SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

MicroRNA-122 (miR-122) is a liver-specific microRNA whose expression is specifically turned on in the mouse liver during embryogenesis, thus it is expected to be involved in liver development. However, the role of miR-122 in liver development and its potential underlying mechanism remain unclear. Here, we show that the expression of miR-122 is closely correlated with four liver-enriched transcription factors (LETFs)—hepatocyte nuclear factor (HNF) 1α, HNF3β, HNF4α, and CCAAT/enhancer-binding protein (C/EBP) α—in the livers of developing mouse embryos and in human hepatocellular carcinoma (HCC) cell lines. Correspondingly, promoter analysis revealed that these LETFs are coordinately involved in the transcriptional regulation of miR-122, and three HNFs directly bind to the miR-122 promoter as transcriptional activators. Using a luciferase reporter system, we identified a group of miR-122 targets involved in proliferation and differentiation regulation. Among these targets, the most prominently repressed target was CUTL1, a transcriptional repressor of genes specifying terminal differentiation in multiple cell lineages, including hepatocytes. We show that CUTL1 expression is gradually silenced at the posttranscriptional level during mouse liver development. Overexpression and knockdown studies both showed that miR-122 repressed CUTL1 protein expression in HCC cell lines. Finally, we show that the stable restoration of miR-122 in HepG2 cells suppresses cellular proliferation and activates the expression of three hepatocyte functional genes, including the cholesterol-7α hydroxylase gene (CYP7A1), a known target of CUTL1 in hepatocytes. Conclusion: Our study provides a model in which miR-122 functions as an effector of LETFs and contributes to liver development by regulating the balance between proliferation and differentiation of hepatocytes, at least by targeting CUTL1. HEPATOLOGY 2010

MicroRNAs (miRNAs) are a family of small, noncoding RNAs that have emerged as posttranscriptional regulators of gene expression in animals and plants.1, 2 Generally, miRNAs bind to the 3′ untranslated region (UTR) of target messenger RNAs (mRNAs) in a sequence-specific manner and mediate translational repression or mRNA degradation.1, 2 Hundreds of miRNAs have been identified that participate in the regulation of various biological processes.3, 4 However, although we have recognized the importance of miRNA-mediated gene regulation, the functions and targets of the majority of miRNAs remain unclear.

Some miRNAs are expressed ubiquitously, whereas others are limited to certain stages in development or to certain tissues and cell types.2, 5–7 Recent studies have demonstrated the essential roles of these specific miRNAs in cell fate specification and embryonic development.8-10 MicroRNA-122 (miR-122) is a highly abundant and liver-specific miRNA that accounts for 70% of the total liver miRNA population, but it is undetectable in other tissues.5 Moreover, the expression of miR-122 is strongly up-regulated in the mouse liver during embryonic development.11 Due to these characteristics, it is hypothesized that miR-122 has important roles in liver function and development. However, except for regulating lipid metabolism,12, 13 the known roles of miR-122 are primarily associated with diseases such as hepatitis C virus (HCV) infection14 and hepatocellular carcinoma (HCC).15, 16 The role of miR-122 in healthy animals is unknown, and the contribution of miR-122 to liver development and its regulatory mechanism have not been determined.

Studies concerning the expression of miR-122 during mouse embryonic development showed that its expression initiates at embryonic day 12.5 (e12.5) and increases with time of development, almost reaching a plateau level just before birth.11 This finding suggests that miR-122 likely regulates certain aspects of liver development, primarily from e12.5 to birth. Previous studies have also shown that the bipotential hepatoblasts differentiate into mature hepatocytes or cholangiocytes (also known as biliary epithelial cells) during the same period.17 miR-122 is primarily expressed in hepatocytes,11 and its activation overlaps with hepatocyte differentiation. Therefore, it is highly likely that miR-122 is involved in hepatocyte differentiation.

Although miR-122 was identified several years ago, the transcriptional regulation of miR-122 remains unknown. The expression of tissue-specific genes is usually controlled by tissue-specific/enriched transcription factors. Therefore, we surmised that miR-122 may be transcriptionally controlled by transcription factors enriched in the liver, such as hepatocyte nuclear factors (HNFs) and CCAAT/enhancer-binding proteins (C/EBPs), which play pivotal roles in regulating the expression of liver-specific genes.17-19

In the present study, we primarily focused on the potential role and mechanism of miR-122 in regulating liver development. First, we searched for transcription factors that control the expression of miR-122. Subsequently, we identified miR-122 target genes that were possibly involved in regulating liver development (proliferation and differentiation). Among the validated targets, we chose the transcriptional repressor CUTL1 (also known as CDP [CCAAT displacement protein], Cut, or Cux-1)20 for further investigation. Finally, we employed a lentiviral-mediated stable expression system to confirm the role of miR-122 in regulating hepatocyte proliferation and differentiation. In combination, we placed miR-122 both upstream and downstream of the known gene regulatory network in liver development, which provides an exciting basis for understanding the regulation of liver development. In addition, our results also shed light on the cause and contribution of the down-regulation of miR-122 in HCCs.

Materials and Methods

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

Detailed materials and methods are described in the Supporting Information.

Liver Tissues and Cell Lines.

Fetal, neonatal, and adult livers were isolated from C57BL/6J mice. Cell lines used were HepG2, Huh7, Sk-hep-1, SMMC-7721, and 293FT.

RNA Oligonucleotides and Antibodies.

miR-122 and control mimics were synthesized by GenePharma (Shanghai, China). Anti-miR inhibitors for miR-122 and negative control were obtained from Ambion. All primary antibodies used for chromatin immunoprecipitation assays and western blot analyses of target genes were obtained from Santa Cruz Biotechnology.

Statistical Analysis.

Quantitative reverse-transcription polymerase chain reaction (qRT-PCR) data are expressed as the mean ± standard deviation (SD) and luciferase data are presented as the mean + SD. The differences between groups were analyzed using one-way analysis of variance, with P < 0.05 considered statistically significant (two-tailed).

Results

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

Expression of miR-122 Is Positively Correlated with Expression of Liver-Enriched Transcription Factors in Both the Livers of Developing Mouse Embryos and Human HCC Cell Lines.

To search for regulators that might control in vivo transcription of miR-122, we focused primarily on the transcription factors that play important roles in regulating hepatocyte differentiation during liver development. Among the six families of liver-enriched transcription factors (LETFs) that have been characterized,18 several LETFs (including C/EBPα, HNF1α, HNF3α, HNF3β, and HNF3γ, and HNF4α,) are essential for expression of the complete repertoire of proteins that define hepatocyte function.21 C/EBPα, HNF1α HNF3β, and HNF4α, were further selected because they are highly abundant in both human and mouse liver (Supporting Fig. 1).

We first investigated whether LETF expression correlated with miR-122 levels in both mouse embryonic livers and human HCC cell lines. The expression of miR-122 was detected by way of northern blot analysis and qRT-PCR. As shown in Fig. 1A,B, consistent with the previous report,11 miR-122 was gradually up-regulated in the fetal liver from e12.5 to birth. Compared with the adult mouse liver, miR-122 was down-regulated to different degrees in human HCC cell lines (Fig. 1C,D). Notably, although the miR-122 level in Huh7 cells was also significantly down-regulated, it was still detected on northern blot analysis (Fig. 1C,D).

thumbnail image

Figure 1. Expression of miR-122 is associated with C/EBPα, HNF1α, HNF3β, and HNF4α in the mouse liver and in HCC cell lines. Mouse livers were isolated from five time points: e12.5, e15.5, e18.5, postnatal day 2 (P2), and adult (8 weeks old). Four human HCC cell lines (HepG2, Huh7, Sk-hep-1, and SMMC-7721) were employed. The total RNA was isolated as indicated, and the same RNA samples were used in different analyses. (A,C) Northern blot analysis of miR-122 in mouse livers (A) and human HCC cell lines (C). A total of 30 μg of RNA was loaded. The same membrane was hybridized sequentially with an miR-122 probe and a U6 probe. (B,D) qRT-PCR assays of miR-122 in mouse livers (B) and human HCC cell lines (D). The relative level of miR-122 was normalized to U6 expression in each sample and then compared with the miR-122 level in the e12.5 liver. (E,F) qRT-PCR assays of four LETFs (C/EBPα, HNF1α, HNF3β, and HNF4α) in mouse livers (E) and HCC cell lines (F). The level of LETFs was normalized to glyceraldehyde 3-phosphate dehydrogenase and then compared with the HNF1α level in the e12.5 liver. qRT-PCR data are expressed as the mean ± SD from three separate experiments performed in triplicate.

Download figure to PowerPoint

As shown in Fig. 1E, qRT-PCR data revealed that the expression levels of C/EBPα, HNF1α, HNF3β, and HNF4α gradually increased with time from e12.5 to birth, after which they either continued to increase at a much slower rate or declined slightly. Additionally, the up-regulation of HNF4α and C/EBPα (70-fold and 40-fold, respectively) was more significant than that of HNF1α and HNF3β (<10-fold). As shown in Fig. 1F, the expression of these four LETFs was also significantly down-regulated in HCC cell lines compared with that in the adult mouse liver. Similarly, their expression levels in Huh7 cells tended to be much higher than the levels in the other cell lines.

Overall, it was evident that the expression of miR-122 was strongly correlated with the expression of HNF1α, HNF3β, HNF4α, and C/EBPα, especially the latter two. The data suggest that these LETFs potentially regulate the expression of miR-122.

C/EBPα, HNF1α, HNF3β, and HNF4α Are Coordinately Involved in Transcriptional Regulation of miR-122.

To investigate whether C/EBPα, HNF1α, HNF3β, and HNF4α are involved in the transcriptional regulation of miR-122, we analyzed the promoter region of miR-122. Previous study has shown that miR-122 is derived from the 3′ end of a 4.7-kb noncoding RNA (hcr) in the woodchuck.11 Comparison of genomic sequences across species revealed that the 5′ end of the woodchuck hcr gene22 located in a highly conserved region (≈160 bp), which was included in the predicted promoter (predicted by the FirstEF program, available at the UCSC Genome Browser [http://genome.ucsc.edu/]) of human primary miR-122 (Supporting Fig. 2). Rapid amplification of complementary DNA ends assay revealed that the transcription start sites of human and mouse primary miR-122 locate at the same region as that of the woodchuck (Supporting Fig. 3), also supporting the prediction.

By scanning the predicted human miR-122 promoter with TransFac (BIOBASE gene-regulation.com, Wolfenbuettel, Germany) and Genomatix (Genomatix, Munich, Germany) software, several potential LETF binding sites were identified (Fig. 2A and Table 1).

thumbnail image

Figure 2. C/EBPα, HNF1α, HNF3β, and HNF4α are involved in the transcriptional regulation of miR-122. (A) Schematic representation of predicted binding sites of the LETFs C/EBPα, HNF1α, HNF3β, and HNF4α in the human miR-122 gene. The number indicates the relative distance from the 5′ end of the primary miR-122 transcript (indicated as +1). In the predicted miR-122 promoter, the conserved portion is shown in dark gray. The putative binding sites of each LETF are indicated by different symbols and numbered from the 5′ end of the miR-122 promoter. The validated sites were filled in black. The predicted human miR-122 promoter was cloned into the pGL3-basic vector upstream of the Firefly luciferase gene, whereas the putative enhancer fragments (E1-E3) were inserted between the promoter and the Firefly luciferase. The detailed information on the sequence and the related position of the putative binding sites is shown in Table 1. (B) Luciferase assays show the effects of each LETF on the reporter construct containing the wild-type miR-122 promoter. The constructs were cotransfected into HepG2 cells with vectors expressing LETFs, as well as the pRL-TK vector as a transfection control. A vector expressing red fluorescence protein (dsRed) serves as a negative control for the LETFs. The relative luciferase activities are the ratio of Firefly/Renilla luciferase normalized to the negative control. The data of the luciferase assays are presented as the mean + SD from three separate experiments. (C) Luciferase assays show the effect of C/EBPα on reporter constructs containing the predicted enhancer fragments (E1-E3). (D) Luciferase assays on the mutant promoter constructs. (E) Chromatin immunoprecipitation assays show the in vivo interaction between three HNFs (HNF1α, HNF3β, and HNF4α) and the miR-122 promoter. Huh7 cell chromatin fragments were immunoprecipitated with antibodies for each HNF or negative control antibody (normal mouse immunoglobulin G). The DNA samples from immunoprecipitates were analyzed by way of PCR using primers specific for the 160-bp conserved region of miR-122 promoter. The picture provided here is the result of agarose gel electrophoresis. (F) qRT-PCR assays show the effects of LETFs on the expression of miR-122 in three HCC cell lines (HepG2, Sk-hep-1, and SMMC-7721). Samples were measured in triplicate.

Download figure to PowerPoint

Table 1. Sequences and Positions of LETF Binding Sites in the miR-122 Promoter and the Noncoding Region
Predicted SitesPosition (5′)Wild-Type SequenceMutant Sequence
  1. The genomic position of the miR-122 promoter is chr18:54263944-54264513. The 5′ end of the primary miR-122 transcript was considered as +1. The binding sites within the conserved portion (−155 to +8) are shown in bold, and the validated binding sites are underlined. The C/EBPα binding sites in the promoter that were not mutated as C/EBPα could not activate the promoter.

F4A-1−502CAAAGTGGGA
F1A-1−488GTTAACCG
BPA-1−405CCAAG
BPA-2−194CCACT
F4A-2−146CAAAGTGGGA
F4A-3−122CTAAGTCGGA
F1A-2−90GATAAAGCGG
F3B-1−55TGTTTCACCC
F1A-3−56GTTTAACCCG
Selected FragmentsPosition (5′ to 3′)Regions Enriched with Putative C/EBPα Binding Sites (5′ to 3′)
E1+4202 to +4545+4238 to +4272; +4513 to +4536
E2+1855 to +2338+1965 to +1981; +2177 to +2248
E3+603 to +1071+1019 to +1065

We first checked if these LETFs could act on the predicted promoter using luciferase reporter assays. As shown in Fig. 2B, the miR-122 promoter activated the expression of the downstream reporter up to 20-fold, indicating that the predicted miR-122 promoter was a true promoter. Moreover, compared with red fluorescence protein, HNF1α, HNF3β, and HNF4α further elevated expression of the reporter through this promoter (Fig. 2B). These data suggest that these three HNFs might directly bind to the miR-122 promoter.

C/EBPα may bind to elements outside of the promoter, including intronic regions.19 Therefore, we further analyzed the noncoding region between the promoter and the miR-122 precursor (+38 to +4811) (Fig. 2A). Three fragments enriched with putative C/EBPα binding sites (Table 1) were cloned into the reporter vector downstream of the miR-122 promoter. As shown in Fig. 2C, C/EBPα activated the reporter gene through the E2 fragment.

Next, we performed mutational analyses on predicted HNF binding sites. Because the multiple putative C/EBPα target sites were arranged in a tandem array, we did not perform mutation analysis on these sites. As shown in Fig. 2D, mutagenesis of certain conserved sites (F4A-3, F3B-1, and F1A-3) abolished the effects of HNFs on the reporter, but mutations of all nonconserved sites made no difference. Remarkably, the F4A-3 site was a crucial site, because its mutation completely abolished the miR-122 promoter function. The F3B-1 and F1A-3 sites overlapped (Table 1), and mutations in these sites eliminated the effects of both HNF1α and HNF3β. These data demonstrate that the HNFs could directly bind to the miR-122 promoter. This conclusion was further confirmed by way of chromatin immunoprecipitation assay. As shown in Fig. 2E, the three HNFs directly bind to the miR-122 promoter in Huh7 cells.

To test whether the LETFs could up-regulate miR-122 expression in HCC cells, we performed overexpression studies. As shown in Fig. 2F, cells transfected with LETF-expressing vectors display an obvious up-regulation of miR-122, especially for C/EBPα. Moreover, this finding is consistently observed in the three cell lines used. Together, these results show that C/EBPα, HNF1α, HNF3β, and HNF4α are involved in the transcriptional regulation of miR-122, which also suggests that miR-122 functions as an effector of these LETFs during liver development.

A Group of Predicted Genes Involved in Proliferation and Differentiation Regulation Are Potential Targets of miR-122 During Liver Development.

Cellular proliferation and differentiation are the two most important processes for organ development.23 Numerous studies have established the pivotal roles of LETFs in the regulation of both processes during liver development.17-19 To search for the functional targets of miR-122, we primarily focused on candidate target genes with the potential to suppress differentiation and/or promote proliferation, which are contrary to the roles of LETFs. Eleven candidate targets were arbitrarily selected from the results predicted by Targetscan 4.2 for further confirmation (Table 2). In addition, CCNG1 and BCL2L2, two known targets of miR-122, were employed as positive controls.16, 24

Table 2. Candidate miR-122 Target Genes
TargetMolecular PropertiesFunctions
  • Eleven candidate targets were selected from the Targetscan 4.2 prediction results based on both their potential functions and the presence of conserved miRNA binding sites. CCNG1 and BCL2L2 are known targets of miR-122. Nine of the targets also present in the PicTar results are shown in bold.

  • *

    References S1-S7 are provided in the Supporting Information.

CUTL1Transcriptional factorRepressor of genes specifying terminal differentiation during development
CTCFTranscriptional factorImplicated in diverse regulatory functions, including the regulation of lineage-specific gene expression
SRFTranscriptional factorPlays a key role in activating immediate early genes and thereby participates in cell cycle regulation, apoptosis, cell growth, and cell differentiation
MAP3K3Protein kinaseComponent of MAPK cascades; central regulator of cell fate during development
MAP3K12Protein kinaseComponent of MAPK cascades; central regulator of cell fate during development
MARK1KinaseMay play a role in cytoskeletal stability; another isoform of MARK (MARKL1) is elevated in HCCs in which the Wnt-signaling pathway was activated
VAV3Nucleotide exchange factorProto-oncogene; exchange factor for GTP-binding proteins RhoA and RhoG, thus playing a pivotal role in many aspects of cellular signaling, coupling cell surface receptors to various effector functions
LAMC1Extracellular matrix proteinPotent modulator of numerous biological processes in development, including cell proliferation, migration, and differentiation; highly expressed in transformed hepatoma cell lines
RAD21Nuclear matrix proteinInvolved in chromosome cohesion during cell cycle, in DNA repair, and in apoptosis; suppression of RAD21 decreases cell growth and enhances cytotoxicity of etoposide and bleomycin in breast cancer cells
CLIC4Chloride ion channel proteinRegulates fundamental cellular processes; suppression of the CLIC family induces apoptosis, enhances tumor necrosis factor α–induced apoptosis, and inhibits tumor growth
MSN (moesin)Linker proteinA member of the ERM family, which function as cross-linkers between plasma membranes and cytoskeletons; the ERM-organized complexes control many cellular activities, including proliferation, cell communication, motility, and differentiation
CCNG1Cyclin proteinKnown target of miR-122; cell cycle regulator
BCL2L2BCL-2 family proteinKnown target of miR-122; antiapoptosis bcl-2 family

The 3′-UTR segments of each target were synthesized and subcloned downstream of the Renilla luciferase in the psiCHECK-2 dual luciferase reporter vector (Fig. 3A), and reporter assays were performed as indicated. Surprisingly, as shown in Fig. 3B, 11 reporters were significantly repressed by miR-122 to different degrees (30%-70% reduction), including the two known targets. MSN (moesin) and serum response factor (SRF) were not significant in this group. These data indicate that most candidate genes could be directly repressed by miR-122. To further confirm this hypothesis, we performed mutational analyses on each predicted site. As expected, all mutant sites lost their response to miR-122 (Fig. 3C), indicating the specificity of the repression.

thumbnail image

Figure 3. Identification of miR-122 target genes. (A) Diagram depicting the 3′-UTR reporter constructs. The wild-type or mutant 3′-UTR fragments (47-bp) were inserted into the psiCHECK-2 vector downstream of the Renilla luciferase. The forward (top) 47-nt 3′-UTR fragments included 7-8 nt seed-matched sites (ACACTCCA, ACACTCC, or CACTCCA), 10-nt 3′ flanking sequences and 29–30 nt 5′ flanking sequences. The mutation was introduced by replacing the 6-nt core seed-matched sites (CACTCC) with its complementary bases (GTGAGG). (B,C) Luciferase assays show the effects of miR-122 on the reporter constructs containing the wild-type (B) and mutant (C) 3′-UTR fragments of each candidate target gene. MT, mutant; WT, wild-type. The 3′-UTR constructs were cotransfected into 293FT cells with either miR-122 or miR-neg expressing vector. The pcDNA6.2-miR-neg plasmid (miR-neg) was used as a negative control for pcDNA6.2-miR-122 (miR-122). The relative luciferase activities are the ratio of Renilla/Firefly luciferase normalized to the negative control for each target group. The data of the luciferase assays are presented as the mean + SD from three separate experiments. The targets are ranked by the repression ratio from wild-type constructs. (D) qRT-PCR assays show the expression of target genes in the mouse liver. Three time points (e12.5, e18.5, and adult) were chosen. The level of target genes was normalized to glyceraldehyde 3-phosphate dehydrogenase and then compared with the HNF1α level in the e12.5 liver. qRT-PCR data are presented as the mean ± SD from triplicate samples.

Download figure to PowerPoint

To investigate whether these genes were expressed in liver cells, we performed qRT-PCR assays to detect their mRNA levels in the mouse liver. As shown in Fig. 3D, mRNAs were expressed in the mouse liver throughout the development, indicating that they are potential functional targets of miR-122 during liver development.

Notably, mRNA levels in the majority of genes (8 of 12) were not inversely correlated with miR-122 expression during liver development (Fig. 3D), suggesting that they might be regulated by miR-122 posttranscriptionally. To verify this hypothesis, we selected seven genes whose mRNA levels were significantly reduced from e18.5 to adulthood for further confirmation, but none of the transcripts was significantly affected by miR-122 (Supporting Fig. 4).

CUTL1 Is an In Vivo Target of miR-122 During Liver Development.

To investigate whether the identified targets play important roles in liver cells, we selected one target for further investigation. CUTL1 was chosen for three reasons: (1) it is the most prominently repressed target of miR-122 in our reporter screening (Fig. 3B); (2) the binding site for miR-122 within its 3′-UTR is the most conserved site among the candidate targets we selected (Supporting Fig. 5); and (3) it is a known transcriptional repressor of genes specifying terminal differentiation during development.20, 25

Western blot analysis revealed that the CUTL1 protein (p200) was clearly detected in early stage mouse embryonic livers (e12.5 and e15.5), whereas it was barely detectable after birth (Fig. 4A). Interestingly, the amount of CUTL1 protein gradually disappeared during the progression of development, which was inversely correlated with the expression of miR-122 (Fig. 1A). Moreover, we also observed a similar correlation between the CUTL1 protein abundance (Fig. 4B) and the miR-122 level (Fig. 1D) in human HCC cell lines. CUTL1 protein was highly expressed in HepG2, 7721, and Sk-hep-1 cells, whereas it was weakly expressed in Huh7 cells.

thumbnail image

Figure 4. CUTL1 is a biological target of miR-122 during liver development. (A,B) Western blotting analysis shows the expression of CUTL1 protein in mouse livers (A) and human HCC cell lines (B). The molecular weight of the full-length CUTL1 protein (p200) is ≈200 kDa. Two CUTL1 antibodies were used, and the same results were obtained. The blots were stripped and reprobed for β-actin as a loading control. (C) Western blotting showing the effect of overexpression (left) or knockdown (right) of miR-122 on CUTL1 protein expression. Overexpression experiments were performed in four HCC cell lines. The knockdown experiment was only performed in the Huh7 cells for their high-level expression of miR-122. HepG2 cells were transfected with 100 nM miR-122 (122) or the control mimic (NC). Huh7 cells were transfected with 100 nM miR-122 inhibitor (122-Inhibitor) or control inhibitor (NC-Inhibitor). All western blotting experiments were repeated at least twice, and reproducible results were obtained.

Download figure to PowerPoint

To confirm that the CUTL1 protein is suppressed by miR-122, we performed both overexpression and knockdown experiments on miR-122 in HCC cells. As shown in Fig. 4C, when HepG2, 7721, and Sk-hep-1 cells were transfected with miR-122 mimics, the CUTL1 protein was significantly reduced. Alternatively, when Huh7 cells were transfected with the miR-122 inhibitor, the CUTL1 protein expression apparently increased (Fig. 4C). These data strongly support the interpretation that CUTL1 is an in vivo target of miR-122 in hepatocytes.

Restoration of miR-122 in HepG2 Cells Suppresses Cellular Proliferation and the Continuous High Level of miR-122 Induces Hepatocyte Differentiation.

To assess whether miR-122 contributes to liver development, we investigated whether miR-122 suppresses cellular proliferation and promotes differentiation, because these functions are contrary to the roles of CUTL1 but are similar to the roles of the LETFs. The HepG2 cell line was employed because it is derived from a hepatoblastoma, which represents an early stage in the cellular lineage pathway26 and is more closely related to immature/undifferentiated hepatoblasts.

We first performed transient transfection studies. Cell cycle analysis revealed that transfection of miR-122 mimics into HepG2 cells only led to a slight change in the distribution of the cell cycle stages, as indicated by the minimal increase in the number of cells in G1/S phase (Supporting Fig. 6A). These data suggest that miR-122 did not regulate the cell cycle progression directly or its effect was long-term and gradual. As hypothesized, the suppression effect of miR-122 on cell proliferation gradually appeared after 10 days of culturing, with a more than 50% decrease in cell number compared with the control (Supporting Fig. 6B).

Considering the limit of transient transfection, we also employed stable overexpression studies using lentiviral vectors. As schemated in Fig. 5A, we generated two different lentiviral vectors. qRT-PCR data showed that the miR-122 level in HepG2 stable cells (G2-122x1 and G2-122x4) increased significantly compared with the control (Fig. 5B). Notably, miR-122 levels increased a further four-fold in G2-122x4 cells compared with G2-122x1 cells.

thumbnail image

Figure 5. Restoration of miR-122 in HepG2 suppressed cellular proliferation and induced differentiation. (A) Schematic diagram showing the structural difference between the two vectors carrying one copy (122x1) or four copies (122x4) of miR-122 precursor, respectively. (B) qRT-PCR analysis shows the expression level of miR-122 in HepG2 stable cells compared with normal HepG2 or Huh7 cells. G2-122x1, infected with lentiviral carrying 122x1; G2-122x4, infected with lentiviral carrying 122x4. (C) Cell growth assays show the effect of the stable expression of miR-122 on the proliferation of HepG2 cells. Stable cells were seeded in 60-mm dishes at 20,000 cells per dish in quadruplicate and maintained in full growth medium for 7 days. The cells were fixed and stained with crystal violet at day 1, 3, 5, and 7. Images shown are representative photomicrographs taken on day 7. Experiments were repeated twice, and reproducible results were obtained. G2-mock, normal HepG2; G2-neg, HepG2 infected with lentiviral expressing miR-neg. (D) Qualitative (left) and quantitative (right) growth assays compare the proliferation rate between G2-mock and G2-122x4 cells. The method of cell growth assays is the same as above. After the photomicrograph was taken, the staining was resuspended in 10% acetic acid and sample absorbance was measured at 630 nm. The relative cell number was calculated, normalized to their samples in day 1 in each respective group. Samples were measured in triplicate. (E) High level of miR-122 resulted in morphological changes in HepG2 cells. Left panel: qRT-PCR assay show the expression level of miR-122 in G2-122x8 (B7) cells. Right panel: photomicrographs show the morphological difference between G2-neg cells and G2-122x8 (B7) cells. G2-122x8 (B7) was a purified clone of HepG2 stable cells infected with lentivirus carrying eight copies (122x8) of miR-122 precursor. The miR-122 level was normalized to that in G2-122x4. G2-neg cells were served as the normal morphological control. Photomicrographs (magnification ×100) were taken during subculturing under both phase contrast (top) and fluorescence (bottom) objectives at the same visual field. (F) qRT-PCR shows the expression of three cytochrome P450 family genes (CYP1A2, CYP2C9, CYP7A1) in HepG2 stable cells. Samples were measured in triplicate.

Download figure to PowerPoint

As shown in Fig. 5C, the proliferation of G2-122x4 and G2-122x1 cells was significantly suppressed, as indicated by the obvious reduction in the cell numbers compared with the control or G2-neg cells. Moreover, the stable cells expressing higher levels of miR-122 (G2-122x4) achieved more significant growth repression (Fig. 5C,D). Because miR-122 levels in G2-122x4 cells were close to the levels in Huh7 cells (Fig. 5B), which were far lower than those in adult liver, we speculated that miR-122 may strongly arrest or even stop the proliferation as long as its abundance reaches a sufficient level.

Considering this finding, we further generated stable cells possessing eight copies of miR-122 precusor (G2-122x8). Because the effectiveness of infection was unequal for each cell, we picked cell clones with bright green fluorescence, which indicates more copies of viral integration. The miR-122 levels in clone B7 increased a further four-fold compared with G2-122x4 (Fig. 5E). Interestingly, at the beginning, these cells grew slowly and required more than 2 weeks to form a clone that was capable of being passaged, whereas normal HepG2 cells required only 10 days. Moreover, the cells became more and more fragile over time. More than two-thirds of the cells died after trypsinization before the cell number was high enough for cryopreservation. Although we did not perform proliferation assays, these data have demonstrated that a continuous high level of miR-122 strongly affects the proliferative capacity of HepG2 cells.

In addition, we observed morphological changes on G2-122x8 cells. The normal HepG2 or G2-neg cells showed a typical epithelial-like morphology (Fig. 5E). Due to the strong refraction, the nuclear and cell boundaries could be seen clearly under a phase contrast microscope. Although growing in high density, they were still orderly organized. In contrast, G2-122x8 cells displayed irregular cell shape, heterogeneous cell size, and disorganized arrangement. The nuclear and cell boundaries could not be seen clearly. Moreover, many dead cells or cell debris with bright green fluorescence were floating above the living cell layer. These phenomena suggest that G2-122x8 cells might have been undergoing apoptosis or/and differentiation.

To determine whether miR-122 promoted hepatocyte differentiation, we quantified the mRNA expression of three cytochrome P450 family genes (CYP1A2, CYP2C9, and CYP7A1) that are hepatic functional proteins specifically expressed in mature hepatocytes.18, 27, 28 Notably, CYP7A1 is a known target of CUTL1 in HepG2 cells.27 As shown in Fig. 5F, in G2-122x4 cells, only the expression of CYP7A1 increased, whereas in G2-122x8 (B7) cells, all three cytochrome P450 genes were significantly up-regulated. This result indicates that the continuous high level of miR-122 eventually induces the differentiation of hepatoblastoma cells. In addition, it suggests that CUTL1 is an important functional target of miR-122. In combination, our studies suggest that the activation of miR-122 plays an important role in guiding hepatocyte differentiation during development.

Discussion

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

This study demonstrates that four LETFs (C/EBPα, HNF1α, HNF3β, and HNF4α) are involved in the transcriptional regulation of miR-122, which could directly regulate a group of target genes involved in proliferation and differentiation regulation. In line with this, restoration of miR-122 in hepatoblastoma cells suppresses cellular proliferation and activates the expression of hepatocyte functional genes. We show that CUTL1 is a biological target of miR-122 during liver development. Our findings support a role of miR-122 in liver development, as shown in Fig. 6. According to this model, miR-122 acts as an important bridge connecting the two different types of regulators that control the balance between the proliferation and differentiation of hepatocytes during liver development.

thumbnail image

Figure 6. Model of the role of miR-122 in liver development. The differentiation of bipotential hepatoblasts into hepatocytes or biliary epithelial cells begins around e12.5 of mouse development. Hepatocyte differentiation is decided by hepatogenic transcription factors, including several HNFs and C/EBPα.17 In undifferentiated hepatoblasts, miR-122 is gradually activated by these LETFs with the progress of liver development (e12.5 to birth). With the increase of miR-122 expression, the levels of differentiation repressors (such as CUTL1) are gradually silenced; therefore, the transcriptional repression of genes specifying terminal differentiation (such as cell cycle inhibitors20, 23 and hepatocyte functional genes) is gradually relieved. Finally, the hepatoblasts gradually exit the cell cycle and differentiate into mature hepatocytes. In addition, a proportion of hepatoblasts may proliferate for a defined period before they begin terminal differentiation to ensure that the adult liver achieves a sufficient cell number.

Download figure to PowerPoint

The transcriptional regulation of the majority of miRNAs is currently unknown. Because miR-122 is the most abundant and specific miRNA in the liver, clarification of its regulatory mechanism is necessary to reveal the transcriptional regulation of liver miRNAs. Here, we provide the first direct evidence that the transcription of miR-122 is regulated by several LETFs. The involvement of several transcription factors in the transcriptional regulation of a single miRNA has not been reported previously. However, this mechanism is a common principle for liver gene regulation.17, 18

While our study was underway, others identified LETFs as central regulatory molecules in gene networks associated with the loss of miR-122 in human HCCs, and their knockdown experiments suggest that miR-122 is under the transcriptional control of HNF1α, HNF3α, and HNF3β.29 However, because their evidence is indirect, their findings could not determine the regulatory system nor the underlying mechanism. For example, HNF4α and C/EBPα, two important regulators of miR-122 identified in our studies, were not found to be significant in their data.

The physiological role of miR-122 in liver development is currently unknown, primarily because no appropriate targets have been identified. Understanding the molecular mechanisms that regulate cellular proliferation and differentiation is a central theme of developmental biology.9, 23 In this report, we identified that a group of genes involved in proliferation and differentiation regulation are miR-122 targets. Several target genes are considered key regulators of development, such as the two transcription factors (CUTL1 and CCCTC-binding factor [CTCF]) and two mitogen-activated protein kinase kinase kinase (MAP3K) members25, 30, 31 that have been shown to be targets of miRNA. Therefore, our work is significant because it provides important clues for understanding the role of miR-122 during liver development.

During the development of a multicellular organism, cells proliferate for a defined length of time before they begin functional differentiation.23 The process of differentiation of primitive cells into more specialized cells involves an increasing restriction in proliferative capacity, culminating in cell cycle exit.23 Precise regulation of terminal cell division is needed to ensure production of proper numbers of differentiated cells at the appropriate time.23 CUTL1, the target we focused on, is a conserved transcriptional repressor that regulates the balance between cell division and differentiation of multiple cell lineages during embryonic development.20, 25 CUTL1 knockout and transgenic mouse models have confirmed this role.25 The majority of homozygous mice die at or shortly after birth due to severe hypoplasia, whereas transgenic mice constitutively expressing CUTL1 develop multiorgan organomegaly (including the heart, kidney, testis, spleen, seminal vesicle, and liver).25 In hepatomegaly, constitutively expressing CUTL1 results in an excessive increase in the number of immature hepatocytes.32 These studies suggest that CUTL1 is necessary for embryonic development at an early stage, whereas failure to turn off its activity leads to excessive proliferation, as well as differentiation blocking of primitive cells. Researchers have determined that CUTL1 activity (also known as HiNF-D binding activity) is down-regulated during fetal liver development, coinciding with the exit from the cell cycle and terminal differentiation.33 However, the mechanism is unclear. Here, we show that CUTL1 expression is silenced posttranscriptionally during mouse liver development, likely due to repression by miR-122. Therefore, our study not only reveals the mechanism regulating CUTL1 during liver development, but also supports the role of miR-122 in the precise regulation of terminal cell division and differentiation of hepatocytes.

In addition, other targets identified in this study are also notable. SRF is a ubiquitous nuclear protein that regulates the activity of many immediate-early genes.34 While our study was underway, SRF was confirmed by others to be a target of miR-122.35 Our western blot data support this finding (Supporting Fig. 7) even though we could not obtain a significant result in the reporter screen. CTCF is a highly conserved transcription factor implicated in diverse regulatory functions.30 Recent studies suggest that CTCF may be a heritable component of an epigenetic system regulating the interplay between DNA methylation, higher-order chromatin structure, and lineage-specific gene expression.30 MAP3K3 and MAP3K12 are components of protein kinase signal transduction cascades that transduce extracellular signals into a wide range of cellular responses (including differentiation, proliferation, and apoptosis) and could therefore be central regulators of cell fate during development.31 Both of the transcription factors and the MAPK pathways regulate a large number of genes20, 30, 31, 34; therefore, miR-122 may modulate the global gene expression profile during liver development through these targets.

The interesting question regarding the role of miR-122 in the adult liver remained unanswered for many years. Due to the abundance of miR-122 in the liver, it is believed to play an important role in the maintenance of the adult liver phenotype. However, the mechanism is unclear. Our data show that miR-122 targets, such as CUTL1 and SRF, are transcriptionally active in the adult liver but their protein expression is almost silenced. Therefore, miR-122 may be needed to suppress those genes that are normally repressed but may be essential in mature hepatocytes. Furthermore, maintenance of cell cycle arrest in terminally differentiated cells is important for tissue architecture and function.23 In the adult liver, the majority of hepatocytes rarely undergo proliferation; approximately one mitotic hepatocyte can be identified per 20,000 hepatocytes throughout the liver acinus.26 Our data show that the restoration of miR-122 expression in HCC cells significantly limits cellular proliferation. Meanwhile, the correlation between the proliferation suppression and the miR-122 level is evident, suggesting that the high abundance of miR-122 may be responsible for limiting the cell cycle of mature hepatocytes.

Great interest was aroused by the evidence that the deregulation of miRNAs correlates with various human cancers.36 miR-122 is particularly notable because it is highly expressed in normal liver but is frequently down-regulated in human HCC.15, 16 Several groups have shown that the down-regulation of miR-122 in HCC cells is correlated with tumorigenic properties (such as growth, antiapoptotic activity, migration, invasion clonogenic survival, replication potential, and tumor formation).16, 24, 29, 35, 37 Our findings suggest that the down-regulation of miR-122 is due to the aberrant expression of LETFs. The latest view suggests that liver cancers arise either by the dedifferentiation of mature cells or by the maturation arrest of stem cells.26 Given the role of miR-122 in the developmental liver, we believe that loss of miR-122 expression may be primarily involved in the misregulation of the balance between cell proliferation and differentiation during hepatocarcinogenesis.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

We thank Stephen J. Elledge (Howard Hughes Medical Institute) for graciously providing the p203 (pPRIME-TREX-GFP-FF3) vector and Shi-Mei Zhuang (Sun Yat-Sen University) for kindly donating the Huh7 cell line.

References

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

Additional Supporting Information may be found in the online version of this article.

FilenameFormatSizeDescription
HEP_23818_sm_Suppfig1.tif12713KSupplementary Figure 1. The expression of all known LETFs in human (A) and mouse (B) tissues. We search the six families of LETFs in the UCSC Genome Browser by Gene Sorter, and their expression data was downloaded and gathered.
HEP_23818_sm_Suppfig2.tif2314KSupplementary Figure 2. BLAT result shows the conservation of primary transcript for miR-122 between various animals. Sequence for the query is the known genomic sequence of woodchuck hcr (primary transcript for miR-122). Sequence information see the reference 22.
HEP_23818_sm_Suppfig3.tif4194KSupplementary Figure 3. 5'RACE results show the transcription start sites of human (Huh7 cell) and mouse pri-miR-122. A. Agarose gel electrophoresis shows the PCR products of 5'RACE. B. Alignment of genomic sequence shows the detailed transcription start sites of pri-miR-122. The translation start sites of pri-miR-122 are underlined. In woodchuck, it located at S1 site. In human (Huh7 cell), pri-miR-122 could start at two sites (S2 and S5), whereas it could start at four sites in mouse (S1, S2, S3 and S4). The numbers indicate the frequency of each site in the sequenced clones. 8 and 12 clones were sequenced for human and mouse RACE products respectively.
HEP_23818_sm_Suppfig4.tif162KSupplementary Figure 4. qRT-PCR assays show the effect of the miR-122 mimic (122) on the expression of target genes in HepG2 cells. The level of target genes was normalized to GAPDH and then compared to the samples transfected with control mimic (NC)
HEP_23818_sm_Suppfig5.tif4099KSupplementary Figure 5. Multiple sequence alignment of the binding site for miR-122 within CUTL1 3'UTR in 43 species. This picture is a result of Blat from the UCSC Genome Browser. The sequence for Blat is the 47-bp fragment of CUTL1 gene that was used for target validation (5'-TGG GTT TTG CAG ACC AGG GTT TGT TTA ATA CAC TCC ATT CTA GGC CA -3', it is also shown in Supplementary Table 3.). The putative binding site for miR-122 seed is shown as green box in the picture.
HEP_23818_sm_Suppfig6.tif3831KSupplementary Figure 6. A. Flow cytometry analysis shows the effect of the miR-122 mimic on the cell cycle of HepG2 cells. Cells were transfected with 50 nM miR-122 mimic (122) or the control mimic (NC) and were collected, stained and analyzed 48 h after transfection. B. Cell growth assay shows the effect of the miR-122 mimic on the proliferation of HepG2 cells. Cells were transfected with 50 nM of the mimics. 24 h following transfection, cells were trypsinized and seeded into 6-well-plates at 3000 cells per well and maintained for 10 days. Then, the cells were fixed and stained by crystal violet.
HEP_23818_sm_Suppfig7.tif680KSupplementary Figure 7. Western blot assay shows the expression of SRF in developing mouse livers, HCC cell lines as well as the effect of over-expression of miR-122 on SRF protein expression. All the samples used were the same as before.
HEP_23818_sm_Supptable.doc159KSupplementary Table 1. Real-time PCR primers for miR-122, U6, LETFs and GAPDH. Some primers are different between mouse (M) and human (H) and the different bases are underlined.
HEP_23818_sm_Suppinfo.doc711KSupporting Information.
HEP_23818_sm_Suppreference.doc27KSupporting Information.

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.