Lipolyzed hypertriglyceridemic serum and triglyceride-rich lipoprotein cause lipid accumulation in and are cytotoxic to cultured human endothelial cells. High density lipoproteins inhibit this cytotoxicity

THROMBOSIS RESEARCH 58; 251-284,199O 0049-3848190 $8.00 + .OO Printed in the USA. Copyright (c) 1990 Pergamon Press pk. All rights reserved.

LIPOLXZED HYPERTRIGLYCERIDEMIC SERUM AND TRIGLYCEiRIDE-RICH LIPOPROTEIN CAUSE LIPJD A CCUMUIATIONINANDAREcyToToxICTo CULTURFDHUMAN ENDOTHELIAL CELLS. HIGH DEiNSlTY LIPOPROTEINS INHIBIT THIscJYlvmxIClTY. M.T. Speidel”, F.M. Booysel, A Abrams, M.A Moore2 and B.H. Chung’2 Division of Cardiovascular Disease’ and Atherosclerosis Research Unit2, Department of Medicine, University of Alabama at Birmingham, Birmingham, Alabama 35294, U.S.A (Received 21.4.1989; accepted in revised form 10.2.1990 by Editor N.U. Bang) ABsrRAcr

The cytotoxic effect of hypertriglyceridemic (HTG) serum and triglyceride-rich lipoprotein (TG-rich lipoprotein) lipolyzed in vitro by purified lipoprotein lipase on cultured human umbilical vein endothelial cells (HUVECs) was studied. When confluent cultures of HUVECs (8.4 x 104/cm2 ) were incubated in the presence of control (non-lipolyzed HTG serum) or lipolyzed HTG serum or TG-rich lipoprotein, the lipolyzed HTG serum or TG-rich lipoprotein was cytotoxic to the HUVECs as indicated by their detachment from the culture dish; the lipolyzed serum at 10% of the culture medium or lipolyzed TG-rich lipoprotein at 75 pg cholesterol/ ml caused the detachment of all (100%) of the cells from the culture dish after a 24 h incubation. Control (non-lipolyzed) HTG serum or non-lipolyzed TG-rich lipoprotein at the same or higher concentration was not cytotoxic to the cells. The HUVECs incubated for 48 h with low (sublethal) doses of lipolyzed TG-rich lipoprotein (lo-50 pg cholesterol/ml) contained massive lipid inclusions; no lipid inclusions were seen within the cells when the culture medium contained control non-lipolyzed TG-rich li~proteins. Finally, when high density lipoprotein (HDL) was added to the culture medium at the same concentration as the cytotoxic lipolyzed TG-rich lipoprotein (75 c(g cholesterol/ ml), the cytotoxic effect of the lipolyzed TG-rich lipoprotein was inhibited. These data suggest that the interaction of endothelial cells with lipolytic remnants of TG-rich lipoprotein may play a role in the pathogenesis of atherosclerosis and that HDL may play an important role in inhibition of the endothelial cell injury produced by the lipolytic remnants of TG-rich lipoprotein.

KEY WORDS: Triglyceride-rich lipoprotein, human endothelial HDL, lipid peroxides, lactate dehydrogenase .

cells, cytotoxicity, lipid inclusions,

‘To whom correspondence should be addressed: Michael T. Speidel M.D., Department Therapy, Harvard Medical School, 50 Binney Street, Boston, MA 02115, U.S.A

251

of Radiation

252

ENDOTHELIUM AND HYPERTRIGLYCERIDEMIA

Vol. 58, No. 3

INTRODUCTION Elevated levels of circulating plasma lipoproteins have been implicated as risk factors for the development of atherosclerosis (1). Data from several clinical studies indicate that an elevation of plasma triglycerides is common in patients with coronary heart disease (2,3), but multivariate analysis of epidemiological data has failed to establish triglyceride as an independent risk factor (4). However, in patients with clinical hypertriglyceridemia, especially those with type III hyperlipidemia, the association of triglycerides and risk of coronary heart disease is very high and suggests a cause-effect relationship (5,6). The triglycerides in plasma are associated mostly with chylomicrons and very-low-density lipoprotein (VLDL) particles, and the chylomicrons and VLDL in plasma carry the triglycerides from exogenous and endogenous sources, respectively. The catabolism of these TG-rich lipoproteins occurs at the surface of the endothelium by the action During the lipolytic process the TG-rich lipoprotein loose their of lipoprotein lipase (7). triglycerides and become remnant particles poor in triglycerides and rich in cholesterol. One consequence of the catabolism of TG-rich lipoprotein in the plasma compartment is a flux of excess surface components of TG-rich lipoprotein into the HDL in plasma (8). It has been suggested that lipolytic remnants of TG-rich lipoprotein in animals on an atherogenic diet are responsible for the development of atherosclerosis in these animals (9). Although the production of lipolytic remnants of TG-rich lipoprotein may occur at the surface of the endothelium d vivi(lO), the effect of lipolytic remnants of TG-rich lipoprotein on endothelial cell function is not clear. Studies on the interaction of VLDL with cultured endothelial cells have indicated that VLDL in serum, in certain pathological or physiological states, was cytotoxic to the cells (11,12,) and human smooth muscle cells (13). A recent report has indicated that HTG serum, lipolyzed in vitro by purified lipoprotein lipase or, in vivo by heparin injection, is cytotoxic to cultured mouse peritoneal macrophages, and that this cytotoxicity was associated with the post-lipolysis remnants products of TG-rich lipoprotein (14). In the present study, we have examined the cytotoxic effect of lipolyzed HTG serum or TG-rich lipoprotein on cultured HUVECs.

MATERIALS

AND MEl’HODS

Human serum and bsma: Fresh samples of fasting and postprandial lipemic serum or plasma were obtained (after acquiring informed consent) from HTG volunteers through the Alabama Regional Blood Center, Birmingham, Al. The postprandial serum and plasma were obtained 3-4 h after a fatty meal. Levels of TG and cholesterol in plasma or serum were determined by enzymatic assay methods using the Boehringer Mannheim triglyceride assay kit (kit No. 348202, BiodynamicsBMC Co., Indianapolis, IN) and cholesterol assay kit (kit No. 12408). Lipoprotein cholesterol profiles of these serum and plasma were examined by the Vertical Autoprofiler (VAP) method as previously described (15). Isolation of lbonroteins and liuoprotein lipase: The VLDL density ( d< 1.006 g/ml) fraction containing VLDL and/or chylomicrons were separated from fasting or postprandial sera from HTG subjects by ultracentrifugation. LDL (d=1.006-1.063) and HDL fractions (d=1.063-1.21 g/ml) were separated from the above sera by the sequential floatation method (16) and were further purified by ultracentrifugation in a vertical rotor (17). The purity of the isolated lipoprotein was assessed by gradient SDS gradient gel electrophoresis (18). Lipoprotein lipase was isolated from bovine milk and purified by affinity column chromatography using heparin-sepharose (Sigma Co. St Louis, MO) as described previously (19). The purified lipoprotein lipase had a specific activity of 10 - 20 mU/pl when an emulsion of [3H]- triolein was used as a substrate. One milliunit of enzyme represents 1 nmole of free fatty acid released per min. at 37’ C. Cell culture: HUVECs

were obtained

from human umbilical cords by mild collagenase

treatment

Vol. 58, No. 3

ENDOTHELIUM AND HYPERTRIGLYCERIDEMIA

253

(20) and cultured to confluency on fibronectin-coated plastic tissue culture flasks as described previously (21). Selected primary cultures were subcultured, pooled and grown to confluency in fibroneetin-coated multiwell (2 cm2) plates. All experiments described here were carried out on confluent first passage cultured HUVECs, representing a pool from 3-4 different cords to minimize variations in cells derived from single cords. Cells were routinely maintained at 37’C in a humidified 95% air-5% CO, atmosphere in a complete culture medium consisting of Medium199 (powder medium containing L-glutamine and Earle’s salts), 25 mM HEPES adjusted to pH 7.4, 2 mM fresh L-glutamine, 120 U/ml penicillin, 100 /@ml streptomycin, 10% heat-inactivated (56’C for 30 min.) fetal bovine serum, 90 /@ml heparin and 100 /.&ml of unpurified bovine endothelial cell growth factor (ECGF). Treatment of H’IG serum or ‘IG-rich liw~rotein: HTG serum or TG-rich lipoprotein was performed, in vitro by adding purified lipoprotein lipase to serum (50 pi/ml) or mixtures containing isolated TG-rich lipoprotein (30 mg/dl cholesterol) and 6% fatty acid depleted bovine serum albumin and incubating the mixture at 37OC for 120 min. HTG sera containing triglycerides levels of 200-500 mg/ dl were used as the substrates for lipoprotein lipase. Usually more than 80% of the triglyceride in HTG serum or TG-rich lipoprotein was hydrolyzed after 2 h incubation at 37°C. The lipolysis mixture containing TG-rich lipoprotein, serum albumin and lipoprotein lipase was subjected to single vertical spin density gradient ultracentrifugation, to separate the remnants of TG-rich lipoprotein from serum proteins or albumin . The lipolyzed HTG serum and its isolated lipoprotein fraction and the remnants of TG-rich lipoprotein was then dialyzed against phosphate buffered saline (0.05 M sodium phosphate-O.15 M NaCl, pH 7.2) to remove KBr and other salts. In one experiment, butylated hydroxytouluene (BHT) as an antioxidant or cupric nitrate as an oxidant agent was included in the dialysis buffer. The level of BHT or copper ions in the dialysis medium was 20 I.IM or 5 /,LM , respectively. The dialyzed samples were then sterilized by filtration through 0.45 pm membranes ( Millipore Corp.Milford, MA). Incubation of serum and lirmxoteins with cultured HUVECk Control (non-lipolyzed) HTG serum and lipolyzed HTG serum was added to HUVECs cultures in serum-free medium (see below) at final concentrations of l%, 2.5%, 5% and 10% (v/v). Control non-lipolyzed TG-rich lipoprotein (0,10,25,50,75 and 100 c(g cholesterol/ml) and lipolyzed and TG-rich lipoprotein (0,10.25,50,75 and 100 pug cholesterol/ml) was added to cultured HUVECs in serum-free medium (Medium 199, with 90 &ml of heparin, 100 /&ml of ECGF, and 0.25% BSA). Control TG-rich lipoprotein or lipolytic remnants of TG-rich lipoprotein was added into culture dish containing serum free medium to give its level to be 0,25,37.5,50,75,100 and 150 c(g cholesterol per ml culture medium. In certain experiments, an increasing amounts of HDL( 0,25,50,75,100 and 150 pg cholesterol/ml) were added to culture dishes containing 75 pg (cholesterol) of lipolyzed TG-rich lipoprotein or BHT at final concentration 20 PM was added into culture medium containing lipolytic remnants of TG-rich lipoprotein. All serum samples to be tested were heat-inactivated at 56’C for 30 min. Determination of wtotoxicityz The total number of cells were determined using phase contrast microscopy and a counting reticle (0.5 x 0.5 cm). Cytotoxicity was expressed as a percentage of the starting cells remaining on the growth surface at various time periods under the various experimental conditions. All data shown in figures and table are a typical representation of 4-5 experiments. In certain experiment, change in the levels of lactate dehydrogenase (LDH) activity from cell-free supernatants were determined spectrophotometrically, as previously described (22). Determination of bid mxoxides : Levels of lipid peroxides of control TG-rich lipoprotein and lipolytic remnants of TG-rich lipoprotein following lipolysis, dialysis and/or incubation with cultured HUVECs were determined by assaying the content of thiobarbituric acid-reacting substance (TBARS), as previously described (23). Briefly, 2 ml of 20% trichloroacetic acid was added to 1 ml of lipoprotein preparation or culture medium containing 100-200 pg (cholesterol) of control or

254

ENDOTHELIUM AND HYPERTRIGLYCERIDEMIA

Vol. 58, No. 3

lipolyzed TG-rich lipoprotein and 0.25% BSA . The samples were incubated for 10 min. at room temperature. After centrifugation of the mixtures at 3000 t-pm for 20 min. the supernatants were decanted and the protein precipitate was washed once with 1 ml of 0.1 M sulfuric acid, 1.5 ml of 0.67% of thiobarbituric acid in 2 M sodium sulfate was added to the mixture. The coupling of lipid peroxides with thiobarbituric acid was carried out by heating the mixtures for 30 min. in a boiling water bath. After cooling, the resulting chromogen was extracted with 2 ml of n-buthyl alcohol. Following subjection of the mixtures at a low speed centrifugation ( 1000 rpm for 10 min.), the absorbance of chromogen in the butanol phase was determined spectrophotometrically at 530 nm. Malonedialdehyde-bisdiethyl acetal (1,1,3,3-tetraetoxypropane) was used as a standard. Oil red 0 staining for liuid inclusions: Following incubation with control nonlipolyzed or lipolyzed TG-rich lipoprotein, HUVECs cultures were washed three times with Dulbecco’s PBS (pH 7.4) and stained with oil red 0 for 15 sec. as previously described (24), rewashed three times with Dulbecco’s PBS (pH 7.0), and the extent of lipid inclusions were determined by light microscopy. Electron Cells and 1% GsO,. Fixed cells floated from the petri dish, stained with uranyl acetate

were fmed with 2 % glutaraldehyde (in 0.2 M phosphate buffer pH 7.0) were dehydrated through a graded ethenol series. After cells were they were infiltrated and embedded in Epon or Polybed, sectioned and and lead citrate, as previously described (25).

RESULTS Effect of kxhzed Hl’G serum on HUVECs. HTG serum obtained from postprandial lipemic donors was lipolyzed and found to be 100% cytotoxic to HUVECs after 24 h when added to the culture medium at a final concentration of 10% (vk) (Figure 1B). Control fasting (non-lipolyzed) HTG serum and control (non-lipolyzed) HTG serum, added to culture medium (1, 2.5, 5 and 10% final concentration) had no cytotoxic effect on cultured HUVECs (Figure 1A).

0

12

24

36

4a

60

72

nur (HOURS)

FIGURE 1. The effect of non-lipolyzed HTG serum (A) and lipolyzed HTG serum (B) on the viability of cultured HUVECs. Specific conditions were: 8.0 x lo4 cells/cm2 in serum-free medium with 0.25% (w/v) BSA The percentage of cells remaining on the growth surface was determined after 6, 12, 24, 36, 48 and 72 h, as described in Methods.

Vol. 58, No. 3

ENDOTHELIUM AND HYPERTRIGLYCERIDEMIA

255

Efket of lioohned ‘IG-rich linomotein on HUVECB Studies of the effect on cultured HUVECs of control (nonlipolyzed TG-rich lipoprotein) and lipolyzed TG-rich lipoprotein from HTG serum showed that 75 c(g cholesterol/ml of lipolyzed TG-rich lipoprotein produced cytotoxicity and the detachment of 100% of the cells from the culture dish. Control non-lipolyzed TG-rich lipoprotein, at a concentration up to 100 lg cholesterol/ml had no apparent cytotoxic effect (Figure 2). FIGURE 2 The effect of lipolized TGrich lipoprotein ( ??) on cultured HUVECs (48 h). Non-lipolyzed TGrich lipoprotein (A) and bovine lipase ( 10 % v/v) (A) were used as controls.

0

10

20

30

40

50

60

70

60

100 110

90

CHOLESTEROL (&ml)

Lactatedehvdrogenase (LJXD activitv in mediunu Changes in number of HUVECs and, level of LDH in the culture dishes containing various levels of control or lipolyzed TG-rich lipoprotein following 48 h incubation were further determined ( Figure 3). As Figure 3 shows, addition of control TG-rich lipoprotein or sublethal doses of lipolyzed TG-rich lipoprotein ( 50 /.rg cholesterol/ml or less) into culture HUVECs had no significant effect on numbers of cells on the growth surface and in level of LDH in culture dishes, but the cytotoxic level ( 75 pg cholesterol/ml) of lipolyzed TG-rich lipoprotein detached the most of cells from the growth surface and caused significant increase of cell-free LDH activity in the culture medium.

,A '"T

\*

0 20

40

60 A

/

cholesterol @9/d)

4 80

FIGURE 3. Cytotoxic effect of lipolyzed TG-rich lipoprotein on cultured endothelial cells as measured by number of cells remaining on the growth surface (A) and LDH realease to the culture medium (B). Cells were incubated for 48 h in serum-free medium containing 0.25 % BSA (w/v) with non-lipolyzed TG-rich lipoprotein ( 0) or lipolized TG-rich lipoprotein ( 4 ). Control cultures ( 0 ) did not non TG-rich contain cholesterol lipoprotein.

Vol. 58, No. 3

ENDOTHELIUM AND HYPERTRIGLYCERIDEMIA

256

in InJvEck. Examination of HWECs following incubation with lipolyzed TGrich lipoprotein for 48 h and subsequent staining of the cells with oil red 0 revealed that the cultured HUVECs incubated with a low, non-toxic dose of lipolyzed TG-rich lipoprotein (50 pg cholesterol/ml) contained numerous lipid inclusions (arrow) (Figure 4A). No lipid inclusions were detected in the cultured HUVECs which had been incubated with control non-lipolyzed TG-rich lipoprotein (50 c(g cholesterol/ml) (Figure 4B).

LiDid accumuIation

FIGURE 4. Phase-contrast photomicrograph of oil red O-stained cultured HUVECs. Cultured HWECs (8.6 x ld cells/cm2) were incubated with 50 fig cholesterol/ml of lipolyzed TG-rich lipoprotein (A) and 50 pg cholesterol/ml of control, non-lipolyzed TG-rich lipoprotein (B) x 520. Representative electron micrographs of lipid inclusions in cultured HUVECs incubated with lipolyzed TG-rich lipoprotein (50 pg cholesterol/ml for 48 h) are shown in Figure 5A. In contrast, the cells which have ken incubated with control (non-lipolyzed) TG-rich lipoprotein have no detectable lipid inclusions (Figure 5B). However, small lipid accumulations were noticed in some cultures of HUVECs incubated with high dosages ( > 100 c(g cholesterol/ml) of non-lipolyzed TGrich lipoprotein ( Figure SC).

100

F

80

-

w 2 z VI E

60 -

& Z

40 -

Y g

26.

bp 0’ 0

, 20

40

60

00

HDL CHOLESTEROL

100 h/ml)

120

140

160

Vol. 58, No. 3

ENDOTHELIUM AND HYPERTRIGLYCERIDEMIA

257

FWXJRE 5. Transmission electron micrographs of cultured HUVECs showing lipid inclusiohs (arrow) incubated with lipolyzed TG-rich lipoprotein (50 pg cholesterol/ml) (A) (original magnification ) x 9,100 and cells incubated with control non-lipolyzed TG-rich lipoprotein (50 pg cholesterol/ml) (B) (original magnification) x 6,700 and control cells (> 100 pg cholesterol/ml) (C) (original magnification) x 5,700.

258

ENDOTHELIUM AND HYPERTRIGLYCERIDEMIA

Vol. 58, No. 3

Inhibition of cvtotoxiciw with HDL: The effect of adding HDL to cultured HUVECs containing cytotoxic levels of lipolyzed TG-rich lipoprotein (75 pg cholesterol/ml) on the viability of HUVECs was examined. Isolated HDL, added at a concentration of 75 pg cholesterol/ml or higher inhibited the cytotoxic effect of the lipolyzed TG-rich lipoprotein (Figure 6). This apparent protective effect of HDL was observed up to 48 h. HDL alone (lo-150 pg cholesterol/ml) had no apparent effect on cultured HUVECs (data not shown).

FIGURE 6. Inhibition of lipolyzed TG-rich lipoprotein cytotoxicity by HDL in cultured HUVECs. Cells were incubated with cytotoxic amounts of lipolyzed TG-rich lipoprotein (75 pg cholesterol/ml) and varying concentrations of HDL for 48 h in serum-free medium as described in Methods. The effect of liDid peroxides on HUVECx In order to determine if the lipid peroxides in lipolyzed TG-rich lipoprotein are a responsible factor for the cytotoxicity, the level of lipid peroxides in control and lipolyzed TG-rich lipoprotein and the effect of antioxidant (BHT) on cytotoxicity of lipolyzed TG-rich lipoprotein was determined. As shown in Table 1, the levels of lipid peroxides of the control and lipolyzed TG-rich lipoprotein before and after dialysis was about the same. The levels of lipid peroxides of both control and lipolyzed TG-rich lipoprotein was significantly increased following its incubation with cultured HUVECs. However,the HUVECs-mediated increase in level of lipid peroxides of control and lipolyzed TG-rich lipoprotein was about the same, suggesting that the lipolysis step has no influence on the susceptibility of lipoprotein to lipid peroxidation. The presence of antioxidant (BHT) during dialysis and during incubation of lipolyzed samples with HUVECs has significantly suppressed the production of lipid peroxides, but failed to inhibit the cytotoxic effect of lipolyzed TG-rich lipoprotein ( Table 1). These data suggest that lipid peroxidation of lipolyzed TG-rich lipoprotein may not be a responsible cytotoxic factor. Further studies on the in vitro oxidation of TG-rich lipoprotein by an oxidizing agent (Cu++ ions) and subsequent interaction of Cu++ -oxidized TG-rich lipoprotein with cultured HUVECs showed that oxidized TG-rich lipoprotein contained about 10 times more lipid peroxides than that of control or lipolyzed TG-rich lipoprotein (Table l), but were less cytotoxic to HUVECs when compared with lipolyzed TG-rich lipoprotein (Figure 2 and 7).

Vol. 58, No. 3

ENDOTHELIUM

AND

259

HYPERTRIGLYCERIDEMIA

TABLE I. Level of thiobarbituric acid reactin substances of TG-rich lipoproteins followlng dBal sis and/or incubation with cultured HUr;ECs

Lipoproteins Control TG-rich Lipoproteins I I I I I * I ” Lipolyred TG-rich Lipoproteins I I I I ” II I II I I

Dlelysts Condltion

Condition of Incubation with HUVECs

TBARS (nM MDA/mg cholesterol) 1.4

PBS PBS+BHT PBS + Cu++ PBS

48h.

-

PBS PBS+ BHT PBS PBS + BHT PBS

-

48h. 48h. 48h. + BHT

2.3 0.9 24.6 5.4 1.2 2.7 1.1 6.3 3.2 2.9

Control or Liplyzed TG-rich lipoproteins separated by density gradient ultracentrifugation were dialyzed against phosphate buffered saline (PBS) or PBS containing 20 HIM BHT or 5 DM Cuperic nitrate for 48 hrs. at 7°C. The level of BHT in culture medium was adjusted to give 1.O KM. TBARS is average of duplicated experiments. The levels of TBARS of incubated culture medium were estimated based on the initial levels of control or lipolyzed TG-rich lipoproteins added into cultrure medium. The initial level of control or lipolyzed TG-rich lipoproteins in culture medium was 100 pg cholesterol per ml culture medium.

1oc 80 60

100 Cholesterol WrN or oxidized TG-rich lipoprotein ( ) on the viability effect of control ( 7 with an indicated amount of contra or Cu++ -oxidized of cultured HWECs. Cells were incu VLDL for L# h in serum-free medium as described in Methods. The levels of lipid peroxides (nM MDA/mg chok+terol) of amtrol and oxidized TG-rich lipoprotein before its addition to culture medium were 2.7 and 21.2 respectively. FIGURE

7.

The

260

ENDOTHELIUM AND HYPERTRIGLYCERIDEMIA

Vol. 58, No. 3

DISCUSSION This study demonstrated that lipolytic remnants of TG-rich lipoprotein produced in are cytotoxic to cultured HUVECs and cause massive lipid inclusions. Since endothelial cell injury is considered to be an early event in atherosclerosis, the cytotoxic remnants of ‘Xi-rich lipoprotein may play an important role in this early process of atherogenesis. Injury to the endothelium, mainly in medium size arteries, is thought to be initiated by a variety of factors such as trauma, toxins, infections, elevated plasma cholesterol, diabetes, smoking and hypertension (26-29). Early case association between the level of control studies indicated a positive, significant univariate triglycerides (TG) and coronary heart disease (2,3), but the effect of elevated TG in serum on endothelial cell function and atherogenesis in humans is less well understood. Several studies have shown that certain VLDL is cytotoxic to cultured endothelial cells (11,12), but the nature of cytotoxic factor(s) associated with this lipoprotein specie has yet to be identified. A number of studies have shown that oxidized low density lipoprotein (LDL) can cause injury in or be cytotoxic to cultured endothelial cells (30,31) and that LDL can be oxidized by the free radicals enumerating from cultured cells (32,33). Clevidence et al.(34) reported that LDL extracted from atherosclerotic lesions of the artery exhibits chemical and physical properties similar to those of oxidized LDL, suggesting that oxidized LDL may be involved in the formation of atherosclerotic lesions in vivo. Although lipid peroxides have been identified in the sera of healthy men, hypercholesterolemic and hypertriglyceridemic subjects (35), sera from these subjects were unable to produce cytotoxicity to cultured endothelial cells (36).It is currently uncertain whether or not the oxidation of LDL or oxidized LDL-induced cytotoxicity (injury of endothelium) occurs in vivo. A number of studies have shown that oxidation of LDL or cytotoxicity induced by oxidized LDL can be inhibited by the presence of HDL or plasma free protein fraction (27,30). Our data indicated that the cytotoxicity produced by the lipolyzed TG-rich lipoproteins may not be a result of lipid peroxidation. Gianturco et al. (11) have previously demonstrated that HTG VLDL, but not the same level of normal VLDL, was cytotolric to cultured bovine aortic endothelial cells. These observations led to the speculation that interaction of VLDL with the vascular endothelium in HTG may cause abnormal intracellular fluxes or accumulation of cholesterol and fatty acids. These studies are consistent with our present observations and suggest a functional abnormality in TG-rich lipoprotein which may be potentially atherogenic and is further exacerbated by lipolysis. Free fatty acids and lysolecithin are known products of lipolysis (37,38) and are potentially cytotoxic. However, we have not yet determined whether the free fatty acids or lysolecithin released during lipolysis are responsible for the cytotoxicity of the lipolyzed HTG serum or TGrich lipoprotein in cultured HUVECs. Previous studies on the effect of lipolyzed serum with cultured macrophages revealed that levels of free fatty acids or lysolecithin, similar to those found at the cytotoxic level of lipolyzed TG-rich lipoprotein were not cytotoxic to the cells (14). The mechanism by which massive lipid inclusions occur in HUVECs exposed to lipolyzed samples obtained from postprandial HTG subjects is still unknown. Since oil red 0 stains both cholesterol esters and triglycerides, we are also not certain of the content of these inclusions in the cytoplasm of HUVECs. Additionally, lipolysis of TG-rich lipoproteins would produce free fatty acids, monoand di-glycerides. The lipolysis products may then enter the HUVECs, perhaps utilizing the plasmalemmal free-fatty acid transporter described by Stremmel et a1.(39). The process of fatty acid transport and its subsequent intra-cellular accumulation, as triglycerides could account, over time, for the marked increase in oil red 0 positive droplets observed. The culture conditions used in the present experiments may produce an acceleration of the lipid-droplet accumulation and cytotoxicity process as compared to the in vivo situation. Whether the lipid accumulation is the direct cause of the cytotoxicity, or if other factors,( eg. di-glyceride accumulation and stimulation of protein kinase C pathways) are involved requires further study. Recent studies by Laughton et al. (40) have shown that sera from patients with significant coronary artery obstruction caused marked cellular accumulation of neutral lipids in cultured arterial smooth cells. This study showed that the serum factor most strongly correlated with intracellular

Vol. 58, No. 3

ENDOTHELIUM

AND HYPERTRIGLYCERIDEMIA

261

accumulation of lipid was Che level of free fatty acids but not the level of total plasma cholesterol or TG or the level of VLDL and LDL. Humans spend a great deal of time in the postprandial HTG state, and the clearance of these postprandial lipoproteins causes a periodic elevation of lipolytic remnant products, including free fatty acids in the circulating blood. We observed that the postlipolysis samples of certain postprandial serum but not the fasting sera from certain normolipidemic subjects were cytotoxic to cultured HUVECs (data not shown). A recent casecontrol study by Simons et al. (41) suggests that in a subset of patients with premature coronary artery disease having normal fasting lipids, a defect in exogenous fat clearance may exist and may play a role in the etiology of the disease. A previous study had revealed that the cytotoxic effect of lipolyzed HTG serum on cultured macrophages allowed localization of cytotoxicity to the remnant-enriched lipoproteins (15). We observed that LDL and HDL isolated from lipolyzed HTG serum, containing lipolytic products of TG-rich lipoprotein, was cytotoxic to cultured HUVBCs, but control HDL and LDL were not (data not shown). Our study revealed that a critical concentration of HDL in culture medium can inhibit the cytotoxicity produced by lipolytic remnants of TG-rich lipoprotein ( Figure 6). Although the mechanism of inhibiting the remnant-associated cytotoxicity by HDL in cultured HUVECs, _ m is not currently clear, the protection of endothelial cell damage resulting from the lipolytic remnants of TG-rich lipoprotein by HDL may be one important mechanism whereby HDL prevents atherosclerosis.

AC!KNOWLEDGEMBN’IS This work was supported in part by National Institutes of Health Grants HL-33500 , HL 37833 and HL-39960 and by the Atherosclerosis Research Unit. We thank Drs. Jere Segrest and Sandra Gianturco for reviewing the manuscript and for invaluable suggestions and comments. The expert technical assistance of Ms. Malera Traylor and Mr Clark Baker are acknowledged. We would also like to express our thanks to Mr.Eugene Arms from the Department of Microbiology for providing the electron micrographs.

REFEXENCES GORDON. T., and MCNAMARU, P.M. Serum 1. KANNEL, W.B., CASTBLLI, W.P., cholesterol, lipoprotein and the risk of coronary heart disease. &. Intern. Med. 74: 1-12, 1971. SCHROIT, A.G., BIERMAN, E.L., and HAZZARD, W.R., 2. GOLDSTEIN, J. L., MOTULSKY, A.G. Hyperlipidemia in coronary heart disease-lipid levels in 500 survivors of myocardial infarction. J.Clin.Invest.52:1533-1577,1973. 3. CARLSON, LA., and BOTTIGER, L.E. Ischemic heart-disease in relation to fasting values of plasma triglycerides and cholesterol. Stockholm prospective study. Lancet 1: 865-868, 1972. 4. KANNEL, W.B., CASTELLI, W.P., and GORDON ,T. Cholesterol in the prediction of atherosclerotic disease. New perspectives based on the Farmingham study. Ann. Intern. Med. 90 (1): 85-91,1979. 5. BRUNZELL, J.D., ALBERS, J.T., CHAIT, A., GRUNDY MCDONALD, G.B. Plasma lipoproteins in familial combined familial hypertriglyceridemia. J.Lipid Res. 24 (2): 147-155, 1983.

, S.M., GROSZEK, E., and hyperlipidemia and monogenic

6. BORRIE,

667, 1969.

P. Type 3 hyperlipoproteinemia.

Brit. Med. J.z665-

262

ENDOTHELIUM AND HYPERTRIGLYCERIDEMIA

Vol. 58, No. 3

7. SCOW, P.O., BRACHET-MACHIE, E.J. and SMITH L.C. Role of capillary endothelium in the clearance of chylomicrons: a model for lipid transport from blood by lateral diffusion in cell membranes. Circ. Res. 39: 149-162, 1976. 8. PATSCH, J.R., GOTTO JR., AM., OLIVERCRONA T., and EISENBERG, S. Formation of high density lipoprotein-2 like particles during lipolysis of very low density lipoproteins in vitro. Proc. Nat. Acad. Sci. U.S.A 75: 4519-4523, 1978. 9. ARBOGAST, B-W., LEE, G. M. and RAYMOND, T.L. In vitro injury of porcine aortic endothelial cells by very low density lipoproteins from diabetic rat serum. -Diabetes 31: 593-599, 1982. 10. ZILVERSMIT, D.B. A proposal linking atherogenesis to the interaction lipoprotein lipase with triglyceride-rich lipoproteins. -Cir. Res. 33:633-637, 1973.

of endothelial

11. GIANTURCO, S.H., ESKIN,S.G., NAVARRI, T.L., SMITH, L.C.,and GOTTO,A.M., Abnormal effect of hypertriglyceridemic very low density lipoproteins on 3-hydroxy 3-methylglutaryl Co A reductase activity and viability of cultured bovine aortic endothelial cells. Biochim. Biophvs. Acta =143-152, 1986. 12. ABRAMS, A, SPEIDEL, M.T., BOOYSE, EM., and SEGREST, J.P. Cytotoxicity of lipolyzed very low density lipoproteins and inhibition of cytotoxicity by high density lipoproteins to endothelial cells in vitro. Fed. Proc. 5: 4313 (Abstr.) 1988. 13. PARKS, J., SPEIDEL, M. T., and SEGREST, J. P., The effect of lipolyzed hypertriglyceridemic human serum and lipolyzed very low density lipoprotein on smooth muscle cells in vitro. Fed. Proc. 4: 4314 (Abstr.) 1988. 14. CHUNG, B. H., SEGREST, J.P., SMITH, K., GRIFFIN, EM., and BROUILLETI’E, C.G. Lipolytic surface remnants of triglyceride rich lipoproteins are cytotoxic to macrophages but not in the presence of high density lipoproteins. A possible mechanism for atherogenesis? J. Clin. Invest. s3: 1363-1374, 1989. I 15. CHUNG, B.H., SEGREST, J.P., CONE, J.T., PFAU, J., GEER, J.C. and DUNCAN, L.A. High resolution plasma lipoprotein cholesterol profiles by a rapid, high volume, semi-automated method. J. Lipid Res. ~1003-1011. 1981. 16. HAVEL, R.J., EDER, HA., and BRAGDON, J.H., The distribution and chemical composition of ultracentrifugally separated lipoproteins in human serum. J. Clin. Invest. %:1345-1353,1962. 17. CHUNG, B.H.,WILKINSON, T., GEER, J.C. and SEGREST, J.P. Preperative and quantitative isolation of plasma lipoproteins: rapid, single discontinuous density gradient ultracentrifugation in vertical rotor. J. Lipid Res. 2:284-291, 1980. 18. LAEMMLI, V. K. Cleavage of structural proteins bacteriophage T4. Nature (London) 227:680-684, 1970.

during

the assembly

of the head of

19. INVERIUS, P. H. and OSTLUND-LINDQUIST, AM. Lipoprotein lipase from bovine milk: Isolation procedure, chemical characterization and molecular weight analysis. J. Biol. Chem. 251:7791-7795, 1976.

Vol. 58, No. 3

ENDOTHELIUM AND HYPERTRIGLYCERIDEMIA

263

20. JAFFE, EA., NACHMAN, R.L., BECKER, C.G., and MINICK ,C.R. Culture of human endothelial cells derived from umbilical veins: identification by morphologic and immunologic criteria. J.Clin. Invest. 2:2745-2756, 1973. 21. BOOYSE, F. M., QUARFOOT, A.J., CHEDIAK, J., MACUAG, T. and STEMERMAN, Characterization and properties of cultured human von Willebrand umbilical vein endothelial Blood 58(4):788-7%, 1981.

M.B. cells.

22. FREEMAN, B. A, YOUNG, S. L., and CRAPO, J. D. Liposome-mediated augmentation superoxide dismutase in endothelial cells prevents oxygen injury. J. Biol. Chem. =12534-12542, 1983. 23. SATOH, K., Serum lipid peroxide in cerebrovascular method. Clin. Cim. Acta 90:37-43, 1978. 24. KRUTH, H.S. Filipin-positive, oil red O-negative cholesterol feeding. Lab. Invest. 50:87-93, 1983.

disorders determined

of

by a new calorimetric

particles in atherosclerotic

lesions induced by

25. ROTH, M.G., COMPANS, R.W., GIUSTI, L., DAVIS, AR., NAYEK, D.P., GETHING, M.J. and SAMBROOK, J. Influenza virus hemoagglutinin expression is polarized in cells infected with recombinant SV-40 viruses carrying cloned hemoagglutinin DNA Cell 33: 435443, 1983. 26. HESSLER, J. H., ROBERTSON, AL., and CHISOLM, G.M. LDL-induced cytotoxicity and its inhibition by HDL in human vascular smooth muscle and endothelial cells in culture. Arteriosclerosis 32:213-229, 1979. 27. TAUBER, J. P., CHENG, J., and GOSPODAROWICZ, D. Effect of high and low density lipoproteins on proliferation of cultured bovine vascular endothelial cells. J. Clin. Invest. @6%-708, 1980. 28. STEINBERG, D. Lipoproteins Arteriosclerosis 2283-301, 1983.

and

atherosclerosis:

A look

back

and a look

29. HERIKSEN, T., EVENSEN, S. A, and CARLANDER, B. Injury to human endothelial in culture induced by low density lipoprotein. Stand. Clin. Invest. 39:351-368, 1979. 30. MOREL, D.W., HESSLER, J.R., and CHISOLM, G.M. Low density lipoprotein induced by free radical peroxidation of lipid. J. Lipid Res. 21070-1076, 1983. 31. HESSLER, J.H., MOREL, D.W., LEWIS, L.J. and CHISOLM, G.H. Lipoprotein lipoprotein-induced cytotoxicity. Arteriosclerosis &357-364, 1983.

ahead.

cells

cytotoxicity

oxidation and

32. STENBRECHER, V.P.,PARTHASARATHY, S., LEAKE, D.S., WI’ITZTUM, J.L., and STEINBERG, D. Modification of low density lipoproteins by endothelial cells involve lipid peroxidation and degradation of low density lipoprotein phospholipids. Proc. Natl. Acad. Sci. USA &l:3883-3887, 1984. 33. HEINECKE, J-W., BAKER, L., ROSEN, H. and CHAIT, A Superoxide-mediated modification 1986. of low density lipoproteins by arterial smooth muscle cells. J. Clin. Invest. x:757-761,

264

ENDOTHELIUM AND HYPERTRIGLYCERIDEMIA

Vol. 58, No. 3

34. CLEVIDENCE, B.A., MORTON, R.E., WEST, G.,DUSEK, D.M. and HOFF, H.F. Cholesterol esterification in macrophages: stimulation by lipoproteins containing apo-B isolated from human aortas. Arteriosclerosis Al%-207, 1984. 35. SZCZEKLIK, A, and GRYGLEWSKI, peroxides and inhibit prostcyclin biosynthesis

R.J. Low density lipoproteins in arteries. Artery 7:488-495,

are carriers for lipid 1980.

36. WALL, R. T., HARLAN, J.M., and STRIKER, G. E. In vitro effects of hyperlipioproteinemic sera on human endothelium: inhibition of endothelial cells migration by familial hypercholesterolemic sera. Thrombosis Res. fl:753-765, 1980. 37. LEAF, A. and 2549-557, 1988.

WEBER,

P.C.

Cardiovascular

effects of n-3 fatty acids.

N. England J. Med.

38. NILSSON-EHLE, P., GARFINKLE, AS. and SCHOTZ, M.S. Lipolytic enzymes and plasma lipoprotein metabolism. Ann. Rev. Biochem. 3667-693, 1980. 39. STREMMEL, W., STROHMAYER, G., BORCHARD, F., KOCHWA, S., and BERK, P.,D., Isolation and partial characterization of a fatty acid binding protein in rat liver plasma membranes. Proc. Natl. Acad. Sci. (U.S.A.) f&4-8, 1985. 40. LAUGHTON, C-W., RUDDLE, D.L., BEDFORD, C.J., and ALDERMAN, E.L. Sera containing elevated non-ester&xl fatty acids from patients with angiographically documented coronary atherosclerosis cause marked lipid accumulation in cultured human smooth muscle-derived cells. Atherosclerosis 2233-246, 1988. 41. SIMONS, L.k,DWYER, T., SIMONS, J.,BERNSTEIN, L., MOCK, P., POONI, N.S., BALASUBRAMANIA, S., BARON, P., BRANSON, J., MORGAN, J., and RAY, P. Chylomicrons and chylomicron remnants in coronary artery disease: a cause control study. Atherosclerosis 65:181185, 1987.