This work was presented in part at the 24th Annual Meeting of the American Society for Bone and Mineral Research, San Antonio, TX, USA, September 20-24, 2002.
The authors have no conflict of interest.
We examined the effect of PGE2 on OC formation from spleen cells treated with M-CSF and RANKL. PGE2 decreased OC number at 5–6 days of culture and increased OC number, size, and resorptive activity at 7–8 days. A selective EP2 receptor agonist mimicked these effects. Deletion of the EP2 receptor or depletion of T-cells abrogated the increase in OC number.
Introduction: Prostaglandin E2 (PGE2) has been reported to increase osteoclast (OC) number in spleen cells cultured with RANKL and macrophage-colony-stimulating factor (M-CSF). In this study, we examined the time course of PGE2 effects on spleen cells cultured with RANKL and M-CSF. We then investigated which PGE receptors and cell types were involved in these effects.
Materials and Methods: Spleen cells were cultured from wildtype C57BL/6 mice and EP2 or EP4 receptor-deficient (−/−) and wildtype (+/+) mice on a mixed genetic background. Spleen cells were cultured with M-CSF and RANKL for 5–9 days with or without PGE2 or selective agonists for the four PGE2 receptors (EP1A, EP2A, EP3A, or EP4A). Some cultures were performed using T-cell-depleted spleen cells. OC number and size were quantitated. OC apoptosis and pit formation were measured at 7 or 8 days.
Results: PGE2 decreased the number of OCs formed in the presence of RANKL and M-CSF at 5–6 days of culture and increased OC number at 8–9 days compared with cultures without PGE2. PGE2 also increased OC size at 7 and 8 days, decreased apoptosis of OC at 7 days, and increased pit formation at 8 days. EP1A or EP4A had no effect on OC. EP3A decreased OC number. EP2A mimicked effect of PGE2. EP2−/− spleen cells showed no increase in OC number in response to PGE2, whereas deletion of EP4 had no effect. Depletion of T-cells abrogated the late increase of OC number.
Conclusions: We conclude that PGE2 has an initial inhibitory effect on OC formation in spleen cell cultures, possibly mediated by both EP2 and EP3 receptors, and a later stimulatory effect, mediated by the EP2 receptor, possibly acting on T-cells.
PROSTAGLANDINS, PARTICULARLY PROSTAGLANDIN E2 (PGE2), are potent stimulators of bone resorption in organ culture and in vitro.(1,2) PGE2 can promote osteoclastogenesis in bone marrow cultures and co-cultures of osteoblastic/stromal cells with spleen cells.(3,4) This effect has been attributed to stimulation of the expression of RANKL in osteoblastic/stromal cells.(5,6) Addition of RANKL and macrophage-colony-stimulating factor (M-CSF) is necessary and sufficient to stimulate osteoclastogenesis in cultures of hematopoietic precursor cells that lack osteoblastic/stromal cells.(5) However PGE2 has been shown to increase further osteoclastogenesis in spleen cell cultures treated with RANKL and M-CSF.(7-9) PGE2 activates four different receptors, EP1R, EP2R, EP3R, and EP4R, and two of these, EP2R and EP4R, have been implicated in the stimulation of RANKL production and osteoclastogenesis in the presence of osteoblastic/stromal cells.(6, 9, 10, 11) We previously reported that the stimulatory effect of PGE2 in spleen cell cultures was absent when the cells were derived from mice lacking EP2R but was not blocked by an EP4R antagonist.(8)
This study was undertaken to examine the effects of PGE2 on hematopoietic cells in greater detail. Time course experiments showed an unexpected biphasic effect of PGE2 on osteoclast (OC) formation, identified as TRACP+ multinucleated cells (TRACP+ MNCs). In murine spleen cell cultures treated only with RANKL and M-CSF, OC number usually peaked at 6 or 7 days and then decreased. Increasing RANKL concentration from 10 to 100 ng/ml did not alter the response. Addition of PGE2 to cultures resulted in a decrease in OC number during the first 5 or 6 days of culture relative to RANKL and M-CSF treatment alone (control) and an increase above control at 8 and 9 days. PGE2 also increased OC size and pit formation and decreased OC apoptosis. To define the mechanism of these effects, we subsequently examined the roles of specific PGE2 receptor agonists, EP2R or EP4R, deletion, and T-cell depletion in this process.
MATERIALS AND METHODS
5- to 12-week-old male C57BL/6 mice from Harlan (Indianapolis, IN, USA) were used in all studies not involving transgenic mice. Mice in which the EP2R gene was disrupted were a gift from Dr Richard Breyer (Vanderbilt University, Nashville, TN, USA) and described previously.(12) EP2R heterozygous (+/−) F1 animals in a C57BL/6 × 129/SvEv background were bred to produce homozygous (−/−) null mice and wildtype (+/+) mice. EP4R knockout mice, originally generated in Dr Beverly Koller's laboratory,(13) were the kind gift of Dr Lydia Pan (Pfizer Central Research, Groton, CT, USA). These mice were bred in a mixed background of 129/Ola, C57BL/6, and DBA/2 to enhance survival of EP4R−/− mice. Wildtype and EP4R+/− mice were interbred every three generations to minimize genetic drift. Mice were genotyped by PCR analysis performed on DNA extracted from the tails as previously described.(12,13) Animals were housed at the Center for Laboratory Animal Care, University of Connecticut Health Center. All animal protocols were approved by the Animal Care and Use Committee of the University of Connecticut Health Center.
PGE2 was obtained from Cayman Chemical (Ann Arbor, MI, USA). Specific agonists for the EP receptors - EP1A, EP2A, EP3A, and EP4A (ONO-DI-004, ONO-AE1-259, ONO-AE-248, and ONO-AE1-329, respectively) - were kindly provided by Ono Pharmaceutical (Osaka, Japan). Recombinant murine RANKL and M-CSF were obtained from R & D Systems (Minneapolis, MN, USA). Culture dishes and plates were purchased from Corning (Corning, NY, USA). αMEM and FCS were from Life Technologies (Grand Island, NY, USA). Leukocyte acid phosphatase A kit for TRACP staining was from Sigma (St Louis, MO, USA). Other chemicals and reagents were of analytical grade and obtained from Sigma.
Spleen cells were isolated from 5- to 12-week-old male mice as described previously.(8,9) Spleen cells were cultured with 10 ng/ml M-CSF and 10 or 30 ng/ml RANKL in αMEM containing 10% heat-inactivated FCS (HIFCS). Cells were plated at 5.0 × 105 cells/well in a 24-well plate (1.0 ml/well) and cultured for 5-9 days. Cultures were fed every 3 or 4 days by replacing 0.5 ml of spent medium with fresh medium. The reagents were added at the beginning of the culture and at the time of medium change. Cultures were maintained at 37°C in a humidified atmosphere of 5% CO2 in air. At the end of culture, adherent cells were fixed with 2.5% glutaraldehyde in PBS for 30 minutes and stained for TRACP. TRACP+ MNCs containing three or more nuclei were counted as OCs. Total number of OCs was counted in each of four to six wells.
OC size was quantitated with a video area measurement system (Boeckler Instruments, Tucson, AZ, USA) using a reflected light microscope. Two fields were randomly chosen in each well, and OC sizes were measured in four wells for each experiment (total eight fields). Sizes of >39 OCs were measured for each experimental point.
Bone resorption and pit formation
Bone resorption was assayed by measuring the ability of cultured spleen cells to form resorption pits on the surface of devitalized bovine cortical bone slices (4.4 × 4.4 × 0.2 mm) as previously described.(14,15) Spleen cells were cultured in wells containing devitalized bone with RANKL and M-CSF (both at 30 ng/ml) and with or without PGE2 (1 μM) at 37°C in αMEM and 10% HIFCS. After the culture, cells were fixed with 2.5% glutaraldehyde in PBS for 30 minutes and stained for TRACP, followed by staining with 1% toluidine blue in 1% borax to observe resorption pits. Resorption pit area on these bone slices was quantitated with a video area measurement system (Boeckler Instruments) using reflected light microscopy.
T-cells were separated from mouse spleen cells using mouse pan T (Thy 1.2) dynabeads (Dynal, Oslo, Norway) for depletion of T-cells as previously described.(16) Using this method, 96-98% of T-cells were removed as assessed by flow cytometry. The remaining cells were cultured at 5 × 105/well, similar to the whole spleen cell cultures described above.
Measurement of DNA content
Spleen cells were cultured in αMEM containing 10% heat-inactivated FCS at 2.0 × 106 cells/well in a 6-well plate (2.0 ml/well) for 5-9 days. At the end of culture, cells from two wells were pooled. DNA was isolated using DNAZOL Reagent (Invitrogen, Carlsbad, CA, USA). DNA content was measured by UV absorption at 260 nm, and sample purity was determined by a 260/280 ratio of 1.4-1.6.
Detection of apoptosis: TUNEL staining
Spleen cells were grown on 8-well chamber glass slides (Laboratory-Tek; Nalge Nunc International, Naperville, IL, USA) at 2.5 or 5 × 105 cells/well as described previously. At 7 or 8 days, cells were fixed with 4% paraformaldehyde in PBS (pH 7.4) for 1 h at room temperature. Fixed cells were washed with PBS, permeabilized with 0.1% Triton X-100 in 0.1% sodium citrate solution for 2 minutes on ice, and incubated with terminal deoxynucleotidyl transferase and fluorescein isothiocyanate (FITC)-labeled dUTP (Roche Diagnostics, Indianapolis, IN, USA) in a humidified chamber in dark at 37°C for 1 h. The cells were washed with PBS, stained for TRACP, and visualized by fluorescence microscopy (Nikon TE 2000-U; MVI, Avon, MA, USA). OCs displaying three or more green-labeled nuclei were scored as apoptotic. Six contiguous fields were screened in each of three wells of both control and PGE2-treated samples to give a total of 50-60 OCs for examination of apoptosis. The percentage of apoptotic OCs was determined as apoptotic/total number of examined OCs.
The results were expressed as means ± SE. Comparison between multiple groups were made by a nonparametric Mann-Whitney U-test. Paired comparisons were made by t-test. For TUNEL experiments, comparisons were made by χ2.
In spleen cell cultures from C57BL/6 mice treated with RANKL and M-CSF (control cultures), OC formation was seen at 5 days and usually peaked at 7 days, decreasing thereafter (Fig. 1A; Table 1). The number of OCs was similar over a wide range of RANKL concentrations. In the presence of 30 ng/ml M-CSF, values at 8 days for 10, 30, and 100 ng/ml RANKL were 250 ± 25, 270 ± 14, and 291 ± 25, respectively. Therefore, either 10 or 30 ng/ml of RANKL was added to the cultures in all other experiments. Cultures treated with PGE2 (1 μM) showed an initial decrease in OC number at 5 and 6 days compared with controls. Then, as OC number declined in control cultures, it continued to increase in PGE2-treated cultures, usually reaching a peak at 8 days. Although the absolute number of OCs varied, particularly at 8 days, a biphasic effect was observed consistently for eight different experiments (Table 1). The mean decrease at 6 days was 58%, and the peak values for OC number at 8 days in the presence of PGE2 were increased by 160% and were equal to or greater than the peak values at 7 days in control cultures. However, the magnitude of the difference at 8 days varied greatly, largely because of the variable decrease in OC numbers in cultures at this time. Treatment of spleen cells with EP2A mimicked the effect of PGE2 (Fig. 1). EP2A showed a dose-related response from 10 nM to 1 μM for both the decrease and increase in OC number (Table 2). The effects of EP2A were abrogated when spleen cells from EP2R−/− animals were used, whereas PGE2 treatment still produced an inhibitory effect but no late increase in OC number (Fig. 2A). The inhibition may have been mediated by the EP3 receptor because EP3A caused a sustained decrease in OC number with no late increase (Fig. 1B). PGE2 still produced an initial decrease and subsequent increase in OC number in cultures from EP4R−/− mice, which were on a different genetic background (Fig. 2B).
Table Table 1.. Effects of PGE2(1 μM) in OC Number in Spleen Cell Cultures Treated With RANKL (30 ng/ml) and M-CSF (10 ng/ml)
Table Table 2.. Effects of EP2A on OC Number in Spleen Cell Cultures From EP2R+/+ and EP2R−/− Mice Treated With RANKL (30 ng/ml) and M-CSF (10 ng/ml)
PGE2 and EP2A not only increased OC number but also resulted in the formation of larger OC (Figs. 3A and 3B). OC size tended to decrease in control cultures after 6 days but was sustained in PGE2- or EP2A-treated cultures. In contrast, EP1A, EP3A, and EP4A did not alter OC size. In additional experiments, OC size was increased by PGE2 or EP2A at 8 days in wildtype but not in EP2R−/− mice (data not shown).
OCs in 7-day cultures on cortical bone slices formed resorption pits (Fig. 4). The total resorption pit area on bovine cortical bone slices was increased 5-fold by treatment with PGE2 (Fig. 4).
The initial decrease in response to PGE2 was associated with decrease in the DNA content of adherent spleen cells in cultures treated with RANKL and M-CSF, which persisted throughout 8 days of culture. EP2A had a similar effect to PGE2, whereas EP3A produced a smaller decrease (Table 3).
Table Table 3.. DNA Content (μg/2 wells) of Adherent Spleen Cells in the Presence of RANKL (30 ng/ml) and M-CSF (10 ng/ml), With and Without PGE2, EP2A, or EP3A (all 1 μM) for 6-8 Days of Culture
When spleen cells from C57BL/6 mice were depleted of their T-cell population, PGE2 produced a sustained inhibition of OC number with no late stimulation (Fig. 5A). EP2A also produced a sustained inhibition with no late stimulation in T-cell-depleted cultures (Fig. 5B). Treatment with EP3A produced an inhibitory effect in both whole spleen cell cultures and cultures depleted of T-cells (Fig. 5B).
Two experiments were performed to examine the effect of PGE2 on apoptosis by TUNEL staining (Table 4; Fig. 6). In experiment 1, spleen cells were plated at 5 × 105 cells/well and grown for 7 days. In experiment 2, cells were plated at 2.5 × 105 cells/well and grown for 8 days. PGE2 decreased the number of OC containing three or more apoptotic nuclei by 67% and 84%, respectively.
Table Table 4.. Percent Apoptosis in Spleen Cell Cultures Measured by TUNEL
PGE2 can not only stimulate OC formation in marrow cultures and combined cultures of spleen cells and osteoblastic cells, but can also increase OC formation in spleen cells cultured with RANKL and M-CSF in the absence of osteoblastic/stromal cells.(7-9) This study was designed to explore further this effect on cells of the hematopoietic lineage. A biphasic effect of PGE2 on osteoclast formation was observed, in that there was a decrease in OC number at 5-6 days and a later increase in OC number at 8-9 days. Our previous study had suggested that the stimulatory effect is mediated by EP2R, and these studies, using a selective EP2 agonist and spleen cells from EP2R−/− mice, confirmed this. However, spleen cells from EP2R−/− cells still showed an inhibitory response to PGE2. Another selective agonist, EP3A, produced inhibition without late stimulation of OC number, and this pathway may have contributed to the early decrease in OC. Both EP2A and EP3A decrease DNA content in spleen cell cultures. Whereas EP4R is a critical receptor in mediating the effect of PGE2 on osteoblasts/stromal cells, the selective agonist, EP4A, did not affect OC number in spleen cell cultures, and both the initial inhibition and the late stimulation were still present in cultures from EP4R−/− animals. EP1A also had no effect on spleen cell cultures in these experiments.
PGE2 and EP2A not only increased OC number but also increased OC size. This could represent increased replication and fusion of OC precursors. However, there was also a decrease in the number of OC containing apoptotic nuclei, indicating that PGE2 prolonged the life span of OC. These PGE2-stimulated OCs were also more active in bone resorption, as indicated by their ability to form a larger area of resorption pits on bovine bone slices.
Whereas the initial inhibitory effect of PGE2, EP2A, and EP3A on OC formation was associated with decreased DNA content in these cultures, the late increase in OC number was not associated with a reversal of this inhibition. However, the cultured spleen cells include many other cell types, and hence, a specific increase in proliferation of osteoclast precursors cannot be ruled out.
In our study, the stimulatory effect of PGE2 or EP2A depended on the presence of T-cells. There are a number of studies implicating a role for T-cells in osteoclastogenesis.(16-24) T-cells can produce both stimulators and inhibitors of bone resorption. T-cell production of RANKL has been implicated in bone loss associated with arthritis.(17) However, in our experiments, RANKL seemed to be at saturating concentrations, because increasing RANKL concentrations from 10-100 ng/ml had little effect on OC number. T-cell production of TNFα has been implicated in the bone loss associated with estrogen deficiency,(18) and TNFα can stimulate OC formation independent of RANKL.(25) TGFβ is another potential stimulator produced by T-cells, although its effects on OC formation and activity are complex.(26,27) T-cells also seem to be required for the inhibitory effect of interleukin (IL)-12 and IL-18 on osteoclast formation from spleen cells or the murine macrophage cell line RAW 264.7.(17,23) Additional inhibitory mediators have been described including granulocyte-monocyte colony stimulating factor (GM-CSF), interferon (IFN)-γ, IL-4, and IL-13, but none of these could be shown to mediate the effects of IL-12 and IL-18.(23)
Our data show that T-cells are critical for the stimulatory effect of PGE2 and EP2A, but not the inhibitory effects of PGE2, EP2A, and EP3A. At present, the mechanisms by which T-cells produce the stimulatory effect are unknown. Our preliminary data suggest that this may be caused by a soluble mediator because supernatants of isolated T-cells previously treated with PGE2 also increase OC number in spleen cultures.(28)
In conclusion, we have shown that PGE2 produces an early decrease in OC number in spleen cell cultures treated with RANKL and M-CSF, followed by an increase in OC number, size, and resorptive activity. Both EP2 and EP3 receptors may mediate the initial decrease, whereas the later increase depends on the presence of the EP2 receptor and may be mediated by an effect on T-cells.
The authors thank Lynn Limeburner and Barbara Capella for help in the preparation of this manuscript and Dr Sun-Kyeong Lee for help in the T-cell depletion experiments. This work was supported by NIH Grants AR18063 (LGR) and DK48361 (CCP).