Title : Backbone regulators of non-small cell lung cancer: A hierarchical network dissection reveals translational control as a central oncogenic axis
Abstract:
Introduction: Non-small cell lung cancer (NSCLC) remains a biologically complex and therapeutically challenging disease, accounting for nearly 85% of lung cancer diagnoses worldwide. After decades of studying tumour systems biology, it has become increasingly evident that oncogenesis is not merely the consequence of isolated mutations, but of coordinated regulatory architectures. We therefore sought to define the systems-level backbone of NSCLC by reconstructing patient-derived transcriptomic networks and identifying deeply conserved regulatory nodes that sustain tumour homeostasis across molecular subtypes.
Materials and Methods: RNA-sequencing datasets from 512 primary NSCLC tumours (including lung adenocarcinoma and lung squamous cell carcinoma) and 108 matched normal lung tissues were analysed across TCGA and two independent validation cohorts. Differentially expressed genes (FDR <0.01, |log2FC| ≥1.5) were integrated with curated protein–protein interaction datasets to construct a high-stringency disease interactome. Network architecture was interrogated using eigenvector and betweenness centrality, k-core decomposition, and multilevel modularity refinement. Regulators were prioritised based on persistence across k-core layers, cross-cohort reproducibility, and survival concordance.
Results and Discussion: The reconstructed NSCLC interactome comprised 2,742 nodes interconnected by 18,936 edges and displayed a robust hierarchical scale-free topology with a dominant 3-core structure. While numerous high-degree hubs were observed, only 14 regulators demonstrated structural persistence across all hierarchical layers and maintained prognostic stability in independent cohorts. Contrary to conventional expectations centred solely on canonical oncogenes, the conserved backbone was enriched for regulators of translational fidelity (EIF4G1, RPL23A), proteotoxic stress buffering (HSPD1, DNAJB11), metabolic coupling (IDH1, PHGDH), and RNA maturation (DDX21, NOP58). Elevated expression of this composite backbone signature was associated with significantly reduced overall survival (median OS 32.4 vs 58.7 months, p<0.001). Functional enrichment revealed coordinated reinforcement of ribosome biogenesis, mitochondrial metabolism, redox adaptation, and epithelial–mesenchymal plasticity. Notably, network perturbation simulations demonstrated that removal of intermediate k-core integrators produced greater global instability than elimination of several classical high-degree hubs, underscoring the underestimated importance of architectural stabilisers in tumour persistence.
Conclusion: This systems-level dissection of NSCLC reveals a conserved regulatory backbone rooted in translational control, metabolic adaptability, and proteostasis rather than exclusively in canonical driver mutations. These findings argue for therapeutic strategies that disrupt tumour network integrity itself—particularly its stabilising cores—thereby shifting the paradigm from gene-centric targeting to architecture-informed intervention.
