Bone formation during injury repair is a multi-step series of events modulated by an integrated cascade of gene expression that initially supports the proliferation stage. The later mineralization stage is associated with the sequential expression of genes that support biosynthesis, organization and mineralization of the bone extracellular matrix. Mineralization requires expression of structural proteins such as collagen type I, osteocalcin, as well as noggin and runx2 which aid in mineralization . Transcriptional control defines the regulatory events necessary for both stages of bone formation . There is a general consensus that during injury GFs are released from the wounded bone matrix and promote healing . In this study, we have documented the relative efficiency of bone growth factors FGF-2, TGFβ, and PGE2 markedly enhanced the synthesis of the total protein content of the dishes (Table 1)
Rate of proliferation was dependent on the specific GF. FGF-2, TGFβ and PGE2 significantly promote growth, with FGF-2 having the highest efficacy and the lowest dose. FGF-2 produced a distinct pattern of gene expression. FGF-2 down regulates genes associated with mineralization while it induces genes associated with proliferation and angiogenesis, a finding supported by observations of others . Since cox-2 had a 27-fold induction by FGF-2, we examined the effect of the COX-2 product, PGE2 on proliferation. We found that PGE2 increased DNA synthesis by 3.3 fold significantly higher than TGFβ, IGF-1, PDGF, suggesting that its induction by FGF-2 helps complete the FGF-2 full induction of osteoblast growth. These data also suggest that FGF-2 may be an important regulator of migration, angiogenesis and proliferation during the first stage of healing a critical defect since it induces mmp3, vegfa and vegfr1 expression. In data not shown, FGF-2 had no effect on expression of mmp-1. Moreover, FGF-2 induced its own message as well as TGFβ, but significantly reduced expression of BMP-2, osteocalcin, noggin, runx2, collagen type I and IGF-1, genes which are associated with mineralization.
As described by others, bone formation is divided into two phases, proliferation and mineralization [2, 26–29]. These two stages are the result of a specific sequential regulation of gene expression from the early phase of osteoblast proliferation to the final steps of mineralization. Once the cells start mineralizing, cell division and DNA synthesis dramatically slow down and eventually cease. When an injury occurs in mineralized tissue, GFs like FGF-2 are released and start a new proliferation stage to heal the defect. The increase in cell replication in a mineralizing cell likely represents a de-differentiation from the mineralizing phase to the growing phase, and increases expression of GFs most likely induce proliferation. Treatment of the mineralized defect model with FGF-2 resulted in gene expression that corresponds to de-differentiation (e.g. significant increases in growth related genes egf-1, fgf-2, cox-2, TGFβ, vegfA, vegfr and mmp3 and down-regulation of mineralizing related genes) . Vegf and vegfr1 are primary regulators of angiogenesis, while MMP3 is thought to play a major role on cell behaviors such as proliferation and migration  which may explain the ability of the FGF-2 to enable the cultured cells to fill the defect void efficiently. The fact that FGF-2 induces its own expression suggests that after injury, the FGF-2 released from the wound matrix could promote it's own expression, making it a feed-forward loop.
Although Figures 1 and 2 demonstrate the relative FGF-2 regulation and sequential expression of growth, angiogenic and chemotactic genes and depresses expression of mineralization-related genes, these figures do not tell us the relative abundance of the genes. In Table 2, we determined the relative abundance of genes in three groups after 24 hours; with or without treatment with FGF-2 or BMP-2. FGF-2 caused a significant increase in abundance of genes associated with proliferation, chemotaxis and angiogenesis. Moreover, the addition of FGF-2 to the mineralized wounded cultures caused a marked decrease in abundance of col1a1 as well as fn, igf-1, noggin, oc, bmp-2 and alp message. In the early stages of mineralization, the major protein (greater than 20%) synthesized by the osteoblast is collagen, however collagen is not a major component of the proliferating cell, suggesting that FGF-2 stimulates proliferation partly through its ability to drastically reduce the relative abundance of a majority of the mineralizing-associated genes. The dramatic reduction of IGF-1 by FGF-2 suggests a role for IGF-1 in mineralization, this is in agreement with findings of others that demonstrated IGF-1 to be a major factor in bone mineralization [31–33] using the IGF-1 null mouse. In contrast, in cells treated with BMP-2, the relative abundance of col1a1, fn, oc, and tgfβ were dramatically induced while BMP-2 had no significant effect on genes related to growth, angiogenesis or chemotaxis. These data suggest that BMP-2 may be the best GF to use for the mineralization stage but not the proliferation stage of bone formation. This finding may help explain studies by others  who discovered that a delayed administration of BMP-2 to a fracture resulted in better repair of critical size defects. It is likely that the delay of BMP-2 treatment allowed a longer period of proliferation prior to BMP-2 promotion of mineralization. Our findings in Table 2, 3 and Figure 3 support the hypothesis that FGF-2 and BMP-2 are required at different stages of bone repair.