Skip to main content

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

Transcriptome analysis of miRNA–lncRNA–mRNA interactions in the malignant transformation process of gastric cancer initiation

Cancer Gene Therapy volume 24, pages 267–275 (2017)Cite this article

Subjects

Abstract

Gastric cancer is a common heterogeneous malignancy that is pathologically characterized by the development of chronic atrophic gastritis and dysplasia of the epithelium. The pathogenic mechanisms at the molecular level are completely unclear. In the present study, we first address the expression network and miRNA–lncRNA–mRNA interaction in the malignant transformation process from normal mucosa and atrophic gastritis to a tumor. First, the long non-coding RNA (lncRNA), micro RNA (miRNA) and messenger RNA (mRNA) expression profiles of normal gastric mucosa, atrophic gastritis, gastric cancer and the peri-tumor tissues were analyzed using microarrays. Then, bioinformatic analyses were used to predict the gene targets and analyse their potential roles in tumorigenesis and progression of gastric cancer. Finally, an intersection analysis of microarray data showed that 10 miRNAs and 11 lncRNAs were detected in the malignant transformation process from normal mucosa tissues and atrophic gastritis to tumorigenesis, and many miRNAs and lncRNAs were novel and had important roles. Meanwhile, the significant pathways and biological functions regulated by the deregulated 21 non-coding genes were enriched. In conclusion, our work provides an important theoretical, experimental and clinical foundation for further research on more effective targets for the diagnosis, therapy and prognosis of gastric cancer.

Download PDF

Introduction

Gastric cancer is a common heterogeneous malignant disease. The incidence and mortality of gastric cancer have fallen to a large degree in some geographic areas and countries according to relevant data, but it remains the most notable cause of cancer death in some countries, such as China, as well as in other ethnic groups in different parts of the world.1, 2, 3 Despite substantial progress in diagnosis and treatment, the 5-year survival rate of gastric cancer remains unsatisfactory.4

The possible occurrence and development of gastric cancer represent complex changeable processes that are associated with bacterial factors, environmental factors, genetic susceptibility or comprehensive effects. Genetic and epigenetic alterations are inescapable and strong risk factors for gastric cancer, but the pathogenic mechanisms underlying the disease at the molecular level are completely unclear. In the past few decades, the role of deregulated genes in gastric cancer has been verified. For example, the key mediators of tumor promotion, the NF-κB and STAT transcription factors, are activated,5, 6 and polymorphisms of proinflammatory genes, such as interleukin 1-β (IL-1 β), tumor necrosis factor-α (TNF-α) and IL-10, confer a manifold increased risk of developing gastric cancer.7 In addition, microarray-based deregulated gene studies show that certain genes are differentially expressed between gastric cancer and healthy human gastric mucosa,8 including non-coding genes.9

Non-coding RNAs, such as micro RNAs (miRNAs) and long non-coding RNAs (lncRNAs), are defined as gene transcripts with limited or no protein-coding capacity.10 Previous studies have stated that non-coding RNAs regulate cell proliferation, differentiation, apoptosis and invasion and are involved in the progression and metastasis of tumors.11, 12, 13 For example, miR-29 and miR-146b as well as lncRNA HOTAIR and lncRNA GAPLINC might have key roles in carcinogenesis.6, 14, 15, 16 Although the roles of non-coding RNAs and their potential mechanisms in gastric cancer have been partially reported, a more specific understanding of the mechanisms of the interactions between non-coding RNAs and protein-coding genes in gastric cancer is needed.

Carcinoma of the stomach can be classified into two main types according to its histological appearance, that is, diffuse and intestinal types, or its original location, that is, non-cardia and cardia cancer.17 Intestinal-type gastric cancer, which is highly prevalent and accounts for a notable proportion of cancer, has an array of gastric mucosa pathologies that range from superficial gastritis to atrophic gastritis, intestinal metaplasia, dysplasia and, ultimately, malignancy.18 Gene polymorphisms and the potential molecular characteristics in the specific evolutionary course of the disease have been reported previously,19, 20 but the majority of efforts have focused on investigating protein-coding genes. Non-coding RNAs, which are abundant and important regulators of coding genes during the multistep development of human cancers, has a potential function in the initiation and development of gastric cancer.

In this study, we mainly focused on the expression network and interactions between non-coding RNAs and coding genes, mainly mRNAs, and the biological pathways revealed in the tumorigenesis and the progression of gastric cancer, which could promote the development of multitargeted prevention and therapies for gastric cancer in the future.

Materials and methods

Clinical specimens

All patients in the medical study came from the Affiliated Provincial Hospital of Anhui Medical University. The ethics committee sanctioned the collection of the specimens, and the patients gave written informed consent. The clinical specimens included five normal gastric mucosal tissues, five severe chronic atrophic gastritis tissues with intestinal metaplasia, five advanced human gastric cancer tissues and five peri-tumor tissues. None of the gastric cancer patients received preoperative treatment, including chemotherapy or radiotherapy. The peri-tumor tissues were taken at a distance of at least 5 cm from the tumor and there were no obvious tumor cells. The biopsies of the clinical specimens were obtained during outpatient visits during the endoscopic procedure and were immediately frozen in liquid nitrogen and then stored at −80 °C. All tissues were diagnosed histologically.

Total RNA extraction and quality control

The total RNA of all specimens was extracted using TRIzol (Invitrogen, Carlsbad, CA, USA) according to the instruction manual. The lncRNAs and miRNAs were purified using the NucleoSpin RNA clean-up Reagent kit (740.948.250, Macherey-Nagel, Düren, Germany) and the mirVana miRNA Isolation Kit (AM1561, Ambion, Carlsbad, CA, USA), respectively. The concentration and quality of the RNA were measured by UV absorbance at 260 and 280 nm (260/280 nm) on a Nanodrop 2000 spectrophotometer (Thermo Scientific, Worcester, MA, USA).

Microarray study of RNA expression in clinical samples

Briefly, for the lncRNA and mRNA, the RNA samples underwent reverse transcription to synthesize double-stranded complementary DNA (cDNA), and these were then transcribed to cRNA in vitro. The cRNA was reverse transcribed into cDNA. The cDNA was fluorescently labeled and was hybridized using a 4 × 180K lncRNA+mRNA Human Gene Expression Microarray V4.0 (CapitalBio Technology, Beijing, China) (GeeDom), which included 38 000 lncRNA probes and 34 000 mRNA probes containing the latest and most comprehensive lncRNA sequence information, based on the latest version of 22 444 sequence information in the GENCODE/ENSEMBL database and integrated with the Refseq, UCSC, H-InvDB, Human lincRNA catalog, NRED, lncRNAdb, RNAdb and 13 other major lncRNA databases and related literature.

For the miRNA, the purified miRNA was labeled and hybridized using Affymetrix miRNA 4.0 (Affymetrix, Santa Clara, CA, USA) (GeneChip), which can detect a total of 30424 mature miRNAs (all species) in the miRBase Release 20 database, including 5214 human miRNAs and 1996 human snoRNAs and snaRNAs. The microarray hybridization, washing and scanning were completed by the CapitalBio Technology Corporation (Beijing, China).

Microarray statistical analysis

Agilent Feature Extraction (v10.7) software (Santa Clara, CA, USA) was used to analyse the hybridization map and extract the data. The chip fluorescence scanning images were saved as.DAT files for analysis by AGCC software (Affymetrix GeneChip Command Console Software). The data normalization and difference analysis used Agilent GeneSpring software. The two groups of sample data underwent a t-test analysis to obtain the corrected P-values and Fold Change values. The criteria for the differentially expressed genes that changed more than twofold, and the differences were considered to be statistically significant at P<0.05. Cluster 3.0 software (Michiel de Hoon, Human Genome Center, University of Tokyo) was also used for the cluster analysis and graphical display. Venny 2.1.0 online software (Oliveros, J.C., CNB-CSIC. http://bioinfogp.cnb.csic.es/tools/venny/index.html) was used to find the intersection, and GraphPad Prism 6 (La Jolla, CA, USA) was used for the graphical display.

Target prediction

Target gene prediction of lncRNAs is divided into cis-prediction and trans-prediction. Based on the co-expression of the lncRNAs and mRNAs (Correlation>0.99 or Correlation<−0.99 and P-value<0.05), the cis-prediction involved the lncRNA–mRNA pair located within 10KB in the genome, and the trans-prediction involved the sequence similarity lncRNA–mRNA pairs after comparing the lncRNA and mRNA sequences (3′UTR) using the BLAST tool.

The target gene prediction of miRNA is based on the analysis of miRWalk2.0 (Heidelberg, Germany), which provides 12 miRNA target gene prediction programs, miRWalk, DIANA-microT v4.0, miRanda-rel2010, miRBridge, miRDB 4.0, miRmap, miRNAMap, PicTar2, PITA, RNA22 v2, RNAhybrid 2.1 and Targetscan 6.2, the prediction results were screened for at least 6 programs predictions.

Pathways enrichment analysis

The enrichment analysis used KOBAS annotation (KEGG Orthology Based Annotation System), which includes seven pathway enrichment analyses (KEGG PATHWAY, PID Curated, PID BioCarta, PID Reactome, BioCyc, Reactome and Panther), five disease analyses (OMIM, KEGG DISEASE, FunDO, GAD, NHGRI) and Gene Ontology (GO). Cytoscape software V3.2.1 Cytoscape software V3.2.1 (San Diego, CA, USA) was used to decipher the pathways network and understand their biological functions, and the data output was received in Excel spreadsheets.

Results

Gene expression profile in atrophic gastritis and gastric cancer

To globally analyse the genetic and epigenetic variations of the initiation of gastric cancer, 20 clinical specimens were collected for microarray analysis, including five human healthy normal gastric mucosal tissues, five severe atrophic gastritis tissues accompanied by intestinal metaplasia, five advanced gastric cancer tissues and five corresponding peri-tumor tissues (Table 1). The lncRNA+mRNA Human Gene Expression Microarray V4.0 and the Affymetrix GeneChip miRNA 4.0 were used to detect the gene transcription profiles of each clinical sample. Finally, the deregulated non-coding and coding genes were selected (at least 2.0-fold changes and P<0.05) in each pairwise comparison; compared with normal tissues, 47 lncRNAs, 62 miRNAs and 103 mRNAs were upregulated and 129 lncRNAs, 24 miRNAs and 112 mRNAs were downregulated in atrophic gastritis tissues, whereas 894 lncRNAs, 196 miRNAs and 1447 mRNAs were upregulated and 534 lncRNAs, 7 miRNAs and 1588 mRNAs were downregulated in gastric cancer tissues (Table 2). In addition, the microarray results revealed several important deregulations as well as the top 10 significantly changed lncRNAs, miRNAs and mRNAs, which are listed in Table 3.

Table 1 Information of clinical specimens for microarray
Full size table
Table 2 The number of significant deregulated genes on microarray in groups
Full size table
Table 3 The top10 significantly deregulated genes in each comparison
Full size table

Expression status of the 21 selected miRNAs and lncRNAs

Moreover, the intersection analysis showed that 10 miRNAs and 11 lncRNAs were detected in the malignant transformation process from normal mucosa tissues and atrophic gastritis to tumorigenesis. Among the 21 deregulated lncRNAs and miRNAs, 6 miRNAs (hsa-miR-4417, hsa-miR-6165, hsa-miR-451a, hsa-miR-421, hsa-miR-18b-5p and hsa-miR-181d-5p) and 2 lncRNA probe sets (p33715 and p29027 also known as H19) were upregulated in the malignant transformation process; 3 miRNAs (hsa-miR-4532, hsa-miR-204-3p and hsa-miR-1281) and 8 lncRNA probe sets (p7811, p6625 also known as MIR22GH, p1457, p8725 also known as CTD-3064H18.6, p38190-v4, p4300, p43909-v4 and p36288-v4) were downregulated progressively. It is interesting that has-miR-424-3p and p38901-v4 were downregulated in atrophic gastritis but upregulated in gastric cancer. In addition, hierarchical clustering of samples revealed the partitioning of samples into three major groups (Figure 1), and the miRNA and lncRNA expression level was further analyzed in the malignant transformation process using the column table directly (Figure 2). The majority of these deregulated genes remain novel; only lncRNA H19, has-miR-451a, downregulated has-miR-421 and has-miR-18b-5p have been reported to be involved in gastric cancer.21, 22, 23, 24 To identify the characteristics of the 11 deregulated lncRNAs, we analyzed their location in the genome, and the results indicated that most of them are long intergenic non-coding RNAs (lincRNAs) (Table 4).

Figure 1
figure 1

The expression of the deregulated genes in atrophic gastritis and gastric cancer. (a) An intersection diagram of the number of deregulated micro RNAs (miRNAs) and long non-coding RNAs (lncRNAs) (at least 2.0-fold changes and P<0.05) detected by microarray. The red circle indicates the number of deregulated genes in atrophic gastritis vs normal gastric mucosa, the yellow circle indicates the number of deregulated genes in gastric cancer vs atrophic, and the green circle indicates gastric cancer vs normal mucosa. The overlapping color sections represent the number of common deregulated genes between corresponding comparison groups. In total, 10 miRNAs and 11 lncRNAs were detected in the malignant transformation process from normal mucosal tissue to atrophic gastritis to tumor tissue. (b) Heat maps showing the expression levels of 10 miRNAs and 11 lncRNAs in each clinical tissue. The right column shows the genes. The red-colored pixels correspond to an increased abundance of the gene in the indicated sample, whereas the green pixels indicate decreased levels.

Full size image
Figure 2
figure 2

The microarray results of the 21 micro RNAs (miRNAs) and long non-coding RNAs (lncRNAs). The expression level of the 21 deregulated miRNAs and lncRNAs based on the microarray and fold change level of these genes in atrophic gastritis, tumor and peri-tumor tissues all compared with normal gastric mucosa tissues are shown as the mean±standard deviations.

Full size image
Table 4 Supplementary information for the deregulated 11 lncRNA both in the malignant transformation process
Full size table

Construction of the miRNA–lncRNA–mRNA co-expression network

A miRNA–lncRNA–mRNA co-expression network of the 21 selected miRNAs and lncRNAs was constructed in our study. First, the target coding genes of the 10 miRNAs and 11 lncRNAs were predicted. The graphical results show the correlation of miRNA–mRNA and lncRNA–mRNA (criteria: Correlation>0.99 or Correlation<−0.99, and P-value<0.05). Using each miRNA or lncRNA as the center of the network, we can clearly see possible regulated target genes (Figures 3a and b). We chose the significant correlation pairs in the 10 deregulated miRNAs and 11 lncRNAs mentioned above to construct the co-expression network. The network analysis showed that FBP1 and IGF2 co-expression with H19, CA2 co-expression with p4300, p36288_v4, has-miR-18b-5p and has-miR-181d-5p, and MT1F co-expression with p4300, has-miR-18b-5p and p43909-v4. S100A8, S100A9 and S100A12 were involved in the interactions among has-miR-204-3p, has-miR-1281 and has-miR-4532. We found that p4300, p33715, has-miR-18b-5p and has-miR-181d-5p were of great importance in regulating the network (Figure 3c).

Figure 3
figure 3figure 3

Network plots of the 21 non-coding RNAs (ncRNAs). (a) The predicted target genes of the deregulated 10 micro RNAs (miRNAs), (b) the predicted target genes of the partly deregulated 11 lncRNAs and (c) the network of the miRNA–lncRNA–mRNA co-relationships. In all the networks, taking the deregulated miRNAs and lncRNAs as the hub, the yellow circles represent lncRNA, the green circles represent miRNA and the red circles represent relative mRNA (Correlation>0.99 or Correlation<-0.99, and P-value<0.05). The relative size of each hub in (c) indicates the degree of connectivity (number of edges) for each gene.

Full size image

Enrichment of the biological functions and pathway networks

We predicted the possible enrichment pathways and biological processes in the malignant transformation process according to KOBAS annotation, and the main functions of the deregulated expressed mRNAs were analyzed using GO analysis. The analysis clearly demonstrated that some important functions were activated by the 21 special miRNAs and lncRNAs. For example, some cell cycling genes, including cyclin-dependent kinases and inhibitors, for example, CDKN2A, were regulated by p4300, H19, has-miR-4532, has-miR-1281 and has-miR-204-3p. The proliferation of the negative feedback regulator DUSP5 was regulated by p4300, H19 and has-miR204-3p.

GO analysis also revealed that the inflammatory process occupied a very prominent feature within the 21 special miRNAs and lncRNAs, including chemokine ligands (for example, CCL3, CCL20, CXCL1 and CXCL8), chemokine receptors (for example, CCR3, CCR4, CCR7 and CXCR2), the NF-kappa-B signaling pathway (for example, CDKN2A) and the TNF pathway (for example, TNFRSF11B). Furthermore, calcium- and zinc-binding proteins (S100A8, S100A9 and S100A12) and Toll-like receptors (for example, TLR5 and TLR8) were all regulated. The network diagrams involving the above genes showed a greater relationship with H19, p4300, p33175, has-miR-4532, has-miR-1281 and has-miR-204-3p.

Moreover, our study showed that 17 potential pathways were enriched in KEGG and 54 were enriched in Reatome from normal tissues to atrophic gastritis. There were 32 KEGG enriched pathways and 154 Reatome enriched pathways in atrophic gastritis to gastric cancer, and there were 35 KEGG enriched pathways and Reatome enriched pathways in normal to gastric cancer. For example, chemical carcinogenesis, metabolism, cytokine-cytokine receptor interaction, and the Gastrin-CREB signaling pathway via PKC and MAPK were all enriched in the above biological processes. Detailed descriptions of the top 10 enriched pathways in each process in which the deregulated coding genes participated in are listed in Table 3. In addition, a partial network among the 21 specific deregulated miRNAs and lncRNAs and the pathways of their predicted target genes were built (Figure 4). From the network of the miRNA–lncRNA pathway, the chemokine, Toll-like receptor, calcium, Wnt, MAPK, Jak-STAT, ErbB signaling pathways were involved. Has-miR-1281, has-miR-4532 and has-miR-204-3p together might regulate the M/G1 transition, suggesting that they may affect cell proliferation via the alter cell cycle; p36288-v4 and MIR22HG were related to cell migration by co-regulating the adherens junction. In addition, H19 and has-miR-18b-5p might regulate the MAPK signaling pathway.

Figure 4
figure 4

Network of miRNA–lncRNA pathways. The network of the specific deregulated micro RNAs (miRNAs), deregulated long non-coding RNAs (lncRNAs) and partly deregulated KEGG pathways of the predicted target genes. Yellow hexagons: 10 deregulated miRNAs. Green hexagons: 11 deregulated lncRNAs in the malignant transformation process.

Full size image

Discussion

Because researches in the field of non-coding RNAs are expanding rapidly, it is becoming clear that lncRNA and miRNA have critical roles in biological functions in malignancies, including in tumorigenesis and the development of gastric cancer. The deregulation of of the expression of lncRNA and miRNA is relevant to the phenotype and growth characteristics of tumors and affects the clinicopathological characteristics, prognosis and survival of patients. Gastric cancer is a common malignancy worldwide, and most cases of this disease are diagnosed at an advanced stage. The early diagnosis rate of gastric cancer remains low in China and other developing countries. Thus, it is more important to elucidate the potential molecular functions and underlying enrichment pathways in the initiation of gastric cancer. Atrophic gastritis is considered the precancerous lesion of gastric cancer, and it is important to explore the modification of genes in atrophic gastritis for the early diagnosis and early intervention of gastric cancer. Specific non-coding RNAs, such as lncRNAs or/and miRNAs, are potential candidates for the diagnosis, treatment or biomarkers of the prognosis in gastric cancer in the future.

Currently, studies on coding genes in gastric cancer,8, 25 non-coding genes in gastric cancer,26, 27 and genes in atrophic gastritis9 have been reported, whereas the effects of the interaction between coding genes and non-coding genes in the malignant transformation process from normal mucosa tissues and atrophic gastritis to tumorigenesis have not been reported to date. In our study, we first simultaneously investigated the lncRNA, miRNA and mRNA expression signatures in the malignant transformation process using a microarray and identified a network of lncRNA–miRNA–mRNA co-regulation. Finally, we analyzed and described the underling biological functions of this network.

According to the results of the microarray, we identified a set of differentially expressed miRNAs, lncRNAs and mRNAs between normal gastric mucosa, atrophic gastritis and tumor and peri-tumor tissues. This deregulation signified the critical and potential roles of these RNAs in the initiation of gastric cancer and indicated that they might have other roles in metastasis and invasion during cancer development. Furthermore, we compared the expression of miRNAs, lncRNAs and mRNAs between tumor tissues and peri-tumor tissues. Because of the higher heterogeneity of the peri-tumor tissue, it was unclear whether these RNA species were inhibited at the molecular level, although no obvious tumor cells were demonstrated histologically. In addition, the paired tumor and peri-tumor tissues came from the same patient, and it cannot be guaranteed that the patient's genome only changed in the tumor region. Therefore, the comparison of tumors and peri-tumors provided supporting validation of the results.

The intersection of the pairwise comparisons among the normal tissues, atrophic gastritis and gastric cancer showed that 10 miRNAs and 11 lncRNAs were deregulated. Some of the altered non-coding RNAs were reported in other studies previously. For example, the overexpression of H19 binding ISM1 in gastric cancer was reported to promote cell proliferation, promoted cell proliferation, migration, invasion and metastasis,22 and another report published that H19 might serve as a promising biomarker for the early detection and prognosis of gastric cancer.28 Has-miR-421 was reported to be upregulated in and a risk factor for gastric cancer in association with the tumor genotype, lymph node metastasis and prognosis. E-cadherin and caspase-3 were identified as targets of miR-421, and low levels of miR-421 indicated a significantly longer overall survival.23, 29 Has-miR-18b-5p was also identified in another microarray study.24 These findings are consistent with our results. The published research showed that has-miR-451a was downregulated and acted as a potential tumor suppressor by targeting TSC1 in gastric cancer,21 whereas in our study, has-miR-451a was obviously upregulated in atrophic gastritis and gastric cancer compared with normal mucosa.

Encouragingly, our study also identified many novel lncRNAs and miRNAs that, until now, have not been linked to gastric cancer or other tumors, and functional analysis proved their important biological roles. For example, the lncRNAs p4300, p33715, has-miR-1281, has-miR-204-3p and has-miR-4532 have more important roles. Their effects on tumorigenesis, progression and metastasis are worthy of further study. We also found that a majority of the deregulated miRNAs (71.79%) were upregulated, which suggests that increased miRNAs have a more important role than decreased miRNAs, but there were no significant results in the analysis of the lncRNAs and mRNAs.

The critical role of lncRNA and miRNA in biological functions is the regulation of mRNA expression. Past studies have suggested that the roles of lncRNAs may be related to neighboring coding genes.30 In our study, we leveraged three strategies for the functional predictions of the deregulated lncRNAs and miRNAs, including predictions of the target coding the miRNA and lncRNA genes, the miRNA–mRNA network and the lncRNA–mRNA network. Using the non-coding RNA as the center, we could clearly see its possible target genes or co-target genes regulated by the related non-coding RNAs. Some of these targeted mRNAs were reported to be aberrantly expressed in gastric cancer, other cancers or other diseases.

Another experiment utilized the miRNA–lncRNA–mRNA co-expression network. This is the first attempt to develop a more comprehensive network that includes lncRNAs, miRNAs and mRNAs to describe the genetics of the development of gastric cancer and from precancerous lesions from normal mucosa. The transcriptional patterns of miRNA and lncRNA are very complex, especially lncRNA, because of the relationship between the location and coding genes may be within the intronic regions of the coding genes, between two coding genes or within overlapping exons in a sense or antisense orientation.31 In addition, the human genome is not a combination of isolated genes but a complex networked system; thus, the co-relationship among different genes seems to be extremely important. Of course, our research also focused on this point. SPEM and TFF have significant roles in the intestinal metaplasia,32 Here, we first showed that TFF3 can be regulated by has-miR-4532, p33715, p4300 and has-miR-181d-5p collectively, demonstrating that these non-coding RNAs have key roles in the malignant transformation process from normal mucosa to atrophic gastritis to tumor.

Lastly, the miRNA–lncRNA pathway network was studied. Differentially expressed genes control the occurrence and development of tumors by regulating signaling pathways and biological processes. Enriched pathways, such as the MAPK signaling pathway, are important in gastric cancer.33 Interferon alpha/beta signaling, the chemokine signaling pathway, proteoglycans in cancer and the adherens junction were regulated by specific miRNAs and lncRNAs in gastric cancer. According to the findings, the dominance of the miRNAs and lncRNAs in gastric carcinogenesis-correlated modules suggests the significance of miRNAs and lncRNAs.

According to a number of studies, it is becoming increasingly clear that miRNAs and lncRNAs can affect the expression of a variety of genes, and insights into their regulatory mechanisms strengthen our understanding of these functions. In our study, whether the novel lncRNAs and miRNAs directly or indirectly regulated target genes remains unknown. In particular, lncRNAs are likely to fall into different classes with different mechanisms. Studies have shown that lncRNAs can activate or repress gene expression through a chromatin recruitment mechanism in cis or in trans. PRC2 is recruited to the X chromosome by both lncRNA RepA and lncRNA Xist, leading to X inactivation in cis,34 and the lncRNA HOTAIR can interact with a histone-modified complex to control HoxD gene cluster expression in trans.35 The lncRNAs can also act as enhancers and promoters to regulate gene transcription,36, 37 and DNA methylation is also very important.38 Many lncRNAs can provide miRNA-binding sites to regulate the expression of protein-encoded genes. For example, the lncRNA GAPLINK contains a miR-211-3p binding site and the target gene CD44 contains a miR-211-3p binding site, promoting CD44 mRNA expression.16 The lncRNA SNHG5 also contains a putative miR-32-binding site that regulates gastric cancer cell proliferation and migration.39 The lncRNA RMRP can modulate the cell cycle by regulating Cyclin D2 expression and by acting on miR-206 in gastric cancer.40 The data reinforce the importance of the lncRNA–miRNA–mRNA network shown in this study; the network also provides guidance for the study of the related mechanisms.

On the basis of the variations of the expression levels of non-coding and coding genes, we originally predicted the potential roles of these miRNAs and lncRNAs in the malignant transformation process from normal mucosa and atrophic gastritis to a tumor. Nevertheless, the exact functions of specific non-coding RNAs need to be evaluated by large-sample experiments over time. We also identified some novel ncRNA–mRNA interactions for exploring the functional mechanisms and clinical features of gastric cancer.

In conclusion, our study identified and analyzed the co-relationship network of miRNA–lncRNA–mRNA in the malignant transformation process from normal mucosa and atrophic gastritis to a tumor and showed that this network has crucial biological roles in the occurrence and development of gastric cancer. This study provides an important theoretical, experimental and clinical foundation for further research on more effective targets for the diagnosis, therapy and prognosis of gastric cancer.

References

  1. Arnold M, Karim-Kos HE, Coebergh JW, Byrnes G, Antilla A, Ferlay J et al. Recent trends in incidence of five common cancers in 26 European countries since 1988: analysis of the European Cancer Observatory. Eur J Cancer 2015; 51: 1164–1187.

    Article  Google Scholar 

  2. Chen W, Zheng R, Baade PD, Zhang S, Zeng H, Bray F et al. Cancer statistics in China, 2015. Cancer J Clin 2016; 66: 115–132.

    Article  Google Scholar 

  3. Siegel RL, Miller KD, Jemal A . Cancer statistics, 2016. Cancer J Clin 2016; 66: 7–30.

    Article  Google Scholar 

  4. Luebeck EG, Curtius K, Jeon J, Hazelton WD . Impact of tumor progression on cancer incidence curves. Cancer Res 2013; 73: 1086–1096.

    CAS  Article  Google Scholar 

  5. Li J, Deng Z, Wang Z, Wang D, Zhang L, Su Q et al. Zipper-interacting protein kinase promotes epithelial-mesenchymal transition, invasion and metastasis through AKT and NF-kB signaling and is associated with metastasis and poor prognosis in gastric cancer patients. Oncotarget 2015; 6: 8323–8338.

    PubMed  PubMed Central  Google Scholar 

  6. Xiang M, Birkbak NJ, Vafaizadeh V, Walker SR, Yeh JE, Liu S et al. STAT3 induction of miR-146b forms a feedback loop to inhibit the NF-kappaB to IL-6 signaling axis and STAT3-driven cancer phenotypes. Sci Signal 2014; 7: ra11.

    Article  Google Scholar 

  7. Tsujimoto H, Ono S, Ichikura T, Matsumoto Y, Yamamoto J, Hase K . Roles of inflammatory cytokines in the progression of gastric cancer: friends or foes? Gastric Cancer 2010; 13: 212–221.

    CAS  Article  Google Scholar 

  8. Ooi WF, Xing M, Xu C, Yao X, Ramlee MK, Lim MC et al. Epigenomic profiling of primary gastric adenocarcinoma reveals super-enhancer heterogeneity. Nat Commun 2016; 7: 12983.

    CAS  Article  Google Scholar 

  9. Zhang X, Peng Y, Jin Z, Huang W, Cheng Y, Liu Y et al. Integrated miRNA profiling and bioinformatics analyses reveal potential causative miRNAs in gastric adenocarcinoma. Oncotarget 2015; 6: 32878–32889.

    PubMed  PubMed Central  Google Scholar 

  10. Alexander RP, Fang G, Rozowsky J, Snyder M, Gerstein MB . Annotating non-coding regions of the genome. Nat Rev Genet 2010; 11: 559–571.

    CAS  Article  Google Scholar 

  11. Bartel DP . MicroRNAs: genomics, biogenesis, mechanism, and function. Cell 2004; 116: 281–297.

    CAS  Article  Google Scholar 

  12. Guttman M, Rinn JL . Modular regulatory principles of large non-coding RNAs. Nature 2012; 482: 339–346.

    CAS  Article  Google Scholar 

  13. Wan X, Ding X, Chen S, Song H, Jiang H, Fang Y et al. The functional sites of miRNAs and lncRNAs in gastric carcinogenesis. Tumour Biol 2015; 36: 521–532.

    CAS  Article  Google Scholar 

  14. Gong J, Li J, Wang Y, Liu C, Jia H, Jiang C et al. Characterization of microRNA-29 family expression and investigation of their mechanistic roles in gastric cancer. Carcinogenesis 2014; 35: 497–506.

    CAS  Article  Google Scholar 

  15. Kogo R, Shimamura T, Mimori K, Kawahara K, Imoto S, Sudo T et al. Long noncoding RNA HOTAIR regulates polycomb-dependent chromatin modification and is associated with poor prognosis in colorectal cancers. Cancer Res 2011; 71: 6320–6326.

    CAS  Article  Google Scholar 

  16. Hu Y, Wang J, Qian J, Kong X, Tang J, Wang Y et al. Long noncoding RNA GAPLINC regulates CD44-dependent cell invasiveness and associates with poor prognosis of gastric cancer. Cancer Res 2014; 74: 6890–6902.

    CAS  Article  Google Scholar 

  17. Lauren P . The two histological main types of gastric carcinoma: diffuse and so-called intestinal-type carcinoma. An attempt at a histo-clinical classification. Acta Pathol Microbiol Scand 1965; 64: 31–49.

    CAS  Article  Google Scholar 

  18. Correa P . Human gastric carcinogenesis: a multistep and multifactorial process—First American Cancer Society Award Lecture on Cancer Epidemiology and Prevention. Cancer Res 1992; 52: 6735–6740.

    CAS  PubMed  Google Scholar 

  19. Xu Q, Wu YF, Li Y, He CY, Sun LP, Liu JW et al. SNP-SNP interactions of three new pri-miRNAs with the target gene PGC and multidimensional analysis of H. pylori in the gastric cancer/atrophic gastritis risk in a Chinese population. Oncotarget 2016; 7: 23700–23714.

    PubMed  PubMed Central  Google Scholar 

  20. Kupcinskas J, Wex T, Link A, Leja M, Bruzaite I, Steponaitiene R et al. Gene polymorphisms of micrornas in Helicobacter pylori-induced high risk atrophic gastritis and gastric cancer. PloS One 2014; 9: e87467.

    Article  Google Scholar 

  21. Riquelme I, Tapia O, Leal P, Sandoval A, Varga MG, Letelier P et al. miR-101-2, miR-125b-2 and miR-451a act as potential tumor suppressors in gastric cancer through regulation of the PI3K/AKT/mTOR pathway. Cell Oncol 2016; 39: 23–33.

    CAS  Article  Google Scholar 

  22. Li H, Yu B, Li J, Su L, Yan M, Zhu Z et al. Overexpression of lncRNA H19 enhances carcinogenesis and metastasis of gastric cancer. Oncotarget 2014; 5: 2318–2329.

    PubMed  PubMed Central  Google Scholar 

  23. Jin X, Yu N . MicroRNA-421 gene polymorphism in gastric carcinoma. Med Sci Monit 2016; 22: 1467–1471.

    CAS  Article  Google Scholar 

  24. Xia T, Liao Q, Jiang X, Shao Y, Xiao B, Xi Y et al. Long noncoding RNA associated-competing endogenous RNAs in gastric cancer. Sci Rep 2014; 4: 6088.

    CAS  Article  Google Scholar 

  25. Yuen ST, Leung SY . Genomics study of gastric cancer and its molecular subtypes. Adv Exp Med Biol 2016; 908: 419–439.

    Article  Google Scholar 

  26. Bibi F, Naseer MI, Alvi SA, Yasir M, Jiman-Fatani AA, Sawan A et al. microRNA analysis of gastric cancer patients from Saudi Arabian population. BMC Genomics 2016; 17: 751.

    Article  Google Scholar 

  27. Fan ZY, Liu W, Yan C, Zhu ZL, Xu W, Li JF et al. Identification of a five-lncRNA signature for the diagnosis and prognosis of gastric cancer. Tumour Biol 2016; 37: 13265–13277.

    CAS  Article  Google Scholar 

  28. Chen JS, Wang YF, Zhang XQ, Lv JM, Li Y, Liu XX et al. H19 serves as a diagnostic biomarker and up-regulation of H19 expression contributes to poor prognosis in patients with gastric cancer. Neoplasma 2016; 63: 223–230.

    CAS  PubMed  Google Scholar 

  29. Ge X, Liu X, Lin F, Li P, Liu K, Geng R et al. MicroRNA-421 regulated by HIF-1alpha promotes metastasis, inhibits apoptosis, and induces cisplatin resistance by targeting E-cadherin and caspase-3 in gastric cancer. Oncotarget 2016; 7: 24466–24482.

    PubMed  PubMed Central  Google Scholar 

  30. Ponjavic J, Oliver PL, Lunter G, Ponting CP . Genomic and transcriptional co-localization of protein-coding and long non-coding RNA pairs in the developing brain. PLoS Genet 2009; 5: e1000617.

    Article  Google Scholar 

  31. Carninci P, Kasukawa T, Katayama S, Gough J, Frith MC, Maeda N et al. The transcriptional landscape of the mammalian genome. Science 2005; 309: 1559–1563.

    CAS  Article  Google Scholar 

  32. Goldenring JR, Nomura S . Differentiation of the gastric mucosa III. Animal models of oxyntic atrophy and metaplasia. Am J Physiol 2006; 291: G999–1004.

    CAS  Google Scholar 

  33. Khan SA, Amnekar R, Khade B, Barreto SG, Ramadwar M, Shrikhande SV et al. p38-MAPK/MSK1-mediated overexpression of histone H3 serine 10 phosphorylation defines distance-dependent prognostic value of negative resection margin in gastric cancer. Clin Epigenet 2016; 8: 88.

    Article  Google Scholar 

  34. Zhao J, Sun BK, Erwin JA, Song JJ, Lee JT . Polycomb proteins targeted by a short repeat RNA to the mouse X chromosome. Science 2008; 322: 750–756.

    CAS  Article  Google Scholar 

  35. Li L, Liu B, Wapinski OL, Tsai MC, Qu K, Zhang J et al. Targeted disruption of Hotair leads to homeotic transformation and gene derepression. Cell Rep 2013; 5: 3–12.

    CAS  Article  Google Scholar 

  36. Yang L, Lin C, Jin C, Yang JC, Tanasa B, Li W et al. lncRNA-dependent mechanisms of androgen-receptor-regulated gene activation programs. Nature 2013; 500: 598–602.

    CAS  Article  Google Scholar 

  37. Trimarchi T, Bilal E, Ntziachristos P, Fabbri G, Dalla-Favera R, Tsirigos A et al. Genome-wide mapping and characterization of Notch-regulated long noncoding RNAs in acute leukemia. Cell 2014; 158: 593–606.

    CAS  Article  Google Scholar 

  38. Wang L, Zhao Y, Bao X, Zhu X, Kwok YK, Sun K et al. LncRNA Dum interacts with Dnmts to regulate Dppa2 expression during myogenic differentiation and muscle regeneration. Cell Res 2015; 25: 335–350.

    Article  Google Scholar 

  39. Zhao L, Han T, Li Y, Sun J, Zhang S, Liu Y et al. The lncRNA SNHG5/miR-32 axis regulates gastric cancer cell proliferation and migration by targeting KLF4. FASEB J 2016; 31: 893–903.

    Article  Google Scholar 

  40. Shao Y, Ye M, Li Q, Sun W, Ye G, Zhang X et al. LncRNA-RMRP promotes carcinogenesis by acting as a miR-206 sponge and is used as a novel biomarker for gastric cancer. Oncotarget 2016; 7: 37812–37824.

    PubMed  PubMed Central  Google Scholar 

Download references

Acknowledgements

The work was supported by the New Century Excellent Talents in University, Ministry of Education of China, to Wang H; the Wanjiang Scholars Program of Anhui Province of China to Wang H; the Public Welfare Technology Application Research Linkage Project of Anhui Province of China, 1604f0804018 to Ding XP.

Author information

Affiliations

  1. Department of Gastroenterology, Affiliated Provincial Hospital of Anhui Medical University, Hefei, China

    Y Mao, R Liu & X Ding

  2. Department of Pharmacy, Anhui Medical University, Hefei, China

    H Zhou

  3. Department of Geriatrics, Affiliated Provincial Hospital of Anhui Medical University, Hefei, China

    S Yin & X Ding

  4. Department of Food and Nutrition Hygiene, School of Public Health, Anhui Medical University, Hefei, China

    Q Zhao

  5. Department of Oncology, The First Affiliated Hospital of Anhui Medical University, Hefei, China

    H Wang

Authors
  1. Y Mao
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. R Liu
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. H Zhou
    View author publications

    You can also search for this author in PubMed Google Scholar

  4. S Yin
    View author publications

    You can also search for this author in PubMed Google Scholar

  5. Q Zhao
    View author publications

    You can also search for this author in PubMed Google Scholar

  6. X Ding
    View author publications

    You can also search for this author in PubMed Google Scholar

  7. H Wang
    View author publications

    You can also search for this author in PubMed Google Scholar

Corresponding authors

Correspondence to X Ding or H Wang.

Ethics declarations

Competing interests

The authors declare no conflict of interest.

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Mao, Y., Liu, R., Zhou, H. et al. Transcriptome analysis of miRNA–lncRNA–mRNA interactions in the malignant transformation process of gastric cancer initiation. Cancer Gene Ther 24, 267–275 (2017). https://doi.org/10.1038/cgt.2017.14

Download citation

  • Received:

  • Revised:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1038/cgt.2017.14

Further reading

Search

Advanced search

Quick links