The gut microbiota composition is shaped by disease activity and individual treatment responses in patients with multiple sclerosis

Multiple sclerosis (MS) is a chronic autoimmune disorder of the central nervous system, with both animal and human studies highlighting a pivotal role for the gut microbiota in its pathogenesis. In this cross-sectional study, we investigated the potential role of gut microbiota in treatment response by analyzing its composition using 16S rRNA sequencing in treatment-naïve patients and those receiving disease-modifying therapies (interferon-beta (IFN-β), fingolimod, or cladribine), compared to healthy controls (HC). We also analyzed differences in gut microbiota composition and the serum levels of biomarkers associated with microbial translocation and inflammation based on treatment response. We found that individuals with clinically isolated syndrome (CIS) and treatment non-responders (NR) had significantly different alpha and beta diversity compared to HC. This effect was present in both IFN-β and fingolimod treatment. Individuals treated with cladribine had significantly different alpha and beta diversity regardless of the treatment outcome. The main differences in abundances in CIS and NR were found in bacteria that produce short-chain fatty acids. These patients also had significantly higher levels of lipopolysaccharide-binding protein and mannose-binding lectin compared to HC suggesting the compromised gut barrier function in multiple sclerosis leading to higher level of microbial translocation in these patients. In summary, we found that the treatment influences gut microbiota. Similar profile of gut microbiota and higher levels of molecules associated with microbial translocation were observed in patients with active disease (CIS and NR), suggesting the higher permeability of their gut barrier leading to pro-inflammatory tunning of their immune system.

GRAPHICAL ABSTRACT

Graphical Abstract. Created with BioRender.com.

1 Introduction

Multiple Sclerosis (MS) is a chronic inflammatory disease characterized by demyelination, axonal damage and neuronal loss driven by autoreactive T and B lymphocytes and macrophages (1). An altered composition of the gut microbiota - dysbiosis, has been observed in patients with both active and progressive MS, as well as in pediatric MS, regardless of treatment status (2–7). The presence of certain microbial taxa has been associated with an increased risk of relapse or MRI activity (7). Changes in gut microbiota composition can alter interactions with the gut-associated lymphoid tissue (GALT), promote intestinal inflammation and barrier disruption, and increase intestinal permeability (“leaky gut”), thereby facilitating translocation of microbes and their products and amplifying systemic immune activation. In this way, the microbiota can contribute to the initiation and propagation of autoimmune processes in distant organs such as the brain (8). Among the best-studied pathways, bacterial short-chain fatty acids (SCFAs) promote the differentiation and function of regulatory T cells (Tregs) and have been associated with improvement in MS disease course. A reduction in SCFA-producing bacteria belonging to Clostridia clusters IV and XIVa has been observed in persons with multiple sclerosis (PwMS), and this decrease may contribute to an imbalance between pro- and anti-inflammatory immune activity in these patients (9, 10). In addition, other microbiota-derived metabolites, including tryptophan metabolites, bile acids and phytoestrogens, are increasingly recognized as important modulators of CNS-directed immunity in MS, although these pathways were beyond the scope of the present work. Currently, there is no cure for MS. The goal of treatment is to prevent disease relapses and the progression of neurological deficits. The current treatment strategy for MS is to start the DMT as early as possible, once a diagnosis of MS or clinically isolated syndrome (CIS) as the first demyelinating event affecting the CNS has been established. First-line treatments of low to moderate efficacy include injectable IFN-β and glatiramer acetate, oral teriflunomide and dimethyl fumarate and selective sphingosin-1-phosphate (S1P) agonists such as ponesimod and ozanimod. Second-line or high-efficacy therapies (HET) represent an ever-growing group of drugs with different mechanisms of action in MS. These include drugs that block immune cell trafficking, such as natalizumab and the non-selective S1P agonist fingolimod (FIN), pulse immune-reconstitution therapy with cladribine (CLA) and alemtuzumab, or anti-CD20 therapy ocrelizumab and ofatumumab. HET is intended for escalation of treatment in MS patients who fail on first-line treatment or as an early option for treatment-naïve patients at high risk of rapid disease progression and poor prognosis (11).

Fingolimod acts as a non-selective S1P functional antagonist. By blocking the S1P signaling pathway, it prevents lymphocyte egress from the lymph nodes and migration to target organs. Another effect of natural S1P pathway blocking could be the stabilization of the gut barrier by reinforcing the intestinal barrier integrity (12).

Cladribine is an antimetabolite that induces an initial reduction in both T and B lymphocyte counts, followed by their repopulation and reconstitution of normal effector function of the immune system (13). Cladribine is also used for the treatment of celiac disease. It has been shown to have a positive effect on the gut immune system (14).

IFN-β was the first Disease Modifying Drug approved for MS treatment, but its mechanism of action remains not fully understood. In addition to decreasing the production of interferon gamma, it has a direct anti-proliferative effect on T cells and influences the secretion of regulatory cytokines (15). Endogenous IFN-β appears to promote an anti-inflammatory response in the gut, reduce inflammation by suppressing IL-1b production and inflammasome activity, and maintain the immune system homeostasis (16).

Although numerous studies have investigated patients on different MS treatments, no specific microbiome composition pattern has been consistently associated with particular DMT. Furthermore, it remains unknown, whether the treatment response correlates with distinct microbiome composition compared to treatment failure. In our study, we therefore combined 16S rRNA profiling with measurements of serum pro-inflammatory molecules and SCFA concentrations to link taxonomic shifts and systemic inflammatory tone to one key microbial metabolic axis. Therefore, the aims of our study were 1) to analyze the gut microbiome composition in individuals with CIS and individuals treated with different types of DMTs and compare them with healthy controls (HC); 2) to find out differences in gut microbiota based on the treatment response and 3) to complement these results with the analysis of serum biomarkers associated with microbial translocation and inflammation.

2 Methods

2.1 Study cohort

Patients and healthy controls (HC) were recruited from a single academic hospital MS center between November 2018 and December 2023. Inclusion criteria were diagnosis of relapsing-remitting multiple sclerosis (RRMS) or individuals with clinically isolated syndrome (CIS), self-identified as white, and age between 18 and 60 years. Exclusion criteria were other autoimmune disease, known cancer or other conditions (e.g., gut disorders, psychiatric and mental disorders) that could affect the gut microbiota. For healthy subjects, exclusion criteria also included a use of antibiotics within three months prior to sampling. All study participants provided written informed consent. This study was approved by the Ethics Committee of General University Hospital in Prague 15/19 Grant AZV VES 2020 VFN.

A total of 143 patients and 81 HC were enrolled, including 47 individuals with CIS and 96 patients with RRMS diagnosed according to the 2017 McDonald diagnostic criteria, who had received disease modifying therapy (DMT) for at least two years. RRMS patients were treated either with IFN-β (38 patients), FIN (37 patients) or CLA (21 patients). Neurological examination and Expanded Disability Status Scale (EDSS) assessment were performed at the first visit, at which time patients provided fasting blood and stool samples. Clinical data, including age, sex, disease duration and EDSS were collected (Table 1). All patients underwent a baseline magnetic resonance imaging (MRI) scan of the brain and cervical spine, with follow-up scans conducted annually to assess the number of new or expanding T2 lesions. A relapse was defined according to the 2017 McDonald diagnostic criteria. The number of relapses, progression in EDSS, and presence of new or enlarging white matter lesions on MRI within the two years prior to study entry were considered to assess NEDA-3 status (No Evidence of Disease Activity, i.e., no relapses, no new/enlarging white matter lesions on MRI, and stable EDSS). Patients who met the criteria for NEDA-3 were classified as treatment responders (R group), while those who did not meet these criteria were classified as treatment non-responders (NR group).

Table 1

Table 1. Summary of anthropometric and clinical parameters from patients with multiple sclerosis and healthy controls.

Stool and blood serum samples from healthy controls and both individuals with CIS and RRMS patients on DMTs were collected as previously described (17). All samples were stored at −80°C until processing.

2.2 Serum biomarkers analysis

Biomarkers related to gut barrier function and inflammatory response or produced in response to microbial translocation were analyzed and quantified in serum using a commercial enzyme-linked immunosorbent assay (ELISA) (Table 2).

Table 2

Table 2. The list of quantified biomarkers in sera.

2.3 Gut microbiota analysis

The gut microbiota was analyzed as previously described (17). Briefly, DNA was isolated from the stool samples using the ZymoBIOMICS DNA Miniprep Kit (Zymo Research, Irvine, CA, USA) according to the manufacturer’s protocol. The V3 and V4 regions of the bacterial 16S rRNA gene were amplified by specific primers (341F GTCCTACGGGNGGCWGCAG and 806R GGACTACHVGGGTWTCTAAT) using the KAPA HiFi HotStart Ready Mix (Roche, Penzberg, Germany), as follows: initial denaturation step 3 min at 95°C followed by 25 cycles at 95°C for 30 s, 55°C for 30 s, 72°C for 30 s with a final elongation step at 72°C for 5 min. PCR products were checked with QIAxcel advanced capillary electrophoresis (QIAgen, Hilden, Germany). Triplicates of the amplicons were pooled and normalized with the SequalPrep™ Normalization Plate Kit (Thermo Fisher Scientific, Waltham, MA, USA), concentrated (Eppendorf centrifugal vacuum concentrator), and purified with DNA Clean & Concentrator Kit (Zymo Research). PCR amplification negative controls, extraction and sequencing positive controls (mock communities; ZymoBIOMICS Microbial Community Standard in linear and logarithmic form (MOCK 6300 and MOCK 6310 Zymo Research, USA) were processed in a similar manner. The amplicon libraries were then ligated with sequencing adapters using the KAPA HyperPlus Kit (Roche), pooled in equimolar concentrations, and sequenced 2x300 bp paired-end read using the MiSeq platform (Illumina, San Diego, CA, USA).

2.4 Sequencing data processing and statistics

In the first step, sample demultiplexing, primer detection, and trimming were performed using Skewer (18). Low-quality reads (expected error rate per paired-end read > 2) were then eliminated. DADA2 was used to denoise the quality-filtered reads and quantify 16S rRNA Amplicon Sequence Variants (ASVs) in each sample (19). Chimeric ASVs were detected and eliminated using UCHIME and the gold.fna reference database (https://drive5.com/uchime/gold.fa) (20). Taxonomical assignment of non-chimeric ASVs was conducted using the RDP classifier with an 80% confidence threshold and the Silva database (release 138) as a reference (21). Chloroplasts and mitochondria, as well as sequences that could not be assigned to a bacterial phylum, were considered contaminants of the diet or sequencing artifacts and were excluded from all downstream analyses. Sequences from technical duplicates were merged for each sample. The ASV abundance matrix (i.e., the number of ASV reads in each sample), ASV sequences, their taxonomic annotations, and phylogeny were merged into a single database along with sample metadata using the phyloseq package in R (R Core Team 2020) (22). To standardize sequencing depth, we rarefied the ASV table for alpha and beta diversity analysis. Rarefaction depth was set to the sample size of the minimal sequencing depth, i.e., where the majority of species were observed within a given number of samples. The alpha diversity was expressed as the Shannon diversity index, observed ASVs, and Faith’s Phylogenetic Diversity and compared using linear regression model adjusted for gender as additional factor. Principal Coordinate Analysis (PCoA) was performed to investigate differences in microbiota composition between samples. We used relative ASV abundance (Bray-Curtis index) and UniFrac (23–26). To evaluate the differences in beta diversity between groups, we used the Permutational Multivariate Analysis of Variance Using Distance Matrices (PERMANOVA) with a vegan package (27). Taxonomical analysis was performed using the microViz package, and relative abundances were calculated and presented (28). Associations between clinical metadata and microbial features were tested in R using the MaAsLin2 package after TSS normalization and LOG transformation of data (29). The ggplot2 package was used for graphical visualization of the data (30). Microbiome analysis sequencing data are available in the European Nucleotide Archive (ENA) (https://www.ebi.ac.uk/ena/browser/home) under accession number PRJEB85230.

The differences in serum biomarker levels between groups were tested using non-parametric Kruskal-Wallis test with Dunn’s multiple comparisons in GraphPad Prism statistical software (version 10, GraphPad Software, San Diego, CA, USA).

3 Results

3.1 Dysbiosis of the gut microbiota is a typical feature of both individuals with CIS and RRMS patients

We first compared the gut microbiota of individuals with CIS, patients with longer history of MS treatment (RRMS) and HC. We found that CIS and RRMS had significantly lower alpha diversity compared to HC, regardless of their treatment history (Figure 1A). And while both groups show significant changes in gut microbiota composition, beta diversity is similar in both groups of MS patients (Figure 1B).

Figure 1

Figure 1. Composition of gut microbiota in newly diagnosed CIS and RRMS patients. (A) Alpha diversity was assessed using the Shannon diversity index, Observed ASVs, Faith’s PD metric. Significant differences among groups were tested using linear regression model adjusted for gender. (B) Principal Coordinate Analysis (PCoA) of beta diversity based on Bray-Curtis dissimilarity metrics of HC, and CIS, RRMS patients. Each point represents one sample and statistical significance was tested with PERMANOVA. (*p<0.05, **p<0.01, ***p<0.001).

3.2 Individuals with low alpha diversity do not respond to therapy

When MS patients were divided into responders and non-responders, the significant decrease in alpha diversity was present only in non-responders (Figure 2A). But the gut microbiota beta diversity of either R and NR was significantly different from HC, as measured by the Bray-Curtis dissimilarity index (Figure 2B). There were no significant differences in beta diversity between responders and non-responders suggesting that while responders regain microbiome richness the disease signature in beta diversity is still present. The differences between R and NR are even documented on the relative abundances of the 18 most abundant taxa Figure 2C). When compared to HC with differential abundance analysis, each group has several specificities. In non-responders, four functionally protective microbial guilds are markedly depleted (i) core SCFA–producing bacteria (e.g. Coprococcus spp., Faecalibacterium spp., Lachnospiraceae ND3007 group, Ruminococcaceae CAG-352, Oscillospiraceae UCG-002 group Chybí závorka. (ii) bacteria contributing to overall richness and ecosystem stability (e.g., Christensenellaceae R7 group, Oscillospiraceae NK4A214 group), (iii) cholesterol-to-coprostanol and secondary-bile-acid converters (e.g. Ruminococcaceae CAG 352 group), and (iv) low-endotoxin Firmicutes (e.g. Erysipelotrichaceae UCG003) while the pathobiont genus Bacteroides expands. Many of bacteria significantly decreased in non-responders were found decreased also in CIS when compared to HC. Christensenellaceae R7 group was able to distinguish NR from newly CIS with AUC = 0.612 (p=0.04). Relative to healthy controls, responders displayed a distinct microbiota signature characterized by a significant decrease of mucin-degrading bacteria (e.g. genera Dorea, Collinsella and Prevotella 9), together with increased abundance of genus Streptococcus, which is capable of producing γ-aminobutyric acid (Figure 2D).

Figure 2

Figure 2. Gut microbial composition according the health status and treatment outcome (A) Alpha diversity was assessed using the Shannon diversity index, Observed spASVs, Faith’s PD metric. The significant differences among groups were tested using the linear regression model adjusted for gender. (B) Principal Coordinate Analysis (PCoA) of beta diversity based on Bray-Curtis dissimilarity metrics and weighted UniFrac, where each point represents patient or healthy control sample, tested with PERMANOVA. (C) Relative abundances of the top 18 most abundant taxa of gut microbiota at the genus level are presented as bar plots for each group. (D) The statistical differences in abundances of genera between the groups and HC were performed using the MaAsLin2 R package. (*p<0.05, **p<0.01, ***p<0.001).

3.3 The difference in dysbiosis between R and NR depends on their treatment

Next, we analyzed which MS treatment failure, IFN-β, FIN or CLA, is preceded with dysbiosis by comparing the gut microbiota composition between R and NR. We found that while alpha diversity is significantly decreased in NR to all treatments. It is similar to HC in responders to IFN-β (R-IFN) and fingolimod (R-FIN), but not in responders to CLA (R-CLA) (Figure 3A). This suggests that while dysbiosis may influence the therapeutic success in IFN or FIN, gut microbiota dysbiosis does not prevent CLA from therapeutic success. Similar situation is documented on beta diversity, when Bray-Curtis dissimilarity index differs from HC in all study groups except the responders to IFN-β and FIN (Figure 3B). At the taxonomic level, differences were already evident when the relative abundance profiles for the 18 most abundant taxa were plotted (Figure 3C, Supplementary Figure 1). When compared to HC, we found that R-IFN higher levels of Streptococcus spp. The relative abundance of Streptococcus spp. and Clostridium sensu stricto 1 was capable to distinguish R-IFN from NR-IFN with AUC = 0.768 (p=0.0047) and AUC = 0.696 (p=0.037), respectively. NR-IFN had lower levels of bacterial taxa Lachnospiraceae ND3007 group, Coprococcus, Escherichia-Shigella, Christensenellaceae R7 group, Lachnospiraceae CAG 56, Intestinimonas compared to HC (Figure 3D). R-FIN had significantly lower level of genus Clostridium sensu stricto 1 than HC. The Oscillibacter spp. was able to distinguish R-FIN from NR-FIN with AUC 0.719 (p=0.0034). NR-FIN had significantly lower levels of bacterial taxa Lachnospiraceae ND3007 group, Lachnospiraceae FCS020 group, Coprococcus, Eggerthellaceae, Eubacterium siraeum group, Intestinimonas, Victivallis and higher levels of Erysipelatoclostridium, Eggerthella, and Negativibacillus as compared to HC. Non-responders to cladribine (NR-CLA) had lower levels of Lachnospiraceae FCS020, Faecalibacterium spp., Lachnospiraceae CAG 56 and higher levels of genera Flavonifractor and Negativibacillus. Next to these findings, some taxa were enriched in both responders and non-responders within the same treatment arm (e.g. Oscillospiraceae in FIN; Lachnospiraceae ND3007 group in IFN), suggesting that some microbial shifts are driven by treatment exposure per se rather than by clinical response.

Figure 3

Figure 3. The composition of the microbiota is influenced more by the response to the therapy than by the medication. (A) Alpha diversity was assessed using the Shannon diversity index, Observed ASVs and Faith’s PD metric. The significant differences among groups were tested using the linear regression model adjusted for gender. *p<0.05, **p<0.01, ***p<0.001 compared to HC. (B) Beta diversity was evaluated using Bray-Curtis dissimilarities, weighted UniFrac and Principal Coordinate Analysis (PCoA) where each point represents patient or healthy control sample, tested with PERMANOVA. (C) Relative abundances of the top 18 most abundant taxa of gut microbiota at the genus level are presented as bar plots for each group. (D) The statistical differences in abundances of genera between the groups were performed using the MaAsLin2 R package. Statistical significance compared to HC is depicted by + (increased abundance) and – (decreased abundances).

3.4 Individuals with clinically isolated syndrome and treatment NR demonstrated increased levels of MBL and LBP

Increasing evidence suggests that there is intestinal barrier dysfunction in MS. This could lead to microbial translocation and an altered neuroinflammatory response. For this reason, we measured molecules associated with microbial translocation (Lipopolysaccharide-binding protein (LBP), soluble CD14 (CD14), Mannose-binding lectin (MBL)) and inflammation (Calprotectin, Osteopontin (OPN), IL-18, Lipocalin-2, Serum Amyloid A (SAA)). The individuals with CIS, had significantly higher levels of MBL and LBP, compared to HC and were able to distinguish them from HC with AUC = 0.726 (p=0.0002) and AUC = 0.710 (p=0.0007), respectively. (Figure 4). NR had significantly higher levels of LBP, CD14, MBL and OPN compared to HC. CD14 and OPN were, however, significantly elevated also in responders as compared to HC (Figure 4). Compared to patients receiving long-term treatment, the CIS patients had significantly higher calprotectin levels (Figure 4). In responders, we found significantly higher levels of IL-18 compared to active disease (CIS, NR) (Figure 4). The level of acute-phase apolipoprotein SAA was significantly higher in individuals with CIS, R and NR group as compared to HC, suggesting low-grade systemic inflammation irrespective of the treatment response status. In contrast, serum Lipocalin-2 concentrations did not differ significantly between HC, individuals with CIS, R or NR group (Figure 4). If we then focus on the individual therapies, patients with IFN-β treatment had higher levels of LBP and CD14 compared to HC regardless of response (Supplementary Figure 2). NR-IFN had higher levels of MBL compared to HC. Treatment with CLA or FIN did not lead to any changes in the levels of LBP, CD14 and MBL (Supplementary Figure 1). Interestingly, NR-FIN had significantly lower level of calprotectin and significantly higher level of OPN than HC. Patients with CLA treatment had significantly higher level of OPN than HC regardless of the treatment response (Supplementary Figure 2). The IL-18 was capable to distinguish R from NR with AUC = 0.697 (p=0.0046).When we used the level of IL-18 to distinguish responders and NR to individual DMTs, we found that IL-18 was able to distinguish treatment response in IFN-β (AUC = 0.810, p = 0.001), FIN (AUC = 0.781, p = 0.047), but had no power to distinguish treatment response in patients treated with CLA (AUC = 0.600, p = 0.56).

Figure 4

Figure 4. The levels of serum LBP, CD14, MBL, calprotectin, osteopontin and IL-18 in patients with CIS and RRMS patients according treatment response. Lipopolysaccharide-binding protein (LBP), soluble CD14 (CD14), Mannose-binding lectin (MBL), calprotectin, osteopontin, interleukin-18 (IL-18), Lipocalin-2, serum amyloid A (SAA). Statistical differences between groups were tested using the Kruskal-Wallis test with Dunn’s multiple comparisons *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001 (HC n=63, CIS n=35, R n=30, NR n=42).

3.5 Lower serum short chain fatty acids in individuals with CIS and non-responders

To assess the differences in abundances of SCFA producing bacteria we quantified short-chain fatty acids (SCFAs) directly in serum. Total serum SCFA concentrations were significantly reduced in individuals with CIS and NR compared with HC. We did not detect a statistically significant difference in serum SCFA levels between R and NR group; however, values in the R group were numerically closer to those of HC (Figure 5).

Figure 5

Figure 5. The levels of serum short chain fatty acids (SCFAs) in patients with CIS and RRMS patients according treatment response. Statistical differences between groups were tested using the Kruskal-Wallis test with Dunn’s multiple comparisons *p<0.05, **p<0.01 (HC n=22, CIS n=32, R n=25, NR n=42).

4 Discussion

In this study, we analyzed the gut microbiota aiming to find out the signatures associated with therapeutic success in MS. First, we found that patients with MS have decreased alpha diversity compared to healthy controls, regardless of if they were freshly identified (CIS) or suffer from MS for more than 2 years. This suggests that MS, and not MS treatment, induces dysbiosis. This is in agreement with multiple previous studies, which also observed a significant decrease in alpha and differences in beta diversity in MS patients compared to HC (31, 32). When taking the treatment status and treatment response into consideration, patients with CIS and NR had significantly lower alpha diversity compared to HC, as measured by Shannon diversity index, Observed taxa and Chao1 indices. Next to alpha diversity, we found significant differences in beta diversity between all study groups when compared to HC. An interesting and distinctive finding was observed when we stratified patients not only by treatment response (NEDA-3 status), but also by DMT type. Compared to HC, NR-IFN and NR-FIN showed significantly lower levels of alpha diversity. Both R-CLA and NR-CLA had decreased levels of alpha diversity. A similar pattern was seen in beta diversity, which was significantly different in NR to all three DMTs (NR-FIN, NR-IFN and NR-CLA), R-CLA and CIS patients compared to HC.

A recently published review identified 13 studies evaluating alpha and beta diversity in MS patients (33). Only five of these studies reported any significant differences. One cross-sectional study found significantly lower alpha diversity in patients with secondary progressive MS (SPMS) versus RRMS patients based on the Simpson diversity index, with a similar but not significant trend in Shannon’s diversity index (31). Another study using the Shannon’s diversity index reported significant reduction in alpha diversity in RRMS patients compared to HC (32). However, while no differences between total RRMS patients and HC were observed in another study, RRMS patients with active disease showed decreased species richness compared to those in remission and HC (34).

In our study, we demonstrated a significant difference in alpha and beta diversity between HC and patients with active MS (CIS and NR). However, no significant difference in alpha diversity was observed between patients with stable MS (R group) and HC. This observation was consistent regardless of the DMT used, except for patients treated with CLA, who showed significantly lower alpha diversity compared to HC.

Differences at the taxonomic level were primarily observed between patients with active MS (CIS and NR) and HC. We identified five bacterial taxa with lower abundance in both active MS groups compared to HC and one with higher abundance. Genus Coprococcus, Erysipelotrichaceae UCG003, Eubacterium coprostanoligenes group, Clostridium sensu stricto 1 and Oscillospiraceae of the NK4A214 group were reduced in both CIS and NR compared to HC, whereas Bacteroides were increased in these groups. This compositional shift is expected to lower butyrate availability, weaken epithelial tight-junction integrity, and raise systemic lipopolysaccharide exposure, collectively creating a pro-inflammatory milieu that diminishes the pharmacodynamic efficacy of MS disease-modifying therapies.

Coprococcus spp. is a butyrate-producing genus within the phylum Firmicutes, and its abundance is inversely correlated with several neuropsychological and neurodegenerative disorders, such as depression, language development disorder in children and Parkinson’s disease (35). A study investigating the role of SCFAs in MS found a dramatic depletion of bacterial genera belonging to the Lachnospiraceae family in MS patients compared to HC. Specifically, the proportions of the SCFAs producers Roseburia spp., Coprococcus spp., and Blautia spp. were reduced. Inverse correlations were also found between the levels of proinflammatory CD4+/IL-17+ T lymphocytes and the proportions of the SCFAs-producing genera, Coprococcus and Ruminococcus. Decreased abundance of Coprococcus spp., Clostridium spp., and an unidentified Ruminococcaceae family was observed in patients with active MS compared to HC along with increased CXCR3+ CD4+ and CD8+ T cells (32). These cell types were also inversely correlated with alpha diversity.

We observed a reduction in other SCFAs-producing bacteria and health-associated taxa in our patients with active MS, including genera Clostridium (in both CIS and NR), Ruminococcus and Dorea (only in the CIS) and Faecalibacterium and Christensenellaceae R7 group (only in the NR group). These species are known to be associated with lower severity in progressive MS (34). A significant reduction in Christensenellaceae R7 group and Ruminococcaceae family has also been reported in patients with pediatric MS, regardless of their treatment status, including DMT naïve and treatment exposed individuals (36).

An increased abundance of Bacteroides spp. was found in CIS and NR group. Similarly, higher levels of Bacteroides have been reported in patients treated with IFN-β and natalizumab compared to untreated patients. Conversely, decreased abundance of Bacteroides spp. was found in patients 6–12 months after initiating ocrelizumab treatment. This may reflect the dual immunomodulatory effects of the Bacteroides genus, which can be anti-inflammatory due to SCFAs production and induction of IL-10 secretion by Polysaccharide A (a component of the Bacteroides capsule), but also pro-inflammatory due to the overproduction and translocation of lipopolysaccharide (LPS) across the blood-brain barrier into the brain. LPS have also been detected in aged human brains around and within Alzheimer’s disease -affected neurons (37, 38).

Unlike our patients with active MS, patients with stable MS (R group) showed lower levels of genera Collinsella, and Prevotella 9 and higher levels of Streptococcus compared to HC. Higher abundance of Collinsella spp. has been previously reported to be associated with higher T2 lesion volume and higher EDSS (39). The decrease in Prevotella spp. abundance in MS patients has already been described with a statistically significant negative correlation demonstrated between the relative abundance of Prevotella spp. and the Th17 cell frequency in the human small intestine (40). In contrast to our findings, a study on relapsing-remitting MS patients with inactive disease reported higher abundance of Prevotella spp. not only in comparison to patients with active disease but also to the HC group. This study proposed a protective role for this beneficial, anti-inflammatory strain suggesting that it may counterbalance the effect of Th17 cell–inducing microbes in the pathogenesis of MS (3).

Increased level of genus Streptococcus has been previously described in Chinese and Japanese patients with MS, Streptococcus abundance negatively correlated with the proportion of T regulatory cells and positively correlated with Th17 cells in the peripheral blood (41). On the other hand, the anti-inflammatory activity of common human commensal Streptococcus salivarius K12 is attributed to its ability to avoid triggering an inflammatory response, achieved through inhibition of the NF-κB pathway, enhancement of type I and II interferon response, and modulation of genes associated with adhesion to the epithelial layer and homeostasis (42).

When analyzing patient groups according to their DMT, we again observed a decrease in recognized or putative SFCAs-producing bacteria in the non-responders’ groups. This included taxa Coprococcus (NR INF and NR FIN groups) (43), Lachnospiraceae (NR FIN, NR CLA) (44), Eggerthellaceae (NR FIN) (45), Christensenellaceae R7 group (NR IFN) (46), Intestinimonas (NR IFN and NR FIN) (47), Faecalibacterium (NR CLA), Eubacterium siraeum group (48) and Victivallis (NR FIN) (49).

To summarize our findings on the abundance of different bacterial taxa in the distinct patient groups based on treatment type, treatment responders showed only few significant differences from HC in microbiome composition. However, a significant decrease in many groups of SCFAs-producing bacteria was observed in patients with active MS. In some groups of SCFAs-producing bacteria, this decrease was shared by the CIS patient group and NR to all types of treatment. Thus, reduction in SCFAs producing bacteria was a common finding in active MS groups as compared to HC or responders to treatment. SCFAs exert an anti-inflammatory and immunoregulatory actions that may help counteract the autoimmune processes underlying MS. By promoting regulatory T cells, decreasing pro-inflammatory mediators through several mechanisms in immune and non-immune cells and supporting barrier integrity, SCFAs show promise as an adjunctive strategy in MS management. This observation was also supported by the quantification of serum SCFAs, where patients with active MS (CIS, NR) has significantly lower levels of SCFAs in comparison with HC while responders had SCFAs levels similar to HC. Next to the changes in SCFA-producing bacteria and SCFAs levels in serum another important microbiota-related pathways are discussed with MS pathogenesis. Circulating bile acid profiles are altered in MS and associate with neuroinflammation, brain/retinal atrophy and disability-worsening risk, and tauroursodeoxycholic acid (TUDCA) supplementation in progressive MS is safe with measurable immunologic and microbiome effects (50–52). Microbial cholesterol-to-coprostanol conversion is a major gut route of cholesterol disposal shaped by community composition; coprostanol-forming taxa associate with lower fecal/serum cholesterol and more favorable lipid profiles, with potential implications for immune-cell lipid signaling (53–55). Microbiota-regulated tryptophan metabolism (kynurenine/indole pathways) is consistently disturbed in MS, including pediatric-onset disease, and influences CNS autoimmunity and T-cell balance (56, 57). Integrating these axes with emerging metabolomics/microbiome approaches for disease staging, progression monitoring and DMT selection/monitoring may help explain inter-individual variability in treatment response (58, 59).

The intestinal barrier is the essential component of the microbiota-gut-brain axis. One of its functions is to prevent translocation of antigens and microbes from the gut lumen. Maintaining gut barrier integrity is crucial for proper function of other components of this axis, including the immune system (60, 61). To assess the gut barrier function in CIS and during different DMTs, we investigated several serum biomarkers related to the immune response associated with microbial translocation, and its regulations.

In our study we found significantly elevated levels of MBL and LBP in patients with active MS (CIS and NR) as compared to HC. However, no significant differences were observed in patients with stable disease (R group). These findings suggested the disruption of gut barrier integrity in MS patients with active disease resulting in microbial translocation followed by inflammatory response. In MS, MBL may contribute to complement system activation, which could result in demyelination, damage of oligodendrocytes and infiltration of macrophages to the CNS (62). Plasma levels of MBL have been found to be significantly increased in patients with MS, myasthenia gravis and Guillain-Barré syndrome compared to healthy controls. Moreover, MBL plasma levels have been significantly correlated with severity of these autoimmune neurological disorders (63). LBP is one of the key components of the innate immune response against bacteria. It binds to LPS on bacterial surfaces and interacts with CD14 (64). A small study investigating pretreatment level of LBP in patients with RRMS reported higher levels compared to controls. At 56 weeks of treatment with natalizumab, these mean values were significantly reduced compared to baseline values. This aligns with our observations of elevated MBL and LBP in patients with active, but not in those with stable MS. Interestingly, we found only slightly elevated level of CD14 in CIS patients as compared to HC. Significantly elevated levels of CD14 were found later in both, stable and active MS patients as compared to HC. It was described that increased serum CD14 correlated inversely with disease activity in relapsing MS patients (65). Our observation cannot support this finding since both responders and non-responders had elevated levels of CD14. Besides the results mentioned above, in our cohort SAA, but not Lipocalin-2 captured the low-grade inflammation associated with early diseases pathogenesis as well as established MS, irrespective to treatment response. Previously it was shown that SAA levels are elevated in relapsing-remitting MS and correlated with MRI activity, suggesting that systemic low-grade inflammation tracks CNS lesion formation (66–68). Mechanistically, SAA can interact with the blood–brain barrier and impair its function, activate microglia and astrocytes, and promote differentiation of pathogenic Th17 cells, thereby amplifying CNS inflammation and neurodegeneration (69, 70) The significantly elevated SAA we observed in CIS and MS (responders and non-responders) compared with healthy controls is therefore consistent with a state of persistent systemic immune activation that is already present at the earliest clinically recognizable stage and remains detectable under DMT, irrespective of treatment response status. The absence of significant differences in serum Lipocalin-2 between healthy controls, CIS, responders and non-responders in our cohort suggests that Lipocalin-2 is not a sensitive peripheral marker of the inflammatory processes captured by SAA in early and treated MS.

IL-18 is a member of IL-1 superfamily, which includes eleven cytokines involved in the initiation and regulation of inflammation associated with most CNS diseases. Both in vitro and in vivo studies have shown that IL-1 cytokines family and receptors are altered in Alzheimer Disease and MS. These alterations in production of cytokines such as IL-1 and IL-18 likely switch their role from supporting neuroprotection and homeostasis to promoting pathological inflammation (71). Higher levels of IL-18 have been previously detected in the serum and cerebrospinal fluid (CSF) of patients with MS (72). Unlike our findings, other study reported significantly higher levels of IL-18 in patients with active MS than patients with inactive MS and HC (73). Significantly higher level of IL-18 in responders compared to CIS and NR, could suggest involvement of IL-18 in repairing processes of damaged tissue, while the disease may be non-active. Besides, it was shown that neuronal IL-18 directs killing of enteric bacterial pathogens and orchestrates mucosal barrier immunity by stimulating antimicrobial protein expression in goblet cells (74). We may only speculate, that high levels of IL-18 in the responders as compared to CIS and NR, could be the mechanism supporting mucosal barrier integrity and preventing the microbial translocation in these patients. Using the level of IL-18 we were able to distinguish responders from NR in both, IFN-β and FIN DMTs.

Calprotectin (S100A8/S100A9) belongs an important group of proteins of the innate immune system that promotes inflammatory responses (75, 76). Several studies have demonstrated significantly higher serum levels of calprotectin in patients with active compared to inactive MS. Serum calprotectin has been therefore proposed as a clinically useful biomarker of innate immune activation and disease activity (77, 78). In our study we found significantly higher levels of serum calprotectin only in CIS patients as compared to responders and NR. Therefore, it could not be used to distinguish treatment responses, even when patients were stratified based on specific DMTs.

Conversely, we observed significantly increased levels of OPN in both treatment responders and non-responders, but not in CIS. OPN is well known to bind to several integrin receptors, including α4β1 integrin, which is expressed on circulating T-cells and is essential for T cell homing in MS. In MS, microglia-derived OPN promotes the influx of lymphocytes, including T cells, increases their survival, but also regulates myelination and oligodendrogenesis (79, 80). A systematic review on studies measuring blood and CSF levels of OPN in MS patients and controls found that MS patients generally had significantly higher levels of OPN in both their CSF and blood when compared to all types of controls. All MS patients, except those with CIS, also exhibited elevated blood OPN levels compared to controls. When considering MS subtypes, individuals with CIS had significantly lower CSF OPN levels than RRMS patients and lower levels in both blood and CSF compared to patients with progressive subtypes of MS. Finally, patients with active MS had significantly higher OPN levels in their CSF compared to patients with stable disease (81). Increased OPN levels in our patients with a long disease duration (with median values of 6 years in IFN-β; 9.5 years in CLA and 16.5 years in FIN) compared to HC, but not in those at the very beginning of their disease, may indicate a distinct role of OPN in the early and advanced stages of MS. We did not find any significant differences between responders and NR. The levels of OPN were significantly higher in both, responders and NR when compared to HC. This could support the previous findings highlighting the role of OPN not only in inflammation but also in tissue repair (82).

5 Conclusions

In summary, we observed significant differences in the gut microbiota composition, both at the alpha and beta diversity levels, in individuals with CIS compared to HC. Upon treatment, only NR exhibited microbiota profiles similar to those of CIS patients, with no significant differences between responders and HC. When analyzing patients based on DMT type, no significant differences in alpha or beta diversity were found in responders and HC, except for those treated with CLA. At the taxonomic level, reduced levels of SCFA-producing bacteria were observed in CIS patients and NR compared to HC, which was supported with quantification of SCFA in serum. Additionally, we detected significantly higher levels of LBP and MBL in CIS patients and NR, indicating increased intestinal barrier permeability in active MS. In responders, a decrease in these biomarkers was observed. This finding can be interpreted in the context of SCFAs, which are anti-inflammatory and whose producers are notably more abundant in treatment responders than in NR.

Data availability statement

The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found below: https://www.ebi.ac.uk/ena, PRJEB85230.

Ethics statement

The studies involving humans were approved by Ethics Committee of General University Hospital. The studies were conducted in accordance with the local legislation and institutional requirements. The participants provided their written informed consent to participate in this study.

Author contributions

VT: Conceptualization, Data curation, Data analysis and interpretation, Writing – original draft, Writing – review & editing. SC: Conceptualization, Data curation, Methodology, Data analysis and interpretation,Visualization, Writing – original draft, Writing – review & editing. ZJZ: Conceptualization, Data curation, Data analysis and interpretation, Methodology, Project administration, Supervision, Visualization, Writing – original draft, Writing – review & editing. TT: Data curation, Investigation, Writing – review & editing. RR: Investigation, Writing – review & editing. TH: Investigation, Writing – review & editing. MK: Conceptualization, Data interpretation, Writing – original draft, Writing – review & editing. MP: Data acquisition, Data curation, Writing – review & editing. PK: Data acquisition, Writing – review & editing. JLP: Data acquisition, Writing – review & editing. IK: Data acquisition, Writing – review & editing. JK: Investigation, Writing – review & editing. HT-H: Conceptualization, Data interpretation, Methodology, Writing – review & editing. EKH: Conceptualization, Data acquisition, Data interpretation, Methodology, Writing – review & editing.

Funding

The author(s) declared that financial support was received for this work and/or its publication. This research has been supported by the Ministry of Education, Youth and Sports of the Czech Republic grant Talking microbes-understanding microbial interactions within One Health framework (CZ.02.01.01/00/22_008/0004597) and by grant from the Ministry of Health of the Czech Republic (NU20-04-00077). Supported by The project National Institute for Neurological Research (Programme EXCELES, ID Project No. LX22NPO5107) - Funded by the European Union – Next Generation EU; Charles University: Cooperatio Program in Neuroscience; General University Hospital in Prague project MH CZ-DRO-VFN64165.

Acknowledgments

The authors thank Alena Kubatova for excellent technical assistance and Martin Mihula for sample preparation. We acknowledge the CF Genomics of CEITEC supported by the NCMG research infrastructure (LM2023067 funded by MEYS CR) for their support with obtaining scientific data presented in this paper.

Conflict of interest

The authors declared that this work was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Generative AI statement

The author(s) declared that generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Supplementary material

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fimmu.2025.1681342/full#supplementary-material

Supplementary Figure 1 | Relative abundances of the top 18 most abundant taxa of gut microbiota at the genus level for every individual in each group.

Supplementary Figure 2 | The levels of serum LBP, CD14, MBL, calprotectin, osteopontin and IL-18 in patients with CIS and RRMS patients according treatment and its response. Lipopolysaccharide-binding protein (LBP), soluble CD14 (CD14), Mannose-binding lectin (MBL), calprotectin, osteopontin, interleukin-18 (IL-18), Lipocalin-2, Serum Amyloid A (SAA). Statistical differences between groups were tested using the Kruskal-Wallis test with Dunn’s multiple comparisons *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001.

Glossary

AD: Alzheimer’s disease

ASVs: Amplicon Sequence Variants

BMI: body mass index

CD: cluster of definition

CIS: clinically isolated syndrome

CLA: cladribine

CNS: central nervous system

CSF: cerebrospinal fluid

CXCR: CXC chemokine receptor

DAMP: damage-associated molecular pattern proteins

DMT: disease modifying therapies

DNA: deoxyribonucleic acid

EAE: Experimental Autoimmune Encephalitis

EBV: Epstein-Barr virus

EDSS: Expanded Disability Status Scale

ELISA: enzyme-linked immunosorbent assay

ENA: European Nucleotide Archive

FIN: fingolimod

GALT: gut-associated lymphoid tissue

HC: healthy controls

HET: high-efficacy therapies

IFN: interferon

IL: interleukin

LBP: lipopolysaccharide-binding protein

LPS: lipopolysaccharide

MBL: mannose-binding lectin

MRI: magnetic resonance imaging

MS: multiple sclerosis

NA: not applicable

NEDA: No Evidence of Disease Activity

NF-κB: Nuclear factor kappa-light-chain-enhancer of activated B cells

NR: non-responders

NR group: non-responders

NR-CLA: non-responders to cladribine

NR-FIN: non-responders to fingolimod

NR-IFN: non-responders to IFN-β

OPN: osteopontin

PCoA: Principal Coordinate Analysis

PCR: polymerase chain reaction

PERMANOVA: Permutational Multivariate Analysis of Variance Using Distance Matrices

R: responders

R group: responders

R-CLA: responders to cladribine

R-FIN: responders to fingolimod

R-IFN: responders to IFN-β

RRMS: relapsing-remitting multiple sclerosis

rRNA: ribosomal ribonucleic acid

SAA: Serum Amyloid A

S1P: sphingosin-1-phosphate

SCFAs: short-chain fatty acids

SFB: segmented filamentous bacteria

SPMS: secondary progressive multiple sclerosis

Th: T helper cells

Tregs: T regulatory cells

References

1. Trapp BD, Peterson J, Ransohoff RM, Rudick R, Mörk S, and Bö L. Axonal transection in the lesions of multiple sclerosis. N Engl J Med. (1998) 338:278–85. doi: 10.1056/NEJM199801293380502

PubMed Abstract | Crossref Full Text | Google Scholar

2. Coufal S, Zakostelska ZJ, Thon T, Roubalova R, Kadleckova D, Salakova M, et al. Ocrelizumab transiently alters microbiota and modulates immune response depending on treatment outcome. iScience. (2025) 28:113872. doi: 10.1016/j.isci.2025.113872

PubMed Abstract | Crossref Full Text | Google Scholar

3. Chen J, Chia N, Kalari KR, Yao JZ, Novotna M, Paz Soldan MM, et al. Multiple sclerosis patients have a distinct gut microbiota compared to healthy controls. Sci Rep. (2016) 6:28484. doi: 10.1038/srep28484

PubMed Abstract | Crossref Full Text | Google Scholar

4. Preiningerova JL, Jiraskova Zakostelska Z, Srinivasan A, Ticha V, Kovarova I, Kleinova P, et al. Multiple sclerosis and microbiome. Biomolecules. (2022) 12. doi: 10.3390/biom12030433

PubMed Abstract | Crossref Full Text | Google Scholar

5. Hoffman K, Brownell Z, Doyle WJ, and Ochoa-Repáraz J. The immunomodulatory roles of the gut microbiome in autoimmune diseases of the central nervous system: Multiple sclerosis as a model. J Autoimmun. (2023) 137:102957. doi: 10.1016/j.jaut.2022.102957

PubMed Abstract | Crossref Full Text | Google Scholar

6. Schumacher SM, Doyle WJ, Hill K, and Ochoa-Repáraz J. Gut microbiota in multiple sclerosis and animal models. FEBS J. (2024) 292:1330–56. doi: 10.1111/febs.17161

PubMed Abstract | Crossref Full Text | Google Scholar

7. Reynders T, Devolder L, Valles-Colomer M, Van Remoortel A, Joossens M, De Keyser J, et al. Gut microbiome variation is associated to Multiple Sclerosis phenotypic subtypes. Ann Clin Transl Neurol. (2020) 7:406–19. doi: 10.1002/acn3.51004

PubMed Abstract | Crossref Full Text | Google Scholar

8. Calvo-Barreiro L, Eixarch H, Montalban X, and Espejo C. Combined therapies to treat complex diseases: The role of the gut microbiota in multiple sclerosis. Autoimmun Rev. (2018) 17:165–74. doi: 10.1016/j.autrev.2017.11.019

PubMed Abstract | Crossref Full Text | Google Scholar

9. Duscha A, Gisevius B, Hirschberg S, Yissachar N, Stangl GI, Dawin E, et al. Propionic acid shapes the multiple sclerosis disease course by an immunomodulatory mechanism. Cell. (2020) 180:1067–80.e16. doi: 10.1016/j.cell.2020.02.035

PubMed Abstract | Crossref Full Text | Google Scholar

10. Chen T, Noto D, Hoshino Y, Mizuno M, and Miyake S. Butyrate suppresses demyelination and enhances remyelination. J Neuroinflamm. (2019) 16:165. doi: 10.1186/s12974-019-1552-y

PubMed Abstract | Crossref Full Text | Google Scholar

11. Selmaj K, Cree BAC, Barnett M, Thompson A, and Hartung HP. Multiple sclerosis: time for early treatment with high-efficacy drugs. J Neurol. (2024) 271:105–15. doi: 10.1007/s00415-023-11969-8

PubMed Abstract | Crossref Full Text | Google Scholar

12. Pilotto S, Zoledziewska M, Fenu G, Cocco E, and Lorefice L. Disease-modifying therapy for multiple sclerosis: Implications for gut microbiota. Mult Scler Relat Disord. (2023) 73:104671. doi: 10.1016/j.msard.2023.104671

PubMed Abstract | Crossref Full Text | Google Scholar

13. Giovannoni G and Mathews J. Cladribine tablets for relapsing-remitting multiple sclerosis: A clinician’s review. Neurol Ther. (2022) 11:571–95. doi: 10.1007/s40120-022-00339-7

PubMed Abstract | Crossref Full Text | Google Scholar

14. Van Gils T, Nijeboer P, Van Wanrooij RL, Bouma G, and Mulder CJ. Mechanisms and management of refractory coeliac disease. Nat Rev Gastroenterol Hepatol. (2015) 12:572–9. doi: 10.1038/nrgastro.2015.155

PubMed Abstract | Crossref Full Text | Google Scholar

15. Hall GL, Compston A, and Scolding NJ. Beta-interferon and multiple sclerosis. Trends Neurosci. (1997) 20:63–7. doi: 10.1016/S0166-2236(96)10071-0

PubMed Abstract | Crossref Full Text | Google Scholar

16. Castillo-Álvarez F, Pérez-Matute P, Oteo JA, and Marzo-Sola ME. The influence of interferon β-1b on gut microbiota composition in patients with multiple sclerosis. Neurologia (Engl Ed). (2021) 36:495–503. doi: 10.1016/j.nrleng.2020.05.006

PubMed Abstract | Crossref Full Text | Google Scholar

17. Schierova D, Roubalova R, Kolar M, Stehlikova Z, Rob F, Jackova Z, et al. Fecal microbiome changes and specific anti-bacterial response in patients with IBD during anti-TNF therapy. Cells. (2021) 10. doi: 10.3390/cells10113188

PubMed Abstract | Crossref Full Text | Google Scholar

18. Jiang H, Lei R, Ding SW, and Zhu S. Skewer: a fast and accurate adapter trimmer for next-generation sequencing paired-end reads. BMC Bioinf. (2014) 15:182. doi: 10.1186/1471-2105-15-182

PubMed Abstract | Crossref Full Text | Google Scholar

19. Callahan BJ, McMurdie PJ, Rosen MJ, Han AW, Johnson AJ, and Holmes SP. DADA2: High-resolution sample inference from Illumina amplicon data. Nat Methods. (2016) 13:581–3. doi: 10.1038/nmeth.3869

PubMed Abstract | Crossref Full Text | Google Scholar

20. Edgar RC, Haas BJ, Clemente JC, Quince C, and Knight R. UCHIME improves sensitivity and speed of chimera detection. Bioinformatics. (2011) 27:2194–200. doi: 10.1093/bioinformatics/btr381

PubMed Abstract | Crossref Full Text | Google Scholar

21. Quast C, Pruesse E, Yilmaz P, Gerken J, Schweer T, Yarza P, et al. The SILVA ribosomal RNA gene database project: improved data processing and web-based tools. Nucleic Acids Res. (2013) 41:D590–6. doi: 10.1093/nar/gks1219

PubMed Abstract | Crossref Full Text | Google Scholar

22. McMurdie PJ and Holmes S. phyloseq: an R package for reproducible interactive analysis and graphics of microbiome census data. PloS One. (2013) 8:e61217. doi: 10.1371/journal.pone.0061217

PubMed Abstract | Crossref Full Text | Google Scholar

23. Jaccard P. Distribution de la Flore Alpine dans le Bassin des Dranses et dans quelques régions voisines. Bull la Societe Vaudoise Des Sci Naturelles. (1901) 37:241–72. doi: 10.5169/seals-266440

Crossref Full Text | Google Scholar

24. Bray JR and Curtis JT. An ordination of the upland forest communities of southern wisconsin. Ecol Monogr. (1957) 27:325–49. doi: 10.2307/1942268

Crossref Full Text | Google Scholar

25. Lozupone C and Knight R. UniFrac: a new phylogenetic method for comparing microbial communities. Appl Environ Microbiol. (2005) 71:8228–35. doi: 10.1128/AEM.71.12.8228-8235.2005

PubMed Abstract | Crossref Full Text | Google Scholar

26. Lozupone C, Lladser ME, Knights D, Stombaugh J, and Knight R. UniFrac: an effective distance metric for microbial community comparison. ISME J. (2011) 5:169–72. doi: 10.1038/ismej.2010.133

PubMed Abstract | Crossref Full Text | Google Scholar

27. Oksanen J, Blanchet FG, Kindt R, Legendere P, Minchin P, and O’Hara RB. Vegan: community ecology package. (2015) R Package Vegan.

Google Scholar

28. Barnett D, Arts I, and Penders J. microViz: an R package for microbiome data visualization and statistics. J Open Source Software. (2021) 6:3201. doi: 10.21105/joss.03201

Crossref Full Text | Google Scholar

29. Mallick H, Rahnavard A, McIver LJ, Ma S, Zhang Y, Nguyen LH, et al. Multivariable association discovery in population-scale meta-omics studies. PloS Comput Biol. (2021) 17:e1009442. doi: 10.1371/journal.pcbi.1009442

PubMed Abstract | Crossref Full Text | Google Scholar

30. Wickham H. ggplot2: elegant graphics for data analysis. J R Stat Soc Ser A: Stat Soc. (2016) 174:245–6. doi: 10.1007/978-0-387-98141-3

Crossref Full Text | Google Scholar

31. Saresella M, Marventano I, Barone M, La Rosa F, Piancone F, Mendozzi L, et al. Alterations in circulating fatty acid are associated with gut microbiota dysbiosis and inflammation in multiple sclerosis. Front Immunol. (2020) 11:1390. doi: 10.3389/fimmu.2020.01390

PubMed Abstract | Crossref Full Text | Google Scholar

32. Choileáin SN, Kleinewietfeld M, Raddassi K, Hafler DA, Ruff WE, and Longbrake EE. CXCR3+ T cells in multiple sclerosis correlate with reduced diversity of the gut microbiome. J Transl Autoimmun. (2020) 3:100032. doi: 10.1016/j.jtauto.2019.100032

PubMed Abstract | Crossref Full Text | Google Scholar

33. Jette S, De Schaetzen C, Tsai CC, and Tremlett H. The multiple sclerosis gut microbiome and disease activity: A systematic review. Mult Scler Relat Disord. (2024) 92:106151. doi: 10.1016/j.msard.2024.106151

PubMed Abstract | Crossref Full Text | Google Scholar

34. Ordoñez-Rodriguez A, Roman P, Rueda-Ruzafa L, Campos-Rios A, and Cardona D. Changes in gut microbiota and multiple sclerosis: A systematic review. Int J Environ Res Public Health. (2023) 20. doi: 10.3390/ijerph20054624

PubMed Abstract | Crossref Full Text | Google Scholar

35. Notting F, Pirovano W, Sybesma W, and Kort R. The butyrate-producing and spore-forming bacterial genus Coprococcus as a potential biomarker for neurological disorders. Gut Microbiome. (2023) 4:e16. doi: 10.1017/gmb.2023.14

PubMed Abstract | Crossref Full Text | Google Scholar

36. Tremlett H, Zhu F, Arnold D, Bar-Or A, Bernstein CN, Bonner C, et al. The gut microbiota in pediatric multiple sclerosis and demyelinating syndromes. Ann Clin Transl Neurol. (2021) 8:2252–69. doi: 10.1002/acn3.51476

PubMed Abstract | Crossref Full Text | Google Scholar

37. Tsai CC, Jette S, and Tremlett H. Disease-modifying therapies used to treat multiple sclerosis and the gut microbiome: a systematic review. J Neurol. (2024) 271:1108–23. doi: 10.1007/s00415-023-12107-0

PubMed Abstract | Crossref Full Text | Google Scholar

38. Zhao Y, Jaber VR, Pogue AI, Sharfman NM, Taylor C, and Lukiw WJ. Lipopolysaccharides (LPSs) as potent neurotoxic glycolipids in alzheimer’s disease (AD). Int J Mol Sci. (2022) 23. doi: 10.3390/ijms232012671

PubMed Abstract | Crossref Full Text | Google Scholar

39. Cantoni C, Lin Q, Dorsett Y, Ghezzi L, Liu Z, Pan Y, et al. Alterations of host-gut microbiome interactions in multiple sclerosis. EBioMedicine. (2022) 76:103798. doi: 10.1016/j.ebiom.2021.103798

PubMed Abstract | Crossref Full Text | Google Scholar

40. Cosorich I, Dalla-Costa G, Sorini C, Ferrarese R, Messina MJ, Dolpady J, et al. High frequency of intestinal T(H)17 cells correlates with microbiota alterations and disease activity in multiple sclerosis. Sci Adv. (2017) 3:e1700492. doi: 10.1126/sciadv.1700492

PubMed Abstract | Crossref Full Text | Google Scholar

41. Zeng Q, Junli G, Liu X, Chen C, Sun X, Li H, et al. Gut dysbiosis and lack of short chain fatty acids in a Chinese cohort of patients with multiple sclerosis. Neurochem Int. (2019) 129:104468. doi: 10.1016/j.neuint.2019.104468

PubMed Abstract | Crossref Full Text | Google Scholar

42. Cosseau C, Devine DA, Dullaghan E, Gardy JL, Chikatamarla A, Gellatly S, et al. The commensal Streptococcus salivarius K12 downregulates the innate immune responses of human epithelial cells and promotes host-microbe homeostasis. Infect Immun. (2008) 76:4163–75. doi: 10.1128/IAI.00188-08

PubMed Abstract | Crossref Full Text | Google Scholar

43. Nogal A, Louca P, Zhang X, Wells PM, Steves CJ, Spector TD, et al. Circulating levels of the short-chain fatty acid acetate mediate the effect of the gut microbiome on visceral fat. Front Microbiol. (2021) 12:711359. doi: 10.3389/fmicb.2021.711359

PubMed Abstract | Crossref Full Text | Google Scholar

44. Ćesić D, Lugović Mihić L, Ozretić P, Lojkić I, Buljan M, Šitum M, et al. Association of gut lachnospiraceae and chronic spontaneous urticaria. Life (Basel). (2023) 13. doi: 10.3390/life13061280

PubMed Abstract | Crossref Full Text | Google Scholar

45. Cornejo-Pareja I, Amiar MR, Ocaña-Wilhelmi L, Soler-Humanes R, Arranz-Salas I, Garrido-Sánchez L, et al. Non-alcoholic fatty liver disease in patients with morbid obesity: the gut microbiota axis as a potential pathophysiology mechanism. J Gastroenterol. (2024) 59:329–41. doi: 10.1007/s00535-023-02075-7

PubMed Abstract | Crossref Full Text | Google Scholar

46. Calderón-Pérez L, Gosalbes MJ, Yuste S, Valls RM, Pedret A, Llauradó E, et al. Gut metagenomic and short chain fatty acids signature in hypertension: a cross-sectional study. Sci Rep. (2020) 10:6436. doi: 10.1038/s41598-020-63475-w

PubMed Abstract | Crossref Full Text | Google Scholar

47. Afouda P, Durand GA, Lagier JC, Labas N, Cadoret F, Armstrong N, et al. Noncontiguous finished genome sequence and description of Intestinimonas massiliensis sp. nov strain GD2(T), the second Intestinimonas species cultured from the human gut. Microbiologyopen. (2019) 8:e00621. doi: 10.1002/mbo3.621

PubMed Abstract | Crossref Full Text | Google Scholar

48. Yao H, Flanagan BM, Williams BA, Wu X, Mikkelsen D, and Gidley MJ. Differential effects of pectin-based dietary fibre type and gut microbiota composition on in vitro fermentation outcomes. Carbohydr Polym. (2024) 339:122284. doi: 10.1016/j.carbpol.2024.122284

PubMed Abstract | Crossref Full Text | Google Scholar

49. Miao C, Xu X, Huang S, Kong L, He Z, Wang Y, et al. The causality between gut microbiota and hypertension and hypertension-related complications: A bidirectional two-sample mendelian randomization analysis. Hellenic J Cardiol. (2024) 83:38–50. doi: 10.1016/j.hjc.2024.02.002

PubMed Abstract | Crossref Full Text | Google Scholar

50. Bhargava P, Smith MD, Mische L, Harrington E, Fitzgerald KC, Martin K, et al. Bile acid metabolism is altered in multiple sclerosis and supplementation ameliorates neuroinflammation. J Clin Invest. (2020) 130:3467–82. doi: 10.1172/JCI129401

PubMed Abstract | Crossref Full Text | Google Scholar

51. Ladakis DC, Harrison KL, Smith MD, Solem K, Gadani S, Jank L, et al. Bile acid metabolites predict multiple sclerosis progression and supplementation is safe in progressive disease. Med. (2025) 6:100522. doi: 10.1016/j.medj.2024.09.011

PubMed Abstract | Crossref Full Text | Google Scholar

52. Erngren I, Vaivade A, Carlsson H, Al-Grety A, Akerfeldt T, Kockum I, et al. Bile acid metabolism in multiple sclerosis is perturbed and associated with the risk of confirmed disability worsening. BMC Med. (2025) 23:212. doi: 10.1186/s12916-025-04041-x

PubMed Abstract | Crossref Full Text | Google Scholar

53. Bubeck AM, Urbain P, Horn C, Jung AS, Ferrari L, Ruple HK, et al. High-fat diet impact on intestinal cholesterol conversion by the microbiota and serum cholesterol levels. iScience. (2023) 26:107697. doi: 10.1016/j.isci.2023.107697

PubMed Abstract | Crossref Full Text | Google Scholar

54. Kenny DJ, Plichta DR, Shungin D, Koppel N, Hall AB, Fu B, et al. Cholesterol metabolism by uncultured human gut bacteria influences host cholesterol level. Cell Host Microbe. (2020) 28:245–57.e6. doi: 10.1016/j.chom.2020.05.013

PubMed Abstract | Crossref Full Text | Google Scholar

55. Kriaa A, Bourgin M, Mkaouar H, Jablaoui A, Akermi N, Soussou S, et al. Microbial reduction of cholesterol to coprostanol: an old concept and new insights. Catalysts. (2019) 9:167. doi: 10.3390/catal9020167

Crossref Full Text | Google Scholar

56. Montgomery TL, Eckstrom K, Lile KH, Caldwell S, Heney ER, Lahue KG, et al. Lactobacillus reuteri tryptophan metabolism promotes host susceptibility to CNS autoimmunity. Microbiome. (2022) 10:198. doi: 10.1186/s40168-022-01408-7

PubMed Abstract | Crossref Full Text | Google Scholar

57. Nourbakhsh B, Bhargava P, Tremlett H, Hart J, Graves J, and Waubant E. Altered tryptophan metabolism is associated with pediatric multiple sclerosis risk and course. Ann Clin Transl Neurol. (2018) 5:1211–21. doi: 10.1002/acn3.637

PubMed Abstract | Crossref Full Text | Google Scholar

58. Bhargava P and Calabresi PA. Metabolomics in multiple sclerosis. Mult Scler. (2016) 22:451–60. doi: 10.1177/1352458515622827

PubMed Abstract | Crossref Full Text | Google Scholar

59. Zahoor I, Rui B, Khan J, Datta I, and Giri S. An emerging potential of metabolomics in multiple sclerosis: a comprehensive overview. Cell Mol Life Sci. (2021) 78:3181–203. doi: 10.1007/s00018-020-03733-2

PubMed Abstract | Crossref Full Text | Google Scholar

60. Tlaskalová-Hogenová H, Stěpánková R, Kozáková H, Hudcovic T, Vannucci L, Tučková L, et al. The role of gut microbiota (commensal bacteria) and the mucosal barrier in the pathogenesis of inflammatory and autoimmune diseases and cancer: contribution of germ-free and gnotobiotic animal models of human diseases. Cell Mol Immunol. (2011) 8:110–20. doi: 10.1038/cmi.2010.67

PubMed Abstract | Crossref Full Text | Google Scholar

61. Aburto MR and Cryan JF. Gastrointestinal and brain barriers: unlocking gates of communication across the microbiota-gut-brain axis. Nat Rev Gastroenterol Hepatol. (2024) 21:222–47. doi: 10.1038/s41575-023-00890-0

PubMed Abstract | Crossref Full Text | Google Scholar

62. Kwok JY, Vaida F, Augst RM, Yu DY, and Singh KK. Mannose binding lectin mediated complement pathway in multiple sclerosis. J Neuroimmunol. (2011) 239:98–100. doi: 10.1016/j.jneuroim.2011.08.018

PubMed Abstract | Crossref Full Text | Google Scholar

63. Farrokhi M, Dabirzadeh M, Dastravan N, Etemadifar M, Ghadimi K, Saadatpour Z, et al. Mannose-binding lectin mediated complement pathway in autoimmune neurological disorders. Iran J Allergy Asthma Immunol. (2016) 15:251–6.

PubMed Abstract | Google Scholar

64. Escribano BM, Medina-Fernández FJ, Aguilar-Luque M, Agüera E, Feijoo M, Garcia-Maceira FI, et al. Lipopolysaccharide binding protein and oxidative stress in a multiple sclerosis model. Neurotherapeutics. (2017) 14:199–211. doi: 10.1007/s13311-016-0480-0

PubMed Abstract | Crossref Full Text | Google Scholar

65. Lutterotti A, Kuenz B, Gredler V, Khalil M, Ehling R, Gneiss C, et al. Increased serum levels of soluble CD14 indicate stable multiple sclerosis. J Neuroimmunol. (2006) 181:145–9. doi: 10.1016/j.jneuroim.2006.09.002

PubMed Abstract | Crossref Full Text | Google Scholar

66. Ristori G, Laurenti F, Stacchini P, Gasperini C, Buttinelli C, Pozzilli C, et al. Serum amyloid A protein is elevated in relapsing-remitting multiple sclerosis. J Neuroimmunol. (1998) 88:9–12. doi: 10.1016/S0165-5728(98)00037-X

PubMed Abstract | Crossref Full Text | Google Scholar

67. Yokote H, Yagi Y, Watanabe Y, Amino T, Kamata T, and Mizusawa H. Serum amyloid A level is increased in neuromyelitis optica and atypical multiple sclerosis with smaller T2 lesion volume in brain MRI. J Neuroimmunol. (2013) 259:92–5. doi: 10.1016/j.jneuroim.2013.03.004

PubMed Abstract | Crossref Full Text | Google Scholar

68. Yokote H, Toru S, Nishida Y, Hattori T, Sanjo N, and Yokota T. Serum amyloid A level correlates with T2 lesion volume and cortical volume in patients with multiple sclerosis. J Neuroimmunol. (2021) 351:577466. doi: 10.1016/j.jneuroim.2020.577466

PubMed Abstract | Crossref Full Text | Google Scholar

69. Erickson MA and Mahankali AP. Interactions of serum amyloid A proteins with the blood-brain barrier: implications for central nervous system disease. Int J Mol Sci. (2024) 25. doi: 10.3390/ijms25126607

PubMed Abstract | Crossref Full Text | Google Scholar

70. Barbierato M, Borri M, Facci L, Zusso M, Skaper SD, and Giusti P. Expression and differential responsiveness of central nervous system glial cell populations to the acute phase protein serum amyloid A. Sci Rep. (2017) 7:12158. doi: 10.1038/s41598-017-12529-7

PubMed Abstract | Crossref Full Text | Google Scholar

71. Boraschi D, Italiani P, Migliorini P, and Bossù P. Cause or consequence? The role of IL-1 family cytokines and receptors in neuroinflammatory and neurodegenerative diseases. Front Immunol. (2023) 14:1128190. doi: 10.3389/fimmu.2023.1128190

PubMed Abstract | Crossref Full Text | Google Scholar

72. Keane RW, Dietrich WD, and De Rivero Vaccari JP. Inflammasome proteins as biomarkers of multiple sclerosis. Front Neurol. (2018) 9:135. doi: 10.3389/fneur.2018.00135

PubMed Abstract | Crossref Full Text | Google Scholar

73. Losy J and Niezgoda A. IL-18 in patients with multiple sclerosis. Acta Neurol Scand. (2001) 104:171–3. doi: 10.1034/j.1600-0404.2001.00356.x

PubMed Abstract | Crossref Full Text | Google Scholar

74. Jarret A, Jackson R, Duizer C, Healy ME, Zhao J, Rone JM, et al. Enteric nervous system-derived IL-18 orchestrates mucosal barrier immunity. Cell. (2020) 180:50–63.e12. doi: 10.1016/j.cell.2019.12.016

PubMed Abstract | Crossref Full Text | Google Scholar

75. Machahua C, Guler SA, Horn MP, Planas-Cerezales L, Montes-Worboys A, Geiser TK, et al. Serum calprotectin as new biomarker for disease severity in idiopathic pulmonary fibrosis: a cross-sectional study in two independent cohorts. BMJ Open Respir Res. (2021) 8. doi: 10.1136/bmjresp-2020-000827

PubMed Abstract | Crossref Full Text | Google Scholar

76. Edgeworth J, Gorman M, Bennett R, Freemont P, and Hogg N. Identification of p8,14 as a highly abundant heterodimeric calcium binding protein complex of myeloid cells. J Biol Chem. (1991) 266:7706–13. doi: 10.1016/S0021-9258(20)89506-4

PubMed Abstract | Crossref Full Text | Google Scholar

77. Olsson A, Gustavsen S, Hasselbalch IC, Langkilde AR, Sellebjerg F, Oturai AB, et al. Biomarkers of inflammation and epithelial barrier function in multiple sclerosis. Mult Scler Relat Disord. (2020) 46:102520. doi: 10.1016/j.msard.2020.102520

PubMed Abstract | Crossref Full Text | Google Scholar

78. Bogumil T, Rieckmann P, Kubuschok B, Felgenhauer K, and Brück W. Serum levels of macrophage-derived protein MRP-8/14 are elevated in active multiple sclerosis. Neurosci Lett. (1998) 247:195–7. doi: 10.1016/S0304-3940(98)00263-8

PubMed Abstract | Crossref Full Text | Google Scholar

79. Rosmus DD, Lange C, Ludwig F, Ajami B, and Wieghofer P. The role of osteopontin in microglia biology: current concepts and future perspectives. Biomedicines. (2022) 10. doi: 10.3390/biomedicines10040840

PubMed Abstract | Crossref Full Text | Google Scholar

80. Chabas D, Baranzini SE, Mitchell D, Bernard CC, Rittling SR, Denhardt DT, et al. The influence of the proinflammatory cytokine, osteopontin, on autoimmune demyelinating disease. Science. (2001) 294:1731–5. doi: 10.1126/science.1062960

PubMed Abstract | Crossref Full Text | Google Scholar

81. Agah E, Zardoui A, Saghazadeh A, Ahmadi M, Tafakhori A, and Rezaei N. Osteopontin (OPN) as a CSF and blood biomarker for multiple sclerosis: A systematic review and meta-analysis. PloS One. (2018) 13:e0190252. doi: 10.1371/journal.pone.0190252

PubMed Abstract | Crossref Full Text | Google Scholar

82. Cappellano G, Vecchio D, Magistrelli L, Clemente N, Raineri D, Barbero Mazzucca C, et al. The Yin-Yang of osteopontin in nervous system diseases: damage versus repair. Neural Regener Res. (2021) 16:1131–7. doi: 10.4103/1673-5374.300328

PubMed Abstract | Crossref Full Text | Google Scholar