Ceramide mediates age-associated increase in macrophage cyclooxygenase-2 expression.

Previously, we showed that macrophages (MØ) from old mice have significantly higher levels of lipopolysaccharide (LPS)-induced prostaglandin E(2) (PGE(2)) production than young mice, due to increased cyclooxygenase-2 (COX-2) mRNA levels. The aim of the current study was to determine the underlying mechanisms of age-associated increase in COX-2 gene expression. The results demonstrate that increased COX-2 mRNA expression in the old mice is due to a higher rate of transcription rather than increased stability of COX-2 mRNA. Furthermore, the results show that LPS-induced ceramide levels from the old mice are significantly higher than those of young mice, whereas there is no age-related difference in concentration of its down stream metabolite, sphingosine. The addition of ceramide in the presence or absence of LPS resulted in a significant increase in PGE(2) production in a dose- and time-dependent manner. Inhibition of ceramide conversion to sphingosine had no effect on this ceramide-induced effect. The ceramide-induced up-regulation in PGE(2) production was mediated through increase in COX activity and transcriptional up-regulation of COX-2 mRNA. Collectively, these data suggest that the age-associated increase in MØ COX-2 mRNA is due to transcriptional up-regulation. Furthermore, this increase in transcription is mediated by higher cellular ceramide concentration in old MØ compared with that of young MØ.

(PGE 2 ) production contributes to the age-associated decline in T cell function (7). We further showed that the age-related increase in lipopolysaccharide (LPS)-induced MØ PGE 2 production is due to increased cyclooxygenase-2 (COX-2) mRNA and protein levels, leading to increased COX enzyme activity (8). PGE 2 is synthesized from the precursor arachidonic acid by the two cyclooxygenase isoenzymes: cyclooxygenase 1 (COX-1), which is a ubiquitously expressed enzyme, and COX-2, which has a low basal expression but is rapidly induced by inflammatory stimuli such as LPS.
Exogenously added membrane-permeable analogues of ceramide (C 2 -and C 6 -ceramide) as well as bacterial SMase augmented IL-1-induced PGE 2 production and COX-2 expression in human dermal fibroblast (23) and granulose cell cultures (24). Subsequent work showed that this effect was specific to COX-2 and not COX-1 (24,25). In addition to ceramide, its metabolite, sphingosine (23), has been shown to increase IL-1induced PGE 2 production. The effect of sphingosine, however, might be due to its retroconversion to ceramide (23). Venable et al. (26) showed that senescent WI38 human fibroblast cells had higher ceramide levels and activity of neutral SMase. Agerelated increases in brain and liver ceramide (27)(28)(29) and neutral SMase levels have also been reported. Since several lines of evidence indicated that LPS and ceramide share a common pathway(s) leading to induction of COX-2, we hypothesized that ceramide mediates the age-associated increase in COX-2 expression and PGE 2 production.
Recent studies have indicated that mitogen-activated protein kinase (MAPK) intermediates such as extracellular signalregulated protein kinase (ERK), c-Jun N-terminal kinases (JNK), and p38 (30) are involved in LPS-as well as ceramideinduced COX-2 gene expression. Moreover, several transcription factors, which bind COX-2 gene-regulatory regions such as AP-1, and cyclic AMP-response element-binding protein have been shown to be regulated by these MAPKs (30). Thus, we further hypothesized that ceramide mediates the age-associated increase in MØ COX-2 mRNA expression via MAPKs.

EXPERIMENTAL PROCEDURES
Animals-Specific pathogen-free male young (4 months) and old (24 months) C57BL/6NIA mice were obtained from National Institute on Aging colonies at Harlan Teklad (Madison, WI). Mice with tumors determined by gross pathological examination were excluded from the study. Mice were housed singly in microisolator cages at a constant temperature (23°C) with a 12-h light-dark cycle and fed laboratory chow and water ad libitum. All conditions and handling of the animals were approved by the Animal Care and Use Committee of the Jean Mayer Human Nutrition Research Center on Aging at 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-Peritoneal exudate cells from thioglycollate (3 ml of 2.98% broth for 3 days)-injected mice were obtained by peritoneal lavage with cold Ca 2ϩ -and Mg 2ϩ -free Hanks' balanced salt solution (Invitrogen). Peritoneal MØ were collected by centrifugation at 200 ϫ g at 4°C for 10 min followed by resuspension in endotoxinfree RPMI 1640 (BioWhittaker, Walkersville, MD) medium supplemented with 10 mM HEPES, 2 mM glutamine (Invitrogen), 100 units/ml penicillin, 100 g/ml streptomycin (Invitrogen), and 2% fetal bovine serum. The cells were plated on cell culture dishes or plates (Falcon Labware, Lincoln Park, NJ) and allowed to adhere for 2 h at 37°C in 5% CO 2 , at which time nonadherent cells were removed by vigorous washing. Peritoneal MØ prepared in this manner were at least 90% MØ, as assessed by expression of Mac-1 and F4/80 cell surface proteins (31). In addition, our preliminary experiments demonstrated a similar ageassociated increase in expression of COX-2 mRNA and PGE 2 production between LPS-stimulated resident and thioglycollate-elicited MØ.
Determination of mRNA Stability-Peritoneal MØ from young and old mice were stimulated with 5 g/ml LPS (Sigma) in RPMI medium for 2 h. The cells were washed twice prior to the addition of 2 g/ml actinomycin D (Sigma). Following the addition of actinomycin D, cells were harvested at 4, 6, and 8 h for Northern blot hybridization.
Northern Blot Hybridization-Total RNA was isolated using the Totally RNA isolation kit (Ambion, Austin, TX). Twenty g of total RNA was electrophoresed on a formaldehyde-containing 1.2% agarose gel and transferred to a nylon membrane (Ambion). Membranes were then prehybridized with ultrahybridization solution (Ambion) for 1 h prior to the addition of [ 32 P]dCTP-labeled COX-2 cDNA (kindly provided by Dr. D. Hwang, Pennington Biomedical Research Center, Baton Rouge, LA). Nylon membranes were then incubated at 42°C for overnight in ultrahybridization solution (Ambion). Membranes were washed twice for 5 min at 42°C in 2 ϫ SSC (0.3 M NaCl, 0.03 M sodium citrate) and 0.1% SDS, followed by washing at room temperature for 5 min in 0.1 ϫ SSC (15 mM NaCl, 1.5 mM sodium citrate) and 0.1% SDS. Washed membranes were then subjected to autoradiography at Ϫ80°C.
COX-2 mRNA and Heteronuclear RNA (hnRNA) RT-PCR-Total RNA was isolated using the Totally RNA isolation kit (Ambion). 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 conditions of PCR for both COX-2 mRNA and COX-2 hnRNA were one cycle for 2 min at 94°C and 30 cycles for 1 min at 94°C followed by 5 min at 55°C. For COX-2 mRNA PCR, mouse exon 8 sense primer (5Ј-ACTCACTCAGTTTGTTGAGTCATTC-3Ј) and exon 10 antisense primer (5Ј-GTAATTGGGATGTCATGATTAGTTT-3Ј) were used to generate 583-bp PCR products. For COX-2 hnRNA PCR, mouse COX-2 intron 8 was sequenced (published in GenBank TM ; accession number AF344876) from the mouse genomic DNA to generate intronspecific sense primer (5Ј-TGTTCTTGTAACATGACACTTAC-3Ј) and antisense primer (5Ј-GAAGTCCTATGTCTTGACCTCATCA-3Ј), which were used in PCR to generate 235-bp products. 18 S rRNA primers and the competimers (4:6 ratio, respectively; Ambion) were used to generate 18 S rRNA PCR products from all of the samples used in COX-2 hnRNA and COX-2 mRNA RT-PCR assays. The amplified PCR products using these primers were then visualized and quantitated using ethidium bromide-stained 1.2% agarose gel electrophoresis.
COX-2 Protein Western Blot-Peritoneal MØ from young mice were incubated in the presence of LPS (5 g/ml), C 2 -ceramide (30 M) (Matreya, Pleasant Gap, PA), or both for 16 h. Total cellular lysates were collected, and 30 g of protein from each sample was electrophoresed in 10% SDS-polyacrylamide gel and transferred to nitrocellulose membranes. The membranes were stained with Ponceau S (Sigma) to con-firm even loading across the samples. After blocking with 5% nonfat dry milk in Tris-buffered saline containing 0.2% Tween 20 overnight, the membranes were incubated with monoclonal goat anti-COX-2 antibody (Santa Cruz Biotechnology, Inc., Santa Cruz, CA) for 1 h. The membranes were rinsed and then incubated with rabbit anti-goat IgG secondary antibody conjugated with alkaline phosphatase (Tropix, Bedford, MA) for 1 h. The membranes were rinsed and incubated in the presence of the substrate (Tropix CDP-Star RTU) for 4 min. The bands were visualized by autoradiography.
Sphingolipid Measurement-Peritoneal MØ from young and old MØ were stimulated with 5 g/ml LPS in serum-free RPMI medium for specified times. The total ceramide levels were determined by thin layer chromatography followed by HPLC analysis. Briefly, the lipids were extracted by the modified Blight and Dyer method as described previously (32) and were analyzed by thin layer chromatography on silica gel using chloroform/methanol/triethylamine/2-propanol/0.25% potassium chloride (30:9:18:25:6, by volume) as the developing solvent. The regions migrating with a standard ceramide (bovine brain; Matreya) were scraped from the TLC plates and eluted with 1 ml of chloroform/ methanol (1:1, by volume) followed by 1 ml of methanol. The combined eluates were dried in vacuo and added to 1 nmol of N-acetyl-C 20sphinganine, as a standard, and ceramide and sphingosine masses were quantified by HPLC as previously described (29). PGE 2 Production and COX Enzyme Activity-Peritoneal macrophages were isolated, pooled, and plated (5 ϫ 10 5 cells/well) as described previously (8). One ml of RPMI with 0 or 5 g/ml LPS and various concentrations of C 2 -ceramide in the presence or absence of ceramidase inhibitor D-MAPP (Sigma) were added to each well in serum-free RPMI medium. Plates were incubated at 37°C in 5% CO 2 for 0, 6, 8, and 12 h, at which times the supernatants were removed and immediately frozen and stored at Ϫ70°C for analysis of ex vivo production of PGE 2 . After the removal of supernatants, cells were layered with 1 ml of RPMI medium containing 30 M arachidonic acid and incubated at 37°C for 10 min for determination of COX activity as described by Fu et al. (33). After 10 min, COX enzyme activity was inhibited with 2.1 mM aspirin. Protein concentration was measured using the bicinchoninic acid protein assay kit (Pierce). Supernatants were immediately removed and stored at Ϫ70°C. PGE 2 was measured by RIA as previously described (31).
Immunoprecipitation and in Vitro Kinase Assays-Whole-cell extracts containing 200 g of total protein were diluted to a 1-ml volume with lysis buffer (0.1% Triton X-100, 20 mM Tris, pH 7.4, 137 mM NaCl, 2 mM EDTA, 25 mM glycerophosphate, 1 mM sodium orthovanadate, 2 mM sodium pyrophosphate, 10% glycerol, 0.4 mM phenylmethylsulfonyl fluoride, leupeptin (1 g/ml), and pepstatin (1 g/ml)). The MAPKs were immunoprecipitated with 20 l of specific antibody to JNK1 (Santa Cruz Biotechnology), ERK1/2 (Santa Cruz Biotechnology), or p38 (ATF-2 polyclonal antibody; New England Biolabs, Beverly, MA) that were precipitated with 15 l of protein A-Sepharose (Zymed Laboratories Inc., San Francisco, CA). The resulting immunoprecipitates were washed extensively with lysis buffer and once with kinase assay buffer (25 mM HEPES, pH 7.6, 20 mM MgCl 2 , 2 mM dithiothreitol, 20 mM glycerol phosphate, 1 mM Na 3 VO 4 ). Substrates for kinase activity assays were myelin basic protein (Sigma) for the p44 ERK kinase, the N-terminal c-Jun obtained from a GST fusion protein purified on glutathione-agarose beads (Amersham Biosciences) for p46 JNK kinase, and ATF-2 protein (New England Biolabs) for p38 kinase. The kinase reaction was performed in kinase assay buffer containing 20 M unlabeled ATP and 10 Ci of [ 32 P]ATP (PerkinElmer Life Sciences) and 1 mg/ml substrate in a 30-l volume. Kinase reactions were performed at 30°C for 15 min and terminated by the addition of Laemmli gel-loading buffer. Activity was visualized by autoradiography of the dried SDSpolyacrylamide gel, and quantitation was assessed by PhosphorImager analysis (Amersham Biosciences).
Statistical Analysis-Data are reported as mean Ϯ S.E. and were analyzed by Student's t test.

RESULTS
Effect of Age on COX-2 mRNA Stability-We previously showed that LPS-induced COX-2 gene expression, 2 h following stimulation (optimal time), is significantly higher in MØ from old mice compared with those from young mice (8). To determine the underlying mechanism of the age-associated increase in COX-2 gene expression, we first determined whether age affected COX-2 stability. Young and old MØ were stimulated with LPS (5 g/ml) for 2 h and washed, and stimulated cells were incubated with transcription inhibitor actinomycin D (2 g/ml) for 0, 4, 6, and 8 h. Cells were then harvested for Northern blot hybridization to determine the level of COX-2 mRNA transcripts. As shown in Fig. 1, the rate of COX-2 mRNA degradation post-actinomycin D treatment showed no difference between the two age groups. These data suggest that differences in COX-2 mRNA stability do not contribute to agerelated increases in COX-2 gene expression. Noteworthy is the minimal LPS-induced COX-2 mRNA degradation with actinomycin D treatment in both young and old MØ (Fig. 1). This was also observed in other recent studies (34,35). Further, we tested other transcription inhibitors such as ␣-amanitin and dichlorobenzimidazole riboside. Treatment of LPS-stimulated MØ with these transcription inhibitors also resulted in a degree of COX-2 mRNA degradation inhibition similar to that observed with actinomycin D treatment (data not shown).
Effects of Age on COX-2 Transcription-Recent studies have shown utilization of hnRNA expression measurements as a substitute for transcription run-on assays (36 -38). To validate that our PCR conditions were within the linear range of PCR product formation, aliquots of PCR products were collected over increments of six PCR cycles up to 49 cycles. Results indicate that PCR conditions used in our study (30 cycles) were within the linear range of PCR product formation (Fig. 2). To determine whether increased COX-2 transcription results in agerelated up-regulation of COX-2 message, we treated MØ of young and old mice with LPS (5 g/ml) for 1 h and extracted total RNA, followed by COX-2 hnRNA RT-PCR assays. As shown in Fig. 3A, COX-2 hnRNA levels are significantly higher in the MØ of old compared with those of young mice, indicating that age-related increase in MØ COX-2 gene expression is due to up-regulation of COX-2 mRNA transcription. To determine whether increased transcription rate is coupled to increased mRNA expression, representative samples of reverse-transcribed COX-2 cDNA were amplified using both hnRNA-specific and mRNA-specific primers. As shown in Fig. 3B, LPS stimulated COX-2 hnRNA levels parallel to the COX-2 mRNA levels. These results suggest that LPS increases COX-2 message by increasing the rate of COX-2 gene transcription.
LPS-induced Ceramide Formation in Young and Old Peritoneal Macrophages-Recent studies have shown that old mice hepatocytes (29) as well as old rat brain (27) have higher levels of ceramide than that of young mice. Further, ceramide has been shown to induce COX-2 gene expression (30,39,40). Based on these findings, we hypothesized that age-related increase in MØ COX-2 message is mediated through higher ceramide concentration in the old mice. Since our previous data showed that LPS-induced COX-2 mRNA is significantly higher in the old mouse MØ following 2-h stimulation with LPS (8), we stimulated MØ with LPS (5 g/ml) at 0, 0.5, and 1 h to test the possible upstream regulatory role ceramide may be playing in LPS-induced and age-associated increases in COX-2 mRNA of old mouse MØ. As shown in Table I, MØ from old mice had higher ceramide generation after 30 min and 1 h stimulation with LPS compared with young mice (p Ͻ 0.05 at 30 min, and p Ͻ 0.1 at 1 h, respectively). Considering that old MØ have a significantly higher maximum LPS-stimulated COX-2 mRNA expression compared with those of young mice at 2 h post-LPS stimulation (8), the age-related differences observed in ceramide concentration at 30 min and 1 h could result in higher expression of COX-2 mRNA in old MØ compared with those of young mice. Taken together, these observations suggest that age-associated up-regulation of COX-2 mRNA may be mediated through higher LPS-induced ceramide levels.
Effects of Ceramide on PGE 2 Production of Young MØ-To further determine the role of ceramide in age-associated upregulation of PGE 2 production, young MØ were cultured in presence of C 2 -ceramide. As shown in Fig. 4A, the exogenous addition of C 2 -ceramide, which increases intracellular ceramide levels, enhanced PGE 2 formation in a time-and dosedependent manner in MØ from young mice. After 12-h stimulation with 30 M C 2 -ceramide, PGE 2 production in young MØ increased up to 5-fold compared with that of control. It is interesting to note that we previously reported a 5-fold difference in LPS stimulated PGE 2 production between young and old mouse MØ (8).
As shown in Fig. 4B, this ceramide-induced increase in PGE 2 production is due to increased COX enzyme activity.
Role of Sphingosine in Ceramide-induced PGE 2 Production-In addition to ceramide, its immediate downstream me-FIG. 1. Effect of age on COX-2 mRNA stability. Peritoneal MØ from 6-and 24-month-old C57BL/6NIA mice were stimulated with 5 g/ml LPS in RPMI medium supplemented with 2% fetal bovine serum for 2 h. MØ were washed twice prior to the addition of actinomycin D (2 g/ml) for 0, 4, 6, and 8 h, at which points cells were collected for total RNA isolation followed by Northern blot hybridization as described under "Experimental Procedures." COX-2 mRNA signals were normalized with 18 S rRNA. Data are reported as mean Ϯ S.E. (n ϭ 4).

FIG. 2. Linearity of PCR product formation.
Peritoneal MØ from young mice were stimulated with LPS (5 g/ml) for 1 h, and total RNA was extracted. Two g of total RNA was reverse-transcribed, and aliquots of PCR products were collected over increments of six PCR cycles up to 49 cycles. The amplified PCR products were then visualized and quantitated using ethidium bromide-stained 1.2% agarose gel electrophoresis.
tabolite sphingosine has been shown to increase expression of COX-2 (23). To determine that ceramide and not its downstream metabolite sphingosine induces the age-associated increase in PGE 2 production, we first stimulated MØ of young and old mice with LPS and measured the intracellular release of sphingosine. Furthermore, we treated young mouse MØ with ceramide in the presence or absence of ceramidase (catalyzes conversion of ceramide to sphingosine) inhibitor, D-MAPP. As shown in Fig. 5A, sphingosine levels were not significantly different between the two age groups. Furthermore, whereas the addition of C 2 -ceramide increased PGE 2 production, treatment with ceramidase inhibitor D-MAPP did not influence this ceramide-induced increase in PGE 2 production (Fig. 5B). In the absence of C 2 -ceramide, however, modest increases in PGE 2 production were observed with D-MAPP treatment, presumably due to accumulation of endogenous ceramide. Taken together, these data suggest that ceramide and not its down stream metabolite sphingosine mediates the age-related increase in MØ PGE 2 production.
Additive Effects of Ceramide and LPS on COX-2 Protein Expression and PGE 2 Production-Based on observations that LPS stimulation results in increased intracellular ceramide concentration (Table I) and that the exogenous addition of ceramide, which increases intracellular ceramide levels, results in increased PGE 2 production (Fig. 4A), we further hypothesized that adding LPS and C 2 -ceramide will have an additive effect on MØ PGE 2 production. To test this hypothesis, we treated young mouse MØ with LPS in the absence or presence of C 2 -ceramide and tested the changes in levels of PGE 2

FIG. 3. Effect of age on COX-2 transcription.
A, effect of LPS on COX-2 transcription. Peritoneal MØ from 6-and 24-month-old C57BL/ 6NIA mice were stimulated with LPS (5 g/ml) in RPMI medium supplemented with 2% fetal bovine serum for 1 h. MØ were collected for total RNA isolation followed by RT-PCR using primers specific for intron 8 of the mouse COX-2 gene. COX-2 hnRNA signals were normalized with PCR products of 18 S rRNA. Data are reported as mean Ϯ S.E. (n ϭ 3); *, p Ͻ 0.05. B, coupling of COX-2 mRNA and COX-2 hnRNA expression by LPS stimulation. The cell culture conditions were the same as in A. Two g of extracted total RNA was reverse-transcribed, and resulting cDNA was reverse-transcribed and amplified using COX-2 mRNA, COX-2 intron 8-specific primers, and 18 S rRNA-specific primers. Data are representative of two independent experiments.

TABLE I Effect of age on LPS-induced MA ceramide levels
Peritoneal MA from 6-and 24-month-old C57BL/6NLA mice were stimulated with LPS (5 g/ml) in serum-free RPMI medium for the specified times. Total ceramide levels were determined by thin layer chromatography followed by reverse-phase high performance liquid chromatography (HPLC) analysis. Data are reported as mean Ϯ S.E. For both young and old, peritoneal MA were pooled to make 6 -9 independent samples. *, p Ͻ 0.05 compared with young MA stimulated with LPS (5 g/ml); #, p Ͻ 0.1 compared with young MA stimulated with LPS.  4. Effects of ceramide on PGE 2 production and COX activity. A, dose response and time course of PGE 2 production in response to ceramide. Peritoneal MØ from young mice were cultured in serumfree RPMI medium and were treated with C 2 -ceramide at the specified doses and times. Cell culture supernatants were removed to measure PGE 2 concentrations by RIA. Medium containing appropriate vehicle was used as the control, and data were corrected accordingly. Data are presented as mean Ϯ S.E., n ϭ 4. *, p Ͻ 0.05 compared with control treatment. B, effect of ceramide on COX activity. Peritoneal MØ from young mice were cultured in serum-free RPMI medium and were treated with C 2 -ceramide (30 M) for 12 h. Levels of COX activity were determined by measuring PGE 2 production by RIA from the cell culture medium 10 min after the addition of exogenous arachidonic acid. Data are expressed as mean Ϯ S.E., n ϭ 4; *, p Ͻ 0.01 compared with control treatment.
production. As shown in Fig. 6A, culturing MØ in the presence of LPS and C 2 -ceramide (30 M) resulted in additive increases in PGE 2 production. These findings further support our hypothesis that LPS-induced ceramide generation mediates ageassociated increase in PGE 2 production. Moreover, we showed that this additive increase in PGE 2 production was due to an increase in COX-2 protein expression (Fig. 6B).
Effect of Ceramide on COX-2 mRNA Expression-To determine whether the ceramide-induced increase in COX activity is due to up-regulation of COX-2 mRNA, we stimulated MØ from young mice with C 2 -ceramide for 0, 0.5, 1, 2, and 4 h and determined COX-2 mRNA levels by RT-PCR using primers specific for COX-2 mRNA. As shown in Fig. 7, the exogenous addition of ceramide results in increased COX-2 mRNA of young MØ expression as early as 30 min post-C 2 -ceramide stimulation.
Effect of Ceramide on COX-2 hnRNA Expression-To determine the mechanism by which ceramide increases COX-2 gene expression in MØ, young mouse MØ were treated with ceramide (30 M) for 0, 0.5, 1, 2, and 4 h, and COX-2 hnRNA levels were measured by RT-PCR using primers specific for intron 8 of the mouse COX-2 gene. As shown in Fig. 8, the addition of C 2 -ceramide for up to 4 h increases COX-2 hnRNA levels in young MØ. To determine whether the ceramide stimulationinduced transcription rate is coupled to increased mRNA expression, representative samples of reverse-transcribed COX-2 cDNA from a time course experiment were amplified using both hnRNA-specific and mRNA-specific primers. As shown in Fig. 8, C 2 -ceramide stimulated increases of COX-2 hnRNA parallel the levels of COX-2 mRNA. These results suggest that ceramide increases COX-2 message by increasing the rate of COX-2 gene transcription.
Effects of LPS Treatment on ERK, JNK, and p38 Kinase Activities of Young and Old Mouse MØ-To determine whether the ceramide-induced age-associated increase in PGE 2 production is mediated through MAPKs, we first compared MAPK activity between young and old LPS-stimulated MØ. Our results showed that there were no statistically significant differences in LPS-induced ERK (Fig. 9A), JNK (Fig. 9B), or p38 (Fig.  9C) kinase activities. These results suggest that the age-related and ceramide-induced increase in COX-2 gene expression is not likely to be mediated through increased MAPK activities. The activity of LPS-treated cells for each animal was normalized to the control sample MAPK activity measurements, and LPS-induced ERK, JNK, and p38 activities showed significant increases (p Ͻ 0.05) compared with that of control values (data not shown).
Although we observed rather wide ranges of Phosphor-Imager density units (i.e. p38 activities of young-control, 11-226; young-LPS, 462-2862; old-control, 23-456; old-LPS, 548 -2859), LPS-stimulated values were consistently 5-25-fold higher when compared with control values (data not shown). Further, as shown in Fig. 9D (representative for all MAPKs), observed LPS-induced ERK activity was significantly higher than the control. DISCUSSION PGE 2 is a proinflammatory eicosanoid, which has been indicated in pathogenesis of cardiovascular diseases, cancer, and inflammation. An age-related increase in PGE 2 production has been demonstrated in animal models (8,41,42) and in humans (43,44). Previously we showed that the increase in MØ PGE 2 production with aging contributes to a decline in T cell-mediated immune function. Furthermore, we demonstrated that the age-associated increase in MØ PGE 2 production is due to increased COX-2 mRNA and protein levels, leading to an increase in COX enzyme activity. Age-related increases in COX-2 gene have implications for many age-associated diseases such as cardiovascular, inflammatory, and neoplastic diseases. Thus, determining causative factors for the up-regulation of COX-2 gene expression is imperative. The aim of our current study was to determine the underlying molecular mechanisms responsible for the age-associated increase in PGE 2 production.
The results demonstrated that the age-related increase in LPS-induced MØ COX-2 expression is due to an increase in transcription of COX-2 message rather than an increase in stability of COX-2 mRNA. COX-2 transcription is regulated by several factors such as cytokines (IL-6, IL-1␤, tumor necrosis factor-␣, IL-10), corticosteroids, and sphingolipids. Comparison of cytokine production between young and old MØ showed that IL-6, IL-1␤, tumor necrosis factor-␣, and IL-10 cannot account for the age-related difference in COX-2 mRNA, since there were no age-associated differences in MØ production of these cytokines (data not shown). Furthermore, we previously showed that the addition of recombinant IL-6 to MØ from young mice or that of anti-IL-6 antibody to MØ from old mice did not influence their ability to produce PGE 2 (45). We did not test the role of glucocorticoids in our system, since glucocorticoids inhibit rather than induce COX-2 gene expression (46 -48), and aging has been shown to be associated with increased glucocorticoid levels (49,50). We have also ruled out the age-FIG. 5. Age-associated increase in MØ PGE 2 production is not mediated by sphingosine. A, LPS-induced sphingosine formation in young and old peritoneal MØ. Peritoneal MØ from 6-and 24-month-old C57BL/6NIA mice were stimulated with 5 g/ml LPS in serum-free RPMI medium for specified times. The total sphingosine levels were determined by thin layer chromatography followed by HPLC analysis. Data are reported as mean Ϯ S.E., n ϭ 3. B, effects of ceramidase inhibitor (D-MAPP) on ceramide-induced PGE 2 production. Peritoneal MØ were cultured and treated in serum-free RPMI medium with C 2ceramide (30 M) for 12 h in the presence or absence of ceramidase inhibitor D-MAPP (20 M). Cell culture supernatants were removed to measure PGE 2 concentrations by RIA. Data are expressed as mean Ϯ S.E., n ϭ 3; *, p Ͻ 0.05; # , p Ͻ 0.1 compared with control treatment. related differences of MØ responsiveness to LPS treatment as a possible contributing factor, since the stimulation of MØ with other stimuli, including phytohemagglutinin and calcium ionophore A-23186, also resulted in age-related differences in PGE 2 production (51). Collectively, these data indicate that the agerelated difference in PGE 2 production might be mediated through the postreceptor events.
Another factor that has been shown to increase COX-2 expression is ceramide. Our results suggest that ceramide medi-ates the age-related increase in COX-2 expression. First we showed that old MØ have significantly higher LPS-induced ceramide levels compared with young mice, whereas there is no age difference in ceramide downstream metabolite sphingosine. Second, using C 2 -ceramide, we demonstrated that increasing ceramide levels in the young mice significantly increases PGE 2 production and COX activity. These effects of ceramide were not altered when ceramide conversion to sphingosine was blocked. Our results further showed that ceramide significantly enhances PGE 2 production by increasing COX-2 mRNA (Fig. 7) and that this increase in COX-2 gene expression is coupled to increased transcription of the COX-2 gene (Fig. 8).
Additive effects of LPS and ceramide in young MØ (Fig. 6, A  and B) strongly support our hypothesis that ceramide mediates the higher LPS-induced COX-2 expression in old MØ. From our previous study (8), we showed that old mouse MØ have higher levels of LPS-stimulated COX-2 mRNA, protein, and PGE 2 production. In a time frame that is consistent with age-associ- FIG. 6. Additive effects of LPS and ceramide on PGE 2 production and COX-2 protein expression. A, additive effects of LPS and C 2 -ceramide on PGE 2 production in young MØ. Peritoneal MØ were cultured in serum-free RPMI medium and treated with LPS in the presence or absence of C 2 -ceramide for 16 h at specified doses. Cell culture supernatants were removed to measure PGE 2 concentrations by RIA. Medium containing appropriate vehicle was used as the control, and data were normalized accordingly. Data are presented as mean Ϯ S.E., n ϭ 4. *, p Ͻ 0.05 compared with control; #, p Ͻ 0.01 compared with LPS alone. B, additive effects of LPS and C 2 -ceramide on COX-2 protein expression in young MØ. The cell culture and LPS and C 2 -ceramide treatment conditions were the same as in A. The resulting signal from the Western blot was detected by chemiluminescence as described under "Experimental Procedures." FIG. 7. Effect of ceramide on COX-2 mRNA expression. Peritoneal MØ from young mice were cultured in serum-free RPMI medium and were treated with C 2 -ceramide (30 M) for 0, 0.5, 1, 2, and 4 h. Two g of extracted total RNA was reverse-transcribed, and the resulting cDNA was amplified using COX-2 mRNA-specific primers or 18 S rRNA-specific primers (as described under "Experimental Procedures"). COX-2 mRNA was normalized using 18 S rRNA PCR products. Data are expressed as mean Ϯ S.E., n ϭ 5. *, p Ͻ 0.05; #, p Ͻ 0.1 compared with control treatment.
FIG. 8. Ceramide increases COX-2 mRNA by increasing COX-2 hnRNA expression. Effects of ceramide on COX-2 hnRNA and coupling of COX-2 mRNA and COX-2 hnRNA expression were tested using peritoneal MØ from young mice that were stimulated with C 2 -ceramide (30 M) for the indicated times. Two g of extracted total RNA was reverse-transcribed and the resulting cDNA was reverse-transcribed and amplified using COX-2 hnRNA-, COX-2 mRNA, or 18 S rRNAspecific primers. Data are representative of two independent experiments. ated up-regulation of COX-2 mRNA, ceramide also accumulated at a significantly higher concentration in the old mouse MØ compared with the MØ of young mice (Table I). Based on these observations, we suggest that LPS-induced higher accumulation of intracellular ceramide results in increased COX-2 mRNA, COX-2 protein, and PGE 2 production in old mice. Although LPS stimulation in MØ from SMase knockout mice would provide further support for our proposed mechanisms, unfortunately only knockout mice for one isoform of SMase (i.e. the acidic form) are available. These animals, however, develop Niemann-Pick disease and die by the age of 10 months and thus would not be suitable to address the role of ceramide in COX-2 up-regulation of aged mice (typically more than 20 months old). Furthermore, the neutral and not the acidic SMase has been indicated in the observed age-related increase of ceramide levels in other tissues. FIG. 9. Effects of LPS treatment on MAPKs activity. Peritoneal MØ from young and old mice were cultured in serum-free RPMI medium in the presence or absence of LPS (5 g/ml) for 30 min. A, anti-ERK antibody-protein A complex was incubated with myelin basic protein in kinase buffer containing [ 32 P]ATP. The kinase reaction was resolved on a 15% SDS-PAGE, and the gel was dried and autoradiographed. Quantitation of activity was done by densitometry measurements of the band representing the phosphorylated substrate. B, immunoprecipitated p38 antibody-protein A complex was incubated with ATF-2, as exogenous substrate, in kinase buffer containing [ 32 P]ATP. p38 activity was measured in a similar manner as in ERK activity. C, immunoprecipitated JNK antibody-protein A complex was incubated with p46 N-terminal c-Jun-GST fusion protein, as exogenous substrate, in kinase buffer containing [ 32 P]ATP. JNK activity was measured in a similar manner as ERK activity. Activity of LPS-treated cells for each animal is normalized to the untreated sample for all MAPK activities. D, representative ERK activity bands from peritoneal young and old MØ control treatment and LPS stimulation. Data are expressed as mean Ϯ S.E., n ϭ 3-8.
The underlying mechanism for higher ceramide levels in old MØ was not determined in this study. The age-associated increase in LPS-induced ceramide generation in the old animals may be due to an increase in the sphingomyelinase activity (28,29,52), due to increased sphingomyelin levels, or due to both. Increases in age-related sphingomyelin levels have been observed in liver (29,53), brain (54,55), and nerve cells (56). Moreover, GSH has been shown to inhibit neutral SMase activity and ceramide formation (57,58). Since we and others have shown that GSH levels decline with age (59,60), it is feasible that the age-related decreases in GSH levels lead to increased neutral SMase activity, which in turn would result in higher ceramide levels in MØ from old mice.
To determine signaling molecules responsible for the ceramide-induced age-associated increases in COX-2 mRNA, we tested the involvement of MAPKs. Although MAPKs have been indicated in LPS-and ceramide-induced COX-2 expression (30,39,61), our results showed that the increase in LPS-induced ceramide generation leading to its subsequent up-regulation of COX-2 is unlikely to be due to increased activation of ERK, JNK, or p38 activities, because we found no differences in these MAPKs activity between young and old MØ. Although we were not able to differentiate the ERK-1 effects from that of ERK-2 or to determine the contribution of nuclear and cytosolic MAPKs separately, currently there is no evidence that these various types of MAPKs are differentially affected by age. Furthermore, our preliminary results indicate that the ceramide effect might be mediated through NF-B activation. 2 In conclusion, we have shown that increased MØ COX-2 gene transcription rather than increased COX-2 mRNA stability is responsible for the age-related increase in the rate of MØ COX-2 gene expression. We further demonstrated that LPSinduced increases in ceramide, but not its down stream metabolite, sphingosine, mediate the age-associated up-regulation of COX-2 mRNA, leading to increased COX-2 protein, enzyme activity, and PGE 2 production. Findings from this study may serve as useful information toward developing therapeutic interventions to inhibit or delay age-associated dysregulation of the immune and inflammatory responses.