Trimetazidine delays autoimmune diabetes in mice by transiently suppressing CD8? T-cell fatty-acid oxidation

Type 1 Diabetes Mellitus (T1D) is an organ specific autoimmune disease, which is characterized by persistent hyperglycemia due to immune-mediated destruction of pancreatic islet ß cells. Shifting immune cell metabolism is an emerging therapeutic concept. Therefore, we tested whether the fatty acid oxidation (FAO) inhibitor trimetazidine (TMZ) can restrain autoreactive immunity and delay T1D in a non-obese diabetic mouse model. Of note, TMZ is one of only three approved drugs directly targeting cellular metabolism. TMZ acutely increased mitochondrial membrane potential, suppressed FAO and curtailed activation and proliferation of human CD8? T cells. In dysglycemic NOD mice, a clinically approved dose of TMZ delayed progression to T1D, lowered mean glycaemia and reduced islet CD4?/CD8? infiltration. scRNA-seq revealed depletion of FAO-high, stress-responsive secretory cells and mitochondrially active stromal cells, indicating improved pancreatic health. However, prolonged exposure elicited compensatory up-regulation of carnitine-palmitoyl-transferase-1A in CD8? subsets, normalized T-cell infiltration and nullified protection. We conclude that TMZ delivers a dampening of initial T-cell pressure and pancreatic stress. Adaptive upregulation of FAO within cells counteracted observed benefits thereby only transiently delaying disease onset. Our work demonstrates that drugs targeting cellular metabolism needs to retain sufficient potency to not be outpaced by immune cells upscaling metabolic pathways. Overall design: Pancreatic tissue was collected from five control mice and five TMZ-treated mice after one week of treatment, initiated when fasting blood glucose reached 150-180 mg/dL. Pancreatic heads were dissected, snap-frozen, and stored in liquid nitrogen until processing. Tissue dissociation and fixation were performed using the Chromium Next GEM Single Cell Fixed RNA Sample Preparation Kit (PN-1000414, 10x Genomics, Pleasanton, CA, USA). Nuclei and single cells were isolated following the protocol “Tissue Fixation & Dissociation for Chromium Fixed RNA Profiling” (CG000553). Samples from each group (control and TMZ-treated) were pooled and assigned one barcode per group for multiplexed analysis. Library preparation was performed using the “Chromium Fixed RNA Profiling for Multiplexed Samples” protocol (CG000527, protocol version F), with the Chromium Fixed RNA Profiling Kit - Mouse Transcriptome (4 reactions × 4 barcodes, PN-1000496), Hybridization & Library Kit (PN-1000415), and the Chromium Next GEM Chip Q Single Cell Kit (16 reactions, PN-1000422), all from 10x Genomics. Sequencing was performed on an Illumina NextSeq 2000 instrument using a NextSeq 2000 P4 flow cell (100 cycles, PN-20100994, Lot #20879630, Illumina, San Diego, CA, USA). Sequencing data were demultiplexed and aligned using Cell Ranger software version 8.0.1 (10x Genomics), with the Mus musculus mm10 (Ensembl 98). Cell Ranger Loupe Browser output files were imported into R (version 4.4.1) for downstream processing. Nuclei with >45% mitochondrial transcript content, nFeature_RNA = 500, or nCount_RNA = 500 were excluded from analysis. Filtered data were normalized, log-transformed, and the resulting matrix was used for all downstream analyses. Data integration was performed using RPCA-based batch correction implemented in Seurat v5 (https://satijalab.org/seurat/articles/seurat5_integration). Principal component analysis was conducted using the first 20 components, and clustering was performed with a resolution of 0.1. The final annotated dataset was exported as a cloupe file and reloaded into the Loupe Browser (10x Genomics) for visualization and manual inspection of clusters. Pseudo-bulk differential expression analysis was performed using the DElegate package (https://github.com/cancerbits/DElegate).