MicroRNA miR-204 and miR-1236 inhibit hepatitis B virus replication via two different mechanisms

Hepatitis B virus (HBV) is a major human pathogen. In this study, we found that miR-204 and miR-1236 were down-regulated in HBV-producing cells, and each could suppress HBV replication. Using a bioinformatic approach and a reporter assay, we identified miR-1236, which can reduce HBV replication and protein production by directly targeting at HBV specific mRNA. In contrast, miR-204 was identified by a microarray approach, and had no effect on HBV RNA and protein production. Surprisingly, miR-204 could inhibit HBV pregenomic RNA encapsidation and capsid assembly. We further demonstrated that HBV suppressed miR-204 expression via activating a host transcription factor STAT3. We established a positive feed-forward loop between HBV, miR-204 and STAT3. Interestingly, miR-204 has been considered as a tumor suppressor in some literature. Since the risk for hepatocellular carcinoma (HCC) is significantly increased in chronic HBV patients, it is possible that chronic suppression of miR-204 by HBV contributes to HCC incidence. Both miR-204 and miR-1236 might be useful for developing new therapeutics against HBV.

and observed significant reduction of at least a dozen microRNAs in HBV-producing cells 8 . By stem-loop qPCR analysis, we validated that the expression of miR-204 was indeed decreased in HBV-producing rat and human hepatocytes (Fig. 1A,B). Furthermore, miR-204 was also significantly reduced in an HBV transgenic mouse model (genotype D, serotype ayw) (Fig. 1C) 10 . In a separate approach via bioinformatics analysis, we identified miR-1236 as a potential anti-HBV miRNA by using the Microinspector target-prediction algorithm. This microRNA of human origin has no counterpart in rodents. It can bind to the 3′ UTR on HBV genome with the highest free energy (−30.8 Kcal/mol) (Tables 1 and 2). Interestingly, we found that hsa-miR-1236 was also reduced in expression in human HBV-producing cell lines (Fig. 1D).
We hypothesized that some of these miRNAs with reduced expression in HBV-replicating hepatocytes, could have a negative effect on HBV replication. This hypothesis could explain their reduced expression levels in several independently established rat and human HBV-producing cell lines (Fig. 1A,B). To evaluate the effect of microRNA on viral replication, we cloned the precursors of miR-204 and miR-1236 in miRNA expression vectors, respectively. HepG2 cells were cotransfected with microRNA expression vectors and an HBV ayw genomic dimer plasmid. Five days post-transfection, the intracellular HBV core particle-associated DNA was isolated and analyzed by Southern blot [11][12][13][14] . As shown in Fig. 2A, miR-204 and miR-1236 could each inhibit HBV DNA replication, without altering the level of HBV RNA expression in the cytoplasm (Fig. 2B). The expressions of miR-204 and miR-1236 were detected by stem-loop PCR analysis (Fig. S1). Next, we examined the effect of these two microRNAs on HBV protein expression, including core protein (HBc) (Fig. 2C), surface antigen (HBsAg) and e antigen (HBeAg) (data not shown). Only miR-1236 and miR-130a 5 (data not shown), but not miR-204 and vector control, could reduce HBV protein expression. To further elucidate the mechanism, we cotransfected an HBV genomic replicon with LNA-miR-204-5p or LNA-miR-1236 into HepG2 cells (Fig. 2D). Knockdown of endogenous miR-204-5p or miR-1236 significantly resulted in increased HBV DNA replication (Fig. 2E) and protein synthesis (Fig. 2F). This LNA experiment confirmed the anti-HBV potential from miR-204-5p and miR-1236.
Using Computer-aided programs, we predicted potential target sites of miR-1236 and miR-204 clustering between nt 1521 and nt 2122 on the HBV ayw genome (Fig. 3A). Results from the 3′ UTR luciferase reporter assay suggested that miR-1236, but not miR-204, significantly reduced the luciferase activity (Fig. 3A). We also performed compensatory mutagenesis by introducing mutations into the seed sequences of miR-1236 and its predicted target site on HBV (Fig. 3B). Only the combination of seed mutant miR-1236 and the target site mutant of HBV significantly restored the inhibitory effect of miR-1236 on the reporter activity (Fig. 3B). This result suggests a direct interaction between miR-1236 and HBV specific RNA. We also investigated whether miR-1236 can target HBV RNA of a different genotype, such as adr. As shown in Fig. 3C, the target sites of miR-1236 appear to be evolutionarily conserved. When the genomic dimer of HBV adr was cotransfected with a miR-1236 expression vector, both viral DNA synthesis (Fig. 3D) and secreted HBsAg (Fig. 3E) were significantly inhibited. Therefore, miR-1236 has a negative effect against HBV subtypes other than ayw.
In Fig. 1A-C, we observed that the expression of miR-204 was lower in HBV producing cell lines and transgenic mice. It is known that the expression of miR-204 can be suppressed by STAT3 activation 15 . In addition, it has been shown that HBx can activate the STAT3 signaling pathway 16,17 . We therefore investigated whether STAT3  Table 2. Computational prediction of cellular microRNAs with a strong potential of binding to whole HBV RNA pregenome # . @ The positions refer to the first nucleotide of microRNA binding sites on HBV ayw genome. # The entire miRbase version 21.0 were screened with HBV (ayw) whole genome sequences using the Microinspector Program. can regulate miR-204 in HBV-producing hepatocytes. As shown in Fig. 4A, the expression of total and phosphorylated STAT3 proteins was strongly increased in HBV-producing cells. When we treated these HBV-producing cells with a STAT3 inhibitor, S31-201, we observed decreased levels of the total and phosphorylated STAT3 proteins in a dose dependent manner (Fig. 4B). It is therefore interesting to ask whether one can restore miR-204  feed-forward loop summarizes the relationships among HBV, STAT3, and miR-204 (Fig. 4G). In this triad diagram, HBV can stimulate STAT3, which in turn can suppress the level of miR-204, leading to enhanced viral gene expression replication. It was puzzling that miR-204 could reduce HBV DNA replication without any apparent reductions in HBV specific RNA and proteins (Fig. 2). We reasoned that a potential target for miR-204 could be at the step of RNA encapsidation and capsid assembly (Fig. 5A). To address this issue, we cotransfected HepG2 cells with HBV plasmid and miR-204 expression vector and examined the assembly efficiency and stability of intracellular core particles of HBV by native agarose gel (Fig. 5B upper panel). As a control, we also monitored the level of core protein monomer by denaturing SDS-PAGE (Fig. 5B, lower panel). While miR-204 significantly reduced the level of intracellular core particles, it had no effect on the total amount of core protein by SDS-PAGE. This result supports the notion that miR-204 could interfere with HBV capsid assembly. We also compared the efficiencies of pgRNA encapsidation with and without miR-204 cotransfection, by ribonuclease protection assay (RPA) using core particle-associated HBV RNA (Fig. 5C). Cotransfection with miR-204 reduced core particle-associated pgRNA, yet without any apparent effect on total cytoplasmic pgRNA. Our results suggest that miR-204 might target an unknown host factor(s) involved in capsid assembly or RNA encapsidation.

Discussion
In this study, we reported two different mechanisms of two cellular miRNAs for their control over HBV replication and gene expression. The microRNA, miR-1236, was found to target a sequence localized within the 3′UTR of HBV specific RNAs, leading to reduced viral gene expression. The other microRNA, miR-204, inhibited HBV pregenomic RNA encapsidation and capsid assembly.
MiR-204 is generated from the sixth intron of the transient receptor potential melastatin 3 (TRPM3) gene 18 . We found miR-204 is expressed in multiple tissues in rats, with the highest level of expression in the eye and cerebellum by stem-loop real time PCR (Fig. S2). The expression level in the rat liver is low. It is possible that miR-204 might target an unknown host factor (s) involved in capsid assembly or RNA encapsidation. At present, approximately 400 validated targets of miR-204 have been documented in literature 19 . There is no easy way to speculate on which known or unknown miR-204 target might be involved in HBV RNA encapsidation and capsid assembly. Interestingly, this microRNA is well known for its important role in the eye retina development, adipogenesis, diabetes, and cancer 20 . In fact, miR-204 is being considered as a tumor suppressor. Since chronic infection with HBV can increase the risk of HCC development, it is possible that chronic suppression of miR-204 by HBV might contribute to the increased HCC incidence in chronic hepatitis B.
Previously, it has been demonstrated that treatment by interleukin-6 (IL-6) and epidermal growth factor stimulated the interaction between HBV enhancer 1 and STAT3 protein, which leads to the activation of HBV gene expression 21 . Therefore, we proposed here that HBV infection can activate the phosphorylation of STAT3, which in turn repressed the expression of miR-204, and thus stimulated HBV gene expression and replication. In this scenario, how STAT3 can suppress the level of miR-204 remains unclear. It is not mutually exclusive that the activated STAT3 can upregulate HBV transcription through its interaction with HBV enhancer 1, and at the same time, repress miR-204 expression. To further study on the regulation of miR-204 will provide more information about the control of HBV replication.
MiR-1236 is an intronic miRNA with no other family member (miRBase version 21). Interestingly, miR-1236 is found only in humans, pantroglodytes and pongopygmaeus, but not in rodents (GeneCards and miRBase version 21). We have no results on the tissue distribution of human miR-1236 at present. However, expression of miR-1236 can be found in various tissues including liver 22 . Previous studies showed that miR-1236 can repress VEGFR-3, RORγ, p21 promoter, and ZEB1 to inhibit inflammatory lymphangiogenesis, ulcerative colitis, cell proliferation, or cell migration/invasion, respectively [23][24][25] . In addition, miRNA-1236 contributes to HIV-1 restriction in monocytes 26 , and has a regulatory role in alpha-fetoprotein (AFP) expression and HCC development in the liver 22 . Recently, miR-1236 has also been shown to be induced by IL-1β and repressed by Twist1 to regulate the inflammatory lymphangiogenesis 23 and hypoxia-induced EMT 27 . In our study, we discovered a novel role of miR-1236 in inhibiting HBV replication by direct targeting at HBV specific RNA. The successful bioinformatics approach to identifying miR-1236 allows us to anticipate that miR-3960, miR-3166, miR-4763, miR-665 and miR-663, are potential candidates for anti-HBV miRNA (Tables 1 and 2).
To date, at least three miRNA candidates with therapeutic potential have been selected into clinical trials 28,29 . One well known example is Miravirsen, a locked nucleic acid-modified (LNA) DNA phosphorothioate antisense oligonucleotide. It can sequester mature miR-122, which is essential to HCV replication. The treatment of Miravirsen in chronic HCV infected patients showed prolonged and dose-dependent repression of HCV RNA level without any apparent adverse effects or viral resistance 29 . The success example of Miravirsen on HCV would certainly encourage a similar miRNA approach for HBV or other viral diseases in the future.
In summary, instead of adopting an evasion strategy like target site mutations, HBV could take an alternative strategy by reducing the expression of hostile cellular microRNAs. Since the effects of miR-1236 and miR-204 on HBV appeared to be independent of HBV genotypes, their therapeutic potential may be broad spectrum, and worth further investigation. Animals. The generation of HBV transgenic mice on an ICR background has been reported previously 10 . The transgene is a 1.3-fold HBV genome (genotype D, serotype ayw). The Tg[HBV1.3] mouse line was used in this study. All animals were housed in a specific-pathogen-free environment in the animal facility of the Institute of Biomedical Sciences, Academia Sinica, Taiwan.

Material and Methods
Cell Culture. Human hepatoma HuH-7 30 , HepG2 31 and rat hepatoma cell line Q7 cells 8 were maintained as described previously [11][12][13][14]32 . Stable Qs2, Qs4, Qs5, and Qs21 cell lines can produce infectious HBV and were established by transfecting Q7 (Morris hepatoma 7777) cells with a genomic dimer of HBV DNA (ayw) 8,9 . Stable UP7-4 and UP7-7cell lines were established by transfecting human HepG2 cells with a genomic dimer of HBV DNA (ayw) 33 . While the phenotypic effects of microRNAs on viral replication are in general stronger in HepG2 Scientific RepoRts | 6:34740 | DOI: 10.1038/srep34740 than HuH-7 cells, the latter are less sticky, easier to trypsinize and grow in culture. Therefore, we used both cell lines with no preference.
Quantitative Real-time PCR. Briefly, 2 μg of total RNA was reverse transcribed into cDNA at 37 °C for 2 hours using random primers and High Capacity cDNA Reverse Transcription kit (Applied Biosystem, Grand Island, NY). The cDNA product was analyzed by 1.5% agarose gel electrophoresis and visualized by staining with HealthView (Genomics BioSci & Tech, Taiwan). The cDNA product was then diluted 100 times for real-time PCR analysis using Power SYBR Green PCR master mix (Applied Biosystem, Grand Island, NY), and the default condition in a 20 μl reaction volume by Applied Biosystems 7500 Real-Time PCR System. Data were analyzed by relative quantification methods (∆∆Ct methods) using 7500 software V2.0.1.
Treatment with STAT3 inhibitor. STAT inhibitor (S31-201) was purchased from Selleckchem. HepG2 cells were seeded in 6-well tissue culture plates at 5 × 10 5 cells/well and cultured at 37 °C for 24 hrs. S31-201 was prepared in 0.1% DMSO and added to culture medium at indicated concentrations. Culture medium was changed every two days before harvest. Southern and Northern blots. HBV core particle-associated DNA and total cellular cytoplasmic mRNAs were analyzed by Southern and Northern blots as described previously 34 . Each lane on the Southern blot gel was loaded with the total amount of core particle-associated viral DNA extracted from each transfected dish, which was seeded with equal cell density one night before transfection. Native agarose gel electrophoresis and Western blot of HBV core particles. HBV core particles and total HBV core protein were analyzed by native agarose gel electrophoresis and Western blot as described previously 34 .
A 392 nt antisense riboprobe was radiolabelled by in vitro transcription using a NotI linearized DNA fragment from a pGEM-T vector containing HBV sequences nt 2150-1820. RNase would preferentially digest single-strand RNA, but has no significant effect on the hybridized double-strand RNA. The protected pgRNA fragment was 330 nt in length. HBV polymerase mutant Y63D is defective in DNA synthesis, yet competent in pgRNA encapsidation 35,36 . This mutant accumulates a higher level of encapsidated pgRNA in the core particles due to its arrested priming of cDNA synthesis.
Scientific RepoRts | 6:34740 | DOI: 10.1038/srep34740 Statistics. Statistical significance was determined using the Student's t test. In all figures, values were expressed as mean ± standard deviation (SD) and statistical significance (p < 0.05) was indicated by an asterisk. The data represent results from at least three independent experiments.