Trypan blue, Dulbecco's Modified Eagle's Medium (DMEM), fetal bovine serum (FBS), streptomycin, penicillin, and trypsin were purchased from Gibco BRL (Grand Island, NY, USA). COX-2, NF-κB p65 antibody, anti-mouse IgG, and anti-rabbit IgG horseradish peroxidase antibody were from Cell Signaling (Boston, MA, USA). The HO-1 antibody, PKA inhibitor KT5720, copper protoporphyrin (CoPP), and zinc protoporphyrin (ZnPP) were purchased from Enzo Life Sciences (Victoria, BC, USA). The β-actin antibody was from Abcam (San Francisco, CA, USA). Dibutyryl cAMP (dbcAMP: N⁶,2'-O-dibutyryladenosine 3':5'-cAMP) was from Calbiochem (San Diego, CA, USA). The RII/AKAP disruptor peptide Ht31 was obtained from Promega (Madison, WI, USA). The RI/AKAP disruptor peptide RIAD was purchased from Anaspec (San Jose, CA, USA). PGE₂, indomethacin, Escherichia coli (055:B5) LPS, sucrose, aprotinin, phenylmethylsulfonyl fluoride (PMSF), dithiothreitol (DTT), sodium dodecyl sulfate (SDS), barium chloride (BaCl), and benzene were obtained from Sigma (St. Louis, MO, USA). All other chemicals utilized in this study were of highest purity available from commercial sources.

The murine macrophage cell line RAW 264.7 was obtained from American Type Culture Collection (ATCC). Cells were cultured in DMEM, supplemented with 100 IU/mL penicillin, 100 µg/mL streptomycin, and 10% heat-inactivated FBS, at 37℃ and 5% CO₂ in a humidified chamber. The medium was changed at 3 days. The cells were allowed to reach confluence, and subcultured using trypsin.

RAW 264.7 macrophages were washed with cold PBS (pH 7.4) and harvested. The cells were homogenized in nine volumes of 0.3 M sucrose, 0.26 unit/mo aprotinin, 0.1 mM PMSF, and 1 mM DTT, using a Teflon-potter homogenizer with 10 strokes at 900-1,000 rpm. The protein samples (30 µg) were boiled at 100℃ for 10 min, and subjected to 10% SDS-polyacrylamide gel electrophoresis (SDS-PAGE). Protein content was measured using the Bradford method (Bio-Rad protein assay). The separated proteins were transferred to a polyvinylidene difluoride (PVDF) membrane (Immunobilon-P: Millipore, Bedford, MA, USA). After transfer, membranes were blocked with a solution of 5% non-fat dried milk in Tris-buffered saline (TBS) containing 10 mM Tris (pH 8.0) and 150 mM NaCl for 1 h and incubated with the respective primary antibodies (COX-2, 1:1,000; HO-1, 1:1,000; NF-κB p65, 1:1,000; β-actin, 1:10,000) overnight at 4℃. After washing three times with TBS containing 0.05% Tween 20 (TBST), the membranes were hybridized with horseradish peroxidase-conjugated secondary antibodies (1:1,000) for 1 h at room temperature and developed with an enhanced chemiluminescence detection kit (Amersham Pharmacia Biotech. Buckinghamshire, UK). Protein bands were visualized on X-ray film. Relative band densities were determined by densitometric analysis, and the ratio of COX-2 and HO-1 to β-actin was calculated. In all instances, density values of bands were corrected by subtracting the background values.

RAW 264.7 macrophages (3×10⁶ cells/plate) were plated in 100 mm plates. Then, cells were treated with 1 µg/mL LPS in the presence or absence of a tested compound for 1 h, and washed with ice-cold PBS buffer and kept on ice for 1 min. The suspension was mixed with buffer A (10 mM HEPES, pH 7.5, 10 mM KCl, 0.1 mM EGTA, 0.1 mM EDTA, 1 mM DTT, 0.5 mM PMSF, 5 mg/mL aprotinin, 5 mg/mL pepstatin, and 10 mg/mL leupeptin) and lysed via three freeze-thaw cycles. Cytosolic fractions were obtained by centrifugation at 12,000×g for 20 min at 4℃. The pellets were resuspended in buffer B (20 mM HEPES, pH 7.5, 0.4 M NaCl, 1 mM EGTA, 1 mM EDTA, 1 mM DTT, 1 mM PMSF, 5 mg/mL aprotinin, 5 mg/mL pepstatin, and 10 mg/mL leupeptin), kept on ice for 40 min, and centrifuged at 14,000×g for 20 min at 4℃. Levels of NF-κB p65 subunit in the nuclear protein extract were determined by Western blotting analysis as described above.

The activity of HO was measured spectrophotometrically by quantifying the release of bilirubin into the culture media. RAW 264.7 macrophages (2×10⁶ cells/plate) were incubated for varying periods of time, and 0.5 mL each culture supernatant was collected. Then, 250 mg BaCl and 0.75 mL benzene were added and the samples were vortexed for 50-60 sec and centrifuged (30 min at 13,000×g). The upper benzene layers were collected and the absorbance at 450 nm with a reference wave-length at 600 nm was measured with a spectrophotometer (excitation coefficient of 27.3 mM^-1·cm^-1).

RAW 264.7 cells on cover slips were fixed with 4% paraformaldehyde solution in PBS (pH 8.4; 15 min, room temperature) and permeabilized using 0.1% Triton X-100 solution in PBS (30 min, room temperature). The non-specific binding sites were blocked with 4% bovine serum albumin (BSA) in PBS for 30 min at room temperature and the NF-κB p65 subunit was detected by incubation with rabbit polyclonal antibody (1:50 dilution) for 1 h. Cells were washed three times and incubated with FITC-conjugated anti-donkey-IgG secondary antibody (1:100; Invitrogen) for 1 h. Moreover, all cells were stained with a proper dilution of 4',6'-diamidino-2-phenylindole (DAPI) for 2 min to counter-stain DNA in the nuclei. After extensive washings with PBS, coverslips were mounted onto the glass slides, and cells were photographed with fluorescence microscopy to determine NF-κB translocation. To analyze NF-κB activity, the regions of interest (ROI) were drawn around the cell and nucleus. Mean pixel intensity for each ROI was obtained using image software (i-Solution Lite, iMTechnology, Korea). The following ratio was calculated for each cell to indicate the relative levels of antibody staining in the cell and nucleus: ratio=(nuclear intensity-background intensity)/(cell intensity-background intensity).

Data is expressed as means±SD of at least 3 independent experiments. Statistical analysis was performed using SigmaPlot software version 9.0 (SigmaPlot Software Inc., San Diego, CA, USA). All results were analyzed by one-way ANOVA followed by a multiple comparison test, Scheffe's test. A P value of <0.05 was considered statistically significant.

To determine the role of endogenous prostanoids in LPS-induced COX-2 expression in RAW 264.7 macrophages, cells were challenged with LPS (1 µg/mL) in the absence or presence of the nonselective COX inhibitor indomethacin (5 µM). Indomethacin significantly reduced the LPS-induced expression of COX-2 protein (Fig. 1A), implicating that an endogenous prostanoid potentiates COX-2 expression. LPS-induced COX-2 expression was enhanced by pretreatment of macrophages with PGE₂, which partially restored the inhibitory effects of indomethacin (Fig. 1A). This suggests that PGE₂ involves the expression of COX-2 enzyme in response to LPS by modulating intracellular signaling events.

Diaz-Munoz et al.9 reported that EP2 signaling is required for the upregulation of COX-2 in activated RAW 264.7 macrophages. Because EP2 generally increases intracellular cAMP, we investigated whether the elevation of the second messenger could augment COX-2 expression. The cAMP analog dibutyryl AMP (1 µM) enhanced LPS-induced COX-2 expression in a manner similar to PGE₂ itself (Fig. 1B). In macrophages, cAMP works mainly by activating PKA.5 To delineate whether PKA involves TLR4-induced COX-2 expression, we first treated cells with LPS in the presence or absence of a PKA-specific inhibitor (KT5720, 1 µM) and then added PGE₂ to confirm that PKA is indeed responsible for the potentiation of COX-2 expression. As shown in Fig. 1C, the PKA inhibitor attenuated LPS-induced COX-2 expression. Indeed, KT5720 also blocked the potentiating effects of exogenous PGE₂. Taken together, these results suggest that cAMP-dependent PKA is responsible for the amplification of COX-2 expression by TLR4. PKA is a holoenzyme that consists of two regulatory (R) subunits, RI and RII, and two catalytic subunits. The R subunits of PKA are important for localization inside the cell. AKAPs bind to both regulatory subunits and serve as scaffold proteins to target PKA for specific microdomains. To determine whether AKAPs are involved in the PKA-dependent enhancement of LPS-induced COX-2 expression, we disrupted the potential PKA RI-AKAP interaction (RIAD peptide) and the PKA RII-AKAP interaction (Ht31 peptide). Cells were treated with PKA RI-specific AKAP disruptor peptide (RIAD, 25 µM) and PKA RII-specific AKAP disruptor peptide (Ht31, 25 µM). Both disruptors attenuated LPS-induced COX-2 expression, although RIAD had a relatively stronger effect on LPS-induced COX-2 expression (Fig. 1D). No additive effect was observed when the two disruptors were used together (data not shown). These data suggest that the enhanced expression of LPS-induced COX-2 by endogenous cAMP signaling depends on AKAPs and the presence of multiple signaling pathways after the interaction between AKAPs and both PKAs.

In response to stress such as hypoxia, cytokines, and infection, HO-1 is induced to have an anti-oxidant and anti-inflammatory function. HO-1 regulates cellular contents with suppression of LPS-induced inflammation.13 To determine the effects of HO-1 on LPS-induced COX-2 expression in RAW 264.7 macrophages, we examined the effect of the HO inhibitor ZnPP (20 µM) or the HO activator CoPP (10 µM) at 6 or 24 h post-treatment. As shown in Fig. 2, enhancement of HO-1 expression by CoPP treatment suppressed LPS-induced COX-2 expression at both time points. In contrast, inhibition of HO-1 expression by the HO inhibitor ZnPP augmented LPS-induced COX-2 expression. Notably, COX-2 expression was not influenced by HO-1 expression level in the absence of LPS. This suggests that induction of HO-1 can inhibit the intracellular signaling events associated with cellular expression of COX-2 enzyme in response to LPS.

HO-1 is induced by TLR4 ligation.18 Because induction of HO-1 inhibits LPS-induced COX-2 expression, we hypothesized that PGE₂-PKA-AKAP signaling could suppress LPS-induced HO-1 expression and cause positive feedback augmentation of COX-2 expression. As shown in Fig. 3, the HO-1 activator CoPP significantly reduced LPS-induced COX-2 expression, and exogenous PGE₂ restored CoPP-inhibited responses to LPS at 6 h. Exogenous PGE₂ attenuated LPS-induced HO-1 expression and modestly blocked the potentiating effects of exogenous CoPP on LPS-induced HO-1. This suggests that suppression of LPS-induced HO-1 expression by the addition of PGE₂ is responsible for the amplification of COX-2 in response to ligation of TLR4. To delineate the role of endogenous PGE₂-PKA-AKAP signaling on LPS-induced HO-1 expression, we challenged cells with LPS in the presence or absence of a PKA inhibitor (KT5720) or AKAP disruptors (Ht31 or RIAD). The PKA inhibitor enhanced LPS-induced HO-1 expression, and these potentiating effects were blocked by exogenous PGE₂ (Fig. 4A). To determine whether AKAPs are involved in the PKA-dependent suppression of LPS-induced HO-1, we disrupted the potential PKA RI-AKAP interactions with RIAD peptide and RII-AKAP interactions with Ht31 peptide. Both AKAP disruptors augmented LPS-induced HO-1 expression to an extent similar to that observed with the PKA inhibitor KT5720 (Fig. 4B). To confirm that PKA is indeed responsible for blocking HO-1 expression, we measured bilirubin generated from HO activity. In the presence of KT5720, bilirubin increased markedly, and this effect was restored by exogenous PGE₂ (Fig. 5). Taken together, these data clearly demonstrate that suppression of HO-1 expression via PGE₂-PKA-AKAP signaling potentiates LPS-induced COX2 expression in RAW 264.7 macrophages.

TLR4 responses are mainly mediated via NF-κB activation.19,20 To evaluate whether induction of HO-1 involves LPS-induced NF-κB activation, we stimulated cells for 1 h with LPS, then observed NF-κB p65 expression in the nuclear fraction via Western blotting. As shown in Fig. 6A, NF-κB p65 proteins translocated to the nucleus only with LPS treatment, and PGE₂ augmented the p65 activation via TLR4 ligation. In addition, the HO-1 activator CoPP suppressed NF-κB p65 expression, and exogenous PGE₂ partially restored its effect. These findings were confirmed by assessing the nuclear translocation of NF-κB p65 subunit by immunofluorescence microscopy and by quantifying the nuclear colocalization of p65 with DAPI. LPS enhanced p65 nuclear translocation in macrophages; however, LPS-induced p65 translocation was attenuated with the HO-1 activator CoPP (Fig. 6B and C). Taken together, these results demonstrate that the induction of HO-1 has an inhibitory effect on LPS-induced COX-2 expression through the attenuation of LPS-induced p65 translocation.