Vol. 266, No. 31, Issue of November 5, pp. 21287-21292,1991 Printed in U.S.A.

THEJOURNAL OF BIOLOGICAL CHEMISTRY 0 1991 by The American Society for Biochemistry and Molecular Biology, Inc.

Lipoprotein(a) Mediates HighAffinity Low Density Lipoprotein Association to Receptor Negative Fibroblasts* (Received for publication, April 7,

1991)

Gert M. KostnerS and HaraldK. Grillhofer From the Institute of Medical Biochemistry, University of Graz, Harrachgasse 21/III, A-8010 Graz, Austria

Lipoprotein(a)(Lp(a))is an acute phase protein with structures of kringle-5 and theprotease domain of Plg. Apo(a) unknown function. Lp(a) binds to low density lipoproexhibits a great size heterogeneity with molecular masses of tein (LDL) receptors, as well as to plasminogen (Plg) 300-900 kDa, which is genetically determined (Utermann et receptors. Preincubation of normal human skin fibro- al., 1987; Gaubatz et al., 1990). blasts with Lp(a) or with apo(a) cause a severalfold Due to the presence of apoB as an integral protein, Lp(a) increase ofLDL binding. Plg and kringle-4 of Plg have binds to the LDL receptor (LDL-R) (Krempler et al., 1983; no effect. LDL receptor-negative fibroblasts respond Armstrong et al., 1990; Steyrer and Kostner, 1990). The role uponpreincubation with apo(a) with high affinity of the LDL-R for the in vivo catabolism of Lp(a) is still binding of LDLwith Kd values that are almost identical with those of LDL binding to the LDL receptor. Incu- unclear. There are, however, strong indications that Lp(a) bation of apo(a)-pretreated fibroblasts with anti-apo(a) plays a role in hemostasis and fibrinolysis (reviewed by Miles completely abolishes the increment of LDL binding. and Plow, 1990); Lp(a) binds to Plg receptors on endothelial The high affinity LDL binding toLDL receptor-nega- cells (Hajjar et al., 1989) and on macrophages (Miles et al., tive fibroblasts could be dissociated by approximately 1989). Lp(a), in addition, interfereswith Plg binding to fibrin 80 and 64%with 6 mg/ml proline and30 mg/ml NaCl, clots (Loscalzo et al., 1990), has a high affinity to fibronectin respectively, but not with dextran sulfate. The Lp(a)- (Salonen et al., 1989) and proteoglycans (Bihari-Varga et al., and apo(a)-triggeredLDL binding to fibroblasts have 1988), and increases the fibrinolysis time triggered by strepno effect on LDL internalization. These findings may tokinase (Karadi et al., 1988; Loscalzo et al., 1990). All these reflect a key function in the role as an acute phase data have been put into the perspective of the pronounced protein andmay berelevant to the high atherogenicity atherogenicity of Lp(a). of Lp(a). Here, we report that preincubation with Lp(a) or apo(a) triggers high affinity binding of LDL to normal, as well as to LDL-R-deficient, fibroblasts; this increment binding of LDL Lp(a)’ is currently one of the most studied human plasma is distinct from interaction with the LDL receptor. lipoproteins (reviewed by Kostner, 1976; Utermann, 1989; EXPERIMENTAL PROCEDURES Scanu andFless, 1990;Miles and Plow, 1990) not only because Cell Cultivation and Binding Studies-These were carried out esof its documented atherogenicity (Hoefler et al., 1988) but sentially as described earlier (Krempler et al. 1983;Hesz et al., 1987). even more so because of its structural homology to plasmin- HSF from healthy normolipemic donors between the fifth and tenth ogen and itspotential linkto thefibrinolytic system (McLean passages were seeded in multitray dishes (6- or 12-well) and grown, et al., 1987; Karadi et al., 1988). Lp(a) is an acutephase in DMEM supplemented with 10% fetal calf serum, in a humidified protein with unknown function (Maeda et al., 1989). Struc- 95% air, 5% CO, atmosphere at 37 “C. In the logarithmic growth turally,Lp(a) maybe best characterized as an LDL-like phase between days 2 and 4,the medium was replaced by DMEM, particle with apoB-100 as the integral protein, with the gly- 10% lipoprotein-deficient fetal calf serum (LPDS), supplemented with various lipoproteins or proteins. After incubation for various coprotein apo(a) attachedto itby one or more disulfide bridges periods of times, the cells were washed, chilled to 4 “C and studied (Fless et al., 1986). Human apo(a) hasbeen cloned and found for LDL binding. Two LDL receptor-deficient fibroblast lines (FH to exhibit striking structural similaritiswith Plg; kringle-4 of cells) were also studied the FH-808 line was purchased from ATCC, Plg is repeated up to 37-fold with a high degree of conservation and the FH-IV-1line was obtained from U. Beisiegel (University of (McLean et al., 1987). In addition, apo(a) containsconserved Hamburg) (Clemens et al., 1986).Both cells are considered to lack

LDL-R. In binding studies using the filter assay (Steyrer and Kostner, 1990) less than 5% of LDL binding in comparison to normal HSF Austrian Science Foundation and by Grant 3382 from the Austrian was observed. LDL was iodinated according to McFarlane (1958),yielding prepNational Bank. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore arations with a specific activity of 350-640 cpm/ng of protein. LDL be hereby marked “advertisement” in accordance with 18 U.S.C. binding was studied by incubation of cells at 4 “C for 1 h with increasing amounts of ”‘I-LDL, followed by extensive washing. The Section 1734 solely to indicate this fact. This paper is dedicated to Prof. Dr. Anton Holasek on the occasion cells were solubilized in 0.3 N NaOH and counted in a y-counter. Aliquots were used for protein quantitation by the Lowry procedure of his 70th birthday. $ To whom correspondence should be addressed. Tel.: 43-316-380- (Lowry et al., 1951). The binding measured according to this procedure is referred to as “total binding.” For measuring “specific bind4200. The abbreviations used are: Lp(a), lipoprotein(a); FH, familiar ing,” the cells were incubated in a similar manner in the presence hypercholesterolemia; HSF, human skinfibroblasts; LDL, low density and absence of a 50-fold excess of inactive (“cold) LDL. Specific lipoprotein; LDL-R, low density lipoprotein receptor; HDL, high binding was calculated by subtracting nonspecific binding from total density lipoprotein; LPDS, lipoprotein-deficient fetal calf serum; Plg, binding. The nonspecific portion amounted to 4 8 % of the specific plasminogen; apo(a), apolipoprotein(a); DMEM, Dulbecco’s modified binding. For studying the impact of Lp(a) on cholesterol biosynthesis, Eagle’s medium; HEPES, 4-(2-hydroxyethyl)-l-piperazineethanesul- cells were incubated with variuos lipoproteins and the incorporation fonic acid; SDS, sodium dodecyl sulfate. of [14C]octanoateinto the nonsaponifiable sterol fraction was meas-

* This work was supported by Grants S-4201and S-4202from the

21287

21288

Lp(a) Mediates

LDL Association Receptor to Negative

ured according to standard procedures (Liscum and Faust, 1987). In order to study the interaction with fibroblasts, purified apo(a) was radiolabeled with lZ5Iaccording to McFarlane (1958), yielding a preparation with a specific activity of approximately 10 cpm/ng of protein. Cells were incubated with radiolabeled apo(a) (50 pg/ml) for 72 h, followed by washing in a manner identical with that used in all other experiments. Cells were solubilized in 0.3 M NaOH, and the trichloroacetic acid-precipitable radioactivity was measured. In order to assess the nature of apo(a)-triggered binding, specific polyclonal antibodies against apo(a) were included in the incubate. After preincubation with apo(a) for 72 h, 100 pg/ml immunoaffinitypurified antibodies were added to the medium, followed by a further incubation period of 3 h at 37 "C. In control experiments, an equal amount of nonimmune IgG prepared by (NH&S04 precipitation was used. Thereafter, normal binding of lZ5I-LDLat 4 "C for 1 h was studied. In order to learn more about the nature of the apo(a)-triggered LDL binding, FH-808 cells were preincubated with 50 pg/ml apo(a) for 72 h followed by incubation with 10 pg/ml Iz5I-labeledLDL at 4 "C for 1 h. Thereafter, cells were incubated for 15 min at 4 "C with the following solutions: (i) 5 mM NaCI, 10 mM HEPES,pH 7.4 (reference buffer), (ii) reference buffer containing 4 mg/ml dextran sulfate (Pharmacia LKB Biotechnology Inc.), (iii) reference buffer containing 5 mg/ml prolin, and (iv) reference buffer containing 30 mg/ml NaCI. After one more wash with reference buffer, the radioactivity remaining associated with the cells was counted after solubilization in 0.3 N NaOH. Ligand Blots and Filter Assays-These assays were performed according to Schneider et al. (1985), as described in detail earlier (Steyrer and Kostner, 1990). In short, thecultured cells were washed and solubilized in 10 mmol/liter Tris-maleic acid, 2 mmol/liter CaClz, 1% Triton X-100, 1 mmol/liter phenylmethylsulfonyl fluoride, and 0.05 mmol/liter leupeptin, pH 6.0. Aliquots of this solution containing 50 pg of protein were separated under nonreducing conditions on 7% polyacrylamide gels containing 0.1% SDS followed by electrophoretic transfer of the proteins onto nitrocellulose in 20% methanol buffer (90 min, 4 "C, 150 mA). The blots were then incubated in 0.15 mmol/ liter NaCI, 2 mmol/liter CaClz containing 3 pg/ml lZ5I-labeledLDL (specific activity, 550-640 cpm/ng protein); this concentration was in the linear range of LDL binding to theblotted receptor. The washed blots were autoradiographed on Amersham Hyperfilm MP for 24 h at -70 "C. The filter assay was performed as described (Schneider et al., 1985). Preparation of Lipoproteins and Proteins-This was performed as described in numerous previous publications (Krempler et al., 1980, 1983; Kostner et al., 1989; Steyrer and Kostner, 1990). Plasma was obtained from a fasting normolipemic subject. LDL was prepared by repeated ultracentrifugation at d 1.025-1.050, HDL3 a t d 1.125-1.210, and Lp(a) at d 1.070-1.110. Lp(a) was normally prepared from pooled plasma. In some cases, apo(a) was prepared from plasma of donors who were monocygotic for different isoforms (Steyrer and Kostner, 1990). The d 1.070-1.110 fraction was chromatographed over an antiLp(a)-containing Sepharose CL-GB column. The purity of the Lp(a) preparation, especiallywith respect to contaminating LDL, was tested by double-antibody Laurell electrophoresis (Steyrer and Kostner, 1990). Apo(a) was prepared by heparin-Sepharose column chromatography after reductive cleavage of Lp(a) with dithiothreitol (Armstrong et al., 1990). All lipoprotein preparations were used within 2 weeks of preparation. Plasminogen and the kringle 4 portion of Plg were electrophoretically pure preparations and were a gift from Dr. Patty, University of Budapest. Electron Microscopy-This was performed as described previously (Hesz et al., 1987). Conjugation of lipoproteins with negatively charged colloidalgoldwas performed according to Handley et al. (1981). Fibroblasts preincubated with various lipoproteins were washed and incubated with gold-labeled LDL for 5-20 min a t 37 "C. After washing, the cells were postfixed for 1 h in 1%Os04, washed, dehydrated, and embedded in Epon 812. Ultrathin sections were cut with an ultratome I11 from PharmaciaLKB Biotechnology Inc., stained with uranyl acetate followed bylead citrate, and examined in a Phillips EM 201. RNA Zsolation and Quantitation-Total RNA from fibroblasts was extracted by a modified method of Chirgwin et al. (1979).Cell cultures in 100-mm Petri dishes were dissolved in 7 ml of GTC buffer (4 mol/ liter guanidinium isothiocyanate, 0.05 mol/liter Tris-HC1, pH 7.5, 0.01mmol/liter EDTA, 2% N-lauroylsarcosine, 1%P-mercaptoethanol) for 2 min. 6 mlof 100% ethanol were added, and RNAwas precipitated overnight at -20 "C. The RNA was pelleted by centrif-

Cells

ugation a t 10,000 X g (20 min) at 4 "C. This whole procedure was repeated once. The pellet was dissolved in 0.8 mlof SDSEB (50 mmol/liter Tris-HC1, pH 9.0,0.1 mol/liter NaCl, 10 mmol/liter EDTA, 0.5% SDS) and extracted twice with 0.6mlof phenol and reextracted with chloroform. RNAwas again precipitated with ethanol, washed, and dryed. For Northern blot analysis, RNA was electrophoresed in formaldehyde/agarose followed by transblotting to nylon membranes. The LDL transcript was prehybridized for 6 hat 42 "C and thenhybridized for 20 h a t 42 "C with the 1.05-kilobase PST-I fragment of LDL-R cDNA (pLDL-3 clone from ATCC). Quantitation by densitometry was performed by comparing the message of the LDL-R with that of actin. RESULTS

In previous experiments, we studied the interaction of Lp(a) in comparison with LDL with normal human skin fibroblasts and noticed high affinity binding to theLDL-R. The affinity of Lp(a) was some 30-40%lower as compared with LDL (Krempler et al., 1983). It was also shown that preincubation of fibroblasts with Lp(a) down-regulated cholesterol biosynthesis by inhibiting the activity of hydroxymethylglutarylCoA reductase. We repeated similar experiments here with affinity-purified Lp(a) that was free of LDL as verified by double-antibody Laurell electrophoresis (Steyrerand Kostner, 1990). At identical cholesterol concentrations in the incubate, and after 48 h of preincubation, Lp(a) was only onehalf to one-thirdas active in down-regulating cholesterol biosynthesis as compared with LDL (Table I). Preincubation of fibroblasts with LDL is known to downregulate the LDL-R number; a similar effect was expected with Lp(a). In order to test this, normal human skin fibroblasts (HSF) were preincubated for 72 h with culture medium containingLPDS supplemented with Lp(a) or othersubstances for comparison. After washing the cells, LDL binding was studied (Fig. 1).As expected, preincubation with 50 pg/ ml LDL led to a down-regulation of '"I-LDL binding and 100 pg/ml apo-E-free HDLB led to an up-regulation. We were, however, surprised to learn that preincubation of fibroblasts with 50 pg/ml Lp(a)-stimulated lZ5I-LDLbinding that was even more striking, as compared with HDL3. This effect was entirely due to the specific apo(a) protein, since apo(a) by itself had essentially the same effect (Fig. 1). Apo(a)-free Lp(a) (Lpa-) behaved like LDL and reduced lZ5I-LDLbinding (data notshown). Because of the structuralhomology, we also studied Plg, as well as the kringle 4 moiety of Plg. None of TABLEI Down-regulation of cholesterol biosynthesisin normal human skin fibroblasts by affinity-purified Lp(a) in comparison with LDL and meuinolin Lp(a) was purified by immune-specific adsorbers, as described earlier (Steyrer and Kostner, 1990) and was free of LDL. On day 4 after plating, normal human skin fibroblasts were incubated for 48 h with LPDS supplemented with various concentrations of LDL, Lp(a), or mevinolin. Thereafter, 2 pCi of [14C]octanoate/mlof medium were added, and the incorporation of label into the nonsaponifiable sterol fraction was determined. The values are expressed in percent of the LPDS value and represent means + S.D. of two experiments carried out in triplicate. Percent of radioactivity in relation to the LipoproteinLPDS-incubatedcells cholesterol Ma) Mevinolin LDL dm1

0 (reference) 5 10 20 40

%

%

mol/liter

%

lo-'

25 f 2 13f 1

100

64 f 5 29 f 4 4 51 + 6+ 42 3122f f21

86 f 7

69+ 5

Lp(a) Mediates

LDL Association Receptor to Negative '

''

:.

100

A. :

i' 0 5' 0

+

.

:

10

t

1

Time course of Apola) and Lp(a)-triggered LDI, binding to normal humanfibroblasts Onday 3 after seeding normal HSF, the culture medium was supplemented with 50 pg/ml of apo(aj or with 100 pg/ml of Lp(a) protein in LPDS, respectively.A t the given time intervals, total LDL binding was studied using 10 pg/ml '"I-LDL. The values for LDL binding to HSF incubated with LPDS alone were measured in comparison. The values are means f S.D. of two experiments performed in triolicate. ND, not determined.

LpW

HDL LPDS Plg,K4

20

30

LDL-protein

LDL

5o

40

21289

TABLEI1

apo-a

200 ,

Cells

Incubation time

uglml

FIG. 1. Binding of LDL to normal HSF. Fibroblasts of normolipemic healthy donors were incubated in DMEM containing 10% fetal calf serum. On day 3 in the logarithmic growth phase, the medium was replaced by DMEM, 10% LPDS supplemented with the following substances: LDL (50 pg/ml), Lp(a) (50 pg/ml), HDLs (100 pg/ml, apoE-free), apo(a) (25 pg/ml), plasminogen (25 pglml), or kringle 4 of Plg ( K 4 , 25 pglmlj. All concentrations are given as protein. After further incubation for 72 h, cells were washed and incubated for 1 h at 4 "C with increasing amounts of ""I-labeled LDL (340 cpm/ng of protein). Finally, cellswere washed, solubilized in 0.3 N NaOH and counted.For each substance used during preincubation, a no-cell blank was subtracted. The values are means of triplicate analyses of one characteristic experiment. The standard deviation

h

6 12 24 48 72 96

LPDS + apo(a)

LPDS + Lp(a)

LPDS

ng '"I-LDL boundfmg of c d l protein

ND

55.7 f 6.5 75.7 f 6.1 121.1 f 8.8 142.3 -C 11.6 162.2 f 12.4 145.9 f 14.3

62.4 f 5.4 68.7 f7.2 72.5 f 6.9 95.8 f 8.6 125.7 f 10.8

51.4 f 4.8 52.6 f 5.1 54.5 f 5.5 53.2 f 4.9 51.8 f 5.0 48.3 f 4.4

was 4 % .

300 250 200

LDL BOUND (ng/mg cell protein)

c

1

150 r

i 100

,

I

50

1-

0-

2.5 uglml 125-1 LDL

10 uglml 125-1 LDL

PREINCUBATION WITH APO-a m0 (REF)

m 2 5

0 5 0

75

1

2

3

4

FIG. 3. Ligand blots of LDL to solubilized LDL receptors. Cultivation and preincubation of cells were carried out as described in the legend to Fig. 1. After a 72-h preincubation, cells were washed and solubilized in 1% Triton X-100-containing buffer.50 pg of extract, were electrophoretically separated in 7% SDS-polyacrylamide gels and transblotted to nitrocellulose. The transblots were incubated with 3 pg/ml "'I-LDL (550 cpm/ng) followed by autoradiography (-70", 24 h). Molecular mass standards were electrophoresed simultaneously. The arrows indicate molecular masses from top to bottom: 232, 140, and 67 kDa,respectively. Extracts from the following preincubated fibroblasts areshown: lane I , LPDS (reference); lane 2, 50 pg/ml LDL; lane 3, 50 pg/ml Lp(a); lane 4, 50 pg/ml apo(aj. All concentrations are given as protein.

100

FIG.2. Influence of increasing amounts of apo(a) on LDL binding to normal HSF. The experiment was carried out as described in the legend to Fig. l. Cells were preincubated for 72 h with DMEM, 10% LPDS containing 0, 25, 50,75,or 100 pg/ml apo(a). After washing, LDL binding was measured by incubation with 2.5 and 10 pg/ml '2sI-LDL, respectively. The valuesare means of triplicate experiments. The bars indicate the S.D. these proteins had any effect. The presence of Plg in the preincubation mixture, together with LPDS and apo(a) did not abolish the effect of apo(a) (data not shown). In subsequent experiments,we repeated these studies with three different normal HSF lines, using Lp(a) and apo(a) from donors with four different isoforms (B-type, S-1, S-2, and S-3, according to Utermann (1989)). In all cases, Lp(a) or apo(a) stimulated '*'I-LDL binding,whereasLDL suppressed it. The stimulation of LDL bindingby different apo(a) isoforms was virtually identical. Themagnitude of apo(a)-triggered Y - L D L binding to fibroblasts was concentration-dependent. Fig. 2 shows that preincubation of fibroblasts with25 pg/ml apo(a) had already had asignificant effect. The stimulation of LDL binding increased almost linearlyfrom 25 to 100 pg/ml apolipoprotein, without reaching a plateau. Higher concentrations were not used because of the limitedaccessibility of apo(a). We also studied the timecourse of the effect of apo(a) and Lp(a) on LDL binding to HSF. TableI1 shows that the time course to trigger LDL binding was different for apo(a) and

Lp(a); whereas LDL binding triggered by apo(a) proceeded almost linearly from 12 to 72 h, followed by a decrease a t 96 h, therewas only amoderate increaseof LDL bindingbetween 12 and 72 h but a maximum a t 96 h. Longer incubation times were not applied because of the changes in cell density a t prolonged incubation periods. At the beginning, we were tempted to conclude from these results that Lp(a) and apo(a) may cause an increase of the LDL-R number on fibroblasts, possibly by stimulating the transcription. We, therefore,preincubatedfibroblasts with LPDS alone or with LPDS supplemented with 50 pg/ml each of apo(a), LDL, HDL, orPig. After 72 h, cells were recovered and quantitative Northern blotswere performed. In relation to LPDS preincubated cells, LDL caused a reduction of the LDL-R mRNA to approximately one-third, and HDL caused a n increase of the message by a factor of 2.2. Apo(a), Lp(a), and Plg had no measurable effects. The influence of apo(a) on LDL-R number was further studied by ligand blot and filter assaymethods. Fig. 3 shows a ligand blot of LDL-R isolated from fibroblasts preincubated with LPDS, LDL, Lp(a), or apo(a). LDL-preincubated cells exhibited a significantly lower signal, whereas the intensity of the autoradiographs obtainedwith Lp(a)- or apo(a)-preincubatedHSF were indistinguishable from that of LPDSpreincubated cells. Quantitative filter assays gave the same results (data not shown). In subsequent experiments, the effect of apo(a) on'*'I-LDL

H-IV-1

Lp(a) Mediates LDL Association to Receptor Negative Cells

21290

association to two LDL-R-deficient cell lines from patients with familiar hypercholesterolemia (FH) were studied. The cells that completely lacked LDL-R were preincubated with LPDS with or without apo(a), and high affinity binding of "'1-LDL in the presence and absence of a 50-fold excess of cold LDL was measured. In comparison, three fibroblast lines from healthy donors were used. FH cells exhibited no measurable high affinity LDL binding after preincubation with LPDS (Fig. 4). Preincubation with 50 pg/ml apo(a) led to a specific binding curve that was comparable with that of LDL binding to normal HSF preincubated with LPDS without apo(a). Scatchard analysis of these experiments revealed an association constant of LDL to apo(a)-preincubated FH cells that was very similar to that of normal HSF preincubated with LPDS (Table 111). The B,,, values for LDL binding to apo(a)-preincubated FH cells depended on the apo(a) concentration used in thepreincubation medium; FH-808 cells preincubated with 30 pg/ml apo(a) exhibited a B,,, similar to that of normal HSF preincubated with LPDS alone. Higher concentrations of apo(a) led to a gradual increase of LDL binding. If normal fibroblasts were preincubated with apo(a), K d values

remained at the same level, whereas B,,, values increased proportionally. From these results,we concluded that during preincubation, Lp(a) or apo(a) associates with normal, as well as with FH, fibroblasts, giving rise to the incremental LDL binding. In order to test thishypothesis, apo(a)-preincubated normal, as well as FH,fibroblasts were treated with polyclonal antibodies against apo(a) before high affinity LDL binding was measured. As shown in Fig. 5, the specific antibodies in fact abolished the incremental LDLbinding, whereas nonimmune IgG had no influence. It was also of interest to get some information on the chemical nature of the apo(a)-triggered LDL binding. FH-808 cells were thus pretreated with apo(a) and loaded with lZ5ILDL, as described under "Experimental Procedures." The dissociation of bound LDL was studied with the agents shown in Table IV. Whereas dextransulfatehad no measurable effect, proline dissociated approximately 80% and NaCl 46% of the associated radioactivity. In order to assess the stoichiometry of the apo(a)-triggered LDL association to fibroblasts, the following experiments

LDL BOUND (ng/mg cell protein)

I

140 I

T

T

120 100

1

T

--t

FH(+a)

00 60

40 20

J

0

n

J

NORMAL FIBROBLASTS

0

10

40

30

20

50 LPDS

LDL-protein ug/ml

FIG. 4. Specific binding of LDL to normal HSF and to FH fibroblastspreincubatedwith LPDS with (+a)or without (-a) apo(a).The experiments were carried out essentially as described in the legend to Fig. 1. Three different normal HSF and two FH cells were preincubated on day 3 after seeding with LPDS and 50 pg/ml apo(a), respectively. After 72 h, specific LDL binding was determined by incubation with Iz5I-LDLin the presence or absence of a 50-fold excess of cold ligand. The points indicate mean values of triplicate analysis; bars indicate the S.D.

TABLE I11 Binding characteristics of LDL to mrmal HSF and to FH cells pretreated with Apo(a) Three different normal skin fibroblasts (n-HSF) and two FH cell lines were pretreated with 30 or 50 pg/ml of apo(a) for 72 h, and K d and BmaX values for LDL binding were calculated according to Scatchard. Binding was determined by incubation for 1 h at 4 "C with T - L D L f a 50-fold excess of cold ligand. The values are means f S.D. of tridicate analvses. FH-808 Parameter

n-HSF -apo(a) +apo(a) -apo(a) +apo(a) -apo(a) +apo(a)

Preincubation with 30 pg/ml apo(a) Kd" 3.4 f 0.9 3.7 f 0.5

Bm.:

64.2 f 6.8 108.7 f 12.2

-b

-

4.0 f 0.3 4.8

55.3

+

-

-

+

4.3 0.6 61.7 f 7.7

Preincubation with 50 pg/ml apo(a)

K,,"

3.4 f 0.9

4.1 f 1.2

Bmuc 64.2 f 6.8 177.0 f 19.7

-

3.9 f 0.4 114.8 f 8.6

FH-808 FIBROBLASTS

10 ug/ml 125-1 LDL

4

-

4.2 f 0.4 127.1 f 10.3

K,, is given in nmol/liter. -, because FH fibroblasts preincubated in LPDS do not bind LDL, no K d or Bmaxvalues were obtained. Bmaxis given in ng of LDL protein/mg of cell protein.

Apo(d60

0 Apo(a)60*Antl-a

Apo(a)604pG

FIG. 5. Influence of specific antibodies against apo(a) on apo(a)-triggered LDL binding to normal HSF and to FH fibroblasts, respectively. The experiments were carried out essentially as described in the legend to Fig. 1. After preincubation for 72 h with LPDS in the presence and absence of 50 pg/ml apo(a), the medium was replaced by DMEM containing none (Apo(a)50),100 rg/ ml affinity-purified anti-apo(a) (Apo(a)BO+Anti-a)or 100 pg/ml nonimmune y-globulin from rabbit (Apo(a)5O+lgC). The black bars ( L P D S ) are control cells preincubated with LPDS in the absence of apo(a). After further incubation for 3 h a t 37 "C, LDL binding was measured by incubation with 10 pg/ml Iz5I-LDLfor 1h at 4 "C in the presence and absence of 50-fold excess of cold ligand. The values represent "specific" binding calculated as total minus unspecific binding and are means of two experiments carried out in triplicate; the bars indicate the S.D.

TABLEIV Dissociation of apo(a) triggered LDL-binding toFH-808cells FH-808 cells were preincubated with 50 pg/ml of apo(a) as described in the legend to Table 11, followed by lz51-LDLbinding (IO pg/ml) at 4 "C for 1 h. Cells were then washed for 15 min with HEPES buffer, pH 7.4, alone or with buffer plus 4 mg/ml dextran sulfate, 5 mg/ml proline, or 30 mg/ml NaCl, respectively. The association of '"1-LDL with FH cells is expressed as a percentage of the values observed for the buffer-washed cells. The results are means f S.D. of two experiments carried out in triplicate. Dissociating agent

Percent radioactivity remaining associated with FH 808

HEPES buffer Dextran sulfate (4 mg/ml) Proline (5 mg/ml) NaCl (30 mg/ml)

100 104 f 5 21 3 54 f 7

%

*

Lp(a) Mediates

LDL Association Receptor to Negative

Cells

21291

were performed. FH-808 cells were incubated a t day 3 after concentrations ranging from I00 mg/dl. Except for seeding with '*'I-labeled apo(a) (10 cpm/ng, 50 pg/ml) for 72 acute phases, where Lp(a) plasma levels increase severalfold h. The medium was removed,and cells were carefully washed. (Maeda et at., 1989), Lp(a) concentrations in hlood are very After solubilization of cells, the amount of trichloroacetic constant and genetically regulated. Turnover studies in huacid-precipitable apo(a)was measured. We found that under mans revealed that theplasma concentration is almost excluthese conditions,FH fibroblasts bound376 f 44 ng of apo(a)/ sively determined by therate of synthesis;thefractional for apo(a)-triggered LDL catabolic rate, on the other hand, is comparahle with that of mg of cell protein. Assuming a RmaX binding of approximately 120 ng/mg of cell protein (Table LDL (Krempleret al., 1980, 1983). Thisled to the assumption IV), a molar ratio of apo(a):apoB on the order of 3 may be thattheLDL-Rmighthe involved in Lp(a)bindingand calculated. removal in a way similar to thatknown for LDL (Brown and Wewere also interested in studying the fate of apo(a)- Goldstein, 1986). Several research groups have studied this triggered LDL binding in normal HSF and to FH fibroblasts. issue using cultured HSF (Florenet al., 1981; Krempler et al., Cells were preincubated with LPDS or with apo(a), and the 1983; Armstrong et al., 1990; SteyrerandKostner, 1990). internalization of colloidal gold-labeled LDL was studied by Most of them revealed high affinity binding to the LDL-R. electron microscopy. By inspecting the electron micrographs with a lower Kd as compared with LDL. More important for of Fig. 6, it becomes apparent that normal HSF internalizea the in oiuo situation may be the findings that transgenicmice great amount of LDL-gold complexes. After 20 min of prein- with an overexpressionof the human LDL-Rgene cataholize cubation a t 37 "C, the majority of gold particles are found in human Lp(a) by a far greater extent than normalmice (Hofsecondary lysosomes; the number of gold particles in lysomann et al., 1990). However, in discordance with the assumpsomes was independent of preincubation of cells by apo(a). tion that the LDL-R playsa role in Lp(a) catabolism are the FH-808 cells, on the other hand, internalize under identical findings that lipid-lowering drugs, such as cholestyramineor conditionsonly very smallamounts of LDL-gold. This mevinolin, which are believed to increase the LDL-R numher amount did not increase if FH-808 cells were preincubated on theliver, have littleeffect on Lp(a)levels and, if anything, with apo(a). The internalization of LDL at 37 "C was also even tendtoincreaseplasmaLp(a) (Vesshy et al., 1982; studied with "'I-labeled lipoproteins, yielding exactlythe Kostner et al., 1989). same results (data not shown). The surface replicas Fig. in 6 Because of these discrepancies, we studied the effects of reveal, in addition, that LDL-gold accumulates in coated pit Lp(a) on human skin fihrohlasts in more detail. From earlier areas on normal HSF, whereas on FH cells pretreated with studies, we knew that Lp(a) binds to the LDL-R and supapo(a), they distribute rather uniformly on surface. the presses the cellular cholesterol biosynthesis to yet a lower extent as compared with LDL (Krempleret 01.. 1983). HowDISCUSSION ever, when we triedtodown-regulate high affinityLDLbinding to HSF by preincubation with Lp(a), we were surAlthough Lp(a)has been knownfor over 25 years,its of LDL hinding. This topic biological function is unknown. Lp(a) is biosynthesized in the prised to notice an actual increase liver (Kraft et al., 1989) and circulates in the bloodstreama t was pursued further using not only Lp(a) hut also apo(a) and several controlsubstances for preincubation. As expected, preincubation of normal fibroblastq with LDL led to a down'q " . . , ? p p y A regulation and preincubation with HDL led to an up-regulation of high affinity LDL hinding. This was due to a modulation of the LDL-R number. Lp(a) surprisingly led to an upregulation of high affinity LDLbinding. It is noteworthy here that LDL binding was measured as total radioactivity associated with cells after incuhation for 1 a t 4 "C (see "Experimental Procedures"). In control experiments, we ascertained that under these conditions