INTRODUCTION

Nitric oxide (NO)¹ is a free radical gas mediating intercellular communications in many mammalian organs. In recent years the importance of NO for the regulation of vascular homeostasis, the involvement in neurotransmission, and the defense against infectious agents has become established (1, 2). Three isoforms of NO synthase (NOS) have been identified and cloned (3, 4). The brain (type I) and endothelial (type III) enzymes are constitutively expressed and their enzymatic activity is regulated by changes in concentration of intracellular free Ca²⁺. The third member of the family of NO synthases is the inducible (type II) NOS. This enzyme is regulated at the transcriptional level and the activity is present at intracellular Ca²⁺ levels. Inducible NOS (iNOS) is expressed in many different cell types and produces high levels of nitric oxide. Excessive formation of NO mediates the bactericidal and tumoricidal actions of macrophages. However, under pathological conditions, high output of NO is associated with tissue damage observed in arthritis, type I diabetes, septic shock, and nephritis (for review, see Refs. 1 and 2).

Glucocorticoids are a class of steroid hormones with pleiotropic effects. At pharmacological concentrations, glucocorticoids are used to prevent and suppress inflammation and the activation of the immune system. Despite their widespread medical use, the precise mechanism(s) for the effectiveness as immunosuppressive and anti-inflammatory drugs is not yet entirely understood. Steroids exert their anti-inflammatory actions mainly by modulation of the transcription of a variety of genes involved in the control of inflammatory processes. These include cytokines and their cellular receptors, adhesion molecules, and enzymes producing inflammatory mediators (for review, see Ref. 5). Inhibition of transcriptional activity of target genes occurs either by binding of the activated glucocorticoid receptor to a negative GRE within the 5-flanking region or by cross-coupling, i.e. direct protein-protein interaction of the glucocorticoid receptor with transcription factors induced under stimulatory conditions (6, 7). Moreover, a few reports suggest that post-transcriptional mechanisms including translation or protein secretion may also be involved in the anti-inflammatory actions of steroids (8, 9).

Several recent publications document the suppression of iNOS expression by glucocorticoids in various cell types such as mesangial cells (10), murine macrophages (11), human endothelial cells (12), insulin-producing RINm5F cells (13), rat hepatocytes (14), murine fibroblasts (15), and human epithelial cells (16). We have shown that dexamethasone blocks iNOS expression in rat renal mesangial cells at transcriptional and post-transcriptional levels (17, 18).

In this report we examined the mechanisms of inhibition of iNOS expression in the IFN--stimulated murine macrophage cell line RAW 264.7 by dexamethasone. The present study establishes that the glucocorticoid suppresses iNOS formation at different levels of iNOS gene expression. We observed a reduction of the transcription rate of the iNOS gene and a decrease in stability of iNOS mRNA causing reduced, but not completely abolished iNOS mRNA steady state levels. More importantly, post-transcriptional mechanisms, notably the translation of iNOS mRNA and the degradation of iNOS protein are involved in the suppression of iNOS expression even when iNOS is already induced. Moreover, we provide evidence that the enzyme mediating increased proteolysis of the iNOS protein by dexamethasone is the cysteine protease calpain I.

EXPERIMENTAL PROCEDURES

The macrophage cell line RAW 264.7 was kindly provided by Dr. F. Erard (Ciba, Basel, Switzerland). Recombinant mouse IFN- was a gift from Dr. G. Adolf (Boehringer Ingelheim, Vienna, Austria). Dexamethasone, N-p-tosyl-L-lysine chloromethyl ketone (TLCK), and L-1-tosylamido-2-phenylalanine chloromethyl ketone (TPCK) were purchased from Sigma. N-Acetyl-Leu-Leu-norleucinal (calpain inhibitor I) and actinomycin D were purchased from Boehringer Mannheim, Germany. Calpain I and II were purchased from Calbiochem-Novabiochem Corp. (Lucerne, Switzerland). Antibodies specific for iNOS were generated using an antigen peptide based on an N-terminal sequence of mouse iNOS as described (17). Polyclonal antibodies for rat calpain II were generated as described (19). The -actin cDNA probe was a gift from Dr. U. Aebi (Maurice E. Müller Institute, Basel, Switzerland).

RAW 264.7 cells were cultured in Macrophage SFM medium (Life Technologies, Basel, Switzerland) supplemented with penicillin (100 units/ml) and streptomycin (100 µg/ml). For stimulation, RAW 264.7 cells were washed twice with phosphate-buffered saline and incubated in DMEM without phenol red (Life Technologies, Basel, Switzerland) supplemented with 0.1 mg/ml fatty acid-free bovine serum albumin (Sigma), with or without agents for the indicated time periods.

NOS activity was measured as nitrite production in RAW 264.7 cell culture supernatants. 100 µl of cell culture supernatant were mixed with 200 µl of Griess reagent (Merck, Darmstadt, Germany). The experiments were repeated four times. The plates were measured in an enzyme-linked immunosorbent assay plate reader (KONTRON Analytical SLT 210) at 550 nm against a calibration curve with sodium nitrite standards. Nitrate was stoichiometrically reduced to nitrite by incubation of sample aliquots for 15 min at 37 °C, in the presence of 0.1 unit/ml nitrate reductase (EC 1.6.6.2; Aspergillus species; Boehringer Mannheim), 50 µM NADPH, and 5 µM FAD, in a final volume of 160 µl. When nitrate reduction was complete, the remaining NADPH (which would interfere with nitrite determination) was oxidized with 10 units/ml lactate dehydrogenase (Boehringer Mannheim) and 10 mM sodium pyruvate, in a final volume of 170 µl for 5 min at 37 °C. The amount of nitrate produced in stimulated cells was approximately 30-35% of that of nitrite.

RAW 264.7 cells were stimulated with IFN- (50 units/ml) in the presence of vehicle, dexamethasone at concentrations of 100 nM and 1 µM or dexamethasone (100 nM) plus RU-486 (1 µM) for 12 h. Then the cells were washed with phosphate-buffered saline and scraped into 1 ml of buffer A (50 mM Tris, pH 7.5, 150 mM NaCl, 10% glycerol, 0.5% Triton X-100, 2 mM EDTA, 2 mM EGTA, 1 µM leupeptin, 0.1 M phenylmethylsulfonyl fluoride). Cells were homogenized in a Dounce homogenizer and centrifuged at 200,000 × g for 30 min at 4 °C. The protein concentration of the lysate was determined by the Bradford protein assay (Bio-Rad). Equal amounts of protein lysate were incubated overnight with 200 µl of a 1:1 (v/v) slurry of a specific polyclonal anti-iNOS antibody coupled to protein A-Sepharose CL-4B. The beads were washed twice with buffer B (50 mM Tris, pH 7.5, 150 mM NaCl, 0.2% Triton X-100, 2 mM EDTA, 2 mM EGTA, 0.1% SDS), twice with buffer B plus 500 mM NaCl, and once with buffer B. Proteins were solubilized from the beads by heating for 5 min at 95 °C in 0.1 ml of Laemmli buffer and subjected to SDS-PAGE (7.5% (w/v) acrylamide gel). Immunoblotting was performed as decribed (17) using an anti-iNOS antibody at a dilution of 1:1000 (v/v).

Total cellular RNA was extracted from RAW 264.7 cells using the Trizol reagent according to the manufacturer's instructions. Samples of RNA (20 µg) were separated on 1% (w/v) agarose gels containing 0.66 M formaldehyde before transfer to GeneScreen membranes. After UV-cross-linking and prehybridization for 4-6 h, the filters were hybridized for 16-24 h to a ³²P-labeled (approximately 1 × 10⁶ cpm/ml) SmaI-digested cDNA insert derived from pGemiNOS coding for murine iNOS (20) or EcoRI-digested cDNA insert derived from pT7-7fN-rat calpain II (19). DNA probes were radioactively labeled with [-³²P]dATP by random priming (Boehringer Mannheim). Hybridization reactions were performed in 50% (v/v) formamide, 5 × SSC, 10 × Denhardt's solution, 1% (w/v) SDS, and 10 µg/ml salmon sperm DNA. Filters were washed three times in 2 × SSC at room temperature for 30 min followed by twice in 0.2 × SSC, 1% (w/v) SDS at 65 °C for 1 h. Filters were exposed for 24 to 48 h to Kodak X-Omat XAR film using intensifying screens. Densitometrical analyses were performed on a Molecular Dynamics densitometer. To assess for variations in RNA loading and transfer, ribosomal RNAs were stained on the blots using methylene blue (21).

RAW 264.7 cells were stimulated with IFN- (50 units/ml) for 12 h. Subsequently, either vehicle or dexamethasone (1 µM) was added and cells were incubated for a further 2 h. Thereafter, actinomycin D (10 µg/ml) was added, total RNA was prepared at the indicated time points and used for Northern hybridization as decribed above.

For the nuclear run-on transcription assay, a nuclei suspension was prepared and mixed with 0.2 ml of 2 × reaction buffer (100 mM Hepes, pH 8.0, 10 mM MgCl₂, 300 mM KCl, 200 units of RNasin (Boehringer Mannheim) per ml per 1 mM each ATP, GTP, and CTP per 150 µCi (1 µCi = 37 kBq) of [³²P]UTP (3000 Ci/mmol; Amersham, Dübendorf, Switzerland) and incubated for 30 min at 30 °C. Transcription was stopped by adding 20 µg of DNase I, followed by 80 µg of proteinase K. The ³²P-labeled RNA was purified by extraction with phenol/chloroform and two sequential precipitations with ammonium acetate. Equal amounts of ³²P-labeled RNA were hybridized in 50% formamide, 5 × SSC, 5 × Denhardt's solution, 1% SDS (1 × SSC = 150 mM NaCl, 15 mM sodium citrate, pH 7.0) at 42 °C for 72 h. Filters contained 10 µg each of linearized plasmids immobilized on GeneScreen membranes (DuPont de Nemours International, Regensdorf, Switzerland) after blotting in 12 × SSPE with a dot-blot apparatus. After hybridization filters were rinsed for 30 min in 2 × SSC at 60 °C, for 5 min in 2 × SSC containing 10 µg of RNase A/ml at 37 °C, and finally for 1 h in 2 × SSC at 37 °C. Filters were air dried and exposed to Kodak X-Omat XAR film for 2 to 4 days using intensifying screens. Densitometrical analyses were performed on a Molecular Dynamics densitometer. The amount of sample hybridizing to -actin was used for normalization.

RAW 264.7 cells were cultured in 150-mm (diameter) culture dishes and stimulated with IFN- (50 units/ml) for 3.5 h. Then vehicle, dexamethasone (100 nM), or a combination of dexamethasone (100 nM) and RU-486 (1 µM) was added for 0, 2, and 4 h, as indicated. Thereafter the medium was aspirated and replaced with pulse medium (DMEM without L- methionine) and the cells were incubated at 37 °C for 45 min for depletion of intracellular stores of methionine. L-[³⁵S]Methionine (300 µCi/dish) in pulse medium was added for 30 min. The medium was removed, the dishes were washed three times with phosphate-buffered saline, and the cells were scraped into 1 ml of buffer A (50 mM Tris, pH 7.5, 150 mM NaCl, 10% glycerol, 0.5% Triton X-100, 2 mM EDTA, 2 mM EGTA, 1 µM leupeptin, 0.1 M phenylmethylsulfonyl fluoride). Cells were homogenized in a Dounce homogenizer and centrifuged at 200,000 × g for 30 min at 4 °C. The protein concentration of the lysate was determined by the Bradford protein assay (Bio-Rad). Equal amounts of protein lysate were incubated overnight with 100 µl of a 1:1 (v/v) slurry of a specific polyclonal anti-iNOS antibody coupled to protein A-Sepharose CL-4B. The beads were washed twice with buffer B (50 mM Tris, pH 7.5, 150 mM NaCl, 0.2% Triton X-100, 2 mM EDTA, 2 mM EGTA, 0.1% SDS), twice with buffer B plus 500 mM NaCl, and once with buffer B. Proteins were solubilized from the beads by heating for 5 min at 95 °C in 0.1 ml of Laemmli buffer and subjected to SDS-PAGE (7.5% (w/v) acrylamide gel). Gels were vacuum dried and exposed to Kodak X-Omat XAR film for 2 to 4 days. Densitometrical analyses were performed on a Molecular Dynamics densitometer.

For metabolic labeling and subsequent immunoprecipitation of calpain, RAW 264.7 cells were incubated with IFN- (50 units/ml), IFN- (50 units/ml) plus dexamethasone (1 µM), or dexamethasone (1 µM) for 12 h in 10 ml of pulse medium (DMEM without L-methionine). Thereafter the medium was aspirated and replaced with pulse medium containing L-[³⁵S]methionine (300 µCi/dish) and incubations were continued in the presence of stimuli for 4.5 h. Lysates were prepared as described above. Immunoprecipitation was performed using a specific polyclonal anti-calpain II antibody coupled to protein A-Sepharose CL-4B and proteins were separated on SDS-PAGE (9% (w/v) acrylamide gel).

RAW 264.7 cells were stimulated with IFN- (50 units/ml) for 6 h. Then dexamethasone (1 µM) was added for 1 h. The medium was aspirated and replaced with 10 ml of pulse medium (DMEM without L-methionine) containing IFN- (50 units/ml) and dexamethasone (1 µM). The cells were starved for L-methionine for 45 min to deplete intracellular stores of methionine. Subsequently the cells were pulsed for 2 h with L-[³⁵S]methionine (300 µCi/dish). Thereafter, pulse medium was aspirated and replaced by 10 ml of chase medium (pulse medium containing 15 mg of unlabeled L-methionine per liter). The cells were further incubated for 0, 2, 4, 8, and 12 h, as indicated. Subsequently, cells were harvested and processed for immunoprecipitation and SDS-PAGE as described above.

In the experiments using protease inhibitors, cells were incubated as above. N-Acetyl-Leu-Leu-norleucinal (calpain inhibitor I, 100 µM) or TLCK (100 µM) were added during the pulse period of 2 h with L-[³⁵S]methionine (300 µCi/dish). Thereafter, pulse medium was aspirated and replaced by 10 ml of chase medium (pulse medium containing 15 mg of unlabeled L-methionine per liter). The cells were further incubated for 0 and 2 h as indicated. Subsequently, cells were harvested and processed for immunoprecipitation and SDS-PAGE as described above.

Statistical analysis was done by Student's t test, and p < 0.05 was used as the criterion for statistical significance.

RESULTS

Incubation of RAW 264.7 cells for 24 h with IFN- (50 units/ml) increases the production of nitrite, one of the stable end products of NO formation, about 66-fold, from 2 ± 0.2 nmol/mg of protein in unstimulated cells to 138 ± 15 nmol/mg of protein (means ± S.D., n = 3). Simultaneous incubation of the cells with IFN- (50 units/ml) and dexamethasone for 24 h dose dependently inhibits the formation of nitrite in cell culture supernatants as shown in Fig. 1. Half-maximal inhibition is observed at 9 ± 0.5 nM dexamethasone (means ± S.D., n = 4).

To assess whether the reduced production of nitrite in dexamethasone-treated RAW 264.7 cells is due to a decrease in iNOS protein levels, immunoblotting experiments using a specific anti-iNOS antibody were performed. Cells were stimulated with IFN- (50 units/ml) in the presence of vehicle or dexamethasone at concentrations of 100 nM and 1 µM or a combination of dexamethasone (100 nM) plus a 10-fold molar excess of the glucocorticoid receptor antagonist RU-486 (1 µM) for 12 h. Subsequently cells were lysed and extracts were used for immunoblotting. As shown in Fig. 2, in unstimulated RAW 264.7 cells no iNOS protein is detected, whereas IFN- strongly induces the formation of detectable iNOS protein of approximately 130 kDa. This induction is blocked under the action of dexamethasone at concentrations of 1 µM and 100 nM, respectively. The effect is reversed upon the simultaneous incubation of the cells with dexamethasone (100 nM) and the glucocorticoid receptor antagonist RU-486 (1 µM; Fig. 2).

Dexamethasone inhibits IFN--stimulated iNOS protein levels.

Northern blot analyses were performed to assess whether the glucocorticoid affects the formation of iNOS mRNA steady state levels. RAW 264.7 cells were stimulated for 24 h with IFN- (50 units/ml) in the presence of vehicle or dexamethasone at concentrations of 1 nM, 10 nM, 100 nM, and 1 µM. As shown in Fig. 3, strong induction of iNOS mRNA is observed upon stimulation of the cells with IFN- (50 units/ml). Dexamethasone (1 µM) decreases the formation of iNOS mRNA, however, only to about 50% when compared with levels obtained with IFN- alone (Fig. 3). The application of lower concentrations of dexamethasone such as 10 and 1 nM does not block the induction of iNOS mRNA levels in IFN--stimulated RAW 264.7 cells. Thus, suppression of iNOS protein levels is only in part caused by a decrease in iNOS mRNA steady state levels.

Dexamethasone decreases IFN--induced iNOS mRNA steady state levels.

Dexamethasone Inhibits iNOS mRNA Formation When Administered before IFN-, However, Fails to Inhibit When Added after IFN-

We further were interested in determining whether dexamethasone would inhibit the formation of iNOS mRNA levels when the drug is administered at different time points relative to IFN-. RAW 264.7 cells were stimulated with dexamethasone (1 µM) for 1, 8, and 15 h prior to addition of IFN- (50 units/ml). Subsequently incubations were continued for 24 h. Total cellular RNA was isolated and subjected to Northern blot hybridization. As shown in Fig. 4, dexamethasone efficiently blocks the induction of iNOS mRNA only when administered 15 h prior to IFN-. Furthermore, RAW 264.7 cells were pretreated with IFN- (50 units/ml) for 1 and 8 h. Subsequently, dexamethasone (1 µM) was added and incubations were continued for a total of 24 h. As shown in Fig. 4, dexamethasone fails to inhibit the induction of iNOS mRNA steady state levels when administered after IFN-.

Effect of dexamethasone on the induction of iNOS mRNA by IFN-.

We performed mRNA stability experiments to assess whether dexamethasone would affect the half-life of IFN--induced iNOS mRNA. RAW 264.7 cells were stimulated with IFN- (50 units/ml) for 12 h. Then vehicle or dexamethasone (1 µM) was added for a further 2 h. Subsequently actinomycin D (10 µg/ml) was added and iNOS mRNA levels were determined after 1, 4, and 12 h. Incubations of RAW 264.7 cells with actinomycin D (10 µg/ml) longer than 12 h were not performed because of severe disturbances of cell viability. Whereas the half-life of iNOS mRNA in this experimental setting was determined to be greater than 12 h (Fig. 5), dexamethasone (1 µM) decreased iNOS mRNA half-life to approximately 8 h.

Nuclear run-on experiments were performed to evaluate whether changes in iNOS gene transcription may contribute to dexamethasone-dependent reduction of iNOS mRNA formation. As shown in Fig. 6, in unstimulated cells virtually no transcriptional activity is determined. Incubation of the cells for 6 h with IFN- (50 units/ml) increases the transcription rate of the iNOS gene. Dexamethasone decreases the transcriptional activity of the iNOS gene by approximately 65% compared with IFN- alone (Fig. 6). Thus, the suppressive effect of dexamethasone on the formation of iNOS mRNA steady state levels occurs by a decrease in half-life of iNOS mRNA and by the interference of the drug with the transcriptional machinery regulating iNOS gene expression.

The foregoing experiments suggest that dexamethasone may affect additional post-transcriptional steps of iNOS expression such as translation of iNOS mRNA and/or stability of the iNOS protein. Thus, we intended to evaluate the effects of dexamethasone on the translation of iNOS mRNA. RAW 264.7 cells were stimulated with IFN- (50 units/ml) for 3.5 h to achieve increased iNOS mRNA levels (data not shown). Then vehicle, dexamethasone (100 nM), or a combination of dexamethasone (100 nM) plus the glucocorticoid receptor antagonist RU-486 (1 µM) in a 10-fold molar excess were added to the cells and incubation continued for 0, 2, and 4 h. Finally, RAW 264.7 cells were pulsed with L-[³⁵S]methionine for 30 min to label newly synthesized proteins. Cell lysates were used for immunoprecipitation with a rabbit polyclonal anti-iNOS antibody and SDS-PAGE as shown in Fig. 7. Whereas IFN--induced iNOS mRNA translation increases with time, administration of dexamethasone (100 nM) causes a reduction of the translation of iNOS mRNA in a time-dependent fashion. This effect can completely be reversed by incubation of the cells with dexamethasone (100 nM) in the presence of a 10-fold molar excess of the receptor antagonist RU-486 (1 µM; Fig. 7). It is noteworthy that anti-iNOS immunoblots of activated RAW 264.7 cells reveal two bands, migrating on SDS-PAGE, with apparent molecular masses of 130 and 115 kDa. Whether the lower band represents a proteolytic degradation product or a post-translationally modified form of iNOS as recently described (22, 23) remains to be elucidated. For quantification only the upper bands were used.

We next examined whether dexamethasone would affect the degradation of the iNOS protein. RAW 264.7 cells were incubated with IFN- (50 units/ml) for 6 h. Then vehicle or dexamethasone (1 µM) was added, cells were starved for L-methionine for 45 min, then pulsed with L-[³⁵S]methionine for 2 h and finally chased with an excess of nonradioactive L-methionine for 0, 2, 4, 8, and 12 h. Subsequently cells were lysed and iNOS protein was immunoprecipitated using a rabbit polyclonal anti-iNOS antibody and separated by SDS-PAGE (Fig. 8). By the end of the pulse (a total of 3 h 45 min of dexamethasone) a difference in the amount of iNOS protein is detectable between IFN--stimulated and IFN- plus dexamethasone-treated cells (Fig. 8). This is most probably caused by the action of dexamethasone during the pulse period. However, the rate of degradation of radioactively labeled iNOS protein during the chase period is drastically accelerated in dexamethasone-treated cells already at 2 h (Fig. 8). Thus, enhanced proteolysis of the iNOS protein, together with a decrease in translation of iNOS mRNA may account for the post-transcriptional part in the dexamethasone-dependent inhibition of IFN--induced iNOS expression in RAW 264.7 cells.

Dexamethasone increases the degradation of the IFN--induced iNOS protein.

To assess the identity of the protease involved in the dexamethasone-induced degradation of the iNOS protein, we performed pulse-chase experiments in the presence of protease inhibitors such as calpain inhibitor I and TLCK.

TLCK is known to irreversibly inhibit serine proteases like trypsin. RAW 264.7 cells were incubated with IFN- (50 units/ml) for 6 h. Then vehicle or dexamethasone (1 µM) was added and cells were starved for L-methionine for 45 min before being pulsed with L-[³⁵S]methionine for an additional 2 h and finally chased with an excess of nonradioactive L-methionine for 0 and 2 h. During the pulse period, calpain inhibitor I (100 µM) or TLCK (100 µM) were added to the cells. Subsequently cells were lysed, iNOS was immunoprecipitated and separated by SDS-PAGE (Fig. 9 and 10). As seen in the previous experiment in Fig. 8, by the end of the pulse (a total of 3 h 45 min of dexamethasone) a difference in the amount of iNOS protein is detectable between IFN--stimulated and IFN- plus dexamethasone-treated cells (IFN- and IFN- + DEX in Figs. 9 and 10) which may be caused by the action of dexamethasone during the pulse procedure. Dexamethasone drastically increases the degradation of iNOS protein (IFN- + DEX in Figs. 9 and 10). When co-administered with calpain inhibitor I (100 µM), a specific inhibitor of the protease calpain I and II, dexamethasone-induced degradation is inhibited as shown in Fig. 9 (IFN- + DEX + CPI). This effect is specific for dexamethasone-induced iNOS protein degradation and is not seen in the absence of the glucocorticoid (IFN- + CPI in Fig. 9). The graph (Fig. 9, bottom) shows the densitometrical analysis. The increase in iNOS protein degradation is expressed as the ratio of IFN- to IFN- plus drug.

The trypsin inhibitor TLCK does neither affect dexamethasone-induced proteolysis of the iNOS protein nor the degradation which occurs by IFN- alone (IFN- + DEX compared with IFN- + DEX + TLCK in Fig. 10). Similar results have been obtained using the chymotrypsin inhibitor TPCK, an irreversible inhibitor of chymotrypsin (data not shown). Interestingly TLCK treatment indicates the disappearance of the lower band while not having an effect on the dexamethasone-induced protease. This may suggest that the process of the putative post-translational modification of the iNOS protein is sensitive to a protease that is inhibited by TLCK. Alternatively, the lower band is due to proteolytic processing of iNOS protein by a protease that is inhibited by TCLK.

Dexamethasone Does Not Alter Calpain mRNA Steady State Levels and Calpain Protein Levels in IFN--stimulated RAW 264.7 Cells

To determine whether dexamethasone would affect calpain gene expression in IFN--stimulated RAW 264.7 cells, Northern blot analyses were performed to measure calpain mRNA steady state levels. RAW 264.7 cells were stimulated with IFN- (50 units/ml) for 24 h in the presence of vehicle or dexamethasone at concentrations of 100 nM and 1 µM. As shown in Fig. 11, considerable high amounts of calpain mRNA can be detected in control RAW 264.7 cells. Incubation of the cells with IFN- (50 units/ml) only slightly increases calpain mRNA steady state levels and these are not altered upon coincubation with dexamethasone (Fig. 11). Moreover, levels of calpain protein are not altered by dexamethasone as assessed by immunoprecipitation using a polyclonal anti-calpain II antibody. RAW 264.7 cells were incubated with IFN- (50 units/ml), IFN- (50 units/ml) plus dexamethasone (1 µM), or dexamethasone (1 µM) alone for 12 h in pulse medium. The cells were labeled with L-[³⁵S]methionine (300 µCi/dish) in the presence of stimuli for 4.5 h. Subsequently cells were lysed, calpain protein was immunoprecipitated using a polyclonal anti-calpain antibody and separated by SDS-PAGE. Incubation of the cells with IFN- (50 units/ml) slightly increases calpain protein levels and these are not altered by dexamethasone (Fig. 12).

Dexamethasone does not affect IFN--induced calpain mRNA steady state levels.

Dexamethasone does not alter IFN--induced calpain protein levels.

DISCUSSION

In recent years it has become evident that high output production of nitric oxide by iNOS is responsible for the development of a variety of diverse pathological events in mammalian organs. Recent research has focused on the development of selective inhibitors of iNOS catalytic activity. The goal is to inhibit excessive formation of NO without interfering with the production of small quantities of NO generated by endothelial and neuronal isoforms (for review, see Refs. 24 and 25). Understanding the mechanisms involved in the induction and modulation of iNOS gene expression is important for the development of pharmacological strategies aiming to selectively prevent excessive NO formation. Among the most widely used drugs in anti-inflammatory therapies, glucocorticoids are highly effective in controlling inflammation and this may be in part be due to their capability to inhibit iNOS expression.

Glucocorticoid inhibition of NO production was first described in cytokine-stimulated mesangial cells (10, 26). Di Rosa et al. (11) first demonstrated that dexamethasone and hydrocortisone also inhibit the production of NO in the lipopolysaccharide and IFN--stimulated macrophage cell line J774. In the present report we have focused on studying the mechanisms by which the synthetic glucocorticoid dexamethasone modulates iNOS expression in IFN--stimulated RAW 264.7 cells.

We found that glucocorticoid acts at different levels of iNOS expression to inhibit the formation of NO. Our data suggest that a combination of decreased transcriptional activity of the iNOS gene and stability of iNOS mRNA causes the reduction of iNOS mRNA steady state levels observed under the action of dexamethasone. A very interesting task is the identification of the transcription factors which are involved in the regulation of the iNOS gene by IFN- and which may be the target of drug action. Recent publications of several groups have shown that the transcription factor IRF-1 is required for the synergistic induction of the iNOS gene in response to lipopolysaccharide and IFN- (27). The contribution of IFN- to iNOS gene transcription in RAW 264.7 cells requires binding of IRF-1 to IRF-E, its cis-acting element within the murine iNOS promoter (27). Furthermore, in IRF-1 / gene knockout mice, iNOS is not expressed (28). However, using a gene knockout model Meraz et al. (29) demonstrated the importance of STAT-1 for the induction of iNOS in macrophages. The authors show that activation of STAT-1 is required for IFN--dependent IRF-1 and iNOS expression, suggesting that STAT-1 has functionally to be placed upstream of IRF-1 and most likely is the main mediator of IFN--dependent signal transduction. Preliminary data obtained in our laboratory by electrophoretic mobility shift analyses using a radioactively labeled STAT-1 consensus oligonucleotide and nuclear extracts from RAW 264.7 cells strongly suggest that dexamethasone interferes with the DNA binding activity of this transcription factor.² This may cause the reduction of iNOS gene transcription observed in IFN--stimulated RAW 264.7 cells.

Most importantly, dexamethasone has pronounced inhibitory effects on iNOS expression in IFN--stimulated RAW 264.7 cells at the post-transcriptional level, notably on the translation of iNOS mRNA and the degradation of iNOS protein. Several reports demonstrate that post-transcriptional mechanisms are involved in the suppression of cytokine synthesis by dexamethasone. In human monocytes, dexamethasone affects multiple levels of interleukin-1 production (8). It slightly increases the interleukin-1 mRNA half-life, causes a moderate inhibition of translation of the mRNA, and has profound effects on the release of mature interleukin-1. Similarily, endotoxin-induced tumor necrosis factor- expression in monocytes is inhibited by dexamethasone at the level of translation of tumor necrosis factor- mRNA whereas the transcription rate of the tumor necrosis factor- gene is only marginally decreased (9).

The first data suggesting that post-transcriptional mechanisms are involved in the regulation of iNOS expression were provided by Vodovotz et al. (30). The authors demonstrated that transforming growth factor ₁ inhibits IFN--induced iNOS expression in mouse peritoneal macrophages at multiple levels, including translation of the iNOS mRNA and stability of the iNOS protein. Interestingly, dexamethasone seems to activate the same post-transcriptional mechanisms as transforming growth factor ₁ to exert its anti-inflammatory effects. We show that dexamethasone reduces the translation of iNOS mRNA and, in addition, increases the degradation of the iNOS protein. These results correspond well to our recent findings obtained in interleukin-1-stimulated rat renal mesangial cells (18). Thus we conclude that the effects of dexamethasone on the translation of iNOS mRNA and stability of the iNOS protein are independent of the cell type or the stimulus used for induction. Furthermore, dexamethasone inhibits iNOS expression even after iNOS has already been expressed. This may have important implications for the clinical treatment of diseases associated with overproduction of NO and strongly argues for the effectiveness of dexamethasone in the acute anti-inflammatory therapy.

To identify the degrading enzyme we performed pulse-chase experiments in the presence of protease inhibitors. Our results strongly suggest that the cysteine protease calpain I is involved in the increased degradation of the iNOS protein caused by dexamethasone. Calpain constitutes a large family of proteolytic enzymes and is classified into two classes, ubiquitous and tissue specific. By limited proteolysis, calpain alters the activity or function of substrate proteins. Thus calpain is regarded as a biomodulator involved in the regulation of many physiological processes such as cell division, signal transduction, and long-term potentiation. Various molecular targets for calpain-dependent proteolysis have been identified. Among them are cytoskeletal proteins, membrane proteins, enzymes and transcription factors (for review, see Refs. 31 and 32). Glucocorticoid-induced expression of calpain was reported in human lymphoid cells (33). In rat L8 myotube cultures, dexamethasone increases the expression of a number of proteases including calpain. These changes may account for the ability of glucocorticoids to induce increased proteolysis in skeletal muscle (34). Moreover, the involvement of calpain in dexamethasone-induced programmed cell death is discussed (35). Our results obtained by Northern blot analyses (Fig. 11) and immunoprecipitation using a polyclonal anti-calpain antibody (Fig. 12) suggest that the amount of calpain mRNA steady state levels and calpain protein levels are considerably high in control RAW 264.7 cells and are not changed upon IFN- exposure. Moreover, these levels are not altered by treatment of the cells with dexamethasone. Therefore, activation of iNOS proteolysis by calpain most likely does not occur by gluocorticoid-dependent calpain gene induction. We speculate that the calmodulin (CaM)-binding site in the iNOS protein may provide the starting point for the enhanced degradation triggered by dexamethasone. A common structural feature of a number of different substrates of calpain is the presence of a calmodulin binding motif. Molinari et al. (36) demonstrated that in the plasma membrane Ca²⁺-ATPase, an accessible CaM-binding region appears to be critical for substrate recognition and proteolysis by calpain. Occupation of the CaM-binding site by CaM significantly decreases the rate of proteolysis (36). Although the iNOS protein has long been demonstrated to bind CaM (37), the function is not yet entirely understood. Recently, Stevens et al. (38) provided data suggesting that CaM is tightly but reversibly bound to iNOS protein in a fashion different from other known CaM-enzyme interactions and that it may be required for enzymatic activity. Our observation may provide a new aspect to understand the function of CaM binding to the iNOS protein. Furthermore, it is of particular interest whether transforming growth factor ₁ also activates the cysteine protease calpain I to increase degradation of iNOS protein. Such studies are currently being performed in our laboratory.