T cell proliferation-related subtypes, prognosis model and characterization of tumor microenvironment in head and neck squamous cell carcinoma

Abstract

Background: Thirty-three synthetic driver genes of T-cell proliferation have recently been identified through genome-scale screening. However, the tumor microenvironment (TME) cell infiltration, prognosis, and response to immunotherapy mediated by multiple T cell proliferation-related genes (TRGs) in patients with head and neck squamous cell carcinoma (HNSC) remain unclear.

Methods: This study examined the genetic and transcriptional changes in 771 patients with HNSC by analyzing the TRGs from two independent datasets. Two different subtypes were analyzed to investigate their relationship with immune infiltrating cells in the TME and patient prognosis. The study also developed and validated a risk score to predict overall survival (OS). Furthermore, to enhance the clinical utility of the risk score, an accurate nomogram was constructed by combining the characteristics of this study.

Results: The low-risk score observed in this study was associated with high levels of immune checkpoint expression and TME immune activation, indicating a favorable OS outcome. Additionally, various factors related to risk scores were depicted.

Conclusion: Through comprehensive analysis of TRGs in HNSC, our study has revealed the characteristics of the TME and prognosis, providing a basis for further investigation into TRGs and the development of more effective immunotherapy and targeted therapy strategies.

Keywords: Immunotherapy; Nomogram; Prognosis; T cell proliferation-related genes; Tumor microenvironment.

Fig. 1

Mutation, CNV, and expression of TRGs in HNSC (a) Mutation of 33 TRGs in 510 patients with HNSC. (b) CNV variation of TRGs. (c) The CNV alterations of TRGs position on 23 chromosomes. (d) Expression of 33 TRGs in normal and HNSC tissues.

Fig. 2

Identification of two distinct TRG subtypes in 771 samples. (a) TRGs interaction in HNSC. The thickness of the lines connecting TRGs indicates the strength of the association between TRGs. Pink means positive correlation, blue means negative correlation. (b) The consensus matrix heatmap defined two TRG clusters. (c) PCA analysis between the two TRG subtypes. (d) Kaplan-Meier curve showed a significant difference in survival time between the two TRG subtypes (log-rank test, p = 0.005). (e) Differences in expression levels of TRGs between two different TRG subtypes.

Fig. 3

Differences in TME between two TRG subtypes (a–b) GSVA of biological Process and pathways between two TRG subtypes. (c) Abundance of multiple infiltrating immune cell types in two TRG subtypes. (d) Expression levels of immune checkpoints between two TRG subtypes. (e) Differences in TME score between two TRG subtypes.

Fig. 4

Identification of three distinct genetic subtypes based on DEGs between two TRG subtypes (a) GO analysis of DEGs. (b) KEGG analysis of DEGs. (c) Three gene clusters were identified by consensus matrix heat maps. (d) Differences in DEGs expression levels of different gene subtypes. (e) Differences in TRGs expression levels among different gene subtypes.

Fig. 5

Establishment of a prognostic risk score. (a) Alluvial diagram of TRG subtypes, gene subtypes, risk score and prognosis. (b–c) The LASSO analysis and partial likelihood deviance of prognostic DEGs. (d) Relationship between three gene subtypes and risk scores. (e) Relationship between two TRG subtypes and risk scores. (f) Expression levels of TRGs between high-risk and low-risk score groups.

Fig. 6

Comprehensive utility and evaluation of risk scores. (a–c) Ranked dot and scatter plots showed risk scores and prognosis, and heatmaps showed the expression of prognostic risk genes between training, testing, and all set. (d–f) Kaplan Meier analysis showed that OS rate of high-risk patients was significantly lower than that of low-risk patients between training, testing, and all set. (g–i) ROC curves were used to evaluate 1-, 3-, and 5-year survival between training, testing, and all set.

Fig. 7

Establishment and validation of a nomogram. (a–b) Univariate and multivariate regression analysis of HNSC patients. (c) Nomogram for predicting the 1-, 3-, and 5-year OS of HNSC patients. (d) Calibration curves indicate the accuracy and specificity of the nomogram for predicting of 1-, 3-, and 5-year OS.

Fig. 8

Differences in TME and immune checkpoint between the two risk score groups. (a) Relationship between risk score and immune infiltrating cells. (b) Correlation between immune cell and risk genes. (c) Differences in immune checkpoint expression between two different risk score groups.

Fig. 9

Comprehensive analysis of the risk score in HNSC. (a) Correlation between risk score and CSC index. (b–c) The waterfall plot of somatic mutation features established with low- and high-risk group. Each column represented an individual patient. The upper barplot showed TMB, the number on the right indicated the mutation frequency in each gene. The right barplot showed the proportion of each variant type. (d) TMB in different risk score groups. (e) Spearman correlation analysis of the risk score and TMB. (f–i) Chemotherapy sensitivity in different risk score groups.

Fig. 10

Expression difference and function analysis of prognostic TRGs in tumor and normal cells. (a) Expression difference of some signature genes between normal and tumor tissues (TCGA). (b) The KM analysis evaluated the prognostic value of 5 TRGs (TCGA). (c) The relative RNA levels of ULBP2, NKX2-3, CLEC3B, CSF2 and FGD3 in tumor and adjacent normal tissues of HNSC patients were detected by q-PCR. (d) The protein expression of ULBP2 gene in 8 pairs of HNSC tumors and adjacent tissues was detected by Western blot. (e–f) The knockdown efficiency of si-ULBP2 was detected by q-PCR and Western blot in 2 HNSC cell lines. (g) MTT experiments showed that knocking down ULBP2 could inhibit cell viability. (h) Colony formation experiment showed that knocking down ULBP2 could decrease colony-forming abilities.