Human Th17 state classification reveals a selective vulnerability in acylcarnitine metabolism in type 2 diabetes

Th17 cells are categorized as non-pathogenic (Th17np) and pathogenic (Th17p), states defined primarily in murine autoimmune models. In contrast, studying these states in humans remains difficult because these subsets lack validated identifiers. We developed a nanovial-based single-cell ELISA to isolate cytokine-secreting human Th17 cells and define Th17np vs Th17p cell transcriptional states. These signatures enabled predictive classification and flow-cytometric identification across tissues and diseases. Application of these strategies with single-cell metabolic modeling and functional profiling revealed that Th17np cells predominate in type 2 diabetes yet exhibit impaired fatty acid oxidation capacity despite elevated carnitine palmitoyltransferase 1A (CPT1A). CRISPR perturbation and pharmacologic inhibition identified acylcarnitine metabolism as a regulator of Th17 state balance and exposed a selective metabolic vulnerability in type 2 diabetes. This framework links human Th17 cell pathogenicity to lipid metabolism.