Time-dependent epigenomic changes during 3D spheroid culture improve the preclinical modeling of non-small cell lung cancer

Background The chance of surviving non-small cell lung cancer (NSCLC) is low, due to the challenge in replicating it in preclinical models, limiting the discovery of effective treatments. Most preclinical models rely on two-dimensional (2D) cancer cell cultures. However, they do not reflect the effects of the 3D tumor architecture on cell proliferation, gene expression, and epigenomic regulation. 3D cell culture models, such as spheroids, can improve the therapeutic predictiveness of preclinical models. Yet, the time of 3D culture where 2D cells will retrieve 3D cell features to faithfully replicate NSCLC remain unknown. Methods We developed a protocol enabling long-term (> 24 days) 3D spheroid culture of NSCLC cell lines. We investigated transcriptomic and epigenomic profiles over time. We compared those changes to 2D cultures and to single-cell transcriptomes of NSCLC patient samples. We performed drug screens to compare the therapeutic vulnerabilities of long-term 3D spheroids to cells in 2D and analyzed their plasticity upon return to 2D culture, at the single-cell transcriptomic and pharmacological levels. Results NSCLC cells within spheroids revealed significant time-dependent epigenomic and transcriptomic changes. While DNA methylation remained stable, we observed time-dependent changes in the methylation and acetylation of histones H3 and H4 that recapitulated those observed in NSCLC. Accordingly, single cell transcriptomic profiles after 24 days of 3D culture were significantly more correlated with those of tumor cells from patients with NSCLC. Long-term spheroid culture increased the stringency of drug screening, inducing resistance to compounds that showed promise in many preclinical studies but failed to obtain approval in NSCLC clinical trials ( e.g. , histone deacetylase (HDAC) inhibitors). We demonstrated novel pharmacological vulnerabilities and synergistic interactions ( e.g ., between protein arginine methyltransferase and HDAC inhibitors). Strikingly, reverting long-term 3D spheroids to 2D culture reversed transcriptomic, epigenetic, and pharmacological signatures within 7 days, demonstrating how variations in cell culture conditions can impact outcomes in preclinical research. Conclusions Our findings demonstrate the value of implementing long-term 3D culture, as compared to short-term spheroids and 2D culture to improve preclinical studies. This method could be adapted to other cancer cell lines to improve preclinical drug development and maximize clinical trial results.