Organoid-based ex vivo reconstitution of Kras-driven pancreatic ductal carcinogenesis

Abstract The organoid culture technique has been recently applied to modeling carcinogenesis in several organs. To further explore its potential and gain novel insights into tumorigenesis, we here investigated whether pancreatic ductal adenocarcinoma (PDA) could be generated as subcutaneous tumors in immunocompromised nude mice, by genetic engineering of normal organoids. As expected, acute induction of KrasG12Din vitro occasionally led to development of tiny nodules compatible with early lesions known as pancreatic intraepithelial neoplasia (PanIN). KrasG12D-expressing cells were enriched after inoculation in the subcutis, yet proved rather declined during culture, suggesting that its advantage might depend on surrounding environments. Depletion of growth factors or concurrent Trp53 deletion resulted in its robust enrichment, invariably leading to development of PanIN or large high-grade adenocarcinoma, respectively, consistent with in vivo mouse studies for the same genotype. Progression from PanIN was also recapitulated by subsequent knockdown of common tumor suppressors, whereas the impact of Tgfbr2 deletion was only partially recapitulated, illustrating genotype-dependent requirement of the pancreatic niche for tumorigenesis. Intriguingly, analysis of tumor-derived organoids revealed that KrasG12D-expressing cells with spontaneous deletion of wild-type Kras were positively selected and exhibited an aging-related mutation signature in nude mice, mirroring the pathogenesis of human PDA, and that the sphere-forming potential and orthotopic tumorigenicity in syngenic mice were significantly augmented. These observations highlighted the relevance of the subcutis of nude mice in promoting PDA development despite its ectopic nature. Taken together, pancreatic carcinogenesis could be considerably recapitulated with organoids, which would probably serve as a novel disease model. We showed that Kras-dependent pancreatic carcinogenesis can be essentially, albeit not perfectly, reconstituted even in a completely ex vivo setting, providing an alternative model to decipher its pathogenesis by comparing with in vivo models, and valuable resources for translational studies.