3D Biomimetic Liver Cancer Model: Diethylnitrosamine-Induced Proteomic Dysregulations in Stromal-Epithelial Milieu

Hydrogel-based three-dimensional (3D) co-culture systems are emerging as biomimetic platforms that more accurately recapitulate tissue architecture and microenvironmental interactions compared to conventional two-dimensional (2D) cultures. This study introduces an engineered 3D liver-like model to investigate compartment-specific responses to the potent hepatocarcinogen Diethylnitrosamine (DEN), with a focus on early events in carcinogenesis and tumor-stroma interactions. AML12 and 3T3 cell lines were treated with DEN or vehicle either in 2D culture or in 3D hydrogels in four experimental groups: (1) DEN-treated AML12 with vehicle-treated 3T3, (2) DEN-treated 3T3 with vehicle-treated AML12, (3) both cell types DEN-treated, and (4) both vehicle-treated. The cultured recombinants were subjected to proteomic profiling via mass spectrometry, followed by bioinformatics analysis and the results were validated through immunocytochemical staining (ICC). Gene ontology analysis revealed that cytoskeletal, RNA metabolism, and scaffold/adaptor proteins were among the most significantly enriched in 3D versus 2D models. Structural proteins emerged exclusively in mixed 3D co-cultures, reinforcing the organotypic nature of the system. Enriched pathways in 3D included intermediate filament organization, actin dynamics, and focal adhesion-pathways closely associated with liver carcinogenesis. Protein-protein interaction analysis demonstrated maximal network complexity in 3D cultures where both compartments were DEN-exposed. Survival analysis further identified poor-prognosis biomarkers (KRT20, KRT15, KRT14) uniquely enriched in this condition. ICC staining supported the proteomic findings. This organoid-like 3D co-culture model provides a physiologically relevant platform for investigating early-stage liver carcinogenesis and highlights the critical role of stromal-epithelial interactions. Its ability to replicate organ-level complexity and generate clinically relevant proteomic signatures supports its utility in translational cancer research and future drug discovery applications.