This article is mentioned in:

Introduction

Colon cancer is the third most common type of cancer, and the leading cause of cancer-related mortality worldwide (1). There is increasing evidence to suggest that the progression of colon cancer is a multiple-step process, involving genetic and epigenetic abnormalities (2). By comparison, epigenetic alterations are more frequently observed, and carry greater effects compared with gene mutations. For example, the silencing micro RNA-137 can target and decrease the expression level of >500 protein coding genes (3).

Long non-coding RNAs (lncRNAs) are a type of non-coding transcript of >200 nucleotides in length. lncRNAs have been associated with a number of physiological and pathological processes in various human diseases, regulating gene expression via epigenetic, transcriptional and post-transcriptional modification (4). Recent studies revealed that lncRNA was an important regulator in colon cancer pathogenesis, modulating numerous processes including cell proliferation, differentiation, tumor migration, invasion and angiogenesis. It has been reported that lncRNA SUMO1P3 stimulated the proliferation, metastasis and angiogenesis of colon cancer cells, suggesting it's potential as a novel prognostic indicator and therapeutic target for the disease (5).

lncRNAs can target different steps in the transcription process, regulating the activation or regression of gene transcription by modulating different transcriptional components, including RNA polymerase (RNAP) II and the DNA duplex (6). These lncRNAs may be part of the regulatory network that, together with transcription factors, regulate the level of gene expression. Reportedly, certain lncRNAs acted as regulators of general transcription factors which were required for the activity of RNAP II (7). These transcription factors, in addition to components of the initiation complex, assemble at promoter regions and participate in transcription elongation and the regulation of cis elements (8); it was reported that an lncRNA transcribed from the dihydrofolate reductase (DHFR) gene was able to form a RNA-DNA triplex in the promoter region of DHFR, inhibiting its transcription by blocking the binding of transcription initiation factor IIB (9). Furthermore, the gene coding lncRNA ANRIL is able to run antisense, silencing P15 with which it overlaps.

However, survival-related lncRNAs, particularly in colon cancer remained to be identified. In order to select therapeutically relevant lncRNAs, a linear regression model was used to identify survival-related lncRNAs in colon cancer and further detect MIR210HG-related genes. Gene regulatory network and pathway analyses revealed that MIR210HG may exert its effects in colon cancer through the regulation of cancer cell metabolism and adhesion. Furthermore, bioinformatics predictions indicated that MIR210HG may affect colon cancer by regulating transcription and post-transcriptional processing.

Materials and methods

Screening of survival-related lncRNAs in colon cancer

The RNA-seq data of colon cancer-associated lncRNAs, which consisted of 471 tumor samples and 41 normal samples, were downloaded from The Cancer Genome Atlas (TCGA) data portal (https://cancergenome.nih.gov/) (10). The data were divided into high and low expression groups according to a cut-off of 75 and 25%, respectively. Homo_sapiens.GRCh38.84 was used for annotation. The RNA-seq data of other gastrointestinal cancers, including esophageal cancer (162 tumor and 11 normal samples), rectal cancer (167 tumor and 10 normal samples) and stomach cancer (375 tumor and 32 normal samples), were also downloaded, and assigned into high and low expression groups using the same classification criteria as in colon cancer. The association between lncRNA expression and patient survival was determined using the Kaplan-Meier method, and evaluated using the log-rank test. The R package survival (https://github.com/therneau/survival) was used to perform these statistical analyses. P<0.05 was set as the significance threshold.

Genes significantly associated with MIR210HG expression in colon cancer

Based on the above screening, the most significant survival-related lncRNA in colon cancer was identified as MIR210HG; this lncRNA was therefore selected for use in further analysis. Associations between the expression of MIR210HG and other genes were calculated using a linear regression model. It was hypothesized that genes whose expression levels were linearly associated with MIR201HG were likely to be located nearby, upstream or downstream. This method was widely used to screen target-related genes. RNA expression was processed using log2(counts +1) prior to analysis. Subsequently, multiple selection criteria (P-value and adj.r2) were applied to determine the most appropriate candidates. P<0.05 and adj.r2 >0.16 were set as the significance threshold. R was used to perform these statistical analyses.

Functional and pathway enrichment analysis

Gene ontology (GO) enrichment analysis (http://www.geneontology.org/) was used to detect biological processes, molecular functions and cellular components. Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis (http://www.kegg.jp/), which is used for pathway investigations, was performed using the Database for Annotation, Visualization, and Integrated Discovery (DAVID; https://david.ncifcrf.gov/, version 6.8) (11–13). P<0.05 was set as the significance threshold.

Investigation of the correlation between MIRE210HG expression and lymphatic metastasis

All clinical data were obtained from TCGA. A normality test, the Kolmogorov-Smirnov test, was performed and a Mann Whitney U test was subsequently used to assess the association between MIRE210HG expression and metastasis in colon cancer. P<0.05 was considered to indicate a statistically significant difference. SPSS Statistics 22.0 (IBM Corp.) was used to perform statistical analysis.

Protein-protein interaction (PPI) network construction and identification of hub genes

All MIR210HG-related genes were imported into the Search Tool for the Retrieval of Interacting Genes/Proteins (STRING) to construct a PPI network (14). A combined score ≥0.4 was set as the significance threshold. The network was visualized using Cytoscape (version 3.5.1, http://www.cytoscape.org/) (15). The top 10 hub genes were filtered using cytoHubba, an app in Cytoscape, using the Maximal Clique Centrality (MCC) method (16).

Module analysis of the PPI network

To investigate the most significant modules, Clustering with Overlapping Neighborhood Expansion (ClusterONE), an app in Cytoscape, was used to screen modules within the PPI network with a minimum size of 6, and the minimum density set to auto (17). P<0.05 was considered to indicate a statistically significant difference. GO and KEGG analyses of genes in the screened modules were performed using DAVID.

Transcription and post-transcriptional modulation prediction

To further investigate the mechanism of MIR210HG in colon cancer, known MIR210HG-protein interactions were summarized using starBase v2.0 (http://starbase.sysu.edu.cn/index.php) (18). Novel lncRNA-protein interactions between MIR210HG and its associated transcription factors were also predicted using RPIseq (http://pridb.gdcb.iastate.edu/RPISeq/index.html) (19). Candidates with a Random Forest classifier value >0.5 and Support Vector Machine classifier value >0.5 were considered to be positive results.

Interaction probabilities generated using RPISeq range between 0 and 1. In performance evaluation experiments, predictions with probabilities >0.5 were considered ‘positive’, indicating that the corresponding RNA and protein were likely to interact. Neighboring mRNAs located within 1 Mbps of the MIR210HG loci were investigated to predict possible cis target genes of MIR210HG. Candidate gene information was downloaded from USCS (GRCh38) (20).

Results

MIR210HG expression correlates with survival in patients with colon cancer

lncRNAs have been reported to contribute to patient survival in multiple cancers, including colon cancer (20), breast cancer (21) and larynx squamous cell carcinoma (22), and therefore, in the present study, survival-related lncRNAs in colon cancer were investigated. RNA-seq and clinical data from colon cancer cohorts (obtained from TCGA) were analyzed to assess the association between various lncRNAs and patient survival. A total of 226 survival-related lncRNAs were selected using P<0.05 as the cut-off threshold (The top 10 survival-related lncRNAs are presented in Table I). A number of lncRNAs were reported to contribute to tumor promotion, including MIR210HG (P=1.19×10−3; log-rank test), which is a diagnostic biomarker in glioma (23). A high expression level of MIR210HG was significantly associated with shorter overall survival in patients with colon cancer, suggesting a potential prognostic role for MIR210HG in the disease (Fig. 1A). However, MIR210HG showed a weak association with survival from esophageal, rectal and stomach cancers, indicating that this particular effect of MIR210HG overexpression is specific to colon cancer (Fig. 1).

Figure 1. MIR210HG survival curves for gastrointestinal cancers. (A) MIR210HG survival curve for colon cancer (P=1.19×10−3; log-rank test). (B) MIR210HG survival curve for esophageal cancer (P=1.00; log-rank test). (C) MIR210HG survival curve for stomach cancer (P=0.96; log-rank test). (D) MIR210HG survival curve for rectal cancer (P=0.93; log-rank test).

Table I. Top 10 survival-related lncRNAs in colon cancer.

Table I.

Top 10 survival-related lncRNAs in colon cancer.

lncRNA	Expression level [log2 (counts +1)]	P-value
RP11-66B24.2	4.35 (1.58, 8.63)	1.08×10−4
MIR210HG	208.82 (111.73, 352.98)	1.19×10−4
RP11-367H1.1	5.29 (2.42, 9.54)	1.20×10−4
CTC-573N18.1	0.68 (0.08, 2.9)	1.30×10−4
LA16c-380A1.1	1.36 (0.14, 3.41)	1.31×10−4
RP11-108K3.1	7.79 (4.3, 13.05)	1.35×10−4
RP11-93I21.3	0.52 (0.08, 2.15)	1.49×10−4
LINC00174	179.71 (122.2, 269.39)	2.11×10−4
RP11-108K3.2	5.38 (1.97, 9.91)	2.96×10−4
AP006621.5	68.84 (44.66, 95.88)	3.80×10−4

[i] lncRNA, long no-coding RNA.

Investigation of MIR210HG-related genes in colon cancer

TCGA RNA-seq expression data for MIR210HG was downloaded, and the correlation between MIR210HG expression and other genes in colon cancer was calculated using a linear regression model. A total of 373 genes with P<0.05 and adj.r2 >0.16 were filtered out for further analysis (Fig. 2). The top 10 genes significantly associated with MIR210HG expression in colon cancer are presented in Table II. A total of 231 and 142 genes were positively, and negatively associated with MIR210HG expression in colon cancer, respectively. Additionally, the known PPIs among these MIR210HG-related candidates were detected using STRING (Fig. 3). A total of 359 nodes were noted in the network, and their interactions are presented in Fig. 3A. To identify the key nodes that most affected this network, MCC analyses were conducted to rank the strongest hub genes. Hub genes act as cores within the regulatory network, and the results revealed complex and strong links with various MIR210HG-related candidates. It was speculated that the expression of these hub genes may be directly regulated by MIR210HG or its downstream targets, and the results of linear regression analysis showed a significant correlation between the expression of the 10 hub genes and MIR210HG (Table III). Modules were regarded as a group of proteins that participate in the same biological process, or composition of the same complex (24). Modularity analysis was conducted, and 2 modules were selected from the constructed PPI network, as illustrated in Fig. 3B and C. GO and KEGG analysis indicated that module 1 may be associated with glycolysis and cell adhesion (e.g. triosephosphate isomerase 1, GAPDH, enolase 2, mucin 1, cell surface associated, egl-9 family hypoxia inducible factor 3, Bardet-Biedl syndrome 10, and phosphofructokinase, liver type), while module 2 may be associated with ubiquitination (e.g. kelch like family member 42, anaphase promoting complex subunit 2, F-box and WD repeat domain containing 5, and F-box and leucine rich repeat protein 15) (Figs. 4B and C; 5B).

Figure 2. Genes significantly associated with MIR210HG expression in colon cancer.

Figure 3. PPI network and modules. (A) Known PPIs among MIR210HG-related candidates. (B) Module 1 and (C) module 2. PPI, protein-protein interaction.

Figure 4. GO functional enrichment analyses. (A) GO functional enrichment analyses of MIR210HG-related candidates. GO functional enrichment analyses results for (B) module1 and (C) module 2. GO, gene ontology.

Figure 5. KEGG pathway enrichment analyses. (A) KEGG pathway enrichment analyses of MIR210HG-related candidates. (B) KEGG pathway enrichment analyses results of modules. KEGG, Kyoto Encyclopedia of Genes and Genomes.

Table II. Top 10 genes significantly associated with MIR210HG expression in colon cancer.

Table II.

Top 10 genes significantly associated with MIR210HG expression in colon cancer.

Gene	P-value	Adj.r2
NDRG1	7.96×10−51	0.38
ALDOA	1.43×10−51	0.38
EGLN3	2.49×10−51	0.36
ALDOC	1.09×10−51	0.35
SLC16A3	5.78×10−51	0.34
PFKFB4	3.23×10−51	0.33
MIR210	1.02×10−51	0.30
SLC6A8	1.85×10−51	0.30
NARF	3.34×10−51	0.30
ANGPTL4	7.93×10−51	0.30

[i] NDRG1, N-myc downstream regulated 1; ALDOA, aldolase, fructose-bisphosphate A; EGLN3, egl-9 family hypoxia inducible factor 3; ALDOC, aldolase, fructose-bisphosphate C; PFKFB4, 6-phosphofructo-2-kinase/fructose-2,6-biphosphatase 4; MIR210, microRNA 210; SLC6A8, solute carrier family 6 member 8; NARF, nuclear prelamin A recognition factor; ANGPTL4, angiopoietin like 4.

Table III. Top 10 hub genes in protein-protein interaction networks.

Table III.

Top 10 hub genes in protein-protein interaction networks.

No.	Hub genes	P-value	Adj.r.squared
1	PKM	6.14×10−26	0.21
2	TPI1	1.01×10−20	0.17
3	PFKL	1.18×10−25	0.21
4	PFKP	9.47×10−27	0.22
5	GAPDH	1.65×10−21	0.17
6	ENO2	7.55×10−27	0.22
7	LDHA	1.88×10−29	0.24
8	ALDOA	1.43×10−50	0.38
9	ALDOC	1.09×10−46	0.35
10	PFKFB4	3.23×10−42	0.33

[i] PKM, pyruvate kinase M1/2; TPI1, triosephosphate isomerase 1; PFKL, phosphofructokinase, liver type; PFKP, phosphofructokinase, platelet; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; ENO2, enolase 2; LDHA, lactate dehydrogenase A; ALDOA, aldolase, fructose-bisphosphate A; ALDOC, aldolase, fructose-bisphosphate C; PFKFB4, 6-phosphofructo-2-kinase/fructose-2,6-biphosphatase 4.

MIR210HG may impact colon cancer by targeting energy metabolism and cell adhesion

In order to determine the possible mechanism of MIR210HG in colon cancer, GO and KEGG pathway analyses were conducted. The GO results revealed that MIR210HG expression was strongly associated with energy metabolism in colon cancer, as evidenced by the enrichment of genes involved in canonical glycolysis (P=1.15×10−8), glycolytic process (P=2.78×10−5) and gluconeogenesis (P=0.001) (Fig. 4A). Additionally, the KEGG results revealed that MIR210HG was significantly associated with the biosynthesis of glycolysis or gluconeogenesis (P=8.33×10−6) and central carbon metabolism in cancer (P=0.003) (Fig. 5A). Furthermore, cell-cell adhesion (P=3.99×10−4), cell-cell adherens junction (P=0.002) and cadherin binding involved in cell-cell adhesion (P=0.001) were notably enriched in biological process, cellular component and molecular function analysis, respectively. A recently published paper suggested that MIR210HG facilitates cancer cell invasion and metastasis in osteosarcoma (25), which is consistent with the findings of the present study. These findings suggest that MIR210HG may exert its effects in colon cancer by modulating energy metabolism and cell adhesion, which are closely associated with the prognosis of patients with colon cancer. Furthermore, a Mann Whitney U test was performed to verify the association between lymphatic metastasis and MIR210HG expression in colon cancer, and a strong association was identified (P=0.02; Fig. 6). These results suggest that MIR210HG may be a modulator of colon cancer cell metastasis, and are consistent with the results of the GO and KEGG enrichment analyses.

Figure 6. Relative expression level of MIR210HG in patients with or without lymphatic invasion. *P<0.05 vs. without lymphatic invasion.

MIR210HG may influence colon cancer via transcription and post-transcriptional processing

To further investigate the mechanism of MIR210HG in colon cancer, various possible interactions between MIR210HG and screened MIR210HG candidates were predicted. Known MIR210HG-protien interactions were summarized using starBase v2.0. Specific genes associated with RNA splicing and degradation were detected (Table IV), including FUS RNA binding protein (FUS). These genes may therefore serve as post-transcriptional regulators in colon cancer. Novel lncRNA-protein interactions were also predicted using RPISeq. Since transcription factors are considered to be important regulators of gene expression, the focus was on their association with MIR210HG (Table V). Predictions for Mediator of RNA polymerase II transcription (MED)15, MED16 and B-cell lymphoma 3 protein (BCL3; probabilities >0.5) were considered to be ‘positive’. It was observed that MED16 and BCL3 were connected to hub genes within the MIR210HG regulatory network, suggesting that they may modulate the network by stimulating these genes.

Table IV. Known MIG210HG-protien interactions via starBase v2.0.

Table IV.

Known MIG210HG-protien interactions via starBase v2.0.

Name	lncRNA	Target sites	BioComplex	ClipReadNum
eIF4AIII	MIR210HG	12	2	31
FUS	MIR210HG	1	1	5
SFRS	MIR210HG	1	1	800
U2AF65	MIR210HG	1	1	2

[i] BioComplex is the number of supporting experiments. ClipReadNum is the read number of CLIP-Seq. lncRNA, long non-coding RNA; eIF4AIII, Eukaryotic initiation factor 4A-III; FUS, FUS RNA binding protein; SFRS, Serine/arginine-rich splicing factor 1; U2AF65, U2 small nuclear RNA auxiliary factor 2.

Table V. Prediction of lncRNA-protein interactions between MIR210HG and candidate transcription factors through RPIseq.

Table V.

Prediction of lncRNA-protein interactions between MIR210HG and candidate transcription factors through RPIseq.

Transcription factors	Isoform	Gene names	RF classifier	SVM classifier
MED15	1	Mediator complex subunit 15	0.9	0.57
MED15	2	Mediator complex subunit 15	0.9	0.59
MED15	3	Mediator complex subunit 15	0.9	0.62
MED16	1	Mediator complex subunit 16	0.8	0.61
MED16	2	Mediator complex subunit 16	0.75	0.61
MED16	3	Mediator complex subunit 16	0.75	0.6
MED16	4	Mediator complex subunit 16	0.8	0.54
MED16	5	Mediator complex subunit 16	0.8	0.48
CDK9	1	Cyclin dependent kinase 9	0.85	0.55
CDK9	2	Cyclin dependent kinase 9	0.85	0.51
BCL3	1	B-cell CLL/lymphoma 3	0.75	0.53
PPP1R13L	1	Protein phosphatase 1 regulatory subunit 13 like	0.85	0.49

[i] RF, Random forrest; SVM, Support Vector Machine.

lncRNAs also interacted with other cis acting elements of promoters or co-expressed genes, thus regulating gene expression via transcription or post-transcriptional processes. We therefore investigated neighboring mRNAs located within 1 Mbps of the MIR210HG locus to predict possible cis target genes. A total of 13 genes were identified (Table VI), 5 of which were members of the MIR210HG regulatory network. One of these genes was HRAS, which is known to be involved in central carbon metabolism in cancer. HRAS is a member of the Ras superfamily of GTPases, which are involved in the regulation of cell division in response to growth factor stimulation. Activated mutant HRAS (Hras-V12) is able to initiate irreversible cell cycle arrest and increase the expression of the tumor suppressors p16 and p53 (26).

Table VI. Prediction of nearby cis target genes of MIR210HG.

Table VI.

Prediction of nearby cis target genes of MIR210HG.

Nearby gene names	Whether in PPI network (1=yes, 0=no)
RNH1	0
LOC101059906	0
HRAS	1
LRRC56	0
LMNTD2	1
RASSF7	1
MIR210	1
LOC143666	0
PHRF1	0
IRF7	1
CDHR5	0
SCT	0
DRD4	0

[i] PPI, protein-protein interaction; RNH1, ribonuclease/angiogenin inhibitor 1; HRAS, HRas proto-oncogene, GTPase; LRRC56, leucine rich repeat containing 56; LMNTD2, lamin tail domain containing 2; RASSF7, Ras association domain family member 7; MIR210, microRNA 210; PHRF1, PHD and ring finger domains 1; IRF7, interferon regulatory factor 7; CDHR5, cadherin related family member 5; SCT, secretin; DRD4, dopamine receptor D4.

Discussion

Lymphatic metastasis is one of the principal factors to affect the prognosis of patients with colon cancer (27). The survival rates of patients with lymphatic metastasis is significantly lower compared with that of patients without lymphatic metastasis (28). Furthermore, it has been identified that metastasis-related factors may be used as prognostic biomarkers or potential therapeutic targets for improving the survival rates of patients with colon cancer (29–31).

In the present study, candidate lncRNAs were screened based on the effects of lncRNA expression on the survival rate of patients with colon cancer. The employed screening method may be more suitable for clinical practice compared with traditional selection methods, and is more likely to be applied to treatment. The same method was used to screen survival-related lncRNAs in breast cancer, and a co-expressed gene network of the lncRNA with the greatest effect, LINC00704 (data not shown), was also constructed. The results of GO and KEGG analyses suggested that LINC00704-related genes are associated with the regulation of cell proliferation, which is in agreement with a previous report (32).

Glycolysis is the primary method by which tumor cells generate energy, even under normoxic conditions (33). Glycolysis provides energy for tumor cells, and so targeting glycolysis may have potential as a method to control tumor proliferation. Changes in glycolysis in tumor cells are associated with resistance to anticancer drugs. For instance, Song et al (34) reported that increased glycolytic activity and low phosphorylation efficiency increases chemo-resistance in patients with acute myeloid leukemia. This may be regulated by glycolysis-related molecules, including hypoxia-inducible factor 1α, hexokinase II, glucose transport 1 and lactate dehydrogenase (34).

In the present study, a number of survival-related lncRNAs in colon cancer were identified. Notably, the upregulation of a tumor promoter lncRNA, MIR210HG, was associated with shorter overall patient survival. To the best of our knowledge, current investigations into the association between MIR210HG and colon cancer prognosis are limited. Here, linear regression analysis was performed and an MIR210HG-related regulatory network was constructed. Specific MIR210HG-related candidates have been reported to be associated with colon cancer cell survival and metastasis, suggesting a possible link between MIR210HG expression and these genes that affects overall survival in patients with colon cancer. MIR210HG also influences osteosarcoma metastasis (25); furthermore, it was revealed that patients with high MIR210HG expression levels were more likely to have lymphatic metastasis. Genes that were not reported may be regarded as novel colon cancer-specific hubs in the MIR210HG regulatory network. Two notable modules and the top 10 hub genes were screened out using modularity analysis and MCC. A number of these hub genes, including fructose-bisphosphate aldolase A (ALDOA) and pyruvate kinase M (PKM)2, were associated with glycolysis, cell migration and apoptosis in colon cancer cells. These hub genes have been demonstrated to affect colon cancer prognosis. For instance, ALDOA was reported to be a key promoting element in colon cancer; ALDOA acts as a significant regulator of hypoxic adaptation in colon cancer cells (35), whilst it has also been reported to contribute to metastasis of colon cancer. Another key hub gene, PKM2 was able to disturb the Warburg effect and increase the sensitivity of cancer cells to tumor necrosis factor-related apoptosis-inducing ligand-induced cell death. Yang et al (36) reported that PKM2 overexpression was able to modulate STAT3 signaling, thereby facilitating the migration of colorectal cancer cells (36). Of these genes, ALDOA was found to have the strongest correlation with MIR210HG (adj.r2=0.4296; P=7.450×10−55), which suggests that it may be a direct target of MIR210HG.

The possible mechanism of interaction between MIR210HG and its co-expressed genes was also predicted. Reports have indicated that MIR210HG may affect colon cancer via transcription and post-transcriptional processing. Certain genes associated with RNA splicing and degradation, including FUS, were reported to interact with MIR210HG, suggesting that they may serve as post-transcriptional regulators in colon cancer. mRNA splicing is a vital step in post-transcriptional regulation, as it can trigger translation and functionally diversify proteins. In part of a study by Zarnack et al (37) experimental verification was carried out and reported a correlation between MIR210HG and U2AF65 (supplementary material; summary by starBase v2.0 database: http://starbase.sysu.edu.cn/index.php). While, according to the present study, MIR210HG and U2AF65 also showed associations (probabilities >0.5 using RPISeq: http://pridb.gdcb.iastate.edu/RPISeq/index.html), which is consistent with Zarnack's experimental results. However, in order to develop novel treatment methods to improve the prognosis of patients with colon cancer, we must improve our understanding of the molecular mechanisms of lncRNAs. In the present study, novel lncRNA-protein interactions were detected using RPISeq, and numerous transcription factors (including MED15, MED16 and BCL3) were considered to be positive candidates. In particular, MED16 and BCL3 were connected to hub genes within the MIR210HG regulatory network, reinforcing our prediction that MIR210HG serves as a transcriptional modulator in colon cancer. Although there is a lack of evidence indicating the specific molecular association between the selected transcription factors and candidates, it was revealed that some transcription factors (including MED15 and Angiopoietin-related protein 4, BCL3 and GAPDH, Cyclin-dependent kinase 9 and HRAS) interact with the hub genes and are strongly correlated with MIR210HG. A total of 5 selected transcription factors were associated with cancer metastasis. Decreased MED15 expression can suppress the transforming growth factor-β/Smad signaling pathway, thereby downregulating the metastasis of highly invasive breast cancer cells (38). BCL3 can also reduce cancer cell metastasis by influencing cell motility (39). Furthermore, neighboring mRNAs located within 1 Mbps of the MIR210HG loci were investigated to predict possible cis element target genes of MIR210HG. A total of 13 genes were identified, 5 of which, including HRAS, were members of the MIR210HG regulatory network. These results provide new information as how MIR210HG regulates the regulatory network and affects survival in patients with colon cancer.

In summary; in the present study survival-related lncRNAs in colon cancer were screened and bioinformatics analyses were performed to investigate the prognostic potential of MIR210HG and its related genes in colon cancer. The findings indicated that MIR210HG may exert its effects in colon cancer by modulating energy metabolism and cell adhesion. Further predictions suggested that MIR210HG may also affect colon cancer via transcription and post-transcriptional processing. Together, these results provide evidence of the transcriptional regulatory network of MIR210HG in colon cancer, and suggest that MIR210HG may serve a potential role as a novel biomarker and therapeutic target. Despite the multiple bioinformatic and statistical methods used in the present study, it does have some limitations, and further studies are required to verify these results. Experimental evidence will be provided in future studies; for example, the high expression level of MR210HG may regulate the expression levels of screened MIR210HG-related genes, or induce the phenotype obtained with bioinformatics analyses (such as the metastasis of colon cancer cells), which will be assessed using cell-based assays with colon cancer cell lines.

Acknowledgements

Not applicable.

Funding

The present study was supported by grants from the National Natural Science Foundation of China (grant nos. 81760094 and 31602111), the Science and Technology Research Project for Youth of Educational Commission in Jiangxi Province (grant no. GJJ160238), the Science and Technology Foundation for Youths of Jiangxi province (grant no. 20171BAB215021) and the Science and Technology Project of Health and Family Planning commission of Jiangxi Province (grant. no. 20175081).

Availability of data and materials

The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Authors' contributions

YL and ZR designed the study. ZR, ZX and ZL performed the data analyses. All authors contributed to the conception of the study and drafted the manuscript. All authors contributed significantly in writing the manuscript and read and approved the final manuscript.

Ethics approval and consent to participate

Not applicable.

Patient consent for publication

Not applicable.

Competing interests

The authors declare that they have no competing interests.

Glossary

Abbreviations

Abbreviations:

References