INTRODUCTION

Several lines of evidence suggest that metabolism of the cis-polyunsaturated fatty acids, arachidonic acid and linoleic acid, by prostaglandin H synthase and lipoxygenases generate metabolites that modulate the EGF¹ mitogenic signal in fibroblasts. In Balb/c 3T3 fibroblasts, mitogenic stimulation by EGF induced the formation of prostaglandin E₂ (1) and the expression of c-myc (2). Inhibition of prostaglandin H synthase partially blocked both mitogenesis and c-myc expression, which was restored and enhanced by the addition of exogenous prostaglandins. In these studies lipoxygenase inhibitors were very effective inhibitors of mitogenesis and also inhibited the metabolism of linoleic acid by these cells (1, 3). We did not detect lipoxygenase metabolites of arachidonic acid, but did observe that linoleic acid was readily metabolized to (13S)-hydroxyoctadecadienoic acid (HODE). Linoleic acid metabolites were potent stimulators of EGF-dependent mitogenesis in these cells (3).

A more extensive investigation was conducted with Syrian hamster embryo fibroblasts (SHE) cells (4). These cells metabolized arachidonic acid to low amounts of prostaglandin E₂ with little or no lipoxygenase metabolites observed. The addition of prostaglandins inhibited EGF-dependent mitogenesis and the expression of the proto-oncogenes, c-myc (5), c-jun, and Jun-B (6). Indomethacin did not inhibit EGF-dependent mitogenesis, while lipoxygenase inhibitors were particularly effective for inhibiting mitogenesis. EGF is a potent mitogen in these cells but does not enhance arachidonic acid metabolism nor the expression of prostaglandin H synthase-2 (7). EGF stimulates the metabolism of exogenous as well as endogenous linoleic acid to (13S)-HPODE/(13S)-HODE catalyzed by an apparent 15-lipoxygenase. The activity of this lipoxygenase is regulated by the EGF receptor tyrosine kinase (7), thus linking the activation of lipoxygenase metabolism of linoleic acid to the tyrosine kinase activity of the epidermal growth factor receptor (EGFR). Other growth factors, including fibroblast growth factor (8) and insulin and platelet-derived growth factor (2), do not activate linoleic acid metabolism, suggesting that the stimulation of linoleic acid metabolism is unique to the EGF signaling pathway.

The SHE cells are used as a model for neoplastic progression since two phenotypes are available, which, when fused with tumor cells, either suppress (supB⁺) or do not suppress (supB) tumorigenic phenotypes in cell-cell hybrids (9-11). The addition of (13S)-HODE or its hydroperoxide precursor, (13S)-HPODE, greatly potentiated the mitogenic response of supB⁺ SHE cells to EGF (4). At concentrations of 10⁹ to 10⁶ M, (13S)-HPODE and (13S)-HODE increased EGF dependent mitogenesis 3-4-fold in the supB⁺ cells with (13S)-HPODE being 10 times more potent than (13S)-HODE. The linoleic acid metabolites did not alter the EGF mitogenic response in the supB variant. In the absence of EGF, these lipid metabolites were not mitogenic. Further structural characterization of the lipid-dependent enhancement of mitogenesis indicated (13S)-HPODE and (13S)-HODE have exclusive activity (11). For example, the (13R)-HODE enantiomer and the analogous 15-lipoxygenase metabolites of arachidonic acid, (15S)-HPETE/15-HETE, were not active (11). This finding suggests a specific interaction of (13S)-HPODE with either a receptor or with a intracellular signal transduction protein of the EGF signaling pathway.

These results indicate that the formation of (13S)-HPODE/(13S)-HODE is a component of the EGF signaling pathway and this lipid metabolite potentiates EGF-dependent mitogenesis in the supB⁺ cell line. Polypeptide growth factors such as EGF regulate cellular growth and metabolism by binding to specific cell surface receptors that have tyrosine kinase activity (12, 13). Occupation of the EGFR by EGF and other ligands causes dimerization of the receptor and stimulates the receptor's intrinsic tyrosine kinase activity (14). The intrinsic tyrosine kinase activity of the EGFR is the critical biochemical signal involved in cellular mitogenic responses to EGF (17-19). The kinase activity autophosphorylates the EGFR, phosphorylates key tyrosine residues present in several signaling proteins, and initiates a series of tyrosine, serine and threonine protein phosphorylations which link the cell surface with the nucleus (15). One of the final steps in the EGF cascade appears to be the activation of the mitogen-activated protein (MAP) kinase family, which in turn regulates other biochemical events including activation of transcription factors (16). One possible explanation for the enhancement of EGF-dependent mitogenesis by (13S)-HPODE/(13S)-HODE is that this lipid metabolite alters phosphorylation events in the EGFR pathway. Furthermore, our findings suggest that important differences may exist in the EGF signaling pathway between these two SHE cell lines and suggest that during the neoplastic progression a mutation or deletion occurred that alters the signaling transduction process. In this report we have tested these hypotheses by examining the EGF signaling pathway in the two SHE cell phenotypes. In the supB⁺ cell line, (13S)-HPODE specifically enhanced the phosphorylation of EGFR and other signaling proteins of the pathway leading ultimately to enhanced MAP kinase activity. The (13S)-HPODE exerts this response by inhibiting the dephosphorylation of the EGF receptor. In contrast, the linoleate metabolite did not modulate receptor dephosphorylation in supB cells and thus did not alter EGF-dependent tyrosine phosphorylation in the supB cell line.

EXPERIMENTAL PROCEDURES

Bovine serum albumin (BSA), IBR modified Dulbecco's modified Eagle's medium (IBR), phosphate-buffered saline (PBS), trypsin, and gentamicin were from Life Technologies, Inc. Fetal calf serum was purchased from HyClone Laboratories (Logan, UT). 75-cm² flasks were obtained from Costar (Cambridge, MA). 100-mm dishes were from Falcon (Plymouth, United Kingdom). (13S)-HPODE, (13S)-HODE, (15S)-HETE, (15S)-HPETE, and (12S)-HETE were from Cayman Chemical Co. (Ann Arbor, MI). (13R)-HODE was obtained from Oxford (Oxford, MI). EGF was obtained from Collaborative Research Associates (Bedford, MA). Tris, glycerol, ecolume and glycine were purchased from ICN (Aurora, IL). Sodium chloride and methanol were from J.T. Baker. The magnesium chloride and trichloroacetic acid were from Mallinckrodt. The acrylamide and bisacrylamide were purchased from Amresco (Solon, OH). The nitrocellulose was obtained from Schleicher & Schuell. The anti-GAP antibody was from Transduction Laboratories (Lexington, KY). The Ig conjugates, Hyperfilm, ECL reagents, [-³³P] adenosine triphosphate (2000 Ci/mmol) and ¹²⁵I-EGF were purchased from Amersham. The anti-MAP antibodies (ERK-1, ERK-2) were from Santa Cruz Biotechnology (Santa Cruz, CA), while anti-GAP antibodies were purchased from Transduction Laboratories (Lexington, KY). The P81 chromatography paper was from Whatman. The ethanol was from Pharmaco (Brookfield, CT). The protein A-Sepharose beads were from Pharmacia Biotech (Uppsala, Sweden). The methyl 2,5-dihydroxycinnamate was from Biomol (Plymouth Meeting, PA). The anti-phosphotyrosine antibody and all other reagents were obtained from Sigma, while a EGFR antibody raised in rabbits against the carboxyl region of rat EGFR was a gift from Dr. G. Clark, NIEHS.

These experiments were performed with Syrian hamster embryo fibroblast cell lines, 10WsupB⁺ 8 (supB⁺) and 10WsupB 1 (supB), as described previously (4). Cells were maintained at 37 °C in a humidified 5% CO₂, 95% air atmosphere. The cells were cultured in IBR containing 10% fetal calf serum and gentamicin (10 µg/ml). Trypsin (0.05%) was used to subculture the cells. Both variants were grown under identical conditions and passaged at 70-80% confluence. Passage 10-17 was used in these experiments. Passage number and degree of confluence (70-80%) were identical in experiments comparing supB⁺ to supB  cells.

After reaching 70-80% confluence on 100-mm dishes, cells were serum-deprived for 20 h to synchronize cell cycles in G_o. All experiments were done in serum-free medium at 37 °C unless otherwise stated. Cells were treated with (13S)-HPODE or other lipids at 10 nM to 1 µM for 1 h prior to the addition of EGF. After stimulation by EGF, cells were harvested at various time points. For the room temperature experiments, the cells were allowed to equilibrate for 30 min before the addition of the linoleic acid metabolite.

Treated cells were washed twice with ice-cold PBS and lysed in boiling sample buffer: 5 mM sodium phosphate, pH 6.8, 2% sodium dodecyl sulfate, 0.1 M dithiothreitol, 10% glycerol, 5% -mercaptoethanol, 0.2% bromphenol blue, 1 mM sodium orthovanadate, 1 mM sodium fluoride. The lysates from an equal number of cells were then sheared five times through a 25-gauge needle, boiled for 5 min, and immediately loaded onto a gel or the protein estimated by the Bradford reagent and the same amount of protein loaded onto the gel. Samples were run on SDS-PAGE (6%, 8%, or 10%) and electrophoretically transferred to a nitrocellulose membrane in 25 mM Tris, 192 mM glycine, 20% methanol with 200 µM sodium orthovanadate using Hoefer electrophoresis equipment. Membranes were blocked in Tris-buffered saline containing 0.2% Tween 20 (TBST) with 5% BSA. The blots were then incubated for 1 h with an anti-phosphotyrosine antibody (1:2000), anti-EGFR antibody (1:2000), or an anti-GAP antibody (1:250) with 1% BSA at room temperature. The blots were washed five times in TBST with 0.1% BSA, then incubated with peroxidase-conjugated anti-mouse Ig (1:10000) or anti-rabbit Ig (1:10000 or 1:2000) in 1% BSA. The blots were again washed five times in TBST with 0.1% BSA and visualized using the Amersham ECL system.

When cells reached 70-80% confluence, they were starved in Dulbecco's modified Eagle's medium without antibiotics containing 0.05% fetal bovine serum overnight. Three dishes of each experimental group were set aside to count the cell number. The medium was aspirated, and the cells were rinsed with 10 ml of PBS containing 0.1% BSA and 2 mM sodium vanadate and lysed with 0.8 ml of ice-cold lysis buffer: 1% Triton X-100, 50 mM Tris, pH 8.5, 150 mM NaCl, 5 mM EDTA, 5 mM PMSF, 50 mM NaF, 2.0 mM sodium vanadate, 20 µg/ml leupeptin, 20 µg/ml aprotinin. After scraping of the cells, the lysate was collected, 50 µl of 5 M NaCl added, sonicated on ice for 7 s twice, and centrifuged in a microcentrifuge for 10 min at 4 °C. The supernatant was transferred to a new tube and an aliquot removed to measure the protein concentration by BCA protein assay reagent. 2.0 mg of cellular protein from 3 × 10⁶ cells was precleared with 5 µl of normal rabbit serum and 100 µl of protein A-Sepharose 4B beads. The mixture was tumbled at 4 °C for 30 min and centrifuged at 10,000 × g for 10 min, and the supernatant removed. To the supernatant was added 7 µl of EGFR antibody, tumbled overnight at 4 °C, and 100 µl of protein A-Sepharose 4B beads added and tumbled for an additional 2 h. After centrifugation at 10,000 × g for 5 min, the pellet was washed sequentially with 1.0 ml of IP₂, IP₃, and IP₄ buffers. (IP₂: 0.5% Triton X-100, 50 mM Tris, pH 8.5, 500 mM NaCl, 1 mM EDTA, 1 mM sodium vanadate, 20 µg/ml leupeptin, 20 µg/ml aprotinin, 40 µg/ml PMSF; IP₃: 0.5% Triton X-100, 50 mM Tris pH 8.5, 150 mM NaCl, 1 mM EDTA, 1 mM sodium vanadate, 20 µg/ml leupeptin, 20 µg/ml aprotinin, 40 µg/ml PMSF; IP₄: 0.1% Triton X-100, 10 mM Tris, pH 8.5). To the final pellet was added equal amounts of 10 mM Tris, pH 8.5, and 2 × protein sample buffer, boiled for 8 min and centrifuged at 10,000 × g for 5 min. The supernatant was frozen at 70 °C for later use on SDS-PAGE (8%).

The EGFR in cell membrane preparations was used to estimate the EGFR kinase activity using a modification of the method described by Griswold-Prenner et al. (20). Briefly, membranes prepared from SHE cells were incubated in 0.25 mM Hepes buffer, pH 7.4, 50 nM EGF, 11 mM MgCl₂, 1 mg/ml angiotensin II, 20 µM ATP, and 10 µCi of [-³³P]ATP. The incubation mixture was incubated 37 °C for the times indicated. The reaction was stopped by the addition of 5% trichloroacetic acid, vortexed, and maintained on ice. The samples were centrifuged for 5 min at 15,000 × g and the supernatant spotted on P-81 phosphocellulose chromatography paper. The paper was washed with 0.2% phosphoric acid six times to remove unincorporated ATP and the filters counted. The incorporation of ³³P into angiotensin II was used as a measure of EGFR kinase activity.

Treated cells were washed twice with ice-cold PBS and lysed in ice-cold lysis buffer: 50 mM HEPES, pH 7.4, 50 mM sodium pyrophosphate, 50 mM sodium fluoride, 50 mM sodium chloride, 5 mM EDTA, 5 mM EGTA, 2 mM sodium orthovanadate, 0.5 mM PMSF, 10 µg/ml leupeptin, and 0.01% Triton X-100. An aliquot (20 µl) of each MAPK antibody and 40 µl of 50% protein A-Sepharose beads were added to each of the cell lysates to immunoprecipitate both p42 and p44 forms. The incubation mixtures from an equal number of cells were allowed to tumble overnight at 4 °C. Immunoprecipitates were washed three times in lysis buffer and then resuspended in 50 µl of the reaction buffer: 50 mM -glycerophosphate, pH 7.3, 1.5 mM EGTA, 0.1 mM sodium orthovanadate, 1 mM dithiothreitol, 10 µM calmidazolium, 10 mM MgCl₂, 10 µg/ml leupeptin, 10 µg/ml aprotinin, 2 µg/ml pepstatin A, 1 mM benzamidine, 0.3 mM adenosine triphosphate, and 0.5 mg/ml myelin basic protein. 10 µCi of [-³³P]ATP was added to the reaction. The reaction was allowed to proceed for 30 min at 37 °C and was quenched by adding 50 µl of ice-cold trichloroacetic acid. The mixture was then spotted on P-81 phosphocellulose chromatography paper and allowed to dry. The filters were washed five times in 0.5% phosphoric acid, followed by a final wash in ethanol (21). The filters were allowed to dry and the radioactivity counted on a Packard 2000CA scintillation counter.

EGFR down-regulation was measured using a modification of the procedure described by Carpenter (22). 80% confluent 60-mm dishes were rinsed twice with IBR and treated with or without 1 µM (13S)-HPODE for 1 h at 37 °C. Cells were then treated with cold EGF (10 ng/ml) plus or minus 1 µM (13S)-HPODE for 0, 15, 30, 60, and 120 min at 37 °C. At the end of incubation period, the plates were washed twice with PBS, and incubated 2 min in 0.2 M acetic acid, 0.5 M NaCl (ice-cold) followed by a second incubation for 20 s to strip cold EGF. Cells were then rinsed twice with PBS and incubated with ¹²⁵I-EGF (1 nM) in IBR plus 0.1% BSA on ice for 2 h. Cells were then washed three times with PBS and scraped, and the bound ¹²⁵I-EGF was quantified by -counting.

70-80% confluent cells in 100-mm dishes were starved for 18 h at 37 °C in IBR containing 0.1% fetal bovine serum with no antibiotics. The media were replaced with Hepes-buffered IBR (0.027% BSA) either containing (13S)-HPODE (the final concentration of ethanol was less than 0.02%) or vehicle. The cells were incubated for 1 h. EGF (10 ng/ml) was added and the cells incubated for 1 min at room temperature. The EGF receptor kinase activity was blocked by addition of 50 µM methyl 2,5-dihydroxycinnamate, a protein-tyrosine kinase inhibitor, in Me₂SO and the cells harvested at the times indicated. To harvest, the cells were moved onto ice, the medium removed and replaced with ice-cold PBS containing BSA (0.1%) and sodium orthovanadate (2 mM). The medium was removed and the cells lysed immediately with boiling sample buffer and boiled for 5 min. The lysate was dispersed with a 25-gauge needle and centrifuged, and the supernatant frozen at 70 °C for Western analysis with the anti-phosphotyrosine monoclonal antibody. Gels were loaded with lysate from 4 × 10⁵ cells.

RESULTS

The addition of EGF to serum-starved SHE cells stimulated a rapid autophosphorylation of the EGFR (170 kDa) as measured by Western analysis using an antibody for phosphorylated tyrosine residues (anti-Tyr(P)). The tyrosine phosphorylation peaked between 0.5 and 1.5 min after the addition of EGF and then rapidly declined (data not shown). An EGF concentration dependence was then determined in the supB⁺ and supB variants measured at the peak of EGFR autophosphorylation. The dose-response relationship between EGF concentration and EGFR autophosphorylation was very similar for both variants (data not shown). These findings are in agreement with previous data that report the number of EGFR, and their affinity for EGF are approximately the same for both SHE cell variants (4). We observed a number of other proteins, which were also tyrosine-phosphorylated in response to EGF. One of the most intense bands migrated on SDS-PAGE at 120 kDa. The tyrosine phosphorylation of the 120-kDa band also showed a concentration-dependent effect of EGF similar to that observed for EGFR autophosphorylation in both variants.

We attempted to estimate the levels of the EGFR in lysates prepared from both the supB⁺ and supB cells by Western analysis, but the low level of the receptor and the poor cross-reactivity with anti-EGFR antibodies prevent adequate estimation. Lysates were then prepared from the same number of cells, the protein estimated, and EGFR immunoprecipitated. The immunocomplexes were then analyzed by Western analysis using anti-EGFR antibody for blotting. As illustrated in Fig. 1A, the level of the EGFR based on protein or cell number was the same in lysates prepared from both SHE cell variants.

Previously, we reported that the addition of the linoleic acid metabolite, (13S)-HPODE/(13S)-HODE, enhanced EGF-dependent mitogenesis in the supB⁺ SHE cells but not in the supB variant (9). Other structurally related lipid metabolites were ineffective (11). (13S)-HPODE was approximately 10 times more potent than (13S)-HODE in enhancing EGF-dependent mitogenesis in supB⁺ cells. Both the hydroperoxide and the alcohol were used in experiments to examine possible modulation of the EGF signaling pathway and the concentrations adjusted to account for the difference in potency.

Since EGF binding to its receptor initiates tyrosine phosphorylation of signaling proteins in the pathway, we measured EGF-dependent tyrosine phosphorylation by Western analysis using an antibody that detects phosphorylated tyrosine residues. The addition of EGF to either the supB⁺ or supB cells increased tyrosine phosphorylation of the EGFR and a band at 120 kDa (Fig. 1B) tentatively characterized as GAP. GAP identification was based on the apparent molecular weight and immunoreactivity with specific GAP antibody. After immunoprecipitation of cell lysates with the anti-GAP antibodies, the 120-kDa tyrosine-phosphorylated band was not detected in subsequent analysis of the supernatant (data not shown). The addition of (13S)-HODE significantly increased EGF-dependent tyrosine phosphorylation in the supB⁺ cells. In the supB variant, the addition of (13S)-HODE caused significantly less increase in tyrosine phosphorylation than observed for the supB⁺ cells with EGF and (13S)-HODE treatment. This finding suggests possible up-regulation of the EGF signaling pathway in supB⁺ cells by (13S)-HODE.

Initiation of the EGF signal transduction pathway leads to the downstream activation of MAP kinases, which are likely to mediate many of the nuclear activation events associated with EGF-dependent mitogenesis (16). MAP kinase is activated by threonine and tyrosine phosphorylation catalyzed by the dual functional MAP kinase kinase (MAPKK or MEK) (23-25). If (13S)-HODE/(13S)-HPODE up-regulates the EGF signaling pathway, then (13S)-HPODE when added to the cells should eventually result in an increase in MAP kinase tyrosine phosphorylation and activity in the supB⁺ and not in the supB variant. We thus measured tyrosine phosphorylation of p42 MAPK and p44 MAPK stimulated by EGF in the presence and absence of (13S)-HPODE, 10 nM to 1 µM. The results for treatment with 1 µM (13S)-HPODE are shown in Fig. 2. In the absence of EGF or on addition of (13S)-HPODE alone, little or no tyrosine phosphorylation of either p42 or p44 MAPK was detected in the quiescent SHE cells. Treatment with 10% fetal calf serum or EGF (10 ng/ml) stimulated tyrosine phosphorylation of both the p42 and p44 proteins. A biphasic response to EGF was observed, with initial enhancement in the tyrosine phosphorylation observed at early times (2-30 min) followed by a down-regulation of the signal. A second phase of increased tyrosine phosphorylation of MAP kinase occurs at 90-100 min. The addition of (13S)-HPODE enhanced the magnitude of MAP kinase tyrosine phosphorylation at all time points examined in the supB⁺ cell line (Fig. 2). Furthermore, tyrosine phosphorylation of MAP kinase was prolonged in the presence of (13S)-HPODE. The addition of (13S)-HPODE did not alter the tyrosine phosphorylation of either p44 or p42 in the supB variant (data not shown). Thus, (13S)-HPODE appears to up-regulate the entire EGF signaling pathway exclusively in supB⁺ cells.

To determine if the observed changes in the phosphotyrosine state of MAP kinase correlates with changes in MAP kinase activity, the activity was measured in immunoprecipitates from supB⁺ SHE cells using anti-MAP kinase antibodies, which react with both the p42 and p44 proteins. Quiescent cells were treated with EGF for 2 or 90 min ± (13S)-HPODE (pretreatment for 1 h prior to EGF addition). The MAP kinase activity of the immunoprecipitates was determined by measuring the incorporation of [-³³P]ATP into myelin basic protein as described under "Experimental Procedures." As shown in Table I, the addition of EGF stimulated the MAP kinase activity as expected. At the 2-min time point, the combination of EGF and (13S)-HPODE increased MAP kinase activity in the immunocomplex 1.5-fold compared with the corresponding cells incubated with EGF alone. After 90 min of treatment, a greater than 2-fold increase in MAP kinase activity was observed in the immunocomplex obtained from cells treated with the lipid metabolite and EGF as compared with cells treated with EGF alone. Thus, the addition of the linoleic acid metabolite to supB⁺ cells increased and prolonged EGF-dependent MAP kinase tyrosine phosphorylation and MAP kinase activity.

Table I. Effect of 13(S)-HPODE on EGF-stimulated MAP kinase activity SHE cells were growth arrested by incubation in serum-free media for 20 h. Cells were then treated with EGF(10 ng/ml) for either 2 min or 90 min. To test the effect of (13S)-HPODE, samples were preincubated with 1 µM (13S)-HPODE for 1 h prior to the addition of EGF. As a positive control, cells were activated with 10% fetal calf serum. Following treatment, cell lysates prepared from an equal number of cells were immunoprecipitated with MAP kinase (ERK-1 and ERK-2) antiserum and the MAP kinase immunocomplexes were incubated with 0.3 mM [-33P]ATP (2000 Ci/mmol) and myelin basic protein (0.5 mg/ml) as substrate for 30 min at 37 °C. Results are mean ± S.D. of triplicate samples. Treatment MAP kinase activity pmol/min/mg protein Serum-free medium 45  ± 5 10% fetal calf serum (5 min) 190  ± 15 EGF (2 min) 105  ± 10 EGF + (13S)-HPODE (2 min) 145  ± 10 EGF (90 min) 40  ± 5 EGF + (13S)-HPODE (90 min) 90  ± 5

One possible explanation for the up-regulation of the EGF signaling pathway in the supB⁺ cells by (13S)-HPODE is the selective increase in the EGFR kinase activity in the supB⁺ cell line. The EGFR kinase activity in SHE cell lysates was estimated by measuring the incorporation of [-³³P]ATP into angiotensin II, which is a good substrate for the EGFR kinase (20). As shown in Fig. 3, little or no difference was observed in the EGFR kinase activity of the two SHE cell lines. Furthermore, incubation in vitro with (13S)-HPODE did not directly alter EGFR kinase activity. Thus (13S)-HPODE does not up-regulate the EGF pathway in the supB⁺ cells by direct activation of the receptor kinase activity.

The linoleic acid metabolite could also act by altering the rate of receptor internalization. To test this hypothesis cells were incubated with EGF in the presence and absence of (13S)-HPODE for different times at 37 °C and EGFR internalization measured by a method published previously (22). The rate and extent of receptor down-regulation was very similar in the supB⁺ and supB cells (data not shown). Pretreatment of the supB⁺ cells with either vehicle (Fig. 4A) or (13S)-HPODE (Fig. 4B) for 1 h prior to the incubation with EGF also did not change the rate and extent of receptor down-regulation. Similar results were obtained with supBcells.

The phosphorylation and dephosphorylation of the EGFR was examined in both cell lines in the presence and absence of the lipid metabolite (Fig. 5). The EGFR was rapidly tyrosine-phosphorylated in response to EGF with maximum phosphorylation observed at 1-2 min after the addition of EGF and then rapidly declined. The duration of EGF-dependent tyrosine phosphorylation appears to be greater in the supB than in the supB⁺ variant.

The addition of (13S)-HPODE 1 h prior to the addition of EGF to quiescent SHE cells augmented EGFR tyrosine phosphorylation in supB⁺ but not in supB cells. In the supB⁺ cells, (13S)-HPODE clearly stimulated an increase in autophosphorylation of EGFR at the 1-10-min time points (Fig. 5). In contrast, (13S)-HPODE did not significantly alter the phosphotyrosine state of the EGFR in supB cells (Fig. 5). Shown in Fig. 5 are the results of a typical experiment in which the cells were treated with 1 µM lipid, which produced maximum enhancement of tyrosine phosphorylation. A concentration-dependent effect was observed from 10 nM to 1 µM (data not shown). This effect of (13S)-HPODE appears to be on both the magnitude and duration of the tyrosine phosphorylation signal in the supB⁺ cells. Treatment of cells with (13S)-HPODE alone did not alter the tyrosine phosphorylation pattern in either SHE supB cell (data not shown). Other corresponding arachidonic acid metabolites (15-HETE and 12-HETE) did not increase tyrosine phosphorylation in EGF stimulated supB⁺ cells. (13R)-HODE, which was inactive for enhancement of EGF-dependent mitogenesis, also did not increase EGF-dependent tyrosine phosphorylation (data not shown). In addition to modulating EGFR tyrosine phosphorylation, the addition of (13S)-HODE or (13S)-HPODE to the supB⁺ cells alters the magnitude and the duration of tyrosine phosphorylation of the protein band at 120 kDa (immunoreactive with anti-GAP antibodies). This observation fits the hypothesis that the linoleic acid metabolite up-regulates the EGF pathway in supB⁺ cells starting with the EGFR and transmitted downstream to GAP, MAP kinase, and other signal transduction proteins.

The biological effect of the linoleic acid metabolite on EGFR autophosphorylation can be demonstrated even more dramatically in experiments conducted at room temperature rather than 37 °C. Shifting to lower temperature slows down the rate of EGFR dephosphorylation and ligand-induced receptor internalization and degradation (14). The room temperature experiments allow one to study the tyrosine phosphorylation signal at much slower kinetics. As shown in Fig. 6 (panel A versus panel C), (13S)-HPODE enhanced EGF-stimulated tyrosine phosphorylation of both EGFR and GAP as compared with cells treated with EGF alone. The action of (13S)-HPODE increased both the magnitude and duration of protein-tyrosine phosphorylation in this analysis. Western immunoblots using antibodies specific for GAP indicated that the level of this protein did not change during the course of the experiment as shown in panel B of Fig. 6. Thus, the (13S)-HPODE-stimulated changes noted with the anti-phosphotyrosine immunoblots appear to be due to an effect on the tyrosine phosphorylation state of GAP and not due to alterations in protein levels. These results suggest that (13S)-HPODE attenuates the dephosphorylation of the EGFR, which subsequently results in enhancement of EGF-dependent tyrosine phosphorylation of key signal transduction proteins in the supB⁺ but not in the supB variant.

The rate of EGFR dephosphorylation was estimated using a modification of the assay described by Bohmer et al. (26). EGFR dephosphorylation occurred very rapidly (Fig. 7) in the supB⁺ cell line. Dephosphorylation of the EGFR appeared to reach completion after 2 min. In contrast, the dephosphorylation of the EGFR was significantly slower in the supB variant. Dephosphorylation was nearing completion at 6-8 min after EGF binding. The difference in the rate of EGFR dephosphorylation observed in this experiment is in agreement with the other data on EGF-dependent tyrosine phosphorylation. The effect of (13S)-HPODE on dephosphorylation of the EGFR of the supB⁺ and supB cells was examined following preincubation of the cells for 1 h with the lipid metabolite prior to the addition of EGF. As shown in Fig. 8, (13S)-HPODE significantly altered the rate of receptor dephosphorylation in the supB⁺ cells. No significant response was observed in the supB cell (data not shown). In the presence of (13S)-HPODE, receptor dephosphorylation for the supB⁺ cells was not complete until 7 min after EGF addition, while in the absence of (13S)-HPODE, dephosphorylation was complete by 3 min. The effect of (13S)-HPODE on EGFR dephosphorylation in the supB⁺ and supB cells at several concentrations of (13S)-HPODE was examined by measuring the EGFR tyrosine phosphorylation 2 min after the addition of EGF. With increasing concentration of (13S)-HPODE, a greater intensity of tyrosine phosphorylation of the receptor was observed (Fig. 9) in the supB⁺ cells, indicating inhibition of receptor dephosphorylation was dependent on the concentration of (13S)-HPODE. Little or no effect of (13S)-HPODE on EGFR dephosphorylation was observed in the supB cells. Thus, (13S)-HPODE appears to up-regulate the EGF pathway in supB⁺ cells by inhibiting dephosphorylation of the EGFR, resulting in an increased activation of the EGFR tyrosine kinase and downstream kinase signaling.

8

5

DISCUSSION

The binding of EGF to its receptor activates the receptor tyrosine kinase activity, which in turn tyrosine phosphorylates other key signaling proteins and begins a series of phosphorylation events that culminates in activation of transcription factors leading to cell division. For Syrian hamster embryo fibroblasts, the addition of EGF to serum-starved quiescent cells initiates cell proliferation (4) and stimulates the metabolism of linoleic acid to (13S)-HPODE. In recent years considerable progress was made in understanding this transduction pathway, but the nature and the importance of linoleic acid metabolism in the signaling pathway is not recognized nor fully understood. The formation of (13S)-HPODE and (13S)-HODE by an apparent 15-lipoxygenase appears to be regulated by the EGFR tyrosine kinase activity (7), suggesting that formation of (13S)-HPODE/(13S)-HODE is a component of the EGF signaling pathway in these cells. The addition of lipoxygenase inhibitors that block the 15-lipoxygenase without effects on the EGFR kinase activity attenuates the EGF response as measured by a mitogenesis assay (4). Furthermore, exogenous (13S)-HPODE and (13S)-HODE enhanced the EGF-dependent mitogenesis 4-5-fold in the SHE supB⁺ phenotype, but did not alter EGF-dependent mitogenesis in the supB variant (4). The linoleic acid metabolites were not mitogens alone, but appeared to specifically enhance the EGF mitogenic signal. Other metabolites of arachidonic acid and linoleic acid formed by either prostaglandin H synthase or lipoxygenases were either inactive or inhibited mitogenesis, suggesting that (13S)-HPODE/(13S)-HODE have exclusive ability to up-regulate the EGF signaling pathway in these cells (11).

The EGF signaling pathway in both SHE cell lines was examined in an effort to determine mechanisms by which the linoleic acid metabolites enhance EGF-dependent proliferation and to explore potential differences in the signaling pathway in the two cell types. No differences were observed in either the number of EGF receptors, the kinase activity of the receptor, or the internalization and degradation of the receptor in the two SHE cell lines. (13S)-HPODE in the presence of EGF did not directly alter either the kinase activity or internalization of EGFR but a more intense tyrosine phosphorylation of the EGFR and GAP was observed in the supB⁺ cells, indicating that (13S)-HPODE up-regulated the EGF signaling pathway at an early event in the cascade. This up-regulation in supB⁺ cells resulted in an increase in MAP kinase tyrosine phosphorylation and MAP kinase activity. In the presence of (13S)-HPODE, the supB⁺ cells appear to resemble the supB variant response to EGF. The up-regulation of the EGF signaling pathway is consistent with previously reported modulation of EGF-dependent mitogenesis in SHE fibroblasts by the linoleic acid metabolite (4). As observed for EGF-dependent mitogenesis in supB⁺ cells, the ability to modulate EGF-dependent tyrosine phosphorylation is exclusive for (13S)-HODE and (13S)-HPODE. Other arachidonic acid and linoleic acid metabolites were inactive. This suggests a specific but uncharacterized biological target for the metabolite.

An important difference in the EGF signaling pathway of supB⁺ and supB cells is the observation of a longer duration of EGF-dependent tyrosine phosphorylation in the supB variant. The extended duration of tyrosine phosphorylation was noted for both EGFR and GAP. EGFR dephosphorylation, measured directly in intact cells, was slower in the supB variant compared with the supB⁺ phenotype, an observation that provides a rational explanation for the more prolonged tyrosine phosphorylation of signaling proteins of the EGFR cascade. Furthermore, incubation of the supB⁺ cells with (13S)-HPODE resulted in a concentration-dependent attenuation of EGFR dephosphorylation, while no effect on EGFR dephosphorylation was observed in the supB cells. Thus, it appears that an important difference between supB⁺ and supB cells is the rate of EGFR dephosphorylation, which also appears to be the site for (13S)-HPODE modulation of the EGF signaling pathway. The mechanism for the attenuation of EGFR dephosphorylation by (13S)-HPODE is not known and is made more complex by a poor general understanding of EGFR dephosphorylation. Negative regulation of growth factor receptor by receptor-directed protein-tyrosine phosphatases has been suggested. Differences in functional activity of tyrosine phosphatases specific for the EGF pathway could provide an explanation for the phenotypic difference between the two SHE cell lines and may be a target for the enhancement of EGFR signaling by (13S)-HPODE. Our current hypothesis is that (13S)-HPODE modulates the interaction of a tyrosine phosphatase with the EGFR. On-going studies are designed to test this hypothesis.

The lack of a response to (13S)-HPODE in the supB variant and the difference in the EGF-dependent tyrosine phosphorylation of the two SHE cell lines suggest that during neoplastic progression of the SHE cells, gene deletions or mutations of EGF signaling proteins have occurred with the loss of the tumor suppressor phenotype (14-16). One of these proteins appears to be the site or the "receptor" responsible for the modulation of the EGF signaling pathway by (13S)-HODE/(13S)-HPODE. Further studies of the EGF signaling pathway in the two SHE cell lines should shed light on the molecular events associated with the loss of the tumor suppressor phenotype and should provide the necessary information to understand how the linoleic acid metabolites modulate the EGF signaling pathway.

Ligand-induced activation of the EGFR tyrosine kinase appears to be the essential biochemical event for further EGF mitogenic signal transduction (17-19). In the work reported here, we demonstrate that the linoleic acid metabolite (13S)-HPODE modulates the tyrosine phosphorylation of the EGFR signaling pathway. Downstream tyrosine phosphorylation of MAP kinase is enhanced. Phosphorylation and subsequent activation of MAP kinase allows for transduction of the mitogenic signal from the cell membrane to the nucleus (5). The cytosolic form of phospholipase A₂ (cPLA₂) is a substrate for MAP kinase, with phosphorylation of cPLA₂ resulting in increased enzymatic activity (27, 28). Thus, (13S)-HPODE enhancement of MAP kinase activity could potentially result in phospholipase activation generating additional lipid mediators involved in a positive feedback loop for mitogenic signaling. In earlier work we found that another cis-polyunsaturated fatty acid, arachidonic acid, can augment MAP kinase activity (21). In these previous studies with vascular smooth muscle cells, arachidonic acid at micromolar levels was found to directly increase MAP kinase tyrosine phosphorylation and activity. This effect was mediated in part by the 15-lipoxygenase metabolite, 15-hydroxy-eicosatetraenoic acid (15-HETE). In these cells the arachidonate compounds stimulated MAP kinase in the absence of growth factors. The implication from this work is a role for arachidonic acid in smooth muscle cell proliferation in both normal and pathological conditions such as atherosclerosis. Our current finding of modulation of EGF-dependent MAP kinase activity by a linoleic acid metabolite in SHE fibroblasts further demonstrates the importance of specific lipid compounds in regulating mitogenesis in multiple cell types and in various physiological processes. (13S)-HPODE appears to be a highly specific modulator of the tyrosine phosphorylation state of the EGF pathway from EGFR autophosphorylation to MAP kinase activity by altering the balance between kinase and phosphatase activities (29-31) associated with this signaling pathway.

The importance of arachidonic acid and linoleic acid metabolites in regulating cell proliferation is supported by both animal and human epidemiology studies. For example, nonsteroidal anti-inflammatory drugs, such as aspirin and sulindac, reduce the incidence and mortality of colon cancer and will induce regression of rectal polyps in patients with familial adenomatous polyposis (32-34). Aspirin and other nonsteroidal anti-inflammatory drugs inhibit the activity of prostaglandin H synthase, which can convert arachidonic acid to prostaglandins but can also metabolize linoleic acid to (9R)- and (13S)-HODE (35-37). A risk factor associated with the etiology of breast cancer arises from animal and human epidemiology studies linking high dietary fat with increased incidence of breast cancer. Linoleic acid is the major polyunsaturated fatty acid consumed in the human diet and has been demonstrated to stimulate cell proliferation and metastasis in human breast carcinoma cells (38, 39). Our investigation indicates that linoleic acid metabolism may play a central role in transduction of the EGF mitogenic signal from the cell surface to the nucleus and may provide a useful model for understanding neoplastic progression at a molecular level.