NF-κB RelB negatively regulates osteoblast differentiation and bone formation

Abstract

RelA-mediated NF-κB canonical signaling promotes mesenchymal progenitor cell (MPC) proliferation, but inhibits differentiation of mature osteoblasts (OBs) and thus negatively regulates bone formation. Previous studies suggest that NF-κB RelB may also negatively regulate bone formation through noncanonical signaling, but they involved a complex knockout mouse model, and the molecular mechanisms involved were not investigated. Here, we report that RelB(-/-) mice develop age-related increased trabecular bone mass associated with increased bone formation. RelB(-/-) bone marrow stromal cells expanded faster in vitro and have enhanced OB differentiation associated with increased expression of the osteoblastogenic transcription factor, Runt-related transcription factor 2 (Runx2). In addition, RelB directly targeted the Runx2 promoter to inhibit its activation. Importantly, RelB(-/-) bone-derived MPCs formed bone more rapidly than wild-type cells after they were injected into a murine tibial bone defect model. Our findings indicate that RelB negatively regulates bone mass as mice age and limits bone formation in healing bone defects, suggesting that inhibition of RelB could reduce age-related bone loss and enhance bone repair.

Keywords: BONE FORMATION; MESENCHYMAL PROGENITOR CELLS; NF-κB; OSTEOBLASTS; RELB.

Fig. 1. Trabecular bone mass increases in RelB−/− mice as they age

(A) H&E-stained sections of tibiae from WT and RelB−/− mice and histomorphometric data showing trabecular bone volume (BV/TV) in the tibial metaphyses (area corresponding to the position of the white bar) and proximal diaphysis (vertical green bar) of 4–14-wk-old mice. n=5–8 mice/group. Metaphysic bone (corresponding to ROI-1 of μCT) was defined as a 0.3 mm-region along with the long-axis of tibia beginning at 0.15 mm apart from growth plate. Proximal diaphysis (corresponding to ROI-2 of μCT) was defined as a 0.6 mm-region under ROI-1. (B) Representative μCT scans and bone volume (BV/TV), trabecular number (Tb.N), trabecular thickness (Tb.Th) and trabecular separation (Tb.Sp) in the metaphysis (Region of Interest-1 (ROI-1) and in the proximal part of the diaphysis (ROI-2) of tibiae from 10–14wk-old mice. n=5–8 mice/group. (C) μCT scans and data from 4^th lumbar vertebrae from 10–14-wk-old mice. n=10 mice/group. (D) μCT scans and data from cortical bone at the junction between RO-1 and -2 from mice in (B). All groups contained male and female mice. * p<0.05; ** p<0.01 vs. control.

Fig 2. RelB−/− mice have a transient increase in bone formation

(A) Histology and histomorphometric analysis of OBs (green arrows) and OB surface (OB.S/BS) in the tibiae of 4-wk-old RelB−/− mice and WT littermates (n=7/group). (B) Double calcein labeling (arrows) in trabeculae in proximal tibial diaphyses corresponding to ROI-2 in Fig. 1 and analysis of dynamic parameters of bone formation: single labeled surface (sLS/BS), double labeled surface (dLS/BS), mineralization surface (MS/BS), mineral apposition rate (MAR) and bone formation rate (BFR) in 4-wk-old RelB−/− and WT mice (n=8/group). (C) Analysis of dynamic parameters of bone formation listed in (B) and OB surface in proximal tibial diaphyses of 7–8-week-old RelB−/− and WT mice (n=5–6/group). (D) Serum osteocalcin (OCal) levels from 4–5- (n=9) and 8–9-week-old (n=8) male and female WT and RelB−/− mice tested by ELISA.

Fig 3. RelB−/− mice have enhanced stromal cell proliferation and OB differentiation in vitro

(A) FACS analysis showing the % of CD45⁻CD105⁺ MPCs in freshly isolated BM cells from 2 month-old WT and RelB−/− mice and BM stromal cells from the mice after being cultured for 7 days with growth-inducing medium. (B) Equal numbers of freshly isolated BM cells from 2 month-old RelB−/− and WT mice were cultured in OB differentiation medium containing 25μg/ml L-ascorbic acid (Vit C) and 5mM β-glycerophosphate (β-GP) for 7 and 12d. ALP⁺ cells colonies (upper panel) and mineralized nodules (lower panel) were evaluated after ALP and von Kossa staining. (C) Expression of ALP and osteocalcin (OCal) was tested using Real-time PCR in BM stromal cells and in differentiating OBs induced by Vit C and β-GP after 6 and 10 days of culture. (D) BM stromal cells generated from WT and RelB−/− mouse were re-seeded in 12-well plate. After the cells were sub-confluent, they were induced for OB differentiation for 5 days and stained for ALP activity. (E) Calvarial pre-OBs generated from 7-day-old pups were cultured in OB differentiation medium for 5 and 10 days and stained for ALP activity. (F) Calvarial pre-OBs were cultured in OB differentiation medium for the indicated times, total RNA was extracted, and expression of ALP, osteocalcin and Runx2 mRNA was tested by real-time PCR. All mice in in vitro experiments were 1.5–2-months-old males or females except for those used for calvarial pre-osteoblasts in (E&F); 3 wells per group. * p<0.05 vs. control.

Fig. 4. RelB−/− bone-derived mesenchymal progenitor cells (bMPCs) have enhanced proliferation and OB differentiation

(A) bMPCs generated from 8 week-old RelB−/− mice and WT littermates were analyzed by FACS using CD45, CD105 and Sca-1 antibodies. (B) 2×10⁴ bMPCs from WT and RelB−/− mice were seeded in 12-well plates for indicated times. The cells were fixed with 10% formalin followed by H&E staining to count cell numbers (4 wells/time-point). (C) Cultured bMPCs at 60–70% confluent were collected and incubated with DAPI (1μg/ml). The cell cycle was analyzed by FACS, and the data presented were from 3 pair of mice. (D) bMPCs from RelB−/− and WT littermates were cultured in OB differentiation medium for 5d and stained for ALP activity followed by eosin counterstaining. ALP+ and total cell numbers were counted and the ALP+/total cell ratio was calculated. (E) 1×10⁴ bMPCs were cultured in 12-well plates for 5 days to sub-confluence and OB differentiation medium was added for 21 days when nodules had formed. Von Kossa staining was performed to quantify nodule area. (F) bMPCs cultured in 60 mm dish were induced for OB differentiation for the indicated times. Expression levels of ALP and osteocalcin (OCal) were tested using Real-time PCR. All mice in in vitro experiments were 1.5–2-months-old males or females; 3–4 wells per group. * p<0.05; ** p<0.01 vs. control.

Fig. 5. RelB directly targets the Runx2 promoter and inhibits Runx2 expression

(A) mRNA (left panel) and protein (right panel) expression of Runx2 were tested by Real-time PCR and Western blot from RelB−/− and WT BM stromal cells cultured with OB differentiation medium for 7d. (B) Construction scheme of mouse Runx2 promoter luciferase (Luc) reporter. The 2-kb Runx2 promoter constructs contain 2 putative κB binding sites. (C) The 2-kb Runx2 promoter Luc reporter was co-transfected with RelB plasmid into C2C12 cells and the relative Luc activity was tested. (D) Site-directed mutagenesis of both κB binding sites 1 and 2 in the 2 kb Runx2 promoter was performed by deleting “ATC” and “ACA”. A luciferase activity assay was performed using C2C12 cells that were co-transfected with a RelB plasmid and the mutated Runx2 reporter. (E) ChIP assays were carried out using an anti-RelB or control IgG antibody on sheared chromatin from WT and RelB−/− MPCs. Immunoprecipitated DNA was analyzed by qPCR using primers covering either NF-κB binding site-1 (left panel) or 2 (middle panel) in the Runx2 promoter region or a pair of un-related primers (right panel) designed in the region that is 3 kb apart from the κB binding sites. Results are expressed as fold-enrichment compared with IgG normalized to input. (F) bMPCs from WT and RelB−/− mice were transfected with a scrambled or Runx2 mouse shRNA sequence for 2 days followed by puromycin selection to kill the uninfected cells. The cells were treated with OB differentiation medium for 5d and stained for ALP activity to measure ALP+ cells (left and middle panel), and mRNA expression of Runx2 in these cells was tested by real-time PCR (right panel). All mice in in vitro experiments were 1.5 to 2 months-old. * p<0.05; ** p<0.01 vs. control.

Fig. 6. RelB−/− MPCs induce bone formation and repair in tibial bone defects more rapidly than WT MPCs

Bilateral 2×5mm cortical defects were made in the anterior proximal tibiae of SCID mice and filled with decalcified bovine bone matrix. 5×10⁵ bMPCs from 7-wk-old RelB−/− and WT mice were injected into the bone matrix in the left and right tibial defects, respectively. The mice were sacrificed 4 and 8 weeks post-surgery and the volume of new bone (BV/TV, expressed as a percentage of the total defect volume) formed in the defects was measured by μCT (A) followed by histomorphometric analysis of the area of newly formed trabecular bone (Trab) and fibrosis tissue (Fib, black arrows) containing spindle-shaped fibroblast-like cells observed in decalcified H&E-stained sections of the bones (B). (C) Representative images of bone in the defect sites 8 weeks post-surgery, and the extent of new bone (with viable osteocytes and covered with a periosteum-like membrane, green arrow) formed on the DBM was quantified and expressed as a % of the total length of implanted DBM. The black arrow shows fibrous tissue covering the acellular DBM. n=5/group. * p<0.05 vs. control.