221

Atherosclerosis, 21 (1977) 221-225 @ Elsevier/North-Holland Scientific Publishers, Ltd.

INTERCELLULAR GLYCOPROTEINS BINDING TO FIBROBLASTS

AND LOW DENSITY

LIPOPROTEIN

J. MICHAEL BOWNESS Department of Biochemistry, (Canada R3E 0W3)

University of Manitoba, Winnipeg, Manitoba

(Received 8 September, 1976) (Accepted 28 September, 1976)

Summary The interactions of low density lipoprotein (LDL) with solubilized subunits of two purified noncollagenous glycoproteins (A and G) of cartilage matrix have been studied in solution using Sepharose chromatography and at the cell surface using fibroblast monolayers. In the presence of A or G, [12’1]LDL forms aggregates of varying size. After pre-incubation with A or G, under conditions which form highly aggregated complexes, there is an increase in the binding of LDL to fibroblasts. Mixing with A under conditions which form smaller complexes produced no increase in binding to fibroblasts. It is suggested that interactions of this type may be involved in the pathogenesis of atherosclerosis. Key words:

Fibroblasts -Intercellular

matrixglycoproteins

-Low

density proteins

Introduction Low density lipoprotein (LDL), the major cholesterol-carrying lipoprotein in human plasma, has been found to bind to the cell surface of fibroblasts and then be transferred to the lysosomes in the interior of the cell. In addition to general or nonspecific processes of binding and transfer, it has been found that there is a specific binding of LDL to normal fibroblast surfaces which is lacking in familial hypercholesterolemia [ 11. A number of observations suggested that intercellular glycoproteins might influence the interactions of LDL with fibroblasts. Firstly, it is known that intercellular structural components such as acid mucopolysaccharides [ 2,3] and This work was supported by the Medical Research Council of Canada.

222

collagen [ 41 are associated with fibroblast cell surfaces and it has been suggested [ 51 that the “fibroblast cell surface glycoprotein” [6] may also be a connective tissue matrix protein. Secondly, there is a striking similarity in composition and properties between apo-LDL and certain non-collagenous structural glycoproteins [ 71; this raises the possibility that they may interact with similar cell surface sites. They resemble each other in amino-acid composition, in their insolubility and the ability of their sub-units to aggregate, and in their interaction with glycosaminoglycans. Thirdly, some diseases of lipid metabolism, such as familial hypercholsterolemia, where there is an increased concentration of circulating LDL, are associated with tendinous xanthomas [8] in which there is an abnormal accumulation of lipid and uptake of LDL in localized regions of connective tissues [9] which normally contain intercellular structural glycoproteins. Materials and Methods As part of a continuing study of intercellular matrix glycoproteins, soluble sub-units were prepared from two insoluble non-collagenous structural glycoproteins of puppy-rib cartilage [lO,ll], by suspending 1 mg of the protein per ml of 5-M guanidine hydrochloride containing either 10 or 50 n-&f dithiothreitol. After rotating the suspensions for 2 days at 25”) they were dialyzed against Krebs-Ringer-barbiturate buffer (KRB), with a pH of 7.2 and containing 6.5 n-&f barbiturate. After centrifugation of the retentate for 45 min at 57,000 X g the supernatant contained 200-300 pg of protein per ml. Evidence that these solubilized sub-units are single proteins is provided by the observation that both A and G show a single band on SDS polyacrylamide gel electrophoresis [lo], with evidence of aggregation in both SDS [lo] and KRB (Fig. 1). LDL and lipoprotein deficient serum (LPDS) were prepared by KBr floatation

TABLE

1

INFLUEN,?? ING

OF

OF

[

Materials

SOLUBILIZED

II LDL

TO

INTERCELLULAR

HUMAN

GLYCOPROTEINS

added

Bound of

[ 12511LDL + disaggregated

1 Iz511

+ aggregated

LDL

A (129 A (110

pg)

label

pg)

1.52(l)

a a

[ ’ 2511 A (12

1.95(l)

fig; aggregated)

(1Opg)

[ 12511LDL

0.15(3) + aggregated

[ 12511

+ G (156

[ ‘2511G

P = 0.03

A (110

pg)

pg)

added

Means

(with

with

fibroblasts

Dishes

without

fibroblasts

0.54(2)

aggregated)

protein for

THE

0.20(2) 0.31(l)

(1Owg)

50 pg LDL a

pg:

ON

1.67(l)

1 ’ 2511 LDL LDL

as % age of total

a Dishes

0.87(3)

[ 1 2511 A (12

G)

0.23(2)

[ 12511LDL+G(156~g) [ 12511G

AND

observations)

0.32(4)

r ’ * ‘11 LDL

(A

0.76(l) (8-10

a t-test

X lo5

comparing

BIND-

FIBROBLASTS

cpm)

were

fibroblasts

mixed with

with

the stated

and without

amounts

aggregated

of the

other

glycoproteins.

materials.

number

223

from pooled human sera [l]. The proteins were labeled with “‘1 using iodine monochloride and the fibroblasts were grown and treated exactly as described by Goldstein and Brown [l] except that McCoys medium 5A [12] was used for the growth of the cells and for the cell binding experiments. The [‘*‘I] LDL and the solubilized A and G were dialyzed against this medium and were then mixed, using the quantities stated in Table 1 in a total of 0.7 ml McCoys medium, and allowed to stand 24 h at 4°C. After mixing with 1.3 ml of 7.5% LPDS serum in McCoys medium the whole volume was added to the fibroblasts in a single dish on Day 7 (of growth after seeding) after removal of the previous medium. The fibroblast dishes were then incubated for 2 h in air at 37°C in a shaker bath, the medium was removed, the cells washed repeatedly and the bound label was determined as described by Goldstein and Brown [ 11. Results and Discussion When freshly prepared, using 50 mM dithiothreitol, solutions of glycoprotein A in KRB show a single peak on Sepharose 2 B (K,, = 0.9) with very little material of high molecular weight. After standing 3 days at 4°C in KRB a small amount of aggregate is formed (curve A, in Fig. 1) but the main peak is still at 0.9 K,,. If A is solubilized using 10 mM dithiothreitol and then kept in 0.5 M guanidine hydrochloride before dialysis against KRB a prominent peak of highly aggregated material (K,, < 0.1) is present (curve A2 in Fig. 1). Glycoprotein G, even when solubilized using 5 M guanidine hydrochloride and 50 mM dithiothreitol, always shows considerable aggregation after dialysis against KRB (curve G1 in Fig. 1). LDL alone gives a single peak with K,, of 0.75 on Sepharose 2B, (LDLI in Fig. l), but when a mixture of [l”I]LDL with disaggregated A was allowed to stand for 2-3 days at 4’C a considerable portion of the label was eluted earlier (K,, < 0.7) and a new peak was formed at K,, 0.63 (curve LDL2). The region of the elution pattern in which this shift occurs corresponds with the region of a shift in [‘*‘I] A pattern (K,, of 0.2-0.7) caused by LDL (curve A3 in Fig. 1) and appears to be due to a fairly stable intermediate complex or aggregation state which involves both LDL and A in its formation. When a mixture of [l*‘I]LDL with aggregated glycoprotein G was allowed to stand for 24 h at 4°C a small highly aggregated peak is formed (curve G2 in Fig. 1). A similar peak (K,, < 0.2) was formed from [12’I]LDL mixed with aggregated A. The results in Table 1 show that both A and G solutions which contain the highly aggregated peak (K,, < 0.2) are capable of increasing the proportion of labelled LDL which binds to fibroblast monolayers in culture. They also increase the surface binding of LDL to the dishes without cells, but this effect is not as great as the effect on binding to the cells. A and G themselves bind strongly to the fibroblasts and less strongly to the dishes in which the cells are cultured. The observation that disaggregated A actually produced a slight decrease in LDL binding to fibroblasts indicates that the formation of large aggregates (K,, < 0.2 on Sepharose 2B) is necessary for the increase in binding and suggests that smaller complexes (K,, 0.2-0.7 in Fig. 1) may interfere with LDL binding. While the complexes observed in the present work are soluble, it is possible

224

LDL 2000 -

Fig. 1. Chromatography on Sepharose 2B of LDL and A or G in KRB buffer. 1 ml solution. containing 1 mg bovine serum albumin plus the stated amounts of the proteins, was applied to a 25 X 110 mm column. except for A2 which was 9 X 300 mm. A1 ( -) = disaggregated [12511A (20 pg); A2 (- - - - - -) = aggregated [12511A (20 pg); A3 (* . . . . *) = disaggregated [‘2511A (20 /a) + LDL (500 pg) for 3 days; = [‘2511G (20 u(p); G2 (-----) = [‘2511LDL (13 c(g) + G (130 pg) for 1 day; LDLl GI (-_) ) = [12511LDL (10 /.tg): LDL2 (------) = [ ‘*‘IlLDL (10 fig) + A (120 pg) for 3 days. Mixing of (---LDL with A or G was done at 4’C at the stated times before application to the column.

that they may be related, as precursors, to the insoluble complexes containing protein, glycosaminoglycan and lipid which are formed by the interaction of an arterial extract with LDL [ 131. This lipoprotein complexing factor formed more insoluble complex with sera from ischemic heart disease patients than from controls. Since the glycoproteins A and G are associated with glycosaminoglycan in cartilage matrix [lo] it is also possible that there may be some relationship between the complexes observed in the present work and the lipoprotein-glycosaminoglycan complexes isolated from human atherosclerotic lesions [ 141. While these results demonstrate that an interaction can occur between intercellular matrix glycoproteins and LDL, both in solution and at the fibroblast cell surface, further work is required to compare the effects of matrix glycoproteins from different sources, to show whether the metabolism of the cells and of the two intercellular components is affected by the interaction, and to study the nature of the interaction. Since it is known that an abnormal retention of LDL in the connective tissue of the arterial wall is a common early event in atherosclerosis [ 151 it seems possible that an “anchorage modulation”

225

[16], involving an interaction between LDL and connective tissue glycoproteins at the cell surface, may be involved in the pathogenesis of the tissue alterations in this condition. Acknowledgements I wish to thank Dr. John Hamerton for the fibroblasts, the sera and Alan Tarr for excellent technical assistance.

Dr. Paul Desjardins for

References J.L. and Brown, M.S., Binding and degradation of low density lipoproteins by cultured 1 Goldstein, human fibroblasts, J. Biol. Chem., 249 (1974) 5153-5162. E. and Fromme, H.G.. Metabolism of sulfated glycosamino2 Kresse, H.. Figura, K. van, Buddecke, glycans in cultivated bovine arterial cells, Hoppe-Seyler’s Z. Physiol. Chem., 356 (1975) 929-941. of mucopolysaccharides by normal and 3 Satoh, C., Duff, R., Rapp, F. and Davidson, E.A., Production transformed cells, Proc. Nat. Acad. Sci. (U.S.A.), 70 (1973) 54-56. characterization of the J.R., Bauer. E.A., Hoyt, R. and Wedner, H.J., Immunologic 4 Lichtenstein, membrane bound collagen in normal human fibroblasts - Identification of a distinct membrane collagen, J. Exp. Med., 144 (1976) 145-154. subunit of elastic fiber microfibrils secreted 5 Muir, L.W., Bornstein, P. and Ross, R.. A presumptive by arterial smooth muscle cells in culture. Europ. J. Biochem., 64 (1976) 105-114. from fibroblasts. Proc. 6 Yamada, K.M. and Weston, J.A., Isolation of a major cell surface glycoprotein Nat. Acad. Sci. (U.S.A.), 71 (1974) 3492-3496. versus structural glycoproteins - Atherosclerosis as an analogue and comI Bowness, J.M., Lipoproteins petitor of normal connective tissue interactions, Med. Hypotheses, 2 (1976) Sept. and hyperlipoproteinemia -A review, Dermatologica, 149 (1974) l-9. 8 Polano, M.K., Zanthomatosis C.C.. Low density lipoprotein accumulation in actively growing xantho9 Scott, P.J. and Winterbourn, mas. J. Atheroscler. Res.. 7 (1967) 207-223. cartilage glycoproteins with aggregating sub10 Shipp, D.W. and Bowness, J.M., Insoluble noncollagenous units, Biochim. Biophys. Acta. 379 (1975) 282-294. 11 Joseph, E.A., Shipp, D.W. and Bowness, J.M., Manuscript in preparation. T.A., Maxwell, M. and Krusc, P.F., Amino acid requirements of the Novikoff hepatoma in 12 McCoy, vitro, Proc. Sot. EXP. Biol. Med., 100 (1959) 115-118. of aortir proteins in the formation 13 Camejo, G., Lopez, A., Vegas, H. and Paoli, H., The participation of complexes between low density lipoproteins and intima-media extracts. Atherosclerosis, 21 (1975) 77-91. B., Pargaorkar, P.S. and Berenson, G.S., Lipopro14 Srinivasan, S.R., Dolan, P., Radhakrishramurthy, tein-acid mucopolysaccharide complexes of human atherosclerotic lesions, Biochim. Biophys. Acta, 388 (1975) 58-70. 15 Smith, E.B., The relationship between plasma and tissue lipids in human atherosclerosis, Advanc. Lipid, Res., 12 (1974) l-49. G.M., Surface modulation in cell recognition and cell growth, Science. 192 (19’76) 21816 Edelman, 226.