Tumor mutational burden, gene expression patterns, and immune-cell deconvolution in testicular germ cell tumors

Introduction

Testicular germ cell tumor (TGCT), the incidence of which has been increasing since the mid-20^th century, is the most prevalent solid carcinoma in males aged 15 to 44 years (1,2). TGCTs are histologically classified into seminoma, nonseminoma, and spermatocytic tumor, with the first two types constituting 98% to 99% of all TGCT cases (1,3). Although radical orchiectomy, cytotoxic chemotherapy, radiotherapy, or their combination can offer cure for most patients with TGCT (2,4), about 15–30% of patients with poor prognosis experience tumor recurrence after surgery and chemotherapy (2). Thus, treatment for TGCT remains a clinical challenge.

In recent years, there has been a growing body of evidence suggesting that tumor mutational burden (TMB) may be an inclusive biomarker that can indicate the number of mutation errors in tumor, demonstrating an association with the tumor microenvironment (TME) (5-7). For instance, Tang et al. identified six genes (ADRA2A, CXCL12, S1PR1, ADAMTS9, F13A1, and SPON1) that were associated with the overall survival of patients with bladder cancer (8). Moreover, a gene signature comprising nine TMB-related genes was found to correlate with the prognosis and TME of patients with clear-cell renal cell carcinoma (9). TMB was also found to be differentially enriched in patients with metastatic breast cancers as compared to those with localized breast cancers (10). However, few studies have been conducted regarding TMB in the context of TGCT, and a better understanding of the relationship between TMB and TGCT is needed (9).

In this study, we aimed to interrogate the association of TMB and the messenger RNA (mRNA) expression profile and immune infiltration of TGCT and identify its potential biomarkers. Datasets from the Gene Expression Omnibus (GEO) (11), The Cancer Genome Atlas (TCGA) (12), and Genotype-Tissue Expression (GTEx) (13) were analyzed. First, TMB-related genes were determined via DESeq2 (14) and were then imported to Gene Ontology (GO) (15,16) and Kyoto Encyclopedia of Genes and Genomes (KEGG) (17) databases, where biological process (BP) and functional pathway were analyzed via gene set enrichment analysis (GSEA). Only TMB-related genes differentially expressed between TGCT and normal control samples were further analyzed via the Search Tool for the Retrieval of Interacting Genes/Proteins (STRING) database (18), through which a protein-protein interaction (PPI) network was developed. Subsequent network analysis, including both degrees, modules, and Kaplan-Meier analyses with reference to progression-free interval (PFI), was applied to identify four hub genes (FOXA2, IRX3, MYH7, and TNNT2). GSEA was also conducted to implement GO and KEGG analyses of the four hub genes. Following this, analysis of the association with clinical characteristics was conducted in TCGA and the GSE99420 dataset (19). The Human Protein Atlas (20,21) database was used to examine the protein levels of the hub genes in TGCT and normal testis samples. The prognostic implications of four hub genes were confirmed via receiver operating characteristic (ROC) analysis and univariate Cox analysis in TCGA and GSE10783 (22) datasets. The CIBERSORT algorithm (23,24), Wilcoxon rank-sum test, and Spearman analysis were used to determine the association between TMB, the four hub genes, and the infiltration levels of 22 immune cell types. We present this article in accordance with the REMARK reporting checklist (available at https://tau.amegroups.com/article/view/10.21037/tau-2025-514/rc).

Methods

TCGA and GTEx data sources

Datasets from the GEO (11), TCGA (12), and GTEx (13) were analyzed. The RNA-sequencing (RNA-seq) data, including both raw counts and fragments per kilobase of transcripts per million mapped reads (FPKM) values of 156 TGCT samples and single-nucleotide variant (SNV) data in the MuTect2 workflow of 144 TGCT samples, were downloaded from TCGA database (12). The FPKM value was converted to transcripts per million (TPM) value (25). Only samples with both RNA-seq data and SNV data were included in downstream analysis, which resulted in a cohort of 128 samples. Moreover, the raw RNA-seq counts of 361 normal control testis samples were obtained from GTEx database (13). We confirmed that there were no duplicate samples in our analysis. In summary, a total of 128 TGCT samples including seminoma and nonseminoma and 361 normal testis samples were used in the subsequent analyses. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments.

GEO data source

The gene expression data of TGCT were profiled by the Illumina HumanHT-12 WG-DASL V4.0 R2 expression beadchip (platform: GPL14951; Illumina, San Diego, CA, USA), which included 30 nonseminoma and 30 seminoma samples derived from the GSE99420 (19) dataset in the GEO (11) database. Expression profiles of 34 TGCT patients via the Affymetrix Human Genome U133A Array (platform: GPL96; Affymetrix, Santa Clara, CA, USA) and the Affymetrix Human Genome U133B Array (platform: GPL97), along with the detailed clinical characteristics in GSE10783, were also downloaded from GEO. Genes in GPL14951 were annotated via the Bioconductor package “illuminaHumanv4.db” (26), and the “hgu133a.db” (27), and “hgu133b.db” (28) annotation packages were applied to annotate genes in GPL96 and GPL97, respectively. The expression level of a gene with multiple probe IDs was considered to be the mean value.

Genomic analysis of TGCT samples

The Bioconductor package “maftools” (29) was used to summarize and visualize the somatic mutation profiles including variant classification (Missense_Mutation, Nonsense_Mutation, Frame_Shift_Del, Frame_Shift_Ins, In_Frame_Del, In_Frame_Ins, Splice_Site, Nonstop_Mutation, Translation_Start_Site), variant type [single-nucleotide polymorphism (SNP), insertion (INS), and deletion (DEL)], and substitution class (T>G, T>A, T>C, C>T, C>G, and C>A) in TGCT samples. A waterfall plot was used to depict the top 20 mutated genes across all TGCT samples. TMB was defined as the total numbers of the mutation errors including base substitutions, INSs, and DELs per megabyte bases. In our study, the length of exons (38 million) was regarded as the captured gene size. The median TMB value was applied to determine TMB-high and TMB-low groups.

Identification of TMB-related genes

DESeq2 (14) was used to identify differentially expressed genes between TMB-high samples and TMB-low samples, with a threshold of |log₂fold change| >1.5 and an adjusted P value <0.05 being applied. A volcano plot generated by ggplot2 (30) was used to depict these differentially expressed genes. Following this, GSEA (31) was used to implement GO (15) and KEGG (17) analysis via the “clusterProfiler” (32) package to examine biological function, with a threshold of minimum gene set size of 15, a maximum gene set size of 500, and an adjusted P value <0.05.

PPI network analysis

TMB-related genes with a significant difference (|log₂fold change| >1 and adjusted P value <0.01) between TGCT and normal control samples were further entered into STRING database, which led to the development of PPI network. The network was further imported into Cytoscape (33) (version 3.8.2) for comprehensive analysis and visualization. Network module analysis and degree analysis were conducted via MCODE (34) and cytoHubba (35), respectively, to determine the key nodes. Key genes were derived by the overlapping of the nodes in modules with density scores >4 and nodes with a degree >5.

Kaplan-Meier analysis and determination of hub genes

Hub genes were defined as key TMB-related genes with prognostic implications. Kaplan-Meier analysis was used to derive hub genes with reference to PFI, which was defined as the interval from diagnosis to first emergence of a new tumor event (36). A new tumor event was defined as either local recurrence confirmed by biopsy, distant metastasis on imaging, or a second primary malignancy diagnosed pathologically. P<0.05 was regarded as being statistically significant. Functional enrichment analysis was further conducted to investigate the BP in GO and functional pathways in KEGG of the hub genes via the “clusterProfiler” (32) package.

Association of hub genes with clinical characteristics

The Wilcoxon rank-sum test was first applied to determine the expression profile of the hub genes for nonseminoma and seminoma cases in TCGA and GSE99420 datasets. The associations with age, T stage, N stage, and M stage of the hub genes were analyzed via the Wilcoxon rank-sum test, Kruskal-Wallis rank-sum test, and Spearman correlation analysis.

Comparison of TGCT to control samples

The Human Protein Atlas was searched to validate the differential expression of hub genes between the TGCT and normal control samples. ROC analysis was used to determine the ability of hub genes and tumor-node-metastasis (TNM) stage to predict the 1-, 3-, 5-, 7-, and 10-year PFI survival probabilities in TGCT. GSE10783 was examined with Univariate Cox analysis to confirm the prognostic implications of the hub genes. P<0.05 was considered to be the cutoff value.

TME analysis

The CIBERSORT tool in R (The R Foundation for Statistical Computing, Vienna, Austria) was used to infer the infiltration levels of 22 immune cell types in TGCT samples. The differential infiltration of immune cells between TMB-high and TMB-low samples and between seminoma and nonseminoma samples was determined via the Wilcoxon rank-sum test. Spearman correlation coefficient analysis was performed to investigate the association of hub genes with immune infiltration genes, with P<0.05 being the cutoff value.

Statistical analysis

All statistical tests were based on a P value <0.05 to indicate statistical significance. The Benjamini-Hochberg method was used to adjust the P value. R software version 4.0.2 was used for the majority of the analyses.

Results

The schematic diagram of this study is provided in Figure 1.

Figure 1 The schematic diagram of the study. GEO, Gene Expression Omnibus; GSEA, gene set enrichment analysis; GTEx, Genotype-Tissue Expression; PPI, protein-protein interaction; RNA-seq, RNA-sequencing; SNV, single-nucleotide variant; STRING, Search Tool for the Retrieval of Interacting Genes/Proteins; TCGA, The Cancer Genome Atlas; TGCT, testicular germ cell tumor; TMB, tumor mutational burden.

Mutation burden profile of TGCT

Somatic mutation analysis conducted via maftools yielded a median of 12 variants in TGCT. Missense mutation was the most frequent variant category, SNPs were more common than were INS or DEL, and C>T was the most common base substitution type, accounting for 1,098 mutations in TGCT (Figure 2A). The top 20 mutated genes among 70 samples, accounting for 54.69% of the total, were further analyzed, which indicated that KIT (16%), KRAS (10%), and TTN (5%) were the most frequently mutated genes in TGCT, with their mutations being more common in seminoma than in nonseminoma (Figure 2B). In addition, the MUC4 gene often contained an in-frame INS mutation (Figure 2B).

Figure 2 Exploration of the genomic characteristics of TGCT. (A) Overview of the mutations in TGCT. (B) Waterfall of the top 20 mutated genes in the TGCT. DEL, deletion; INS, insertion; SNP, single-nucleotide polymorphism; SNV, single-nucleotide variant; TGCT, testicular germ cell tumor.

Determination of TMB-related genes

Differential expression analysis via DESeq2 between the TMB-high and TMB-low samples indicated that 286 genes were upregulated and 91 genes downregulated in the TMB-high group relative to the TMB-low group (Figure 3A). Functional enrichment analysis of the BP in GO revealed that these genes were mainly involved in B-cell receptor signaling pathway, small nuclear RNA (snRNA) metabolic process, and snRNA processing (Figure 3B). KEGG analysis showed that TMB-related genes were most related to antigen processing and presentation, intestinal immune network for immunoglobulin A (IgA) production, and viral protein interaction with cytokine and cytokine receptor (Figure 3B).

Figure 3 Identification of TMB-related genes by DESeq2. (A) Volcano plot for the differentially expressed genes between the TMB-high and TMB-low samples with later-identified hub genes (FOXA2, IRX3, MYH7, and TNNT2) labeled and circled in yellow. (B) GSEA was applied to analyze the BP and functional pathways (KEGG) of the TMB-related genes. BP, biological process; GO, Gene Ontology; GSEA, gene set enrichment analysis; IgA, immunoglobulin A; KEGG, Kyoto Encyclopedia of Genes and Genomes; snRNA, small nuclear RNA; TMB, tumor mutational burden.

Structure of the PPI network and the identification of hub genes

By comparing the expression levels between TGCT and normal control samples, we identified a total of 276 TMB-related genes with a significant difference. These were subsequently imported into the STRING database to build a PPI network comprising 210 nodes and 339 edges (Figure 4). An enrichment P value <1.0e−16 indicated that this network had significantly more interactions than expected. Module analysis via MCODE extracted 8 key clusters, among which 3, including 37 genes, had density scores >4 (Figure 5A). Network analysis via cytoHubba identified 48 genes with a degree >5. By overlapping these genes, we derived 33 key genes for Kaplan-Meier analysis. Ultimately, we found four genes, FOXA2 [hazard ratio (HR) =2.64; 95% confidence interval (CI): 1.26–5.53; P=0.007], IRX3 (HR =2.04; 95% CI: 0.99–4.22; P=0.049), MYH7 (HR =2.14; 95% CI: 1.06–4.34; P=0.03), and TNNT2 (HR =2.05; 95% CI: 1.00–4.22; P=0.046), that were significantly associated with PFI in TGCT (Figure 5B).

Figure 4 The development of the PPI network via the STRING database. Only the TMB-related genes differentially expressed between TGCT and normal control samples were imported into the STRING database. A network including 210 nodes and 339 edges was constructed. PPI, protein-protein interaction; STRING, Search Tool for the Retrieval of Interacting Genes/Proteins; TGCT, testicular germ cell tumor; TMB, tumor mutational burden.

Figure 5 Determination of the hub genes. (A) Module analysis by MCODE extracted a total of eight clusters; three modules demonstrating density scores >4 are depicted. (B) MCODE and cytoHubba identified 33 key genes. Kaplan-Meier analysis further identified four genes significantly associated with PFI in TGCT. CI, confidence interval; HR, hazard ratio; PFI, progression-free interval; TGCT, testicular germ cell tumor.

Associations of hub genes with clinical characteristics

GSEA was first applied to examine the BP of hub genes: FOXA2 and IRX3 had the strongest association with complement activation and B-cell receptor signaling pathway, respectively, while MYH7 and TNNT2 had the strongest association with ribosomal large-subunit biogenesis (Figure 6). KEGG analysis indicated that all the hub genes were associated with graft-versus-host disease (Figure 6). The expression levels of four hub genes in different histologic types were compared via the Wilcoxon rank-sum test, which indicated four hub genes with significant differences between the nonseminoma and seminoma in TCGA dataset (FOXA2: P<0.001; IRX3: P<0.001; MYH7: P=0.02; TNNT2: P<0.001) (Figure 7A); however, only FOXA2 (P=0.008) and TNNT2 (P=0.03) differed significantly in the GSE99420 dataset (Figure 7B). Furthermore, the Wilcoxon rank-sum test, Kruskal-Wallis rank-sum test, and Spearman correlation analysis were performed to clarify the relationship of the hub genes and clinical traits. The Kruskal-Wallis rank-sum test indicated that FOXA2 and MYH7 were associated with tumor stage (P=0.01) and N stage (P=0.03), respectively; meanwhile, the Wilcoxon rank-sum test and Spearman correlation analysis indicated that both IRX3 (age: P=0.007; M stage: P=0.009) and TNNT2 (age: P=0.04; M stage: P=0.006) were correlated with M stage and age (Figure 7C).

Figure 6 Exploration of the function of the four hub genes. GSEA was applied to analyze the BP and functional pathways (KEGG) for FOXA2, IRX3, MYH7, and TNNT2. BP, biological process; GO, Gene Ontology; GSEA, gene set enrichment analysis; KEGG, Kyoto Encyclopedia of Genes and Genomes.

Figure 7 Association of clinical traits with hub genes. (A,B) The expression levels of the four hub genes in different histologic types were compared via the Wilcoxon rank-sum test: (A) TCGA dataset and (B) GSE99420 dataset. (C) The association of clinical characteristics with hub genes was analyzed. Wilcoxon rank-sum test or Kruskal-Wallis rank-sum test and Spearman correlation analysis were used to determine the association with discrete variables and continuous variables, respectively. A P value <0.05 was used as the cutoff value. TCGA, The Cancer Genome Atlas.

Importance of hub genes to TGCT

The differential mRNA expression levels of hub genes between TGCT and normal control samples in TCGA and GTEx data cohorts were all significantly different (FOXA2: P=8.28e−90; IRX3: P=3.30e−124; MYH7: P=9.28e−72; TNNT2: P=1.18e−14) (Figure 8A). Subsequently, the Human Protein Atlas indicated that both the protein levels of FOXA2 and TNNT2 were upregulated in the TGCT samples as compared to normal testis samples, while that of MYH7 was downregulated (Figure 8B). Unfortunately, no immunohistochemical results for IRX3 in either TGCT or normal testis were found in the Human Protein Atlas database. Moreover, ROC analysis indicated that IRX3 and TNNT2 only achieved good prediction for the 1-year PFI [IRX3: area under the curve (AUC) =0.63; TNNT2: AUC =0.63] (Figure 9A) and 3-year PFI (AUC; IRX3: AUC =0.56; TNNT2: AUC =0.57) (Figure 9B) as compared with TNM stage. In predicting the 1-, 3-, 5-, 7-, and 10-year PFI, FOXA2 (1-year: AUC =0.67; 3-year: AUC =0.60; 5-year: AUC =0.60; 7-year: AUC =0.5; 10-year: AUC =0.61) and MYH7 (1-year: AUC =0.56; 3-year: AUC =0.60; 5-year: AUC =0.57; 7-year: AUC =0.65; 10-year: AUC =0.65) (Figure 9A-9E) achieved good performance. Univariate Cox analysis in GSE10783 further indicated the prognostic implications of IRX3 (P=3.90e−04) and FOXA2 (P=9.45e−04) (Figure 9F).

Figure 8 Differential expression of hub genes between TGCT and normal control samples. (A) Differential mRNA level based on TCGA and GTEx data cohorts. (B) Differential protein level based on the Human Protein Atlas. No immunohistochemical results for IRX3 in either the TGCT or normal testis were found in the Human Protein Atlas database. Samples were stained with HE, and representative images were acquired at 20× magnification in the first line and 10× magnification in the second line. FOXA2: (normal) https://www.proteinatlas.org/ENSG00000125798-FOXA2/tissue/testis#img; (tumor) https://www.proteinatlas.org/ENSG00000125798-FOXA2/cancer/testis+cancer#img. MYH7: (normal) https://www.proteinatlas.org/ENSG00000092054-MYH7/tissue/testis#img; (tumor) https://www.proteinatlas.org/ENSG00000092054-MYH7/cancer/testis+cancer#img. TNNT2: (normal) https://www.proteinatlas.org/ENSG00000118194-TNNT2/tissue/testis#img; (tumor) https://www.proteinatlas.org/ENSG00000118194-TNNT2/cancer/testis+cancer#img. FC, fold change; GTEx, Genotype-Tissue Expression; HE, hematoxylin and eosin; mRNA, messenger RNA; TCGA, The Cancer Genome Atlas; TGCT, testicular germ cell tumor.

Figure 9 Prognostic implications of hub genes in TGCT. (A-E) ROC analysis in predicting (A) 1-, (B) 3-, (C) 5-, (D) 7-, and (E) 10-year PFI probability in TGCT with hub genes and TNM stage system as the single predictor. (F) Univariate Cox analysis was applied to GSE10783. A P value <0.05 served as the cutoff value. ***, P<0.001. AUC, area under the curve; PFI, progression-free interval; ROC, receiver operating characteristic; TGCT, testicular germ cell tumor; TNM, tumor-node-metastasis.

Analysis of the TME in TGCT

CIBERSORT indicated that the TME of TGCT was infiltrated by various immune cells (Figure 10A). The Wilcoxon rank-sum test indicated that the infiltration characteristics between nonseminoma and seminoma differed greatly for 12 types of immune cells (Figure 10A). We also compared the immune cell infiltration between the TMB-high and TMB-low groups, which suggested that the abundance of resting memory CD4⁺ T cells, resting natural killer (NK) cells, and resting mast cells was associated with TMB (Figure 10B). In addition, we also examined the association between immune infiltration and hub genes using Spearman correlation analysis, which indicated that all hub genes were positively correlated with the abundance of M2 macrophages in TGCT (Figure 10C,10D). FOXA2, IRX3, MYH7, and TNNT2 were found to be correlated with 7, 8, 5, and 12 types of immune cells, respectively (Figure 10C,10D). The correlation in the subgroups (seminoma and nonseminoma) is presented in Figure 11A,11B, with FOXA2 demonstrating a limited correlation with immune infiltration (M2 macrophages and plasma cells in nonseminoma; plasma cells in seminoma); meanwhile, IRX3 and TNNT2 demonstrated a stronger association with infiltration abundance in nonseminoma than in seminoma, and MYH7 had a significantly positive correlation with immune infiltration in seminoma (Figure 11A,11B).

Figure 10 Infiltration of immune cells in TGCT. CIBERSORT was applied to determine the immune cell abundance in TGCT. (A,B) The differences (A) between nonseminoma and seminoma samples and (B) between TMB-high and TMB-low samples were compared via the Wilcoxon rank-sum test. (A) Immune cell infiltration levels were depicted in a bar plot. (B) Infiltration levels were compared via boxplot. A P value <0.05 served as the cutoff value. *, P<0.05; **, P <0.01; ***, P<0.001. (C) Spearman correlation analysis was conducted in the TGCT group (all TGCT samples including both nonseminoma and seminoma), a nonseminoma group, and a seminoma group. The Spearman correlation matrix was plotted. The blue, white, and red colors indicate the negative correlations [−1], noncorrelations [0], and positive correlations [1], respectively; nonsignificant gene-cell pairs are labeled with a cross. (D) Point plot of the significant gene-cell pairs in the TGCT group (Spearman correlation analysis, P value <0.05). NK, natural killer; TGCT, testicular germ cell tumor; TMB, tumor mutational burden.

Figure 11 Correlation between hub genes and immune infiltration cells. Point plot of the significant gene-cell pairs (Spearman correlation analysis, P value <0.05) for the (A) seminoma group and (B) nonseminoma group.

Discussion

TGCT, which is the most common cancer in young males, caused 71.0 thousand incidences and 9.5 thousand mortalities in 2018 (37,38). This study examined the association of TMB on the gene expression profile and immune infiltration in TGCT using GEO, TCGA, and GTEx data.

A total of 377 TMB-related genes were identified via DESeq2. GSEA revealed that these genes were related to immune-related function pathways, such as the B-cell receptor signaling pathways in GEO and with antigen processing and presentation, intestinal immune network for IgA production, and viral protein interaction with cytokines and cytokine receptors in KEGG (Figure 3B). Overall, this suggests that TMB may be correlated with the immune landscape of TGCT. PPI network analysis and Kaplan-Meier analysis identified four hub genes (FOXA2, IRX3, MYH7, and TNNT2) for TGCT, after which functional enrichment analysis indicated that all hub genes (FOXA2, IRX3, MYH7, and TNNT2) were highly correlated with immune processes (allograft rejection, graft-versus-host disease, B-cell receptor signaling pathway, complement activation, and classical pathway and humoral immune response mediated by circulating immunoglobulin) (Figure 6). ROC analysis identified IRX3 and TNNT2 as predictors of early PFI (1- and 3-year survival probability), while MYH7 was found to be more suitable for later PFI prediction (7- and 10-year survival probability); finally, FOXA2 performed well in predicting all stages of PFI (1-year: AUC =0.67; 3-year: AUC =0.60; 5-year: AUC =0.60; 7-year: AUC =0.5; 10-year: AUC =0.61) as compared with the TNM staging system (Figure 9A-9E). Univariable Cox analysis of GSE10783 further indicated the significant prognostic ability of IRX3 and FOXA2 but not of TNNT2 and MYH7, which may partly be attributed to the relatively small number of samples.

Among the four hub genes, FOXA2 is known to be transcription factor of the forkhead family plays and to play a key role in the development of tissues (39,40). FOXA2, despite mainly functioning as a tumor-suppressor gene in cancer, has been reported to contribute to the metastasis of liver, lung, breast cancer, as well as oral cancer and glioma (41-45); moreover, it has been associated with the prognosis of oral squamous cell carcinoma, intrahepatic cholangiocarcinoma, lung cancer, and breast cancer (46-49). FOXA2 was also found to interact with the MAPK signaling pathway to mediate the progression of numerous tumor types (48,50,51). Unfortunately, no study regarding FOXA2’s role in TGCT has been conducted. Although FOXA2 is mainly downregulated in cancers for which it acts as a tumor-suppressor. In our study, FOXA2 was similarly upregulated in TGCT as indicated by both mRNA and protein levels (Figure 8A,8B) and served as a risk factor in TCGA (Figure 5B) and GSE10783 (Figure 9F) datasets. Therefore, additional studies should be conducted on the mechanism underlying FOXA2’s role in different cancers.

IRX3 belongs to the Iroquois homeobox gene family and is associated with embryonic development and tumor progression. The knockdown of IRX3 was reported to regulate the differentiation of acute myeloid leukemia (52); in addition, IRX3 has been implicated in the emergence of glioblastoma, hepatocellular, and breast cancer (53-55). Moreover, studies in immunopharmacology have found that IRX3 and IRX2 can induce an increase in the number of T cells in both animals and humans (56-58). This suggests that IRX3 exerts immune-related activity, which is in line with our findings (Figure 6).

Few studies on MYH7 and TNNT2 have been performed. MYH7 was reported to be highly upregulated in patients with esophageal carcinoma and multidrug resistance (59) and to be important in patients with prostate cancer and the PTEN mutation (60). A large-scale, whole-exome sequencing study found that both MYH7 and TNNT2 are pathogenic variants in certain cancer (61). However, no studies on MYH7 or TNNT2 in the context of TGCT have been conducted previously. Therefore, further in vivo and in vitro studies should be performed to clarify the function of MYH7 and TNNT2 in TGCT.

GSEA indicated that TMB and TMB-related hub genes (FOXA2, IRX3, MYH7, and TNNT2) were all highly associated with immune-related function. Therefore, we further examined the association of TMB with the infiltration of immune cells in TGCT. We found that resting memory CD4⁺ T cells, resting NK cells, and resting mast cells were significantly associated with TMB. Among these three immune cell types, CD4⁺ T cells, characterized as class II-restricted and tumor-specific, have been reported to innately infiltrate the TME and exert an anticancer effect with the assistance of CD8⁺ T cells, with the secretion of type 1 cytokines, or with the direct killing of tumor cells (62). NK cells in the TME have been reported to be repressed by the selection/editing of poorly immunogenic tumor cells (63). Moreover, a recent study in bladder cancer has revealed the association of TMB with the abundance of resting NK cells and resting mast cells (8). These findings support the potential importance of memory resting CD4⁺ T cells, resting NK cells, and resting mast cells in tumors.

Our study comprehensively examined the association of TMB with gene expression and immune infiltration in TGCT via analyses of GEO, TCGA, and GTEx cohorts and the Human Protein Atlas. Nevertheless, certain limitations to our study should be acknowledged. First, additional experimental validation of our findings still needs to be completed, and the molecular mechanism underlying the relationship of FOXA2, IRX3, MYH7, and TNNT2 with TGCT remains to be clarified through in vitro and in vivo experiments. In addition, these findings should be validated in other TGCT datasets containing both DNA sequencing and mRNA expression information.

Conclusions

We conducted a comprehensive analysis of TMB’s association with TGCT and identified four hub genes (FOXA2, IRX3, MYH7, and TNNT2) related to the prognosis of TGCT. Our findings contribute to a deeper understanding of the progression of TGCT and present potential biomarkers or drug targets for this malignancy.

Acknowledgments

None.

Footnote

Reporting Checklist: The authors have completed the REMARK reporting checklist. Available at https://tau.amegroups.com/article/view/10.21037/tau-2025-514/rc

Peer Review File: Available at https://tau.amegroups.com/article/view/10.21037/tau-2025-514/prf

Funding: This work was supported by Tianjin Health Science and Technology Project (No. TJWJ2025QN018).

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://tau.amegroups.com/article/view/10.21037/tau-2025-514/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Singla N, Bagrodia A, Baraban E, et al. Testicular Germ Cell Tumors: A Review. JAMA 2025;333:793-803. [Crossref] [PubMed]
2. Belfield J, Findlay-Line C. Testicular Germ Cell Tumours-The Role of Conventional Ultrasound. Cancers (Basel) 2022;14:3882. [Crossref] [PubMed]
3. Islam R, Heyer J, Figura M, et al. T cells in testicular germ cell tumors: new evidence of fundamental contributions by rare subsets. Br J Cancer 2024;130:1893-903. [Crossref] [PubMed]
4. Necchi A, Bratslavsky G, Corona RJ, et al. Genomic Characterization of Testicular Germ Cell Tumors Relapsing After Chemotherapy. Eur Urol Focus 2020;6:122-30. [Crossref] [PubMed]
5. Jardim DL, Goodman A, de Melo Gagliato D, et al. The Challenges of Tumor Mutational Burden as an Immunotherapy Biomarker. Cancer Cell 2021;39:154-73. [Crossref] [PubMed]
6. Strickler JH, Hanks BA, Khasraw M. Tumor Mutational Burden as a Predictor of Immunotherapy Response: Is More Always Better? Clin Cancer Res 2021;27:1236-41. [Crossref] [PubMed]
7. Wang P, Chen Y, Wang C. Beyond Tumor Mutation Burden: Tumor Neoantigen Burden as a Biomarker for Immunotherapy and Other Types of Therapy. Front Oncol 2021;11:672677. [Crossref] [PubMed]
8. Tang X, Qian WL, Yan WF, et al. Radiomic assessment as a method for predicting tumor mutation burden (TMB) of bladder cancer patients: a feasibility study. BMC Cancer 2021;21:823. [Crossref] [PubMed]
9. Yakirevich E, Patel NR. Tumor mutational burden and immune signatures interplay in renal cell carcinoma. Ann Transl Med 2020;8:269. [Crossref] [PubMed]
10. O'Meara TA, Tolaney SM. Tumor mutational burden as a predictor of immunotherapy response in breast cancer. Oncotarget 2021;12:394-400. [Crossref] [PubMed]
11. Clough E, Barrett T, Wilhite SE, et al. NCBI GEO: archive for gene expression and epigenomics data sets: 23-year update. Nucleic Acids Res 2024;52:D138-44. [Crossref] [PubMed]
12. Genomic Data Commons. Data Portal. [Cited 2020 Mar 27]. Available online: https://portal.gdc.cancer.gov/
13. GTEx Portal. [Cited 2020 Mar 27]. Available online: https://www.gtexportal.org/home/
14. Love MI, Huber W, Anders S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol 2014;15:550. [Crossref] [PubMed]
15. Ersoz NS, Bakir-Gungor B, Yousef M. GeNetOntology: identifying affected gene ontology terms via grouping, scoring, and modeling of gene expression data utilizing biological knowledge-based machine learning. Front Genet 2023;14:1139082. [Crossref] [PubMed]
16. Sayols S. rrvgo: a Bioconductor package for interpreting lists of Gene Ontology terms. MicroPubl Biol 2023; [PubMed]
17. Jin Z, Sato Y, Kawashima M, et al. KEGG tools for classification and analysis of viral proteins. Protein Sci 2023;32:e4820. [Crossref] [PubMed]
18. Szklarczyk D, Kirsch R, Koutrouli M, et al. The STRING database in 2023: protein-protein association networks and functional enrichment analyses for any sequenced genome of interest. Nucleic Acids Res 2023;51:D638-46. [Crossref] [PubMed]
19. Gao F, Xu Q, Jiang Y, et al. A novel DNA methylation signature to improve survival prediction of progression-free survival for testicular germ cell tumors. Sci Rep 2023;13:3759. [Crossref] [PubMed]
20. Uhlén M, Fagerberg L, Hallström BM, et al. Proteomics. Tissue-based map of the human proteome. Science 2015;347:1260419. [Crossref] [PubMed]
21. Christopher JA, Stadler C, Martin CE, et al. Subcellular proteomics. Nat Rev Methods Primers 2021;1:32. [Crossref] [PubMed]
22. Korkola JE, Houldsworth J, Feldman DR, et al. Identification and validation of a gene expression signature that predicts outcome in adult men with germ cell tumors. J Clin Oncol 2009;27:5240-7. [Crossref] [PubMed]
23. Chen B, Khodadoust MS, Liu CL, et al. Profiling Tumor Infiltrating Immune Cells with CIBERSORT. Methods Mol Biol 2018;1711:243-59. [Crossref] [PubMed]
24. Li Z, Guo Z, Cheng Y, et al. Robust partial reference-free cell composition estimation from tissue expression. Bioinformatics 2020;36:3431-8. [Crossref] [PubMed]
25. Chisanga D, Liao Y, Shi W. Impact of gene annotation choice on the quantification of RNA-seq data. BMC Bioinformatics 2022;23:107. [Crossref] [PubMed]
26. Dunning M, Lynch A, Eldridge M. Illumina HumanHT12v4 annotation data (chip illuminaHumanv4). Bioconductor. Available online: https://bioconductor.org/packages/release/data/annotation/html/illuminaHumanv4.db.html
27. Carlson M. Affymetrix Affymetrix HG-U133A Array annotation data (chip hgu133a). Bioconductor. Available online: https://bioconductor.org/packages/release/data/annotation/html/hgu133a.db.html
28. Roson-Burgo B, Sanchez-Guijo F, Del Cañizo C, et al. Insights into the human mesenchymal stromal/stem cell identity through integrative transcriptomic profiling. BMC Genomics 2016;17:944. [Crossref] [PubMed]
29. Atzeni R, Massidda M, Pieroni E, et al. A Novel Affordable and Reliable Framework for Accurate Detection and Comprehensive Analysis of Somatic Mutations in Cancer. Int J Mol Sci 2024;25:8044. [Crossref] [PubMed]
30. Wickham H. ggplot2: Elegant Graphics for Data Analysis. New York: Springer-Verlag; 2016.
31. Gene Set Enrichment Analysis (GSEA) User Guide. [Cited 2021 Jan 8]. Available online: https://www.gsea-msigdb.org/gsea/doc/GSEAUserGuideFrame.html
32. Yu G, Wang LG, Han Y, et al. clusterProfiler: an R package for comparing biological themes among gene clusters. OMICS 2012;16:284-7. [Crossref] [PubMed]
33. Doncheva NT, Morris JH, Holze H, et al. Cytoscape stringApp 2.0: Analysis and Visualization of Heterogeneous Biological Networks. J Proteome Res 2023;22:637-46. [Crossref] [PubMed]
34. Burke DF, Bryant P, Barrio-Hernandez I, et al. Towards a structurally resolved human protein interaction network. Nat Struct Mol Biol 2023;30:216-25. [Crossref] [PubMed]
35. Ma Z, Zhong P, Yue P, et al. Identification of immune-related molecular markers in intracranial aneurysm (IA) based on machine learning and cytoscape-cytohubba plug-in. BMC Genom Data 2023;24:20. [Crossref] [PubMed]
36. Pradat Y, Viot J, Yurchenko AA, et al. Integrative Pan-Cancer Genomic and Transcriptomic Analyses of Refractory Metastatic Cancer. Cancer Discov 2023;13:1116-43. [Crossref] [PubMed]
37. Ferlay J, Colombet M, Soerjomataram I, et al. Estimating the global cancer incidence and mortality in 2018: GLOBOCAN sources and methods. Int J Cancer 2019;144:1941-53. [Crossref] [PubMed]
38. Bray F, Ferlay J, Soerjomataram I, et al. Global cancer statistics 2018: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin 2018;68:394-424. [Crossref] [PubMed]
39. Morgan E, Arnold M, Gini A, et al. Global burden of colorectal cancer in 2020 and 2040: incidence and mortality estimates from GLOBOCAN. Gut 2023;72:338-44. [Crossref] [PubMed]
40. Koch S. Regulation of Wnt Signaling by FOX Transcription Factors in Cancer. Cancers (Basel) 2021;13:3446. [Crossref] [PubMed]
41. Orstad G, Fort G, Parnell TJ, et al. FoxA1 and FoxA2 control growth and cellular identity in NKX2-1-positive lung adenocarcinoma. Dev Cell 2022;57:1866-1882.e10. [Crossref] [PubMed]
42. Chen Z, Xiang L, Hu Z, et al. Epigenetically silenced linc00261 contributes to the metastasis of hepatocellular carcinoma via inducing the deficiency of FOXA2 transcription. Am J Cancer Res 2021;11:277-96. [PubMed]
43. Zhang Z, Yang C, Gao W, et al. FOXA2 attenuates the epithelial to mesenchymal transition by regulating the transcription of E-cadherin and ZEB2 in human breast cancer. Cancer Lett 2015;361:240-50. [Crossref] [PubMed]
44. Bow YD, Wang YY, Chen YK, et al. Silencing of FOXA2 decreases E-cadherin expression and is associated with lymph node metastasis in oral cancer. Oral Dis 2020;26:756-65. [Crossref] [PubMed]
45. Ding B, Liang H, Gao M, et al. Forkhead Box A2 (FOXA2) Inhibits Invasion and Tumorigenesis in Glioma Cells. Oncol Res 2017;25:701-8. [Crossref] [PubMed]
46. Shen S, Wang G, Shi Q, et al. Seven-CpG-based prognostic signature coupled with gene expression predicts survival of oral squamous cell carcinoma. Clin Epigenetics 2017;9:88. [Crossref] [PubMed]
47. Yadegar N, Dadashi Z, Shams K, et al. The Prominent Role of miR-942 in Carcinogenesis of Tumors. Adv Biomed Res 2022;11:63. [Crossref] [PubMed]
48. Shen J, Zhou Y, Zhang X, et al. Loss of FoxA2 accelerates neoplastic changes in the intrahepatic bile duct partly via the MAPK signaling pathway. Aging (Albany NY) 2019;11:9280-94. [Crossref] [PubMed]
49. Huang C, Liu J, Xiong B, et al. Expression and prognosis analyses of forkhead box A (FOXA) family in human lung cancer. Gene 2019;685:202-10. [Crossref] [PubMed]
50. Yue J, López JM. Understanding MAPK Signaling Pathways in Apoptosis. Int J Mol Sci 2020;21:2346. [Crossref] [PubMed]
51. Kciuk M, Gielecińska A, Budzinska A, et al. Metastasis and MAPK Pathways. Int J Mol Sci 2022;23:3847. [Crossref] [PubMed]
52. Nagel S. The Role of IRX Homeobox Genes in Hematopoietic Progenitors and Leukemia. Genes (Basel) 2023;14:297. [Crossref] [PubMed]
53. Rong L, Li N, Zhang Z. Emerging therapies for glioblastoma: current state and future directions. J Exp Clin Cancer Res 2022;41:142. [Crossref] [PubMed]
54. Akbari ME, Gholamalizadeh M, Doaei S, et al. FTO Gene Affects Obesity and Breast Cancer Through Similar Mechanisms: A New Insight into the Molecular Therapeutic Targets. Nutr Cancer 2018;70:30-6. [Crossref] [PubMed]
55. Azzam SK, Alsafar H, Sajini AA. FTO m6A Demethylase in Obesity and Cancer: Implications and Underlying Molecular Mechanisms. Int J Mol Sci 2022;23:3800. [Crossref] [PubMed]
56. Naylor PH, Quadrini K, Garaci E, et al. Immunopharmacology of thymosin alpha1 and cytokine synergy. Ann N Y Acad Sci 2007;1112:235-44. [Crossref] [PubMed]
57. Rai R, Thrasher AJ, Cavazza A. Gene Editing for the Treatment of Primary Immunodeficiency Diseases. Hum Gene Ther 2021;32:43-51. [Crossref] [PubMed]
58. Naylor PH, Hadden JW. Preclinical studies with IRX-2 and thymosin alpha1 in combination therapy. Ann N Y Acad Sci 2010;1194:162-8. [Crossref] [PubMed]
59. Tian L, Yang L, Zheng W, et al. RNA sequencing of exosomes revealed differentially expressed long noncoding RNAs in early-stage esophageal squamous cell carcinoma and benign esophagitis. Epigenomics 2020;12:525-41. [Crossref] [PubMed]
60. Sun J, Li S, Wang F, et al. Identification of key pathways and genes in PTEN mutation prostate cancer by bioinformatics analysis. BMC Med Genet 2019;20:191. [Crossref] [PubMed]
61. Horiuchi Y, Matsubayashi H, Kiyozumi Y, et al. Disclosure of secondary findings in exome sequencing of 2480 Japanese cancer patients. Hum Genet 2021;140:321-31. [Crossref] [PubMed]
62. Hadrup S, Donia M, Thor Straten P. Effector CD4 and CD8 T cells and their role in the tumor microenvironment. Cancer Microenviron 2013;6:123-33. [Crossref] [PubMed]
63. Vitale M, Cantoni C, Pietra G, et al. Effect of tumor cells and tumor microenvironment on NK-cell function. Eur J Immunol 2014;44:1582-92. [Crossref] [PubMed]

(English Language Editor: J. Gray)