MacKenzie, K. L. | bioRxiv Science

bioRxiv2025

Potent synthetic lethality between PLK1 and EYA-family inhibitors in tumours of the central and peripheral nervous system

The Eyes Absent family of protein phosphatases (EYA1-4) are aberrantly expressed and tumour-promoting across many devastating cancers of neurological origin affecting both children and adults. It has recently been demonstrated that EYA1 and EYA4 promote tumour cell survival by increasing the active pool of Polo-like kinase 1 (PLK1) molecules. This discovery provides a rationale for the therapeutic combination of EYA inhibitors with direct, ATP-competitive, PLK1 inhibitors. Here, we demonstrate potent and synergistic effects of EYA and PLK1 inhibition in cancer cell lines that overexpress EYA1 and/or EYA4, including in neuroblastoma and glioblastoma models. We identify decreases in PLK1 activity and RAD51 foci formation, and increases in mitotic arrest and cell death, as mechanistic contributors to combination sensitivity. Combined EYA and PLK1 inhibition is also effective in glioblastoma stem cell models that overexpress EYA1/EYA4 and specifically targets the cancer stem cell state. Finally, through multi-omic correlational analysis, we identify high levels of the NuRD complex and SOX9 as contributors to combination treatment sensitivity. Overall, this work identifies a novel synthetic lethal combination therapy with potential utility across a wide range of neurological cancers.

cancer biology↗

bioRxiv2024

Federated deep learning enables cancer subtyping by proteomics

Artificial intelligence applications in biomedicine face major challenges from data privacy requirements. To address this issue for clinically annotated tissue proteomic data, we developed a Federated Deep Learning (FDL) approach (ProCanFDL), training local models on simulated sites containing data from a pan-cancer cohort (n=1,260) and 29 cohorts held behind private firewalls (n=6,265), representing 19,930 replicate data-independent acquisition mass spectrometry (DIA-MS) runs. Local parameter updates were aggregated to build the global model, achieving a 43% performance gain on the hold-out test set (n=625) in 14 cancer subtyping tasks compared to local models, and matching centralized model performance. The approachs generalizability was demonstrated by retraining the global model with data from two external DIA-MS cohorts (n=55) and eight acquired by tandem mass tag (TMT) proteomics (n=832). ProCanFDL presents a solution for internationally collaborative machine learning initiatives using proteomic data, e.g., for discovering predictive biomarkers or treatment targets, while maintaining data privacy. Statement of SignificanceA federated deep learning approach applied to human proteomic data, acquired using two distinct proteomic technologies from 40 tumor cohorts from eight countries, enabled accurate cancer histopathological subtyping while preserving data privacy. This approach will enable privacy-compliant development of large-scale proteomic AI models, including foundation models, across institutions globally.

cancer biology↗