Longitudinal Metabolomics in Amyotrophic Lateral Sclerosis Implicates Impaired Lipid Metabolism.

Abstract

Amyotrophic lateral sclerosis (ALS) is a fatal neurodegenerative disease characterized by altered metabolome and energy homeostasis, manifesting with body mass index changes and hypermetabolism-both prognostic of disease progression and survival. The cross-sectional ALS metabolome has been characterized, but longitudinal correlations to functional decline are lacking. We longitudinally evaluated metabolomes from ALS plasma and terminal postmortem spinal cord and brain motor cortex tissue. We constructed 3 plasma models. A linear mixed effects model correlated all metabolite levels across all timepoints to their corresponding functional scores. An interaction model predicted a longitudinal change in function from baseline metabolites, whereas a progression model identified metabolites linked to a 20% or 50% drop in function. In postmortem samples, differential metabolites in onset versus second spinal cord segments served as a surrogate of disease progression. Mendelian randomization assessed potential causality from metabolites. In plasma, all models primarily selected lipid metabolites and sub-pathways, in addition to amino acids, xenobiotics, and various less frequently selected pathways. Among lipids, fatty acids and sphingomyelins were predominant, along with plasmalogens, phosphatidylcholines, and lysophospholipids. Sex interaction findings were nominal. In the spinal cord, sphingomyelin and long-chain saturated and monounsaturated fatty acids were more abundant in the onset segment tissue, whereas phosphatidylcholines and phosphatidylethanolamines were less abundant. Mendelian randomization suggested that impaired carnitine and short chain acylcarnitine metabolism may be genetically determined in ALS, along with various antioxidant derivatives. Our findings suggest metabolomic changes primarily involving different lipid classes and carnitine metabolism may underscore ALS severity and progression. ANN NEUROL 2025;98:19-34.