GEMMs as preclinical models for testing pancreatic cancer therapies

Abstract

Pancreatic ductal adenocarcinoma is the most common form of pancreatic tumour, with a very limited survival rate and currently no available disease-modifying treatments. Despite recent advances in the production of genetically engineered mouse models (GEMMs), the development of new therapies for pancreatic cancer is still hampered by a lack of reliable and predictive preclinical animal models for this disease. Preclinical models are vitally important for assessing therapies in the first stages of the drug development pipeline, prior to their transition to the clinical arena. GEMMs carry mutations in genes that are associated with specific human diseases and they can thus accurately mimic the genetic, phenotypic and physiological aspects of human pathologies. Here, we discuss different GEMMs of human pancreatic cancer, with a focus on the Lox-Stop-Lox (LSL)-Kras(G12D); LSL-Trp53(R172H); Pdx1-cre (KPC) model, one of the most widely used preclinical models for this disease. We describe its application in preclinical research, highlighting its advantages and disadvantages, its potential for predicting clinical outcomes in humans and the factors that can affect such outcomes, and, finally, future developments that could advance the discovery of new therapies for pancreatic cancer.

Keywords: Co-clinical trials; Drug development; Drug discovery; PDAC; Pancreatic ductal adenocarcinoma; Preclinical mouse models.

Fig. 1.

Progression of pancreatic cancer in the KPC model recapitulates the human disease. (A) In KPC mice, the conditional expression of mutant Kras^G12D and Trp53^R172H is controlled by a pancreas-specific Cre (Pdx1-cre in the model described here). In the absence of Cre, a transcriptional and translational STOP cassette flanked by loxP sites (LSL) silences the expression of mutant Kras^G12D and Trp53^R172H. When Cre is expressed in the pancreas, the STOP cassette is excised and the mutant alleles are expressed. The coloured boxes represent exons, and the asterisk (*) indicates the exon in which the mutation is present. The concomitant expression of mutant Kras and Trp53 in the murine pancreas results in pre-invasive pancreatic intraepithelial neoplasia (PanIN), which progress to pancreatic ductal adenocarcinoma (PDAC). (B) Haematoxylin and eosin (H&E) staining of normal pancreatic ducts in wild-type mice showing that they consist of a single layer of flat, low cuboidal epithelial cells with basal nuclei (arrows). (C-G) H&E staining in KPC mice. (C) In PanIN-1B, papillary or micropapillary projections develop in the ducts (arrows), otherwise these lesions are similar to PanIN-1A (not shown here). (D) More advanced PanIN-2 is characterized by nuclear abnormalities, including loss of polarity (black and white arrows), nuclear overcrowding, enlarged nuclei and rare mitoses (white arrow). (E) PanIN-3, ductal carcinoma in situ, is the highest grade of neoplasm and is associated with several abnormalities, including: papillary architecture with loss of nuclear polarity; occasional aberrant mitoses; nuclear abnormalities; large prominent nucleoli; and cribriforming (small clusters of epithelial cells budding into the lumen and necrosis in the lumen) (arrows). (F) PDAC, the resulting carcinoma, exhibits a glandular phenotype with duct-like structures of varying degrees of differentiation, and can exhibit adenosquamous or sarcomatoid histology. Substantial nuclear abnormalities occur and glands appear embedded in the tumour stroma (arrowheads) with completely random organization (arrows). Tumour cells can be observed next to arteries, with perineural and vascular invasion often seen. Necrotic debris can be seen in the lumen of the gland. (G) In advanced disease, metastatic spread is common, particularly to the liver (Liver met). Metastases often exhibit a glandular histology similar to well-differentiated PDAC. Arrow shows a metastatic deposit in mouse liver.

Fig. 2.

Schema of the drug discovery and development pipeline. The initial stages of the pipeline focus on the discovery and validation of newly identified disease-associated drug targets. Target validation is carried out using in vitro and in vivo approaches to confirm the relevance of the target in the disease being studied. This is followed by the identification of lead compounds through high-throughput or focused screens of chemical libraries or of naturally occurring molecules, or by structure-based rational drug design. The next stage is lead optimization, where the identified compound is subjected to chemical modifications to improve its pharmacological properties. Optimized lead compounds are then carried forward into preclinical testing, where pharmacology, toxicity and efficacy are assessed. Preclinical testing can occur in vitro, in 2- or 3-dimensional cell culture assays, or in vivo, in either xenografts or animal models, including genetically engineered mouse models (GEMMs). Promising therapies identified here are taken forward into the clinic. Preclinical models thus provide a bridge to the clinic, and are a requisite part of the drug development pipeline. For more detailed descriptions of the different stages of this pipeline, we refer readers to several recent reviews (Herter-Sprie et al., 2013; Hughes et al., 2011; Kamb et al., 2007; Kenakin, 2003; Pritchard et al., 2003).

Fig. 3.

Enrolment scheme for chemoprevention and intervention studies in KPC mice. In preclinical studies using LSL-Kras^G12D; LSL-Trp53^R172H; Pdx1-Cre (KPC) mice, different approaches are used to address different clinical questions. Grey arrows indicate no intervention; green arrows indicate pre-treatment monitoring; and blue arrows indicate treatment assessment. (A) Chemoprevention studies aim to evaluate dietary compounds or therapeutic agents that prevent tumour initiation or slow/arrest tumour development. Mice are enrolled between weaning and 6 weeks of age and, at this stage, usually present with early-stage PanIN. Treatments can be assessed at pre-determined time points or can continue until end point (to evaluate survival). (B) In early intervention studies, which are used to test anti-metastatic therapies, treatment is initiated when mice are 10-12 weeks old, when they commonly have early and late PanINs and occasional tumours. Treatment can last for a fixed period or can continue until end point. (C) Later intervention studies are performed on animals bearing established tumours and are thus relevant for identifying treatments that can reverse, slow or arrest cancer once fully established. These studies require more elaborate monitoring of mice, including manual palpation and ultrasound to monitor tumour size and progression. Treatment can begin when tumours reach the enrolment size for a study. Depending on tumour size, treatment can be short (9-11 days) or long (up to 45 days) (see main text for more detail). (D) Optimal design for intervention studies in KPC mice, incorporating serial sampling to allow pre- and post-treatment assessments of tumour and blood samples.

Fig. 4.

Effects of tumour size on response to treatment in the KPC model. Gemcitabine imparts a small yet significant survival benefit in mice with smaller tumours (our unpublished observations). Mice were enrolled on study when tumours reached 3-6 mm mean diameter. Mice were treated with either saline or gemcitabine (100 mg/kg body weight) intraperitoneally twice per week until end point. 12 mice were enrolled per cohort. The median survival post-enrolment is indicated for each cohort in the figure. The Log-rank test was conducted using GraphPad Prism (P=0.0236).