Lipid metabolism-related miRNAs with potential diagnostic roles in prostate cancer

Abstract

Background Prostate cancer (PCa), the second most prevalent solid tumor among men worldwide, has caused greatly increasing mortality in PCa patients. The effects of lipid metabolism on tumor growth have been explored, but the mechanistic details of the association of lipid metabolism disorders with PCa remain largely elusive. Methods The RNA sequencing data of the GSE45604 and The Cancer Genome Atlas-Prostate Adenocarcinoma (TCGA-PRAD) datasets were extracted from the Gene Expression Omnibus (GEO) and UCSC Xena databases, respectively. The Molecular Signatures Database (MSigDB) was utilized to identify lipid metabolism-related genes. The limma R package was used to identify differentially expressed lipid metabolism-related genes (DE-LMRGs) and differentially expressed microRNAs (DEMs). Moreover, least absolute shrinkage and selection operator (LASSO), extreme gradient boosting (XGBoost), and support vector machine-recursive feature elimination (SVM-RFE) were applied to select signature miRNAs and construct a lipid metabolism-related diagnostic model. The expression levels of selected differentially expressed lipid metabolism-related miRNAs (DE-LMRMs) in PCa and benign prostate hyperplasia (BPH) specimens were verified using quantitative real-time polymerase chain reaction (qRT‒PCR). Furthermore, a transcription factor (TF)-miRNA‒mRNA network was constructed. Eventually, Kaplan‒Meier (KM) curves were plotted to illustrate the associations between signature miRNA-related mRNAs and TFs and overall survival (OS) along with biochemical recurrence-free survival (BCR). Results Forty-seven LMRMs were screened based on the correlation analysis of 29 DE-LMRGs and 56 DEMs, in which 27 LMRMs were stably expressed in the GSE45604 dataset. Subsequently, receiver operating characteristic (ROC) curves and machine learning methods were employed to develop a lipid metabolism-related diagnostic signature, which may be of diagnostic value for PCa patients. qRT‒PCR results showed that all seven key DE-LMRMs were differentially expressed between PCa and BPH tissues. Eventually, a TF-miRNA‒mRNA network was constructed. Conclusions These results suggested that 7 key diagnostic miRNAs were closely related to PCa pathological processes and provided new targets for the diagnosis and treatment of PCa. Moreover, CLIC6 and SCNN1A linked to miR-200c-3p had good prognostic potential and provided valuable insights into the pathogenesis of PCa.

Establishing and clinically validating a machine learning model for predicting unplanned reoperation risk in colorectal cancer

“…To efficiently screen feature variables associated with colorectal cancer postoperative unplanned reoperation, we used three different machine learning methods: support vector machine (SVM)[ 16 ] least absolute shrinkage and selection operator (LASSO) regression[ 17 ], and extreme gradient boosting (XGBoost)[ 18 ].…”

Targeting Dysregulated Lipid Metabolism in Cancer with Pharmacological Inhibitors

“…Seven key lipid metabolism-related miRNAs (LMRMs), including miR-148-3p and miR-375-3p, exhibit significant diagnostic potential for prognosis and diagnosis in prostate cancer (PCa). A TF-miRNA-mRNA regulatory network involving SNAI2, GATA3, and target genes like CLIC6 and SCNN1A, is significant in PCa progression and patient prognosis [172]. In HCC, miR-122 promotes lipid accumulation by targeting several genes involved in lipid oxidation, including CPT1A and acyl-CoA oxidase 1 (ACOX1) [173].…”

Diagnostic significance of dysregulated miRNAs in T-cell malignancies and their metabolic roles

“…A recent computational biology study using machine learning has identified seven key miRNAs associated with mRNAs involved in lipid metabolism that are differentially expressed between prostate cancer and its benign counterpart. The differential expression of this lipid metabolism-related signature was also validated by PCR and in publicly available datasets ( 14 ). In addition to forming a molecular signature, miRNA expression profiles can identify distinct metabolic phenotypes within a disease spectrum.…”