Making experimental rodents obsolete for research into Barrett's oesophagus - an in vitro model molecularly mimicking disease

Oesophageal carcinoma, especially adenocarcinoma has the fastest increasing incidence rate in the Western World with an average increase of around 5% per year in the Netherlands. Metastatic disease remains unusually lethal thus prevention of disease is considered the way forward. Oesophageal adenocarcinoma develops from a condition known as Barrett’s oesophagus. In Barrett’s oesophagus as a consequence of exposure of the oesophagus to bile and stomach acid a transdifferentiation occurs in which the squamous epithelium of the oesophagus acquires columnar characteristics and becomes phenotypically similar to more distal gastrointestinal tissues, like the colon. Molecular analysis, however, has shown that this is only a partial transdifferentiation and that Barrett’s oesophagus combines oesophageal characteristics and properties of more distal epithelia. Patients with Barrett’s oesophagus are regularly screened by gastroendoscopy for the development of oesophageal cancer, which is highly invasive and burdensome for the patients involved. Despite screening efforts a substantial number of patients still develops oesophageal cancer. Currently it can be estimated that there are 15,809 patients with clinically confirmed Barrett’s oesophagus in the Netherlands, but this is likely a gross underestimation of the true incidence. Thus there is a clear societal need for models of Barrett’s oesophagus that would allow testing for better preventive strategies. The by far most widely used animal models for Barrett’s oesophagus are surgical and are performed in either rats or mice. In such models a surgical bypass or destruction of existing antireflux mechanisms in experimental rodents is created and the subsequent exposure of the oesophageal epithelium to acid and/or bile creates an oesophagitis that finally develops into a Barrett’s oesophagus-like condition. Problems with these surgical models are mainly ethical, practical and economical : the low efficacy of the model (many -if not most- animals will not survive the surgical procedure hampering progress of investigations and substantially adding to costs) and the investigators need to undergo long, extensive and costly training to be able to perform the procedure. Ethically the severe distress these experimental animals suffer is problematic. The squamous epithelium in the murine oesophagus and forestomach is keratinized, while in humans the oesophagus is lined by a non-keratinized squamous epithelium. Hence developmental pathways are likely quite different between mice and humans, as they involve different epithelial structures and thus translation from animals to humans is difficult. One great limitation to basic laboratory research into the pathogenesis of Barrett's oesophagus and associated oesophageal adenocarcinoma has been the lack of reliable cell lines and cell culture model systems. We have generated embryonal stem cells that overexpress HOXA13 upon induction with doxycycline. Embryonic stem cells have important advantages over traditional immortalized approaches as the former have a normal karyotype and are untransformed as dramatically illustrated by the potential to grow a new animal from a single embryonic stem cell. Importantly, embryonic stem cells are pluripotent. If subjected to a squamous oesophageal differentiation protocol, the doxycycline-induced HOXA13 embryonic stem cells do not develop into a squamous epithelial cells but in columnar epithelial cells with a Barrett’s phenotype. We hypothesise that these cultures represent a novel model that very closely mimics true human Barrett’s oesophagus. Through this proposal we aim to prove that this notion is correct by comparing systems biology information from these induced Barrett’s cultures to clinical Barrett’s oesophagus, also exploiting different induction protocols in the embryonic stem cells. We expect to able to show that this new model is vastly superior to existing models of Barrett’s oesophagus (whose inadequacies are explained above), making the current rodent models obsolete. Firstly, by comparing kinome profiles and expression profiles of cultures in the presence and the absence doxycycline deep insight into the consequences of HOXA13 expression will be generated. In the second phase we shall compare proteomic and transcriptomic profiles of clinical Barrett’s oesophagus to HOXA13-induced squamously-differentiated embryonic stem cell derivatives. In the last phase, we aim to document the usefulness of the new model for clinical and industrial research. This will be done by comparing (using Pepscope’s systems medicine platforms) induction of HOXA13 in vehicle treated cultures to those exposed to chemopreventive medications. In conjunction these experimentations will establish a novel model for Barrett’s oesophagus evidently superior to current animal models and provide the community with a body of literature documenting this usefulness, which should make current animal experimentation obsolete.