Cuproptosis combines immune landscape providing prognostic biomarker in head and neck squamous carcinoma

Abstract

Head and neck squamous carcinomas (HNSC) are the seventh most common cancer around the world. Treatment options available today have considerable limitations in terms of efficacy. Identifying novel therapeutic targets for HNSC is, therefore, urgently needed. As a novel determined regulated cell death (RCD), Cuproptosis is correlated with the development, treatment response, and prognosis of various cancer. However, the potential role of Cuproptosis-related genes (CRGs) in the tumor microenvironment (TME) of HNSC remains unclear. To figure out whether TME cells and Cuproptosis could better predict prognosis, in this study, we analyzed the expression, mutation status, and other clinical information of 502 HNSC patients by dividing them into four clusters based on their CRGs and TME cell expression. Utilizing the LASSO-Cox method and bootstrap, we established Prognostic Cuproptosis and TME classifier, which were significantly associated with prognosis, pathways, clinical features, and immune cell infiltration in TME of HNSC. To go further, the subgroup Cup low/TMEhigh displayed a better prognosis than any others. Two GEO datasets demonstrated the proposed risk model's clinical applicability. Our GO enrichment analyses proved the conjoint effect of Cuproptosis and TME on tumor angiogenesis, proliferation, and so on. Single-cell analysis and Immunotherapy profile then provided a foundation for determining the molecular mechanisms. It revealed the prognostic risk score positively correlated with T cell activation and natural killer (NK) recruiting. As far as we know, this study is the first time to explore the involvement of CRGs regulation in the TME of HNSC. In a word, it is vital to use these findings to develop new therapeutic strategies.

Keywords: Cuproptosis; Head and neck squamous carcinoma; Prognosis; Therapeutic responses; Tumor microenvironment.

Fig. 1

The construction and validation of Cup-TME model. Independent HNSC patient's cohort with complete clinical information, TCGA-HNSC, which were used to establish the prognostic TME score and Cuproptosis score, respectively. Bootstrap-multicox regression analyses of 22 TME cells and 398 Cuproptosis-related genes were performed. In the end, we got 6 TME cells. As well, 14 Cuproptosis-related genes were used for the establishment of Cuproptosis score. A Cup-TME classifier which integrated the TME and Cuproptosis scores classified all patients into three different subgroups: Cup^low/TME^high, mixed, and Cup^high/TME^low. Based on the Cup-TME classifier, the differences in prognosis, pathway enrichment analysis, clinical subtype features, tumor mutational burden, and tumor molecular characteristics were investigated in several patient subgroups. Another independent cohort (GSE42743 and GSE31056) were used to further validate the classifier's performance.

Fig. 2

The construction of the Cuproptosis Scoring (Cup-score) model and TME-score model. (A). Kaplan-Meier survival analysis of the two strata. Patients with high Cup-score HNSC had a worse prognosis. The yellow line represents the high Cup-score, and the blue line represents the low Cup-score. P < 0.001 (B). Kaplan-Meier survival analysis of patients with low TME-score had a worse prognosis. The yellow line represents the high TME-score, and the blue line represents the low TME-score. P < 0.001 (C). Violin plot shows that Cup_score is differentially expressed in different cell types. (D), (E). Bubble plots show ligand-receptor pairs difference. P < 0.01 was considered significant. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

Fig. 3

Identification of modules associated with the clinical traits of head and neck cancer. (A). Dendrogram of all differentially expressed genes clustered based on the measurement of dissimilarity (1-TOM). The color band shows the results obtained from the automatic single-block analysis. (B). Heatmap of the correlation between the module eigengenes and subgroups of HNSC. We selected the MEblue, MEyellow and MEbrown block for subsequent analysis. (C), (D), (E). Functional enrichment analysis for the hub genes in the royal blue, yellow and brown module by Metascape analysis, respectively. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

Fig. 4

The construction and validation of the Cup-TME classifier. (A). Kaplan-Meier overall survival curves of TCGA-HNSC training cohorts (n = 502) divided into three different subgroups based on the Cup-TME classifier. Log-rank test, P < 0.001. (B). Prognostic efficiency ability evaluation of this model by introducing the receiver operating characteristic (ROC) curve. (C). Time-dependent ROC curves at separately three year, five years, and seven years. (D), (E). The forest plot shows the results of the univariate Cox regression analysis and multivariate cox analysis, respectively. (F). FGSEA reveals main regulated pathways of the differences in the Cup-TME classifier.

Fig. 5

Comparison of immune-related markers and therapy responses prediction based on Cup-TME classifier. (A). The 23 normalized immune activity scores obtained by TIP. (B) HLA gene-set is differentially expressed between these subtypes. (C). The immune checkpoint gene-set is differentially expressed between these subtypes. (The Cup^low/TME^high, mixed and Cup^high/TME^low subgroups are represented as purple, blue and yellow, respectively.) (D). The different percentages of immunotherapy responder among Cup^low/TME^high and Cup^high/TME^low based on Cup-TME classifier. (E). Comparison of Cup-scores between immunotherapy responder and nonresponder. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

Fig. 6

Survival analysis of the subgroup of clinical factors and functional analysis. (A), (B), (C), (D). Kaplan–Meier survival curves in subgroup analyses according to sex, age, grade, stage. (A. sex (P < 0.005); B. grade (P < 0.005); C. stage (P≤0.054); D. age (P < 0.005)) (E), (F), (G), (H). Functional analysis between Cup^low/TME^high and responder of patients under immunotherapy. (E. down in Cup^low/TME^high; F. down in responders; G. up in Cup^low/TME^high; H. up in responders. Each small polygon corresponds to a single KEGG pathway, and the size correlates with the ratio between the subgroups. Proteomaps (https://bionic-vis.biologie.uni-greifswald.de/)) (I). Kaplan-Meier overall survival curves of the independent validation GEO cohort (n = 171) stratified into three different subgroups based upon the Cup-TME classifier (Cup^low/TME^high, Mixed, Cup^high/TME^low). Log-rank test, P < 0 0.001.