High TRAF3IP3 Level Predicts Poor Prognosis of Patients with Gliomas

Purpose: TRAF3IP3 is involved in the maturation of immune cells, the development of immune tissues and the immune response of the body. TRAF3IP3 is shown to expressed in a variety of malignant tumor cell lines. Down-regulated expression of TRAF3IP3 in malignant melanoma can inhibit tumor growth. However, the role of TRAF3IP3 in glioma is still unknown. In this study, we aimed to study the relationship between TRAF3IP3 and glioma based on TCGA data. Method: We used the Wilcoxon rank sum test to compare the expression of TRAF3IP3 in glioma and normal tissues. Subsequently, Kruskal-Wallis test, Wilcoxon rank sum test, and logistics regression were used to evaluate the relationship between TRAF3IP3 and clinicopathological parameters of glioma patients. GSEA was used to verify the key signal pathways involved in TRAF3IP3. We used the ssGSEA method to analyze the relationship between the expression level of TRAF3IP3 and the immune inltration in the glioma tumor microenvironment. Finally, we used Kaplan-Meier and COX regression to evaluate the prognostic value of TRAF3IP3. Results: TRAF3IP3 transcription level was highly expressed in gliomas(P<0.001). And the high expression of TRAF3IP3 and WHO grade(OR=3.57(2.42-5.34), P<0.001), IDH status (OR=4.79(3.40-6.83), P<0.001), 1P /19q codeletion (OR=0.07(0.04-0.11), P<0.001), EGFR status (OR=2.77(1.65-4.81), P<0.001), histological type (OR=3.64(2.48-5.43), P<0.001), age (OR=1.64(1.13-2.41), P=0.01), and primary therapy outcome (OR=2.29(1.47-3.61), P<0.001) were signicantly correlated. GSEA showed that six signaling pathways were signicantly enriched in the TRAF3IP3 high expression phenotype group, including JAK STAT signaling pathway, interferon-γ signaling pathway, apoptosis, P53 signaling pathway, PD-1 signaling pathway, and CTLA4 signaling pathway. ssGSEA showed that the expression of TRAF3IP3 was signicantly positively correlated with the inltration of Macrophages, Th17 cells, etc. Multivariate COX regression showed that TRAF3IP3 was an independent prognostic factor for glioma OS (HR=2.169(1.301-3.615), P=0.003). Kaplan-Meier analysis showed that high expression of TRAF3IP3 was associated with worse PFS(HR=2.39(1.39-3.01), P<0.001), DFS(HR=3.02(2.27-4.01), P<0.001) and OS(HR=2.87(2.20-3.75), P<0.001). Conclusion: TRAF3IP3 may play an important role in the occurrence and development of glioma, and may be a potential biomarker for the diagnosis and prognosis of glioma. gliomas on we Our study revealed that high expression of in predicted poor and for leukemia, carcinoma, serous cystadenocarcinoma, pancreatic cancer, prostate cancer, testicular cancer, thyroid cancer, skin melanoma, thymic carcinoma, endometrial carcinoma


Introduction
Glioma is the most common type of tumor in the central nervous system, accounting for 50%-60% of the primary intracranial tumors [1]. In recent years, the incidence of glioma is gradually increasing. It is a type of intracranial tumors with signi cant heterogeneity, usually presenting as diffuse and invasive growth of intracranial masses. From a basic histopathological point of view, gliomas can easily be divided into two categories, low grade (grade I/II) and high grade (grade III/IV) [2]. Due to the particularity of intracranial tumor sites and the limitative treatment methods, the therapeutic effect of glioma, especially advanced glioblastoma, is unsatisfactory. The median overall survival time for glioblastoma is currently 14.6 months, and less than 10% of patients survive more than 5 years [3]. With the development of research on tumor molecular mechanism, more and more molecular genetic characteristics play an important guiding role in tumor diagnosis, treatment and prognosis assessment. Although a great deal of work has been done in the molecular biology of glioma in recent years and considerable progress has been made, it has not been of great help to the improvement of patients' quality of life and the extension of their survival time [4].
TRAF3IP3 also known as T3JAM, is a protein-coding gene located at 1q32.2 [5]. Its protein products can speci cally interact with tumor necrosis factor receptor-related factor 3 (TRAF3). TRAF3IP3 and TRAF3 combine to form detergent-insoluble complex and mediate Jun N-terminal kinase (JNK) signaling pathway activation. The current researches on TRAF3IP3 mainly focus on the immune system, which is mainly involved in regulating the development of T lymphocytes and B lymphocytes in the immune organs [6][7][8].
Previous studies had shown that TRAF3IP3 was highly expressed in the vascular system of breast cancer tissue specimens, suggesting that TRAF3IP3 may be involved in the occurrence and development of breast cancer [9]. Nasarre P, et al [10]found that the expression of TRAF3IP3 in malignant melanoma primary lesion and lymph node metastases were up-regulated by microarray analysis. Then they found that down-regulation of TRAF3IP3 expression in tumor cells could promote apoptosis of tumor cells, and decrease the density of microvascular. Furthermore, they also found that down-regulation of TRAF3IP3 was associated with increased levels of interferon-gamma, which inhibited the growth of tumor cells and the survival of tumor blood vessels.
All of these indicate that TRAF3IP3 may be involved in the occurrence and development of malignant melanoma, promoting the growth and metastasis of malignant melanoma. At present, there is still a lack of research on TRAF3IP3 in glioma. In this study, we aimed to investigate the relationship between TRAF3IP3 and glioma, and to analyzed the prognosis and diagnostic value of TRAF3IP3 in glioma. To achieve this objective, we analyzed the expression of TRAF3IP3 in gliomas based on TCGA. Then, the signaling pathways that TRAF3IP3 may be involved in glioma were analyzed through GSEA. Furthermore, ssGSEA method was used to analyze the relationship of in ltration of immune cells and the expression level of TRAF3IP3 in glioma. Subsequently, we further analyzed the value of TRAF3IP3 expression level in the prognosis assessment of glioma. Our study revealed that high expression of TRAF3IP3 in gliomas predicted poor prognosis and may be a potential molecular target for gliomas.
In this process, we discarded those data without clinical information. The format of the data being downloaded was level 3 HTSeq-FPKM. Subsequently, we converted these level 3 HTSeq-FPKM format data into transcripts per million reads (TPM) format for further analysis. In further analysis, we identi ed a single unavailable or unknown clinical information as missing values. The median of the expression level of TRAF3IP3 was used as the cut-off value, and the tumor samples were divided into high expression group and low expression group.

Gene Set Enrichment Analysis
Based on the gene expression matrix, divided into high expression group and low expression group, GSEA can analyzes the difference in signal pathway enrichment between the two groups, and infers the phenotype and signal pathway that may be associated with genes [14]. In this study, we used the R package clusterPro ler for GSEA analysis and visualization [15]. We analyzed the functions and signal pathways with signi cant differences between the high and low expression groups. GSEA performed 1000 permutations of the genome per analysis. The statistically signi cant indicators and thresholds were set as adjusted P value<0.05 and FDR q value<0.25.

Immune in ltration analysis
Based on the gene markers of the 24 immune cells mentioned in the literature, we used the ssGSEA method of the GSVA package to evaluate the in ltration of 24 immune cells in the tumor, and then used spearman correlation analysis to analyze the correlation between the TRAF3IP3 and the relative in ltration abundance of immune cells [16][17][18]. Furthermore, Wilcoxon rank sum test was used to analyze the differences in the in ltration levels of various immune cells between the TRAF3IP3 high and low expression groups, and P<0.05 was considered statistically signi cant.

Statistical Analysis
In this study, all statistical analysis used R software (v.3.6.2). We used the Wilcoxon rank sum test to compare the expression of TRAF3IP3 in normal samples of GTEx combined with TCGA and the expression of TRAF3IP3 in tumor samples of TCGA. Then we used Kruskal-Wallis rank sum test, Wilcoxon rank sum test, and logistics regression methods to compare the relationship between the expression of TRAF3IP3 and the clinicopathological parameters. Furthermore, we used Cox regression to analyze the prognostic factors of gliomas. The variables with P<0.1 in the single-factor Cox regression will be included in the multi-factor Cox regression. Finally, we used the Kaplan-Meier method to evaluate the prognostic value of TRAF3IP3 in gliomas. The primary end points of prognostic analysis included PFS, DFS and OS. In the above research results, P<0.05 was considered statistically signi cant.

Clinical Characteristics
We collected the clinical data of 670 glioma patients contained in the TCGA database, including WHO classi cation, IDH status, 1P/19q codeletion, primary therapy outcome, gender, race, histological type, EGFR status, PIK3CA status and age. In the subsequent correlation analysis, we found that TRAF3IP3 expression level was signi cantly correlated with WHO classi cation, IDH status, 1P /19q codeletion, primary therapy outcome, histological type, EGFR status and age (Table 1).
TRAF3IP3 involved signaling pathways based on GSEA By using GSEA, the signaling pathways involved in low and high TRAF3IP3 expression data sets were identi ed, and signi cant differences (FDR<0.25 and P.adjust<0.05) were demonstrated in the rich set of MSigDB (C2.all.v7.0.symbols.GMT) collection. Six signaling pathways were found to be signi cantly enriched in the TRAF3IP3 high-expression phenotype group, including interferon gamma signaling pathway, PD-1 signaling pathway, JAK STAT signaling pathway, CTLA4 signaling pathway, P53 signaling pathway and apoptosis signaling pathway (Table 2, Figure 1A-F). All of these suggested the role of TRAF3IP3 in the development of glioma.

Relationship between TRAF3IP3 expression and immune cell in ltration
We used ssGSEA to analyze the relationship between TRAF3IP3 expression level and immune cells in ltration in glioma tumor microenvironment. Spearman correlation analysis revealed a signi cant positive correlation between TRAF3IP3 expression and the in ltration of multiple immune cells, such as macrophages, Th17 cells (Figure 2A-B, D). Then we used the Wilcoxon rank sum test to nd that, compared with the TRAF3IP3 lowexpression group, the in ltration level of the above immune cells in high TRAF3IP3 expression group was signi cantly increased, and the differences were statistically signi cant P<0.001 ( Figure 2C,E).

Relationship between TRAF3IP3 expression and clinicopathological parameters of glioma
Wilcoxon rank sum test was used to compare the expression differences of TRAF3IP3 between the normal samples of GTEx combined TCGA and the unpaired tumor samples of TCGA. It was nally concluded that the expression of TRAF3IP3 in low grade glioma brain glioma, glioblastoma, breast ductal carcinoma, colon cancer, rectal gland carcinoma, lung adenocarcinoma, adrenal cortical carcinoma, bladder urothelial carcinoma, bile duct carcinoma, esophageal cancer, head and neck squamous cell carcinoma, renal clear cell carcinoma, papillary carcinoma, acute myelogenous leukemia, lung squamous carcinoma, ovarian serous cystadenocarcinoma, pancreatic cancer, prostate cancer, testicular cancer, thyroid cancer, skin melanoma, thymic carcinoma, endometrial carcinoma and sarcoma of uterus showed a remarkably high (P<0.05)( Figure  3A). And it was also found that TRAF3IP3 was highly expressed in glioma, and the results were statistically signi cant (P<0.001) ( Figure 3B). Wilcoxon rank sum test showed that high expression of TRAF3IP3 was signi cantly correlated with higher WHO grade, worse primary therapeutic outcome, older age, wild-type IDH, mutant EGFR, no combined loss of 1P / 19Q, and worse histological type(P<0.001) ( Figure 4A-G). The relationship between the clinicopathological features of glioma and TRAF3IP3 expression was analyzed by using logistics regression. The results suggested that high expression of TRAF3IP3 was signi cantly  Table 3). These results suggested that the high expression of TRAF3IP3 in gliomas was associated with poorer clinicopathological parameters.

Univariate and multivariate analyses of survival
We analyzed the diagnostic e cacy of TRAF3IP3 for glioma by receiver operating characteristic (ROC). The results suggested that TRAF3IP3 might be a potential diagnostic molecule(AUC=0.766 CI(0.746-0.786))( Figure 5A). Univariate cox regression analysis showed that the high expression of TRAF3IP3 was associated with poorer OS in glioma patients (HR=2.870(2.196-3.750), P<0.001) (  (HR=2.169(1.301-3.615), P=0.003) were independent prognostic factors in OS ( Table   4). As shown in the Figure 5B Discussion TRAF3IP3, also known as the JNK activation modulator (T3JAM), interacts with the isoleucine zipper of TRAF3 through the coiled-coil domain to up-regulate the JNK pathway [5]. Previous studies have shown that TRAF3IP3 is a gene regulated during B cell development, which promotes autophagy and prevents B cell apoptosis in the marginal zone [6]. In addition, TRAF3IP3 recruits mitogen/extracellular signal-regulated kinase (MEK) to the Golgi apparatus, thereby promoting the interaction of MEK and its activator BRAF and T cell development [7]. TRAF3IP3 on the trans-Golgi network can also regulate thymic Natural Killer T (NKT) 2 cells maturation through the MEK / extracellular signal-regulated kinase (ERK) signaling pathway [8]. Compared with aggressive SVR cells, the level of TRAF3IP3 protein in benign MS1 cells is lower, and the down-regulation of TRAF3IP3 is related to the inhibition of tube formation [10]. TRAF3IP3 is overexpressed in the vasculature of breast tumors and is a marker for patients with skin T-cell lymphoma at risk for progression [9,19]. Previously, researchers found that the down-regulation of TRAF3IP3 in melanoma cells signi cantly increased interferongamma, which is a known tumor growth inhibitor with anti-proliferation and immunomodulatory effects [10]. They also measured the expression level of TRAF3IP3 in a group of tumor cell lines, and it was found that the expression level of TRAF3IP3 in U-87 (glioblastoma) was signi cantly increased [10]. However, the research about the speci c role of TRAF3IP3 in glioma has not been reported. This study analyzed the expression of TRAF3IP3 in gliomas from the data based on the TCGA using bioinformatics methods, and analyzed the relationship between its expression and the clinical pathological parameters of patients, and the prognostic value of TRAF3IP3 in gliomas. We found that in gliomas, elevated TRAF3IP3 expression is associated with the clinicopathological characteristics of progression, and is associated with shorter survival time and worse prognosis.
In order to further study the function of TRAF3IP3 in glioma, we used GSEA to analyze the signal pathways that TRAF3IP3 may participate. We found that the signal pathways enriched in the high expression phenotype of TRAF3IP3 mainly include JAK-STAT signaling pathway, interferon-gamma signaling pathway, apoptosis, P53 signaling pathway, PD-1 signaling pathway, and CTLA4 signaling pathway. It has already been proved that JAK-STAT signaling pathway and P53 signaling pathway played an important role in the occurrence and development of a variety of malignant tumors [20,21]. The abnormality of the apoptosis signaling pathway accelerates occurrence of malignant tumors [20]. More interestingly, we found that the high expression of TRAF3IP3 was involved in the PD-1 signaling pathway and CTLA4 pathways in gliomas. Abnormal activation of these two signal pathways give rise to the immune escape of malignant tumor cells and blocking these two pathways is also the basic principle of immunotherapy, which has become a research hotspot. However, in glioma, the application of immune checkpoint inhibitors has not achieved the same effect as in other tumors, such as malignant melanoma and non-small cell lung cancer [22]. The combined use of two immune checkpoint inhibitors, as well as the combined use of immune checkpoint inhibitors and bevacizumab, failed to signi cantly improve the survival of patients with glioma, nor did it improve the survival rate [23]. Based on these and compared with other malignant tumors, we hypothesize that the PD-1 signaling pathway and CTLA-4 signaling pathway in glioma may generate more interaction networks with other signaling pathways through certain molecules, such as TRAF3IP3. So, we further speculate that the reduced expression of TRAF3IP3 may improve the e cacy of immunotherapy.
Previous studies have shown that in ltrating immune cells in malignant tumors have an important impact on the clinical characteristics of human cancers [24]. Different immune cell in ltration will have different effects on the clinical outcome of patients. Among them, the in ltration of macrophages, dendritic cells, and helper T cells in the center and the invasive edge of the tumor are related to the patient's resistance to radiotherapy and chemotherapy, as well as the shortening of survival [25]. In this study, we used the ssGSEA method to analyze the relationship between the expression level of TRAF3IP3 and the immune in ltration in the glioma tumor microenvironment. We found that the expression of TRAF3IP3 was signi cantly positive with the in ltration of Macrophages, aDCs, iDCs, Th17 cells, etc. Compared to the TRAF3IP3 low expression group, the in ltration levels of the above-mentioned immune cells in the TRAF3IP3 high expression group were signi cantly higher. These indicate that TRAF3IP3 may not only participate in the occurrence and development of gliomas by affecting the PD-1 signaling pathway and CTLA-4 signaling pathway, but also may involve in the regulation of immune cell in ltration in the tumor microenvironment of gliomas, which will affect the prognosis of patients with glioma.
Studies have shown that the method of using mRNA to predict protein expression is far from perfect, and due to the limitations of our research design, it is impossible to clearly assess the correlation between TRAF3IP3 mRNA expression and TRAF3IP3 protein expression in this report [26]. In addition, we may have missed other important signal pathways related to TRAF3IP3, and related pathways need to be further studied. In order to further study the mechanism of TRAF3IP3 in glioma, we have carried out some wet laboratory work.
In short, this is the rst study on the expression and related functions of TRAF3IP3 in gliomas. TRAF3IP3 is highly expressed in gliomas. Highly expressed TRAF3IP3 predicts a worse survival period for patients with gliomas. TRAF3IP3 may be a new biomarker for glioma, which may guide drug development and improve patient prognosis.     JAK STAT signaling pathway(c), CTLA4 signaling pathway(d), P53 signaling pathway(e) and apoptosis(f).

Figure 2
The expression level of TRAF3IP3 was associated with the immune in ltration in the tumor microenvironment. The forest plot show the correlation between TRAF3IP3 expression and immune cell subsets(a). We used spearman correlation to analyze the correlation between TRAF3IP3 expression and macrophages(b), Th17 cells(d). Wilcoxon rank sum test was used to analyze the difference in the in ltration level of macrophages(c), Th17 cells(e) between the high and low TRAF3IP3 expression groups, and the results were statistically signi cant (P < 0.001).