Aging is a major risk factor for chronic diseases, and calorie restriction (CR) is a robust non-pharmacological intervention that can extend health span in multiple species. Alternative splicing (AS) generates multiple RNA isoforms from a single pre-mRNA and becomes dysregulated with age; intriguingly, prior work suggests that CR can attenuate age-associated splicing noise. What has remained unclear is whether AS responses to CR are coordinated across tissues and whether they scale with the level of restriction.
In a new study published in Life Metabolism, Prof. John R. Speakman and colleagues analyzed multi-tissue RNA sequencing (RNA-seq) data from mice undergoing graded CR and show that a 3-month CR intervention induces a largely tissue-specific, yet functionally convergent, AS program that is largely independent of transcriptional (gene expression) changes.
Male C57BL/6J mice were maintained for 3 months on graded CR (10%, 20%, 30%, or 40% restriction) and compared with a 12-h ad libitum feeding control. The authors analyzed RNA-seq data from six tissues, including the epididymal white adipose tissue (eWAT), liver, hypothalamus, gastrocnemius muscle, testes, and stomach, quantifying differential gene expression (DE) and differential alternative splicing/differential transcript usage (DAS/DTU) at isoform resolution (Figure 1).
The gene-expression response to CR was strongly tissue-dependent, with the eWAT showing the greatest number of differentially expressed genes (DEGs), followed by the muscle and the liver. At 40% CR, only two genes (H2-Aa and H2-Eb1; both being MHC class II components) were consistently down-regulated across all six tissues, consistent with a systemic shift away from inflammatory antigen-presentation programs. In contrast, the AS response scaled with CR level across tissues but involved largely distinct loci in each tissue: approximately 94% of loci showing DTU were not differentially expressed, highlighting largely independent regulation of splicing versus transcription. Notably, the testes displayed a pronounced AS response despite relatively modest DE changes. Despite limited overlap of specific DAS loci between tissues, functional enrichment of DAS/DTU genes converged on shared processes—including the mitochondria and oxidative phosphorylation, ribosome/translation, and RNA and protein quality-control pathways—supporting the idea of a functionally integrated, cross-tissue program. A small subset of loci showed cross-tissue isoform switches, including Gna13 and Nfe2l2, as well as genes linked to endosomal sorting and extracellular vesicle biology (e.g., Arrdc4 and Pdcd6ip).
Overall, the study supports a model in which AS is a dose-responsive component of the CR adaptation, potentially intersecting with hallmarks of aging such as RNA/protein homeostasis. The authors note several important limitations: short and heterogeneous sequencing read lengths and depths across tissues, relatively small group sizes, a male-only cohort, a short (3-month) intervention, and the lack of functional validation. Future work leveraging long-read sequencing, both sexes, and mechanistic perturbations will be needed to determine which CR-responsive isoform changes are causal versus correlative.
