Analysis of the whole transcriptome of deep vein thrombosis and immune-metabolic characteristics reveals the lipid...

Deep vein thrombosis (DVT) is a common venous thrombotic disease associated with substantial morbidity and mortality. Its onset and progression are closely related to coagulation abnormalities and are also profoundly influenced by inflammatory responses and immune regulatory networks. However, the molecular mechanisms underlying DVT in the context of systemic inflammation, particularly the key processes driven by coordinated immune-metabolic interactions, remain to be elucidated.

Twenty clinical parameters were integrated to systematically characterize patients with DVT. Whole-transcriptome sequencing of peripheral blood samples was performed to profile differentially expressed mRNAs, lncRNAs, miRNAs, and circRNAs. Core diagnostic markers were identified using an integrated machine learning strategy combining random forest, LASSO, and SVM-RFE. Single-sample gene set enrichment analysis (ssGSEA) was used to deconvolute immune microenvironment features, and quantitative metabolic pathway analysis was performed to explore the metabolic basis of neutrophil functional polarization. Finally, an independent external validation cohort (GSE48000) was used to evaluate the robustness of the key mechanisms and their potential value in recurrence risk stratification.

Clinical analysis showed that D-dimer levels were significantly elevated in patients with DVT, and that the systemic inflammatory indices AISI and NLR were also increased, indicating an inflammatory phenotype characterized by enhanced myeloid stress and relative lymphocyte suppression. Whole-transcriptome sequencing identified a candidate biomarker panel centered on F7, MTRNR2L10, and SYT16, and the resulting diagnostic model showed excellent discriminatory ability. Immune deconvolution further identified Neu_NETotic_Hub as a key pathogenic subgroup in DVT, which was significantly enriched in the DVT group and closely associated with high expression of the core driver gene PADI4. Metabolic analysis showed that neutrophils in the DVT state exhibited marked lipid metabolic reprogramming and enhanced reactive oxygen species (ROS) generation, both of which were strongly positively correlated with the NETotic score (r > 0.8), suggesting that they may constitute an important metabolic basis for immunothrombosis. External validation showed that both the lipid metabolic score and the NETs-related score increased stepwise with recurrence risk grade, and that a combined model based on these two metrics showed good discriminatory ability for identifying high-risk recurrence states AUC =0.825.

This study revealed an immune-metabolic coupling axis in which lipid metabolic reprogramming drives neutrophils toward a pathogenic NETotic phenotype, suggesting that this axis may represent an important mechanism promoting the onset, progression, and recurrence risk of DVT. These findings not only deepen our understanding of the biological basis of immunothrombosis in DVT, but also provide a theoretical foundation for early diagnosis, recurrence risk stratification, and targeted interventions aimed at neutrophil immunometabolism.