Abstract 6662: Deciphering molecular network complexity of oncogenic PIK3CA through proximity labeling

PIK3CA mutations are major drivers for oncogenesis. Mutant PIK3CA continuously activates the PI3K pathway, transmitting oncogenic signals within transformed cells. Despite decades of efforts to target oncogenic PIK3CA for cancer therapy, the precise molecular and functional machinery remains incomplete due to the molecular complexity of the PI3K pathway. Mapping proximity proteins in the oncogenic PI3K pathway and identifying mutant signals will signify a deep understanding of oncogenesis and a broad range of therapeutic applications. In this study, we show that PIK3CA mutant cells establish a unique molecular program that network with distinct proteins, differing from cells with wild-type (WT) PIK3CA. We have developed a bioengineering approach to decode protein-protein interactions of oncogenic PIK3CA and identify key molecular partners, potentially unveiling therapeutic vulnerabilities in cancer cells. First, we applied the proximity labeling technique, based on the fusion of a protein ligase to a protein of interest (POI), featuring proximity-dependent biotinylation of interacting proteins. The biotin ligase BioID2 was utilized to integrate WT PIK3CA or gain-of-function hotspot mutations in E545K or H1047R residues. To optimize the labeling protocol and biotinylation conditions, the BioID2 plasmids were fused with each POI, incorporating multiple combinations of protein linkers and tags, including myc, and red fluorescence protein. A total of 13 plasmids were constructed in mammalian expression vectors, including pcDNA3.1 vector for transient expression, and pHAGE-puro vector for generating stable cell lines. Next, we used fluorescence microscope to screen RFP signals for transfection efficiency and employed immunofluorescence to verify the situations of protein expression and biotinylation with specific antibodies and fluorescence-conjugated streptavidin. Finally, streptavidin pulldown and LC-MS were utilized to identify the proximity proteins. Our technical approach confirmed a high level of reliability and consistency in LC-MS data between POI replicates. We identified approximately 1,000 proximity proteins, of which 436 were shared between samples labeled with WT and mutant PIK3CA. We found that both H1047R and E545K mutants demonstrated altered interactions with the protein GATAP2A. Furthermore, PIK3CA H1047R uniquely interacted with proteins S100A7, FLG2, LGALS7, and PPIB, while E545R specifically mapped proximal proteins SEPTIN7, SERBP1, SNW1, and RNPS1. To further validate our methods, we engineered two cancer cell lines (DHSA-1426 and COSB) to induce PIK3CA H1047R mutations using CRISPR/Cas9. Our ongoing work involves labeling proximity proteins with mutant PIK3CA in these malignant cells, enabling temporal and spatial molecular sensing of oncogenic PIK3CA and its functional interactome in living cancer cells. Citation Format: Rong Li, Donghee Lee, Md Abdullah, Eslenur Nipa, Jong Hyuk Kim. Deciphering molecular network complexity of oncogenic PIK3CA through proximity labeling [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 6662.