Tissue-Resolved Transcriptome-Wide Association Study Reveals Telomere Maintenance, Cellular Senescence, Synaptic Refinement, and Reduced Insulin Drive as Transcriptional Mechanisms of Human Lifespan

Human lifespan is a complex polygenic trait shaped by common variants that appear to act mainly through protection from age-related disease rather than through a single general ageing program. To move from association toward biological mechanism, we performed a tissue-resolved transcriptome-wide association analysis using the common-effect-size parental-lifespan GWAS from Timmers et al. (2019; 1,012,240 parental lifespans) together with tissue-level S-PrediXcan results from subcutaneous adipose tissue, liver, and skeletal muscle. Across 42 curated gene sets, we combined gene-level association statistics with Stouffer Z aggregation, permutation testing, Wilcoxon testing, leave-one-out influence analysis, leave-three-out sensitivity analysis, and pooled family-wise false-discovery-rate correction.The strongest positive signal was the Reactome telomere pathway, with a meta Stouffer Z of 4.259, permutation p = .0201, and pooled family-wise FDR = .0402. This was the only finding to survive the pooled FDR threshold. Gene Ontology cellular senescence was also positively enriched, with a meta Stouffer Z of 3.610 and permutation p = .04539, and remained significant after removal of its three largest drivers. Negative meta-level enrichments were observed for KEGG long-term potentiation, regulation of synaptic plasticity, glucose-responsive insulin secretion, and energy-dependent regulation of mTOR by LKB1/AMPK. Synapse pruning, adult microglial markers, and KEGG apoptosis showed positive enrichment.The leading-edge genes pointed to telomere and DNA-replication maintenance genes including POLR2A, POLA2, POLD2, TERF1, DNA2, and POLD3; senescence-associated genes including YPEL3, MAP2K4, ATM, and CDKN1A; and synaptic genes including PPP3R1, SHISA6, CALM3, and HRAS. Only four of 11 Z-significant gene sets remained significant after removal of their top three drivers: KEGG long-term potentiation, KEGG apoptosis, cellular senescence, and regulation of synaptic plasticity. By contrast, NAD⁺ and nicotinate-nicotinamide metabolism were not significant at the meta-tissue level in the current framework, although residual tissue-specific and gene-level signals remain suitable for follow-up.These findings support a transcriptional architecture of human lifespan centered on telomere and DNA-maintenance programs, cellular senescence-related stress responses, synaptic refinement, and reduced glucose-stimulated insulin drive. The results do not support a dominant pan-tissue NAD⁺ salvage signal and should be interpreted as pathway-level associations requiring replication and causal validation.