Pancreatic cancer (PC) is anticipated to be second only to lung cancer as the leading cause of cancer-related deaths in the United States by 2030. Surgery remains the only potentially curative treatment for patients with pancreatic ductal adenocarcinoma (PDAC), the most common form of PC. Multiple recent preclinical studies focus on identifying effective treatments for PDAC, but the models available for these studies often fail to reproduce the heterogeneity of this tumor type. Data generated with such models are of unknown clinical relevance. Patient-derived xenograft (PDX) models offer several advantages over human cell line-based in vitro and in vivo models and models of non-human origin. PDX models retain genetic characteristics of the human tumor specimens from which they were derived, have intact stromal components, and are more predictive of patient response than traditional models. This review briefly describes the advantages and disadvantages of 2D cultures, organoids and genetically engineered mouse (GEM) models of PDAC, and focuses on the applications, characteristics, advantages, limitations, and the future potential of PDX models for improving the management of PDAC.

Gemcitabine is used to treat pancreatic cancer (PC), but is not curative. We sought to determine whether gemcitabine + a BET bromodomain inhibitor was superior to gemcitabine, and identify proteins that may contribute to the efficacy of this combination. This study was based on observations that cell cycle dysregulation and DNA damage augment the efficacy of gemcitabine. BET inhibitors arrest cells in G1 and allow increases in DNA damage, likely due to inhibition of expression of DNA repair proteins Ku80 and RAD51. BET inhibitors (JQ1 or I-BET762) + gemcitabine were synergistic in vitro, in Panc1, MiaPaCa2 and Su86 PC cell lines. JQ1 + gemcitabine was more effective in vivo than either drug alone in patient-derived xenograft models (P < 0.01). Increases in the apoptosis marker cleaved caspase 3 and DNA damage marker γH2AX paralleled antitumor efficacy. Notably, RNA-seq data showed that JQ1 + gemcitabine selectively inhibited HMGCS2 and APOC1 ~6-fold, compared to controls. These proteins contribute to cholesterol biosynthesis and lipid metabolism, and their overexpression supports tumor cell proliferation. IPA data indicated that JQ1 + gemcitabine selectively inhibited the LXR/RXR activation pathway, suggesting the hypothesis that this inhibition may contribute to the observed in vivo efficacy of JQ1 + gemcitabine.

ABC toxin complexes (Tcs) are tripartite complexes that come together to form nano-syringe-like translocation systems. ABC Tcs are often compared with Bacillus thuringiensis (Bt) toxins, and as such, they have been highly studied as a potential novel pesticide to combat growing insect resistance. Moreover, it is possible to substitute the cytotoxic hypervariable region with alternative peptides, which promise potential use as a novel peptide delivery system. These toxins possess the unique ability to form active chimeric holotoxins across species and display the capability to translocate a variety of payloads across membrane bilayers. Additionally, mutagenesis on the linker region and the receptor binding domains (RBDs) show that mutations do not inherently cause a loss of functionality for translocation. For these reasons, Tcs have emerged as an ideal candidate for targeted protein engineering. However, elucidation of the specific function of each RBD in relation to target receptor recognition currently limits the use of a rational design approach with any ABC Tc. Additionally, there is a distinct lack of targeting and biodistribution data for many Tcs among mammals and mammalian cell lines. Here, we outline two separate strategies for modifying the targeting capabilities of the A subunit (TcA) from Xenorhabdus nematophilus, Xn-XptA2. We identify novel structural differences that make Xn-XptA2 different than other characterized TcAs and display the modular capabilities of substituting RBDs from alternative TcAs into the Xn-XptA2 scaffold. Finally, we show the first, to our knowledge, biodistribution data of any TcA in mice.