The class of small RNA species [20–24 nucleotides (nt)] known as microRNAs (miRNAs) was discovered relatively recently. These molecules degrade or disrupt the translation of target mRNAs by binding to complementary sites in the mRNA 3′ untranslated region (UTR) [1–5]. The functional importance of miRNAs is suggested by evolutionary conservation across widely diverse phyla [6–9]. Further, miRNAs are associated with diverse biological phenomena, such as cell growth, apoptosis, development, differentiation, and tumorigenesis. For example, Lu et al. [10] have reported that miRNA expression profiles reflect the developmental lineage and differentiation state of tumors. Poorly differentiated tumors were more accurately classified by their miRNA profiles than by their mRNA profiles, suggesting that miRNA profiling may be useful in cancer diagnostics. Further, some miRNAs are associated with oncogenesis. For example, the miR-17-92 miRNA cluster is often overexpressed in B-cell lymphomas [11] and lung cancer [12]. This cluster has been identified as a potential human oncogene and named oncomiR-1 [11]. In addition, miRNAs are involved in other diseases, including nervous system disorders [13, 14], diabetes [15–17], infectious diseases [18–20], cardiovascular diseases [21–23], asthma [24], autoimmune diseases [25–27], and arthritis [25, 26, 28, 29].
Here, we review miRNA biology and the methodological background of miRNA analyses.
The multistep process of miRNA biogenesis begins with transcription, which is carried out primarily by RNA polymerase II to produce long transcripts [primary-miRNA (pri-miRNA)] from miRNA-encoding genes [30]. There are at least two types of miRNA promoters: some miRNAs have their own promoters, but others are located within an mRNA or intron and are transcribed as part of the host gene. Pri-miRNAs are processed in the nucleus by the RNase III enzyme Drosha, resulting in precursor molecules called precursor-miRNAs (pre-miRNAs) that are approximately 70 nt long with stem-loop structures containing 2-nt 3′ overhangs [31]. Pre-miRNAs translocate through nuclear pores to the cytoplasm via Exportin-5 [32–36] and are further processed to mature miRNAs by the RNase III enzyme Dicer [37, 38]. Biogenesis pathways other than this canonical miRNA pathway have also been reported. Some pre-miRNAs are generated by the RNA splicing machinery and lariat-debranching enzyme (the mirtron pathway), which yield pre-miRNA-like hairpins [39]. The mirtron pathway was first discovered in Drosophila [39] and is also found in mammals [40]. The Dicer-independent miRNA biogenesis pathway is another noncanonical pathway, in which pre-miRNA is incorporated into Argonaute (Ago) complexes and cleaved by the slicer activity of Ago to generate an intermediate 3′ end, which is then further trimmed [41].
RNA-induced silencing complexes (RISCs) incorporate miRNAs, which then bind to the 3′ UTR of target mRNAs. Ago1, Ago2, Ago3, and Ago4, members of the Ago protein family involved in RISC function, play central roles in small RNA-mediated gene regulation [42]. Although recognition of mRNA targets is sequence-specific, the miRNA may not be completely complementary to its target. When the miRNA and its target sequence are 100% complementary, the mRNA target is digested by Ago2, which is the only Ago protein with slicer activity. When the miRNA and its target sequence are not 100% complementary, mRNA function is suppressed by translation inhibition or RNA instability. The seed sequence (2–8 nt from the miRNA 5′ end) is important for recognition between the miRNA and its target [43]. The seed sequence and corresponding region of its target are 100% complementary.
Many miRNAs show tissue-specific expression patterns [44–46] and play a crucial role in tissue development or homeostasis [44–46]. For example, targeted disruption of miR1–2, which is expressed in cardiac myocytes, causes a severe phenotype in heart development [47, 48]. Mice deficient forbic/miR-155 display increased lung airway remodeling [49, 50]. Several miRNAs have been associated with diseases, such as cancer and viral infections [51]. One of the human diseases in which miRNAs are involved is the neuropsychiatric disorder Tourette’s syndrome (TS) [52]. In some TS patients, the 3′ UTR of SLITRK1 containing a miR-189-binding site is mutated. This dysregulation of SLITRK1 by miR-189 is implicated in the onset of TS.
miRNA expression levels are also altered in primary human tumors [10, 53]. Differential expression patterns of miRNA profiles in various types of tumors suggest these miRNAs can be used as diagnostic markers or/and therapeutic targets [10, 44]. The miR-15a and miR-16-1 genes are located on the chromosome region 13q14, which is deleted in most cases of chronic lymphocytic leukemia [54]. These miRNAs target B cell lymphoma 2 (Bcl2), an antiapoptotic gene, suggesting that loss of miR-15a and miR-16-1 in B cells may lead to the inhibition of apoptosis, giving rise to malignancies [55]. miRNAs with oncogenic potential are expressed from the miR17-92 locus 13q31, which is amplified in some tumors, such as B-cell lymphoma [11, 56]. These miRNAs support a shift from apoptosis toward proliferation by the downregulation of E2F1. miR-372 and miR-373 are expressed specifically in testicular germ cell tumors [57]. In part, their function may be mediated by targeting the tumor suppressor LATS2 [57].
There is accumulating evidence that miRNAs are important in the immune system. Using genetic deletion and transgenic approaches, miR-155 has been shown to specifically regulate T-helper cell differentiation to produce an optimal T-cell-dependent antibody response. Also, the overexpression of miR-150 in hematopoietic stem cells, followed by bone marrow transplantation, altered B-cell maturation [58].
Expression profiling of 200 miRNAs in human monocytes has revealed that several miRNAs (miR-146a/b, miR-132, and miR-155) are endotoxin-responsive genes that can also be induced by cytokines [59]. By means of promoter analysis, miR-146a was found to be a nuclear factor (NF)-κB-dependent gene [59]. Importantly, miR-146a and miR-146b were predicted to base-pair with sequences in the 3′ UTRs of the TRAF-6 and IRAK-1 genes, and these UTRs inhibited the expression of a linked reporter gene [59]. These observations suggest that miR-146 downregulates Toll-like receptor and cytokine signaling.
In addition, it has recently been reported that miR-146 plays an important role in the development and activation of T-reg, a critical regulator of autoimmune disease pathogenesis [60]. We reported that miR-146 was significantly increased in rheumatoid arthritis synovium compared with osteoarthritis synovium and normal synovium [26, 61]. Consistent with the potential role of miR-146 functions in macrophages to attenuate inflammatory signals, the introduction of miR-146 in vitro inhibited tumor necrosis factor (TNF)-α-dependent inflammatory responses (our preliminary data). However, it is still unknown whether the introduction of miR-146 in arthritis can diminish rheumatic disease progression and be used as a therapeutic tool.
Accumulating evidence for the role of miRNAs in cancer, metabolic diseases, and viral infections suggests that they could be a new class of drug targets. Antisense oligonucleotide approaches for inhibiting miRNA function and small interfering (si) RNA-like technologies for the replacement of miRNAs are currently being explored as tools for uncovering miRNA biology and as potential therapeutic agents. Therapeutic trials using silencing miRNAs have recently been conducted [62].
Decreased expression of both miR-133 and miR-1, which belong to the same transcriptional unit, has been observed in mouse models and in human cardiac hypertrophy. In vivo inhibition of miR-133 by a single infusion of an antagomir caused marked and sustained cardiac hypertrophy, suggesting its potential therapeutic application in heart disease [62, 63].
Recently, we have provided evidence that a cartilage-specific miRNA, “miR-140”, plays a critical role in cartilage homeostasis, and we showed that mice overexpressing miR-140 in chondrocytes were resistant to osteoarthritis changes [64]. Further research on miRNAs, such as the methods outlined below, should provide us with a strategy to utilize miRNAs as therapeutic tools in rheumatoid arthritis and osteoarthritis.
With the following modifications, Northern blot analysis can be used to detect miRNAs: a polyacrylamide gel (approximately 15%) is used instead of an agarose gel to separate the small miRNAs; blotting is performed electrically; and 32P-labeled locked nucleic acid (LNA) probes [65, 66] are often used to increase sensitivity. Northern blot analysis is not very sensitive for detecting miRNAs, but provides the advantage of detecting both pre-miRNA and miRNA in a single experiment so that an abundance ratio can be determined.
There are three types of miRNA expression profiling assays, which are based on hybridization, RT-PCR, and sequencing.
Microarrays are hybridization-based assays that are widely used for miRNA profiling. Microarrays for miRNAs are supplied by several companies, including Exiqon (Vedbaek, Denmark) (miRCURY LNA miRNA array) and Agilent (Santa Clare, CA) (miRNA microarray). Although this is a simple and efficient method to obtain miRNA expression profiles, only known miRNAs can be detected.
Stem-loop RT-PCR miRNA assays are available in low-density formats array from Life Technologies Corporation. With this array, 384 miRNA RT-PCR assays can be carried out on a single plate. This method is fast, sensitive, and easy, but requires several dedicated instruments.
Whole-mount and section in situ hybridization are also used for miRNA detection. The procedures are similar to mRNA in situ hybridization techniques. Digoxygenin (DIG) is incorporated into the nucleic acid probe, but labeling methods for miRNA differ from those used for mRNA. To detect mRNA, DIG is incorporated into the probe as it is synthesized using RNA polymerase (internal labeling), but to detect miRNA, DIG is added to the 3′ end of an LNA-containing probe using terminal deoxynucleotidyl transferase. Although in situ hybridization is a powerful method for the identification of cellular or tissue miRNA expression, the sensitivity of this method is relatively low. To detect pri-miRNAs or pre-miRNAs, the conventional in situ hybridization method designed for mRNA can be used without modification. A detailed protocol has been published by Kloosterman et al. [71].
Because mRNA targets are specifically recognized by miRNAs with both perfect and imperfect complementary matching, it is difficult to identify these targets based on sequence information alone. Web-based programs that predict targets of miRNAs include miRanda (http://www.microrna.org/microrna/home.do), PicTar [73] (http://pictar.mdc-berlin.de/cgi-bin/PicTar_vertebrate.cgi), TargetScan [74] (http://www.targetscan.org/), MicroCosm (http://www.ebi.ac.uk/enright-srv/microcosm/cgi-bin/targets/v5/search.pl) [75, 76], and PITA (http://genie.weizmann.ac.il/pubs/mir07/mir07_prediction.html) [77]. These online tools use methods based on combinations of target features, such as seed sequence matching and the evolutionary conservation of 3′ UTR sequences of candidate genes. This in-silico approach to miRNA target identification is easy and quick; however, it is necessary to verify candidate targets by biochemical methods and ascertain whether the miRNA and candidate targets are co-expressed in the cell type of interest.
One function of miRNAs is reducing the stability of target mRNAs. Levels of a target mRNA can be reduced by overexpression of the cognate miRNA. Thus, an miRNA target may be detected by comparing mRNA expression profiles in cell lines overexpressing the cognate miRNA with cell lines that do not express the miRNA. However, it is still necessary to determine whether the targets are authentic using biochemical methods.
The luciferase assay is the biochemical method most widely used to identify miRNA targets. This method uses a luciferase reporter to detect inhibition of translation and mRNA degradation, but it cannot distinguish between these two processes. In-silico analysis and/or microarray assays are typically performed first to identify candidate targets of the miRNA. A complete or partial 3′ UTR sequence, or the miRNA target site is cloned into the 3′ UTR of the luciferase gene contained in a plasmid (test plasmid). A control plasmid is then constructed by generating one or two point mutations in the seed sequence of the candidate target sequence. The test plasmid or control plasmid is then transfected into a cell line with or without the miRNA overexpression vector, and luciferase assays are carried out in parallel. When the candidate is an authentic target, luciferase activity is lower in miRNA-overexpressingsss cells containing the control plasmid compared with those containing the test plasmid. An miRNA antagonist (see “Loss-of-function assays”) can be also used, but in this case, luciferase activity is higher in cells containing the control plasmid.
Gain-of-function experiments are performed by transfecting a plasmid containing a constitutive promoter (e.g., CMV or CAG) to overexpress a pri-miRNA or a pre-miRNA sequence. Viral vectors can also be used, or the pre-miRNA itself can be transfected. Life Technologies Corporation offers the Pre-mi miRNA Precursor Molecule, an miRNA mimic that is chemically synthesized as a modified double-stranded oligonucleotide.
Loss-of-function experiments against miRNAs have also been reported. Krützfeldt et al. [84] reported the long-term specific suppression of miR-16, miR-122, miR-192, and miR-194 in liver, lung, kidney, heart, intestine, fat, skin, bone marrow, muscle, ovaries, and adrenal glands of mice by the intravenous administration of a corresponding 21- to 23-nt antagomir (antisense 2′-O-methyl oligoribonucleotide). The antagomirs also work in transient transfection of cultured cells [15, 85–88]. Elmén et al. [89] reported that liver-expressed miR-122 was antagonized in the African green monkey by a phosphate-buffered saline-formulated LNA-modified oligonucleotide administered by intravenous injection. The LNA-modified oligonucleotide is also effective in transient transfection assays in cultured cells. These miRNA inhibitors show potential for clinical applications.
Several companies offer reagents for miRNA antagonism, including Exiqon A/S (miRCURY LNA microRNA Inhibitors) and Life Technologies Corporation (Anti-miRNA miRNA Inhibitors).
Most miRNAs can be downregulated by Dicer gene knockout in embryonic stem cells. However, because mouse embryos lacking Dicer die at 7.5 days post coitum, a conditional knockout mouse can be used for tissue-specific or cell type-specific Dicer knockout experiments [90].
The introduction of stem-loop RT-PCR and next-generation sequencers has increased the sensitivity and dynamic range of miRNA detection and profiling. The sensor method is also sensitive, but it requires expensive and time-consuming transgenic mouse production. An in-situ method for detecting miRNA that combines easy handling with high sensitivity is needed.
One of the most difficult issues in miRNA research is the identification of miRNA targets. In-silico methods are rapid and simple, but require validation by biochemical experiments. Thus, it may be necessary to develop an easy screening assay based on biochemical methods for the identification of miRNA targets.
Results of miRNA studies show promise for clinical applications, such as miRNA detection for diagnostics and gain-of-function/loss-of-function therapy. Therefore, the continuing development and improvement of miRNA techniques is crucial.