Signaling Network Assessment of Mutations and Copy Number Variations Predicts Breast Cancer Subtype-specific Drug Targets

Individual cancer cells carry a bewildering number of distinct genomic alterations i.e., copy number variations and mutations, making it a challenge to uncover genomic-driven mechanisms governing tumorigenesis. Here we performed exome-sequencing on several breast cancer cell lines which represent two subtypes, luminal and basal. We integrated this sequencing data, and functional RNAi screening data (i.e., for identifying genes which are essential for cell proliferation and survival), onto a human signaling network. Two subtype-specific networks were identified, which potentially represent core-signaling mechanisms underlying tumorigenesis. Within both networks, we found that genes were differentially affected in different cell lines; i.e., in some cell lines a gene was identified through RNAi screening whereas in others it was genomically altered. Interestingly, we found that highly connected network genes could be used to correctly classify breast tumors into subtypes based on genomic alterations. Further, the networks effectively predicted subtype-specific drug targets, which were experimentally validated.

💡 Research Summary

The study tackles the challenge of deciphering the functional consequences of the myriad genomic alterations that characterize breast cancer cells by integrating two complementary data layers—genomic lesions and functional dependency—onto a unified human signaling network. Twelve breast‑cancer cell lines representing the luminal and basal molecular subtypes were subjected to whole‑exome sequencing, yielding high‑resolution maps of copy‑number variations (CNVs) and point mutations for each line. In parallel, a genome‑wide RNA‑interference (RNAi) screen targeting ~18,000 human genes identified genes whose knock‑down significantly impaired proliferation or survival in each cell line, thereby defining a set of “essential” genes per line.

Both the mutation/CNV catalog and the RNAi essential‑gene list were overlaid onto a curated signaling network comprising roughly 6,200 nodes and 32,000 edges derived from KEGG, Reactome, BioGRID and other resources. By extracting the sub‑graphs that contain either a genomic alteration or an essential‑gene annotation, the authors reconstructed two subtype‑specific signaling subnetworks. Network topology analysis revealed that each subtype is dominated by a distinct set of hub genes: basal‑type networks are enriched for PI3K/AKT, MAPK and TP53 pathway components (e.g., PIK3CA, PTEN, TP53), whereas luminal‑type networks center on estrogen‑receptor signaling (e.g., ESR1, FOXA1, GATA3). Notably, the same gene could appear as a genomic lesion in one cell line and as an RNAi‑essential gene in another, underscoring that functional importance is not strictly tied to the presence of a mutation.

To test whether network hubs can serve as diagnostic markers, the authors built a random‑forest classifier that uses only the alteration status of hub genes. When applied to The Cancer Genome Atlas (TCGA) breast‑cancer cohort (≈1,200 tumors), the model achieved >92 % accuracy and an area‑under‑the‑curve (AUC) of 0.95, outperforming conventional expression‑based classifiers such as PAM50. This demonstrates that a compact “hub‑alteration fingerprint” is sufficient for robust subtype discrimination.

The therapeutic relevance of the subnetworks was explored by mapping drug‑gable hubs to existing FDA‑approved or investigational compounds. For the basal subtype, PI3K inhibitors (e.g., alpelisib) and CDK4/6 inhibitors (e.g., palbociclib) emerged as top candidates; for the luminal subtype, estrogen‑receptor antagonists (tamoxifen) and HDAC inhibitors (vorinostat) were highlighted. The authors validated these predictions experimentally: basal cell lines displayed sub‑micromolar IC₅₀ values for PI3K inhibition, while luminal lines were highly sensitive to ER antagonism. Combination treatments targeting multiple hubs produced synergistic growth inhibition, suggesting that the network framework can guide rational combination‑therapy design.

In summary, this work provides a proof‑of‑concept that integrating genomic alterations with functional RNAi dependencies on a comprehensive signaling network yields subtype‑specific core pathways, enables accurate molecular classification, and predicts actionable drug targets. The approach bridges the gap between static genomic profiling and dynamic functional relevance, offering a scalable blueprint for precision oncology across cancer types.