Objectives: Experimental data based on cell line-derived xenograft models (cell xenograft) seldom reproduce the clinical situation, and therefore we demonstrated here the superiority of a murine model involving transplantation of human pancreatic cancer tissue fragments (tumor graft), focusing on the histological features and drug delivery characteristics.
Methods: Tumor pieces from 10 pancreatic cancer patients were transplanted into SCID (severe combined immunodeficient) mice. Histological characteristics of tumor grafts, including morphology, desmoplastic reaction, and vascularization, were compared with those of cell xenografts. Drug delivery was evaluated by quantifying the concentrations of injected drug, and the results were compared with its histological features.
Results: Eight of the 10 transplanted tumors successfully engrafted. Histological comparisons between tumor grafts and cell xenografts revealed the following: the amount of stroma was more (22.9% ± 11.8% vs 10.8% ± 5.4%; P < 0.05), vessel-cancer cell distance was longer (35.3 ± 39.0 vs 3.9 ± 3.1 μm; P < 0.001), and microvessel density was lower (6.8 ± 1.9 vs 10.8 ± 2.1 vessels/0.4 mm(2); P < 0.05) in tumor grafts. Drug concentrations in tumor grafts were lower than those in cell xenografts (3.3 ± 1.2 vs 6.0±0.2 μg/mL; P = 0.003), and the differences were correlated with the histological differences.
Conclusions: Pancreatic tumor grafts better reproduce the histological nature of clinical cancer and thus provide a more realistic model that is applicable for pharmacokinetic studies.