MicroRNA and Brain Tumors

What are miRNAs?

MiRNAs are a class of highly conserved small non-coding RNAs made up of single-stranded RNA molecules (approximately 20-25 nucleotides long) which powerfully negatively regulate gene expression by targeting mRNAs at the post-transcriptional level^[1]. It is generally believed that miRNAs have the following fundamental characteristics: i) they exist in all eukaryocytes; ii) have no open-reading frame (ORF); iii) the average length of the miRNAs is 22 nucleotides; iv) mature miRNAs are single-stranded; v) there are stable stem-loop structures in the miRNA precursors; vi) the coding gene of most miRNAs is conservative. About 84% of the miRNAs consist of 21~23 nucleotides. There is a phosphate group at the 5’ end and and a hydroxyl group at the 3’ end of the miRNA, while there is usually a 2nt overhang at the 3’ end. These unique structures distinguish the miRNAs from oligonucleotides and RNA denaturing frames^[2-9].

When Lee et al.^[2] first identified the member of the miRNA family, lin-4, in Caenorhabditis elegans, most investigators viewed this small molecule as an oddity only in worm genetics. Recently studies confirmed that miRNAs widely exists in a series of species including worms, flies, flowering plants, Arabidopsis thaliana and mammalians^[10,11]. Over the past 5 years, more than 3,000 miRNAs have been identified in vertebrates, flies, worms, plants and even in viruses^[12,13]. Employing a bioinformatics approach, it is suggested that there might be approximately one thousand human miRNAs while 326 miRNAs have been confirmed. MiRNAs account for ~ 3% of the total number of genes^[14] and are considered to be one of the biggest gene families. This feature may provide a broad perspective for miRNA research. It has been speculated that each miRNA is able to target ~ 200 transcripts directly or indirectly, whereas more than one miRNA can converge on a single protein coding gene target^[15]. Therefore, as a novel class of gene regulators with an enormous effect on gene expression, miRNAs may provide an important target for tumor prevention and therapeutics.

How do miRNAs become transcriptbed and mature?

MiRNAs are transcribed by RNA polymerase II (pol II) to generate a long primary miRNAs (pri-miRNAs) ^[16]. The pri-miRNAs that serve as the nascent primary transcripts of a miRNA coding gene are usually several kilobases long, and contain certain stem-loop hairpin structures. pri-miRNAs are often considered as immature mRNAs. The longer molecules are excised to ~ 80 nucleotide hairpin precursors (pre-miR-NAs) by the nuclear RNase III Drosha. The nuclear RNase III Drosha is a dsRNA-sepcific ribonuclease which contains a double stranded RNA-binding domain and two RNase III domains that play an important role to mediate the initiation step (‘cropping’) of pri-miRNAs^[17-19].

The two precursors (pri-miRNAs and pre-miR-NAs) are produced in the nucleus while the miRNAs are matured in the cytoplasm. The export of pre-miRNAs from the nucleus is very crucial to their maturation process. Many investigators have shown a great deal of interest in the mechanism by which pre-miRNAs are exported out of the nucleus. This process is mediated by exportin-5 and its cofactor Ran-GTP^[20]. Exportin-5 is a member of the key karyopher in family that mediates nucleomolecular nucleocytoplastic trafficking. The exportin-5/Ran-GTP complex has a high affinity for pre-miRNAs. Ran GTP has a higher concentration in the nucleus and it promotes a combination of pre-miRNAs with the exportin-5. The GTP is hydrolysed to GDP and the exportin-5/Ran-GDP complex releases the pre-miRNAs in the cytoplasm^[21]. The exported pre-miRNAs are further processed to a short double strand miRNA (~22 nucleotides) by Dicer. Dicer is poly-domain RNase III protein which combines ATP in the recognition and processing of miRNAs^{[16, 21, 22]}. It was originally confirmed to generate siRNA, the miRNA homologues. Dicer has been shown to be an evolutionally conserved protein that exists in all eukaryotic organisms, including plants, worms, flies and mammals^{[2, 23, 24]}. The maturation of miRNAs and siRNAs are similar, and the two pathways share a series of biogenesis steps^[21,22]. The mature miRNAs are mostly derived from one arm of the stem-loop hairpin precursors, and are released from the miRNA-miRNA duplex while the rest of the nucleotide sequences of the hairpin precursors are denatured^[25].

Taken together, miRNAs are processed in different localizations and the regulation of miRNAs can be conducted of at multiple levels. Knowledge of the gene regulation network by miRNA is still limited. Elucidation of gene regulation derived from miRNAs will require application of both transcriptomics and proteomics approches^[21-25].

How do miRNAs regulate gene expression?

Following their processing by Dicer in the cytoplasm, the miRNAs can regulate gene expression by silencing target mRNAs. Although the mechanism of miRNAs regulation is still unknown, we might obtain some insight by examing the best-studied miRNA lin-4, which negatively regulates its target, lin-14 by suppressing its expression^[26]. MiRNAs are able to guide the RISC complex to the target mRNAs and direct the large protein complex (RNA-induced silencing complex, RISC) to down-regulate gene expression by mRNA cleavage or translational repression. This process depends on the specific RISC complex assembled, and the degree of complementarity between a miRNAs and its mRNA target.

In eukaryocytes, there is only one strand of the miRNA-miRNA duplex; the miRNA strand can integrate the RISC while the complimentary strand miR-NA of the duplex is disposed by two mechanisms. In mammalians, the degree of complementarity in most miRNAs enables binding to the 3’untranslated region (3’-UTR). This binding is not as perfect as in plants^[27,28].

The “miRNA seed” that is encompassed by short 6~8 bases at the 5’ end of the mature miRNAs is responsible for searching for the complementarity to sequences in the 3’ UTR of all expressed genes. Most of the miRNAs suppress gene function based on partial complementarity, i.e., one miRNA may target more than one mRNA, and many miRNAs may act on a single mRNA, coordinately regulating gene expression, especially in mammalian cells and tissues. The mechanisms by which the miRNAs execute their biological function in animals and plants are significantly different. For most animal miRNAs, they do not inhibit the target mRNAs transcription, but suppress the target gene-coding protein translation. For example, lin-4 only reduces the expression level of the LIN-14 protein but not its mRNA level^[26]. In plants, this mechanism is not predominant. The base pairing between plant miRNAs and their corresponding targets is nearly perfect, and in order to direct mRNA cleavage, their complementary sites are located in the exons instead of targeting the 3’ UTRs^[29].

MiRNAs are associated with tumorigenesis

Cancer is a complex disease characterized by alterations in genes encoding oncogenic and tumor suppressor proteins. Increasing evidence suggests that miRNAs are extensively involved in tumorigenesis. Calin et al.^[30] reported that about half of the known human miRNAs are located in cancer-associated genomic regions, including fragile sites of chromosomes, common breakpoint regions, minimal regions of heterozygosity (LOH) loss, minimal regions of amplification, and hotspots for papilloma virus integration sites. For many miRNAs, DNA copy changes correlated with miRNA transcript expression, suggesting that the high-frequency abnormalities of the DNA copy number often occur in miRNA-containing regions throughout epithelial tumors.

The genomic association between miRNAs and human cancers was first reported in the study of human B-cell chronic lymphocytic leukemia, characterized by carrying 13q14.3 deletions, in which two clustered miRNA genes, miR-15a and miR-16-1, are involved. The encoding region of the miRNA cluster miR-15a-16 is located in 13q14.3, and has been confirmed by an array of comparative genomic hybridizations (aCGH). Mir-15a and mir-16-1 function as tumor suppressors that negatively regulate the expression of the anti-apoptotic gene, BCL-2. This might be an important genetic event in human tumorigenesis, including follicular lymphoma, lung cancer and B-CLL^[30,31].

Another most well studied miRNA, let-7 also functions as tumor suppressor, and may inhibit lung cancers by regulating oncogenes and/or genes that control cell differentiation or apoptosis, especially the Ras genes^[32]. The microarray data profile from the study of over 167 miRNAs demonstrated that underexpressed let-7 can be treated as a molecular biomaker in lung cancer^[33]. It has been reported that the overexpression of let-7 in human cancer resulted in decreased level of Ras as compared to the normal tissue. In C. elegan, let-7 can negatively regulate Ras gene expression by targeting its 3’UTR complementarity sites. This might provide a possible mechanism for oncogenesis of human cancer because about 15~30% of Ras gene mutations are found in human cancer. Mayr et al.^[34,35] reported that low expression of let-7 causes anchorage-independent growth, the most common characteristic of oncogenic transformation. The possible mechanism is that low expression of let-7 failed to repress oncogene High Mobility Group A2 (Hmga2). Taken together, let-7 may be a potential therapeutic target for treatment of lung cancer caused by activating mutations of the RAS gene.

Regarding digestive system tumors, it has been reported that genetic aberration of RAS and overexpression of C-MYC at the translational level are frequently involved in colon cancers. The let-7 family of miRNAs includes 14 isomers, and each isomer is usually located in a different chromosome. Let-7a-1 resides on chromosome 9q22.3 which is frequently deleted in colon cancers. Akao et al.^[36] has proposed the hypothesis that let-7 negatively regulates Ras and C-myc expression at the translation level by obtaining growth suppression after transfecting let-7a-1 miRNA precursor into the human colon tumor cells.

Studies have indicated that multiple mechanisms are involved in the silencing of ERα in ERα-negative breast cancers, including mutations in the 3’UTR of the ESR1 gene. Upregulated miR-206 suppress the endogenous ERα level, whereas a low level of ERα facilitates cancerous mutations in human breast cancers. In this regard, during the progression of human breast cancers the role of miR-206 has shifted from the classic tumor suppressor gene to one that contributes to tumorigenic progression known as “oncogenic addiction”^[37].

Elevated expression of mature miRNA has been identified in primary lymphoma and tumor cell lines. Mir-17-92, which is located at 13q31.3, intron 3, is the best-characterized evidence of a miRNA acting as a mammalian oncogene. The frequency of the mir-17-92 polycistron region deletion was significantly higher in the epithelial cancers than in the normal tissues^[38]. It has been reported that the apoptotic level was lower in B-cell lymphoma cells that overexpress mir-17-92^[28]. The possible explanation of the intricate mechanism is that the primary transcripts of the cluster are overexpressed in lymphomas and thereby negatively regulate E2F1 expression to slow down tumor development in B-cell lymphomas, while providing feed back to regulate the interaction with oncogene c-myc in order to accelerate oncogenesis^[35].

In addition, over-expressed miR-29 reduced the anti-apoptotic factor Mcl-1 expression causing facilitation of oncogenesis of cholangiocarcinoma; overexpression of miR-200 and underexpression of miR-127 may fail to repress the proto-oncogenes (RAB35, ERBB4, MAP3K3, VASP6, KIT, FOSB, NOTCH1) which contribute to papillary thyroid carcinoma; miR-34a functions as a potential tumor suppressor by inducing apoptosis in neuroblastoma cells^[39-41].

MiRNAs and human brain tumors

Glioblastoma multiforme (GBM) is the most common and highly invasive brain tumor. With the rapid increase of miRNAs knowledge, a considerable effort has been made to identify the miRNA profiles of these poorly-differentiated high-grade tumors.

Ciafrè et al.^[4] performed microarray analysis on miRNA expression profiles in 10 glioblastoma cell lines and 9 glioblastoma freshly resected samples, and identified a series of overexpressed miRNAs including: miR-10b, miR-130a, miR-221, miR-125b-1, miR-125b-2, miR-9-2, miR-21, miR-123 and miR-25. The control samples were resected from peripheral areas at an average distance of 2 cm away from the visible tumors. Furthermore, Ciafrè et al.^[4] reported a group of down-regulated miRNAs including, miR-128a, miR-181c, miR-181a, and miR-181b. Similar results were also reported by Barrad’s group^[42]. Among those miRNAs, miR-10b and miR-221 were the only 2 miRNAs that over-expressed in both glioblastoma cell lines and tissue samples. It’s generally believed that the over-expressed miRNA (miR-221) and the under-expressed miRNAs (miR-128, miR-181a and miR181b) can be considered to be a specific molecular signature of GBM^[4]. However, the relationship between the WHO classification of gliomas and the expression level of certain miRNAs is still unknown.

MiR-21 over-expression is one of the common features of glioblastoma multiformes. Chan et al.^[43] found that miR-21 functions as an antiapoptotic factor in GBMs by knocking down miR-21 with antisense oligonucleotides. The expression level of miR-21 was 1.5~2 fold higher in most organs (liver, heart, ovary, kidney etc.) compared to that in normal brain. Northern blot analysis identified that both miR-21 (5~100 fold) and pre-miR-21 5(5~30 fold) were higher in glioma samples than the normal brain tissue (e.g.: fetal brain, adult cortex and adult white matter) as well as in the early-passage of glioma cell cultures established from resected fresh samples. These findings suggest that miR-21 plays an important role in the pathogenesis of brain tumors^[43].

The suppression of miR-21 was able to increase the activity of the caspase family and induce an apoptotic process. The caspase 3/7 activity of glioma cells was enhanced after being treated with the LNA-miR-21. Accordingly, the number of apoptotic cells increased significantly when the cells were treated with LNA-miR-21 and 2’OMe-miR-21^[43]. They proposed that an abnormal increase in expression of miR-21 may suppress apoptotic-related genes via miRNA expression directly or through other pathways. Overall, these studies demonstrated that miR-21 is not only one of the micro-oncogenes, but also is a potential therapeutic target for human GBM^[43].

The studies of knocking down the expression of miR-21 in human glioma in vivo are limited. Si et al.^[44] observed a satisfactory therapeutic effect of anti-miR-21 in a human breast cancer xenograft model. They suggested that anti-miR-21 therapy can induce tumor cell apoptosis possibly through targeting Bcl-2.

MiR-15a and miR-16-1 genes are located at chromosome 13q14, a region that is frequently deleted in many human tumors^[31,45-47]. Bottoni et al.^[47] confirmed that, miR-15a and miR-16-1 are expressed at lower levels in pituitary adenomas compared to normal pituitary tissue. Moreover, their expression in-versely correlated with the tumor diameter and RARS (arginyl-tRNA synthetase) expression, but positively correlated with p43 secretion. The results suggest that these genes may, at least in part, regulate tumor growth.