TGFα is required for full expression of the ... - Semantic Scholar

Download PDF
3 downloads 0 Views 222KB Size Report
Comment
Hairong Geng1, Paul H.Naylor2, Julie Dosescu1, associated with cell transformation (4,5) and tumorigenesis in animal models (6–8) and humans (9,10).
Carcinogenesis vol.21 no.4 pp.567–572, 2000

TGFα is required for full expression of the transformed growth phenotype of NIH 3T3 cells overexpressing ornithine decarboxylase

Hairong Geng1, Paul H.Naylor2, Julie Dosescu1, Magdalena Skunca1, Adhip P.N.Majumdar2,3 and Jeffrey A.Moshier1,3,4 1Center

for Molecular Medicine and Genetics and 2Department of Internal Medicine, Wayne State University School of Medicine and 3John D.Dingell VA Medical Center, Detroit, MI 48201, USA 4To

whom correspondence should be addressed at: William Tyndale College, 35700 W. Twelve Mile Rd, Farmington Hills, MI 48331, USA Email: [email protected]

Ornithine decarboxylase (ODC) overexpressed from a heterologous promoter drives the tumorigenic transformation of NIH 3T3 cells and provides a model to investigate the underlying molecular mechanisms. These transformed cells, designated NODC cells, exhibit elevated levels of epidermal growth factor receptor (EGFR) tyrosine kinase (Tyr-k) activity relative to control transfected cells and inhibition of EGFR Tyr-k activation suppresses the transformed growth phenotype of these cells. Thus, ODCinduced transformation of NIH 3T3 cells appears to be mediated, at least in part, by enhanced signaling through the EGFR pathway. Here we extend these studies by evaluating: (i) the effects on growth regulation of overexpressing ODC in EGFR-deficient NIH 3T3 cells; (ii) the potential role of TGFα in mediating the EGFR-dependent transformation of NIH 3T3 cells by ODC. Disruption of EGFR–TGFα interactions either by deleting EGFR, by treatment with anti-TGFα neutralizing antibody or by transfection with a TGFα antisense expression vector suppressed acquisition of the full transformed growth phenotype. Specifically, the loss of contact inhibition and the capacity for clonogenic growth appear more dependent on EGFR–TGFα interactions than anchorage-independent growth in ODC-overexpressing cells. ODC overexpression does not alter the amount, localization or secretion of TGFα. Thus, TGFα is not the ODC-responsive component of the EGFR signaling pathway but appears to be critically involved in development of the transformed phenotype of NODC cells. Introduction Putrescine, spermidine and spermine, collectively called polyamines, are ubiquitous aliphatic polycations which play critical, though largely undefined, roles in cell proliferation and differentiation (1–3). It is not surprising, therefore, that disruption of the intricate mechanisms regulating polyamine metabolism has severe consequences on cell function. For example, elevated expression of ornithine decarboxylase (ODC), the first key enzyme in polyamine biosynthesis, has long been intimately Abbreviations: DMEM, Dulbecco’s modified Eagle’s medium; EGF, epidermal growth factor; EGFR, epidermal growth factor receptor; FBS, fetal bovine serum; ODC, ornithine decarboxylase; PBS, phosphate-buffered saline; TGFα, transforming growth factor α; Tyr-k, tyrosine kinase. © Oxford University Press

associated with cell transformation (4,5) and tumorigenesis in animal models (6–8) and humans (9,10). More directly, ectopic overexpression of ODC results in the loss of normal growth controls and the tumorigenic transformation of mesenchymal and epithelial cells under certain conditions (11–14). In our studies, ODC overexpressed from a β-actin promoter directed the tumorigenic transformation of NIH 3T3 fibroblasts, as demonstrated by the loss of contact inhibition and the acquisition of anchorage independence, efficient clonogenic growth and tumorigenicity in athymic mice (11). However, the mechanisms by which ODC overexpression deregulates normal growth controls have yet to be elucidated. We reported that ODCtransformed NIH 3T3 cells displayed elevated levels of epidermal growth factor receptor (EGFR) tyrosine kinase (Tyrk) activity. Down-regulation of EGFR Tyr-k activity by the pharmacological inhibitor tyrphostin-25 or by expression of a dominant negative EGFR mutant suppressed the transformed growth phenotype of these cells (15). Thus, ODC-induced transformation of NIH 3T3 cells appears to be mediated, at least in part, by elevated signaling through the EGFR pathway. How ODC overproduction modulates EGFR Tyr-k activity is not known. Aberrant activation or production of EGFR and its ligand transforming growth factor α (TGFα) are strongly implicated in the development and progression of neoplastic diseases (16,17). TGFα was discovered and purified on the basis of its ability to promote transformation of normal cultured fibroblasts (18) and most oncogene-transformed NIH 3T3 fibroblasts overproduce TGFα (19). However, several types of rodent fibroblasts, including NIH 3T3 cells, are not transformed by the overproduction of either EGFR or TGFα alone, but only when the receptor and ligand are co-overexpressed (16,17,20,21); a common condition in human tumors and tumor-derived cell lines (22). The coupling of EGFR/TGFα deregulation and tumorigenesis led us to evaluate the potential role of TGFα in mediating the EGFR-dependent transformation of NIH 3T3 cells by ODC overexpression. The results presented here indicate that TGFα production is not the ODC-responsive component of the EGFR signaling pathway but that endogenous levels of TGFα are critical to the fully transformed growth phenotype of ODC-overexpressing NIH 3T3 cells. Materials and methods Cell culture and transfection conditions NODC and NLK cell strains were generated in our laboratory by stably transfecting NIH 3T3 cells with ODC expression vector pβ-ODC or control vector, respectively, as previously described (11). NIH 3T3w cells, an EGFRdeficient variant of NIH 3T3 cells (23), were kindly provided by Dr M.I. Greene (University of Pennsylvania, Philadelphia, PA). All cell lines were cultured in Dulbecco’s modified Eagle’s medium (DMEM) with high glucose and 10% fetal bovine serum (FBS) (Life Technologies, Gaithersburg, MD) at 37°C in 5% CO2. Cell transfections were performed by the standard calcium phosphate precipitation protocol (24) using the CellPhect transfection kit (Pharmacia LKB Biotechnology, Piscataway, NJ) according to the manufacturer’s instructions. NIH 3T3w cells were transfected with plasmid pβ-ODC and selected

567

H.Geng et al. in DMEM with 10% FBS containing 400 µg/ml geneticin (Life Technologies). NODC-2 cells were co-transfected with either TGFα antisense expression plasmid pCMV-AS-TGF (a generous gift of Dr G.Filmus, University of Toronto, Canada) or control vector pRc/CMV and vector pMSG (Pharmacia LKB Biotechnology) for positive selection in DMEM with 10% FBS containing xanthine, hypoxanthine, aminopterin and mycophenolic acid as described (25). For either selection protocol, the medium was changed every 3 days for 2 weeks to remove dead cells. Some selected cells were isolated with cloning cylinders and the rest were pooled as a transfected population. Tumorigenic growth assays For contact inhibition assays, cells were seeded at 1⫻105 per T-25 flask or 60 mm plate in DMEM with 10% FBS. Culture medium was replaced every 3 days for 2–3 weeks, at which time cells were washed with phosphatebuffered saline (PBS), fixed in methanol for 4 min, dried and stained with 6% Giemsa solution. Cells were then washed with distilled water, dried and photographed. For the TGFα neutralization studies, 5 µg/ml of anti-TGFα antibody (Santa Cruz Biotechnology, Santa Cruz, CA) or mouse IgG (Sigma Chemical Co., St Louis, MO) was added 16 h after seeding and fresh antibody was added with each change of medium. Anchorage-independent growth assays were performed in soft agar as described (11). Briefly, cells were seeded at a concentration of 1⫻103 cells/ 35 mm plate in DMEM supplemented with 18% FBS, 6.7 mM sodium bicarbonate, 1.25 mM sodium pyruvate, 0.05 mM essential and non-essential amino acids, 2 mM glutamine, 0.1 mM serine, 0.15 mM asparagine, 10 U/ml penicillin, and 120 µg/ml streptomycin gelled with 0.3% agarose (SeaPlaque; FMC, Rockland, ME) over a 0.9% agarose underlayer containing 2.5 ml of conditioned medium and 4.5 ml of the above DMEM-enriched medium. Enriched medium (100 µl/plate) was added to each culture the day after seeding and then every 4 days. Colonies were counted 2 weeks after seeding. For clonogenic assays on plastic plates, 1⫻103 cells/60 mm plate were seeded in DMEM with 10% FBS. For TGFα neutralization studies, 5 µg/ml of anti-TGFα antibody (Santa Cruz Biotechnology) or mouse IgG (SigmaAldrich, St Louis, MO) was added the next day. Cells were refed every 3 days with fresh medium and antibody. Colonies were counted after staining as described above. EGFR activity assay EGFR Tyr-K activity was determined by measuring the incorporation of 32P into the synthetic polymer L-Glu-L-Tyr (4:1) from [γ-32P]ATP as previously described (15). The results are expressed as pmol 32P incorporated/mg protein. For TGFα induction, cell extracts were incubated with 10 nM TGFα for 15 min at 4°C before immunoprecipitation. Western blot analyses Exponentially growing cells were solubilized in lysis buffer (1% Triton X100, 0.5% NP-40, 10% glycerol, 50 mM HEPES, pH 7.2, 100 mM NaCl, 1 mM Na3VO4, 1 mM PMSF, 10 µg/ml leupeptin, 10 µg/ml aprotinin and 544 µM iodoacetamide) for 15 min at 4°C. The scraped lysate was clarified by centrifugation at 1⫻104 g for 10 min. For membrane and cytosol preparations, the cell pellet was resuspended in homogenization buffer (50 mM HEPES, pH 7.2, 0.1 mM EDTA, 0.04% Brij 35, 4 mM dithiothreitol), sonicated for 10 s three times and centrifuged at 3⫻104 g for 30 min at 4°C. The supernatant was recovered as the cytosol preparation. The pellet was resuspended in homogenization buffer, solubilized for 30 min at 4°C and fractionated by centrifugation at 1⫻104 g for 10 min at 4°C. The membrane fraction was recovered from the supernatant. Protein concentrations were determined by the method of Bradford (26). For TGFα detection, aliquots of cell lysate were fractionated on 15% SDS– PAGE gels and transferred to PolyScreen polyvinylidene difluoride membrane (Dupont NEN Research Products) using a Trans-Blot semi-dry electrophoretic transfer apparatus (Bio-Rad, Hercules, CA). Transfer buffer was composed of 48 mM Tris, 39 mM glycine, 20% methanol and 1.3 mM SDS. Membranes and gels were subsequently stained with Ponceau S and Coomasie blue, respectively, to evaluate the equity of protein loading and transfer efficiency. Membranes were air dried, blotted in 5% non-fat milk in PBS containing Tween 20 (9.4 mM sodium phosphate, pH 7.4, 145 mM NaCl, 0.5% Tween-20) overnight at 4°C, incubated with mouse anti-TGFα (Santa Cruz Biotechnology) and horseradish peroxidase-conjugated anti-mouse IgG (ICN Pharmaceuticals) for 1.5 h at room temperature. Afterwards, membranes were incubated with ECL chemiluminescence reagent (Dupont NEN) for 1 min and exposed to X-ray film. Signal intensities were quantitated using a Personal Densitometer SI (Molecular Dynamics, Sunnyvale, CA). Several westerns were performed for each study utilizing multiple, independent cell extracts. TGFα ELISA Conditioned medium was collected from cells established in culture in DMEM with 10% FBS for 2 days, washed twice with DMEM and cultured in DMEM

568

Fig. 1. EGFR Tyr-K activity levels of ODC-transformed NODC and control transfected NLK cells. Cell extracts were incubated with anti-EGFR antibody and the immunoprecipitates assayed for Tyr-k activity by measuring the transfer of 32P from [γ-32P]ATP to the synthetic polymer L-Glu-L-Tyr (4:1). The results are expressed relative to the Tyr-k activity of control NLK-3 cells. Each value represents the mean ⫾ SEM. *Values differing significantly from those of NLK-3 cells (P ⬍ 0.05).

with 0.5% bovine calf serum for 48 h. Fifty microliters of conditioned medium was incubated overnight with an equivalent volume of anti-TGFα monoclonal antibody (1:4000 dilution; Santa Cruz Biotechnology) at 4°C. The amount of free antibody was measured by adding the sample to a TGFα-coated microtiter well that was blocked with 7% non-fat milk and 0.5% bovine serum albumin. After an overnight incubation at 4°C, the plate was washed with PBS, incubated with biotinylated anti-mouse IgG (1:2000 dilution; Vector Laboratories) at room temperature for 1 h, washed again and then incubated with avidin– biotin–alkaline phosphatase complex (1:1000 dilution; Vector Laboratories) at room temperature for 1 h. After final washings with PBS and then D buffer (1% diethanolamine and 500 µM MgCl2), 200 µl of p-nitrophenylphosphate (1 mg/ml) in D buffer was added to each well and the color intensity at 405 nm was measured in an automated microtiter plate reader. Varying concentrations of TGFα (Oncogene Research Products, Cambridge, MA) were run concurrently as standards.

Results EGFR Tyr-k activities in three ODC-transformed NIH 3T3 subclones (NODC-2, -5 and -6) and control subclones (NLK2 and -3) were measured to determine whether the correlation between ODC overexpression and EGFR Tyr-k activation was a common feature of independently derived NODC clones. Figure 1 shows that EGFR Tyr-k activities were at least 3fold greater in each NODC subclone compared with the control NLK subclones. The amplitude of the EGFR Tyr-k response in ODC-transformed fibroblasts may be influenced by other cellular parameters since there is no direct quantitative correlation between the levels of ODC and EGFR Tyr-k activity in these clones. To evaluate the extent to which ODC-induced transformed growth is dependent on EGFR Tyr-k activation, we examined the effect of ODC overexpression on the growth of NIH 3T3w cells, a NIH 3T3 cell line lacking EGFR (23). NIH 3T3w cells stably transfected with pβ-ODC were designated NWODC cells and were shown to have a 4- to 5-fold increase in ODC activity compared with control transfected NIH 3T3w cells (data not shown). The results in Table I show that parental NIH 3T3w cells exhibited a greater capacity for anchorageindependent growth than control NLK-3 cells. Therefore, even though the total number of anchorage-independent colonies formed by NWODC cells was 78% of control NODC-2 cells, the increase in colonies was 9-fold for NWODC compared

TGFα and ODC-induced transformation

Table I. Characteristics of NIH 3T3w cells transfected with ODC expression vector

Table II. Effects of TGFα neutralizing antibody on NODC-2 tumorigenic growth

Cell line

Anchorage-independent colonies (no./plate, mean ⫾ SEM)a

Foci (no./plate, mean ⫾ SEM)b

Cell line

Anchorage-independent colonies (no./plate, mean ⫾ SEM)a

Foci (no./plate, mean ⫾ SEM)b

NLK-3 NODC-2 NIH 3T3w NWODC

1.0 254.0 22.0 197.5

0 20.7 ⫾ 3.5 0 3.8 ⫾ 1.9c

Untreated TGFα Ab IgG

202.3 ⫾ 7.2 46.5 ⫾ 1.5 203.0 ⫾ 1.0

11.5 ⫾ 0.5 1.0 ⫾ 0.0 NDc

⫾ ⫾ ⫾ ⫾

0.0 5.7 1.4 14.8c

aCells

in DMEM-enriched medium were seeded at a density of 1⫻103 cells/ 35 mm plate in 0.3% agarose over a 1 ml layer of 0.9% agarose. bCells in DMEM ⫹ 10% FBS were seeded at a density of 1⫻105 cells/T-25 flask. cSignificantly different from NODC-2 cells (P ⬍ 0.001).

in DMEM ⫹ 10% FBS were seeded at a density of 1⫻103 cells/60 mm plate (n ⫽ 3). bCells in DMEM ⫹ 10% FBS were seeded at a density of 1⫻105 cells/T-25 flask (n ⫽ 3). cND, not determined. aCells

Fig. 3. Western blot analysis of TGFα protein levels in NODC-asT cells. Extracts of NLK-3 and NODC-2 cells and NODC-2 cells stably transfected with an antisense TGFα expression plasmid (NODC-asT) were fractionated on the same 15% SDS–PAGE gel with mature TGFα peptide (5.6 kDa) included as a positive control. After transfer to PVDF membrane, the blot was incubated with mouse anti-TGFα and horseradish peroxidaseconjugated anti-mouse IgG. Afterwards, membranes were incubated with chemiluminescence reagent and exposed to film as described in Materials and methods. The lanes shown were from the same blot and the same exposure but were rearranged for presentation. Fig. 2. Reduced clonogenic growth of NODC-2 cells treated with TGFα neutralizing antibody. NODC-2 cells were seeded in DMEM with 10% FBS at a concentration of 1⫻103 cells/60 mm plate and either 5 µg/ml of TGFα neutralizing antibody or mouse IgG was added to the culture the next day. Cells were refed every 3 days with fresh medium and antibody. After 1 week, cells were washed with PBS, fixed in methanol and stained with 6% Giemsa solution.

with a 250-fold increase for NODC-2 cells. On the other hand, foci formation was not detected for either control NLK-3 or NIH 3T3w cells. The number of foci formed by NWODC cells was 18% of those produced by NODC-2 cells. Thus, the lack of EGFR interferes with at least two in vitro growth phenotypes associated with the ODC-induced transformation of NIH 3T3 cells. We next focused on the role of TGFα in the deregulated growth of ODC-transformed NIH 3T3 cells. We first determined the effect of disrupting EGFR and TGFα interactions with TGFα neutralizing antibody on the transformed growth of NODC-2 cells. In our hands, TGFα neutralizing antibody inhibits EGFR Tyr-k activity as well as incorporation of [3H]thymidine into the DNA of cells stimulated with exogenous TGFα (data not shown). The results in Figure 2 and Table II show that the anti-TGFα antibody reduced the clonogenic growth of NODC cells to 25% of untreated and IgG-treated cultures. In addition, non-ODC-transformed NODC cells grew

very poorly and displayed morphological changes under the control conditions used for clonogenic growth. The results in Table II also show that foci formation by NODC-2 cells was efficiently suppressed by the anti-TGFα antibody as well. Taken together, these results suggest that TGFα plays a role in the transformed growth of NODC cells and that a critical portion of the elevated EGFR Tyr-k activity detected in these cells is ligand dependent. In a complementary approach, we evaluated the impact of TGFα antisense RNA expression on growth regulation of NODC-2 in vitro. NODC-2 cells were stably transfected with either pCMV-AS-TGF vector, which constitutively expresses TGFα antisense RNA from a CMV promoter (27), or control vector pRc/CMV without insert. Selected transfected colonies, represented by clones NODC-asT and NODC-Rc, respectively, were harvested and characterized. As shown by densitometric scanning of the representative western blot in Figure 3, the level of TGFα precursor protein (18 kDa) in NODC-asT cells is reduced by 50–66% compared with the levels detected in NLK-3, NODC-2 and NODC-Rc cells (not shown). The mature form of TGFα (5.6 kDa) is usually secreted and was not detected in any of the NIH 3T3 variant sublines. The effect of lowered TGFα precursor production in NODC-2 cells was correlated with decreased tumorigenic growth, as shown by 569

H.Geng et al.

Table III. Effect of TGFα antisense expression on NODC-2 tumorigenic growth Cell line

Anchorage-independent colonies (no./plate, mean ⫾ SEM)a

Foci (no./plate, mean ⫾ SEM)b

NLK-3 NODC-2 NODC-Rc NODC-asT

1.7 ⫾ 0.3 145.7 ⫾ 3.0 117.4 ⫾ 3.9 47.3 ⫾ 1.8c

0 21.4 ⫾ 1.4 14.8 ⫾ 1.6 0c

aCells

in DMEM-enriched medium were seeded at a density of 1⫻103 cells/ 35 mm plate in 0.3% agarose over a 1 ml layer of 0.9% agarose. bCells in DMEM ⫹ 10% FBS were seeded at a density of 1⫻105 cells/T-25 flask. cSignificantly different from NODC-2 and NODC-Rc cells (P ⬍ 0.001).

Fig. 5. Analysis of TGFα levels detected in NODC-2 and NLK-3 conditioned medium. Exponentially growing cells were washed and the medium replaced with DMEM containing 0.5% bovine calf serum. Aliquots of medium were removed 48 h later and TGFα levels were determined by competitive ELISA. Medium alone was analyzed for background levels of TGFα. The concentration of TGFα is given as ng/ml medium. The values shown represent the means ⫾ SEM for each sample. *P ⬍ 0.05 compared with medium values and P ⬎ 0.05 compared with NLK-3 values.

by competitive ELISA. As shown in Figure 5, TGFα is secreted into the medium by both NODC-2 and NLK-3 cells, however, the amount secreted by ODC-transformed cells is not statistically higher than that produced by control cells. This result was confirmed by a capture ELISA assay (data not shown). Discussion

Fig. 4. Western blot analysis of TGFα protein levels and localization in NODC-2 and NLK-3 cells. Exponentially growing cells were lysed and the membrane and cytosolic fractions recovered as described in Materials and methods. Aliquots of total cell extracts or membrane and cytosol fractions were separated by 15% SDS–PAGE and probed with anti-TGFα antibody as described in Figure 3.

the data in Table III. NODC-asT cells exhibited complete restoration of growth regulation by contact inhibition and a reduction in anchorage-independent growth of 60 and 66% compared with NODC-Rc and NODC-2 cells, respectively. Similar to the results of experiments directly inhibiting EGFR, the perturbation TGFα–EGFR interactions appear to be more important to the loss of contact inhibition than to anchorageindependent growth of ODC-transformed NIH 3T3 cells. A comparison of TGFα protein levels in NLK-3 and NODC2 cells in Figure 3 suggests that ODC-induced transformation of NIH 3T3 cells does not result in enhanced production of TGFα. However, the amount of TGFα on the cell membrane or secreted into the culture medium may be affected by ODC overexpression. We compared the levels of TGFα protein in the cytosol and membrane fractions of NLK-3 and NODC-2 cells by western blot analysis. Densitometric scanning of the representative western blot in Figure 4 shows that there is no difference in the total amount of TGFα precursor or in the relative levels of TGFα precursor in cytosolic or membrane fractions of non-transformed and ODC-transformed NIH 3T3 fibroblasts. These results were supported by studies in which total TGFα levels were determined by immunoprecipitation of radiolabeled TGFα from NLK-3 and NODC-2 cells incubated in medium containing [35S]methionine (data not shown). The levels of TGFα secreted by NLK-3 and NODC-2 cells into the culture medium over a 48 h period were quantitated 570

We have extended our previous findings that ODC-induced growth transformation of NIH 3T3 fibroblasts is mediated, in part, by signaling through the EGFR pathway (15). The demonstration that EGFR Tyr-k activity is elevated in three independent ODC-transformed NIH 3T3 subclones but not in transfected control cells reduces the possibility that the correlation between ODC and EGF-R Tyr-k activity levels is due to an insertional position effect or to random heterogeneity of EGFR Tyr-k levels in the parental NIH 3T3 cell population. We also showed that NIH 3T3 cells lacking EGFR (NIH 3T3w cells) are not fully transformed by ODC overexpression. The absence of functional EGFR interfered to a greater extent with the ODC-induced loss of contact inhibition than with the acquisition of anchorage-independent growth. This may merely reflect the increased proclivity of NIH 3T3w cells for anchorage-independent growth relative to NIH 3T3 cells from which these cells were derived over a decade ago. Alternatively, ODC induction of anchorage-independent growth may be less dependent on EGFR activation than other parameters of cell transformation, such as loss of contact inhibition or clonogenic growth. In support of the latter interpretation, NODC-2 cells treated with TGFα neutralizing antibody or transfected with a TGFα antisense expression vector also reacquired ‘normal’ growth control, as measured by the reversal of contact inhibition, clonogenic growth and anchorage-independent growth. Again, contact inhibition and clonogenic growth were more strongly correlated with TGFα levels than was anchorageindependent growth. These results may be explained by recent findings that TGFα appears to function at the point of growth initiation, by moving non-cycling cells into the cell cycle, rather than by supporting the growth of already cycling cells (28). The transformed growth parameters displayed by NODC cells which most rely on cell cycle reinitiation, namely foci formation and clonogenic growth, are the parameters which are most dependent on EGFR/TGFα signaling. We conclude from these and a previous study (15) that elevated signaling

TGFα and ODC-induced transformation

through the EGFR/TGFα pathway is necessary for ODCoverexpressing NIH 3T3 cells to exhibit a fully transformed phenotype in vitro. Despite the critical role of TGFα in maintaining the transformed phenotype of NODC cells, ODC overexpression did not significantly alter the amount or localization of precursor TGFα within the cells or the level of mature growth factor secreted. These results differ from the reported effects on TGFα expression in NIH 3T3 cells transformed by a wide variety of viral oncogenes and protooncogenes (19). NIH 3T3 cells transformed with v-Ki-ras, v-mos, v-fes, v-src, activated c-met and c-trk all secreted 5- to 10-fold higher levels of TGFα protein than control NIH 3T3 cells. In addition, there was a concomitant 75–95% reduction in 125I-labeled epidermal growth factor (EGF) binding to these transformed cells compared with parental cells, possibly reflecting enhanced internalization of ligand–receptor complexes. However, the link between oncogenic transformation and elevated TGFα expression was uncoupled in two cases. SV40-transformed NIH 3T3 cells, like NODC-2 cells, did not produce significantly higher levels of TGFα and a morphologically flat Ki-ras cellular revertant produced as much TGFα as the fully transformed Ki-ras NIH 3T3 cells. Thus, elevated TGFα expression is a common feature of the oncogenic transformation of NIH 3T3 cells, but is not obligatory. The dependence on TGFα production for maintaining transformed growth phenotypes was not evaluated in this panel of oncogene-transformed NIH 3T3 cells. We consistently observe that ectopically overexpressed ODC stimulates EGFR Tyr-k activity in NIH 3T3 cells with no detectable changes in EGFR mRNA or protein level (15) or EGF binding levels (unpublished results). In contrast, Ho¨ llta¨ and associates reported that ODC-transformed NIH 3T3 cells down-regulate multiple growth factor receptor protein tyrosine kinases, including EGFR (12,29). These seemingly divergent observations are likely due to the fact that ODC is overexpressed as much as 100-fold in their model compared with the ~5-fold overexpression in our model. It is not clear what changes in polyamine metabolism, regulation or localization mediate the transformation of NIH 3T3 cells by ODC deregulation, but the mechanism appears to be dependent upon the cooperative activation of EGFR when ODC overexpression is modest. A portion of this activation appears to be dependent on endogenous levels of TGFα and the remainder on unknown processes associated with ODC overexpression. Recently, Hamilton and Wolfman (30) showed that EGFR activation was necessary to produce a fully transformed phenotype in mouse fibroblasts when activated Ha-ras was minimally expressed, but not when Ha-ras was overexpressed to ‘supraphysiological’ levels. They concluded that high expression ‘activated all Ras-dependent signaling cascades in a persistent but promiscuous manner’, whereas low Ha-ras expression stimulates a subset of signaling cascades which depend on the cooperative activity of TGFα-activated EGFR to affect transformation. Whether these observations and interpretations are pertinent to ODC-induced cell transformation remains to be tested. In any event, differences in signaling and growthrelated processes induced by a given oncogene may be real and reflect alternative mechanisms influenced by varying expression levels. We have not identified the component(s) of the EGFR signaling pathway which is responsive to ODC overexpression in the multiple independent NODC cell strains. To date, only

EGFR Tyr-k activation is altered in these cells. We see no effect of ODC overexpression on any other EGFR signaling parameter examined, including: short-term TGFα production, secretion or localization; EGFR mRNA or protein levels; EGF binding levels. Moreover, if EGFR activation was directly regulated by ODC or its by-products, we would predict that EGFR Tyr-k activity levels should revert to normal levels when NODC cells are treated with low concentrations of DFMO (50 µM) sufficient to return ODC to parental cell levels and to suppress in vitro transformed growth parameters (15). That this does not occur may be due to the typically slow response of polyamine metabolism to ODC modulation (31) or to an indirect effect on EGFR signaling. Faaland et al. (32) showed that exogenous polyamines modulated EGFR Tyr-k activity in intact A431 cells 16–24 h after treatment, but did not influence EGFR levels, EGF binding activity or directly interact with membrane bound EGFR. They postulated that polyamines may act indirectly to modulate EGFR, possibly by regulating phosphatases, which in turn are important regulators of EGFR function. Further experiments are required to determine whether phosphatases or other signaling components are responsible for the cooperative activation of EGFR required for the full transformation phenotype exhibited by NIH 3T3 cells overexpressing ODC within physiologically relevant levels. Acknowledgement This work was supported by a VA Merit Review grant from the Department of Veterans Affairs (J.A.M.).

References 1. Marton,L.J. and Pegg,A.E. (1995) Polyamines as targets for therapeutic intervention. Annu. Rev. Pharmacol. Toxicol., 35, 55–91. 2. Morgan,D.M. and Wallace,H.M. (1994) Polyamines in clinical and basic science. Biochem. Soc. Trans., 22, 845–898. 3. McCann,P.P., Pegg,A.E. and Sjoerdsma,A. (1987) Inhibition of Polyamine Metabolism: Biological Significance and Basis for New Therapies. Academic Press, New York, NY. 4. Ho¨ llta¨ ,E., Auvinen,M. and Andersson,L.C. (1993) Polyamines are essential for cell transformation by pp60v–src: delineation of molecular events relevant for the transformation phenotype. J. Cell Biol., 122, 903–914. 5. Ho¨ llta¨ ,E., Sistonen,L. and Alitalo,K. (1988) The mechanism of ornithine decarboxylase deregulation in c-Ha-ras oncogene-transformed NIH3T3 cells. J. Biol. Chem., 263, 4500–4507. 6. O’Brien,T.G., Simsiman,R.C. and Boutwell,R.K. (1975) Induction of the polyamine biosynthetic enzymes in mouse epidermis by tumor promoting agents. Cancer Res., 35, 1662–1670. 7. Kingsnorth,A.N., King,W.W.K., Diekema,K.A., McCann,P.P., Ross,J.S. and Malt,R.A. (1983) Inhibition of ornithine decarboxylase with 2difluoromethylornithine: reduced incidence of diamethylhydrazine-induced colon tumors in mice. Cancer Res., 43, 2545–2549. 8. Homma,Y., Ozona,S., Numata, I., Seidenfeld,J. and Oyasu,R. (1985) Inhibition of carcinogenesis by α-difluoromethylornithine in heterotopically transplanted rat urinary bladders. Cancer Res., 45, 648–652. 9. Luk,G.D. and Baylin,S.B. (1984) Ornithine decarboxylase as a biological marker in familial colonic polyposis. N. Engl. J. Med., 311, 80–83. 10. Giardiello,F.M., Hamilton,S.R., Hylind,L.M., Yang,V.W., Tamez,P. and Casero,R.A. (1997) Ornithine decarboxylase and polyamines in familial adenomatous polyposis. Cancer Res., 57, 199–201. 11. Moshier,J.A., Dosescu,J., Skunca,M. and Luk,G.D. (1993) Transformation of NIH/3T3 cells by ornithine decarboxylase overexpression. Cancer Res., 53, 2618–2622. 12. Auvinen,M., Paasinen,A., Anderssen,L.C. and Ho¨ llta¨ ,E. (1992) Ornithine decarboxylase is critical for cell transformation. Nature, 360, 355–358. 13. Megosh,L., Gilmour,S.K., Rosson,D., Soler,A.P., Blessing,M., Sawicki,J.A. and O’Brien,T.G. (1995) Increased frequency of spontaneous skin tumors in transgenic mice which overexpress ornithine decarboxylase. Cancer Res., 55, 4205–4209. 14. Clifford,A., Morgan,D., Yusoa,S.H., Soler,A.P. and Gilmour,S. (1995) Role

571

H.Geng et al. of ornithine decarboxylase in epidermal tumorigenesis. Cancer Res., 55, 1680–1686. 15. Moshier,J.A., Malecka-Panas,E., Geng,H., Dosescu,J., Tureaud,J., Skunca,M. and Majumdar,A.P.N. (1995) Ornithine decarboxylase transformation of NIH/3T3 cells is mediated by altered epidermal growth factor receptor activity. Cancer Res., 55, 5358–5365. 16. Khazaie,K., Schirrmacher,V. and Lichtner,R.B. (1993) EFG receptor in neoplasia and metastasis. Cancer Metastasis Rev., 12, 255–274. 17. Lee,D.C., Fenton,S.E., Berkowitz,E.A. and Hissong,M,A, (1995) Transforming growth factor α: expression, regulation and biological activities. Pharmacol. Rev., 47, 51–85. 18. Rosenthal,A., Linquist,P.B., Bringman,T.S., Goeddel,D.V. and Derynck,R. (1986) Expression in rat fibroblasts of a human transforming growth factor-α cDNA results in transformation. Cell, 46, 301–309. 19. Ciardiello,F., Valverius,E.M., Colucci-D’Amato,G.L., Kim,N., Bassin,R.H. and Salomon,D.S. (1990) Differential growth factor expression in transformed mouse NIH-3T3 cells. J. Cell. Biochem., 42, 45–57. 20. Finzi,E., Fleming,T., Segatto,O., Pennington,C.Y., Bringman,T.S., Derynck,R. and Aaronson,S.A. (1987) The human transforming growth factor coding sequence is not a direct-acting oncogene when overexpressed in NIH 3T3 cells. Proc. Natl Acad. Sci. USA, 84, 3733–3737. 21. DiMarco,E., Pierce,J.H., Fleming,T.P., Kraus,M.H., Molloy,C.J., Aaronson,S.A. and DiFiore,P.P. (1989) Autocrine interaction between TGFα and the EGF-receptor: quantitative requirements for induction of the malignant phenotype. Oncogene, 4, 831–838. 22. Derynck,R. (1992) The physiology of transforming growth factor-α. Adv. Cancer Res., 58, 27–52. 23. Kokai,Y., Myers,J.N., Wada,T., Brown,V. I., LeVea,C.M., Davis,J.G., Dobashi,K. and Greene,M.I. (1989) Synergistic interaction of p185C-neu and EGF receptor leads to transformation of rodent fibroblasts. Cell, 58, 287–292. 24. Gorman,C. (1985) High efficiency gene transfer into mammalian cells. In

572

Glover,D.M. (ed.) DNA Cloning: A Practical Approach. IRL Press, Washington, DC, Vol. 2, pp. 143–190. 25. Joh,T., Darland,T., Samuels,M., Wu,J.X. and Adamson,E.D. (1992) Regulation of epidermal growth factor receptor gene expression in murine embryonal carcinoma cells. Cell Growth Differ., 3, 315–325. 26. Bradford,M. (1976) A rapid and sensitive method for the determination of microgram quantities of protein utilizing the principles of dye binding. Anal. Biochem., 72, 248–254. 27. Filmus,J., Pollak,M.N., Cairncross,J.G. and Buick,R.N. (1985) Amplified, overexpressed and rearranged epidermal growth factor receptor gene in a human astrocytoma cell line. Biochem. Biophys. Res. Commun., 131, 207–215. 28. Wang,D., Li,W., Jiang,W., Humphrey,L.E., Howell,G.M. and Brattain,M.G. (1998) Autocrine TGFα expression in the regulation of initiation of human colon carcinoma growth. J. Cell. Physiol., 177, 387–395. 29. Auvinen,M., Laine,A., Paasinen-Sohns,A., Kangas,A., Kangas,L., Saksela,O., Andersson,L.C. and Ho¨ ltta¨ ,E. (1997) Human ornithine decarboxylase-overproducing NIH3T3 cells induce rapidly growing, highly vascularized tumors in nude mice. Cancer Res., 57, 3016–3025. 30. Hamilton,M. and Wolfman,A. (1998) Oncogenic Ha-ras-dependent mitogen-activated protein kinase activity requires signaling through the epidermal growth factor receptor. J. Biol. Chem., 273, 28155–28162. 31. Davis,R.H., Morris,D.R. and Coffino,P. (1992) Sequestered end products and enzyme regulation: the case of ornithine decarboxylase. Microbiol. Rev., 56, 280–290. 32. Faaland,C.A., Laskin,J.D. and Thomas,T.J. (1995) Inhibition of epidermal growth factor-stimulated EGF receptor tyrosine kinase activity in A431 human epidermoid carcinoma cells by polyamines. Cell Growth Differ., 6, 115–121. Received June 21, 1999; revised November 2, 1999; accepted November 19, 1999