The FASEB Journal • Research Communication
Inhibition of prostaglandin E2 synthesis by SC-560 is independent of cyclooxygenase 1 inhibition Christian Brenneis,* Thorsten J. Maier,* Ronald Schmidt,* Annette Hofacker,* Lars Zulauf,* Per-Johan Jakobsson,† Klaus Scholich,* and Gerd Geisslinger*,1 *Pharmazentrum Frankfurt, ZAFES, Klinikum der Johann Wolfgang Goethe-Universita¨t Frankfurt, Frankfurt, Germany; and †Karolinska Hospital, Stockholm, Sweden Prostaglandin E2 (PGE2) produced by cyclooxygenase-2 (COX-2) and microsomal prostaglandin E2 synthase-1 (mPGES-1) plays an important role in the pathophysiology of inflammation, pain, and fever. We investigated the actions of TNF␣ toward stimulation of PGE2 synthesis in primary spinal cord neurons. TNF␣ induced COX-2 and mPGES-1 expression in neurons, followed by formation of PGE2, which was blocked by a selective COX-2 inhibitor. Surprisingly, the “selective COX-1” inhibitor SC-560 completely inhibited TNF␣-induced PGE2 synthesis in neurons at nanomolar concentrations. Moreover, SC-560 inhibited PGE2 and thromboxane A2 synthesis in human monocytes and platelets with IC50 of 1.8 nM and 2.5 nM, respectively. SC-560 treatment neither altered TNF␣induced COX-2 or mPGES-1 expression nor did the addition of the calcium ionophore A23187 or arachidonic acid reverse the inhibition by SC-560. Moreover, no influence of SC-560 on PGE2 synthase activities or PGE2 transport was seen. Most importantly, SC-560 blocked TNF␣-induced PGE2 synthesis in COX-1-deficient spinal cord neurons, demonstrating a COX-1independent inhibition of PGE2 synthesis. Although SC-560 inhibited LPS-induced PGE2 synthesis in neurons and RAW264.7 macrophages in whole cell assays, no inhibition was observed in lysates of the same cells. Taken together our data demonstrate that SC-560 acts at least in some cell types as an unselective COX inhibitor despite its selectivity toward COX-1 under cell-free conditions.—Brenneis, C., Maier, T. J., Schmidt, R., Hofacker, A., Zulauf, L., Jakobsson, P-J., Scholich, K., Geisslinger, G. Inhibition of prostaglandin E2 synthesis by SC-560 is independent of cyclooxygenase 1 inhibition. FASEB J. 20, 1352–1360 (2006) ABSTRACT
Key Words: PGE2 䡠 cyclooxygenase 䡠 TNF␣
Pge2 regulates key functions in the reproductive, gastrointestinal, neuroendocrine, immune, and central nervous system. PGE2 is synthesized by two cyclooxygenase (COX) isoforms, COX-1 or COX-2, through conversion of arachidonic acid to PGH2 (1, 2). COX-1 is constitutively expressed in most tissues and is believed to produce the PGE2 necessary for homeostasis (3). Although COX-2 is also found to be constitutively 1352
expressed in some tissues (e.g., kidney, spinal cord), COX-2 is found to be dramatically up-regulated in response to inflammatory signals and is therefore believed to play an important role in the PGE2 production involved in pathophysiological processes (4). To dissect the role of the two COX isoforms in the different tissues as well as in physiological and pathophysiological processes selective inhibitors play a dominant role. While several COX-2-selective substances, such as celecoxib, rofecoxib, or valdecoxib, have been developed and are also partly used clinically, the only substance that exhibits a strong COX-1 selectivity is SC-560. This inhibitor is widely used to investigate the roles of COX-1 and -2 and has been reported to inhibit COX-1 activity with an IC50 of 9 nM and COX-2 activity with an IC50 of 6.3 M (5, 6). An important role of COX-2-derived PGE2 is the regulation of inflammatory processes. Pathophysiological and neuropathic pain states are often associated with increased tumor necrosis factor ␣ (TNF␣) concentrations in the spinal cord (7–9). One of the consequences of a stimulation by TNF␣ is an increased PGE2 synthesis and an up-regulation of COX-2 and mPGES-1 expression (10, 11). The PGE2 subsequently released has an important role in the generation of hyperalgesia and allodynia (12, 13). Here, we aimed to investigate the actions of TNF␣ toward stimulation of PGE2 synthesis in primary spinal cord cells to determine which cell types in the spinal cord respond to TNF␣ with an up-regulation of COX-2 and mPGES-1 as well as an increased PGE2 production. The selective COX-1 and COX-2 inhibitors, SC-560 and rofecoxib, respectively, were used to determine the specific roles of the two COX isoforms in the up-regulation of PGE2 synthesis. We found that the selectivity for COX isoforms of SC-560 seems only to apply for cell-free conditions whereas in certain cellular systems SC-560 seems to act as an unselective COX inhibitor.
1 Correspondence: Pharmazentrum Frankfurt, ZAFES, Klinikum der Johann Wolfgang Goethe-Universita¨t Frankfurt, Theodor-Stern-Kai 7, Frankfurt 60590, Germany. E-mail:
[email protected] doi: 10.1096/fj.05-5346com
0892-6638/06/0020-1352 © FASEB
MATERIALS AND METHODS
Immuncytochemical analysis
Tissue preparation and primary cell cultures
Primary cells were plated on poly-l-lysine-coated glass coverslips. Immuncytochemistry was performed as described previously (15). Cells were incubated for 1 h with the primary antibody (Ab), followed by 1 h incubation with FITC-labeled goat anti-rabbit or Cy3-labeled goat antimouse Ab (Sigma). All primary antibodies except the NeuN Ab (Chemicon, Temecula, CA, USA) were purchased from Cayman Chemical Company.
Pregnant Sprague Dawley rats were purchased from Charles River Wiga GmbH (Sulzfeld, Germany). For cell culture, all substances were purchased from Life Technologies, Inc. (Karlsruhe, Germany) unless specified otherwise. Whole spinal cords were prepared from embryos of pregnant rats 16 to 18 days postcoitus and directly transferred to ice-cold HBSS containing CaCl2 and MgCl2. Then the spinal cords were treated with trypsin-EDTA and collagenase (500 U/ml, Biochrom AG, Germany), followed by adjacent mechanical separation using a 1 ml Gilson pipette. To obtain neuronenriched cultures, the cell suspension was plated onto 3.5 cm dishes that had been coated with poly-l-lysine and incubated for 2 h in neurobasal medium containing B-27 supplement (Invitrogen, Carlsbad, CA, USA), 10% fetal calf serum (FCS), 100 U/ml penicillin, 100 g/ml streptomycin, and 2 mM l-glutamine at 37°C and 5% CO2. After the cells became adherent, the medium was replaced and 0.01 g/ml murine nerve growth factor (Invitrogen) was added. After 2 days the same medium without FCS was supplemented with 1 M uridine, 1 M cytosine--d-arabin-furanoside, and 0.1 M 5-fluoro-3-desoxyuridine (Sigma, St. Louis, MO, USA) to inhibit microglia and astrocyte proliferation. For the generation of neuron-free culture the cell suspension was cultivated in RPMI 1640 ⫹ GlutaMAX, 10% FCS, 100 U/ml penicillin, and 100 g/ml streptomycin at 37°C and 5% CO2. The cells were grown for 7 days and passaged twice to eliminate neurons from the culture. Knockout mice Female homozygous (COX-1⫺/⫺) or heterozygous (COX1-/⫹) mice (14) were crossed with homozygous (COX-1⫺/⫺) male mice. Primary neuronal cultures were prepared as described above. Genotypes of embryos from heterozygous females were identified by polymerase chain reaction (PCR) as described previously (14) and the absence of COX-1 protein was confirmed by Western blot analysis. RAW 264.7 cell culture The murine macrophage cell line RAW 264.7 was cultured at 37°C in RPMI 1640 ⫹ GlutaMAX supplemented with 10% FCS, 100 IU/ml penicillin, and 100 g/ml streptomycin. Stimulation of primary cells Stimulation of neuronal cells was done 5 to 6 days after culturing the cells in serum-free medium. 50 g/ml recombinant rat TNF␣ (PeproTech, London, UK) or 5 ng/ml lipopolysaccharide (Sigma) were added. COX inhibitors rofecoxib (COX-2 inhibitor) (Witega, Berlin, Germany) and SC-560 (COX-1 inhibitor) (Sigma or Witega, Berlin, Germany) were given simultaneously with the stimulators. To determine the effects of SC-560 after addition of arachidonic acid or after activation of phospholipases, cells were treated with TNF␣ (50 ng/ml) for 15 h in the absence of SC-560. Then the medium was changed and SC-560 was added for 10 min, followed by induction with 5 M arachidonic acid (Cayman, Ann Arbor, MI, USA) for 30 min or 2 M A23187 (Sigma) for 2 h. For determination of prostaglandins, medium was directly taken and frozen at – 80°C. For Western blots, cells were harvested in boiling SDS sample, heated at 100°C for 10 min, and frozen at – 80°C until later use. INHIBITION OF PGE2 SYNTHESIS BY SC-560
Western blot Cell lysates (30 g protein) were separated on a 15% SDSpolyacrylamide gel. After blotting, COX-1, COX-2, mPGES-1, mPGES-2, and c-PGES proteins were detected with polyclonal antibodies (Cayman). Anti-HSP-70 (BD Biosciences, Heidelberg, Germany) or anti-ERK-1/2 (Promega, Madison, WI, USA) antibodies were used to control for equal loading. Determination of eicosanoid concentrations PGE2 and TXB2 ELISA: PGE2 and TXB2 concentrations in the medium were determined using the enzyme immunoassay (EIA) kit from Assay Designs (Ann Arbor, MI, USA). Thin-layer chromatography lipids were extracted in 1 (metabolic labeling) or 4 (in vitro synthesis) volumes of chloroform containing 1% formic acid, followed by centrifugation at 18,000 g for 5 min. The chloroform phase was collected and evaporated under nitrogen at 45°C. The remaining lipids were resuspended in 50 l chloroform and spotted on an ALUGRAM SIL G/UV silica gel 60 plate (Macherey Nagel, Du¨ren, Germany). Lipids were separated in a thin-layer chromatography (TLC) separation chamber with benzene/ dichlormethane/acetic acid (70/35/1.75) (Roth, Karlsruhe, Germany) and visualized with the PhosphorImager BAS-1500 (Fuji-Film, Vienna, Austria). To identify prostaglandins, spots were compared by their Rf values with pure substances purchased from Cayman. COX-1 and -2 assay with enriched human platelets and monocytes Platelets and monocytes were isolated from human blood as described previously (16, 17) with the following modifications. Tromboxane B2 (TXB2) production by human platelets: to obtain platelet-rich plasma, 20 ml nonheparinized human blood was centrifuged at 300 g for 20 min. The plasma was centrifuged again at 1000 g and the pellet was washed and resuspended in 4 ml RPMI 1640 ⫹ GlutaMAX, including 100 U/ml penicillin and 100 g/ml streptomycin. 7 ⫻ 105 cells were preincubated in a volume of 100 l with various inhibitor concentrations (SC-560, rofecoxib) for 5 min. The reaction was started by addition of arachidonic acid to a final concentration of 0.25 mg/ml and stopped after 10 min with 4 volumes of 1% formic acid in chloroform. Lipids were extracted, then stored at – 80°C until further use. PGE2 production by human monocytes: monocytes were isolated from human blood using the CPT vacutainer system (BD, Heidelberg, Germany). 2 ⫻ 105 cells/well were plated on 24-well dishes and incubated with RPMI 1640 ⫹ GlutaMAX, including 10% FCS, 100 U/ml penicillin, and 100 g/ml streptomycin, for 4 h. Then the medium was replaced by serum-free medium containing LPS and the indicated inhibitor concentrations (SC-560, rofecoxib) and incubated for another 22 h. PGE2 concentrations were determined in the medium as described above. 1353
Whole blood assay To evaluate COX-1-mediated TXB2 production, fresh human blood containing no anticoagulants was mixed with the corresponding inhibitor concentrations and allowed to clot for 1 h at 37°C (18). Samples were then centrifuged with 12,000 g for 5 min and TXB2 concentrations were determined as described above. To determine COX-2 inhibition, 100 g/ml LPS was added to fresh human heparinized blood (18). Then the inhibitors were added at the indicated concentration to 250 l blood and incubated for 24 h at 37°C. Thereafter the blood was centrifuged with 12,000 g for 5 min and PGE2 was determined from the supernatants as described above. Metabolic labeling Cells were incubated with 1 M radioactive arachidonic acid [1-14C] (Moravek Biochemicals, Brea, CA, USA) for 8 h. Then the cells were washed with PBS and stimulated as indicated for 15 h. The medium (1 ml) was extracted with one volume chloroform containing 1% formic acid. After drying the pellets, eicosanoids were resuspended in 50 l chloroform and subjected to TLC. Cell-free prostaglandin synthesis Cells were scraped in 0.1 M Tris-HCl pH 7.5 and homogenized by sonification. Reactions were performed in a final volume of 200 l containing 100 g protein and, where indicated, SC-560 or rofecoxib. The reaction was started with the addition of arachidonic acid [1-14C] (Moravek Biochemicals) to a final concentration of 5 M. After 30 min incubation at 37°C, the reaction was stopped by adding 800 l chloroform containing 1% formic acid (Roth, Karlsruhe, Germany). PGES enzyme activities were determined as described previously (19).
RESULTS To characterize TNF␣-induced PGE2 synthesis in primary embryonic cultures from spinal cords, we monitored the expression levels of PGE2 synthesizing enzymes as well as PGE2 production itself. Stimulation of neuron-enriched cultures with TNF␣ caused an upregulation of COX-2 and mPGES-1. COX-2 expression was up-regulated 5 h after TNF␣ application (Fig. 1A), while up-regulation of mPGES-1 expression started only 10 h after TNF␣ stimulation. The time lag seen between the induction of COX-2 and mPGES-1 is in accordance with a previous observation using stimulated fibroblasts (20). COX-1, cPGES, and mPGES-2 were detected in unstimulated cells and their expression levels were not significantly altered by TNF␣-stimulation (Fig. 1A). Incubation with TNF␣ caused a strong increase in PGE2 synthesis between 10 and 18 h, which is in line with the up-regulation of COX-2 and mPGES-1 (Fig. 1B). To investigate which cell type mainly produces PGE2, neuron-enriched and neuron-free cultures were prepared in parallel as described in Materials and Methods. Immuncytochemical analysis using anti-NeuN Ab showed that the neuron-enriched cultures consisted of 70 – 80% neurons whereas no neurons were found in 1354
Vol. 20
July 2006
the neuron-free cultures (data not shown). TNF␣ increased COX-2 and mPGES-1 expression in the neuronenriched but not in the neuron-free cultures (Fig. 1C). In accordance with these data, only neuron-enriched cultures showed increased PGE2 levels after TNF␣ stimulation (Fig. 1D). Immuncytochemical analysis showed a preferentially non-neuronal expression for COX-1 whereas COX-2, cPGES, mPGES-1 and mPGES-2 seem to be expressed in both neurons and non-neuronal cells (Fig. 1E). So far the data show that TNF␣ induces COX-2 and mPGES-1 expression specifically in neurons, whereas non-neuronal cells do not respond with up-regulation of COX-2/mPGES-1 and an increased PGE2 synthesis. To study in more detail the contribution of the two different COX isoforms in the TNF␣ response, we used the selective COX-1 inhibitor SC-560 and the selective COX-2 inhibitor rofecoxib. As expected, rofecoxib inhibited TNF␣-induced PGE2 synthesis in neurons (Fig. 2A). Surprisingly, SC-560 also inhibited effectively in the nanomolar range TNF␣-induced PGE2 synthesis. To exclude the possibility of impurities in the substances or effects due to breakdown products, we compared the effect of SC-560 obtained from two different companies (Sigma and Witega, Berlin, Germany), which gave identical results. Using LC-MS/MS, we confirmed that a substance with the expected molecular weight was in samples of SC-560 from both sources (data not shown). Additionally the fragmentation patterns of SC-560 samples from both sources were identical. Since according to the manufacturers policies the structure of both substances was verified by NMR spectroscopy and mass spectrometry after synthesis, we can exclude the possibility that the observed effect of SC-560 is a result of a contamination or breakdown of the product. To rule out the possibility that SC-560 was used in a concentration range where it loses its selectivity, we verified the published IC50 values of SC-560 for COX-1 and -2 by determining its effect on thromboxane B2 synthesis of enriched human platelets and PGE2 synthesis of LPS-stimulated human monocytes, respectively. SC-560 reduced COX-1-dependent TXB2 synthesis in human platelets with a similar IC50 value of 2.5 nM as published (5), whereas rofecoxib had only little effect on TXB2 synthesis (Fig. 2B). SC-560 reduced COX-2-dependent PGE2 synthesis in LPS-stimulated human monocytes with a similar IC50 value (1.8 nM) as COX-1-dependent TXB2 synthesis. In contrast, the IC50 value of rofecoxib was with 41 nM in the previously published range (Fig. 2C). Notably, the inhibition of COX-1 by SC-560 and rofecoxib in the “whole blood assay” was similar to values derived in assays with enriched platelets (Fig. 2D). In contrast, the IC50 values for COX-2 inhibition by rofecoxib and SC-560 in the “whole blood assay” were shifted to higher concentrations (IC50 for COX-2: 120 and 128 nM, respectively; Fig. 2E). This shift most likely depends on the protein concentration in the medium and has been described (18, 21, 22). Taken together, the data demonstrate that
The FASEB Journal
BRENNEIS ET AL.
Figure 1. TNF␣ induces PGE2 synthesis in neurons. A) Neuron-enriched cultures were incubated with 50 ng/ml TNF␣ for the indicated times, then subjected to Western blot analysis to detect relative protein amounts. B) Same as panel A except PGE2 concentrations in the medium were determined. Data are shown as the mean ⫾ se and are representatives of 5 experiments. C) Neuron-enriched and neuron-free cultures were incubated with 50 ng/ml TNF␣ for 15 h. Lysates prepared from the cells were subjected to Western blot analysis to detect relative protein amounts. D) Same as panel C except that PGE2 concentrations in the medium were determined. Data are shown as the mean ⫾ se from triplicate determinations. E) Immunofluorescence staining for COX-1, COX-2, cPGES, mPGES-1, mPGES-2 (green), NeuN (red), and merged images of neuronenriched cultures and incubated with 50 ng/ml TNF␣ for 15 h.
SC-560 loses its COX-1 selectivity under cell culturing conditions although some COX-1 selectivity of SC-560 is seen in the “whole blood assay.” Therefore, we investigated the mechanism by which SC-560 may inhibit TNF␣-induced PGE2 synthesis in neurons. Since PGE2 synthesis has been implied in up-regulation of COX-2 (23), we investigated whether SC-560 may prevent the TNF␣-induced up-regulation of COX-2 and mPGES-1. However, Western blot analysis showed no significant differences in the up-regulation of COX-2 and mPGES-1 expression in the absence and presence of SC-560 (Fig. 3A). Then we examined whether SC-560 inhibits other enzymes that are involved in the synthesis of PGE2. First we tested whether SC-560 blocks the TNF␣-induced PGE2 synthesis by decreasing the release of arachidonic acid from the INHIBITION OF PGE2 SYNTHESIS BY SC-560
membrane. Addition of the calcium ionophore A23187 to neurons that were prestimulated with TNF␣ evoked an additional increase in PGE2 synthesis that was completely inhibited by SC-560 and rofecoxib (Fig. 3B). Since endogenously produced arachidonic acid did not reverse inhibition by SC-560, we analyzed whether exogenously added arachidonic acid can rescue TNF␣evoked PGE2 synthesis. Here we determined PGE2 concentrations that were released over 30 min by cells prestimulated with TNF␣. Thus, long-term effects such as changes in the gene expression pattern are excluded. Addition of 0.5 M arachidonic acid resulted in an increased PGE2 release that was more pronounced in TNF␣-prestimulated cells than in unstimulated cells (Fig. 3C). However, again PGE2 synthesis was completely and concentration-dependently inhibited by 1355
Figure 2. Nanomolar concentrations of SC-560 and rofecoxib completely block the TNF␣ induced PGE2 release in primary embryonic spinal cord cells. A) Incubation of neuron-enriched cultures with 50 ng/ml TNF␣ for 15 h. Inhibitors were added in parallel to the stimulation. PGE2 levels were determined from cell medium by EIA kit. Data are shown as the mean ⫾ sd and are representatives from 4 independent experiments. *P ⬍ 0.05 as compared to TNF␣ alone. B) Washed human platelets were preincubated with different inhibitor concentrations for 5 min, then stimulated with 0.25 mg/ml AA for 10 min. TXB2 was extracted and determined as described in Materials and Methods. C) Human monocytes were incubated with 5 g/ml LPS for 22 h and PGE2 was measured directly from medium by EIA kit. Samples were performed in duplicate and data shown as the mean ⫾ sd. D) Human blood was mixed with the corresponding inhibitor concentrations and allowed to clot for 1 h. TXB2 was measured from plasma by EIA kit. Data shown are the mean ⫾ se from 5 independent determinations. E) Human whole blood was incubated with 100 g/ml LPS and COX inhibitors for 24 h. PGE2 was measured directly from plasma by EIA kit. Samples were performed in triplicate and data shown as the mean ⫾ se. IC50 values were derived using the four-parameter logistic fit with the software program SigmaPlot.
nanomolar concentrations of SC-560 (Fig. 3C), ruling out the possibility that a decrease in the availability of arachidonic acid by SC-560 is the primary reason for its 1356
Vol. 20
July 2006
effect on PGE2 synthesis. Similar results were seen with arachidonic acid concentrations ranging between 0.5 and 10 M (data not shown).
The FASEB Journal
BRENNEIS ET AL.
Figure 3. SC-560 does not alter calcium-dependent PLA2 activity, arachidonic acid release, PGES activity or PGE2 transport . A) Western blot analysis of neuron-enriched cultures incubated for 15 h in parallel with 50 ng/ml TNF␣ and 1 M SC-560. B) Neuron-enriched cultures were prestimulated with 50 ng/ml TNF␣ for 15 h. After washing, cells were induced with 2 M A23187 for 2 h and supplemented with 10 nM SC-560 or 30 nM rofecoxib. PGE2 concentrations were measured in the medium. Data are shown as the mean ⫾ sd. C) Same as panel B, except cells were induced with 500 nM arachidonic acid for 30 min instead of A23187. D) PGES activity was determined using microsomal preparations of Il 1 stimulated A 549 cells as described previously (19). Data are shown as the mean ⫾ sd and are representatives from 3 experiments. E) Neuron-enriched cultures were prestimulated with 50 ng/ml TNF␣ for 15 h, followed by 30 min incubation with fresh medium containing 500 nM arachidonic acid and the corresponding SC-560 concentrations. For determination of PGE2, medium was used directly, while PGE2 in cells was first extracted with chloroform. Data shown are the mean ⫾ sd. *P ⬍ 0.05 as compared to TNF␣ and A23187 (B) or TNF␣ and AA (C), or TNF␣ alone (E).
Next, we tested whether SC-560 is able to inhibit the enzymatic activity of mPGES-1 using vesicles isolated from A549 cells. These vesicles contain mainly mPGES-1 and also small amounts of cPGES and mPGES-2 (19). SC560 at concentrations up to 100 M had no significant effect on PGES enzyme activity (Fig. 3D). In contrast, 5 M of the unspecific mPGES-1 inhibitor MK886 reduced PGES activity by ⬎80%. Finally, we investigated whether SC-560 inhibits the PGE2 transport out of neurons by comparing intra- and extracellular PGE2 concentrations in SC-560-treated neurons. A block of the PGE2 transport would result in increased intracellular PGE2 concentrations as described previously (24). However, SC-560 decreased intra- and extracellular PGE2 concentrations with similar IC50 values (Fig. 3E) without resulting in PGE2 accumulation in the cells. Hence, so far our data show no influence of SC-560 on up-regulation of COX-2 or INHIBITION OF PGE2 SYNTHESIS BY SC-560
mPGES-1, the release of arachidonic acid from the membrane, PGE2 synthase activity, or the transport of PGE2 out of neurons. Therefore, it seems likely that the inhibitory effect of SC-560 on the TNF␣-stimulated PGE2 release from spinal cord neurons is based on the inhibition of not only COX-1, but mainly COX-2. To examine whether COX-1 is the primary target of SC-560, we compared the effect of SC-560 on TNF␣stimulated embryonic spinal cord neurons from wildtype (WT) mice and COX-1-deficient mice. The genotypes of all embryos were controlled by PCR (14), and the neurons were tested after the experiments for COX-1 expression by Western blot. As shown in Fig. 4, no difference in the TNF␣-induced PGE2 synthesis was seen between WT and COX-1-deficient mice, suggesting again that the increase of PGE2 is COX-2 mediated. Most important, 10 nM SC-560 completely inhibited the TNF␣-induced PGE2 synthesis independent of the 1357
rofecoxib, respectively (6). Most important, SC-560 had little effect on the prostaglandin synthesis whereas a strong inhibition by rofecoxib was seen (Fig. 5D) showing that relative low concentrations of SC560 are not affecting COX-2 under in vitro conditions. Thus, our findings suggest that SC-560 is a selective COX-1 inhibitor under cell-free conditions but acts as an unselective COX inhibitor in the cellular environment.
Figure 4. SC-560 inhibits TNF␣-induced PGE2 synthesis in COX-1-deficient spinal cord neurons. Neuron-enriched cultures from embryos of mice with different genetic backgrounds were incubated with 50 ng/ml TNF␣ and 10 nM SC-560 for 15 h. PGE2 concentrations in the medium were determined. Data are shown as the mean ⫾ se representing at least 20 embryos. For each genotype, cells were dissected from embryos of at least 3 different mothers. *P ⬍ 0.01 as compared to TNF␣ alone.
genotype (Fig. 4). Thus this finding rules out COX-1 inhibition as the primary reason for SC-560-mediated inhibition of TNF␣-stimulated PGE2 synthesis. Published data show a selective COX-1 inhibition by SC-560 under cell-free conditions (5, 6) that is in conflict with the observed COX-1 independent effect we found in the TNF␣-stimulated neurons and in the blood assays. Therefore, we compared the specificity of SC-560 in cells and in a cell-free assay using cell lysates from the same cells. First we tested whether SC-560 also inhibits LPS-stimulated PGE2 synthesis in the spinal cord neurons as well as in RAW264.7 macrophages. Metabolic labeling of both cell lines showed that rofecoxib and low concentrations of SC-560 both inhibited LPS-induced PGE2 synthesis. Notably, no shift in the prostaglandin synthesis from PGE2 to other prostaglandins was observed (Fig. 5A–C) as would be expected after an inhibition of PGE2 synthases (25). After that we generated lysates from LPS-stimulated RAW264.7 macrophages that express high amounts of COX-2 and relative low amounts of COX-1 (data not shown). To visualize only prostaglandins produced in vitro, the reaction was started by addition of radioactive arachidonic acid and visualized using TLC. The inhibitor concentrations used in this assay were around the published IC90 concentrations to allow specific but complete inhibition of COX-1 and -2 by SC-560 and
DISCUSSION In an approach using immunocytochemistry and cellselective culturing conditions, we showed that TNF␣ induces COX-2 and mPGES-1 expression as well as PGE2 synthesis in spinal cord neurons. Using a pharmacological approach to investigate the roles of COX-1 and -2 in the TNF␣-evoked PGE2 response, we were surprised to find that besides the selective COX-2 inhibitor rofecoxib, SC-560 also inhibited the TNF␣evoked PGE2 release at nanomolar concentrations. Furthermore, in assays using enriched platelets or monocytes, SC-560 inhibited COX-1 and -2 with similar potencies (IC50 of 2.5 nM and 1.8 nM, respectively) while the COX-2 selective inhibitor rofecoxib exhibited under the same assay conditions a strong COX-2 selectivity. In a “whole blood assay” system, SC-560 showed some COX-1 selectivity (IC50 ratio COX-2/COX-1: 15). The selectivity of rofecoxib as a COX-2 inhibitor was, however, significantly greater than that of SC-560 as a COX-1 inhibitor (IC50 ratio COX-1/COX-2: 83). As in assays using enriched blood cells, SC-560 inhibited COX-2 in the “whole blood assay” with a potency equal to that of rofecoxib (128 nM vs. 120 nM, respectively). Thus, it seems that COX inhibition and COX selectivity of SC-560 strongly depend on the conditions in the medium, as described by Warner et al. (21), as well as the cellular environment as demonstrated by our experiments that compare COX inhibition by SC-560 in cultured RAW macrophages and lysates of the same cells. Several possible mechanisms could underlie the missing COX-1 selectivity of SC-560 under the cell culturing conditions. For once it has been reported that COX-1 activity can be necessary for the induction of the up-regulation of COX-2 or mPGES-1 (20, 23). However, this possibility can be ruled out since SC-560 did not
Figure 5. Effect of SC-560 on prostaglandin synthesis in vitro and in cells. A) Neuronenriched cells were metabolically labeled as described in Materials and Methods. Cells were stimulated with 5 g/ml LPS and the indicated SC-560 concentrations for 15 h. Prostaglandins were separated using TLC as described in Materials and Methods. B, C) Same as in panel A except that RAW macrophages were used instead of neuronal cells. D) RAW macrophages were prestimulated with 5 g/ml LPS for 15 h, then lysed. Prostaglandin synthesis was done in the presence of the indicated inhibitor concentrations as described in Materials and Methods. 1358
Vol. 20
July 2006
The FASEB Journal
BRENNEIS ET AL.
alter TNF␣-induced up-regulation of COX-2 and mPGES-1 in rat neurons, and TNF␣-evoked PGE2 release in COX-1-deficient neurons was still efficiently blocked by SC-560. Especially the sensitivity of COX-1deficient neurons for SC-560 clearly demonstrates a COX-1-independent inhibition of PGE2 production by SC-560. Other possibilities as to how SC-560 may inhibit PGE2 synthesis or PGE2 release could be an inhibition of the release of arachidonic acid from the membrane, the conversion of PGH2 to PGE2 by PGE2 synthases, or the release of PGE2 from the cells. However, increasing the intracellular calcium concentration to activate phospholipases or direct addition of arachidonic acid to the neurons did not reverse the inhibition by SC-560 suggesting that the target of SC-560 is not upstream of the COX isoforms. SC-560 did not inhibit enzymes downstream of COX in the PGE2 synthesis cascade, since no inhibition of the enzyme activity of the three known PGE2 synthases in the presence of SC-560 and no shift in the prostaglandin production, as would be expected after the inhibition of one or more PGE2 synthases (25), was observed. Nor could we confirm an inhibition of the PGE2 transport out of the cell by SC-560. Finally, since the inhibitory actions of SC-560 can be seen within 30 min, changes in the transcriptional regulation of other genes seem highly unlikely. Thus, our data suggest that SC-560 acts as an unselective COX inhibitor. Indeed, recently Rouzer et al. speculated that SC-560 might act as an unselective COX inhibitor after observing that SC-560 inhibited LPSinduced PGE2 and 6-keto-PGF1a synthesis in murine peritoneal macrophages (26). Since an unselective COX inhibition by SC-560 is in conflict to published data demonstrating a strong COX-1 selectivity for SC560 (IC50 value of ca. 9 nM for COX-1 and 6.3 M for COX-2) (5), we compared the actions of SC-560 in cells with its activity in an cell-free assay using lysates. Here, SC-560 completely blocked the PGE2 production in RAW264.7 macrophages used as control but did not inhibit prostaglandin synthesis in an in vitro assay using lysates of the same cells. In conclusion, our data suggest that the selectivity of SC-560 for COX-1 and -2 changes with the reaction conditions and that the strong COX-1 selectivity seen under cell-free conditions is lost when SC-560 is applied under cell culture conditions to primary neurons and macrophages. A similar increase in the potency to inhibit COX-2 activity has been described for cell-free and cellular conditions with paracetamol (27). The underlying mechanism that leads to the increased potency of SC-560 toward COX-2 in cells is unclear. A mechanism for COX-2 inhibition by diarylheterocyclic derivatives has been described, consisting of three sequential reversible steps (6). The first two kinetically distinct steps describe the binding of the inhibitors to COX-2, and were similar for SC-560 and several selective COX-2 inhibitors such as valdecoxib or celecoxib. However, diarylheterocycles are able to form a tightly bound enzyme inhibitor complex in the active INHIBITION OF PGE2 SYNTHESIS BY SC-560
site and a side pocket of COX-2 due to their phenylsulfonamide or a phenylsulphone group. SC-560 cannot form this tight complex due to the absence of a phenylsulfonamide or a phenylsulphone group, and is easily released from the binding pocket of COX-2. However, the binding pocket is partly formed by a loop consisting of a series of alpha helices that are part of the membrane binding domain (MBD) of COX-2 and may regulate the accessibility of the binding pocket. The very narrow dimensions of the aperture within the COX channel suggest that the MBD may undergo significant conformational changes during substrate entry and product exit (28, 29). The flexibility of the MBD may be much less pronounced in a cellular environment as compared to cell-free conditions and may cause a decrease in the release rate for SC-560. Thus, since COX-2-selective inhibitors and SC-560 have similar binding kinetics (6), a decreased release rate of SC-560 from the binding pocket of COX-2 would result in a higher inhibitory potency of SC-560 toward COX-2. Taken together, our findings raise the question of whether SC-560 is a useful tool to differentiate the roles of COX-1 and -2 in cellular systems or animal models. Since it is not clear whether SC-560 loses its selectivity for COX-1 generally in a cellular environment or only in specific cell types (here: rat spinal cord neurons, murine RAW 264.7, human platelets and monocytes), a careful analysis regarding its selectivity is necessary in the various cell and animal models. Especially in studies where the effects of SC-560 were compared with selective COX-2 inhibitors but not with nonselective COX inhibitors, it may be difficult to distinguish between selective COX-1 inhibition and unselective COX inhibition by SC-560. The work was supported by DFG grant GE 695/2–1. We thank Prof. G. Schneider for helpful discussion and Prof. R. Nu¨sing for supplying the COX-1 knockout mice.
REFERENCES 1. 2. 3.
4.
5.
6.
Ivanov, A. I., and Romanovsky, A. A. (2004) Prostaglandin E2 as a mediator of fever: synthesis and catabolism. Front. Biosci. 9, 1977–1993 Murakami, M., Nakatani, Y., Tanioka, T., and Kudo, I. (2002) Prostaglandin E synthase. Prostaglandins Other Lipid Mediat. 6869, 383–399 Tanioka, T., Nakatani, Y., Kobayashi, T., Tsujimoto, M., Oh-ishi, S., Murakami, M., and Kudo, I. (2003) Regulation of cytosolic prostaglandin E2 synthase by 90-kDa heat shock protein. Biochem. Biophys. Res. Commun. 303, 1018 –1023 Trebino, C. E., Stock, J. L., Gibbons, C. P., Naiman, B. M., Wachtmann, T. S., Umland, J. P., Pandher, K., Lapointe, J. M., Saha, S., Roach, M. L., Carter, D., Thomas, N. A., Durtschi, B. A., McNeish, J. D., Hambor, J. E., Jakobsson, P. J., Carty, T. J., Perez, J. R., and Audoly, L. P. (2003) Impaired inflammatory and pain responses in mice lacking an inducible prostaglandin E synthase. Proc. Natl. Acad. Sci. USA 100, 9044 –9049 Smith, C. J., Zhang, Y., Koboldt, C. M., Muhammad, J., Zweifel, B. S., Shaffer, A., Talley, J. J., Masferrer, J. L., Seibert, K., and Isakson, P. C. (1998) Pharmacological analysis of cyclooxygenase-1 in inflammation. Proc. Natl. Acad. Sci. USA 95, 13313–13318 Walker, M. C., Kurumbail, R. G., Kiefer, J. R., Moreland, K. T., Koboldt, C. M., Isakson, P. C., Seibert, K., and Gierse, J. K. (2001) A three-step kinetic mechanism for selective inhibition
1359
7.
8.
9. 10.
11.
12.
13. 14.
15.
16. 17.
18.
1360
of cyclo-oxygenase-2 by diarylheterocyclic inhibitors. Biochem. J. 357, 709 –718 DeLeo, J. A., Rutkowski, M. D., Stalder, A. K., and Campbell, I. L. (2000) Transgenic expression of TNF by astrocytes increases mechanical allodynia in a mouse neuropathy model. NeuroReport 11, 599 – 602 Schafers, M., Svensson, C. I., Sommer, C., and Sorkin, L. S. (2003) Tumor necrosis factor-alpha induces mechanical allodynia after spinal nerve ligation by activation of p38 MAPK in primary sensory neurons. J. Neurosci. 23, 2517–2521 Watkins, L. R., Milligan, E. D., and Maier, S. F. (2003) Glial proinflammatory cytokines mediate exaggerated pain states: implications for clinical pain. Adv. Exp. Med. Biol. 521, 1–21 Fournier, T., Fadok, V., and Henson, P. M. (1997) Tumor necrosis factor-alpha inversely regulates prostaglandin D2 and prostaglandin E2 production in murine macrophages. Synergistic action of cyclic AMP on cyclooxygenase-2 expression and prostaglandin E2 synthesis. J. Biol. Chem. 272, 31065–31072 Kojima, F., Naraba, H., Miyamoto, S., Beppu, M., Aoki, H., and Kawai, S. (2004) Membrane-associated prostaglandin E synthase-1 is up-regulated by proinflammatory cytokines in chondrocytes from patients with osteoarthritis. Arthritis Res. Ther. 6, R355–365 Samad, T. A., Moore, K. A., Sapirstein, A., Billet, S., Allchorne, A., Poole, S., Bonventre, J. V., and Woolf, C. J. (2001) Interleukin-1beta-mediated induction of Cox-2 in the CNS contributes to inflammatory pain hypersensitivity. Nature 410, 471– 475 Vanegas, H., and Schaible, H. G. (2001) Prostaglandins and cyclooxygenases in the spinal cord. Prog. Neurobiol. 64, 327–363 Langenbach, R., Morham, S. G., Tiano, H. F., Loftin, C. D., Ghanayem, B. I., Chulada, P. C., Mahler, J. F., Lee, C. A., Goulding, E. H., Kluckman, K. D., et al. (1995) Prostaglandin synthase 1 gene disruption in mice reduces arachidonic acidinduced inflammation and indomethacin-induced gastric ulceration. Cell 83, 483– 492 Ehnert, C., Tegeder, I., Pierre, S., Birod, K., Nguyen, H. V., Schmidtko, A., Geisslinger, G., and Scholich, K. (2004) Protein associated with Myc (PAM) is involved in spinal nociceptive processing. J. Neurochem. 88, 948 –957 Demasi, M., Caughey, G. E., James, M. J., and Cleland, L. G. (2000) Assay of cyclooxygenase-1 and 2 in human monocytes. Inflamm. Res. 49, 737–743 Riendeau, D., Percival, M. D., Brideau, C., Charleson, S., Dube, D., Ethier, D., Falgueyret, J. P., Friesen, R. W., Gordon, R., Greig, G., Guay, J., Mancini, J., Ouellet, M., Wong, E., Xu, L., Boyce, S., Visco, D., Girard, Y., Prasit, P., Zamboni, R., Rodger, I. W., Gresser, M., Ford-Hutchinson, A. W., Young, R. N., and Chan, C. C. (2001) Etoricoxib (MK-0663): preclinical profile and comparison with other agents that selectively inhibit cyclooxygenase-2. J. Pharmacol. Exp. Ther. 296, 558 –566 Gierse, J. K., Zhang, Y., Hood, W. F., Walker, M. C., Trigg, J. S., Maziasz, T. J., Koboldt, C. M., Muhammad, J. L., Zweifel, B. S., Masferrer, J. L., Isakson, P. C., and Seibert, K. (2005) Valde-
Vol. 20
July 2006
19.
20.
21.
22.
23.
24.
25.
26.
27. 28.
29.
coxib: assessment of cyclooxygenase-2 potency and selectivity. J. Pharmacol. Exp. Ther. 312, 1206 –1212 Thoren, S., and Jakobsson, P. J. (2000) Coordinate up- and down-regulation of glutathione-dependent prostaglandin E synthase and cyclooxygenase-2 in A549 cells. Inhibition by NS-398 and leukotriene C4. Eur. J. Biochem. 267, 6428 – 6434 Kojima, F., Naraba, H., Sasaki, Y., Beppu, M., Aoki, H., and Kawai, S. (2003) Prostaglandin E2 is an enhancer of interleukin1beta-induced expression of membrane-associated prostaglandin E synthase in rheumatoid synovial fibroblasts. Arthritis Rheum. 48, 2819 –2828 Warner, T. D., Vojnovic, I., Bishop-Bailey, D., and Mitchell, J. A. (2006) Influence of plasma protein on the potencies of inhibitors of cyclooxygenase-1 and -2. FASEB J. 20, 542–544, Epub 2006 Jan 10 Palomer, A., Cabre, F., Pascual, J., Campos, J., Trujillo, M. A., Entrena, A., Gallo, M. A., Garcia, L., Mauleon, D., and Espinosa, A. (2002) Identification of novel cyclooxygenase-2 selective inhibitors using pharmacophore models. J. Med. Chem. 45, 1402–1411 Hinz, B., Brune, K., and Pahl, A. (2000) Prostaglandin E(2) up-regulates cyclooxygenase-2 expression in lipopolysaccharidestimulated RAW 264.7 macrophages. Biochem. Biophys. Res. Commun. 272, 744 –748 Reid, G., Wielinga, P., Zelcer, N., van der Heijden, I., Kuil, A., de Haas, M., Wijnholds, J., and Borst, P. (2003) The human multidrug resistance protein MRP4 functions as a prostaglandin efflux transporter and is inhibited by nonsteroidal antiinflammatory drugs. Proc. Natl. Acad. Sci. USA 100, 9244 –9249 Trebino, C. E., Eskra, J. D., Wachtmann, T. S., Perez, J. R., Carty, T. J., and Audoly, L. P. (2005) Redirection of eicosanoid metabolism in mPGES-1-deficient macrophages. J. Biol. Chem. 280, 16579 –16585, Epub 2005 Feb 18 Rouzer, C. A., Kingsley, P. J., Wang, H., Zhang, H., Morrow, J. D., Dey, S. K., and Marnett, L. J. (2004) Cyclooxygenase-1dependent prostaglandin synthesis modulates tumor necrosis factor-alpha secretion in lipopolysaccharide-challenged murine resident peritoneal macrophages. J. Biol. Chem. 279, 34256 – 34268 Graham, G. G., and Scott, K. F. (2005) Mechanism of action of paracetamol. Am. J. Ther. 12, 46 –55 Loll, P. J., Picot, D., Ekabo, O., and Garavito, R. M. (1996) Synthesis and use of iodinated nonsteroidal antiinflammatory drug analogs as crystallographic probes of the prostaglandin H2 synthase cyclooxygenase active site. Biochemistry 35, 7330 – 7340 Garavito, R. M., Malkowski, M. G., and DeWitt, D. L. (2002) The structures of prostaglandin endoperoxide H synthases-1 and -2. Prostaglandins Other Lipid Mediat. 68 – 69, 129 –152
The FASEB Journal
Received for publication November 7, 2005. Accepted for publication March 3, 2006.
BRENNEIS ET AL.
