Introduction
Hepatocellular carcinoma (HCC) is one of the most common malignancies and one of the leading causes of cancer-related death worldwide [
1,
2]. Despite recent advances in disease management and treatment, HCC patients still have a very dismal long-term prognosis. For advanced-stage HCC patients, the overall 5-year survival rate is less than 5% [
3]. The main challenges in the treatment of HCC involve intrahepatic recurrence and metastasis, which simultaneously predict poor outcome for HCC patients. The identification of critical players that suppress these processes may lead to novel therapeutic targets for improving the prognosis of these patients. Various previous studies have identified some key signaling transduction cascades that are implicated in the progression, invasion, and metastasis of HCC, such as the EGFR/PI3K pathway [
4], the RhoGTPase/Rho-effector pathway [
5], the SAPK/JNK pathway [
6], and the Ras/MAPK pathways [
7]. However, the underlying molecular mechanisms of HCC metastasis are far from fully understood.
In the past decade, increasing attention has been paid to the important biological and pathological roles of microRNAs (miRNAs), a type of small conserved RNA molecule of approximately 17–22 nucleotides in length [
8,
9]. As post-transcriptional regulators of gene expression, these small non-coding RNAs complementarily bind to the 3′ untranslated region (3′-UTR) of their target messenger RNAs (mRNAs), leading to either the degradation of mRNAs or the inhibition of protein translation. miRNAs are involved in the regulation of most cellular processes, including cell proliferation, migration, and apoptosis [
10]. In recent years, numerous studies have shown that aberrant expression of miRNAs is associated with the development and progression of various types of cancer, including HCC [
10], and some of these miRNAs function as tumor suppressor genes or oncogenes [
11,
12]. MiR-146a is located on human chromosome 5q34, which is a region that is often deleted in human tumors [
13], and has been reported to be aberrantly expressed in several cancers. For example, it is downregulated in breast [
14], hormone-refractory prostate cancer [
15], and pancreatic cancer [
16], which indicates that miR-146a is a potential tumor suppressor. MiR-146a has also been shown to inhibit cancer cell metastasis by targeting IRAK-1 [
17‐
19]. However, it remains unknown whether miR-146a is involved in the regulation of HCC cell invasion and how this miRNA mediates invasive inhibition.
In this study, we discovered that miR-146a is downregulated in human HCC and likely suppresses hepatoma cell invasion and metastasis by downregulating VEGF expression through 2 signaling pathways. First, miR-146a mediates the upregulation of APC, which leads to the inhibition of the nuclear accumulation of β-catenin, and second, this miRNA inhibits HAb18G/CD147 (HAb18G) expression, a direct target of miR-146a, which consequently downregulates NF-κB p65. These findings provide a framework for better understanding the pathogenesis of HCC.
Materials and methods
Ethics statement and clinical specimens
Frozen and paraffin-embedded tissues from 53 HCC cases were collected from XiJing Hospital of Fourth Military Medical University, with patients provided prior consent and approval of Fourth Military Medical University. The study protocol was approved by the Ethics Committee of the Fourth Military Medical University. The sample enrollment criteria included those with HCC diagnosed by two independent pathologists, detailed information on clinical presentation and detailed follow-up data for at least 3 years. All patients had been treated with a combination of surgery and platinum-based chemotherapy.
DNA isolation and bisulfite treatment and BS-PCR DNA Sequencing
DNA was isolated using the QIAGEN DNAesasy Blood and Tissue kit following the guidelines of the manufacturer. DNA was modified sodium-bisulfite using the EpiTect Bisulfite Kit (QIAGEN) according to the manufacturer’s instructions. The CpG sites of miR-146a promoter region were amplified by PCR containing bisulfite DNA template, Hotstartaq, bisulfite specific primers following established procedure. The detail information of primers is listed in Additional file
1: Table S1. After PCR reaction. DNA fragment were purified using the Cycle pure kit from Omega bio-tek. The purified PCR fragments were cloned into pGEM-T easy vector (Promega), and individual clones were sequenced.
Cell culture and nuclear-cytoplasmic fractionation
The human HCC cell lines SMMC-7721, Huh-7, and HepG2 and the human embryonic kidney cell line 293 T (HEK 293 T) were purchased from Shanghai Institute for Biological Sciences (Shanghai, China) and routinely cultured in Dulbecco’s Modified Eagle Medium (DMEM) (GIBCO BRL, Grand Island, NY) containing 10% fetal bovine serum (Hyclone Laboratories, Logan, UT) at 37°C in a humidified atmosphere of 5% CO2. Nuclear-cytoplasmic fractionation was performed using Nuclear and Cytoplasmic Extraction Reagents (Beyotime, Wuhan, China) according to the manufacturer’s protocol.
RNA extraction and quantitative real-time polymerase chain reaction (qPCR)
Total RNA was isolated from cultured cell lines and tissue samples using TRIzol Reagent (Invitrogen, CA), according to the manufacturer’s instructions. For the enrichment of normal liver cells, paraffin-embedded normal liver tissue samples from 11 hepatic hemangioma patients were microdissected on HE-stained sections using laser capture microdissection under a microscope (Leica, LMD7000), according to the method described previously [
20]. Normal liver tissues were also collected, and subsequent RNA isolation was performed using the NucleoSpin® FFPE RNA kit (Macherey-Nagel) according to the manufacturer’s protocol.
Reverse transcription (RT) reactions were performed using the Prime-Script RT reagent kit (TaKaRa, Dalian, China). The qPCR was performed using a SYBR Premix Ex Taq™ II kit (Takara). For miR-146a detection, mature miR-146a was reverse transcribed, with specific RT primers that were designed as previously described [
21], and normalized by RUN6. The sequences of the PCR and RT primers are listed in Additional file
1: Table S1.
Pathway-specific expression array
The human tumor metastasis RT2 profiler PCR array PAHS-028Z (QIAGEN) was used to assess the effect of miR-146a on the expression of 84 known metastasis-related genes. Total RNA was isolated either from miR-146a-transfected SMMC-7721 cells or from negative control-transfected SMMC-7721 cells using the HP Total RNA kit (Omega bio-tek). Total RNA was reverse transcribed into cDNA using the RT2 First Strand Kit (QIAGEN) with RT2 qPCR master mix containing SYBR Green (QIAGEN), according to the supplier’s instructions. The real-time PCR reaction was performed with a CFX384 real-time system (BIO RAD). Four reference genes with the lowest standard deviations across replicates were used in the analysis (B2M, GAPDH, HPRT1, and RPLP0). Expression profiles were obtained from 3 independent experiments.
Cell apoptosis and cell cycle analysis
For cell apoptosis analysis, the cells were resuspended in 1× Binding Buffer, and 5 μl of Annexin FITC Conjugate and 10 μl of Propidium Iodide Solution were separately added to each cell suspension. The stained cells (1 × 105) were then analyzed using a FACScalibur flow cytometer (BD Biosciences).
For cell cycle analysis, the cells were harvested by trypsinization, washed in ice-cold phosphate-buffered (PBS), and then fixed in 70% ice-cold ethanol. The cells were washed with PBS and resuspended in Staining Solution (50 μg/mL propidium iodide, 1 mg/mL of RNase A, and 0.1% Triton X-100 in PBS). Cell cycle profiles of 2 × 105 cells were analyzed using a FACScalibur flow cytometer (BD Biosciences).
Cell proliferation assay
SMMC-7721, HepG2, and Huh-7 cells were plated in 96-well plates at 3 × 103 cells per well after transfection with miR-146a or anti-miR-146a for 48 h. Cells were cultured for 24, 48, 72, and 96 h, and then the absorbance at 450 and 630 nm was measured after incubation with 10 μl of WST-1 (Roche Applied Science, Indianapolis, IN, USA). Each assay was performed in triplicate.
Immunofluorescence analysis
Cells were fixed in 4% paraformaldehyde at room temperature for 15 min, followed by permeabilization in 1 × PBS containing 0.2% Triton X-100 for 10 min at room temperature. The cells were blocked in blocking solution (1 × PBS containing 10% normal goat serum) for at least 2 h. After being briefly washed, the cells were incubated with a rabbit anti human β-catenin or NF-κB p65 antibody at 4°C for overnight. After being washed, a goat anti-rabbit IgG conjugated with Alexa Fluor 488 (Invitrogen) was incubated with the cells at room temperature for 1 h. DAPI (Sigma) was used to stain the DNA. The fluorescence images were acquired using a laser scanning microscope (A1, Nikon, Japan).
Immunohistochemical staining
Fresh-frozen tissue samples were formalin-fixed, paraffin-embedded, and then cut into 5-μm sections. Immunohistochemical staining for the HAb18G [
22], APC and VEGF was performed, and the expression level was independently evaluated by 2 senior pathologists according to the proportion and intensity of positive cells. A score of 0 (no staining), 1 (any percentage with weak intensity or <30% with intermediate intensity), 2 (>30% with intermediate intensity or <50% with strong intensity) or 3 (>50% with strong intensity) was assigned to each sample.
Vector construction
The expression vectors for miR-146a (pcDNA3.1-miR-146a) and HAb18G (pcDNA3.1-HAb18G) were constructed by cloning the miR-146a precursor sequences amplified from the genomic DNA of SMMC 7721 cells and the HAb18G open reading frame (ORF) sequence (without 3′-UTR) amplified from the total RNA of SMMC 7721 cells into the multiple cloning site of the mammalian expression vector pcDNA3.1 (Invitrogen, CA). In addition, we generated 2 luciferase reporter vectors with a wide-type (pGL3-HAb18G wt) or mutant (pGL3-HAb18G mut) 3′-UTR fragment of HAb18G inserted downstream of the stop codon of the firefly luciferase gene in the pGL3 vector (Promega). The 3′-UTR fragment of HAb18G for both pGL3-HAb18G wt and pGL3-HAb18G mut contained putative binding sites for miR-146a, whereas the mutant 3′-UTR of pGL3-HAb18G mut carried a mutated sequence in the complementary site for the seed region of miR-146a. The sequences of all primers used in vector construction are provided in Additional file
1: Table S1. All constructed vectors were confirmed by direct sequencing.
Wound migration and invasion assay
Cells transfected with miR-Ctrl/miR-146a, or nonrelative control moleculars (NC)/antagomiR-146a were plated in 24-well culture plates at a density of 105 or 2 × 105 (Huh-7, hepG2) cells per well, and incubated for 24 h or 48 h (Huh-7, hepG2) to reach confluence. Using a 200 μl tip, a wound was made in the monolayer (at time 0). The cells were then washed with PBS and incubated. The distance between the two sides of the wound was measured with an Olympus CX71 microscope (Olympus). The distance between the two sides of the wound after 20-72 h of migration was divided from the distance at time 0 and represented on a graph.
The invasion abilities of cultured cells were measured using an in vitro transwell assay with modified Boyden chambers containing polycarbonate filters (Millipore, MA), according to the manufacturer’s instructions. Cells transfected with miR-146a/miR-Ctrl, or antagomiR-146a/nonrelated control molecules (NC) were plated 24 h after transfection in serum-free medium and allowed to invade towards a 10% FBS medium for 24 h, or 48 h. Cells that remained on top of the filter were scrubbed off, and those that invaded the underside of the filter were fixed and stained with crystal violet.
Generation of SMMC-7721-miR-146a stable cell lines
Pre-miR-146a were amplified by PCR using cDNA from SMMC-7721 cells and cloned into pcDNA3.1 vector. The pcDNA-miR-146a and the empty vector alone were transfected into SMMC-7721 cells using lipofectamine 2000 (Invitrogen). At 48 h post-transfection the cells were culture in complete medium with 400 μg/ml G418 for 4 weeks.
An experimental metastasis model in athymic nude mice was developed using the HCC cell line SMMC-7721, which has relatively strong
in vivo invasive and metastatic properties, as previously described [
23]. Briefly,
nude mice were anesthetized with pentobarbital and a transverse incision was made in the left flank through the skin and peritoneum. The spleen was carefully exposed and 2 × 10
6 viable SMMC-7721 cells transfected with pcDNA3.1 or pcDNA-pre-miR-146a were injected under the spleen capsule via a 27-gauge needle. Six weeks after the injection, the mice were sacrificed under anesthesia and tumor metastasis was examined under a stereo microscope.
Luciferase reporter assay
The 3′-UTRs of Ehmt2, mpzl1, rhoA, kif2c, and HAb18G were amplified by PCR and cloned downstream of the luciferase gene in the pGL3 reporter vector (Promega). Cells (3 × 104) were seeded in triplicate in 24-well plates and allowed to settle for 24 h. Then, approximately 100 ng of pGL3-HAb18-3′-UTR (wt or mut) and 1 ng of pRL-TK Renilla plasmid (Promega) were transfected into cells using Lipofectamine 2000 (Invitrogen) according to the manufacturer’s recommendations. Luciferase and Renilla signals were measured 48 h after transfection using the Dual Luciferase Reporter Assay Kit (Promega) according to the manufacturer’s protocol. Three independent experiments were performed, and the data are presented as the mean ± SD.
Western blotting
Western blotting analysis was performed according to the standard protocol described previously [
22].
The samples were subjected to SDS-PAGE and transferred onto a polyvinylidene fluoride membrane. The primary antibodies used in this study were as follows: anti-HAb18G (1:4,000 [
24], prepared by our lab), rabbit-anti-β-catenin (1:500, Santa Cruz), rabbit-anti-APC (1:500, Boster), rabbit-anti-VEGF (1:400, Boster), rabbit-anti-NF-κB p65 (1:400, Boster), and an anti-α-tubulin antibody as a loading control. The secondary antibodies used were either goat anti-mouse or goat anti-rabbit IgG (PIERCE), depending on the primary antibody used.
Statistical analysis
Statistical significance was evaluated using the Student’s t-test for paired comparisons. All values are expressed as the mean ± SD. P values <0.05 (using a 2-tailed paired t-test) were considered to indicate significantly significant differences between 2 groups of data. Non-metastasis time data were represented using Kaplan Meier curves and differences were compared by means of the pairwise log-rank test.
Discussion
Human miR-146a is embedded on chromosome 5q34, which is a region that is often deleted in HCC [
13], and has been reported to be aberrantly expressed in several cancers. Other reports have shown that miR-146a functions as a tumor suppressive miRNA in several kinds of tumors, including pancreatic cancer, and breast cancers cancer [
16,
30].
Our observation suggests that miR-146a was significantly downregulated in HCC cell lines as compared to normal liver tissue, which was consistent with our finding that miR-146a was expressed at lower levels in HCC tissues compared to adjacent non-cancerous tissues (Figure
1A). It may play a role in HCC tumoregenisis as a potential tumor suppressor. MiR-146a is also downregulated in metastatic HCC in comparison with non-metastatic HCC from patients (Figure
1F). These results suggest that miR-146a expression is often inhibited at earlier stages of cancer progression and that its expression level contributes to metastatic dissemination. Our observation also suggests that miR-146a is partially regulated by methylation. MiR-146a expression was restored in HCC cell lines upon administration of DNA methylation inhibitor 5-aza-2′-deoxycytidine (Figure
2A). Furthermore, we analyzed the methylation level of CpG sites in miR-146a promoter by bisulfite sequencing, our results suggest that the hypermethylation of the miR-146a promoter may be associated with down-expression of miR-146a in HCC tissues. Unfortunately, we could not reveal the apparent relationship involved in the methylation level and metastatic in HCC tissues in this study. However, it was notable that miR-146a promoter methylation level at one of seven the CpG site may related to metastatic (Figure
2D). Further studies using larger sample size are needed to reveal the relationship between miR-146a methylation and miR-146a expression, and HCC metastasis. MiR-146a was shown to directly inhibit expression of UHRF1 [
31], an epigenetic regulator and coordinate tumor suppressor gene silencing via DNA methylation [
32] in several cancers. Our results conjecture that repression of miR-146a may promote self-methylation by upregulation of UHRF1 in HCC. To identify this hypothesis, we compared the expression of UHRF1 mRNA by qRT-PCR and the methylation level by bisulfite sequencing in SMMC-7721 cells transfected with miR-Ctrl or miR-146a. The miR-146a treatment cells exhibited low expression of UHRF1 and lower methylation levels at seven CpG sites within the miR-146a promoter in comparison to the controls (Additional file
8: Figure S6C and D). These results hinted that there is a negative feedback loop in miR-146a expression. The further mechanism and function of this feedback regulation in HCC need to be further studied.
We testified the putative tumor suppressor function of miR-146a in human HCC by in vitro and in vivo assays. Ectopic expression of miR-146a could inhibit the invasion and metastasis of HCC cells (Figure
3B and C), and this effect was not due to suppressed cell cycle or increased apoptosis (Additional file
3: Figure S2). However, downregulated miR-146a was able to promote cell migration and invasion ability in 2 HCC cell lines (Huh-7 and HepG2) (Figure
3E and F). In nude mice, HCC cancer cells overexpressing miR-146a displayed significantly fewer metastatic loci than control cancer cells (Figure
3G,H and Additional file
2: Figure S1G). Collectively, our data indicate that miR-146a functions as a tumor suppressor in HCC. Our results imply that miR-146a may function as a negative regulator or tumor suppressor for cell invasion and metastasis in HCC.
The molecular mechanisms involved in miR-146a-mediated repression of metastasis are not well understood. In our study, we found that miR-146a inhibits HCC invasion and metastasis partly through the upregulation of APC and the downregulation of VEGF (Additional file
4: Figure S3 and Figure
4A). Adenomatous polyposis coli (APC), a tumor suppressor, interacts with β-catenin to inhibit Wnt signaling, is involved in multiple cellular functions and processes, including cell migration [
33]. Intact APC exerts its anti-tumor effect through the accelerated degradation of β-catenin or by regulating β-catenin nuclear export and the inhibition of signal transduction by the lymphoid enhancer factor–T cell factor (LEF-TCF) family of transcription factors [
34]. Here, we found that ectopic miR-146a expression increased APC protein levels and caused β-catenin to relocalize from the nucleus to the cytoplasm (Figure
4C,D, Additional file
5: Figure S4A and B). These findings indicate that miR-146a may inhibit the Wnt/β-catenin pathway by repressing nuclear β-catenin accumulation, which consequently leads to the partial downregulation of VEGF. However, the molecular mechanism involved in the upregulation of APC expression by miR-146a requires further investigation.
To date, several targets of miR-146a have been identified, such as IRAK-1 [
17,
18]. EGFR [
16], UHRF1 [
31]. On the basis of our bioinformatics analysis and sequential experiments, we demonstrated that HAb18G is an additional miR-146a target in HCC cells. There are several lines of evidence to support this conclusion. First, based on our analysis using publicly available algorithms (TargetScan and miRanda), we found that HAb18G mRNA is a theoretical target gene of miR-146a, with potential binding sites conserved in multiple species (Additional file
8: Figure S6B). Importantly, our luciferase reporter assay results confirmed that miR-146a overexpression significantly downregulated luciferase activity by directly targeting the 3′-UTR of HAb18G mRNA. However, this effect was eliminated when the nucleic acids in the HAb18G 3′-UTR targeted by miR-146a were mutated. Second, HAb18G protein expression was significantly decreased in miR-146a-overexpressing SMMC-7721 cells as compared to the NC. Moreover, impaired miR-146a expression resulted in increased HAb18G protein levels in HepG2 and Huh-7 cells. Third, to further elucidate mechanisms underlying the tumor suppressive effect of miR-146a, we overexpressed HAb18G, but without its endogenous 3′-UTR in SMMC-7721-miR-146a cells and found that the invasion was increased, which suggest that the regulation of miR-146a on migration and invasion in HCC cells is related to HAb18G inhibition. We established the miR-146a-
HAb18G axis by rescue experiment and the reverse correlation between miR-146a and HAb18G expression in HCC samples.
We observed decreased VEGF levels after HAb18G siRNA transfection and also found that elevated levels of HAb18G mRNA lacking the 3′-UTR were able to upregulate VEGF expression. Therefore, miR-146a could downregulate VEGF levels through the inhibition of HAb18G, leading to the suppression of cancer cell invasion and metastasis. However, the molecular mechanism involved in the upregulation of APC expression by miR-146a requires further investigation.
It has been reported that miR-146a can regulate NF-κB p65 signaling by targeting IRAK-1 [
17,
18] or EGFR [
16]. Importantly, we also observed decreased NF-κB p65 levels after reexpressing miR-146a or after HAb18G siRNA transfection (Figures
4A and
5F). Altogether, our data indicate that miR-146a can indirectly downregulate NF-κB p65 signaling by suppressing multiple molecules, resulting in the inhibition of cancer cell invasion and metastasis. On the other hand, Ectopic expression of NF–κB p65 or β-catenin in SMMC-7721-miR-146a cells reversed the effect of miR-146a to inhibit VEGFA expression (Figure
4E) and promote cell invasion (Additional file
6: Figure S7B). We also noted that decreased NF-κB p65 expression by siRNA could attenuate the HAb18G-induced SMMC-7721-miR-146a cells and repressed cells invasion (Additional file
6: figure S7C, D). These results indicate that miR-146a repress VEGFA though β-catenin and NF–κB signal pathway.
Moreover, we demonstrated that miR-146a downregulation in HCC cells led to the upregulation of VEGF via 2 signaling pathways: 1) the repression of APC expression, leading to β-catenin accumulation in nucleus, and 2) a direct reduction in HAb18G expression, which consequently promoted the expression of NF-κB p65. These findings will contribute to our understanding of the molecular mechanism by which miR-146a influences tumorigenesis and may aid in the development of novel cancer therapy strategies.
Acknowledgements
We thank Dr. Chenggong Liao (Cell Engineering Research Center, Fourth Military Medical University) and Dr. Li Xu (College of Life Sciences and Bioengineering, Beijing Jiaotong University) for their technical assistance and critical review of the manuscript.
Financial support
This work was supported by grant from the National Natural Science Foundation of China (81101569), the National Science and Technology Major Project (2013ZX09301301), and the National Basic Research Program (973 Program, 2009CB521705).
Writing assistance
The manuscript was edited by American Journal Experts, USA.
Competing interests
The authors declare they have no competing interests.
Authors’ contributions
CZN and ZZ contributed to concept design, discussed results and wrote the manuscript. ZZ also performed cell culture, methylation assay, cell migration and invasion assay, ZY and SXX performed animal experiment, protein isolation and immunoblot analysis, IHC and FCM; MX performed Bioinformatics analysis. All authors read and approved the final manuscript.