Expression and correlation of TRAF family genes

The analysis of 33 cancer types revealed that TRAF2, TRAF3, TRAF4, and TRAF7 are predominantly overexpressed, whereas TRAF1, TRAF5, and TRAF6 exhibit lower expression levels (Fig. 1A). Detailed examination indicated that the highest expressions of TRAF1, TRAF2, TRAF3, TRAF5, and TRAF7 were in Cholangiocarcinoma (CHOL), while TRAF4 was most expressed in Uterine Corpus Endometrial Carcinoma (UCEC) and TRAF6 in Glioblastoma multiforme (GBM) (Fig. 1B). Furthermore, the study investigated the interrelations within the TRAF gene family. A significant positive correlation was observed between TRAF3 and TRAF6 (Correlation coefficient = 0.37, Fig. 1C), whereas TRAF4 showed a notable negative correlation with TRAF6 (Correlation coefficient = − 0.22, Fig. 1C).

Fig. 1

Expression and correlation of TRAF family genes across various cancers. A Depicts the expression levels of TRAF family genes in different tumor types. B Showcases the specific expression levels of TRAF family genes in each cancer type, with red indicating overexpression and green indicating underexpression. C Correlations among TRAF family genes, where blue signifies positive correlation and red signifies negative correlation. The size of each circle is proportional to the absolute value of the correlation coefficient

We sourced TRAF family gene expression RNA sequencing data (TCGA data) from the UCSC database, specifically examining TRAF1 expression across various cancer types (Table 1). Our analysis revealed elevated TRAF1 expression in several cancers, notably CHOL, colon adenocarcinoma (COAD), head and neck squamous cell carcinoma (HNSC), kidney renal clear cell carcinoma (KIRC), kidney renal papillary cell carcinoma (KIRP), liver hepatocellular carcinoma (LIHC), lung adenocarcinoma (LUAD), and stomach adenocarcinoma (STAD). Conversely, reduced TRAF1 expression was observed in Bladder urothelial carcinoma (BLCA), Kidney chromophobe (KICH), and Uterine corpus Endometrial carcinoma (UCEC) (Fig. 2A). Our investigation further identified that TRAF2 demonstrated increased expression in a variety of cancer types, including BLCA, breast invasive carcinoma (BRCA), CHOL, COAD, esophageal carcinoma (ESCA), HNSC, KICH, KIRC, KIRP, LIHC, LUAD, LUSC, prostate adenocarcinoma (PRAD), rectum adenocarcinoma (READ), STAD, and UCEC. However, a diminished expression of TRAF2 was observed in thyroid carcinoma (THCA) (Fig. 2B). In addition, our data revealed that TRAF3 was more prominently expressed in cancers, such as CHOL, ESCA, HNSC, KICH, KIRC, LIHC, LUSC and STAD. In contrast, lower levels of TRAF3 expression were noted in BRCA, COAD, GBM, and THCA (Fig. 2C). Elevated expression of TRAF4 was observed in a range of cancer types, including BLCA, BRCA, CHOL, COAD, ESCA, GBM, HNSC, KIRP, LIHC, LUAD, LUSC, PRAD, READ, STAD, and UCEC. In contrast, lower expression levels of TRAF4 were noted in KICH and KIRC (Supplementary Fig. 1A). Furthermore, TRAF5 showed increased expression in BRCA, CHOL, COAD, GBM, HNSC, KIRC, LIHC, LUAD, LUSC, READ, and STAD, while its expression was reduced in BLCA, KICH, PRAD, and UCEC (Supplementary Fig. 1B). In addition, TRAF6 exhibited higher expression in CHOL and STAD, but lower expression in BLCA, BRCA, COAD, KIRP, LUAD, LUSC, PRAD, THCA, and UCEC (Supplementary Fig. 1C). Finally, TRAF7 was more highly expressed in BLCA, BRCA, CHOL, COAD, ESC), GBM, HNSC, KIRP, LIHC, LUAD, LUSC, PRAD, READ, STAD, THCA, and UCEC, with a reduced expression in KIRC (Supplementary Fig. 1D).

Table 1 Differential expression of TRAF family genes in different cancer typesFig. 2

Differential expression of TRAF family genes in pan-carcinoma and para-carcinoma. A TRAF1. B TRAF2. C TRAF3. Red represents tumour and blue indicates normal tissue. (*P < 0.05, **P < 0.01 and ***P < 0.001)

Prognostic value of TRAF family genes in pan-cancer

To elucidate the prognostic significance of TRAF family genes across various cancers, we conducted a comprehensive analysis utilizing multiple databases. Kaplan–Meier survival analyses revealed a notable association between the expression levels of TRAF family genes and patient prognosis in certain cancers. Specifically, in some cancer types, a trend was observed, where patients with higher TRAF family gene expression exhibited an increased survival rate compared to those with lower expression. Conversely, in other cancer forms, the data indicated that elevated TRAF family gene expression correlated with a decreased survival rate, highlighting a complex and variable impact of these genes on cancer prognosis (Table 1). In this study, TRAF1 was found to function as an oncogene in several cancers. Notably, in COAD (OS: n = 448, p = 0.030, Fig. 3A), KIRC (OS: n = 531, p = 0.012, Fig. 3B), Brain Lower Grade Glioma (LGG) (OS: n = 524, p < 0.001, Fig. 3C), and THYM (OS: n = 118, p = 0.038, Fig. 3D), elevated expression of TRAF1 was associated with poorer patient outcomes. Conversely, in HNSC (OS: n = 501, p = 0.030, Fig. 3E), mesothelioma (MESO) (OS: n = 84, p = 0.049, Fig. 3F), pancreatic adenocarcinoma (PAAD) (OS: n = 177, p = 0.003, Fig. 3G), and skin cutaneous melanoma (SKCM) (OS: n = 457, p < 0.001, Fig. 3H), TRAF1 expression correlated with a protective effect, indicating its role as a potential inhibitor in these cancer types. TRAF2 was identified as an oncogenic factor in several cancers. It played a significant role in adrenocortical carcinoma (ACC) (OS: n = 79, p = 0.001, Fig. 3I), COAD (OS: n = 448, p = 0.049, Fig. 3J), LGG (OS: n = 524, p = 0.002, Fig. 3K), LIHC (OS: n = 368, p = 0.027, Fig. 3L), and MESO (OS: n = 84, p < 0.001, Fig. 3M). Interestingly, in STAD (OS: n = 350, p = 0.005, Fig. 3N), TRAF2 exhibited an oncostatic role. Furthermore, TRAF3 demonstrated carcinogenic properties in ACC (OS: n = 79, p < 0.001, Fig. 3O), LIHC (OS: n = 368, p = 0.010, Fig. 3P), THYM (OS: n = 118, p = 0.021, Fig. 3Q), and Uveal Melanoma (UVM) (OS: n = 80, p < 0.001, Fig. 3R). Conversely, it showed a protective effect in LGG (OS: n = 524, p < 0.001, Fig. 3S) and PAAD (OS: n = 177, p = 0.020, Fig. 3T). Finally, TRAF4 was associated with oncogenic effects in ACC (OS: n = 79, p = 0.003, Supplementary Fig. 2A), HNSC (OS: n = 501, p = 0.023, Supplementary Fig. 2B), KIRC (OS: n = 531, p = 0.005, Supplementary Fig. 2C), LIHC (OS: n = 368, p = 0.032, Supplementary Fig. 2D), and Sarcoma (SARC) (OS: n = 260, p = 0.036, Supplementary Fig. 2E). In addition, in BLCA (OS: n = 406, p = 0.015, Supplementary Fig. 2F), TRAF4 also demonstrated oncogenic properties. In this study, TRAF5 was implicated in oncogenic processes in several cancers. Specifically, it contributed to oncogenesis in ACC (OS: n = 79, p < 0.001, Supplementary Fig. 2G), KIRC (OS: n = 531, p = 0.001, Supplementary Fig. 2H), KIRP (OS: n = 286, p = 0.007, Supplementary Fig. 2I), LGG (OS: n = 524, p < 0.001 Supplementary Fig. 2 J), and LIHC (OS: n = 368, p = 0.011, Supplementary Fig. 2 K). However, it demonstrated oncostatic effects in BLCA (OS: n = 406, p = 0.003, Supplementary Fig. 2L) and UCEC (OS: n = 544, p = 0.028, Supplementary Fig. 2 M). Moreover, TRAF6 was found to exert inhibitory effects on cancer progression in KIRC (OS: n = 531, p < 0.001, Supplementary Fig. 2N) and READ (OS: n = 158, p = 0.049, Supplementary Fig. 2O). Finally, TRAF7 was associated with carcinogenic activity in several cancer types, including KIRC (OS: n = 531, p = 0.003, Supplementary Fig. 2P), LGG (OS: n = 524, p < 0.001, Supplementary Fig. 2Q), LUSC (OS: n = 493, p = 0.028, Supplementary Fig. 2R), Ovarian Serous Cystadenocarcinoma (OV) (OS: n = 378, p = 0.036, Supplementary Fig. 2S), PAAD (OS: n = 177, p = 0.037, Supplementary Fig. 2 T), and SKCM (OS: n = 457, p = 0.002, Supplementary Fig. 2U).

Fig. 3

Kaplan–Meier survival curves comparison of high and low expression of TRAF family gene in pan-cancer. OS survival curves of EGFR in different cancers. A COAD (OS: n = 448, p = 0.030). B KIRC (OS: n = 531, p = 0.012). C LGG (OS: n = 524, p < 0.001). D THYM (OS: n = 118, p = 0.038). E HNSC (OS: n = 501, p = 0.030). F MESO (OS: n = 84, p = 0.049). G PAAD (OS: n = 177, p = 0.003). H SKCM (OS: n = 457, p < 0.001). I ACC (OS: n = 79, p = 0.001). J COAD (OS: n = 448, p = 0.049). K LGG (OS: n = 524, p = 0.002). L LIHC (OS: n = 368, p = 0.027). M MESO (OS: n = 84, p < 0.001). N STAD (OS: n = 350, p = 0.005). O ACC (OS: n = 79, p < 0.001). P LIHC (OS: n = 368, p = 0.010). Q THYM (OS: n = 118, p = 0.021). R UVM (OS: n = 80, p < 0.001). S LGG (OS: n = 524, p < 0.001). T PAAD (OS: n = 177, p = 0.020)

In our comprehensive analysis utilizing COX regression (Fig. 4), we observed distinct prognostic implications of TRAF family genes across various cancer types. Notably, TRAF1 emerged as a protective prognostic factor in BRCA, HNSC, PAAD, SKCM, and UCEC (HR < 1, P < 0.05, Fig. 4, Table 2). Conversely, TRAF1 demonstrated a negative prognostic impact in COAD, KIRC, LGG, and THYM (HR > 1, P < 0.05, Fig. 4, Table 2). Moreover, TRAF2 was characterized as a low-risk gene in Cervical squamous cell carcinoma and endocervical adenocarcinoma (CESC), lymphoid neoplasm diffuse large b-cell lymphoma (DLBC), and STAD (HR < 1, P < 0.05, Fig. 4, Table 2). Conversely, TRAF2 was identified as a high-risk gene in ACC, COAD, KIRC, LGG, LIHC, MESO, and pheochromocytoma and paraganglioma (PCPG) (HR > 1, P < 0.05, Fig. 4, Table 2). TRAF3, on the other hand, was deemed a low-risk gene in LGG and PAAD (HR < 1, P < 0.05, Fig. 4, Table 2), while it was classified as a high-risk gene in ACC, Acute Myeloid Leukemia (LAML), LIHC, THYM, and UVM (HR > 1, P < 0.05, Fig. 4, Table 2). Furthermore, TRAF4 was recognized as a low-risk gene in BLCA (HR < 1, P < 0.05, Fig. 4, Table 2). In contrast, it was noted as a high-risk gene in ACC, KIRC, LGG, LIHC, and UCEC (HR > 1, P < 0.05, Fig. 4, Table 2). In addition, TRAF5 was categorized as a low-risk gene in BLCA and SARC (HR < 1, P < 0.05, Fig. 4, Table 2), whereas it emerged as a high-risk gene in ACC, KIRC, KIRP, LGG, PCPG, and UVM (HR > 1, P < 0.05, Fig. 4, Table 2). In our findings, TRAF6 was characterized as a low-risk gene in KIRC and READ (HR < 1, P < 0.05, Fig. 4, Table 2). However, it was identified as a high-risk gene in PAAD (HR > 1, P < 0.05, Fig. 4, Table 2). Furthermore, TRAF7 emerged as a high-risk gene across multiple cancer types, specifically in ACC, KIRC, LGG, LUAD, PAAD, and SKCM (HR > 1, P < 0.05, Fig. 4, Table 2).

Fig. 4

Correlation between TRAF family gene expression and patient survival across various cancer types, as determined by COX regression analysis. It features distinct colored lines, each representing a different TRAF gene. These lines indicate the risk value associated with each gene in different tumor types, hazard ratio < 1 is indicative of a low-risk gene, while HR > 1 signifies a high-risk gene. This visual representation aids in the clear understanding of the complex relationship between TRAF gene expression and cancer prognosis

Table 2 TRAF family gene was associated with the prognosis risk of different cancers by COX analysis

To further delineate the prognostic significance of TRAF family genes in a pan-cancer context, we conducted a validation study using the Kaplan–Meier plotter. This analysis reinforced our findings, revealing that TRAF1, TRAF2, TRAF4, and TRAF5 play an inhibitory role in BLCA (Fig. 5A). In addition, TRAF1 and TRAF3 were identified as suppressor oncogenes in BRCA (Fig. 5B). In the context of CESC, both TRAF1 and TRAF2 emerged as suppressor oncogenes (Fig. 5C). For ESCA, TRAF2 and TRAF5 were recognized as suppressor genes (Fig. 5D). Finally, TRAF2 and TRAF7 were found to inhibit OV (Fig. 5E). In the context of HNSC, our findings indicate that TRAF1 and TRAF5 act as tumor suppressors, whereas TRAF2 emerges as an oncogenic factor (Fig. 5F). In KIRP, both TRAF3 and TRAF4 are identified as carcinogenic factors, contrastingly, TRAF5 functions as an inhibitor (Fig. 5G). Within KIRC, our analysis shows that TRAF1, TRAF2, TRAF4, TRAF5, and TRAF7 play oncogenic roles, while TRAF6 appears to serve as an oncostatic factor (Fig. 5H). In the case of LIHC, TRAF2, TRAF3, TRAF4, TRAF5, and TRAF7 are identified as suppressor genes (Fig. 5I). In addition, patients with heightened TRAF1 expression in TGCT are associated with a poorer prognosis compared to those with lower expression levels (Fig. 5J). In LUAD, our analysis identified TRAF2 and TRAF7 as oncogenic factors, whereas TRAF1 and TRAF6 were found to play oncostatic roles (Fig. 5K). Regarding PCPG, TRAF1 and TRAF6 were determined to be carcinogenic factors (Fig. 5L). In LUSC, TRAF2, TRAF6, and TRAF7 were recognized as inhibitory factors (Fig. 5M). In STAD, patients exhibiting high TRAF6 expression were associated with a worse prognosis, while those with elevated TRAF2 and TRAF4 expressions had a more favorable prognosis compared to individuals with low expression levels (Fig. 5N). In PAAD, TRAF2 and TRAF7 were characterized as carcinogenic, in contrast, TRAF1, TRAF3, and TRAF5 acted as inhibitors (Fig. 5O). In the case of UCEC, a dismal prognosis was associated with high TRAF4 expression, whereas patients with high expressions of TRAF1 and TRAF5 had better prognoses compared to those in the low-expression group (Fig. 5P). In patients diagnosed with SARC, elevated expression levels of TRAF4 correlate with an unfavorable prognosis. Conversely, increased levels of TRAF1, TRAF5, and TRAF6 are associated with a more favorable prognosis, in contrast to patients exhibiting lower expression levels of these factors (Fig. 5Q). Similarly, in READ cases, a high TRAF4 expression is indicative of an adverse prognosis. On the other hand, elevated TRAF6 expression suggests a more positive outcome when compared to patients with lower expression levels of this factor (Fig. 5R). Furthermore, in THCA, a high expression of TRAF3 is linked to a detrimental prognosis, whereas increased TRAF5 expression correlates with a better prognosis relative to those with lower expression levels (Fig. 5S). Finally, in THYM patients, heightened expression of both TRAF1 and TRAF3 is associated with a less favorable prognosis when compared to patients with lower expression levels of these factors (Fig. 5T).

Fig. 5

Overall survival curves comparing the high and low expression of TRAF family gene in various cancer types in Kaplan–Meier plotter database. A TRAF1, TRAF2, TRAF4 and TRAF5 in BLCA. B TRAF1and TRAF3 in BRCA. C TRAF1 and TRAF2 in CESC. D TRAF2 and TRAF5 in ESCA. E TRAF2 and TRAF7 in OV. F TRAF1, TRAF4, and TRAF5 in HNSC. G TRAF3, TRAF4 and TRAF5 in KIRP. H TRAF1, TRAF2, TRAF4, TRAF5, TRAF6 and TRAF7 in KIRC. I TRAF2, TRAF3, TRAF4, TRAF5 and TRAF7 in LIHC. J TRAF1 in TGCT. K TRAF1, TRAF2, TRAF5 and TRAF7 in LUAD. L TRAF1 and TRAF6 in PCPG. M TRAF2, TRAF6 and TRAF7 in LUSC. N TRAF2, TRAF4 and TRAF6 in STAD. O TRAF1, TRAF3, TRAF5, TRAF6 and TRAF7 in PAAD. P TRAF1, TRAF4 and TRAF5 in UCEC. Q TRAF1, TRAF4, TRAF5 and TRAF6 in SARC. R TRAF4 and TRAF6 in READ. S TRAF3 and TRAF5 in THCA. T TRAF1 and TRAF3 in THYM

In our comprehensive analysis, we explored the association between TRAF family gene expression and overall cancer prognosis using the PrognoScan online database. This database compiles data sourced from the GEO database. For a detailed exposition of these findings, refer to Table S2. Upon examining the outcomes from various databases, a spectrum of findings emerged regarding the TRAF family's expression in diverse cancer types, as summarized in Table 3. While some discrepancies were noted, a portion of the data exhibited consistency. These variations in findings may stem from differences in data collection methodologies and underlying theoretical frameworks that account for distinct biological characteristics of each cancer type. Notwithstanding these discrepancies, the congruent data present a more robust indication of the TRAF family's prognostic significance in oncology.

Table 3 Association between TRAF family gene high expression and pan-cancer in different databaseAssociation of TRAF family gene expression with TME, stemness score, and immune subtypes in pan-cancer and selected cancers

The TME plays a critical role in tumor development, metastasis, and response to therapy. In this context, we investigated the relationship between TRAF family gene expression and various aspects of the TME. We utilized the ESTIMATE method to derive stromal and immune scores, and also calculated tumor purity, as detailed in Table S3. Our analysis revealed that TRAF family gene expression demonstrates significant positive or negative correlations with both stromal (Fig. 6A) and immune (Fig. 6B) scores across a range of cancers. In a pan-cancer analysis, the expression of TRAF family genes exhibited significant correlations, both positive and negative, with RNAss (Fig. 6C) and DNAss (Fig. 6D). Specifically, within PAAD, TRAF1 and TRAF3 demonstrated positive correlations with stromal, immune, and estimated scores while showing negative correlations with tumor purity. Conversely, TRAF2, TRAF4, and TRAF7 displayed positive correlations with stromal, immune, and estimated scores, and negative correlations with tumor purity. Furthermore, a negative correlation of TRAF1 with DNAss was observed, in contrast to the positive correlations of TRAF2 and TRAF7. In addition, TRAF1, TRAF3, and TRAF5 showed negative correlations with RNAss, whereas TRAF2, TRAF4, and TRAF7 exhibited positive correlations.

Fig. 6

Correlation of TRAF family gene expression with tumor microenvironment, Stemness score in pan-cancer. A, B, C, D TRAF family gene expression associated with stromal score, immune score, ESTIMAT score and tumour purity in different cancers. E, F TRAF family gene expression associated with DNAss and RNAss in different cancers. Red dots indicate a positive correlation, and blue dots indicate a negative correlation. G Correlation analysis of TRAF family gene expression with Stemness score, tumor microenvironment in PAAD

In their comprehensive immunogenomic study, Thorsson et al. analyzed over 10,000 tumors encompassing 33 distinct cancer types from the TCGA database. This analysis led to the identification of six immune subtypes: wound-healing (C1), IFN-γ dominant (C2), inflammatory (C3), lymphocyte depleted (C4), immunologically quiet (C5), and TGF-β dominant (C6) [18]. These subtypes have significant implications for both prognosis and immune regulation. Building upon this foundation, this study investigated the association between TRAF family genes and these immune subtypes. Our findings reveal that all seven TRAF family genes are correlated with specific immune subtypes (Fig. 7A). Within the context of PAAD, we discovered associations between immune subtypes and five TRAF genes, namely TRAF1, TRAF2, TRAF3, TRAF4, and TRAF7 (Fig. 7B).

Fig. 7

TRAF family gene expression level of different immune subtype in pan-cancer and KIRC. A All seven genes of the TRAF family were associated with immune subtypes in pan-cancer. B TRAF family gene expression level in different immune subtypes in PAAD

Correlation between the expression of TRAF family genes and drug sensitivity

We explore the relationship between TRAF family gene expression and drug sensitivity. Utilizing the CellMiner database, we accessed sensitivity data for FDA-approved drugs and those currently in clinical trials. Our analysis demonstrates a distinct correlation between TRAF family gene expressions and drug sensitivity, as depicted in Fig. 8 and Table S4. Specifically, TRAF1 shows a negative correlation with the sensitivity of METHOTREXATE, Fluorouracil, and Malacid (Fig. 8B, F, G). In contrast, TRAF2 exhibits a positive correlation with the sensitivity of Gemcitabine, Triapine, Floxuridine, Fludarabine, 6-THIOGUANINE, and Acrichine (Fig. 8D, H, J, L, P), but a negative correlation with Mithramycin and Depsipeptide (Fig. 8E, O). TRAF5 correlates positively with the sensitivity of PX-316, Dexrazoxane, and Hypothemycin (Fig. 8A, K, N). TRAF6 is positively correlated with PX-316 sensitivity (Fig. 8C), and TRAF7 shows a positive correlation with Floxuridine sensitivity (Fig. 8I).

Fig. 8

Correlation between the expression of TRAF family genes and drug sensitivity. TRAF1 is negatively correlated with sensitivity of METHOTREXATE, Fluorouracil and malacid B, F, G. TRAF2 was positively correlated with Gemcitabine, Triapine, Floxuridine, Fludarabine, 6-THIOGUANINE and Acrichine sensitivity D, H, J, L, P, but negatively correlated with Mithramycin and Depsipeptide sensitivity E, O. TRAF5 was positively correlated with PX-316, Dexrazoxane and HYPOTHEMYCIN sensitivity (A, K, N). TRAF6 was positively correlated with the sensitivity of PX-316 (C). TRAF7 was positively correlated with Floxuridine sensitivity (I)

Expression of TRAF family genes in PAAD

Leveraging data from the TCGA database for pancreatic cancer and corresponding normal tissue data sourced from GTEx, extracted via the UCSC database, we conducted a comparative analysis. This analysis revealed that TRAF1, TRAF2, TRAF3, TRAF4, TRAF6, and TRAF7 exhibit higher expression levels in cancer tissues, whereas TRAF5 is expressed at lower levels in these tissues (Fig. 9A). Furthermore, our correlation analysis with clinicopathological features indicates a significant association of TRAF6 expression with the TNM stage of pancreatic cancer. In addition, TRAF4, TRAF5, TRAF6, and TRAF7 demonstrate correlations with the degree of pancreatic cancer differentiation (Fig. 9B, C).

Fig. 9

TRAF family genes in PAAD. A Expression level of TRAFs in TCGA–PAAD and GTEx-pancreas. B TRAF6 was differentially expressed in the TNM stage of PAAD. C TRAF4, TRAF5, TRAF6, and TRAF7 were differentially expressed in the degree of differentiation of PAAD. D mRNA expression of TRAF genes in 40 PAAD tissues and para-cancerous tissues. E–L Diagnostic value analysis of TRAFs. ns: not statistically significant, *: p < 0.05, **: p < 0.01, ***: p < 0.001

In addition, we employed qRT–PCR to validate the expression levels of TRAF family genes in PAAD. Analyzing 40 pairs of PAAD and adjacent non-tumorous tissues, our findings indicate that the relative mRNA expression levels of TRAF4, TRAF6, and TRAF7 are elevated in PAAD tissues. Conversely, TRAF2 and TRAF5 exhibit higher expression in adjacent non-tumorous tissues (Fig. 9D). Notably, the expression levels of TRAF1 and TRAF3 in PAAD tissues do not significantly differ from those in adjacent non-tumorous tissues (Fig. 9D). These observations align closely with the results discussed in the previous sections.

Furthermore, this study conducted an in-depth evaluation of the diagnostic potential of TRAF family genes in pancreatic cancer. Data for TCGA–PAAD and normal pancreatic tissues from GTEx were downloaded from the UCSC XENA platform. We employed ROC analysis using the "pROC" package in R and visualized the results with "ggplot2". The analysis reveals that all TRAF family genes exhibit high accuracy in differentiating between normal and tumor tissues in pancreatic cancer, as evidenced by the ROC curves presented in Fig. 9E–L.

TRAF6 promotes proliferation, migration and invasion of pancreatic cancer cells in vitro.

To further explore the role of TRAFs in pancreatic cancer, we first verified the correlation between TRAFs expression and prognosis in 40 pairs of pancreatic cancer tissue samples. The results showed that only TRAF6 was associated with pancreatic cancer prognosis (Fig. 10A–E). Subsequently, we examined TRAF6 expression in HPNE and PDAC cell lines (BxPC-3, CFPAC-1, MiaPaCa-2, and PANC-1). We found that TRAF6 was highly expressed in PDAC cells, and PANC-1 and MiaPaCa-2 cell lines were used for further studies (Fig. 10F). To assess the biological function of TRAF6 in PDAC cells, we designed and synthesized short interfering RNA to knockdown TRAF6 expression and verified the effectiveness of TRAF6 knockdown by qRT–PCR (Fig. 10G). Clone formation experiments showed that down-regulation of TRAF6 inhibited proliferation (Fig. 10H). In addition, the results of wound healing assay and transwell assay showed that knockdown of TRAF6 significantly reduced the migration and invasion ability of PDAC cells (Fig. 10I, J). Taken together, these observations illustrate the oncogenic role of TRAF6 in PDAC cells.

Fig. 10

TRAF6 promotes the proliferation, migration and invasion of PDAC cells in vitro. A–E Kaplan–Meier survival curves comparison of high and low expression of TRAFs in pancreatic cancer. F Expression of TRAF6 in PDAC cells. G qRT–PCR analysis of TRAF6 in PANC-1 and MiaPaCa-2 cells transfected with TRAF6 siRNAs. H Colony formation assays were performed in PDAC cells transfected with TRAF6 siRNAs. I, J Wound healing assays and Transwell assays were used to investigate the migratory and invasive ability of TRAF6-knockdown PDAC cells. (Values are expressed as the means ± SDs; *P < 0.05, **P < 0.01 and ***P < 0.001)