Dynamic lipidome alterations associated with human health, disease and ageing

Lipids, a diverse class of molecules, play crucial roles in various biological processes, including cell membrane structure, signaling, and energy metabolism. Despite their importance, our understanding of human lipid diversity, inter-individual variations, and temporal changes during health and disease remains limited. This knowledge gap hinders our ability to fully understand processes like aging and the roles of lipids in health and disease. High-throughput omics technologies, particularly mass spectrometry (MS), offer powerful tools to investigate the complex lipidome. While genomics and proteomics have been extensively studied using next-generation sequencing and MS, respectively, lipidomics has lagged due to the chemical diversity of lipids. This study aims to address this gap by performing comprehensive longitudinal lipidomic profiling to characterize lipidome dynamics across health and disease states in humans.

Existing literature highlights the critical roles of lipids in maintaining metabolic homeostasis and mediating inflammatory processes. Studies have linked abnormal lipid profiles (dyslipidemia) to various diseases, including metabolic syndrome, type 2 diabetes (T2D), cancer, and cardiovascular and neurodegenerative diseases. The interplay between lipids and inflammation is particularly complex, with lipids involved in both the induction and resolution of inflammation. Previous research has focused on individual lipid species or classes, but a comprehensive understanding of the dynamic interplay of the entire lipidome across various health states and over time is lacking. This study builds upon previous work by utilizing high-throughput quantitative lipidomics to investigate the lipidome's dynamic nature in a large, longitudinal cohort.

The study utilized a longitudinal cohort of 112 participants with either insulin resistance (IR) or insulin sensitivity (IS), followed for up to 9 years (average 3.2 years). Plasma samples were collected at regular intervals, with increased frequency during illness or stress. A high-throughput quantitative lipidomics pipeline (Lipidyzer) employing a triple-quadrupole mass spectrometer and differential mobility separation (DMS) was used to characterize the lipidome, identifying and quantifying >1,000 lipid species across 16 subclasses. To ensure accurate quantification, a mix of 54 deuterated spike-in standards was included. Data were processed rigorously, with quality control measures implemented to ensure high data quality. Statistical analyses included variance decomposition, t-distributed stochastic neighbor embedding (t-SNE), weighted gene correlation network analysis (WGCNA), linear mixed models, and enrichment analyses. The study also incorporated clinical laboratory measurements, medical records, and cytokine profiling.

The study revealed several key findings: 1. **High Individuality of Lipid Signatures:** Lipid signatures were highly individualized, with substantial inter-individual differences observed, particularly among triacylglycerols (TAGs), sphingomyelins (SMs), hexosylceramides (HCERs), and cholesterol esters (CEs). 2. **Lipid-Clinical Measure Associations:** WGCNA identified lipid modules strongly associated with clinical measures. Modules enriched in ceramides (CERs) and PEs, as well as small and large TAGs, showed positive associations with T2D markers, inflammatory markers, and negative associations with HDL levels. Conversely, a module enriched in phosphatidylethanolamines with alkenyl ether (PE-P) and alkenyl ether substituent containing PEs (PE-O) correlated with higher HDL and lower fasting insulin. 3. **Lipidome Disruption in IR:** Individuals with IR exhibited distinct lipid signatures, with increased levels of TAGs, DAGs, and subsets of CERs. Ether-linked PEs (PE-P), in contrast, were associated with lower SSPG levels, suggesting a potential link between oxidative stress, inflammation, and IR. Lipid-clinical measure associations also differed significantly between IR and IS groups. 4. **Dynamic Lipidome Changes During Viral Infections:** Significant lipidome changes were observed during respiratory viral infections (RVIs), with alterations in various subclasses, including ether-linked PEs and TAGs containing saturated fatty acids (SFAs). Distinct temporal patterns were identified, suggesting shifts in energy metabolism and inflammation resolution. IR/IS status also modulated the response to infections. 5. **Age-Associated Lipidome Changes:** Most lipid subclasses increased with age, notably CERs, SMs, LPCs, and CEs. Ageing was accompanied by increased SFAs and MUFAs, and reduced PUFAs and omega-3 FAs. IR accelerated ageing-related lipid changes. Sex dimorphism was also evident. 6. **Cytokine-Lipidome Networks:** Many significant associations were observed between lipids and cytokines/chemokines. TAGs showed strong positive associations with GM-CSF and leptin, while lyso-species of PEs and PCs showed fewer and less central roles. FA composition influenced these associations. LPCs exhibited both pro- and anti-inflammatory associations, depending on context.

This study significantly advances our understanding of the human lipidome and its dynamic role in health and disease. The findings highlight the importance of considering the entire lipidome, rather than focusing solely on conventional clinical lipid profiles, for a comprehensive assessment of metabolic health. The identification of highly individualized lipid signatures underscores the potential for personalized medicine approaches in the prevention and treatment of metabolic disorders. The associations between specific lipid subclasses and clinical measures, particularly in the context of IR and viral infections, offer promising avenues for biomarker discovery and therapeutic interventions. The intricate cytokine-lipidome networks provide further mechanistic insights into immune regulation and inflammation.

This large-scale longitudinal study provides a comprehensive characterization of dynamic lipidome changes associated with health, disease, and aging. The findings reveal highly individualized lipid signatures, identify key lipids associated with clinical measures and disease states, and uncover intricate cytokine-lipidome networks. This work highlights the potential of targeted lipidomics for biomarker discovery and personalized interventions in metabolic health, inflammation, and aging. Future research should focus on exploring the mechanistic roles of identified lipids and developing targeted interventions based on these findings.

While this study is comprehensive, several limitations should be considered. The cohort, although diverse, was biased towards middle-aged, highly educated participants from Northern California. Some analyses were based on smaller sample sizes. The lipidomics pipeline, while comprehensive, did not resolve all lipid molecular identities. The study used descriptive rather than predictive models. Finally, correlations do not imply causality, and lifestyle factors could also influence the findings.