Inhibitory and Stimulatory Effects of Lactacystin on Expression of Nitric Oxide Synthase Type 2 in Brain Glial Cells

Expression of inflammatory nitric oxide synthase (NOS2) is mediated by transcription factor NFκB. By using the specific proteasome inhibitor lactacystin to examine IκB degradation, we observed a paradoxical increase in lipopolysaccharide- and cytokine-dependent NOS2 expression at low concentrations or when lactacystin was added subsequent to cytokines. Lactacystin reduced the initial accumulation of NOS2 mRNA but reduced its subsequent decrease. Lactacystin increased NOS2 promoter activation after 24 h, but not after 4 h, and similarly prevented initial NFκB activation and at later times caused NFκB reactivation. Lactacystin reduced initial degradation of IκB-α and IκB-β, however, at later times selectively increased IκB-β, which was predominantly non-phosphorylated. Expression of full-length rat IκB-β, but not a carboxyl-terminal truncated form, inhibited NOS2 induction and potentiation by lactacystin. Lactacystin increased IκB-β expression in the absence of NOS2 inducers, as well as expression of heat shock protein 70, and the heat shock response due to hyperthermia increased IκB-β expression. These results suggest that IκB-β contributes to persistent NFκB activation and NOS2 expression in glial cells, that IκB-β is a stress protein inducible by hyperthermia or proteasome inhibitors, and that delayed addition of proteasome inhibitors can have stimulatory rather than inhibitory actions.

In brain, inflammatory responses of astroglial cells occur during disease, infection, trauma, and ischemia. These responses include release of pro-inflammatory cytokines as well as synthesis and release of nitric oxide (NO). 1 In astrocytes, NO is primarily biosynthesized by the calcium-independent isoform of NO synthase (NOS2) which is normally not present but whose expression is induced by a variety of inflammatory stimuli. Increasing evidence points to a role for NOS2-derived NO in the pathogenesis of a variety of human neurological diseases and brain trauma (1). In human brain, astrocytes express NOS2 during demyelinating diseases (2)(3)(4), cerebral ischemia (5,6), viral infection (7), traumatic brain injury (8), Alzheimer's disease (9), and possibly during AID's dementia (10). Beneficial aspects of preventing or reducing NOS2 expression have been demonstrated in animal studies of experimental autoimmune encephalomyelitis (EAE) (11) and cerebral ischemia (12) demonstrating that suppression of astroglial NOS2 can be of therapeutic value in the prevention of neurological damage.
Primary cultures of astrocytes from rat (13,14), mouse (15), and human (16) have been used extensively to characterize the inductive process of astroglial NOS2 expression. In vitro studies have demonstrated that stimulation of cells with bacterial endotoxin lipopolysaccharide (LPS) or with a combination of cytokines which nominally includes IL1-␤ leads to de novo expression of NOS2. Extensive characterization of glial NOS2 expression and regulation has also been carried out with the rat C6 glioma cell line (17,18), which shares many properties with primary astrocyte cultures, including expression and regulation of the same NOS2 gene, mRNA, and protein (17).
As found for many other cell types, the induction of astroglial NOS2 requires activation of transcription factor NFB (19), as determined by use of pharmacological inhibitors of NFB and by analysis of the NOS2 promoter region (20 -26). The activation pathway of NFB has been well characterized (19,(27)(28)(29)(30). NFB is a dimeric complex consisting of two members of the Rel protein family including p50, p52, Rel A (p65), Rel B, and the oncogene c-rel. NFB is maintained in the cytoplasm by association with members of the IB protein family (36) whose ankyrin repeat domains bind to, and mask, nuclear localization sites present in NFB subunits. Inflammatory stimulation by cytokines or lipopolysaccharides (LPS) results in activation of IK kinases, leading to serine phosphorylation of IBs at their amino termini (28). Phosphorylated IBs are ubiquinated at amino-terminal lysine residues (31) which targets then for degradation by the 26 S proteasome (28,30). This allows the rapid translocation of NFB to the nucleus, where it binds to B DNA motifs present in a variety of promoter regions including that of NOS2 (26) as well as of IBϪ␣ (32). The de novo transcription of IBϪ␣ results in a rapid increase of IBϪ␣ protein expression, which can reassociate with active NFB (both in the cytoplasm as well as in the nucleus) thus reducing ongoing NFB activity. The expression patterns and phosphorylation kinetics of IB proteins are therefore key to the overall level of NFB activation. However, the characterization of IB proteins in astrocytes is limited (33)(34)(35).
Recently, the cloning of the second major member of the IB family, IBϪ␤, has allowed molecular characterization of the interaction of this isoform with NFB (29). In contrast to IBϪ␣, the IBϪ␤ gene is not induced upon NFB activation, presumably due to lack of B-binding sites in its promoter region (29). Instead, alternative signals, not yet well characterized, lead to increased IBϪ␤ transcription and de novo expression of hypophosphorylated IBϪ␤. Non-phosphorylated IBϪ␤ can bind to NFB but does not mask its nuclear localization site nor transcription activation domain. NFB⅐IBϪ␤ complexes can enter the nucleus and exhibit transcriptional activity, and it has been proposed that this mechanism is responsible for persistent NFB activation (37)(38)(39)(40). Phosphorylation of IBϪ␤ at its carboxyl-terminal PEST domain (41) converts IBϪ␤ to an inhibitory molecule similar in properties to IBϪ␣. Thus, the phosphorylation state of IBϪ␤ dictates its role as an inhibitory versus a stimulatory molecule. It is not yet clear what maintains the hypophosphorylated state of newly made IBϪ␤ during persistent activation, but it has been suggested that LPS and IL-1␤ induce an IBϪ␤-specific phosphatase (38).
We have previously used a nonspecific serine protease inhibitor (TPCK) of the 26 S proteasome, as well as the more selective peptide aldehyde inhibitor benzyloxycarbonyl-Ile-Glu (O-tϪbutyl)ϪAla-leucinal (ZIE), to investigate NFB activation in astrocytes and C6 cells (22). ZIE inhibited NOS2 induction only when present during the initial times of incubation with NOS2 inducers, whereas TPCK inhibited NOS2 expression regardless of when it was added to cells, pointing to effects of TPCK on events other than initial NFB activation. To further examine the role of the NFB⅐IB system in induction of glial NOS2 expression, we have now used lactacystin (43,44), a recently identified microbial product that is a specific inhibitor of the 26 S proteasome. In contrast to peptide aldehyde inhibitors such as ZIE, lactacystin shows no inhibition of any other proteases tested (including thrombin, cathepsin B, calpains I and II, trypsin, chymotrypsin, and papain) (44). In the present study, we demonstrate that NOS2 expression in brain glial cells is blocked by lactacystin, validating previous reports that proteasome activity is requisite for NOS2 induction. Surprisingly, we observed that low doses of lactacystin, or addition of high doses of lactacystin at times after the initial stimulation with NOS2 inducers, caused potentiation rather than inhibition of NOS2 expression. The increased NOS2 expression was accompanied by an enhanced re-expression of IBϪ␤, but not IBϪ␣. Since lactacystin induces a heat shock response (HSR) in these cells (Ref. 45, and Fig. 8 below), we tested the possibility that IBϪ␤, as previously shown for IBϪ␣ (46,47,60), is a stress protein. Our results suggest that sustained NFB activation, mediated by increases in IB-␤ expression, is important for glial cell NOS2 expression, which may be potentiated by proteasome inhibitors due to induction of a HSR.
Cells-C6 glioma and human U172 astrocytoma cells were grown in DMEM containing 10% FCS and antibiotics (penicillin and streptomycin). Cells were passaged once a week into 6-or 96-well plates and used after 3-4 days at which point they were 90 -95% confluent. Primary astrocytes were from cerebral cortices of postnatal day 1 Harlan Sprague-Dawley rats prepared in 96-well plates as described previously (13). Media were changed every 3 days. After 2 weeks growth in complete media (DMEM with 10% FCS) the cultures consisted of 95-98% astrocytes and 2-3% microglia.
NOS2 Induction and Activity Assay-NOS2 expression was induced in C6 cells by overnight incubation in DMEM containing 1% FCS and NOS2 inducers as indicated (final concentrations of inducers were as follows: LPS, 1 g/ml; IL-1␤, 2 ng/ml; IFN 20 units/ml). NOS2 was induced in primary astrocyte cultures in DMEM containing 1% FCS and LPS. After 24 h, NOS2 induction was assessed indirectly by the accumulation of nitrite in the culture media. An aliquot of media (100 l) was mixed with 50 l of Griess reagent (48), incubated at room temperature for 10 min, and then the absorbance at 546 nm determined. Solutions of NaNO 2 served as standards. Background nitrite accumulation in unstimulated cells (cells incubated overnight in the absence of cytokines or LPS) was not significantly different from nitrite accumulation measured in media alone.
Preparation of Stable Transfected C6 Cell Lines-A 2,168-bp fragment of the rat NOS2 promoter was amplified from Harlan Sprague-Dawley liver genomic DNA by PCR using primers derived from the published rat NOS2 promoter sequence (49) (forward, 5Ј-CAG CCA AGT ATT CCA AAG CAA-3Ј, corresponding to bases 1108 -1127; reverse, 5ЈϪAGT CCA GTC CCC TCA CCA A-3Ј, corresponding to bases 3259 -3277), and its identity was confirmed by DNA sequence analysis. Briefly, the 2.2-kb fragment was ligated into the pGL3 basic luciferase vector (Promega, Madison, WI) and co-transfected into C6 cells with plasmid pCMV-Neo (containing a neomycin resistance gene) using Lipofectin reagent (Life Technologies, Inc.). After 2 days, stable transfectants were selected by growth in 0.4 mg/ml G418 (Sigma). The resulting cell line, C6-2.2, has a low level of basal luciferase activity, which can be induced between 4-and 10-fold upon incubation with LPS plus IFN. 2 Luciferase Assays-Cells were incubated with the indicated NOS2 inducers in DMEM containing 1% FCS and the indicated concentrations of lactacystin. After varying the incubation times, the media were removed, and the cells were washed once with cold phosphate-buffered saline. To prepare lysates, 50 l of CHAPS buffer (10 mM CHAPS, 10 mM Tris, pH 7.4) were added to each well, the plate frozen at Ϫ80°C, thawed, and shaken on a rotary shaker for 10 -15 min at room temperature. Aliquots of cell lysates (10 -20 l) containing equal amounts of protein (10 -20 g) were placed into wells of an opaque white 96-well microplate. An equal volume of luciferase substrate (Steady Glo reagent, Promega) was added to all samples, and the luminescence was measured in a microplate luminometer (Rosys-Anthos, Austria). The data are presented as the percentage of luciferase activity measured in the presence of NOS2 inducers and lactacystin, relative to the activity of control cells (incubated in media alone).
Preparation of Nuclear Extracts-Nuclear extracts were prepared from C6 cells as described previously (33). Following stimulation with NOS2 inducers, cells were washed in cold phosphate-buffered saline, pelleted, and resuspended in hypotonic buffer (10 mM HEPES, pH 7.6, 1 mM EDTA, 10 mM KCl, 1 mM DTT) and protease inhibitors (10 g/ml aprotinin and leupeptin, 100 M TPCK, 1 mM PMSF). After 15 min on ice, Nonidet P-40 was added to a final concentration of 0.6%, and the lysates were incubated a further 5 min and then centrifuged for 15 min at 12,000 ϫ g to pellet the nuclei. The nuclei were washed once in the same buffer by gentle resuspension and centrifugation. Nuclear extracts were prepared by extraction of the nuclear pellet for 15 min at 4°C in 50 -100 l of 400 mM NaCl, 10 mM HEPES, pH 7.9, 1 mM EDTA, 1 mM DTT, and 1 mM PMSF. Extracts were used immediately or frozen at Ϫ80°C until use.
Electrophoretic Mobility Shift Assay (EMSA)-Double-stranded B consensus binding sites were obtained by annealing complementary B site oligonucleotides (sequence 5Ј AGT TGA GGG GAC TTT CCC AGG C-3Ј). The resulting double-stranded B DNA was radiolabeled with 32 P using T4 kinase, and unincorporated nucleotide was removed by ethanol precipitation. Aliquots of nuclear extracts (1-2 g) were incubated in binding buffer (50 mM NaCl, 200 ng of poly(dI⅐dC), 1 mM DTT, 5% glycerol, 10 mM HEPES, pH 7.9, 1 mM PMSF, 50 g/ml aprotinin) containing 20 -50,000 cpm of B probe. After 20 min incubation at room temperature, the samples were applied to a 5% non-denaturing acrylamide gel that has been pre-run for 1 h at 100 V at 4°C. The running buffer and gel buffer were both 0.5ϫ TBE (Tris/borate/EDTA). Samples were electrophoresed for 2.5 h at 200 V in a cold box, after which the gels were dried and exposed to x-ray film at room temperature for up to 2 days. For competition assays, the binding reaction was carried out in the presence of a 100-fold excess of non-radioactive B probe or mutated B probe. For supershift EMSA, the nuclear extracts were preincubated for 10 min at room temperature with 1 g of antibodies to NFB subunits p50 or p65.
Heat Shock Procedure-C6 cells were incubated for indicated times (0 -60 min) in a cell culture incubator equilibrated to 43 Ϯ 0.5°C, 95% humidity, and 5% CO 2 . Following incubation, the cells were returned to a 37°C incubator and allowed to recover for 4 h before preparation of cell lysates for Western blot analysis.
Preparation of Protein Lysates and Western Blot Analysis-After incubation with NOS2 inducers and proteasome inhibitors, cells were washed twice with ice-cold phosphate-buffered saline, scraped from the dishes, and collected by centrifugation (3,000 ϫ g for 5 min). The cells were resuspended directly into 5 volumes of 8 M urea. Aliquots were either frozen at Ϫ80°C or immediately mixed with an equal volume of a 2ϫ gel sample buffer (2ϫ buffer: 124 mM Tris-Cl, pH 6.8, 0.2% SDS, 10% ␤Ϫmercaptoethanol, 10 mM EDTA, 50% glycerol) and boiled for 5 min. Protein samples (approximately 20 g) were separated through SDS-10% polyacrylamide gels. After electrophoresis, proteins were transferred to polyvinylidene difluoride membranes by semi-dry electrophoretic transfer. The membranes were blocked in Tris-buffered saline containing 0.1% Tween 20 (TBST), and 5% dry milk (1 h), rinsed, and incubated with primary antibodies in TBST containing 0.5% bovine serum albumin overnight with gentle shaking at 4°C. The primary antibody was removed, membranes washed 4 times in TBST, and 0.1 g/ml peroxidase-labeled goat secondary antibodies added for 1 h. Following 4 washes in TBST, bands were visualized by incubation in enhanced chemiluminescence reagents and exposure to x-ray film. Quantitative assessment of band intensities was obtained by imaging of autoradiographs with an Alpha Innotech (Temucula, CA) Imaging 2000 system. Band intensities were determined from autoradiographs obtained from at least two different exposure times, and background intensities (determined from an equal-sized area of the film immediately below the band of interest) were subtracted.
For analysis of the phosphorylation state of IB-␤, an aliquot of whole cell lysate was diluted 5-fold to reduce urea concentration, calf intestine alkaline phosphatase (CIP) buffer added, and samples incubated for 30 min at 37°C with 10 units of CIP. Samples were separated by SDS-10% polyacrylamide electrophoresis gels that also contained 4 M urea, which we found increases band resolution. Gels were run at 50 V for approximately 18 h with 4 changes of the running buffer.
RT-PCR Analysis of NOS2 mRNA-Total cytoplasmic RNA was prepared from cells by the Nonidet P-40 lysis procedure, and levels of NOS2 mRNA were estimated by competitive RT-PCR assay (50). The primers used for NOS2 detection were 1704F (5Ј CTG CAT GGA ACA GTA TAA GGC AAA C-3Ј), corresponding to bases 1704 -1728; and 1933R (5Ј CAG ACA GTT TCT GGT CGA TGT CAT GA-3Ј), complementary to bases 1908 -1933 of rat iNOS cDNA sequence (50) which yield a 230-bp product. PCR was done in the presence of 2 fg of a 180-bp NOS2 CIS (50) which is amplified with the same efficiency as the cDNA template. The mRNA levels of glyceraldehyde-3-phosphate dehydrogenase were determined in parallel aliquots of cDNA, with primers 796F (5Ј GCC AAG TAT GAT GAC ATC AAG AAG) and 1059R (5Ј TCC AGG GGT TTC TTA CTC CTT GGA) which yield a 264-bp product. PCR was initiated by a hot start method, and conditions were 35 cycles of denaturation at 93°C for 35 s; annealing at 63°C for 45 s; and extension at 72°C for 45 s; followed by 10 min at 72°C done in a Hybaid Thermoreactor (Franklin, MA) controlled by tube temperature. PCR products were separated by electrophoresis through 2% agarose gels containing 0.1 g/ml ethidium bromide. Band intensities were determined using the Alpha Infotech 2000 software.
Cloning of Rat IB-␤ cDNA and Plasmid Constructions-A search of the rat dbEST data base identified EST AI176319 (566 bp) as 92% identical to the corresponding region of the mouse IB-␤ cDNA. This clone is part of rat Unigene cluster Rn.8395 that contains 11 overlapping rat ESTs. The longest contig spans the entire coding region of rat IB-␤ and is contained in rat EST AI574838. The insert in this EST contains a 5Ј-untranslated regions of 68 bp, a coding region of 1077 bp, a 3Ј-untranslated region of 54 bp, and a short poly(A) ϩ tract. The DNA sequence of rat IB-␤ (GenBank TM accession number AF246634) is 92% identical to the mouse IB-␤, and the predicted protein of 359 residues is 92% identical to the mouse IB-␤ protein. The coding region was subcloned into the pcDNA3.1 expression vector (Invitrogen, Carlsbad, CA) to give p3.1-IB␤, and the expressed protein was detected with an antibody against mouse IBϪ␤ (Santa Cruz SC-20) when transiently expressed in Chinese hamster ovary cells (not shown), confirming its identity as rat IB-␤ cDNA. The carboxyl-terminal 54 residues of the rat IB-␤ , containing a PEST sequence and casein kinase II phosphorylation site, was deleted by PCR methods and subcloned back into pcDNA3.1 to yield p3.1-⌬IB-␤.
Expression of IB-␤ in C6 -2.2 Cells-Cells were transfected with the either p3.1-IB-␤ or p3.1-⌬IB-␤ using the Lipofectin 2000 reagent (Life Technologies, Inc.). Briefly, cells were washed 3 times with HBSS, and serum-free medium without antibiotic was used for transfections. The Lipofectin-DNA complex (a mixture of 10 g of plasmid DNA and 25 g of Lipofectin) was prepared according to the manufacturer's recommendations, and varying amounts were added to C6 -2.2 cells at approximately 60% confluency in 96-well plates. After 6 h at 37°C, the cells were washed 3 times with HBSS, and then fresh DMEM containing 10% FBS was added. After 24 h expression, the media were removed, and the cells were incubated with LPS and IFN-␥ as described above to induce NOS2 promoter expression, which was assessed by measurement of luciferase activity after 8 or 24 h.
Data Analysis-Experiments measuring nitrite production and promoter activation were performed at least three times on different cell preparations, each with three or more samples. PCR analysis, Western blots, and EMSAs were repeated at least twice with different preparations of RNA or extracts. Data are presented as mean Ϯ S.E. and are analyzed by one-way or two-way ANOVA. Statistical significance between groups was determined by Dunnett's multiple comparison analysis, and differences were considered significant at p values Ͻ0.05.

RESULTS
Biphasic Effects of Proteasome Inhibitors on Nitrite Production-As previously reported (13), incubation of primary cultures of rat astrocytes with LPS induced NOS2 expression as indicated by the accumulation of nitrites, a stable metabolite of NO, in the cell culture media (Fig. 1A). Simultaneous incubation with the selective proteasome inhibitor ZIE (23) reduced the LPS-dependent nitrite production when present at concentrations of 30 nM and above, with complete inhibition observed at concentrations of 200 nM and above (not shown). This is similar to the reported inhibition of NOS2 expression by ZIE in C6 cells (22) and in smooth muscle cells (23) where ZIE was used at micromolar concentrations. We reproducibly noticed that at lower concentrations of ZIE, there was no inhibition observed, and in some cases there was an increase in the nitrites produced. Although the increase in astrocyte nitrite production did not reach statistical significance, there was a tendency toward increases at ZIE concentrations near 7.5 nM. Since ZIE can inhibit the activity of other proteases (43,44), we tested the highly specific proteasome inhibitor lactacystin that does not cross-react with any other protease tested (44). As the case with ZIE, inhibition of astroglial nitrite production by lactacystin followed a biphasic pattern. Decreased nitrite levels were observed at concentrations of 1.65 M, with complete inhibition at concentrations of 13.5 M and above. At lower doses of lactacystin, there was again a tendency to increase in nitrite levels, with a statistically significant increase (220 Ϯ 25% of control values, n ϭ 4) observed at roughly 800 nM (p Ͻ 0.05 versus control cells). These results confirm that NOS2-dependent nitrite production in astrocytes requires proteasome activation, as previously concluded using less specific proteasome inhibitors, and suggest that low doses of proteasome inhibitors potentiate, rather than inhibit, NOS2 expression.
Incubation of rat C6 glioma cells with LPS plus IFN results in expression of the same NOS2 as astrocytes (17,18) (Fig. 1B). As observed in primary astrocytes, both ZIE and lactacystin inhibited C6 cell NOS2 expression in a dose-dependent manner. C6 nitrite production was significantly reduced versus control cells at ZIE concentrations greater than 40 nM and at lactacystin concentrations greater than 3.4 M. At lower doses of either of these inhibitors there was an increase in nitrite production compared with control cells. The maximum increase in nitrite production over control levels was seen at 7.5 nM ZIE (40 Ϯ 4% increase) and between 100 and 800 nM lactacystin (approximately 60 -70% increase versus control cells). These results confirm that low doses of proteasome inhibitors cause a paradoxical increase rather than decrease in NOS2 expression. To confirm that these effects were not species-specific, we tested the effects of lactacystin on NOS2 expression in human astrocytoma U172 cells (Fig. 1C). In these cells, NOS2 is maximally induced by the cytokine combination of IFN and IL-1␤ (64). As found for both primary astrocytes and C6 cells, cytokinedependent nitrite accumulation was potentiated in these cells at the lower doses (80 -300 nM) lactacystin, and inhibited at higher doses.
To determine if lactacystin needed to be added simultaneously with NOS2 inducers, we incubated C6 cells with inhibitory amounts of lactacystin either at the same time or up to 6 h after addition of the NOS2 inducers (Fig. 2). Delaying lactacystin addition until 3 h after LPS and IFN greatly decreased its inhibitory capacity, while delaying until 6 h after NOS2 inducers reversed its effects from inhibitory to stimulatory. These results suggest that the inhibitory action of lactacystin is mediated at an early event during NOS2 induction, whereas the potentiating effect involves later events.
Effect of Lactacystin on NOS2 Expression-To determine if changes in nitrite production were accompanied by changes in NOS2 expression, we measured NOS2 mRNA levels following incubation with lactacystin (Fig. 3). As previously shown (17) appearance of NOS2 mRNA. After 24 h of incubation in LPS and IFN, NOS2 mRNA levels were reduced versus levels observed at 4 h. Estimation of mRNA levels by comparison to an internal competitive standard revealed a 4 -5-fold reduction in NOS2 mRNA levels occurring between 4 and 24 h. In the presence of 1 M lactacystin, NOS2 mRNA levels at 4 h were much lower than control levels; however, those levels were only slightly reduced (10 -20%) after 24 h of incubation. Incubation with 10 M lactacystin greatly reduced the LPS/cytokine-dependent NOS2 mRNA accumulation at 4 h, and this was further reduced after 24 h. In the same samples levels of glyceraldehyde-3-phosphate dehydrogenase mRNA were not reduced by lactacystin after 4 h and only slightly (less than 10%) reduced after 24 h. These results demonstrate that lactacystin inhibits initial NOS2 mRNA accumulation and that low, but not high, doses prevent the subsequent decrease in mRNA levels normally seen. However, whether this is due to increased NOS2 mRNA stability and/or increased NOS2 gene transcription is not yet known (but see below, Fig. 4).
Since changes in NOS2 mRNA levels could be due to alteration of NOS2 promoter activity, we tested the effects of lacta-cystin on activation of a 2.2-kb fragment of the rat NOS2 promoter attached to the luciferase reporter gene and stably transfected into C6 cells (C6 -2.2 cells, Fig. 4). In these cells, the basal level of luciferase activity was increased 4-fold after 6 h with LPS plus IFN. The presence of lactacystin during this 6-h period caused a dose-dependent decrease in luciferase activity, indicating that lactacystin decreases initial LPS plus IFN-dependent NOS2 promoter activation, with similar potency to the inhibition of nitrite production seen in astrocytes or C6 cells. After 20 h incubation with LPS plus IFN, the NOS2 promoter was activated roughly 3-fold compared with basal activity (i.e. 20 h in the presence of media alone). In contrast to results obtained after 6 h, co-incubation with lactacystin during the 20-h incubation did not inhibit promoter activation except at the highest concentration tested (27 M). At lower concentrations, there was a tendency to increase promoter activation, with the increases at 3.4 and 6.8 M being statistically different from control values. These results suggest that increased nitrite production at low doses of lactacystin are due to a delayed increase in NOS2 promoter activation occurring at least 6 h after initial addition of the NOS2 inducers.
Biphasic Effects of Lactacystin on NFB Activation-We have previously shown that in glial cells, as in other cell types, transcription from the NOS2 promoter is mediated by activation of transcription factor NFB (33,51). We used EMSAs to determine if changes in NFB activation correlated with the observed dual effects of lactacystin (Fig. 5). A 5-h incubation with LPS and IFN led to appearance of two major DNA-protein complexes (labeled III and IV), and the levels of both were decreased, but still present, after 24 h of incubation (Fig. 5A).
In the presence of 1 M lactacystin, the levels of both these complexes were decreased at 5 h compared with control cells. However, after 24 h of incubation the level of complex III was greatly increased, whereas complex IV was further decreased. At a higher concentration of lactacystin (10 M), formation of complex III was completely abrogated after 5 h of incubation, whereas levels of complex IV were reduced but still present. After 24 h of incubation, complex III was once again detectable. The composition of NFB present in complexes III and IV were determined by supershift EMSAs (Fig. 5B). Incubation with LPS and IFN for 5 h led to the appearance of complexes III and IV, as well as several minor bands (complexes I, II, and V), the formation of all could be blocked by co-incubation with excess B oligonucleotide but not mutated B oligonucleotide. Supershift EMSAs indicated that complex III contained both p50 and p65 subunits, whereas complex IV did not contain p50 but did contain p65. These results demonstrate that lactacystin inhib- its the initial activation of NFB p50:p65 heterodimers, as well as causing their delayed re-activation.
Effects of Lactacystin on IB Levels-Since NFB activation requires degradation of inhibitory IB proteins, we examined the effects of lactacystin on IBϪ␣ and IBϪ␤ protein levels (Fig. 6). C6 cells displayed a basal level of both IB-␣ and -␤. After a 1-h incubation with LPS and IFN, the levels of both isoforms were decreased compared with basal levels. In the presence of lactacystin, the loss of both isoforms was reduced when measured at this time point. Re-expression of both IB proteins was detected at 6 h of incubation with LPS and IFN. However, in the presence of lactacystin the reappearance of IBϪ␣ was reduced, whereas the reappearance of IBϪ␤ was greatly increased. After 24 h of incubation in LPS and IFN, the levels of IBϪ␤ were equal to or greater than initial levels, whereas levels of IBϪ␣ were still reduced compared with zero time values. The presence of lactacystin did not modify IBϪ␣ levels when measured at 24 h, whereas IBϪ␤ levels were increased by lactacystin. These results indicate that lactacystin delayed the degradation of pre-existing IB proteins. Additionally, incubation with lactacystin led to an increase in IBϪ␤, but not IBϪ␣, levels at later times.
The above results demonstrate that lactacystin selectively increased IB-␤ expression. Since increases in the levels of hypophosphorylated IB-␤ can increase and prolong NFB activation, we determined the effects of lactacystin on the phosphorylation state of IB-␤ (Fig. 7). As previously shown (39), alkaline phosphatase treatment of cell extracts to remove phosphoryl groups resulted in increased electrophoretic mobility of IB-␤, confirming that non-phosphorylated IB-␤ migrates faster than phosphorylated forms and indicating that in quies-cent C6 cells the majority of IB-␤ was phosphorylated. Following incubation with LPS and IFN, the levels of IB-␤ decreased in a time-dependent manner, and by 4 h a hypophosphorylated form having increased mobility was apparent. This form was still present at 6 h but not increased versus 4 h. In the presence of lactacystin, the loss of IB-␤ was attenuated, and a hypophosphorylated, faster migrating species could be detected as early as 1 h after incubation. Between 2 and 6 h of incubation, the IB-␤ in the lactacystin-treated cells was primarily non-phosphorylated, and overall levels were greater than those seen in the non-treated cells. These results indicate that lactacystin promoted the re-expression of hypophosphorylated IB-␤ at earlier times and at greater levels than did LPS plus IFN alone.
To determine if overexpression of IB-␤ could reverse the effects of lactacystin, we transfected C6 cells with a cDNA encoding full-length rat IB-␤ or a truncated IB-␤ lacking the carboxyl-terminal PEST site and casein kinase II phosphoryl- ation sites (Fig. 8). Expression of full-length IB-␤ dose-dependently inhibited the induction of the NOS2 promoter upon incubation with LPS and IFN. Similarly, the augmentation of NOS2 promoter induction due to co-incubation with lactacystin was dose-dependently blocked by expression of full-length IB-␤. By contrast, expression of a carboxyl-terminal deleted IB-␤ only weakly inhibited NOS2 promoter induction, both in the presence and absence of lactacystin. These results demonstrate that the effects of lactacystin on NOS2 expression can be blocked by overexpression of IB-␤ and require the PEST region and casein kinase II phosphorylation sites.
The observation that lactacystin increased IBϪ␤ but not IBϪ␣ levels after a 6-h incubation with LPS and IFN suggested that lactacystin might be directly increasing IBϪ␤ expression. To test this, we examined if lactacystin had any effect on IBϪ␤ in the absence of inflammatory stimulation (Fig. 9). At all time points examined, IBϪ␤ levels were increased by lactacystin versus incubation in media alone (which led to a slight increase in IBϪ␤ levels over the course of 24 h). In extracts from the same cells, levels of IBϪ␣ were increased by lactacystin after 6 and 24 h of incubation but not after 1 h of incubation. These data indicate that lactacystin, even in the absence of inflammatory stimulation, causes an increase IBϪ␤ as well as IBϪ␣ protein levels although with distinct kinetics.
Induction of the Heat Shock Response by Lactacystin-To examine further the mechanism by which lactacystin modifies NOS2 expression and IBϪ␤ expression, we tested lactacystin effects on heat shock protein (HSP) expression, since it has been shown that proteasome inhibitors can promote HSP expression (45,(52)(53)(54) and since HSP70 can suppress NOS2 expression (22,51,55,56). C6 cells were incubated with an inhibitory dose of lactacystin (6.5 M), and cell lysates were analyzed by Western blots for the presence of HSP70 after different periods of incubation (Fig. 10). HSP70 was detected at low levels in control cells, as well as in cells incubated for up to 1 h with lactacystin. However, after 6 h of incubation, cells treated with lactacystin expressed easily detectable HSP70 levels. Incubation with NOS2 inducers alone did not induce HSP70 expression, although the combination of lactacystin with LPS/IFN induced a higher level of HSP70 than lactacystin alone. After 20 h incubation, HSP70 levels were further increased by lactacystin, as well as by the combination of LPS/ IFN plus lactacystin. These results demonstrate that lactacystin induces HSP70 expression in glial cells and suggests that HSP70 expression at early times may contribute to inhibition FIG. 7. Lactacystin induces accumulation of hypophosphorylated IB-␤. A, equal aliquots (approximately 10 g) of whole cell lysate from untreated C6 cells were incubated with calf intestinal alkaline phosphatase (CIP, 10 units) for 30 min at 37°C, and then proteins were analyzed by Western blots with antibody to IB-␤. Treatment with calf intestinal alkaline phosphatase increased the electrophoretic mobility of IB-␤ (arrows), whereas the mobilities of minor cross-reacting proteins (arrowheads) were not affected. B, total cell lysates were prepared from C6 cells incubated with 1 g/ml LPS plus 20 units/ml IFN-␥ for the indicated times, in the presence of 0 or 1 M lactacystin (Lacta). Aliquots containing equal amounts of protein were analyzed by Western blots with antibody to IB-␤. Incubation with LPS and IFN-␥ (upper gel) led to loss of pre-existing IB-␤ (arrow), and re-expression was not as pronounced at 6 h as for the experiment shown in Fig. 6. However, in the presence of lactacystin (bottom gel), the initial loss of IB-␤ was delayed, and re-expression was apparent by 6 h. 2 cells that contain a 2.2-kb rat NOS2 promoter driving expression of the luciferase reporter gene were transfected with the indicated amounts of rat IB-␤ or ⌬IB-␤ expressing plasmid. Transfection of cells with empty vector alone (up to 1 g) had no effect on NOS2 promoter activation. After 24 h, 1 g/ml LPS plus 20 units/ml IFN-␥ was added to the cells, and NOS2 promoter activation was measured 24 h later. The fold activation by LPS and IFN-␥ was 2.0-fold over basal levels (slightly less than the 3-fold shown in Fig. 4) and is given as 100%. The results shown are from one of three independent experiments, each done in duplicate or triplicate, and the variability between experiments was less than 10%. IB-␤ Is a Heat Shock Protein-The observation that lactacystin induces a HSR in glial cells, as well as induces IBϪ␤ expression, suggested to us that IBϪ␤ could be responding to the HSR. To test this hypothesis, we examined IB-␣ and IB-␤ levels in C6 cells after thermal induction of a HSR (Fig. 11). As previously reported (51), 20 min of hyperthermia induced IBϪ␣ expression when measured 4 h after heat shock. In the same cells, we found that a 20-min heat shock also increased IBϪ␤ expression. These results demonstrate that IBϪ␤, as is the case for IBϪ␣, is induced by the HSR suggesting that IBϪ␤ is also a stress protein. DISCUSSION Although NOS2 expression in astrocytes and glial cells requires NFB activation, the precise role of distinct members of the IB family in mediating glial NFB activation and thus NOS2 expression has not been fully characterized. In the current study we have extended our previous results showing that inhibition of the 26 S proteasome inhibits astroglial and C6 cell NOS2 expression (22). We have found that incubation with low doses of the highly specific proteasome inhibitor lactacystin led to a paradoxical increase, rather than decrease, in NOS2 expression. This conclusion is supported by our observations that low doses of lactacystin increased nitrite accumulation, NOS2 mRNA levels, and NFB activation when measured after 24 h of incubation. That the potentiating effect is mediated by a delayed event, and not by an early event in NOS2 expression, is further supported by the observation that high doses of lactacystin added after NOS2 inducers also led to an increase in nitrite accumulation. Finally, examination of the degradation and re-expression patterns of IBϪ␣ and IB-␤ suggests that lactacystin may be potentiating NOS2 expression due to a reduction in the re-expression of IBϪ␣ and/or an increase in the re-expression of IB-␤ (Fig. 6), both of which could contribute to an overall increase in the levels of sustained NFB activity. The finding (Fig. 7) that lactacystin increased expression of a hypophosphorylated IB-␤ suggests that non-phosphorylated IB-␤ contributes to increased NFB activation and NOS2 expression. These findings suggest that changes in the relative levels of IB proteins dictate the overall level and duration of NFB activation, which can result in either inhibitory or stimulatory effects, and point out that using proteasome inhibitors to block inflammatory gene activation could, if added subsequent to initial inflammatory activation, result in stimulatory rather than inhibitory actions.
Our results showing NOS2 promoter activation (Fig. 4) and NFB activation (Fig. 5B) at 20 -24 h after the addition of inflammatory stimuli indicate that stimulation of glial cells with LPS and IFN leads to long lasting activation of NFB. Sustained NFB activation has previously been suggested to be necessary for macrophage NOS2 activation since inhibition of delayed but not initial NFB activation decreased LPS-dependent NOS2 expression (26). These observations are consistent with the findings that NFB activation due to stimulation with LPS (or IL1-␤), in contrast to other inducers such as tumor necrosis factorϪ␣ or phorbol 12-myristate 13-acetate, results in a persistent activation rather than a rapid transient response (29,39). The kinetics of rapid, transient NFB activation are such that by 3-5 h after inducer addition, there remains little or no active NFB in the nucleus as determined by electrophoretic mobility shift assays (EMSAs). In contrast, incubation of cells with LPS or IL1-␤ results in "persistent" NFB activation, characterized by a slow, gradual diminution of active NFB over the course of 24 -48 h. We therefore conclude that glial NOS2 induction involves the sustained activation of NFB.
Until recently, it has not been clear how the known regulation of IBϪ␣ could account for sustained NFB activation. Following initial NFB activation, newly synthesized IBϪ␣ translocates to the nucleus and binds to existing NFB⅐DNA complexes causing repression of ongoing transcription, dissociation of NFB from DNA, and export of inactive NF⅐IB complex back to the cytoplasm. The net effect is rapid abrogation of transcriptional activity. Observations made in cells that are characterized by constitutive NFB activation (B-cells, human immunodeficiency virus-infected cells, and cancer cells) suggested that differences in the levels and stability among IB family members could strongly influence the overall response to inflammatory stimuli, and suggested the IBϪ␤ protein as a dual positive/negative regulator of NFB activity. Following cloning of IBϪ␤ (29), it was discovered that during persistent but not transient NFB activation, there is re-expression of IBϪ␤. Whereas in some cases newly IBϪ␤ is rapidly phosphorylated at its carboxyl terminus, and functions like IBϪ␣ to inhibit NFB activation (57), further studies demonstrated that during persistent activation, newly synthesized IBϪ␤ is not phosphorylated. Instead, hypophosphorylated IBϪ␤ associates with NFB (and prevents NFB association with inhibitory IB Ϫ␣) and acts as a chaperone to bring NFB to the nucleus where it is transcriptionally active (29,37,38,40,57). Thus, the phosphorylation state of IBϪ␤ dictates its role as an inhibitory versus a stimulatory molecule. Our results are consistent with this model, since treatment of cells with LPS and IFN led to a re-expression of predominantly hypophosphorylated IBϪ␤, whose expression was further increased by lactacystin, providing evidence that glial NOS2 expression involves persistent NFB activation.
It is not yet clear what are the signals causing re-expression of IBϪ␤, since in contrast to IBϪ␣, IBϪ␤ does not respond to NFB activation. However, several reports indicate that LPS and certain cytokines, which induce IBϪ␤ expression, induce a HSR both in vivo and in vitro (58,59), suggesting that IBϪ␤ may be responding as a stress protein. In our studies, incubation of cells with lactacystin, in the absence of LPS or cytokines, induced the expression of HSP70, consistent with previous observations that proteasome inhibitors induce a HSR (42,45,(52)(53)(54). In the same cells, we also observed that lactacystin alone induced IBϪ␤ expression. In view of the findings that IB-␣ is an HSP (47), the above considerations suggest that IBϪ␤ is also a stress protein and that increased IBϪ␤ expression by lactacystin is due to establishment of an HSR. Our findings that brief hyperthermia-induced IB-␤ expression is consistent with the possibility that IB-␤ is an HSP. Howevern it is not yet known whether or not the promoter region of IB-␤ contains a heat shock element, which can be identified in the promoter region of IBϪ␣ (60). We suggest that in cells stimulated with LPS and cytokines, lactacystin enhances the HSR FIG. 11. IB-␣ and IB-␤ are heat shock proteins. C6 cells were incubated at 43°C for the indicated times and then allowed to recover at 37°C for 4 h. At that time, aliquots of whole cell protein lysates were analyzed by Western blot for levels of IB-␣ (left panel) and IB-␤ (right panel). This experiment was repeated once with similar results. initiated by LPS thus leading to an increase in IBϪ␤ expression. The mechanism by which lactacystin, or other proteasome inhibitors, can induce an HSR is not clear; however, it has been suggested that the proteasome constitutively degrades a protein which otherwise would activate heat shock factors (61).
Our results suggest that the re-expression of both IBϪ␣ and IBϪ␤ is critical to determine the duration and magnitude of NOS2 expression, and that inhibition of the proteasome by lactacystin exerts multiple effects that can lead to suppression or augmentation of overall NFB activation and NOS2 expression. To explain the dual effects, we propose the following model. In the absence of lactacystin, incubation of C6 cells with a combination of LPS plus IFN causes the rapid loss of preexisting IBs (both ␣ and ␤) leading to NFB activation. At later times, re-expression of IBϪ␣ occurs that sequesters active NFB thus reducing overall levels of NFB activity. There is also delayed expression of IBϪ␤, which contributes to sustained NFB activation, due to prevention of inhibition by IBϪ␣, as proposed for persistent NFB activation in other cell types.
When lactacystin is present at low doses, partial inhibition of proteasomal degradation of IBs permits low levels of initial NFB activation to occur, sufficient to activate initial NOS2 transcription, as well as cause limited new expression of IBϪ␣, although less than what occurs in the absence of lactacystin. The low dose of lactacystin is sufficient to induce IBϪ␤ expression, which reduces inhibition of the initially activated NFB by residual or newly made IBϪ␣. Overall, the net result is an increased amount of NFB-dependent transcription compared with that observed in the absence of lactacystin. At higher concentrations, lactacystin potently inhibits the initial degradation of pre-existing IBs by the proteasome and thus prevents the initial NFB activation and any NOS2 transcription. Although the lactacystin also induces de novo IBϪ␤ expression, there is no pre-existing activated NFB for association with and prolongation of transcriptional activity. Overall, the net effect is the absence of NOS2 expression. However, at later times, limited proteolysis of IBs occurs, due to either some reversal of proteasome inhibition by lactacystin and/or to de novo synthesis of proteasome catalytic subunits. This results in a delayed and limited amount of NFB activation to occur. At these times, the presence of lactacystin-induced IBϪ␤ increases the extent of NFB activation by preventing any inhibition by remaining or newly expressed IBϪ␣. Overall, there is a low but significant degree of NFB activation occurring at later times, which is sufficient to induce NOS2 expression, as indicated by our observations of delayed appearance of NOS2 promoter activity and mRNA in a high dose of lactacystin.
In summary, our results using lactacystin are consistent with a role for persistent NFB activation in NOS2 expression, mediated by increases in IB-␤ expression. The findings that lactacystin can have contrasting effects on NOS2 expression are reminiscent of findings that proteasome inhibitors can either inhibit or induce apoptosis, depending upon the concentration used (62). Our findings that lactacystin induces both IB-␤ expression, which may contribute to sustained NFB activation, and at the same time HSP70 which can reduce NFB activation and NOS2 expression (22,46,47,(51)(52)(53)(54) indicate multiple interacting levels of NFB regulation. Our observations that even high doses of lactacystin can potentiate NOS2 expression, if added to cells subsequent to initial NOS2 induction, raise concerns about the therapeutic use of proteasome inhibitors to reduce inflammatory damage (54, 63), since in many cases inhibitors would be administered to an already existing inflammatory condition that could result in exacerbation, rather than attenuation, of the response.