Ceramide Kinase Mediates Cytokine- and Calcium Ionophore-induced Arachidonic Acid Release*

Despite the importance of prostaglandins, little is known about the regulation of prostanoid synthesis proximal to the activation of cytosolic phospholipase A2, the initial rate-limiting step. In this study, ceramide-1-phosphate (C-1-P) was shown to be a specific and potent inducer of arachidonic acid (AA) and prostanoid synthesis in cells. This study also demonstrates that two well established activators of AA release and prostanoid synthesis, the cytokine, interleukin-1β (IL-1β), and the calcium ionophore, A23187, induce an increase in C-1-P levels within the relevant time-frame of AA release. Furthermore, the enzyme responsible for the production of C-1-P in mammalian cells, ceramide kinase, was activated in response to IL-1β and A23187. RNA interference targeted to ceramide kinase specifically down-regulated ceramide kinase mRNA and activity with a concomitant decrease of AA release in response to IL-1β and A23187. Down-regulation of ceramide kinase had no effect on AA release induced by exogenous C-1-P. Collectively, these results indicate that ceramide kinase, via the formation of C-1-P, is an upstream modulator of phospholipase A2 activation. This study identifies previously unknown roles for ceramide kinase and its product, C-1-P, in AA release and production of eicosanoids and provides clues for potential new targets to block inflammatory responses.

Eicosanoids are synthesized de novo from arachidonic acid (AA) 1 released from membranes in response to agonists such as cytokines and growth factors. The production of AA is the initial rate-limiting step in eicosanoid biosynthesis; thus, the regulation of phospholipases, specifically phospholipase A 2 , is of key importance in this pathway (1,2). At the Golgi apparatus, endoplasmic reticulum, and nuclear membrane, PLA 2 hydrolyzes membrane phospholipids to produce AA, beginning eicosanoid biosynthesis (1,2). The AA produced is utilized by either lipoxygenases to produce leukotrienes or by cyclooxygenases (COXs) to produce prostanoids (1,2).
The COX-2 pathway of prostanoid synthesis has been established as an important therapeutic target for the treatment of inflammatory disorders (3). More recently, it has been found that inhibitors of COX-2 reduce the growth and metastasis capabilities of some cancers. Specifically, in mouse models of familial adenomatous polyposis, it has been shown that crosses with knockout mice of either cytosolic phospholipase A 2 or COX-2 reduced the size and metastatic capability of tumors (4 -6). However, despite the value of COX-2 inhibitors for the treatment of disease states, inhibition of COX-2 is unable to reduce AA release that may lead to excess production of leukotrienes and thromboxanes and, thus, side effects. In addition, AA production leads to glutathione depletion and reactive oxygen species generation. For these reasons, PLA 2 as an earlier therapeutic target in the pathway of eicosanoid production is of increased importance. Consequently, identification of components in the pathways proximal to PLA 2 activation is of key importance and high priority.
The main component of the venom from Loxosceles reclusus (brown recluse spider) is the enzyme sphingomyelinase D (SMase D) (7), which hydrolyzes sphingomyelin to produce ceramide-1-phosphate (C-1-P). The pathology of a wound generated from the bite of this spider consists of an intense inflammatory response mediated by AA and prostaglandins (8 -10). The production of endogenous C-1-P by the action of SMase D raised the intriguing possibility that C-1-P may act as an endogenous and proximal activator of PLA 2 and the subsequent inflammatory response, mediated by prostaglandins.
In this study C-1-P was found to be a potent and specific inducer of AA release and prostanoid synthesis in cells. This study further demonstrates that the activation of the enzyme that produces C-1-P in mammalian cells, ceramide kinase, is necessary for AA release and prostanoid synthesis in response to agonists. Thus, for the first time, a specific biology for the C-1-P and the enzyme that generates it, ceramide kinase, is demonstrated. Furthermore, this study also identifies a missing link in the induction of AA liberation in response to agonists.
Quantification of Arachidonic Acid Release-A549 cells (5 ϫ 10 4 ) were labeled overnight with 5 Ci/ml [ 3 H]AA (5 nM) (PerkinElmer Life Sciences). Cells were washed and incubated in DMEM supplemented with 2% fetal bovine serum for 2 h. After treatment, medium was transferred to 1.5-ml polypropylene tubes and centrifuged at 10,000 ϫ g, and [ 3 H]AA cpm was determined by scintillation counting. Results were controlled for an equivalent number of cells quantified by MTT assay as described (11) and by verification of total AA labeling by scintillation counting. For the experiments in Fig. 4, purified SMase D was obtained from Dr. Alfred H. Merrill, Jr., and SMase C (purified from Bacillus cereus) was obtained commercially from Sigma.
Quantification of Prostaglandin E 2 and Thromboxane B 2 Synthesis-For measurement of PGE 2 and thromboxane B 2 (TXB 2 ) levels, aliquots of media were taken at the indicated time points and assayed according to the manufacturer's instructions using the prostaglandin E 2 and thromboxane B 2 monoclonal EIA kit from Cayman Chemical. Briefly, media containing PGE 2 or TXB 2 compete with PGE 2 or TXB 2 acetylcholinesterase conjugate for a limited amount of monoclonal antibody. The antibody-PGE 2 or -TXB 2 conjugate binds to a goat-anti-mouse antibody previously attached to the 96-well plate. The plate is washed to remove unbound reagents, and then the substrate for acetylcholinesterase is provided. The concentration of each prostanoid in a sample is inversely proportional to the yellow color produced. Results were controlled for an equivalent number of cells quantified by the MTT assay.
Quantification of Ceramide-1-Phosphate Levels by Pulse-labeling with D-e-C 6 NBD-ceramide-Two hours before Bligh-Dyer extraction, cells were chased with 20 M D-e-C 6 NBD-ceramide (Matreya). At the appropriate times, lipids were extracted by the Bligh-Dyer method and analyzed for NBD-C-1-P formation using TLC analysis, and the results were normalized to total lipid phosphate as described (12).
Ceramide Kinase Assay-Ceramide kinase activity was measured as described previously with slight modifications (13). Briefly, 2 ϫ 10 5 cells were treated with A23187 (1 g/ml) for 5 min or IL-1␤ (2.5 ng/ml) for 3.5 h, washed with phosphate-buffered saline (ice-cold), and resuspended in a 20 mM HEPES (pH 7.2), 50 mM NaCl, 50% glycerol, 1 mM dithiothreitol, and protease inhibitors (10 g/ml each leupeptin, aprotinin, trypsin, chymotrypsin, and 1 mM phenylmethylsulfonyl fluoride). Cells were disrupted by two 15-s pulses with a Fisher 550 Sonic Dismembranator at setting 2. Each sample (10 g) was assayed for cera-FIG. 1. The effects of natural ceramide-1-phosphate on AA release (Panels A and B) and PGE 2 synthesis (Panels C and D) are time-dependent. A549 cells (5 ϫ 10 4 ) in 24-well plates were labeled overnight with 5 Ci/ml [ 3 H]arachidonic acid (5 nM). Cells were washed and placed in DMEM supplemented with 2% fetal bovine serum for 2 h. Cells were then treated with 2.5 M D-e-C 16 ceramide-1-phosphate (solubilized in 98% ethanol, 2% dodecane) for the indicated times. Results were controlled for an equivalent number of cells quantified by MTT assay and by verification of total AA labeling by scintillation counting. The results are presented as dpm of [ 3 H]arachidonic acid/ml of media for vehicle control (OE) and C-1-P (2.5 M) (q). In parallel experiments, the synthesis of PGE 2 was assayed, and the results are presented as pg of PGE 2 /ml of media controlled for an equivalent number of cells by MTT assay for vehicle control (OE) and C-1-P (2.5 M) (q). Data presented in this figure are representative of six separate determinations on three separate occasions mide kinase activity by incubation with NBD-C 12 ceramide/cardiolipin/ ␤-octylglucoside and 500 M ATP in 20 mM HEPES (pH 7.2), 50 mM NaCl, and 1 mM dithiothreitol for 15 min at 30°C. After Bligh-Dyer extraction, lipids were separated by thin layer chromatography, and NBD-C-1-P was visualized/quantified by fluorescent phosphorimaging.
RNAi Design and Transfections-Knock-downs of ceramide kinase were performed essentially as described using sequence-specific small RNAi reagents (14) and the human ceramide kinase RNAi starting 142 nucleotides from start codon (UGCCUGCUCUGUGCCUGUAdTdT and UACAGGCACAGAGCAGGCAdTdT). All sequences were evaluated against the data base using the NIH blast program to test for specificity (Xeragon, a Qiagen company). A549 cells (3 ϫ 10 4 ) were transfected with the 21-nucleotide duplexes using OligofectAMINE (Invitrogen) as recommended by the manufacturer. After the 4-h transfection, cells were treated and assayed for AA release or PGE 2 synthesis at 26 h post-transfection.
Reverse Transcriptase-PCR-For evaluating the levels of ceramide kinase mRNA and ribosomal 18 S RNA, 1 g of total RNA was reversetranscribed using Superscript II reverse transcriptase (Invitrogen) and oligo(dT) as the priming agent. After a 1-h incubation at 43.5°C, the reactions were stopped by 70°C heating for 15 min. Template RNA was then removed using RNase H (Invitrogen).
For evaluating ceramide kinase expression, an upstream 5Ј primer to ceramide kinase (5Ј-TCGTCTGTGTCGGCGGAGAT-3Ј) and a downstream 3Ј primer (3Ј-GCAAGACCCAACCACCGTTTC-5Ј) were designed to amplify ceramide kinase mRNA crossing at least one exon/ intron boundary to eliminate the possibility of amplifying genomic DNA. Using these primers 10% of the reverse transcriptase reaction was amplified for 15 cycles (94°C, 30s; 58°C, 30 s; 72°C, 1 min) using Platinum Taq DNA polymerase (Invitrogen). For ribosomal 18 S RNA, primers were obtained from Clontech Laboratories, and RNA was amplified for 10 cycles following the standard protocol provided by the company. with a maximal early increase of ϳ2-fold after 20 min followed by reuptake and a sustained increase in AA release after 1 h (Fig. 1, panels A and B). Increased PGE 2 synthesis was observed only after 1 h, corresponding with the sustained increase in AA liberation (Fig. 1, panels C and D). This effect of C-1-P on AA release and PGE 2 synthesis was dose-responsive with an EC 50 of 400 nM with a concentration as low as 100 nM, inducing a significant increase in AA release (Fig. 2, panels A  and B). Therefore, C-1-P induces a time-and dose-dependent increase in both AA and PGE 2 release.

Natural and Endogenous
Because C-1-P stimulated AA release, the production of the COX-1 product, TXB 2 , was also examined after C-1-P treatment (note: A549 cells do not express lipoxygenases or cytochrome P450 (15,16)). C-1-P (1 M) also dramatically increased TXB 2 by 2.5-fold after 2 h and 11-fold after 4 h (Fig. 3, panel A), suggesting that C-1-P induces the activation of a PLA 2 species upstream of cyclooxygenases. This was further demonstrated by using L929 fibroblasts that do not express COX-2. In these cells, PGE 2 synthesis/release was not observed in response to C-1-P (Fig. 3, panel B), but significant AA release of 225% over basal release was detected after 1 h of C-1-P treatment (1 M) (Fig. 3, panel B). Therefore, C-1-P induces a time-and dose-dependent increase in both AA and PGE 2 release in A549 cells that express high levels of COX-2, but is specific for AA release, as COX-1 products are produced and cells lacking COX-2 do not produce PGE 2 in response to C-1-P. Furthermore, C-1-P had no effect on the expression of COX-2 as judged by Western immunoblotting (data not shown).
To examine whether the effect of C-1-P on AA synthesis was specific, A549 cells were treated with closely related lipids, dioleoylphosphatidic acid (PA), dioleoylglycerol (DAG), D-erythro-sphingosine-1-phosphate (S-1-P), and D-erythro-C 16 ceramide (CER), for 2 h. Only C-1-P induced significant AA release (Fig. 4, panel A). Because D-erythro-C 16 ceramide induced a small increase in AA release, we examined the effects of long term treatment of D-erythro-C 16 ceramide and C-1-P. Unlike C-1-P, D-erythro-C 16 ceramide showed no further increase in AA release (Fig. 4, panel B). Thus, the effect of C-1-P on AA release in cells is lipid-specific.
To determine whether endogenously produced C-1-P can induce AA release, A549 cells were treated with SMase D, which hydrolyzes membrane sphingomyelin (SM) to generate C-1-P (data not shown). SMase D treatment resulted in an ϳ3-fold increase in AA release (Fig. 4, panel C). To demonstrate that the enhancement of AA release was a specific effect of C-1-P generation in response to SMase D, A549 cells were treated with SMase C, which hydrolyzes membrane SM to generate ceramide directly (data not shown) and remove the substrate for SMase D (Fig. 4, panel C) (17). SMase C alone only induced modest AA release, which could be the result of the action of ceramide or further conversion to C-1-P (18). Importantly, pretreatment with SMase C inhibited AA release in response to SMase D (Fig. 4, panel C), demonstrating that SM is required for the action of SMase D and that C-1-P is the primary product of SM hydrolysis responsible for the action of SMase D on AA release. Thus, C-1-P is a potent effector of AA release in A549 cells, and more importantly, generation of endogenous C-1-P can induce significant AA release.
The Generation of Endogenous Ceramide-1-Phosphate and Activation of Ceramide Kinase Coincides with AA Release in response to IL-1␤ and A23187-If C-1-P functions as a signaling molecule in the pathway to AA release and eicosanoid synthesis, then an increase in the levels of C-1-P should be observed before or coinciding with AA release in response to agonists. The cytokine, IL-1␤, and the calcium ionophore, A23187, are potent and well established inducers of AA release (19,20). IL-1␤ began to increase AA release after 3.5-4 h of treatment in A549 cells, whereas A23187 increased AA release within minutes (Fig. 5, panels A and B). Using pulse labeling with NBD-D-e-C 6 ceramide, it was found that IL-1␤ invoked a significant increase of 242% in C-1-P levels after 4 h (Fig. 5,  panel A). A23187 invoked a significant increase in C-1-P levels within 5 min of treatment, reaching a maximal increase in C-1-P levels of ϳ4-fold after 10 min (Fig. 5, panel B). Thus, IL-1␤ and A23187 increase C-1-P levels within the relevant time frame of AA release.
Little is known about the biosynthesis of C-1-P in vivo. C-1-P could be generated through either the action of a SMase D-type enzyme or ceramide kinase. Mammalian cells are known to contain ceramide kinase activity, which has been recently identified by molecular cloning (21); however, no mammalian SMase D-like activity has been described or identified. Moreover, the preceding results on the generation of C-1-P are more consistent with the activation of ceramide kinase, since treat- . At the appropriate times, media was taken for AA analysis, and lipids from the cells were extracted by the Bligh-Dyer method followed by analysis of NBD-C-1-P formation. Data are presented as percent control of C-1-P levels (q) and AA release (f). Data are representative of three separate determinations on two separate occasions. Panel C, IL-1␤ and calcium ionophore increase ceramide kinase activity. A549 cells (5 ϫ 10 5 ) were treated with either 2.5 ng/ml IL-1␤ for 3.5 h or 1 g/ml A23187 for 5 min. Cells were then washed, lysed in assay buffer, and assayed for ceramide kinase activity. Data are presented as percent control of ceramide kinase activity (arbitrary densitometry units of fluorescent NBD-C-1-P produced/min/mg of total cellular protein). Data are representative of three separate determinations on two separate occasions determinations on three separate occasions. Panel C, the effects of endogenously generated ceramide 1-phosphate on AA release. A549 cells (5 ϫ 10 4 ) were treated with either 0.5 units/ml SMase D for 8 h, 100 milliunits/ml SMase C for 8.5 h, or pretreated for 30 min with 100 milliunits/ml SMase C followed by the addition of 0.5 units/ml SMase D for 8 h. The results are presented as normalized dpm of [ 3 H]arachidonic acid/ml of media controlled for an equivalent number of cells by MTT assay. Data are representative of four separate determinations on two separate occasions ment with agonists enhanced the phosphorylation of NBDceramide to NBD-C-1-P. Therefore, we examined the activation and the role of ceramide kinase in agonist-induced prostanoid synthesis. Treatment of cells with IL-1␤ (3.5 h) or A23187 (5 min) increased ceramide kinase activity by 187 and 216%, respectively (Fig. 5, panel C). Thus, known agonists of AA liberation also activate ceramide kinase within the relevant time frame of AA release.
Down-regulation of Ceramide Kinase Using RNAi Technology Inhibits AA Release and PGE 2 Synthesis in Response to IL-1␤ and A23187-To demonstrate a role for endogenous ceramide kinase in regulating PLA 2 activation/AA release in response to agonists, we used RNAi technology to specifically down-regulate ceramide kinase (14). RNAi specifically decreased the mRNA levels of ceramide kinase by 73%, and endogenous ceramide kinase activity was decreased 67% without effects on cell number (Fig. 6). Mock transfection and transfection of control (scrambled) RNAi had no effect on the mRNA levels or the activity of ceramide kinase (data not shown). Although ceramide kinase was cloned because of its similarity to sphingosine kinase 1 (21), there was no decrease in immunoreactive levels of sphingosine kinase-1 observed with RNAi targeted to ceramide kinase (data not shown). Importantly, down-regulation of ceramide kinase by RNAi inhibited AA release in response to A23187 by 72% and AA release in response to IL-1␤ by 78%. When normalized to the down-regulation of ceramide kinase activity and mRNA expression induced by RNAi, this is equivalent to a complete abolishment of AA liberation in response to these agonists. This effect of ceramide kinase RNAi was also dose-responsive, with an IC 50 of ϳ30 nM, indicating the high potency of the action of RNAi (Fig. 7, panels  A-C). Furthermore, ceramide kinase RNAi also inhibited PGE 2 synthesis in response to IL-1␤ and A23187 to the same extent in A549 cells (data not shown). In another cell type, L929 fibroblasts, AA release, and PGE 2 synthesis in response to IL-1␤ were also inhibited more than 60% by RNAi targeted to ceramide kinase (data not shown). Ceramide kinase RNAi did not affect the distal "machinery" of PLA 2 activation/AA release as transfection of ceramide kinase RNAi had no effect on AA release in response to C-1-P treatment (Fig. 7, panel D). Thus, ceramide kinase is a necessary enzyme in the induction of AA liberation in response to cytokines and calcium-mobilizing agonists in cells.

DISCUSSION
In this study, we demonstrate that C-1-P is a potent modulator of AA release and PGE 2 production. Furthermore, endogenous C-1-P and the enzyme that generates C-1-P, ceramide kinase, were necessary for PLA 2 activation in response to cytokines and calcium ionophore. Thus, a specific biology has been established for C-1-P, and for the first time a physiological role for ceramide kinase, the enzyme that catalyzes the formation of C-1-P, has been determined. Moreover, our results show ceramide kinase to be a potentially important regulatory component of inflammatory responses because AA release is crucial for its conversion to eicosanoids. Therefore, ceramide kinase and C-1-P have been established as new "players" proximal to the liberation of AA.
The sphingomyelin-derived metabolites, ceramide, sphingosine, and sphingosine-1-phosphate are emerging as an important class of bioactive lipid mediators. Although the metabolite of ceramide, C-1-P, was identified more than a decade ago (22,23), only tantalizing hints of its possible functions have since appeared (24 -31). Micromolar concentrations of C-1-P could induce an increase in DNA synthesis in cells (24,25). Shayman and co-workers (28) also demonstrate that during phagocytosis in neutrophils ceramide kinase activity is increased followed by an increase in C-1-P, but recently Rile et al. (30) reported that C-1-P levels were not increased in response to cell stimulation in neutrophils. Therefore, a biological role for ceramide kinase in the phagocytosis pathway of neutrophils remains to be defined. This study convincingly establishes a biological role for C-1-P as a proximal mediator of AA liberation and adds it to a growing list of bioactive sphingolipid metabolites. In contrast to previous reports on C-1-P, this study demonstrates that C-1-P induces a biological effect (liberation of AA) in the nanomolar range, and the generation of endogenous C-1-P by the action of SMase D alone was sufficient to elicit this biological response in cells. Furthermore, the recent cloning of ceramide kinase (21) allowed for the specific down-regulation of the enzyme using RNAi technology, and thus, for the first time a biological role for ceramide kinase has been established. Therefore, the product of ceramide kinase, C-1-P, fulfills several of the following key criteria of a bioactive lipid. First, C-1-P levels FIG. 6. Ceramide kinase RNAi down-regulates endogenous ceramide kinase activity and mRNA levels. Panel A, the effect of ceramide kinase RNAi on endogenous ceramide kinase activity. A549 cells (5 ϫ 10 4 ) were transfected with the 21-nucleotide duplexes for ceramide kinase or scrambled control sequence. After the 4 h transfection, cells were incubated in normal growth media for 26 h. Cells were then washed, lysed in assay buffer, and assayed for ceramide kinase activity. Data are presented as percent control ceramide kinase activity with control RNAi (Control) or ceramide kinase RNAi (CerK RNAi). Data are representative of three separate determinations on two separate occasions. Panel B, the effect of ceramide kinase RNAi on the mRNA levels of ceramide kinase. Following the same procedure as outlined in panel A, total RNA was extracted and analyzed via logarithmic reverse transcriptase-polymerase chain reaction for human ceramide kinase and 18 S ribosomal RNA. Data are presented as percent control after the densitometry of ceramide kinase mRNA was normalized to 18 S ribosomal RNA levels with control RNAi (Control) or ceramide kinase RNAi (CerK RNAi). Data are representative of three separate determinations on two separate occasions are regulated in response to agonists. Second, exogenous C-1-P induces a specific biochemical and cellular response (release of AA), and this action of C-1-P demonstrates lipid specificity. Third, endogenous C-1-P reproduces these effects specifically. Last, the generation of C-1-P is required for AA release in response to inflammatory agonists. Because the presented study has identified C-1-P as a new messenger lipid in biological systems, the search for a direct intracellular target of C-1-P is now warranted.
Previous reports of sphingolipids regulating inflammatory responses give insight into a possible intracellular target of C-1-P. Klapisz et al. (32) report that SMase C and exogenous ceramide enhanced activation of cytosolic PLA 2 (cPLA 2 ) in response to A23187, suggesting that ceramide might enhance calcium-dependent translocation of cPLA 2 to membranes. Sato and co-workers (33) then reported that both ceramide and SMase C treatment did indeed enhance calcium-dependent translocation of cPLA 2 . Because ceramide and SMase C alone FIG. 7. Ceramide kinase regulates AA release in response to IL-1␤ and calcium ionophore. Panel A, the effect of ceramide kinase RNAi on AA release in response to IL-1␤. A549 cells (5 ϫ 10 4 ) were transfected with the 21-nucleotide duplexes for ceramide kinase or scrambled control sequence. After the 4-h transfection, cells were incubated overnight. Cells were washed and placed in DMEM supplemented with 2% fetal bovine serum for 2 h. After 26 h post-transfection, cells were treated with IL-1␤ (2.5 ng/ml). Data are presented as AA release (dpm) after 4 h of treatment for control RNAi plus vehicle (f), ceramide kinase RNAi plus vehicle (q), control RNAi plus IL-1␤ (2.5 ng/ml) (OE), and ceramide kinase RNAi plus IL-1␤ (2.5 ng/ml) (ࡗ). Data are representative of six separate determinations on three separate occasions. Panel B, the effect of ceramide kinase RNAi on AA release in response to 5 min of treatment with A23187. A549 cells (5 ϫ 10 4 ) were transfected with the 21-nucleotide duplexes for ceramide kinase or scrambled control sequence. After the 4 h transfection, cells were labeled overnight with 5 Ci/ml [ 3 H]arachidonic acid (5 nM). Cells were washed and placed in DMEM supplemented with 2% fetal bovine serum for two h. After 26 h post-transfection, cells were treated with A23187 (1 g/ml) for 5 min and assayed for AA release. Data are presented as AA release (dpm) for control RNAi plus vehicle (f), ceramide kinase RNAi plus vehicle (q), control RNAi plus A23187 (1 g/ml) (OE), and ceramide kinase RNAi plus A23187 (1 g/ml) (ࡗ). Data are representative of six separate determinations on three separate occasions. Panel C, the effect of ceramide kinase RNAi on AA release in response to 30 min of treatment with A23187. Following the same procedure as outlined in panel B for ceramide kinase RNAi, cells were treated with A23187 (1 g/ml) for 30 min and assayed for AA release. Data are presented as AA release (dpm) for control RNAi plus vehicle (f), ceramide kinase RNAi plus vehicle (q), control RNAi plus A23187 (1 g/ml) (OE), and ceramide kinase RNAi plus A23187 (1 g/ml) (ࡗ). Data are representative of six separate determinations on three separate occasions. Panel D, C-1-P overcomes the inhibition of ceramide kinase RNAi on AA release. Following the same procedure as outlined in panel B, cells transfected with control RNAi (Control) and ceramide kinase RNAi (CerK RNAi) were treated with C-1-P (2.5 M) for 5 h, and AA release (dpm) was measured. Data are representative of three separate determinations on two separate occasions did not elicit an increase in cPLA 2 translocation in these reports, conversion of ceramide to C-1-P in response to A23187 may explain these early findings and suggests cPLA 2 as a possible intracellular target of C-1-P. In this regard, we demonstrated that A23187 induces the activation of ceramide kinase and the accumulation of C-1-P, and because this study shows that C-1-P is the actual mediator of AA liberation, the conversion of ceramide to C-1-P is a likely explanation of the synergism observed between ceramide generation and enhanced cPLA 2 activation. Interestingly, ceramide kinase is a Ca 2ϩ -dependent enzyme like cPLA 2 (13, 34 -37), and ceramide kinase also demonstrates very similar tissue expression as cPLA 2 (34 -37). Because cPLA 2 contains a calcium-dependent lipid binding domain similar to the C2 domain of protein kinase C (38 -41), C-1-P may possibly act indirectly to activate cPLA 2 since exogenous C-1-P has recently been reported to induce calcium mobilization in cells (29).
If the inference that C-1-P indirectly activates cPLA 2 , the presented study may have "global" implications bridging the gap between Ca 2ϩ -dependent activation of cPLA 2 and the two other reported mechanisms of cPLA 2 activation, phosphorylation of Ser-505 and -727 and phosphatidylinositol 4,5-bisphophatedependent activation of cPLA 2 (42)(43)(44)(45)(46). Phosphorylation of Ser-505 and -727 is required for both calcium ionophore-and cytokine-induced AA release (47), and several reports suggest that the mitogen-activated protein kinase pathway is responsible for modulating the phospho state of cPLA 2 (43,48,49). This, therefore, correlates well with the report demonstrating that C-1-P treatment activates the mitogen-activated protein kinase pathway (31). The inhibitor of serine-threonine protein phosphatases, okadaic acid, has also been shown to induce AA release via cPLA 2 in a calcium-independent manner, possibly through effects on the phospho-state of the enzyme (42). Okadaic acid may be an artificial means of mimicking the intracellular role of endogenously generated C-1-P since unpublished findings from our laboratory demonstrate that C-1-P is a potent inhibitor (IC 50 ϭ 50 nM) of protein phosphatase-1 and protein phosphatase 2A. Thus, the generation of C-1-P may act indirectly to induce/enhance the phosphorylation of Ser-505 and Ser-727 of cPLA 2 by inhibiting protein phosphatases and simultaneous activation of the mitogen-activated protein kinase pathway. A role for ceramide kinase and C-1-P in cPLA 2 activation may also have implications in phosphatidylinositol 4,5-bisphophate-dependent activation of cPLA 2 since the recent cloning of ceramide kinase identified a PH domain in its N terminus (21); thus, one can hypothesize that phosphatidylinositol 4,5-bisphophate may activate ceramide kinase through its PH domain, generating C-1-P and eliciting cPLA 2 activation indirectly, as alluded to above.
Based on the results of this study, a pathophysiologic link can now be inferred between the production of C-1-P and inflammatory responses mediated by AA and prostaglandins observed from the bite of a brown recluse spider. C-1-P may also play a role in the reported link between PGE 2 , neurotransmitters, and airway epithelial inflammation since the highest levels of C-1-P are found in synaptic vesicles (13,23). In conclusion, this study provides strong evidence for ceramide kinase and C-1-P as major regulators/inducers of inflammatory responses and may have implications in the development of therapeutics for inflammatory disorders.