Tumor Necrosis Factor-a Induces Expression of Monocyte Chemoattractant JE via fos and jun Genes in Clonal Osteoblastic MC3T3-El Cells*

The mechanism by which circulating monocytes are attracted to sites of bone remodeling is unknown. We now report that tumor necrosis factor-a (TNF-a), a potent osteotrophic cytokine, was stimulatory for expression of the monocyte chemoattractant JE gene in osteoblastic MC3T3-El cells. TNF-a stimulated this JE gene expression transcriptionally. The presence of JE gene product in conditioned medium of the cyto- kine-treated cells was evidenced by an immunoprecipitation assay with antiserum specific for JE/MCP-1. The stimulated JE gene expression was markedly inhibited by H-7, a potent inhibitor of protein kinase C. Phorbol 12-myristate 13-acetate induced the JE gene expression, and the cytokine-induced JE gene expres- sion was down-regulated by the phorbol ester pretreatment. TNF-a induced expression of both early protoon- cogenes, c-fos and c-jun, in the cells. Antisense oligonucleotides to these oncogenes significantly inhibited by densitometry and is expressed as a percentage of the maxi- mum. An identical experiment, independently performed, triplicate cultures. An identical exDeriment. indeDendentlv performed, gave similar results.

whose amino acid sequence is 68% homologous to the human monocyte chemoattractant MCP-1 (8)(9)(10). MCP-1 is chemotactic for monocytes but not for neutrophils (11)(12)(13)(14). Therefore, this cytokine has an apparently unique property because other well characterized chemotactic factors respond to both monocytes and neutrophils. In view of these findings, the JE gene may thus play a critical role in attracting monocytes to the sites of bone tissues and to the sites of inflammation or injury of the tissues. In accordance with this point, it is interesting to examine whether the osteoblasts express the J E gene and secrete its product into the extracellular environment. We recently showed that transforming growth factor-@ was able to induce gene expression and production of t h e J E gene in mouse osteoblastic . Because TNF-a' is a potent osteotrophic cytokine (16)(17)(18)(19), it is possible that this cytokine might also elicit such a chemoattractant.
In this paper, we demonstrate that TNF-a stimulates induction of the JE gene expression via expression of c-fos and c-jun genes after protein kinase C activation in the MC3T3-E l cells and secretes its gene product into the extracellular fluid. Also, we suggest that the JE gene product may be an important cytokine that contributes to remodeling and inflammation of bone tissues.

MATERIALS AND METHODS
TNF-a-Recombinant human TNF-a produced in Escherichia coli was supplied by the Suntory Institute for Biomedical Research (Osaka, Japan), was purified to homogeneity (>98.9%, determined by SDS-polyacrylamide gel electrophoresis analysis), and had a specific activity of 2.0 X lo6 units/mg protein. The TNF-a preparation contained a trace amount of lipopolysaccharide (0.01 ng/mg protein).
Anti-mouse JE Antiserum-Anti-mouse JE antiserum was prepared from a rabbit immunized with partially purified mouse JE produced by pmt2-JEl0-transfected COS cells and was kindly supplied by Dr. B. J . Rollins of the Dana-Farber Cancer Institute (Boston, MA). The characteristics and specificity of the antiserum were described previously (8).
Cell Culture-Clonal osteoblastic MC3T3-El cells derived from C57BL/6 mouse calvaria (kindly provided by Dr. H. Kodama of the Ohu University, Koriyama, Japan) were cultured in a-MEM (Flow Laboratories, McLean, VA) with 10% fetal calf serum (HyClone, Logan, UT) in plastic dishes at 37 "C and 5% Cog in air and subcultured every 3 days as described previously (20). The cells (3 X lo5) The abbreviations used are: TNF-a, tumor necrosis factor-a; BSA, bovine serum albumin; fMLP, formyl-methionyl-leucyl-phenylalanine; PMA, phorbol 12-myristate 13-acetate; BtzcAMP, dibutyryl cyclic AMP; H-7,1-(5-isoquinolinesulfonyl)-2-methylpiperazine dihydrochloride; HA 1004, N-(2-guanidinoethyl)-5-isoquinolinesulfonamide hydrochloride; TRE, 12-0-tetradecanoylphorbol-13-acetate-responsive element; MEM, minimum Eagle's medium. Incubation Time(hr) FIG. 1. JE gene expression in TNF-a-treated MC3T3-El cells. A, cells were incubated with or without TNF-a (100 ng/ml), and then total RNA was prepared a t various times after the initiation of the cytokine treatment. B, cells were incubated with various doses of TNF-a, and total RNA was prepared after 1 h. Northern blot analysis was performed with JE, and &actin cDNAs were used as probes. An identical experiment, independently performed, gave similar results.  Generation of monocyte chemotactic activity of TNF-a-treated MC3T3-El cells and neutralization of the chemotactic activity by anti-mouse JE antiserum. A, cells were incubated with or without various concentrations of TNF-a. The conditioned media were harvested at 24 h after initiation of the cytokine treatment. B, conditioned medium of cells treated for 24 h with TNF-a (100 ng/ml) was incubated for 2 h a t 37 "C with anti-mouse J E antiserum (10001) or normal rabbit serum. Each test sample was applied to the monocyte chemotactic assay. The results are expressed as mean f S.D. of triplicates. The dashed line represents the activity of lo-* M fMLP, and the straight line represents the control level. Three identical experiments, independently performed, gave similar results. Cont., control.
were cultured in a-MEM with 10% fetal calf serum in 90-mm plastic dishes until near confluence a t 37 "C and 5% COS in air. The cells were then washed and incubated for 24 h in serum-free a-MEM, and then they were treated for the indicated periods in serum-free a-MEM with and without test samples a t various concentrations.
Fractionation of Conditioned Medium by Gel Filtration-Confluent monolayers of the cells were incubated with or without TNF-a (100 ng/ml) in serum-free a-MEM. After incubation for 24 h, the culture supernatants (conditioned medium) were collected, combined, and then concentrated 150-fold with polyethylene glycol 20,000. The concentrated conditioned medium was applied to a Superose prep 12 (Pharmacia) column equilibrated with phosphate buffer (pH 7.2).
Monocyte Chemotactic Assay-Human peripheral blood mononuclear cells were isolated by the method using Ficoll-Hypaque (Pharmacia) and then suspended in 1% BSA containing Gey's balanced salt solution a t a concentration of 5 X lo6 cells/ml. Monocyte chemotactic activity was measured in a 48-well microchamber apparatus with a 5-pm pore size polycarbonate filter (Nucleopore, Costar Corp., Cambridge, MA) as described by Falk et al. (21). After the microchamber had been incubated for 90 min at 37 "C, the filter was removed and fixed in methanol, and then it was stained with Diff-Quik (Kokusai Shiyaku, Kobe, Japan). Monocytes that had migrated through the bottom of the filter were counted in 10 high power fields (x~OOO), and the chemotactic activity was expressed as the mean & S.D. of the total number of migrating monocytes/oil immersion field in triplicate assays. In each experiment, fMLP (Sigma) was used as a positive control.
In a neutralization test, the conditioned medium and a peak fraction for monocyte chemotaxis prepared by gel filtration as described above were treated for 2 h with or without normal rabbit serum or anti-mouse J E antiserum diluted in Cey's buffer containing 1% BSA. Thereafter, the treated sample was incubated for 90 min with human peripheral blood mononuclear cells and then assessed for chemotactic activity.
cDNA Hybridization Probe-Plasmids containing the mouse J E The cells were incubated for 1 h with or without TNF-a (100 ng/ml), and then their nuclei were incubated for 30 min in the presence of 5'-[a-32P]UTP, after which the RNA was isolated. Transcriptional activity assay was performed with J E and @actin cDNAs. pBR322, the vector plasmid, was used as a negative control (Cont.). An identical experiment, independently performed, gave similar results. cDNA sequence were provided by Dr. C. D. Stiles (Harvard University, Boston, MA). Also, plasmids with c-fos and &actin cDNA were obtained from Oncor (Gaithersburg, MD) and those with c-jun came from the American Type Culture Collection (Rockville, MD). The methods used for plasmid preparation were described previously (22).

Mechanism of J E Expressit
Preparation of RNA and Northern Blot Analysis-Total cellular RNA was extracted by the guanidine isothiocyanate procedure (23). It was subjected to 1% agarose electrophoresis and blotted onto a nylon membrane (MSI Magnagraph, Westboro, MA). Then, the membranes were baked and prehybridized for 18 h at 42 "C in 50% formamide, 1% SDS, 5 X SSC (1 X SSC = 0.15 M NaCI, 0.015 M sodium citrate, pH 7.0), 1 X Denhardt's solution (0.02% Ficoll, 0.02% BSA, 0.02% polyvinylpyrrolidone), 0.25 mg/ml denatured salmon on in Osteoblasts by TNF-CY sperm DNA, and 50 mM sodium phosphate buffer (pH 6.5). Restriction fragment hybridization to each cDNA probe labeled with 5'-[a-32P]dCTP (Amersham Corp.) by use of a multiprimed DNA labeling system (Amersham Corp.) was carried out overnight at 42 "C in fresh Denhardt's solution containing 10% dextran sulfate. After hybridization, the membranes were washed sequentially with 2 X SSC/O.l% SDS, 1 X SSC, and 0.1 X SSC, each at 55 "C for 15 min. The membranes were dried and exposed to x-ray film (Eastman) at -70 "C. &Actin was used as an internal standard for quantification of total mRNA in each lane of the gel.
Nuclear Transcriptional (Run-on) Assay-MC3T3-El cells (5 X 10') were treated for 60 min with or without TNF-a (100 ng/ml), and their nuclei were isolated as described previously (24). Transcription initiated in intact cells was allowed to proceed for 30 min at 30 "C in the presence of ~'-[cY-~'P]UTP, and the RNA was isolated and hybridized to slot-blotted cDNA probes (5 pglslot). Blots were hybridized for 72 h and autoradiographed for 3 days. The 8-actin gene was used as an internal standard.

GAA-GCC-CGA-GAA-3') and jun (5'-CGT-TTC-CAT-CTT-TGC-
Preparation of Oligonucleotides-Antisense fos (5"TGC-GTT-AGT-3') unsubstituted oligodeoxynucleotides were synthesized on an automated solid-phase DNA synthesizer (Applied Biosystems, Inc., Foster City, CA) and purified by high performance liquid column the first 18 bases after the AUG sequence of fos (25) and jun (26) chromatography. These nucleotide sequences were complementary to mRNAs, and the corresponding sense oligonucleotides were also synthesized, purified, and then used as a control. These oligonucleotides penetrated into the cells without any treatment, as described by Lokekt al. (27).
Other Rewents-Prostadandin E,. PMA. forskolin. and BkcAMP were purchased from Sigm-a. H-7 and HA 1004 were'obtained from Seikagaku Kougyo (Tokyo, Japan). Curcumin was from Nakarai Tesque, Inc. (Kyoto, Japan). Because the JE gene expression was not induced by lipopolysaccharide (1-100 ng/ml) from E. coli, was not inhibited by the combination of TNF-a and polymyxin B (50 units/ ml), a potent inhibitor of lipopolysaccharide, and was not affected by a combination of the cytokine with the lipopolysaccharide at 0.01, 0.1, or 1 ng/ml, lipopolysaccharide contaminant in the TNF-a preparation cannot be responsible for TNF-a-induced JE gene expression (data not shown).

TNF-a Induces JE Expression in
We then tested for the presence of the monocyte chemoattractant, JE, in the culture supernatant of TNF-a-treated cells. Confluent monolayers of the cells were treated for 24 h with various doses of TNF-a, and then monocyte chemotactic activity and the presence of t h e J E in the conditioned medium were assessed. Monocyte chemotactic activity in the conditioned medium increased in a dose-dependent manner, and the marked activity was neutralized significantly by antimouse JE antiserum (Fig. 2, A and B). Although the data are not shown, TNF-a by itself did not act as a monocyte attractant.

Secretion of Monocyte Chemoattractant JE in Conditioned
Medium by TNF-a-treated MC3T3-El Cells-We fractionated the conditioned medium by MC3T3-El cells by Superose prep 12 gel filtration and then assessed each fraction for the presence of the chemoattractant, JE, using antiserum specific for the mouse JE/MCP-1 protein. Fig. 3A shows the profiles of monocyte chemotactic activity in the fractions of media

FIG. 5. Role of protein kinase C in TNF-a-induced JE gene expression in MC3T3-El cells. A , the cells were
pretreated for 1 h with or without H-7 (50 p~) or HA 1004 (50 PM) and, thereafter, incubated with or without TNF-a (100 ng/ml), and then total RNA was prepared at 1 h after the initiation of incubation. B, the cells were incubated with PMA (100 nM), and total RNA was prepared at the indicated times after the initiation of the incubation. C, the cells were pretreated for 12 h with or without PMA (500 nM) and then incubated in the presence or the absence of TNF-a (100 ng/ml). The total RNA was prepared at 1 h after the initiation of the

FIG. 6. TNF-a-induced c-foo and c-jun gene expression in MC3T3-El cells.
A , the cells were treated with or without TNF-CY (100 ng/ml), and then the total RNA was prepared at the indicated times after the initiation of the incubation. Northern blot analysis was performed with c-fos and c-jun, and &actin cDNAs were used as probes. B, quantification of the c-fos and c-jun gene expression was done by densitometry and is expressed as a percentage of the maximum. An identical experiment, independently performed, gave similar results.
conditioned by TNF-a-treated and -untreated cells. A major peak for the chemotactic activity was detected in fractions corresponding to a molecular mass of 26 kDa. Although minor peaks were also observed, the antiserum had no effect on the chemotactic activity (data not shown). On the other hand, no significant chemotactic activity was detected in fractions of conditioned medium from TNF-a-untreated cell cultures.
We then determined by a neutralization test using antimouse J E antiserum whether the peak fraction for monocyte chemotactic activity depended on the JE gene product. As seen in Fig. 3B, most of the chemotactic activity of the peak fraction was neutralized by the antiserum. These results show that the JE gene product is a predominant cytokine for monocyte chemoattractant of MC3T3-El cells induced by TNF-a. Furthermore, the presence of the JE gene product was examined by immunoprecipitation assay with anti-mouse JE antiserum. Fig. 3C shows that TNF-a-treated MC3T3-El cells secrete the JE gene product.
TNF-a Stimulates Transcriptionally J E Gene Expression in MC3T3-EI Cells-To ascertain whether TNF-a-induced JE gene expression in the cells depended on transcriptional activation, we tested the transcriptional activity by a run-on

TABLE I
Antisense oligonucleotides to c-fos and c-jun inhibit TNF-a-induced monocyte chemotactic activity of MC3T3-El cells The cells were incubated in the presence or absence of 2.5 p M antisense and sense oligonucleotide to c-fos or c-jun, respectively.
After 3 h, TNF-CY (100 ng/ml) was added. Then, after 24 h, their cell medium was collected, and monocyte chemotactic activity was meas- assay using nuclei isolated from the cells treated with TNF-a (100 ng/ml). Figure 4 shows that TNF-a stimulated the transcriptional activity of the JE gene. No significant change was detected in the level of @-actin mRNA, and no hybridization to the vector plasmid was evident.

Role of Protein Kinase C in TNF-a-induced JE Gene
Expression of MC3T3-El Cells-Many cellular stimulators transduce signals to the nucleus by activatingphosphorylation of cellular proteins. At first, we examined the effects of inhibitors of protein kinases, H-7 and HA 1004, on the JE gene expression in TNF-a-treated cells. The cytokine-induced JE gene expression was inhibited completely by 50 pM H-7, an inhibitor having relatively high specificity for protein kinase C, but was not affected by HA 1004, one that is relatively specific for CAMP-dependent protein kinases (Fig.  5A). Although the data are not shown, prostaglandin E2 (100 ng/ml), forskolin (100 nM), and BtzcAMP (100 nM) were unable to induce the JE gene expression.
The results described above suggest the involvement of

A a-induced c-jun gene expression in MC3T3-El cells but not c-fos gene expression. Cells were treated or not
c-fos treated with curcumin (20 p~) in the presence or absence of TNF-CY (100 ng/ ml), and then the total RNA was prepared at 5

min ( A ) and 15 min ( B ) after
Actin the initiation of the cytokine treatment.

Northern blot analysis was performed
TNF-a(lOOng/mI) -+ +with c-fos and c-juri, and p-actin cDNAs were used as probes.
Curcumin ( protein kinase C for TNF-a-induced J E gene expression in the cells. Therefore, we next examined whether the JE gene expression is inducible by PMA, a potent activator of protein kinase C. As shown in Fig. 5B, marked J E gene expression was observed when the cells were treated for 1 h with 100 nM PMA. However, the gene expression was short-lived inasmuch as the expression decreased rapidly after 3 h. Furthermore, the involvement of protein kinase C in the induction of J E gene by TNF-a was examined in the cells pretreated with PMA because such pretreatment is well known to down-regulate the activity of protein kinase C and desensitize the signal transduction by the enzyme. TNF-ainduced J E gene expression was down-regulated in the cells pretreated for 12 h with 500 nM PMA (Fig. 5, C and D). These results suggest that TNF-a stimulates the J E gene expression in the cells by activating protein kinase C.
TNF-a Induces c-fos and c-jun Gene Expression in MC3T3-El Cells-Because PMA also induces gene expression of c-fos and c-jun through protein kinase C and promotes transcriptional activation of many genes, we examined the inducing effect of TNF-a on expression of these oncogenes in the cells.
As shown in Fig. 6, A and B, marked induction of c-fos gene expression by TNF-a appeared at 5 min after initiation of the treatment and, thereafter, decreased gradually, returning to the basal level by 45 min. Furthermore, the cytokine also induced c-jun mRNA, which was at its maximum at 15 min.
TNF-a-induced Monocyte Chemotactic Activity of MC3T3-El Cells Is Inhibited by Antisense Oligonucleotide of c-fos and c-jun Genes-We examined whether one or both of these protooncogenes are necessary for TNF-a-induced monocyte chemotactic activity using antisense c-fos and c-jun oligonucleotides. The cells were pretreated for 3 h with each antisense and/or sense oligonucleotide (2.5 p~) and, thereafter, treated with TNF-a (100 ng/ml). The conditioned medium then was harvested at 24 h after initiation of their treatments and assessed for monocyte chemotactic activity. As shown in Table I, simultaneous addition of antisense oligonucleotides of c-fos and c-jun genes to the cells inhibited TNF-a-induced monocyte chemotactic activity by about 47%. Although a significant inhibition was observed, such a high inhibitory effect was not observed by the separate addition of antisense c-fos or c-jun alone. On the other hand, neither sense oligonucleotide inhibited the monocyte chemotactic activity.
A c-jun/AP-1 Inhibitor, Curcumin, Inhibits c-jun, JE Gene Expression, and Monocyte Chemotactic Activity of TNF-atreated MC3T3-EI Cells-Recently, Huang et a1. (28) indicated that curcumin, a potent inhibitor of tumor promotion, inhibits gene expression of c-jun induced by PMA and suppresses the PMA-induced TRE binding activity of c-jun/AP-1 protein but does not affect PMA-induced c-fos gene expression. Therefore, we examined the effect of the inhibitor on expression of these protooncogenes in MC3T3-El cells induced by TNF-a. As shown in Fig. 7, TNF-a-induced c-jun gene expression was dramatically inhibited by curcumin. However, the inhibitor did not affect the cytokine-induced cfos gene expression.
We used this inhibitor substance to demonstrate the functional role of the c-jun gene in the J E gene expression and the monocyte chemotactic activity of the TNF-a-induced cells. Curcumin inhibited the JE gene expression induced by TNF-a (100 ng/ml) and also significantly reduced the monocyte chemotactic activity in conditioned medium of the cytokine-treated cells (Fig. 8, A and B ) .

DISCUSSION
Monocytes are often observed near areas undergoing bone resorption. Molecules attracting monocytes to bone tissues may be of both physiological and pathological importance in remodeling and inflammation of bone tissues. Although the precise mechanism by which such attractants act is unknown, it is clear that the local generation of a specific chemoattractant for cells of the monocyte lineage is required.
In the present study, we have demonstrated that TNF-a rapidly induces the expression of the monocyte chemoattrac-tant JE gene of mouse osteoblastic MC3T3-El cells, which subsequently release their gene product extracellularly. Our study (15) showed that transforming growth factor& a potent regulator for remodeling and inflammation of bone tissues, was able to induce gene expression and production of JE in the cells. These observations have suggested the physiological and pathological importance of JE in bone tissues, though it is as yet unknown whether the JE gene product is expressed in the bone matrix and plays a functional role in vivo.
Although it is well known that TNF-a regulates gene expression and function of the osteoblastic cells (29-33), the possible expression of monocyte chemoattractant in TNF-atreated osteoblastic cells has not been studied. We have demonstrated here that TNF-a induces gene expression of the monocyte chemoattractant JE in MC3T3-El cells. The JE gene expression was closely associated with the generation of monocyte chemotactic activity because the chemotactic activity in conditioned medium of TNF-a-treated cells was detected in a major peak corresponding to a molecular mass of 26 kDa in gel filtration and was neutralized almost completely with anti-mouse J E antiserum. Furthermore, immunoprecipitation assay confirmed that TNF-a-treated MC3T3-E l cells secrete the J E gene product. We also observed minor peaks having monocyte chemotactic activity in the gel filtration. Although the substances responsible for this chemotactic activity are as yet unknown, extracellular matrix proteins such as collagen, osteocalcin, and fibronectin may be candidates for such activity. These observations strongly suggest that the JE gene is a major monocyte chemoattractant induced in the cells by TNF-a. Also, TNF-a induced the JE gene expression in a primary osteoblast-enriched cell population.' Therefore, the JE gene may be an important factor in the monocyte recruitment in bone remodeling and in the inflammatory reactions of bone tissues.
Many inflammatory cytokines bind to specific cell surface receptors and regulate gene expression through the activation of intracellular messenger systems. Some investigations (34-37) have demonstrated mechanisms of signal transduction expression of cytokine gene by TNF-a. However, because the mechanism of gene expression of the JE gene by TNF-a has not been demonstrated, we explored this mechanism in MC3T3-El cells. TNF-a induced a high level of J E mRNA in the cells at 1 h after initiation of the treatment. The high gene expression was completely inhibited by 50 ~L M H-7 but was not affected by 50 PM HA 1004. PMA also markedly induced the J E gene expression. Because desensitization with PMA is known to result in down-regulation of protein kinase C in cells, such an experimental design was used to determine the role of protein kinase C in TNF-a-induced expression of t h e J E gene in MC3T3-El cells. In our studies, TNF-ainduced J E gene expression was significantly inhibited in the cells pretreated for 12 h with PMA. On the other hand, PGE2, forskolin, and Bt'cAMP, potent activators of CAMP-dependent protein kinases, were unable to induce the gene expression. These results suggest that TNF-a induces the JE gene expression through a protein kinase C-dependent pathway in the cells. It has been well demonstrated that PMA-induced protein kinase C is involved in the regulation of c-jos and cjun. Therefore, we explored the possible involvement of c-fos and c-jun protooncogenes in TNF-a-induced J E expression in MC3T3-El cells.
The sequence of the murine and rat JE gene promoter has been reported (38, 39). The promoter contains TRE, which binds the transcription factor AP-1. The c-jun gene product, AP-1, binds to TRE with high affinity and may mediate increased transcription after activation of protein kinase C.
The formation of c-fos and c-jun heterodimers is very important for high binding affinity for TRE. In this study, we showed that TNF-a markedly induced expression of c-fos and c-jun protooncogenes in MC3T3-El cells. Addition of both antisense oligonucleotides of c-fos and c-jun genes to the cells significantly inhibited TNF-a-induced monocyte chemotactic activity. Because these antisense oligonucleotides were made complementary to the region near the initiation sequence of c-fos and c-jun mRNA, they are presumed to prevent the translation of messages by forming a duplex with the mRNA. Antisense to either c-fos or c-jun alone exhibited a weaker inhibitory activity for TNF-a-induced monocyte chemotactic activity. Some recent studies (40-44) have shown that c-fos/ c-jun heterodimers bind to TRE 25 times more efficiently than c-jun homodimers. These findings support our results that both antisense oligonucleotide of these protooncogenes inhibited more strongly than the oligonucleotide of c-jos or cjun alone. As described above, because J E is a major factor in TNF-a-induced chemotactic activity, the inhibitory effect of antisense oligonucleotides of these protooncogenes on the cytokine-induced monocyte chemotactic activity may act to regulate the JE gene transcription. However, these observations do not provide direct evidence for a functional role of cjun/AP-l in JE gene expression induced by TNF-a. To demonstrate this point, we examined the effect of curcumin, a potent inhibitor of c-jun/AP-1, on TNF-a-induced J E expression. Huang et al. (28) showed that curcumin can downregulate transcriptional activation of the c-jun gene by PMA and can block the increase in TRE binding activity of c-jun/ AP-1 protein. Curcumin (10 WM) significantly inhibited TNFa-induced c-jun gene expression in MC3T3-El cells but did not affect c-fos gene expression, and the inhibitor markedly inhibited the J E expression. These results strongly suggest the importance of TRE binding of c-jun/AP-1 protein in the JE gene expression induced by TNF-a.
In conclusion, our present study has demonstrated that TNF-a induces expression of c-fos and c-jun protooncogenes after activation of protein kinase C and stimulates monocyte chemoattractant JE expression of osteoblastic MC3T3-EI cells.