Culture Dimensionality Modulates Gallium Maltolate Response in Glioblastoma: Comparative Analyses in 2D and 3D Models

Gallium maltolate (GaM) targets iron-dependent processes in glioblastoma (GBM), but responses vary with the model context. We evaluated GaM across established (A-172, U-87 MG) and patient-derived (3005, 3019, 3034, 3048, 3073) GBM lines in 2D and 3D using viability modeling (IC10/IC50/IC90), transferrin receptor (TFRC) quantification, oxygen consumption rate (OCR), and PCA/PLS-DA-guided metabolomics with false discovery rate (FDR) and variable importance in projection (VIP)-based selection. GaM reduced viability in all models, but the impact of 3D culture on IC50 was line-specific rather than uniformly increasing resistance: classical/proneural patient-derived lines (3005, 3019, 3048) showed equal or lower IC50 in 3D compared with 2D, 3073 showed minimal change, whereas the mesenchymal-like line 3034 displayed a marked IC50 increase in 3D. Basal TFRC levels correlated with IC50 in 2D but not 3D, indicating that the TFRC alone does not predict GaM response once microenvironmental constraints are introduced. Instead, a broader phenotype involving TFRC/CD44/MGMT and TFR2 expression is associated with 3D sensitization versus protection. OCR was markedly suppressed in A-172, U-87 MG, 3048, and 3073, particularly in 3D, while 3005 and 3019 were more respiration-resilient. Multivariate analyses showed treatment-dominant separation in 3005 and 3048, format dominance in 3019 and 3034, and time effects in A-172/U-87 MG/3073. A concise metabolic signature consisting of tryptophan, methionine, uracil, and allantoin indicated coordinated perturbations in amino acid, nucleotide, and redox pathways. These findings support complementary 2D and 3D patient-derived GBM models for mechanistic studies and the predictive evaluation of GaM.