Abstract 4048: 3D bioprinted cancer models for drug screening versus 2D and <i>in vivo</i> models: A comparison study

Abstract 3D models are increasingly used to perform preclinical drug screening. 3D bioprinted models offer a promising alternative to better mimic the in vivo setting. In this work we compared molecular characteristics and drug sensitivity of two HER2 positive human cancer lines, the breast cancer model BT-474 and the ovarian cancer model SK-OV-3. Both models were a) cultured in classical 2D systems, with analyses during the exponential growth phase; b) bioprinted using a fibrinogen/alginate/gelatin bioink and characterized at different time points; c) implanted subscutaneously in CB17-SCID mice and treated once tumors were established (median volume #100 mm3). All models were compared using RNA sequencing. Drug sensitivity was evaluated against HER2 specific agents (lapatinib, trastuzumab-emtansine) and paclitaxel in 2D and 3D models and against paclitaxel in tumor models. Tumor bearing mice were treated with different paclitaxel doses ranging between 2.5 mg/kg (ineffective) to 30 mg/kg (T/C ratio of 0.4). Paclitaxel accumulation in tumor cells was quantified in all models exposed to various concentrations of drug. IC50 values were found to be approximately 3 to 10-fold higher in 3D models than in 2D models for all treatments tested. Analysis of 3D models at different time points using non disruptive (Presto Blue) or disruptive (flow cytometry analysis) methods showed active proliferation of tumor cells up to 40 days after bioprinting, as documented by Ki67 staining. In a standard assay objects were printed on day 0, exposed to treatments on day 4 and analyzed up to day 11. Addition of cancer associated fibroblasts (CAFs) in the bioink reduced sensitivity of tumor cells to therapy. Coculture with adipocytes or in the presence of adipocyte conditioned medium also showed reduced sensitivity to therapy. Tumor cell concentrations of paclitaxel was quantified by mass spectrometry at various time points and concentrations in the 2D and 3D settings as well as in in vivo tumor models. RNA profiling of the different models identified a category of genes which were highly expressed in the 2D setting only, while genes specifically expressed in in vivo models tended to be also more highly expressed in 3D models. Differentially expressed genes were confirmed using RT-qPCR and are currently being explored at the proteomic level. Gene Ontology analysis suggested that genes involved in cell adhesion, focal adhesion and extracellular-matrix receptor interactions tended to be less expressed in 2D models. These results suggest that 3D bioprinted models better mimic the in vivo situation than 2D models and should be systematically used in the preclinical drug screening process of novel agents, in order to reduce and better choose relevant murine models. Citation Format: Aurélie Cadiou, Eva-Laure Matera, Doriane Mathé, Marine Fellmann, Jérôme Guitton, Kamel Chettab, Christophe Marquette, Charles Dumontet. 3D bioprinted cancer models for drug screening versus 2D and in vivo models: A comparison study [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2026; Part 1 (Regular Abstracts); 2026 Apr 17-22; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2026;86(7 Suppl):Abstract nr 4048.