Nicotinamide Mononucleotide (NMN) Prevents Age-Associated Transcriptional Drift in a Tissue-Dependent Manner:...

This study is a secondary computational reanalysis of the publicly available GSE85718 microarray dataset generated during the long-term nicotinamide mononucleotide (NMN) study. NMN has been reported to improve several age-sensitive physiological traits in mice, including energy metabolism, insulin sensitivity, plasma lipid profiles, and skeletal-muscle mitochondrial function. However, the transcriptional mechanisms behind these effects remain less clear than the broader nicotinamide adenine dinucleotide (NAD+) and sirtuin-centered model often used to explain NMN biology. The present analysis used a per-tissue age-by-treatment interaction model to test whether NMN modifies the rate of age-associated transcriptional change rather than simply shifting expression at one age. The model was expression ~ age × treatment, with the age-by-treatment term used as the central test. A gene was considered an NMN-rescue candidate only when it changed with age in control mice and showed an opposite-signed interaction term, consistent with NMN shifting old-age expression toward the young-control state. No individual gene reached genome-wide false discovery rate (FDR) <0.05 for age, NMN at six months, or the interaction term in skeletal muscle, liver, or white adipose tissue (WAT). Therefore, all gene-level results should be treated as hypothesis-generating. Using relaxed nominal criteria, 421 rescue candidates were identified in skeletal muscle, 355 in liver, and 397 in WAT. Only 35 genes were rescued in at least two tissues, 26 of which were direction-consistent, and none were rescued in all three tissues. Ras-related protein Rab-11A (RAB11A) emerged as the strongest cross-tissue candidate, with rescue in skeletal muscle and WAT, high confidence in at least one tissue, consistent directionality, and involvement in 39 gene set enrichment analysis (GSEA) leading-edge terms, largely related to trafficking and cellular transport. Carnitine palmitoyltransferase 2 (CPT2) was the only mitochondrial gene among the robust cross-tissue candidates and was consistently rescued in skeletal muscle and WAT, supporting a focused fatty-acid oxidation and substrate-handling hypothesis rather than broad mitochondrial activation. At the pathway level, liver showed the clearest signal: NMN was associated with suppression of fatty-acyl-coenzyme A (CoA) and long-chain fatty-acyl-CoA metabolic programs. These findings do not establish that NMN prevents transcriptional aging or that RAB11A or CPT2 mediates the physiological effects of NMN. Instead, they identify tissue-specific, testable candidates from a secondary reanalysis of an existing animal dataset. In particular, they support moving beyond a generic "NAD+ improves mitochondria" model toward testable mechanisms involving cellular logistics, membrane recycling, and substrate utilization.