Background: Haemophagocytic lymphohistiocytosis (HLH) is a life-threatening hyperinflammatory syndrome, characterised by cytotoxic lymphocyte defects and uncontrolled activation/proliferation of CD8 ⁺ T cells and macrophages. In HLH, dampened high- and low-density lipoprotein cholesterol (HDL-C/LDL-C) levels and hypertriglyceridemia have been reported, suggesting lipid metabolism disturbances. However, comprehensive studies into the immunometabolism of adults with HLH (aHLH) is lacking.

Objectives: This study first used metabolomic comparisons between adults hospitalised with HLH and clinically relevant disease mimics and then integrated metabolomics, whole-blood transcriptomics, and detailed clinical information from matched aHLH patients to characterise disease mechanisms, define the immunometabolic landscape, and identify potential therapeutic targets.

Methods: Blood from patients with aHLH diagnosed using HScore and expert clinical review and stratified as hospitalised (n=25) or recovered (n=18) at time of sample collection was analysed using serum metabolomics (NMR spectroscopy, n=250 metabolites) and matched whole blood RNA sequencing. Serum metabolomics was additionally performed in patients with HLH disease mimics, including sepsis, malignancy, and rheumatological disease (systemic lupus erythematosus, rheumatoid arthritis, and systemic juvenile idiopathic arthritis), for comparative analyses. Metabolomic data was analysed using supervised machine learning, univariate logistic regression to stratify patients and metabolite enrichment analysis to identify dysregulated metabolic pathways. Results were validated by applying metabolic module analysis on differentially expressed genes between hospitalised and recovered aHLH patients. RNA-sequencing data was further analysed using gene-set enrichment (GSEA) and deconvolution (CIBERSORTx) analysis to identify dysregulated pathways and inferred cell type proportions associated with the gene expression profiles of adults hospitalised with HLH. Lastly, metabolomic, transcriptomic and clinical data were integrated together using multi-omic factor analysis (MOFA).

Results: Metabolomic data stratified hospitalised from recovered patients with 89-100% accuracy. Lipid metabolites indicative of dyslipidaemia and lipoprotein remodelling towards triglyceride enrichment and cholesterol depletion were elevated in hospitalised adults with HLH, suggesting impaired cholesterol efflux and reverse cholesterol transport in aHLH, potentially leading to lipid accumulation within immune cells. Lipids reflecting this phenotype, including HDL-triglycerides (AUC=0.908), LDL-triglycerides (AUC=0.964), intermediate-density lipoprotein triglycerides (IDL-TG; AUC=0.964), very low-density lipoprotein cholesterol (VLDL-C; AUC=0.873) in addition to saturated fatty acids, AUC=0.850), effectively discriminated adults hospitalised with HLH from disease mimics (sepsis, malignancy and rheumatological disease). This signature was validated by metabolic module analysis on differentially expressed genes, which highlighted upregulated pathways associated with lipid, glucose and amino acid metabolism: triacylglycerol, fatty acid biosynthesis, glycosphingolipid metabolism, glycolysis, tyrosine, and leucine degradation (padj<0.0001). GSEA further identified significantly upregulated mTORC1 signalling (padj=4.27×10 ⁻⁴ ) in hospitalised patients, together with mTORC1-regulated processes including glycolysis, oxidative phosphorylation, cholesterol homeostasis, fatty acid metabolism and cell proliferation-associated pathways. Deconvolution analysis revealed mTORC1 activation across immune subsets, with altered lipid metabolism strongly associated with monocytes exhibiting an M2 macrophage-like phenotype. Features included evidence of enhanced lipid uptake via scavenger receptor expression/biosynthesis, impaired LDL receptor degradation, and altered lipid droplet turnover, potentially driving inflammasome hyperactivation. Finally, MOFA identified a metabolomic-transcriptomic signature that effectively stratified hospitalised from recovered patients (p=3.40×10 ⁻⁷ ), after adjusting for age, sex, ethnicity, underlying cause for HLH and treatment. This signature correlated significantly with HLH-associated clinical parameters (ferritin, triglycerides, albumin, haemoglobin, lymphocyte, platelet, neutrophil and red blood cell counts) and highlighted therapeutic targets that could attenuate mTORC1 and promote cholesterol efflux from cells.

Conclusions: mTORC1-driven dysregulation of lipid and energy metabolism may sustain hyperinflammation in aHLH by fuelling pathogenic processes such as cell proliferation and inflammasome activation. Integrated metabolomic and transcriptomic profiling stratifies disease state among aHLH patients and highlights lipid-related signatures with therapeutic potential, while metabolomic profiling identified potential lipid biomarkers that could distinguish aHLH from disease mimics.

REFERENCES: NIL.

Acknowledgments: NIL.

Disclosure of Interests: None declared.