High expression of COPB2 predicts adverse outcomes: A potential therapeutic target for glioma

Abstract Aims To evaluate the clinical significance of coatomer protein complex subunit beta 2 (COPB2) in patients with glioma using a bioinformatics analysis. Methods Oncomine, GEO, and The Cancer Genome Atlas databases were used to examine the COPB2 transcript levels in glioma tissues. Gene expression profiles with clinical information from low‐grade glioma and glioblastoma (GBM) projects were analyzed for associations between COPB2 expression and clinicopathologic characteristics. Kaplan‐Meier survival and Cox regression analyses were used for survival analysis. Gene set enrichment analysis (GSEA) was conducted to screen the pathways involved in COPB2 expression. Gene set variation analysis (GSVA) and correlograms were performed to verify the correlations between COPB2 and inflammatory responses. Canonical correlation analyses examined whether COPB2‐high patients have more infiltrating inflammatory and immune cells. Results COPB2 was highly expressed in gliomas and high COPB2 expression correlated with shorter overall survival time and several poor clinical prognostic variables. GSEA indicated that some immune‐related pathways and other signaling pathways in cancer were associated with the COPB2‐high phenotype. The GSVA and canonical correlation analysis demonstrated that COPB2 expression was closely linked to inflammatory and immune responses, and higher immune cell infiltration. Conclusions COPB2 may be a potential prognostic biomarker and an immunotherapeutic target for glioma.

molecular classifications of glioma have been established. However, effective and reliable biomarkers that could predict poor prognosis and direct treatment strategies are rare. Thus, the identification of efficient neuropathological biomarkers and therapeutic targets is urgently required.
The coatomer protein complex subunit beta 2 (COPB2), also known as beta prime-COP or beta-Cop, is one of the seven subunits that form coatomer complex I. 4 COPB2 serves as a mediator in the process of protein synthesis which transports proteins from the endoplasmic reticulum (ER) to the Golgi apparatus. 5 Recently, COPB2 has been viewed as a new oncogene in many cancer types.
Bhandari et al 6 reported that the upregulation of COPB2 in breast cancer is associated with age and lymph node metastasis in a validated cohort and promotes tumor cell proliferation and invasion.
Further, a recent study demonstrated that COPB2 could be a potential target gene in prostate cancer. 5  Thus, this research aimed to reveal the association between COPB2 and glioma and explore the potential prognostic value of COPB2 in patients with glioma based on The Cancer Genome Atlas (TCGA), Oncomine, and the Gene Expression Omnibus (GEO) databases. The results indicated that COPB2 was significantly overexpressed in glioma tissues compared with nontumor tissues and that high COPB2 expression was correlated with higher WHO grade, shorter overall survival (OS) time, and several poor clinical prognostic variables. Gene set enrichment analysis (GSEA) showed that some immune-related pathways and other signaling pathways in cancer were associated with the COPB2 high expression phenotype, shedding light on the molecular mechanisms underlying the onset and progression of glioma. Gene set variation analysis (GSVA) and canonical correlation analysis demonstrated COPB2 expression was closely linked to a higher infiltration of immune cells, as well as inflammatory and immune responses.

| Public database and bioinformatics analysis
The transcript level of COPB2 in different cancers was ascertained by the Oncomine database (https ://www.oncom ine.org/resou rce/ main.html), 8 with a threshold set as such-top gene rank 10%, fold change >2, and P-value <1E-4. The microarray data of patients with glioma were downloaded from the GEO (https ://www.ncbi.nlm.nih. gov/geo) 9 public database under accession number GSE16011. 10 Gene expression profile data containing clinical information from low-grade glioma and GBM projects (HTSeq-FPKM) were obtained from TCGA database (http://cance rgeno me.nih.gov/). 11 The data from TCGA were further analyzed for associations between COPB2 expression and clinicopathologic characteristics in glioma.
F I G U R E 1 A, COPB2' expression level in cancers in Oncomine Database: the left box in red indicated the number of datasets with COPB2 hyperexpression and the right box in blue indicated the number of datasets with COPB2 hypo expression after comparing cancerous and normal tissues. B, C, TCGA cohort and GSE16011 dataset from GEO support the findings that indicate COPB2 upregulation in glioma

| Gene set enrichment analysis and gene set variation analysis
To investigate the potential mechanisms underlying the interaction of COPB2 expression on glioma progression, a GSEA 12 was conducted to screen out whether some biological pathways showed statistically significant differences between high and low COPB2 expression groups. For each analysis, gene set permutations were implemented 1000 times. Gene sets with a false discovery rate (FDR) <0.05 and normal P-value <.05 were viewed as significantly enriched. Moreover, GSVA 13 was performed to transform gene expression values into scores for inflammatory response metagene sets, followed by the application of correlograms to further verify correlations between COPB2 and these metagenes.

| Statistical analysis
The statistical analyses were performed utilizing R software v3.5.1.
Descriptive statistics were used to summarize the molecular and clinical characteristics of patients in the TCGA database. To analyze potential relationships between COPB2 and clinicopathologic features, Mann-Whitney U and logistic regression tests were used.
The Kaplan-Meier method and Cox regression analyses were used to compare the impact of COPB2 expression on the OS of TCGA patients alongside with other clinical variables. The remaining correlations between COPB2 expression and inflammatory and immune cell types were detected by using canonical correlation analysis in GraphPad Prism 7 and SPSS 25.0. In all statistical analyses conducted, a P-value <.05 was viewed as statistically significant.

| Glioma COPB2 transcript levels in different databases
Firstly, the transcript levels of COPB2 in different cancers were analyzed. The Oncomine database (one of the main functions of which is gene expression differential analysis) was used to explore the expression of COPB2 mRNA in different cancers ( Figure 1A), and 189 datasets, including 33 144 samples, were included. Relative to normal clinical specimens, COPB2 indicated significant hyper-expression in bladder, brain and central nervous system, breast, esophageal, head and neck, lung, lymphoma, sarcoma, and other cancers, but hypo-expressed in leukemia ( Figure 1A), suggesting that the high expression of COPB2 is common in various types of cancer. The detailed expression profile was summarized in Table S1.
In glioma, 675 glioma patients with COPB2 expression profile data were obtained from TCGA COPB2 is significantly upregulated in tumor tissues relative to nontumor tissues ( Figure 1B, P < .001). In addition, we also used the GSE16011 dataset from GEO database for the purpose of validation ( Figure 1C, P < .001). The results indicated increased transcript levels of COPB2 in glioma.

| TCGA glioma patient characteristics
As the TCGA database contains sufficient glioma samples, we only selected this database for further analysis of the association between gene and clinical characteristics. A total of 1114 cases (Table 1)

| Association with COPB2 expression and clinicopathologic features
To explore the expression pattern of COPB2 in gliomas, mRNA expression profiles from the TCGA database were obtained and analyzed. As shown in Figure 2(A-H), increased expression of COPB2 correlated significantly with tumor grade (P < .001), histological type (P < .001), age (P < .001), KPS (P = .0260), tumor status (P < .001), and vital status (P < .001). Despite the lack of significant differences, a correlated trend was observed for IDH1 mutation (P = .114) and family history of cancer (P = .219).
Univariate analysis using logistic regression revealed that COPB2 expression (ground on median expression value) was linked to poor prognostic clinicopathologic variables (

| COPB2-related signaling pathways based on GSEA
As many signaling pathways contribute to tumor initiation and progression, the poor prognosis of COPB2-high may be related to the numerous signaling pathways activated in glioma. GSEA was utilized to recognize signaling pathways involved in glioma between low and high COPB2 expression cohorts. Significant differences (normalized P < .05, FDR < 0.05) were observed in the enrichment of the MSigDB Collection (kegg.v6.2.symbols.gmt).
Several signaling pathways-especially inflammation-and immunity-related pathways-were enriched in the COPB2 high expression phenotype, including B-cell receptor, T-cell receptor, natural killer (NK) cell-mediated cytotoxicity, antigen processing and presentation, Fc gamma R-mediated phagocytosis, cytokinecytokine receptor interaction, leukocyte transendothelial migration, and other pathways in cancer (please see Figure 3 and Table 4).

| COPB2-related inflammatory response
To better comprehend COPB2-related inflammatory activities, seven immune system-related metagene clusters (comprising 104 genes) 14 that serve as surrogate markers of different immunological cell types were employed (Data S1). COPB2 expression, metagenes expression, age, gender, vital status, tumor grade, and histology of samples were displayed on a heat map in Figure 4.   Figure 5A).

| Relationship between COPB2 and infiltrating immune cells
Previous studies reported that tumor-infiltrating immune cells may represent a crucial pathophysiological factor in the onset and progression of glioma. 15 We examined the relationship between

| D ISCUSS I ON
In our study, the quantitative results indicated that COPB2 had higher expression levels in most cancers compared with normal tissues in the Oncomine database. In addition, COPB2's high expression in patients with glioma was further validated in the TCGA and GEO databases. RNA sequencing data from TCGA were also obtained and analyzed. COPB2 expression levels were correlated  In the last decade, sequencing analysis has been applied comprehensively to explore the molecular mechanisms implicated in the process of disease progression. 20 Recently, some reports found significant differences in COPB2's expression in various tumors.
Underexpression of COPB2 could downregulate the EMT-related protein N-cadherin and vimentin which may promote breast tumor cell invasion. 6 Knockdown of COPB2 led to cell apoptosis by inhibiting the RTK signaling cascade molecules in gastric cancer. 21 In lung adenocarcinoma, 7 patients with COPB2-high had worse survival status than COPB2-low. In the present study, we demonstrated that the overexpression of COPB2 in glioma was correlated with advanced clinicopathologic characteristics and predicts worse outcomes.
These findings indicated that COPB2 may be regarded as a promising target for cancer gene therapy. COPB2 expression was observed to be positively associated with the interferon gene sets (Figures 4 and 5A). Thus, it makes sense to try to combine anti-PD-1 therapy and anti-COPB2 therapy to amplify the efficacy of treatments typically used in isolation.
Some limitations were present in this study: First, the number of patients incorporated in the univariate and multivariate Cox analyses patients were reduced as many patients had missing integrated data on all variables; second, only a small number of healthy samples were used as controls, so additional studies are needed to balance sample size; and lastly, laboratory studies should be carried out to elucidate the precise mechanisms of COPB2 overexpression in human glioma and clarify its relationship with poor prognosis and immunomodulation.

| CON CLUS ION
As far as we are aware, at present, this is the first study to explore the prognostic value of COPB2 in patients with glioma. This study revealed that COPB2 expression was upregulated in glioma samples and was related to adverse outcomes. We also found that COPB2 may represent an important factor in the immunomodulation of the glioma immune microenvironment. Therefore, taken together, COPB2 may act as a potential biomarker of prognosis and immunotherapeutic target for glioma.

CO N FLI C T O F I NTE R E S T
The authors declare no conflict of interest.
TA B L E 5 Secreted and membrane immunosuppressive molecules expressed by glioma cells

CD70
Membrane proteins Inducing apoptosis of T and B cells from PBMCs. 32

FASL
Inducing apoptosis of FAS-expressing T cells. 33 HLA-G Inhibiting proliferation, cytotoxicity by interaction with inhibitory receptors expressed on effector lymphocytes. 34