Multi-omics identification of a microbial metabolite driving OXPHOS-mediated epithelial damage in experimental autoimmune encephalomyelitis

Abstract

Bowel syndrome is a prevalent and debilitating symptom in patients with multiple sclerosis (MS), substantially impairing their quality of life. However, the underlying mechanisms of MS development remain poorly understood. In this study, we demonstrated that intestinal epithelial cells (IECs) and the mucosal barrier were disrupted during experimental autoimmune encephalomyelitis (EAE) induction, driven by the inhibition of mitochondrial oxidative phosphorylation (OXPHOS). Proteomic analysis confirmed alterations in OXPHOS complexes, with a pronounced decrease in the expression of cytochrome c oxidase and ATP synthase subunits in small intestinal epithelial cells (sIECs). We identified a gut microbiota-derived metabolite that induces IEC dysfunction by downregulating OXPHOS protein complexes. Specifically, metabolomic analysis revealed an enrichment of phenyllactic acid (PLA), a phenolic acid typically produced by Lactobacillus murinus, in the cecal contents of EAE mice. Our findings indicate that PLA actively downregulates OXPHOS complexes and restrains maximal mitochondrial respiration. Using a multi-omics approach, this study elucidated a potential mechanism by which gut microbiota dysbiosis observed in EAE mice compromises IEC integrity and disrupts the mucosal barrier.

INTRODUCTION

Multiple sclerosis (MS) is an autoimmune disease of the central nervous system (CNS) characterized by demyelination and neurodegeneration as the primary symptoms (1). Patients with MS often experience secondary symptoms, such as bowel dysfunction, reducing their quality of life (1, 2). The experimental autoimmune encephalomyelitis (EAE) mouse model mimics the key pathological and clinical features of MS. EAE is induced by immunizing susceptible mice with myelin-derived antigens, triggering an inflammatory response that replicates the pathophysiology of MS (3). Patients with MS and EAE mice harbor gut dysbiosis, and this microbiota shift worsens EAE symptoms, highlighting the role of the gut–brain axis in disease pathogenesis (4-6).

Intestinal epithelial integrity and function are vital for gut homeostasis and immune regulation. Intestinal epithelial cells (IECs) directly interact with gut microbiota and immune cells, rendering them vulnerable to perturbations in the gut ecosystem of EAE mice (7). Continuous IEC proliferation and differentiation ensure rapid gut lining turnover and regeneration, which is critical for barrier function against luminal contents (8). Dysregulation of these processes can compromise gut barrier function, increase permeability and microbial product translocation, and affect systemic immunity.

Intestinal mitochondria maintain gut microbiota, metabolism, and immune homeostasis (9). IEC mitochondrial dysfunction alters gut microbiota composition, contributing to local and systemic inflammation, including CNS inflammation (9). Mitochondrial oxidative phosphorylation (OXPHOS) involves electron transport chains and adenosine triphosphate (ATP) synthesis, influencing IEC proliferation, differentiation, and function (10, 11). A metabolic gradient exists along the crypt–villus axis, reflecting changes in cellular energy demands (12). Intestinal stem cells primarily use glycolysis for rapid proliferation (13). As they migrate and differentiate into spe-cialized cells, their metabolism shifts and relies heavily on mitochondrial OXPHOS (14). This metabolic transition is crucial because differentiated IECs have high energy requirements for absorbing nutrients. Therefore, mitochondrial and OXPHOS disruption contribute to intestinal dysfunction and CNS-related diseases (9, 14).

Phenyllactic acid (PLA) is a microbiota-derived phenolic acid with antimicrobial, antioxidant, and anti-inflammatory activities. PLA is primarily produced through phenolic compound fermentation by lactic acid bacteria, particularly Lactobacillus murinus (15). PLA levels are altered in patients with MS, with lower concentrations found in the serum of those with benign MS than in those with relapsing-remitting MS or disease duration under 10 years (16). Emerging research suggests that PLA and other phenolic acids may modulate cellular metabolic pathways, affecting oxidative stress and mitochondrial health (17, 18). Understanding how PLA affects mitochondrial function could provide insights into its therapeutic potential, particularly in chronic diseases associated with mitochondrial dysfunction.

This study demonstrated that IEC dysfunction occurs during EAE induction, coinciding with mitochondrial OXPHOS pathway disruption. We demonstrated that PLA, a gut microbiota-derived metabolite, reduces the expression of OXPHOS protein complexes. Our findings suggest that understanding how microbial metabolites influence mitochondrial energy metabolism and IEC function provides novel insights into the pathogenesis of intestinal diseases and therapeutic strategies to restore mitochondrial function. Investigating this relationship could offer new approaches for understanding and treating auto-immune diseases.

RESULTS

Small intestinal epithelium damage upon EAE induction

Intestinal damage in autoimmune diseases occurs when the immune system aberrantly targets the gastrointestinal tract. To investigate whether the intestinal epithelial architecture is altered following EAE induction, a comparative pathophysiological analysis of intestinal tissues between EAE-induced and naïve control mice was conducted. Active EAE was induced using MOG_35–55 peptide, and clinical scores were monitored for 28 days. The results demonstrated that the mice developed a robust EAE phenotype (Fig. 1A). Small intestinal and colonic tissues were harvested for histopathological analysis using hematoxylin and eosin (H&E) staining at the initial peak of the disease symptoms. Histopathological examination revealed that small intestinal tissues from the EAE group exhibited significant structural damage, characterized by disrupted crypt-villus formation, increased inflammatory cell infiltration, and epithelial erosions, resulting in significantly elevated histopathological scores compared to those of the naïve control group (Fig. 1B).

Fig. 1.

Experimental autoimmune encephalomyelitis (EAE) mice exhibit intestinal epithelial damage. (A) Mean clinical scores of mice fol-lowing EAE induction with MOG_35–55 peptide. (B, C) Representative H&E staining and histopathology score of distal small intestine sections (B) and colon sections (C). (D) Representative PAS staining of distal small intestine sections and quantification of goblet cell numbers per crypt. (E) Representative Ki67 immunofluorescence staining of distal small intestine sections and quantification of Ki67^＋ cell numbers per crypt. Data are analyzed using Student’s t-test and represented as mean ± standard deviation (naïve n = 8, EAE n = 8), P < 0.01; **P < 0.001; ***.

In contrast, the histopathological scores of colonic tissues from EAE mice were comparable to those of naïve control mice (Fig. 1C). Furthermore, Periodic Acid-Schiff (PAS) staining revealed markedly depleted goblet cells, a differentiated secretory cell lineage responsible for mucin secretion, in the small intestinal tissues of EAE mice (Fig. 1D). Conversely, increased proliferative activity within small intestinal crypts following EAE induction was evidenced by enhanced Ki67-positive cell intensity and number, which predominantly represent intestinal stem cells (ISCs) and trans-amplifying cells (TAs) (Fig. 1E). This phenomenon likely represents a compensatory regenerative mechanism in response to intestinal epithelial injury. Collectively, these findings demonstrate that the small intestinal epithelium undergoes considerable structural and functional alterations following EAE induction.

Downregulation of proteins related to mitochondrial OXPHOS in the intestinal epithelial cells of EAE mice

Differentiated IEC subsets depend on mitochondrial OXPHOS as their metabolic pathway to support cellular differentiation and function (14, 19). In contrast, ISCs and TAs predominantly rely on glycolysis for energy supply, with supplemental contributions from OXPHOS (13, 20). As EAE mice exhibit abnormal villous formation and alterations in differentiated cell lineages, especially goblet cells, while inducing proliferative marker expression, we hypothesized that small intestinal damage following EAE induction may result from intestinal mitochondrial OXPHOS disruption. To test this hypothesis, we performed a comprehensive proteomic analysis of freshly isolated small intestinal epithelial cells (sIECs) from EAE and naïve control mice at the peak of EAE symptoms. Proteomic analysis revealed differential expressions of 4,371 proteins in sIECs derived from EAE mice compared to naïve controls, with 2,230 proteins downregulated and 2,141 proteins upregulated (Fig. 2A). To validate alterations in sIEC lineage composition, we first confirmed increased expression of Lyz, a Paneth cell marker (Supplementary Fig. 1A), consistent with previous reports demonstrating Paneth cell enrichment in EAE mice (21). Other lineage markers in the villi were also reduced upon EAE induction, including Vil and Dclk1, which are markers of enterocytes and tuft cells, respectively (Supplementary Fig. 1A). The expression of proliferative markers, including Ki67 and proliferating cell nuclear antigen, was increased, indicating active proliferation of sIECs derived from EAE mice (Supplementary Fig. 1B). Furthermore, stemness markers Olmf4 and Msi1 were enriched in EAE mice, suggesting that ISC stemness was induced to activate compensatory mechanisms in response to IEC damage. Concurrently, we observed a slight reduction in Muc2 and Fut2 expression, which are established markers of goblet cells (Supplementary Fig. 1A).

Fig. 2.

EAE mice-derived cecal content reduces OXPHOS complex protein expression in small intestinal organoids. (A) Representative images of small intestinal organoids at day 3, quantification of surface area ratio, and budding of organoids cultured with cecal content from naïve or EAE mice. (B) Western blot analysis of OXPHOS complex protein expression in small intestinal organoids cultured with cecal content from naïve or EAE mice. HSP-90 served as the loading control. Data present results from at least two independent biological experiments.

Furthermore, OXPHOS complex-associated proteins exhibited significant alterations in EAE-derived sIECs. Proteins associated with Complex I (NADH-ubiquinone oxidoreductase), Complex III (ubiquinol-cytochrome c oxidoreductase), Complex IV (cytochrome c oxidoreductase), and Complex V (ATP synthase) were downregulated, whereas proteins associated with Complex II (succinate-ubiquinone oxidoreductase) were upregulated (Fig. 2B). Specifically, seven Complex I subunit proteins were upregulated, whereas 15 Complex I-related proteins were downregulated in EAE mice (Fig. 2B). Two subunit proteins and three assembly proteins of Complex II were enriched upon EAE induction (Fig. 2B). For Complex III, four subunit proteins showed decreased expression in EAE mice (Fig. 2B). Regarding Complex IV, which represents the terminal protein complex in the electron transport chain in OXPHOS, two proteins were upregulated. In comparison, nine proteins were downregulated, including cytochrome c oxidase subunit 1 (P00397), subunit 4 (P19783), subunit 5A (P12787), subunit 5B (P19536), subunit 6A1 (P43024), subunit 6C (Q9CPQ1), subunit 7A1 (P56392), subunit 7A2 (P48771), and subunit 7C (P17665) (Fig. 2B, C). Finally, the expression of one Complex V subunit, the primary contributor to ATP production, was induced, while the expression of one Complex V membrane protein and four subunits, including ATP synthase subunit alpha (Q03265), subunit e (Q06185), subunit f (P56135), and subunit g (Q9CPQ8), were significantly inhibited in the sIECs derived from EAE mice compared to those derived from naïve mice (Fig. 2B, D). The distinct expression patterns observed in Complex II may represent a compensatory mechanism in response to a deficiency in Complex I protein expression (22).

Additionally, our results revealed that antioxidant proteins, glutathione peroxidase 2 and glutathione reductase, and xanthine dehydrogenase, which is associated with ROS production, were downregulated in sIECs derived from EAE mice compared to those derived from naïve mice (Supplementary Fig. 1C). These findings collectively suggest that mitochondrial OXPHOS disruption occurs in the small intestinal epithelial cells following EAE induction.

Depletion of mitochondrial OXPHOS in small intestinal organoids treated with EAE mice-derived cecal content

Patients with MS and EAE mice exhibit intestinal dysbiosis accompanied by metabolic alterations that may influence disease progression (6). Furthermore, through complex host interactions, the gut microbiota produces diverse metabolites that serve as critical modulators of epithelial homeostasis (23). Therefore, we hypothesized that alterations in the gut microbiota and microbial metabolites may contribute to OXPHOS dysregulation in sIECs following EAE induction. To test this hypothesis, we employed intestinal organoid cultures, which are three-dimensional in vitro models of the native intestine (24). The organoids were treated with cecal contents obtained from either EAE or naïve mice. Cecal content treatment did not affect organoid growth or bud formation (Fig. 3A). Moreover, Lgr5 expression level was upregulated, while the expression level of Ki67 was moderately induced in organoids treated with cecal content derived from EAE mice (Supplementary Fig. 2A). This result also strengthens the previous notion that the induction of proliferative markers observed in EAE mice is a compensatory mechanism for damaged IECs (Fig. 1E, Supplementary 1A). In addition, the expression level of Muc2 was reduced in organoids treated with EAE-derived cecal content compared to that in naïve-derived cecal content (Supplementary Fig. 2A).

Fig. 3.

Oxidative phosphorylation (OXPHOS) complex expression in small intestinal epithelial cells is altered following EAE induction. (A) Volcano plot of differentially expressed proteins in small intestinal epithelial cells. Downregulated and upregulated proteins were identified using a 1.2-fold change cutoff. (B) Heatmap of differentially expressed proteins related to the OXPHOS complexes of naïve and EAE mice. (C, D) Peak intensity of proteins related to OXPHOS Complex IV (C) and V (D). Data are analyzed using Student’s t-test and represented as mean ± standard deviation (naïve n = 7, EAE n = 7) mice, P < 0.05; *P < 0.01; **P < 0.001; ***P < 0.0001; ****.

To determine whether microbial metabolites in cecal content induce changes in intestinal epithelial cell mitochondrial OXPHOS, the OXPHOS-related proteins expression was assessed using western blot analysis. The intensity of protein bands corresponding to mitochondrial OXPHOS complexes V, III, IV, and I was reduced in intestinal organoids treated with cecal content from EAE mice compared to those treated with cecal content from naïve mice (Fig. 3B). These results suggest that alterations in microbial metabolites within the cecal content may induce mitochondrial OXPHOS dysfunction in intestinal epithelial cells.

In addition, to determine whether alterations in microbial metabolites affect mitochondrial ROS production in IECs, we performed quantitative mRNA expression analysis of ROS production-related enzymes (Xo and Nox1) and antioxidant-related enzymes (Sod1, Gpx2, and Gsr) in the intestinal or-ganoids treated with cecal content derived from EAE or naïve mice. The results demonstrated that the expression levels of Nox1 were significantly increased, whereas the expression levels of Sod1 and Gpx2 were significantly downregulated in organoids treated with EAE-derived cecal content compared to those treated with cecal content from naïve mice (Supplementary Fig. 3). These findings suggest that alterations in microbial metabolite levels directly promote mitochondrial ROS production while simultaneously diminishing IEC antioxidant capacity.

Metabolomic analysis identified elevated 3-phenyllactic acid levels in EAE mice

Our results demonstrated that IECs treated with EAE-derived cecal content exhibited mitochondrial OXPHOS dysfunction. Therefore, we investigated specific changes in microbial metabolites following EAE induction that may contribute to alterations in mitochondrial OXPHOS. Metabolomic analysis of cecal contents from EAE or naïve mice was performed to identify these changes. Our analysis revealed that 250 microbial metabolites were downregulated and 285 were upregulated in the cecal contents of EAE mice compared to naïve controls (Fig. 4A). The levels of phenolic acids, including 3-phenyllactic acid (PLA), ferulic acid, and cinnamic acid, were significantly elevated in the cecal content of EAE mice compared to that in naïve mice (Fig. 4B, C). Among these phenolic acids, PLA is a microbiota-derived metabolite produced primarily by lactic acid bacteria. L. murinus has been previously identified as the primary producer of PLA in the small intestine (15). Therefore, we quantified L. murinus abundance in stool samples using qPCR. Our analysis showed that L. murinus abundance was higher in the stool of EAE mice than in naïve mice (Fig. 4D). These findings suggest that elevated levels of microbial metabolites, particularly PLA, may contribute to the mitochondrial OXPHOS dysfunction observed in IEC during EAE.

Fig. 4.

Elevated 3-phenyllactic acid (PLA) was identified using metabolomic analysis in EAE mice. (A) Volcano plot of differentially expressed metabolites from untargeted metabolomic analysis of cecal content from naïve and EAE mice. Downregulated and up-regulated metabolites were identified using a 1.5-fold change cutoff. (B) Heatmap of phenolic acids in cecal content comparing EAE mice to naïve mice. Data are normalized using auto-scaling. (C) Peak intensity of PLA in cecal content from naïve and EAE mice. (D) Relative abundance of Lactobacillus murinus in the stool samples of naïve and EAE mice. Data present results from at least two independent biological experiments, P < 0.05; *.

Microbiota-derived 3-phenyllactic acid directly impairs mitochondrial OXPHOS in intestinal epithelial cells

Given the elevated concentration of PLA and its producers in EAE mice, we further investigated whether PLA induces mitochondrial OXPHOS dysfunction using small intestinal organoids. PLA treatment did not affect organoid growth or bud formation. Lgr5 and Ki67 expression levels were up-regulated, whereas Muc2 was downregulated in organoids treated with 1 μM PLA compared to the vehicle. (Supplementary Fig. 2B). These data suggest that PLA promotes ISC stemness and compensatory activation, similar to the effects observed in EAE mouse-derived cecal content (Fig. 5A). Subsequently, we determined the expression of OXPHOS protein complexes in PLA-treated organoids using western blot analysis. We found that the overall expression levels of protein complexes, particularly Complexes I, III, and IV, were reduced following PLA treatment compared to vehicle treatment (Fig. 5B). To further validate the mitochondrial OXPHOS defect, the oxygen consumption rate (OCR) of the Caco-2 intestinal epithelial cell line treated with either PLA or vehicle was measured. Seahorse analysis showed that PLA-treated Caco-2 cells had decreased OCR, especially at maximal respiration (Fig. 5C), which corresponded with reduced OXPHOS Complex IV. Collectively, these results suggest that PLA, which is elevated in EAE mice, can directly inhibit mitochondrial OXPHOS in IECs by suppressing the expression of protein complexes I, III, and IV.

Fig. 5.

PLA mediates OXPHOS dysfunction in intestinal organoids and intestinal epithelial cells. (A) Representative images of small intestinal organoids at day 3, quantification of surface area ratio, and budding of organoids treated with vehicle PLA. (B) Western blot analysis of OXPHOS complex protein expression in small intestinal organoids treated with vehicle or PLA. HSP-90 served as the loading control. (C) Seahorse analysis of OCR with basal respiration, maximum respiration, ATP production, and spare respiratory capacity graph in Caco-2 cells treated with either PLA or vehicle. Data present results from at least two independent biological experiments, P < 0.05; *.

DISCUSSION

Bowel dysfunction, including constipation, fecal incontinence, and diarrhea, is a prevalent clinical manifestation in patients with MS and significantly impacts their quality of life (2). However, the underlying pathophysiological mechanisms remain poorly understood. Using an EAE murine model, our study demonstrated that EAE mice exhibit sIECs dysfunction, primarily characterized by compromised intestinal barrier integrity and enhanced cellular proliferation. Following EAE induction, alterations in gut microbiota-derived metabolite levels were observed. PLA was enriched in the cecal content of EAE mice and contributed to the disruption of the mitochondrial OXPHOS pathway in IECs (Fig. 6). These findings suggest a potential mechanism for intestinal barrier dysfunction due to altered gut microbiota metabolism and provide insights into bowel dysfunction in patients with MS.

Fig. 6.

PLA mediates OXPHOS dysfunction in small intestinal epithelial cells of EAE mice.

sIECs comprise diverse IEC subsets, including stem cells, Paneth cells, goblet cells, and enterocytes. Female EAE mice exhibit elevated lysozyme secretion from Paneth cells due to dopamine D2 receptor activation, and lysozyme administration worsens disease severity (21). Our findings reveal an additional mechanism by which IEC subset populations are altered in EAE mice, driven by changes in microbiota metabolism. Treatment with cecal contents enriched in microbial metabolites reduced the number of mucus-producing goblet cells and increased the number of Ki67^＋ proliferating cells and Paneth cells. The potential role of IEC dysfunction in modulating EAE pathogenesis warrants further investigation.

sIECs’ respiration predominantly relies on OXPHOS to support nutrient absorption and IEC differentiation, whereas stem cells use energy from glycolysis and OXPHOS for proliferation (14). Our study revealed reduced expression of OXPHOS-dependent IEC lineages, such as goblet cells, enterocytes, and tuft cells (19), with increased expression of glycolysis-reliant lineages (20), including ISCs and Paneth cells (11, 13). Furthermore, EAE mice exhibited mitochondrial OXPHOS dysfunction, characterized by decreased expression of respiratory subunits. Similar reductions in OXPHOS complex subunit levels have been observed in neurons and immune cells of patients with MS (25, 26). This dysfunction may result from several factors, including mitochondrial ROS accumulation, as previously documented in EAE mice (27). However, whether this dysfunction directly results from ROS accumulation remains unclear.

We observed that OXPHOS protein destabilization was driven by EAE-associated gut microbiota metabolites, as cecal content treatment in organoids similarly induced OXPHOS complexes downregulation. Phenolic acids were significantly enriched in the cecal content of EAE mice compared to that of naïve mice, with PLA showing a 30-fold elevation. In clinical samples, PLA levels were detected in the serum of patients with relapsing-remitting MS (RRMS) and those with MS duration of less than 10 years, compared with patients with benign MS. However, no differences in PLA concentrations were observed between healthy donors and patients with RRMS or secondary progressive MS (16). Conversely, another study reported that patients with MS had lower serum PLA levels than healthy controls (28). Despite reduced serum PLA levels, intestinal PLA concentrations may follow different patterns, as patients with MS typically have intestinal dysbiosis (4-6). Furthermore, L. murinus, a major PLA producer, was more abundant in EAE mice. The Lactobacillus comprises 6% of duodenal bacteria but only 0.3% of colonic bacteria (29), supporting small intestinal-selective IEC damage. These findings suggest that PLA abundance at specific anatomical sites may yield distinct outcomes, and its presence in the small intestine may specifically correlate with bowel dysfunction in altered gut microbiota contexts.

Phenolic acids induce ROS production and reduce the oxygen consumption rate (OCR) in endothelial cells (30). Conversely, other studies have reported that PLA may enhance mitochondrial function and ameliorate certain diseases (18). Although the role in mitochondrial respiration has yielded contradictory results (17), previous research has shown that PLA increases ROS accumulation and downregulates Complex I and IV mitochondrial proteins in Rhizopus oryzae (31). Our results showed that PLA treatment induces mitochondrial OXPHOS dysfunction by reducing the OXPHOS complexes expression. Therefore, additional studies are needed to investigate whether microenvironmental factors under EAE conditions influence PLA’s role in modulating mitochondrial function. Besides PLA, other phenolic acids were also elevated in EAE mice cecal content. However, we did not explore whether these metabolites exert similar effects to OXPHOS complexes as PLA.

Our data revealed that PLA disrupted mitochondrial OXPHOS respiration by downregulating the expression of OXPHOS complexes. However, whether the PLA directly induces enterocytes and ISC dysfunction was not addressed. PLA has been identified as a peroxisome proliferator-activated receptor gamma (PPARγ) ligand, and PPARγ activation impairs Lgr5^＋ ISC function (32). We observed elevated PPARγ expression in sIECs of EAE mice at the peak of the disease (data not shown). Therefore, PLA may disrupt IEC homeostasis via distinct mechanisms: OXPHOS inhibition or PPARγ interaction. PLA’s role in altering IECs’ barrier function, particularly its effects on IEC stemness, requires further investigation.

In conclusion, our study demonstrates a mechanism by which IEC dysfunction occurs in the EAE model, potentially reflecting bowel dysfunction in patients with MS. IEC disruption in EAE mice is mediated by mitochondrial OXPHOS dysfunction, potentially driven by gut microbiota-derived metabolite PLA. This study enriches our understanding of the gut–brain axis in the EAE model, revealing how CNS neuro-inflammation impacts IECs and mucosal barrier integrity.

MATERIALS AND METHODS

EAE induction

Nine-week-old C57BL/6L female mice were purchased from ORIENT Bio Korea (Korea) and maintained in a specific pathogen-free animal facility at the Soonchunhyang Institute of Medi-bio Science. All experimental procedures were conducted in accordance with institutional guidelines and approved protocols. Active EAE induction was conducted using the Hooke Kit (Hooke Laboratories, EK-2110). Each mouse received a subcutaneous (s.c.) injection of 200 μg MOG_35–55 peptide emulsified in 200 μl of Complete Freund’s Adjuvant. Pertussis toxin (200 ng per mouse; List Biological Laboratories; supplied with Hooke Kit) was intraperitoneally (i.p) administered at 2 h and 24 h post-immunization. EAE clinical progression was evaluated daily as previously described (33). Disease progression was quantified as the mean clinical score calculated from the day of immunization (day 0) through the experimental endpoint (day 27).

Intestinal histopathology

Distal small intestine and colon tissues were harvested from naïve and EAE mice. After fat removal, the specimens were carefully washed with phosphate-buffered saline (PBS; Corning, NY). The tissues were subsequently fixed in Carnoy’s solution for 2 h at room temperature, and then moved to 20% sucrose (BioShop, Canada) dissolved in PBS overnight at 4°C. Specimens were embedded in OCT compound (Sakura, UK).

For hematoxylin and eosin (H&E) staining, intestinal tissues were sectioned at 7 μm thickness using a cryostat. Sections were stained with hematoxylin, counterstained with eosin, and dehydrated using a graded ethanol-xylene series. Slides were coverslipped using Permount mounting medium (Fisher Scientific, Belgium). Histological evaluation was performed using a scoring system assessing four parameters: epithelial surface erosion, distortion of villous and crypt architecture, presence of infiltrating inflammatory cells, and reduction in villous height. Each parameter was scored on a scale of 0-5 as previously described (34).

Goblet cell staining

Tissue sections prepared as described above were subjected to alcian blue (AB) and Periodic Acid-Schiff (PAS) staining for mucus detection. The number of positively stained goblet cells was quantified per individual crypt. Cell counts were performed systematically across multiple representative fields to ensure accurate quantification.

Immunofluorescence staining

Five-micrometer-thick sections of OCT-embedded distal small intestinal tissue were used for immunofluorescence analysis. Tissue sections were permeabilized with 0.05% Tween-20 (Sigma-Aldrich, MO) in PBS (PBS-T), followed by blocking with 5% bovine serum albumin (BSA; Sigma-Aldrich, MO) in PBS-T for 1 h at room temperature. Primary antibody incubation was performed using rabbit anti-Ki67 antibody (Cell Signaling Technology, MA) diluted in 1% BSA in PBS-T overnight at 4°C. Sections were washed with PBS and incubated with Alexa Fluor 488-conjugated anti-rabbit IgG secondary antibody (A11008, Life Technologies, MD) diluted in 1% BSA in PBS-T for 1 h at room temperature. Nuclei were counterstained with 4’,6-diamidino-2-phenylindole (DAPI; Sigma-Aldrich, MO). The fluorescence imaging was performed using a confocal laser scanning microscope (LSM 710; Carl Zeiss, Germany). Quantification of Ki67-positive cell number per crypt was conducted using ImageJ software (NIH, USA; version 1.54).

Intestinal epithelial cell isolation

IECs were isolated from small intestinal tissues of naïve and EAE mice according to a previous established protocol (35). Small intestinal tissues were harvested, cleaned, and sectioned into small fragments. Specimens were thoroughly washed with ice-cold PBS, then transferred to tubes containing 20 ml of digestion solution (5 mM ethylenediaminetetraacetic acid (EDTA; Sigma-Aldrich, MO) and 1 mM dithiothreitol (DTT; Sigma-Aldrich, MO) in ice-cold PBS. Tubes were agitated at 150 rounds per minute (rpm) at 37°C for 20 min. Subsequently, tissues were transferred to a tube containing 5 ml of 30 mM EDTA in ice-cold PBS, agitated at 200 rpm and at 37°C for 10 min, and the digested solution was filtered through a 70 μm cell strainer. The filtered solution was centrifuged at 1,000 × g at 4°C for 10 min. The resulting pellet was washed with PBS containing calcium and magnesium (Corning, NY) supplemented with 10% fetal bovine serum (FBS; Thermo Fisher Scientific, MO). The IEC pellet was resuspended in complete growth medium consisting of RPMI-1640 (Corning, NY) supplemented with 10% FBS, 1% penicillin–streptomycin (Corning, NY), 1% MEM nonessential amino acids (Corning, NY), and 0.1% β-mercaptoethanol (Thermo Fisher Scientific, Waltham, MO).

Mass spectrometry sample preparation

Chilled urea lysis buffer, containing 8 M urea, 75 mM NaCl, 50 mM Tris (pH 8.0), 1 mM EDTA, 2 μg/ml aprotinin, 10 μg/ml leupeptin, 1 mM PMSF, 1:100 (v/v) phosphatase inhibitor cocktail 2, 1:100 (v/v) phosphatase inhibitor cocktail 3, and 10 mM NaF (4°C, 300 μl) was added to each cell sample. The lysis buffer was mixed on the highest setting for 5-10 s, then incubated at 4°C for 15 min twice. The lysis buffer solution was centrifuged at 10,000 × g at 4°C for 10 min to clear the lysate. Proteins were denatured with 5 mM DTT at 25°C for 1 h, and then free thiol groups were alkylated with 12.5 mM iodoacetamide at 25°C for 1 h in the dark. Protein samples were treated with sequencing-grade trypsin (Promega, V5111) at a 1:20 enzyme-to-protein ratio (wt/wt) overnight at 25°C after dilution in 2 M urea with 50 mM Tris-HCl, pH 8.0. The digestion was quenched by adding formic acid to a final concentration of 3%. The total peptide concentration of each sample was measured by the bicinchoninic acid (BCA) assay after desalting the digest by solid-phase extraction (SPE; Waters [Sep-Pak, tC18 cartridges]) cleanup. The same amount of peptide from each sample, including two pooling samples made from 14 samples, was then labeled with TMT-16plex reagents according to the manufacturer’s protocol. After 1 h incubation, 5% aqueous hydroxylamine solution was added to stop the labeling reaction. The 16 TMT-labeled samples were combined, dried, and reconstituted in water containing 0.1% formic acid. The combined TMT 16plex-labeled sample was desalted using SPE, dried, and dissolved in water.

High pH reversed-phase liquid chromatography for peptide fractionation

The dissolved TMT 16 plex labeled sample was fractionated using an XBridge BEH Shield RP18 Column (130 Å, 2.5 μm, 4.6 × 150 mm, Waters) on NexeraXR HPLC (Shimadzu) as previously described (36). A total of 40 or 42 fractions were collected using an FRC-10A fraction collector (Shimadzu), after the elution started with an interval of 1 min for each fraction. The 40 fractions were concentrated into 20 fractions. The concatenated fractions were dried and kept at −80°C until LC-MS/MS analysis.

Liquid chromatography tandem mass spectrometry (LC-MS/MS)

Twenty fractions were analyzed using an LC-MS/MS system consisting of an UltiMate 3000 RSLnano System (Thermo Fisher Scientific) and an Orbitrap Fusion Lumos mass spectrometer with a nano-electrospray source (EASY-Spray Sources), as previously described with minor modifications (36). The LC was performed with the following gradient: 5% B, 15% B over 6 min; after 9 min, 20% B over 75 min; 30% B for 10 min; 95% B over 1 min; constant hold for 8 min; 5% B over 1 min. To produce an electrospray, 1800 V was used.

Protein identification and quantitation

The Integrated Proteomics Pipeline using built-in search engines (IP2, version 6.5.5, Integrated Proteomics) was used for data analysis with the UniProt-reviewed mouse protein database (downloaded 27th September, 2023) as previously described, with some modifications (36). ProLucid (37) was used to identify the peptides. Tandem mass tag modification (+ 304.2071) at the N-terminus and lysine residue by the labeling reagent and carbamidomethylation at cysteine were chosen as static modifications. The output data files were filtered and sorted to compose the protein list using DTASelect (The Scripps Research Institute, USA), with at least 2 peptide assignments per protein for protein identification and a false-positive rate < 0.01 (38).

A quantitative analysis was conducted using the Census module in the IP2 pipeline (Integrated Proteomics Pipeline Ver. 6.5.5, USA) with only the unique peptides (36, 39). Perseus platform (version 1.6.15.0) was used following data processing. The protein intensity was entered into Perseus, transformed to log2, and normalized to the column median for protein quantification. An analysis of variance (ANOVA) test was performed to select significant proteins with permutation FDR < 0.05 across groups.

Cecal content supernatant preparation

The cecal contents were lyophilized overnight to obtain dried pellets. Dried cecal contents (100 mg) were reconstituted in 1 ml of serum-free DMEM/F12 medium (Gibco, MA) and vortexed for 1 h at room temperature. The suspension was centrifuged at 4,000 rpm for 10 min, and the resulting supernatants were filtered through 0.22 μm syringe filters.

Intestinal organoid culture, treatment of cecal contents, and metabolites

The small intestines of naïve mice were dissected, sectioned into 0.5 cm pieces, and washed with ice-cold PBS to remove luminal debris. For intestinal crypt isolation, the tissues were incubated with Gentle Cell Dissociation Reagent (StemCell Technologies, MA) and filtered through a 70 μm cell strainer. Isolated crypts were mixed with Matrigel (BD Biosciences, NJ) at a 1:1 ratio (vol/vol) and plated in a dome-shaped configuration in 48-well plates. Following Matrigel polymerization by incubation at 37°C for 10 min, 400 μl Intesticult^TM OGM mouse basal medium (StemCell Technologies, MA) was added to each well. Organoid formation was monitored daily using a light microscope, and the culture medium was replaced every 2-3 days. Organoids were passaged every 5-7 days and harvested using Cell Recovery Solution (Corning, NY). Organoid pellets were washed twice with ice-cold PBS prior to analysis.

For cecal content and metabolite treatment, organoids were cultured in Intesticult^TM OGM mouse basal medium with 0.01% cecal content supernatant, 1-100 μM PLA (Sigma-Aldrich, MO), or vehicle for 72 h, after which organoids were collected for analysis. Organoid surface areas and budding were calculated using ImageJ software (NIH, USA; version 1.54).

Western blotting

Organoids were collected and incubated with lysis buffer (50 mM Tris-HCl, pH 7.5, 5 mM EDTA, 150 mM NaCl, 1% NP-40, 0.5% sodium deoxycholate) supplemented with 1x protease inhibitor cocktail (BMS, Korea) on ice for 30 min. The lysate was subsequently centrifuged at 13,000 rpm for 10 min at 4°C. Protein concentration in samples was determined using Bio-Rad protein assay (Bio-Rad, CA). Twenty micrograms of protein were mixed with Laemmli sample buffer (Bio-Rad, CA), and samples were denatured at 50°C for 10 min. Protein samples were subjected to sodium dodecyl sulfate poly-acrylamide gel electrophoresis (SDS-PAGE) and transferred to polyvinylidene difluoride membrane (PVDF, Amersham^TM, GB). PVDF membranes were incubated with either total OXPHOS rodent WB antibody cocktail (1:200 dilution; ab110413, Abcam, CA) or HSP90 alpha/beta antibody (1:1000 dilution; SC-13119, Santa Cruz, TX). Membranes were subsequently incubated with anti-mouse HRP-conjugated secondary antibody (BioLegend, CA). Proteins were visualized by using enhanced chemiluminescence (ECL, Bio-Rad, CA).

Metabolomic analysis

Cecal contents from naïve and EAE mice were collected at disease peak for metabolomic analysis. Metabolite extractions were performed using liquid-liquid extraction with some modifications (40). Briefly, 400 μl of chloroform and methanol (2:1 ratio) was added to the cecal contents, which were homogenized using a TissueLyzer (Qiagen), followed by centrifugation for 15 min. Nonpolar metabolites were recovered from the organic phase, whereas polar metabolites were recovered from the aqueous phase. The collected fractions were dried using vacuum centrifugation and reconstituted in 50 μl of mobile phase A (0.1% formic acid in water) prior to liquid chromatography-mass spectrometry (LC-MS/MS).

The LC-MS/MS system was equipped with a Dionex Ultimate3000 and an LTQ-Orbitrap XL (Thermo Fisher Scientific) in positive and negative ion modes. For the organic phase solution, a reverse-phase column (Pursuit5 C18, 3 μm, 150 × 2.1 mm) was used, with 0.1% formic acid in water as mobile phase A and 0.1% formic acid in methanol as mobile phase B, following the gradient previously described (40). For the aqueous phase, a hydrophilic interaction liquid chromatography column (Waters XBridge BEH amide, 2.5 μM, 2.1 × 150 mm) was used as previously described (40). Both LC experiments were performed at 200 μl/min and 25°C.

Volcano plots were generated to visualize the significantly different metabolites between cecal contents from naïve and EAE-induced mice using the EnhancedVolcano R package with a |fold change| ≥ 1.5 and P-value < 0.05.

Quantitative polymerase chain reaction (qPCR) analysis

Stool samples from naïve and EAE mice were collected aseptically at the peak of the disease. Stool DNA was extracted according to the manufacturer’s protocol using the Exgene Stool DNA Mini Kit (GeneAll, Korea). qPCR was performed using SYBR Green Real-time PCR Master Mix Kit (TOYOBO, Japan) with Lactobacillus murinus 16s rRNA-specific primers: forward (5’-GCA ATG ATG CGT AGC CGA AC-3’) and reverse (5’- GCA CTT TCT TCT CTA ACA ACA GGG-3’). PCR reactions were conducted using the ABI StepOne Plus real-time PCR system (Applied Biosystems^TM, CA). Target gene expression was normalized to universal 16S rRNA gene expression using primers 334F (5’- ACT CCT ACG GGA GGC AGC AGT-3’) and 514R (5’- ATT ACC GCG GCT GCT GGC-3’).

Seahorse analysis

Caco-2 cells (50,000 cells) were cultured in Minimum Essential Medium (Cellgro, NY) supplemented with 10% FBS (Corning, NY), 1% penicillin/streptomycin (Corning, NY), 25 mM HEPES (Gibco, MA), and 25 mM NaHCO₃ (Gibco, MA) and were plated on the Seahorse XFp cell culture miniplates. Caco-2 cells were treated with either 100 μM PLA or vehicle for 24 h. Prior to Seahorse assay, the medium was changed to Seahorse XF DMEM medium supplemented with Seahorse XF supplements, including glucose, glutamine, and pyruvate (Agilent Technologies, CA). Seahorse was performed using a Seahorse XF HS Mini Analyzer (Agilent Technologies, CA) according to the manufacturer’s protocol, with 2 μM oligomycin (Sigma-Aldrich, MO), 0.025 μM Carbonyl cyanide p-trifluoromethoxyphenylhydrazone (FCCP, Sigma-Aldrich, MO), and 0.5 μM rotenone/antimycin (Sigma-Aldrich, MO).

Statistical analysis

Statistical analyses of general observations were conducted using GraphPad Prism 10 (GraphPad, CA) with a Student’s t test. Values are presented as the mean ± standard deviation (SD). P < 0.05 was considered statistically significant.