MKP-4 suppresses hepatocarcinogenesis by targeting ERK1/2 pathway

Background Mitogen-activated protein kinase phosphatases-4 (MKP-4) is reported to exert a prognostic merit in hepatocarcinogenesis. However, the underlying molecular mechanisms have not been clearly defined. Methods Immunoprecipitation-mass spectrometry (IP-MS) approach was used to identify interactive proteins with MKP-4. Western blot and immunohistochemistry were employed to detect proteins in HCC tissues. Cell counting kit-8, colony formation, Edu incorporation and sphere formation assays were performed to investigate functions of MKP-4/ERK1/2 interaction. Tumor xenografts in nude mice were used to determine effects in vivo. Results Extracellular signal-regulated kinase 1 and 2 (ERK1/2) were identified as binding partners of MKP-4. Knockdown of MKP-4 increased cell proliferation and cancer stem cell (CSC) traits while upregulation of MKP-4 or pre-incubation with ERK1/2 inhibition reversed these effects. Mechanistically MKP-4 negatively regulated phosphorylation of ERK1/2 and reduced expressions of CyclinD1 and c-Myc. Both xenograft tumor models and clinical analysis of HCC patients indicated that lower expression of MKP-4 and higher expressions of ERK1/2 were associated with worse prognosis. Conclusions MKP-4-mediated dephosphorylation of ERK1/2 might serve as a novel tumor-suppressive mechanism and provide a potential therapy for HCC.


Background
Hepatocellular carcinoma (HCC) is one of the most common cancers and the third leading cause of cancerrelated deaths worldwide [1,2]. Although great progress has been made in therapeutic strategies [3], 5-year survival of HCC is no more than 10% [4,5]. Identification of molecular mechanisms involved in HCC is of particular significance.
Mitogen-activated protein kinases (MAPKs) are important signal transduction molecules which regulate a variety of cellular processes, including cell differentiation, proliferation and apoptosis [6,7]. MAPK kinases promote activation of MAPKs by phosphorylation of threonine/serine residues [8]. Dual-specificity phosphatases (DUSPs), which selectively dephosphorylate threonine/serine and tyrosine residues on MAPKs, negatively regulate signal transduction of MAPK cascades [9,10]. Mitogen-activated protein kinase phosphatases (MKPs) are a subgroup of DUSPs, including 11 members. Recent researches indicated that MKP-4 inhibited the progression of colorectal cancer, gastric cancer and clear cell renal cell carcinoma [11,12]. It triggers cellular enlargement, microtubule disruption, G2/M-associated cell death, and some features of mitotic catastrophe in epidermal carcinogenesis [13]. Our previous studies revealed MKP-4 as a potential tumor suppressor in hepatocellular carcinoma [14]. However, precise mechanisms remain poorly understood.
Here we confirmed that MKP-4 could interact with ERK1/2 and negatively regulate ERK1/2 pathway through dephosphorylating ERK1/2 in liver tumor cells and xenograft tumor models. We also demonstrated that lower expression of MKP-4 was correlated with higher expressions of ERK1/2 and p-ERK1/2 in HCC tissues. On the basis of these findings, we conclude that MKP-4 may suppress hepatocarcinogenesis by targeting ERK1/2 pathway.

Mass spectrometry assay
HCC tissues were drew with immunoprecipitation lysis buffer (25 mM Tris-HCl (pH 7.5), 150 mM NaCl, 1 mM EDTA, and 1% NP-40, pH 7.8) and pre-clarified with protein G Sepharose (Sigma) for 2 h. Protein (100 mg) was immunoprecipited with anti-MKP-4 antibody at 4 °C overnight. The complexes were retrieved with protein The results of mass spectrometry in HCC tissues. b Verification of the interaction between MKP-4 and ERK1/2 in HCC tissues using immunoprecipitation assay. c Reciprocal immunoprecipitation of MKP-4 and ERK1/2 in HepG2 cells. Lysates of HepG2 cells were immunoprecipitated with anti-MKP-4, anti-ERK1/2 antibodies or control IgG. The immunoprecipitates were subjected to western blot analysis with anti-ERK1/2 and anti-MKP-4 antibodies. d Immunofluorescence analysis of MKP-4 and ERK1/2 in HepG2 cells. HepG2 cells were subjected to immunofluorescence assay using anti-MKP-4 and anti-ERK1/2 antibodies. Scale bar: 50 μm G Sepharose for another 2 h. The precipitations were washed three times and then loaded onto 10% polyacrylamide gel and stained with coomassie brilliant blue. The gels were cut and then analyzed for the interacted proteins using an LTQ mass spectrometer (Thermo, San Jose, CA). The peptide maps were clustered and aligned using clustering parameters. The peptide clusters were aligned with Mascot identification files to assign sequence identity. Protein identifications were accepted if they could be established at 95% probability and contained at least two unique identified peptides.

Cell counting kit-8 assay
Cell proliferation assay was performed by cell counting kit-8 (CCK-8) solution according to the manufacturer's protocol. Liver tumor cells were firstly plated at a density of 2 × 10 4 cells per well in 100 μl volume in a 96-well plate. Then cells were incubated with 90 μl complete DMEM medium and 10 μl CCK-8 reagent (Dojindo, Kumamoto, Japan) under different treatments after cell adherence. Cells were incubated for 2 h at 37 °C and the absorbance was measured at 490 nm and 630 nm using a microplate reader (Bio-Rad).

Colony formation assays
For colony formation assays, liver tumor cells (500 cells per well) were plated in 6-well culture plates. After 2 weeks, the surviving colonies (50 cells per colony) were counted after staining with 0.5% crystal violet for 30 min.

Edu incorporation assay
Cells were plated into a 96-well plate and then labeled with 20 μM Edu overnight. After labeling and washing, cells were fixed with formaldehyde rinsed and stained with Alexa488-azide for 20 min. After washing three times with PBS with 0.5% Triton X-100, the cells were stained with 10 μM Hoechst 33,342 for 30 min. The cells were washed again and imaged by fluorescence microscopy.

Sphere formation assay
HepG2 and SK-Hep1 cells were incubated in anchorageindependent conditions for tumor sphere formation assay. Liver tumor cells were seeded into 6-well plates and maintained in serum-free medium. Basic fibroblast growth factor (b-FGF; 10 ng/ml; R&D Systems) and fresh epidermal growth factor (EGF; 20 ng/ml; R&D Systems) were added every other day. The radius of each tumor spheroid and the number of tumor spheres were was measured using NIS-Elements Microscope Imaging Software (Nikon, Tokyo, Japan) after 2 weeks.

Xenograft mouse model
Five-week-old female nude mice purchased from Shanghai SLAC Animal Center were raised in a pathogen-free condition. A total of 2 × 10 6 HepG2-shMKP-4; HepG2-MKP-4 or HepG2-control cells were re-suspended in 200 μl PBS and injected subcutaneously into the nude mice. The tumor volume was measured for 7 days with a vernier caliper and calculated on the basis of the following formula: volume (mm 3 ) = length × width × height × 0.52.) [24]. The mice were sacrificed 28 days after injection and the tumors were removed and weighted. The experimental protocol was approved by the Committee on Animals Care and Use of Nantong University.

Immunohistochemistry
For immunohistochemically analysis, HCC sections were deparaffinized and rehydrated with graded ethanol, then soaked in EDTA (1 mmol/L, pH 8.0) and heated to 121 °C to retrieve the antigen. After rinsing with phosphate-buffered saline (PBS, pH 7.2), 0.3% Hydrogen peroxide was applied to block endogenous peroxide activity for 20 min, 10% goat serum was applied to block any nonspecific reactions for 1 h. After washing with PBS, the sections were incubated with the primary antibody overnight. All sections were processed using the peroxidase-anti-per-oxidase method (Dako, Hamburg, Germany). The slides were counterstained with DAB (0.1% phosphate buffer solution, 0.02% diaminobenzidine tetrahydrochloride, and 3% H 2 O 2 ) dehydrated, and fastened with resin mount. Finally, the slides were examined with a Leica CTR5000 microscope (Leica Microsystems, Wetzlar, Germany).

Statistical analysis
All statistical analyses were performed using SPSS (Statistical Product and Service Solutions) 20.0 software package. Statistical analyses of continuous variables were performed by Student's t test. Paired t-tests were used to compare xenograft tumor size and MKP-4 expressions in paired clinical samples. Pearson's Chi square test was performed to evaluate associations between MKP-4, ERK1/2 and p-ERK1/2 expressions and clinicopathological factors. Kaplan-Meier plots and log-rank tests were used for overall survival analysis. Multivariate analysis was constructed using the Cox proportional hazards model. P < 0.05 was considered statistically significant. All statistical tests were two-sided.

Interaction between MKP-4 and ERK1/2 in HCC
ERK1/2 were identified as novel binding partners of MKP-4 in HCC tissues (Fig. 1a). Immunoprecipitation assay was performed in HCC tissues and HepG2 cells to validate the interaction of MKP-4/ERK1/2 (Fig. 1b,  c). In addition, immunofluorescence staining revealed that MKP-4 and ERK1/2 proteins were co-localized in cytoplasm of HepG2 cells (Fig. 1d), which provided further support for a functional interplay.

MKP-4 regulates phosphorylation of ERK1/2 in liver tumor cells
Since p-ERK1/2 is the active form of ERK1/2 and plays a vital role in tumor progression, we speculated whether MKP-4 could regulate ERK1/2 phosphorylation. We employed western blot analysis and found that MKP-4 expression was obviously down-regulated in liver tumor cells, as compared with LO2 hepatocytes (Fig. 2a). We used RNA interference to knockdown MKP-4 expression in HepG2 or SK-Hep1 cells and found that MKP-4 siRNA5 exerted the best interfering efficiency (Fig. 2b). Moreover, Myc-tagged MKP-4 was employed to upregulate MKP-4 expression in liver tumor cells. After that, we detected the expressions of p-ERK1/2 and downstream genes in different treated cells as shown in Fig. 2c. The results indicated that expressions of p-ERK1/2, CyclinD1 and c-Myc were decreased by overexpression of MKP-4 or pre-incubation of 10 μM PD98059 by 24 h while expressions of the above genes were increased by MKP-4 interference. These data suggested that MKP-4 could regulate phosphorylation of ERK1/2 and ERK1/2 pathway in liver tumor cells.

MKP-4 inhibits cell proliferation and cancer stem cell (CSC) traits through ERK1/2 pathway
We investigated biological effects of the interaction between MKP-4 and ERK1/2 following different treatments. Colony formation, CCK-8 assays and Edu assays indicated that proliferation of HepG2 and SK-Hep1 cells were significantly increased after MKP-4 depletion, whereas overexpression of MKP-4 impaired the capacity of cell proliferation. Moreover, treatment with 10 μM PD98059, an antagonist of ERK kinases by 24 h, abrogated the pro-proliferative effect of MKP-4 depletion in liver tumor cells (Fig. 3a-c). CSCs can form spheres in the absence of serum under low adherence conditions. Therefore, we evaluated the ability of different treated cells to grow spheres under serum-free conditions. The ability of tumor sphere formation was decreased in MKP-4-overexpressing cells or MKP-4-lacking cells with the inhibition of ERK1/2 pathway whereas the ability was enhanced accompanied by the decreased expression of MKP-4 (Fig. 3d).

MKP-4 suppresses tumor growth in vivo through the modulation of ERK1/2 signaling
We then examined the effect of MKP-4 on HCC progression using subcutaneous xenograft model. As shown in Fig. 4a, tumors in HepG2-MKP-4 group grew much slower than HepG2-control and HepG2-shMKP-4 groups. Both tumor volumes and weights in HepG2-MKP-4 group were significantly lower than the other two groups 28 days after the subcutaneous implantation (Fig. 4b, c). The volumes of MKP-4-overexpressing tumors increased slower than other groups, as indicated by tumor growth curves (Fig. 4d). These results indicated that depletion of MKP-4 significantly promoted the progression of HCC in vivo. To further clarify whether the interaction between MKP-4 and ERK1/2 is involved in tumor progression, we detected the expressions of MKP-4, ERK1/2 and p-ERK1/2 using western blot and immunohistochemistry analyses. The results showed that MKP-4 significantly decreased the phosphorylation levels of ERK1/2 ( Fig. 4e-g). Together, these findings implicated that MKP-4 suppresses growth of HCC in nude mice via regulation of ERK1/2 pathway.

Expressions of MKP-4, ERK1/2 and p-ERK1/2 in HCC tissues
To further determine the relationship between MKP-4, ERK1/2 and p-ERK1/2, we analyzed the expressions in eight paired HCC and adjacent non-tumorous tissues using western blot analysis. Our results revealed significantly lower expression of MKP-4 and higher expression of ERK1/2, p-ERK1/2 in HCC tissues than in the nontumorous tissues (Fig. 5a, b). Furthermore, we performed immunohistochemical analysis to detect the expression of MKP-4, ERK1/2, p-ERK1/2 and Ki-67 in 160 HCC specimens and found that expression of MKP-4 was frequently downregulated while the expressions of ERK1/2, p-ERK1/2 and Ki-67 were elevated in tumorous samples compared with non-tumorous tissues (Fig. 5c, d).

Relationship between expressions of MKP-4, ERK1/2, p-ERK1/2 and clinicopathological factors of HCC
To reveal the correlation between protein expressions and clinical characteristics, clinical samples were divided into low and high expression groups according to immunohistochemical evaluation. As shown in Table 1, expressions of MKP-4, ERK1/2 and p-ERK1/2 were correlated with tumor differentiation (P < 0.001, P = 0.004 and P = 0.042), microvascular invasion (P = 0.006, P = 0.031 and P = 0.078) and TNM stage (P = 0.008, P = 0.013 and P = 0.014). However, there was no significant relationship between other prognostic factors, such as age, gender, tumor size, tumor number, Child-Pugh score, tumor encapsulation, HBsAg and serum AFP level. Furthermore, univariate analysis showed that tumor differentiation, TNM stage, MKP-4 expression, ERK1/2 expression and p-ERK1/2 expression were significantly associated with patients' survival (Tables 2, 3). In addition, further studies showed that MKP-4 expression was positively correlated with ERK1/2 and p-ERK1/2 expression in HCC tissues (Table 4). Kaplan-Meier survival curves indicated that low expression of MKP-4 and high expression of ERK1/2, p-ERK1/2 were significantly associated with poor overall survival (Fig. 5e).

Discussion
Hepatocellular carcinoma, especially diagnosed at an advanced stage, has been considered to be one of the most fatal cancers [25]. Novel molecular targets for diagnosis and therapy of HCC are urgently needed to improve HCC prognosis. Recently, deregulation of MAPK pathways has been identified to play a vital role in the pathogenesis of HCC [26][27][28]. Therefore, it is vital to seek potential mechanisms underlying deregulation of MAPK pathways in HCC initiation and progression.
MAPK pathways, highly conserved in the majority of eukaryotes, play key roles in cellular developmental and physiological processes by delivering extracellular signals into nuclei [29]. Aberrant activation of MAPK pathways are reported to be associated with development of tumors [30]. Undergoing a cascade of sequential phosphorylation events mediated by upstream MEK kinases, phosphorylation of MAPKs on threonine and tyrosine residues can be activated [31]. MKP-4 is a member of (See figure on next page.) Fig. 5 Expressions and prognostic roles of MKP-4/ERK1/2 in HCC patients. a Western blot analysis revealed a lower expression of MKP-4 or higher expressions of ERK1/2, p-ERK1/2 in hepatocellular carcinoma (T) and adjacent non-tumorous tissues (N). b The bar chart demonstrates the ratio of MKP-4, ERK1/2 and p-ERK1/2 to GAPDH by quantitative analysis. *P < 0.05 compared with adjacent normal liver tissue. c Immunohistochemical analysis of MKP-4, ERK1/2 p-ERK1/2 and Ki-67 expressions in paraffin-embedded tissue sections. Scale bar: 100 μm. d Relationship between MKP-4, ERK1/2, p-ERK1/2 and Ki-67 expression in HCC patients. Scatter plot of them with regression line showing a significant correlation using the Pearson's test (P < 0.01). e Kaplan-Meier survival curve according to MKP-4, p-ERK1/2 and ERK1/2 in 160 HCC patients (P < 0.05, log-rank test) MAPK phosphatases which is composed of two domains, MAPK-binding domain in N-terminal whereas the dual-specificity phosphatase domain in C-terminal [32]. ERK1/2 are critical members of MAPKs and involved in plenty of fundamental cellular processes by regulating the phosphorylation of various substrates [33]. In our study, we detected interaction and explored functions of MKP-4/ERK1/2 in HCC both in vivo and in vitro. We speculates that association of MKP-4 to ERK1/2 is MBPdependent through direct binding of the two proteins and this will be further confirmed by truncation analysis or GST pull-down. ERK1/2 can translocate into nucleus and promote transcription by phosphorylation in HCC, while combination of MKP-4 and ERK1/2 greatly reduces the entry of p-ERK1/2. This result is consistent with MKP-1, which is downregulated and controls ERK1/2 phosphorylation in HCC [34,35]. Interestingly, evidence here showed that MKP-4 also negatively regulates total protein level of ERK1/2. MKP-4 may affects protein stability of ERK1/2 followed by effects on phosphor-dynamics of ERK1/2. We will perform additional experiments regarding the protein stability of ERK1/2 by MKP-4 in further study.
A small subset within tumour bulk which was defined as tumour-initiating cells (TICs), are considered to be source of tumors including HCC [36]. Liver TICs are reported to be responsible for tumorigenesis and intervention of TIC self-renewal can be a potential treatment in HCC [37]. To determine effect of MKP-4/ERK1/2 interaction in self-renewal potential of liver TICs, we performed sphere formation and validated the promotion of liver TICs self-renewal by MKP-4/ERK1/2 interaction. Since c-Myc which has been recognized as a vital regulator of stem cell biology can serve as a link connecting malignancy and stem cells [38], we detected its expression in different treated cells and tissues from xenograft mice. Our results demonstrated that depletion of MKP-4 increased c-Myc expression, while overexpression of MKP-4 decreased its expression. In consequence, we speculated that interaction of MKP-4 and ERK1/2 inhibit self-renew of liver tumor cells and HCC initiation partly through the transcription factor c-Myc which is a downstream target gene of ERK1/2 pathway. Although lots of transcription factors and signal pathways have been reported to participate in stem cell self-renewal. Due to limitation of time and money, we have not carried out a systematic and comprehensive study in this aspect and just found such a phenomenon. We will do further study in the future.
Our results demonstrate that MKP-4 was downregulated in HCC and that lower expressions of MKP-4 were closely related to higher expressions of ERK1/2 and p-ERK1/2, which are indicators of poor prognosis in HCC. DNA methylation of promoter-associated CpG islands can function as a potential mechanism of silencing tumor suppressor genes in numerous cancers, including HCC [39]. Hypermethylation of CpG islands in the promoter region of tumor suppressor genes is a major event in the development of many cancers [40]. MKP-4 also acts as a tumor suppressor gene in many other cancers in addition to HCC and it has been reported that promoter methylation of DUSP9 in human gastric  [11,12]. This may be one of the reason for decreased expression of MKP-4 in HCC.

Conclusion
We demonstrate that MKP-4 inhibits the occurrence and development of HCC through directly promoting the dephosphorylation of ERK1/2 and decreasing expression of CyclinD1 and c-Myc (Fig. 6). Thus, we supposed the dephosphorylation of ERK1/2 by MKP-4 may act as a promising therapeutic strategy in HCC.