Background: Nanoparticles (NPs) hold enormous promise for the targeted delivery of therapeutics for cancer, but clinical translation is lacking largely due to limited tumor accumulation. Tumor heterogeneity and NP complexity make it challenging to deconvolute individual factors that contribute to NP-cell interactions. To address this, we developed a competition assay leveraging 500 stably DNA-barcoded adherent cancer cell lines annotated with multi-omic data from the Broad Institute (PRISM cells) to investigate cell association patterns across a library of NPs. We hypothesize that simultaneous screening of hundreds of cancer cell lines will identify factors underlying differential NP-cancer cell interactions.

Methods: We synthesized a library of 40 fluorescently-labeled NPs comprising clinical and experimental formulations. Clinical formulations included liposomal doxorubicin and irinotecan analogs and liposomal or poly(lactide-co-glycolide, PLGA) NPs with and without polyethylene glycol (PEG); these are either FDA-approved or in clinical trials. Experimental formulations included liposomal and PLGA cores electrostatically coated with a range of native and synthetic polymers as well as polystyrene NPs of varying sizes and surface chemistries. Fluorescent antibodies -in free form or NP-conjugated—were included as validation compounds. PRISM cells were pooled and incubated with NPs prior to fluorescence-activated cell sorting (FACS) to bin cells based on strength of NP association. After cell lysis, DNA barcodes were amplified and sequenced. Using appropriate controls to adjust for baseline barcode abundance, we generated an association score for each NP-cell line pair. Next, we performed multi-omic univariate analyses and applied a random forest algorithm to identify factors predictive of NP-cancer cell association.

Results: After pooled screening of PRISM cells, we consistently identified cancer cell lines based on strength of NP-association across technical and biologic replicates. Using antibodies and antibody-conjugated NPs targeting epidermal growth factor receptor (EGFR), we identified EGFR gene and protein expression as highly significant hits, validating our ability to robustly identify relevant biomarkers. Additional hits were evaluated based on strength and direction of association to identify predictive biomarkers by formulation. We also employed k-means clustering to investigate hits across NP formulations, identifying highly interconnected protein association networks that elucidate likely mechanisms of NP-cancer cell association.

Conclusions: We report a new pooled screening platform to investigate factors influencing NP-cancer cell interactions. We validated the screen by identifying known biomarkers, and also identified new predictive biomarkers that may pave the way for more effective nanotherapeutics.

Citation Format: Joelle P. Straehla, Natalie Boehnke, Mustafa Kocak, Melissa Ronan, Hannah Safford, Matthew G. Rees, Jennifer A. Roth, Angela N. Koehler, Paula T. Hammond. Development of a multi-omic, pooled cancer cell screen for nanoparticle delivery [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2021; 2021 Apr 10-15 and May 17-21. Philadelphia (PA): AACR; Cancer Res 2021;81(13_Suppl):Abstract nr 309.