Osteoprotegerin (OPG)/osteoclastogenesis inhibitory factor (OCIF) is a soluble member of the tumor necrosis factor receptor family of proteins and plays an important role in the negative regulation of osteoclastic bone resorption. Whether OPG/OCIF circulates in human blood and how its level changes under pathological conditions is not known. To address these issues, a panel of monoclonal antibodies was generated against recombinant OPG/OCIF and screened for reactivity with solid-phase monomeric and homodimeric forms of the recombinant protein. Antibodies that showed high affinity for both forms of OPG/OCIF and those that selectively recognized the homodimer were identified, enabling development of two types of sensitive enzyme-linked immunosorbent assay (ELISA): one that detects both forms of OPG/OCIF equally and one specific for the homodimer. Characterization of circulating OPG/OCIF with these ELISAs revealed that the protein exists in human serum mainly in the monomeric form. The serum concentration of OPG/OCIF increased with age in both healthy Japanese men and women, and was significantly higher in postmenopausal women with osteoporosis than in age-matched controls. Within the osteoporotic group, serum OPG/OCIF concentrations were higher in patients with low bone mass. Serum OPG/OCIF concentrations were also significantly increased in those postmenopausal women with a high rate of bone turnover, as determined by increased serum bone-specific alkaline phosphatase and urinary excretion of pyridinoline and deoxypyridinoline. The results suggested that circulating OPG/OCIF levels are regulated by an age-related factor(s) and that the increased serum concentration may reflect a compensative response to enhanced osteoclastic bone resorption and the resultant bone loss rather than a cause of osteoporosis.
The quality and quantity of bone are maintained by a continuous and coordinated process of remodeling that involves bone resorption and subsequent bone formation. Multinucleated osteoclasts are primarily responsible for bone resorption.(1-3) Differentiation and activation of osteoclasts are controlled by systemic hormones as well as by local factors produced by various cell types in the bone microenvironment.(1-3) Although several factors, including PU.1,(4) macrophage colony-stimulating factor,(5) c-Fos,(6) microphthalmia locus-associated transcription factor,(7) NF-κB,(8) c-Src,(9) and cathepsin K,(10) have been identified as those playing important roles in the development and function of osteoclasts, mostly on the basis of the osteopetrotic phenotype of the respective gene knockout mice, the molecular mechanism of osteoclastogenesis is not fully understood.
Osteoporosis is a systemic skeletal disease characterized by excessive bone resorption, typically in association with estrogen deficiency after menopause, and leads to low bone mass and microstructural deterioration.(11) Various cytokines that regulate osteoclastic bone resorption, including tumor necrosis factor (TNF),(12) interleukin-1,(12) interleukin-6,(13) and transforming growth factor-β,(14) are thought to be important in the pathogenesis of osteoporosis, although the precise mechanisms of their contributions remain to be determined.
Osteoclastogenesis inhibitory factor (OCIF) is a cytokine that we recently purified from the conditioned medium of human IMR-90 embryonic fibroblasts and that specifically inhibits osteoclastogenesis.(15) Complementary DNA encoding OCIF has been cloned, and analysis of the cDNA sequence revealed that OCIF is a soluble member of the TNF receptor (TNFR) family of proteins.(16) Simonet et al.(17) independently cloned a cDNA encoding osteoprotegerin (OPG), a protein that contributes to the regulation of bone density. Comparison of the predicted amino acid sequences of OCIF and OPG revealed that the two proteins are identical. The systemic administration of recombinant OPG (rOPG)/rOCIF results in a marked increase in bone mineral density (BMD), associated with a decrease in the number of active osteoclasts, in normal rats,(16) and it both prevents bone loss and restores bone strength in ovariectomized rats.(17) The results of our in vitro studies suggest that OPG/OCIF inhibits osteoclastogenesis by interrupting the intercellular signaling between osteoblastic stromal cells and osteoclast progenitors.(16)
Although OPG/OCIF is capable of inhibiting bone resorption in vivo,(16,17) its roles in normal bone remodeling as well as in the pathogenesis of osteoporosis remain unknown. To address these important issues, we have established a panel of monoclonal antibodies (MAbs) to human rOPG/rOCIF and developed two types of sensitive enzyme-linked immunosorbent assays (ELISA): one that can detect both the homodimeric and monomeric forms of OPG/OCIF and one that specifically detects the homodimer. With these assay systems, we have demonstrated that OPG/OCIF circulates in human blood mainly as a monomer, that serum concentrations of OPG/OCIF increase with age in both healthy men and women, and that these concentrations are significantly higher in postmenopausal women with osteoporosis than in age-matched normal controls.
MATERIALS AND METHODS
Homodimeric and monomeric forms of human rOPG/rOCIF were purified from conditioned medium of Chinese hamster ovary cells by successive chromatography on HiLoad/SP (Pharmacia, Uppsala, Sweden), Sulfate-Cellulofine-m (Seikagaku Kogyo, Tokyo, Japan), and Resource S (Pharmacia) columns, respectively.(18) A murine myeloma cell line, P3 × 63Ag.8.653, and a murine osteoblastic cell line, ST2, were obtained from American Type Culture Collection (Rockville, MD, U.S.A.) and Riken Cell Bank (Ibaraki, Japan), respectively. Normal human serum was obtained from BioWhittaker (Walkersville, MD, U.S.A.). A substrate for peroxidase (POD), tetramethylbenzidine (TMB) soluble reagent, and TMB stop buffer were from ScyTek (Logan, UT, U.S.A.).
Preparation of MAbs to OCIF
A mixture of the monomeric and the homodimeric forms of rOPG/rOCIF (1:1) was emulsified with an equal volume of complete Freund's adjuvant (Difco, Detroit, MI, U.S.A.) and then injected intraperitoneally into 8-week-old BALB/c mice (Charles River Japan, Yokohama, Japan) at a dose of 10 μg per mouse (in a volume of 200 μl) three times at 1-week intervals. Seven days after the third injection, the rOPG/rOCIF emulsion was injected at a dose of 5 μg of rOPG/rOCIF per mouse. Four days later, the animals received an intravenous administration of rOPG/rOCIF in PBS at a dose of 5 μg per mouse. Titers of antibodies to OPG/OCIF in the blood of the immunized animals were determined with a solid-phase ELISA (see below).
Splenocytes from the animals showing high titers of antibodies were fused to P3 × 63Ag.8.653 mouse myeloma cells, as previously described.(19,20) Hybridomas secreting MAbs were screened with ELISA systems based on solid-phase monomeric and homodimeric forms of rOPG/rOCIF. Ninety-six-well plates (MaxiSorp; Nalge Nunc, Reskilde, Denmark) were coated with 50 μl of either the homodimeric or the monomeric form of rOPG/rOCIF (2 μg/ml in 0.1 M sodium bicarbonate, pH 9.6) at 4°C overnight, and then incubated for 2 h at room temperature with 300 μl of a blocking solution containing 50% BlockAce (Snow Brand Milk Products, Tokyo, Japan). After washing three times with phosphate-buffered saline (PBS) containing 0.1% polysorbate 20 (PBS-P), both 40 μl of calf serum and 10 μl of hybridoma-conditioned medium were added to each well, and the plates were incubated for 2 h at room temperature. After washing five times with PBS-P, 50 μl of POD-labeled goat anti-mouse immunoglobulin G (IgG) antibody (KPL, Gaithersburg, MD, U.S.A.) diluted 5000-fold with PBS containing 25% BlockAce were added to each well, and the plates were incubated for 2 h at room temperature. After washing six times with PBS-P, 100 μl of TMB substrate reagent was added to each well, and the plates were incubated for 5 minutes at room temperature. TMB stop buffer (100 μl) was then added to each well, and absorbance at 450 nm was measured with a Microplate Reader NJ-2000 (Nippon InterMed, Tokyo, Japan).
The dissociation constant (Kd) of each MAb for the homodimeric or the monomeric form of rOPG/rOCIF was determined with a liquid-phase ELISA as previously described,(21) and was calculated from the equation
where a0 is the concentration of antigen (OPG/OCIF homodimer or the monomer) and A0 and A represent the absorbances measured for each MAb in the absence and presence of antigen, respectively.
To test the ability of MAbs to neutralize OPG/OCIF activity, both MAbs and rOPG/rOCIF were dissolved in α-minimal essential medium containing 10% fetal bovine serum to final concentrations of 500 ng/ml and 10 ng/ml, respectively, and incubated at 37°C for 1 h. The biological activity of OPG/OCIF was evaluated by the ability to suppress the formation of osteoclast-like cells in cocultures of ST2 mouse osteoblastic stromal cells and spleen cells in the presence of 10 nM 1α,25-dihydroxyvitamin D3 and 100 nM dexamethasone.(15)
Development of ELISAs for OPG/OCIF
An anti-OPG/OCIF MAb designated OI-19 was dissolved in 0.1 M sodium bicarbonate (pH 9.6) to a final concentration of 10 μg/ml, and 100 μl of the solution were added to each well of 96-well plates (MaxiSorp; Nalge Nunc). The plates were maintained at 4°C overnight, after which the solution in each well was replaced with 300 μl of a blocking solution (50% BlockAce in water) and the plates were incubated for 2 h at room temperature. After washing three times with PBS-P, 100 μl of rOPG/rOCIF standard or test sample, prepared by serial dilution with 0.2 M Tris-HCl (pH 7.4) containing 40% BlockAce, normal mouse IgG (10 μg/ml; Cappel, Aurora, OH, U.S.A.), and 0.1% polysorbate 20, was added to each well. The plates were incubated for 2 h at room temperature and then washed six times with PBS-P, after which 100 μl of POD-labeled anti-OPG/OCIF MAb designated OI-4, diluted 1000-fold with 0.1 M Tris-HCl (pH 7.4) containing 25% BlockAce, normal mouse IgG (10 μg/ml), and 0.1% polysorbate 20, was added to each well. After incubation for an additional 2 h at room temperature, the plates were washed six times with PBS-P, 100 μl of TMB substrate reagent was added to each well, and the plates were incubated for 20 minutes at room temperature. TMB stop buffer (100 μl) was added to each well, and absorbance at 450 nm was measured with a Microplate Reader (Nippon InterMed).
An ELISA that specifically detects the homodimeric form of OPG/OCIF was constructed as described above, with the exception that a MAb designated OI-26 was used in place of OI-19 as the solid-phase antibody.
The study population comprised a total of 242 consecutive unrelated Japanese subjects, including 56 men aged 25–93 years (54.6 ± 19.7 years, mean ± SD) and 186 women aged 23–90 years (64.9 ± 13.9 years), who visited outpatient clinics or were admitted to National Chubu Hospital (Obu City, Aichi Prefecture, Japan) between April 1996 and September 1997. Excluded from the study were subjects with disorders known to cause abnormalities of bone metabolism, including diabetes mellitus; thyroid, parathyroid, and other endocrinological diseases; renal diseases; and rheumatoid arthritis. Individuals who had taken drugs such as estrogen, progesterone, glucocorticoids, bisphosphonates, and active vitamin D3 were also excluded. Control subjects were selected from healthy volunteers recruited in the Obu City area or individuals who visited the hospital and did not show any serious diseases or take drugs known to affect bone and calcium metabolism. The study protocol was approved by the Committee on the Ethics of Human Research at National Chubu Hospital, and written consent was obtained from all subjects.
Measurement of BMD
BMD at the lumbar spine (L2–L4) was measured by dual-energy X-ray absorptiometry (DXA) with a DPX instrument (Lunar Corp., Madison, WI, U.S.A.); the coefficient of variation (CV) of these measurements was <1.5%. All DXA scans were reviewed by orthopedic surgeons, and aortic calcification and osteoarthritic changes were evaluated by plain X-ray films before DXA. To obtain accurate data for lumbar spine BMD, BMD values in regions where aortic calcification and/or osteoarthritic changes had been marked were excluded. The diagnosis of osteoporosis was based on the criteria recommended by the Committee for Diagnostic Criteria of Osteoporosis of the Japanese Society for Bone and Mineral Research,(22) which are either an L2–L4 BMD of <80% of the young adult (20–44 years of age) reference mean in the presence of nontraumatic vertebral fracture, or an L2–L4 BMD of <70% of the young adult reference mean in the absence of nontraumatic vertebral fracture. Individuals with an L2–L4 BMD of 70–80% of the young adult reference mean in the absence of vertebral fracture were diagnosed with osteopenia. On the basis of these criteria, 73 subjects were diagnosed with osteoporosis and 39 individuals with osteopenia. The L2–L4 BMD of normal controls was >80% of that of the young adult reference mean.
Measurement of biochemical markers of bone turnover
Venous blood and urine samples were collected in the early morning after an overnight fast. Blood samples were centrifuged at 1600g for 15 minutes at 4°C, and serum was separated and stored at −30°C until assay. Urine samples were collected in plain tubes and stored at −30°C.
As markers of bone resorption, urinary pyridinoline (Pyr) and deoxypyridinoline (Dpyr) were measured by high-performance liquid chromatography with a fluorescence detector.(23) The minimum detection limit for both substances was 0.166 pmol/ml, and the intra- and interassay CV values were ≤3.5 and ≤11.7% for Pyr, and ≤8.4 and ≤10.5% for Dpyr, respectively. The concentrations were corrected for urinary creatinine (Cr) and expressed as picomoles per micromole of Cr. The serum concentration of osteocalcin was measured with an immunoradiometric assay kit (Mitsubishi Chemical, Tokyo, Japan). The detection limit of this assay system was 1 ng/ml, and the intra- and interassay CV values were ≤3.7 and ≤5.1%, respectively. The activity of serum bone-specific alkaline phosphatase (BAP) was measured with an immunoassay kit (Metra Biosystems, Mountain View, CA, U.S.A.). The cross-reactivity with serum isoenzymes of ALP from liver, placenta, and intestine was 3–8, 0, and 0.4%, respectively. The detection limit of the assay system was 0.7 U/l, and the intra- and interassay CV values were ≤5.8 and ≤7.6%, respectively.
Data are expressed as means ± SD. Correlation between the serum concentration of OPG/OCIF and age, BMD, or biochemical markers of bone metabolism was detected by simple regression analysis. Clinical and laboratory data were compared among normal controls and subjects with osteopenia or osteoporosis by one-way analysis of variance and Scheffe's multiple range test. Unpaired Student's t-test was used to compare data between two groups. Qualitative data were compared by the chi-square test. A level of p < 0.05 was considered statistically significant.
Preparation and characterization of MAbs to OPG/OCIF
Approximately 300 hybridomas established from five separate fusions were found to produce MAbs to OPG/OCIF, and 33 of these MAbs that displayed substantial binding, even at high dilution of conditioned medium, in the solid-phase ELISA system were selected (Table 1). Antibodies OI-3 and OI-27 are classified as IgG2b and IgG2a, respectively, whereas the others are all of the IgG1 isotype. The Kd values of the established MAbs ranged from 0.00175 to 148 nM for the homodimeric form of rOPG/rOCIF and from 0.107 to 2530 nM for the monomer (Table 1). Some of the MAbs (for example, OI-4, OI-14, OI-19, OI-20, and OI-35) showed high affinity for both forms of rOPG/rOCIF, whereas others (for example, OI-12, OI-13, OI-21, OI-26, and OI-30) recognized the homodimer preferentially or selectively. Only OI-1 cross-reacted with murine rOPG/rOCIF (data not shown). Some of the MAbs reacted with ΔD567,(24) an OPG/OCIF mutant that lacks three (domains 5–7) of the seven domains of the molecule, indicating that the epitopes recognized by these MAbs are located in domains 1–4. Two MAbs, OI-1 and OI-4, were capable of neutralizing OPG/OCIF activity by >80% (Table 1).
Table Table 1.. Characterization of Mabs Generated Against Human rOPG/rOCIF
Development of ELISAs for OPG/OCIF
We attempted to select suitable combinations of MAbs for two-site ELISAs for OPG/OCIF. When each MAb was tested after conjugation with POD, OI-4 was the most suitable for the detection of rOPG/rOCIF captured by all MAbs except OI-4 itself. With regard to the capture antibodies, OI-19 and OI-26 were most suitable for the recognition of both forms of rOPG/rOCIF and the homodimeric form, respectively. Thus, an ELISA for “total” OPG/OCIF, which detects both the homodimer and the monomer, was constructed with OI-19 as the capture antibody and OI-4 as the POD-labeled antibody; the detection range for total OPG/OCIF was 32.5–500 pg/ml (Fig. 1, upper panel). An ELISA that specifically detects the homodimeric form of OPG/OCIF was constructed with OI-26 as the capture antibody and OI-4 as the POD-labeled antibody (Fig. 1, lower panel). When the minimum detection limit was determined as the lowest concentration of rOPG/rOCIF exhibiting an absorbance twice that of the blank, the sensitivity of both ELISA systems averaged 32.5 pg/ml. Given that human serum samples were diluted with an equal volume of assay buffer [0.2 M Tris-HCl, pH 7.4, containing 40% BlockAce, normal mouse IgG (10 μg/ml), and 0.1% polysorbate 20] to minimize the effect of potential interfering substances, concentrations of total OPG/OCIF or the homodimer as low as 65 pg/ml can be measured.
To estimate the relative proportions of monomeric and dimeric OPG/OCIF in human blood, we diluted the homodimeric standard with either assay buffer alone or buffer containing 50% normal human serum and subjected the resulting samples to both ELISA systems (Fig. 2). The presence of 50% human serum in the assay buffer resulted in an upward shift in the standard curve in the ELISA for total OPG/OCIF (Fig. 2, upper panel), the extent of the shift corresponding to the amount of OPG/OCIF contained in the serum. In contrast, the presence of 50% human serum did not alter the standard curve for the homodimer-specific ELISA (Fig. 2, lower panel), suggesting that human serum does not contain substantial amounts of the homodimer. These data indicate that OPG/OCIF exists in human serum mainly as the monomeric form and that the ELISA for total OPG/OCIF is suitable for measurement of OPG/OCIF in human serum. Similar results were obtained with serum samples from 18 additional normal individuals and from 6 patients with osteoporosis (data not shown). The ELISA for total OPG/OCIF detected both forms of rOPG/rOCIF at concentrations between 32.5 and 500 pg/ml with a recovery of 93.4–102.6% (monomeric form of OPG/OCIF, Fig. 3A; homodimeric form of OPG/OCIF, data not shown). When serum samples from normal controls and patients with osteoporosis were diluted 2- to 16-fold with the assay buffer, the ELISA for total OPG/OCIF showed a linear response over the given range of dilution (Fig. 3B). To validate the basal levels of OPG/OCIF in human serum, we have incubated normal human serum with an agarose gel alone or the agarose gel coupled with an anti-OCIF antibody, OI-1 or OI-20, a MAb other than those used for the ELISA for total OPG/OCIF. After removing the gel, the concentration of OPG/OCIF in the serum was determined by the ELISA for total OPG/OCIF. The concentration of OPG/OCIF was under the minimum detection limit in the antibody-treated serum, but was almost the same level in both the untreated serum and the serum treated with an agarose gel alone. The results indicated that the ELISA for total OPG/OCIF specifically detects the basal levels of OPG/OCIF in human serum.
Serum concentrations of OPG/OCIF in healthy individuals and patients with osteoporosis
With the ELISA for total OPG/OCIF, we determined the serum concentration of OPG/OCIF in healthy individuals. Serum OPG/OCIF concentrations increased with age in both healthy men and women (Figs. 4A and 4B). There was a significant positive correlation (r = 0.829, p < 0.0001, n = 56) between the serum concentration of OPG/OCIF and age in men (Fig. 4A). Similarly, for women, a significant correlation (r = 0.769, p < 0.0001, n = 61) was detected between the serum concentration of OPG/OCIF and age (Fig. 4B). There was no statistically significant difference in the concentration of OPG/OCIF between men and women of the same age. Renal function of these subjects was normal, and there was no relation between the concentration of OPG/OCIF and serum Cr (data not shown).
To examine the possible relation between circulating OPG/OCIF concentration and bone mass, we determined the serum concentration of OPG/OCIF in postmenopausal Japanese women diagnosed as normal or with osteopenia or osteoporosis on the basis of L2–L4 BMD. There was no significant difference in age, years after menopause, or smoking status among the three groups (Table 2). Body weight in normal controls was significantly greater than that in individuals with osteopenia or those with osteoporosis, and the frequency of nontraumatic vertebral fracture was significantly higher in patients with osteoporosis than in the other two groups. Serum concentrations of calcium or osteocalcin, serum activity of BAP, and urinary Pyr excretion did not differ significantly among the three groups, but urinary Dpyr excretion was significantly higher in patients with osteoporosis than in normal controls (Table 2). The L2–L4 BMD was significantly lower in patients with osteoporosis than in subjects with osteopenia or normal controls, and L2–L4 BMD in individuals with osteopenia was also significantly lower than that in normal controls (Table 2).
Table Table 2.. BMD and Other Characteristics of Postmenopausal Women Diagnosed as Normal or with Osteopenia or Osteoporosis
The serum concentration of OPG/OCIF was significantly higher in patients with osteoporosis than in age-matched normal controls (Fig. 5). When the osteoporotic group was further divided into age-matched subgroups of lower (<0.7 g/cm2 for L2–L4 BMD or <0.85 g/cm2 for total body BMD) and higher bone mass, the serum concentration of OPG/OCIF was significantly higher in the former subgroups (2.1 ± 0.5 ng/ml, n = 36, p = 0.002 for L2–L4 BMD and 2.2 ± 0.6 ng/ml, n = 34, p = 0.03 for total body BMD) than in the latter subgroups (1.8 ± 0.3 ng/ml, n = 37, for L2–L4 BMD; and 1.9 ± 0.4 ng/ml, n = 33 for total body BMD). Serum OPG/OCIF concentrations were also significantly higher (p = 0.03) in osteoporotic women with a history of nontraumatic vertebral fracture than in those without such a history (data not shown). Simple regression analysis of combined data for BMD from normal controls and patients with osteopenia or osteoporosis revealed an inverse correlation between the serum concentration of OPG/OCIF and L2–L4 BMD (r = 0.248, p = 0.001, n = 167) or total body BMD (r = 0.355, p < 0.0001, n = 162).
When the relation between the serum concentration of OPG/OCIF and biochemical markers of bone metabolism was analyzed in the osteoporotic women, a positive correlation was detected between OPG/OCIF concentration and serum activity of BAP (Fig. 6A), or urinary excretion of Pyr (Fig. 6B) or Dpyr (Fig. 6C). Similar results were obtained in normal controls: the concentration of OPG/OCIF was positively correlated with serum BAP activity (r = 0.318, p = 0.026, n = 49), serum osteocalcin (r = 0.301, p = 0.033, n = 50), urinary Pyr (r = 0.464, p = 0.0008, n = 49), or Dpyr (r = 0.404, p = 0.004, n = 49).
OPG/OCIF is present as both a monomer with an apparent molecular mass of 60 kDa and a homodimer of 120 kDa in conditioned media of human fibroblasts(15) and Chinese hamster ovary cells transfected with an expression vector comprising OPG/OCIF cDNA.(17,18) A naturally occurring soluble form of Fas, another member of the TNFR family, has been detected in serum of patients with systemic lupus erythematosus, suggesting that the soluble Fas protein functions as an antagonist for the membrane-bound Fas ligand.(25) In addition, a soluble form of TNFR generated by limited proteolysis inhibits the biological activity of both TNF-α and TNF-β.(26-28) It is therefore possible that OPG/OCIF also functions as an antagonist for a ligand that triggers osteoclastogenesis and plays an important role in bone remodeling through the regulation of osteoclastic bone resorption. Recently, we and another group have molecularly cloned a ligand for OPG/OCIF.(29,30) The ligand termed osteoclast differentiation factor is a member of the membrane-associated TNF ligand family that stimulates both development and activation of osteoclasts.(29,30)
Although OPG/OCIF had been shown to inhibit bone resorption in vivo,(16,17) it remained unknown whether OPG/OCIF circulates in human blood or whether it plays any role in bone physiology or pathology, especially in diseases characterized by excessive bone resorption. In an attempt to develop a sensitive ELISA capable of measuring the concentration of OPG/OCIF in human blood, we prepared several MAbs to human rOPG/rOCIF. In screening these MAbs, we selected those that exhibited substantial reactivity with solid-phase antigen even in the presence of 80% calf serum, indicative of a high-affinity interaction with antigen. Further characterization of the MAbs with both monomeric and homodimeric forms of rOPG/rOCIF revealed that some showed a high affinity for both forms, whereas others preferentially or specifically recognized the homodimer. In addition, several of the antibodies were capable of neutralizing the ability of OPG/OCIF to inhibit osteoclastogenesis.
The OPG/OCIF molecule contains seven domains, and the NH2-terminal portion (domains 1–4) shows substantial homology to the extracellular domain of members of the TNFR family of proteins.(16,17) Most MAbs that bound to ΔD567, an OPG/OCIF mutant consisting of only domains 1–4,(24) had the ability to neutralize OPG/OCIF activity (Table 1), which is consistent with our previous observation that the NH2-terminal portion of OPG/OCIF (domains 1–4) is sufficient to inhibit osteoclastogenesis in vitro.(24) Some MAbs neutralized OPG/OCIF activity despite the fact that they showed no binding to the mutant (Table 1), raising the possibility that they might induce a conformational change in the molecule by binding to the COOH-terminal portion. The MAb designated OI-1 neutralized the biological activity of OPG/OCIF by blocking its binding to sites on ST2 cells treated with 1α,25-dihydroxyvitamin D3.(16) In contrast, MAb OI-5 and OI-7 neither neutralized OPG/OCIF activity nor blocked its binding. The neutralizing MAbs may provide useful tools for studying the molecular mechanism by which OPG/OCIF inhibits osteoclastogenesis.
We developed two sensitive ELISAs, one that detects both the homodimeric and monomeric forms of OPG/OCIF to equal extent and one that specifically detects the homodimer. Using these assays and homodimeric rOPG/rOCIF diluted with human serum, we demonstrated that OPG/OCIF circulates in human blood mainly as a monomer, suggesting that the ELISA for total OPG/OCIF is suitable for determining the serum concentration of OPG/OCIF. The half-life of the OPG/OCIF monomer in α-phase is significantly longer than that of the homodimer,(18) suggesting that the homodimer may tend to redistribute from the circulation to tissues more rapidly than the monomer. Alternatively, the monomer may represent a degradation product of the homodimer, given that it lacks several amino acids in the COOH-terminal region, including Cys379 which participates in an intermolecular disulfide bond.(18) Our sensitive and differential ELISA systems may provide useful tools for gaining further insight into the molecular forms of OPG/OCIF produced by various tissues(16,17) as well as their relative amounts and physiological roles.
We have now demonstrated that the serum concentration of OPG/OCIF increases with age in both healthy men and women. Interestingly, the increase in serum OPG/OCIF concentrations appears to be accelerated after 50–60 years of age in men as well as in women. An important question is what age-dependent factor(s) regulates serum OPG/OCIF levels. At present, however, there is no further information about the mechanism of regulation as well as the major source of circulating OPG/OCIF. Since OPG/OCIF mRNA is expressed in a variety of tissues, including lung, kidney, and heart,(17) it may well be that multiple tissues contribute together to circulating OPG/OCIF. In view of our recent observations that OPG/OCIF is produced and secreted by various osteoblastic and bone marrow stromal cell lines,(31) OPG/OCIF produced in the skeletal tissue may also be released into the circulation.
We further showed in the current study that serum OPG/OCIF concentrations are significantly higher in postmenopausal women with osteoporosis than in age-matched normal controls and that within the osteoporotic group, OPG/OCIF concentrations are higher in women with lower BMD, suggesting that the concentration of OPG/OCIF changes as a function of bone mass. Moreover, high circulating concentrations of OPG/OCIF were associated with higher serum BAP activity and urinary excretion of Pyr and Dpyr in osteoporotic as well as normal women, which is consistent with the concept that serum OPG/OCIF reflects a high bone-turnover state. Taken together with the findings that OPG/OCIF-deficient mice develop osteoporosis(32,33) and that exogenously administered OPG/OCIF protects against osteoporosis in estrogen-deficient animals,(17) the increase in serum OPG/OCIF concentrations in osteoporotic women may reflect a compensative response to the increased osteoclastic bone resorption and the resultant bone loss caused by estrogen deficiency, rather than a cause of osteoporosis.
In conclusion, measurement of OPG/OCIF by the sensitive ELISA developed in the present study should provide a useful tool for studying pathophysiological roles of OPG/OCIF, the mechanism by which OPG/OCIF expression is regulated, and how serum OPG/OCIF concentrations change with the progression of osteoporotic disease and in response to therapeutic intervention. The utility of our OPG/OCIF ELISA may warrant further studies of larger populations with diverse ethnic backgrounds and of other bone disorders characterized by excessive bone resorption, including hypercalcemia in malignancy, hyperparathyroidism, Paget's disease, and rheumatoid arthritis.
This work was supported in part by a Research Grant for Longevity Sciences from the Ministry of Health and Welfare of Japan (to K.I. and Y.Y.).