Hypothesis

With age, human osteoblasts show declining O‑glycosylation of osteocalcin at the position analogous to mouse Ser8, shortening its half‑life and shifting its signaling at the adrenal GPR158 receptor from Gαs‑driven cAMP production toward β‑arrestin‑biased pathways. This bias reduces glucocorticoid synthesis while weakening the exercise‑induced IL‑6/osteocalcin feed‑forward loop that supports muscle performance. The loss of stable, O‑glycosylated osteocalcin also triggers osteocytes to release more sclerostin, which acts as a proinflammatory cytokine that further suppresses adrenal steroidogenesis.

Mechanistic Basis

• O‑glycosylation loss: In mice, redundant GalNAc‑Ts (T1/2/3/6) add O‑GlcNAc to Ser8, prolonging osteocalcin half‑life independent of carboxylation[3]. Humans lack the equivalent site and rely on carboxylation for stability; aging lowers γ‑glutamyl carboxylase activity, decreasing the pool of undercarboxylated, active osteocalcin[2].
• Receptor bias: GPR158 can signal via Gαs (cAMP) or β‑arrestin scaffolds. Short‑lived, poorly modified osteocalcin favors β‑arrestin recruitment, which engages MAPK phosphatases that blunt STAR expression, cutting cortisol output[1].
• Muscle‑bone loop disruption: Exercise‑induced IL‑6 stimulates osteoclast resorption to release osteocalcin[5]. With less circulating osteocalcin, IL‑6 signaling falls, lowering muscle IGF‑1 and impairing performance[2].
• Sclerostin rise: Osteocytes sense low mechanical loading and low osteocalcin, upregulating sclerostin secretion. Secreted sclerostin binds LRP5/6 on adrenal cortical cells, activating NF‑κB and raising IL‑1β, which suppresses CYP11B1 expression[4].

Testable Predictions

1. Older adults will exhibit a lower ratio of O‑glycosylated to total osteocalcin in plasma than young adults, measurable by lectin (VVA) enrichment followed by mass spectrometry.
2. Pharmacologic O‑glycosylation of recombinant human osteocalcin will restore cAMP signaling in H295R adrenal cells and increase cortisol secretion, whereas the unmodified form will preferentially activate β‑arrestin pathways.
3. Inhibiting β‑arrestin recruitment with a biased antagonist (e.g., a pepducin) will rescue cortisol production in aged mouse adrenal explants despite low endogenous osteocalcin.
4. Neutralizing sclerostin with antibodies will blunt the age‑related rise in adrenal IL‑1β and strengthen the correlation between circulating osteocalcin and muscle strength in elderly cohorts.

Experimental Design

• Human cohort: Enroll 30 young (20‑30 y) and 30 older (65‑75 y) participants. Collect fasting plasma, perform VVA lectin pull‑down for O‑GlcNAc, quantify osteocalcin by ELISA and targeted MS. Correlate the O‑glycosylated/total ratio with morning cortisol, post‑exercise IL‑6, and hand‑grip strength.
• In vitro: Treat H295R cells with recombinant human osteocalcin (unmodified) or enzymatically O‑glycosylated osteocalcin. Measure cAMP (ELISA), β‑arrestin recruitment (BRET), CYP11B1 mRNA (qPCR), and cortisol (RIA). Use siRNA against β‑arrestin2 to confirm bias.
• Animal study: Aged C57BL/6 mice receive weekly injections of O‑glycosylated osteocalcin or vehicle for 4 weeks. Assess adrenal weight, zona fasciculata thickness, serum cortisol, treadmill endurance, and serum sclerostin. Include a group receiving a β‑arrestin biased antagonist.
• Intervention arm: In a parallel set, give aged mice anti‑sclerostin antibody and examine whether osteocalcin‑cortisol coupling is restored.

Falsifiability

If O‑glycosylated osteocalcin fails to raise cAMP or cortisol in aged human adrenal cells, or if sclerostin neutralization does not improve the osteocalcin‑muscle relationship, the hypothesis would be falsified. Conversely, supportive data would link age‑dependent PTM loss to adrenal‑muscle uncoupling via receptor bias and sclerostin‑driven inflammation.