Comprehensive Analysis of the Prognostic Value of ITGB Superfamily Members in Gastric Cancer Using Bioinformatics Analysis


 Background: Gastric cancer (GC) is one of the most common malignant tumors worldwide with a high incidence and mortality, making it urgent to search for potential biomarkers to improve the prognosis of patients. Integrin β (ITGB) superfamily members have been reported to participate in the biological process in various tumors. However, the expression and prognostic value of ITGBs in GC have not been clarified. In this study, the expression and roles of ITGBs in stomach adenocarcinoma (STAD) were investigated with ONCOMINE, UALCAN, The Human Protein Atlas, GEPIA, Kaplan-Meier Plotter, NetworkAnalyst, cBioPortal, STRING, GeneMANIA, David 6.8 and TIMER. Results: The transcriptional levels of ITGB1, 2, 4, 5, 6 and 8, and the protein expression of ITGB4, 5, 6 and 8 were significantly elevated while the transcriptional level of ITGB7 was significantly reduced in GC. High mRNA expression of ITGB1, 3 and 5, and low mRNA expression of ITGB4 were associated with short overall survival (OS), and high mRNA expression of ITGB6 was associated with short disease-free survival (DFS) in STAD patients. We also identified DAB2 as the potential transcription factor for ITGB3 and ITGB5, and TFAP2A for ITGB4 in STAD. Moreover, the functions of ITGB superfamily members were primarily related to regulation of actin cytoskeleton, focal adhesion and PI3K-Akt signalling pathway. The expression of ITGBs was significantly correlated with the infiltration of diverse immune cells, B cell, CD8+ T cell, CD4+ T cell, macrophage, neutrophil and dendritic cell.Conclusion: Taken together, our study indicated that ITGB superfamily members could be a potential prognostic target in the development of anti-STAD therapeutics. However, further experimental studies are still necessary to elucidate the mechanism of ITGBs.


Introduction
Gastric cancer (GC) is the fth most common malignancy and it is the third leading cause of cancerrelated death, with over 1 million estimated new cases and approximately 800000 deaths in 2018 [1]. GC is a heterogeneous group of tumors with diverse morphologies, molecular models and histogenesis, of which stomach adenocarcinoma (STAD) accounts for 95% [2]. Since GC still constitutes a global health problem, preventions such as proper diet, early diagnosis and follow-up proper treatments lead to reduction of incidence [3,4]. However, it remains the fact that the frequent presence of advanced stage of gastric cancer at rst diagnosis, which often narrows the options for the surgical treatment, leading to a worse outcome of patients [5]. Therefore, there is a urgent need to discover the prognostic candidates for gastric cancer's diagnosis and treatment.
Integrin-β (ITGB) superfamily , a member of integrin, is made up of eight members in human body, ITGB1-8 [6]. ITGB superfamily members have been reported to interact with extracellular matrix and activate signal transduction cascades, regulating various cellular processes including proliferation, carcinogenesis and immune responses [7][8][9][10]. Zhuang et al reported that ITGB superfamily is closely associated with the prognosis of patients with pancreatic cancer [11]. The correlations between the ITGB superfamily members and the prognosis of ovarian cancer were also explored [12]. However, the picture of prognostic and oncologic characteristics of the ITGB superfamily members remains poorly explored in STAD.
In this study, a comprehensive bioinformatics analysis of the mRNA expression of ITGB superfamily members and their potential as biomarkers such as tumor progression, prognosis, gene alterations and immune in ltration in STAD was explored to provide additional data for clinicians to select the appropriate therapeutic schedules, which might lead to a better outcome for patients.

Materials And Methods
Differentially expressed ITGBs at the transcriptional and protein level ONCOMINE (www.oncomine.org) is an integrated online oncogene database and data-mining platform, providing powerful, genemo-wide expression analysis [13]. In this study, a p-value < 0.05, a fold change of 2 and a gene rank in the top 10% were set as the signi cant thresholds to explore the aberrant transcriptional levels of ITGBs in different cancer types. The difference in the expression of ITGBs in STAD was analyzed by student's t test.
As a comprehensive, user-friendly and interactive online database, UALCAN (http://ualcan.path.uab.edu/) helps users gather valuable information and data [14]. The expression of ITGBs was analyzed in the "TGCA Gene analysis" module and the "Stomach adenocarcinoma" dataset, in which a statistically signi cant difference was considered when the p value was < 0.05.
The HPA database (https://www.proteinatlas.org/) aims to map all the human proteins in cells, tissues and organs, allowing users to conduct the exploration of human proteome [15]. In this study, immunohistochemistry images of protein expression of ITGBs in STAD and normal tissues were directly visualized by HPA.
Associations between the expression of ITGBs and outcomes of STAD patients GEPIA (http://gepia.cancer-pku.cn/) is a web server based on TCGA and GTEx data, providing functions such as differential expression analysis, pathological stage analysis, correlation analysis and survival analysis [16]. The pathological stage analysis and survival analysis of ITGBs in STAD, Student's t test was used to generate a p value and the survival analysis was performed using a Kaplan-Meier curve.
Kaplan-Meier Plotter (http://kmplot.com/analysis/) allows users to evaluate the effect of mRNA, miRNA or protein on the survival in breast, ovarian, lung and gastric cancer [17]. The prognostic value of ITGBs in patients with STAD was explored and exhibited together with the HR index, 95% CI, p value and number at risk in the gures. The p value cutoff was 0.05.
Predicted potential transcription factors (TFs) for prognosis-associated ITGBs in STAD NetworkAnalyst (https://www.networkanalyst.ca/) is a comprehensive network visual analytics platform specialized in transcriptome pro ling and network analysis for gene expression data [18].
The predicted potential TFs of prognosis-associated ITGBs were explored and visualized as a network picture. TFs regulating two or more prognosis-associated ITGBs were considered as key TFs, and selected to conduct co-expression analysis with their corresponding ITGBs with GEPIA. The differential expression analysis of predicted TFs of tumor and normal tissues were also performed with GEPIA. The threshold of Pearson correlated coe cient > |0.4|, p value < 0.05. The OS survival analysis of TFs was shown as Kaplan-Meier curves.

Gene alterations and protein-protein interaction (PPI) network of ITGBs
The genetic alteration of ITGBs in STAD was obtained from cBioPortal (http://www.cbioportal.org/) based on TCGA database, which is an open-access resource for the exploration of multidimensional cancer genomics data sets, allowing users to study the molecular pro les and clinical attributes of cancers [19]. STRING (https://www.string-db.org/) provides physical and functional associated protein-protein interactions, covering more than 24 million and 5 thousand organisms [20]. The PPI network of differentially expressed ITGBs was carried out and 20 relative genes were obtained for further analysis. GeneMANIA (http://genemania.org/) provides quick and convenient access to predicting the function of user-interested genes and gene sets [21]. The functions of differentially expressed ITGBs and associated molecules were explored with GeneMANIA.

Function enrichment analysis of ITGBs
The Gene Ontology (GO) enrichment and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis of ITGB superfamily members and their 20 relative genes were conducted through David 6.8 (https://david.ncifcrf.gov/), a comprehensive, functional annotation online tool helping users understand biological meaning of submitted genes [22]. The top 10 items of each module were visualized as bubble diagrams with Hiplot (https://hiplot.com.cn/), a scienti c data analysis and visualization tool.
GO enrichment comprises biological process (BP), cellular components (CC) and molecular function (MF).
Immune in ltration analysis of ITGBs TIMER (https://cistrome.shinyapps.io/timer/) is comprehensive web tool, providing systematical analysis of immune in ltration of diverse cancer types [23]. The correlations of ITGB superfamily members expression with immune in ltration levels in STAD was performed in "Gene" module, and exhibited as the scatterplot.

Results
Aberrant mRNA and protein expressions of ITGB superfamily members in patients with STAD ONCOMINE database was utilized in order to research the mRNA expression of ITGB superfamily members in patients with STAD. Compared with normal tissues, the mRNA expressions of ITGB1, ITGB2, ITGB4, ITGB5 and ITGB8 were signi cantly elevated, while the mRNA expression of ITGB7 was signi cantly reduced in STAD tissues (FIGURE 1). Data in TABLE 1 was consistent with the results above.
For instance, in the Chen dataset the transcriptional levels of ITGB1 were signi cant higher in gastric intestinal type adenocarcinoma and gastric mixed adenocarcinoma, with the fold changes of 2.130 and 3.145, respectively [24]. In the DErrico dataset, the transcriptional level of ITGB2 was signi cantly upregulated in diffused gastric adenocarcinoma with the fold change of 2.318 [25]. In the Cho dataset, the transcriptional levels of ITGB7 were signi cantly reduced in gastric mixed adenocarcinoma, diffused gastric adenocarcinoma and gastric adenocarcinoma with the fold changes of -5.811, -5.450 and -4.872, respectively [26]. Next, the mRNA expression patterns of ITGB superfamily members were measured by the UALCAN database. The results showed that the mRNA expressions of ITGB1 (p = 1.78e-04), ITGB2 (p = 1.06E-10), ITGB4 (p = 1.63E-12), ITGB5 (p = 1.67E-09), ITGB6 (p = 6.17E-08) and ITGB8 (p = 1.64E-12) were signi cantly elevated in STAD vs normal tissues (FIGURE 2). These results were almost consistent with those from ONCOMINE.
After analyzing the mRNA expression of ITGB superfamily members in STAD, the protein expression of ITGB superfamily members were further explored with the Human Protein Atlas. We found that the protein levels of ITGB4, ITGB5, ITGB6 and ITGB8 were signi cantly up-regulated in STAD tissues (FIGURE 3).
Finally, the correlation between the expressions of ITGB superfamily members and the pathological stage of STAD patients were explored with GEPIA database. As shown in FIGURE 4, there was a signi cant correlation between the pathological stages and the expression of ITGB2 (p = 0.0143) and ITGB7 (p = 0.0375). These data suggested that ITGB2 and ITGB7 might play a signi cant role in the tumorigenesis and progression of STAD.

Prognostic value of ITGB superfamily members in patients with STAD
The prognostic value of ITGB superfamily members in STAD patients including overall survival (OS) and disease-free survival (DFS) was investigated with GEPIA. The curves in FIGURE 5A suggested that patients with low transcriptional levels of ITGB1 (p = 0.029) and ITGB3 (p = 0.039) were signi cantly associated with longer OS. And high transcriptional level of ITGB4 (p = 0.038) was associated with longer OS. The results in FIGURE 5B showed that patients with low transcriptional level of ITGB6 (p = 0.011) were associated with longer DFS.
The Kaplan-Meier Plotter database was also utilized to analyze the correlation between the mRNA expressions of the ITGB superfamily members and the OS of patients with STAD. As shown in the Therefore, out study implied that DAB2 might synchronously promote the transcription of ITGB3 and ITGB5, and TFAP2A might promote the expression of ITGB4.
Genetic alteration, expression and PPI network of ITGB superfamily members in patients with STAD The genetic alterations of ITGB superfamily members in patients with STAD were analyzed by cBioPortal online tool. As a result, ITGB1, ITGB2, ITGB3, ITGB4, ITGB5, ITGB6, ITGB7 and ITGB8 were altered in 9, 8, 7, 11, 8, 7, 8 and 11% of the queried STAD samples, respectively (FIGURE 8A). Low mRNA expression was the most common alteration in these samples (FIGURE 8B).
Then the PPI network analysis of ITGB superfamily members was conducted to explore the potential interactions among them by STRING. There were 13 nodes and 68 edges in the PPI network (FIGURE 8C). The ITGB superfamily members were associated with pathways such as regulation of actin cytoskeleton and arrhythmogenic right ventricular cardiomyopathy. The GeneMANIA was also utilized to reveal the functions of ITGB superfamily members and their relative molecules (such as WIF1, ATRNL1, EGFL7, TNC, TNR and TENM4), primarily associated with extracellular matrix organization, extracellular structure organization, receptor complex and integrin complex (FIGURE 8D).

Functional enrichment analysis of ITGB superfamily members in patients with STAD
Firstly, the 20 genes related to ITGB superfamily members were obtained from STRING. Then the functional enrichment analysis of these 28 genes was conducted by DAVID 6.8 and visualized by Hiplot tool. From the FIGURE 9A, the top 10 most highly enriched GO items in the BP category were integrinmediated signaling pathway, extracellular matrix organization, leukocyte migration, cell-matrix adhesion, cell adhesion, heterotypic cell-cell adhesion, cell adhesion mediated by integrin, viral entry into host cell, leukocyte cell-cell adhesion and endodermal cell differentiation, which were signi cantly regulated by ITGBs in STAD. The integrin complex, cell surface, plasma membrane, focal adhesion, receptor complex, extracellular exosome, external side of plasma membrane, ru e membrane, cell-cell adherens junction and hemidesmosome were the top 10 most highly enriched items in the CC category (FIGURE 9B). In the MF category, the ITGB superfamily members and their relative genes were mainly enriched in integrin binding, receptor activity and virus receptor activity (FIURE 9C).
As shown in FIGURE 9D, the top 10 KEGG pathways affected by ITGBs in STAD were regulation of actin cytoskeleton, focal adhesion, hypertrophic cardiomyopathy, dilated cardiomyopathy, ECM-receptor interaction, arrhythmogenic right ventricular cardiomyopathy, PI3K-Akt signaling pathway, proteoglycans in cancer, cell adhesion molecules and pathways in cancer.

Immune cell in ltration of ITGB superfamily members in patients with STAD
The proliferation and progression of cancer cell is associated with immune cell levels. Therefore, the associations between ITGB superfamily members and immune cell in ltration were explored by TIMER database (FIGURE 10). The expression of ITGB1 was positively associated with the in ltration of CD4 + T cell, macrophage and dendritic cell. The expressions of ITGB2 and ITGB3 were both signi cantly positively related to the in ltration levels of all six immune cells, B cell, CD8 + T cell, CD4 + T cell, macrophage, neutrophil and dendritic cell. There was a remarkably negative correlation between the expression of ITGB4 and the in ltration of CD8 + T cell, CD4 + T cell and dendritic cell. ITGB5 expression was positively associated with the in ltration of CD4 + T cell, macrophage and dendritic cell, and negatively associated with the in ltration of B cell. We also found the higher the in ltration levels of B cell, the higher the expression of ITGB6. The expression of ITGB7 was signi cantly positively related to the in ltration of CD8 + T cell, CD4 + T cell, macrophage, neutrophil and dendritic cell. ITGB8 expression was positively associated with the in ltration of B cell.

Discussion
Gastric cancer displays marked molecular heterogeneity with aggressive behavior and treatment resistance, exhibiting a high mortality rate with an average 5-year survival rate of 20% in most areas of the world [27,28]. The exploration of potential prognostic biomarkers is necessary for improving the bad outcomes for patients with STAD.
Integrins-β (TIGB) superfamily, made up of ITGB1, ITGB2, ITGB3, ITGB4, ITGB5, ITGB6, ITGB7 and ITGB8, is a member of integrin, which is a kind of cell cohesion protein, playing an important role in cell migration, hemostasis and immune response [29]. Besides, integrin takes part in mediating the initiation, progression and metastasis of tumor [30,31]. ITGB1 overexpression can promote the progression of prostate cancer through upregulating the expression of caveolin-1 [32]. ITGB2 enhances the glycolysis activity through PI3K/AKT/mTOR pathways, which plays a signi cant role in cancer associated broblasts in the promotion of oral squamous cell carcinoma [33]. It has been reported that ITGB3dependent uptake of extracellular vesicles, for which focal adhesion kinase is required as a key role, could promote the metastasis of breast cancer cells [34]. ITGB4 and ITGB6 targeted immunotherapy is a promising strategy, showing a great effect on the inhibition of the progression and metastasis of tumors [35,36]. Studies have also suggested that ITGB7 and ITGB8 have the potential to mediate the progression of tumors, paving the way for a novel therapeutic approach for the treatment for cancers [37][38][39].
In this study, the mRNA expression of ITGB superfamily members in STAD was rstly explored. The transcriptional levels of ITGB1, ITGB2, ITGB4, ITGB5, ITGB6 and ITGB8 were signi cantly up-regulated and the transcriptional level of ITGB7 was signi cantly down-regulated in the STAD tissues compared with normal tissues. Besides, through analyzing the protein expression of ITGB superfamily members in STAD by HPA, we found that the protein expressions of ITGB4, ITGB5, ITGB6 and ITGB8 were higher in STAD tissues than normal tissues. Next, results suggested that the expression of ITGB2 and ITGB7 increased as the tumors progressed. STAD patients with low expression of ITGB1, ITGB3 and ITGB5 and high expression of ITGB4 were signi cantly associated with better overall survival. And the low expression of ITGB6 in STAD patients was related to better disease-free survival. These data demonstrate that the differentially expressed ITGB superfamily members may play a signi cant role in the tumorigenesis, progression and prognosis of STAD. TF functions in the transcriptional level as key regulators in the gene regulation [40]. Our study suggests that ITGB1, ITGB3, ITGB4, ITGB5 and ITGB6 may serve as potential prognostic biomarkers for STAD. Therefore, the predicted TFs for the prognosis-associated ITGBs were explored based on the Network Analysis database. Results identi ed DAB2 as the potential TFs for ITGB3 and ITGB5, and TFAP2A as the potential TF for ITGB4. Both of DAB2 and TFAP2A were obviously up-regulated in STAD. The overexpression of DAB2 was signi cantly associated with shorter OS, consistent with the results that overexpression of ITGB3 and ITGB5 was related to shorter OS. And the same conclusion can be found that overexpression of TFAP2A and ITGB4 was signi cantly associated with longer OS. Therefore, we supposed that DAB2 might enhance the expression of ITGB3 and ITGB5, and TFAP2A might enhance the expression of ITGB4. Our study may provide deeper insight into the potential mechanisms of ITGB3, ITGB4 and ITGB6 in gastric adenocarcinoma.
As well as the differential expression of genes in STAD, gene mutations and epigenetic alterations play a non-negligible role in the progression of tumors [41]. There were frequent genetic alterations in the ITGB superfamily members in STAD, where the low mRNA expression was the most alteration. The gene alteration analysis may also help to understand the role of ITGBs in STAD.
Then the function of ITGB superfamily members was assessed using GO enrichment analysis and KEGG pathway enrichment analysis. As expected, the functions of ITGBs and relative 20 genes are mainly related to regulation of actin cytoskeleton, focal adhesion and PI3K-Akt signaling pathway. Actin cytoskeleton is formed by polymerization of globular actin monomers to form micro laments, and participated in several essential cellular processes including endocytosis, intracellular transport, cell morphogenesis and cell motility [42,43]. Moreover, the regulation of actin cytoskeleton plays a signi cant role in biological processes associated with the invasion and metastasis of tumors [44,45]. Focal adhesion is reported to promote the progression of cancers when activated by follistatin-like protein [46]. The key role of PI3K-Akt signaling pathway in the tumorigenesis, proliferation and progression in cancers has also been proven [47,48]. These data indicate that ITGB superfamily members could be potential drug therapeutic targets.
Integrins could mediate the migration and location of immune cells [49,50]. More and more evidence suggested that immune cell in ltration is the key role affecting the progression and recurrence of the tumor [51]. In our study, the expression of ITGB superfamily members was signi cantly associated with the in ltration of six immune cells, B cells, CD8 + T cells, CD4 + T cells, macrophages, neutrophils and dendritic cells, indicating that ITGBs could affect the prognosis of tumors through mediating the immune status.
There were some limitations in our study. The analysis involved in the study were all based on different online databases. And further studies consisting of cell experiments and clinical studies are required to validate the results and explore the potential mechanisms underlying the roles of ITGB superfamily members in STAD.

Conclusion
In conclusion, our study indicated that differentially expressed ITGBs could be potential prognostic biomarkers for the outcomes of patients with STAD, might providing novel insights for the design of new anti-STAD therapeutics.  The expression of ITGB superfamily members in STAD (UALCAN). The expression of ITGB1, 2, 4, 5, 6 and 8 was signi cantly elevated in STAD tissues compared with normal tissues (p < 0.05).

Figure 3
The protein expression of ITGB superfamily members in STAD vs normal tissues (HPA). The protein expression of ITGB4, 5, 6 and 8 was higher in STAD than normal tissues.

Figure 4
The correlations between the pathological stage of STAD and the expression of ITGB2 and ITGB7 (GEPIA) (p < 0.05).     The GO and KEGG pathway enrichment analysis of ITGB superfamily members and their 20 relative genes (David 6.8). (A-C) Bubble plots of GO enrichment in BP terms, CC terms and MF terms; (D) Bubble plot of KEGG pathway enrichment terms.

Figure 10
Correlations between ITGB superfamily members and the in ltration of immune cells in STAD (TIMER).
The mRNA expression of ITGB superfamily members was signi cantly related to the immune in ltration