Identification of a chromatin regulator signature and potential prognostic ability for adrenocortical carcinoma

Abstract

Objective: Adrenocortical carcinoma (ACC) is a rare malignant tumor. Chromatin regulators (CRs) can drive epigenetic changes, which have been considered as one of the most vital hallmarks of tumors. This study aimed to explore the CR signature for ACC in order to clarify the molecular basis of ACC's pathogenic mechanism and provide novel methods to diagnose and treat ACC clinically. Methods: This study obtained transcriptome sequencing datasets of ACC patients and sequencing data on normal adrenal tissues in TCGA and GTEx databases, respectively. Meanwhile, prognostic genes were selected through Lasso and Cox regression analyses. Using the transcriptome sequencing datasets of ACC patients downloaded from the GEO database to finish validation, we performed Kaplan-Meier (KM) analysis for evaluating the differential survival between low- and high-risk groups. Then, this work constructed the risk model for predicting ACC prognosis. TIMER 2.0 was employed to assess the differences in immune infiltration between the two groups. Furthermore, this work adopted the R package "pRRophetic" for exploring and estimating the sensitivity of patients to different chemotherapeutic agents. Results: A 5-CR model was established to predict ACC survival, and the CR signature was confirmed as a factor in order to independently predict ACC patient prognosis. In addition, a nomogram composed of the risk score and clinical T stage performed well in the prediction of patients' prognosis. Differentially expressed CRs (DECRs) were mostly associated with the cell cycle, base excision repair, colon cancer, gene duplication, homologous recombination, and other signaling pathways for the high-risk group. As for the low-risk group, DECRs were mainly enriched in allograft rejection, drug metabolism of cytochrome P450, metabolism of xenogeneic organisms by cytochrome P450, retinol metabolism, and other signaling pathways. According to TIMER analysis, the immune infiltration degrees of endothelial cells, M2 macrophages, myeloid dendritic cells, CD4⁺ Th1 cells, NKT cells, and M0 macrophages showed significant statistical differences between the high- and low-risk groups, and high infiltration levels of M0 and M2 macrophages were more pronounced in higher T stage (T3 and T4), N stage (N1), and clinical stages (III and IV). In addition, high-risk cases exhibited higher sensitivity to etoposide and doxorubicin. Additionally, low-risk patients had significantly decreased expression of RRM1 compared with high-risk cases, suggesting the better effect of mitotane treatment. Conclusion: This study identified the DECRs, which might be related to ACC genesis and progression. The pathways enriched by these DECRs were screened, and these DECRs were verified with excellent significance for estimating ACC survival. Drug sensitivity analysis also supported the current clinical treatment plan. Moreover, this study will provide reliable ideas and evidence for diagnosing and treating ACC in the clinic.

Keywords: adrenocortical carcinoma; chromatin regulator; diagnosis; prognosis; treatment.

FIGURE 1

20 DECRs selected from the intersection of DEGs and CRs (A). Univariate Cox regression of eight DECRs that showed a prognostic value (B). KM survival analysis of eight DECRs that showed a prognostic value (C–J).

FIGURE 2

Lasso Cox regression of DECRs (A). Expression of five selected DECRs in different clinical features groups (B). KM survival analysis of low- and high-risk patients in test cohort (C). ROC analysis of clinical stage, clinical T stage, clinical N stage and riskscore in test cohort (D). KM survival analysis of low- and high-risk patients in validation cohort (E). ROC analysis of stage and riskscore in validation cohort (F).

FIGURE 3

Riskscore in different gender, clinical stage, clinical T stage and clinical N stage.

FIGURE 4

A nomogram constructed by clinical T stage and riskscore in test cohort (A). Calibration curve predicting 1-, 3-, and 5-year patient prognosis (B–D).

FIGURE 5

A nomogram constructed by riskscore in validation cohort (A). Calibration curve predicting 1-, 3-, and 5-year patient prognosis (B–D).

FIGURE 6

GO_BP, CC and MF analysis of 20 DECRs.

FIGURE 7

KEGG analysis of low- and high-risk groups (A,B). GSEA analysis of BP, CC and MF in low- and high-risk groups (C–H).

FIGURE 8

Heat map of CR signature with immune infiltration by TIMER analysis.

FIGURE 9

RRM1 expression in low- and high-risk groups (A). Etoposide, cisplatin and doxorubicin senstivity in low- and high-risk groups (B–D).