The large intestine is a multifunctional organ that acts as a selective barrier preventing the passage of potentially harmful substances into the body while ensuring water and electrolytes absorption. Perturbation of intestinal barrier can induce a proinflammatory response. In individuals with genetic predispositions this can result in chronic inflammation and a defect in epithelial regeneration leading to incomplete abnormal healing, as in the case of inflammatory bowel diseases (IBD). Current IBD treatments only focus on controlling the inflammation symptoms and fail to restore epithelial healing. Despite research efforts, the mechanisms underlying abnormal epithelial regeneration remain poorly understood. This is particularly difficult to study in traditional mice models since they provide restricted access to the epithelium for real-time longitudinal observations. New approach methodologies (NAMs), such as in vitro artificial colonic microphysiological systems (MPS) that faithfully reproduce complex in vivo systems, are therefore essential tools to improve our understanding of the human gut physiology and pathologies.
Although some models have been developed to represent the gut architecture and physical cues, such as continuous flow and topography, very few simultaneously recapitulate the 3D tissue architecture, nutrient flow control and in particular, matrix stiffness. However, this parameter is critical, as human colon tissue displays distinct mechanical properties under physiological and pathological conditions, the latter presenting a higher matrix stiffness. In that context, the objective of this work was to further develop, characterize and validate a human micro-physiologicalsystem that accurately represents in vitro the 3D human colonic epithelium topography and matrix stiffness while allowing microenvironment control of both luminal and stromal compartments separately thanks to a microfluidic set-up.
The first part of this work optimized and characterized a collagen I/polyacrylamide interpenetrating hydrogel to mimic the extracellular matrix and reproduce physiological (4kPa) or inflammatory (26kPa) colon stiffness. We evaluated bulk stiffness, fibrillar collagen organization, and diffusion of cell-relevant molecules. The system accurately replicated crypt topography, formed a dense, homogeneous collagen network, and enabled diffusion of molecules up to 70kDa from a channel mimicking vascular nutrient supply. Optimized microfluidic flow supported long-term culture of Caco-2 cells under both stiffness conditions. Cultures were highly polarized, showed stiffness-dependent morphology, and exhibited higher expression of enterocyte maturation markers (FABP1, ALPI, KRT20, VIL1) compared to standard 2D Transwell cultures.
Finally, the second part focused on optimizing cell culture conditions to complexify the model by integrating: i) human primary cells from colorectal organoids derived from patient biopsies to represent different intestinal cell populations and ii) fibroblasts-epithelial cocultures to take into account the mesenchyme contribution on tissue homeostasis. In conclusion, this novel device combining colon-like crypt topography, matrix stiffness and the possibility of microfluidic control allowed us to recapitulate differential cell behaviour in response to matrix stiffness and promote maturation of Caco-2 cells. Integration of primary cells in the system will permit a better understanding of the tissue both in physiological and pathological conditions, opening the door to more precise and predictive biomedical studies.