Activation of Vascular Endothelial Growth Factor Receptor 2 in a Cellular Model of Loricrin Keratoderma*

Loricrin is a major constituent of the epidermal cornified cell envelope. Recently, heterozygous loricrin gene mutations have been identified in two dominantly inherited skin diseases, Vohwinkel syndrome with ichthyosis and progressive symmetric erythrokeratoderma, collectively termed loricrin keratoderma. We generated stable HaCaT cell lines that express wild-type (WT) loricrin and a mutant form found in Vohwinkel syndrome with ichthyosis, using an ecdysone-inducible promoter system. The cells expressing the mutant loricrin grew more rapidly than those expressing WT loricrin after induction for 5 days. Confocal immunofluorescence microscopy revealed that phospho-Akt occurred in the nucleolus where the mutant loricrin was also located. The level of activity of Akt kinase was about nine times higher in cells with the mutant than in those with WT loricrin. ERK1/2, the epidermal growth factor receptor, vascular endothelial growth factor (VEGF) receptor 2 and Stat3 were all phosphorylated in cells with the mutant loricrin. The docking proteins, Gab1 and c-Cbl, were also tyrosine-phosphorylated in these cells. Furthermore, chromatin immunoprecipitation assays showed that Stat3 protein bound to the VEGF promoter in cells with the mutant. Thus, this study suggests that VEGF release and the subsequent activation of VEGF receptor 2 link loricrin gene mutations to rapid cell proliferation in a cellular model of loricrin keratoderma.

The stratum corneum functions as a barrier both to protect against environmental insults and prevent water loss. These functions are mainly attributed to cornified cell envelope (1,2) formed beneath the plasma membrane in terminally differentiating stratified squamous epithelia. It provides a vital physical barrier in mammals and consists of a 10-nm-thick layer of highly cross-linked insoluble proteins. Its components are several epidermis-specific structural proteins, involucrin, cystatin A, and loricrin (OMIM 152445); several small proline-rich proteins, trichohyalin, profilaggrin, repetin, hornerin, elafin, and profilaggrin-related proteins; S100 family proteins, and some desmosomal proteins and keratins, assembled by the catalytic action of transglutaminases.
Loricrin (Latin for "lorica," a protective shell or cover) is incorporated into the scaffold formed with involucrin, envoplakin, and periplakin. Loricrin consists of many tandem quasirepeats in the form of aliphatic (glycine/serine/cysteine) n loops, which are interspaced by glutamine/serine-rich domains (3)(4)(5). Recently, unique heterozygous, insertional mutations in the loricrin gene have been found to cause some congenital skin abnormalities (6 -14). Clinically, the diagnosis for such a condition can be Vohwinkel syndrome with ichthyosis (OMIM 604117), progressive symmetric erythrokeratoderma (OMIM 602036), or congenital ichthyosiform keratoderma born as a collodion baby. The clinical features originally described by Vohwinkel in 1929 include the following: (i) honeycomb-like palmoplantar keratoderma accompanying small "honeycomb" depressions; (ii) starfish-like hyperkeratosis and hyperkeratotic knuckle pads on dorsal parts of hands; and (iii) pseudoainhums of the fingers and/or toes leading to autoamputation. If these signs are associated with hearing impairment, the diagnosis is classic (hearing loss-associated) Vohwinkel syndrome (OMIM 124500: deafness, congenital, with keratopachydermia and constrictions of fingers and toes) caused by a mutation in the connexin 26 gene (GJB2). Vohwinkel syndrome caused by an insertional loricrin mutation is currently termed loricrin keratoderma (LK) 2 (OMIM 604117) (15)(16)(17). Patients from nine families with four different mutations have been reported so far. The most frequent mutation, 730insG, has been found in families from the United Kingdom, Japan, and Italy.
We have previously shown that the expression of wild-type (WT), but not a mutant, loricrin causes programmed cell death in HaCaT keratinocytes (18). We have demonstrated that WT loricrin-transfected HaCaT keratinocytes are susceptible to programmed cell death caused by the activation of caspase-14.
Although such a function of WT loricrin is plausible, it was not possible to quantify biochemical changes occurring in these cells due to the low frequency of transient transfections. Hence, we created stable human keratinocyte cell lines in which WT and mutant loricrin are expressed in an inducible manner using an ecdysone-inducible promoter system (19).
Here, we demonstrate that overexpression of the mutant loricrin causes the release of vascular endothelial growth factor (VEGF) and transforming growth factor-␣ (TGF-␣) from HaCaT keratinocytes and the subsequent activation of vascular endothelial growth factor receptor 2 (VEGFR 2). We speculate that the activation of VEGFR 2 by an autocrine/paracrine pathway links loricrin gene mutations to rapid cell proliferation in a cellular model of LK.

EXPERIMENTAL PROCEDURES
Plasmid Construction-Genomic DNA containing the entire coding region of WT loricrin and mutant loricrin was subcloned into the pIND/V5-His vector (Invitrogen) (3)(4)(5). The most frequent mutation, 730insG, was chosen for this study. The sequence of each of the plasmid constructs was verified by the dideoxynucleotide chain termination method using the 377 DNA sequencing system (Applied Biosystems Inc., Foster City, CA).
Cell Culture, Plasmid Transfection, and Establishment of Inducible Cell Lines-The ecdysone-inducible mammalian expression system from Invitrogen was used (19). The culture and transfection of HaCaT cells were carried out as described previously with minor modifications (20). Briefly, cells were plated on 35-or 60-mm culture dishes at a density of 4 ϫ 10 5 cells/ml 24 h before the transfection and cultured in Dulbecco's modified Eagle's medium (450 mg/dl glucose) supplemented with 10% (v/v) fetal bovine serum. A portion, 2 g for 35-mm dishes and 10 g for 100-mm dishes of pVgRXR, mock, WT loricrin, or mutant loricrin in pIND/V5-His vector, was transfected into cells with Lipofectamine Plus reagent (Invitrogen) according to the manufacturer's instructions. Forty eight h after transfection, the selection of cells using Zeocin was begun, i.e. HaCaT cells were first stably transfected with pVgRXR (Zeocin-resistant) using Lipofectamine Plus reagent (Invitrogen), and then the cells in which protein expression was well regulated by muristerone A were selected through the transient expression of pIND/V5-His WT keratin 14. The transfected WT keratin 14 was stained with an anti-V5 antibody (Invitrogen). The selected cells (EcR-4) were then transfected with pIND/V5-His WT loricrin and pIND/V5-His mutant loricrin to generate stable cell lines expressing the WT and mutant loricrin. Individual clones were selected using 0.4 mg/ml Zeocin and 0.4 mg/ml G418. Zeocin and G48-resistant clones were then screened for transgene expression induced with muristerone A (Invitrogen) by microscopic observation of V5 staining and immunoblot analysis using the anti-V5 antibody (Invitrogen). Experiments described utilized the permanently transfected cells (referred to as WL-1, WL-13, VL-2, and VL-5 cells); however, similar results were obtained with multiple independent transient transfections.
DNA Synthesis Labeling and Detection-5-Bromo-2Ј-deoxyuridine (BrdUrd) labeling and detection were conducted with BrdUrd labeling and detection kit (Sigma). Induced cells on glass coverslips were incubated with BrdUrd labeling medium for 30 min at 37°C and fixed with ethanol. Detection was with anti-BrdUrd mouse monoclonal antibody and fluorescein isothiocyanate-conjugated antibody to mouse IgG. Propidium iodide (Sigma) was used for nuclear detection. Immunofluorescent images were viewed with a confocal laser microscope (Bio-Rad).
Akt Kinase Assay-The Akt kinase assay was performed with the Cell Signaling Technology Akt kinase assay kit. After 20 h of treatment with muristerone A, mock-transfected, WL-1, and VL-5 cells at ϳ80 -90% confluency in a 100-mm culture dish containing 10 ml of growth medium were lysed with 1ϫ Cell Lysis Buffer plus 1 mM phenylmethylsulfonyl fluoride. The cells were incubated on ice for 5 min. Cells were scraped from the dish into a microcentrifuge tube, sonicated, and microcentrifuged for 10 min at 4°C. Next, 20 l of immobilized Akt antibody bead slurry was added to 200 l of cell lysate. Incubation was done overnight at 4°C with rotation. After centrifugation of the cell lysate/immobilized antibody in the microcentrifuge, the anti-Akt antibody-bead-Akt complex was washed with 1ϫ Cell Lysis Buffer and 1ϫ Kinase Buffer. The pellet was suspended in 50 l of 1ϫ Kinase Buffer supplemented with 1 l of 10 mM ATP and 1 g of GSK-3 fusion protein, and incubation was conducted for 30 min at 30°C. The reaction was terminated with 25 l of 3ϫ SDS sample buffer, and immunoblotting was conducted with the anti-phospho-GSK-3␣/␤ (Ser-21/9) antibody. Protein expression was quantified by measuring the density of immunostaining by NIH Image software.
Chromatin Immunoprecipitation Assay-Chromatin immunoprecipitation assays were performed with the ChIP assay kit (Upstate Biotechnology). After 20 h of treatment with muristerone A of mock-transfected, WL-1, and VL-5 cells at ϳ80 -90% confluency in a 100-mm culture dish containing 10 ml of growth medium, formaldehyde was added at a final concentration of 1%. The cells were incubated at 37°C for 10 min, then scraped from the dish into a microcentrifuge tube, and centrifuged. The cell pellet was resuspended in SDS lysis buffer containing protease inhibitor. The cell lysate was then sonicated on wet ice. The cell lysate was presorbed by incubation with 60 l of protein G-agarose for 1 h at 4°C followed by centrifugation for 1 min at 3,000 ϫ g. The supernatant was incubated with 1 mg/ml anti-RNA polymerase II antibody, anti-RhoA rabbit polyclonal antibody, anti-Stat3 rabbit polyclonal antibody, and normal mouse IgG overnight at 4°C with rotation. Next, 60 l of protein G-agarose was added to each microcentrifuge tube, and incubation was further conducted for 1 h at 4°C with rotation. Protein G-agarose-antibody-protein-chromatin complexes were washed with wash buffer and TE buffer. Protein-DNA complexes were eluted with elution buffer; 8 l of 5 M NaCl was added to 200 l of eluate, and incubation was further conducted at 65°C overnight to reverse the DNA-protein cross-links. RNase A (1 l) was added to all tubes, and incubation was done for 30 min at 37°C. Subsequently, 4 l of 0.5 M EDTA, 8 l of 1 M Tris-HCl, and 1 l of proteinase K were added, and incubation was conducted at 45°C for 2 h. Next, DNA was purified using spin columns in the kit. The purified DNA was subjected to PCR with primers specific for a 130-bp region (Ϫ913 to Ϫ783) spanning the Stat3-binding site (Ϫ848) in the VEGF promoter (24). The sequences of the primers used were as follows: VEGF forward (ϩ), 5Ј-CTGGCCTGCAGA-CATCAAAGTGAG-3Ј, and VEGF reverse (Ϫ), 5Ј-CTTC-CCGTTCTCAGCTCCACAAAC-3Ј. PCR was run for 38 cycles (94°C for 30 s, 58°C for 30 s, and 72°C for 1 min), and final products were resolved on a 2.5% agarose gel.
TGF-␣ Measurement-The amount of TGF-␣ in the culture medium was measured with a quantitative sandwich enzyme immunoassay technique using a Quantikine TGF-␣ immunoassay kit (R & D Systems, Minneapolis, MN) according to the manufacturer's instructions. Briefly, 50 l of culture medium was added to each plate coated with goat polyclonal anti-TGF-␣ antibody and incubated for 2 h at room temperature on a horizontal shaker. After four washes of the plates with the washing buffer supplied, horseradish peroxidase-conjugated goat polyclonal anti-TGF-␣ antibody was added, and plates were incubated for 2 h at room temperature to allow formation of the polyclonal antibody-TGF-␣-polyclonal antibody-horseradish peroxidase complex. After four washes with the washing buffer, the plates were incubated with a chromogenic solution containing tetramethylbenzidine and H 2 O 2 for 30 min. The reaction was terminated by the addition of 0.36 N H 2 SO 4 , and absorbance at A 450 was measured with a spectrophotometer (ELNX96, TFB, Tokyo, Japan).
VEGF Measurement-The amount of VEGF in the culture medium was measured with a quantitative sandwich enzyme immunoassay technique using a Quantikine VEGF immunoassay kit (R & D Systems) according to the manufacturer's instructions. Briefly, 50 l of culture medium was added to each plate coated with goat polyclonal anti-VEGF antibody and incubated for 2 h at room temperature on a horizontal shaker. After four washes with the washing buffer supplied, horseradish peroxidase-conjugated goat polyclonal anti-VEGF antibody was added, and plates were incubated for 2 h at room temperature to allow formation of the polyclonal antibody-VEGFpolyclonal antibody-horseradish peroxidase complex. After four washes with the washing buffer, the plates were incubated with a chromogenic solution containing tetramethylbenzidine and H 2 O 2 for 30 min. The reaction was terminated by the addition of 0.36 N H 2 SO 4 , and absorbance at A 450 was measured with a spectrophotometer (ELNX96, TFB, Tokyo, Japan).
Statistics-All experiments were performed at least three times. All values are presented as means Ϯ S.E. One-way ANOVA with Tukey-Kramer multiple comparisons test was used, and the results are presented in Fig. 1D. One-way ANOVA with Dunnett's multiple comparisons test was used, and the results are shown in Fig. 1E and Fig. 10, A and B. A p value of less than 0.05 was considered statistically significant.

Generation of Keratinocyte Cell Lines That Express WT and
Mutant Loricrin-The clones with the highest levels of expression and tightest regulation were selected and named WL-1, WL-13, VL-2, and VL-5. WL-1 and WL-13 cell lines expressed WT loricrin. VL-2 and VL-5 cell lines expressed mutant loricrin. To examine whether WT and mutant loricrin were indeed induced to express by muristerone A in the WL-1, WL-13, VL-2, and VL-5 cell lines, we performed immunoblot analyses with the anti-V5 antibody and anti-mutant loricrin antibody. Immunoblotting using the anti-V5 antibody showed that the WT (35 kDa) and mutant (42 kDa) were expressed. The analysis using the anti-mutant loricrin antibody showed that the mutant was specifically expressed in VL-2 and VL-5 cells (Fig. 1A). The amounts of WT loricrin and mutant loricrin were dependent on the dose of muristerone A (Fig. 1B). To know proliferative capacity, we treated WL-1 cells, VL-5 cells, and a control (mock-transfected) cell line with 1 mM muristerone A and conducted BrdUrd labeling and detection. Ten thousand cells were chosen for the count. BrdUrd-positive cells were counted by blinded examiners. Histogram vertical axis depicts percentage of BrdUrd-positive cells. Percentage of BrdUrd-positive cells was higher in VL-5 cells than mock or WL-1 cells (Fig. 1, C and D). To further examine proliferative capacity, we treated WL-1 and VL-5 cells and a control (mocktransfected) cell line with 1 mM muristerone A and counted cell numbers daily based on trypan blue staining. The proliferation of VL-5 cells was significantly increased (Fig. 1E).

Expression of Mutant Loricrin Up-regulates Phosphorylation of Epidermal Growth Factor Receptor and Vascular Endothelial
Growth Factor Receptor 2-To determine whether the EGFR and VEGFR 2 are phosphorylated in WL-1 and VL-5 cells, we conducted immunoblot analyses with the anti-phospho-EGFR (Tyr-992) antibody, anti-phospho-EGFR (Tyr-1068) antibody, and anti-phospho-VEGFR 2 (Tyr-1175) antibody. EGFR and VEGFR 2 were both phosphorylated in the control (mocktransfected) cell line, WL-1 cells, WL-13 cells, VL-2 cells, and VL-5 cells 2 h after the addition of muristerone A. In contrast, EGFR and VEGFR 2 continued to be phosphorylated only in VL-2 and VL-5 cells 20 h after the addition of muristerone A (Fig. 2). We were unable to detect phosphorylation of EGFR or VEGFR 2 on noninduced WL-1, WL-13, VL-2, VL-5, or HaCaT cells (data not shown).
Phosphorylation of Akt and Effect of ⌬p85 Transfection on Akt Activity in VL-5 Cells-To determine whether Akt is phosphorylated in mock-transfected, WL-1, and VL-5 cells, we conducted an immunoblot analysis with the anti-phospho-Akt (Ser-473) antibody. Akt kinase was phosphorylated only in VL-5 cells 20 h after the addition of muristerone A (Fig. 3A). To identify signal transduction pathways relevant to EGFR-or VEGFR 2-dependent VL-5 cell proliferation, we first determined the effect of a transient transfection with SR␣-⌬p85 (Fig.  3B). Phosphoinositide 3-kinase (PI3K) is a heterodimer consisting of regulatory 85 kDa (p85) and catalytic 110 kDa (p110) subunits. ⌬p85 protein lacks a binding site for p110. Overexpression of ⌬p85 acts to suppress endogenous PI3K activity. When SR␣-⌬p85 was transfected, Akt kinase activity was reduced, suggesting that PI3K is required for phosphorylation of Akt in VL-5 cells (Fig. 3B).
Effect of PI3K Inhibitors on Akt Activity and Distribution of Phospho-Akt in VL-5 Cells-We also examined the effect of PI3K inhibitors on Akt activity. Akt activity is represented by the intensity of bands on the phospho-GSK-3␣/␤ (Ser-21/9) immunoblots. Wortmannin is a specific inhibitor of PI3K. LY294002 acts in vivo as a highly selective inhibitor of PI3K. When wortmannin was used at 0.2 M, phospho-Akt (Thr-308) was not detected. When wortmannin was used at 1 M, neither phospho-Akt (Ser-473) nor phospho-Akt (Thr-308) was detected. When LY294002 was used at a concentration of 10 M, both phospho-Akt (Ser-473) and phospho-Akt (Thr-308) were detected. When LY294002 was used at 50 M, neither phospho-Akt (Ser-473) nor phospho-Akt (Thr308) was detected (Fig.  4A). The density of immunostaining was measured in five experiments and quantified with NIH Image software. The level of Akt kinase activity was nine times higher in VL-5 cells than the control cell lines (mock-transfected and WL-1 cells). The activity of Akt kinase was suppressed by treatment with wortmannin or LY294002 (Fig. 4B). We double-stained WL-1 and VL-5 cells with anti-phospho-Akt (Ser-473) and anti-V5 antibodies. Phospho-Akt (Ser-473) distributed in the cytoplasm and nucleus in WL-1 and VL-5 cells. A substantial portion of phospho-Akt (Ser-473) colocalized with aggregates of the V5-tagged mutant loricrin (arrows) (Fig. 4C, VL-5). These results, when taken together, indicate that the expression of mutant, but not WT, loricrin in keratinocytes leads to PI3K-de-pendent activation of Akt. This phenomenon appears to take place in the nucleus of these cells.
Adapter Proteins Gab1 and c-Cbl Are Involved in the Signaling Pathway-To examine whether tyrosine-phosphorylated Gab-1 is associated with PI3K (25), we conducted immunoprecipitation and immunoblot experiments. Gab1 was immunoprecipitated with the anti-Gab1 antibody, transferred to a membrane, and immunoblotted with the anti-phosphotyrosine (anti-Tyr(P)), anti-Gab1, anti-p85, or anti-SHP-2 antibody (Fig. 5A). The immunoprecipitates of Gab1 from VL-5 cells contained p85 and SHP-2, indicating that Gab1 forms a complex with SHP-2 and p85 in response to the expression of mutant loricrin in VL-5 cells. The immunoblot experiment with anti-phospho-Gab1 (Tyr-307) and anti-phospho-Gab1 (Tyr-627) showed that Gab1 is phosphorylated at tyrosine 307 and tyrosine 627 in VL-5 cells (Fig. 5A). Like Gab1, c-Cbl is known to be involved in the regulation of receptor signaling (26). To explore the involvement of c-Cbl in response to the expression of mutant loricrin in VL-5 cells, we conducted immunoprecipitation and immunoblot experiments. c-Cbl was immunoprecipitated with the anti-c-Cbl antibody, transferred to a membrane, and immunoblotted with the anti-Tyr(P), anti-c-Cbl, anti-Grb2, or anti-Shc antibody. c-Cbl was also found to be tyrosinephosphorylated only in VL-5 cells following the addition of muristerone A (Fig. 5B). Furthermore, the immunoprecipitates of c-Cbl from VL-5 cells contained Grb2 and Shc. The immunoblot experiment with the anti-phospho-c-Cbl (Tyr-731) and anti-phospho-c-Cbl (Tyr-774) antibodies showed that c-Cbl is phosphorylated at tyrosine 731 and tyrosine 774 in VL-5 cells (Fig. 5B).
ERK1/2, Not p38 MAPK or SAPK/ JNK, Is Phosphorylated in VL-2 Cells and VL-5 Cells-The transcription of the loricrin gene is regulated by the binding of protein factors to an AP-1 consensus site in the proximal portion of the loricrin promoter (27). Because MAPKs are upstream activators of AP-1 transcription factors, we examined whether MAPKs are involved in the regulation of versus mock and VL-5; **, p Ͻ 0.05 versus mock and WL-1, by one-way ANOVA with Tukey-Kramer multiple comparisons test. E, to examine whether there is a difference in proliferation among cell lines, we treated mock, WL-1, and VL-5 cells with 1 mM muristerone A and counted the cells daily using trypan blue staining. VL-5 cells proliferated more actively than the mock cells. WL-1 cells proliferated less actively than the mock cells. Statistical significance was determined using the one-way ANOVA with Dunnett's multiple comparisons test (n ϭ 5). Differences were considered statistically significant at p Ͻ 0.05. *, p Ͻ 0.05 versus mock and VL-5; **, p Ͻ 0.05 versus mock and WL-1. loricrin expression. The level of phosphorylated ERK was elevated at 20 h in VL-2 and VL-5 cells (Fig. 6A). The levels of phosphorylated JNK and p38 MAPK were almost unchanged during the experimental period in mock-transfected, WL-1, WL-13, VL-2, and VL-5 cells (Fig. 6, B and C).
Stat3 Up-regulates VEGF Expression through the VEGF Gene Promoter-Because the Jak/Stat3 pathway is critical to the leptin-induced activation of VEGF and VEGFR 2 (28,29), we investigated whether Stat3 protein could directly bind to the Stat3-binding site in the VEGF promoter by conducting ChIP assays (30). This technique allows for the detection of specific genomic DNA sequences that are associated with a particular transcription factor in intact cells. As shown in Fig. 7A, an association of Stat3 with the VEGF promoter in VL-5 cells was detected. Immunoprecipitation with a Stat3 antibody followed by PCR using oligonucleotide primers that amplify a 130-bp region spanning the Stat3-binding site at Ϫ848 within the VEGF promoter yielded a 130-bp band, demonstrating the specificity of the interaction between Stat3 and the VEGF promoter. In contrast, PCR amplification of the murine albumin promoter (which does not possess Stat3 sites) from all samples revealed no enrichment for nonspecific DNA sequences in the Stat3 immunoprecipitation reaction. We next investigated whether Stat3 was phosphorylated or not. The immunoblot experiment with anti-phospho-Stat3 (Tyr-705) and anti-phospho-Stat3 (Ser-727) showed that Stat3 is phosphorylated at tyrosine 705 and serine 727 in VL-5 cells (Fig. 7B).
Stimulation of VEGFR 2 by an Autocrine/Paracrine Mechanism-The binding of VEGF to VEGFR 2 leads to the tyrosine phosphorylation of the Grb2-bound adapter Gab1. Because VEGFR 2 is expressed in keratinocytes (31), the possible involvement of the VEGF pathway in mock-transfected, WL-1, WL-13, VL-2, and VL-5 cells was examined. We measured the amount of VEGF in the medium with a solid phase enzyme amplified sensitivity immunoassay.  (Fig. 9A). We then assessed the expression of VEGF protein in keratinocytes by conducting an immunoblot analysis and found levels to be significantly higher in VL-5 cells than in mock-transfected or WL-1 cells (Fig. 9B). These results suggest that the expression of mutant loricrin stimulates the synthesis and release of VEGF in cultured VL-2 and VL-5 cells.

DISCUSSION
Loricrin is a major component of the cornified cell envelope of terminally differentiated epidermal keratinocytes. Mutations in the loricrin gene have been identified in Vohwinkel syndrome with ichthyosis and progressive symmetric erythrokeratoderma (6 -17). A transgenic animal model system expressing a mutant loricrin showed a Vohwinkel syndrome-like phenotype (34). To acquire a causative relationship in signal transduction between loricrin mutations and LK, we developed a stable cell line expressing a mutant loricrin using the ecdysone-inducible system. To our knowledge, this study is the first to show that the expression of a mutant loricrin caused the activation of VEGFR 2 and EGFR in a cellular model of LK (VL-2 and VL-5 cells), where Akt is markedly phosphorylated. Akt kinase activity was nine times the control level. When the mutant was induced to express, the adapter proteins Gab-1 and c-Cbl were FIGURE 4. Effect of PI3K inhibitors on Akt activity and distribution of phospho-Akt in VL-5 cells. A, to determine whether PI3K is really required for VL-5 cells to proliferate, we used two PI3K inhibitors at various concentrations and conducted an immunoblot analysis. Wortmannin and LY294002 were added immediately after the treatment with 1 M muristerone A. We also conducted an Akt kinase activity assay. Akt activity is represented as the intensity of bands on the phospho-GSK-3␣/␤ (Ser-21/9) immunoblot. When wortmannin was used at 0.2 M, phospho-Akt (Thr-308) was not detected. When this concentration was increased to 1 M, neither phospho-Akt (Ser-473) nor phospho-Akt (Thr-308) was detected. When LY294002 was used at 50 M, neither phospho-Akt (Ser-473) nor phospho-Akt (Thr-308) was detected. B, density of immunostaining was measured in five experiments and quantified with NIH Image software. The level of Akt activity was nine times higher in VL-5 cells than the control cells (mock and WL-1 cells). C, representative double stainings with anti-phospho-Akt (Ser-473) antibody (green) and anti-V5 antibody (red) of WL-1 and VL-5 cells. 4Ј,6-Diamidino-2-phenylindole (DAPI) was used for nuclear detection.   both phosphorylated in VL-5 cells. Stat3 and ERK1/2 were both phosphorylated as well. VEGF and TGF-␣ levels were increased in LK model cells. VEGF inhibitors, CBO-P11 and JE-11, hindered LK model cell proliferation.
LK model cells showed significant phosphorylation of EGFR and VEGFR 2 at 20 h after induction. TGF-␣ along with epidermal growth factor and amphiregulin are ligands for the EGFR (35). EGFR and VEGFR 2 are membrane-bound receptor tyrosine kinases. On binding ligands, these receptors form homo-and heterodimers leading to the autophosphorylation of tyrosine residues in the cytosolic domains of the proteins. The phosphorylated tyrosine residues become docking sites for signaling molecules such as Gab1 and c-Cbl that activate cellular signaling pathways regulating a number of cellular processes, including proliferation and survival. We detected phosphorylated forms of Gab1 and c-Cbl only in LK model cells and observed increased TGF-␣ secretion and prolonged EGFR phosphorylation. VEGF secretion and prolonged VEGFR 2 phosphorylation were also observed. VEGF is a dimeric glycoprotein and a hypoxia-inducible endothelial cell mitogen (36). Although VEGF had been thought to be highly specific for endothelial cells, it has become increasingly clear that it also elicits responses in nonendothelial cell types such as keratinocytes. In normal human skin, VEGF is expressed and secreted by epidermal keratinocytes (37). Originally, the keratinocytederived VEGF was thought to act in a paracrine manner on vascular hyperpermeability and angiogenesis during wound healing. Yang et al. (31) found that HaCaT cells expressed all five known VEGF receptors and coreceptors. Most importantly, neutralizing VEGFR 2 could block the VEGF-induced proliferation and migration of HaCaT cells. In addition, Lichtenberger et al. (38) found that VEGF could be considered as a potent growth factor for epidermal tumors. They demonstrated that VEGFR signaling is cell-autonomously required in skin tumor cells to stimulate their proliferation in an autocrine-and angiogenesis-independent manner and that VEGFR and EGFR signaling synergize in neoplastic cells to promote tumor growth. We observed that simultaneous use of VEGF and EGFR inhibitors blocked VL-5 cell proliferation significantly. This blockade effect may be attributed to a synergic action of EGFR and VEGF signaling (38).
The differences in cell proliferation between control (mock and WL-1 cells) and cellular model of LK (VL-5 cells) was rather small, and one wonders whether it explained the clinically seen hyperproliferation of LK. Perhaps there may be a signaling loop from keratinocytes to dermal fibroblasts and back in vivo. One candidate of cytokines secreted by keratinocytes may be interleukin-1␤. Another candidate of growth factors secreted by fibroblasts may be keratinocyte growth factor. . ChIP assays and immunoblotting using anti-phospho-Stat3 antibody. A, ChIP assay was performed in VL-5 cells stimulated with 1 mM muristerone A for 20 h using either an anti-Stat3-specific antibody, an irrelevant antibody (anti-Rho), or no antibody. PCR primers were designed to yield a 130-bp product, which includes the Stat3-binding site (Ϫ848) of the VEGF promoter. As a negative control, a PCR using primers for the mouse albumin gene was included in these experiments. Input lanes represent 0.02% of total chromatin used in ChIP assays. Similar data were obtained in five experiments. B, mock, WL-1, and VL-5 cells were stimulated with 1 mM muristerone A for 20 h. The cells were lysed and processed for anti-phospho-Stat3 (Tyr-705) (top) and anti-phospho-Stat3 (Ser-727) (middle) immunoblotting (n ϭ 4).  It is well known that interleukin-1␤ is the most potent inducer of keratinocyte growth factor in fibroblasts through the c-Jun pathway (39). Such signaling loop may partly explain the hyperproliferation observed in LK.
VEGF stimulation drives the formation of a complex between Gab1 and PI3K, a heterodimer consisting of regulatory 85-kDa (p85) and catalytic 110-kDa (p110) subunits. This interaction takes place via the phosphorylated YMXM motifs of Gab1 and the Src homology region 2 domains of p85. We conducted immunoprecipitation experiments and found Gab1 to be associated with p85 and SHP-2 (Fig. 5). This together with the results of ⌬p85 transfection experiments (Fig. 3B) suggest PI3K to be necessary for TGF-␣ and VEGF signaling in VL-5 cells.
Akt is activated by PIP 3 phospholipids and is thus sensitive to inhibition by the PI3K inhibitors wortmannin and LY294002. The activation of Akt requires binding to PIP 3 via the pleckstrin homology domain, phosphorylation on the activation loop of Thr-308 by 3-phosphoinositide-dependent kinase-1 (PDK1), and also phosphorylation within the C terminus at Ser-473. We detected Akt phosphorylated at serine 473 and threonine 308 in VL-5 cells. When a dominant-negative form of p85 (⌬p85) was transfected into VL-5 cells, Akt kinase activity was suppressed. Our results thus provide evidence of the activation of a PIP 3and Akt-mediated pathway in the VL-5 cells. We also found that phospho-Akt colocalized with mutant loricrin. Although how is not clear at present, mutant loricrin might trigger the activation of PIP 3 and the Akt-mediated pathway.
Suzuki et al. (40) reported that a keratinocyte-specific deficiency of PTEN resulted in epidermal hyperplasia and hyperk-eratosis. PTEN is a lipid phosphatase whose major substrate is PIP 3 . They showed that the phosphorylation of Akt and MAPK (ERK1/2) was increased in PTEN-deficient keratinocytes (k5Pten flox/flox ) compared with k5Pten ϩ/ϩ or k5Pten flox/ϩ keratinocytes. We thus speculate that the activation of Akt and MAPK in LK model cells is related to the hyperkeratosis observed in LK.
Although transgenic mice expressing a mutant form of loricrin were created, the precise molecular mechanism inducing LK phenotypic changes in the epidermis has not been resolved as yet. There has been controversy over the function of mutant loricrin. Suga et al. (34) insisted that it conferred a gain of function to the system. However, Ishida-Yamamoto et al. (21) thought that mutant loricrin worked as a dominant-negative disrupter. Mutant loricrin gains nuclear localization signals and is transported to the nucleoli. We detected some phospho-Akt (Ser-473) colocalized with mutant loricrin at the nucleoli. The precise interaction of phospho-Akt (Ser-473) and mutant loricrin needs to be elucidated.
The blocking of VEGF is an attractive way to treat LK because LK model cells synthesize and secrete VEGF into the culture medium. We found that VEGF inhibitors, CBO-P11 and JE-11, plus PD 153035 inhibited the proliferation of VL-5 cells and attenuated the phosphorylation of VEGFR 2. CBO-P11 is also known to inhibit the VEGF-stimulated phosphorylation of MAPK (33). Thus, CBO-P11 plus PD 153035 or JE-11 plus PD 153035 are promising candidates for the development of new inhibitors useful for the treatment of LK.
In conclusion, this study highlights a novel role for the activation of VEGFR 2 and EGFR in the hereditary keratinizing disorder LK. The release of VEGF and TGF-␣ and subsequent activation of VEGFR 2 and EGFR by an autocrine/paracrine pathway link loricrin gene mutations to the rapid cell proliferation in the HaCaT LK cellular model. It would be of interest to examine the involvement of this mechanism in the pathogenesis of other hereditary keratinizing disorders.