In murine 3T3 fibroblasts, different second messenger pathways resulting in the induction of NO synthase II (iNOS) converge in the activation of transcription factor NF-kappaB.

Transcription factor NF-kappaB is essential for the induction of nitric oxide synthase (NOS) II (iNOS) by bacterial lipopolysaccharide in murine macrophages (Xie, Q. W., Kashiwabara, Y., and Nathan, C. (1994) J. Biol. Chem. 269, 4705-4708). In 3T3 fibroblasts, agents other than cytokines are efficacious inducers of NOS II expression. In addition to cytokines such as interferon-gamma or tumor necrosis factor-alpha, protein kinase C-stimulating agents such as tetradecanoylphorbol-13-acetate, or cyclic AMP-elevating agents such as forskolin and 8-bromo-cAMP markedly increased NOS II mRNA (measured by Sl nuclease and RNase protection analyses), NOS II protein (determined by Western blotting), and NOS activity (measured by chemiluminescence detection of NO2-). Transforming growth factor-beta1 (which is an inhibitor of NOS II induction in other cell types) potentiated NOS II mRNA expression produced by all inducing agents listed, whereas dexamethasone, pyrrolidine dithiocarbamate and 3,4-dichloroisocoumarin (inhibitors of NF-kappaB activation) suppressed NOS II mRNA induction in response to all stimulants. In electrophoretic mobility shift assays, nuclear protein extracts from 3T3 cells stimulated with any of the inducing agents significantly slowed the migration of an NF-kappaB-binding oligonucleotide, whereas nuclear extracts from untreated control cells did not. These experiments indicate that NF-kappaB is the key control element for the induction of NOS II in response to at least three different second messenger pathways in 3T3 cells.

Transcription factor NF-B is essential for the induction of nitric oxide synthase (NOS) II (iNOS) by bacterial lipopolysaccharide in murine macrophages (Xie, Q. W., Kashiwabara, Y., and Nathan, C. (1994) J. Biol. Chem. 269, 4705-4708). In 3T3 fibroblasts, agents other than cytokines are efficacious inducers of NOS II expression. In addition to cytokines such as interferon-␥ or tumor necrosis factor-␣, protein kinase C-stimulating agents such as tetradecanoylphorbol-13-acetate, or cyclic AMPelevating agents such as forskolin and 8-bromo-cAMP markedly increased NOS II mRNA (measured by S1 nuclease and RNase protection analyses), NOS II protein (determined by Western blotting), and NOS activity (measured by chemiluminescence detection of NO 2 ؊ ).

Transforming growth factor-␤1 (which is an inhibitor of NOS II induction in other cell types) potentiated NOS II mRNA expression produced by all inducing agents listed, whereas dexamethasone, pyrrolidine dithiocarbamate and 3,4-dichloroisocoumarin (inhibitors of NF-B activation) suppressed NOS II mRNA induction
in response to all stimulants. In electrophoretic mobility shift assays, nuclear protein extracts from 3T3 cells stimulated with any of the inducing agents significantly slowed the migration of an NF-B-binding oligonucleotide, whereas nuclear extracts from untreated control cells did not. These experiments indicate that NF-B is the key control element for the induction of NOS II in response to at least three different second messenger pathways in 3T3 cells.
Nitric oxide (NO) 1 is a short-lived bioactive molecule partic-ipating in the physiology and/or pathophysiology of many organ systems (1). The expression of the inducible isoform of nitric oxide synthase (NOS II or iNOS) is regulated mainly at the transcriptional level (2). Inflammatory stimuli such as bacterial lipopolysaccharide (LPS) and cytokines induce the expression of this enzyme in many cell types. Interestingly, in some cells, agents other than cytokines are efficacious inducers of NOS II expression. For example, in rat mesangial cells, cAMPelevating agents stimulate NOS II expression (3). Phorbol ester induction of NOS II has been reported for rat peritoneal macrophages (4). In murine BALB 3T3 fibroblasts, NOS II is expressed in response to forskolin, dibutyryl cAMP, or tetradecanoylphorbol-13-acetate (TPA) (5).
Analyses of the cloned murine NOS II promoter (6 -8) have revealed the presence of numerous consensus sequences for the binding of transcription factors. Of these potentially relevant transcription factors, nuclear factor-B (NF-B) (6,9) and interferon regulatory factor (10,11) have been shown to be functionally important for NOS II induction. The molecular mechanisms utilized by other second messenger pathways are still unclear. In rat mesangial cells, the inhibitor of NF-B activation, pyrrolidine dithiocarbamate (PDTC), blocked NOS II expression induced by interleukin-1␤ (IL-1␤), but not the expression stimulated by 8-bromo-cAMP, suggesting two different induction pathways (12).
In the current study, we attempted to induce NOS II expression in 3T3 fibroblasts via four different second messenger pathways, namely receptor tyrosine kinase, protein kinase C, protein kinase A, and protein kinase G. We characterized the induction processes with modulators of NOS II induction such as transforming growth factor-␤1 (TGF-␤1), dexamethasone, PDTC, and 3,4-dichloroisocoumarin (DCI). The experiments indicate that all NOS II-inducing second messenger pathways are modulated in the same way and all converge in the activation of NF-B as an essential transcription factor.
Cloning of a Murine NOS II and a Murine ␤-Actin cDNA Fragment-Total RNA was isolated by guanidinium isothiocyanate/phenol/chloroform extraction (16) from RAW 264.7 cells induced with 1 g/ml LPS for 16 h. Two g of this RNA were annealed with 0.5 g of oligo(dT) primer (Pharmacia) and reverse-transcribed with Superscript reverse transcriptase (Life Technologies, Inc.) following the manufacture's instructions. Reverse transcriptase-generated cDNA encoding for murine NOS II and murine ␤-actin were amplified using PCR. Oligonucleotide primers for NOS II and ␤-actin were: GACAAGAGGCTGCCCCCC (sense), and GCTGGGAGTCATGGAGCCG (antisense); GTGGGCCGCTCTAG-GCACCAA (sense), and CTCTTTGATGTCACGCACGATTTC (antisense), respectively. They generated PCR fragments corresponding to the murine NOS II cDNA (17) (positions 2612-3170) and murine ␤-actin (18) (positions 25-564). PCR was performed in 100 l of Taq polymerase buffer (Pharmacia), containing 0.2 mM dNTPs, 1.5 mM MgCl 2, 2 units of Taq polymerase, 50 pmol oligonucleotide primers, and reverse transcriptase products (0.10 of the reverse transcriptase reaction). After a initial denaturation step of 95°C for 5 min, 30 cycles were performed (1 min at 95°C, 1 min at 60°C, and 1 min at 72°C). The final extension period at 72°C was 10 min. Amplified cDNA fragments (NOS II, 559 base pairs; ␤-actin, 540 base pairs) were cloned into the EcoRV site of pCR-Script (Stratagene) using the Sure Clone ligation kit (Pharmacia), generating the cDNA clones pCR_NOS II_mouse and pCR_␤-actin_mouse. DNA sequences of the cloned PCR products were determined from plasmid templates using the dideoxy chain termination method with the T7 sequencing kit (Pharmacia).
Preparation of DNA and Antisense RNA Probes-To generate radiolabeled DNA probes for S1 nuclease protection analysis, the cDNA clones pCR_NOS II_mouse and pCR_␤-actin_mouse were restricted with NcoI or BglII, dephosporylated (with calf intestinal alkaline phosphatase, Boehringer Mannheim), extracted with phenol/chloroform, and concentrated by ethanol precipitation. Fifty to one hundred ng of this DNA were labeled with [␥-32 P]ATP using polynucleotide kinase (Pharmacia). The radiolabeled DNA was separated from unincorporated radioactivity using NucTrap probe purification columns (Stratagene). To generate radiolabeled antisense RNA probes for RNase protection assays, the cDNA clones pCR_NOS II_mouse and pCR_␤-actin_mouse were linearized with NcoI or BstEII, extracted with phenol/chloroform, and concentrated by ethanol precipitation. One half of a microgram of this DNA was in vitro transcribed using T3 RNA polymerase (Pharmacia) and [␣-32 P]UTP. After a 1-h incubation, the template DNA was degraded with DNase I for 15 min. The radiolabeled RNA was purified using NucTrap probe purification columns (Stratagene). S1 Nuclease Protection Analyses and RNase Protection Analyses-S1 nuclease protection analyses were performed as described (19,20). Briefly, after a denaturation step at 85°C for 30 min, 20 g of total RNA isolated by the guanidinium isothiocyanate/phenol/chloroform extraction method (16) were hybridized at 52°C for 16 h with 75,000 cpm labeled NOS II DNA probe and 30,000 cpm labeled ␤-actin DNA probe in hybridization buffer (40 mM Pipes, pH 6.4, 400 mM NaCl, 1 mM EDTA, 80% formamide) in a volume of 30 l. The S1 nuclease digestion was started by adding 310 l of digesting buffer (280 mM NaCl, 4.5 mM Zn(CH 3 COO Ϫ ) 2 , pH 4.5, 30 g/ml denatured salmon sperm DNA, and 300 units/ml S1 nuclease). After 20 min at 37°C, the reaction was stopped by adding 65 l of stop-buffer (2.5 M NH 4 -acetate, 50 mM EDTA), followed by a phenol/chloroform extraction. The reaction products were concentrated by ethanol precipitation and analyzed by electrophoresis in denaturing urea-polyacrylamide gels (8 M urea, 6% polyacrylamide gel electrophoresis). The electrophoresis buffer was 1 ϫ TBE (1.08% Tris, pH 8.3, 0.55% boric acid, and 20 mM EDTA). The gels were electrophoresed for 2-3 h, dried, and exposed to x-ray films. The protected DNA fragments of NOS II and ␤-actin were 380 and 150 nucleotides, respectively. RNase protection assays were performed with RNase ONE™ according to the manufacturer's instructions (Promega). Briefly, following denaturation, 20 g of total RNA (prepared as described above) were hybridized with 100,000 cpm labeled NOS II antisense RNA probe and 10,000 cpm labeled ␤-actin antisense RNA probe at 51°C for 16 h in a volume of 30 l. Then the mixture was digested with 5 units of RNase ONE™ for 1 h at room temperature in 300 l. The reaction was stopped with 1% SDS, and the samples were concentrated and electrophoresed as described for the S1 nuclease protection analysis. The protected RNA fragments of NOS II and ␤-actin were 184 and 108 nucleotides, respectively.
Electrophoretic Mobility Shift Assay (EMSA)-NF-B binding activity in the nuclei of control 3T3 fibroblast-or RAW 264.7 cells, or cells incubated for 3 h with one of the NOS II-inducing agents mentioned above were determined by EMSA using the Promega gel shift assay system. Nuclear proteins were extracted from the cells by detergent lysis (21). Ten g of nuclear protein were incubated with 17.5 fmol of 32 P-labeled double-stranded oligonucleotide containing a motif for NF-B binding (5Ј-AGTTGAGGGGACTTTCCCAGGC-3Ј). In some experiments, 1.75 pmol of an oligonucleotide with the putative NF-B binding sequence of the murine NOS II promoter (5Ј-CAACTGGG-GACTCTCCCTTTG-3Ј) were added. The DNA-protein complexes were analyzed on 5% polyacrylamide gels (electrophoresis buffer: 6.7 mM Tris/HCl, pH 7.5, 3.3 mM sodium acetate, 1 mM EDTA), dried, and autoradiographed.
Measurement of NO Production by Chemiluminescence-Confluent 3T3 fibroblasts were cultured for 18 h in DMEM containing 10% fetal bovine serum. Control cells received no additions to the medium; other cells were incubated with LPS (1 g/ml), TNF-␣ (10 ng/ml), TPA (50 ng/ml), or forskolin (100 M). After 18 h, the cell supernatants were collected and aliquots were deproteinized with 2 volumes of ethanol. Following centrifugation, 200 l of the supernatant were injected into a collection chamber containing 100 mM KI in 10 mM sulfuric acid. This strong reducing environment converts NO 2 Ϫ (and nitrosyl compounds) back to NO. A constant stream of N 2 gas carried the NO into a nitric oxide analyzer (Sievers, Boulder, CO) where the NO was reacted with ozone, resulting in the emission of light. The light emission is proportional to the NO formed; standard amounts of NO 2 Ϫ were used for calibration.

Stimulation of Different Second Messenger Pathways
Induced NOS II mRNA Expression-In 3T3 cells, NOS II mRNA was markedly induced with IFN-␥ (100 units/ml) or TNF-␣ (10 ng/ml) (Fig. 1). NOS II expression was also enhanced with TPA (50 ng/ml) or the cAMP-elevating agents forskolin (100 M) or 8-bromo-cAMP (1 mM) (Fig. 2). In contrast, 8-bromo-cGMP (1 mM) was ineffective as a stimulator of NOS II induction (Fig. 2). The phosphodiesterase inhibitor IBMX (250 M) also produced a marked induction of NOS II mRNA in 3T3 cells (Fig. 3). This can be explained by the increase in cAMP, but not cGMP (cf. Fig. 2). LPS (up to 1 g/ml) showed little efficacy in inducing NOS II mRNA (Fig. 3). Thus the stimulation of the receptor tyrosine kinase pathway (by INF-␥, TNF-␣, and possibly LPS), the stimulation of the protein kinase C pathway (by TPA), and the stimulation of the protein kinase A pathway (by forskolin, 8-bromo-cAMP, and IBMX) all induced the transcription of NOS II mRNA in 3T3 fibroblasts.
Because double protected bands for NOS II mRNA were seen in some of the S1 nuclease analyses, RNase protection assays were performed on the same RNAs (using antisense RNA probes derived from the same NOS II cDNA fragment). The RNase protection assays yielded single protected bands and quantitatively similar results as obtained in the S1 nuclease protection assays (data not shown). Therefore, the double bands seen in the S1 assays are unlikely to represent two different NOS II mRNAs.
Small amounts of NOS II mRNA were detected even in the absence of cytokines or stimulants (cf. Figs. 1 and 2). The same phenomenon has been described previously for murine 3T3 cells (5) and human DLD-1 epithelial cells (22). It may either represent a low constitutive expression of this isoform or an autocrine/paracrine induction of these cells by endogenous cytokines (22).
In 3T3 cells, INF-␥ alone produced a marked induction of NOS II mRNA (Fig. 1). There is controversy as to whether INF-␥ alone can induce NOS II in RAW 264.7 macrophages. While this has been described by some authors (8) The signal transduction pathways effective in inducing NOS II expression vary considerably between cell types (and probably species). In many cells, stimulation of the protein kinase C pathway has little or no effect on NOS II induction by itself, but potentiates cytokine induction (24 -26). In 3T3 fibroblasts, it is an efficacious inducing pathway by itself. Stimulators of the protein kinase A pathway alone have been shown to promote NOS II expression in vascular smooth muscle cells and rat mesangial cells (12,27,28). On the other hand, protein kinase A activation seems to inhibit NOS II induction in rat RINm5F insulinoma cells (26). Thus the NOS II-inducing mechanisms seem to be cell-specific, and the stimulation pattern observed in the present study (Figs. 1-3) is unique to 3T3 cells.
Stimulation of Different Second Messenger Pathways in 3T3 Cells Increased NOS II Protein Expression-Similar to the NOS II mRNA expression induced by the various signal transduction pathways, expression of NOS II immunoreactive protein was stimulated by TNF-␣ (10 ng/ml), LPS (1 g/ml), TPA (50 ng/ml), or forskolin (100 M) (Fig. 4). Noninduced 3T3 cells showed no NOS II immunoreactivity in Western blots (n ϭ 3, not shown).

Stimulation of Different Second Messenger Pathways in 3T3
FIG. 1. Double S1 nuclease protection analysis using cDNA probes for murine NOS II and ␤-actin (for standardization). RNAs were prepared from untreated 3T3 fibroblasts (control, Co) and

Cells Enhanced NO 2
Ϫ Production-Incubation of 3T3 fibroblasts with TNF-␣, TPA, or forskolin markedly enhanced the NO 2 Ϫ content in the supernatant of the cells (Fig. 5). LPS was a much weaker stimulant of 3T3 cell NO 2 Ϫ production (Fig. 5). This indicates that NOS II protein and activity is also induced via the receptor tyrosine kinase, protein kinase C, and protein kinase A pathways.

Stimulation of Three Different Second Messenger Pathways in 3T3 Cells Induced Proteins with NF-B Binding Activity-
NF-B is a multisubunit transcription factor that can rapidly activate the expression of genes involved in immune and acute phase responses (29). NF-B is composed mainly of proteins with molecular weights of 50 kDa (p50) and 65 kDa (p65). Both types of proteins share significant homology with the protooncogene c-rel (30 -32). The proteins p50, p65, and c-Rel can interact with each other and, following activation, bind the NF-B response element as homo-or heterodimers (33) (consensus sequence: GGGRNNYYCC) (34). In its unstimulated form, NF-B is present in the cytosol bound to the inhibitory protein I-kB. After induction of cells by a variety of agents, NF-B is released from I-kB and translocated to the nucleus. Agents that have been described as NF-B activators include mitogens, cytokines, and LPS, TPA, and cAMP (29,35). EMSA experiments shown in Fig. 6 demonstrated that nuclear extracts of untreated 3T3 cells contained low concentrations of proteins that bind an oligonucleotide containing the NF-B response element. Incubation of 3T3 cells either with TPA (50 ng/ml), TNF (10 ng/ml) or 8-bromo-cAMP (1 mM) markedly increased the NF-B binding activity (Fig. 6). In 3T3 fibroblasts, TNF-␣ was the most efficacious inducer of NF-B binding activity tested. TPA and cAMP-elevating agents (8-bromo-cAMP or forskolin) were less efficacious in inducing NF-B binding activity; there were no significant differences in efficacy between these two classes of agents. This parallels the NOS II mRNA and NOS II protein expression as well as the NOS activity stimulated by these compounds. The protein-DNA interaction was totally prevented in all cases with an excess of unlabeled double-stranded oligonucleotide containing the NF-B site of the murine NOS II promoter (Fig. 6 and data  not shown). These data suggest that, in 3T3 cells, the receptor tyrosine kinase pathway, the protein kinase A pathway, and the protein kinase C pathway stimulate the activation of transcription factor NF-B. While cytokines such as TNF-␣ can activate NF-B in most cell types, there is a marked inter-cell variability for the protein kinase A and C pathways. For example, in murine RAW264.7 cells, neither the protein kinase A pathway nor the protein kinase C pathway are able to stimulate this transcription factor; they even inhibit NF-B-dependent reporter gene expression in response to LPS (36). In human Jurkat T cells, the protein kinase C pathway, but not the protein kinase A pathway activates NF-B (37). Conversely, in murine J774 macrophages, activators of protein kinase A are effective stimulators of NF-B, whereas protein kinase C activators failed to stimulate this transcription factor (38).
Effect of Dexamethasone on NOS II mRNA Expression-Glucocorticoids such as dexamethasone have been known for some years to inhibit cytokine induction of NOS II activity in various cell types (such as endothelial cells, macrophages, and smooth muscle cells (39 -42)). More recently, this inhibition has also been demonstrated at the mRNA level in several cell types (5,43,44). In a recent communication, Kunz et al. (45) demonstrated in rat mesangial cells that dexamethasone prevented the induction of NOS II activity in response to IL-1␤ and dibutyryl cAMP. Interestingly, NOS II mRNA levels were only reduced when dibutyryl cAMP was used as the inducing agent, but not after IL-1␤. Consequently, these authors postulated that dexamethasone acts at different levels, depending on the stimulus used to suppress NOS II induction in rat mesangial cells (45). In the current study we examined the effect of dexamethasone (5 M) on NOS II mRNA expression in 3T3 cells. We found that the steroid was equally effective against inductions produced by LPS, TPA or IBMX (Fig. 3). Also the NOS II mRNA inductions in response to TNF-␣ (10 ng/ml) or INF-␥ (100 units/ml) were markedly inhibited by dexamethasone (5 M) (n ϭ 3, not shown).
Effect of TGF-␤1 on NOS II mRNA Expression-TGF-␤1 is an inhibitor of NOS II induction in mouse macrophages and rat vascular smooth muscle cells (42, 43, 46 -48). On the other hand, in 3T3 cells and in bovine retinal pigmented epithelial cells, TGF-␤1 has been described as a stimulator of cytokineinduced NOS II mRNA induction (43,49). Also in the current experiments, TGF-␤1 (2 ng/ml) potentiated NOS II mRNA production irrespective of the second messenger pathway used for induction (Fig. 7).
Inhibition of NF-B Activation Blocks NOS II mRNA Induction-The activation of NF-B can be blocked by thiol compounds such as PDTC or diethyldithiocarbamate, which leave the DNA binding activity of other transcription factors (e.g. SP1, Oct, and CREB) unaffected (50). PDTC or diethyldithiocarbamate have been shown to prevent the induction of NOS II in LPS-induced murine macrophages (9, 51) and rat alveolar macrophages (52). Eberhardt et al. (12) reported that PDTC inhibits the induction of NOS II expression in response to IL-1␤, but not to dibutyryl cAMP. They concluded that in rat mesangial cells cAMP-stimulated NOS II expression is activated through a transcription factor different from NF-B. In the current series of experiments in 3T3 cells, PDTC prevented the induction of the NOS II mRNA expression in response to all inducing compounds used (Fig. 8). Also DCI, a serine protease inhibitor, which blocks NF-B activation by inhibiting proteolytic degradation of I-B (53), blocked (by over 90%) NOS II mRNA expression induced by INF-␥ (100 units/ml), TNF-␣ (10 ng/ml), TPA (50 ng/ml), and forskolin (100 M) (n ϭ 3, data not shown). This confirms the results of our EMSA experiments and indicates that in 3T3 fibroblasts NF-B is essential for NOS II induction in response to different second messengers. Interestingly, the inhibition of NOS II induction by dexamethasone (described above) is likely to reflect its ability to inactivate NF-B (20).
In conclusion, our data demonstrate that in 3T3 cells at least three different signal transduction pathways can stimulated NOS II mRNA expression, namely the cytokine/receptor tyrosine kinase pathway, the cAMP/protein kinase A pathway, and the protein kinase C pathway. All these pathways seem to converge in the activation of the essential transcription factor NF-B, which increases the transcription of the NOS II gene. Densitometric analyses of the gels demonstrated that TPA and 8-bromo-cAMP were about equieffective in stimulation NF-B binding activity, whereas TNF-␣ induced the largest increase in NF-B binding activity. Nuclear extracts from LPS-induced RAW 264.7 macrophages are known to contain NF-B binding activity (9) and were used as positive controls (RAW). RAW 264.7 macrophages were induced with bacterial lipopolysaccharide (1 g/ml, LPS), and the same nuclear protein extract from RAW 264.7 cells was tested in the presence of a 100-fold excess of an oligonucleotide containing the NF-B binding site of the murine NOS II promoter (competition experiment, LPS comp.). The gels are representative of three experiments yielding similar results.
FIG. 7. S1 nuclease protection analysis using cDNA probes for murine NOS II and ␤-actin (for standardization). RNAs were obtained from 3T3 cells stimulated with various agents in the absence and presence of transforming growth factor-␤1 (2 ng/ml, TGF-␤1). TGF-␤1 alone did not produced any NOS II induction. The following stimulating agents were used: bacterial lipopolysaccharide (1 g/ml, LPS) and LPS in the presence of TGF-␤1; tetradecanoylphorbol-13acetate (50 ng/ml, TPA) and TPA in the presence of TGF-␤1; the phosphodiesterase inhibitor isobutylmethylxanthine (250 M, IBMX) and IBMX in the presence of TGF-␤1. T, tRNA control; M, molecular weight markers (pGl 2 -Basic restricted with HinfI). The gel is representative of three experiments with similar results. Densitometric analysis demonstrated that the NOS II signal in the LPS ϩ TGF-␤1 lane (after correction by the ␤-actin signal) was 130% of the NOS II signal in the LPS lane; the NOS II signal in the TPA ϩ TGF-␤1 lane was 230% of the NOS II signal in the TPA lane, and the NOS II signal in the IBMX ϩ TGF-␤1 lane was 330% of the NOS II signal in the IBMX lane.