Abstract 3916: Unbiased phospho-proteomic profiling of mouse breast cancer models with DIA mass spectrometry refines CanPath prototype, a platform for predictive cancer pathway modeling

Background CanPathPro is designed to build and validate a combined experimental and systems biology platform, which will be used in testing cancer signaling hypotheses. It combines highly defined mouse and organotypic experimental systems, high-dimensional data including next generation sequencing and quantitative proteomics, and computational models for data integration, visualization and modelling. Mouse cancer models are characterized by quantitative transcriptome, quantitative mass spectrometry including phospho-proteome, histopathology and histochemistry. Phospho-proteomic data has been obtained using data independent acquisition (DIA) and used to refine signaling models in the mouse cell lines. Currently, the model integrates modules of the signal transduction pathways Egfr/Erbb2, Fgfr, insulin, Akt, Mtor, Myc, Ras, Hippo, Met, Tgfbr, Il6, integrin, Wnt, apoptosis and cell cycle regulation integrating in total about 380 different genes as well as related proteins and phospho-proteins Methods Two mouse mammary gland model cell lines, (Cdh1-fl/AKT1[E17K] and Cdh1-fl + PTEN-fl) were each treated with DMSO, a Pik3ca inhibitor (Wortmannin), or an Akt inhibitor (MK-2206). Cells were lysed and proteins were denatured, followed by reduction, alkylation and digestion with trypsin. The resulting peptides were desalted and enriched for phosphopeptides with TiO2 beads and cleaned up for mass spectrometry. A phosphopeptide library was generated from pooled phosphoenriched samples using LC-MS/MS shotgun measurements and included 22,893 phosphosites from 3,549 protein groups. DIA data was acquired on a Q Exactive HF mass spectrometer with a gradient length of 60 - 120 minutes on a C18 Easy LC 1200 nano-liquid chromatography system. The DIA data was extracted and processed with Spectronaut 11 (Biognosys) for analysis. Results In the AKT1[E17K] samples, 13,396 peptides (21,862 phospho-peptides) were quantified in the DIA runs and 12,297 peptides (19,928 phospho-peptides) were quantified in PTEN-fl samples. Under treatment with MK2206 and Wortmanin, 859 phosphopeptides from 548 protein groups were significantly changed across all comparisons in the AKT1[E17K] samples and with the same treatments 2,276 phosphopeptides from 976 proteins were significantly changed in the PTEN-fl cells. Based on this data 11 functionally relevant new phosphosites have been added to the model including ones on: EIF4B, FOX03, MAP2K4, PAK1, RAF1, and ULK1. Conclusions Phospho-proteomic profi