The ideal scaffold for tissue engineering provides an architecture similar to natural tissue. For cardiac tissue repair, this three dimensional environment can be modelled on the myocardial extracellular matrix (ECM). To approximate the material composition of the ECM, collagen type I can be used, as it constitutes the majority of the cardiac ECM in vivo. Furthermore, it can be easily extracted and processed, and offers a range of potential cell binding motifs. However, although its biophysical properties can be controlled via cross-linking, recent reports have suggested that this process has detrimental effects on bioactivity. The aim of this work is to optimise the properties of collagen scaffolds, modelled on the cardiac ECM, by varying cross-linking and freeze-drying conditions and thereby create an environment in which cardiomyocytes exhibit physiological behaviour.

As its starting point, this thesis investigates porcine myocardial ECM. Tissue was decellularised using a combination of chemical and physical agents and the architecture and biophysical properties of the resulting cell-free samples were characterised. It was found that these samples comprised interconnected, isotropic pore structures, showed the capacity to swell i.e. increase their weight by) 2,800 % in 14 days, exhibited no mass loss over 14 days and had a Young’s modulus (E) of 8.0 +/- 0.4 kPa.

To mimic the observed properties, bovine collagen type I (Collagen Solutions, plc) was freeze-dried to produce isotropic pores and stabilised using a N-hydroxy-succinimide and 1-ethyl-3-(3-dimethylaminopropyl) carbodiimidehydrochloride, (NHS/EDC) cross-linking protocol. It was hypothesised that by tailoring the degree of cross-linking, the biophysical properties of the scaffolds could be modified to match the properties observed in the cardiac ECM. A mass of 0.69 g of NHS and 1.15 g of EDC, when combined in 100 ml of 96 % ethanol was defined as a cross-linking solution of 100 % concentration (100 % XL). Based on this definition, scaffolds were treated using solutions diluted to 10 %, 3 % and 1 %. Samples treated in 96 % ethanol alone, were defined as 0 % XL. It was found that while 100 % XL scaffolds possessed properties most similar to the ECM (swelling capacity of 4,200 % , mass loss of 10.8 % over 14 days, E = 6.9 +/- 0.5 kPa), cells showed limited attachment and survival at this cross-linking degree. By comparison, on 3 % XL samples, cell attachment and survival were significantly improved. In addition, the 3 % XL samples possessed sufficient stability to be used in cell culture (swelling capacity of 4,500 %, mass loss of 32.8 % over 14 days, E = 3.6 +/- 0.2 kPa).

In vivo, cardiomyocytes have a spindle-shaped morphology and populate elongated, anisotropic pores. It was hypothesised that this porous structure could be imitated by imposing increased temperature gradients on the collagen suspension during freezing. Finite element modelling was used to simulate the freeze-drying process and indicated that the temperature gradient during freezing could be increased from 4.4 °C to 23 °C, over the 10 mm sample height, using polycarbonate moulds with conductive bases. Experimentally this led to successful creation of scaffolds with anisotropic pores and demonstrated that temperature gradient is ultimately responsible for anisotropic pore architecture.

In a final set of experiments, isotropic and anisotropic 3 % XL scaffolds were seeded, in two different experiments, with either myoblasts or hESC-derived cardiomyocytes. The myoblast cell line was used to investigate cell access to the scaffolds and it was found that cells were able to invade isotropic scaffolds. However, in anisotropic scaffolds with horizontally aligned pores, invasion was limited to the top 60 µm of the samples, leading to an increased cell density in this area. This was subsequently confirmed with the hESC-derived cardiomyocytes. Seven days after cell seeding, isotropic scaffolds showed non-coordinated contractions, while anisotropic scaffolds with horizontally aligned pores, exhibited large-scale, coordinated contractions in the high cell density areas, deforming scaffolds up to 1170 nm over a length of 10 mm along the direction of the pores.

It has been demonstrated that, based on the characteristics of the cardiac ECM, three-dimensional scaffolds can be created which allow cardiomyocytes to connect and exhibit physiological behaviour. These scaffolds show considerable potential in the field of regenerative medicine and the insights gained during their development will contribute significantly to the further advancement of the field.