GC is a gastroenterological malignancy with high rates of prevalence and mortality[1–3, 11]. Therefore, sensitive and specific biomarkers of GC are urgently needed to be detected. In the present study, bioinformatic methods are promising methods to analyze the critical genes and pathways, which might provide novel clues for diagnosis, therapy, and prognosis of GC. We integrated seven gene expression profile datasets from different groups and used R software and bioinformatics to deeply analyze these datasets. DEGs PPI network was successfully constructed via the STRING online database and Cytoscape software. Seven vital regulated genes including HMMR, SPP1, FN1, CCNB1, CXCL8, MAD2L1, and CCNA2 were screened from the PPI network complex by cytoHubba plug-in of cytoscape.
Through Kaplan Meier plotter analysis, we found that most of the selected genes were associated with a significantly worse survival, except SPP1. The expression of the genes was higher in GC samples than normal samples by GEPIA analysis. Importantly, using qRT-PCR, we could validate the higher mRNA expression of the selected genes based our bioinformatics analysis; most selected genes, except SPP1 and FN1, were upregulated in tumor tissue. They showed the same trend in expression as predicted by bioinformatics verifying the accuracy of our method. In the light of important roles in cells, the selected hub genes in GC (HMMR, CCNB1, CXCL8, MAD2L1, and CCNA2) may represent potential prognostic biomarkers and/or therapeutic targets for GC.
For a more in-depth understanding of these DEGs, we analyzed the selected genes for GO and KEGG enrichment analyses and found that ‘Cell cycle’ signaling pathways was significant enriched. HMMR, CCNB1, MAD2L1 and CCNA2 play important roles in cell cycle. HMMR, a cell surface hyaluronan receptor and mitotic spindle protein and the driver of tumor progression [12] [13] [14], plays an important role in the modulation of motor activities and the maintenance of genome stability[15, 16]. High expression of HMMR significantly correlates with tumor relapse [17, 18] and predicts poorer prognosis in GC patients. Furthermore, HMMR has been identified as a promising target for antibody therapy to block the extracellular function of HMMR on the surface of tumor cells[19], which might be a potential prognostic marker or therapeutic target against the disease. The protein encoded by CCNB1 gene is an important monitoring protein in mitosis, which is necessary for proper controlling the cell cycle at the G2/M transition phase[20]. Previous studies have reported that the CCNB1–Cdk1 complex is a key regulator of mitotic entry[21]. Recently, increasing evidence demonstrated that CCNB1 was over-expressed in considerable cancers with poor prognosis, including hepatocellular carcinoma[22, 23], breast Cancer [24, 25], and pancreatic cancer [26, 27].The expression of CCNB1 is often used to estimate prognosis after treatment with anticancer drugs[22, 28].
The function of MAD2L1 is to maintain the separation state of chromosomes during the dissociation of mitotic chromosomes and spindle, and to play a role in the checkpoint during mitosis [29, 30]. Abnormal regulation of MAD2L1 is associated with chromosomal instability and a large number of aneuploidy, which can lead to tumorigenesis [31]. Studies have found that MAD2L1 is overexpressed in lung adenocarcinoma tissues, and the overexpression of MAD2L1 may indicate poor prognosis and increased risk of tumor recurrence in patients, which can be used as a prognostic marker for lung adenocarcinoma[29]. Our bioinformatics analysis showed that MAD2L1 was highly expressed in tumor tissues compared with normal tissues. MAD2L1 is a pro-oncogene which is upregulated in GC [32, 33], and we need to further study its specific mechanism. The protein encoded by CCNA2 belongs to the highly conserved cyclin family, whose members function as regulators of the cell cycle at the G1/S and the G2/M transitions[34]. CCNA2 is overexpressed in several human cancers and closely related to tumor progression and shorter survival in lung, breast, and colorectal cancer [35–41]. CCNA2 overexpression was an indicator of poor prognosis. Limited by few studies about evaluating the expression and prognostic role of CCNA2 in GC patients, more efforts are necessary to confirm expression pattern and prognostic role of CCNA2 in GC patients.
CXCL8 is a member of the CXC chemokine family that acts as an important multifunctional cytokine to modulate tumor proliferation, invasion and migration in an autocrine or paracrine manner. Neovascularization, which provides a basis for fostering tumor growth and metastasis, is now recognized as a critical function of CXCL8 in the tumor microenvironment[42]. CXCL8 signaling axis also plays an indispensable role in colorectal carcinoma[43, 44], renal cell carcinoma [45], pancreatic cancer[46], thyroid tumors [47, 48], gastric cancer [49, 50], and lymphomas [51]. Interruption of the related signaling pathways may thus provide promising therapeutic avenues for tumors. Studies have found that CXCL8 is predominantly secreted by macrophages and contributes to the immunosuppressive microenvironment by inducing PD-L1 + macrophages in GC[52]. CXCL8 inhibitors may drive antitumor response, providing potential therapeutic effects for patients with gastric cancer.
To further screen the TFs in hub genes, we constructed a gene-TF regulatory network and found IRF1, ETV4, KLFs, and SMAD5 that were meaningful in GC. It was reported that MTMR2 mediated epithelial-mesenchymal transition through the IFNγ/STAT1/IRF1 pathway to promote GC invasion and metastasis [53]. KIF2A expression is a potential target for GC therapy, which can be upregulated by transcription factor ETV4 [54]. Krüppel-like factors (KLFs) have been extensively investigated in multi-cancers[55], which plays a significant role in GC progression and could be a new therapeutic target for GC patients. Interestingly, SMAD5 was frequently altered in human GC[56]. The intricate interaction between TFs and other hub genes made great contribution to the development of cancer.
Studied have proved that Toll-like receptor (TLR) signaling pathways play important roles in development of GC. TLR signaling pathways are involved in innate and adaptive immunity responses[57] and activation of both inflammatory and carcinogenic processes[58]. Thus, the pattern of the host’s immune response beyond genetic and environmental factors is also essential for understanding the pathology of GC[59]. TLRs, a class of transmembrane receptors[60], play an important role in defense against Helicobacter pylori (H. pylori) widely known as a class I carcinogen in GC[60, 61]. Therefore, the abnormal expression of TLRs is closely related to tumorigenesis and cancer progression and a better understanding of TLRs will provide new diagnostic or predictive markers for the diagnosis of GC.
We failed to validate SPP1 as a DEG in our fresh GC samples, which may be as a result of the small sample size and inter-sample variation. The protein encoded by SPP1 helps osteoclasts to bind to the mineralized bone matrix, which plays an important role in tumorigenesis, invasion and metastasis [10, 62]. Overexpressed SPP1 expression had been confirmed in various types of cancers[63–67]. A study based on gastric cancer cell lines indicated that the elevated expression of SPP1 is a critical determinant of poor prognosis[68]. In addition, in a recent study, SPP1 rs4754 polymorphism was observed to be associated with the risk of gastric cancer and has an important effect in gastric carcinogenesis[69].
All above, we found that high expression of 5 validated hub genes should promote the progress of GC patients, suggesting that their antagonism may improve the prognosis of GC. Although some of these genes were found before, our study could validate and explain the expression status of these genes and their impact on prognosis in GC again. These findings provide a set of useful driving genes and key pathways of cancers, which are worth future investigating for novel therapeutic targets, a prognostic evaluation index, and the detailed pathogenesis of them in GCs.
However, there were several limitations of the present study. Firstly, validation with qRT-PCR study need more tumor and adjacent normal tissues samples. Second, more experiments, such as immunohistochemistry and Western blot, should be conducted to confirm the protein levels in GC.