MicroRNA-142-3p and microRNA-142-5p are downregulated in hepatocellular carcinoma and exhibit synergistic effects on cell motility

Abstract

MicroRNAs (miRNAs), an important class of small non-coding RNAs, regulate gene expression at the post-transcriptional level. miRNAs are involved in a wide range of biological processes and implicated in different diseases, including cancers. In this study, miRNA profiling and qRT-PCR validation revealed that miR-142-3p and miR-142-5p were significantly downregulated in hepatocellular carcinoma (HCC) and their expression levels decreased as the disease progressed. The ectopic expression of miR-142 significantly reduced HCC cell migration and invasion. Overexpression of either miR-142-3p or miR-142-5p suppressed HCC cell migration, and overexpression of both synergistically inhibited cell migration, which indicated that miR-142-3p and miR-142-5p may cooperatively regulate cell movement. miR-142-3p and miR-142-5p, which are mature miRNAs derived from the 3′- and 5′-strands of the precursor miR-142, target distinct pools of genes because of their different seed sequences. Pathway enrichment analysis showed a strong association of the putative gene targets of miR-142-3p and miR-142-5p with several cell motility-associated pathways, including those regulating actin cytoskeleton, adherens junctions, and focal adhesion. Importantly, a number of the putative gene targets were also significantly upregulated in human HCC cells. Moreover, overexpression of miR-142 significantly abrogated stress fiber formation in HCC cells and led to cell shrinkage. This study shows that mature miR-142 pairs collaboratively regulate different components of distinct signaling cascades and therefore affects the motility of HCC cells.

Keywords

hepatocellular carcinoma / microRNA / metastasis / cytoskeletal reorganization

Introduction

Altered microRNA (miRNA) expression has been frequently implicated in human cancers. Over the past years, a substantial number of oncogenic and tumor suppressive miRNAs has been identified and evaluated to gain important insights into the underlying mechanisms of carcinogenesis [ 1]. The gene encoding miR-142 is located on the human chromosome 17. RNA deep-sequencing analysis revealed that the miR-142-3p and miR-142-5p strands, which originated from the precursor miR-142, is associated with the RNA-induced silencing complex (RISC); these strands are equally efficient in guiding the RNA interference (RNAi) machinery to the respective target mRNAs [ 2]. The aberrant expression of miR-142-3p and miR-142-5p has been found in multiple solid cancers. On the one hand, miR-142-5p was significantly downregulated in gastric cancer and associated with tumor recurrence [ 3]. miR-142-5p has also been shown to regulate several oncogenic signaling pathways by targeting MAPK, Wnt, VEGF, and Notch [ 3]. On the other hand, miR-142-3p was significantly downregulated in colon cancer and underexpression of which was associated with poor differentiation and tumor size. miR-142-3p may function as a tumor suppressor by targeting stem cell or cancer stem cell markers, such as CD133, ABCG2, and LGR5, in colon cancer cells [ 4]. Overexpressing miR-142-3p in colon cancer cells arrested cell cycle at G₁ phase and enhanced chemosensitivity to 5-fluorouracil [ 5]. miR-142-3p also suppressed cell proliferation in pancreatic ductal adenocarcinoma by targeting HSP70 [ 6].

Hepatocellular carcinoma (HCC) is a prevalent cancer with a poor prognosis and a high mortality rate. Hepatocarcinogenesis involves accumulation of genetic and epigenetic alterations, including miRNA deregulation [ 7]. In the present study, we investigated the expression levels of miR-142-3p and miR-142-5p in human HCC cells. Both miR-142-3p and miR-142-5p were significantly downregulated in human HCCs and cooperatively modulated the Rho/ROCK signaling pathway, thereby suppressing HCC cell migration. This study suggests that miR-142 is a metastatic-suppressive miRNA in HCC.

Materials and methods

Patient samples and HCC cell lines

Seventy pairs of human HCC samples and their corresponding non-tumorous (NT) liver samples were obtained from surgical resection of patients with HCC at Queen Mary Hospital, Hong Kong. Eleven normal liver (NL) samples were also acquired from resection of liver metastasis in patients with colonic cancer. The clinicopathological features of patients with HCC are listed in Supplementary Table 1. The criteria for resectability included absence of distant metastasis or main portal vein thrombosis, anatomically resectable disease, and adequate liver function reserve. None of the patients had received other therapies, such as chemoembolization or chemotherapy, before resection. Tumors were completely excised with clear microscopic margin and no residual tumors were revealed by CT scan or angiography 1 month after surgery. NT liver samples were collected at least 1 cm away from the corresponding tumors. The samples were snap-frozen in liquid nitrogen and stored at -80 °C. H and E staining was performed to confirm tissue homogeneity in HCC and NT liver sample blocks. The use of human tissue in this study was approved by the Institutional Review Board of the University of Hong Kong and the Hospital Authority. The HCC cell lines SMMC-7721 and BEL-7402 used in this study were obtained from the Shanghai Institute of Cell Biology, Chinese Academy of Sciences. The cells were cultured in Dulbeccco’s modified Eagle’s medium-high glucose supplemented with 10% fetal bovine serum (FBS) and antibiotics.

RNA extraction and quantitative real-time RT-PCR (qRT-PCR)

Total RNA was extracted with TRIZOL reagent, and cDNA was synthesized using 10 ng of total RNA through TaqMan® MicroRNA assays (Applied Biosystems, Foster city, CA) specific for miR-142-3p, miR-142-5p, and RNU44 (endogenous control). All qRT-PCR reactions were performed using TaqMan Universal PCR Master Mix (2×) in the 7900HT Fast real-time PCR systems (Applied Biosystems). miRNA expression of each clinical sample was measured in triplicate. The relative expression levels of miR-142-3p and miR-142-5p were determined with the following formula: ∆Ct= Ct (miR-142) - Ct (RNU44) and ∆∆Ct= ∆Ct (T) - ∆Ct (NT), respectively.

Transient transfection of miR-142-3p and miR-142-5p

miR-142-3p, miR-142-5p, and scramble control were obtained from Ambion (Foster City, CA). The day before transfection, 2×10⁵ HCC cells were seeded in each well of a 6-well plate. miRNA (10 µM) was transfected into HCC cells by using X-tremeGene siRNA transfection reagent (Roche, Indianapolis, IN) according to the manufacturer’s manual. The cells were harvested 48 h after transfection for subsequent functional studies and assays.

Establishment of miR-142 overexpressing stable clones

miR-142-specific forward primer (5′GGCGCTAGC-AAGAGGAAGTGGGGG3′) and reverse primer (5′GGCGGATCC-TCGTCCTTGACGCTG 3′) flanking the 3′ and 5′ ends of the miR-142 precursor were used for PCR amplification. PCR product was purified and subsequently cloned into the lentivirus expression vector pCDH-CMV-MCS-EF1-GFP-Puro (System Biosciences, Mountain View, CA). For establishing miR-142 overexpressing stable cells, lentivirus was packaged in 293FT cells by co-transfection of lentiviral expression plasmids and pPACKH1 packaging plasmid mix (System Biosciences) with Lipofectamine Reagent Plus (Invitrogen, Foster City, CA). Recombinant lentivirus particles were harvested 48 h post transfection and used to infect the parental BEL-7402 and SMMC-7721. Stably transduced HCC cells were selected using 2 μg/ml of puromycin.

Cell proliferation assay

miR-142 overexpressing HCC cells were seeded in a 24-well plate with 5×10⁴ HCC cells per well in triplicate. The number of cells per well was quantified using Z1 COULTER COUNTER® Cell and Particle Counter (Beckman coulter, Brea, CA). Cells were counted daily for 8 days.

Colony formation assay

HCC cells overexpressing miR-142 were seeded in a 6-well plate with 1×10⁴ cells per well and maintained in a culture medium supplemented with 10% FBS at 37 °C humidified CO₂ chamber for 14 days. Surviving colonies were fixed with absolute methanol and stained with 0.1% crystal violet. AlphaEaseFC^TM software (Alpha Innotech, San Leandro, CA) was used to quantify the number of colonies.

Cell migration assay

Briefly, 5×10⁴ HCC cells were resuspended in 200 μl of serum-free culture medium and gently loaded into the upper chamber of the transwell. 500 μl culture medium supplemented with 10% FBS was filled into the lower chamber. The transwell system was incubated in a humidified CO₂ chamber at 37 °C for 16 h. Non-migrated cells in the inner side of the upper chamber were removed by cotton swabs. Migrated cells attached on the outer side of the transwell membrane were first fixed with 100% methanol for 10 min and then stained with 0.1% crystal violet for 30 min. Migrated cells were counted in three random field images of the transwell membrane by using the AlphaEaseFC^TM software (Alpha Innotech).

Cell invasion assay

A thin layer of Matrigel ^TMMatrix (BD Bioscience) was used to coat the inner side of the transwell to mimic cell migration and breakdown of the basement membrane matrix during metastatic events. Experimental procedures for cell invasion assay are similar to that for cell migration assays.

Immunofluorescence staining

Briefly, 1×10⁵ cells were seeded onto coverslips and allowed to grow for 24 h. The cells were serum starved with FBS-depleted culture medium for another 24 h and then replenished with 10% FBS culture medium for 30 min to induce stress fiber formation. The cells were then fixed with 4% paraformaldehyde in 1×PBS and permeabilized with 0.2% Triton X-100 in PBS. The fixed cells were incubated with Texas-Red conjugated phalloidin (1:2000; Sigma-Aldrich, St. Louis, MO) at room temperature for 1 h. The cells were then counterstained with DAPI (1:100; Calbiochem, San Diego, CA). The coverslips were mounted with Vectashield antifade mountant (Vector Laboratories, Burlingame, CA). Images of the cells were captured under 100×magnification by using a fluorescence microscope connected to a charge-coupled device camera (Leica, Wetzlar, Germany). Cells strongly stained with phalloidin were quantified by randomly counting 100 cells.

Scanning electron microscopy (SEM)

SEM was performed to reveal the morphology of miR-142 overexpressing and control cells. Briefly, 1×10⁵ cells were seeded onto coverslips and allowed to grow for 24 h. The cells were fixed with 2.5% glutaraldehyde in distilled water for 1 h and dehydrated stepwise by using an ascending series of ethanol. Dehydrated cells were subjected to critical point drying, mounted on an aluminum stub by using a silver paste, and then coated with a very thin film of palladium prior to SEM examination. Images were captured under 12 000×magnification by using Hitachi S-4800 FEG SEM (Tokyo, Japan).

In silico miRNA target prediction and pathway enrichment analysis

Putative gene targets of miR-142-3p and miR-142-5p were predicted using three bioinformatics algorithms, namely, TargetScan 5.0 (http://www.targetscan.org) [ 8], Pictar (http://pictar.mdc-berlin.de) [ 9], and MicroT4.0 (http://diana.cslab.ece.ntua.gr/microT) [ 10]. Pathway enrichment analysis was performed using the DIANA-miRPath 2.0 software developed by the DNA Intelligent Analysis (DIANA) laboratory (http://www.microrna.gr/miRPathv2) [ 11]. An enrichment statistical score (negative natural logarithm of the P value) was generated for each specific Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway [ 12].

Results

Frequent downregulation of miR-142-3p and miR-142-5p in primary HCCs

We previously compared the miRNA expression profiles of 20 pairs of primary HCC samples and their corresponding NT liver samples [ 13] and identified a subset of frequently deregulated miRNAs in human HCCs. In this study, we further analyzed these deregulated miRNAs and found that a pair of concurrently expressed sister miRNAs, namely, miR-142-3p and miR-142-5p, was significantly downregulated in primary HCCs (miR-142-3p, P = 0.0009; miR142-5p, P = 0.0013, paired t-test) (Supplementary Fig. 1).

We validated the expression levels of miR-142-3p and miR-142-5p through qRT-PCR in an expanded cohort of 70 pairs of primary HCC and their corresponding NT liver samples. Both miR-142-3p and miR-142-5p were consistently and significantly downregulated in primary HCCs (P<0.0001, paired t-test) (Fig. 1A). Both miR-142-3p and miR-142-5p were also underexpressed in 71.4% (50/70) of the HCC cases (Fig. 1B). We also observed a significant and positive correlation between the expression levels of miR-142-3p and miR-142-5p in both the NT liver samples (P<0.0001, R² = 0.7971, linear regression) and primary HCC tissue (P<0.0001, R² = 0.7385, linear regression) (Fig. 1C). This finding indicated that both the 3′ and 5′ miRNA strands, which originated from the miR-142 precursor, were concurrently expressed.

Expression changes of miR-142-3p and miR-142-5p in multistep hepatocarcinogenesis

HCC development involves multiple steps evolving from chronic hepatitis and cirrhosis to early and advanced HCC. To evaluate the expression changes of miR-142-3p and miR-142-5p during HCC progression, we segregated 70 NT liver samples into chronic hepatitis (CH) and liver cirrhosis (CL), whereas the 70 primary HCC samples were further segregated into early HCC (pTNM stages I and II) and advanced HCC (pTNM stages III and IV). Eleven NL samples were also included for comparison. The expression levels of both miR-142-3p and miR-142-5p gradually decreased from NL to CH and/or CL to early and advanced HCC. These finding suggest that miR-142-3p and miR-142-5 play critical roles in multistep hepatocarcinogenesis (Fig. 1D).

Establishment of miR-142 overexpression model

To characterize the molecular functions of miR-142 in HCC, we used a lentiviral transduction system to stably overexpress miR-142 in BEL-7402 and SMMC-7721 cells. As the entire miR-142 precursor was cloned into the lentiviral expression vector, stable cells established using this approach simultaneously expressed miR-142-3p and miR-142-5p. GFP expression was observed as a reporter in stable cells, which indicated the successful transduction of the lentiviral expression vectors (Supplementary Fig. 2A). qRT-PCR was performed using individual TaqMan probes to validate overexpression of miR-142-3p and miR-142-5p in stable HCC cells (Supplementary Fig. 2B). The results confirm the successful establishment of miR-142 overexpressing HCC cell lines for subsequent functional characterization.

Effect of miR-142 on cell proliferation

Cell proliferation and colony formation assays were performed to assess the effects of overexpressing miR-142 on HCC cell proliferation and clonogenicity, respectively. Overexpression of miR-142 in HCC cells did not significantly affect cell proliferation (Supplementary Fig. 3A) and the number of colonies formed by HCC cells. However, colonies formed by miR-142 overexpressing BEL-7402 cells were smaller than those of the empty vector (EV) cells (Supplementary Fig. 3B). These results suggest that miR-142 does not regulate cell proliferation but minimal affects clonogenicity in HCC.

Overexpression of miR-142 inhibited cell migration in HCC

Transwell cell migration assay was performed to assess and quantify the effects of miR-142 on HCC cell migration. HCC cells were allowed to migrate through the transwell membrane by using 10% FBS as chemoattractant. Overexpression of miR-142 in BEL-7402 and SMMC-7721 cells remarkably attenuated cell migration compared with the EV control (P<0.0001, t-test) (Fig. 2). The result suggests that miR-142 inhibits cell motility in HCC.

Overexpression of miR-142 inhibited HCC cell invasion

Cell invasion assay was performed to assess and quantify the effects of overexpressing miR-142 on cell invasiveness. As a thin layer of extracellular matrix was coated in the inner side of the transwell membrane, HCC cells have to break the matrix in order to migrate through the membrane toward the chemoattractant. The invasiveness of miR-142 overexpressing BEL-7402 cells was significantly inhibited compared with that of the EV control (P<0.0001, by t-test) (Supplementary Fig. 4). These results suggest that miR-142 regulates metastatic events, such as HCC cell migration and breakdown of the basement membrane matrix, to invade distant sites.

Additive inhibitory effects of miR-142-3p and miR-142-5p on HCC cell migration

Given that miR-142-3p and miR-142-5p strands were simultaneously expressed in the lentiviral expression system, we could not distinguish the individual effects of miR-142-3p and miR-142-5p on HCC cells. As such, synthetic miRNA molecules were employed to ectopically express either miR-142-3p or miR-142-5p in HCC cells. Overexpression of miR-142-3p or miR-142-5p alone in BEL-7402 cells significantly attenuated cell migration across the transwell membrane (P<0.0001, t-test) (Fig. 3), which indicated that either miR-142-3p or miR-142-5p reduces HCC cell migration. Overexpression of both miR-142-3p and miR-142-5p in BEL-7402 cells exhibited additive inhibitory effect on cell migration (compared with miR-142-3p, P = 0.0007; compared with miR-142-5p, P = 0.0002; t-test) (Fig. 3). The drastic inhibitory effect induced by miR-142-3p and miR-142-5p suggests their possible coordination in regulating cell motility.

miR-142-3p and miR-142-5p regulated distinct pools of gene targets

The combinatorial inhibitory effect of miR-142-3p and miR-142-5p on cell migration indicates a possible coordination between miR-142-3p and miR-142-5p. As the combinatorial effect may be mediated by the respective downstream genes targeted by miR-142-3p and miR-142-5p, we predicted the putative targets of miR-142-3p and miR-142-5p by using the TargetScan 5.0 algorithm. The predictions were based on three basic principles, namely, sequence complementarity, RNA hybrid thermostability, and conservation degree of the miRNA binding sites. A total of 330 and 764 putative targets of miR-142-3p and miR-142-5p, respectively, were predicted. Of these targets, only 46 genes were predicted to be commonly regulated by both miR-142-3p and miR-142-5p. This finding supports that despite being generated from the same precursor, miR-142-3p and miR-142-5p harbor different seed sequences and thus target distinct pools of genes (Supplementary Fig. 5).

Putative targets of miR-142-3p and miR-142-5p were enriched in pathways associated with cell motility

DIANA-miRPath 2.0 software was used to investigate the underlying molecular mechanisms of the combinatorial effect of miR-142-3p and miR-142-5p. Pathway enrichment analysis was performed using the miR-142-3p and miR-142-5p gene targets predicted by the TargetScan 5.0, Pictar, and MicroT4.0 algorithms. A total of 330, 166, and 328 putative miR-142-3p targets were predicted by TargetScan 5.0, Pictar, and MicroT4.0, respectively; of which, 101 targets were commonly predicted by the three algorithms. For miR-142-5p, 764, 301, and 719 putative targets were predicted by TargetScan 5.0, Pictar, and MicroT4.0, respectively; of which, 133 putative targets were commonly predicted by the three algorithms (Supplementary Fig. 6). An enrichment statistical score for each KEGG pathway was then generated with the DIANA-miRPath algorithm. Analysis revealed that 234 putative targets of miR-142-3p and miR-142-5p were enriched in 65 KEGG pathways. The top 30 pathways associated with miR-142-3p and miR-142-5p are shown in Table 1. Notably, several cancer pathways, such as colorectal cancer (Pathway ID hsa05210), pancreatic cancer (Pathway ID hsa05212), chronic myeloid leukemia (Pathway ID hsa05220), renal cell carcinoma (Pathway ID hsa05211), glioma (Pathway ID hsa05214), and acute myeloid leukemia (Pathway ID hsa05221), were enriched as shown in the analysis. Some well-established cancer-related signaling pathways, including mTOR (Pathway ID hsa04150), TGF-β (Pathway ID hsa04350), MAPK (Pathway ID hsa04010), and non-canonical Wnt signaling (Pathway ID hsa04310), were also significantly enriched. Specific cell motility-associated pathways, including those regulating actin cytoskeleton (Pathway ID hsa04810), adherens junction (Pathway ID hsa04520), and focal adhesion (Pathway ID hsa04510), were significantly associated with the putative gene targets of miR-142-3p and miR-142-5p. This prediction was consistent with our observation on overexpressing miR-142-3p and miR-142-5p in HCC cell lines. Thus, the pathway enrichment analysis on the putative targets of miR-142-3p and miR-142-5p confirms that mature miR-142 pairs collaboratively influence multiple important biological processes by regulating different components of distinct signaling cascades and thus promote cancer development and progression.

Putative targets of miR-142-3p and miR-142-5p were upregulated in HCC

The pathway enrichment analysis results suggested that miR-142-3p and miR-142-5p modulate different pathways through their downstream targets and thus contribute to cancer development and metastasis. miR-142 was significantly associated with cell motility-associated pathways, which further strengthened our experimental findings. In particular, a large number of genes involved in actin cytoskeleton signaling cascades were significantly enriched (Fig. 4A). For instance, the Rho family of small GTPases, including Rho, Rac, and Cdc42, has been reported to regulate the organization of actin cytoskeleton through their downstream protein kinase effectors, such as ROCK and PAK, together with other actin regulatory proteins, such as mDia and WASP. The RhoGTPase signaling pathway has been frequently found to be upregulated in human cancers. Several studies provided strong evidence of the roles of this RhoGTPase signaling pathway in regulating cytoskeleton and cancer metastasis [ 14, 15]. To verify the prediction results of the in silico analysis, we evaluated the expression levels of the individual components of the actin cytoskeleton signaling cascade in human HCCs. By using the RNA sequencing data obtained from the Cancer Genome Atlas (TCGA), we examined the expression levels of the actin cytoskeleton signaling components. The expression levels of several signaling molecules, such as ACTN4, DIAPH3, FGF13, ITGAV, MYH9, NCKAP1, PIP4K2C, RAC1, RHOA, and SSH2, were consistently and significantly elevated in HCC tissue compared with those in the NT liver samples (ACTN4, P<0.0001; DIAPH3, P<0.0001; FGF13, P<0.0001; ITGAV, P<0.0001; MYH9, P = 0.002; NCKAP1, P = 0.0216; PIP4K2C, P<0.0001; RAC1, P = 0.0002; RHOA, P<0.0001; and SSH2, P = 0.0059; t-test) (Fig. 4B). The inverse expression trends of these molecules with miR-142 suggested the possible regulatory relationship between them. Therefore, these findings suggest a combinatorial effect of miR-142-3p and miR-142-5p in suppressing HCC cell motility by regulating various signaling cascades. Downregulation of miR-142-3p and miR-142-5p in HCC may also mediate the activation of cell motility-associated pathways to facilitate HCC metastasis.

Overexpression of miR-142 disrupted actin cytoskeleton in HCC cells

Predictions from the pathway enrichment analysis prompted us to investigate the effect of miR-142 on actin cytoskeleton organization. Cytoskeletal reorganization, as exemplified by stress fiber formation and polymerized actin, are important for the contractile motion of cancer cells. To determine whether miR-142 regulates actin cytoskeleton and influences cell motility in HCC, we performed phalloidin staining to visualize stress fiber formation in miR-142 overexpressing HCC cells. GFP was visualized as a reporter gene to confirm the stable expression of miR-142 or to confirm the presence of vector control in the HCC cells. A marked reduction in stress fiber formation was observed in miR-142 overexpressing SMMC-7721 cells compared with the control (Fig. 5A). BEL-7402 cells overexpressing miR-142 exhibited different cell morphology and markedly reduced cell cytoplasm compared with the control cells (Fig. 5B). SEM was conducted to investigate the effect of overexpressing miR-142 on HCC cell morphology. In agreement with the immunofluorescence staining, SEM results revealed that overexpression of miR-142 in BEL-7402 and SMMC-7721 cells resulted in significant cell shrinkage (Fig. 5C) possibly because of miR-142-3p and miR-142-5p disrupted the stress fiber network.

Discussion

miR-142-3p and miR-142-5p have been reported to be commonly downregulated in several types of solid cancers, including gastric, colon, and pancreatic [ 3, 5, 6]. Downregulation of miR-142-3p and miR-142-5p in these cancers is frequently associated with a poor prognosis and short survival of cancer patients. In this study, we revealed that miR-142-3p and miR-142-5p were frequently (>70%) downregulated in HCC samples and their expression levels showed a stepwise decrease from NL to CH and CL to early HCC and eventually advanced HCC. Hence, miR-142-3p and miR-142-5p are involved in multistep HCC carcinogenesis.

miR-142 has been shown to impair cell proliferation and induce cell cycle arrest in other types of cancer by targeting various oncogenes [ 5, 6]. However, in the present study, overexpression of miR-142 did not significantly affect the proliferation and viability of HCC cells but significantly attenuated the migration and invasion of HCC cells. Our findings are consistent with that of Wu et al., who demonstrated the inhibitory effect of overexpressing miR-142-3p on the migration and invasion of QGY-7703 and SMMC-7721 cells [ 4]. In our previous miRNA profiling study, the expression of both miR-142-3p and miR-142-5p significantly reduced in primary HCCs to venous metastases [ 13]. Taken together, our current findings suggest that miR-142 regulates cell motility-associated events.

miR-142-3p and miR-142-5p are a sister miRNA pair that originates from the same miRNA precursor. According to the canonical biogenesis pathway of miRNA, only one strand in the stem-loop structure of a precursor miRNA is preferentially selected, based on thermodynamic stability, to be incorporated in the RISC, which will guide the RNAi machinery to target RNA; the other passenger strand is destroyed [ 16, 17]. However, recent research showed that both the 5′- and 3′-strands, which originated from opposite arms of the same precursor miRNA, could be concurrently expressed. Therefore, the preferential strand selection during miRNA biogenesis is not universally followed [ 18]. Based on the human miRNAs listed in the miRBase, 152 of the miRNAs are paired and contain both the 5′- and 3′-strands [ 19].

Despite that sister miRNA pairs share the same precursor, they do not share similar seed sequences and thus, theoretically, can regulate different pools of target genes. As suggested by the “Target-Two-Sets-of-Genes-With-One-Pre-miRNA” model proposed by Ro et al., the number of gene targets regulated by each pre-miRNA may be higher and thus allow miRNA to augment their regulatory capacity [ 18]. Moreover, distinct gene targets regulated by sister miRNA pairs may be involved in intricate regulatory networks or interact with each other to coordinately mediate physiological changes. On the one hand, miR-17-3p and miR-17-5p were demonstrated to synergistically promote the proliferation, migration, survival, morphogenesis, and colony formation of HCC cells by independently targeting PTEN, vimentin, or GalNT7 [ 20]. On the other hand, the 3′ UTR of each gene may harbor multiple miRNA binding sites; thus, a single target gene can be simultaneously regulated by both sister miRNAs via independent interactions in the binding sites in their 3′ UTRs. For example, tumor-suppressive miR-582-5p and miR-582-3p were shown to regulate similar set of genes, including PGGT1B, LRRK2, and DIXDC1, thereby reducing the proliferation and invasion of human bladder cancer cells [ 21]. Therefore, the repressive ability of single-precursor miRNAs can be enhanced by generating two mature miRNAs to bind to the same target.

In the present study, we speculated that sister miRNA pairs generated from pre-miR-142 regulate different pools of target genes, depending on their respective seed sequences, and coordinately mediate certain physiological changes by regulating a network of genes in HCC. Our findings demonstrated that miR-142-3p and miR-142-5p collaboratively regulate cell motility in HCC. In silico analysis results revealed that miR-142-3p and miR-142-5p regulate distinct pools of gene targets; of these putative targets, RACI, HSP70, CD133, and LGR5 have already been experimentally validated by previous studies [ 4]. A small number of the targets, including RhoGEF, integrin αV, and integrin β8, were predicted to harbor binding sites for both miR-142-3p and miR-142-5p, suggesting a possible simultaneous regulation by miR-142-3p and miR-142-5p.

Pathway enrichment analysis was performed to investigate the possible coordination between miR-142-3p and miR-142-5p. The putative targets of miR-142-3p and miR-142-5p were significantly enriched in a number of cancer pathways, which is consistent with that reported in previous studies. For instance, miR-142 was aberrantly expressed in colorectal cancer, pancreatic cancer, and acute myeloid leukemia, whose pathways were also significantly enriched in our analysis [ 5, 6, 22]. More importantly, the putative targets of miR-142-3p and miR-142-5p were shown to be highly enriched in cell motility-associated pathways, including those regulating actin cytoskeleton, adherens junction, and focal adhesion, which is consistent with our experimental findings. The prediction of the enrichment analysis was further supported by the increased expression levels of the signaling components that regulate the actin cytoskeleton; these components included ACTN4, DIAPH3, FGF13, ITGAV, MYH9, NCKAP1, PIP4K2C, RAC1, RHOA, and SSH2 in HCC. Finally, prediction of the in silico analysis was supported experimentally, as evidenced by the disruption of the stress fiber network and collapse of the cytoplasm of HCC cells by the ectopic overexpression of miR-142. Taken together, our findings indicate that miR-142-3p and miR-142-5p may collaboratively regulate different targets involved in cell motility-associated pathways and eventually inhibit cell motility in HCC.

Conclusions

In this study, miR-142-3p and miR-142-5p were significantly and frequently (71.4%) downregulated in HCC compared with that in the corresponding NT liver samples. The expression levels of miR-142-3p and miR-142-5p progressively decreased from NL to CH and CL, and then to early HCC and eventually advanced HCC. This finding indicated the critical involvements of miR-142-3p and miR-142-5p in multistep HCC development and progression. Functionally, overexpression of miR-142-3p and miR-142-5p showed a combinatory effect on attenuating cell migration by disrupting cytoskeleton organization in HCC cells. Pathway enrichment analysis further suggested the involvements of miR-142-3p and miR-142-5p in modulating key pathways associated with cell motility, as well as their collaborative relationship in regulating the motility of HCC cells.

References

[1] Esquela-Kerscher A, Slack FJ. Oncomirs—microRNAs with a role in cancer. Nat Rev Cancer2006; 6(4): 259–269

[2] Landgraf P, Rusu M, Sheridan R, Sewer A, Iovino N, Aravin A, Pfeffer S, Rice A, Kamphorst AO, Landthaler M, Lin C, Socci ND, Hermida L, Fulci V, Chiaretti S, Foà R, Schliwka J, Fuchs U, Novosel A, Müller RU, Schermer B, Bissels U, Inman J, Phan Q, Chien M, Weir DB, Choksi R, De Vita G, Frezzetti D, Trompeter HI, Hornung V, Teng G, Hartmann G, Palkovits M, Di Lauro R, Wernet P, Macino G, Rogler CE, Nagle JW, Ju J, Papavasiliou FN, Benzing T, Lichter P, Tam W, Brownstein MJ, Bosio A, Borkhardt A, Russo JJ, Sander C, Zavolan M, Tuschl T. A mammalian microRNA expression atlas based on small RNA library sequencing. Cell2007; 129(7): 1401–1414

[3] Zhang X, Yan Z, Zhang J, Gong L, Li W, Cui J, Liu Y, Gao Z, Li J, Shen L, Lu Y. Combination of hsa-miR-375 and hsa-miR-142-5p as a predictor for recurrence risk in gastric cancer patients following surgical resection. Ann Oncol2011; 22(10): 2257–2266

[4] Wu L, Cai C, Wang X, Liu M, Li X, Tang H. MicroRNA-142-3p, a new regulator of RAC1, suppresses the migration and invasion of hepatocellular carcinoma cells. FEBS Lett2011; 585(9): 1322–1330

[5] Shen WW, Zeng Z, Zhu WX, Fu GH. miR-142-3p functions as a tumor suppressor by targeting CD133, ABCG2, and Lgr5 in colon cancer cells. J Mol Med (Berl)2013; 91(8): 989–1000

[6] MacKenzie TN, Mujumdar N, Banerjee S, Sangwan V, Sarver A, Vickers S, Subramanian S, Saluja AK. Triptolide induces the expression of miR-142-3p: a negative regulator of heat shock protein 70 and pancreatic cancer cell proliferation. Mol Cancer Ther2013; 12(7): 1266–1275

[7] Wong CM, Kai AK, Tsang FH, Ng IO. Regulation of hepatocarcinogenesis by microRNAs. Front Biosci (Elite Ed)2013; 5: 49–60

[8] Lewis BP, Burge CB, Bartel DP. Conserved seed pairing, often flanked by adenosines, indicates that thousands of human genes are microRNA targets. Cell2005; 120(1): 15–20

[9] Chen K, Rajewsky N. Natural selection on human microRNA binding sites inferred from SNP data. Nat Genet2006; 38(12): 1452–1456

[10] Maragkakis M, Alexiou P, Papadopoulos GL, Reczko M, Dalamagas T, Giannopoulos G, Goumas G, Koukis E, Kourtis K, Simossis VA, Sethupathy P, Vergoulis T, Koziris N, Sellis T, Tsanakas P, Hatzigeorgiou AG. Accurate microRNA target prediction correlates with protein repression levels. BMC Bioinformatics2009; 10(1): 295

[11] Papadopoulos GL, Alexiou P, Maragkakis M, Reczko M, Hatzigeorgiou AG. DIANA-mirPath: integrating human and mouse microRNAs in pathways. Bioinformatics2009; 25(15): 1991–1993

[12] Kanehisa M, Goto S. KEGG: Kyoto encyclopedia of genes and genomes. Nucleic Acids Res2000; 28(1): 27–30

[13] Wong CM, Wong CC, Lee JM, Fan DN, Au SL, Ng IO. Sequential alterations of microRNA expression in hepatocellular carcinoma development and venous metastasis. Hepatology2012; 55(5): 1453–1461

[14] Wong CC, Wong CM, Tung EK, Man K, Ng IO. Rho-kinase 2 is frequently overexpressed in hepatocellular carcinoma and involved in tumor invasion. Hepatology2009; 49(5): 1583–1594

[15] Fukui K, Tamura S, Wada A, Kamada Y, Sawai Y, Imanaka K, Kudara T, Shimomura I, Hayashi N. Expression and prognostic role of RhoA GTPases in hepatocellular carcinoma. J Cancer Res Clin Oncol2006; 132(10): 627–633

[16] Khvorova A, Reynolds A, Jayasena SD. Functional siRNAs and miRNAs exhibit strand bias. Cell2003; 115(2): 209–216

[17] Schwarz DS, Hutvágner G, Du T, Xu Z, Aronin N, Zamore PD. Asymmetry in the assembly of the RNAi enzyme complex. Cell2003; 115(2): 199–208

[18] Ro S, Park C, Young D, Sanders KM, Yan W. Tissue-dependent paired expression of miRNAs. Nucleic Acids Res2007; 35(17): 5944–5953

[19] Kozomara A, Griffiths-Jones S. miRBase: integrating microRNA annotation and deep-sequencing data. Nucleic Acids Res2011; 39(Database issue): D152–D157

[20] Shan SW, Fang L, Shatseva T, Rutnam ZJ, Yang X, Du W, Lu WY, Xuan JW, Deng Z, Yang BB. Mature miR-17-5p and passenger miR-17-3p induce hepatocellular carcinoma by targeting PTEN, GalNT7 and vimentin in different signal pathways. J Cell Sci2013; 126(Pt 6): 1517–1530

[21] Uchino K, Takeshita F, Takahashi RU, Kosaka N, Fujiwara K, Naruoka H, Sonoke S, Yano J, Sasaki H, Nozawa S, Yoshiike M, Kitajima K, Chikaraishi T, Ochiya T. Therapeutic effects of microRNA-582-5p and <?Pub Caret?>-3p on the inhibition of bladder cancer progression. Mol Ther2013; 21(3): 610–619

[22] Wang F, Wang XS, Yang GH, Zhai PF, Xiao Z, Xia LY, Chen LR, Wang Y, Wang XZ, Bi LX, Liu N, Yu Y, Gao D, Huang BT, Wang J, Zhou DB, Gong JN, Zhao HL, Bi XH, Yu J, Zhang JW. miR-29a and miR-142-3p downregulation and diagnostic implication in human acute myeloid leukemia. Mol Biol Rep2012; 39(3): 2713–2722

RIGHTS & PERMISSIONS

Higher Education Press and Springer-Verlag Berlin Heidelberg