LncRNA DSCAM-AS1 promotes colorectal cancer progression by acting as a molecular sponge of miR-384 to modulate AKT3 expression

Down Syndrome Cell Adhesion Molecule antisense1 (DSCAM-AS1), a novel long non-coding RNA (lncRNA), reportedly contributes to the development and progression of several cancers. There is a lack of information on its biological role and regulatory mechanism with respect to colorectal cancer (CRC). Here, we discovered that the expression of DSCAM-AS1 exhibited a significant upregulation in CRC tissues and cell lines in comparison with the corresponding control. Increased DSCAM-AS1 expression was associated with poor prognosis for those diagnosed with CRC. Loss-of function assay illustrated that knockdown of DSCAM-AS1 resulted in significant inhibition of cell proliferation, invasion and migration in vitro, and impaired tumor growth in vivo. MicroRNA-384(miR-384) was directly targeted by DSCAM-AS1 in CRC cells, and repression of DSCAM-AS1 inhibited the expression of AKT3, a known target of miR-384 in CRC. In addition, repression of miR-384 or overexpression of AKT3 could partially rescue the inhibitory effect of DSCAM-AS1 knockdown on CRC progression. In summary, DSCAM-AS1 exerted an oncogenic role in CRC by functioning as a competing endogenous RNA of miR-384 to bring about regulation of AKT3 expression. These results implied that DSCAM-AS1 might be a novel therapeutic target for patients suffering from CRC.


INTRODUCTION
Colorectal cancer (CRC), a commonly diagnosed digestive malignant tumor, is the third highest reason for tumorassociated mortality around the world [1]. In spite of great efforts being made to develop effective strategies against CRC, the outcomes for patients is unsatisfactory [2,3]. Thus, studying the potential mechanisms involved in occurrence and progression of CRC is crucial to explore novel targets for diagnosis and treatment of this disease.
Spanning over 200 nucleotides in length, long non-coding RNA (lncRNA) are a class of RNA transcripts with little protein-coding potential [4]. Emerging evidence demonstrates that lncRNAs play a crucial regulatory role among various cellular processes, like cell proliferation, invasion, apoptosis and cycle [5,6]. Studies show that lncRNAs are associated with initiation and development of various cancers [7]. Many lncRNAs are verified to play tumor suppressor or oncogenic role in the progression of CRC [8,9], suggesting that lncRNAs could serve as a diagnosis marker and therapy agent.
Despite recently studies demonstrated that DSCAM-AS1 expression was upregulated in CRC and was involved in CRC proliferation and invasion [17,18], the function and underlying mechanism of DSCAM-AS1 in CRC progression remains largely unknown.
Studies show that LncRNAs can depict competing endogenous RNAs (ceRNAs) or natural microRNA (miRNA) sponges that bring about modulation of miRNAs [19,20]. MiRNAs (small non-coding RNAs: 18-25 nucleotides), bring about inhibition of translation or degradation of target mRNAs when it binds to the 3′untranslated regions (3′-UTR) of target genes [21]. MicroRNAs (miRNAs) was reported to play crucial roles in multiple cancer processes [22]. Nevertheless, it remains largely unclear whether DSCAM-AS1 can serve as ceRNA of miRNAs to regulate CRC progression.
In our study, analysis of DSCAM-AS1 expression in CRC tissues and its relationship with clinicopathologic characteristics of CRC patients was carried out. Functional roles of DSCAM-AS1 on CRC cell growth and metastasis were determined by numerous experiments. Moreover, the regulatory mechanism of DSCAM-AS1 in CRC was investigated by confirming whether it serves as a ceRNA of miRNA to modulate CRC progression.

Upregulation of DSCAM-AS1 and its correlation with poor prognosis in patients with CRC
The expression of DSCAM-AS1 in 56 CRC tissues and adjacent normal tissues was detected, and we found a significant increase in its expression in CRC tissues ( Figure 1A). We also detected the expression of DSCAM-AS1 in CRC cell lines and found that in 4 CRC-derived cell lines (LOVO, PKO, SW480 and HT29) DSCAM-AS1 expression was significantly higher than in the normal human colon epithelial cell line NCM460 ( Figure 1B).
To investigate the correlation between DSCAM-AS1 and clinicopathological features of patients with CRC, the 56 patients were split into two groups based on the median value: DSCAM-AS1 high group and DSCAM-AS1-low group. As shown in Table 1, high DSCAM-AS1 group was associated with advanced clinical stage and lymph node metastasis. In addition, Kaplan-Meier curve and log rank test showed that overall survival (OS) was significantly shorter in the high DSCAM-AS1 expression group relative to low DSCAM-AS1 group ( Figure 1C).

DSCAM-AS1 knockdown inhibited migration and invasion of CRC cells
The effect of DSCAM-AS1 knockdown on cell invasion was studied using transwell assay and its effect on migration was determined by the wound healing assay.  As seen in Figure 3A, down-regulation of DSCAM-AS1 markedly decreased cell migration ability of LOVO and HT29 cells. DSCAM-AS1 depletion also inhibited the cell invasion ability when compared to sh-NC group ( Figure 3B). These indicated that downregulation of DSCAM-AS1 suppressed CRC cell metastasis.

DSCAM-AS1 is a sponge for miR-384
Emerging studies reveal lncRNA could directly bind to miRNAs and function as a molecular sponge during tumorigenesis [19,20]. We further investigated the regulatory mechanism of DSCAM-AS1 in CRC. Through bioinformatics analysis using starBase V2 tool, we found that there are miR-384 binding sites in the DSCAM-AS1 sequence ( Figure 4A). To test this predication, luciferase reporter assay was carried out and we found that miR-384 over-expression resulted in a significant suppression of the activity of WT-DSCAM-AS1 reporter plasmid in LOVO and HT29 cells (Figure4B). Furthermore, RIP assay demonstrated that DSCAM-AS1 and miR-384 were remarkably clustered in Ago2 immunoprecipitate in comparison with the IgG-pellet, indicating they enriched in the same RNA-induced silencing complex (RISC) ( Figure 4C). Furthermore, qRT-PCR analysis showed that DSCAM-AS1 knockdown increased miR-384 expression in LOVO and HT29 cells (Figure4D), while overexpression of miR-384 inhibited DSCAM-AS1 expression in LOVO and HT29 cells ( Figure 4E). Moreover, we discovered that there was a downregulation of miR-384 expression in CRC tissues ( Figure 4F), and there was negative correlation with DSCAM-AS1 in CRC tissues ( Figure 4G).

Knockdown of DSCAM-AS1 impeded CRC tumor growth in nude mice
The tumor xenograft assay was done to study the impact of DSCAM-AS1 knockdown on CRC tumor growth in vivo. As seen in Figure 6A-6C, nude mice injected with sh-DSCAM-AS1/LOVO cells had smaller tumors, both in volume and weight, when compared to mice injected with sh-NC/LOVO cells. Moreover, we found that the Ki-67 positive cells were significantly decreased in sh-DSCAM-AS1/LOVO group when compared to sh-NC/LOVO group ( Figure 6D). We also found that DSCAM-AS1 and AKT3 expression was downregulated, while miR-384 was upregulated in xenograft tumor of sh-DSCAM-AS1/LOVO group when compared to sh-NC/LOVO group ( Figure 6E-6H). The in vivo

DISCUSSION
In recent decades, evidence has highlighted that the dysregulation of lncRNAs on CRC could be a leading cause for tumor process [8,9]. Studies show that DSCAM-AS1 has a role in progression of several cancers, and that it can function as an oncogenic lncRNA in these cancers [10][11][12][13][14][15][16]. Although recently studies demonstrated that DSCAM-AS1 expression was upregulated and played a crucial role in CRC, the functional roles and underlying mechanism of DSCAM-AS1 in CRC cells remains largely unknown [17,18]. We discovered that there is an upregulation of DSCAM-AS1 in CRC tissues and cell lines. Increased DSCAM-AS1 had a positive correlation with advanced clinical stage, lymph node metastasis and poor overall survival, indicating that it may act as an oncogene. Results of loss-of-function experiments revealed that the knockdown of DSCAM-AS1 in CRC cells can inhibit migration, cell proliferation, and invasion in vitro, as well as cause suppression of tumor growth in vivo, It is well known that lncRNAs function as ceRNAs and interact with miRNAs [19,20]. DSCAM-AS1 was reported to serve as a ceRNA for sponging several miRNAs to regulate the development of tumors [10,13,15,17]. For example, it was reported that DSCAM-AS1 promotes proliferation and decreases apoptosis of breast cancers cells by regulating miR-204-5p/RRM2 axis [15]. Huang et al. revealed the association between DSCAM-AS1 and poor clinical prognosis, and that it contributes to promoting melanoma progression by sponging miR-136 [13]. Ma et al. indicated that DSCAM-AS1 served as a ceRNA of miR-137 and caused regulation of EPS8, bringing about promotion of cell reproduction and suppression of cell apoptosis in tamoxifen resistance breast cancer cells [24]. Through bioinformatics analysis, miR-384 was found to be a potential target of DSCAM-AS1. Previous studies reported that miR-384 functioned as a tumor suppressor in multiple human malignancies [25][26][27]. In CRC, miR-384 expression was significantly downregulated, and overexpression of miR-384 suppressed the CRC growth and metastasis [23,28]. Our study exhibited the regulatory relationship between DSCAM-AS1 and miR-384 through luciferase reporter activity and RIP assays. Our study also indicated that miR-384 expression was decreased in CRC tissues, which was consistent with previous results [23,28]. In addition, there was a negative correlation between miR-384 expression and DSCAM-AS1 in CRC. It is important to note that miR-384 inhibitor partially reversed the inhibitory effect caused by DSCAM-AS1 knockdown in CRC cells. Therefore we suggest that DSCAM-AS1 promoted CRC progression by sponging for miR-384. Previous studies showed that DSCAM-AS1 could sponge miR-216b and miR-144-5p in CRC [17,18]. Linked with our results, DSCAM-AS1 could sponge multiple miRNAs in CRC progression.
Various examples in literature have demonstrated that lncRNA functioned as a ceRNA of miRNA that brings about modulation of depression of miRNA's target gene expression [29]. Previous study showed that AKT3 is targeted by miR-384 in CRC cells [23]. Here, we showed that in CRC cells, down-regulation of DSCAM-AS1 significantly reduced AKT3 expression, while overexpression of AKT3 or downregulation of miR-384 could reserve this trend. Furthermore, AKT3 expression was upregulated, and its expression was positive correlated with DSCAM-AS1 and negative correlated with miR-384 in CRC tissues. Overexpression of AKT3 in CRC cells reversed the effect caused by knockdown of DSCAM-AS1 on proliferation and invasion. These results implied that DSCAM-AS1 acts as a ceRNA of miR-384, bringing about modulation of AKT3 expression, thereby promoting the progression of CRC.
In summary, we identified that lncRNA DSCAM-AS1 is linked to tumor metastasis and poor OS of CRC patients. DSCAM-AS1 promoted the progression of CRC by playing the role of a ceRNA of miR-384 to modulate AKT3 expression. Thus, DSCAM-AS1 has the potential to be a diagnosis marker as well as a therapeutic target for CRC.

Clinical tissues
Harvesting of 56 clinical CRC samples and adjoining normal tissues were carried out at Department of Gastrointestinal Colorectal and Anal Surgery, China-Japan Union Hospital of Jilin University between the time period January 2011 to January 2012. None of the patients had received any anti-tumor therapy before surgery. Written informed consent was acquired from the patients. Research approval was got from "Ethics Committee of Jilin University (Changchun, China).

Cell culture and transfection
Normal human colon epithelial cell line NCM460 and four human CRC cell lines (LOVO, PKO, SW480 and HT29) were bought from American Type Culture Collection (ATCC; Manassas, VA, USA). Along with supplementation of 10% fetal bovine serum (FBS, Gibco; MA, USA), cells were grown in Dulbecco's modified Eagle's medium (DMEM; Gibco) containing with 100 U/ml penicillin (Invitrogen, CA, USA) and 1 μg/ml streptomycin (Invitrogen). All cells were maintained in a humidified incubator at 37 °C with 5% CO2.
AKT3 overexpression plasmid (pCDNA-3.1) was granted from Dr Li (Jilin University) was transfected into CRC cells using lipofectamine 3000 (Invitrogen) as per company protocol. Transfection efficiency was examined using real time quantitative PCR (qRT-PCR) at 48 h after transfection.

RNA extraction, reverse transcription and quantitative PCR
Total RNA was extracted with TRIzol reagent (Invitrogen) from CRC tissues and cell lines. The quality and concentration of RNA were assessed at 260/280 nm by the use of a Nanodrop Spectrophotometer (ND-2000, Thermo, USA). With a Prime Script Kit, reverse-transcription of RNA samples(1μg) into cDNA was carried out (Takara, China). Quantitative PCR reactions were done with SYBR Green PCR Kit (Roche, Germany) under the ABI Prism 7500 system (Applied Biosystems, USA). GAPDH was used for normalization of DSCAM-AS1 and AKT3 mRNA, while U6 was used for normalization of miR-384. All primers sequences are shown in Table 2 [12,23]. 2 −ΔΔCt method was used to examine gene expression levels [30].

Wound healing assay
Cell migration determination was carried out by this assay according to a previous a study [31]. Briefly,

Transwell invasion assay
Following trypsinization, cells were seeded with Matrigel-coated (BD Biosciences, USA) transwell filters in a 24-well plate (50,000 cells/well) in serumfree medium. The lower chambers contained medium with 20% FBS. Following 24h incubation, fixation of invaded CRC cells were carried out with methanol for 30 min followed by staining for a duration of 15 min with 0.1% crystal violet. The number of invasive cells was determined by counting the stained cells at five fields selected by random with an inverted microscope (Olympus, Japan) at 200 × magnification.

Bioinformatic analysis and luciferase reporter assay
StarBase2.0 was used to predict the potential binding sites of miRNAs on DSCAM-AS1 [32]. Generation of Wildtype (WT) DSCAM-AS1 with potential miR-384 binding sites were inserted into a luciferase reporter vector psi-CHECK-2 (Promega, USA). In addition, a site-directed mutagenesis kit (Tiangen, Beijing, China) was used to generate a mutated form of this vector termed MT-DSCAM-AS1.Co-transfection of CRC cells with luciferase plasmids and miR-384 mimics or miR-NC were carried out followed by culturing for 48 h. Dual-Luciferase Reporter Assay System (Promega) was used to determine the luciferase activities, following the manufacturer's instructions.

RNA immunoprecipitation (RIP) assay
RNA-Binding Protein Immunoprecipitation Kit (Millipore, USA) was used as per manufacturer's instructions. LOVO and HT29 cells were lysed and treated with RIP buffer with magnetic beads that had undergone conjugation with human anti-human argonaute 2 (Ago2) antibody (Millipore, USA). Mouse IgG (Millipore) was used for negative control. Immunoprecipitated RNA was extracted using TRIzol reagent after the protein was digested using Proteinase K buffer. qRT-PCR was carried out for detection of DSCAM-AS1 and miR-384 as above-mentioned.

Established of xenograft model
All animal experimental protocols and surgical procedures were approved by the "Animal Care and Use Committee of Jilin University (Changchun, China)". Ten male athymic nude mice (5-week-old, 18-20g) were obtained from the "Laboratory Animal Center of Jilin University (Changchun, China)". A total of 2 × 10 6 LOVO cells that had been stably transfected using either sh-DSCAM-AS1 or sh-NC were implanted subcutaneously into 5-week old nude BALB/c mice (n = 5/group). Using the formula below, tumor volumes were determined every 5 days: Tumor volume = (Length × Width 2 )/2. The mice were sacrificed after a duration of 30 days and the tumors were removed and weighed for further studies.

Immunohistochemistry (IHC)
The tumors tissues were fixed with 10% neutral buffered formalin and embedded in paraffin. Then Paraffin-embedded tissue sections (4-μm) were deparaffinized, rehydrated and immunostained for detection of Ki-67 (a proliferation marker) expression levels. Staining of slides with Ki-67 antibody (1:400 dilution; ab16667, Abcam) was carried out. EnVision FLEX High pH 9.0 Visualization System (DAKO) was then used, followed by incubating with streptavidin horseradish peroxidase (LSAB kit; Dako, Denmark) and staining with 3, 3-diaminobenzidine (DAB). Staining of sections with hematoxylin was then done after which it was dehydrated, mounted, and photographed with a light microscope (Olympus).

Statistics analysis
SPSS 19.0 (Armonk, USA) and GraphPad Prism software 5.01 (La Jolla, USA) was used for statistical analysis. To analyze the differences, Student's t-test and one-way ANOVA was carried out. A P value less than 0.05 were considered as significance.
Correlations between DSCAM-AS1, AKT3 and miR-384 were conducted by Pearson's correlation. Kaplan-Meier curve and log rank test was utilized to determine survival rate. All data was shown as mean ± standard deviation (SD) with at least 3 replicates measurements.