Genome-wide mutation profiling in TETs
We acquired the mutation data of a total of 123 TET patients from the TCGA database. The mutation data were downloaded and visualized using the “maftools” package in R software. Waterfall plot showed higher frequency of gene mutations in patients with TET, such as GTF2I (33%), HRAS (8%), TTN (5%), MUC16 (3%), TP53 (3%) and so on (Fig. 1A, H). In gene mutation correlation diagrams, most genes were independent and a few genes were synergistic. HRAS and GTF2I had a high synergistic correlation (p < 0.001) (Fig. 1B). Missense mutations were the most common type of mutation in patients with TETs (Fig. 1C), SNP occurred more frequently than insertion or deletion (Fig. 1D), and C > T was the predominant mutation type detected (Fig. 1E). In addition, the number of mutated bases per sample was shown in Fig. 1F. In Fig. 1G, the mutation types were shown in different colors in box diagram.
Figure 1. Landscape of mutation profiles in TET samples. A, Mutation information of each gene in each sample was shown in the waterfall plot. The annotation of mutation types were shown at the bottle with various colors and the number of mutation burden was listed in the bar chart above the legend. B, The relationship between mutated genes C-E, Based on statistical calculations of different types of mutations, where missense mutations accounted for the majority, SNP occurred more frequently than deletion or insertion, and C > T was the most common type of SNV. F-G, Illustration of tumor mutation burden in per samples. H, The top 10 mutated genes in TET.
TMB was associated with prognosis
In order to explore the relationship between TMB and the prognosis of patients with TETs, we downloaded the prognostic information of the patients and plotted a Kaplan-Meier curve. The results indicated that a low TMB was associated with a better clinical outcome of patients with TETs (Fig. 2).
Figure 2. Kaplan– Meier curve indicated that a low TMB was associated with a better prognosis.
Genetic changes associated with TMB and functional pathway analysis
To study the DEGs associated with TMB in TET patients, we divided the patients with TET into high TMB and low TMB groups. The heatmap showed TOP 40 DEGs in two TMB groups (Fig. 3A). The GO functional analysis revealed that, these mutant genes mainly enriched in plasma membrane signaling receptor complex, cell-cell junction, cell junction assembly and actin binding (Fig. 3B). In KEGG pathway analysis, PI3K-Akt signaling pathway, cytokine-cytokine receptor interaction, human papillomavirus infection, Rap1 signaling pathway and focal adhesion were top 5 signaling enriched (Fig. 3C). In addition, the GSEA results suggested that patients in high-TMB group tend to be more associated with tumor-related signaling pathways, including focal adhesions, ErbB signaling, ECM-receptor interaction and TGF-β signaling pathway. (Fig. 3D).
Figure 3 Genetic changes associated with TMB and functional pathway analysis
A, The top 40 selected differentially expressed genes to be exhibited in heatmap graph with |log (FC) > 1| and FDR < 0.05. B-C, GO and KEGG pathway analysis for mutated genes. D, GSEA results showed the top TMB-related signaling axis, including focal adhesions, ErbB Signaling, ECM-receptor interaction and TGF-β signaling pathway. GO, Gene Ontology; KEGG, Kyoto Encyclopedia of Genes and Genomes; TMB, tumor mutation burden.
Identification and Kaplan-Meier survival analysis of hub TMB-related immune genes
We screened 97 differential immune genes from high TMB and low TMB groups, and then utilized univariate cox analysis to screened 13 survival related genes among these 97 differentially expressed immune genes (Table 1). Then, the multivariate cox regression analysis was performed to further select 3 hub TMB-related immune genes that are highly associated with prognosis. Higher expression levels of HCK correlated positively with poor prognosis, while lower levels of CD1B and CD1E correlated with a worse prognosis (Figure,4A-C). In addition, we utilized the 3 genes to establish TMB-related signature (TMBRS) model. The AUC of ROC was 0.729, indicating that the high predictive accuracy for our identified TMBRS model (Fig. 4D). The patients were divided into two groups according to the median value of risk score, and patients in high-risk group had poor prognosis (Fig. 4D).
Table 1
13 prognostic mutation related genes by univariate Cox regression analysis
Gene | HR | HR.95L | HR.95H | CoxPvalue |
CD1B | 0.992676 | 0.986658 | 0.99873 | 0.017808 |
CD1C | 0.97868 | 0.960149 | 0.99757 | 0.027146 |
CD1E | 0.992612 | 0.986518 | 0.998743 | 0.018253 |
HCK | 1.005474 | 1.001646 | 1.009316 | 0.005033 |
CCR9 | 0.878058 | 0.772374 | 0.998202 | 0.04687 |
RORC | 0.871029 | 0.77387 | 0.980386 | 0.022124 |
SH2D1A | 0.977006 | 0.95721 | 0.99721 | 0.025918 |
GRAP2 | 0.936357 | 0.889272 | 0.985936 | 0.012489 |
TRAJ61 | 0.313034 | 0.121074 | 0.809341 | 0.016556 |
TRAV18 | 0.247726 | 0.072607 | 0.84521 | 0.025845 |
TRBV14 | 0.858547 | 0.737851 | 0.998986 | 0.048485 |
TRBV18 | 0.893928 | 0.801581 | 0.996915 | 0.043851 |
TRBV19 | 0.941754 | 0.888898 | 0.997754 | 0.041723 |
Figure 4 Identification and Kaplan-Meier survival analysis of hub TMB-related immune genes.
A-C, Kaplan-Meier analysis with log-rank test for identified hub TMB-related immune genes. D-E, Construction and assessment of TMBRS for TETs (the AUC of ROC curve was 0.729), where patients with higher TMBRS conferred poor survival outcomes.
Associations of 3 hub TMB-related immune genes expression with immune cells infiltration
Then multivariate Cox is utilized to further identify hub independent risk characteristics and establish a risk model. We further screened three hub survival-related genes from 13 prognostic mutation related genes to construct our prognostic model. The 3 hub TMB-related immune genes expression between the high- and the low-TMB groups were shown in volcano plot in Fig. 5A. Among these, CD1B and CD1E were lowly expressed in high-TMB group, while HCK was highly expressed. More importantly, we further assessed the underlying relationships of the expression of these hub genes with immune infiltrates in TET microenvironment.The expression of HCK was negatively correlated with the infiltration of CD8+ T cell and CD4+ T cell, while the levels of CD1B and CD1E were positively correlated with immune infiltrates. These immune infiltrates include B cell, CD8+ T cell, CD4+ T cell, macrophages and dendritic cells (Fig. 5B-D). These suggest that, low-expression CD1B, CD1E and high-expression HCK in tumors can inhibit immuno-immersion.
Figure 5 Associations of 3 hub TMB-related immune genes expression with immune cells infiltration. A, Expression analysis of 3 hub TMB-related immune genes with |log (Foldchange) > 1| and FDR < 0.05; B-D, The relationship of 3 hub TMB-related immune genes with immune infiltrates.
Tumor infiltrating immune cells (TIICs) in high-and low-TMB groups
To investigate the correlation between TIICs and TMB in TETs, we first used CIBERSORT to calculate infiltration of 22 immune cells in the patients with TETs (Fig. 6A). Besides, the Wilcoxon rank-sum test indicated that native CD4+ T cells, plasma cells, activated memory CD4+ T cell, follicular helper T cells and regulatory T cells were higher infiltrating in low-TMB group, while native B cells, activated NK cells, resting mast cells, activated dendritic cells, M0, M1and M2 macrophages showed higher infiltrating levels in high-TMB group (Fig. 6B).
Figure 6 TIICs in high-and low-TMB groups. A Summary of estimated fractions of 22 immune cell subtypes from the CIBERSORT algorithm. B, TIICs associated with TMB. Red means high TMB and green means low TMB
Association of top mutated genes in TET with immune infiltrates
GTF2I (33%) and HRAS (8%) are top mutated genes, which have synergistic mutation in patients with TET (Fig. 1H). The mutated GTF2I and HRAS can inhibit the infiltration levels of B cells, CD8+ T cells, CD4+ T cell and dendritic cells (Fig. 7A, B). Moreover, mutated HRAS can also inhibit the immune infiltration of macrophage (Fig. 7B).These results suggested that mutated GTF2I and HRAS might promote tumor development by inhibiting anti-tumor immune response, resulting in a poor prognosis for patients.
Figure 7 Association of top mutated genes in TET with immune infiltrates. A-B, Mutated GTF2I and HRAS can inhibit immune infiltration levels of B cells, CD8+ T cells, CD4+ T cell and dendritic cells. (P-value Significant Codes: 0 ≤ *** < 0.001≤ ** < 0.01 ≤ * < 0.05 ≤. < 0.1)