Novel cuproptosis-related long non-coding RNA signature to predict prognosis in prostate carcinoma

• Published in: BMC cancer Vol. 23; no. 1; pp. 105 - 17
• Main Authors: Cheng, Xiaofeng, Zeng, Zhenhao, Yang, Heng, Chen, Yujun, Liu, Yifu, Zhou, Xiaochen, Zhang, Cheng, Wang, Gongxian
• Format: Journal Article
• Language: English
• Published: London BioMed Central 30.01.2023; BioMed Central Ltd; Springer Nature B.V; BMC
• Subjects: Algorithms; Analysis; Androgens; Angiogenesis; Apoptosis; Biomedical and Life Sciences; Biomedicine; Cancer; Cancer Research; Cancer therapies; Carcinoma; Care and treatment; Cell death; Copper; Cuproptosis; Data mining; Diagnosis; Gene expression; Gene mutations; Genetic aspects; Genomics; Health aspects; Health Promotion and Disease Prevention; Humans; Immune checkpoint; Immunotherapy; Learning algorithms; LncRNA; Machine learning; Male; Medical prognosis; Medicine/Public Health; Metastases; Metastasis; Methods; Microenvironments; Microsatellite Instability; Mutation; Nomograms; Non-coding RNA; Oncology; Patients; Point mutation; Prognosis; Prognostic signature; Prostate; Prostate cancer; Prostate carcinoma; Prostatic Neoplasms - genetics; Risk groups; RNA; RNA, Long Noncoding - genetics; Surgical Oncology; Tumor Microenvironment; Tumors; China; Taiwan; LncRNA; Prognostic signature; Prostate carcinoma; Cuproptosis; Machine learning
• ISSN: 1471-2407; 1471-2407
• DOI: 10.1186/s12885-023-10584-0

Background Cuproptosis, an emerging form of programmed cell death, has recently been identified. However, the association between cuproptosis-related long non-coding RNA (lncRNA) signature and the prognosis in prostate carcinoma remains elusive. This study aims to develop the novel cuproptosis-related lncRNA signature in prostate cancer and explore its latent molecular function. Methods RNA-seq data and clinical information were downloaded from the TCGA datasets. Then, cuproptosis-related gene was identified from the previous literature and further applied to screen the cuproptosis-related differentially expressed lncRNAs. Patients were randomly assigned to the training cohort or the validation cohort with a 1:1 ratio. Subsequently, the machine learning algorithms (Lasso and stepwise Cox (direction = both)) were used to construct a novel prognostic signature in the training cohorts, which was validated by the validation and the entire TCGA cohorts. The nomogram base on the lncRNA signature and several clinicopathological traits were constructed to predict the prognosis. Functional enrichment and immune analysis were performed to evaluate its potential mechanism. Furthermore, differences in the landscape of gene mutation, tumour mutational burden (TMB), microsatellite instability (MSI), drug sensitivity between both risk groups were also assessed to explicit their relationships. Results The cuproptosis-related lncRNA signature was constructed based on the differentially expressed cuproptosis-related lncRNAs, including AC005790.1, AC011472.4, AC099791.2, AC144450.1, LIPE-AS1, and STPG3-AS1. Kaplan–Meier survival and ROC curves demonstrate that the prognosis signature as an independent risk indicator had excellent potential to predict the prognosis in prostate cancer. The signature was closely associated with age, T stage, N stage, and the Gleason score. Immune analysis shows that the high-risk group was in an immunosuppressive microenvironment. Additionally, the significant difference in landscape of gene mutation, tumour mutational burden, microsatellite instability, and drug sensitivity between both risk groups was observed. Conclusions A novel cuproptosis-related lncRNA signature was constructed using machine learning algorithms to predict the prognosis of prostate cancer. It was closely with associated with several common clinical traits, immune cell infiltration, immune-related functions, immune checkpoints, gene mutation, TMB, MSI, and the drug sensitivity, which may be useful to improve the clinical outcome.