MicroRNAs (miRs) are recently discovered molecules that regulate entire intracellular pathways at a posttranscriptional level through RNA-RNA binding. miRs are evolutionarily conserved, approximately 22-nucleotide-long RNAs that are encoded in the genome. The majority of miRs reside in introns of protein-coding genes.1 Similar to messenger RNAs (mRNAs), most miRs are transcribed by RNA polymerase II, which yields primary microRNA transcripts (pri-miRs) of various lengths. pri-miRs are initially processed by the ribonuclease III enzyme Drosha in the nucleus. The cleavage of pri-miRs releases small, approximately 65-nucleotide-long, stem-loop-structured molecules called precursor microRNAs (pre-miRs). After they undergo exportin-mediated translocation into the cytoplasm, pre-miRs are further processed into approximately 22-nucleotide-long RNA duplexes (stems of the pre-miRs) composed of mature miR and miR* strands, which are known as guide and passenger strands, respectively. Mature miRs are loaded into an argonaute-containing RNA interference-induced silencing complex (RISC). This miR/RISC complex is the effector of miR-mediated gene repression activity2, 3 (Fig. 1) and mediates posttranscriptional gene repression by facilitating degradation and/or inhibiting translation of target mRNAs.
The human genome has been predicted to encode approximately 1000 miRs. Although many miRs are ubiquitously expressed, others are expressed in a tissue and cell type-specific manner; this suggest a pivotal role in differentiation and cell fate determination. Functionally, miRs are thought to down-regulate gene expression via two nonexclusive mechanisms: inhibiting translation and facilitating degradation of target mRNAs. In mammalian cells, the seed region of miRs (2-8 nucleotides) is the primary determinant of target recognition. miRs bind to the 3′-untranslated regions of the target transcripts. The majority of miR/mRNA interactions require perfect complementarity between the seed region of miRs and their target mRNAs, whereas the pairing requirement outside the seed region is significantly less stringent.4 miRs sharing an identical sequence in the seed region are grouped into families, and miRs from the same family bind to and regulate the expression of essentially the same set of target mRNAs.5 Because the primary target recognition determinant is 7 to 8 nucleotides long in the seed region, a single miR can potentially regulate hundreds of target mRNAs. This notion was further validated by a transcriptome analysis of tissues isolated from mice with targeted deletion of miRs. Among the multiple downstream mRNA targets, the degree of regulation of each individual transcript is usually small (<2-fold).6, 7 It is believed that the additive/synergistic effect of regulating multiple mRNAs in the same pathway translates into significant biological outcomes and phenotypes.
Similar to mRNA expression, miR expression has been found to be dysregulated in disease tissues in comparison with normal tissues. These dysregulated miRs represent a novel pool of therapeutic targets and biomarkers, including those in tissue fibrosis.8, 9 For example, the miR-29 family of miRs is down-regulated in a mouse model of cardiac fibrosis following myocardial infarction. miR-29 is encoded in two separate genomic loci yielding four mature miRs (29a, 29b1, 29b2, and 29c). Multiple extracellular matrix (ECM) genes, including many isoforms of the collagen superfamily, are among the top-ranked predicted targets of miR-29. Experimentally, an miR-29 mimic led to repression whereas anti-miR-29 resulted in de-repression of many ECM genes in cultured cardiac fibroblasts. Despite the modest regulation of the individual gene, it has been postulated that the coordinated repression of multiple genes in the ECM pathway results in a strong biological outcome.10
In the current issue of Hepatology, Roderburg and colleagues11 set out to identify changes in miR expression in two different mouse models of liver fibrosis: carbon tetrachloride and bile duct ligation. As expected, many miRs were dysregulated in response to these fibrosis-inducing injuries. Among them, all three members of the miR-29 family were significantly down-regulated in response to both of these models. The authors further extended these observations by demonstrating miR-29 down-regulation in human cirrhotic liver samples.
Because miRs are expressed in a cell type-specific manner, the authors also determined the relative expression levels of miR-29 in the four major liver cell types (hepatocytes, stellate cells, Kupffer cells, and endothelial cells). An intriguing finding from this experiment showed that miR-29b expression was more than 100 times higher in hepatic stellate cells versus the other three cell types. Upon exposure to fibrotic stress, miR-29b was down-regulated in stellate cells and hepatocytes, whereas it was up-regulated in Kupffer cells and endothelial cells. Activated hepatic stellate cells are key mediators of fibrosis because they are reported to be the major cell type producing collagen and other ECM proteins in the injured liver. Along with the notion that miR-29 directly suppresses the expression of various ECM mRNAs, it has been postulated that miR-29 down-regulation directly leads to the overexpression of ECM gene products during fibrosis. These results also suggest that although miR profiling of RNA isolated from whole organ lysate is useful, the additional effort to obtain samples from purified cell types will provide clearer insights into the molecular pathophysiology. These observations are also consistent with previous studies demonstrating that the increased fibrillar collagen expression in liver fibrosis is primarily posttranscriptional.12 Mechanistically, the authors showed that the treatment of hepatic stellate cells with transforming growth factor β (TGF-β) suppressed miR-29 expression; this provided evidence that the fibrogenic effect of TGF-β is mediated in part through the down-regulation of miR-29. TGF-β3 has also been reported to be a direct target of miR-29.13 Such cross-regulatory relationships between miRs and their cognate mRNA targets are commonly observed. In this case, miR-29 acts as a feed forward switch: the fibrogenic signal initiates TGF-β-induced miR-29 down-regulation. Reduced miR-29 activity further de-represses TGF-β expression and results in amplification of the fibrogenic signal.
The miR-29 down-regulation observed during fibrosis directed the authors to investigate its utility as a circulating biomarker for liver fibrosis. Emerging evidence suggests that miRs are found in lipid-enclosed particles in the serum called exosomes.14, 15 During tissue injury, the release of tissue-specific miRs (e.g., miR-122 and miR-208 during hepatic and cardiac injuries, respectively) into the circulation has been reported.16, 17 Despite the ubiquitous expression of miR-29 and its modest expression level in the liver among a list of organs tested, this study demonstrated that the amount of circulating miR-29 was significantly inversely correlated with the advancement of fibrotic stages. Altogether, the utility of miR-29 as a circulating biomarker of fibrosis in the liver and other organs warrants further investigation.
The repression of ECM genes by miR-29 and its down-regulation during fibrosis strongly suggest that miR-29 down-regulation contributes to the pathogenesis of fibrosis, and the reintroduction of miR-29 could be a novel therapeutic strategy for fibrosis. To test such a hypothesis, one must determine the therapeutic effect of an miR-29 mimic in vivo. Currently, an miR mimic is composed of an RNA duplex similar to that of the small interfering RNA employed in RNA interference. The double strandedness of the RNA duplex configuration is believed to be essential for the efficient loading of the miR guide strand into the RISC complex. In order to increase stability and improve cellular uptake of the miR duplex, it is formulated with lipid nanoparticles (LNPs).18 To date, the LNP-mediated delivery of an RNA duplex limits its tissue distribution primarily to the liver (including hepatic stellate cells). The characteristics of an LNP-enclosed miR-29 mimic render liver fibrosis an attractive disease indication for the initial clinical proof of experiments. It is believed that similar underlying mechanisms are involved in the development of fibrosis in different organs. Further advances in oligonucleotide delivery technology will enable the evaluation of whether an miR-29 mimic could be an effective therapy for fibrotic conditions of other organs.