Alterations in Circulating Fatty Acid Are Associated With Gut Microbiota Dysbiosis and Inflammation in Multiple Sclerosis

Background: Butyric acid (BA) is a short-chain fatty acid (SCFA) with anti-inflammatory properties, which promotes intestinal barrier function. Medium-chain fatty acids (MCFA), including caproic acid (CA), promote TH1 and TH17 differentiation, thus supporting inflammation. Aim: Since most SCFAs are absorbed in the cecum and colon, the measurement of BA in peripheral blood could provide information on the health status of the intestinal ecosystem. Additionally, given the different immunomodulatory properties of BA and CA the evaluation of their serum concentration, as well as their ratio could be as a simple and rapid biomarker of disease activity and/or treatment efficacy in MS. Methods: We evaluated serum BA and CA concentrations, immune parameters, intestinal barrier integrity and the gut microbiota composition in patients with multiple sclerosis (MS) comparing result to those obtained in healthy controls. Results: In MS, the concentration of BA was reduced and that of CA was increased. Concurrently, the microbiota was depleted of BA producers while it was enriched in mucin-degrading, pro-inflammatory components. The reduced serum concentration of BA seen in MS patients correlated with alterations of the barrier permeability, as evidenced by the higher plasma concentrations of lipopolysaccharide and intestinal fatty acid-binding protein, and inflammation. Specifically, CA was positively associated with CD4+/IFNγ+ T lymphocytes, and the BA/CA ratio correlated positively with CD4+/CD25high/Foxp3+ and negatively with CD4+/IFNγ+ T lymphocytes. Conclusion: The gut microbiota dysbiosis found in MS is possibly associated with alterations of the SCFA/MCFA ratio and of the intestinal barrier; this could explain the chronic inflammation that characterizes this disease. SCFA and MCFA quantification could be a simple biomarker to evaluate the efficacy of therapeutic and rehabilitation procedures in MS.


INTRODUCTION
Multiple sclerosis (MS) is a chronic disease of the central nervous system (CNS) characterized by demyelination and mediated by auto-reactive immune processes directed against neural tissues. Inflammation drives and accompanies MS, and this disease has repeatedly been shown to be characterized by augmented activation of TH1 and TH17 lymphocytes and downregulation of Treg cells. The etiopathogenesis of MS is still only partly understood, but a number of recent publications suggest that alterations of the intestinal microbiota play a role, even if the underlying mechanisms are unclear (1)(2)(3)(4). Compared to healthy individuals, the gut microbial ecosystem of MS patients is indeed enriched with Methanobrevibacter (5), an archaeal genus associated with inflammatory and autoimmune diseases (6), and Akkermansia, a mucin-degrading bacterium capable of favoring pro-inflammatory T lymphocyte responses,7 but is depleted of Parabacteroides distasonis (7), bacterial species that support T cell differentiation into Treg cells (8). An additional body of data indicates that the relative abundance of the Lachnospiraceae and Ruminococcaceae families, which include short-chain fatty acid (SCFA) producers, as well as Bacteroides fragilis, Butyricimonas, and Prevotella is considerably reduced in MS (4,5,(9)(10)(11). The hypothesis that these alterations play a role in the pathogenesis of MS has recently been reinforced by results showing that dietinduced changes in the gut microbiota resulted in a significant increase in Lachnospiraceae that correlated with an augmented expression of anti-inflammatory immune cells and a significant reduction of the relapse rate, as well as of the expanded disability status score (EDSS) (12).
Butyric, acetic and propionic acids are the main SCFAs, i.e., carboxylic acids containing 2-5 carbon atoms, which are produced in the proximal colon by bacterial fermentation of non-digestible carbohydrates (13,14). Butyric acid (BA), in particular, is endowed with immunomodulatory properties that are mediated by histone deacetylase (HDAC) inhibition (15)(16)(17) or through the activation of the metabolite-sensing G protein-coupled receptors GPR41, GPR43, and GPR109A (18)(19)(20)(21)(22)(23). Thus, SCFA-mediated stimulation of Treg cell differentiation is driven by HDAC inhibition, while IL-10 production is mediated by GPR41 or GPR43 activation (23). Notably, recent results have shown that SCFAs could mediate the activation of pro-inflammatory immune circuits as well, depending on the cytokine milieu and immunological context (23)(24)(25)(26). In contrast to SCFAs, medium-chain fatty acids (MCFAs) mostly derive from the diet, even if the liver contributes to their production via peroxisomal beta-oxidation of long-chain fatty acids. MCFAs control carbohydrate and lipid tissue metabolism as well as the production of mitochondrial energy (27) and antagonize the anti-inflammatory activities of SCFAs, as they favor TH1 and TH17 differentiation. The immunological roles of BA and MCFAs, particularly of caproic acid (CA), have been confirmed in animal models (28,29). In these models, BA has been shown to stimulate the maturation of fully functional Treg cells, and MCFAs have been confirmed to have an effect that antagonizes that of SCFAs, as they enhance TH1 and TH17 cell differentiation (29). In addition to its direct anti-inflammatory effect, BA also downregulates inflammation secondarily to its pivotal role in maintaining the integrity of the gastrointestinal (GI) epithelial barrier through regulation of mucus production and tight junction expression (13,30,31). Therefore, BA prevents alterations of intestinal permeability and the consequent translocation of lipopolysaccharide (LPS) from the gut lumen to the peripheral blood. As repeatedly demonstrated in the setting of other diseases, such as HIV infection (32), increased amounts of LPS in peripheral blood drive chronic inflammation by binding to toll-like receptor (TLR) 4, an activatory receptor expressed by lymphoid cells.
Since over 90% of SCFAs are absorbed in the cecum and colon, and only 5-10% of them are excreted in the feces (33)(34)(35)(36), the measurement of BA in peripheral blood could provide information on the health status of the intestinal ecosystem (37). Additionally, given the many beneficial functions of BA as an immune-modulator, its monitoring could work as a simple and rapid biomarker of disease activity and/or treatment efficacy in MS. Herein, we analyzed serum concentrations of BA and CA, as well as immune parameters and parameters of integrity of the GI epithelial barrier, and profiled the gut microbiota in MS patients and healthy controls (HC).

Individuals Enrolled in the Study
Thirty-eight patients with a diagnosis of relapsing-remitting (RR) or secondary-progressive (SP) Multiple Sclerosis (MS) (20 females and 18 males; median age = 47 years, IQ = 42-57) who are followed by the Multiple Sclerosis Rehabilitation Unit of the Don Carlo Gnocchi Foundation in Milan, Italy, were enrolled in the study. Inclusion criteria were age >18 years and disease stability for >6 months prior to enrollment. Main exclusion criteria were: (1) changes in disease-modifying treatment (DMD) in the 6 months prior to enrollment; (2) use of corticosteroids in the 6 months prior to enrollment; (3) presence of significant co-morbidities, including arterial hypertension, cerebrovascular disorders, heart or pulmonary diseases, diabetes, endocrine, gastrointestinal, or psychiatric diseases. Patients were following a free diet (western type) as determined by an expert dietician through interviews. At the time of enrolment and of biological sampling none of the patients was using antibiotics; none of the patients had used any antibiotic in the 3 months preceding enrolment and biological sampling. Stable disease was diagnosed on the basis of brain and spinal cord magnetic resonance imaging (MRI) with gadolinium showing no areas of enhancement at the time of enrolment. Median disease duration was 19 years (range: 15-24 years); disability level, as assessed by the median Kurtze Expanded Disability Status Scale (EDSS) score was 5.3 (range: 3-6). Finally, a group of 38 sex-and agematched HC (median age = 48 years; range 33-62; 20 females and 18 males) was enrolled as well in the study. HC were also following a free, western style diet and had not taken antibiotics or probiotics in the 3 months prior to sampling. Italian controls have been specifically selected to reduce any bias related to lifestyle or, more generally, to the geographical effect, for which the strongest associations with microbiota variation have recently been shown (38). The Ethics Committee of the Don Carlo Gnocchi Foundation approved the study protocol; all the enrolled subjects signed an informed consent. The clinical and demographic characterization of the MS patients enrolled in the study is presented in Supplementary Table 1.

Serum
Serum was collected in vacutainer tubes containing serum separator (Becton Dickinson & Co., Rutherford, NJ, USA), centrifuged at 3,000 rpm for 10 min, and stored at −80 • C until use.

Analysis by LC-MS/MS of Serum Fatty Acids
Butyric and caproic acids were extracted and analyzed by LC-MS/MS according to the protocol described in Dei Cas (39). Briefly, serum (50 µl) was deproteinized by isopropanol (100 µl) and fatty acids derivatized with nitrophenylhydrazine. The extract was analyzed on an HPLC Dionex 3000 UltiMate system (Thermo Fisher Scientific, MA, USA) coupled to a tandem mass spectrometer AB Sciex 3200 QTRAP (AB Sciex, Milan, Italy) operated by multiple reaction monitoring under negative electrospray ionization.

Blood Sample Collection and Cell Separation
Whole blood (10 ml) was collected in vacutainer tubes containing ethylenediaminetetraacetic acid (EDTA) (Becton Dickinson & Co.). Peripheral blood mononuclear cells (PBMCs) were separated on lympholyte separation medium (Cedarlane, Hornby, Ontario, CA, USA) and washed twice in PBS at 1,500 rpm for 10 min; viable leukocytes were determined with Bio-Rad TC20 Automated Cell Counter (Bio-Rad Laboratories, Hercules, CA, USA).

Flow Cytometry Analysis
PBMCs were analyzed to identify regulatory T cells (Tregs: CD4+CD25++FOXP3+), TH1 (CD4+IFNγ+), TH17 (CD4+IL-17+) and inducible regulatory T cells (TH R 1: CD4+IL-10+) using a Beckman-Coulter GALLIOS flow cytometer equipped with a 22 mW Blue Solid State Diode laser operating at 488 nm and with a 25 mW Red Solid State Diode laser operating at 638 nm, and interfaced with Kaluza analysis software. Two hundred thousand cells were acquired and gated on lymphocyte FSC and SSC properties. Isotype control or single fluorochrome-stained preparations were used for color compensation.

Gut Microbiota Analysis
Total bacterial DNA was extracted from stool samples of 35 MS patients, as previously described (40). The V3-V4 hypervariable region of the 16S rRNA gene was PCR-amplified using the 341F and 785R primers (41) with Illumina overhang adapter sequences as previously reported (40) PCR products of about 460 bp were purified using a magnetic bead-based system (Agencourt AMPure XP; Beckman Coulter) and indexed by limited-cycle PCR using Nextera technology. Indexed libraries, further cleaned up as described above, were pooled at equimolar concentrations, denatured and diluted to 6 pmol/l. Sequencing was performed on an Illumina MiSeq platform using the 2 × 250 bp protocol, according to the manufacturer's instructions (Illumina, San Diego, CA, USA). Sequence reads were deposited in the National Center for Biotechnology Information Sequence Read Archive (NCBI SRA; BioProject ID PRJNS633233).

Bioinformatics and Statistical Analysis
Quantitative data were not normally distributed (Shapiro-Wilk test) and are therefore summarized as median and interquartile range (IQR; 25th and 75th percentiles). Comparisons between groups were performed using a two-tailed Mann-Whitney test for independent samples. Kruskal-Wallis analysis of variance was utilized for each variable. Statistical correlations between the immunological parameters and BA, CA, or the BA/CA ratio were investigated by the Spearman correlation coefficient and 95% confidence limits performed by Fisher's Z transformation. Data analysis was performed using the MEDCALC statistical package (MedCalc Software bvba, Mariakerke Belgium).
As for microbiota analysis, paired-end reads were processed using a pipeline combining PANDAseq (42) and QIIME 2 (43). High-quality reads were filtered and clustered into Amplicon Sequence Variants (ASVs) at 99% similarity through an openreference strategy performed with DADA2 (44). Singleton ASVs were discarded and chimeras were identified using ChimeraSlayer (45) and then filtered out. Taxonomy was assigned using the vsearch classifier (46) against the Greengenes database as a reference (release May 2013). 16S rRNA gene sequencing data of MS patients were compared to publicly available data of age-and sex-matched healthy Italian subjects [20 subjects: MG-RAST ID 17761 (47), 15 subjects: MG-RAST ID 7058 (48)]. Genus-level community composition was generated for all combined cohorts. Alpha diversity was measured using the Shannon and Simpson indices (estimating evenness and richness), while beta diversity was computed based on Jaccard similarity and visualized on a Principal Coordinates Analysis (PCoA) plot. The significance of the separation between the study groups was tested by a permutation test with pseudo-F ratios using the function adonis in the R package, vegan (49). Bar plots were built using the packages made4 (50) and vegan. Kendall rank correlation test was used to assess the associations between genus-level relative abundances and EDSS, levels of T lymphocytes, LPS, I-FABP and fatty acids in MS patients. Statistics was performed using R Studio 1.0.44 on R software version 3.3.2 (https://www.r-project.org/) implemented with the packages stats and vegan. P-values were corrected for multiple comparisons using the Benjamini-Hochberg method when appropriate. A p-value ≤ 0.05 was considered statistically significant.

Serum Concentration of Butyric and Caproic Acids
The concentration of butyric acid (BA) and caproic acid (CA) was analyzed in the serum of all individuals enrolled in the study, by liquid chromatography coupled to mass spectrometry (LC-MS). The results showed that BA was significantly reduced (median: HC = 907 ng/ml; MS = 752 ng/ml; p < 0.0001) whereas CA was significantly increased (median: HC = 181 ng/ml; MS = 863 ng/ml; p < 0.0001) in MS compared to HC. Accordingly, the BA/CA ratio was significantly reduced in MS (ratio: 0.9) compared to HC (ratio: 5; p < 0.0001). These data are shown in Figure 1.

T Lymphocyte Functional Subpopulations
BA regulates Treg lymphocyte development whereas CA is known to support TH1 and TH17 differentiation. To verify possible associations between BA and CA and these lymphocyte subsets Treg, TH1 and TH17 were measured in freshly isolated and unstimulated PBMCs from all patients and controls. The results indicated that Treg lymphocytes (CD4+/CD25 high /Foxp3+) were decreased in MS (median: 0.1%) compared to HC (median: 0. 8%; p < 0.0001). On the contrary, both TH1 (CD4+/IFNγ+) (median: HC = 0.01%; MS = 0.1%; p < 0.0001) and TH17 (CD4+/IL-17+) (median: HC = 0.0%; MS = 0.3%; p < 0.0001) T lymphocytes were significantly increased in MS compared to HC. CD4+/IL-10 positive T lymphocytes were also measured in all individuals; these cells were not significantly different when MS and HC were compared. These results are shown in

Microbial Translocation and Gut Barrier Permeability
Since BA is known to maintain the integrity of the gastrointestinal (GI) epithelial barrier, parameters of GI permeability were subsequently measured. The results showed that the plasma concentration of lipopolysaccharide (LPS) (median: HC = 0.3 Eu/ml; MS = 0.7 Eu/ml; p = 0.001) and intestinal fatty acid-binding protein (I-FABP) (median: HC = 336 pg/ml; MS = 715 pg/ml; p < 0.0001) were significantly increased in MS compared to HC (Figure 4). LPS translocates from the intestinal lumen to the peripheral circulation when the integrity of the GI barrier is altered; I-FABP is released into circulation in case of enterocyte damage and intestinal ischemia.    These results therefore suggest the presence of damage to the gut barrier in MS patients.

Gut Microbiota Layout in MS
The gut microbiota (GM) of MS patients was finally profiled and compared with that of age-and sex-matched HC to verify whether the differences in SCFAs and MCFAs observed in these patients could be the result of changes in the gut microbial ecosystem. The 16S rRNA gene sequencing yielded a total of 1,520,873 high-quality reads, ranging from 23,759 to 80,331 per sample, clustered into 3,221 Amplicon Sequence Variants (ASVs). No differences in alpha diversity were observed between MS patients and HC, except for a slightly lower Shannon index value in the former (Supplementary Figure 1). On the other hand, when stratifying MS patients according to the severity of the disease, we found significantly lower diversity in secondary-progressive MS (SPMS) compared to relapsingremitting MS (RRMS) and HC (inverse Simpson index; p ≤ 0.01, Wilcoxon test) (Figure 5A). Although not significant, a similar trend was observed for the Shannon index (p ≤ 0.1). The Principal Coordinates Analysis (PCoA) of inter-individual variation, based on the Jaccard similarity index, revealed a significant separation between MS patients and HC (p < 1 × 10 −4 , permutation test with pseudo-F ratios), while no significant differences were detected between MS subtypes (i.e., SPMS and RRMS) (p = 0.4) (Figure 5B and Supplementary Figure 1). In line with recent literature (7), the GM composition of MS patients showed increased relative abundance of Akkermansia, as well as a depletion of Parabacteroides (Figure 6A). Compared to HC, in the GM of MS patients we found a dramatic depletion of bacterial genera belonging to the Lachnospiraceae family. In particular, the proportions of the well-known SCFA producers Roseburia, Coprococcus, and Blautia were reduced by 4.5, 3.4, and 2 times, respectively, in the GM of MS vs. HC (p ≤ 0.02, Wilcoxon test) (Supplementary Figure 1). On the other hand, the MS microbiota showed a 7.4 and 2.4-fold increase in Collinsella and [Eubacterium], respectively (p ≤ 0.05) (Supplementary Figure 1). By focusing the analysis by disease subtype (Figure 6B), we found some commonalities, including the decrease in the SCFA producers mentioned above, but also specific microbial signatures. In particular, SPMS patients were found to be characterized by greater relative abundance of Akkermansia and Collinsella (p = 0.002), and a decrease in Dorea (p = 0.003). On the other hand, Parabacteroides was significantly under-represented in RRMS patients (p = 0.003), who also showed a reduction in [Ruminococcus] and Lachnospira (p ≤ 0.0008), a marked increase in Streptococcus (p = 0.006), and an increasing trend for [Eubacterium] (p = 0.07).
Correlations between the relative abundances of bacterial taxa and EDSS, levels of T lymphocyte subpopulations, LPS and I-FABP in MS patients were next specifically sought ( Figure 7A). Interestingly, the plasma levels of LPS and I-FABP were negatively correlated with the relative abundance of the well-known probiotic genus, Bifidobacterium (respectively, p = 0.003 and 0.04, tau = −0.363 and −0.252, Kendall rank correlation test). Inverse correlations were also found between the levels of CD4+/IL-17+ T lymphocytes and the proportions of the SCFA-producing genera, Coprococcus and Ruminococcus (p = 0.02, tau = −0.277 and −0.28, respectively). In contrast, a positive correlation was found between the levels of this T cell subpopulation and Prevotella (p = 0.02, tau = 0.295), whose relative abundance showed a 2.7-fold increase in MS compared to HC (2.9 ± 0.9 vs. 1.1 ± 0.6%), even if in the absence of statistical significance. Furthermore, levels of CD4+/IL10+ T lymphocytes were negatively correlated with Akkermansia (p = 0.04, tau = −0.275). As for fatty acids, a negative correlation was found between [Eubacterium] and the serum BA levels (p = 0.004, tau = −0. 35). No significant correlations were found between the proportions of bacterial taxa and EDSS.
With specific regard to disease subtypes (Figure 7B), it is worth noting that in SPMS patients we found a positive correlation between CD4+/IFNγ+ T lymphocytes and [Ruminococcus] (p = 0.05, tau = 0.427), a well-known mucindegrading gut microbe to be associated with Crohn's disease as well as other inflammatory disorders (51). A positive correlation was also found between the serum BA/CA ratio and Lachnospira (p = 0.04, tau = 0.466), consistent with the known ability of this bacterial genus to produce BA. As for RRMS, as expected based on available literature (52), Roseburia was found to positively correlate with the levels of CD4+/IL10+ T lymphocytes (p = 0.04, tau = 0.341). Another positive correlation was found between Parabacteroides and the serum levels of I-FABP (p = 0.03, tau = 0.366).

DISCUSSION
Multiple sclerosis (MS) is a chronic and progressive autoimmune disease characterized by inflammation whose etiopathogenesis is still unclear. Among the different hypothesis suggested to be involved in the pathogenesis of MS the inside-out model states that the disease is initially triggered by oligodendrocytes injury and/or death, possibly caused by oxidative stress (53). This would result in the release of myelin antigens into the periphery that would activate autoreactive B and T lymphocytes; these immune cells would migrate back into the central nervous system (CNS) and initiate inflammation (54). Inflammation is indeed observed throughout disease progression and is driven both by immune cell infiltration of the CNS and, directly, by CNS cells that are activated against the tissue damage that accumulates during the natural history of the disease.
Alterations in the composition of the gut microbiota have recently been held to be at least partially responsible for MS-associated inflammation (55). One additional set of results showed that butyrate-producing bacteria are reduced in MS (12). These bacteria are responsible for fermenting non-digestible carbohydrates in the proximal colon to generate SCFAs, including butyric acid (BA). The observation that BA drives Treg differentiation (56,57) and maintains the integrity of the GI epithelial barrier (58)(59)(60)(61)(62), thus preventing microbial translocation and LPS-driven triggering of TLR4-mediated signaling, underlines the immunological importance of these bacteria and their byproducts, SCFAs, in down-modulating inflammation. BA can nevertheless induce inflammatory response as well (23,(63)(64)(65), and an elevated BA concentration in a proinflammatory milieu was shown to possibly result in stimulation of IFNγ-producing CD4+ T cells by inhibiting histone deacetylase activity (63), activating GPR41 or GPR43 (23,64), or promoting survival of CD8+ T cells (65).
Even more recently, MCFAs, which are derived from the diet and can be generated by hepatic peroxisomal beta-oxidation of long-chain fatty acids, were shown to support the differentiation of TH1 and TH17 and to suppress that of Treg lymphocytes (29). Caproic acid (CA), in particular, seems to be prototypically endowed with such pro-inflammatory properties, which result from the activation of p38 MAPK signaling (29). Notably, CA can also be directly produced by some of the bacteria that form the microbiota. Prevotella, in particular, was shown to be able to generate CA. It is interesting to observe that these bacteria were indeed increased in MS patients, suggesting its possible role in the increased CA production seen in these patients.
Herein, we analyzed serum concentrations of BA and CA in a group of MS patients and observed that BA was significantly reduced and CA significantly augmented compared to healthy controls (HC). This peculiar SCFA/MCFA alteration correlated with the immunological profile expected to be supported by such a variation, i.e., an increase in TH1 and TH17 and a decrease in Treg lymphocytes, as it was observed in MS patients. The reduction in BA concentration had previously been described in MS patients with chronic progressive disease (66), and both effector and IL-10+ T cells were shown to be induced by SCFAs and dietary fiber in experimental autoimmune encephalomyelitis (EAE), the animal model of MS. These results suggested that modulation of SCFA production could represent a novel adjunct to therapeutic approaches in autoimmune CNS diseases (66).  (14)(15) cytosolic protein found in mature enterocytes of the small and large intestine that is released when the integrity of the cell membrane is compromised (58). These results therefore suggest that the reduction of BA seen in MS is likely also responsible for impaired GI permeability (62). We do not know whether the augmented I-FABP concentrations observed in MS are the consequence of functional alterations of the small or of the large intestine; as SCFAs are produced in the large intestine it is nevertheless tempting to hypothesize that such I-FABP increments are the consequence of damages that affect the large intestine as a result of dysbiosis. Few studies address I-FABP in the setting of MS. Some data indicated that I-FABP as well as ileal bile acid binding protein (IBABP) are increased in MS patients, (67,68), whereas other data did not observe any differences between patients and controls (69). Results herein lend support to the idea that intestinal damages, as indicated by alterations of LPS and I-FABP, are indeed present in MS; further analyses will be needed to sort out these discrepancies The observed alteration of the SCFA/MCFA ratio could thus support MS-associated inflammation with at least two distinct mechanisms: a direct one, which is secondary to the skewing of T lymphocyte functional subset differentiation toward those subsets that support inflammation, and an indirect one driven by alterations in GI tract integrity and microbial translocation. As it has been extensively shown in the setting of another chronic and inflammatory disease, HIV infection (32), increased plasma levels of LPS lead to systemic immune activation and inflammation that is the consequence of LPS binding to TLR4, an activation protein expressed on different cell types. Notably, in EAE, the animal model of MS, alterations of the intestinal microbial ecosystem have also been shown to result in impaired permeability of the blood-brain barrier and activation of microglia and astrocytes (58).
Consistent with the alterations in the SCFA/MCFA ratio and supporting the above assumptions, the gut microbial ecosystem of MS patients was characterized by a dysbiotic layout that included: (1) a decrease in SCFA producers belonging to the Lachnospiraceae family, i.e., Roseburia, Coprococcus, and Blautia; (12) (2) an increased relative abundance of Akkermansia, a mucin degrader capable of inducing pro-inflammatory responses in human PBMCs and mono-colonized mice; (5, 7) (3) the depletion of Parabacteroides, a bacterial genus with antiinflammatory properties shown to stimulate the maturation of IL-10-expressing CD4+CD25+ T cells in humans (7) and IL-10+Foxp3+ Tregs in mice; (7) and (4) an increased proportion of Collinsella, which has recently been shown to correlate with the production of the pro-inflammatory cytokine IL-17A and with altered gut permeability in MS (9). It is interesting to note that when we stratified patients based on the MS subtype (i.e., RRMS and SPMS), two distinct types of dysbiosis emerged. In particular, reduced microbial biodiversity and an overabundance of Akkermansia and Collinsella were specifically observed in patients with SPMS compared to HC. As briefly discussed above, these bacterial genera have already been associated with MS and shown to exacerbate its symptoms, because of their ability to induce pro-inflammatory responses and compromise the integrity of the intestinal epithelial barrier, thus contributing to aggravate a condition of chronic inflammation (5,7,9). On the other hand, the ecosystem configuration observed in patients with RRMS showed a different, somewhat less pronounced dysbiosis, with levels of diversity comparable to those of HC and mainly characterized by an overabundance of Streptococcus and depletion of Parabacteroides. It should be noted that this dysbiotic profile had already been found in cohorts of RRMS patients and suggested to influence the population of T lymphocytes and promote inflammation (4,9,70). In particular, the abundance of Streptococcus spp. has been shown to correlate positively with the proportion of TH17 cells while negatively with Tregs (71), thus representing a potential key factor in the development and/or reactivation of the disease. CD4+/IL-17+ T lymphocytes were negatively correlated with Coprococcus, a bacteria known to be depleted in MS patients (72) and recognized as a beneficial commensal taxon, but correlated positively with Prevotella, whose role on the host physiology is instead more controversial. Some studies point in fact to its pro-inflammatory role in autoimmune diseases (73)(74)(75), while a very recent work indicates that Prevotella histicola can suppress EAE as efficiently as the disease-modifying drug Copaxone (76). This stresses the need for species-level analysis in future microbiome-based studies. On the other hand, the CD4+/IL-10+ T cell subpopulation was found to negatively correlate with Akkermansia, an organism that, as mentioned above, can exacerbate MS symptoms, probably either directly, by shifting immune responses toward a TH1 phenotype, or indirectly, by interacting with other bacteria and reducing their ability to drive Treg differentiation (77). In line with the available literature (52), the CD4+/IL-10+ T cell subpopulation was positively correlated with Roseburia, even if only in RRMS patients. On the other hand, in SPMS patients, CD4+/IFNγ+ T lymphocytes were found to positively correlate with [Ruminococcus], consistent with its ability to degrade mucus and induce inflammatory responses (51).
As for fatty acids, a negative correlation was found between serum BA levels and the relative abundance of [Eubacterium], a bacterial genus significantly increased in MS and comprising potential opportunistic pathogens that could affect the mucus layer (75,78). On the contrary, a positive correlation was found between the serum BA/CA ratio and Lachnospira in SPMS patients, consistent with the known ability of this bacterial genus to produce BA. As expected, a negative correlation was also observed between the proportions of the health-promoting genus Bifidobacterium and serum levels of I-FABP and LPS. Several in vitro and in vivo studies in fact show that probiotics, including Bifidobacterium spp., are associated with improved barrier function and reduced metabolic endotoxemia (79,80).
Overall, it is thus tempting to suggest that the disbyotic profile of the MS-associated gut microbiota justifies the SCFA/MCFA alterations described herein, possibly shedding light on the genesis of the inflammatory milieu that characterizes and accompanies MS. If this assumption is correct, then therapeutic/dietary interventions aimed at restoring a physiological (i.e., eubiotic) gut microbial ecosystem should be considered in MS therapy. The results obtained in the EAE animal model support this idea, since the administration of a multistrain probiotic has been shown to result in prophylactic and therapeutic efficacy that was associated with a reduced degree of inflammation (81,82). Even more interestingly, in the same animal model, a SCFA-rich diet has been shown to increase the frequency of peripheral Tregs and ameliorate the clinical course of the disease, whereas mice given MCFA-or LCFA-rich diets showed an aggravated disease progression (29).

CONCLUSIONS
Taken together, our results suggest a rationale explaining the etiology of MS-associated inflammation, possibly supporting the idea that alterations in the gut microbial ecosystem play a role in inflammatory autoimmune conditions, including MS, and reinforce the idea that interventions aimed at restoring microbiota eubiosis could be integrated into current therapeutic and rehabilitative strategies for MS. These are preliminary results and our study has limitations mostly stemming from the relatively small sample size and the heterogeneity of the enrolled patients in regard to treatment status. Additionally, we showed the presence of gut barrier damage and bacterial translocation in MS patients compared to controls, but the limited number of enrolled individuals did not allow us to verify possible differences in these parameters in patients with diverse disease phenotypes. Further studies will clarify these issues.

DATA AVAILABILITY STATEMENT
The datasets generated for this study are available on request to corresponding author. Sequence reads were deposited in the National Center for Biotechnology Information Sequence Read Archive (NCBI SRA; BioProject ID PRJNS633233).

ETHICS STATEMENT
The studies involving human participants were reviewed and approved by The Ethics Committee of the Don Carlo Gnocchi Foundation. The patients/participants provided their written informed consent to participate in this study.

AUTHOR CONTRIBUTIONS
MC designed the research, drafted, and edited the manuscript. MS designed and performed research and drafted the manuscript. LM, LP, VR, and Ad'A selected patient groups, collected clinical data, organized patient enrolment, and collected blood sample. MB, ST, and PB performed gut microbiota experiments, analyzed the data, and drafted the manuscript. GR, EC, MDC, and RP performed chemical experiments, analysed, and wrote chemical data. FP, FL, and IM performed immunological experiments, analyzed results, and prepared the figures. All authors contributed to the article and approved the submitted version.