[CANCER RESEARCH 48, 850-855, February 15, 1988]

Transforming Growth Factors Produced by Normal and NeoplasticallyTransformed Rat Liver Epithelial Cells in Culture1Chi Liu, Ming-Sound Isao,2 and Joe W. Grisham 'Department of Pathology, Montreal General Hospital and McGill University, Montreal, Quebec, H3G 1A4, Canada [C. L., M-S. T.and Department of PathologyUniversity of North Carolina, Chapel Hill, North Carolina 27599-7525 J.W.G.]

ABSTRACT

The secretion of transforming growth factors (TGFs) a and bynormal, chemically transformed, and malignant rat liver epithelial celllines was investigated. The WB-F344 normal cultured rat liver epithelialcell line does not secrete an epidermal growth factor-like (putative!)TGF-a) activity, but several clonal cell strains derived from WB-F344cells which had been treated with JV-methyl-/V'-nitro-AT-nitrosoguanidine,especially those that expressed high levels of y-glutamyltranspeptidase,secreted TGF-a-like activity into their conditioned media. Cell linesobtained from tumors which were produced by these cell strains variedin their abilities to secrete TGF-a, even though they all expressed highlevels of Y-glutamyltranspeptidase activity. When two of the non-TGF-a-secreting tumor cell lines were transplanted into isogeneic rats, thetumors that formed contained high levels of TGF-a-like activity. Although epidermal growth-factor (hence, TGF-a also) inhibited the proliferation of several of these tumor cell lines in monolayer cultures, thisgrowth factor often paradoxically stimulated the anchorage-independentgrowth of the same cell lines. In contrast to TGF-a-like activity, all celllines/strains released TGF-/9 activity into their conditioned media. However, while both normal or chemically transformed cell strains typicallyproduced the inactive form of TGF-/9, the tumor cell lines tended toproduce activated TGF-0 de novo. Anchorage-independent growth of celllines that produced active TGF-0 was either stimulated, inhibited, orunaffected by TGF-/3. Cell lines that were inhibited by TGF-/3 concurrently produced TGF-a which was usually able to overcome the negative"automne" effect of TGF-/3. We conclude that both TGF-a and TGF-0,singly or in combination, are variously involved in the growth of transformed rat liver epithelial cells. TGF-a has a predominantly positiveautocrine action on the growth of rat liver epithelial tumor cell lines. The"paracrine" effect of TGF-/2 may be at least as important as its autocrine

effect in the growth of these transformed epithelial cell lines.

INTRODUCTION

TGFs can be defined operationally as growth factors whichconfer the transformed phenotype (anchorage-independentgrowth) on untransformed (anchorage-dependent) fibroblasticindicator cells in vitro (1, 2). TGF was first isolated from theconditioned medium of murine sarcoma virus-transformedmouse fibroblasts and initially was called sarcoma growth factor(3). Subsequent studies using high-performance liquid chro-matography demonstrated that sarcoma growth factor actuallyconsisted of two different TGFs, a and (4). While TGF-a,which is homologous in its structure and action to EGF, ispredominantly a growth-stimulatory polypeptide (5), severalinvestigators have shown that TGF-0 inhibits the anchorage-dependent and/or -independent growth of most cultured cells,

Received 5/28/87; revised 11/6/87; accepted 11/16/87.The costs of publication of this article were defrayed in part by the payment

of page charges. This article must therefore be hereby marked advertisement inaccordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1Supported by grants from Medical Research Council of Canada [MA-9595(M. S. T.)] and the NIH [CA29323 (J. W. G.)].2Scholar of the Medical Research Council of Canada.

3To whom requests for reprints should be addressed.*The abbreviations used are: TGF, transforming growth factor; EGF, epider

mal growth factor; GGT, 7-glutamyl transpeptidase; IMEM-ZO, improved minimal medium with zinc salt; HEPES, 4-(2-hydroxyethyl)-l-piperazineethane sul-fonic acid; MNNG, W-methyl-W-nitro-yV-nitrosoguanidine; CFE, colony-formingefficiency.

especially epithelial cells (1,6, 7). Furthermore, since most ifnot all normal and neoplastically transformed cells synthesizeand secrete TGF-0 (1, 2, 8), as well as express the surfacereceptor (1, 9, 10) to this growth factor, the role of TGF-/3 inthe regulation of the growth of normal and neoplastic cells isunclear.

Cultured rat liver epithelial cells have been used extensivelyto study events and mechanisms in neoplastic transformationof epithelial cells (11-16). We have previously reported thatEGF (hence, also TGF-a) can induce or enhance the colony-forming ability in soft agar of chemically treated or transformedrat liver epithelial cells (17). In fact, the emergence of EGF-induced anchorage-independent growth capacity is the earliestand possibly the most reliable phenotype to indicate the imminent tumorigenic transformation of cultured rat liver epithelialcells. On the contrary, TGF-/8 inhibits DNA synthesis andproliferation of normal cultured rat liver epithelial cells (18,19), but TGF-/3 does not inhibit the proliferation of an alia toxinB,-transformed cell line ( 18). TGF-/3 neither induces anchorage-independent growth of normal rat liver epithelial cells noraffects the growth of aflatoxin B,-treated cells in soft agar (18).Differential sensitivity of malignant versus normal cells to theinhibitory effect of TGF-/3 has been proposed to be one of themechanisms by which TGF-exerts a role in tumorigenesis(18, 20). In this study we analyze the production of TGFs bynormal rat liver epithelial cell lines, by clones of chemicallytransformed rat liver epithelial cells, and by cell lines derivedfrom tumors produced by the transformed epithelial cells, andwe examine the roles of these growth factors in the "autocrine"regulation of the growth of these cells.

MATERIALS AND METHODSTissue culture plates were obtained from Falcon (Becton-Dickinson,

Mississauga, Ontario). Richter's IMEM-ZO was obtained from IrvineScientific (Irvine, CA). Fetal bovine serum and receptor grade EGFwere from Gibco (Burlington, Ontario). TGF-/was purchased from R& D (Minneapolis, MN). [3H]Methylthymidine and 125Iwere, respectively, from 1CN (Montreal, Quebec) and Amersham Canada Ltd.(Oakville, Ontario).

Cell Lines. The establishment and phenotypic characteristics of allof the cell lines used have been described previously. Briefly, the WB-F344 normal rat liver epithelial cell line was isolated from an adultmale Fischer 344 rat (21). This cell line was tumorigenically transformed by eleven episodic treatments with 5 Mg/ml of MNNG (16).Clonal cell strains (GGT-positive and -negative strains) were isolatedfrom this transformed cell line based on their expression of GGTenzyme (22). These cell strains were transplanted into 1-day-old isogeneic rats for tumorigenicity testing, and tumor cell lines were isolatedfrom some of the tumors formed (23). Tumor cell lines were namedaccording to the cell strains from which they were derived, i.e., GP6TBoriginated from one of the tumors produced in a rat injected with GP6cells. To induce second generation tumors, these tumor cell lines weretransplanted into the s.c. tissue of newborn Fischer 344 rats and 4-6wk later the tumors were harvested and stored at -80C prior toextraction of TGFs.

Conditioned Medium. Cells were grown in 20 100-mm tissue cultureplates in Richter's IMEM-ZO containing 10 mM HEPES and 10%

850

Research. on December 2, 2014. 1988 American Association for Cancercancerres.aacrjournals.org Downloaded from

TGFS PRODUCED BY CULTURED RAT LIVER EPITHELIAL CELLS

fetal bovine serum. After the cultures had become confluent, they werewashed once with Ca2+-Mg2+-free Hanks' balanced salt solution andthen incubated in 10 ml of serum-free IMEM-ZO. One day later, themedium on each plate was discarded and replaced by 10 ml of freshserum-free medium. The conditioned medium was collected every 2 to3 days until a total volume of 1 liters was accumulated from eachculture. Collected media were always stored frozen at -20C.

Each batch of conditioned medium was concentrated approximately20 times using the Amicon CH2 Hollow Fiber concentrator with amolecular-weight cutoff of 3500. The concentrated medium was extensively dialyzed with 1% (v/v) acetic acid solution using Spectrophormembrane molecular-weight cutoff of 3500) and further concentratedwith Amicon ullnifilt ration cells to a volume of approximately 10 ml.After ultracentrifugation at 100,000 x g, the supernatant was sterilizedby passage through 0.2 urn membrane filter and then lyophilized. Thepowder was solubilized in 4 IHM HC1 to a final concentration ofapproximately 2 mg/ml.

Extraction of TGFs from Tumor Tissue. This procedure was performed exactly according to Roberts et al. (24). The final extract waslyophilized to dryness and reconstituted with 4 HIM HC1 to a finalconcentration of 2 mg/ml.

Bioassay of TGFs. TGFs were assayed using NRK-c49 cells (American Type Culture Collection, Rockville, MD) as the indicator cells.The soft agar medium was prepared according to DeLarco and Todaro(3) with two modifications. The medium used was Richter's IMEM-ZO with or without insulin at 4 Mg/ml, and the top agar layer was 0.5%rather than 0.3%. IMEM-ZO containing HEPES buffer allows a bettercontrol of the pH of the agar medium than does Dulbecco's modifiedEagle's minimal essential medium, and the 0.5% top agar layer alsofacilitates handling of plates and counting of colonies. Our preliminaryresults have shown that increased consistency of the agar does not altersignificantly the CFEs of NRK-C49 cells in the presence or absence ofEGF.! A total of 100 /gof protein from processed conditioned mediumwas added to each plate, colonies were allowed to form during culturefor 14 days, and the colonies were stained overnight at 37"C"with asolution containing /3-NADH and nitroblue-tetrazolium at concentrations of 0.5 mg/ml each. Colonies were counted with a dissectingmicroscope fitted with an ocular micrometer. Control plates withoutadded protein prepared from conditioned media were always performedin parallel in each experiment. For the assay of TGF-/3, EGF was alsoadded to the agar medium at a final concentration of 2 ng/ml.

The assay for anchorage-independent growth of rat liver epithelialcells has been described previously (12). In most cases, plates wereincubated for 2 to 4 wk before the colonies were stained and counted.

'"I Labeling of EGF. lodo-Gen (Pierce Chemical Co., Rockford, IL)was used to catalyze the iodination reaction. The I2SI labeling wasperformed according to the instructions of the company. Five tgofEGF were reacted with 500 //Ci of I25Ifor 30 min. The unbound '"Iwas removed by elution through a prepacked disposable ExcelluloseGF-5 desalting column (Pierce) with phosphate-buffered saline containing 0.1% bovine serum albumin. The fractions were checked for '"I-EGF by assaying specific binding to A-431 cells in 24-well cultureplates. The fractions containing bioactive '"I-EGF were pooled andstored frozen in the eluting buffer at a specific activity of approximately20,000 cpm per mg of '"I-EGF.

Competitive I2!I-EGF Binding. This assay was performed accordingto Lin et al. (25). Approximately IO5 A-431 cells were plated in eachwell of 24-well tissue culture plates, and binding was evaluated 1-2days later. Cells were rinsed 3 times with ice-cold serum-free Dulbecco'smodified Eagle's minimal essential medium containing 0.1% bovineserum albumin and 50 m\i HEPES, pH 7.4, and then incubated with0.5 ml of this medium for 45 min at 4C.Subsequently, 200 gofprotein derived from conditioned media or 1 igof unlabeled EGF (tomeasure nonspecific binding) were added. A total of 20,000 cpm of'"I-EGF was then added and incubated further for 2 h at 4C,at whichtime the medium was removed and cells were rinsed twice with 2 ml ofice-cold phosphate-buffered saline containing 0.1% bovine serum albumin. Cells were then solubilized in 1 ml of l N NaOH at 37Cfor

18-24 h and the solution was counted in a gamma counter (Nuclear,Chicago, IL).

DNA Synthesis. The effects of EGF and TGF-/3 on the proliferationof rat liver epithelial cell lines cultured in monolayers were measuredby incorporation of [3H]thymidine into the acid-precipitable fraction ofcells. Cells (5 x IO4) were plated in each replicate of 60-mm tissueculture plates in IMEM-ZO containing 10% fetal bovine serum. Twodays later, the medium was changed to fresh medium containing EGF(10 ng/ml) and/or TGF-/3 (1 ng/ml). [3H]Thymidine incorporation wasperformed as described previously (26).

RESULTS

Effect of Insulin on CFE of NRK-c49 Cells. Insulin alone didnot induce anchorage-independent growth of NRK-c49 cellsnor did it significantly increase the CFE of NRK cells inducedby EGF (Table 1). However, insulin almost doubled the CFEof NRK cells in soft agar containing EGF and TGF-0. Hence,insulin increases the sensitivity of the assay for detection ofTGF-,but this hormone does not significantly alter the resultson the bioassay for EGF-like (or TGF-a) growth factor. In mostof the experiments described below, except when otherwisespecified, IMEM-ZO with insulin was used as the medium inwhich to grow NRK-c49 cells in soft agar.

Production of TGFs by Transformed and Nontransformed RatLiver Epithelial Cells. Table 2 shows the activities of TGFs inmedia conditioned by parental normal WB-F344 cells, thetransformed cell strains clonally derived from the parental lineafter it was treated with MNNG, and the corresponding tumorcell lines. All of the cell strains and lines tested, including thewild-type WB-F344 cells, produced TGF-/3. The conditionedmedium in which WB-F344 cells were cultured did not containEGF-like activity. Among the clonal cell strains, most of theGGT-positive strains produced EGF-like (TGF-a) activity,whereas most of the GGT-negative strains did not. There wasno quantitative correlation between the specific activity of GGTexpressed by these cells and the level of TGF-a-like activitypresent in their conditioned media. The tumor cell lines, whichwere all GGT positive, varied in their secretion of TGF-a-likeactivity. The results demonstrate that while TGF-/3 is producedubiquitously by transformed and nontransformed rat liver epithelial cells, TGF-a is not. However, while most GGT-positivecells excrete TGF-a-like activity into the medium, some, including some tumor cell lines, did not.

Since cultured GN3TG and GN6TF cells excreted a lowamount or no TGF-a-like activity, we determined whether theyalso produced TGF-a-like activity when they formed tumors invivo. Table 3 shows that tumors formed both by GN3TG andGN6TF cells contained significant amounts of TGF-a-like activity. In fact, when the acid-ethanol extracts of cultured tumorcells were tested for EGF-like activity, all three cell lines whichwere poor or nonsecretors of TGF-a contained significantamounts of cell-associated EGF-like/TGF-a activity (Table 4).

Nature of TGF-/3 Produced by Cells. In our initial experiments

Table 1 Effect of insulin on the colony-transforming efficiency ofNRK-c49 cellsTwo thousand NRKc49 cells were suspended in agar medium containing

IMEM-ZO with or without 4 ig/mlof insulin and in the presence or absence of2 ng/ml EGF and/or 200 pg of TGF-/3. Colonies were allowed to form in 14days. Values represent the mean SD of triplicate plates.

5 Unpublished data.

GrowthfactorNone

EGFEGF + TGF-/3NoneEGFEGF + TGF-/3Insulin

100iim0

1 266 3

01 1

15319Total041

12295 7

056 2

578 98851

Research. on December 2, 2014. 1988 American Association for Cancercancerres.aacrjournals.org Downloaded from

TGFS PRODUCED BY CULTURED RAT LIVER EPITHELIAL CELLS

Table 2 TGF activities in conditioned media ofnontransformed, transformed,and malignant tumor cell lines of rat liver epithelial cells

Table 5 Effect of acid treatment on the activities of transforming growth factorsin media of cultured rat liver epithelial cell lines

EGF-like (TGF-)activity TGF-ff activityConditioned No. of colonies inmedia added absence of EGF

100fmidine into acid-precipitable fraction was assayed. Valuesrepresent the mean SD of the triplicate plates.

['HIThymidine incorporation (x 10 ')

Control EGF TGF-0 EGF -I-TGF-piWB-F344-PIO33.0WB-F344-P2351.0GP8TB6.3GN3TG21.1GN5TA9.9GN6TF7.3GP6TB

31.53.7

22.84.4103.90.67.12.215.1O.I10.00.84.01.8

27.20.6"

5.6 0.33.544.6 2.21.0

5.90.31.112.51.10.68.9 0.20.6" 3.1 1.11.1

24.1 1.514.6

1.551.45.16.40.19.50.69.2

0.71.50.121.92.0

F344 cells and three of the transformed cell strains (GN1, GP8,GP9) secreted an inactive form of TGF-/3 which could beactivated by acid treatment, whereas all three tumor cell lines(GN3TG, GP6TB, GN6TF) tested produced TGF-/3 that wasactive de novo.

Effect of EGF and TGF-/8 on Anchorage-dependent Proliferation. Depending on the passage number, the proliferation ofnormal rat liver epithelial cells may be either inhibited orstimulated by EGF (Table 6). However, TGF-/3 inhibited theproliferation of both early- and late-passage normal rat liverepithelial cells, but the latter were affected to a much lesserdegree. The enhanced growth-stimulatory effect of EGF onlate-passage WB-F344 cells was abolished by TGF-/3 (Table 6).The anchorage-dependent proliferation of two of the tumor celllines (GP8TB and GN5TA) was not affected by either EGF orTGF-/3, but in three other cell lines (GN3TG, GN6TF, andGP6TB) both EGF and TGF-0 suppressed their growth.

Effect of EGF and TGF-/? on Anchorage-independent Proliferation. The effect of growth modulators on the proliferationof rat liver epithelial cells may be altered by the extracellularsubstrate on or in which they grow (29). Therefore we alsostudied the effect of EGF and TGF-/3 on the proliferation insoft agar of some of these tumor cell lines (Table 7). Neither

852

Research. on December 2, 2014. 1988 American Association for Cancercancerres.aacrjournals.org Downloaded from

TGFS PRODUCED BY CULTURED RAT LIVER EPITHELIAL CELLS

Table 7 Effect ofEGF and TGF-0 on the colony-forming ability in soft agar of some tumor cell linesApproximately IO3 to 5 x IO3cells/plate of these tumor cells were suspended in 0.5% agar media containing IMEM-ZO with insulin and 10% fetal calf serum.

The experiment was performed on triplicate plates with addition of either EGF (2 ng/ml) or TGF-/3 (0.2 ng/ml) or EGF (2 ng/ml) plus TGF-/3 (0.2 ng/ml). Thecontrol plates contained no growth factors. Colonies of 100 pm in diameter were counted in 14 days. The values represent the mean SD.

CelllineGN3TGGN5TAGN6TFGP6TBNo.

ofcellsper Factor(s)

plateaddedIO3NoneEGFTGF-,3EGF

+TGF-/3IO3NoneEGFTGF-/3EGF

+TGF-/3IO3NoneEGFTGF-/3EGF

+TGF-/35x IO3 NoneEGFTGF-/3EGF

+ TGF-/J100/jm3031196417531465185120161107119941515241541793313217232

160176Total170

166708925950594

51169+25325295012274333572539152270

122744686

33317010

237920P

value

TGFS PRODUCED BY CULTURED RAT LIVER EPITHELIAL CELLS

induced in NRK-c49 cells by protein derived from conditionedmedia. This correlation strongly suggests that the EGF-likeactivity observed in the conditioned media is due to TGF-a.Some of the conditioned media which showed EGF-like activityin the bioassay and/or radioreceptor assay also gave positiveresults with the radioimmunoassay for TGF-a (BioTope, Belle-vue, WA) (data not shown). Six of nine (67%) tumor cell linesestablished from the various clonal strains of carcinogen-treatedrat liver epithelial cells also secreted into the medium a TGF-a-like activity. In three tumor cell lines which secreted no orvery small amounts of TGF-a-like substance in culture, TGF-a-like activity was detected in the tumors which grew in vivoafter the transplantation of these cells into isogeneic rats and/or in the acid-ethanol extracts of the cultured tumor cells. Theseobservations reflect the inability of these cells to excrete theTGF-a, rather than a lack of synthesis of this growth factor bythe transformed cells in culture. Derynck et al. (8) have alsorecently reported that some cultured human tumor cells synthesized TGF-a but were not able to excrete this growth factorin culture medium. The significance of nonsecreted cytoplasmicTGF-a is currently unknown.

In several tumor cell lines, the proliferative effects of EGFor TGF-a and TGF-differed depending on whether the cellswere growing anchorage dependently or independently. Similarfindings also have been reported in rat liver epithelial cell linestreated with retinoic acid (29). At this time it is unclear whichgrowth condition better reflects the in vivo responsiveness oftumor cells. Nevertheless, it is important to note that in thetumorigenic rat liver epithelial cell lines studied, EGF (andTGF-a) was mainly a mitogen, whereas TGF-/3 might eitherstimulate or inhibit the anchorage-independent growth of thesecells. However, in none of these cell lines was the overall effectof the protein extracted from media which they conditionednegative in terms of growth effect (i.e., an autocrine inhibitionof cell growth was not observed).

The results of this study provide important insights concerning the physiological roles of TGFs produced by normal andmalignant cells. Although most normal cells which have beenstudied produce TGF-/3, this growth factor is secreted in a latentor inactive form and, hence, exerts no significant biologicaleffect. This study shows that epithelial tumor cells often secreteactive TGF-0 and in some instances this substance may exertan autocrine growth-stimulatory effect. In other instances whenthe cells are themselves susceptible to inhibition by TGF-/S,TGF-a may also be produced and override the effect of TGF-,resulting in a positive overall autocrine effect. In such casesand in instances in which the cells that secrete active TGF-/3are themselves insensitive to its growth effect, the main physiological role of TGF-|8 may lie in its "paracrine" function. Theparacrine effects of TGFs may be at least as important, if notmore so, as their autocrine effects in the regulation of malignantneoplastic growth. These paracrine effects include (a) desmo-plasia as a result of a stimulation of the proliferation of fibro-blasts and synthesis of collagen by TGF-a and/or TGF-/3 (32,33), (/>) induction of angiogenesis and neovascularization byTGF-a (34), (c) suppression of proliferation of surroundingnormal cells by TGF-/S to cause a differential growth ratebetween neoplastic and normal cells, and (d) alteration of hostimmune response to tumor cells by the suppressive action ofTGF-/J on lymphocytes (35).

ACKNOWLEDGMENTSWe wish to thank Wendy J. Segall and D. O'Connor for their

excellent secretarial assistance and Diana Lee for some technical help.

REFERENCES

1. Roberts, A. B., and Sporn, M. B. Transforming growth factors. Cancer Surv.,4: 683-705, 1985.

2. Goustin, A. S., Leof, E. B., Shipley, G. D., and Moses, H. L. Growth factorsand cancer. Cancer Res., 46: 1015-1019, 1986.

3. De Larco, J. E., and Todaro, G. J. Growth factors from murine sarcomavirus-transformed cells. Proc. Nati. Acad. Sci. USA, 75:4001-4005, 1978.

4. Anzano, M. A., Roberts, A. B., Smith, J. M., Sporn, M. B., and De Larco,J. E. Sarcoma growth factor from conditioned medium of virally transformedcells is composed of both type a and type transforming growth factors.Proc. Nati. Acad. Sci. USA, 80:6264-6268, 1983.

5. Carpenter, G., and Cohen, S. Epidermal growth factor. Ann. Rev. Biochem.,48: 193-216, 1979.

6. Roberts, A. B., Anzano, M. A., Wakefield, L. M., Roche, N. S., Stern, D.F., and Sporn, M. B. Type beta transforming growth factor: a bifunctionalregulator of cellular growth. Proc. Nati. Acad. Sci. USA, 82:119-123, 1985.

7. Sporn, M. B., and Roberts, A. B. Autocrine growth factors and cancer.Nature (Lond.), 313: 745-747, 1985.

8. Derynck, R., Goeddel, D. V., Ullrich, A., Gutterman, J. U. Williams, R. D.,Bringman, T. S., and Berger, W. H. Synthesis of messenger RNAs fortransforming growth factors a and and the epidermal growth factor receptorby human tumors. Cancer Res., 47: 707-712, 1987.

9. Tucker, R. F., Branum, E. L., Shipley, G. D., Ryan, R. J., and Moses, H. L.Specific binding to cultured cells of 125I-labeledtype beta transforming growthfactor from human platelets. Proc. Nati. Acad. Sci. USA, 81: 6757-6761,1984.

10. Massagu,J., and Like, B. Cellular receptors for type-|8 transforming growthfactor. J. Biol. Chem., 260:2636-2645, 1985.

11. Williams, G. M., Elliot, J. M., and Weisburger, J. H. Carcinoma aftermalignant conversion in vitro of epithelial like cells from rat liver followingexposure to chemical carcinogens. Cancer Res., 33: 606-612, 1973.

12. Montesano, R., Saint Vincent, L., Drevon, C., and Tomatis, L. Productionof epithelial and mesenchymal tumors with rat liver cells transformed invitro. Int. J. Cancer, 16: 550-558, 1975.

13. Schaeffer, W., and Heintz, N. H. A diploid rat liver cell culture. IV: Malignanttransformation by Aflotoxin Bl. In Vitro, 14:418-427, 1978.

14. Brown, J. D., Wilson, M. J., and Poirier, L. A. Neoplastic conversion of ratliver epithelial cells in culture by ethionine and s-adenosylethionine. ( 'ircinogenesis (Lond.), 4: 173-177, 1983.

15. Shimada, T., Furukawa, K., Kreiser, D. M., Cawein, A., and Williams, G.M. Induction of transformation by six classes of chemical carcinogens inadult rat liver epithelial cells. Cancer Res., 43: 5087-5092, 1983.

16. Tsao, M.-S., Grisham, J. W., Nelson, K. G., and Smith, J. D. Phenotypicand karyotypic changes induced in cultured rat hepatic epithelial cells thatexpress the "oval" cell phenotype by exposure to A'-methyl-W-nitro-A'-nitrosoguanidine. Am. J. Pathol., 18:306-315, 1985.

17. Tsao, M.-S., Earp, H. S., and Grisham, J. W. Gradation of carcinogen-induced capacity for anchorage-independent growth in cultured rat liverepithelial cells. Cancer Res., 45: 4428-4432, 1985.

18. McMahon, J. B., Richards, W. L., del Campo, A. A., Song, M.-K. H., andThorgeirsson, S. S. Differential effects of transforming growth factor-/3 onproliferation of normal and malignant rat liver epithelial cells in culture.Cancer Res., 46:4665-4671, 1986.

19. Lin, P., Liu, C., Tsao, M.-S., and Grisham, J. W. Inhibition of proliferationof cultured rat liver epithelial cells at specific cell cycle stages by transforminggrowth factor-/3. Biochem. Biophys. Res. Commun., 143:26-30, 1987.

20. Shipley, G. D., Pittelkow, M. R., Wille, J. J. Jr., Scott, R. E., and Moses,H. L. Reversible inhibition of normal human prokeratinocyte proliferationby type-/3transforming growth factor inhibitor in serum-free medium. CancerRes., 46: 2068-2071, 1986.

21. Tsao, M.-S., Smith, J. D., Nelson, K. G., and Grisham, J. W. A diploidepithelial cell line from normal adult rat liver with phenotypic properties of"oval" cells. Exp. Cell Res., 154: 38-52, 1984.

22. Tsao, M.-S., Grisham, J. W., Chou, B. B., and Smith, J. D. Clonal isolationof populations of GGT-positive and GGT-negative cells from rat liver epithelial cells chemically transformed in vitro. Cancer Res., 45: 5134-5138,1985.

23. Tsao, M.-S., and Grisham, J. W. Phenotypic modulation during tumorigen-esis by clones of transformed rat liver epithelial cells. Cancer Res., 47:1282-1286, 1987.

24. Roberts, A. B., Lamb, L. C., Newton, D. L., Sporn, M. B., de Larco, J. E.,and Todaro, G. J. Transforming growth factors: isolation of polypeptidesfrom virally and chemically transformed cells by acid/ethanol extraction.Proc. Nati. Acad. Sci. USA, 77:3494-3498, 1980.

25. Lin, Q., Blaisdell, J., O'Keefe, E., and Earp, H. S. Insulin inhibits theglucocorticoid-mediated increase in hepatocyte EGF binding. J. Cell Physiol.,119:267-272, 1984.

26. Tsao, M.-S., Saunders, G. H. S., and Grisham, J. W. Regulation of growthof cultured hepatic epithelial cells by transferrin. Exp. Cell Res., 171:52-62,1987.

27. Lawrence, D. A., Pircher, R., Krycve-Martinerie,C., and Jullien, P. Normalembryo fibroblasts release transforming growth factors in a latent form. J.Cell Physiol., 121: 184-188, 1984.

28. Lawrence, D. A., Pircher, R., and Jullien, P. Conversion of a high molecularweight latent /3-TGF from chicken embryo fibroblasts into a low molecularweight active /3-TGF under acidic conditions. Biochem. Biophys. Res. Commun., 33:1026-1034, 1985.

29. Tsao, M.-S., Blaisdell, J., Chou, B. B., Earp, H. S., and Grisham, J. W.Inconsistent effects of M-trans retinoid acid on different clones of chemicallytransformed rat liver epithelial cells. Carcinogenesis (Lond.), *: 497-501,1987.

854

Research. on December 2, 2014. 1988 American Association for Cancercancerres.aacrjournals.org Downloaded from

TGFS PRODUCED BY CULTURED RAT LIVER EPITHELIAL CELLS

30. Massagu,.1.. Kelly, H . and Mottola, C. Stimulation by insulin-like growth Fauci, A. S. Transforming growth factor type 0: rapid induction of fibrosisfactors is required for cellular transformation by type-0-transforming growth and angiogenesis in vivoand stimulation of collagen formation in vitro. Proc.factor. J. Biol. Chem., 260:4551-4554, 1985. Nati. Acad. Sci. USA, S3:4167-4171, 1986.

31. Yamaoka, K., and Hirai, R. Characterization of intra- and extracellular 34. Schreiber, A. II., Winkler, M. E., and Derynck, R. Transforming growthtransforming growth factors from normal and avian sarcoma virus-trans- factor-:a more potent angiogenic mediator than epidermal growth factor,formed rat cells. Biochem. Intern., 13: 863-872, 1986. Science (Wash. DC), 232:1250-1253, 1986.

32. Ignotz, R. A., and Massagu,J. Transforming growth factor-/^ stimulates the 35. Kehrl, J. F!.. Wakefield, I.. M.. Roberts, A. B., Jakowlew. S., Alvarez-Mon,expression of fibronectin and collagen and their incorporation into the M., Derynck, R., Sporn, M. B., and Fauci, A. S. Production of transformingextracellular matrix. J. Biol. Chem., 261:4337-4345, 1986. growth factor-0 by human T lymphocytes and its potential role in the

33. Roberts, A. B., Sporn, M. B.. Assoian, R. K., Smith, J. M., Rche,N. S., regulation of T cell growth. J. Exp. Med., 163: 1037-1050, 1986.Wakefield, L. M., Heine, V. I., Liotta, L. A., Falanga, V., Kehrl. J. H., and

855

Research. on December 2, 2014. 1988 American Association for Cancercancerres.aacrjournals.org Downloaded from

1988;48:850-855. Cancer Res

Chi Liu, Ming-Sound Tsao and Joe W. Grisham

Neoplastically Transformed Rat Liver Epithelial Cells in CultureTransforming Growth Factors Produced by Normal and

Updated version

http://cancerres.aacrjournals.org/content/48/4/850Access the most recent version of this article at:

E-mail alerts related to this article or journal.Sign up to receive free email-alerts

SubscriptionsReprints and

.pubs@aacr.orgDepartment atTo order reprints of this article or to subscribe to the journal, contact the AACR Publications

Permissions

.permissions@aacr.orgDepartment atTo request permission to re-use all or part of this article, contact the AACR Publications

Research. on December 2, 2014. 1988 American Association for Cancercancerres.aacrjournals.org Downloaded from