Long noncoding RNA CRNDE promotes colorectal cancer cell proliferation via epigenetically silencing DUSP5/CDKN1A expression

Evidence indicates that long non-coding RNAs (lncRNAs) play a critical role in the regulation of tumor cellular processes, such as proliferation, apoptosis, and metastasis. LncRNA CRNDE (Colorectal Neoplasia Differentially Expressed) is located at human chromosome 16 and has been found overexpressed in a variety of cancers including colorectal cancer (CRC). In this paper, we report that lncRNA CRNDE expression was remarkably upregulated in CRC tissues and that lncRNA CRNDE overexpression was positively correlated with advanced pathological stages and larger tumor sizes. In addition, the knockdown of CRNDE significantly suppressed proliferation and caused apoptosis of CRC cells both in vitro and in vivo. Furthermore, RNA immunoprecipitation and chromatin immunoprecipitation assays demonstrated that lncRNA CRNDE could epigenetically suppress the expressions of dual-specificity phosphatase 5 (DUSP5) and CDKN1A by binding to EZH2 (the key components of Polycomb repressive complex 2 (PRC2)), thus promoting CRC development. In conclusion, our data suggest that the lncRNA CRNDE promotes the progression of CRC and is a potential therapeutic target for CRC intervention.

Although considerable headway has been achieved in the diagnosis and therapy of colorectal cancer (CRC) in the past decade, it still causes significant morbidity and mortality, especially in the developed world. [1][2][3] According to the latest cancer statistics, CRC is the third most common cancer and third leading cause of cancer death worldwide. 2 Unfortunately, a great majority of the patients are diagnosed at an advanced stage that has passed the optimal time for a radical operation. 4 As is known to all, CRC carcinogenesis is a complicated biological process involving multiple genomic variation and biological processes. 5 Therefore, further study of the molecular mechanisms involved in the development and progression of CRC is essential to gain the targeted strategies to ease the burden of the disease.
With the development of whole-genome sequencing technology, it has gradually been revealed that protein-coding genes constitute only small portions (∼2%) of the genome encode proteins, whereas the vast majority of the human genome is transcribed into noncoding RNAs including micro-RNA genes, long noncoding RNA (lncRNA) genes, and pseudogenes. 6,7 Recently, lncRNAs as the new class of noncoding RNA molecules have become a hot topic because of their abnormal expression in human diseases, especially in tumor. [8][9][10][11] The role of lncRNAs in tumorigenesis and development has emerged from the water and been identified as the future diagnostic, therapeutic, or prognostic cancer biomarkers, including for CRC. [12][13][14] For instance, lncRNA-linc00152 downregulated by miR-376c-3p restricts viability and promotes apoptosis of colorectal cancer cells; 15 a high expression of lncRNA-NEAT1 (nuclear paraspeckle assembly transcript 1) predicted an unfavorable prognosis in patients with digestive system malignancies. 16 Our previous studies also demonstrated that lncRNA-HOXA-AS2 (HOXA cluster antisense RNA 2) represses CDKN1A and KLF2 expression transcription by binding with EZH2 (enhancer of zeste homolog 2) and LSD1 in CRC; 17 lncRNA-LOC554202 could induce apoptosis via the caspase cleavage cascades in CRC cells; 18 lncRNA-HOTTIP (HOXA transcript at the distal tip) promotes colorectal cancer growth partially via silencing of CDKN1A expression. 19 LncRNA CRNDE (Colorectal Neoplasia Differentially Expressed) is located at human chromosome 16 and has been found overexpressed in a variety of cancers including CRC. [20][21][22] A large number of studies have shown that CRNDE plays an important role in the occurrence and development of CRC, indicating a poor prognosis. 23 The upregulation of CRNDE expression has been shown to promote cell proliferation, migration, and invasion while inhibiting apoptosis of CRC. 24,25 Nevertheless, the underlying molecular mechanisms involved in CRNDE interactions are complex and diverse. It may function as a competing endogenous RNA to promote CRC metastasis and oxaliplatin resistance by sponging miR-136, 25 or via miR-181a-5p-mediated regulation of Wnt/ β-catenin signaling promotes colorectal cancer cell proliferation and chemoresistance. 24 However, the above studies have singly pointed that CRNDE plays an oncogene function by regulating miRNA. The more effective and complex mechanisms of CRNDE deserve further study.
Given the importance of CRNDE in CRC, in the present study we investigated that CRNDE expression was remarkably upregulated in CRC tissues and that CRNDE overexpression was positively correlated with advanced pathological stages and larger tumor sizes. In addition, the knockdown of CRNDE significantly suppressed proliferation and caused apoptosis of CRC cells both in vitro and in vivo. Furthermore, RNA immunoprecipitation and chromatin immunoprecipitation assays demonstrated that lncRNA CRNDE could epigenetically suppress the expressions of DUSP5 (a negative regulator of mitogen-activated protein kinase (MAPK) signaling pathway) and CDKN1A by binding to EZH2 (the key components of Polycomb repressive complex 2 (PRC2)), thus promoting CRC development. In conclusion, our data suggest that the lncRNA CRNDE promotes the progression of CRC and is a potential therapeutic target for CRC intervention.

Results
CRNDE is upregulated in human CRC tissues and is positively correlated with larger tumor size, advanced TNM stage, and lymph node metastasis. We first analyzed the expression level of the CRNDE in human CRC tissues by using microarray data downloaded from the Gene Expression Omnibus (GEO; GSE21510), and found that the CRNDE expression level was significantly upregulated in CRC tissues compared with that in normal tissues ( Figure 1a). The results of the CRC database downloaded from TCGA are consistent with the above findings (Supplementary Figure S1). Then, we examined CRNDE expression levels in 80 CRC tissue samples and matched adjacent normal tissues by performing quantitative real-time PCR (qRT-PCR) analysis. Similarly, this showed that CRNDE expression was significantly upregulated in CRC tissues compared with that in their normal counterparts ( Figure 1b). Increased CRNDE expression was also validated in CRC cells, including DLD1, HCT116, SW480, and LOVO cells (Po0.05) (Figure 1c). To assess the significance of CRNDE expression in CRC, the clinicopathological features are summarized in Table 1. As shown in Figures 1d and e, upregulated CRNDE expression is closely related to increased tumor size and advanced TNM stage in the CRC patients. These findings imply that CRNDE may function as an oncogene involved in the tumorigenesis of CRC.
SiRNA selection for CRNDE silencing or overexpression in CRC cells. To observe the functional relevance of CRNDE in CRC cells, we chose the DLD1 and HCT116 cell lines for further studies because they exhibited the highest expression levels compared with that in the normal colon epithelium cell line (HCoEpiC) (Po0.05) (Figure 1c). Next, we designed three different CRNDE short interfering RNAs (siRNAs) and one overexpression plasmid for transfection into the above two CRC cell lines. The qRT-PCR analysis was used to detect the transfection efficiency (Figures 1e and f). Of the three siRNAs, si-CRNDE 1 and 2 exhibited the better silencing capacity. Meanwhile, CRNDE expression was increased in all two CRC cell lines compared with the negative control following transfection with pcDNA-CRNDE. Thus, we selected si-CRNDE 1, 2, and pcDNA-CRNDE in all further experiments for CRNDE silencing or overexpressing, respectively.
CRNDE promotes CRC cell proliferation in vitro. To further determine the potential biological role of CRNDE in CRC cell proliferation, we performed a MTT assay. The results showed that the growth vitality of DLD1 and HCT116 cells transfected with siRNA-CRNDE 1 or 2 was significantly inhibited compared with control groups. In contrast, CRNDE overexpression promoted the cell growth capacity (Po0.05, Figure 2a). The colony formation assays were performed to detect the tumorigenicity of CRC cells. The results suggest that the CRC cell colony numbers significantly decreased following knockdown of CRNDE but increased in CRC cells overexpressing CRNDE (Po0.05, Figures 2b and c). Taken together, these findings indicate that CRNDE can markedly promote the proliferation of CRC cells.
Downregulation of CRNDE promotes G1 arrest and causes apoptosis in CRC cells in vitro. To assess whether the effects of CRNDE on CRC cell proliferation are the result of CRNDE-mediated changes in cell cycle progression, we examined cell cycling in CRC cells by flow cytometry. The results of si-CRNDE 1, 2, or si-NC transfection for 48 h showed that CRNDE knockdown increased the percentage of cells in G0/G1 phase and decreased the percentage of cells in S and G2/M phase compared with control cells (Po0.05) ( Figure 3a). As shown in Figure 3b, overexpressing CRNDE promoted an obvious reduction in the G0/G1 phase, with a significant arrest in the number of cells in the S phase (Po0.05). Consistent with this finding, the results of western blot analysis showed that the protein levels of CDK2 (cyclin-dependent kinase 2) and CDK6 were significantly changed in si-CRNDE-treated cells (Figures 3c and d), confirming that CRNDE is involved in cell-cycle regulation. Moreover, we performed flow cytometric analysis to characterize the level of apoptosis of CRC cells that also reflect CRC cell growth vitality. The results demonstrated that the silencing of CRNDE expression significantly increased in CRC cell apoptosis (Po0.05, Figure 3e). These investigations suggest that CRNDE exerts critical influences on CRC cells by affecting both the cell cycle and apoptosis.

Knockdown of CRNDE inhibits CRC tumorigenesis in vivo.
To determine whether the level of CRNDE expression could affect CRC tumorigenesis in vivo, sh-CRNDE or negative control-transfected HCT116 cells were subcutaneously inoculated into male nude mice. At 15 days after injection, the tumors formed in the sh-CRNDE group were dramatically smaller than that in the negative control group (Po0.05) (Figure 4a). Correspondingly, the tumor volumes and weights were obviously decreased compared with the controls (Po0.05) (Figures 4b and c). Moreover, the qRT-PCR analysis confirmed that the expression level of CRNDE remained markedly reduced in the sh-CRNDE group (Po0.05) (Figure 4d). The immunohistochemistry These results indicate that CRNDE is significantly associated with CRC cell proliferation in vivo.

CRNDE knockdown increases the expression of genes involved in cell proliferation.
To obtain unbiased findings on the CRNDE-associated pathway, we assessed the gene expression profiles of CRC cells in which CRNDE expression was suppressed. Specifically, we performed RNA transcriptome sequencing from control and CRNDE-depleted DLD1 cells. DLD1 cells were treated with si-NC or si-CRNDE 1 for 48 h. To further study the involved pathways activated by CRNDE, we analyzed the associated genes using data collected from the Gene Ontology (GO) database ( Figure 5a). The most prominent GO categories were related to cell proliferation, suggesting that these biological processes are particularly affected in CRNDE-knockdown cells (Figure 5b). The GO results were essentially in agreement with our experimental findings. The qRT-PCR was then used to confirm the changes in the levels of several upregulated or downregulated mRNAs involved in cell proliferation. The results showed that knockdown of CRNDE increased DUSP5, CDKN1A, EMP1, RAET1E, and CDKN2B expression, but decreased RPL23A, EIF4E2,QPRT, ACSS1, and APCDD1 expression (Figure 5c). To observe the key downstream mediators of CRNDE in CRC cells, we chose the DUAP5 and CDKN1A mRNAs for further studies because they exhibited the highest expression levels in CRNDEknockdown cells. The qRT-PCR analysis confirmed that the expression level of DUAP5 and CDKN1A remained markedly increased in the si-CRNDE group (Po0.05) (Figure 5d). In addition, western blot analysis demonstrated that DUAP5 and   CRNDE epigenetically silences DUSP5 and CDNK1A transcription by binding with EZH2. Although CRNDE exhibits vital biological functions in CRC, the underlying molecular mechanisms involved in its interactions are complex and diverse. CRNDE is previously found to be overexpressed in CRC and promotes its cell proliferation via acting as endogenous competing RNAs (ceRNAs) for miRNAs. However, its other molecular mechanism and downstream targets involved in regulation of CRC cell phenotype are not known.
To investigate the potential mechanism and downstream targets of CRNDE in CRC cells, we first analyzed its distribution in CRC cells and found that CRNDE is mostly located in nucleus (Po0.05) (Figure 6a), implying that CRNDE may be involved in transcriptional regulation. Then, we predicted the possible binding proteins of CRNDE by literature and found that it could bind to histone modification enzymes. Recently, several studies have concluded that ∼ 20% of lncRNAs can regulate downstream target genes by binding with PRC2. PRC2 is a methyltransferase that trimethylates H3K27 to suppress the transcription of specific genes. The RIP assay was performed to further investigate whether CRNDE could bind to PRC2. The results showed that CRNDE could interact with EZH2 and SUZ12 (suppressor of zeste 12 homolog) (the two key components of PRC2) in DLD1 and HCT116 cells (Figure 6b). The RNA pull-down experiments also confirmed CRNDE could interact with EZH2 in DLD1 cell line (Figure 6c). Our findings indicate that CRNDE may epigenetically inhibit downstream target genes by binding to EZH2.
Furthermore, we found that the expression levels of DUSP5 and CDKN1A were increased in si-EZH2-treated CRC cells (Po0.05) (Figure 6d), implying DUSP5 and CDKN1A may be the key downstream genes of CRNDE, and CRNDE can inhibit its expression by binding to EZH2. The results of chromatin immunoprecipitation (ChIP) assays showed that EZH2 could directly bind to DUSP5 and CDKN1A promoter regions and induce the histone H3 lysine 27 trimethylation (H3K27me3) modification in DLD1 and HCT116 cells. Knockdown of CRNDE resulted in reduced EZH2 binding and H3K27me3 occupancy of the DUSP5 and CDKN1A promoter (Po0.01)  (Figure 7a). We performed MTT and colony formation assays to determine cell viability. The results revealed that the cell vitality of DLD1 and HCT116 cells transfected with pcDNA-DUSP5 was significantly inhibited compared with empty vector-treated groups (Po0.05) (Figures 7b and c). Then, the flow cytometric analysis demonstrated that the upregulation of DUSP5 expression significantly increased in CRC cell apoptosis (Po0.05, Figure 7d). These data suggest that DUSP5 inhibited the proliferation and promoted apoptosis of CRC cells.
DUSP5 silencing potentially involves the oncogenic function of CRNDE. To validate whether CRNDE regulates CRC cell proliferation by silencing DUSP5 expression, rescue assays were performed. DLD1 and HCT116 cells were cotransfected with pcDNA-DUSP5 and pcDNA-CRNDE, and the MTT and colony formation assay results indicated that pcDNA-DUSP5 and pcDNA-CRNDE co-transfection partially rescues pcDNA-CRNDE-promoted proliferation ability (Po0.05) (Figures 8a-c). These results indicated that the effect of CRNDE on CRC partially involves targeting DUSP5.

Discussion
LncRNAs are broadly defined as endogenous cellular RNAs molecules longer than 200 nt in length with limited or no protein-coding capacity and capable of regulating gene expression 26 at various levels, including chromatin modification, transcription, and post-transcriptional processing. 9,27,28 Recently, more evidence has emerged that dysregulation of lncRNAs may be an important class of pervasive genes and the future diagnostic, therapeutic, or prognostic cancer biomarkers involved in carcinogenesis and development. 13,29 For instance, our previous studies showed that lncRNA SPRY4-IT1 (SPRY4 intronic transcript 1) could be a growth regulator, and promoted the proliferation of estrogen receptor (− ) human is an unfavorable prognostic factor and regulates proliferation and apoptosis in colorectal cancer. 31 LncRNA CRNDE was long considered to an oncogene in the occurrence and development of CRC, indicating a poor prognosis. 20,23 The underlying molecular mechanisms involved in CRNDE interactions are complex and diverse. Recent studies have revealed the multilayered biological function of CRNDE cell proliferation, metastasis, and chemoresistance by regulating miRNAs. 24,25 Against this background, there is an urgent need to identify CRNDE and investigate its more effective and complex mechanisms. Here, we revealed a novel mechanism and thoroughfare of CRNDE in CRC cells.
In this study, we found that the expression of CRNDE in CRC tissues was highest by the network data, suggesting that CRNDE plays a crucial role in CRC and may be an independent clinical marker in CRC prognosis and therapy. Then, we identified this result by tissue validation and found that CRNDE overexpression was positively correlated with advanced pathological stages and larger tumor sizes. In addition, the knockdown of CRNDE significantly suppressed proliferation of CRC cells both in vitro and in vivo by causing G1 arrest and promoting apoptosis.
We analyzed the downstream target genes of CRNDE by microarray analysis, and the data showed that DUSP5 and CDKN1A might be the downstream mediators via affecting the cell proliferation pathways. DUSP5 is a negative regulator of MAPK signaling pathway and has recently been identified as a tumor suppressor in several human malignancies, including CRC. [32][33][34][35] Recent studies found that the importance of feedback mechanism is illustrated by the fact that expression of oncogenic BRAFV600E, a feedback-insensitive mutant RAF kinase, reprograms DUSP5 into a cell-wide ERK inhibitor that facilitates cell proliferation and transformation. 36 CDKN1A (p21), one of the CDK inhibitor, is expressed ubiquitously and plays multifarious roles in many cellular processes. 37 Our previous studies have also shown that CDKN1A is a critical molecule for inhibiting cell proliferation in CRC cells. 17 Here, we also demonstrate that DUSP5 can act as a tumor suppressor and is silenced by CRNDE in CRC cells. Meanwhile, CRNDE can also suppress the expression of CDKN1A.
Lots of evidence suggested that the importance of lncRNAs in human cancer may be associated with their abilities to impact cellular functions through regulating target gene expression through binding to various histone modification enzymes. 29,38,39 PRC2, which consists of EZH2, SUZ12, and EED (embryonic ectoderm development), is a typical methyltransferase for histone H3K27me3. 40 For example, lncRNA HOXA11-AS could promote proliferation and invasion of  29 long intergenic noncoding RNA 00511 acts as an oncogene in non-small-cell lung cancer by binding to EZH2 and suppressing p57. 41 To further grope for the potential molecular mechanisms involved, we performed the nucleocytoplasmic separation experiments and found that CRNDE RNA is mainly distributed in the nucleus, indicating that CRNDE may exert transcriptional regulation function. The RIP and RNA pull-down assays revealed that CRNDE could directly bind with EZH2 in DLD1 and HCT116 cells to silence DUSP5 and CDKN1A expression. The results of ChIP analysis demonstrated that EZH2 could directly bind to DUSP5 and CDKN1A promoter regions and induce H3K27me3 modification in CRC cells. These results revealed that CRNDE promotes CRC cell proliferation is a manner that is dependent on the regulation of DUSP5 and CDKN1A expression by binding to EZH2.
In conclusion, our study provides the evidence that CRNDE was upregulated in CRC tissues and its high expression level may be associated with CRC patients' progression of disease. The promotion role of CRNDE on CRC cell proliferation and tumorigenesis may partly via epigenetically restraining DUSP5 and CDKN1A transcription by binding with EZH2 and inhibiting EZH2-mediated methylation modification of these target genes. CRNDE may serve as a potential target for diagnosis and treatment in human CRC.

Materials and Methods
Differential expression analysis. CRC gene expression data were downloaded from the GEO data set. The independent data set from GSE21510 (http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc = GSE21510) was included in this study and normalized using the Robust Multichip Average. After we had downloaded probe sequences from GEO or the microarray manufacturers, Blast +2.2.30 was used to reannotate the probes in the GENCODE Release 20 sequence databases (http://www.gencodegenes.org/) for lncRNA.
Tissue collection and ethics statement. A total of 80 paired tumor tissues and matched normal tissues were collected from CRC patients who received surgical treatment between January 2011 and December 2014 at Second Affiliated Hospital of Nanjing Medical (China). None of the patients received any local or systemic treatment before surgery. The pathological stage, grade, and nodal status were appraised by an experienced pathologist. All experiments were approved by the research ethics committee of Nanjing Medical University, China. Written informed consent was obtained from all patients. The animal experiments were performed with the approval of the institutional committee for animal research and conformed to national guidelines for the care and use of laboratory animals.
Cell lines and culture conditions. The human CRC cell lines (DLD1, HCT116, SW480, and LOVO cells) and the human colonic epithelial cells HCoEpiC were purchased from American Type Culture Collection (Manassas, VA, USA). The cells were cultured in Dulbecco's modified Eagle's medium (DMEM), (Invitrogen, Grand Island, NY, USA) in humidified air at 37°C with 5% CO 2 . All media were supplemented with 10% fetal bovine serum (FBS), 100 U/ml penicillin, and 100 mg/ml streptomycin (Invitrogen).
RNA extraction and qRT-PCR analyses. Total RNA was extracted from tissues or cultured cells using TRIZOL reagent (Invitrogen). For qRT-PCR, RNA was     Tumor formation assay in a nude mouse model. Male athymic BALB/ c nude mice (4 weeks old) were maintained under pathogen-free conditions and manipulated in accordance with the protocols approved by the Shanghai Medical Experimental Animal Care Commission. HCT116 cells were stably transfected with sh-CRNDE or empty vector, harvested from 6-well cell culture plates, washed with PBS, and resuspended at a concentration of 1 ×108 cells/ml. A total of 100 μl of suspended cells was subcutaneously injected into a single side of the posterior flank of each mouse. Tumor growth was examined every 3 days, and tumor volumes were calculated using the following formula: 0.5 × length × width 2 . At 2 weeks after cell injection, the mice were killed. The subcutaneous weight of each tumor was measured and the tumors were used for further analysis. The protocol was approved by the committee on the ethics of animal experiments of the Nanjing Medical University.
IHC analysis. Tumor tissue samples were immunostained for H&E and Ki-67 as described previously. 17 The expression was considered to be positive when ⩾ 50% cancer cells were stained.
Microarray data analysis. RNA extraction and purification, RNA amplification and labeling, array hybridization, and data acquisition were performed as described in the Affymetrix (Santa Clara, CA, USA) manufacturer's instructions. Affymetrix Human U133 Plus 2.0 whole genome chips were used. Array hybridization and wash was performed using GeneChip Hybridization, Wash and Stain Kit (cat. no. 900720, Affymetrix) in Hybridization Oven 645 (cat. no. 00-0331-220V, Affymetrix) and Fluidics Station 450 (cat. no. 00-0079, Affymetrix) following the manufacturer's instructions. Slides were scanned by GeneChip Scanner 3000 (cat. no. 00-00212, Affymetrix) and Command Console Software 4.0 (Affymetrix) with default settings. Raw data were normalized by MAS 5.0 algorithm, Gene Spring Software 11.0 (Agilent Technologies, Santa Clara, CA, USA).
Subcellular fractionation location. The separation of nuclear and cytosolic fractions was performed using the PARIS Kit (Life Technologies, Carlsbad, CA, USA) according to the manufacturer's instructions.
RNA immunoprecipitation (RIP) assay. RIP experiments were performed using the Magna RIP RNA-Binding Protein Immunoprecipitation Kit (Millipore, Billerica, MA, USA) following the manufacturer's instructions. Antibody for RIP assays of EZH2, SUZ12, or control IgG were from Millipore.
RNA pull-down assays. CRNDE RNAs were in vitro transcribed using T7 RNA polymerase (Ambio Life, Austen, TX, USA) that were then purified using the RNeasy Plus Mini Kit (Qiagen) and treated with RNase-free DNase I (Qiagen). Transcribed RNAs were biotin-labeled with the Biotin RNA Labeling Mix (Ambio Life). Positive, negative, and biotinylated RNAs were mixed and incubated with DLD1 cell lysates. Magnetic beads were added to each binding reaction, followed by incubation at room temperature. Then, the beads were washed with washing buffer. The eluted proteins were detected by western blot analysis.
ChIP assay. CRC cells were treated with formaldehyde and incubated for 10 min to generate DNA-protein crosslinks. Cell lysates were then sonicated to generate chromatin fragments of 200 to 300 bp and immunoprecipitated with H3K27me3 and EZH2-specific antibody (CST) or IgG as control. Precipitated chromatin DNA was recovered and analyzed by qRT-PCR. The primer sequences used for the studies are presented in Supplementary Table S3.
Statistical analysis. All experiments were repeated at least three times with each sample in triplicate. Sample sizes for relevant experiment were determined by power analysis. All data were expressed as mean ± S.D. and analyzed using Student's t-test to compare two groups of in vitro and in vivo data using the SPSS 17.0 software program (version 17.0, SPSS Inc, Armonk, NY, USA). A value of Po0.05 was considered to be statistically significant.

Conflict of Interest
The authors declare no conflict of interest.