INTRODUCTION

Nitric oxide (NO),¹ a bioactive free radical, is involved in various physiological and pathological processes in many organ systems (1). At low concentration NO has been shown to play a role in neurotransmission and vasodilation and NO secreted at higher concentrations is implicated in having a role in the pathogenesis of stroke and other neurodegenerative diseases (2). This is of particular importance in conditions associated with infiltrating macrophages and production of proinflammatory cytokines such as demyelinating conditions (e.g. multiple sclerosis, experimental allergic encephalopathy, and X-adrenoleukodystrophy) and in ischemic and traumatic injuries (3-6). NO is enzymatically formed from L-arginine by nitric-oxide synthase (NOS). Basically, the NOS are classified into two groups. One type, constitutively expressed (cNOS) in several cell types (e.g. neurons, endothelial cells) is regulated predominantly at the post-transcriptional level by calmodulin in a calcium-dependent manner (2). In contrast, the inducible form (iNOS), expressed in various cell types including smooth muscle cells, macrophages, keratinocytes, hepatocytes, and brain cells is induced in response to a series of proinflammatory cytokines including interleukin-1 (IL-1), tumor necrosis factor- (TNF-), interferon- (IFN-), and bacterial lipopolysaccharide (LPS) (7-10). The cytokine induced production of iNOS and NO in astrocytes and microglia has been implicated in oligodendrocyte degeneration in demyelinating diseases (4, 6) and neuronal death during trauma (11).

Efforts at understanding the mechanism of signal transduction cascade for induction of iNOS in response to LPS and cytokines are an active area of investigation. Despite a large number of observations describing the induction of NO by cytokines, the molecular events leading to the induction of iNOS in cytokines induced astrocytes are not understood. The differential effect of Rolipram, an inhibitor of type IV phosphodiesterase on the production of TNF- and NO in a murine macrophage cell line suggests a role for cAMP in LPS-induced immune response (12). An increase in the intracellular cAMP concentration by Rolipram increased the production of NO and decreased the production of TNF- in a LPS-stimulated murine macrophage cell line (12). The cAMP-dependent protein kinase (protein kinase A, PKA) is an integral constituent of the protein kinase cascade that links a number of extracellular signals to a variety of cellular functions (13). Some of the effects of PKA are known to antagonize the effects stimulated by growth factors whose receptors are protein tyrosine kinases (e.g. the effect of cAMP on insulin-stimulated glycogen and triacylglyceride formation, and the antagonism of mitogen-activated protein kinase pathway by PKA) (14-16). Inhibition of LPS- and cytokine-mediated NO production by tyrosine kinase inhibitors suggests the involvement of putative protein tyrosine kinase(s) in the signaling events transducing iNOS expression (17, 18). The inhibition of iNOS induction by an inhibitor of NF- (pyrrolidine dithiocarbamate) but not the expression stimulated by 8-bromo-cAMP suggested the participation of more than one pathway in the induction of iNOS (19).

To understand the mechanism of LPS- and cytokine-mediated induction of iNOS and its possible role in neurodegenerative diseases, we examined the intracellular events for the induction of iNOS in astrocytes. We report here that cAMP-dependent protein kinase (PKA) significantly inhibits LPS- and cytokine-mediated activation of NF- and the expression of iNOS in rat primary astrocytes. However, cAMP analogues stimulate the expression of iNOS and activation of NF- in rat peritoneal macrophages. To our knowledge, this is the first example where cAMP regulates the activation of NF- and induction of iNOS differentially in two cell types (astrocytes and macrophages) of the same animal species (rat).

MATERIALS AND METHODS

Recombinant rat IFN-, DMEM/F-12 medium, fetal bovine serum, and NF- DNA-binding protein detection kit were from Life Technologies, Inc. Human IL-1 was from Genzyme. Mouse recombinant TNF- was obtained from Boehringer Mannheim, Germany. LPS (Escherichia coli), NADPH, FAD, tetrahydrobiopterin, Leu-Arg-Arg-Ala-Ser-Leu-Gly (Kemptide), and Dowex 50W were from Sigma. N^G-Methyl-L-arginine, forskolin, 1,9-dideoxyforskolin, 8-Br-cAMP, (S_p)-cAMP, isoproterenol, 3-isobutyl-1-methylxanthine, PKItide, and antibodies against mouse macrophage iNOS were obtained from Calbiochem. L-[2,3,4,5,-H³]Arginine was purchased from Amersham. [-³²P]ATP (3000 Ci/mmol) was from DuPont NEN.

Astrocytes were prepared from rat cerebral tissue as described by McCarthy and DeVellis (20). Cells were maintained in DMEM/F-12 medium containing 10% fetal bovine serum. After 10 days of culture astrocytes were separated from microglia and oligodendrocytes by shaking for 24 h in an orbital shaker at 240 rpm. The shaking was repeated two more times after a gap of 1 or 2 weeks time before subculturing to ensure the complete removal of all the oligodendrocytes and microglia. Cells were trypsinized, subcultured, and used in various experiments as described. Cells were stimulated with LPS or different cytokines in serum-free DMEM/F-12 medium. C₆ glial cells obtained from ATCC was also maintained and induced with different stimuli as described.

Synthesis of NO was determined by assay of culture supernatants for nitrite, a stable reaction product of NO and molecular oxygen. Briefly, 400 µl of culture supernatant was allowed to react with 200 µl of Griess reagent (9, 10) and incubated at room temperature for 15 min. The optical density of the assay samples was measured spectrophotometrically at 570 nm. Fresh culture media served as the blank in all experiments. Nitrite concentrations were calculated from a standard curve derived from the reaction of NaNO₂ in the assay.

NOS activity was measured directly by production of L-[2,3,4,5-³H]citrulline from L-[2,3,4,5-³H]arginine (10). In these experiments, 50 µl of astrocyte homogenate was incubated at 37 °C in the presence of 50 mM Tris-HCl (pH 7.8), 0.5 mM NADPH, 5 µM FAD, 5 µM tetrahydrobiopterin, and 12 µM L-[2,3,4,5-³H]arginine (118 mCi/mmol) in a total volume of 200 µl. The reactions were stopped by the addition of 800 µl of ice-cold 20 mM HEPES (pH 5.5) followed by the addition of 2 ml of Dowex 50W equilibrated in the same buffer. The samples were then centrifuged and the concentration of L-[³H]citrulline was determined in the supernatant by liquid scintillation counting. Protein was measured by the procedure of Bradford (21).

Cell extracts were assayed for PKA activity as described (22) by measuring the phosphorylation of Kemptide (0.17 mM) in the presence or absence of PKI peptide (15 µM). PKA activity was calculated as the amount of Kemptide phosphorylated in the absence of PKI peptide minus that phosphorylated in the presence of PKI peptide.

Following 24 h incubation in the presence or absence of different stimuli, astrocytes were scraped off, washed with Hank's buffer, and homogenized in 50 mM Tris-HCl (pH 7.4) containing protease inhibitors (1 mM phenylmethylsulfonyl fluoride, 5 µg/ml aprotinin, 5 µg/ml pepstatin A, and 5 µg/ml leupeptin). After electrophoresis the proteins were transferred onto a nitrocellulose membrane, and the iNOS band was visualized by immunoblotting with antibodies against mouse macrophage iNOS and ¹²⁵I-labeled protein A. 

Stimulated primary astrocytes were taken out from culture dishes directly by adding Ultraspec-II RNA reagent (Biotecx Laboratories Inc.) and total RNA was isolated according to the manufacturer's protocol. For Northern blot analyses, 20 µg of total RNA was electrophoresed on 1.2% denaturing formaldehyde-agarose gels, electrotransferred to Hybond-Nylon Membrane (Amersham), and hybridized at 68 °C with ³²P-labeled cDNA probe using Express Hyb hybridization solution (Clontech) as described by the manufacturer. The cDNA probe was made by polymerase chain reaction amplification using two primers (forward primer: 5-CTC CTT CAA AGA GGC AAA AAT A-3; reverse primer: 5-CAC TTC CTC CAG GAT GTT GT-3) (23).² After hybridization filters were washed two or three times in solution I (2 × SSC, 0.05% SDS) for 1 h at room temperature followed by solution II (0.1 × SSC, 0.1% SDS) at 50 °C for another hour. The membranes were then dried and exposed with x-ray films (Kodak). The same filters were stripped and rehybridized with probes for glyceraldehyde-3-phosphate dehydrogenase. The relative mRNA content for iNOS was measured after scanning the bands with a Bio-Rad (Model GS-670) imaging densitometer.

Nuclear extracts from stimulated or unstimulated astrocytes (1 × 10⁷ cells) were prepared using the method of Dignam et al. (25) with slight modification. Cells were harvested, washed twice with ice-cold phosphate-buffered saline, and lysed in 400 µl of buffer A (10 mM HEPES, pH 7.9, 10 mM KCl, 2 mM MgCl₂, 0.5 mM dithiothreitol, 1 mM phenylmethylsulfonyl fluoride, 5 µg/ml aprotinin, 5 µg/ml pepstatin A, and 5 µg/ml leupeptin) containing 0.1% Nonidet P-40 for 15 min on ice, vortexed vigorously for 15 s, and centrifuged at 14,000 rpm for 30 s. The pelleted nuclei were resuspended in buffer B (20 mM HEPES, pH 7.9, 25% (v/v) glycerol, 0.42 M NaCl, 1.5 mM MgCl₂, 0.2 mM EDTA, 0.5 mM dithiothreitol, 1 mM phenylmethylsulfonyl fluoride, 5 µg/ml aprotinin, 5 µg/ml pepstatin A, and 5 µg/ml leupeptin). After 30 min on ice, lysates were centrifuged at 14,000 rpm for 10 min. Supernatants containing the nuclear proteins were diluted with 20 µl of modified buffer C (20 mM HEPES, pH 7.9, 20% (v/v) glycerol, 0.05 M KCl, 0.2 mM EDTA, 0.5 mM dithiothreitol, and 0.5 mM phenylmethylsulfonyl fluoride) and stored at 70 °C until use. Nuclear extracts were used for the electrophoretic mobility shift assay using the NF- DNA-binding protein detection system kit (Life Technologies, Inc.) according to the manufacturer's protocol.

RESULTS

Primary astrocytes in serum-free DMEM/F-12 were treated with different activators and inhibitors of PKA 15 min before the addition of 1 µg/ml LPS. Fig. 1 shows that the compounds (forskolin, 8-bromo-cAMP, and (S_p)-cAMP) known to increase intracellular cAMP inhibited the LPS-stimulated NO production as nitrite (Fig. 1A), iNOS activity as conversion of arginine to citrulline (Fig. 1B), expression of iNOS protein (Fig. 1C) and mRNA (Fig. 1D), and activated the PKA activity (Fig. 1E). The inactive forskolin analogue, 1,9-dideoxyforskolin (10 µM), neither inhibited the LPS-induced iNOS activity nor stimulated the PKA activity (Table I). Other PKA activators like -adrenergic receptor agonist, isoproterenol (10 µM), and cAMP phosphodiesterase inhibitor, 3-isobutyl-1-methylxanthine (1 mM), also inhibited LPS-stimulated NO production and iNOS activity (Table I). On the other hand, LPS-stimulated NO production, iNOS activity, and expression of iNOS protein and mRNA were increased by PKA inhibitors (H-89 and (R_p)-cAMP) (Fig. 1). However, in the absence of LPS neither PKA activators nor PKA inhibitors had any effect on the production of NO (data not shown). This inhibition of NO production by cAMP was not only confined to astrocytes, but forskolin was also found to inhibit LPS- and cytokine-induced NO production in rat C₆ glial cells (Fig. 2). The decrease in LPS-induced iNOS expression with the increase in cAMP level and the increase in LPS-induced iNOS expression with the decrease in cAMP level clearly delineate cAMP and cAMP-dependent protein kinase as important regulators of iNOS biosynthesis in glial cells.

Table I. Inhibition of LPS-induced nitrite accumulation in rat primary astrocytes by different cAMP agonists Primary astrocytes were cultured for 24 h in serum-free DMEM/F-12 with the listed reagents; nitrite concentration in the supernatants was then measured as described under "Materials and Methods." Concentration of reagents were: LPS, 1.0 µg/ml; forskolin, 10 µM; 1,9-dideoxyforskolin, 10 µM; isoproterenol, 10 µM; 1-isobutyl-1-methylxanthine (IBMX), 1 mM; Rolipram, 10 µM. All the cAMP agonists were added to the cells 15 min prior to the addition of LPS. Data are mean ± S.D. of three different experiments. Stimuli Nitrite Inhibition nmol/mg/24 h % Control 3.2  ± 0.4 LPS 31.4  ± 3.6 0 LPS + forskolin 4.7  ± 1.2 85 LPS + dideoxyforskolin 31.2  ± 4.1 LPS + isoproterenol 8.1  ± 2.3 74 LPS + IBMX 6.8  ± 1.3 78 LPS + Rolipram 6.2  ± 1.2 80

Astrocytes were incubated with different concentrations of forskolin 15 min before the addition of 1 µg/ml LPS, and after 24 h the iNOS activity was measured as nitrite concentrations in the supernatant and conversion of arginine to citrulline in the cellular homogenates (Fig. 3). The level of nitrite and iNOS activity were inhibited to a similar degree at all the concentrations of forskolin tested. The lowest dose of forskolin found to inhibit iNOS activity and NO production significantly (by 30%) was 0.1 µM. At 10 µM forskolin, NO production and iNOS activity were inhibited by about 90%. Higher doses of forskolin (50-100 µM) did not result in further significant inhibition of iNOS (data not shown). This may be due to the fact that PKA was already completely activated in extracts of cells incubated with 10 µM forskolin. The PKA activity increased with the increase in forskolin concentration. The reciprocal relationship of production of NO and iNOS activity with PKA activity supports the conclusion that PKA may play a pivotal role in the regulation of iNOS expression in astrocytes.

Primary astrocytes were stimulated with TNF-, IL-1, and IFN- alone or in different combinations for 24 h and iNOS was measured. TNF-, IL-1, and IFN- individually were able to induce iNOS activity, protein, and mRNA, however, when tested in combinations between them or with LPS, the magnitude of induction was significantly higher (Figs. 4 and 5). Forskolin, the activator of PKA, completely inhibited the cytokine-induced expression of iNOS, whereas H-89, a specific inhibitor of PKA, stimulated the cytokine-induced expression of iNOS (Fig. 4). Similarly, the induction of iNOS by several combinations of cytokines and LPS were also inhibited by forskolin (Fig. 5) suggesting that augmentation of the cellular levels of cAMP and the activation of cAMP-dependent protein kinase may represent a general counter-regulatory mechanism for down-regulation of iNOS expression in astrocytes.

Since cAMP derivatives inhibited the LPS- and cytokine-induced NO production in rat primary astrocytes, we examined the effect of these derivatives on NO production in rat resident macrophages. In contrast to the inhibition of NO production observed in astrocytes, both forskolin and 8-bromo-cAMP stimulated the LPS-induced NO production in macrophages (data not shown). Previously other investigators have also shown the stimulation of LPS-induced production of NO by forskolin or 8-Br-cAMP in rat macrophages (12, 26), mesangial cells (27), and murine 3T3 fibroblasts (28). On the other hand, H-89, the specific inhibitor of PKA, inhibited LPS-induced NO production in macrophages, suggesting also the involvement of the PKA pathway in LPS-induced NO production in macrophages. These results suggest that cAMP may regulate the LPS-induced expression of iNOS in astrocytes and macrophages by different mechanisms.

To investigate the mechanism of activation of iNOS by PKA possibly by post-translational phosphorylation, the homogenates of astrocytes or macrophages stimulated with 1.0 µg/ml LPS for 24 h in serum-free DMEM/F-12 were used as substrates for phosphorylation by a catalytic subunit of PKA. Phosphorylation of homogenates of either astrocytes or macrophages by different amounts of a catalytic subunit of PKA for different time intervals did not result in any change in iNOS activity as measured by the formation of L-[³H]citrulline from L-[³H]arginine (data not shown). These observations suggest that PKA does not directly activate or inhibit iNOS from either astrocytes or macrophages by post-translational phosphorylation.

Differential Regulation of LPS-induced NF- Activation by cAMP in Rat Primary Astrocytes and Macrophages

Increase in the levels of intracellular cAMP inhibits the induction of NO production in astrocytes but stimulates the production of NO in macrophages. Since the activation of NF- is necessary for the induction of iNOS (29, 30), we examined the effect of cAMP on LPS-induced activation of NF- in astrocytes and macrophages to understand the basis of this differential regulation of NO production by cAMP. Both astrocytes and macrophages were treated with forskolin or H-89, alone or with LPS, and the nuclear proteins were extracted. NF- activation was evaluated by the formation of a distinct and specific complex in a gel-shift DNA-binding assay. Treatment of astrocytes or macrophages with 1.0 µg/ml LPS resulted in the activation of NF- (Fig. 6, A and B). This gel shift assay detected a specific band in response to LPS that was competed off by an unlabeled probe. Forskolin or H-89 alone at different concentrations failed to induce NF- in astrocytes. However, in astrocytes, forskolin markedly inhibited the LPS-induced activation of NF-, whereas H-89 stimulated this activation (Fig. 6A). On the other hand, in macrophages, consistent with previous observations (31, 32), forskolin alone induced the DNA binding activity of NF- and stimulated LPS-induced activation of NF-, and H-89, the specific inhibitor of PKA, inhibited LPS-induced activation of NF-. These results demonstrate differential regulation of NF- activation by cAMP in astrocytes and macrophages.

Effect of cAMP on LPS-induced NF- activation in astrocytes and macrophages.

DISCUSSION

Recent studies have provided evidence that cAMP induces the expression of iNOS in LPS- and cytokine-stimulated glomerular mesangial cells (33), smooth muscle cells (34), cardiac myocytes (35), murine 3T3 fibroblasts (28), and peritoneal macrophages (26). The increase in cAMP as a result of inhibition of phosphodiesterase (an enzyme that degrades cAMP), or an exogenous supply of cAMP derivatives or compounds that enhance intracellular levels of cAMP cause induction of iNOS. Contrary to these results we have observed that intracellular levels of cAMP negatively regulate the expression of iNOS in LPS or cytokine-stimulated primary astrocytes from rat brain. This conclusion is based on the following observations. 1) LPS induced the expression of iNOS mRNA, protein, and activity as compared with no effect on the PKA activity in astrocytes. Treatment of LPS-stimulated astrocyte cultures with forskolin or 8-Br-cAMP or (S_p)-cAMP results in an increase in PKA activity and the down-regulation of the expression of iNOS mRNA, protein, and activity. 2) Inhibition of the PKA activity with H-89 or with (R_p)-cAMP results in an increase in the expression of iNOS mRNA, protein, and activity in LPS-stimulated astrocytes. The reciprocal relationship between activation of PKA by forskolin and inhibition of NO production and iNOS activity also supports the conclusion that cAMP levels negatively regulate the expression of iNOS in astrocytes. Similar results were observed with rat C₆-glial cells. In contrast, in agreement with previous observations (12, 26) the treatment of LPS-stimulated macrophages with forskolin, 8-Br-cAMP, and (S_p)-cAMP augmented the NO production. Moreover, addition of PKA inhibitors (H-89 or (R_p)-cAMP) to LPS-stimulated macrophages resulted in inhibition of NO production.

A comparative study of induction of iNOS by different pathways in 3T3 fibroblast cells reported that stimulation of PKC by tetradecanoylphorbol-13-acetate or activation of PKA by cAMP elevating agents or cytokine/receptor tyrosine kinase (e.g. transforming growth factor-1) converge in the activation of NF- (28). The presence of the consensus sequence for the binding of NF- (36, 37) and interferon regulatory factor (29, 38) in the iNOS gene have been shown to be functionally important for the induction of iNOS. Although, the signal transduction pathways effective in iNOS induction converge at NF- activation, they differ significantly in different cell types. For example, in murine RAW 264.7 cells, neither the PKA nor the PKC pathways are able to activate NF- (39). In cardiac myocytes, iNOS is activated by the PKA pathway (35), whereas, in human Jurkat T-cells the PKC pathway but not the PKA pathway activates NF- (40). Studies reported in this article clearly establish that the induction of iNOS in both macrophages and astrocytes converge at the activation of NF-. However, interestingly, the intracellular events for the activation of NF- in astrocytes and macrophages are quite different. In astrocytes, activators of PKA inhibit LPS-induced NF- activation as well as induction of iNOS and inhibitors of PKA enhance LPS-induced activation of NF- and induction of iNOS. This is contrary to the observations with macrophages (Fig. 6) (31, 32). In macrophages, cAMP enhances the activation of NF- and the induction of iNOS.

The signaling events transduced by LPS and cytokines for the induction of iNOS are not completely established so far. LPS is shown to bind cell-surface receptor CD14 (41) and induce iNOS probably via activation of NF- (29, 30). The inhibition of LPS-mediated NF- activation by cAMP in astrocytes and stimulation of NF- activation by cAMP in macrophages clearly suggest that the differential effect of cAMP on NO production in astrocytes and macrophages is due to the differential effect on NF- activation. The basis for this differential regulation of NF- by cAMP in astrocytes and macrophages is not understood at the present time. One potential intracellular target of LPS signaling in cells is the activation of the MAP kinase pathway. It has been shown by several groups that augmentation of intracellular cAMP blocks the signaling pathway from Ras to MAP kinase in cells such as fibroblasts and fat cells by phosphorylation of Raf (an upstream member of MAP kinase pathway) (15, 16, 22). Phosphorylation of Raf decreases its affinity for and prevents its interaction with Ras, another upstream proximal member of the MAP kinase pathway (15, 16). This lack of interaction effectively prevents translocation of Raf-1 to the membrane by Ras and subsequent activation of Raf by other as yet unidentified molecules (15, 16). If a similar mechanism also exists in astrocytes, the inhibition of LPS-induced activation of NF- and expression of iNOS by augmentation of intracellular cAMP in rat primary astrocytes may involve the inhibition of a LPS-mediated signal transduction event, i.e. activation of Raf. In addition, it has been reported that Ik can be phosphorylated by Raf, resulting in release of a p50/p65 heterodimer (42). Thus inhibition of Raf may lead to a decrease in the release of p50/p65 heterodimer and its translocation into the nucleus. However, in macrophages, cAMP alone activates NF- and also stimulates LPS-mediated activation of NF- suggesting that the regulation of NF- by cAMP in macrophages is entirely different from that in astrocytes. Increase in intracellular cAMP may inhibit the activation of Raf in macrophages, however, the activation of NF- may not involve Raf signaling. In addition to inhibiting the MAP kinase signaling pathway, cAMP may also modulate the activation of NF- as well as the transcription of iNOS via a cAMP-dependent protein kinase-mediated phosphorylation of the cAMP-response element-binding protein (13). Therefore, it will be interesting to study the role of cAMP in the signaling pathways from Ras to MAP kinase in astrocytes or macrophages to explain the observed differential regulation of NF- and iNOS and to explore the role of cAMP-response element-binding protein, if any, in LPS- and cytokine-mediated activation of NF- and expression of iNOS.

NO is a diffusible free radical that plays many roles in diverse pathological conditions. Once iNOS is induced in vitro, NO release can continue for hours or days (43-45) and its presence can be detrimental to neurons and oligodendrocytes (4, 46). NO and peroxynitrite (reaction product of NO and O₂) are potentially toxic molecules that may mediate toxicity through the formation of iron-NO complexes of iron containing enzyme systems (47), oxidation of protein sulfhydryl groups (48), nitration of proteins and nitrosylation of nucleic acid, and DNA strand break (49). Although monocytes/macrophages are the primary source of iNOS in inflammation, LPS and other cytokines also induce a similar response in astrocytes (50, 51). Astrocytes are the major cell population in the central nervous system, therefore, induction of iNOS in astrocytes may be an important source of NO in central nervous system inflammatory disorders associated with neuronal and oligodendrocyte death (4, 24). The studies reported in this article indicate that use of cAMP analogues or cAMP phosphodiesterase inhibitors may represent a possible avenue of research for therapeutics directed against the nitric oxide-mediated brain disorders, particularly in demyelinating conditions.