EXPERIMENTAL

CELL

RESEARCH

201, 119-125 (19%)

Transforming Growth Factor @I-Specific Binding Proteins on Human Vascular Endothelial Cells REIKO HIRAI*

AND KAZUHIKO

*Tokyo Metropolitan Institute of Medical Science, Honkomagome, and tTokyo Metropolitan Institute of Gerontology, Sakae-cho,

KAJI? Bunkyo-ku, Itabashi-ku,

Tokyo, Japan; Tokyo, Japan

ger RNA expression has been found in various tissues [7] and the recombinant expression of TGFP3 in mammalian cells has been performed [8]. TGF/31 is the isoform originally described and the one expressed most generally in both normal and malignant cells and tissues. Furthermore, TGFPl is stored in human platelets [9], perhaps in order to act on vascular cells at high concentrations during times of emergency. TGFP2 shares 72% amino acid sequence homology with TGFPl. Human TGF/32 was first isolated from malignant cell lines [5,6]; subsequently, like TGFPl, its expression has been demonstrated in a variety of mammalian tissues. In classical assay systems for TGF/3, such as soft agar colony formation of murine fibroblasts and growth inhibition of mink epithelial cells, TGFP2 exhibits similar activity to TGFPl [4]. However, recent studies using various cell systems have shown differences between the functions of TGF/31 and TGF/32 [ll-131. TGFP binds to various types of cells with high affinity [ 141. Cross-linking studies with radiolabeled TGFP reveal that TGFP binds to several membrane proteins and that the binding patterns vary depending on the types of cell as well as on the TGFP isoform [4, X-171. Three types of TGFP binding proteins have been identified on a number of epithelial and fibroblastic mammalian cells: two glycoproteins, 53 kDa (type I) and 70-100 kDa (type II) in size, and a high molecular weight component of 200-400 kDa (type III) [2] that has been identified as a proteoglycan with a lOO- to 120-kDa core protein [18]. The primary structure of the type III receptor has been deduced from cDNA cloned recently [ 19,201. The function of the TGFP binding proteins will be identified by cDNA cloning and expression studies. However, the possibility that type I and/or type II components function as TGFP receptors mediating signal transduction has been suggested by the characterization of TGF/3 binding proteins in mutant cells that lack responses to TGFP [21, 221. In various cell types, including vascular endothelial cells, types I and II have been found, while type III is not detectable [12, 16, 231. In addition to these three TGFP binding proteins, various cell surface components interacting with TGFP have been demonstrated: a 60-kDa protein that binds activin and inhibin

Transforming growth factor /? (TGFj3) regulates the growth of human umbilical vein endothelial cells (HUVEC) differently depending on the isoform of TGF/3 and the culture conditions. The cells are resistant to growth inhibition by TGFB when the cells are cultured on substratum coated with gelatin. However, the proliferation of HUVEC cultured on substratum without a gelatin coating is inhibited by TGFB, depending on the isoform and concentration of TGFj3. Binding assays with “‘ITGFBl reveal that HUVEC contain a single class of high-affinity (& = 4.4 PM) TGF@l binding sites with 8500 sites per cell. Affinity cross-linking studies demonstrate that HUVEC express 180 and 80 kDa TGFfll binding sites that do not bind TGFfl2. The reduction and the removal of glycosaminoglycans does not affect the electrophoretic mobility of the ISO-kDa binding protein cross-linked with “‘1-TGFfll. Therefore, the 1 SOkDa TGF@l binding protein is not related to the type III TGFj3 receptor, but might be a novel TGFBl-specific receptor/binding protein expressed on vascular endothelial cells. The expression of TGF/31 binding sites is not affected by the presence or absence of the gelatin coating on the culture substratum. The data suggest that a gelatin coating does not regulate the susceptibility of HUVEC to TGF@l at the level of the receptor/binding proteins, and that growth inhibition of HUVEC by TGF@l is linked to the regulation of extracellular matrices required for the interaction between the cells and 0 1992 Academic Press. Inc. the substratum.

INTRODUCTION

Transforming growth factor@ (TGFP) is a multifunctional polypeptide growth factor that regulates differentiation, cellular functions, and proliferation of a variety of cell types in both positive and negative ways depending on the microenvironment [l, 21. TGFP is a dimeric polypeptide with subunits composed of 112 amino acids located on the C-terminal of the precursor [3]. At present, among five known isoforms, TGF/31 and TGFP2 [4-61, have been identified as the protein from natural mammalian sources, although TGFP3 messen119

All

Copyright 0 1992 rights of reproduction

0014.4827/92 $5.00 by Academic Press, Inc. in any form reserved.

120

HIRAI

AND

[24] in rat pituitary tumor cells, a group of disulfidelinked TGFpl-specific binding proteins of 320-170 kDa in rat glomeruli [25], a 400-kDa glycoprotein in bovine liver [26], and TGFpl-specific 150- to MO-kDa glycoproteins in several cell lines [27]. TGFP has been shown to regulate proliferation and many functions of vascular endothelial cells from various sources [13, 28-311. The effect of TGFP alters depending on the concentration and isoform of TGFP, as well as on the cell type and culture conditions. Because the majority of endothelial cell lines require the addition of growth factors belonging fibroblast growth factors and extracellular matrices for their proliferation in u&-o, it is likely that at least part of the effect of TGFP on cells is linked to the regulation of the susceptibility and processing of fibroblast growth factors and extracellular matrices. In this report we describe the effect of TGFP and the expression of TGFP binding proteins on human umbilical vein endothelial cells (HUVEC). In HUVEC, which retain the ability to produce extracellular matrices, TGFP inhibits proliferation depending on the isoform and concentration. The presence of extracellular matrices counteracts the inhibition by TGFP. Cross-linking reveal 180- and 80-kDa studies using iz51-TGFpl TGFpl-specific binding proteins on the cells. The relationship between receptor expression and the susceptibility to TGFP is discussed. MATERIALS Cells and culture conditions.

AND

METHODS

HUVEC were isolated from umbilical cords by the method of Jaffe et al. 1321, using 0.1% trypsin and 0.02% EDTA instead of collagenase, and maintained in thymidine-deprived MCDB104 medium (Kyokuto Seiyaku Kogyo Co., Tokyo, Japan) supplemented with 10% fetal calf serum @era-Lab, Sussex, England), 10 rig/ml EGF. (human recombinant; Upstate Biotechnology, Lake Placid, NY), 100 pg/ml heparin (Sigma Chemical Co., St. Louis, MO), and 69 rig/ml ECGF. ECGF was a partially purified acidic FGF preparation obtained from bovine brains by heparin affinity chromatography, according to the method described by Lobb and Fett [33]. HUVEC were maintained in the plastic tissue culture plates pretreated with 0.1% gelatin (Sigma Chemical Co.) in phosphate-buffered saline (PBS; 0.8% NaCl, 0.02% KCl, 0.295% Na,HPO,. 12H,O, 0.02% KH,PO,) for 30 min at room temperature and then washed once with PBS. HUVEC at 6 to 15 population-doubling levels, the cells used in these experiments, can proliferate without gelatin coating, although better proliferation was obtained on substratum coated with gelatin. Human embryonic lung diploid fibroblasts, TIG-7 [34], were maintained in HD medium (Ham F12:Dulbecco’s modification of Eagle’s medium; 1:l; Sigma Chemical Co.) containing 15 mM 4-(2-hydroxyethyl)-1-piperazine-ethane-sulfonic acid (Hepes) supplemented with 10% fetal calf serum. Normal rat kidney (NRK) clone 49F cells were obtained from American Type Culture Collection through Dainippon Seiyaku Co. (Osaka, Japan). A mink lung epithelial cell line, MvlLu, was obtained from Dr. H. Yoshikura, Tokyo University. NRK49F and MvlLu were maintained in HD medium containing 15 mM Hepes supplemented with 5% fetal calf serum. TGF-(3s. The iZ51-labeled TGFPl (human recombinant; BoltonHunter labeled) was obtained from DuPont/NEN Research Products

KAJI

(Wilmington, DE). TGFPl and TGF/32 derived from porcine platelets were purchased from R&D Systems (Minneapolis, MN). These TGFP preparations were titrated for their biological activities in the colony formation assay on NRK49F (for ‘251-labeled TGF@l) and the growth inhibition assay on MvlLu (for porcine TGFPl and $2). All TGFP preparations had similar activities on the indicator cells. Proliferation assay. HUVEC were plated in 24.well plates (Costar, Cambridge, MA) with or without gelatin coating at a density of 2 X lo4 cells per well (2 cm*) with 1 ml MCDB104 medium containing fetal calf serum, EGF, ECGF, and heparin. Cells were allowed to attach and spread for 16 h before the addition of TGF@s. Four days later cells were trypsinized and counted with a Coulter counter. Synthesis of DNA was assayed by incorporation of [‘*‘I]iododeoxyuridine (NEN), added at a concentration of 0.5 &i/well, into a 5% trifluoroacetic acid-insoluble fraction between 18 and 24 h after addition of TGFP. Binding assay with ‘25Z-TGF(31. In the studies on TGFPl receptor/ binding proteins, cells were plated in 6-well plates (Costar) at a density of 1 x lo5 cells per well (8 cm’) with 2 ml growth medium and incubated for 2 or 3 days. Binding assays were performed essentially as described by Massague and Like [15]. Briefly, cells were washed and then incubated for 2 h in HD medium supplemented with 0.2% bovine serum albumin (BSA) and 15 mM Hepes, pH 7.2. After washing twice with binding buffer [128 mM NaCl, 5 mM KCI, 1.2 mM CaCI,, 1.2 mM MgCl,, 50 mM Hepes, pH 7.4,5 mglml BSA, 1 mM phenylmethylsulfonyl fluoride (PMSF), 10 pg/ml leupeptin, 10 fig/ml pepstatin A], cells received appropriate amounts of iz51-TGF@l with or without a 50-fold excess of unlabeled TGFPl in 0.5 ml binding buffer and incubated for 4 h at 4’C on a rocking platform. Cells were washed five times with binding buffer, once with sucrose buffer (0.25 M sucrose, 1 mM EDTA, 10 mM Tris-HCl, pH 7.4,1 mM PMSF, 10 pg/ml leupeptin, 10 pg/ml pepstatin A), and solubilized in 0.5 ml solubilization buffer (1% Triton X-100, 1 mM EDTA, 1 mM PMSF, 10 wg/ml leupeptin, 10 fig/ml pepstatin A, 10 mM TrisHCl, pH 7.0) at 4°C for 40 min. Affinity labeling of TGFPI binding proteins with ‘25Z-TGF@I. TGFPl binding sites on cells were cross-linked to iz51-TGFfll using disuccinimidyl suberate (DSS; Pierce, Rockford, IL). Cells were culas described above for tured, washed, and incubated with “?-TGF@l binding assay. In some experiments, instead of preincubation in HD medium, cells were washed briefly with acid washing buffer (0.1% acetic acid, 150 mM NaCl, 0.1% BSA) according to the method described by Glick et al. [35] to remove cell surface TGFP and serum factors. After washing five times with binding buffer, the cells were treated for 15 min with 0.3 mM DSS in 0.5 ml binding buffer. Then, the cells were washed once with sucrose buffer and solubilized in lOO200 ~1 solubilizing buffer for 40 min at 4°C. The solubilized material was clarified by centrifugation at 10,OOOgfor 10 min and sodium dodecyl sulfate (SDS) and glycerol were added to final concentrations of 1 and lo%, respectively. SDS-polyacrylamide gel electrophoresis (PAGE) was performed on linear 5-20% acrylamide gradient gels (ATTO, Tokyo, Japan). Rainbow protein molecular weight markers (Amersham, Buckinghamshire, England) containing myosin (200 kDa), phosphorylase b (97 kDa), BSA (69 kDa), ovalbumin (46 kDa), carbonic anhydrase (30 kDa), trypsin inhibitor (21.5 kDa), and lysozyme (14.3 kDa) were used as standards. After electrophoresis, the gels were fixed with 30% methanol in 10% acetic acid for 30 min, dried, and subjected to autoradiography with X-OMAT AR (Kodak, Rochester, NY) or Fuji Imaging Plates (Fuji Photo Film Co., Kanagawa, Japan) followed by analysis using a Bio-Image Analyzer BAS2000 (Fuji Photo Film Co.). Treatment of cells with glycosaminoglycan-cleaving enzymes. The removal of glycosaminoglycans from cells was performed by a modification of the method described by Cheifetz et al. [18]. Monolayers of HUVEC and TIG-7 cells were washed with HD containing 0.2% BSA;

TGFfll-BINDING

PROTEINS

ON HUMAN

ENDOTHELIAL

121

CELLS

doubling levels. To study whether the growth inhibition of HUVEC grown on the substratum without gelatin coating is a primary or secondary effect of TGFP, DNA synthesis in the cells from 18 to 24 h after addition of TGFP was examined. As seen in Table 1, the inhibitory effect was not a primary one, because both TGF/3 isoforms had little effect on DNA synthesis in HUVEC at doses that inhibited cell number growth of the cells counted after 4 days incubation. On the other hand, TGFP inhibited the growth of TIG-7 fibroblasts directly by preventing DNA synthesis. The Expression

0

20 78 313 1250 5000 TGFBl (pghl)

0

20

78 313 12505000 TGF82 (pg/ml)

FIG. 1. The effect of TGFPl and TGFP2 on the proliferation of HUVEC. Cells were plated in 24.well plates with (0) or without (0) gelatin coating at a density of 2 X 104/well, received TGFPl or TGFPP, and cultured in 1 ml MCDB104 supplemented with 10% fetal calf serum, 10 rig/ml EGF, 69 rig/ml ECGF, and 100 pg/ml heparin. After 4 days, cells were trypsinized and counted with a Coulter counter.

received 25 milliunits chondroitinase ABC (Seikagaku Kogyo, Tokyo, Japan), 2.5 milliunits heparitinase (Seikagaku Kogyo), or a combination of both enzymes in 0.5 ml binding buffer; and were incubated at 37°C for 3 h. Cells were washed 5 times with ice-cold binding buffer and then affinity-labeled with 1251-TGF@1. RESULTS

The Effect of TGF(3 on the Growth of HUVEC The effect of TGFP on the growth of HUVEC varied with the culture substratum, isoform, and concentration of TGFP. As shown in Fig. 1, neither TGFPl nor TGFP2 affect the proliferation of the cells plated on the substratum coated with gelatin. On the other hand, the growth of HUVEC was inhibited by both TGFP isoforms when the cells were plated on plastic tissue culture plates without gelatin coating. Half-maximal inhibition occurred at 0.03-0.05 rig/ml. Unexpectedly, the inhibition by TGFPl was less at doses higher than 1.25 rig/ml, and TGFPl at 5 rig/ml did not affect the proliferation of HUVEC. This recovery from growth inhibition at high concentrations was specific for TGFPl. Proliferation was blocked completely in the culture receiving 5 rig/ml TGFP2 as seen in Fig. 1, and no recovery was observed with TGFP2 at concentrations greater than 20 rig/ml (data not shown). The unresponsiveness of HUVEC grown on gelatin to TGFP and the difference in responses to TGFP isoforms were observed in repeating experiments using HUVEC at 6, 10, and 13 population-

of TGF(3 Binding

Proteins in HUVEC

To examine the correlation between the susceptibility to TGFPl and the expression of TGFPl binding proteins on the cells, binding kinetics analysis of lz51TGFPl to HUVEC and affinity labeling studies of TGFPl binding proteins in HUVEC were performed. The binding of 1251-TGFfil was saturable as seen in Fig. 2A. Scatchard analysis, shown in Fig. 2B, revealed that HUVEC express a single class of high affinity (& = 4.4 PM) binding sites with 8500 sites per cell. Cross-linking analysis demonstrated that HUVEC expressed two classes of TGF/31 binding sites (Fig. 3A). From the molecular masses of the complexes composed of cell surface components and ‘251-TGFpl, the TGFPl binding proteins on HUVEC were estimated as 180 and 80 kDa. The addition of 25-fold excess unlabeled TGFP2, which effectively decreased affinity labeling of three classes of TGFPl binding sites in TIG-7 fibroblasts, had little effect on labeling of both TGFPl binding sites in HUVEC. There were no differences in the amounts of ‘251-TGF/R binding (data not shown) or in the affinity labeling patterns between HUVEC with (Fig. 3A) or without (Fig. 3B) washing with mild acid, indicating that there are no masked binding sites by endogenous TGFPl or serum factors such as 01~macroglobulin [36, 371. The effect of gelatin coating of the substratum on the expression of TGFPl binding proteins was analyzed, because, as seen in Fig. 1, the substratum strongly influences the effect of TGFPl. No differences were found in the labeling patterns between HUVEC grown without (Fig. 3B) and with (Fig. 3C) gelatin. These results suggest that the expression of TGFPl binding sites is not linked to a susceptibility to TGFPl that depends on the gelatin coating of the substratum. Characteristics

of TGFpl

Binding

Sites in HUVEC

The bulk of the 180- and 80-kDa TGFPl binding proteins in HUVEC did not form disulfide-linked complex, because as seen in Figs. 4A and 4B, the mobility on SDS-PAGE of each TGF/31 binding sites was unchanged after (A) and before (B) reduction. The high molecular weight components eluted at the 250- to 300-

122

HIRAI

III

510 20

AND

KAJI

I

io

40

' 120

' 160

1251-TGF61add (fmol)

0.1

0.4

Bound~fmol/f.6x

4.0

10 lO%U.s)

FIG. 2. Binding of ‘x51-TGFpl (A) and the Scatchard plot of the binding (B) to HUVEC. Specifically bound iz51-TGFfll by subtracting nonspecific binding in the presence of 50-fold excess unlabeled TGFfll (w) from total binding (A).

kDa region disappeared after reduction. Densitometry analysis for the radioautography in Fig. 4B revealed that the 250- to 300-kDa regions contained less than 20% ‘251-TGF/31 cross-linked to cellular components (data not shown). It is assumed that reduction of affinity labeled samples resulted in the dissociation of a 1251labeled TGF@l subunit from the cross-linked complex and the loss of about one half of radioactivity, because ‘251-TGFfi1 used in this study contains less than one “‘1

per molecule. Therefore, to examine TGFPl binding sites in HUVEC which received small amounts of labeled ligand, the dose effect of 1251-TGFfil on the affinity labeling of the binding sites was analyzed on SDS-

161

[Al CELLS GELATIN ACID WASh

TABLE 1 Effect of TGFP Isoforms on the Growth of HUVEC and TIG-7 Cells

(@) were obtained

HUVEC + - 81 82

TIG-7 + - 81

82

HUVEC

ICI HUVEC + -

-81 82

- 81 82

kDa ZOO-

97 -

TGF@ None Pl

P2

DNA synthesis 24 h after TGFP addition (cpm/well)

Number of cells on the 4th day (X10-4/cm2)

rig/ml

HUVEC

TIG-7

HUVEC

1556 0.078 5 0.078 5

1502 1872 1640

6338 2050 526 2349 686

1679

2.33 1.72 2.49 1.44 0.84

6946-

TIG-7 1.94 0.78 0.53 1.18 0.61

Note. The cells were plated in 2-cm* wells at a density of 2 X 104/ well with 0.5 ml culture medium, MCDBlO4 supplemented with ECGF, EGF, heparin, and 10% fetal calf serum for HUVEC and Dulbecco’s MEM containing 15 mM Hepes and 10% fetal calf serum for TIG-7 cells. TGFfll or TGFPP was added after 3 h incubation. The per well from 18 cells were pulsed with 0.5 pCi 1261-iododeoxyuridine to 24 h after addition of TGFP. Number of cells was counted on the 4th day of incubation.

FIG. 3. Affinity labeling of TGFfil binding proteins in HUVEC. HUVEC were cultured for 3 days in 8 cm’-wells precoated without (A and B) or with (C) gelatin incubated with 1.6 ng iz51-TGF@l (63 aCi/ pg) with or without 40 ng unlabeled TGFfll or TGFBZ, and treated with DSS. To remove cell surface TGF/3 and serum factors, mild acid wash was performed before affinity labeling as described under Materials and Methods (A). After extraction with Triton X-100 and reduction with 5% 2-mercaptoethanol, affinity-labeled proteins were electrophoresed on 5-20% SDS-polyacrylamide gels and subjected to radioautography using X-OMAT AR film. For comparison, the TGFPl binding proteins on TIG-7 fibroblasts were analyzed on the same gel (A).

TGF@l-BINDING

[Al REDUCED 123 - ,L/= ‘4

PROTEINS

Bl NONREDUCED 123

ON HUMAN

ENDOTHELIAL

123

CELLS

TGFbl TGF@ -x20x50x200 flx50x200x4Do

ICI NONREDUCED 1234567

200-

4

97-

M

6946’ 30-

FIG. 5.

FIG. 4. The effect of reduction on the pattern of affinity-labeled TGFPl binding proteins (A and B) and the dose effect of iz51-TGF@l on the affinity labeling of the TGFPl binding proteins (C) in HUVEC. (A and B) HUVEC cultured in &cm2 wells without gelatin coating and anawere affinity labeled with 0.5 ng ‘“51-TGF@l (115 &i/~g) lyzed on SDS-PAGE with (A) or without (B) reduction with 5% 2mercaptoethanol. Proteins in 5, 10, and 20 ~1 extract were applied to lanes 1, 2, and 3 in (A) and (B), respectively. (C) HUVEC grown without gelatin were affinity labeled with iz51-TGF@l at various concentrations; 0.06 (lane l), 0.125 (lane 2), 0.25 (lane 3), 0.5 (lane 4), 1 (lane 5), 2 (lane 6), and 4 rig/ml (lane 7). SDS-PAGE samples were nonreduced. The patterns of affinity-labeled proteins were obtained using the Bio-Image Analyzer.

PAGE under nonreduced condition. As seen in Fig. 4C, the 180- and 80-kDa TGF/?l binding proteins detected in HUVEC received small amounts of lz51-TGFfll that strongly inhibit proliferation (0.125 and 0.25 rig/ml), as well as in the cells that received ‘251-TGF@1 that does not inhibit the proliferation (4 rig/ml). Addition of 20- to 400-fold excess unlabeled TGFP2 did not affect the total binding of ‘251-TGFpl to HUVEC, whereas a 20-fold excess unlabeled TGFPl competed 88% of lz51-TGF/31 binding to the cells (data not shown). Neither 180- nor 80-kDa TGF/31 binding proteins in HUVEC bound TGFPB; the affinity labeling of the 180- or 80-kDa sites with [‘251]TGFfll was not prevented by the presence of up to 400-fold excess unlabeled TGFp2 as seen in Fig. 5. The 180-kDa TGF@l Binding Protein on HUVEC Not Contain Glycosaminoglycans

VEC. ence were sites.

Effect of TGFP2 on the binding of ‘Z61-TGF~1 to HUCells were affinity labeled with 0.5 ng 12SI-TGF@1 in the presof indicated amount of TGFPl or TGFPZ. SDS-PAGE samples nonreduced. Arrows indicate 180- and SO-kDa TGFPl binding Radioautograph was prepared using Bio-Image Analyzer.

virtually complete conversion of the component into the 140 kDa products. The results indicate that the 180-kDa TGFPl binding protein in HUVEC does not contain glycosaminoglycan chains and suggests that the 180-kDa site is not related to type III TGFP receptor. DISCUSSION

Under our culture conditions, the effect of TGFP on the growth of HUVEC changed depending on the substratum for culture, as well as on the concentration and isoform (Bl and B2) of TGFP. TGFP does not affect proliferation of the cells plated on gelatin coating. On the contrary, cells plated on substratum without gelatin coating were highly sensitive to growth inhibition by TGFPl and TGFp2, with a half-maximal dose of 0.030.05 rig/ml. The growth inhibition of HUVEC by TGFP is not a primary effect because TGF/3 does not prevent HUVEC entering S-phase. The results suggest the possibility that TGF/3 regulates the proliferation via synthe-

HUVEC - C H&H

TIG-7 -

C

HC+H

Does

The removal of glycosaminoglycans from HUVEC by incubation with chondroitinase ABC, heparitinase, or a combination of both enzymes did not affect the binding of ‘251-TGF/31 to the 180-kDa site as shown in Fig. 6. Under the same condition, both enzymes induced the decrease in the molecular mass of the 250- to 300-kDa component in TIG-7 fibroblasts, and treatment of the cells with both enzymes in combination resulted in a

FIG. 6. Effect of the removal of glycosaminoglycans on the TGF/31 binding proteins in HUVEC and TIG-7 fibroblasts. Cells were treated with chondroitinase ABC (C), heparitinase (H), or a combination of both enzymes (C + H) at 37°C for 3 h and affinity labeled with ‘251-TGFP as described under Materials and Methods. The 180-kDa (in HUVEC, A) and 250- to 300-kDa (in TIG-7, A) TGFPl binding sites were indicated. Radioautographs from top of gels to the 97-kDa region were shown.

124

HIRAI

AND

sis of extracellular matrices. Differences in the TGFP isoforms could be detected only at doses higher than 1.25 rig/ml in cells plated without gelatin coating. Cells receiving TGFPl at doses higher than 1.25 rig/ml recovered from the growth inhibition, but no recovery was observed in cells receiving TGFp2. The addition of TGFPl at doses greater than 1.25 rig/ml resulted in the same effect on HUVEC with that induced by gelatin coating of substratum. The HUVEC used in this study retained the ability to produce the extracellular matrices required for proliferation; however, cells plated on substratum coated with gelatin exhibited not only better proliferation but also resistance to the growth inhibition produced by TGFP. The mechanism for the recovery from growth inhibition induced by higher doses of TGFPl is not known, but it is suggested that the regulation of extracellular matrices plays a key role in the control of TGFfll-induced growth of HUVEC. Cross-linking studies demonstrated MO- and 80-kDa TGF/31-specific binding proteins on HUVEC. The molecular weights of these proteins are not changed under reducing conditions. The addition of excess unlabeled TGF(32 had no effect on the binding of lz51-TGF/31 to these proteins. The two major TGF(31 binding proteins in HUVEC appear to bind TGFPl with similar affinities. The results suggest the possibility that TGFPl transduces both inhibitory and stimulatory signals through a common set of receptors. The 80-kDa proteins appeared to resemble the type II receptor for TGFP expressed ubiquitously in mammalian cells [2] including TIG-7 fibroblast. On the other hand, the 180kDa TGFpl-binding protein is distinct from type III TGFP receptor because it does not contain glycosaminoglycans. In addition, the 180-kDa protein appears to differ from TGF/31-specific binding proteins observed in rat glomeruli [25] and bovine liver [26] in the electrophoretie mobility and susceptibility to the reducing agent. Other TGFfll-specific binding proteins detected in various cultured cells [27] have common characteristics with the 180-kDa protein, although their affinities for TGF(31 are much lower than that of the 180-kDa protein in HUVEC. The 180-kDa protein in HUVEC might be a novel TGF/31-specific receptor/binding protein expressed in vascular endothelial cells. In HUVEC, as much as a 400-fold excess of TGFP2 did not affect the affinity labeling of the 180- and 80kDa TGF-fil-binding sites, although a 20-fold excess TGFPl effectively prevented labeling. As far as tested in our laboratory, four different human and bovine vascular endothelial cell lines do not express type III receptor, and a 50-fold excess TGFP2 does not compete with 1251TGFPl for binding to these cells (data not shown). In contrast, a 50-fold excess TGFP2 completely prevents the labeling of type I and II receptors as well as type III receptor in various cell lines including TIG-7 fibro-

KAJI

blasts. It appears likely that HUVEC possess specific or preferential TGFP receptors for each isoform. Studies on TGFP2 receptor/binding proteins in HUVEC are in progress. The authors thank Dr. Margaret D. Ohto for her help in preparing this manuscript. This work was supported by grants from the Japanese Ministry of Education, Science, and Culture.

REFERENCES 1.

Sporn, M. B., Roberts, A. B., Wakefield, L. M., and de Crombrugghe, B. (1987) J. Cell Biol. 105,1039-1045.

2.

Massague, J. (1990) Annu. Reu. Cell Biol. 6, 597-641.

3.

Derynck, R., Jarret, J. A., Chen, E. Y., Eaton, D. H., Bell, J. R., Assoian, R. K., Roberts, A. B., Sporn, M. B., and Goeddel, D. V. (1985) Nature 316, 701-705.

4.

Cheifetz, S., Weatherbee, J. A., Tsang, M. L.-S., Anderson, J. K., Mole, J. E., Lucas, R., and Massague, J. (1987) Cell 48,409-415.

5.

Ikeda, T., Lioubin, M. N., and Marquardt, try 26, 2406-2410.

6.

de Martin, R., Harndler, B., Hofer-Warbinek, R., Gaugitsch, H., Wrann, M., Schlusener, H., Seifert, J. M., Bodmer, S., Fontana, A., and Hofer, E. (1987) EMBO J. 6, 3673-3677.

I.

Miller, D. A., Lee, A., Matsui, Y., Chen, E. Y., Moses, H. L., and Derynck, R. (1989) Mol. Endocrinol. 3, 1926-1934.

8.

Graycar, J. L., Miller, D.A., Arrick, B. A., Lyons, R. M., Moses, H. L., and Derynck, R. (1989) Mol. Endocrinol. 3, 1977-1986.

9.

Assoian, R. K., Komoriya, A., Meyers, C., Miller, Sporn, M. B. (1983) J. Biol. Chem. 258, 71557160.

H. (1987) Biochemti-

D. M., and

10.

Miller, D. A., Lee, A., Pelton, R. W., Chen, E. Y., Moses, H. L., and Derynck, R. (1989) Mol. Endocrinol. 3, 1108-1114.

11.

Rosa, F., Roberts, A. B., Danielpour, D., Dart, L. L., Sporn, M. B., and Dawid, I. B. (1988) Science 239, 783-785.

12.

Ohta, M., Greenberger, J. S., Anklesaria, P., Bassols, Massague, J. (1987) Nature 329, 539-541.

13.

Merwin, J. R., Newman, J. (1991) Am. J. Pathol.

14.

Wakefield, L. M., Smith, D. M., Matsui, T., Harris, C. C., and Sporn, M. B. (1987) J. Cell Biol. 105,965-975. Massague, J., and Like, B. (1985) J. Biol. Chem. 260,2636-2645.

15.

W., Beall, L. D., Tucker, 138, 37-51.

A., and

A., and Madri,

16.

Segarini, P. R., Rosen, D. M., and Seyedin, Endocrinol. 3, 261-271.

17.

Segarini, P. R., Roberts, A. B., Rosen, D. M., and Seyedin, S. M. (1987) J. Biol. Chem. 262, 14,655-14,662.

18.

Cheifetz, S., Andres, J. L., and Massague, Chem. 263, 16,984-16,991.

19.

Lopez-Casillas, F., Cheifetz, S., Doody, J., Andres, J. L., Lane, W. S., and Massague, J. (1991) Cell 67, 785-795.

20.

Wang, X.-F., Lin, H. Y., Ng-Eaton, E., Downward, J., Lodish, H. F., and Weinberg, R. A. (1991) Cell 67, 797-805. Boyd, F. T., and Massague, J. (1989) J. Biol. Chem. 264,22722278.

21. 22.

S. M. (1989) Mol.

J. (1988) J. Biol.

Laiho, M., Weis, F. M. B., and Massague, J. (1990) J. Biol. Chem. 265, 18,518-18,524.

TGFpl-BINDING

PROTEINS

ON HUMAN

23.

Massague, J., Cheifetz, S., Endo, T., and Nadal-Ginard, (1986) Proc. Nutl. Acad. Sci. USA 83,8206-8210.

24.

Cheifetz, S., Ling, N., Guillemin, Biol. Chem. 263, 17,225-17,228.

25.

MacKay, K., Robbins, A. R., Bruce, M. D., and Danielpour, (1990) J. Biol. Chem. 265, 9351-9356.

26.

O’Grady, P., Kuo, M.-D., Baldassare, J. J., Huang, S. S., and Huang, J. S. (1991) J. Biol. Chem. 266,8583-8589. Mackay, K., and Danielpour, D. (1991) J. Biol. Chem. 266, 9907-9911.

21.

R., and Massague,

28.

Heimark, R. L., Twardzik, Science 233, 1078-1080.

29.

Jennings, J. C., Mohan, S., Linkhart, Baylink, D. J. (1988) J. Cell. Physiol.

Received November 18, 1991 Revised version received March

D. R., and Schwartz,

11, 1992

B.

30.

J. (1988) J.

31.

S. M. (1986)

T. A., Widstrom,

137, 167-172.

D.

R., and

32. 33. 34. 35. 36. 37.

ENDOTHELIAL

CELLS

125

Pepper, M. S., Belin, D., Montesano, R., Orci, L., and Vassalli, J.-D. (1990) J. Cell Biol. 111, 743-755. Myoken, Y., Kan, M., Sato, G. H., McKeehan, W. L., and Sato, D. (1990) Exp. Cell Res. 191,299-304. Jaffe, E. A., Nachman, R. L., Becker, C. G., and Minck, C. R. (1973) J. Clin. Znuest. 53, 2745-2756. Lobb, R. P., and Fett, J. W. (1984) Biochemistry 23,6295-6299. Yamamoto, K., Kaji, K., Matsuo, M., Shibata, Y., Tasaki, Y., Utakoji, T., and Ooka, H. (1991) Exp. Gerontol. 26, 525-540. Glick, A. B., Danielpour, D., Morgan, D., Sporn, M. B., and Yuspa, S. H. (1990) Mol. Endocrinol. 4, 46-52. O’Connor-McCourt, M. D., and Wakefield, L. M. (1987) J. Biol. Chem. 262, 14,090-14,099. LaMarre, J., Hayes, M. A., Wollenberg, G. K., Hussaini, I., Hall, S. W., and Gonias, S. L. (1991) J. Clin. Invest. 87, 39-44.