Skip to main content
NCBI home page
eGastroenterology logoLink to eGastroenterology
. 2026 Jan 22;4(1):e100296. doi: 10.1136/egastro-2025-100296

Reduced fibre-fermenting capacity of gut microbes in multiple sclerosis may result in prebiotic dietary fibre β-fructan promoting inflammation and CNS damage

Stephanie L Tollenaar 1,0, Reihane Khorasaniha 2,0, Juan Jovel 3, Ismaila Ba 4, Athalia Voisin 2, Richard Miller 2, Hana Olof 2, Ramsha Mahmood 2, Ruth Ann Marrie 2,5, Erin Strachan 6, Luisa P Soares 6, Christopher Cheng 7,8, Jonathan Janveaux 2,7,8, Deenaz Zaidi 7,8, Charles N Bernstein 2, Christine Bonner 9, Amit Bar-Or 10, Emmanuelle Waubant 11, E Ann Yeh 12, Morag Graham 9, Douglas L Arnold 13, Julia O’Mahony 14, Brenda L Banwell 15, Feng Zhu 16,17, Ali I Mirza 18, Soheila Karimi-Abdolrezaee 19, Sue Tsai 6, Helen Tremlett 18, Kevin McGregor 20, Benjamin P Willing 1,8, Heather Armstrong 2,7,8,✉
PMCID: PMC12829359  PMID: 41585377

Abstract

Background

Some people with multiple sclerosis display changes in their gut microbiota with separate evidence suggesting that symptoms may worsen following a high-fibre diet. We hypothesised that in people with multiple sclerosis whose gut microbiota are less able to efficiently ferment dietary fibres, unfermented β-fructans induce inflammation.

Methods

Diet data (n=48 individuals with multiple sclerosis, n=78 unaffected controls) and stool microbiome data (n=31 individuals with multiple sclerosis, n=61 unaffected controls) were previously collected from participants. Daily fibre subtype intakes were calculated and compared with faecal shotgun metagenomic sequencing in paediatric onset multiple sclerosis and unaffected persons. Response to unfermented β-fructans was examined in a germ-free experimental autoimmune encephalomyelitis (EAE) mouse model (unable to ferment fibres). Mice were fed β-fructans or control fibre diet beginning at symptom onset (day 14). EAE scores and weights were recorded daily. Intestinal and central nervous system tissues were collected at two endpoints to examine inflammatory responses and demyelinating lesions.

Results

Individuals with paediatric onset multiple sclerosis consumed less β-fructans (2.4 g/day±0.3 SD; p<0.05) than unaffected participants (3.6 g/day±0.4), which coincided with differences in the gut microbiota including lower fibre fermenting enzymes. Mice exposed to unfermented β-fructans sustained worsened EAE symptoms (day 20–28; p<0.05), immune activation in the gut and immune activation plus demyelinating lesions in the spinal cord compared with mice on control diet.

Conclusions

The gut microbiota of individuals with paediatric-onset multiple sclerosis showed reduced fibre fermenting properties, and our animal findings suggest that unfermented β-fructans can worsen demyelination and promote gut–brain axis immune activation. Lower β-fructan consumption was observed among participants with paediatric-onset multiple sclerosis. Future longitudinal studies are warranted to confirm the findings uncovered in this manuscript.

Keywords: Multiple Sclerosis; Gastrointestinal Microbiome; Dietary Fibre; Brain-Gut Axis; Encephalomyelitis, Autoimmune, Experimental

WHAT IS ALREADY KNOWN ON THIS TOPIC

  • Some people with multiple sclerosis display altered gut microbiota, with separate evidence suggesting high-fibre diet worsens symptoms. If gut microbiota do not fully ferment dietary β-fructan fibres, β-fructans can induce inflammation in gut diseases. The impacts of altered gut microbiome on diet-host interactions in multiple sclerosis have not been studied.

WHAT THIS STUDY ADDS

  • Persons with multiple sclerosis consume less β-fructan fibres than unaffected participants. Lower β-fructan consumption was associated with differences in the gut microbiota. The presence of unfermented dietary β-fructan fibres induced immune activation in the gut along with immune activation and damage in the central nervous system in a mouse model of multiple sclerosis.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

  • Although fibres are typically thought to be beneficial, some dietary fibres may have detrimental effects in select persons with multiple sclerosis who lack fermentative microbial activities. Persons with low fibre-fermenting microbiota who consume high amounts of prebiotic β-fructans may experience worsened multiple sclerosis symptoms. It is possible that low amounts of β-fructans from whole foods may still be tolerated or beneficial, while higher levels of β-fructans may have a negative impact.

Introduction

Multiple sclerosis is a disease of the central nervous system (CNS) in which neurological disability results from demyelination, inflammation and axonal degeneration.1 Multiple sclerosis involves interactions between genetic and environmental factors, including diet and the gut microbiome.2 Gut microbiota compositional differences, known as dysbiosis, influence immune dysfunction in multiple sclerosis,3 4 but specific aspects of the dysbiosis require further study. Diet has one of the greatest influences on microbiota composition and functions.

Unlike other dietary factors, dietary fibres are not digested by the host; benefits of a high-fibre diet rely on microbial fermentation in the colon, leading to production of beneficial by-products such as short-chain fatty acids (SCFAs).5 Studies using the experimental autoimmune encephalomyelitis (EAE) mouse model have demonstrated that SCFAs ameliorate disease course, suppress demyelination and enhance remyelination.6 7 In people with multiple sclerosis, SCFA levels are reduced, which has been associated with differences in circulating immune cells and serum cytokine levels.8 9 Lower production of SCFAs and loss of fibre-fermenting microbiota in multiple sclerosis10,12 suggest that fibre fermentation may be impeded in multiple sclerosis, likely due to altered microbiota.

We previously showed that loss of fibre fermenting capacity in persons with inflammatory bowel disease results specifically in proinflammatory responses to β-fructan fibres, not other tested fibre subtypes; response was driven by β-fructan interactions with toll-like receptor-2 (TLR2) and activation of NOD-containing, LRR-containing and pyrin domain-containing protein 3 (NLRP3).13 In this setting, the inflammatory response is driven by an increase in the amount of fibre remaining in the gut due to lack of fermentation and a subsequent decrease in by-products of fermentation, including the anti-inflammatory SCFAs, butyrate and propionate. Interestingly, persons with multiple sclerosis also display changes in TLR2 and loss of fibre-fermenting microbes.14,17

Ketogenic and low-carbohydrate diets have been shown to possibly be beneficial and well-tolerated by people with multiple sclerosis.18 19 However, an unintended consequence of such diets may lead to some persons reducing intake of dietary fibre and other essential vitamins and nutrients in attempts to avoid carbohydrates.20 One cross-sectional UK study concluded that people with multiple sclerosis may be sensitive to fibre consumption.15 Among 2410 people with multiple sclerosis, intake of the recommended daily fibre amount was significantly associated with higher self-care-related problems, pain and problems carrying out usual activities, versus persons consuming a low-fibre diet.

Research increasingly focuses on the role of fibre fermentation in the gut and the subsequent impact on immune-mediated diseases. The benefits of dietary fibres consistently correlate with successful fibre fermentation.21 Different microbes ferment different types of dietary fibres;5 22 however, whether altered microbiota in multiple sclerosis impacts fermentation of specific fibres is unknown. In this study, we aimed to determine whether the intake of one dietary fibre subtype, β-fructans, was different in paediatric onset multiple sclerosis versus age-matched unaffected controls and to what extent β-fructan ingestion adversely impacted a murine EAE model of multiple sclerosis lacking fibre fermentation capabilities.

Materials and methods

Participant selection and data sources

Details of recruitment and cohort characteristics were previously published.17 In the present study, we used published dietary and stool shotgun metagenomics data previously collected from persons ≤22 years old enrolled in a Canadian Pediatric Demyelinating Disease Network Study (CPDDN; online supplemental table 1). Paediatric onset multiple sclerosis was defined as those who met McDonald criteria (2017),23 with symptom onset <18 years of age. They were treatment-naïve or exposed to interferon-β or glatiramer acetate, at the time of the present substudy. Participants had not taken an antibiotic or corticosteroids within the preceding 30 days. The CPDDN advertised to recruit unaffected youth (controls). Demographics were collected using a questionnaire and other detailed clinical data were obtained using standardised forms, completed by site neurologists and facilitated by a study coordinator.16 In the current study, we obtained access to diet data and stool shotgun metagenomics sequencing. We started with those participants from the previously published CPDDN cohort who had complete dietary intake data (n=48 individuals with multiple sclerosis, n=78 unaffected controls), and next evaluated paired microbiota data from those who had both complete dietary data along with paired stool shotgun metagenomics data available (n=31 individuals with multiple sclerosis, n=61 unaffected controls).

Dietary fibre consumption analysis

Dietary intake was evaluated using the Block Kids Food Screener (NutritionQuest), a validated Food Frequency Questionnaire (FFQ) containing 77 food items to reflect the prior week’s diet.24 The questionnaire was completed within 1 week of stool sample collection. Individuals with unreasonably low or high energy intake, that is, <500 or >3500 kcal/day, were excluded from the diet data analysis under suspicion of lack of accurate recall. The Canadian Nutrient File online database was applied along with the dietary fibre data set13 22 to calculate the daily kcal and the consumption of dietary fibre subtypes including inulin, fructooligosaccharide (FOS), pectin and arabinoxylan (AX). We applied our validated approach to calculate consumption of each of the most abundant dietary fibre subtypes from the collected diet data, as previously.13 22 Both g/day and kcal adjusted consumption were calculated (inulin, FOS, pectin and AX).

Stool data processing, gene annotation and phylogenetic profiling

Stool was previously processed using the Zymo Quick-DNA Fecal/Soil Microbe Miniprep Kit (D6010) and shotgun metagenomic was performed using the Illumina NextSeq platform.17 Here, we assessed previously collected stool shotgun metagenomics data.

Sequences were trimmed with fastq-mcf using a quality score threshold of 20 and allowing a minimal length of trimmed sequences of 100 bp. Trimmed sequences were taxonomically classified with MetaPhlAn325 using a custom database containing the archaea, bacteria, viruses, plasmid, human and UniVec_Core RefSeq data sets supplemented with the Genome Taxonomy Database. Pseudoalignments were filtered with a confidence threshold of 0.1 (at least 10% of kmers in each query sequence should map to the reference sequence). Taxa accumulating less than 1% of counts in each sample were filtered out to increase the robustness of quantification.

Functional profiling was carried out with HUMAnN326 27 using only end1 of the trimmed sequences described above. Classification was done against the UniRef protein database clustered with a similarity threshold of 90%. Initially, HUMAnN3 identifies species present in the microbiome and then aligns reads against the pangenomes and conducts searches of in-silico translated sequences that were initially unclassified to quantify gene families and metabolic pathways.26

Differential accumulation analysis was performed with the R V.4.2 package DESeq2.28 Count data were transformed using the harmonic mean.

EAE mouse model of multiple sclerosis

Female germ-free C57BL/6 mice aged 7–8 weeks were acclimated to their cages for 1 week and housed with up to five mice per cage. EAE was induced using the Hooke Lab vaccine (MOG35–55/PTX; EK-2110). Detailed procedures were performed according to the manufacturer’s instructions. The mice were evaluated blindly every 1–2 days for weight along with using a standard five-point scoring scale (online supplemental table 2) to assess symptoms of EAE such as weakness in limbs, loss of coordination or balance, altered gait or difficulty walking, tail weakness or limpness, weight loss and general lethargy. On symptom onset (day 14), mice were randomly assigned to one of two experimental diets either receiving a diet containing β-fructans as the sole fibre (synergy-1 Beneo 5% β-fructan; n=10 EAE; n=9 control) or a balanced control diet with non-fermentable cellulose as the sole fibre (AIN-76A 5% cellulose; n=11 EAE; n=5 control). Feed was provided ad libitum and body weight was measured every 1–2 days. On day 29, the mice were euthanised and the tissues from the gut (cecum and colon) and CNS (brain and spinal cord) were dissected and either submersed in cold phosphate-buffered saline (PBS) and processed for flow cytometry or single-cell sequencing (scSeq) or were fixed in 4% paraformaldehyde solution before preparation of formalin-fixed paraffin-embedded (FFPE) slides. FFPE CNS and gut tissues from EAE C57BL/6 mice were sectioned at a thickness of 5 µm, deparaffinised in xylene and rehydrated in ethanol for all staining. Digesta samples were tested to confirm germ-free status at the end of the study using aerobic and anaerobic culture as well as by PCR.

LFB myelination staining and analysis

Staining was performed using 0.1% Luxol fast blue (LFB; 0.5% glacial acetic acid, in methanol) incubated for 16 hours at 56°C in a water bath. Slides were washed in ethanol followed by deionised water, then placed in a 0.05% sodium carbonate solution for 3 min, followed by a 30 s immersion in 70% ethanol. Slides were counterstained with haematoxylin for 90 s and washed under running deionised water. Subsequently, slides were immersed in 0.3% acetic acid alcohol. Slides were processed using ethanol and xylene and mounted using dibutyl phthalate xylene (DPX) mounting media. Slides were imaged at 40× magnification using the Nanozoomer digital slide scanner (Hamamatsu) and images were analysed using NDP.view2 software. LFB was blindly scored manually29 to define the extent of demyelinating lesions by two researchers using Fiji.30

H&E staining and histology

Tissues were stained with haematoxylin for 90 s then differentiated with 0.3% acetic acid alcohol and counterstained with eosin for 10 s. Slides were dehydrated in ethanol and cleared with xylene before mounting using DPX mounting media. Slides were imaged at 40× magnification using the Nanozoomer digital slide scanner (Hamamatsu) and images were analysed using NDP.view2 software. Blinded intestinal scoring was conducted independently by two researchers using Fiji. A validated scoring method was used to score (0–4) the intestinal inflammation based on specific schemes for the small intestine (scheme 6) and colon (scheme 4).31 Four different sections were scored and averaged for each tissue.

Immunohistochemistry staining and analysis

Citrate antigen retrieval (10 mM Citric Acid, 0.05% Tween 20, pH 6.0) was performed on tissue slides, then blocked with 2% goat serum for 1 hour. Primary antibodies were applied overnight at 4°C in a humidified chamber; CXCR1 (1:10,000, PA5-27184, Thermo Fisher Scientific, Waltham, MA, USA), IL-17A (1: 40, ab79056, Abcam, Cambridge, UK) and CD4 (1:250, 1:500, 14-9766-80, Clone 4SM95, Thermo Fisher Scientific) in 1% bovine serum albumin (BSA). Slides were incubated in 0.3% hydrogen peroxide for 5 min, followed by horseradish peroxidase (HRP)-conjugated secondary antibodies (Goat-anti-rabbit IgG; ab214880; Abcam for CXCR1 and Goat-anti-rat IgG; ab214882; Abcam for IL-17A and CD4) for 1 hour. Tissues were counter-stained using 3,3'-diaminobenzidine (DAB) and haematoxylin before final rinse steps in ethanol and xylene. Slides were mounted using DPX mounting medium. Stained slides were scanned at 40× magnification using the Nanozoomer digital slide scanner (Hamamatsu) and viewed with NDP.view2 software. Cell staining was quantified using Fiji software.32 Colour deconvolution for haematoxylin-DAB was applied to eliminate background interference. The threshold settings were adjusted to focus on stained areas for CD4 and CXCR1. The watershed tool was used to identify and count specific cells, and the number of particles was quantified.32 The average scores were also calculated and compared between mouse groups.

Single-cell RNA sequencing

Raw FASTQ data is publicly available at the National Center for Biotechnology Information (NCBI) Sequence Read Archive portal under accession number PRJNA1298367. Processed single-cell RNA sequencing (scRNA-seq) data and analysis results are freely available on request.

Following the culling at day 23 (3 days post symptom peak and mid-inflammatory cycle), the intestinal tissue and spinal cord were cut in half longitudinally, and half of each was used for scSeq RNA analysis of three mice consuming a β-fructan diet and in three mice consuming a control diet. Tissue segments were gently cut into approximately 0.5 cm segments to begin manual tissue digestion, followed by chemical digestion (Hanks’ Balanced Salt Solution (HBSS) medium, 4% DNase I, 1% Collagenase MA and 0.5% AOF BP Protease). Tissues were digested in chemical digestion buffer at 37°C for 10 min followed by gentle pipetting and repeated chemical digestion to a maximum of four times. Resulting supernatants were passed through 70 µm cell strainers into 1 mL of cold foetal calf serum (FCS) and 5 mL cold fluorescence activated cell sorting (FACS) buffer (PBS+2% FCS). Cells were pelleted by centrifugation and haemocytometer counting was performed with a desired concentration of 5×106 cells/mL. Samples were FACS sorted by the University of Alberta Flow Cytometry Core on a Sony MA900 (100 µm nozzle) using propidium iodide stain and 50 000 cells were collected per sample. Samples were delivered to the University of Alberta Advanced Cell Exploration Core. The 10x scSeq RNA libraries were prepared using the 10x Genomics Chromium Next GEM Single Cell 3’ Gene Expression V.3.1 (Dual Index) and quality control was assessed on a Bioanalyzer. Sequencing was performed by Novogene.

Transcript abundance in scRNA-seq libraries was assessed with CellRanger V.8.0.1 (10x Genomics), using as reference the mouse genome GRCm38/mm10 and the corresponding GTF file. Each library was sequenced in triplicate and they were quantified separately. Quantification results were obtained as filtered feature-barcode matrices.

Count matrices generated by CellRanger were postprocessed using a custom computational pipeline implemented in Python V.3.8. The analysis pipeline employed several bioinformatics tools, including Scanpy V.1.9.333 and AnnData.34 Initially, replicates for each sample were merged. For each tissue (cecum and spine cord), control and β-fructan were merged. Merged data sets were filtered to exclude cells where the percentage of mitochondrial RNA surpassed 10% and the number of detected genes was larger than 200 and smaller than 6000. Doublets were identified and removed using Scrublet.35 Data were normalised to median total counts, logarithmised and feature selection was conducted using the top expressing 2000 genes.

Cell types were identified using CellTypist.36 Cecum and spine cord cells were classified using the models ‘Adult_Mouse_Gut.pkl’ and ‘Mouse_Whole_Brain.pkl’, respectively. No model for the spinal cord is available in CellTypist. Data sets were subjected to principal component analysis37 and Leiden graph clustering.38 Uniform Manifold Approximation and Projection (UMAP) embedding was conducted using ‘predicted labels’ or ‘majority voting’ for colouring plots and those plots were finally used to curate annotations.

For differential expression analysis, for each tissue, control and β-fructan groups were contrasted using the Wilcoxon rank-sum test as implemented in Scanpy. Genes were considered significantly differentially expressed if they met the following criteria: (1) absolute log2 fold change >0.25 and (2) adjusted p value <0.05. P values were adjusted for multiple testing using the Benjamini-Hochberg procedure. Differentially expressed genes were visualised using volcano plots, with significant genes highlighted and top-genes labelled. The plots display log2 fold changes on the x-axis and -log10 (adjusted p-value) on the y-axis.

All the above-described computational analyses were performed using Python V.3.8. The computational pipeline used the following packages: Scanpy, AnnData, Scrublet, SciPy, Pandas, NumPy, Seaborn, Matplotlib, AdjustText, CellTypist and Leidenalg. Analyses were executed in Google Colab with data stored on Google Drive.

Gene Ontology (GO) enrichment analysis was conducted to identify biological processes, molecular functions (MFs) and cellular components associated with differentially expressed genes. Analyses were performed separately for upregulated and downregulated genes using the clusterProfiler package in R.39 Significant differentially expressed genes were partitioned into upregulated (log2 fold change >0) and downregulated (log2 fold change <0) sets. The enrichGO function from clusterProfiler was applied to each gene set using the org.Mm.eg.db annotation database with gene symbols as the key type. GO enrichment was performed separately for MF ontology categories with the following parameters: p value adjustment method set to Benjamini-Hochberg, p value cut-off of 0.05 and q value cut-off of 0.05. Results were visualised using dot plots of the top 10 significantly enriched GO terms for both upregulated and downregulated gene sets. The enrichment results, including GO terms, gene counts, p values and adjusted p values, were saved as tab-separated files.

Flow cytometry

Following culling on day 23 (3 days post symptom peak and mid-inflammatory cycle), the intestinal tissue and spinal cord were cut in half longitudinally. Half of each of the intestine and the spinal cord was used for flow cytometry analysis of three mice who had β-fructan diet and in three mice fed a control diet.

The caecum was excised and cut longitudinally and sectioned into approximately 1 cm segments. Cecum segments were incubated in 10 mL stripping buffer (HBSS+2% foetal bovine serum (FBS), 5 mM EDTA, 15 mM HEPES) at 37°C with shaking for 20 min (×2), rinsed in 10 mL of ice cold wash buffer (HBSS+2% FBS, 15 mM HEPES), minced with scissors to create a pulp, transferred into 10 mL digestion buffer (RPMI+10% FBS, Penn/Strep, L-glutamine, 1 mM NaPyruvate, β-mercaptoethanol (β-ME), 50 U/mL DNase, 0.5 mg/mL Collagenase IV), incubated at 37°C for 45 min with shaking and filtered through a 70 µm filter with a cold PBS wash, followed by resuspension and enumeration. Spinal cord tissue was similarly minced and digested. Single cell suspensions (1×106) were pelleted in a V-bottom 96 well plate and blocked with Fc-block in FACS buffer (1/100 dilution) for 10 min on ice, followed by staining for surface antigens (online supplemental table 3) with diluted antibodies (1/250 in FACS buffer, 4°C, 30 min). Cells were washed, permeabilised (FoxP3 fix/perm buffer, eBiosciences, Thermo Fisher Scientific) and stained for intracellular antigens according to manufacturers’ recommendations. Cells were analysed on an Aurora spectral cytometer. FlowJo (BD, V.10.8) was used to analyse all flow cytometry data.

Statistical analysis

To assess the associations between FOS, paediatric onset multiple sclerosis (case or control), and microbe abundances, analyses were performed in R V.4.2 using the edgeR package. This package allows normalisation using the ‘Trimmed Mean of M-values’ method and accounts for over-dispersion in the abundance data. Each microbe was included as an outcome variable and predictors included FOS level, paediatric onset multiple sclerosis diagnosis, as well as their interaction. A two-way analysis of variance test was used to assess significant differences among groups in GraphPad Prism V.10 for animal work, while comparison between paediatric onset multiple sclerosis and controls was performed using the analysis of covariance test in SPSS V.26 which was further adjusted for age. All p values were corrected with the Benjamini-Hochberg method and comparisons with corrected p values <0.05 were considered significant.

Results

Participant summary

Dietary data from 48 paediatric onset multiple sclerosis and 78 unaffected controls were examined from a previously published cohort.17 Of these participants, only 31 paediatric onset multiple sclerosis participants and 61 unaffected controls had paired stool shotgun metagenomics data available for analysis. Cohort characteristics are shown in online supplemental table 1. All cases had relapsing–remitting multiple sclerosis. The multiple sclerosis cases and controls were similar in age at diet data collection (average 15.3 and 14.2 years, respectively) and age at stool sample procurement (average 16.5 and 14.9 years, respectively). The average age of multiple sclerosis symptom onset was 15.1 years for the total cohort and 15.7 years for those who had both diet and stool. Females represented 81% of multiple sclerosis cases and 71% of controls for the full cohort and 84% of multiple sclerosis cases and 64% of controls for those with both diet and stool data.

Lower consumption of dietary β-fructan fibres is associated with reduction in microbial metabolites involved in β-fructan fermentation in paediatric onset multiple sclerosis

Individuals with paediatric onset multiple sclerosis consume similar amounts of total fibre based on FFQ reported intake compared with unaffected controls, yet they did consume less β-fructans (FOS and inulin; p<0.05) versus controls (figure 1A). On testing the association of microbiota findings with dietary fibre intake, we found β-fructan fermenting microbes (eg, Bifidobacterium) and hydrolysing enzymes (eg, xyloglucan oligosaccharide beta-galactosidase, xylan 1,4-beta-xylosidase) were lower in paediatric onset multiple sclerosis than in unaffected controls (figure 1B–D). Lower consumption of β-fructan was associated with a higher abundance of Methanobrevibacter spp., Escherichia spp. and Desulfovibrio (figure 2A). Consumption of β-fructan was differentially associated with microbiota composition in unaffected controls compared with paediatric onset multiple sclerosis (figure 2B). Higher β-fructan intake in controls was associated with higher abundance of key fibre-fermenting microbiota, including several Lactobacillus spp., while intake of β-fructan in paediatric onset multiple sclerosis was primarily associated with Enterobacter spp., Methanobrevibacter spp. and Prevotella spp.

Figure 1. Reduced consumption of dietary β-fructan fibres is associated with reduction in microbial metabolites involved in β-fructan fermentation. (A) Dietary fibre subtypes including β-fructan (FOS, inulin), pectin and AX were assessed using diet data from individuals with paediatric onset MS (n=48) and unaffected controls (n=78). Microbiota composition and functions were confirmed by shotgun metagenomics in stool collected from a subset of individuals with paediatric onset MS (n=31) and control participants (n=61) with paired diet and stool data. Shotgun metagenomics data were analysed by (B) MetaPhlAn3 (microbiota composition) and (C) HUMAnN3 (microbiota functions). (D) Microbial metabolites identified by shotgun metagenomics to be significantly altered in MS were presented. *p<0.05, **p<0.01, ***p<0.001. AX, arabinoxylan; FC, fold change; FOS, fructooligosaccharide; HC, healthy control; MS, multiple sclerosis; ns, not significant.

Figure 1

Open in a new tab

Figure 2. Predictive analysis demonstrated a correlation between consumption of β-fructan and gut microbiota taxonomy. Consumption of dietary β-fructan fibres was correlated with microbiota taxa from shotgun metagenomic analysis of stool samples and was presented as (A) the association of the log2 FC in microbiota taxa relative abundance with increasing β-fructan consumption (g/day) by one unit; and (B) log2 FC >0 indicates that the effect of increased β-fructan consumption (g/day) on microbiota taxa is higher in MS, while log2 FC <0 indicates that the effect of increased β-fructan consumption (g/day) on microbiota taxa is higher in HC. FC, fold change; HC, healthy control; MS, multiple sclerosis.

Figure 2

Open in a new tab

Unfermented β-fructans worsen CNS demyelination and multiple sclerosis-like symptoms along with both CNS and gastrointestinal immune activation in an EAE mouse model

Direct response to unfermented β-fructans was examined in a germ-free EAE mouse model (unable to ferment fibres; figure 3A). Unfermented β-fructans induced worsened EAE symptoms (day 20–29; p<0.05; figure 3B) and a higher burden of CNS demyelinating lesions compared with EAE on the control cellulose fibre diet (figure 3C). There was no impact of β-fructans on animal weight compared with control diets. Non-EAE control mice showed no response to β-fructans.

Figure 3. β-fructan diet caused worsened symptom score and increased demyelinating lesions in a germ-free EAE mouse model (unable to ferment fibres). Both C57BL/6 control mice and EAE mice were included in the study. (A) All mice were fed β-fructan diet (n=10 EAE; n=9 control) or control fibre diet (n=11 EAE; n=5 control) beginning at symptom onset (day 14) until 28 days. (B) Animals were weighed and assessed for EAE symptoms daily. (C) At endpoint, spinal cord tissue was collected, paraffin embedded and stained using LFB to identify demyelinating lesions (red arrows). *p<0.05, **p<0.01, ***p<0.001. B, β-fructan diet; C, control diet; CNTRL, control; EAE, experimental autoimmune encephalomyelitis; LFB, Luxol fast blue; MS, multiple sclerosis.

Figure 3

Open in a new tab

As β-fructans are large molecules, they do not pass the intestinal barrier, suggesting that the interactions between β-fructans and the host occur in the gut. We found no significant difference in the caecum (online supplemental figure 1A) and colon (online supplemental figure 1B) histological scores between β-fructan and control diets in either EAE or non-EAE control mice. To examine if there was any impact of β-fructans on known immune markers in the CNS and gut relevant to multiple sclerosis,3 we performed immunohistochemistry using a series of antibodies and uncovered significantly higher levels of CXCR1 (figure 4A) and CD4 (figure 4B) in CNS, and higher levels of CXCR1 in the colon (figure 4C), but not in cecum (figure 4D).

Figure 4. In an EAE mouse model, reduced consumption of dietary β-fructan fibres is associated with reduction in microbial metabolites involved in β-fructan fermentation. All mice were fed a β-fructan diet (n=10 EAE; n=9 control) or control fibre diet (n=11 EAE ; n=5 control) beginning at symptom onset (day 14) until 28 days. At endpoint, spinal cord, cecum and colon tissues were collected, paraffin embedded and stained using IHC to identify possible immune changes including (A) CXCR1 in the spinal cord of EAE mice compared with control diet, (B) CD4 in the spinal cord of EAE mice compared with control diet and (C) CXCR1 in the colon of EAE mice compared with control diet, (D) CXCR1 in the cecum and CD4 in the cecum and colon. (E) Multiplex ELISA of mouse serum in EAE mice consuming the β-fructan diet compared with the control diet. *p<0.05, **p<0.01, ***p<0.001. B, β-fructan diet; C, control diet; CNTRL, control; EAE, experimental autoimmune encephalomyelitis; IHC, immunohistochemistry.

Figure 4

Open in a new tab

Consumption of dietary β-fructan fibres in EAE mice is associated with altered immune cell populations and functions in the spinal cord and cecum

We performed flow cytometry and scSeq on tissues collected 3 days postsymptom peak, which is reflective of the inflammatory phase of EAE. As the initial site of response to dietary β-fructan, the cecum displayed trends towards elevated T helper cell 1 (Th1) cell abundance in β-fructan diet mice compared with control diet mice (figure 5A,B). The top six significantly upregulated gene pathways identified in the cecum of EAE mice consuming β-fructan diet, compared with control diet, included GTPase regulator activity, nucleoside-triphosphatase regulator activity, glycosyltransferase activity, hexosyltransferase activity, UDP-glycosyltransferase activity and cytokine binding (figure 5C). The top six downregulated gene pathways included structural constituent of ribosome, ribosomal RNA binding, cadherin binding, messenger RNA 5’-UTR binding, ubiquitin-protein transferase inhibitor activity and ubiquitin-protein transferase regulator activity (figure 5C). In the spinal cord, β-fructan intake in EAE displayed trends towards higher CD45+ cells, microglia, CD11c+/CD69+ cells and CD11chi/Tbet+cells (figure 6A,B). The top six significantly upregulated gene pathways in the spinal cord of mice consuming the β-fructan diet included primary active transmembrane transporter activity, active transmembrane transporter activity, oxidoreduction-driven active transmembrane transporter activity, electron transfer activity, NADH dehydrogenase (ubiquinone) activity and NADH dehydrogenase (quinone) activity; the top six downregulated gene pathways included methylated histone binding, methylation-dependent protein binding, modification-dependent protein binding, histone binding, protein transmembrane transporter activity and oxidoreductase activity (figure 6C). Other cell populations measured by flow cytometry did not display any difference in difference between β-fructan and control diets (online supplemental figure 2).

Figure 5. Consumption of dietary β-fructan fibres in EAE mice is associated with altered immune cell populations in the caecum. EAE mice were fed β-fructan diet (n=3) or control fibre diet (n=3) beginning at symptom onset (day 14) until 23 days. At endpoint cecum tissues were collected for (A) single cell sequencing cell profiles, (B) quantitative flow cytometry of immune cell populations and (C) single-cell RNA sequencing–based differential gene expression and molecular function enrichment analyses. DC, dendritic cell; EAE, experimental autoimmune encephalomyelitis; MF, molecular function; mRNA, messenger RNA; NK, natural killer; pDC, plasmacytoid dendritic cell; rRNA, ribosomal RNA; Th1, T helper cell 1; Treg, regulatory T cell; UMAP, Uniform Manifold Approximation and Projection.

Figure 5

Open in a new tab

Figure 6. Consumption of dietary β-fructan fibres in EAE mice is associated with altered immune cell populations in the spinal cord. EAE mice were fed β-fructan diet (n=3) or control fibre diet (n=3) beginning at symptom onset (day 14) and maintained until 23 days. At the endpoint, spinal cord tissues were collected for (A) single cell sequencing cell profiles, (B) quantitative flow cytometry of immune cell populations and (C) single-cell RNA sequencing-based differential gene expression and molecular function enrichment analyses. EAE, experimental autoimmune encephalomyelitis; MF, molecular function; UMAP, Uniform Manifold Approximation and Projection; VLMC, vascular leptomeningeal cell.

Figure 6

Open in a new tab

Discussion

While previous studies in both human cohorts and animal models have suggested that a high-fibre diet is associated with worsened multiple sclerosis,15 40 no prior studies have examined mechanisms of fibre–microbiome–host interactions or assessed the impacts of specific fibre subtypes in multiple sclerosis. Consistent with our previous findings in persons with inflammatory bowel disease,13 here we found that paediatric onset multiple sclerosis participants consume less β-fructan fibres than unaffected children and adolescents, which coincided with differences in the gut microbiota, including lower fibre fermenting capacity. In contrast, higher consumption of β-fructans was associated with beneficial microbiota such as Lactobacillus spp. in unaffected controls. The same beneficial relationship was not found in paediatric onset multiple sclerosis, which is in line with prior evidence that suggests that the extent to which select dietary fibres impact the host in autoimmunity is highly influenced by an individual’s microbiota and differs significantly between healthy and autoimmune disease states.41,43 However, the paediatric cohort in the present study involved collection of samples and data at only one time point, which does not allow us to define whether changes in the gut microbiome or dietary fibre intake occur first in paediatric onset multiple sclerosis. While microbiota shifts have been identified in our patient cohort,41 44 longitudinal studies are needed with experimental approaches such as ex vivo culture of stool microbiota with different dietary fibres to better identify the capacity of microbiota from people with multiple sclerosis to use these molecules.

To overcome these limitations, we used a murine model that offers evidence of why a high-fibre diet may not always be beneficial in multiple sclerosis. We designed this murine experiment based on our prior work in inflammatory bowel disease,1341,43 which showed lower consumption of β-fructans could be due to lower fermentation by gut microbiota, resulting in greater amounts of unfermented β-fructans interacting with host cells of the gut, driving TLR2-NLRP3 activation and inflammatory cytokine secretion in the gut. More recent studies have supported our findings, demonstrating that high levels of dietary β-fructans can induce type-2 inflammation in the gut and lung in mouse models.41,43 Here we used a germ-free EAE mouse model to mimic this lower fibre-fermenting capacity. Mice were exposed to diet after EAE symptom onset to examine the impacts of β-fructans in the setting of diagnosed multiple sclerosis where the gut microbiome is known to differ from healthy individuals. In EAE mice exposed to β-fructans, immune activation was seen in the gut and immune activation plus demyelinating lesions in the spinal cord, along with worse EAE symptoms, compared with control diet-fed mice. This indicates a heightened inflammatory response in the CNS, suggesting unfermented β-fructans may induce neuroinflammation in the setting of multiple sclerosis. Neuroinflammation and demyelination in multiple sclerosis likely occur when the immune system mistakenly targets myelin in the CNS; this leads to the breakdown of the blood–brain barrier, allowing immune cells to infiltrate the CNS.45 46 Intestinal barrier dysfunction and greater intestinal permeability are common at the onset of EAE, allowing for increased interactions between luminal contents and gut immune cells.47 These findings suggest that the impacts of unfermented β-fructans may involve host immune changes reflective of multiple sclerosis, as unfermented β-fructans only induced immune changes and inflammation in EAE mice and not in non-EAE control mice.

Mechanistically, we uncovered that unfermented β-fructan promotes immune activation of pathways with known involvement in the multiple sclerosis gut-brain axis.3 CXCR1 was higher in both the cecum and spinal cord of EAE mice on a β-fructan diet, while CD4 was only significantly higher in the spinal cord. CXCR1+ macrophages are known to be involved in gut–brain axis activation in multiple sclerosis and are typically more abundant in patients with multiple sclerosis and EAE mice.3 48 CXCR1 knockout mice display reduced sensitivity to EAE.48 These data support our findings that β-fructan-induced increase in CXCR1+ cells would be associated with worsened EAE outcomes. CXCR1+ macrophages also play critical roles in chemoattraction and T cell activation through the production of cytokines that can influence T cell differentiation and function.49 50 Multiple sclerosis is mediated by myelin-reactive CD4+ Th cells, including primarily Th1 and Th17 lineages.51 52 Thus, our results suggest that activation of macrophages in the gut through interactions with unfermented β-fructans may result in inflammatory cytokine signalling that drives infiltration and activation of T cells in the CNS, ultimately promoting demyelination and worsened multiple sclerosis symptoms. This is in line with our findings in inflammatory bowel disease, indicating a key role for macrophages in β-fructan response in the gut.13 While our results suggest that several immune cell populations are likely involved in this response to β-fructans, further research is needed to confirm the complex mechanisms of response as the scRNA-seq and flow cytometry experiments in this study were not well powered and validation studies have not been performed to confirm a mechanistic role of different cell types. Furthermore, the germ-free EAE mice lack gut microbes entirely, which does not adequately reflect the natural microbiome or immune settings of persons with multiple sclerosis. Future studies should aim to use humanised mice colonised with human gut microbiota along with clinical investigations to determine precision responses to different types of dietary fibres, including the β-fructans.

Conclusion

The benefits of certain fibres are well studied in select disease settings; however, the benefits of dietary fibres depend on the presence of fibre-fermenting gut microbes.5 22 In individuals lacking these microbes, evidence has shown that certain types of fibre (eg, β-fructan) can remain unfermented, triggering inflammation and worsening symptoms.13 41 43 These findings, along with the established gut–brain axis,3 53 suggest that gut dysbiosis may extend the complex impact of fibres beyond the gut, potentially linking diet, microbiota and autoimmune conditions like multiple sclerosis. Though causality cannot be inferred, findings from this study support initial evidence that if β-fructans are not broken down by gut microbes and an individual continues to consume high amounts of these typically beneficial fibres, they may induce worsened multiple sclerosis symptoms and result in people with multiple sclerosis avoiding foods that make them feel worse. It is possible that low amounts of β-fructan from foods such as garlic and banana may be tolerated or beneficial,54 even in those with reduced fibre fermentation capacity, while higher levels of prebiotic β-fructans found in commercial prebiotics, supplements or certain grocery products may have a negative impact until the gut microbiome’s fermentation potential is increased. These findings demonstrate that further, more clinically relevant investigations are warranted to fully understand the interaction between specific fibres, microbes and the host in multiple sclerosis and how to support clinical improvements in the gut microbiome and host through precision nutrition approaches in the future.

Supplementary material

online supplemental figure 1
egastro-4-1-s001.pdf (366.7KB, pdf)
DOI: 10.1136/egastro-2025-100296
online supplemental figure 2
egastro-4-1-s002.pdf (309.8KB, pdf)
DOI: 10.1136/egastro-2025-100296
online supplemental table 1
egastro-4-1-s003.docx (21.9KB, docx)
DOI: 10.1136/egastro-2025-100296

Acknowledgements

We graciously acknowledge the support of the Center of Excellence of Inflammation and Immunity Research (CEGIIR), Advanced Cell Exploration Core and the flow cytometry core facilities at the University of Alberta, along with the Manitoba Centre for Proteomics and Systems Biology, University of Manitoba for staff and infrastructure supports for the completion of this study. The presented study also builds off a previously published cohort involving the prior collection of creation of data analysed in this study. The study ‘From bugs to brains: the gut microbiome in pediatric MS’ was supported by MS Canada’s Multiple Sclerosis Scientific Research Foundation (#EGID: 2636; PI: HT) and is a substudy of the Canadian Pediatric Demyelinating Disease Network, CPDDN, study (PIs: BLB, AB-O, RAM, EAY, DLA, was also funded by MS Canada’s Multiple Sclerosis Scientific Research Foundation). The authors are grateful for all of the involvement of the participants, especially children and youth with multiple sclerosis, and their parents without whom this study would not have been possible. The authors also acknowledge the important contributions of: the Tremlett team (University of British Columbia); Thomas Duggan in facilitating study set-up, coordination and data collection; Bonnie Leung for study coordination; Michael Sargent (Department of Internal Medicine and the University of Manitoba IBD Clinical and Research Centre laboratory, Winnipeg, MB, Canada) for managing the stool biobank and Dr Jessica D Forbes (University of Toronto, Toronto, Canada) for assisting with the original grant application. We are also grateful to the investigators and study teams at each site who participated in the CPDDN study, including the study manager (Dr JO’M) and study members Dr Natalie Knox and Dr Yinshan Zhao.

Footnotes

Funding: Canada Research Chairs (HA; CRC-2021-00172) MS Canada (HA; 1030817).

Prepub: Prepublication history and additional supplemental material for this paper are available online. To view these files, please visit the journal online (https://doi.org/10.1136/egastro-2025-100296).

Provenance and peer review: Not commissioned; externally peer reviewed.

Patient consent for publication: Not applicable.

Ethics approval: Details of clinical participant recruitment and cohort characteristics were previously published with publicly available data reassessed in the present study.16 Human ethics was previously approved and only publicly available data was used in the present study (The SickKids Research Ethics Board #1000017648; University of British Columbia’s Clinical Research Ethics Board #H15-03330). Animal ethics was approved by the University of Alberta Research Ethics Office (AUP00004214).

Data availability free text: Sequencing data are publicly available from the NCBI SRA portal under the accession PRJNA1298367; there are no restrictions on data availability. All data are available in the main text or the supplementary materials.

Patient and public involvement: Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.

Data availability statement

Data are available in a public, open access repository.

References

  • 1.Oudejans E, Luchicchi A, Strijbis EMM, et al. Is MS affecting the CNS only? Lessons from clinic to myelin pathophysiology. Neurol Neuroimmunol Neuroinflamm. 2021;8:e914. doi: 10.1212/NXI.0000000000000914. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Olsson T, Barcellos LF, Alfredsson L. Interactions between genetic, lifestyle and environmental risk factors for multiple sclerosis. Nat Rev Neurol. 2017;13:25–36. doi: 10.1038/nrneurol.2016.187. [DOI] [PubMed] [Google Scholar]
  • 3.Correale J, Hohlfeld R, Baranzini SE. The role of the gut microbiota in multiple sclerosis. Nat Rev Neurol. 2022;18:544–58. doi: 10.1038/s41582-022-00697-8. [DOI] [PubMed] [Google Scholar]
  • 4.Yoon H, Gerdes LA, Beigel F, et al. Multiple sclerosis and gut microbiota: Lachnospiraceae from the ileum of MS twins trigger MS-like disease in germfree transgenic mice-An unbiased functional study. Proc Natl Acad Sci U S A. 2025;122:e2419689122. doi: 10.1073/pnas.2419689122. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Armstrong H, Mander I, Zhang Z, et al. Not All Fibers Are Born Equal; Variable Response to Dietary Fiber Subtypes in IBD. Front Pediatr. 2021;8:620189. doi: 10.3389/fped.2020.620189. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Melbye P, Olsson A, Hansen TH, et al. Short-chain fatty acids and gut microbiota in multiple sclerosis. Acta Neurol Scand. 2019;139:208–19. doi: 10.1111/ane.13045. [DOI] [PubMed] [Google Scholar]
  • 7.Chen T, Noto D, Hoshino Y, et al. Butyrate suppresses demyelination and enhances remyelination. J Neuroinflammation. 2019;16:165. doi: 10.1186/s12974-019-1552-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Trend S, Leffler J, Jones AP, et al. Associations of serum short-chain fatty acids with circulating immune cells and serum biomarkers in patients with multiple sclerosis. Sci Rep. 2021;11:5244. doi: 10.1038/s41598-021-84881-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Olsson A, Gustavsen S, Nguyen TD, et al. Serum Short-Chain Fatty Acids and Associations With Inflammation in Newly Diagnosed Patients With Multiple Sclerosis and Healthy Controls. Front Immunol. 2021;12:661493. doi: 10.3389/fimmu.2021.661493. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Ordoñez-Rodriguez A, Roman P, Rueda-Ruzafa L, et al. Changes in Gut Microbiota and Multiple Sclerosis: A Systematic Review. Int J Environ Res Public Health. 2023;20:4624. doi: 10.3390/ijerph20054624. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Schoeps VA, Zhou X, Horton MK, et al. Short-chain fatty acid producers in the gut are associated with pediatric multiple sclerosis onset. Ann Clin Transl Neurol. 2024;11:169–84. doi: 10.1002/acn3.51944. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Lin Q, Dorsett Y, Mirza A, et al. Meta-analysis identifies common gut microbiota associated with multiple sclerosis. Genome Med. 2024;16:94. doi: 10.1186/s13073-024-01364-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Armstrong HK, Bording-Jorgensen M, Santer DM, et al. Unfermented β-fructan Fibers Fuel Inflammation in Select Inflammatory Bowel Disease Patients. Gastroenterology. 2023;164:228–40. doi: 10.1053/j.gastro.2022.09.034. [DOI] [PubMed] [Google Scholar]
  • 14.Tremlett H, Fadrosh DW, Faruqi AA, et al. Gut microbiota in early pediatric multiple sclerosis: a case-control study. Eur J Neurol. 2016;23:1308–21. doi: 10.1111/ene.13026. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Coe S, Tektonidis TG, Coverdale C, et al. A cross sectional assessment of nutrient intake and the association of the inflammatory properties of nutrients and foods with symptom severity in a large cohort from the UK Multiple Sclerosis Registry. Nutr Res. 2021;85:31–9. doi: 10.1016/j.nutres.2020.11.006. [DOI] [PubMed] [Google Scholar]
  • 16.Tremlett H, Zhu F, Arnold D, 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. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Mirza AI, Zhu F, Knox N, et al. Metagenomic Analysis of the Pediatric-Onset Multiple Sclerosis Gut Microbiome. Neurology (ECronicon) 2022;98:e1050–63. doi: 10.1212/WNL.0000000000013245. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Ortí JER, Cuerda-Ballester M, Sanchis-Sanchis CE, et al. Exploring the impact of ketogenic diet on multiple sclerosis: obesity, anxiety, depression, and the glutamate system. Front Nutr. 2023;10:1227431. doi: 10.3389/fnut.2023.1227431. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Brenton JN, Lehner-Gulotta D, Woolbright E, et al. Phase II study of ketogenic diets in relapsing multiple sclerosis: safety, tolerability and potential clinical benefits. J Neurol Neurosurg Psychiatry . 2022;93:637–44. doi: 10.1136/jnnp-2022-329074. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Adam-Perrot A, Clifton P, Brouns F. Low-carbohydrate diets: nutritional and physiological aspects. Obes Rev. 2006;7:49–58. doi: 10.1111/j.1467-789X.2006.00222.x. [DOI] [PubMed] [Google Scholar]
  • 21.Zhang F, Fan D, Huang J, et al. The gut microbiome: linking dietary fiber to inflammatory diseases. Medicine in Microecology . 2022;14:100070. doi: 10.1016/j.medmic.2022.100070. [DOI] [Google Scholar]
  • 22.Khorasaniha R, Olof H, Voisin A, et al. Diversity of fibers in common foods: Key to advancing dietary research. Food Hydrocoll. 2023;139:108495. doi: 10.1016/j.foodhyd.2023.108495. [DOI] [Google Scholar]
  • 23.Thompson AJ, Banwell BL, Barkhof F, et al. Diagnosis of multiple sclerosis: 2017 revisions of the McDonald criteria. Lancet Neurol. 2018;17:162–73. doi: 10.1016/S1474-4422(17)30470-2. [DOI] [PubMed] [Google Scholar]
  • 24.Hunsberger M, O’Malley J, Block T, et al. Relative validation of Block Kids Food Screener for dietary assessment in children and adolescents. Matern Child Nutr. 2015;11:260–70. doi: 10.1111/j.1740-8709.2012.00446.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Lu J, Rincon N, Wood DE, et al. Metagenome analysis using the Kraken software suite. Nat Protoc. 2022;17:2815–39. doi: 10.1038/s41596-022-00738-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Franzosa EA, McIver LJ, Rahnavard G, et al. Species-level functional profiling of metagenomes and metatranscriptomes. Nat Methods. 2018;15:962–8. doi: 10.1038/s41592-018-0176-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Beghini F, McIver LJ, Blanco-Míguez A, et al. Integrating taxonomic, functional, and strain-level profiling of diverse microbial communities with bioBakery 3. Elife. 2021;10:e65088. doi: 10.7554/eLife.65088. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Love MI, Huber W, Anders S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol. 2014;15:550. doi: 10.1186/s13059-014-0550-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.García-García ÓD, Weiss T, Chato-Astrain J, et al. Staining Methods for Normal and Regenerative Myelin in the Nervous System. Methods Mol Biol. 2023;2566:187–203. doi: 10.1007/978-1-0716-2675-7_15. [DOI] [PubMed] [Google Scholar]
  • 30.Khodanovich MY, Anan’ina TV, Krutenkova EP, et al. Challenges and Practical Solutions to MRI and Histology Matching and Measurements Using Available ImageJ Software Tools. Biomedicines. 2022;10:1556. doi: 10.3390/biomedicines10071556. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Erben U, Loddenkemper C, Doerfel K, et al. A guide to histomorphological evaluation of intestinal inflammation in mouse models. Int J Clin Exp Pathol. 2014;7:4557–76. [PMC free article] [PubMed] [Google Scholar]
  • 32.Schneider CA, Rasband WS, Eliceiri KW. NIH Image to ImageJ: 25 years of image analysis. Nat Methods. 2012;9:671–5. doi: 10.1038/nmeth.2089. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Wolf FA, Angerer P, Theis FJ. SCANPY: large-scale single-cell gene expression data analysis. Genome Biol. 2018;19:15. doi: 10.1186/s13059-017-1382-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34.Virshup I, Rybakov S, Theis FJ, et al. anndata: Annotated data. bioRxiv. 2021 doi: 10.1101/2021.12.16.473007. [DOI] [Google Scholar]
  • 35.Wolock SL, Lopez R, Klein AM. Scrublet: Computational Identification of Cell Doublets in Single-Cell Transcriptomic Data. Cell Syst. 2019;8:281–91. doi: 10.1016/j.cels.2018.11.005. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Xu C, Prete M, Webb S, et al. Automatic cell-type harmonization and integration across Human Cell Atlas datasets. Cell. 2023;186:5876–91. doi: 10.1016/j.cell.2023.11.026. [DOI] [PubMed] [Google Scholar]
  • 37.Ahlmann-Eltze C, Huber W. Comparison of transformations for single-cell RNA-seq data. Nat Methods. 2023;20:665–72. doi: 10.1038/s41592-023-01814-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38.Traag VA, Waltman L, van Eck NJ. From Louvain to Leiden: guaranteeing well-connected communities. Sci Rep. 2019;9:5233. doi: 10.1038/s41598-019-41695-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39.Wu T, Hu E, Xu S, et al. clusterProfiler 4.0: A universal enrichment tool for interpreting omics data. The Innovation (Camb) 2021;2:100141. doi: 10.1016/j.xinn.2021.100141. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40.Sen MK, Ge A, Liao E, et al. Impact of a high fibre diet on immune cells and the multiple sclerosis animal model. The Journal of Immunology. 2023;210:66. doi: 10.4049/jimmunol.210.Supp.66.11. [DOI] [Google Scholar]
  • 41.Bonazzi E, Bretin A, Vigué L, et al. Individualized microbiotas dictate the impact of dietary fiber on colitis sensitivity. Microbiome. 2024;12:5. doi: 10.1186/s40168-023-01724-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 42.Arifuzzaman M, Collins N, Guo CJ, et al. Nutritional regulation of microbiota-derived metabolites: Implications for immunity and inflammation. Immunity. 2024;57:14–27. doi: 10.1016/j.immuni.2023.12.009. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 43.Arifuzzaman M, Won TH, Li T-T, et al. Inulin fibre promotes microbiota-derived bile acids and type 2 inflammation. Nature. 2022;611:578–84. doi: 10.1038/s41586-022-05380-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 44.Hobson CA, Vigue L, Naimi S, et al. MiniBioReactor Array (MBRA) in vitro gut model: a reliable system to study microbiota-dependent response to antibiotic treatment. JAC Antimicrob Resist . 2022;4:dlac077. doi: 10.1093/jacamr/dlac077. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 45.Ellwardt E, Zipp F. Molecular mechanisms linking neuroinflammation and neurodegeneration in MS. Exp Neurol. 2014;262:8–17. doi: 10.1016/j.expneurol.2014.02.006. [DOI] [PubMed] [Google Scholar]
  • 46.Engelhardt B. The blood-central nervous system barriers actively control immune cell entry into the central nervous system. Curr Pharm Des. 2008;14:1555–65. doi: 10.2174/138161208784705432. [DOI] [PubMed] [Google Scholar]
  • 47.Nouri M, Bredberg A, Weström B, et al. Intestinal barrier dysfunction develops at the onset of experimental autoimmune encephalomyelitis, and can be induced by adoptive transfer of auto-reactive T cells. PLoS One. 2014;9:e106335. doi: 10.1371/journal.pone.0106335. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 48.Zhuang W, Zhou J, Zhong L, et al. CXCR1 drives the pathogenesis of EAE and ARDS via boosting dendritic cells-dependent inflammation. Cell Death Dis . 2023;14:608. doi: 10.1038/s41419-023-06126-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 49.Korbecki J, Gąssowska-Dobrowolska M, Wójcik J, et al. The Importance of CXCL1 in Physiology and Noncancerous Diseases of Bone, Bone Marrow, Muscle and the Nervous System. Int J Mol Sci. 2022;23:4205. doi: 10.3390/ijms23084205. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 50.Guerriero JL. Macrophages: Their Untold Story in T Cell Activation and Function. Int Rev Cell Mol Biol. 2019;342:73–93. doi: 10.1016/bs.ircmb.2018.07.001. [DOI] [PubMed] [Google Scholar]
  • 51.Moser T, Akgün K, Proschmann U, et al. The role of TH17 cells in multiple sclerosis: Therapeutic implications. Autoimmun Rev. 2020;19:102647. doi: 10.1016/j.autrev.2020.102647. [DOI] [PubMed] [Google Scholar]
  • 52.Jadidi-Niaragh F, Mirshafiey A. Th17 cell, the new player of neuroinflammatory process in multiple sclerosis. Scand J Immunol. 2011;74:1–13. doi: 10.1111/j.1365-3083.2011.02536.x. [DOI] [PubMed] [Google Scholar]
  • 53.Kasarello K, Cudnoch-Jedrzejewska A, Czarzasta K. Communication of gut microbiota and brain via immune and neuroendocrine signaling. Front Microbiol. 2023;14:1118529. doi: 10.3389/fmicb.2023.1118529. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 54.Li N, Han X, Ruan M, et al. Prebiotic inulin controls Th17 cells mediated central nervous system autoimmunity through modulating the gut microbiota and short chain fatty acids. Gut Microbes. 2024;16:2402547. doi: 10.1080/19490976.2024.2402547. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

    This section collects any data citations, data availability statements, or supplementary materials included in this article.

    Supplementary Materials

    online supplemental figure 1
    egastro-4-1-s001.pdf (366.7KB, pdf)
    DOI: 10.1136/egastro-2025-100296
    online supplemental figure 2
    egastro-4-1-s002.pdf (309.8KB, pdf)
    DOI: 10.1136/egastro-2025-100296
    online supplemental table 1
    egastro-4-1-s003.docx (21.9KB, docx)
    DOI: 10.1136/egastro-2025-100296

    Data Availability Statement

    Data are available in a public, open access repository.

    Articles from eGastroenterology are provided here courtesy of The First Hospital of Jilin University and BMJ

    RESOURCES