SAM-Dependent m6A MethylationMetabolic Feedback Loop Drives Liver Fibrosis Progression: A Public Database-Based Multi-Omics Study

Abstract

Background and Aims: Liver fibrosis is a critical pathological transition in chronic liver disease, predisposing patients to cirrhosis
and hepatocellular carcinoma (HCC). S-adenosylmethionine (SAM), the universal methyl donor synthesized by methionine adenosyltransferase 1A (MAT1A) in the liver, serves as the obligate co-substrate for N6-methyladenosine (m6A) RNA methylation. A shift from MAT1A to
MAT2A expression occurs during liver injury and reduces hepatic SAM availability. Whether this metabolic reprogramming rewires the m6A
epitranscriptomic landscape and creates a self-reinforcing feedback loop that drives fibrosis progression has not been systematically investigated. The present study addresses this question using integrative multi-omics analysis of publicly available databases. Method: Transcriptomic datasets GSE84044 (n = 124, hepatitis B virus [HBV]-related, fibrosis stages S0S4), GSE149601 (n = 79, non-alcoholic steatohepatitis
[NASH], F0F4), GSE103580 (n = 82, alcohol-associated liver disease [ALD], F0F4), GSE33650 (n = 41, hepatitis C virus [HCV]-related,
F0F4), and GSE162694 (n = 143, NASH, F0F4) were retrieved from the Gene Expression Omnibus (GEO). The single-cell RNA sequencing (scRNA-seq) dataset GSE136103 (>100, 000 cells, healthy vs. cirrhotic liver) was used for cell-type-specific analysis. One-carbon metabolism and SAM biosynthesis pathway activity was evaluated by gene set variation analysis (GSVA). Differential expression of m6A regulators
was assessed using the limma package. Weighted gene co-expression network analysis (WGCNA) was performed on the merged GSE84044
and GSE162694 cohort (n = 267). Proteinprotein interaction (PPI) networks were constructed using the STRING database (v12.0). Intercellular communication was analyzed with CellChat v1.6.0. Transcription factor regulatory networks were inferred with pySCENIC v0.12.1. A
seven-gene m6ASAM axis signature score was constructed and validated across three independent cohorts. Result: MAT1A was progressively downregulated across all six bulk transcriptomic cohorts (log2FC = ?1.84, FDR < 0.001 in GSE84044), while MAT2A was consistently upregulated (log2FC = +1.21, FDR < 0.001), defining an etiology-independent metabolic switch. One-carbon metabolic pathway activity
declined progressively from F0 to F4 (mean GSVA score: +0.31 0.14 vs. ?0.28 0.17; p < 0.001). The m6A methyltransferases METTL3
and METTL14 were upregulated, while the erasers FTO and ALKBH5 were downregulated in advanced fibrosis. The reader protein YTHDF1
(log2FC = +0.77, FDR < 0.001) emerged as the central hub (degree centrality k = 23) in the PPI network. WGCNA identified a pro-fibrotic
turquoise module (847 genes; r = +0.73 with fibrosis stage, p < 0.001) reciprocally correlated with a hepatocyte differentiation-associated blue
module harboring MAT1A. Single-cell transcriptomic analysis showed that MAT2A and YTHDF1 were markedly co-induced in activated
hepatic stellate cells (HSCs) relative to quiescent HSCs (both p < 0.001), while MAT1A was severely suppressed in cirrhotic hepatocytes (p <
0.001). The m6ASAM axis composite signature score achieved AUROCs of 0.87, 0.83, and 0.80 for discriminating advanced fibrosis in the
three cohorts (all p < 0.001). Conclusion: This multi-omics study establishes a SAM-dependent m6A methylationmetabolic feedback loop as
a previously unrecognized driver of liver fibrosis progression. These findings identify YTHDF1, MAT2A, and the SMAD3/SP1 transcriptional
axis as potential targets for epitranscriptomic intervention.