Longevity & AgingResearch PaperOpen Access

Aging Bone's Hidden Signaling Flaw Traced to a Single Phosphorylation Switch

Phosphoproteomics reveals how P70S6K phosphorylation of AFF4 drives gene-specific transcription in bone—and fails in aged, insulin-resistant mice.

Saturday, September 19, 2026 1 view
Published in Nat Commun
Glowing molecular scaffold of AFF4 protein amid chromatin strands inside an osteoblast, with phosphorylation sites highlighted in amber light.

Summary

Researchers mapped insulin signaling in mouse bone for the first time using phosphoproteomics, comparing young lean mice to old obese insulin-resistant mice. They discovered hundreds of dysregulated phosphorylation sites and used a zebrafish genetic screen to identify key regulators of bone development. The most striking finding was that the transcriptional scaffold protein AFF4 is phosphorylated at serine 831 by P70S6K in response to insulin—and this modification is blunted in aged, insulin-resistant bone. This single phosphorylation event recruits chromatin remodelers ENL and AF9 to activate specific genes, linking insulin resistance to defective transcriptional elongation in osteoblasts.

Detailed Summary

Insulin's role in bone health extends well beyond skeletal maintenance—bone-derived signals help govern whole-body glucose metabolism. Yet until now, no study had mapped the full insulin signaling network in bone tissue or determined how aging and insulin resistance rewire it. This study fills that gap with the first global phosphoproteomic and proteomic analysis of insulin action in bone.

The researchers compared 10-week-old lean, insulin-sensitive and 73-week-old obese, insulin-resistant C57BL/6J male mice. After fasting and injecting saline or insulin, tibiae were rapidly harvested and processed for multiplexed stable-isotope phosphoproteomics. The aged mice showed classic metabolic dysfunction—elevated fasting glucose, high HOMA-IR, reduced lean mass, increased adiposity—alongside lower trabecular bone volume and cortical thickness confirmed by micro-CT. Western blotting confirmed that insulin-stimulated Akt phosphorylation was preserved in aged bone, but RPS6 phosphorylation was attenuated, pointing to a post-Akt signaling defect.

Phosphoproteomics identified thousands of phosphopeptides, with hundreds significantly altered by insulin or age. Kinase activity inference revealed broad suppression of P70S6K and mTORC1 substrate activity in aged bone. To prioritize which dysregulated phosphoproteins matter functionally for the skeleton, the team ran a zebrafish CRISPR/morpholino screen targeting 68 candidates and assessed bone growth, development, and mineralization. Several genes scored as essential for skeletal formation, including AFF4—the core scaffold of the Super Elongation Complex (SEC), which releases paused RNA Polymerase II into active transcription.

Deep mechanistic work showed that AFF4 serine 831 (S831) is an insulin-dependent P70S6K phosphorylation site. This site was robustly phosphorylated in young bone after insulin stimulation but markedly attenuated in aged, IR bone. In cultured osteoblasts made insulin-resistant via palmitate treatment, S831 phosphorylation was similarly defective. ChIP-seq and GRO-seq experiments revealed that loss of S831 phosphorylation was associated with reduced transcriptional elongation at a discrete set of genomic loci—not globally, but gene-specifically. Mechanistically, phospho-S831 promotes recruitment of the YEATS domain-containing proteins ENL and AF9 to the SEC. These chromatin readers bind crotonylated histones (a histone acylation mark associated with active transcription), thereby increasing local chromatin accessibility and driving gene activation. Importantly, this mechanism operates independently of P-TEFb kinase activity, identifying a novel parallel pathway for elongation control.

The study's findings reframe how insulin resistance affects gene regulation: not only through blunted kinase cascades at the level of initiation, but through defective phosphorylation of elongation machinery, causing locus-specific transcriptional silencing. This could explain why insulin-resistant bone fails to upregulate specific anabolic and metabolic gene programs even when proximal insulin signaling (e.g., Akt) appears intact.

Key Findings

  • Aged insulin-resistant mouse bone shows rewired phosphorylation, with suppressed P70S6K/mTORC1 substrate activity despite intact Akt signaling.
  • A zebrafish CRISPR/morpholino screen of 68 bone-enriched phosphoproteins identified AFF4 as essential for skeletal development and mineralization.
  • P70S6K phosphorylates AFF4 at serine 831 in an insulin-dependent manner; this is blunted in aged and insulin-resistant osteoblasts.
  • Phospho-S831 AFF4 recruits YEATS-domain proteins ENL/AF9 to crotonylated histones, driving gene-specific transcriptional elongation.
  • Insulin resistance causes locus-specific transcriptional elongation defects in bone, independent of P-TEFb kinase activity.

Methodology

Male C57BL/6J mice aged 10 or 73 weeks were fasted, then injected with insulin or saline; tibiae were harvested and processed for multiplexed stable-isotope phosphoproteomics and proteomics. Candidate phosphoproteins were functionally validated via CRISPR/morpholino zebrafish screens for bone development, and mechanistic studies used ChIP-seq, GRO-seq, co-immunoprecipitation, and in vitro kinase assays in primary and cultured osteoblasts.

Study Limitations

The in vivo model uses only male mice, limiting generalizability to females and humans. The study identifies gene-specific elongation defects but does not fully catalog which downstream genes drive the bone phenotype. Zebrafish screens, while informative, do not perfectly replicate mammalian bone biology.

Enjoyed this summary?

Get the latest longevity research delivered to your inbox every week.

Enter your email to subscribe: