Selenization of S. cerevisiae increases its protective potential in experimental autoimmune encephalomyelitis by triggering an intestinal immunomodulatory loop

Multiple sclerosis is an autoimmune disease that affects the myelinated central nervous system (CNS) neurons and triggers physical and cognitive disabilities. Conventional therapy is based on disease-modifying drugs that control disease severity but can also be deleterious. Complementary medicines have been adopted and evidence indicates that yeast supplements can improve symptoms mainly by modulating the immune response. In this investigation, we evaluated the therapeutic potential of Saccharomyces cerevisiae and its selenized derivative (Selemax) in experimental autoimmune encephalomyelitis (EAE). Female C57BL/6 mice submitted to EAE induction were orally supplemented with these yeasts by gavage from day 0 to day 14 after EAE induction. Both supplements determined significant reduction in clinical signs concomitantly with diminished Th1 immune response in CNS, increased proportion of Foxp3+ lymphocytes in inguinal and mesenteric lymph nodes and increased microbiota diversity. However, Selemax was more effective clinically and immunologically; it reduced disease prevalence more sharply, increased the proportion of CD103+ dendritic cells expressing high levels of PD-L1 in mesenteric lymph nodes and reduced the intestinal inflammatory process more strongly than S. cerevisiae. These results suggest a clear gut-brain axis modulation by selenized S. cerevisiae and suggest their inclusion in clinical trials.

Multiple sclerosis is an autoimmune disease that affects the myelinated central nervous system (CNS) neurons and triggers physical and cognitive disabilities. Conventional therapy is based on diseasemodifying drugs that control disease severity but can also be deleterious. Complementary medicines have been adopted and evidence indicates that yeast supplements can improve symptoms mainly by modulating the immune response. In this investigation, we evaluated the therapeutic potential of Saccharomyces cerevisiae and its selenized derivative (Selemax) in experimental autoimmune encephalomyelitis (EAE). Female C57BL/6 mice submitted to EAE induction were orally supplemented with these yeasts by gavage from day 0 to day 14 after EAE induction. Both supplements determined significant reduction in clinical signs concomitantly with diminished Th1 immune response in CNS, increased proportion of Foxp3 + lymphocytes in inguinal and mesenteric lymph nodes and increased microbiota diversity. However, Selemax was more effective clinically and immunologically; it reduced disease prevalence more sharply, increased the proportion of CD103 + dendritic cells expressing high levels of PD-L1 in mesenteric lymph nodes and reduced the intestinal inflammatory process more strongly than S. cerevisiae. These results suggest a clear gut-brain axis modulation by selenized S. cerevisiae and suggest their inclusion in clinical trials.
Multiple sclerosis (MS) is considered the most common inflammatory and demyelinating disease of the central nervous system (CNS). Despite some controversy, there is a large body of evidence that MS is an autoimmune pathology, mainly mediated by specific immune response against the myelin sheath that wraps around nerve fibers 1 . T cell subsets, specially T helper (Th)-1 and Th17, cytotoxic T and innate cells as macrophages and microglia have been deeply implicated in MS pathobiology 2 . Details of MS immunopathogenesis have been revealed with the help of experimental autoimmune encephalomyelitis (EAE), which is an animal model induced mainly in mice by immunization with CNS antigens associated with complete Freund´s adjuvant and frequently with pertussis toxin 3 .

Yeast supplementation increases Foxp3 + T cells in inguinal lymph nodes.
To understand the mechanism involved in this protection, we initially characterized the CD4 + T cell composition in the spleen and inguinal lymph nodes. The number of total leukocytes was significantly lower in spleen samples from Selemax supplemented EAE mice (p = 0.0369) (Fig. 4a). On the other hand, after the stimulation with MOG 35-55, splenocytes from EAE, EAE/S. cerevisiae and EAE/Selemax groups produced similar amounts of IFN-γ, TNF-α, IL-2, IL-6 and IL-10 ( Fig. 4b-d,f,g). The only exception was IL-17A that was produced in lower levels by Selemax supplemented animals than the S. cerevisiae supplemented ones (p = 0.0064) (Fig. 4e). Even though the proportion of T helper cells (Th -CD3 + CD4 + ) in inguinal lymph nodes was similar in all experimental groups (Fig. 4i), the proportion of regulatory T cells (Treg -CD25 + Foxp3 + ) was similarly upregulated by both, S. cerevisiae (p = 0.0044) and Selemax (p = 0.0027) (Fig. 4h,j). Nonetheless, the ratio between Treg and Th cells was higher in the Selemax group (p = 0.0069) (Fig. 4k).
Selenization increases the tolerogenic potential of S. cerevisiae. Considering that mesenteric lymph nodes (MLN) play a major role as an inducing site of intestinal tolerance 30,31 , our next step was to evaluate the effect of these yeasts in parameters indicative of immunological tolerance. The total leukocyte number and the percentage of CD3 + CD4 + T cells in MLN was similar among groups (Fig. 5a,b). Nonetheless, Selemax supplementation enhanced the proportion of CD4 + CD25 + Foxp3 + cells (Fig. 5c) and the ratio between Treg and Th cells (Fig. 5d) in comparison to EAE group (p = 0.0191 and p = 0.0617). In addition, the evaluation of dendritic cell (DC -CD11c + MHCII High ) subsets (Fig. 5e) and the expression of PD-L1 in the subset of tolerogenic Figure 1. Experimental design. C57BL/6 mice received daily oral doses of Saccharomyces cerevisiae or selenium-enriched yeast (Selemax) for 14 days starting on EAE induction (a). Body weight and clinical score were assessed daily for 15 days. On the 15th day, which corresponded to the acute disease phase, fresh stools were collected for metagenomic sequencing and the animals were then euthanized (arrow). Samples from gutbrain axis and peripheral lymphoid organs were collected to evaluate immunological parameters (b). The mRNA expression and cytokine levels were assessed on CNS (brain plus spinal cord) whereas inflammatory infiltration and demyelination were assessed on lumbar spinal cord. Inflammatory infiltration, mRNA expression and cytokine levels were assessed on small intestine and the proportion of T cells and dendritic cells subsets were evaluated in mesenteric lymph nodes. Peripheral immunological parameters as T cell and cytokine levels were evaluated in inguinal lymph nodes and spleen, respectively. This figure was created using images from Servier Medical Art Commons Attribution 3.0 Unported License (https ://smart .servi er.com) and image was modified with Adobe Photoshop v. 22 www.nature.com/scientificreports/ DC (CD11b -CD103 + ) ( Fig. 5f) indicated that selenization can increase the tolerogenic potential of S. cerevisiae. Concerning CD11b + DC subsets, the proportion of CD11b + CD103 -DC ( Fig. 5g) was significantly reduced in both EAE/S. cerevisiae (p = 0.0363) and EAE/Selemax (p = 0.0044) groups and CD11b + CD103 + DC (Fig. 5h) was reduced only in EAE/Selemax (p = 0.0141). The percentage of CD11b -CD103 + tolerogenic DC was higher in both EAE/S. cerevisiae (p = 0.0379) and EAE/Selemax (p = 0.0095) groups in comparison to EAE group (Fig. 5i). Despite this, tolerogenic DC from EAE/Selemax expressed a significantly higher amount of PD-L1 (p = 0.0197) in comparison to EAE group (Fig. 5j). The same profile was observed in non-EAE mice (control group). Only Selemax supplementation increased the percentage of CD4 + CD25 + Foxp3 + cells (p = 0.0190) and CD11b -CD103 + tolerogenic DC (p = 0.0502) (Supplementary Fig. 1a and 1b, respectively). (j) and IL-10 (k) were assessed in the supernatants. Statistical analysis was performed by one-way ANOVA and subsequent Tukey's test (a,b,d-j) and Pearson correlation (c). All data were expressed as the mean ± SD and statistical differences were represented by *p < 0.05 and **p < 0.01. Two independent experiments were combined, n = 7-8 mice/group. The graphs were created using GraphPad Prism v. 8 www.nature.com/scientificreports/ Representative gating strategy to assess regulatory T cells (Treg -CD25 + Foxp3 + ) in inguinal lymph nodes (h). Total cells from inguinal lymph nodes were collected and analyzed by flow cytometry to determine the proportion of T helper cells (Th -CD3 + CD4 + ) (i) and Treg in total Th cells (j). The proportion of T subsets was used to calculate the ratio between Treg and Th cells (k). Statistical analysis was performed by one-way ANOVA and subsequent Tukey's test. All data were expressed as the mean ± SD and statistical differences were represented by *p < 0.05 and **p < 0.01. Two independent experiments were combined, n = 7-10 mice/group. The graphs were created using GraphPad Prism v. 8 www.nature.com/scientificreports/ S. cerevisiae and Selemax determine distinct changes in the microbiota. Modulation of intestinal immunity by S. cerevisiae has been mainly attributed to its ability to change microbiota 32 . Based on this evidence, we used high-throughput sequencing of faecal samples, expressed by operational taxonomic units (OTUs), to investigate the effect of supplementation with S. cerevisiae and its selenized derivative on the composition of microbiota. The β-diversity metric, represented by the principal coordinate analysis (PCoA), indicated that microbiota was different among the three groups (Fig. 6a). Permutational multivariate analysis of variance (PERMANOVA) revealed that the bacterial community structure in EAE group was different from EAE/S. cerevisiae (p = 0.029) or EAE/Selemax (p = 0.011) groups and between the two supplemented groups (p = 0.014). The proportion of CD11b + CD103 -(g), CD11b + CD103 + (h) and CD11b -CD103 + (i) DC subsets in total DC. The MFI of PD-L1 was evaluated in CD103 + CD11b -DC subset (j). Statistical analysis was performed by one-way ANOVA and subsequent Tukey's test. All data were expressed as the mean ± SD and statistical differences were represented by *p < 0.05 and **p < 0.01. Two independent experiments were combined, n = 6-8 mice/group. The graphs were created using GraphPad Prism v. 8 www.nature.com/scientificreports/ Regarding α-diversity metrics, were evaluated the absolute abundance, richness, evenness, Shannon-Wiener diversity index and Faith's phylogenetic diversity index. EAE group had increased absolute abundance (absolute number of bacteria in the sample) and richness (the number of OTUs) in comparison to CTL group (p = 0.0167 and p = 0.0098) (Fig. 6b,c). At the family level these differences were mainly related to the detection of OTUs in Christensenellaceae, Coriobacteriaceae, Erysipelotrichaceae, Lachnospiraceae, Ruminococcaceae and S24-7 in EAE group but not in CTL group; whereas at genus level, the same was observed concerning Adlercreutzia, Anaerostipes, Coprococcus, Dorea, Oscillospira and Ruminococcus genera. The supplementation with Selemax decreased the absolute abundance in EAE mice (p = 0.0084) ( Fig. 6b) but did not alter the richness. Additionally, the supplementation with Selemax increased evenness (p = 0.0309) and both supplementations, S. cerevisiae and Selemax, increased Shannon-Wiener diversity index (p = 0.0279 and p = 0.0234, respectively) in comparison to CTL group (Fig. 6d,e). The phylogenetic diversity (Faith's phylogenetic diversity index) did not differ between EAE and supplemented groups (Fig. 6f). Lastly, we analyzed the relative abundance of taxonomic groups in the supplemented EAE-mice. At the phylum level (Fig. 6g), Bacteroidetes relative abundance in EAE mice was higher than EAE mice supplemented with S. cerevisiae (p = 0.0044) or Selemax (p = 0.0135) whereas the Firmicutes phylum was lower in EAE group in comparison to EAE/S. cerevisiae (p = 0.0110) or EAE/Selemax (p = 0.0163) groups. At the family level, we observed an increase in the relative abundance of Lactobacillaceae in EAE/S. cerevisiae (p = 0.0017) and Clostridiaceae in EAE/Selemax group (p = 0.0176) in comparison to EAE group (Fig. 6h). When differences were evaluated at the genus level, only S. cerevisiae supplementation triggered a significant microbiota modulation. The Statistical analysis was performed by one-way ANOVA and subsequent Tukey's test (b-f) and two-way ANOVA and subsequent Tukey's test (g-i). All data were expressed as the mean ± SD and statistical differences were represented by *p < 0.05 and **p < 0.01; # p < 0.05 between EAE and EAE/S. cerevisiae; & p < 0.05 between EAE and EAE/Selemax. One experiment representative of two independent experiments (a) and two independent experiments were combined (b-i), n = 4-9 mice/group. PCoA was created using QIIME2 v.2020.2 (https :// view.qiime 2.org/), the graphs were created using GraphPad Prism v. 8 www.nature.com/scientificreports/ genus Turicibacter was reduced (p = 0.0024), whereas the Lactobacillus one was increased (p < 0.0001) in EAE/S. cerevisiae in comparison to EAE group (Fig. 6i).

Selemax controls intestinal inflammation better than S. cerevisiae. Considering that both yeasts
increased the diversity of microbiota and altered the relative abundance of some bacterial groups, we evaluated whether microbiota modulation was related to different degrees of intestinal inflammation. When compared to the healthy control group, the villus of EAE animals were enlarged, presented reduction in length and displayed a significant increase in mononuclear and polymorphonuclear cells, suggesting the presence of neutrophils infiltrating the lamina propria as well as the submucosa. In addition, an influx of such polymorphonuclear cells was detected in the gut lumen of EAE mice (arrows in the second panel), which is indicative of severe inflammation. S. cerevisiae and Selemax-supplemented mice presented a significant reduction in the cell infiltrate in the lamina propria however, the intestinal structure was still altered in comparison to non-EAE group (Fig. 7). Selemax supplementation was slightly more effective in attenuating gut inflammation. Notably, we also detected the presence of yeasts in both treated groups (as indicated in the upright insert in the third panel) suggesting, therefore, their gut accumulation after treatment. Gut-homing of myelin-specific Th17 has been described as a pivotal step for the development of EAE 33 . Assays involving gene expression in intestinal samples and cytokine production by gut explant cultures were then performed to evaluate if supplementation was affecting the local level of T cell subsets, including Th17 and Treg, and of immunoregulatory molecules. A significant reduction was observed in the expression of Tbx21 (p = 0.0151) (Fig. 8a), Rorc (p = 0.0338) (Fig. 8b) and Gata3 (Th2) (p = 0.0015) (Fig. 8c) mRNA levels in Selemax supplemented EAE-mice compared to non-supplemented EAE-animals. No differences were observed among the three experimental groups concerning the expression of genes related to Treg cells as Foxp3 (Fig. 8d), Tgf (Fig. 8e) and Ido (indoleamine 2,3-dioxygenase) (Fig. 8i). The expression of pattern recognition receptors (PRRs) involved in the interaction with S. cerevisiae was also evaluated. Both products significantly reduced the expression of Tlr4 (p = 0.0451) (Fig. 8h) but not any other PRRs measured, like Clec7 (Fig. 8f) and Tlr2 (Fig. 8g).
Corroborating with the findings of gene expression, gut explant immersion cultures from EAE mice supplemented with Selemax showed reduced production of IFN-γ (p = 0.0654) (Fig. 8k), IL-17A (p = 0.0074) (Fig. 8m) and IL-10 (p = 0.0513) (Fig. 8j) when compared to the untreated EAE group. Similarly, the production of IFN-γ (p = 0.0309), TNF-α (p = 0.0369) and IL-17A (p = 0.0027) were significantly reduced in S. cerevisiae supplemented group (Fig. 8k-m), compared to the EAE group. No differences were observed in IL-6 ( Fig. 8n) and IL-2 ( Fig. 8o) production in both supplemented groups in comparison to EAE animals. Gene expression and cytokine production were used to calculate the fold change of data from supplemented groups in relation to the EAE group (Fig. 8p). This overview of fold change clearly revealed that Selemax supplementation improved the benefits of S. cerevisiae in the context of EAE. As can be observed, the selenized yeast decreased most inflammatory parameters and simultaneously increased tolerogenic indicators in the intestinal environment, as Foxp3 (p = 0.0010) and Ido (p = 0.0161) (Fig. 8p).

Discussion
In the current study, we evaluated the therapeutic potential of Saccharomyces cerevisiae and its selenized derivative named Selemax to control EAE development. Both yeast preparations were clearly able to control clinical disease evolution. To the best of our knowledge, this is the first report supporting the usage of S. cerevisiae and Selemax to control experimental MS. This seems a relevant contribution because yeast-rich diets or yeast supplementation have been used for a long time worldwide, in a series of settings, without side effect records. S. cerevisiae is one of the most well-known yeast being predominantly employed for food fermentation and used for centuries in bread-making and production of alcoholic beverages such as wines and beers 34 . Besides the use of wild type strains, engineered S. cerevisiae strains are being largely employed as commercially profitable cell systems for application in bioprocesses aiming the synthesis of biopharmaceuticals and therapeutic proteins 35,36 . Another application of S. cerevisiae is its enrichment with essential elements for the organism's homeostasis, such as iron 37 and selenium (Se) 38,39 . The enrichment of S. cerevisiae yeast with Se has been previously reported www.nature.com/scientificreports/ and increased its immunomodulatory potential. The oral administration of S. cerevisiae protected mice against mucositis and reduced the inflammatory response in the intestinal tissue and this effect was potentiated when it was enriched with Se 40 . Se has antioxidant and anti-inflammatory properties which are particularly relevant in CNS diseases 41 . Only a few reports have addressed the relevance of Se levels in MS patients. Significantly lower serum levels were described in one study 42 whereas no alterations were found in two other investigations, in comparison to healthy controls 43,44 . In the present work, no differences were observed concerning serum Se concentration between non-EAE and EAE-mice. As expected, only Selemax supplemented group presented higher serum level of Se. Of note, these high Se levels were not associated with an evident renal or hepatic toxicity. We hypothesized that the Se-enriched yeast could be a safe approach to achieve modulation of both, microbiota and autoimmune response in the context of EAE. Our initial findings already suggested that selenization was improving S. cerevisiae ability to control EAE development; this was indicated by the lower paralysis prevalence in this experimental model. Clinical protection in both supplemented groups was associated with less inflammation and demyelination in the CNS. Besides, a more detailed analysis of the cells eluted from this tissue confirmed the lower inflammatory process in the Selemax supplemented EAE-mice. They presented a reduced number of leukocytes in the CNS, including Th1 cells, and produced less encephalitogenic cytokines. In addition, the level of TAU mRNA was statistically higher in the EAE/Selemax group. Tau belongs to the microtubule-associated proteins family, essential for microtubule assembly in neurons. Usually found in axons, the protein can also be translocated to the cell body and dendrites and the abnormal deposition of modified tau in the neurons is a common aspect of many neurodegenerative diseases called "tauopathies" 45 . The role of this protein has also been investigated in EAE/MS. Mice devoid of tau protein presented higher susceptibility to neuronal damage, suggesting that tau integrity is relevant to avoid axonal damage in EAE mice 46 . In MS patients, the lower levels of tau in the cerebrospinal fluid have been correlated with parenchymal brain loss 47 and may denote axonal degeneration 48 . In addition, we observed a lower expression of tgf in CNS samples from Selemax supplemented EAE-mice. It has been described that early production of TGF-β1 by glial cells creates a dangerous environment for the initiation of autoimmune inflammation whereas pharmacological inhibition of TGF-β signaling ameliorated the disease signs and reduced pathogenic T cells and IL-6 expression in the CNS 49 . The reduction of neuroinflammation and increased TAU expression in CNS observed in EAE-mice supplemented with Selemax support the enhanced protective effect of Se-enriched yeast in relation to non-enriched S. cerevisiae yeast and reinforce the relevance of our findings. Tgf (e), Clec7 (f), Tlr2 (g), Tlr4 (h) and Ido (i) were performed in small intestine samples. Gut explants of small intestine were cultured for 6 h and the production of IL-10 (j), IFN-γ (k), TNF-α (l), IL-17A (m), IL-6 (n) and IL-2 (o) were assessed in the supernatants. The gene expression and cytokine levels were used to calculate the fold change in relation to EAE group (p). Statistical analysis was performed by one-way ANOVA and subsequent Tukey's test (a-o) and two-way ANOVA and subsequent Tukey's test (p). All data were expressed as the mean ± SD and statistical differences were represented by *p < 0.05 and **p < 0.01; & p < 0.05 between EAE and EAE/Selemax. Two independent experiments were combined, n = 7-8 mice/group. The graphs were created using GraphPad Prism v. 8 www.nature.com/scientificreports/ The ensuing experiments were done to investigate if these distinct levels of protection were due to differential immunomodulatory ability on the spleen or the regional lymph nodes. Even though a few differences were observed between S. cerevisiae and Selemax supplemented mice concerning the number of spleen cells and proportion of Treg (CD25 + Foxp3 + ), they did not seem relevant enough to explain the most pronounced protective effect of selenized yeast. However, the data obtained from the MLN revealed a higher percentage of Treg and tolerogenic DC (CD11b -CD103 + ) in Selemax than in S. cerevisiae-supplemented mice. These findings are in consonance with literature data that supports the contribution of these cells to control inflammation in EAE and MS 20,50,51 . In addition, these tolerogenic DC expressed higher levels of PD-L1. The contribution of PD-L1 expression in DC as a mechanism to limit the initial activation of myelin-reactive CD4 + T cells was already reported in EAE 52 . Also, among four distinct DC subsets expressing CD103 and/or PD-L1 in MLN, only CD11b -CD103 + PD-L1 High DC subset was a higher inducer of Foxp3 + T cells 53 . These findings may sugest that selenization is acting, at least partially, by increasing the DC/Treg pathway. Findings of previous literature reports support this viewpoint. In an iodine-induced autoimmune thyroiditis model, the percentage of Treg cells and the expression of Foxp3 mRNA were increased by treatment with sodium selenite in drinking water 54 . As reviewed by Nettleford and Prabhu 55 , Se acts on intestinal immunity through upregulation of the nuclear hormone receptor peroxisome proliferator-activated receptor gamma leading to differentiation of Treg and inhibition of Th1 and Th17. Se also inhibits NF-κB, decreasing, therefore, the production of pro-inflammatory cytokines that are required for T helper cell differentiation.
To better understand the mechanism by which these two supplements were being protective and possibly help to elucidate the greater effectiveness of Selemax, we analyzed the faecal microbiota. The results clearly indicated that both supplements were able to increase microbiota diversity in EAE-mice, but only Selemax was able to reduce the enhanced absolute abundance associated with the disease. A diversified microbiota is generally considered beneficial to the health 56,57 ; conversely, the decline in microbiota diversity has been linked to the rise of modern diseases as obesity, asthma, food allergies, diabetes, inflammatory bowel disease, and cognitive disorders 58 . A low-diversity microbiota contributes to decreasing microbiota resilience and stability and changes the growth of some microbes relative to others 59 . Recently, a number of studies showing that MS patients and EAE mice exhibit gut dysbiosis disclosed the gut microbiome as a relevant player in the development of this disease [60][61][62][63][64] . In the present work, EAE development was associated with an increased absolute abundance and richness and the detection of some bacteria not observed in non-EAE mice. EAE mice exhibited dysbiosis associated, at least partially, to the detection of an increased abundance of Dorea and Ruminococcus genera, similarly to what is observed in patients with MS 62,63 . Considering that the disease causes dysbiosis related to bacterial abundance, we evaluated if yeast supplementation altered the relative abundance of bacterial community at different taxonomic levels. At the phylum level, both supplements increased Firmicutes and decreased Bacteroidetes phyla. Considering that the Bacteroidetes phylum has been related to an inflammatory profile 65 , the reduction of this taxon may have contributed to protection. A higher abundance of Lactobacillaceae and Clostridiaceae families, both from the Firmicutes phylum, was detected in faecal samples of S. cerevisiae and Selemax supplemented EAE groups, respectively. Considering that members of Lactobacillaceae and Clostridiaceae family were significantly decreased in MS patients 63 , their increase in supplemented EAE groups could be related to protection. Results observed in experimental diseases, including EAE, support this possibility. For instance, the oral administration of heat-killed Candida kefyr to EAE mice resulted in disease protection associated with higher Lactobacillales abundance and increased Treg proportion 20 . The direct role of Lactobacillus in inflammation has been linked to the differentiation of Treg cells in MLN of mice fed with L. casei 66 and related to autoimmunity suppression after remodelling microbiota with L. reuteri administration in Treg-deficient scurfy mice 67 and in EAE mice 21 . Similarly, the oral administration of 17 Clostridia strains to adult mice provided bacterial antigens and a TGF-β-rich environment that helped Treg cells expansion and differentiation and attenuated other experimental diseases, as colitis and allergic diarrhea 68 . The induction of Treg by Clostridia has been attributed to signals delivered by CD103 + DC to drive naïve CD4 + T cells to differentiate into Treg cells 69 . At genus level, only supplementation with S. cerevisiae increased Lactobacillus and reduced Turicibacter. Despite this, both supplements significantly increased Firmicutes phylum and families of this phylum, and, as mentioned above, members of these families have been related to the regulation of inflammation.
Having in mind that MLN are a pivotal site in intestinal immunity 70 , that DC traffic from the intestine to these lymph nodes 71 , and that Th1 cells migrate to the gut contributing to the inflammatory environment that favours Th17 differentiation during EAE development 72 , we evaluated the effect of supplementation on the intestinal inflammatory condition. The histopathological analysis revealed that the intestinal inflammation degree was lower in supplemented animals being this effect more evident in the EAE/Selemax group. This finding was reinforced by other assays performed with gut explants that showed lower levels of proinflammatory gene expression and of cytokine production in supplemented EAE mice. Some of these parameters, as IL-6, IL-10, IL-17A, Tbx21, Rorc, and Gata3 were more downmodulated in EAE/Selemax mice than in EAE/S. cerevisiae ones. The decreased expression of toll-like receptor (TLR)-4 in both supplemented groups seems meaningful. Considering that the expression of Tlr4 is low at the small intestine in a steady-state condition 73 , we believe that most of its elevated expression in EAE mice could be due to the presence of inflammatory cells, mainly CD4 + T lymphocytes and γδT cells. This possibility is supported by data that indicates a T cell homing stage during EAE development 33,72 . In the EAE model, the absence of TLR4 solely in CD4 + T cells significantly decreased disease signs; the authors also found that Tlr4 -/-γδT cells presented defective IFN-γ and IL-17 production 74 . In addition to the reduction of innate and adaptive components of the immune system, Selemax also increased the fold change of Foxp3 and Ido mRNA expression in the small intestine. This seems interesting because it has been shown that encephalitogenic cells migrating into the gut-associated lymphoid tissue can be restrained there, expelled into gut lumen and even transformed into Treg which are effective against CNS autoimmunity 75,76 . CNS and gut are connected via the gut-brain axis with bidirectional interactions between the CNS and the gastrointestinal tract, including microbial Scientific Reports | (2020) 10:22190 | https://doi.org/10.1038/s41598-020-79102-7 www.nature.com/scientificreports/ community 77,78 . These bidirectional interactions enable the brain to influence gastrointestinal functions as well as the immune functions through the cross-talk amongst the gut, the nervous system, and the immune system 79 .
In summary, S. cerevisiae increased Lactobacilaceae whereas selenium-enriched yeast (Selemax) increased Clostridiaceae and both supplements increased Treg proportion in MLN and controlled EAE development. However, the prevalence of the disease was lower in Selemax supplemented group and only these mice showed a higher PD-L1 expression on CD103 + DC in MLN and a significant reduction of IL-17A in the CNS and IL-17A and Rorc in the small intestine. In this context, Selemax supplementation seems to be a formulation worthwhile to be considered for MS therapy once it is endowed with modulatory properties towards the microbiota and the immune system. It is important to mention, however, that Se status can be affected by several factors, including genetic variation. This polymorphism affects mainly the selenoproteins involved in its metabolism 80 . These genetic differences must be checked and considered in any preventive or therapeutic measure involving supplementation with Selemax or other organic Se sources.
Altogether our results show that selenization of S. cerevisiae improved its ability to control EAE development. These data also suggest that a significant part of this protective effect seems to occur at the intestinal level, including the generation of a tolerogenic scenario at the MLN and small intestine and modifications of the gut microbiota. In this context, this selenized yeast is a strong candidate to be trialed, in the future, as a CAM in MS patients.

Methods
Experimental design. C57BL/6 mice were allocated into six experimental groups: control (CTL)-healthy animals (n = 6); S. cerevisiae-healthy animals daily supplemented with Saccharomyces cerevisiae (n = 8); Selemax-healthy animals daily supplemented with selenized S. cerevisiae (n = 8); experimental autoimmune encephalomyelitis (EAE)-animals immunized with myelin oligodendrocyte glycoprotein peptide (MOG  to develop encephalomyelitis (n = 16); EAE/ S. cerevisiae-immunized with MOG and daily supplemented with S. cerevisiae (n = 18), and EAE/Selemax-immunized with MOG and daily supplemented with selenized S. cerevisiae (n = 21). Four independent experiments were combined. Supplementation was orally delivered by gavage for 14 days, once a day. CTL and EAE groups received the diluent in the same conditions. Body weight and clinical score were evaluated daily for 15 days. On the 15th day, that corresponded to the acute phase of EAE, fresh faecal samples were collected for metagenomic sequencing. The animals were then euthanized with an overdose of ketamine-xylazine and blood samples were collected through cardiac puncture. Mice were then perfused intracardially with sterile Phosphate Buffered Saline (PBS) before samples collection. Lumbar spinal cord was used for histopathological analysis and real time polymerase chain reaction (PCR) assays and brain plus cervical and thoracic spinal cord eluted cells were cultured to assess cytokine production. Spleen cells were cultured to analyze peripheral cytokine production and inguinal lymph node cells were used to evaluate T cell subsets by flow cytometry. T and dendritic cell subsets present in mesenteric lymph nodes were also analyzed by flow cytometry. Pieces (4-5 mm) of small intestine (jejunum and ileum) were freshly used for cultures to assess cytokine production, fixed for histopathological analysis or frozen for further PCR assays. The experimental design is summarized in Fig. 1  Animals. Six-week-old female C57BL/6 mice were purchased from a specific pathogen-free facility at the  82,83 : no signs (0); limp tail (1); hind legs weakness (2); partially paralyzed hind legs (3); complete hind legs paralysis (4); and complete paralysis or death (5). Body weight loss was calculated considering the difference between the initial weight (at the time of EAE induction-day 0) and the one at the acute disease phase (day 15). (YSC 11,111) and Selemax, a selenium-enriched S. cerevisiae-YSC111, were provided by Biorigin (Zilor, Lençóis Paulista, SP, Brazil). Selemax production was performed according to Oraby et al. 39 . Both supplements were diluted in sterile water and administered once a day to the animals by gavage for 14 days. The daily dose of both supplements was 21.34 mg. In the case of Selemax this dose contained 50 µg of Se that, according to Vieira et al. 84 , was therapeutically effective in experimental arthritis. The composition of both supplements is described in Table 2 Gut explant immersion cultures. Cytokine production by immune cells from lamina propria was checked by using gut explant immersion cultures which were done with small fragments from jejunum and ileum. This approach was adapted from Bareiss et al. 85 and Randall et al. 86 . Briefly, the entire jejunum and ileum were aseptically removed and placed in sterile Petri dishes with cold HBSS containing 1% of antibiotic/antimycotic (Sigma-Aldrich). The cleaned sections (4 -5 mm), devoid of faeces and Peyer's patches, were sectioned longitudinally to expose the mucosal surface and placed in a 48-well plate. Each well contained two explants immersed in 0.5 ml of RPMI 1640 medium supplemented with 10% heat-inactivated fetal calf serum (Gibco, United States), 4 mM of l-glutamine (Sigma-Aldrich), 1% of sodium pyruvate (Sigma-Aldrich), 1% of nonessential amino acids (Sigma-Aldrich) and 2% antibiotic/antimycotic (Sigma-Aldrich). The cultures were incubated in a humidified incubator at 37 °C and 5% CO 2 for up to 6 h after which the supernatants were stored at − 20 °C for further cytokine quantification.

Supplementation with S. cerevisiae and Selemax. Saccharomyces cerevisiae
RT-qPCR analysis. RNA   Flow cytometry. The percentage of regulatory T (CD3 + CD4 + CD25 + FoxP3 + ) and DC (CD11c + MHCII High ) were evaluated in total cells from the MLN by flow cytometry. Samples were incubated with PerCP-Cy5.5 labelled anti-mouse CD3 (clone 145-2C11), PE-Cy7 labelled anti-mouse CD4 (clone GK1.5) and APC labelled anti-mouse CD25 (clone PC61.5) for T cell subsets panel or with PerCP-Cy5.5 labelled anti-mouse CD11b (clone M1/70), APC-Cy7 labelled anti-mouse CD11c (clone N418), APC labelled anti-mouse MHCII (clone M5/114.15.2), PE labelled anti-mouse CD103 (clone 2E7) and PE-Cy7 anti-mouse PD-L1 (clone MIH5) for DC panel during 30 min at 4 °C. Intracellular Foxp3 transcription factor was detected using anti-mouse/rat Foxp3 Staining Set PE (FJK-16s) (eBiosciences, San Diego, CA, USA) according to manufacturer's instructions. After staining, the cells were washed, resuspended in flow cytometry buffer and fixed in paraformaldehyde 1%. Flow cytometry was performed using a FACS Canto II (BD Biosciences) from Institute of Biosciences (UNESP, Botucatu, SP, Brazil), the compensation of acquisition data was performed using OneComp eBeads Compensation Beads (eBiosciences) and data were analyzed with FlowJo software (BD Biosciences). Gating strategy for T cell and DC subsets was showed in Supplementary Fig. 2 Metagenomic DNA was amplified using the V4 region of the 16S rRNA gene that was amplified with regionspecific primers that included the Illumina flowcell adapter sequences. The reverse amplification primer also contained a twelve-base barcode sequence that supports pooling of different samples for Illumina sequencing. After cluster formation, the amplicons were sequenced with custom primers. These sequencing primers were designed to be complementary to the V4 amplification primers to avoid sequencing of the primers, and the barcode was read using a third sequencing primer in an additional cycle. Amplification primers (forward GTG CCA GCMGCC GCG GTAA and reverse GGA CTA CHVGGG TWT CTAAT) were adapted from Caporaso et al. 88 protocol to include nine extra bases in the adapter region of the forward amplification primer that supported paired-end sequencing. Libraries were normalized and pooled to 4 nM based on qPCR values. Pooled samples were denatured and diluted to a final concentration of 10 pM with a 20% PhiX (Illumina) control. Sequencing was performed using the MiSeq Reagent Kit V3 in the Illumina MiSeq System. All 47 samples were multiplexed and sequenced on the MiSeq using 2 × 150 bp paired-end sequencing. Analysis was performed with QIIME2 version 2020.2 89 . Briefly, low quality sequences and chimeras were discarded with DADA2 pipeline with standard parameters. Alfa rarefaction and beta diversity were performed with a sampling of 100,000 reads from each sample. Taxonomy was assigned in QIIME against the GreenGenes ribosomal RNA gene database (GreenGenes 13_8 99% OTUs from 515F/806R region).

Statistical analysis.
With the exception of the microbiota, normality of data was analyzed by Shapiro-Wilk test and significant outliers were determined by Grubbs' test. Comparisons between three groups were made by One-Way ANOVA or Two-Way ANOVA multiple comparisons followed by Tukey's multiple comparisons test or Holm-Sidak's multiple comparisons test and the disease prevalence was analyzed by Log-rank (Mantel-Cox) test. Data analysis and figures were made in GraphPad Prism 7 (GraphPad Software Inc., San Diego, California, United States) and values of p < 0.05 were considered statistically significant. Permutational multivariate analysis of variance (PERMANOVA) tests of the gut microbiota on the unweighted UniFrac distances. Alpha diversity indices were calculated using PAST 3.26 software (https ://www.softp edia.com/get/Scien ce-CAD/PAST.shtml ).