Kinetic Lipidomics: Quantifying in vivo changes in lipid metabolism using metabolic labeling

Lipid metabolism reflects the dynamic balance between metabolic turnover and concentration. Kinetic mass spectrometry (MS) enables direct quantification of molecular turnover in vivo. Previous work has shown that MS-based kinetic proteomics has provided powerful insights into proteome regulation. Analogous lipidome-wide kinetic measurements remain limited by challenges in defining molecule-specific labeling behavior. Here, we extend kinetic MS to untargeted lipidomics. Isotope labeling with deuterated water (²H₂O) is commonly used for monitoring turnover of palmitate and other select lipids by measuring labeling of stable C-H positions with deuterium (²H). Here, we extend the deuterium-incorporation model underlying these targeted lipid turnover assays to support untargeted analysis of all detectable lipids. This allows us to empirically quantify the effective fraction of endogenous synthesis (A _syn) and the turnover rate ( $k$ ) across hundreds of lipid species simultaneously. One central barrier to lipidome-wide kinetic modeling is determining the endogenous number of deuterium-labeling sites for each molecule ( $n_L$ ) which is required to estimate $A_{syn}$ and $k$ accurately. The $n_L$ value is an essential component of biological kinetic assays. In kinetic proteomics, curated amino acid $n_L$ libraries enable peptide-level modeling by summing sequence-specific labeling-site values, but comparable resources are lacking for lipids and may not generalize across metabolic states or non-mammalian systems. Yet, gaps remain for lipids and for amino acids in modified metabolic conditions or non-mammalian biologies. Here, we empirically determine lipid $n_L$ values and validate the process with peptides against an $n_L$ library. To evaluate this strategy in a biologically relevant setting, we applied it to brain tissue from transgenic mice expressing human ApoE isoforms, where altered lipid transport and metabolism are implicated in Alzheimer's disease risk. These data validate the method in a clinically relevant context and suggest that genotype-dependent metabolism can alter empirically determined lipid n_L values.