Neuromyelitis optica spectrum disorder (NMOSD) is a severe inflammatory demyelinating disease of the central nervous system, characterized by recurrent attacks of transverse myelitis and optic neuritis (1). Although the etiology of NMOSD remains elusive, aquaporin-4 (AQP4) has been identified as a disease-specific serum autoantibody marker (2). Growing evidence indicates that environmental factors play a significant role in the pathogenesis and progression of inflammatory and autoimmune diseases.

The gut microbiota, often called the second brain, may influence brain activity through the gut-microbiota-brain axis under both physiological and pathological conditions. The gut microbiome contributes substantially to these disorders by affecting the immune system and metabolic pathways (3). Dysbiosis of gastrointestinal microbiota affects the differentiation of proinflammatory T-cells and may enhance organ-specific autoimmunity (4). Extensive homology between gut microbiota and AQP4 protein demonstrates molecular mimicry in the pathogenesis of NMOSD (5).Compared with healthy controls (HCs), antibody responses against Helicobacter pylori and gastrointestinal antigens were more frequently observed in NMOSD patients, suggesting that alterations in the gastrointestinal environment may contribute to the initiation or progression of NMOSD (6).Our previous research found that NMOSD patients had increased abundances of the pathogenic genera Streptococcus and Flavonifractor compared with HCs (7).

Previous studies have primarily examined group differences at a cross-sectional level, which may not fully account for treatment-related effects. To address this limitation, we conducted a case-control study comparing gut microbiome and metabolomic profiles between treatment-naïve NMOSD patients and HCs. Additionally, we performed longitudinal sampling (before and six months after immunosuppressant therapy) to evaluate how these medications alter the gut microbiota in NMOSD patients.

Our analysis included twenty-one treatment-naïve NMOSD patients (NT group) and twenty-one matched healthy controls (HCs). Fecal samples were collected from all participants at enrollment. Of these 21 NMOSD patients, 5 subsequently underwent re-examination of fecal samples six months after immunosuppressant treatment (IST group). Table 1 summarizes the detailed demographic information and clinical characteristics of the participants. The clinical manifestations and severity of each NMOSD patient are shown in Supplementary Table S1.

In the present study, we identified a striking gut microbiota imbalance in treatment-naïve NMOSD patients compared to HCs, along with dysregulated microbiota-associated metabolic pathways. Community analysis and β-diversity revealed significant differences in microbial composition between the two groups. The α-diversity of fecal microbiota in NMOSD patients was significantly higher than that in HCs, suggesting a potential link between gut microbiota and NMOSD pathogenesis. Our findings align with prior Chinese cohort studies (18, 19) reporting altered fecal microbiome α-diversity in NMOSD patients versus HCs. Although higher microbiota diversity is generally beneficial, elevated diversity and richness have also been observed in diseases like depression (20) and autism (21).

In our study, an overgrowth of opportunistic pathogens (e.g., Streptococcus and Ruminococcus) characterizes gut dysbiosis in NMOSD patients. Prior studies in Chinese NMOSD cohorts also reported associations with Streptococcus (7, 18, 19). Notably, Gong et al. found Streptococcus abundance positively correlated with disease severity (18). The immunomodulatory mechanism of Streptococcus remains unclear; however, its overrepresentation correlates with reduced short-chain fatty acid (SCFA) levels in NMOSD patients, which may enhance CD4+ T cell activity and pro-inflammatory responses (18, 22). In addition, increased Streptococcus abundance may promote inflammation by elevating pro-inflammatory cytokines (TNF-α, IL-6, and IFN-γ), which can also induce Th1/Th17 cell differentiation in humans (20, 23). Furthermore, Streptococcus species such as S. pyogenes were shown to reduce Treg frequency in tumor microenvironments and impair Treg function via APC-derived IL-12 (24). These findings suggest Streptococcus may play a pivotal role in NMOSD pathogenesis. This contrasts with prior reports implicating Clostridium perfringens through molecular mimicry (25). We propose that genetic and environmental differences between Caucasian and Asian populations may underlie these discrepant findings.

We observed increased Ruminococcus abundance in the gut microbiota of NMOSD patients, though its mechanistic role requires further investigation. Xie et al. similarly reported elevated Ruminococcaceae levels during NMOSD acute phases (19). Ruminococcus is an anaerobic bacterial genus with context-dependent roles in disease. While some species promote health through SCFA production, others may exacerbate disease when overabundant (26). This genus is enriched in multiple disorders, including Crohn’s disease, inflammatory bowel disease (IBD), spondylarthritis, and asthma (27). Notably, IBD patients exhibit transient Ruminococcus blooms correlating with disease activity (28), and its abundance associates with pro-inflammatory cytokines (e.g., IL-6, TNF-α) and LPS levels (29). Together with prior studies, our findings suggest that Streptococcus and Ruminococcus overgrowth may contribute to NMOSD pathogenesis via inflammatory mechanisms.

Furthermore, our data revealed depletion of beneficial commensal microbes (e.g., Faecalibacterium) in NMOSD patients. As a dominant bacterial genus in the healthy human gut, Faecalibacterium suppresses inflammation by producing SCFAs and IL-10. It also enhances gut barrier integrity, inhibiting pathogen invasion (30). The roles of Pseudomonas and Ralstonia in NMOSD remain unclear and warrant investigation. These findings suggest that gut microbiota dysbiosis, characterized by diminished commensal bacteria and increased opportunistic pathogens, may contribute to NMOSD pathogenesis through dysregulation of inflammatory mediators. However, further studies are needed to establish whether this dysbiosis initiates NMOSD or arises secondary to the disease.

To date, little is known about the potential link between gut microbiome metabolic pathways and NMOSD pathogenesis. Functional analyses revealed that gut microbiota alterations may contribute to NMOSD development through associated metabolic pathways. In this study, metabolomic profiling of fecal samples identified 19 significantly altered metabolites in treatment-naïve NMOSD patients. Furthermore, PICRUSt analysis uncovered 10 dysregulated microbiome-associated metabolic pathways related to NMOSD. Notably, elevated fecal levels of glucogenic amino acids (including D-proline, L-glutamate, and L-methionine) and glucogenic/ketogenic amino acids (particularly L-tyrosine) were observed in NMOSD patients. These findings suggest potential disruptions in glucose and energy metabolism, as these amino acids may serve as alternative energy substrates. Glutamate metabolism has been previously reported to modulate both resting and activated T-cell function (31). Proline, a key component of mucus glycoproteins in colonic epithelial cells (32), showed elevated levels in stool samples, potentially indicating alterations in mucin synthesis and intestinal barrier function. These findings suggest a potential link between amino acid metabolism and gut homeostasis. We hypothesize that impaired amino acid absorption may lead to their accumulation in the intestinal lumen. Most notably, the significantly enriched aminoacyl-tRNA biosynthesis pathway, involving L-methionine, L-tyrosine and D-proline, indicates dysregulated amino acid turnover and protein biosynthesis in NMOSD patients.

Our analysis revealed skewed vitamin B metabolism in fecal samples from NMOSD patients. Costliow et al. suggested that under conditions of thiamine limitation, the biosynthesis of thiamine is crucial for the growth and competitive fitness of B. thetaiotaomicron (33), a Bacteroides species (phylum Bacteroidetes) in the human gut microbiota. We hypothesize that elevated fecal thiamine levels in NMOSD patients may sustain microbial survival by altering gut vitamin B metabolism. However, the immunomodulatory role of vitamin B remains poorly understood. Investigating vitamin B-mediated immune regulation could provide insights into NMOSD pathogenesis and reveal novel therapeutic targets.

Metabolites associated with purine metabolism were dysregulated in NMOSD patients compared to HCs, including hypoxanthine, adenine, and adenosine. Adenosine is a key immunosuppressive mediator secreted by Treg cells, which suppresses self-reactive immune responses, promotes transplant tolerance, and mitigates autoimmune disorders (34–36). Additionally, lipid metabolism, which regulates critical cellular processes such as proliferation, differentiation, inflammation, and apoptosis, was significantly altered in NMOSD patients. Key perturbed metabolites in this pathway included linoelaidic acid, palmitoyl ethanolamide, and 5(S)-HETE. These findings suggest a potential link between lipid carbon-chain metabolism and NMOSD pathology. Thus, our study provides mechanistic insights into how gut microbiota-derived metabolic changes may contribute to NMOSD pathogenesis.

Notably, metabolism can be influenced by drug intake (37). Intriguingly, in the present study, not only did the gut microbiome composition in treatment-naïve NMOSD patients closely resemble that of HCs after immunosuppressive therapy, but principal component analysis also revealed significant differences in overall metabolic profiles between NMOSD patients and HCs. Following therapy, the gut microbial metabolites in these patients became nearly identical to those in HCs. This suggests that metabolic alterations may arise not only from the disease itself but also from therapeutic interventions. A limitation of this study is the small cohort of treatment-naïve NMOSD patients, which limits the statistical power to precisely evaluate medication-induced changes in fecal metabolites. Future studies should enroll a larger treatment-naïve cohort to better delineate treatment-specific metabolic shifts.

Nevertheless, our study has several limitations. First, as a single-center case-control study, the sample size was limited, particularly in the longitudinal immunosuppressive treatment group. This precluded robust subgroup analyses to assess how disease status or clinical presentation specifically affects microbiota abundance. Second, although early longitudinal data from five patients suggested clinical benefits from immunosuppressants, the scarcity of serial samples prevented a comparative analysis of microbial or metabolomic dynamics between treatment responders and non-responders. Third, while cross-sectional analyses indicate that therapy is associated with a microbiota shift towards a state resembling healthy controls, the most direct evidence would come from paired pre- and post-treatment samples. The limited availability of such longitudinal specimens restricted our ability to perform this definitive within-patient analysis. Therefore, our findings necessitate further validation in larger, multi-center prospective studies with comprehensive longitudinal sampling. In addition, dietary habits may influence microbial community dynamics; thus, standardized dietary questionnaires need to be incorporated into gut microbiota research. Finally, the follow-up period of this longitudinal cohort study was limited to 6 months, which may not fully capture the long-term effects of immunosuppressive therapy on gut microbiota in NMOSD patients.

In conclusion, our research reveals that intestinal dysbiosis and aberrant metabolic pathways exist in NMOSD patients. Immunotherapy may shift the gut microbiota of NMOSD patients toward that of HCs. These findings could not only help elucidate the underlying mechanism by which microbiome dysbiosis contributes to NMOSD pathogenesis but also provide insights into potential individualized treatments for the disease. Further studies with larger sample sizes and extended follow-up durations are warranted to confirm our findings.