JOURNAL OF CELLULAR PHYSIOLOGY 148:220-227 (1991)

Growth Modulation by Epidermal Growth Factor (EGF) in Human Colonic Carcinoma Cells: Constitutive Expression of the Human EGF Gene SHUANG HUANG, PIN-FANG LIN, DOMINIC FAN, JANET E. PRICE, JOSEM. TRUJILLO,AND SUBHAS CHAKRABARTY* Division of Laboratory Medicine ( S .H.,].M.T.,S.C.) and Department of Cell Biology (D.F., J.E.P.) University of Texas, M.D. Anderson Cancer Center, Houston, Texas 77030; Department of Virology, Bristol-Myers Syuibb Co., Wallingford, Connecticut 06492 (P.3.L). The functional role of epidermal growth factor (EGF) in epithelium-derived human colonic carcinoma cells was investigated by transfection with plasmid pUCDS3, which contained synthetic human EGF encoding sequences, into two human colonic carcinoma cell types with dissimilar phenotypic properties: the moderately differentiated and growth factor-responsive Moser and the highly metastatic KM12SM cells. The Moser cells exhibited a proliferative response to treatment with exogenous EGF, while the KM12SM cells did not. The constitutive expression of the human EGF gene in these colonic carcinoma cell types resulted in elevated expression of EGF mRNA, with concurrent production and secretion of a large amount of EGF, and downmodulation of transforming growth factor-alpha (TGF-alpha) secretion. Growth stimulation and down-modulation of both high and low affinity EGF receptors were observed in the EGF-transfected Moser clones. Results of experiments using anti-EGF and anti-EGF-receptor antibody to block the proliferation of EGF-transfected Moser clones suggested that autocrine stimulatory mechanisms involving both EGF and TGF-alpha were operative in these cells. By comparison, a growth-inhibitory effect, with no apparent EGF receptor modulation, was observed in the EGF-transfected KM12SM clones. Both the parental and EGF-transfected KM12SM clones possessed fewer EGF receptors than the Moser cells, and anti-EGF or anti-EGF-receptor antibody did not affect the cells’ growth properties. These results suggested that the mechanisms of growth inhibition in the EGF-transfected KM12SM clones were non-autocrine or intracellular in nature. Thus, constitutive expression of the human EGF gene in two phenotypically different, epithelium-derived human colonic carcinoma cells resulted in divergent altered growth characteristics.

The regulation of cell growth and differentiation involves complex molecular processes in which various polypeptide growth factors play critical roles in a n autocrine or paracrine manner (Gospodarowicz, 1983; Heldin and Westermark, 1984: Moses et al., 1987; Coffey et al., 1987, 1986; Laiho and Keski-Oja, 1989; Kim et al., 1990). Epidermal growth factor (EGF), first identified in and isolated from the mouse submaxillary gland, is a small polypeptide growth factor of 53 amino acid residues and a molecular weight of about 6,000 (Cohen, 1962; Cohen and Carpenter, 1975). Human EGF shares extensive sequence homology with mouse EGF, and both forms of EGF have similar biological properties in vitro (Donaldson et al., 1989). EGF stimulates cellular proliferation and mitogenesis in most tissue-cultured cells. This biological effect is mediated by the binding of EGF with its specific cell-surface receptors and culminates in biochemical events that 0 1991 WILEY-LISS, INC.

lead to DNA synthesis and cellular proliferation (Donaldson et al., 1989). It has been postulated that abnormal production of growth-stimulatory factors or abnormally high expression of receptors for these growth factors, or both, may allow normal cells to escape the normal constraint of growth control, bringing on cellular transformation (Temin, 1970; De Larco and Todaro, 1978; Todaro et al., 1982;). Indeed, transfection of the human EGF gene into mesenchymally derived fibroblastic cells, with a n appropriate expression vector, results in the constitutive expression and secretion of EGF by these cells and the induction, via a n autocrine mechanism, of a transformed phenotype (Stern et al., 1987). Overexpression of the EGF receptor gene in Received February 18, 1991; accepted April 18, 1991

*To whom reprint requestsicorrespondence should be addressed.

EGF ACTION IN HUMAN COLONIC CARCINOMA CELLS

fibroblastic cells via transfection has also been shown to contribute to the transformed phenotype (Velu e t al., 1987). Epithelium-derived human colonic carcinoma cells are known to produce and respond to a variety of polypeptide growth factors (Coffey et al., 1987, 1986; Chakrabarty et al., 1989a, 1990; Anzano et al., 1989; Schroy et al., 1990). However, the functional role of these growth factors in control of growth and differentiation in colonic carcinoma cells is largely unknown. To examine the functional role of EGF in these cells, human EGF encoding sequences were transfected and constitutively expressed in two human colonic carcinoma cell lines, Moser and KM12SM. The Moser cells are moderately differentiated and responsive to differentiation-inducing agents and polypeptide growth factors (Chakrabarty et al., 1989a, 19901, whereas the KMl2SM cells are highly metastatic (Morrikawa et al., 1988; Fan et al., 1989). EGF-transfected clonal cells expressed a high level of EGF mRNA and secreted EGF into the media. In the Moser cells, constitutive expression and secretion of EGF resulted in an accelerated rate of proliferation with a concurrent down-modulation of EGF cell-surface receptors and down-modulation of transforming growth factor-alpha (TGF-alpha) secretion. Growth stimulation by EGF in the EGF-transfected Moser cells was found to be associated with autocrine mechanisms involving both EGF and TGF-alpha. In the KMl2SM cells, constitutive expression and secretion of EGF resulted in inhibition of cell growth with no apparent modulation of cell-surface EGF receptors. However, down-modulation of TGF-alpha secretion was also observed. Growth inhibition in the EGF-transfected KMl2SM cells was found to be non-autocrine or intracellular in nature.

221

SV40 polyadenylation site, sequences encoding a mouse immunoglobulin heavy-chain signal peptide (to facilitate secretion of translated product), and synthetic human EGF encoding sequences. The pUCDS3 DNA was transformed into E . coli JM109 and selected for ampillicin resistance. Amplified pUCDS3 DNA was isolated from bacterial cells by the alkaline-lysis and CsC1-gradient centrifugation methods (Ausubel et al., 1989). The amplified plasmid DNA was characterized by restriction endonuclease digestion followed by agarose gel electrophoresis, which showed the presence of a unique EcoRIiSal I fragment of about 180 bp containing the EGF sequences, and the approximately 300 bp fragment containing the promoterienhancer sequences after digestion with Xba I (Stern et al., 1987). The control plasmid pUCDS5, also generously provided by Dr. Stern, was constructed in which a 4-bp insertion causes a frameshift resulting in premature translation termination in the signal peptide (Stern et al., 1987). Plasmid pSV2-neo (Southern and Berg, 1982), containing the bacterial neomycin (G418) resistant gene, was similarly transformed and amplified in E. coli JM109. Amplified plasmid DNA was similarly prepared for cotransfection with either pUCDS3 DNA or pUCDS5 DNA into target cells. Target colonic carcinoma cells were cotransfected with pUCDS3 or pUCDS5 (22.5 bg) and pSV2-neo (2.5 pg) DNA by the calcium-phosphate precipitation method (Chen and Okayama, 1988). Sixteen to 24 h r later, the cells were given a “glycerol shock” (Parker and Stark, 1979) by being swirled gently with medium containing 15% glycerol for 2 to 3 min. The cells were then washed with media and usually subcultured by trypsinization and seeded into new culture flasks with medium containing the selection antibiotic G418 (2.5 mgiml for the Moser and 1.0 mgiml for KMl2SM cells) and cultured in the presence of this drug for 2-3 weeks. G418 resistant cells MATERIALS AND METHODS were cloned with the aid of cloning cylinders or by the Cells and reagents limiting-dilution method, or both. Cloned cells were The human colonic carcinoma cells Moser and cultured and maintained in medium containing G418 KMl2SM were maintained in monolayer culture, a t (0.5 mgiml for Moser and 0.25 mg/ml for KMl2SM 37°C and 5% CO, atmosphere, using supplemental cells). To eliminate drug effect, all experiments were McCoy’s 5A medium containing 5% fetal bovine serum. performed in the absence of G418; cells were washed The biological properties of these cells have been with medium and seeded into medium in the absence of described i n detail previously (Chakrabarty et al., drug. 1989a, 1990; Morrikawa et al., 1988; Fan et al., 1989). RNA analysis by Northern hybridization Human EGF was purchased from Boehringer ManCellular RNA was extracted and prepared by the nheim Biochemicals (Indianapolis, IN). Monoclonal anti-human EGF antibody was purchased from Onco- guanidinum extraction and CsCl centrifugation gene Science (Manhasset, NY). Monoclonal anti- method (Ausubel et al., 1989), followed by electrohuman EGF-receptor antibody was purchased from phoretic fractionation in 1.2% agarose-formaldehyde UBI (Lake Placid, NY). 1251-Protein A was purchased gel, and blotted onto nitrocellulose membrane (Ausubel from ICN (Costa Mesa, CAI. Restriction endonucleases et al., 1989). A synthetic antisense oligonucleotide (20 bp corresponding to the 5‘-end of the coding sequence of were purchased from Boehringer Mannheim. the mature human EGF protein) was used as an EGF EGF expression vector and transfection probe. The synthetic EGF probe was labeled with 32Pby The EGF expression vector designated pUCDS3 was the 5’-end labeling method (Maniatis et al., 19821, a generous gift from Dr. David Stern of the Department using gamma 32P-ATP and T4 polynucleotide kinase of Pathology, Yale University School of Medicine. The (Promega, Madison, WI). A 32P-labeledGAPDH cDNA construct of pUCDS3 has been described in detail probe was prepared using the Promega Prime-a-Gene previously (Stern et al., 1987). Briefly, pUCDS3 con- labeling system and was used as a positive control sists of a murine leukemia virus long terminal repeat probe for the amount of RNA loaded per lane in the (LTR) that contains promoterienhancer sequences, a n agarose gel.

222

HUANG ET AL

Analysis of EGF and TGF-alpha secretion

lz5-IEGF binding

A solid-phase radioimmunoassay (RIA), using monoclonal anti-human EGF antibody (Oncogene), was developed to analyze levels of EGF secretion from the cells. The RIA was performed according to previously described procedures (Chakrabarty et al., 1981) with some modification. Briefly, cells were cultured to 7080%confluency in 25 cm2 flasks, after which the cells were washed with serum-free medium and cultured for another 3 days in serum-free medium. The conditioned media were conjugated to RIA star tubes (Nunc, Denmark), blocked with 3% bovine serum albumin (Sigma RIA grade), and incubated with antihuman EGF antibody. After being washed, the tubes were incubated with a n appropriately diluted rabbit antimouse Ig antibody, and the amount of antihuman EGF antibody bound to the RIA tubes was assessed by incubation with 1251-labeledProtein A (ICN). Amount of radioactivity bound to the RIA tubes was determined with a Beckman gamma counter. For background controls, the procedure was performed in the absence of anti-EGF antibody or using a n irrevelant mouse monoclonal antibody, or both. A calibration curve of varying amounts of human EGF (1to 50 ng) conjugated to the RIA tubes was used to quantitate the amount of EGF secreted into the medium. Amount of EGF secreted into medium was normalized to nanogram of secreted EGF per milligram of cell-lysate protein. EGF secreted into the medium was also quantitated by a n enzyme-linked immunosorbent assay (ELISA) inhibition method as described previously (Varani and Chakrabarty, 1990). Briefly, serum-free conditioned medium was used to inhibit binding of anti-EGF antibody to insolubilized EGF on the ELISA wells. Amount of EGF in the medium was estimated from a calibration curve using varying amount of EGF to inhibit binding of anti-EGF antibody to insolubilized EGF. Amount of EGF in medium was normalized to nanogram of EGF per milligram of cell-lysate protein. TGF-alpha in serum-free conditioned medium was analyzed by a commercially available human TGFalpha RIA kit (Biomedical Technologies Inc., Stoughton, MA). This assay was performed according to instructions provided by the manufacturer.

This assay was performed with cells in monolayer growth in 24-well culture plates as previously described (Chakrabarty et al., 1989b; Varani and Chakrabarty, 1990). Human EGF was labeled with NalZ5I using the previously described chloramine-T procedure (Chakrabarty et al., 1990; Varani and Chakrabarty, 1990). Briefly, 200,000 cells were seeded into 24-well tissue-culture plates and allowed to attach overnight at 37°C in a humidified incubator. The cells were then washed twice with ice-cold binding buffer (serum-free McCoy’s medium containing 0.1% bovine serum albumin), followed by incubation with a n acid-salt solution containing 0.2 M acetic acid and 0.5 M NaCl for 5 min a t 0°C. The acid wash was used to remove endogenous EGF bound to its cell-surface receptors (Zheng and Goldsmith, 1990). The cells were then washed twice with binding buffer and varying amounts of radiolabeled EGF were added, after which the cells were incubated for 3.5 h r a t 4°C and shaken gently on a horizontal shaker. Following this, the cells were washed 5 to 6 times with ice-cold binding buffer and solubilized in 0.5 ml of a solution containing 0.1 M NaOH and 1% Triton X-100 at room temperature. The entire contents of the wells were removed to separate test tubes. The wells were then rinsed with 0.5 ml of the same solution, combined with the previously removed contents, and the total radioactivity removed from each well was determined by counting the tubes in a Beckman gamma counter. Background nonspecific binding was determined using 100-fold excess of unlabeled EGF in the assay as described previously (Chakrabarty et al., 198913; Varani and Chakrabarty, 1990). Amount of isotope that was bound nonspecifically was usually about 10% of the total amount of isotope bound in the absence of unlabeled EGF. Data were analyzed by the Scatchard method using a microcomputer and the LIGAND program; they were evaluated for best fit in a two-site ligand-binding model as described previously (Gladhaug and Christofferson, 1989; Eldar e t al., 1990; Lichtner and Schirrmacher, 1990).

Cell growth and proliferation Cell growth and proliferation were assessed by counting cells with the aid of a hemocytometer. Briefly, 100,000 cells were seeded into six-well tissue-culture plates and allowed to grow for a varying number of days. The cells were then removed from the tissueculture wells by trypsinization and the cells’ growth was determined by counting them with a hemocytometer. To assess the effect of anti-EGF or anti-EGF-receptor antibody on cell proliferation, cells were similarly seeded in the absence or presence of various concentrations of antibody. Cell number, after differing periods of growth, was determined with a hemocytometer. An irrelevant monoclonal antibody to nucleolar protein B23 (Chakrabarty et al., 19901, which was used as control, had no effect on the cells’ proliferation.

RESULTS EGF secretion and cell proliferation Four Moser and two KMl2SM G418-resistant, pUCDS3-transfected clones were isolated and established in tissue culture. These cells are listed in Table 1 and their levels of EGF secretion were characterized and compared to that of their parental and control plasmid-transfected cells. Although no detectable EGF was secreted into the medium by parental or control transfected cells according to RIA, a low level of EGF secretion was detectable with the more sensitive ELISA inhibition assay. Amount of EGF secreted by parental and control transfected cells ranged from 0.5 to 1.8 ngimg of cellular proteins, while EGF-transfected Moser clones secreted EGF in the concentration range of 5-14 ngimg of cellular protein (Table 1). EGFtransfected KM12SM cells secreted EGF in the concentration range of 6 1 0 ngimg of cellular protein, with slightly higher values obtained from ELISA (Table 1). Northern hybridization analysis revealed enhanced

223

EGF ACTION IN HUMAN COLONIC CARCINOMA CELLS TABLE 1. Levels of EGF secretion (ng EGF secreted/mg cellular moteind' Cell lines Moser parental Control transfected EGF-I EGF-I1 EGF-111 EGF-IV KM12SM parental Control transfected EGF-I EGF-I1

I

I1

TI1

IV

ND ND 10.9 14.7 6.7 14.6 ND ND 4.7 10.1

ND ND 11.8 15.3 5.7 13.7 ND ND 4.3 10.2

1.4 0.5 12.4 16.3 7.7 12.0

0.8 1.5 10.7 16.1 7.5 11.3 1.8 1.6 6.8 9.1

1.5 1.6 7.4 9.9

'In experiments I and 11, the amounts of EGF in serum-free conditioned media (conditioned for 3 days) were determined by solid-phase RIA using monoclonal antihuman EGF antibody as described in Materials and Methods. Each value represents average of triplicatedeterminations. ND =not detectable. In experiments 111and IV, the amountsof EGF were determined by ELISAinhibitionmethod using the same antihuman EGF antibody as described in Materials and Methods. Each value represents average of quadruplrcate determinations.

EGF mRNA expression in all EGF-transfected clonal cells when compared with the parental cells in which the EGF mRNA was barely detectable (Fig. 1). In view of the mitogenic effect of EGF in tissuecultured cells, we compared the rate of proliferation of the various EGF-transfected cells with that of their parental lines in tissue culture. Figure 2 shows the growth curves of the Moser cells and their EGF transfectants. All transfectants had an enhanced rate of proliferation in comparison with that of the parental Moser or control transfected Moser cells. A maximum increase in growth of about 40% was observed for the EGF-transfected cells on the fourth day of culture. Treatment with exogenous EGF stimulated the parental Moser cells' proliferation in a dose-dependent manner (Fig. 3) which could be blocked completely by anti-EGF antibody (not shown). This result suggested the presence of an EGF autocrine stimulatory mechanism in the EGF-transfected clones. If such a mechanism was responsible for increased growth, anti-EGF antibodies should block the proliferative effect of the secreted EGF. Figure 4 shows that anti-EGF antibody exerted different degrees of growth inhibition on these cells. Most responsive were the Moser EGF-111 cells, which demonstrated a maximum of 53% growth inhibition, and least responsive were the Moser EGF-IV cells, in which only a maximum of 12% growth inhibition was observed (Fig. 4). Anti-EGF antibody did not inhibit the proliferation of the parental Moser cells (not shown). An opposite effect on growth was observed for the EGF-transfected KMl2SM cells (Fig. 5). Both EGFtransfected clones were growth-inhibited by 35-40%, compared with parental or control-transfected cells after 4 days of culture. Unlike the Moser cells, treatment with exogenous EGF did not affect the KM12SM cells' proliferation, and anti-EGF antibody had no effect on the growth of the the KMl2SM parental nor the EGF-transfected KMl2SM clones (Figs. 3, 4).Under identical conditions, a lower range of EGF concentration (0.01-0.1 ng) also did not affect the growth of the KMl2SM cells. This result suggested an intracel-

Fig. 1. Northern analyses of EGF mRNA. Lanes 1 4 3 Moser and EGF-transfected Moser clones. Lanes 7-10 KMl2SM and EGFtransfected KMlZSM clones. Two micrograms of RNA from each cell line was fractionated by agarose-formaldehyde gel electrophoresis and blotted onto nitrocellulose membrane. A Autoradiograph of the membrane hybridized with 32P-synthetic EGF oligonucleotide probe. B Autoradiograph of same membrane rehybridized with 32P-GAPDH cDNA probe after first probe was stripped with 50% formamide in 0.9 M NaCl and 0.09 M sodium citrate at 65°C for 30 min and the membrane washed with this salt solution without formamide. Lane 1, Moser; lane 2, Moser transfected with pSV2-neo; lane 3, Moser EGF-I; lane 4,Moser EGF-11; lane 5, Moser EGF-111; lane 6 , Moser EGF-1% lane 7, KM12SM; lane 8, KMl2SM transfected with pSV2-neo; lane 9, KMl2SM EGF-I; and lane 10, KMl2SM EGF-11.

1.6 1.4 v)

1.2

-

Q)

3

.-=

z

*

-

10.8

-

Q, 0 0.6 -

*

cn

0 -I

0.40.2

-

0

1

2

4

3

5

Days After Seeding

-

Moser

8

EGF-II

+

*

Control EGF-Ill

+&

EGF-I EGF-IV

Fig. 2. Growth of Moser and Moser cells transfected with EGF gene. Cell numbers were determined by counting cells in tissue culture wells with the aid of a hernocytometer as described in Materials and Methods. Values shown represent average of quadruplicate determinations with standard error of *2%. Control, PUCDS5 transfected cells.

224

HUANG ET AL.

mm1m and ~rarrsfectmb

Maser 601

' rc

EGF-I ~ E G F - 1 1

HEW-111BEGF-IV 4

40

0

30 -d

-

20 3 c c

10

bp

1 .o 0.5 1.5 Anti-EGF Antibody (ug)

"

EGF (ng/ml) Fig. 3. Effect of exogenous EGF on growth of Moser and KMl2SM cells. Cell-growth assay was performed as described in Materials and Methods. Cells (100,000) were seeded into 6-well tissue culture plates and allowed to attach overnight. Wells were replenished with fresh medium in the absence (controls) or presence of the indicated concentrations of EGF and incubated in a humidified incubator a t 37°C for 4 days. Cell number in EGF-treated wells was expressed as percentage of cell-growth stimulation in comparison with cell number in control wells. Values shown represent average of triplicate determinations.

Fig. 4. Effect of anti-EGF antibody on growth of EGF-transfected Moser clones, KMlZSM, and its EGF transfectants. Assay was performed similarly to assay described in Figure 3. After 4 day culture period, cell number in antibody-treated wells was compared with that of control wells (no antibody). Cell number in antibody-treated wells was expressed as percentage of cell-growth inhibition in comparison with cell number in control wells. Values shown represent average of quadruplicate determinations with standard error of 23%.

1.2

5 lular or other mechanisms of growth inhibition, one that bypassed the ligand-receptor interaction step.

TGF-alpha secretion and proliferation Many types of transformed cells, including human colon carcinoma cells, produce the EGF-like molecule TGF-alpha, which mediates its action through EGF receptors (Coffey et al., 1987; Anzano et al., 1989; Donaldson et al., 1989). Because levels of TGF-alpha secretion might also influence growth properties by autocrine stimulation, we went on to examine whether growth alterations in the EGF-transfected cells were associated with altered levels of TGF-alpha secretion. Table 2 shows that the Moser parental cells produced a higher amount of TGF-alpha (24 ngimg of protein) than the KM12SM cells (16 ngimg). We observed significant reduction in TGF-alpha production in all EGF-transfected cell lines in comparison with their parental or control-transfected cells (Table 2). Percentages of reduction of TGF-alpha for the EGF-transfected Moser cells ranged from 43 to 71% and from 55 to 60% for the EGF-transfected KM 12SM cells. In spite of the down-modulation of TGF-alpha secretion, EGF-transfected cells still secreted TGF-alpha in the concentration range of 7-13 ng per mg of cellular protein (Table 2). Anti-EGF-receptor antibody was then used to block the binding of both EGF and TGF-alpha to their cell-surface receptors; and its effect on cell proliferation was determined. In a 3 day culture period, anti-EGF-receptor antibody inhibited proliferation of the Moser cells and Moser EGF-transfected

1-

0.8 -

0.6 -

0.4

-

0.2 -

0'

0

++-

I

I

I

I

1

2

3

4

5

Days After Seeding

KMlZSM

+

Control

EGF-I

€k

EGF-II

Fig. 5. Growth of KMlZSM and KMl2SM cells transfected with EGF gene. Assay was performed according to procedures described for Figure 2. Control, PUCDS5 transfected cells.

clones in a dose-dependent manner (Fig. 6). Most responsive were the EGF-I11 and EGF-IV cells in which the cells were growth inhibited by over 40% (Fig. 6). By comparison, under identical conditions, anti-EGF-re-

EGF ACTION IN HUMAN COLONIC CARCINOMA CELLS

225

TABLE 2. Levels of TGF-alpha secretion (ng TGF-alpha secreted/mg cellular proteins) Cell lines Moser parental Control transfected EGF-I EGF-I1 EGF-I11 EGF-IV KMl2SM parental Control transfected

EGF-I EGF-I1

I

I1

Average

22.7 24.9 9.2 13.7 6.9 9.9 14.6 13.7 6.5 7.4

25.4 26.0 8.9 13.3 6.9 9.5 18.2 16.3 6.5 7.1

24.0 25.4 9.1 13.5 6.9 9.7 16.4 15.0 6.5 7.3

% reduction

J z Y 62 43 71 59

5

L

40

c3

'k 30 C 0

60 55

'TGF.alpha in serum-free conditioned medium (conditioned for 3 days) was analyzed by human TGF-alphaRIA kit a s statedin Materialsand Methods. Results of two experiments are shown. Each value representing average of quadruplicate determinations. Percentages of reduction of TGF-alpha secretion were calculated using parental cells' secretion levels as 100%.

ceptor antibody did not exert a growth-inhibitory effect on the parental KMl2SM cells and KMl2SM EGFtransfected clones (Fig. 6).

EGF receptors Since extracellular EGF mediates its action through binding to its cell-surface receptors, and cell-surface receptor density may altered the mitogenic response to EGF (Gill and Lazar, 1981; Bravo et al., 1985;), we analyzed EGF binding to its cell-surface receptors by Scatchard analyses using lz5I-EGF. Table 3 summarizes the results of these EGF-binding studies. In the EGF-transfected Moser clones, the expression of both high and low affinity receptor subtypes were downmodulated when compared with parental and control transfected cells, while a slight decrease in Kd values were observed for both types of receptors (Table 3). By comparison, the KM12SM cells possessed lower EGFreceptor density than the Moser cells, and no obvious receptor alterations were seen in the EGF-transfected KMl2SM cells (Table 3). DISCUSSION This study showed th at the constitutive expression of the human EGF gene in epithelium-derived human colonic carcinoma cells stimulated the growth of moderately differentiated Moser cells and reduced the growth of highly metastatic KMl2SM cells. Unlike transfection of the EGF gene into normal fibroblasts (Stern et al., 19871, no significant morphologic alteration was observed in any of the EGF-transfected cells (not shown). Three of the Moser pUCDS3-transfected clones produced EGF at a concentration of 10-14 ng/mg of protein, whereas one of the clones, EGF-111, secreted only about half of that amount. Yet no significant differences in the growth of these cells (EGF-transfected Moser clones) were observed. In the parental Moser cells, TGF-alpha played a dominant role in autocrine growth stimulation since cell proliferation was reduced by anti-EGF-receptor antibody but not by anti-EGF antibody. By comparision, the enhanced rate of proliferation of Moser EGF-I11 cells could be inhibited by anti-EGF antibody. These results were highly suggestive of a n autocrine mechanism of growth stim-

-4

20

4 f 5 10 be

n.

0.01 0.05 0.1 Anti-EGF-Receptor

0.5

1.0 2.5 Antibody (ug/rnl)

Fig. 6 . Effect of anti-EGF-receptor antibody on growth of Moser, Moser EGF-transfected clones, KMlZSM, and its EGF transfectants. Assay was performed similarly to assay described in Figure 4. After 3 day culture period, cell number in antibody-treated wells was compared with that of control wells (no antibody). Cell number in antibody-treated wells was expressed as percentage of cell-growth inhibition in comparison with cell number in control wells. Values shown represent average of quadruplicate determinations with standard error of 23%.

ulation by EGF in this clone of EGF-transfected cells. Anti-EGF antibody was able to block the proliferation of the Moser EGF-I, -11,and -1V cells by 35,25, and 12%, respectively. However, a higher level of growth inhibition was achieved with anti-EGF-receptor antibody (blocking both EGF and TGF-alpha), suggesting that TGF-alpha also played a role in autocrine growth stimulation in these cells. This was especially true with the EGF-IV cells, which were inhibited by over 40% with anti-EGF receptor antibody, albeit only 12% growth inhibition was obtained with anti-EGF antibody. These data suggested that some of the EGFtransfected Moser cells showed enhanced responsiveness to the growth-stimulatory effect of TGF-alpha, albeit the down-modulation of TGF-alpha secretion. Enhanced responsiveness to TGF-alpha in the EGF-IV cells might be a mechanism through which altered proliferation rate was achieved. A growth-inhibitory effect was observed in the KMl2SM cells as a consequence of constitutive expression of the EGF gene. The parental KMl2SM cells did not respond to exogenous EGF, and both anti-EGF and anti-EGF-receptor antibody did not affect proliferation of the KMl2SM and pUCDS3-transfected cells. These results demonstrated that these cells had lost their ability to respond to growth factors (EGF and TGFalpha) at the cell surface and that non-autocrine or intracellular growth-inhibitory mechanisms, bypassing ligand-receptor interaction, were likely to exist. Interleukin-3 has been shown to stimulate cell growth by intracellular action (bypassing ligand-receptor interaction a t the cell surface) when a n appropriate construct of its gene was transfected into interleukin3-dependent murine 32D cells (Dunbar e t al., 1989).

226

HUANG ET AL.

TABLE 3. lZ51-EGFbinding to human colonic carcinoma cells' Binding sites/cell Cell lines

4.26 4.39 .~ 0.83 1.50 1.20 0.60 1.12 1.35 0.80 1.32

Moser ~~

Contra1 - .......

~

EGF-I EGF-I1 EGF-I11 EGF-IV KMl2SM Control EGF-I EGF-I1 ~~

Kd (M) LA (XIO@)

HA ( ~ 1 0 ~ )LA ( ~ 1 0 ~ )HA (XIO-'~) 3.59 3.59 0.72 1.03 0.95 0.68 1.09 1.15 1.10 1.03

6.16 7.29 3.68 4.55 5.00 3.64 2.97 4.10 1.77 2.33

1.19 0.82 0.57 0.74 0.79 0.91 4.90 5.11 4.56 2.89

~

'Values presented were derived from Scatchard analyses of the binding data as described in Materials and Methods. HA and LA refer to high and low affinity, respectively.

divergent altered growth properties. Growth-stimulatory and growth-inhibitory effects were observed in the moderately differentiated Moser cell line and the highly metastatic KMl2SM cell line, respectively. Growth stimulation by EGF in the Moser cells was associated with a n autocrine mechanism involving both EGF and TGF-alpha, whereas growth inhibition in the KMl2SM cells was non-autocrine or intracellular in nature.

ACKNOWLEDGMENTS This work was partly supported by USPHS grant CA 47775 awarded to S.C.

LITERATURE CITED KMl2SM EGF-I1 cells secreted twice the amount of EGF than the EGF-I cells (Table l ) , and their growth inhibition was comparatively higher (Fig. 5). It is not known if any correlation exists between levels of EGF secretion and degree of growth inhibition in these cells. Overall, we encountered much more difficulty in establishing viable pUCDS3-transfected KMl2SM clones than in establishing clones for the Moser cells. Many clones did not grow and eventually died. It is possible that a high level of EGF expression in the KMl2SM cells was cytotoxic. The growth-inhibitory action of EGF has been reported in epidermal carcinoma A431 cells, which have a n unusually high density of cell-surface EGF receptors (Gill and Lazar, 1981; Bravo et al., 1985). EGF-transfected KMlZSM clones did not show a significant alteration in EGF receptor expression; in fact, the KM12SM cells had a n overall lower density of cellsurface EGF receptor by comparison to that of the parental Moser cells (Table 3). Thus, growth inhibition and cell-surface receptor density were not related in the EGF-transfected KMlBSM clones. As has been shown in NIH-3T3 cells, intracellular mechanisms can activate the EGF-associated signal transduction pathway, in the absence of receptor down-modulation, and result in cellular transformation (Wells et al., 1990). In human colonic carcinoma cells, post-receptor mechanisms seem to be important in governing the type of responses elicited by transforming growth factor-beta (Chakrabarty et al., 1989a, 1990). We are now investigating the intracellular mechanisms of EGF action in these cells. By comparison with EGF-transfected KMl2SM clones, cell-surface EGF receptors were down-modulated in all EGF-transfected Moser clones. The parental Moser cells responded to treatment with exogenous EGF. In tissue culture, exogenous EGF binds to its cell-surface receptors and results in receptor downmodulation (Carpentier et al., 1982). The continuous secretion of EGF and exposure of EGF receptor to its ligand, in the EGF-transfected Moser cells, could account for the down-modulation of receptor in these cells. Clearly, the constitutive expression of the human EGF gene in two phenotypically different, epitheliumderived human colonic carcinoma cells resulted in

Anzano, M.A., Rieman, D., Prichett, W., Bowen-Pope, D.F., and Russel, G. (1989) Growth factor production in human colon carcinoma cell lines. Cancer Res., 49:2898-2904. Ausubel, F.M., Brent, R., Kingston, R.E., Moore, D.D., Seidman, J.G., Smith, J.A., and Struhl, K. (1989) Current Protocols in Molecular Biologv. Greene Publishing Associates and Wilev-Interscience. John%iley and Sons, New qork. Bravo R., Burckhardt, J., Curran, T., and Muller, R. (1985) Stimulation and inhibition of growth by EGF in different A431 cell clones is accompanied by rapid induction of c-fos and c-myc proto-oncogenes. EMBO J., 4:1193-1197. Carpentier, J.-L., Gorden, P., Anderson, R.G.W., Goldstein, J.L., Brown, M.S., Cohen, S., and Orci, L. (1982) Co-localization of '"I-epidermal growth factor and ferritin-low density lipoprotein in coated pits: a quantitative electron microscopic study in normal and mutant human fibroblasts. J . Cell Biol., 95r73-77. Chakrabarty, S. Ekramoddoullah, A.K.M., K i d , F.T., and Sehon, A.H. (1981) Allergenic relationships of two different antigens of Kentucky Blue Grass pollen. Int. Arch. Allergy Appl. Immunol., 66:13-20. Chakrabarty, S., Jan, Y., Brattain, M.G., Tobon, A,, and Varani, J . (1989a) Diverse cellular responses elicited from human colon carcinoma cells by transforming growth factor-B. Cancer Res., 49r2112-2117. Chakrabarty, S., J a n , Y., Levine, A,, McClenic, B., and Varani, J. (198913)Fibronectidlaminin and their receptors in aberrant growth control in Ha-Ras oncogene transformed and epidermal growth factor gene transformed FR3T3 cells. Int. J . Cancer, 44:325-331. Chakrabarty, S., Fan, D., and Varani, J . (1990) Regulation of differentiation and cellular proliferation in human colon carcinoma cells by transforming growth factor beta 1 and beta 2. Int. J. Cancer, 46:493499. Chen, C.A., and Okayama, H. (1988) Calcium phosphate-mediated gene transfer: a highly efficient transfection system for stably transforming cells with plasmid DNA. Bio Techniques, 6:632-637. Coffey, R.J., Jr., Shipley, G.D., and Moses, H.L. (1986) Production of transforming growth factors by human colon cancer cells. Cancer Res., 46: 116P1169. Coffey, R.J. Jr., Goustin, AS., Soderquist, A.M., Shipley, G.D., Wolfshohl, J., Carpenter, G., and Moses, H.L. (1987) Transforming growth factor alpha and beta expression in human colon cancer cell lines: implications for a n autocrine model. Cancer Res., 47:45904594. Cohen, S. (1962) Isolation of a mouse submaxillary protein gland protein accelerating incisor eruption and eyelid opening in the newborn animal. J Biol. Chem., 237:1555-1562. Cohen, S.. and Caruenter, G. (1975) Human euidermal Prowth factor: isolation and chemical and biological properties. P r k . Natl. Acad. Sci. U.S.A., 721317-1321. De Larco, J.E., and Todaro, G.J. (1978)A human fibrosarcoma cell line producing multiplication-stimulating activity (MSA)-related peptides. Nature, 272:356-358. Donaldson, R.W., Nishibe, S., and Carpenter, G. (1989) The structure and physiology of epidermal growth factor and its receptor. In: Advances in Growth Hormone and Growth Factor Research. E.E. Muller, D. Cocchi, and V. Locatelli, eds. Pythagora Press, Roma, and Springer Verlag, Berlin, pp. 166-180. Dunbar, C.E., Browder, T.M., Abrams, J.S., and Nienhuis, A.W.

EGF ACTION IN HUMAN COLONIC CARCINOMA CELLS (1989) Cooh-terminal-modified interleukin-3 is retained intracellularly and stimulates autocrine growth. Science, 245:1493-1496. Eldar, H., Zisman, Y., Ullrich, A,, and Livneh, E. (1990) Overpression of protein kinase C alpha-subtype in Swissi3T fibroblast causes loss of both high and low affinity receptor numbers for epidermal growth factor. J . Biol. Chem., 265:13290-13296. Fan, D., Chakrabarty, S., Seid, C., Bell, C.W., Schackert, H., Morikawa, K., and Fidler, I.J. (1989) Clonal stimulation or inhibition of human colon carcinomas and human renal carcinomas mediated by transforming growth factor B1. Cancer Commun., 1:117-125. Gill, G.N., and Lazar C.S. (1981) Increased phosphotyrosine content and inhibition of proliferation in EGF-treated A431 cells. Nature, 293t305-307. Gladhaug, I.V., and Christofferson, T. (1989) n-Butyrate and dexamethasone synergistically modulate the surface expression of epidermal growth factor receptors in cultured rat hepatocytes. FEBS Lett., 243:21-24. Gospodarowicz, D (1983) Growth Factors and their action in vivo and in vitro. J. Pathol., 141:201-233. Heldin, C-H., and Westermark, B. (1984) Growth factors: mechanism of action and relation to oncogenes. Cell, 37:9-20. Kim, S.J., Angel, P., Lafgastin, R., Hattori, K., and Sporn, M.B. (1990) Autoinduction of transforming growth factor beta 1is modulated by the AP-1 complex. Mol. Cell. Biol., lOr1492-1497. Laiho, M., and Keski-Oja, J . (1989) Growth factors in the regulation of pericellular proteolysis: a review. Cancer Res., 49:2533-2553. Lichtner, R.B., and Schirrmacher, V. (1990) Cellular distribution and biological activity of epidermal growth factor receptors in A432 cells are influenced by cell-cell contact. J. Cell. Physiol., 144t303-312. Maniatis. T.. Fritsch, E.F., and Sambrook, J . (1982) Molecular Cloning: A Laboratory Manual. Cold Spring Harbor Laboratory. Cold Spring Harbor, NY, pp. 122-123. Morrikawa, K., Walker, S.M., Jessup, J.M., and Fidler, I.J. (1988) In vivo selection of highly metastatic cells from surgical specimens of different primary colon carcinomas implanted into nude mice. Cancer Res., 48:1943-1948. Moses, H.L., and Pittelkow, M.R. (1987) Production and auto-induction of transforming growth factor-alpha in human keratinocytes. Nature, 328:8117-8120.

227

Moses, H.L., Coffey, R.J., Jr, Leof, E.B., Lyons, R.M., and Keski-Oja, J. (1987) Transforming growth factor beta regulation of cell proliferation. J. Cell. Physiol. [Suppl.], 5:l-7. Parker, B.A., and Stark, G.R. (1979) Regulation of simian virus 40 transcription: sensitive analysis of the RNA species present early in infections by virus or viral DNA. J . Virology, 31:360-367. Schroy, P., Rifkin, J., Coffey, R.J., Winawer, S., and Friedman, E. (1990) Role of transforming growth factor B1 in the induction of colon carcinoma differentiation by hexamethylene bisacetamide. Cancer Res., 50:261-265. Southern, P.J., and Berg, P. (1982) Transformation of mammalian cells to antibiotic resistance with a bacterial gene under control of the SV40 early region promoter. J. Mol. Appl. Genet., I:327-341. Stern, D.F., Hare, D.L., Cecchini, M.A., and Weinberg, R.W. (1987) Construction of a novel oncogene based on synthetic sequences encoding epidermal growth factor. Science, 235:321-324. Temin, H.M. (1970) Control of multiplication of uninfected rat cells and rat cells converted by murine sarcoma virus. J. Cell. Physiol., 75:101-120. Todaro, G.J., Marquart, H., Twardzik, D.R., Johnson, P.A., Fryling, C.M., and De Larco, J.E. (1982) Transforming growth factors produced by tumor cells. In: Tumor Cell Heterogeneity. A.H. Owens, D.S. Coffey, and S.B. Baylin, eds. Academic Press, New York, pp. 205-224. Varani, J., and Chakrabarty, S. (1990) Fibronectinifibronectin receptor modulation during transformation and differentiation of mouse AKR fibroblasts. J. Cell. Physiol., 143:445454. Velu, T.J., Beguinot, L., Vass, W.C., Willingham, M. C., Merlino, G.T., Pastan, I., and Lowy, D.R. (1987) Epidermal growth factor-dependent transformation by a human EGF receptor proto-oncogen. Science, 238:140%1410. Wells, A,, Welsh, J.B., Lazar, C.S., Wiley, H.S., Gill, G.N., and Rosenfeld, M.G. (1990) Ligand-induced transformation by a noninternalizing epidermal growth factor receptor. Science, 247:962-964. Zheng, Z.-S., and Goldsmith, L.A. (1990) Modulation of epidermal growth factor receptors by retinoic acid in Me180 cells. Cancer Res., 50:1201-1205.