Inhibition of acyl-CoA: cholesterol acyltransferase in chinese hamster ovary (CHO) cells by short-chain ceramide and dihydroceramide

BB

ELSEVIER

Biochimica et Biophysica Acta 1256 (1995) 39-46

Biochi~ic~a et BiophysicaActa

Inhibition of acyl-CoA:cholesterol acyltransferase in chinese hamster ovary (CHO) cells by short-chain ceramide and dihydroceramide * Neale D. Ridgway * Departments of Pediatrics and Biochemistry and Atlantic Research Centre, Dalhousie University, 5849 Unit~ersityAvenue, Halifax, Nova Scotia, B3H 4H7, Canada Received 19 May 1994; revised 17 October 1994; accepted 23 December 1994

Abstract

The biological activity of ceramide, an intermediate in the synthesis and catabolism of sphingolipids, has been shown to be mimicked by short-chain N-acyl analogues. A potential role for ceramide in modulating cholesterol esterification was investigated using a series of short-chain ceramides and dihydroceramides. AcyI-CoA:cholesterol acyltransferase (ACAT) in CHO cells was inhibited rapidly ( < 30 min) and in a dose-dependent fashion by two N-acyl analogues of naturally occurring D-erythro-ceramide, N-acetyl-sphingosine (o-erythro-C2-ceramide) and N-hexanoyl-sphingosine (D-erythro-C6-ceramide). At 10 /xM D-erythro-C2-ceramide, esterification of cholesterol was inhibited by 95% in CHO cells grown in delipidated serum, and 80-85% in cells grown in 25-hydroxycholesterol or human low-density lipoprotein (LDL). D-erythro-C2-Ceramide did not inhibit [14C]oleate-labelling of triacylglycerol and phospholipid. Inhibition of cholesterol esterification in cells and isolated membranes required the D-erythro (2S,3R) configuration (the L-threo isomer of C2-ceramide was not inhibitory) and an N-acyl group (sphingosine and sphinganine did not inhibit). DL-erythro-C2-Dihydroceramide was also a potent ACAT inhibitor in isolated membranes (IC50 0.2 ~M) and cells indicating lack of requirement for a 4-trans double bond. Consistent with results for C2-ceramides, DL-threo-C2-dihydroceramide was not inhibitory in cells or in vitro. Long-chain ceramide and N-palmitoyl-dihydroceramide did not inhibit ACAT in isolated membranes. Compared to D-erythro-C2-ceramide, D-erythro-C 6- and C4-ceramide were slightly weaker inhibitors of ACAT in isolated membranes. Thus, N-acyl chain length could influence inhibition, either by altering the effective concentration of ceramide in membranes or affinity for the ACAT enzyme. Short-chain ceramides and dihydroceramides are the first ACAT inhibitors described with structural similarity to a naturally occurring compound. Keywords: Acyl-CoA:cholesterol acyltransferase; Ceramide; Dihydroceramide; CHO cell

1. Introduction

Ceramide, the common precursor of sphingomyelin and complex glycosphingolipids, consists of a long-chain base (usually D-erythro-sphingosine) and an N-acylated fatty acid. As part of the synthesis of new sphingolipids, ceramide is made in the endoplasmic reticulum and trans-

~ This work was supported by a grant and scholarship from the Medical Research Council of Canada, and an establishment grant from Izaak Walton Killiam Children's Hospital. Abbreviations: ACAT, acyl-CoA:cholesterol acyltransferase; BSA, bovine serum albumin; C8-APPD, L-threo-2-(N-octanoyl-amino)-lphenyl-l,3-propanodiol; C2-ceramide, N-acetyl-sphingosine; C2-dihydroceramide, N-acetyl-sphinganine; C4-ceramide, N-butyl-sphingosine; C 6ceramide, N-hexanoyl-sphingosine; L-MAPP, L-threo-2-(N-myristoylamino)-l-phenyl-l-propanol; D-MAPP, D-erythro-2-(N-myristoyl-amino)1-phenyl-l-propanol; TNF, tumor necrosis factor. * Corresponding author. Fax: + 1 (902) 494 1394. 0005-2760/95/$09.50 © 1995 Elsevier Science B.V. All rights reserved SSDI 0 0 0 5 - 2 7 6 0 ( 9 5 ) 0 0 0 0 9 - 7

ported to the Golgi apparatus where it is a substrate for sphingomyelin synthase [1] and various glycosyltransferases [2]. Sphingolipids are transported through the Golgi via vesicle-mediated processes and deposited in the plasma membrane [3,4]. However, there is evidence for alternate routes of sphingomyelin transport and synthesis [5,6], as well as evidence for the localization of sphingomyelin synthase activity in the plasma membrane and Golgi apparatus [1]. Since 6 0 - 9 0 % of sphingomyelin is found in the plasma membrane of cultured cells [7,8], there appears to be a mechanism for compartmentalization of this lipid with little redistribution to other intracellular membranes. There is mounting evidence of a relationship between the cholesterol and sphingolipid content of cells. Such a relationship was originally observed for various disease states [9,10]. Sphingomyelin is known to accumulate in atherosclerotic plaques [11], and in tissues from individuals with Niemann-Pick disease Type II [12], both primary

N.D. Ridgway / Biochimica et Biophysica A cta 1256 (1995) 39-46

40

disorders of cholesterol metabolism. In the Niemann-Pick Type I disorder, the result of acid sphingomyelinase deficiency, there is a modest increase in tissue cholesterol [12]. At the cellular level, plasma membrane sphingomyelin is key to maintaining the cellular gradient of cholesterol, which is also enriched in plasma membrane [13], and indirectly controlling cholesterol regulation. This has been demonstrated by the hydrolysis of plasma membrane sphingomyelin with bacterial sphingomyelinase with resultant mobilization of a cholesterol pool that was esterified and down-regulated cholesterol synthesis [7]. Similarly, activation of sphingomyelin hydrolysis by tumour necrosis factor (TNF)-c~ increased cholesterol esterification [14]. Artificially elevating the sphingomyelin content of fibroblasts with sphingomyelin liposomes has the opposite effect: increased cholesterol synthesis and reduced esterification [15]. Thus, varying the concentration of sphingomyelin seems to directly effect sterol balance in the cell. Preferential association of cholesterol and sphingomyelin has been demonstrated in model membranes [16] and inferred from studies showing co-localization in plasma membrane [7,8,13]. Depletion or supplementation of cellular SM would also increase the concentration of sphingolipid metabolites such as ceramide and long-chain bases. As an example, treatment of cells with bacterial sphingomyelinase or cytokines have been shown to transiently elevate cellular ceramide concentrations [17]. However, the involvement of ceramide in cholesterol mobilization, regulation of cholesterol synthesis and ACAT activation has not been addressed. In this study, a role for ceramide in regulating cholesterol esterification was investigated using short-chain ceramide and dihydroceramide analogues. Short-chain (dihydro)ceramides 2 have been used extensively to mimic the activity of endogenous ceramide in other biological systems [17,18]. Inhibition of A C A T activity by short-chain (dihydro)ceramide analogues was investigated in cultured CHO cells and in vitro. Inhibition was dependent on the stereochemical configuration and N-acyl chain length of the (dihydro)ceramides.

and [1-14C]oleoyl-CoA were from DuPont-New England Nuclear. Silica-gel G thin-layer chromatography plates were from BDH. Acetic and hexanoic anhydride were purchased from Aldrich Chemical. I>MAPP, L-MAPP and C8-APPD were gifts from Dr. Yusuf Hannun, Division of Hematology/Oncology, Duke University Medical Centre, Durham, NC. LDL (1.018-1.063 g / m l ) was prepared by ultracentrifugation of donor human plasma supplied by the Canadian Red Cross. Fetal calf serum was delipidated by ultracentrifugation following density adjustment to 1.21 g / m l with NaBr. Delipidated fetal calf serum was dialyzed against phosphate buffered saline (150 mM NaC1, 10 mM sodium phosphate, pH 7.4), sterile filtered and stored at - 2 0 ° C.

2.2. Cell culture CHO-K1 cells were grown in monolayer at 37°C in an atmosphere of 5% CO 2 in Dulbecco's modified Eagle's Medium containing 5% fetal calf serum (medium A) supplemented with proline (34 /xg/ml). Unless otherwise indicated 60 mm dishes were seeded at a density of 175000 cells in 3 ml of medium A on day 0. On day 3 medium was exchanged for 2 ml Dulbecco's modified Eagle's Medium containing 5% delipidated fetal calf serum (medium B) supplemented with proline (34 /xg/ml). Experiments were started on day 4, 18-24 h after addition of medium B. (Dihydro)ceramides and analogues were dissolved in ethanol and added to warm medium containing serum. Ethanol concentrations did not exceed 0.1% and controls contained ethanol solvent alone. In some experiments ceramides were prepared as a 1:1 (mol/mol) complex with BSA [19]. The complex was prepared by adding C2-ceramide or related compound, dissolved in ethanol at 10 mM, to 1 mM bovine serum albumin in Tris-HC1 (pH 7.4). After mixing, the 1 mM complex of BSA and (dihydro)ceramide was warmed at 37°C for 10 min and added directly to cultured cells. Both methods of (dihydro)ceramide presentation to cells gave similar results.

2.3. Preparation of (dihydro)ceramides 2. Materials and methods

D-erythro-Sphingosine, 2.1. Materials D-erythro-Sphingosine, DL-erythro-dihydrosphingosine, DL-threo-dihydrosphingosine, ceramide (from bovine brain sphingomyelin) and BSA (fatty acid-free) were obtained from Sigma Chemical. N-Palmitoyl-dihydroceramide was purchased from Serdary, London, Ont. 25-hydroxycholesterol was from Steraloids, Wilton, NH. [l-14C]Oleate

1 Refers to c e r a m i d e s and d i h y d r o c e r a m i d e s .

oL-erythro-dihydrosphingosine

and DL-threo-dihydrosphingosine were acylated by treatment with acetic or hexanoic anhydride as previously described [20,21]. C 2- and C6-ceramides or dihydroceramides were purified by thin-layer chromatography in diethyl ether/methanol (9:1, v / v ) and stored in 2 m g / m l solutions in chloroform/methanol (19:1, v / v ) at - 2 0 ° C under N 2. The threo isomer of C2-ceramide was prepared by oxidation of acetylated D-erythro-sphingosine with chromic anhydride in pyridine [20]. The 3-keto product was purified by thin-layer chromatography in chloroform/methanol (93:7, v / v ) and reduced with NaBH 4 in alkaline methanol [20,21]. The racemic mixture

41

N.D. Ridgway / Biochimica et Biophysica Acta 1256 (1995) 39-46

was resolved into D-erythro and L-threo components by thin-layer chromatography in diethyl ether/methanol (9:1,

~

v/v).

a~ o l . 5

2.4. A C A T assays

The incorporation of [1-14 C]oleate into cholesteryl[1lnC]oleate in cultured CHO cells was measured as previously described [22]. Briefly, a 10 mM [1-t4C]oleate/BSA complex was prepared (7000-8000 d p m / n m o l ) and added to dishes at a final concentration of 0.1 mM. Cells were cultured for 30 or 60 min at 3 7 ° C and lipids extracted directly from the culture dish with hexane/isopropanol (3:2, v / v ) for 30 min. Lipid extracts were dried under N: and separated by thin-layer chromatography in a solvent system of hexane/diethyl ether/acetic acid (90:30:1, v / v ) . Cholesteryl ester, triacylglycerol and phospholipids were visualized by exposure to iodine vapours or autoradiography and quantitated by liquid scintillation counting. A C A T activity in cell membranes was assayed by measuring the conversion of [1-14C]oleoyl-CoA into cholesteryl[1-14C]oleate [23]. Dishes of cells were rinsed twice with 2 ml of phosphate-buffered saline, scraped in the same buffer and collected by centrifugation at 2000 × g for 5 min. Cells were homogenized in 20 mM Tris-HC1 (pH 7.7) and 10 mM EDTA (buffer A) by 10 passages through a 22-gauge needle. The homogenate was centrifuged at 100 000 × g for 1 h, and the membrane fraction (pellet) resuspended in buffer A using a glass dounce homogenizer. A C A T assays were in a final volume of 100 /xl containing 2 5 - 5 0 p,g membrane protein and 0.2 mg BSA in buffer A. Assays were preincubated for 2 min at 37 ° C, the reaction initiated by the addition of 10 /xl of 5 0 0 / x M [1-14 C]oleoyl-CoA (60 d p m / p m o l ) , incubated for 10 min at 37 ° C and terminated by the addition of 2 ml of chloroform/methanol (1:1, v / v ) . Neutral lipids were extracted by the method of Folch et al. [24], and cholesteryl ester and triacylglycerol were separated and quantitated as described above for the incorporation of [1-~4C]oleate into cell monolayers.

2

aE~

,

,

,

~

.

_\

© []

NA 2S-0H

O [] •

NA 25-OH LDL

O

_oa. o,=

~O.S -5 o

0 1.2

~0.8 o

~"-- 0.4 ~ , . 5 0.2

0

I

2.5

~"

2

~-1.5 ~.~

1

5E

~_,~ 0.5 I

0

2

i

I

i

4 6 8 Ceramide (p_M)

I

10

12

Fig. 1. Inhibition of cholesterol esterification in CHO cells by C 2- and C6-ceramide. Cells were cultured as described in Section 2. Panel A and B, cells received medium B containing 25-hydroxycholesterol(25-OH, 2.5 /zg/ml) human LDL (LDL, 50 p.g/ml) or no addition (NA) for 5 h. D-erythro-C2-Ceramide/BSA complex (0-10 /xM) was added to cells and, after incubating for 1 h, cholesteryl ester and triacylglycerolsynthesis was measured by a 1 h pulse with 0.1 mM [1-14C]oleate. Panel C, cells were treated with medium B containing 25-hydroxycholesterol(2.5 p.g/ml) for 5 h followedby a 1 h incubation with 0-10 p.M O-erythro-C2and C6-ceramide as described for panel A. Triacylglycerol(C2-ceramide [D] and C6-ceramide [A]) and cholesteryl ester (C2-ceramide [0] and C6-ceramide [0]) synthesis measured by a 1 h incubation with 0.1 mM [1-14C]oleate. All results are the means of duplicate determinations from a representative experiment.

3. R e s u l t s

Short-chain analogues of ceramide offer the advantage of increased solubility in aqueous environments and greater uptake by cultured cells compared to their long-chain counterparts. Thus the cellular concentration of 'ceramide' can be elevated independent of processes that increase ceramide endogenously or alter sphingolipid synthesis or catabolism. To determine what effects short-chain ceramides have on cholesterol homeostasis, CHO cells were treated with D-erythro-C2-ceramide and A C A T activity measured by [1-14 C]oleate incorporation into cholesteryl esters. As shown in Fig. 1A, o-erythro-C2-ceramide produced a dose-dependent inhibition of cholesteryl esterifica-

tion. A concentration of 10 /zM D-erythro-C2-ceramide produced 95% inhibition of cholesteryl ester synthesis in cells grown in delipidated serum. The same concentration of D-erythro-C2-ceramide in cells pretreated with LDL or 25-hydroxycholesterol resulted in 80% inhibition of esterification. Synthesis of [1-a4C]-labelled-triacylglycerol was virtually unchanged over the concentration range of Derythro-C2-ceramide (Fig. 1B), as was [1-14 C]oleate-labelling of total phospholipids (results not shown). Another short-chain ceramide analogue, D-erythro-C6-ceramide , also inhibited cholesteryl ester synthesis in a dose-dependent fashion and had little effect on triacylglycerol synthesis (Fig. 1C).

N.D. Ridgway / Biochimica et Biophysica Acta 1256 (1995) 39-46

42

Time dependence of D-erythro-C2-ceramide inhibition of cholesterol esterification is shown in Fig. 2. D-erythroCz-Ceramide (10 /xM) inhibited cholesterol esterification very rapidly. Even at the earliest times tested with a [1-14C]oleate-pulse (30 min), cholesteryl ester synthesis was inhibited by 90%. Shorter incubation times could not be used due to reduced incorporation of [1-J4C]oleate into cholesteryl ester (30 min [1-tac]oleate treatment gave approximately one-half the activity seen at 60 rain). Cholesteryl ester synthesis recovered slightly by 5 h, probably due to metabolism of D-erythro-C2-ceramide. The mechanism and specificity of short-chain ceramide effects on cholesterol esterification in CHO cells were investigated. Inhibition of HL-60 cell growth and induction of differentiation by ceramide and related analogues [25], and inhibition of yeast growth [26], are sensitive to the stereochemical configuration at the two chirai carbons of ceramide. Stereo-specificity is also seen for inhibition of cholesteryl ester synthesis in CHO cells (Fig. 3). Derythro-C2-Ceramide (10 ~M), but not L-threo-Ce-ceramide, inhibited ester synthesis indicating that the naturally occurring 2S,3R configuration is required for activity. Likewise, 10 /zM of a racemic mixture of DL-erythro-C2dihydroceramide, but not DL-threo-C2-dihydroceramide, inhibited esterification to the same extent as D-erythro-C2ceramide. Also shown in Fig. 3 are the effects of three compounds structurally related to ceramide. 2-(N-Myristoylamino)-l-phenyl-l-propanols (MAPP) have been shown to mimic ceramide effects on HL-60 cells [25], but not yeast [26]. D-MAPP, which has the same stereochemical configuration as D-erythro-C2-ceramide at its two chiral centres, inhibited cholesteryl ester synthesis, but LMAPP did not. C8-APPD, which did not mimic D-erythroC2-ceramide effects in HL-60 cells [25], was also an ineffective inhibitor of cholesterol esterification in this study. D-erythro-Sphingosine and DL-erythro-dihydrosphingosine did not significantly inhibit cholesterol esteri-

1.6 ..~ ~ 1 . 4 "~ 1.2

£3

~0.8 ~ "~ 0.6 0 EO.4, %.£= 0.2 0 0

1

2

3 4 Time (h)

5

6

Fig. 2. Time-dependent inhibition of cholesteryl esterification by C 2ceramide. Cell culture and lipid analysis were performed as described in Section 2. Cells were incubated with medium B containing 10 /zM o-erythro-C2-ceramide for up to 5 h. In the last 30 min of each treatment, cells were pulsed with 0.1 mM [1-14 C]oleate and cholesteryl ester ( 0 ) and triaeylglycerol ([2]) synthesis measured. Results are the means of triplicate determinations from a representative experiment.

1.8

~ ~,1.5

l.z

~.. v 0.3 0

!

•

Fig. 3. Stereospecific inhibition of cbolesteryl ester synthesis in CHO cells by ceramide analogues. Cells received 2 ml of medium B containing 10 txM of the various ceramide analogues dissolved in ethanol: NA, no addition (ethanol vehicle); Spo, spbingosine; Spa, sphinganine; E-C2-Cer, D-erythro-C2-ceramide; T-C2-Cer, L-threo-C2-ceramide; E-Cz-DHC, DLerythro-C2-dihydroceramide; T-C2-DHC, DL-threo-C2-dihydroceramide; L-MAPP, u-myristoylaminophenyl propanol; o-MAPP, D-myristoylaminophenyl propanol; C8-APPD, octanoylaminophenyl-l,3-propanodiol. During the last 1 h of a 2 h incubation at 37 ° C the synthesis of cholesteryl ester (shaded) and triacylglycerol (diagonal lines) was measured by a 1 h pulse with 0.1 mM [1-14C]oleate. Results are the means of three experiments + S.D.

fication demonstrating the requirement for an N-acyl group. In these experiments both C2-ceramides and D-MAPP caused a 10-15% increase in triacylglycerol synthesis. However, this increase was not significant or reproduced in other experiments. Cholesteryl ester synthesis is catalyzed by the microsomai enzyme ACAT. Results of [1-~4C]oleate-labelling studies in intact cells suggested this enzyme was directly inhibited by C2-(dihydro)ceramides. When membranes from CHO cells treated with D-erythro-Cz-ceramide for 2 h were assayed for ACAT activity using [1-14 C]oleoyl CoA, significant inhibition of activity was seen for 3 different culture conditions that effect ACAT activity and sterol balance; delipidated serum, LDL addition and 25-hydroxycholesterol addition (Fig. 4). Consistent with results from intact cells (Fig. 1), greater inhibition was observed in membranes from cells grown in delipidated serum (80%) then in LDL-treated cells (60%). Treatment of cells for 12 h with 25-hydroxycholesterol increased ACAT activity in membranes; however, the degree of inhibition by Derythro-C2-ceramide was the same as cells grown in delipidated serum. As shown in Fig. 4, D-erythro-Cz-ceramide displays potent and persistent inhibition of ACAT. This inhibition could be duplicated by adding D-erythro-C2-ceramides (10 /xM) directly to ACAT assays of CHO cell membranes from cells grown in human LDL (Fig. 5A) or fetal calf serum (Fig. 5B). It should be noted that ACAT specific activity in membranes from cells grown in delipidated fetal calf serum was 20 p m o l / m i n per mg protein (Fig. 4),

N.D. Ridgway / Biochimica et Biophysica Acta 1256 (1995) 39-46 60 .~_ 50

.:..:..:..:

D o

:.L; " 40

¢..'¢.:¢/~

?.-.--.:......-.

< 30 ~--.~_ ~ < E 20

?:'(6,.'<

=:.:.:.:.:.:.:=?.'.:

E 310

NA-

NA+

LDL+

LDL-

25-0H- 25-0H+

Fig. 4. ACAT activity in membranes from C~-ceramide-treated CHO cells. CHO cells were incubated in medium B containing 25-hydroxycholesterol (25-OH, 2.5 txg/ml), human LDL (LDL, 50 /xg/ml) or no addition (NA). After 6 h, cells received fresh medium B containing LDL, 25-hydroxycholesterol or no addition with ( + ) or without ( - ) 10 /xM I>erythro-C2-ceramide dissolved in ethanol. After 2 h at 37~ C, cells were harvested and ACAT activity assayed in cell membranes as described under Section 2. Results are the means of three separate experiments+ S.D.

120 A

"•1

O0

I--._~

.

o

~

~0

~B

~i o

~i o

~

~

~

43

50-60 p m o l / m i n per mg in whole fetal calf serum (Fig. 5B) and 100 p m o l / m i n per mg in human LDL-supplemented delipidated serum (Fig. 5A). The pattern of in vitro inhibition (Fig. 5A) with the various ceramides is almost identical to results for intact cells (Fig. 3). However, t-MAPP, which did not inhibit in cells, is now seen to have some inhibitory activity in vitro. This inhibition could be non-specific since the threo isomers of C2-ceramide and C2-dihydroceramide, and the two long-chain bases, also display some inhibitory activity (10-20%) compared to untreated membranes. As was the case with cultured cells, D-erythro-C2-ceramide and DL-erythro-C2-dihydroceramide inhibited ACAT activity in isolated membranes by 90-95%. Long-chain ceramide and dihydroceramide were tested as ACAT inhibitors in isolated membranes (Fig. 5B). Ceramide (derived from bovine sphingomyelin) and synthetic N-palmitoyl-dihydroceramide did not inhibit ACAT when dissolved in ethanol and added directly to the assay. D-Erythro-C 6- and C4-ceramide (10 /xM) were also less effective ACAT inhibitors compared to o-erythro-C 2ceramide. Short-chain ceramides were found not to inhibit ACAT activity via activation of protein phosphatase 2A [27] and subsequent enzyme dephosphorylation since okadaic acid (a potent protein phosphatase 2A inhibitor) did not prevent D-erythro-C2-ceramide inhibition of ACAT in cells or in vitro (results not shown). Also, oL-erythro-C2-dihydroceramide has been shown not to activate protein phosphatase 2A [27], but this analogue clearly inhibits ACAT activity to the same extent as D-erythro-C2-ceramide (Figs. 3 and 5). ACAT activity was titrated with increasing concentrations of C2-ceramide and C2-dihydroceramide at saturating oleoyl-CoA (50 /~M) to determine the IC50 in isolated

70 ,-60 e-

B l

11

I

50

~8 >

~40

~~" ¢j"

.;--.?..':?::;-

120 _>, i~ 1oo "5 < 8O I--

..?:.:i.:--:.-

~ 80

~.:.:::

~

~?~?:?~

<

30

g

Q.

"-- 10

?.".777-~~

40 20

0 z

(.)~

t,.jv

u"

,

CI

Fig. 5. Inhibition of ACAT activity in CHO cell membranes by short-chain ceramides and dihydroceramides. Membranes (50 p,g) isolated from CHO cells grown in medium B plus human LDL (50 /.~g/ml, panel A) or medium A (panel B) for 12 h were incubated with various ceramide analogues dissolved in ethanol (10 /xM) or ethanol with no additions (NA) for 2 rain prior to assaying for ACAT activity. Abbreviations are as described in Fig. 3: SM-Cer, ceramide from bovine sphingomyelin; PaI-DHC, N-palmitoyl-dihydroceramide; E-C4-Cer, D-erythro-C4-ceramide; E-C6-Cer, D-erythro-C6-ceramide. Results are the means of three separate experiments_+ S.D.

0.01

0.1 1 10 Ceramide analogue (~tM)

Fig. 6. Concentration-dependent inhibition of ACAT activity in isolated CHO cell membranes by C2-ceramide and C2-dihydroceramide. Membranes were isolated from CHO cells grown 12 h in human LDL (50 txg/ml). Membranes (50 /zg) were incubated with the indicated concentrations of D-erythro-Cz-ceramide ( 0 ) , oL-erythro-C2-dihydroceramide ( [] ), L-threo-C2-ceramide ( • ) or DL-threo-C2-dihydroceramide ( • ) for 2 min prior to starting the ACAT assay by addition of [l-]4C]oleoyl CoA. Ethanol concentrations in all assays were fixed at 1% (v/v). Results are the means of duplicate assays from a representative experiment. The control value for ACAT activity was 107.9 pmol/min per mg protein.

N.D. Ridgway / Biochimica et Biophysica Acta 1256 (1995) 39-46

44

membranes. D-erythro-C2-Ceramide and DL-erythro-C2-dihydroceramide displayed IC50 values of 0.2 and 0.5 /zM, respectively (Fig. 6). D-erythro-C2-Dihydroceramide displayed slightly greater inhibition at lower concentrations, but both compounds inhibited 90% of ACAT activity at 5 txM. L-threo-C2-Ceramide and DL-threo-dihydroceramide inhibited activity by 20% at 10 /xM. Substrate/velocity curves for oleoyl-CoA using CHO membranes were highly irregular thus making it impossible to determine the kinetic mechanism of inhibition by short-chain (dihydro)ceramides.

4. Discussion

Ceramide is a key intermediate in the synthesis and catabolism of sphingolipids, as well as an important lipid effector for some aspects of cell growth and differentiation. Studies of ceramide metabolism are hampered by limited solubility of this lipid in aqueous media and poor uptake by cultured cells. This can be circumvented by replacement of the N-acyl chain of ceramide with an acetyl or hexanoyl group. These short-chain ceramides have proved to be useful in delineating sphingolipid and ceramide localization and trafficking in cultured cells [4,28-30] and in numerous systems have been shown to mimic the activity of endogenous long-chain ceramide in signal transduction processes [18]. In this study short-chain (dihydro)ceramides and a related analogue were shown to inhibit ACAT. Inhibition of ACAT by short-chain (dihydro)ceramides has several important features: (1) Derythro-C2-ceramide [2S,2R] and DL-erythro-C2-dihydroceramide [2S,3R][2R,3S] were effective inhibitors while the L-threo [2S,3S] and DL-threo [2S,3R][2R,3R] isomers were not, (2) an N-acyl group was required for activity since sphingosine and sphinganine were not inhibitory and (3) a 4-trans double bond was not necessary for activity. Short-chain ceramides are inhibitory at concentrations close to endogenous ceramide levels in CHO cells. Treatment of CHO cells or membranes with 5 to 10 /.tM o-erythro-C2-ceramides was found to inhibit esterification by 90-95%. Based on the uptake of the radiolabelled compound, it was calculated that the concentration of D-erythro-C2-ceramide in CHO cells treated with 10 /zM of this analogue was 50 p m o l / n m o l lipid phosphorus [21]. Since it was reported that uptake of D-erythro-C2-ceramide in HL-60 cells is concentration dependent up to 15 /xM [31], inhibitory concentrations of short-chain ceramide are within the range reported for endogenous ceramide in some cultured cells [13]. For example, endogenous ceramide in CHO cells grown in 5% fetal calf serum is 8 - 1 0 p m o l / n m o l lipid phosphorus 3, while addition of 5 /.~M

2 N.D. Ridgway, unpublished results.

D-erythro-C2-ceramide in the culture medium will give an intracellular concentration for this analogue of approx. 25 p m o l / n m o l lipid phosphorus. Inhibitory concentrations of short-chain ceramides are probably reached very rapidly in cultured cells. D-erythro-C2-Ceramide was taken up by 2 - 3 min in HL-60 cells [31] and was maximal by 10 min for D-erythro-C4-ceramide in CHO cells [21]. Rapid uptake of the short-chain ceramides probably results in ACAT inhibition in CHO cells well before the earliest time point (30 min) that could accurately measure activity in this study. Inhibition of ACAT by the D-erythro isomers of C 6 and C2-ceramide suggests a role for endogenous ceramide in ACAT regulation. In this study, no direct attempt was made to elevate endogenous ceramide levels. However, incubation of CHO cells with exogenous long-chain bases might be expected to increase ceramide concentrations since the rate-limiting step in the synthetic pathway is bypassed [32]. In this study, long-chain bases did not effect ACAT activity in cells or in vitro. Whether long-chain bases actually increase ceramide concentrations is unclear. Incubation of Swiss 3T3 cells with 15 /zM D-erythrosphingosine has been shown to elevate ceramide by 6-fold in 4 h [33], but had no effect on ceramide in another study using the same cell line [34]. In vitro addition of ceramide (from bovine brain sphingomyelin) or synthetic N-palmitoyl-dihydroceramide to ACAT assays of CHO membranes did not affect cholesterol esterification. Also, a limited comparison of inhibitory activity and acyl chain length showed C6-ceramide to be a slightly weaker inhibitor than C2-ceramide. These results were not unexpected as longchain ceramides and C6-ceramide are more nonpolar and may not exchange into isolated membranes as readily as C2-ceramide. Similarly, ceramides with N-acyl groups of 14 carbons or longer have been shown to be poor activators of protein phosphatase 2A compared to shorter chain analogues [27]. It was previously shown that hydrolysis of plasma membrane SM results in ACAT activation and an increase in cellular ceramide [7,14]. Increased cholesterol esterification seems to result from liberation of a SM associatedcholesterol pool in the plasma membrane that is available for esterification and feed-back suppression of cholesterol synthesis. Any role for ceramide in modulating ACAT activity has not been investigated. Whereas sphingomyelin hydrolysis increases cholesterol esterification, the present study s h o w s that t r e a t m e n t with short-chain (dihydro)ceramides inhibits ACAT activity. This apparent paradox may be simply explained by assuming endogenous ceramide does not affect ACAT activity and results for short-chain analogues are related to chain length modification and increased solubility. However, other explanations for lack of inhibition by long-chain ceramides are possible. Firstly, it is questionable whether ceramide that is generated at the plasma membrane via sphingomyelin breakdown ever reaches the endoplasmic reticulum where

N.D. Ridgway / Biochimica et Biophysica Acta 1256 (1995) 39-46

ACAT is presumed to reside [35]. If ceramide generated by receptor-stimulated sphingomyelin breakdown is destined for resynthesis to sphingolipids, this reaction can occur in the plasma membrane or Golgi apparatus [36]. A recent study has shown that resynthesis of SM following exogenous sphingomyelinase treatment is restricted to recycling endosomes [37]. Similarly, sphingomyelinase treatment of Jurkat T-cells did not inhibit cell proliferation while ceramide generated by other treatments was inversely correlated with DNA synthesis [38]. Therefore, the sphingomyelin cycle of degradation and resynthesis could be compartmentalized to a limited set of membranes, including the plasma membrane and endocytic pathway, and ceramide may not gain access to other organelles such as the endoplasmic reticulum. Secondly, inhibition of ACAT by ceramide may be over-ridden by increased cholesterol availability to the enzyme. There is some evidence for this in our study as ACAT activity in CHO cells treated with LDL was suppressed to a lesser degree by 10 /zM Derythro-C2-ceramide then cells grown in delipidated serum, and membranes from D-erythro-C2-ceramide-treated cells that received LDL contained relatively more ACAT activity then C2-ceramide-treated controls grown in delipidated serum. These results indicate that the cholesterol concentration within the cell may dictate, to a limited degree, the extent of inhibition by cell permeable ceramide analogues. The partial resistance to D-erythro-C2-ceramide inhibition afforded by increased cellular sterol is not evident in similar experiments using isolated membranes where ACAT activity is inhibited by 90-95% following addition of D-erythro-C2-ceramide and DL-erythro-C2-dihydroceramide to membranes from LDL-treated CHO cells. Thirdly, if ceramide modulates ACAT activity, its effects should be confined to the site of ceramide synthesis in the endoplasmic reticulum [39]. These ideas are speculative but can be tested by examining the intracellular trafficking of ceramide, mass of this lipid in subcellular membranes and other agents that alter ceramide concentrations. In comparison to known ACAT inhibitors, short-chain (dihydro)ceramides are moderately potent and unique because of similarity to naturally occurring compounds. Fatty acid amide (58-035 [40] and 57-118 [41]), fatty acid anilide (PD 128042 [42]) and phenylakylurea [43] inhibitors of ACAT do not have structural similarity to ceramide other then general hydrophobic character and the presence of an fatty acyl amide moiety. Interestingly, malinamide, a linoleamide inhibitor of ACAT [44,45], and D-MAPP [25] both have a N-acyl phenylamino moiety. 57-118 and PD 128042 are competitive inhibitors of ACAT, and have IC50 of approximately 0.2 /xM using isolated microsomes [41,42]. More recently developed inhibitors have IC50 of less than 20 nM [44,45]. In comparison, 0.5 /~M C2-ceramide inhibited cholesterol esterification by 50% in assays of CHO membranes, and C2-dihydroceramide was a slightly more potent (IC50 0.2 /xM). Short-chain cell p e r m e a b l e analogues of

45

(dihydro)ceramide mimic many of the cellular responses evoked by elevation of endogenous ceramide. These compounds are also potent effectors of metabolic pathways, in particular cholesterol esterification, and are degraded to precursors that enter the de novo sphingolipid biosynthetic pathway [21,46]. The potential for short-chain (dihydro)ceramides to affect metabolic pathways makes them useful reagents to study (dihydro)ceramide activity, but caution should be taken when interpreting the action of these compounds given their potent biological activities.

Acknowledgements Thanks to Tom Lagace for providing excellent technical assistance, Robert Zwicker for maintaining and culturing CHO cells and Ketan Badiani for critical review of this manuscript.

References [1] Koval, M. and Pagano, R.E. (1991) Biochim. Biophys. Acta 1082, 113-125. [2] Van Echten, G. and Sandhoff, K. (1993) J. Biol. Chem. 268, 5341-5344. [3] Lipsky, N.G. and Pagano, R.E. (1985) J. Cell Biol. 100, 27-34. [4] Helms, .I.B., Karrenbauer, A., Wirtz, K.W.A., Rothman, J.E. and Wieland, F.T. (1990) J. Biol. Chem. 265, 20027-20032. [5] Kallen, K.-J., Quinn, P. and Allen, D. (1993) Biochim. Biophys. Acta 1166, 305-308. [6] Shiao, Y.-J. and Vance, J.E. (1993) J. Biol. Chem. 268, 2608526092. [7] Slotte, J.P. and Biermann, E.L. (1988) Biochem. J. 250, 653-658. [8] Lange, Y., Swaisgood, M.H., Ramos, B.V. and Steck, T.L. (1989) J. Biol. Chem. 264, 3786-3793. [9] Merrill, A.H., Jr. and Jones, D.D. (1990) Biochim. Biophys. Acta 1044, 1-12. [10] Kolesnick, R.N. (1991) Prog. Lipid Res. 30, 1-38. [11] Bottcher, C.F.J. and Van Gent, C.M.J. (1961) J. Atheroscler. Res. 1, 36-46. [12] Spence, M.W. and Callahan, ,I.W. (1989) in The Metabolic Basis of Inherited Disease (Scriver, C.R., Beaudet, A.L., Sly, W.S. and Valle, D., eds.), Vol. 2, pp. 1655-1676, McGraw-Hill, New York, NY. [13] Lange, Y. (1992) ,i. Lipid Res. 33, 315-321. [14] Chatterjee, S. (1994) J. Biol. Chem. 269, 879-882. [15] Kudchodkar, B.J., Albers, J.J. and Bieman, E.L. (1983) Atherosclerosis 46, 353-364. [16] Mclntosh, T.J., Simon, S.A., Needham, D. and Huang, C.-h. (1992) Biochemistry 31, 2012-2020. [17] Hannun, Y.A. and Linardic, C.A. (1993) Biochim. Biophys. Acta 1154, 223-236. [18] Hannun, Y.A. (1994) J. Biol. Chem. 269, 3125-3128. [19] Pagano, R.E. and Martin, O.C. (1988) Biochemistry 27, 4439-4445. [20] Gaver, R.C. and Sweeley, C.C. (1966) J. Am. Chem. Soc. 88, 3643-3647. [21] Ridgway, N.D. and Merriam, D.L. (1995) Biochim. Biophys. Acta, in press. [22] Goldstein, J.L., Basu, S.K. and Brown, M.S. (1983) Methods Enzymol. 98, 241-260. [23] Metherall, J.E., Ridgway, N.D., Dawson, P.A., Goldstein, J.L. and Brown, M.S. (1991) J. Biol. Chem. 266, 12734-12740.

46

N.D. Ridgway / Biochimica et Biophysica Acta 1256 (1995) 39-46

[24] Folch, J., Lees, M., Sloane-Stanley, G.H. and Randall, R.J. (1951) J. Biol. Chem. 226, 497-50. [25] Bielawska, A., Linardic, C.M. and Hannun, Y.A. (1992) J. Biol. Chem. 267, 18493-18497. [26] Fishbein, J.D., Dobrowsky, R.T., Bielawski, A., Garrett, S. and Hannun, Y.A. (1993) J. Biol. Chem. 268, 9255-9261. [27] Dobrowsky, R.T., Kamibayashi, C., Mumby, M.C. and Hannun, Y.A. (1993) J. Biol. Chem. 268, 15523-15530. [28] Rosenwald, A.G. and Pagano, R.E. (1993) J. Biol. Chem. 268, 4577-4579. [29] Rosenwald, A.G. and Pagano, R.E. (1993) Adv. Lipid Res. 26, 101-142. [30] Karrenbauer, A., Jeckel, D., Just, W., Birk, R., Schmidt, R.R., Rothman, J.E. and Wieland, F.T. (1990) Cell 63, 259-267. [31] Bielawska, A., Crane, H.M., Liotta, D., Obeid, L.M. and Hannun, Y.A. (1993) J. Biol. Chem. 268, 26226-26232. [32] Merrill, A.H., Jr. and Jones, D.D. (1990) Biochim. Biophys. Acta 1044, 1-12. [33] Hauser, J.M.L., Buehrer, B.M. and Bell, R.M. (1994) J. Biol. Chem. 269, 6803-6809. [34] Olivera, A., Buckley, N.E. and Spiegel, S. (1992) J. Biol. Chem. 267, 26121-26127. [35] Suckling, K.E. and Stange, E.F. (1985) J. Lipid Res. 26, 647-671.

[36] Futerman, A.H., Stieger, B., Hubbard, A.L. and Pagano, R.E. (1990) J. Biol. Chem. 265, 8650-8657. [37] Kallen, K.-J., Allen, D., Whatmore, J. and Quinn, P. (1994) Biochim. Biophys. Acta 1191, 52-58. [38] Borchardt, R.A., Lee, W.T., Kalen, A., Buckley, R.H., Peters, C., Schiff, S. and Bell, R.M. (1994) Biochim. Biophys. Acta 1212, 327-336. [39] Mandon, E.C., Van Echten, G., Birk, R., Schmidt, R.R., Sandhoff, K. (1991) Eur. J. Biochem. 198, 667-674. [40] Ross, A.C., Go, K.J., Heider, J.G. and Rothblat, G.H. (1984) J. Biol. Chem. 259, 815-819. [41] Heider, J.G., Pickens, C.E. and Kelly, L.A. (1983) J. Lipid Res. 24, 1127-1134. [42] Field, F.J., Albright, E. and Mathur, S. (1991) Lipids 26, 1-8. [43] Trivedi, B.K., Holmes, A., Stoeber, T.L., Blankey, C.J., Roark, W.H., Picard, J.A., Shaw, M.K., Essenburg, A.D., Stanield, R.L. and Krause, B.R. (1993) J. Med. Chem. 36, 3300-3307. [44] Sliskovic, D.R. and White, A.D. (1991) Trends Pharmacol. Sci. 12, 194-199. [45] Natori, K., Okazaki, Y., Nakajima, T., Hirohashi, T. and Aono, S. (1986) Jpn. J. Pharmacol. 42, 517-523. [46] Landcrson, R.C., Monsey, J.D., Allin, J. and Silbert, D.F. (1993) J. Biol. Chem. 268, 7650-7659.