Multi-Omics and Machine Learning-Based Characterization of the Lactylation Microenvironment and Biomarker Identification in Crohn’s Disease Intestinal Fibrosis

Crohn’s disease (CD) is characterized by transmural inflammation and intestinal fibrosis, in which metabolic reprogramming may contribute to fibrotic remodeling through lactate-associated epigenetic and transcriptional regulation, but key cellular states and candidate biomarkers remain unclear. Therefore, we integrated single-cell RNA sequencing (scRNA-seq) with a lactylation-associated transcriptional score to estimate lactate/lactylation-related transcriptional activity in CD intestinal tissues. High-dimensional weighted gene co-expression network analysis (hdWGCNA) and an integrated machine learning framework identified core lactylation-associated biomarkers, which were validated in clinical tissue and a TNBS-induced mouse model. Additionally, we evaluated the ability of these core genes to predict anti-TNF-α treatment response in an independent clinical cohort, and analyzed cellular communication and trajectories, using CellChat, Monocle 2, and spatial transcriptomics. In this study, an enterocyte state with high lactylation-associated transcriptional scores was identified. CALD1 and CALM1 were identified as core transcriptional biomarkers associated with this high-lactylation-score epithelial state. A diagnostic nomogram exhibited robust predictive performance (AUC > 0.75, p < 0.05). Notably, increased CALM1 expression was associated with favorable anti-TNF-α response (AUC > 0.70, p < 0.05). Mechanistically, CALD1/CALM1 dysregulation was associated with enhanced GDF15- and MIF-related signaling, potentially contributing to endothelial and macrophage-associated microenvironmental remodeling. Together, CALD1 and CALM1 represent lactylation-score-associated transcriptional biomarkers that may link metabolic reprogramming, epithelial dysfunction, and fibrotic remodeling in CD, serving as candidate diagnostic and treatment-response-associated biomarkers.