Gastric cancer is one of the most common malignant tumors around the world and it has high incidence and mortality. It is estimated that the mortality of GC reaches 70% and is obviously higher than other epidemic diseases (28). With the improvement of standards of hygiene and the eradication of H. pylori, the incidence rate of GC has shown a decline trend and changed from the fourth most common cancer to the fifth (29, 30). However, the mortality rate of GC has not been significantly decreased. Due to the application of radical surgery and chemotherapy, the five-year survival rate of early GC can increase up to 90%, but the prognosis of advanced GC is very poor (31). It’s because that advanced GC missed the best therapeutic time window and has the potential for distant metastasis (32). Meanwhile, most patients with early GC do not display any symptom is also an important cause and it means that GC patients are difficult to detect at an early stage (33). Although surgical techniques are improving, novel chemotherapy and targeted therapy are developing, there is still no effective treatment for advanced gastric cancer (10). Therefore, the identification of the effective biomarkers for GC and further exploration of their correlative biological functions could help us to detect GC at early stage and understand the pathogenesis of GC, which in turns lead to the emergence of novel targeted therapy.
In this study, three gene expression profiles of GC (GSE27342, GSE63089, GSE33335) were utilized to select DEGs between GC tissues and noncancerous tissues. A total of 35 genes were identified, including 11 up-regulated genes and 24 down-regulated genes. GO and KEGG pathway enrichment analysis were executed to investigate the biological functions of DEGs. The results of GO functional enrichment analysis indicated that 35 genes were mainly enriched in the process of digestion, regulation of biological quality, response to hormone and steroid hormone, and homeostatic process. GC itself is a malignant tumor in the digestive system and most of risk factors are associated with digestion and diet. There are many studies have revealed that obesity acts as a risk factor of GC and could affect digestion (8, 34). While fat has metabolic capacity and could secret leptin and insulin-like growth factor (IGF), the change of leptin and IGF has been reported to be associated with gastrointestinal cancers (35). Moreover, Zhou et al. has proven that the abnormal expression of gastrointestinal hormones like hypergastrinemia have correlation with GC (36). Some previous studies have suggested that the production of steroid hormone could affect the development of GC (37, 38). Estrogen receptor (ER) acts as steroid hormone receptor and the dysregulation of ERs is associated with GC. ERα36 is a kind of ER, and it could increase the expression of cyclin D1 to promote the proliferation of GC cells (39). In addition, the role of estrogen in GC is related to its concentration. Low concentration of estrogen can increase the expression of ERα36 and promote GC cells growth, but high concentration of estrogen is on the contrary (40). Homeostatic process has also been demonstrated that is associated with GC (41). Autophagy is a vital homeostatic process and it plays both tumor-suppressor role and tumor-promoter functions in GC (42). One cause of GC development is autophagy could suppress GC cells at early stages and promote the growth of existing tumor (43).
The findings from KEGG pathway enrichment analysis showed DEGs were significantly enriched in the secretion of gastric acid and collecting duct acid, leukocyte transendothelial migration, ECM-receptor interaction and pancreatic secretion. It has been reported that gastric acid secretion is associated with GC and chronic atrophic gastritis which is regarded as a risk factor of GC (44, 45). The migration characteristics of leukocytes are essential to promote the systemic immune response (46). Enarsson et al. found that the transendothelial migration of T cells in GC patients was reduced and speculated that reduced transendothelial migration of T cells might help GC cells to evade immune responses and was beneficial for carcinogenesis (47). ECM, its full name is extracellular matrix, and the abnormal changes of ECM contents are associated with the formation of cancer (48). The component of ECM like hyaluronan could play roles in promoting GC development (49). The relevant studies of the above GO and KEGG analysis have been confirmed to be associated with GC, which is also consistent with our results.
According to the PPI network and a total of 10 hub genes are obtained from it. Among these genes, COL1A1, BGN, SPP1 and THY1 were up-regulated and ATP4A, TFF2, GIF, SST, GHRL and ATP4B were down-regulated. We further explore these genes and analyze their possible biological functions. The result showed that ATP4A and ATP4B had correlation with most of enriched pathway. ECM-receptor interaction and focal adhesion were associated with COL1A1 and BGN, which two genes have the highest score in MCC method. Subsequently, we validated the expression level of these hub genes by using GEPIA website. The result as shown in Figure 6, COL1A1, BGN, SPP1, THY1 in GC tissues had an increased expression. While ATP4A, TFF2, GIF, SST, GHRL and ATP4B had a decreased expression level. The verification of expression level in the present study supports the screened hub gene and makes this study more credible. Finally, the Kaplan–Meier survival analysis indicated that up-regulated COL1A1, BGN and THY1, down-regulated TFF2 and SST were associated with a poor OS. It provides a possible research trend of GC, which is worth further exploring.
COL1A1, the full name is collagen type I α 1, and its protein is collagen type I which is a major component of ECM (50). There are many studies have been reported that the overexpression of COL1A1 is related to a variety of cancers, such as colorectal cancer, breast cancer and GC (51-53). Liu et al. found that the abnormal expression of COL1A1 in breast cancer cells promoted cancer metastasis and knockout of COL1A1 could inhibit this process (54). In colorectal cancer patients, overexpressed COL1A1 affected metastasis by regulating the WNT/PCP signaling pathway (51). As a component of ECM, high expression of COL1A1 could increase the stiffness of ECM and contribute to cancer development (55). Furthermore, miR-129-5p has been proved to inhibit the proliferation of GC and distant metastasis by suppressing the expression of COL1A1 (52). COL1A1 has also been recognized as a potential biomarker of GC for early detection at many times (56, 57).
BGN is also the component of ECM, and the upregulation of BGN has been reported as a biomarker for multiple cancers (58). A recent study demonstrated that BGN has a high expression in GC patients and BGN could interact with Toll like receptor 2 and 4 secreted by GC cells to activate the VEGF pathway and promote angiogenesis, which is beneficial to the progression of cancer (59). Therefore, BGN could be regarded as a biomarker of GC that is correlated with poor prognosis. THY-1 acts as a cell surface protein with immunoglobulin domain and it exhibits a high level in GC patients (60, 61). There are some studies has revealed that the formation of GC was associated with THY1 (62). The activation of Notch signaling pathway can promote cancer growth and THY1 stimulates Notch signaling pathway through the regulation of RhoA (63, 64). Furthermore, THY1 has been proven to be a target gene of miR-140-5p (65). Wu et al. has indicated that miR-140-5p can inhibit the development of gastric cancer by inhibiting the expression of THY1 and inactivating Notch signaling pathway (65). According to the above studies, THY1 can be used as a target for the treatment of GC, and it is worth further exploring its value in early detection.
In addition, TFF2 and SST have been reported as tumor suppressors, and low level of TFF2 and SST were found in GC patients (66-68). TFF2 is a short peptide that secreted by gastric mucous neck cells and is helpful for cellular protection of gastrointestinal tract (69). However, the exact mechanism between this protective effect and GC has not been elucidated. Recent studies have indicated that TFF2 could inhibit the development of GC and tumor invasion by interacting with Sp3, therefore down-regulated TFF2 might loss its anti-tumor effect (70). SST, an inhibitory gut peptide, it plays role in regulating cell proliferation and its analog octreotide has been used to inhibit the GC metastasis (71, 72). In GC patients, SST has a decreased expression and its promoter DNA methylation level is higher than normal patients (73). Low expression of SST promotes GC growth, while the inhibition of GC invasion and metastasis by overexpression of SST has been confirmed in vivo experiments (67). Accordingly, we speculate that promoter DNA methylation might regulate the expression of SST and further promotes GC growth through mechanisms that are not yet clear. Although the mechanism of TFF2 and STT in GC has not been elucidated, they are still worthy of being used as biomarkers for early diagnosis of GC.