Despite the known importance of COX-2/PGE2 on fracture healing and bone repair, this study is the first to define the role of the EP1 receptor in this process. Fractures in EP1−/− mice have accelerated chondrogenesis, chondrocyte maturation, endochondral bone formation, osteoblast differentiation, and bone remodeling. Gene expression studies from fracture callus tissues and bone marrow mesenchymal stem cell cultures demonstrated an enhanced rate of chondrocyte and osteoblast differentiation in EP1−/− mice. The effect appears to be directly due to the EP1 receptor because neither differences in EP2 or EP4 receptor expression nor activity, as measured by intracellular cAMP levels, were observed between wild-type and EP1−/− bone marrow cells. The findings show that the EP1 receptor is a negative regulator in bone repair. Since both the EP2 and EP4 receptors accelerate bone repair, the findings suggest that the overall effect of COX-2/PGE2 on the process of healing is determined by the relative activation of these various receptors.
Nonsteroidal anti-inflammatory drugs (NSAIDs), which are used widely as pain killers in fractures and other conditions, act by inhibiting COX-1 and COX-2. These drugs have been reported to impair the rate of fracture healing in humans.(2
) Prior work by others and us has clearly established a role for COX-2/PGE2
in fracture healing in animal models.(7
exerts anabolic and catabolic effects through its four receptor subtypes EP1 through EP4.(16
) Activation of EP2 and EP4 stimulates the production of cAMP through Gs, whereas activation of EP3 results in a decrease in cAMP levels through Gi, Gq, or Gs depending on the EP3 isoform.(16
) In contrast, the EP1 receptor is involved in regulating intracellular calcium levels. All four receptor subtypes are expressed in the fracture callus area (data not shown). Injection of EP2 or EP4 agonist immediately after fracture accelerates the healing process.(20
) EP2 and EP4 induce bone formation through the PKA pathway in contrast to EP3, which has been shown to inhibit bone formation in vitro.(15
) These experiments advance our understanding of the role of COX-2/PGE2
on fracture healing by showing that in addition to the EP2 and EP4 receptors, the EP1 receptor also plays a major role in fracture healing.
Fracture healing is a complex process in which bone injury results in the recruitment of mesenchymal stem cells to the site of injury, with subsequent proliferation and differentiation into bone-forming cells.(10
) Our findings show that the EP1 receptor acts as a negative regulator of this process because cell differentiation and fracture healing are accelerated in mice deficient in the EP1 receptor. Our in vivo findings in the femur fracture model established that EP1−/−
mice have accelerated chondrogenesis, chondrocyte maturation, and endochondral bone formation. Furthermore, EP1−/−
mice have enhanced osteoblast differentiation in the fracture callus.
The in vivo findings were confirmed in cultures of bone marrow stem cells in which osteoblast differentiation and mineralization of EP1−/−
bone marrow cells both were accelerated compared with cells from wild-type mice. The mesenchymal stem cell experiments were performed in marrow-derived cells that are only one component of the stem cell population that contributes to fracture healing.(36
) Since it is possible that other mesenchymal stem cell populations, including those from the periosteum, surrounding musculature, and vascular-derived stem cells could behave differently, the findings will need confirmation in other cell populations undergoing osteoblast differentiation. The finding that the expression levels of the transcription factors Runx2
were increased in cells and fracture calluses from EP1−/−
mice suggests an important role for the EP1 receptor in cell fate determination because both these transcription factors regulate osteoblastogenesis from mesenchymal precursor cells.(37
) As an inhibitor of osteoblast differentiation from mesenchymal stem cells, EP1 has an important role in the tissue response to bone injury.
An apparent paradox involving the COX-2/PGE2
signaling pathway is that while marked effects are observed on fracture repair, essentially no abnormalities are observed on skeletal development. While COX-2−/−
mice have a profound impairment in the differentiation of cartilage and bone from mesenchyme, limb development proceeds normally in COX-2−/−
) These observations suggest that while reparative processes recapitulate many of the events observed during development, the repair process depends on the activation of unique signaling pathways. In a similar manner, postnatal growth and development also have been shown to depend on unique signaling pathways. An example is Smad3−/−
mice, which have normal skeletons at birth and grow normally until approximately 3 weeks of age, at which time profound abnormalities of endochondral bone formation develop.(39
) Thus it is not surprising that the EP1−/−
mice develop normally. While recent work in our laboratory has established that COX-2−/−
mice have subtle differences in bone morphology and bone mass, to date, no differences in bone metabolism have been identified in the EP1−/−
mice used in this study.(31
There are two different models of EP1
gene deletion. The EP1−/−
mice initially described were produced by homologous recombination using a targeting vector that replaced a 671-base-pair sequence involving exon 2.(41
) However, since the EP1
receptor gene locus in mouse overlaps with the PKN
protein kinase gene on the antiparallel strand, exon 2 deletion also resulted in disruption of PKN
) PKN is activated by rho GTPase and by fatty acids, including arachidonate, and has a C-terminal region that is highly homologous to protein kinase C.(43
) Although the role of PKN in bone development has not been studied, it is possible that PKN regulates bone formation independent of EP1. In order to disrupt EP1 while sparing the PKN locus, Guan and colleagues deleted EP1
by introducing a premature in-frame stop codon into exon 2 by nucleotide substitution.(31
) Since the latter mice were used in this study, the findings depend completely on the EP1 receptor.
In addition to enhanced bone formation, our findings also suggest that fracture remodeling was accelerated in EP1−/− mice compared with wild-type mice. We performed cell culture experiments to determine whether deficiency of EP1 was associated with the potential for accelerated osteoclast formation through a cell-autonomous process. In cell culture, spleen cells from EP1−/−and wild-type mice had similar rates of osteoclastogenesis following treatment with M-CSF and RANKL. In contrast, the EP1−/− fracture calluses and cultures of EP1−/− bone marrow stem cells had both accelerated and increased expression levels of RANKL. Thus the data support a primary role for enhanced osteoblast differentiation and RANKL expression in the enhanced osteoclastogenesis and remodeling observed in the fractures in EP1−/− mice.
Consistent with these histologic and molecular observations, µCT analysis of fracture callus bone volume and mineral density (Supplemental Fig. 1
) demonstrated an accelerated reduction in callus volume in EP1−/−
mice at 21 days, with a trend of increased mineral density, also suggesting accelerated remodeling of the fracture callus.
Biomechanical testing at 28 days following the fracture confirmed that the fractures in EP1−/−
mice healed more efficiently and with increased torsional strength and rigidity (Supplemental Fig. 1
). The ultimate torque and torsional rigidity both were increased significantly in the EP1−/−
fractures. This confirms the fact that although EP1−/−
fractures have reduced callus area compared with wild-type mice late in the healing process, the fractures have enhanced biomechanical strength owing to increased bone remodeling. Since torsional strength in torsion is inversely related to the cross-sectional area of the callus,(44
) the increased strength despite the reduced callus area and volume in the EP1−/−
fracture callus is indicative of higher-quality bone.
Compensatory induction of EP2 or EP4 signaling is a possible explanation for why EP1−/− mice exhibit increased bone repair. However, no differences were found between wild-type and EP1−/− bone marrow cells in either EP2 or EP4 receptor level or intracellular cAMP activity following stimulation with PGE2. This suggests that the phenotype of accelerated fracture healing in EP1−/− mice is independent of changes in EP2 or EP4 signaling. The findings support a primary role for the EP1 receptor as a negative regulator of bone repair.
All together, these experiments show that the EP1 receptor is a negative regulator of fracture healing and demonstrate that the overall effect of COX-2/PGE2 signaling in fracture healing depends on the complex integration of signals from the various EP receptors. Our results suggest that inhibition of the EP1 receptor may increase fracture healing and support the importance of further studies to define the expression, regulation, and effects of this receptor in mesenchymal stem cells, chondrocytes, and osteoblasts during reparative processes.