Plastic compression of collagen is based on unidirectional expulsion of fluid from hydrated collagen gel. The process results in dense collagen sheet, with higher density of collagen at the fluid leaving surface (FLS) than non-FLS. Compression process is completely cell-independent and at the same time cell-friendly. However, engineered tissues should replicate not only components of tissues in vivo (extracellular matrix and cells) but also their complex micro-architecture. Therefore the aim of this work was to develop collagen-based scaffolds with controllable micro-architecture for biomedical and tissue engineering applications using plastic compression (PC) of collagen. The objectives of this project were: i. to test formation of progressively opening channels in the PC collagen, ii. to investigate stable and predictable PC collagen patterning, iii. to adapt PC method in a upward-flow system as a route to process automation, iv. to investigate formation of channels using in layered PC collagen constructs. Two approaches were used in this work. Firstly, internal channels were introduced using lost fibre approach, where soluble glass fibres are incorporated in the scaffold and leave channel when dissolved. Shape and potentially progression of the channels’ opening is controlled by the shape of the template. The shape of the fibres was altered from cylindrical to conical in a controlled manner and incorporated into the PC constructs, resulting in conically-shaped channels, giving predictable internal 3D structures. The second approach relied on formation of dense collagen zone at the fluid leaving surface of the compressed collagen constructs. Formation of the densely packed collagen zone at the fluid leaving surface is essential for stable and faithful pattern formation in the process of micro-moulding. This finding has been applied in a novel upward-flow compression system to create channels using a ‘roofing’ technique. ‘Roof’ is formed by a compression of a new collagen gel on top of a patterned one; process results in open lumen channels. This appears to be due to a combination of the small dimension of the grooves in the base layer and viscosity of the collagen in the upper layer. This work demonstrates a new, previously unknown level of subtlety by which collagen fibrils can be packed and aggregated due to directional fluid flow. The outcome of this work is important for understanding pattern formation in PC collagen in vitro and potentially tissue morphogenesis in vivo. It also introduces new generation of implantable living tissue equivalents with complex micro-architecture. The multi-well compression technique has already been implemented in semi-automative working station for biomedical applications.

The aim of this project was to understand the key mechanisms which control and initiate the process(es) by which collagen connective tissues extend, remodel and grow under the influence of external tensile loads. Inguinal herniation was used as a model of natural cell mediated bio-creep to investigate the mechanisms behind such a process since the transversalis fascia (TF); a thin collagen connective tissue undergoes dramatic extension during herniation. It was hypothesized that the TF expands due to a process of growth and remodelling rather than stretch or injury and repair. The relationship between functional tissue mechanical properties, cell responses to mechanical forces and the architecture of the ECM were investigated. TEM imaging of TF collagen fibril ultrastructure demonstrated no differences between normal and hernia specimens and therefore no evidence of microdamage/ repair. An important finding was that hernia TF was not mechanically weaker or thinner than normal TF tissue, indicating it retains its normal material properties and thickness despite dramatic extension during herniation. In addition, the TF was identified to have anisotropic mechanical properties with greater strength and stiffness in the transverse anatomical plane, a finding which has not previously been reported for the TF. The ability of normal and hernia TF fibroblasts to contract and remodel a collagen matrix were almost identical, suggesting that hernia cells retain ‘normal’ cell behaviour in terms of their cytoskeletal and cell-motor functions. Analysis of TF fibroblast creep responses using an in vitro 3D bio-creep model demonstrated, to our knowledge, the first direct functional (mechanical) defect in hernia cell behaviour, since hernia cells did not demonstrate any capacity to mediate any measurable bio-creep. Our investigations of TF mechanics, structure and cellular behaviour suggest that TF expansion in inguinal herniation occurs due to growth/remodelling rather than stretch or micro-rupture and repair but the exact mechanisms remain inconclusive until further studies are carried out.

Osteoarthritis (OA) is a common, debilitating disease, involving degeneration of cartilage and bone in synovial joints. Subtle changes in the molecular structure of subchondral bone matrix occur and may precede gross morphological changes in the osteoarthritic joint. In this thesis, the analytical technique Raman Spectroscopy (which uses a monochromatic light source to probe chemical composition) is used to explore the hypothesis that subchondral bone changes occur prior to and during joint degeneration. The question is approached by looking at excised tibial plateaus from patients undergoing total knee replacement for advanced OA of the knee and comparing them with tibial plateaus from healthy joints. The samples were analysed with Raman spectroscopy, peripheral quantitative computed tomography (pQCT) and chemical analysis, to compare collagen alpha chains. The results show that bone matrix changes, related to OA, can be detected in the subchondral bone prior to overt cartilage damage, by Raman spectroscopy. These data provide support that chemical changes in bone can be related to the initiation of, or predisposition towards, joint degeneration. The results demonstrate that Raman spectroscopy should be further developed as a future tool to provide screening for early detection of joint degeneration based on correlating molecular-specific modifications in the subchondral bone.

Tissue engineering has traditionally relied on the use of scaffolds as inert, durable materials for seeded cells to re-grow. However, a paradigm shift in the role of scaffolds has become necessary towards bio-functional ‘devices’ which directly mimic native tissue matrix. It is key to control extracellular matrix remodelling and tissue structure by building control cues into the initial cell support matrix. The aim of this study was to test the effect and predictability of physical cues, engineered into 3D native collagen scaffolds by a cell-independent fabrication method, Plastic Compression (PC). Our findings indicated that fluid expulsion during collagen hydrogel compression produced anisotropic structuring and could be modelled as an ultrafiltration process. A groove/ridge topography engineered on collagen scaffolds through pattern-template-embossing influenced endothelial cell attachment/orientation and keratinocyte stratification in culture. Matrix collagen density and stiffness were directly related to its hydration level and could be controlled by limiting the extent of compression. Human-Dermal-Fibroblast (HDF) proliferation was proportional to matrix stiffness. In addition HDFs, seeded evenly within a PC collagen stiffness gradient, migrated and accumulated at the stiff end after 6 days. Bi-layer collagen matrices underwent cell-mediated integration, but despite higher cell migration across the interface in compliant than in stiff matrices at 24hrs, there was no significant difference in interface adhesive strength at 1 week. Core O2 tension in 3D spiral constructs directly correlated with total cell number along the diffusion path, i.e. consumption path length. This model was used to engineer local cell-mediated hypoxia in 3D constructs to generate populations of hypoxia-induced-signalling cells which produced angiogenic factor protein cascades. This in turn induced directed, functional micro-vascular ingrowth in vitro and in vivo. These data indicate how directing physical cues can be built into the structure of biomimetic, tissue-like scaffolds. This helps to understand intricate cell-matrix behaviours without reliance on complex biological control mechanisms and points the way to using simple physical cues for tissue formation.

The main issue associated with revision total hip replacements (rTHRs) is how to generate new bone and restore bone stock for fixation of the revision stem. Bone tissue engineering (BTE) seeks the generation of constructs ex vivo in order to replace damaged or lost bone. The aim of this thesis was to develop a bone tissue-engineered construct with a calcium-phosphate (CaP) coated porous metal scaffold seeded throughout its structure with mesenchymal stem cells (MSCs) in order to enhance new bone formation at rTHRs. The study had in vitro and in vivo phases. For the in vitro phase, CaP coatings by biomimetic and electrochemical methods on the surface of titanium and tantalum discs were investigated and seeded with MSCs under static culture conditions. Different coating methods produced different morphologies and compositions with biomimetic coatings enhancing MSCs growth while the electrochemical ones enhanced their osteogenic potential. An electrochemically CaP coated porous titanium cylinder was seeded with MSCs and dynamically cultured in a perfusion bioreactor, showing an increased MSCs proliferation and osteogenic differentiation and an even distribution of cells throughout the scaffolds compared to statically cultured constructs. Tissue-engineered constructs in the perfusion bioreactor were evaluated in vivo by implantation in the medial femoral condyle of sheep with and without gap. Their osseointegration and implant-bone fixation strength were compared to non tissue-engineered constructs. The results showed that the addition of MSCs to the scaffolds did not significantly increase osseointegration or implant-bone fixation strength. However, in the defects with gap the tissue-engineered constructs showed a higher implant-bone contact area and therefore higher forces were necessary to push the tissue-engineered implants out of the bone in the defects with gap than for the non tissue-engineered ones. In conclusion, BTE can be applied in order to develop constructs with a clinical application in rTHRs where a lack of bone stock is problematic.

Tendon injury is common in humans and horses, and incidence increases with age. The high-strain energy storing equine superficial digital flexor (SDFT) is injured more frequently than the low-strain positional common digital extensor (CDET). However, previous work indicated that matrix turnover is greater in the CDET than in the SDFT. It was hypothesised that matrix turnover is programmed by the cells’ strain environment; therefore high-strain energy storing tendons would have a lower rate of matrix turnover than low-strain positional tendons and the rate of matrix turnover would decrease with increasing age. The rate of matrix turnover was investigated by measuring the potential of the cells to synthesise and degrade matrix proteins, measuring the half-life of the collagenous and non-collagenous matrix proteins and assessing collagen turnover at the protein level. In vitro cell phenotype was also assessed in 2D and 3D culture and the effect of load on cells within native tissue was determined. The results show that turnover of collagenous and non-collagenous matrix proteins is differentially regulated in the functionally distinct SDFT and CDET. CDET tenocytes show greater potential for collagen turnover, whereas SDFT tenocytes have a greater potential for proteoglycan turnover; differences that are also present at the protein level. The differences in cell phenotype identified in vivo were lost in 2D and 3D culture, but tendon organ culture resulted in the maintenance of tenocyte phenotype. The cells’ ability to turnover the matrix does not decrease with increasing age, but collagen within the SDFT appears to become more resistant to degradation with ageing. This results in the accumulation of partially degraded collagen within the SDFT which may have a detrimental effect on tendon mechanical properties. These findings will help to elucidate the mechanisms behind the development of age-related tendinopathy and will be of use when developing treatment regimes.

Abstract: A New Zealand White (NZW) rabbit model was used to investigate the effect of cyclical cytotoxic chemotherapy on the structural and material properties of regenerate bone. This attempted to reproduce the biological situation encountered in a human adolescent with a primary malignant bone tumour, treated by surgical resection and either adjuvant or neo-adjuvant cyclical chemotherapy with bone transport to reconstruct the skeletal defect. General Hypothesis It is possible to produce normal bone by distraction osteogenesis in the presence of cyclical cytotoxic therapy. Materials and Methods: Forty immature male rabbits were divided equally into 2 groups. Each received 2 cycles of either cis-platinum / adriamycin or normal saline, with a tibial osteotomy and lengthening at 12-weeks of age. The timing of the cytotoxic drugs differed between groups in an attempt to simulate an adjuvant and neo-adjuvant dose schedule. Results: A reproducible animal model was developed, appropriate doses of cis-platinum and adriamycin were determined and it was demonstrated that surgical lengthening was possible in animals receiving chemotherapy. There were no differences in the physical characteristics of the regenerate or lengthened bone in either arm of the study. In the group that received 2-cycles of chemotherapy before lengthening (neoadjuvant group), there was a significant reduction in bone mineral concentration (BMC), bone mineral density (BMD) and volumetric bone mineral density (vBMD), assessed by dual X- ray absorptiometry (DXA). There was no effect on the structural properties assessed by compression testing In the group that received chemotherapy before and during lengthening (adjuvant group), there was no effect on mineralisation but a reduction in energy to yield and yield strain was demonstrated. Conclusion: These findings should be interpreted with caution, as the animals did not have malignant bone tumours and were given a limited drug regimen. The study did not demonstrate any consistent effect on the properties of regenerate bone but the assessment did not include histological analysis. Further work is needed to investigate the mechanism in which these agents affect distraction osteogenesis and this will require a dynamic assessment of bone formation and more sophisticated analysis of regenerate structure.

The hypothesis for this thesis was that using cultured autologous chondrocytes would lead to repair of full thickness defects with a hyaline type cartilage reparative tissue producing a significant improvement in pain and joint function in both the short and medium term. It was also hypothesised that the cover to contain the implanted cells is only a containment device so can be biologically inert resulting in no difference between Autologous Chondrocyte Implantation (ACI) and Matrix Assisted Chondrocyte Implantation (MACI). For this study autologous cultured chondrocytes were re-implanted under (ACI) or within (MACI) an inert type I/III collagen membrane. Patients were clinically assessed for up to seven years by standardised objective and subjective scores, as well as undergoing a second arthroscopy at one year to assess the regenerating tissue within the defect. All patients treated had full thickness chondral defects (1-12 cm^2).and were aged between .15-55 years age. The majority of patients had undergone at least one surgical procedure prior to referral for this technique, most commonly arthroscopy. The objective and subjective scores used showed a significant improvement post-surgery and the Short Form 36 proved to be sufficiently sensitive to demonstrate perceived health benefit from ACI at one year. ACI also resulted in an increase in post-operative score in patient previously treated with microfracture. The study showed that defect site, duration of symptoms, gender, defect size, and pre-operative score all affected the post-operative score. Histological assessment of the repair tissue showed that the regenerate is fibrocartilaginous but continues to adapt with time post-surgery resulting in a tissue more like normal articular cartilage. However, the type of regenerate does not significantly affect the post-operative patient score. Standard histological techniques showed that the regenerate contained collagen type IIA and B, collagen X, proteoglycans and S100. The results showed that the reparative tissue is showing features of hyaline cartilage but its architectural structure is not yet formed and the on the superficial surface only fibrous tissue is found architectural structure is not yet formed and the on the superficial surface only fibrous tissue is found. Both ACI and MACI produced significant improvements in knee function when compared to pre-operative levels (p<0.0001), with continued improvement in outcome for up to seven years, but the rate of clinical improvement in the MACI group was superior. There was, however, a greater tendency in the MACI group for fibrocartilaginous repair. This may be the reason why the MACI group had an inferior post-operative score at one year post-implantation compared to ACI group. However, by two years the MACI score surpassed the ACI group, possibly indicating a slower rate of maturation of the MACI regenerate. In summary, ACI for the repair of full thickness defects of the knee produces a repair with a tendency to form hyaline-like articular cartilage. Subjective and objective scores demonstrate sequential improvement for up to seven years demonstrating the durability of this technique in this group of patients.

Studies into the pathogenesis of systemic sclerosis (SSc) skin fibrosis to date have concentrated on dermal changes in the disease. Little attention has been paid to the epidermis in SSc. Epithelial-fibroblast interactions are believed to regulate wound healing and contribute to a number of fibrotic diseases. Recent proteomic data from our laboratory reveals altered keratinocyte (KC) specific proteins in SSc skin consistent with a wound healing phenotype of the disease epidermis. I therefore studied SSc KCs focusing on differentiation and KC-fibroblast interaction. I found that KC maturation is altered in SSc with abnormal persistence of cytokeratins 1, 10 and 14 into suprabasal layers. Cytokeratins 6 and 16, induced in wound healing KCs, were shown to be expressed in SSc epidermis. In addition, IL-1, a pivotal cytokine involved in KC and fibroblast events post epidermal injury, and its downstream signalling phosphoproteins p38 and JNK were elevated in SSc epidermis. I went on to study the effect of SSc epidermis on normal human fibroblasts. I found that SSc epidermis promoted fibroblast activation in an ET-1, TGF-β, and IL-1 dependent fashion. I suggest a double paracrine loop initiated by KC-derived IL-1 as a mechanism for epidermal-dermal co-activation in the disease, similar to that previously demonstrated for wound healing. There is a need for developing antifibrotic agents targeting epithelium-derived factors and their signalling pathways. I went on to study normal epidermal wound healing. A paradox during epithelial repair is that KCs proliferate despite a TGF-β dominated environment, which is known to be anti-proliferative. Our laboratory previously showed that prostanoids antagonise TGF-β-dependent events in human cells. The induction of prostanoids following injury could transiently free KCs from the anti-proliferative effects of TGF-β. I test this hypothesis by confirming transient induction of epidermal COX II and PGE2 following injury. I also show that PGE2 antagonises the anti-proliferative and pro-migratory effects of TGF-β on KCs. My work supports a model where induction of epidermal wound edge COX II leads to antagonism of TGF-β and allows KCs to proliferate prior to migration over the wound.

Bone is a composite material comprising an organic collagen scaffold surrounded by mineral platelets. The weight ratio of mineral to collagen is not constant, and hence the mechanical properties of the material can vary. In this thesis, the analytical technique Raman spectroscopy (which uses inelastically scattered photons to probe chemical composition) is used to explore the hypothesis that the mechanical properties of bone material are adapted in differently loaded regions by a subtle tuning of the collagen chemistry, which controls the mineral to collagen ratio. The question is first approached by looking at Raman spectra of a range of bony materials which are adapted to different functions. Bone material composition in different loading environments within individual bone organs is then examined, and finally, collagen Raman band profiles from a range of bony material are examined to search for variations which would signify differences in collagen chemistry. The results of the experiments show that for the functionally adapted bony materials a correlation exists between the Raman spectra and mechanical properties. Raman data from further experiments show that there are anatomical regions near the ends of long bones where the mineral to collagen ratio is lower than average (that this may be another adaptation of mineral/collagen ratio to function is discussed). Finally, analyses of Raman band profiles are presented which show that there are differences in the collagen chemistry of the various functionally adapted bony materials. The work furthers our understanding of the relationship between the mechanical properties of bone material and its Raman spectrum, a relationship which is of great importance if Raman spectroscopy is to be developed as an in vivo technique. Understanding the regional variations in mineral to collagen ratio will be vital if Raman spectroscopy is to be utilised to measure the mechanical integrity of whole bone organs. The elucidation of the band profile differences may provide a therapeutic target for the management of bone conditions i.e. enabling the control of the mechanical properties of bone; further research is required to determine what specific aspect of the collagen chemistry they represent, and to establish the causal relationship from the spectra to mechanical properties.