Analysis of gene copy number in esophageal patient-derived tumor xenografts.

Abstract

e13648 Background: Patient-derived tumor xenograft (PDX) models are a valuable resource for studying cancer biology and antitumor drug evaluation. The suitability of tumor models in vivo depends on how accurately they mimic a human disease and reproduce the histotype and molecular genetic features of a human tumor. The purpose of the study was to create a PDX model of human cancer and analyze its characteristics. Methods: PDX models of esophageal cancer were obtained by transplanting a tumor fragment from a patient with esophageal squamous cell carcinoma to the BALB/c Nude athymic mice (n = 10 for one PDX generation). Preservation of the tumor histotype was confirmed histologically (hematoxylin and eosin staining). 5 PDX were generated. An analysis of the relative copy number of the YAP1 and KDM6A genes (Real-Time qPCR) in xenograft and donor tumor tissues was performed in each generation. Results: A decrease in the copy numbers of the YAP1 and KDM6A genes by 3.8 and 2.2 times, respectively, was observed in PDX tumor samples (F3 generation) compared to normal donor tissues (p < 0.05). This trend maintained in F4 and F5 generation PDX samples. No changes in the copy numbers of the YAP1 and KDM6A genes were detected in PDX tumor samples in F1 and F2 generations. A cluster analysis (Hierarchical Clustering, Euclidean distance) demonstrated that samples of the first and second generations of esophageal cancer PDX models were closest to the patient tumor tissues in gene copy numbers. Conclusions: Later generations of esophageal cancer PDX models are characterized by changes in the genes copy numbers due to changes in the tumor and clonal selection. Early PDX generations better reproduce genetic, molecular and morphological features of tumors.