Abstract 293: Identification of a predictive biomarker for liposomal nanoparticle delivery through pan-cancer pooled screening

Abstract

Abstract Background: Nanoparticles are a modular technology with great promise for cancer treatment, but one obstacle for clinical implementation is a lack of predictive biomarkers to identify patient groups most likely to benefit from specific nanoformulations. We developed a massively parallel pooled screen to investigate genomic factors underlying NP-cancer cell engagement, and report here the identification of a predictive biomarker that is both formulation-specific and cancer lineage agnostic. Methods: We interrogated the interactions of 35 fluorescent nanoparticle formulations (15 liposomal and 25 polymeric) against 488 pooled, barcoded cancer cell lines using fluorescence-activated cell sorting and binned cells based on strength of association. After sequencing, the relative barcode abundance in each bin was used to generate an association score for each nanoparticle-cell line pair. We identified features predictive of NP-cancer cell association by interfacing this score with multi-omic data from the Cancer Cell Line Encyclopedia. Results: Using machine learning model predictions, we identified thousands of candidate biomarkers both across and within formulations. Expression of the lysosomal transporter SLC46A3 was the top ranked random forest feature for all liposome formulations and was prioritized for further study. Univariate analyses confirmed the significance of this candidate biomarker and indicated a negative relationship, such that low expression of SLC46A3 is highly correlated with high nanoparticle association. This inverse relationship held true across all 22 cancer lineages screened and was recapitulated in a non-pooled screen of 13 cancer cell lines with a range of native SLC46A3 expression. To determine whether modulating SLC46A3 expression is sufficient to negatively regulate uptake of liposomes, we developed a toolkit of engineered cell lines by knocking out SLC46A3 in high-expressing T47D cells and inducing overexpression in low-expressing LOXIMVI cells. Consistent with our hypothesis, we show that SLC46A3 expression is inversely correlated with uptake of liposomal, but not polymeric, nanoparticles. To evaluate the potential clinical utility of SLC46A3 as a biomarker, we tested in vivo delivery of an FDA-approved nanoparticle analog, the drug-free version of liposomal irinotecan, to LOXIMVI flank tumors via intratumoral injection and intravenous administration. For both administration methods, we show that low tumor expression of SLC46A3 correlates with significantly improved delivery of liposomal nanoparticles. Conclusions: We identified SLC46A3 as a negative regulator of liposomal nanoparticle delivery to cancer cells both in vitro and in vivo. Given that liposomal nanoparticles comprise the majority of FDA approvals for cancer indications, we propose SLC46A3 may be a clinically actionable biomarker for patient stratification. Citation Format: Joelle P. Straehla, Natalie Boehnke, Hannah Safford, Mustafa Kocak, Matthew G. Rees, Melissa Ronan, Danny Rosenberg, Charles H. Adelmann, Raghu R. Chivukula, Jaime H. Cheah, Hojun Li, Jennifer A. Roth, Angela N. Koehler, Paula T. Hammond. Identification of a predictive biomarker for liposomal nanoparticle delivery through pan-cancer pooled screening [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 293.