Osteomimetic matrix components alter cell migration and drug response in a 3D tumour-engineered osteosarcoma model.

Osteosarcoma management continues to lack the appropriate prognostic tools to assign personalised treatment. This leaves non-responders to standard care vulnerable to recurring disease and pulmonary metastases. Developing 3D in vitro disease models to serve as a test bed for personalised treatment is a promising approach to address this issue. This study describes the generation of 3D osteosarcoma models termed "tumouroids", which are geometrically compartmentalised to reproduce the bone cancer mass and its surrounding. Although the tumour microenvironment impacts osteosarcoma in many ways, this model focussed on interrogating the influence of a biomimetic matrix on tumour cell behaviour. The 3D matrix was supplemented with the bone-marrow proteins laminin, fibronectin and NuOss® bone granules. This led to increased invasion of osteosarcoma cell aggregates from within the bone-like matrix into the surrounding acellular bone marrow-like ECM. The presence of bone granules also yielded an atypical molecular profile of osteosarcoma cells, suggesting malignant metabolic reprogramming. Changes include decreased MMP-9 (p < 0.05) and increased PTEN (p < 0.05), MCP-1 (p < 0.01) and MCT-4 (p < 0.05) gene expression. This complex 3D biomimetic composition also changed cellular responses to doxorubicin, a common chemotherapeutic agent used to treat osteosarcoma, and reproduced key issues of in vivo treatment like drug penetrance and doxorubicin-induced bone toxicity. This work highlights the importance of a biomimetic matrix in 3D osteosarcoma models for both basic and translational research. STATEMENT OF SIGNIFICANCE: This study describes the generation of 3D osteosarcoma models termed "tumouroids", which are geometrically compartmentalised to reproduce the bone cancer mass and its environment. Utilising this novel model, specific parameters of osteosarcoma growth and invasion were investigated. Osteosarcoma cell lines proliferate at a slower rate, exhibit malignant metabolic reprogramming, and respond to drug intervention at lower concentrations of doxorubicin hydrochloride in matrix-complex compared to basic tumouroids. As such, this study provides evidence that the tumour microenvironment impacts osteosarcoma in many ways. The osteosarcoma tumouroid described herein may form the basis of a personalised-medicine strategy, which will allow the testing of drug effectiveness similar to that used for antibiotic selection for pathogenic bacteria.

A combined three-dimensional in vitro-in silico approach to modelling bubble dynamics in decompression sickness.

The growth of bubbles within the body is widely believed to be the cause of decompression sickness (DCS). Dive computer algorithms that aim to prevent DCS by mathematically modelling bubble dynamics and tissue gas kinetics are challenging to validate. This is due to lack of understanding regarding the mechanism(s) leading from bubble formation to DCS. In this work, a biomimetic in vitro tissue phantom and a three-dimensional computational model, comprising a hyperelastic strain-energy density function to model tissue elasticity, were combined to investigate key areas of bubble dynamics. A sensitivity analysis indicated that the diffusion coefficient was the most influential material parameter. Comparison of computational and experimental data revealed the bubble surface's diffusion coefficient to be 30 times smaller than that in the bulk tissue and dependent on the bubble's surface area. The initial size, size distribution and proximity of bubbles within the tissue phantom were also shown to influence their subsequent dynamics highlighting the importance of modelling bubble nucleation and bubble-bubble interactions in order to develop more accurate dive algorithms.

Quantification of cell-bubble interactions in a 3D engineered tissue phantom.

Understanding cell-bubble interactions is crucial for preventing bubble related pathologies and harnessing their potential therapeutic benefits. Bubbles can occur in the body as a result of therapeutic intravenous administration, surgery, infections or decompression. Subsequent interactions with living cells, may result in pathological responses such as decompression sickness (DCS). This work investigates the interactions that occur between bubbles formed during decompression and cells in a 3D engineered tissue phantom. Increasing the tissue phantoms' cellular density resulted in decreased dissolved O2 (DO) concentrations (p = 0.0003) measured using real-time O2 monitoring. Direct microscopic observation of these phantoms, revealed a significant (p = 0.0024) corresponding reduction in bubble nucleation. No significant difference in growth rate or maximum size of the bubbles was measured (p = 0.99 and 0.23). These results show that bubble nucleation is dominated by DO concentration (affected by cellular metabolism), rather than potential nucleation sites provided by cell-surfaces. Consequent bubble growth depends not only on DO concentration but also on competition for dissolved gas. Cell death was found to significantly increase (p = 0.0116) following a bubble-forming decompression. By comparison to 2D experiments; the more biomimetic 3D geometry and extracellular matrix in this work, provide data more applicable for understanding and developing models of in vivo bubble dynamics.

Injectable system for spatio-temporally controlled delivery of hypoxia-induced angiogenic signalling.

While chronically ischaemic tissues are continuously exposed to hypoxia, the primary angiogenic stimulus, they fail to appropriately respond to it, as hypoxia-regulated angiogenic factor production gradually undergoes down-regulation, thus hindering adaptive angiogenesis. We have previously reported on two strategies for delivering on demand hypoxia-induced signalling (HIS) in vivo, namely, implanting living or non-viable hypoxic cell-matrix depots that actively produce factors or act as carriers of factors trapped within the matrix during in vitro pre-conditioning, respectively. This study aims to improve this approach through the development of a novel, injectable system for delivering cell-free matrix HIS-carriers. 3D spiral collagen constructs, comprising an inner cellular and outer acellular compartment, were cultured under hypoxia (5% O2). Cell-produced angiogenic factors (e.g. VEGF, FGF, PLGF, IL-8) were trapped within the nano-porous matrix of the acellular compartment as they radially diffused through it. The acellular matrix was mechanically fragmented into micro-fractions and added into a low temperature (5 °C) thermo-responsive type I collagen solution, which underwent a collagen concentration-dependent solution-to-gel phase transition at 37 °C. Levels of VEGF and IL-8, delivered from matrix fractions into media by diffusion through collagen sol-gel, were up-regulated by day 4 of hypoxic culture, peaked at day 8, and gradually declined towards the baseline by day 20, while FGF levels were stable over this period. Factors captured within matrix fractions were bioactive after 3 months freeze storage, as shown by their ability to induce tubule formation in an in vitro angiogenesis assay. This system provides a minimally invasive, and repeatable, method for localised delivery of time-specific, cell-free HIS factor mixtures, as a tool for physiological induction of spatio-temporally controlled angiogenesis.

First implantable device for hypoxia-mediated angiogenic induction.

Delayed or inadequate vascularisation is one of the major factors leading to tissue infarction and poor graft survival. Current vascularisation strategies that rely on delivering single growth factors have proved ineffective or hard to control in practise. An alternative approach has been identified by this group that relies on stimulation of physiological angiogenic factor cascades by engineering local cell-hypoxia, within a nano-fibrillar collagen material. Here we report on a novel, practical and effective implantable device for delivering engineered angiogenic signalling, on demand. Human dermal fibroblast-seeded dense-collagen depots were pre-conditioned under physiological cell-generated hypoxia to up-regulate production of key angiogenic factors, including HIF1α and VEGF(165). The level of VEGF(165) protein retained within depots (indicating general angiogenic factor production) was directly correlated to the duration of pre-conditioning. Angiogenic factor delivery from pre-conditioned, non-viable depots rapidly induced an angiogenic response within endothelial cell-seeded constructs in vitro, while implanted acellular 3D constructs incorporating such angiogenic depots in their core were infiltrated with perfused vessels by 1 week in vivo, at which stage non-angiogenic implants were minimally perfused. Depot stability, tuneability of cell/matrix composition with long clinical experience of the collagen material, together with cost effectiveness, make this angiogenic therapy a promising addition to a clinician's tool kit for improving local tissue perfusion.

Switching off angiogenic signalling: creating channelled constructs for adequate oxygen delivery in tissue engineered constructs.

A major question in biomimetic tissue engineering is how much of the structure/function of native vasculature needs to be reproduced for effective tissue perfusion. O2 supplied to cells in 3D scaffolds in vitro is initially dependent upon diffusion through the scaffold and cell consumption. Low O2 (3%) enhances specific cell behaviours, but where O2 is critically low (pathological hypoxia) cell survival becomes compromised. We measured real-time O2 in 3D scaffolds and introduced micro-channelled architecture to controllably increase delivery of O2 to cells and switch off the hypoxic response. Simple static micro-channelling gives adequate perfusion and can be used to control cell generated hypoxia-induced signalling.

Controlling physiological angiogenesis by hypoxia-induced signaling.

The full sequence of signals leading to new blood vessel formation is a physiological response to tissue hypoxia through upregulation of angiogenic factor cascades. Controlled initiation of this mechanism for therapeutic/engineered angiogenesis must rely on precisely localized hypoxia. Here we have designed a 3D in vitro model able to test the effect and predictability of spatially positioned local hypoxic stimuli using defined cell depots within a 3D collagen matrix. Cell-mediated hypoxia was engineered using human dermal fibroblasts (HDFs), to generate a local population of Hypoxia-Induced Signaling (HIS) cells. HIS cell depots released angiogenic factors which induced directional endothelial cell (EC) migration and tubule formation in a spatially defined assay system. Non-hypoxic baseline control cultures induced minimal EC migration with little tubule formation. Furthermore, depots of HIS cells, positioned in the core of 3D collagen constructs directed host vessel in-growth deep into the implant by 1 week, which was at least 7 days earlier than in non-hypoxia pre-conditioned constructs. The functionality of in vivo vascularisation was verified by real-time monitoring of O2 levels in the core of implanted constructs. These findings establish the angiogenic potential of HIS cells applicable to in vitro tissue modeling, implant vascularization and engineering predictable angiogenic therapies.

Spatially defined oxygen gradients and vascular endothelial growth factor expression in an engineered 3D cell model.

Tissue hypoxia results in rapid angiogenesis in vivo, triggered by angiogenic proteins, including vascular endothelial growth factor (VEGF). Current views of tissue viability are founded on whether 'deeper-lying' cells receive sufficient nutrients and oxygen for normal activity and ultimately survival. For intact tissues, levels of such essential nutrients are governed by micro-vascular perfusion. However, there have been few effective quantitatively defined 3D models, which enable testing of the interplay or interdependence of matrix and cell density, and path diffusion on oxygen consumption in vitro. As a result, concepts on cell vulnerability to low oxygen levels, together with the nature of cellular responses are ill defined. The present study has adapted a novel, optical fibre-based system for in situ, real-time oxygen monitoring within three-dimensionally-spiralled cellular collagen constructs, which were then unfurled to enable quantitative, spatial measurements of VEGF production in different parts of the same construct exposed to different oxygen levels. A VEGF response was elicited by cells exposed to low oxygen levels (20 mmHg), primarily within the construct core.

3-D in vitro model of early skeletal muscle development.

An understanding of the mechanical and mechano-molecular responses that occur during the differentiation of mouse C2C12 [corrected] myoblasts in 3-D culture is critical for understanding growth, which is important for progress towards producing a tissue-engineered muscle construct. We have established the main differences in force generation between skeletal myoblasts, dermal fibroblasts, and smooth muscle cells in a 3-D culture model in which cells contract a collagen gel construct. This model was developed to provide a reproducible 3-D muscle organoid in which differences in force generation could be measured, as the skeletal myoblasts fused to form myotubes within a collagen gel. Maintenance of the 3-D culture under sustained uni-axial tension, was found to promote fusion of myoblasts to form aligned multi-nucleate myotubes. Gene expression of both Insulin Like Growth Factor (IGF-1 Ea) and an isoform of IGF-1 Ea, Mechano-growth factor (IGF-1 Eb, also termed MGF), was monitored in this differentiating collagen construct over the time course of fusion and maturation (0-7 days). This identified a transient surge in both IGF-1 and MGF expression on day 3 of the developing construct. This peak of IGF-1 and MGF expression, just prior to differentiation, was consistent with the idea that IGF-1 stimulates differentiation through a Myogenin pathway [Florini et al., 1991: Mol. Endocrinol. 5:718-724]. MGF gene expression was increased 77-fold on day 3, compared to a 36-fold increase with IGF-1 on day 3. This indicates an important role for MGF in either differentiation or, more likely, a response to mechanical or tensional cues.