < .001 vs T. e1813532-8 L. MESTRE ET AL. No significant variation was detected in the num­ administration of this probiotic, during the chronic ber of CD4+ T cells in the MLNs under the different phase improved the motor disability likely asso­ conditions studied (Figure 6(a)). However, the ciated to changes in gut microbiota. release of IL-17 by Th17 polarized lymphocytes Differences in the composition of the fecal derived from TMEV-mice was stronger respect to microbiota at the genus level have been reported sham mice (p < .05) (Figure 6(b)). By contrast, sig­ in a variety of studies on MS patients.4,5,18,19 In our nificantly less IL-17 was produced by CD4+ T cells study TMEV-mice showed changes in gut micro­ obtained from the MLNs of TMEV-mice that biota population compatible with MS disease like received Vivomixx (p < .05 vs TMEV-mice). decreases of Prevotella, Bacteroides, Parabacteroides or Coprobacillus relative abundance or an increase of Anaerostipes.18,24,25 Vivomixx treatment tends to Discussion decrease Anaerostipes, Dorea, Oscillospira, In the last decade, evidence has emerged of Enterobacteraceae or Ruminococcus while a relationship between gut microbiota and MS, Bacteroides, Odoribacter, Lactobacillus, or both in patients and in animal models.4,5,18,19 In Sutterella were increased. These changes would be addition, there is significant support for the gut associated with the beneficial effects of probiotic microbiota playing a critical role in shaping the treatment. For example, Dorea have been consid­ immune system.20–22 Several reports described, in ered to be part of the healthy gut microbiota or EAE models, a therapeutic effect of multi-strain patients on disease-modifying treatment show probiotics15 and specific bacterial strains like increased abundance of Sutterella and Prevotella.5 Bacteroides,9,10 Prevotella11 or Lactobacillus. In the TMEV-IDD model we found that mice However, the involved mechanisms for their bene­ receiving oral Vivomixx had fewer CNS infiltrated fits seem to be different. For example, while leukocytes, particularly B cells, without significant L. paracasei DSM13434 and L. plantarum DSM changes in the proportion of CD4+ and CD8+ 15312 increased TGF-β1 in the blood, T cells, Treg cells, or Foxp3+CD39+ T cells. One L. plantarum DSM 15313 enhanced blood levels of the pathways by which the commensal micro­ of IL-27.12 Until date the only probiotic considered biome affects host immunity at distal sites is by a kind of medical food is Vivomixx23.Here, we regulating the activity of Treg cells through the proved the efficacy of Vivomixx treatment in ectonucleotidase CD39.10 Particularly, CD39+ cells a model of progressive MS whereby oral have been associated with the suppression of Figure 6. Vivomixx affects the Th17 responses in the MLNs. CD4 T cells isolated from MLN were polarized to Th17 in culture. The expression of IL-17 was measured by ELISA, and the data represent the mean ± SEM of the change relative to the S or T mice from two experiments performed in triplicate: *p < .001 vs S; #p < .05 vs T. GUT MICROBES e1813532-9 neuroinflammation by Bacteroides fragilis in the present study, the protective effects against EAE EAE model26 and its polysacharide A was not able following Lactobacillus helveticus seems to involve to protect against EAE in CD39-deficient mice.10 decreased IL-6 and reduced Th-17 cells in the spinal Here, TMEV infection decreased CD39 Treg cells cord, without changes in Tregs.14 population in the CNS, in accordance with our Preclinical studies suggest that the microbiome previous results27 and previous data from the per­ plays a critical role in regulating microglial matura­ ipheral blood of MS patients.28 Unlike what we tion and function.33 Indeed, microglia from TMEV- might expect, probiotic treatment did not reverse infected mice administered Vivomixx undergo mor­ the decreased CD39+ Tregs suggesting that CD39 phological changes, which could reflect changes in activity is not the main pathway involved in the their functional state, in accordance with the emer­ anti-inflammatory effects of Vivomixx. Indeed, our gent hypothesis that microglia may adopt distinct results are consistent with data from MS patients CNS region-specific phenotypes.34 In the spinal who did not show changes in the frequencies of cord, the pro-inflammatory environment of TMEV- CD4, CD8 or CD39+T cells following Vivomixx.17 mice shifted to an anti-inflammatory profile follow­ It is known that in EAE, encephalitogenic T cells ing Vivomixx supplementation. upregulate their CD4 cell surface expression follow­ In an attempt to find a mechanism that may ing antigen recognition and that this was associated explain the changes in the CNS of TMEV-mice with EAE development, for example after transfer that receive oral Vivomixx we assessed plasma acet­ to naïve animals.29 Therefore, although oral ate and butyrate. On the whole, it seems that the Vivomixx did not affect the frequency of CD4+ reduction in bacteria producing SCFAs may be T cells, the fact that this probiotic restricted the a common characteristic of MS patients.18 A high surface expression of CD4 on T cells suggest its fiber diet or SCFA supplementation are reported to possible contribution to its beneficial effects in ameliorate EAE35 and, there is growing evidence TMEV mice. supporting that circulating metabolites of the Recent reports show that gut microbiota and microbiota can influence CNS and immune cell microbial products can directly regulate B-cell activity,33 including microglial function.27,36 One development, activation, and differentiation.30 In of the most significant findings of the present addition, the induction of Breg cells in an autoim­ study was the increased butyrate and acetate levels mune context was seen to depend on a healthy in the plasma of TMEV-mice that received microbiota.31 Notably, in our model of progressive Vivomixx which may contribute to the altered MS, Vivomixx increased the frequency of Breg cells, immune CNS responses and the improved motor unveiling a different mechanism of action to that activity. Indeed, several reports indicate the bene­ exerted by other probiotics in EAE models where ficial effects of SCFA suppressing inflammation in Treg cells have been specifically implicated.26,32 the EAE model.35,37,38 Furthermore, a recent find­ Nonetheless, Bregs would be involved in the bene­ ing shows that specific SCFAs exert an epigenetic ficial effect of Vivomixx but functional studies are modulation of Breg responses.39,40 Particularly, needed to be sure that these cells are working in an EAE mice that were subjected to the transference effective way. There are many crucial differences of Breg cells, expanded in vitro in presence of between EAE and the TMEV model. First, the pentanoate, did not develop the classical symptoms. TMEV ability to maintain the inflammation into Furthermore, butyrate increased the levels of the CNS compartment, with minor effects in the 5-Hydroxyindole-3-acetic acid, a metabolite of ser­ periphery unlike that observed in EAE. On this way, otonine neurotransmitter which activates the aryl- most of studies noticed the induction of IL-10 pro­ hydrocarbon receptor (AhR), a newly discovered ducing Tregs in the periphery as the main mechan­ transcriptional marker for Breg function.41 Here, ism for EAE protection when mice received many we speculate that Breg response as well as the probiotics as Lactobacilli12 or Prevotella Histicola.11 reduction of IL17 production by Th17 mesenteric The second possibility comes from the different polarized cells may be associated with acetate and strain of mice used in the models of MS. In EAE butyrate increments following Vivomixx adminis­ models that use SJL/L mice, like we used in the tration to TMEV-mice. e1813532-10 L. MESTRE ET AL. Although, there is no clear evidence of 4-aminobutyrate pathways, our finding about a relationship between SCFA and Bregs at CNS increased Odoribacter abundance would be asso­ level, astrocytes express AhR that, when activated ciated with the increased levels of butyrate in limit CNS inflammation. In the EAE model the TMEV-mice treated with Vivomixx. results by Quintana’s group42 suggest a link The subtle changes promoted by Vivomixx at between metabolites derived from dietary trypto­ spleen level are in accordance with the compart­ phan by the commensal gut bacteria and the activa­ mentalization of inflammation that occurs in pro­ tion of AhR signaling in astrocytes. It is worth gressive MS, characterized by intrinsic CNS noting that our understanding of the molecular processes driven by resident cells like microglia pathways involved in the effects of probiotics and and astrocytes. As such, the main immunological SCFAs on CNS inflammation, in MS and its animal changes in the chronic phase of TMEV-IDD linked models, is still very limited. This statement also to the manipulation of the gut microbiota by oral includes microglial cells which are thought to be Vivomixx happen within the CNS. Although pro­ targeted by SCFAs,36 although their underlying biotics do not modify the host microbiome in mechanisms remain incomplete too. On this line, a robust and persistent manner, they can influence more recently, a new pathway in which microbial gut immunity and homeostasis.47 Such effects were metabolites may limit the pathogenic activity of evident in the TMEV model through the CNS microglia and astrocytes has been identified.43 immune responses, microgliosis and astrogliosis, Then, we would like to highlight the reduction of IL-17 production by ex vivo mesenteric CD4+ astrogliosis and microgliosis we found in Vivomixx T cells, and more importantly, the severity of treated-TMEV mice. motor dysfunction. An important issue to consider is that MS disease Regarding probiotics as therapeutic agents, modifying therapies may alter the microbiome VSL#3 administered to a small number of RRMS directly. As an interesting example, the abundance patients leads to an anti-inflammatory peripheral of Prevotella was increased in MS patients treated immune response.17 It seems clear that both, dis­ with glatiramer acetate taking into account that MS ease promotion and amelioration may be induced reduced the frequency of these bacteria.44 by specific components of the gut microbiota that Furthermore, in a humanized transgenic EAE pre­ interact bi-directionally with the host immune sys­ clinic model of MS, the combination of human tem. This concept implies that selective experimen­ Prevotella histicola and glatiramer acetate did not tal strategies manipulating specific subsets of have a synergistic effect, but Prevotella was as effec­ bacteria or balancing specific gut microbiota may tive as glatiramer in suppressing the disease.11 be useful as a therapeutic approach in MS. Our While dogma dictates that Firmicutes are findings may provide a starting point for further responsible for butyrate and Bacteroidetes are mechanistic studies that help to develop adjuvant responsible for acetate and propionate production, treatments complementary to the available thera­ there is evidence for alternative pathways.45 There peutic options for progressive MS patients. are non-canonical butyrate-producing pathways that use amino acids as primary substrates, includ­ ing the lysine, 4-aminiobutyrate, pathways. Materials and Methods Considering metagenomic analysis of samples Animals and TMEV infection from healthy human patients from the Human Microbiome Project a significant overlap between TMEV-infected susceptible female SJL/J mice the genes involved in the lysine and 4-aminobuty­ (Charles River) were maintained under conven­ rate pathways and those expressed by members of tional controlled conditions. The right brain hemi­ the phylum Bacteroidetes was found for butyrate sphere of 6–8 weeks-old mice was inoculated with production. Then, it was predicted that these both 2 × 106 plaque forming units (pfu) of the Daniel pathways can contribute to the butyrogenic poten­ (DA) strain of TMEV). Sham control mice received tial of Bacteroidetes.46 Because members of the the vehicle alone. Each experimental group was Odoribacter genus dominated the lysine and housed in different cages and in different isolated GUT MICROBES e1813532-11 racks in the same room. Each cage contained 5 mice Metabolite analysis subjected to the same treatment, receiving the same Blood samples in EDTA-treated BD vacutainer food and sterile water, and manipulated by the were centrifuged for 10 min at 1500 g in a refriger­ same person. All experiments were performed in ated centrifuge within two hours of extraction and strict accordance with EU (Directive 2010/63/EU) aliquots were stored at −20°C. Acetate and butyrate and Spanish regulations (Royal Decree 53/2013 levels were assessed as described previously.27 BOE n° 34 and “Comunidad de Madrid” decree: Validation of the method was carried out ES 280790000184). The Ethics Committee on using13C-acetate and13C-butyrate standards. For Animal Experimentation at the Instituto Cajal both analytes the recoveries were above 90% and approved all the procedures employed in this the precision, in terms of relative standard devia­ study (protocol number: 2013/03 CEEA-IC). tion, was below 10%. Treatments Ex vivo experiments on CD4 T cells 8 TMEV-mice were orally gavaged with 3 × 10 cfu CD4+ T cells were isolated from the MLN using (100 μl) of Vivomixx for 70 to 85 days post infec­ a commercial cell isolation kit and following the tion (dpi) (TVx group; n = 10). Vivomixx, a multi- manufacturer’s instructions (Miltenyi, Biotec Inc). species probiotic is composed of Lactobacillus para­ Cells from each animal were cultured and polarized casei DSM 24734, Lactobacillus plantarum DSM to Th17 cells in RPMI supplemented with: 2-β 24730, Lactobacillus acidophilus DSM 24735, Mercaptoethanol (50 μM); immobilized anti- Lactobacillus delbruckeii subspecies bulgaricus mouse CD3ε; anti-mouse CD28; anti-mouse IFN- DSM 24734, Bifidobacterium longum DSM 24736, γ; anti-mouse IL-4 (all from BD biosciences); Bifidobacterium infantis DSM 24737, hTGF-β and recombinant murine IL-6 (from Bifidobacterium breve 24732 and Streptococcus ther­ Peprotech); and recombinant murine IL-23 mophilus DSM 24731. Sham or TMEV-mice admi­ (BioLegend). The cells were stimulated 3 days nistered the vehicle (S group, n = 10 or T group, later for 5 hours with PMA (100 ng/ml) and iono­ n = 10, respectively) and Sham mice administered mycin (500 ng/ml, both from Tocris Bioscience). Vivomixx (SVx group, n = 5). The supernatants were collected and stored at −20° C for IL-17 measurement by solid phase sandwich ELISA using Quantikine kits (M1700; R&D systems Evaluation of motor function Inc), according to the manufacturer’s instructions. Mice were followed for the progression of demye­ The assay’s sensitivity was 5 pg/ml. linating disease by evaluating spontaneous motor activity on day 70 and 85 pi using an activity cage (Activity Monitor System Omnitech Electronics, Tissue processing and immunohistochemistry Inc.; Colombus, OH, USA) and motor coordination Anesthetized mice were perfused transcardially by rotarod test as we previously described.48 with saline and the spinal cord processed as pre­ viously described.48 Microglia were labeled with a primary antibody against Iba-1 (ionized calcium Sample collection binding adaptor molecule 1, 1:1,000: Wako Fecal samples freshly obtained from each mouse Chemical Pure Industry), while reactive astrogliosis on day 70 and 85 pi were stored at −80°C. On day was assessed by GFAP (1:2000: Sigma) and 85 pi, spleen and MLN were removed from Vimentin (1:1000: Sigma) labeling. Five spinal anesthetized mice (Dolethal; 50 mg/kg body cord sections were examined per animal and at weight) for flow cytometry and ex vivo CD4-T cell least three mice per group were used to quantify studies. Then, anesthetized mice were perfused the immunostained area using ImageJ software transcardially with saline to obtain the brain and (NIH, Bethesda). The white matter area in the spinal cord for further FACS analysis. ventral and lateral horns was analyzed by e1813532-12 L. MESTRE ET AL. immunofluorescence on a Leica TC SP5 confocal anti-CD3; PerCP-Cy5.5-conjugated anti-CD45; microscope. Representative images of the surface of PerCP-Cy5.5-conjugated anti-B220; PE-Cy7- a single microglial cell were obtained with Imaris conjugated anti-CD39; PE-Cy7-conjugated anti- software. Astroglial activation defined as the CD8a; PE-Cy7-conjugated anti-CD19; APC anti- amount of Vimentin relative to the GFAP staining CD3; APC-Cy7-conjugated anti-CD4; APC-Cy7- (Vimentin-GFAP overlapping) was analyzed using conjugated anti-CD11b (all from BD the Manders overlap coefficient (MoC) that quan­ Pharmingen); APC anti-CD3; and PE-Cy7- tifies the degree of co-localization of the two conjugated anti-CD39 (both from eBioscience). fluorophores. To detect Foxp3, the cells were suspended in the fixation/permeabilization buffer and stained with a PE anti-Foxp3 antibody (BD Pharmingen). RNA extraction, reverse transcription and RT-PCR Cell viability was assessed with the LIVE/DEAD Total RNA from the cervical spinal cord was Fixable Green Dead Cell Stain Kit or LIVE/ extracted and then transcribed into cDNA as we DEAD Fixable Red Dead Cell Stain Kit (Life previously described.49 Real-time PCR (RT-PCR) Technologies Corporation). At least 10,000 was performed with SYBR® and the following oli­ events were acquired in each experiment on gonucleotide primers mIL-1β, forward 5ʹ-TGG a FACSaria flow cytometer (BD Biosciences) TGT GTG ACG TTC CCA TT-3ʹ, reverse 5ʹ-TCC and the data were analyzed using FlowJo soft­ ATT GAG GTG GAG AGC TTT C-3ʹ; mTNF-α, ware (v.10; Tree Star). forward 5ʹ-AGA GGC ACT CCC CCA AAA GA-3ʹ, reverse 5ʹ-CGA TCA CCC CGA AGT TCA GT-3ʹ; 16S rRNA analysis of the microbial community mIL-4, forward, 5ʹ-CCACGGATGCGACAAAA ATC-3ʹ, reverse 5ʹ-GACGTTTGGCACATCCA Fecal DNA was isolated using the QIAamp DNA TCTC-3ʹ; mIL-10, forward 5ʹ-TGA ATT CCC Stool Mini Kit (Qiagen; cat. no. 51504) and follow­ TGG GTG AGA AGC TGA-3ʹ, reverse 5ʹ-TGG ing the stool pathogen detection protocol. DNA CCT TGT AGA CAC CTT GGT CTT-3ʹ; mIL-6, was extracted following the manufacturer’s forward 5ʹ-TCC AGA AAC CGC TAT GAA GTT instructions with the modifications described C-3ʹ, reverse 5ʹ-CAC CAG CAT CAG TCC CAA previously,27 and the quality of the reads was ana­ GA-3ʹ; and mRps29 forward 5ʹ-GCC GCG TCT lyzed with FastQC software (http://www.bioinfor GCT CCA A-3ʹ, reverse 5ʹ-ACA TGT TCA GCC matics.babraham.ac.uk/projects/fastqc/). The CGT ATT TGC-3ʹ. PCR amplification were microbiome analysis of the paired-end reads was initiated incubating at 50°C for 2 min and 95°C performed using the QIIME2 v2018.8 software for 10 min, and amplification was performed over (Qualitative Insights into Microbial Ecology)50 40 cycles at 95°C for 15 s and 60°C for 1 min. The and the sequences were grouped as “sequence var­ cDNA samples were analyzed in triplicate on an iants” (SVs), equivalent to the 100% Operational Applied Biosystems PRISM 7500 Sequence detec­ Taxonomic Units (OTUs). Representative tion system, normalizing the mRNA expression to sequences from each cluster were queried against the Rps29 gene in each sample and quantifying gene the GreenGenes database (version 13_8) to obtain expression by the 2−ΔΔCt method. The results are taxonomic assignments at a 97% sequence identity expressed relative to the Sham mice for each time using BLAST. These representative sequences point. were then used for phylogenetic analysis and for diversity calculations. Flow cytometry antibodies Statistical analysis A CNS or spleen leukocyte suspension was obtained as described previously.49 Freshly iso­ The data were expressed as the mean ± standard lated cells (106) were incubated with an anti- error or the mean (SEM) and they were analyzed mouse CD16/CD32 FC (eBioscience) and labeled using GraphPath Prism5 Software. One-Way with anti-mouse antibodies: PE anti-CD25; PE ANOVA followed by the Bonferroni post hoc GUT MICROBES e1813532-13 test were used to determine the statistical sig­ 2. Olsson T, Barcellos LF, Alfredsson L. Interactions nificance. For non-parametric analyzes, between genetic, lifestyle and environmental risk factors a Kruskall-Wallis and Mann-Whitney U test for multiple sclerosis. Nat Rev Neurol. 2017;13:25–36. doi:10.1038/nrneurol.2016.187. were applied. 3. Berer K, Mues M, Koutrolos M, Rasbi ZA, Boziki M, Johner C, Wekerle H, Krishnamoorthy G. Commensal microbiota and myelin autoantigen cooperate to trigger Acknowledgments autoimmune demyelination. Nature. 2011;479:538–541. doi:10.1038/nature10554. We thank Laura Ramos and Nieves López de Andradas for 4. Chen J, Chia N, Kalari KR, Yao JZ, Novotna M, Paz their excellent technical support. We are grateful to Dr. Moses Soldan MM, Luckey DH, Marietta EV, Jeraldo PR, Rodríguez (Department of Immunology and Neurology, Chen X, et al. Multiple sclerosis patients have Mayo Clinic/Foundation, Rochester, MN, USA) for the gener­ a distinct gut microbiota compared to healthy ous gift of the Theiler’s virus DA strain. We also thank the controls. Sci Rep. 2016;6:28484. doi:10.1038/srep28484. CBMSO-Genomics & Massive Sequencing service, CBMSO- 5. Jangi S, Gandhi R, Cox LM, Li N, von Glehn F, Yan R, Cytometry service, the Fundación Parque Científico de Patel B, Mazzola MA, Liu S, Glanz BL et al. Alterations Madrid and the Instituto Cajal-Microscopy Unit for their of the human gut microbiome in multiple sclerosis. Nat brilliant technical assistance. Commun. 2016;7:12015. doi:10.1038/ncomms12015. 6. Ochoa-Reparaz J, Mielcarz DW, Ditrio LE, Burroughs AR, Foureau DM, Haque-Begum S, Author contributions Kasper LH. Role of gut commensal microflora in the development of experimental autoimmune LM and FJC conducted the research, performed the experi­ encephalomyelitis. J Immunol. 2009;183:6041–6050. ments and evaluate the results. AF and MM collaborate in doi:10.4049/jimmunol.0900747. FACs assay; GA realized immunohistochemistry studies; JLLG 7. Yokote H, Miyake S, Croxford JL, Oki S, Mizusawa H, and HE performed the metabolite analysis; LMV, CE and LCB Yamamura T. NKT cell-dependent amelioration of reviewed the manuscript. LM and CG designed and supervised a mouse model of multiple sclerosis by altering gut the study, discussed the experimental setting and wrote the flora. Am J Pathol. 2008;173:1714–1723. doi:10.2353/ manuscript. ajpath.2008.080622. 8. Lee YK, Menezes JS, Umesaki Y, Mazmanian SK. Proinflammatory T-cell responses to gut microbiota Disclosure of Potential Conflicts of Interest promote experimental autoimmune encephalomyelitis. Proc Natl Acad Sci U S A. 2011;108(Suppl No potential conflicts of interest were disclosed. 1):4615–4622. doi:10.1073/pnas.1000082107. 9. Ochoa-Reparaz J, Mielcarz DW, Ditrio LE, Burroughs AR, Begum-Haque S, Dasgupta S, Kasper Funding DL, Kasper LH. Central nervous system demyelinating disease protection by the human commensal bacter­ This work was supported by grants from the Ministerio de oides fragilis depends on polysaccharide A expression. Economía y Competitividad [MINECO SAF2016-76449-R] J Immunol. 2010;185:4101–4108. doi:10.4049/ and the Red Española de Esclerosis Múltiple [REEM: RD16/ jimmunol.1001443. 0015/0001; RD16/0015/0004; RD16/015/0021], sponsored by 10. Wang Y, Begum-Haque S, Telesford KM, Ochoa- the Fondo de Investigación Sanitaria (FIS) Reparaz J, Christy M, Kasper EJ, Kasper DL, Robson SC, Kasper LH. A commensal bacterial product elicits and modulates migratory capacity of CD39+CD4 ORCID T regulatory subsets in the suppression of Francisco J. Carrillo-Salinas http://orcid.org/0000-0002- neuroinflammation. Gut Microbes. 2014;5:552–561. 7936-5859 doi:10.4161/gmic.29797. Ana Feliú http://orcid.org/0000-0002-1157-1525 11. Shahi SK, Freedman SN, Murra AC, Zarei K, Sompallae R, Luisa María Villar http://orcid.org/0000-0002-9067-3668 Gibson-Corley KN, Karandikar NJ, Murray JA, Mangalam AK. Prevotella histicola, a human gut com­ mensal, is as potent as COPAXONE(R) in an animal References model of multiple sclerosis. Front Immunol. 2019;10:462. doi:10.3389/fimmu.2019.00462. 1. Ebers GC. Environmental factors and multiple sclerosis. 12. Lavasani S, Dzhambazov B, Nouri M, Fak F, Buske S, Lancet Neurol. 2008;7:268–277. doi:10.1016/S1474- Molin G, Thorlacius H, Alenfall J, Jeppsson B, 4422(08)70042-5. Weström B, et al. A novel probiotic mixture exerts e1813532-14 L. MESTRE ET AL. a therapeutic effect on experimental autoimmune ence­ 24. Camara-Lemarroy CR, Metz LM, Yong VW. Focus on phalomyelitis mediated by IL-10 producing regulatory the gut-brain axis: multiple sclerosis, the intestinal bar­ T cells. PLoS One. 2010;5(2):e9009. doi:10.1371/journal. rier and the microbiome. World J Gastroenterol. pone.0009009. 2018;24:4217–4223. doi:10.3748/wjg.v24.i37.4217. 13. He B, Hoang TK, Tian X, Taylor CM, Blanchard E, 25. Schepici G, Silvestro S, Bramanti P, Mazzon E. The gut Luo M, Bhattacharjee MB, Freeborn J, Park S, microbiota in multiple sclerosis: an overview of clinical Couturier J, et al. Lactobacillus reuteri reduces the trials. Cell Transplant. 2019;28:1507–1527. doi:10.1177/ severity of experimental autoimmune encephalomyeli­ 0963689719873890. tis in mice by modulating gut microbiota. Front 26. Telesford KM, Yan W, Ochoa-Reparaz J, Pant A, Immunol. 2019;10:385. doi:10.3389/fimmu.2019.00385. Kircher C, Christy MA, Begum-Haque S, Kasper DL, 14. Yamashita M, Ukibe K, Matsubara Y, Hosoya T, Sakai F, Kasper LH. A commensal symbiotic factor derived from Kon S, Arima Y, Murakami M, Nakagawa H, Miyazaki T, Bacteroides fragilis promotes human CD39 + Foxp3 + et al. Lactobacillus helveticus SBT2171 attenuates experi­ T cells and T reg function. Gut Microbes. mental autoimmune encephalomyelitis in Mice. Front 2015;6:234–242. doi:10.1080/19490976.2015.1056973. Microbiol. 2017;8:2596. doi:10.3389/fmicb.2017.02596. 27. Mestre L, Carrillo-Salinas FJ, Mecha M, Feliu A, 15. Kwon HK, Kim GC, Kim Y, Hwang W, Jash A, Sahoo A, Espejo C, Alvarez-Cermeno JC, Villar LM, Guaza C. Kim J-E, Nam JH, Im S-H. Amelioration of experimental Manipulation of gut microbiota influences immune autoimmune encephalomyelitis by probiotic mixture is responses, axon preservation, and motor disability in mediated by a shift in T helper cell immune response. a model of progressive multiple sclerosis. Front Clin Immunol. 2013;146:217–227. doi:10.1016/j. Immunol. 2019;10:1374. doi:10.3389/fimmu.2019.01374. clim.2013.01.001. 28. Fletcher JM, Lonergan R, Costelloe L, Kinsella K, 16. Kouchaki E, Tamtaji OR, Salami M, Bahmani F, Moran B, O’Farrelly C, Tubridy N, Mills KHG. CD39 Daneshvar Kakhaki R, Akbari E, Tajabadi-Ebrahimi + Foxp3 + regulatory T cells suppress pathogenic Th17 M, Jafari P, Asemi Z. Clinical and metabolic response cells and are impaired in multiple sclerosis. J Immunol. to probiotic supplementation in patients with multiple 2009;183:7602–7610. doi:10.4049/jimmunol.0901881. sclerosis: a randomized, double-blind, placebo- 29. Li J, Ridgway W, Fathman CG, Tse HY, Shaw MK. High controlled trial. Clin Nutr. 2017;36(5):1245–1249. cell surface expression of CD4 allows distinction of CD4 doi:10.1016/j.clnu.2016.08.015. (+)CD25(+) antigen-specific effector T cells from CD4 17. Tankou SK, Regev K, Healy BC, Tjon E, Laghi L, (+)CD25(+) regulatory T cells in murine experimental Cox LM, Kivisäkk P, Pierre IV, Hrishikesh L, autoimmune encephalomyelitis. J Neuroimmunol. Gandhi R, et al. A probiotic modulates the microbiome 2007;192:57–67. doi:10.1016/j.jneuroim.2007.09.004. and immunity in multiple sclerosis. Ann Neurol. 30. Kim M, Kim CH. Regulation of humoral immunity by 2018;83(6):1147–1161. doi:10.1002/ana.25244. gut microbial products. Gut Microbes. 2017;8:392–399. 18. Miyake S, Kim S, Suda W, Oshima K, Nakamura M, doi:10.1080/19490976.2017.1299311. Matsuoka T, Chihara N, Tomita A, Sato W, 31. Rosser EC, Oleinika K, Tonon S, Doyle R, Bosma A, Kim S-W, et al. Dysbiosis in the gut microbiota of Carter NA, Harris KA, Jones SA, Klein N, Mauri C, et al. patients with multiple sclerosis, with a striking Regulatory B cells are induced by gut microbiota-driven depletion of species belonging to clostridia XIVa interleukin-1beta and interleukin-6 production. Nat and IV clusters. PLoS One. 2015;10(9):e0137429. Med. 2014;20:1334–1339. doi:10.1038/nm.3680. doi:10.1371/journal.pone.0137429. 32. Ochoa-Reparaz J, Kasper LH. The influence of 19. Mielcarz DW, Kasper LH. The gut microbiome in mul­ gut-derived CD39 regulatory T cells in CNS demyeli­ tiple sclerosis. Curr Treat Options Neurol. 2015;17:344. nating disease. Transl Res. 2017;179:126–138. doi:10.1007/s11940-015-0344-7. doi:10.1016/j.trsl.2016.07.016. 20. Wu HJ, Wu E. The role of gut microbiota in immune 33. Fung TC, Olson CA, Hsiao EY. Interactions between homeostasis and autoimmunity. Gut Microbes. the microbiota, immune and nervous systems in health 2012;3:4–14. doi:10.4161/gmic.19320. and disease. Nat Neurosci. 2017;20:145–155. 21. Honda K, Littman DR. The microbiota in adaptive doi:10.1038/nn.4476. immune homeostasis and disease. Nature. 34. De Biase LM, Schuebel KE, Fusfeld ZH, Jair K, Hawes IA, 2016;535:75–84. doi:10.1038/nature18848. Cimbro R, Zhang H-Y, Liu Q-R, Shen H, Xi Z-X, et al. 22. Colpitts SL, Kasper LH. Influence of the gut micro­ Local cues establish and maintain region-specific pheno­ biome on autoimmunity in the central nervous system. types of basal ganglia microglia. Neuron. 2017;95(2):341– J Immunol. 2017;198:596–604. doi:10.4049/ 56 e6. doi:10.1016/j.neuron.2017.06.020. jimmunol.1601438. 35. Haghikia A, Jorg S, Duscha A, Berg J, Manzel A, 23. Cheng FS, Pan D, Chang B, Jiang M, Sang LX. Probiotic Waschbisch A, Hammer A, Lee D-H, May C, Wilck N, mixture VSL#3: an overview of basic and clinical studies et al. Dietary fatty acids directly impact central nervous in chronic diseases. World J Clin Cases. system autoimmunity via the small intestine. Immunity. 2020;8:1361–1384. doi:10.12998/wjcc.v8.i8.1361. 2015;43:817–829. doi:10.1016/j.immuni.2015.09.007. GUT MICROBES e1813532-15 36. Erny D, Hrabe de Angelis AL, Jaitin D, Wieghofer P, the aryl hydrocarbon receptor. Nat Med. Staszewski O, David E, Keren-Shaul H, Mahlakoiv T, 2016;22:586–597. doi:10.1038/nm.4106. Jakobshagen K, Buch T, et al. Host microbiota con­ 43. Rothhammer V, Borucki DM, Tjon EC, Takenaka MC, stantly control maturation and function of microglia Chao CC, Ardura-Fabregat A, de Lima KA, Gutiérrez- in the CNS. Nat Neurosci. 2015;18:965–977. Vázquez C, Hewson P, Staszewski O, et al. Microglial doi:10.1038/nn.4030. control of astrocytes in response to microbial 37. Mizuno M, Noto D, Kaga N, Chiba A, Miyake S, metabolites. Nature. 2018;557:724–728. doi:10.1038/ Ashour HM. The dual role of short fatty acid chains in s41586-018-0119-x. the pathogenesis of autoimmune disease models. PLoS 44. Mangalam A, Shahi SK, Luckey D, Karau M, Marietta E, One. 2017;12:e0173032. doi:10.1371/journal. Luo N, Choung RS, Ju J, Sompallae R, Gibson-Corley K, pone.0173032. et al. Human gut-derived commensal bacteria suppress 38. Chitrala KN, Guan H, Singh NP, Busbee B, Gandy A, CNS inflammatory and demyelinating disease. Cell Rep. Mehrpouya-Bahrami P, Ganewatta MS, Tang C, 2017;20:1269–1277. doi:10.1016/j.celrep.2017.07.031. Chatterjee S, Nagarkatti P, et al. CD44 deletion leading 45. Macfarlane S, Macfarlane GT. Regulation of short-chain to attenuation of experimental autoimmune encephalo­ fatty acid production. Proc Nutr Soc. 2003;62:67–72. myelitis results from alterations in gut microbiome in doi:10.1079/PNS2002207. mice. Eur J Immunol. 2017;47(7):1188–1199. 46. Vital M, Howe AC, Tiedje JM, Moran MA. Revealing doi:10.1002/eji.201646792. the bacterial butyrate synthesis pathways by analyzing 39. Luu M, Pautz S, Kohl V, Singh R, Romero R, Lucas S, (meta)genomic data. mBio. 2014;5(2):e00889. Hofmann J, Raifer H, Vachharajani N, Carrascosa LC, doi:10.1128/mBio.00889-14. et al. The short-chain fatty acid pentanoate suppresses 47. van Baarlen P, Wells JM, Kleerebezem M. Regulation of autoimmunity by modulating the metabolic-epigenetic intestinal homeostasis and immunity with probiotic crosstalk in lymphocytes. Nat Commun. 2019;10:760. lactobacilli. Trends Immunol. 2013;34:208–215. doi:10.1038/s41467-019-08711-2. doi:10.1016/j.it.2013.01.005. 40. Sanchez HN, Moroney JB, Gan H, Shen T, Im JL, Li T, 48. Mestre L, Correa F, Arevalo-Martin A, Molina-Holgado E, Taylor JR, Zan H, Casali P. B cell-intrinsic epigenetic Valenti M, Ortar G, Marzo V, Guaza C. Pharmacological modulation of antibody responses by dietary modulation of the endocannabinoid system in a viral fiber-derived short-chain fatty acids. Nat Commun. model of multiple sclerosis. J Neurochem. 2020;11:60. doi:10.1038/s41467-019-13603-6. 2005;92:1327–1339. doi:10.1111/j.1471-4159.2004.02979.x. 41. Rosser EC, Piper CJM, Matei DE, Blair PA, 49. Carrillo-Salinas FJ, Mestre L, Mecha M, Feliu A, Del Rendeiro AF, Orford M, Alber DG, Krausgruber T, Campo R, Villarrubia N, Espejo C, Montalbán X, Catalan D, Klein N, et al. Microbiota-derived metabo­ Álvarez-Cermeño JC, Villar LM. Gut dysbiosis and lites suppress arthritis by amplifying aryl-hydrocarbon neuroimmune responses to brain infection with receptor activation in regulatory B cells. Cell Metab. Theiler’s murine encephalomyelitis virus. Sci Rep. 2020;31(4):837–51 e10. doi:10.1016/j.cmet.2020.03.003. 2017;7:44377. doi:10.1038/srep44377. 42. Rothhammer V, Mascanfroni ID, Bunse L, 50. Caporaso JG, Kuczynski J, Stombaugh J, Bittinger K, Takenaka MC, Kenison JE, Mayo L, Chao -C-C, Bushman FD, Costello EK, Fierer N, Peña AG, Patel B, Yan R, Blain M, et al. Type I interferons and Goodrich JK, Gordon JI, et al. QIIME allows analysis microbial metabolites of tryptophan modulate astrocyte of high-throughput community sequencing data. Nat activity and central nervous system inflammation via Methods. 2010;7:335–336. doi:10.1038/nmeth.f.303.