Journal list menu
Tumor suppressor microRNAs: A novel non-coding alliance against cancer
Abstract
Tumor initiation and progression are the outcomes of a stepwise accumulation of genetic alterations. Among these, gene amplification and aberrant expression of oncogenic proteins, as well as deletion or inactivation of tumor suppressor genes, represent hallmark steps. Mounting evidence collected over the last few years has identified different populations of non-coding RNAs as major players in tumor suppression in almost all cancer types. Elucidating the diverse molecular mechanisms underlying the roles of non-coding RNAs in tumor progression might provide illuminating insights, potentially able to assist improved diagnosis, better staging and effective treatments of human cancers. Here we focus on several groups of tumor suppressor microRNAs, whose downregulation exerts a profound oncologic impact and might be harnessed for the benefit of cancer patients.
1 Introduction
In most epithelial tissues, cancer develops through separate and interrelated steps of clonal expansion, genetic diversification, and clonal selection. During cancer development, cancer cells acquire diverse biological capabilities that are conferred by numerous genetic and epigenetic modifications [1]. In recent years, different high-throughput platforms have been used for the genomic, transcriptomic, proteomic, and epigenetic analyses to search for new biomarkers involved in cancer and to bring new insights into the several aspects of cancer pathophysiology [1]. In addition to the classical transcriptional cell regulators involved in cancer development, a class of non-coding RNAs, termed microRNAs (miRNAs) has emerged as critical regulators of gene expression acting predominantly at the post-transcriptional level. MiRNAs were first identified through their ability to regulate developmental processes, such as developmental timing and cell fate transitions [2]. Subsequently, miRNAs have been studied in relation to cancer development. A large number of miRNAs that map to specific regions of the human genome have been shown to be frequently deleted or amplified in cancer [3]. Several lines of evidence indicate that miRNAs might be differentially expressed in cancer cells, in which they form unique expression patterns or signatures [4]. Sevignani and colleagues reported a significant association between the chromosomal location of miRNAs and those of mouse susceptibility loci that influence the development of solid tumors [5]. Dysregulation of miRNAs in cancer can occur through both epigenetic changes, including aberrant DNA methylation and histone modification [6], and genetic alterations. These two biological mechanisms can affect the production of the primary RNAs, their processing to the mature miRNA forms, and/or interactions with mRNA targets [7].
More recent studies indicate that mutations affecting proteins involved in the processing and maturation of miRNA, such as TARBP2, DICER1 and XPO5, can also lead to overall reductions in miRNA expression [8-10]. Consistent with these observations, miRNAs are thought to act mainly as tumor suppressor genes, and their deregulation is currently recognized as a common feature of human cancers. Later on, additional data indicated that the expression of miRNAs is mainly downregulated in tumor tissues, as compared to corresponding healthy tissues, which supported the role of miRNAs as primarily tumor suppressors [4, 8, 9, 11, 12]. In the same vein, there is evidence that an extensive downregulation of miRNAs is one of the first outcomes of the stimulation of signaling cascades downstream to specific growth factor receptors implicated in a number of human cancers, including breast cancer [13]. For example, EGF signaling rapidly and simultaneously induces an extensive downregulation of multiple miRNAs, reflecting coordinated regulation at the level of miRNA synthesis, processing or degradation [13].
Along with the dominance of tumor suppressor microRNAs, several well-characterized oncogenic miRNAs have been reported in tumors. An interplay between RNA-binding proteins and oncogenic miRNAs, which drive expression of proto-oncogenes or maintenance of stem cell phenotypes, contributes to human cancer [14]. One example relates to oncogenic receptors for growth factors, such as the EGF-receptor (EGFR/ErbB) family of receptor tyrosine kinases, the expression of which is regulated by several miRNAs [15]. In the same vein, signaling pathways are ideal candidates for miRNA-mediated regulation, owing to the sharp dose-sensitive nature of their effects. For instance, EGFR activation induces miR-7 expression through a RAS-MYC pathway. In support of this, MYC binds to and activates the miR-7 promoter and ectopic miR-7 promotes cell growth and tumor formation in lung cancer cells [16]. Thus, in addition to the EGFR/ErbB family, oncogenic miRNAs (onco-miRs) affect the responsiveness of cells to signaling molecules, such as transforming growth factor-beta, WNT and Notch [17]. miRNAs control not only cellular proliferation and programmed cell death, but also dissemination of tumor cells and colonization of distant organs (metastasis). Indeed, some miRNAs are associated with the invasive and metastatic phenotype of breast and other cancer cell lines or metastatic tumor tissues [18, 19].
MiRNAs are also deregulated upon exposure to both metabolic cancer risk factors and exposures to carcinogenic substances [20, 21]. Thus, miRNAs may represent at the same time both predictors and players of cancer development. A number of life-style factors (e.g., diet rich in fats and refined carbohydrates) and pathological conditions (e.g., obesity), often related to inflammation and cancer, result in deregulation of specific miRNAs [22-25]. In addition, there is evidence of an altered expression of miRNAs in relation to the exposure to well-known carcinogenic substances such as asbestos, formaldehyde and cigarette smoke in lung and hepatic tissue [26, 27]. In regard to this evidence, one study examined the expression of 484 miRNAs in the lungs of rats exposed to environmental cigarette smoke for 4 weeks. It was found that 126 miRNAs were down-regulated at least 2-fold and 24 miRNAs were downregulated more than 3-fold [28].
In this review, we highlight the contribution of miRNA modulation, in particular prevalent downregulation of specific miRNAs, to cancer development. Due to space consideration, this review concentrates on a selected group of tumor suppressor microRNAs. Table 1 lists some additional molecules within this category, which we do not discuss in the main text. They include miR-34, a p53 target gene [29, 30], miR-31, an inhibitor of metastasis [31], as well as miR-205 [32], miR-375 [33], miR-203 [34-36], as well as miR-15a [37-39]. Because many downregulated miRNAs function as tumor suppressors, better understanding of the biological mechanisms underlying their modulation will likely enable new strategies for prevention, early detection and therapy of cancer.
| microRNA | Target | Function | References |
|---|---|---|---|
| miR-34 | CDK4 cylinE2 c-MetCDK6NotchHMGA2BCL-2SIRT1AXL | Induces cell cycle arrest and apoptosisA p53 target geneEpigenetically silences in some tumors | [29, 30] |
| miR-205 | E2F1LAMC1 | Induces senescence and reduces cell proliferationDirectly transactivated by wild type p53 | [32] |
| miR-375 | PDK114-3-3fAEG-1IGF1-receptor | Suppresses cell growth through the AKT pathwayEpigenetically silenced by DNA methylation | [33] |
| miR-31 | Integrin alpha5RhoAMMP16RadixinWAVE3 | Inhibits motility and invasivenessInduces anoikis | [135] |
| miR-203 | Caveolin 1LASP1 | Reduction of proliferation and inhibition of migration | [34-36] |
| miR-15a | Cyclin D1WNT3ACryptoBCL2 | Inhibition of ovarian, breast and lung cell proliferation and invasionInduces apoptosis | [37-39] |
1.1 MiR-10b3p: the early arm of the miR-10b locus
The so-called miRNA-10b locus is located on chromosome 2, within the cluster of the HOXD genes, in an intergenic region between HOXD4 and HOXD8 [40]. Processing by Drosha and Dicer transforms the RNA product of the miRNA-10b locus into a 22-nucleotide RNA duplex that contains two distinct 5′ phosphorylated strands with 3′ overhangs (Fig. 1 ). The functional strand of the duplex, referred to as the guide strand, is miR-10b5p while the other, passenger strand generates miR-10b3p. MiR-10b5p was originally identified as a molecules down-regulated in primary breast tumors, compared to normal breast tissues [41]. Similarly, downregulation of miR-10b5p by promoter hyper-methylation has been reported in gastric tumors [42]. The Weinberg's group later reported that miR-10b5p acts as a metastasis-supporting miRNA, due to its ability to favor migration and invasion of breast cancer cells [43, 44]. In line with this, miR-10b5p targets the HOXD10 gene, a repressor of several modulators of cell migration [44]. The expression of miR-10b5p is tightly controlled by the transcription factor Twist, a well-established regulator of epithelial-to-mesenchymal transition (EMT). Increased expression of miR-10b5p was detected in the vasculature of breast IDC (invasive ductal carcinoma) grade III tumors, compared to lower expression in DCIS (ductal carcinoma in situ) [45]. The pleiotropic activity of miR-10b5p could also rely on its ability to target the expression of diverse tumor suppressors, including TP53, HODX10, PAX6, NOTCH1 and FOXO3 (see Table 2 ) [46].
Biagioni et al. originally reported that the expression of miR-10b3p was down-regulated in breast tumors, relative to matched peri-tumoral tissues [12]. This downregulation occurred, at least in part, through the methylation of CpG islands located within the regulatory regions of the miR-10b locus [12]. Ectopic expression of miR-10b3p inhibited proliferation of breast cancer cell lines and reduced the size of xenografted breast tumors. Three pivotal proteins involved in the control of cell proliferation, namely BUB1, PLK1, and CCNA2, were shown to serve as targets of miR-10b3p. Accordingly, intratumoral injection of a mimic of miR-10b3p reduced the expression of BUB1, PLK1 and CCNA2 proteins [12]. The prognostic role of miR-10b3p and of its target was evaluated in the MEATBRIC dataset. This analysis included 1286 breast cancers from 5 different subtypes: HER2+ (127 patients), basal-like (209 patients), luminal A (479 patients), luminal B (312 patients), and normal-like (151 patients), for which both mRNA and miRNA data were available [12]. mRNAs and miRNAs were measured for each tumoral and normal samples of the METABRIC dataset. Kaplan-Meyer analysis revealed a significant association between low expression levels of miR-10b3p and poor disease-specific survival [12]. This association was not evidenced for the augmented expression of miR-10b5p. The combined application of the COSMIC algorithm [47] and miRanda predictions uncovered 15 target mRNAs of miR-10b3p [12]. Among those targets three, BUB1, PLK1 and CCNA2 were confirmed, and additional cell cycle related targets were also identified. Increased expression of BUB1, PLK1 and CCNA2 was associated with poor survival (Table 2) [12] .
These findings have several implications to the roles played by miR-10b in breast tumorigenesis. Presumably, the expression of miR-10b3p is altered in the early stages of mammary cell transformation. This could lead to aberrant cell proliferation, mediated by increased expression of the cell cycle related targets of miR-10b3p. In line with early alterations, downregulation of miR-10b3p expression appears to occur independently from the subtype of breast cancer, suggesting that it might represent an event preceding specification of breast cancer subtypes. Interestingly, the regulation of the expression of the two strands derived from the miR-10b locus is controlled by the combination of epigenetic and transcriptional events. While downregulation of miR10b-3p occurs through methylation of CpG islands, the transcription factor TWIST up-regulates the expression of miR-10b5p [44]. This might underlie mechanistically the dual and opposite roles of the miR-10b locus: Early in breast tumorigenesis miR-10b3p downregulation leads to aberrant cell proliferation, while TWIST-mediated transcriptional induction of miR-10b5p contributes, as a late step, to shape a metastatic phenotype.
Unlike downregulation of miR-10b3p, which occurs independently from the breast cancer subtype, up-regulation of miR-10b5p might be specifically selected in highly metastatic breast tumors. Thus, waves of miR-10b3p and 5p targets might tune the pro-proliferative and pro-metastatic activities of the aberrantly activated miR-10b locus. Intriguingly, miR-10b3p (previously named miR-10b∗) could represent a prototype of miRNAs derived from passenger strands, which target specific mRNAs and exert biological activities as efficiently as those originated from guide strands. While the role of miR-10b5p is relatively well documented in different human cancers, that of miR-10b3p is poorly investigated. This also indicates that the miR-10b locus, via downregulation of the 3p strand or up-regulation of 5p plays a pivotal role in breast cancer establishment or dissemination. Once, the molecular events responsible for aberrant activation of the miR-10b locus in tumors will be fully understood, they might hold promise for novel therapeutic strategies. This might turn true also for passenger derived tumor suppressor miRNAs.
1.2 Let-7c acts as a tumor suppressor miRNA
The let-7 family is one of the most ancient and conserved groups of miRNAs, showing high conservation across species from Caenorhabditis elegans to mammals [48]. In humans, the let-7 family is comprised of ten members (let-7a, let-7b, let-7c, let-7d, let-7e, let-7f, let-7g, let-7i, miR-98 and miR202), which differ in their nucleotide sequences. Some of the isoforms appear in multiple copies in the genome, hence a number is added as a suffix to their name (e.g., let-7a-1, let-7a-2 and let-7a-3) (reviewed in [49]). Interestingly, many of the let-7 miRNAs are located in fragile sites and specific genomic regions that relate to cancer [50]. Thus, for example, in the human genome, the cluster let-7a-1/let-7f-1/let-7d is included in a frequently deleted region of chromosome 9 (region B at 9q22.3). In breast carcinomas, a region of LOH (loss of heterozygosity) at 11q23 was shown to harbor the cluster miR-125b1/let-7a-2/miR-100, and the cluster miR-99a/let-7c/miR-125b-2 resides at 21p11.1, a region of frequent HD (homozygous deletions) in lung cancers. Furthermore, let-7a-1, let-7f-1, let-7d and miR-202 are located close to class II homeotic genes in the human HOX gene clusters [50].
The let-7 family plays crucial roles in cellular differentiation and in development, and it displays specific temporal and spatial expression patterns during development of several species [48, 51]. For example, in mammals, the level of let-7 increases during embryogenesis and during brain development [52]. In concordance with developmental roles for the let-7 family, members of this family are also involved in cancer, and as discussed below, many of them act as tumor suppressor miRNAs. Downregulation of most let-7 family miRNAs was demonstrated in several types of human cancer, including lung, head and neck squamous cell carcinoma, breast, melanoma, ovarian and prostate cancers. Interestingly, silencing all let-7 family members in ovarian cancer cell lines increased cell survival, invasion and adhesion [53]. The targets of let-7 are also conserved from worms to humans, and they include known oncogenic transcripts, such as MYC and RAS [54]. Interestingly, the viral homolog of KRAS contains a single nucleotide polymorphism (SNP) in its potential let-7 binding site, which increases the risk for lung, oral and colorectal cancers [55-57]. Another cancer related target of the let-7 family is the high-mobility group AT-hook 2 (HMGA2) oncoprotein, a chromatin-associated non-histone protein that affects transcription by modulating chromatin's architecture [58].
Let-7c, a member of the let-7 family, functions as a tumor suppressor in several types of cancer (Fig. 2 ). It regulates EMT [55] and targets various oncogenes and cancer related genes [59], such as: N-RAS [54], c-MYC [60], HMGA2 [58], MMP11, PBX2, PBX3, TRIB2, ITGB3, TGFBR1 [61], BCL-XL and MAP4K3 (see Table 3 ). In support of a role for let-7c as a tumor suppressor, its reduced expression was demonstrated in tumors and cultured cells from prostate, lung, colorectal and hepatocellular tumors [60, 62-65]. Additionally, downregulation of let-7c was associated with poor prognosis in colorectal cancer, where let-7c expression was significantly lower in patients presenting lymph node involvement or distant metastases, compared to patients without any detectable metastasis [65]. In non-small cell lung carcinoma (NSCLC), low expression of let-7c was associated with poor patient survival, as well as with tumor spread, venous invasion and advanced disease stages [64]. Interestingly, let-7c is encoded by chromosomal locus 21q21, which shows loss of heterozygosity in lung cancer [66].
| Tissue | Target | Function | References |
|---|---|---|---|
| Lung cancer | RASTRIB2 | Inhibition of proliferation | [54, 136] |
| Bcl-xL | Inhibition of chemo- or radioresistance and EMT | [55] | |
| ITGB3MAP4K3 | Inhibition of migration and invasion | [64] | |
| Acute myeloid leukemia | PBX2 | Inhibition of leukemogenesis | [62] |
| Uterine leiomyoma | HMGA2 | Inhibition of proliferation | [58] |
| Colorectal cancer | MMP11PBX3 | Inhibition of metastatization | [65] |
| Prostate cancer | MYC | Suppression of androgen receptor via regulation of MYC | [60] |
| Kidney fibrosis | TGF-beta receptor 1 | Inhibition of fibrosis by suppression of a receptor for TGF-beta | [61] |
The maturation process of miRNAs of the let-7 family, including that of let-7c, is regulated by the RNA-binding proteins called Lin28 and Lin28B, which employ several different regulatory mechanisms [67]. Lin28 is expressed in the cytoplasm and blocks processing of let-7 by Dicer through direct binding to the terminal loop of pre-let-7 and by recruiting the terminal uridyl transferase (TUTase) Zcchc11/TUT4 to catalyze the oligouridine tail, thus marking pre-let-7 for degradation [68, 69]. Lin28B is found mainly in the nucleolus, and upon binding to pri-let-7 it blocks the activity of the microprocessor complex through TUTase-independent mechanism (reviewed by [70]). Interestingly, miRNAs of the let-7 family have been suggested to reciprocally downregulate the expression of Lin28 and Lin28B and thus increase the levels of mature let-7 family miRNAs [71, 72]. Let-7c's expression levels can also be regulated by the peroxisome proliferator-activated receptor α (PPARα) in hepatocellular carcinoma cells [63].
Manipulation of let-7c expression in cancer cell lines further established its role as a tumor suppressor. For example, overexpression of let-7c decreased, whereas depletion of let-7c increased cell proliferation and clonogenicity of prostate cancer cells in vitro. Let-7c also exerted an anti-proliferative effect in vivo, when tested by intratumoral injection, which significantly reduced tumor size in xenografts of human prostate cancer cells [60]. According to recent studies, let-7c can also function as a metastasis suppressor. In highly metastatic colorectal adenocarcinoma cells, ectopic over-expression of let-7c led to reduced migration and invasion in vitro, and almost completely inhibited tumor growth and metastasis. When tested in weakly metastatic cells, let-7c inhibition resulted in increased cell motility and invasion. These effects of let-7c on motility were likely mediated by targeting of KRAS, MMP11 and PBX3 [65]. Additionally, ectopic let-7c expression in chemotherapy-resistant lung adenocarcinoma cells reduced the number of metastatic nodules in lung and liver, probably via inactivation of the AKT pathway [55].
In addition to the AKT pathway, let-7c controls several other cancer-related signaling pathways. Thus, let-7c plays an important role in the regulation of androgen receptor (AR) signaling by directly targeting MYC, thereby controlling prostate cancer proliferation. Accordingly, the expression of let-7c and AR are negatively correlated in human prostate cancer, where AR, MYC and Lin28 expression levels are high while let-7c expression level is low [60]. In acute leukaemia, let7-c targets PBX2, a homeodomain protein, which upon aberrant expression enhances HoxA9-dependent leukemogenesis, and promotes granulocytic differentiation [62]. Let-7c may also suppress cell growth and control cancer pathogenesis by regulating the mitogen-activated protein kinase (MAPK) pathway. In lung adenocarcinoma, let-7c directly targets TRIB2, which in turn activates the downstream components, C/EBP-α and a phosphorylated p38-MAPK [73]. Another pathway by which let-7c can regulate the MAPK pathway is through its role in controlling NRAS expression [54]. One of the key metastasis-driving processes is EMT [74], which is affected by several transcriptional switches, including a let-7c regulated switch. Docetaxel-resistant lung adenocarcinoma cells are characterized by fibroblast-like morphology and adhesion, which are typical of the mesenchymal phenotype. However, upon expression of the let-7c precursor, these cells gained an epithelial phenotype, which might contribute to the restoration of chemo- and radio-sensitivity. This effect of let-7c is probably achieved through direct targeting of BCL-XL [55].
In summary, let-7c plays a crucial role in cancer pathogenesis through targeting key cancer-related proteins and acting as a suppressor of both tumor growth and tumor spread. Thus, let-7c might be considered an attractive candidate for drug-induced manipulation in cancer therapy.
1.3 miR-223 in cell differentiation and in tumor suppression
MiR-223 was initially identified as a miRNA nearly exclusively expressed in bone marrow, such that its functional role in the regulation of the cell fate determination of human hematopoietic progenitors cells (HPCs) rapidly emerged [75, 76]. Recently, Vian and colleagues evidenced that in human CD34 + HPCs undergoing unilineage differentiation/maturation, miR-223 is up-regulated during granulopoiesis rather than during monocytopoiesis and is maintained at low levels during erythropoiesis [77]. Interestingly, miR-223 overexpression in human CD34 + HPCs favors granulopoiesis and impairs erythroid and monocytic/macrophagic differentiation [77].
The fine-tuning of miR-223 expression levels during hematopoietic differentiation of HPCs, as well as of myeloid cell lines, into erythroid, granulocytic and monocytic/macrophagic lineages is the result of the coordinated recruitment and function of lineage-specific transcription factors (TFs) on two different regulatory regions of the miR-223 promoter [76, 77]. For example, during granulocytic differentiation of human myeloid progenitors, the induction of miR-223 expression is transcriptionally regulated by competitive binding of two transcription factors (TFs), Nuclear Factor I (NFI-A) and CCAAT/enhanced-binding protein alpha (C/EBPα), to the proximal miR-223 regulatory region. NFI-A maintains miR-223 transcription at low levels, while its replacement by C/EBPα results in miR-223 induction and granulocytic differentiation [78]. NFI-A was also identified as a target for miR-223 at transcriptional and translational level, thus establishing a feedback regulatory circuitry in the control of granulopoiesis [76, 79]. Interestingly, miR-223 expression is downregulated in different subtypes of acute myeloid leukemia (AML), which represents the clonal expansion of hematopoietic precursors blocked at different stages of differentiation [80]. In particular, primary blasts carrying the t(8;21) generating AML1/ETO, the most common acute myeloid leukemia-associated fusion protein, present very low expression levels of miR-223. In these cells the AML1-ETO oncoprotein, which recruits an epigenetic silencing complex consisting of HDACs, DNMTs, and methyl-CpG-binding proteins on the AML1 binding site of the miR-223 promoter, links the epigenetic silencing of a miRNA locus to the differentiation block of this acute myeloid leukemia subtype [80]. Of note, the enforced expression of miR-223 in primary AML blasts and in AML cell lines affects cell cycle progression and enhances granulocyte differentiation [76, 77, 80, 81].
The tumor suppressor function of miR-223 in acute myeloid leukemia is further supported by a recent study showing that the induction of miR-223 inhibits the translation of the cell-cycle regulator E2F1 and significantly down-regulates the proliferation rate of myeloid progenitors cells [81]. Of note, E2F1 transcriptionally represses miR-223 gene in AML cells, suggesting that miR-223 functions as a key regulator of myeloid cell proliferation is strictly linked with E2F1 activity in a mutual negative feedback loop.
Exosomes or micro-vesicles are emerging as important mediators of intercellular cross-talk for the regulation of a variety of biological functions, such as cellular communication, proliferation and differentiation [82]. Micro-vesicle transfer represents a novel mechanism by which infiltrating mononuclear phagocytes may contribute to cellular activation, survival, and immune function. MiR-223 was evidenced to be the most highly expressed miRNA in the macrophage-derived micro-vesicles that are able to induce cellular differentiation when added to naive monocytes, supporting a functional amplification loop to enhance immune function (Fig. 3 A) [83].
In addition to the exosome-mediated transfer of nucleic acids, tissue macrophages were also shown to be able to transfer miR-223 to hepatocarcinoma cells (HCCs) in a manner that required intercellular contact and gap junction communication (Fig. 3B) [84]. Functionally, the transfer of miR-223 from macrophages to HuH7 cells resulted in the inhibition of proliferation of these HCCs cancer cells, thus highlighting intercellular transfer of miRNA from immune cells as a new, possible mechanism of defense against neoplastic cell transformation or tumor growth [84]. The transfer of miR-223 influences protein expression in HuH7 cells. Specifically, miR-223 decreases the expression of Stathmin-1 (STMN1) and insulin-like growth factor-1 receptor (IGF-1R), which both influence cellular proliferation and can support the growth of tumors [85, 86]. STMN1 is a key microtubule-regulatory protein that controls microtubule dynamics, cell proliferation and, in particular, the S-phase of the cell cycle. High-levels of STMN1 have been associated with increased histologic grading, shorter patient survival times, and increased drug resistance in different tumors. STMN1 is a protein that is usually present at low levels in healthy hepatocytes but is expressed at high levels in hepatocarcinomas [86]. In HCC cell lines, a strong inverse relationship between STMN1's mRNA and miR-223 expression was observed and a substantial reduction in STMN1 protein was demonstrated upon restoration of miR-223 expression, resulting in a consistent inhibitory effect on cell viability [86].
Consistent with these lines of evidence, it was recently reported that modulation of miR-223/STMN-1 pathway represents another way by which mutant p53 increases cellular resistance to chemotherapeutic drugs [87]. The induction of mutant p53R175H in breast and colon cancer cell lines decreases miR-223 expression. Mutant p53R175H, together with the transcriptional repressor ZEB-1, binds to the miR-223 promoter and decreases miR-223 expression, resulting in an up-regulation of STMN-1 and in an increased cellular resistance to chemotherapeutic agents. On the contrary, ectopic expression of miR-223 can lower the levels of STMN-1 and sensitize breast and colon cancer cell lines expressing mutant p53 to treatment with DNA-damaging drugs [87].
By targeting IGF-1R and the cyclin-dependent kinase 2 (CDK2), miR-223 functions as a potent tumor suppressor also in the Lewis lung carcinoma (LLC) cell line [88]. Ectopic expression of miR-223 suppresses proliferation, invasion and tumorigenicity of LLC cells, induces G2/M phase arrest and inhibits tumor growth in vivo, providing the basis for novel therapies targeting IGF-1R in the treatment of NSCLC [88].
The inhibition of cancer cell proliferation by miR-223 was recently reported also in colorectal cancer [89]. The colorectal cancer HCT116 cells express very low levels of endogenous miR-223 and overexpression of miR-223 in these cells reduces mRNA and protein expression levels of FOXO1, whose abnormal expression or activation can result in aberrant apoptotic pathways, proliferation, and cell cycle regulation. Interestingly, miR-223 overexpression mainly increases the un-phosphorylated FOXO1 protein, its nuclear localization, as well as cyclin D1/p21/p27 at either mRNA or protein accumulation and inhibition of tumor cell proliferation [89].
Although several experimental lines of evidence support the involvement of miR-223 in cell differentiation and tumor suppression, miR-223 was also reported to be up-regulated in some tumors and to contribute to leukemogenesis in specific disease subtypes (T-ALL), indicating that the biological effect of miR-223 strongly depends on the cellular context where this miRNA specifically performs its function [90, 91] (see Table 4 ).
| Tissue | Target | Function | References |
|---|---|---|---|
| Acute myeloid leukemia | E2F1 | Inhibition of proliferation | [81] |
| Hepato-carcinoma | STMN1IGF-1RABCB1 | Inhibition of proliferation and of multidrug resistance | [86, 85, 137] |
| Lewis lung carcinoma | CDK2 | Inhibition of proliferation and invasion | [88] |
| Colorectal cancer | FOXO1 | Inhibition of proliferation | [89] |
| Glioblastoma | PAX6 | Proliferation invasion | [138] |
| Breast cancer | Caprin-1 | Inhibition of proliferation and invasion | [139] |
| Breast and colon cancer | STMN1 | Inhibition of chemoresistance | [87] |
| Gastric cancer | EPB41L3 | Invasion and metastatization | [140] |
| Esophageal carcinoma | ARTN | Inhibition of invasion and metastatization | [141] |
1.4 miR-145 in tumor suppression
miR-145 is located at chromosome 5q33.1 and is usually transcribed in a bicistronic primary transcript with miR-143 [92]. Of note, miR-145 represents one of the miRNAs that are highly expressed in normal tissues but they are down-regulated in several human cancers, including colorectal and breast cancer [41, 93]. Although the downregulation of miR-145 expression in human cancer may result from genetic aberrations, as occurs in hematological malignancies associated with the 5q- syndrome phenotype [94], a major mechanism for the modulation of miR-145 expression is represented by transcriptional and post-transcriptional control. Interestingly, transcriptional contribution to the regulation of miR-145 expression was reported in breast and in colon cancer, cell lines where p53, by interacting with a consensus sequence in the miR-145 promoter, transcriptionally induces miR-145 expression, promoting the post-transcriptional downregulation of MYC and consequently the inhibition of tumor cell growth [95]. In line with these results, in prostate cancer tissues and in cell lines, miR-145 was found to be silenced through the methylation of its promoter, and miR-145 silencing was significantly correlated to the status of the p53 gene [96]. A recent study also reported that the CCAAT/enhancer binding protein beta (C/EBPβ) is able to counteract the p53-mediated induction of miR-145. In fact, C/EBPβ induces the transcriptional downregulation of miR-145 expression by interacting with a CCAAT binding site located within the miR-145 promoter; this downregulation seems to be independent from p53 and it involves the AKT pathway [95, 97].
Concerning miR-143/miR-145 processing, it was recently reported that p53 itself might be involved in the post-transcriptional maturation of several miRNAs with growth-suppressive functions, in response to DNA damage, including miR-143/miR-145. In particular, the p53 protein, through an association with the DEAD-box RNA helicase p68 (also known as DDX5), interacts with the Drosha processing complex, thus supporting the processing of primary miRNAs to precursor miRNAs. On the contrary, p53 mutated proteins interfere with the functional assembly between the Drosha complex and p68, inducing an attenuation of miRNA processing activity [98]. Moreover, it was also recently reported that the DEAD-box RNA helicase 6 (DDX6), that is highly expressed in most malignant cells, post-transcriptionally down-regulates both miR-143 and miR-145 expression. In human gastric cancer cells, DDX6 protein, that is abundantly expressed and accumulated in processing bodies (P-bodies) containing many proteins involved in mRNA turnover, negatively regulates the RNA stability of the bicistronic primary transcript resulting in the downregulation of both miR-143 and miR-145 [99]. Several reports, supporting miR-145's action as a tumor suppressor in different tumor types, have revealed that the downregulation of miR-145 expression is associated with neoplastic cell growth and proliferation, as well as with cancer cell migration, invasion and metastasis (Table 5 ) [100-104],
| Tissue | Target | Function | References |
|---|---|---|---|
| Colon cancer | IRS-1YES and STAT1c-Myc | Growth and proliferation | [105, 106] |
| p70S6K1 | Growth and angiogenesis | [95] | |
| JNK and MAPK pathways | Colony and tumor formation | [107] | |
| CD44, KLF5, KRAS, BRAF | Transformation properties | [118] | |
| Lung cancer | EGFR, NUDT1 | Proliferation and tumorigenesis | [108] |
| Breast cancer | MUC1JAM-A and fascin | Migration and invasion | [100, 101] |
| N-RAS and VEGFARTKN | Angiogenesis, invasion and growth | [110] | |
| p53 pathway and ER-alpha | Proliferation and apoptosis | [111] | |
| Zeb2 and Klf4 | Cancer stem-like cells function | [112] | |
| Mesothelioma | OCT4 | Growth and clonogenicity | [116] |
| Pancreatic cancer | KRAS and RREB1 | KRAS-mediated tumorigenesis | [117] |
| Prostate cancer | BNIP3FSCN1ERG | Proliferation, migration, invasion, EMT and metastasis | [120, 104, 122] |
In colon cancer, the tumor suppressor function of miR-145 was initially related to the downregulation of the insulin receptor substrate-1 (IRS-1). Of note, in human colon cancer cells miR-145, by targeting the 3′ UTR of IRS-1, dramatically down-regulates the IRS-1 protein, thereby inducing cell growth arrest [105]. In colon cancer cell lines, YES and STAT1 factors were also evidenced as additional miR-145 targets [106]. Interestingly, in colon cancer tissues, low miR-145 expression level is inversely correlated to p70S6K1 protein levels. On the contrary, the forced expression of miR-145, by targeting p70S6K1, resulted in the downregulation of two downstream molecules of p70S6K1 pathway, the VEGFA and HIF-1α proteins, thus inhibiting tumor growth and angiogenesis [107].
The tumor suppressive function of miR-145 on cancer cell proliferation was also recently reported in lung cancer. By targeting the EGFR and nucleoside diphosphate linked moiety X-type motif 1 (NUDT1), miR-145 inhibits lung adenocarcinoma cell proliferation and lung tumorigenesis [108]. Of note, the prognostic value of miR-145 was originally evidenced in lung cancer, where the low expression miR-145 is significantly correlated with a worse prognosis and survival of adenocarcinoma patients [109].
In breast cancer tissues, miR-145 expression is inversely correlated with the stage of malignancy. In particular, miR-145, through the post-transcriptional regulation of N-RAS and VEGFA expression, exhibits inhibitory roles in tumor angiogenesis, invasion and tumor growth [110]. Accordingly, in the human breast cancer cell line MCF-7, miR-145 induction was shown to promote the inhibition of cell growth and the induction of apoptosis by targeting the Rho-effector rhotekin (RTKN) [111]. A death-promoting loop between miR-145 and TP53 was also identified in MCF7 breast cancer cells expressing wild-type TP53. miR-145 indeed activates the p53 pathway, resulting in the promotion of apoptosis, and sustaining in turn a further induction of miR-145 expression. In the same context, miR-145 may also down-regulate estrogen receptor-alpha (ER-α) protein expression, whereby a miR-145 re-expression therapy was proposed, at least for the subgroup of patients with ER-a-positive and/or TP53 wild-type tumors [112]. Interestingly, the re-expression of miR-143/miR-145 was also shown to suppress cellular growth and to support the apoptosis of epithelial cancer cells by enhancing p53 activity via MDM2 turnover [113].
The involvement of miR-145 in an integrated transcriptional regulatory circuit together with TFs and chromatin-modifying activities that support the growth and function of breast cancer stem-like cells (CSCs) recently emerged [114]. Accordingly, previous results evidenced that expression of miR-145 is low in self-renewing human embryonic stem cells (hESCs) but highly up-regulated during differentiation. Increased miR-145 expression inhibits hESC self-renewal, represses expression of pluripotency genes, such as OCT4, SOX2, and KLF4, and induces lineage-restricted differentiation [115].
In malignant pleural mesothelioma (MPM) cells, miR-145 has the potential to modulate many pro-tumorigenic features, including growth, clonogenicity and tumor engraftment in vivo. Interestingly, in MPM cells miR-145 targets directly OCT4 and, indirectly, the EMT-promoting target ZEB1. Of note, the levels of miR-145 and OCT4 are inversely correlated in vivo. Promoter hyper-methylation may contribute to the low levels of miR-145 in both MPM cells and in malignant MPM tissues. Importantly, miR-145 downregulation has been proposed to differentiate benign from malignant mesothelial tissues [116].
In pancreatic cancer cells, the activation of the KRAS signaling pathway consistently leads to the repression of the miR-143/miR-145 cluster and this is necessary to maintain the tumorigenic potential of these cancer cells. The downregulation of miR-143/miR-145 expression requires the RAS-responsive element-binding protein (RREB1), which represses the miR-143/miR-145 promoter. Both KRAS and RREB1 transcripts are direct targets of these miRNAs (see Table 5), demonstrating the existence of a feed-forward pathway that potentiates KRAS-mediated tumorigenesis [117].
More recently, the RREB1 protein was found to be overexpressed in colorectal adenocarcinoma tumors and cell lines, where the expression of the miR-143/miR-145 primary transcript is inversely related to that of RREB1. RREB1 negatively regulates expression of the miR-143/miR-145 cluster in a KRAS-dependent manner, thus establishing a complex network of regulation through which the miR-143/miR-145 cluster is able to modulate KRAS signaling [118]. In line with this, additional direct and indirect miR-143/miR-145 target genes have been reported, and they belong to the growth factor receptor-mitogen-activated protein kinase network, as well as to the p53 signaling pathway, further supporting a contribution of miR-143/miR-145 to the cell signaling pathways involved in colorectal tumorigenesis [119].
In prostate cancer miR-145 is consistently downregulated. A significant inverse correlation between the expression of miR-145 and that of the BNIP3 protein was observed in prostate cancer and in in benign prostate tissues. Accordingly, the overexpression of miR-145 in prostate cancer PC-3 and DU145 cells significantly down-regulated BNIP3, reduced cell growth, and increased cell death. As aforementioned for breast and colon cancer, the overexpression of wild-type p53 resulted in the up-regulation of miR-145 expression also in PC-3 cells, with consequent pro-apoptotic effects [120]. Wild-type p53 induces up-regulation of miR-145 expression and the inhibitory effects of wild-type p53 on migration, invasion, EMT and stemness of PC-3 cells were reversed by anti-miR-145. These results suggest that loss of wild-type p53 may promote bone metastasis of prostate cancer, at least partially through miR-145 downregulation, and resulting in increased EMT and stemness of cancer cells [121]. Moreover, the ectopic expression of miR-145 in LNCaP and DU145 cell lines led to a reduction in the expression of the ERG protein, suggesting that downregulation of miR-145 associated with prostate cancer may contribute to the increased expression of several ERG isoforms that are frequently observed in this tumor type [122] .
In conclusion, several lines of evidence indicate that miR-145 may be largely considered a miRNA with tumor suppressor activity, which is involved in the regulation of tumor growth, cell invasion and metastasis by targeting multiple cancer related genes, thus offering miR-145 as a novel therapeutic target for cancer therapy.
1.5 MiR-204: a key player in development and in tumor suppression
MiR-204 is encoded by the cancer associated genomic region (CAGR), the 9q21.1-q22.3 locus that exhibits high frequency of loss of heterozygosity in diverse human cancers [123]. MiR-204 is an intragenic miRNA, which is located within the transient receptor potential melastatin-3 (TRPM3) gene belonging to the family of transient receptor potential (TRP) channels [124].
MiR-204 activity is highly involved in vertebrate lens development and its loss is evidenced in different human cancers (Fig. 4 ). Banfi's group has shown that miR-204 is highly expressed in retinal pigment epithelium, lens, ciliary body and neural retina [125]. Its activity is required for the proper development of lens and optic cups [125]. The TF Meis2 is a main target of the developmental activity of miR-204. Aberrant expression of Meis2 released by miR-204 downregulation leads to lens abnormalities, microphthalamia, and eye coloboma [125]. Interestingly, miR-204 and its host gene TRMP3 are transcriptionally co-regulated by the developmental TF PAX6 [126]. The analysis of genes aberrantly expressed in PAX6 mutants during eye development revealed that miR-204 target genes are highly represented. This led to identification of novel developmental target genes of miR-204, such as Sox11, a member of the SoxC family of neuronal TFs, which plays a critical role in normal eye development. Intriguingly, PAX6 exerts its transcriptional program either by modulating directly the expression of its targets genes, or by indirectly modulating the expression of miR-204 that controls that of mRNA targets. This gives rise to a complex regulatory network that tunes and integrates coding and non-coding gene expression to pursue proper development.
Growing evidence indicates that miR-204 downregulation is a common alteration in different types of human tumors. MiRNA expression profiling of three different subsets of gastric cancer patients revealed that miR-204 expression was down-regulated in tumoral specimens when compared to matched peri-tumoral tissues [127]. TRPM3 gene loss was evidenced in a large fraction of the analysed gastric patients. Hence, this might be one of the molecular mechanisms underlying miR-204 downregulation in human cancers [127-129]. Notably, gastric cancer patients can be grouped according to the extent of miR-204 downregulation. Those characterized by a severe reduction (more than 0.5-fold) had the worst survival when compared to those with mild downregulation of miR-204 (less than 0.5-fold) [127]. The pivotal anti-apoptotic protein, BCL-2, was shown to be a target of miR-204. Reduced expression of miR-204 paired with increased staining of BCL-2 in gastric cancer patients [127]. BCL-2 ectopic expression counteracted the pro-apoptotic role of miR-204 in response to 5-Fluorouracil [127]. Downregulation of miR-204 in gastric cancers was also associated with increased expression of the class III histone deacetylase SIRT1. The targeting of SIRT1 by miR-204 ectopic expression favored mesenchymal to epithelial transition (MET) phenotypes and suppressed anoikis resistance of gastric cancer cells [130]. Notably, network modelling performed in head and neck tumors by the combination of gene expression data with inheritable cancer traits and risk factor loci uncovered 18 targets of miR-204 that are mainly involved in the development of metastasis [131]. Interestingly, ectopic expression of miR-204 not only reduced the levels of its target genes but also resulted in the inhibition of metastatic phenotype of head and neck cancer cell lines [131]. Altogether the localization of miR-204 linked the 9q21.1-q22.3 CAGR locus, a very well known risk factor for head and neck tumors, and the identified target genes, thus accounting for a bona-fide tumor suppressor micro-RNA. Downregulation of miR-204-5p was also evidenced in endometrial carcinomas (EC), when compared to normal endometrium [132]. This might result from the aberrant expression of the neurotrophic receptor tyrosine kinase B (TrkB), which is a target of miR-204 [132]. Ectopic expression of TrkB led to the accumulation of phospho-STAT3 that was recruited to a specific binding site of the miR-204's host gene, TRMP3, and might account for miR-204 downregulation in EC [132]. Loss of miR-204 expression through promoter methylation was evidenced in both glioma and neural stem cells [133]. This led to enhanced migration of glioma cells and to the acquisition of a stem cell-like phenotype. Indeed, attenuation of promoter methylation increased miR-204 expression in glioma cells and ectopic expression of miR-204 suppressed the tumorigenic potential of glioma cells. Mechanistic investigation revealed that miR-204 targets the expression of SOX4, a stemness TF, and EphB2, a receptor that promotes migration [133].
MiR-204 is a transcriptional target of the von Hippel-Lindau tumor suppressor gene (VHL) that is lost in the largest fraction of clear cell renal cell carcinomas (ccRCC) [134]. VHL-induced expression of miR-204 resulted in increased expression of short transcripts of TRMP3, thus indicating that miR-204 is not, at least in ccRCC, transcriptionally co-regulated with the large transcript encoding the full-length TRMP3 protein [134]. MiR-204 expression is clearly reduced in ccRCC with known VHL status when compared to matched normal kidney samples and its reconstitution led to growth inhibition in vitro and in vivo [134]. The transcriptional tumor suppressor axis VHL/miR-204 inhibited macroautophagy. This occurs through the ability of miR-204 to directly target LC3B and VHL-induced expression of the parolog, LC3C, which caused growth suppression.
It appears increasingly clear, that both in development and in cancer the downregulation of miR-204 disables coordinated transcriptional networks and instigates unscheduled signaling pathways (see Table 6 ). These might contribute to developmental diseases or to aberrant proliferation, metastasis and poor response to conventional anticancer treatments. The identification of additional miR-204 targets and the dissection of the epigenetic events regulating its expression in normal and in malignant tissues might provide intriguing insights, which would make miR-204 an appealing and hopefully druggable target.
| Tissue | Target | Function | References |
|---|---|---|---|
| Eye | Meis2Sox11 | Lens and retina development | [126, 125] |
| Gastric cancer | BCL-2SIRT-1 | Apoptosis MET | [127, 130] |
| Head and neck cancer | Multiple targets | Inhibition of metastatization | [131] |
| Endometrial carcinoma | TrKB | Inhibition of clonogenic growth, migration and invasion | [132] |
| Glioma | Sox4EphB2 | Inhibition of stem-cell likephenotype and migration | [133] |
| Renal cell carcinoma | LC3B | Inhibition of macroautophagy | [134] |
| Pancreatic cancer | Mcl-1 | Apoptosis | [142] |
| Breast cancer | IL-11 | Inhibition of metastatization | [143] |
1.6 Perspectives and open questions
Collectively, the reviewed groups of tumor suppressor miRNAs are emerging as major players of the cellular response to different types of oncogenic insults. Together with coding RNAs, epigenetic modifications and other mechanisms, this response can lead to clonal expansion, migration and invasion, as well as chemoresistance of a given tumor. Along with ever improving understanding of the concerted alterations in miRNAs, many cardinal questions remain open. For example, it is logical that tumor suppressor miRNAs maintain complex crosstalks with protein coding tumor suppressor RNAs in order to fortify barriers to oncogenicity, but the details of this interplay are currently unknown. Also unknown are the molecular determinants underlying activation of intragenic and intergenic tumor suppressor miRNAs upon oncogenic insults. The multiplicity of RNA targets of each tumor suppressor miRNA poses a pivotal challenge. Moreover, it is conceivable that the tumor cell context influences target specificity, and consequently determines biological effects. These and additional issues will require deeper, perhaps more integrated understanding of tumor suppressor miRNAs.
Conflicts of interest
The authors declare that they have no conflict of interest.
Acknowledgements
Contribution of Associazione Italiana per la Ricerca sul Cancro-Rome Oncogenomic Center to G.B., of Epigen to G.B., and of FIRB (Investment Fund for Basic Research) to G.B. was greatly appreciated. Yarden's research is supported by the National Cancer Institute, the German-Israeli Project Cooperation (DIP), the Israel Cancer Research Fund and the Dr. Miriam and Sheldon G. Adelson Medical Research Foundation.
