Age‑dependent osteoblast carboxylase dysfunction drives divergent hydroxyapatite crystal remodeling in bone versus tooth via localized vitamin K recycling

Amara Afriyie· Mar 23, 2026

Mechanism: Aging reduces NAD+ in osteoblasts, impairing vitamin K recycling and osteocalcin carboxylation, leading to brittle bone crystals, while odontoblasts maintain function. Readout: Readout: NAD+ boosters improve bone osteocalcin carboxylation and reduce bone crystal size from ~70nm to ~45nm.

Hypothesis

Aging impairs the vitamin K cycle in osteoblasts but not in odontoblasts, producing tissue‑specific carboxylation states of osteocalcin that oppositely regulates hydroxyapatite crystal growth (6, 2).

Mechanistic Rationale

Osteocalcin must be γ‑glutamyl carboxylated to bind hydroxyapatite tightly and restrain crystal elongation (6).
In bone, aged osteoblasts show ↓ NAD⁺, which reduces the activity of NAD⁺‑dependent vitamin K epoxide reductase (VKOR) homologs, limiting recycling of vitamin K hydroquinone needed for carboxylation (7).
Consequently, osteocalcin remains undercarboxylated (ucOC), loses its inhibitory effect, and allows uncontrolled crystal growth → larger, more brittle hydroxyapatite (1, 5).
In teeth, odontoblasts maintain NAD⁺ levels and express a distinct VKOR isoform (VKORC1L1) that is resistant to age‑related decline, preserving carboxylated osteocalcin (cOC) that continues to suppress crystal growth, yielding smaller, more crystalline apatite with age (4, 8).

Testable Predictions

Rescue experiment – Chronic NAD⁺ booster (e.g., nicotinamide riboside) or menadione supplementation in aged mice will:
- Increase serum cOC/ucOC ratio in bone marrow aspirates (1).
- Reduce mean hydroxyapatite length in femoral cortical bone from ~70 nm to ~45 nm (measured by TEM) without altering incisor crystal dimensions (3, 4).
Cell‑specific inhibition – Conditional knockout of VKORC1L1 in odontoblasts of young mice will:
- Raise ucOC levels in dentin extracts.
- Increase incisor hydroxyapatite length from ~30 nm to ~50 nm, mimicking the aged‑bone phenotype (4, 6).
Correlation in humans – In post‑menopausal women, plasma NAD⁺ levels will positively correlate with serum cOC (r > 0.4, p < 0.01) and inversely correlate with femoral cortical crystal size (derived from high‑resolution pQCT), while no such relationship will be observed for dentin crystal size estimated via micro‑Raman spectroscopy of extracted teeth (1, 3, 8).

Falsifiability

If NAD⁺ supplementation fails to raise the cOC/ucOC ratio or does not change bone crystal size, or if odontoblast‑specific VKORC1L1 loss does not alter dentin crystal dimensions, the hypothesis is refuted. Likewise, a lack of correlation between NAD⁺ and cOC in human cohorts would contradict the proposed mechanistic link.

Experimental Approach (brief)

Use aged (24‑month) C57BL/6 mice treated with NR (400 mg/kg/day) or vehicle for 12 weeks.
Harvest femurs and incisors; quantify osteocalcin isoforms by ELISA; assess crystal size via transmission electron microscopy and synchrotron XRD.
Generate Odont‑Cre;Vkorc1l1^fl/fl mice for odontoblast‑specific VKORC1L1 deletion; analyze dentin histology and micro‑hardness.
Human pilot study: recruit 60 post‑menopausal women; measure plasma NAD⁺, serum cOC/ucOC, femoral cortical crystal size (HR‑pQCT), and dentin crystal size (micro‑Raman on exfoliated teeth).

Mechanistic Rationale

Osteocalcin must be γ‑glutamyl carboxylated to bind hydroxyapatite tightly and restrain crystal elongation (6).

In bone, aged osteoblasts show ↓ NAD⁺, which reduces the activity of NAD⁺‑dependent vitamin K epoxide reductase (VKOR) homologs, limiting recycling of vitamin K hydroquinone needed for carboxylation (7).

Consequently, osteocalcin remains undercarboxylated (ucOC), loses its inhibitory effect, and allows uncontrolled crystal growth → larger, more brittle hydroxyapatite (1, 5).

In teeth, odontoblasts maintain NAD⁺ levels and express a distinct VKOR isoform (VKORC1L1) that is resistant to age‑related decline, preserving carboxylated osteocalcin (cOC) that continues to suppress crystal growth, yielding smaller, more crystalline apatite with age (4, 8).

Testable Predictions

Rescue experiment – Chronic NAD⁺ booster (e.g., nicotinamide riboside) or menadione supplementation in aged mice will:

Increase serum cOC/ucOC ratio in bone marrow aspirates (1).
Reduce mean hydroxyapatite length in femoral cortical bone from ~70 nm to ~45 nm (measured by TEM) without altering incisor crystal dimensions (3, 4).

Cell‑specific inhibition – Conditional knockout of VKORC1L1 in odontoblasts of young mice will:

Raise ucOC levels in dentin extracts.
Increase incisor hydroxyapatite length from ~30 nm to ~50 nm, mimicking the aged‑bone phenotype (4, 6).

Correlation in humans – In post‑menopausal women, plasma NAD⁺ levels will positively correlate with serum cOC (r > 0.4, p < 0.01) and inversely correlate with femoral cortical crystal size (derived from high‑resolution pQCT), while no such relationship will be observed for dentin crystal size estimated via micro‑Raman spectroscopy of extracted teeth (1, 3, 8).

Falsifiability

Experimental Approach (brief)

Use aged (24‑month) C57BL/6 mice treated with NR (400 mg/kg/day) or vehicle for 12 weeks.

Harvest femurs and incisors; quantify osteocalcin isoforms by ELISA; assess crystal size via transmission electron microscopy and synchrotron XRD.

Generate Odont‑Cre;Vkorc1l1^fl/fl mice for odontoblast‑specific VKORC1L1 deletion; analyze dentin histology and micro‑hardness.

Human pilot study: recruit 60 post‑menopausal women; measure plasma NAD⁺, serum cOC/ucOC, femoral cortical crystal size (HR‑pQCT), and dentin crystal size (micro‑Raman on exfoliated teeth).