An integrated approach identifies IFN-regulated microRNAs and targeted mRNAs modulated by different HCV replicon clones

Infection with hepatitis C virus (HCV) represents the major cause of liver disease, affecting more than 170 million individuals worldwide. After a sub-clinical phase, greater than 80% of patients progress to persistent HCV infection, the leading cause of chronic liver disease associated with cirrhosis and hepatocellular carcinoma [1, 2].

In the last years, microarray technology provided a comprehensive analysis of alterations in gene expression induced by HCV and revealed important processes of virus-host interactions [3–7]. Interestingly, microarray studies indicated that HCV stimulates the endogenous Type I Interferon (IFN-α/β) pathway as suggested by activation of IFN-stimulated genes (ISGs) [8–15]. Recently, it has been proposed that also microRNAs (miRs), a class of small non-coding regulatory RNAs, are involved in the antiviral pathway induced by IFN-β treatment. The synthetic introduction of five IFN-β-induced miRs into HCV replicon cells may simulate the antiviral effect of IFN-β blocking HCV replication and infection. These five miRs (miR-196, miR-296, miR-351, miR-431 and miR-448) likely induced an antiviral state either through alteration of gene expression and/or directly targeting HCV RNA, as was demonstrated for two of them (miR-196 and miR-448) [16].

Although HCV activates the endogenous IFN-α/β pathway it conversely shows an impressive ability to induce persistent infections. Indeed, it is also clear that HCV has evolved several mechanisms to control the IFN antiviral response, inhibiting the pathway at different levels [17]. Recently, it has been suggested that an improper pre-activation of ISGs in the liver of HCV infected patients may hinder the antiviral response. The discovery of a genetic polymorphism in the interleukin 28B (IL28B) region on chromosome 19 of HCV patients depicted a more complex virus-host interaction. The IL28B non-CC variant has been associated with non-response to the IFN therapy and with lower rates of spontaneous clearance of HCV infection. The poor-response variant is also associated with higher intrahepatic expression level of ISGs [18–20].

A missing aspect in this scenario is the study of the effect produced by HCV on the expression of IFN-β-induced miRs. This is a relevant issue to understand how the virus can suppress the innate antiviral signaling and induce a persistent infection.

In a previous paper, we identified a common transcriptional response of Huh-7 cells to different clones of full-length HCV replicon [21]. Although a more advanced HCV cell culture models that release HCV viral particles has been developed [22–24], the replicon system has the advantage of taking into account the cellular gene expression variability of different HCV replicon cell clones. This approach allows searching for modulated genes shared by all clones, which are likely to be strictly needed for viral replication in different cellular contexts.

On this basis, we used the replicon system to identify IFN-regulated miRs that are modulated by HCV RNA replication. In particular, we analyzed the expression profile of 24 selected miRs in IFN-β-treated Huh-7 cell line and in three cell clones carrying a full-length HCV replicon (clones 21-5, 22-6 and 21-7). Among the identified 16 miRs modulated in the 21-5 clone, 3 miRs showed concordant expression when analyzed in the two other HCV replicons. By a combined approach, based on bioinformatic prediction and microarray analysis, we also identified 37 genes, targeted by the 3 miRs, which are involved in pathways and biological processes potentially implicated in the control of antiviral response by HCV infection.

In the present study we analyzed the effect of HCV replication on the expression of selected miRs involved in the IFN-pathway. In particular, we identified 3 miRs that are equally modulated by HCV in three HCV replicon clones and by IFN treatment. Moreover, we also identified 37 out of 83 predicted target genes, differentially expressed in HCV replicon cells, which are most likely functional targets of these 3 miRs: in fact they showed an inverse expression relationship with the level of the 3 miRs, as described for true targets. These genes could be implicated in regulation of the host response to HCV.

About one half predicted targets did not show the expected inverse expression relationship with miR level, but this result is not surprising. First, computational prediction by miRgator was based on predictions from the widely used programs miRanda, TargetScanS and PicTar (see Materials and Methods). To date, these programs, as well as any other available prediction program, still have a very high false-positive rate, estimated to be up to 40% [34, 35]: i.e. up to 40% genes predicted to be targeted by a miR are not, actually, true targets, and they will not show any inverse relationship. Second, each mRNA is usually targeted by multiple miRs and, in addition, each mRNA has its own set of modulating miRs. Whether or not a change in the level of a single miR can prevail on the effect of the other miRs simultaneously targeting the same mRNA in physiological conditions is, at present, poorly understood. However, it is possible in some cases it will produce no effect and, thus, even a true target will not show the expected inverse relationship.

The recent discovery of the roles of miRs in many human diseases suggests that studies exploring the relationship between HCV and miRs/mRNA may provide new insights into host cell response to HCV infection. Importantly, this approach also gives the opportunity to identify viral mechanisms that control the antiviral defense. Recently, it has been demonstrated that five IFN-β-modulated miRs (miR-196, miR-296, miR-351, miR-431 and miR-448) showed significant effect on HCV replication and at least two of them (miR-196 and miR-448) are directly targeting the HCV genomic RNA [16]. Moreover, it seems that level of miR-122 is inversely correlated with the antiviral defense [16, 36, 37].

Our expression analysis revealed that miR-196a was down-regulated in all three HCV replicon clones. Thus, at least for this component of the pathway, it seems that modulation of IFN-miRs may be altered by HCV in replicon cells. Interestingly, this miR targets the HCV RNA [16], thus, down-regulation of miR-196a may indirectly influence viral replication also by up-regulation of specific target genes. Accordingly, 11 genes, controlled by miR-196a, showed by microarray analysis an inverse expression relationship suggesting that they can be likely considered functional targets of miR-196a. Moreover, gene ontology analysis of the 11 genes highlights that some of them are really involved in pathways such as, extracellular matrix constitution (LAM4), oxidative stress (HADH2, LTB4DH) and cytoskeletal network (CENPE), which are relevant for HCV RNA replication [7, 38, 39].

As for the IFN-β-regulated miR-296, miR-351, miR-431, miR-448 and miR-122 (16) our data indicate that their expression in different HCV replicon clones is either not concordant (miR-296, and miR-122a) or not detected (miR-351, miR-431 and miR-448). In particular, miR-122a was down-regulated in 21-5 and 21-7 clones but its level was not modified in clone 22-6 (data not shown) while miR-296 was down-regulated in 21-5 clone and up-regulated in clone 22-6 and 21-7, respectively (data not shown). Moreover, miR 351, miR-431 and miR-448 were not detected in all clones examined supporting, at least for miR-448, what found in human biopsies where miR-448 is totally absent [37].

Interestingly, we observed that two miRs, miR-142-3p and miR-128a, were up-regulated and down-regulated respectively by IFN-β treatment of Huh-7 cells and in HCV replicon-expressing cells. Although modulation of these 2 miRs by IFN-β has never been described before, it is consistently observed in all clones suggesting that it may be part of the endogenous IFN response to HCV. The IFN pathway is a highly regulated process and several controls has been evolved to activate and turn off this pathway. In this scenario, the temporal modulation of specific miRs seems to represent one of the control elements [40]. It is important to note that this process cannot properly occur in cells sustaining HCV replication. In this case a chronic up or down-regulation of IFN-miRs, likely induced by the virus, may negatively affect the control of the pathway finally improving the efficacy of the antiviral effectors. It would be interesting to investigate whether the experimental use of miR inhibitors (for miR-142-3p) or miR mimics (for miR-128a) could influence the control of the endogenous IFN system.

Among the 37 predicted target genes showing an inverse expression relationship with the 3 miRs, four genes were identified as Interferon Regulated Genes (IRGs) according to the INTERFEROME database [32]. One of these genes (HNMT) is a predicted target gene of miR-196a while the other three (XPO1, PMPCB and HMGB1) are all targeted by miR-142-3p. Importantly, in autoimmune diseases the high mobility group box 1 (HMGB1) protein was identified as a component of immune complex-containing DNA or RNA, which may act as endogenous IFN-β inducer [41]. Down-regulation of HMGB1 gene (fold change 0, 58; p-value 0, 01) in all HCV replicon clones suggests that it might contribute to impair the activation of the IFN signaling. Currently, the role of these four IRGs in the IFN response to HCV replication is unknown. Thus, unraveling their contribution to the regulation of the IFN response may reveal new mechanisms of viral persistence.

Gene Ontology annotations of the 37 common genes also revealed the presence of two genes, UBE2E3 and ATAD2 targets of miR-128a and miR-142-3p respectively, which are involved in the Ubiquitin proteasome pathway. The contribution of this pathway to HCV subversion of the IFN response has never been investigated. This is a quite interesting issue as several viruses use the ubiquitin proteasome system to destabilize proteins, such as IRF3 and STAT proteins, that are important for transcription of Interferon and Interferon-stimulated genes [42].

In attempt to validate our data, we found that 4 out of 37 genes, targeted by the 3 miRs, were also modulated, in a concerted fashion, in HCV genotype 2a chimeric virus J6/JFH microarray datasets [31] supporting the biological relevance of our results. In addition, 6 genes (4 up-regulated: CEACAM1, DDIT3, SCARB2, ZFAND2A; and 2 down-regulated: C9orf19 and RPL5) selected from the HCV clones microarray dataset were found to be modulated in a same way in liver biopsies of patients showing non-CC IL28B polymorphism [20]. This polymorphism is not a good predictor of response to IFN therapy and it is also associated with higher level of ISG expression in the liver and propension to chronicity. So, it can be speculated that modulation of IFN signature as mediated by common miRs in replicon cells, can mirror de-regulation of the IFN signaling proposed for such patients.

In the present study we used the HCV replicon system to identify IFN-regulated miRs that are modulated by HCV RNA replication. By a combined approach, based on Real-Time PCR, bioinformatic prediction and microarray analysis, we identified 3 IFN-β-regulated miRs and 37 genes, which are likely their functional targets, commonly modulated by HCV in three replicon clones. Gene ontology classified the 37 genes into functional categories potentially implicated in the control of antiviral response by HCV infection. The future design of siRNAs directed against some of these genes and the use of miRs and antimiRs may provide an experimental background for the development of therapeutic strategies aimed at the recovering of protective innate responses in HCV infections.