Jan 24, 1990 - NANCY E. HYNES,'* DANIELA TAVERNA,1 INA MARIA HARWERTH,1 ...... Bates, S. E., N. E. Davidson, E. M. Valverius, C. E. Freter, R. B..
Vol. 10, No. 8
MOLECULAR AND CELLULAR BIOLOGY, Aug. 1990, p. 4027-4034
0270-7306/90/084027-08$02.00/0 Copyright C 1990, American Society for Microbiology
Epidermal Growth Factor Receptor, but Not c-erbB-2, Activation Prevents Lactogenic Hormone Induction of the 1-Casein Gene in Mouse Mammary Epithelial Cells NANCY E. HYNES,'* DANIELA TAVERNA,1 INA MARIA HARWERTH,1 FORTUNATO CIARDIELLO,2 DAVID S. SALOMON,2 TADASHI YAMAMOTO,3 AND BERND GRONER' Friedrich Miescher Institute, P.O. Box 2543, 4002 Basel, Switzerland'; Laboratory of Tumor Immunology and Biology, National Cancer Institute, National Institutes of Health, Bethesda, Maryland 208922; and Department of Oncology, Institute of Medical Science, University of Tokyo, Tokyo 108, Japan3 Received 24 January 1990/Accepted 30 April 1990
The HCll cell line was isolated from mammary gland cells of pregnant mice. The cells displayed a normal phenotype and retained some characteristics of mammary epithelial cell differentiation. After treatment with the lactogenic hormones prolactin and glucocorticoids, the HCll cells expressed the milk protein (-casein. Various oncogenes were transfected and expressed in HCll cells. The oncogenes were tested for their transformation ability and for their effects upon the differentiation of the HCll cells. All of the oncogenes tested, including activated human Ha-ras, human transforming growth factor-a, activated rat neuT, and human c-erbB-2 activated by a point mutation in the transmembrane domain, caused transformation of the HCll cells, as shown by tumor formation in nude mice. HCll cells expressing the neuT and activated c-erbB-2 genes synthesized I-casein in response to lactogenic hormones, whereas those expressing the Ha-ras or transforming growth factor-a oncogenes were no longer able to respond to the lactogenic hormones. This inhibition of 0-casein production occurs at the transcriptional level and in the transforming growth factor-atransformed cells is due to an autocrine mechanism involving the activation of the epidermal growth factor receptor. This suggests that, although the c-erbB-2 and epidermal growth factor receptors are structurally quite similar, their activation has different effects upon mammary epithelial cell differentiation.
of the human c-erbB-2 gene were able to respond to lactogenic hormones and express the ,-casein gene. We observed that Ha-ras and TGF-a-expressing HC11 cells were inhibited in ,B-casein production. This inhibition occurred at the transcriptional level. In the TGF-a-transformed cells the mechanism of inhibition involves an autocrine activation of the epidermal growth factor (EGF) receptor. The results suggest that in the HC11 mammary gland cells the EGF receptor and the c-erbB-2 (neu) receptor have different signaling properties. This might reflect different functions for these structurally similar receptors in mammary epithelial cells. The activation of the EGF receptor inhibited lactogenic hormone-induced differentiation, whereas an activated c-neu gene or c-erbB-2 gene did not interfere with this process.
The mammary gland undergoes a complex pattern of growth and differentiation. The epithelial cells of the ducts proliferate and differentiate under the hormonal influence of pregnancy. Multiple steroid and peptide hormones cooperate in this process, which ultimately leads to the induction of milk protein synthesis (41, 46). We have reduced the complexity of these interactions by developing an in vitro culture system (3). A clonal derivative of the COMMA-1D cells (12) obtained from the mammary gland of pregnant BALB/c mice was established. This cell line (HC11) maintains important features of mammary epithelial cell differentiation and lactogenic hormone responsiveness. The HC11 mammary gland cells respond to prolactin and glucocorticoids and express the milk protein ,-casein. The hormones act in a synergistic fashion to regulate transcription from the ,B-casein promoter (15). Various oncogenes have been implicated in human and rodent mammary tumor development (7, 26, 29, 37, 40). Normal signals controlling cell growth and differentiation are disturbed during the transformation process. We used the HC11 mammary cell culture system to study the influence of oncogene expression upon tumorigenicity and differentiation represented by the ability of the cells to respond to lactogenic hormones. Although all of the oncogenes caused malignant transformation of the cells, as measured by tumor formation in nude mice, differences were observed in their effects upon HC11 cell differentiation. Cells expressing an activated Ha-ras oncogene or a transfected transforming growth factor-a (TGF-a) plasmid were inhibited in 3-casein production. Cells transfected with the transforming variant of the rat c-neu gene (neuT) or with a mutated activated copy *
MATERIALS AND METHODS Growth and transfection of HCll cells. HCll cells (3) are clonally derived from the COMMA-1D mouse mammary gland cell line (12). HC11 cells were grown in RPMI 1640 medium containing 8% heat-inactivated fetal calf serum, 5 ,ug of insulin per ml, and 10 ng of EGF per ml (growth medium). 1-Casein was induced in HC11 cells by the following protocol. Confluent cultures were maintained for 2 days in growth medium followed by 2 to 4 days of treatment with induction medium (RPMI 1640, 8% fetal calf serum, 5 ,ug of ovine prolactin per ml, 1 ,uM dexamethasone, and 5 ,ug of insulin per ml) (15). In some experiments an EGF receptor blocking monoclonal antibody A4, obtained from H. Masui and J. Mendelsohn, was used. In this case, 100 to 200 ,ug of the antibody per ml was added directly to the induction medium. HC11 cells were transfected by the calcium phosphate precipitation technique (17). Then 1 x 106 to 2 x 106 cells
Corresponding author. 4027
4028
HYNES ET AL.
cotransfected in Dulbecco modified Eagle medium (DMEM) containing 8% fetal calf serum, 5 ,ug of insulin per ml, and 10 ng of EGF per ml. The precipitate contained 0.5 ptg of pSV2neo, 5 ,ug of a 0-casein promoter-chloramphenicol acetyltransferase (CAT) expression plasmid, and 5 ,ug of each individual expression plasmid (see below). The transfected cells were selected by growth for 10 to 14 days in growth medium containing 200 jig of G418 per ml. Plasmids. The following expression plasmids were transfected into the HCll cells. pHa-ras, received from R. Weinberg, was cloned from the EJ bladder carcinoma cell line and encodes an activated Ha-ras protein (36); pSVTGFa, received from R. Derynck, expresses the human TGF-a cDNA under the transcriptional control of the simian virus 40 early promoter (32, 35); pSV2neuN and pSV2neuT, obtained from R. Weinberg, express, respectively, the rat c-neu normal and transforming proteins, under the control of the simian virus 40 promoter (4); pSV2 erbB-2 (N) and pSV2erbB2(VE) express, respectively, the human c-erbB-2 normal and transforming proteins under the control of the simian virus 40 promoter (27). The activating mutation in the pSV2erbB-2(VE) is a replacement of the Val at position 659 for a Glu and mimics the mutation found in the transmembrane domain of the activated neuT at amino acid 664 (4). The plasmid pSV2neo (38), which confers resistance to the aminoglycoside antibiotic G418 and a P-casein promoterCAT expression plasmid (pplcat) were cotransfected with each of the above-described expression plasmids. Plasmid p,BlCAT, received from W. Doppler, contains 2.3 kilobases (kb) of the 5'-flanking region of the rat 1-casein gene (15) and the coding region of the CAT gene. RNA isolation. Poly(A)+ RNA was isolated and fractionated on 1.85 M formaldehyde-1.2% agarose gels and transferred to a nylon membrane (Gene Screen) in lOx SSC (1 x SSC is 0.15 M NaCI plus 0.015 M sodium citrate). The RNA was cross-linked to the membrane by UV irradiation (Stratagene) and hybridized with 32P-labeled cDNA probes (9). Immunoblot analysis. Lactogenic hormone-induced cultures of HCll cells were analyzed for P-casein protein by using a protein blotting technique (42). Cells were washed with phosphate-buffered saline and homogenized in a buffer containing 10 mM N-2-hydroxyethylpiperazine-N'-2-ethanesulfonic acid (pH 7.4), 10 mM NaCl, 1% Triton X-100, and 1 mM EDTA. Debris was removed by centrifugation at 12,000 x g for 10 min. Then 75 ,ug of protein extract was separated by electrophoresis on 12% sodium dodecyl sulfate-polyacrylamide gels. Proteins were electroblotted onto PVDF membranes (Millipore Corp.) (20). 3-Casein was detected with a rabbit anti-mouse ,B-casein antibody (3), a gift of E. Reichmann, and '25I-labeled protein A. Monoclonal antibody preparation. A monoclonal antibody, FRP5, specific for the human c-erbB-2 protein, was prepared as follows. Mice were immunized with intact SKBR-3 human breast tumor cells, which express very high levels of the c-erbB-2 protein (23, 45). Hybridomas were screened for c-erbB-2-specific antibodies in two steps. First, supernatants were used to immunoprecipitate proteins from SKBR-3 cell extracts. Second, the precipitated proteins were separated on sodium dodecyl sulfate-polyacrylamide gels, blotted onto membranes, and treated with 21N antibody, which specifically detects human c-erbB-2 (19). A more detailed description of FRP5 will be presented elsewhere by I. M. Harwerth (unpublished data). Immunoprecipitation of [p5S]methionine-labeled proteins. The human c-erbB-2 protein was detected in transfected HC11 cells labeled overnight in methionine-free DMEM were
MOL. CELL. BIOL.
containing 100 ,uCi of [35S]methionine (trans 35S label; ICN Radiochemicals) per ml. Cells were lysed in a buffer containing 50 mM Tris hydrochloride (pH 7.5), 5 mM ethylene glycol bis(,-aminoethyl ether)-N,N,N',N'-tetraacetic acid, 1% Triton X-100, 150 mM NaCl, 1 mM phenylmethylsulfonyl fluoride, 80 ,ug of aprotinin per ml, 50 ,g of leupeptin per ml, and 4 pug of pepstatin per ml. Cell debris was removed by centrifugation at 12,000 x g for 10 min. c-erbB-2 protein was immunoprecipitated for 2 h at 4°C with the mouse monoclonal antibody FRP5. The lysates were incubated for an additional hour with rabbit anti-mouse antiserum, and then the immunocomplexes were collected by using protein ASepharose, washed, boiled for 5 min in electrophoresis buffer, and separated on an 8% sodium dodecyl sulfatepolyacrylamide gel. The gel was fluorographed after salicylic acid treatment (8). CAT assays. For a determination of CAT activity, HCll cells were lysed in situ with a solution of 0.25 M Tris hydrochloride (pH 7.8)-0.5% Triton X-100 (2). The extract was centrifuged at 10,000 x g for 5 min, and the supematant was heated for 10 min at 60°C. Denatured protein was pelleted at 10,000 x g for 5 min, and samples of the supernatant were taken for a determination of the protein concentration and CAT enzyme activity (16). [14C]chloramphenicol (20 ,uM) was used in the incubation mixture, and reaction products were resolved by thin-layer chromatography. Conversion was determined by measuring the radioactivity in the nonacylated and acylated forms of chloramphenicol. Radioimmunoassay for TGF-a. Transfected HCll cells were grown in serum-free medium (DMEM-Hams F12 [1:1] containing 10 ,ug of insulin and 10 ,ug of transferrin per ml). The conditioned medium was collected after 2 days and concentrated, and the level of immunoreactive TGF-a was determined by using a liquid-phase competitive radioimmunoassay with a polyclonal rabbit anti-rat TGF-a antiserum as described previously (33). The rat synthetic '251-labeled TGF-a, rabbit anti-TGF-a antiserum, and other reagents were purchased from Biotope (Seattle, Wash.). Tumorigenicity assay. Transfected HC11 cells were selected and grown in medium containing 200 ,ug of G418 per ml. Pools or individual clones were trypsinized, washed, and suspended in 0.2 ml of phosphate-buffered saline. Then 1 x 106 to 3 x 106 cells were injected subcutaneously into nude mice. The animals were monitored for tumor growth at least once per week (24). RESULTS Oncogene introduction into HCll cells. HC11 cells are cloned mouse mammary epithelial cells that retain important features of differentiation and hormonal responsiveness (3). Confluent cultures of HCll cells express the milk protein 0-casein in response to the lactogenic hormones prolactin and glucocorticoids. The cells retain a normal phenotype and do not form tumors in nude mice. HCll cells can be efficiently transfected, and it is thus possible to test the effects of exogenously introduced oncogenes upon both their transformation and differentiation. Activated human Ha-ras, the TGF-a gene, the normal and activated rat neu genes, and the normal and activated human c-erbB-2 genes were cotransfected with pSV2neo into HCll cells. The choice of oncogenes was governed by their consistent activation in human and rodent mammary tumors. About 30%o of primary human breast cancers overexpress the c-erbB-2 protein (7, 18, 37, 45, 48). It is likely that this protein is involved at an
EFFECTS OF ONCOGENES ON MAMMARY EPITHELIAL CELLS
VOL. 10, 1990
Kb 5.0
1 2 Kbv 3 4 Kb 5 6 7
W.-
6.2
4029
TGFa RIA (ng/10' cells/48hrs)
2.3
FIG. 1. Northern blot analysis of poly(A)+ RNA isolated from HC11 transfectants. Poly(A)+ RNA (5 1Lg per lane) was fractionated on denaturing formaldehyde gels, transferred to GeneScreen, and hybridized to 32P-labeled probes. Lanes: 1 and 2, TGF-a; 3 and 4, Ha-ras; 5 through 7, neuT. The RNA was isolated from the following cells: lanes 1, 5, and 6, untransfected HC11 cells; 3, control transfected HC11 cells; 2, 4, and 7, respectively, TGF-a-, Ha-ras-, and neuT-transfected HC11 cells.
early stage in the development of breast cancer (21, 44). The human c-erbB-2 gene and the rat homolog, the c-neu gene, were introduced into HCll cells. The point-mutated transforming neuT gene and a point-mutated human c-erbB-2 gene, c-erbB-2(VE), were expressed in HCll cells. The c-erbB-2(VE) gene has a mutation causing a Val-to-Glu change at amino acid 659 (27). An equivalent change is found at position 664 of the rat NeuT protein (4). A point mutated Ha-ras gene has been found in a high percentage of chemically induced rodent mammary tumors (40). Transgenic mice expressing the activated Ha-ras gene under the control of promoters transcribed in mammary gland cells also develop mammary tumors (43). These observations indicate a causal role for the Ha-ras oncogene in the process of tumorigenesis. The human TGF-a gene was transfected into HCll cells. It has previously been shown that Ha-ras transformation of NOG-8 (31) and NMuMG mouse mammary epithelial cells causes the cells to secrete elevated levels of TGF-aL (11, 33). Enhanced levels of TGF-a may be responsible for some of the changes associated with Ha-ras transformation. In addition, TGF-a expression has been detected in rodent and human mammary tumors (6, 26, 29). Detection of oncogene RNA or protein. Expression of the transfected genes was tested in the HCll cells. Fifty or more independently transfected G418-resistant cell clones were pooled and expanded. Three different methods were used to show that the transfected HCll cells expressed the correct gene product. mRNA was electrophoresed and blotted onto membranes, and the membranes were hybridized with a TGF-a probe (Fig. 1, lanes 1 and 2), an Ha-ras probe (lanes 3 and 4) or a neuT probe (lanes 5 through 7). The TGF-a-transfected cells (lane 2) expressed the expected transcript of 2.3 kb (32, 35). The smaller transcript at 1.8 kb is unidentified. The endogenous mouse TGF-oa transcript in HCll cells was 5.0 kb (lane 1). The Ha-ras-transfected HCll cells (lane 4) expressed the expected 1.4-kb transcript (36). This probe did not detect any mouse transcripts in control HCll cells (lane 3). HCll cells transfected with the neuT plasmid (lane 7) expressed the expected 6.2-kb mRNA (4). Hybridization of an unidentified 2.0-kb RNA was also observed. The 5.4-kb transcript that hybridized with the neuT probe represents the endogenous mouse c-erbB-2/neu transcript (Hynes, unpublished results). A specific radioimmunoassay was used to detect immunoreactive TGF-oa. Conditioned medium was collected from HCll cells transfected with the TGF-a expression plasmid and from cells transfected with the neuT, neuN, and Ha-ras plasmids and, as a control, pUC18. Samples of concentrated conditioned medium were assayed for TGF-a. Medium from
pUCGS
NeuT
neuN
TGFa
Ha-ras
Transfected Plasmid FIG. 2. TGF-a secretion from transfected HCll cells. Media from pools of HCll cells transfected with the indicated plasmid were conditioned for 48 h, concentrated, and assayed by radioimmunoassay for TGF-a as described in Materials and Methods.
cells transfected with the TGF-a gene contained the highest levels of protein (Fig. 2). The cells transfected with the Ha-ras plasmid also secrete elevated TGF-a levels as compared with the control pUC18-transfected cells or with the cells expressing neuT or neuN. We conclude that the transfected TGF-ot plasmid is expressed in the HCll cells and that the expression of the activated Ha-ras but not neuT in HCll cells causes increased expression of the endogenous TGF-a protein. NOG-8 mammary epithelial cells transformed by Ha-ras also secrete elevated levels of TGF-oa, whereas those transformed by neuT do not (10, 11, 33). The enhanced TGF-a production might represent a general aspect of ras transformation, as has previously been shown in fibroblasts (1). The transfected human c-erbB-2 protein was detected in HCll cells by using a mouse monoclonal antibody, FRP5. This antibody specifically recognizes the extracellular domain of the human c-erbB-2 protein (Harwerth, unpublished results). Transfected HCll cells were metabolically labeled with [35S]methionine, cell monolayers were solubilized, and the c-erbB-2 protein was quantitatively immunoprecipitated with FRP5. The precipitates were electrophoresed on sodium dodecyl sulfate-polyacrylamide gels and autoradiographed. Figure 3, lane 1, shows the c-erbB-2 protein present in a pool of c-erbB-2(N)-transfected HCll cells. The cerbB-2 protein was also immunoprecipitated from two individually transfected cell clones containing the c-erbB-2 (VE) product (lane 2) or the c-erbB-2 (N) product (lane 3). The FRP5 monoclonal antibody did not precipitate the endogenous mouse c-erbB-2 protein present in control HCll cells (lane 4). Transformation parameters measured in transfected HCll
4030
MOL. CELL. BIOL.
HYNES ET AL.
MWxlo.3 1
2
4
3
MW Qi 31 2 3 1 2 3 1
2
2
26 v
U
180_
2 3
bi
..
A
B
36 1 2
D
C
2
1
1
E
2
26
116
FIG. 3. Immunoprecipitation of c-erbB-2 from HC11 transfectants. Cells were metabolically labeled with [35S]methionine, lysed, and incubated with FRP5, a monoclonal antibody specific for the human c-erbB-2 protein. Extracts are from the following cells: lane 1, the pool of c-erbB-2(N) transfectants; 2, clone R2 11 expressing c-erbB-2(VE); 3, clone Ri 11 expressing c-erbB-2(N); 4, control untransfected HC11 cells. Molecular weights in thousands are shown on the left.
cells. We determined the consequence of oncogene expression in the HC11 cells by testing two criteria: tumorigenicity in nude mice and growth in the absence of exogenous EGF (Table 1). Cells (1.5 x 106 to 3 x 106) were injected subcutaneously into nude mice, and tumor formation was monitored. HC11 cells containing only the pSV2neo plasmid served as a control and were not tumorigenic. Pools of HC11 cells expressing the neuT, Ha-ras, TGF-a, and c-erbB-2(VE) constructs caused tumor formation. The latency and frequency varied. Pools of HC11 cells expressing activated Ha-ras had the shortest latency and produced the fastestgrowing tumors. HC11 cells expressing TGF-a had the longest latency, and a small tumor was observed after 3 weeks in one of two animals. Three individual cell clones expressing various amounts of the c-erbB-2(VE) protein were also tumorigenic within 1 week of injection. In contrast, pools of HC11 cells expressing the c-erbB-2(N) protein were nontumorigenic (Table 1). Individual cell clones expressing the c-erbB-2(N) protein were also nontumorigenic over a 4-week interval (data not shown). EGF is a potent mitogen for HC11 cells (Taverna, unpubTABLE 1. Transformation parameters measured in HC11 transfectants Transfected plasmid
Tumorigenicity in Cell density in medium nude micea lacking EGFb (% of No. with Latency maximum) tumors/total (days)
Pools
pSV2neo pSV2neuT pSVneuN pHa-ras pSVTGF-a pSV2-erbB-2(N) pSV2-erbB-2(VE) Clones R2 2 pSV2-erbB-2(VE) R2 5 pSV2-erbB-2(VE) R2 11 pSV2-erbB-2(VE)
0/3 3/3 NDC 2/2 1/2 0/2 1/2d 2/2 1/1 2/2
14
27 81 19 51 107 ND ND
7 7 7
ND ND ND
10 7 >21
106 and 3.0 x 106 cells were injected. cells were plated in medium with or without EGF, and the final cell
a Between 1.5 x b
The
density was determined as (total cell number without EGF/total cell number with EGF) x 100. c ND, Not determined. d One animal died at 2 weeks.
A
F
G
FIG. 4. Western immunoblot analysis of P-casein in lactogenic hormone-induced HC11 transfectants. Cell lysates were prepared, electrophoresed, and electroblotted, and the P-casein protein was detected by using a specific antibody followed by '251-labeled protein A treatment. The lysates are from HC11 cells transfected with the following: A, pUC18; B, neuN; C, neuT; D, c-erbB-2(N); E, c-erbB-2(VE); F, TGF-oa; G, Ha-ras. The HC11 transfectants were treated with the following: lanes 1, insulin; 2, DIP; 3, DIP plus EGF. Each lane contained 75 ,ug of protein extract. Molecular weights in thousands are shown on the left.
lished observation). In its absence, the HC11 cells reach a lower final cell density. HC11 cells expressing pSV2neuT, pHa-ras, and pSVTGF-a are less dependent than pSV2neoor pSV2neuN-transfected cells upon exogenous EGF addition to the medium (Table 1). Thus, the tumorigenic HC11 transfectants are also altered in their in vitro growth properties. Lactogenic hormone response in transfected HCll cells. Pools of transfected HC11 cells were tested for their ability to respond to the lactogenic hormones dexamethasone, insulin, and prolactin (DIP). The cells were grown to confluency and maintained for 2 days in growth medium before treatment in the hormone-containing induction medium. This protocol has been shown to promote efficient utilization of the P-casein promoter (15) and leads to high levels of mouse ,-casein protein production. Cell extracts were prepared from pools of transfected HC11 cells, and the level of P-casein protein was analyzed by using a protein blotting technique (Fig. 4). HC11 cells transfected with the control plasmid pUC18 as well as the neuN-, neuT-, erbB-2(N)-, and erbB-2(VE)-expressing plasmids synthesized high levels of the ,-casein protein in response to the lactogenic hormones (lanes 2 in Fig. 4A through E, respectively). The expression of the c-neu or c-erbB-2 products, in their normal or activated forms, did not interfere with the lactogenic hormone response. In contrast, in HC11 cells expressing TGF-a and those expressing the transforming Ha-ras protein little or no Icasein protein was detected after lactogenic hormone treatment (lanes 2 in Fig. 4F and G, respectively). One possible explanation for the inhibition of 3-casein synthesis in these cells is the activation of the EGF receptor due to enhanced TGF-a production in the TGF-a- and Ha-ras-transfected cells (Fig. 2). When EGF was included in the DIP medium, 1-casein induction was also strongly inhibited in HC11 cells transfected with pUC18, neuN, and neuT (lanes 3 in Fig. 4A through C, respectively). These results suggest that activation of the EGF receptor either by exogenous addition of EGF or by an autocrine mechanism involving the production of TGF-a is incompatible with lactogenic hormone induction of P-casein expression. To further characterize the mechanism of inhibition in the TGF-a- and Ha-ras-transformed cells, we made use of an
EFFECTS OF ONCOGENES ON MAMMARY EPITHELIAL CELLS
VOL. 10, 1990
1 2
3
4
1
TABLE 2. Induction of ,3-casein CAT expression in transfected HCll cells after hormone additiona
c
B
A
2
3
1
4031
2
3
FIG. 5. Western blot analysis of 3-casein in lactogenic hornoneinduced HCll cells treated with a monoclonal antibody (A4) specific for the EGF receptor. Cell lysates were prepared, electrophoresed, and electroblotted, and the 3-casein protein was detected by using a specific antibody followed by '25N-labeled protein A treatment. (A) The lysates are from HCll cells treated with the following: lane 1, DIP; 2 and 3, DIP plus EGF; 4, DIP, EGF, and 100 ,g of A4 per ml. (B) The lysates are from TGF-a transfectants treated with the following: lane 1, DIP; 2, DIP plus 100 ig of A4 per ml; 3, DIP plus 200 ,ug of A4 per ml. (C) Ha-ras transfectants treated as in panel B, lanes 1 through 3. The amount of protein extract loaded was 40 Fg in panel A and 30 ,ug in panels B and C. In each panel the ,B-casein protein is indicated with a bar. A different gel system was used for panels B and C; hence the ,3-casein protein migrated differently from that in panel A.
antibody directed against the EGF receptor. The monoclonal antibody A4 binds to the extracellular domain of the mouse EGF receptor and inhibits its ligand-induced phosphorylation (H. Masui, T. Aldrich and J. Mendelsohn, Proc. Am. Assoc. Cancer Res. 30:251, 1989). This blocking activity was confirmed by the results shown in Fig. 5A. HC11 cells treated with DIP medium containing EGF synthesize much lower levels of 3-casein protein than do control cells (Fig. 5A; compare lanes 2 and 3 with lane 1). This inhibition was partially reversed by the addition of 100 ,ug of monoclonal antibody A4 per ml to the medium containing DIP and EGF (Fig. 5A, lane 4). Figures 5B and C show the results obtained with TGF-ca- and Ha-ras-transformed cells, respectively. After treatment of the DIP-induced TGF-a-transformed cells with 100 ,ug and 200 ,ug of A4 per ml, increasing amounts of 1-casein protein were synthesized (Fig. 5B, lanes 2 and 3). The presence of 100 and 200 ,ug of monoclonal antibody A4 per ml had no effect upon the low level of ,-casein protein synthesized in DIP-treated Ha-ras transfectants (Fig. 5C). From these results we conclude that it is possible to block at least partially the autocrine circuit in TGF-a-transfected HCll cells and to restore their lactogenic hormone sensitivity. This does not appear to be the case in Ha-ras-transformed cells and suggests that these transformed cells have undergone alterations in addition to TGF-a production that render them insensitive to lactogenic hormones. Oncogenes inhibit the P-casein promoter utilization upon lactogenic hormone induction. It has previously been shown that the lactogenic hormones regulate transcription of the f-casein promoter (3, 15). A ,-casein promoter-CAT construct (po1-CAT) that contains 2,300 nucleotides of 5'flanking region from the rat 3-casein gene responds to the synergistic action of dexamethasone and prolactin (15). Thus, we examined the effects of oncogene expression upon the transcription of the ,B-casein promoter by using the pBl-CAT gene, which was cotransfected with the pSV2neo and each of the oncogene-expressing plasmids into HC11 cells. CAT activity was measured in stably transfected, confluent cultures after lactogenic hormone induction (Table 2). The CAT assays were performed twice on pools of transfected HC11 cells, and the induction ratio was determined by comparison of the basal CAT activity in the presence of insulin with the activity in the presence of DIP. The value of the pUC18-transfected control cells induced with DIP was used as 100% activity. Two different pools of cells transfected with the Ha-ras plasmid showed lower
CAT activity'
Induction (fold)
% of maximal
DIP + EGF
0.044 0.030 0.289 0.028 0.278
13.3 2.3 9.6 9.9
100 17.3 72.2 74.4
Transfected plasmid
Insulin
pUC18
0.019 0.252
pSV2neuN pSVneuT (pool 1) pHa-ras
DIP
activity
0.011
3.7
27.8
(pool 1) pSVTGF-a
0.029 0.166
5.7
42.8
pUC18
0.035 0.652
0.652
18.6 5.6 15.9
100 30.1 85.5
0.012 0.052
4.3
23.1
0.042 0.260
6.2
33.3
0.125
5.5
29.6
0.003
0.195
pSV2neuT (pool 2) pHa-ras (pool 1) pHa-ras (pool 2) pSVTGF-a
0.041
0.683
The results of two independent experiments are shown. b Nanomoles per minute per milligram of protein in cells grown with the indicated hormones. a
levels of CAT activity (23 to 33% of maximal activity), HCll cells expressing TGF-a showed lower CAT activity (30 to 43%), and control pUC18-transfected cells induced with DIP in the presence of EGF also showed lower levels of CAT activity (17 to 30%). In contrast, cells expressing neuT or neuN showed 72 to 85% of the CAT activity seen in control cells. This suggests that transcription from the P-casein promoter is impaired in HCll cells treated with EGF as well as in Ha-ras- and TGF-a-transformed HCll cells. DISCUSSION We have shown that the Ha-ras, TGF-a, neuT, and c-erbB-2(VE) oncogenes, which have all been implicated in mammary cancer, can malignantly transform the HC11 mammary epithelial cell line. But by examining both growth and differentiation parameters, we have observed that these oncogenes not only employ different mechanisms for transformation but also have different effects upon the cellular differentiation. The transformation of the HC11 cells by Ha-ras and TGF-a results in part from the activation of the EGF receptor. Conditioned medium from Ha-ras- and TGF-aexpressing cells contains elevated levels of secreted TGF-a. The NOG-8 mammary epithetial cell line transformed by a dexamethasone-inducible mouse mammary tumor virus long terminal repeat-Ha-ras construct shows Ha-ras-dependent increased TGF-a production (11). The transformed NOG-8 cells have reduced EGF binding, most likely due to TGFa-induced down-regulation of the EGF receptors. This autocrine loop may be one of the mechanisms important in Ha-ras transformation of mammary epithelial cells. But Shankar et al. (35) have observed that the tumorigenicity of TGF-a-transfected NOG-8 cells is dependent upon the level of growth factor production. The ras-transformed HC11 cells secrete less TGF-a than do the TGF-a-transformed cells, thus making it unlikely that their tumorigenicity is due only to elevated growth factor production. The Ha-ras- and TGF-ca-transformed cells have lost their ability to differentiate and produce P-casein. At least in the
4032
HYNES ET AL.
TGF-a-transformed cells, the mechanism that leads to the inhibition of DIP-induced ,-casein production appears to be the activation of the EGF receptor by secreted TGF-a. Control transfected HC11 cells treated with EGF during the DIP induction also fail to synthesize ,B-casein. In both cases partial sensitivity to lactogenic hormones could be restored by blocking the EGF receptor activity with the monoclonal antibody A4. The level of P-casein synthesized by the TGF-a-transfected cells in the presence of the A4 antibody was lower than that observed in the HC11 cells treated with the antibody. This could have various explanations. Mature TGF-a is cleaved from a precursor protein that is embedded in the plasma membrane. Recently, it has been shown that this membrane-bound TGF-a precursor is biologically active (47). Due to its close proximity to the EGF receptor, the TGF-a precursor may be able to activate its receptor directly on the membrane before being processed and shed into the medium. Thus, the TGF-a synthesized by the transfected cells may be more difficult to block than the exogenous EGF, since its concentration on the membrane is most likely higher than the concentration of monoclonal antibody A4 in the cell culture medium. This may explain why the presence of increasing concentrations of monoclonal antibody in the DIP medium led to increasing amounts of 1-casein protein. In EGF-treated HC11 cells, different amounts of A4 were tested (50 to 200 ,uglml), but the level of ,-casein protein remained constant (data not shown). The lactogenic hormone sensitivity of the Ha-ras-transformed HC11 cells could not be restored by treatment with the EGF receptor-blocking antibody. This is despite the fact that these cells secrete less TGF-a than do the TGF-a-transfected cells and suggests that other factors important for the differentiation of HC11 cells are altered in the Ha-ras-expressing cells. Some human breast tumors have elevated EGF receptor levels and high TGF-a expression (14, 29). This suggests that the autocrine mechanism at work in the Ha-ras- and TGFa-transformed HC11 cells may mimic the situation in these tumors, where the overall result would be enhanced growth as well as maintenance of an undifferentiated phenotype. The block in 1-casein production occurs, at least partly, at the level of transcription from the P-casein promoter. Transfected cells expressing P-casein CAT reporter plasmids, in addition to Ha-ras or TGF-a, displayed lower DIP-induced CAT activity than control pUC18-transfected cells. The DIP inducibility of the P-casein CAT construct is also negatively affected by exogenous EGF addition. Dexamethasone and prolactin are the important components in the lactogenic hormone mixture (15). These hormones have different modes of action. The dexamethasone-mediated response on the 1-casein promoter appears to be indirect. The kinetics of its action are slow, and no glucocorticoid response element has been detected (W. Doppler, W. Hoeck, P. Hofer, B. Groner, and R. Ball, Mol. Endocrinol., in press). We have observed that the mouse mammary tumor virus long terminal repeat-CAT reporter gene in Ha-ras- and TGF-a-transformed cells responds normally to the action of dexamethasone, suggesting that the activity of the glucocorticoid receptor is not impaired (data not shown). Since its action is indirect, other glucocorticoid-sensitive gene products in the Ha-ras- and TGF-a-transformed cells or the EGF-treated control cells may be affected. The events that follow binding of prolactin to its cell surface receptor have not been characterized. In HC11 cells it has a rapid effect upon 3-casein transcription (15). It will be interesting to see whether in Ha-ras- or TGF-a-transformed cells or in EGFtreated cells the altered activity of the ,B-casein gene pro-
MOL. CELL. BIOL. moter can be traced to
changes in the presence or binding of nuclear transcription factors. ras oncogenes have been introduced into other cell lines that can be induced to differentiate. Microinjection of the activated H-ras protein into rat pheochromocytoma cells (PC-12) promoted their morphological differentiation into neuronlike cells (5). Human bronchial epithelial cells, which can be induced by TGF-,1 or serum to undergo squamous cell differentiation, become unresponsive to these agents upon introduction of an activated Ki-ras oncogene (30). Undifferentiated skeletal muscle myoblasts fuse to form differentiated multinecleated myotubes. This process is accompanied by both down-regulation and up-regulation of specific gene products. The oncogenic ras protein interferes with these processes (28, 39). Sternberg et al. (39) have shown that ras blocks the accumulation of transcriptional regulatory factors required for the induction of muscle creatine kinase, one of the genes induced during myogenesis. Taken together, the results show that, depending upon the cells, activated ras proteins can either promote or interfere with differentiation, which suggests that the response to a ras-mediated signal is an intrinsic property of a particular cell type. The c-erbB-2 protein is overexpressed in a high percentage of human breast tumors (7, 18, 37, 45, 48). It is probable that high levels of the protein confer a growth advantage upon these tumor cells. In NIH 3T3 mouse fibroblasts there is a correlation between the level of c-erbB-2 tyrosine kinase activity and cellular transformation (34). Elevated kinase activity can be achieved by low levels of activated c-erbB-2 or by overexpression of the normal protein. In both cases NIH 3T3 cells become transformed (13, 22). HC11 cells were transfected with a plasmid expressing normal human cerbB-2 as well as with a plasmid expressing a point-mutated activated version of c-erbB-2. All of the cells expressing low or high levels of the activated c-erbB-2 protein formed rapidly growing tumors within 1 week of injection. HC11 cells transfected with the c-erbB-2(N) expression plasmid were not tumorigenic in nude mice. The transfected cells that we analyzed, however, did not contain the extremely high levels of c-erbB-2 protein that are detected in tumor cell lines such as SKBR3 (23, 45). The activated neuT or c-erbB-2(VE) proteins are capable of transforming HC11 cells but do not inhibit DIP-induced ,-casein production. This suggests that the intracellular pathways activated by the c-erbB-21neu protein are at least partially distinct from those stimulated by ras transformation and by EGF receptor activation. Lee et al. (25) have introduced an EGF receptor/c-erbB-2 chimeric protein into NIH 3T3 fibroblasts. The chimeric protein responds to EGF. Cellular substrates phosphorylated on tyrosine residues have been identified and compared with those seen after activation of the EGF receptor. Although most of the substrates appeared to be shared by the two receptors, at least one was specific for the c-erbB-2 kinase. The HC11 cells may prove to be useful in characterizing the signal transduction pathways activated by these receptors in mammary epithelial cells and may help clarify subtle differences in the function of closely related receptor tyrosine kinases. ACKNOWLEDGMENTS We acknowledge the technical assistance of U. Stiefel and P. Hofer. We thank W. Wels and W. Filipowicz for their useful comments on the manuscript and H. Masui and J. Medelsohn for the A4 monoclonal antibody.
VOL. 10, 1990
EFFECTS OF ONCOGENES ON MAMMARY EPITHELIAL CELLS
LITERATURE CITED 1. Anzano, M. A., A. B. Roberts, J. E. Delarco, L. M. Wakefield, R. K. Assoian, N. S. Roche, J. M. Smith, J. E. Lazarus, and M. B. Sporn. 1985. Increased secretion of type 1B transforming growth factor accompanies viral transformation of cells. Mol. Cell. Biol. 5:242-247. 2. Ausubel, F. M., R. Brent, R. E. Kingston, D. D. Moore, J. G. Seidman, J. A. Smith, and K. Struhl (ed.). 1987. Current protocols in molecular biology. John Wiley & Sons, Inc., New York. 3. Ball, R. K., R. R. Friis, C. A. Schonenberger, W. Doppler, and B. Groner. 1988. Prolactin regulation of ,B-casein gene expression and of a cytosolic 120-kd protein in a cloned mouse mammary epthelial cell line. EMBO J. 7:2089-2095. 4. Bargmann, C. I., M.-C. Hung, and R. A. Weinberg. 1986. Multiple independent activations of the neu oncogene by a point mutation altering the transmembrane domain of P185. Cell 45:649-657. 5. Bar-Sagi, D., and J. R. Feramisco. 1985. Microinjection of the ras oncogene protein into PC12 cells induces morphological differentiation. Cell 42:841-848. 6. Bates, S. E., N. E. Davidson, E. M. Valverius, C. E. Freter, R. B. Dickson, J. P. Tam, J. E. Kudlow, M. E. Lippman, and D. S. Salomon. 1988. Expression of transforming growth factor a and its messenger ribonucleic acid in human breast cancer: its regulation by estrogen and its possible functional significance. Mol. Endocrinol. 2:543-555. 7. Berger, M. S., G. W. Locher, S. Saurer, W. J. Gullick, M. D. Waterfield, B. Groner, and N. E. Hynes. 1988. Correlation of C-erbB-2 gene amplification and protein expression in human breast carcinoma with nodal status and nuclear grading. Cancer Res. 48:1238-1243. 8. Chamberlain, J. 1979. Fluorographic detection of radioactivity in polyacrylamide gels with the water soluble fluor, sodium salicylate. Anal. Biochem. 98:132-135. 9. Church, G. M., and W. Gilbert. 1984. Genomic sequencing. Proc. Natl. Acad. Sci. USA 81:1991-1995. 10. Ciardiello, F., N. E. Hynes, N. Kim, E. M. Valverius, M. E. Lippman, and D. S. Salomon. 1989. Transformation of mouse mammary epithelial cells with the Ha-ras but not the neu oncogene results in a gene dosage-dependent increase in transforming growth factor-a production. FEBS Lett. 250:474-478. 11. Ciardiello, F., N. Kim, N. E. Hynes, R. Jaggi, S. Redmond, D. S. Liscia, B. Sanfilippo, G. Merlo, R. Callahan, W. R. Kidwell, and D. S. Salomon. 1988. Induction of transforming growth factor a expression in mouse mammary epithelial cells after transformation with a point-mutated c-Ha-ras protooncogene. Mol. Endocrinol. 2:1202-1216. 12. Danielson, K. G., C. J. Oborn, E. M. Durban, J. S. Butel, and D. Medina. 1984. Epithelial mouse mammary cell line exhibiting normal morphogenesis in vivo and functional differentiation in vitro. Proc. Natl. Acad. Sci. USA 81:3756-3760. 13. Di Fiore, P. P., J. H. Pierce, M. H. Kraus, 0. Segatto, C. R. King, and S. A. Adronson. 1987. erbB-2 is a potent oncogene when over-expressed in NIH/3T3 cells. Science 230:1132-1139. 14. DiMarco, E., J. H. Pierce, T. P. Fleming, M. H. Kraus, C. J. Molloy, S. A. Aaronson, and P. P. DiFiore. 1989. Autocrine interaction between TGF-a and the EGF-receptor: quantitative requirements for induction of the malignant phenotype. Oncogene 4:831-838. 15. Doppler, W., B. Groner, and R. K. Ball. 1989. Prolactin and glucocorticoid hormones synergistically induce expression of transfected rat 1-casein gene promoter constructs in a mammary epithelial cell line. Proc. Natl. Acad. Sci. USA 86:
19. 20. 21.
22.
23. 24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
104-108.
16. Gorman, C. M., L. F. Moffat, and B. H. Howard. 1982. Recombinant genomes which express chloramphenicol acetyltransferase in mammalian cells. Mol. Cell. Biol. 2:1044-1051. 17. Graham, F. L., and A. J. van der Eb. 1973. A new technique for the assay of infectivity of human adenovirus 5 DNA. Virology 52:456-457. 18. Guerin, M., M. Barrois, M.-J. Terrier, M. Spielmann, and G. Riou. 1988. Over-expression of either c-myc or c-erbB-2/neu
34.
35.
4033
proto-oncogenes in human breast carcinomas: correlation with poor prognoses. Oncogene Res. 3:21-31. Gullick, W. J., M. S. Berger, P. L. P. Bennett, J. B. Rothbard, and M. D. Waterfield. 1987. Expression of the c-erbB-2 protein in normal and transformed cells. Int. J. Cancer 40:246-254. Gultekin, H., and K. H. Heermann. 1988. The use of polyvinylidenedifluoride membranes as a general blotting matrix. Anal. Biochem. 172:320-329. Gusterson, B. A., L. G. Machin, W. J. Gullick, N. M. Gibbs, T. J. Powles, C. Elliott, S. Ashley, P. Monaghan, and S. Harrison. 1988. c-erbB-2 expression in benign and malignant breast disease. Br. J. Cancer 58:453-457. Hudziak, R. M., J. Schlessinger, and A. Ullrich. 1987. Increased expression of the putative growth factor receptor p185HER2 causes transformation and tumorigenesis of NIH/3T3 cells. Proc. Natl. Acad. Sci. USA 84:7159-7163. Hynes, N. E., H. A. Gerber, S. Saurer, and B. Groner. 1989. Over-expression of the c-erbB-2 protein in human breast tumor cell lines. J. Cell. Biochem. 39:167-173. Hynes, N. E., R. Jaggi, S. C. Kozma, R. Ball, D. Muellener, N. T. Wetherall, B. W. Davis, and B. Groner. 1985. New acceptor cell for transfected genomic DNA: oncogene transfer into a mouse mammary epithelial cell line. Mol. Cell. Biol. 5:268-272. Lee, J., T. J. Dull, I. Lax, A. Schlessinger, and A. Ullrich. 1989. HER2 cytoplasmic domain generates normal mitogenic and transforming signals in a chimeric receptor. EMBO J. 8:167173. Liu, S. C., B. Sanfilippo, I. Perroteau, R. Derynck, D. S. Salomon, and W. R. Kidwell. 1987. Expression of transforming growth factor alpha (TGF alpha) in differentiated rat mammary tumors: estrogen induction of TGF alpha production. Mol. Endocrinol. 1:683-692. Masuko, T., K. Sugahara, M. Kozono, S. Otsuki, T. Akiyama, T. Yamamoto, K. Toyoshima, and Y. Hashimoto. 1989. A murine monoclonal antibody that recognizes an extracellular domain of the human c-erbB-2 protooncogene product. Jpn. Cancer Res. 80:10-14. Olson, E. N., and Y. G. Capetanaki. 1989. Developmental regulation of intermediate filament and actin mRNAs during myogenesis is disrupted by oncogenic ras genes. Oncogene 4:907-913. Perroteau, I., D. S. Salomon, M. DeBertoli, W. R. KidweUl, P. Hazarika, R. Pardue, J. Dedman, and J. Tam. 1986. Immunological detection and quantitation of alpha-transforming growth factors in human breast carcinoma cells. Breast Cancer Res. Treat. 7:201-210. Reddel, R. R., Y. Ke, M. E. Kaighn, L. Malan-Shibley, J. F. Lechner, J. S. Rhim, and C. C. Harris. 1988. Human bronchial epithelial cells neoplastically transformed by v-Ki-ras: altered response to inducers of terminal squamous differentiation. Oncogene Res. 3:401-408. Redmond, S. M. S., E. Reichmann, R. G. Muller, R. R. Friis, B. Groner, and N. E. Hynes. 1988. The transformation of primary and established mouse mammary epithelial cells by p21-ras is concentration dependent. Oncogene 2:259-265. Rosenthal, A., P. B. Lindquist, T. S. Bringman, D. V. Goeddel, and R. Derynck. 1986. Expression in rat fibroblasts of a human transforming growth factor alpha cDNA results in transformation. Cell 46:301-309. Salomon, D. S., I. Perroteau, W. R. KidweUl, J. Tam, and R. Derynck. 1987. Loss of growth responsiveness to epidemal growth factor and enhanced production of alpha-transforming growth factors in ras-transformed mouse mammary epithelial cells. J. Cell. Physiol. 130:397-409. Segatto, O., C. R. King, J. H. Pierce, P. P. di Fiore, and S. A. Aaronson. 1988. Different structural alterations upregulate in vitro tyrosine kinase activity and transforming potency of the erbB-2 gene. Mol. Cell. Biol. 8:5570-5574. Shankar, V., F. Ciardieflo, N. Kim, R. Derynck, D. S. Liscia, G. Merlo, B. C. Langton, D. Sheer, R. Callahan, R. H.1 Bassin, M. E. Lippman, N. Hynes, and D. S. Salomon. 1989. Transformation of an established mouse mammary epithelial cell line
4034
36. 37.
38.
39.
40.
41. 42.
HYNES ET AL.
following transfection with a human transforming growth factor alpha CDNA. Mol. Carcinogen. 2:1-11. Shib, C., and R. A. Weinberg. 1982. Isolation of a transforming sequence from a human bladder carcinoma cell line. Cell 29:161-169. Slamon, D. S., G. M. Clark, S. G. Wong, W. J. Levin, A. Ullrich, and W. L. McGuire. 1987. Human breast cancer: correlation of relapse and survival with amplification of the HER-2/neu oncogene. Science 235:177-182. Southern, P. J., and P. Berg. 1982. Transformation of mammalian cells to antibiotic resistance with a bacterial gene under control of the SV40 early region promoter. J. Mol. Appl. Genet. 1:327-341. Sternberg, E. A., G. Spizz, M. E. Perry, and E. N. Olson. 1989. A ras-dependent pathway abolishes activity of a muscle-specific enhancer upstream from the muscle creative kinase gene. Mol. Cell. Biol. 9:594-601. Sukumar, S., V. Notario, D. Martin-Zanca, and M. Barbacid. 1983. Induction of mammary carcinomas in rats by nitrosomethylurea involves malignant activation of H-ras-1 locus by single point mutations. Nature (London) 306:658-661. Topper, Y. J., and C. S. Freeman. 1980. Multiple hormone interactions in the developmental biology of the mammary gland. Physiol. Rev. 60:1049-1106. Towbin, H., T. Staehelin, and J. Gordon. 1979. Electrophoretic transfer of proteins from polyacrylamide gels to nitrocellulose sheets: procedure and some applications. Proc. Natl. Acad. Sci. USA 76:4350-4354.
MOL. CELL. BIOL. 43. Tremblay, P. J., F. Pothier, T. Hoang, G. Tremblay, S. Brownstein, A. Liszauer, and P. Jolicoeur. 1989. Transgenic mice carrying the mouse mammary tumor virus ras fusion gene: distinct effects in various tissues. Mol. Cell. Biol. 9:854-889. 44. Van de Vjver, M. J., J. L. Peterse, W. J. Moot, P. Wisman, J. Lomans, 0. Dalesio, and R. Nusse. 1988. Neu-protein overexpression in breast cancer: association with comedo-type ductal carcinoma in situ and limited prognostic value in stage II breast cancer. N. Engl. J. Med. 319:1239-1245. 45. Van de V"ver, M., R. van de Bersselaar, P. Devilee, C. Corenelisse, J. Peterse, and R. Nusse. 1987. Amplification of the neu (c-erbB-2) oncogene in human mammary tumors is relatively frequent and is often accompanied by amplification of the linked c-erbA oncogene. Mol. Cell. Biol. 7:2019-2023. 46. Vonderhaar, B. K. 1988. Regulation of development of the normal mammary gland by hormones and growth factors, p. 251-266. In M. E. Lippman and R. Dickson (ed.), Breast cancer cellular and molecular biology. Kluwer Academic Publishers, Boston. 47. Wong, S. T., L. F. Wincheli, B. K. McCune, H. S. Earp, J. Tebixdo, J. Massague, B. Herman, and D. C. Lee. 1989. The TGF-a precursor expressed on the cell surface binds to the EGF receptor on adjacent cells, leading to signal transduction. Cell 56:495-806. 48. Zhou, D., H. Battifora, J. Yokata, T. Yamamoto, and M. J. Kline. 1987. Association of multiple copies of the c-erbB-2 oncogene with spread of breast cancer. Cancer Res. 47:61236125.
