Thrombospondin Secretion by Cultured Human ... - Europe PMC

American Journal of Pathology, Vol. 129, No. 2, November 1987 Copyright © American Association of Pathologists

Thrombospondin Secretion by Cultured Human Glomerular Mesangial Cells GREGORY J. RAUGI, MD, PhD, and DAVID H. LOVETT, MD

From the Medical Service, Veterans Administration Medical Center, Seattle, Washington, and the Divisions ofDermatology and Nephrology, Department ofMedicine, University of

Washington, Seattle, Washington

Thrombospondin is a high-molecular-weight glycoprotein constituent of extracellular matrices of several cells in culture. Immunoreactive thrombospondin is also present within the normal human renal mesangium. To determine whether this thrombospondin could be a synthetic product ofthe intrinsic glomerular mesangial cells, we examined cultured human glomerular mesangial cells for the ability to synthesize and secrete thrombospondin. Well-characterized human mesangial cells were found to synthesize and secrete thrombospondin, as determined by specific immuno-

staining at the light- and electron-microscopic levels. Furthermore, metabolically labeled thrombospondin was immunoprecipitated from the conditioned medium of cultured cells. These studies suggest that the thrombospondin present within the normal mesangium is of intrinsic glomerular cell origin. Mesangial thrombospondin may be an important mediator ofcellular function, particularly in disease states characterized by intrinsic glomerular cell proliferation. (Am J Pathol 1987, 129:364-372)

THROMBOSPONDIN, a high-molecular-weight glycoprotein first described as a normal constituent of the alpha granules of platelets," 2 recently has been found to be a secretory product and constituent ofthe extracellular matrix of several mesenchymal cells in culture (for review, see Lawler3). In close analogy to fibronectin, thrombospondin is a multifunctional macromolecule composed of domains. Thrombospondin has been shown to interact with a large number of plasma, cellular, and extracellular matrix elements. Heparin,4 Type V collagen,5 and fibrinogen6-8 interact with thrombospondin by binding to various specific domains ofthe molecule. Laminin,5 fibronectin,9 von Willebrand factor,5 histidine-rich glycoprotein,'0 and plasminogen" also demonstrate specific affinities for thrombospondin. Additionally, thrombospondin binds to cellular membranes, an event which in platelets appears to involve the specific sulfated membrane glycolipid galactosylceramide-31sulfate (sulfatide).'2 The rate of thrombospondin secretion by cultured endothelial cells and lung fibroblasts is dependent upon the degree of cellular confluency and the proliferative status of the cells. Actively proliferating cells

secrete more thrombospondin per cell than do quiescent cells in confluent cultures.'3 Thrombospondin secretion by growth-factor-deficient rat aortic smooth muscle cells cultures is stimulated by platelet-derived growth factor (PDGF). 14 Heparin and related polyanions that inhibit proliferation of rat vascular smooth muscle cells'5l'7 also inhibit incorporation of newly synthesized thrombospondin into the extracellular matrix.'4 It has therefore been suggested that matrixassociated thrombospondin plays a role in regulating the proliferative response of aortic smooth muscle cells to injury.'4 Wight et al'8 observed focal deposits of immunoreactive thrombospondin in the mesangium of normal human glomeruli as well as in a discontinuous pattern at the base of renal tubules and in Bowman's capsule.

364

Supported by Veterans Administration General Medical Research Funds and by a grant from the Northwest Kidney Foundation. Accepted for publication June 24, 1987. Address reprint requests to Gregory J. Raugi, MD, PhD, Medical Service, Veterans Administration Medical Center, 1660 Columbian Way South, Seattle, WA 98108.

Vol. 129 * No. 2

They suggested that the thrombospondin found in these locations may not be entirely the result of platelet activation and release. Given that other mesenchymal cell types can secrete thrombospondin, it was of interest to determine whether the intrinsic glomerular mesangial cell population also possessed this capability. Mesangial cell proliferation is a prominent feature of many forms of glomerular disease, and the demonstration of a relationship between cell proliferation and thrombospondin secretion, as previously found with vascular smooth muscle cells, would be a finding of potential pathophysiologic significance. It was therefore elected to determine whether cultures of proliferating human mesangial cells were capable of thrombospondin synthesis and secretion.

Materials and Methods Establishment of Human Mesangial Cell Cultures Human kidney was obtained at the time of elective nephrectomy from patients with various forms of well-circumscribed renal neoplasia. Nonneoplastic tissue (confirmed by frozen section) was obtained by sharp dissection and homogenized. The glomeruli were recovered by passage through serially graded sieves, as has been reported in the isolation of rat glomeruli.19 Tubular contamination was less than 1%, as determined by phase-contrast microscopy. The glomeruli were washed twice in phosphate-buffered saline (PBS) and incubated with bacterial collagenase (Type IV, Sigma Chemical Co., St. Louis, Mo) using concentrations and conditions precisely as previously reported in detail for rat glomeruli.'9 The resultant glomerular remnants were washed twice in PBS and plated in growth medium at 37 C in a humidified 5% CO2 atmosphere. The growth medium consisted of RPMI 1640 supplemented with 20% heat-inactivated fetal calf serum (Gibco, Grand Island, NY), 50 U/ml penicillin, 50,ug/ml streptomycin, 300 ,ug/ml L-glutamine, 5 ,ug/ml transferrin, 5 ,ug/ml bovine insulin, and 5 ng/ml selenous acid (ITS, Collaborative Research, Lexington, Mass). Cellular outgrowths appeared within 3-5 days and consisted of straplike cells growing in interwoven bundles. Clusters of cells with typical mesangial morphology were picked out, replated, and grown to confluency. The cells were used between the sixth and eighth passages for this study. Characterization of Mesangial Cells Confluent cells grew in monolayers, and there was no morphologic evidence for the presence of either glomerular endothelial or epithelial cells, which grow

MESANGIAL CELLS SECRETE THROMBOSPONDIN

365

as monolayers of polygonal cells.20 Cells with either the functional or morphologic features of macrophages or monocytes were not observed. Further characterization of mesangial cells was accomplished by specific immunostaining for Factor VIII antigen, common leukocyte antigen, and myosin antigen. For conventional light-microscopic evaluation, cells were grown on glass coverslips to subconfluency. After three washes in cold PBS, the cells were fixed for 20 minutes at room temperature with 3.7% paraformaldehyde in PBS supplemented with 0.1 mM CaCl2 and 1 mM MgCl2 and stained with the Richardson stain.

Growth and Labeling of Cells After the cells had reached subconfluent density, the growth medium was removed, the cells were washed once in sterile PBS, and the medium was replaced with methionine-deficient, serum-free Dulbecco's modified Eagle's medium (DMEM) supplemented with ,B-aminopropionitrile (64 mg/l), ascorbic acid (50 mg/l), penicillin (10,000 U/l), and streptomycin (100 mg/l). After 1 hour, this medium was removed and replaced by identical medium supplemented with 35S-methionine (50 ,uCi/ml) for 18-24 hours. The medium was harvested, and 0.1 volume of a cocktail of protease inhibitors was added. The inhibitors and their final concentrations were as follows: ethylenedinitrilotetraacetic acid (EDTA), 2.5 mM; N-ethyl-maleimide (NEM), 10 mM; and phenylmethanesulfonic acid (PMSF), 0.2 mM. The medium was stored at -20 C until used.

Preparation of Human Platelet Thrombospondin Platelet thrombospondin was prepared from human platelets obtained from the Puget Sound Blood Center according to previously published methods, but with the following modifications. All buffers after the initial platelet washing steps contained 1 mM CaCl2. Washed platelets were activated by addition of A23 187 (Calbiochem) to a final concentration of 10 ,uM. The aggregated platelets were removed by centrifugation; PMSF and NEM in final concentrations of 2 and 10 mM, respectively, were added to the supernatant fraction to retard proteolysis, before loading on a column of Sepharose CL-4B (Pharmacia, Piscataway, NJ). The fractions containing thrombospondin were pooled and further purified by affinity chromatography on heparin-Sepharose 6B. Protein-containing fractions eluted with 0.6 M NaCl were pooled and checked for purity by sodium

dodecyl sulfate-polyacrylamide gel electrophoresis

366

RAUGI AND LOVETN

(SDS-PAGE). Protein was measured by the Lowry method2' as modified by Loomis and Stahl.22 Preparation and Characterization of Antisera Rabbits (R&R Rabbitry, Bellingham, Wash) were given weekly injections of thrombospondin suspended in Freund's complete adjuvant (first injection) or incomplete adjuvant (subsequent injections). The IgG fraction of pooled sera from several bleedings was prepared by repeated ammonium sulfate precipitation. Contaminating activity against fibronectin and fibrinogen was removed by affinity chromatography on columns of Sepharose 4B covalently linked to human fibronectin or fibrinogen. The purified antiserum was tested using an ELISA method as described elsewhere23 and was found to have no detectable activity against fibronectin, fibrinogen, von Willebrand factor, bovine serum albumin, or collagen Types I, III, IV, and V. Western blotting of the purified antiserum against a crude supernatant fraction of stimulated platelets revealed immunostaining of a single band that comigrated with an authentic thrombospondin standard (data not shown). Anti-myosin was the generous gift of Dr. Groschel-Stewart (Darmstadt, FRG).

Immunofluorescence Microscopy Cells were grown on glass slides (Costar). All steps were carried out at 4 C unless otherwise noted. Cells were washed two times with PBS containing 0.1% bovine serum albumin (BSA) and fixed with 3% paraformaldehyde for 1 hour. After washing three times in PBS-BSA, the cells were permeabilized by dipping in methanol at -60 C for 15 seconds. After two more washes in PBS-BSA, cells were incubated with preabsorbed anti-thrombospondin IgG at a dilution of 1 200 (25.4 ,ug/ml) in PBS-BSA for 45 minutes. The cells were washed with PBS-BSA three times for 5 minutes and incubated with fluorescein isothiocyanate (FITC)-conjugated goat anti-rabbit IgG (Cappel, Cochraneville, Pa) at a dilution of 1: 200 in PBS-BSA for an additional 45 minutes. After three more 5-minute washes in PBS, the cells were mounted with aqueous Immu-mount and examined using a Leitz phase microscope equipped with epi-illumination. Controls run at the same time included preabsorbed nonimmune IgG at the same IgG concentration as anti-thrombospondin, and FITC-conjugated goat anti-rabbit IgG in the absence of a first antibody. In addition, anti-thrombospondin antiserum (25.4 ,ug/ml) was incubated with purified thrombospondin (500,ug/ml) for 72 hours before use in immunostaining experiments.

AJP * November 1987

Immunoperoxidase Staining Human mesangial cells were cultivated on glass coverslips. Fixation was performed with paraformaldehyde as described above. After fixation and three washes with PBS supplemented with 0. 1% BSA (PBSBSA), the cells were permeabilized by treatment with 0.01% Triton X- 100 for 90 seconds and washed three more times. The fixed, permeabilized cells were incubated with the primary antibody (anti-thrombospondin or anti-myosin) overnight at 1:100 dilution in PBS-BSA at 4 C. After three washes, the cells were incubated overnight with peroxidase-conjugated goat anti-rabbit IgG (Cappel) at 1:100 dilution in PBSBSA at room temperature. After washing three times in Tris-buffered saline (TBS, Tris-HCl [pH 7.6, 50 mM] and NaCl [0.9% wt/vol]), the cells were postfixed with glutaraldehyde in 0.1 M sodium cacodylate, pH 7.6, containing 5% sucrose for 15 minutes at room temperature, washed twice with TBS and once with Tris buffer (50 mM Tris-HCl, pH 7.6). The cells were then incubated with peroxidase substrate (30 mg 3,3'diaminobenzidine tetrahydrochloride [DAB] in 50 ml Tris buffer supplemented with 0.01 ml H202 and 0.25 ml 8% NiCl2 for 10-12 minutes and washed twice with Tris buffer and twice with water. The cells were counterstained with Mayer's hematoxylin, washed, dehydrated, and mounted. Controls included peroxidase-conjugated goat anti-rabbit IgG used in the absence of a primary antibody.

Radioimmunie Precipitation Radioimmune precipitation was carried out using medium conditioned by human mesangial cells for 18-24 hours in the presence of 35S-methionine. Before radioimmune precipitation was carried out, the medium was mixed with a small volume of gelatinSepharose 4B for removal of most of the fibronectin, which, by forming a complex with thrombospondin, interfered with the precipitation reaction. Rabbit anti-thrombospondin IgG, lO[ul, was added to 0.5 ml fibronectin-depleted metabolically labeled conditioned medium and mixed for 2 hours at 20 C. Goat anti-rabbit IgG (Sigma, St. Louis, Mo), 25 ,ul, was then added and incubated an additional 2 hours at 20 C. The immunoprecipitate was collected by centrifugation and washed three times in a solution containing Tris-HCl (pH 7.5, 50 mM), NaCl (150 mM), SDS (0.1% wt/vol), Tween-20 (0.05% vol/vol), and NaN3 (0.02% wt/vol) and once in the same solution without detergents. The washed precipitate was solubilized in 10 ,u1 0.05 N HCI; 30 ,ul electrophoresis sample buffer24 containing 10% (vol/vol) ,8-mercaptoethanol was added, and the resulting solution was neutralized

Vol. 129

*

No. 2

with 10 u1 of 0.05 N NaOH, heated at 95 C for 5 minutes, and subjected to electrophoresis on 7.5% polyacrylamide gels. Labeled proteins were visualized by autoradiography on Kodak X-Omat film. Immunoelectron Microscopy HMC were grown in plastic tissue culture flasks to subconfluence. All steps were carried out at 4 C unless otherwise stated. Cells were washed three times with TBS and fixed in the flasks with 3% glutaraldehyde for 45 minutes. Aldehydes were quenched by rinsing three times (one minute each) with TBS supplemented with 50 mM glycine. Cells were permeabilized by exposure to absolute ethanol at -60 C for 15 seconds, washed three times in TBS with 0.1% BSA (TBS-BSA), and exposed to preabsorbed rabbit antithrombospondin IgG at a dilution of 1: 200 (25.4 ,g/ml) in TBS-BSA overnight. After three washes with TBS-BSA, cells were exposed to peroxidase-conjugated goat anti-rabbit IgG (Cappel, Cochraneville, Pa) at a dilution of 1: 100 in TBS-BSA for 2 hours at 20 C. After washing three times with TBS-BSA, the cells were fixed for 15 minutes with 1.5% glutaraldehyde in 0. 1 M sodium cacodylate buffer, washed three times with TBS, and incubated for 30 minutes with 0.1% DAB in 50 mM Tris-HCl and 0.1 ml of 1% H202 at 20 C. The cell layers were postfixed in 2% osmium 20 C. The cell layers were postfixed in 2% osmium tetroxide in cacodylate buffer for 30 minutes. After washing, dehydration was accomplished in a graded series of ethanol. The cultures were infiltrated with Epon embedding mixture, removed by gentle scrap-

Figure 1-Light micrograph of human mesangial cells. Cells were grown on glass disks in complete medium until subconfluent, fixed, and stained with toluidine blue. Mesangial cells grow in interweaving bundles. (X350)


367

ing, and pelleted by centrifugation. After polymerization, thin sections were cut with diamond knives and mounted on copper grids. The sections were stained with 2% aqueous uranyl acetate and lead citrate and examined with a JEOL electron microscope. Controls included cells incubated with peroxidase-conjugated second antibody alone as well as incubations with DAB/H202 alone.

Results The human mesangial cells used in these studies demonstrated in a uniform manner the phenotypic characteristics of this cell type. As is the case with cultured rat mesangial cells, the human cells closely resemble vascular smooth muscle cells. As seen in Figure 1, subconfluent cultures consisted of predominantly straplike cells growing in interwoven bundles. Figure 2 shows the results of immunostaining by the peroxidase method with a rabbit antibody to myosin. Reaction product prominently decorates longitudinally oriented cytoplasmic filaments characteristic of the intrinsic mesangial cell.'9'20 Immunofluorescence microscopic studies of these cells with rabbit anti-human thrombospondin IgG revealed specific cytoplasmic staining. Figure 3 shows a representative field demonstrating prominent intracellular and perinuclear immunostaining in a granular pattern. In order to assess the specificity of immunostaining, simultaneous experiments were performed using anti-thrombospondin antibody preincubated with thrombospondin antigen. Under conditions of antigen excess, immunostaining was al-

368

RAUGI AND LOVETA

AJP

o

November 1987 v

'Vw

e;I

.9

'Z OY

'4i

Odo

.if .14

Figure 2-Immunoperoxidase staining of human mesangial cells with rabbit anti-myosin IgG. Cells were grown on glass coverslips until subconfluent. After fixation and permeabilization, the cells were stained with specific antibody followed by peroxidase-conjugated goat anti-rabbit IgG and processed as outlined in Materials and Methods. Specific immunostaining of longitudinally arranged cytoplasmic filaments compatible with myosin is demonstrated. (Xl 000)

Figure 3-Immunofluorescence staining of human mesangial cells with rabbit antithrombospondin IgG. Cells were grown on glass coverslips until subconfluent. After fixation and permeabilization the cells were stained with specific antibody followed by FITC-conjugated goat anti-rabbit IgG. Specific staining compatible with both membrane and intracellular localization of thrombospondin antigen is demonstrated. (Xl 30)

Vol. 129 * No. 2

most completely abolished. There was no immunostaining using nonimmune IgG in the same concentration. The impressions obtained by the immunofluorescence studies were confirmed by immunoperoxidase staining. Figure 4 shows the results of immunostaining with anti-thrombospondin. Reaction product decorates the perinuclear cytoplasm in a granular pattern clearly distinct from the pattern observed with anti-myosin antibody, as shown in Figure 2. These results were confirmed and extended by experiments using immunoelectron microscopy. Figure 5a shows the membrane localization of thrombospondin antigen. There is an even distribution ofreaction product along the plasma membranes of the cells as well as along the membranes of the prominent filopodial processes, which are seen in both longitudinal section and cross-section in this figure. The intracellular thrombospondin antigen was localized to the abundant rough endoplasmic reticulum present in these cells (Figure 5b). Controls employing either second


369

antibody alone (Figure 5c) or DAB/peroxide alone (data not shown) revealed no detectable surface or intracellular deposition of reaction product, demonstrating the specificity of the findings with the rabbit anti-thrombospondin antibody. In order to provide independent confirmation that thrombospondin was indeed a synthetic product of human mesangial cells, experiments were performed for immunoprecipitating metabolically labeled thrombospondin from conditioned culture medium. As seen in Figure 6 (Lane 1) mesangial cells secrete a large number of higher molecular weight proteins, including fibronectin. After substantive removal of the fibronectin with gelatin-Sepharose (Lane 2), immune precipitation with specific antibody to thrombospondin resulted in recovery of a single labeled protein with a molecular weight of 190 kd after reduction with,-mercaptoethanol (Lane 3). The radiolabeled protein comigrated with purified platelet thrombospondin on Coomassie blue-stained gels, confirming its identity as thrombospondin protein. In

.1

N

'I

. ." ..

.s.1.. r

..

V,.

t.

.e . I

.1

-.0

PI

I 0

Figure 4-Immunoperoxidase staining of human mesangial cells with rabbit anti-thrombospondin IgG. Cells were grown and processed as described in the legend to Figure 2. Specific immunostaining of thrombospondin antigen is demonstrated in the perinuclear cytoplasm and is clearly distinct from the pattem observed with anti-myosin antibody, as shown in Figure 2. (Xl 000)

370

RAUGI AND LOVETNv

AJP * November 1987

1

2

3 4

A -'.-TSP

B

-43 Figure 6-Biosynthetic labeling and immunoprecipitation of thrombospondin. Human mesangial cells were biosynthetically labeled with 35S-methionine, and the labeled proteins were analyzed on 7.5% polyacrylamide gels after reduction with 8-mercaptoethanol. Lane 1, synthetic profile of secreted proteins. Lane 2, synthetic profile of secreted proteins following absorption of fibronectin on gelatin-Sepharose. Lane 3, precipitated, biosynthetically labeled proteins after treatment with specific anti-thrombospondin IgG. Lane 4, precipitated, biosynthetically labeled, proteins after treatment with nonimmune IgG. The arrow indicates the position of migration of purifed thrombospondin standard.

C

tiserum was used, no radiolabeled proteins were recovered (Lane 4).

Figure 5-Immunoelectron microscopy.

A-The membrane location of thrombospondin antigen is detailed by the arrows showing granular reaction product on the cell surface and on the filopodial processes. B-The localization of thrombospondin antigen in the prominent rough endoplasmic reticulum. C-A representative field when anti-thrombospondin antibody was omitted. (A, X35,000; B, x20,OOO; C, X19,000).

some experiments an additional faint band with an apparent molecular weight of 160 kd after reduction

with,8-mercaptoethanol was seen. The latter probably represents a previously described degradation product of native thrombospondin. When nonspecific an-

Discussion Thrombospondin is a glycoprotein of Mr 450,000 consisting of three similar, if not identical, chains. The chains are linked by disulfide bonds near the carboxyl end (reviewed by Lawler3). Thrombospondin was first described in platelets and later shown to be a component of the a-granules.226 It is released after activation of platelets and in the presence ofCa2+ binds to their surfaces.25 Platelets adhere to fibronectin-coated substrates via thrombospondin,9 but the role of thrombospondin in platelet binding to the subendothelial matrix under conditions of flow has

Vol. 129 * No. 2

been questioned.27 It does, however, appear to act as the "endogenous platelet lectin" and participates in the process of platelet aggregation.28'29 Thrombospondin is also secreted by cells in culture; some secreted thrombospondin binds specifically to components of the extracellular matrix.' Immunologically detectable thrombospondin was previously found to be present within normal human kidney, with a glomerular pattern characteristic of a mesangial localization.'8 Although this pattern of staining was believed to represent synthesis and matrix deposition by the intrinsic mesangial cell population, the contribution of platelets as the source of the thrombospondin could not be excluded with these techniques. These experiments were undertaken in order to clarify this issue. Carefully characterized populations of human mesangial cells were used for these studies. Detailed morphologic characterization of the fine structure and biochemical analysis of the secretory products of these cells are the subject of separate reports (in preparation). Conventional immunofluorescence microscopic studies of cultured mesangial cells using specific anti-thrombospondin rabbit IgG demonstrated both surface and intracellular localization of antigen. The localization of antigen by immunoelectron microscopy to the rough endoplasmic reticulum is consistent with ongoing biosynthesis, an impression confirmed by the studies using biosynthetic labeling and immunoprecipitation of labeled thrombospondin. In cultures of fibroblasts and endothelial cells, net thrombospondin secretion per cell is greater in proliferating as compared with quiescent cultures.'3 Majack, et all4 have shown that thrombospondin secretion by growth-arrested rat aortic smooth muscle cells is stimulated by platelet-derived growth factor. Heparin and related polyanions are inhibitors of growth for smooth muscle cells.'5'-7 Heparin blocks the mitogenic response of vascular smooth muscle cells to PDGF but does not affect the amount of thrombospondin secreted.'4 Instead, heparin shifts the distribution of thrombospondin from the extracellular matrix to the medium. It has been suggested that matrix-associated thrombospondin acts as an "integrator" for the opposing actions of PDGF and heparin-like glycosaminoglycans in regulating cell proliferation. Mesangial cells phenotypically resemble vascular smooth muscle cells in a number of respects, including responsiveness to PDGF and inhibition of growth by heparin and related glycosaminoglycans.30 Thus, it is possible that mesangial cell secretion may be subject to similar regulatory factors. The relationship of thrombospondin synthesis and subsequent incorporation into the ma-


371

trix to the proliferative status of the cell may be of potential significance in the analysis of the pathogenesis of glomerular diseases characterized by cellular proliferation and matrix expansion.

References 1. Baenziger NL, Brodie GN, Majerus PW: A thrombinsensitive protein of human platelet membranes. Proc Natl Acad Sci (USA) 1971, 68:240-243 2. Baenziger NL, Brodie GN, Majerus PW: Isolation and properties of a thrombin-sensitive protein of human platelets. J Biol Chem 1972, 247:2723-2731 3. Lawler J: The structural and functional properties of thrombospondin. Blood 1986, 67:1197-1209 4. Lawler JW, Slayter HS, Coligan JF: Isolation and characterization of a high molecular weight glycoprotein from human blood platelets. J Biol Chem 1978,

253:8609-8616

5. Mumby SM, Raugi GJ, Bornstein P: Interactions of thrombospondin with extracellular matrix proteins: Selective binding to type V collagen. J Cell Biol 1984, 98:646-652 6. Leung LLK, Nachman RL: Complex formation of platelet thrombospondin with fibrinogen. J Clin Invest 1982, 70:542-549 7. Tuszynski GP, Srivastava S, Switalska SI, Holt JC, Cierniewski CS, Niewiarowski S: The interaction of human platelet thrombospondin with fibrinogen. J Biol Chem 1985, 260:12240-12245 8. Bale MD, Westrick LG, Mosher DF: Incorporation of thrombospondin into fibrin clots. J Biol Chem 1985, 260:7502-7508 9. Lahav J, Schwartz MA, Hynes RO: Analysis of platelet adhesion with a radioactive chemical crosslinking reagent: Interaction of thrombospondin with fibronectin and collagen. Cell 1982, 31:253-262 10. Leung LLK, Harpel PC, Nachman RL: Complex formation of platelet thrombospondin with histidine rich glycoprotein. J Clin Invest 1984, 73:5-12 11. Silverstein RL, Leung LLK, Harpel PC, Nachman RL:

12.

13. 14.

15. 16. 17.

18.

Complex formation of platelet thrombospondin with plasminogen: Modulation of activation by tissue activator. J Clin Invest 1984, 74:1625-1633 Roberts DD, Haverstick DM, Dixit VM, Frazier WA, Santoro SA, Ginsburg V: The platelet glycoprotein thrombospondin binds specifically to sulfated glycolipids. J Biol Chem 1985, 260:9405-9411 Mumby SM, Abbott-Brown D, Raugi GJ, Bornstein P: Regulation of thrombospondin secretion by cells in culture. J Cell Physiol 1984, 120:280-288 Majack RA, Cook SC, Bornstein P: Platelet-derived growth factor and heparin-like glycosaminoglycans regulate thrombospondin synthesis and deposition in the matrix by smooth muscle cells. J Cell Biol 1985, 101:1059-1070 Clowes AW, Karnovsky MJ: Suppression by heparin of smooth muscle proliferation in injured arteries. Nature 1977, 265:625-626 Hoover RL, Rosenberg RD, Haering W, Karnovsky MJ: Inhibition of rat smooth muscle cell proliferation by heparin: II. In vitro studies. Circ Res 1980,47:578583 Majack RA, Clowes AW: Inhibition of vascular smooth muscle cell migration by heparin-like glycosaminoglycans. J Cell Physiol 1984, 118:253-256 Wight TN, Raugi GJ, Mumby SM, Bornstein P: Light microscopic immunolocation of thrombospondin in

372

19. 20. 21.

22.

23.

24. 25.

26.

RAUGI AND LOVET[

human tissues. J Histochem Cytochem 1985, 33:295302 Lovett DH, Sterzel RB, Kashgarian M, Ryan JL: Neutral proteinase activity produced in vitro by cells of the glomerular mesangium. Kidney Int 1982, 23:342-347 Striker GE, Striker U: Biology of disease: Glomerular cell culture. Lab Invest 1985, 53:122-131 Lowry OH, Rosebrough NJ, Farr AL, Randall RJ: Protein measurement with the Folin phenol reagent. J Biol Chem 1951, 193:265-275 Loomis TC, Stahl WL: A rapid, flexible method for biochemical assays using a microtiter plate reader and a microcomputer: Application for assays of protein, Na,K-ATPase, and K-p-nitrophenylphosphatase. Int J Biomed Computing 1986, 18:183-192 Raugi GJ, Mumby SM, Abbott-Brown D, Bornstein P: Thrombospondin: Synthesis and secretion by cells in culture. J Cell Biol 1982, 95:351-354 Laemmli UK: Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 1970, 227:680-685 Phillips DR, Jennings LK, Prasanna HR: Ca"-mediated association of glycoprotein G (thrombin-sensitive protein, thrombospondin) with human platelets. J Biol Chem 1980, 255:11629-11632 Nurden AT, Kunicki TS, Dupuis D, Soria C, Caen JP: Specific protein and glycoprotein deficiencies in platelets isolated from two patients with the gray platelet syndrome. Blood 1982, 59:709-718

AJP * November 1987

27. Houdijk WPM, deGroot PG, Nievelstein PFEM, Sakariassen KS, Sixma JJ: Subendothelial proteins and platelet adhesion: Von Willebrand factor and fibronectin, not thrombospondin, are involved in platelet adhesion to extracellular matrix of human vascular endothelial cells. Arteriosclerosis 1986, 6:24-33 28. Jaffe EA, Leung LLK, Nachman R, Levin RI, Mosher DF: Thrombospondin is the endogenous lectin of human platelets. Nature 1982, 295:246-248 29. Dixit VM, Haverstick DM, O'Rourke KM, Hennessy SW, Grant GA, Santoro SA, Frazier WA: A monoclonal antibody against human thrombospondin inhibits platelet aggregation. Proc Natl Acad Sci USA 1985,

82:3472-3476 30. Castellot JJ, Hoover RL, Harper PA, Karnovsky MJ: Heparin and glomerular epithelial cell-secreted heparinlike species inhibit mesangial-cell proliferation. Am J Pathol 1985, 120:427-435

Acknowledgments We are grateful to the Puget Sound Blood Center for supplying platelets, to Lolan Cheng and Barbara Dunn for technical assistance, to Dr. Groschel-Stewart (Darmstadt, FRG) for anti-myosin, to Dr. Gerald Roth (Seattle) for his gift of von Willebrand factor, and to Drs. Paul Bomstein and Helene Sage (Seattle) for gifts of the collagens.