Background
Lung cancer is the leading cause of cancer-related deaths worldwide, despite advances in lung cancer therapy, the average 5-year survival rate is only 18% [
1,
2]. A majority of the deaths associated with lung cancer are due to secondary disease or metastatic progression [
3]. Lung adenocarcinoma (LUAD) is a major lung cancer that is in a locally advanced or metastatic stage at the time of diagnosis, which leaves no time for early detection or treatment [
4]. An early and accurate diagnosis may warrant timely treatment to potentially decrease the mortality. Therefore, molecular mechanisms that may help understand metastases of LUAD should be investigated to contribute to early diagnosis, better treatment and better overall prognosis of this disease.
Numerous published articles demonstrate that dysregulated genes are essential for initiation and progression of lung cancer. In recent years, the development of microarray technology has served as an effective measure to identify differentially expressed genes [
5], and provides new insight into the alteration of gene expression during tumorigenesis [
6]. Differentially expressed genes can be found through different experimental treatments, and their biological functions can be speculated via known information. The application of high-throughput miRNA profling methods, such as RNA sequencing and microarrays, has enabled researchers to identify a group of miRNAs as biomarkers in cancer diagnosis [
7]. MicroRNAs (miRNAs) are a class of single stranded, non-coding RNAs of 19–25 nucleotides [
8], which transcriptionally or post-transcriptionally regulate gene expression through binding to targeted mRNAs and influence the degradation and translation of mRNA [
9]. Accumulating evidence suggests that aberrant levels of miRNAs are linked to proliferation, angiogenesis, and metastasis in various human malignancies [
10]. Besides, miRNA has incurred broad attention as a targeting choice in cancer therapies [
11] or as the diagnostic or prognostic markers [
12]. Long non-coding RNAs (lncRNAs) are non-coding RNAs ranging in length from 200 nucleotides to ~ 100 kb [
13]. LncRNAs have mechanistically diverse functions in the cell, and in the nucleus, LncRNA has been shown to regulate gene expression in either cis or trans by recruiting a chromatin-modifying complex to the promoter of a target gene [
14,
15]. Besides, it has been reported that lncRNAs are key competing endogenous RNAs (ceRNAs) harboring miRNA response elements (MREs) and serve as ceRNAs to exchange with mRNAs via competitively binding to common miRNAs [
16].
The Cancer Genome Atlas (TCGA) is one prominent example of the renowned public databases which provides a platform of RNA sequencing with mRNA, miRNA and lncRNA data of various cancers. By integratively analysis RNA-Seq and miRNA-Seq data of LUAD samples from TCGA database, we successfully corhorted a set of differentially expression genes (DEGs), miRNAs (DEMs) and lncRNAs (DELs) underlying LUAD metastasis and non-metastasis samples. Based on intersection mRNAs, gene ontology (GO), Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment and co-expression analysis were conducted. Furthermore, we performed a receiver operating characteristic (ROC) analysis to investigate the diagnostic value of intersection mRNAs. Finally, miRNA-mRNA-lncRNA network were constructed. Our study might provide a meaningful contribution to exploring the mechanisms of LUAD metastasis and candidate diagnostic biomarkers and therapeutic targets.
Discussion
Even though there was a new development of immunotherapy and targeted therapy of tumors, once the systematic metastasis of cancer cell occurred, the 5-year survival rate would decrease at less by 10% [
27], and the survival of LUAD is far from satisfactory. Early diagnosis plays a critical role in the prevention and treatment of cancer, including LUAD. And researchers are continuously seeking for new biomarkers or targets for LUAD prevention, diagnose and treatment.
RNA-seq data and microarray-based expression profiling data provide a more comprehensive and accurate understanding of carcinogenesis and cancer progression at the molecular level. In our study, we analyzed the differences in the mRNAs, miRNAs, lncRNAs expression according to raw sequencing data of LUAD metastasis and non-metastasis samples from TCGA. We found that 1019 DEGs, 54 DEMs and 21 DELs are differently expressed in LUAD metastasis samples compared to non-metastasis samples. Combined with DEMs targeted mRNAs, we obtained 915 intersection mRNAs which were used to conduct GO and KEGG pathway analysis. According to the results of GO enrichment, the most significantly up-regulated genes were associated with signal transduction, while the most significantly down-regulated genes were mainly involved in mitotic cell cycle. KEGG pathway analysis showed that pathways corresponded to up-regulated transcripts were mainly related to cytokine–cytokine receptor interaction, pathways corresponded to down-regulated transcripts were mainly related to fanconi anemia pathway.
Generally, the common influence of interaction in interacting genes could not only decrease the complexity of biological network, but also benefit to explore meaningful biological information for the researchers, providing further scientific basis for therapy and study of disease [
28]. In our study, the co-expression gene network based on the intersection genes provided an insight of correlation between genes. Here, we found that some genes, such as SASH3, WAS and CD53, had connections with many other gene nodes in LUAD. They might be considered as important biomarkers which account for the metastasis mechanism of LUAD. In addition, survival analysis determined that 114 genes increased or decreased expression is significantly associated with LUAD patients shorter OS or RFS, suggesting that these genes may be valuable predictive factors for LUAD patient’s survival. Finally, 14 target genes were obtained by combining significant survival associated genes and were incorporated into the gene co-expression network.
In the target genes, five genes (ARHGAP30, IL16, PTPRC, SASH3 and TRAF3IP3) were with the degree over 20, which implied the five genes not only had better important regulatory value for the network, but also had diagnostic value of LUAD metastasis. ARHGAP30, a previously uncharacterized RhoGAP domain-containing protein, as a candidate Wrch-1-binding protein. ARHGAP30 is closely related to the Cdc42-specific RhoGAP CdGAP [
29], and together they form a subgroup of the RhoGAP proteins. Wang et al. [
30] identified ARHGAP30 served as a key regulator for p53 acetylation, and suggested ArhGAP30 as both prognostic marker and potential therapeutic target for colorectal cancer. Besides, ARHGAP30 was found had significantly improved OS of pancreatic ductal adenocarcinoma [
31]. Interleukin-16 (IL-16) is a pro-inflammatory cytokine [
32] and chemo attractant for a broad variety of immune cell types with CD4 co-receptors [
33]. Serum IL-16 levels have been associated with other cancers, such as multiple myeloma [
34], gastric cancer [
35], and colorectal cancer. Besides, IL-16 was reported to may act as a key mediator in pre metastatic niches that drives the establishment of lung metastasis and may represent a suitable therapeutic target [
36]. PTPRC, also known as CD45, encodes a member of the protein tyrosine phosphatase (PTP) family, which comprises proteins commonly activated in tumors [
37]. A previous report suggested that up-regulation of PTPRC resulted in high levels of inflammatory cytokines [
38]. It is reported that PTPRC was predicted to interact with CXCR4, and PTPRC might also play a role in colon cancer metastasis [
39]. SASH3 encodes a signaling adapter protein, containing a SLY motif in the N-terminal region, a SH3 motif and a SAM motif in the C-terminal region [
40]. SAM families of receptors are known to play a role in many developmental processes including cell migration, neuronal formation and angiogenesis [
40]. Schieffer et al. [
41] revealed SASH3 as a hub gene was highly correlated with diverticulitis patients compared to non-diverticulosis controls. TRAF3IP3, also known as TRAF3-interacting JNK Activating Modulator (T3JAM), was originally identified as a protein that interacts specifically with Tumor necrosis factor receptor-associated factor 3 (TRAF3) to activate JNK in human kidney cells [
42]. TRAF3IP3 is expressed in the immune system and participates in cell maturation, tissue development, and immune response. Nasarre et al. [
43] identified new functions of TRAF3IP3 in melanoma and angiogenesis, emphasizing its physiological relevance as a potential target for cancer therapy.
Moreover, our study identified that 2 lncRNAs (LOC96610 and ADAM6), 22 miRNAs and 4 mRNAs (LAX1, DERL3, MEI1 and CPNE5) were involved in the miRNA-lncRNA-mRNA interaction network. LOC96610, located at 22q11.22, the official symbol is BMS1P20 (BMS1, ribosome biogenesis factor pseudogene 20). As a survival-related lncRNA, BMS1P20 was found significantly correlated with the pathogenesis, development and metastasis of liver hepatocellular carcinoma [
44]. Furthermore, Sui et al. [
45] reported that BMS1P20 positively correlated with overall survival of LUAD. ADAM6 (ADAM metallopeptidase domain 6), located at 14q32.33. ADAM is a family of membrane proteins involved in cell–cell adhesion and cell–matrix adhesion. It is characterized by a disintegrin and metalloprotease domain with an epidermal growth factor-like region and harbors both adhesion and proteolytic domains implicated in integrin function and matrix degradation [
46,
47]. It is reported that an exploratory biomarker panel derived from ADAM6 conferred prognostic utility for melanoma recurrence and death [
48].
Conclusions
In summary, our findings documented that 1019 DEGs, 54 DEMs and 21 DELs were differently expressed in LUAD metastatic samples compared with non-metastatic samples. Among these altered mRNAs, many are significantly associated with LUAD patient’s survival time, and might play critical roles in LUAD metastasis. Our study highlights the important roles of mRNAs, miRNAs and lncRNAs in LUAD metastasis and may provide useful candidates as diagnostic markers and potential targets for LUAD therapy. The present study also has a few limitations, for example, the data used were obtained from TCGA, rather than directly from LUAD patients, thus a series of verification experiments must be performed to confirm our results. Overall, our findings will improve our understanding of the molecular mechanisms of LUAD and aid in finding potential targets for diagnostic and therapeutic usage.
Authors’ contributions
ZJ and QDC designed and edited this study; LLF, PML and XWH searched the databases and collected full-text papers; FZR, WT and LJY extracted and analysed the data; ZYK and LWP wrote the manuscript. All authors read and approved the final manuscript.