Precision oncology aims to find the most beneficial therapeutic strategies for individual patient1. Currently, numerous preclinical models such as in vitro primary culture, in vitro organoid culture2, and patient-derived xenografts (PDX) in mice before or after organoid culture are proposed for diagnosis and to screen/assess the potential therapeutic choices3. PDX model generated by the injection of human primary cancer cells into immune-compromised mice, is one of the most promising tools for personalized drug screening in clinical oncology3,4. Unlike the cultured cell line in vitro, PDX models usually preserve the integrity and heterogeneity of the in vivo tumor environment, better mimicking the diversity and idiosyncratic characteristics of different tumor patients, and therefore, may predict the potential medical outcome of patients4. However, the generation of PDX models in mice requires high quality patient samples and months of time to gather sufficient cells and models for multi group experiments, and the cellular/genetic compositions of the xenograft may drift from those of the original patient’s biopsy. The success rate for establishing mice PDX model is also low, making it difficult to be broadly implemented in clinical practice. For the patients carrying rapidly progressed cancers like pancreatic cancer, they may not be able to obtain valuable information from the PDX experiments in time.
In the past few years, zebrafish has been reported to be potential hosts for not only CDX (cell-derived tumor xenograft) models, but also PDX models5,6,7,8,9,10. As a vertebrate model animal, zebrafish harbors sufficient similarities with mammals in both genetics and physiology, with two significant advantages: transparency and small in size11. Zebrafish is also highly fecundity, and hundreds of inbred larvae can be obtained within a few days from a single pair of adults12. Several studies have employed zebrafish to generate both transgenic and xenograft models of cancer diseases13,14. Compared to mice xenografts, zebrafish xenografts allow tracking at single cell resolution. A certain amount of human tissues is capable of generating hundreds of zebrafish PDX models (zPDXs), while may only be sufficient to generate a couple of mice PDX models15,16. Besides, the zebrafish larvae at 2-5 dpf already develop complete circulatory systems and metabolic organs such as liver and kidney, but not the immune system17, while the remaining yolk sac is a natural 3D medium, ideal for drug screening, drug resistance tests and tumor migration observations6,18,19,20,21.
With an ultimate attempt to use zPDX as a screening/testing platform for clinical use, here, we describe an optimized proposal for zPDX model of pancreatic cancer, which allows the in vivo candidate drug assessment within a short time using fewer cells at lower costs. Compared to the previous references about zPDX6,9,10, we introduced several optimizations to make the system more feasible and reliable for clinical personalized diagnosis: 1) pre-sorting different cell groups in the primary tumor tissues and stabilizing primary cells for one week before further experiments; 2) labeling the human cells and enhancing the cell viability in xenograft via lentivirus-based genetic modification; 3) optimizing the zebrafish culture condition in both nutriment supplements (glucose and glutamine) and temperature; 4) quantifying the drug responses of different cell types in a comparative manner. We also made changes to the injection solution by adding several supplementary materials. Altogether, those improvements provide the possibility to quickly generate a more patient-like xenograft in zebrafish hosts that can be used as a reliable tool to assess the response of candidate drugs.