INTRODUCTION

Interleukin-6 (IL-6),¹ a cytokine produced by cells of the osteoblast and osteoclast lineages, increases osteoclast recruitment and consequently bone resorption (1-5). IL-6 is believed to mediate the effects of selected hormones on bone resorption and is considered in part responsible for the bone loss observed in conditions of estrogen or androgen deficiency and to play a role in the hypercalcemia of malignancy (2-4, 6-8). Although IL-6 appears to play a central role in bone resorption, its mechanism of action is poorly understood, and its effects on bone matrix degradation are not known. In co-culture systems of mouse osteoblasts and bone marrow cells, IL-6 stimulates the formation of multinucleated osteoclast-like cells in the presence of the IL-6-soluble receptor (IL-6sR) but not in its absence (9). The IL-6sR is a 55-kDa soluble protein generated by proteolytic cleavage of the IL-6 membrane-bound receptor or by translation of an alternatively spliced RNA (10-13). The proteolytic cleavage site of the IL-6 receptor is immediately adjacent to the transmembrane domain between residues Gln-357 and Asp-358 (13). Since the IL-6sR is present in the systemic circulation and allows for the effects of IL-6 on bone resorption, it may be relevant to the actions of IL-6 in normal and abnormal bone remodeling. Agents that modify bone resorption frequently alter bone collagen degradation by changing the expression of matrix metalloproteinases and their inhibitors (14-17). Consequently, in addition to its stimulatory effects on bone resorption, IL-6 and its soluble receptor might alter bone collagen degradation and the expression of collagenase and its inhibitors by skeletal cells.

Matrix metalloproteinases are a family of related proteolytic enzymes including collagenases, gelatinases, and stromelysins (18-20). Collagenases cleave fibrillar collagen at neutral pH and are considered important in matrix remodeling. Three collagenases have been described as follows: collagenase 1, secreted by stimulated human fibroblasts and osteoblasts and by human chondrocytes from osteoarthritic cartilage; collagenase 2, secreted by neutrophils; and collagenase 3, secreted by human breast carcinoma cells, human chondrocytes, and rat osteoblasts (14, 16, 19, 21-23). Although human osteoarthritic chondrocytes express collagenase 1 and 3, unstimulated normal human osteoblasts do not secrete detectable levels of either collagenase² (21, 23). However, certain human osteosarcoma cell lines synthesize collagenase and normal human osteoblasts exposed to parathyroid hormone, and selected cytokines synthesize collagenase 1 (21). Normal rat osteoblasts and rat osteosarcoma cells express collagenase 3 but are not known to secrete detectable amounts of collagenase 1 (14, 16, 22). Type II collagen is preferentially hydrolyzed by collagenase 3, and collagenases 1, 2, and 3 degrade fibrillar type I collagen with similar efficiency (24). These observations suggest a role for collagenase 3 in connective tissue remodeling in the skeleton.

IL-6 was reported not to induce collagenase 1 or stromelysin but to increase the expression of the tissue inhibitor of metalloproteinases (TIMP) 1 in human skin fibroblasts, synoviocytes, and chondrocytes (25-27). However, its effects on the expression of collagenase and TIMP 1 in skeletal cells or on the expression of TIMP 2 and 3 in skeletal or nonskeletal cells have not been reported. Despite a lack of an effect on matrix metalloproteinase expression in nonskeletal cells, we postulated that since IL-6 enhances bone resorption, it might regulate the expression of collagenase and of TIMPs in rat osteoblasts.

In the present study, we examined the actions of IL-6 in the presence and absence of the IL-6sR, on collagenase 3 transcripts, and immunoreactive collagenase levels and on the expression of TIMPs 1, 2, and 3 in cultures of osteoblast-enriched cells from 22-day fetal rat calvariae (Ob cells) and determined possible mechanisms involved. Effects of IL-6 on collagenase mRNA levels also were examined in the murine osteoblastic cell line MC3T3 and in skin fibroblasts.

EXPERIMENTAL PROCEDURES

The culture method used to obtain Ob cells was described in detail previously (28). Parietal bones were obtained from 22-day-old fetal rats immediately after the mothers were sacrificed by blunt trauma to the nuchal area. (This project was approved by the Institutional Animal Care and Use Committee of Saint Francis Hospital and Medical Center.) Cells were obtained by five sequential digestions of the parietal bone using bacterial collagenase (CLS II, Worthington). Cell populations harvested from the third to the fifth digestions were cultured as a pool. About 80% or more of these cells were previously shown to display osteoblastic characteristics (28, 29). Ob cells were plated at a density of 8,000-12,000 cells/cm² and cultured in a humidified 5% CO₂ incubator at 37 °C until reaching confluence (about 50,000 cells/cm²). Cells were cultured in Dulbecco's modified Eagle's medium supplemented with nonessential amino acids and 10% fetal bovine serum (both from Summit Biotechnology, Fort Collins, CO). MC3T3 cells are a mouse osteoblastic cell line derived from fetal mouse calvaria and were grown in -minimal essential medium supplemented with 10% fetal bovine serum under similar conditions (30). Ob or MC3T3 cells were grown to confluence, transferred to serum-free medium for 20-24 h, and exposed to test or control medium in the absence of serum for 2-24 h as indicated in the text and legends. For the nuclear run-on experiment, Ob cells were grown to subconfluence, trypsinized, replated, and grown to confluence when they were serum-deprived and exposed to test or control solutions for 2-6 h. Skin fibroblasts, obtained from 22-day-old fetal rats by collagenase digestion, were grown in Dulbecco's modified Eagle's medium in the presence of 10% fetal bovine serum, trypsinized, and passaged three to four times before testing (31, 32). Fibroblasts were serum-deprived for 20-24 h prior to exposure to test or control medium in the absence of serum. Recombinant human IL-6 and IL-6sR (R&D Systems Inc., Minneapolis, MN) were dissolved in phosphate-buffered saline containing 0.1% bovine serum albumin and tested independently or added simultaneously to the culture. Cycloheximide (Sigma) was added directly to the culture medium. 5,6-Dichlorobenzimidazole riboside (DRB) and indomethacin (both from Sigma) were dissolved in ethanol and diluted 1:200 and 1:1000, respectively, in Dulbecco's modified Eagle's medium. An equal amount of solvent was added to control cultures. At the end of the incubation, the medium was harvested in the presence of 0.1% polyoxyethylene sorbitan monolaurate (Pierce) and stored at 80 °C prior to immunoblot analysis. The cell layer was extracted for RNA analysis and stored at 80 °C or nuclei were obtained by Dounce homogenization for the nuclear run-on assay.

Total cellular RNA was isolated with guanidine isothiocyanate, at acid pH, followed by a phenol-chloroform extraction and isopropyl alcohol and ethanol precipitations (33) or by RNeasy kit per manufacturer's instructions (Qiagen, Chatsworth, CA). The RNA recovered was quantitated by spectrophotometry, and equal amounts of RNA from control or test samples were loaded on a formaldehyde-agarose gel following denaturation. The gel was stained with ethidium bromide to visualize RNA standards and ribosomal RNA, before and after transfer, documenting equal RNA loading of the samples. The RNA was blotted onto GeneScreen Plus charged nylon (DuPont NEN). Restriction fragments containing a 2.6-kilobase (kb) interstitial collagenase 3 cDNA (kindly provided by Cheryl Quinn, St. Louis, MO), a 1.6-kb rat stromelysin 1 cDNA (kindly provided by Lynn Matrisian, Nashville, TN), a murine 1.6-kb stromelysin 3 cDNA (kindly provided by Paul Basset, Strasbourg, France), an 825-base pair (bp) murine TIMP 1 cDNA, a 700-bp murine TIMP 2 cDNA, a 750-bp murine TIMP 3 cDNA (all TIMP cDNAs kindly provided by Dylan Edwards, Calgary, Alberta, Canada), and a 750-bp murine 18 S ribosomal RNA cDNA (American Type Culture Collection, Rockville, MD) were labeled with [-³²P]deoxycytidine triphosphate (dCTP) and [-³²P]deoxyadenosine triphosphate (dATP) (specific activity of 3,000 Ci/mmol; DuPont NEN), using the random hexanucleotide primed second strand synthesis method (22, 34-39). Hybridizations were carried out at 42 °C for 16-72 h. Post-hybridization washes were performed in 1 × saline sodium citrate at 65 °C for collagenase 3, stromelysin 1 and 3, TIMP 1, 2, and 3 cDNAs, and in 0.1 × saline sodium citrate at 65 °C for 18 S ribosomal RNA. The bound radioactive materials were visualized by autoradiography on Kodak X-AR5 or Biomax film (Eastman Kodak) or DuPont reflection film employing intensifying screens. Relative hybridization levels were determined by densitometry. Northern analyses shown are representative of three or more cultures.

Collagenase heterogeneous nuclear RNA (hnRNA) was analyzed by RT-PCR using a sense primer 5-CATTCAGCTATTCTGGCCAC-3, spanning nucleotides 27-46 of exon 1 of the rat collagenase 3 gene and an antisense primer, 5-AAAAGACCAGAACAACCAGC-3, spanning nucleotides 61-80 of intron 1 to yield a 186-bp product (22, 40). RNA was extracted as described for Northern analysis, and samples were treated with amplification grade DNase I according to manufacturer's (Life Technologies, Inc.) instructions to remove potentially contaminating DNA. RNA (1 µg) was copied into cDNA using Moloney murine leukemia virus reverse transcriptase (Life Technologies, Inc.) and the antisense primer according to manufacturer's instructions except that Taq polymerase buffer was used instead of reverse transcriptase buffer (41). A DNA standard was synthesized by PCR amplification of plasmid DNA pGL2-Basic (Promega Corp., Madison, WI) using the rat collagenase hnRNA primer set and low stringency annealing conditions, as described by Forstr (42). The cDNA and 0.05 attomol DNA standard were amplified by PCR using 24 cycles of 94 °C for 1 min, 59 °C for 1 min, and 72 °C for 1 min in the presence of Taq polymerase (Life Technologies, Inc.), 0.15 µmol of sense and antisense primers, and 5 µCi [-³²P]dCTP. PCR products were resolved on an 8% polyacrylamide gel (Gel-Mix 8, Life Technologies, Inc.), containing 100 mM Tris borate, 1 mM EDTA, and visualized by autoradiography. The amplification protocol yielded products that were within the linear range for both the collagenase hnRNA and the standard. Data on hnRNA are representative of three cultures.

To examine changes in the rate of transcription, nuclei were isolated by Dounce homogenization in a Tris buffer containing 0.5% Nonidet P-40. Nascent transcripts were labeled by incubation of nuclei in a reaction buffer containing 500 µM each of ATP, CTP, and guanidine triphosphate, 150 units of RNasin (Promega), and 250 µCi of [-³²P]uridine triphosphate (3,000 Ci/mmol; DuPont NEN) (43). RNA was isolated by treatment with DNase I and proteinase K, followed by phenol-chloroform extraction and ethanol precipitation. Linearized plasmid DNA containing about 1 µg of cDNA was immobilized onto GeneScreen Plus by slot blotting according to manufacturer's directions (DuPont NEN). The plasmid vector pGL2-Basic (Promega) was used as a control for nonspecific hybridization, and a 750-bp murine 18 S ribosomal RNA cDNA was used to estimate uniformity of radioactive counts applied to the membrane. Equal counts/min of [³²P]RNA from each sample were hybridized to cDNAs using the same conditions as for Northern blot analysis and were visualized by autoradiography.

Medium samples were fractionated by polyacrylamide gel electrophoresis using denaturing and nonreducing conditions and transferred onto Immobilon P membranes (Millipore, Bedford, MA). After blocking with 2% bovine serum albumin, the membranes were exposed to a 1:1000 dilution of rabbit antiserum raised against rat collagenase 3 (kindly provided by J. Jeffrey, Albany, NY), previously characterized for specificity and immunoreactivity, followed by the addition of goat anti-rabbit IgG conjugated to horseradish peroxidase (44). The blots were washed and developed with a horseradish peroxidase chemiluminescence detection reagent (DuPont NEN), visualized by autoradiography on DuPont Reflection film employing Reflection intensifying screens, and analyzed by densitometry. Data shown are representative of three cultures.

To assess gelatinase activity, aliquots of conditioned medium were extracted with methanol and chloroform, dried, and resuspended in sample buffer containing 2% sodium dodecyl sulfate. Samples were loaded on a 7.5% polyacrylamide gel containing 1 mg/ml gelatin and fractionated by electrophoresis as described (21). The gels were washed twice with a 2.5% Triton X-100 (Sigma) solution, rinsed with water, and incubated overnight in 50 mM Tris, 10 mM CaCl₂ buffer at 37 °C. Proteolytic activity was visualized by staining the gels with 1% Coomassie Blue.

Data on collagenase mRNA decay were analyzed by linear regression, and the slopes of the regression lines obtained for control and treated cells were compared for significant differences using the method of Sokal and Rohlf (45).

RESULTS

Northern blot analysis of total RNA from Ob cells revealed a rat collagenase 3 transcript of 2.9 kb (Fig. 1). Continuous treatment of Ob cells with IL-6 at 100 ng/ml caused a small increase in collagenase steady state transcripts. However, in the presence of IL-6sR a marked increase was observed. This increase initially occurred after 2 h of exposure to IL-6 and IL-6sR, and the absolute effect was virtually maximal after 6 h (Fig. 1). Confirming prior observations, a decrease in collagenase 3 mRNA was noted in control cultures after 24 h, but this did not preclude the stimulatory effect of IL-6 and its soluble receptor (46). Since the effect of IL-6 and IL-6sR was sustained despite the decrease in control collagenase expression, the relative stimulation by IL-6 and IL-6sR after 24 h was greater than after 6 h. The effect of IL-6 alone was dose-dependent and of limited magnitude, and continued exposure of Ob cells to IL-6 at 100-300 ng/ml for 6 or 24 h increased collagenase transcripts by 1.5- or 3.0-fold, respectively (Figs. 2 and 3). In the presence of IL-6sR, IL-6 increased collagenase mRNA levels at doses as low as 1 ng/ml, and the effect was maximal at 100 ng/ml (Fig. 3). IL-6sR alone increased collagenase 3 transcripts at concentrations of 125 and 250 ng/ml by 7- to 8-fold (Fig. 4). This effect was magnified in the presence of IL-6, and IL-6sR at 31-250 ng/ml in the presence of IL-6 at 100 ng/ml increased collagenase 3 mRNA levels in Ob cells by 50-100-fold. IL-6 at 100 ng/ml and IL-6sR at 50 ng/ml for 24 h increased the levels of immunoreactive interstitial collagenase 3 in the culture medium of Ob cells, and when tested together they caused a greater than 300-fold increase as determined by Western blot analysis (Fig. 5). Collagenase was identified by co-migration with a purified rat procollagenase 3 standard. Colloidal gold staining of the blot revealed equal amounts of protein in control and treated cultures (data not shown). The precise extent of the IL-6 and IL-6sR effect on collagenase mRNA and protein levels, was difficult to determine by densitometry due to the magnitude of the effect as compared with control cultures.

To determine whether the effect of IL-6 and its soluble receptor on rat interstitial collagenase mRNA levels was dependent on protein synthesis, confluent cultures of Ob cells were treated with IL-6 and IL-6sR in the presence or absence of cycloheximide at doses known to inhibit protein synthesis (47). Cycloheximide at 3.6 µM did not prevent the stimulatory effect of IL-6 and IL-6sR on collagenase 3 mRNA alone or in combination (Fig. 6). To determine whether the effects of IL-6 and IL-6sR were due to changes in prostaglandin synthesis, IL-6 and its soluble receptor were tested in the presence and absence of indomethacin at 10 µM. Indomethacin did not modify the expression of collagenase 3 in control Ob cell cultures and did not prevent the effect of IL-6 in the presence of IL-6sR on collagenase transcripts (Fig. 7).

To determine whether the effects of IL-6 on collagenase mRNA levels were due to changes in transcript stability, Ob cells were exposed to control or IL-6 and IL-6sR-containing medium for 1 h and then treated with the RNA polymerase II inhibitor DRB at 75 µM, in the presence or absence of IL-6 and IL-6sR for 2, 4, and 8 h (48). The half-life of collagenase mRNA in transcriptionally arrested Ob cells was 4 h in control and test cultures (Fig. 8). Slope analysis indicated that IL-6 and IL-6sR did not change the stability of collagenase mRNA. IL-6 caused a small increase in interstitial collagenase hnRNA in Ob cells as determined by RT-PCR, but in the presence of its soluble receptor, collagenase 3 hnRNA was increased by 5- and 8-fold at 2 and 6 h, respectively (Fig. 9). The effect was sustained for 24 h, when a decrease in collagenase hnRNA levels was observed in control cultures. This resulted in a relative greater stimulation of approximately 30-fold in collagenase hnRNA by IL-6 and IL-6sR. Co-amplification of an exogenous DNA standard, designed to use the same set of primers, revealed uniform PCR efficiency, and omission of the reverse transcription step resulted in no signal proving lack of DNA contamination. To confirm that IL-6 modified collagenase mRNA by transcriptional mechanisms, the rate of gene transcription was measured by nuclear run-on assays. IL-6 and IL-6sR increased the transcriptional rate of the collagenase gene after 2 h by 5-fold and after 6 h by 33-fold (Fig. 10).

To determine whether the effect of IL-6 and its soluble receptor was specific for collagenase 3, we examined their actions on other matrix metalloproteinases in Ob cell cultures. Transcripts of stromelysin 1 were not detectable by Northern analysis in control or IL-6 and IL-6sR-treated Ob cells. IL-6 at 100 ng/ml in the presence of IL-6sR at 50 ng/ml for 2, 6, or 24 h did not cause a detectable increase in stromelysin 3 mRNA levels in Ob cells (not shown). Confirming observations in normal human osteoblasts, conditioned medium from control Ob cell cultures contained a gelatin-degrading enzyme migrating with an approximate molecular mass of 72 kDa but did not contain a 92-kDa gelatinase (21). IL-6 at 100 ng/ml in the presence of its soluble receptor at 50 ng/ml for 24 h increased the levels of the 72-kDa gelatinase by about 3- to 4-fold (Fig. 11).

The effect of IL-6 and its soluble receptor on collagenase 3 expression was also observed in the cell line MC3T3, a homogeneous source of non-transformed mouse osteoblastic cells. IL-6 at 100 ng/ml and IL-6sR at 50 ng/ml enhanced collagenase 3 mRNA levels after 2 h, and the effect continued to increase after 6 and 24 h, when IL-6 and IL-6sR caused a marked increase in collagenase 3 mRNA levels (Fig. 12). IL-6 at 100 ng/ml by itself stimulated collagenase 3 mRNA levels in MC3T3 cells, but IL-6sR at 50 ng/ml alone did not increase collagenase mRNA, and this is in agreement with the results obtained in Ob cells at the same concentrations of IL-6 and soluble receptor (Figs. 3, 4, and 12). Confirming our previous observations, rat skin fibroblasts expressed limited levels of collagenase 3 mRNA (32) (Fig. 13). Neither IL-6 at 100 ng/ml nor IL-6sR at 50 ng/ml for periods of 2, 6, or 24 h modified collagenase 3 transcripts in skin fibroblasts. However, when tested in combination they increased collagenase 3 mRNA by 2-, 5-, and 10-fold after 2, 6, and 24, h respectively.

Confirming prior observations, Ob cells expressed a TIMP 1 transcript of 0.9 kb, two TIMP 2 transcripts of 1.0 and 3.5 kb, respectively, and two TIMP 3 transcripts of 2.5 and 4.5 kb, respectively (46). IL-6 at 100 ng/ml in the presence of its soluble receptor at 50 ng/ml increased the expression of TIMP 1 after 6 h, an effect that was sustained for 24 h (Fig. 14). IL-6 and IL-6sR had no effect on TIMP 2 expression, and a minimal increase in TIMP 3 mRNA levels after 6 h of exposure was observed (Figs. 15 and 16).

DISCUSSION

The present investigation was undertaken to determine whether IL-6 regulates collagenase 3 and TIMP 1, 2, or 3 expression in cultures of rat Ob cells. IL-6 by itself caused a small increase in collagenase 3 transcripts, but in the presence of its soluble receptor, IL-6 caused a significant time- and dose-dependent stimulation of collagenase 3 mRNA and protease levels. IL-6 also increased the expression of TIMP 1 transcripts and had a minimal stimulatory effect on TIMP-3 mRNA levels. The effect of IL-6 and IL-6sR on collagenase mRNA levels was not dependent on de novo protein synthesis and was observed in the presence and absence of cycloheximide. Although IL-6 has been shown to induce prostaglandin E₂ and prostaglandins enhance collagenase synthesis in rat UMR-106 osteoblastic cells, the stimulation of collagenase 3 by IL-6 and IL6sR in Ob cells was not dependent on prostaglandin synthesis (49, 50). IL-6 and IL-6sR did not alter collagenase mRNA stability in transcriptionally arrested Ob cells and increased collagenase hnRNA levels and the rate of transcription of the collagenase 3 gene. These results indicate that IL-6, in the presence of its soluble receptor, stimulates rat collagenase 3 expression by transcriptional mechanisms.

The effect of IL-6 and its soluble receptor was relatively specific for collagenase 3 expression, since they had only a modest effect on gelatinase activity. Our studies confirm previous observations by other investigators demonstrating that the predominant gelatinase secreted by osteoblastic cells in culture is the 72-kDa gelatinase (21, 51). The stimulatory effect of IL-6 on 72-kDa gelatinase levels is not surprising since other agents known to induce collagenase expression, such as IL-1, also increase gelatinase production by skeletal cells (51). It is possible that IL-6 induces collagenase 1 expression in human osteoblasts. This effect was not tested in the present studies because rat osteoblasts do not express collagenase 1 (14, 16, 22). IL-6 in the presence of its soluble receptor increased collagenase 3 mRNA levels in the osteoblastic MC3T3 cell line to a similar extent as in Ob cells, confirming that IL-6 stimulates collagenase synthesis in osteoblasts. Although IL-6 was reported not to regulate collagenase expression in other cell systems, its effect is not unique to osteoblasts and, when tested with its soluble receptor, an effect was noted in skin fibroblasts (25-27). However, it is important to indicate that the magnitude of the effect was about 10 times greater in osteoblasts than in fibroblasts. Our studies demonstrate that while IL-6 alone does not stimulate collagenase transcripts in skin fibroblasts, in the presence of its soluble receptor it was stimulatory. This suggests that failure of other investigators to detect an effect was related to the fact that IL-6 was tested in the absence of IL-6sR.

IL-6 induces the transcription of the collagenase 3 gene in osteoblasts, but the gene sequences responsible for the effect have not been determined. In other cells, IL-6 activates the synthesis of the AP-1 family of transcription factors and induces NF-IL-6, a member of the CCAAT enhancer binding protein family, acute phase response factor, and octamer binding proteins (52-55). Examination of the collagenase 3 promoter region reveals potential binding sites for NF-IL-6, AP-1, and acute phase response factor in the base pair 818 to +1 region (40). It is possible that IL-6 acts by inducing or activating transcription factors that bind to one or more of these sequences. In fact, the induction of collagenase 3 by parathyroid hormone in rat osteosarcoma cells is mediated through an AP-1 binding site (40).

IL-6 stimulates bone resorption by increasing osteoclast recruitment and differentiation, and in selected culture systems this effect is observed only in the presence of IL-6sR (9). The effect of IL-6 on osteoclast recruitment requires the presence of osteoblasts and depends on IL-6 receptors expressed by osteoblastic cells and not by osteoclast progenitors (9, 56). Consequently, it is not surprising that IL-6 regulates collagenase expression in osteoblasts and that its effect depends on the presence of IL-6sR. The IL-6 receptor consists of two transmembrane proteins, an 80-kDa ligand binding protein and 130-kDa glycoprotein (gp130) involved in signal transduction. IL-6 binds to the 80-kDa surface receptor, and the complex associates with two gp130 molecules (13). The IL-6sR binds IL-6 with similar affinity as the membrane-bound receptor and the complex can elicit IL-6 signaling in cells expressing gp130 and not cell surface-associated receptor (13, 55, 57). This indicates that the IL-6sR is sufficient to allow for IL-6 actions. It is important to note that shedding of gp130 also occurs, and IL-6, IL-6sR, and soluble gp130 can associate and form a ternary complex (13). The stimulatory activity of a soluble receptor is unusual since soluble receptors for other cytokines, such as leukemia inhibitory factor, compete with the membrane-bound receptor and prevent their biological effects (58).

The modest effect of IL-6 by itself on collagenase synthesis may be because of limited IL-6 receptor expression by the osteoblast or because of receptor down-regulation and internalization immediately following IL-6 binding to the cell. This may be a mechanism to prevent overstimulation by IL-6 (55). The addition of the IL-6sR may allow for continuous stimulation by IL-6 and a supramaximal effect, since IL-6sR binds IL-6 with similar affinity as the membrane-bound receptor and mediates IL-6 signaling (13, 55, 57). The soluble receptor by itself had an effect on collagenase expression in Ob cells confirming that these cells secrete IL-6 in quantities sufficient to have biological activity (59). An alternative explanation may be that the soluble receptor itself initiates signaling in the absence of IL-6 and that ligand binding greatly increases the efficiency of activation. The effects of IL-6sR on collagenase 3 expression were observed at doses similar to those detected in human serum, suggesting its relevance to physiological or pathological conditions (13). Serum concentrations of IL-6sR are elevated in conditions of increased bone remodeling, such as multiple myeloma and estrogen deficiency (60, 61). Reports on the effects of IL-6 on bone resorption have been contradictory, and some investigators have failed to demonstrate a stimulatory effect (62). A stimulation of collagenase 3 by IL-6 and IL-6sR would support a role for this cytokine in bone resorption.

There is a similarity between our results and those observed in other skeletal cells since IL-6 alone does not stimulate the formation of osteoclast-like cells, but in the presence of its soluble receptor it has a marked stimulatory effect (9). These observations, as well as those from the present studies, suggest that skeletal cells express limited levels of IL-6 receptors under basal or unstimulated conditions. Glucocorticoids have been shown to induce IL-6 receptors in epithelial and osteoblastic cells, and it is possible that the actions of IL-6 on skeletal remodeling occur under specific conditions of IL-6 receptor up-regulation (56, 63). In fact, the induction of osteoclast differentiation by IL-6 depends on the presence of IL-6 receptors on osteoblastic cells, and glucocorticoids allow for IL-6 to induce osteoclast differentiation (56). This also suggests that IL-6 may be relevant to the actions of glucocorticoids on bone resorption. However, IL-6 probably does not mediate the actions of these steroids on collagenase expression in osteoblasts, because in osteoblasts IL-6 and IL-6sR induce collagenase at the transcriptional level, whereas the stimulatory effect of glucocorticoids on this gene occurs by post-transcriptional mechanisms (32).

The synthesis of collagenase 1 and 3 by human and rat osteoblasts, respectively, is regulated by systemic hormones and by cytokines present in the bone microenvironment. Consequently, the apparent constitutive level of collagenase expression by the osteoblast is in fine balance and depends on the exposure of the osteoblast to factors that stimulate and factors that inhibit collagenase synthesis (16, 46, 50, 64). Recently, we demonstrated that insulin-like growth factors I and II are autocrine down-regulators of collagenase 3 in Ob cells (65). Therefore, accumulation of insulin-like growth factors I or II in the bone microenvironment results in decreased collagenase synthesis by the osteoblast. In addition, the state of osteoblastic maturation may result in changes in collagenase expression (21, 66). Furthermore, agents that stimulate the replication of cells of the osteoblastic lineage increase collagenase expression, whereas agents that induce osteoblastic cell differentiation, such as bone morphogenetic proteins, decrease collagenase synthesis (16, 64, 66, 67).

In conclusion, the present studies demonstrate that IL-6 and its soluble receptor increase rat collagenase 3 expression in osteoblasts by transcriptional mechanisms and increase the expression of TIMP 1 transcripts. It is probable that these effects play a role in the degradation of the collagen matrix and in bone remodeling.