Machine-learning-assisted multi-omics profiling identifies CALB1 as a subtype-associated regulator of PI3K/AKT signaling and malignant progression in non-small cell lung cancer
Year: 2026
Type: article
Abstract: Background: Non-small cell lung cancer (NSCLC) is molecularly heterogeneous, and computationally nominated biomarkers often lack sufficient functional validation. CALB1 is a calcium-binding protein, but its clinical relevance, cellular source, and signaling role in NSCLC remain unclear. This study aimed to identify subtype-associated regulators in NSCLC through a machine-learning-assisted multi-omics framework and to characterize the functional and signaling relevance of CALB1 in malignant progression.Methods: NSCLC transcriptomic cohorts were used to define metabolic-epigenetic remodeling (MER) subtypes. Machine-learning-assisted Cox modeling was then applied to identify subtype-associated prognostic regulators. CALB1 was selected for further validation using public cohorts, in-house NSCLC specimens, quantitative real-time polymerase chain reaction (qRT-PCR), western blotting, immunohistochemistry, single-cell RNA sequencing, proteomic and phosphoproteomic profiling, and functional assays. CALB1 was silenced in A549 and PC9 cells to assess its effects on PI3K/AKT activation, proliferation, migration, invasion, and epithelial-mesenchymal transition (EMT)-associated phenotypes. Alpelisib was used as a PI3Kα inhibitor to examine the pathway-level involvement of PI3K/AKT signaling in CALB1-associated phenotypes. In vivo relevance and alpelisib responsiveness were further evaluated using a four-group subcutaneous A549 xenograft model comprising sh-Control + vehicle, sh-CALB1 + vehicle, sh-Control + alpelisib, and sh-CALB1 + alpelisib groups.Results: Three MER subtypes were identified, with MER2 showing the poorest prognosis and distinct immune-metabolic features. CALB1 was retained in the prognostic model and positively correlated with model-derived risk scores. CALB1 was upregulated in NSCLC tissues, associated with advanced clinicopathological features and unfavorable prognosis, and mainly enriched in malignant epithelial cells. CALB1-high tumors showed enrichment of proliferative, metabolic, DNA repair, and PI3K/AKT-related programs. CALB1 knockdown induced coordinated proteomic and phosphoproteomic changes involving proliferation, migration, extracellular matrix organization, secretory trafficking, and PI3K/AKT signaling. Functionally, CALB1 silencing reduced p-PI3K and p-AKT levels and suppressed proliferation, migration, invasion, and EMT-associated phenotypes. PI3Kα inhibition with alpelisib attenuated CALB1 overexpression-associated malignant phenotypes in vitro. In vivo, stable CALB1 knockdown suppressed xenograft growth and enhanced the antitumor response to alpelisib, with the sh-CALB1 + alpelisib group showing the greatest reductions in tumor volume and tumor weight.Conclusions: This study identifies CALB1 as a subtype-associated regulator in NSCLC through a machine-learning-assisted multi-omics framework. CALB1 may promote malignant progression partly through PI3K/AKT-related signaling. Stable CALB1 knockdown increased the in vivo antitumor response to alpelisib, supporting CALB1 as a potential modulator of PI3Kα inhibitor sensitivity.
Institutions: Shanghai Jiao Tong University, Shanghai Chest Hospital
Language: English
Subfield: Pulmonary and Respiratory Medicine
Field: Medicine
Domain: Health Sciences
Open Access status: diamond
OpenAlex ID: W7223885460