miR-429 suppresses tumor migration and invasion by targeting CRKL in hepatocellular carcinoma via inhibiting Raf/MEK/ERK pathway and epithelial-mesenchymal transition

Tumor metastasis is one of the main causes of hepatocellular carcinoma (HCC) high mortality. CRKL (v-crk sarcoma virus CT10 oncogene homologue (avian)-like) play important roles in tumor metastasis, however, the exact role and underlying mechanism of CRKL in HCC is still unknown. In our study, we demonstrated miR-429 negatively regulated CRKL expression via selectively binding to CRKL-3′-UTR at 3728–3735 bp site by post-transcriptionally mediating its functionality. Re-expression and silencing of miR-429 was remarkably effective in suppressing and promoting HepG2 cell migration and invasion in vitro. Knockdown or overexpression of CRKL exhibited similar effects as the overexpression or silencing of miR-429, whereas, CRKL overexpression (without the 3′-UTR) abrogated miR-429-induced inhibition on HepG2 migration and invasion. Moreover, miR-429-CRKL axis affected HepG2 migration and invasion potentials by regulating the adhesion ability, cytoskeleton F-actin expression and arrangement of HepG2. Furthermore, interference of Raf/MEK/ERK pathway and EMT contributed to miR-429-CRKL axis mediated metastasis inhibition. Nevertheless, miR-429 could not inhibit HepG2 proliferation through CRKL/c-Jun pathway. Taken together, our data demonstrated that miR-429 might function as an antimetastatic miRNA to regulate HCC metastasis by directly targeting CRKL via modulating Raf/MEK/ERK-EMT pathway. The newly identified miR-429-CRKL axis represents a novel potential therapeutic target for HCC treatment.

Consequently, miR-429 overexpression and miR-429 supression decreased endogenous CRKL level by 55.0 ± 1.5% (P = 0.0089, Fig. 1C) and increased endogenous CRKL level by 44.3 ± 1.0% (P = 0.0038, Fig. 1D) in HepG2 cells, respectively. While interestingly, miR-429 dysexpression showed no effect on the endogenous expression level of CRKL mRNA in HepG2 cells (Fig. 1C,D). These results suggested CRKL was a direct downstream target of miR-429 via direct binding to site 2 in its 3′-UTR by post-transcriptionally mediating its functionality. miR-429 expression is inversely correlated with CRKL expression. To further confirm CRKL is a target of miR-429, we further detected the endogenous expression levels of miR-429 and CRKL in 12 paris of matched HCC and corresponding nontumor liver tissues. WB results showed that CRKL protein expression levels were significantly higher in HCC tissues than in corresponding nontumor liver tissues (66.7%, P = 0.0362, Fig. 2A). Moreover, qRT-PCR results showed that miR-429 expression levels were significantly lower in HCC tissues than that in corresponding nontumor liver tissues (P = 0.0295, Fig. 2B). Meanwhile, we also observed a highly significant negative correlation between the expression level of miR-429 and CRKL (R 2 = 0.5905, P = 0.0035, Fig. 2C). Our results further demonstrated that miR-429 expression is negatively correlated with CRKL expression. miR-429 deregulation affects the in vitro migration and invasion of HepG2. The effects of miR-429 expression level change on the proliferation, colony forming, migration and invasion properties of HepG2 cells were measured by both down-regulating and up-regulating miR-429 in HepG2. A synthetic double-stranded miR-429 mimic and single-stranded LNA-miR-429 were transiently transfected into HepG2 cells to increase and decrease miR-429 expression. qRT-PCR showed miR-429 expression level was increased by 6373600 ± 10220% (P = 0.0034) in HepG2-miR-429 cells compared with HepG2-miR-NC (Fig. 3A) and decreased by 76.2 ± 16.7% (P = 0.0103) in HepG2-LNA-miR-429 cells compared with HepG2-LNA-miR-NC cells (Fig. 3B). miR-429 deregulation showed no influence on the proliferation ability determined by MTT (Fig. 3C) and colony formation ability determined by plate colony forming assay (Fig. 3D) of HepG2 cells. Our results indicated that miR-429 was uncritical for HepG2 growth.
CRKL promotes the proliferation and colony forming capacities of HepG2. CRKL level positively correlates with the in vitro proliferation of HepG2 cells. No proliferation difference was observed between HepG2-shNC and HepG2 cells by MTT assay. While, the stable CRKL knockdown resulted in decreased proliferation abilities by ~50.8% and 56.4% (P < 0.001, Fig. 7A) of HepG2-shCRKL cells at the inoculation time intervals of 72 h and 96 h than HepG2-shNC cells, respectively. Consequently, CRKL overexpression increased the proliferation of HepG2 cells. The relative proliferation capacities of HepG2-PCDH-CRKL cells were measured ~17.0% and 39.2% higher than HepG2-PCDH cells at the time intervals of 72 h (P < 0.05) and 96 h (P < 0.001, Fig. 7B). No difference was determined between HepG2-PCDH and HepG2 cells (Fig. 7B).  CRKL expression level is also positively associated with the colony forming ability of HepG2 cells. Plate colony formation assay indicated that CRKL downregulation reduced the colony forming efficiency of HepG2 cells. The number of formed colonies of HepG2-shCRKL cells was 123.3 ± 6.0 per view that was only about 32.3% of that of HepG2-shNC cells (382.0 ± 14.8, P < 0.0001, Fig. 7A). The colony size of HepG2-shCRKL cells was smaller than HepG2-shNC cells. Comparable colony forming capacity was detected for HepG2-shNC and HepG2 cells. Consistently, CRKL overexpression improved colony forming capacity of HepG2, the number of HepG2-PCDH-CRKL cells was 601.7 ± 7.5 per view that was 57.2% higher than that of HepG2-PCDH cells (382.7 ± 9.5, P < 0.0001, Fig. 7B). The size of colony of HepG2-PCDH-CRKL cell was enlarged than HepG2-PCDH due to CRKL overexpression. No difference was determined between HepG2-PCDH and HepG2 cells.

CRKL overexpression reversed miR-429-induced inhibition on HepG2 migration and invasion.
Is CRKL a direct functional mediator of miR-429-inhibited migration and invasion for HepG2 cells? Could the in vitro phenotypes associated with miR-429 deregulation be reversed by opposite CRKL expression? We performed a "rescue" experiment by co-tansfecting miR-429 mimic and PCDH-CRKL vectors into HepG2 cells. miR-429 could downregulate endogenous CRKL expression level in HepG2 cells, while, it could only downregulate the CRKL expression level in HepG2 cells co-transfected with miR-429 and PCDH-CRKL vector to certain extent ( Fig. 8A) as PCDH-CRKL expression vector does not have 3′-UTR domain. Meanwhile, exogenous CRKL overexpression markedly counteracted the suppression effect of miR-429 on the migration and invasion capacities of HepG2 cells (Fig. 8B, P < 0.001). The above results proved that miR-429 inhibited HepG2 migration and invasion by downregulating CRKL through binding with its CRKL-3′-UTR. CRKL was a functional effector of miR-429 mediated anti-tumor effect.

miR-429-CRKL axis mediates tumor cell migration and invasion via Raf/MEK/ERK and EMT.
The underlying molecular mechanism of miR-429-CRKL axis in HCC is unknown. Current work linked miR-429-mediated action on biological behaviours of HepG2 via Raf/MEK/ERK and EMT by targeting CRKL.
We found CRKL knockdown or miR-429 overexpression resulted in the similar trend on downregulating the protein expression levels of Raf, p-Raf, p-MEK and p-ERK2 in HepG2 cells (Fig. 9A). Consistently, CRKL overexpression or miR-429 silencing increased the protein expression levels of Raf, p-Raf, p-MEK and p-ERK2. No changes were observed for Ras and ERK1/2 (Fig. 9A). Clearly, miR-429-CRKL axis mediates the biological properties of HepG2 via Raf/MEK/ERK signaling pathway.
miR-429-CRKL axis influence the EMT property of HepG2 cells. We examined the expression level changes of EMT markers, E-cadherin, N-cadherin and Vimentin in HepG2 cells following CRKL or miR-429 dysexpression. qRT-PCR assay showed CRKL knockdown or miR-429 overexpression increased the expression level of epithelial marker E-cadherin (Fig. 9B), and decreased the expression levels of mesenchymal marker N-cadherin (Fig. 9C) and Vimentin (Fig. 9D). CRKL overexpression or miR-429 silencing decreased E-cadherin expression (Fig. 9B),  c-Jun signaling pathway is critical for cell growth. CRKL knockdown decreased and overexpression increased the mRNA level of c-Jun, while miR-429 overexpression and silencing showed no influence on c-Jun expression (Fig. 10). Interestingly, CRKL knockdown or overexpression significantly decreased or increased the protein expression levels of c-Jun, p-c-Jun (73), p-c-Jun (243), JNK, p-JNK by 57.7%, 56.1%, 45.2%, 3.6-fold or 4.4-fold, 6.0-fold, 3.6-fold, 78.7%, 1.9-fold (Fig. 10). While, miR-429 overexpression or silencing only slightly decreased or increased the protein levels of c-Jun, p-c-Jun (73), p-c-Jun (243), JNK, p-JNK by 26.1%, 43.2%, 27.5%, 12.1%, 35.1% or 23.0%, 49.2%, 52.3%, 12.7%, 15.9% (Fig. 10). Our results showed that the deregulate degree of miR-429 on c-Jun pathway was significantly weaker than CRKL, this difference might contribute to explain our previous results that the overexpression or silencing of miR-429 rarely influenced the proliferation and colony formation  The linkages of Raf/MEK/ERK pathway and EMT to miR-429-and/or CRKL-mediated tumor migration and invasion were further validated using PD98059, a specific pERK1/2 inhibitor, for specific signaling blocking. The treatment of HepG2-miR-429 and HepG2-PCDH-CRKL cells with 20 μg/ml PD98059 resulted in reduced protein expression of pERK2 (Fig. 11A), reduced mRNA levels of N-cadherin and Vimentin, and increased mRNA level of E-cadherin (Fig. 11B). As in Fig. 11C, miR-429 overexpression significantly inhibited the migration and invasion potentials of HepG2 cells, which were further enhanced in the presence of PD98059. The presence of PD98059 tremendously reversed the promotion of CRKL overexpression on the migration and invasion potentials of HepG2 cells (Fig. 11D). These results concluded miR-429 and CRKL regulates HepG2 cell migration and invasion via Raf/MEK/ERK-EMT pathway, which also revealed the mutual influence relationship between miR-429 and CRKL, providing a novel insight for investigating the combined power of miR-429 and CRKL in HCC field.

Discussion
miRNAs are attractive candidates for human malignancies 27 . Considerable attention is globally attracted in understanding the role of miRNAs in the tumorigenesis, diagnosis, classification and prognosis of a variety of miR-429 abnormal expression has been linked to osteosarcoma, renal cancer, ovarian cancer, glioma, breast cancer, oral squamous cell carcinoma, gastric cancer, esophagus cancer, cervical cancer, bladder cancer, lung cancer, prostate cancer, colon cancer and liver cancer 30 . It shows suppression or promotion effects on the tumor development, invasion, metastasis, apoptosis and drug-resistance 31 depending on the tumor type and subtype. It is a potential indicator for the diagnosis, treatment and prognosis of certain tumors.
As a mutli-functional adaptor protein in signal transduction, CRKL misregulation is involved in a variety of cancers 21 . It is an attractive target for diagnostics, treatment and prognosis of certain cancers 22,32 . Previously reported that CRKL was a potential target gene of miR-429, miR-429 reduced CRKL protein expression in breast cancer MDA-MB-231 cells 25 , further study found miR-429 directly targeting to CRKL-3′-UTR by luciferase reporter assay in cervical cancer cells 26 , indicating CRKL was a direct target gene of miR-429. However, the exact role and underlying molecular mechanism of miR-429-CRKL axis in HCC is still unknown. In our study, we reported miR-429 directly binding to site 2 in CRKL-3′-UTR regulated HCC cell migration and invasion via Raf/ MEK/ERK-EMT pathway.
We found CRKL with two putative binding sites of 1898-1904 and 3728-3735 at CRKL-3′-UTR region for miR-429 by bioinformatics analysis, our results showed miR-429 directly targeting to CRKL-3′-UTR at 3728-3735 bp site (Fig. 1B), which was consistent with previous report 33 . miR-429 could decrease CRKL protein expression level in breast cancer 25 and cervical cancer cells 26 , consistently, our results showed miR-429 overexpression decreased endogenous CRKL protein expression level (Fig. 1C) and miR-429 silencing increased endogenous CRKL protein expression level (Fig. 1D) in HepG2 cells. However, the effect of miR-429 deregulation on CRKL mRNA expression level is still unknown, we found miR-429 dysexpression showed no effect on endogenous CRKL mRNA expression level in HepG2 cells (Fig. 1C,D). The results suggested CRKL was a direct downstream target of miR-429 via direct binding to site 2 in its 3′-UTR by post-transcriptionally mediating its functionality. Our "Rescue" experiments indicated miR-429 could not downregulate exogenous CRKL expression on account of PCDH-CRKL without 3′-UTR, which further verified miR-429 directly targeting CRKL-3′-UTR to downregulate its expression. Our results demonstrated miR-429 could negatively regulate CRKL expression in vitro, we also further demonstrated CRKL protein expression inversely correlates with miR-429 expression in HCC tissues (Fig. 2). Previous studies have reported that CRKL could affect cancer cell proliferation 21 . In our study, CRKL also indeed affected HepG2 cells proliferation and colony abilities, CRKL knockdown significantly inhibited HepG2 cells proliferation and colony abilities (Fig. 7A), while, CRKL overexpression promoted HepG2 cells proliferation and colony abilities (Fig. 7B). However, re-expression or silencing of miR-429 cannot affect HCC cells  (Fig. 3C) and colony formation abilities (Fig. 3D). CRKL was only downregulated or upregulated about 55% (Fig. 1C) and 44% (Fig. 1D) by miR-429 mimics or inhibitors, respectively, while, CRKL downregulated or upregulated about 97% or 127% in our obtained monoclonal HepG2-shCRKL or transiently transfected HepG2-PCDH-CRKL cells, we speculate the underlying cause of this phenomenon is the deregulate degree of CRKL expression level was not enough, indicating CRKL only deregulated to a certain degree, which can affect HepG2 cells proliferation ability. We further found that CRKL knockdown or overexpression significantly decreased or increased the protein levels of c-Jun, p-c-Jun (73), p-c-Jun (243), JNK, p-JNK, while, miR-429 overexpression or silencing only slightly decreased or increased the protein levels of c-Jun, p-c-Jun (73), p-c-Jun (243), JNK, p-JNK (Fig. 10). We think the deregulate degree of CRKL by miR-429 was not enough to significantly affected on c-Jun pathway, so the deregulate degree of miR-429 on c-Jun pathway was significantly weaker than CRKL, indicating the deregulate degree of miR-429 on c-Jun pathway was not enough to regulate the proliferation and colony formation abilities of HepG2 cells. Although miR-429 was confirmed to target CRKL directly, we think that the inhibitory effect of miR-429 on HepG2 cells proliferation through the CRKL/c-Jun pathway is not powerful.
In contrast, CRKL contributed to miR-429-mediated metastasis inhibition. We demonstrated that re-expression of miR-429 was remarkably effective in suppressing HepG2 cells migration and invasion in vitro (Fig. 3E), in addition, silencing of miR-429 was significantly promoted HepG2 cells migration and invasion abilities (Fig. 3F). Furthermore, CRKL knockdown or overexpression exhibited similar effects as the overexpression or silencing of miR-429 (Fig. 3C,D). Adhesion is an important step in tumor metastasis, miR-429 overespression inhibited the in situ lymph node and extracellular matrix FN adhesion potentials (Fig. 4), consistently, CRKL knockdown also inhibited the in situ lymph node and extracellular matrix FN adhesion potential (Fig. 6). F-actin depolymerization is associated with cell movement, our results showed miR-429 overexpression resulted in an obvious decrease of F-actin microfilament (Fig. 4), meanwhile, CRKL knockdown also resulted in obviously decreased of F-actin microfilament (Fig. 6). Our results demonstrated miR-429-CRKL axis affected HepG2 cell migration and invasion potentials by regulating the adhesion ability, cytoskeleton F-actin expression and arrangement of HepG2 cell.
Are the miR-429-mediated influences on HepG2 properties linked to CRKL dysexpression? Is CRKL a direct functional mediator of miR-429 on HepG2 cells migration, invasion and associated in vitro phenotypes? The "rescue" experiment found miR-429 could not downregulate exogenous CRKL expression on account of PCDH-CRKL without 3′-UTR (Fig. 8A). Meanwhile, CRKL overexprssion counteracted the inhibition effect of miR-429 on HepG2 cells migration and invasion (Fig. 8B). The above results once again proved that CRKL is a functionally relevant effector of miR-429 mediated antimetastatic effect, and miR-429 inhibited HepG2 cells migration and invasion only by directly targeting CRKL-3′-UTR to downregulate its expression. miR-429-CRKL axis, a new antimetastatic regulator for HCC shown to significantly suppress HCC invasion and metastasis in vitro and it maybe a novel potential therapeutic target for HCC treatment. These findings open a novel avenue to investigate the molecular mechanism of HCC progression and to develop potential therapeutics against HCC. However, no studies have reported a definitive mechanism for miR-429-CRKL axis in HCCs. Our results showed that miR-429 and CRKL affected the expression level of Raf/MEK/ERK pathway-and EMT-related molecules (Fig. 9), we speculated miR-429-CRKL axis might mediate tumor migration and invasion via regulating Raf/MEK/ERK-EMT pathway. We validated the potential involvement of Raf/MEK/ERK-EMT using a specific ERK inhibitor PD98059 (Fig. 11A), the expression level of E-cadherin increased and N-cadherin, Vimentin decreased after blocking Raf/MEK/ERK pathway by ERK inhibitor PD98059 (Fig. 11B), indicating Raf/MEK/ ERK pathway could regulate EMT. Furthermore, PD98059 exacerbated the effect of miR-429 overexpression on HepG2 cells migration and invasion abilities (Fig. 11C), PD98059 weakened the effect of CRKL overexpression on HepG2 cells migration and invasion abilities (Fig. 11D). The results indicated that miR-429-CRKL axis regulated HepG2 cell migration and invasion via Raf/MEK/ERK-EMT pathway.
Current work showed the deregulation of CRKL or miR-429 could induce both the expression level changes of Raf and p-Raf in HepG2 cells. In fact, except for the involvement of expression level change of p-Raf in a canonical regulation fashion in diseases, the differential expression of total Raf in human tumors or in cancer cells induced by other molecule has commonly been regarded as signal transduction change from Raf/MEK/ ERK. In a lot of published papers, if the up-stream molecules and down-stream molecules for Raf or p-Raf were found deregulated in the detection systems, the authors commonly declared that investigated genes, proteins, miRNA or drugs worked through Raf pathway [34][35][36][37] . Raf is the key point in the Raf/MEK/ERK pathway, the Raf family consists of three isoforms: A-Raf, B-Raf, and Raf-1, much of the unprecedented research involving Raf was carried out on Raf-1 38 . It was reported that the overexpression of Raf-1 was significantly associated with a shorter DFS (disease free survival) and a poor OS (overall survival) of HCC patients as an independent risk factor for HCC recurrence and death 36 . These reports were consistent with our results obtained from current work, and CRKL-deregulation associated with expression changes of Raf and p-Raf in CML patients' specimens and K562 cells (unpublished) as well as in ccRCC patients' tissues and 786-O and ACHN cells (unpublished). In gastric cancer (GC), by targeting SLC34A2, the overexpression of miR-939 in GC cells significantly decreased Raf-1, p-MEK1/2 and p-ERK expression levels for suppressing the chemoresistance and metastasis of GC 34 . MiR-7 was reported to be able to inhibit glioblastoma growth by simultaneously interfering with the PI3K/ATK and Raf/ MEK/ERK pathways. MiR-7 markedly decreased the expression of Raf-1 39 . Moreover, compared with normal tissues, Raf-1, ERK and p-ERK were overexpressed in malignant breast tumor tissues. MiR-195 and miR-497 inhibited corresponding breast cancer cells proliferation and invasion via decreasing Raf-1, ERK, p-ERK expression levels 35 . Taken together, our results highly implicated that CKRL or miR-429 deregulation affected Raf/MEK/ ERK signal transduction in mediating the metastatical behaviours of HepG2. Current work provides certain new clue to the regulation mechanism for miR-429-CRKL functionality in tumorigenicity. In conclusion, we demonstrated that miR-429 could significantly inhibit HCC cell invasion and metastasis by targeting CRKL, which is a functional target of miR-429. The novel action mechanism was outlined in Fig. 12. miR-429 downregulated CRKL expression at the post-transcriptional protein translation level by directly targeting its 3′-UTR, CRKL downregulation inhibited the expression level of Raf, p-Raf, p-MEK and p-ERK2, and then suppressed migration and invasion through inhibiting EMT by increasing the epithelial marker E-cadherin expression, and decreasing the mesenchymal marker N-cadherin and Vimentin expression; Meanwhile, CRKL promoted tumor cell proliferation through upregulating the expression level of c-Jun; Nevertheless, miR-429 could not inhibit tumor cell proliferation through the CRKL/c-Jun pathway. The newly identified miR-429-CRKL axis provides new insight into the pathogenesis of HCC and represents a potential therapeutic target for diagnosis and treatment of HCC.

Materials and Methods
Patients and tissue samples. A total of 12 pairs of frozen human HCC tissues and their coprresponding non-tumor liver tissues were collected from Division of Hepatobiliary and Pancreatic Surgery, Department of Surgery, The Second Affiliated Hospital of Dalian Medical University, Dalian, China. None of the patients had received radiotherapy or chemotherapy before surgery. There were 8 men and 4 women, 5 patients with the age of ≥ 60 years, 7 patients with the age of < 60 years. Tissue specimens were frozen in liquid nitrogen immediately after surgical resection and stored at −80 °C until use. The use of human tissues and study protocol was approved by the Medical Ethics Committee of Dalian Medical University, and informed consent was obtained from all patients. All experiment methods were performed in accordance with the relevant guidelines and regulations.
To overexpress CRKL in HepG2 cells, the full-length coding sequence of CRKL was first amplificated by RT-PCR using the forward primer 5′-TATCTAGAGCCACCATGTCCTCCGCCAGGTT-3′ and reverse primer 5′-CGTTCGAAGGGCCTCGTTTTCATCTGGGTTT-3′, then cloned into the BstB I and Xba I sites of PCDH-EF1-MCS-T2A-Puro vector. The recombinant PCDH-EF1-MCS-T2A-Puro-CRKL expression vector was used for overexpressing CRKL in HepG2 cells. MTT and colony forming assays. The influences of miR-429 and CRKL level changes on HepG2 proliferation was determined by MTT assay. The cells from each group were seeded into a 96-well plate at 3 × 10 3 cells/well in 100 μl RPMI-1640 supplemented with 15% FBS, incubated at 37 °C, 5% CO 2 for 24, 48, 72 and 96 h separately, incubated in MTT solution (5 mg/ml) by replacing culture medium at 37 °C, 5% CO 2 for 4 h in darkness. The supernatants were then removed and 150 μl DMSO was added to dissolve for formazan crystals. The absorbance at 492 nm were measured using a microplate reader (Thermo, USA) for cell density quantification. Results were obtained from triplicate measurements.
The deregulation effects of miR-429 and CRKL on HepG2 colony forming ability was performed by colony formation assay. The dissociated 1000 cells in 2 ml RPMI-1640 supplemented with 15% FBS were seeded in 6-well plate and then kept at 37 °C with 5% CO 2 for 10 d until visible cell colonies appeared. The colonies were washed with PBS, fixed with methanol for 30 min and stained with 0.05% crystal violet for 1 h at RT. The colonies were observed and larger than 50 cells were counted using an upright light microscope (Olympus, Japan) with magnification of 100x. Triplicate experiments were performed for each assay.
In vitro cell migration and invasion assays. The effect of miR-429 and CRKL deregulations on the migration and invasion abilities of HepG2 cells were examined using 24-well-plate transwell chamber assay. Briefly, 1 × 10 4 cells in 200 μl serum-free RPMI-1640 were seeded onto the upper compartment of transwell with 8 μm pore size polycarbonate filters (Corning, USA). The chambers were then placed into 24-well plates containing 600 μl RPMI-1640 with 20% FBS and incubated for 48 h at 37 °C with 5% CO 2 . For invasion assay, the inserts were coated with 50 μl extracellular matrix gel (ECM, Sigma, USA) which 1:8 dilution with RPMI-1640, and incubated at 37 °C for 1 h. 1 × 10 4 cells in 200 μl serum-free RPMI-1640 were seeded onto the upper compartment of transwell. The chambers were then placed into 24-well plates containing 600 μl RPMI-1640 with 20% FBS and incubated for 48 h at 37 °C with 5% CO 2 . The non-migrated and non-invaded cells on the upper surface of the insert were removed by swabbing, the migrated and invaded cells onto the lower surface were fixed with methanol for 30 min, stained with 0.5% crystal violet for 1 h, washed with PBS, counted using an upright light microscope (Olympus, Japan) with 100x. Five field views were randomly counted and averaged.
Extracellular matrix adhesion assay. The effect of miR-429 and CRKL deregulations on the adhesion capacity of HepG2 to fibronectin (FN, Millipore, USA) was examined. Each well of a 96-well plate were first coated with 50 μl 50 μg/ml FN by incubating at RT for 1 h and blocked with 50 μl 1% BSA for 1 h at 37 °C. 1 × 10 3 cells in 100 μl RPMI-1640 containing 15% FBS were seeded into each well of a 96-well plate and cultured at 37 °C, 5% CO 2 for 2 h. The non-adherent cells were washed with PBS, incubated in 5 mg/ml MTT solution at 37 °C with 5% CO 2 in darkness for 4 h. Then, 150 μl DMSO was added into each well to dissolve formazan. The absorbance at 492 nm was measured using a microplate reader (Thermo, USA) for cell number quantification. The results were averaged from triplicate independent experiments. In situ cell adhesion potential to lymph node (LN) assay. The effect of miR-429 and CRKL on HepG2 adhesion ability to LNwas examined by in situ cell adhesion assay. Fresh LNs taken from 615 mice were frozen at −20 °C, embedded by optimum cutting temperature compound (OCT) and sectioned into 10 μm slices. The frozen LN slices were merged in 200 μl RPMI 1640 supplemented with 15% FBS containing 2 × 10 5 cells from each group, incubated at 37 °C with 5% CO 2 for 24 h, washed with ice-cold PBS for 3 times, fixed in 95% ethanol for 5 min, washed with distilled water for 3 s and stained with hematoxylin eosin (HE). The number of adherent cells was counted using an upright light microscope (Olympus, Japan) with 100 × by randomly selecting five fields.
F-actin cytoskeleton staining assay. FITC-Phalloidinstaining assay was performed to investigate the influence of miR-429 and CRKL on the cytoskeleton structure of HepG2 cells. The 0.17 mm thick round cover glasses with 25 mm in diameter were placed into 6-well plates, then 3 × 10 4 cells from each group were seeded into cover glasses and cultured for 24 h at 37 °C in humidified incubator with 5% CO 2 . The cover glasses were then washed with PBS for 2 times, fixed in 4% paraformaldehyde for 10 min at RT, washed with PBS for 3 times, permeabilized with acetone, and continuously incubated in 200 μl 400 nM Phalloidin-FITC containing 1% BSA for 30 min in the dark at RT. The cover glasses were washed with PBS for 3 times and immediately imaged for examining the actin cytoskeleton organization and filaments using an oli microscope (Olympus, Japan) at five randomly selected visual fields of 1000x.
Quantitative real-time RT-PCR(qRT-PCR) assay. Total RNA was extracted from each group cells using Trizol TM reagent (Invitrogen, USA) and reversely transcribed into cDNA using PrimeScript TM RT Kit with gDNA Eraser (Takara, Japan). qRT-PCR was then performed using FastStart universal SYBR Green Master (ROX) (Roche, USA) with an Applied Biosystems StepOne TM Real-Time PCR System (Life, USA). snRNA U6 and β-actin (ACTB) were used as internal references for miR-429 and mRNA, respectively. The relative expression levels of miR-429 and targeting genes among different cell lines, and in paired tumor and non-tumor tissues were comapred using 2 −△△CT method. The specific primers for CRKL, c-Jun, E-cadherin, N-cadherin, Vimentin and ACTB were provided in Table 1.
Data processing and statistical analysis. The data were processed as mean ± standard deviations of at least three independent experiments. The differences between groups were evaluated by unpaired Student's t-test analysis. Differences with the values of P ≤ 0.05 were significant.  Table 1. Synthesized sequences of primers for targeting genes.