Zinc finger transcription factor 191, directly binding to β-catenin promoter, promotes cell proliferation of hepatocellular carcinoma

Authors


  • Potential conflict of interest: Nothing to report

Abstract

Activation of β-catenin, the central effector of the canonical wingless-type (Wnt) pathway, has been implicated in hepatocellular carcinoma (HCC). However, the transcription regulation mechanism of the β-catenin gene in HCC remains unknown. Here we report that human zinc finger protein 191 (ZNF191) is a potential regulator of β-catenin transcription. ZNF191, a Krüppel-like protein, specifically interacts with the TCAT motif, which constitutes the HUMTH01 microsatellite in the tyrosine hydroxylase (TH) gene ex vivo. We demonstrate that ZNF191 is significantly overexpressed in human HCC specimens and is associated with growth of human HCC cells. Global profiling of gene expression in ZNF191 knockdown human hepatic L02 cells revealed that the important Wnt signal pathway genes β-catenin and cyclin D1 messenger RNAs (mRNAs) are significantly down-regulated. In agreement with transcription level, β-catenin and cyclin D1 proteins are also down-regulated in transient and stable ZNF191 knockdown L02 and hepatoma Hep3B cell lines. Moreover, significant correlation between ZNF191 and β-catenin mRNA expression was detected in human HCCs. Promoter luciferase assay indicated that ZNF191 can increase transcription activity of the full-length β-catenin (CTNNB1) promoter, and nucleotide (nt)-1407/-907 of the CTNNB1 promoter exhibited the maximum transcriptional activity. Electrophoretic mobility shift assay showed that purified ZNF191 protein can directly bind to the CTNNB1 promoter, and the binding region is located at nt-1254/-1224. Finally, we demonstrate that the key binding sequence of ZNF191 in vivo is ATTAATT. Conclusion: ZNF191 can directly bind to the CTNNB1 promoter and activate the expression of β-catenin and its downstream target genes such as cyclin D1 in hepatoma cell lines. This study uncovers a new molecular mechanism of transcription regulation of the β-catenin gene in HCC. (HEPATOLOGY 2012;55:1830–1839)

Hepatocellular carcinoma (HCC) is a primary cancer of the liver and the fifth most common cancer worldwide, which is predominant in developing countries, with nearly 600,000 deaths each year worldwide.1 Various risk factors have been associated with HCC, including infection with hepatitis B virus and/or hepatitis C virus,2 aflatoxin B intake,3 heavy alcohol intake,4 hemochromatosis.5 The pathogenesis of the development and progression of HCC is far from being clear presently, and several cellular signal transduction pathways are involved in HCC, such as wingless-type (Wnt)/β-catenin, p53, pRb, mitogen-activated protein kinase (MAPK), Ras pathway.6 Of these pathways activated in HCC, the canonical Wnt pathway is one of most frequently reported.1

In canonical Wnt signaling pathway, β-catenin is the central player. Under unstimulated conditions, β-catenin is phosphorylated by interactions with glycogen synthase kinase 3β (GSK-3β), and forms a destruction complex with axin and the adenomatous polyposis coli protein (APC).7, 8 Mutations in the N-terminal region of β-catenin can prevent its phosphorylation and subsequent degradation, and this stabilizes the protein and the mutant protein accumulates in the nucleus, and causes an elevated level of constitutive transcriptional activation by β-catenin/TCF complexes, which contributes to liver carcinogenesis.9 In HCC aberrant activation of the canonical Wnt//β-catenin signaling pathway includes mutations in β-catenin, Axin1, Axin2, or APC genes.10-12 However, some studies have revealed that 35%-80% of HCCs with β-catenin nuclear and cytoplasmic accumulation is not associated with these gene mutations. This phenomenon implies that the pathway may be activated by some other mechanisms.9, 13, 14

β-Catenin accumulation leads to activation of target genes, such as cyclin D1, c-Myc, implicated in human cancer.15-17 In addition to numerous studies that focused on β-catenin protein stabilization and subcellular localization, some studies reported that β-catenin messenger RNA (mRNA) levels were elevated in human cancers including HCC.16, 18 This suggests that transcription deregulation of the β-catenin gene itself may be an important factor during tumor development. However, only several transcription factors have been identified with high-affinity binding to the CTNNB1 promoter, such as AP1, LEF/TCF, NKX2-5, TRβ,16, 19, 20 which have been reported to be involved in some physiological and pathophysiological processes. However, the mechanism of transcription regulation of β-catenin gene in HCC remains unknown.

The human zinc finger protein 191 (ZNF191), a Krüppel-like protein, specifically interacts with the TCAT motif, which constitutes the HUMTH01 microsatellite in the tyrosine hydroxylase (TH) gene ex vivo.21, 22 However, the target genes and the binding sequence of this transcription factor in vivo remains unknown. In this work we demonstrate for the first time that transcription factor ZNF191, overexpressed in human HCC, can positively regulate transcription of β-catenin by way of directly binding to the CTNNB1 promoter. We further demonstrate that the key binding sequence of ZNF191 in vivo is ATTAATT. The findings suggest a possible role of ZNF191 in association with cell proliferation of HCC.

Abbreviations

β2-MG, β-microglobulin; CCND1, cyclin D1; CTNNB1, β-catenin; HCC, hepatocellular carcinoma; mRNA, messenger RNA; PCR, polymerase chain reaction; RNAi, RNA interference; SD, standard deviation; siRNA, small interfering RNA; Wnt, wingless-type; ZNF191, zinc finger transcription factor 191.

Materials and Methods

Additional experimental procedures are described in the Supporting Information.

Microarray Analysis.

Total RNA of cell line L02 with transient and stable ZNF191 knockdown was analyzed with Affymetrix HG_U133_ Plus2.0 microarrays. Data quality control was performed with Affymetrix Microarray Suite 5.0 and was normalized with Robust Multichip Analysis software. Genes with a negative or positive fold change of 1.5 times or more in the transient and stable knockdown group were further analyzed. Microarray datasets were annotated and analyzed with Database for Annotation, Visualization, and Integrated Discovery (DAVID) bioinformatics resources.

Electrophoretic Mobility Shift Assay (EMSA).

5′ biotin-labeled oligonucleotides were synthesized and labeled by Invitrogen, China. Various length fragments of CTNNB1 P(-1407/-957) were amplified by reverse-transcription polymerase chain reaction (RT-PCR) with the synthesized probes. DNA binding activity was detected using a LightShift Chemiluminescent EMSA kit (Pierce, USA). For competition assays, unlabeled oligonucleotides were included in the binding reaction. The sequences of the oligonucleotides used for these binding studies are listed in Supplemental Table 1.

Chromatin Immunoprecipitation (ChIP) Assay.

For in vitro ChIP assay, HEK-293T were transfected with 10 μg expression vector pCMV-Myc-ZNF191 and control vector and harvested 48 hours later. Physical associations between mammalian expression ZNF191 and CTNNB1 promoter were analyzed using a ChIP assay kit (Millipore, USA). The in vivo ChIP assay was performed using Hep3B cells. The crosslinked protein-DNA complexes were immunoprecipitated with corresponding antibodies. Isolated DNA was subjected to RT-PCR and analyzed by running agarose gels, stained with ethidium bromide, followed by ultraviolet (UV) visualization. The primers used in the amplification are listed in Supplemental Table 1.

Results

ZNF191 is Up-Regulated in HCCs.

We identified ZNF191 as one of the genes that are up-regulated (14 out of 18) in human liver cancers in comparison to adjacent noncancerous tissues during our initial screening for HCC relevant genes by northern blot analysis (Fig. 1A). In agreement with the higher mRNA level of ZNF191 in HCC, the ZNF191 protein was also up-regulated in four out of six HCC (Fig. 1B). These findings were supported by quantitative real-time PCR analysis of ZNF191 mRNA expression in 44 paired HCCs (Fig. 1C). In all, 22 of 44 (50%) cases showed significant up-regulation of ZNF191in HCC, 17 of 44 (38.6%) cases showed no alteration, and only 5 of 44 (11.4%) cases showed reduction. Thus, we demonstrated by various approaches that both ZNF191 mRNA and protein are frequently overexpressed in human HCCs. Finally, western blot analysis revealed that the ZNF191 protein was also readily detected in the majority of HCC cell lines examined (Fig. 1D).

Figure 1.

Increased expression of ZNF191 in human HCC. (A) ZNF191 mRNA expression levels were analyzed in 18 paired HCC specimens with their corresponding noncancerous specimens by northern blot. (B) ZNF191 protein expression levels were analyzed in another six paired HCCs by western blot. (C) ZNF191 mRNA expression levels were analyzed in 44 paired HCCs (nos. 1-44, numbered according to relative expression level from low to high) by real-time PCR and the 2-ΔΔCt method. Values are expressed as log2 transformed relative fold decrease or fold increase in the mRNA expression with respect to the adjacent nontumorous tissues after normalization to β2-MG. A 2-fold change threshold was set for identifying significant changes in gene expression. (D) Expression of ZNF191 protein in 13 HCC cell lines was examined by western blot.

Knockdown of ZNF191 Suppresses Cell Growth of Human HCC Cell Lines.

To obtain insight on ZNF191 function, we employed a loss-of-function approach to assess the role of ZNF191 in HCC cell growth. We constructed hairpin RNA expression vectors pSUPER-EGFP-si-ZNF191 for functional ZNF191 small interfering RNAs (siRNAs) to assess the long-term effect of ZNF191 knockdown on growth of HCC cells in vitro and in vivo. ZNF191 stable knockdown clones and control clones of L02 and Hep3B cell lines were selected for further analysis (Fig. 2A). As shown in Fig. 2B, ZNF191 stable knockdown induced a reduction of the cell number in the S phase. Consistent with cell cycle results, 3-(4,5-dimethyl-thiazol-2yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium (MTS) assays showed that cell numbers of stable ZNF191 knockdown clones decreased versus controls as the culture time was prolonged (Fig. 2C). In xenograft mouse models, as shown in the left panels of Fig. 2D,E, ZNF191 knockdown resulted in a significant decrease in the volume of L02 and Hep3B tumors. Consistently, the size and weight of ZNF191 knockdown tumors were much smaller than control tumors (Fig. 2D,E, middle and right panels). Taken together, these data suggest that stable knockdown of ZNF191 suppresses cell proliferation and that ZNF191 is associated with cell growth of human HCC cell lines.

Figure 2.

ZNF191 is associated with cell growth of human HCC cell lines. (A) Left, 2 ZNF191 stable knockdown clones (pS-si-Z2, pS-si-Z7) of L02 cell line and 1 control clone (pS-Scram) were identified and analyzed by western blot. Right, two ZNF191 stable knockdown clones (pS-si-Z1, pS-si-Z2) of Hep3B cell line and one control clone were also identified and detected. (B) Flow cytometry analysis of cell cycle of L02 (left) and Hep3B (right) with ZNF191 stable knockdown. The fractions of viable cells in the G0-G1, S, and G2-M phases of the cell cycle were quantified by flow cytometry analysis of propidium iodide-stained cells. Data represent the mean ± SD, n = 3, *P < 0.05, **P < 0.01. (C) Effect of ZNF191 stable knockdown on cell proliferation of L02 or Hep3B cells, as measured by MTS assay in a 7-day culture period. (D) Effect of ZNF191 on growth of L02 xenograft in nude mice. Xenografts were established by subcutaneous injection of indicated clones to the nude mice. Six mice were used in each group (n = 6). The left panel shows the tumor volume of xenografts at the indicated days after injection. The middle panel shows the resected tumors arising from individual nude mouse. Weight of the resected tumors is shown in the right panel. (E) Effect of ZNF191 on growth of Hep3B xenograft in nude mice. Statistical analysis shows dramatic inhibitory effect of ZNF191 on tumor weight (**P < 0.01).

Microarray Analysis of Gene Expression Pattern in ZNF191 Knockdown Cell Line L02 Reveals that ZNF191 Is Associated with Wnt Signal Pathway.

In order to explore the role of ZNF191 in HCC, we searched for ZNF191 target genes by using transient and stable knockdown strategies in L02 with microarray analyses. Figure 3A shows that endogenous ZNF191 protein level was substantially reduced at 48 hours posttransfection of L02 cells with ZNF191 siRNAs (Si-ZNF191) when compared with scrambling control siRNAs (Scram-si). Then we compared two groups of transcriptome of L02 cells: transient knockdown group (Si-ZNF191 versus Scram-si) and stable knockdown group (pS-si-Z7 versus pS-Scram) with oligo-microarray (Affymetrix HG_U133_ Plus2.0). After statistical selection of the transcripts regulated in both groups, and functional annotations of genes among the transcript using DAVID, we demonstrated that in total 152 genes were regulated. The main regulated genes are listed in Supporting Table 1. The microarray results were confirmed by real-time PCR analysis of five selected genes of interest: ZNF191, CTNNB1, CCND1, HSPA9, BMP1 (Fig. 3B).

Figure 3.

Microarray analysis of gene expression pattern in transient ZNF191 knockdown cell line L02 reveals that ZNF191 is associated with the Wnt signal pathway. (A) The effect of transient knockdown of ZNF191 expression in the L02 cell line was assessed by western blot. Mock indicated cells have not transfected with any siRNA complexes. (B) Five regulated genes in L02 in response to ZNF191 transient (left panel) and stable (right panel) knockdown with microarray analysis was confirmed by real-time PCR. (C) The effect of overexpression of ZNF191 on the Wnt/β-catenin signaling pathway in HEK-293T cells was assessed by luciferase activity assays. HEK-293T cells were transfected with various amounts (0, 100, 200, 400, 600 ng/well) of pCMV-Myc-ZNF191 and Top-flash-Luc. pRL-SV40 was used as the internal control in each transfection. Mock pCMV-Myc was used as a negative control. Data represent the mean ± SD, n = 3, *P < 0.05, **P < 0.01.

The altered genes of ZNF191 knockdown are involved in processes of the regulation of cell cycle, signal transduction, transcription, cartilage/skeletal development, etc. Among them, several of the altered genes are involved in the Wnt signal pathway, including CTNNB1, CCND1, etc. The mRNA level of CTNNB1 and CCND1 was significantly down-regulated in both groups. Next we performed a luciferase activity assay to examine the effect of overexpression of ZNF191 on Wnt signal pathway activity. Figure 3C shows that ZNF191 alone can increase Wnt-responsive TCF/LEF reporter Top-flash-Luc activity by 3.16-fold in HEK-293T cells. Moreover, this activation was in a dose-dependent manner.

ZNF191 Positively Regulates β-Catenin Expression in Human HCC Cells.

Because ZNF191 knockdown can decrease mRNA levels of CTNNB1 and CCND1 in L02 cells, and ZNF191 can promote Wnt signal pathway activity, it is not clear whether ZNF191 regulates the expression level of β-catenin and cyclin D1 proteins. We performed a series of western blots analyzing their relationships. As expected, β-catenin and cyclin D1 protein levels increased in L02 cells with transient overexpression of ZNF191 protein (Fig. 4A). Consistently, in transient and stable ZNF191 knockdown L02 and Hep3B cells, β-catenin and cyclin D1 proteins were down-regulated as the endogenous ZNF191 protein level decreased when compared with controls (Fig. 4B,C). Moreover, ZNF191 siRNA resistant complementary DNA (cDNA) can rescue the suppression of β-catenin and cyclin D1 proteins in transient ZNF191 knockdown L02 and Hep3B cells (Supporting Fig. 1).

Figure 4.

ZNF191 can positively regulate β-catenin expression in human HCC cells. Western blot analysis of ZNF191, β-catenin, cyclin D1, and β-actin (loading control) in (A) L02 cells transfected with pCMV-Myc-ZNF191 and mock pCMV-Myc vectors; (B) L02 (left) and Hep3B (right) cells with ZNF191 transient knockdown; (C) L02 (left) and Hep3B (right) cells with ZNF191 stable knockdown. (D) The mRNA level of ZNF191, β-catenin, cyclin D1 in 44 pairs of human HCC and paracancerous tissues was examined by real-time PCR. (E) Scatter plot showing the correlation of ZNF191 with β-catenin and cyclin D1 in 44 human HCC tissues. Data were analyzed using the paired t test, and correlation analysis was performed using Pearson's correlation test. (F) Transient knockdown of ZNF191 in HCC cells inhibits serum induction of β-catenin mRNA expression. L02 and Hep3B cells were transiently transfected with control siRNA (Scram-si) or ZNF191 siRNA (Si-ZNF191), then at 24 hours after transfection cells were incubated with free-serum medium for another 20 hours. After that, cells were treated with 10% fetal bovine serum (FBS) as a mitogenic factor for the indicated timepoints. Expression of the β-catenin transcript in L02 (left) and Hep3B (right) cells was analyzed by real-time PCR. Data represent the mean ± SD, n = 3, *P < 0.05, **P < 0.01.

To further confirm the correlation of ZNF191 and β-catenin in HCC in vivo, we analyzed mRNA expression of the ZNF191, β-catenin and its downstream genes in the Wnt/β-catenin pathway (cyclin D1 and c-Myc) in the 44 pairs of human HCCs, as mentioned above. Up-regulation of ZNF191 was concomitant with enhanced β-catenin expression (Fig. 4D), and significant statistical correlation was observed between the two genes (Fig. 4E). Like cyclin D1 (Fig. 4E, right), statistically significant correlation was also observed between ZNF191 and c-Myc, but not the nontarget gene of the Wnt pathway, STAT4 (signal transducer and activator of transcription 4) (Supporting Fig 2). Simultaneously, we investigated the mutational status of the exon 3 of the β-catenin gene, which encodes the GSK-3β phosphorylation site of the β-catenin gene. The results showed wildtype of β-catenin exon 3 sequences in all tumors (Supporting Fig. 3).

Given that ZNF191 regulates β-catenin mRNA and protein expression and is associated with the proliferation of HCC cells and that de novo synthesis of β-catenin mRNA can be induced by serum,23 we used serum as a mitogenic factor to induce HCC cells to proliferate and analyzed β-catenin mRNA expression in L02 and Hep3B cells. A rapid and marked induction of β-catenin mRNA was observed after serum treatment, which was maximal at 8 hours in L02 cells and 4 hours in Hep3B cells after stimulation and declined afterward (Fig. 4F). However, these responses were largely inhibited when ZNF191 was transiently knocked down (Fig. 4F). The results suggest that ZNF191 may act as a mediator of serum induction of β-catenin mRNA expression in HCC cells.

ZNF191 Increases Transcription Activity of the CTNNB1 Promoter.

It is clear that ZNF191 can positively regulate mRNA and protein levels of β-catenin. Next we sought to determine the mechanism of this regulation. To this end we assessed whether overexpression of ZNF191 has any effect on transcription activity of the CTNNB1 promoter. Promoter luciferase assay indicated that ZNF191 can increase the transcription activity of the full-length CTNNB1 promoter (PGL3-HBCP, gift of Prof. R.H. Dashwood, Oregon State University) by about 3.5-fold compared with transfecting control vector (Fig. 5A). Furthermore, this activation was in a dose-dependent manner (Fig. 5B). Compared with the full-length isoform of ZNF191 (ZNF191-FU), the short isoform of ZNF191 (ZNF191-NF, without C2H2 zinc finger domain) had no activation effect on CTNNB1 promoter (Fig. 5B). This result suggests that ZNF191 exerts this activation function role by way of C2H2 zinc finger domain. Because cyclin D1 is the downstream gene of β-catenin, we assessed the effect of ZNF191 on CCND1 promoter. Figure 5C shows that ZNF191 increased CCND1 promoter by 6.4-fold. Mutation in the LEF/TCF site (the binding site of β-catenin) of the CCND1 promoter resulted in a much lower increase (3.2-fold) in transcription activity. In vivo ChIP assays showed that ZNF191 cannot directly bind to the CCND1 promoter (-962CD1), including the LEF/TCF site of the CCND1 promoter (Supporting Fig. 4). The results suggest that activation of CCND1 promoter by ZNF191 is through β-catenin, but not through direct binding of endogenous ZNF191 to the promoter.

Figure 5.

ZNF191 activates CTNNB1 promoter. (A) The effect of ZNF191 on transcriptional activity of full-length CTNNB1 promoter PGL3-HBCP [P(-2692/+93)] was examined in HEK-293T cells by luciferase reporter assays. HEK-293T cells were transfected with PGL3-Basic, PGL3-HBCP plus mock pCMV-Myc vectors, and PGL3-HBCP plus pCMV-Myc-ZNF191vectors, respectively. The relative luciferase activities were obtained against the promoter activities of transfecting PGL3-HBCP plus mock pCMV-Myc vectors. (B) ZNF191 activates CTNNB1 promoter in a dose-dependent manner, ZNF191-NF served as negative controls. HEK-293T cells cotransfected with PGL3-HBCP, pRL-SV40, various amounts (0, 100, 200, 400, 600 ng/well) of pCMV-Myc-ZNF191-FU or pCMV-Myc-ZNF191-NF vectors, and mock pCMV-Myc vectors. The relative luciferase activities were measured as described above. (C) The effect of ZNF191 on transcriptional activity of wildtype CCND1 promoter (-962CD1) and mutant type CCND1 promoter (mTCF, mutant in LEF/TCF site). (D) Schematic view of the luciferase reporter constructs containing various lengths of the 5′-flanking region of the β-catenin gene. (E) Identification of response regions for ZNF191 in the CTNNB1 promoter. HEK-293T cells were transfected with pCMV-Myc-ZNF191 and the reporter constructs as indicated. The value represents the mean ± SD of three independent transfection experiments, each performed in triplicate, *P < 0.05, **P < 0.01.

Next, in order to identify ZNF191 response regions in the CTNNB1 promoter, various lengths of CTNNB1 5′-flanking region (Fig. 5D) were transfected into HEK-293T cells with pCMV-Myc-ZNF191 to determine the promoter transcriptional activities. The luciferase reporter assay indicated that the construct P(-1407/+93) exhibited the maximum luciferase activity, which was much higher than that of P(-2692/+93) and P(-1907/+93). P(-907/+93) and P(-409/+93) constructs displayed modest promoter activity (Fig. 4E). These results suggest that nucleotide (nt)-1407/-907 of the CTNNB1 promoter region is indispensable to elicit transcriptional response for ZNF191.

ZNF191 Directly Binds to the CTNNB1 Promoter.

The finding that potential binding sites for ZNF191 are located at nt-1407/-907 of CTNNB1 promoter region prompted us to determine whether ZNF191 is directly binding to the CTNNB1 promoter. With delicate analysis of the nucleotide sequences of the 5′-flanking region (-1467/-907) of the β-catenin gene (Fig. 6A, top), we found that sequences ATTAATT at nt-1244 of the CTNNB1 promoter are similar to ATTCATT (within three repetitions of [TCAT] motif, TCATTCATTCAT, defined previously as ZNF191 interacting motif21). We hypothesized that ZNF191 may directly bind to the CTNNB1 promoter at this candidate site (Fig. 6A). To this end we expressed recombinant ZNF191 protein24 (Supporting Fig. 5), and confirmed the authenticity of the protein by mass spectrometry (Supporting Table 3). We employed an EMSA assay to examine their binding capabilities. Various length fragments (183 bp, 260 bp, and 119 bp) of CTNNB1 P(-1407/-957) were amplified with designed primers and prepared for EMSA (Fig. 6A). As we expected, purified ZNF191 protein bound to P(-1407/-1224) and P(-1254/-994) (183 bp and 260 bp, respectively), both of which have a common 30-bp sequence (nt-1254/-1224) containing the candidate ZNF191 binding sequence ATTAATT (Fig. 6B). Purified ZNF191 protein bound to annealed double-stranded 30-bp oligomer F2/R1 (Fig. 6C). Then we mutated the seven key nucleotides to AAAATAA (Fig. 6A, bottom), compared with high-affinity binding of wildtype F2/R1 to purified ZNF191 protein. We found that mutF/R has no binding capacity to ZNF191 (Fig. 6C). Next we performed a ChIP assay to examine the physical interaction of ZNF191 to the CTNNB1 promoter in vitro and in vivo. Figure 6D (left panel) shows that the DNA sequence harboring ectopic expression ZNF191 protein in the CTNNB1 promoter was immunoprecipitated in HEK-293T cells. Figure 6D (right panel) shows direct binding of endogenous ZNF191 to the promoter in Hep3B cells. These results further confirm that ZNF191 can directly bind to the CTNNB1 promoter.

Figure 6.

ZNF191 directly binds to CTNNB1 promoter. (A) Top, the nucleotide sequences of the 5′-flanking region (-1467/-907) of the β-catenin gene. The candidate ZNF191 binding site is boxed, and primers F1, R1, F2, R2, F3, R3 used in EMSA analysis are underlined. Bottom, ZNF191 binding sequence of the first intron of the tyrosine hydroxylase (TH) gene and its similar sequence in CTNNB1 promoter. Arrow indicates the base changes of mutant sequences in mutF/R and in the CTNNB1 promoter: mP(-2692/+93) and mP(-1407/+93) for further analysis. (B) DNA binding activity of purified ZNF191 protein to various length fragments (183 bp, 260 bp, and 119 bp) of CTNNB1 P(-1407/-957) was detected by EMSA. Arrow indicates specific bands of purified ZNF191 binding to the various length fragments. (C) DNA binding activity assay of purified ZNF191 protein to 30-bp annealed F2/R1 oligonucleotides of CTNNB1 promoter and its mutant (mut F/R). Arrow indicates specific band of purified ZNF191 binding to the oligonucleotides. (D) ZNF191 binding to CTNNB1 promoter was examined by in vitro (left panel) and in vivo (right panel) ChIP experiments as described in Materials and Methods. (E) The effect of ZNF191 on transcriptional activity of mutated CTNNB1 promoters. Activity of CTNNB1 promoters [P(-2692/+93), P(-1407/+93)], and their mutated types [mP(-2692/+93), mP(-1407/+93)] was measured by luciferase reporter assays as described in Fig. 5.

To further examine whether the candidate binding sequences ATTAATT are the key sequences of ZNF191 binding to the CTNNB1 promoter, we examined the effect of ZNF191 on transcriptional activity of mutated type CTNNB1 promoters mP(-2692/+93) and mP(-1407/+93) with mutating key sequences ATTAATT to AAAATAA. As demonstrated in Fig. 6E, the activation ability of mutated CTNNB1 promoters by ZNF191 was markedly reduced compared with that of wildtype promoters. Taken together, these data suggest that ZNF191 can directly bind to the CTNNB1 promoter, and the key sequences of binding site are ATTAATT.

Discussion

Human ZNF191, also known as ZNF24,25 belongs to the SCAN domain subfamily of Krüppel-like zinc finger transcription factors,21, 22 which shows 94% identity to its mouse homolog zinc finger protein 191 (Zfp191).26 In recent years several groups have studied the functions of the gene, which is involved in embryonic development,27-29 hematopoiesis,30 and is a negative regulator of VEGF.25 Khalfallah et al.27, 29 reported that ZNF191/Zfp191 mRNA principally expressed in proliferative area, especially in the early stages of the human or mouse embryonic ventricular zone of brain and spinal cord. ZNF191 knockdown in human neural progenitors can inhibit proliferation and leads to exit from the cell cycle. Li et al.28 generated mice that are deficient in Zfp191 and found that homozygous Zfp-/- embryos cannot survive beyond embryonic day 7.5 without clear cause of lethality. These results suggest that ZNF191 plays an important role in cell proliferation and differentiation during embryonic development. In this study we demonstrate for the first time that ZNF191 is overexpressed in human HCCs (Fig. 1A-C). Through loss of function studies, we found that stable knockdown of ZNF191 suppresses cell growth of human HCC cell lines L02 and Hep3B in vitro and in vivo (Fig. 2). Together, these results show that ZNF191 may be associated with cell proliferation of human HCC. Our findings are consistent with previous studies that ZNF191 may play an essential role in cell proliferation during embryonic development.27-29

Next, through microarray analysis we found that ZNF191 can regulate Wnt/β-catenin pathway in the L02 cell line (Fig. 3B,C). β-Catenin, the key gene of the pathway, and its target gene cyclin D1, were positively regulated by ZNF191 (Figs. 3-5). Previous studies showed ZNF191 can specifically interact with the intronic polymorphic TCAT repeat in the TH gene, the microsatellite HUMTH01. Allelic variations of HUMTH01 correlated with quantitative and qualitative changes in the binding by ZNF191 and the minimal binding motif is a (TCAT)3 repeat.21 With this hint, we finally identified that purified ZNF191 protein can directly bind to the CTNNB1 promoter, and the key binding sequence of ZNF191 in vivo is ATTAATT (Figs. 5, 6). The identified new binding sequence will throw new light on exploring novel target genes of ZNF191 in vivo, and will be very important in studying biological function of the transcription factor.

Previous studies have reported several transcription factors responsible for transcription of β-catenin.16, 19, 20 In this study, with a series of methods (Figs. 5, 6), we identified a new member, ZNF191, as a positive transcription factor that directly regulates the expression of β-catenin gene in hepatoma cell lines. The findings that up-regulation of ZNF191 is closely correlated with elevation of β-catenin in HCC specimens, and ZNF191 can activate β-catenin in HCC cell lines (Figs. 1-4), suggests that ZNF191 may function through up-regulating β-catenin and its downstream target cyclin D1 in HCC, and therefore promote tumor cell proliferation in vivo. The results may explain the phenomenon that β-catenin mRNA levels were elevated in some HCC.18 Thus, we observed a novel mode of mechanism involved in the control of β-catenin abundance in HCC in addition to known proteasomal-mediated degradation.7, 8 It is worth noting that the mechanism for up-regulated expression of ZNF191 in HCC remains unknown and warrants further investigation.

Ancillary