The interaction of Lp(a) with normal and LDL-receptor-deficient human skin fibroblasts

ELSEVIER SCIENCE IRELAND

Chemistry and Physics of Lipids 67/68 (1994) 153-159

Chemistry and Physics of LIPIDS

The interaction of Lp(a) with normal and LDL-receptordeficient human skin fibroblasts G e r t M. K o s t n e r * , H a r a l d G r i l l h o f e r Institute of Medical Biochemistry, University of Graz, A-8010 Graz, Austria

(Accepted 12 November 1992)

Abstract The role of LDL receptors in the in vivo catabolism of Lp(a) is still a matter of controversy. Since Lp(a) binds LDL with high affinity, it was essential for this study to separate Lp(a) quantitatively from all other apo-B and apo-Econtaining lipoproteins. This was achieved by the addition of proline as a dissociating agent to all buffers during Lp(a) preparation. Normal human skin fibroblasts pre-incubated with 40 mg/ml of Lp(a) downregulated cholesterol biosynthesis by approx. 35%; the same amount of LDL caused a 90% reduction. Cholesterol biosynthesis of LDL-receptordeficient fibroblasts was not affected at all by LDL, yet Lp(a) exhibited a similar effect as in normal fibroblasts (32% reduction). An LDL-receptor-independent uptake of Lp(a) into fibroblasts must therefore be postulated. We also studied the degradation of Lp(a) in normal fibroblasts in comparison with LDL. Pure Lp(a) was only slightly degraded in relation to LDL. If fibroblasts were pre-incubated with small amounts of LDL, Lp(a) degradation was enhanced by a factor of 3-5. This effect was even more pronounced in fibroblasts pre-incubated with mevinolin. Thus the LDL receptor may play an indirect role in Lp(a) catabolism. The significance of these findings for the in vivo metabolism of Lp(a) remains to be established. Key words: Lp(a) degradation; Cholesterol biosynthesis; Lp(a) purification; Familial hypercholesterolemia

1. Introduction

Lp(a) is a lipoprotein that may be considered as an LDL or Lp particle with apo-a attached to it via * Corresponding author, lnstitut ffir Medizinische Biochemie, Harrachgasse 21, 8010 Graz, Austria. Abbreviations: FCS, fetal calf serum; FH, familial hypercholesterolemia; FH-Fb, fibroblasts from patient with familial hypercholesterolemia; HSF, human skin fibroblasts; LDL-R, low-density lipoprotein receptor; LPDS, lipoprotein-deficient serum; TCA, trichloroacetic acid.

a disulfide bridge. Thus the protein moiety consists of two apolipoproteins, B-100 and apo-a. B-100 is the ligand for the LDL receptor (LDL-R), and thus plasma levels of LDL and other apo-Bcontaining lipoproteins are strongly affected by the receptor-mediated catabolism. Whether or not this may apply also for Lp(a) is a subject of controversy: most in vitro studies of numerous laboratories demonstrate that Lp(a) binds to LDL receptors, yet with a reduced affinity as compared with LDL (Havekes et al., 1981; Krempler et al.,

0009-3084/94/$07.00 © 1994 Elsevier Science Ireland Ltd. All rights reserved. SSDI 0009-3084(93)02210-1

154

G.M. Kostner, H. Grillhofer / Chem. Phys. Lipids 67/68 (1994) 153-159

1983; Armstrong et al., 1990; Steyrer and Kostner, 1990). In vivo the situation is more controversial and far from being clear: lipid-lowering drugs, which are believed to increase the number of receptors in the liver, as for example lovastatin, hardly affect Lp(a) levels to a significant degree (Kostner et al., 1989). In familial hypercholesterolemic patients lacking LDL receptors, on the other hand, Lp(a) is increased approximately threefold in relation to the particular isoform of apo-a as compared with normals, suggesting that LDL-R does play a role in Lp(a) catabolism. In a recent report we have shown that Lp(a) as well as apo-a bind with high affinity not only to normal human skin fibroblasts (HSF) but also to fibroblasts of patients with a complete lack of LDL-R (Kostner and Grillhofer, 1991). This Lp(a) binding, however, was not affected by monoclonal antibodies against LDL-R, nor was the bound Lp(a) internalized to a significant degree. It was, however, of interest to note that surface-bound Lp(a) as well as apo-a-mediated high-affinity LDL binding to normal and LDL-R-deficient cells. The fraction of LDL bound through apo-a, however, was prevented from being internalized and degraded. In the present study we pursued this issue further and asked the question whether or not Lp(a) may influence cholesterol biosynthesis in the HSF of normal patients as well as of patients suffering from familial hypercholesterolemia (FH-Fb). In addition, degradation of Lp(a) after incubation with fibroblasts was studied and compared with that of LDL. As Lp(a) is very hard to separate from residual impurities of LDL, we developed a new procedure for the preparation of uncontaminated Lp(a). 2. Material and methods

2.1. Cell cultivation and measurement of cholesterol biosynthesis Cell studies were carried out essentially as described earlier (Kostner and Grillhofer, 1991; Krempler et al., 1987). Human skin fibroblasts (HSF) from normal donors or from an FH patient

were seeded in multi-tray Petri dishes and grown to 60-70% confluency in DMEM containing 10% FCS in a humidified incubator (95% air, 5% CO2) at 37°C. The medium was then replaced with DMEM containing 10% lipoprotein-deficient FCS (LPDS), and cells were further incubated for 24 h. Cholesterol biosynthesis was now measured according to Innerarity et al. (1986). The medium was again replaced with DMEM/10% LPDS supplemented with various concentrations of LDL or Lp(a); after incubation for 6 h, 2 #Ci/dish of 14-C octanoate were added and the incubation proceeded for an additional 18 h. After washing with PBS, the cells were solubilized in 1 ml 0.3 M NaOH, hydrolyzed for 1 h at 90°C and extracted twice with 1 ml each of hexane:isopropanol 3:2. The lipids were separated by thin-layer chromatography using petrol ether:diethyl ether:acetic acid 50:50:1, and the band corresponding to free cholesterol was scraped off and counted in a liquid scintillation counter. All values were normalized to 1 mg cell protein.

2.2. Lipoprotein degradation Lipoproteins taken up by receptor-mediated pathways are lysosomally degraded, and the free amino acids are secreted into the culture medium. In order to study this, Lp(a) and LDL were radiolabelled with 125I according to McFarlane (1985), yielding a specific activity of 100-300 counts/min/ng protein. The lipoproteins were incubated with fibroblasts pre-incubated for 48 h with LPDS in the presence or absence of 10 -6 M mevinolin, and the TCA-soluble material accumulating in the medium during 2 h incubation at 37°C was counted in a "r-counter. In some cases LPDS and mevinolin pre-treated cells were incubated for 60 min at 4°C with 5 /~g/ml of LDL. Cells were then washed with ice-cold PBS and incubated with 125I LDL or 125I Lp(a) for 1 h at 4°C and 2 h at 37°C, and degradation was measured by counting the TCA-soluble material in the medium. In all these experiments, non-cell blanks were subtracted from the results, amounting to <20% of the cell values. In addition, lipoprotein internalization was measured according to Goldstein et al. (1983).

155

G.M. Kostner, /t. Grillhofer/ Chem. Phys. Lipids 67/68 (1994) 153-159 2.3. Preparation o f lipoproteins

All lipoproteins were isolated from pooled plasma of normolipemic donors with elevated Lp(a) concentrations (40-80 mg/dl). The donors exhibited mostly the S I or $2 isoform according to U t e r m a n n (Seed et al., 1990). L D L was isolated by

repeated ultracentrifugation in a density range 1.025-1.055. For the purification of Lp(a), approx. 500 ml of pooled plasma was adjusted to d 1.060 by adding solid NaCI and ultracentrifuged at 140 000 x g for 24 h. The supernatant was removed by tube slicing and used for the purification of LDL. The infranatant was adjusted to d 1.125 with

0,2

~

0,2

i '1

0,1

30

60

90

FRACTION No.

~t

Fig. 1. Lp(a) was purified by (i) ultracentrifugation, (ii) Biogel A-I 5m column chromatography and (iii) DEAE cellulose column chromatography (see Methods). The elution pattern of this latter step is shown using an NaC1 gradient from 0 to 0.2 M. The fractions were analyzed by double antibody Laurell electrophoresis (bottom). The top portion of the gels contains anti-LpB and the bottom portion of the gel anti-apo-a. In the first hole, the d 1.060-1.125 fraction was applied; the holes of the left plate contain samples of peak 1 (fractions 30-50) and those of the plate on the right contain samples from peak 2 (fractions 52-70).

156

G.M. Kosmer, H. Grillhofer/ Chem. Phys. Lipids 67/68 (1994) 153-159

NaBr and further centrifuged for 24 h at 160 000 × g. The floating lipoproteins contained >90% of the Lp(a) plus residual LDL and all HDL 2. As it is known that Lp(a) has a high affinity to LDL, yielding mixed Lp(a):LDL aggregates (Trieu et al., 1991), all further experiments were performed in the presence of 0.5-0.1 M Pro. This latter amino acid has been shown by Trieu et al. (1991) to interfere with aggregate formation of Lp(a). In a first step, the d 1.060-1.125 fraction was concentrated by vacuum dialysis to 20 ml and chromatographed in portions of 10 ml over a column, 100 × 2 cm, packed with Biogel A-15 M in 0.05 M Tris/0.5 g/l EDTA, 0.1 M Pro, pH 8.5 (buffer A). The eluate consisted of three peaks: first peak, Lp(a); second peak, LDL; third peak, HDL. To the Lp(a) peak we added 0.25 mol/l Pro and subjected it to chromatography in portions of 25 mg over a DEAE-cellulose column in buffer A, containing 1.5 M urea. The lipoproteins were eluted with a NaCI gradient in the range 0.0-0.2 mol/l. The eluting lipoproteins were checked for purity by double antibody rocket electrophoresis. Fractions devoid of LDL were pooled, concentrated to 5-10 mg/ml and stored for a maximum of 1 week at 4°C. Apo-a was isolated by the Heparin-Sepharose method after reductive cleavage of Lp(a) with DTT, as described earlier (Sattler et al., 1991). All chemicals were p.a. reagents from E. Merck, Darmstadt. 3. Results

Fig. 1 shows the elution pattern of the Lp(a) peak obtained from Biogel-A15 M column over DEAE cellulose and the corresponding double antibody rockets. The applied lipoproteins eluted in several peaks. Peak I consisted mainly of LDL plus some Lp(a) and peak II of pure Lp(a); the other peaks represented mixtures of Lp(a) with small amounts of contaminating serum proteins and apolipoproteins. Peak II was used for subsequent experiments.

3.1. Regulation of cholesterol biosynthesis by Lp(a) and LDL Normal as well as FH-HSF were incubated in

the presence or absence of Lp(a) or LDL with 14-C octanoate, and the cholesterol biosynthesis was measured as described in Material and methods. Fig. 2A exhibits the downregulation of cholesterol biosynthesis by LDL and Lp(a), respectively. As is known from previous work, LDL downregulated cholesterol biosynthesis in a concentrationdependent manner. At 40 #g/ml LDL-cholesterol, cholesterol biosynthesis was suppressed by almost 90%. LDL-free Lp(a) also affected cholesterol biosynthesis, yet 40/~g/ml of Lp(a)-cholesterol yielded only an approx. 35% reduction. Similar experiments were repeated using receptor-deficient fibroblasts (Fig. 2B). The particular cell line used (FH-808) lacked LDL-R completely, as determined by monoclonal antibodies (Kostner and Grillhofer, 1991). As one would ex-

2

cpm/mg CELL PROTEIN

0

0

0

~

~H

1500 1000 5OO

0

2.5 5 10 20 l REFERENCEI Lp(a) ~ LDL

1200,cpm/mg CELL 1000 800 600 400 200 0 2.5

40

PROTEIN

5

10

20

40

LIPOPROTEIN CHOLESTEROL (ug/ml) Fig. 2. (A) 14-C Octanoate incorporation into cholesterol of NHSF. Influence of Lp(a) and LDL on cholesterol biosynthesis. Normal HSF were pre-incubated for 48 h with LPDS followed by an incubation for 24 h with increasing amounts of lipoproteins. Then 14-C octanoate was added, and the incorporation of radioactivity in cholesterol was measured. The results are means ± S.D. from triplicate analysis of a representative experiment. (B) 14-C Octanoate incorporation into cholesterol of FH-HSF. Identical experiment as shown in A, but with FHFb (FH-808).

G.M. Kostner, H. Grillhofer / Chern. Phys. Lipids 67/68 (1994) 153-159

pect, the addition of LDL to the medium up to 40 /~g/ml had no influence on the incorporation of octanoate in the non-saponifiable sterol fraction. Lp(a) on the other hand caused a comparable effect as observed with normal fibroblasts. At 40 /~g/ml of Lp(a)-cholesterol, cholesterol biosynthesis was downregulated by 32%.

trol experiments. Fig. 4 shows the results. The de-. gradation of Lp(a) in the absence of LDL was very low, as expected, and could in fact be increased by a factor of 3-5 if cells were pre-incubated with 5-10 /~g/ml of LDL. Pro-treatment of cells with mevinolin increased the effect of LDL on Lp(a) degradation. We measured in addition the internalization of Lp(a) in the presence and absence of LDL and obtained comparable results, namely that Lp(a) internalization in the presence of LDL was significantly higher than in its absence (data not shown).

3.2. Degradation of Lp(a) and LDL In another set of experiments, the degradation of Lp(a) in comparison with LDL by normal as well as FH-HSF was studied. Here again LDL was degraded by normal HSF to a significant degree, whereas Lp(a) showed only a degradation of borderline significance (Fig. 3). LDL as well as Lp(a) degradation by FH-HSF was not significantly different from non-cell blanks (data not shown). Since it is known that Lp(a) binds to LDL with high affinity (Trieu et al., 1991), we tested the possibility that LDL, once bound to the LDL-R, may mediate the Lp(a) transport into the cells by a 'pick-a-back' type of mechanism. In order to test this, normal HSF were pro-incubated at 4°C with LDL and the excess of LDL was washed off, followed by incubation with Lp(a). In order to stimulate the expression of LDL-R, some cells were pro-incubated with 10 -6 M mevinolin in con-

4

157

4. Discussion

There is a long discussion going on concerning the interaction of Lp(a) with cell surfaces, in particular with the specific LDL receptor. Most if not all of the previous work was hampered by the fact that Lp(a) uncontaminated by LDL cannot be prepared by conventional procedures. Even homogeneous fractions in the ultracentrifuge or by steric exclusion column chromatography do represent complexes of Lp(a) with LDL, which may be dissociated in the electric field and demonstrated by double antibody rocket electrophoresis. Even affinity purified Lp(a) preparations using specific immunoadsorbers or Lys-Sepharose proved to be

~m LPDS

DEGRADATION:ug/mg cell protein

S MEV.

LPDS+Spg LDL m MEV.+Spg LDL 1

9

J

Z

z

8

,7,7 I-

06 44 ¢j

1

0

10

20

30

ug LIPOPROTEIN CHOLESTEROL 125-1 LDL

~

125ol Lp(a)

Fig. 3. Degradation of Lp(a) and LDL by normal HSF. The experimental procedure is described in the Methods section. LPDS pro-treated cells were incubated for 48 h with 1251labelled Lp(a) or LDL, and the TCA-soluble material was measured. The values represent means of three experiments carried out in triplicates. The coefficient of variation was < 15%. All values are corrected for non-cell blank degradation.

40

o 5

lO

Lp(a)-C (pg/ml)

Fig. 4. Influence of LDL on the degradation of Lp(a). Cells were pro-incubated for 48 h with LPDS ± 10 -6 M mevinolin, washed and further incubated for 1 h at 4°C with LPDS ± 5 pg/ml of LDL. 5 or 10/zg/ml of lzsI-labelled Lp(a) was then added, and the degradation was measured as described in Methods. The values are means ± S.D. of two experiments carried out in triplicate. All values are corrected for non-cell blank degradation.

158

G.M. Kostner, H. Grillhofer / Chem. Phys. Lipids 67/68 (1994) 153-159

contaminated by variable amounts of LDL. In order to overcome this problem, we took advantage of the observation made by Trieu et al. (1991) that Lp(a) binding to LDL may be overcome by the addition of Pro. In fact we had to develop a multi-step purification procedure in the presence of 0.1 M Pro to yield Lp(a) of satisfactory purity. The loss of Lp(a) during this procedure, however, was in the range of 60-70%. Before starting our experiments we made sure in each single case by double antibody rocket electrophoresis that Lp(a) was free of LDL. Normal HSF when pre-incubated with Lp(a) downregulated cholesterol biosynthesis by a maximum of 25% (at 40/~g/ml), which was rather low compared with LDL, which caused an almost 90°/,, reduction. At first we considered the possibility that this downregulation was caused by LDL-Rmediated binding and uptake of Lp(a). Surprisingly, LDL-R-deficient fibroblasts also downregulated cholesterol biosynthesis almost to the same degree as normal fibroblasts if pre-incubated with Lp(a), whereas LDL had no effect whatsoever. From our results we are unable to say by what mechanism the downregulation may be mediated. In previous studies we demonstrated that Lp(a) binds with high affinity to connective tissue proteins (Bihari-Varga et al., 1988), which are produced in high quantities by fibroblasts. Other investigators demonstrated that Lp(a) also has a high affinity to fibronectin (Salonen et al., 1989), tetranectin (Kluft et al., 1989) and collagen. There is a possibility that Lp(a) is bound to such substances followed by internalization and degradation. Another possibility could be that Lp(a) binds to LRP, the plasminogen receptor or any other receptor specific for proteins other than B-100. A final speculation is that there exists a cell surface receptor that might be specific for apo-a. Such possibilities are currently under investigation in our laboratory. Also noteworthy at this point is a recent paper by Williams et al. (1992), who demonstrated that heparinase-treated fibroblasts lose their ability to bind Lp(a). When non-treated cells were incubated with lipoprotein lipase from cow milk, a twofold to fivefold increase of Lp(a) binding and degradation was observed. The physiological significance of this observation remains to be demonstrated.

Assuming that Lp(a) does not bind to the LDLR to an appreciable degree, we are left with the unexplained observation that Lp(a) is increased approximately threefold in patients suffering from LDL receptor defects (Seed et al., 1990). In order to provide a feasible explanation for this observation, we conducted some experiments where Lp(a) was incubated in the presence and absence of LDL with normal HSF. In order to stimulate LDL-R expression, HSF were also pre-incubated with mevinolin. Knowing that Lp(a) binds LDL with high affinity, we argued that LDL might facilitate Lp(a) uptake by a 'pick-a-back' mechanism. Our results clearly demonstrate that this in fact is the case: LDL-free Lp(a) were only slowly degraded, whereas in the presence of LDL, Lp(a) uptake was stimulated by a factor of 3-5. It must be mentioned at this point, however, that if HSF were pre-incubated at 37°C with greater amounts of LDL, the Lp(a) degradation was reduced in a similar manner as LDL degradation (data not shown). This was caused by the fact that pre-incubation at 37°C with LDL downregulated the LDL-R number and thus also the 'pick-a-back' uptake of Lp(a). The extent to which our results may be relevant to the in vivo situation is the subject of speculation at the present time. In vivo, Lp(a) is most probably complexed to LDL to a variable degree. Whether plasma proteins interfere with this interaction cannot yet be answered. Such Lp(a)-LDL aggregates without doubt are recognized by the LDL-R and taken up and degraded by the 'Brown and Goldstein' pathway. This might be an explanation for the increased Lp(a) values in FH patients. There must exist, however, additional pathways for Lp(a) clearance, as FH patients exhibit a higher fractional catabolic rate for Lp(a) as compared with LDL (Krempler et al., 1980). Whether or not this pathway is operative on fibroblasts or mainly in specific organs, as suggested by our group at the First Lp(a) Conference in Chicago (Kostner, 1990), remains to be established.

5. Acknowledgements This work was supported by grants from the Austrian Research Foundation (grant no. S-4602) and the Kurt und Senta Herrmann Stiftung. The

G.M. Kostner, H. Grillhofer / Chem. Phys. Lipids 67/68 (1994) 153-159

expert technical assistance of Margarete Fruhmann is appreciated. 6. References Armstrong, V.W., Harrach, B., Robenek, H., Helmhold, M., Walli, A.K., and Seidel, D., 1990, Heterogeneity of human lipoprotein Lp(a): cytochemical and biochemical studies on the interaction of two Lp(a) species with the LDL receptor. J. Lipid Res. 31,429-441. Bihari-Varga, M., Gruber, E., Rotheneder, M., Zechner, R., and Kostner, G.M., 1988, Interaction of lipoprotein(a) and low density lipoprotein with glycosaminoglycans from human aorta. Arteriosclerosis 8, 851-857. Goldstein, J.L., Basu, S.K., and Brown, M.S., 1983, Receptormediated endocytosis of low-density lipoprotein in cultured cells. Methods Enzymol. 98, 241-260. Havekes, L., Vermeer, B.J., Brugman, T., and Emeis, J., 1981, Binding of Lp(a) to the low density lipoprotein receptor of human fibroblasts. FEBS Lett. 132, 169-173. Innerarity, T.L., Pitas, R.E., and Mahley R.W., 1986, Lipoprorein receptor interactions. Methods Enzymol. 129,542-565. Kluft, C., Jie, A.F.H., Los, P., DeWit, E., and Havekes, L., 1989, Functional analogy between lipoprotein(a) and plasminogen in the binding to the kringle 4 binding protein tetranectin. Biochem. Biophys. Res. Comm. 161,427-433. Kostner, G.M., 1990, The physiological role of Lp(a), in: A.M. Scanu (Ed.), Lipoprotein(a), Academic Press, New York, pp. 183-204. Kostner, G.M., and Grillhofer, G., 1991, Lipoprotein(a) mediates high affinity low density lipoprotein association to receptor negative fibroblasts. J. Biol. Chem. 266, 21287-21292. Kostner, G.M., Gavish, D., Leopold, F., Bolzano, K., Weintraub, M.S., and Breslow, J.L., 1989, HMG-CoA reductase inhibitors lower LDL but increase Lp(a) levels: a study of 24 individuals treated with Simvastatin or Lovastatin. Circulation 80, 1313-1319.

159

Krempler, F., Kostner, G.M., Bolzano, K., and Sandhofer, F., 1980, Turnover of lipoprotein Lp(a) in man. J. Clin. Invest. 65, 1483-1490. Krempler, F., Kostner, G.M., Roscher, A., and Sandhofer, F., 1983, Studies on the role of specific cell surface receptor on the removal of Lp(a) in man. J. Clin. Invest. 71, 1431-1441. Krempler, F., Kostner, G.M., Friedl, W., Paulweber, B., Bauer, H., and Sandhofer, F., 1987, Lipoprotein binding to cultured human hepatoma cells. J. Clin. Invest. 80, 401-408. McFarlane, A.S., 1985, Efficient trace labelling of proteins with iodine. Nature 182, 53-57. Salonen, E., Jauhiainen, M., Zardi, L., Vaheri, A., and Ehnholm, C., 1989, Lipoprotein(a) bind to fibronectin and has serine protease activity capable of cleaving it. EMBO J. 8, 4035-4040. Sattler, W., Kostner, G.M., Waeg, G., and Esterbauer, H., 1991, Oxidation of lipoprotein Lp(a): a comparison with low density lipoproteins. Biochim. Biophys. Acta 1081, 65-74. Seed, M., Hoppichler, F., Reavelev, D., McCarthy, S., Thompson, G.R., Boerwinkle, E., and Utermann, G., 1990, Relation of serum lipoprotein(a) concentration and apolipoprotein(a) phenotype to coronary heart disease in patients with familial hypercholesterolemia. New Engl. J. Med. 322, 1494-1499. Steyrer, E., and Kostner, G.M., 1990, The interaction of lipoprotein Lp(a) with the B/E receptor. J. Lipid Res. 31, 1247-1253. Trieu, V.N., Zioncheck, T.F., Lawn, R.M., and McConathy, W.J., 1991, Interaction of apolipoprotein(a) with apolipoprotein B-containing lipoproteins. J. Biol. Chem. 266, 5480-5485. Williams, K.J., Fless, G.M., Petrie, K.A., Snyder, M.L., Brocia, R.W., Swenson, T.L., 1992, Mechanism by which lipoprotein lipase alters cellular metabolism of lipoprotein(a), low density lipoprotein, and nascent lipoproteins. J. Biol. Chem. 267, 13284-13292.