Abstract 7445: Deep learning prioritizes cancer mutations that alter protein nucleocytoplasmic shuttling to drive tumorigenesis

Abstract Genetic variants can affect protein function by driving aberrant subcellular localization. However, due to limitations in existing computational tools, comprehensive analysis of how mutations promote tumor progression by influencing nuclear localization is currently lacking. Here, for the first time, we systematically characterized potential shuttling-attacking mutations (SAMs) across cancers. The deep learning model Predicting Shuttling-Attacking Mutations (pSAM) was the leading tool constructed for the ab initio decoding of protein sequence determinants essential for subcellular shuttling, including classical nuclear localization signals and unconventional sequences such as fragments embedded in DNA/RNA-binding domains. Leveraging cancer mutations across 11 cancer types, we found that SAMs constitute about 6.7% of the mutations and enrich functional genetic variations and critical genes in cancer; for example, SAMs in transcription factors prevent them from entering the nucleus and subsequently cause abnormal expression of downstream targets. We experimentally validated a dozen SAMs, among which R14M in PTEN, P255L in CHFR, R674C in NFE2L1 and R144C in KAT8 were identified to disrupt the nuclear localization through interfering their interactions with importins. Further studies confirmed that the altered nucleocytoplasmic shuttling triggered by SAMs in PTEN and CHFR rewired the downstream signaling pathways and eliminated their function of suppressing cancer progression. Thus, this study greatly contributes to the understanding of the molecular details of nucleocytoplasmic shuttling and reveals the variants mediating aberrant protein subcellular localization and dysfunction in cancers and diseases. Citation Format: Ze-Xian Liu, Kai Yu, Yongqiang Zheng, Jin-Fei Lin. Deep learning prioritizes cancer mutations that alter protein nucleocytoplasmic shuttling to drive tumorigenesis [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2025; Part 1 (Regular Abstracts); 2025 Apr 25-30; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2025;85(8_Suppl_1):Abstract nr 7445.