The Role of Prostaglandin E 2 on Osteoblast Proliferation Induced by Hydroxyapatite

Objective: Prostaglandin E 2 (PGE 2 ) plays a crucial role in regulating bone cell differentiation and proliferation. The aim of the present study was to determine whether PGE 2 may regulate osteoblast proliferation induced by hydroxyapatite. Materials and Methods: Osteoblasts (HOS cell line) pretreated with cyclooxygenase (COX) inhibitors (indomethacin, aspirin and nimesulide) were then cultured. The cells were also pre-treated with or without nimesulide and then cultured with PGE 2 . The cell cultures were also treated with SQ22536 (adenylyl cyclase inhibitor) and added with Db-cAMP (cAMP analog), and/or PGE 2 . KT5720 [protein kinase A (PKA) inhibitor . ], Db-cAMP and/or forskolin (adenylyl cyclase activator)-treated cultures were used to assess the role of PKA. The role of EP2 and/or EP4 was determined by using EP2 antagonist (PF-04418948) and EP4 antagonist (L-161,982) with PGE 2 . All cells were cultured with or without hydroxyapatite. The levels of PGE 2 and cAMP were detected from the culture supernantants and the cell proliferation was assessed colorimetrically. Results: Nimesulide and indomethacin but not aspirin suppressed partially the cell proliferation but fully PGE 2 production. PGE 2 abrogated nimesulide-mediated suppression of cell proliferation. The cell proliferation was enhanced by low but suppressed by high concentration of PGE 2 . Moreover, the SQ22536-mediated suppression of cell proliferation was abolished by Db-cAMP but not PGE 2 . Conversely, PGE 2 , Db-cAMP or forskolin failed to eliminate KT5720-mediated suppression of cell proliferation. The effect of PGE 2 on cell proliferation and cAMP levels was mediated predominantly via EP2 and to a lesser extent, EP4. The results of the controls for all experiments were significantly lower than hy-droxyapatite-stimulated cell cultures. Conclusion: These results suggest that PGE 2, acting via a COX-2-, cAMP-PKA- and both EP2 and EP4-dependent pathway may partially regulate hydroxyapatite-induced human osteoblasts in an autocrine fashion.


Introduction
Prostaglandin E 2 (PGE 2 ), a prostanoid produced by the action of cyclooxygenases (COX-1 and COX-2) on arachidonic acid, is known as a potent regulator on bone remodeling, since it has ability to activate osteoclast and osteoblast differentiation and proliferation by binding with its receptors, i.e., EP1, EP2, EP3, and EP4 [1]. It seems plausible that all four PGE 2 receptors may play a regulatory role in the dynamic bone remodeling. EP2 and/or EP4 bound by PGE 2 or its analogs are shown to be the receptors that activate the cAMP-PKA pathway leading to COX-2 activities and bone formation [2] [3]. On the other hand, binding PGE 2 on cell surface EP1 or EP3 may result in the reduction of bone formation [3] [4], although others demonstrated that EP2 and/or EP4 may also play a role in bone resorption [4] [5].
Bone remodeling at the site of implantation seems also to be under regulation of PGE 2 . Regardless of the implant surface topography, PGE 2 suppressed the conical Wnt signaling pathway of osteoblasts leading to downregulation of cell differentiation on the surface of implants [6]. However, others indicated that PGE 2 production by osteoblasts plated on the rough surface is lower than that by cells plated on the smooth surface and thus, the cell number and differentiation are increased on the former surface [7] [8], suggesting that the regulatory role of PGE 2 on implant-induced osteoblast differentiation and proliferation may be dependent on the implant surface microstructure. The effect of PGE 2 on bone remodeling may also be dependent on the concentration of this cytokine. This notion was based on the fact that low and high concentration of PGE2 induce bone formation and destruction in the implanted site, respectively [9].
Potent osteoconductive properties of hydroxyapatite make it a widely popular biomaterial for orthopaedic and dental implants [10]. Bone formation in the implanted hydroxyapatite involves complex regulatory roles of soluble mediators and intracellular signaling pathways initiated by the binding between osteoblast-expressed surface integrin molecules and its ligand, e.g., arginine-glycineaspartic acid (RGD) motif, of the extracellular matrix proteins covering hydroxyapatite surfaces [11]. Previous studies showed that hydroxyapatite stimulates human osteoblast proliferation via cell surface integrin αV and under the regulation of nitric oxide (NO) [12] [13] [14] [15]. Moreover, PGE 2 was produced by osteoblasts cultured on hydroxyapatite [16] [17]. Therefore, the aim of the present study was to determine whether hydroxyapatite-stimulated human osteoblast proliferation may be directly under the regulation of PGE 2 .
After harvesting and washing, a single cell suspension (1 × 10 6 cells/ml) was prepared in the above medium. In order to elucidate the role of COXs, indomethacin (a non-specific COX inhibitor), aspirin (a COX-1 inhibitor) and nimesulide (a COX-2 inhibitor) were used and dissolved in DMSO [17]. Aspirin was a gift from PT. Bayer, Jakarta-Indonesia. One million cells were incubated in 1 ml of the culture medium containing various concentrations of the respective COX inhibitors for 30 minutes at room temperature. After washing, two hundred microliters of cell suspension containing 2 × 10 5 cells/well in the 96-well culture plates were cultured with or without hydroxyapatite and incubated for 3 days at 37˚C in a humidified atmosphere and 5% CO 2 [12].
The next study was to determine the effect of exogenous PGE 2 on the cell cul- SQ22536, an adenylyl cyclase inhibitor, and dibutyryl cAMP (Db-cAMP), a cAMP analog, used to determine the role of adenylyl cyclase were dissolved in distilled water to obtain 1 mM of solution [15]. Both materials were purchased from Sigma. One million cells were incubated in 1 ml of the culture medium containing 100 μM of SQ22536 for 30 minutes at room temperature [15]. After washing, 200 μL of cell suspension containing 2 × 10 5 cells/well were cultured with or without hydroxyapatite stimulation as described above. In some wells, the cultures pre-treated with or without SQ22536 were added with 10 μM of PGE 2 and/or 10 μM of Db-cAMP. All cultures were incubated for 3 days.
Further experiment was carried out to determine the role of protein kinase A (PKA). KT5720, a PKA inhibitor, and forskolin, an adenylyl cyclase activator, were dissolved in DMSO, whereas Db-cAMP was diluted in distilled water to obtain 1 mM of a stock solution. All of them were then filter sterilized. One million cells per one milliliter culture medium were incubated with KT5720 (1 µM) and/or forskolin (10 µM) for 2 hours at room temperature. After washing, the cells were cultured with or without hydroxyapatite. PGE 2 10 μM and/or of Db-cAMP (10 μM) were added into some wells of the cultures. Again, the cell cultures were incubated for 3 days.
In order to delineate the role of EP2 and EP4, PF-04418948, an EP2 antagon- The data were statistically analyzed by a one-way analysis of variance followed by Fischer's least square differences (SPSS co., Chicago, USA).

The Role of Cyclooxygenases (COXs)
The role of COX isoforms on osteoblast proliferation stimulated with hydroxyapatite were studied by using COX inhibitors, i.e., indomethacin, a non-specific COX inhibitor, aspirin, a COX-1 inhibitor, and nimesulide, a COX-2 inhibitor.
As seen in Figure 1(a), only pre-treatment with indomethacin and nimesulide but not aspirin resulted in partial suppression of hydroxyapatite-stimulated osteoblast proliferation in a dose-dependent fashion as compared with the COX inhibitor-untreated cell proliferation (P < 0.05). In contrast, indomethacin and nimesulide but not aspirin did inhibit fully the production of PGE 2 by cells stimulated with hydroxyapatite (P < 0.05) (Figure 1(b)). Partial suppression of osteoblasts proliferation but profound inhibition of PGE 2 production due to indomethacin or nimesulide could also be observed in the osteoblast cultures without hydroxyapatite stimulation (P < 0.05).

The Effect of Exogenous PGE2
When nimesulide-pretreated cells were cultured with or without hydroxyapatite and added with various concentration of PGE 2 , the results showed that the sup-

The Role of Adenylyl Cyclase
Addition of either PGE 2 or Db-cAMP increased the proliferation of cells cultured with or without hydroxyapatite as compared with that of the untreated cells (P < 0.05) ( Figure 4). Interestingly, the proliferation of SQ22536-pretreated cells cultured with or without hydroxyapatite was significantly lower as com-

The Role of Protein Kinase A (PKA)
A PKA inhibitor, KT5720, was used to determine the central role of PKA on the mechanism of PGE 2 on osteoblast proliferation induced by hydroxyapatite. The results showed that additional exogenous PGE 2 , Db-cAMP, or forskolin enhanced the proliferation of cells stimulated with or without hydroxyapatite (P > 0.05) ( Figure 5). In contrast, KT5720 did suppress the cell proliferation stimulated with hydroxyapatite (P > 0.05). Suppressed cell proliferation mediated by KT5720 on the cell cultures with or without hydroxyapatite could not be abolished by adding exogenous PGE 2 , Db-cAMP or forskolin alone or combination of PGE 2 and Db-cAMP or forskolin (P > 0.05). The results indicated that the activation of PKA may prerequisite for the action of PGE 2 on the osteoblast proliferation stimulated with or without hydroxyapatite.

The Role of EP2 and EP4
The role of EP2 and EP4 on the action of PGE 2 toward the osteoblast proliferation stimulated with or without hydroxyapatite were studied by using their antagonists, i.e., PF-044189448 and L-161,982, respectively. When the cells were pretreated with PF-044189448 or L-161,982 and then cultured with or without hydroxyapatite, the cell proliferation and cAMP production were significantly lower as compared with that treated with or without PGE 2 (P < 0.05) ( Figure   6(a) and Figure 6

Discussion
The present study showed that whilst nimesulide does fully inhibit PGE 2 pro-  [17], the exact reason(s) by which partial cell proliferation was due to COX-2 inhibition remains unclear. During bone formation, osteoblast differentiation and proliferation require not only PGE 2 but also other growth factors/mediators such as NO and insulin-like growth factor-1 (IGF-1) [18]. Thus, it is conceivable that partial cell proliferation with or without hydroxyapatite as a That adenylyl cyclase inhibitor-mediated suppression of cell proliferation was abolished by cAMP analog suggests that osteoblast proliferation may be dependent on the activation of cAMP and that hydroxyapatite stimulation on osteoblast proliferation may amplify the cAMP pathway as also previously documented [14] [15]. Interestingly, failure of exogenous PGE 2 to overcome the SQ22536-  [25], IL-6 [26] and bone matrix metalloproteinase-1 (MMP-1) [27] is regulated by the cAMP pathway may validate the current study. Furthermore, since hydroxyapatite-induced osteoblast proliferation was also a c-AMP-dependent manner [15], the assumption that PGE 2 action on osteoblast proliferation was to augment the cAMP pathway initially activated by hydroxyapatite should not be ruled out (Figure 7).  [27], or regulating its AMP-activated protein kinase kinase (AMPK) [28] are all a PKA-dependent mechanism. Yet again, a possibility that the effect of PGE on osteoblast proliferation induced by hydroxyapatite is also to amplify the same signal transduction pathway is imminent. It should be noted, however, that partial suppression of hydroxyapatite-stimulated osteoblast proliferation by a COX-2 inhibitor was observed in the current study. Therefore, it seems plausible that although COX-2 was inhibited, osteoblast-derived PKA Figure 7. A simplified model of PGE 2 action on hydroxyapatite-stimulated osteoblast proliferation. Following the binding between extracellular matrix-derived RGD motifs covering hydroxyapatite (HA) surfaces [11] and osteoblast-expressed surface αV integrin molecules [12], adenylyl cyclase and hence, a cAMP pathway is generated (1) thereby leading to PKA activation. Multiple signals encoding the production of cyclooxygenes-2 (COX-2) enzyme and other mediators such as endothelial nitric oxide synthase (eNOS) and growth factors are generated by activated PKA. The action of COX-2 enzyme catalyzes the production PGE 2 from arachidonic acid and acts on the hydroxyapatitestimulated osteoblasts in an autocrine fashion via both EP2 and EP4, thereby amplifying the existing signals of the cAMP-PKA pathways (2) and hence, up-regulating cell proliferation. Note: a bold arrow represents a dominant signal.  [27] that involve in osteoblast proliferation ( Figure 7). This contention is in accordance with wealth of evidences indicating that PKA may act as a crossroad-like site where multiple signals could be generated to activate multiple genes encoding multiple mediators [29]. Further studies are required to validate the above contention.
Out of four receptor subtypes specific for PGE 2 , EP2 and EP4 are known to activate the cAMP-PKA pathway [1]. Therefore, the next experiment was to determine whether the role of PGE 2 on hydroxyapatite-stimulated osteoblast proli-  [3]. Surprisingly, previous reports demonstrating that the stimulatory effect of PGE 2 on hydroxyapatite-induced osteoblast proliferation is mediated via EP2 and EP4 are still lacking. The work of Sanuki and colleagues [30] indicating that application of mechanical forces onto human osteoblasts (Saso-2 cells) results in an increased PGE 2 production and EP2/EP4 expression may support the results of this study.
Furthermore, the present study also indicated that the usage of EP2 by PGE 2 was predominant over that of EP4 in osteoblast proliferation with or without hydroxyapatite. It was previously found that EP2 agonists on murine osteoblasts induce cAMP production much higher than EP4 agonists [31] and that EP2 and EP4 act independently due to their differences in generating specific MAPK signaling pathways [32]. Thus, PGE 2 -elevated osteoblast proliferation and cAMP production of the cell cultures with or without hydroxyapatite may be due to greater signal transduction generated from activated EP2 than EP4 (Figure 7).
However, this contention needs to be investigated further.

Conclusion
The PGE 2 production and partial cell proliferation by human osteoblasts stimulated with hydroxyapatite were a COX-2-dependent mechanism. The cell proliferation cultured with hydroxyapatite was enhanced and suppressed by low and high concentration of PGE 2 , respectively. The adenylyl cyclase inhibitor-suppressed cell proliferation was abolished by cAMP analog but not exogenous PGE 2 . Suppression of hydroxyapatite-induced cell proliferation by a PKA inhibitor could not be eliminated by cAMP analog and/or adenylyl cyclase activator. The COX-2-mediated PGE 2 acted on hydroxyapatite-stimulated osteoblast proliferation via both EP2 and EP4 but the former receptor was more dominant than the latter one. Therefore, the current study suggests that human osteoblast proliferation induced by hydroxyapatite may be partially regulated by PGE 2 acting in an autocrine fashion via a COX-2-, cAMP-PKA-and both EP2 and EP4-dependent