The hepatitis C virus (HCV) belongs to the Flaviviridae
family and encodes a small enveloped, positive-stranded RNA genome of 9.6 kb [1
]. HCV infects approximately 170 million people worldwide and infection often leads to serious chronic liver diseases including liver cirrhosis, liver failure and hepatocellular carcinoma [2
]. This is attributed in part to the remarkable ability of the virus to establish persistent infections [4
]. Current medical treatment for HCV involves the antiviral agent type 1 IFN combined with ribavirin, but only ~55% of treated patients have a sustained virologic response, i.e., the absence of detectable HCV RNA 24 weeks after cessation of therapy [8
]. Clearly, there is a desperate need for new anti-HCV therapies.
miRNAs have been implicated as potential new targets for HCV therapy. miRNAs are a class of small non-coding RNA molecules of 20–22 nucleotides that regulate gene expression through translational repression and mRNA degradation [10
]. miRNAs play key roles in regulating gene expression in a variety of organisms and are involved in crucial physiologic and pathologic processes [11
]. In vitro studies have shown that the liver-specific miR-122 is required for HCV RNA replication [14
], whereas IFN-induced miRNAs miR-196 and miR-448 directly target HCV genomic RNA for inhibition of viral replication [15
]. However, a recent report showed that there was no correlation between miR-122 expression and viral load in human subjects with chronic HCV undergoing IFN therapy, cautioning the extrapolation of in vitro
observations to human subjects with HCV infection [16
]. As a new player among gene regulation mechanisms, the functions of miRNAs have not been fully clarified. The comprehensive delineation of the relationships between HCV infection and cellular miRNAs is crucial to better understand HCV pathogenesis and to develop novel therapeutic strategies. To the best of our knowledge, there has not been a systematic study using high-throughput technology to analyze the role of cellular miRNAs during HCV infection.
Precise identification of miRNA targets is essential for the functional characterization of individual miRNAs and a better understanding of complex human diseases. Multiple computational approaches have been developed to predict miRNA-target relationships using sequence information [17
]. However, accurate prediction of physiologically active miRNA targets remains challenging. The observation that many miRNAs cause mRNA degradation of their targets provides opportunities to develop new approaches for target identification and validation using high-throughput expression profiling. Gene expression profiling data has been used to identify functional targets [18
], and to improve target predictions [20
]. When the expression of miRNAs and mRNAs are simultaneously profiled across different conditions, the miRNA and the mRNAs that it targets for degradation should exhibit an inverse expression relationship. A successful strategy for miRNA target identification has been reported using the inverse relationships between miRNAs and mRNAs inferred from the paired expression profiles across different conditions [22
]. The utility of inverse correlation for miRNA target identification has been further demonstrated in more recent studies, such as surveys of miRNA and mRNA expression in human cell lines [23
] and miRNA expression and protein abundance in rat kidneys [24
]. This approach, unlike miRNA transfection, avoids artificial conditions needed to perturb gene expression in the systems of interest.
Identification of functional modules has greatly advanced our understanding of complex biological networks [25
]. A single miRNA can regulate a large number of target genes in mammalian cells [27
], and multiple miRNAs may regulate the same target [28
]. To understand the many-to-many regulatory relationships in complex cellular systems, attempts have been made to predict miRNA regulatory modules. Yoon and De Micheli introduced the concept of such modules, or groups of miRNAs and target genes that are believed to participate cooperatively in post-transcriptional gene regulation [29
]. Their modules are related only to miRNA-mRNA binding information at the sequence level. To improve module prediction, two different computational approaches have been proposed to integrate mRNA and miRNA expression profiles [30
], using the same published expression dataset [32
]. Both approaches introduced the measurement of coherent expression among miRNAs or mRNAs, but not between miRNAs and mRNAs. Recently, a graphical model approach was used to predict miRNA regulatory modules in Arabidopsis
]. However, none of the above explicitly considers the inverse expression relationships for module prediction, which has been very effective in terms of uncovering functional miRNA-target relationships.
In this report, we present an integrative strategy for inferring HCV-associated miRNA-mRNA regulatory modules, by combining the inverse expression relationships between miRNAs and mRNAs and computational target predictions at the sequence level. We generated, for the first time, a systematic profiling of cellular miRNA expression during HCV infection in human livers. We inferred inverse expression relationships by simultaneous microarray profiling of host miRNA and mRNA expression across 30 human liver biopsies, including samples from HCV-infected and uninfected individuals. Using our integrative computational approach, we identified 38 miRNA-mRNA regulatory modules that were associated with HCV infection. We analyzed the functional roles of those modules at the systems level through the integration of a large protein interaction network. We show that the expression of multiple cellular miRNAs was altered and provide evidence that miRNAs are involved in a combinatorial and modular fashion in the regulation of host responses during HCV infection. Together, these results provide novel insights into regulatory mechanisms at the miRNA level during HCV infection, and our analytical approach shows the utility of an integrative strategy that may be applied to the study other complex human diseases for the identification of miRNA regulatory modules.