ELSEVIER Atherosclerosis 120 (1996) 1999208

atherosclerosis

Oxidized lipid-mediated alterations in proteoglycan metabolism in cultured pulmonary endothelial cells

Santhini Ramasamy”,‘, David W. Lipkeb, Gibert A. Boissonneault’, Hongtao Guob, Bernhard Hennig””

“Department of’ Nutrition und Food Science, University oj Kentucky, Lexington, KY 40506, USA

“College of Pharmacy, Uniaersit!, of’ Kentucky, Lexington. KY 40506, USA ‘Department of‘ Clinical Sciences, Unirersity cf Kentucky. Lesington. KY 40506. USA

Received 21 July 1994: revision received 6 July 1995: accepted 24 August 1995

Abstract

Compared to cholesterol or linoleic acid (18:2), oxidized lipids such as cholestan-3/I,Scc,6P-triol (triol) and hydroperoxy hnoleic acid (HPODE) markedly impair endothelial barrier function in culture [Hennig and Boissonneault, 1987; Hennig et al. 19861. Because proteoglycans contribute to vascular permeability properties, the effects of cholesterol and 18:2 and their oxidation products, trio1 and HPODE, on endothelial proteoglycan metabolism were determined. While cholesterol was without effect, a concentration-dependent decrease in cellular proteoglycans (measured by ?S incorporation) was observed after exposure to triol. Compared to control cultures, cholesterol reduced mRNA levels for the proteoglycans, perlecan and biglycan. Trio1 had a similar effect on biglycan but not on perlecan mRNA levels. Compared to 18:2, 1,3 and 5 ,uM HPODE depressed cellular proteoglycans. Perlecan mRNA levels were reduced more by HPODE when compared to 18:2. Biglycan mRNA levels were reduced by 3 PM, but not by 5 PM HPODE. These data demonstrate that oxidized lipids such as trio1 and HPODE can decrease cellular proteoglycan metabolism in endothelial monolayers and alter mRNA levels of major specific proteoglycans in a concentration-dependent manner. This may have implications in lipid-mediated disruption of endothelial barrier function and atherosclerosis.

Keywords: Endothelial cells; Oxidized lipids; Proteoglycans; Atherosclerosis

1. Introduction

It is well established that one of the major functions of the vascular endothelium is it’s abil-

* Corresponding author, 204 Funkhouser Building, Univer- sity of Kentucky. Lexington, KY 40506 0054. USA. Tel.: (606) 257 6880; F,ix: (606) 257 3707.

’ Present .address: Department of Cardiology, Emory Uni- versity, Atlanta, GA 30322.

ity to act as a selectively permeable barrier to plasma components and that atherosclerosis in- volves alterations in the barrier properties of the vascular endothelium [ 1,2]. Certain lipid oxidation derivatives such as oxidized cholesterol, e.g.,

0021-91 SO~9~~~Sl5.00 Q 1996 Elsevier Science Ireland Ltd. All rights reserved

ssDl0021-9150(95)05702-x

200 S. Ramasamy rt al. / Atherosclerosis 120 (1996) 199-208

cholestan-3P,5ir,6/?-triol (triol) [3], and oxidized linoleic acid (18:2), hydroperoxy linoleic acid (HPODE) [4], were shown to damage the en- dothelium and to decrease endothelial barrier function when compared to their unoxidized counterparts. The effect of these and certain other lipids on endothelial cell integrity was reviewed recently by Hennig and Alvarado [5].

The barrier property of the vascular endothe- lium is primarily dependent on the ability of these cells to produce prominent basement membrane containing specific proteins, glycoproteins and proteoglycans. Of special interest are the proteo- glycans, as these molecules have unique physico- chemical properties which make them a major determinant of the permeability characteristics of the vascular endothelium. Proteoglycans consist of a protein core to which sulfated carbohydrate chains (glycosaminoglycans) are covalently at- tached. The classification of these molecules is quite complex due to the differences in the carbo- hydrate chains and the core protein structure [6,7]. Generally, heparan sulfate proteoglycan (HSPG) and chondroitin sulfate/dermatan sulfate (CSPG/DSPG) refer to the proteoglycans having heparan and chondroitin/dermatan sulfate as the predominant carbohydrate chains attached to the core protein, respectively. Proteoglycans have been shown to play a role in cell-to-cell and cell-to-matrix interactions due to their affinity to certain extracellular matrix components such as collagen, laminin and fibronectin [8-lo].

The effect of lipids on proteoglycan metabolism in endothelial cells is not well understood. There are few reports where alterations in proteoglycan metabolism were examined after treating cultured endothelial cells with low density lipoprotein (LDL). Vijayagopal et al. [ll] observed an accu- mulation as well as a decrease in the amount of 35S-labeled proteoglycans in human umbilical vein endothelial cells, depending on the concentrations of LDL treatment. A decrease in basement mem- brane proteoglycans has been noted in porcine aortic endothelial cells treated with LDL [12]. Recently, Ramasamy et al. [13] showed that when endothelial cells were treated with high concentra- tions of linoleic acid, there was a disruption of the barrier function of the endothelium and a de-

crease in the amount as well as anionic properties of the heparan sulfate proteoglycans associated with the cell monolayers. Our earlier findings suggest that the oxidized lipids, trio1 [3] and HPODE, [4] cause a significant disruption of en- dothelial barrier function when compared to the respective unoxidized lipids. The effect of these lipids on endothelial barrier function may also be mediated by changes in proteoglycan metabolism.

In the present study, porcine pulmonary artery- derived endothelial cells were exposed to similar concentrations of cholesterol or 18:2, or their oxidized counterparts, and the incorporation of radioactive sulfate into media- and cell-associated proteoglycans was studied. In addition, alter- ations in the expression of specific proteoglycans, namely, perlecan and biglycan, were measured in response to the lipid treatment because of their ability to interact with basement membrane com- ponents. Perlecan, a basement membrane-associ- ated HSPG synthesized by endothelial cells [14], is known to interact with several basement mem- brane components [ 151. Biglycan is a cell surface- associated CSPG [ 161 and could participate in the assembly of the extracellular matrix [17,18].

2. Materials and methods

2.1. Cell culture Endothelial cells were harvested from porcine

pulmonary arteries and cultured in medium (M199; Gibco Laboratories, Grand Island, NY) containing 10% fetal calf serum (FCS) (Hyclone Laboratories, Logan, UT) according to Hennig et al. [19]. The endothelial cultures were character- ized based on their uniform cobblestone morphol- ogy and the expression of angiotensin-converting enzyme (ACE) activity. Cells from passages 5- 15 were used for the experiments.

2.2. Proteoglycan determination 2.2.1. Experimental media, metabolic labeling

and harvesting of cultures. Endothelial cells were plated into &well dishes at a density of 2.0 x 10’ cells/well. Two days after plating when a conflu- ent monolayer was formed, cells were exposed to M 199 + 5% FCS (control) or Ml99 + 5% FCS

S. RamasamJ* et al. I Atherosc~lerosis 120 (1996) 199-208 201

supplemented separately with triol, HPODE, as well as their respective unoxidized lipids such as cholesterol or 18:2 at different doses and/or time points. Simultaneously, the cells also received 20 PCi of Na~‘SO,/ml media (ICN Biomedicals, Costa Mesa., CA). Triol- (Steraloids, Wilton, NH) and cholesterol (Sigma)-containing media were prepared by the methods described by Hennig and Boissonneault [3]. 18:2 (Nu-Check-Prep, Elysian, MN)-containing media was prepared according to Hennig et al. [19]. HPODE was prepared from 18:2 using soybean lipoxygenase according to the procedure reported by Hennig et al. [4]. The HPODE concentration was determined based upon the absorbance value at 234 nm and calcu- lated as absorbance at 234 nm x (extinction coefficient, 2.52 x lo4 M) - ’

The proteoglycans were extracted from the medium and cell monolayers (includes intracellu- lar, pericellular and subendothelial matrix) by the method published by Yanaghishita et al. [20]. Briefly, the culture medium (2.0 ml) was removed into a tube containing solid guanidine hydrochlo- ride (4 M, final concentration) and N-ethyl maleimide (10 mM) at 4°C. The cell monolayer was washed twice with 0.5 ml ice cold Hanks, and the washes were pooled with the media. The third 0.5 ml wash with Hanks was discarded. Cellular proteoglycans were extracted with 2.5 ml of ice cold guanicline hydrochloride buffer (4 M, pH 5.8) containing 2% Triton X-100, 10 mM N-ethyl maleimide, 5 mM phenyl methyl sulfonyl fluoride, 0.1 M 6-amino hexanoic acid, 5 mM benzamidine hydrochloride, 50 mM ethylenediamine tetraacetic acid (EDTA) and 50 mM sodium acetate for 16 h. The samples were stored at - 70°C until further analysis.

2.2.2. Isolation and purification of proteoglycans Samples from each culture medium as well as cell layer were loaded onto Sephadex G-25 columns (PDlO columns, Pharmacia LKB Biotechnology, Inc., Piscataway, NJ) in order to remove un- incorporated sulfate from radiolabeled macro- molecules. The column was first equilibrated with 8 M urea buffer (pH 6.0), containing 0.05 M sodium acetate, 0.15 M sodium chloride and 0.5% Triton X- 100. About 2.5 ml of samples were loaded onto the column and the radiolabeled

proteoglycans were eluted with 3.5 ml (void vol- ume) of 8 M urea buffer solution. The Sephadex purified samples were mixed with carrier chon- droitin and heparan sulfate and precipitated overnight with cetyl pyridinium chloride (CPC) to a final concentration of 1%. The samples were centrifuged at 400 x g for 20 min and the precipitate was dissolved in 0.2 M NaCl, 50 mM Tris-HCl buffer (pH 7.5) containing 0.2% Triton X-100. An aliquot (0.2 ml) was mixed with 10 ml of counting cocktail, 3a70B (Research Products International Corp., Mount Prospect, IL) and total radioactivity incorporated into proteo- glycans was determined using a 1500 Tri-Carb Packard liquid scintillation counter.

For protein determination, a duplicate set of cells were plated simultaneously and treated similarly as described above for proteoglycan analysis; however, 35S was excluded from the media. After 24 h exposure to experimental media, the cell layers were washed free of serum with phosphate buffered saline (PBS) (pH 7.4), dissolved in 0.5% sodium dodecyl sulfate (SDS) and protein was determined according to Lees and Paxman [21].

2.3. Northern blot analysis Endothelial cells grown to confluence on PlOO

dishes were treated with Ml99 + 5% FCS (con- trol) or M 199 + 5% FCS supplemented sepa- rately with triol, HPODE, as well as their respective unoxidized lipids such as cholesterol or 18:2 at different doses. Cells from three PlOO plates were pooled for each concentration. Cells were washed briefly with warm PBS and total RNA was extracted by adding 2.0 ml of a guanidine isothiocyanate solution (4 M guanidine isothiocyanate, 5 mM sodium citrate, 0.1 M B- mercaptoethanol, 0.5% sodium sarcosyl sulfate; pH 7.0) sequentially to each plate within a group. Purification of total RNA and Northern blot analysis were performed according to the methods of Claycomb and Lanson [22]. Hybridization and washes were performed under stringent conditions as described by Claycomb and Lanson [22]. The cDNA probes for perlecan and biglycan were received from Noonan et al. [23] and Fisher et al. P41, respectively. To quantitate differences in

202 S. Ramasamy et al. / Atherosclerosis 120 (1996) 199-208

steady state levels of mRNA for perlecan and biglycan, scanning densitometry was performed on autoradiograms of blots using a TLC Scanner II (CAMAG Scientific Inc., Wilmington, NC) coupled to an SP4290 Integrator (CAMAG Scien- tific Inc.). All values were normalized to densito- metric scans of the 28s ribosomal band from a negative of the original gel.

2.4. Statistical analysis The data were analyzed using SAS (Statistical

Analysis System). Comparisons between treat- ments were made by analysis of variance (ANOVA) with post-hoc comparisons of the means made by Fischer’s least significance differ- ence method. Statistical probability of P < 0.05 was considered significant.

3. Results

Fig. 1 indicates the effect of 10 or 20 PM of trio1 (A) or cholesterol (B) treatment on the per- cent incorporation of radioactive sulfate into pro- teoglycans in the media and cell monolayers when compared to control cultures. The total amount of radiolabeled proteoglycans in the media and cell monolayers in the control cultures were 17.86 f 0.93 x lo’, 11.93 + 0.37 x lo4 counts/min/ mg protein, respectively. Trio1 treatment caused a dose-dependent decrease in the incorporation of radioactive sulfate into macromolecules in the cell monolayers. In the media, however, the amount of radiolabeled proteoglycans was not signifi- cantly altered by treatment with triol. In contrast to triol, exposure of endothelial monolayers to cholesterol did not change 35S incorporation into proteoglycans, either in media or in cell mono- layers (Fig. 1B).

The temporal effect of trio1 (20 PM) on media (A) and cellular proteoglycans (B) is illustrated in Fig, 2. When compared to the corresponding control, trio1 treatment caused an increase in 35S- labeled proteoglycans in the media only at 24 h. There was an initial increase in radiolabeled pro- teoglycans associated with cell monolayers (6 h) followed by a significant decrease at 12 and 24 h in response to triol.

Fig. 3 shows the effect of 1, 3 or 5 PM HPODE

(A) or 3, 5 ,uM 18:2 (B) treatment on the percent incorporation of radioactive sulfate into proteo- glycans in the media and cell monolayers when compared to control cultures. The total amount of radiolabeled proteoglycans in the media and cell monolayers in the control cultures were 17.86 f 0.93 x lo’, and 11.93 + 0.37 x lo4 counts/min/ mg protein, respectively. There was a significant decrease in the media-associated proteoglycans only at 3 PM concentration. Similar to triol, HPODE treatment resulted in a concentration-

0 10

Trio1 (PM)

150 B 0 Media proteoglycans H Cellular proteoglycans

0 10 20

Cholesterol (PM) Fig. 1. Effect of 10 or 20 PM trio1 (A) or cholesterol (B) on the percent incorporation of % into proteoglycans in the media and cell monolayers when compared to control cultures. Ra- dioactivity was expressed per mg protein. Values are mean i S.E.M., with n = 6. *Significantly different when compared to control cultures.

S. Ramasamy et nl. /I( Athrrosclerosis 120 (I 996) 199-208 203

T4 0 Control El Trio1 (2OuM) *

. I ,

ences in the amount of 35S-labeled proteoglycans in the media at any of the time points studied. However, HPODE exposure decreased the amount of radioactivity associated with cell monolayer proteoglycans at 12 and 24 h.

Fig. 5 shows the scanning densitometric analy- sis for perlecan and biglycan mRNA levels in endothelial cultures treated with trio1 (A) and cholesterol (B). Perlecan mRNA levels were in- creased in cultures treated with low doses of trio1 (10 PM), followed by no change at higher concen- tration of trio1 treatment (20 /lM). Trio1 exposure

6 12 24

Time (hr) A 0 Media proteoglycans &I Cellular proteoglycans IL 15

e- I q Control

0 El Trio1 (20 PM)

3 6 12 24 150 c I 0 Media proteoglycans

t&l Cellular proteoglycans Time (hr) Fig. 2. Time response effect of trio1 (20 /tM) on incorporation

of “S into proteoglycans secreted into the media (A) or associated wit:? the cell monolayers (B). Radioactivity was

expressed per mg protein. Values are mean f S.E.M., with

n = 3. *Significantly different when compared to control cul- tures at corresponding time point.

dependent decrease in the amount of radiolabeled proteoglycans associated with cell monolayers. Similar concentrations of 18:2 treatment did not change the radioactive sulfate incorporation into proteoglycans either in media or in cell monolay- ers (Fig. 3E).

Fig. 4 represents the temporal effect of HPODE (5 PM) on the amount of “S-labeled proteogly- cans in the media (A) and the cell monolayers (B). When compared to the corresponding control, HPODE treatment produced no significant differ-

0 0 3

18:2 (PM)

Fig. 3. Etfect of I, 3 or 5 HIM HPODE (A) or 3, 5 /tM linoleic

acid, IS:2 (B) on the percent incorporation of “S into proteo- glycans in the media and cell monolayers when compared to

control cultures. Radioactivity was expressed per mg protein. Values are mean i S.E.M., with II = 6. *Significantly diKerent when compared to control cultures.

S. Ramasamy et al. / Atherosclerosis 120 (1996) 199-208

Cl Control q HPODE (5 PM)

cultures. The effect was more prominent at lower concentrations of HPODE. Biglycan levels were decreased at lower concentrations followed by no change in the expression of biglycan at higher concentrations. When compared to control, 182 treatment caused a decrease in perlecan mRNA levels at 3 PM and an increase in biglycan mRNA levels at 5 PM.

6 12 24

200 ]A I q Perlecan

1 El Biglycan

Time (hr) 12-

.B Cl Control lo- q HPODE (5 PM) *

8-

6-

,&%;: *,\,*,-,* >:,:+: \‘.‘.‘.‘. .;.;.;.;. :+:,:,: .‘\‘.‘.‘. .;.;.;.I

III-

:,:,:,:,: .‘.‘.‘i. .‘.‘.‘.‘. ::::::::: .‘.‘.‘.‘. ::::::::: ::::::::: ,,,, Control Low High

Trio1 Treatment

0 Perlecan kl Biglycan 3 6 12 24

Time (hr)

Fig. 4. Time response effect of HPODE (5 PM) on the incorporation of s5S into proteoglycans secreted into the me- dia (A) or associated with the cell monolayers (B). Radioactiv- ity was expressed per mg protein. Values are mean k S.E.M., with n = 3. *Significantly different when compared to control cultures at corresponding time point.

resulted in a steady decline in the mRNA levels coding for biglycan. Different from triol, choles- terol treatment caused a progressive decrease in both perlecan and biglycan mRNA levels.

Scanning densitometric analysis for perlecan and biglycan mRNA levels in endothelial cultures treated with HPODE (A) and 18:2 (B) are shown in Fig. 6. HPODE treatment caused a decrease in the expression of perlecan mRNA at both 3 and 5 PM concentrations when compared with control

Control Low High

Cholesterol Treatment

Fig. 5. Scanning densitometric values for perlecan and bigly- can mRNA levels in endothelial cultures treated with trio1 (A) and cholesterol (B). Values were corrected for densitometric values of the 28s ribosomal bands of corresponding lanes taken from a negative of the original gel. Low and high concentrations represent IO and 20 ,uM concentrations of the respective lipid.

S. Ramasamy et al. / Atherosclerosis 120 (1996) 199-208 205

III Perlecan 1 KI Biglycan

Control Low High

HPODE Treatment

Cl Perlecan

Control Low High

18:2 Treatment

1

Fig. 6. Scanning densitometric values for perlecan and bigly- can mRNA levels in endothelial cultures treated with HPODE (A) and 18:2 (B). Values were corrected for densitometric values of the 28s ribosomal bands of corresponding lanes taken from a negative of the original gel. Low and high concentrations represent 3 and 5 ,uM concentrations of the respective lipid.

4. Discussion

Impairment in endothelial barrier function or loss in endothelial integrity is considered an ini- tiating event in the pathogenesis of atherosclero- sis. Previous reports from our laboratory showed that when cultured endothelial monolayers were exposed either to trio1 [3,25] or HPODE [4], there was a significant impairment in endothelial barrier function co.mpared to their unoxidized counter-

parts, cholesterol or linoleic acid (18:2), respec- tively. The cause for the loss in endothelial integrity due to these lipids, however, is not well understood. The present study focused on evalu- ating the effect of these lipids on proteoglycan metabolism because proteoglycans play an impor- tant role in regulating the permeability character- istics of the endothelium. Because oxidized lipids are more potent in causing an impairment in endothelial barrier function relative to unoxidized lipids, we compared the effects of similar concen- trations of oxidized as well as unoxidized lipids on proteoglycan metabolism. The alterations in the distribution of radiolabeled proteoglycans be- tween the media and cell monolayers of cultures as well as the changes in mRNA levels of specific proteoglycans, i.e., perlecan and biglycan, were measured.

Upon exposure to oxidized lipids, both trio1 and HPODE caused a significant reduction in the amount of radioactive proteoglycans associated with the cell monolayers. The protocol employed in the methods in harvesting the cellular proteo- glycans extracts the radiolabeled proteoglycans from the pericellular, intracellular and suben- dothelial matrix. The observed reduction in cell- associated proteoglycans by oxidized lipids is similar to the reduction in cell-associated proteo- glycans by a-D-xyloside (a known proteoglycan synthesis inhibitor). Like oxidized lipids, /I-D-

xyloside treatment also caused a significant im- pairment in endothelial barrier function [13]. The oxidized lipid-induced reduction in the cell-associ- ated proteoglycans might in part mediate the de- creased endothelial barrier function caused by these lipids [3,4,25]. We have also reported an impairment in endothelial barrier function in cells treated with IS:2 with a significant decrease in the cell-associated proteoglycans [ 131. However, the concentration of 18:2 used in that report was several fold higher than the fatty acid concentra- tion used in the present study (5 PM vs. 90 PM). At the concentrations tested here, the unoxidized lipids, both 18:2 and cholesterol, did not affect the endothelial barrier function [3,4] and caused no alterations in the incorporation of radioactive sul- fate into the cell-associated proteoglycans.

206 S. Ramasamy et al. / Athrrosclero@s 120 (1996) 1995208

To understand further the mechanisms associ- ated with lipid-mediated alterations in proteogly- can metabolism, we attempted to investigate the effect of oxidized lipids on the expression of some specific proteoglycans, namely perlecan and bigly- can. There are distinct differences in the way by which the mRNA levels of specific proteoglycans are altered by these lipids. Perlecan is a heparan sulfate proteoglycan found in the subendothelial matrix of endothelial cells and binds with fibronectin and other extracellular matrix com- pounds [14,15]. In addition to perlecan, biglycan, a cell surface-associated dermatan sulfate/chon- droitin sulfate proteoglycan, also plays an impor- tant role in interacting with basement membrane components [16- 18,261. The steady state levels of perlecan and biglycan mRNAs are altered by trio1 and HPODE in a dose-dependent manner. We observed an overall decrease in the expression of perlecan in cultures treated with HPODE. Bigly- can mRNA levels were decreased only after 3 ,uM HPODE treatment. On the other hand, trio1 treat- ment caused an increase in the expression of perlecan only at a lower concentration (10 ,uM). Biglycan mRNA levels were decreased at both 10 and 20 ,uM concentrations. In contrast to our expectations, we found changes in perlecan mRNA and biglycan mRNA levels in cells treated with unoxidized lipids without affecting the total amount of 35S-labeled proteoglycans. Thus, we do not find a direct correlation between the mRNA levels of these two specific proteoglycans and the total amount of “S-labeled proteoglycans.

The discrepancy between the mRNA levels of perlecan and biglycan and the total amount of 35S-labeled proteoglycans may be explained by the following reasons. First, the total amount of radi- olabeled proteoglycans measured in this study depends upon the balance between the synthesis and degradation as well as the sulfation pattern of the proteoglycans. It is not clear at this point, once the proteoglycans are synthesized, if there exists a dose-dependent and lipid-specific degrada- tion mechanism involved under the conditions studied. Endothelial matrix proteoglycans could be degraded by heparanase, glucuronidase, sulfa- tases, elastase and myeloperoxidase-H,O,-chloride system [27-301. Decrease in cell-associated pro-

teoglycans without a decrease in perlecan mRNA levels could be attributed to an increase in elastase activity in endothelial cultures treated with triol, as observed in our earlier report [31]. Second, the measured “S-labeled proteoglycans represent not only perlecan and biglycan but also a few other proteoglycans such as decorin, glypican, syndecan and fibroglycan that could be synthesized by en- dothelial cells [15]. McCarthy et al. [32] character- ized a chondroitin sulfate proteoglycan in Reichert’s membrane as a widespread basement membrane component. It remains to be deter- mined if the impairment in endothelial barrier function is related to this basement membrane-as- sociated proteoglycan. Third, once the core proteins of specific proteoglycans are made and glycosaminoglycan chains are anchored to the core proteins, there could be lipid-specific changes in the post-translational mechanisms, e.g., sul- fation of the proteoglycans, as the last step in proteoglycan synthesis [28]. Thus, the net charge of these molecules depends on a complex coordi- nation of regulation in gene expression for the protein core, the number and length of the gly- cosaminoglycan chains, as well as the degree of sulfation. Studies show that when P-D-xylosides were added in increasing concentrations, the num- ber of core proteins available for glycosaminogly- can chain formation was increased, but the length of the synthesized chains were decreased [33,34]. At the same time, when cycloheximide was added, protein synthesis was inhibited, resulting in a re- duction in the number of chains but an increase in the length of the glycosaminoglycans synthesized on the available core proteins [3.5]. Thus, the protein core expression and synthesis play an important role in determining the length and number of the glycosaminoglycan chains by regulating the number of polymer modifying enzymes [36]. Since the glycosyl transferases and sulfotransferases (enzymes responsible for glycosaminoglycan chain synthesis) are mem- brane-bound enzymes, it is speculated that these lipids could affect the activity of these enzymes by altering the membrane fluidity [25,37,38]. Thus, more complex mechanisms are involved in the regulation of proteoglycan metabolism and addi- tional studies are required to understand the lipid specific effects.

Alterations of the HSPG component in the basement membrane in the pathogenesis of some diseases have been investigated. The presence of HSPG/CSPG in the basement membrane con- tributes to the anionic charge properties to this membrane. HSPG is considered a major charge determinant in glomerular basement membrane and plays a crucial role in the pathogenesis of nephrotic syndrome [39]. For example, a defi- ciency of HSPG and a corresponding enhanced glomerular permeability was observed in patho- logical conditions such as diabetic nephropathy [40] and congenital nephrotic syndrome [41]. Interestingly, the reduction of HSPG was found in streptozotocin-induced diabetes [42] and the HSPG levels were brought back to control values by insulin treatment [43].

During oxidative stress, 18:2, present in LDL. can easily get oxidized to HPODE and could further contribute to the development of atherosclerosis by decreasing endothelial barrier function and by causing foam cell formation. Although cholesterol is not easily oxidized, like unsaturated fatty acids, trio1 is found in certain processed cholesterol-rich foods and deep fried fats [3]. Thus the lipid-induced changes on proteo- glycan metabolism could be due to differences in the proteogl.ycan synthesis and degradation, in addition to 1:heir effect on glycosylation and core protein expression. In conclusion, our study sug- gests that the decrease in endothelial barrier func- tion due to l.ipid treatment might be mediated by alterations in proteoglycan metabolism. Oxidized lipids could potentiate these effects and this might have implications in atherosclerosis.

Acknowledgement

Supported in part by grant lPO1 HL36552 from the National Institute of Health and the Kentucky Agricultural Experiment Station (article number 94-9 133).

References

[I] Ross R, Marker L. Hyperlipidemia and atherosclerosis. Chronic hyperlipidemia initiates and maintains lesions by endothelial desquamation and lipid accumulation. Science

1976;193:1094.

PI

[41

[51

[61

171

PI

PI

[]()I

1111

[]?I

[I31

[I41

iI51

[I61

['71

DiCorleto PE, Chisolm GM. Participation of the en- dothelium in the development of atherosclerotic plaque.

Prog Lipid Res 1986;25:365. Hennig B. Boissonneault GA. Cholestan-3/1’,5r.6P-trio1

decreases barrier function of cultured endothelial cell

monolayers. Atherosclerosis 1987;68:255. Hennig B, Enoch C, Chow CK. Linoleic acid hydroperox-

ide increases the transfer of albumin across cultured

endothelial monolayers. Arch Biochem Biophys 1986;248:353.

Hennig B. Alvarado A. Nutrition and endothelial cell integrity: Implications in atherosclerosis. Prog Food Nutr

Sci 1993:17:119.

Hook M, Kjellen L. Johansson S, Robinson J. Cell-sur-

face glycosaminoglycans. Ann Rev Biochem 1984;53:847.

Hassell JR. Kimura JH. Hascall VC. Proteoglycan core protein families. Ann Rev Biochem 1986;55:539.

Gallagher JT, Lyon M, Steward WP. Structure and func- tion of heparan sulphatc proteoglycans. Biochem J

1986:236:31X

Ruoslahti E. Structure and biology of proteoglycans. Annu Rev’ Cell Biol 1988:4:229.

Wight TN. Cell biology of arterial proteoglycans. Arte-

riosclerosis 1989;9: I. Vijayagopal P. Srinivasan SR, Dalferes Jr ER. Radhakr-

ishnammurthy B. Berenson GS. Effect of low density

lipoproteins on the synthesis and secretion of proteogly- cans by human endothelial cells in culture. Biochem J

1988:255:639.

Olgemoller B. Schleicher ED, Schwaabe S, Guretzki H-J. Gerbitz KD. High concentrations of low density lipo-

protein decrease basement membrane associated heparan

sulfate proteoglycan in cultured endothelial cells. FEBS Letters 1990;264:37.

Ramasamy S. Boissonneault GA, Lipke DW. Hem@ B.

Proteoglycans and endothelial barrier function: Effect of linoleic acid exposure to porcine pulmonary artery en-

dothelial cells. Atherosclerosis 1993:103:279.

Saku T, Furthmayr H. Characterization of the major

hcparan sulfate proteoglycan secreted by bovine aortic endothelial cells in culture: Homology to the large molcc-

ular weight molecule of basement membranes. J Biol Chem 1989:264:35 14.

Mertens G. Cassiman J-J, Van den Bcrghe H. Vermylen

J, David G. Cell surface heparan sulfate proteoglycans from human vascular endothelial cells: Core protein char-

acterization and antithrombin III binding properties, J

Biol Chem 1992;267:20435. Bianco P. Fisher LW, Young MF. Termine JD, Robey

PG. Expression and localization of the two small proteo- &cans biglycan and decorin in developing human skelc-

tal and non-skeletal tissues. J Histochem Cytochem

1990;38: 1549.

Kinsella MG. Wight TN. Isolation and characterization of dermatan sulfate proteoglycans synthesized by cultured

bovine aortic endothelial cells. J Biol Chem 1988;2h: 19222.

S. Ramasamy et al. /Atherosclerosis 120 (1996) 199-208 208

V81

u91

[201

PII

P21

~231

]241

PI

1261

~271

WI

~291

[301

Kinsella MG, Wight TN. Formation of high molecular weight derrnatan sulfate proteoglycan in bovine aortic endothelial cell cultures: Evidence for transglutaminase- catalyzed cross-linking to fibronectin. J Biol Chem 1990;265:17891. Hennig B, Shasby DM, Fulton AB, Spector AA. Expo- sure to free fatty acid increases the transfer of albumin across cultured endothelial monolayers. Arteriosclerosis 1984;4:489. Yanaghishita M, Midura RJ, Hascall VC. Proteoglycans: Isolation and purification from tissue cultures. In: Gins- burg V (ed). Methods in Enzymology, Vol. 138, Academic Press, New York, 1987;278-289. Lees MB, Paxman S. SDS-Lowry protein estimation for lipid containing samples. Anal Biochem 1972;47: 184. Claycomb WC, Lanson NA Jr. Proto-oncogene expres- sion in proliferating and differentiating cardiac and skele- tal muscle. Biochem J 1987;247:701. Noonan DM, Horigan EA. Ledbetter SR, Vogeli G, Sasaki M, Yamada Y, Hassell JR. Identification of cDNA clones encoding different domains of the basement membrane heparan sulfate proteoglycan. J Biol Chem 1988;263: 16379. Fisher LW, Termine JD, Young MF. Deduced protein sequence of bone small proteoglycan I (biglycan) shows homology with proteoglycan II (decorin) and several non- connective tissue proteins in a variety of species. J Biol Chem 1989;264:457 1. Ramasamy S, Boissonneault GA, Hennig B. Oxysterol-in- duced endothelial cell dysfunction in culture. J Am Coll Nutr 1992;11:532. Jarvelainen HT, Kinsella MG, Wight TN, Sandell LJ. Differential expression of small chondroitin/dermatan sul- fate proteoglycans, PGI/biglycan and PGII/decorin, by vascular smooth muscle and endothelial cells in culture. J Biol Chem 1991;266:23274. Godder K, Vlodavsky A, Eldor A, Weksler BB, Haimovitz-Friedman A, Fuks Z. Heparanase activity in cultured endothelial cells. J Cell Physiol 1991;148:274. Sampaio LO, Dietrich CP, Colburn P, Buonassisi V, Nader HB. Effect of monensin on the sulfation of hep- aran sulfate proteoglycan from endothelial cells. J Cell Biochem 1992;50: 103. Klebanoff SJ, Kinsella MG, Wight TN. Degradation of endothelial cell matrix heparan sulfate proteoglycan by elastase and the myeloperoxidase-H,O,-chloride system. Am J Pathol 1993;143:907. Vijayagopal P. Enhanced synthesis and accumulation of proteoglycans in cholesterol-enriched arterial smooth muscle cells. Biochem J 1993;294:603.

[31] Toborek M, Hennig B. Vitamin E attenuates induction of elastase-like activity by tumor necrosis factor-x, cholestan-38, 5c(, 6/J-trio1 and linoleic acid in cultured endothelial cells. Clin Chim Acta 1993;215:201.

[32] McCarthy KJ, Accavitti MA, Couchman JR. Immuno- logical characterization of a basement membrane-specific chondroitin sulfate proteoglycan. J Cell Biol 1989;109:3187.

[33] Robinson HC, Brett MJ, Tralaggan PJ, Lowther DA,

1341

[351

1361

1371

[381

1391

1401

[411

~421

]431

Okayama M. The effect of D-xylose, B-D-xylosides and P-D-galactosides on chondroitin sulfate biosynthesis in embryonic chicken cartilage. Biochem J 1975;148:25. Lohmander S, Madsen K, Hinek A. Secretion of proteo- glycans by chondrocytes. Influence of colchicine, cytocha- lasin B and P-D-xyloside. Arch Biochem Biophys 1979;192:148. Kimura JH, Caputo CB, Hascall VC. The effect of cyclo- heximide on synthesis of proteoglycans by cultured chon- drocytes from the swarm rat chondrosarcoma. J Biol Chem 1981;256:4368. Coster L, Hernnas J, Malmstrom A. Biosynthesis of dermatan sulfate proteoglycans: The effect of fi-D-xy- loside addition on the polymer modification process in fibroblast cultures. Biochem J 1991;276:533. Chatelain P, Laurel R, Gillard M. Effect of amiodarone on membrane fluidity and Na + /K + ATPase activity in rat-brain synaptic membranes. Biochem Biophys Res Commun 1985;129:148. Tanaka T, Hidaka T, Ogura R, Sugiyama M. Changes of membrane fluidity and Na + ,K + -ATPase activity dur- ing cellular differentiation in the guinea pig epidermis. Arch Dermatol Res 1988;280:29. Mynderse LA, Hassell JR, Kleinman HK, Martin GR, Martinez-Hernandez A. Loss of heparan sulfate proteo- glycan from glomerular basement membrane of nephrotic rats. Lab Invest 1983;48:292. Carrie BJ, Myers BD. Proteinuria and functional charac- teristics of the glomerular barrier in diabetic nephropathy. Kidney Int 1980;17:669. Vernier RL, Klein DJ, Sisson SP, Mahan JD, Oegema TR, Brown DM. Heparan sulfate-rich anionic sites in the human glomerular basement membrane. Decreased con- centration in congenital nephrotic syndrome. N Engl J Med 1983;309:1001. Kanwar YS, Rosenzweig LJ, Linker A, Jakubowski ML. Decreased de novo synthesis of glomerular proteoglycans in diabetes: Biochemical and autoradiographic evidence. Proc Nat1 Acad Sci USA 1983;80:2272. Rohrbach DH, Wagner CW, Star VL, Martin GR, Brown KS, Yoon J-W. Reduced synthesis of basement membrane heparan sulfate proteoglycan in streptozo- tocin-induced diabetic mice. J Biol Chem 1983;258:11672.