Neutral sphingomyelinase action stimulates signal transduction of tumor necrosis factor-alpha in the synthesis of cholesteryl esters in human fibroblasts.

We have investigated biochemical mechanisms of tumor necrosis factor (TNF)-alpha signaling in cultured human skin fibroblasts. We found that TNF-alpha signaling may involve activation of a cell membrane neutral sphingomyelinase (N-SMase) in that within 2.5-5 min of treatment of cells with TNF-alpha there was a 2-fold increase in the activity of N-SMase compared to control. This reaction led to the hydrolysis of sphingomyelin as evidenced by a decrease in sphingomyelin mass and in the radioactivity associated with [14C]choline-labeled sphingomyelin. This was accompanied by a 4-fold increase in the formation of cholesteryl [14C]oleate within 2.5 min of incubation with TNF-alpha. This reaction also stimulated the mobilization of cell surface-associated [3H]cholesterol and its utilization in the synthesis of [3H]cholesteryl esters via acyl coenzyme-A cholesterol acyltransferase (ACAT). Gas chromatographic analysis revealed that the cellular level of cholesteryl esters increased about 2.5-3-fold following treatment with TNF-alpha compared to control. Cholesteryl ester synthesis was compromised upon incubation of cells with antibody against N-SMase and remained unaltered with TNF-beta and fibroblast growth factor. Furthermore, TNF-alpha-mediated stimulation of cholesteryl ester synthesis was compromised by incubation of cells with an inhibitor of ACAT. These findings suggest a possible biological role of N-SMase in the signal transduction of TNF-alpha in the synthesis of cholesteryl esters in human fibroblasts.

(1FNI-y (4, 5) presumably by activating SMase. O n the other hand, in renal proximal tubular cells gentamicin an aminoglycoside antibiotic, decreases the activity of N-SMase in vitro and in viuo in m a n (6, 7) that contribute to increased levels of sphingomyelin. The TNF-a-mediated stimulation of sphingomyelinase activity is an early event in monocytic differentiation of human leukemic (HL-60) cells (4, 5).
We have recently reported that preincubation of human skin fibroblasts with exogenous purified N-SMase stimulates the * This work was supported by Grant R01-DK-31722 from the National Institute of Health. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. The abbreviations used are: N-SMase, neutral sphingomyelinase; TNF-a, tumor necrosis factor, ACAT, acyl-CoA-cholesterol-fatty acyltransferase (EC 1.1.1.88); FGF, fibroblast growth factor. synthesis of cholesteryl esters (€49). The present studies sought to establish whether TNF-a can activate N-SMase in human skin fibroblasts and whether this activation is accompanied by the stimulation of cholesteryl ester synthesis. Our studies show that TNF-a can stimulate N-SMase activity in fibroblasts that is accompanied by decreased radioactivity associated with sphingomyelin and increased synthesis of cholesteryl esters derived from cell membrane cholesterol.
Human recombinant tumor necrosis factor-a was purchased from Promega (Madison, WI). Human recombinant TNF-P and fibroblast growth factor were purchased from Genzyme (Cambridge, MA). Compound 58-035, an ACAT inhibitor (3-decyldimethylsilyl~-N-[24-methylphenyll-l-phenylethyl propanamide), was provided by Sandoz, East Hanover, NJ. Other chemicals were obtained from Sigma. All experiments were performed in duplicate or triplicate. Each experiment was done two or three times.
Incubation of Human Fibroblasts with %mor Necrosis Factor-a and Measurement of Neutral Sphingomyelinase Actiuity-Human skin fibroblasts were obtained from healthy volunteers. Cells were grown in Eagle's minimum essential medium containing 10% fetal bovine serum (Hyclone, Logan, UT). Approximately lo5 cells were seeded in six-well plastic trays and grown to confluence for 6 days in the growth medium. Next, fresh medium without serum was added, and cells were incubated (2TNF-a) for various time periods. Cells were harvested in Tris glycine buffer, pH 7.2, and stored frozen until N-SMase assays were performed as described previously (6).

Incubation of Cells with ['4CICholine-
To determine the effects of TNF-a on the catabolism of sphingomyelin, cells were labeled with [14Clcholine (5 pCi/ml) for 48 h. Medium was removed, and cells were washed with phosphate-buffered saline. Next, cells were incubated with and without TNF-a (10 units, 60 pg/ml medium) at the indicated time, washed with phosphate-buffered saline, harvested, and centrifuged. The cell pellets were extracted with organic solvents. The individual phospholipids were separated on silica gel HL plates by use of chlorofodmethanollformic acid (65:25:4, v/v) as the developing solvent (10). The chromato plate was calibrated with authentic phospholipid standards. After chromatography, the gel area corresponding to sphingomyelin was scraped, and the radioactivity was measured. The recovery of radioactivity was about 906.
Incubation of Cells with ['"ClOleic Acid-To study the effects of TNF-a, TNF-8, and FGF (acidic) on cholesteryl oleate synthesis, cells were incubated with ["Cloleate (2 pCi/ml, complexed with albumin) for 1 h at 37 "C. Next, cells were incubated with and without various concentrations of TNF-a, TNF-P, and FGF and antibody against N-SMase, as described in individual experiments. Then, the cells were harvested and homogenized, and suitable aliquots were saved for protein assays according to Lowry et al. (11). Total lipids were extracted from the remainder of the cell homogenates and separated by thin layer chromatography by use of heptane/ethylether/acetic acid (90:15:1, v/v) as the developing solvent (12).  Incorporation of PHlCholesterol into Cellular Plasma Membranes "In order to label the cellular plasma membranes with [3Hlcholesterol, fibroblast cultures were incubated with [3Hlcholesterol (5 pCi/ml) for 1 h at 37 "C. The cells were washed with medium. Then they were incubated with TNF-a. The incorporation of [3Hlcholesterol into cholesteryl esters was measured as described above. In another set of dishes, cells were incubated with ACAT inhibitor at a final concentration of 50 p g / d in ethanol for 30 min prior to the addition of the TNF-a. Following incubation for 1 h at 37 "C, cholesteryl ester synthesis was measured as described above.
Gas Liquid Chromatography of Cholesteryl Esters-The level of cholesteryl esters in cells incubated with *TNF-cx was measured using stigmasterol as an internal standard following minor modifications (13) of the procedures of Ikekawa et al. (14).

Measurement of Phospholipid Levels in Cells Following Incubation
with TNF-w-The level of phosphatidylcholine, and sphingomyelin was measured in cells incubated fTNF-a according to the procedure of Bartlett (15). incubation with TNF-a. Thereafter, the activity of N-SMase did not increase further. In contrast, TNF-B (50 unitdm1 medium) did not alter the activity of N-SMase in fibroblasts (data not shown).

Effects of Time of Incubation with TNF-a on N-SMase Activity, Incorporation of [*4ClCholine into Sphingomyelin, and Phospholipid Levels in Fibroblasts-
When cells were prelabeled with [14C]choline, washed, and incubated with TNF-a, within 2.5 min we observed a 30% decrease in the radioactivity incorporated into sphingomyelin compared with control (Fig. 1B). Incubation of cells with TNF-a up to 60 min further decreased the incorporation of [14C]choline into sphingomyelin on the order of 40% compared with control. TNF-a concurrently and consistently decreased the cellular levels of sphingomyelin on the order of 30% over the 2.5-60 min period compared to control (Fig. 1C). For

Effects of Time of Incubation with TNF-a on the Synthesis of Cholesteryl Ester from Exogenous [14C101eate-
We observed a marked increase in the formation of cholesterol oleate (derived from exogenous [14C]01eate) within 2.5-5 min of incubation of cells with TNF-a (Fig. 2). In that, we observed a 4-fold increase in the formation of cholesterol [14C]01eate during this time period in cells incubated with TNF-a compared to control. Incubation of cells with TNF-a for 30-60 min did not significantly increase the formation of cholesterol [14C]oleate further (Fig.  2).

Effects of Time of Incubation with TNF-a on the Utilization of Cell Surface Cholesterol for FH]Cholesteryl Ester Synthesis
-As shown in Fig. 3, the synthesis of cholesteryl esters from plasma membrane-derived [3H]cholesterol in control fibroblasts over a period of 60 min was insignificant. In contrast, in cells incubated with TNF-a (10 unitdml) there was a significant time-dependent increase in the utilization of plasma membrane-derived [3H]cholesterol for the synthesis of L3H1cholesteryl esters. Within 2.5 min of incubation of cells with TNF-a there was a rapid increase in the synthesis of cholesteryl esters from cell membrane derived [3H]cholesterol. Thereafter, there was a modest increase in cholesteryl ester synthesis in cells incubated with TNF-a up to 60 min. Fig. 4, TNF-a exerted a concentration-dependent increase in the activity of N-SMase in fibroblasts. The maximum increase in N-SMase activity (a 2.7-fold increase compared with control) occurred upon incubation of cells with 10 units (60 ng) of TNF-dm1 medium. At higher concentrations of TNF-a (25-50 unitdml medium), there was only a 2-fold increase in N-SMase activity compared with control.  Fig. 1. After incubation for 1 h at 37 "C, cells were harvested, and N-SMase activity was measured as described (6). Two experiments in triplicate were performed. The results of the first experiment are presented. The bars represent the range for individual treatment. The activity of N-SMase in control cells was on the order of 0.7 pmol/h/mg protein.

Effects of TNF-a Concentration on the Activity of N-SMase in Fibroblasts-As shown in
Effects of TNF-(I; TNF-p, and FGF (acidic) Concentration on the Synthesis of Cholesteryl Oleate-In human skin fibroblasts, TNF-a increased the synthesis of cholesteryl oleate as a function of concentration (Fig. 5) of incubation. Amaximum of 5-fold increase in cholesteryl oleate synthesis compared with control occurred upon incubating cells with 50 units (300 ng) of TNFd m 1 medium for 1 h at 37 "C (Fig. 5). In contrast, TNF-/3 and FGF did not alter the synthesis of cholesteryl oleate in these cells (Fig. 5).
Effects of TNF-a Concentration on the Cellular Level of Cholesteryl Esters-& shown in Fig. 6, in normal human fibroblasts TNF-ru exerted a concentration-dependent increase in the mass of cholesteryl esters. Maximum increase in cholesteryl ester levels (2.7-fold higher than control) occurred upon incubation with 25 units (150 ng) of TNF-dm1 medium. Incubation of cells with higher concentration of TNF-a did not increase the mass of cholesteryl ester in cells, further.
Effects of ACAT Znhibitor a n d Co-incubation with N-SMase and TNF-a on the Synthesis of Cholesteryl Esters-Incubation

DISCUSSION
The major findings of this study are summarized in the hypothetical model in Fig. 7. First, exogenous TNF-a increases the endogenous activity of N-SMase in human fibroblasts. Second, cell membrane sphingomyelin is hydrolyzed to ceramide and phosphocholine due to N-SMase action. Third, the depletion of cell membrane sphingomyelin may lead to the mobilization of cell membrane cholesterol. Fourth, the membranederived cholesterol is converted to cholesteryl esters via ACAT.
Previous studies have shown that sphingomyelin has a high affinity for cholesterol and keeps cholesterol in the plasma membrane compartment of mammalian cells (1). Previous studies have also revealed a strong relationship between N-SMase action and cholesteryl ester synthesis (8,9,16)  studies. Human skin fibroblasts have an approximately equal amount of cholesterol and sphingomyelin (70 pg/mg protein), and TNF-a: and N-SMase action can decrease the level of sphingomyelin on the order of 30% (as shown in Fig. 1C) and up to 50%, respectively (9). Thus, about 21-35 pg of sphingomyelin may be hydrolyzed upon treatment of cells with TNF-a or N-SMase, respectively. Assuming that an approximately equal amount of cholesterol was removed from the cell surface, then almost 10 pg was used for cholesteryl ester synthesis (Fig. 6). Such findings are consistent with previously published data (9, 16) in which exogenous sphingomyelinase was also shown to decrease cellular levels of sphingomyelin and stimulate cholesteryl ester synthesis in an approximately proportional fashion. Moreover, when we prelabeled fibroblasts with [3Hlcholesterol, that has been previously shown to be esterified by ACAT (16), and incubated such cells with TNF-a, within 2.5 min (when almost maximum stimulation of N-SMase had occurred) (Fig. hi), we observed a marked increase in [3Hlcholesteryl ester formation (Fig. 3). During this time period, we also ob- Other studies have shown that cholesterol released from the cell surface due to N-SMase-mediated depletion of sphingomyelin levels is not egressed in the presence of high density lipoproteins in the cultured medium (18). Rather, it is destined for esterification via ACAT. This tenet is further supported by our finding that ACAT inhibitor could compromise the stimulatory effect of TNF-a. Previously, a dramatic increase in the activity of ACAT was reported in fibroblasts incubated with exogenous N-SMase (16). Thus, the cell surface cholesterol is utilized for cholesteryl ester synthesis via ACAT. However, we and others have previously shown that the catabolic products of sphingomyelin, e.g. ceramide, phosphocholine, and sphingosine, have no effects on cholesteryl ester synthesis (9, 16).
Our studies reveal that antibody against N-SMase (2) compromised the stimulatory effects of TNF-a on cholesteryl ester synthesis. We have previously shown that antibody against N-SMase does not alter cell viability (9). Moreover, incubation of cells with exogenous N-SMase plus corresponding antibody compromised the stimulatory effects of N-SMase on low density lipoprotein receptor activity and cholesteryl ester synthesis (9).
Whereas, preimmune serum I g G did not alter this reaction.
Studies from our laboratory in human PT cells (19) and in human skin fibroblasts' reveal that N-SMase is localized on the surface of these cells. Such topology may provide TNF-a direct access to N-SMase. However, recently Shutze et al. (20) have suggested that TNF-a-mediated sphingomyelin breakdown may be carried out by an acidic sphingomyelinase. On the other hand, the cytosolic origin of N-SMase in HL-60 cells has been ~u g g e s t e d .~ Clearly, studies are required to address this issue further.
In summary, our findings indicate that by controlling the microenvironment of the cell surface in which cholesterol and sphingomyelin reside, biomodulators (e.g. TNF-a) but not TNF-p and FGF may facilitate the regulation of cholesterol and sphingomyelin metabolism. This is an important finding as it suggests a possible biological role of cell surface N-SMase in the signal transduction of TNF-a in normal human fibroblasts. Further studies will be required to explain the detailed biochemical mechanisms involved in this process. Such studies are underway in our laboratory.