Exploration of prognostic biomarkers in head and neck squamous cell carcinoma microenvironment from TCGA database

Abstract

Background: Immune checkpoint blockade (ICB) therapies have redefined human cancer treatment, including for head and neck squamous cell carcinoma (HNSCC). However, clinical responses to various immune checkpoint inhibitors are often accompanied by immune-related adverse events (irAEs). Therefore, it is crucial to obtain a comprehensive understanding of the association between different immune tumor microenvironments (TMEs) and the immunotherapeutic response.

Methods: The research data were obtained from The Cancer Genome Atlas (TCGA) database. We applied RNA-seq genomic data from tumor biopsies to assess the immune TME in HNSCC. As the TME is a heterogeneous system that is highly associated with HNSCC progression and clinical outcome, we relied on the Estimation of Stromal and Immune cells in Malignant Tumor tissues using Expression data (ESTIMATE) algorithm to calculate immune and stromal scores that were evaluated based on the immune or stromal components in the TME. Then, the Tumor Immune Dysfunction and Exclusion algorithm (TIDE) was used to predict the benefits of ICB to each patient. Finally, we identified specific prognostic tumor-infiltrating immune cells (TIICs) by quantifying the cellular composition of the immune response in HNSCC and its association to survival outcome, using the CIBERSORT algorithm.

Results: Utilizing the HNSCC cohort of the TCGA database and TIDE and ESTIMATE algorithm-derived immune scores, we obtained a list of microenvironment-associated lncRNAs that predicted different clinical outcomes in HNSCC patients. We validated these correlations in a different HNSCC cohort available from the TCGA database and provided insight into the prediction of response to ICB therapies in HNSCC.

Conclusions: This study confirmed that CD8⁺ T cells were significantly associated with better survival in HNSCC and verified that the top five significantly mutated genes (SMGs) in the TCGA HNSCC cohort were TP53, TTN, FAT1, CDKN2A, and MUC16. A high level of CD8⁺ T cells and high immune and stroma scores corresponded to a better survival probability in HNSCC.

Keywords: Head and neck squamous cell carcinoma (HNSCC); The Cancer Genome Atlas (TCGA); immune checkpoint blockade (ICB); tumor microenvironment (TME); tumor-infiltrating immune cell (TIIC).

Figure 1

Comparison of lncRNA expression profiles between HNSCC tumor samples and normal samples. (A) Heatmap of differentially expressed lncRNA in normal and tumor tissue. Differentially expressed lncRNA are lncRNA show an adjusted P value ≤0.0001. Genes shown in red are upregulated and genes in blue are downregulated. (B) The significantly upregulated and downregulated lncRNA in HNSCC tumor samples and normal samples are shown in the volcano plot (log₂ fold change >1.0, P<0.01). HNSCC, head and neck squamous cell carcinoma; lncRNA, long noncoding RNA.

Figure 2

LASSO Cox regression model construction and survival analysis. (A) λ selection by 10-fold cross-validation. The shrinkage curve of elastic net in selecting signature lncRNA. Adjusting parameter (lambda) screening in the LASSO regression model, lambda with lowest partial likelihood deviance were selected. Solid vertical lines represent partial likelihood deviance ± standard error. The dotted vertical lines are drawn correspond to the optimal values by minimum criteria and 1- standard error criteria. The value of 0.6457363 was chosen for λ by 10-fold cross-validation with the minimum criteria. (B) Survival curve of low- and high-risk groups stratified by their risk score. LASSO, Least Absolute Shrinkage and Selection Operator; lncRNA, long noncoding RNA.

Figure 3

The selected 22 signature lncRNA and its association with predicted survival probability. (A) Consensus clustering matrix of 508 samples from TCGA dataset for k=3. (B) Survival difference among three groups of patients. Three groups of patients show significant survival difference. lncRNA, long noncoding RNA; TCGA, The Cancer Genome Atlas.

Figure 4

Immune therapy analysis. (A) The expression level of CD8 is significantly related the survival of patients. (B) Expression level of CD8 in different HNSCC stages. HNSCC, head and neck squamous cell carcinoma.

Figure 5

TMB score and mutation status of HNSCC patients. (A) TMB for every patient. (B) Oncoplot for the top 50 mostly mutated genes. TMB, tumor mutational burden; HNSCC, head and neck squamous cell carcinoma.

Figure 6

The relationship between HNSCC tumor immune status and prognosis. (A) Distribution of stromal score among different stage of tumors. (B) Distribution of immune score among different stage of patients. (C) The effects of stromal score on the survival of patients. (D) Prediction of HNSCC patients’ prognosis corresponding with different immune score. HNSCC, head and neck squamous cell carcinoma.

Figure 7

The landscape of immune infiltration in HNSCC. (A) The difference of immune infiltration in HNSCC samples. (B) Heat map of the 22 immune cell fractions in TCGA-HNSCC cohort. The vertical axis depicts the clustering data of samples which were divided into two discrete groups. (C) Survival plot of median of CD8 T cells. (D) Survival plot of median of CD4⁺ memory T cells. (E) Survival plot of median of follicular helper T cells. (F) Survival plot of median of regulatory T cells. The P values are from log-rank tests. HNSCC, head and neck squamous cell carcinoma; TCGA, The Cancer Genome Atlas.