Introduction

Colorectal cancer is a common malignancy, accounting for 11% of all cancers in the United States with an estimated 146,970 new cases and 49,900 deaths expected in the year 2009 [1]. Although recent advances in surgical and adjuvant therapies and dietary and screening programs have facilitated an overall decline in the mortality of colon cancer, many colorectal cancer patients still develop metastatic disease and ultimately die [2]. Improved methods for early diagnosis, new biomarkers for aggressive disease, and effective systemic treatment of metastatic disease are urgently needed to improve survival of these patients. Increased understanding of the molecular mechanisms for altered gene expression in colorectal cancer is likely to aid in development of novel strategies for better clinical care of the patients. Specifically, microRNAs (miRNAs) are a class of small noncoding RNAs 18–22 nucleotides long that post-transcriptionally silence protein expression through binding to complementary target mRNAs and thereby target them for degradation or inhibit them from translating into protein [3]. To date, many studies have shown that miRNAs play an important role in cell growth, differentiation, apoptosis, and carcinogenesis through their tumorigenic and anti-tumorigenic functions [4–6]. It was found that a number of miRNAs were differentially expressed in colon cancer cells and tissues and that some of them play an important role in colon cancer development and progression [7, 8]. Therefore, it is insightful to study the regulatory role of miRNA and to understand their functions and the underlying molecular mechanisms in colon cancer.

Nasopharyngeal carcinoma-associated gene 6 (NGX6) was originally cloned from nasopharyngeal carcinoma cells that harbored a loss of heterozygosity on chromosome 9 [9]. Further studies found that NGX6 is a novel putative metastasis suppressor gene in nasopharyngeal carcinoma cells and NGX6 protein contains two transmembrane domains and one EGF-like domain. Induction of NGX6 expression significantly decreased the growth, motility, and invasion of nasopharyngeal carcinoma cells [10]. More recently, NGX6 expression was found to be downregulated in different stages of colon cancer and 93.8% of metastatic colon cancer tissues were shown to exhibit significantly decreased NGX6 expression [11]. Xenograft and invasion assays showed that restoration of NGX6 gene expression reversed aggressiveness of colon cancer cells and arrested the colon cancer cells at the G0/G1 cell cycle phase [12]. In our previous study, we identified the NGX6 tumor suppressor mechanism for both transcriptional and protein level regulation, which negatively regulated the expression and activities of Wnt/β-catenin and EGFR-mediated JNK pathways [13]. We were, therefore, interested in determining whether and how NGX6 is able to regulate expression of miRNAs in colon cancer and to discover novel targets for future control of colon cancer in clinic. In this study, we used NGX6-transfected and vector-only transfected colon cancer cells to screen differentially expressed miRNA by means of a miRNA oligonucleotide array; the functions of miRNA-target genes was then analyzed to elucidate the tumor suppression mechanisms of NGX6.

Materials and methods

Cell lines and reagents

A poorly differentiated colon cancer cell line, HT-29, was obtained from the Cell Line Center of Xiangya Medical School. The stable NGX6-expressed and vector-only HT-29 sublines had been previously established in our lab by transfection of pcDNA3.1(+)/NGX6 vector or the control vector, respectively. These cells were cultivated in Dulbecco’s modified Eagle’s minimum essential medium (DMEM) with 10% fetal bovine serum (FBS) at 37°C in a humidified atmosphere of 95% air and 5% CO2. Calf intestine alkaline phosphatase (CIP) and T4 RNA ligase were obtained from GE Healthcare (Barrington, IL, USA). Dimethyl sulfoxide (DMSO) was from Sigma-Aldrich (St Louis, MO, USA). MicroBioSpin 6 columns were from Bio-Rad (Hercules, CA, USA). miRNA microarray, miRNA labeling reagents, hybridization kit, and buffers were from Agilent (Santa Clara, CA, USA). DNase/RNase-free distilled water and mirVana RNA isolation kit were from Ambion (Austin, TX, USA). RNase-free DNase, miScript reverse transcription kit, miScript SYBR® green PCR kit, and primers were obtained from Qiagen (Germantown, MA, USA).

RNA isolation and hybridization

The cells were seeded and grown for 3 days and total RNA was then isolated from the cells using an RNA isolation kit according to the miRNA microarray manufacturer’s instructions (Shanghai Array Biotechnology Company, Shanghai, China). Next, the 3′ ends of the RNA were labeled with the 3′-phosphate of 3′,5′-cytidine bisphosphate-Cy3 (Cy3 is a fluorescence dye) and these quantified and labeled RNAs were then hybridized to the miRNA microarray chips in 16 ml of hybridization buffer containing 15% formamide, 0.2% SDS, 3× SSC, and 50× Denhardt’s solution at 42°C overnight. On the following day, the miRNA microarray slides were washed at 42°C in a wash buffer containing 0.2% SDS and 2× SSC for 5 min followed by an additional 5 min wash in 0.2× SSC buffer. After that, these slides were subjected to scanning and data analysis.

Chip scanning and data analysis

The array slides were scanned using a LuxScan 10K/A double-channel laser (CapitalBio Corp, Beijing, China), and the data were analyzed by the software GenePixPro 4.0 and Significance Analysis of Microarrays (SAM, version 2.1) to determine differentially expressed miRNAs between NGX6-expressed and vector-only transfected HT-29 cells.

Quantitative real-time reverse transcription polymerase chain reaction (Q-RT-PCR)

Q-RT-PCR was used to verify NGX6-regulated miRNA expression. Briefly, RNA was isolated from the stable NGX6 or vector-only transfected HT-29 cells and 50 ng of each RNA was reversely transcribed into cDNA to be used as templates for PCR amplification. The PCR was carried out in a mixture of 3 mM MgCl2, 0.25 M of each primer, and 5 ng cDNA; amplification occurred at an initial denaturation cycle of 95°C for 10 min followed by 40 cycles of 95°C for 15 s, 60°C for 30 s, and 72°C for 30 s in a Bio-Rad real-time PCR system using the miScript SYBR® Green PCR kit from Qiagen. Quantification of gene expression was performed using the comparative threshold cycle (ΔΔC T) method according to the manufacturer’s protocol. The primers used have been listed in Table 1.

Table 1 Primers used to detect miRNA expression for Q-RT-PCR
Full size table

Bioinformatic analyses of miRNA-targeted genes

The miRNA-targeted genes were predicted by online-available software (i.e., Sanger, Target Scan, and MicroRNA). The gene ontology (GO) database (www.geneontology.org) was then utilized to analyze the functions of these target genes and to create the network of miRNAs and their target genes with respect to target gene functions.

Results

miRNA array analysis

Total RNA isolated from pcDNA3.1(+)/NGX6/HT-29 and pcDNA3.1(+)/HT-29 cells was quantified and qualified and the 3′ ends of the RNAs were labeled with the 3′-phosphate of 3′,5′-cytidine bisphosphate-Cy3; these labeled and quantified RNA probes were then hybridized to microarray chips containing 866 human miRNAs. Data mining indicated that 14 miRNAs were differentially expressed between NGX6-transfected and vector-only transfected cells. The cut-off point of twofold was used to identify differential up-regulation and 0.5-fold for down-regulation. In particular, miR-126, miR-142-3p, miR-155, miR-552, and miR-630 were found to be upregulated by NGX6 transfection, whereas miR-146a, miR-152, miR-205, miR-365, miR-449, miR-518c*, miR-584, miR-615, and miR-622 appeared to be downregulated in response to NGX6 transfection (Table 2).

Table 2 Differential expression of miRNAs in NGX6-transfected HT-29 cells
Full size table

Real-time PCR confirmation of these differentially expressed miRNAs after NGX6 transfection

To verify the results of our miRNA microarray analysis, we performed Q-RT-PCR analysis of a sampling of the differentially expressed miRNAs. We designed and synthesized primers for 14 miRNAs. Among these 14 miRNAs, 11 had been downregulated and 3 upregulated after NGX6 gene transfection. Our Q-RT-PCR confirmed 85.7% of the miRNA microarray data. In particular, miR-146a, miR-152, miR-205, miR-365, miR-449, miR-518c, miR-584, miR-615, miR-622, miR-126, miR-142-3p, and miR-155 matched well with the miRNA microarray data, whereas miR-552 and miR-630 were upregulated in miRNA array data but was downregulated according to Q-RT-PCR data (Fig. 1a; Table 3).

Fig. 1
figure 1

Identification and regulated-gene network of NGX6-induced differentially expressed miRNAs. a Q-RT-PCR analysis of miRNA expression in NGX6- and vector-only transfected colon cancer cells. The stable NGX6- and vector-only transfected HT-29 sublines and parental cells were grown and RNA from the cells was then isolated and subjected to Q-RT-PCR analysis of miRNA expression. The experiments were in triplicate and repeated once. The data were normalized to β-actin gene. Top panel, miR-142-3p; bottom panel, miR-152. In each panel: left: PCR amplification data; middle: PCR melt peak data; right: quantitative data; the data were normalized to the β-actin gene. Experiments were performed in triplicate and repeated once. b The gene network mediated by NGX6-induced differentially expressed miRNAs. After NGX6 gene transfection, 12 miRNAs were found to be differentially expressed and bioinformatic analyses showed that target genes of miR-615 are unclear, while the another 11 miRNAs produced a total of 254 target genes. After that, we then mapped their regulatory network based on the miRNAs and their target genes with a JAVA software. The triangles represent the genes (such as HEY2, SGCD, and SMAP1) that are regulated by four different miRNAs. The pentagons are genes that are second-ranked, including NOTCH2, PBX2, VAMP5, FGD2, SGCB, HDAC9, JPH1, and CTPS2 and regulated by three different miRNAs. The rounds show genes that are regulated by only single miRNA, such as DBNL, MAP4K5, IGF1, MIB2, AKT1, and EGFL7

Full size image
Table 3 Data comparison between Q-RT-PCR and microarray analysis for the differentially expressed miRNAs after NGX6 transfection
Full size table

Prediction of miRNA-targeted genes and their functions

Next, we analyzed the target genes that may be regulated by these 12 miRNAs. Bioinformatic analyses identified a total of 254 genes were targeted by 11 miRNAs, and the target genes of miR-615 are unclear. We then mapped the regulatory networks associated with the miRNAs and their target genes (Fig. 1b). Degree values were assigned to genes in accordance with regulation by individual miRNAs; hence, a higher degree value signified association with more miRNAs. On the map, the triangles represent the higher degree genes (such as HEY2, SGCD, and SMAP1) that are regulated by four different miRNAs and were considered to be most important in this network. The pentagons represent genes that are regulated by three different miRNAs (including NOTCH2, PBX2, VAMP5, FGD2, SGCB, HDAC9, JPH1, and CTPS2). The rounds show genes that were associated with regulation by only a single miRNA (DBNL, MAP4K5, IGF1, MIB2, AKT1, and EGFL7).

After this, we sought to determine the potential functions of each of these genes. According to the GO database, the major functions associated with these genes are apoptosis, mobility, migration, hydrolysis activity, and molecular signaling through JNK and Notch pathways (Table 4). In other words, NGX6 likely plays an important role in regulation of apoptosis, mobility, migration, and activity of JNK and Notch pathways through NGX6-mediated miRNA expression.

Table 4 Classification and functions of the genes that are regulated by the 11 differentially expressed miRNAs
Full size table

Discussion

In the current study, we compared miRNA expression in NGX6-transfected HT-29 cells to vector-only transfected cells and found that 12 miRNAs were differentially expressed. Bioinformatic analyses of 11 miRNAs yielded a total of 254 potential target genes, and the target genes of miR-615 are unclear. The further investigation indicated that these target genes together formed a regulatory network to modulate tumor cell apoptosis, mobility, migration, hydrolysis activity, and molecular signaling through JNK and Notch pathways. Taken together, the results from the current study suggest that NGX6 plays an important role in regulation of apoptosis, mobility, migration, hydrolysis activity, and molecular signaling via JNK and Notch pathways through NGX6-mediated miRNA expression. Gaining a detailed understanding of these miRNA and their targeted genes will aid in the development of novel strategies to improve colon cancer treatment and outcome.

Full-length NGX6 cDNA was previously cloned in our laboratory and our studies to date have demonstrated that NGX6 is a putative tumor suppressor gene in colon cancer and nasopharyngeal cancer [10–12]. In our previous study, we have screened and profiled a number of different genes that were differentially expressed in NGX6 overexpressed nasopharyngeal cancer cells using a cDNA array. Our data showed that 55 genes, or expression sequence tags, were differentially expressed in response to NGX6 transfection and that some of these genes are associated with adhesion, cell cycle and transcriptional regulation of gene expression [14]. In colon cancer, NGX6 expression is known to be significantly reduced and this reduction has been associated with colon cancer metastasis. Recently, we found that the NGX6 gene has two alternative spliced isoforms and the shorter form is the major-expressed one in colon epithelia and its reduction was associated with distant metastasis of colon cancer. Nude mouse xenograft and invasion assays revealed that over-expression of NGX6 inhibited tumor cell invasion and metastasis capacity in colon cancer [15]. Moreover, NGX6 was able to block angiogenesis in a chicken embryo chorioallantoic membrane model. NGX6 protein can down-regulate expression of RhoA, catenin α2, ezrin, and nm23-H1 in nasopharyngeal cancer. Ezrin is an adaptor protein situated at the interface of the cell membrane and cytoskeleton, and it is involved in regulation of cell migration and adhesion. NGX6 protein binds to ezrin through its N-terminus to inhibit cell membrane and cell adhesion to the basal membrane [16]. We also found that NGX6 anti-tumor activity is associated with suppression of Wnt/β-catenin and EGFR-mediated JNK pathways at both transcriptional and translational levels. However, it is unclear how NGX6 regulated the expression of miRNA at the post-transcriptional level. In our current study, we identified five miRNAs that were upregulated by NGX6 and nine miRNAs that were downregulated by NGX6. Q-RT-PCR confirmed that 12 miRNAs matched well with the miRNA microarray data. Among these 12 miRNAs, the target genes of miR-615 are unclear, the other 11 differentially expressed miRNAs and their target genes may form a cohesive regulatory network that affects apoptosis, mobility, migration, hydrolysis activity, and molecular signaling through JNK and Notch pathways in the cells.

In particular, these 11 miRNAs play tumor-suppressing and tumor-promoting roles in human cells. For example, miR-126 has anti-tumor activities in lung and cervical cancers [17, 18] and was upregulated in colon cancer cells by NGX6 transfection in the current study. miR-365 has been shown to be expressed at high levels in different human cancers with tumor-promoting activities [19], and we found it to be downregulated by NGX6. However, miR-146a has been shown to have high expression levels in thyroid cancer and cervical cancer [18, 20], and it can inhibit metastasis of breast cancer cells [21]. In our current study, miR-146a was downregulated by NGX6. Again, miR-205 has been found to be expressed at low levels in breast cancer tissues and have anti-tumor activities [22, 23], but its expression is increased in endometrial cancer, non-small cell lung cancer, and ovarian cancer [24–27]. Our current study showed that miR-205 was suppressed by NGX6. The variant expression pattern of miR-146a and miR-205 in different cancers suggests that the same miRNA might have different functions in different types of cancer, which may reflect differentially expressed target genes in different cancer cells. Therefore, their role in colon cancer and NGX6-regulation of their expression merits further investigation.

It is well known that miRNA can post-transcriptional regulate target gene expression. In the current study, we utilized web-based query tools (Sanger, TargetScan, and MicroRNA) to predict the target genes of 12 differentially expressed miRNAs; these genes, together, may represent a cohesive regulatory network to mediate apoptosis, mobility, migration, hydrolysis activity, and molecular signaling through JNK and Notch pathways in cells. We found that 12 genes from our cohort were targeted by multiple miRNAs. For example, HEY2 can be regulated by miR-146a, miR-152, miR-205, and miR-449, while SGCD and SMAP1 can be regulated by miR-146a, miR-152, miR-365, and miR-449. NOTCH2 was regulated by miR-146a, miR-205, and miR-449, while NR2C2 was regulated by miR-126 and miR-155. Thus, our data confirmed that a gene, like NGX6, can regulate expression of multiple miRNAs, while a single miRNA can target multiple genes. Furthermore, a single target gene can be coordinately regulated by multiple miRNAs to form a complex and dynamic regulatory network. However, further investigation to confirm the relationship in the cells is necessary in order to define the exact roles of these miRNAs in cancer development, progression, and metastasis. Nonetheless, our data clearly suggest that NGX6 most likely exerts its tumor suppressor functions by regulating such a complex network.

In addition, our current study also analyzed and predicted the biological functions of the miRNA-targeted genes. We found that among these target genes, AKT1, BAX, BCL2L13, BID, CYCS, and BCL6 were known to mediate cell apoptosis; ACTG1, IGF1 BMP4, BMP7, FGD2, FGD3, FGD4, MTSS1, and CAD were related with cell mobility [28, 29], and CITED2, DRD1, JAG1, CXCR4, LAMA2, and LAMA5 were involved in cell migration [30, 31]. Moreover, SIRT1and SIRT2 exhibit hydrolase activities. In addition, among these target genes, MAP4K5, DBNL, and BIRC7 are critical along the JNK pathway, which is itself, involved in cell proliferation and apoptosis [32]. Previous studies have demonstrated that NGX6 was able to suppress activation of the EGFR-mediated JNK pathway. The current study revealed that NGX6 inhibition of JNK pathway genes can occur, at least partially, through the regulation of miRNA expression. Again, HEY2, NOTCH1, NOTCH2, JAG1, and JAG2 were identified as important signaling molecules in the Notch pathway, which is involved in cell differentiation and tumor formation [33]. Notch pathway is recognized to be aberrantly activated in colon cancer cells [34]; recent studies have shown that the over-expression of Notch can inhibit activation of Wnt pathway genes to inhibit tumor formation. Our previous reports found that NGX6 can block the Wnt pathway. Thus, we speculate that NGX6 may inhibit activation of Wnt by activating Notch at the post-transcriptional level. However, further study is required to definitively determine whether NGX6 inhibits tumor formation through activation of the Notch pathway.