Integrative Network-Based Transcriptomic Analysis Identifies Niclosamide as a Candidate Repositioned Drug for Breast...

Breast cancer (BC) is a highly heterogeneous malignancy, and current treatments often suffer from toxicity, limited selectivity, and high cost. This study aimed to integrate transcriptome-level data, multi-layered network analysis, and drug repositioning strategies to identify candidate diagnostic and prognostic biomarkers for BC and propose potential repositioned drug candidates.

Differentially expressed genes (DEGs) were identified from the GSE42568 dataset (|log₂FC| > 1 and p < 0.05). Functional enrichment analyses were conducted using GO and KEGG. Three biological interaction layers - protein-protein interactions, transcription factors, and miRNA-mRNA interactions were constructed, and hub nodes were identified using topological metrics. Kaplan-Meier analyses assessed survival associations. PCA evaluated sample separation across datasets in GSE42568, GSE113865, and GSE22820. Drug repositioning was performed using L1000CDS², and in vitro validation of a top drug candidate (niclosamide) and an exploratory comparative compound (amitriptyline) was performed using MCF-7 cells, including viability and combination assays.

A total of 4266 DEGs were identified. Network analyses revealed 37 hub signatures, 11 of which-ESR1, RECQL4, FOS, BCL2, CXCL8, TRIM25, EGR1, CDH1, KRAS, PTGS2, and IL6 were associated with survival outcomes. PCA demonstrated clear separation between healthy and BC samples. Drug repositioning identified eight candidates, with niclosamide as the top hit. In vitro assays showed marked reduction in cell viability at 5 µM niclosamide and 25 µM amitriptyline after 24 h treatment. No significant additive effect was observed in the combination treatment.

This integrative approach revealed candidate BC-specific biomarkers and identified niclosamide as a potential repositioned therapeutic. These findings remain exploratory and do not provide definitive clinical evidence. Further validation across additional models and clinical settings is required.