Monolayer culture alters EGFR inhibitor response through abrogation of microRNA-mediated feedback regulation

Abstract

Ex vivo drug screening is a potentially powerful tool for the future of cancer care, but the accuracy of results is contingent on the culture model. Both monolayer (2D) and spheroid (3D) culture systems offer advantages, but given the differences in mechanical environment, we hypothesized that that the suitability of one system over another would be critical for screening drugs with mechanical targets in mechanical tissues. HCC827 lung adenocarcinoma cells were challenged with EGFR tyrosine kinase inhibitors in monolayer and spheroid culture. RNA sequencing was performed on cells in both conditions to assess culture-induced transcriptional changes that could account for differences in drug response and differences in EGFR expression detected by immunostain. A microRNA microarray was performed to assess culture-induced differences in regulation of microRNA, and the impact of miR-146a-5p on drug response was verified by inhibition. Results were confirmed in human lung adenocarcinoma tissue. HCC827 spheroids were resistant to erlotinib and gefitinib, but significantly more sensitive in 2D culture. RNA-seq and immunostaining show a discrepancy in EGFR transcript and protein expression between the two conditions, which we attribute to miR-146a-5p. This microRNA targets EGFR and is differentially expressed between 2D and 3D culture. Inhibition of miR-146a-5p significantly increased erlotinib cytotoxicity, but validation in patient-derived spheroids suggests that the effect may be mutation-specific. Analysis of RNA-seq data suggests that cells in 2D culture become highly dependent on EGFR signaling to drive proliferation and cell spreading, resulting in a misleading level of sensitivity to EGFR TKIs, while the same cells in spheroid culture retain microRNA-driven EGFR feedback regulation that leaves them less vulnerable to EGFR inhibition. These findings underscore the need for close scrutiny of culture-induced effects on drug target regulation in model design for ex vivo drug screening.