Background
Gastric cancer (GC) is the fourth most common type of cancer and the second leading cause of cancer deaths worldwide [
1,
2]. At present, treatment of GC involves surgery, radiotherapy, chemotherapy and molecular targeted therapy [
3]. Tumor metastasis and recurrence in patients with GC are considered to be the most significant determinants for treatment failure and mortality [
4]. The mechanisms underlying tumor metastasis are very complex, and appear to involve multiple steps [
5,
6]. There is thus an urgent need to identify the molecular constituents of these mechanisms that could be targeted to improve the treatment of GC.
Maternal embryonic leucine zipper kinase (MELK), a member of the sucrose-non-fermenting (SNF1)/AMPK family of serine-threonine kinases, is a cell cycle dependent protein kinase [
7,
8]. MELK is conserved across several species including
Xenopus (pEg3) [
9], murine (MPK38) [
7] and human (KIAA0175) [
10], and plays a key functional role in multiple cellular processes such as the proliferation, cell cycle progression, mitosis, and spliceosome assembly [
8,
11‐
15]. Molecularly, MELK interacts with and phosphorylates Ser323 of CDC25B to regulate G2/M progression [
8]. The zinc finger protein ZPR9 can also be phosphorylated by MELK to enable its translocation into the nucleus, where it interacts with B-Myb, leading to its increased transcriptional activity [
16]. Recent studies also show that MELK is frequently elevated in multiple human tumors such as prostate cancer [
17], breast cancer [
18], glioblastoma multiforme [
19] and medulloblastoma [
20], and is correlated with a poor prognosis [
21]. Indeed, MELK has recently emerged as an oncogene and a biomarker overexpressed in multiple cancer stem cells [
20,
22,
23], and so is considered a potential therapeutic target [
24,
25]. Knockdown of MELK inhibited proliferation, colony formation and survival of cancer stem cells [
20,
26]. In prostate cancers with high Gleason scores, MELK expression was elevated and its inhibition by RNAi detailed putative functions in chromatin modification, embryonic development, and cell migration [
17]. In breast cancer, MELK has been found to interact with Bcl-G
L through its amino-terminal region and suppress apoptosis [
18]. Study also implied that MELK was involved in the resistance of colorectal cancer cells to radiation and 5-FU [
27].
The FAK/Paxillin pathway plays an important role in cell migration and invasion [
28]. Upon activation of its upstream pathways, FAK binds SH2 domains of Src family kinases, which promotes Src kinase activity through a conformational change and then activates downstream signals to regulate cell motility, invasion, survival and proliferation [
29,
30]. Activated FAK can phosphorylate various adaptor proteins such as paxillin, which is a multidomain protein located in focal adhesion complexes and connects extracellular matrices to the cytoskeleton [
31,
32]. The paxillin signaling hub controls the dynamics of focal adhesion assembly and disassembly through protein interactions and phosphorylation events. The FAK/Paxillin pathway also regulates small Rho GTPases, an important family of small GTPases [
33]. These proteins, including RhoA, Rac1 and Cdc42, act as molecular switches that cycle between an active GTP-bound and an inactive GDP-bound forms, and play important roles in cytoskeletal reorganization [
34]. Paxillin phosphorylation leads to enhanced Rac1 activity and decreased RhoA activity [
35,
36]. In addition, recent studies have indicated that FAK signaling can promote matrix-degrading invasive behavior through a pathway involving the c-Jun NH2-terminal kinase and MMP-mediated pathways [
37].
Here we demonstrate that MELK expression is elevated in tumor-derived primary human gastric tissues compared to normal controls at both mRNA and protein levels. This enhanced expression of MELK is shown to be associated with pleiotropic effects in gastric cancer cells, including increased cell proliferation, migration, and invasion. Finally, we show that MELK can regulate RhoA activity and promote cell migration and invasion via the FAK/Paxillin pathway.
Discussion
Recent studies have shown that MELK plays an important role in tumorigenesis and tumor development [
8,
12,
17,
24]. However, the exact mechanism has not been established. Here, we showed that MELK expression was up-regulated in gastric tumors. MELK knockdown and overexpression models demonstrated its role in regulating cell proliferation, cell cycle progression, and chemoresistance to 5FU. Furthermore, we observed that MELK regulated the activity of RhoA and promote cell migration and invasion via the FAK/Paxillin pathway.
Many studies have shown that MELK is highly expressed in tumors and this expression is correlated with tumor grade and prognosis. For example, Gray et al. examined 3600 normal and 1701 cancer tissues by oligonucleotide microarray analysis, including breast, cervix, colorectal, esophagus, kidney, liver, and ovary cancers [
24]. Although MELK has been the focus of many cancer-related studies, most of these lacked data related to the protein level, and few have investigated gastric cancer. We examined here MELK expression in clinical tissue samples and cell lines, and found that MELK mRNA and protein expression were both elevated in tumor tissues. We also analyzed the correlation between MELK expression and clinicopathological parameters and found that MELK protein expression was higher in well differentiated and intestinal type gastric cancers.
Previous studies also suggested that MELK expression was elevated in cancer stem cells (CSCs) and could promote CSCs growth, differentiation and self-renewal [
11,
23,
39,
40]. In agreement with this earlier work, we found that MELK had wide-spread effects involving chemoresistance, cell proliferation, migration, invasion and cytoskeleton regulation in gastric cancer cells. We speculate that this is owing to the effect of MELK on gastric CSCs. In particular, knockdown of MELK dramatically suppressed tumor growth in vivo, although the effect was not so significant in vitro. Besides blocking the cell cycle, the main reason could be that MELK suppressed the proliferation of gastric CSCs. We also found that MELK expression was elevated after treatment with different concentrations of 5-FU, and MELK could regulate apoptosis induced by 5-FU. As chemoresistance is closely correlated with CSCs and MELK is generally regarded as a marker of CSCs, we propose that MELK expression might be elevated in gastric CSCs and is closely related to chemoresistance in gastric cancer. We also hypothesize that MELK might be a potential target for chemotherapy, but further study is needed to support this. Furthermore, it appeared as though the effect of MELK overexpression was not as dramatic as its knockdown. This might be due to the already relative high expression of MELK in gastric cancer. Interestingly, MELK knockdown and overexpression both resulted in increased cell populations in the G2/M phase. Several studies also found these contradictory results [
8,
24,
41]. In these, it was suggested to be owing to a delay of G2/M. In addition, MELK might be necessary for cells to through a G2/M checkpoint but its expression and activity are both suppressed once the cell has overcome this checkpoint.
The migration and invasion of cancer cells involves a host of processes and the interaction of multiple genes [
42]. Previous studies indicated that MELK plays an important role in tumors, but the mechanism was unclear. In this study we found that MELK knockdown or overexpression decreased or increased Tyr397, Tyr576/577, and Tyr925 phosphorylation of FAK and Tyr118 phosphorylation of paxillin, respectively. The FAK inhibitor and MELK inhibitor could both partly reverse the up-regulatory effect on migration and invasion caused by MELK overexpression. This indicated that MELK could be an upstream regulator of FAK. Furthermore, we found that MELK prevented gastrin-stimulated FAK and paxillin phosphorylation, which may be an important factor for tumorigenesis. These data strongly support the role of MELK in regulating cell migration and invasion via FAK/Paxillin pathway.
As MELK could regulate the phosphorylation of paxillin, we also analyzed its effect on the cytoskeleton by IFC. We found that MELK knockdown decreased the amount of actin stress fibers and filopodia, while MELK overexpression resulted in an increase in these structures. We speculated that it might be due to the elevation of Cdc42 and RhoA activity. The phosphorylation of FAK Tyr397 can promote Rac1 activity via the Crk/Dock180/ELMO complex [
33,
43]. Furthermore, phosphorylated Tyr31/Tyr118 of paxillin can bind to p120RasGAP, which releases the inhibitory interaction of p120RasGAP with p190RhoGAP and then suppresses RhoA activity [
36]. However, we found that knockdown of MELK inhibited RhoA activity, whereas its overexpression promoted RhoA activity. In addition, there was no effect on the activities of Rac1 and Cdc42, which seemed rather contradictory. Similarly, Zhai [
44] also found that the overexpression of FAK increased RhoA activity and p190RhoGEF phosphorylation in neuronal cells. Our data thus provide additional evidence that FAK could have both positive and negative effects on RhoA. Thus, the effect of MELK through RhoA or FAK/paxillin may affect cell migration and invasion.
In summary, our data indicate that MELK expression is elevated in gastric cancer. MELK plays an important role in the regulation of cell proliferation, cell cycle progression, chemoresistance, migration, invasion, and cytoskeleton assembly. Furthermore, MELK was found to promote cell migration and invasion via the FAK/Paxillin pathway, which could thus be a potential focus of future therapy against gastric cancer.
Methods
Ethical statement
Written informed consent was obtained from all participants. The study was approved by the Human Research Ethics Committee of Ruijin Hospital, School of Medicine, Shanghai Jiao Tong University (Permit number: HREC 08–028). All animal experiments were approved by the Laboratory Animal Ethics Committee of Ruijin Hospital (Permit Number: 2013062) and performed in accordance with the Guide for the Care and Use Laboratory Animals of Ruijin Hospital, School of Medicine, Shanghai Jiao Tong University.
Cell lines and chemicals
Gastric cancer cell lines SGC7901, BGC823, MKN45, MKN28, NCI-N87, AGS and an immortalized normal gastric epithelial cell line GES-1 were preserved in the institute. Cells were cultured in RPMI-1640 medium containing 10% fetal calf serum with 100U/ml penicillin and 100ug/ml streptomycin (GIBCO BRL) and maintained at 37°C in a humidified atmosphere of 5% CO2. FAK inhibitor and gastrin were from Millipore. MELK inhibitor OTSSP167 was from MedChem Express.
Tissues
Gastric cancer tissues were obtained from 150 patients who underwent radical gastrectomy between 2006 and 2008 at the Department of Surgery, Ruijin Hospital, Shanghai, China. All samples were confirmed by pathological diagnosis. All tissue samples were formalin-fixed and paraffin-embedded. Eighty pairs of tissue samples from patients were processed into tissue arrays and confirmed by a pathologist.
Plasmids construction and transfection
For MELK knockdown, the target sequence was 5′-GGATCTCAACCAAGCACATAT-3′, the negative control sequence was 5′-GTTCTCCGAACGTGTCACGT-3′. pGPU6/GFP/Neo (GenePharma) was used for shRNA plasmid construction. Plasmids were transfected into gastric cancer cells using Lipofectamine 2000 (Invitrogen). For 3′UTR of MELK knockdown, the target sequence was 5′-GCCTACATAAAGACTGTTA-3′, the negative control sequence was 5′-GTTCTCCGAACGTGTCACGT-3′. MELK cDNA ORF (Origene Technologies) was cloned into the pL/ERES/GFP plasmid (Novobio) for lentivirus production.
qPCR (quantitative-PCR)
Total RNA was isolated using Trizol reagent (Invitrogen) and cDNA was obtained using a reverse transcription kit (Promega) according to the manufacturer’s instructions. qPCR was performed using the Applied Biosystems 7900HT sequence detection system (Applied Biosystems) and Universal PCR Master Mix (Applied Biosystems). Relative expression was calculated with GAPDH using the 2-ΔCt and -ΔΔCt method. The primers for MELK were 5′-CATTAGCCCTGAGAGGCGGTGC-3′ (fwd) and 5′-GCCCGTCTCTGGCAGAACCCTT-3′ (rev). The primers for GAPDH were 5′-TTGGCATCGTTGAGGGTCT-3′ (fwd), and 5′-CAGTGGGAACACGGAAAGC-3′ (rev).
Immunohistochemistry staining
Immunohistochemistry (IHC) staining was performed as previously reported [
45]. Polyclonal anti-MELK was used at a dilution of 1:150 (Sigma). The slides were evaluated by a single board-certified pathologist (RRT) without clinicopathologic information. The percentage of positive cells was divided into five grades (percentage scores): <5% (0), 5-25% (1), 25-50% (2), 50-75% (3), 75-100% (4). The intensity of staining was divided into four grades (intensity scores): no staining (0), weak staining (1), moderate staining (2) and strong staining (3). MELK staining positivity was determined by the following formula: overall score = percentage score × intensity score. The overall score ≤ 3 was defined as negative, and >3 as positive.
Immunoblotting
Cells were lysed using RIPA cell lysis buffer (Kangwei) supplemented with protease inhibitor cocktail (Cell Signaling Biotechnology). The amount of total protein was quantified using a protein assay kit (Bio-Rad). Protein samples were loaded onto 12.5% SDS-PAGE gels and then transferred onto PVDF membranes. The membranes were blocked in TBS-T buffer containing 5% non-fat dry milk and hybridized with a primary antibody. Paxillin (Tyr118) antibody was from Abcam, GAPDH antibody was from Kangchen Bio-tech, and all other primary and secondary antibodies were from Cell Signaling Biotechnology. Finally, membranes were incubated with HRP-conjugated secondary antibody. Protein bands were visualized using ECL reagent (Thermo) on a Tanon detection system.
Cell proliferation assay
Cell proliferation was assayed using Cell Counting Kit-8. Cells were cultured in a 96-well plate at a concentration of 1 × 104 cells/ml; OD450 was measured 2 h after adding CCK-8 at 0, 1, 2, 3 and 4 days.
Flow cytometry
For cell cycle analysis, cells were harvested and fixed in 70% ice-cold ethanol at 4°C overnight and then incubated with 100 μg/ml RNase at 37°C for 20 min. After staining with 50 μg/ml propidium iodide, cell cycle analysis was performed by fluorescence flow cytometry on a FACScan machine (Beckman Instruments). For apoptotic analysis, cells were washed and stained using an Annexin V/PI double staining kit (BD Biosciences) according to the manufacturer’s protocol.
Cell migration, invasion and wound healing assays
Cell migration and invasion were analyzed using a transwell chamber assay (Corning). For migration, cells cultured with serum-free medium were added to the upper chamber and medium containing 10% fetal calf serum was added to the lower chamber. For the invasion assay, insert membranes were coated with diluted Matrigel (BD Biosciences). After culture, the insert membranes were fixed and stained with 0.1% Crystal violet. Permeating cells were visualized on an Olympus BX50 microscope (Olympus Opticol Co) and Nikon Digital Sight DS-U2 (Nikon). For the wound healing assay, cells were wounded with a pipette tip and then cultured with fresh DMEM medium containing 1% fetal calf serum. Wound closing was observed every 24 h.
Immunofluorescence staining
Cells were cultured on cover slips for 24 h. The coverslips were then washed with PBS and fixed in 4% paraformaldehyde for 15 min at room temperature. Monolayers were washed with PBS, then permeabilized with 0.5% Triton X-100 and blocked with 5% BSA for 1 h. To visualize the cytoskeleton and nuclei, cells were stained with rhodamine phalloidin antibody (1:150, Cytoskeleton) and 4′-6-diamidino-2-phenylindole (DAPI, 0.5 μg/ml). Images were acquired using an Olympus BX50 microscope (Olympus) and a Zeiss LSM510 confocal microscope (40X oil lens; Carl Zeiss).
Rho GTPase assay
Rho GTPases were measured using the Rho GTPases activation Assay Combo Biochem Kit (Cytoskeleton) according to the manufacturer’s instructions. Briefly, cells were washed with ice cold PBS and then lysed in ice cold lysis buffer. After quantification of protein concentrations, 650 μg of cellular extracts were incubated with 10 μg Rhotekin-RBD or PAK-PBD affinity beads. The beads were then pelleted and washed. After adding 2 × Laemmli of sample buffer, GTP-bound RhoA/Rac1/CDC42 was detected by immunoblotting.
Male BALB/c nude mice (Institute of Zoology, Chinese Academy of Sciences) were housed in a specific pathogen-free (SPF) environment. 1 × 10
6 cells were subcutaneously injected into twenty 4-week-old male nude mice (five mice each group) and 2 × 10
6 cells were intraperitoneally injected into forty 5-week-old male nude mice (ten mice each group). Tumor length (L) and width (W) were measured every 5 days with calipers and tumor volume was calculated using the equation: volume = (W + L)/2 × W × L × 0.5236 [
46]. Mice were sacrificed under anesthesia 30 days after injection. Tumor grafts were fixed, embedded and stained using MELK and Ki-67 antibody (Dako, dilution 1:50) by IHC. Furthermore, peritoneal nodules were visualized under microscope.
Statistical analysis
Results were shown as mean ± standard deviation (SD). Differences in frequency of MELK expression and the correlation with clinicopathological parameters were analyzed by the Pearsonχ2 test. Differences between experimental groups were assessed by the Student’s t test or one-way ANOVA. A two-tailed value of P < 0.05 was deemed as statistically significant. Statistical analyses were performed using SPSS 19.0 software (SPSS Inc).
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
TD, BL, ZZ and LS conceived and designed this work. TD, YQ, JL, HL, and QZ performed experiments and analyzed data. TD, BL, MY and CL interpreted the data and wrote the manuscript. All authors read and approved the final manuscript.