New Results

Human microbiome-derived peptide affects the development of experimental autoimmune encephalomyelitis via molecular mimicry

Xin Ma, Jian Zhang, Qianling Jiang, Yong-Xin Li, Guan Yang
doi: https://doi.org/10.1101/2024.07.05.602171

Abstract

Background Gut commensal microbiota has been identified as a potential environmental risk factor for multiple sclerosis (MS), and numerous studies have linked the commensal microorganism with the onset of MS. However, little is known about the mechanisms underlying the gut microbiome and host-immune system interaction.

Results Here, we introduce the concept of molecular mimicry to address this issue by mining human microbial-derived peptides based on their similarity to the MHC II-TCR binding pattern of self-antigens. We analyzed 304,246 human microbiome genomes and 103 metagenomes collected from the MS cohort and identified 731 nonredundant analogs of myelin oligodendrocyte glycoprotein peptide 35-55 (MOG35-55). Of note, half of these analogs could bind to MHC II and interact with TCR through structural modeling of the interaction using fine-tuned AlphaFold. Among the 8 selected peptides, the peptide (P3) derived from human gut commensal Akkermansia muciniphila shows the ability to activate MOG35-55-specific CD4+ T cells in vitro and exacerbate the development of experimental autoimmune encephalomyelitis (EAE) in mice. Furthermore, dendritic cells could process and present P3 to MOG-specific CD4+ T cells and activate these cells. Collectively, our data suggests the potential involvement of a MOG35-55-mimic peptide derived from the gut microbiota as a molecular trigger of EAE pathogenesis.

Conclusions Our findings offer direct evidence of how microbes can initiate the development of EAE, suggesting a potential microbiome-based therapeutic target for inhibiting the progression of MS.

Introduction

Multiple sclerosis (MS) is an autoimmune disease characterized by the involvement of autoreactive T cells targeting myelin antigens, leading to the destruction of myelin and damage to axons, resulting in neurologic syndromes and physical disability (1). The global prevalence of MS stands at approximately 2.8 million individuals (2), with an increasing incidence observed in developing countries and among children (3). Although the etiology of MS remains unknown, both genetic and environmental factors have been implicated in its development (1). Accumulating evidence suggests that gut microorganisms play a pathogenic role in MS (4, 5). Research has found notable differences in the composition of gut bacteria between MS patients and healthy individuals (6, 7). Similar alterations in the gut microbiota, including elevated levels of Akkermansia muciniphila (A. muciniphila), have also been observed in experimental autoimmune encephalomyelitis (EAE), an animal model employed to investigate MS (8). However, how the alterations in the gut microbiota might affect MS/EAE development remains largely unknown.

Molecular mimicry is one of the leading mechanisms in linking the role of microbiota in immune reactivity in autoimmune diseases and cancer (9–13). Reassuringly, it has been reported that microbiota-derived peptides resembling myelin basic protein (MBP) can activate autoreactive T cells in MS (14–16). MBP, along with proteolipid protein (PLP) and myelin oligodendrocyte glycoprotein (MOG), are myelin antigens targeted by HLA-DR-restricted T cells, leading to MS development (1). Of note, MOG has been considered a primary target of cellular and humoral immune responses in multiple diseases due to its special extracellular immunoglobulin domain, which is directly accessible to binding partners including antibodies (17–19). Furthermore, MOG 35-55 (MOG35-55) peptide has been widely used to induce EAE. Intrigued by these findings, we hypothesized that microbiome-derived peptides resembling the MOG35-55 epitope could trigger the autoimmune response responsible for initiating EAE.

To address this hypothesis, we analyzed human microbiome genomes and metagenome databases collected from MS cohort. Our aim was to identify microbial peptides with MHC II-TCR recognition patterns resembling those of MOG35-55, which is presented by the MHC class II I-Ab and induces chronic EAE in C57BL/6J mice (20). Subsequently, we identified, synthesized, and tested 8 of these peptides for their ability to activate MOG35-55-specific T cells. Our findings revealed that a predicted peptide (P3) derived from the gut commensal organism A. muciniphila can activate mouse MOG35-55-specific T cells in vitro, albeit to a lesser extent compared to the MOG35-55 peptide. Additionally, P3 immunization can exacerbate the development of EAE. We also observed some cross-reactivity between MOG35-55 peptide and CD4+ T cells obtained from mice immunized with P3. Furthermore, dendritic cells (DCs) were able to process P3 and present it to CD4+ T cells isolated from 2D2 mice expressing a MOG-specific T cell receptor. Our findings suggest that microbial peptide derived from the gut commensal, resembling MOG35-55, may mimic the native myelin peptide and potentially contribute to the EAE development.

Results

Well-defined peptide-MHC II-TCR binding motifs guided microbiome-derived epitope discovery from the human microbiome

To uncover the microbiome-derived epitopes resembling the immunodominant MOG35-55 peptide, we designed a genome mining workflow involving an analysis of a vast dataset comprising 304,246 human microbiome genomes (21, 22) and 103 metagenomic samples from a MS study cohort (23) (Fig 1A). We first extracted 404,181 MOG35-55 analogs (Supplementary table 1) from these resources, followed by a filtering process that retained 21,557 analogs (3,946 non-redundant sequences) possessing well-defined binding residues at specific sequential spacing (Fig 1B, Supplementary table 2).

Figure 1.
Figure 1. Overview of MOG35-55 analogs identified in the human microbiome.

(A) The workflow of human microbiome-derived peptides discovery. (B) The proportion of MOG35-55 analogs with well-defined binding motifs. (C) Chord diagram showing the taxonomy distribution of prioritized analogs. The values in brackets show the total number of analogs with corresponding binding motifs. per genome in each phylum. The counts of analogs in each phylum were normalized by the genome counts. The scale was the proportion of analogs in each motif (upper) or in each phylum. Of note, the genomes less than 10 counts from Firmicutes_I, Firmicutes_B, Bacillota_B, Cyanobacteria, Desulfobacterota_I, Fusobacteriota, Pseudomonadota, Euryarchaeota, Halobacterota, Patescibacteria, and 19 unclassified contigs were merged into Others.

To gain insight into the phylogenetic distribution of MOG35-55 analogs derived from the human microbiome, we examined the 21,557 analogs with well-defined binding motifs. Those analogs exhibited a wide distribution across 24 phyla, with the occurrence of a specific motif per genome varying from 0 to 1 within each phylum (Fig 1C). Of note, 10 phyla with fewer than 10 genomes and 19 unclassified bacteria-derived contigs at the phylum level were grouped as “Others”. Interestingly, Verrucomicrobiota demonstrated a higher likelihood of encoding analogs featuring the YRxxFxRVx motif. Besides, these analogs were predominantly encoded by species such as A. muciniphila or A. muciniphila.B, which are enriched in patients with MS (24).

Microbiome-derived peptide-MHC II-TCR complexes structure-prediction

Following the general model of antigen presentation (25, 26), we employed a fine-tuned AlphaFold with a binding classifier (27) to discriminate ligands with mouse MHC II molecules (H2-Ab) binding core peptides. Subsequently, the binding core peptides were subjected to modeling the peptide-MHC II-TCR complex by TCRdock (28). First, the performance of fine-tuned AlphaFold and TCRdock was assessed using the MOG35-55 peptide as a test case (Fig 2). MOG35-55 peptide was further chopped into 10 unique putative binding core peptide candidates (Fig 2A) for subsequent analyses. This was necessary because the current state-of-the-art machine learning-based structure predictors (27–30) for the peptide-MHC II-TCR complex are designed to predict binding between a 9-mer binding core peptide and its receptor (MHC II or TCR). Finally, the binding core peptide of MOG35-55, with the specific sequence “WYRPPFSRVVH”, was predicted to be MHC II binders (Fig 2A), and this binding core peptide exhibited reliable interactions with four TCRs that contained different gene segments (Fig 2B). It is noteworthy that this binding core peptide with the sequence “WYRPPFSRVVH” could interact with mouse MHC II and TCR is consistent with the previous studies (20, 31). Inspired by these results, we integrated these two methods to prioritize the microbiome-derived epitopes.

Figure 2.
Figure 2. The confidence of hMOG35-55 binding core peptide-MHC II and peptide-MHC II-TCR complexes predicted TCRDock.

The scatter plots show the pLDDT value of a peptide versus an averaged inter-chain PAE of peptide-MHC-II complexes (A) and peptide-MHC II-TCR complexes (B) for binding core peptide candidates chopped from human MOG35-55. The predicted MHC II binding peptides (pLDDT ≥ 90, inter-PAE ≤ 4.34) are highlighted in red. The sequence of binding core peptide candidates is listed on the right side of panel A. The modeling of a peptide with pLDDT≥ 70 is considered as confident.

To mimic the process of proteolytical degradation of exogenous proteins before their presentation to MHC II in antigen presentation (26), we truncated the corresponding peptide windows (∼ 21 residues) containing the binding motifs along these 3,946 analogs, resulting in 731 nonredundant putative ligand sequences (Fig 3A, Supplementary table 2, 3). These ligands were further chopped into 5,783 unique putative binding core peptide candidates (Supplementary table 3) for subsequent analyses. Consequently, we identified 493 binding core peptides from 439 ligands as MHC II binders (Fig 3B, Supplementary table 4). In addition, we successfully modeled interactions between peptide-MHC II and 4 types of TCRs for 97.8% of the MHC II binders using 4 peptide-MHC-TCR complexes templates (Supplementary Fig 1 and Fig 2, Supplementary table 4). To determine the binding specificity, binding motifs are generated from these binding core peptides and characterized by anchor positions (Fig 3C and 3D). The motifs revealed that Y/F, R, F, R, and V residues are enriched at well-defined anchor positions (loc_1, loc_2, loc_5, loc_7, loc_8) in this study. Surprisingly, several noncanonical binding motifs starting with “L”, “A”, “I”, “S”, “N”, “H”, “Q”, “V”, “T”, “M”, “W”, and others were also observed (Fig 3C). These findings suggest that microbiome-derived epitopes could interact with MHC II and TCRs with diverse binding characteristics.

Figure 3.
Figure 3. Polymorphism in 9-mer binding core peptides of the predicted peptide-MHC II complex.

(A) The stacked bar plot shows the number of nonredundant ligands with 11 amino acids residues truncated from analogs, colored with different motifs. (B) The scatter plot shows the pLDDT value of a binding core peptide (corresponding to local structural accuracy) versus an averaged MHC II-peptide complex’s inter-chain PAE score (corresponding to the peptide-MHC II interactions). The predicted binder is highlighted in red. (C) The frequencies of amino acid residues (rows) at specific positions (columns) among the predicted MHC II binding peptides. The loc_1 ∼ loc_9 represent the binding core region among the binder, and loc_0 and loc_10 represent the flanking region of binders. (D) Representative sequence logos illustrating peptide residue preference of MHC II. Each logo is labeled with the starting amino acid residue of binding core peptides, and the number of binding core peptides used to generate the logos is paired to the corresponding cells in panel [B]. The logos are colored by the chemical polarity of the amino acid residue. In panels C and D, sequence/residue numbers are 0-indexed: they start at 0 and end with N-1, where N is the length of the sequence.

Microbiome-derived peptide stimulates MOG35-55-specific T cell

In order to investigate the potential modulatory effects of microbiome-derived peptides resembling MOG35-55 peptide sequence on the autoimmune response that triggers EAE, a total of 8 peptides (Fig 3 and 4A) were selected from a pool of 439 ligands. These 8 peptides were confidently modeled in pMHC II-TCR interactions (Fig 3), and their selection was based on four specific criteria. Firstly, ligands that contained the “YRxxFxRVx” motif were considered. Secondly, ligands that could be detected in the NBDC database were included (specifically, P1, P2, P4, P5, P6, and P8). Thirdly, ligands collected from species related to MS were chosen (P3 and P5) (24). Finally, ligands containing noncanonical binding motifs starting with “F” and “L” were selected (P1, P3, P7, and P8), as these motifs were found to be most enriched in the pool of 493 binding core peptides. To investigate the potential of microbe-derived peptides to activate MOG35-55-specific T cell, we stimulated splenocytes isolated from MOG35-55-induced EAE mice with MOG35-55 peptide and 8 selected peptides (Fig 4A). Of the tested 8 microbiome-derived peptides, P3 demonstrated the capacity to stimulate the production of IL-17 and IFN-γ by MOG35-55-specific T cells (Fig 4B-D, supplementary Fig 3A). It is noteworthy that P3 was identified as originating from A. muciniphila, a gut commensal that is significantly enriched in the gut microbiota of individuals with MS (5).

Figure 4.
Figure 4. Microbiome-derived peptide can stimulate MOG35-55-specific T cells from EAE mice.

(A) The origin and sequence of microbial peptides tested in this study. The amino acids highlighted in green are the dominant MHC II-binding residue, while major TCR contact residues are colored in red. The amino acids colored in blue represent different residue between human MOG35-55 and mouse MOG35-55. The asterisk highlights that P5 is a peptide predicted in both Alistipes putredinis and A. muciniphila. (B, C and D) MOG35-55–specific CD4+ T cells collected from the spleen of EAE mice were treated with each of the selected 8 microbial peptides. PBS and MOG35-55 served as negative control and positive control; Irrelevant peptide was set as scrambled peptide control. Flow cytometry was used to determine IFN-γ and IL-17 production (n = 3). Representative FACS plots (gated on CD3+ and CD4+ cells) for the population of IFN-γ- or IL-17-secreted CD4+ T cells that were stimulated with mimic peptides or MOG35-55 (20 μg/ml final concentration). Only data from P3, which induced the highest cytokine-producing cell population was shown (B), along with the frequency (C) and cell counts (D) of each population among CD4+ T cells (Red triangle specifically point P3 treatment group). Data shown are the average ± SEM.

P3 exacerbate EAE development in mice and shows bidirectional cross reactivity with MOG35-55 peptide in vitro

To further explore the impact of microbiome-derived peptide on EAE development, C57BL/6J mice were immunized with a combination of P3 and a low dose of MOG35-55. We found that P3 could exacerbate EAE development in mice (Fig 5A). At 17 days post-immunization, the spinal cord tissues were collected, and flow cytometry as well as Hematoxylin and Eosin (H&E) or Luxol Fast Blue (LFB) staining were performed to determine the impact of P3 on the severity of inflammation and demyelination. As shown in Fig 5B, increased leukocyte infiltration was observed in the MOG35-55 + P3 group, implicating more severe immune infiltration in this group. Moreover, more severe inflammation was observed in the white matter of spinal cord from mice immunized with P3 and MOG35-55 compared to the control mice (Fig 5C). Quantitative analysis revealed that the inflammatory scores (H&E score) in the MOG35-55 + P3 group were significantly higher than the MOG35-55 + control group (Fig. 5C). Additionally, the demyelination level (LFB score) showed an increasing trend in mice immunized with MOG35-55 + P3 compared to the control mice (Fig. 5D).

Figure 5.
Figure 5. P3 immunization exacerbates EAE development in vivo.

(A) Mean EAE clinical scores and incidence rate from each group of mice. (B) Spinal cord samples were collected at day 17 post-immunization and the leukocyte population within the CNS was assessed by flow cytometry. The cell count of leukocytes were significantly higher in MOG35-55+P3 group than MOG35-55+control group (n = 8). (C, D) Spinal cord tissue sections were obtained at day 17 post-immunization and stained with H&E (C) or LFB (D). The inflammatory level (H&E scores) and demyelination level (LFB scores) were quantified (n=4-5). Data shown are the average ± SEM. *p<0.05.

Splenocytes from P3 immunized mice (10-day post-immunization) were collected to further test the cross-reactivity between P3 and MOG35-55. Following an 18-hour stimulation of P3-specific CD4+ T cells either with P3 or MOG35-55, the cells exhibited a robust immune response to P3 and a relatively strong response to MOG35-55 as indicated by the production of IFN-γ (Fig 6). These data suggest that the microbial peptide P3 and MOG35-55 can stimulate a bidirectional cross-reactive immune response.

Figure 6.
Figure 6. P3 shows bidirectional cross-reactivity with MOG35-55.

(A, B and C) Splenocytes from P3 immunized mice (n=5) were stimulated with either P3 (positive control), MOG35-55, or PBS (negative control) in a concentration of 40 μg/ml to detect the cross-reactivity between P3 and MOG35-55, and activation of CD4+ T cells was assessed by measuring the IL-17- and IFN-γ-producing population using flow cytometry. Representative FACS plots (gated on CD3+ and CD4+ cells) demonstrate the population of activated CD4+ T cells (A), along with the frequency (B) and cell counts (C) of IL-17+ and INF-γ+ population among CD4+ T cells. Data shown are the average ± SEM. *p<0.05, **p < 0.01, ***p < 0.001.

DCs process microbiome-derived peptide P3 and further activate 2D2 CD4+ T cell proliferation

Given that P3 stimulated MOG35-55-specific CD4+ T cells in vitro, we investigated whether DCs are capable of processing and presenting P3 to MOG35-55-specific T cells. Our pMHC II-TCR model predicted an interaction between the 11-residue windows with the core 9-mer residues truncated from microbiome-derived peptide and the MHC II-TCR complex (Fig 7A and 7B), leading us to hypothesize that this peptide could be presented by DCs. To test this hypothesis, we performed an antigen presentation assay using DCs isolated from the spleens of C57BL/6J mice. The DCs were primed with either PBS, P3, or MOG35-55, and then co-cultured with sorted 2D2 CD4+ T cells. The activation status of CD4+ T cells was later assessed by quantifying the expression of the early T cell activation marker CD69.

Figure 7.
Figure 7. P3 stimulates 2D2 CD4+ T cell proliferation through DCs presentation in vitro.

(A) Predicted structure of peptide-MHC II-TCR complex with binding core peptide from P3. TCR alleles same from complex template 3c60 (PDB ID). Surface representation is displayed in lighter shades. (B) Close-up view of peptide-TCR interactions (upper) and peptide-MHC II (bottom) in panel D, colored amino acid residues represent the dominant MHC II-binding residue (green) and major TCR recognition residue (red). Peptide residues and TCR alpha CDR3 (orange) and TCR beta CDR3 (hot pink) residues were shown in sticks, and residues from MHC II interacting with peptide were shown in slate blue. Hydrogen bonds are highlighted in black dashed lines. (C, D) The histogram FACS plots (C) and quantification (D) of CFSE labeled 2D2 CD4+ T cells proliferation, as well as the frequency of CD69+ cells within the CD4+ T cells population, following co-culture with DCs pulsed with PBS, P3, or MOG35-55. Dash line in histogram represents the proliferation level of PBS control. Data shown are the average ± SEM. **p < 0.01, ****p < 0.0001.

Furthermore, the proliferation of CD4+ T cells was evaluated by monitoring the changes in the population of CFSE-labeled cells. The results demonstrated that P3 could be presented by DCs and subsequently activate the proliferation of 2D2 CD4+ T cells (Fig 7C). However, P3 induces a weaker CD4+ T cell activation and proliferation than MOG35-55 (Fig 7D), which may explain the lower incidence and clinical score observed in P3-induced EAE (Supplementary Fig 4).

Discussion

The gut microbiota has been implicated in the development of autoimmune diseases. Investigations of MS patients and the EAE mouse model have revealed that the gut microbiome is a crucial factor in disease progression and severity (4, 5). Alterations in gut bacterial composition, including reductions in butyrate-producing bacteria and increases in Akkermansia and Clostridia populations have been consistently observed in MS patients (5, 7). Moreover, transfer of microbiota from MS patients to animal models has been shown to increase disease incidence, suggesting that alterations in the microbiome may actively contribute to the development of MS (32). Germ-free and antibiotic-treated mice have shown resistant to EAE, further supporting the necessity of the microbiome for disease induction (33, 34).

Despite significant advances in understanding the role of the gut microbiota in MS pathogenesis, the underlying mechanisms leading to disease development remain poorly understood. Continuous sampling of gut microbial antigens by innate immune cells enhances the probability of presenting specific antigens to antigen-specific T cells. The presentation of these antigens is influenced by polymorphisms in MHC genes, suggesting a potential role for host-commensal microorganisms cross-reactivity in the development and persistence of autoimmunity among genetically predisposed individuals. Molecular mimicry has been proposed as a possible mechanism by which microorganisms can induce autoimmune diseases (15). According to this hypothesis, T cells that respond to an epitope derived from an infectious agent may cross-react with a self-antigen that share sequence or structural similarities with the original microbial peptide (35). Evidence supporting this hypothesis has been observed in numerous autoimmune diseases, such as antiphospholipid syndrome and Type 1 Diabetes (36, 37). Given the diverse range of microorganisms in the gut, peptides derived from gut flora could potentially trigger the development of MS. The activation of autoreactive T cells is a critical component of the early autoimmune response in individuals who ultimately develop MS. While multiple potential antigens exist, MOG is a major EAE antigen in H-2b mice and a possible autoantigen in MS (20). The binding pattern of the MOG-MHC II-TCR complex has been characterized, revealing that tyrosine (Y) in position 40 of the MOG35-55 peptide is the dominant MHC II-binding residue, occupying the p1 pocket. Meanwhile, arginine (R) 41, phenylalanine (F) 44, R46, and valine (V) 47 are the major amino acids involved in TCR recognition (20). Research suggests that low-affinity hotspot mimicry, which involves a subset of crucial residues within the TCR-binding footprint, rather than high-affinity structural mimicry, may be a more common contributor to the initiation of autoimmune diseases (14, 20). In our study, we focused on identifying MOG35-55 mimics based on the similarity of essential residues within the TCR-binding footprint, rather than relying on the direct amino acid similarity between microbial-derived peptides and self-antigens, as reported in studies on Type 1 Diabetes (37).

By leveraging the expanding genome databases for human gut microbes, we identified 8 microbial peptides that exhibit significant analogy in their antigen-MHC II-TCR binding patterns with MOG35-55 peptide. Our findings indicated that microbiome derived P3 can activate CD4+ T cells specific to MOG35-55 peptide in mice. Interestingly, our bioinformatic analysis revealed that this peptide was derived from A. muciniphila, suggesting a potential impact of A. muciniphila on EAE development. Previous studies have explored the link between A. muciniphila and EAE or MS, such as the findings of Cekanaviciute et al., who observed that A. muciniphila promotes Th1 differentiation in human peripheral blood mononuclear cells (5). Moreover, previous studies have demonstrated that a peptide from A. muciniphila shares a similar sequence with a MS associated self-antigen, guanosine diphosphate (GDP)–L-fucose synthase. This shared peptide sequence has the ability to stimulate the proliferation of CD4+ cerebrospinal fluid-infiltrating T cells from MS patients (16).

In this study, we optimized the process of screening self-antigen mimics from gut microbiota by integrating the binding affinity to MHC II and TCR into the selection criteria, thereby enhancing the accuracy of identifying self-antigen analogs in the gut microbiome. Furthermore, our findings suggest that the enrichment of A. muciniphila in the gut microbiota of EAE mice and MS patients may have pathological implications as this bacterium has the capability of generating MOG35-55 analogs. However, our method utilizing metagenomic data can only annotate genome to species level, and further determination is needed to identify the specific A. muciniphila strain capable of producing MOG35-55 analogs.

A. muciniphila, a next-generation probiotic, has shown positive effects on several diseases including obesity and diabetes (38). However, recent studies have revealed an enrichment of A. muciniphila in the gut microbiome of MS patients. Furthermore, transplantation of fecal samples from MS patients into the intestines of EAE mice leads to more severe EAE symptoms (5), indicating the need to assess the potential beneficial or harmful effects of A. muciniphila from multiple perspectives. Although multiple strains of A. muciniphila have been identified (39), it remains unclear whether the observed effects in different disease contexts are attributable to strain differences. The predominant use of metagenomics and 16S sequencing in numerous studies exploring the relationship between various diseases and A. muciniphila has impeded the identification of specific target strains, as these methods primarily facilitate gene annotation at the species level, presenting challenges in strain-level identification. Our study suggests that A. muciniphila may contribute to the development of EAE due to its capacity to generate MOG35-55-resembling peptides. However, more evidence are needed to determine the target strain that generates MOG35-55 analogs. Future studies will also focus on elucidating the role of synergistic or antagonistic effect among different microbial populations in autoimmune diseases. This study highlights the necessity for further research on the correlation between gut microbiota and autoimmune diseases in the growing trend of widespread probiotic consumption.

In summary, our findings prove a mechanism of molecular mimicry in which a particular sequence within commensal gut microbe can imitate crucial residues in the MOG-MHC II-TCR binding footprint, consequently initiating or modifying the immune response associated with EAE development. While our study focused only on MOG35-55 mimics, other myelin antigens, such as MBP and PLP, may also serve as targets for autoreactive CD4+ T cells (40). Given the enormous number of microbial peptides generated by the gut microbiota, it is possible that other microbial peptides may exist and can mimic these antigens and activate related autoantigen reactive T cells. Meanwhile, it’s important to note that the induction of autoimmune diseases through molecular mimicry is complicated and a single factor acting alone is unlikely to trigger the disease. Instead, accompanying host genetic susceptibility is considered to be linked with an abnormal immune response triggered by one or more microbial antigens or molecules. Nonetheless, this discovery may provide a therapeutic target and an opportunity to impede the progression of MS.

Conclusion

We have investigated the role of molecular mimicry in the link between microbial flora and EAE development. We have identified MOG35-55 mimics derived from gut commensal cross-react with MOG35-55-specific CD4+ T cells and significantly exacerbate the development of EAE in mice. Notably, antigen-presenting assays confirmed that DCs can process and present this gut commensal derived peptide to MOG35-55-specific CD4+ T cells. Our findings offer direct evidence of how microbes can initiate the development of EAE, suggesting a potential microbiome-based therapeutic target for inhibiting the progression of MS.

Material and Methods

Identify the human MOG35-55 peptide analog in the human microbiome genomes and human-associated metagenomic samples

In brief, we first collected and curated human microbiome-derived peptide resources, which included human microbiome genomes and human-associated metagenomics samples. In detail, we collected 286,997 human microbiome genomes from Unified Human Gastrointestinal Genome dataset (UHGG, v1.0) (41), and 17249 genomes from CIBIO species-level genome bins (21). Taxonomic assignments were performed for genomes from CIBIO by GTDB-tk (42). In addition, we downloaded 103 metagenomes from MS patients in National Bioscience Database Center (NBDC) Human Database with the accession number of hum0197 (43). The data obtained from the UHGG and NBDC were acquired in accordance with human research ethics guidelines and were approved by the Human Ethics sub-Committee of the City University of Hong Kong (jcc2223ay001). For these metagenomic samples, we applied metagenomic assembly to obtain contigs using MEGAHIT (44) for each sample separately. Taxonomic classification was further conducted for the resulting contigs using CAT (v5.2.3) (45). Taxonomic assignments for genomes or contigs based on the Genome Taxonomy Database (46). Finally, all open reading frames were predicted in each genome or contig using Prodigal-short (47). Human MOG35-55 sequence analog searching was conducted by BLASTp with an e-value threshold of 0.001 in a pool of proteins from the above Human microbiome-derived peptide resources.

The structure-based prediction of peptide-MHC II-TCR binding

Identified analogs were first filtered based on whether they contained the motifs with [YR..F.RV.], [YR..F.R..], [YR..F..V.], [YR….RV], and [Y…F.RV.] are involved in the peptide-MHC-TCR interactions. Next, a special peptide region (∼ 21 residues) containing the above binding pattern was truncated from these filtered analogs. The yielded nonredundant truncated peptides were subjected to further analyses.

The peptide-MHC II-TCR binding prediction was performed in two steps: the peptide-MHC II binding specificity prediction and TCR: peptide-MHC II interactions modeling. Because the advanced peptide-MHC II structure prediction tools often predict the binding structure composed by the binding core of peptides (normally with 9 residues) and receptors (MHC II or TCR) (27–30, 48), these nonredundant truncated peptides were further chopped into a series of consecutive 11-residue windows along peptide with the putative core nine-mer residues and their immediate neighboring residues. The resulting 11-residue windows were subjected to a fine-tuned AlphaFold to predict peptide-MHC II binding with the following parameters: -- model_names model_2_ptm_ft --ignore_identities (27). Peptide-MHC II templates (4p23_MH2_H2ABa_H2ABb.pdb,6mng_MH2_H2ABa_H2ABb.pdb,1muj_MH2_H2A Ba_H2ABb.pdb) in the PDB template set were used (27). Mean Inter-chain lowest predicted aligned error (PAE) terms corresponding to the peptide-MHC interactions < 4.34 and peptides’ per-residue confidence score (pLDDT) corresponding to local structural accuracy > 90 were used to classify MHC II binder and non-binder. Next, peptide-MHC II-TCR binding for the predicted binders was modeled by a specialized version of the neural network predictor AlphaFold and TCR–pMHC complex templates-based method, TCRDock with ‘model_2_ptm’ and ‘--benchmark’ parameter set (28). Based on TCRDock’s workflow, three simulations were conducted for each target peptide-MHC II-TCR complex, and finally chose the model with the lowest PAE between the TCR and pMHC. Given the hypervariable gene segments in TCR alpha and beta chain largely determine the structure modeling (49), we took TCR:pMHC II structural templates (PDB IDs: 3c60, 3rdt, 3c5z, 6mnn), pairing with four types of mouse TCR (Supplementary table 5), to model the peptide-MHC II-TCR structures. The predicted structures were visualized by PyMoL.

Mice

Eight-week-old C57BL/6J mice were obtained from the Laboratory Animal Research Unit (LARU) of City University of Hong Kong. The 2D2 TCR transgenic mice C57BL/6-Tg (Tcra2D2, Tcrb2D2) 1Kuch/J were purchased from The Jackson Laboratory and bred in LARU. Mice were randomly allocated to each treatment group. All experimental procedures were approved by the Animal Ethics Committee of the City University of Hong Kong and performed in accordance with LARU Guidelines.

EAE induction

For induction of active EAE, mice (8-week-old) were immunized subcutaneously at two sites on the back flank area with 200 μg of MOG35–55 (MEVGWYRSPFSRVVHLYRNGK) peptide (Thermo Fisher) emulsified in 200 μl incomplete Freund’s adjuvant (BD bioscience) supplemented with 2 mg/ml heat-killed Mycobacterium tuberculosis (BD bioscience). On days 0 and 2 after immunization, mice received intraperitoneal 400 ng of pertussis toxin (Calbiochem). The same immunization procedure was employed to detect the encephalitogenic capacity of P3, using a combination of P3 and MOG35–55 (P3: 200 μg; MOG35–55: 50μg;). The other group was administered same dose of a combination of irrelevant peptide (set as a control of P3) and MOG35–55. EAE severity was scored daily using the standard scale: 0, no disease; 1, loss of tail tone; 2, hind limb weakness; 3, hind limb paralysis; 4, hind limb paralysis and forelimb paralysis or weakness; and 5, moribund/death (50). To detect the cross reactivity between MOG35-55 and synthesized peptides, mice were injected with peptide by following the similar immunization procedure (DG peptides, China). The same immunization protocol was employed to induce EAE using P3.

Antigen recall assay

T cell antigen recall response was assessed in cells harvested from the spleen of EAE mice 10 days after the challenge. Mice were anesthetized by 3% isoflurane inhalation and then euthanized by cervical dislocation. Spleens were then collected from the euthanized mice and crushed through a 70-μm cell strainer and centrifuged at 400 × g for 5 min at 4 °C, then incubated in red blood cell lysis buffer (Thermo Fisher) for 5 min. After centrifuge, isolated cells were plated at a density of 1x106 cells/well and stimulated with synthesized peptide or MOG35–55 (20 μg/ml final concentrations) in cell culture media (RPMI1640, 10% FBS, 2 mM L-glutamine, 10 mM Hepes, 50 μM 2-mercaptoethanol, 100 IU penicillin, and 100 μg/mL streptomycin), in a total volume of 200 μL per well. Cells were stimulated for 18h at 37 °C and 5% CO2 atmosphere, then analyzed by flow cytometry for cytokine (IFN-γ, IL-17A) producing cells.

Histopathology

Inflammation and demyelination of spinal cords was assessed at day 17 after immunization. The lumbar enlargement of the spinal cords was isolated from animals and instantly fixed in 4% paraformaldehyde for a minimum of 48 h, then embedded in paraffin and sectioned at 5-μm thickness. Sections were stained with H&E (Solarbio) or LFB (Abcam) to evaluate inflammation and demyelination. The sections were scanned using a NanoZoomer S60 Digital slide scanner (Hamamatsu). The degrees of inflammation (H&E scores) were calculated as 0, normal; 1, cellular infiltrates partially around meninges and blood vessels; 2, 1–10 lymphocyte infiltrates in the parenchyma; 3, 11–100 lymphocyte infiltrates in the spinal cord; 4, 100-300 lymphocyte infiltrates in the spinal cord; 5, 300-1000 lymphocyte infiltrates in the spinal cord; and 6, over 1000 lymphocyte infiltrates in the spinal cord (51). The severity of demyelination (LFB scores) was analyzed as 0, normal; 1, myelin sheath injuries in rare areas; 2, a few areas of demyelination; 3, confluent perivascular or subpial demyelination; 4, massive demyelination involving one half of the spinal cord; and 5, extensive demyelination involving the whole spinal cord (51).

Antigen presentation assays

To obtain stimulator DCs, single cell suspensions from the spleen collected from C57BL/6J mice were prepared by mechanical disruption on a 70-μm cell strainer. Cell suspensions were resuspended in MACS buffer (0.5% BSA, 2 mM EDTA in PBS) after lysing red blood cells. CD11c+ cells were then negatively selected according to the manufacturer’s instructions using the mouse Pan Dendritic Cell Isolation Kit (Miltenyitec). Isolated DCs were pulsed with either PBS, P3, or MOG35-55 at 20 μg/ml final concentrations, using 2 x 104 cells in 96-well plate for 1 h at 37 °C. Cells were then washed extensively and used for co-culture with responder 2D2 CD4+ T cells. Responder MOG35-55-specific CD4+ T cells were isolated from 2D2 TCR transgenic mice spleen by using CD4+ T cells Isolation Kit (Miltenyitec). Sorted CD4+ T cells were labeled with Carboxyfluorescein succinimidyl ester (CFSE, Thermo Fisher) following the user guide. 105 CFSE labeled 2D2 CD4+ T cells were then seeded in triplicate in a round bottom 96-well plate with 2 x 104 stimulated DCs (5:1 ratio). DCs and T cells were cultured together for 72□h at 37 °C and 5% CO2 atmosphere, then cells were prepared for flow cytometry to detect T cell proliferation.

Flow cytometry

Single cell suspensions from the spleen were obtained by mechanical disruption on a 70-μm cell strainer, then incubated in red blood cell lysis buffer for 5 min. For the staining process, dead cells were stained with Propidium Iodide (Sigma-Aldrich) or Live/Dead Ghost dye (Tonbo Biosciences). For staining of surface markers, cells were incubated in fluorescently labeled antibodies (CD45, CD3, CD4, CD69, TCR Vβ11) for 30 min in FACS buffer (PBS containing 2% FBS) at 4 °C. For intracellular staining, cells were washed with 2 mL wash buffer, then fixed and permeabilized by using Cytofix/Cytoperm™ Fixation/Permeabilization Kit (BD Biosciences) as manufacturer’s instructions. Fixed cells were incubated in fluorescently labeled antibodies (IFN-γ, IL-17A) for 30 min in permeabilization buffer at 4 °C. All data were collected on a BD Celesta flow cytometer (BD Biosciences) and analyzed using FlowJo software (v.10.2, BD Biosciences).

Statistical analysis

Data were analyzed using GraphPad Prism software and reported as mean ± SEM unless otherwise noted. The two-tailed unpaired Student t test was used for samples with equal variances, whereas Welch’s t test was employed for samples with unequal variances. Statistical significance was defined as a p-value <0.05.

Declarations

Ethics approval and consent to participate

The data obtained from the UHGG and NBDC were acquired in accordance with human research ethics guidelines and were approved by the Human Ethics sub-Committee of the City University of Hong Kong (jcc2223ay001). And all experimental procedures were approved by the Animal Ethics Committee of the City University of Hong Kong and performed in accordance with LARU Guidelines.

Consent for publication

Not applicable.

Availability of data and materials

All study data are included in the article and/or Supplemantary files.

Competing interests

The authors declare no competing interest.

Funding

This work is supported by the National Natural Science Foundation of China (82371350).

Authors’ contributions

Conceptualization: GY, YXL; Methodology: XM, JZ, QLJ; Investigation: XM, JZ; Visualization: XM, JZ; Supervision: GY, YXL; Writing: XM, JZ, YXL, GY.

Acknowledgements

Not applicable.

Abbreviations

MS
Multiple sclerosis
MOG
Myelin oligodendrocyte glycoprotein
MOG35-55
Myelin oligodendrocyte glycoprotein peptide 35-55
EAE
Experimental autoimmune encephalomyelitis
DCs
Dendritic cells
A. muciniphila
Akkermansia muciniphila
MBP
Myelin basic protein
PLP
Proteolipid protein
UHGG
Unified Human Gastrointestinal Genome dataset
NBDC
National Bioscience Database Center
PAE
Predicted aligned error
pLDDT
Peptides’ per-residue confidence score
H&E
Hematoxylin and Eosin
LFB
Luxol Fast Blue

Reference

  1. 1.↵
    Sospedra M, Martin R. Immunology of multiple sclerosis. Annu Rev Immunol. 2005;23:683–747.
    OpenUrlCrossRefPubMedWeb of Science
  2. 2.↵
    Walton C, King R, Rechtman L, Kaye W, Leray E, Marrie RA, et al. Rising prevalence of multiple sclerosis worldwide: Insights from the Atlas of MS. Multiple Sclerosis Journal. 2020;26(14):1816–21.
    OpenUrl
  3. 3.↵
    Alroughani R, Boyko A. Pediatric multiple sclerosis: a review. BMC neurology. 2018;18:1–8.
    OpenUrl
  4. 4.↵
    Lee YK, Menezes JS, Umesaki Y, Mazmanian SK. Proinflammatory T-cell responses to gut microbiota promote experimental autoimmune encephalomyelitis. Proceedings of the National Academy of Sciences. 2011;108(supplement_1):4615–22.
    OpenUrlAbstract/FREE Full Text
  5. 5.↵
    Cekanaviciute E, Yoo BB, Runia TF, Debelius JW, Singh S, Nelson CA, et al. Gut bacteria from multiple sclerosis patients modulate human T cells and exacerbate symptoms in mouse models. Proceedings of the National Academy of Sciences. 2017;114(40):10713–8.
    OpenUrlAbstract/FREE Full Text
  6. 6.↵
    Chen J, Chia N, Kalari KR, Yao JZ, Novotna M, Paz Soldan MM, et al. Multiple sclerosis patients have a distinct gut microbiota compared to healthy controls. Scientific reports. 2016;6(1):28484.
    OpenUrl
  7. 7.↵
    Jangi S, Gandhi R, Cox LM, Li N, Von Glehn F, Yan R, et al. Alterations of the human gut microbiome in multiple sclerosis. Nature communications. 2016;7(1):12015.
    OpenUrl
  8. 8.↵
    Johanson DM, Goertz JE, Marin IA, Costello J, Overall CC, Gaultier A. Experimental autoimmune encephalomyelitis is associated with changes of the microbiota composition in the gastrointestinal tract. Scientific reports. 2020;10(1):15183.
    OpenUrl
  9. 9.↵
    Rojas M, Restrepo-Jiménez P, Monsalve DM, Pacheco Y, Acosta-Ampudia Y, Ramírez-Santana C, et al. Molecular mimicry and autoimmunity. Journal of autoimmunity. 2018;95:100–23.
    OpenUrlCrossRefPubMed
  10. 10.
    Greiling TM, Dehner C, Chen X, Hughes K, Iñiguez AJ, Boccitto M, et al. Commensal orthologs of the human autoantigen Ro60 as triggers of autoimmunity in lupus. Science Translational Medicine. 2018;10(434):eaan2306.
    OpenUrlAbstract/FREE Full Text
  11. 11.
    Naghavian R, Faigle W, Oldrati P, Wang J, Toussaint NC, Qiu Y, et al. Microbial peptides activate tumour-infiltrating lymphocytes in glioblastoma. Nature. 2023:1–11.
  12. 12.
    Kalaora S, Nagler A, Nejman D, Alon M, Barbolin C, Barnea E, et al. Identification of bacteria-derived HLA-bound peptides in melanoma. Nature. 2021;592(7852):138–43.
    OpenUrlPubMed
  13. 13.↵
    Gil-Cruz C, Perez-Shibayama C, De Martin A, Ronchi F, Van Der Borght K, Niederer R, et al. Microbiota-derived peptide mimics drive lethal inflammatory cardiomyopathy. Science. 2019;366(6467):881–6.
    OpenUrlAbstract/FREE Full Text
  14. 14.↵
    Harkiolaki M, Holmes SL, Svendsen P, Gregersen JW, Jensen LT, McMahon R, et al. T cell-mediated autoimmune disease due to low-affinity crossreactivity to common microbial peptides. Immunity. 2009;30(3):348–57.
    OpenUrlCrossRefPubMedWeb of Science
  15. 15.↵
    Wucherpfennig KW, Strominger JL. Molecular mimicry in T cell-mediated autoimmunity: viral peptides activate human T cell clones specific for myelin basic protein. Cell. 1995;80(5):695–705.
    OpenUrlCrossRefPubMedWeb of Science
  16. 16.↵
    Planas R, Santos R, Tomas-Ojer P, Cruciani C, Lutterotti A, Faigle W, et al. GDP-l-fucose synthase is a CD4+ T cell–specific autoantigen in DRB3* 02: 02 patients with multiple sclerosis. Science translational medicine. 2018;10(462):eaat4301.
    OpenUrlAbstract/FREE Full Text
  17. 17.↵
    Lerch M, Bauer A, Reindl M. The Potential pathogenicity of myelin oligodendrocyte glycoprotein antibodies in the optic pathway. Journal of Neuro-Ophthalmology. 2023;43(1):5.
    OpenUrl
  18. 18.
    Marignier R, Hacohen Y, Cobo-Calvo A, Pröbstel A-K, Aktas O, Alexopoulos H, et al. Myelin-oligodendrocyte glycoprotein antibody-associated disease. The Lancet Neurology. 2021;20(9):762–72.
    OpenUrl
  19. 19.↵
    Reindl M, Waters P. Myelin oligodendrocyte glycoprotein antibodies in neurological disease. Nature Reviews Neurology. 2019;15(2):89–102.
    OpenUrl
  20. 20.↵
    Petersen TR, Bettelli E, Sidney J, Sette A, Kuchroo V, Backstrom BT. Characterization of MHC- and TCR-binding residues of the myelin oligodendrocyte glycoprotein 38-51 peptide. Eur J Immunol. 2004;34(1):165–73.
    OpenUrlCrossRefPubMed
  21. 21.↵
    Pasolli E, Asnicar F, Manara S, Zolfo M, Karcher N, Armanini F, et al. Extensive Unexplored Human Microbiome Diversity Revealed by Over 150,000 Genomes from Metagenomes Spanning Age, Geography, and Lifestyle. Cell. 2019;176(3):649–62 e20.
    OpenUrlCrossRefPubMed
  22. 22.↵
    Almeida A, Nayfach S, Boland M, Strozzi F, Beracochea M, Shi ZJ, et al. A unified catalog of 204,938 reference genomes from the human gut microbiome. Nature biotechnology. 2021;39(1):105–14.
    OpenUrlPubMed
  23. 23.↵
    Kishikawa T, Ogawa K, Motooka D, Hosokawa A, Kinoshita M, Suzuki K, et al. A metagenome-wide association study of gut microbiome in patients with multiple sclerosis revealed novel disease pathology. Frontiers in cellular and infection microbiology. 2020;10:585973.
    OpenUrl
  24. 24.↵
    Zhou X, Baumann R, Gao X, Mendoza M, Singh S, Sand IK, et al. Gut microbiome of multiple sclerosis patients and paired household healthy controls reveal associations with disease risk and course. Cell. 2022;185(19):3467–86. e16.
    OpenUrlCrossRef
  25. 25.↵
    Ishina IA, Zakharova MY, Kurbatskaia IN, Mamedov AE, Belogurov AA, Jr., Gabibov AG. MHC Class II Presentation in Autoimmunity. Cells. 2023;12(2).
  26. 26.↵
    Pishesha N, Harmand TJ, Ploegh HL. A guide to antigen processing and presentation. Nat Rev Immunol. 2022;22(12):751–64.
    OpenUrl
  27. 27.↵
    Motmaen A, Dauparas J, Baek M, Abedi MH, Baker D, Bradley P. Peptide-binding specificity prediction using fine-tuned protein structure prediction networks. Proc Natl Acad Sci U S A. 2023;120(9):e2216697120.
    OpenUrl
  28. 28.↵
    Bradley P. Structure-based prediction of T cell receptor:peptide-MHC interactions. Elife. 2023;12.
  29. 29.
    Marzella DF, Parizi FM, van Tilborg D, Renaud N, Sybrandi D, Buzatu R, et al. PANDORA: A Fast, Anchor-Restrained Modelling Protocol for Peptide: MHC Complexes. Front Immunol. 2022;13:878762.
    OpenUrl
  30. 30.↵
    Reynisson B, Barra C, Kaabinejadian S, Hildebrand WH, Peters B, Nielsen M. Improved Prediction of MHC II Antigen Presentation through Integration and Motif Deconvolution of Mass Spectrometry MHC Eluted Ligand Data. J Proteome Res. 2020;19(6):2304–15.
    OpenUrlCrossRef
  31. 31.↵
    Lourbopoulos A, Deraos G, Matsoukas MT, Touloumi O, Giannakopoulou A, Kalbacher H, et al. Cyclic MOG(35)(-)(55) ameliorates clinical and neuropathological features of experimental autoimmune encephalomyelitis. Bioorg Med Chem. 2017;25(15):4163–74.
    OpenUrl
  32. 32.↵
    Berer K, Gerdes LA, Cekanaviciute E, Jia X, Xiao L, Xia Z, et al. Gut microbiota from multiple sclerosis patients enables spontaneous autoimmune encephalomyelitis in mice. Proceedings of the National Academy of Sciences. 2017;114(40):10719–24.
    OpenUrlAbstract/FREE Full Text
  33. 33.↵
    Berer K, Mues M, Koutrolos M, Rasbi ZA, Boziki M, Johner C, et al. Commensal microbiota and myelin autoantigen cooperate to trigger autoimmune demyelination. Nature. 2011;479(7374):538–41.
    OpenUrlCrossRefPubMedWeb of Science
  34. 34.↵
    Ochoa-Repáraz J, Mielcarz DW, Ditrio LE, Burroughs AR, Foureau DM, Haque-Begum S, et al. Role of gut commensal microflora in the development of experimental autoimmune encephalomyelitis. The Journal of Immunology. 2009;183(10):6041–50.
    OpenUrl
  35. 35.↵
    Ochoa-Repáraz J, Kirby TO, Kasper LH. The gut microbiome and multiple sclerosis. Cold Spring Harbor perspectives in medicine. 2018;8(6).
  36. 36.↵
    Ruff WE, Dehner C, Kim WJ, Pagovich O, Aguiar CL, Andrew TY, et al. Pathogenic autoreactive T and B cells cross-react with mimotopes expressed by a common human gut commensal to trigger autoimmunity. Cell host & microbe. 2019;26(1):100–13. e8.
    OpenUrlCrossRefPubMed
  37. 37.↵
    Girdhar K, Huang Q, Chow I-T, Vatanen T, Brady C, Raisingani A, et al. A gut microbial peptide and molecular mimicry in the pathogenesis of type 1 diabetes. Proceedings of the National Academy of Sciences. 2022;119(31):e2120028119.
    OpenUrlCrossRef
  38. 38.↵
    Cani PD, Depommier C, Derrien M, Everard A, de Vos WM. Akkermansia muciniphila: paradigm for next-generation beneficial microorganisms. Nature Reviews Gastroenterology & Hepatology. 2022;19(10):625–37.
    OpenUrl
  39. 39.↵
    Guo X, Zhang J, Wu F, Zhang M, Yi M, Peng Y. Different subtype strains of Akkermansia muciniphila abundantly colonize in southern China. Journal of applied microbiology. 2016;120(2):452–9.
    OpenUrlCrossRef
  40. 40.↵
    Greer JM. Autoimmune T-cell reactivity to myelin proteolipids and glycolipids in multiple sclerosis. Multiple sclerosis international. 2013;2013.
  41. 41.↵
    Almeida A, Nayfach S, Boland M, Strozzi F, Beracochea M, Shi ZJ, et al. A unified catalog of 204,938 reference genomes from the human gut microbiome. Nat Biotechnol. 2021;39(1):105–14.
    OpenUrlPubMed
  42. 42.↵
    Chaumeil PA, Mussig AJ, Hugenholtz P, Parks DH. GTDB-Tk v2: memory friendly classification with the genome taxonomy database. Bioinformatics. 2022;38(23):5315–6.
    OpenUrl
  43. 43.↵
    Kishikawa T, Ogawa K, Motooka D, Hosokawa A, Kinoshita M, Suzuki K, et al. A Metagenome-Wide Association Study of Gut Microbiome in Patients With Multiple Sclerosis Revealed Novel Disease Pathology. Front Cell Infect Microbiol. 2020;10:585973.
    OpenUrl
  44. 44.↵
    Li D, Liu CM, Luo R, Sadakane K, Lam TW. MEGAHIT: an ultra-fast single-node solution for large and complex metagenomics assembly via succinct de Bruijn graph. Bioinformatics. 2015;31(10):1674–6.
    OpenUrlCrossRefPubMed
  45. 45.↵
    von Meijenfeldt FAB, Arkhipova K, Cambuy DD, Coutinho FH, Dutilh BE. Robust taxonomic classification of uncharted microbial sequences and bins with CAT and BAT. Genome Biol. 2019;20(1):217.
    OpenUrlCrossRefPubMed
  46. 46.↵
    Parks DH, Chuvochina M, Rinke C, Mussig AJ, Chaumeil PA, Hugenholtz P. GTDB: an ongoing census of bacterial and archaeal diversity through a phylogenetically consistent, rank normalized and complete genome-based taxonomy. Nucleic Acids Res. 2022;50(D1):D785–D94.
    OpenUrl
  47. 47.↵
    Santos-Aberturas J, Chandra G, Frattaruolo L, Lacret R, Pham TH, Vior NM, et al. Uncovering the unexplored diversity of thioamidated ribosomal peptides in Actinobacteria using the RiPPER genome mining tool. Nucleic Acids Res. 2019;47(9):4624–37.
    OpenUrl
  48. 48.↵
    Mikhaylov V, Levine AJ. Accurate modeling of peptideMHC structures with AlphaFold. 2023.
  49. 49.↵
    La Gruta NL, Gras S, Daley SR, Thomas PG, Rossjohn J. Understanding the drivers of MHC restriction of T cell receptors. Nat Rev Immunol. 2018;18(7):467–78.
    OpenUrlCrossRefPubMed
  50. 50.↵
    Shahi SK, Freedman SN, Dahl RA, Karandikar NJ, Mangalam AK. Scoring disease in an animal model of multiple sclerosis using a novel infrared-based automated activity-monitoring system. Scientific reports. 2019;9(1):19194.
    OpenUrl
  51. 51.↵
    Quan M-Y, Song X-J, Liu H-J, Deng X-H, Hou H-Q, Chen L-P, et al. Amlexanox attenuates experimental autoimmune encephalomyelitis by inhibiting dendritic cell maturation and reprogramming effector and regulatory T cell responses. Journal of neuroinflammation. 2019;16:1–17.
    OpenUrl
Back to top
Posted July 10, 2024.
Download PDF

Supplementary Material

Email
Human microbiome-derived peptide affects the development of experimental autoimmune encephalomyelitis via molecular mimicry
Xin Ma, Jian Zhang, Qianling Jiang, Yong-Xin Li, Guan Yang
bioRxiv 2024.07.05.602171; doi: https://doi.org/10.1101/2024.07.05.602171
Twitter logo Facebook logo LinkedIn logo Mendeley logo
Citation Tools

Subject Area