BMP3: to be or not to be a BMP

Background: Bone morphogenetic proteins (BMPs) are osteogenic but also have diverse functions during development. BMP3 is a major component of osteogenin, which has osteogenic activity. However, recombinant BMP3 (rhBMP3) has no apparent osteogenic function, raising the possibility that BMP3 has no bone-inducing activity or that the recombinant material is not properly processed. To resolve this apparent discrepancy, we utilized a retroviral system to study the effects of BMP3 in vitro. In addition, we generated Bmp3-deficient mice to elucidate the function of BMP3 in vivo.

Methods: Retroviral as well as mammalian expression constructs were utilized to express BMP3 and to create BMP3 conditioned medium. Alkaline Phosphatase (ALP) activity and transcriptional response assays were used to monitor the ability of BMP3 to elicit either a BMP-like or a transforming growth factor beta (TGF-beta)/activin-like response in osteoblastic cell lines. Finally, mice deficient in BMP3 were generated to investigate BMP3 function in vivo.

Results: BMP3 was unable to induce an osteogenic response in W-20-17, MC3T3-E1, or C3H10T1/2 cells, although all three cell lines were responsive to BMP2. However, BMP3 inhibited responsiveness to BMP2 in these assays, suggesting that BMP3 antagonizes BMP2 signaling. This inhibition did not occur through inhibition of binding of BMP2 to its receptors. BMP3 activated the TGF-beta/activin-pathway in these cells, suggesting that BMP3 exerts its inhibiting effects by activating a signaling pathway that antagonizes the BMP pathway. To examine the potential functional consequences of BMP3 action, Bmp3-/- mice, which lack BMP3, were generated. On an outbred genetic background, Bmp3-/- mice are viable and show no obvious skeletal phenotype as embryos or neonates. However, adult mice exhibit twice as much trabecular bone as do their wild-type littermates. This observation is consistent with our in vitro observations suggesting that BMP3 is an inhibitor of osteogenesis in vitro and of bone formation in vivo.

Conclusions: BMP3 is an inhibitor of osteogenic BMPs and can signal through a TGF-beta/activin pathway. The ability of BMP3 to antagonize BMP2 activity may thus be a consequence of competition for signaling components common to TGF-beta/activin and BMP pathways. BMP3, the most abundant BMP in demineralized bone, may therefore play an essential role as a modulator of the activity of osteogenic BMPs in vivo.

Clinical relevance: Therapies to accelerate bone healing usually utilize administration of exogenous BMP either in recombinant form or by gene therapy approaches. It is conceivable that the potency of osteogenic BMPs would be increased by inhibiting the activation of antagonistic signaling pathways or by increasing levels of rate-limiting signaling components shared by both BMP and TGF-beta/activin pathways.