REGγ is critical for skin carcinogenesis by modulating the Wnt/β-catenin pathway

Here we report that mice deficient for the proteasome activator, REGγ, exhibit a marked resistance to TPA (12-O-tetradecanoyl-phorbol-13-acetate)-induced keratinocyte proliferation, epidermal hyperplasia and onset of papillomas compared with wild-type counterparts. Interestingly, a massive increase of REGγ in skin tissues or cells resulting from TPA induces activation of p38 mitogen-activated protein kinase (MAPK/p38). Blocking p38 MAPK activation prevents REGγ elevation in HaCaT cells with TPA treatment. AP-1, the downstream effector of MAPK/p38, directly binds to the REGγ promoter and activates its transcription in response to TPA stimulation. Furthermore, we find that REGγ activates Wnt/β-catenin signalling by degrading GSK-3β in vitro and in cells, increasing levels of CyclinD1 and c-Myc, the downstream targets of β-catenin. Conversely, MAPK/p38 inactivation or REGγ deletion prevents the increase of cyclinD1 and c-Myc by TPA. This study demonstrates that REGγ acts in skin tumorigenesis mediating MAPK/p38 activation of the Wnt/β-catenin pathway. REGγ, a proteasome activator, is associated with multiple oncogenic pathways in human cancers and can promote the degradation of multiple proteins including p53. Here the authors highlight a potential role for REGγ in skin cancer and propose a molecular mechanism linking p38 MAPK and Wnt signalling.

R EGg (also known as PA28g, PSME3 and Ki antigen) is a member of the 11S family of proteasome activators of the 20S core proteasome. It can promote the degradation of multiple proteins including p53 and MDM2 in an ubiquitin-and ATP-independent manner [1][2][3] . Previous studies found that REGg is overexpressed in numerous cancers including breast, thyroid, lung and liver cancers [4][5][6] , suggesting a potential role for REGg in tumorigenesis.
Non-melanoma skin cancers comprised basal cell carcinoma and squamous cell carcinoma that are the most common of all human cancers. The phorbol ester TPA is a skin tumour promoter, resulting in activation of protein kinase C (PKC) isotypes by mimicking diacylglycerol (DAG) function 7 . Serial activation of some PKCs induced by TPA is crucial for the initiation and progression of skin tumours 8 . PKC isoforms are known to activate multiple MAPK components including ERK, JNK and p38 (ref. 9). Abnormalities in the MAPK pathway play a critical role in the development and progression of skin cancer by targeting AP-1 (cJun and cFos) 10 . In a majority of SCC, Ras GTPase and MAPK cascade are activated 11,12 .
A previous study found that the Wnt/b-catenin/TCF signalling pathway is constitutively activated in non-melanocytic skin tumours 13 . Moreover, abnormal and nuclear localization of b-catenin have been observed in human skin cancer 14 . C-Myc, a downstream target of b-catenin, can regulate p21 and is involved in Ras-driven epidermal tumorigenesis 15 . Thus, it seems that different signalling pathways may connect to each other, contributing to skin tumorigenesis. However, cross-talk between the MAPK signalling pathway and the Wnt /b-catenin signalling pathway has not been demonstrated during the development of skin cancer.
In the present study, we demonstrate that REGg promotes skin tumour development in a two-stage skin carcinogenesis mouse model. TPA stimulation upregulates expression of REGg via the MAPK/p38 signalling pathway. Subsequently, overexpressed REGg activates the b-catenin/Wnt signalling pathway through degradation of GSK-3b, leading to abnormal cell proliferation and concurrent papilloma formation. Skin tumour initiation in REGg -/mice is markedly inhibited, suggesting that REGg is critical for skin tumorigenesis.

Results
REGc is critical for skin tumorigenesis in a mouse model. The role of REGg in skin tumour formation was examined in animals by chemical carcinogenesis. Compared with REGg þ / þ mice that had papillomas as early as 9 weeks after the first TPA application, REGg -/mice required at least 15 weeks for tumour induction. On average, REGg þ / þ mice developed 6.6 tumours per mouse with 20 weeks of TPA treatment, whereas REGg -/mice developed only 1.1 tumours per mouse (two-tailed Student's t-test, n ¼ 10, Po0.001) (Fig. 1a,b). The volume of tumours in REGg -/mice was markedly reduced in comparison with that in REGg þ / þ mice (two-tailed Student's t-test, n ¼ 10, Po0.001) (Fig. 1c), suggesting that REGg functions as an oncogenic promoter for papilloma formation in a DMBA/TPA tumour model.
All skin tumours derived from REGg þ / þ and REGg -/mice were characterized as well-differentiated squamous cell papillomas (Fig. 1d). There were no significant histological differences in tumours from REGg þ / þ and REGg -/mice. The number of Ki-67-positive cells in skin tumours from REGg þ / þ mice was higher than that in REGg -/mice by immunohistochemistry (Fig. 1e, Supplementary Fig. 1a), indicating a tumorigenic potential from action of REGg.
We then examined the effects on epidermal cell proliferation of a single topical treatment of TPA and found that TPA-treated REGg þ / þ mouse skin exhibited a marked increase in epidermal thickness (Fig. 1f) along with an increased number of Ki67positive cells ( Supplementary Fig. 1b,c), compared with REGg -/mice. Thus, the resistance to tumour formation in REGg-deficient skin is, at least partially, related to decreased proliferation in epidermal cells. These results suggest an early action of REGg in carcinogenesis.
TPA treatment promotes REGc expression in skin. In the TPA-induced mouse cancer and cultured cell models, we found an increased expression of REGg at mRNA and protein levels ( Fig. 2a-d). To verify this, we generated a luciferase reporter driven by the REGg-promoter (REGg-Luc) and tested TPAinduced expression of this reporter in HaCaT cells. We found that TPA was able to promote REGg-luciferase activity in HaCaT cells, reflecting that TPA-induced transcription of REGg is time and dose-dependent (Fig. 2e,f). Immunohistochemical data also showed elevated levels of REGg in TPA-treated skin and TPA-induced tumour tissues compared with those in normal skin ( Supplementary Fig. 2a), validating the increased expression of REGg after TPA exposure in skin. Together, these data raise a possibility that REGg may be a critical factor mediating TPA-induced skin tumour promotion.
p38 is essential for TPA-induced overexpression of REGc. It is known that TPA activates MAPK/p38 through multiple PKCs 16 . To address how TPA induced abnormal expression of REGg, we measured the activities of MAPK pathways. The levels of phosphorylated p38 and cJun were increased in TPA-treated mouse skin compared with untreated controls. More marked elevation of p-p38 and p-cJun was observed in TPA-induced tumours (Fig. 3a). It appears the levels of active p38 and cJun are positively correlated with the expression of REGg (Fig. 3a).
To determine whether activation of the MAPK/p38/AP-1 pathway is required for increased REGg expression in response to TPA, we tested the action of p38 inhibitor, SB202190, on REGg expression. The specific p38 inhibitor effectively prevented TPA-induced increase of REGg mRNA and protein expression in HaCaT cells (Fig. 3b,c). It is noteworthy that p38 and cJun levels can be enhanced by TPA treatments in the skin of both REGg þ / þ and REGg -/mice (Fig. 3a), indicating that the MAPK/p38/AP-1 pathway acts upstream of REGg. However, the activity of JNK (a different MAPK member) was unchanged in the skin before and after TPA treatment ( Supplementary Fig. 2b), indicating a pathway-specific action by TPA.
To address the relation between REGg and p38/AP-1 signalling pathways, we performed semi-quantitative analysis of these protein levels with IHC staining of human SCC samples. REGg was highly expressed in 98.2% SCC of tumour samples ( Supplementary Fig. 2c). Consistently, positive staining of p-cJun was observed in 80% of SCC samples and high levels of p-p38 staining in 84.2% SCC samples ( Supplementary Fig. 2c). Thus, the expression of REGg and AP-1 signal molecules was highly correlated in human SCC tumours ( Supplementary Fig. 2c). Together, these results suggest that the MAPK/p38/AP-1 pathway may be responsible for TPA-induced regulation of REGg transcription.
AP-1 promotes REGc transcription by binding to its promoter. To define which regulatory region(s) in the REGg promoter is required for transcriptional response to TPA, we generated a series of deletion mutants in the promoter region of the REGg-Luc construct. Since the ( À 700/ À 500) regions had minimal REGg-Luc activity compared with the À 1,300/ À 1,100 region, the REGg UTR regions between À 1,100 and À 900 should contain an element responsive to TPA (Fig. 4a). The fact that AP-1 binds to TPA response element (TGACTCA) within target genes 17 facilitated our finding a TPA response element (TRE) in the REGg gene ( À 997/ À 991). A TRE-deleted REGg-Luc construct was generated and expression of this construct in HaCaT cells had no response to TPA treatment (Fig. 4b), suggesting that AP-1 could be a critical regulator for REGg transcription.
To elucidate AP-1-mediated regulation of REGg, EMSA was performed using a TRE probe and HaCaT cell extracts with or without TPA stimulation. We found increased binding of the TRE probe to AP-1 in lysates from cells treated with TPA, indicating that AP-1 is a transcription factor binding to the TRE in REGg. In the EMSA competition experiments, AP-1 association with the TRE probe was effectively competed by oligonucleotides with intact TRE sequence, but not by mutant oligonucleotides (Fig. 4c). Importantly, silencing p38 expression or application of p38 inhibitor blocked AP-1 binding to REGg promoter ( Supplementary Fig. 3a,b).
To address whether AP-1 acts on the transcription of REGg in response to TPA in the cell, we performed CHIP assay using primers flanking the TRE in the 5 0 -UTR of REGg. As expected, we found that both cJun and c-Fos were recruited to the corresponding region in the CHIP analysis (Fig. 4d). Taken together, these results indicate that TPA treatment enhances the activity of AP-1, which directly upregulates REGg transcription in skin cells.
REGc potentiates Wnt signalling via degradation of GSK-3b. To understand how REGg overexpression leads to skin carcinogenesis following TPA induction, we evaluated the Wnt signalling pathway that has been suggested to play an important role in the progression of skin cancers. We found higher levels of b-catenin in the skin tumour of REGg þ / þ mice than in REGg -/mice by western blot and immunohistochemical analysis (Fig. 5a,b). Consistently, expression of b-catenin target genes, including CyclinD1 and c-Myc, was increased in REGg þ / þ skin but lowered or had no change in REGg -/skin cells (Fig. 5a,b). Furthermore, we observed a significant increase in b-catenin, CyclinD1 and c-Myc in REGg þ / þ skin cells treated with TPA ( Supplementary Fig. 3c). Importantly, both IHC and Western blot data showed an inverse correlation between GSK-3b and b-catenin levels in skin tumour samples from REGg þ / þ and REGg -/mice ( Fig. 5a,b). Taken together, these results demonstrate that REGg promotes Wnt/b-catenin signalling and enhances the expression of Wnt target genes.
As GSK-3b is required for phosphorylation and subsequent degradation of b-catenin, we tested whether REGg may directly regulate the stability of GSK-3b. Strikingly, the level of GSK-3b in the skin of REGg -/mice was much higher than that in REGg þ / þ mice (Fig. 6a). We found no differences in the mRNA level of GSK-3b between REGg KO and WT skins (Fig. 6b). Furthermore, we found physical interactions between REGg and GSK-3b as defined by a co-immunoprecipitation assay following transient expression of FLAG-REGg and GSK-3b, alone or in combination with a control vector in 293T cells (Fig. 6c). Importantly, co-immunoprecipitation assays using lysates from HaCaT cells demonstrated clear interaction between endogenous REGg and GSK-3b (Fig. 6d). These results suggest that GSK-3b is a target of the REGg proteasome.
To gain additional insight into the mechanism of REGgmediated GSK-3b degradation, we examined the capacity of faster degradation of GSK-3b in the presence of REGg (Fig. 6f,g). Using cells inducibly expressing a WT REGg or an enzymatically inactive mutant (N151Y) REGg, we demonstrated that only cells expressing functional REGg could repress GSK-3b expression. (Supplementary Fig. 3d). Therefore, we conclude that GSK-3b is a direct target of the REGg proteasome.
In REGg -/skin, p-p38 and p-cJun were increased while b-catenin and its target genes were not elevated ( Supplementary Fig. 4a). We also observed that expression of REGg, b-catenin, c-Myc and CyclinD1 were significantly reduced when p38 activity was inhibited ( Supplementary Fig.  3c). Silencing cJun or p38 by a si-RNA significantly reduced the expression of REGg, b-catenin, c-Myc and CyclinD1 ( Supplementary Fig. 4b-d), indicating a regulatory pathway linking AP-1, REGg and Wnt/b-catenin downstream effectors. Interestingly, the reduction of REGg, b-catenin, c-Myc and CyclinD1 in the cJun-knockdown cells was effectively restored with overexpression of REGg ( Supplementary Fig. 5a). In a mouse xenograft model, tumours derived from REGg-knockdown A549 cells were significantly smaller than those from A549 cells with normal REGg expression. Cells expressing a constitutively active b-catenin (S37Y mutant 19 ) can 'rescue' REGg depletion and produce much larger tumours than REGg knockdown alone (Fig. 7b, Supplementary Fig. 5b). Similar results were observed for HaCaT cells (Fig. 7c, Supplementary  Fig. 5c,d). MTT assay showed that HaCaT cells expressing the active b-catenin proliferated faster than normal cells ( Supplementary Fig. 5e). Furthermore, we found that REGg, p-p38 and p-cJun were highly expressed in some human skin tumours compared with normal skin controls (Fig. 7d, Supplementary Fig. 2c). In contrast, GSK-3b exhibited a much lower expression level in human skin SCC tumours than in normal skin samples (Fig. 7d). Quantified analysis of IHC for human SCC tumours showed REGg was positively correlated to p-cJun and b-catenin but negatively associated with GSK-3b ( Supplementary Fig. 2c). Taken together, out data show that REGg has a novel action, bridging the MAPK/p38/ AP-1 pathway and the Wnt/b-catenin pathway in skin tumorigenesis.

Discussion
In the present study, we analysed the potential role of REGg in mouse skin carcinogenesis using WT and REGg -/model. We demonstrate that REGg functions as an oncogene for skin tumorigenesis in a two-stage skin carcinogenesis model, in which TPA-induced overexpression of REGg is dependent on the activation of MAPK/p38 signalling pathway. AP-1 complex can bind directly to the TRE upstream the REGg promoter and enhance REGg transcription, ultimately activating the Wnt/bcatenin signalling pathway by augmenting the degradation of GSK-3b. Notably, REGg acts as a critical factor bridging the MAPK/p38 and the Wnt/b-catenin signalling pathways during skin carcinogenesis, indicating that targeting REGg could be an alternative approach for skin cancer therapy.
As TPA can substitute for naturally occurring diacylglycerol (DAG) for an effective activation of PKC, it is used as a classical skin tumour promoter 20 . PKCs have been implicated as mediators for MAPK/p38 activation 21,22 . In addition to mediating TPA-induced skin cancer, p38 is also critical for UV-induced skin carcinogenesis 17,23,24 . Mice deficient for p38d exhibit a marked resistance to TPA-induced skin tumour development by impairing ERK1/2-AP-1 and STAT3 pathways 25 . A previous study reported that AP-1-induced involucrin gene expression may be one mechanism in skin tumour development 9 . In this study, we demonstrate that p38 activation is required for REGg-mediated signalling relays and skin carcinogenesis. Increased transcription of REGg via the p38/AP-1 signalling pathway reveals an additional mechanism that significantly contributes to the activation of the Wnt/b-catenin signalling and skin tumorigenesis. Despite a positive correlation between MAPK/p38 activation and REGg overexpression, JNK activation is not observed following TPA stimulation in our study, indicating some specificity in the transduction of TPA-induced carcinogenic signals 26 . A similar phenomenon has been reported; HaCaT-shN and HaCaT-shR cells were treated with CHX (100 mg ml À 1 ) for indicated times followed by western blotting. (g) Quantitated results in (f) were plotted against indicated time periods to indicate dynamic changes (two-tailed Student's t-test, n ¼ 3).
MMP-1 upregulation and TIMP-1 downregulation are p38 (but not ERK or JNK) MAPK-dependent 27 . This is the first study demonstrating that the MAPK/p38 signalling pathway is involved in the regulation of REGg expression.
Interestingly, both Wnt and MAPK/p38 signalling pathways are involved in carcinogenesis and tumour progression, yet possible synergistic contribution to skin cancer is unknown. In the past several years, the cross-talk between the p38/MAPK and the Wnt/b-catenin signalling pathways have been noticed [28][29][30] with little mechanistic explanation. In this study, we show for the first time that the MAPK/p38/Ap-1 signalling pathway is coupled with the Wnt/b-catenin signalling via the REGg proteasome system, promoting skin tumour development. This conclusion is also endorsed by our studies using human skin cancer samples. b-Catenin stabilization is involved in upregulation of PKC activity, reflecting a feedback regulation between these two pathways 31 . This may explain our observation that activation of MAPK/p38/AP-1 is stronger in WT than that in REGg -/skin after days of TPA treatment (Fig. 3a). It is worth emphasizing that REGg-KO mice also develop a few tumours after 20 weeks of TPA treatments, but with delayed onset, in agreement with the importance of MAPK/p38 signalling for additional targets, such as Stat3, AKT and NFkB 32 . Indeed, combined inhibition of p38 and AKT signalling pathways abrogates cyclosporine A-mediated pathogenesis of aggressive skin SCCs 33 . Thus, additional studies are necessary to elucidate the cross-talk between the p38/MAPK signalling pathway and other signalling molecules during skin carcinogenesis in REGg-deficient mice.
Given that phosphorylation of b-catenin by GSK-3b triggers b-TrCP-mediated degradation of b-catenin 34 , we tested whether REGg might regulate components in the Wnt/b-catenin pathway. REGg can directly interact with GSK-3b and promote the degradation of GSK-3b in a ubiquitin-independent manner. Thus, the activation of Wnt/b-catenin signalling pathway followed by REGg overexpression is dependent on the proteolysis of GSK-3b. Interestingly, we found REGg deficiency had more profound effects on b-catenin upregulation than on GSK-3b in some experiments. As CK1 is a priming kinase to synergize with GSK-3b for b-TRCP-mediated destruction of b-catenin 35 , our previous finding that REGg can interact with CK1 and promoted its degradation 21 may interpret this phenomenon.
We have shown that REGg can promote the degradation of cancer suppressor p53 and p21 and thus is likely involved in the regulation of cell cycle and cell growth 2,36 . Although p21 is a regulator in skin cancer development 15,37 , no striking differences were found in the expression of p53 and p21 between the REGg þ / þ and REGg -/skin ( Supplementary Fig. 6a). In addition, TPA treatment could induce p53 expression in skin of both REGg þ / þ and REGg -/mice ( Supplementary Fig. 6b), indicating that p53 and p21 might be less important factors in the TPA-induced skin cancer model. Whether REGg-dependent regulation of p53 and p21 occurs in subsets of skin cell types in response to environmental changes deserves further investigation.
In summary, our data support a model in which TPA stimulation upregulates the expression of REGg through the Conversely, REGg deletion blocks the linkage between the MAPK/p38/AP-1 signalling and the Wnt/b-catenin signal pathways in response to TPA, attenuating tumour formation (Fig. 8).
Thus, REGg acts as a critical molecule bridging the MAPK/p38/ AP-1 pathway and the Wnt/b-catenin pathway during the development of skin carcinogenesis.
Cell culture and treatment. HaCaT cells were purchased from ATCC. The HaCaT stable cell lines were generated by integration of retroviral shRNA vectors specific for REGg or a control gene (GFP) from OriGene (Rockville, MD). The 293-REGg-inducible cell line was previously reported 2 . All cell lines were cultured in DMEM supplemented with 10% fetal bovine serum (Gibco). TPA dissolved in acetone at 0.5 mg ml À 1 was added to HaCaT cells culture media at a final concentration of 50 ng ml À 1 . An MAPK/p38 inhibitor SB202190 dissolved in DMSO was added to the culture media (10 mM) 1 h before TPA treatment.
Mice. REGg -/mice with C57BL/6 genetic background were acquired from Dr John J. Monaco 38 and bred in the Animal Core Facility following procedures approved by the Institutional Animal Care and Use Committee. Genotyping of REGg þ / þ and REGg -/mice was carried out by PCR analysis of genomic DNA as described previously 16 .
DMBA/TPA treatment. All animal experiments were performed with 8-week-old female REGg þ / þ and REGg -/mice. Skin tumours were generated as described 39,40 , with a single dose of DMBA (50 mg per 100 ml acetone; Sigma-Aldrich) for the first week followed by repetitive TPA applications (6.25 mg per 100 ml acetone; Sigma-Aldrich) every 48 h for 20 weeks. Tumours were assessed weekly and mice were killed 20 weeks after the final treatment.
Western blot analysis and in vitro proteolytic analysis. Western blot analysis of proteins extracted from cells was performed as described 2 . Equivalent amounts of total protein from each sample were loaded and immuneblots were analysed using primary antibodies specific for REGg, p21, p53, GSK-3b, p-p38, p-cJun, p-c-Fos, p38, cJun, b-catenin, c-Myc, GAPDH and CyclinD1 (1:1,000 dilutions) overnight at 4°C. After incubation with a fluorescent-labelled secondary antibody (1:5,000 dilutions), specific signals for proteins were visualized by a LI-COR Odyssey Infrared Imaging System.
In vitro proteolytic analysis was performed using 10 ml of purified GSK-3b, 0.25 mg of purified 20S proteasome and 1 mg of REGg heptamers for the indicated times in 50 ml reaction volume at 30°C. An aliquot of the reaction was analysed by western blotting (Supplementary Figs 7,8).
Immunohistochemistry and H&E staining. Tumours or normal skin samples were fixed with 4% paraformaldehyde for 3 days and were then dehydrated through a graded series of ethanol and embedded in paraffin. Four-mm sections were cut in a microtome (Leica, Germany) and then stained with haematoxylineosin (H&E).
For immunostaining, slides were heated in a microwave oven at 92°C for 20 m in a citrate buffer. The samples were then incubated overnight at 4°C with the primary antibodies at 1:250. Subsequently, they were incubated for 1 h with biotinylated goat anti-rabbit antibody IgG and then for 30 m with Streptavidin-HRP peroxidase. Colour reaction product was visualized by using diaminobenzidine (DAB)-H 2 O 2 as a substrate for peroxidase. All sections were counterstained with haematoxylin.
Luciferase assay. After transfection with pGL3-REG-g luciferase vector or the pGL3-Basic vector, HaCaT cells were lysed in the luciferase cell culture lysis buffer (Promega). Cell lysates were vortexed and then centrifuged in 4°C at 13,000 g for 2 min. Supernatant was added with the luciferase assay substrate (60-80 ml). Luminescence was measured as relative light units, twice for each lysate, taking the reading of luciferase assay using a LUMIstar OPTIMA (BMG Labtech). Each assay was repeated for at least three times.
ChIP assay and electrophoretic mobility shift assay. For ChIP experiments, cells treated with or without TPA were collected and lysed in lysis buffer (1% SDS, 10 mM EDTA, protease inhibitors and 50 mM Tris-HCl (pH 8.1)). The lysates were then sonicated to result in DNA fragments of 200-1,000 bp in length. Cellular debris was removed by centrifugation and the lysates were diluted 1:10 in ChIP dilution buffer (0.01% SDS, 1.1% Triton X-100, 1.2 mM EDTA, 16.7 mM NaCl, protease inhibitors and 16.7 mM Tris-HCl (pH 8.1)). The samples were immunoprecipitaed with 5 ml of indicated antibodies (IgG, cjun, c-Fos) overnight at 4°C with rotation. DNA-protein immunocomplexes were isolated with 60 ml of protein A agarose beads for 2 h at 4°C. The beads were washed, eluted in 250 ml elution buffer (1% SDS and 100 mM NaHCO 3 ) and crosslinkes were reversed by adding NaCl to a final concentration of 200 mM and incubating overnight at 65°C. The DNA was recovered by phenol/chloroform/isoamyl alcohol (25/24/1) extractions and precipitated with 0.1 volume of 3 M sodium acetate (pH 5.2) and 2 volumes of ethanol using glycogen as a carrier. PCR amplification of the genomic fragments was performed with specific primers flanking putative binding sites on the REGg promoter. The PCR products were electrophoresed in agarose gels and visualized by ethidium bromide. ChIP primer sequences are as follows: forward: 5 0 -CATGTT GAAATACTTGTA-3 0 ; reverse: 5 0 -TCTTCATGCACCCATTCA-3 0 .
MTT assay and xenograft animal model. MTT assay was performed by seeding cells in a 96-well plate at 2.5 Â 10 3 cells per well and were cultured for 36 h. Then cells were incubated with MTT solution at 37°C for 2 h. Absorbance (490 nm) was measured and analysed.
For Xenograft animal models, A549 (ShR) cells or HaCaT (ShR) cells were transfected with pcDNA3.1-b-catenin (S37Y) for 48 h and hygromycin used to screen the monoclonal cells. Female BALB/c nude mice at the age of 5 weeks were prepared. Cells were implanted into the dorsal flanking sites of nude mice at In response to TPA, activation of the MAPK/p38/AP-1 signalling pathway stimulates overexpression of REGg which promotes the degradation of GSK-3b, leading to accumulation of b-catenin and activation of Wnt/bcatenin signalling pathway. Overexpression of proliferation-related genes contributes to, at least in part, abnormal proliferation of keratinocytes and skin tumour development. REGg acts as an oncogene in skin carcinogenesis by linking the MAPK/p38/AP-1 signalling pathway with the Wnt/b-catenin signalling pathway.
2 Â 10 6 cells in 100 ml per spot. Four weeks after injection, mice bearing tumours were killed for the assessment of tumour size and immunohistological examination.
Data collection and statistical analysis. The statistical data were obtained by GraphPad Prism 5.0 software. The intensity of the western blot results was analysed by densitometry using Bio-Rad Quantity One 4.4.0 software. The results were expressed as the mean ± s.d. Statistical analysis was performed using two-tailed, paired Student's t-test or two-way ANOVA. A P value of less than 0.05 was considered statistically significant.