Bimodal Regulation of Ceramidase by Interleukin-1β

Interleukin 1β (IL-1β) induces the hydrolysis of sphingomyelin (SM) to ceramide (Cer) in primary cultures of rat hepatocytes, and Cer has been proposed to play a role in the down-regulation of cytochrome P450 2C11 (CYP2C11) and induction of α1-acid glycoprotein (AGP) by this cytokine (Chen, J., Nikolova-Karakashian, M., Merrill, A. H. & Morgan, E. T. (1995)J. Biol. Chem. 270, 25233–25238). Nonetheless, some of the features of the down-regulation of CYP2C11 do not fit a simple model of Cer as a second messenger as follows:N-acetylsphinganine (C2-DHCer) is as potent asN-acetylsphingosine (C2-Cer) in suppression of CYP2C11; the IL-1β concentration dependence for SM turnover is different from that for the increase in Cer; and the increase in Cer mass is not equivalent to the amount of SM hydrolyzed nor the time course of SM hydrolysis. In this article, we report that these discrepancies are due to activation of ceramidase by the low concentrations of IL-1β (∼2.5 ng/ml) that maximally down-regulate CYP2C11 expression, whereas higher IL-1β concentrations (that induce AGP) do not activate ceramidase and allow Cer accumulation. This bimodal concentration dependence is demonstrated both by in vitro ceramidase assays and in intact hepatocytes using a fluorescence Cer analog, 6-((N-(7-nitrobenz-2-oxa-1,3-diazol-4-yl)amino)-Cer (NBD-Cer), and following release of the NBD-fatty acid. IL-1β increases both acid and neutral ceramidase activities, which appear to be regulated by tyrosine phosphorylation because pretreatment of hepatocytes with sodium vanadate increases (and 25 μmgenistein reduces) the basal and IL-1β-stimulated ceramidase activities. Since these findings suggested that sphingosine (and, possibly, subsequent metabolites) is the primary mediator of the down-regulation of CYP2C11 by IL-1β, the effects of exogenous sphingosine and C2-Cer on expression of this gene were compared. Sphingosine was more potent than C2-Cer in down-regulation of CYP2C11 when added alone or with fumonisin B1 to block acylation of the exogenous sphingosine. Furthermore, the suppression of CYP2C11 by C2-Cer (and C2-DHCer) is probably mediated by free sphingoid bases, rather than the short chain Cer directly, because both are hydrolyzed by hepatocytes and increase cellular levels of sphingosine and sphinganine. From these observations we conclude that sphingosine, possibly via sphingosine 1-phosphate, is a mediator of the regulation of CYP2C11 by IL-1β in rat hepatocytes and that ceramidase activation provides a “switch” that determines which sphingolipids are elevated by this cytokine to produce multiple intracellular responses.

The lipid backbones (e.g. ceramide, sphingosine, and sphingosine 1-phosphate) of sphingolipids are highly bioactive compounds that have the potential to serve as second messengers in the regulation of cell growth, differentiation, diverse cell behaviors, and cell death (for recent reviews, see Refs. [1][2][3]. Signaling via such products of sphingolipid turnover is exemplified by the formation of ceramide (Cer) 1 upon activation of sphingomyelinase(s) by cytokines (tumor necrosis factor-␣ and interleukin 1␤ (IL-1␤)) (4, 5), 1␣,25-dihydroxyvitamin D 3 (6), and a wide range of other agents (7)(8)(9), and by the formation of sphingosine 1-phosphate by activation of sphingosine kinase by platelet-derived growth factor (10). It has been proposed that cells activate sphingomyelinase in response to cytokines, whereas growth factors also activate ceramidase and sphingosine kinase and thereby choose between the formation of Cer versus sphingosine 1-phosphate (and other bioactive metabolites) (11). Nonetheless, current knowledge is still quite fragmentary about the full spectrum of sphingolipid metabolites that are produced in response to various stimuli and how the activities of the key regulatory enzymes are controlled.
We have investigated the involvement of sphingolipid metabolites in the inflammatory response of hepatocytes to IL-1␤ (12) and found that this cytokine stimulated sphingomyelin (SM) hydrolysis and increased cellular levels of Cer. Furthermore, addition of a short chain Cer (N-acetylsphingosine, C2-Cer) induced the expression of ␣ 1 -acid glycoprotein (AGP) and down-regulated cytochrome P450 2C11 (CYP2C11), which mimicked the effects of IL-1␤ on these genes. By these criteria, SM metabolites (possibly Cer) appear to mediate the response of hepatocytes to IL-1␤, as has been seen in other systems (5,13).
Nonetheless, not all of the observations with rat hepatocytes were consistent with Cer acting as the mediator of CYP2C11 down-regulation (12). The concentrations of IL-1␤ that suppressed CYP2C11 (ϳ2.5 ng/ml, ED 50 ϭ 1 ng/ml) did not cause an increase in Cer mass (whereas Cer was elevated by the higher levels of IL-1␤ that induced AGP). Furthermore, both C2-Cer and N-acetylsphinganine (C2-DHCer) down-regulated CYP2C11 (which is not typically seen in systems where Cer is a mediator) (14,15), and only C2-Cer induced AGP. These discrepancies, plus the detection of some increase in free sphingosine in IL-1␤-treated hepatocytes (12), suggested that this cytokine may be activating ceramidase as well as sphingomyelinase. This article describes the dose-dependent activation of ceramidase in rat hepatocytes, apparently via phosphorylation and dephosphorylation, and demonstrates that sphingosine is more potent than Cer in down-regulation of CYP2C11. These findings establish that downstream metabolites of Cer, e.g. sphingosine or sphingosine 1-phosphate, are responsible for the down-regulation of CYP2C11 by IL-1␤ in hepatocytes.
Cell Culture-The tissue culture dishes were treated with Matrigel (6.3 mg/ml) as described previously (12). Hepatocytes were isolated from ether-anesthetized rats by in situ collagenase perfusion (18), and the cells (3.5 ϫ 10 6 per plate; viability Ͼ80%) were plated in 3 ml of Waymouth's medium containing insulin (0.15 M) as the only hormone. Cultures were maintained for 5 days at 37°C in 5% CO 2 atmosphere with replacement of the medium every 48 h, commencing 3 h after plating. This protocol is used because expression of CYP2C11 mRNA is almost completely lost when hepatocytes are initially placed in culture but is restored after 5 days in culture on Matrigel (18).
Cell Treatments-The cells were treated on day 5 with the indicated concentrations of IL-1␤, which was first prepared as a concentrated stock solution in 0.1% bovine serum albumin (BSA) in phosphatebuffered saline (PBS) and diluted with culture medium immediately before use. The control cells were treated with the same concentrations of BSA and other vehicles (i.e. solvents used for delivery of the sphingolipids). Unless otherwise stated, the cells were incubated with IL-1␤ for 45 min prior to analysis of the enzymatic activities or metabolites. To test the effects of exogenously added ceramides on cellular free long chain bases, the hepatocytes were treated with 30 M C2-Cer or C2 DHCer for 1 and 4 h. They were delivered to the cells from stock solutions in ethanol (12). To determine the effects of different sphingolipids on CYP2C11, the cells were treated with up to 30 M sphingosine, C2-Cer, and/or 25 M fumonisin B 1 for 24 h. In these experiments C2-ceramide was added from 60 mM stock in Me 2 SO. Sphingosine was delivered to the cells as BSA complex in a molar ratio of 1:1. This complex was formed by injecting 10 l of 100 mM sphingosine (stock solution in ethanol) into 1 ml of 1 mM BSA in PBS, vigorously vortexing, and further incubating for 30 min at 37°C. For the experiments using various inhibitors, the cells were preincubated with 1 mM sodium vanadate or 25 M genistein for 1 h, or with 50 M DL-threo-sphinganine (as a BSA complex) for 10 min, prior to addition of IL-1␤. In all experiments the control groups were treated with the respective vehicles, and in the cases of dose response, the concentration of the vehicles was kept one and the same between the groups. Hepatocytes were harvested at appropriate times with 1 ml of PBS, and aliquots were taken for protein measurement. In the case of sodium vanadate or genistein pretreatment, the harvest buffer was supplemented with the respective inhibitor, and these inhibitors were kept throughout the procedure.
Measurements of Sphingolipid Mass-The lipids were extracted by the method of Bligh and Dyer (19), modified as described previously (20), and were analyzed by thin layer chromatography on Silica gel 60 plates (20 ϫ 20 cm) using chloroform:methanol:triethylamine:2-propyl alcohol:0.25% potassium chloride (30:9:25:18:6, by volume) as the developing solvent. The regions migrating with standard Cer were scraped from the plate, and the SM was visualized with I 2 or 50% H 2 SO 4 and quantitated by phosphate assay (12). To quantitate the mass of Cer, 5 nmol of N-acetyl-C 20 -sphinganine was added to the unknown Cer sample on the chromatoplate, and then the lipids were eluted from the silica with 1 ml of chloroform:methanol (1:1, by volume) followed by 1 ml of methanol. The combined eluates were dried in vacuo, and the Cer mass was quantitated by HPLC of the long chain bases released after an acid hydrolysis in 0.5 M HCl in methanol at 65°C for 15 h. Free long chain bases were analyzed as described in Ref. 21.
In Vitro Enzyme Assays-The cells were scraped from two dishes (each of which contained approximately 1.0 mg of cellular protein), pooled, and recovered by centrifugation at 300 ϫ g for 5 min. They were lysed in 1.0 ml of 0.2% Triton X-100 in 10 mM Tris, pH 7.4 (supplemented where appropriate with 1 mM sodium vanadate or 25 M genistein), for 10 min on ice. The lysates were homogenized with three passes through a 25-gauge needle, and 10-l aliquots were taken for protein assay. NBD-Cer or NBD-SM was added to the lysates to give a concentration of 20 M (from 10 mM stock solutions in ethanol), and they were incubated at 4°C for 10 min to allow the substrate to equilibrate among the micelles. Aliquots of this mixture (containing approximately 0.1 mg of protein and a final concentration of NBD-substrate of approximately 3 M) were added to 5 mM MgCl 2 in 10 mM Tris, pH 7.4 (for the neutral sphingomyelinase), 10 mM Tris, pH 7.4 (for neutral ceramidase), 0.5 M acetate buffer, pH 4.5 (for acidic sphingomyelinase and ceramidase), or 10 mM Hepes, pH 9.5 (for alkaline ceramidase), for a final volume of 300 l. All incubation buffers contained 0.2% Triton X-100 and were supplemented with sodium vanadate or genistein for the respective treatments. After incubation for 1 h at 37°C, the reaction was stopped by adding 1 ml of the mobile phase used in the subsequent HPLC analysis, and the products were analyzed by HPLC (as described below). Product formation was linear over this time of incubation and proportional to the amount of added protein for up to 0.3 mg/assay.
Analysis of Ceramidase in Situ Using NBD-Cer-On the 4th day in culture, the hepatocytes were changed to new medium containing 6 M NBD-Cer (added to the medium by ethanol injection as described previously (22)). After 12 h, the medium was changed, and the cells were placed in new medium with or without IL-1␤ at the concentrations indicated in the text and figures. After 45 min, the cells were harvested in cold 0.5 ml of PBS, and 1.5 ml of methanol:water:phosphoric acid, 850:150:1.5 (by volume), was added to the cells, which were then incubated at 37°C for 1 h with shaking. After removal of insoluble material by centrifugation in a table top clinical centrifuge (at 3,000 rpm for 10 min), an aliquot of the supernatant was analyzed by HPLC (as described below).
HPLC Analysis of NBD-Lipids-For analysis of NBD-ceramide hydrolysis, the samples were injected onto a normal phase Silica column (Nova Pak, Bio-Rad) and eluted with hexane:chloroform:methanol:water:triethylamine (240:560:180:11:7, by volume) at a flow rate of 2 ml/min. Under these conditions, the NBD-fatty acid elutes at 1.7 min and NBD-ceramide at 3.2 min.
For analysis of SM hydrolysis, a second mobile phase (hexane:chloroform:methanol:water:85% phosphoric acid (281:650:300:30:4, by volume) (23) at a flow rate of 2 ml/min was used. In this system, NBD-SM elutes at 7.8 min and NBD-Cer and NBD-Fatty acid appear as a single peak at 1.6 min.
When analysis of all three NBD-lipids was needed they were analyzed using a reverse-phase column (Nova Pak, C18, Bio-Rad), eluted with methanol:water:phosphoric acid (850:150:1.5, by volume) as the mobile phase (at 2 ml/min). In this system, the elution times of the NBD-fatty acid, NBD-Cer, and NBD-SM are 1.5, 10.3, and 12.3 min, respectively. The resolution of NBD-Cer and NBD-SM diminished as the column aged using this system.
The NBD fluorescence was analyzed with excitation at 455 nm and emission at 530 nm. The mass of the NBD compounds was calculated by comparison with the fluorescence of the NBD-lipids standards.
Isolation of Total RNA and Slot-Blot Assays-Total hepatocyte RNA was prepared by the acid phenol extraction method (24). The relative abundances of CYP2C11 mRNA in total RNA was measured by slot-blot hybridization assay as described previously (12), using a full-length cDNA for CYP2C11. Bound probe was assayed by autoradiography and densitometric scanning. All results were normalized to the content of poly(A ϩ ) RNA and measured by probing slot-blots with an oligo(dT) 30 probe. The amount of total RNA were previously determined to be in the range giving a linear response. (12) demonstrated that IL-1␤ induced SM turnover in rat hepatocytes which (in combination with other data in that study) indicated that SM metabolite(s) are involved in the downregulation of expression of CYP2C11 and induction of AGP by this cytokine. It was puzzling, however, that Ͻ1 ng/ml of IL-1␤ induced SM turnover, but Cer mass did not increase until addition of Ͼ1 ng/ml (cf. Figs. 2 and 3 in Ref. 12). This discrep-ancy in dose-response alerted us to examine other differences in dose response, and these observations (from Refs. 12 and 25) are summarized in Fig. 1. CYP2C11 is suppressed by lower concentrations of IL-1␤ (ϳ2.5 ng/ml; ED 50 ϭ 1 ng/ml) than are required to induce AGP (which occurs in two phases, one at 0.5-2 ng/ml IL-1␤ and a second, larger increase at Ͼ5 ng/ml) (Fig. 1, left panels) (25); furthermore, maximal suppression of CYP2C11 is apparent with 10 -30 M C2-Cer, whereas induction of AGP does not exhibit saturation behavior at 30 M (Fig.  1, right panels) (12).

Dose Dependence and Kinetics of IL-1␤-induced Changes in the Mass of Endogenous Sphingolipids-Our previous study
Thus, to obtain a better understanding of the relationship between SM turnover and appearance of Cer (and sphingosine), hepatocytes were treated with 5 ng/ml IL-1␤, and these lipids were quantified (Fig. 2). 2 There was a significant reduction in SM mass (ϳ0.75 nmol) within 15 min, whereas Cer mass did not increase until 45 min, and there was a small (0.1 nmol), but statistically significant, decline in Cer at the 15-min time point. Even at the maximum increase in Cer mass (0.5 nmol), this accounted for less than half of the total turnover of SM (Ͼ1.25 nmol). Some of this difference could be accounted for by an elevation in the mass of sphingosine (by ϳ0.25 nmol at 45 min).
The goals of our next studies were to explain why Cer mass does not increase upon addition of the low concentrations of IL-1␤ that induce SM turnover and suppress CYP2C11 and to determine if this is relevant to the regulation of CYP2C11 by sphingolipids. One possibility is that SM is hydrolyzed to another metabolite, such as ceramide 1-phosphate or sphingosine phosphorylcholine; however, neither of these were detected on the thin layer chromatoplates (data not shown). Because there was an elevation in sphingosine (albeit only at a later time point), it appeared that Cer was undergoing hydrolysis; therefore, the effect of IL-1␤ on ceramidase activity was determined.

Effects of IL-1␤ on Ceramidase Activity Assayed in Vitro-
Ceramidase activity was assayed at acidic, neutral, and basic pH (Fig. 3) because it has been suggested that there are multiple forms of this enzyme (27,28). Under basal conditions, the specific activity was greatest at pH 5.0, followed by pH 9.0, and a low, but measurable activity was obtained at pH 7.2. Addition of 2.5 ng/ml IL-1␤ to the hepatocytes 45 min prior to assay of ceramidase in vitro resulted in a 10-fold increase in the neutral activity and 2-fold increases in the activities at acidic and alkaline pH. However, when 5 ng/ml IL-1␤ was added to the cells, the acidic and neutral activities were not elevated above the controls, and the activity at alkaline pH was increased only 1.4-fold. The magnitude of these changes varied from experiment to experiment; however, the trends shown in Fig. 3 were seen consistently.
Effects of IL-1␤ on Ceramidase Activity Measured in Situ Using NBD-Cer-Because these results were somewhat surprising, and in vitro assays of lipid metabolizing enzymes can be complicated by many factors (delivery of substrates, etc.), ceramidase activation was examined using a method that can evaluate Cer turnover in situ. In this assay, hepatocytes are incubated overnight with NBD-Cer, which results in formation of some NBD-SM and NBD-hexanoic acid, but a substantial amount of NBD-Cer remains in the cells, and its subsequent hydrolysis in response to IL-1␤ can be analyzed by HPLC with quantitation of the products by fluorescence (as described under "Experimental Procedures"). 3 This assay has an advantage over measurement of sphingosine that NBD-hexanoic acid is not reutilized, whereas sphingosine can be removed by further metabolism (30,31).
When varying concentrations of IL-1␤ were added to the hepatocytes after the overnight preincubation with NBD-Cer, there was a significant decline in the amount of NBD-Cer and 2 We should note that particular care was taken to ascertain whether these changes are due to the addition of new medium to the cells because this has been found to induce sphingolipid biosynthesis and turnover in other systems (26). In the experiment depicted in Fig. 2, there were no significant changes in sphingolipids in cells that were changed to new medium without IL-1␤. Nonetheless, such was observed occasionally in other experiments; therefore, this should be routinely checked in experiments in which the culture medium is changed during the addition of an agonist. an essentially stoichiometric increase in the amount of NBDhexanoic acid (Fig. 4). The concentration dependence of the IL-1␤-induced changes in NBD-Cer and NBD-fatty acid was similar to the results of the assays of ceramidase in vitro (i.e. displayed a maximum at 2.5 ng/ml IL-1␤ and thereafter declined until no turnover was apparent at 7.5 ng/ml). This is remarkably good quantitative agreement, considering the many factors that differ between the in vitro and in situ assays (e.g. physical state of the enzymes and substrates, etc.).

Effects of Inhibitors of Tyrosine Phosphorylation and Dephosphorylation on Ceramidase Activity Assayed in Vitro-
To characterize the activation of ceramidase more thoroughly, the effects of incubating hepatocytes with sodium vanadate and genistein were determined because Coroneos et al. (11) have reported that ceramidase activity is increased in mesangial cells by platelet-derived growth factor, and the activation is blocked by inhibitors of tyrosine kinases. Sodium vanadate elevated basal ceramidase activities at pH 5 and 7.2 almost 6-fold (compared with the basal activities without IL-1␤ or vanadate, shown by the arrows in Fig. 5, left panel), whereas there was no noticeable change in the activity at pH 9. There were statistically significant increases in activity at each pH upon addition of 2.5 ng/ml IL-1␤, and these activities were higher than in the absence of vanadate (cf. Fig. 3). At 5 ng/ml, IL-1␤ caused little (or no) increase in the ceramidase activities at pH 5 or 7.2 (versus vanadate addition alone), whereas the activity at pH 9 was still elevated significantly, as was seen in the absence of vanadate (cf. Fig. 3).
When hepatocytes were treated with 25 M genistein, the basal ceramidase activities at pH 5 and 9 were approximately the same as for hepatocytes in the absence of this inhibitor of a number of tyrosine kinases (32, 33) (Fig. 5, right panel). Genistein blunted the increases in ceramidase in response to 2.5 ng/ml IL-1␤ at each pH, but the alkaline activity at 5 ng/ml IL-1␤ was affected less.
All together, these results suggest that ceramidase activities are increased by tyrosine phosphorylation (either on this enzyme or a up-stream activator), and 2.5 ng/ml IL-1␤ increases ceramidase activity via tyrosine phosphorylation. The activity at pH 9 appears to be regulated somewhat differently than the acidic and neutral activities because the basal activity was not increased by vanadate, although the stimulated activity was somewhat higher in the presence of vanadate. Assay of Sphingomyelinase Activity in Vitro-For comparison, in vitro assays of the acidic and neutral sphingomyelinase were conducted with hepatocytes incubated with IL-1␤ in the presence and absence of sodium vanadate or genistein (Fig. 6). Sodium vanadate reduced the basal sphingomyelinase activity at pH 5 by about 30% and at pH 7.2 by Ͼ80%; genistein pretreatment significantly increased the basal activity of the neutral sphingomyelinase. These results suggest that tyrosine phosphorylation of the neutral enzyme (or an up-steam regulator) suppresses its activity. Treatment of the cells with IL-1␤ caused only small increases in sphingomyelinase activity at either pH; however, when genistein was present, both 2.5 and 5 ng/ml IL-1␤ increased neutral sphingomyelinase activity significantly (Fig. 6, right panel). Although these responses to IL-1␤ are somewhat difficult to interpret, they confirm that this cytokine can increase sphingomyelinase activity, which is consistent with its induction of SM turnover in intact hepatocytes (Ref. 12 and Fig. 2).
Suppression of CYP2C11 by Sphingosine-The finding that ceramidase is activated by the concentrations of IL-1␤ that suppress CYP2C11 gene expression suggests that sphingosine (or a subsequent) metabolite may be the actual modulator of CYP2C11 rather than Cer per se. To test this hypothesis, the effects of exogenous sphingosine and C2-Cer on CYP2C11 were compared. As shown in Fig. 7, sphingosine was more potent than C2-Cer in down-regulating CYP2C11; furthermore, addition of an inhibitor of ceramide synthase (fumonisin B 1 ) did not block the effects of sphingosine. Therefore, the effects of sphingosine on CYP2C11 are not due to its conversion to Cer.
To determine if the suppression of CYP2C11 by C2-Cer (and C2-DHCer) could be due to hydrolysis of these short chain derivatives to sphingosine and sphinganine, respectively, hepatocytes were incubated with C2-Cer or C2-DHCer, and the amounts of the free sphingoid bases were measured. As shown in Fig. 8, there were significant elevations in free sphingosine and sphinganine; furthermore, cells treated with C2-Cer showed an elevation in both sphingosine and sphinganine (which implies that the elevated sphingosine causes accumulation of endogenous sphinganine). When hepatocytes were incubated with radiolabeled C2-Cer (data not shown), both degradation products and more complex sphingolipids were observed; therefore, it is evident that these short chain Cer analogs alter the cellular levels of free sphingoid bases, which could suppress CYP2C11.
This study did not attempt to establish if free sphingoid bases or subsequent metabolites (such as sphingosine 1-phosphate) are the ultimate "signals" for suppression of CYP2C11.
Nonetheless, it is likely that sphingosine 1-phosphate is formed in IL-1␤-treated cells since the amounts and time course of sphingosine accumulation did not account for the turnover of SM and Cer (Fig. 2). To test the likelihood that sphingosine is undergoing phosphorylation, hepatocytes were incubated with an inhibitor of sphingosine kinase (DL-threo-sphinganine) (34), which (alone) caused a 10% increase in the amount of sphingosine. When IL-1␤ was added to hepatocytes pretreated with DL-threo-sphinganine, the sphingosine mass increased by 60% (Fig. 9); therefore, it is likely that the sphingosine that is produced by ceramidase undergoes rapid metabolism.

DISCUSSION
This study has established that IL-1␤ not only induces SM hydrolysis in hepatocytes but also activates ceramidase in a highly concentration-dependent manner. This accounts for the lack of accumulation of Cer at concentrations of IL-1␤ that increase SM turnover (i.e. Ͻ4 ng/ml) and provides the first evidence (as far as we are aware) for IL-1␤ signaling via two sphingolipid mediators: sphingosine (or sphingosine 1-phosphate) in the suppression of CYP2C11 and Cer in the induction of AGP. Fig. 10 summarizes the data in support of these conclusions. The induction of AGP exhibited all of the usual characteristics of signaling via Cer: C2-Cer (but not C2-DHCer) increases AGP (12), and there are significant decreases in SM and increases in Cer at the concentrations of IL-1␤ that stimulate maximal expression of this gene. 4 Furthermore, although there is some induction of AGP at low concentrations of IL-1␤, most occur at the higher levels where there is activation of sphingomyelinase but not ceramidase. In contrast, CYP2C11 expression occurs at concentrations of IL-1␤ where ceramidase is activated (and Cer mass increases are not seen) and is more potently regulated by sphingosine than C2-Cer; furthermore, CYP2C11 is suppressed by both C2-Cer and C2-DHCer, which is probably due to elevation of cellular sphingosine and sphinganine. The latter observation was one of the first observations that suggested to us that Cer might not be the direct mediator of the down-regulation of CYP2C11 because Cer-mediated signal transduction pathways are typically sensitive to the presence or absence of the 4,5-trans-double bond of the sphingosine backbone (2), whereas many intracellular systems are affected comparably by sphingosine and sphinganine (35).
It is not clear why IL-1␤ affects CYP2C11, AGP, and sphingolipid metabolism with this bimodal concentration dependence. Hepatocytes might contain more than one IL-1␤ receptor (types I and II receptors, as well as soluble receptors, have been reported for this cytokine, and rat hepatocytes have been reported to have a third class of high affinity receptor for IL-1␤) (36) or a single receptor may be coupled to different downstream signal transduction machinery. These options cannot be distinguished at this time; however, bimodal effects of IL-1␤ are not unique to rat hepatocytes because low concentrations of IL-1␤ stimulate insulin secretion in pancreatic islets, whereas high concentrations are inhibitory (37).
IL-1␤ signaling is thought to involve a coordinate activation of protein kinases and inhibition of phosphatases (38). Inhibition of protein phosphatases (including phosphotyrosine phosphatase 1B) (39) is considered to be a critical early event in the regulation of IL-1␤ action. These aspects of IL-1␤ signaling may be related to the modulation of ceramidase activity because the finding that genistein and vanadate affect the in vitro activities of the acidic and neutral ceramidase suggests that these enzyme(s) are modulated by tyrosine phosphorylation, either directly or at upstream step(s) in activation of the enzymes. The acidic ceramidase has been recently purified (29), cloned, and sequenced (40) and has at least one tyrosine (Tyr-305) that is flanked by amino acids that are commonly found in 4 The dose-response data in Fig. 1 suggest that AGP induction occurs in two phases. This discussion refers mainly to the concentrations of IL-1␤ that result in the greatest increase in AGP mRNA (i.e. Ͼ5 ng/ml), which are also beyond the levels that achieve maximal suppression of CYP2C11.

FIG. 8. Cellular level of sphingosine and sphinganine in hepatocytes incubated with C2-ceramide and C2-dihydroceramide.
Hepatocytes were incubated with 30 M the indicated sphingolipids for the time shown, and then the amounts of free sphinganine and sphingosine were assayed by HPLC as described under "Experimental Procedures." The ** denotes changes that are statistically significant from the control with p Յ 0.005, and * corresponds to p Յ 0.05.  10. A scheme that summarizes the observations that indicate that sphingolipid signaling to down-regulate CYP2C11 in rat hepatocytes involves ceramidase activation and production of sphingosine (and, possibly, subsequent metabolites) compared with the induction of AGP by ceramide. The scheme is based on the previously observed (12) differences in concentration dependence for regulation of these two genes (by both IL-1␤ and C2-Cer) and C2-Cer and C2-DHCer to down-regulate CYP2C11, whereas induction of AGP is selective for C2-Cer; it incorporates the new observations on SM turnover and the sphingomyelinase activity of IL-1␤-treated cells, the finding that only the lower concentrations of IL-1␤ increase ceramidase activity, and that exogenous sphingosine is more potent than C2-Cer in down-regulation of CYP2C11 (as well as that exogenous C2-Cer and C2-DHCer elevate sphingosine and sphinganine). tyrosine phosphorylation sites (41), i.e. two acidic amino acids (glutamate and aspartate) at positions 2 and 5, and an arginine at position 7 toward the N-terminal. Ceramidase activation has been observed previously upon stimulation of mesangial cell growth by platelet-derived growth factor (and to be inhibited by genistein and vanadate as in our study) (11). In addition, ceramidase activation has been inferred to occur in adult cardiac myocytes treated with TNF-␣, since there is a 1.4-fold elevation of sphingosine (42); therefore, this may be a common mechanism for regulation of this aspect of sphingolipid signaling.
Based on numerous studies of the intracellular systems that are affected by exogenous addition of sphingolipids to cells in culture (35), the target(s) that could be involved in the downregulation of CYP2C11 by sphingosine or a subsequent metabolite (such as sphingosine 1-phosphate) include protein kinase C (43), sphingosine-activated protein kinases (44), extracellular signal-regulated kinases (45), transcriptional factor AP-1 (46), and retinoblastoma (rB) protein (47). Only a few of these have been correlated with changes in endogenous sphingoid bases: the inhibition of protein kinase C in cells where sphingoid bases (including the 1-phosphates) have been elevated (48), and the induction of intracellular calcium release and activation of AP-1 transcription factor, when the levels of sphingosine 1-phosphate (and sphingosine) are increased by growth factors (10,46). Although little is known about the transcriptional regulation of CYP2C11, AP-1 has been implicated in the acute phase response in the liver (49), and epidermal growth factor has been shown to down-regulate CYP 2C11 in hepatocytes (50) thus suggesting possible involvement of the Raf/MEK-pathway.
As this present study has shown, a given agonist (IL-1␤) can affect both Cer and sphingosine production in a highly concentration-dependent manner, apparently to achieve multiple, and differential, intracellular responses. It is not known if this complex behavior occurs in other cell types since most studies of sphingolipid signaling have focused on the production of only one product (usually Cer) and have often utilized a narrow range of agonist concentrations. However, given the large number of bioactive sphingolipid metabolites that are now known, and the diversity of the systems that are affected by them, it is likely that such complexity in sphingolipid signaling will be common.