Development of an endoplasmic reticulum stress-related signature with potential implications in prognosis and immunotherapy in head and neck squamous cell carcinoma

Abstract

Background: Head and neck squamous cell carcinoma (HNSCC) is a multisite malignancy that responds well to immunotherapy. Despite the initial enthusiasm, the clinical benefits of immunotherapy in HNSCC patients are overall limited. Endoplasmic reticulum stress (ERS) has been indicated to play a key role in the process of anti-tumor immune response mediation. However, ERS-related biomarkers which can accurately predict prognosis and immunotherapy response in HNSCC are still lacking.

Methods and results: In this study, we identify and validate an ERS-related signature comprises of six genes (ASNS, EXOSC6, BAK1, TPP1, EXOSC8, and TATDN2) that can predict the prognosis of HNSCC patients. GSEA analysis indicates that the ERS-related signature is significantly correlated with tumor immunity in HNSCC. Moreover, the infiltration of naive B cells and CD8 + T cells are significantly diminished in patients with high-risk scores compared to those with low-risk scores, while macrophages and activated mast cells are remarkably enhanced. Furthermore, the ERS-related signature also displays a tremendous potential for predicting immunotherapy response in HNSCC.

Conclusions: Our study identifies an ERS-related signature that can predict the prognosis of HNSCC patients and highlights its potential value as a predictive biomarker of immunotherapy response, potentially enabling more precise and personalized immunotherapy response and paving the way for further investigation of the prognostic and therapeutic potentials of ERS.

Keywords: Biomarker; Endoplasmic reticulum stress; Head and neck squamous cell carcinoma; Immunotherapy; Prognosis.

Fig. 1

The patient flow diagram

Fig. 2

Activity of ERS is significantly different between tumor and normal samples in HNSCC. (A) Significant difference of ERS scores between normal and tumor samples in HNSCC. P-values are based on the Mann–Whitney test. (B) Heat map showing the expression levels of the 28 ERS-related genes differently expressed between normal and tumor samples in HNSCC from the TCGA.

Fig. 3

Construction of the ERS-related prognostic signature for OS of HNSCC in the TCGA dataset. (A) 19 genes related to OS of HNSCC were obtained by utilizing univariate Cox regression. (B) Kaplan–Meier curves of OS of low- and high- risk groups stratified by the risk score in HNSCC. (C) The distributions of risk scores and patient survival status for HNSCC. (D) ROC curve for the 5-year OS prediction by the ERS-related signature. (E) Kaplan–Meier curves of RFS, DSS and PFS of low- and high- risk groups stratified by the risk score in HNSCC, respectively. (F) ROC curves for the 5-year RFS, DSS and PFS prediction by the ERS-related signature, respectively

Fig. 4

The relationship of the ERS-related signature with clinical characteristics of HNSCC patients. (A) The strip chart of risk score and clinical characteristics for patients with HNSCC in TCGA datatset. (B) Pie charts showing the Chi-squared test of clinicopathologic factors for low- and high- risk groups in HNSCC samples from TCGA dataset

Fig. 5

Validation of the ERS-related prognostic signature in GEO dataset. (A) Kaplan–Meier curves of OS of low- and high- risk groups stratified by the risk scores in HNSCC in GEO dataset. (C) The distributions of risk scores and patient survival status for HNSCC in TCGA dataset. (D) ROC curve for the 5-year OS prediction by the ERS-related signature in GEO dataset

Fig. 6

The ERS-related signature is an independent prognostic factor in HNSCC patients. Univariate (A) and multivariate (B) Cox regression of prognosis factor for OS of HNSCC patients. (C) Kaplan–Meier analysis of OS for HNSCC patients stratified by sex, age, ethnicity, T, stage, grade, cancer status and lymphnode neck dissection status

Fig. 7

The ERS-related signature shows strongly association with immune surveillance against cancer cells in HNSCC. (A) The biological functions of the ERS-related signature in HNSCC using GSEA analysis based on the curated gene sets c5.go.bp.v7.4.symbols. (B) The difference of immune cell infiltration between low- and high- risk groups based on the ERS-related signature in HNSCC using CIBERSORT algorithm. P-values are based on the Mann–Whitney test. * P < 0.05, **P < 0.01, ***P < 0.001

Fig. 8

The ERS-related signature is a potential predictor for immunotherapy response in HNSCC. (A) The correlation of the ERS-related signature with the expression levels of check-point genes in HNSC from the TCGA. R: Spearman’s correlation coefficient. (B) The expression levels of check-point genes among HNSCC samples grouped by risk score in TCGA dataset. Figure S1. Different expression levels of ASNS (A) and EXOSC8 (B) between normal and tumor samples in HNSCC