Tissue injury in inflammation involves the release of several cytokines that activate sphingomyelinases and generate ceramide. In the lung, the impaired metabolism of surfactant phosphatidylcholine (PC) accompanies this acute and chronic injury. These effects are long-lived and extend beyond the time frame over which tumor necrosis factor (TNF)-α and interleukin-1β are elevated. In this paper, we demonstrate that in H441 lung cells these two processes, cytokine-induced metabolism of sphingomyelin and the inhibition of PC metabolism, are directly interrelated. First, metabolites of sphingomyelin hydrolysis themselves inhibit key enzymes necessary for restoring homeostasis between sphingomyelin and its metabolites. Ceramide stimulates sphingomyelinases as effectively as TNF-α, thereby amplifying the sphingomyelinase activation, and TNF-α, ceramide, and sphingosine all inhibit PC:ceramide phosphocholine transferase (sphingomyelin synthase), the enzyme that restores homeostasis between sphingomyelin and ceramide pools. Second, ceramide inhibits PC synthesis, probably because of its effects on CTP:phosphocholine cytidylyltransferase, the rate-limiting enzymatic step in de novo PC synthesis. The data presented here suggest that TNF-α may be an inhibitor of phospholipid metabolism in inflammatory tissue injury. These actions may be amplified because of the ability of metabolites of sphingomyelin to inhibit the pathways that should restore the normal ceramide-sphingomyelin homeostasis.
- tumor necrosis factor-α
- pulmonary surfactant
- cytidine 5′-triphosphate
the sphingomyelin pathway interacts in a large number of cellular activities including differentiation and mitosis, gene transcription, and cell death (26). These multiple actions result in part from the ubiquitous presence of receptors that initiate the signaling pathway, activated by several ligands. These include cytokines that are released at the onset of inflammatory injury (26). The central molecule in this pathway is ceramide (15).
There are numerous reports, many in the older literature, that show that changes in cellular metabolism accompany inflammation (12,27); indeed, some of the earliest reports on tumor necrosis factor (TNF)-α emphasized its effects in promoting cachexia (6). In the inflammatory response of the lung resulting from acute traumatic injury or infection, these events may affect the metabolism of phospholipids required for the maintenance of the proper amount and composition of pulmonary surfactant (30). The resulting condition leads to alveolar collapse, interference with gas transfer, hypoxia, and, in 40–60% of the patients, death (25).
Recent work (1, 34) indicates that TNF-α inhibits phosphatidylcholine (PC) synthesis, but there is very little published work on the mechanism by which TNF-α exerts this effect. It has been shown that TNF-α given to type II cells results in increased levels of ceramide and sphingosine (22), suggesting that TNF-α could be acting through one of these sphingolipids. Furthermore, in another recent study, Mallampalli and coworkers (21) showed that TNF-α inhibited PC synthesis, principally by its actions on CTP:phosphocholine cytidylyltransferase (CT). These changes appeared to be mediated through a decrease in the amount of CT, possibly through ubiquitin-proteasome processing. The work described in this and the companion paper (3) confirms and extends these observations. In this paper, we concentrate on describing the role of the sphingomyelin metabolites formed as a result of the stimulation of TNF-α on this overall process. In the companion paper, we investigate the signaling pathway used by ceramide to inhibit CT activity.
We report here that ceramide, generated as a consequence of signaling initiated by TNF-α, may have a central role in these metabolic perturbations. We present evidence that ceramide or a metabolite of ceramide inhibits regulatory events that would be expected to restore the normal balance between sphingomyelin and ceramide concentrations. In this regard, we report that TNF-α and ceramide and sphingosine, products of sphingomyelin metabolism initiated by TNF-α, inhibit the key enzyme responsible for the restoration of sphingomyelin-ceramide homeostasis, PC:ceramide phosphocholinetransferase (sphingomyelin synthase). Furthermore, ceramide activates acidic and neutral sphingomyelinases rather than inhibiting the products. These dual actions of ceramide (and/or ceramide metabolites) promote elevated ceramide pools, thereby reinforcing the sustained injury. We then show that at least two metabolites of sphingomyelin hydrolysis, ceramide and sphingosine, mimic exactly the actions of TNF-α (23) in that they diminish the synthesis of PC, reduce the content of cellular PC, and inhibit CT, the rate-limiting enzymatic step in PC synthesis (29). These results suggest that the effects of TNF-α are through ceramide pathway intermediates and that the pathways contain self-reinforcing controls that could prolong the elevated levels of ceramide and the undesirable metabolic effects.
We used H441 cells for these experiments. NCI-H441 [American Type Culture Collection (ATCC)] is a human adenocarcinoma thought to be derived from small-airway epithelial cells. These cells synthesize several of the constituents of surfactant (31) and have been widely used for studies in this area (5, 44). The cells were obtained from ATCC at the 50th passage, expanded in 10% fetal bovine serum (FBS)-McCoy's 5A medium with 50 μg/ml of gentamicin, frozen, and maintained in small aliquots in liquid nitrogen. Cells used for these experiments were maintained in the same medium. Experiments were conducted in six-well plates when cells had reached ∼50–60% confluence. In all protocols, the medium was changed to McCoy's 5A without FBS 24 h before the start of the experiment.
These experiments studied the effects of TNF-α and two sphingolipids, C2 ceramide and sphingosine. TNF-α was added to the medium dissolved in McCoy's 5A; ceramide and sphingosine were dissolved in DMSO. When DMSO was used, control cells received the same amounts of vehicle as the experimental cells. In none of the experiments was the amount of added DMSO >0.1 volume percent (1:1,000 vol/vol). Sphingolipids were delivered to cells in medium without serum in concentrations used extensively in the literature (8, 11, 17,39).
In two completely separate experiments, we monitored cell viability by trypan blue exclusion (2) after giving the cells 10 μM C2 ceramide or 10 μM sphingosine. Both reagents had similar effects. After 2 h, cell dye exclusion was not different from that in control cells [92 ± 2.6% (SE)]. At 4 h, viable cells decreased to 83 ± 2.3%, and at 6 h, to 76 ± 6.9%. These changes are comparable to those reported for similar concentrations of ceramide and sphingosine in media both with and without added albumin or serum (8, 10, 17, 39).
Sphingomyelinase (EC 18.104.22.168) activity.
We measured acidic and neutral sphingomyelinase activity in H441 cells using the method described by Quintern and Sandhoff (32). Assay of neutral sphingomyelinase activity was performed in a buffer containing 20 mM HEPES (pH 7.4), 10 mM MgCl2, 2 mM EDTA, 5 mM dithiothreitol, 10 mM β-glycerophosphate, 750 μM ATP, 1 mM phenylmethylsulfonyl fluoride, 10 μM pepstatin, 10 μM leupeptin, and 0.2% Triton X-100. After incubation for 5 min on ice, cells were homogenized by repeated passage through an 18-gauge needle. Nuclei and cell debris were removed by centrifugation at 800 g. The protein concentration in the cell lysate was measured with a bicinchoninic acid assay (Pierce) with BSA as standard. Fifty micrograms of protein were incubated for 2 h in a shaking water bath at 37°C in a buffer (100 μl final volume) containing 20 mM HEPES (pH 7.0), 1 mM MgCl2, and 0.02 μCi [N-methyl-14C]sphingomyelin. At the end of the reaction period, the [14C]phosphosphocholine produced from [14C]sphingomyelin was extracted with 800 μl of chloroform-methanol (2:1 vol/vol) and 250 μl of water. [14C]phosphosphocholine in the upper water phase was determined by scintillation counting. Acidic sphingomyelinase activity was measured identically but in a buffer of 250 mM sodium acetate and 1 mM EDTA (pH 5.0).
PC:ceramide phosphocholinetransferase activity.
We assayed phosphocholinetransferase activity according to Marggraf and Kanfer (24). Cells were lysed in hypotonic, cold 1 mM MgCl2 and homogenized with a Dounce homogenizer. After a low-speed centrifugation for 5 min to pellet cell debris, the supernatant was further centrifuged at 100,000 g for 30 min at 4°C. The pellet containing the membrane fraction was used for the enzyme assay. Phosphocholine transferase activity was determined by measuring the quantity of 3H-labeled sphingomyelin produced from phosphatidyl[N-methyl-3H]choline substrate. The reaction mixture contained 0.05 μCi of phosphatidyl[N-methyl-3H]choline, 10 mM HEPES (pH 7.4), 3 mM MnCl2, 1% fatty acid-free BSA, and 50 μg of membrane protein in a total volume of 100 μl. In some experiments, we also added 2 mg/ml of C8 ceramide to ensure that a substrate limitation would not affect the results (24). The incubations were carried out for 6 h in a shaking water bath at 37°C and were terminated by the addition of 1 ml of 0.2 M methanolic NaOH. The contents were heated at 55°C for 30 min to hydrolyze glycerolipids and cooled to 20°C. Fifteen micrograms of cold sphingomyelin carrier were added to each sample, immediately followed by the addition of 0.5 ml of 0.45 M HCl. The lower organic phase containing sphingomyelin was separated by TLC on silica gel G with two solvent systems: solvent 1, chloroform-methanol (95:5 vol/vol), and solvent 2, methanol-2-propanol-0.25% KCl-triethylamine (9:25:6:18 vol/vol). Radiolabeled sphingomyelin was scraped and counted by liquid scintillation counting. Activity in the cells at time 0 was taken as the control value.
CDPcholine:1,2-diacylglycerol cholinephosphotransferase (EC22.214.171.124) activity.
We used the method of Cornell (9). We scraped cells and resuspended them in 0.5 ml of a buffer of 20 mM Tris, pH 7.0, 0.1 M NaCl, 1 mM dithiothreitol, and 1 mM phenylmethylsulfonyl fluoride. We homogenized the cell suspension with a Dounce homogenizer and sonicated it for 10 min at 4°C. The lysate was centrifuged for 5 min at 800g, and the supernatant was further centrifuged at 100,000g for 60 min. The membrane fraction was resuspended in 75 μl of homogenization buffer and used in the enzyme assay. The reaction contained 50 mM Tris · HCl, pH 8.5, 10 mM MgCl2, 0.5 mM EGTA, 2.4 mM sn-1,2-diolein in 0.15% Tween 20, 0.05 mCi of CDP[methyl-14C]choline, and 50 μl of membrane suspension in a final volume of 100 μl. We incubated the reaction mixture at 37°C in a shaking water bath for 30 min and terminated the reaction with the addition of 1.5 ml of methanol-chloroform (2:1 vol/vol). We extracted 14C-labeled PC into the organic phase of a Bligh-Dyer solvent partition and quantified it with scintillation counting. The results were normalized for the amount of protein in the assay mixture.
CT (EC 126.96.36.199) activity.
The assay for CT activity was done on the membrane fraction with the method of Vance et al. (41) as modified by Weinhold and Feldman (43). We incubated 5–20 μg of protein in a shaking water bath for 1 h at 37°C in the reaction buffer [50 mM Tris-Cl (pH 6.5), 12 mM magnesium acetate, 10 μl of [methyl-14C]phosphocholine, and 3 mM CTP] in a total volume of 100 μl. At the end of the reaction time, we added 400 μl of water and 500 μl of charcoal saturated with 20 mM phosphosphocholine. After a vigorous mixing, the contents were centrifuged at 2,800 rpm, and the supernatants were passed over small charcoal columns. The charcoal, which adsorbs [14C]CDPcholine, was then washed with 1 ml of water (3 times) to remove nonspecific radioactivity. The adsorbed radioactive CDPcholine was eluted with 500 μl of formic acid and counted in a scintillation counter. The assay was linear over a concentration of 0–7 μg of protein (r 2 = 0.9). In initial experiments, we compared activities in the membrane fraction either with or without the addition of 100 μM sonicated egg PC-oleic acid (1:1 molar ratio) to the assay mixture. The results in both protocols were nearly identical, and the data reported here are from assays without added liposomes. Liposomes added to the cytosolic fractions markedly enhanced measured activities as expected.
The method used was according to Krug and Kent (19). We scraped and homogenized cells in a buffer of 0.4 M Tris · HCl, pH 7.3, 50 mM CaCl2, 1 mg/ml of BSA, and 0.6 M NH4SO4. We incubated ∼50 μg of cell protein with a substrate of 1,2-di[1-C14]hexadecanoyl-l-3 PC and conducted the reaction for 1 h at 37°C. We terminated the reaction by adding 0.2 M Na2EDTA, pH 7.3, and extracted the lipids. We added carrier lipids of diacylglycerol (DAG) and lysophosphatidylcholine (lysoPC) and resolved them with TLC on silica gel thin-layer plates. DAG was separated with the chromatographic system of Freeman and West (13), with a solvent system of ethyl ether-benzene-ethanol-acetic acid [40:50:2:0.2] followed by ethyl ether-hexane [6:94]. 1,2-DAG was clearly separated from 1,3-DAG and cholesterol in this solvent system. The lipids were visualized by spraying with 2-(p-toluidino)naphthalene-6-sulfonic acid (TNS), and the 1,2-DAG was spot scraped and counted. LysoPC was separated with the method of Touchstone et al. (40) as described inQuantification of phospholipids. Phospholipase (PL) C was assayed as the amount of radioactive DAG formed per hour per microgram of protein; PLA2 was assayed as the amount of lysoPC.
Chromatography of sphingolipids.
Sphingomyelin, ceramide, and sphingosine, when needed together for the analysis of an experiment, were isolated by the following procedures. Extracted lipids were subjected to mild base hydrolysis as described by Marggraf and Kanfer (24). Sphingolipids were isolated from hydrolyzed glycerolipids by solvent partition and separated by TLC with chlororoform-acetone-methanol-acetic acid-water (150:40:20:20:10 vol/vol). Sphingomyelin, ceramide, and sphingosine pools were localized by TNS spray and, if radioactive, were scraped and quantified by measuring the amount of radioactivity. In this chromatographic system, sphingomyelin, C6 ceramide or endogenous ceramides, C2 ceramide, and sphingosine were resolved.
Quantification of phospholipids.
We quantified PC by gas chromatography after extraction, separation by TLC, and quantitative transesterification with an internal standard of diheptadecanoyl PC. PCs were isolated with the following protocol. Lipids in cells and media were extracted and dissolved in a small volume of 1:1 (vol/vol) chloroform-methanol. The phospholipids were separated by silica gel TLC with the following chromatographic system: first, 95:5 (vol/vol) chloroform-methanol to remove neutral lipids followed by the solvent system B of Touchstone et al. (40). Phospholipids were detected by a TNS spray, and the PC band was recovered by scraping and was transesterified with 1% H2SO4 in methanol for 2 h at 70°C together with an internal standard of a known amount of diheptadecanoyl PC. Methyl esters were extracted with hexane, and PC content was quantified by gas-liquid chromatography.
Reagents were from Sigma or Calbiochem. Radioactive materials were from Amersham or New England Nuclear.
Statistical evaluation was done by ANOVA with the use of Statview (Abacus Concepts, Berkeley, CA). Significance was assumed whenP < 0.05, with P < 0.1 shown for informational purposes.
Effects of ceramide on endogenous sphingolipid content.
We first showed that H441 cells respond to TNF-α with the expected activation of sphingomyelinases. The results are shown in Fig.1. A relatively low concentration of TNF-α (10 ng/ml) activated both acidic and neutral sphingomyelinases to >150% of control values within 2 h after administration, and these activities remained elevated for 4 h. We next investigated the effects of ceramide on sphingomyelinase activity, thinking that ceramide might limit the extent of sphingomyelin hydrolysis. The results in Fig. 1 indicate the opposite: 10 μM C2ceramide activated both acidic and neutral sphingomyelinases as effectively as 10 ng/ml of TNF-α, with maximum activation from 2 through 4 h. The effect was dose dependent and specific because dihydro-C2 ceramide had no effect (dihydro-C2ceramide, 103% of control value at 2 h).
Because of the somewhat surprising finding that ceramide itself activated sphingomyelinases when tested in vitro, we wished to confirm in vivo that the rate of hydrolysis of sphingomyelin was increased in cells given ceramide. We labeled cells overnight withl-[3-3H]serine. Afterward, we removed the medium containing the serine label and replaced it with medium containing 10 μM C2 ceramide or solvent (there was nol-[3-3H]serine in the medium at this point) and followed the changes in radiolabeled sphingomyelin for 24 h. The results are shown in Fig. 2. Cells given C2 ceramide had contents of radiolabeled sphingomyelin that decreased more rapidly in the first 5 h of the experiment than did those of control cells, consistent with an activation of sphingomyelinases by ceramide. By 1 h, the level of radiolabeled sphingomyelin was lower in the treated than in the control cells. The results of corresponding analyses of the endogenous pool of ceramide (C2 ceramide was excluded by chromatography) were consistent with these findings; endogenous ceramide pools were consistently higher in cells given C2 ceramide than in control cells, but the differences were not statistically significant until 3 h.
The homeostasis between ceramide and sphingomyelin is principally maintained through the balance between sphingomyelinase and PC:ceramide phosphocholine transferase, an enzyme that catalyzes the hydrolysis of PC to DAG and the transfer of choline phosphate to ceramide to resynthesize sphingomyelin (15, 24). The effects of ceramide on this key regulatory activity, measured in vitro, are shown in Fig. 3. C2ceramide (10 μM) inhibited PC:ceramide phosphocholinetransferase to ∼50% of control levels in 4 h; 25 μM C2 ceramide accelerated and intensified the effect (data not shown).
The time course of the effects of C2 ceramide on PC:ceramide phosphocholine transferase was consistent with the possibility that C2 ceramide is metabolized to a product that is then responsible for the inhibition of PC:ceramide phosphocholine transferase. Ceramide has three principal metabolic fates: the synthesis of sphingomyelin by PC:ceramide phosphocholine transferase, the synthesis of complex cerebrosides, and its hydrolysis to sphingosine. Although we recognized that we could not identify and test all possible sphingomyelin metabolites for their effects on PC:ceramide phosphocholine transferase, especially the numerous cerebrosides, we did evaluate the effects of sphingosine. In addition, we tested the effects of TNF-α on PC:ceramide phosphocholine transferase because TNF-α must, necessarily, duplicate the effects of these sphingomyelin metabolites. The results are also presented in Fig.3 and indicate that both 10 μM sphingosine and 100 ng/ml of TNF-α are as effective as 10 μM C2 ceramide in inhibiting PC:ceramide phosphocholine transferase.
We verified the ceramide-induced inhibition of PC:ceramide phosphocholine transferase in intact cells using the procedure described by Luberto and Hannun (20), which measures the rate of sphingomyelin resynthesis after its hydrolysis in vivo. We labeled the sphingolipid pools in H441 cells withl-[3-3H]serine overnight. The next morning, we treated the cells with 200 mU/ml of bacterial sphingomyelinase for 30 min to hydrolyze and reduce sphingomyelin content. Afterward, we removed the medium containing sphingomyelinase and the serine label and replaced it with medium containing 10 μM sphingosine or solvent (control) together with new l-[3-3H]serine and followed the resynthesis of sphingomyelin for 24 h. The results are seen in Fig. 4 A. In control cells, sphingomyelin was reduced by nearly 75% after treatment with sphingomyelinase for 30 min (compare time 0data in Fig. 4 with those in Fig. 2). In the recovery period, sphingomyelin resynthesis was evident within 17 h (P < 0.05 compared with time 0), and sphingomyelin content was restored to pretreatment levels within 17 h. In contrast, in cells given sphingosine, there was no resynthesis of sphingomyelin within 24 h.
We repeated this experiment using 10 μM C2 ceramide in two separate protocols. In the first, we repeated the exact protocol used with sphingosine; l-[3H]serine was in the medium for both the overnight equilibration as well as during the treatment with C2 ceramide. This procedure should provide a reliable estimate of total sphingomyelin and ceramide pools at all times. In the second protocol, we removed thel-[3H]serine during the period when the cells were receiving C2 ceramide and cell samples were being taken. This procedure allowed the analysis of the metabolic fate of prelabeled sphingomyelin. The results in both protocols indicated that C2 ceramide enhanced endogenous sphingomyelinase activity and inhibited the resynthesis of sphingomyelin that was observed in control cells (Fig. 4, B and C).
Finally, we labeled cells to steady state by overnight treatment withl-[3-3H]serine and then challenged them with 10 μM C2 ceramide or 10 μM sphingosine to measure changes in the ceramide, sphingosine, and sphingomyelin pools. The results are seen in Fig. 5. Cells treated with both 10 μM C2 ceramide and sphingosine had higher steady-state amounts of ceramide but decreased levels of sphingomyelin within 3–5 h of the administration of C2ceramide. These effects were increased with longer periods of treatment with ceramide. In both protocols, however, counts per minute in the sphingosine pools were low compared with those in ceramide and were not different in control and treated cells.
Effects of ceramide on PC content and metabolism.
Mallampalli et al. (21) have reported that betamethasone given to adult rats decreased sphingomyelinase activity in lung tissue, and, as expected, this effect was associated with an increase in sphingomyelin and a decrease in sphingosine. Furthermore, they found that sphingosine was a competitive inhibitor of CT, thereby implying that an activation of sphingomyelinase would result in an inhibition of CT. We tested this assumption by performing experiments on the effects of TNF-α on CT activity and on the synthesis of PC. A relatively low dose of TNF-α (10 ng/ml) reduced CT activity to 50% of control values within 6 h, and the effect was persistent through at least 24 h. C2 ceramide and sphingosine were equally effective, but dihydro-C2 ceramide had no inhibitive effect (Fig. 6).
TNF-α and ceramide have been reported to activate PLA2and PLC (45), and this was confirmed (Fig.7). PLA2 was activated within 2 h after treatment with ceramide and PLC or PLD (EC 188.8.131.52; see below) within 2–3 h. Enzyme activities thought to be PLC might actually be mistaken for either PC:ceramide phosphocholine transferase (20) or PLD acting together with phosphatidic acid phosphohydrolase because in our assay system, they both use the same substrate (PC) and generate the same product (DAG). In these cells, as shown above, C2 ceramide inhibited PC:ceramide phosphocholine transferase, suggesting that the ceramide-induced increase in labeled DAG was not from that source. We did not attempt to distinguish between PLC and PLD because this information was not central to the questions posed in this study.
The activation of PC hydrolytic enzymes and the concurrent inhibition of CT would be expected to result in a decrease in the amount of cellular PC and a suppression of PC synthesis. These were the effects that were observed. When cells were exposed to 10 μM C2ceramide for 4 h or more, there was a 22.6 ± 2.8% (SE) average reduction in PC content for times from 4 through 24 h. These small but relatively consistent changes were statistically significant at all times (P < 0.05) and are consistent with another report (4) showing that PC content undergoes smaller changes than would be predicted from the changes in CT. (The results are normalized to the amount of cellular protein present in each sample to account for possible differences in cell number in control and treated cultures.)
We then investigated ceramide-induced changes in the metabolism of cellular PC. Cells were pretreated with 10 μM C2 ceramide for times from 1 to 24 h. They were then given [32P]orthophosphate, and the incorporation of orthophosphate into PC was followed for 5 h. Figure8 A shows the kinetics of the incorporation of [32P]orthophosphate at 1, 3, and 5 h after preexposure to 10 μM C2 ceramide for the times shown. C2 ceramide was present in the medium during this labeling period. These data are from one of four similar experiments. In Fig. 8 B, we have combined the data from the four experiments using the 5-h labeling period. Ceramide reduced the incorporation of [32P]orthophosphate into PC by 40% after preexposure to ceramide for 1 h; it inhibited PC synthesis by 60% after 24 h.
Are the effects of ceramide a nonspecific inhibition of the enzymes of the Kennedy pathway?
We considered that the effects of C2 ceramide on PC synthesis might be from the inhibition of multiple sites in the pathway of de novo synthesis, perhaps symptomatic of a generalized reduction of cell function due to developing apoptosis or necrosis. To assess this, we assayed the activity of the last enzyme in the Kennedy pathway, CDPcholine:1,2-DAG cholinephosphotransferase, reasoning that if the inhibition of PC synthesis by ceramide was nonspecific, this activity should be inhibited as was that of CT. This was not found. C2 ceramide (10 μM) had no demonstrable effect on the activity of CDPcholine:1,2-DAG cholinephosphotransferase at any time period between 2 and 24 h (P > 0.10; data not shown). These results indicate that ceramide does not affect all of the enzymes of the Kennedy pathway and suggest that the results are not due to the nonspecific inhibition of PC from cell death.
In this paper, we present data that directly link two commonly observed features of chronic tissue injury, the elaboration of increased amounts of cytokines and the perturbation of metabolic regulation. We conclude that TNF-α activates pathways that reduce PC concentration in H441 cells and that markedly inhibit the de novo synthesis of new PC. Metabolic products of sphingomyelin probably modulate these effects on PC synthesis, and in these cells, the signaling likely involves the inhibition of CT activity. This regulation through CT may not be ubiquitous for all cells. In a recent paper, Bladergroen et al. (7) studied the effects of C6 ceramide on PC synthesis in Rat-2 fibroblasts. As in our findings, these investigators found that ceramide (C6ceramide) inhibited PC synthesis, although doses of 25 μM or higher were required. Directly in opposition to our results, however, they found that C6 ceramide had no effect on CT; rather it inhibited CDPcholine:1,2-DAG cholinephosphotransferase by >50%. We have no explanation for these differing results. They are unlikely to be a result of technical differences, at least as discerned by the description of the enzyme assay methodology, which was identical to ours. We note that Bladergroen et al. used a fibroblast, whereas our studies were done with immortalized epithelial cells. This raises the interesting possibility that the mode of action of ceramide on PC synthesis may differ with cell type, a surprising but certainly not impossible phenomenon.
We used three separate protocols to demonstrate that sphingomyelin pools are reduced and endogenous ceramide is elevated in cells treated with either C2 ceramide or sphingosine. These protocols were based on steady-state labeling of sphingolipid pools with l-[3-3H]- serine, a technique shown to duplicate results obtained by the conventional assay of sphingolipid mass (10, 11, 36, 38). The findings and conclusions drawn from all three protocols were completely consistent. We have not identified which of the sphingomyelin metabolites is responsible for the inhibition of sphingomyelin resynthesis; indeed, several may be acting. Changes in sphingosine pools were not evident after steady-state labeling for 24 h (Figs. 4 and 5) or with quasi-steady-state labeling for 4 h withl-[3H]serine and treatment with C2 ceramide (data not shown). However, this may not be expected if the sphingosine were to act through the initiation of a signaling pathway or be metabolized to another active sphingomyelin metabolite. In such circumstances, the changes in sphingosine mass might be small and transient. For instance, sphingosine phosphate has been shown to affect the activities of cell cycle intermediates and to regulate mitosis and may participate in other signaling pathways, but the actual sphingosine-derived molecules have not been identified (36).
A perplexing aspect of a purported signaling scheme involving TNF-α is the relatively short time span in which increased levels of active TNF-α are observed (for example see Ref. 33). TNF-α and interleukin (IL)-1β both generally peak within 1 h, whereas the inhibition of surfactant synthesis is notable for days. TNF-α and IL-1β have frequently been shown to initiate transcription of other growth factors (37), and we have shown that this may be responsible for the increased amounts of hepatocyte growth factor seen in chronic lung injury in rats in response to 100% O2 (42). The elaboration of new regulating factors, therefore, may be an explanation of the longer range effects of TNF-α. The results presented here, however, provide yet another mechanism for the duration and intensity of the actions of TNF-α. Ceramide itself, or a further metabolic product of ceramide, modulates the principal enzymes needed for the restoration of sphingomyelin-ceramide homeostasis; i.e., it further activates sphingomyelinases and inhibits PC:ceramide phosphocholine transferase. We have not identified which intermediate(s) is responsible for these actions; it could be sphingosine or one further downstream. Whatever the ceramide metabolite, these actions would be expected to prolong the elevation in the ceramide pool and, consequently, extend the time and duration of the ceramide effect. In one set of experiments, we measured the resynthesis of sphingomyelin from ceramide when C2 ceramide was given as a purported metabolic perturbant and then removed. The inhibited resynthesis was evident through at least 8 h after removal of the C2ceramide from the cell medium (we did not follow the experiment beyond 8 h).
An increasing number of reports are emerging that indicate that changes in PC metabolism are a common feature of the pathophysiology of experimental chronic lung injury (18) or adult respiratory distress syndrome, especially when it is associated with inflammation induced by infection (for examples, see Refs.14, 28, 31). A concomitant elevation of cytokines is generally observed (35). The data presented here suggest that the presence of TNF-α and IL-1β may not only be coincidental but rather causal and self-reinforcing. Moreover, these effects may not be unique to lung cells. The pathways of sphingomyelin metabolism are common to all cells (15,26), and many cells use the Kennedy pathway for the de novo synthesis of PC. We expect, therefore, that other tissues experiencing inflammatory injury may also share the changes in PC and sphingolipid content and metabolism that were found in lung cells.
This work was supported by National Heart, Lung, and Blood Institute Grant HL-52664.
Address for reprint requests and other correspondence: R. J. King, Dept. of Physiology, Univ. of Texas Health Science Ctr., 7703 Floyd Curl Dr., San Antonio, TX 78284-7756 (E-mail:).
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2001 the American Physiological Society