Identification of a centrosome-related prognostic signature for breast cancer

• Published in: Frontiers in oncology Vol. 13; p. 1138049
• Main Authors: Fang, Zhou, Gao, Zhi-Jie, Yu, Xin, Sun, Sheng-Rong, Yao, Feng
• Format: Journal Article
• Language: English
• Published: Switzerland Frontiers Media S.A 22.03.2023
• Subjects: breast cancer; centrosome; CRSS; immune; machine algorithm; Oncology; prognostic model; CRSS; breast cancer; immune; centrosome; prognostic model; machine algorithm
• ISSN: 2234-943X; 2234-943X
• DOI: 10.3389/fonc.2023.1138049

As the major microtubule organizing center in animal cells, the centrosome is implicated with human breast tumor in multiple ways, such as promotion of tumor cell immune evasion. Here, we aimed to detect the expression of centrosome-related genes (CRGs) in normal and malignant breast tissues, and construct a novel centrosome-related prognostic model to discover new biomarkers and screen drugs for breast cancer. We collected CRGs from the public databases and literature. The differentially expressed CRGs between normal and malignant breast tissues were identified by the DESeq2. Univariate Cox and LASSO regression analyses were conducted to screen candidate prognostic CRGs and develop a centrosome-related signature (CRS) to score breast cancer patients. We further manipulated and visualized data from TCGA, GEO, IMvigor210, TCIA and TIMER to explore the correlation between CRS and patient outcomes, clinical manifestations, mutational landscapes, tumor immune microenvironments, and responses to diverse therapies. Single cell analyses were performed to investigate the difference of immune cell landscape between high- and low-risk group patients. In addition, we constructed a nomogram to guide clinicians in precise treatment. A total of 726 CRGs were collected from the public databases and literature. , , were screened as the prognostic genes in breast cancer. Next, we constructed a centrosome-related prognostic signature and validated its efficacy based on the genes for predicting the survival of breast cancer patients. The high-risk group patients had poor prognoses, the area under the ROC curve for 1-, 3-, and 5-year overall survival (OS) was 0.77, 0.67, and 0.65, respectively. The predictive capacity of CRS was validated by other datasets from GEO dataset. In addition, high-risk group patients exhibited elevated level of mutational landscapes and decreased level of immune infiltration, especially T and B lymphocytes. In terms of treatment responses, patients in the high-risk group were found to be resistant to immunotherapy but sensitive to chemotherapy. Moreover, we screened a series of candidate anticancer drugs with high sensitivity in the high-risk group. Our work exploited a centrosome-related prognostic signature and developed a predictive nomogram capable of accurately predicting breast cancer OS. The above discoveries provide deeper insights into the vital roles of the centrosome and contribute to the development of personalized treatment for breast cancer.