Cytokine-Stimulated Expression of Inducible Nitric Oxide Synthase by ...

0013.7227/95/$03.00/O Endocrinology Copyright 0 1995 by The Endocrine

Vol. 136, No. 12 Printed m U.S.A. Society

Cytokine-Stimulated Expression of Inducible Nitric Oxide Synthase by Mouse, Rat, and Human OsteoblastLike Cells and Its Functional Role in Osteoblast Metabolic Activity* MIKA HUKKANEN?, FRANCIS J. HUGHES, LEE D. K. BUTTERY, STEVEN S. GROSS, TOM J. EVANS, SHAUN SEDDON, VALENTINA RIVEROS-MORENO, IAIN MACINTYRE$, AND JULIA M. POLAK Department of Histochemistry (M.H., L.D.K.B., J.M.P.), Royal Postgraduate Medical School, Du Cane Road, London W12 ONN, United Kingdom; Department of Periodontology (F.J.H., S.S.), London Hospital Medical College, London El 2AD, United Kingdom; William Harvey Research Institute (I.M.), St. Bartholomew$ Medical College, London EC1 6BQ, United Kingdom; Department of Pharmacology (S.S.G.), Cornell University Medical College, New York, New York 10021; Department of Infectious Diseases and Bacteriology (T.J.E.), Royal Postgraduate Medical School, London W12 ONN, United Kingdom; and Wellcome Research Laboratories (V.R.-M.), Langley Court, Beckenham, Kent BR3 3BS, United Kingdom ABSTRACT

NOS activity and by dexamethasone. IL-lp, TNF-ol, and bacterial lipopolysaccharide were found to have weak stimulatory effects on nitrite production on their own. However, IL-lp and TNF-cx showed strong synergy with IFN-7, but, surprisingly, lipopolysaccharide was found to exert potent inhibitory effects on IFN-y-induced nitrite synthesis. Basal production of nitrite and induction of its synthesis was similarly observed with primary rat osteoblasts as well as ROS 17/2.8, MC3T3-El, and MG-63 cell lines. Cytokine-induced NO production significantly reduced osteoblast activity, as was evidenced by inhibition of DNA synthesis, cell proliferbtion, alkaline phosphatase activity, and osteocalcin production. The results provide evidence for a basal expression of iNOS activity and show that the iNOS messenger RNA, protein, and enzyme activity are all induced by cytokines across the species. The data further suggest that osteoblast-derived NO may have an important role in mediation of localized bone destruction associated with inflammatory bone diseases such as rheumatoid arthritis. (Endocrinology 136: 5445-5453, 1995)

Recent evidence suggests that the production of nitric oxide (NO) may have important roles in the regulation of osteoblast and osteoclast metabolism. The present study was performed to investigate the effects of interleukin-lp (IL-lp), tumor necrosis factor-a (TNF-a), and interferon-y (IFN-y) on the expression of inducible NO-synthase (iNOS) and to measure high-output production of NO by primary rat osteoblasts and osteoblastic cell lines ROS 17/2.8, MC3T3-El and MG-63. In addition, we have investigated if NO may mediate some of the effects of these cytokines on osteoblast metabolism. Northern blots and immunocytochemistry revealed time-dependent iNOS messenger RNA and protein expression in primary rat osteoblasts in response to cytokine treatment. Reverse transcription polymerase chain reaction amplified an 807-base pair (bp) product from ROS 17/2.8 cells, which had a size and restriction enzyme-cut pattern identical to that predicted for authentic rat iNOS. Nitrite accumulation in culture medium was induced by IFN-7 in a time- and dosedependent manner and inhibited by cotreatment with inhibitors of

B

ONE metabolism is closely regulated by hormones and cytokines, which have effects on both bone resorption and deposition. Under physiological conditions, these processesare carefully coordinated such that deposition is coupled to resorption. In metabolic and inflammatory bone diseases,these processesmay become uncoupled, resulting in loss of bone massas can be seenin postmenopausal osteoporosis, Paget’s diseaseof bone, rheumatoid arthritis, and periodontal disease.Proinflammatory cytokines such as inReceived June 20, 1995. Address all correspondence and requests for reprints to: Mika Hukkanen, Ph.D., Department of Histochemistry, Royal Postgraduate Medical School, Hammersmith Hospital, Du Cane Road, London W12 ONN, United Kingdom. * Support in part was obtained from the BLJI’A Medical Foundation and from the Medical Research Council (Grant No. G9409774MA). t Recipient of the Emil Aaltonen Research Foundation Fellowship. $ Supported by the Pinewood Foundation.

5445

terleukin-lb (IL-l/3) and tumor necrosis factor-a (TNF-a) can act as powerful local stimulatory signals for bone resorption (1, 2). These factors function via interactions with receptors on the osteoblasts,which modulate osteoclastfunction by the release of unknown signals named osteoclast resorption stimulating activity (3). IL-lb and TNF-a may also inhibit osteoblast activity, suggesting a mechanism for the uncoupling of bone resorption and deposition during inflammation. Mechanisms by which these factors may act are not yet well understood. Nitric oxide (NO) is a cell-signaling molecule with diverse roles in the regulation of vascular tone, neurotransmission, host defense, and stimulation-secretion coupling (4, 5). NO is produced from a guanidino-nitrogen of L-arginine and dioxygen by at leastthree different isoforms of NO-synthases (NOS), arising from distinct genes (5). Two of these genes encode NOS isoforms that are closely regulated by levels of

5446

EXPRESSION

OF NOS BY OSTEOBLASTS

intracellular calcium and are constitutively expressed in cells of endothelial and neuronal origin. A third NOS gene encodes a protein whose expression is induced by several cytokines via a mechanism involving de nova gene transcription and translation. The inducible isoform of NOS (iNOS) produces large quantities of NO and can modulate the activity of key metabolic enzymes by covalent-binding to heme-iron and iron-sulfur complexes in enzyme-active centers. This isoform has been demonstrated in several cell types relevant to the musculoskeletal system including vascular smooth muscle cells, endothelial cells, macrophages, granulocytes, and fibroblasts (4, 5). All isoforms require NADPH, tetrahydrobiopterin, flavin adenine dinucleotide, and flavin mononucleotide, as cofactors (6-8). NO has been shown to function on osteoclasts in a similar manner to calcium by causing retraction of the cell; this major change in the cell behavior is thought to precede movement of the cell to a fresh site of bone to be resorbed. This action indicates a role for NO in osteoclastic motility and, therefore, resorption activity (9). Several groups have reported a significant effect of prototypic NOS inhibitors in suppression of arthritis and local bone destruction in experimental rat models (10,ll). These findings have suggested a powerful role for NO in induction of osteoclast activity but, paradoxically, more detailed studies on osteoclast function have shown that NO may have a preferential role as an inhibitor of osteoclast activity both in vitro and in viva (9, 12, 13). Thus, it is becoming increasingly clear that the regulation of NO production and its effects, whether autocrine or paracrine, are very complex and most probably bidirectional, depending on whether regulated by hormonal or proinflammatory factors. The osteoblast is the principal cell able to modulate osteoclastic bone resorption activity, and the production of NO by osteoblasts has recently been demonstrated by us and others (13-15). In this study, we aimed to investigate the effects of IL-lp, TNF-a, and IFN-y on the expression of iNOS messenger RNA (mRNA) and protein and production of NO by osteoblast-like cells derived from mouse, rat, and human species. Further, we have investigated the hypothesis that NO functions as a postreceptor effector molecule, mediating the effects of these cytokines on osteoblast DNA replication, proliferation, and differentiation. Materials

and Methods

Cell cultures The well characterized osteoblast-like cell lines ROS 17/2.8, MG-63, and MC3T3-El were used in all studies. The ROS 17/2.8 line is derived from a rat osteosarcoma, MG-63 from a human osteosarcoma, and MC3T3-El is a clonally-derived mouse line. Primary osteoblast-like cells were obtained by sequential collagenase digestion of rat calvaria (RC cells) from neonatal rats according to the method of Wong and Cohn (16). Osteoblast-enriched cell populations derived from the third to fifth digestion were used for studies. In addition, rat primary vascular smooth muscle cells and mouse macrophage-like cell line J774 were used in some experiments. All cells were cultured in 75 cm2 culture flasks in alpha modified Eagle’s medium supplemented with 15% fetal bovine serum, 50 &ml v ascorbic acid, penicillin, and streptomycin (a-MEM). Cells were maintained at 37 C in a humidified atmosuhere of 95% air and 5% CO,. and at confluence cells were either subcul&red at a ratio of 1:5 by suspezding with trypsin/EDTA or were used for experiments. RC and vascular

Endo. 1995 Vol 136 . No 12

smooth muscle cells were used as primary cultures only. Confluent cell cultures were used in all experiments unless otherwise stated. For experiments, cells were stimulated with recombinant human interleukin-lS (Sigma Chemical Co., Poole, UK; specific activity 5 x lo7 IU/mg) and recombinant human TNF-cu (Sigma; specific activity 1 x lo7 IU/mg). Recombinant mouse IFN-y (Genzyme, West Malling, Kent, UK; specific activity 1 X lo7 U/mg) was used for stimulation of RC cells, ROS17/2.8, MC3T3-El, vascular smooth muscle and J774 cells, and recombinant human IFN-y (Genzyme; specific activity 2.5 X lo7 U/mg) for MG-63 cells. Lipopolysaccharide (LPS) was from Escherichia coli (Olll.B4; Sigma) and was used at concentrations of O-100 pg/ml.

RNA isolation

and Northern

blot analysis

Primary rat osteoblast cultures (RC cells) were incubated with IL-lp (100 U/ml), TNF-(U (1 rig/ml), IFN-7 (200 U/ml), or with a combination of the cytokines for 0, 6, 12, or 24 h. Total RNA was extracted from RC cells by a modified guanidinium isothiocyanate method (RNAzol; Biotex Laboratories, Houston, TX). All RNA samples had an A260/A28L1 ratio > 1.50. For Northern blots, aliquots containing 20 PLg of total RNA were electrophoresed in 1% agarose gel containing 3% formaldehyde. RNAs were blot-transferred to Hvbond-N hvbridization transfer membranes (Amersham International, kmersham,‘Bucks, UK) and were baked at 80 C to cross-link the RNA. Membranes were prehybridized for 6 h at 42 C in a solution of 50% formamide, 5 X SSPE, 2 X Denhardt’s reagent, and 0.1% SDS. The membranes were then hybridized for 16 hat 42 C with 50 ng of 32P-labeled (Megaprime random DNA labeling system, Amersham) iNOS probe [602 base pair (bp) complementary DNA (cDNA) fragment corresponding to bp 1084-1686 of the mouse macrophage iNOS] in the same buffer and temperature as for prehybridization. After stringency washes, the membranes were exposed to autoradiographic film (Reflection NEF-485; DuPont NEN Products, Stevenage, Herts, UK) for 6 days at -70 C.

Reverse transcription

polymerase

chain

reaction

(RT-PCR)

Total RNA was extracted from ROS 17/2.8, MC3T3-El, mouse macrophage-like cells (J774 cell line), and primary rat vascular smooth muscle cells as described above. The first strand cDNA was synthesized using random primers and Murine-Moloney Leukemia Virus reverse transcriptase @omega, Madison, WI) using 1 big of RNA template followed by PCR amgification using synthet; genespecific prime& for mouse iNOS (17). Primers used were: forward 21-mer. 5’-CTGCAGGTCTTTGACGCTCGG-3’; reverse 21.mer, 5’-GTGGAACACAGGGGTGATGCT-3’. PCR amplification was performed using a DNA PCR kit (Perkin Elmer Cetus, Norwalk, CT) according to the after schedule: denaturation, annealing, and elongation at 95, 55, and 72 C for 30 set, 30 set and 1 min, respectively, for 30 cycles. PCR products were electrophoresed on a 1.5% agarose gel containing ethidium bromide and visualized by UV-induced fluorescence.

Immunocytochemistry

and antiserum

Four-well microscope slides were coated with 5 Kg/ml fibronectin and lOO-~1 droplets of RC cells and ROS 17/2.8 cell suspensions containing 2 X lo4 cells/ml in (u-MEM were placed on the slide wells and incubated for 24 h, followed by a further 72 h incubation with or without IFN-y (200 U/ml) or with a combination of IL-l/3 (100 U/ml), TNF-ol (lng/ml), and IFN-y (200 U/ml). Cytokines were added into the culture medium 24,48, or 72 h before termination of experiment by fixation of the cells in 1% paraformaldehyde. Cells were treated with 0.3% hydrogen peroxide in methanol for 20 min, then rinsed in PBS. The antiserum used was raised in rabbits against the peptide sequence QNGSPQLLTGTAQNVPESLDKLHVT (~49, position 48-71) selected from the published sequence of the mouse macrophage iNOS (18). The characterization of the antiserum has been described elsewhere (19). After incubation with the p49 antiserum, diluted 1:750, the cells were incubated with biotinylated goat antirabbit, avidin-biotin peroxidase complex, and immunoreactivity was visualized by incubaiing the cells in-a solution of 3,3’diaminobenzidine (Sigma). Total cells and those that showed immunoreactivity were counted manually in duplicates from four ran-

EXPRESSION domly selected microscopic that were immunoreactive

Determination

fields to obtain the proportion for the p49 antiserum.

of nitrite

OF NOS BY OSTEOBLASTS

of total cells

production

To measure the production of nitrite, a stable end-product of NO oxidation, cells were grown in 96-well plates and incubated with cytokines, LPS (O-100 wg/ml), L-arginine analogs NG-monomethyl-L-arginine (L-NMA; Sigma) (10-1000 FM) or No-nitro-L[]-ARGININE-METHYL ESTER (L-NAME; Sigma) (1 mM), with or without added L-arginine (l-20 mM) or dexamethasone (1 X 10e8 M). Nitrite was measured according to the method described by Green and co-workers (20,21). Briefly, 50 ~1 of medium was incubated with 50 ~1 of Griess reagent consisting of 1% sulphanilamide and 0.1% naphthylethylenediamine dihydrochloride in 2% phosphoric acid for 10 min at room temperature. The absorbance was then measured at 540 nm and nitrite concentrations were determined from a standard curve made using a solution of sodium nitrite dissolved in medium. Fresh culture medium was used as a control to zero the absorbance. In initial experiments, stimulations were carried out on subconfluent cultures, but later experiments were carried out on confluent cultures.

DNA synthesis

and cell proliferation

of alkaline

phosphatase

of osteocalcin

activity

Effects of NO on osteoblastic cell alkaline phosphatase activity were determined in ROS 17/2.8 and MC3T3-El cultures. Cells were grown until confluent in 30-mm culture dishes and then exposed to a combination of IL-lp (100 U/ml), TNF-a (1 rig/ml), and IFN-y (200 U/ml) with or without 1 mM L-NMA for 48 h. In some experiments, the NO-inhibitor L-NAME was used instead of L-NMA. Cells were then harvested in Tris buffer, pH 7.6, with 0.1% nonidet NP40 (Sigma), sonicated, centrifuged, and aliquots of the supernatants were analyzed for the total alkaline phosphatase activity by an automated analysis (Monarch 2000, Monarch Instrumentation Laboratories, Lexington, MA). Total protein concentration in supernatants was determined by Bradford assay.

production

Effects of NO on osteocalcin synthesis were determined in cultures of human MG-63 osteoblast-like cells. Confluent cell cultures were treated with 5 X 10m9 M 1,25-dihydroxyvitamin D3 together with treatments as described for alkaline phosphatase assays above for 48 h, after which medium was collected in PBS containing 400 Kallikrein inhibiting units/ml aprotinin (Sigma) and stored at -70 C until assayed. Aliquots of medium were analyzed for osteocalcin by RIA kit for human osteocalcin according to the manufacturer’s protocol (CIS Bio International, Gif-sur-Yvette, France).

Statistical

analysis

All data represent the mean values obtained from a minimum of two independent experiments with at least three replicates in each experiment. Data were analyzed using Student’s unpaired t test, or, where more than two groups were compared, a one-way analysis of variance (ANOVA) with Tukey’s or Bonferroni’s posttests. A P value less than 0.05 was taken as a significant difference between the pairs. The values are presented throughout as mean + SEM.

Results

assays

To assess the effects of NO production on osteoblast function, cells were incubated with cytokines with or without inhibitors of NOS enzymatic activity, and effects on DNA synthesis and cell proliferation were determined. To assess the effects of NO on cell proliferation, three replicates of ROS 17/2.8 and MC3T3-El cells were plated in 30-mm culture dishes at a concentration of lo4 cells/ml and continuously exposed to a combination of IL-lfl(100 U/ml), TNF-cu (1 rig/ml), and IFN-y (200 U/ml) with or without 1 mM L-NAME. The increase in cell numbers was determined in replicate cultures after 0, 1, 3, and 5 days by suspending the cells in trypsin/EDTA and counting total cell numbers by electronic cell counter (Coulter Electronics, Luton, Bedfordshire, UK). In some experiments, proliferation was assessed by measuring DNA synthesis. Mid log-phase RC cell cultures in 96-well plates were stimulated with cytokines i 1 mM L-NAME for 24 hand 10 FM bromodeoxyuridine (BrdU; Sigma) added to the medium for the last 6 h of culture. BrdU incorporation was then measured by a modified enzyme-linked immunosorbent assay technique using a commercially available kit (Cell Proliferation Assay, Boehringer Mannheim, East Sussex, UK). In further experiments, human MG-63 cells were plated in 30.mm culture dishes at a concentration of lo4 cells/ml and were grown for 4 days so that they were in mid-late log-phase of growth. For the last 24 h, cells were incubated with 100 U/ml IL-If3 -C 1 mM L-NMA or with L-NMA only; for the final 2 h cells were pulse labeled with 10 FM BrdU. Cells were then suspended in trypsin/EDTA and cytospin preparations were made on glass microscope slides. Preparations were immunostained with a monoclonal antibody to BrdU (DAKO) using the avidin-biotin peroxidase complex detection system with 3,3’diaminobenzidine substrate. All nuclei were then counterstained with the fluorescent DNA dye 4’.6diamidino-2-phenylindole dihydrochloride (DAPI) (Sigma), and the total number of cells and numbers of BrdU-labeled cells were counted in representative microscopic fields by image analysis (Seescan Ltd., Cambridge, UK)) to determine a labeling index (proportion of total cells that were in S phase).

Determination

Determination

5447

Analysis of iNOS and RT-PCR

mRNA

levels by Northern

blots

RC osteoblast cultures were exposed to IL-lp, TNF-a, IFN-y, or to a combination of the cytokines for 0,6,12, or 24 h. Northern blot analysis, using a 602-bp cDNA fragment corresponding to bp 1084-1686 of the mouse macrophage iNOS, showed induction of iNOS mRNA in cultures exposed to a combination of all three cytokines; treatment of primary osteoblasts with single cytokines failed to reveal any detectable mRNA levels. The mRNA was characterized as a single band of approximately 4.4 kilobases (kb), and its expression was found to be maximal at 12 h of stimulation followed by a decline after 24 h of stimulation (Fig. 1A). RT-PCR, using specific primers for iNOS, amplified an 807-bp PCR product from ROS 17/2.8 osteoblast-like cells. Low levels of iNOS mRNA were detectable from unstimulated cells. After incubation for 6 h with a combination of IFN-y and LPS, a substantial increase in iNOS mRNA was observed (Fig. 1B). The 807-bp PCR product from the ROS 17/2.8 cells was confirmed as NOS, in that it was cut into the predicted size fragments by EcoRI (155 and 652 bp; Fig. 1B). Similarly, RT-PCR, using the same set of primers, amplified an 807 bp PCR product from the mouse osteoblastic cell line MC3T3-El that was cut into three fragments (155, 307, and 345 bp) by EcoRI as predicted for mouse iNOS (not shown). Omission of the reverse transcription step resulted in no amplification product. The primers used in the present study were based on the published mouse iNOS cDNA sequence (17). Nonetheless, using this set of primers, we amplified the predicted 807-bp product not only from a mouse macrophage-like J774 cells but also from rat primary vascular smooth muscle cells that we confirmed by subcloning and sequencing and found to be identical to the recently cloned iNOS cDNA from cytokine-activated rat vascular smooth muscle cells (22). The diagnostic pattern of restriction cuts with EcoRI of iNOS PCR products was identical for both rat cells (ROS 17/2.8 and primary vascular smooth muscle cells), and mouse cells (MC3T3-El and J774 macrophage-like cells), respectively (not shown).

EXPRESSION

5448

OF NOS BY OSTEOBLASTS

Endo Vol 136.

l

1995 No 12

TABLE 1. Time-course study on effects of cytokines on iNOS protein expression by RC cells Control

Treatment

24

h

48

h

72

h

IFN- y 25.8 (1.8) 58.6 (4.7)” 80.4 (1.6)” 26.3 (2.7) IFN-y + IL-16 25.8 (1.8) 72.3 (2.8ja 91.7 (0.9)a 34.0 (2.8) + TNF-ol IFN-y was used at 200 U/ml, IL-16 at 100 U/ml, and TNF-a! at 1 rig/ml. Results are percentage ratios of cells immunoreactive for p49 iNOS antiserum and were counted in duplicates from four randomly selected fields. Mean percentage + SEM are shown. a P < 0.001 by ANOVA and Bonferroni’s posttest.

STD

CON

L/I

UI + Eco RI

tion, the percentage ratio of cells immunoreactive for the p49 antiserum declined dramatically, there being no differences in protein expression compared with the control cultures (P > 0.05; Table 1). Controls in which nonimmune rabbit serum was used instead of the p49 antiserum did not show any immunoreactivity. Nitrite

2176 1766

-

653 517 453 394 296 234

-

as a measure

of

NOS activity

In initial experiments IFN--y was found to induce nitrite production by ROS 17/2.8 cells in a time- and dose-dependent manner in both subconfluent and confluent cultures (Fig. 2). Unstimulated ROS 17/2.8 cellsproduced nitrite near to the lower threshold levels of detection, but this was not found to change significantly over the time-course (Fig. 2). In subconfluent cultures, IL-1 p or TNF-a! alone did not have any significant effects. L-WA, the prototypic inhibitor of NOS activity, was effective in blocking IFN-7 induced nitrite production at concentrations of 10-1000 PM, with optimal effects

FIG. 1. Induction of iNOS mRNA by immunostimulants. A, Timecourse study of rat primary osteoblast (RC) iNOS mRNA induction after exposure to a combination of IL-16 (100 U/ml), TNF-ol (1 r&ml), and IFN-7 (200 U/ml) for 6-24 h or from unstimulated control cells (0 h). Northern blot analysis showed significant induction of iNOS mRNA after 6 h of treatment, maximal expression was seen at 12 h followed by a decline after 24 h of stimulation. The mRNA band was seen as a single band at approximately 4.4 kb corresponding to authentic iNOS mRNA (4.5 kb). An identical pattern was obtained in one additional independent experiment. B, RT-PCR analysis of osteoblastic (ROS 17/2.8) iNOS mRNA. Total RNA was prepared from untreated cells (CON, lane 2) or from cells 6 h after treatment with IFN--r (50 &ml) and LPS (30 pg/ml) (L/I, lane 3). RNA (1 pg) was amplified by RT-PCR using primers for iNOS. Lane 4 (L/I + EcoRI) shows the restriction cut of the PCR product by EcoRI. DNA size markers are shown in the extreme left lane. An identical pattern of amplification was obtained in two additional experiments.

Analysis of iNOS protein immunocytochemistry

production

expression

60 -

by

Immunocytochemistry, using antibodies to a peptide sequence specific for mouse macrophage iNOS, showed weak to moderate labeling of many osteoblasts(RC and ROS 17/ 2.8 cells) in unstimulated cultures. Incubation of cells with IFN--y alone, or with a combination of IL-l& TNF-a and IFN--y, resulted in a time-dependent increase in numbers of labeled cells, reaching a maximal level after 48 h of stimulation (P < 0.001; Table 1). The immunoreactivity was characterized as granular staining in those cells with strong labeling, others showed a more diffuse and weaker cytoplasmic pattern of immunoreactivity. After 72 h of stimula-

70

-

60

-

50

-

40

-

30

-

20

-

10

-

01 0

24

46

72

Time (hours)

2. Time-course and dose-response study of nitrite production by confluent cultures of ROS 17/2.8 cells after the addition of 10-1000 U/ml of IFN-y in the culture medium for 12-72 h. Nitrite accumulation reached its maximum level after 24 h for 100 U/ml IFN-y (open circles) and 48 h for 10 (closed rhombi) and 1000 U/ml of IFN-y (closed squares). This time point was selected for further experiments on nitrite production. Unstimulated cells showed basal level nitrite production which did not change significantly during the time (open FIG.

squares).

EXPRESSION

OF NOS BY OSTEOBLASTS

seenat a concentration of 100pM (3 2 0.7 PM). The inhibitory effect of L-NMA was reversed by L-arginine in a dose-dependent manner (Fig. 3). All subsequent experiments were carried out on confluent cultures as these proved to be more responsive for cytokine-induced nitrite production. The production of nitrite was compared among the cell lines ROS17/2.8, MC3T3-El, and MG-63 and osteoblast enriched RC cultures. In these experiments, confluent cultures were exposed to lower concentrations of cytokines than used previously in nitrite assays, including IL-lp (100 U/ml), TNF-(r (lng/ml), and IFN--y (200U/ml) for 48 h before nitrite determination. All cell types tested showed low levels of basal nitrite production, which was significantly induced by cytokine stimulation (Table 2). In each case,maximum stimulation was seen with stimulation by a combination of the cytokines. Cytokine-induced nitrite production was inhibited by addition of 1 mM L-NAME, and this inhibitory effect was reversed by the addition of 20-fold excessof L-arginine. In two of the lines tested (ROS 17/2.8 and MG-63), maximum stimulation was seen in the presence of exogenous excess L-arginine, although L-arginine treatment alone did not alter nitrite levels above the control values (data not shown). Furthermore, treatment of cells with 1 X lop8 M dexamethasone either totally (ROS 17/2.8 and MG-63) or partially (MC3T3-El and RC cells) inhibited nitrite production. Comparison of responsesbetween different cell lines showed that MC3T3-El and RC cells produced substantially more nitrite than the other cell types. Furthermore, MG-63 cells showed a marked responseto IL-l/3 and TNF-a stimulation on their own, which was not seen in the other cell lines examined (Table 2). As IL-lp and TNF-(r have been shown to synergistically

TABLE nitrite

2. Effects in confluent

L-arginine

addition

(mM)

FIG. 3. Dose-response study of [scapl-arginine on the reversal of inhibition of nitrite production (50 rig/ml IFN-7) caused by 300 FM L-NMA in subconfluent cultures of ROS 17/2.8 cells. The effects of L-NMA were studied at a concentration range of O-1000 pM. A concentration of300 pM was found to cause the maximal effect, which was reversed by addition of L-arginine in a dose-dependent manner. See also Table 2 for data on the effects of L-NAME on inhibition of nitrite production in different cell lines and primary rat osteoblast cultures.

of IL-l& osteoblast

TNF-o, and IFN-y cell cultures

Treatment

ROS

1712.8

RC

Control L-NAME IL-16 TN@-o IFN- y All All + L-NAME All + L-NAME

3.2 0.0 5.2 12.7 17.7 18.8 0.0 34.2

(2.4) (0.8) (4.4) (2.3j (3.8jb (2.0jb (1.1) (2.5)

9.7 (2.7) 5.1C1.9) 11.3 (3.7) 15.6 i4.8j 70.8 (6.3) 101.2 (1.7)” 22.6 (4.4) 90.9 (2.2Y

All

+A%

+ Dex

0.0 (1.0)

66.5 (3.0)

on production

MC3T3-El

26.1 13.0 32.2 48.8 82.8 117.0 43.7 116.0

(3.6) (3.5) (2.5) (3.3) (6.8)’ (S.lP (1.5) (7.6)”

42.3 (7.5)

of

MG-63

15.0 16.3 34.5 33.8 25.3 34.2 11.5 68.3

(4.4) (4.2) (4.1)” (1.1)” (3.9) (3.8)” (5.1) (4.1)

19.3 (5.0)

All, Combination of IL-16 (100 U/ml), TNF-o (1 ng/mlj, and IFN-7 (200 U/ml); L-NAME was used at 1 mM concentration; Arg, 20 mM L-arginine; Dex, 1 X lOme M dexamethasone. Nitrite measurements were performed after 48 h of stimulation and the results are shown in pM concentrations. Results are from two independent experiments with five replicates in each experiment. Mean k SEM are shown. a P < 0.05, b P < 0.01, ’ P < 0.001 by ANOVA and Bonferroni’s posttest (significantly different from controls).

augment the effects of IFN-7 and increaseNO production in other cell types, it was of interest to investigate whether such mechanismswould apply also in osteoblast cultures. For this purpose, RC cells were stimulated with a range of concentrations of IL-l/3 or TNF-a and with a suboptimal dose of IFN-7 (100 U/ml). Although both IL-lp and TNF-(Uon their own generally had small effects on nitrite production, all three cytokines showed strong synergistic effects when used at suboptimal concentrations with each other (Fig. 4, A and B). Incubation of osteoblasts with bacterial LPS alone (10 pg/ml) had a weak stimulatory effect on nitrite production. Surprisingly, IFN-y-induced nitrite production was paradoxically reduced by coincubation with LPS (10 pg/ml) (Fig. 5). Lower LPS concentrations (0.01-1.0 pg/ml) did not affect nitrite production significantly (data not shown). DNA replication

3

5449

and cell proliferation

Stimulation with all cytokines significantly inhibited DNA synthesis in RC cells. However the addition of L-NAME partially reversed the effects of IFN-y and the combination of cytokines on DNA synthesis (Fig. 6). As MG-63 cells, unlike other cell types, showed significant nitrite production in response to IL-lp treatment alone, we investigated whether this responseis related to cell proliferation. In BrdU pulse-labeling experiments with MG-63 mid-late log-phase cells, control cultures showed a labeling index of 46 2 2.8% (mean t SEM). Treatment with IL-l/3 for 24 h reduced the labeling index to 29 ? 4.4%. L-NMA reversed the effects of IL-l/3 to control values (41 -C3.7%, P > 0.1). Treatment of ROS 17/2.8 cells with a combination of cytokines caused a significant decrease in cell proliferation compared with controls at days 3 and 5 (Fig. 7). However, when L-NAME was added with the cytokine mixture, or added alone to cultures, there was no difference in proliferation rate compared with the controls. Cytokine stimulation also reduced cell proliferation in MC3T3-El cultures, which was also reversed by L-NAME treatment to the control levels (data not shown).

5450

EXPRESSION

I

120-

loo-

Ei

+ 0 U/ml IL-1

0

q

+lO U/ml IL-1 +lOO U/ml IL-1

ka

+lOOO U/ml IL-1

OF NOS BY OSTEOBLASTS

Endo. Vol 136

1995 No 12

l

BO-

Ill .+: L .z

60-

40 20

-

10 -

TNF 1 rig/ml

COlWCJ

-

IFN 100 U/ml

I

0

B

24

1

1

48

72

Time (hours) FIG. 5. Time-course of the effect of 10 U/ml)-induced nitrite production in ROS bation of osteoblasts with LPS alone (open ulatory effect on nitrite production, but cells with IFN-+y plus LPS (closed squares) production compared with that of IFN-7

pg/ml LPS on IFN-y (100 17/2.8 cell cultures. Incusquares) had a weak stimcotreatment of ROS 17/2.8 resulted in reduced nitrite alone (closed rhombi).

I

q

Treatment

0

+ L-NAME

# -I-

0 IL-1 100 U/ml

C0nt10l

IFN 100 U/ml

Treatment FIG. 4. cultures. TNF-a treatment TNF-cu icantly

Synergistic effects of cytokines on nitrite production in RC cell A, Addition of varying concentrations of IL-16 to control, (1 rig/ml), and IFN-7 (100 U/ml). *, Significantly different from without IL-l& B, Addition of varying concentrations of to control, IL-16 (100 U/ml), and IFN-y (100 U/ml). *, Signifdifferent from treatment without TNF-CX. TNF

Alkaline

phosphatase

ALL

Treatment

activity

In confluent ROS 17/2.8 cultures, 48 h treatment with a combination of cytokines resulted in inhibition of alkaline phosphatase activity of 20% compared with control values (control, 5.26 + 0.05 IU/mg protein; IL-lp, TNF-a plus IEN-7, 4.2 ? 0.049 IU/mg, P < 0.01). Addition of L-NAME reversed the effects of the cytokines to control values (5.51 -C 0.23%;P > 0.1). Treatment of cultures with L-NAME only did not alter the alkaline phosphatase activity from the control values (5.6 ? 0.32%; P > 0.1). Similar results were obtained using MC3T3-El cultures, with cytokine treatment reducing the activity of alkaline phosphatase by 27%, which was reversed by the addition of L-NAME (data not shown).

IFN

6. Effect of IL-16, TNF-ol, and IFN--r on DNA synthesis in RC cell cultures. BrdU incorporation was measured as arbitrary units by optical density. Cells were incubated with medium only (control), medium supplemented with IL-16 (100 U/ml), TNF-(Y (1 rig/ml), IFN-7 (200 U/ml) or with a combination of cytokines (All), or medium with cytokines plus 1 mM L-NAME. Mean ? SEM of three replicates. *, Significantly different from controls; #, significant effect of L-NAME.

FIG.

Osteocalcin

synthesis

In confluent cultures, MG-63 cells exposed to 1,25-dihydroxyvitamin D, treatment with a combination of IL-lp, TNF-a, and IFN-y resulted in 22% inhibition of osteocalcin synthesis, as assessedby its levels released into the culture

EXPRESSION

0

1

3

OF NOS BY OSTEOBLASTS

5

Day

FIG. 7. Time-course of the effect of IL-l& TNF-a, and IFN-7 on proliferation of ROS 17/2.8 cells. Cells were incubated with medium only (control). medium suaalemented with a combination of IL-18 (100 U/ml), +~F-cK (1 ng/&j, and IFN-y (200 U/ml) (all), or medium with a combination of cytokines plus 1 mM L-NAME (all + L-NAME). Mean ?ZSEM of four replicates. *, P < 0.05 by ANOVA and Tukey’s posttest. 20

1

I

*

Control

All

All + L-NAME

I

I

L-NAME

Treatment

FIG. 8. Effect of IL-l& TNF-(u, and IFN--y on osteocalcin production in confluent human MG-63 cell cultures. Cells were continuously treated for 48 h with medium only (control), medium supplemented with IL-lp (100 U/ml), TNF-(U (1 rig/ml), and IFN-7 (200 U/ml) (all), medium with a combination of cytokines plus 1 mM L-NAME (all + L-NAME), or L-NAME alone. Mean i SEM offour replicates. *, P < 0.05 by ANOVA and Tukey’s posttest. medium (Fig. 8). Addition of L-NAME reversed this inhibitory effect to within the control values. Addition of L-NAME alone without cytokines resulted in increased osteocalcin production compared with those obtained from the control cultures.

Discussion This investigation provides clear evidence for the production of NO by osteoblasts in response to cytokine stimulation. The time-course of iNOS mRNA in stimulated cells is demonstrated by Northern blots and protein expression by immunocytochemistry using iNOS specific probes and anti-

5451

bodies. Production of nitrite, the stable end product of the L-arginine/NO generating pathway, is similarly induced by cytokine stimulation in a time- and dose-dependent manner. The involvement of NO in the nitrite accumulation is supported by the demonstration of inhibition of nitrite production by the synthetic L-arginine analogs L-NMA and L-NAME. The reversal of cytokine-induced effects on osteoblast metabolism by the inhibitors of NOS enzymatic activity further supports the involvement of NO in the regulation of osteoblast metabolism. In addition, the data suggest that iNOS expression and NO production is inhibited by the glucocorticoid dexamethasone. This is consistent with previous reports demonstrating the inhibition of iNOS in macrophages and other cells such as vascular smooth muscle cells (23,24) when exposed to dexamethasone. In contrast to this, Palmer et al. (25) have reported the expression of an isoform of iNOS in rabbit chondrocytes that is induced by IL-l/3 and is not inhibited by glucocorticoids. Our data suggest that the inducible enzyme present in osteoblasts of mouse, rat, and human origins is similar, or closely related, to the enzyme reported in other cell types such as macrophages and vascular smooth muscle cells. These results support other reports that have very recently appeared demonstrating the production of NO by osteoblasts (12,15,26) and provide valuable new data on the regulation of iNOS expression by immunostimulants and information on the possible role of NO as a mediator of cytokine-induced effects on osteoblast metabolism. In the present series of experiments, it was found that the osteoblast-like cells expressed the iNOS protein and enzyme activity, albeit at low levels, before cytokine stimulation. For example, immunocytochemical analysis showed that approximately 26% of the unstimulated RC cells were immunoreactive for the antiserum raised against a peptide fragment of mouse macrophage iNOS. This was quite unexpected because in other cell types, to our knowledge, this basal protein production has not been found. To study more closely whether the osteoblasts might have an isoform of iNOS not reported earlier, we performed RT-PCR with RNA isolated from ROS 17/2.8 cells and MC3T3-El cells. We showed that low levels of iNOS mRNA were detectable in both unstimulated ROS 17/2.8 and MC3T3-El cells and that levels were substantially increased after treatment with immunostimulants. Restriction analysis of the PCR products with EcoRI showed that the diagnostic pattern was identical between rat cells (ROS 17/2.8 cells and vascular smooth muscle cells) and mouse cells (MC3T3-El cells and J774 macrophages). Therefore, it is very likely that the difference in DNA sequence of the PCR product between rat and mouse osteoblasts is due to species differences rather than their origin from distinct genes. Although the production of nitrite in response to cytokine stimulation is only an indirect measurement of NO production, the results obtained are consistent with the more direct measure of induction and time-course of iNOS mRNA and protein and are further supported by the evidence that nitrite production is inhibited by the L-arginine analogues L-NMA and L-NAME and by dexamethasone. Nitrite production was increased in response to a range of concentrations of cytokines and demonstrate clear synergistic effects between com-

5452

EXPRESSION

OF NOS BY OSTEOBLASTS

binations of IL-lj3, TNF-cr and IFN-y. The magnitude of response seen in confluent cultures varied between different cell lines tested and, for example, in the human MG-63 cells, IL-lp and TNF-a had a greater effect than in other osteoblast lines or primary cultures tested. Further work is required to determine if these data represent an interspecies variation or if this is related specifically to the cell lines, independent of species. LPS inhibited IFN-y-induced nitrite production by ROS 17/2.8 cells, although it had a weak stimulatory effect on nitrite production by its own. This finding was unexpected given the ability of LPS to synergize with IFN-y on NO production in other cell types (4). One plausible explanation, although not a subject of this study, is that osteoblasts respond to LPS with high production of other concomitant factors such as TGFP, which has been shown to destabilize iNOS mRNA, decrease its translation, and increase the degradation of iNOS protein (27, 28). The results of the functional assays of osteoblast metabolism in the presence or absence of prototypic inhibitors of NOS activity strongly suggest that NO is an important effector of cytokine-induced regulation of osteoblast function. Cytokines are known to have complex effects on bone metabolism, which include effects on both osteoblast and osteoclast function. In osteoblast cultures, IL-l/3 inhibits alkaline phosphatase expression (29, 301, collagen synthesis (29, 31), osteocalcin synthesis (30), and formation of mineralized bone nodules (32,33). There are conflicting data on the effects of IL-l/3 on osteoblast proliferation including reports of both inhibition (34) and stimulation (30) according to the cell lines and experimental conditions used. Taken together, these data suggest that IL-lp is capable of inhibiting osteoblast differentiation and bone formation and are consistent with reports of inhibition of bone formation in viva (35). TNF-a! and IFN-y also inhibit osteoblast metabolism in vitro and may act synergistically with each other and with IL-1P (2,36-39). Data from the current experiments support the inhibitory effects of these cytokines on osteoblast metabolism as determined by DNA replication, cell proliferation, alkaline phosphatase activity, and osteocalcin synthesis. This is consistent with the known synergy between these cytokines and indicates that NO mediates many of their actions. It is perhaps surprising that the effects of the cytokines were often completely inhibited by inhibitors of NOS activity given the known responses to stimulation with IL-lp, for example, include stimulation of prostaglandin and lipoxygenase metabolite synthesis, tissue plasminogen activator, and induction of other cytokines such as IL-6. These factors in turn are also known to have effects on osteoblast metabolism (40-42). The relationship between these responses to cytokine stimulation and NO production in osteoblasts remains to be investigated. Bone resorption and bone deposition are carefully coupled during physiological bone remodeling and increased synthesis of proinflammatory cytokines such as IL-lp, and TNF-a may result in a decoupling of osteoblast-osteoclast communication (1, 2). The net effect of this activity in inflammatory bone diseases such as rheumatoid arthritis is a severe loss of bone. Furthermore, recent preliminary evidence from experiments investigating levels of IL-l/3 receptor antagonist and TNF-c~ soluble receptors suggests that

Endo . 1995 Vol 136 . No 12

IL-l/3 and TNF-o( may also have an important role in the etiology of bone loss seen in postmenopausal osteoporosis (43). The data presented here suggest that NO formation is an important postreceptor effector mechanism of these processes although the precise role of NO in inflammatory and metabolic bone diseases remains to be investigated further. However, in support of this hypothesis, we have recently reported evidence that human preosteoclastic cell line FLG 29.1 and isolated neonatal rat osteoclasts express both constitutive calcium-dependent and cytokine-induced isoforms of NOS mRNA and protein and contain NADPH-diaphorase activity. In that report, we show that calcium-elicited osteoclast contraction/retraction mechanism is mediated by NO and that L-NMA treatment strongly inhibits physiological osteoclastic resorption activity, suggesting a basal minimum requirement for NO in bone resorption. In contrast to this, we also found that aminoguanidine enhanced osteoclastic bone resorption when the cells were treated with LPS before addition of this NOS enzyme inhibitor in the medium (44). The presence of both isoforms of NOS and the contrasting effects of pharmacological manipulation on osteoclast function is indicative of complex regulation of cell activity and further testifies for the role of NO in both maintaining normal bone homeostasis and mediating proinflammatory cytokine actions on bone resorption. These new findings strongly support the hypothesis that NO mediates cell-to-cell signaling between osteoblasts and osteoclasts in addition to its role as an intracellular signaling molecule involved in mediation of calcium and cytokine effects in both osteoblasts and osteoclasts. In conclusion, we have demonstrated the induction of iNOS protein and its mRNA in several osteoblast cell lines and primary cultures in response to treatment with immunostimulants. Stimulation of cells resulted in the production of the NO end-product nitrite, and this effect was inhibited by the use of the prototypic NOS inhibitors L-NMA and L-NAME and by the glucocorticoid dexamethasone. The functional data obtained suggest that NO is an important mediator of cytokine-induced effects on osteoblast metabolism and further suggest that NO production may have a role in the localized bone destruction associated with inflammatory bone diseases such as rheumatoid arthritis. Acknowledgments We are grateful to Dr. S. Moncada, Wellcome Research Laboratories, for providing his research facilities for our disposal and for valuable discussions of this work. We thank Drs. J. Burrin, Y. Hattori, S. Hill, and T. Umeda for carrying out various analyses during the project.

References 1. Gowen M, Wood DD, Ihrie EJ, McGuire MKB, Russell RGG 1983 An interleukin-1 like factor stimulates bone resorption in vitro. Nature 306:378-380 2. Bertolini DR, Nedwin G, Bringman D, Smith D, Mundy GR 1986 Stimulation of bone resorption and inhibition of bone formation in vitro by human tumour necrosis factors. Nature 319:516-519 3. McSheehy PMJ, Chambers TJ 1986 Osteoblastic cells mediate osteoclastic responsiveness to PTH. Endocrinology 118:824-828 4. Moncada S, Palmer RMJ, Higgs EA 1991 Nitric oxide: physiology, pathophysiology, and pharmacology. Pharmacol Rev 43:109-142

EXPRESSION

OF NOS BY OSTEOBLASTS

5. Nathan C 1992 Nitric oxide as a secretory product of mammalian cells. FASEB J 6:3051-3064 6. Stuehr DJ, Cho HJ, Kwon NS, Weise MF, Nathan CF 1991 Purification and characterization of the cytokine-induced macrophage nitric oxide synthase: an FAD- and FMN-containing flavoprotein. Proc Nat1 Acad Sci USA 88:7773-7777 7. Mayer B, John M, Heinzel B, Werner ER, Wachter H, Schultz G, BShme E 1991 Brain nitric oxide synthase is a biopterinand flavincontaining multi-functional oxido-reductase. FEBS Lett 288:187-191 8. Gross SS, Levi R 1992 Tetrahydrobiopterin synthesis. An absolute requirement for cytokine-induced nitric oxide generation by vascular smooth muscle. J Biol Chem 26725722-25729 9. MacIntyre I, Zaidi M, Towhidul Alam ASM, Datta HK, Moonga BS, Lidbury I’S, Hecker M, Vane JR 1991 Osteoclastic inhibition: an action of nitric oxide not mediated by cyclic GMP. Proc Nat1 Acad Sci USA 882936-2940 10. McCartnev-Francis N. Allen TB. Mizel DE, Albina IE. Xie QW, Nathan &, Wahl SM’1993 Suppression of aithritis b; an inhibito; of nitric oxide synthase. J Exp Med 178:749-754 11. Stefanovic-Racic M, Meyers K, Meschter C, Coffey JW, Hoffman RA, Evans CH 1994 N-Monomethyl arginine, an inhibitor of nitric oxide synthase, suppresses the development of adjuvant arthritis in rats. Arthritis Rheum 37~1062-1069 12. LGwik CWGM, Nibbering PH, van de Ruit M, Papapoulos E 1994 Inducible production of nitric oxide in osteoblast-like cells and in fetal mouse bone explants is associated with suppression of osteoclastic bone resorlltion. I Clin Invest 93:1465-1472 13. Kasten TP, Collin:Osdoby I’, Pate1 N, Osdoby P, Krunowski M, Misko TP, Settle SL, Currie MG, Nickels A 1994 Potentiation of osteoclast bone-resorption activity by inhibition of nitric oxide synthase. Proc Nat1 Acad Sci USA 91:3569-3573 14. Hukkanen M, Hughes FJ, Springall DR, Riveros-Moreno V, Moncada S, Polak JM 1993 Osteoblasts express inducible NO-synthase which in turn mediates osteoblast function. Endothelium [Suppl] 1:ll (Abstract) 15. Ralston SH. Todd D. Helfrich M. Beniamin N. Grabowski I’S 1994 Human ostkoblast-like cells prohuce’nitric oxide and express inducible nitric oxide synthase. Endocrinology 135:330-336 16 Wong GL, Cohn DV 1974 Separation of parathyroid hormone and calcitonin sensitive cells from non-responsive bone cells. Nature 252:713-715 17 Lyons CR, Orloff GJ, Cunningham JM 1992 Molecular cloning and functional expression of an inducible nitric oxide synthase from a murine macrophage cell line. J Biol Chem 2626370-6374 18 Hevel JM, White KA, Marletta MA 1991 Purification of the inducible murine macrophage nitric oxide synthase. J Biol Chem 266: 22789-22791 19 Hamid Q, Springall DR, Riveros-Moreno V, Chanez I’, Howarth P, Redington A, Bousquet J, Godard I’, Holgate S, Polak JM 1993 Nitric oxide synthase induction in bronchial biopsies from asthmatics. Lancet 3421510-1513 20 Green LC, Wagner DA, Glogowski J, Skipper DL, Wishnak JS, Tannerbaum SF 1982 Analvsis of nitrate, nitrite and [15N] nitrate in biological fluids. Anal Bioihem 128:131-138 21 Ding AH, Nathan CF, Stuehr DJ 1988 Release of reactive nitrogen intermediates and reactive oxygen intermediates from mouse peritoneal macrophages. Comparison of activating cytokines and evidence for independent prdduction. J Immunoi141:2407-2412 22 Hattori Y. Gross SS 1993 GTP cvclohvdrolase I mRNA is induced by LPS in vascular smooth mu&e: characterization, sequence and relationship to nitric oxide synthase. Biochem Biophys Res Commun 195:435-441 71 Di Rosa M, Radomski M, Camuccio R, Moncada S 1990 Glucocorticoids inhibit the induction of nitric oxide synthase in macrophages. Biochem Biophys Res Commun 172:1246-1252 24. Rees DD, Cellek S, Palmer RMJ, Moncada S 1990 Dexamethasone prevents the induction by endotoxin of a nitric oxide synthase and L

,

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

36

37 38

39.

40.

41.

42. 43.

b”.

44.

5453

the associated effects on vascular tone. An insight into endotoxin shock. Biochem Biophys Res Commun 173:541-547 Palmer RMJ, Hickery MS, Charles IG, Moncada S, Bayliss MT 1993 Induction of nitric oxide synthase in human chondrocytes. Biochem Biophys Res Commun 193:398-405 Damoulis PD, Hauschka PV 1994 Cytokines induce nitric oxide production in mouse osteoblasts. Biochem Biophys Res Commun 201:924-931 Centrella M, Canalis E 1985 Transforming and non-transforming growth factors are present in medium conditioned by fetal rat calvariae. Proc Nat1 Acad Sci USA 82:7325-7339 Vodovotz Y, Bogdan C, Paik J, Xie QW, Nathan C 1993 Mechanisms of suppression of macrophage nitric oxide release by transforming growth factor-p. J Exp Med 178:605-613 Ikeda E, Kusaka M, Hakeda Y, Yokota K, Kumegawa M, Yamamoto S 1988 Effects of interleukin-l/3 on osteoblastic clone MC3T3-El cells. Calcif Tissue Int 43:162-166 Evans DB, Bunning RA, Russell RGG 1990 The effects of recombinant human interleukin-10 on cellular proliferation and the production of prostaglandin E2, plasminogen activator, osteocalcin and alkaline phosphatase by osteoblast-like cells derived from human bone. Biochem Biophys Res Commun 166:208-216 Beresford JN, Gallagher JA, Gowen M, Couch M, Poser J, Wood D, Russell RGG 1984 The effects of monocyte-conditioned medium and interleukin-1 on the synthesis of collagenous and non-collagenous proteins by mouse bone and human bone cells in vitro. Biochem Biophys Acta 801:58-65 Stashenko I’, Dewhirst FE, Rooney ML, Desjardin LA, Heeley JD 1987 Interleukin-lp is a potent inhibitor of bone formation in vitro. J Bone Miner Res 2:559-565 Ellies LG, Aubin JE 1990 Temporal sequence of interleukin-1 mediated stimulation and inhibition of bone formation by isolated fetal rat calvaria cells in vitro. Cytokines 2:430-437 Hanazawa S, Ohmori Y, Amano S, Kitano S 1986 Human purified interleukin-1 inhibits DNA synthesis and cell growth of osteoblastic cell line MC3T3-El, but enhances alkaline phosphatase activity. FEBS Lett 203:279-284 Nguyen L, Dewhirst FE, Hauschka PV, Stashenko P 1991 Interleukin-lfl stimulates bone resorption and inhibits bone formation in vivo. Lymphokine Cytokine Res 10:15-21 Smith DD, Gowen M, Mundy GR 1987 Effects of interferon-y and other cytokines on collagen synthesis in fetal rat bones. Endocrinology 120:2494-2499 Canalis E 1987 Effects of tumour necrosis factor on bone formation in vitro. Endocrinology 121:1596-2002 Centrella M, McCarthy TL, Canalis E 1988 Tumor necrosis factor-a inhibits collagen synthesis and alkaline phosphatase activity independently of its effect on DNA synthesis in osteoblast-enriched bone cell cultures. Endocrinology 123:1442-1448 Nanes MS, McCoy WM, Marx SJ 1989 Inhibitory effects of TNF-a and interferon-y dn DNA and collagen synthesis by rat osteosarcoma cells (ROS 17/2.8). Endocrinoloav 124:339-345 Raisz LG, Martin TJ 1983 Prostaglan&ns in bone and mineral metabolism. In: Peck WA (ed) Bone & Mineral Research, vol. 2. Elsevier/North Holland, Amsterdam, pp 286-310 Meghji S, Sandy JR, Scutt AM, Harvey W, Harris M 1988 Stimulation of bone resorption by lipoxygenase metabolites of arachidonic acid. Prostaglandins 36:139-141 Hughes FJ, Howells GL 1993 Interleukin-6 inhibits bone formation in vitro. Bone Miner 21:21-28 Hannon RA, Blumsohn A, Galatti H, Feener H, Russell RGG, Eastell R 1993 Cytokine receptors, receptor antagonists and postmenopausal bone loss. J Bone Miner Res [Suppl 1]8:5152 (Abstract) Brandi ML, Hukkanen M, Umeda T, Moradi-Bidhendi N, Bianchi S, Gross SS, Polak JM, MacIntyre I 1995 Bidirectional regulation of osteoclast function by nitric oxide synthase isoforms. Proc Nat1 Acad Sci USA 92:2954-2958