Cyanogen bromide, activated Sepharose, Sephacryl S-400 HR, and Sephacryl S-500 HR were .... slots in agarose submarine gels. The gels, approximately 3 ...
Vol. 268, No. 19, Issue of July 5, pp. 14131-14137,1993 Printed in U.S.A.
THE JOURNAL OF BIOLOGICAL CHEMISTRY 0 1993 by The American Society for Biochemistry and Molecular Biology, Inc.
Binding of Low Density Lipoproteins byProteoglycans Synthesized by Proliferating and QuiescentHuman Arterial Smooth Muscle Cells* (Received for publication, September 21, 1992, and in revised form, January 25, 1993)
Germhn CamejoSO, GunnarFagerSQ,Birgitta Rosengrent, Eva Hurt-Camejot,and Goran Bondjerst From the $Biochemistry Department, Preclinical Research Laboratories, Astra Hiissle, Molndal, S-431 83, Sweden and (i Wallenberg Laboratow. DeDartment of Heart andLung Disease, Uniuersity of Gothenburg, Sahlgren’s Hospital, Gothenburg, Sweden “
I
~
Chondroitin sulfate-rich proteoglycans secreted by arterial intima smooth muscle cells appear involved in low density lipoprotein entrapment and modification. Hypothetically, such a process may contribute to atherogenesis. We compared composition and size of those proteoglycans synthesized by proliferating and resting human arterial smooth muscle cells for which low density lipoprotein had affinity. Lipoprotein-binding proteoglycans secreted by proliferating cells were larger than those of resting cells ( M . = 1.1 X lo6 uersus 0.8 X 10’). This was primarily caused by increased M, of the chondroitin sulfate chains (6 X lo4 uersus 3.5 X lo4).The glycosaminoglycan chains of the proteoglycans from both cells were madeofmore than 90% chondroitin 6-sulfate and chondroitin 4-sulfate in a 6:4 ratio. Affinity chromatography indicated that low density lipoprotein had a higher affinity withthe proteoglycans synthesized by proliferating cells than those from resting cells. Measured with gel mobility shift assay, the apparent affinity constant of low density lipoproteins for proteoglycans from proliferating cells was %fold higher than that for proteoglycans from resting cells. This increased affinity appeared related to the higher relative proportion of proteoglycans with longer glycosaminoglycan chains secreted by the proliferating cellsthan those secreted by the resting cells.
Proteoglycans of the arterialintima appear to be responsible for the formation of complexes with lipoproteins that contain apoB-100. This may bea step inthe normal exchange of components between the circulating plasma and thearterial wall. However, during atherogenesis, this process seems to contribute to continuous focal deposition of cholesterol-rich lipoproteins, mainly low density lipoprotein (LDL)’ in lesions (for reviews, see Refs.1-3). Proteoglycans of the arterial intima are mainly synthesized by smooth muscle cells and endothelial cells. During development of atherosclerotic lesions, arterial smooth muscle cells (ASMC) which are normally quiescent start to proliferate in the intima. Here these cells synthesize most of the extracellular matrix components *This research was partiallysupported by Grant 4531 of the Swedish Medical Research Council and grants from the Swedish Heart and Lung Foundation. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. The abbreviations used are: LDL, low density lipoprotein; ASMC, arterial smooth muscle cells; PG(s), proteoglycan(s); CSPG, chondroitin sulfate proteoglycan; GAG, glycosaminoglycan; C6S, chondroitin 6-sulfate; C4S, chondroitin 4-sulfate; CS, chondroitin sulfates; DS, dermatan sulfates; PDGF, platelet-derived growth factor.
’
that include: collagen(s), elastin, proteoglycan(s), and other glycoproteins (1). The increase in extracellular matrix in lesions is accompanied by deposition of apoB-100 lipoproteins rich in cholesterol esters (1-3). Formation of complexes of LDL with the extracellular matrix canbe a major contributor to lipoprotein accumulation in atheromas. The rest of the apoB-100 lipoprotein in lesions appears associated with macrophages and proliferating ASMC (4,5). Once immobilizedin the extracellular matrix, apoB-100 lipoproteins seem to be the target of oxidative and hydrolytic modifications that increase their rate of uptake by macrophages. These phenomena may be contributors to formation of foam cells that arecharacteristic of atherosclerotic lesions (6-8). In animal models and humans, at sites of lesion development and lipoprotein deposition, there is also focal increase of chondroitin sulfate proteoglycans (CSPG) and dermatan sulfate proteoglycans (911).A possible cause-effect relationship between these processes has been suggested (2, 12-14). The shift of ASMC from a quiescent to a proliferating condition in the intima appearsassociated with a change from acontractile phenotype toasynthetic one (15, 16). Such modulation can be simulated in vitro by manipulations on nutrients or the addition of specific growth factors to cell cultures (16). Proliferating ASMC from arteries of bovines, primates, and humanssynthesize more sulfated proteoglycans than quiescent cells (17-19). In addition, structural changes of CSPG andheparan proteoglycans take place in aortic ASMC in culture when they are induced to proliferate (1720). Recently, it was shown in cultured ASMC from Macaca nemestrina that platelet-derived growth factor (PDGF), besides increasing the rate of synthesis of the large Versicanlike CSPG, also increases their hydrodynamic size (21). These findings may have implications in atherogenesis because first, PDGF could be one of the main growth factors involved in the formation of the intimal population of SMC in lesions (1, 22), and, second, because affinity of CSPG with LDL is positively correlated with its hydrodynamic size (23). If similar phenomena are present in proliferating lesions, they may contribute to the focal co-accumulation of CSPG and apoBcontaining lipoproteins. This hypothesis led us to explore the LDL-binding properties of proteoglycans synthesized by quiescent and proliferating human ASMC. The results indicated that proliferating cells synthesized relatively more of a large LDL-binding CSPG than resting cells. In addition, LDL bound to the CSPG of proliferating cells with a %fold higher affinity constant than to the CSPG synthesized by resting cells. The increased affinity of LDL with the CSPG from proliferating cells appears related to thelarger hydrodynamic size of the proteoglycan. This appears to be caused by longer chondroitin sulfate chains in the CSPG of proliferating cells than those in the CSPG secreted by resting cells.
14131
14132
Lipoprotein Bindingby Smooth Muscle Cell Proteoglycans EXPERIMENTAL PROCEDURES
Materials Guanidine HCl, grade I, Hepes, Triton X-100, N-ethylmaleimide, r-aminocaproic acid, benzamidine HCl, phenylmethylsulfonyl fluoride, papain, cetylpyridinium bromide, and ethylaminohexanoic acid were purchased from Sigma. Dextran sulfate with average molecular weights of 8,000, 50,000, and 500,000 and heparin grade I was also purchased from Sigma. Chondroitin 4-sulfate (C4S), chondroitin 6sulfate (C6S), and dermatan sulfate (all super special grade), chondroitinase ABC (EC 4.2.2.4.) protease-free, chondroitinase AC I (EC 4.2.2.5), heparinase (EC 4.2.2.7), heparitinase (EC 4.2.2.8), and the chondro-disaccharide standards kit for high performance liquid chromatography were purchased from Seikagaku Kogyo Co. (Tokyo, Japan). Cyanogen bromide, activated Sepharose, Sephacryl S-400 HR, and Sephacryl S-500 HR were bought from Pharmacia Fine Chemicals (Uppsala, Sweden). Cell culture media, trypsin, fetal bovine calf serum, and culture vessels were from Flow Laboratories (Irvine, Scotland). Na[%3]S04(25-40 Ci/mg) and ~-[4,5-~H]leucine(120190 mCi/mmol) were from Amersham Sweden AB. NuSieve GTG agarose was from FMC Bioproducts (Vallensbaek, Denmark). Human low density lipoprotein (LDL), density 1.019-1.063 g/ml, was prepared by differential centrifugation as described (24).
and proliferating cells. The LDL-binding proteoglycans were dialyzed extensively against water and lyophilized. Affinity of LDL for Proteoglycans Evaluated by Gel Mobility Shift Assay
Affinity of LDL for proteoglycans and glycosaminoglycans, in terms of an apparent affinity constant (K,) expressed as molarity of LDL, was measured bygel mobility shift assay. This procedure, presently used mostly for evaluation of interactions between proteins and DNA, was applied to study the association between LDL and the proteoglycans and glycosaminoglycans synthesized by the human ASMC (27). The following modification was developed. Fixed amounts (0.4-0.5 pg) of the LDL-binding, 35S-labeledproteoglycans, obtained from the affinity chromatography runs, measured as chondroitin sulfate (26) were used. The proteoglycans were mixed with increasing concentrations of LDL in polypropylene conical microcentrifuge tubes. Usually between 8,000 and 10,000 cpm of the PGs or GAGs in 5- to 20-4 aliquots were mixed with 2- to 20-pl aliquots of LDL. In addition, 10 pl of glycerol was added and the volumes were completed to 80 pl with buffer 10 mM Hepes, 140 mM NaCl, 2 mM MgC12,2 mM CaC12,pH 7.2. The PGs, GAGs, and theLDL were preequilibrated in the same buffer. The final concentrations of LDL were in the range 0.05 to 0.50 pM. The tubes were incubated at 20 "C for 1 h and then 10 p1 from each tube was applied in duplicate to slots in agarose submarine gels. The gels, approximately 3 mm in Cell Culture thickness, were made of NuSieve agarose dissolved in a buffer conUntransformed, mycoplasma-free, human ASMC were studied in taining 5 mM Hepes, 2 mM CaC12,2 mM MgC12,pH 7.4. The electrothe 5th to 8thpassages. The cells were isolated from the inner media phoresis step was run for 2 h at 60 V, constant voltage, with buffer of human uterine arteriesas previously described (25). The cells were recirculation. The gels were fixed in 0.1% cetylpyridinium bromide subcultured in 75-cmZdishes. Split ratios of 2:3 and 1:3 of the cells overnight, dried, and quantitatively evaluated in a Berthold digital from the same original dish were used to establish proliferating and autoradiograph (Berthold Laboratories, Wilband, Germany). In this quiescent cultures, respectively. The cells were maintained in Way- system, free PGs and GAGs moved 3 to 4 cm from the origin and the mouth's medium plus 10% v/v human serum and 10% v/v fetal calf free LDL approximately 0.5 cm. Binding of the sulfated polysacchaserum for 3 days. After this period, to obtain proliferation, the medium rides to LDL was indicated by a reduction in their electrophoretic was changed to basal minimal Eagle's-diploid plus 10%human serum mobility. At saturating concentrations of LDL, most of the PGs and and 10% fetal calf serum. To obtain quiescent (resting) cells, the GAGS formed complexes with LDL that did not enter thegel or were medium was changed to basal minimal Eagle's-diploid with only 1% associated with the LDL band. By measuring the radioactivity of the bovine serum albumin. After 3 days in these media, the cells were faster moving PGs or GAGS band, their free fraction was directly placed in fresh media with the same composition plus 25 pCi/ml [35S] measured and the bound fraction calculated from the total measured sulfate (40-50 pCi/mol) and 10 pCi/ml [3H]leucine (Ci/mmol). The in samples containing no LDL. This allowed measurement of the cells remained in these media for 48 h. During culture, the cells were apparent affinity constant (KO)according to the equation: K. = [F/ maintained at 37 "C in an atmosphere of 5% COZ in air. After this (Lo(1- F ) ) ] where F is the fraction of radioactive PG or GAG bound period, the media (aproximately 10 ml) were collected, and the cells and Lo is the total molar concentration of LDL, in large excess over werewashed once with phosphate-buffered saline. The cells were that of PGs orGAGS (28). The system is designed for the interactions dissolved in 5 ml of 25 mM NHdOH containing 0.5% Triton X-100. to takeplace at or nearto physiological ionic strength conditions and Both to the media and the cell fraction was immediately added a the separation step totake place at conditions that maximize them. mixture of protease inhibitors with the following final concentrations: Partial Characterization of the LDL-binding Proteoglycans 5 mM c-aminocaproic acid, 0.1 mM phenylmethylsulfonyl fluoride, 5 mM N-ethylmaleimide, and 5 mM benzamidine HCl. The fractions Nature of the GAGS-The fractions containing the PGs bound by were dialyzed (4 X 1 liter) against a buffer containing 10 mM Hepes, the Sepharose-LDL columns were dissolved in 1.0 ml of 5 mM Hepes, 20 mM NaC1, pH 7.2, and the protease inhibitors mixture. Dialysis 5 mM CaCIZ,2 mM MgC12, 0.1% bovine serum albumin, pH 7.4, and against this buffer was required to run the samples through the 0.1 ml of this solution was subjected to hydrolysis by adding 0.1 unit/ subsequent affinity chromatography step. ml of chondroitinase ABC, chondroitinase AC, heparitinase, or heparinase. The different fractions were incubated overnight at 37 "C. Affinity Chromatography on Sepharose-LDL Columns After treatment with the enzymes, the fractions were run again Sepharose-LDL columns (5 X 1 cm diameter) were prepared from through the Sepharose-LDL. This step allowed quantitatively evaluLDL bound to cyanogen bromide-activated Sepharose according to ation of the effect of the different enzymes on the binding of the PGs the manufacturer's procedure. To protect the proteoglycan-binding to LDL. Additionally, it allowed identification of the GAG responsible arginine- and lysine-rich regions of apoB-100 from being blocked by for this property. Hydrodynamic Size of PGs and GAGS-Gel exclusion chromatogreacting with the activated gel, the coupling reaction was conducted raphy was performed in 100 X 1cm columns packed with Sephacrylusing 5 mg/ml LDL-protein per ml of gel and in the presence of 50 pg/ml heparin. The column was washed with 50 ml of 2 M NaCl to 400 and 500 HR (high resolution) at a flow of 12 cm2/h, using either dissociate the heparin from the covalent bound LDL and then equil- for dissociating conditions buffer 0.1 M Tris-HC1, 0.1 M Na2S04, 25 ibrated in buffer 10 mM Hepes, 20 mM NaCI. Direct measurements mM EDTA, 4 M guanidine HCl, pH 7.0, or for nondissociating of glycosaminoglycans (26) in the gel suspension indicated that the conditions 20 mM Hepes, 140 mM NaCl, 4 mMCaC12, 2 mM MgClz. heparin remaining was less than 1 pg/mg bound LDL. A similar These columns were calibrated with C4S,C6S, DS, and dextran column containing no LDL and blocked with ethanolamine served as sulfate of known average Mr. The elution patterns of these polysaccontrol. Aliquots (2-5 ml) of the media and cell fractions in buffer 10 charides were followedby the 1,9-dimethylene blue assay (26). GAGS mM Hepes, 20 mM NaCl and containing the 35S-and 3H-labeled were released from the proteoglycans by reductive @-elimination(29). proteoglycans were passed through the columns. The unbound frac- Other analytical procedures for characterization of proteoglycans, tions were collected in the first 10 ml and then a gradient from 20 glycosaminoglycans,and lipoproteins have been described (30). mM to 250 mM NaCl was used to elute the bound material into 1-ml RESULTS fractions. The fractions containing bound PGs were pooled, dialyzed extensively against water, and lyophilized. In one experiment, the Affinity Chromatography of Secreted Proteoglycam-After peaks containing the media PGs that were bound to theimmobilized binding to Sepharose-LDL columns, the total proteoglycans LDL were separated into two subfractions at the peak maximum. This was done with proteoglycans secreted into the media by resting secreted by or associated, respectively, with the resting and
Cell Proteoglycans
Lipoprotein Smooth Binding Muscle by proliferating cells were eluted with an NaCl linear gradient. Fig. 1shows that a higher NaCl molarity was required to elute the bulk of the 35S-labeledmacromolecules secreted by proliferating cells than therespective PGs of the resting cells. This was also the case for cell-associated PGs (not shown). The covalent binding of LDL to Sepharose improved markedly the capacity of the column for PGs. However, the positions at which the peaks were eluted were not different in columns prepared with and without heparin. This indicates that heparin and the cell proteoglycans recognized similar regions of LDL and confirmed that little heparin remained noncovalently associated with the Sepharose-LDL. The specific activity of the PGs secreted by the cell types was similar (8 to 10 X lo5 cpm/pg of GAG). Therefore, the results suggest that the differences in elution pattern were caused either by an increase in the amount of PGs with high affinity that were synthesized by the proliferating cells, or that they synthesized proportionally less of the PGs with lower affinity. Similar patterns as those in Fig. 1 were obtained with 5 out of 6 different batches of proliferating and resting cells originating from similar explants (not shown). For both cell phenotypes, 80-90% of the 35Sin the secreted or cell-associated macromoleculeswere bound to the columns. This was the case independently of the aliquot volumes of media or cell extract loaded in the column (2-10 ml). These results indicate that the Sepharose-LDL columns were notsaturated by the amounts used in the experiments. As mentioned under “Experimental Procedures,” the LDL-binding proteoglycans secreted in the media from one preparation of proliferating and resting cells were divided into two subfractions at the peak maxima (Fig. 1).This corresponds to pooled fractions 5 to 8 (PG-Rest I) and fractions 9 to 16 (PG-Rest 11) of the resting cells elution profile (Fig. 1).The profile of the proliferating cells was divided at fractions corresponding to the first half of the peak (fractions 9-15, PG-Pro1 I) and the second half (fractions 16-24, PG-Pro1 11). Such subfractions were used for evaluation of their affinity with LDL and of their hydrodynamic size (see below). Effect of Chondroitinases and Heparitinase on Binding to Sepharose-LDL-To characterize the GAGs responsible for the binding to Sepharose-LDL, those fractions containing PGs that were retained by the first passage through the columns were pooled, dialyzed, lyophilized, and treated with the different lyases. After incubation with the enzymes, the
fractions were bound and eluted again from the SepharoseLDL columns. It can be observed in Fig. 2 (panels A and E ) that chondroitinase ABC and AC converted more than 90% of the 35S-containing GAGs from proliferating cells to nonbinding fragments. On the other hand, heparitinase had little effect (panel C ) . In this case, more than 90% of the %labeled macromolecules were still retained by the column. Notice the difference in ordinates of profiles A and B with profile C. No effect of heparinase was observed in the profiles (not shown). Similar results were obtained with the LDLbinding fractions obtained from resting cells (Fig. 3). The experiment indicated that most of the PGssynthesized by the cells responsible for the interaction with the Sepharose-LDL column contained GAGs made mainly of C6S and C4S. These PGs appear to contain less than 10% DS or heparan. High performance liquid chromatography of disaccharides obtained after the enzyme treatments confirmed that theLDL-binding GAGs synthesized by both cells were made of C6S and C4S isomers in a 6:4 ratio. We could not detect the presence of oversulfated unsaturated disaccharides. Such findings indicate that the extentof sulfation inLDL-binding CSPGs from both cells was similar. Affinity of Secreted PGs and GAGS for LDL-The association between LDL and the CSPGs and GAGs system was analyzed quantitatively bygel mobility shift assay. This method allowed evaluation of binding in terms of the apparent affinity constant (Ka)at or near to physiological ionic conditions which can be expressed as LDL molarity. However, it should be indicated that our lack of knowledge of the stateof aggregation of the PGs and GAGs in the gel does not allow us to interpret this KOin molecular terms. The results obtained with PGs from two subcultures and two LDL samples are presented in TableI. LDL showed a 3-fold higher affinity with CSPG from proliferating cells than with the corresponding fraction from quiescent cells. LDL had a 5-8-fold higher affinity with the intact CSPG of either cell phenotype than with the corresponding GAGs obtained by reductive P-elimination. In addition,LDL showed a 3- to 4-fold higher affinity with GAG from proliferating cells than with those of the resting cells. The results indicate that the difference in binding affinity between the CSPG of the two cell phenotypes originates in the GAGs. These GAGs and their assembly into PGs provide a matrix with which LDL interacts more avidly than those of resting cells. We evaluated also the binding to
,
80
PG-Rest I, ,-PG-Rest
FIG. 1. Affinity chromatography of ‘“S-labeled macromolecules synthesized by proliferating and resting human arterial smooth muscle cells. Macromolecules synthesized and secreted into the medium by proliferating cells (sofidcircles) and resting cells (empty circles) wereloadedand eluted from a column to which Low density Lipoprotein was covalently attached to Sepharose. The boundmacromolecules were eluted by alinearNaClgradient (dotted tine). The macromolecules associated with the cell fraction were similarly fractionated.PC-Rest I, PG-Rest II, PG Pro1 I , and PG-Pro1 II indicate the pooled fractions used to characterize subclasses PGs.
II
1 o Rest. Cells
0 0
0
000
60-
0 0
Prolif. Cells
.
0
0
0 0
0
40-
0 )
0
X
0
.
0
3
”
0
0I
0
I
...
0
..250
/
.0
0
k z
14133
200
5E
v
-- 1 5 0 ~
)/
0
m z
O..
9**
0
00
/./
ONO
20--
”
0
0
100
0
0
0e #0/
0
0.
O
0 0
/’O
0O 0
-- 50
0
0
LPG-Prol
L P G - P g l I d 0 000 0
0
4
8
12
16
Fraction No. (ml)
24 20
’
28
0
Lipoprotein Binding
14134
A
by Smooth Muscle
Cell Proteoglycans
loo0
1
1000
,
..../-.
i
...
200
5
160
.......
16
....... 160 1 20
I
4 T
40 0
a
c ...........
0
1
240
160
9
4 x z
I
80
120
n
C
i
0 0
............. !240 200
...._ /
0
40
0
0
4
E
1 2
1 6 24 20
28
Fraction No.(ml)
FIG. 2. Affinity chromatography of 36S-labeledmacromolecules secreted by proliferating cells after treatment with chondroitinase AC, chondroitinase ABC, and heparitinase. The macromolecules synthesized by proliferating cells and those that were bound to the Sepharose-low density lipoprotein column (Fig. 1) were subjected to extensive treatment with chondroitinase AC (panel A ) , chondroitinase ABC (panel B ) , and heparitinase (panel C ) . The treated samples were refractionated with a Sepharose-LDL column also using a linear NaCl gradient (dotted lines).
LDL of the proteoglycan subfractions PG-Rest I, PG-Rest 11, PG-Pro1 I, and PG-Pro1 I1 obtained after affinity chromatography of the secreted PGs from a subculture of resting and proliferating cells (Fig. 1).The data obtained indicated that LDL hadsimilaraffinities for overlapping regions of the profiles for PGs from resting and proliferating cells (PG-Rest 11, PG-Pro1 I), Ka= 11.2 & 3.2 and 13.7 f 3.2 p"', respectively. For thoseregions of the peaks in Fig. 1 not overlapping (PG-Rest I and PG-Pro1 II), LDL showed different apparent affinity constants ( K , = 4.2 & 1.4 and 21.8 & 3.6 @ - I , respectively). In each preparation, LDL showed higher affinity with those PGs eluted at high NaCl concentrations. Hydrodynamic Size of LDL-binding CSPGs and GAGsExclusion gel chromatography in dissociative conditions was
........ 0 0
oooooo
0
oono.
,
4
8
,
,
1 2
,
,
,
1 6
.oOo?~Q~Qo 20
24
28
Fraction No.(ml)
FIG. 3. Affinity chromatography of '%-labeled macromolecules secreted by resting cells after treatment with chondroitinase AC, chondroitinase ABC, and heparitinase. The macromolecules synthesized by resting cells and those that were bound to the Sepharose-low density lipoprotein column (Fig. 1) were subjected to extensive treatment with chondroitinase AC (panel A 1, chondroitinase ABC (panel B),and heparitinase (panel C). Fractionation conditions were as in Fig. 2.
used to evaluate the hydrodynamic size profiles of the LDLbindingCSPGs produced by the resting and proliferating cells. The most prominent component from the PGs of the proliferating cells showed a Kay of 0.13 f 0.03 ( n = 4) in Sephacryl-400 HR (Fig. 4A). The CSPG from resting cells run in the same column (Fig. 4B) showed that the most prominent component had a K., = 0.23 f 0.05 ( n = 4). This may represent a difference from 1.1 X lo6 to 0.8 X lo6 in M , of the main components, based on the calibration curve provided by the manufacturers of Sephacryl for random polymers. The GAGs liberated by reductive p-elimination were
Lipoprotein Binding
by Smooth Muscle Cell Proteoglycans
TABLEI Affinity constants of LDL for proteoglycans and glycosaminoglycans secreted by resting and proliferating cells evaluated by gel mobility shift analysis
14135
900,
-
I
700
v)
Y)
*-
2
,uMLDL
0.7 f 0.3 6.9 f 1.6 Resting 3.7 f 1.8 22.7 f 7.2 Proliferating 0.8 f 0.4 5.8 f 3.3 2 Resting 3.2 f 1.9 18.8 f 5.0 Proliferating a Measurements were carried out using two separated preparations of human ASMC obtained from two separate subcultures. Also, the lipoprotein preparations used were from two different donors. The values are presented as averages & S.D. of 6 measurements carried out in the same gel. 1
700-
T
X
e
r
300
100
0 30
40
50
60
70
90
100
conditions of the glycosaminoglycans from LDL-binding proteoglycans secreted by proliferating (filled circles)and resting cells (empty circles).The GAGs were obtained by reductive 0elimination. The chromatographic conditions were the same as those used in Fig. 3.
A
.
35s 3H
4'.L
1 1 0
VT
Prolif. Cells
0
o Rest. Cells
0
0
0
30
40
50
60
70
80
90
350
100
0
I x
X
80
Fraction No. (mi) FIG.5. Gel exclusion chromatography under dissociative
.::.
600 ..
500
0
I
,
I
0
50
0
0 0
-10
30
40
50
60
70
ao
so
i
loo
Fraction No. (ml) FIG.6. Gel exclusion chromatography under associative 30
40
50
60
70
80
90
100
Fraction No. (ml)
conditions of LDL-binding proteoglycans secreted by proliferating (filled circles) and resting cells (empty circles).The proteoglycans obtained from the cell culture media by affinity chromatography on Sepharose-LDL were chromatographed in Sephacryl500 HR in 10 mM Hepes buffer, pH 7.2, containing 150 mM NaCl and 4 mM CaC12.
FIG.4. Gel exclusion chromatography under dissociative conditions of LDL-binding CSPG from proliferating (A) and resting cells ( B ) .CSPGs from the two cell phenotypes that were isolated by affinity chromatography from the cell culture media were dissolved in buffer with 4 M guanidine HCI and analyzed in a column of Sephacryl-400 HR (100 X 1 cm). The circles are the 35S-associated secreted by the proliferating cells is larger because it contains radioactivity, and the squares the one from [3H]leucine in the core chains of chondroitin sulfate longer than the chains present protein. in the CSPG secreted by the restingcells. Although the above also examined by gel exclusion chromatography in Sephacryl400 HR in dissociative conditions. The results are presented in Fig. 5. The K., of the chondroitin sulfate GAG from the proliferating cells was 0.35 f 0.02 ( n = 3). This wasvery similar to that of a CS standard with M , average of 6 X lo4. The Kavof the resting cell GAG was 0.43 f 0.04 (n = 3), which was similar to that obtained in the same column with CS of M , average = 3.5 X lo4. The 3H-labeled core proteins of the LDL-binding CSPGs of the resting and proliferating cells were obtained after extensive treatment with chondroitinase AC and ABC and were examined by SDS-gel electrophoresis in 4% polyacrylamide. No differences in the banding patterns were observed and the M , of the main band was -4 x lo6. The resultsindicatethattheLDL-bindingCSPG
gel exclusion chromatography in dissociating conditions provides information about the CSPG monomers, the extent of aggregation present in aphysiological environment is not known. To gainsome insightintothisaspect,theLDLbinding CSPG from resting and proliferating cells was examined by gel exclusion chromatography in Sephacryl-500 H R in associative conditions (buffer 10 mM Hepes, 150 mM NaC1,4 mM CaC12).In this column, themajor fraction of the CSPG from proliferating cells was eluted with a KaV= 0.27, much earlier than that of quiescent cells (Kav= 0.66), thus indicating that without denaturing agents the LDL-binding CSPG from proliferating cells has a higher tendency to form large aggregates than those of resting cells (Fig. 6). We evaluated also thehydrodynamic size of thenonoverlapping subfractions (PG-Pro1 I1 and PG-Rest I) obtained from one
14136
Lipoprotein Binding
by Smooth Muscle Cell Proteoglycans
of theLDL-affinitychromatographyexperiments (Fig.1). The results confirmed that those subpopulationsof secreted proteoglycans thatareelutedfirst from theLDL-affinity column by the NaCl gradient have a lower size range than those eluted later. The size distribution of the nonoverlapping subfractions (PG-Rest I, PG-Pro1 11) are shown in Fig. 7. Those subfractions overlapping in the LDL-affinity column showed similar elution patterns in gel exclusion chromatography (not shown).
increased with respect to those synthesized by nontreated cells (21). Our experimental design intended to look a t alterations associated with proliferation of only those PGs with affinity with LDL in human ASMC. The results indicate that most of the PGs secreted by these cells are capable of interacting as large PGs with human LDL and that they can classified be that contain mainly isomeric C6S and C4S glycosaminoglycan chains with similar extents of sulfation. We used a cell culture model in which the phenotypic modulationfrom quiescent to proliferating cells depends on PDGF supplied by the added DISCUSSION serum (25, 33, 34). Inthese cells, we observed a similar Proliferation consistently stimulates the production of se- alteration in the large CSPG provoked by proliferation as cretedextracellularmatrix macromolecules by vascular that observed in M. nenestrina cells (21). This was evidenced smooth muscle cells in tissue andcell cultures (1). Synthesis as a n increase of approximately 20% in the Mr of the PG of CSPG is stimulated by increased activity of the enzymes caused mainly by the increase in length of the CS chains. that catalyze this process (17). Growth factors seem to be theHowever, contrary to the findings in Macaca cells, we did not main modulators of cell proliferation and of the consequent observe alterations in the ratio of the C6S to C4Sglycosamisynthesis of PGs of bovine (31), human (32), and primate noglycans. Proteoglycans from aorta,rich in C6S, andof large molecular weight have been shown repeatedly to have high cells (21). However, recently it was shown that PDGF and affinity for LDL (30, 35). Furthermore, Wagner et al. (36) transforming growth factor-I@, acting together, can induce found that the most prominent PG associated with human increased CSPG biosynthesis without proliferation in primate ASMC. Actingseparately, these factorsprovoke proliferation. atherosclerotic lesion is a CSPG larger than the equivalent component isolated from normal intima-media. In relation to Such proliferation is accompanied by a marked increase in of steady state levels of the mRNA for the Versican-like core this, Alves and Mourao (23) demonstrated that the length protein. Also, the chain length of the chondroitin sulfate is the chondroitin sulfate chain of proteoglycans isolated from human aorta is a major determinant of the affinity of PGs and GAGs for LDL.OurresultsindicatethattheCSPG synthesized and secreted by proliferating cells bound LDL A more efficiently than that secreted by quiescent cells. This appearsrelatedto arelativelyhigher proportion of high PG-Pro1 I I molecular weight CSPG secreted by the proliferating than in the resting cells. The present experiments, however, do not allow discrimination if this was caused by an increasedsecretion of the large CSPG or a reduced secretion of a population 0 of smaller CSPG in the proliferating cells. The resultssuggest that the CSPG synthesized by proliferating smooth muscle cells in lesions may have a special capacity to retain apoB100-containing lipoproteins. Themobility shift assay, as applied here, is based on the simplified assumption that LDL interacts througha homogeneous single classof binding sites. Previous results from our laboratory, in which the sequences of apoB-100 responsiblefor theassociationwitharterial CSPG were identified, suggest that this may be the case (37, 38). Thepresent mobility shiftanalysisresultsappeared especially significant because the affinityof LDL for the PGs 09 1 0 and GAGs was evaluated in ionic conditions thatmay resemI I ble those presentin the arterial intima.Differences in neither PG-Rest I 0 / O 600 susceptibility to thelyases tested nor in the nature or propor1 0 I tion of the unsaturated disaccharides obtained were observed. I X This supports theconclusion that the lengthof CS chains in the PG synthesized by the proliferating cells is the main factor responsible for the increased affinity with LDL. Probably without denaturing agents, as in the band shift analysis, the CSPG withlarger CS chains presentsa matrix with which apoB lipoproteins could interact more efficiently. Extension of conclusions supported by the present experiments to the situation in acomplex tissue as the arterial intima can only be speculative. However, evidence that proFraction No. (rnl) teoglycans may contribute to entrapment and modification of (1-5, 8-14). Our FIG. 7. Gel exclusion chromatography under dissociative LDLduringatherogenesisissubstantial conditions of the proteoglycan subfractions (PG-Rest I and results suggest the additionalpossibility that atsites of lesion PC-Prol 11) obtained from the LDL-affinity chromatography development, when smooth muscle cells begin to proliferate (Fig. 1). The proteoglycan subfractions obtained from the LDL- in the intima, thecomposition of extracellular proteoglycans agarose column were dialyzed against H 2 0 ,lyophilized, and dissolved forwhich LDL could have special in dissociating buffer. The chromatographic conditions were similar mav be shifted to one to those of Fig. 4. affinity. The hypothetical existence of such microenviron-
Lipoprotein Binding
by Smooth Muscle Cell Proteoglycans
ments could contribute both to focal LDL accumulation in the intima and to facilitate its modification and subsequent uptake by resident macrophages (4,5, 7, 24, 30, 32). REFERENCES 1. Wight, T. N. (1989) Arteriosclerosis 9, 1-20
2. Berenson, G. S., Radhakrishnamurthy, B., Srinivasan, S. R., Vijayagopal, P., Dalferes, E. R., Jr., and Sharma,C. (1984) Exp. Mol.Pathol. 41,267287 3. Camejo, G. (1982) A d a Lipid Res. 1 9 , 1-53 4. Rosenfeld, M. E., Carew, T. E., von Hodenberg, E., Pitman, R. C., Ross, R., and Steinberg, D. (1992) Arteriosclerosis Thrornb. 12,985-995 5. Nievelstein, P. F., Fogelman, A. M., Mottino, G., and Frank, J. S. (1991) Arteriosclerosis Thromb. 11,1795-1805 6. Hoff, H. F., and O’Neil, J. (1991) Arteriosclerosis Thromb. 11, 1209-1222 7. Yla-Herttuala, S. (1991) Ann. Med. 23, 561-567 8. Hurt-Camejo, E., Camejo, G., Rosengren, B., Lbpepez, F., Ahlstrom, C., Fager, G., and Bondjers, G. (1992) Arteriosclerosis Thrornb. 1 2 , 569-583 9. Hoff, H., and Bond, G. (1983) Artery 12,104-116 10. Hoff, H., and Wagner, W. D. (1986) Atherosclerosis 61,231-236 11. Alavi, M. Z., Wasty, F., Li, Z., Galis, Z., Ismail, N., and Moore, S. (1992) Atherosclerosis 95, 59-67 12. Faber, M. (1949) Arch. Pathol. 48, 342-361 13. Bihari-Varga, M., Gergely, J., and Gero, S. (1964) J . Atheroscler. Res. 4, 106-109 14. Hollander, W. (1976) Exp. Mol. Pathol. 26, 106-120 15. Schwartz, S., Campbell, J. R., and Campbell, J. H (1986) Circ. Res. 68, 427-444 16. Thyberg, J., Hedin, U., Sjolund, M., Palmberg, L., and Bottger, B. (1990) Arteriosclerosis 10,966-990 17. Hollman, J., Thiel, J., Schmidt, A,, and Buddecke, E. (1986) Exp. Cell Res. 167,484-494
14137
18. Wi ht, T. N., Kinsella, M. G., Lark, M. W., and Potter-Perigo, S. (1986) 8iba Found. Symp. 124,241-259 19. Edwards, I. J., and Wagner W. D. (1988) J . Biol. Chern. 263,9612-9620 20. Schmidt, A., and Buddecke‘ E. (1990) Eur. J. Cell Biol. 52,229-235 21. Schonherr, E., Jhalaeine;, H. T., Sandell, L., and Wight, T. N. (1991) J. Biol. Chem. 266,1740-1746 22. Ross, R. (1991) Trends Cardiouasc. Med. 1,277-282 23. Alves, C. S., and Mourao, P. A. (1988) Atherosclerosis 73, 113-124 24. Hurt-Camejo, E., Bondjers,G., and Camejo, G. (1990) J. Lipid Res. 31,43454 25. Fa er, G., Hansson, G., Ottosson, P., Dahllof, B., and Bondjers, G. (1988) Cell Res. 176,319-335 26. Gckferg, R. L., and Kolihas, L. M. (1990) Connect. Tiss. Res. 24,265-275 27. Rezvin A. (1989) BioTechni ues 7,346-355 Shea, M., and Ackers, K. (1986) Methods 28. Brenoditz, M., Senear D. Enzymol. 130,132-181 29. Lindhal U. Backstrom G., Jansson, L., andHallen, A. (1973) J. Biol. Cheml 248,7234-724i 30. Hurt-Camejo E. Came’o G. Rosengren B. Lbpez, F., Wiklund, O., and Bondjers, G. (i990) ih Res. 31,188711398 31. Vijayagopal, P., Ciolino, H.g., and Berenson,G. S. (1992) Biochirn. Biophys. Acta 1136,129-140 32. Chen, J.-K., Hoshi, H., and Mckeehan, W. L. (1991) In Vitro Cell Deu. Biol. 27A, 6-12 33. Fager, G., Camejo, G., and Bondjers, G. (1992) In Vitro Cell Deu. Biol. 28A, 168-175 34. Fa er, G., Camejo G. Olsson U. Ostergren-Lunden, G., and Bondjers, G. t1992) In Vitro %ell Deu. B$l. 28A, 176-180 35. Srinivasan, S. R., Vijayago al, P ,Eberle, K , Radhakrishnamurthy, B., and Berenson, G. S. (1989) fiochim. Biophys:Acta 1006, 159-166 36. Wagner, W. D., Salisbury, B. G., and Rowe, A. H. (1988) Arteriosclerosis
2,
2 ti
__.
R~ 4117-41 7
37. Camejo, G., Olofsson, S.-O., Lbpez, F., Carlsson, P., and Bondjers, G. (1988) Arteriosclerosis 8, 368-377 38. Olsson, U., Camejo, G., Olofsson, S.-O., and Bondjers, G. (1991) Biochirn. Biophys. Acta 1097,31-44
