BMP Signaling Mediated by BMPR1A in Osteoclasts Negatively Regulates Osteoblast Mineralization Through Suppression of Cx43

Abstract

Osteoblasts and osteoclasts are well orchestrated through different mechanisms of communication during bone remodeling. Previously, we found that osteoclast-specific disruption of one of the BMP receptors, Bmpr1a, results in increased osteoblastic bone formation in mice. We hypothesized that BMPR1A signaling in osteoclasts regulates production of either membrane bound proteins or secreted molecules that regulated osteoblast differentiation. In our current study, we co-cultured wild-type osteoblasts with either control osteoclasts or osteoclasts lacking BMPR1A signaling activity. We found that loss of Bmpr1a in osteoclasts promoted osteoblast mineralization in vitro. Further, we found that the expression of Cx43/Gja1 in the mutant osteoclasts was increased, which encoded for one of the gap junction proteins connexin 43/gap junction alpha 1. Knockdown of Gja1 in the mutant osteoclasts for Bmpr1a reduced osteoblastic mineralization when co-cultured. Our findings suggest that GJA1 may be one of the downstream targets of BMPR1A signaling in osteoclasts that mediates osteoclast-osteoblast communication during bone remodeling. J. Cell. Biochem. 118: 605-614, 2017. © 2016 Wiley Periodicals, Inc.

Keywords: BMPR1A; CONNEXIN 43/GJA1; GAP JUNCTION; OSTEOBLAST; OSTEOCLAST.

Fig. 1

Bmpr1a was efficiently disrupted in undifferentiated and differentiated osteoclasts. (A) Breeding was set up between the mice indicated. Bone marrow cells were harvested and Cre activity was induced during culture. (B) Levels of recombination were evaluated by genomic PCR (fx1 and fx4) and quantitative genomic PCR for exon 4. (C) BMMCs were differentiated using 20 ng/ml M-CSF and 50 ng/ml RANKL and treated with TM for different days, followed by DNA extraction. Recombination levels were assessed by genomic PCR using primers fx1 and fx4. A 180-bp amplicon indicated Cre-dependent recombination, whereas an over 1.5-kb amplicon indicated no recombination. (D) BMMCs were treated with or without TM for 4 days in the presence of M-CSF only (undifferentiated) or differentiation medium (M-CSF and RANKL) (differentiated), followed by DNA extraction. Deletion efficiency of Bmpr1a locus was quantified by genomic real-time quantitative PCR. Levels of the presence of exon 4 were compared with the same stock of cells that were not treated with TM. (E) Control and mutant BMMCs were differentiated using M-CSF and RANKL for 4 days, and Cre activity was induced using tamoxifen at the same time. Then cells were stimulated with or without 100 ng/ml BMP-2 for 30 min, followed by cell lysis. P-SMAD1/5/9 level was detected by Western blot, and GAPDH was used as a loading control. (F) The scheme for osteoclast-osteoblast co-culture. BMMCs were expanded using M-CSF for 4 days and differentiated using M-CSF and RANKL for another 4 days. A total of 1 × 10⁵ osteoclasts/well were added to osteoblast culture every 3 days in osteogenic differentiation medium. cont, control, genotyped as Bmpr1a fx/+, Ubi-CreER™ (+)/(−); mut, mutant, genotyped as Bmpr1a fx/−, Ubi-CreER™ (+)/(−); het, heterozygous; cKO, conditional knockout; bp, base pair; TM, (Z)-4-hydroxytamoxifen; ex, exon; BMMC, bone marrow mononuclear cell; OC, osteoclast; OB, osteoblast. *P<0.05; **P<0.01.

Fig.2

Mutant osteoclasts deficient for Bmpr1a promoted osteoblast mineralization. (A)After co-culture for 14days, cells were fix and stained for alizarin red. Comparison was done in triplicate. (B) The number of mineralized nodules was counted. The percentage of nodule area was measured and calculated. The staining was dissolved and the optical density was measured. Bar=2 mm. cont, control; mut, mutant; OC, osteoclast; OB, osteoblast. *P<0.05; **P<0.01.

Fig. 3

Osteoclastic expression profiles of candidate genes mediating osteoclast–osteoblast communication were changed. BMMCs were differentiated for 4 days, then cells were stimulated with or without BMP-2 for 6h followed by RNA extraction. Quantitative RT–PCR was performed. Gene expression levels in control osteoclasts without BMP treatment were calculated as 1. cont, control; mut, mutant; OC, osteoclast. *P<0.05; **P<0.01.

Fig. 4

Cx43/Gja1 was efficiently knocked down in osteoclasts by adenovirus treatment. (A) BMMCs were differentiated using M–CSF and RANKL for 4 days, then cells were stimulated with BMP-2 for different periods followed by RNA extraction. The expression of Cx43/Gja1 was evaluated by quantitative RT–PCR. (B) Representative images of osteoclasts 48 h after transfection, showing GFP signals (lower panel) and the corresponding bright field (upper panel). (C) The expression of Cx43/Gja1 after transfection. (D) Protein level of Cx43/GJA1 after transfection. (E) The scheme for co-culture when osteoclasts were treated with adenovirus. BMMCs were expanded using M–CSF for 4 days and differentiated using M–CSF and RANKL for another 4 days. Two days after differentiation, osteoclasts were transfected with adenovirus. A total of 1×10⁵ osteoclasts/well were added to osteoblast culture every 3 days in osteogenic differentiation medium. cont, control; mut, mutant; OC, osteoclast; BF, bright field; GFP, green fluorescent protein; Ad, adenovirus; Scrb, scrambled. *P<0.05; **P<0.01.

Fig. 5

Mutant osteoclasts promoted osteoblast mineralization in the co-culture, and knockdown of Cx43/Gja1 in mutant osteoclasts decreased osteoblast mineralization. (A) Cells were stained with alizarin red. Comparison was done in triplicate. (B) The number of mineralized nodules was counted. The percentage of nodule area was measured and calculated. The staining was dissolved and the optical density was measured. cont, control; mut, mutant; OC, osteoclast; OB, osteoblast; Ad, adenovirus; Scrb, scrambled. *P<0.05; **P<0.01.