A step-by-step microRNA guide to cancer development and metastasis
- 1.1k Downloads
Cancer is one of the leading causes of mortality. The neoplastic transformation of normal cells to cancer cells is caused by a progressive accumulation of genetic and epigenetic alterations in oncogenes, tumor suppressor genes and epigenetic regulators, providing cells with new properties, collectively known as the hallmarks of cancer. During the process of neoplastic transformation cells progressively acquire novel characteristics such as unlimited growth potential, increased motility and the ability to migrate and invade adjacent tissues, the ability to spread from the tumor of origin to distant sites, and increased resistance to various types of stresses, mostly attributed to the activation of genetic stress-response programs. Accumulating evidence indicates a crucial role of microRNAs (miRNAs or miRs) in the initiation and progression of cancer, acting either as oncogenes (oncomirs) or as tumor suppressors via several molecular mechanisms. MiRNAs comprise a class of small ~22 bp long noncoding RNAs that play a key role in the regulation of gene expression at the post-transcriptional level, acting as negative regulators of mRNA translation and/or stability. MiRNAs are involved in the regulation of a variety of biological processes including cell cycle progression, DNA damage responses and apoptosis, epithelial-to-mesenchymal cell transitions, cell motility and stemness through complex and interactive transcription factor-miRNA regulatory networks.
The impact and the dynamic potential of miRNAs with oncogenic or tumor suppressor properties in each stage of the multistep process of tumorigenesis, and in the adaptation of cancer cells to stress, are discussed. We propose that the balance between oncogenic versus tumor suppressive miRNAs acting within transcription factor-miRNA regulatory networks, influences both the multistage process of neoplastic transformation, whereby normal cells become cancerous, and their stress responses. The role of specific tumor-derived exosomes containing miRNAs and their use as biomarkers in diagnosis and prognosis, and as therapeutic targets, are also discussed.
KeywordsMicroRNAs (miRNAs or miRs) Stages of tumorigenesis Cancer Transcription factor-miRNA regulatory networks Epithelial-to mesenchymal cell transition (EMT) Exosomes Biomarkers
Carcinogenesis is a multistep process by which normal cells progressively acquire genetic and epigenetic alterations to become malignant, invasive and metastatic cancer cells. These alterations provide a cell with the ability to bypass cellular senescence and to replicate indefinitely, thereby producing a tumor mass. Cancer cells, apart from immortality, which is a property characteristic of embryonic stem cells, also acquire a number of features that enable them to develop a primary tumor and later to migrate, invade adjacent tissues, enter the circulation and disseminate to distant organs to establish metastases. The sum of these cancer cell features has been conceptualized and is now widely accepted as the hallmarks of cancer [1, 2]. Cancers are among the leading causes of human mortality worldwide, with ~14 million new cases and > 8 million deaths annually, while the expected numbers of new cases for the next 20 years have been estimated to rise up to 70% . Carcinogenesis involves the accumulation of oncogenic driver mutations and inactivating mutations in tumor suppressor genes along with epigenetic changes leading to major alterations in gene expression and cell physiology . Increasing evidence over the past decade has shown that miRNAs play crucial roles in the regulation of cellular senescence and in different stages of the multistep process of cancer initiation and progression [4, 5, 6, 7, 8, 9, 10].
MiRNAs constitute a class of small, endogenous, noncoding RNAs (ncRNAs) of ~22 nucleotides in length. MiRNAs are transcribed by RNA polymerases (Pol) II and III, generating primary transcripts (pri-miRNAs). The pri-miRNAs contain one or more miRNAs and are 5′ capped and polyadenylated. A minor group of miRNAs that is associated with Alu repeats is transcribed by Pol III [11, 12, 13, 14, 15, 16, 17]. Pri-miRNAs are processed in the nucleus by a microprocessor complex composed of two proteins, the double-stranded RNA-specific RNase Drosha and the RNA-binding protein DGCR8 (DiGeorge Syndrome Critical Region 8), thereby generating a precursor miRNA (pre-miRNA), which is further cleaved into a mature miRNA by the RNAse III enzyme Dicer .
MiRNAs regulate gene expression at the posttranscriptional level, acting as negative regulators of mRNA translation and/or stability by binding to complementary sequences in the 3′ untranslated region (3′ UTR) of their target mRNAs. Since most target sites on a mRNA show only partial base complementarity with their corresponding miRNAs, individual miRNAs may target a multitude of different mRNAs to inhibit their translation into polypeptides. In cases of perfect complementarity, cleavage of the target mRNA is induced [11, 13, 14, 15, 19, 20]. Moreover, individual mRNAs may contain multiple binding sites for different miRNAs, resulting in a complex regulatory network . Whereas some miRNAs regulate specific individual targets, others can function as master regulators. Thus, key miRNAs may simultaneously regulate the expression levels of hundreds of genes, or cooperatively regulate their mRNA targets [22, 23, 24].
MiRNAs play important roles in regulating a variety of normal biological processes such as cell proliferation, differentiation and death [13, 21, 23, 25, 26]. MiRNAs also play crucial roles in oncogenesis by regulating cell proliferation, apoptosis, growth behaviour, metabolism and stress responses, acting as either oncogenes (oncomirs) or tumor suppressors [4, 5, 6, 7, 10, 21, 24, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36]. Hence, miRNAs may affect the initiation, development and/or progression of cancer through alteration of the expression levels of their target genes [4, 5, 7, 8, 34, 35, 36, 37]. Over-expression, amplification, or loss of epigenetic silencing of a gene encoding an oncogenic miRNA, which targets one or more tumor suppressor genes, could inhibit the activity of an anti-oncogenic pathway. In contrast, deletion, subtle mutations or epigenetic silencing of tumor suppressive miRNAs that normally repress the expression of one or more oncogenes might lead to increased oncogenic activity [6, 7, 21, 25, 38, 39, 40]. Given the emerging oncogenic and tumor suppressive roles of miRNAs in cancer initiation, development and progression, manipulating miRNA expression levels may be a potential therapeutic strategy against cancer [6, 21, 39].
The interacting transcription factor - miRNA networks, p53 - miRNA and NF-κB - miRNA. Examples of oncogenic or tumor suppressive miRNAs of either the p53 - miRNA or the NF-κB - miRNA networks are presented. Relevant references can be found in the text. Symbol (+) denotes miRNA up-regulation while (−) denotes miRNA down-regulation by p53 or NF-κB, in their respective rows
They are induced by p53. They promote apoptosis or growth arrest by post-transcriptional repression of target genes, including c-MYC, MYCN, CCND1, CCNE2, E2F3, CDK4, CDK6, MET, BCL-2, NOTCH1, HMGA2, SIRT1 in a context-dependent manner. MiR-34 represses the expression of SIRT1 which is a negative regulator of p53 through acetylation, leading to the induction of p53 and miR-34 thereby creating a feedback loop.
miR-34a, b, c
Snail1, is an NF-κB regulated transcription factor and a potent EMT inducer which is involved in a reciprocal feedback loop with miR-34: Snail1 inhibits the transcription of miR-34 family members, whereas miR-34 inhibits the translation of Snail1. In addition Snail1 was shown to repress p53, through the formation of a tri-molecular complex Snail1/HDAC1/p53, which deacetylates active p53 to promote its proteasomal degradation.
A family of miRNAs induced by p53. It represses the mesenchymal transcriptional regulator Zeb1, and inhibits metastasis
A family of miRNAs induced by p53. It targets MDM2
It is induced by p53. It regulates cell cycle by down-regulating G0/G1 proteins.. It targets CCND1, CCNE, CDK6, CDC27, CGI-38, CYCE, DMTF1, HMGA2, MCL1, MYB, NGN2, VEGF, WNT3A, BCL2, CAPRIN1 and CARD10.
They act as inhibitors of proliferation, metastasis and invasion. The tumor-suppressive miR-143 secreted by non-tumorigenic prostate cells inhibits the growth exclusively of prostate cancer cells in vitro and in vivo. They target MYC, ERK5, HOXA9, KRAS, PARP8.
It is induced by p53. It is also induced by pro-inflammatory cytokines such as IL-6 and IL-8 following replicative senescence or OIS, but prevents excessive SASP by targeting IRAK1 mRNA, via a negative feedback loop that restrains IL1R-IRAK1 signaling and impairs NF-κB activity. Hence, miR-146a induced by p53 suppresses canonical NF-κB signaling via a negative feedback loop.
MiR-10b is induced by Twist1, an NF-κB regulated gene, by binding to an E-box sequence that is located within the putative miR-10b gene promoter. Hence, the NF-κB-dependent Twist1-induction of miR-10b may be another mechanism by which NF-κB contributes to metastasis.
NF-κB regulated oncogenic miRNA implicated in several cancer types that promotes cell proliferation, angiogenesis, migration and metastasis. MiR-21 may function as a negative feedback regulator of the NF-κB signaling pathway leading to suppression of inflammation; Targets BCL2, MASPIN, PDCD4, PTEN, TPM1, RECK, RASA1, EGFR, HER2/neu, TIMP3
MiR-155 is induced by NF-κB and it acts as a negative feedback regulator of NF-κB signaling, inflammation and immunity. MiR-155 may also control the expression of both IKKβ and IKKε which either leads to repression or at least the limitation of NF-κB activation, constituting a negative feedback loop. It is also induced by mutant but not wild-type p53 and targets ZNF652, a novel zinc-finger DNA-binding transcription repressor of key drivers of invasion and metastasis. Hence, miR-155 is regulated by both NF-κB and mutant p53. MiR-155 promotes EMT and metastasis and also influences the tumor microenvironment; Targets AGTR1, AID, IKBKE, TP53INP1, SOCS, SHIP1 and CEBPB.
Mir-301a is a member of the miR-301a/miR-454 oncogenic miRNA cluster. MiR-301a enhances NF-κB activity by down-regulating the NF-κB repressing factor (NKRF) and functioning in a positive feedback loop; Targets the Hippo signaling, PTEN, Tumor protein-53 induced nuclear protein 1 (TP53INP1).
Member of the miR-301 family which is an NF-κB regulated miRNA that, at least in bladder cancer, sustains NF-κB activation by decreasing the expression of CYLD. MiR-130b repressed by hot spot p53 mutants in endometrial cancer. It functions as oncomir or tumor suppressive miRNA in a cell context dependent manner; Targets Hippo signaling, PTEN, TP53INP1.
NF-κB induced oncogenic miRNAs which down-regulates several tumor suppressors such as the Cip/Kip family members of cyclin-dependent kinase inhibitors p27 and p57, and PTEN to promotes cancer cell proliferation. MiR-221/222 impairs TRAIL-dependent response. Targets c-KIT, P27, CDKN1B, P57, CDKN1C, ESR1.
It is regulated by NF-κB, and it exerts cell context-dependent effects. It-acts upstream of p53 as a p53-repressive miRNA. MiR-125b represses p53 by binding to the 3′ UTR of p53 mRNA, and directly represses novel targets in the p53 network including the apoptosis regulators Bak1, Puma and Tp53INP1, showing the anti-apoptotic role of miR-125b in oncogenesis. However, it is down-regulated by DNA damage-induced by genotoxic drugs allowing p53 to take part in the DNA damage response. It target ERBB2, ERBB3, LIN28, LIN41 and TNFSF4.
NF-κB regulated family of miRNAs that promote cell proliferation, oncogenic transformation and evasion of apoptosis. It is repressed in senescent cells. It facilitates cell cycle progression at the G1/S transition. It targets AIB1, AML1, BIM1, CTGF, CDKN1A, E2F1, E2F2, E2F3, N-MYC, HIF-1A, PTEN, TGFBR2, TSP1 and Rb2/P130.
It is regulated by NF-κB and it acts as a metastasis-promoting miRNA which also negatively regulates NF-κB, forming a negative feedback loop. It targets E-cadherin (CDH1) and promotes EMT, cell motility and invasiveness.
NF-κB regulated miRNA. IL-6 activates STAT3 which induces miR-21 and miR-181b-1, leading to inhibition of PTEN and CYLD tumor suppressors, respectively. This inhibition also leads to NF-κB activation. Hence, it forms a positive feedback loop which sustains NF-κB activation and links inflammation to transformation and metastasis. It contributes to the acquisition of different hallmarks of cancer. MiR-181a/b overexpression has also been linked to aberrant activation of major pathways including IL6/Stat3, TGFβ (by targeting TIMP3) and WNT/β-catenin, suggesting that it may represent a crucial mechanism fuelling neoplastic transformation. It targets TIMP3, CYLD, PTEN and p27Kip1. MiR-181a/b negatively regulates the expression and activity of ATM and dampens the DNA damage response in breast cancer cells which are rendered sensitive to treatment with PARP1 inhibitors.
It is a tumor suppressive miRNA regulated by NF-κB (responsive to IKK2 inhibitors). It targets Lin-28, Lin-41, ERBB2 and ERBB3.
A family of tumor suppressor miRNAs which is negatively regulated by NF-κB. They act as repressors of cell growth. A SNP in K-RAS 3′ UTR (let-7a binding site) increases NSCLC risk (cancer predisposition); Targets: CCND1, CDC25a, CDC34, CDK6, CRDBP, DICER, HMGA2, HOXA9, IMP-1, ITGB3, MYC,RAS, TLR4 and E2F2.
IKK-regulated miRNAs that inhibit NF-κB signaling
It inhibits IKKβmRNA translation in ovarian cancer cells
It is regulated by NF-κB and it inhibits IKKβ mRNA translation (forming a negative feedback loop) leading to inhibition of cell proliferation, migration and invasion in vitro.
MiR-15a, miR16 miR-223
They act as negative regulators of IKKα. They also act as inhibitors of non-canonical NF-κB pathway during macrophage differentiation.
MiRNAs with dual roles, acting as tumor suppressors and tumor promoters. Some miRNAs are listed with a dual role in cancer initiation and progression, having both tumor suppressive and tumor promoting functions, in a cell context-dependent manner
Tumor suppressor function
Tumor promoting function
It promotes senescence of lung HDFs  and EC cells , and it inhibits angiogenesis in response to SCF by targeting the SCF receptor c-kit, and it suppresses endothelial cell migration and proliferation in vitro .
It is involved in the initiation of the malignant process, in a context-specific manner. MiR-221/222 are overexpressed in many human cancers, acting as oncogenes via down-regulation of many tumor suppressors such as p27, p57 and PTEN [71, 72, 73, 74, 75, 266, 267].
In normal FTSE cells, miR-182 overexpression triggers cellular senescence via p53-mediated up-regulation of p21 .
In p53-compromised FTSE cells, miR-182 overexpression functions as an oncomir .
It is up-regulated during replicative and DNA damage-induced senescence of skin HCA2 and BJ HDFs, but not of lung IMR-90 and WI-38 HDFs .
Decrease miR-146a levels result in the de-repression of NOX4, increased ROS generation and DNA damage-induced senescence . Thus, miR-146 may act as an oncomir by suppressing SASP via a negative feedback loop.
It inhibits ErbB2-induced malignant transformation of human breast epithelial cells. Trastuzumab treatment results in the up-regulation of miR-155 and in a marked reduction of ErbB2 expression in ErbB2-positive breast cancer cells .
It targets Zeb-1, and inhibits Zeb1-dependent EMT and cancer cell invasion .
Epigenetic silencing of miR-130b contributes to the development of endometrial  and ovarian  cancers. It also acts as a tumor suppressor in prostate cancer by down-regulating MMP2 , suggesting a cell context-specific tumor suppressive action.
It acts as oncomir in
1. bladder cancer, by decreasing the expression of cylindromatosis (CYLD), an endogenous inhibitor of NF-κB activation 
2. glioblastoma by inhibiting Hippo signaling 
3. esophageal squamous cell carcinoma by inhibiting PTEN  and
4. liver cancer, by directly targeting TP53INP1, promoting self-renewal of CD133+ cancer stem cells and conferring chemoresistance 
1. In cutaneous squamous cell carcinoma, it down-regulates MMP13 and inhibits cancer cell proliferation, migration and invasion 
2. In ovarian cancer, it targets Bcl-3 and suppresses cell proliferation 
3. In bladder cancer cells it targets E2F3 and suppress colony formation in vitro and tumor development in nude mice 
4. In liver cancer it targets oncogenic Lin28B2 and suppresses cancer cell proliferation and metastasis 
5. In melanoma, it targets c-jun and inhibits cell proliferation 
It is a NF-κB-driven miRNA  acting upstream of p53 as a p53-repressive miRNA.
6. In osteosarcoma, it down-regulates STAT3 and suppresses cancer cell proliferation and migration 
8. In AML, it targets canonical NF-κB signaling and inhibits cell proliferation and invasion by inducing a G2/M cell cycle phase arrest and apoptosis 
It can promote metastasis, depending on the cell model used .
It induces MET, leading to colonization in distant sites .
It acts as a tumor suppressor by inducing cell cycle arrest, and by promoting apoptosis .
It regulates angiogenesis, promotes invasion and metastasis under hypoxic conditions .
2 Aspects of miRNA biogenesis
MiRNA biogenesis and maturation is a two-step process that takes place in the nucleus and subsequently the cytoplasm, both involving cleavage events catalyzed by two ribonuclease III endonucleases, Drosha and Dicer, respectively.
The miRNA genes are located both in intergenic (exonic or intergenic miRNAs) and intronic (intronic miRNAs) regions of the genome. In humans, the majority of canonical miRNAs is encoded by introns of non-coding or coding genes, whereas some miRNAs are encoded by exonic regions. Intronic miRNAs that are located within a host gene in the same orientation are transcribed along with the primary transcript, driven by the same promoter. In contrast, exonic miRNAs are thought to rely on their own promoters. The transcription of a specific miRNA gene may thus be regulated either by its own specific promoter or by a promoter from a neighbouring gene. MiRNA genes are transcribed by RNA Pol II and III, the majority being transcribed by RNA Pol II, to produce a pri-miRNA which is processed to generate pre-miRNAs that undergo a series of cleavage events to form mature miRNAs. MiRNAs may be transcribed as individual genes or as polycistronic clusters (~30% of miRNAs). These clusters encompass several miRNA genes located adjacent to each other on the chromosome, which are transcribed as one long pri-miRNA transcript and subsequently processed into various pre-miRNAs [11, 12, 13, 14, 15, 51].
Intergenic (exonic) miRNAs are transcribed by RNA Pol II or III generating a pri-miRNA molecule. The initiation step (cropping) is mediated by a microprocessor complex in the nucleus, which is comprised of the RNase III endonuclease Drosha and the double-stranded RNA binding protein DGCR8. This microprocessor complex recognizes the distinct secondary hairpin structure of the pri-miRNA and specifically cleaves both strands at 11 base pairs (bp) from the base of the stem loop, releasing a 60–70-nucleotide (nt) pre-miRNA intermediate. This pre-miRNA has a short stem which is recognized by the nuclear export factor exportin 5 (Exp5 or XPO5)-Ran-GTP, and is exported from the nucleus to the cytoplasm where RNase III Dicer catalyses the second processing step to cleave off the loop of the pre-miRNA in order to produce miRNA duplexes. These miRNA:miRNA* duplexes, which encompass 22-nt strands from each arm of the original hairpin, associate with Argonaute proteins (Ago2) such that the miRNA strand usually becomes stably incorporated while the miRNA* strand dissociates and is degraded. The resulting leading functional strand of the mature miRNA is loaded together with the Ago2 proteins into the RNA-induced silencing complex (RISC), which is comprised of Ago2, TARBP2, PACT and Dicer, where it guides RISC to silence target mRNAs through mRNA cleavage, translational repression or deadenylation.
Intronic miRNAs are transcribed as part of a pre-mRNA by RNA Pol II. Canonical intronic miRNAs are processed co-transcriptionally before splicing. For the splicing of miRNA-containing introns, the splicing complex and Drosha, which cleaves the miRNA hairpin, are required. The pre-miRNA enters the miRNA biogenesis pathway, whereas the rest of the transcript undergoes pre-mRNA splicing and produces mature mRNA for protein synthesis. Non-canonical intronic small RNAs, called mirtrons, are produced from spliced introns after debranching and, due to their similarity to pre-miRNAs, bypass the Drosha-processing step [11, 12, 13, 14, 15].
Most mammalian miRNA genes have multiple isoforms (paralogs), probably as a result of gene duplications, which often have identical six nucleotide sequences at nucleotide positions 2–7 relative to the 5′ end of the miRNA, known as seed sequence, which is crucial for base pairing with the target mRNA (target specificity). Hence, a functional redundancy of paralogs is manifested. The processing of a pre-miRNA into mature miRNAs results in at least two mature 19–23 nt long miRNAs, named miR-X-5p originating from the 5′-end and miR-X-3p originating from the 3′-end of the pre-miRNA. The mature miRNA base-pairs with its target mRNA to direct gene silencing via mRNA cleavage or translation repression, depending on the level of complementarity between the miRNA and its mRNA target [11, 12, 13, 14, 15, 51].
3 Impact of miRNAs in cancer
The roles of ncRNAs in tumorigenesis have been studied most widely for miRNAs, which may act as oncogenes or tumor suppressors by targeting one or more different mRNAs. The genes encoding miRNAs may also be mutated or epigenetically altered, as well as suppressed or activated by transcription factors leading to changes in their expression [36, 37, 52].
Tumor-specific genetic alterations in the miRNA-processing machinery, such as in the genes encoding TARBP2 , DICER1  and XPO5/Exp5 , have strongly highlighted the relevance of the miRNA biogenesis pathways in cellular transformation, and miRNA deregulation in cancer . Exosomes secreted by cancer cells contain proteins and nucleic acids, including miRNAs. Breast cancer-secreted exosomes contain e.g. miRNAs associated with the RISC-loading complex along with Dicer, AGO2, and TRABP2, and display a cell-independent capacity to process pre-miRNAs into mature miRNAs. Importantly, exosomes derived from cells and sera of breast cancer patients facilitate the tumorigenic conversion of non-tumorigenic epithelial cells in a Dicer-dependent manner . The miRNA expression profile of human tumors is characterized by a general defect in miRNA production, which results in global miRNA down-regulation [7, 21, 37, 57].
MiRNAs have been classified with regard to their role in cancer as ‘the Good, the Bad and the Ugly’ . The ‘Good’ miRNAs are innocent bystanders in the oncogenic transformation process, whose expression profile may be used for cancer diagnosis or prognosis. The ‘Bad’ miRNAs are causally linked to tumorigenesis and directly modify tumor suppressor or oncogenic pathways. The ‘Ugly’ miRNAs, whose inappropriate loss or gain destabilizes the cellular identity of a tumor, indirectly result in enhanced phenotypic variability and tumor progression .
Several studies have revealed miRNA mutations leading to cancer predisposition [36, 52, 58]. For example, a single-nucleotide polymorphism (SNP) in miR-146 has been found to contribute to thyroid cancer , and a SNP in let-7, specifically in a site critical to Ras binding, has been found to increase more than two-fold the risk to develop non-small cell lung cancer (NSCLC) . MiR-15 and miR-16 deregulation, which is observed in most B-cell chronic lymphocytic leukemias, results from a chromosome 13q14 deletion . Deregulation of miRNA expression in cancer may also be due to epigenetic modifications such as DNA methylation . Epigenetic deregulation through DNA hypermethylation of miR-34a has been found to be associated with p53-dependent cancer predisposition , while epigenetic alterations in miR-192 expression have been found to contribute to the progression of pancreatic cancer . It has also been found that alterations in the epigenetic regulation of the miR-200 family may be involved in the epithelial-to-mesenchymal cell transition (EMT) of cancer cells. CpG island hypermethylation silences these miRNAs in human tumors, causing an up-regulation of the zinc finger E-box-binding homeobox (HOX) 1 (Zeb1) and 2 (Zeb2) transcriptional repressors. These, in turn, suppress the expression of E-cadherin (CDH1) and promote EMT . Additionally, MiR-127, which targets the proto-oncogene BCL-6, has been found to be silenced in several cancer cell types . Epigenetically regulated miRNAs have also been identified by profiling DNMT1- and DNMT3b-deficient colorectal cancer cells compared to normal colonic cells. Interestingly, it was found that miR-124a, embedded in a large CpG island, was the only unmethylated miRNA in normal cells but hypermethylated and silenced in tumors. MiR-124a targets cyclin D kinase 6 (CDK6), which mediates phosphorylation of the RB tumor suppressor protein. The epigenetic silencing of miR-124a thus leads to CDK6-mediated phosphorylation and inactivation of pRB . In keeping with this, it has been shown that transient inhibition of HNF4α initiates hepatocellular transformation through a miRNA inflammatory feedback loop circuit containing miR-124. HNF4α binds strongly to the miR-124 promoter in liver cancer cells and inhibits its expression. It has been reported that systemic administration of miR-124, which modulates inflammatory signaling, prevents and suppresses hepatocellular carcinogenesis by inducing tumor-specific apoptosis without toxic side effects . Thus, miR-124 may act as a tumor suppressor in liver cancer.
The tumor suppressive miR-128a, which acts by targeting the Bmi-1 oncogene, has been found to be significantly down-regulated in medulloblastoma . Down-regulation of Bim-1 by miR-494, which is regulated by ERK1/2, confers resistance to TRAIL-induced apoptosis of NSCLC cells . MiR-221 has been found to be up-regulated in papillary thyroid tumors, while thyroid tissues from normal individuals exhibited lower levels of this miRNA . Similar findings were reported in normal and cancerous breast tissues [71, 72, 73]. Specific miRNAs, such as miR-221, may be involved in cancer initiation, in a context-specific manner, or in specific cancer progression stages [74, 75, 76]. Context-specific effects have e.g. been reported for miR-182-5p, which has been found to function as an oncogene in breast, ovarian and bladder cancer, but as a tumor suppressor in lung cancer. Similarly, miR-17-5p acts an oncogenic miRNA in hepatocellular and colorectal carcinomas, whereas it behaves as a tumor suppressor in cervical cancer. It has also been reported that MiR-200 can either promote or inhibit metastasis, depending on the cell model used . These context-dependent effects may be explained by the differential expression of target genes between tissues and to the involvement of miRNAs in complex interacting regulatory networks.
It has been proposed that alterations in miRNAs play critical roles in the pathophysiology of all human cancers . Many oncogenic miRNAs that have been found to be overexpressed in cancer cells are, at least in part, responsible for sustaining a high proliferation rate and for evading apoptosis through epigenetic deregulation [21, 38, 40]. A classic paradigm is miR-21, which has oncogenic properties in many types of cancer. MiR-21 has been found to be overexpressed in hepatocellular carcinoma (HCC) cells targeting the phosphatase and tensin homologue (PTEN) tumor suppressor. It has been reported that inhibition of miR-21 in HCC cells results in decreased proliferation through up-regulation of PTEN . It has also been shown that miR-21 expression may be increased and predict a poor survival in NSCLC. Overexpression of miR-21 has been found to enhance K-Ras-induced NSCLC tumorigenesis through inhibition of negative regulators of the Ras-MEK/ERK pathway and inhibition of apoptosis, while genetic deletion of miR-21 was found to partially protect against tumor formation .
MiRNA expression profiles differ between tumors and the corresponding normal tissues of origin, as also between tumor types [24, 31, 32, 33, 35, 36]. By profiling the miR transcriptome, miRNA expression signatures have been identified that are associated with specific cancer developmental steps, from the normal to the metastatic stage, and with the acquisition of cancer-specific hallmarks in a prototypical mouse model of cancer . In this model, down-regulation of members of the miR-200 family in metastases was found to relieve repression of the mesenchymal transcription factor Zeb1, which suppresses CDH1 . MiRNA expression profiles have also been shown to change between immortal and dysplastic oral cell lines [81, 82], during cervical cancer progression from normal tissue to squamous cell carcinoma, involving ~70 miRNAs , and during prostate cancer progression .
Several studies have shown that miRNAs may concurrently target multiple effectors in common pathways involved in cell proliferation, survival and differentiation [52, 58, 85]. There are several examples of cancer pathways that are regulated at multiple points by miRNAs, including protein kinase and transcription factor signaling pathways, cell cycle and apoptosis regulatory pathways, and cytoskeleton assembly and dynamics regulating pathways . It has, for example, been shown that the miR-16 family negatively regulates cellular growth and cell cycle progression by down-regulating transcripts that function cooperatively to control G0/G1 cell cycle phase progression . There are also several examples of co-regulation of common biological processes by multiple miRNAs, particularly by those from a miRNA family or a cluster, such as the miR-17-92 cluster .
4 Cancer initiation: acquisition of somatic mutations in the evolution of cancer
Defining the mechanisms whereby miRNAs contribute to cancer development requires a thorough understanding of the types of mutations in specific protein coding genes that lead to cancer onset and progression. Mutations are not only the drivers of evolution, but also the initiators of malignant phenotypes. The acquisition and accumulation of somatic mutations, including base substitutions, insertions, deletions, gross DNA rearrangements and amplifications are known to lead to neoplastic transformation . Somatic mutations may be the result of unrepaired DNA damages induced by mutagens or by inappropriate DNA repair. Specific somatic mutations are selected for in vivo, resulting in cancer cell clones that possess diverse survival and proliferation advantages, as well as resistance to chemotherapy. Somatic mutations occurring in the first step of cancer development are cancer-initiating mutations, also known as ‘driver mutations’, which cause clonal expansion [88, 89, 90, 91, 92, 93, 94, 95].
Analyses of cancer genomes have led to the identification of several driver mutations, and provide strong evidence that they contribute to cancer development [96, 97] (http://www.sanger.ac.uk/genetics/CGP/Census/). The analysis of lung adenocarcinomas has, for example, led to the identification of a set of 26 high frequency dominantly acting mutant genes, including oncogenic K-Ras, EGFR and STK11 (LKB1), and inactivating mutations in the TP53 tumor suppressor gene [98, 99, 100, 101]. Other studies have revealed a intra-tumor heterogeneity of somatic mutations in NSCLC  and other cancer types . Comprehensive genomic characterization of squamous cell lung cancers (SqCC) revealed driver mutations in the CDKN2A and RB1 genes in 72% of the cases . Similarly, genomic analyses of small cell lung cancer (SCLC) have identified key somatic driver mutations, including activating SOX2 mutations and inactivating TP53 and RB1 mutations [104, 105, 106]. Somatic driver mutations have also been identified in colorectal cancer, including mutations in the APC, K-Ras and the phosphoinositide 3-kinase (PI3K) genes , in melanoma, including mutations in the B-RAF, TP53, PTEN and CDKN2A genes , and in several other types of cancer. In addition to driver mutations, which confer growth advantage and are positively selected during cancer evolution, other mutations known as ‘passengers’ do occur that are not selected for and do not confer clonal growth advantage. As such, they do not contribute to cancer development but may be transmitted along with driver mutations to daughter cancer cells [88, 89, 90, 91, 92].
Combinations of mutations with relatively weak tumor-promoting effects, termed ‘mini-driver mutations’, may substitute for a major driver mutation, especially in the presence of genomic instability or exposure to mutagens . Since these minidriver mutations provide a minute selective advantage to the cancer cell, they tend to accumulate as tumorigenesis proceeds . Somatic passenger mutations in cancers, including mutations in promoters, enhancers, repressors, insulators, miRNAs, long ncRNAs and untranslated regions, may constitute minidriver mutations [37, 40, 109, 110].
The acquisition of cancer-initiating driver mutations often triggers senescence, a state of stable cell cycle arrest, particularly in the case of oncogenic mutations [111, 112, 113, 114, 115, 116, 117]. Senescence has, for instance, been detected in premalignant, but not malignant, lung lesions in mice expressing mutant K-Ras  or B-RAF V600E , as also in human and mouse melanocytes and melanocytic naevi expressing mutant B-RAF V600E [119, 120]. Despite oncogene-induced senescence (OIS), which is believed to be a cancer initiating barrier [111, 114] and other tumor suppressive mechanisms, benign cancers may still develop into overt malignancies [113, 115, 118]. Therefore, the next step of a cell that has acquired an oncogenic mutation towards malignancy is to bypass cellular senescence.
5 On the road to immortality: miRNAs involved in senescence bypass to premalignancy
A cancer-associated miRNA may act as a bystander, a regulatory or a causative factor in tumor development. mRNA expression profiles may be altered by miRNAs that are up-regulated or down-regulated during replicative stress-induced (SIPS) or oncogene-induced (OIS) premature senescence. Most of these miRNAs are still not functionally characterized and may undergo altered regulation as a consequence of senescence. However, some miRNAs have been found to be associated with the promotion of senescence, while others have been found to suppress or bypass senescence and to induce the growth of various cell types of different origins via different mechanisms [7, 9, 71, 85, 124, 125, 126, 127], as in the case for the let-7 family of tumor suppressor miRNAs  or the miR-17-92 cluster of oncogenic miRNAs [124, 129], respectively. In addition to oncomirs, many miRNAs possess tumor suppressive activities by acting as inducers of senescence, referred to as senescence-associated miRNAs (SA-miRNAs or SA-miRs) [125, 127]. These SA-miRs may be induced through the classical p53 and pRB tumor suppressor pathways, or through epigenetic regulatory mechanisms. Cells acquiring oncogenic driver mutations undergo oncogene-induced replication stress leading to premature senescence that is established by de-repression of the INK4/ARF locus, which acts as a sensor to link stress with pRb (p16-pRB) and p53 (p14Arf-p53-p21) tumor suppressor networks [116, 117, 130, 131]. Hence, the tumor suppressor pathways themselves may regulate miRNA expression.
The miR-17-92 cluster, also referred to as oncomiR-1, represents a polycistron precursor transcript derived from the miR-17-92 gene and contains six tandem stem-loop hairpin structures that ultimately yield six mature miRNAs belonging to four seed families, i.e., the miR-17 family (miR-17 and miR-20a), miR-18 family (miR-18a), miR-19 family (miR-19a and miR-19b-1) and miR-92 family (miR-92a-1). MiR-18 exhibits a significant sequence homology with miR-17 and miR-20, despite one nucleotide difference within the respective seed regions. Ancient gene duplications have given rise to two miR-17-92 cluster paralogs in mammals, miR-106a-363 (miR-106a, miR-18b, miR-20b, miR-19b-2, miR-92a-2, miR-363) and miR-106b-25 (miR-106b, miR-93, miR-25), each of which only contains homologous miRNAs to a subset of the miR-17-92 components. The miR-17-92 gene cluster is frequently overexpressed in human cancers and has been shown to promote cell proliferation, to evade apoptosis and to induce oncogenic transformation [124, 129, 132, 133]. It has been shown that the miR-17-92 cluster, and in particular the miR-17/20a seed family, can block Ha-Ras V12-induced senescence of BJ and WI-38 normal human diploid fibroblasts (HDFs) by directly targeting p21Waf1/Cip1 (CDKN1A), whereas it failed to evade c-Myc-induced apoptosis. In addition, it has been found that miR-17/20a can cooperate with oncogenic E1A and Ha-Ras V12 to confer tumorigenicity in nude mice .
A genome-wide short interfering RNA (siRNA) screen and a miRNA screen were employed to identify mediators of OIS. Initially, it was shown that siRNA-mediated knockdown of p21Waf1/Cip1 could rescue human mammary epithelial cells (HMECs) from Ras G12V-induced senescence. A total of 28 miRNAs was found to prevent Ras G12V-induced growth arrest, among which all of the miR-106b family members, as well as several other miRNAs such as miR-130b (miR-301a family), members of the miR-302 family (miR-302a, miR-302b, miR-302c and miR-302d), miR-512-3p and miR-515-3p, with seed sequences very similar to that of the miR-106b family members. It was found that overexpression of all these miRNAs could rescue HMECs from Ras G12V-induced senescence by preventing up-regulation of p21Waf1/Cip1 .
In addition to OIS, miRNA microarray-based expression analyses, followed by quantitative real-time PCR (qRT-PCR) validation, revealed changes in miRNA expression profiles associated with SIPS in primary cultures of dermal HDFs and human trabecular meshwork (HTM) cells. Twenty-five miRNAs associated with SIPS were identified. Specifically, SIPS in both HTM and HDF cells was found to be associated with a significant down-regulation of four members of the miR-15 family and five members of the miR-106b family, including miR-106a-363 and miR-106b-25. SIPS was also found to be associated with the up-regulation of two miRNAs, miR-182 and miR-183, from the miR-183-96-182 cluster. Exogenous expression of miR-106a inhibited the up-regulation of p21Waf1/Cip1 associated with SIPS, while exogenous expression of a miR-106a antagomir led to increased p21Waf1/Cip1 expression in senescent cells . MiR-182 is considered to be a ‘ROSmiR’ as its expression can be up-regulated by excess intracellular reactive oxygen species (ROS) [136, 137]. While most ‘ROSmiRs’ are regulated by p53 and induce SIPS, miR-182 is regulated by β-catenin. It has been shown that in normal fallopian tube secretory epithelial (FTSE) cells miR-182 overexpression triggers cellular senescence via p53-mediated up-regulation of p21Waf1/Cip1. In contrast, in p53-compromised FTSE cells, miR-182 overexpression no longer enhances p21Waf1/Cip1 but functions as an oncomir . In addition, miR-182 has been found to down-regulate BRCA1, the expression of which is commonly decreased in sporadic breast tumors and correlates with a poor prognosis of the patients. The reduced expression of BRCA1 induced by miR-182 impairs homologous recombination-mediated repair, and renders cells hypersensitive to ionizing radiation and poly (ADP-ribose) polymerase 1 (PARP1) inhibitors, a phenotype similar to the BRCA1-deficient phenotype . Thus, miR-182 can function as a SA-miRNA or an oncomir in a cell context-dependent manner (Table 2).
A functional genetic screen led to the identification of two oncogenic miRNAs, miR-372 and miR-373 which, when expressed in BJ HDFs bearing oncogenic Ha-Ras and wild-type TP53, can bypass OIS. These miRNAs were shown to inhibit the p53 pathway by neutralizing p53-mediated CDK inhibition through direct inhibition of the tumor suppressor LATS2. Indeed, siRNA-mediated down-regulation of LATS2 was found to mimick to a certain extent the growth advantage that the miRNAs conferred to oncogenic Ras-expressing HDFs, but not as potently as the expression of miR-372 and miR-373 .
Another genetic screen led to the identification of oncogenic miR-378a-5p, in addition to several other miRNAs that were previously shown to play a role in senescence, as a negative regulator of BRAF oncogene-induced senescence of TIG3 HDFs. In this setting, miR-378a-5p was found to reduce the expression of several senescence markers, including p16INK4A and senescence-associated-β-galactosidase (SA-β-Gal). Subsequently, several miR-378a-5p targets were identified that may underlie the mechanism by which miR-378a-5p can delay OIS .
In addition, it was found that miR-34a and miR-146 are up-regulated during BRAF-induced senescence of TIG3 HDFs. MiR-34a acts as a tumor suppressor and as an inducer of senescence by down-regulating the E2F pathway [140, 141, 142]. The induction of miR-34a during OIS was found to be independent of p53 but, instead, to be mediated by the ETS family transcription factor ELK1, causing c-Myc suppression and inhibition of cell proliferation. However, in this cellular system inhibition of miR-34a alone did not prevent senescence, suggesting that miR-34a is an important, but not sole, player in BRAF-OIS in these cells . MiR-34a triggers senescence and elicits tumor suppression by direct targeting of the 3′ UTR of the NOTCH1, E2F1, CDK6 and SIRT1 mRNAs . Thus, hypermethylation resulting in silencing of the miR-34a locus was proposed to represent an oncogenic alteration in different cancer cell types, including lung, breast, colon and kidney cancer . Hence, epigenetic silencing of a miRNA gene may represent another mechanism to bypass senescence.
MiR-146a and miR-146b (miR-146a/b) were found to be up-regulated during replicative and DNA damage-induced senescence of skin HCA2 and BJ HDFs, but not of lung IMR-90 and WI-38 HDFs . In normal cells, senescence, a state of stable and irreversible cell cycle arrest, occurs after a finite number of divisions. It can be triggered by intrinsic and extrinsic stimuli including telomere shortening, and oxidative, genotoxic or oncogenic stresses, leading to induction of a DNA damage response (DDR), and the development of a senescence-associated secretion phenotype (SASP) characterized by inflammatory cytokine secretion such as IL-6 and IL-8 [117, 130, 131]. Induction of miR-146a/b has been detected in skin HDFs exhibiting a robust SASP phenotype, but not in lung HDFs. Ras-induced senescence of lung HDFs was, however, found to lead to an increase in IL-6 secretion and miR-146a/b expression, suggesting a link between miR-146a/b expression and the robustness of inflammatory cytokine secretion . Moreover, ectopic expression of miR-146a/b in HCA2 HDFs did not induce growth arrest or senescence, but suppressed IL-6 and IL-8 secretion by down-regulating IRAK1, a crucial downstream component of the IL-1 receptor (IL1R) signal transduction pathway, which regulates the expression of both inflammatory cytokines (IL-6 and IL-8). Indeed, it was found that IL-1α neutralizing antibodies can abolish both miR-146a/b expression and IL-6 secretion. Thus, by targeting the IRAK1 mRNA, miR-146a/b, which is induced in response to rising inflammatory cytokine levels following replicative senescence or OIS, a negative feedback loop can be created that restrains IL1R-IRAK1 signaling and impairs NF-κB activity [47, 146], thereby limiting senescence-associated inflammatory cytokine production and restraining excessive SASP activity [117, 145]. However, miR-146a, which appears to function in a negative feedback loop to suppress SASP, was found to be expressed at 5-fold higher levels in early passage telomerized BJ (BJ-hTERT) HDFs, relative to early passage wild-type cells. It was also found to be markedly up-regulated (10-fold higher) in late passage BJ-hTERT cells compared to senescent BJ HDFs, indicating that miR-146a could be regulated independently of senescence, and that its expression was elevated in cells with an extended lifespan . Indeed, microarray-based expression profiling identified miR-146a as a pro-proliferative miRNA in human umbilical vein endothelial cells. It prevented senescence by down-regulating the expression of the NAPDH oxidase subunit NOX4 via targeting its 3′ UTR. During replicative senescence the miR-146a levels were found to decrease resulting in a de-repression of NOX4 and, presumably, increased ROS generation leading to DNA damage-induced senescence . Thus, miR-146 induced by inflammatory cytokines may act as an oncomir by suppressing SASP via a negative feedback loop and ROS generation.
A group of 15 miRNAs was found to be down-regulated in senescent human WI-38 and MRC-5 HDFs harboring wild-type p53. Interestingly, these 15 miRNAs represent three clusters including the miR-106b/93/25 polycistron (a paralog of miR-19-92) residing within an intron of the cell-cycle gene ‘minichromosome maintenance protein 7’ (MCM7), miRs-17/18a/19a/20a/19b-1/92a-1 (miR-17-92 polycistron) and miRs-106a/18b/20b/19b-2/92–2 (miR-106a-92 polycistron, paralog of miR-19-92), which are clustered on chromosome X. Additional well-represented miRNAs in the clusters include miR-155 and the miR-15b/16 polycistron. These miRNAs were found to be transcriptionally activated by E2F, whereas p53 was found to exert its repression through E2F inhibition. These p53-repressed miRNAs were found to be down-regulated during senescence, in agreement with a reduced E2F activity in senescent cells . Similar to p53 inactivation, overexpression of representative miRNAs of this group were found to promote cell proliferation and to delay senescence. Taken together, these results suggest that miRNAs and transcription factors such as E2F1 and p53 may cooperate in complex transcriptional and post-transcriptional loops that control major cellular processes such as cell proliferation and apoptosis. In the current feed-forward loop, p53 acts to inhibit both E2F1 and a group of proliferation-inducing miRNAs, leading to inhibition of cell proliferation and senescence .
The p53 tumor suppressor pathway is induced during replicative senescence, SIPS and OIS, but also in response to DNA damage. The p53 system senses environmental stress and modulates miRNAs at the levels of transcription and processing [28, 149]. Upon DNA damage or oncogenic stress, p53 down-regulates the expression of oncomiR-17-92 and induces the transcription of miR-34a, miR-34b and miR-34c from distinct genomic loci. These miRNAs share and repress a number of target mRNAs to promote growth arrest and apoptosis [150, 151]. TP53 also induces the expression of several other miRNAs (miR-16, miR-103, miR-143, miR-145, miR-26a, and miR-206), causing a decrease in cellular proliferation rate. For instance, miR-16 targets the CCND and CCNE mRNAs . Interestingly, most cancer-associated p53 mutations are located in a domain that is required for both miRNA processing and transcriptional activity. Thus, loss of p53 transcription and miRNA processing may contribute to tumor progression [28, 149].
It has been shown in primary mouse and human fibroblasts that senescence epigenetically silences the histone demethylase KDM2B and induces the expression of the tumor suppressor miRNAs let-7b [128, 152, 153] and miR-101 [154, 155], which target the EZH2 histone H3K27 methyltransferase, the functional enzymatic component of the Polycomb Repressive Complex 2 (PRC2). Sustained expression of KDM2B can bypass cellular senescence in primary mouse embryonic fibroblasts (MEFs) and promote immortalization by silencing these miRNAs through locus-specific histone H3K36me2 demethylation, leading to EZH2 up-regulation. It has been found that overexpression of the SA-miR let-7b down-regulates EZH2, induces premature senescence, and counteracts immortalization of MEFs driven by KDM2B. The KDM2B-let-7-EZH2 pathway also contributes to the proliferation of immortal Ink4a/Arf-null MEFs suggesting that, next to its anti-senescence role in primary cells, this histone-modifying enzyme functions more broadly in the regulation of cellular proliferation [154, 156]. It has also been shown that the KDM2B-miR-101-EZH2 pathway regulates the proliferation, migration and angiogenesis of MEFs and several cancer-derived cell lines in response to FGF2 .
It has become clear that senescence bypass may be caused by overexpression of oncomirs or down-regulation or silencing of SA-miRs. Even though different mechanisms for senescence induction exist, a pathway analysis for targets of 12 SA-miRs revealed that they can potentially regulate core senescence processes, including cell cycle regulation, regardless of the stimuli involved . Importantly, most miRNAs may have either converging targets or elicit opposing effects in cell cycle regulation. Hence, the expression of tumor suppressive SA-miRs versus tumor promoting miRNAs in a cell may lead to either growth arrest, apoptosis or proliferation. Consequently, global miRΝΑ expression profiles should be taken into account when studying the process of cancer cell homeostasis. As a paradigm, it has been shown that a distinct miRNA profile exists in prostate cancer cell progenitors, in which multiple tumor-suppressive miRNAs including miR-34a, let-7b and miR-141 were found to be down-regulated, whereas tumor promoting miRNAs including miR-301 and miR-452 were found to be overexpressed . This profile enabled the proliferation of prostate cancer cells, whereas overexpression of let-7 led to inhibition of proliferation and tumor growth by inducing a G2/M cell cycle arrest .
A study aiming at miRNA expression dynamics using miR transcriptome profiling led to the identification of miRNA expression signatures associated with different stages of cancer progression and with the acquisition of cancer hallmarks in a prototypical mouse pancreatic neuroendocrine cancer model . It was shown that in hyperplastic/dysplastic islets representing a hyperproliferative stage, nine miRNAs were up-regulated at least two-fold and one was down-regulated 2.5-fold. Interestingly, six of the nine up-regulated miRNAs, miR-17-5p, miR-201, miR-92, miR-106a, miR-106b and miR-25, that belong to the well-characterized miR-17-92 cluster paralogs (see above), were found to possess oncogenic properties and to be associated with increased proliferation . Interestingly, miRNA up-regulation was first detected in hyperplastic islets and was further found to be modest in angiogenic islets. Additional up-regulated hyperplastic signature miRNAs included miR-142-3p, miR-155 and miR-15. In contrast, it was found that the pancreatic tumor suppressor miR-410  was down-regulated in the hyperplastic/dysplastic islets compared to normal pancreatic islets . While miR transcriptome alterations are difficult to model during the development of primary human cancers, a survey of published cancer miRNA profiling studies revealed that the majority of altered miRNAs in the mouse neuroendocrine tumors were, in fact, similarly affected in a number of human tumor types . In several additional studies miRNA expression profiles during different stages of cancer progression in several different cancer types, including prostate cancer , squamous lung cell cancer , NSCLC , oral cancer [81, 82] and cervical cancer  have been compared. The results from these studies suggest that the miRNA transcriptome may be altered in instructive ways during distinct stages of the multistep process of carcinogenesis.
Collectively, these studies show that miRNA expression profiles may be altered, with some miRNAs being up-regulated and others being down-regulated, following replicative senescence, SIPS or OIS. Oncogenic miRNAs were found to bypass or delay senescence by affecting tumor suppressor pathways. In contrast SA-miRs were found to re-enforce senescence by targeting cell cycle and epigenetic regulators. These miRNAs are either mutated or epigenetically silenced during cancer progression. Hence, the balance between oncomiRs and SA-miRs is considered to play a critical role in senescence bypass, premalignancy and ultimately overt malignancy.
6 From premalignant lesions to malignant tumors: impact of specific miRNAs on angiogenesis
The aquisition of an immortal phenotype and the induction of tumor angiogenesis are two key events in cancer development. Normal and pre-malignant tissues generally exhibit a less molecular complexity than cancerous cells. Human premalignant lesions, including naevi, colon and lung adenomas, prostatic and cervical intraepithelial neoplasias exhibit features of senescence in vitro and in vivo. However, a subset of these pre-malignant lesions eventually progresses to tumors in situ and then to malignant invasive tumors, suggesting that premalignant cells can either bypass or escape the senescent response (reversion from a senescent state), or both . Some of the genes and miRNAs (both oncomirs and SA-miRs) as well as the molecular mechanisms involved in senescence bypass and malignant progression have been discussed above (Fig. 2).
While the induction of cell proliferation, in conjunction with the mechanisms evading or disrupting senescence, are fundamental steps in cancer development, another key step is the acquisition of enhanced angiogenic activity. Benign tumors progress to malignancy by acquiring several hallmarks of cancer such as sustained proliferative signaling, evasion of growth suppressors, replicative immortality, development of genomic instability and an inflammatory state . However, benign tumors can be maintained only if they have an adequate blood supply. Histological analyses of premalignant non-invasive lesions have revealed that angiogenesis is induced early during the multistage development of invasive cancers both in animal models and in humans [165, 166]. MiRNAs play important roles in normal endothelial cell (EC) function and angiogenesis [167, 168, 169, 170], as well as in tumor angiogenesis, which involves an angiogenic switch from vascular quiescence [171, 172, 173, 174, 175]. Tumor angiogenesis is a central process in cancer development, since without an appropriate vasculature the cancer cells inside the tumor will suffer from hypoxia and a diminution of nutrients leading to stress and, finally, cell death.
MiRNA profiling in human umbilical vein endothelial cells (HUVECs) has revealed that miR-221/222, miR-21, miR-126, the let-7 family, and the miR-17-92 and miR-23-24 clusters, are highly expressed in these cells. MiR-126 is the only miRNA known to be expressed specifically in the endothelial lineage and in hematopoietic progenitor cells [167, 176, 177]. An additional set of 15 over-expressed miRNAs was found to putatively target angiogenic factors and receptors. It has also been reported that the anti-angiogenic miR-221 and miR-222 (miR-221/222) inhibit angiogenic responses to stem cell factor (SCF) by targeting the SCF receptor c-kit, and suppress endothelial cell migration and proliferation in vitro . These miRNAs are also up-regulated during EC  and lung HDF  senescence. In addition, in an in vitro cell model of human nasopharyngeal carcinoma cells cultured under hypoxia, 96 miRNAs were identified through computational analysis as possible regulators of vascular endothelial growth factor (VEGF) and/or other angiogenic factors, acting either in concert or competitively .
Distinct miRNAs, such as let-7f, miR-27b and miR-130a, have been recognized as pro-angiogenic factors, whereas some other miRNAs, such as miR-378 and the miR-17-92 cluster, have also been found to be involved in tumor angiogenesis . In K-Ras-transformed p53-null mouse colonocytes, c-Myc expression has been found to stimulate in vivo tumor growth and vascularization in a non cell autonomous manner by up-regulating the expression of the miR-17-92 cluster. The proposed mechanism underlying this phenomenon was that two angiogenic inhibitors, thrombospondin-1 (Tsp1) and connective tissue growth factor (CTGF) were targeted by two miRNAs of the miR-17-92 cluster, miR-19 and miR-18, respectively. These findings illustrate that different miRNAs derived from one cluster or family can synergize to achieve a certain biological outcome [9, 181].
MiR-126 is transcribed from an intron of the Egfl7 gene that encodes an EC-derived secreted peptide, which acts as a chemo-attractant and inhibitor of smooth muscle cell migration. MiR-126 is an important key regulator of angiogenesis, known as angiomiR. Loss-of-function studies in mice and zebrafish have revealed an important function of miR-126 in governing vascular integrity and angiogenesis, due to defective EC function in response to the angiogenic factors VEGF and FGF. The pro-angiogenic action of miR-126 is mediated, at least in part, by promoting MAP kinase and PI3K signaling in response to VEGF and FGF, through targeting negative regulators of these signaling pathways. These negative regulators include the Sprouty-related EVH domain-containing protein Spred-1 and the PI3K regulatory subunit 2 (PIK3R2/p85-b), which are intracellular inhibitors of AKT and RAF kinases [167, 168, 169, 170]. MiR-126 may be epigenetically silenced in many human cancers, including breast, gastric and colon cancers. Loss of miR-126 has been associated with a reduced metastasis-free survival in recurrent breast cancer patients. Mechanistically, miR-126 regulates endothelial cell recruitment to metastatic breast cancer cells by inhibiting several pro-angiogenic factors, including IGFBP2, PITPNC1 and MERTK . In addition, it has been found that miR-126 levels may inversely correlate with microvessel density in lung cancer, suggesting that miR-126 is an important regulator of angiogenesis in multiple tumor types .
MiR-210, which is positively regulated at the transcriptional level by hypoxia-inducible factors, is a crucial regulator of angiogenesis and EC survival in response to hypoxia. MiR-210 regulates multiple normoxic genes and tumor-associated pathways, while its overexpression in normoxic ECs has been found to stimulate angiogenesis and VEGF-induced cell migration. Conversely, it was found that blocking of miR-210 via anti-miRs inhibited tube formation stimulated by hypoxia, as well as cell migration in response to VEGF. MiR-210 silencing has also been found to inhibit cell growth and to induce apoptosis under both normoxic and hypoxic conditions. Ephrin-A3, a receptor tyrosine kinase implicated in cell migration and adhesion, is another target of miR-210 that may regulate the hypoxia response .
In a prototypical mouse pancreatic neuroendocrine cancer model of neoplastic progression, 11 miRNAs were found to be differentially altered during the angiogenic switch from hyperproliferative pancreatic islets with a quiescent vasculature to angiogenic pancreatic islets. These included several up-regulated miRNAs of the miR-17-92 cluster (miR-17, miR-20a, miR-19b), miR-21, miR-146, miR-424 and miR-126, the expression levels of which were also increased in the hyperproliferative stage, and moderately increased in the angiogenic islets . MiR-21, commonly overexpressed in many tumor types including NSCLC [79, 184], breast , liver , colon  and human endocrine pancreatic tumors, was also found to be induced in the angiogenic pancreatic islets. MiR-21 targets PTEN, thus enhancing HIF-1α and VEGF-A expression and angiogenesis . MiR-146, like miR-155, exhibits a high expression in activated macrophages and may be expressed in immune cell types implicated in the angiogenic switch [47, 188]. MiR-424 promotes angiogenesis by inhibiting cullin 2 (CUL2), a scaffold protein critical for the assembly of the ubiquitin ligase system, thereby stabilizing and increasing HIF-1α levels during hypoxia .
Interestingly, it was found that the miR-126 precursor produced comparable amounts of 5p and 3p miRNA species . The two mature miRNAs derived from pre-miR-126, miR-126 and miR-126*, are known to act in cooperation to directly target the expression of stromal cell-derived factor-1 alpha (SDF-1α)/Chemokine (C-X-C motif) ligand 12 (CXCL12) through two unique binding sites in the Sdf-1α 3′ UTR, and subsequently the chemokine (C-C motif) ligand 2 (CCL2), in cancer cells. Through these inhibitory effects, they were found to suppress the migration and recruitment of mes-enchymal stem cells and inflammatory monocytes into tumor xenografts, ultimately suppressing lung metastasis of breast cancer cells. In the same report, miR-126/-126* appeared to exert metastasis suppressive activities predominantly at the primary tumor site, rather than in the metastatic nodules. This pair of miRNAs may be down-regulated in cancer due to host gene (Egfl7) promoter hypermethylation, thus illustrating how changes in gene methylation within tumor cells can affect their microenvironment via a miRNA-mediated mechanism [173, 174, 175, 190].
Collectively, these studies show that specific miRNAs, acting alone or in concert, play crucial roles in tumor angiogenesis enabling the evolution of benign tumors to metastatic malignant cancers, and that the ultimate balance between pro- and anti-angiogenic miRNAs regulates tumor angiogenesis.
7 Primary tumor development and stress responses
The growth and survival of cancer cells within a primary tumor depends on their potential to: (i) control their microenvironment composed of multiple parenchyma and stromal cell types with a complex histological organization, (ii) avoid destruction by immune system responses, as the immune inflammatory cells present in tumors can include both tumor-promoting as well as tumor-killing subclasses of immune effector cells, and (iii) invade into the surrounding tissues through the activation of complex biological programs leading to enhanced cell proliferation, increased resistance to cell death, increased cell motility and induction of extracellular matrix (ECM) degrading enzymes [1, 94, 173, 174, 175, 191]. These properties, along with the ability to survive stress, are controlled by pleiotropic, stress-induced transcription factor signaling pathways that are finely regulated by miRNAs (Fig. 2) through complex transcription factor (TF)-miRNA regulatory networks [25, 28, 41, 42, 43, 87, 94]. The most prominent examples are the p53-miRNA [25, 28, 43, 44, 45] and the NF-κB-miRNA [28, 46, 47, 48, 49, 50] regulatory networks, which interact with each other to determine the multistep process of tumor progression (Table 1).
7.1 MiRNA networks impacting the p53 pathway
The p53-miRNA regulatory network comprises the genes encoding the miR-34, miR-107, miR-200, miR-15/16 and miR-192/194/215 families, which are all directly induced by p53. TP53 can promote the processing of specific pri-miRNAs to pre-miRNAs in order to increase the levels of mature miRNAs with growth suppressive functions, such as miR-16-1, miR-143 and miR-145, in response to DNA damage, while mutant p53 interferes with miRNA processing [45, 192] (Table 1). The p53-miRNA regulatory network contributes to tumor suppression by controlling multiple processes, including cell cycle progression, survival, metabolism, EMT, stemness and angiogenesis [44, 45, 193].
The miR-34 family of miRNAs  promotes senescence [141, 143], growth arrest and apoptosis in response to DNA damage [28, 192]. The p53/miR-34 axis is involved in cell cycle regulation as it has been found that miR-34 induces a G1 cell cycle phase arrest by directly down-regulating the expression of cyclin-dependent kinase 4 (CDK4), CDK6, cyclin (Ccn) D1, CcnE2, E2F3, CREB, c-Myc and N-Myc. TP53 also uses miR-145 to down-regulate c-Myc and other G1/S cell cycle phase regulators, such as CDK4 and CDK6 [28, 43, 44, 45, 144, 151, 194]. MiR-34a functions as a SA-miR in cancer cells treated with chemotherapeutic agents. MiR-34a is induced by the DNA-damaging agent adriamycin and inhibits the growth of HCT116 colorectal cancer cells with a senescence-like phenotype. These effects of miR-34a are due to E2F1 down-regulation and p53 induction . Expression analyses have shown that miR-34a may be down-regulated in colon cancers compared to normal tissues. MiR-34 may act via a positive feedback loop involving miR-34, sirtuin1 (SIRT1) and p53. MiR-34 represses the expression of SIRT1, a NAD-dependent deacetylase that negatively regulates p53, leading to the induction of p53 and miR-34, which subsequently targets SIRT1. In tumors this feedback loop may be disrupted through the silencing of miR-34 genes via CpG methylation and/or mutation of p53 . MiR-34 may also induce apoptosis by repressing Bcl-2 expression .
The p53/miR-34 axis has also been implicated in cancer progression. MiR-34 deregulation has been found to occur in several epithelial cancers, melanomas, neuroblastomas, leukemias and sarcomas, in the presence or absence of p53 mutations. Besides decreased miR-34 expression due to inactivating p53 mutations, the miR-34-encoding genes themselves may also be subject to mutational or epigenetic inactivation in cancer, resulting in enhanced cancer cell proliferation and resistance to apoptosis [45, 144, 151]. In a K-Ras G12D/p53R172H mouse lung cancer model, it was found that miR-34 can prevent both lung cancer initiation and progression .
The p53-induced miRNAs miR-192, miR-194 and miR-215 directly inhibit MDM2 expression. This inhibition creates a tumor suppressive, positive feedback loop, which is impaired in several cancers due to inactivating p53 or miRNA mutations , or by as yet unknown mechanisms as in the case of multiple myeloma .
In addition to p53-induced miRNAs, there are also miRNAs that mediate inhibition of p53 expression by directly interacting with the 3′ UTR of the p53 mRNA. Such TP53-repressive miRNAs may, therefore, function upstream of p53, constituting clinically relevant oncogenes. For example, miR-125b has been found to repress p53, resulting in a decreased apoptosis by targeting BAK1, PUMA and IGFBP3 [198, 199, 200]. Elevated expression of miR-125b was found to be associated with increased tumor size and invasion in colorectal cancer samples, and to correlate with a poor prognosis and a decreased survival. In contrast, miR-449, which is regulated by E2F1 but not by p53, can indirectly activate p53 by directly suppressing the expression of SIRT1 . MiR-122 also leads to an indirect up-regulation of p53 by inhibiting the recruitment of the PP2A phosphatase to MDM2, resulting in a decreased MDM2 activity and, as a result, an increased p53 level and activity .
From these data is is clear that p53-induced miRNAs, or those activating p53, may induce growth arrest mainly by down-regulating cell cycle effectors, and that loss of p53-induced miRNAs or activation of miRNAs that suppress p53 expression leads to enhanced cancer cell proliferation, survival and resistance to apoptosis in response to DNA-damaging chemotherapeutic agents.
7.2 MiRNA networks impacting the NF-κB pathway
NF-κB transcription factors are critical regulators of pro-inflammatory and stress-like responses regulating genetic programs that sustain cell growth, increase cancer cell survival, enhance cell motility and regulate EMT and ECM homeostasis [202, 203, 204, 205, 206, 207]. However, some of the NF-κB target genes do not encode proteins but rather miRNA precursors that may either positively or negatively regulate a variety of biological processes . The NF-κB-miRNA regulatory network comprises several miRNAs with diverse biological functions related to cell cycle regulation, survival, metabolism, inflammation, cell motility, EMT and stemness. It is also implicated in the regulation of DNA damage responses, and in inflammation-linked epigenetic switches through interactions with other signaling systems during cancer development [48, 49, 50, 208, 209, 210, 211, 212, 213, 214, 215, 216]. NF-κB can drive the expression of miR-146a, miR-21, miR-155, miR-147b, miR-301a, miR-221/222, miR-125b-1, miR-23b-27b24–1, miR-30b and the miR-17-92 cluster. The NF-κB subunit p65 binds directly to promoter elements of these miRNA genes, as has been shown by chromatin immunoprecipitation analyses and confirmed by luciferase reporter assays. Additional NF-κB regulated miRNAs include miR-9 (a negative feedback regulator of NF-κB signaling), miR-181b, miR-125a and let-7 [49, 50, 211, 217, 218] (Table 1).
In addition to the regulation of miRNAs by NF-κB, a few studies have also revealed an involvement of miRNAs in the regulation or function of the upstream NF-κB activating kinases IKKβ and IKKα. For example, miR-199a has been found to negatively regulate the expression of IKKβ in ovarian cancer cells by directly binding on three target sequences in the 3′ UTR of the IKKβ mRNA, thereby inhibiting its translation. MiR-199a inhibits the secretion of pro-inflammatory cytokines, thereby altering the tumor microenvironment and causing suppression of tumor progression and chemoresistance . IKKβ is also targeted by miR-497, which directly binds on the 3′ UTR of the IKKβ mRNA in prostate cancer cells leading to inhibition of cell proliferation, migration and invasion . Moreover, it has been shown that miR-497 can be induced by IL-1β in a NF-κB-dependent manner, suggesting that NF-κB controls the expression of miR-497, which subsequently reduces the production of IKKβ through a negative feedback loop . MiR-15a, miR-16 and miR-223 have been shown to negatively regulate IKKα. These miRNAs can bind to the 3′ UTR of the IKKα mRNA, leading to changes in the expression of non-canonical NF-κB target genes during macrophage differentiation . Interestingly, miR-223 can be down-regulated by gain-of-function mutant p53 protein, which recruits the transcriptional repressor Zeb1 which directly binds to miR-223 to suppress its transcriptional activity . MiR-223 has been reported to target stathmin-1 (STMN-1), an oncoprotein that confers resistance to chemotherapeutic drugs associated with a poor clinical prognosis. Mutant p53 induced suppression of miR-223 and the consequent up-regulation of STMN-1 represents a mechanism that contributes to the chemoresistance of p53 mutant breast and colon cancer cells in response to treatment with DNA-damaging chemotherapeutic drugs . Although these studies were focused on miRNAs that regulate IKKα or IKKβ, there are no reports dealing with IKK-regulated miRNAs and their impact on cancer cell development.
MiR-146a controls inflammatory responses and has been implicated in senescence and cancer development [47, 50, 143, 145, 188]. Several lines of evidence suggest that miR-146a can function as a potent inhibitor of inflammation by negatively regulating canonical NF-κB signaling, leading to the suppression of pro-inflammatory gene expression. It has been shown that miR-146a null mice display severe tissue inflammation and increased basal cytokine production, and that they develop primary tumors in secondary lymphoid organs, suggesting that this miRNA functions as a tumor suppressor . Accordingly, down-regulation or loss of miR-146a expression has been observed in a number of human neoplasms, including natural killer (NK)/T-cell lymphomas, and thyroid, breast, gastric, pancreatic, cervical and prostate cancers. Since miR-146a, induced by inflammatory cytokines, acts as a feeback inhibitor of NF-κB signaling and inflammation, it may function as a molecular link between inflammation and cancer development [49, 50, 224].
Reduced expression of miR-146a in human breast cancer cells has been found to result in up-regulation of MMP-1, uPA and uPAR and to enhance the migratory and invasive activities of these cells in brain metastases . Similarly, miR-146a has been found to target MMP16 in human colon cancer cells and to decrease the motility and invasiveness of these cells . Breast cancer metastasis suppressor 1 (BRMS1) functions by inhibiting the activities of various genes, including NF-κB, EGFR and urokinase-type plasminogen activator (uPA)  via the induction of miR-146a . MiR-146a, which has been shown to be down-regulated in androgen-independent prostate cancer CaP cells, inhibits proliferation and hyaluronan-induced cell invasion through preventing the up-regulation of Rho-activated protein kinase 1 (ROCK1) . In a mouse model of pancreatic neuroendocrine cancer progression, miR-146a was initially found to be up-regulated during the transition from the hyperproliferative to angiogenic islet stage, after which its expression dramatically decreased in the metastatic stage . MiR-146a may also determine the BRCA phenotype in breast cancer and, thereby, the aggressiveness of the disease [138, 230]. Collectively, these data show that miR-146a may suppress canonical NF-κB signaling via a negative feedback loop, resulting in decreases in cancer cell motility and invasion by targeting the expression of EGFR and extracellular matrix degrading enzymes.
The miR-21 gene is strongly induced during inflammatory responses. Macrophages deficient in RelA/p65 and Myd88 are unable to up-regulate miR-21 in response to lipopolysaccharide (LPS) . MiR-21 expression can also be induced by the pro-inflammatory cytokine IL-6 via the STAT3 transcription fac-tor, which binds to three consensus sites within the miR-21 promoter . Transcription factors and miRNAs may both play roles in feedback loops that link inflammation with cancer. It was found that an inflammatory signal from non-transformed cells to cancer cells can initiate a positive feedback loop involving NF-κB, Lin28, let-7 miRNA and IL-6. It has been found that activation of the src oncogene can trigger an inflammatory response, which is mediated by NF-κB. This in turn up-regulates Lin28 which down-regulates the tumor suppressor let-7 miRNA. As let-7 normally represses the expression of IL-6, this reduction in let-7 results in IL-6 overproduction leading to NF-κB activation, further propagating the cycle of events and establishing a feedback loop that sustains the oncogene-induced transformed state . This loop can be triggered by an additional mechanism involving miR-21 and miR-181b-1. IL-6 can activate STAT3 which induces miR-21 and miR-181b-1, leading to inhibition of the tumor suppressors PTEN and CYLD, respectively. This inhibition also leads to NF-κB activation . Thus, NF-κB and STAT3 may act in concert with miR-21 and miR-181b-1 up-regulation concomitant with let-7 down-regulation, forming a positive feedback loop/switch that links inflammation to cancer. Also, miR-181a/b overexpression has been linked to aberrant activation of major pathways including the IL-6/STAT3 , TGFβ (by targeting TIMP3) [231, 232] and WNT/β-catenin  pathways involved in breast and liver tumorigenesis, suggesting that it may represent a crucial mechanism fuelling neoplastic transformation. Indeed, the miR-181 family has been found to be deregulated in several other cancers including pancreas, prostate, gastric and colon cancer, by targeting various tumor suppressors, including TIMP3, CYLD, PTEN and p27Kip1 [208, 232, 234]. Moreover, it has been shown that miR-181a/b can negatively regulate the expression and activity of ATM and, thereby, dampen the DNA damage response in breast cancer. Thus, aggressive breast cancer cells exhibiting elevated miR-181a/b expression levels are rendered sensitive to treatment with PARP1 inhibitors . As such, unscheduled miR-181a/b expression may represent a common step in cellular transformation, contributing to the acquisition of different hallmarks of cancer.
MiR-21 may function as a negative feedback regulator of the NF-κB signaling pathway leading to suppression of inflammation [236, 237]. MiR-21 may also promote inflammation and activate NF-κB during cellular transformation and cancer development, thus acting as an oncomir. Indeed, miR-21 is one of the most consistently up-regulated miRNAs in many different types of cancer, via several mechanisms [238, 239, 240]. An analysis of consensus sequences within the miR-21 promoter region identified several conserved enhancer elements, including binding sites for AP-1, C/EBP-α, nuclear factor I (NFI), p53 and SRF, suggesting that oncogenic transformation may affect the expression of miR-21. The transcription of miR-21 has for example been found to be induced by Ras/ERK, AP-1, NF-κB and STAT3, and to be repressed by NFI and C/EBP-α [214, 215, 238, 239, 240]. Activation of miR-21 by AP-1 in response to the Ras oncoprotein has been found to down-regulate the expression of its target gene, the programmed cell death protein 4 (PDCD4), that contributes to increased AP-1 activity and thyroid cell viability . MiR-21 targeting PDCD4 also results in increased canonical NF-κB activity, as PDCD4 is a negative regulator of NF-κB , and a reduced PDCD4 expression has been reported in several cancers . The role of miR-21 in tumor development was also investigated in a mouse K-Ras G12D lung cancer model, combined with loss-of-function and gain-of-function miR-21 alleles. It was found that overexpression of miR-21 enhanced tumorigenesis and that genetic deletion of miR-21 partially protected against tumor formation. MiR-21 is known to drive tumorigenesis by inhibiting both negative regulators of the Ras/MEK/ERK pathway and apoptosis. In this mouse study, it was shown that miR-21 targets antagonists of the Ras pathway (SPRY1, SPRY2, BTG2 and PDCD4), resulting in its activation and in inhibition of apoptosis. This auto-regulatory mechanism that is involved in Ras-induced transformation suggests that miR-21 may play a decisive role in facilitating Ras-induced transformation . Whether miR-21 can suppress OIS has, however, not been investigated yet. Sprouty2 (SPRY2), a protein that affects cellular outgrowth, branching and migration and that is down-regulated in a number of cancers expressing high miR-21 levels, has been found to act as a direct miR-21 target in colon cancer SW480 cells. MiR-21 also targets the tumor suppressor PTEN and its downstream PI3-kinase pathway in human cholangiocarcinoma and HCC cells, as well as in several other cancer cell types .
MiR-155 is a product of the non coding B-cell integration cluster (BIC) transcript, which is activated by insertion of avian leukosis virus (ALV) in the promoter region of the BIC locus in B-cell lymphomas. MiR-155 controls myeloid development and inflammatory cytokine production in myeloid cells through targeting SOCS1 and Src homology-2 domain containing inositol-5 phosphatase (SHIP1), a negative inhibitor of the PI3K/Akt pathway. Thus, miR-155 has been found to be involved in myeloid and lymphoid malignancies, as also in other cancer types [49, 50, 211, 214, 227]. In a K-Ras G12V-induced mouse model of pancreatic cancer, it was found that K-Ras G12V-induced up-regulation of miR-155 can promote cell proliferation through ROS generation. MiR-155 induction was, however, found to be blocked by NF-κB signaling inhibitors suggesting that miR-155 may play an important role in oncogenic K-Ras transformation mediated by cellular redox regulation . It has been found that neurotensin can induce miR-155 expression via NF-κB signaling and stimulate the growth of HCT-116 colon cancer cell xenograft tumors. Subsequently, it was found that inhibition of miR-155 can slow down the growth of colon cancer in vivo . MiR-155 also acts as a negative feedback regulator of NF-κB signaling, inflammation and immunity. MiR-155 target genes include Fas-associated death domain protein (FADD), IκB kinase epsilon (IKKε), Receptor (TNFR superfamily)-interacting serine threonine kinase 1 (Ripk1) and PU.1. In addition, it has been found that splenocytes of Eμ-miR-155 transgenic mice, which overexpress miR-155 specifically in B cells, exhibit lower levels of IKKβ transcripts than their wild-type counterparts. Thus, miR-155 may control the expression of both IKKβ and IKKε, which leads to repression or at least reduced NF-κB activation, constituting a negative feedback loop . Recent studies have shown that miR-155 expression in several breast cancer cell lines was inversely correlated with the level of the ErbB2 oncogene, which is implicated in breast carcinogenesis. MiR-155 can inhibit ErbB2-induced malignant transformation of human breast epithelial cells via two mechanisms. First, miR-155 can directly down-regulate ErbB2 and, second, it can suppress ErbB2 expression by targeting a miR-155 regulatory site (MRE) within the coding region of HDAC2, a transcriptional activator of ErbB2. Treatment of human breast cancer cells with trastuzumab has been found to result in miR-155 up-regulation and in a marked reduction of ErbB2 expression in ErbB2-positive breast cancer cells . Thus, miR-155 appears to have a dual role in cancer progression (Table 2).
In contrast, miR-155 is known to target a conserved sequence motif in the 3′ UTR of TRF1 (a component of the shelterin complex that protects chromosome ends and regulates telomere length and integrity). This miR-155-dependent reduction in TRF1 expression results in reduced telomere elongation and increased telomere damage, telomere fragility and chromosome instability. In contrast, it has been found that reducing miR-155 expression improves telomere function and genomic stability. MiR-155 has been found to be frequently up-regulated in human breast cancer samples, and to correlate with reduced TRF1 protein levels and a poor prognosis in estrogen receptor-positive breast cancers. Thus, miR-155 acts as a clinically relevant, novel telomere regulator that drives telomere fragility and genomic instability by repressing TRF1 expression . Hence, miR-155 and miR-181a/b are two NF-κB regulated genes involved in DNA damage and genomic instability.
Additional studies have shown that p53 mutants such as p53R248Q or p53R282W can induce the up-regulation of miR-155 [246, 247] leading to EMT, breast cancer cell migration and invasion . In the absence of mutant p53, p63 binds directly to the consensus p63 response element within the promoter of the miR-155 host gene to negatively regulate miR-155 expression. Mutant p53 enhances miR-155 expression by inhibiting p63 binding. MiR-155 targets ZNF652 by binding to its 3′ UTR. ZNF652 is a novel zinc-finger DNA-binding transcription repressor of key drivers of invasion and metastasis, such as TGFB1, TGFB2, TGFBR2, EGFR, SMAD2 and VIM, as demonstrated by ChIP assays. Silencing of ZNF652 in epithelial cancer cells was found to promote invasion into matrigel. Importantly, loss of ZNF652 expression in primary breast tumors has been found to be significantly correlated with an increased local invasion and to define a population of breast cancer patients with metastatic tumors [193, 246].
Additionally, miR-155 appears to affect the tumor microenvironment by influencing tumor-associated macrophages (TAMs) [174, 248]. In ovarian cancer, miR-155 has been found to contribute to the reprogramming of normal fibroblasts into cancer-associated fibroblasts (CAFs), which exhibit increased colony formation, motility and invasion capacities by affecting the expression of chemokines, particularly CCL20, CCL5 and IL8/CXCL8. Triple transfection of normal fibroblasts with anti-miR-31, anti-miR-214 and pre-miR-155 (miR-CAFs) was found to lead to the generation of CAFs, the major constituents of the tumor stroma, secreting CCL5, a key tumor-promoting factor that has been identified as a miR-214 target . The generation of CAFs is crucial as they promote cancer cell invasion, proliferation and metastasis. CAFs secrete cytokines and chemokines, which stimulate receptor tyrosine kinase signaling and EMT programs, while producing a characteristic extracellular matrix that promotes the attachment and invasion of tumor cells [94, 174, 175, 191, 250, 251].
MiR-301a belongs to a miRNA family that consists of five members, miR-130a, miR-130b, miR-301a, miR-301b and miR-454, which share the same seed sequence and which can potentially silence the expression of the same target genes. The miR-301a/miR-454 gene cluster is located in an intron of the protein-coding gene FAM33A, the product of which functions in chromosome segregation during mitosis . The miR-301a promoter contains a single RelA/p65 binding site whose mutation abolishes TNF-induced promoter activity. MiR-301a enhances NF-κB activity by down-regulating the NF-κB repressing factor (NKRF) and the Homeobox gene Gax (also known as Meox2), which function in a positive feedback loop [213, 253]. NKRF is a nuclear repressor that interacts with specific negative regulatory elements (NRE) within the promoters of some NF-κB responsive genes, including IL-8, IFNβ and NOS2A, and suppresses their basal expression [49, 50, 211, 253]. MiR-301a expression has been found to be up-regulated in a number of cancers, including pancreatic ductal adenocarcinoma (PDAC), HCC, lung and colon cancer [49, 50, 211, 253, 254]. MiR-301a-3p functions as an oncogene in PDAC  and in laryngeal squamous cell carcinoma (LSCC)  by targeting SMAD4. Furthermore, it has been shown that miR-301a deficiency can suppress Kras G12D-driven lung tumorigenesis in a mouse lung cancer model by activating NF-κB and STAT3, and also reduce the severity of experimental colitis and chemically-induced colon tumorigenesis . Recent studies have shown that the NF-κB regulated miR-130b acts as an oncomir in bladder cancer that sustains NF-κB activation by decreasing the expression of Cylindromatosis (CYLD), a K63-specific DUB and endogenous inhibitor of NF-κB activation . MiR-130b has also been found to act as an oncomir in glioblastoma by inhibiting Hippo signaling  and in esophageal squamous cell carcinoma by inhibiting PTEN . In liver cancer, it has been found that miR-130b promotes the generation and self-renewal of CD133+ cancer stem cells (CSCs, or tumor initiating cells, TICs) via directly targeting TP53-induced nuclear protein 1 (TP53INP1), thereby conferring chemoresistance . However, it has been shown that exogenous expression of hot spot p53 mutants, including p53R273H, p53R175H or p53C135Y, in a p53-null endometrial cancer cell line represses the expression of miR-130b (a negative regulator of Zeb1) by directly binding to its promoter, and triggers Zeb1-dependent EMT and cancer cell invasion . It has been found that epigenetic silencing of miR-130b contributes to the development of endometrial  and ovarian  cancers, suggesting that miR-130b acts as a tumor suppressor in these cancer types. MiR-130b has also been shown to act as a tumor suppressor in prostate cancer by down-regulating MMP2 . These results indicate that miR-130b may function either as an oncomir or as a tumor suppressor in a cell context-dependent manner (Table 2).
The miR-221/222 gene cluster is located on chromosome Xp11.3 . The miR-221 and miR-222 genes are paralogs, separated by a distance of 726 bp. They are transcribed as a single long noncoding pri-miR-221/222 RNA precursor, which subsequently undergoes processing to generate miR-221 and miR-222. The miR-221/222 promoter encompasses two separate distal regions which can bind the NF-κB subunit p65 and c-jun. These transcription factors act cooperatively to efficiently drive miR-221/222 transcription, a mechanism through which oncogenic transcription factors contribute to cancer development and progression . Indeed, miR-221/222 are overexpressed in a large variety of human cancers, including glioblastomas and prostate, breast, liver and NSCLC cancers, in which they have been shown to play their oncogenic roles via the down-regulation of various tumor suppressors such as p27, p57, PTEN, as well as many others [71, 72, 73, 74, 75, 266, 267]. It has, however, been found that miR-221 expression may be increased during the senescence of endothelial cells and human lung fibroblasts and that this miRNA may also function as a SA-miR in normal cells [85, 147]. Hence, the miR-221/222 cluster appears to play a dual role (Table 2).
The miR-125 family encompasses a group of highly conserved miRNAs, consisting of two sub-families (miR-125a and miR-125b), which have distinct seed regions and chromosomal locations. The miR-125a locus, consisting of miRNA125a-3p and miRNA-125a-5p, is located at 19q13, while miR-125b, consisting of miR-125b-1 and miR-125-2, is transcribed from two loci located on chromosomes 11q23 (hsa-miR-125b-1) and 21q21 (hsa-miR-125b-2). Both miR-125a and miR-125b are NF-κB-regulated, and they appear to constitutively activate the canonical NF-κB pathway by targeting TNFα-induced protein 3 (TNFAIP3, A20) . The miR-125 family members have amply been associated with cancer development [269, 270]. MiR-125a is part of a miRNA cluster that also contains miR-99b and let-7e, which are located on human chromosome 19q13, a region that is frequently deleted in primary gliomas, especially oligodendrogliomas [270, 271]. The expression of miR-125a has been found to be induced in macrophages in response to fungal and bacterial challenges. Importantly, this up-regulation can be blocked by IKK2 (IKKβ) inhibitors . MiR-125a has been shown to act as a tumor-suppressor in several cancers including HCC, NSCLC, ovarian and breast cancer , and multiple myeloma . Germline mutations in miR-125a have been found to predispose to the development of breast cancer . Targets that have been identified for miR-125a include Lin-28, Lin-41, ERBB2 and ERBB3 [270, 271].
Pri-miR-125b-1 and pri-miR-125b-2 encode the same mature miRNA, miR-125b, which is a bona fide RelA/p65-regulated miRNA . MiR-125b-1 has been implicated in some chromosomal translocations, i.e., t(11;14) (q24;q32) and t(2;11)(p21;q23), which may lead to B-cell acute lymphoid leukemia (B-ALL), myelodysplasia and acute myeloid leukemia (AML), respectively. MiR-125b is highly expressed in a subset of AML, and cooperates with other oncogenes such as BCR-ABL to promote leukemogenesis by activating an autocrine loop involving VEGFA [275, 276].
MiR-125b also appears to act as a tumor suppressor in several malignancies [270, 271]. It has been found that miR-125b can down-regulate MMP13 in cutaneous squamous cell carcinoma and inhibit cancer cell proliferation, migration and invasion . In ovarian cancer, it can target Bcl-3, an IκB family gene, and by doing so suppresseses cancer cell proliferation . In bladder cancer cells, miR-125b has been found to target E2F3, and to suppress colony formation in vitro and tumor development in vivo in nude mice . MiR-125b may also target oncogenic Lin28B2 and suppress human liver cancer cell proliferation and metastasis . It has also been found to inhibit melanoma cell proliferation by targeting c-jun  and to suppress proliferation and migration of osteosarcoma cells through down-regulation of STAT3 . Moreover, it has been shown that miR-125a and miR-125b cooperate to inhibit the expression of erbB2 and erbB3 by blocking the activation of ERK1/2 and AKT, leading to suppression of the proliferation of breast cancer cells [283, 284]. Recently, it was shown that miR-125b can suppress human AML cell proliferation, by inducing a G2/M cell cycle phase arrest, and cell invasion and promote apoptosis in a dose-dependent manner by targeting canonical NF-κB signaling . Repression of miR-125b in ovarian cancer has been shown to promote cell migration and invasion .
MiR-125b may also act upstream of p53 as a TP53-repressive miRNA . MiR-125b can repress TP53 by binding to the 3′ UTR of its mRNA and, thus, affect novel targets in the p53 network, including the apoptosis regulators Bak1, Puma and TP53INP1 [198, 199, 200]. MiR-125b also appears to be involved in the DDR, as it is down-regulated by genotoxic drugs, thereby allowing p53 activation and execution of the DDR cascade [198, 200]. MiR-125b may thus represent another level of controlling crosstalk between NF-κB and p53, which plays a pivotal role in determining the cellular response to stress. MiR-125b has a dual role, acting as an oncomir or a tumor suppressive miRNA in a cell context-dependent manner (Table 2) .
The miR-23a/27a/24–2 cluster, which includes miR-23a, miR-27a and miR-24, has been found to be dramatically down-regulated in acute erythroid leukemia (AEL) patients. Forced expression of miR-23a, miR-27a and miR-24, which induces apoptosis and erythropoiesis, has been found to inhibit adverse growth and partly relieve the leukemic symptoms of AEL patients, acting synergistically to target multiple members of the oncogenic gp130-JAK1-Stat3 pathway, and to promote the expression and activity of GATA1 inhibition via PU.1, thereby improving erythroid differentiation . In contrast, other recent studies have shown that the miR-23a level is increased in AML patients and correlates with a marked reduction in the expression of RAF kinase inhibitor protein (RKIP), a seminal regulator of intracellular signaling, exhibiting both anti-metastatic and anti-tumorigenic properties . MiR-24 was also found to promote the survival of hematopoietic cell lines, including myeloid and B cell lines, as well as primary hematopoietic cells . Additional studies have shown that miR-27a, and its coordinately expressed cluster (miR-23a ∼ miR-27a ∼ miR-24-2), may be down-regulated in AML cell lines and primary samples compared to hematopoietic stem-progenitor cells. In addition, it was found that miR-27a may function as a tumor suppressor in AML cells by reducing the expression of the anti-apoptotic 14–3-3θ protein, which interacts with and negatively regulates pro-apoptotic proteins such as Bax and Bad . While, miR-27a may act as a tumor suppressive miRNA in leukemia [287, 290], other studies have shown that it may also act as an oncomir in other human cancers including breast cancer [291, 292].
In summary, it has been shown that the majority of NF-κB-regulated miRNAs is up-regulated in several types of cancer. They may act as oncomirs by promoting cancer cell growth, migration and invasion, and by affecting the tumor microenvironment. Some miRNAs may, however, act as tumor suppressors by regulating cell cycle progression, telomere function, DNA damage and genome stability, and are implicated in the crosstalk between the NF-κB and p53 transcription factors. The interplay of TF-miRNA regulatory networks, like the interacting p53-miRNA and NF-κB-miRNA networks, with both oncogenic and tumor suppressive miRNAs affecting multiple cell cycle effectors and apoptosis regulating genes, plays a crucial role in progression towards malignancy and in the outcome of stress responses such the DNA damage responses of cancer cells elicited by irradiation or chemotherapeutic drugs.
8 Metastasis: epithelial-to-mesenchymal cell transition and cancer stem cells
Metastasis is a process by which cancer cells originating from a malignant primary tumor spread and colonize distant sites, establishing secondary tumors. Metastasis is of major clinical relevance as it is the main cause of lethality in cancer patients, accounting for at least 90% of all cancer-associated deaths. Metastatic dissemination occurs via a sequence of events, also known as the ‘invasion-metastasis cascade’ which is supported by functions of the cancer cells themselves, the tumor stroma and its microenvironment [173, 191, 293, 294]. This process includes the local invasion of primary tumor epithelial cells (carcinoma in situ to invasive carcinoma), followed by entry into the circulation through lymphatic or blood vessels known as intravasation. It subsequently involves their infiltration into distant tissues/organs and invasion in the parenchyma of such tissues (extravasation), followed by colonization and formation of micrometastases, some of which develop into macroscopic metastases [1, 94, 295, 296, 297, 298]. Several cell types within the tumor microenvironment, including cancer-associated fibroblasts (CAFs), immune and endothelial cells and pericytes, express several miRNAs that can mediate changes in the tumor microenvironment and influence cancer growth, angiogenesis, EMT and metastasis .
In solid tumors of epithelial origin metastasis is initiated by a multifaceted developmental program referred to as EMT, which provides primary tumor cells with the abilities to invade, resist apoptosis and disseminate to distant sites. During the course of the invasion-metastasis cascade, the EMT phenotype may be activated transiently or stably, and to several degrees, in carcinoma cells through the activity of a variety of transcription factors. EMT involves the loss of an epithelial cell phenotype and the acquisition of a mesenchymal-like phenotype, which is critical for cancer cell migration and metastasis, and functional loss of the epithelial cell marker E-cadherin, an event that is considered to be a hallmark of EMT. Hence, the best-studied transcriptional modulation during EMT involves the E-cadherin (CDH1) gene promoter, which possesses E-box elements that regulate transcriptional repression in mesenchymal cells. Several EMT-inducing pleiotropic zinc finger and bHLH transcription factors have been identified, including Snail1, Slug, Twist1 and Zeb1/2, that repress CDH1 expression [1, 94, 250, 251, 294, 296, 297, 299, 300, 301]. Importantly, these critical EMT regulators and CDH1 repressors act as direct or indirect (via HIF) downstream NF-κB targets [209, 251, 293, 297, 301, 302, 303, 304].
An extensive crosstalk among EMT-inducing TFs and miRNAs that can act as pro- or anti-metastatic factors has been documented [251, 296, 297, 301, 305, 306, 307] (Fig. 2). Several studies have shown that members of the miR-200 family (miR-200a, miR-200b, miR-200c, miR-141 and miR-429) and miR-205 can negatively regulate the EMT program in various cancer-derived cell lines by targeting the 3′ UTRs of Zeb1 and Zeb2 EMT-inducing TFs. In addition, the expression levels of miR-205 and members of the miR-200 family have been found to be inversely correlated with the expression level of vimentin, a mesenchymal cell marker. Conversely, exogenous expression of members of the miR-200 family has been found to promote a mesenchymal-to-epithelial cell transition (MET), a procedure that reverses EMT and is important for cancer cell colonization once these cells have infiltrated the parenchyma at distant metastatic sites [251, 308, 309, 310, 311]. Colorectal cancer (CRC) is a heterogeneous disease  and four consensus molecular subtypes (CMSs) of CRC have been identified of which the mesenchymal CMS4 subtype exhibits the worst prognosis. The miR-200 family has been identified as the most powerful determinant of CMS4-specific gene expression . In a prototypical mouse model of pancreatic neuroendocrine cancer, it was found that the miR-200 family was markedly down-regulated in metastases, thereby relieving repression of the mesenchymal transcription factor Zeb1 .
Hypoxia is a condition that cancer cells may experience within the tumor microenvironment during malignant progression, and this condition may control metastasis . The hypoxia-regulated transcription factor HIF1α has been found to induce the expression of miR-210 in several cancer types . MiR-210 is an intronic miRNA located within the AK123483 locus . MiR-210 overexpression may repress the initiation of tumor growth, playing a regulatory role in the adaptation of cancer cells to hypoxia, while miR-210 may act as a tumor suppressor under normoxic conditions . Expression of DICER, an enzyme involved in miRNA processing, has been shown to be down-regulated by hypoxia through an epigenetic mechanism that involves inhibition of the oxygen-dependent H3K27me3 demethylases KDM6A/B, resulting in silencing of the DICER promoter. Subsequent reduced miRNA processing activity was found to lead to derepression of the miR-200 target Zeb1, to stimulate EMT and to ultimately result in the acquisition of stem cell phenotypes in human mammary epithelial cells, thus uncovering a relationship between oxygen-sensitive epigenetic regulators, miRNA biogenesis and tumor stem cell phenotypes that may underlie a poor outcome in breast cancer . Despite these studies and the fact that miR-210 up-regulation is frequently observed in various types of cancer, the role of miR-210 in tumorigenesis may vary, and it is still unclear whether it functions as an oncomir or as a tumor suppressive miRNA, since miR-210 has been found to have multiple targets and to operate in complex networks [174, 318] (Table 2).
Members of the SNAIL superfamily of zinc finger transcription factors (Snail1 and Snail2/Slug), are also known to act as key inducers of EMT, cell motility and stemness. Snail1 can directly repress the transcription of CDH1 and activate the expression of the ZEB genes through different mechanisms, including direct activation of Zeb1 and induction of a natural antisense transcript (NAT) which favors the translation of Zeb2 mRNA [297, 310, 319]. Regulatory networks linking different EMT inducers to miRNA regulators are currently emerging . Several studies have e.g. uncovered a reciprocal feedback loop between the ZEB family of EMT inducers (Zeb1 and Zeb2) and the miR-200 family as an inducer of epithelial differentiation. Within this ZEB-miR-200 feedback loop, ZEB inhibits the transcription of miR-200 family members and miR-200 family members inhibit the translation of Zeb1, thereby mutually controling their expression [320, 321]. While Zeb1 can induce EMT and a stem cell-like phenotype by directly inhibiting the expression of CDH1 and its own repressor miR-200, miR-200 can induce differentiation by targeting its own repressor Zeb1 and directly inhibiting the translation of stem cell factors and stem cell-associated epigenetic regulators, such as BMI1 [320, 321] and SUZ12 . Zeb1 is a strong inducer of tumor cell invasion and has been found to be necessary for metastasis in animal models . It is overexpressed in a large number of human cancer types and is associated with a poor prognosis. Intriguingly, also miR-200 can be overexpressed in certain cancer types, and its expression level may also be associated with a poor prognosis . A possible molecular explanation for these apparently contradicting results is that, although miR-200 down-regulation enhances dissemination, its re-expression may induce MET which is crucial for metastatic colonization and macro-metastasis formation [297, 324].
Other potent EMT inducers, such as Snail1, are not targeted by miR-200 and, therefore, are not directly controlled by the ZEB-MiR-200 feedback loop. Snail1 has been found to be involved in another feedback loop with miR-34. Snail1 can inhibit the transcription of miR-34 family members, whereas miR-34 can inhibit the translation of Snail1 [325, 326]. However, Snail1 and miR-200c may act antagonistically in EMT induction via a reciprocal regulatory loop [297, 327, 328, 329]. Thus, the ZEB-miR-200 and the Snail1-miR-34 double negative feedback loops are linked together via miR-200 to regulate EMT, cell motility and stemness . However, Snail1, a NF-κB target gene [302, 330, 331], has been shown to repress p53 through the formation of a tri-molecular complex, Snail1/HDAC1/p53, which deacetylates active p53 to promote its proteasomal degradation, which can lead to an increased expansion and activity of tumor-initiating cells in breast cancer . Thus, in addition to the NF-κB-driven role of miR-125b as a repressor of p53 (discussed above) , the Snail1-dependent direct repression of both p53 activity and the p53-miR-34 axis provides another level of control by which NF-κB may suppress p53-mediated effects, resulting in EMT, cell motility, stemness and metastasis. In keeping with this, it has been shown that miR-34a suppression is required for IL6-driven EMT in a colorectal cancer model. This IL-6/STAT3/miR-34a feedback loop also operates in primary colorectal tumors and is associated with metastases in patients [307, 332] (Table 1).
The members of the miR-200 family of miRNAs and miR-34 are p53-induced miRNAs of the p53-miRNA regulatory network which negatively regulates cell plasticity, stemness and cell motility [45, 325, 326, 327, 333], suggesting that p53 controls EMT and metastasis through multiple miRNA-dependent routes [297, 334, 335]. Hence, inactivating p53 mutations and alterations in the expression of miRNAs through epigenetic modifications may be required to maintain EMT and metastasis . In tumor cells, p53 is frequently inactivated by mutations  whereas the miR-34a and miR-34b/c genes may be silenced by CpG methylation, correlating with metastasis and a poor survival [45, 83, 144, 151, 326]. Resulting increases in Snail1 expression presumably lock cells in a mesenchymal state and promote metastatic processes .
MiR-185 may function as an anti-metastatic miRNA, and a negative correlation between miR-185 and colorectal cancer (CRC) has recently been documented . STIM1 (stromal interaction molecule 1), an endoplasmic reticulum Ca2+ sensor that triggers store-operated Ca2+ entry activation, was recently found to be highly expressed in invasive CRC-derived cell lines and in primary CRC tissues compared to adjacent noncancerous tissues. Silencing of STIM1 was found to result in reduced CRC cell metastasis in vitro and in vivo, whereas enhanced expression of STIM1 was found to promote CRC cell metastasis by inducing EMT. It was also found that miR-185 directly targets STIM1 by binding to its 3′ UTR, and suppresses STIM1-induced EMT of CRC-derived cell lines and their migrative and invasive abilities . Previous studies have shown that miR-185 can induce cell cycle arrest and inhibit cell proliferation in NSCLC  and CRC by targeting RhoA and Cdc42 expression , and in other human cancers including breast cancer by targeting the Six1 oncogene which encodes a homeobox protein . MiR-185 can also inhibit breast cancer cell proliferation by targeting VEFF-A . Hence, miR-185 acts as a metastasis suppressor.
Recent studies have revealed a molecular mechanism by which oncogenic K-Ras, which is activated in over 90% of PDAC pancreatic cancers, suppresses miR-489 to drive tumor progression and metastasis. It was shown that K-Ras, by acting through canonical NF-κB signaling, can activate the transcription factor YY1. This results in repression of the tumor suppressor miR-489 which targets the extracellular matrix factors ADAM9 and MMP7 and, by doing so, inhibits cell migration and metastasis of PDAC cells .
MiR-506 may also act as a metastasis suppressor. This miRNA was found to be able to induce G1/S cell cycle phase arrest, and to enhance the apoptosis and chemosensitivity of PDAC cells by targeting sphingosine kinase 1 (SPHK1) that activates the SPHK1/Akt/NF-κB signaling pathway, which is often activated in pancreatic cancer .
Several studies have shown that metastatic growth may be initiated by the suppression of miR-126 and miR-335 in breast and colorectal carcinomas. The expression of miR-126 is frequently lost in colon and breast cancers and targets PI3K signaling, whereas miR-335 suppresses cell migration and metastasis by targeting the progenitor cell transcription factor SOX4 and the extracellular matrix component tenascin C. A well-defined miR-335 gene signature comprises six genes (COL1A1, MERTK, PLCB1, PTPRN2, TNC and SOX4) whose expression was found to be suppressed by miR-335 but to be elevated in metastatic cells, and to be associated with a decreased metastasis-free survival in breast cancer patients .
MiR-150 has been identified as yet another negative regulator of EMT in esophageal squamous cell carcinoma (ESCC) cells through Zeb1 targeting, and down-regulation of miR-150 has been associated with a poor prognosis of ESCC patients . Inhibitors of metastasis and invasion also include miR-143 and miR-145 in prostate cancer , miR-29b in prostate and liver cancer by targeting VEGF-A and PDGF, and miR-139 in HCC by targeting Rock2 . Tumor-suppressive miR-143 secreted by the non-cancerous prostate cell line PNT-2 has been found to inhibit the growth exclusively of prostate cancer PC-3M cells in vitro and in vivo .
The miR-148a, miR-34b/c and miR-9 genes have been found to specifically undergo hypermethylation induced silencing in metastatic cancers compared to normal tissues. Silencing of the promoters of tumor suppressor miRNAs due to DNA methylation has been found to contribute to the development of metastases in many human cancer types including melanoma, colon and breast cancer. Introduction of miR-148a and miR-34b/c in epigenetically inactivated cancer cells inhibited their motility, reduced tumor growth, and inhibited metastasis formation in xenograft models, correlating with down-regulation of several oncogenic miRNA target genes, such as c-MYC, E2F3, CDK6 and TGIF2 [21, 349].
The heterochronic let-7 family of miRNAs (Let-7a/b/c/d/e/f/g/i and miR-98) has also been implicated in EMT . One member of let-7 family, let-7c has for example been shown to play a role in prostate and lung cancer development [351, 352]. Let-7c overexpression inhibits the growth of multiple human lung cancer cell lines in vitro, the growth of lung cancer cell xenografts in immunodeficient mice and in vivo tumor growth in the lungs of mice expressing K-Ras G12D, thereby providing strong, direct evidence for its role as a tumor suppressor . In general, the let-7 family may act as tumor suppressors by repressing certain oncogenes, including those encoding Ras family members and HMGA2 [152, 354, 355]. Additional studies have shown that certain members of the let-7 family, such as let-7i, may be down-regulated by mutant p53. The lung-specific mutant p53R273H has e.g. been shown to down-regulate tumor suppressive let-7i in the p53-null lung cancer-derived cell line H1299. Conversely, stable knockdown of endogenous mutant p53 has been found to increase let-7i levels in the breast MDA-MB-231, colon HT-29 and pancreatic MIA-PaCa2 cancer cell lines . Reduction of let-7i by mutant p53 increases the migration and metastasis of MDA-MB-231 cells. Some of the let-7i targets include E2F5, LIN28B, c- MYC and NRAS. Moreover, let-7i expression has been found to negatively correlate with the mutant p53 status in the NCI-60 cancer panel and in breast cancer patients, suggesting its clinical relevance and a role for the mutant p53/let-7i axis in various types of cancer [193, 247].
In addition to the known metastasis-suppressing miRNAs, including miR-29b, miR-31, miR-126, miR-139, miR-143, miR-145, miR-148, miR-150, the miR-200 family members, miR-205, miR-335, let-7 and the p53-regulated miRNAs, miR-34, miR-103/107 and miR-206, also several metastasis-promoting miRNAs have been identified, including miR-373, miR-520c, miR-21, miR-9, miR-103/107 and miR-10b [306, 347]. Interestingly, miR-31, a metastasis suppressor , also targets the NF-κB-inducing kinase NIK, and it has been shown to undergo polycomb-mediated epigenetic silencing leading to the aberrant activation of non-canonical NF-κB signaling in adult T cell leukemia (ATL) cells [357, 358].
MiR-373 and miR-520c have been found to stimulate the migration and invasion of breast cancer cells both in vitro and in vivo by down-regulating the expression of the CD44 adhesion molecule, which is enriched in tumor-initiating cells of mammary carcinoma . The action of miR-373 contrasts to that of miR-34a. Although CD44 has also been found to be targeted by miR-34a in prostate cancer cells, miR-34a inhibits cancer stem cell regeneration and metastatic potential, thus acting as a metastasis inhibitor . In addition, miR-373 was initially identified through a genetic screen as a factor required to bypass OIS of human dermal fibroblasts (HDFs) by interfering with the p53 pathway , while miR-34a acts as a tumor suppressor and as an inducer of senescence [141, 143].
MiR-224 is part of a miRNA cluster that together with miR-452 is intronically located in the GABRE gene, which encodes the epsilon subunit of the γ-aminobutyric acid (GABA) A receptor, being under the control of common regulatory elements. MiR-224/452-mediated down-regulation of the tumor suppressor TXNIP was found to be essential for E2F1-induced EMT, cell motility and invasion of malignant melanomas .
The NF-κB regulated oncogenic miR-21  also plays a role in invasion and metastasis in various types of cancer [238, 239, 306], including colorectal , non-small cell lung  and breast cancer , which all exhibit a significant up-regulation of miR-21. The pro-metastatic effects of miR-21 can be due to the targeting a multitude of genes, such as the tumor suppressor PDCD4, an antagonist of the Ras pathway in colorectal cancer , tropomyosin 1 (TPM1) in breast cancer  and the PTEN/AKT signaling pathway in NSCLC , all acting as efficient inhibitors of invasion and metastasis. Additional miR-21 targets include components of the p53 network, i.e., CDC25A, NFIB and maspin, and regulators of matrix metalloproteinases such as TIMP3 and RECK .
MiR-9, an NF-κB-regulated miRNA [28, 50] that is enriched in c-MYC-amplified human tumors, was found to modulate tumor angiogenesis through the regulation of VEGF-A . In breast cancer cells, miR-9 was found to be induced by c-MYC and MYCN and to target CDH1, to promote EMT and increase cell motility and invasiveness. E-cadherin down-regulation by miR-9 was found to activate β-catenin signaling, thereby further up-regulating VEGF-A expression and increasing tumor angiogenesis. Additionally, it has been shown that overexpression of miR-9 in non-metastatic breast tumor cells enables them to form lung micrometastases in mice, while inhibiting miR-9 in highly malignant cells blocks their metastatic potential .
In contrast, p53-inducible miR-103/107 has been shown to directly target DICER1 mRNA. Exogenous expression of miR-103/107 has been found to enhance the migration and to promote EMT and metastatic dissemination of non-aggressive breast cancer cells, whereas loss of miR-103/107 opposes the migration and metastasis of malignant cells. Although these observations may seem inconsistent with p53-induced tumor suppression by miR-103/107, it is likely that p53 induces miR-107 and, by doing so, down-regulates DICER1 to suppress p53-induced miRNA synthesis. It is relevant to note here that miR-103/107 has been found to down-regulate miR-200 levels .
Another example of a miRNA with pro-metastatic properties is miR-10b [305, 306]. MiR-10b has been found to be highly expressed in cultured metastatic cancer cells and in metastatic breast tumors derived from patients, and has been associated with mesenchymal features and invasive properties. Exogenous expression of miR-10b in two non-metastatic breast cancer cell lines followed by implantation into mammary fat pads of immunodeficient mice resulted in the formation of mammary tumors and lung metastases . Subsequently, it was noted that the level of miR-10b not only correlates with the metastatic potential of various tumor cell types, but also reflects the expression pattern of Twist1, a pleiotropic transcription factor that positively regulates cancer metastasis. Exogenous expression of Twist1 in human mammary epithelial cells (HMECs) was found to induce miR-10b by binding to an E-box sequence that is located within the putative miR-10b gene promoter, providing strong evidence that miR-10b can be directly regulated by Twist1. Twist1-induced miR-10b expression in HMECs was found to induce cell invasion and metastasis by suppressing the translation of HOXD10, leading to increased protein levels of a small pro-metastatic G protein, RhoC, that plays a key role in facilitating cell motility and invasiveness [305, 306, 366]. Twist1 is also a NF-κB regulated gene , hence in addition to the NF-κB/Snail1/p53-miR-34 axis, the NF-κB-dependent Twist1 induction of miR-10b may represent another mechanism by which NF-κB contributes to metastasis.
Induction of the EMT program by EMT-inducing TF-miRNA regulatory networks results in the generation of cancer cells with stem cell-like characteristics that have a propensity to invade surrounding tissue and to display resistance to therapy. Cancer stem cells (CSCs) displaying a CD44+/CD24− cell surface stem cell-like phenotype may act as tumor initiating cells  and contribute to intratumoral heterogeneity [251, 294, 297]. CSCs migrating away from the primary tumor site enter the circulation to localize into distant tissues/organs. These circulating tumor cells (CTCs), which can be detected in the blood of patients with advanced primary carcinomas, extravasate and invade into the parenchyma of distinct tissues. The occurrence of CTCs and extravasated, disseminated tumor cells (DTCs) correlates with an increased aggressiveness and metastasis, and a decreased time to relapse [251, 294, 297]. Several miRNAs are involved in the generation and maintenance of CSCs [369, 370], including those induced by p53  and NF-κB [49, 50]. Indeed, the miRNA profile of CSCs is different from that of its corresponding non-CSCs. In addition, many miRNAs have been shown to regulate the self-renewal and differentiation properties of CSCs. For example, miR-93 has been found to inhibit the CSC phenotype and metastasis in breast and colon cancer by targeting SOX4. MiR-200a has been found to inhibit ovarian and breast CSCs by targeting Zeb1/2, leading to suppression of migration and invasion, while miR-34a has been found to inhibit CSCs and metastasis in lung, colon and prostate cancers by targeting CD44. In liver and prostate cancer, miR-145 has been shown to affect CSC self-renewal by targeting Oct4. Some miRNAs can down-regulate the expression of the pluripotency transcription factors Nanog, Oct4, Sox2 and Klf4. For example, miR-134 has been found to down-regulate Nanog in brain tumors and prostate cancer, miR-145 to down-regulate Oct4 in HCC and prostate cancer, and miR-7 to down-regulate Klf4 in breast cancer [369, 370]. Clearly, the link between EMT and CSCs is regulated by complex TF-miRNA regulatory networks. Again, the equilibrium between the p53-miRNA and NF-κB-miRNA regulatory networks seems to play a crucial role in the control of metastatic progression [347, 370].
Exosomes are small nanovesicles secreted from a variety of cell types and can circulate in biological fluids such as urine and plasma. They play important roles in intercellular communication, induce physiological changes in recipient cells by transferring lipids, proteins and nucleic acids including miRNAs, and have been implicated in a number of human diseases, including cancer, contributing to tumor progression and metastasis [371, 372]. It has e.g. been shown that glioblastoma cells release exosomes containing mRNA, miRNA and angiogenic proteins. These exosomes were found to stimulate the proliferation of a human glioma-derived cell line and also tubule formation by endothelial cells . These glioblastoma-secreted exosomes, which were found to contain and transfer the tumor-specific EGFRvIII, an oncogenic form of EGFR in gliomas, are also known as oncosomes . Importantly, they were also found to contain 11 miRNAs known to be abundant in gliomas, including let-7a, miR-15b, miR-16, miR-19b, miR-20, miR-21, miR-26a, miR-27a, miR-92, miR-93 and miR-320, which correlated well with the tumor profile . Hence, these tumor-derived exosomes may act as multicomponent delivery vehicles to transfer genetic information and signaling proteins to recipient cells in the tumor microenvironment.
Recently, it was shown that breast cancer cell-derived exosomes can transfer miRNAs to immortalized, non-tumorigenic mammary cells (MCF-10A) and stimulate them to become cancerous. Among the miRNAs identified in cancer cell exosomes, many exhibit oncogenic, pro-metastatic and pro-angiogenic properties, including miR-10a, miR-27a, miR-155, miR-373, and in particular miR-10b and miR-21. This potentially expands the array of mechanisms by which cancer spreads . Importantly, it has been shown that breast cancer-secreted exosomes contain key enzymes involved in miRNA biogenesis such as Dicer, TAR (trans-activation response) RNA-binding protein (TRBP), Ago2, and the membrane protein CD43, which plays a role in accumulating Dicer in cancer exosomes. The Dicer-containing cancer exosomes process precursor miRNAs into mature miRNAs (including oncomirs) over time and, by doing so, promote the oncogenic transformation of MCF-10A cells . Similar studies have shown that miR-200-containing exosomes can promote breast cancer cell metastasis to the lungs in vivo by altering target gene expression and inducing EMT .
MiRNAs have also been identified in lung cancer-derived exosomes . Through NanoString analysis, nine miRNAs were detected in exosomes derived from two human lung cancer-drived cell lines with the oncomirs miR-21, miR-27b and miR-29a being expressed at significantly higher levels . In addition, it was found that miR-21 and miR-29a in lung cancer-released exosomes can reach and bind as ligands to receptors of the Toll-like receptor (TLR) family, i.e., murine TLR7 and human TLR8, in immune cells. This binding leads to activation of TLRs, TLR-mediated NF-κB activation and secretion of the pro-metastatic inflammatory cytokines TNFα and IL-6, which in turn may lead to tumor growth and metastasis . Thus, miRNAs in cancer-derived exosomes may act in a paracrine manner to activate TLRs in the surrounding macrophages to release cytokines, alter the tumor microenvironment and promote tumor growth and spread [376, 377]. Hence, these miRNAs may serve as potential targets for cancer treatment. Quantitative and stoichiometric analyses of the miRNA contents of exosomes have, however, shown that regardless the source, on average, there is less than one molecule of a given miRNA present per exosome. This stoichiometry of miRNAs and exosomes suggests that most individual exosome preparations do not carry biologically significant numbers of miRNAs. Hence, these studies may challenge our perspective on exosome-mediated miRNA-based intercellular communication .
Collectively, these studies show that EMT, cancer cell migration, invasion and metastasis are regulated by miRNAs. MiRNAs acting as metastasis suppressors which target inducers of EMT and cancer stem cells, are either down-regulated by oncogenic pathways or mutated, lost or epigenetically silenced during cancer development. In contrast, metastasis-promoting miRNAs stimulate cancer cell migration, invasion and metastasis by evading tumor suppressive pathways, targeting cell adhesion molecules and regulators of extracellular matrix-degrading enzymes, stimulating angiogenesis, and promoting EMT and the generation of cancer stem cells. Some of these anti- or pro-metastatic miRNAs may also be transferred by exosomes.
9 Concluding remarks: emergence of miRNAs as novel putative therapeutic targets?
Following extensive research during the last years, miRNAs have emerged as regulators of major cellular processes. Specific miRNAs can be considered as tumor promoting or tumor suppressive (Figs 1 and 2). Their significance in cancer biology has been discussed in parallel to the function of classical oncogenes and tumor suppressor genes, with particular emphasis on complex TF-miRNA regulatory networks. We believe that an equilibrium of miRNA expression, as shown in cases of the interconnected p53-miRNA and NF-κB-miRNA networks, may influence the neoplastic transformation of a normal cell to a malignant, invasive and metastatic cancer cell, and also alter the overall behavior of cancer cells including their ability to undergo EMT and metastasize to distant sites. Exhaustive and detailed bioinformatics analysis of the TF-miRNA regulatory networks could provide further insight into the (de)regulation of gene expression, signaling and metabolic pathways in cancer versus normal cells, which may lead to novel therapeutic approaches. This notion may be important for the treatment of specific cancer types as one miRNA may function as an oncomir in one cancer type and as a tumor suppressive miRNA in another. Therefore, the study of specific miRNA signatures of cancer cells may prove useful for their early diagnosis and for the design of novel, cancer cell type-specific therapeutic approaches. Systems biology approaches and analyses of gene expression networks that include miRNA targets may offer novel insights on the role of miRNAs in the various processes of malignant transformation [19, 213, 305, 306, 369, 370] sparking hope for the development of novel chemotherapeutic agents targeting specific types of cancer, specific cancer clones or different cancer clones sharing the same miRNA signatures.
MiRNA expression signatures in human cancers can be used as diagnostic, prognostic and predictive biomarkers, as also for the identification of therapeutic in vivo targets [8, 58, 379, 380]. Due to its complexity and the enormous numbers of interaction sites between miRNAs and their mRNA targets, several rules for functional miRNA targeting have been proposed through a systematic and exhaustive evaluation of all functional miRNA site types . Current strategies for inhibitory miRNA targeting are mainly based on antisense oligonucleotides, including small interfering RNAs and antagomirs, and miRNA sponges which are RNAs that contain multiple tandem binding sites for a miRNA of interest. Additional miRNA therapeutic strategies include the use of small-molecule drugs that target specific miRNAs (SMIRs). MiRNA mimetics, which are double-stranded synthetic RNAs that mimic endogenous miRNAs, are also being used to restore the expression and function of specific miRNAs. These miRNA-based therapeutic approaches are currently at different stages of development, including preclinical studies and Phase I and IIa trials .
The first cancer-targeting miRNA drug, i.e., MRX34, a liposome-based miR-34 mimic with a diameter of ~120 nm (developed by Mirna Therapeutics) has entered Phase I clinical trials in patients with advanced HCC or metastatic liver cancer in April 2013. While preclinical studies showed that tail vein injection of MRX34 reduced tumor growth and significantly enhanced survival, with a favorable safety profile, in orthotopic HCC mouse models , the phase I clinical trial was terminated in 2016 due to serious immune related adverse events.
Tumor-derived exosomes that contain specific miRNAs may be used as cancer biomarkers and may provide information on cancer diagnosis and prognosis through a blood test. They may also contribute to the design of therapeutic strategies. One approach may be to block the release of cancer-produced exosomes, hence the transfer of ongogenic miRNAs to other cells in the tumor microenvironment. Ceramide-dependent secretion is known to be one of the mechanisms by which exosomic miRNAs can be released, and small molecule inhibitors of neutral sphingomyelinase such as GW4869 that play a key role in ceramide biosynthesis can effectively block the release of miRNA-containing exosomes [377, 380]. Anti-miR strategies can also be used to interfere with the function of secreted miRNAs in the recipient cell . A combination of miRNA therapeutic approaches together with chemotherapeutic agents may serve as another strategy. It is anticipated that future strategies aimed at interfering with the various aspects of exosomic miRNAs in the cell of origin, in the tumor microenvironment and/or in the recipient cell will have beneficial effects in controlling cancer development and metastasis.
This research has, in part, been co-financed by the European Union (European Social Fund - ESF) and Greek national funds through the Operational Program “Education and Lifelong Learning” of the National Strategic Reference Framework (NSRF) - Research Funding Program: THALES (CancerTFs, MIS 379435), Investing in knowledge society through the European Social Fund to E.K. (Grand number 80799), in part, by the Broad Medical Research Program at the Chron’s & Colitis Foundation of America and the Pancreatic Cancer UK-Research Innovation Fund (RIF2016_A08), by a research grant in Biomedical Sciences from FONDATION SANTÉ to E.R. and E.K., and a State (IKY) fellowship of Excellence for Postdoctoral Research in Greece - SIEMENS programme to G.M. and E.K.
Compliance with ethical standards
Conflict of interest
The authors declare no conflict of interest.
- 29.F.d.A. di Fagagna, A direct role for small non-coding RNAs in DNA damage response. Trends Cell Biol 24, 171–178 (2014)Google Scholar
- 47.K.D. Taganov, M.P. Boldin, K.J. Chang, D. Baltimore, NF-kappaB-dependent induction of microRNA miR-146, an inhibitor targeted to signaling proteins of innate immune responses. Proc Natl Acad Sci USA 103, 12481–12486 (2006)Google Scholar
- 53.S.A. Melo, S. Ropero, C. Moutinho, L.A. Aaltonen, H. Yamamoto, G.A. Calin, S. Rossi, A.F. Fernandez, F. Carneiro, C. Oliveira, B. Ferreira, C.G. Liu, A. Villanueva, G. Capella, S. Schwartz Jr., R. Shiekhattar, M. Esteller, A TARBP2 mutation in human cancer impairs microRNA processing and DICER1 function. Nat Genet 41, 365–370 (2009)PubMedPubMedCentralCrossRefGoogle Scholar
- 54.D.A. Hill, J. Ivanovich, J.R. Priest, C.A. Gurnett, L.P. Dehner, D. Desruisseau, J.A. Jarzembowski, K.A. Wikenheiser-Brokamp, B.K. Suarez, A.J. Whelan, G. Williams, D. Bracamontes, Y. Messinger, P.J. Goodfellow, DICER1 mutations in familial pleuropulmonary blastoma. Science 325, 965 (2009)PubMedPubMedCentralCrossRefGoogle Scholar
- 55.S.A. Melo, C. Moutinho, S. Ropero, G.A. Calin, S. Rossi, R. Spizzo, A.F. Fernandez, V. Davalos, A. Villanueva, G. Montoya, H. Yamamoto, S. Schwartz Jr., M. Esteller, A genetic defect in exportin-5 traps precursor microRNAs in the nucleus of cancer cells. Cancer Cell 18, 303–315 (2010)PubMedCrossRefGoogle Scholar
- 56.S.A. Melo, H. Sugimoto, J.T. O'Connell, N. Kato, A. Villanueva, A. Vidal, L. Qiu, E. Vitkin, L.T. Perelman, C.A. Melo, A. Lucci, C. Ivan, G.A. Calin, R. Kalluri, Cancer exosomes perform cell-independent microRNA biogenesis and promote tumorigenesis. Cancer Cell 26, 707–721 (2014)PubMedPubMedCentralCrossRefGoogle Scholar
- 57.P. Carninci, T. Kasukawa, S. Katayama, J. Gough, M.C. Frith, N. Maeda, R. Oyama, T. Ravasi, B. Lenhard, C. Wells, R. Kodzius, K. Shimokawa, V.B. Bajic, S.E. Brenner, S. Batalov, A.R. Forrest, M. Zavolan, M.J. Davis, L.G. Wilming, V. Aidinis, J.E. Allen, A. Ambesi-Impiombato, R. Apweiler, R.N. Aturaliya, T.L. Bailey, M. Bansal, L. Baxter, K.W. Beisel, T. Bersano, H. Bono, A.M. Chalk, K.P. Chiu, V. Choudhary, A. Christoffels, D.R. Clutterbuck, M.L. Crowe, E. Dalla, B.P. Dalrymple, B. de Bono, G. Della Gatta, D. di Bernardo, T. Down, P. Engstrom, M. Fagiolini, G. Faulkner, C.F. Fletcher, T. Fukushima, M. Furuno, S. Futaki, M. Gariboldi, P. Georgii-Hemming, T.R. Gingeras, T. Gojobori, R.E. Green, S. Gustincich, M. Harbers, Y. Hayashi, T.K. Hensch, N. Hirokawa, D. Hill, L. Huminiecki, M. Iacono, K. Ikeo, A. Iwama, T. Ishikawa, M. Jakt, A. Kanapin, M. Katoh, Y. Kawasawa, J. Kelso, H. Kitamura, H. Kitano, G. Kollias, S.P. Krishnan, A. Kruger, S.K. Kummerfeld, I.V. Kurochkin, L.F. Lareau, D. Lazarevic, L. Lipovich, J. Liu, S. Liuni, S. McWilliam, M. Madan Babu, M. Madera, L. Marchionni, H. Matsuda, S. Matsuzawa, H. Miki, F. Mignone, S. Miyake, K. Morris, S. Mottagui-Tabar, N. Mulder, N. Nakano, H. Nakauchi, P. Ng, R. Nilsson, S. Nishiguchi, S. Nishikawa, F. Nori, O. Ohara, Y. Okazaki, V. Orlando, K.C. Pang, W.J. Pavan, G. Pavesi, G. Pesole, N. Petrovsky, S. Piazza, J. Reed, J.F. Reid, B.Z. Ring, M. Ringwald, B. Rost, Y. Ruan, S.L. Salzberg, A. Sandelin, C. Schneider, C. Schonbach, K. Sekiguchi, C.A. Semple, S. Seno, L. Sessa, Y. Sheng, Y. Shibata, H. Shimada, K. Shimada, D. Silva, B. Sinclair, S. Sperling, E. Stupka, K. Sugiura, R. Sultana, Y. Takenaka, K. Taki, K. Tammoja, S.L. Tan, S. Tang, M.S. Taylor, J. Tegner, S.A. Teichmann, H.R. Ueda, E. van Nimwegen, R. Verardo, C.L. Wei, K. Yagi, H. Yamanishi, E. Zabarovsky, S. Zhu, A. Zimmer, W. Hide, C. Bult, S.M. Grimmond, R.D. Teasdale, E.T. Liu, V. Brusic, J. Quackenbush, C. Wahlestedt, J.S. Mattick, D.A. Hume, C. Kai, D. Sasaki, Y. Tomaru, S. Fukuda, M. Kanamori-Katayama, M. Suzuki, J. Aoki, T. Arakawa, J. Iida, K. Imamura, M. Itoh, T. Kato, H. Kawaji, N. Kawagashira, T. Kawashima, M. Kojima, S. Kondo, H. Konno, K. Nakano, N. Ninomiya, T. Nishio, M. Okada, C. Plessy, K. Shibata, T. Shiraki, S. Suzuki, M. Tagami, K. Waki, A. Watahiki, Y. Okamura-Oho, H. Suzuki, J. Kawai, Y. Hayashizaki, The transcriptional landscape of the mammalian genome. Sci 309, 1559–1563 (2005)CrossRefGoogle Scholar
- 59.K. Jazdzewski, S. Liyanarachchi, M. Swierniak, J. Pachucki, M.D. Ringel, B. Jarzab, A. De la Chapelle, Polymorphic mature microRNAs from passenger strand of pre-miR-146a contribute to thyroid cancer. Proc Natl Acad Sci USA 106, 1502–1505 (2009)Google Scholar
- 60.L.J. Chin, E. Ratner, S. Leng, R. Zhai, S. Nallur, I. Babar, R.-U. Muller, E. Straka, L. Su, E.A. Burki, A SNP in a let-7 microRNA complementary site in the KRAS 3′ untranslated region increases non–small cell lung cancer risk. Cancer Res 68, 8535–8540 (2008)PubMedPubMedCentralCrossRefGoogle Scholar
- 61.G.A. Calin, C.D. Dumitru, M. Shimizu, R. Bichi, S. Zupo, E. Noch, H. Aldler, S. Rattan, M. Keating, K. Rai, L. Rassenti, T. Kipps, M. Negrini, F. Bullrich, C.M. Croce, Frequent deletions and down-regulation of micro- RNA genes miR15 and miR16 at 13q14 in chronic lymphocytic leukemia. Proc Natl Acad Sci USA 99, 15524–15529 (2002)Google Scholar
- 62.N. Samuel, G. Wilson, M. Lemire, B.I. Said, Y. Lou, W. Li, D. Merino, A. Novokmet, J. Tran and K.E. Nichols, Genome-wide DNA methylation analysis reveals epigenetic dysregulation of MicroRNA-34A in TP53-associated cancer susceptibility. J Clin Oncol JCO676940 (2016)Google Scholar
- 63.S.K. Botla, S. Savant, P. Jandaghi, A.S. Bauer, O. Mücke, E.A. Moskalev, J.P. Neoptolemos, E. Costello, W. Greenhalf and A. Scarpa, Early epigenetic down-regulation of microRNA-192 expression promotes pancreatic cancer progression. Cancer Res 0390.2015 (2016)Google Scholar
- 66.A. Lujambio, S. Ropero, E. Ballestar, M.F. Fraga, C. Cerrato, F. Setien, S. Casado, A. Suarez-Gauthier, M. Sanchez-Cespedes, A. Git, I. Spiteri, P.P. Das, C. Caldas, E. Miska, M. Esteller, Genetic unmasking of an epigenetically silenced microRNA in human cancer cells. Cancer Res 67, 1424–1429 (2007)PubMedCrossRefGoogle Scholar
- 67.M. Hatziapostolou, C. Polytarchou, E. Aggelidou, A. Drakaki, G.A. Poultsides, S.A. Jaeger, H. Ogata, M. Karin, K. Struhl, M. Hadzopoulou-Cladaras, D. Iliopoulos, An HNF4alpha-miRNA inflammatory feedback circuit regulates hepatocellular oncogenesis. Cell 147, 1233–1247 (2011)PubMedPubMedCentralCrossRefGoogle Scholar
- 69.G. Romano, M. Acunzo, M. Garofalo, G. Di Leva, L. Cascione, C. Zanca, B. Bolon, G. Condorelli, C.M. Croce, MiR-494 is regulated by ERK1/2 and modulates TRAIL-induced apoptosis in non-small-cell lung cancer through BIM down-regulation. Proc Natl Acad Sci USA 109, 16570–16575 (2012)Google Scholar
- 70.H. He, K. Jazdzewski, W. Li, S. Liyanarachchi, R. Nagy, S. Volinia, G.A. Calin, C.-g. Liu, K. Franssila, S. Suster, The role of microRNA genes in papillary thyroid carcinoma. Proc Natl Acad Sci USA 102, 19075–19080 (2005)Google Scholar
- 71.C. le Sage, R. Nagel, D.A. Egan, M. Schrier, E. Mesman, A. Mangiola, C. Anile, G. Maira, N. Mercatelli, S.A. Ciafre, M.G. Farace, R. Agami, Regulation of the p27(Kip1) tumor suppressor by miR-221 and miR-222 promotes cancer cell proliferation. EMBO J 26, 3699–3708 (2007)PubMedPubMedCentralCrossRefGoogle Scholar
- 74.P. Pineau, S. Volinia, K. McJunkin, A. Marchio, C. Battiston, B. Terris, V. Mazzaferro, S.W. Lowe, C.M. Croce, A. Dejean, miR-221 overexpression contributes to liver tumorigenesis. Proc Natl Acad Sci USA 107, 264–269 (2010)Google Scholar
- 75.M. Garofalo, G. Di Leva, G. Romano, G. Nuovo, S.S. Suh, A. Ngankeu, C. Taccioli, F. Pichiorri, H. Alder, P. Secchiero, P. Gasparini, A. Gonelli, S. Costinean, M. Acunzo, G. Condorelli, C.M. Croce, miR-221&222 regulate TRAIL resistance and enhance tumorigenicity through PTEN and TIMP3 downregulation. Cancer Cell 16, 498–509 (2009)PubMedPubMedCentralCrossRefGoogle Scholar
- 76.G. Roscigno, C. Quintavalle, E. Donnarumma, I. Puoti, A. Diaz-Lagares, M. Iaboni, D. Fiore, V. Russo, M. Todaro, G. Romano, R. Thomas, G. Cortino, M. Gaggianesi, M. Esteller, C.M. Croce, G. Condorelli, MiR-221 promotes stemness of breast cancer cells by targeting DNMT3b. Oncotarget 7, 580–592 (2016)PubMedCrossRefGoogle Scholar
- 82.N.K. Cervigne, P.P. Reis, J. Machado, B. Sadikovic, G. Bradley, N.N. Galloni, M. Pintilie, I. Jurisica, B. Perez-Ordonez, R. Gilbert, P. Gullane, J. Irish, S. Kamel-Reid, Identification of a microRNA signature associated with progression of leukoplakia to oral carcinoma. Hum Mol Genet 18, 4818–4829 (2009)PubMedCrossRefGoogle Scholar
- 85.G.S. Markopoulos, E. Roupakia, M. Tokamani, G. Vartholomatos, T. Tzavaras, M. Hatziapostolou, F.O. Fackelmayer, R. Sandaltzopoulos, C. Polytarchou, E. Kolettas, Senescence-associated microRNAs target cell cycle regulatory genes in normal human lung fibroblasts. Exp Gerontol 96, 110–122 (2017)PubMedCrossRefGoogle Scholar
- 86.P.S. Linsley, J. Schelter, J. Burchard, M. Kibukawa, M.M. Martin, S.R. Bartz, J.M. Johnson, J.M. Cummins, C.K. Raymond, H. Dai, N. Chau, M. Cleary, A.L. Jackson, M. Carleton, L. Lim, Transcripts targeted by the microRNA-16 family cooperatively regulate cell cycle progression. Mol Cell Biol 27, 2240–2252 (2007)PubMedPubMedCentralCrossRefGoogle Scholar
- 97.T.J. Hudson, W. Anderson, A. Artez, A.D. Barker, C. Bell, R.R. Bernabe, M.K. Bhan, F. Calvo, I. Eerola, D.S. Gerhard, A. Guttmacher, M. Guyer, F.M. Hemsley, J.L. Jennings, D. Kerr, P. Klatt, P. Kolar, J. Kusada, D.P. Lane, F. Laplace, L. Youyong, G. Nettekoven, B. Ozenberger, J. Peterson, T.S. Rao, J. Remacle, A.J. Schafer, T. Shibata, M.R. Stratton, J.G. Vockley, K. Watanabe, H. Yang, M.M. Yuen, B.M. Knoppers, M. Bobrow, A. Cambon-Thomsen, L.G. Dressler, S.O. Dyke, Y. Joly, K. Kato, K.L. Kennedy, P. Nicolas, M.J. Parker, E. Rial-Sebbag, C.M. Romeo-Casabona, K.M. Shaw, S. Wallace, G.L. Wiesner, N. Zeps, P. Lichter, A.V. Biankin, C. Chabannon, L. Chin, B. Clement, E. de Alava, F. Degos, M.L. Ferguson, P. Geary, D.N. Hayes, A.L. Johns, A. Kasprzyk, H. Nakagawa, R. Penny, M.A. Piris, R. Sarin, A. Scarpa, M. van de Vijver, P.A. Futreal, H. Aburatani, M. Bayes, D.D. Botwell, P.J. Campbell, X. Estivill, S.M. Grimmond, I. Gut, M. Hirst, C. Lopez-Otin, P. Majumder, M. Marra, J.D. McPherson, Z. Ning, X.S. Puente, Y. Ruan, H.G. Stunnenberg, H. Swerdlow, V.E. Velculescu, R.K. Wilson, H.H. Xue, L. Yang, P.T. Spellman, G.D. Bader, P.C. Boutros, P. Flicek, G. Getz, R. Guigo, G. Guo, D. Haussler, S. Heath, T.J. Hubbard, T. Jiang, S.M. Jones, Q. Li, N. Lopez-Bigas, R. Luo, L. Muthuswamy, B.F. Ouellette, J.V. Pearson, V. Quesada, B.J. Raphael, C. Sander, T.P. Speed, L.D. Stein, J.M. Stuart, J.W. Teague, Y. Totoki, T. Tsunoda, A. Valencia, D.A. Wheeler, H. Wu, S. Zhao, G. Zhou, M. Lathrop, G. Thomas, T. Yoshida, M. Axton, C. Gunter, L.J. Miller, J. Zhang, S.A. Haider, J. Wang, C.K. Yung, A. Cros, Y. Liang, S. Gnaneshan, J. Guberman, J. Hsu, D.R. Chalmers, K.W. Hasel, T.S. Kaan, W.W. Lowrance, T. Masui, L.L. Rodriguez, C. Vergely, D.D. Bowtell, N. Cloonan, A. deFazio, J.R. Eshleman, D. Etemadmoghadam, B.B. Gardiner, J.G. Kench, R.L. Sutherland, M.A. Tempero, N.J. Waddell, P.J. Wilson, S. Gallinger, M.S. Tsao, P.A. Shaw, G.M. Petersen, D. Mukhopadhyay, R.A. DePinho, S. Thayer, K. Shazand, T. Beck, M. Sam, L. Timms, V. Ballin, Y. Lu, J. Ji, X. Zhang, F. Chen, X. Hu, Q. Yang, G. Tian, L. Zhang, X. Xing, X. Li, Z. Zhu, Y. Yu, J. Yu, J. Tost, P. Brennan, I. Holcatova, D. Zaridze, A. Brazma, L. Egevard, E. Prokhortchouk, R.E. Banks, M. Uhlen, J. Viksna, F. Ponten, K. Skryabin, E. Birney, A. Borg, A.L. Borresen-Dale, C. Caldas, J.A. Foekens, S. Martin, J.S. Reis-Filho, A.L. Richardson, C. Sotiriou, G. Thoms, L. van't Veer, D. Birnbaum, H. Blanche, P. Boucher, S. Boyault, J.D. Masson-Jacquemier, I. Pauporte, X. Pivot, A. Vincent-Salomon, E. Tabone, C. Theillet, I. Treilleux, P. Bioulac-Sage, T. Decaens, D. Franco, M. Gut, D. Samuel, J. Zucman-Rossi, R. Eils, B. Brors, J.O. Korbel, A. Korshunov, P. Landgraf, H. Lehrach, S. Pfister, B. Radlwimmer, G. Reifenberger, M.D. Taylor, C. von Kalle, P.P. Majumder, P. Pederzoli, R.A. Lawlor, M. Delledonne, A. Bardelli, T. Gress, D. Klimstra, G. Zamboni, Y. Nakamura, S. Miyano, A. Fujimoto, E. Campo, S. de Sanjose, E. Montserrat, M. Gonzalez-Diaz, P. Jares, H. Himmelbauer, S. Bea, S. Aparicio, D.F. Easton, F.S. Collins, C.C. Compton, E.S. Lander, W. Burke, A.R. Green, S.R. Hamilton, O.P. Kallioniemi, T.J. Ley, E.T. Liu, B.J. Wainwright, International network of cancer genome projects. Nature 464, 993–998 (2010)PubMedCrossRefGoogle Scholar
- 98.L. Ding, G. Getz, D.A. Wheeler, E.R. Mardis, M.D. McLellan, K. Cibulskis, C. Sougnez, H. Greulich, D.M. Muzny, M.B. Morgan, L. Fulton, R.S. Fulton, Q. Zhang, M.C. Wendl, M.S. Lawrence, D.E. Larson, K. Chen, D.J. Dooling, A. Sabo, A.C. Hawes, H. Shen, S.N. Jhangiani, L.R. Lewis, O. Hall, Y. Zhu, T. Mathew, Y. Ren, J. Yao, S.E. Scherer, K. Clerc, G.A. Metcalf, B. Ng, A. Milosavljevic, M.L. Gonzalez-Garay, J.R. Osborne, R. Meyer, X. Shi, Y. Tang, D.C. Koboldt, L. Lin, R. Abbott, T.L. Miner, C. Pohl, G. Fewell, C. Haipek, H. Schmidt, B.H. Dunford-Shore, A. Kraja, S.D. Crosby, C.S. Sawyer, T. Vickery, S. Sander, J. Robinson, W. Winckler, J. Baldwin, L.R. Chirieac, A. Dutt, T. Fennell, M. Hanna, B.E. Johnson, R.C. Onofrio, R.K. Thomas, G. Tonon, B.A. Weir, X. Zhao, L. Ziaugra, M.C. Zody, T. Giordano, M.B. Orringer, J.A. Roth, M.R. Spitz, I.I. Wistuba, B. Ozenberger, P.J. Good, A.C. Chang, D.G. Beer, M.A. Watson, M. Ladanyi, S. Broderick, A. Yoshizawa, W.D. Travis, W. Pao, M.A. Province, G.M. Weinstock, H.E. Varmus, S.B. Gabriel, E.S. Lander, R.A. Gibbs, M. Meyerson, R.K. Wilson, Somatic mutations affect key pathways in lung adenocarcinoma. Nature 455, 1069–1075 (2008)PubMedPubMedCentralCrossRefGoogle Scholar
- 99.B.A. Weir, M.S. Woo, G. Getz, S. Perner, L. Ding, R. Beroukhim, W.M. Lin, M.A. Province, A. Kraja, L.A. Johnson, K. Shah, M. Sato, R.K. Thomas, J.A. Barletta, I.B. Borecki, S. Broderick, A.C. Chang, D.Y. Chiang, L.R. Chirieac, J. Cho, Y. Fujii, A.F. Gazdar, T. Giordano, H. Greulich, M. Hanna, B.E. Johnson, M.G. Kris, A. Lash, L. Lin, N. Lindeman, E.R. Mardis, J.D. McPherson, J.D. Minna, M.B. Morgan, M. Nadel, M.B. Orringer, J.R. Osborne, B. Ozenberger, A.H. Ramos, J. Robinson, J.A. Roth, V. Rusch, H. Sasaki, F. Shepherd, C. Sougnez, M.R. Spitz, M.S. Tsao, D. Twomey, R.G. Verhaak, G.M. Weinstock, D.A. Wheeler, W. Winckler, A. Yoshizawa, S. Yu, M.F. Zakowski, Q. Zhang, D.G. Beer, I.I. Wistuba, M.A. Watson, L.A. Garraway, M. Ladanyi, W.D. Travis, W. Pao, M.A. Rubin, S.B. Gabriel, R.A. Gibbs, H.E. Varmus, R.K. Wilson, E.S. Lander, M. Meyerson, Characterizing the cancer genome in lung adenocarcinoma. Nature 450, 893–898 (2007)PubMedPubMedCentralCrossRefGoogle Scholar
- 100.M. Imielinski, A.H. Berger, P.S. Hammerman, B. Hernandez, T.J. Pugh, E. Hodis, J. Cho, J. Suh, M. Capelletti, A. Sivachenko, C. Sougnez, D. Auclair, M.S. Lawrence, P. Stojanov, K. Cibulskis, K. Choi, L. de Waal, T. Sharifnia, A. Brooks, H. Greulich, S. Banerji, T. Zander, D. Seidel, F. Leenders, S. Ansen, C. Ludwig, W. Engel-Riedel, E. Stoelben, J. Wolf, C. Goparju, K. Thompson, W. Winckler, D. Kwiatkowski, B.E. Johnson, P.A. Janne, V.A. Miller, W. Pao, W.D. Travis, H.I. Pass, S.B. Gabriel, E.S. Lander, R.K. Thomas, L.A. Garraway, G. Getz, M. Meyerson, Mapping the hallmarks of lung adenocarcinoma with massively parallel sequencing. Cell 150, 1107–1120 (2012)PubMedPubMedCentralCrossRefGoogle Scholar
- 102.E.C. de Bruin, N. McGranahan, R. Mitter, M. Salm, D.C. Wedge, L. Yates, M. Jamal-Hanjani, S. Shafi, N. Murugaesu, A.J. Rowan, E. Gronroos, M.A. Muhammad, S. Horswell, M. Gerlinger, I. Varela, D. Jones, J. Marshall, T. Voet, P. Van Loo, D.M. Rassl, R.C. Rintoul, S.M. Janes, S.M. Lee, M. Forster, T. Ahmad, D. Lawrence, M. Falzon, A. Capitanio, T.T. Harkins, C.C. Lee, W. Tom, E. Teefe, S.C. Chen, S. Begum, A. Rabinowitz, B. Phillimore, B. Spencer-Dene, G. Stamp, Z. Szallasi, N. Matthews, A. Stewart, P. Campbell, C. Swanton, Spatial and temporal diversity in genomic instability processes defines lung cancer evolution. Sci 346, 251–256 (2014)CrossRefGoogle Scholar
- 103.Cancer Genome Atlas Research Network, Comprehensive genomic characterization of squamous cell lung cancers. Nature 489, 519–525 (2012)Google Scholar
- 104.M. Peifer, L. Fernandez-Cuesta, M.L. Sos, J. George, D. Seidel, L.H. Kasper, D. Plenker, F. Leenders, R. Sun, T. Zander, R. Menon, M. Koker, I. Dahmen, C. Muller, V. Di Cerbo, H.U. Schildhaus, J. Altmuller, I. Baessmann, C. Becker, B. de Wilde, J. Vandesompele, D. Bohm, S. Ansen, F. Gabler, I. Wilkening, S. Heynck, J.M. Heuckmann, X. Lu, S.L. Carter, K. Cibulskis, S. Banerji, G. Getz, K.S. Park, D. Rauh, C. Grutter, M. Fischer, L. Pasqualucci, G. Wright, Z. Wainer, P. Russell, I. Petersen, Y. Chen, E. Stoelben, C. Ludwig, P. Schnabel, H. Hoffmann, T. Muley, M. Brockmann, W. Engel-Riedel, L.A. Muscarella, V.M. Fazio, H. Groen, W. Timens, H. Sietsma, E. Thunnissen, E. Smit, D.A. Heideman, P.J. Snijders, F. Cappuzzo, C. Ligorio, S. Damiani, J. Field, S. Solberg, O.T. Brustugun, M. Lund-Iversen, J. Sanger, J.H. Clement, A. Soltermann, H. Moch, W. Weder, B. Solomon, J.C. Soria, P. Validire, B. Besse, E. Brambilla, C. Brambilla, S. Lantuejoul, P. Lorimier, P.M. Schneider, M. Hallek, W. Pao, M. Meyerson, J. Sage, J. Shendure, R. Schneider, R. Buttner, J. Wolf, P. Nurnberg, S. Perner, L.C. Heukamp, P.K. Brindle, S. Haas, R.K. Thomas, Integrative genome analyses identify key somatic driver mutations of small-cell lung cancer. Nat Genet 44, 1104–1110 (2012)PubMedPubMedCentralCrossRefGoogle Scholar
- 105.C.M. Rudin, S. Durinck, E.W. Stawiski, J.T. Poirier, Z. Modrusan, D.S. Shames, E.A. Bergbower, Y. Guan, J. Shin, J. Guillory, C.S. Rivers, C.K. Foo, D. Bhatt, J. Stinson, F. Gnad, P.M. Haverty, R. Gentleman, S. Chaudhuri, V. Janakiraman, B.S. Jaiswal, C. Parikh, W. Yuan, Z. Zhang, H. Koeppen, T.D. Wu, H.M. Stern, R.L. Yauch, K.E. Huffman, D.D. Paskulin, P.B. Illei, M. Varella-Garcia, A.F. Gazdar, F.J. de Sauvage, R. Bourgon, J.D. Minna, M.V. Brock, S. Seshagiri, Comprehensive genomic analysis identifies SOX2 as a frequently amplified gene in small-cell lung cancer. Nat Genet 44, 1111–1116 (2012)PubMedPubMedCentralCrossRefGoogle Scholar
- 106.J. George, J.S. Lim, S.J. Jang, Y. Cun, L. Ozretic, G. Kong, F. Leenders, X. Lu, L. Fernandez-Cuesta, G. Bosco, C. Muller, I. Dahmen, N.S. Jahchan, K.S. Park, D. Yang, A.N. Karnezis, D. Vaka, A. Torres, M.S. Wang, J.O. Korbel, R. Menon, S.M. Chun, D. Kim, M. Wilkerson, N. Hayes, D. Engelmann, B. Putzer, M. Bos, S. Michels, I. Vlasic, D. Seidel, B. Pinther, P. Schaub, C. Becker, J. Altmuller, J. Yokota, T. Kohno, R. Iwakawa, K. Tsuta, M. Noguchi, T. Muley, H. Hoffmann, P.A. Schnabel, I. Petersen, Y. Chen, A. Soltermann, V. Tischler, C.M. Choi, Y.H. Kim, P.P. Massion, Y. Zou, D. Jovanovic, M. Kontic, G.M. Wright, P.A. Russell, B. Solomon, I. Koch, M. Lindner, L.A. Muscarella, A. la Torre, J.K. Field, M. Jakopovic, J. Knezevic, E. Castanos-Velez, L. Roz, U. Pastorino, O.T. Brustugun, M. Lund-Iversen, E. Thunnissen, J. Kohler, M. Schuler, J. Botling, M. Sandelin, M. Sanchez-Cespedes, H.B. Salvesen, V. Achter, U. Lang, M. Bogus, P.M. Schneider, T. Zander, S. Ansen, M. Hallek, J. Wolf, M. Vingron, Y. Yatabe, W.D. Travis, P. Nurnberg, C. Reinhardt, S. Perner, L. Heukamp, R. Buttner, S.A. Haas, E. Brambilla, M. Peifer, J. Sage, R.K. Thomas, Comprehensive genomic profiles of small cell lung cancer. Nature 524, 47–53 (2015)PubMedPubMedCentralCrossRefGoogle Scholar
- 107.E. Hodis, I.R. Watson, G.V. Kryukov, S.T. Arold, M. Imielinski, J.P. Theurillat, E. Nickerson, D. Auclair, L. Li, C. Place, D. Dicara, A.H. Ramos, M.S. Lawrence, K. Cibulskis, A. Sivachenko, D. Voet, G. Saksena, N. Stransky, R.C. Onofrio, W. Winckler, K. Ardlie, N. Wagle, J. Wargo, K. Chong, D.L. Morton, K. Stemke-Hale, G. Chen, M. Noble, M. Meyerson, J.E. Ladbury, M.A. Davies, J.E. Gershenwald, S.N. Wagner, D.S. Hoon, D. Schadendorf, E.S. Lander, S.B. Gabriel, G. Getz, L.A. Garraway, L. Chin, A landscape of driver mutations in melanoma. Cell 150, 251–263 (2012)PubMedPubMedCentralCrossRefGoogle Scholar
- 112.J. Bartkova, N. Rezaei, M. Liontos, P. Karakaidos, D. Kletsas, N. Issaeva, L.V. Vassiliou, E. Kolettas, K. Niforou, V.C. Zoumpourlis, M. Takaoka, H. Nakagawa, F. Tort, K. Fugger, F. Johansson, M. Sehested, C.L. Andersen, L. Dyrskjot, T. Orntoft, J. Lukas, C. Kittas, T. Helleday, T.D. Halazonetis, J. Bartek, V.G. Gorgoulis, Oncogene-induced senescence is part of the tumorigenesis barrier imposed by DNA damage checkpoints. Nature 444, 633–637 (2006)PubMedCrossRefGoogle Scholar
- 138.P. Moskwa, F.M. Buffa, Y. Pan, R. Panchakshari, P. Gottipati, R.J. Muschel, J. Beech, R. Kulshrestha, K. Abdelmohsen, D.M. Weinstock, M. Gorospe, A.L. Harris, T. Helleday, D. Chowdhury, miR-182-mediated downregulation of BRCA1 impacts DNA repair and sensitivity to PARP inhibitors. Mol Cell 41, 210–220 (2011)PubMedCrossRefGoogle Scholar
- 141.H. Tazawa, N. Tsuchiya, M. Izumiya, H. Nakagama, Tumor-suppressive miR-34a induces senescence-like growth arrest through modulation of the E2F pathway in human colon cancer cells. Proc Natl Acad Sci USA 104, 15472–15477 (2007)Google Scholar
- 145.D. Bhaumik, G.K. Scott, S. Schokrpur, C.K. Patil, A.V. Orjalo, F. Rodier, G.J. Lithgow, J. Campisi, MicroRNAs miR-146a/b negatively modulate the senescence-associated inflammatory mediators IL-6 and IL-8. Aging 1, 402–411 (2009)Google Scholar
- 150.L. He, X. He, L.P. Lim, E. de Stanchina, Z. Xuan, Y. Liang, W. Xue, L. Zender, J. Magnus, D. Ridzon, A.L. Jackson, P.S. Linsley, C. Chen, S.W. Lowe, M.A. Cleary, G.J. Hannon, A microRNA component of the p53 tumour suppressor network. Nature 447, 1130–1134 (2007)PubMedPubMedCentralCrossRefGoogle Scholar
- 154.A. Tzatsos, P. Paskaleva, S. Lymperi, G. Contino, S. Stoykova, Z. Chen, K.K. Wong, N. Bardeesy, Lysine-specific demethylase 2B (KDM2B)-let-7-enhancer of zester homolog 2 (EZH2) pathway regulates cell cycle progression and senescence in primary cells. J Biol Chem 286, 33061–33069 (2011)PubMedPubMedCentralCrossRefGoogle Scholar
- 155.R. Greussing, M. Hackl, P. Charoentong, A. Pauck, R. Monteforte, M. Cavinato, E. Hofer, M. Scheideler, M. Neuhaus, L. Micutkova, C. Mueck, Z. Trajanoski, J. Grillari, P. Jansen-Durr, Identification of microRNA-mRNA functional interactions in UVB-induced senescence of human diploid fibroblasts. BMC Genomics 14, 224 (2013)PubMedPubMedCentralCrossRefGoogle Scholar
- 176.Z. Li, P. Chen, R. Su, Y. Li, C. Hu, Y. Wang, S. Arnovitz, M. He, S. Gurbuxani, Z. Zuo, A.G. Elkahloun, S. Li, H. Weng, H. Huang, M.B. Neilly, S. Wang, E.N. Olson, R.A. Larson, M.M. Le Beau, J. Zhang, X. Jiang, M. Wei, J. Jin, P.P. Liu, J. Chen, Overexpression and knockout of miR-126 both promote leukemogenesis. Blood 126, 2005–2015 (2015)PubMedPubMedCentralCrossRefGoogle Scholar
- 177.E.R. Lechman, B. Gentner, S.W. Ng, E.M. Schoof, P. van Galen, J.A. Kennedy, S. Nucera, F. Ciceri, K.B. Kaufmann, N. Takayama, S.M. Dobson, A. Trotman-Grant, G. Krivdova, J. Elzinga, A. Mitchell, B. Nilsson, K.G. Hermans, K. Eppert, R. Marke, R. Isserlin, V. Voisin, G.D. Bader, P.W. Zandstra, T.R. Golub, B.L. Ebert, J. Lu, M. Minden, J.C. Wang, L. Naldini, J.E. Dick, miR-126 Regulates Distinct Self-Renewal Outcomes in Normal and Malignant Hematopoietic Stem Cells. Cancer Cell 29, 602–606 (2016)PubMedCrossRefGoogle Scholar
- 183.P. Fasanaro, Y. D'Alessandra, V. Di Stefano, R. Melchionna, S. Romani, G. Pompilio, M.C. Capogrossi, F. Martelli, MicroRNA-210 modulates endothelial cell response to hypoxia and inhibits the receptor tyrosine kinase ligand Ephrin-A3. J Biol Chem 283, 15878–15883 (2008)PubMedPubMedCentralCrossRefGoogle Scholar
- 185.L.-X. Yan, X.-F. Huang, Q. Shao, M.-Y. Huang, L. Deng, Q.-L. Wu, Y.-X. Zeng, J.-Y. Shao, MicroRNA miR-21 overexpression in human breast cancer is associated with advanced clinical stage, lymph node metastasis and patient poor prognosis. RNA 14, 2348–2360 (2008)PubMedPubMedCentralCrossRefGoogle Scholar
- 189.G. Ghosh, I.V. Subramanian, N. Adhikari, X. Zhang, H.P. Joshi, D. Basi, Y.S. Chandrashekhar, J.L. Hall, S. Roy, Y. Zeng, S. Ramakrishnan, Hypoxia-induced microRNA-424 expression in human endothelial cells regulates HIF-alpha isoforms and promotes angiogenesis. J Clin Invest 120, 4141–4154 (2010)PubMedPubMedCentralCrossRefGoogle Scholar
- 190.Y. Zhang, P. Yang, T. Sun, D. Li, X. Xu, Y. Rui, C. Li, M. Chong, T. Ibrahim, L. Mercatali, D. Amadori, X. Lu, D. Xie, Q.J. Li and X.F. Wang, miR-126 and miR-126* repress recruitment of mesenchymal stem cells and inflammatory monocytes to inhibit breast cancer metastasis. Nat Cell Biol 15, 284–294 (2013)Google Scholar
- 197.F. Pichiorri, S.S. Suh, A. Rocci, L. De Luca, C. Taccioli, R. Santhanam, W. Zhou, D.M. Benson Jr., C. Hofmainster, H. Alder, M. Garofalo, G. Di Leva, S. Volinia, H.J. Lin, D. Perrotti, M. Kuehl, R.I. Aqeilan, A. Palumbo, C.M. Croce, Downregulation of p53-inducible microRNAs 192, 194, and 215 impairs the p53/MDM2 autoregulatory loop in multiple myeloma development. Cancer Cell 18, 367–381 (2010)PubMedPubMedCentralCrossRefGoogle Scholar
- 200.M.T. Le, N. Shyh-Chang, S.L. Khaw, L. Chin, C. Teh, J. Tay, E. O'Day, V. Korzh, H. Yang, A. Lal, J. Lieberman, H.F. Lodish, B. Lim, Conserved regulation of p53 network dosage by microRNA-125b occurs through evolving miRNA-target gene pairs. PLoS Genet 7, e1002242 (2011)PubMedPubMedCentralCrossRefGoogle Scholar
- 214.K. Bakirtzi, M. Hatziapostolou, I. Karagiannides, C. Polytarchou, S. Jaeger, D. Iliopoulos, C. Pothoulakis, Neurotensin signaling activates microRNAs-21 and -155 and Akt, promotes tumor growth in mice, and is increased in human colon tumors. Gastroenterology 141, 1749–1761 e1741 (2011)PubMedPubMedCentralCrossRefGoogle Scholar
- 216.C. Polytarchou, D.W. Hommes, T. Palumbo, M. Hatziapostolou, M. Koutsioumpa, G. Koukos, A.E. van der Meulen-de Jong, A. Oikonomopoulos, W.K. van Deen, C. Vorvis, O.B. Serebrennikova, E. Birli, J. Choi, L. Chang, P.A. Anton, P.N. Tsichlis, C. Pothoulakis, H.W. Verspaget, D. Iliopoulos, MicroRNA214 Is Associated With Progression of Ulcerative Colitis, and Inhibition Reduces Development of Colitis and Colitis-Associated Cancer in Mice. Gastroenterology 149, 981–992 e911 (2015)PubMedPubMedCentralCrossRefGoogle Scholar
- 218.A. Drakaki, M. Hatziapostolou, C. Polytarchou, C. Vorvis, G.A. Poultsides, J. Souglakos, V. Georgoulias, D. Iliopoulos, Functional microRNA high throughput screening reveals miR-9 as a central regulator of liver oncogenesis by affecting the PPARA-CDH1 pathway. BMC Cancer 15, 542 (2015)PubMedPubMedCentralCrossRefGoogle Scholar
- 237.R.T. Marquez, E. Wendlandt, C.S. Galle, K. Keck, A.P. McCaffrey, MicroRNA-21 is upregulated during the proliferative phase of liver regeneration, targets Pellino-1, and inhibits NF-kappaB signaling. Am J Physiol Gastrointest Liver Physiol 298, G535–G541 (2010)PubMedPubMedCentralCrossRefGoogle Scholar
- 243.S. Costinean, N. Zanesi, Y. Pekarsky, E. Tili, S. Volinia, N. Heerema, C.M. Croce, Pre-B cell proliferation and lymphoblastic leukemia/high-grade lymphoma in E(mu)-miR155 transgenic mice. Proc Natl Acad Sci USA 103, 7024–7029 (2006)Google Scholar
- 245.R. Dinami, C. Ercolani, E. Petti, S. Piazza, Y. Ciani, R. Sestito, A. Sacconi, F. Biagioni, C. le Sage, R. Agami, R. Benetti, M. Mottolese, C. Schneider, G. Blandino, S. Schoeftner, miR-155 drives telomere fragility in human breast cancer by targeting TRF1. Cancer Res 74, 4145–4156 (2014)PubMedCrossRefGoogle Scholar
- 247.M. Subramanian, P. Francis, S. Bilke, X.L. Li, T. Hara, X. Lu, M.F. Jones, R.L. Walker, Y. Zhu, M. Pineda, C. Lee, L. Varanasi, Y. Yang, L.A. Martinez, J. Luo, S. Ambs, S. Sharma, L.M. Wakefield, P.S. Meltzer, A. Lal, A mutant p53/let-7i-axis-regulated gene network drives cell migration, invasion and metastasis. Oncogene 34, 1094–1104 (2015)PubMedCrossRefGoogle Scholar
- 254.X. Ma, F. Yan, Q. Deng, F. Li, Z. Lu, M. Liu, L. Wang, D.J. Conklin, J. McCracken, S. Srivastava, A. Bhatnagar, Y. Li, Modulation of tumorigenesis by the pro-inflammatory microRNA miR-301a in mouse models of lung cancer and colorectal cancer. Cell Discov 1, 15005 (2015)PubMedPubMedCentralCrossRefGoogle Scholar
- 257.X. Cui, C. Kong, Y. Zhu, Y. Zeng, Z. Zhang, X. Liu, B. Zhan, C. Piao, Z. Jiang, miR-130b, an onco-miRNA in bladder cancer, is directly regulated by NF-kappaB and sustains NF-kappaB activation by decreasing Cylindromatosis expression. Oncotarget 7, 48547–48561 (2016)PubMedPubMedCentralCrossRefGoogle Scholar
- 260.S. Ma, K.H. Tang, Y.P. Chan, T.K. Lee, P.S. Kwan, A. Castilho, I. Ng, K. Man, N. Wong, K.F. To, B.J. Zheng, P.B. Lai, C.M. Lo, K.W. Chan, X.Y. Guan, miR-130b Promotes CD133(+) liver tumor-initiating cell growth and self-renewal via tumor protein 53-induced nuclear protein 1. Cell Stem Cell 7, 694–707 (2010)PubMedCrossRefGoogle Scholar
- 268.S.W. Kim, K. Ramasamy, H. Bouamar, A.P. Lin, D. Jiang, R.C. Aguiar, MicroRNAs miR-125a and miR-125b constitutively activate the NF-kappaB pathway by targeting the tumor necrosis factor alpha-induced protein 3 (TNFAIP3, A20). Proc Natl Acad Sci USA 109, 7865–7870 (2012)Google Scholar
- 273.M. Leotta, L. Biamonte, L. Raimondi, D. Ronchetti, M.T. Di Martino, C. Botta, E. Leone, M.R. Pitari, A. Neri, A. Giordano, P. Tagliaferri, P. Tassone, N. Amodio, A p53-dependent tumor suppressor network is induced by selective miR-125a-5p inhibition in multiple myeloma cells. J Cell Physiol 229, 2106–2116 (2014)PubMedCrossRefGoogle Scholar
- 275.M. Bousquet, M.H. Harris, B. Zhou, H.F. Lodish, MicroRNA miR-125b causes leukemia. Proc Natl Acad Sci USA 107, 21558–21563 (2010)Google Scholar
- 276.J. Liu, B. Guo, Z. Chen, N. Wang, M. Iacovino, J. Cheng, C. Roden, W. Pan, S. Khan, S. Chen, M. Kyba, R. Fan, S. Guo, J. Lu, miR-125b promotes MLL-AF9-driven murine acute myeloid leukemia involving a VEGFA-mediated non-cell-intrinsic mechanism. Blood 129, 1491–1502 (2017)PubMedCrossRefGoogle Scholar
- 277.N. Xu, L. Zhang, F. Meisgen, M. Harada, J. Heilborn, B. Homey, D. Grander, M. Stahle, E. Sonkoly, A. Pivarcsi, MicroRNA-125b down-regulates matrix metallopeptidase 13 and inhibits cutaneous squamous cell carcinoma cell proliferation, migration, and invasion. J Biol Chem 287, 29899–29908 (2012)PubMedPubMedCentralCrossRefGoogle Scholar
- 286.L. Yang, X. Zhang, Y. Ma, X. Zhao, B. Li, H. Wang, Ascites promotes cell migration through the repression of miR-125b in ovarian cancer. Oncotarget (2017). doi: 10.18632/oncotarget.16846
- 287.R. Su, L. Dong, D. Zou, H. Zhao, Y. Ren, F. Li, P. Yi, L. Li, Y. Zhu, Y. Ma, J. Wang, F. Wang, J. Yu, microRNA-23a, −27a and −24 synergistically regulate JAK1/Stat3 cascade and serve as novel therapeutic targets in human acute erythroid leukemia. Oncogene 35, 6001–6014 (2016)PubMedCrossRefGoogle Scholar
- 288.S. Hatzl, O. Geiger, M.K. Kuepper, V. Caraffini, T. Seime, T. Furlan, E. Nussbaumer, R. Wieser, M. Pichler, M. Scheideler, K. Nowek, M. Jongen-Lavrencic, F. Quehenberger, A. Wolfler, J. Troppmair, H. Sill, A. Zebisch, Increased expression of miR-23a mediates a loss of expression in the RAF kinase inhibitor protein RKIP. Cancer Res 76, 3644–3654 (2016)PubMedPubMedCentralCrossRefGoogle Scholar
- 301.T. Ni, X.Y. Li, N. Lu, T. An, Z.P. Liu, R. Fu, W.C. Lv, Y.W. Zhang, X.J. Xu, R. Grant Rowe, Y.S. Lin, A. Scherer, T. Feinberg, X.Q. Zheng, B.A. Chen, X.S. Liu, Q.L. Guo, Z.Q. Wu, S.J. Weiss, Snail1-dependent p53 repression regulates expansion and activity of tumour-initiating cells in breast cancer. Nat Cell Biol 18, 1221–1232 (2016)PubMedCrossRefGoogle Scholar
- 304.G. Storci, P. Sansone, S. Mari, G. D'Uva, S. Tavolari, T. Guarnieri, M. Taffurelli, C. Ceccarelli, D. Santini, P. Chieco, K.B. Marcu, M. Bonafe, TNFalpha up-regulates SLUG via the NF-kappaB/HIF1alpha axis, which imparts breast cancer cells with a stem cell-like phenotype. J Cell Physiol 225, 682–691 (2010)PubMedPubMedCentralCrossRefGoogle Scholar
- 311.D.L. Gibbons, W. Lin, C.J. Creighton, Z.H. Rizvi, P.A. Gregory, G.J. Goodall, N. Thilaganathan, L. Du, Y. Zhang, A. Pertsemlidis, Contextual extracellular cues promote tumor cell EMT and metastasis by regulating miR-200 family expression. Genes Dev 23, 2140–2151 (2009)PubMedPubMedCentralCrossRefGoogle Scholar
- 312.M. Hahn, R. de Voer, N. Hoogerbrugge, M. Ligtenberg, R. Kuiper, A. Geurts van Kessel, The genetic heterogeneity of colorectal cancer predisposition-guidelines for gene discovery. Cell Oncol 39, 491–510 (2016)Google Scholar
- 313.E. Fessler, M. Jansen, E.M.F. De Sousa, L. Zhao, P.R. Prasetyanti, H. Rodermond, R. Kandimalla, J.F. Linnekamp, M. Franitza, S.R. van Hooff, J.H. de Jong, S.C. Oppeneer, C.J. van Noesel, E. Dekker, G. Stassi, X. Wang, J.P. Medema, L. Vermeulen, A multidimensional network approach reveals microRNAs as determinants of the mesenchymal colorectal cancer subtype. Oncogene 35, 6026–6037 (2016)PubMedPubMedCentralCrossRefGoogle Scholar
- 315.A. Giannakakis, R. Sandaltzopoulos, J. Greshock, S. Liang, J. Huang, K. Hasegawa, C. Li, A. O'Brien-Jenkins, D. Katsaros, B.L. Weber, C. Simon, G. Coukos, L. Zhang, miR-210 links hypoxia with cell cycle regulation and is deleted in human epithelial ovarian cancer. Cancer Biol Ther 7, 255–264 (2008)PubMedCrossRefGoogle Scholar
- 317.T. van den Beucken, E. Koch, K. Chu, R. Rupaimoole, P. Prickaerts, M. Adriaens, J.W. Voncken, A.L. Harris, F.M. Buffa, S. Haider, M.H. Starmans, C.Q. Yao, M. Ivan, C. Ivan, C.V. Pecot, P.C. Boutros, A.K. Sood, M. Koritzinsky, B.G. Wouters, Hypoxia promotes stem cell phenotypes and poor prognosis through epigenetic regulation of DICER. Nat Commun 5, 5203 (2014)PubMedPubMedCentralCrossRefGoogle Scholar
- 320.Y. Shimono, M. Zabala, R.W. Cho, N. Lobo, P. Dalerba, D. Qian, M. Diehn, H. Liu, S.P. Panula, E. Chiao, F.M. Dirbas, G. Somlo, R.A. Pera, K. Lao, M.F. Clarke, Downregulation of miRNA-200c links breast cancer stem cells with normal stem cells. Cell 138, 592–603 (2009)PubMedPubMedCentralCrossRefGoogle Scholar
- 321.U. Wellner, J. Schubert, U.C. Burk, O. Schmalhofer, F. Zhu, A. Sonntag, B. Waldvogel, C. Vannier, D. Darling, A. zur Hausen, V.G. Brunton, J. Morton, O. Sansom, J. Schuler, M.P. Stemmler, C. Herzberger, U. Hopt, T. Keck, S. Brabletz, T. Brabletz, The EMT-activator ZEB1 promotes tumorigenicity by repressing stemness-inhibiting microRNAs. Nat Cell Biol 11, 1487–1495 (2009)PubMedCrossRefGoogle Scholar
- 325.N.H. Kim, H.S. Kim, X.Y. Li, I. Lee, H.S. Choi, S.E. Kang, S.Y. Cha, J.K. Ryu, D. Yoon, E.R. Fearon, R.G. Rowe, S. Lee, C.A. Maher, S.J. Weiss, J.I. Yook, A p53/miRNA-34 axis regulates Snail1-dependent cancer cell epithelial-mesenchymal transition. J Cell Biol 195, 417–433 (2011)PubMedPubMedCentralCrossRefGoogle Scholar
- 327.J.G. Gill, E.M. Langer, R.C. Lindsley, M. Cai, T.L. Murphy, M. Kyba, K.M. Murphy, Snail and the microRNA-200 family act in opposition to regulate epithelial-to-mesenchymal transition and germ layer fate restriction in differentiating ESCs. Stem Cells 29, 764–776 (2011)PubMedPubMedCentralCrossRefGoogle Scholar
- 329.R. Perdigao-Henriques, F. Petrocca, G. Altschuler, M.P. Thomas, M.T. Le, S.M. Tan, W. Hide, J. Lieberman, miR-200 promotes the mesenchymal to epithelial transition by suppressing multiple members of the Zeb2 and Snail1 transcriptional repressor complexes. Oncogene 35, 158–172 (2016)PubMedCrossRefGoogle Scholar
- 332.M. Rokavec, M.G. Oner, H. Li, R. Jackstadt, L. Jiang, D. Lodygin, M. Kaller, D. Horst, P.K. Ziegler, S. Schwitalla, J. Slotta-Huspenina, F.G. Bader, F.R. Greten, H. Hermeking, IL-6R/STAT3/miR-34a feedback loop promotes EMT-mediated colorectal cancer invasion and metastasis. J Clin Invest 124, 1853–1867 (2014)PubMedPubMedCentralCrossRefGoogle Scholar
- 334.C.J. Chang, C.H. Chao, W. Xia, J.Y. Yang, Y. Xiong, C.W. Li, W.H. Yu, S.K. Rehman, J.L. Hsu, H.H. Lee, M. Liu, C.T. Chen, D. Yu, M.C. Hung, p53 regulates epithelial-mesenchymal transition and stem cell properties through modulating miRNAs. Nat Cell Biol 13, 317–323 (2011)PubMedPubMedCentralCrossRefGoogle Scholar
- 343.J. Li, H. Wu, W. Li, L. Yin, S. Guo, X. Xu, Y. Ouyang, Z. Zhao, S. Liu, Y. Tian, Z. Tian, J. Ju, B. Ni, H. Wang, Downregulated miR-506 expression facilitates pancreatic cancer progression and chemoresistance via SPHK1/Akt/NF-kappaB signaling. Oncogene 35, 5501–5514 (2016)PubMedPubMedCentralCrossRefGoogle Scholar
- 346.X. Peng, W. Guo, T. Liu, X. Wang, X.a. Tu, D. Xiong, S. Chen, Y. Lai, H. Du, G. Chen, Identification of miRs-143 and-145 that is associated with bone metastasis of prostate cancer and involved in the regulation of EMT. PloS One 6, e20341 (2011)Google Scholar
- 349.A. Lujambio, G.A. Calin, A. Villanueva, S. Ropero, M. Sanchez-Cespedes, D. Blanco, L.M. Montuenga, S. Rossi, M.S. Nicoloso, W.J. Faller, W.M. Gallagher, S.A. Eccles, C.M. Croce, M. Esteller, A microRNA DNA methylation signature for human cancer metastasis. Proc Natl Acad Sci USA 105, 13556–13561 (2008)Google Scholar
- 357.M. Yamagishi, K. Nakano, A. Miyake, T. Yamochi, Y. Kagami, A. Tsutsumi, Y. Matsuda, A. Sato-Otsubo, S. Muto, A. Utsunomiya, K. Yamaguchi, K. Uchimaru, S. Ogawa, T. Watanabe, Polycomb-mediated loss of miR-31 activates NIK-dependent NF-kappaB pathway in adult T cell leukemia and other cancers. Cancer Cell 21, 121–135 (2012)PubMedCrossRefGoogle Scholar
- 364.L. Ma, J. Young, H. Prabhala, E. Pan, P. Mestdagh, D. Muth, J. Teruya-Feldstein, F. Reinhardt, T.T. Onder, S. Valastyan, F. Westermann, F. Speleman, J. Vandesompele, R.A. Weinberg, miR-9, a MYC/MYCN-activated microRNA, regulates E-cadherin and cancer metastasis. Nat Cell Biol 12, 247–256 (2010)PubMedPubMedCentralGoogle Scholar
- 367.C.W. Li, W. Xia, L. Huo, S.O. Lim, Y. Wu, J.L. Hsu, C.H. Chao, H. Yamaguchi, N.K. Yang, Q. Ding, Y. Wang, Y.J. Lai, A.M. LaBaff, T.J. Wu, B.R. Lin, M.H. Yang, G.N. Hortobagyi, M.C. Hung, Epithelial-mesenchymal transition induced by TNF-alpha requires NF-kappaB-mediated transcriptional upregulation of Twist1. Cancer Res 72, 1290–1300 (2012)PubMedPubMedCentralCrossRefGoogle Scholar
- 368.M. Al-Hajj, M.S. Wicha, A. Benito-Hernandez, S.J. Morrison, M.F. Clarke, Prospective identification of tumorigenic breast cancer cells. Proc Natl Acad Sci USA 100, 3983–3988 (2003)Google Scholar
- 373.J. Skog, T. Wurdinger, S. van Rijn, D.H. Meijer, L. Gainche, M. Sena-Esteves, W.T. Curry Jr., B.S. Carter, A.M. Krichevsky, X.O. Breakefield, Glioblastoma microvesicles transport RNA and proteins that promote tumour growth and provide diagnostic biomarkers. Nat Cell Biol 10, 1470–1476 (2008)PubMedPubMedCentralCrossRefGoogle Scholar
- 376.M. Fabbri, A. Paone, F. Calore, R. Galli, E. Gaudio, R. Santhanam, F. Lovat, P. Fadda, C. Mao, G.J. Nuovo, N. Zanesi, M. Crawford, G.H. Ozer, D. Wernicke, H. Alder, M.A. Caligiuri, P. Nana-Sinkam, D. Perrotti, C.M. Croce, MicroRNAs bind to toll-like receptors to induce prometastatic inflammatory response. Proc Natl Acad Sci USA 109, E2110–E2116 (2012)Google Scholar
- 378.J.R. Chevillet, Q. Kang, I.K. Ruf, H.A. Briggs, L.N. Vojtech, S.M. Hughes, H.H. Cheng, J.D. Arroyo, E.K. Meredith, E.N. Gallichotte, E.L. Pogosova-Agadjanyan, C. Morrissey, D.L. Stirewalt, F. Hladik, E.Y. Yu, C.S. Higano, M. Tewari, Quantitative and stoichiometric analysis of the microRNA content of exosomes. Proc Natl Acad Sci USA 111, 14888–14893 (2014)Google Scholar