JOURNAL OF CELLULAR PHYSIOLOGY 142:15-20 (1990)

Secretion of a TCF-P-Like Growth Inhibitor by Normal Rat Mammary Epithelial Cells In Vitro STEPHEN P. ETHIER* AND ROCHELLE M. V A N DE VELDE Brea5t Cancer Group, Michigan Cancer Foundation, Detroit, Michigan 4820 7

We have examined conditioned medium (CM) from cultures of normal rat mammary epithelial (RME) cells for growth factor activity on fresh RME cell cultures. RME cell-derived CM contained potent growth inhibitory activity toward fresh RME cell cultures when the medium was acidified by dialysis against 1 % acetic acid prior to concentration. Dialysis of the CM at neutral pH resulted in CM that had growth stimulatory activity and no inhibitory activity. The acid-activated growth inhibitor was heat and acid stable, protease sensitive, and eluted from a Bio-Gel p60 column with a peak of activity in t h e 28 kDa range. Incubation of the acidified-concentrated CM with neutralizing antiserum (affinity purified IgG) against transforming growth factor (TGF)-@completely abolished the inhibitory activity of the CM. Furthermore, RME cell growth in the presence of the growth inhibitor plus TGF-6 antiserum was greater than that observed in growth medium alone. Subsequent experiments demonstrated that addition of TGF-@anti5erum alone to serum-free medium enhanced RME cell growth, whereas addition of n o n i m m u n e IgG was without effect even at 25-fold higher concentrations. Zymographic analysis of RME-CM revealed the presence of plasminogen activator proteases that may mediate the partial activation of the latent growth factor. These results indicate that normal RME cells secrete a latent TGF-@-likegrowth factor into conditioned medium. Furthermore, the results indicate that some of the latent growth factor is activated in situ and contributes to the growth potential of the cells in primary culture in an autocrine manner. Transforming growth factor type beta (TGF-P) is a multifunctional growth factor that is secreted by a variety of cell types both in vivo and in vitro. Among the many biological activities attributed to TGF-P is a potent growth inhibitory activity toward cultured epithelial cells and to a lesser extent fibroblasts (reviewed in Sporn et al., 1987). Cultured epithelial cells that are growth inhibited by TGF-P include bronchial epithelial cells (Masui et al., 19861, epidermal keratinocytes (Shipley et al., 19861, hepatocytes (Carr et al., 1986; McMahon et al., 1986; Braun et al., 19881, and mammary epithelial cells (Silberstein and Daniel, 1987). The potential importance of the growth inhibitory activity of TGF-P is amplified by the finding that many cell types growing in culture synthesize and secrete this growth factor in a latent form. Lawrence et al. (1984, 1985) demonstrated that chicken, mouse, and human fibroblasts secrete a form of TGF-P that is only biologically active after transient acidification. Shipley et al. (1986) demonstrated the presence of TGF-P in conditioned medium of cultured human prokeratinocytes. More recently, Wakefield et al. (1987) demonstrated the presence of latent TGF-(3 in conditioned medium from ten tumor-derived cell lines and one normal cell type (human bronchial epithelial cells). Human breast cancer cell lines also secrete TGF-Plike factors. Knabbe et al. (1987) reported that conditioned medium from the human breast cancer cell line C 1990 WILEY-LISS, INC.

MCF-7 contained TGF-P activity that could directly enhance the growth of NRK cells in soft agar. This result was interpreted to indicate that the growth factor was secreted in a biologically active form and not in the latent form. These workers further postulated a n autocrine role for TGF-P toward MCF-7 cells based on the observation that antiestrogen treatment, which is growth inhibitory for MCF-7 cells, results in increased synthesis of TGF-P. However, Arteaga e t al. (1988) reported that MCF-7 cells obtained from several sources are not growth inhibited by exogenous purified TGF-P. Thus, whereas MCF-7 cells may secrete a TGF-p like growth factor, it is unclear whether this factor affects these cells in a n autocrine manner. In this report we demonstrate that primary cultures of normal rat mammary epithelial (RME) cells growing in serum-free medium secrete a n acid-activated factor with potent growth inhibitory activity toward RME cells. This growth inhibitor shares biochemical properties with TGF-P, and its biological activity is completely abolished by incubation with neutralizing TGF-

Received March 17, 1989; accepted August 14, 1989. *Stephen P. Ethier is now at Department of Radiation Oncology, University of Michigan Medical School, 1331 E. Ann St, Box 0582, Ann Arbor, MI 48109. Address reprint requests there.

16

ETHIER AND VAN DE VELDE

p antiserum. Furthermore, our data indicate that some of the TGF-p in the culture medium is present in a n active form and influences the growth potential of the cells in primary culture. Thus, these results represent the first report that cultured normal mammary epithelial cells secrete a TGF-P like factor that affects RME cell growth in a n autocrine manner. MATERIALS AND METHODS Methods for the preparation of RME cells for primary culture have been described previously (Ethier, 1986). Briefly, the inguinal mammary glands of 45-50-dayold inbred female Lewis rats are removed, minced, and incubated overnight in type I11 collagenase (200 units/ ml) at 37°C in a shaking water bath. The collagenase dissociation is followed by two 15-minute differential sedimentations a t unit gravity followed by a 45-minute incubation in 0.05% Pronase at 37°C. The cells are washed and the epithelial cells as small aggregates are separated from stromal cells by differential centrifugation on a discontinuous Percoll gradient followed by differential plating on tissue culture plastic. The cells are counted and plated onto type I collagen coated tissue culture plates a t a density of 2 x 10” cells per dish. The cells are plated in hormone and growth factor supplemented medium containing 2% FBS to allow cell attachment. Twenty-four hours later the cells are switched to serum-free medium. Also at this time the number of attached cells per plate is determined. The complete serum-free medium consists of Hepesbuffered Ham’s F12 supplemented with bovine serum albumin (BSA, 1mgiml), transferrin (Tf, 5 pgiml), ethanolamine (EA, 5 mM), insulin (5 pgiml), hydrocortisone (HC, 1 pgiml), epidermal growth factor (EGF, 10 ngiml), cholera toxin (CT, 100 ngiml), prolactin (M, 5 pg/ml), progesterone (P, 100 ngiml), triiodothyronine (T3, 10 nM), sodium selenite (Se, 50 nM), gentamycin ( 5 pgiml), and fungizone (0.5 pgiml). The cells are maintained in a humidified atmosphere of 90% air and 10% CO, (media pH = 7.1-7.2) and the medium is changed three times per week.

Preparation of CM RME cells were grown to near confluence, the medium aspirated, and the cells washed three times with calcium and magnesium-free (cmf) Hanks’ balanced salt solution. The cells were incubated in Hepes-buffered Ham’s F12 medium supplemented with BSA, Tf, EA, and HC for 24 hours for conditioning. The conditioned medium was removed, clarified by centrifugation, and dialyzed a t 4°C against 20 volumes of 1% acetic acid for 24 hours with three changes. The CM was then lyophilized to dryness and resuspended in a sufficient volume of phosphate buffered saline (PBS) to yield a 20-fold concentrate. To assay concentrated CM for biological activity, RME cells were seeded into 60 mm culture dishes and allowed to attach overnight. Following attachment, cells were switched to growth factor supplemented serum-free medium. At each feeding 100 ~1 of concentrated CM were added to each dish (5 ml medium per dish) and the cells were grown for 10 days (3 feedings) and compared to cells grown in serum-free medium alone. The number of cells per dish at 24 hours postplating and a t the end of primary culture was deter-

mined by counting isolated nuclei using a Coulter Counter as described previously (Ethier et al., 1987). For experiments involving TGF-P antiserum, RME cells were seeded into 35 mm plates at lo5 cells per dish and allowed to attach overnight. Cells were then switched to growth factor supplemented serum-free medium (2 mlidish) with or without CM and antiserum and grown for 5 days. Trypsin treatment of RME-CM Samples of acidified-concentrated CM were treated with trypsin (100 pgiO.1 ml) for 3 hours at 37°C. The reaction was stopped by addition of soybean trypsin inhibitor (200 pgiO.1 ml). Control samples were treated concurrently with trypsin and trypsin inhibitor for three hours at 37°C. The trypsin-trypsin inhibitor mixture was pre-incubated for 2 hours at room temperature before exposure to the conditioned medium. Gel filtration of RME-CM One milliliter of concentrated CM was neutralized and applied to a 1 x 80 cm Bio-Gel p60 column (bed volume = 65 ml) that was equilibrated and eluted with PBS. The column was run a t a flow rate of 8 mlihour and 5 ml fractions were collected. Individual fractions were concentrated tenfold by lyophilization and reconstitution with deionized water containing 0.1 mg/ml BSA. Zymographic analysis of RME-CM To determine if plasminogen activator activity was

present in RME-CM, 50 pl aliquots of CM were run on a 10% SDS-PAGE gel in the presence of casein and plasminogen. After electrophoresis, the gel was washed three times with 2.5% Triton = X 100 over a 90-minute period t o remove the SDS. The gel was then incubated overnight in 0.1 M glycine pH 8.3 a t 37°C and then stained with Amido Black. Following destaining, clear bands could be visualized against. a blue background that represented areas of proteolysis of casein. Materials Insulin, HC, M, P, EA, Hepes, BSA, and Tf were obtained from Sigma Chemical Company (St. Louis, MO). EGF was obtained from Collaborative Research (Waltham, MA), CT from Schwartz-Mann (Spring Valley, NY), Ham’s F12 and FBS from Hazelton Biological (Lenexa, KS). Gentamycin and fungizone were obtained from Gibco (Grand Island, NY), type I collagen from The Collagen Corporation (Palo Alto, CA), Pronase from Calbiochcm-Behring (La Jolla, CA), and type I11 collagenase from Cooper Biomedical (Malvern, PA). Bio-Gel was obtained from Bio Rad (Richmond, CA), and TGF-P antibodies were obtained from R and D Systems (Minneapolis, MN). RESULTS Preparation of RME-CM To prepare CM for examination of biological activity, primary cultures of RME cells were grown to near confluence in collagen coated 60 mm culture dishes a s previously described (Ethier, 1985). The cells were seeded in 2% serum containing medium for 24 hours to allow cell attachment and then switched to growth fact o r supplemented serum-free medium for 12 days with

17

TGF-S SECRETION BY NORMAL RME CELLS

TABLE 1. Growth inhibitory activity in RME-CM Dialysis medium 1% acetic acid (pH 4.0) 1 m M h e w s (DH 7.2)

Fold-increase in cell No. SF-medium' SF-medium + CM2 45.8 1.4 19.2 31.6

'Cells grown for 10 days in serum-free (SF) medium without concmlrated CM. 'Cells grown for 10 days in serum-free medium plus 100 p1 neutral or acidifiedconcentrated CM per dish.

four changes of media. The cultures were then washed three times with cmf-Hanks' balanced salt solution and incubated for 24 hours in serum-free, growth factor free medium for conditioning. The CM was collected, centrifuged and dialyzed extensively against 1%acetic acid, lyophilized, and reconstituted in a volume of PBS that yielded a 20-fold concentrate. The acidified-concentrated CM was clarified by centrifugation and filter sterilized. Biological activity of RME-CM To examine the biological activity of the RME-CM, 100 ~1 aliquots were added to fresh primary RME cell cultures a t each feeding. The data in Table 1 indicate that addition of acidified-CM had a profound inhibitory effect on RME-cell growth. By contrast, preparation of 20-fold concentrated RME-CM with dialysis a t neutral pH did not yield any inhibitory activity but rather was growth stimulatory. Thus, the growth inhibitory activity present in RME-CM was in a latent form that was activated by acid treatment. Furthermore, neutralization of the acidified-CM did not reverse the growth inhibitory activity (not shown). The acidified-concentrated CM inhibited RME cell growth in a concentration dependent manner as is shown in Figure 1. Experiments were then performed to insure that the latent growth inhibitor present in the CM was secreted by the RME cells. Acidified-concentrated serum-free medium that was not exposed to cells did not contain any growth inhibitory activity. However, acidified-concentrated 2% serum containing medium (attachment medium) did contain potent growth inhibitory activity (not shown). To determine if the growth inhibitory activity found in RME-CM could be accounted for by residual serum-derived inhibitor, a n experiment was performed in which 60 mm collagen coated plates without cells were incubated with serum containing medium for 24 hours and then switched to serum-free medium and changed on the same schedule as parallel cultures containing RME cells. After 12 days, acidified-concentrated CM was prepared from the plates without cells (sham CM) and compared to RME-CM, and the results are given in Figure 2. RME-CM contained potent growth inhibitory activity, and there was no residual inhibitory activity present in the collagen coated plates without cells. Characterization of CM g r o w t h inhibitor The RME-derived growth inhibitor is heat stable and protease sensitive. Incubation of RME-CM for 30 minutes at 56°C or for 3 minutes a t 100°C had no effect on its biological activity (Fig. 3). Trypsin treatment of CM resulted in approximately 30% loss of inhibitory activ-

In 0 F

x I

9

n

o ! 1

0

I

I

200

300

I

0

"lCM

100 .

.

1

2

"

1

'

.

1

4

.

.

6

1

8

Milligrams protein Fig. 1. Concentration dependence of acidified-concentrated CM on RME cell growth in primary culture. RME cells were grown in growth factor supplemented serum-free medium in the presence or absence of varying amounts of acidified-concentrated RME-CM. The amount of protein present at each CM level tested was determined by the BloRad protein assay. The number of cells per dish was determined by counting isolated nuclei.

ity (not shown). To estimate the molecular weight of the RME-derived growth inhibitor, 1 ml of the concentrated CM was neutralized and run on a Bio Gel P-60 column. Column fractions were reconcentrated, filter sterilized, and tested for growth inhibitory activity. Thc RME-derived growth inhibitor eluted a s two peaks of activity, one with a n apparent molecular weight of 28 kDa and a second peak at 100% of the column volume (Fig. 4). This second peak of biological activity is likely to present the same growth inhibitor that has stuck to the column since we have observed similar results following gel filtration of other small polypeptide growth factors. These results indicate that the RME-derived growth inhibitor has biochemical properties similar to those of TGF-p (Moses et al., 1981). Therefore, experiments were performed to determine if neutralizing anti-serum against TGF-P could block the growth inhibitory activity of RME-CM. Aliquots of acidified-concentrated CM were incubated with varying concentrations of affinity purified polyclonal TGFantibodies (IgG) or non-immune serum for 30 minutes at room temperature and then added to RME cultures as usual. The results indicate that TGF-p antibodies abolished the growth inhibitory activity of the RME-CM and the non-immune serum was without effect (Fig. 5). Furthermore, RME-cell growth in the presence of acidified-concentrated CM plus TGF-P antibodies was greater than that observed in the control cultures grown in the absence of the growth inhibitor. This result suggests that neutralizing TGF-P antibodies not only block the inhibitory activity of the exogenous growth inhibitor, but also that of endogenous activity produced in situ. To test this possibility further, experiments were carried out in which RME cells were grown in primary culture in the presence or absence of TGF-P antibodies and the results are shown in Figure 6. The results indicate that purified 'I'GF-P antibodies enhanced RME cell growth in primary culture and sim-

18

ETHIER AND VAN DE VELDE J

ln

9-

;

O v) ’Q

0

*

7

X

7-

I

E P

K

W

5-

3 -I

3-

n u1

0

AF

SF-MEDIUM SHAM-CM

0

RME-CM

Fig. 2. Influence of acidified-concentrated RME-CM versus shamCM on growth of RME cells in primary culture. RME cells were grown for 10 days in growth factor-supplemented serum-free medium or in serum-free medium supplemented with 100 plidish of either acidifiedconcentrated RME-CM or sham-CM. The number of cells per dish was determined by counting isolated nuclei. AF=the number of attached cells 24 hours after plating. Results are the mean number of cells per dish for three dishes plus or minus the standard deviation.

5

15

10

25

20

FRACTION NUMBER

Fig. 4. Gel-filtration chromatography of RME-derived growth inhibitor. Acidified-concentrated RME-CM was neutralized and run on a 1 x 80 cm Bio-Gel p60 column and eluted with PBS. Individual fractions were concentrated tenfold and tested for growth inhibitory activity on individual dishes of fresh RME cell cultures. Each point represents the number of cells per dish for one dish per fraction.

I

2 , ln

2

z

IT

0

r

X

x I

I

cn

P [” K

I

v)

0 K

1

1

W

n

W

n

u)

-I -I

u) -I

-I

W

w u

0

0 0

SF-MED

SF-MED

RME-CM

56

CM

IOOug

50ug

20ug

NIS

100

Fig, 3. Heat stability of RME-derived growth inhibitor. Acidifiedconcentrated RME-CM was heated at 56°C for 30 minutes or at 100°C for 3 minutes and tested for growth inhibitory activity on fresh RME cell cultures. Results are the mean number of cells per dish for three dishes plus or minus the standard deviation.

ilar results were obtained in three separate experiments. To insure that the growth stimulatory effects of the TGF-P antibodies was specific, a n experiment was performed in which RME cells were grown in the presence of non-immune IgG that was purified in a manner similar to that of the TGF-p antibodies. The data in Figure 7 show that this non-immune IgG did not enhance RME cell growth even when present a t 25-fold higher concentrations. Since the TGF-P antibodies only recognize the activated form of the growth factor (Keski-Oja et al., 1987), this result indicates that some of the latent TGF-p-like activity secreted by RME-cells is activated in situ and influences the growth of RME cells in a n autocrine manner. One mechanism by which latent TGF-P may be activated is hy the action of proteases (Lyons et al., 1988). To determine if the RME-CM contained plasminogen activator activity, aliquots of the CM were electrophoresed through casein gels in the presence or ab-

Fig. 5. Effect of TGF-P antiserum on inhibitory activity of acidifiedconcentrated RME-CM. CM was incubated with varying amounts of neutralizing TGF-P antiserum for 30 minutes a t room temperature and then tested for growth inhibitory activity on RME cell cultures. SF-MED =growth in complete serum-free medium (three plates); CM =growth in SF-MED plus acidified concentrated CM (three plates!; 100, 50, 20 kg, = growth in the presence of CM that was incubated with the indicated amount of TGF-B antiserum (one plate each): NIS =growth in the presence of CM that was incubated with 100 kl of non-immune rabbit serum.

sence of plasminogen. Clear bands in the stained gels, indicative of casein proteolysis, were detected in the lanes loaded with CM and no bands were detected in the lane loaded with non-conditioned medium (Fig. 8). Furthermore, no proteolysis was detected in identical gels run in the absence of plasminogen. Thus, plasminogen activator-like proteases are present within RMECM. Keski-Oja et al. (1989) have recently reported substantial activation of latent TGF-p by purified urokinase plasminogen activator. Thus, the results shown in Figure 8 represent a possible mechanism for the activation of the latent TGF-p present in RME-CM.

DISCUSSION The results of these experiments indicate that RME cells secrete a latent TGF-p-like growth factor into CM.

TGF-P SECRETION BY NORMAL RME CELLS

v)

0

3.0

-

?

0

X

v

19

SF-MED rluglrnlab

I

5

0

2.0-

U

W

a 1.0-

0

2

4

6

a

10

DAYS IN CULTURE

Fig. 6. Influence of TGF-p antiserum on RME cell growth in primary culture. RME cells were grown in growth factor supplemented serum-free medium in the presence or absence of 4 pgiml affinity purified TGF-p antiserum.

u)

T

0 7

SF-MED

4

10

20

40

100

Fig. 7. Influence of non-immune IgG on RME cell growth in primary culture. RME cells were grown for 10 days in primary culture in growth factor supplemented serum-free medium in the presence nr absence of varying amounts of affinity purified non-immune rabbit

IgG.

The RME cell-derived factor has potent growth inhibitory activity toward RME cells, is irreversibly activated by acid treatment, is heat stable, protease sensitive, and elutes from a gel filtration column with a peak activity of approximately 28 kDa. The inhibitory activity is completely abolished by incubation with neutralizing TGF-P antibodies. The results also suggest that some of the latent TGF-P-like growth inhibitor secreted by RME cells is activated in sztu. RME cell growth in the presence of affinity purified TGF-P antibodies was consistently enhanced over that observed in growth factor supplemented serum-free medium. Similarly purified non-immune IgG had no effect on cell growth even a t 25-fold higher concentrations. Thus, our data provide direct evidence for a n autocrine activity of the TGF-P secreted by the RME cells. Two recent reports indicate that latent TGF-P can be activated by the action of proteases. Lyons et al. (1988) have demonstrated partial activation of the latent growth factor by plasmin treatment. More recently, Keski-Oja et al. (1989) reported that purified urokinase

Fig. 8. Zymographic analysis of M E - C M . 50 p1 aliquots of RMECM or non-conditioned medium were electrophoresed through casein gels in the presence or absence of plasminogen. Clear bands in the RME-CM lane indicate proteolysis of casein.

plasminogen activator is a potent activator of latent TGF-P. Zymographic analysis of the RME cell conditioned medium obtained in our experiments indicated the presence of plasminogen activator proteases that may mediate the activation of latent TGF-P secreted by RME cells. Active TGF-p has been detected in the conditioned medium of other cell types. Terzaghi and McKeown (1986) demonstrated the presence of a TGFP-like growth inhibitor in CM of normal rat tracheal epithelial cells that had direct activity on partially transformed tracheal cells. In other studies, Knabbe et al. (1987) demonstrated the presence of active TGF-P in

20

ETHIER AND VAN DE VELDE

MCF-7 cell CM following exposure of the cells to antiestrogens. Synthesis and secretion of latent TGF-p appears to be a common property of a wide variety of cultured normal and neoplastic cells of both epithelial and mesenchymal origin. TGF-P has been shown to be secreted by various human breast cancer derived cell lines (Knabbe et al., 1987; Arteaga e t al., 1988). Furthermore, there is indirect evidence that active TGF-P may have a n autocrine influence on growth of MCF-7 cells; however, this point is controversial. The findings with human breast cancer cell lines have been interpreted to indicate a potential role for TGF-P in human breast cancer biology. However, our results indicate that synthesis and secretion of TGF-p-like growth factors is a function of proliferating normal mammary epithelial cells. Thus, a n understanding of the role that this growth factor may play in the growth potential of neoplastic mammary cells must await a more clear understanding of the physiological role of TGF-P in regulating normal cellular proliferation. In summary, cultured normal RME cells secrete a latent TGF-P-like growth factor that, once activated, has potent growth inhibitory activity on RME cells. Furthermore, some of that inhibitory activity appears to be activated in situ and contributes to the growth potential of the cells in a n autocrine manner.

ACKNOWLEDGMENTS The authors gratefully acknowledge the assistance of Dr. Barkley Butler for performing the zymographic analysis of the RME-CM, and to Ms. Kimberly Lyons for technical assistance. This work was supported by PHS grant CA 40064 from the National Cancer Institute.

LITERATURE CITED Arteaga, C.L., Tandon, A.K., Van Hoff, D.D., and Osborn, C.K. (1988) Transforming growth factor-6: Potential autocrine growth inhibitor of estrogen receptor-negative human breast cancer cells. Cancer Res., 48t3898-3904. Braun, L., Mead, J.E., Pazica, M., Mikumo, R., Bell, G.I., and Fausto, N. (1988) Transforming growth factor-p mRNA increases during liver regeneration: A possible paracrine mechanism of growth regulation. Proc. Natl. Acad. Sci. U.S.A., 85:1539-1543. Carr, B.I., Hayashi, I., Branum, E.L., and Moses, H.L. (1986) Inhibi-

tion of DNA synthesis in rat hepatocytes by platlet-derived type-@ transforming growth factor. Cancer Res., 46t2330-2334. Ethier, S.P. 11985) Primary culture and serial passage of normal and carcinogen-treated rat mammary epithelial cells in uitro. J.N.C.I., 74:1307-1318. Ethier, S.P., Kudla, A, and Cundiff. K.C. (1987) The influence of hormone and growth factor interactions on the proliferative potential of normal rat mammary epithelial cells in vitro. J . Cell. Physi d . , f32r161-167. Keski-Oia, J., Lyons, R.M., and Moses, H.L. (1987) Immunodetection and modulatiin of cellular growth with antibodies against native transforming growth factor-p. Cancer Res., 47t6451-6458. Keski-Oja, J., Laiho, M., and Lohi, J. (1989) Activation of latent cell derived transforming growth factor-p by the plasminogen activator urokinase. J. Cell Biol., I07r50a. Knabbe, C., Lippman, M.E., Wakefield, L.M., Flanders, K.C., Kasid, A., Derynk, R., and Dickson, R.B. (1987) Evidence that transforming growth factor-p is a hormonally regulated negative growth factor in human breast cancer cells. Cell, 48r417-428. Lawrence, D.A., Pircher, R., Kryceve-Martinerie, C., and Jullian, P. (1984) Normal embryo fibroblasts release transforming growth factors in a latent form. J. Cell. Physiol., 121t184-188. Lawrence, D.A., Pircher, R., and Jullian, P. (1985) Conversion of a high molecular weight latent P-TGF from chicken embryo fibroblasts into a low molecular weight active p-TGF under acidic conditions. Biochem. Hiophys. Res. Commun., 133r1026-1034. Lyons, R.M., Keski-Oja, J., and Moses, H.L. (1988) Proteolytic activation of latent transforming growth hctor-p from fibroblast conditioned medium. J. Cell Biol., IMr1659-1665. Masui, T., Wakefield, L.M., Lechner, J.F., LaVeck, M.A., Sporn, M.B.: and Harris, C.C. (1986) Type-p transforming growth factor is the primary differentiation inducing serum factor for normal human bronchial epithelial cells. Proc. Natl. Acad. Sci. U.S.A., 83r24382442. McMahon, J.B., Richards, W.L., del Campo, A.A., Song, M.K.H., and Thorgeirsson, S.S. (1986) Differential effects of transforming growth factor-p on proliferation of normal and malignant rat liver epithelial cells in culture. Cancer Res., 46:4665-4671. Moses, H.L., Branum, E.L., Proper, J.A., and Robinson, R.A. (1981) Transforming growth factor production by chemically transformed cells. Cancer Res., 41r2842-2848. Shipley, G.C., Pittelkow, M.R. Wille, J.J.,Scott, R.E., andMoses, R.L. (1986) Reversible inhibition of normal human prokeratinocyte proliferation by type-p transforming growth factor-growth inhihitor in serum-free medium. Cancer Res., 46.2068-2071, Silberstein, G.B., and Daniel, C.W. (1987) Reversible inhibition of mammary gland growth by transforming growth factor-p. Science, 2379-293. Sporn, M.B., Roberts, A.B., Wakefield, L.M., and deCrombrugghe, B. (1987) Some recent advances in the chemistry and biology of transforming growth factor-p. J . Cell Biol., 105:1039-1045. Terzaghi, M., and McKeown, C. (1986) Inhibition of carcinogen altered rat tracheal epithelial cells by normal epithelial cell conditioned medium. Cancer Res., 46:917-921. Wakefield, L.M., Smith, D.M., Masui, T., Harris, C.C., and Sporn, M.B. (1987) Distribution and modulation of the cellular receptor for transforming growth factor-p. J. Cell Biol., 145t965-975.