Abstract 2775: Establishment of patient primary gastric cancer xenograft models for test of anticancer agents

Background. Human xenograft tumor models established by inoculating human tumor cell lines into immunodifficient mice have been routinely used for preclinical test of anticancer agents. But tumor cell lines have a relatively low transplantability, which resulted in a limited number of tumor models available for selection of right tumor modes for test of targeted therapeutics. Recently, we have developed a large number of patient primary gastric cancer xenograft models by transplanting human fresh gastric tumor fragments into nude mice, which have been employed for preclinical test of clinically used drugs and novel anticancer agents for their chemosensitivity screening. Mehtods. The fresh gastric tumor samples were collected from local hospitals. The tumor fragments of 2-3 mm were subcutaneously implanted in the flanks of nude mice by trocar needle. Sixteen tumor fragments were grafted into four mice from one patient tumor tissue (passage 0). The clinically used drugs included 5-FU, gemcitabine, and doxorubicin. The histology and gene sequence of the established primary gastric tumor models were analyzed and compared with patients’ original tumors. Results. A total of 110 patient gastric cancer samples were implanted into nude mice; and 30 primary tumor models have been established with a tumor taking rates of 27% for the first passage. The tumor taking rates were higher in the later passages ranged from approximately 70-95%. The therapeutic efficacy of the test anticancer drugs was consistent with their clinical findings. The patients’ primary gastric cancer xenografts from all 5 passages retained a similarity in architecture, histopathological morphology, and genomic mutation status to their patients’ original tumors. Conclusions. The patient primary tumor model system can provide a larger number of models within same tumor histological type for selection of right models for preclinical testing novel agents based on their anticancer mechanism. The primary tumor models retain a similarity in histology and genomic mutation status to their patients’ original tumors. They may predict more relevant clinical response rate and higher correlation with clinical findings than use of traditional xenograft models established from long-term cultured cancer cell lines. Especially, they have advantages for test of target-oriented therapeutics in new drugs discovery and development programs. Citation Format: Ning Zhang, Shengyi You, Wenwei Li, Wen Zhou, Yong Liu, Rui Zh