To test whether miR-181s were transcriptional targets of Wnt/β-catenin, we first examined the expression of β-catenin and miR-181s in five different HCC cell lines. Figure 1A and 1B showed that β-catenin and miR-181s were concordantly expressed in these cell lines. Moreover, we tested the association of miR-181 and β-catenin using Klotho mice where loss of Klotho, a secreted Wnt antagonist, leads to β-catenin activation 11 . Consistently, levels of all 4 mature miR-181s in the liver tissues of Klotho mice were higher than those of wild-type mice (figure 1C). Therefore, both in vivo and in vitro data demonstrate that the expression of miR-181s is positively correlated with the Wnt/β-catenin signaling pathway.

To examine whether Wnt/β-catenin signaling could directly induce miR-181 expression, we first used HuH7 cells with stable WNT10B expression (HuH7 WNT10B R8) 12 . Compared to control HuH7 cells, HuH7 WNT10B R8 had 6.2-, 4.1-, 8.5-, and 4.4- fold higher levels of miR-181a, miR-181b, miR-181c and miR-181d, respectively (figure 2A). Moreover, conditioned media derived from HuH7 WNT10B R8 also significantly induced the expression of miR-181s in HuH7 cells (figure 2A). Second, lithium chloride (LiCl), an activator of the Wnt/β-catenin pathway through inhibiting glycogen synthase kinase (GSK)-3β then consequently stabilizing β-catenin 13 , was applied to further examine the role of Wnt/β-catenin pathway in regulating miR-181s. We found that LiCl significantly induced expression of all 4 miR-181 transcripts in HuH7 (figure 2B) and HuH1 cells (data not shown) when compared to NaCl control treated cells. Additionally, miR-181s were significantly induced by ectopic expression of a constitutively activated β-catenin or a wild type Tcf4, a co-transcriptional activator of β-catenin, in HuH7 (figure 2C) and HuH1 cells (data not shown). As a control, a dominant negative Tcf-4 mutant failed to do so. These data indicate that Wnt/β-catenin signaling activation directly induces miR-181 expression.

Several known proteins such as adenomatosis polyposis coli (APC) can promote β-catenin degradation 14 . To investigate the effect of Wnt/β-catenin inactivation on the expression of miR-181s, we used HT29-APC cells where the APC gene is under the control of a zinc-activated metallothionein promoter 14 . The levels of all four miR-181s were significantly decreased in HT29-APC cells following addition of zinc chloride (ZnCl₂), but not in control HT29-GAL cells (figure 3A). Furthermore, β-catenin siRNA was employed to suppress Wnt/β-catenin activity in HuH1 and HuH7 cells. Consistently, declined β-catenin expression by siRNA (Figure 3B and 3C) significantly reduced the expression of all four miR-181 transcripts (figure 3D). These data demonstrate that an inactivation of Wnt/β-catenin signaling results in an inhibition of miR-181 expression.

Studies have shown that the β-catenin/Tcf complex transcriptionally regulates their target gene expression through binding to a consensus core TCF/LEF-binding site (5'-A/T A/T CAAAG-3') within the promoter regions 15 16 . GenBank database analyses indicate that miR-181a-1 and miR-181b-1 are located in the intron 2 of a novel host gene (RP11-31E23.1-001) on 1q31 (data not shown), while miR-181a-2 and miR-181b-2 are located in the intron 2 of a novel transcript, RP11-348K2.1-001, on 9q33 (figure 4A). We searched the core TCF/LEF binding site within 4kb genomic sequences of the predicted promoter regions of these two genes. Seven TCF/LEF binding elements were identified in the promoter region of RP11-348K2.1-001 (figure 4A), and three elements were in the promoter region of RP11-31E23.1-001 (data not shown). MiR-181c and miR-181d are present in the intergenic region of chromosome 19 and consequently no promoter region is available for analysis. Furthermore, RP11-348K2.1-001 was selected for chromatin immuoprecipitation (ChIP) assay using Tcf4 antibody in Hep3B cells where both β-catenin and miR-181s are highly elevated. We designed two amplicons within a 200-bp window, i.e., amplicon U (in the upstream of transcription site) and amplicon D (in the downstream of transcription site) (figure 4B). The results showed that Tcf4 specifically bound to both amplicons in the miR-181a-2/miR-181b-2 promoter (figure 4C), indicating that the miR-181a-2/miR-181b-2 transcript is a direct transcriptional target of Wnt/β-catenin canonical signaling pathway in HCC.

Constitutively active β-catenin expression vector (pCI-NEO-βcateninXL) (mutated at S33Y), wild-type Tcf4 expression vector (pcDNA/myc-hTcf4) and dominant negative Tcf4 expression vector (pcDNA/myc-ΔN-hTcf4) were generous gifts from Dr. Bert Vogelstein 20 . β-catenin siRNA and negative control siRNA were purchased from Qiagen. Anti-β-catenin monoclonal antibody (BD Transduction Laboratories), anti-β-actin monoclonal antibody (Sigma) and anti-Tcf-4 monoclonal antibody (Chemicon) were used in this study. Lithium chloride, sodium chloride and zinc chloride were commercially available (Sigma).

Human HCC cell lines (Hep3B, HuH7, HuH1, MHCC97 and Smmc7721) were cultured as previously described 6 . HT29-APC, HT29-GAL cells were kindly provided by Dr. Bert Vogelstein and routinely cultured as described 14 15 . To induce APC's expression, we incubated HT29-APC cells with 100 μmol/L of zinc chloride for 24 and 48 hours. HuH7 cells with stably expressed WNT10B (WNT10B clone R8) were maintained in our lab 12 . WNT10B conditioned media were prepared from HuH7 WNT10B R8. Three wild-type and two Klotho knock out mice were kindly given by Dr. Toren Finkel. Liver tissues were dissected from these mice and immediately stored at -80°C until RNA extraction. Transfection of siRNAs and plasmids was done using Lipofectamine 2000 (Invitrogen) according to the manufacturer's instrument. A total of 100nM siRNAs or 2 μg/ml plasmids were used for each transfection. Total RNA was extracted three days after transfection.

Total RNA was extracted using Trizol (Invitrogen). Reverse transcription was done using High Capacity cDNA Reverse Transcription Kit for mRNAs and TaqMan MicroRNA Reverse Transcription Kit for miRNAs. The expression of mRNAs and mature miRNAs was measured using Taqman Gene Assay specific for CTNNB1 and Taqman MicroRNA Assays specific for miR-181a, miR-181b, miR-181c and miR-181d (Applied biosystems). The experiments were performed in triplicate. The Taqman gene assay for 18S and Taqman MicroRNA Assay for U6 RNA were used to normalize the relative abundance of mRNA and miRNA, respectively.

Total protein was extracted as previously described 6 . They were electrophoresed on 10% Tris-Glycine Gel (Invitrogen) and transferred to nitrocellulose membrane (Invitrogen). The membrane was incubated with anti-β-catenin and anti-β-actin, followed by anti-mouse IgG conjugated to horseradish peroxidase. The bands were developed using enhanced chemiluminescence detection reagents (Amersham).

ChIP assay was done using the Chromatin Immunoprecipitation Assay Kit (Upstate), Hep3B cells and anti-Tcf4 antibody according to the manufactures' instrument. The oligonucleotides flanking transcription start site and predicted Tcf4 binding sites were designed within ~200-bp windows in the 700-bp upstream of the transcription start site (amplicon U), and in the 500-bp windows downstream of transcription start site (amplicon D) for detecting miR-181a-2 and miR-181b-2 promoter. The primers were 5'-aaa caa agc aac tgc cat gta ct-3' (sense) and 5'-agg gtc ctt ggg gtt tat ttt-3' (antisense) for amplicon U, and 5'-ggt tgc ttt cca ttc ttt gc-3' (sense) and 5'-ttt ggc cac aca aaa cta gag at-3' (antisense) for amplicon D. The PCR was done using Platinum Taq Polymerase High Fidelity (Invitrogen).

Student's t-test (two-tailed) was used for statistical analysis of comparative data between different groups. p < 0.05 was considered to be statistically significant. Data were presented as the mean ± standard deviation (S.D.) and the analyses were performed using GraphPad Prism software 5.0 (GraphPad Software, San Diego, CA).