A hypoxia risk score for prognosis prediction and tumor microenvironment in adrenocortical carcinoma

Abstract

Background: Adrenocortical carcinoma (ACC) is a rare malignant endocrine tumor derived from the adrenal cortex. Because of its highly aggressive nature, the prognosis of patients with adrenocortical carcinoma is not impressive. Hypoxia exists in the vast majority of solid tumors and contributes to invasion, metastasis, and drug resistance. This study aimed to reveal the role of hypoxia in Adrenocortical carcinoma and develop a hypoxia risk score (HRS) for Adrenocortical carcinoma prognostic prediction. Methods: Hypoxia-related genes were obtained from the Molecular Signatures Database. The training cohorts of patients with adrenocortical carcinoma were downloaded from The Cancer Genome Atlas, while another three validation cohorts with comprehensive survival data were collected from the Gene Expression Omnibus. In addition, we constructed a hypoxia classifier using a random survival forest model. Moreover, we explored the relationship between the hypoxia risk score and immunophenotype in adrenocortical carcinoma to evaluate the efficacy of immune check inhibitors (ICI) therapy and prognosis of patients. Results: HRS and tumor stage were identified as independent prognostic factors. HRS was negatively correlated with immune cycle activity, immune cell infiltration, and the T cell inflammatory score. Therefore, we considered the low hypoxia risk score group as the inflammatory immunophenotype, whereas the high HRS group was a non-inflammatory immunophenotype. In addition, the HRS was negatively related to the expression of common immune checkpoint molecules such as PD-L1, CD200, CTLA-4, and TIGIT, suggesting that patients with a lower hypoxia risk score respond better to immunotherapy. Conclusion: We developed and validated a novel hypoxia risk score to predict the immunophenotype and response of patients with adrenocortical carcinoma to immune check inhibitors therapy. These findings not only provide fresh prognostic indicators for adrenocortical carcinoma but also offer several promising treatment targets for this disease.

Keywords: Adrenocortical carcinoma; hypoxia; immunotherapy; risk score; tumor microenvironment.

FIGURE 1

Hypoxia clusters correlated with immune phenotypes. (A) The TCGA-ACC cohort was divided into two distinct cluster when k = 2. (B) Survival analysis between cluster 1 and cluster 2. Cluster 1, red. Cluster 2, yellow. (C) Comparison of immune cycle activity involved in seven steps between cluster 1 and cluster 2. Cluster 1, red. Cluster 2, blue. (D) Comparison of various immune cell infiltration score between cluster 1 and cluster 2. Cluster 1, red. Cluster 2, blue. (E) The enrichment score of IFN-γ pathway in cluster 1 and cluster 2. |NES| > 1 and p < 0.01 were considered as statistically significant.

FIGURE 2

Hypoxia related DEGs and functional analysis. (A) A heatmap depicts the between DEGs cluster1 and cluster 2. Higher expression DEGs with are displayed in red, and lower expression are displayed in blue. (B) A volcano plot depicts the DEGs between cluster 1 and cluster 2. DEGs with log₂(FC)≥1 were shown in red while the genes with log₂ (FC)≤ -1 were shown in blue, and the genes with indiscriminate expression were shown in gray. (C–F) GO and KEGG analysis of DEGs between cluster 1 and 2. (C) The pathway in GO functional enrichment comparison between cluster 1 and cluster 2. (D) The pathway in KEGG functional enrichment comparison between cluster 1 and cluster 2.

FIGURE 3

Development of HRS and the role in clinical prognosis prediction. (A) LASSO coefficient profiles of 143 hypoxia-related prognostic DEGs (B) Ten-fold cross-validation for tuning parameter selection in the LASSO model. The two dotted vertical lines are drawn at the optimal value using the minimum criteria. Optimal hypoxia genes with the best discriminative capability (6 in number) were selected for generating the HRS (C) Forest plot of hazard rations for six optimal hypoxia-related prognostic genes. (D) Survival analysis between the two different risk score group. Risk score high is shown red and risk score low group is shown yellow. (E) The predictive accuracy of the HRS for survival. (F) Results of univariate Cox analysis by integrating the HRS and clinicopathological characters. (G) The nomogram used to predict the 12-month,36-month, 60-month overall survival.

FIGURE 4

External validation of the hypoxia risk score. (A, B) Validation of the hypoxia risk score in GSE76019 (C, D) Validation of the hypoxia risk score in GSE19750 (E, F) Validation of the hypoxia risk score in GSE 33371.

FIGURE 5

Differences in immunological characteristic between HRS groups. (A, B) Spearman correlation analysis of HRS with activity of cancer immunity cycle and immune cell in TME analyzed by ssGSEA. The positive correlation is shown in solid line. The negative correlation is shown in dotted line. The association strength was represented by the thickness of the lines. The different colors of the lines represent different p-values. (C) The associations between the HRS and the several anti-tumor immune cell in six different algorithms. (D) The correlations between HRS and immune checkpoint. (E) The correlations between HRS and T cell inflamed score. (F) A heatmap was drawn to depict the differences in cytotoxic effector molecule between high-risk score group and low-risk group.