Eur. J. Biochem. 93, 51 - 5 5 (1979)

Sialic-Acid Content of Low-Density Lipoproteins Controls Their Binding and Uptake by Cultured Cells Ivan FILIPOVIC, Giinter SCHWARZMANN, Wilfried MRAZ, Herbert WIEGANDT, and Eckhart BUDDECKE Institute of Physiological Chemistry, University of Miinster, and Institute of Physiological Chemistry, University of Marburg (Received September 28, 1978)

The (high-affinity receptor)-mediated uptake of homologous low-density (low-g) lipoproteins by cultured human arterial smooth muscle cells or human skin fibroblasts is controlled by the sialic acid content of low-g lipoprotein particles. This conclusion is derived from the following results . 1. Gangliosides incubated with native low-@lipoproteins associate with low-e lipoprotein particles. Low-Q lipoproteins modified by associated GL,,l, GGtetl,and GGtet2b Gctet3 gangliosides are internalized by arterial smooth muscle cells at a rate up to 80 % lower than native 1ow-Q lipoproteins or those preincubated with desialized gangliosides. 2. The inhibitory effect of gangliosides is specific for high affinity uptake and not detectable on skin fibroblasts deficient in low-@-lipoproteinreceptor. 3. Desialyzed 1ow-Qlipoproteins are internalized by smooth muscle cells up to 100 % faster than native 1ow-Qlipoproteins, the enhancement of uptake corresponding to the degree of desialization.

+

Homologous 1ow-Q lipoproteins are internalized by cultured human fibroblasts through an endocytotic process that requires high-affinity binding of lipoproteins to specific receptors located in coated pits on the cell surface [1,2] and a recognition site on the apoprotein component of lipoprotein, Arginine residues within apoprotein B and an arginine-rich protein have been identified as significant recognition markers of low-@lipoproteins [3]. The apoprotein component of low-g lipoproteins has been shown to contain 5-97! carbohydrates consisting of galactose, mannose, N-acetyl glucosamine and sialic acid, which form branched oligosaccharides with N-acetylneuraminic acid as the nonreducing terminus [4]. The function of the carbohydrate moiety in low-g lipoprotein is not clear at present, but the assumption that the uptake of 1ow-Qlipoproteins by cultured cells might involve the interaction of the carbohydrate moiety of 1ow-g lipoproteins with cell membranes has been excluded [5]. Abbreviatzons. Abbreviations for lipids are given according to [16-181. GLas1, G Mor~ I13NeuAc-LacCer;GCM1, GMIor 113NeuAc-GgOse4Cer; Gctet 2b, GDlb or I13(NeuAc)z-GgOse4Cer; GGret3, GTIa or IV3NeuAc,I13(NeuAc)z-GgOse4Cer. Enzyme. Neuraminidase or acylneuraminyl hydrolase (EC 3 2.1.18).

The present report demonstrates that surface associated sialic acid residues of low-@-lipoprotein particles, either present as part of the prosthetic carbohydrate group of the apoprotein or as associated exogenous gangliosides might regulate the internalization of homologous low-@lipoproteins by cultured human aortic smooth muscle cells.

MATERIAL AND METHODS Reagents and their sources included : Cow's milk lactoperoxidase, 160 U/mg (Boehringer, Mannheim), neuraminidase from Vibrio comma, 500 U/mg (Behringwerke, Marburg) dicetylphosphate, octadecyl amine, and phosphatidic acid (Serva, Heidelberg) I2'I carrier-free (Radiochemical Centre, Amersham, U.K.). Low- Density Lipoproteins

Isolation and '251-labelling of human 1ow-e lipoproteins were performed as described previously [6]. Specific activity of 1251-labelled low-g lipoproteins was in the range of 80000- 120000 counts x min-l (pg low+ lipoproteins protein)-'. Association of gangliosides by low-g lipoproteins was achieved by incubation of 10 - 20 pg 1251-labelledlow-g lipopro-

52

Sialic-Acid-Controlled Uptake of Low-Density Lipoproteins by Cultured Cells

tein with 5 - 100 pg ganglioside in a total volume of 150- 300 p1 for the specified period at room temperature. For desialization 100- 500 pg low-q lipoproteins was incubated with 25 - 100 U neuraminidase in 0.5 mI of 50 mM sodium acetate buffer pH 5.0. Incubation was stopped by adding 0.2 mM phosphate buffer pH 7.4 and rapidly cooling the mixture to 4°C. The desialyzed low-g lipoproteins were separated on a 70 x 0.8-cm Sepharose 6B column at 4 "C and concentrated by ultradialysis. Freee sialic acid and low+ lipoprotein-bound sialic acid were determined according to [7], the latter after acid hydrolysis. The total neuraminic acid content of human 1ow-Q lipoproteins used in the reported experiments was found to be in a range of 10- 17 pg/mg low-g lipoproteins.

+

a

30

20

10

.--

A

i

C

E

> o

.v)

z .-m ._ -:m: 20

0

Q

C

m

[gr

U

Gangliosides

Isolation and purification of gangliosides and their oligosaccharide components were performed as described previously [8]. For studies of ganglioside binding to 1ow-Q lipoproteins, [3H]Gctetl 1151 were used. 100-200 pg ,GL,,~was treated with 50 U of neuraminidase, separated from free sialic acid by gel chromatography and lyophilized. Cultured Cells

Human aortic smooth muscle cells, and skin fibroblasts were grown from intimal-medial explants of human aorta thoracia and from the skin of a normal adult and cultivated as described previously [6]. Human skin fibroblasts deficient in 1ow-Q lipoprotein receptor (GM361) were purchased from the Institute for Medical Research (Camden, N. J., U.S.A.). Incubation Experiments

Cells grown to confluency in 25 crn2 plastic dishes were washed once with Hank's solution after which 3 ml of fresh medium were added. The medium contained 10% lipoprotein-free human serum and 5 20 pg 1ow-e lipoproteins/ml as native low-g lipoproteins or preincubated with the specified substances or desialyzed by neuraminidase. After a 6-h incubation period the medium was removed, the cell monolayers were washed 7 times with 3 ml of Hank's solution containing 0.1 % bovine serum albumin and the surface-bound and internalized radioactivity were determined as described previously [6]. RESULTS The experiments described are based on the following points. Firstly, at low concentration of low-g lipoproteins high-affinity binding to specific receptor sites is rate-limiting in the process by which cultured

2 .a -m

10

I

m

TQ

O C

20

10

0

0

2

4 6 8 Fraction number

12

Fig. 1. Analysi.7 by ultracentrifugation of the attachment of 3Hlabelled Ccrer1 to l o w e lipoproteins. The gangliosides and ganglioside-lipoprotein mixtures were preincuhated at 37 "C for 2 h. Centrifugation was performed at 4 'C with a gradient of 10 - 25 % sucrose. The samples were layered on top of a cushion of 5 5 % sucrose. After 100 h at 15000 x g , fractions were collected from the hottoin of the tubes. (A) 'H-labelled ganglioside Gctet 1 (0.1 mM); (B) 3H-labelled ganglioside GcIet1 (0.1 mM) low-@lipoproteins (1 pM); (C) 3H-lahelled ganglioside GGtetl (0.1 mM) + l o w - ~ lipoproteins (10 pM)

+

smooth-muscle cells or fibroblasts internalize 1ow-Q lipoproteins. Therefore all low-@-lipoproteinbinding and uptake experiments were undertaken at a concentration of 10- 20 pg low-g lipoproteins/ml medium. Secondly, gangliosides incubated with native l o w - ~ lipoproteins for 1 h become associated with the lipoprotein. On incubation of low-g lipoproteins and gangliosides in a molar ratio of 1 : 10 to 1 : 100 and subsequent separation of low-g lipoproteins from the non-incorporated gangliosides by ultracentrifugation about 50 mol gangliosides per mol low-g lipoproteins appears associated with lipoprotein. A typical experiment is shown in Fig. 1.

I. Filipovic, G. Schwarzmann, W. Mraz, H. Wiegandt, and E. Buddecke

53

100 U

a, -

2 m -

80

To'O H

N C .+ a

60

.G u o c

40

::

.* .Q -

z ?3

:c n ._ E

20

z

H C

0 0

0 20 40 60 Ganglioside (nrnoi / rnl)

0

2b 40 60 Ganglioside (nrnol / m i )

Fig. 2. Inhibition of' cell surface binding ( A ) and internalization ( B ) of low-^ lipoproteins us a function of preincubation of low-g lipoproteins with dqferent gangliosides. Preincubation with gangliosides G L , , ~(0)and Gctct2(A) and Gctet2band Gctet3(0).The data were derived from a comparison of native '251-labelledlow-g lipoproteins and '251-labelled low-@lipoproteins preincubated with the indicated gangliosides for 1 h at room temperature. The cell culture experiments were performed with 10 Fg low-@lipoproteins/ml medium at 37 "C for 6 h

Low-@lipoproteins with the gangliosides G ~ t ~ t l or G L , , ~or a mixture of Gct,,2b G ~ t ~ t 3ina molar ratios up to 1:2500 are bound to and internalized by aortic smooth-muscle cells at a significantly lower rate than nontreated 1ow-Q lipoproteins. At 10 pg 1ow-Qlipoproteins/ml incubation medium the depression of binding and uptake of 1ow-e lipoproteins is concentration-dependent and reaches 70 - 80 :d inhibition at 60 - 80 nmol ganglioside/ml medium 0 ' (Fig.2). It appears that the depressive effect of the 0 15 30 45 60 gangliosides depends on their sialic acid content, the Time of preincubation (rnln) gangliosides GGlet2band Gciei3bearing two and three Fig. 3 . Inhibition of low-g lipoproteins uptake by human aortic smooth N-acetylneuraminic acid residues being more efmuscle cells as influenced by time-dependent preincubation of l o w y fective at equal molar concentration than the ganlipoproteins with GG,,,I gangliosides. 10 pg '251-labelled low-g glioside GLacl.The ganglioside effect, however, was lipoproteins and 50 pg GGtetlwere preincubated in 0.2 ml 0.15 M sodium chloride at room temperature. At indicated periods the reduced with increasing low-@-lipoproteinconcentrations and no longer detectable at 100 pg 1ow-Q preincubated mixture and 1.8 ml medium were added to the cells and incubation was continued for further 6 h at 37 'C lipoproteins/ml medium. The inhibitory effect of gangliosides on the rate of low-@-lipoproteinbinding and uptake increases with increasing preincubation time (Fig. 3). In contrast, 1ow-e lipoproteins and their ability to depress the preincubation of smooth muscle cells with gangliosides internalization of 1ow-e lipoproteins (Fig.4). has virtually no influence on the binding and interIn cultured fibroblasts from a subject with the homozygous form of familial hypercholesterinemia nalization rate of low-@-lipoproteins. the high-affinity binding and uptake of low-@lipoBoth the sialic acid component and the nonpolar moiety of gangliosides appear to be necessary reproteins is deficient [6], but an internalization of quirements for the inhibitory effect. Thus, other low-@lipoproteins still proceeds (although at a reduced charged amphipathic substances such as 0.1 mM rate) by nonspecific low-affinity processes. In contrast dicetylphosphate, phosphatidic acid and octadecylto the high-affinity uptake, however, this type of amine were ineffective. Likewise, at 0.1 mM conceninternalization is not altered when low-@lipoproteins tration, neither the desialyzed gangliosides nor neumodified by exogenous gangliosides are offered to the raminosyl-lactose, N-acetylneuraminic acid or the receptor-deficient cells instead of native low-Q lipoproteins. The data of Fig. 5 reveal that low-Q-liposialooligosaccharide component of monosialogangliotetraose caused a depression of low-@-lipoprotein protein uptake by receptor-deficient cells is depressed at most by 10 % of control values, thus indicating that binding and uptake. There is a correlation between the sialic acid the inhibitory effect of gangliosides specifically refers to the receptor-mediated uptake. content of gangliosides used for preincubation with

+

54

Sialic-Acid-Controlled Uptake of Low-Density Lipoproteins by Cultured Cells

NeuAc released f r o m GLac 1

('lo)

Fig. 4. Uptake of 1ow-e lipoproteins, preincubated with purtially or , human aortic smooth muscle cells. completely desialized G L ~ Iby Ganglioside-bound N-acetylneuraminic acid was released by neuraminidase treatment over a period of 10- 180 min, and afterwards the glycolipid was recovered by gel filtration. Other conditions as in Fig. 1

0

60

c

._ ? -

52."

-k -12 4 0 c

H

5 ~

_0 5Q :e

20 40 60 80 NeuAc released from l o w - p lipoproteins

100

(%)

Fig. 6. Uptake of neuraminidase treated '2SI-lubellecilow-g lipoproteins by humun aortic smooth muscle cells. 6.2 mg IOW-Q lipoproteins protein and 750 U neuraminidase were incubated for 3 h at 37 "C in a volume of 2 ml. 0.25 nil of the incubation mixture were withdrawn at 15, 30, 45, 60, 90, 120, 150 and 180 min for isolation of low-@lipoproteins (see Methods) and neuraminic acid analysis. Cells were incubated in the presence of 10 pg/of partkally desialyzed 1ow-e lipoproteins/ml

L

.-

0

c

: E20 .._ 131 ._ c -

L

u c

0

0

40

20 G,,,

60

80

1 (nrnol/rnl)

Fig. 5. Efyect ofgunglioside G L o c lon the uptuke of low-@lipoproteins by normal (0) and low-@-lipoprotein-receptordeficient (A) human skin fibroblasts. Cells were incubated for 6 h in the presence of 10 pg '251-labelled low-@ lipoproteins/rnl medium preincubated l I h (see Fig.2) with the indicated concentrations of G L d c for

Assuming a molecular weight of 2 500000 for the native low-g lipoproteins, one can calculate that 1 mol of low-g lipoproteins might bear 20 mol sialic acid residues [4]. Fig.6 shows that sialic acid residues can be removed by microbial neuraminidase and that the uptake rate of 1ow-Qlipoproteins increases proportionally to the increase in the degree of desialization. Desialized low-g lipoproteins (10 yg/ml medium), however, behaved as untreated low-g lipoproteins, when preincubated with 0.05 mmol gangliosides/ml at room temperature for 1 h.

DISCUSSION The chemical nature of the cellular receptor for low-g lipoproteins is unknown, but evidence has been provided that positively charged residues of lipoproteins associated with the apoprotein B or an arginine-rich protein are implicated in the binding of low-g lipoproteins to the cell surface receptor [3].

Our results indicate that modification of 1ow-Q lipoproteins by removing or introducing sialic acid residues markedly enhances or depresses binding and uptake of low-g lipoproteins by vascular smooth-muscle cells or fibroblasts. Some possible functions of carbohydrates in glycoproteins have been proposed in the past, such as the secretion from the cell [lo] or the regulation of their catabolism. Morel1 et al. [ l l ] have shown that removal of sialic acid from glycoproteins decreases their biological half-life, but similar studies using native and desialized human low-g lipoproteins labelled with I''' did not show significant differences in the respective rates of disappearance in rats [ 5 ] . However, in view of the known species specificity of receptormediated low-g-lipoproteins internalization [ 121 these results would not argue against a functional role of sialic acid in the lipoprotein catabolism. A functional incorporation of gangliosides into liposomes has been described [9]. Assuming a similar mechanism of ganglioside-lipoprotein interaction, it is suggested that the bulk of negatively charged sialic acid residues of the gangliosides are distributed on the surface of the lipoprotein, while the majority of nonpolar groups are buried in the interior. The observation that other negatively charged amphipathic molecules, such as dicetylphosphate, had no influence could be explained by the assumption that either the observed effect is specific for neuraminic acid or that the terminal negative charge has to be linked to the nonpolar part of the low-g-lipoprotein molecule by a 'spacer'.

55

I. Filipovic, G. Schwarzmann, W. Mraz, H. Wiegandt, and E. Buddecke

Lee and Alaupovic [I31 have shown that 1ow-elipoproteins are made up of a heterogeneous population with respect to apoprotein and lipid composition, and that low-g lipoprotein from hyper-. lipaemic subjects show important structural differences that appear to be under nutritional, metabolic and genetic control [13,14]. The charge heterogeneity of human low-g lipoproteins and the detection of higher isoelectric points for low-g lipoproteins from hyperlipoproteinaemia type-I1 patients [I41 *indicate that the surface charge distribution of low-@ lipoproteins might be one of the controlling factors for specific uptake and degradation of serum low-g lipoproteins. Since the surface net charge is the result of negatively and positively charged residues, it is conceivable that both the number of arginyl and sialic acid residues may inff uence the half-life and hence the ‘atherogenicity’ of serum low-g lipoproteins. This study was supported by the Deutsche Forschungsgemeinschnft (SFB 104 and Wi 299/19).

REFERENCES 1. Anderson, R. G. W., Goldstein, J. L. Sr Brown, M. S. (1977) Nature (Lond.) 270, 695-699.

2. Anderson, R. G . W., Brown, M. S. & Goldstein, J. L. (1977) Cell, 10, 164. 3. Mahley, R. W., Innerarity, T. L., Pitas, R. E., Weisgraber, K. H., Brown, J . H. Sr Gross, E. (1977) J . Biol. Chem. 25, 7279 - 7287. 4. Swaminathan, N. Sr Aladjem, F. (1976) Biochemistry, 15, 1516- 1522. 5. Swaminathan, N. Sr Aladjem, F. (1974) Fed. Proc. Fed. Am. Soc. E.xp. Biol. 33, 1585. 6. Filipovic, I. Sr Buddecke, E. (1977) Lipids, Z2, 1069-1077. 7. Warren, L. (1959) J . B i d . Chem. 234, 1971-1975. 8. Wiegandt, H. Sr Bucking, H. W. (1970) Eur. J . Biochem. I S , 289 292. 9. Ohsawa, T., Nagai, Y . Sr Wiegandt, H. (1977) Jap. J . Exp. Med. 47,221-222. 10. Eylar, E. H. (1965) J . Theor. Biol. 10, 89. 11. Morell, A. G., Gregoriadis, G., Schcinberg, I. H . , Hickmanu, J. B Ashwell, G . (1971) J . B i d . Chem. 246, 1461. 12. Goldstein, J. L. Sr Brown, M. S . (1974) J . Biol. Chem. 249, 5133 -5162. 13. Lee, D. J. Sr Alaupovic, P. (1974) Atherosclerosis, 19,501-520. 14. Ghosh, S . M., Basu, M. K. s( Schweppe, J. S. (1973) Proc. Soc. Exp. Biol. Med. 142, 1322-1325. 15. Schwarzmann, G. (2978) Biochim. Biophys. Actn, 529, 106114. 16. Wiegandt, H. (1973) Hoppe-Seyler’s Z. Physiol. Chem. 354, 1049-1056. 17. The Nomenclature of Lipids (1977) Eur. J . Biochem. 79, 1121. 18. Svennerholm, L. (1963) J . Neurochem. 10, 613. -

I. Filipovic and E. Buddecke*, Institut fur Physiologische Chemic der Westfalischen Wilhelms-Universitat Munster, WaldeyerstraDe 15, D-4400 Munster,,Federal Republic of Germany G. Schwarzmann, W. Mraz, and H. Wiegandt, Institut fur Physiologische Chemie I der Philipps-Universitat Marburg, Lahnberge, D-3550 Marburg/Lahn, Federal Republic of Germany

* To whom correspondence should be addressed.