Ceramide-induced and Age-associated Increase in Macrophage COX-2 Expression Is Mediated through Up-regulation of NF-κB Activity*

We have shown that the age-associated increase in lipopolysaccharide (LPS)-stimulated macrophages (Mφ) prostaglandin E2 (PGE2) production is because of ceramide-induced up-regulation of cyclooxygenase (COX)-2 transcription that leads to increased COX-2 expression and enzyme activity. To determine the mechanism of the age-related and ceramide-dependent increase in COX-2 transcription, we investigated the role of various transcription factors involved in COX-2 gene expression. The results showed that LPS-initiated activations of both consensus and COX-2-specific NF-κB, but not AP-1 and CREB, were significantly higher in Mφ from old mice than those from young mice. We further showed that the higher NF-κB activation in old Mφ was because of greater IκB degradation in the cytoplasm and p65 translocation to the nucleus. An IκB phosphorylation inhibitor, Bay 11-7082, inhibited NF-κB activation, as well as PGE2 production, COX activity, COX-2 protein, and mRNA expression in both young and old Mφ. Similar results were obtained by blocking NF-κB binding activity using a NF-κB decoy. Furthermore, NF-κB inhibition resulted in significantly greater reduction in PGE2 production and COX activity in old compared with young Mφ. Addition of ceramide to the young Mφ, in the presence or absence of LPS, increased NF-κB activation in parallel with PGE2 production. Bay 11-7082 or NF-κB decoy prevented this ceramide-induced increase in NF-κB binding activity and PGE2 production. These findings strongly suggest that the age-associated and ceramide-induced increase in COX-2 transcription is mediated through higher NF-κB activation, which is, in turn, because of a greater IκB degradation in old Mφ.

It is well documented that T cell-mediated immune function declines in old animals and elderly humans compared with their young counterparts (1,2). The age-associated dysregula-tion in macrophages (M) 1 contributes to the impaired T cell function with aging. We, as well as others, have demonstrated that immune cells, including M, from old animals and humans produced more PGE 2 than those from their young counterparts (3)(4)(5)(6)(7). We further showed that the increased PGE 2 production by M contributes to the decline in T cell-mediated function with aging (8).
Cyclooxygenase (COX) is the rate-limiting enzyme that catalyzes the conversion of arachidonic acid (AA) to PG endoperoxide (PGH 2 ), which is further converted to different PGs and thromboxane. COX is hence a key factor in PG synthesis. Two isoforms of COX have been identified: a constitutive form, COX-1 (9,10), and the inducible counterpart, COX-2 (11,12). We have demonstrated that the age-associated increase in M PGE 2 production is because of higher COX activity in M from old mice compared with those from young mice. This increased COX activity is, in turn, a result of increased expression of COX-2 protein and mRNA (13). In a recent study, we further demonstrated that the age-related increase in COX-2 mRNA was because of a higher level of ceramide in old M compared with those of young, which induced up-regulation of COX-2 transcription (14). In addition, we showed that the effect of ceramide was not mediated through components of the mitogen-activated protein kinase pathway, c-Jun NH 2 -terminal kinase, extracellular signal-regulated kinase, or p38 (14).
To determine the mechanism of the age-related and ceramide-induced increase in COX-2 transcription, we investigated the role of various nuclear transcription factors that are involved in COX-2 gene expression. The binding sites for several nuclear transcription factors, such as nuclear factor B (NF-B), nuclear factor interleukin-6, and cAMP-responsive element (CRE), have been identified on the promoter region of the COX-2 gene (15)(16)(17). A number of studies have suggested that NF-B activation and binding to its cognate site on the COX-2 promoter region are required to induce COX-2 expression (16,18,19). Dysregulation of NF-B activation has been indicated in certain inflammatory diseases (20 -22), in which COX-2-catalyzed prostaglandin production may play an important role. It is thus feasible that the ageassociated increase in COX-2 expression may be mediated through a corresponding change in the regulation of NF-B with aging. However, to date, age-related changes in M NF-B activity and the role of NF-B in age-related upregulation of COX-2 have not been demonstrated.
Along with NF-B, another redox-sensitive transcription factor, activator protein-1 (AP-1), has been shown to be involved in COX-2 transcriptional regulation (23,24). Although an independent AP-1 binding site has not been recognized on the COX-2 promoter, it was reported that the binding site for AP-1 in the COX-2 promoter is a CRE binding site (15,24,25). A number of studies suggest that binding of nuclear proteins, such as CREbinding protein (CREB) and c-Jun, to CRE, an element of COX-2 promoter, induces COX-2 transcription (26,27). Accordingly, we examined the roles of NF-B, AP-1, and CREB in the age-associated up-regulation of M COX-2 transcription.
Our recent study showed that the intracellular concentration of ceramide was higher in LPS-stimulated M of old mice compared with those of young mice and that this increased level of ceramide mediates the age-associated up-regulation of COX-2 transcription (14). Whereas the effect of ceramide on regulation of transcription factors has not been well defined, previous studies showed that ceramide induced AP-1 (28) and NFB activation (29). We hypothesize that up-regulated COX-2 transcription with aging is because of altered activation of transcription factors involved in COX-2 expression by ceramide. We demonstrate here that of the three transcription factors studied, only NF-B contributes to the age-associated up-regulation of COX-2. The ageassociated increase in NF-B activation is because of enhanced IB degradation in the cytoplasm, resulting in increased nuclear translocation of activated NF-B. Ceramide induces increased COX-2 activation and the consequent PGE 2 production through up-regulating NF-B activation.

EXPERIMENTAL PROCEDURES
Animals-Specific pathogen-free male young (4 -6 months) and old (22-24 months) C57BL/6NIA mice were obtained from National Institute on Aging colonies at Harlan Teklad (Madison, WI). Mice exhibiting skin lesions, visible tumors, or splenomegaly were excluded from the study. Mice were housed individually in microisolator cages at a constant temperature (23°C) with a 12-h light-dark cycle and were fed autoclaved mouse chow Harlan 7012 (Harlan Teklad) and water ad libitum. All conditions and handling of the animals was approved by the Animal Care and Use Committee of the Jean Mayer Human Nutrition Research Center on Aging, Tufts University, and were in accordance with guidelines provided by the National Institutes of Health Guide for the Care and Use of Laboratory Animals.
Peritoneal Macrophage Isolation-Mice were injected per peritoneum with 3 ml of 2.98% thioglycollate to elicit M. Three days later, the mice were euthanized via CO 2 asphyxiation. Peritoneal exudate cells were obtained by peritoneal lavage with cold Ca 2ϩ -and Mg 2ϩ -free Hanks' balanced salt solution (Sigma). Cells were centrifuged and resuspended in RPMI 1640 medium (BioWhittaker, Walkersville, MD) supplemented with 10 mM HEPES (Sigma), 2 mM glutamine (Invitrogen), 100 units/ml penicillin, and 100 g/ml streptomycin (Invitrogen), and 2% fetal bovine serum. Peritoneal exudates were enriched for M by their adherence to tissue culture-treated plastic dishes or plates for 2 h at 37°C in 5% CO 2 . Nonadherent cells were removed by vigorous washing and the remaining cells were at least 90% macrophages as assessed by the expression of cell surface markers Mac-1 and F4/80. Cells were monitored for their general condition and viability throughout the study as assessed by morphology, adherence, and trypan blue exclusion. No cytotoxicity was observed in treated cells compared with untreated controls. We previously conducted a number of studies using resident macrophages (8,13,30,31). Because of a large number of M that were necessary for various experiments and the limited number of resident M obtainable from each mouse, we used thioglycollate-elicited M. Prior to use of thioglycollate in our experiments, we compared the magnitude and pattern of responsiveness between resident and thioglycollate-elicited M. Although there are differences between the two types of cells in their ability to response to certain stimulation agents, the relative response pattern, the age-related difference, as well as the response to in vitro intervention are the same between these two cell types. Particularly and of relevance to this study, thioglycollate-elicited M showed the age-associated difference in COX-2 expression similar to that observed with resident M (8,13,30).
Preparation of Cytosolic and Nuclear Extracts-M (2 ϫ 10 7 ) in a 10-cm dish were incubated overnight in serum-free RPMI 1640 medium. The cells were washed and then stimulated by lipopolysaccharide (LPS, Escherichia coli serotype 0111:B4, Sigma) at 5 g/ml for various lengths of time. This concentration of LPS was used because our testing experiments indicated it to be optimal for production of PGE 2 and nitric oxide. A parallel experiment was conducted using IL-1␤ (R & D Systems, Minneapolis, MN) at 50 ng/ml as stimulator. For the NF-B inhibition study, the cells were preincubated with an inhibitor of IB-␣ phosphorylation, Bay 11-7082 (Biomol Research Laboratories, Plymouth Meeting, PA) for 30 min, or a NF-B decoy (see "NF-B Decoy Approach" below) for 24 h before LPS stimulation. To increase intracellular ceramide levels, cell-permeable C2-ceramide (30 M) (Matreya, Pleasant Gap, PA) was added to the cell cultures with or without the presence of LPS and the cells were incubated for different times. The concentration of 30 M was chosen based on our previous study in which different doses of ceramide were used and shown to induce an efficient COX-2 expression and PGE 2 production at this level (14). At the end of the stimulation period, the cells were washed with cold PBS and then collected with a cell scraper. The cells were resuspended in a hypotonic buffer (10 mM HEPES, pH 7.9, 2 mM MgCl 2 , 10 mM KCl, 0.1 mM EDTA, 1 mM dithiothreitol, 0.5 mM phenylmethysulfonyl fluoride, and 0.5% Nonidet P-40) and incubated on ice for 10 min. After the cell lysates were centrifuged at 15,000 ϫ g for 1 min, the supernatants were collected as a cytosolic fraction and stored at Ϫ70°C. The remaining pellets were resuspended in a high salt buffer (50 mM HEPES, pH 7.9, 300 mM NaCl, 50 mM KCl, 0.1 mM EDTA, 1 mM dithiothreitol, 0.5 mM phenylmethysulfonyl fluoride, and 10% glycerol) and incubated in rotation at 4°C for 30 min. The nuclear lysate was centrifuged at 15,000 ϫ g at 4°C for 30 min. The supernatant was collected as a nuclear fraction and stored at Ϫ70°C.
Electrophoretic Mobility Shift Assay (EMSA)-For the NF-B binding activity assay, both consensus and COX-2 promoter-specific sequences were used. A double-stranded oligonucleotide containing an NF-B consensus sequence (5Ј-AGTTGAGGGGACTTTCCCAGG C-3Ј) was purchased from Promega (Madison, WI). A COX-2-specific NF-B binding oligonucleotide (distal, Ϫ408/Ϫ388, 5Ј-GAGGTGAGGGGATT-CCCTTAG-3Ј) and its complementary sequence were synthesized by the Tufts University Core Facility laboratory and were annealed before labeling. For AP-1 and CREB binding assays, the double-stranded consensus oligonucleotides (AP-1, 5Ј-CGCTTGATGAGTCAGCCGGA-A-3Ј and CREB, 5Ј-AGAGATTGCCTGACGTCAGAGAGCTAG-3Ј) were purchased from Promega. All the oligonucleotides were end labeled using [␥-32 P]ATP (3000 Ci/mmol, PerkinElmer Life Sciences) and T4 polynucleotide kinase (Promega). 32 P-Labeled probes were purified using MicroSpin G-25 columns (Amersham Biosciences). For each reaction, nuclear extracts (2 g of protein) were incubated with labeled oligonucleotide in the presence of binding buffer (Promega) at room temperature for 20 min. To ensure specificity of probe binding, a 50-fold excess of unlabeled (cold) and mutant oligonucleotides were added to the nuclear samples and incubated for 10 min before the labeled oligonucleotide was added. In supershift assays, antibodies specific for the p65 or p50 subunit of NF-B (Santa Cruz Biotechnology, Santa Cruz, CA) were added to the binding reaction and incubated for 30 min at room temperature before the labeled oligonucleotide was added. Protein-DNA complexes were resolved at 350 V for 1 h in 4% polyacrylamide gels and visualized using Kodak x-ray film. Bands were quantified by ChemiImager (Alpha Innotech Corp., San Leandro, CA).
Western Blot-The cytosolic and nuclear samples were prepared as described under "Preparation of Cytosolic and Nuclear Extracts" and were used for IB-␣ and p65 detection, respectively. For COX-2 and inducible nitric-oxide synthase (iNOS) detection, M were preincubated with or without Bay 11-7082 for 30 min and then stimulated by LPS (5 g/ml) for 16 h. Total cellular lysates were collected and 25 g of protein from each sample was electrophoresed in a 10% SDS-polyacrylamide gel and transferred to nitrocellulose membranes. After blocking with 5% nonfat dry milk in TBS containing 0.1% Tween 20 overnight, the membranes were incubated with the respective antibodies (all from Santa Cruz Biotechnology) for 1 h. The membranes were rinsed and then incubated with the corresponding secondary antibodies conjugated with alkaline phosphatase (Tropix, Inc., Bedford, MA) for 1 h. After being rinsed, the membrane was incubated in a Chemiluminescent Detection System (Tropix) for 4 min and then exposed to film. The equal loading across the samples was first estimated by staining the membranes with Ponceau S (Sigma) and further confirmed by reprobing the stripped membranes with ␤-actin antibody (Sigma). All bands were quantified by ChemiImager (Alpha Innotech). The bands of interest molecules were normalized with ␤-actin bands and presented as relative density ratio. PGE 2 Production and COX Enzyme Activity-Peritoneal M (1 ϫ 10 6 cells/well) were plated to 24-well plates and isolated by adherence as described above. The cells were preincubated with or without Bay 11-7082 for 30 min before being stimulated by 5 g/ml LPS (Sigma), 30 M ceramide (Matreya), or both for 12 to 24 h. After the supernatants were removed and stored at Ϫ70°C for analysis of accumulated production of PGE 2 , the cells were layered with 1 ml of medium containing 30 M AA and incubated at 37°C for 10 min for determination of COX activity as described by Fu et al. (32). Total cellular COX activity can be measured by adding excess exogenous AA to M, because the intracellular enzyme pool is saturated with the substrate and is functioning at maximal velocity. After 10 min, 2.1 mM aspirin was added to inactivate the COX enzyme activity. Supernatants were immediately removed and stored at Ϫ70°C. Cells were then incubated with 1 M NaOH for 5 min, at which time the supernatant was removed and stored at Ϫ20°C for protein analysis by the bicinchoninic acid protein assay kit (Pierce). PGE 2 was measured by radioimmunoassay (RIA) as previously described (4).
COX-2 mRNA Reverse Transcriptase-PCR-M were preincubated with or without Bay 11-7082 for 30 min and then stimulated by LPS (5 g/ml) for 4 h. Total RNA was isolated using the Totally RNA Isolation kit (Ambion, Austin, TX). Two g of total RNA was reverse-transcribed to first-strand cDNA using random hexamer, and amplified by PCR using the Superscript amplification kit (Invitrogen). The PCR conditions for COX-2 mRNA were 1 cycle for 2 min at 94°C, followed by 30 cycles of 1 min at 94°C and 5 min at 55°C. Mouse exon 8 sense primer (5Ј-ACTCACTCAGTTTGTTGAGTCATTC-3Ј) and exon 10 antisense primer (5Ј-GTAATTGGGATGTCATGATTAGTTT-3Ј) were used to generate 583-bp PCR products. To normalize the COX-2 mRNA reverse transcriptase-PCR results, 18 S rRNA primers and the competitors at a ratio of 4:6 (Ambion) were used to generate 18 S rRNA PCR products from the same cDNA samples used in COX-2 mRNA PCR assays. Our PCR conditions for both murine COX-2 mRNA and 18 S rRNA were tested to be within the linear range of PCR product formation (14). The PCR products were resolved by electrophoresis in an ethidium bromidestained 1.2% agarose gel and the bands were visualized by ethidium bromide staining and quantified using ChemiImager (Alpha Innotech).
NF-B Decoy Approach-The COX-2-specific NF-B binding oligonucleotide (distal, 5Ј-GAGGTGAGGGGATTCCCTTAG-3Ј) and its complementary sequence (5Ј-CTAAGGGAATCCCCTCACCTC-3Ј), and their mutated counterparts (5Ј-GAGGTGAGGGCCTTCCCTTAG-3Ј and 5Ј-CTAAGGGAAGGCCCTCACCTC-3Ј) were custom synthesized by Qiagen Operon (Alameda, CA). The underlined letters denote phosphorothioated bases and the bold letters mark mutations. The complementary oligonucleotides were annealed to double strands by heating at 90°C for 5 min and then cooling down to room temperature within 3 h. To be efficiently delivered to the cells, the double-stranded oligonucleotides were mixed with LipofectAMINE reagent (Invitrogen) and incubated at room temperature for 40 min. The complex was then added to the cell cultures at 0.5 to 10 nM for oligonucleotides and 5 g/ml for LipofectAMINE reagent. The cells were incubated in antibiotics and serum-free RPMI 1640 medium for 24 h, after which the cells were washed and then stimulated for varied times depending on the purpose.
Statistical Analysis-Data were analyzed using SYSTAT statistical software (SYSTAT 10, 2000; Evanston, IL). Paired Student's t test was used to determine the effect of incubation time and inhibitor. The difference between two age groups was assessed using nonpaired Student's t test. Results are expressed as mean Ϯ S.E. Significance was set at p Ͻ 0.05.

NF-B Binding Activity in Murine M Increases with Ag-
ing-We previously showed that the age-associated increase in PGE 2 production is because of the ceramide-induced up-regulation of COX-2 transcription with aging (14). Of the transcription factors found in the promoter region of the COX-2 gene, NF-B is the most intensively studied and several investigations have linked its activity to the activation of the COX-2 gene (16,18). Although the age-related up-regulation of NF-B activation has been shown in rat gastric mucosa (33), kidney (34), liver, heart (35), and brain (36), its binding activity in M, a major source of PGE 2 , has not been compared between young and old animals. In this study, the peritoneal M from young and old mice were stimulated with LPS for different times as indicated in Fig. 1. NF-B binding activity in nuclear fractions was assessed by EMSA. Fig. 1A shows the results obtained using an NF-B consensus oligonucleotide as the binding motif. Without stimulation, there was no detectable binding activity in either young or old M. At every time point following LPS stimulation, the NF-B binding activity was higher in old M compared with that of young M. To assure the specificity of the binding, a 50-fold concentrated unlabeled (cold) NF-B consensus oligonucleotide was added to compete with the 32 Plabeled NF-B oligonucleotide. The binding was competed out in both young and old M at peak time (2 h). When a mutated NF-B oligonucleotide was added, however, the binding was not affected, further indicating the sequence specificity. To determine the composition of the NF-B proteins in the binding complex, we used antibody supershift in EMSA. As seen in Fig.  1A, the binding complex was shifted by antibodies to p65 and p50 units. Next, we used a COX-2-specific NF-B oligonucleotide in EMSA. Similar to the experiments in which the consensus oligonucleotide was used, a consistently higher binding activity was observed in old compared with young M (Fig. 1B). The specificity was examined using the samples from old M at peak stimulation time (2 h) and 50-fold cold oligonucleotide was shown to completely compete out the band. The supershift assay yielded similar results to those observed in the assay using the consensus oligonucleotide.
AP-1 and CREB Binding Activity in M Does Not Change with Aging-After establishing an age-related difference in NF-B binding activity, we examined whether two other transcription factors, AP-1 and CREB, which have been shown to be involved in COX-2 regulation, are affected by the aging process. Culture conditions and LPS doses were the same as those used in the NF-B gel shift assay. Both AP-1 and CREB were activated by LPS treatment, but the induction was less potent than that seen in NF-B so that the autoradiography required 5-8fold longer exposure times than that with the NF-B probe. The binding activity for both AP-1 and CREB peaked at around 1 h poststimulation and no significant difference was detected between young and old mice in either AP-1 or CREB activation (data not shown).
Old M Have Higher IB Degradation and p65 Translocation Than Young M-To determine the mechanism of the age-related increase in NF-B activation, we examined the two key steps preceding the NF-B binding activity: IB degradation in the cytoplasm and NF-B translocation to the nucleus. The results indicated that there was no difference between young and old M in expression of IB under resting conditions. However, after LPS stimulation, there was greater IB-␣ degradation in old M than that in young M. The degradation also occurred faster, as shown at 15 min after stimulation, and appeared to recover more slowly when compared with young M (Fig. 2A). We then examined the p65 appearance in nuclear extracts following its translocation from the cytoplasm. As shown in Fig. 2B, p65, which was not detected in the nucleus before stimulation, gradually increased during the time from 15 min to 2 h after LPS stimulation, followed by a decline between 2 and 4 h. These results indicate that the increased NF-B binding activity with age is because of different rates of IB degradation, and subsequent NF-B translocation, two immediate events prior to the binding of the NFB to its target genes.
An IB Inhibitor Prevents NF-B Activation in Both Young and Old Mouse M-As no age-related change was observed in AP-1 and CREB activation, the relationship between NF-B activation and age-related increase in COX-2 expression was further investigated. Because increased NF-B activation in old M was associated with a higher IB phosphorylation and the consequent degradation, to inhibit NF-B activation, we used Bay 11-7082, an inhibitor of IB phosphorylation, and therefore NF-B activation (37). This inhibitor has been shown not to have as broad an effect as do other NF-B activation inhibitors such as aspirin, caffeic acid, and N-acetylcysteine. To demonstrate that the effect of Bay 11-7082 is specific to NFB activation, we also examined its effect on AP-1 and CREB activation. As shown in Fig. 3B, neither AP-1 nor CREB binding activity was affected by Bay 11-7082 under the same condition as that used for determination of NF-B activation. These results indicate that Bay 11-7082 can be used as a tool to determine the role of NF-B activation in age-related up-regulation of COX-2 expression.
Inhibition of NF-B Preferentially Reduces PGE 2 Production and COX Activity in Old M-Of the transcription factors that control COX-2 transcriptional activation, NF-B was the only one that exhibited the age-related increase. Therefore, we next examined whether changing NF-B activation would alter the LPS-stimulated production of PGE 2 , a representative COX product in M. After M from either young or old mice were preincubated with IB inhibitor Bay 11-7082 for 30 min, the cells were stimulated by LPS for 24 h and PGE 2 production was then determined. As shown in Fig. 4A, PGE 2 production was low in unstimulated cells and greatly increased with LPS stimulation. This LPS-induced PGE 2 production was 5-fold higher in old compared with young mice. LPS-induced PGE 2 production was inhibited by Bay 11-7082 in a dose-dependent manner between 0.5 and 2 M, whereas at a dose of 5 M, no further inhibition was observed. Because old M have higher NF-B activation and COX-2 expression than those of young M, we next examined the effect of NF-B inhibition on the ability of young and old M to produce PGE 2 . Results indicated that when NF-B activation was reduced with Bay 11-7082, PGE 2 production in old M was inhibited more significantly (p Ͻ 0.05) compared with that of young M (80 versus 58, 27 versus 12, and 39 versus 13% of their control levels were observed in young and old mice in the presence of Bay 11-7082 at 1, 2, and 5 M, respectively). COX is the rate-limiting enzyme in pros-

FIG. 4. Inhibition of NF-B activation dose dependently reduces M PGE 2 production and M COX activity.
The peritoneal M from young and old mice were pretreated with increasing concentrations of IB phosphorylation inhibitor Bay 11-7082 for 30 min and then stimulated with LPS (5 g/ml) for 24 h at 37°C. The supernatant was collected and analyzed for PGE 2 production. After the supernatant was collected for PGE 2 production, the cells were washed and then further incubated in the medium containing 30 M AA for 10 min at 37°C. Aspirin (2.1 mM) was added at the end of incubation to terminate the reaction. The supernatant was collected and analyzed for the PGE 2 synthesized utilizing exogenous AA to assess the COX enzyme activity. PGE 2 concentrations in the samples were determined using RIA and adjusted for total cell protein. The results for PGE 2 production and COX enzyme activity are shown in A and B, respectively. The data are mean Ϯ S.E. of four independent experiments in each of which a duplicate measurement was conducted. The bars bearing different letters within the same case (lower or upper) represent significant difference with p Ͻ 0.05. The lowercase and uppercase letters represent young and old mice, respectively.
taglandin biosynthesis and we have demonstrated that increased COX activity is the major contributing factor to the age-related increase in PGE 2 production (13). We, therefore, examined COX activity in cells treated under the same condition as in the test for PGE 2 production. As shown in Fig. 4B, LPS-stimulated M from old mice have significantly higher COX activity than those from young mice. NF-B inhibition by Bay 11-7082 reduced COX activity in a dose-dependent manner. Furthermore, the inhibition of COX activity in old M was more significant (p Ͻ 0.05) compared with that of young M (75 versus 48, 36 versus 12, and 3 versus 0.2% of their control levels were observed in young and old mice in the presence of Bay 11-7082 at 1, 2, and 5 M, respectively).
To rule out the possibility that Bay 11-7082 may directly inactivate the COX enzyme rather than inhibit transcriptional activation of COX-2 through reducing NF-B activation, we also tested its direct effect on COX activity by adding Bay 11-7082 to cultures after LPS stimulation. After 24 h of LPS stimulation, COX-2 would be fully activated. The presence of Bay 11-7082 for 30 min thereafter should not change the levels of COX-2 enzyme, but would be adequate to affect enzyme activity if it did have a direct effect on the enzyme. The results showed that addition of Bay 11-7082 after LPS stimulation did not change COX activity in either young or old M (data not shown), thus a direct effect on COX enzyme activity can be ruled out.

Age-related Increase in PGE 2 Production as Well as the Involvement of NF-B Is Not Limited to LPS as Stimulant-
Increased PGE 2 production with age is not limited to that stimulated by LPS. Previously we showed that in addition to LPS, calcium ionophore or T cell mitogens also stimulated more PGE 2 production in splenocytes of old mice or peripheral blood mononuclear cells of elderly humans compared with their young counterparts (4 -6). To further confirm this in the elicited peritoneal M, we conducted a dose-response experiment using IL-1␤, another common stimulant of COX-2. Fig. 5A shows that IL-1␤ dose dependently induced PGE 2 production in both young and old M, but old M produced significantly more PGE 2 than young M in response to IL-1␤ stimulation. Next, we stimulated the cells with 50 ng/ml IL-1␤ in the presence of Bay 11-7082 and found that IL-1␤-stimulated PGE 2 production was also inhibited by Bay 11-7082 in a dose-dependent manner (Fig. 5B). The patterns of response in young and old M were similar to those when LPS was used as a stimulant. Similar results were obtained when COX activity was evaluated (data not shown).
Inhibition of NF-B Reduces COX-2 mRNA and Protein Levels-Because our previous studies (13,14) showed that the age-related increase in COX activity is because of increased expression of the COX-2 mRNA and protein, we determined COX-2 mRNA and protein levels in cells that were incubated with Bay 11-7082 prior to LPS stimulation. As shown in Fig.  6A, unstimulated M had very low expression of COX-2. However, LPS significantly induced COX-2 expression and the LPSstimulated COX-2 expression was higher in old compared with young M. Inhibition of NF-B activation dose-dependently reduced the COX-2 mRNA expression in both young and old M. The change in COX-2 protein levels was generally in accordance with the change in COX-2 mRNA levels, although the dose response was not as pronounced as seen in mRNA expression (Fig. 6B). These results demonstrate that there is an age-dependent increase in activation of NF-B, which results in higher transcription of the COX-2 gene, increased COX-2 mRNA, COX-2 protein, and greater PGE 2 production.
Nitric oxide and iNOS have been shown to be up-regulated with age (31,38). The promoter region of the murine iNOS gene has a NF-B binding site (39). Thus, to further prove the link between NF-B activation and its target genes in the context of age-related events, we chose to measure iNOS protein expressed by young and old M under the same condition used for COX-2 determination. As demonstrated in Fig. 6C, LPS-induced iNOS expression was higher in old rather than in young M and was also inhibited by Bay 11-7082 in a dose-dependent manner.
Ceramide Increases LPS-induced Activation of NF-B in Young M-Our recent work (14) showed that old M produce higher levels of intracellular ceramide, compared with those from young mice after LPS stimulation. Furthermore, we showed that increasing ceramide levels in young M, by adding exogenous ceramide, significantly increased COX-2 expression. This effect was specific to ceramide and did not depend on its downstream metabolite, sphingosine (14). Because no age-related difference in mitogen-activated protein kinase activity (14) or AP-1 and CREB activation was observed, we hypothesized that NF-B mediates ceramide-induced COX-2 up-regulation. In this study, we added exogenous ceramide to the young M and determined its effect on NF-B activation. First, we tested the effect of different concentrations of ceramide on NF-B activation. The results showed that incubating cells in the presence of ceramide for 2 h induced, although not as strongly as LPS, NF-B activation in a dose-dependent manner (Fig. 7A). We then tested the time course of ceramide-induced NF-B activation in the absence or presence of Bay 11-7082. As shown in Fig. 7B, ceramide induced NF-B activation at all time points tested. This ceramide-induced NF-B activation was prevented by addition of the IB phosphorylation inhibitor Bay 11-7082. Furthermore, we determined the effect of ceramide on LPS-stimulated NF-B activation and demonstrated that LPS induced higher NF-B activation in the cells supplemented with ceramide compared with those treated with vehicle control (Fig. 7C). It should be mentioned that ceramide by itself is a weak inducer of NF-B activation and a much longer exposure time was needed to obtain a comparable band density to that seen with LPS. However, ceramide had an additive effect on LPS-induced NF-B activation.
To confirm that the altered NF-B activation by ceramide or Bay 11-7082 is coupled to the changes in COX-2 activation, we measured PGE 2 production under the same condition. As shown in Fig. 8, addition of exogenous ceramide increased PGE 2 production and this ceramide-induced increase was prevented by inhibiting NF-B activation. Similar to that seen in the NF-B binding assay, addition of ceramide increased LPSstimulated PGE 2 production, an effect that was also prevented by inhibiting NF-B activation.
NF-B Decoy Blocks NF-B Binding Activity and Inhibits PGE 2 Production and COX-2 Expression-To further confirm the role of NF-B, we repeated some of the above experiments by employing NF-B decoy as an alternative and more specific approach to block NF-B binding to COX-2 promoter. Use of FIG. 6. Inhibition of NF-B dose dependently inhibits expression of COX-2 mRNA and protein, and iNOS protein in young and old mouse M. Peritoneal M from young and old mice were pretreated with increasing concentrations of IB phosphorylation inhibitor Bay 11-7082 for 30 min and then stimulated with LPS (5 g/ml) at 37°C for 4 and 16 h for mRNA and protein assays, respectively. A, the isolated total RNA (2 g) was used to generate first-strand cDNA and then COX-2 mRNA level was determined using PCR as described under "Experimental Procedures." To normalize COX-2 mRNA, the 18 S rRNA PCR products were generated from the same cDNA samples used in COX-2 mRNA PCR assays. The results are representative of three independent experiments. Total cell lysates (25 g of protein per lane) were used to determine COX-2 (B) or iNOS (C) protein using Western blot analysis. After COX-2 or iNOS bands were visualized, membranes were stripped and reprobed with the antibody to ␤-actin to serve a loading control. Results are representative of four independent experiments for COX-2 and two independent experiments for iNOS.

FIG. 7. Ceramide induces by itself, and also enhances LPSstimulated NF-B binding activity, which is inhibited by blocking NF-B activation.
A, peritoneal M from young mice were incubated in the presence of ceramide at different concentrations as indicated for 2 h at 37°C. B, peritoneal M from young mice were preincubated with or without Bay 11-7082 (5 M) for 30 min. Ceramide (30 M) was added to the cells and incubation was continued at 37°C for the additional times as indicated. C, in a separate experiment, M were stimulated with LPS (5 g/ml) during the same time course in the presence or absence of ceramide (30 M). The nuclear extracts were prepared as described under "Experimental Procedures." Nuclear extracts (2 g) were incubated with 32 P-labeled COX-2-specific NF-B oligonucleotide and subjected to an EMSA. The experiments were repeated twice and similar results were obtained. the NF-B decoy has been shown to successfully suppress COX-2 expression (40). The COX-2-specific NF-B decoys have the identical sequences to those used for EMSA but modified on the 3 bases at each end by phosphorothioation to prevent being digested in the cells. The NF-B decoy competes with the COX-2 promoter for binding to the activated NF-B dimers and thus, block COX-2 gene activation. As shown in Fig. 9A, NF-B decoy dose-dependently inhibited LPS-induced NF-B binding activity while its mutated form did not have any effect. The role of NF-B in the additive effect of ceramide on LPS-induced PGE 2 production was further confirmed by data shown in Figs. 9 and 10. As seen in Fig. 9, NF-B decoy but not the mutant abrogated the ceramide and LPS-induced NF-B binding activity (Fig. 9B).
Furthermore, we examined whether this blocked NF-B binding would similarly affect PGE 2 production and COX-2 expression. Mutated NF-B decoy did not have a significant effect on PGE 2 production (data not shown). NF-B decoy, however, significantly inhibited LPS-stimulated PGE 2 production both in the presence and absence of ceramide (Fig. 10A). The changes in PGE 2 production was associated with similar changes in COX-2 protein expression (Fig. 10B). DISCUSSION We previously showed that the higher PGE 2 production by M from old mice was because of their increased expression of COX-2 mRNA (13). Furthermore, we showed that the increase in COX-2 mRNA was because of transcriptional up-regulation of COX-2 (14). Transcription factors NF-B, AP-1, and CREB have been indicated in regulation of COX-2 activation. In this study, we tested the involvement of these transcription factors in the age-associated and ceramide-induced up-regulation of COX-2 activation. First we compared the activation levels of NF-B, AP-1, and CREB in peritoneal M from young and old mice. We found that in response to LPS stimulation, old M had significantly higher consensus as well as COX-2-specific NF-B binding activity compared with young M. However, neither AP-1 nor CREB showed any significant change with age. The age-associated change in transcription factor activation has not been well studied and in particular, no information was available regarding their changes with age in M. Kim et al. (34) reported that NF-B binding activity in rat kidney increased with age. Helenius et al. (35) showed an age-related increase in the nuclear binding activity of NF-B, but not those of Sp1 and AP-1, in rat liver and heart. Increased NF-B and AP-1 activation with age was also observed in rat gastric mucosa cells (33). M are the major source of inflammatory mediators including PGE 2 . These M-originated mediators are involved in the pathogenesis of inflammatory, cardiovascular, and neoplastic diseases (41)(42)(43). Most of these diseases are more prevalent in the elderly population. Because NF-B has been indicated in regulation of various M-originated mediators, our finding that consensus NF-B activation in M is up-regulated with aging might shed light on the mechanism of age-related changes in other M-originated mediators such as nitric oxide and proinflammatory cytokines.
Binding of activated NF-B to the COX-2 promoter region has been suggested to be necessary for COX-2 transcriptional activation (16,18,19). To confirm this and also determine its role in the age-associated increase of COX-2 activation, we inhibited NF-B activation by employing an IB phosphorylation inhibitor, Bay 11-7082. In agreement with the result reported by others (37), this inhibitor effectively prevented NF-B activation in this study. Furthermore, the NF-B inhibition dose dependently, up to 2 M, reduced LPS-induced PGE 2 production in both young and old M. However, inhibition of NF-B activation resulted in a significantly larger reduction in PGE 2 synthesis in old M compared with that in young M. Because old M have higher NF-B activity as well

FIG. 9. NF-B decoy blocks LPS or LPS plus ceramide induced NF-B binding activity.
A, NF-B decoy or mutated NF-B decoy at indicated concentrations were prepared as described under "Experimental Procedures." Peritoneal M from young mice were preincubated in the presence of the decoys for 24 h at 37°C. The culture medium was then replaced with new medium containing LPS (5 g/ml) and incubated for 2 h. The nuclear extracts were prepared as described under "Experimental Procedures." Nuclear extracts (2 g) were incubated with 32 P-labeled COX-2-specific NF-B oligonucleotide and subjected to an EMSA. B, the cells were preincubated with only one high concentration (10 nM) of NF-B decoy or its mutant. LPS (5 g/ml) were than added to stimulate the cells in the presence or absence of ceramide (30 M). The experimental procedures are the same as described above in A.
as PGE 2 production compared with young M, these results further confirm the involvement of NF-B in the age-related up-regulation of COX-2. Because PGE 2 synthesis is determined by both substrate availability and COX activity, we measured COX activity by providing excessive exogenous arachidonic acid so that substrate availability would not be a limiting factor. The results showed dose-dependent inhibition of COX activity by Bay 11-7082. The degree of inhibition was significantly higher in old M compared with that in young M.
To stimulate M for the activation of COX-2 and PGE 2 production, we have mainly used LPS. It has been questioned whether the age-related difference in COX-2 expression is a phenomenon specific for LPS, merely reflecting the difference in LPS signal transduction at its receptor level. This is not likely because we previously observed an age-related increase in PGE 2 production when different immune cells from both mice and humans and several other stimulating agents were used (4,6). In fact, M Toll-like receptor 4 expression decreases with aging (57). Thus, the age-related up-regulation of LPSstimulated COX-2 activation involves a post-receptor signal transduction event, such as NF-B, as suggested by this study. This was further strengthened in this study by the observation that when IL-1␤ was used in place of LPS to stimulate M, those from old mice have significantly higher PGE 2 production compared with those from young mice. Furthermore, the IL-1␤-induced increase in PGE 2 production was abrogated by inhibiting NF-B activation.
If age-associated up-regulation of COX-2 is mediated through increased NF-B activation, it will be predicted that other NF-B target genes may also demonstrate an up-regula-tion with age and their expression could also be suppressed when NF-B activation is blocked. We chose iNOS as such a candidate to substantiate this speculation. Increased NO production and iNOS expression have been shown in old compared with young murine M (31,38). The murine iNOS gene has an NF-B binding site in its promoter region and pyrrolidine dithiocarbamate, an NF-B inhibitor, blocks both NF-B activation and NO production in LPS-stimulated M (39). Increased iNOS expression has also been shown in vascular smooth muscle cells from old rats compared with those from young rats and this age-related up-regulation is associated with NF-B activation (44). In this study, we confirmed the previous finding by showing higher iNOS expression in old M compared with young M. More importantly, we further demonstrated that blocking NF-B activation reduced iNOS expression. This observation added further support for the involvement of NF-B as a mechanism underlying the age-associated up-regulation of certain genes and their products.
The prototypical and most abundant form of NF-B complex is the p65 and p50 heterodimer (45). In most resting cells, NF-B is sequestered in the cytoplasm as an inactive precursor in complex with the inhibitory protein IB. By binding to the p65 component, IB inhibits transactivation of the p65 and p50 heterodimer, and thus blocking the translocation of the dimer to the nucleus (46,47). In response to activation signals, IB is phosphorylated and degraded, allowing NF-B release and further translocation to the nucleus where it regulates gene expression. Thus, the degradation of IB and translocation of NF-B are closely linked to the activity of NF-B binding to DNA. Although several members of the IB family have been identified, the best characterized IB is IB-␣. In the current study, we found that upon LPS stimulation, old M had increased degradation of IB-␣ in the cytoplasm, which was accompanied by an increased appearance of p65 in the nucleus. These results strongly suggest that the increased degradation of IB-␣ and the subsequent p65 translocation to the nucleus are the main contributors to the increased NF-B binding activity seen in old M compared with that in young M. An age-related decrease in cytoplasmic IB-␣ and an age-related increase in nuclear p65 was observed in unstimulated rat kidney tissue by Kim et al. (34). However, another study showed that NF-B activation increased with age but IB-␣ and p65 levels were unaffected in the rat gastric mucosa (33). IB is phosphorylated by the action of IB kinase (IKK) (48). The activation of IKK is in turn mediated by phosphorylation through NFB-inducing kinase (49,50). The cause of increased IB degradation was not determined in this study and is the subject of our future investigation. However, our recent work (14) suggested that ceramide might be involved in the ageassociated increase in IB degradation. We showed that, following LPS stimulation, M from old mice generated significantly more ceramide than those from young mice. Furthermore, we showed that C2-ceramide significantly increased LPS-induced COX activity and COX-2 expression in young M; this effect of ceramide was not mediated through mitogenactivated protein kinase. In this study, we showed that ceramide dose dependently induced NF-B activation in young M. This effect of ceramide was blocked by the IB phosphorylation inhibitor, Bay 11-7082. Addition of ceramide increased LPSstimulated NF-B activation, which was also inhibited by Bay 11-7082. These changes in NF-B activation caused by ceramide, in the presence or absence of LPS, were mirrored by those in PGE 2 production. Bay 11-7082 treatment prevented the ceramide-induced effect on both NF-B activation and PGE 2 production in the presence or absence of LPS. In this study, we also used a COX-2-specific NF-B decoy as an alter- FIG. 10. NF-B decoy inhibits ceramide, LPS, or LPS plus ceramide-induced PGE 2 production through suppressing COX-2 expression. A, the peritoneal M from young mice were preincubated with NF-B decoy or its mutant for 24 h. LPS (5 g/ml), ceramide (30 M), or both were then added to stimulate the cells for 12 h. The supernatants were collected and analyzed for PGE 2 production. PGE 2 concentrations in samples were determined using RIA and adjusted for total cell protein. Data are mean Ϯ S.E. of four independent experiments. * indicates a significant difference at p Ͻ 0.05. B, the cell treatments were similar to those described above. Total cell lysates (25 g of protein per lane) were used to determine COX-2 protein levels using Western blot analysis. After COX-2 bands were visualized, membranes were stripped and reprobed with the antibody to ␤-actin to serve as a loading control. native approach to block NF-B activation. The decoy competes with the COX-2 promoter for binding to the activated NF-B dimers and as a result, blocks COX-2 gene activation. The results obtained using the NF-B decoy were similar to those obtained with Bay 11-7082. Taken together, because old M have higher levels of ceramide compared with young M and because increasing the level of ceramide in young M enhances NF-B activation, COX-2 expression, and PGE 2 production, all of which are abrogated by Bay 11-7082 or an NF-B decoy, these data strongly suggest that the age-associated increase in COX-2 transcription is because of the ceramide-induced upregulation of NF-B activation.
Determining NF-B activation after reducing ceramide levels in old M would have provided further support for our proposed mechanism, however, the approach is not feasible at the present time. Although knockout mice for acidic sphingomyelinase are available, these animals develop Neimann-Pick disease and die by the age of 10 months, making them unsuitable to address the role of ceramide in NF-B up-regulation in aged mice (typically more than 20 months old). In addition, the neutral, but not the acidic sphingomyelinase has been indicated to be responsible for the age-related increase of ceramide levels in other tissues. However, specific neutral sphingomyelinase inhibitors are not commercially available.
The mechanism for the ceramide-induced increase in NF-B activation has not been determined and will be the subject of our future study. It has been suggested that ceramide may induce NF-B activation through its effect on the isotype of protein kinase C (PKC-). PKC-was shown to activate NF-B through phosphorylation of IB-␣ in NIH-3T3 fibroblasts (51). Furthermore, overexpression of PKC-positively modulates IKK␤ activity, whereas the transfection of a PKC-dominant negative mutant severely impairs the activation of IKK␤ (52). PKC--deficient mice have impaired NF-B activation (53). It was reported that PKC-can be activated in vitro by ceramide and in vivo by sphingomyelinase, which produces ceramide, in NIH-3T3 fibroblasts (54). Finally, ceramide was shown to induce the translocation of PKC-to both the nucleus and membrane in rat astrocytes (55) and hepatocytes (56).
In summary, our data demonstrated for the first time, that NF-B activation is up-regulated with aging in murine M, and that this age-associated up-regulation of NF-B activation mediates the higher age-associated expression of COX-2. Increased IB degradation and p65 translocation with aging represent important determinants of the increased NF-B activation observed in old M. Combined with our previous study (14), in which ceramide was shown to mediate the age-associated up-regulation of COX-2 transcription, the current study suggests that increased ceramide levels in old M induces higher NF-B activation, leading to increased COX-2 transcription. These findings will help to further understand the mechanism of the age-associated increase in COX-2 expression and associated diseases.