Nonalcoholic fatty liver disease (NAFLD) is a prevalent condition affecting a significant portion of the global population, encompassing a spectrum of liver damage from simple steatosis to advanced fibrosis and cirrhosis. While typically associated with obesity, NAFLD also occurs in non-obese individuals, often displaying similar pathological severities. Understanding the pathogenesis of non-obese NAFLD is crucial because it represents a significant portion of the NAFLD population. Visceral fat, diet, and genetics have been implicated, but a comprehensive understanding remains elusive. The gut microbiome, a complex community of microorganisms residing in the intestines, has emerged as a critical factor in various metabolic diseases, including NAFLD. Gut dysbiosis, or an imbalance in the gut microbiota, is linked to changes in beneficial bacteria, bile acid composition, immune responses, and metabolic byproducts. This altered gut-liver axis is implicated in the progression of chronic liver diseases, including NAFLD, with implications for fibrosis development. However, prior research has yielded inconsistent results regarding specific microbial taxa associated with NAFLD severity and fibrosis stage, possibly due to variations in study populations and methodologies. This study aimed to investigate the association between histological NAFLD severity and stool microbial alterations in a well-characterized Asian NAFLD cohort, specifically focusing on the differences between obese and non-obese individuals, and to identify potential microbial markers for diagnosing non-obese NAFLD.

Numerous studies have explored the connection between gut microbiota and NAFLD, showing associations between gut dysbiosis and disease progression. These studies have highlighted the roles of reduced beneficial short-chain fatty acid (SCFA)-producing bacteria, altered bile acid profiles, increased immune responses, elevated ethanol production, and altered metabolism of phosphatidylcholine. The gut-liver axis is increasingly recognized as central to NAFLD pathogenesis, with changes in the gut microbiota influencing liver inflammation and fibrosis. However, findings concerning specific microbial taxa associated with disease severity and fibrosis stage have been inconsistent across studies. This inconsistency may be attributed to several factors including regional differences in microbiota composition and the influence of BMI. The current study aims to address these inconsistencies by examining a well-characterized Asian NAFLD cohort with a focus on the distinction between obese and non-obese participants and leveraging advanced technologies like 16S rRNA gene amplicon sequencing and metagenomic shotgun sequencing.

This study employed a multi-faceted approach involving human subject analysis and in vivo validation using mouse models. The human study comprised 171 biopsy-proven NAFLD subjects and 31 non-NAFLD controls of Asian descent, categorized by obesity status (BMI <25 vs. BMI ≥25). Subjects were further stratified into three subgroups based on fibrosis severity (F0-F4). 16S rRNA gene amplicon sequencing was used to analyze the gut microbiome composition in stool samples. Data were analyzed using univariate and multivariate statistical methods (MaAsLin) to identify associations between specific microbial taxa and fibrosis severity, adjusting for confounding factors like age, sex, BMI, diabetes, and cirrhosis. Stool metabolite profiles (bile acids and SCFAs) were assessed using Q-TOF and GC-FID systems. Network analysis explored interactions between microbial taxa and metabolites. For external validation, a publicly available Western NAFLD cohort was utilized. Metagenomic shotgun sequencing was performed on a subset of non-obese subjects to identify fibrosis-related species. In vivo validation involved administering four representative bacterial species from *Ruminococcaceae* and *Veillonellaceae* to three different mouse NAFLD models (MCD diet, CDAHFD diet, and *db/db* mice) to assess their impact on liver damage. Liver histology, serum markers (ALT, AST), fibrogenic gene expression, and bile acid levels were analyzed to evaluate the effects of the administered bacteria. Statistical analyses included nonparametric tests (Mann-Whitney, Kruskal-Wallis), and receiver-operating characteristic (ROC) curve analysis to assess the diagnostic performance of identified microbial markers.

The study revealed distinct gut microbiome alterations associated with fibrosis severity, primarily in non-obese NAFLD subjects. Microbial diversity significantly decreased with increasing fibrosis severity in the non-obese group, unlike the obese group. *Veillonellaceae* abundance increased, while *Ruminococcaceae* abundance decreased with increasing fibrosis severity, specifically in non-obese individuals. These associations were confirmed using multivariate analysis (MaAsLin), adjusting for age, sex, BMI, diabetes, and cirrhosis. Stool levels of primary bile acids and propionate were significantly elevated in non-obese subjects with more severe fibrosis. Network analysis showed strong inverse interactions between *Veillonellaceae* and *Ruminococcaceae* only in non-obese subjects. A combination of *Veillonellaceae*, *Ruminococcaceae*, and four stool metabolites (cholic acid, chenodeoxycholic acid, ursodeoxycholic acid, and propionate) achieved a high AUC (0.939) for diagnosing significant fibrosis in non-obese subjects, exceeding the diagnostic power of the FIB-4 index. The findings were validated using an independent Western NAFLD cohort, showing similar trends in *Veillonellaceae* and *Ruminococcaceae* abundance in relation to fibrosis severity in non-obese subjects. Metagenomic analysis identified *Ruminococcus bromii*, *Megamonas hypermegale*, and *M. funiformis* as key species associated with fibrosis. *Ruminococcus faecis* administration in mouse models showed a protective effect on liver damage, alleviating fibrosis and reducing fibrogenic gene expression. This effect was observed across three different mouse NAFLD models.

This study highlights the importance of considering obesity status when investigating the gut microbiome's role in NAFLD pathogenesis. The distinct microbiome and metabolite signatures associated with fibrosis severity in non-obese NAFLD patients suggest the potential for developing non-invasive diagnostic tools based on microbiome profiling and metabolite analysis. The strong association between specific bacterial taxa (*Veillonellaceae*, *Ruminococcaceae*) and metabolites with fibrosis severity underscores the complex interplay between the gut microbiota and liver disease progression. The protective effect of *R. faecis* in mouse models provides evidence for a potential therapeutic avenue targeting specific gut microbes to mitigate NAFLD-associated fibrosis. The validation of the findings in an independent Western cohort adds to the generalizability of these results. Future studies should focus on longitudinal investigations to determine the causal relationships and predictive value of these markers.

This study demonstrates distinct gut microbiome and metabolite signatures associated with fibrosis severity in non-obese NAFLD individuals. The identified bacterial taxa and metabolites offer potential as non-invasive diagnostic markers for significant fibrosis. *Ruminococcus faecis* shows promise as a therapeutic target. Further longitudinal studies are needed to confirm the causal relationships and predictive value of these markers. Future research could explore personalized interventions targeting the gut microbiome to prevent or treat NAFLD-associated fibrosis.

The study's cross-sectional design limits causal inference. The substantial number of diabetic patients and the use of anti-diabetic medications might have influenced the microbiome findings. The validation cohort used different fibrosis assessment methods. The smaller sample size in certain subgroups could have affected the statistical power. The lack of clear differences in obese NAFLD subjects might be due to the overriding influence of obesity on microbial composition.