Defining Functional Hub Genes in Basal-like Breast Cancer Networks

The behaviour of complex biological systems emerges from the coordinated activity of networked molecular components. In this context, gene regulatory networks (aka gene coexpression networks) offer insights into the regulation of gene expression programs. In cancer, aberrant gene expression underlies molecular and clinical features, and identifying key networked transcriptional regulators may enable targeted therapeutic interventions. However, computationally inferred regulatory nodes have so far hardly been experimentally validated. Here we combined gene expression network analysis with gene perturbation experiments to test whether computationally identified hub genes act as upstream regulators of their coexpression modules in breast cancer. To better capture the context-dependent nature of gene regulation and minimize confounding effects from inter-subtype heterogeneity, we also constructed subtype-specific networks. Using the METABRIC transcriptomic dataset of primary breast tumours, we identified clinically-informative gene modules in the highly aggressive basal-like subtype. Candidate regulatory hubs were prioritized based on network centrality, and their functional relevance was assessed both in silico and in vitro. CRISPR-mediated knockout of selected hub genes resulted in coordinated down-regulation of module genes and impaired cellular functions, demonstrating causal links between hub gene function, module expression and phenotypic outcome. Moreover, we observed a significant correlation between the transcriptional impact of each knockout and its functional effects—highlighting the biological relevance of coexpression modules and supporting the hypothesis that their structure reflects functional dependencies.