Epidermal and Hepatocyte Growth Factors, but Not Keratinocyte Growth Factor, Modulate Protein Kinase C (cid:1) Translocation to the Plasma Membrane through 15( S )-Hydroxyeicosatetraenoic Acid Synthesis* □ S

Activation of protein kinase C (PKC) involves its re-cruitment to the membrane, where it interacts with its activator(s). We expressed PKC (cid:1) fused to green fluorescent protein and examined its real time translocation to the plasma membrane in living human corneal epithelial cells. Upon 10 min of stimulation with epidermal and hepatocyte growth factors (EGF and HGF), PKC (cid:1) translocated to the plasma membrane. Keratinocyte growth factor did not stimulate PKC (cid:1) translocation up to 1 h after stimulation. Pretreatment with the 15-lipoxygen-ase metabolite, 15( S )-hydroxyeicosatetraenoic acid (15( S )-HETE), followed by EGF or HGF, produced faster translocation of PKC (cid:1) detectable at 2 min. However, the same concentration of 15( S )-HETE alone did not stimulate translocation. 15( S )-Hydroperoxyeicosatetraenoic acid and 5( S )-HETE did not affect growth factor-induced translocation of PKC (cid:1) . PD153035, a specific inhibitor of tyrosine kinase activity of the EGF receptor, completely blocked PKC (cid:1) translocation induced by EGF. PD98059,

* This work was supported by United States Public Health Service Grants R01 EY06635 and R01 EY04928 from the NEI, National Institutes of Health and by the Neurobiotechnology Program of Louisiana. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
We have shown recently that the 12-lipoxygenase metabolite, 12(S)-HETE, a product of arachidonic acid, is involved in EGF-induced proliferation of rabbit corneal epithelial cells (13). 12(S)-HETE is the major LOX metabolite produced in rabbit cornea after injury, whereas human corneal epithelial (HCE) cells express mainly 15(S)-HETE, a product of the 15-LOX (14,15).
In the eye, 15(S)-HETE stimulates migration of endothelial cells in the retina (16) and secretion of ocular mucin, which could be important in the treatment of dry eye (17,18). However, the role of 15(S)-HETE in corneal epithelium has not yet been determined.
Using a construct containing full-length PKC␣ tagged to green fluorescent protein (PKC␣-GFP), we monitored its subcellular localization upon EGF, HGF, and KGF stimulation in HCE cells. We also studied the effects of 15(S)-HETE on PKC␣ translocation.
Our experiments showed that, although EGF and HGF induced movement of PKC␣ to the membrane, KGF did not. In addition, we demonstrated that 15(S)-HETE is a second messenger involved in PKC␣ translocation affected by EGF and HGF.

EXPERIMENTAL PROCEDURES
Materials-Human EGF was obtained from Sigma. Human recombinant double-chain HGF was a gift from Genentech (San Francisco, CA). Human recombinant KGF was from Upstate Biotechnology, Inc. (Lake Placid, NY). The mitogen-activated protein kinase/extracellular signal-regulated kinase kinase (MEK) inhibitor PD98059, the specific inhibitor of the tyrosine kinase activity of the EGF receptor, PD153035, and PKC␣ inhibitor Go6976 were obtained from Calbiochem. Mouse monoclonal phosphorylated ERK1/2 (p-ERK1/2) antibody was from Sigma. Anti-ERK1 antibody was from BD Transduction Laboratories (San Jose, CA). For immunofluorescence staining and Western blot, anti-phospho-cPLA 2 (p-cPLA 2 ), anti-cPLA 2, and anti-PKC␣ antibodies were obtained from Cell Signaling Technology (Beverly, MA) and Santa Cruz Biotechnology, Inc. (Santa Cruz, CA). Anti-glyceraldehyde-3-phosphate dehydrogenase was from Research Diagnostics Inc. Anti-rabbit Ig (donkey) or anti-mouse Ig (sheep) fluorescein-linked secondary antibodies and ECL Western blotting system were obtained from Amersham Biosciences. Thermo Plate, a thermal stage to maintain the cells at 37°C, was purchased from Nikon Inc. All SDS-PAGE reagents were from Bio-Rad. The horseradish peroxidase protein marker detection kit was from New England Biolabs (Beverly, MA), and the biotinylated protein ladder detection pack was from Cell Signaling Technology. [ 3 H]Arachidonic acid (AA) (specific activity 189 Ci/mmol) was purchased from PerkinElmer Life Sciences, and 5(S)-, 12(S)-, 15(S)-HETE, 15(S)-HpETE, and AA were from Cayman Chemical Company (Ann Arbor, MI). The LOX inhibitor cinnamyl 3,4-dihydroxy-␣-cyanocinnamate (CDC) was from BIOMOL International L.P. (Plymouth Meeting, PA). Organic solvents were of HPLC grade (Fisher). The CyQuant cell proliferation assay kit was from Molecular Probes (Eugene, OR).
Cell Culture-Immortalized HCE cells that express EGF, HGF, and KGF receptors were obtained from Dr. Roger Beuerman (Department of Ophthalmology, Louisiana State University Health Sciences Center) and maintained in serum-free keratinocyte growth medium (KGM; Clonetics, BioWhittaker Europe) supplemented with the appropriate growth factors and antibiotics, essentially as described earlier (9). Experiments using HCE were performed between passages 25-45.
Expression of PKC␣-GFP-The PKC␣-GFP plasmid (a gift from Dr. Rosario Rizzuto, University of Ferrara, Italy) is a construct of pcDNA3 containing the PKC␣ gene fused with GFP. HCE cells were cultured in one-well Lab-Tek glass chamber slides or 35-or 60-mm cell culture dishes (Corning Incorporated, NY) in KGM as needed and allowed to grow to 50 -60% confluence. At this stage, the cells were transfected with 2 g/dish plasmid using FuGENE 6 transfection reagent (Roche Applied Sciences) as described in the manufacturer's protocol. Briefly, the cells were incubated with FuGENE 6 and PKC␣-GFP DNA (1:3 ratio) overnight in KGM to allow delivery of PKC␣-GFP into the cells. An equal volume of FuGENE6 was added to control cells (it had no adverse effect on cells). The transfected cells were fed with fresh KGM after 12 h and incubated at 37°C for another 24 h to allow the expression of PKC␣-GFP. Similarly, cells transfected with vector containing only GFP but no PKC␣ genes were used as negative controls. The cells were starved in keratinocyte basal medium (KBM: KGM without any growth supplements) for 16 -18 h prior to the experiment. The cells were examined under fluorescence microscope to confirm the desired transfection efficiency (50 -60%).
Real Time Translocation of PKC␣-GFP to the Plasma Membrane-Transfected cells were starved as mentioned above, and slides were secured on a Thermo Plate kept at 37°C. The cells were stimulated with growth factors (HGF and KGF (20 ng/ml) or EGF (10 ng/ml)), and the images were recorded prior to treatment, control (t ϭ 0 min), at 60-s intervals up to 5 min, and at 5-min intervals up to 30 min. The initial microscope stage position was not changed so that the same group of cells could be continuously monitored. In some experiments, the cells were pretreated for 30 min with 1 M 15(S)-HETE or inhibitors: PD153035 (20 M), CDC (10 M; CDC is a 12-lipoxygenase inhibitor at lower concentration and inhibits 15-LOX at higher concentrations; IC 50 ϭ 3.33 M), or PD98059 (25 M). The images were recorded as described above. HCE cells were then stimulated with the indicated growth factors. The images were recorded by fluorescence microscope (Nikon Eclipse TE200), with a software-controlled shutter to minimize GFP photobleaching, and an attached Nikon digital camera (DXM1200), at 20ϫ magnification in a dark room using Meta Vue 5.0 (Nikon Inc.) imaging software.
Immunofluorescence Staining of Phospho-cPLA 2 /cPLA 2 -HCE cells were seeded in 4-well Lab-Tek chamber slides and allowed to grow to 50 -60% confluence. The cells were starved in KBM medium for 36 h and stimulated with EGF (10 ng/ml) for the indicated times. After a brief wash with prechilled PBS, the cells were fixed and permeabilized with 4% paraformaldehyde for 30 min at room temperature. This was followed by extensive washes with PBS and further incubation for 60 min at room temperature with 1% bovine serum albumin and 10% normal goat serum in PBS to suppress nonspecific binding of IgG. The cells were subsequently incubated with primary antibodies against anti-phospho-cPLA 2 (1:100) and anti-cPLA 2 (1:100) at 4°C for 16 h in PBS supplemented with 1.5% normal blocking serum. After three washings with PBS, the cells were incubated with fluorescein-linked Ig anti-mouse (sheep, 1:50) or anti-rabbit (donkey, 1:50) at room temperature for 45 min and washed extensively with PBS. The cell nuclei were stained by 2 M Hoechst 33258 (Molecular Probes, Eugene OR) for 30 min at room temperature. Slides were washed twice with PBS, mounted in aqueous mounting fluid (Lerner Laboratories), and examined under a fluorescence microscope with appropriate filters. In all experiments, negative controls were incubated with mouse or rabbit fluoresceinlinked IgG, in the absence of specific primary antibody.
Western Blotting-HCE cells were incubated at 37°C overnight in KBM followed by stimulation with EGF (10 ng/ml) for 1, 2, 5, 10, and 15 min. In some experiments, the cells were preincubated with 25 M PD98059 for 30 min prior to stimulation with EGF. Inhibitors were dissolved in Me 2 SO, and the same concentration Me 2 SO (0.01%) was added to controls. For PKC␣ translocation studies, cytosol and membrane fractions were prepared from PKC␣-GFP-transfected cells as described (6) with some modifications. The translocation of PKC␣-GFP was examined by Western blotting using anti-PKC␣ antibodies. Equal loading was determined by anti-glyceraldehyde-3-phosphate dehydrogenase antibodies. Activation of cPLA 2 and ERK1/2 in HCE was evaluated using phospho-specific antibodies (12). Briefly, 30 g of protein/ well was separated on SDS-PAGE (10% gel) and then transferred to nitrocellulose membrane using a Bio-Rad Mini Trans Blot transfer unit. The membranes were blocked with Tris-buffered saline (20 mM Tris-HCl, 150 mM NaCl, pH 7.4, 0.05% Tween 20) containing 5% bovine serum albumin for p-cPLA 2 and 5% nonfat dry milk for p-ERK1/2 and PKC␣ immunoblotting; incubation with specific primary antibodies followed. The membranes were washed five times (5 min/wash) with Tris-buffered saline (0.05% Tween 20) and further incubated with appropriate secondary antibodies. The membranes were stripped using a standard ECL kit protocol and reprobed with anti-cPLA 2 or anti-ERK1 antibodies. The separated proteins were visualized by an ECL kit according to the manufacturer's protocol. Intensities of the respective bands were quantified by densitometric analysis (Bio-Rad Molecular Analyst program).

Reverse Phase HPLC Analysis of [ 3 H]AA and Its
Metabolites-HCE cells (5 ϫ 10 5 ) were incubated overnight with 1 Ci of [ 3 H]AA in KBM medium to allow the incorporation of fatty acids into membrane lipids. After three washes with KBM, the cells were stimulated with EGF (10 ng/ml) for 5 and 15 min in KBM supplemented with 0.0125% bovine serum albumin fraction V. The medium was collected in glass tubes pretreated with Sigmacote (Sigma). The cells were washed three times, scraped in prechilled PBS (pH 7.4), and homogenized by 20 pulses through a 20-gauge needle. Homogenates were centrifuged at 100,000 ϫ g for 60 min at 4°C, and supernatants (cytosolic fraction) were collected. Medium and the cytosolic fraction from the cells were acidified to pH 3.0 with formic acid (8.8%) followed by three extractions with ethyl acetate (1:2 v:v) containing 2 g of unlabeled 5, 12, and 15(S)-HETE and 4 g of AA as carriers. The lipid extracts were dried under N 2 and resuspended in a small volume of methanol to store under N 2 at Ϫ80°C. For HPLC studies, the extracts were dried under N 2 and resuspended in 45 l of mobile phase (acetic acid:MeOH:H 2 O 1:78:21 v:v:v) (13) delivered by a Agilent 1100 series Quat Pump (Hewlett-Packard) and separated on an Ultermex 5 C-18 (Phenomenex, CA) reverse phase column at a flow rate of 1 ml/min. In all studies, the retention times of 5, 12, and 15(S)-HETE and AA were confirmed by co-migration with the added unlabeled standards determined by UV spectra (DAD1A Signal 205 nm for AA and DAD1D Signal 235 nm for HETEs). Fractions from the HPLC eluate were collected every minute and counted on a 1414 Win Spectral DSA-based liquid scintillation counter (Wallac, Turku, Finland) to analyze product formation.
Cell Proliferation Assay-HCE cells were seeded (5000 cells/well) into 96-well microplates and allowed to attach overnight. The cells were serum-starved for 24 h, then treated with EGF (10 ng/ml) and/or 15(S)-HETE (1.0 M) and/or inhibitors CDC (3 M) and Go6976 (200 nM), and incubated at 37°C for 48 h. For these experiments CDC was used at similar concentration to IC 50 to be able to reverse the inhibitory effect with 1.0 M 15(S)-HETE. Each condition was performed in octuplicate, and cell proliferation was determined by a CyQuant cell proliferation assay kit as described earlier (13,19). To examine reversal of 15-LOX inhibition by 15(S)-HETE, the cells were co-incubated with 15(S)-HETE and CDC along with EGF. Fluorescence was measured on a Fluoroskan Ascent FL plate reader (Labsystems) with 485-nm excitation and 538-nm emission maximum filters. In experiments with inhibitors, the controls were supplemented with the same concentration of vehicle (Me 2 SO or ethanol, final concentration less than 0.2%).
Statistical Analysis-The significance of data was analyzed by Student's t test. Values of p Ͻ 0.05 were considered significantly different. ing to track PKC␣-GFP movement in living cells. The HCE cells were transfected with PKC␣-GFP and starved overnight as described under "Experimental Procedures" and then secured on the Thermo Plate at 37°C. Transfected cells without stimulation showed marked expression and distribution of GFP-PKC␣ in the cytosol, whereas the nuclei appeared to be devoid of enzyme expression (Fig. 1A). Following EGF treatment, the increase in PKC␣ intensity in the plasma membrane was first detected at 10 min when compared with control cells (Fig. 1A). Additional accumulation of PKC␣ in the plasma membrane occurred by 15 min and was sustained up to 30 min. Observation of different fields at that time showed a similar pattern of PKC␣ movement. Cells transfected with vector containing GFP but no PKC␣ gene were used as a negative control and showed no changes in GFP distribution upon EGF stimulation (data not shown). To further confirm the results, we performed Western blotting and found a significant translocation of PKC␣-GFP and endogenous PKC␣ to membrane fractions at 15 min (Fig. 1B). When PKC␣-GFP-transfected HCE cells were incubated with 1 M 15(S)-HETE up to 30 min, there were no changes in PKC␣ distribution. However, addition of EGF after 30 min of 15(S)-HETE preincubation triggered a rapid PKC␣ translocation that was noticeable after 1 min of stimulation, peaked at 5 min, and was sustained until 30 min (Fig. 1C). To examine the specificity of 15(S)-HETE, the cells were preincubated with the intermediate hydroperoxide 15(S)-HpETE (500 nM or 1 M) and then stimulated with EGF. The addition of 15(S)-HpETE did not influence EGF-induced PKC␣ translocation (Fig. 1D). Similarly, 5(S)-HETE (500 nM or 1 M; data not shown) did not affect EGF-mediated PKC␣ translocation. Neither 15(S)-HpETE nor 5(S)-HETE alone changed PKC␣ localization. A video depicting the real time change in PKC␣ localization to the plasma membrane upon EGF stimulation in the presence of 15(S)-HETE is available on-line and linked to Fig. 1E.

15(S)-HETE Accelerated EGF-mediated
To confirm the specificity of EGF-induced translocation of PKC␣, the cells were pretreated with PD153035, an inhibitor of tyrosine kinase activity of the EGF receptor, for 30 min prior to stimulation with EGF. Incubation with PD153035 inhibited EGF-induced PKC␣ translocation, and up to 15 min, PKC␣-GFP distribution was limited to the cytoplasmic compartment. The further addition of 15(S)-HETE did not reverse this inhibition ( Fig. 2A). To determine whether there was involvement of the MEK/ERK1/2 mitogen-activated protein kinase pathway in this response, the cells were preincubated with PD98059 for 30 min followed by stimulation with EGF. Incubation of cells with PD98059 alone did not change PKC␣ distribution. EGFmediated translocation of PKC␣ to the plasma membrane was inhibited upon pretreatment with PD98059. The addition of 15(S)-HETE after 15 min of EGF stimulation reversed PD98059-mediated inhibition of PKC␣ translocation (Fig. 2B).
To further determine the role of 15(S)-HETE, HCE cells were preincubated with the LOX inhibitor CDC for 30 min followed by stimulation with EGF. Inhibition of 15-LOX blocked EGFinduced PKC␣ translocation to the plasma membrane. Moreover, there was no noticeable difference in PKC␣-GFP distribution when compared with control cells up to 30 min (Fig. 2C). The addition of 15(S)-HETE alone did not induce translocation of PKC␣ to the plasma membrane, but when combined with EGF, 15(S)-HETE reversed the inhibitory effect of CDC and significantly translocated PKC␣ to the plasma membrane at 5 min (Fig. 2D).
Differential Effect of HGF and KGF Stimulation on PKC␣ Translocation to the Plasma Membrane-We next examined by real time imaging how these two paracrine growth factors affect PKC␣ movement in PKC␣-GFP-transfected HCE cells. HGF (20 ng/ml) produced a small change in the distribution of PKC␣ in the plasma membrane at 10 min, an effect that was sustained for 30 min (Fig. 3A). Pretreatment with 15(S)-HETE (1 M) for 30 min and stimulation with HGF induced a noticeable translocation of PKC␣ to the plasma membrane as early as 1 min (Fig. 3B). Inhibition of the MEK/ERK1/2 pathway by preincubation with PD98059 blocked HGF-induced PKC␣ translocation to the plasma membrane (Fig. 3C). KGF (20 ng/ml), on the other hand, did not produce any change in PKC␣ localization up to 30 min (Fig. 3D). A higher concentration (50 ng/ml) of KGF did not induce PKC␣ translocation (data not shown). Similarly, HGF was tested at higher concentrations, with no difference from the response elicited by 20 ng/ml HGF. As a positive control, cells treated with KGF were stimulated with 12-O-tetradecanoylphorbol-13-acetate (200 nM) for 10 min, which induced marked translocation of PKC␣ to the plasma membrane (Fig. 3D), which is in agreement with previous reports in human corneal epithelial cells (12) as well as other cell types (20).
EGF Induced a Prompt cPLA 2 and ERK1/2 Activation-Because EGF was more potent than HGF, EGF was used to study the involvement of cPLA 2 and the MEK/ERK1/2 pathway in PKC␣ translocation. HCE cells were treated with EGF and analyzed for cPLA 2 and ERK1/2 activation by immunofluorescence staining and Western blotting. Immunostaining with anti-p-cPLA 2 showed very low levels of p-cPLA 2 staining in control cells. EGF stimulation produced a rapid increase in p-cPLA 2 staining in the plasma membrane at 1 min (as indicated by arrows in Fig. 4A). A significant increase in p-cPLA 2 staining was noticed at 2 min, localized in the plasma membrane and perinuclear regions (indicated by arrows in enlarged Fig. 4A). Increased p-cPLA 2 staining was also present in cy- FIG. 4. EGF induces a rapid ERK1/2 activation, which precedes cPLA 2 phosphorylation. HCE cells were plated in 4-well Lab-Tek glass chamber slides or 60-mm dishes and treated with EGF (10 ng/ml) for the indicated times and analyzed by immunostaining and Western blotting for cPLA 2 and ERK1/2 activation. A, cells were stimulated with EGF and immunostained for p-cPLA 2 and cPLA 2 to analyze the activation and localization of cPLA 2 . The cells were counterstained for nuclei with Hoechst reagent, and an overlay is shown. An enlarged image of EGF at 2 min clearly shows plasma and perinucelar localization of p-cPLA 2 as indicated by arrows. Total cPLA 2 staining did not noticeably change with these treatments. B, cells were stimulated with EGF alone for the indicated times or pretreated with PD98059 for 30 min before stimulation with EGF for 5 min and immunoblotted with anti-phospho-cPLA 2 or cPLA 2 antibodies. Stimulation of p-cPLA 2 is indicated as fold increases as compared with the control (upper panels). In another experiment, immunoblotting was performed with anti-p-ERK1/2 and anti-ERK1 antibodies (lower panels). The total cPLA 2 and ERK1/2 did not change with similar treatments. The data represent three separate experiments.
tosol. The last panels of Fig. 4A show that total cPLA 2 staining was not changed upon EGF stimulation at 2 min compared with control cells. Cells were counter-stained with Hoechst 33258 for nuclear staining. Western blot analysis also showed EGF stimulation of p-cPLA 2 at 2 min (2.2-fold) and 5 min (2.8-fold) compared with controls, which was maintained until 10 min. The membranes were stripped and reprobed for total cPLA 2, which did not change (Fig. 4B, upper panels). Similarly, HCE cells were evaluated for ERK1/2 activation upon EGF stimulation. An increase in p-ERK1/2 occurred by 1 min and was sustained at 10 min. To confirm equal loading, the membrane was stripped and reprobed for total ERK1/2, which did not show any changes (Fig. 4B, lower panels). Preincubation with PD98059 inhibited phosphorylation of cPLA 2 (to basal levels) as well as of ERK1/2, indicating that EGF stimulation of ERK1/2 plays an important role in cPLA 2 activation.

15(S)-HETE Reversed the Effect of LOX Inhibitor on EGFinduced Cell
Proliferation-To investigate the functional involvement of EGF-induced PKC␣ activation and the role of the 15-LOX/15(S)-HETE pathway, we performed a cell prolifera-tion assay. The HCE cells were seeded overnight and starved for 24 h before stimulation with EGF for 48 h with or without inhibitors. Vehicle (ethanol and/or Me 2 SO) or 15(S)-HETE alone (controls) did not affect proliferation at the same concentrations. EGF stimulation produced a 40% increase in proliferation (Fig. 6, *, p Ͻ 0.01) compared with control cells. When CDC was added along with EGF, cell proliferation was significantly inhibited (Fig. 6, **, p Ͻ 0.05) compared with EGF alone. 15(S)-HETE completely reversed the inhibitory effect of CDC (Fig. 6, ***, p Ͻ 0.05, compared with CDC ϩ EGF). Blockade of PKC␣ activity with Go6976 also significantly inhibited cell proliferation (Fig. 6, **, p Ͻ 0.05) compared with EGF. Because neither CDC nor Go6976 could completely inhibit proliferation, we added EGF in the presence of both inhibitors. This reduced cell proliferation to nearly control levels, suggesting an additive effect (Fig. 6, #, p Ͻ 0.05, compared with EGF ϩ CDC or EGF ϩ Go6976).

DISCUSSION
In the present study, we investigated changes in the subcellular localization of PKC␣ upon EGF, HGF, and KGF stimulation as well as the underlying signaling mechanisms in corneal epithelial cells transfected with PKC␣-GFP. Our results show that both EGF and HGF induced PKC␣-GFP translocation to the plasma membrane. Translocation of PKC␣ was more potent with EGF than with HGF, probably because of a stronger ERK1/2 activation compared with HGF (9). In contrast, KGF did not induce translocation of PKC␣. Both HGF and KGF are paracrine growth factors that induce proliferation of corneal epithelial cells (13), but there are no reports indicating how HGF and KGF differ in their signaling mechanisms. We have previously shown that activation of ERK1/2 by HGF is more potent and long lasting (60 min) compared with that by KGF, which returns to basal levels at 30 min (12). These differences may in part explain why KGF did not induce noticeable PKC␣ translocation in real time experiments. Unlike EGF and HGF, KGF does not stimulate corneal cell migration (21), but the three growth factors stimulate corneal cell proliferation and wound healing (9,10,13). Our results suggest that in the case of KGF, the signaling cascade does not involve PKC␣ translocation to the plasma membrane.
Translocation of PKC␣ was clearly demonstrated at 15 min with both growth factors, and in the case of EGF, blocking the receptor abolished this translocation, which was not reversible by the addition of 15(S)-HETE. Corneal epithelial injury induces phosphorylation of the EGF receptor and ERK1/2 activation in HCE cells, and blockade of both delays wound healing (9,22). Recent reports indicate that conventional and novel PKCs, such as PKC␣, ␤I, ␤II, ␦, and ⑀, are not involved in EGF-induced ERK1/2 activation and/or cell proliferation in corneal epithelial cells (12). Because PKC␣ expression as well as activity is increased in proliferating epithelium after injury (6,7), this raises the possibility that PKC␣ interaction/stimulation with growth factors is not by way of direct activation by PKC␣ of ERK1/2 (12). Our data show that the inhibition of ERK1/2 abolished EGF-and HGF-stimulated PKC␣ translocation and that the addition of 15(S)-HETE reversed this inhibition in the presence of EGF. Blocking ERK1/2 activation also inhibited cPLA 2 activation to almost basal levels, suggesting that the ERK1/2 pathway is involved in the activation of cPLA 2 and 15(S)-HETE production. Thus, ERK1/2 activation is an important signal in the translocation and activation of PKC␣ by these growth factors, as depicted schematically in Fig. 7.
Recently we showed that EGF, HGF, and KGF stimulate 12(S)-HETE production in rabbit corneal epithelial cells through expression of a platelet-type 12-lipoxygenase, which induces cell proliferation (13). 12(S)-HETE is the main lipoxygenase metabolite increased in rabbits after corneal injury (14).
In lens epithelial cells EGF and 12(S)-HETE activate PKC␣ and ␤ (23). Human corneal epithelium contains mainly 15-LOX activity (15), and here, we demonstrate that the eicosanoid acts as a modulator of EGF-and HGF-induced PKC␣ translocation. The addition of 15(S)-HETE before EGF or HGF stimulation accelerated the response in HCE cells, and PKC␣ translocated to the membrane as early as 2 min, compared with 15 min with EGF or HGF stimulation alone. This effect is specific, and a hydroperoxy form of 15(S)-HETE, 15(S)-HpETE, or 5(S)-HETE did not induce PKC␣ translocation. Treatment of the cells with the same concentrations of 15(S)-HETE alone did not produce PKC␣ translocation, indicating that this is not the sole target of the signaling cascade leading to PKC␣ translocation by EGF and HGF. This is in contrast to prior results in lens epithelial cells, which showed that 12(S)-HETE induced PKC␤ translocation to the plasma membrane (24). In our hands, higher concentrations of 15(S)-HETE (Ͼ2 M) were needed to produce PKC␣ translocation. These results also suggest that the lipoxygenase metabolites could be selective in the activation of different PKC isoforms and that their effects may vary depending on cell type.
EGF induced a rapid synthesis of 15(S)-HETE that in part was retained intracellularly. This is in agreement with previous studies showing that EGF and insulin induce AA release to produce 5-, 12-, and 15(S)-HETE, which enhance mouse mammary epithelial cell proliferation (25). In fact, inhibiting the synthesis of 15(S)-HETE abolished the translocation of PKC␣ induced by EGF and HGF, suggesting the need for the growth factors to activate the signaling cascade involving ERK1/2-cPLA 2 to produce PKC␣ activation. These changes, along with added 15(S)-HETE, could increase the intracellular pool of 15(S)-HETE and accelerate the translocation of PKC␣. The mechanism by which 15(S)-HETE activates translocation of PKC␣ to the plasma membrane is not known. One possibility is that part of endogenous pool of 15(S)-HETE is esterified to FIG. 7. Schematic representation of the role of 15(S)-HETE in growth factor-induced PKC␣ translocation. Corneal injury induces the release of growth factors (e.g. EGF and HGF) that activate ERK1/2, which stimulates phosphorylation of cPLA 2 and its movement to the plasma and nuclear membranes, where it releases AA from phospholipids. AA is then converted to 15(S)-HETE by 15-LOX, possibly by 15-LOX-2 in the nuclear membrane (28,29). 15(S)-HETE then triggers PKC␣ translocation to the plasma membrane and its subsequent activation. Activated PKC␣ induces cell proliferation and promotes wound healing (6, 7). some of the phospholipid components of the inner layer of the plasma membrane (e.g. phosphatidylethanolamine, phosphatidylserine, and phosphatidylinositol) and induces changes in the properties of the membrane that facilitate the translocation of PKC␣. In fact, in vitro experiments using human tracheal epithelial cells showed selective incorporation of 15(S)-HETE into the sn-position of phosphatidylinositol and selective activation of PKC␣ after formation of diacylglycerol-containing 15(S)-HETE (26). We have previously shown that in rabbit corneal epithelial cells, 12(S)-HETE is rapidly esterified to the membrane phospholipids (27).
To further understand the dynamics of the ERK1/2 and cPLA 2 signaling pathway leading to AA release and 15(S)-HETE production, we also analyzed the time course of EGFinduced activation of ERK1/2 and cPLA 2 . Stimulation by EGF induced ERK1/2 activation that was followed by cPLA 2 phosphorylation and its accumulation in the plasma membrane as well as in perinuclear regions. Translocation of p-cPLA 2 to the perinuclear region has been reported in response to calcium (28). In addition, recent studies showed a 15-LOX-2 enzyme in human corneal epithelial cells localized in the nuclear membrane (29). Blockade of the ERK1/2 pathway significantly reduced p-cPLA 2 , suggesting that ERK1/2 is involved in cPLA 2 activation in HCE cells. The active cPLA 2 then releases AA from membrane phospholipids, which is converted to 15(S)-HETE by the action of 15-LOX. This stimulates PKC␣ translocation to the membrane (Fig. 7). The inhibition of EGFstimulated HCE cell proliferation when the lipoxygenase was blocked and the converse action by 15(S)-HETE demonstrate the functional role of the eicosanoid. It is interesting to note that inhibition of PKC␣ or 15-LOX does not completely block proliferation, suggesting that other pathways are involved in the proliferative actions of these two growth factors. In fact, complete inhibition was obtained by blocking 15(S)-HETE as well as PKC␣ activation.
In conclusion, this study shows a new signaling mechanism by which EGF and HGF, but not KGF, mediate PKC␣ translocation to the plasma membrane. We demonstrate that there is growth factor selectivity in inducing PKC␣ translocation, utilizing an ERK1/2/cPLA 2 /15-LOX pathway. PKC␣, once localized in the plasma membrane, is eventually activated (3) and induces cellular proliferation of corneal epithelial cells after injury. This demonstrates, for the first time, a functional role for 15(S)-HETE in human corneal epithelial cells.