Induction of Interleukin 18 Expression from Human Peripheral Blood ...

Vol. 267, No. 20, Issue of July 15, pp. 14183-14188,1992 Printed in U.S.A.

THEJOURNAL OF BIOLOGICAL CHEMISTRY 0 1992 by The American Society for Biochemistry and Molecular Biology, Inc.

Induction ofInterleukin 18 Expression from Human Peripheral Blood Monocyte-derived Macrophagesby 9-HydroxyoctadecadienoicAcid* (Received for publication, October 21, 1991)

George KuS, Craig E. Thomas, Ann L. Akeson, and RichardL. Jackson From the Marion Merrell Dow Research Institute, Cincinnati, Ohio 45215

Oxidatively modified low density lipoproteins(LDL) haps by macrophages or endothelial cells, plays a major role have recently been proposed to play a role in athero- in the pathogenesis of atherosclerosis (1).With regard to genesis by promoting foam cell formation and endothe-foam cell formation, these modified LDL are recognized and lial cell toxicity. The purposeof the present study wasinternalized by a scavenger receptor on macrophages, resultto determine whether modified LDL could also induce ing in lipid accumulation. Since this receptor is not downmacrophage release of interleukin 18 (IL-lB), a cyto- regulated by cytoplasmic cholesterol (2), internalization of kine which enhances vascular smooth muscle cell pro- lipoprotein cholesterol results in massive accumulation of liferation, another feature of the atherosclerotic proc- cholesteryl esters withinthe macrophages leading to foam cell ess. LDL were oxidativelymodified by incubation with formation (3). either Cu2+ (Cu2+-LDL) or humanperipheral blood The proliferation of smooth muscle cells in the atherosclemonocyte-derived macrophages (M-LDL). Incubation rotic lesion is promoted by a number of growth factors (4). of these modified LDL with macrophages(6 X 10‘ cells/ One of these growth factors is interleukin-I@(IL-16) (5-8),a culture) resulted in a dose-dependent induction of IL18 release. At 300 p g protein/ml, Cu2+-LDL and M- cytokine known to be released by activated macrophages. LDL induced422 and333 pg of IL-l@/culture,respec- Minimally modified LDL induce monocyte colony-stimulattively. Saponified Cu2+-LDLand M-LDL were shown ing factor and monocyte chemotactic protein-1 expression in endothelial cells, and these factors are released during the to contain 9- and13-hydroxyoctadecadienoicacid (HODE), lipid oxidation products of linoleate. When atherogenic process (9). Since minimally modified LDL have (3 X lo6cells/ been shown to contain low levels of thiobarbituric acid-reactested for activity in macrophage culture culture), it was found that 9-HODE and 13-HODE tive substances (TBARS), which are derived primarily from ) the release of 122 lipid hydroperoxides (lo), we hypothesized that lipid oxida(final concentration 33 p ~ induced and 43 pgof IL-l@/culture, respectively, whereas un- tion products may play a role in cytokine induction. In this treated cells released only 4pg of IL- lb/culture. Incu- regard, a recent report has shown that probucol, a lipophilic bation of macrophages with cholesteryl-9-HODE also antioxidant,inhibits ex vivo lipopolysaccharide (LPS)-ininduced IL-18 release; however, the degree of induc- duced IL-lP release from murine peritonealmacrophages (11). tion of IL- 18release by 9-HODE or its cholesteryl ester Oxidized lipids have also been shown to be involved in LPS relative to modified LDL suggests that other compo- stimulation of macrophages (12) and protein kinase C actinents in oxidized LDL may also contribute to IL-18 vation (13). These various observations indicate that products induction. 9-HODE was rapidly taken up by macro- of lipid peroxidation may be involved in thesignal transducphages, and the kinetics were similarIL-to18 release. leading to IL-16 release. Since oxidized derivatives of A 1.6- to 6-fold increase in the level of IL-18 mRNA tion linoleic acid and cholesteryl linoleate, i.e. hydroxyoctadecawas detected as little as 3-h post-9-HODE treatment. The induction of IL- 18 release from humanmonocyte- dienoic acids (HODEs) and cholesteryl-HODEs, are present derived macrophages by 9-HODE and cholesteryl-9- in human atheroma (14-16), it was of interest to determine HODE suggests a role for modified LDL, and its asso- whether these lipid oxidation products present in oxidatively modified LDL could stimulate macrophages to release IL-1P. ciated linoleate oxidation products, in vascular smooth The present studies demonstrate that 9-HODE or its corremuscle cell proliferation. sponding cholesteryl ester induces IL-16 release, implicating a pathological connection between oxidized LDL and smooth muscle cell proliferation in atherosclerotic lesions. Recent evidence suggests that the cell-mediated oxidative EXPERIMENTALPROCEDURES modification of plasma low density lipoproteins (LDL),’ per-

* The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. $ To whom correspondence should be addressed Marion Merrell Dow Research Inst., 2110 E. Galbraith Rd., Cincinnati, OH 45215. Tel.: 513-948-3141;Fax: 513-948-7472. The abbreviations usedare: LDL, low density lipoproteins; ALDL, acetylated LDL; Cu2+-LDL,copper-oxidized LDL; M-LDL, macrophage-oxidized LDL; ELISA, enzyme-linked immunosorbent assay; HODE, hydroxyoctadecadienoic acid; HPODE, hydroperoxyoctadecadienoic acid; IL-16, interleukin 18; LPS, lipopolysaccharide; TBARS, thiobarbituric acid-reactive substances; TNFa, tumor necrosis factor a;HPLC, high performance liquid chromatography.

Isolation of Mononuclear Cells-Human peripheral blood was collected in 10 mM sodium citrate from healthy volunteers. Erythrocytes and neutrophils were removed by low speed centrifugation in Leucoprep tubes (Becton Dickinson, Oxnard, CA) according to the manufacturer’s suggested protocol. The resulting mixture of platelets and mononuclear cells was incubated in tissue culture dishes (3 X lo6 cells/well of 24-well plate, Corning, Corning, NY) for 1 h at 37 “C, after which nonadherent plateletsand other cells were removed.Fresh medium RPMI 1640 (GIBCO) was then added to the adherent cells. Flow cytometric analyses showed that 99.5%of the adherent cells were capable of internalizing a fluorescent acetylated LDL (ALDL) probe, suggesting a macrophage-like phenotype bearing the scavenger receptor (data not shown). All experiments were performed with freshly adhered cells in the absence of serum. Preparation of Modified LDL-Human plasma was obtained from

14183

14184

9-HODE-induced IL-1B Release

a localblood center. To each pool of plasma (18-20 donors) was added 1 mM EDTA, 0.5 mM phenylmethylsulfonyl fluoride, 50 units/ ml aprotinin, and 0.01% sodium azide prior to LDL fractionation. LDL were isolated by preparative ultracentrifugation in KBr (d = 1.019-1.063 g/ml) as described previously (17). Five different batches ofLDLwereused inthis study. Prior to oxidation, EDTA was removed from the LDL by dialysis against phosphate-buffered saline (PBS), pH7.4, at 4 "C in the dark. Copper-oxidizedLDL (CuZ+-LDL) were prepared by incubating LDL (2 mg LDL protein/ml) with 10 p M CuSO, for 16 h at 37 "C.Typically, Cu2+-LDLcontained 40 nmol of TBARS/mg of protein. Excess CuS04 was removed by dialysis. Macrophage-modified LDL (M-LDL) were prepared by incubating LDL (2 mg LDL protein/ml) with human peripheral blood monocytes (adherent cells) for 24 h at 37 "C; M-LDL contained 10-12 nmol of TBARS/mg of protein. M-LDL were re-isolated from the culture medium by preparative ultracentrifugation in KBr (d = 1.006-1.15 g/ml). ALDL were prepared according to the method of FraenkelConrat (18).To control for possible endotoxin contamination during dialysis and re-isolation, modified LDL were routinely passed over a prepacked endotoxin affinity column (Detoxigel, Pierce Chemical Co.). As shown by the Limulus polyphemus amebocyte lysate test (Sigma; sensitivity 1 ng of endotoxin), this Detoxigel column has the capacity of removing 2 mg of endotoxin. Furthermore, the affinity of endotoxin for the column is unaffected by the presence of LDL (data not shown). Since endotoxin is equally potent in inducing IL-10 and tumor necrosis factor a (TNFa), andsince oxidized LDL only stimulate IL-lP but not TNFa release (see below), a TNFa enzyme-linked immunosorbent assay (ELISA, sensitivity 20 pg) was used routinely for oxidized LDL-stimulated culture supernatants to monitor for potential endotoxin contamination. In every experiment in which Cu'+-LDL, M-LDL, and ALDLwere compared with native LDL, LDL were from the same lipoprotein preparation. Human IL-10and TNFa ELISA kits were from Cistron (Pinebrook, NJ), andboth were used according to the manufacturer's suggested protocol. Lactate dehydrogenase levels were routinely measured by a clinical analyzer (Coulter, Hialeah, FL). Determination of Lipid Peronidation Products-Lipid peroxidation products were determined by HPLC analysis of saponified lipids. For LDL and modified LDL, the lipids were extracted by dropwise addition of the sample (0.4 mg of protein) into 2 ml of ether:ethanol (3:l) with mixing. The protein was pelleted by centrifugation and the supernatant fraction decanted and dried under nitrogen. The dried lipids were reconstituted in 0.5 mlof ethanol to which was added 0.05 ml of 10 N NaOH, followed by heating a t 60 "Cfor 20 min to generate the free acids. The solution was neutralized by the addition of 0.03 ml of glacial acetic acid and dried under nitrogen. To the lipid was added 1 ml of water and then 1 ml of heptane. Following mixing and centrifugation at 1,800 rpm for 5 min, 0.8 mlof the heptane layer was removed, dried under nitrogen, and re-solubilized into 0.15 ml of heptane. Fatty acid oxidation products were separatedusing anormalphase silica column (Zorbax Sil, 25 cm X 4.6 mm, 5 pm) according to Teng andSmith (19) with minor modification. The mobile phase was a quaternary mixture of heptane:diethyl ether:isopropanol:acetic acid (100:10:0.9:0.1) run at a flow rate of 2 ml/min; conjugated dienes were monitored at 234 nm. Standards of linoleate hydroperoxides, i.e. 9(S)and 13(S)-HPODE, and linoleate hydroxides, i.e. 9(S)- and 13(S)HODE (cis and trans isomers, Cayman Chemicals, Ann Arbor, MI) in ethanol were dried under nitrogen, re-solubilized in heptane, and subjected to chromatography as described for the saponified lipid fractions. For determination of conjugated dienes in cells, the lipids were extracted as described by Schade et al. (20) with slight modifications. Cells (25 X IO6)were lysed with 0.75 ml of 0.25% deoxycholate and then thelysate was transferred to 13X 100-mm polypropylene tubes. The plates were washedwith 0.75 ml of water and thewash combined with the deoxycholate lysate. This mixture was then extracted with 1.8 mlof ch1oroform:methanol:aceticacid (2OO:lOOl) followed by centrifugation at 1,500 rpm for 5 min. The chloroform layer (0.9 ml) was removed and dried under nitrogen. The lipids were saponified as described above and subjected to chromatography. Quantitation was achieved by comparison to a standard curve (0-1 pg) generated from the commercially available standards of 9(S)- and 13(S)-HPODE and 9(S)- and 13(S)-HODE. Preparation of Cholesteryl-9-HODE-Cholesteryl-9-HODEwas prepared by a modification of the method described by Peers and Coxon (21) for 2-linoleoyl-1,3-dipalmitoylglycerol. Briefly, 50mgof cholesteryl linoleate (Nu-chek Prep, Elysian, MN) were dissolved in heptane and mixed with 5 mg of a-tocopherol (Sigma) in a 16 X 125-

mm glass screw cap test tube. The mixture was dried to an oil under nitrogen and the tube flushed with oxygen and sealed. The oil was heated at 40 "C for 5 days, and the material was then dissolved in 5 ml of ethanol. Twenty-five mg of sodium borohydride were added and the mixture shaken gently a t room temperature for 10 min. The ethanol was evaporated, and 1 ml of water was added. The lipid was extracted into heptane and subjected to semi-preparative HPLC using a Spherisorb 5 Si1 column (22.5 X 250mm, Phenomenex, Rancho Palos Verdes, CA). The mobile phase consisted of heptane:ethyl ether:isopropanol:acetic acid (1002:l:O.l); the flow rate was 8 ml/ min. Two major peaks exhibiting significant diene conjugation, as monitored by absorbance a t 234 nm, were collected. Following removal of the mobile phase under vacuum, the residues weredissolved in heptane. Fast atom bombardment-mass spectrometry indicated both products to have a molecular weight identical to the hydroxylated derivative of cholesteryl linoleate. The identity of regio-specificityof the products was confirmed by saponification and analytical HPLC analysis of the free fatty acids as described above. One product contained a fatty acid which co-migrated with authentic 13-HODE and the other with 9-HODE. The amount of each product was determined by comparison of the areas of the free fatty acids to a standard curve prepared from standards of 9- and 13-HODE. Heptane wasremoved from each isolated fraction under nitrogen and the cholesteryl ester products reconstituted in ethanol at a final concentration of 1 mg/ml (3.3 mM with respect to HODE content). Determination of ZL-10 mRNA-TotalRNAwas isolated from human peripheral blood monocytes using guanidine thiocyanate and phenol/chloroform extractions by a modification of the method of Chomczynski and Sacchi (22). Briefly, cells were washed twice with cold PBS andreleased from the plate with a cell scraper into 1 ml of PBS containing 40 units of RNasin (Stratagene, La Jolla, CA). The cells were pelleted and resuspended in 1 ml of GTC extraction buffer (4 M guanidine thiocyanate, 25 mM sodium citrate, pH 7.0, 40 mM 2mercaptoethanol, and 0.5% sodium lauryl sarcosine). The cellular lysate was sheared by passing three times through an18-gauge needle and four to five times through a25-gauge needle.The lysate was kept on ice and extracted by addition of 0.2 ml of 1 M sodium acetate, 1.0 ml of water-saturated phenol, and 0.2 ml of ch1oroform:isopropanol (49:l) with mixing on a vortex after each addition and for 30 s at the end. The tubes were immediately centrifuged for 20 min at 4 "C and 10,000 rpm. The aqueous phase was transferred to a new tube and the RNA precipitated, following addition of an equal volume of cold isopropanol at -80 "C for 10 min. The pellet was redissolved in 200 pl of Tris-EDTA buffer (10 mM Tris, 1 mM EDTA, pH 7.5) with 4 units RNasin. The RNA was further purified by extracting one time with an equal volume of ch1oroform:isoamyl alcoho1:phenol (49:1:50) and one time with an equal volume of ch1oroform:isoamyl alcohol (24:l). The RNA wasprecipitated at -80 "C for 10 min or -20 "C for 1 h or longer following addition of 0.1 volume of 3 M sodium acetate and 2 volumes of cold ethanol. After centrifugation, the RNA was resuspended in 25 p1 of the Tris-EDTA buffer and was quantitated spectrophotometrically. The relative cellular levels of IL-10 were determined by Northern blot analysis of 3 or 10 pg of total RNA. The RNA was electrophoresed in a 1%agarose/formaldehyde gel (23) and blot-transferred to Nytran (Schleicher and Schuell). An EcoRIIXbaI 480-base pair portion of the mature IL-10 cDNA (Beckman, Fullerton, CA) and a PstIIXbaI 516-base pair portion of human skeletal muscle actin cDNA (24) were used as hybridization probes and were radiolabeled by random priming with [32P]dCTP(25). Following autoradiography, band intensities were determined with a densitometer (Molecular Dynamics, Sunnyvale, CA). RESULTS

Incubation of humanperipheral blood monocytes with Cu2+-LDL or M-LDL for 24 h induced a dose-dependent increase of IL-1@release into the medium (Fig. 1). At the highest concentration tested (300 pg protein/ml), Cu2'-LDL and M-LDL induced an average of 422 f 8 and 333 k 7 pg of IL-lp/culture (6 X lo6 cells/culture), respectively. Control LDL which were subjected to incubation at 37 "C for 24 h but in the absence of Cu2+or cells did not induce IL-lP release, as compared with untreated cells (65 pg of IL-lplculture at 300 pg of LDL protein). ALDL differs from Cu2'-LDL and M-LDL in that itis modified by acetylation of lysine residues

9-HODE-induced IL-10 Release (26), whereas Cu2+-LDLand M-LDL are modified by oxidative processes. However, ALDL (300 pg protein/ml), which contained less than 0.5 nmol of TBARS/mg of protein and undetectable levels of conjugated dienes, did not induce IL16 release. TNFa was not detected (sensitivityof assay 20 pg) in any culture supernatants (data not shown). In a total of five separate experiments, Cu2+-LDL induced a 3-15-fold increase in IL-lP release when compared with medium control, whereas M-LDL induced a 4-8-fold increase (Table I). HPLC analysis of the saponified lipid fraction from Cu2+LDL and M-LDL showed the presence of 9- and 13-HODE and 9- and 13-HPODE,products of linoleate oxidation (Fig. 2). In Fig. 2A, peaks 1-4 represent standards of 13-HPODE, g-HPODE, 13-HODE, and 9-HODE, respectively. Cu2'-LDL (Fig. 2B) contained eight major peaks representing isomers of the four products, whereas M-LDL contained primarily 9HODE and 13-HODE (3.06 and 5.90 pg/mg LDL protein, respectively) (Fig. 2C). Prior to oxidation, freshly isolated LDL contained less than 10% of these lipid dienes relative to M-LDL (Fig. 2 0 ) . To assess the contribution of these oxidized lipids to the release of IL-lP by modified LDL, macrophages (3 x lo6cells/ culture) were incubated with authentic 9- and 13-HODE. As shown in Fig. 3, at 33 PM the HODEs significantly increased IL-lP release relative to medium control. Interestingly,9HODE induced a much greater release of IL-lp than 13HODE (122 uersus 43 pg of IL-l@/culture,respectively), suggesting a specificity in the signal transduction pathway leading to IL-lP induction. The absolute amounts of IL-1p in500

OJ

0

50

100

150 250 200

300

Protein ilg;ml

FIG. 1. Stimulation of IL-lfl release by Cu2+-LDL and MLDL. 6 X lo6 humanperipheral blood monocyte-derivedmacrophages were incubated in 24-multiwell plates with medium RPMI 1640 alone (O),LDL (0),ALDL (A), Cu*+-LDL (O), or M-LDL (W) for 24 h. Secreted IL-10 was measured with an IL-10 ELISA. Data are mean k S.E. of quadruplicate cultures.

14185

duced by 9-HODE varied among donors. However, IL-1P levels were consistently higher than that of control cultures, but were 2-8-fold lower than that of LPS-stimulated cells (Table 11). Although free 9-HODE can be produced and released by endothelial cells from linoleic acid via a cyclooxygenasedependent mechanism, in the context of atherosclerosis (27), the majority of linoleate-derived 9-HODE present inoxidized LDL occurs as esters of cholesterol or phospholipid. Therefore, cholesteryl-9-HODE was synthesized and its effects on IL-lP induction determined. Both cholesteryl-9-HODE and 9-HODE induced a similar dose-dependent increase in IL-lP release (Fig. 4). The plateau observed with cholesteryl-9HODE may reflect its relative insolubility in the culture medium. To determine the kinetics of IL-1P induction, macrophages were incubated with 9-HODE for 30, 60, 120, and 300 min, after which culture supernatants were collected for IL-1P determination. Analysis of the culture supernatants showed that extracellular IL-lP was detectable as early as 5 h after the addition of 9-HODE (Fig. 5 A ) . To determine whether the kinetics of induction correlated with the uptake of 9-HODE, the hydroxy fatty acid was quantitated both in the cell and the medium. HPLC analysis of cells at these various times showed a time-dependent association of 9-HODE with the cells (Fig. 5B). The sum of cell-associated 9-HODE and 9HODE in themedium was similar to the amountof 9-HODE recovered from cell-free medium at each time point, suggesting that 9-HODE was not significantly metabolized by the cells during this time period. We next determined whether 9-HODEaffected the expression of IL-1P by macrophages. By Northern blot analysis of total RNA, the levels of IL-10 mRNA in unactivated peripheral monocytes were undectable. However, adherence to plastic induced detectable levels of IL-lP mRNA in mediumtreated control cells (Fig. 6), consistent with earlier reports (28, 29). In the experiment shown in Fig.6, after a 3-h exposure to 33 PM 9-HODE, cells from a single donor had a 2-fold increase of IL-lp mRNA level compared with the untreated cells from the same donor. For nine different macrophage preparations tested (Fig. 7), IL-lPmRNAs increased from 1.5- to 6-fold with a mean of 3.67 ? 1.6 in response to 9-HODE. In Fig.7, the relative levels of IL-1p mRNA in vehicle- and 9-HODE-treated cells were determined by normalizing tothe level of actin mRNA in each sample as detected on the same blot. DISCUSSION

In the present study, we demonstrate that oxidized LDL, but neither ALDL nor native LDL, were capable of inducing

TABLEI Induction of IL-16, bv " oxidized LDL Experiment

LPS"

ALDL

Donor

IL-lP

M-LDL Vehicle CuZ'-LDL

LDL pg/culture*

1

2 3 4 5 6 7 8

R. S. T. D. T . M. J. M. M. T. M. K. B. G. A. M.

33 f 5 I f 2 84 & 2 100 f 26 33 f 3 74 k 9 95 f 4 22 k 3

42 +. 4' 4 f Id 141 f 25' 37 f 6B 82 f Sh 83 f 28' 20 f 6'

The LPS concentrationwas 1 pg/ml/culture. Data are meanf S.E. of triplicate cultures. ' The lipoprotein concentrations were 250,' 300: 100,' 200,' 250:

228 f 14 228 k 30' 504 k 27' 486 f 2@ 355 k 2Oh

54 f 22'

55 k 2d 369 f 39'

72 f 23' 124 k 10' 32 f 11"

904 f 9

384 f 4' 166 k 7'

100," 200,' and 400' pg of protein/ml/culture, respectively.

14186

9-HODE-induced IL-lp Release 1

A

1 2

120 100

2

80

0

a"

= 60 5

40 20 0

FIG. 3. HODE-induced stimulation of IL-10release by human peripheral blood monocyte-derived macrophages. Cul-

F

tures of macrophages (3 X lo6) in 24-multiwell plates were incubated with 33 PM of 9-HODE (uertical bar) or 13-HODE (horizontal bar). Eight lotsof 9- and 13-HODEhave been tested. 10 pl of 95% ethanol (vehicle) to I-ml culture (final 1%ethanol) did not affect the pH of the medium, cell viability, nor the cellular IL-la response to LPS. Concentrations of 9- and 13-HODEbelow 50 pM were not cytotoxic, as indicated by the release of lactate dehydrogenase. Data are mean f S. E. of quadruplicate culturesof a representative experiment.

r

r

c

t2 c

[I

[I

5 c

TABLE I1 Induction of IL-1B bv 9-HODE Experiment

IL-10

Donor Vehicle

9-HODE"

LPSb

pgfculture'

1

I

0

10

I

I

20

30

TIME (min)

FIG. 2. HPLC analysis of lipid peroxidation products. A, 13(S)-HPODE (l),9(S)-HPODE (2), 13(S)-HODE (3), and 9(S)HODE (4) standards were separated usinga normalphase silica column. B , saponified lipid profile from Cu2+-LDL. All four, 9- and 13-HODE and 9- and 13-HPODE,were detected. C, saponified lipid profile from M-LDL.The two major lipid peroxidation products found in M-LDLwere 9- and13-HODE. D, saponified lipid profile of native LDL.

IL-1p release from human peripheralblood monocyte-derived macrophages. The lack of effect with ALDL suggests that binding to the scavenger receptor alone is insufficient to induce IL-la release and that other components of oxidized LDL are involved in the stimulation of IL-lB release. Since oxidized LDL did not induce TNFa release, and since endotoxin induces both IL-1p and TNFa release, it is unlikely that IL-1/3 release induced by oxidized LDL was due to endotoxin contamination. In an effort to identify component(s) of oxidized LDL to which this biological activity could be attributed, the lipid fraction was separated and characterized by HPLC. Analysis of the saponified lipid fraction from Cu2'-LDL showed the presence of both Z,E- and E,E-isomers of 9- and 13-hydroxy and 9- and 13-hydroperoxy derivatives of linoleate, as expected from free radical-mediated linoleate oxidation (19). A recent report hascharacterized the lipid fraction of Cu2+-LDL by HPLC and gas chromatography-mass spectrometry analyses (30). In agreement with our results, it was found that

1 2 3 4 5 6 7 8 9

R.

R.S. T. M. D.M. C.B. E. M. E. H. M.R. D.D. D.

2 6 6 28 & 6 12&2 17+3 20+ 3 21 f 10 7+ 1 lo+ 1 20 + 4

1 1 73 +7 34 6f 3 372 + 53 2 4 7 t 17 212+ 13 159 + 15 208+ 65 115+ 11 155+ 11 120+ 7

611f54 626f44 4 1 6 f 38 881 f 16 4 3 4 f 15

33 pM.

* 200 ng/mI culture. e

Data are mean+ S.E. of triplicate cultures.

350/

3ooJ

0

11

22

33

44

Concentrailon UM

FIG. 4. Dose-dependent induction o f IL-lasecretion by 9HODE and cholesteryl-9-HODE. Cultures of macrophages in 24multiwell plates were incubated with 11,22,33, or 44 PM of 9-HODE (0)or cholesteryl-9-HODE (0).Data are mean -t S.E. of triplicate cultures of a representative experiment.

67% of the oxidized linoleate products were present as 9-and 13-derivatives. Additionally, conjugated dienes derived from arachidonate are also generated by oxidation, although at approximately 10% the level of linoleate oxidation products.

9-HODE-induced IL-IP Release

14187

8oTA

1.8 E 1.4 ._ c 3

3 1.0

/

a

0.6 0.2 I L - l a rnRNA I i:\

Aclm rnRNA

FIG.6. 9-HODE induction of IL-Ij3 mRNA expression. Human peripheral hlood monocyte-derived macrophages were cultured in 60-mm culturedishes. Medium RPMI 1640 ( o p n hnrs) or9-HODE (3.7 p M ) (closed bars) was added to cultures for 3 h, after which total RNAs wereisolated and were electrophoresed in a 1;' agaroseformaldehyde gel and blotted onto Nytran. I I . - l P mRNA and actin mRNA levels were determined hy Northern hlot analysis with labeled cDNA probes. The relative hand intensities were quantitated hy densitometric scanning of the autoradiograph. Data are from a representative experiment with cells from a single donor. I 2 P -

i,

1.

,

0

o-

!\r:::" 0

30

81

"1)-

~. fiO

.~

~~

~

120

.~ ~~

~

~ . " ~

.

"~ ~~

300 Mlnules

FIG. 5. Kinetics of 9-HODE-induced IL-IS release. Human peripheral hlood monocyte-derived macrophages (20 X 10") were cultured in 60-mm tissue culture dishes. 9-HODE (10 pl of a stock solution of 3.3 mM in 9.5% ethanol, final concentration was 33 p M ) was added to cultures a t 0 h. A , time-dependent secretion of IL-10 incubated with 9-HODE (0)or medium RPMI 1640 (0).R, timedependent changesof 9-HODE levels in cells (0). culture supernatant ( A ) , or cell-free medium ('>). The sumof 9-HODE in cells and culture supernatants is represented hy open squares. Data are mean 2 S.E. of triplicate cultures of a representative experiment.

W

sz :E E r

n .c c; o

m

-

b E U

t j E

@.=

--m g a

> .-

r

I

U

1

2

3

4

5

6

7

8

9

Donor

FIG. 7. Increase in IL-IDmRNA levels by 9-HODE in cells I t mustalso beconsidered that cyclic endoperoxidesand from multiple donors. Monocyte-derived macrophages from nine aldehydic compounds which are not detectable as conjugated donors were treated with 33 p M 9-HODE for 3 h and total R S A dienes are generated during LDL oxidation. The potential isolated as descrihed in the legend of Fig. 6 and under "Experimental contribution of these additional products to IL-1P release by Procedures." The Northern hlots were first probed for II,-ld mRSA oxidized LDL is suggested by the demonstration that0.1-0.3 levels and then for actin mRNA levels. Following autoradiography. mg of M-LDL, containing0.3-0.9 pg of 9-HODE, induced IL- handintensities were determinedand levels of IL-IdmHSA in lp release comparable with 9.5 pg (33 p ~ of) free 9-HODE. control (open bars) and 9-HODE-treated samples (closed hats) normalized to the levels of actin mRNA in the same samples. For each This may be explained, in part, by the limited uptake of 9- donor, the control samplewas assigned 8 relative I L l d mRNA/actin HODE by macrophages which achieved a level of 0.1-0.2 pg/ MRNA level of 1. For the samples from the nine donors tested. the 10" cells after a 5-h incubation (Fig. 5B).Uptake of M-LDL mean level of IL-10 mRNA induction hv 9-HODE WRS 3.67 k 1.6 or Cu2+-LDLvia the scavenger receptor maybe more efficient (mean k S.D.). at delivering 9-HODE to the cells, or alternatively,other components present in modified LDL may also contribute to cells incubated with 9- or 13-HPODE. As few conjugated dienes were observable in modified LDL I L - l a release. withoutsaponification, it is likely thatoxidation involves Incontrastto Cu"-LDL, saponifiedM-LDLcontained mainly 9- and 13-HODE in agreement with preliminary recholesteryl esters or phospholipids. Structural identification sults reported by Rankin et al. (31). This difference is con- of oxygenated fattyacid derivatives in human atherosclerotic sistent with the ability of cells to reduce 9- and 13-HPODE plaques revealed the presence of oxidized cholesterol esters totheir respective HODE. Wewere unable to accurately (16). It cannotbe determined from these studies whether LDL lipid oxidation by macrophages is an oxygen radical-mediated determine the abilityof HPODEs to induce IL-lP, as they are unstable in medium RPMI 1640 and are degraded to uniden- process or is lipoxygenase catalyzed (32), since both mechatified products (data not shown). They exhibited very limited nisms result in the formation of oxidized linoleate products. ability to induce IL-ls (data not shown) which reflected the Support for free radical-mediated lipid oxidation is provided net result of continual "deterioration" of HPODEs and cell- by demonstration that the oxidized lipids exist as a racemic mixture of Z,E- and E,E-isomers of 9- and 13-HODE(34). I t mediated conversion of HPODEs to HODEs. In agreement, no HPODE and only a small level of HODE was found in cannot be ruled out, however, that the initial lipid oxidation

14188

9-HODE-induced IL-lP Release

productsarestereoandlor regio-specific hydroperoxides University) concerning the methodology for the preparation of chowhich are converted to racemic mixtures via subsequent tran- lesteryl-9-HODE. sition metal-catalyzed reactions. REFERENCES The 9-HODE presentin modified LDL would exist predom1. Steinberg, D., Parthasarath , S , Carew, T. E., Khoo, J. C., and Witztum, J. L. (1989) N. E 8 l . J. d d . 320,915-924 inantly ascholesteryl-9-HODE which we have shown can also 2. Brown, M. S., and oldstem, J. L. (1983) Annu. Reu. Eiochem. 6 2 , 2233fil induce IL-1p release (Fig. 4). Since cholesteryl-9-HODE could 3. Go%tein. J. L.. Ho. Y. K.. Basu. S. K.. and Brown. M. S. (1979) Proc. conceivably be de-esterified by macrophage lysosomal esterNatl. Acad. Sci. U.’S. A . 76, 333-337 ’ 4. Ross, R. (1986) N. Engl. J. Med. 314,488-500 ases to 9-HODE, 9-HODE could be an intermediate metabo5. Libby, P., Warner, S. J. C., and Friedman, G. B. (1988) J. Clin. Inuest. 8 1 , lite in cholesteryl-9-HODE stimulation of IL-lp release. Ex6. periments are in progress to determine whether de-esterifi7. cation of cholesteryl-9-HODE by the macrophage is required 8. for induction of IL-16 release. It also remains to be ascertained 9. whether 9-HODE serves as a precursor or substrate for the ultimate activating agent. Kinetic experiments showed that within 5 h of stimulation, the amount of 9-HODE recovered 10. from the medium and cells was similar to that added at zero 11. time, indicating that 9-HODE was not significantly metabo- 12. lized and that 9-HODE could be the intracellular signal. 13. However, radiolabeled 9-HODE is needed to determine whether the cell-associated 9-HODE is derived from the me- 14. dium or is a product of oxidation of endogenous cellular 15. linoleic acid. Precedence for adirect role of 9-HODEin 16. activationhas been provided by the finding that HODE activates rat brainproteinkinase C (11) by fulfilling the 17. requirement for phospholipid. A similar role in induction of 18. 19. IL-lP expression may be envisioned. 20. Northern blot analyses of aortic lesions obtained from 21. cholesterol-fed non-human primates show a significant in- 22. crease of IL-lp mRNA levels compared with aortas from 23. Focus 1 0 , l - 5 , P., Ponte, P., Okayama, H., Engel, J., Blau, H., and Kedes, L. control animals (33). In addition, a640-fold increase of IL-lp 24. Gunnin (19837 Mol. Cell. Biol. 3 , 787-791 mRNA levels was found in human atherosclerotic lesions (34). 25. Feinberg, A. P., and Vogelstein, B. (1983) Anal. Biochem. 132,6-11 M. S., Basu, S. K., Falck, J. R., Ho, Y. K., and Goldstein, J. L. These observations support a potentialatherogenic role of IL- 26. Brown, (1980) J . Supramol. S t r u t . 13,67-81 10 and is consistent with reports that IL-10 stimulates vas- 27. Kaduce, T. L., Figard, P. H., Leihr,R., and Spector, A. A. (1989) J. Biol. Chem. 264,6823-6830 cular smooth muscle cell proliferation (5, 6). Oxidized LDL 28. Haskill, S., Johnson, C., Eierman, D., Becker, S., and Warren, K. (1988) J. Immunol. 140,1690-1694 have been shown previously to promote cholesteryl ester 29. Unanue E. R., and Allen P. M. (1987) Science 236,551-557 accumulation in macrophages (4) and to be cytotoxic toward 30. Lenz, M . L., Hu hes, H.,’Mitchell,,J. R., Via, D. P., Guyton, J. R., Taylor, A. A,, Gotto, M., Jr., and Smlth, C. V. (1990) J. L ~ p dRes. 3 1 , 1043endothelial cells, vascular smooth muscle cells, and fibroblasts 1n m (35,36). The present demonstrationthat oxidized LDL induce 31. RaG& S. M., Parthasarathy, S., and Steinberg. D.(1990) Arteriosclerosis 10,828a the expression and release of IL-1p further supports a critical 32. McNally, A. K., Chisolm, G. M., 111, Morel, D. W., and Cathcart, M. K. (1990) J. Immunol. 145,254-259 role of oxidized LDL in atherogenesis. Accordingly, inhibition 33. Ross R Masuda J. Raines, E. W., Gown, A. M., Katsuda, S., Sasahara, of LDL oxidation may be a plausible therapeutic approach M.’, Malden, L.’T.,’Masuko, H.,and Sato, H. (1990) Science 2 4 8 , 10091012 for prevention of atherosclerosis. 34. Wang, A. M., Doyle, M. V., and Mark, D. F. (1989) Proc. Natl. Acad. Sci.

1.

Acknouledgrnents-We thank David Ohlweiler for assistance in the HPLC analyses and Rosanne Dennin for typing of the manuscript. We appreciate the suggestion of Dr. Alan R. Brash (Vanderbilt

U. S. A . 86,9717-9721 35. Evenorn, S. A., Galdal, K. S., and Nilsen, E. (1983) Atherosclerosis 49,2330 36. Morel, D. W., Hessler, J. R., and Chisolm, G. M. (1983) J. Lipid Res. 2 4 , 1070-1076