Changes of Serum Soluble Receptor Activator for Nuclear Factor-κB Ligand after Glucocorticoid Therapy Reflect Regulation of Its Expression by Osteoblasts

Abstract

Context:

Osteoporosis is a serious complication of systemic glucocorticoid therapy. The role of serum soluble receptor activator for nuclear factor-κB ligand (RANKL) in glucocorticoid-induced osteoporosis remains unclear.

Objective:

The objective of the study was to clarify the influence of serum soluble RANKL on the osteoprotegerin (OPG)/RANKL/receptor activator for nuclear factor-κB system in patients with systemic autoimmune diseases receiving glucocorticoid therapy.

Patients and Methods:

Sixty patients (40 women) with systemic autoimmune diseases who received initial glucocorticoid therapy with prednisolone (30–60 mg/d) plus bisphosphonate therapy were prospectively enrolled. Serum soluble RANKL and OPG levels were measured at 0, 1, 2, 3, and 4 wk after starting glucocorticoid therapy. The effects of dexamethasone on production of RANKL and OPG mRNA and protein by cultured normal human osteoblasts were evaluated by RT-PCR and ELISA, respectively.

Results:

The mean serum soluble RANKL level of the patients was unchanged by glucocorticoid therapy. Because the distribution of serum soluble RANKL was bimodal, the patients were stratified into two groups. Serum soluble RANKL decreased significantly in the higher soluble RANKL group (≥0.16 pmol/liter), whereas it increased significantly in the lower soluble RANKL group. The mean serum OPG level of the patients decreased significantly. Bone mineral density increased in the higher soluble RANKL group after starting glucocorticoid therapy, whereas it decreased in the lower soluble RANKL group. In cultures of unstimulated human osteoblasts, RANKL mRNA expression was increased and OPG mRNA was decreased by dexamethasone. Up-regulation of RANKL and OPG mRNA by IL-6 was suppressed by dexamethasone.

Conclusion:

Serum soluble RANKL might be a useful marker of bone remodeling in patients with systemic autoimmune diseases receiving glucocorticoid therapy.

Glucocorticoids are widely used to treat a variety of diseases, including systemic autoimmune diseases. Although glucocorticoids can improve the outcome for patients with these diseases, various side effects of long-term treatment, such as osteoporosis, have become an important problem (1–4). Receptor activator for nuclear factor-κB ligand (RANKL) and osteoprotegerin (OPG) have been identified as secreted cytokines that play an important role in regulating bone resorption by osteoclasts (5–8). Glucocorticoids decrease bone density through multiple mechanisms, including inhibition of sex steroid hormones, inhibition of gastrointestinal absorption and renal reabsorption of calcium, promotion of parathyroid hormone secretion, and inhibition of bone formation by suppressing osteoblasts together with stimulation of bone resorption via effects on osteoclasts mediated by changes in the production of RANKL and OPG (1–4). The OPG/RANKL/receptor activator for nuclear factor-κB system is one of the important mechanisms involved in various kind of osteoporosis (9–11). In general, glucocorticoids stimulate RANKL production (12–15) and inhibit OPG production (12–16) in vitro, thereby enhancing bone resorption. However, the pathophysiological role of serum soluble RANKL (sRANKL) in vivo has not been clarified yet. In the present study, we prospectively measured serum sRANKL and OPG levels before and after the initiation of glucocorticoid therapy in patients with systemic autoimmune diseases. We also performed in vitro studies to assess the effects of glucocorticoids on primary cultured human osteoblasts.

Patients and Methods

Patients and study protocol

Patients and healthy controls were recruited at Toho University Omori Hospital and the Research Center for Clinical Pharmacology of Kitasato University, respectively. This study was approved by the Ethics Committee at Toho University Omori Medical Center (approval no. 21-61) and at Research Center for Clinical Pharmacology of Kitasato University (approval no. 06104). This was a prospective observational study that involved 60 patients with systemic autoimmune diseases, including 21 patients with systemic lupus erythematosus (SLE), 15 patients with polymyositis (PM)/dermatomyositis (DM), 19 patients with vasculitis syndrome, and five patients with adult-onset Still's disease (AOSD). Patients who were starting treatment with prednisolone at doses from 30 to 60 mg daily [mean daily dose: 45.2 ± 1.9 mg (sem)] based on our standard therapeutic regimen were eligible for this study. All patients gave written informed consent before enrollment. None of them had received any treatment for their diseases at the time of enrollment.

In guidelines on the management and treatment of glucocorticoid-induced osteoporosis of the Japanese Society for Bone and Mineral Research, a bisphosphonate was recommended as a first-line drug for prevention of glucocorticoid-induced osteoporosis (17). We then adopted administration of a bisphosphonate (alendronate 35 mg/wk or risedronate 17.5 mg/wk) as a supplemental drug for our regimen of high-dose glucocorticoid therapy. This regimen was used for all patients in this study. None of the patients received supplemental calcium and vitamin D during this study.

Fasting morning blood samples were collected prospectively just before the patients started treatment and after 1, 2, 3, and 4 wk of glucocorticoid therapy. Serum samples were immediately frozen at −80 C until the measurement of markers of bone metabolism. The healthy control subjects were matched for sex, menopausal status, and age (±5 yr). Serum samples were immediately frozen at −80 C until assay.

Serum biochemical markers

As inflammation markers, serum levels of C-reactive protein (CRP; Sekisui Medical, Tokyo, Japan) and IL-6 (Fujirebio Inc., Tokyo, Japan) were measured by the latex-enhanced nephelometric method and enzyme immunoassay, respectively. Serum 25-hydroxyvitamin D (25-OHD; Diasorin Inc., Stillwater, MN) was measured by RIA. Serum levels of soluble RANKL (sRANKL; Biomedica, Vienna, Austria) and OPG (Biomedica) were determined by ELISA. As bone formation markers, serum levels of osteocalcin (OC; Mitsubishi Kagaku Bioclinical Laboratories, Tokyo, Japan) and procollagen type I N-terminal peptide (PINP; Orion Diagnostica, Espoo, Finland) were determined by immunoradiometric assay. The serum level of undercarboxylated OC (ucOC; Sanko Junyaku Co., Ltd., Tokyo, Japan) was measured by an electrochemiluminescence immunoassay and bone alkaline phosphatase (BAP; Quidel, San Diego, CA) was measured by an enzyme immunoassay. As bone resorption markers, serum levels of the N-telopeptide crosslinked of type I collagen (NTX; Inverness, Princeton, NJ) and tartrate-resistant acid phosphatase isoform 5b (TRACP-5b; DS Pharma Biomedical Co., Ltd., Tokyo, Japan) were measured by enzyme immunoassay.

Measurement of bone mineral density (BMD)

Before starting glucocorticoid therapy, the BMD of the lumbar spine (L2–4) was measured by dual-energy X-ray absorptiometry using Discovery A (Hologic, Waltham, MA), being automatically calculated from the bone area (square centimeters) and bone mineral content (grams) and was expressed in grams per square centimeters. BMD was measured again after 15 ± 4 months of glucocorticoid treatment and the percent change from baseline was calculated.

Cell culture

Normal human osteoblasts (NHOst) were obtained from Lonza Inc. (Williamsport, PA) and were maintained at 37 C in osteoblast growth medium (Lonza) supplemented with 10% fetal bovine serum, 50 μg/ml gentamycin sulfate, 2.5 μg/ml amphotericin-B, and 50 mm l-ascorbic acid under a humidified atmosphere of 5% CO₂ in air. NHOst were resuspended in 5 ml of osteoblast growth medium supplemented with 1% (vol/vol) fetal bovine serum at 1.0 × 10⁵ cells/ml and then cultured for 24 h in 35-mm dishes.

Measurement of RANKL and OPG production

To evaluate the effect of dexamethasone on RANKL and OPG production, NHOst were incubated with or without 100 ng/ml IL-6 (R&D Systems Inc., Minneapolis, MN) plus 100 ng/ml soluble IL-6 receptor (sIL-6R; R&D Systems) for 24 h and then were coincubated with dexamethasone (0, 10, 100, or 1000 nm). Culture supernatants were collected, centrifuged, and stored at −80 C for subsequent analysis. RANKL and OPG concentrations in the culture supernatants were measured in duplicate by ELISA (Biomedica) according to the instructions of the manufacturer.

Assessment of RANKL and OPG gene expression

NHOst were cultured under various conditions in medium containing 1% (vol/vol) fetal bovine serum, and RNA was extracted by using an RNeasy minikit (QIAGEN GmbH, Hilden, Germany) according to the manufacturer's instructions. RT-PCR was performed with a SuperScript first-strand synthesis system (Invitrogen Corp., Carlsbad, CA) according to the manufacturer's instructions, using 1 μg total cellular RNA as the template. Equal amounts of each reverse-transcribed product were amplified by PCR with HotStarTaq polymerase (QIAGEN). The sequences of the primers were as follows: RANKL, sense 5′-GCCAGTGGGAGATGTTAG-3′ and antisense 5′-TTAGCTGCAAGTTTTCCC-3′; OPG, sense 5′-GCTAACCTCACCTTCGAG-3′, and antisense 5′-TGATTGGACCTGGTTACC-3′; β-actin (endogenous control), sense 5′-CCTCGCCTTTGCCGATCC-3′, and antisense 5′-GGATCTTCATGAGGTAGTCAGTC-3′. The RANKL to β-actin and OPG to β-actin ratios were determined for semiquantification of mRNA expression. PCR for amplification of RANKL cDNA (486 bp) involved 40 cycles of 94 C for 30 sec, 55 C for 30 sec, and 72 C for 30 sec, whereas 30 cycles under the same conditions were used to amplify OPG cDNA (324 bp). PCR for β-actin cDNA (626 bp) was performed with 28 cycles of 95 C for 30 sec, 56 C for 30 sec, and 72 C for 30 sec. The amplified cDNA fragments were resolved by electrophoresis on 2% (wt/vol) agarose gel and were detected under UV light using an LAS-3000 (Fujifilm Corp., Tokyo, Japan) after staining of the gel with ethidium bromide.

Statistical analysis

Statistical analysis was performed with Prism version 5.0 software (GraphPad Software, San Diego, CA). Numerical data were expressed as both the mean ± sem and the median with the interquartile range. An assessment of the changes during glucocorticoid treatment was performed with a Friedman's test followed by a Dunnett's multiple comparison test. To compare two groups, the Mann-Whitney U test or Student's t test was applied for numerical data and the χ² test (or Fisher's exact test) was used for categorical data. Multiple comparisons were performed by the Kruskal-Wallis test. The level of significance was set at P < 0.05.

Results

Profile of the subjects

Table 1 shows the demographic and clinical data of the patients and healthy controls. There were no significant differences of body mass index (BMI) and BMD between the patients and the matched controls. As shown in Table 1, serum CRP and IL-6 levels were significantly increased, whereas the serum 25-OHD level was significantly decreased in the patients with systemic autoimmune diseases before glucocorticoid therapy when compared with those in the healthy controls. Nine of 60 patients were withdrawn during 15 months due to death (n = 5) or hospital transfer (n = 4). Fifty-one patients continued to receive prednisolone at 15 months except for a patient with AOSD. The mean daily and cumulative doses of prednisolone during 15 months were 9.5 ± 4.0 (sem) and 9296 ± 372 mg, respectively. There were no significant differences in the mean initial, daily, and cumulative doses of prednisolone between the higher and lower sRANKL groups (Table 2).

Table 1.

Demographics and clinical data of the study population at baseline

	Patients (n = 60)	Healthy controls (n = 60)
Age (yr)	55.2 ± 2.7	55.2 ± 1.5
Number of males/females	20/40	20/40
Postmenopausal women (%)	23 (57.5)	23 (57.5)
BMI (kg/m2)	21.7 ± 0.6	22.1 ± 2.6
BMD (g/cm2)	0.96 ± 0.03	0.92 ± 0.02
Serum markers
CRP (mg/dl)	4.11 ± 0.61a	0.05 ± 0.00
	2.15 [0.30–7.05]b	0.05 [0.04–0.07]
IL-6 (pg/ml)	34.4 ± 5.7a	2.2 ± 0.2
	17.2 [8.7–37.4]b	1.8 [1.4–2.5]
25-OHD (ng/ml)	16.8 ± 1.1a	26.2 ± 0.9
	16.0 [10.0–21.0]b	26.0 [22.0–32.0]
OC (ng/ml)	3.30 ± 0.49a	6.18 ± 0.31
	2.35 [1.35–3.68]b	5.70 [4.35–7.68]
ucOC (ng/ml)	1.56 ± 0.20a	3.84 ± 0.28
	1.01 [0.55–2.31]b	3.48 [2.31–4.86]
BAP (U/liter)	16.0 ± 0.8	20.8 ± 1.0
	14.3 [11.5–19.3]	19.4 [14.3–24.4]
PINP (μg/liter)	44.5 ± 3.4	54.7 ± 2.6
	37.6 [27.1–51.9]	55.2 [39.8–70.0]
NTX (nmolBCE/liter)	15.1 ± 0.6	14.6 ± 0.9
	14.7 [11.9–16.7]	13.6 [11.3–17.6]
TRACP-5b (mU/dl)	198.7 ± 12.6	182.7 ± 10.8
	183 [136–227]	178 [121–244]

Data are mean ± sem and median [25th to 75th percentile range].

^aP < 0.05 for the patients compared with healthy controls by Student's t test.

^bP < 0.05 for the patients compared with healthy controls by Mann-Whitney U test.

Table 2.

Demographics and clinical data of higher and lower sRANKL groups

	Higher sRANKL group (n = 20)	Lower sRANKL group (n = 40)
Age (yr)	53.2 ± 4.2	56.2 ± 2.7
Number of males/females	6/14	14/26
Postmenopausal women (%)	7 (50.0)	16 (61.5)
BMI (kg/m2)	20.6 ± 0.7	22.3 ± 0.7
BMD (g/cm2)	0.96 ± 0.04	0.95 ± 0.03
Systemic autoimmune disease
SLE	8/21 (38%)	13/21 (62%)
PM/DM	1/15 (7%)	14/15 (93%)
Vasculitis syndrome	7/19 (37%)	12/19 (63%)
AOSD	4/5 (80%)	1/5 (20%)
Serum markers
CRP (mg/dl)	5.62 ± 1.03a	3.91 ± 0.93
	5.10 [2.20–7.75]b	1.55 [0.23–4.15]
IL-6 (pg/ml)	43.2 ± 12.2	31.1 ± 6.5
	24.5 [9.9–66.1]	15.8 [7.3–3.4]
sRANKL (pmol/liter)	0.434 ± 0.099a	0.018 ± 0.005
	0.251 [0.178–0.479]b	0.000 [0.000–0.023]
OPG (pmol/liter)	6.63 ± 0.56	5.62 ± 0.34
	5.95 [4.52–8.21]	5.15 [4.16–6.83]
25-OHD (ng/ml)	15.2 ± 1.4	17.8 ± 1.4
	14.5 [10.0–17.8]	17.0 [10.5–22.0]
OC (ng/ml)	1.93 ± 0.34	2.56 ± 0.23
	1.6 [1.0–2.8]	2.3 [1.5–3.7]
ucOC (ng/ml)	1.24 ± 0.36	1.66 ± 0.23
	0.73 [0.41–1.35]	1.18 [0.59–2.49]
BAP (U/liter)	14.6 ± 1.4	16.7 ± 1.0
	12.9 [10.7–14.6]	15.7 [12.8–20.1]
PINP (μg/liter)	39.4 ± 4.1	47.1 ± 4.6
	30.6 [25.4–51.7]	39.2 [29.9–57.0]
NTX (nmolBCE/liter)	14.4 ± 1.0	15.4 ± 0.7
	14.5 [11.4–16.1]	15.0 [12.0–17.3]
TRACP-5b (mU/dl)	190.9 ± 24.2	201.5 ± 13.7
	181 [103–220]	185 [137–238]
Initial glucocorticoid dose (mg/d)	45.0 ± 2.0	45.8 ± 1.6
Mean glucocorticoid dose (mg/d) at 15 months	9.0 ± 0.9	9.7 ± 0.7
Cumulative glucocorticoid dose (mg)	9210 ± 566	9319 ± 449

Data are mean ± sem and median [25th to 75th percentile range].

^aP < 0.05 for higher sRANKL group compared with lower sRANKL group by Student's t test.

^bP < 0.05 for higher sRANKL group compared with lower sRANKL group by Mann-Whitney U test.

Serum sRANKL, OPG, and inflammatory markers

As shown in Fig. 1A, the serum sRANKL level of the patients with systemic autoimmune diseases before glucocorticoid therapy was significantly higher than that of the healthy controls. The mean serum sRANKL level of the patients was unchanged by glucocorticoid therapy (Fig. 1B). There was no significant difference in serum sRANKL level [median (25th to 75th percentile range)] [SLE, 0.034 (0.000–0.186), PM/DM, 0.003 (0.000–0.085), vasculitis syndrome, 0.001 (0.000–0.178), and AOSD, 0.229 (0.094–0.366) pmol/liter, Kruskal-Wallis test, P = 0.2378] among these systemic autoimmune diseases. As shown in Fig. 1C, the serum OPG level of the patients before glucocorticoid therapy was significantly higher than that of the healthy controls, whereas the mean serum OPG level of the patients decreased significantly after starting glucocorticoid therapy (Fig. 1D). There was no significant difference in serum OPG level [SLE, 5.23 (3.90–6.50), PM/DM, 4.59 (4.38–5.61), vasculitis syndrome, 5.83 (4.48–9.31), and AOSD, 5.30 (3.80–3.73) pmol/liter, P = 0.2413] among these systemic autoimmune diseases.

Fig. 1.

Serum levels of sRANKL (A and B), OPG (C and D), CRP (E), and IL-6 (F) in the patients and healthy controls. Data are expressed as the median with 25th to 75th percentiles in brackets (A and C). *, P < 0.05; **, P < 0.0001 by the Mann-Whitney U test. Changes of serum sRANKL (B), OPG (D), CRP (E), and IL-6 (F) during glucocorticoid therapy are shown. Data are expressed as the mean ± sem (B, D, E, and F). **, P < 0.0001 by Friedman's test followed by Dunnett's multiple comparison test.

As shown in Fig. 1 (E and F), the mean serum CRP and IL-6 levels of the patients were significantly decreased by glucocorticoid therapy. There were significant differences in the serum levels of CRP [SLE, 1.4 (0.2–2.9), PM/DM, 0.3 (0.1–1.5), vasculitis syndrome, 7.2 (3.1–12.2), and AOSD, 7.9 (0.6–14.4) mg/dl, P < 0.0001] and IL-6 [SLE, 12.0 (7.3–21.6), PM/DM, 11.0 (6.4–17.4), vasculitis syndrome, 46.1 (18.0–121.0), and AOSD, 26.2 (18.5–147.1) pg/ml, P < 0.0005] among these diseases at baseline.

Serum sRANKL stratification

Because we found that the distribution of serum sRANKL was bimodal, the patients were stratified into two groups. A cutoff level for the definition of the higher serum sRANKL group was 0.16 pmol/liter. With this definition, the number of the higher sRANKL group was 20 patients, whereas that of the lower sRANKL group was 40 patients, respectively. Table 2 compares the demographic and clinical data of the higher and lower sRANKL groups. There were no significant differences of sex, menopausal status, age, BMI, BMD, serum IL-6, serum 25-OHD, serum OPG, bone formation markers (OC, ucOC, BAP, and PINP), and bone resorption markers (NTX and TRACP-5b) between the higher and lower sRANKL groups. Baseline serum CRP levels were significantly elevated in the higher sRANKL group compared with those in the lower sRANKL group (Table 2). The percentage of patients with SLE and PM/DM in the higher sRANKL group tended to be lower than those patients with vasculitis syndrome and AOSD.

As shown in Fig. 2A, the mean serum sRANKL level of the higher sRANKL group was significantly decreased by glucocorticoid therapy. In contrast, the mean serum sRANKL level of the lower sRANKL group showed a significant increase after the start of glucocorticoid therapy (Fig. 2B).

Fig. 2.

Changes of serum sRANKL in the higher (A) and lower (B) sRANKL groups during glucocorticoid therapy. Data are expressed as the mean ± sem. *, P < 0.05 by Friedman's test followed by Dunnett's multiple comparison test.

Bone turnover markers

As shown in Table 1, the baseline serum OC and ucOC levels of the patients were significantly lower than those of the healthy controls. However, the baseline levels of the other bone formation markers (BAP and PINP) and bone resorption markers (NTX and TRACP-5b) were not significantly different between the patients and healthy controls.

Bone mineral density

There was no significant difference of baseline in BMD between the higher and lower sRANKL groups (Table 2). As shown in Fig. 3, BMD [median (25th to 75th percentile range)] increased from baseline by 0.93% (−1.23–5.51%) with glucocorticoid therapy in the higher sRANKL group, whereas it decreased by −2.03% (−5.99 to 0.38%) in the lower sRANKL group. The difference in the change of BMD between the two groups was statistically significant (P < 0.05). There were no statistically significant differences in the baseline BMD (mean ± sem, SLE, 1.02 ± 0.05; PM/DM, 0.91 ± 0.04; vasculitis syndrome, 0.94 ± 0.04; AOSD, 0.94 ± 0.18 g/cm², P = 0.3744) and in the changes of BMD [median (25th to 75th percentile range)] [SLE, −2.77 (−6.05 to −0.26), PM/DM, −1.91 (−13.02 to 4.83), vasculitis syndrome, 1.00 (−3.21 to 5.51), and AOSD, −0.82% (−2.09 to 0.45%), P = 0.2580] among systemic autoimmune diseases.

Fig. 3.

Changes of BMD due to glucocorticoid therapy in the higher and lower sRANKL groups. Data are expressed as the median with the 25th to 75th percentiles in brackets. P < 0.05 by the Mann-Whitney U test.

RANKL and OPG mRNA expression

We also investigated the effect of dexamethasone on RANKL and OPG mRNA expression in vitro. As shown in Fig. 4 (A and B), RANKL mRNA expression was increased by addition of dexamethasone to cultures of unstimulated NHOst. RANKL mRNA expression was also increased by incubation of these cells with IL-6 plus soluble IL-6 receptor, whereas IL-6-stimulated up-regulation of RANKL mRNA expression was suppressed by the addition of dexamethasone.

Fig. 4.

Effect of dexamethasone and IL-6 on RANKL and OPG expression by NHOst. RANKL, OPG, and β-actin mRNA were detected in NHOst by RT-PCR (A). RANKL mRNA expression was normalized by that of β-actin mRNA (B). OPG mRNA expression was normalized by that of β-actin mRNA (C). Detection of OPG protein was by ELISA (D).

As shown in Fig. 4 (A and C), OPG mRNA expression was decreased by addition of dexamethasone to unstimulated NHOst, whereas it was increased by incubation with IL-6 plus soluble IL-6 receptor. The IL-6-stimulated up-regulation of OPG mRNA expression was attenuated by dexamethasone.

RANKL and OPG protein expression

We attempted to measure RANKL protein levels in NHOst culture supernatants by ELISA, but RANKL was not detectable with the assay kit. As shown in Fig. 4D, the OPG protein level in the culture medium of unstimulated NHOst was decreased by incubation with dexamethasone. Secretion of OPG protein was significantly increased by IL-6 plus soluble IL-6 receptor, whereas IL-6-stimulated increase of OPG protein secretion was suppressed by addition of dexamethasone.

Discussion

In the present study, the mean serum sRANKL level of all 60 patients with systemic autoimmune diseases did not change after the initiation of glucocorticoid therapy. This finding is similar to that reported by von Tirpitz et al. (18), who did not detect any significant change of serum sRANKL after 1–12 wk of glucocorticoid therapy in 27 patients with Crohn's disease. In both their subjects and ours, the distribution of baseline serum sRANKL levels was skewed and the range was wide. Unlike von Tirpitz et al., we also carried out a more detailed investigation after dividing the patients into two groups with higher and lower baseline serum sRANKL levels. This revealed a difference in the response to glucocorticoid therapy, with the serum sRANKL level decreasing in the higher sRANKL group and increasing in the lower sRANKL group.

At present, it is not clear whether sRANKL has effects on bone metabolism that are similar to the actions of RANKL in vitro. sRANKL is a truncated ectodomain that is released from the cell-bound form by enzymatic cleavage and is a soluble molecule similar to that produced by recombinant technology or that secreted by activated T cells (19). Injection of sRANKL causes a rapid increase of serum calcium levels due to both activation of osteoclasts and enhanced generation of these cells (20). Therefore, it is possible that the changes of serum sRANKL detected in this study had a clinically important influence on bone metabolism.

Our study showed that BMD increased after the start of glucocorticoid therapy in the higher sRANKL group, whereas it decreased in the lower sRANKL group. From this result and the differing behavior of serum sRANKL in the two groups, the risk of osteoporosis related to glucocorticoid therapy seems to be greater in the lower sRANKL group. Our findings also suggest that the risk of osteoporosis might be predicted by measuring the serum sRANKL before glucocorticoid therapy. All patients in this study received a bisphosphonate. Mechanism of action of bisphosphonate is considered to improve bone resorption by the suppression of osteoclast. All patients should have a favorable result on increasing BMD; however, the changes of BMD were different from each other. We then suggested that the differences in the changes of BMD during glucocorticoid therapy were caused at least in part by the changes of serum sRANKL.

CRP was also significantly higher in the higher sRANKL group compared with that in the lower sRANKL group. It is known that CRP tends to be low in patients with SLE (21) and PM/DM (22), regardless of disease activity. Because patients with SLE or PM/DM dominated the lower sRANKL group, the difference of CRP between the two sRANKL groups might have been due to a difference in the mix of systemic autoimmune diseases. In contrast, serum IL-6 level had a trend to be increased, but the difference was not statistically significant between the two sRANKL groups. It is well established that the elevated serum IL-6 levels in systemic autoimmune diseases may be more useful index for active disease than CRP (23). We showed that no statistically significant difference was observed in the serum sRANKL level among these diseases. The mechanisms of serum sRANKL production and regulation remained to be studied.

Our in vitro study demonstrated dual regulation of RANKL mRNA expression by dexamethasone in NHOst with or without IL-6 stimulation. In general, glucocorticoids enhance RANKL mRNA expression by osteoblast-like cancer cell lines (12–15). Regarding the mechanism involved in the increase of RANKL mRNA promoted by dexamethasone, Kondo et al. (15) reported that the expression of RANKL mRNA was up-regulated through the glucocorticoid-responsive element half-site or the activator protein-1 site. However, there have been no previous reports about the suppressive effect of dexamethasone on cytokine-stimulated osteoblasts that we detected in this study. We found that RANKL mRNA expression was up-regulated by adding IL-6 to cultures of osteoblasts, which was similar to the report of O'Brien et al. (24). We also found that dexamethasone suppressed this up-regulation of RANKL mRNA by IL-6, and these in vitro findings corresponded to the results of our clinical study.

The mean serum OPG level of our patients with systemic autoimmune diseases showed a significant decrease after they started glucocorticoid therapy. Similarly, it has been reported that serum OPG levels are significantly decreased by glucocorticoid therapy in patients with Crohn's disease (18) or chronic glomerulonephritis (25). Moreover, our in vitro studies showed that the expression of OPG mRNA by cultured NHOst was decreased after adding dexamethasone, irrespective of whether the cells received IL-6 stimulation. Because a decrease of OPG promotes osteoclast differentiation and activation, our experimental and clinical data concerning OPG may help to explain the mechanism of glucocorticoid-induced osteoporosis.

In the present study, the serum 25-OHD level of the patients with systemic autoimmune diseases before glucocorticoid therapy was significantly lower than that of the healthy controls. Cutolo (26) reported that serum 25-OHD level was decreased in patients with rheumatoid arthritis and SLE when compared with that in healthy controls. It was suggested that low serum levels of 25-OHD might be partially related to prolonged daily darkness, different genetic background, and nutritional factors in these patients.

We found that serum OC and ucOC levels were already low before the initiation of glucocorticoid therapy in our patients with active systemic autoimmune diseases. Kim et al. (27) reported that the serum OC level tended to decrease in healthy subjects who were immobilized for 14 d, so it might be possible that the relative immobility of patients with active disease contributes to their lower serum OC and ucOC levels.

In conclusion, the OPG/RANKL/RANK system seems to have an important role in the mechanism of glucocorticoid-induced osteoporosis, and sRANKL might be a useful marker of bone remodeling in patients with systemic autoimmune disease receiving glucocorticoid therapy.