3.1 TRIP13 is highly expressed in most types of human cancer

To ascertain the effects of human TRIP13 on the immunotherapy response, we analyzed TRIP13 mRNA expression via TIMER 2.0 by comparing cancer tissue and normal tissue. TRIP13 was found to have significantly higher transcription levels (P < 0.001) in several cancer types, including bladder (BLCA), breast (BRCA), colorectal (COAD), esophageal (ESCA), head and neck (HNSC), kidney clear cell (KIRC) and papillary (KIRP), liver (LIHC), lung adenocarcinoma (LUAD) and squamous cell (LUSC), prostate (PRAD), rectal (READ), and uterine (UCEC) cancers. However, for cholangiocarcinoma (CHOL) and glioblastoma (GBM), the number of normal tissue samples used for comparison was small (9 and 5 samples, respectively), so it is recommended that more normal tissue samples be included in future studies. Moreover, TRIP13 was highly expressed in CESC (cervical squamous cell carcinoma and endocervical adenocarcinoma) (P < 0.01), PAAD, and thyroid carcinoma (THCA) (P < 0.05) (Fig. 1A).

Fig. 1

TRIP13 mRNA and protein are highly expressed in most types of cancer. A We used TIMER2.0 to assess TRIP13 expression. If there is no other explanation, * in this article means P < 0.05, database, we compared ** means P < 0.01 and *** means P < 0.001. B In the TCGA and GTEx databases, we compared the mRNA expression of TRIP13 in tumor and normal tissues. C We present TRIP13 protein expression in breast cancer, KIRC, colon cancer, hepatocellular carcinoma, LUAD, OV, PAAD, and UCEC samples via CPTAC. D From left to right, the mRNA expression of ACC, BRCA, KICH, KIRC, KIRP, LIHC, LUAD, and THCA is shown according to the major pathological stages (stages I, II, III, and IV)

In GEPIA2, we confirmed that TRIP13 was highly expressed in most types of cancer, including DLBC (diffuse large B-cell lymphoma), OV (ovarian serous cystadenocarcinoma), SARC (sarcoma), SKCM (skin cutaneous melanoma), THYM (thyroid carcinoma), and UCS (uterine carcinosarcoma). However, in LAML (acute myeloid leukemia) and TGCT (testicular germ cell tumors), TRIP13 was expressed at low levels in cancer tissues (Fig. 1B). From the perspective of protein expression, we verified the high TRIP13 expression in nine tumors (breast cancer, clear cell RCC (i.e., KIRC), colon cancer, ovarian cancer, hepatocellular carcinoma, LUAD, OV, PAAD, and UCEC) with CPTAC datasets (Fig. 1C). We also revealed that TRIP13 was expressed differently according to diverse pathological stages in patients with ACC, BRCA, KICH, KIRC, KIRP, LIHC, LUAD, and THCA (Fig. 1D). In particular, in ACC, KICH, KIRC, KIRP, and LUAD, the expression of TRIP13 was positively correlated with pathological stage, which illustrated that TRIP13 might be involved in the progression of these types of cancer.

We also obtained immunohistochemistry and immunofluorescence results from the HPA database to validate TRIP13 expression and cellular localization. By immunohistochemistry, we detected high TRIP13 expression in the tumor tissue of patients with COAD, LIHC, and LUSC compared with normal tissue (Fig. 2A–C). This confirmed the results derived from TIMER 2.0. Furthermore, we found that more TRIP13 localized to the nucleus via immunofluorescence staining analysis of A549 cells (Fig. 2D). A549 cells are derived from non-small cell lung cancer, so the immunofluorescence staining results supported the differential TRIP13 expression shown in Fig. 1 and the immunohistochemistry results. SiHa cells represent cervical carcinoma, and their TRIP13 expression confirmed the differential expression of TRIP13, as shown in Fig. 1 (Fig. 2E). We also performed real-time RT‒qPCR for verification in vitro (Fig. 2F). Compared with the corresponding normal groups (HK-2, L02, and HPDE), the tumor groups (786-O, MHCC-97H, and PANC-1) presented significantly increased expression of TRIP13 (P < 0.01). Especially in LIHC, the increase in relative TRIP13 expression was the most notable. Taken together, these findings suggest that validating TRIP13 expression provides strong evidence that TRIP13 functions as a novel prognostic biomarker.

Fig. 2

High TRIP13 expression was detected via immunohistochemistry in A normal colon tissue and colon adenocarcinoma tissue, B normal liver tissue and liver hepatocellular carcinoma tissue, and C normal lung tissue and lung squamous cell carcinoma tissue. Immunofluorescence revealed TRIP13 expression in the endoplasmic reticulum, microtubules, nucleus, and merged landscape in D A495 cells and E SiHa cells. F The relative expression of TRIP13 in HK-2, L02, HPDE, 786-O, MHCC-97H, and PANC-1 cells was determined via real-time quantitative PCR. The normal cell lines used were HK-2, L02, and HPDE; the cancer cell lines used were 786-O, MHCC-97H, and PANC-1. Normalization gene: β-actin. Number of samples in each group: 3. ** indicates P < 0.01. T-test was used for the comparation

3.2 High TRIP13 expression is related to poor prognosis

To prove that high TRIP13 expression is related to poor prognosis in cancer patients, cancer patients were divided into two groups on the basis of the expression level of TRIP13 (the median level was the cutoff value). As shown in Fig. 3A, high expression of TRIP13 was related to poor OS in ACC, KICH, KIRP, LGG (brain lower grade glioma), LIHC, LUAD, MESO (mesothelioma), and SKCM patients within the TCGA project. In other words, TRIP13 was a risk factor for cancers according to the univariate Cox analysis as a forest plot embedded in the heatmap from GEPIA2 (Fig. 3A and Supplementary Fig. 1). Although the P values in the forest map and survival map were slightly different, they were both less than 0.05, resulting from database differences. For KICH and PAAD, the P values obtained from GEPIA 2 were more significant than 0.05, which requires a larger sample size for confirmation. For UCEC (P = 0.0017), the P value from GDC was more significant.

Fig. 3

High TRIP13 expression was related to a poor prognosis in the TCGA cohort. We analyzed the expression of TRIP13 in different tumors in the TCGA cohort for OS (A) and DFS (B) via the GEPIA2 website

DFS analysis supported the OS results in patients with ACC, KIRC, KIRP, LGG, and LIHC (Fig. 3B and Supplementary Fig. 2), but in LUAD, MESO and SKCM, high TRIP13 expression did not seem to be associated with poor prognosis in the TCGA cohort. A comparison of the PFS and OS results revealed that ACC, KIRC, KIRP, LGG, LIHC, LUAD and MESO patients with high TRIP13 expression had a poor prognosis (Supplementary Fig. 3). We also found that PCPG, PRAD, SARC, and UCEC patients with high TRIP13 expression in the TCGA cohort likely also had a poor prognosis. In terms of OS and DSS, high TRIP13 expression in ACC, KIRC, KIRP, LGG, LUAD, MESO, and SKCM patients was related to a poor prognosis. In the SARC and UCEC cohorts, patients with high TRIP13 expression also had a poor prognosis, as shown in Supplementary Fig. 4.

Various definitions of prognostic markers reveal different clinical features associated with the effects of TRIP13 on prognosis. Overall, we believe that there is a relationship between high TRIP13 expression and poor prognosis, especially in patients with ACC, KIRC, KIRP, and LGG, but in specific types of tumors, recurrence or death after treatment likely depends on complicated interactions between the tumor tissue and the whole-body TRIP13, which still requires further investigation.

3.3 TRIP13 amplification and mutation drive poor prognosis and immunotherapy response alterations through antigenic changes

Genetic alterations cause protein structure alterations, resulting in changes in the antigens on the cell surface that probably affect the prognosis via immunoreaction. Figure 4A shows that the highest frequency of TRIP13 changes (~ 15%) occurred in LUSC patients, with amplification as the dominant type, which might be associated with poor prognosis. The frequency of TRIP13 alteration in the ACC was second highest. In HNSC, DLBC, MESO, PAAD, and THCA, amplification was the only type of alteration of TRIP13 (~ 4%). In most other cancers, such as ACC, ESCA, LUAD, BLCA, and OV, amplification was also the dominant type of alteration (Fig. 4A). Figure 4B shows the sites, types, and case numbers of the TRIP13 mutation. Missense mutations accounted for the greatest number of TRIP13 gene mutations, which were induced by T268I/N/S alterations in SKCM, UCEC, and LUAD cases (Fig. 4B). As a result, T (threonine) was replaced by I (isoleucine), N (asparagine), or S (serine) at the 268th site of the TRIP13 protein. Figure 4C shows the 3D structure of the TRIP13 protein and highlights the 268th site. Taking the ACC as an example, we showed that, compared with the wild-type group, the TRIP13 alteration group had a poorer prognosis [OS (P < 0.001), DFS (P = 0.0287), PFS (P < 0.001), and DSS (P < 0.001)] (Fig. 4D). Furthermore, the expression of TRIP13 in the TRIP13 alteration group was significantly greater than that in the wild-type group in ACC (P = 0.034), verifying the association between the mutation and expression of TRIP13 (Supplementary Fig. 5). The same conclusion was reached for two other cancers, PRAD and SARC. The reason we were unable to obtain DFS data for patients with PRAD was the low number of cases.

Fig. 4

Mutation characteristics of TRIP13 are associated with poor prognosis and immunotherapy response. The cBioPortal website was used to explore the characteristics of the TRIP13 gene mutation. The frequencies of mutation type (A) and site (B) are shown. The 3D structure of TRIP13 (C) is shown, along with the site with the highest alteration frequency (T268I/N/S). The underlying associations between mutation state and the DFS, PFS, OS, and DSS of ACC, PRAD, and SARC (D) patients were analyzed via the cBioPortal website. E In addition, Spearman correlation analysis between the TMB, MSI, and expression of TRIP13 in GDC was performed. The acronyms for each disease in Fig. 3A are as follows: lung squamous cell carcinoma (LUSC), invasive carcinoma (ACC), esophageal carcinoma (ESCA), lung adenocarcinoma (LUAD), bladder urothelial carcinoma (BLCA), ovarian serous cystadenocarcinoma (OV), cervical squamous cell carcinoma (CESC), skin cutaneous melanoma (SKCM), stomach adenocarcinoma (STAD), sarcoma (SARC), uterine corpus endometrial carcinoma (UCEC), uterine carcinoma (UCS), head and neck squamous cell carcinoma (HNSC), liver hepatocellular carcinoma (LIHC), breast invasive carcinoma (BRCA), diffuse large B-cell lymphoma (DLBC), glioblastoma multiforme (GBM), brain lower grade glioma (LGG), colorectal adenocarcinoma (COAD), testicular germ cell tumor (TGCT), kidney renal clear cell carcinoma (KIRC), mesothelioma (MESO), kidney renal papillary cell carcinoma (KIRP), prostate adenocarcinoma (PRAD), pancreatic adenocarcinoma (PAAD), and thyroid carcinoma

TMB and MSI have been used as reliable markers for predicting the immunotherapy response. With respect to TMB, as shown in the left half of Fig. 4E, high expression of TRIP13 in ACC, LUAD, STAD, PAAD, PRAD, BRAC, SARC, LGG, KIRC, BLCA, SKCM (P < 0.001), LUSC, KICH (P < 0.01) and CHOL (P < 0.05) was related to high TMB. However, high TMB in ESCA (P < 0.01) and THYM (P < 0.001) patients was associated with low TRIP13 expression. In patients with STAD, LUSC, UCEC (P < 0.001), ACC, BLCA, LIHC, and PRAD (P < 0.05), MSI was positively correlated with high TRIP13 expression (the right half of Fig. 4E). However, the relationships between high TRIP13 expression and MSI in patients with DLBC (P < 0.01), PCPG, and CESC (P < 0.05) were negative. TMB contributed to the poor prognosis of ACC, LUAD, LGG, KIRC, and SKCM patients because of high TRIP13 expression. For ACC, LUSC, and LIHC, high TRIP13 expression was associated with poor prognosis. Although TMB and MSI seem to be indirect markers of the immunotherapy response, we revealed that TRIP13 was associated with immunoreaction via genetic alterations, at least for a few types of cancer.

3.4 TRIP13 Affects immune infiltration associated with the immunotherapy response and poor prognosis

We then investigated how TRIP13 affects the immunotherapy response, leading to a poor prognosis related to high TRIP13 expression. In this study, we focused on immune infiltration in the tumor microenvironment, which supported the fundamentals of immunotherapy. Both tumor cell accumulation and the tumor microenvironment, which contains endothelium cells (ECs), fibroblasts, structural components, and osmotic immune cells, affect tumor development, invasion, and metastasis [37]. CAFs (cancer-associated fibroblasts) have been reported to act as regulators of the tumor microenvironment through interactions with immune cells [38], which might also be involved in the high TRIP13 expression related to poor prognosis. With the aid of the XCELL algorithm, the results revealed that high expression of TRIP13 was positively related to EC infiltration in most cancers, among which KIRC, LIHC, LUAD, and SKCM were confirmed to be related to poor prognosis (Fig. 5C). Moreover, the results also revealed that the higher the expression of TRIP13 in LGG was, the lower the degree of EC infiltration. However, TGCT and BRCA patients with high TRIP13 expression had fewer CAFs in the TME (Fig. 5A and B). It was previously demonstrated that high TRIP13 expression in these two cancers was not associated with poor prognosis. In terms of immune cell infiltration and TRIP13 expression, CD4 + Th2 cell infiltration, as well as common lymphoid progenitor infiltration, was negatively associated with high TRIP13 expression in almost all cancers (Fig. 5C). Interestingly, high TRIP13 expression was associated with different patterns of immune cell infiltration, even in patients with a poor prognosis, as indicated by OS, PFS, DFS, and DSS. For example, in patients with KIRC and LUAD, high TRIP13 expression is related to poor prognosis and is accompanied by insufficient CD8 + T cells in KIRC and additional dendritic cell activation in LUAD. We speculated that various infiltration patterns might result from the inhibition or overexpression of several immune checkpoints. From the perspective of immune molecular modulators, we further developed a heatmap of immune checkpoints associated with TRIP13 by performing co-expression analysis. For ACC, CESC, GBM, LAML, LUSC, PCPG, SKCM, STAD, TGCT, and THYM, high TRIP13 expression was estimated to be positively correlated with the expression of several immune checkpoints, which might lead to severe immunosuppression, promoting tumorigenesis (Fig. 6A). For KIRC, LIHC, LUAD, and THCA, immune checkpoints were suppressed in patients with high TRIP13 expression. In addition, TRIP13 expression in CHOL, DLBC, MESO, OV, USC, and UVM did not correlate with any of the reported checkpoint genes (Fig. 6A). The activation of immune checkpoints was also noted to be associated with high expression of TRIP13, as disturbances in immune cell function provide insufficient antitumor effects.

Fig. 5

TRIP13 is involved in the immune infiltration of various tumors. The infiltration of CAFs (A) is shown separately. The underlying association between the expression of TRIP13 and the infiltration level of CAFs (B) in the TCGA cohort was analyzed via a variety of algorithms. The heatmap of the immune score and TRIP13 expression in different cancer tissues generated via the xCell algorithm is presented (C)

Fig. 6

TRIP13 affects checkpoint genes and immune infiltration. The heatmap shows the relationship between TRIP13 expression and immune checkpoint-related gene expression in different cancer tissues (A). Violin plots of the distribution of immune infiltration in mutant vs. wild-type tumors are shown (B)

We explored the effect of mutation on immune cell infiltration, possibly accounting for the high TRIP13 expression-related poor prognosis and alterations in immunotherapy efficacy. Via the mutation module of TIMER 2.0, we compared the infiltration ratios of different immune cells in the different samples with or without TRIP13 mutation. We first explored T cells and their related subtypes. As shown in Fig. 6B, TRIP13 gene mutation increased the degree of T-cell gamma delta infiltration (P < 0.05) in LUSC. Mutation of the TRIP13 gene in BLCA, COAD, and KIRC led to increased infiltration of CD8 + T cells (P < 0.05).

Among the few cancer types mentioned above, TRIP13 expression was positively correlated with poor prognosis in KIRC, supporting the findings of at least one OS, DFS, PFS, and DSS. Mutations in TRIP13 altered CD8 + T-cell and T-cell gamma delta infiltration, potentially leading to unresponsiveness to immunotherapy, while the innate immune cell alterations caused by TRIP13 mutations could be considered novel drug targets for monotherapy.

3.5 TRIP13-correlated genes associated with poor prognosis

To investigate the association between the TRIP13 gene and the molecular mechanism related to tumorigenesis, we distinguished the proteins involved in TRIP13-binding and the genes correlated with TRIP13 expression to perform pathway enrichment analysis. The 50 TRIP13-binding proteins are presented as an interaction network in Fig. 7A. The top 100 genes from the GEPIA2 tool were included according to the expression of TRIP13. Cross-analysis of the above two types of genes revealed three crossover proteins, namely, CDC20, RAD1, and MAD2L1 (Fig. 7B). From GEPIA2, we chose six genes strongly related to TRIP13 expression: CCT5 (R = 0.67), BRD9 (R = 0.63), BRIX1 (R = 0.60), NUP155 (R = 0.58), NSUN2 (R = 0.57), KIF2C (R = 0.57) (P < 0.001), and the above-mentioned three intersecting genes: CDC20 (R = 0.5), RAD1 (R = 0.49) and MAD2L1 (R = 0.54) (Fig. 7C). We then utilized the gene expression data of OS in cancer patients to construct a predictive model for prognosis through the LASSO Cox regression model of ACC patients (Fig. 7E). Through multivariate Cox regression analysis, we derived the risk score formula as follows: risk score = (0.2888)*TRIP13 + (0.4426)*CDC20. Using the previously obtained Λ values, we divided the cancer patients into two groups according to their expression levels (Fig. 7E). Patients in the high-expression subgroup died more frequently, and the survival time was shorter. We used R software to draw a transient ROC curve and calculated the AUC at 1, 3, and 5 years to estimate the model’s prediction performance. K‒M curves indicated that patients in the higher groups had notably shorter OS than those in the lower groups did (Fig. 7D). The AUC reached 0.868, 0.94, and 0.885 at one, three, and five years, respectively, demonstrating that our model had relatively satisfactory value in the prediction of different follow-up durations. This model also demonstrated that the essential role of TRIP13-related genes, especially CDC20, in TRIP13 was correlated with poor prognosis.

Fig. 7

TRIP13-related genes related to poor prognosis are highlighted. A We selected 50 TRIP13-binding proteins through the STRING tool. B We cross-analyzed TRIP13-binding genes and related genes. C We utilized GEPIA2 to calculate the correlation ratios between TRIP13 and selected genes, including CCT5, BRD9, BRIX1, NUP155, NSUN2, KIF2C, CDC20, RAD1 and MAD2L1. D K‒M curves indicating survival differences among patients in different expression groups in the ACC, KICH and LUAD cohorts. E A LASSO Cox regression model was subsequently built with the help of ten prognostic genes. Finally, we performed time-dependent ROC analysis of our model with the AUC in the TCGA cohort and the corresponding K‒M chart of OS in different expression groups. F Heatmap of the associations between the above nine selected genes and various cancers in TCGA. G KEGG and GO pathway analyses were performed for the TRIP13-binding gene and interaction gene

Furthermore, we obtained the K‒M curves of different types of cancer patients from the K‒M plotter. For example, we generated three K‒M curves to confirm that there was a positive relationship between high TRIP13 expression and poor prognosis in patients with ACC, KICH, and LUAD (Fig. 7D). A strong correlation between TRIP13 and the above six genes is shown as a heatmap in Fig. 7F. In the enrichment analysis, TRIP13 and coexpressed genes were highly related to the mitotic cell cycle process, regulation of the cell cycle process, and the cell cycle, highlighting that the TRIP13 and TRIP13-related genes CDC20, RAD1, and MAD2L1 might be involved in the above processes, which in turn lead to tumorigenesis and proliferation (Fig. 7G).

3.6 TRIP13 acts as a novel biomarker for the immunotherapy response

Here, we aimed to find evidence for the predictive effect of TRIP13 on the efficacy of specific immunotherapies. With the help of the TIDE website, by comparing TRIP13 with standardized biomarkers, we evaluated the biomarker correlation of TRIP13 and the predictive efficacy AUC in different immunotherapy cohort studies (Fig. 8A). We found that out of 25 immune checkpoint blockade (ICB) sub-cohorts, nine had an AUC > 0.5 when TRIP13 alone was used. In addition, we found that TRIP13 predicted only a greater value than did TMB and that TRIP13 was comparable to T. Clonality (AUC > 0.5 in 8 ICB subgroups) and was lower than TIDE and MSI. Score, CD274, CD8, IFNG, B. Clonality, Merck18. In the cohort of patients receiving ganitumab plus paclitaxel treatment for breast cancer, we also discovered that the responders had higher levels of TRIP13 expression than did the non-responders (Supplementary Fig. 6). This evidence illustrated that TRIP13 could act as a potential immunotherapy response prediction marker alone (Supplementary Table 2). In addition, we analyzed the impact of CTL numbers with different TRIP13 expression levels on patient survival. Among the LAML and UCEC cohorts, lower CTL infiltration was related to a worse prognosis, especially in patients with higher TRIP13 expression (Fig. 8B and C).

Fig. 8

TRIP13 predicted the immunotherapy response. A The bar graph indicates the correlation between TRIP13 and standardized biomarkers of immune evasion of tumors in the ICB sub-cohort. B, C Prognostic K‒M curves related to cytotoxic T lymphocyte (CTL) infiltration and TRIP13 expression

Finally, we compared the changes in the expression levels of FDA-approved drug target genes after TRIP13 mutation. We found that the expression levels of RRM1, RXRA, SLC12A4, FASN, and TYMS (Supplementary Fig. 7A–E) increased with TRIP13 mutation in UCEC. However, in LUSC, the expression of HMGCR decreased (Supplementary Fig. 7F). These findings may provide insight and further research directions for curing patients with this disease.