Author:

Abstract:

The field of tissue engineering, as well as the study of many diseases, relies on in vitro models in which obtaining proper vascularization is key. In the last decades, researchers have discovered that chemical signaling alone cannot guarantee the proper growth of new blood vessels from pre-existing vasculature (angiogenesis). Rather, a combination of chemical and mechanical cues is crucial in vascular biology. Through mechanotransduction pathways, cells can sense and respond to mechanical stimuli. An important aspect of angiogenesis is the exertion of forces on the extracellular matrix (ECM) by endothelial cells. While the field has developed well-established Traction Force Microscopy (TFM) techniques to quantify cellular forces within 2D in vitro models, these cannot be applied to study processes that evolve in a three-dimensional ECM, such as angiogenesis. The few existing TFM methods applicable to 3D in vitro models are often validated in simplified scenarios (bypassing inherent sources of error such as microscopy noise or bead density) and the codes are either not available or they require high programming expertise. These methods do not necessarily fulfill fundamental and universal laws such as the equilibrium of forces in the hydrogel. Moreover, in vitro systems that make use of fibrillar hydrogels such as collagen are very common in the field. However, all the existing 3D TFM methods model the inherently discrete ECM (composed of discrete interconnected fibers) as a continuum medium. This leads to homogenization procedures that limit the resolution at which forces can be recovered. As a result of all these factors, 3D TFM remains a marginally used methodology with questionable applicability. In this thesis, an advanced in silico simulation framework that incorporates complex cell geometries, simulation of microscopy images of varying bead densities, and different focal adhesion sizes and distributions have been developed. This framework can be used to quantify the expected range of errors from a specific TFM workflow. A novel inverse traction recovery method that imposes the fulfillment of equilibrium with real and known forces acting in the hydrogel was also developed. The simulation framework was used to compare the accuracy of this method to that of a forward method, which does not make use of regularization. Results showed a two-fold increase in traction recovery accuracy for the inverse method. Moreover, TFMLAB, an open-source Matlab-based software platform that includes all the necessary computational steps for straightforwardly calculating 3D cellular forces in 3D synthetic matrices was developed. Specific focus was put on making it easy to use by non-technical users. Within the functionalities of this toolbox, one can automatically generate a 3D image-based Finite Element mesh for any cellular culture system and compute the tractions from the measured displacements. A novel way of performing 3D TFM in fibrillar hydrogels is also presented. The methodological formulations of the so-called, data-driven 3D TFM, are provided. This methodology is based on building a discrete fiber mechanical model that incorporates geometrical information of the imaged region. After adapting the inverse method to the new model of the hydrogel, in silico simulations were conducted to again prove its superiority to the forward method. This establishes a potential new paradigm for 3D TFM in fibrillar hydrogels. Finally, a demonstration of the usefulness of the computational tools developed in this thesis is provided by applying them to calculate forces within two different 3D in vitro models. First, angiogenic sprouting in a disease context is investigated by calculating forces in a 3D in vitro vascular invasion model for both wild-type and mutant endothelial cells representative of Cerebral Cavernous Malformations (CCM). Mutant cells displayed higher traction forces than wild-type cells suggesting an increase in cell contractility upon depletion of the CCM genes. Second, the flexibility of the developed methods is showcased by calculating 3D forces around growing human Neural Tube Organoids (hNTOs). Results suggest that the stress distribution around growing hNTOs is not homogeneous and that different regions of the organoids can be subject to specific mechanical stimuli during growth, enabling future explorations of the implications with regards to stem cell differentiation and patterning.