Role of Heparan Sulfate Proteoglycans in the Binding and Uptake of ...

THE JOURNAL OF BIOLOGICAL CHEMISTRY 0 1993 by The American Society for Biochemistry and Molecular Biology, Inc. Vol. 268, No. 14, Issue of May 15, pp. 10160-10167,1993

Printed in U. S. A.

Role of Heparan Sulfate Proteoglycans in the Binding and Uptake of Apolipoprotein E-enriched Remnant Lipoproteins by Cultured Cells*

(Received for publication, October 14, 1992, and in revised form, January 22, 1993)

Zhong-Sheng JiS, Walter J. Brecht, R. Dennis Miranda, M. Mahmood Hussains, Thomas L. InneraritySll, and Robert W. MahleySlI 11 ** From the Gladstone Institute of Cardiovascular Disease, the $Cardiovascular Research Institute, the Departments of TPathology and IlMedicine, University of California, S u n Francisco, California 94141-9100

Addition of apolipoprotein (apo) E to rabbit B-very low density lipoproteins (8-VLDL) has been shown to result in a marked enhancement of their binding and uptake by various cell types. Apolipoprotein E binds to lipoprotein receptors and proteoglycans. To distin- guish between apoE binding to these sites, cells were treated with heparinase. Heparinase treatment of receptor-negative familial hypercholesterolemic (FH) fibroblasts and human hepatoma cells (HepG2) released 30-40% of newly synthesized cell surface s%- labeled proteoglycans and decreased the binding of 8- VLDL+apoE to FH and normal fibroblasts and HepG2 cells by more than 80%. Furthermore, heparinase treatment significantly decreased the uptake of fluorescently labeled &VLDL+apoE by HepG2 cells and decreased cholesteryl ester synthesis in FH fibroblasts by 75%. Likewise, canine chylomicron remnants enriched in apoE demonstrated enhanced binding that was 80% inhibited by heparinase treatment of HepG2 cells. Heparinase treatment did not affect j3-VLDL (without added apoE) or low density lipoprotein (LDL) binding to these cells or the binding activity of 6- VLDL+apoE to the LDL receptor-related protein (LRP) or to the LDL receptor on ligand blots. Chinese hamster ovary (CHO) mutant cells lacking the synthesis of either heparan sulfate (pgsD-677) or all proteoglycans (pgsA-745) did not display any enhanced binding of the p-VLDL+apoE. By comparison, wild-type CHO cells demonstrated enhanced binding of 8- VLDL+apoE that could be abolished by treatment with heparinase. These mutant cells and wild-type CHO cells possessed a similar amount of LRP, as determined by ligand blot analyses and by cY2-macroglobulin binding, and possessed a similar amount of LDL receptor activity, as determined by LDL binding. Therefore, we would interpret these data as showing that heparan sulfate proteoglycan may be involved in the initial binding of the apoE-enriched remnants with the subsequent involvement of the LRP in the uptake of these lipoproteins. It remains to be determined whether the heparan sulfate proteoglycan can function by itself in both the binding and internalization of the apoE-enriched remnants or whether the proteoglycan is part of a complex with LRP that mediates a two-step proc-

* This work was supported in part by National Institutes of Health Program Project Grant HL41633. 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 accord- ance with 18 U.S.C. Section 1734 solely to indicate this fact.

§ Present address: Dept. of Biochemistry, Medical College of Penn- sylvania, 3300 Henry Ave., Philadelphia, PA 19129.

** To whom correspondence should be addressed Gladstone Inst. of Cardiovascular Disease, P. 0. Box 419100, San Francisco, CA 94141-9100. Tel.: 415-826-7500; Fax: 415-285-5632.

ess, i.e. binding and subsequent internalization by the receptor.

Chylomicrons are synthesized by the intestine to transport dietary triglycerides and cholesterol. After these very large lipoproteins enter the circulation through the thoracic duct, their triglyceride core is hydrolyzed, converting them to chylomicron remnants (for review, see Ref. 1). The chylomicron remnants are then cleared from the blood, primarily by the liver. Several steps in the hepatic uptake of chylomicron remnants from the blood are well defined, whereas other aspects of the pathways remain to be elucidated. Biochemical and genetic investigations have shown that apolipoprotein (apo)’ E is the ligand that mediates the uptake of remnants by the liver (for review, see Refs. 2 and 3). Individuals with receptor-defective mutant apoE or no apoE accumulate very low density lipoprotein (VLDL) remnants and chylomicron remnants, collectively referred to as P-VLDL, in their plasma (for review, see Refs. 3 and 4). These remnant lipoproteins are also induced in mammals fed a high cholesterol diet.

Although there is no doubt about the importance of apoE in chylomicron remnant clearance, other aspects of the pathway are less certain. The postulated initial binding of remnants in the space of Disse has not been clearly demonstrated (3). Studies have shown that although the low density lipoprotein (LDL) receptor may participate in the uptake of chylomicron remnants (5), these receptors are not essential for the rapid clearance of radioactive chylomicron remnants injected into animals (for review, see Refs. 3 and 6). Several studies have indicated that the LDL receptor-related protein (LRP), which is identical with the a2-macroglobulin receptor (7) , is involved in the clearance of apoE-enriched remnants (8-12). However, the kinetics of the initial phase of the clearance suggested to us that some other component is re- sponsible for the initial sequestration of chylomicron remnants.

We postulated that a candidate site for the initial binding of chylomicron remnants is heparan sulfate proteoglycans on the surface of hepatocytes (for review, see Ref. 3). Immuno- chemical studies have demonstrated the presence of large quantities of heparan sulfate proteoglycans in the space of Disse (13). The major proteins involved in chylomicron ca- tabolism, including apoE, lipoprotein lipase, and hepatic lipase, have all been shown to bind to heparan sulfate proteogly-

The abbreviations used are: apo, apolipoprotein; VLDL, very low density lipoprotein(s); FH, familial hypercholesterolemia; LDL, low density lipoprotein(s); CHO, Chinese hamster ovary; LRP, LDL receptor-related protein; FBS, fetal bovine serum; DiI, 1,l’-dioctadecyl-3,3,3’,3’-tetramethylindocarbocyanine.

10160

Remnant Lipoprotein Binding to Heparan Sulfate 10161

cans (1,2). Moreover, lipoprotein lipase potentiates the binding of chylomicron remnants to the LRP receptor (14). Thus, by binding chylomicron remnants, cell surface heparan sulfate proteoglycans could bring all the participants in the chylomicron remnant pathway together in the space of Disse where lipases could further hydrolyze the lipoprotein particles and where apoE secreted by hepatocytes could be accessible for association with the lipoproteins (3). The remnants enriched with apoE may then be more readily internalized by the LRP receptor.

The purpose of the present study was to determine whether remnant lipoproteins bind to heparan sulfate proteoglycans on t he cell surface and whether the heparan sulfate proteoglycan-remnant interaction plays a major role in the uptake of remnants by cultured cells i n uitro.

MATERIALS AND METHODS

Preparation of Lipoproteins-Rabbit 0-VLDL (d = 1.006 g/ml) were isolated from New Zealand White rabbits fed a high fat, high cholesterol diet for 4 days. Rabbit chow (Purina Complete Blend 5315, Purina Mills, St. Louis, MO) was supplemented with 10% coconut oil (Sigma) and 0.5% cholesterol (Zeigler Brothers, Gardners, PA), as described by Kowal et al. (8). The lipoproteins were centri- fuged twice at d = 1.006, filtered, and either iodinated by the method of Bilheimer et al. (15) or biotinylated by the method of Kowal et al. (8). Canine chylomicron remnants were obtained from the thoracic duct lymph of normal mongrel dogs (20-30 kg), as described (16). Chylomicron remnants were prepared by injecting the chylomicrons (200 mg of triglyceride/kg of body weight) into functionally hepatec- tomized rabbits and allowed to circulate for 30 min. The remnants were isolated from the rabbit plasma by ultracentrifugation, washed twice at d = 1.006 by ultracentrifugation, and then subjected to 3-4% agarose column chromatography (16). The remnants were iodinated by the iodine monochloride method (17). Human LDL (d = 1.02-1.05 g/ml) were obtained by centrifugation of blood from fasted normal volunteers as described (17) and iodinated by the iodine monochloride method (17).

Labeling of LDL and 0-VLDL with l,l'-dioctadecyl-3,3,3',3'-tetramethylindocarbocyanine (DiI) was carried out by adding the lipoproteins (1-2 mg) to 2 ml of d > 1.21 g/ml lipoprotein-deficient plasma (0.5 mg/ml), filtering the mixture (0.45-pm filter for LDL, 0.8-wm for 0-VLDL), adding 100 p1 of DiI in dimethyl sulfoxide (3 mg/ml), and incubating in the dark for 8 h at 37 "C (12). The lipoproteins were overlaid with phosphate-buffered saline and centri- fuged at 40,000 rpm for 16 h at 4 "C. The LDL and 0-VLDL fractions were dialyzed against saline. The lipoproteins were then filtered before use.

The human apoE3 and rabbit apoE used in this study were prepared by the procedure described (18). The 0-VLDL or chylomicron remnants plus excess exogenous apoE (0-VLDL+apoE) and chylomicron remnants+apoE were incubated at 37 "C for 1 h, as described earlier (lo), before use.

Anti-human LDL Receptor Antibody-Rabbit anti-bovine LDL receptor antibody was prepared following the procedure described previously (19).

az-Macroglobulin-Active az-macroglobulin (provided by Dr. F. R. Maxfield, Columbia University, New York), was labeled with IODO- GEN (Pierce Chemical Co.) as described (12). The az-macroglobulin was activated by treatment with methylamine.

Cultured Cells-Human hepatoma cells (HepG2) were maintained in minimal essential medium containing 10% fetal bovine serum (FBS). The LDL receptor-negative fibroblasts were obtained from persons with familial hypercholesterolemia (FH) with a 10-kb dele- tion in the LDL receptor gene (provided by Dr. J. Davignon, Clinical Research Institute, Montreal) and were grown in Dulbecco's modified Eagle's medium supplemented with 10% FBS. The FBS was changed to 10% human lipoprotein-deficient serum 48 h before each experiment. Mutant pgsA-745 (xylose transferase-deficient) Chinese hamster ovary (CHO) cells and mutant pgsD-677 (AT-acetylglucosamine transferase- and glucuronic acid transferase-deficient) CHO cells were provided by Dr. J. D. Esko (University of Alabama, Birming- ham). The former strain does not produce any glycosaminoglycans, and the latter does not produce any heparan sulfate, replacing the missing heparan sulfate with chondroitin sulfate (20). The wild-type

and mutant CHO cells were grown in F12 medium containing 7.5% FBS.

Tissue Culture Assays-Binding experiments were performed at 4 "C for 3 h (17). Nonspecific binding was determined by adding 50- or 100-fold excess unlabeled ligands. After the incubation with the lipoproteins, the cells were washed and then dissolved in 0.1 N NaOH, and their radioactivity and protein concentration were measured (21).

To determine cholesteryl ester formation, cells were incubated with or without lipoproteins at 37 "C for 3 h, and then 0.2 mM ['*C]oleate was added into the incubation medium for an additional 2 h. The esterified cholesterol in the cells was extracted and separated by thin layer chromatography after the incubation (17). The amount of ["CC] oleate incorporated was determined by scintillation counting.

Fluorescently labeled lipoprotein uptake was determined by incubating the cells with DiI-labeled LDL or 0-VLDL at 37 "C for 1-4 h. The cell monolayers were washed with 0.1 M phosphate-buffered saline (pH 7.4). In certain studies, after the cells were incubated with the lipoproteins, they were treated with 0.005% trypsin to remove extracellular DiI-labeled lipoproteins or those bound to the cell surface. Then the cells were fixed with 4% paraformaldehyde and observed under a fluorescent microscope (Zeiss MC63).

The heparinase (EC 4.2.2.7) (Sigma, heparinase I, catalog number H 2519) used to treat the cells in culture was dissolved in sterile 0.15 M NaCl and used immediately. Periodically, heparinase activity was assayed using the colorimetric assay described by Khan and Newman (22). Heparinase was added to a stock solution of the specific tissue culture media used for each cell type and then added to the cells. The activity of the heparinase and the time of incubation at 37 "C was varied as reported for each experiment. The cells in culture were washed three times with media to remove heparinase, and the specific assay was performed. Heparitinase (EC 4.2.2.8) (ICN, catalog number 190102) was solubilized and used in the assays as described for heparinase.

Ligand Blot Analysis-Membrane proteins of CHO cells and rat liver partially purified by DEAE-cellulose chromatography (8, 23) were separated on 3-8% or 3-20% sodium dodecyl sulfate-polyacryl- amide gels and then transferred to Immobilon-P membranes (Milli- pore, Bedford, MA). The membrane was cut into several strips. The strips were incubated with 50 mM Tris, pH 8.0, containing 15 mM NaC1, 2 mM CaClZ, and 50 mg/ml bovine serum albumin (Buffer 1) for 1 h, followed by incubation with biotin-labeled 0-VLDL+apoE or 'zSI-labeled az-macroglobulin for 1-2 h at room temperature or 37 "C. The biotin-labeled P-VLDL+apoE strips were then washed, and 1 million cpm of 'z51-labeled streptavidin (Pierce Chemical Co.)/ml was applied for 1 h at room temperature. All strips were then washed, dried, and subjected to autoradiography (12). Some nitrocellulose strips were treated with heparinase (0, 1, 5, 15, 30, 80, or 200 units/ ml incubated at 37 "C for 2 h) and were used in ligand blot analyses. Streptavidin was labeled with '*'I using IODO-GEN as described by Kowal et al. (8).

Rat Liuer Proteoglycan Isolation-Proteoglycans were isolated from rat liver as described by Soroka and Farquhar (24). Before isolation of the proteoglycans from the rats, they were injected with 2 mCi of [35S]sulfate, and the animals were sacrificed 4 h after the last injec- tion. The extract was fractionated by ion exchange chromatography using Q-Sepharose (24). The crude liver proteoglycan was used in the studies reported. Ligand blot analysis was performed as described earlier.

RESULTS

Effects of Heparinase on Binding and Uptake of Lipoproteins by Cultured Cells-Excess apoE enhanced 0-VLDL binding to HepG2 cells, FH fibroblasts, and normal fibroblasts. How- ever, after these cells were treated with increasing concentra- tions of heparinase, the enhanced binding of P-VLDL+apoE was markedly inhibited. As shown in Fig. lA , the enhanced binding of P-VLDL+apoE to HepG2 cells was decreased by approximately 67% with 0.5 units/ml of heparinase (37 "C, 2 h). At 3 units/ml of heparinase, only 18% of the binding of @- VLDL+apoE remained. This is approximately the level of binding observed with P-VLDL without added apoE. In an additional study, P-VLDL+apoE binding to HepG2 cells was decreased 74% at 0.5 units/ml and 82% at 3 units/ml of heparinase. Furthermore, as shown in Fig. lB, the enhanced binding of P-VLDL+apoE to FH fibroblasts, which are devoid

10162 Remnant Lipoprotein Binding to Heparan Sulfate

FIG. 1. Effects of heparinase treatment on lipoprotein binding to cultured cells. A , the effect of increasing concentration of heparinase treatment (37 "C, 2 h) on the binding of @- VLDL, 0-VLDL+apoE, and LDL to HepG2 cells. Binding was conducted at 4 "C for 3 h using lZ6I-LDL (m, 2 pg of protein/ml), Iz5I-@-VLDL ( 0 , 2 pg of protein/ml), and 1251-@-VLDL+apoE (0, 2 pg of P-VLDL protein/ml + 3 pg of human apoE3). Inset, effect of time of incubation of the HepG2 cells with 3 units of heparinase on the binding of p- VLDL+apoE. B, the effect of increasing concentration of heparinase on the binding of @-VLDL+apoE to human FH fibroblasts. Binding was conducted at 4 "C for 3 h using '251-@-VLDL+apoE ( 0 , 2 pg of p-VLDL protein/ml + 3 pg of human apoE3). Data are presented as means f S.D. obtained in three separate experiments.

250i 200

A.

Heparinase (units/ml)

1007 B.

I I I I I I

0.5 1 .o 1.5 2.0 2.5 3.0 Heparinase (units/ml)

1007 B.

of LDL receptors, was decreased by more than 90% when these cells were treated with heparinase (37 "C, 2 h). The time course of the effect of heparinase (3 units/ml, 37 "C) on HepG2 cells is shown in Fig. 1A (inset). Inactivation of heparinase activity by heat treatment (80 "C, 30 min) abolished the ability of heparinase to decrease the binding of p- VLDL+apoE to cultured cells (not shown).

To establish that the results obtained with P-VLDL were representative of those obtained with chylomicron remnants, canine chylomicron remnants (1 pg of remnant protein) and chylomicron remnants with added human apoE (1 pg of remnant + 1.5 pg of apoE) were used in binding studies with HepG2 cells. The binding of the chylomicron remnants enriched in exogenous apoE was 19 times that of the remnants without added apoE (345 versus 18 ng of remnant protein/mg of cell protein). Treatment of the cells with heparinase (3 units/ml, 37 "C, 2 h) decreased the binding of the remnants+apoE by approximately 80% (data not shown), but had no effect on the binding of the remnants without added apoE. These results are essentially identical with those obtained with the p-VLDL+apoE.

Heparinase treatment of HepG2 cells did not significantly affect the binding of P-VLDL (without added apoE) or LDL (Fig. 1A). Furthermore, as shown in Fig. 2 A , even when heparinase was increased to 100 units/ml, there was no significant decrease in either P-VLDL (without added apoE) or LDL binding. In addition, the increased concentration of heparinase (100 units/ml) did not further reduce the binding of p-VLDL+apoE to HepG2 cells. As expected, the binding of P-VLDL (without added apoE) and LDL to the cells was inhibited by the LDL receptor antibody by 80 and 85%, respectively (Fig. 2B). However, the binding of the p- VLDL+apoE to the HepG2 cells was inhibited about lo%,

Heparinase (unitdml)

confirming that the enhanced binding of P-VLDL+apoE was not mediated primarily by the LDL receptor.

Experiments were also performed to determine the effects of heparinase treatment on the uptake of fluorescently labeled lipoproteins by HepG2 cells using DiI, a visual marker for lipoprotein uptake by cells in culture. Fig. 3A shows that DiI- labeled P-VLDL+apoE was internalized by HepG2 cells, re- sulting in intense punctate fluorescent labeling of the cells. The intensity of the DiI fluorescence associated with P- VLDL+apoE uptake was significantly diminished in the HepG2 cells treated with heparinase (3 units/ml, 37 "C, 2 h) (Fig. 3B); however, no apparent changes were observed in DiI-labeled LDL and /I-VLDL fluorescence in heparinase- treated HepG2 cells (Fig. 3, C-F).

Previous studies have shown that [14C]oleate incorporation into cholesteryl [14C]oleate is a sensitive assay for the degradation of LDL internalized by the LDL receptor (17) and of P-VLDL+apoE presumably internalized by the LRP/a2-macroglobulin receptor (8, 11, 12). In FH fibroblasts, P-VLDL (without added apoE) did not increase cholesteryl ester formation, whereas the addition of apoE to P-VLDL stimulated cholesteryl ester synthesis 40-fold (8). As shown in Fig. 4, P- VLDL+apoE enhanced cholesteryl ester formation in FH fibroblasts, and this markedly decreased in the cells after heparinase treatment. Heparinase treatment had no effect on the base-line synthesis of cholesteryl ester in the absence of added lipoproteins (Fig. 4).

Effect of Heparinase on the LRP and LDL Receptors- Consideration was given to the possibility that heparinase treatment directly affected the LRP/az-macroglobulin receptor. Studies were performed in FH fibroblasts and HepG2 cells to determine whether heparinase altered the binding of '251-a2-macroglobulin to these cells in culture. As shown in


250 r . U A.

5 e 2 , - 25

0 Heparinase (3 units) .- m No heparinase

d= 2: a, 100 - gg

Z h l 5 0 - 0

Heparinase (100 units)

5 0 - VI- .- o m m

LDL P-VLDL P-VLDL +

APO-E

U 250- B.

C , g.5 200

- a, .- 2 150 d=

- 0 Anti-LDL receptor

nm 100 2:

? ? 5 0 -

-

0 =No antibody

.- Q m - r

0 LDL P-VLDL P-VLDL

+ APO-E

FIG. 2. Effects of heparinase treatment and LDL receptor antibody on the binding of '"I-LDL, 12'I-&VLDL, a n d IzaI-j3-

VLDL+apoE to HepG2 cells. A , effect of heparinase treatment on "'1-LDL, '2sI-P-VLDL, and '2sI-P-VLDL+apoE to HepG2 cells. The cultured cells were treated with 3 or 100 units/ml of heparinase (37 "C, 2 h). The binding was performed a t 4 "C for 3 h using "'1- LDL (2 pg of protein/ml), '251-P-VLDL (2 pg of protein/ml), and '"1- 0-VLDL+apoE (2 pg of protein/ml + 3 pg of human apoE3). These data represent two separate experiments. B, effect of LDL receptor antibody on the binding of lZ5I-LDL, '2sI-p-VLDL, and 12sI-& VLDL+apoE to HepG2 cells. The HepG2 cells were incubated with 100 pg/ml of rabbit anti-human LDL receptor antibody a t 4 "C for 1 h. The '"I-LDL, '2sI-fl-VLDL, or '*'I-P-VLDL+apoE were then added to the incubation medium, and the cells were incubated a t 4 "C for an additional 2 h. Data are presented as the means -C S.D. obtained in three separate experiments.

Fig. 5, for a representative study, heparinase treatment did not significantly affect the binding. A summary of the data from several studies using FH fibroblasts and HepG2 cells is presented in Table I. Additional studies demonstrating the lack of an effect of heparinase treatment on a*-macroglobulin binding in normal fibroblasts are presented later in the text.

Furthermore, studies were performed to determine whether heparinase treatment had an effect on P-VLDL+apoE binding to the LRP/a2-macroglobulin receptor on ligand blots. A crude preparation of rat liver membranes containing the LRP and LDL receptors was transferred to blots, incubated with or without heparinase, and probed with biotin-labeled @- VLDL+apoE. There does not appear to be a major effect of heparinase treatment on p-VLDL+apoE binding to either the LRP or the LDL receptor even a t high levels of heparinase (30 units/ml) (Fig. 6). In additional studies, heparinase a t 200 units/ml did not appear to affect p-VLDL+apoE binding (not shown).

Effect of Heparinase on Cell Surface and Hepatic Proteogly- cans-To demonstrate that heparinase treatment decreased @-VLDL+apoE binding to cultured cells by the hydrolysis of glycosidic linkages in heparan sulfate (25, 26), we established that heparinase treatment of FH fibroblasts and HepG2 cells resulted in the removal of 35S-labeled cell surface proteoglycans under the conditions used in our studies (3 units/ml, 37 "C, 2 h). Cultured cells were incubated with [35S]sulfate for

FIG. 3. Effect of heparinase treatment on fluorescently labeled lipoproteins by HepG2 cells. A , uptake of DiI-labeled p- VLDL+apoE (5 pg of Ij-VLDI, protein + 7.5 pg of human apoE/ml) by untreated HepG2 cells. H , uptake of DiI-labeled p-VLDL+apoE by cells treated with heparinase (3 units/ml, 37 "C, 2 h). C, uptake of DiI-labeled p-VLDL (5 pg of p-VLDL protein/ml) by untreated cells. D, uptake of DiI-labeled fl-VLDL by cells treated with heparinase (3 units/ml, 37 "C, 2 h). E, uptake of DiI-labeled LDL (5 pg of LDL protein/ml) by untreated cells. F, uptake of DiI-labeled LDL by cells treated with heparinase (3 units/ml, 37 "C, 2 h). A-F, original mag- nification, X 68. In this series of studies, the cells were fixed directly with paraformaldehyde and observed. In other studies, after incubation of the cells with fluorescently labeled lipoproteins, the cells were treated with 0.005% trypsin to remove cell surface or matrix bound lipoproteins. There was no difference in the results of trypsin-treated or -untreated cells.

2.5 r T al (TIE "

zz 2.0 - 2 5 No heparinase

2, 1.0- $3 $ E 6 s

vk 1.5 - 0 Heparinase (3 units) "

E 0.5 -

0 No

Lipoproteins P-VLDL

+ ApoE

FIG. 4. Effect of heparinase treatment on cholesterol esterification in FH fibroblasts. The cells were incubated in the absence or presence of heparinase (3 units/ml, 37 "C, 2 h). They were then incubated with no added lipoproteins or with p-VLDL+apoE (IO pg of P-VLDL protein + 15 pg of human apoE/ml). Data are presented as means & S.D. obtained in four separate studies.

2 or 4 h to label the cell surfaces and treated with heparinase, and then the percent decrease in cell surface ["S]sulfate was determined. As shown in Table 11, approximately 30-40% of the total incorporated radioactivity was released from the surface of these cells by heparinase treatment. Studies conducted in parallel at the same time showed that @- VLDL+apoE binding decreased, as shown previously for these cells (Fig. 1).

A crude preparation of rat liver matrix proteoglycans was

10164 Remnant Lipoprotein Bin

r - FH Fibroblasts HepG2 FIG. 5. Effect of heparinase treatment on the binding of a2-

macroglobulin to HepGP cells and FH cells. The cultured cells were incubated in the absence or presence of heparinase ( 3 units/ml, 37 "C, 2 h) and washed with media four times. The cells were incubated with '2sII-~2-macroglobulin for 3 h a t 4 "C.

TABLE I Effect of heparinase treatment on the binding of oc2-macroglobulin

'2sII-n2-macroglobulin binding

No treatment Heuarinase treatment

nglmg cell protein; mean -C S.D.

n = 3 22.0 f 4.3" 20.3 f 4.5" n = 3 33.0 f. 8.5' 34.0 f 10.1'

n = 3 6.0 f. 1.3" 6.1 f. 1.1" n = 4 9.0 k 2.5' 8.6 f 1.8'

FH fibroblasts

HepG2 cells

a 1 pg of cu~-macroglobulin/ml. ' 4 pg of 0,-macroglobulin/ml.

LRP.

LDL, receptor

1 2 3 4 5 FIG. 6. Effect of heparinase treatment on &VLDL+apoE

binding to the LRP/a2-macroglobulin receptor on ligand blots. The 8-VLDL+apoE (10 pg of p-VLDL protein + 10 pg of rabbit apoE/ml) was incubated with the rat liver membrane fraction blotted on nitrocellulose. The binding was detected using "'1-streptavidin. Lane I represents the control (no heparinase treatment). Lanes 2-5 were pretreated with 1, 5, 15, and 30 units of heparinase/ml for 2 h at 37 "C, respectively.

TABLE I1 Effect of heparinase treatment on the release of

"S-labeled proteoglycans % of total 35S-labeled cell surface

proteoglycans released

FH fibroblasts

HepG2 cells n = 3

n = 3 n = l

%

40.5 f 12"

29.7 f. 7.6" 42'

' Cells were incubated with [35S]sulfate for 2 h before treatment

'Cells were incubated with [D5S]sulfate for 4 h before treatment with heparinase for 2 h at 37 "C ( 3 units/ml).

with heparinase for 2 h a t 37 "C ( 3 units/ml).

)ding to Heparan Sulfate

isolated (24) and transferred to nitrocellulose membranes. Before isolation of the hepatic proteoglycans, the rats were injected with ["S]sulfate to prelabel the proteoglycans. The blots were incubated with or without heparinase and probed with biotin-labeled P-VLDL+apoE. As previously described (24), hepatic proteoglycans display a broad spectrum of sizes ranging from M, z 150-300 kDa. As shown in Fig. 7, lune 1, the ["S]sulfate distribution on the blot represented a broad band. Following heparinase treatment (50 units/ml, 37 "C, 2 h), there was a marked reduction in the "S-sulfate-labeled material, demonstrating a reduction in the hepatic proteoglycans (Fig. 7, lune 2) . Incubation of the rat liver matrix proteoglycans on the nitrocellulose membranes with biotin- labeled p-VLDL+apoE revealed marked binding of these lipoproteins to the blots (Fig. 7, lune 3 ) . However, when the nitrocellulose membranes were treated with heparinase (50 units/ml, 37 "C, 2 h), there was a significant reduction in the binding of P-VLDL+apoE (Fig. 7, lune 4 ) . Thus, the hydrolysis of proteoglycans by heparinase results in a parallel inhibition of P-VLDL+apoE binding.

P- VLDL+ApoE Binding to Proteoglycun-deficient Cells- To confirm that cell surface proteoglycans were involved in the binding of P-VLDL+apoE, we compared the binding activity of wild-type CHO cells with that of mutant CHO cells that were defective in cell surface proteoglycans. The CHO pgsD-677 cells are deficient in heparan sulfate, whereas CHO pgsA-745 cells are deficient in all cell surface proteoglycans (20).

Wild-type CHO cells were used as controls and were shown to bind P-VLDL and to display enhanced binding of P- VLDL+apoE (Table 111). Furthermore, the enhanced P- VLDL+apoE binding to the wild-type CHO cells was inhibited by heparinase treatment (3 units/ml, 37 "C, 2 h). The results were essentially identical with those for fibroblasts (Table 111) and HepG2 cells (previously discussed). Similar

R

500 kDa -

200 kDa +

1 2 3 4 FIG. 7. Nitrocellulose blots of rat liver proteoglycans dem-

onstrating the effect of heparinase treatment on the removal of [""Slsulfate-labeled heparan sulfate and the effect of heparinase treatment on the inhibition of binding of B- VLDL+apoE to the proteoglycans. Lane I is an autoradiograph of the [:'sS]sulfate-labeled heparan sulfate proteoglycans on the blot. Lane 2 demonstrates that heparinase treatment (50 units/ml, 37 "C, 2 h) removes a large component of the radiolabeled proteoglycans. Lane 3 shows the binding of biotin-labeled 8-VLDL+apoE (10 pg of protein + 10 pg/ml) to the proteoglycans on the blot. Lane 4 demonstrates that heparinase treatment (50 units/ml, 37 "C, 2 h) inhibits 8-VLDL+apoE binding to the proteoglycans. Biotin-labeled 8-VLDL was visualized by using a Vectastain ABC kit (Vector Laboratories, Inc., Burlingame, CA) for alkaline phosphatase, and following the manufacturer's directions, for Western blotting.


TABLE 111 Effect of heparinase treatment on binding of p-VLDL and p-VLDL + apoE to wild-type and mutant CHO cells and human fibroblasts

,8-VLDL" 0-VLDL + apoE"

Cells 0.8 pg/ml 2.0 pg/ml 0.8 pg/ml 2.0 pg/ml

Control Heparinase Control Heparinase Control Heparinase Control Heparinase treatment treatment treatment treatment

Wild-type CHO Experiment 1 49.3 57.1 58.4 64.7 209.2 47.4 426.0 57.3 Experiment 2 42.0 39.5 40.0 37.0 131.6 38.2 424.0 53.4

Human fibroblasts ( n = 4) 47.1 f 12.2 52.9 f 15.5 358 f 103.9 62 f 25 Mutant msD-677 CHO 47.1 f 12.2 49.1 f 16.3 75.8 f 23.4 80.4 f 28.6 44.0 f 10.5 46.0 f 10.8 78.8 f 15.3 75.0 f 17.7 .-

Mutant pgsA-745 CHO 55.7 f 11.6 55.3 f 13.3 84.2 f 22.3 80.0f 28.6 41.0 -t 9.7 42.0 f 9.1 62.0 * 16.4 68.0 f 17.1 ( n = 9)

( n = 9) Amount of lipoprotein binding at 4 "C for 3 h (ng/mg of cell protein).

TABLE IV Effect of heparinase treatment on binding of az-macroglobulin and

LDL to wild-type and mutant CHO cells and normal fibroblasts '261-o12-macroglobulin pglml)a

Control Heparinase treatment

Control Heparinase treatment

Human fibroblasts 27.5 -t 15.9 29.8 f 18 123 f 50 120 k 48 ( n = 8 ) ( n = 8 ) ( n = 4 ) ( n = 4 )

Cells (1 pg/ml)"

Wild-type CHO 5.0 f 1.1 4.7 f 1.0 48 f 7.9 49 t 10.5 ( n = 5 ) ( n = 5 ) ( n = 4 ) ( n = 4 )

Mutant pgsA-745 5.9 f 2.1 6.5 -t 2.8 41 f 17.3 46 f 17.4 CHO ( n = 8 ) ( n = 8 ) ( n = 5 ) ( n = 5 )

MutantpgsD-677 5.0 f 1.9 5.6 f 2.4 32 t 10 38 f 7.5 CHO ( n = 8 ) ( n = 8 ) ( n = 4 ) ( n = 4 ) ' Amount of radiolabeled az-macroglobulin or LDL binding at 4 "c

for 3 h (ng/mg of cell protein).

results were obtained for @-VLDL (without added apoE) binding to the CHO pgsD-677 and CHO pgsA-745 cells, and there was no effect of heparinase treatment on the low level @-VLDL binding (Table 111). However, the CHO pgsD-677 and CHO pgsA-745 cells did not display enhanced binding of @-VLDL+apoE (Table 111).

Because the LRP/a,-macroglobulin receptor has been shown to bind and internalize @-VLDL+apoE (8, 10, 12), it is possible that the lack of enhanced binding of the @- VLDL+apoE by the CHO pgsD-677 and pgsA-745 cells was caused by inactive or absent receptors. Wild-type CHO cells have been shown to express LRP/az-macroglobulin receptor, although at lower levels than observed for fibroblasts (27). This observation has been confirmed in wild-type CHO cells and established for the CHO pgsD-677 and CHO pgsA-745 cells. As shown in Table IV, the wild-type and mutant CHO cells bound similar amounts of lZ5I-labeled a*-macroglobulin. This binding was not affected by heparinase treatment. Fur- thermore, ligand blots of solubilized CHO wild-type, pgsD- 677, andpgsA-745 cells revealed a band in the correct position for the LRP by 45Ca detection (data not shown). Likewise, the wild-type and mutant CHO cells possessed the LDL receptor and binding of lZ5I-LDL was not affected by heparinase treatment (Table IV).

Effect of Heparitinase on Binding to Cultured Cells-In other studies, heparitinase, which catalyzes the cleavage of the a-N-acetyl-D-glucosamine linkage in heparan sulfate, was used. The binding of @-VLDL (without added apoE) was similar in all cell types and was unaffected by heparitinase treatment (Table V). On the other hand, @-VLDL+apoE resulted in enhanced binding that was markedly inhibited by

TABLE V Effect of heparitinase treatment on binding of 6- V L D L and p - V L D L

+ apoE to HepG2 cells, normal fibroblasts, and FH fibroblasts @-VLDL + apoE

CdlQ B-VLDL (O" (0.8 pg/ml + 1.2 pg/ml)" "__I

Control Heparitinase Control Heparitinase

HepG2 22.2 f 2.1 20.0 f 1.4 70.5 f 7.1 19.9 f 1.0

Fibroblasts 51.5 f 6.1 49.4 f 5.6 214.3 f 32.6 60.5 f 7.5

FH fibroblasts 303.2 f 19.6 22.3 f 4.4

( n = 4)

( n = 4)

( n = 4) a Amount of lipoprotein bound to the cells a t 4 "C for 3 h (ng/mg

of cell protein).

heparitinase treatment (Table V). These results are essentially identical with those obtained with heparinase.

DISCUSSION

The present study indicates that heparinase- and heparitinase-sensitive sites on the surface of cultured cells represent major sites for the binding of apoE-enriched remnant lipoproteins. These enzymes have been shown to cleave heparan sulfate (25,26). Heparinase treatment of HepG2 cells, normal human fibroblasts, and FH fibroblasts resulted in, respectively, about 80, 80, and 95% inhibition of binding of the @- VLDL+apoE. This inhibition of binding represents approximately all of the enhanced binding of the apoE-enriched @- VLDL. Furthermore, heparinase treatment decreased the uptake of labeled @-VLDL+apoE by HepG2 cells and decreased cholesteryl ester formation in FH cells. On the other hand, heparinase treatment had no effect on the binding and uptake of @-VLDL (without added apoE) and LDL by these cells.

The indication that heparan sulfate proteoglycans plays a major role in the binding and uptake of remnant lipoproteins was greatly strengthened by the observations in the CHO mutant cells that displayed defects in the biosynthetic pathways of glycosaminoglycans or heparan sulfate proteoglycans. These cells failed to demonstrate enhanced binding of p- VLDL+apoE. However, these mutant cells did possess binding activity for az-macroglobulin and did possess the LRP/ az-macroglobulin receptor.

Previously, it has been shown that (I-VLDL enriched in apoE were bound and taken up by FH fibroblasts through the LRP receptor (for review, see Ref. 28). Treatment of these cells with anti-LRP-515, which results in a decreased expres- sion of these receptors on the cell surface, abolished the enhanced uptake of the @-VLDL+apoE as determined by the stimulation of cholesteryl ester formation (10). Herz et al. (29) showed that the 39-kDa protein associated with the LRP

10166 Remnant Lipoprotein Binding to Heparan Sulfate

can inhibit the uptake of P-VLDL+apoE by FH fibroblasts and can prevent the binding of P-VLDL+apoE to the LRP on ligand blots. In addition, we demonstrated that a2-macroglobulin can compete for apoE-enriched remnant lipoprotein uptake by the liver in uiuo, suggesting that the LRP was a physiologically important receptor involved in remnant clearance (12). Recently, lactoferrin, a 76-kDa protein, has been shown to bind to the LRP (30) and to block the uptake of remnant lipoprotein in vitro and in uivo (30-33). Thus, several lines of evidence indicate that the LRP/a2-macroglobulin receptor is involved in the uptake of remnant lipoproteins (3, 28, 30).

The present study suggests that cell surface heparan sulfate proteoglycans and the LRP may be operating as a complex and that there may be an involvement of both heparan sulfate proteoglycans and the LRP in the enhanced binding and uptake of apoE-enriched remnant lipoproteins. Treatment of various cells with heparinase or heparitinase, both of which cleave heparan sulfate, prevents both the enhanced binding and uptake of the apoE-enriched remnants; likewise, the CHO mutants do not display the enhanced binding or internalization of the apoE-enriched remnants. Our postulate is that heparan sulfate proteoglycans serve as the initial binding site on the cell surface for the remnants with apoE and then present the remnants to the LRP for internalization by the cells. Previous studies, using the anti-LRP (515) that prevents enhanced cholesterol esterification (10) (i.e. prevents enhanced uptake) are not inconsistent with the results of the present study. Furthermore, we would suggest that the ability of the 39-kDa protein to prevent binding to the LRP on a ligand blot is also not inconsistent since the LRP is in a denatured form on the membrane. In addition, the ability of lactoferrin to inhibit the enhanced uptake of remnant lipoproteins by cultured cells and in vivo by the liver (30-33) is difficult to interpret because, as we have shown, lactoferrin binds both to heparan sulfate proteoglycans and the LRP.'

Support for this model of a linkage between heparan sulfate proteoglycan binding and receptor-mediated uptake comes from studies of basic fibroblast growth factor cell biology. The binding of basic fibroblast growth factor to cells requires that it interact with heparan sulfate proteoglycans as an essential first step before interaction with the basic fibroblast growth factor receptor (34-36). In CHO cells lacking heparan sulfate, the basic fibroblast growth factor has no biological activity even in the presence of its receptor (34).

Previously, we postulated that the initial step in the clearance of chylomicron remnants from the plasma was by sequestration in the space of Disse (3). The space of Disse has been shown to contain an abundance of proteoglycans, primarily heparan sulfate (13). Furthermore, the existence of high levels of apoE in the space of Disse (37) suggests a possible physiological role for apoE enrichment of remnant lipoproteins in their uptake by the liver. I t is entirely possible that apoE secreted by the liver causes an increase in the local concentration of apoE. Earlier, it was suggested that high local levels of apoE might serve to capture lipids and lipoproteins in the vicinity of cells expressing the LDL receptor or LRP (2,3,28). This may also occur in the space of Disse. The presence of an abundance of heparan sulfate proteoglycans and apoE in the space of Disse may represent the pathway whereby lipoproteins may become sequestered and prepared for internalization and degradation in the hepatocytes.

Numerous studies have demonstrated that lipoprotein lipase can mediate the binding of a variety of lipoproteins to the surface of cultured cells (38-42). Beisiegel et al. (14) demon-

* Z.-S. Ji and R. W. Mahley, unpublished results.

strated that lipoprotein lipase enhances the interaction of apoE-containing lipoproteins to HepG2 cells and FH fibroblasts and that the interaction is mediated by LRP. Several studies have now suggested that lipoprotein lipase can en- hance the cellular interaction of both apoB- and apoE-containing lipoproteins (VLDL, LDL, and lipoprotein(a)) through the binding of lipoprotein lipase to cell surface heparan sulfate proteoglycans (40-42). This has been demonstrated using heparinase (40, 42) or heparitinase (41) and by studies of mutant CHO cells deficient in or lacking cell surface heparan sulfate proteoglycans (41, 42). Whereas Williams et al. (41) and Mulder et al. (40) showed that the lipoprotein lipase-proteoglycan pathway is independent of the LDL receptor, it is less clear whether the interaction is independent of the LRP. Furthermore, the physiological role of the lipoprotein lipase-proteoglycan pathway in chylomicron metabolism is unclear. The present study, using apoE-enriched lipoproteins, demonstrates the importance of heparan sulfate proteoglycans in mediating cell surface binding as one pathway for enhancing lipoprotein uptake. The question remains as to whether the apoE-proteoglycan pathway can result in the uptake of remnants independent of the involvement of receptors or whether the proteoglycans bind the remnants to the cell surface, facilitating the transfer of the lipoproteins to the receptors for actual receptor-mediated endocytosis.

Acknouledgments-We thank Dr. Jeffrey D. Esko for providing mutantpgsA-745 andpgsD-677 CHO cells, Dr. Frederick R. Maxfield for preparation of a2-macroglobulin, Dr. Jean Davignon for providing human LDL receptor-negative fibroblasts, Dr. Karl H. Weisgraber for providing human and rabbit apoE, Sylvia Richmond for manu- script preparation, and Liliana Jach for graphics.

REFERENCES

1.

2. 3. 4.

5.

6. 7.

8.

9.

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

Brunzell, J. D. (1989) in The Metabolic Basis of Inherited Disease (Scriver, C. R., Beaudet, A. L., Sly, W. S. and Valle, D., eds) 6th Ed., pp. 1165- 1180 MrC.mw-Hill. New York - - - -, -. . - - - - . . . ""

Mahley, R. W. (1988) Science 240,622-630 Mahley, R. W., and Hussain, M. M. (1991) Curr. Opin. hipidol. 2,170-176 Mahlev. R. W.. and Rall. S. C.. Jr. (1989) in The Metabolrc Basrs of Inhercted

I - - - - ~

~~~~

Dis&e (Schver, C. R . , Beaudet, A. L., Sly, W. S. and Valle, D., eds) 6th

Choi. S. Y.. Fone. L. G.. Kirven. M. J.. and Cooper, A. D. (1991) J. clin. Ed., pp. 1195-1213, McGraw-Hill, New York

Inbest. 8 8 , 1173-1181 Brown, M. S., and Goldstein, J. L. (1983) J. Clin. Inuest. 7 2 , 743-747 Strickland, D. K., Ashcom, J. D., Williams, S., Burgess, W. H., Migliorini,

Kowal. R. C.. Herz. J.. Goldstein. J. L.. Esser. V., and Brown, M. S. (1989) M., and Argraves, W. S. (1990) J. Biol. Chem. 266 , 17401-17404

.. ~ ~

Proc.Natl.'Acad.'SCi. U. S. A. 8 6 , 5810-5814

Nature 341,162-164 Beisiegel, U., Weber, W., Ihrke, G., Herz, J., and Stanley, K. K. (1989)

Herz. J., Kowal, R. C., Ho, Y. K., Brown, M. S., and Goldstein, J. L. (1990)

Kowal, R. C., Herz, J., Weisgraber, K. H., Mahley, R. W., Brown, M. S., J. Biol. Chem. 2 6 6 , 21355-21362

Hussain, M. M., Maxfield, F. R., MCs-Oliva, J., Tabas, I., Ji, Z.-S., Inner- and Goldstein, J. L. (1990) J. Biol. Chem. 266,10771-10779

Stow, J. L., KjBllen, L., Unger, E., Hiiiik, M., and Farquhar, M. G. (1985) arity, T. L., and Mahley, R. W. (1991) J. Biol. Chem. 2 6 6 , 13936-13940

Beisiegel, U., Weber, W., and Bengtsson-Olivecrona, G. (1991) Proc. Natl. J. Cell Biol. 100 , 975-980

Bilheimer, D. W., Eisenberg, S., and Levy, R. I. (1972) Biochim. Biophys. Acad. Sci. U. S. A. 88,8342-8346

Hussain, M. M., Mahley, R. W., Boyles, J. K., Fainam, M., Brecht, w. J.,

Goldstein, J. L., Basu, S. K., and Brown, M. S. (1983) Methods Enzymol.

Acta 260,212-221

and Lindquist, P. A. (1989) J. Biol. Chem. 264,9571-9582

98,241-260 Rall, S. C., Jr., Weisgraber, K. H., and Mahley, R. W. (1986) Methods

Wernette-Hammond, M. E., Garcia, Z., Arnold, K. S., and Innerarity, T. Enzymol. 128,273-287

L. (1989) Arteriosclerosis 9 , 501-510 20. Esko, J. D. (1991) Curr. Opin. Lipidol. 3, 805-816 21. Lowry, 0. H., Rosebrough, N. J., Farr, A. L., and Randall, R. J. (1951) J.

22. Khan, Y., and Newman, S. A. (1991) A d . Biochem. 196,373-376 23. Schneider, W. J., Beisiegel, U., Goldstein, J. L., and Brown, M. S. (1982)

24. Soroka, C. J., and Farquhar, M. G. (1991) J. Cell Biol. 113 , 1231-1241 25. Yang, V. C., Linhardt, R. J., Bernstein, H., Cooney, C. L., and Langer, R.

26. Nader H. B. Dietrich, C. P., Buonassisi, V., and Colburn, P. (1987) Proc.

27. Yamashiro, D. J., Borden, L. A,, and Maxfield, F. R. (1989) J. Cell. Physiol.

Bwl: Chem. 193,2651215

J. Biol. Chem. 267,2664-2673

(1985) J. Biol. Chem. 260,1849-1857

Nati Acadl Sci. U. S. A. 84,3565-3569

139,377-382

28.

29.

30.

31.

32.

33.

34.

35.

Remnant Lipoprotein Binding to Heparan Sulfate Brown. M. S.. Herz. J.. Kowal. R. C.. and Goldstein. J. L. (1991) Curr.

Opin. Lipidol. 2 , 65-72

(1991) J. Biol. C k m . 266,21232-21238

J. Biol. Chem. 268,26172-26180

, .

Herz, J., Goldstein, J. L., Strickland, D. K., Ho, Y. K., and Brown, M. S.

Willnow, T. E., Goldstein, J. L., Orth, K., Brown, M. S., and Herz, J. (1992)

Huettinger, M., Retzek, H., Eder, M., and Goldenberg, H. (1988) Clin. Bwchpm. 2 1.87-92 ~ ~~ . ~ ~~

van Dijk, M. C. M., Ziere, G. J., Boers, W., Linthorst, C., Bijsterbosch, M.

Huettinger, M., Retzek, H., Hermann, M., and Goldenberg, H. (1992) J.

Yayon, A., Klagsbrun, M., Esko, J. D., Leder, P., and Ornitz, D. M. (1991)

Klagsbrun, M., and Baird, A. (1991) Cell 67,229-231

,_ "

K., and van Berkel, T. J. C. (1991) Biochern. J . 279,863-870

Biol. C k m . 267,18551-18557

Cell 64,841-848

36.

37.

38.

39.

40.

41.

42.

10167 Rapraeger, A. C., Krufka, A., and Olwin, B. B. (1991) Science 252 , 1705-

1 7 n ~ Hamilton, R. L., Wong, J. S., Guo, L. S. S., Krisans, S., and Havel, R. J.

Chajek-Shaul, T., Friedman, G., Stein, O., Olivecrona, T., and Stein, Y. (1990) J. Lipid Res. 3 1 , 1589-1603

Oka, K., Wang-Iverson, P., Paterniti, J. R., Jr., and Brown, W. V. (1989) (1982) Biochm. Btophys. Acta 712,200-210

Mulder, M., Lombardi, P., Jansen, H., van Berkel, T. J. C., Frants, R. R., Ann. N . Y. Acad. Sci. 6 5 6 , 173-180

and Havekes, L. M. (1992) Biochem. Biophys. Res. Cornmun. 186 , 582- 587

. . "-

Williams, K. J., Fless, G. M., Petrie, K. A,, Snyder, M. L., Brocia, R. W.,

Eisenberg, S., Sehayek, E., Olivecrona, T., and Vlodavsky, I. (1992) J. Clin. and Swenson, T. L. (1992) J. Biol. Chem. 267 , 13284-13292

Inuest. 90,2013-2021

Role of Heparan Sulfate Proteoglycans in the Binding and Uptake of ...

Documents

Transcript of Role of Heparan Sulfate Proteoglycans in the Binding and Uptake of ...