Abstract
Abstract
Objective
This study aims to investigate the expression profiling characteristics of N6-methyladenosine (m6A) RNA modification regulators in sepsis, their associations with glycolytic metabolic reprogramming, and their preliminary prognostic value.
Methods
We integrated multiple sepsis transcriptomic cohorts (GSE65682, GSE95233, GSE54514) from the GEO database to systematically analyze the differential expression of 15 m6A regulators. Gene Set Variation Analysis (GSVA) scores, weighted gene co-expression network analysis (WGCNA) networks, and Spearman correlation analyses were utilized to assess the associations between m6A regulators and glycolytic activity. The regulation of glycolytic genes by m6A modifications was confirmed using ALKBH5 knockout models (GSE198316, GSE224650) and MeRIP-seq data (GSE225143). A prognostic model was developed by integrating CIBERSORT immune cell infiltration analysis, single-cell RNA sequencing data (GSE147363), and LASSO-Cox regression.
Results
Most m6A regulators are significantly downregulated in patients with sepsis, while IGF2BP2 and IGF2BP3 are upregulated. ALKBH5 clusters with IGF2BP2 within the same WGCNA module and exhibits a positive correlation with glycolytic activity (rho=0.592,
P
=1.0 × 10⁻
29
). The differentially expressed genes (DEGs) resulting from ALKBH5 knockout significantly overlap with sepsis-associated DEGs (
n
=124; OR=2.48,
P
=5.10 × 10⁻
11
), which are enriched in the HIF-1 signaling pathway and glycolytic processes. Independent validation confirmed a directional consistency of 89.6%. MeRIP-seq data indicated m6A-mediated associations with glycolytic genes, including ENO1, GAPDH, HK2, and LDHA. Single-cell analysis demonstrated a positive correlation between ALKBH5 and most glycolytic genes in CD14+ monocytes. A prognostic model based on four m6A genes (HNRNPC, YTHDF1, YTHDF2, VIRMA) achieved a C-index of 0.750 in the training set and 0.709–0.718 in external validation sets.
Conclusion
These integrated analyses suggest that ALKBH5-associated m6A modification is correlated with sepsis-associated glycolytic reprogramming and support a biologically plausible, hypothesis-generating ALKBH5-associated glycolytic program that requires prospective and experimental validation. The m6A-based four-gene model showed preliminary prognostic potential for 28-day mortality, but its external reproducibility was incomplete and requires validation in larger, prospectively collected and clinically harmonized cohorts.