Aberrant Regulation of RAD51 Promotes Resistance of Neoadjuvant Endocrine Therapy in ER-positive Breast Cancer

Abstract

Breast cancer is one of the most common malignant cancers affecting females. Estrogen receptor (ER)-positive breast cancer is responsive to endocrine therapy. Although current therapies offer favorable prospects for improving survival, the development of resistance remains a severe problem. In this study, we explored the resistance mechanisms of ER-positive breast cancer to neoadjuvant endocrine therapy. Microarray data of GSE87411 contained 109 pairs of samples from Z1031 trial, including untreated samples and post-treated samples with neoadjuvant aromatase inhibitor (AI) therapy. The differentially expressed genes (DEGs) were obtained from two different comparisons: untreated samples versus post-treated samples with AIs, and post-treated samples sensitive versus resistant to AIs. Multiple bioinformatic methods were applied to evaluate biological function, protein-protein network and potential binding between target protein and aromatase inhibitor. Then, regulation of gene expression, DNA methylation and clinicopathological factors of breast cancer were further analyzed with TCGA data. From GSE87411 dataset, 30 overlapped DEGs were identified. Cell division was found to be the main function of overlapped DEGs by functional enrichment and gene ontology (GO) analysis. RAD51 recombinase (RAD51), a key protein of homologous recombination, was detected to interact with BReast CAncer genes 2 (BRCA2). Moreover, according to the docking simulation, RAD51 might potentially bind to AIs. Overexpressed RAD51 was associated with hypermethylation of BRCA2, resistance to AIs and poor overall survival of patients with ER-positive breast cancer. Furthermore, RAD51 was found to be a better indicator than MKI67 for predicting resistance in neoadjuvant setting. The results indicated that methylation of BRCA2 led to incomplete suppression on RAD51, which caused an increased expression of RAD51, subsequently AI-resistance and poor prognosis in ER-positive breast cancer. RAD51 could be a new candidate used as a predicative marker and therapeutic target in neoadjuvant endocrine treatment.

Figure 1

The overlapped DEGs were associated with AI-resistance in ER-positive breast cancer. (A) The DEGs of GSE87411 dataset were shown as cluster heatmaps. The DEGs were generated from T-group (untreated samples versus post-treated samples with AIs) and R-group (post-treated samples sensitive versus resistant to AIs). The red points represented upregulated genes (logFC>1), while the green points represented downregulated genes (logFC<−1). The criteria of P value < 0.05 was considered to be significantly different. DEGs, differentially expressed genes; AI, aromatase inhibitor; FC, fold change. (B) The volcano plots showed DEGs in T-group and R-group. In T-group, 82 genes were discovered to be downregulated after AI-treatment (T-down), while none was identified to be upregulated. In R-group, 48 DEGs were obtained. Among them, 6 downregulated genes (R-down) and 42 upregulated genes (R-up) were identified. The red points represented upregulated genes, while the green points represented downregulated genes. The DEGs were screened out with the criteria of P value < 0.05 and |log fold change (FC)| >1. Down, downregulated genes; Up, upregulated genes. (C) The Venn diagram demonstrated different overlaps of genes among 5 groups: T-group, R-group, T-down, R-up and R-down. 30 overlapped DEGs was found to be significantly downregulated after neoadjuvant AI-treatment (T-down), and highly upregulated in AI-resistant samples compared with AI-sensitive samples (R-up). T-down, downregulated genes of T-group; R-up, upregulated genes of R-group; R-down, downregulated genes of R-group.

Figure 2

GO annotation and functional enrichment of DEGs revealed 25 genes involved in cell division. (A) Based on the analysis of GO annotation and functional enrichment, AI-resistant associated DEGs were divided into three categories including biological processes, cellular component and molecular function. Cell division (GO: 0051301) was the most significant biological process. DEGs, differentially expressed genes; GO, gene ontology. (B) GO biological process analysis of AI-resistant associated DEGs was performed by using Metascape with the criteria of P value < 0.01. The 25 genes of 30 overlapped DEGs significantly involved in cell division. (C) Network plot of the relationships among GO terms. Nodes represented enriched terms colored by its cluster identifier.

Figure 3

The overall survival of patients with breast cancer or ER-positive breast cancer. The overall survival curve of CDC25C, ERCC6L and RAD51 in breast cancer or ER-positive breast cancer. High expression of RAD51, CDC25C or ERCC6L was related to poor overall survival of patients with breast cancer or specified ER-positive breast cancer during the early follow-up period (before 15 years). The P values < 0.05 was considered to be significantly different.

Figure 4

The network pharmacology-based prediction of RAD51 and AIs. (A) The docking scores (pKd/pKi) of docking simulation was performed by SystemsDock and shown as bar graph. The docking capacity of Anastrozole, Letrozole or Exemestane for RAD51 was different. (B) The 2-dimensional protein-ligand interactions were analyzed by using docking simulation.

Figure 5

RAD51 was an indicator of poor survival and AI-resistance in breast cancer. (A) Representative result of RAD51 expression. Highly expressed RAD51 was found in breast cancer tissues compared to normal breast tissues by immunohistochemistry. (B) Highly expressed RAD51 was found in breast cancer tissues (n = 1097) compared to normal breast tissues (n = 114) by analyzing TCGA data. (C) An increasing expression of RAD51, in ascending order, was found in luminal (n = 566), Her2-positive (n = 37) and triple-negative breast cancer tissues (n = 116) compared with normal breast tissues (n = 114). (D) Positive correlation was found between the expression of RAD51 and MKI67. E. Expression of RAD51 and MKI67 in AI-sensitive samples and AI-resistant samples was shown as box plot. Higher expression of RAD51 and MKI67 was detected in AI-resistant samples than in AI-sensitive samples. In AI-resistant samples, RAD51 had higher expression than MKI67.

Figure 6

The methylation of BRCA2 was associated with aberrant expression of RAD51. (A) Expression of BRCA2 was negatively regulated by the methylation of BRCA2 in breast cancer (P < 0.05). Highly expressed RAD51 was detected in breast cancer samples with high-methylated BRCA2 by analyzing the TCGA data (P < 0.05). (B) In ER-positive breast cancer, decreased expression of BRCA2 and increased expression of RAD51 were found in high-methylated BRCA2 samples by analyzing the TCGA data (P < 0.05, P < 0.01, separately).