A multiomic investigation of lung adenocarcinoma molecular subtypes

Background: Lung adenocarcinoma—an aggressive and life-threatening malignancy—is a type of non–small-cell lung cancer. Despite medical advancements, the prognosis of lung adenocarcinoma remains unfavorable, likely because of its heterogeneous nature. Furthermore, few subtype-specific treatments are available for lung adenocarcinoma. This study was conducted to explore the molecular subtypes of lung adenocarcinoma. Methods: We performed a joint analysis of transcriptome and proteome data from East Asian patients with lung adenocarcinoma (nonsmokers, 86.5%). Results: Four novel subtypes were identified based on distinct molecular characteristics: subtypes I, II, III, and IV. In patients with subtype I lung adenocarcinoma, eukaryotic translation initiation factor 4 gamma 1 activates cell proliferation; inhibiting this factor suppresses tumor growth, and reducing its level induces autophagy. Subtype II is characterized by Kristen rat sarcoma viral oncogene homolog-activating oncogenesis; the onset age of this subtype is the lowest among all subtypes. Subtype III manifests as an advanced disease at diagnosis; it is characterized by a core serum response-related oncogenic signature, which indicates poor overall survival in Western patients with lung cancer. Subtype IV is more common in men than in women; it has astroglial characteristics. A Connectivity Map analysis revealed that the oncogenic expression patterns corresponding to subtypes I, II, III, and IV can be reversed by the inhibitors of Inhibitor of κB (IκB) kinase (eg, withaferin A), mammalian target of rapamycin (eg, everolimus), Src proto-oncogene (Src) (eg, saracatinib), and Transforming Growth Factor (TGF)-β/Smad (eg, LY-364947), respectively. Conclusion: This study introduced an innovative multiomics data analysis pipeline. Using this approach, we successfully identified four molecular subtypes of lung adenocarcinoma and their candidate therapeutic agents. The newly identified subtypes can be combined with the current biomarkers to generate a comprehensive roadmap for treatment decision-making.