|Home | About | Journals | Submit | Contact Us | Français|
MicroRNAs (miRNAs) play essential roles in many physiological and pathological processes, including tumor development, by regulating the expression of a plethora of mRNAs. Although the importance of miRNAs in tumorigenesis is well established, only recently have reports elucidated miRNAs as promoters or suppressors of metastasis. The miR-200 family has been shown to inhibit the initiating step of metastasis, epithelial-mesenchymal transition (EMT), by maintaining the epithelial phenotype through direct targeting of transcriptional repressors of E-cadherin, ZEB1 and ZEB2. These findings shed light into a miRNA-mediated regulatory pathway that influences EMT in a developmentally and pathologically relevant setting.
Since the discovery of microRNAs (miRNAs) in mammals in 2001,1–3 there has been a tremendous surge in effort to catalogue and attribute biological functions to these small (~21–23 nt), non-coding RNA molecules. These efforts have reaped great rewards as our understanding of the role of miRNAs in normal physiology and pathology has increased significantly. Aside from participating in normal physiological processes such as muscle differentiation,4 miRNAs also function in the onset and/or progression of several pathologies such as Alzheimer’s disease,5,6 dilated cardiomyopathy,7 Huntington’s disease8 and cancer.9,10
MiRNAs are initially synthesized by RNA polymerase II as long primary transcripts, which are subsequently capped and polyadenylated.11 These transcripts are processed into ~70 nt stem-loop pre-miRNAs by Drosha RNase III endonuclease12 and are transported out of the nucleus by Ran-GTP/exportin 5.13 Pre-miRNAs are further processed by Dicer in the cytoplasm to yield a ~21–23 nt duplex.14 One strand of the duplex is incorporated into the RNA-induced silencing complex (RISC) and is used to regulate the expression of target genes. Binding of miRNAs to the 3' UTR of mRNAs with perfect or near-perfect complementary sequences induces mRNA degradation, whereas imperfect complementarity often induces translational repression. The seed sequence of miRNAs, representing 7–8 nt in the 5' end, is critical for efficient targeting and miRNAs harboring similar seed sequences can theoretically regulate the expression of a similar subset of genes.
In addition to deregulated expression of protein-coding genes, alterations in the expression of non-coding genes, such as miRNA genes, have also been documented as contributors to cancer pathology.9,15 More than 50% of human miRNAs are located in fragile chromosomal regions that are prone to mutations during tumor progression.16 In addition, recent expression profiling analyses have uncovered miRNA signatures that have the power to classify certain human cancers.17–19 More importantly, functional characterization has revealed a role for miRNAs as oncogenes (miR-21, miR-155, miR-17-92 cluster) or tumor-suppressor genes (miR-34a, let-7, miR-15a, miR-16) through the silencing of target tumor-suppressor or oncogenic protein-coding genes, respectively (Fig. 1). Interfering with miRNA processing also has been shown to enhance experimental tumorigenesis, further implicating the role of miRNAs in cancer.20 Although there is a clear role of miRNAs in oncogenesis, the contribution of miRNAs to malignant progression of human tumors has only recently been investigated and characterized.
Metastasis is a multi-step process by which cells from the primary tumor invade adjacent stroma, enter the systemic circulation, translocate and arrest at distal capillaries, extravasate and finally proliferate to form distant secondary tumors.21 Several recent reports have elucidated the role of certain miRNAs as promoters22–25 or suppressors26 of metastasis using a variety of approaches (Fig. 1). Tavazoie et al.26 performed array-based miRNA profiling of the MDA-MB-231 human breast cancer cell parental population as well as its highly bone- or lung-metastatic derivatives to uncover the role of miR-335 and miR-126 as metastasis suppressors in human breast cancer. In contrast, Ma et al.24 identified miR-10b as a pro-metastatic miRNA of breast cancer by functionally testing miRNAs deregulated in breast cancer previously reported by Iorio et al.27 In a recent study, Huang et al.23 identified miR-373 and miR-520c as pro-metastasis miRNAs by applying a forward genetic screen to search for miRNAs that can induce migration and invasion in the non-migratory and non-metastatic MCF-7 cell line. It is clear from these studies and others that miRNAs may influence multiple steps of metastatic cascade, such as tumor cell migration,23,24,26 invasion22–26 and intravasation.22,24 However, the initiating, and perhaps the most critical step in malignant conversion of tumor cells is thought to be the activation of a cellular program in tumor cells that drives epithelial-mesenchymal transition (EMT). EMT is a process whereby epithelial tumor cells are stimulated by extracellular cytokines, such as TGFβ, or intracellular cues, such as oncogenic Ras, to lose their epithelial polarity and gain mesenchymal phenotypes with increased migratory and invasive capabilities.28,29 One of the molecular hallmarks driving this transition is the functional loss of E-cadherin (CDH1), a cell adhesion protein and a major constituent of adherens junctions thought to be a suppressor of migration/invasion during carcinoma progression.30 The characterization of different modes of E-cadherin regulation, such as epigenetic silencing and transcriptional repression, during tumor progression has enhanced our understanding of the mechanisms that drive malignancy. In light of the importance of E-cadherin as a suppressor of metastatic progression, elucidating how miRNAs regulate epithelial cell plasticity and ultimately EMT during tumor progression may hold tremendous therapeutic potential in reducing the incidence of metastasis. However, since the aforementioned studies primarily used mesenchymal-like cell lines—which may have already undergone the EMT process—for their initial miRNA screens, it is unlikely that the identified miRNAs are important regulators of EMT.
During metazoan embryogenesis, EMT plays a key role in the formation of various tissues and organs such as the neural crest, heart, musculoskeletal and peripheral nervous systems. However, during adult life, only a certain subset of cells retain the ability to undergo EMT-keratinocytes, for example, during wound healing. Epithelial tumor cells often activate latent embryonic programs such as EMT to acquire malignant properties, including enhanced motility and invasiveness during tumor progression (Fig. 1). Several mouse models of breast cancer progression have implicated EMT as a promoter for metastasis through enhancing motility and invasive properties.31–33
The EMT process requires a complex genetic program coordinated in a large part by the nuclear translocation of several transcriptional repressors of CDH1. The characterization of the zinc-finger factor snail (SNAI1) as a transcriptional repressor of CDH1 and an inducer of EMT34,35 partially uncovered the molecular mechanisms that govern tumor invasion.36 Since then, many other transcriptional repressors of CDH1 have been discovered and because of their potential importance in tumor malignancies, there is a tremendous effort in characterizing their molecular function.29
ZEB1 (ZFHX1A/TCF8/δEF1/Nil-2α) and ZEB2 (ZFHX1B/SIP1), members of the ZEB family, are such CDH1 repressors that have been implicated in EMT, tumorigenesis and metastasis.37–41 ZEB1 and ZEB2 can interact with DNA binding sites composed of bipartite E-boxes (CACCT and CACCTG).42 The CDH1 promoter contains such E-box sequences and it has been shown that ectopic expression of ZEB factors in mammary gland epithelial cells43 or MDCK cells,44 is sufficient to induce dissociation of adherens junctions43,44 and enhance invasiveness/motility respectively.44 Both ZEB factors appear to have common tissue expression patterns during development and seemingly have redundant roles,45 although some differences in tissue expression45 and functionality have been reported.36,46
The ZEB factors can regulate the expression of various EMT- and tumor-related genes. For example, they have been shown to repress the expression of genes encoding proteins critical to maintaining the epithelial phenotype such as E-cadherin, plakophilin 2 and ZO3.47 Conversely, the ZEB factors can also activate the expression of genes promoting migratory/invasive phenotypes, such as the pro-invasion gene MMP2.36,48
Although a great deal is known of the signaling pathways that regulate the expression of the EMT-promoting snail superfamily members SNAI1 and SNAI2 (also know as slug), little information is available regarding the specific regulation of the ZEB factors during tumor progression.36 It has been elucidated that SNAI1 is an activator of ZEB1,49 and that SNAI1 can indirectly regulate the expression of ZEB2 by inducing the expression of a natural antisense transcript of ZEB2 that promotes the translation of its own mRNA during SNAI1-induced EMT50 (Fig. 1). In addition to Snail proteins, miRNAs may also serve to regulate the expression of the ZEB factors to influence the epithelial phenotype (Fig. 1). In fact, a compilation of data from a few studies suggests a negative correlation between the expression of miR-200s with that of the ZEB factors, suggesting a miR-200 mediated targeting of ZEB factors in embryonic tissues.18,51,52
The miR-200 family consists of five members organized as two clusters, miRs-200b/a/429 and miRs-200c/141, on chromosomes 1 and 12 in humans and 4 and 6 in mice. Analysis of the miR-200 expression levels in various cell lines in our laboratory suggests that miRNAs from each cluster are co-expressed whereas the expression of miRNAs from the two clusters do not appear to be highly correlated (Korpal M et al., unpublished result).53,54 The five miR-200 family miRNAs contain very similar seed sequences, with the seed sequence of miR-200b/c/429, AAUACU, differing from the seed sequence of miR-200a/141, AACACU, by only one nucleotide. Although target prediction algorithms predict a significant difference in the spectrum of genes targeted by miR-200b/c/429 and miR-200a/141, microarray analyses of cells transfected with members of these subfamilies indicate a much higher degree of overlap in target genes, suggesting that multiple members of the miR-200 family may target a large common subset of genes to enhance the efficiency of genetic regulation.54
Several recent developmental studies show an enrichment of the miR-200 family in differentiated epithelial tissues. Choi et al.55 showed that the miR-200 family is required for the differentiation of olfactory progenitor cells in the zebrafish and mouse. Similarly, an independent report demonstrated that this family is highly expressed in skin epidermal cells during mouse skin morphogenesis.56 In support, other reports have found that miR-200a and miR-200b are expressed in the ectoderm and the endoderm but largely excluded from the mesoderm during chick embryo development.57 In zebrafish embryos, miR-200a, miR-200b and miR-141 are expressed in the epidermis, proctodeum, the lateral line organs, and the sensory epithelial structures that can sense chemicals such as nose epithelium and taste buds, as assessed by in situ hybridization.58
These data strongly imply an underlying correlation between the miR-200 family and the epithelial phenotype, as well as a negative correlation with ZEB factors during embryonic development. Collectively, these studies imply that the miR-200 family might regulate CDH1 expression through targeting of the ZEB factors (Fig. 1). Two reports in 2007 functionally linked miR-200b and miR-200c with the epithelial phenotype through targeting of ZEB1/ZEB2 in a developmental setting59 and in cancer cell lines.59,60 Christoffersen et al.59 showed that miR-200b and ZEB2 show overlapping expression patterns in the mouse brain and in vitro assays revealed that miR-200b targets ZEB2 through binding to multiple sites in the 3' UTR. In addition, Hurteau et al.60 reported the targeting of ZEB1 by miR-200c in cancer cell lines and showed that ectopic expression of miR-200c enhanced CDH1 expression and promoted an epithelial-like morphology. Although these two reports were the first to show targeting of ZEB factors by the miR-200 family, they only partially uncovered the full potential of this family of miRNAs to regulate the epithelial phenotype and the underlying clinical implications.
Three recent reports overwhelmingly linked the miR-200 family with the epithelial phenotype, EMT, and its inverse process, mesenchymal-epithelial transition (MET).53,54,61 Inhibition of endogenous miR-200 expression levels was sufficient to induce EMT54,61 and ectopic expression induced MET in normal and cancer cell lines through direct targeting of ZEB1/253,54,61 and reduced motility and aggressiveness of breast carcinoma cells.53,54 Interestingly, ectopic expression of even single miR-200 members was sufficient to target ZEB1/ZEB2 and induce CDH1 expression, albeit with different efficacies.53,54 Park et al.54 reported a significant correlation between the expression of miR-200 and the E-cadherin/Vimentin ratio across all NCI60 cells, suggesting that the expression levels of the miR-200 family is a powerful, universal regulator of the epithelial phenotype of cancer cells. In clinical patient samples, there is a significant correlation between E-cadherin and miR-200c expression in primary serous papillary ovarian tumors54 and the expression of miR-200 family is significantly lower in sarcomatoid metaplastic breast tumors as compared to the more epithelial ductal breast carcinomas.61 These data collectively suggest that the miR-200 family may serve as a potential therapeutic target to reduce the rates of invasive carcinomas.
Since many diseased states have been correlated with deregulated expression of various miRNAs, developing strategies to manipulate miRNA expression levels in vivo may hold considerable potential in the treatment of previously intractable diseases. Methods of silencing miRNA expression in vivo via antagomirs62 and miRNA sponges63 or overexpressing miRNAs using viral/non-viral vectors, in vivo have been tested but need to be optimized.
Other strategies for influencing miR-200 expression in tumors in vivo are to manipulate the activity of transcription factors (TFs) that regulate the expression of these miRNA genes. Previous studies have shown that miRNA promoters can express protein coding reporter genes and it has been shown that a translocation event forced MYC to be expressed from a miRNA promoter in an aggressive B-cell leukemia.64 Taken together, these data suggest that regulation of miRNA expression may involve many of the same TFs that regulate the expression of protein-encoding genes. A recent report by Ma et al.24 showed that Twist, another CDH1 suppressor, could induce the expression of miR-10b by binding directly to its promoter. Interestingly, Burk et al.65 recently showed that ZEB1, a target of the miR200 family, could bind directly to the promoter and suppress the expression of miR-141 and miR-200c, triggering a feed forward loop that stabilizes EMT and promotes invasion of cancer cells. In addition to binding sites for ZEB1, the promoter element also contains Snail binding sites and overexpression of SNAI1 reduces expression of miR-141 and miR-200c, although to a lesser extent65 (Fig. 1). This suggests that multiple EMT factors may coordinately reduce miR-200 family expression and disrupt the epithelial phenotype, thereby driving the EMT process.
Human tumors upregulate the normally cytostatic cytokine TGFβ late in tumor progression, enhancing the expression of EMT inducers such as the ZEB factors, to promote invasiveness. Since the miR-200 family members are repressed during TGFβ-induced EMT in NMuMG53 and MDCK cells,61 it is likely that enhanced TGFβ production by primary tumors may reduce expression of the miR-200 family by enhancing the expression of ZEB factors, thus promoting EMT and invasiveness (Fig. 1). Indeed, TGFβ mRNA is highly expressed in NCI60 cell lines that express high levels of ZEB1/2 and vimentin.54 In light of this, the simplest therapeutic strategy may be to disrupt TGFβ-signaling activity in advanced stage tumors to reduce the likelihood of EMT occurring through the action of the miR-200 family in tumor cells in vivo.
The recently described role of the miR-200 family in regulating EMT enhances our understanding of the regulatory pathways influencing this key developmental and pathological process, opening up new avenues for therapeutic intervention in patients with localized primary lesions. Therapeutic targeting of the miR-200 family may be achieved at many levels, the simplest strategy of which being disrupting TGFβ-signaling activity using various inhibitors currently in preclinical and/or clinical trials.
This work is supported by the US Army Medical Research and Material Command (W81XWH-06-1-0481) with additional support from the American Cancer Society (RSG MGO-110765), the Susan Kormen for the Cure (BCTR0503765). MK is supported by a pre-doctoral fellowships from the Natural Sciences and Engineering Research Council of Canada and the Department of Defense. We thank Dr. Yong Wei, Nilay Sethi and Mario A. Blanco for their helpful suggestions and discussion.