Glycolytic-Cholesterol Subtypes of Severe Asthma Reveal Distinct Immune-Inflammatory and Metabolic Phenotypes

Severe asthma (SA) is a heterogeneous disease with unmet therapeutic needs. Metabolic dysregulation involving glycolysis and cholesterol synthesis is implicated in its pathogenesis. This study aimed to systematically profile the glycolysis-cholesterol synthesis axis in SA to identify distinct metabolic subtypes and explore their potential therapeutic implications.

We analyzed sputum mRNA expression from 93 SA patients and 16 healthy controls (HC) in the U-BIOPRED cohort (GSE76262). First, differentially expressed genes (DEGs) in the glycolysis-cholesterol synthesis axis were identified between SA patients and HC. Unsupervised consensus clustering was then applied to these DEGs within the SA cohort to define metabolic subtypes. Immune microenvironment features were characterized using single-sample gene set enrichment analysis. A random forest (RF) model was built to distinguish subtypes, and candidate therapeutic compounds were screened for each subtype via the Connectivity Map (CMap) database based on subtype-specific gene signatures.

This study revealed that, 47 genes in the glycolysis-cholesterol synthesis axis were dysregulated in SA when compared to HC. Clustering of SA based on these DEGs revealed two distinct metabolic subtypes (Cluster 1, n=52; Cluster 2, n=41). Cluster 1 exhibited an up-regulated axis with an immune landscape enriched for NK and gamma-delta T cells, linked to oxidative phosphorylation. Conversely, Cluster 2 displayed a down-regulated axis with enrichment of eosinophils, neutrophils, mast cells, and Th17 cells, associated with calcium and cAMP signaling. An RF model utilizing PPP2CB and SEH1L achieved accurate subtype discrimination (AUC=0.888), validated by decreased protein expression in an SA murine model. Exploratory drug screening suggested dipeptidyl peptidase inhibitor and mTOR inhibitor were identified as candidate compounds for Cluster 1, while leukotriene receptor antagonist and calcium channel blocker were suggested for Cluster 2 via CMap analysis.

This study reveals two novel metabolic subtypes in SA, which offers a classification model and suggests potential targeted therapeutics for personalized management.