EXPERIMENTAL

CELL

RESEARCH

186,12%129

(1990)

ffects of Epidermal Growth Factor and Transforming on Rat Heart Endothelial Cell Anchorage-De and -Independent Growth1 DANIEL L. MOORADIAN~ AND Department

of Pathology,

Wayne

State

University

This report describes the effects of epidermal growth factor (EGF) and transforming growth factor-/31 (TGF,61) on the anchorage-dependent and -independent growth of rat heart endothelial cells (RHE-1A). When RHE-IA cells were grown in monolayer culture with medium containing 10% fetal bovine serum (FBS) supplemented witb epidermal growth factor (0.1-100 ng/ ml), growth was stimulated fivefold when compared to that of cells grown in medium containing 10% FBS alone. The stimulatory effect of EGF on RHE-1A cell monolayer growth was dose-dependent and half-maximal at 5 rig/ml. The addition of TGF-@l in the range 0.1-10 rig/ml bad no effect on RHE-1A cell monolayer growth when added to medium containing 10% FBS alone or 10% FBS supplemented with EGF (50 rig/ml). RHE- IA cells failed to grow under anchorage-independent conditions in 0.3% agar medium containing 10% FBS. In the presence of EGF, however, colony formation increased dramatically. The stimulatory effect of EGF was dose-dependent in the range 0.1-100 rig/ml and was half-maximal at 5 rig/ml. In contrast to its effects under anchorage-dependent conditions, TGF-@I (0.1-10 rig/ml) antagonized the stimulatory effects of EGF on RHE-1A cell anchorage-independent growth. The inhibitory effect of TGF-81 was dose-dependent and half-maximal at 0.1 rig/ml. EGF-induced RHE-1A soft agar colonies were isolated and reinitiated in monolayer culture. They retained the cobblestone morphology and contact-inhibition characteristic of normal vascular endothelial cells. Each of the clones continued to express Factor VIII antigen. These findings suggest that TGF-/3 may influence not only endothelial cell proliferation but also anchorage dependence. These effects may in turn be of relevance to endothelial cell growth and Zingi0geneSi.S i?Z viva. 0 1990 Academic PWSS, 1~.

$3.00 0 1990 by Academic Press, of reproduction in any form

of Medicine,

Detroit,

Michigan

48201

’ To whom reprint requests should be ad~daessed at: Department Laboratory Medicine and Pathology, University of Minnesota, 609, Mayo Building UMHC, Minneapolis, MN 55455. 122

Inc. reserved.

School

A. DIGLIO

Epidermal growth factor (EGF) and transforming growth factor-p (TGF-/3) influence the growth of vascular endothelial cells. Human umbilical vein endothelial cells are responsive to EGF but bovine endothelial cells from various sites neither bind nor respond to EGF [I]. In uiuo, EGF is an effective angiogenic stimulus in the rabbit cornea1 neovascularization assay [Z]. In addition, TGF-or (an analog of EGF) is a more potent angiogenic mediator than EGF when tested in the hamster cheek pouch assay [31. Transforming growth factor-p is a ubiquitous 25kDa polypeptide that stimulates the anchorage-independent growth of normal cells ]4] an is now recognized as an important bifunctional modulator of growth and differentiation [5]. Interest in its effects on endothelial cell proliferation stems from reports that subcutaneous injections of TGF-/I induce a~gioge~esis and fibrosis [6] and that TGF-P accelerates wound healing ipz uiuo [7]. In vitro, TGF-P transiently inhibits endothelial DNA synthesis and migration [8] and also blocks the stimulatory effects of acidic and basic FGF on the growth of bovine aortic endothelial cells and adrenal cortexderived endothelial cells ]9, IO]. studied the effects of TGF-/!I on receptors by bovine aortic en dial cells, human umbilical vein endotbelial cells, rat heart endothelial cells. TGF-fi inhibited the monolayer growth of all three endothelial populations and hibition was associated with a decrease in the nu of high-affinity EGF receptors, although the total er of EGF receptors remained unchanged. The effects of TGF-fi on more (i.e., angiogencomplex aspects of endotbeliaB esis in vitro) have been ambigu er et al. [12] reported that TGF-p inhibited the phorbol ester-induced invasion of collagen matrices by ~~~i~~a~y endothelial cells. In contrast, Madri et al. [33] demonstrated that

* This work is part of a doctoral thesis by Daniel L. Mooradian to fulfill the requirements of the Graduate School of Wayne State University. This work was supported in part by U.S. Public Health Service Grant HL-23603 and by a postdoctoral fellowship from the Minnesota Affiliate of the American Heart Association.

0014-4827/90 Copyright All rights

CLEMENT

ctor-

of Box

EGF/TGF-@1

EFFECTS

TGF-P did not inhibit the growth of microvascular endothelial cells embedded in collagen gels, but did elicit the formation of tube-like structures. Endothelial cells derived from various sources have an inherent capacity for anchorage-independent growth and this capability may be essential to the proliferation and migration of endothelial cells during angiogenesis [14]. Despite these observations and the fact that EGF and TGF-P are potent modulators of anchorage-independent growth, the effects of EGF and TGF-Pl on the growth of vascular endothelial cells under anchorage-independent conditions have not been previously studied. In this paper, we report that EGF is a potent stimulator of clonal rat heart endothelial cell growth under anchorage-dependent and -independent conditions. While TGF-@l did not inhibit the monolayer growth of rat heart endothelial cells in the presence of 10% FBS alone or in the presence of 10% FBS plus EGF (50 rig/ml), it was a potent inhibitor of EGF-induced anchorage-independent growth. MATERIALS

AND

METHODS

Cell cultures. Normal rat kidney fibroblasts, clone 49F (NRK-49F, CRL 1570), were obtained from the American Type Culture Collection (Rockville, MD). These cells are sensitive to EGF and TGF-P and were used as positive controls in our anchorage-independent growth assays. Only cells at passage levels < 30 were used and all cultures were monitored daily for signs of spontaneous transformation. Rat heart endothelial cell clone RHE-1A was isolated in our laboratory and has been characterized previously [15]. In brief, the identification of RHE-1A cells as endothelial in origin was based on their characteristic cobblestone morphology and positive staining for Factor VIII antigen. All stock cultures were maintained in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal bovine serum (FBS), 25 pg/ml gentamicin, and 250 pg/ml L-glutamine (GIBCO Laboratories, Grand Island, NY). Medium was changed every second day. Mouse submaxillary epidermal growth factor (EGF) was obtained from Sigma Chemical Co. (St. Louis, MO). Platelet-derived porcine transforming growth factor-p1 (TGF-/31) was obtained from R&D Systems, Inc. (Minneapolis, MN). Cellproliferation assay. Rat heart endothelial cells (RHE-1A) were seeded in 24-well trays (2.1 cm’/well) (Falcon Plastics, Cockesville MD) at 2 X lo4 cells/well containing DMEM supplemented with 10% FBS and appropriate concentrations of EGF (l-100 rig/ml) or TGF/31 (0.1-10 rig/ml). Medium was changed on Day 3. Cells were counted using a hemocytometer on days 1,3,5, and 7. Experiments were routinely performed in triplicate. Anchorage-independent grozuth assay. The effect of EGF and TGF-01 on the anchorage-independent growth of RHE-1A cells was assayed in 24-well trays (2.1 cm’/well). A O.&ml base layer of DMEM containing 10% FBS, 10% tryptose phosphate broth (TPB), and 0.87% agar was covered with 0.5 ml of an overlay of DMEM containing 10% FBS, 10% TPB, 0.35% agar, and 5 X lo3 RHE-1A cells. TGF-Pl (0.1-10 rig/ml) or EGF (l-100 rig/ml) was added to the overlay where appropriate. Experiments were performed in triplicate. Colonies greater than 60 pm in diameter were counted after 2 weeks. The anchorage-independent growth of normal rat kidney cells (NRK-49F) under identical conditions was also measured and served as a positive control. Isolation and characterization of EGF-induced RHE-IA soft agar subclones. RHE-1A cells (passage 24) were seeded in agar as de-

ON

RHE-1A

123

CELLS

106:

105:

1044 0

1

2

3

4

Time

(days)

5

6

7

FIG. 1. Cells were and EGF and 7. (0) 50 rig/ml; terminations.

Effect of EGF on the monolayer growth of RHE-1A cells. seeded at 2 X lo4 cells/well in DMEM containing 10% FBS (1 to 100 rig/ml). Cell counts were performed on days 1,3,5 Without EGF; (0) 1 rig/ml; (A) 5 rig/ml; (A) 10 rig/ml; (0) (m) 100 rig/ml). Points represent the means of triplicate deStandard deviations were less than 10%.

scribed for the anchorage-independent growth assay. EGF (4 rig/ml, half-maximal dose) was added to stimulate RHE-1A cell anchorageindependent growth. On Day 14, several large colonies (2100 pm diam) were selected. These were drawn into the tip of a sterile Pasteur pipet and transferred aseptically to individual wells of a 24-well plate. A small amount of DMEM containing 10% FBS was then added to each well. Within l-2 days, cells began to grow out from the transplanted agar matrix onto the surrounding plastic. Several subclonesdesignated RHE-lA/SAl, RHE-lA/SAB, and RHE-lA/SA3-were propagated for several weeks and maintained in a manner identical to the parental strain (RHE-1A). The ability of these subclones to grow under anchorage-independent conditions was measured in the presence and absence of EGF (4 rig/ml) and compared with the parental strain (RHE-1A). The expression of Factor VIII antigen by these clones was also measured as previously described [15].

RESULTS

EGF stimulated the growth of RHE-1A cells when added to DMEM containing 10% FBS (Fig. 1). While an increase in the rate of RHE-1A cell growth was observed, the most striking feature of EGF growth stimulation was an increase in saturation density displayed by confluent RHE-1A cells when grown in the presence of EGF. This increase was dose-dependent in the range 0.1 to 100 ng/ ml and was half-maximal at 5 rig/ml (Fig. 2). When RHE-1A cells were grown in DMEM containing 10% FBS supplemented with TGF$l, no alteration in growth was observed (Fig. 3). This was unexpected in light of the reported inhibitory effect of TGF-fil on endothelial cell monolayer growth. In order to measure the ability of TGF-Pl to modulate the stimulatory effects of EGF, cells were seeded in DMEM containing 10% FBS and exposed to a combination of EGF (50 rig/ml) and TGF-@l (0.1 to 10 rig/ml). In this concentration range,

MOORADIAN

AND

DIGLIC 800

600

1

1

50.0

q/ml

EGF

plus

500 z i=

600

400

E B

300

E FiT -ii! :

400

% 0

200

200

100

0

0 1

0

5

10

EGF added

50

MS

100

I

0

1

1

0

(ngiml)

TGF-/31 did not block the effects of EGF on RHE-1A cell anchorage-dependent growth or saturation density (Fig. 4). Laug et aE. 1141 have suggested that vascular endotheha1 cells have an inherent capacity for anchorage-independent growth. It was therefore of interest to determine whether RHE-1A cells wouId be responsive to EGF and TGF-fll under anchorage-independent conditions. RHE-1A coIony formation in the presence of 10% FBS alone was very low (less than 0.1% of cells seeded). EGF

I

FES

0.1 TGF-13

FIG. 2. Effect of EGF on the saturation density of RHE-1A cells. Cells were seeded at 2 X lo4 cells/well in DMEM containing 10% FBS and EGF (1 to 100 rig/ml). Counts were performed on Day 7. Data are expressed as the means + one standard deviation of triplicate determinations.

IO4

10%

I

I

I

I

I

I

2

3

4

5

6

7

Time

(days)

10.0

(ngiml)

FIG. 4. Effect of TGF-pl on RHE-1 cell growth stimulation by EGF. Cells were seeded at 2 X 104 cells/well in DMEM comaining 10% FBS and 50 rig/ml EGF. TGF-/31 was also added at 0.1 to 10 rig/ml. Cell counts were performed 7 days later. Data are expressed as the means t one standard deviation of triplicate determinations.

in the range of 0.1 to 100 ng/mI stimulated RHE-1A cell anchorage-independent growth in a dose-dep manner (Fig. 5) and was haIf-ma~maI at 4-5 EGF. The dramatic effect of EGF on RHE-PA cell anchorage-independent growth was rea ily apparent (Fig. 6). The ability of TGF-Bl to modulate EGF-induced RHE-IA cell anchorage-independent growth was then assessed. When experiments were carried out in the presence of 4 rig/ml EGF (haIf-~ax~rnaI dose), TGF-01 was found to be an effective inhibitor ofanchorage-independent growth at 0.1 to 10 ng/mI (Fig. 7). The inhibitory effect of TGF-Pl was ~~~~-~axi~~~ at 1 rig/ml TGF-

,

0

1

5 EGF added

FIG. 3. Effect of TGF-fil on Cells were seeded at 2 X lo* cells/well and TGF-~31 (0.1 to 10 rig/ml). Cell 3, 5 and 7. ((0) Without TGF-01; 1 rig/ml; (A) 5 rig/ml; (A,) 10 rig/ml). triplicate determinations. Standard

1.0 added

RHE-IA cell monolayer growth. in DMEM containing 10% FBS counts were performed on days 1, (0) 0.1 rig/ml; (0) 0.5 rig/ml; (m) Points represent the means of deviations were less than 10%.

.I 0

50

100

(rig/ml)

FIG. 5. EGF-induced RHE-1A cell anchorage-independent Cells were suspended in agar as described under Materials and ods. EGF was present at 1 to 100 rig/ml. Colonies greater than in diameter were counted after 2 weeks. Results are expressed means +- one standard deviation with n = 12.

growth. Meth60 pm as the

EGF/TGF$l

EFFECTS

ON

FIG. 6. Phase-contrast micrographs of RHE-1A cell anchorage-independent (B) colony formation in DMEM + 10% FBS supplemented with 4 rig/ml EGF.

,Lll. Identical experiments were performed using normal rat kidney cells (NRK-49F) whose anchorage-independent growth potential in serum, EGF, and TGF-fl are well established [ 161. NRK-49F cells did not grow under anchorage-independent conditions in the presence of 10% FBS alone, or in the presence of 10% FBS supplemented with EGF or TGF-@l, as would be expected if the FBS used in these experiments contained significant levels of TGF-/3 or EGF, respectively (data not shown). EGF-induced RHE-1A soft agar subclones were isolated to determine whether the phenotypic changes induced by EGF were transient or permanent. Following reinitiation into monolayer culture, each of the subclones were found to have the cobblestone morphology characteristic of normal endothelial cells (Fig. 8). Three such clones, designated RHE-lA/SAl, RHE-lA/SAB,

RHE-1A

125

CELLS

growth in agar. (A) Colony Original magnification, 40X.

formation

in DMEM

+ 10% FBS;

and RHE-lAJSA3, were propagated for further study. These clones were strictly contact-inhibited and no evidence of overcrowding at confluence was evident even after several weeks in culture with twice weekly changes of DMEM containing 10% FBS. In addition, each of these clones continued to express Factor VIII antigen as determined by indirect immunofluorescence (data not shown). However, these subclones were not identical to parental RHE-1A cells. RHE-lA/SA3, for example, exhibited a pattern of growth at confluence reminiscent of the “sprouting” morphology described by Schwartz [ 171 in bovine aortic endothelium. This pattern was not observed in either of the other subclones. The ability of these subclones to grow under anchorage-independent conditions also differed from that of the parental RHE1A cells (Fig. 9). Each of the subclones exhibited a

126

MOORADIAN 600 -q

.-5 5 2

TGF-R

FIG. 7. Effect of TGF-Pl on pendent growth of RHE-1A cells. along with 4 rig/ml EGF. Colonies counted after 2 weeks. Results are dard deviation with n = 3.

added

(ngimi)

the EGF-induced anchorage-indeTGF-fil was added at 0.1 to 5 rig/ml larger than 60 pm in diameter were expressed as the means * one stan-

greater capacity for anchorage-independent growth-in the absence of EGF-than did RHE-1A cells. In addition, clones SAl and SA2 exhibited twofold greater colony formation in the presence of 4 rig/ml EGF than the parent strain, while colony formation by clone SA3 was not significantly different from the parental cell strain. DISCUSSION

This paper describes the effects of EGF and TGF-01 on the growth of rat heart endothelial cells under anchorage-dependent and -independent conditions. We report that the monolayer (anchorage-dependent) growth of rat heart endothelial cells (RHE-1A) grown in DMEM containing 10% FBS increased in a dose-dependent manner in the presence of EGF. These findings are consistent with reports by several other investigators. Gospodarowicz demonstrated that EGF is a potent mitogen for retinal capillary endothelial cells [l]. Takehara [ 111 reported that rat heart endothelial cells express the EGF receptor and are sensitive to the mitogenic effects of EGF. In uiuo, EGF [2] and TGF-(Y (an EGF analog) [3] promote angiogenesis. However, regional differences in endothelial cell sensitivity to the mitogenic effects of EGF do exist. Heimark et al. [8] demonstrated that vascular endothelial cell sensitivity to EGF in vitro was associated with differences in the expression of the EGF receptor. The effect of EGF on the monolayer growth of RHElA cells is probably the result of its direct effect on DNA synthesis and cell division [lS]. However, the experiments we have described were conducted in the presence of serum. Under these conditions RHE-1A cells are al-

AND

DIGLIO

ready growing and the effects of E linked to mitogenesis. For loss of density-dependent gr cells and induces the expre [II, 205 as well. The most striking effect of EGF -lA cell monolayer growth was an increase in ion density (Fig. 3) consistent with a decrease in contact-inhibition of growth. While EGF is known to act in concert with other growth factors to stimulate the anchorage-independent growth of a variety of normal and transformed cells [21], the ability of EGF to stimulate rat, heart endothelial cell anchorage-independent g --to the best knowledge-has not previ been reported. servation that RHE-IA cells can grow under an independent conditions is consisten and co-workers [14] that normal en elial cells were capable of anchorage-independent g h under conditions in which other normal cell types failed to grow. The effects of TGF-P on the gro Lb of vascular endothelial cells in monolayer culture ve been well documented. Heimark et al. [S] demo trated that TGF-0 transiently inhibits DNA syntbesis and migration ofbovine endothelial cells in mono1 culture. FraterSchroder et al. [9], reported that T -p blocks the stimulatory effects of acidic and basic F on bovine aortic endothelial cell monolayer growtk while Baird and Durkin [lo] showed similar effects in bovine adrenal cortexderived endothelial cells in. vitro. In vim, TGF-fi has been shown to induce fibrosis and angiogenesis [6] and to accelerate wound healing [7] 1 TGF-/Yl did not inhibit R cell monolayer 10% FBS alone, growth in DMEM supplement nor did it inhibit the increase in -LA cell saturation density which occurred in the presence of EGF. The discrepency between our data and that of other investigators may be due in large part to differences in culture condition. We have used TGF-/31 at concentrations between 0.1 and 10 rig/ml, a c ce~t~~ti~n range in which maximal effects are observ in a variety of cells [5, 9121. However, the activity of TG by the concentration of other grow in serum. The inhibitory effect of tic endothelial cell DNA synthe et al. [8] was observed in the presence of 10% adult vine serum and disappeared The complement of growth factors presence in adult bovine serum may differ substantially from that of fetal bovine serum. In addition, t~a~~ie~t growth inhibition might not be apparent in growth assays lasting 5-1 days. The inhibitory effect of TGF-fl on rat heart endothelial cell monolayer growth reported by Takehara et al. [II] was observed in the presence of 2.5% S. This concenFBS used here tration of serum is well below the I and may represent a smaller growth stimulatory signal that is more easily overcome by TGI?-@I. In addition,

EGF/TGF$l

EPIG. 8. Photomicrographs of RHE-1A soft agar clones cl01 nes are contact-inhibited and retain normal cobblestone “Sp 8routs” in cultures of clone RHE-lA/SAB. (A) RHE-lA/SAl,

EFFECTS

ON

RHE-1A

CELLS

127

following reinitiation into monolayer culture. Postconfluent RHE-1A soft agar morphology with no indication of overgrowth. Note appearance of endotk relial (B) RHE-lA/SAS, (C) RHE-lA/SA3. Original magnification 200X.

128

MOORADIAN 700 -I

RHE-1A

RHE-lA/SAl

RHE-IAISAE

RHE-IAISAJ

Anchorage-independent growth of RHE-IA soft agar clones. Parental RHE-1A cells and soft agar clones in agar containing 10% FBS t 4 rig/ml EGF. Colonies larger than 60 pm in diameter were counted after 2 weeks. ((m) Without EGF, (U) with 4 rig/ml EGF). Results are expressed as the means k one standard deviation with n = 12.

serum itself contains platelet-derived TGF-P [22]. While platelet-derived TGF-P present in serum is largely inactive [23] and therefore should not significantly influence growth, its potential effects cannot be ignored. We have measured the effects of FBS on the anchorage-independent growth of cells known to be sensitive to EGF and TGF-01 [ 161. Normal rat kidney fibroblasts did not grow under anchorage-independent conditions in DMEM containing 10% FBS alone or in DMEM containing 10% FBS and 4 rig/ml EGF, as would be expected if it contained significant amounts of endogenous TGF-P (data not shown). The endogenous production of growth factors by RHE-1A cells themselves may also play a role in the insensitivity of RHE-1A cells to TGF-fil. Diglio et al. [15] have shown that parental RHE cells-from which RHEIA cells were derived-grow well in low serum concentrations and may produce autocrine growth factors. These may effectively increase the concentration of exogenous TGF-/3 necessary for growth inhibition. EGF has been shown to enhance the transformation of normal cells [24, 251. It was therefore necessary to verify that the effect of EGF on colony formation by RHE-1A cells grown under anchorage-independent conditions was reversible. When soft agar subclones were reinitiated into monolayer culture they retained the cobblestone morphology and contact-inhibition characteristic of normal vascular endothelial cells in uitro. They also continued to express Factor VIII antigen, an important endothelial cell marker [26]. Each clone exhibited an enhanced capacity for anchorage-independent growth in the absence of EGF and one of the three clones studied exhibited enhanced sensitivity to EGF under anchorage-independent conditions. These data suggest that RHE-1.A cells, while clonal in origin, comprise di-

AND

DIGLIQ

verse subpopulations, with di ring sensitivities to EGF. Selection of subclones in agar in the presence of EGF yielded a population enrich in EGF-sensitive cells. Similar selection could be us to isolate populations resistant to the inhibitory elects of TGF-0 and these would be of interest in studies of the molecular mechanisms of TGF-@ on endothe While the mechanism responsi HE-1A cell growth differential effects of TGF-@1 on under anchorage-dependent and -i tions suggest that TGF-PI. may affect anchorage-independent growth via complex mechanisms not directly related to its inhibitory effects on EGF mitogenesis. EGF can stimulate the production of prote ses and protease inhibitors [27] that have been associated with anchorage-independent gxowth &ease production is also subject to modulation -p 1293. Further studies will be needed to determine the mechanism(s) by which these potent growth factors influence RI-BE-IA cell anchorage-independent growth.

1. 2.

Gospodarowicz, D., Brown, K. S., Birdwell, B. R. (1978) J. Cell Biol. 77’,774-1%. Gospodarowicz, D., Bialecki, II., and Thakral, Eye Res. 28,501-514.

T. K. (1979)

Schreiber, A. B., Winkler, 232,X250-1253.

4.

De Larco, J. E., and Todaro, 6. J. (1978) Proc. Natl. Acad. Sci. USA 75,4001-4005. Roberts, A. B., Anzano, M. A., Wakefield, L. M., N. D. F., and Sporn, M. B. (1985) Proc. Acad. Sci. USA 123. Roberts, A. B., Sporn, M. B., Assoian, R. K.; Smith, J. M., Roche, N. S., Wakefield, L. M., Heine, U. I., Liotta, L. A., Falanga, V., Kehrl, J. II., and Fauci, A. S. (1986) Proc. N&E. Acad. Sci. USA 83,4167-4171. Sporn, M. B., Roberts, A. B., Shull, J. H., Smith, J. M., Ward, J. M., and Sodek, J. (1983) Science 219,1329-1331.

6.

7. 8. 9. 10. 11. 12. 13. 14. 15. 16.

R. (1986)

Exp.

3.

5.

M. E., and Derynck,

C. R., and Zetter,

Science

Heimark, R. L., Twardzik, D. R., and Schwartz, S. M. (1986) Science 233,1078-1080. Frater-Schroder, M., Muiler, G., Birchmeier, W., and Bohlen, P. (1986) Biochem. Biophys. Res. Comm. 137(l), 295-302. Baird, A., and Durkin, T. (1986) Biochem. Biophys. Res. Comm. 139(l), 476-482. Takehara, K., LeRoy, E. C., and Grotendorst, G. R. (1987) Cell 49,415-422. Muller G., Behrens, J., Nussbaumer, LJ., Bobien, P., and Bircbmeier, W. (1987) Proc. Natl. Acad. Sci. lJSA 84,5600-5604. Madri, J. A., Pratt, B. M., and Tucker, A. M. (1988) J. Cell. Biol. 106,1375-1384. Laug, W. E., Tokes, Z. A., Benedict, W. F., and Sorgente, N. (1980) J. Cell Bid. 84,281-293. Diglio, C. A., Grammas, P., Giacomeili, F., and Wiener, 9. (1988) Tissue & Cell 20(4), 477-492. Anzano, M. A., Roberts, A. B., Meyers, C. A., Komoriya, A., Lamb, L. C., Smith, J. M., and Spron, M. B. (1982) Cancer Res. 42,4776-4778.

EGF/TGF-fll 17.

Schwartz,

S. M. (1978)

18.

Carpenter, 193-216.

19.

Westermark,

G., and

B. (1976)

S. (1979)

Biochem.

Annu.

Biophys.

Rev.

Biochem.

48,

Res.

Commun.

69,

T. (1984)

Nu-

878-883. 20. Muller,

R., Bravo, R., Burckhardt, ture (London) 312,716-720.

J., and Curran,

D. B., Sporn, M. B., De Fed. Proc. Fed. Amer. Sot.

22. Assoian,

R. K., Komoriya, A., Meyers, C. A., Miller, D. M., and M. B. (1983) J. Biol. Chem. 258(11), 7155-7160.

A. B., Lamb, L. C., Newton, Larco, J. E., and Todaro, G. J. (1983) Exp. Biol. 42,2621-2626.

Received Revised

April 26,1989 version received

September

24. 25.

4,1989

RHE-1A

CELLS

129

L. M., Smith, D. M., Flanders, K. C., and Sporn, M. B. (1988) J. Biol. Chem. 263(16), 7646-7654. Harrison, J., and Auersperg, N. (1981) Science 213,218-219. Reynolds, V. H., Boehm, F. H., and Cogen, S. (1965) Surg. Forum

16,108-109. 26. Jaffe, E. A., Hoyer, 27.

21. Roberts,

Sporn,

ON

23. Wakefield,

14,966-980.

In Vitro

Cohen,

EFFECTS

28. 29.

L. W., and Nachman, R. L. (1973) J. Clin. Invest. 52,2757-2764. Eaton, D. B., and Baker, J. B. (1983) J. Cell Biol. 97,323-328. Cook, J. R., and Chen, J.-K. (1988) J. Cell. Physiol. 136, 188193. Edwards, D. R., Murphy, B., Reynolds, J. J., Whitham, S. E., Dochertz, A. J. P., Angel, P., and Heath, J. K. (1987) EMBO J.

6,1899-1904.