Identification of potential immunotherapy biomarkers for breast cancer by bioinformatics analysis

Abstract

Breast cancer is a serious malignancy with a high incidence worldwide and a tendency to relapse. We used integrated bioinformatics analysis to identify potential biomarkers in breast carcinoma in the present study. Microarray data, 127breast tumor samples and 23 non-tumor samples, received from the Gene Expression Omnibus (GEO) dataset; 121 differentially expressed genes (DEGs) were selected. Functional analysis using DAVID revealed that these DEGs were highly gathered in endodermal cell differentiation and proteinaceous extracellular matrix. Five bioactive compounds (prostaglandin J2, tanespimycin, semustine, 5182598, and flunarizine) were identified using Connectivity Map. We used Cytoscape software and STRING dataset to structure a protein-protein interaction (PPI) network. The expression of CD24, MMP1, SDC1, and SPP1 was much higher in breast carcinoma tissue than in Para cancerous tissues analyzed by Gene Expression Profiling Interactive Analysis (GEPIA) and ONCOMINE. Overexpression ofCD24, MMP1, SDC1, and SPP1 indicated the poor prognosis in breast carcinoma patients analyzed by Kaplan-Meier (KM) Plotter. Immunohistochemistry microarray was used to further confirm that protein expression of CD24, MMP1, SDC1, and SPP1 was much higher in tumor sections than in Para cancerous tissues. Hub genes expression at the protein level was correlated tothe breast cancer subtype and grade. Furthermore, immunity analysis showed that CD24, MMP1, SDC1, and SPP1 were potentially associated with five immune cell types infiltration (CD8+ T cells, CD4+ T cells, neutrophils, macrophages,and dendritic cells) by TIMER. Thus, this study indicates potential biomarkers that could have applications in the development of immune therapy for breast cancer. However, further studies are required for verifying these results in vivo and vitro.

Keywords: Breast cancer; bioinformatics analysis; biomarkers; immunotherapy; prognosis.

Figure 1. DEGs in breast cancer and non-cancer tissues

(A) Volcano plot, represented DEGs in breast cancer tissues and non-tumor samples in GSE26910 dataset. (B)Volcano plot, represented DEGs in breast cancer tissues and non-tumor tissues in GSE42568 dataset. Red dots, indicate genes highly induced in breast cancer; green dots indicate genes greatly reduced in breast cancer, blue dots indicate non-DEGs. (C)Venn diagram represented the downregulated overlapping DEGs from GSE26910 and GSE42568 datasets. (D) Venn diagram represented the upregulated overlapping DEGs from GSE26910 and GSE42568 datasets.

Figure 2. PPI network and module analysis

(A) PPI network. (B) Top module cluster analyzed by MCODE. (C) Top two module clusters analyzed by MCODE.

Figure 3. Survival analysis for hub genes

(A) 15 hub genes expressed much higher in breast cancer tissues than normal breast tissues by GEPIA. (B)9 hub genes was correlated with OS in breast cancer patients. (C) MMP1, CD24, SDC1, and SPP1 was expression correlated with RFS in breast cancer patients. Abbreviation: OS, overall survival; RFS, relapse-free survival. *,P<0.01.

Figure 4. OS curves for MMP1, CD24, SDC1, and SPP1 expression in breast cancer subtypes

(A) Luminal A; (B) luminal B; (C) triple-negative; (D) HER2+.

Figure 5. GSEA was applied to identify enriched biological processes for the four key genes (MMP1, CD24, SDC1 and SPP1) with highly expressed samples

Figure 6. Correlations of expression of four key genes (MMP1, CD24, SDC1, and SPP1) with immune cell infiltration

Figure 7. Evaluation of MMP1, CD24, SDC1, and SPP1 expression in breast cancer tissue and normal tissue

(A) MMP1, CD24, SDC1, and SPP1 expression was up-regulated in breast cancer tissue compared with normal tissue in the Oncomine dataset. (B) MMP1, CD24, SDC1, and SPP1 protein expression was up-regulated in breast cancer tissue compared with normal tissue according to tissue microarray analysis.