Potential conflict of interest: Nothing to report.
MicroRNAs (miRNAs) are approximately 22-nucleotide noncoding RNAs that constitute silencers of target gene expression. Aberrant expression of miRNA has been linked to a variety of cancers, including hepatocellular carcinoma (HCC). Hepatitis C virus (HCV) infection is considered a major cause of chronic liver disease and HCC, although the mechanism of virus infection–associated hepatocarcinogenesis remains unclear. We report a direct role of miRNAs induced in HCV-infected primary human hepatocytes that target the tumor suppressor gene DLC-1 (a Rho GTPase-activating protein), which is frequently deleted in HCC, and other solid human tumors. MicroRNA miR-141 that targets DLC-1 was accentuated in cells infected with HCV genotypes 1a, 1b, and 2a. We present several lines of evidence that efficient HCV replication requires miR-141–mediated suppression of DLC-1. An increase in miR-141 correlated with the inhibition of DLC-1 protein in HCV-infected cells. Depletion of miR-141 with oligonucleotides complementary to the miRNAs inhibited virus replication, whereas artificially increased levels of intracellular miR-141 enhanced HCV replication. HCV-infected hepatocytes showed enhanced cell proliferation that can be countered by overexpression of DLC-1. Conclusion: The collective results of this study suggest a novel mechanism of HCV infection–associated miRNA-mediated regulation of a tumor suppressor protein that has the ability to influence cell proliferation and HCV infection–mediated liver cancer. (HEPATOLOGY 2011)
MicroRNAs (miRNAs) originate from highly structured primary transcripts of RNA Pol II genes by way of two-step processing events involving RNase III type nucleases. Primary miRNA transcripts are processed in the nucleus by the RNase III type endonuclease Drosha into precursor and exported to the cytoplasm by exportin 5, to be secondarily cleaved into miRNA duplexes by the cytoplasmic RNase type III Dicer. The resulting miRNA duplexes are incorporated into the RNA-induced silencing complex, where one of the miRNA strands, the passenger, is degraded, while the guide strand complementary to the target messenger RNA (mRNA) serves in target selection and silencing, either by degradation (in case of perfect base complementarity) or inhibition of translation (in case of imperfect sequence complementarity).1 Thus, the expression of miRNAs in cell type–specific fashion shapes mRNA profiles.
Hepatitis C virus (HCV) is among the most successful of human pathogens. HCV persists in the vast majority of infected individuals as a major cause of chronic hepatitis, liver cirrhosis, and hepatocellular carcinoma (HCC) worldwide. The HCV genome is a positive-sense ≈9.6-kb RNA consisting of a single open reading frame that encodes a large polyprotein complex that is proteolytically cleaved to produce 10 viral proteins. The highly basic N-terminal one-third includes core, envelope glycoproteins E1 and E2, and the integral transmembrane protein p7. The remaining two-thirds of HCV polyprotein include nonstructural proteins NS2, NS3, NS4A, NS4B, NS5A, and NS5B. The NS5B protein functions as RNA-dependent RNA polymerase.2 HCV infection triggers expression of host genes of innate antiviral defense whose levels vary widely among patients and possibly with different degrees of liver fibrosis and cirrhosis,3 suggesting that HCV can both trigger and control host defenses during viral infection. Because HCV infection is critically linked to the development of HCC, a major challenge in understanding hepatocarcinogenesis is to identify functionally relevant cellular mRNAs that are targeted by miRNAs.4, 5 We surveyed miRNAs that are accentuated in HCV-infected primary human hepatocytes and observed up-regulated miRNAs that target the DLC-1 tumor suppressor, which is frequently deleted in HCC.
DLC-1 encodes a Rho-GTPase activating protein and is a candidate tumor suppressor gene located on chromosome 8p21.3-22. DLC-1 is recurrently deleted in HCC and other human tumors.6DLC-1 was originally isolated and characterized from human HCC by polymerase chain reaction (PCR)-based subtractive hybridization.7 The GTPase activity of DLC-1 is specific for RhoA, a member of the Ras family of oncogenes.8 Restoration of DLC-1 in hepatoma cell lines lacking DLC-1 showed reduced cell proliferation as well as reduced metastatic activity.9 Xue et al.6 examined a mosaic mouse model to demonstrate that the loss of DLC-1 in hepatoblasts coexpressing Myc and lacking p5310, 11 promotes cell transformation in vitro and/or tumorigenesis in vivo. These studies demonstrated that loss of DLC-1, when combined with other oncogenic lesions, promotes HCC in vivo.
The chronic role of DLC-1 as tumor suppressor has been established on the basis of its inactivation by deletion, point mutations, or promoter hypermethylation. However, it is less clear how HCV infection, a major etiologic agent of HCC, acutely targets DLC-1 expression in human hepatocytes. We report that the miRNAs miR-141 and miR-200a are accentuated in HCV-infected human primary hepatocytes and can target DLC-1 mRNA to suppress protein expression. This miRNA-mediated regulation may represent an early event in HCV tumorigenesis in primary human hepatocytes.
Long-term cultures of primary human hepatocytes were maintained in a defined medium that supported productive replication of HCV as described.12
RNA Isolation, Reverse-Transcription, and Quantitative PCR.
Total cell RNA was extracted using TRIzol LS Reagent (Invitrogen) according to the manufacturer's instructions and processed for viral RNA quantitation by way of nested reverse-transcription polymerase chain reaction (RT-PCR). Nested PCR amplification of HCV was performed using a set of external and internal primers for HCV genotypes 1a, 1b, and 2a as described.13
Western Blot Analysis of DLC-1 Protein.
Primary hepatocyte cultures were transfected with HCV genotype 1a (HCV1a) genomic RNA (1.0 μg/106 cells) in triplicate. At the indicated times thereafter, the total cell proteins were separated by 4%-12% gradient gel electrophoresis and transferred to nitrocellulose membranes for immunoblotting. The blots were first blocked with 5% nonfat dry milk in Tris-HCl buffer (pH 7.5) containing 0.1% Tween-20 and then probed with the primary antibodies against DLC-1 protein (mouse monoclonal anti-human; BD Bioscience) for 1 hour. After extensive washes, the blots were incubated with secondary antibodies conjugated with horseradish peroxidase for 1 hour. Protein bands were developed by way of enhanced chemiluminescence reagent and were detected using X-ray film.
miRNA Detection by Way of RNA-Primed, Array-Based Klenow Enzyme Assay.
Primary human hepatocyte cultures were transfected with genomic RNAs of HCV genotypes 1a, 1b, and 2a (1 μg/106 cells) using FuGENE6 (Roche). On day 6 postinfection, the small RNA (≤200-nucleotide) fraction was enriched from HCV-infected cell RNA using a mirVana isolation kit (Ambion). Four micrograms of each sample together with positive control (synthetic Arabidopsis thaliana mir-157a, which is not present in the human genome) was spiked in and was hybridized to the microarray slide (BioMicro System). After 16 hours, the hybridized microarray was washed with a standard sodium citrate solution to remove unhybridized probes. After 3 hours of Klenow exonuclease-1 incubation, exo(-) Klenow enzyme was added to extend the miRNAs hybridized to the chip-attached templates in a primer extension step. During this step, biotinylated dATP was incorporated as a final portion of the extension through the designed polythymidine region. Detection of this template-hybridized miRNA was performed using streptovidin-conjugated Alexa-fluor-555, which binds to the biotinylated stretch of A's at the 3′-end of the captured miRNA. Fluorescence data sets were collected using GenePix 4000 scanner (Axon). Details of the procedure are described in Yeung et al.14
Flow Cytometry of Ki67 Nuclear Antigen Fluorescence Stained with Fluorescein Isothiocyanate.
Primary hepatocytes were transfected with HCV1a genomic RNA (1 μg/106 cells) in triplicate. Parallel cultures were transfected with DLC-1 complementary DNA (cDNA) expression vector (50 ng/106 cells for 6 hours) prior to transfection with HCV 1a genomic RNA. Six days posttransfection, the cells were released with 0.05% trypsin treatment and were resuspended at 104/100 μL in (phosphate-buffered saline containing 2% fetal bovine serum) processed for Ki67 immunostaining (BD Biosciences) according to the manufacturer's instructions.
Infection of Naïve Hepatocytes with Virus from Cell Culture Media.
Primary human hepatocytes were transfected with HCV genotypes 1a, 1b, and 2a (1 μg/106 cells) as described.12 Virus released in the culture medium was filtered through 0.25-μm filters from infected cells.12 Viral RNA replication was evaluated at indicated times after infection as outlined above, and the efficiency of virus released in the culture media was validated using the World Health Organization's HCV standards (Acrometrix, Benicia, CA).
Luciferase Reporter Assay.
Primary human hepatocyte culture was cotransfected with luciferase reporter containing DLC-1 3′ untranslated region (UTR) (50 ng/106 cells), miR-141 (50 nM/106 cells, antagomir) or miR-141 (50 nM/106 cells, Mimic) using Lipofectamine 2000 (Invitrogen). Luciferase assays (Promega) were performed on the third day after transfection according to the manufacturer's instructions.
Statistical Analysis and Reproducibility.
The results are given as the mean ± SE. Statistical analysis of the data was performed using the Student t test, Fisher's exact test, or otherwise as described.
Replication of HCV Genotypes 1a, 1b, and 2a Alters the Expression of miR-141 and miR-200a in Primary Human Hepatocytes.
To assess virus infection-associated changes in host gene expression, we analyzed alterations in miRNAs in primary human hepatocytes infected with HCV genotypes 1a, 1b, and 2a (Supporting Information Fig. 1). We measured the changes in miRNA levels following infection with full-length genomic clones of HCV using the RNA-primed array-based Klenow extension procedure.14 Total cell RNAs from HCV-infected and mock-infected cells, 6 days postinfection, were hybridized to microchips affixed with the sequences of approximately 350 known human miRNAs. As shown in Fig. 1, the levels of miRNA expression in primary hepatocytes infected with the three major strains of HCV were accentuated to various degrees. Consistently, cells infected with the HCV strains showed induced expression of miR-141, miR-200a, and miR-200b and miR-200c to a lesser degree (Fig. 2; Supporting Information Table 1). miR-141 levels increased concomitantly with HCV infection. We chose miR-141 for further studies because it was consistently the most enriched miRNAs up-regulated in HCV-infected cells. The levels of miR-141 that are consistently accentuated in HCV infection is summarized in (Table 1).
Table 1. miRNAs Constantly Up-regulated in Cells Infected with HCV Genotypes 1a and 1b
HCV Genotype 1a
HCV Genotype 1b
HCV Genotype 1a
HCV Genotype 1b
Changes in miRNA Landscape in HCV-Infected Cells.
Previous studies have analyzed alterations in miRNA in liver cancer tissues compared with adjacent normal tissues. For the most part, miRNA expression appears to be accentuated in HCC, with down-regulated miRNAs in HCC being rare.15, 16 Earlier studies of HCV-infected hepatocytes indicated that virus replication depends on host cell miR-12217–19 or the cellular RNA interference machinery in general.20 Because both miR-141 and miR-200a share a target sequence, we chose to study the biological effects of miR-141 up-regulated during HCV infection. We focused next on validating the target of miR-141 that is induced in HCV-infected cells.
Rational miRNA Target Predictions.
miRNAs act upon their mRNA targets by way of imperfect base pair complementarity, and subsequently modulate the translation of the mRNA. The most important determinant for targeting efficacy is the strength of the interaction at the 6- to 7-bp seed sequence at the 5′ end of the miRNA.21–24 We used bioinformatics methods to determine a small number of candidate mRNA targets for the four miRNAs (miR-141, miR-200a, miR-200b, miR-200c) based on the following criteria: (1) strength of the miRNA-target duplex, particularly at the seed position; (2) combinatorial effects of multiple binding sites of the same miRNA or of multiple miRNAs targeting the same gene; and (3) conservation of miRNA target sites in multiple related species, which increases their likelihood to be functional.21, 25 Using Pictar software,26 we searched for genes containing multiple miR-141, miR-200a, miR-200b, and miR-200c target sites that are conserved in humans, chimpanzees, mice, rats, and dogs. Whereas Pictar reported 100-200 target genes for each of the miRNAs individually, the combinatorial search revealed a more specific set of 65 potential genes (Supporting Information Table 2). We narrowed down the set of targets further based on the functional characterization of genes, prioritizing those likely associated with HCV infection and liver disease. DLC-1 (GenBank accession no. NM_006094) was the second highest scoring gene reported by Pictar that harbored multiple binding sites for all four miRNAs. DLC-1 has been shown to be inactivated in HCC7, 27 and to serve as a tumor suppressor gene,28, 29 and thus represents an important choice for further analysis.
Target Validation of miRNAs (miR-141 and miR-200a).
Validation of miRNAs (miR-141 and miR-200a) target recognition was based on luciferase reporter vectors containing the 3′-UTR of DLC-1 mRNA. We observed that the introduction of miR-141 and miR-200a mimics inhibited luciferase, whereas the transfection of miRNA antagomirs (2'-O-methyl–modified antisense oligonucleotides) restored luciferase expression (Supporting Information Fig. 2). The results suggest that the DLC-1 3′-UTR indeed harbors target sequences for miR-141 and miR-200a, and that alterations in the miRNA levels could regulate intracellular DLC-1 expression. miR-141 and miR-200a share identical 5′-seed sequences (Supporting Information Fig. 3); these studies have focused on the biological validation of miR-141–targeted DLC-1 expression and its effect on HCV replication.
To validate whether the increase in intracellular miR-141 during HCV infection targets DLC-1 expression, we introduced luciferase DLC-1 3′-UTR reporter in HCV1a-infected cells, with or without miR-141 antagomirs (Fig. 3). Expression of DLC-1 3′-UTR luciferase was down-regulated in HCV1a-infected cells. The expression of DLC-1 in HCV-infected cells was restored when miR-141 antagomirs were introduced by way of cotransfection (Fig. 3). The results suggest that DLC-1 expression in HCV1a-infected cells is regulated by intracellular miR-141.
Increased miR-141 Inhibits DLC-1 Protein in HCV-Infected Cells.
We next examined whether increased miR-141 in HCV-infected cells reduced DLC-1 protein in host cells. Western blot analysis of HCV-infected hepatocytes (infected either with HCV genotypes 1a, 2a, or the JFH1 strain) showed reduced DLC-1 protein levels (between 50% and 60% within 72 hours postinfection) compared with uninfected cells (Fig. 4).
Next, we validated the effects of miR-141 on DLC-1 expression (in uninfected and HCV-infected hepatocytes) by either depleting miR-141 with antagomirs or artificially increasing the miR-141 levels by transfection with miR-141 mimic oligonucleotides (Fig. 5). Increasing miR-141 inhibited DLC-1 protein in uninfected cells (Fig. 5, lanes 2 and 5); whereas depleting miR-141 with miR-141 antagomirs derepressed DLC-1 expression (Fig. 5, lanes 3 and 6). There was no further inhibition of DLC-1 in HCV-infected cells upon addition of the miR-141 mimic (Fig. 5, lane 5), presumably because the miR-141 target sites within DLC-1 3′-UTR are saturated with the increased levels of miR-141. These findings suggest that miR-141 regulates DLC-1 protein expression inside cells. The inhibition of DLC-1 protein was not accompanied by a parallel decrease in DLC-1 mRNA, suggesting that miR-141 primarily targets translational inhibition of DLC-1. Artificially increasing intracellular miRNA through the transfection of the miR-141 mimics in HCV-infected hepatocytes did, however, reduce DLC-1 mRNA (Fig. 6C), suggesting that a vast abundance of miR-141 could also disrupt the stability of DLC-1 mRNA.
HCV Replication Depends on miR-141 Modulated DLC-1.
Next, we examined whether alterations in DLC-1 protein levels in HCV-infected human primary hepatocytes influence viral RNA level. We assessed the changes in HCV RNA levels within the infected cells, as well as in virions released in culture media of HCV1a-infected hepatocytes. Changes in HCV RNA were quantified by way of nested RT-PCR13 of infected cells' RNA (Fig. 6A), and the effects of miR-141 modulation on virus released in the culture media of HCV1a-infected hepatocytes were analyzed by way of quantitative RT-PCR (Fig. 6B). The results represent genomic equivalents of HCV RNA normalized with World Health Organization standards for HCV. We depleted intracellular miR-141 through transfection with the miR-141 antagomirs or artificially increased miR-141 inside cells through transfection with miR-141 mimics. The depletion of miR-141 resulted in inhibition of HCV RNA replication in infected hepatocytes, whereas artificially increasing miR-141 resulted in increased viral RNA replication (Fig. 6A, lanes 1-3). Thus, HCV RNA replication in infected hepatocytes appears to be inversely related to the intracellular level of miR-141 and perhaps to its targeted DLC-1 (Figs. 5 and 6).
The efficiency of virus released into the culture media of HCV1a-infected primary hepatocytes (Fig. 6B) appears to be quantitatively more severely affected by the depletion of miR-141. Similarly, the increase in virus released into the culture medium in response to artificially increased miR-141 (through transfection with miR-141 mimics) was higher than the increase in HCV RNA within the infected cells (compare the percentage inhibition of HCV RNA replication and the released viral RNA in Figs. 6A and 6B). Although the reasons for the quantitative differences in miR-141 modulated HCV replication and mature virus particles are not entirely clear, the collective results suggest that HCV replication in infected hepatocytes relies on miR-141–mediated depletion of tumor suppressor DLC-1. The reciprocal relation between the cellular DLC-1 protein level and miR-141 in HCV-infected cells suggests that virus replication modulates the abundance of DLC-1 tumor suppressor protein, which subsequently influences the efficiency of viral RNA replication and the release of mature virus particles from infected hepatocytes. However, these findings do not support a direct role of either miR-141 or DLC-1 protein in the regulation of HCV replication.
DLC-1 Level in HCV-Infected Cells Modulates Cell Proliferation.
Functional validation of the role of DLC-1 as a tumor suppressor has been examined based on its effect on cell growth.28 We next asked whether intracellular changes in DLC-1 protein influence the propagation of HCV-infected primary hepatocytes. Cell proliferation was analyzed by way of immunostaining for Ki67 nuclear antigen (Fig. 7). The Ki67-positive cells showed an increase in HCV1a-infected cells compared with uninfected hepatocytes; the increased cell proliferation was countered by increasing DLC-1 protein with the transfection of DLC-1 expression vector (Fig. 7A, upper panel). The relative numbers of Ki67-positive cells (in uninfected cells and HCV1a-infected cells with and without DLC-1 cDNA transfection) is shown in Fig. 7B. In addition, we quantified the enhanced proliferative capacity of HCV1a-infected primary hepatocytes by way of fluorescence-activated cell sorting analysis of Ki67-labeled cells (Fig. 7C). The results showed an approximately 40% increase in cell proliferation of HCV1a-infected hepatocytes (3 days posttransfection). The increased cell proliferation is neutralized when the DLC-1 level was artificially increased through transfection with a DLC-1 expression vector.
Recent studies support the oncogenic role of miRNAs in different human neoplasms including HCC,30 glioblastoma,31 urinary bladder cancer,32 papillary tumors of the thyroid,33 and pancreatic cancer.34 However, the role of miRNA-mediated oncogenesis remains unclear for HCV infection. We present several lines of evidence that HCV infection results in the induction of miR-141, which silences DLC-1 expression, a tumor suppressor gene that is frequently deleted in HCC and other solid human tumors.
The results presented in this study support a link between HCV replication and altered expression of miR-141 that target a tumor suppressor gene frequently deleted in HCC. Tumor suppressor genes can influence oncogenic virus replication by negatively regulating pro-oncogenic signaling proteins.6 NF1 inhibits the Ras signaling pathway, which is deregulated in many cancers and can be a potential therapeutic target. Phosphatase and tensin homologue inhibits the phosphoinositide 3-kinase (PI3K) pathway, and inhibitors of PI3K components such as PI3K, AKT, and mTORs have been similarly pursued for cancer therapy.11 The results presented here support a model of HCV-associated hepatocarcinogenesis, based on miRNA-mediated silencing of tumor suppressor DLC-1. The intracellular induction of miR-141 by HCV appears to translationally inhibit the tumor suppressor DLC-1, whose depletion promotes cell proliferation. Although our findings support the role of DLC-1 in HCV-associated hepatocarcinogenesis, they do not by themselves support a direct role of DLC-1 in regulating HCV replication, nor do they rule out possible contribution of other tumor suppressor genes.
Dependence of HCV replication on miRNAs has been debated in earlier studies.18, 19, 30 In recent studies, Fornari et al.4 have presented evidence that miRNA-221 induced in HCC tissues promotes tumorigenesis by targeting the CDK inhibitors CDKN1C/p57 and CDKN1B/p27. Our findings support the argument that miR-141–targeted suppression of tumor suppressor DLC-1 may promote the initial stages of HCV-associated hepatocarcinogenesis and cell proliferation.
Efficient HCV replication is correlated with miR-141–mediated reduction of DLC-1 protein in virus-infected cells. The reciprocal relationship between miR-141 and DLC-1 protein levels in HCV-infected cells suggests that virus replication is favored in cells with reduced levels of DLC-1 protein, although, the exact mechanism by which miR-141 or DLC-1 modulate virus replication is not clear. We verified the tumor suppressor function of DLC-1 based on the observations that reduced level of DLC-1 in HCV-infected cells increased cell proliferation, whereas artificially increasing DLC-1 protein levels in HCV-infected cells countered the increased cell proliferation.
We thank Nicholas Popescu (National Cancer Institute) for DLC-1 cDNA and helpful discussions, Sita D. Gupta (Uniformed Services University of the Health Sciences) for help with the manuscript, and Wenjie Bao for help with western blot analysis. We also thank Teresa Hawley for assistance with flow cytometry data analysis and Rahul Vanjani and Siva Balasubramanian for help with earlier stages of the study.