SEARCH

SEARCH BY CITATION

Keywords:

  • microRNAs;
  • lncRNAs;
  • EZH2;
  • cancer;
  • chromatin modifications

Abstract

  1. Top of page
  2. Abstract
  3. MicroRNAs Biology
  4. MiRNAs–EZH2 Interactions
  5. LncRNAs Biology
  6. LncRNAs–EZH2 Interactions
  7. Conclusions
  8. Acknowledgments
  9. References

A large amount of data indicates that non-coding RNAs represent more than the “dark matter” of the genome. Both microRNAs and long non-coding RNAs are involved in several fundamental biologic processes, and their deregulation may lead in oncogenesis. Interacting with the Polycomb-repressive complex 2 subunit EZH2, they could affect the expression of protein-coding genes and form feedback networks and autoregulatory loops. They can also form networks with upstream and downstream important factors, in which EZH2 represent the stabilizing factor of the pathway. As such non-coding RNAs affect the epigenetic modifications leading to malignant transformation.

Histone modifications represent one of the main epigenetic changes affecting biological processes which can be epigenetically inherited.[1] Histone modifications may affect chromatin structure by changing amino acid charge, thus unfolding chromatin, whereas they can also bind to non-histone proteins via specific domains. Histone modifications have the ability to separate the genome in euchromatin and heterochromatin and to enhance gene transcription, DNA replication and repair and chromosome condensation.[2] They play an essential role in normal development and disruption of the histone modification machinery results in developmental defects, early embryonic lethality, cancer, neurodevelopmental disorders, neurodegenerative and neurological disorders and autoimmune diseases.[3] Nevertheless, the connection between the density of CpG islands and histone modifications suggests a dependence of histone modifications on the DNA sequence.[5] Changes in histone modifications appear in early stages of tumorigenesis and accumulate during cancer development indicating their role in the malignant transformation process. Global loss of monoacetylation and trimethylation of histone H4 might represent a common characteristic of cancer cells resembling the global DNA hypomethylation and the CpG hypermethylation.[6] Histone methyltransferases (HMTs) are key components of the epigenetic machinery, and cancer cells are characterized by deregulated expression of the epigenetic machinery including aberrant expression of HMTs, demethylases, acetyltransferases and deacetylases.[7] Moreover, genetic alterations of the histone-modifying enzymes have been observed in a tissue-dependent manner with important differences between hematological malignancies and solid cancers.[8]

The Polycomb group proteins (PcGs) represent important regulators of epigenetic changes and are well conserved among species. PcGs are capable of binding and repressing promoters of protein-coding target genes involved in cell development and differentiation of several cell types.[9] PcG can maintain both normal and cancer stem cell populations by repressing the expression of tumor suppressors involved in fundamental signaling pathways, and their inhibition could enhance the activity of chemotherapeutic drugs in cancer subtypes.[10-12] PcGs form multiprotein complexes: the Polycomb-repressive complexes PRC1 and PRC2. The two complexes regulate chromatin structure mainly through post-translational modification of histones such as monoubiquitylation of lysine 119 of histone H2A (H2AK119ub) and methylation of lysine 27 of histone H3 (H3K27me2/3), respectively. The PRC2 complex comprises several components including the core components: enhancer of zeste 1 (EZH1), enhancer of zeste 2 (EZH2), embryonic ectoderm development (EED) and suppressor of zeste 12 (SUZ12). It also comprises Rbap46/48, PCL1, PCL2, PCL3, JARID2 and AEBP2, which are cofactors and PRC2 enzymatic activity modulators. PRC2 is capable of establishing and restoring histone methylation through the action of its components EZH2 and EZH1, respectively.[13] EZH2 is a HMT necessary for the formation of the H3K27me3 mark resulting in the consequent recruitment of the PRC1 complex, and its depletion results in embryonic death from gastrulation defects.[14] EZH2 is involved in cell lineage commitment, and its expression levels are high in embryonic stem cells (ESCs) and decrease upon terminal cell differentiation. By interacting with downstream and upstream important factors, EZH2 is involved in osteogenesis, myogenenesis, neurogenesis, hematopoiesis, adipogenesis and hepatogenesis.[15] The interplay between EZH2 and DNA methyltransferases (DNMTs) links H3K27 and CpG methylation leading to DNA hypermethylation and consequent silencing of genes that have previously acquired H3K27me3. EZH2 also recruits histone deacetylases again resulting in gene transcriptional repression in cancer cells.[16] In fact, this interaction between the epigenetic machineries might result in progressive recruitment of DNMTs to PcG target genes resulting in transcriptional silencing therefore promoting tumorigenesis. EZH2 has a dual role in oncogenesis acting either as an oncogene or as a tumor suppressor. Moreover, EZH2 is abnormally overexpressed during disease initiation, at the level of cancer stem cell and progression in a vast number of cancer types.[17] In cancer, EZH2 expression can be regulated at transcriptional, post-transcriptional and post-translational levels by transcription factors, oncogenic fusion genes and chromatin repressors and through phosphorylation in a cell-cycle-dependent signaling. In turn, EZH2 directly targets and modulates several factors involved in oncogenesis.[18] Ectopic expression of EZH2 in murine models affected the hematopoietic stem cell (HSC) population and their repopulating ability, altered gene expression programs in HSC, aberrantly produced mature myeloid cells and compromised hematopoietic function with final result the development of myeloproliferative disease suggesting a stem cell-specific oncogenic role of EZH2 in myeloid malignancies.[19] EZH2 gain of function mutations have also been observed in a subset of BCL2-rearrenged non-Hodgkin lymphomas (NHLs) arguing in favor of its oncogenic function.[20] On the other hand, loss-of-function mutations have been reported in myelodysplastic syndromes (MDS) and other myeloid malignancies, suggesting that EZH2 acts as a tumor suppressor in myelopoiesis in contrast to the previous findings.[19] Most recently, loss-of-function somatic mutations of EZH2 were observed in T-cell acute lymphoblastic leukemia (T-ALL) indicating their role in disease pathogenesis. It was also demonstrated that EZH2 mutations resulted in upregulation of NOTCH1 target genes. Interestingly, it was observed that NOTCH1-mediated oncogenic transformation is associated with the loss of H3K27me3 mark on NOTCH1 target gene promoters and not in the whole T-ALL genome indicating that loss of the specific histone mark characterizes the oncogenic function of NOTCH1 in T-ALL.[23]

A rapidly growing body of data describes the importance of EZH2 in cancer biology, whereas most recent reports illustrate the interplay between EZH2 and non-coding RNAs. This review represents an effort to accumulate the current knowledge on the topic highlighting the contribution of EZH2 in pathogenetic pathways.

MicroRNAs Biology

  1. Top of page
  2. Abstract
  3. MicroRNAs Biology
  4. MiRNAs–EZH2 Interactions
  5. LncRNAs Biology
  6. LncRNAs–EZH2 Interactions
  7. Conclusions
  8. Acknowledgments
  9. References

MicroRNAs (miRNAs) are small ncRNAs ∼22 nt long and are distributed in all human chromosomes except for the Y chromosome. They are found in clusters and are mainly transcribed as polycistronic primary transcripts. Their biogenesis is a multistep process involving a complex protein network, and an extensive description of their biogenesis is beyond the scope of this review. Briefly, the circuit of proteins involved in the canonical biogenesis pathway comprise the RNA polymerase II Drosha, the cofactor DGCR8, the nuclear transport receptor family member exportin5, the protein Dicer with cleavage properties and the Argonaute protein. Alternative biogenesis pathways could be mainly Drosha/DGCR8 independent or Dicer independent.[24] They are involved in fundamental biological processes such as development, differentiation, proliferation and apoptosis. Through their oncogenic or tumor-suppressive function, they are involved in oncogenesis of solid and blood cancers.[26-29] MiRNAs also play important role in the chromatin structure control by regulation at the post-transcriptional level of the chromatin modifying enzymes including PcG subunits.[30]

MiRNAs–EZH2 Interactions

  1. Top of page
  2. Abstract
  3. MicroRNAs Biology
  4. MiRNAs–EZH2 Interactions
  5. LncRNAs Biology
  6. LncRNAs–EZH2 Interactions
  7. Conclusions
  8. Acknowledgments
  9. References

The first evidence of miRNAs and EZH2 interaction was reported by Varambally et al.[32] who reported that miR-101 decreased expression correlated with an increase in EZH2 expression and H3K27me3 levels contributing to increased proliferation rates and invasion ability of prostate cancer cells, and disease progression. MiR-101 by affecting EZH2 expression in both transcript and protein levels contributed to the epigenetic modifications observed in cancer.[32] MiR-205 was also found to have a causative role in prostate cancer pathogenesis as it was able to induce morphologic changes of the cell and enhanced migratory and invasive properties of prostate cancer cells. It also affected epithelial to mesenchymal transition, and inversely correlated with EZH2, suggesting a tumor suppressor role of miR-205 probably mediated by EZH2.[33] MiR-101 and miR-26a targeted the 3′-UTR of EZH2, and their expression was decreased in prostate cancer cell lines. Ectopic expression of both miRNAs repressed EZH2 mRNA translation and H3K27 methylation affecting cell proliferation and invasion. Proteins located upstream of miR-101 like the hypoxia inducible factor (HIF-1 b) and androgen receptor (AR) were evaluated as miR-101 regulators. Prostate cancer cell lines treated with agents that induce HIF-1 a expression exhibited an increase of HIF-1 a/HIF-1 b expression, concomitant increase of EZH2 expression and decreased expression of miR-101. As such HIF, AR, mir-101 and EZH2 compose a HIF-1 a/HIF-1 b/AR-miR-101-EZH2 signaling pathway in prostate cancer pathogenesis and progression.[34]

A recent study established the coordinated downregulation between the histone demethylase KDM2B and EZH2 in mouse embryonic fibroblasts undergoing senescence. Enforced expression of KDM2B epigenetically repressed the expression of the tumor suppressor miRNAs let-7 b and mir-101 through direct demethylation of H3K36. Silencing of these specific miRNAs resulted in increased EZH2 expression and contributed to the immortalization of cells. Thus, deregulation of the KDM2B-let-7 b/miR-101-EZH2 axis could find application in the pathogenesis of cancer by modulating growth control and could represent a potential therapeutic target.[35] An inverse correlation between the increased expression of EZH2 and let-7 family members was also observed in prostate cancer cells.[36] Autonomous or fibroblast growth factor-2 (FGF-2) driven upregulation of the histone demethylase NDY1-induced chromatin modifications and led to miR-101 repression. Moreover, FGF-2 and vascular endothelial growth factor (VEGF) downregulated miR-101 expression and induced EZH2 by reproducing the chromatin modifications induced by NDY1 on miR-101. EZH2 attributed a causative role in the FGF-2-driven cell proliferation and migration. It was demonstrated that FGF-2 upregulated NDY1, which has the ability to recruit EZH2. In turn, EZH2 repressed miR-101 composing the FGF-2-NDY1/EZH2-miR-101-EZH2 axis, which is active both in normal cell proliferation and in human cancer proliferation and in survival, migration and angiogenesis. EZH2 is upregulated through post-transcriptional mechanisms by NDY1 through the repression of miR-101. NDY1 recruited EZH2, and their cooperation was essential for the silencing of miR-101, whereas either NDY1 or EZH2 could achieve miR-101 silencing when acting alone. This elegant work established the existence of a feedback regulatory network between miR-101 and EZH2, with EZH2 being the stabilizing factor of the pathway.[37]

In an effort to better understand the pathogenesis of an aggressive cancer as the squamous cell carcinoma of the head and neck (HNSCC), Banjeree et al.[38] evaluated the interaction between the tumor suppressor rap1GAP, miR-101, and EZH2. EZH2 mRNA and protein levels were increased in human HNSCC cells compared with normal keratinocytes and induced cell proliferation and invasion. The tumor suppressor rap1GAP involved in the biology of HNSCC represents an EZH2 negatively regulated gene at both mRNA and protein levels. MiR-101 was downregulated in the cases in which EZH2 expression was upregulated and rap1GAP was silenced. Induced overexpression of miR-101 led to reversion of EZH2 and rap1GAP expression. EZH2 upregulation enhanced H3K27me3 methylation on rap1GAP promoter resulting in its repression, and also promoted CpG promoter hypermethylation. This study identified a miR-101-EZH2-rap1GAP network and documented the ability of PRC2 subunits to act as DNMTs recruiter enhancing CpG methylation.[38] MiR-138 decreased expression in HNSCC cell lines promoted epithelial to mesenchymal transition. MiR-138 deregulated expression inversely correlated with EZH2 expression which in turn was associated with a decreased expression of the downstream target E-cadherin involved in epithelial to mesenchymal transition.[39] In a similar way, Carvalho et al.[40] reported the existence of a miR-101-EZH2-E-cadherin pathway preferentially associated with the intestinal type of gastric cancer.

EZH2 represented a downstream target of the underexpressed miR-101 also in epithelial ovarian cancer. Moreover, the deregulated ovarian cancer tumor suppressor p21waf1/cip1 promoter interacted with EZH2 and that association was inhibited when miR-101 expression was restored. Thus, miR-101 inhibits tumor growth by indirect regulation of p21waf1/cip1 via down-regulation of EZH2.[41] MiR-101 was downregulated in hepatoma cell lines and exhibited a drug-induced overexpression, which is protein kinase c-PKCa dependent. MiR-101 enforced expression resulted in downregulation of EZH2 and H3K27 methylation also in a PKCa-dependent pattern. Interestingly, the use of the cell cycle inhibitor TPA (12-O-tetradecanoylphorbol 13-acetate)-induced expression of miR-101, which was under the ERK signaling pathway control.[42] EZH2 expression levels were increased and negatively associated with miR-101 and miR-128 a expression in bladder transitional cell carcinoma, gastric cancer, glioblastomas, non-small cell lung cancer samples and cell lines, intraductal papillary mucinous neoplasm of the pancreas and adult T-cell leukemia/lymphoma cells (ATL).[43-47]

EZH2 overexpression is also associated with a more aggressive type of nasopharyngeal carcinoma (NPC), and is inversely associated with the expression of three downregulated miRNAs; miR-26a, miR-98 and miR-101. In fact, exogenous expression of the three miRNAs resulted in decreased EZH2 transcript levels acting as negative regulators of EZH2.[48] MiR-26a downregulation in NPC cell lines and tumor samples was also verified by others. MiR-26a-forced expression resulted in EZH2 dose-dependent downregulation at both mRNA and protein levels, suggesting a post-transcriptional regulation. MiR-26a cell growth inhibition and cell cycle arrest effects were in part mediated by the EZH2 repression. Moreover, a subset of cell cycle regulators including cyclin E2, cyclin D3, cyclin-dependent kinase 4 and 6 and cell-cycle inhibitors p14ARF, p16INK4A were indirectly regulated by miR-26a in a EZH2-dependent manner, suggesting a role of miR-26a and EZH2 in cell cycle regulation.[49] The same tumor suppressor effect of miR-26a in EZH2 expression was observed in breast cancer, pancreatic cancer, lung cancer and rhabdyosarcoma. That interaction affected cell apoptosis, cell proliferation, migration, invasion and cancer stem cell function.[50-53] MYC-dependent repression of miR-26a has been established in murine lymphoma cell lines and in human Burkitt lymphoma cell lines. There was also a difference in miR-26a expression between chronic lymphocytic leukemia and Burkitt samples suggesting the MYC-dependence of miR-26a. Its expression inversely correlated with EZH2 expression, while induced miR-26a led to over-representation of several EZH2 target genes supporting the existence of a MYC-miR-26a-EZH2-target genes axis in MYC-driven lymphomas.[54] The same effect of EZH2 dependence on MYC via miR-26a has also been suggested in prostate cancer samples.[55]

MiR-214 was found downregulated in human breast cancer cell lines, and its expression inversely correlated with EZH2 expression at both mRNA and protein levels in a dose-dependent manner affecting cell proliferation. Interestingly, in some cases miR-214 underexpression was not associated with a concomitant decreased miR-101 expression, suggesting that the former miRNA may affect EZH2 expression in a miR-101 independent way.[56]

Knockdown of Dicer, the key protein in miRNAs biogenesis, resulted in increased EZH2 expression, suggesting a generalized EZH2 control by tumor suppressor miRNAs. In addition, EZH2 negatively regulated several miRNAs as their expression was restored after EZH2 knock down in prostate and breast cancer cell lines. EZH2 regulated miRNAs expression through epigenetic repression by inducing H3K27me at the specific loci. These miRNAs seemed to possess tumor suppressor functions and they modulated in vitro cancer cell invasiveness, tumor growth and a cancer stem cell phenotype by acting on PRC1 proteins. Most importantly, the EZH2-regulated miRNAs (miR-181 a/b, miR-200 a/b/c and miR-203) inhibited the expression of two PRC1 member proteins suggesting the existence of a feedback loop. The higher expression of EZH2 and the corresponding lower EZH2-regulated miRNAs levels in metastatic prostate cancer correlated with the high levels of PRC1 expression levels. These data suggested the existence of a miRNAs-EZH2-miRNAs-PRC1 axis in advanced prostate cancer.[57]

The synthetic derivative of curcumin diflourinated-curcumin CDF with known antitumoral activity is capable of inhibiting EZH2 expression and the cell growth and migration in pancreatic cancer cell lines and induced the expression of let-7 a/b/c/d, miR-26a, miR-101, miR-146a and miR-200b. Moreover, a direct interaction between these miRNAs and EZH2 was established as EZH2 knock down with siRNA resulted in re-expression of the mRNAs. Moreover, miR-101 overexpression led to decreased expression of EZH2 and transcription factors, suggesting an important role in the regulation of cancer stem cell marker genes in pancreatic cancer cells.[58]

MiR-31 was the most downregulated miRNA in primary ATL cells among the 59 miRNAs evaluated. MiR-31 by binding to 3′-UTR of NF-κB induced kinase (NIK) negatively regulated NIK function, and therefore negatively correlated with NF-κB activity. Thus, miR-31 plays an important role in the negative regulation of the non-canonical NF-κB pathway through the modulation of NIK. The final result of the forced miR-31 expression and the subsequent NF-κB inhibition was the suppressed ATL cell growth and the enhanced apoptosis. MiR-31 levels inversely correlated with EZH2, H3K27me3 and H3K9me2 levels. Moreover, the transcription factor YY1 regulated EZH2 localization, increased HeK27me3 levels and initiated miR-31 suppression. This interaction indicates that histone methylation may be causative factor for the loss of miR-31 in ATL. EZH2 knockdown resulted in increased apoptosis of leukemic cells while additional expression of NIK inhibited this phenomenon suggesting the interaction between PRC2 and NF-κB pathway. These data suggest the existence of a YY1/EZH2/SUZ12-miR-31-NIK- NF-κB pathway at least in ATL and breast cancer and also highlight the contribution of epigenetic modifications in the oncogenic activation of NF-κB.[59]

The putative tumor suppressor miR-124 presented lower levels in hepatocellular carcinoma (HCC) compared to normal cells and enhanced the expression of EZH2 mRNA. That interaction affected cell growth, migration, epithelial to mesenchymal transition, invasion and metastasis in HCC (Fig. 1).[60]

LncRNAs Biology

  1. Top of page
  2. Abstract
  3. MicroRNAs Biology
  4. MiRNAs–EZH2 Interactions
  5. LncRNAs Biology
  6. LncRNAs–EZH2 Interactions
  7. Conclusions
  8. Acknowledgments
  9. References

Long non-coding RNAs (lncRNAs) are more than 200 nucleotides long and encode no proteins. They can induce local gene silencing in cis or in trans throughout the genome. They also bind chromatin modification complexes, recruit chromatin modifying factors and act as scaffolds for chromatin modifiers altering histone marks.[61] LncRNAs are also involved in genomic reprogramming and in the formation of nuclear structures. They are crucial for normal tissue maintenance and function, and they are aberrantly expressed in several cancer subtypes.[62-64] Several lncRNAs are essential for maintaining the pluripotent state by being downstream targets of the Oct4 and Nanog transcription factors, while others are necessary for repressing linage-specific gene expression and are enriched for EZH2 and SUZ12 subunits in mouse ESCs.[65] LncRNAs may themselves be epigenetically regulated through EZH2-driven H3K27 promoter methylation, proposing a transcription regulation mechanism of lncRNAs.[67] LncRNAs are in physical association with miRNAs given that miRNAs loci are found in coding and non-coding regions of the genome as well.[68] LncRNAs have the ability of cross-talking with miRNAs forming regulatory networks within the transcriptome[69] and some also possess enhancer-like functions affecting the expression of neighboring coding genes.[70] An interesting example of interaction between lncRNAs and miRNAs is represented by the DLK1-MEG3 imprinted domain, which contains the tumor suppressor lncRNAs (MEG3) and several miRNAs involved in cancer biology.[71] Moreover, miR-29 was experimentally found to target the DNMTs, and therefore indirectly modulated MEG3 expression in hepatocellular carcinoma.[72] Furthermore, miRNAs hosted in that Dlk1-Dio3 imprinted domain target three genes of the PRC2 complex forming a feedback regulatory loop between protein coding and ncRNAs of the cluster in fully pluripotent stem cells. These unique features of the Dlk1-Dio3 imprinted domain may have clinical application given that modulation of these miRNAs by pharmacological means may lead in inhibition of the abnormally overexpressed EZH2 in cancer and also suggest that miRNAs hosted among the lncRNAs of the imprinted domain are key regulators of stem cell gene expression forming feedback regulatory networks.[73]

Zhao et al., using the RIP-seq method in murine ESCs, demonstrated that the imprinted lncRNA Meg3 directly and specifically binds PRC2, and most probably recruits the EZH2 subunit to Dlk1 affecting its expression, providing evidence that lncRNAs may control in cis the expression of protein coding genes within imprinted domains.[74]

It has also been demonstrated that the lncRNA Xist interacts with EZH2 and with YY1, which acts as a docking protein and traps the Xist-EZH2 complex, resulting in the X chromosome inactivation arguing in favor of the cis acting nature of Xist.[75]

Recent extensive reviews on the biology, function, involvement in disease, and targeting mechanisms of lncRNAs have been published,[76] and a census of cancer-related lncRNAs, expression patterns and function in cancer has been published as well.[78]

LncRNAs–EZH2 Interactions

  1. Top of page
  2. Abstract
  3. MicroRNAs Biology
  4. MiRNAs–EZH2 Interactions
  5. LncRNAs Biology
  6. LncRNAs–EZH2 Interactions
  7. Conclusions
  8. Acknowledgments
  9. References

An increasing amount of evidence suggests that lncRNAs are key players in cancer biology exhibiting oncogenic or tumor suppressive functions or both. The cancer-specific expression pattern of lncRNAs highlights their potential use as cancer diagnostic biomarkers in clinical practice. LncRNA-based anti-cancer therapies utilizing either the interference RNA technology or the synthesis of small molecules that interact with the lncRNAs are being developed. LncRNAs may also harbor single nucleotide polymorphisms, and genome-wide studies may identify subjects at risk of cancer development, response to treatment and outcome.[79] Oncogenic lncRNAs interact with the epigenetic machinery modulating important cellular pathways leading to oncogenesis. Furthermore, antisense lncRNAs can induce epigenetic silencing of their sense protein-coding partners or they can control protein-coding genes at the post-transcriptional level resulting in malignant transformation. Otherwise, oncogenic lncRNAs can induce the expression of other protoncogenes involved in malignant transformation and metastasis. Tumor suppressor lncRNAs can be induced by p53, DNA damage or oncogenic stress and disruption of their expression can lead to malignant transformation.[81]

Less data are available regarding the association between lncRNAs and EZH2. They interact upstream and downstream with important transcription factors as in the case of the lncRNA P21 associated ncRNA DNA damage activated (PANDA) which is induced by p53, while it inhibits apoptotic gene expression by direct sequestration of the transcription factor NF-YA from occupying target genes promoters.[83]

Long intergenic ncRNAs (lincRNAs) act as molecular scaffolds, providing binding surfaces to histone modifying enzymes, and therefore, specify the pattern of histone modifications on target genes. LincRNAs can coordinate histone modifications by binding to multiple histone-modifying enzymes.[84] This could be the case of HOX transcript antisense intergenic RNA (HOTAIR), which was first demonstrated to mediate epigenetic silencing in trans through direct or indirect interaction with PRC2 subunits.[85] HOTAIR is capable of inducing H3K27me3 and suppressing the HOXD locus and other target genes as directly targets EZH2 subunit. HOTAIR expression levels were higher in colorectal cancerous tissue samples compared to non-cancerous counterparts, and the higher expression corresponded to a more aggressive phenotype with a less differentiated histology, liver metastasis and a reduced overall prognosis. Importantly, it was established that HOTAIR overexpression induced EZH2 localization in colorectal cancer cells, and gene signatures with HOTAIR-induced EZH2 occupancy were significantly enriched in colorectal cancer cells. Thus, HOTAIR overexpression is associated with PRC2 targeting, and their cooperation regulates the expression of target genes in colorectal cancer.[86] The mechanism through which HOTAIR affects PRC2 occupancy to target genes throughout the whole genome is not completely clarified but it seems that HOTAIR might nucleate Polycomb domains. However, the HOTAIR-chromatin interaction is associated with PRC2 relocalization and target gene silencing, and its function of chromatin binding seems to be EZH2-independent.[87] HOTAIR expression was increased in metastatic breast cancer lines and human primary breast cancer samples, and was associated with increased metastasis rates and a poor overall prognosis as an independent prognostic factor. HOTAIR promoted colony growth and cell invasion in vitro while promoted lung metastasis in a xenograft model. HOTAIR overexpression enhanced EZH2 occupancy, and H3K27me3 and epigenetically regulated several target genes involved in breast cancer progression inhibition especially in aggressive breast tumors. Interestingly, EZH2 knockdown reversed the ability of HOTAIR to promote cell invasion, although EZH2 depletion itself did not affect cell invasiveness. These findings suggested that EZH2 subunit is necessary for HOTAIR to promote invasiveness. Nevertheless, EZH2 upregulation promoted invasion but concomitant depletion of HOTAIR inhibited the ability of EZH2 to induce cell invasion suggesting a functional interplay between PRC2 and HOTAIR in promoting cancer invasiveness. HOTAIR also modulated and repressed PRC2 target genes expression, while these target genes were re-expressed when PRC2 was depleted.[88] A higher HOTAIR expression was also verified in pancreatic cancer samples and correlated with more advanced disease stages, and overall survival. In the same study it was demonstrated that HOTAIR induced gene suppression in a EZH2-dependent and EZH2-independent mechanisms as well, and it might regulate different sets of genes in pancreatic cancer in contrast to breast cancer.[89]

Yang et al.[90] identified 174 lncRNAs aberrantly expressed in HBV + HCC. In particular, the lncRNA-high expression in HCC (lncRNA-HEIH) was overexpressed in HCC and cirrhotic tissues compared to healthy liver tissues. LncRNA-HEIH overexpression was more profound especially in cases of cancer recurrence, conferred a poorer prognosis, and was highly connected to two other lncRNAs and 16 protein coding genes that were all involved in tumor growth and chemoresistance. The lncRNA-HEIH was experimentally demonstrated to interact directly with EZH2, and that interaction was essential for the repression of EZH2-regulated target genes. The lncRNA-HEIH overexpression resulted in decreased expression of important cell cycle regulators such as p16, while the suppressed expression of these genes by lncRNA-HEIH was reversed when EZH2 was downregulated. However, it seemed that the lncRNA-HEIH expression level was not the only necessary condition determining the association of lncRNA-HEIH and EZH2 for cell cycle regulation, suggesting the role of other still unidentified factors.[90]

The recently discovered lncRNA PCAT-1 was upregulated in high-grade localized and metastatic prostate cancer samples. PCAT-1 and EZH2 expression were nearly mutually exclusive. EZH2 pharmacologic and shRNA inhibition resulted in an increase of PCAT-1 expression levels in cell lines indicating that is an EZH2 target. Interestingly, EZH2 knockdown resulted in a downregulation of PCAT-1 target genes arguing in favor of its transcription repressor function. Moreover, PCAT-1 target genes were up-regulated upon PCAT-1 knockdown, while EZH2 knockdown indirectly downregulated PCAT-1 target genes due to PCAT-1 upregulation. These findings support the notion of the EZH2-PCAT-1 gene regulation.[91]

The antisense lncRNA ANRIL which overlaps the INK4b/ARF/INK4a locus was upregulated in prostate cancer cells compared to normal prostate cells and correlated with higher levels of EZH2 occupancy near the promoters of INK4a and ARF genes. There was established an interaction between EZH2-induced H3K27 methylation and ANRIL-based silencing of the INK4b/ARF/INK4a locus by the PRC1 CBX7 subunit. These findings suggested that ANRIL lncRNA recruits both PRC2 and PRC1 subunits (EZH2 and CBX7, respectively) to achieve INK4b/ARF/INK4a silencing, and that lncRNAs directly contributes in the epigenetic transcriptional repression observed in cancer.[92]

Conclusions

  1. Top of page
  2. Abstract
  3. MicroRNAs Biology
  4. MiRNAs–EZH2 Interactions
  5. LncRNAs Biology
  6. LncRNAs–EZH2 Interactions
  7. Conclusions
  8. Acknowledgments
  9. References

MiRNAs and lncRNAs are involved in cellular homeostasis affecting cell proliferation, cell growth, differentiation and apoptosis, while their disruption leads in malignant transformation of the cells. EZH2 is a core PRC2 subunit, has fundamental role in the epigenetic modifications and is involved in oncogenesis. Several miRNAs bind directly to the 3′-UTR of EZH2 through which they exert their function. MiRNAs are capable of regulating EZH2 expression at both transcriptional and post-transcriptional levels. Through the interaction with EZH2 they affect H3K27 methylation and in consequence they modulate cell biology. MiRNAs and EZH2 form important regulatory and feedback pathways in which EZH2 plays the role of the pathway stabilizing factor. LncRNAs and EZH2 are in tight association affecting each other's expression and the expression of target genes. Given that lncRNAs demonstrate promoter CpG methylation[82] it would be interest to evaluate whether EZH2 recruits DNMTs to lncRNAs promoters in a lncRNA-lncRNA functional circuitry. The interaction of EZH2 and ncRNAs defines the way through which ncRNAs affect chromatin structure, protein coding genes expression, and modulate pathways contributing to oncogenesis. It has been demonstrated that lncRNAs act as endogenous decoys for miRNAs affecting their distribution on their target genes. LncRNAs deregulated expression modulated the expression of miRNAs target mRNA in a direct manner, whereas miRNAs expression inversely correlated with target mRNA expression.[93] Again, it could be hypothesized that EZH2 has a mediator role in this interplay, and if an EZH2 role exists and is experimentally verified, epigenetic modifications will also affect the competing endogenous RNA world.

image

Figure 1. Schematic representation of inhibitory/activating interactions between EZH2, microRNAs and other upstream and downstream factors. EZH2 is under the negative control of several miRNAs and it also negatively affects other miRNAs itself. In particular, EZH2 forms feedback loops with miR-101 and miR-26a which are under the negative control of HIF-1 a/b and MYC respectively. EZH2 inhibits cell cycle regulators, tumor suppressor genes (rap1GAP), and molecules involved in epithelial to mesenchymal transition like E-cadherin. For a more detailed description see text.

Download figure to PowerPoint

References

  1. Top of page
  2. Abstract
  3. MicroRNAs Biology
  4. MiRNAs–EZH2 Interactions
  5. LncRNAs Biology
  6. LncRNAs–EZH2 Interactions
  7. Conclusions
  8. Acknowledgments
  9. References