Inhibitory Effect of Glucocorticoid for Osteoblast Apoptosis ... - CiteSeerX

0013-7227/98/$03.00/0 Endocrinology Copyright © 1998 by The Endocrine Society

Vol. 139, No. 4 Printed in U.S.A.

Inhibitory Effect of Glucocorticoid for Osteoblast Apoptosis Induced by Activated Peripheral Blood Mononuclear Cells TOMOKI NAKASHIMA, HITOSHI SASAKI, MASAHIKO TSUBOI, ATSUSHI KAWAKAMI, KAORU FUJIYAMA, TAKESHI KIRIYAMA, KATSUMI EGUCHI, MASATAKA ICHIKAWA, AND SHIGENOBU NAGATAKI Department of Hospital Pharmacy (T.N., H.S., M.I.), and The First Department of Internal Medicine (M.T., A.K., K.F., T.K., K.E., S.N.), Nagasaki University School of Medicine, 1–7-1 Sakamoto, Nagasaki 852, Japan ABSTRACT Recent studies suggest a protective effect of glucocorticoid against progression of bone erosion and periarticular osteoporosis in patients with rheumatoid arthritis (RA), although this steroid hormone itself is believed to increase bone loss. To understand the antagonistic effect of glucocorticoid for osteopenic process in RA patients, we examined the effect of dexamethasone on Fas-mediated apoptosis of cultured human osteoblasts induced by either anti-Fas IgM or activated peripheral blood mononuclear cells (PBMC). Human osteoblastic cell line MG63 and primary osteoblast-like cells obtained from biopsy specimens were used in this study. PBMC isolated from healthy donors were cultured with or without recombinant interleukin-2 (rIL-2) followed by 12-O-tetradecanoyl-phorbol 13-acetate (PMA) with ionomycin in the presence or absence of dexamethasone. Fas was functionally expressed on MG63 and primary osteoblast-like cells, and treatment of these cells with dexamethasone affected neither Fas expression nor anti-Fas IgM-induced apoptosis. Activated PBMC expressing membrane-type Fas ligand (mFasL) efficiently killed both MG63 and primary osteoblasts-like cells, and the addition of human Fas chimeric protein (hFas-Fc) significantly diminished the cytotox-

O

STEOPOROSIS of juxtaarticular bone as well as bone erosion are cardinal diagnostic criteria for RA (1–3). Although the exact cause of osteopenic process has not been identified, several studies have suggested that accumulation of activated mononuclear cells in the synovium of affected joints may play a role in this process (2). High concentrations of bone resorbing cytokines interleukin (IL)-1b, IL-6, and tumor necrosis factor (TNF)-a derived from activated mononuclear cells are present in the synovial fluid of RA patients (4 –7). In the same patients, the degree of periarticular osteoporosis is proportional with the disease activity (8, 9), whereas the severity of bone destruction in the joints is proportional to the concentration of IL-1 produced from the synovium (10). These findings suggest that periarticular bone loss and bone erosion are directly associated with the inflammatory process in RA joints. In this regard, glucocorticoid-induced osteoporosis remains a common and important problem in rheumatoid diseases, but controversy continues about the relative safety of low-dose glucocorticoid Received September 8, 1997. Address all correspondence and requests for reprints to: Hitoshi Sasaki, Ph.D., Associate Professor, Department of Hospital Pharmacy, Nagasaki University School of Medicine, 1–7-1 Sakamoto, Nagasaki 852, Japan.

icity, indicating that interactions between mFasL of activated PBMC and Fas on human osteoblasts induce apoptosis of the latter. Although dexamethasone did not affect apoptosis of MG63 and primary osteoblast-like cells induced by anti-Fas IgM, treatment of activated PBMC with dexamethasone markedly inhibited both mFasL expression and cytotoxicity of these cells against human osteoblasts, suggesting that dexamethasone preferentially acts not on osteoblasts but PBMC. Cultured supernatants from activated PBMC induced apoptosis of human osteoblasts and the addition of hFas-Fc also inhibited the cytotoxicity of the supernatants. In addition, soluble form FasL (sFasL) was detected in the supernatants of activated PBMC. Furthermore, both the cytotoxicity and sFasL concentration of cultured supernatants of activated PBMC incubated with dexamethasone was significantly lower than that in the absence of dexamethasone. Our data suggest that glucocorticoid suppresses the apoptotic process of osteoblasts by inhibiting the expression of both mFasL and sFasL derived from activated PBMC, mediating a protective effect against periarticular bone loss and bone erosion in inflammatory arthritis such as RA. (Endocrinology 139: 2032–2040, 1998)

therapy, particularly with regard to its effect on bone, which should be carefully weighed against the beneficial effects of controlling synovitis and minimizing functional impairment (3). Glucocorticoid itself induces bone loss by several mechanisms (2, 3, 11); however, recent reports suggested a protective effect for glucocorticoid against bone erosion and periarticular bone loss in RA patients based on its suppressive effect on inflammation (8, 12). Apoptosis of bone cells could be involved in the progression and regression of osteoporosis (13–17). In vitro studies showed that apoptosis of osteoclasts could be induced by the addition of estradiol and vitamin K2, which are therapeutic agents used for osteoporosis (15, 16). In addition, apoptosis of murine osteoblasts, the cells that enhance new bone formation (18), is accelerated by TNF-a (13, 14). These data suggest that the induction of osteoclast apoptosis and the prevention of osteoblast apoptosis are the desirable mechanisms to treat osteoporosis patients. Previous reports showed that apoptosis is tightly controlled by various gene products (19). Among these, Fas, a member of the tumor necrosis factor receptor gene family, has been identified as the putative molecule capable of transducing apoptotic signals into cells (20). In addition, FasL, a natural ligand of Fas, has also been cloned and its binding to the cognate Fas

2032

REGULATION OF OSTEOBLAST APOPTOSIS BY GLUCOCORTICOID

receptor induces apoptosis of target cells (21). We also recently showed that Fas-mediated apoptosis of cultured human osteoblasts is induced by both anti-Fas IgM and activated peripheral blood mononuclear cells (PBMC) strongly expressing mFasL (17). Immunohistological examinations suggest the presence of Fas-mediated apoptosis of synovial cells and endothelial cells in rheumatoid synovium (22, 23). In addition, infiltration of activated PBMC expressing mFasL into the rheumatoid synovium has also been identified (24), suggesting the importance of Fas-mediated apoptosis of osteoblasts in the progression of periarticular osteoporosis and bone erosion in RA patients. Furthermore, mFasL is cleaved by metalloproteinases and released as a soluble form (sFasL), which can also induce apoptosis of target cells (25, 26). We have recently shown that sFasL as well as mFasL derived from activated PBMC induce apoptosis of human osteoblasts (17). In the present study, we examined the possible role of glucocorticoid in the interactions between human osteoblasts and activated PBMC via Fas/FasL pathways. Materials and Methods Cell culture Human osteoblastic cell line MG63 and human primary osteoblastlike cells obtained from normal human bone who were undergoing corrective surgery following accidental injury, were used in the present study. None of the patients had any known metabolic bone disease or endocrine disorder. A signed consent was obtained from each patient. The production of osteocalcin from these cells was determined before the present study, using an ELISA assay (data not shown), and alkaline phosphatase (ALP) activity of these cells was assayed by the method of Lowry et al. (27), showing that ALP activity of the cells was clearly augmented by the addition of 1,25-(OH)2D3 (Fig. 1). In addition, almost all cells were positively stained when the cells were incubated with 2-amino-2-methyl-1,3-propandiol buffer (Wako Pure Chemical Industries, Osaka, Japan) containing naphthol AS-MX phosphatase and fast blue RR salt (data not shown, Sigma Chemical Co., St. Louis, MO). Dexamethasone was used as a glucocorticoid in the present experiments. PBMC from healthy donors were isolated as described previously (17).

2033

Briefly, mononuclear cells were isolated from heparinized peripheral blood by Ficoll-Conray gradient centrifugation (Daiichi Pharmaceutical Co., Tokyo, Japan). The cells were depleted of adherent cells by incubating the cell suspensions in Petri dishes (Falcon 3003, Becton Dickinson, Oxnard, LA) for 2 h. Both unstimulated and activated PBMC were used in the present experiments. To prepare activated PBMC, nonadherent cells were collected and cultured for 7 days in RPMI1640 supplemented with 10% FBS (Gibco Laboratories, Grand Island, NY) and containing 500 IU/ml of rIL-2 (Takeda Pharmaceutical Co., Osaka, Japan). After cultivation, the cells were collected and further stimulated by PMA (10 ng/ml, Sigma) and ionomycin (500 ng/ml, Sigma) for 24 h in RPMI1640 supplemented with 5% FBS. To examine the effect of dexamethasone for the activation of PBMC, dexamethasone was added to the culture during these 8 days of culture. After incubation, cultured supernatants were also harvested and filtered through a 0.45-mm filter (Millipore Corp., Bedford, MA).

Flow cytometric analysis of Fas and Bcl-2 on human osteoblasts We examined the expression of Fas on cultured MG63 and primary osteoblast-like cells as described previously (17). Briefly, the cells were cultured for 48 h in RPMI1640 supplemented with 2% FBS in the presence or absence of various concentrations of dexamethasone. After incubation, MG63 and primary osteoblast-like cells were detached by the addition of 0.265 mm EDTA. The cells were washed twice with PBS containing 1% FBS. Fas expression was detected by an indirect immunofluorescence method using antihuman Fas monoclonal antibody (mAb, IgG1, MBL, Nagoya, Japan) and phycoerythrin (PE)-conjugated antimouse IgG (MBL) as the second reagent. In brief, MG63 and primary osteoblast-like cells were incubated with saturating amounts of anti-Fas mAb for 30 min at 4 C, washed three times with PBS, and resuspended in PE-conjugated antimouse IgG. After incubation for 30 min at 4 C, the cells were washed, and Fas expression was determined using a flow cytometer (Epics-Profile II, Coulter Immunology, Hialeah, FL). To determine the expression of Bcl-2 on MG63 and primary osteoblast-like cells, the cells were permeabilized with digitonin, as described previously (28, 29). After confirming the adequacy of permeabilization by trypan blue uptake, the permeabilized cells were incubated for 30 min at 4 C with fluorescein isothiocyanate (FITC)-conjugated mouse antihuman Bcl-2 mAb (Dako Japan Co., Kyoto). After incubation, the expression of Bcl-2 on the cells was analyzed using a flow cytometer (Epics-Profile II).

Induction of apoptosis of MG63 and primary osteoblast-like cells by anti-Fas IgM Cultured MG63 and primary osteoblast-like cells were examined for anti-Fas IgM-induced apoptosis using 51Cr release assay and Hoechst 33258 dye (Wako) staining. 51Cr (Amersham International, Amersham, UK)-labeled MG63 and primary osteoblast-like cells (5 3 103/well) cultured in the presence or absence of various concentrations of dexamethasone for 48 h were incubated for additional 8 h with either control mouse IgM (1,000 ng/ml, Seikagaku Co., Tokyo, Japan) or anti-Fas IgM (1,000 ng/ml, MBL) in 96-well flat-bottom microtiter plates (Falcon 3072, Becton Dickinson) in a total volume of 200 ml of RPMI1640 supplemented with 2% FBS. After incubation, the plates were centrifuged, and 100 ml-aliquots of the supernatants was assayed for radioactivity using a gamma counter. The spontaneous release of 51Cr was determined by incubating the target cells with the medium alone, whereas the maximum release was determined by adding Triton X-100 to a final concentration of 1%. The percentage of specific lysis was determined as follows:

FIG. 1. Effect of 1,25-(OH)2D3 on ALP activity of primary osteoblastlike cells. Various concentrations of 1,25-(OH)2D3was added to the culture for 72 h, and ALP activity of the cells was determined as described in the text. *, P , 0.01, vs. ALP activity in the absence of 1,25-(OH)2D3.

Experimental 51Cr release 2 Spontaneous 51Cr release Lysis (%) 5 3 100 Maximum 51Cr release 51 2 Spontaneous Cr release To detect apoptotic cells with Hoechst 33258 dye staining, MG63 and primary osteoblast-like cells treated with either control mouse IgM or anti-Fas IgM were fixed with 2% glutaraldehyde solution for 10 min and stained with 0.2 mm Hoechst 33258 dye to visualize the localization of

2034


Endo • 1998 Vol 139 • No 4

FIG. 2. Expression of functional Fas on MG63 and primary osteoblast-like cells. A, Fas expression on MG63 and primary osteoblast-like cells was determined by a flow cytometer as described in the text. Numbers in parenthesis are the percentage of positive cells. B, AntiFas IgM-induced apoptosis of MG63 and primary osteoblast-like cells determined by Hoechst 33258 dye staining. A, MG63 treated with control IgM. B, MG63 treated with anti-Fas IgM. C, primary osteoblast-like cells treated with control IgM. D, Primary osteoblast-like cells treated with anti-Fas IgM. Data are representative examples of four experiments. Note that the fragmentation of nuclei and condensation of chromatin in the cells was determined in B and D.

DNA. The cells were examined under a fluorescence microscope to determine fragmentation of nuclei and/or condensation of chromatin (AHB-LB, Olympus, Tokyo, Japan).

Cytotoxic activity of PBMC and cultured supernatants against MG63 and primary osteoblast-like cells Detection of the cytotoxic activity of PBMC and cultured supernatants toward human osteoblasts was performed as previously described (17). Briefly, either unstimulated or activated PBMC with or without dexamethasone treatment were cocultured with 51Cr-labeled MG63 or primary osteoblast-like cells at an effector: target ratio of 20 (5 3 103of MG63 or primary osteoblast-like cells and 1 3 105 of PBMC/well) in 96-well microtiter plates (Costar 3779, Cambridge, MA) in a total volume of 200 ml of RPMI1640 supplemented with 5% FBS. After incubation for 4 h, the

plates were centrifuged, and the percentage of specific release was determined as mentioned above. We also examined the cytotoxic activity of cultured supernatants using 51Cr-labeled MG63 and primary osteoblast-like cells. In this case, 51 Cr-labeled MG63 and primary osteoblast-like cells were incubated with cultured supernatants from either unstimulated or activated PBMC cultured in the presence or absence of dexamethasone in microtiter plates for 8 h as previously described (17). After incubation, the plates were centrifuged and the percentage of 51Cr release in supernatants was calculated using the following formula:

Experimental 51Cr release 2 51Cr release cultured in medium alone Lysis (%) 5 3 100 Maximum 51Cr release 51 2 Cr release cultured in medium alone


2035

TABLE 1. Effect of dexamethasone on Fas and Bcl-2 expression of MG63 A: Fas expression Dexamethasone (mM)

(%) of positive cells Mean fluorescence intensity

Control

0.001

0.01

0.1

1

95.7 6 2.7 1.01 6 0.13

95.1 6 2.3 1.02 6 0.14

94.6 6 2.1 0.96 6 0.10

94.7 6 1.7 0.98 6 0.08

94.3 6 1.9 0.91 6 0.11

Control

0.001

0.01

0.1

1

95.9 6 2.1 3.49 6 0.12

95.2 6 1.8 3.49 6 0.10

95.3 6 1.1 3.49 6 0.10

95.7 6 2.0 3.41 6 0.08

95.5 6 1.7 3.34 6 0.13

B: Bcl-2 expression Dexamethasone (mM)

(%) of positive cells Mean fluorescence intensity

MG63 were cultured with or without various concentrations of dexamethasone for 48 h. After incubation, the expression of Fas and Bcl-2 on these cells was examined as described in the text. Data are the mean 6 SD of four experiments.

Detection of sFasL in cultured supernatants from PBMC. Levels of sFasL in cultured supernatants from PBMC were determined by a sandwich enzyme-linked immunosorbent assay kit using two antihuman FasL mAbs (MBL). The optical density of each well is measured at 450 nm using a microplate reader (ImmunoReader NJ-2001, InterMed, Tokyo, Japan). The concentration of sFasL is calibrated from a dose-response curve based on reference standards.

Proliferation assay of MG63 and primary osteoblast-like cells

FIG. 3. Dexamethasone did not change anti-Fas IgM-induced apoptosis of MG63 determined by 51Cr release assay. MG63 were cultured with various concentrations of dexamethasone for 48 h. After incubation, the cells were labeled with 51Cr, further incubated with either control IgM or anti-Fas IgM, and the cytotoxicity was determined as described in the text. Data are the mean 6 SD of five experiments.

In some experiments, either hFas-Fc, which specifically interferes with the interaction between Fas and FasL (kindly provided by Dr. Shigekazu Nagata, Osaka Bioscience Institute and Department of Genetics, Osaka University Medical School), or human IgG (Seikagaku Co.), was added, and 51Cr release was also determined.

The proliferative response of MG63 and primary osteoblast-like cells was determined using a 3H-thymidine incorporation assay. Briefly, the cells (5 3 103/well) were plated in 96-well flat-bottomed microtiter plates (Falcon 3072) in RPMI1640 supplemented with 2% FBS with or without various concentrations of dexamethasone for 48 h. Twenty-four hours before terminating the cell culture, each well was pulsed with 0.5 mCi of 3H-thymidine (New England Nuclear, Boston, MA) and harvested on glass filter, using a semiautomatic cell harvester (Labo Mash, Labo Science, Tokyo). The radioactivity of each sample was determined in a liquid scintillation counter (Aloka, LSC-5100, Tokyo, Japan).

Statistical analysis Data were expressed as mean 6sd. Differences between groups were tested for statistical significance using the ANOVA. A P value less than 0.05 was selected as the level of significance.

mFasL expression on PBMC

Results Fas expression and anti-Fas IgM-induced apoptosis of MG63 and primary osteoblast-like cells

mFasL expression on PBMC was examined by Western blot analysis. For this purpose, unstimulated or activated PBMC with or without dexamethasone were collected after cultivation and lysed by the addition of lysis buffer (50 mm Tris buffer, pH 8, 150 mm NaCl, 0.02% sodium azaide, 0.1% SDS, 100 mg/ml PMSF, 1 mg/ml aprotinin, 1% NP-40, 0.5% sodium deoxycholate) for 20 min at 4 C, and insoluble materials were removed by centrifugation at 13,000 rpm for 30 min at 4 C. The supernatant was collected and the protein concentration was determined by the Bio-Rad (Melville, NY) protein assay kit. An identical amount of protein for each lysate (20 mg/well) was subjected to 10% SDS-PAGE. Proteins were transferred to a nitrocellulose filter. The filter was blocked for 1.5 h using 5% nonfat dried milk in PBS containing 0.1% Tween 20 (PBS-T), washed with PBS-T and incubated at room temperature for 1 h in 1:1000 dilution of mouse antihuman FasL mAb (Transduction Laboratories, Lexington, KY). The filter was washed with PBS-T and incubated with 1:1000 dilution of donkey antimouse IgG, coupled with horseradish peroxidase. The enhanced chemiluminescence (ECL) system (Amersham) was used for detection.

We first examined the expression of Fas on MG63 and primary osteoblast-like cells using flow cytometry. As described previously (17), Fas was strongly expressed on MG63 and primary osteoblast-like cells (Fig. 2A). In the next step, we investigated whether Fas expressed on these cells was functional by examining the development of anti-Fas IgMinduced apoptosis of MG63 and primary osteoblast-like cells. Although MG63 and primary osteoblast-like cells did not show spontaneous apoptosis (A and C of Fig. 2B), treatment of the cells with anti-Fas IgM-induced morphological changes characteristic of apoptosis (e.g. nuclear condensation/fragmentation) as determined by Hoechst 33258 dye staining (B and D of Fig. 2B). These results indicated the expression of functional Fas on the surface of human osteoblasts.

2036


Endo • 1998 Vol 139 • No 4

FIG. 5. Cytotoxic activity of PBMC against MG63 and primary osteoblast-like cells. 51Cr-labeled MG63 and primary osteoblast-like cells were cultured with either unstimulated or activated PBMC at an effector:target ratio of 20, and the cytotoxicity of PBMC was examined as described in the text. Values are the mean 6 SD of four experiments. *, P , 0.01, vs. activated PBMC in the absence of hFas-Fc.

Effects of dexamethasone on apoptosis and proliferative response of human osteoblasts

The function of mouse osteoblasts such as proliferation and collagen synthesis is influenced by glucocorticoid (11). Therefore, we investigated the effect of dexamethasone on Fas expression and anti-Fas IgM-induced apoptosis of human osteoblasts. As shown in Table 1A, Fas expression on MG63 was not affected by the addition of dexamethasone. Furthermore, dexamethasone did not influence anti-Fas IgMinduced apoptosis of MG63 as determined by 51Cr release assay (Fig. 3). In addition, dexamethasone failed to change spontaneous apoptosis as well as proliferation of these cells (Fig. 4). Because Bcl-2, a proto-oncogene product, inhibits the apoptotic process mediated by Fas (30), we examined the effect of dexamethasone on Bcl-2 expression on human osteoblasts. Because Bcl-2 is an integral membrane protein which lies within the cell rather than on the surface (31), permeabilized MG63 and primary osteoblast-like cells were used. After permeabilization with digitonin, almost all cells were positive with anti-a-tubulin mAb (data not shown). This finding indicated that the technique for permeabilization was suitable for cytoplasmic staining of Bcl-2. Under these conditions, we examined Bcl-2 expression on MG63. As shown in Table 1B, dexamethasone did not influence Bcl-2 expression on MG63. The expression of Fas and Bcl-2 of primary osteoblast-like cells was also not affected by the addition of dexamethasone (data not shown).

FIG. 4. Dexamethasone affected neither spontaneous apoptosis nor proliferation of MG63. A, MG63 were incubated with various concentrations of dexamethasone for 48 h. After incubation, the apoptotic

cells were examined by Hoechst 33258 dye staining. A, MG63 treated without dexamethasone. B, MG63 treated with 0.01 mM of dexamethasone. C, MG63 treated with 1 mM of dexamethasone. Data are representative examples of five experiments. Note that the cells showing the apoptotic appearance were not observed. B, The proliferative response of MG63 was examined by 3H-thymidine incorporation as described in the text. Data are the mean 6 SD of four experiments.


2037

FIG. 6. Effect of dexamethasone on cytotoxicty and mFasL expression of PBMC. PBMC were activated with rIL-2 followed by PMA and ionomycin in the presence or absence of various concentrations of dexamethasone. After incubation, the cytotoxicty and mFasL expression of PBMC were examined as described in the text. A, Cytotoxicity of PBMC. Values are the mean 6SD of six experiments. *, P , 0.05; **, P , 0.01, vs. control (no dexamethasone). B, mFasL expression determined by Western blot analysis. Results are representative examples of six experiments.

FIG. 7. Cytotoxic activity of cultured supernatants from PBMC against MG63 and primary osteoblast-like cells. 51Cr-labeled MG63 and primary osteoblast-like cells were incubated with cultured supernatants of either unstimulated or activated PBMC and the cytotoxicity was examined as described in the text. Values are the mean 6 SD of four experiments. *, P , 0.01, vs. activated PBMC in the absence of hFas-Fc.

Effect of dexamethasone on cytotoxicity of activated PBMC against human osteoblasts

As previously described (17), although unstimulated PBMC did not kill MG63 and primary osteoblast-like cells in vitro, a significant cytotoxicity toward the latter was induced by activated PBMC (Fig. 5). In addition, hFas-Fc markedly inhibited cytotoxicity (Fig. 5), indicating the importance of Fas/FasL interactions in inducing apoptosis of human os-

FIG. 8. Inhibitory effect of dexamethasone on cytotoxicity of cultured supernatants of activated PBMC determined by 51Cr release assay. PBMC were activated with rIL-2 followed by PMA and ionomycin in the presence or absence of various concentrations of dexamethasone. After incubation, the cytotoxicity of cultured supernatants of PBMC against MG63 and primary osteoblast-like cells was determined as described in the text. *, P , 0.01, vs. activated PBMC incubated without dexamethasone. Values are the mean 6 SD of six experiments.

teoblasts. Therefore, we examined the effect of dexamethasone on activated PBMC-induced cytotoxicity against MG63 and primary osteoblast-like cells. For this purpose, PBMC were stimulated with rIL-2 followed by PMA and ionomycin in the presence or absence of dexamethasone, followed by examination of cytotoxicity toward human osteoblasts. As shown in Fig. 6, dexamethasone suppressed both the cyto-

2038


Endo • 1998 Vol 139 • No 4

FIG. 10. Dexamethasone inhibited sFasL production from activated PBMC. PBMC were activated with rIL-2 followed by PMA and ionomycin in the presence or absence of various concentrations of dexamethasone. After incubation, sFasL concentration in cultured supernatants was examined as described in the text. *, P , 0.01, vs. activated PBMC incubated without dexamethasone. Values are the mean 6SD of five experiments. FIG. 9. Inhibitory effect of dexamethasone on cytotoxicity of cultured supernatants of activated PBMC determined by Hoechst 33258 dye staining. MG63 were incubated with cultured supernatants of activated PBMC with or without dexamethasone, and the apoptotic cells were examined by Hoechst 33258 dye staining as described in the text. A, MG63 incubated with cultured supernatants from unstimulated PBMC. B, MG63 incubated with cultured supernatants from activated PBMC. C, MG63 incubated with cultured supernatants from activated PBMC cultured with 0.01 mM of dexamethasone. D, MG63 incubated with cultured supernatants from activated PBMC cultured with 1 mM of dexamethasone. Results are representative examples of five experiments. Note that the number of the cells showing the apoptotic appearance was dose dependently inhibited by dexamethasone.

toxicity and mFasL expression of activated PBMC in a dosedependent fashion. However, the viability of activated PBMC determined by trypan blue uptake, and the detection of hypodiploid DNA was not increased by the treatment of dexamethasone (data not shown), indicating that dexamethasone did not increase the apoptotic process of activated PBMC in the experiment. These data suggest that dexamethasone preferentially inhibited the activation of PBMC since it did not influence anti-Fas IgM-induced apoptosis of human osteoblasts. Dexamethasone inhibited the cytotoxicity of cultured supernatants from activated PBMC toward human osteoblasts

We have previously demonstrated that cultured supernatants of activated PBMC contain sFasL and induce the apoptosis of human osteoblasts as well as mFasL (17). Thus, we examined the effect of dexamethasone on cultured supernatants from activated PBMC. As shown in Fig. 7, cultured supernatants from activated PBMC efficiently killed human osteoblasts, but such activity was significantly reduced by the addition of hFas-Fc (Fig. 7). Furthermore, the cytotoxicity of supernatants from activated PBMC incubated with dexamethasone against human osteoblasts was significantly

lower than that from activated PBMC cultured in the absence of dexamethasone (Fig. 8). The inhibition of apoptosis of human osteoblasts induced by cultured supernatants in the presence of dexamethasone was confirmed by Hoechst 33258 dye staining (Fig. 9). sFasL was clearly detected in cultured supernatants from activated PBMC and the concentration of sFasL was dose dependently suppressed by dexamethasone (Fig. 10). In addition, cytokines themselves including TNF-a, IL-1b, or IL-6 did not induce apoptosis of human osteoblasts as determined by Hoechst 33258 dye staining (Fig. 11). Discussion

There is a considerable debate, based on data from crosssectional studies, as to whether glucocorticoid increase or decrease bone loss in patients with RA (2). Part of the difficulty in interpreting the effect of steroids is that they are frequently reserved for patients with severe RA who are the most disabled and who may lose bone quickly. Recent studies suggest that bone mineral density (BMD) of hand in patients with RA, especially in early RA patients, correlates not with glucocorticoid therapy but with disease activity (8, 9). Furthermore, the use of daily prednisolone in the same patients reduces joint destruction of the hands (12). Since periarticular osteoporosis usually coexists with bone erosion of the affected joints in RA patients (32), these data suggest that glucocorticoid primarily suppress both periarticular bone loss and bone erosion in patients with RA by its inhibitory effect on inflammation including cellular interactions. Although the precise mechanisms causing osteoporosis are not known at present, estradiol and vitamin K2, two agents used in the treatment of these patients, are known to induce apoptosis of osteoclasts (15, 16). In addition, proinflammatory cytokine TNF-a, which is thought to play an important role in RA (4), acts on murine osteoblasts and induces apoptosis of these cells (13, 14). Osteoblasts organize the extracellular matrix and enhance new bone formation


2039

FIG. 11. Cytokines did not affect spontaneous apoptosis of MG63. MG63 were incubated with rTNF-a (100 IU/ml), rIL-1b (20 IU/ml), or rIL-6 (1,000 IU/ml) for 8 h. After incubation, the apoptotic cells were examined by Hoechst 33258 dye staining as described in the text. A, MG63 cultured in the absence of cytokine. B, MG63 cultured with TNF-a. C, MG63 cultured with IL-1b. D, MG63 cultured with IL-6. Results are representative examples of five experiments. Note that the cells showing the apoptotic appearance were not observed.

(18). These data indicate that the regulation of the life span of bone cells through apoptosis could be critical for the induction or prevention of osteoporosis in RA. Kitajima et al. (33) recently reported the presence of osteoblast apoptosis of affected joints in vivo in RA patients. In this regard, we previously reported that Fas is functionally expressed on human osteoblasts, and the interactions between FasL of activated PBMC and Fas on human osteoblasts induce apoptosis of the latter (17). Therefore, to understand the possible role of glucocorticoids on osteoblasts/PBMC interactions, we examined the effect of dexamethasone on Fasmediated apoptosis of human osteoblasts induced by activated PBMC. As previously reported (17), activated PBMC strongly express mFasL and kill human osteoblasts. The addition of hFas-Fc significantly suppressed the cytotoxicity of activated PBMC, indicating that mFasL of these cells is a strong candidate molecule for the induction of Fas-mediated apoptosis of osteoblasts. Although cultured human osteoblasts did not show spontaneous apoptosis, Fasmediated apoptosis was induced by the addition of anti-Fas IgM as previously reported (17). Therefore, the effect of glucocorticoid on anti-Fas IgM-induced apoptosis of human osteoblasts was initially examined. While glucocorticoids affect the rat osteoblast functions such as collagen synthesis (11), dexamethasone did not change Fas expression or antiFas IgM-induced apoptosis in the present study. In addition, the expression of Bcl-2, which is known to inhibit Fas-mediated apoptosis (30), did not change following the addition of dexamethasone. Furthermore, neither spontaneous apoptosis nor proliferation of MG63 and primary osteoblast-like

cells were affected by the addition of dexamethasone. In contrast, both the cytotoxicity of activated PBMC toward osteoblasts and mFasL expression of these cells were suppressed by the addition of dexamethasone in a dose-dependent fashion. Such inhibition was detected when as little as 1029–1028 m of dexamethasone, the concentration could be measured in the synovium of steroid-treated RA patients (34) was used. These data indicate that glucocorticoid preferentially acts on PBMC and inhibits mFasL expression, thus down-regulating the cytotoxicity of activated PBMC toward human osteoblasts. The nature of the osteolytic lesion of affected joints in RA patients, many of which arise in contact with synovial granulation tissues containing activated PBMC (32), suggests that dexamethasone suppresses bone destruction by inhibiting the activation of PBMC. In addition to the direct cellular interactions, humoral factors may play an important role in the development of periarticular osteoporosis as well as bone erosion in RA patients. Shimizu et al. (35) reported the absence of mononuclear cell infiltration in juxtaarticular bone in patients with RA, although bone resorption may be identified. IL-1, IL-6, and TNF-a are found in high concentrations in the synovial fluid of patients with RA (4 – 8). Because these cytokines stimulate bone resorption process (36 –38), it is possible that humoral factors, including cytokines, may act on bone cells and increase bone loss. From this view point, we previously showed that cultured supernatants of activated PBMC induce apoptosis of human osteoblasts (17). Because the addition of hFas-Fc significantly suppressed the cytotoxicity of these supernatants, sFasL is a strong candidate molecule

2040


present in the supernatants of activated PBMC. In the present study, the cytotoxicity of cultured supernatants of dexamethasone-treated activated PBMC was significantly lower than that of control activated PBMC. As shown in the present study, the addition of dexamethasone alone to osteoblasts did not affect the apoptotic process of these cells, and in addition, IL-1b, TNF-a, and IL-6 did not directly induce apoptosis of human osteoblasts. Furthermore, the concentration of sFasL in cultured supernatants of activated PBMC was significantly inhibited by dexamethasone. Thus, our data demonstrated that the production of certain soluble factors, such as sFasL, derived from activated PBMC, which induces apoptosis of osteoblasts, was suppressed by dexamethasone, indicating that this effect may explain the inhibition of progression of periarticular osteoporosis and bone resorption in RA patients. We showed in the present study that dexamethasone inhibited the apoptotic process of human osteoblasts by suppressing the activation of PBMC. Accumulation of activated mononuclear cells expressing mFasL has been demonstrated in RA synovium (24), and in addition, apoptotic change of osteoblasts in periarticular regions of RA patients were detected by nick end-labeling method (33). Although the concentrations of dexamethasone used in the present experiments may inhibit the functions of human osteoblasts such as collagen synthesis, the direct inhibitory effects of glucocorticoid may be overcome by its protective effects for osteoblast apoptosis in particular inflammatory situations such as affected joints of RA, and thus suppresses the progression of periarticular osteoporosis and joint destruction. Acknowledgment We thank Dr. Shigekazu Nagata for providing hFas-Fc. We also thank Dr Hideaki Sakai, Department of Pharmacology, Nagaski University School of Dentistry, for technical advice and helpful suggestion.

References 1. McCarty DJ 1993 Pathology of rheumatoid arthritis and allied disorders. In: McCarty DJ, Koopman WJ (eds) Arthritis and Allied Conditions, ed. 12. Lea & Febiger, PA, vol 1:781– 809 2. Joffe I, Epstein S 1991 Osteoporosis associated with rheumatoid arthritis: Pathogenesis and management. Semin Arthritis Rheum 20:256 –272 3. Henderson NK, Sambrook PN 1996 Relationship between osteoporosis and arthritis and effect of corticosteroids and other drugs on bone. Curr Opin Rheumatol 8:365–369 4. Arend WP, Dayer JM 1990 Cytokines and cytokine inhibitors or antagonists in rheumatoid arthritis. Arthritis Rheum 33:305–315 5. Farahat MN, Yanni G, Poston R, Panani GS 1993 Cytokine expression in synovial membranes of patients with rheumatoid arthritis. Ann Rheum Dis 52:870 – 875 6. Endo H, Akahoshi T, Takagishi K, Kashiwazaki S, Matsushima K 1991 Elevation of interleukin-8 (IL-8) levels in joint fluids of patients with rheumatoid arthritis and the induction by IL-8 of leukocyte infiltration and synovitis in rabbit joints. Lymphokine Cytokine Res 10:245–252 7. Xu WD, Firestein GS, Taetle R, Kaushansky K, Zvaifler NJ 1989 Cytokines in chronic inflammatory arthritis. Granulocyte-macrophage colony-stimulating factor in rheumatoid synovial effusions. J Clin Invest 83:876 – 882 8. Gough AKS, Lilley J, Eyre S, Holder RL, Emery P 1994 Generalised bone loss in patients with early rheumatoid arthritis. Lancet 334:23–27 9. Peel NFA, Spittlehouse AJ, Bax DE, Eastell R 1994 Bone mineral density of the hand in rheumatoid arthritis. Arthritis Rheum 37:983–91 10. Miyasaka N, Sato K, Goto M, Sasano M, Natsuyama M, Inoue K, Nishioka K 1988 Augmented interleukin-1 production and HLA-DR expression in the synovium of rheumatoid arthritis patients: Possible involvement in joint destruction. Arthritis Rheum 31:480 – 486 11. Lukert BP, Raisz LG 1990 Glucocorticoid-induce osteoporosis: pathogenesis and management. Ann Intern Med 112:352–364

Endo • 1998 Vol 139 • No 4

12. Kirwan JR 1995 The effect of glucocorticoids on joint destruction in rheumatoid arthritis. N Engl J Med 333:142–146 13. Kitajima I, Nakajima T, Imamura T, Takasaki I, Kawahara K, Takioka T, Okano T, Soejima Y, Abeyama K, Maruyama I 1996 Induction of apoptosis in murine clonal osteoblasts expressed by human T-cell leukemia virus type I tax by NF-kB and TNF-a. J Bone Miner Res 11:200 –210 14. Kitajima I, Soejima Y, Takasaki I, Beppu H, Tokioka T, Maruyama I 1996 Ceramide-induced nuclear translocation of NF-kB is a potential mediator of the apoptotic response to TNF-a in murine clonal osteoblasts. Bone 19:263–270 15. Kameda T, Miyazawa K, Mori Y, Yuasa T, Shiokawa M, Nakamaru Y, Mano H, Hakeda Y, Kameda A, Kumegawa M 1996 Vitamin K2 inhibits osteoclastic bone resorption by inducing osteoclast apoptosis. Biochem Biophys Res Commun 220:515–519 16. Hughes DE, Dai A, Tiffee JC, Li HH, Mundy GR, Boyce BF 1996 Estrogen promotes apoptosis of murine osteoclast mediated by TGF-b1. Nature Med 2:1132–1136 17. Kawakami A, Eguchi K, Matsuoka N, Tsuboi M, Koji T, Urayama S, Fujiyama K, Kiriyama T, Nakashima T, Nakane PK, Nagataki S 1997 Fas and Fas ligand interaction is necessary for human osteoblast apoptosis. J Bone Miner Res 12: 1637–1646 18. Rodan GA, Rodan SB 1995 The cells of bone. In: Riggs BL, Melton LJ (eds) Osteoporosis: Etiology, Diagnosis, and Management, ed. 2. Lippincott-Raven Publishers, Philadelphia, 1–39 19. Thompson CB 1995 Apoptosis in the pathogenesis and treatment of disease. Science 267:1456 –1462 20. Itoh N, Yonehara S, Ishii A, Yonehara M, Mizushima S, Sameshima M, Hase A, Seto Y, Nagata S 1991 The polypeptide encoded by the cDNA for human cell surface Fas antigen Fas can mediate apoptosis. Cell 66:233–243 21. Suda T, Takahashi T, Golstein P, Nagata S 1993 Molecular cloning and expression of the Fas ligand, a novel member of the tumor necrosis factor family. Cell 75:1169 –1178 22. Nakajima T, Aono H, Hasunuma T, Yamamoto K, Shirai T, Hirohata K, Nishioka K 1995 Apoptosis and functional Fas antigen in rheumatoid arthritis synoviocytes arthritis. Arthritis Rheum 38:485– 491 23. Firestein GS, Yeo M, Zvaifler NJ 1995 Apoptosis in rheumatoid arthritis synovium. J Clin Invest 96:1631–1638 24. Hoa TTM, Hasunuma T, Aono H, Masuko K, Kobata T, Yamamoto K, Sumida T, Nishioka K 1996 Novel mechanisms of selective apoptosis in synovial T-cell of patients with rheumatoid arthritis. J Rheumatol 23:1332–1337 25. Tanaka M, Suda T, Takahashi T, Nagata S 1995 Expression of the functional soluble form of human Fas ligand in activated lymphocytes. EMBO J 14:1129 –1135 26. Kayagaki N, Kawasaki A, Ebata T, Ohmoto H, Ikeda S, Inoue S, Yoshino K, Okaumura K, Yaagita H 1995 Metalloproteinase-mediated release of human Fas ligand. J Exp Med 182:1777–1783 27. Lowry OH, Roberts NR, Wu ML, Hixon WS, Crawford EJ 1954 The quantitative histochemistry of brain. J Biol Chem 233:19 –37 28. Kawakami A, Eguchi K, Matsuoka N, Tsuboi M, Kawabe Y, Ishikawa N, Ito K, Nagataki S 1996 Thyroid-stimulating hormone inhibits Fas antigen-mediated apoptosis of human thyrocytes in vitro. Endocrinology 137:3163–3169 29. Kawakami A, Eguchi K, Matsuoka N, Tsuboi M, Kawabe Y, Aoyagi T, Nagataki S 1996 Inhibition of Fas antigen-mediated apoptosis of rheumatoid synovial cells in vitro by transforming growth factor-b1. Arthritis Rheum 39:1267–1276 30. Itoh N, Tsujimoto Y, Nagata S 1993 Effect of Bcl-2 on Fas antigen-mediated cell death. J Immunol 151:621– 627 31. Hockenbery D, Nunez G, Milliman C, Schreiber RD, Korsmeyer SJ 1990 Bcl-2 is an inner mitochondrial membrane protein that blocks programmed cell death. Nature 348:334 –336 32. Hough AJ Jr 1993 Pathology of rheumatoid arthritis and allied disorders. In: McCarty DJ, Koopman WJ (eds) Arthritis and Allied Conditions, ed 12. Lea & Febiger, PA, vol 1:737–761 33. Kitajima I, Shao Q, Soejima Y, Maruyama I, Ceramide-induced AP-1 activation is a mediator of apoptotic response to TNF-a in periarticular osteoblasts in patients with rheumatoid arthritis. Program of the 61st Annual Meeting of The American College of Rheumatology, Washington, DC, 1997, p S78 (Abstract) 34. Facui AS, Dale DC, Balow JE 1976 Glucocorticosteroid therapy: mechanisms of action and clinical considerations. Ann Intern Med 84:304 –315 35. Shimizu S, Shiozawa S, Shiozawa K, Imura S, Fujita T 1985 Quantitative histologic studies on the pathogenesis of perarticular osteoporosis in rheumatoid arthritis. Arthritis Rheum 28:25–31 36. Manolagas SC 1995 Role of cytokines in bone resorption. Bone [Suppl 2] 17:635– 675 37. Ishimi Y, Miyaura C, Cheng HJ, Akatsu T, Abe E, Nakamura Y, Yamaguchi A, Yoshiki S, Matsuda T, Hirano T, Kishimoto T, Suda T 1990 IL-6 is produced by osteoblasts and induces bone resorption. J Immunol 145:3297–3303 38. Lerner UH, Ohlin A 1993 Tumor necrosis factors-a can stimulate bone resorption in cultured mouse calvariae by a prostaglandin-independent mechanism. J Bone Miner Res 8:147–155