The Role of Glycogen Synthase Kinase 3B in the Transformation of Epidermal Cells

Research Article

Cuiling Ma,1,3 Jian Wang,4 Ying Gao,3 Tian-Wen Gao,3 Gang Chen,1 Kimberly A. Bower,1 Mohammed Odetallah,1 Min Ding,1,2 Zunji Ke,5 and Jia Luo1,5

1Department of Microbiology, Immunology and Cell Biology, West Virginia University School of Medicine, Robert C. Byrd Health Sciences Center; 2National Institute for Occupational Safety and Health, Morgantown, West Virginia; Departments of 3Dermatology and 4Obstetrics and Gynecology, Xijing Hospital, Xian, P.R. China; and 5Institute for Nutritional Sciences, Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences, Shanghai, P.R. China

Abstract number of external signals (3). GSK3h activity is regulated by site- B B specific phosphorylation. Full activity of GSK3h generally requires Glycogen synthase kinase 3 (GSK3 ) is a multifunctional 216 serine/threonine kinase. We showed that the expression of phosphorylation on Tyr , and conversely, phosphorylation at 9 h h GSK3B was drastically down-regulated in human cutaneous Ser inhibits GSK3 activity. GSK3 is a negative regulator of Wnt/ h squamous cell carcinomas and basal cell carcinomas. Due -catenin signaling (1, 4). Because some oncogenic transcription to its negative regulation of many oncogenic proteins, we factors [e.g., activator protein-1 (AP-1)] and proto-oncoproteins h h hypothesized that GSK3B may function as a tumor suppressor (i.e., -catenin) are putative GSK3 substrates for phosphorylation- h during the neoplastic transformation of epidermal cells. We dependent inactivation (4), it has been hypothesized that GSK3 tested this hypothesis using an in vitro model system, JB6 may interfere with cellular neoplastic transformation and tumor mouse epidermal cells. In response to epidermal growth factor development. However, there are only limited studies examining h (EGF) or 12-O-tetradecanoylphorbol-13-acetate (TPA), the the involvement of GSK3 in tumor development; the findings h promotion-sensitive JB6 P+ cells initiate neoplastic transfor- are sometimes contradictory (5–8). The role of GSK3 during mation, whereas the promotion-resistant JB6 PÀ cells do not. tumorigenesis remains unclear. Carcinogenesis is a complex JB6 PÀ cells expressed much higher levels of GSK3B than JB6 process that can be divided experimentally into three stages, P+ cells; JB7 cells, the transformed derivatives of JB6, had the namely, initiation, promotion, and progression. Initiation is least amount of GSK3B. The activity of GSK3B is negatively associated with irreversible, carcinogen-mediated DNA mutation. regulated by its phosphorylation at Ser9. EGF and TPA induced In contrast, promotion is a reversible process in which there are strong Ser9 phoshorylation in JB6 P+ cells, but phosphoryla- increases in the rate of cell replication and/or alterations in gene tion was seen at a much lesser extent in JB6 PÀ cells. EGF expression. Progression represents the final genetic changes and TPA-stimulated Ser9 phosphorylation was mediated by associated with the conversion of benign tumors into fully phosphoinositide-3-kinase (PI3K)/Akt and protein kinase C malignant cells. Skin cancer is the most common cancer worldwide (PKC) pathways. Inhibition of GSK3B activation significantly (9). Our understanding of the mechanisms underlying the stimulated activator protein-1 (AP-1) activity. Overexpression development and progression of skin tumors is still fragmentary. of wild-type (WT) and S9A mutant GSK3B in JB6 P+ cells JB6 P+ mouse epidermal cells (Cl 41), originally derived from suppressed EGF and TPA-mediated anchorage-independent primary mouse epidermal cells, offer an excellent model to growth in soft agar and tumorigenicity in nude mice. Over- investigate the molecular events that are associated with tumor expression of a kinase-deficient (K85R) GSK3B, in contrast, promotion. These cells undergo a response analogous to second- potentiated anchorage-independent growth and drastically stage tumor promotion in mouse skin when treated with various enhanced in vivo tumorigenicity. Together, these results tumor promoters. For example, exposure of JB6 P+ cells to indicate that GSK3B plays an important role in skin epidermal growth factor (EGF) or 12-O-tetradecanoylphorbol-13- tumorigenesis. [Cancer Res 2007;67(16):7756–64] acetate (TPA) induces the phenotype of anchorage-independent growth and tumorigenicity in vivo (10–12). In contrast, JB6 PÀ cells Introduction are promotion resistant; EGF and TPA fail to initiate neoplastic transformation in these cells (10, 13). JB6 cells and their derivatives h h Glycogen synthase kinase 3 (GSK3 ) is a serine/threonine have been extensively used as an in vitro model for the promotion kinase that was first identified as a critical mediator in glycogen h of neoplastic transformation (10, 12–15). It has been shown that metabolism and insulin signaling. It is now known that GSK3 is three signaling pathways are involved in the transformation of JB6 a multifunctional kinase; more than 40 proteins are substrates of P+ cells, namely, phosphoinositide-3-kinase (PI3K)/Akt, PKC and GSK3h, including transcription factors, cell cycle/survival regu- mitogen-activated protein (MAP)/extracellular signal-regulated lators and oncogenic/proto-oncogenic proteins (1, 2). Unlike most kinase (ERK) kinase 1 (MEK1)/Erk (13–16). Activation of these protein kinases, GSK3h is constitutively active in resting cells signaling pathways by EGF or TPA results in AP-1 transactivation, and undergoes a rapid and transient inhibition in response to a which is essential for the transformation of JB6 P+ cells (13, 14, 16). In this study, we compared the expression of GSK3h in human nonmelanoma skin cancers (cutaneous squamous cell carcinomas Requests for reprints: Jia Luo, West Virginia University School of Medicine, P.O. and basal cell carcinomas) to normal skin tissues. Using the JB6 Box 9177, Morgantown, WV 26506. Phone: 304-293-7208; Fax: 304-293-7823; E-mail: cell system, we have investigated the role of GSK3h in neoplastic [email protected]. I2007 American Association for Cancer Research. transformation and delineated the signal pathways that regulate doi:10.1158/0008-5472.CAN-06-4665 GSK3h activity.

Cancer Res 2007; 67: (16). August 15, 2007 7756 www.aacrjournals.org

j Materials and Methods at 37 C with an atmosphere of 5% CO2 for 10 to 14 days, and the number of induced cell colonies was counted under a microscope. Colonies Materials and cell cultures. 2 All antibodies except antiactin antibody containing eight or more cells were counted in four 0.5-cm areas were obtained from Cell Signaling Technology, Inc.. Antiactin antibody was randomly chosen with respect to distance from the center of the well, and purchased from Santa Cruz Biotechnology. PKC inhibitors (GF10203X or the count was multiplied by the appropriate factor to give the colony h bisindolylmaleimide I and Go6976), PKA inhibitor (H89), GSK3 inhibitor number per well. (TDZD-8), and MEK1 inhibitor (PD98059) were purchased from Calbio- Tumorigenicity in nude mice. To evaluate in vivo tumorigenicity, h chem. PI3K inhibitors (LY294002 and wortmannin), GSK3 inhibitor 5-week-old male nude mice (BALB/c nu/nu, f25 g; Charles River (SB216763), LiCl, and lactacystin were purchased from Sigma Chemical Laboratories) were used. JB6 P+ cells and stable transfectants expressing Co.. c-Jun-NH2-kinase (JNK) inhibitor (D-JNKI) was purchased from Alexis various GSK3h constructs were treated with EGF or TPA (0 or 10 ng/mL) for Biochemicals. 6 days and then dissociated from monolayer cultures by trypsinization. À JB6 P+ mouse epidermal cell line (Cl 41), JB6 P and transformed JB7 Cells were counted and centrifuged at 1,500 rpm for 5 min and resuspended cells were grown in EMEM containing 10% fetal bovine serum (FBS), in PBS. An aliquot of cells (5 Â 106 in 100 AL of PBS) that were treated with A j 2 mmol/L L-glutamine, and 25 g/mL gentamicin at 37 C with 5% CO2. JB7 EGF or TPA (0 or 10 ng/mL) was directly injected to both flanks of the cells were derived from soft agar colonies of JB6 P+ treated with TPA for animals. One injection per flank was done for each mouse. Eight animals 3 weeks. These cells form colonies in soft agar and display tumorigenicity were used for each treatment group. Mice were maintained in a pathogen- in vivo . The stable transfectants of JB6 P+ cells expressing AP-1-luciferase free environment; food and water were given ad libitum. Seven weeks after reporter (Cl 41 AP-1) have been previously described (17, 18). the initial injections, the length (L) and width (W) of the s.c. tumor mass B Human skin samples and immunohistochemical study of GSK3 . were measured by calipers, and the tumor volume (TV) was calculated as Human skin tissues were obtained from surgical specimens at the described by Yaguchi et al. (20): TV = 0.5 Â L Â W2. At the end of the Department of Dermatology, Xijing Hospital (Xi’an, China). The protocol experiments, mice were sacrificed using a CO2 chamber. Animal housing for collecting human tissues was approved by the Ethical Committee of and all procedures followed the NIH Guide for the Care and Use of Xijing Hospital. The specimens were formalin fixed and paraffin embedded. Laboratory Animals and were approved by the West Virginia University The samples comprised of 31 primary cutaneous squamous cell carcinomas Animal Care and Use Committee. Every effort was made to reduce the and 12 basal cell carcinomas. The median patient age was 61 years. Tumor number of animals and their suffering. diagnoses were established through pathologic evaluation of paraffin- Measurement of AP-1 activity. AP-1transactivationinJB6P+ embedded tissues stained with H&E. None of the patients received radiation epidermal cells was determined by assaying the activity of the luciferase or chemotherapy before the operation. Eight samples of histologically reporter (17, 18). The assay accurately measures AP-1 transactivation normal adult skin tissues were collected as controls. (13, 18). Briefly, JB6 P+ cells expressing AP-1-luciferase reporter (Cl 41 AP-1) Immunohistochemistry for GSK3h was done by the avidin-biotin indirect were cultured in 96-well plates and grown in a medium containing 10% FBS. immunoperoxidase method. Briefly, 4-Am-thick sections were dewaxed, The plates were incubated at 37jC in a humidified atmosphere of 5% rehydrated, and incubated with 0.3% hydrogen peroxide for 30 min to CO2. Subconfluent cultures were maintained in a medium containing block endogenous peroxidase activity. Sections were microwave treated in 0.1% FBS for 24 h and treated with or without various protein kinase 0.01 mol/L citrate buffer (pH, 6.0) at 700 W for 10 min and rinsed with inhibitors 30 min before exposure to EGF or TPA. After treatment, cellular 0.01 mol/L TBS. Sections were incubated with normal horse serum for protein was extracted with a 1Â lysis buffer supplied in the luciferase assay 20 min and then with primary antibodies (dilution 1:100) overnight at 4jC. kit (Promega), and luciferase activity was measured with a monolight After rinsing in TBS, sections were incubated with biotinylated secondary luminometer (3010, Analytical Luminescence Laboratory). AP-1 activity antibodies at room temperature for 30 min, followed by an avidin-biotin- (luciferase activity) was calculated and expressed relative to the untreated peroxidase complex (Fisher Scientific) for 30 min. The reaction was cultures. visualized with 3,3¶-diaminobenzidine tetrahydrochloride (0.5 mg/mL, Statistical analysis. Differences among treatment groups were tested Sigma Chemical Co.) supplemented with 0.01% hydrogen peroxide. Sections using an ANOVA. Differences in which P was < 0.05 were considered were counterstained with Harris hematoxylin. statistically significant. In cases where significant differences were detected, Establishing stable transfectants. Stable transfectants expressing specific post hoc comparisons between treatment groups were examined various GSK3h constructs were established as previously described (19). with Student-Newman-Keuls tests. V5-tagged GSK3h constructs (wild-type, S9A, and K85R) carried by vector pcDNA3 were generous gifts from Dr. Thilo Hagen (University Hospital Nottingham). Cell transfection was carried out with LipofectAMINE 2000 Results reagent (Invitrogen) according to the manufacturer’s instructions. Stable Expression of GSK3B inhumanskintissues.We first cell clones expressing exogenous GSK3h were screened by the treatment of examined the expression of GSK3h and its phsophorylated forms G418 (600 Ag/mL) for 3 to 4 weeks. Positive clones were verified by the [pGSK3h (Ser9 and Tyr216)] in histologically normal skin specimens. expression of V5 as well as the evidence of GSK3h overexpression. The clones expressing the highest levels of exogenous GSK3h were selected for In all eight normal samples examined, a strong expression of h h 9 subsequent experiments. GSK3 and pGSK3 (Ser ) was observed in the keratinocytes 216 Immunoblotting. The immunoblotting procedure to detect phosphor- (Fig. 1A); the immunostaining of pGSK3h (Tyr ) was either very 9 ylation and expression of signal proteins was done as previously described weak or negative. GSK3h and pGSK3h (Ser ) were also expressed (17). To control for the loading, the blots were stripped and reprobed with in keratinocytes of patients with cutaneous squamous cell an antiactin antibody (Santa Cruz Biotechnology). In some cases, the carcinomas or basal cell carcinomas; however, the immunostaining density of immunoblotting was quantified with the software of Quantity of GSK3h and pGSK3h (Ser9) in patients with skin carcinomas One (Bio-Rad Laboratories). was generally weaker than that of age- and sex-matched normal Anchorage-independent growth. Anchorage-independent growth of subjects (Fig. 1B). In all tumor specimens (31 cases of primary JB6 P+, JB7 cells, and stable transfectants expressing various GSK3h cutaneous squamous cell carcinomas and 12 cases of basal cell constructs was determined by a previously described method (17). The cell growth matrix consists of two layers of basal medium Eagle (BME) agar in carcinomas), cutaneous squamous cell carcinomas and basal cell h h 9 six-well culture trays. The base layer (2 mL) contained 10% FBS and 0.5% carcinomas expressed much less GSK3 and pGSK3 (Ser ) than BME agar. The top layer (0.5 mL) contained 10% FBS, 0.33% BME agar, and adjacent nontumor-bearing keratinocytes (Fig. 1B). The expression 216 the suspension of cells (0.5 Â 104). EGF (10 ng/mL) or TPA (10 ng/mL) of pGSK3h (Tyr ) in the skin specimens of cancer patients was were applied in both top and bottom layers. The cultures were maintained consistently negative (data not shown). www.aacrjournals.org 7757 Cancer Res 2007; 67: (16). August 15, 2007

Figure 1. Immunohistochemical study of GSK3h expression in normal human skin tissues and tumor-bearing tissues. A, the expression of GSK3h and phosphorylated forms (Ser9 and Tyr216) was examined in paraffin-embedded normal human skin tissues as described in Materials and Methods. a, expression of GSK3h. b, expression of phosphorylated GSK3h (Ser9) [pGSK3h (Ser9)]. c, expression of pGSK3h (Tyr216). d, human colon cancer tissues were used as a positive control for immunostaining of pGSK3h (Tyr216). B, a representative microphotograph shows GSK3h and pGSK3h (Ser9) immunostaining in tumor-bearing skin tissues and histologically normal skin tissues obtained from age- and sex-matched subjects. SCC, squamous cell carcinomas; BCC, basal cell carcinomas. Arrows, normal keratinocytes; arrowheads, tumor tissues. Bar, 50 Am.

Expression of GSK3B in JB7, JB6 P+ and PÀ cells. JB6 cells JB6 P+ cells, we have established JB6 cells stably expressing wild- and their derivatives offer an excellent system to study the role of type (WT), S9A, or K85RGSK3 h. S9A mutant is unable to be GSK3h in the transformation of epidermal cells. Because GSK3h phosphorylated at Ser9 and, therefore, resistant to inhibitory was apparently down-regulated in skin carcinomas, we sought to regulation; K85Rmutant is kinase deficient and functions as a compare the expression of GSK3h among tumor promotion- dominant negative protein (6, 22). Overexpression of these resistant JB6 PÀ cells, promotion-sensitive JB6 P+ cells, and exogenous GSK3h proteins was verified by immunoblotting using transformed JB7 cells. Among these cells, JB7 had the least amount either anti-GSK3h or V5 antibody (Fig. 2C). The levels of cyclin of GSK3h, whereas JB6 PÀ cells expressed the highest levels of D1 were correlated with the status of GSK3h expression. JB6 GSK3h; the levels of GSK3h in JB6 P+ were intermediate (Fig. 2A). P+ cells overexpressing S9A expressed the least amount of cyclin Cyclin D1 is a substrate of GSK3h; GSK3h may regulate the D1, whereas cells overexpressing K85Rhad the highest expression expression of cyclin D1 by transcriptional activation or controlling (Fig. 2C). Cells overexpressing S9A were less sensitive to EGF- its degradation (21). The expression of cyclin D1 in these cells was induced pGSK3h (Ser9) and up-regulation of cyclin D1 (Fig. 2D), inversely correlated to the levels of GSK3h (Fig. 2A). Two potent indicating that these cells were more resistant to the negative tumor promoters for JB6 cells, EGF and TPA, induced pGSK3h regulation of GSK3h. In contrast, cells overexpressing K85Rwere (Ser9;Fig.2B). EGF produced a rapid (5 min) induction of pGSK3h more sensitive to EGF-induced up-regulation of cyclin D1 (Fig. 2D), (Ser9). TPA-mediated pGSK3h (Ser9) was slower but more verifying a dominant negative role of K85Rmutant. GSK3 h may sustained. EGF and TPA had little effect on pGSK3h (Tyr216; data regulate the expression of cyclin D1 by transcriptional activation not shown). Although JB6 PÀ cells expressed high levels of GSK3h or controlling its degradation (21). Lactacystin, a proteasome (Fig. 2A), they were less sensitive to the induction of EGF and TPA inhibitor, failed to reverse S9A-induced down-regulation of of pGSK3h (Ser9;Fig.2B), indicating that they were more resistant cyclin D1, whereas it enhanced K85R-mediated up-regulation to negative regulation. of cyclin D1 (Fig. 2E). The results suggested that GSK3h regulated GSK3B mediates the transformation of JB6 P+ cells. To cyclin D1 expression in JB6 cells primarily through transcriptional determine whether GSK3h is involved in the transformation of control.

Cancer Res 2007; 67: (16). August 15, 2007 7758 www.aacrjournals.org

The effect of GSK3h on the transformation of JB6 P+ cells was received one injection. There were eight animals for each first determined by anchorage-independent growth in soft agar. injection group. The number and the size of s.c. tumors were As shown in Fig. 3A, EGF and TPA stimulated the formation of measured 7 weeks following initial injection. EGF- or TPA- cell colonies in soft agar. EGF- and TPA-induced cell colonies exposed JB6 P+ cells expressing control vectors formed s.c. were significantly suppressed by the overexpression of WT and tumors in nude mice (Fig. 3C). For example, eight animals S9A GSK3h. In contrast, EGF- and TPA-induced anchorage- received injections of EGF-exposed JB6 P+ cells (each animals independent growth was drastically enhanced by the over- received two injections; one in each flank), a total of 11 tumors expression of K85Rmutant. It was noted that JB6 P+ cells formed, and the average volume of each tumor was 1157 F 105 expressing K85Rformed some small cell colonies (contained less mm3 (Table 1). However, in the eight animals injected with EGF- than eight cells) in the absence of EGF and TPA. These small cell exposed cells expressing WT or S9A GSK3h, only three and two colonies were not scored. The results suggested that GSK3h was a tumors formed, respectively. In addition, the average volume of negative regulator of cell transformation. JB7 cells formed tumors was significantly smaller. The tumor volumes in animals colonies in soft agar (Fig. 3B). Two inhibitors of GSK3h (LiCl injected with cells expressing WT or S9A GSK3h were 632 F 79 and SB216763) significantly enhanced anchorage-independent and 478 F 81 mm3, respectively. In contrast, EGF-exposed cells growth of JB7 cells. In contrast, overexpression of GSK3h in JB7 overexpressing K85Rshowed enhanced tumorigenicity in nude cells suppressed anchorage-independent growth (Fig. 3B). To mice; a total of 14 tumors formed, and the average volume of further assess the role of GSK3h in tumorigenicity, we injected each tumor was 3756 F 279 mm3 (Table 1). Similar results were EGF- or TPA-exposed JB6 P+ cells expressing WT, S9A, or K85R obtained with the treatment of TPA (Table 1). It was noted that GSK3h to nude mice and evaluated the formation of s.c. tumors. three small tumors (368 F 65 mm3) formed in the nude mice that Cells were injected to both flanks for each mouse, and each flank received an injection of cells expressing K85Rthat was not

Figure 2. Expression of GSK3h in JB6 cells and their derivatives and establishment of cells expressing ectopic GSK3h mutants. A, The expression of GSK3h and cyclin D1 in transformation-sensitive JB6 P+ cells, transformation-resistant JB6 PÀ cells, and transformed JB7 cells was examined with immunoblotting. B, JB6 P+ and PÀ cells cultured in serum-free media were treated with EGF (10 ng/mL) and TPA (10 ng/mL) for a specified period. The expression of phosphorylated GSK3h (Ser9) was examined with immunoblotting. C, establishment of JB6 cells stably expressing wild-type, S9A, and K85R GSK3h was carried out as described in Materials and Methods. The expression of V5, GSK3h, and cyclin D1 in these cells was examined with immunoblotting. D, JB6 cells expressing various GSK3h mutants were treated with EGF (10 ng/mL) for a specified period. The expression of phosphorylated GSK3h (Ser9) and cyclin D1 was examined with immunoblotting. E, JB6 cells expressing various GSK3h mutants were treated with lactacystin (Lac, 0 or 10 Amol/L) for 1 h. The expression of cyclin D1 was examined with immunoblotting. www.aacrjournals.org 7759 Cancer Res 2007; 67: (16). August 15, 2007

Figure 3. Role of GSK3h in the transformation of JB6 P+ cells in vitro and in vivo. A, anchorage-independent growth of JB6 P+ cells expressing various GSK3h mutants. JB6 P+ cells stably expressing wild-type, S9A, and K85R GSK3h, which were grown in a matrix of soft agar, were exposed to EGF (0 or 10 ng/mL) or TPA (0 or 10 ng/mL). Cell colonies were scored after 14 d of incubation at 37jC in an atmosphere of 5% CO2 as described in Materials and Methods. The number of soft agar colonies/104 cells in the untreated control was arbitrarily designated as 1. The numbers of colonies in experimental groups were expressed as an arbitrary unit relative to the untreated control group. The experiment was replicated four times. *, P < 0.05 denotes a statistically significant difference from untreated controls. #, P < 0.05 denotes a statistically significant difference from EGF- or TPA-treated JB6 cells expressing the control vector. B, role of GSK3h in anchorage-independent growth of JB7 cells. The effect of GSK3h inhibitors (LiCl, 20 mmol/L; and SB216763, 10 Amol/L) on anchorage-independent growth of JB7 cells or JB7 cells overexpressing wild-type GSK3h (JB7/GSK3) was evaluated as described above. *, P < 0.05, statistically significant difference from untreated JB6 P+ and JB7/GSK3 cells. #, P < 0.05, statistically significant difference from JB7 cells. C, influence of GSK3h on tumorigenesis in nude mice. Nude mice were s.c. inoculated with JB6 P+ cells and their derivatives stably expressing wild-type, S9A, and K85R GSK3h constructs. The cells were treated with EGF or TPA (0 or 10 ng/mL) for 6 d. For each experimental group, there were eight animals. Seven weeks following inoculation, the number of s.c. tumor masses in each animal was scored, and the volume of tumors was measured by calipers as described in Materials and Methods. A representative photo shows s.c. tumors induced by EGF treatment in nude mice. 1, JB6; 2, JB6 + EGF; 3, S9A; 4, S9A + EGF; 5, K85R; 6, K85R + EGF. TPA-induced tumors were not shown. Arrows, s.c. tumors. exposed to EGF or TPA. Thus, the results obtained from the nude Both EGF and TPA induced phosphorylation of Akt and PKC mice agreed with anchorage-independent growth. (Fig. 4A). The profiles of EGF- and TPA-mediated phosphorylation, PKC and PI3K/Akt mediate EGF- and TPA-induced inacti- however, were different. EGF-induced Akt phosphorylation was vation of GSK3B. Previous studies show that EGF and TPA rapid (5 min) and strong, whereas TPA-mediated Akt phosphor- activated PI3K/Akt, PKC, and MEK1/Erk pathways in JB6 P+ cells ylation was modest and more persistent. As detected by an (13, 14, 23). These pathways are potential upstream components of antibody recognizing phosphorylated panPKC, EGF-induced phos- GSK3h. We sought to determine whether EGF- and TPA-induced phorylation of panPKC was weaker than that induced by TPA. inactivation of GSK3h was mediated by these signaling pathways. The profiles of EGF- and TPA-induced pGSK3h (Ser9) were also

Table 1. Name and volume of tumor (mm3) in nude mice

Control, n (volume, mm3) EGF, n (volume, mm3) TPA, n (volume, mm3)

JB6 0 (0) 11 (1,157 F 105) 11 (902 F 112) WT 0 (0) 3 (632 F 79)* 4 (424 F67)* S9A 0 (0) 2 (478 F 81)* 2 (358 F 56)* K85R3 (368 F 65)* 14 (3,756 F 279)* 15 (2,371 F 199)*

NOTE: The number and the mean volume of subcutaneous tumors were calculated. *P < 0.05, denotes a statistically significant difference from parental JB6 P+ cells treated with EGF or TPA.

Cancer Res 2007; 67: (16). August 15, 2007 7760 www.aacrjournals.org

Downloaded from cancerres.aacrjournals.org on September 29, 2021. © 2007 American Association for Cancer Research. GSK3b Regulates the Transformation of Epidermal Cells different. EGF induced a rapid and strong phosphorylation of inhibitor in down-regulating TPA-induced pGSK3h (Ser9). The GSK3h (Ser9), the maximal phosphorylation occurred between inhibitors for MEK1, JNK, and PKA had little effect on EGF- and 5 and 15 min after EGF treatment. TPA-mediated phosphorylation TPA-induced pGSK3h (Ser9;Fig.4B and C). These results indicated of GSK3h (Ser9) was gradual; it became evident at 15 min, and that PI3K/Akt and PKC mediated EGF- and TPA-induced pGSK3h maximal phosphorylation occurred at 3 to 6 h. Inhibitors of PI3K (Ser9), although the profiles of regulation were different. blocked EGF-stimulated pGSK3h (Ser9) in JB6 P+ cells; an inhibitor Like PKC and PI3K/Akt, ERKs are critical regulators of the of PKC also decreased EGF-stimulated pGSK3h (Ser9), but to a transformation of JB6 cells (13). MEK1 inhibitor (PD98059) did not lesser extent (Fig. 4B and C), suggesting that the PI3K/Akt pathway affect EGF- and TPA-mediated pGSK3h (Ser9;Fig.4B and C), played a major role in EGF regulation of pGSK3h (Ser9). On the suggesting that ERK was not involved in GSK3h inactivation in other hand, inhibitor of PKC was more effective than the PI3K JB6 cells. Reversely, GSK3h may inhibit ERK activation (24). We

Figure 4. Signaling pathways that regulate EGF- and TPA-induced pGSK3h (Ser9). A, JB6 P+ cells, cultured in serum-free media, were treated with EGF (10 ng/mL) and TPA (10 ng/mL) for a specified period. The expression of phosphorylated Akt, PKC, and GSK3h (Ser9) was examined with immunoblotting. The phosphorylated Akt was detected with an antibody recognizing pAkt (Thr308); the phosphorylated PKC was detected with an antibody recognizing p-panPKC. B, JB6 P+ cells were pretreated with inhibitors for 30 min and exposed to EGF or TPA for a specified period. The expression of pGSK3h (Ser9) was examined with immunoblotting. PD, PD98059, 50 Amol/L, MEK1 inhibitor; JNKi, 1 Amol/L, inhibitor for JNK; H89, 10 Amol/L, inhibitor for PKA; Bis, bisindolylmaleimide I, 1 Amol/L, inhibitor for PKC; LY, LY294002, 10 Amol/L, inhibitor for PI3K. C, the relative amounts of pGSK3h (Ser9) were measured microdensitometrically. The experiment was replicated thrice. D, effect of EGF on the phosphorylation of ERK in JB6 P+ cells expressing various GSK3h mutants. JB6 P+ cells stably expressing wild-type, S9A, and K85R GSK3h, which were grown in serum-free media, were exposed to EGF (0 or 10 ng/mL) for a specified period. The expression of phosphorylated ERK was determined by immunoblotting. www.aacrjournals.org 7761 Cancer Res 2007; 67: (16). August 15, 2007

Figure 5. Role of GSK3h in AP-1 transactivation. A, JB6 P+ epidermal cells stably expressing AP-1 luciferase reporter were pretreated with two GSK3h inhibitors, LiCl (20 mmol/L) and TDZD8 (10 Amol/L) or PI3K inhibitor (LY,10Amol/L) and PKC inhibitor (Bis,1Amol/L) for 30 min and then exposed to EGF (10 ng/mL) or TPA (10 ng/mL) for 12h. The activity of AP-1 was measured by a luciferase assay as described in Materials and Methods. The activity of AP-1 was expressed relative to untreated cultures. The experiment was replicated thrice. *, P < 0.05, statistically significant difference from untreated JB6 P+ cells. #, P < 0.05, statistically significant difference from paired EGF- or TPA-treated JB6 cells. B, JB6 P+ epidermal cells stably expressing AP-1 luciferase reporter were transfected with either wild-type, S9A-mutated GSK3h constructs or an siRNA for GSK3h for 48 h. The activity of AP-1 was measured as described above. The experiment was replicated thrice. *, P < 0.05, statistically significant difference from cells transfected with an empty vector. therefore sought to determine whether GSK3h affected ERK subjects. The decreased immunostaining for pGSK3h (Ser9) likely activation in JB6 P+ cells. As shown in Fig. 4D, EGF elicited similar results from the down-regulation of total GSK3h expression. Mouse phosphorylation of ERK in cells expressing WT, S9A, and K85R epidermal (JB6) cells have been extensively used as an in vitro GSK3h; furthermore, two inhibitors of GSK3h (LiCl and TDZD8) model for studying the promotion of neoplastic transformation failed to affect ERK phosphorylation (data not shown), indicating (10, 12–15). Consistent with the observations in human skin that GSK3h was not involved in ERK activation. tissues, the expression levels of GSK3h in JB6 cells are correlated to GSK3B is involved in the regulation of AP-1. It has been the stage or potential of cell transformation. JB7 cells, the shown that AP-1 activity is essential for the transformation of JB6 transformed derivatives of JB6, have the least amount of GSK3h, P+ cells (12, 13, 23). We sought to determine whether GSK3h whereas promotion-resistant JB6 PÀ cells express the highest levels regulates the activation of AP-1. As shown in Fig. 5A, two inhibitors of GSK3h; the levels of GSK3h in promotion-sensitive JB6 P+ cells of GSK3h (LiCl and TDZD8) stimulated the basal as well as EGF- are intermediate. In addition, tumor promoters EGF and TPA and TPA-induced activation of AP-1. In contrast, inhibitors for PI3K induce strong phosphorylation of GSK3h at Ser9 in JB6 P+ cells, and PKC decreased EGF- and TPA-induced activation of AP-1. indicating an inactivation of GSK3h. On the other hand, JB6 Furthermore, we showed that overexpression of wild-type and S9A PÀ cells are much less responsive to the negative regulation of GSK3h significantly inhibited AP-1 activity; in contrast, down- GSK3h. The involvement of GSK3h in skin tumorigenesis is further regulation of GSK3h by small interfering RNA (siRNA) enhanced shown by the modulation of GSK3h activity in JB6 P+ cells. AP-1 activity (Fig. 5B). Together, these results indicated that GSK3h Overexpression of WT and S9A mutant GSK3h inhibits EGF- and was a negative regulator of AP-1. TPA-mediated anchorage-independent growth in soft agar, as well as in vivo tumorigenicity in nude mice. In contrast, overexpression of a kinase-deficient K85RGSK3 h drastically potentiates EGF- and Discussion TPA-mediated anchorage-independent growth and greatly enhan- GSK3h is implicated in many biological processes, including ces in vivo tumorigenicity in response to EGF. Overexpression of a embryonic development, cell differentiation, and cell survival/cell kinase-deficient K85RGSK3 h also has a modest promoting effect cycle control (4, 25). Because many oncogenic transcription factors on anchorage-independent growth in soft agar and in vivo (e.g., c-Jun, c-Myc) and proto-oncoproteins (i.e., h-catenin) are tumorigenicity of JB6 P+ cells in the absence of EGF. These results substrates of GSK3h for phosphorylation-dependent inactivation verify that GSK3h negatively regulates epidermal cell transforma- (4), GSK3h may play a role in oncogenesis. However, information tion, and modulating GSK3h activity is sufficient to affect cell regarding the involvement of GSK3h in tumorigenesis is limited, transformation. and the function of GSK3h during cell transformation and cancer We further show that GSK3h is a negative regulator of AP-1 progression remains unclear. transactivation in JB6 P+ cells; the inhibitors of GSK3h increase Here, we show a decreased expression of GSK3h and pGSK3h basal as well as EGF- and TPA-mediated AP-1 transactivation. This (Ser9) in human nonmelanoma skin cancers (cutaneous squamous result is consistent with previous findings using other cells, which cell carcinomas and basal cell carcinomas) compared with adjacent show that GSK3h activation results in the inhibition of AP-1 normal keratinocytes. In addition, the immunostaining for GSK3h activity (26, 27). AP-1 is a heterodimeric transcription factor and pGSK3h (Ser9) in keratinocytes of patients with cutaneous complex composed of a Jun family member and a FOS family squamous cell carcinomas or basal cell carcinomas is generally member that binds the TRE DNA sequence (5¶-TGAGTCA-3¶). It weaker than keratinocytes of age- and sex-matched normal is involved in a variety of cellular processes, including growth,

Cancer Res 2007; 67: (16). August 15, 2007 7762 www.aacrjournals.org

addition to negatively regulating AP-1 activity, GSK3h is a well- known inhibitor for Wnt/h-catenin signaling (1). Hyperactivity of the Wnt/h-catenin signaling pathway is associated with the development of a number of tumors as well as malignancies (35–39). In terms of skin tumors, Wnt signaling is mainly implicated in the development and progression of melanomas (40). From the present study, however, it is not clear whether Wnt/ h-catenin signaling is also involved in GSK3h-induced suppression of skin tumorigenesis. Other studies also support a role of GSK3h as a ‘‘tumor suppressor.’’ For example, expression of a kinase-inactive GSK3h in adult mouse mammary glands promotes mammary tumorigenesis, indicating that antagonism of GSK3h activity is oncogenic for mammary epithelial cells (6). In a mouse epidermal multistage Figure 6. Diagram of the role of GSK3h in skin tumorigenesis. EGF and TPA 9 carcinogenesis model, Leis et al. (41) show a dramatic increase in activate PI3K/Akt and PKC, which induce phosphorylation of GSK3h (Ser ) and h 9 inactivate GSK3h. Inactivation of GSK3h stimulates oncogenic transcription pGSK3 (Ser ) in late papillomas and squamous cell carcinomas. 216 factors, such as AP-1. EGF and TPA can also activate AP-1 through the Furthermore, a significant decrease in pGSK3h (Tyr ) is observed MEK1/Erk pathway. Hyperactivity of AP-1 promotes skin tumorigenesis. in squamous cell carcinoma samples (41), indicating an inactivation of GSK3h during mouse skin carcinogenesis. Together, these h apoptosis, and differentiation (28). The mouse epidermal multi- observations support the notion that GSK3 is a negative regulator h stage carcinogenesis model provides a well-defined system for of skin tumorigenesis; down-regulation or inactivation of GSK3 is examining the transformation of squamous epithelial cells to oncogenic for epidermal cells. benign squamous papillomas and their subsequent progression The mechanisms underlying skin tumorigenesis are complex and into squamous cell carcinomas (29). With this carcinogenesis involve interactions among multiple signal cascades and various h model system, a number of studies show that AP-1 activity is transcription factors. Our study clearly shows that GSK3 is an h required for skin tumorigenesis as well as malignant transforma- important component in the cascades, and modulation of GSK3 tion (30–33). These findings are confirmed by in vitro studies that expression/activity is sufficient to alter the transformation h show that AP-1 activity is essential for the transformation of JB6 potential of epidermal cells. Thus, GSK3 is a target for developing h P+ cells (12, 13, 23). Three signaling pathways, namely, PI3K/Akt, prevention/intervention strategies. GSK3 has emerged as an PKC, and MEK1/Erk, have been shown to regulate EGF- and TPA- attractive therapeutic target for the treatment of multiple induced AP-1 transactivation and transformation of JB6 P+ cells neurologic diseases such as Alzheimer’s and stroke. Lithium has (13, 14, 23). We show here that PI3K/Akt and PKC are upstream of been used as a mood stabilizer to treat bipolar mood disorder, and 9 h GSK3h, and both EGF- and TPA-stimulated pGSK3h (Ser )is other inhibitors of GSK3 have entered clinical trials for diabetes. mediated by PI3K/Akt and PKC. Thus, suppressing GSK3h activity The potential role of these inhibitors in tumorigenesis should be is one of the mechanisms for EGF- and TPA-induced AP-1 considered. transactivation (Fig. 6). The interaction between Erk and GSK3h has been previously reported (24, 34); however, we fail to find such interaction in JB6 cells. Based on these findings, we propose a Acknowledgments signal cascade in which GSK3h plays an important role in Received 12/19/2006; revised 4/13/2007; accepted 5/14/2007. regulating AP-1 activity and cell transformation (Fig. 6): EGF and Grant support: NIH (AA015407), the National Natural Science Foundation of China (30470544, 30471452, and 30570580), and the Scientific Research Foundation for TPA activate PI3K/Akt, PKC, and MEK1/Erk pathways. Activation the Returned Overseas Chinese Scholars that is sponsored by the State Education of PI3K/Akt and PKC results in the inhibition of GSK3h, which is Ministry. Dr. Z.-J. Ke was also supported by the One Hundred Talents Program of the a negative regulator of AP-1. In the meantime, EGF- and TPA- Chinese Academy of Sciences. The costs of publication of this article were defrayed in part by the payment of page mediated ERK activation can also cause AP-1 transactivation. charges. This article must therefore be hereby marked advertisement in accordance Hyperactivity of AP-1 initiates the transformation process. In with 18 U.S.C. Section 1734 solely to indicate this fact.

References Wnt signaling and mammary tumorigenesis. Cancer Res 11. Colburn NH, Former BF, Nelson KA, Yuspa SH. 2005;65:5792–801. Tumour promoter induces anchorage independence 1. Doble BW, Woodgett JR. GSK-3: tricks of the trade for a 7. Ougolkov AV, Fernandez-Zapico ME, Savoy DN, irreversibly. Nature 1979;281:589–91. multi-tasking kinase. J Cell Sci 2003;116:1175–86. Urrutia RA, Billadeau DD. Glycogen synthase kinase- 12. Dong Z, Crawford HC, Lavrovsky V, et al. A dominant 2. Jope RS, Johnson GV. The glamour and gloom of 3h participates in nuclear factor nB-mediated gene negative mutant of jun blocking 12-O-tetradecanoyl- glycogen synthase kinase-3. Trends Biochem Sci 2004;29: transcription and cell survival in pancreatic cancer cells. phorbol-13-acetate–induced invasion in mouse kerati- 95–102. Cancer Res 2005;65:2076–81. nocytes. Mol Carcinog 1997;19:204–12. 3. Frame S, Cohen P. GSK3 takes centre stage more than 8. Shakoori A, Ougolkov A, Yu ZW, et al. Deregulated 13. Huang C, Ma WY, Young MR, Colburn N, Dong Z. 20 years after its discovery. Biochem J 2001;359:1–16. GSK3h activity in colorectal cancer: its association with Shortage of mitogen-activated protein kinase is respon- 4. Manoukian AS, Woodgett JR. Role of glycogen synthase tumor cell survival and proliferation. Biochem Biophys sible for resistance to AP-1 transactivation and trans- kinase-3 in cancer: regulation by Wnts and other Res Commun 2005;334:1365–73. formation in mouse JB6 cells. Proc Natl Acad Sci U S A signaling pathways. Adv Cancer Res 2002;84:203–29. 9. Pons M, Quintanilla M. Molecular biology of malig- 1998;95:156–61. 5. Cao Q, Lu X, Feng YJ. Glycogen synthase kinase-3h nant melanoma and other cutaneous tumors. Clin 14. Huang C, Schmid PC, Ma WY, Schmid HH, Dong Z. positively regulates the proliferation of human ovarian Transl Oncol 2006;8:466–74. Phosphatidylinositol-3 kinase is necessary for 12-O- cancer cells. Cell Res 2006;16:671–7. 10. Bernstein LR, Colburn NH. AP1/jun function is tetradecanoylphorbol-13-acetate–induced cell transfor- 6. Farago M, Dominguez I, Landesman-Bollag E, et al. differentially induced in promotion-sensitive and resis- mation and activated protein 1 activation. J Biol Chem Kinase-inactive glycogen synthase kinase 3h promotes tant JB6 cells. Science 1989;244:566–9. 1997;272:4187–94. www.aacrjournals.org 7763 Cancer Res 2007; 67: (16). August 15, 2007

15. Nomura M, Ichimatsu D, Moritani S, et al. Inhibition 23. Dong Z, Huang C, Ma WY. PI-3 kinase in signal blocks okadaic acid-induced skin tumor promotion. of epidermal growth factor-induced cell transformation transduction, cell transformation, and as a target for Cancer Res 2002;62:3044–7. and Akt activation by caffeine. Mol Carcinog 2005;44: chemoprevention of cancer. Anticancer Res 1999;19: 32. Young MR, Li JJ, Rincon M, et al. Transgenic mice 67–76. 3743–7. demonstrate AP-1 (activator protein-1) transactivation 16. Dhar A, Young MR, Colburn NH. The role of AP-1, 24. Wang Q, Zhou Y, Wang X, Evers BM. Glycogen is required for tumor promotion. Proc Natl Acad Sci NF-nB and ROS/NOS in skin carcinogenesis: the JB6 synthase kinase-3 is a negative regulator of extra- U S A 1999;96:9827–32. model is predictive. Mol Cell Biochem 2002;234–235: cellular signal-regulated kinase. Oncogene 2006;25: 33. Zenz R, Scheuch H, Martin P, et al. c-Jun regulates 185–93. 43–50. eyelid closure and skin tumor development through 17. Ma C, Wang J, Luo J. Exposure to asphalt 25. Mazor M, Kawano Y, Zhu H, Waxman J, Kypta RM. EGFRsignaling. Dev Cell 2003;4:879–89. fumes activates activator protein-1 through the Inhibition of glycogen synthase kinase-3 represses 34. Ding Q, Xia W, Liu JC, et al. Erk associates with and phosphatidylinositol 3-kinase/Akt signaling pathway androgen receptor activity and prostate cancer cell primes GSK-3h for its inactivation resulting in upregu- in mouse epidermal cells. J Biol Chem 2003;278: growth. Oncogene 2004;23:7882–92. lation of h-catenin. Mol Cell 2005;19:159–70. 44265–72. 26. Nikolakaki E, Coffer PJ, Hemelsoet R, Woodgett JR, 35. Behrens J, Lustig B. The Wnt connection to 18. Ma C, Wang J, Luo J. Activation of nuclear factor nB Defize LH. Glycogen synthase kinase 3 phosphorylates tumorigenesis. Int J Dev Biol 2004;48:477–87. by diesel exhaust particles in mouse epidermal cells Jun family members in vitro and negatively regulates 36. Dihlmann S, von Knebel Doeberitz M. Wnt/h- through phosphatidylinositol 3-kinase/Akt signaling their transactivating potential in intact cells. Oncogene catenin–pathway as a molecular target for future anti- pathway. Biochem Pharmacol 2004;67:1975–83. 1993;8:833–40. cancer therapeutics. Int J Cancer 2005;113:515–24. 19. Chen G, Ma C, Bower KA, Ke Z, Luo J. Interaction 27. Troussard AA, Tan C, Yoganathan TN, Dedhar S. Cell- 37. Ilyas M. Wnt signalling and the mechanistic basis of between RAX and PKR modulates the effect of ethanol extracellular matrix interactions stimulate the AP-1 tumour development. J Pathol 2005;205:130–44. on protein synthesis and survival of neurons. J Biol transcription factor in an integrin-linked kinase- and 38. Karim R, Tse G, Putti T, Scolyer R, Lee S. The Chem 2006;281:15909–15. glycogen synthase kinase 3-dependent manner. Mol Cell significance of the Wnt pathway in the pathology of 20. Yaguchi S, Fukui Y, Koshimizu I, et al. Antitumor Biol 1999;19:7420–7. human cancers. Pathology 2004;36:120–8. activity of ZSTK474, a new phosphatidylinositol 3-kinase 28. Eferl R, Wagner EF. AP-1: a double-edged sword in 39. Luu HH, Zhang R, Haydon RC, et al. Wnt/h-catenin inhibitor. J Natl Cancer Inst 2006;98:545–56. tumorigenesis. Nat Rev Cancer 2003;3:859–68. signaling pathway as a novel cancer drug target. Curr 21. Germain D, Russell A, Thompson A, Hendley J. 29. Dlugosz A, Merlino G, Yuspa SH. Progress in Cancer Drug Targets 2004;4:653–71. Ubiquitination of free cyclin D1 is independent of cutaneous cancer research. J Investig Dermatol Symp 40. Weeraratna AT. A Wnt-er wonderland—the com- phosphorylation on threonine 286. J Biol Chem 2000;275: Proc 2002;7:17–26. plexity of Wnt signaling in melanoma. Cancer Metasta- 12074–9. 30. Saez E, Rutberg SE, Mueller E, et al. c-Fos is required sis Rev 2005;24:237–50. 22. Le Floch N, Rivat C, De Wever O, et al. The for malignant progression of skin tumors. Cell 1995;82: 41. Leis H, Segrelles C, Ruiz S, Santos M, Paramio JM. proinvasive activity of Wnt-2 is mediated through a 721–32. Expression, localization, and activity of glycogen syn- noncanonical Wnt pathway coupled to GSK-3h and c- 31. Thompson EJ, MacGowan J, Young MR, Colburn N, thase kinase 3h during mouse skin tumorigenesis. Mol Jun/AP-1 signaling. FASEB J 2005;19:144–6. Bowden GT. A dominant negative c-jun specifically Carcinog 2002;35:180–5.

Cancer Res 2007; 67: (16). August 15, 2007 7764 www.aacrjournals.org

Cuiling Ma, Jian Wang, Ying Gao, et al.

Cancer Res 2007;67:7756-7764.

Updated version Access the most recent version of this article at: http://cancerres.aacrjournals.org/content/67/16/7756

Cited articles This article cites 41 articles, 12 of which you can access for free at: http://cancerres.aacrjournals.org/content/67/16/7756.full#ref-list-1

Citing articles This article has been cited by 10 HighWire-hosted articles. Access the articles at: http://cancerres.aacrjournals.org/content/67/16/7756.full#related-urls

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/67/16/7756. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.