The use of biomaterials for orthopedic implants has significantly increased over the past decades, offering promising solutions as bone substitutes. However, challenges such as implant-associated infections and aseptic loosening remain, as biomaterials trigger an immune response upon implantation. Recent insights have highlighted the crucial role of this immune response in bone regeneration, shifting the focus toward understanding the dynamic interplay between immune cells and osteogenic cells. Macrophages, as key regulators of inflammation and bone regeneration, closely interact with human mesenchymal stem cells (hMSCs), influencing their cellular behavior. As a result, there is increasing interest in designing biomaterials that simultaneously support osteogenesis and modulate immune responses to facilitate enhanced implant integration. However, most studies investigating 3D-printed biomaterials and their effects on hMSCs and macrophage behavior rely on monoculture or indirect co-culture models, thereby neglecting direct cell-cell interactions. This limitation creates a gap between in vitro models and the in vivo environment.
This study addresses this gap by developing a direct co-culture model of hMSCs and THP-1-derived M0 macrophages on 3D-printed Ti-6Al-4V biomaterials to investigate their interactions at the implant-tissue interface. By examining cell adhesion, morphology, and cytokine secretion, this research provides insights into how direct cell-cell interactions and biomaterial properties influence early-stage cellular responses.

To interpret the results of the co-culture model, both cell types were first characterized in monoculture to assess their proliferation, adhesion, and morphology affected by the titanium substrate using immunofluorescence staining and scanning electron microscopy (SEM). In addition, TNF-a and IL-6 cytokine secretion by M0 macrophages was assessed via Enzyme-Linked Immunosorbent Assay (ELISA) to evaluate macrophage polarization. Following this, the mixed culture medium was optimized, and seeding densities and ratios were determined to ensure the viability of both the hMSCs and macrophages in co-culture. M0 macrophages were seeded first to allow attachment before introducing hMSCs, mimicking in vivo conditions where macrophages are the first to arrive at the implant site. Co-culture effects on cell morphology and cytokine secretion were analyzed using immunofluorescence staining, SEM, and ELISA. Additionally, hMSCs were cultured on both dense and porous 3D-printed Ti-6Al-4V substrates in monoculture to assess the effects of substrate porosity on early-stage cell adhesion and morphology.

The findings of this study showed the significant influence of the 3D-printed Ti-6Al-4V substrate on cell morphology, suggesting that it can override the effects of cell-cell interactions and paracrine signaling. As a result, macrophages adopted a more pro-inflammatory morphology, while hMSCs exhibited a more spread-out shape, which may be indicative of early osteogenic differentiation. In addition, cytokine secretion profiles in monoculture showed a trend toward an M1-like macrophage phenotype when the cells were cultured on titanium. In contrast, co-culture conditions led to a shift toward a more pro-repair environment, characterized by reduced TNF-asecretion and increased IL-6 production. This suggests that hMSCs modulate the macrophage response toward a more pro-repair phenotype.
A comparison between dense and porous 3D-printed Ti-6Al-4V substrates revealed no statistically significant differences in hMSC morphology after 7 days, indicating that substrate geometry has only a minor effect on early-stage cell adhesion and morphology.
Overall, these findings highlight the potential of the developed co-culture model for studying osteoimmunomodulation on titanium biomaterials, contributing to bridging the gap between in vitro models and in vivo conditions.

The rising demand for orthopedic implants, driven by rising life expectancy and medical advancements, highlights the need for improved implant-tissue integration. Initial cellular interactions, such as adhesion, migration, and differentiation, are critical for implant success, with surface topography playing a key role. However, existing methods for analyzing focal adhesions (FAs) and cell behavior are inconsistent, often relying on manual or proprietary tools that limit reproducibility. This study addresses these limitations by developing an automated, open-source, and multiparametric workflow for high-throughput 3D analysis of cell morphology and FA dynamics, enabling standardized quantification of cellular responses to biomaterial surfaces. This project aimed to create such a workflow for analysis of cellular characteristics. It was then applied to study cell behavior on titanium surfaces, polished titanium (pTi) and dry-etched titanium (deTi). After cell culturing, fixation, immunostaining, high-resolution imaging using spinning disk confocal microscopy was used to quantify parameters, this included cell and nucleus area, aspect ratio (AR), nucleus height, volume, and FA characteristics (area and number) using FIJI/ImageJ. Validation against manual quantification confirmed its accuracy and efficiency, reducing analysis time by at least fourfold. This workflow was applied to analyze MC3T3-E1 preosteoblasts and human mesenchymal stem cells (hMSCs) on pTi and deTi surfaces. Results showed distinct differences, with hMSCs demonstrating greater adaptability to rougher deTi surfaces, while MC3T3-E1 cells remained more confined. Statistical analysis via two-way ANOVA confirmed surface type as the primary factor influencing cell behavior, with time and surface-time interactions varying in significance across parameters. These results suggested that hMSCs adapt more easily to rougher surfaces, as evidenced by increased spreading, more organized actin structures, and distinct FA formation, highlighting their potential for biomaterial applications due to their ability to dynamically remodel their shape and focal adhesions, which may enhance mechanosensing and tissue integration. The developed workflow offers a valuable tool for advancing research into biomaterial cell interactions and optimizing implant design for improved osseointegration. Additionally, it can be expanded to investigate a broader range of biomaterials and cell types, further enhancing its applicability in tissue engineering and regenerative medicine.

The physical, chemical, and biological versatility and high water content of hydrogels have been taken advantage of to fabricate hydrogels that mimic the extracellular matrix (ECM) structure of native tissues and applied in the field of cell and tissue engineering to provide the cells with a 3D scaffold structure. Especially in tissue engineering, the creation of a suitable hydrogel scaffold with hard and soft interfaces that is capable of both supporting cell growth and stimulation for guided differentiation for tissue interface studies is a primary concern for advancements in tissue engineering. The ideal scaffold with hard and soft interfaces can be achieved by finetuning its biological, biochemical, structural, and mechanical properties for optimal cell proliferation and differentiation.

In this thesis, a magneto-responsive hydrogel scaffold composed of gelatin (Gel, 2.5%), alginate (Alg, 5%), and iron oxide microparticles (10% w/v) was developed and mechanically and rheologically characterized before, during and after the application of a uniaxial static magnetic field. The magneto-responsive hydrogel scaffolds were created through multi-material 3D printing using magnetic and non-magnetic hydrogel inks. The magnetic inks contained magnetic particle (MP) inclusions within its polymer network while the non-magnetic hydrogel ink had no MPs. The 3D printing process allowed for a local control in the magnetic and non-magnetic hydrogel distribution to create hard and soft hydrogel interfaces. The printability and shape fidelity of various ink compositions were evaluated, so that the final composition of Gel:Alg ratio of 1:2 (2.5%:5.0%) with 10% MP was chosen for further mechanical and rheological characterization. The magnetic hydrogel scaffold exhibited magnetorheological properties as it mainly increased the effective Young’s modulus, storage modulus, and damping factor, and decreased viscosity under a uniaxial static magnetic field application.

In tissue engineering, the developed hydrogel scaffold, which is locally responsive to magnetic cues, shows great potential for creating scaffolds capable of continuously stimulating embedded cells in a non-contact manner. As a proof of concept, a bi-layered, multi-material hydrogel scaffold was created with increased surface area attachment points between each hydrogel material to mimic the osteochondral tissue interface. The increased surface area between both layers was achieved through a checkered pattern design with alternating magnetic and non-magnetic hydrogel sections printed alongside each other.

Modifying the surface topography of biomaterials is a promising approach to guide macrophages’ fate to ensure good integration of biomaterials in the host and reduce the risk of implant-associated infections (IAI). High-aspect-ratio titanium nanopillars have shown both bactericidal and osteoconductive properties; however, their effect on immunomodulation is limited researched. Understanding these immunomodulatory effects is important, as the immune response plays a key role in regulating bone regeneration and implant success.

This study investigates the behaviour of J774A.1 macrophages on nanopatterned Dry Etched Titanium (DETi) and polished Titanium (pTi) surfaces, with a focus on cell morphology, migration, proliferation, and cytoskeletal organization. Therefore, stimulated pro-inflammatory (M1) macrophages were cultured on the two substrates and on glass for 48 hours. For morphological evaluation of the cells, the samples were imaged by high resolution fluorescence microscopy after fixation and staining for actin, nucleus and vinculin. Migration was assessed on DETi and pTi samples by creating wounds on the seeded surfaces, measuring the wound closure time. Furthermore, live imaging of M1 macrophages on glass was performed to evaluate their migration and proliferation behaviour.

Liver transplantation remains the only definite cure for end-stage liver disease, yet the demand for donor livers far surpasses their supply, resulting in substantial wait-list mortality rates. This has led to the exploration of extended criteria donor (ECD) livers, which often exhibit compromised function and susceptibility to post-transplantation complications. Ex vivo normothermic machine perfusion (NMP), emerges as a promising approach to assess liver quality, extend preservation duration, and therefore reduce the supply-and-demand imbalance and post-transplantation complications. Moreover, long-term (> 24 hours) NMP could enable treatment, repair, and regeneration of liver grafts and serve as a valuable platform for drug testing and disease modeling. To achieve long-term NMP, it is crucial to recreate the in vivo physical conditions during ex vivo perfusion. The aim of this thesis was to incorporate liver movement during porcine liver NMP and assess the effect of liver movement on liver function and tissue integrity.

Therefore, a new liver reservoir including liver movement was developed and subsequently used to perform porcine liver NMP experiments. Slaughterhouse procured porcine livers (n = 4), were perfused via the hepatic artery (HA) and portal vein (PV) under the established movement condition for 360 minutes. After 120 and 300 minutes of perfusion, indocyanine green (ICG), a fluorescent dye, was dosed, and samples were taken from the arterial circulation and bile to study the clearance capacity of the liver. Hourly samples of the perfusate and bile were taken for blood gas analysis and measurement of injury markers, to assess the general viability and functionality of the liver. And tissue samples were taken at the end of perfusion to study tissue integrity.

Liver movement was established by alternatively inflating and deflating two balloons underneath the liver, with both inflation and deflation lasting 8 seconds. The 6-hour porcine liver NMP experiments under the established movement condition showed liver viability and functionality in terms of glucose metabolism, lactate and bilirubin clearance, stable levels of injury markers, and continuous bile production. The ICG clearance capacity of the liver showed to be improved, although not significantly, under the movement condition, with a mean perfusate disappearance rate (PDR$_\text{ICG}$) of 30.6 $\pm$ 11.7 \% per minute, compared to the static condition (11.0 $\pm$ 3.3 (Zeist) and 21.8 $\pm$ 13.8 (Leiden) \% per minute), after 300 minutes of NMP. The macroscopic appearance and histological analysis of the liver revealed some non-perfused areas on the bottom of the liver, but overall, the liver tissue was intact, and no major hepatocellular damage occurred after 5$-$7 hours of NMP under the movement condition.

A novel liver reservoir including liver movement was established to study the sole effect of movement during NMP. During the 6-hour porcine liver NMP experiments under the established movement condition, the livers showed proper clearance capacity and tissue integrity. Although inclusion of movement did not result in a significant improvement with respect to liver function, it is hypothesized that movement will prolong the viability and functionality of livers when performing NMP for more than 24 hours.

Surface micro- and nanotopographies have shown potential in inducing osteogenic differentiation. Titanium nanopillars can enhance osteogenic differentiation of human mesenchymal stromal cells (hMSCs). While the exact mechanisms by which osteogenic differentiation is induced remain unclear, mechanotransduction is thought to play an important role. This study investigated the effect of titanium nanopillars (bTi) on integrins, involved in the first phase of mechanotransduction, during the adhesion of human mesenchymal stromal cells (hMSCs). Titanium nanopillars were produced by inductively coupled plasma reactive ion etching (ICP RIE). The resultant nanopillars did not hinder cell adhesion but resulted in distinct cell morphologies, namely a predominant polygonal shape on the planar titanium control (fTi) and a stellate shape on bTi. A slower adaptation of cells on bTi relative to fTi was suggested by delayed cell spreading and enhanced filopodia development. RNA was isolated based on an optimised protocol developed in this study. Gene expression analysis revealed that all 10 integrin subunits involved in osteogenic differentiation were expressed by cells on both fTi and bTi during the 48 hours of culture. ITGA1 and ITGA4 were upregulated on bTi relative to fTi after 24 hours of culture. The upregulation of ITGA4 suggested increased migration of cells on bTi. Interestingly, ITGA2, ITGAV, and ITGB3 showed a peak at 24 hours on bTi. Immunocytochemical analysis was performed based on the protocol established in this study. The preliminary immunocytochemical analyses suggested fibrillar adhesions including integrin subunit α5 on fTi and both α2 and α5 on bTi. Further investigation of the mechanotransduction pathway may elucidate any correlation between integrin expression, actin reorganisation, and osteogenic differentiation of hMSCs cultured on such nanopillars and provide valuable insights for finding earlier osteogenic markers to be used for the evaluation of osteogenic nanotopographies.

Due to the limited availability of suitable livers for transplantation and the increasing use of suboptimal donor livers, which are more prone to ischemia-reperfusion injury (IRI), there is an urgent need for novel graft-ameliorating therapies. The promising capacities of (liver-derived) mesenchymal stromal cells ((L-)MSC) in mitigating IRI and their integration with hypothermic oxygenated machine perfusion (HOPE) present a novel paradigm for augmenting graft quality. However, still little research exists into the effects on the liver of (L-)MSC administered during HOPE. Moreover, human livers for research are scarce and, therefore, studies to show safety and efficacy are challenging. To address these challenges, this study investigated the use of porcine precision-cut liver slices (PCLS) as a tool to model the administration of MSC during HOPE. First, optimisation of PCLS procurement and culture was undertaken, exploring methods such as the use of well inserts and different culture media to enhance the retained structural integrity of the slices. Subsequently, a 24-hour co-culture experiment of porcine PCLS (n=3 livers, n=135 PCLS) and L-MSC was conducted, where L-MSC were added at the beginning of a two-hour hypothermic oxygenated culture. The findings of this experiment demonstrated the ability of MSC to adhere to the liver tissue during the two hours in a hypothermic oxygenated environment. Additionally, the effect of MSC on PCLS viability, morphology, cell death, regeneration, liver injury markers, functionality, and gene expression profiles was investigated. While no significant differences between co-cultured and control PCLS were observed, there was a clear trend towards increased inflammation, cell damage, and cell death in the co-cultured slices. Overall, this research contributes to the understanding of the potential implications of (L-)MSC administration during hypothermic oxygenated perfusion and highlights the utility of PCLS as an intermediate experimental model for studying transplant-related IRI.

Vascularisation is a critical factor in large-scale tissue engineering, as it supports the survival of cells in large tissue constructs. In native tissue, a hierarchical vascular network is present where larger vessels transport nutrients and oxygen throughout the body, and smaller capillaries facilitate blood-tissue exchange. In large-scale engineered tissue, such a functional hierarchical vascular network is necessary for cell survival. In this study, we aimed to create a high-density capillary network using biocompatible hydrogels for tissue engineering purposes.

To achieve this, we first synthesised alginate-tyramine (ATA) by modifying the alginate backbone to contain tyramine groups. Then, we created monodisperse ATA microgels through in-air microfluidics (IAMF). These microgels were in turn used to form a microporous annealed particle (MAP) structure through photocrosslinking, which resulted in a structure containing the desired high-density microporous network.

Next, we showed that we could use embedded 3D (emb3D) printing techniques to fabricate channels in the ATA hydrogel, which we stabilised through photocrosslinking. These macroscale channels proved to be open and perfusable.

To create a stable interface between the microgels and hydrogel, we pre-saturated the microgels with a polymeric solution that forms an aqueous two-phase system (ATPS) with the ATA hydrogel. The ATPS and pre-saturation successfully counteracted capillary forces, keeping the hydrogel outside the microgel compartment. At the same time, the ATPS allowed the hydrogel and microgels to be in contact with each other, at which interface photocrosslinking was allowed to establish a connected structure.

By combining these approaches, we were able to successfully assemble a perfusable hierarchical porous network with micro- and macroscale vasculature. Specifically, we fabricated a macroscale channel using emb3D printing methods in a bath with both hydrogel and microgel components, which we then solidified through photocrosslinking. The resulting MAP was perfused through the macroscale inlet with fluorescent particles, demonstrating successful perfusion through the printed channel.

To date, the fabrication and perfusion of a multiscale vascular tree have not been demonstrated in this manner. Our study demonstrates the possibility for multiscale vascular emb3D printing through a microporous bath and highlights the potential of this innovative approach to overcome current limitations in tissue engineering.

Cholangiocarcinoma (CCA) is a rare but aggressive type of primary liver cancer with a dismal prognosis. CCA commonly metastasizes to the lungs and lymph nodes, significantly reducing overall survival. However, a mechanistic understanding of how CCA invades these metastatic sites remains lacking. This is partly due to the failure of current models to recapitulate the complexity of tissue-specific environments for metastatic CCA. Patient-derived tumor organoids are promising models for studying cancer in vitro, including CCA. However, organoids are frequently cultured in basement membrane extract (BME), which does not represent the native microenvironment. Decellularized scaffolds can serve as a culture environment for investigating tumor-extracellular matrix (ECM) interactions. Nonetheless, decellularized scaffolds are often utilized in conjunction with cancer cell lines, lacking critical characteristics of a growing tumor in vivo. Therefore, this study aims to decellularize and characterize human-derived lung (n=3) and lymph nodes (n=17), and recellularize these with patient-derived CCA organoids (CCAOs) (n=3) to establish an in vitro model for studying the interaction between epithelial tumor cells and the microenvironment of the metastatic site. Decellularization resulted in acellular scaffolds with preserved ECM components. Local ECM heterogeneity was shown by the macro- and micro-scale mechanical properties as determined by rheology and micro-indentation. The CCAOs showed adherence and growth when combining them with the decellularized lung ECM (dLECM). The metabolic activity of CCAOs was diminished when cultured in dLECM compared to those cultured in BME, and was dictated both by the organoid line and the dLECM donor. Distinct CCAO gene expression profiles were observed between the two culture environments (BME and dLECM). This elucidates the effect of the culture environment on the behavior of CCAOs in an in vitro model. In conclusion, the convergence of CCAOs with their organ-specific location of metastasis obtained by decellularization, provides a valuable tool for integrating the ECM of the metastatic location in an in vitro model for the mechanistic study of cancer metastasis.

That mechanical properties of the extracellular environment can influence cell behaviour has already been established. Recent studies indicate that human mesenchymal stem cells are affected by the surface curvature of the underlying substrate. However, how meso-scale substrate curvature affects cell behaviour is still not clear. This study utilised the finite element method to simulate a prestressed human mesenchymal stem cell conforming and attaching to a flat control substrate, concave hemispherical and concave cylindrical substrates with curvature radii from 300 µm to 75 µm. The cell model comprises the actin cortex, cytoskeleton, and nucleus modelled with a hyperelastic material definition and 30 linear elastic stress fibres prestressed with a force of 10 nN. The effects of surface curvature on cellular traction forces, cell height, nuclear aspect ratio and actin cortex was studied. The vertical traction forces were observed to be 70% and 40% lower for the hemispherical and cylindrical substrates of the highest curvatures compared to the flat control substrate, respectively. Cellular traction forces towards the cell periphery were roughly 10-20% higher than the more central cellular traction forces independent of substrate curvature. Stresses in the actin cortex were observed to increase by 290% and 220% from the flat control substrate to the hemispherical and cylindrical substrates of the highest curvatures, respectively. These results indicate that the cell is more sensitive to hemispherical substrates than cylindrical substrates. The results also support in vitro observations where hMSCs are seen to span hemispherical substrates and avoid continuous contact.

Delayed healing or non-union of bone fractures are still clinically and economically relevant problems. Bone healing includes pro- and anti-inflammatory phases, which proved crucial for successful healing. Being able to modulate these inflammatory responses could potentially promote bone fracture healing. Dendritic cells play important roles in the immune response after bone fracture, due to their immune regulating abilities via cytokine secretion and antigen presentation. Therefore, DCs are a target for immunomodulatory therapies. One way of modulating cell behaviour can be altering osmolality in cell microenvironments. This thesis investigates in vitro if environmental osmolality can modulate the inflammatory phenotype of monocyte-derived dendritic cells (moDCs) in 2D and 3D microenvironments.
To exclude adverse effects such as cell death by hypo- or hyper-osmotic conditions, a range was established where moDCs maintain high viability and metabolic activity. Hypo-osmotic medium was established by diluting iso-osmotic cell medium with deionized water. Hyper-osmotic medium was established in two ways, by supplementing iso-osmotic cell medium with either ionic sodium chloride (NaCl) or inert polyethylene glycol (PEG). moDCs maintained high viability and metabolic activity between 210-385 mOsm/kg for NaCl and PEG. The conditions for the hypo- iso- and hyper-osmotic media for all cell culture experiments were defined as 220, 281 and 381 mOsm/kg, respectively.
In 2D cell culture, hypo- or hyper-osmolality did not stimulate a significant change in surface marker expression of activation markers CD40, CD80, CD83, CD86, CD197 and HLA-DR compared to the non-activated control. This suggests that osmolality alone was not sufficient to modulate the inflammatory phenotype. Hereafter, the effect of osmolality on maturation with Lipopolysaccharide (LPS) was investigated. Exposure to hypo- or hyper-osmolality in 2D moDC culture with LPS as maturation factor, showed no significant difference in expression of CD40, CD80, CD83, CD86, CD197 and HLA-DR compared to the activated control. However, a trend was observed in hyper-osmotic medium with PEG and NaCl, where moDCs showed a decrease of CD40 and HLA-DR expression in the PEG condition and a decrease of CD197 in the NaCl condition. This suggests that hyper-osmolality could attenuate the activating capacity of LPS and depending on the osmolyte, can potentially decrease T cell activation or migration of moDCs. Functional analyses such as mixed lymphocyte reaction and migration assays may elucidate if hyper-osmolality can affect moDC T cell activation and migration. For cell culture in 3D microenvironment, the mechanical properties of fracture hematoma were modelled using alginate hydrogels to gain an understanding of the behaviour of DCs during bone healing. Low molecular weight alginate was used to mimic the stress relaxation behaviour. Stiffness was tuned trough ionic crosslinking concentration. The resulting gels had a stiffness of 8±0.3 kPa and stress-relaxation half-time of 199.7±10.8 s. In these 3D microenvironments during LPS maturation, hypo- or hyper-osmolality did not show a significant change in expression of CD40, CD80, CD83, CD86, CD197 and HLA-DR compared to the activated control. This suggests that moDCs were not sensitive to osmolality during maturation with LPS in 3D microenvironments. Comparing the outcomes of 3D to 2D cell culture, moDCs appeared less sensitive to changes in environmental osmolality during LPS maturation in 3D than on 2D substrates.

Overall, for both the 2D and 3D studies, a larger cohort of donors is needed to improve confidence in the results. Based on the trends observed, the outcome of the study does not indicate that osmolality could be used in 3D microenvironments to modulate the inflammatory phenotype of moDCs, as activation markers remained largely unaffected. In 2D cell culture, hyper-osmolality could modulate the inflammatory phenotype of moDCs during maturation with LPS through downregulation of activation markers. Osmolality should be considered when designing immunomodulatory treatments, as osmolality might affect DC maturation in 2D cell culture.

To this day, liver transplantation is the only effective treatment for end-stage liver failure. Unfortunately, the scarcity of liver donors leads to waitlist mortality, pushing the need for alternative treatment options. In vitro models are essential tools in research on alternatives for transplantation. One of the promising models is the hepatobiliary organoid model because these three-dimensional cultures model (elements of) the native tissue structure and function. However, a limitation of these models is that they currently resemble only one liver cell type. The aim of this research was therefore to combine different cell types, i.e., biliary organoids and mesenchymal stromal cells (MSCs), in one culture model.

The organoids were obtained from liver-biopsy-derived intrahepatic biliary epithelium and are therefore named intrahepatic cholangiocyte organoids (ICOs). The liver-specific MSCs (L-MSCs) were isolated from perfusion fluid of donor livers collected at liver transplantation procedures. Decellularized liver tissue was used to include extracellular matrix (ECM) in the co-culture models. To assess the most optimal culture conditions, three experimental setups were tested. In the first setup, ICOs were cultured in different concentrations of L-MSC derived conditioned medium (CM) in commercially available basement membrane extract (BME) (1A). This setup was also used to differentiated the ICOs towards hepatocytes (1B). In the second setup, various concentrations of ICOs and L-MSCs were cultured in an indirect (2A) and direct (2B) BME model. The third setup included liver-derived ECM to replace the BME. Cells were cultured similar to model 2B (3A) and ICOs were added after a pre-culture of L-MSCs (3B). The morphology of both cell types, the visible interaction between cell types, and the expression of hepatocyte, cholangiocyte, MSC and proliferation markers were analyzed.

The effect of L-MSCs and L-MSC-derived CM on ICO formation, morphology, and growth patterns, was small. Direct cell-cell contact was rare in the BME co-cultures, but instances of close contact between two cell types were observed in both the indirect (2A) and direct (2B) setups. In the ECM models (3A-B), ICO cells formed polarized monolayers that pushed the L-MSCs aside, indicating that the cell types did not mix well. ICOs expressed the cholangiocyte markers cytokeratin (CK) 7 and 19 in both mono- and co-cultures, whereas L-MSCs were most likely CK7 and CK19 negative. Variation in gene expression levels of hepatocyte, cholangiocyte, MSC, and proliferation markers was observed, but no statistically significant differences were found between conditions. Nevertheless, the down-regulatory effect of CM on the expression of hepatocyte markers in differentiated ICOs showed a consistent trend. Additionally, the presence of L-MSCs resulted in a higher mean expression of MSC markers CD90, CD105 and Vimentin, and proliferation marker KI-67.

Although no clear effects of L-MSCs on ICO growth were observed in this study, the created co-culture models form a promising base for future research. The models include adult human liver-derived components and are therefore a step towards the reconstruction of the hepatic microenvironment. In addition, if other (liver-derived) cell types were to be introduced to the co-culture models, L-MSCs have the potential to enhance the cell function of these new cells. Last, the models can be used as a base for future ICO co-culture studies, since the L-MSCs can be replaced by other cell types.

Background: Despite the considerable development of the field of orthopedic implants in the past century, complications including poor bone ingrowth and implant associated infection (IAI) persist to this day, causing a huge burden to millions of patients and the healthcare systems worldwide. Additive manufacturing (AM) and the subsequent surface biofunctionalization and antibacterial element incorporation are promising techniques to achieve dual antibacterial and osteogenic functionalities within a bone implant. In addition, macrophages are an immune cell type that are known to be essential for the implant success in the body, by having an intimate crosstalk with mesenchymal stem cells (MSCs) in the process of new bone formation. However, the behaviour of these immune cells is affected by environmental cues, including the implant surface properties. Therefore, this work investigated for the first time the effects of AM titanium implants biofunctionalized via plasma electrolytic oxidation (PEO) and incorporated with silver nanoparticles (AgNPs) on the human mesenchymal stem cells (hMSCs) co-cultured in vitro with human macrophages. Specifically, the paracrine effects of immune cells on the hMSCs osteogenic differentiation were studied by the development of an indirect co-culture model.
Materials and methods: AM Ti-6Al-4V implants were biofunctionalized via PEO and AgNPs incorporation and the surface characterization was performed by a scanning electron microscope (SEM) and ion release analysis. The effects of such implants on the hMSCs osteogenic differentiation in vitro were subsequently evaluated by measuring the mineralization and osteogenic gene expression. In addition, a macrophage-hMSCs indirect co-culture system was developed in order to study the effects of the macrophage polarization induced by the implants on the hMSCs osteogenic differentiation by means of a paracrine communication. The macrophage polarization was characterized by measuring the cytokine secretion pattern with an ELISA assay and gene expression.
Results: PEO modification of AM implants created TiO₂ surfaces with interconnected micro and nanoporosities and the incorporation of AgNPs. The single-culture of hMSCs on PEO and PEO + Ag implants revealed their ability to promote the osteogenic differentiation and no detrimental effects were observed in this process by the presence of AgNPs. The immunological evaluation of the co-culture system revealed similar polarization patterns when macrophages were cultured on PEO and PEO + Ag surfaces. In addition, factors secreted by polarized macrophages did not elicit a paracrine effect on the co-cultured hMSCs, neither enhancing nor inhibiting their osteogenic differentiation.
Discussion and conclusions: Based on the findings gathered in this study, the incorporation of AgNPs in the PEO layers, under the conditions used in this work, did not compromise the osteogenic differentiation and mineralization of the hMSCs. In addition, PEO + Ag surfaces did also not cause any detrimental effects on the paracrine communication of human macrophages on hMSCs. Therefore, further investigations regarding the osteoimmunomodulatory potential of these biofunctionalized AM implants are worth performing, in an attempt to achieve an important additional biofunctionality next to their proven osteogenic and antibacterial activity. These future researches should include further optimization of the PEO surface morphology and chemistry as well as the co-culture model used in this study.

Studies have shown that pillars of specific dimensions and spatial arrangement can promote osteogenic differentiation of stem cells and kill bacteria that cause infections. Other studies have shown that surface curvature can also serve as a mechanical cue to modulate cell behavior. Therefore, the fabrication of pillars on curved surfaces would allow researchers to investigate the synergistic effect of pillars and curvature on cell behavior. In this project, a process was developed based on thermal nanoimprint lithography (TNL) and dry etching techniques for the fabrication of pillars into curved substrates made of hard materials. To this aim, a fused silica specimen containing sub-micron pillars was used as the master mold in a molding process to replicate the pillars as pits into a hybrid polydimethylsiloxane (PDMS) mold. The dimensions and morphology of the replicated patterns were assessed using a scanning electron microscope (SEM). Thereafter, TNL was employed to imprint the pits of the hybrid PDMS mold on the surface of the desired planar/ curved substrates. Finally, etching processes were employed to transfer the patterns into the bulk of the substrates. The process was first developed on planar fused silica substrates and then, on curved substrates. The interspace of the resultant pillars on the planar substrates was no more than 3% different than the interspace of the original pillars on the master mold. The diameter was also close to the values of the diameter of the original pillars (the maximum difference was 23%). The height of the pillars differed slightly (mo more than 16%) for the specific process conditions that were used. On the curved substrates, the interspace increased by 5% and 10%, and the diameter by 57% and 11%.

Cartilage degeneration is a major cause of chronic disability and its treatment usually involves highly invasive procedures. To address this issue, three-dimensional (3D) bioprinting was introduced as a promising tissue engineering approach. However, the structures fabricated with this method are static and lack the dynamic response of native tissues. Moreover, the fabrication of curved or tubular structures remained challenging, especially in the case of soft tissues. Four-dimensional (4D) bioprinting is a newly emerged, next generation biofabrication technology capable of resolving the aforementioned challenges.
In this thesis, an advanced 4D bioprinting method based on multi-material extrusion of two hydrogel-based (bio)inks is reported. This approach allows the fabrication of bilayered scaffolds made from a bottom part of a HA-Tyr precursor and a composite hydrogel top part from alginate and HA-Tyr (Alg/HA-Tyr). The scaffolds demonstrated self-bending upon immersion in aqueous solutions that was driven by the different swelling capacity of the two inks. Compatibility of the inks with extrusion printing was verified by rheological characterization. Control of the obtained curvature was achieved by tuning the infill density, printing angle and layer thickness. Moreover, the effect of crosslinking time as well as the influence of the swelling solvent type in the degree of bending was also investigated. Finally, human mesenchymal stem cells (hMSCs) were mixed with the Alg/HA-Tyr ink and self-bending living scaffolds were bioprinted. The shape-shifted scaffolds were capable of supporting a high cell viability for up to 14 days and remained viable for up to 4 weeks.

Bioprinting is a promising technique that has the ability to generate complex tissue structures for articular cartilage (AC) repair and regeneration. However, most biomaterials used for bioprinting are incapable of representing the complexity of native extracellular matrix (ECM). In this study, a novel cartilage specific bioink has been developed with tyramine modified hyaluronic acid (THA) and decellularized ECM (dECM) of AC for threedimensional bioprinting, combining the printable qualities of THA with AC dECM to create an optimized microenvironment. First, the influence of increasing concentrations of
AC dECM on the rheological and printability properties was assessed. Next, human mesenchymal stem cells (hMSCs) were incorporated into bioinks composed of 2.5% w/v THA and AC dECM (0.0%-1.0% w/v). The THA-AC dECM bioinks have been successfully bioprinted into stable scaffolds with viable cells and were cultured for 28 days in chondrogenic differentiation medium. Subsequently, the effect of combining AC dECM with THA was evaluated by measuring the cell viability, cell differentiation and AC tissue deposition. The ink compositions with AC dECM (0.5% and 1.0% w/v) caused an increase in ink viscosity, yield stress and damping factor compared to the THA ink, but still showed shear thinning behavior and provided good printability. Spread cell morphology was observed for the scaffolds with addition of AC dECM compared to AC dECM free control with only THA, indicating enhanced cell adhesion to the hydrogel network. Moreover, a higher production of sulfated glycosaminoglycans (sGAGs) was observed for the scaffolds where AC dECM was added. The THA and AC dECM bioink composition showed promising results for AC tissue engineering and warrants further research into the development of bioinks with tissue-specific inductive properties.

Three-dimensional (3D) bioprinting studies the initial condition of printed constructs, which makes construct static, and neglects the dynamic changes in natural tissue conformation during tissue repair, disease progression and regeneration processes. Therefore, the concept of four-dimensional (4D) bioprinting has been introduced that allows for the fabrication of scaffolds that can transform in time. The extracellular matrix stiffness is one component involved in the in vivo dynamic microenvironment of tissues such as cartilage. Nonetheless, the implementation of dynamic stiffness in in vitro 3D cell culture has not been intensively investigated yet, in particular not in 4D bioprinting for cartilage applications. Therefore, the aim of this study is to create a method for the temporal stiffening of a human mesenchymal stem cell-laden scaffold in 4D bioprinting for cartilage applications. To do this, inks made of tyramine-modified hyaluronic acid (THA), and THA combined with sodium alginate (SA) (THA-SA) were studied on their ability to stiffen days after being printed through photocrosslinking or ionic crosslinking (secondary crosslinking). The method that generated the greatest increase in stiffness was evaluated on the effect of time on stiffening and the effect of stiffening on the swelling behaviour and the internal structure of the scaffold. Next, the 4D printing method was translated to 4D bioprinting and the influence of the temporal stiffening method on the mechanical properties, cell viability, gene expression and matrix deposition was determined. This study demonstrated that in 4D printing, secondary photocrosslinking of THA and secondary ionic crosslinking of THA-SA resulted into scaffolds with a significant increase in stiffness. The method for secondary ionic crosslinking of THA-SA (2.5% w/v THA, 1.5% w/v SA) scaffolds achieved the largest increase in stiffness and was able to temporally stiffen at day 1, 3 or 5. This increase in stiffness was associated with a decrease in swelling ratio and a maintained internal microstructure. For the translation to 4D bioprinting, the method was adjusted for cell culture conditions and resulted in an increase in stiffness. However, the stiffening was not yet reproducible and therefore optimalization of the method will be required. Nevertheless, in the first week of cell culture, the bioprinted scaffolds achieved high cell viability and an equal cell distribution. At day 7, the gene for collagen type II was not expressed and cartilage-specific matrix deposition was not detected. With future studies focusing on optimizing the introduced 4D bioprinting protocol, this novel method for the ionic secondary crosslinking of THA-SA may function as a platform for studying the effect of temporal matrix stiffening on the behaviour of the embedded cells in fundamental research or as a stimulus for tissue formation in tissue engineering.

Purpose: Patients with tracheal lesions that exceed half of the trachea’s total length require a tracheal substitute. Tissue engineering, using either synthetic materials or decellularized tracheal tissue, opens up new possibilities for generating tracheal substitutes. Decellularization, a procedure in which the tissue's immunogenic cellular material is removed while the extracellular matrix (ECM) is preserved, seems the most promising approach. The majority of tracheal decellularization methods, however, are reliant on harsh chemicals and require lengthy wash procedures, resulting in damage to the ECM. To address these issues, a supercritical carbon dioxide (scCO2) decellularization approach has been suggested as an alternative solution due to its ability to both decellularize and sterilize tissues while leaving no toxic residues and requiring less treatment time. Therefore, the aim of this thesis was to compare scCO2 treatment with a chemical decellularization method, which is the current gold standard reported in literature, for creating a decellularized porcine tracheal scaffold with good cytocompatibility.  Methods: A total of five different protocols were tested that varied in decellularization and sterilization methods used. Decellularization efficiency was evaluated in terms of DNA content, histological appearance, and retention of ECM components. Additionally, mechanical tensile testing and scanning electron microscopy were used to assess the effects of the different decellularization protocols. Further, decellularized scaffolds were recellularized with fibrin-encapsulated porcine adipose derived stem cells to assess the cytocompatibility of the scaffolds. Results: The highest reduction in DNA content was observed when samples were subjected to the detergent-enzymatic protocol, followed by sterilization with gamma irradiation, and when samples were subjected to scCO2 treatment, followed by washing with sodium hydroxide. The latter protocol, however, also negatively impacted the ECM, whereas good preservation of ECM components was seen with the DEM protocol. DNA and histological analysis showed that treatment with scCO2 in combination with a hydrogen peroxide (H2O2) washing step was unable to completely decellularize porcine tracheas. Static surface seeding of the decellularized scaffolds led to poor cell adherence. Cells encapsulated in fibrin and seeded on the tracheal scaffolds were able to adhere and survive, showing the cytocompatibility of the decellularized scaffolds.  Conclusions: Decellularization with scCO2 in combination with a H2O2 washing step was not successful in completely decellularizing the porcine tracheas, and possibly requires the use of a co-solvent or secondary agent for successful decellularization. For recellularization of decellularized tracheal scaffolds, the use of fibrin as a cell carrier is an effective and simple seeding method. The findings in this thesis open up new avenues for potential optimizations in future research.

In the last two decades, it has been generally accepted that the interaction between bone and immune cells is critical in the process of bone healing and regeneration. Steering this immune response with the development of biomaterials could lead to implant acceptance in the long term. The understanding of osteoimmunological responses is of high importance in the development of new biomaterials. In order to validate these biomaterials, better in vitro techniques are required that could mimic this inflammatory response immediately after implantation in more detail. This study investigated the effect of a direct co-culture model on a prototype nanopattern. In addition the effect of microflows, that are normally present in vivo, induced inside a microfluidic chip, on mono and co-cultures was investigated in order to improve in vitro research.
Static mono- and co-cultures with a 1:1 (CO11) and 1:2 (CO12) ratio of murine pre-osteoblasts MC3T3-E1s (OBs) and M1 stimulated murine Macrophages J774A.1 (+ LPS & IFN-gamma) were cultured under static conditions for 4 weeks to see the effect of co-culturing on inflammatory markers (TNF-alpha, PGE2, and IL-10) and osteogenic markers (ALP, Runx2 & Alizarin Red Staining). It was found that the use of a co-culture did not have a positive effect on OB differentiation. The introduction of a prototype nanopattern (500 nm height and 300 nm width) that induces a macrophage polarization shift towards an anti-inflammatory phenotype, did have a positive effect on Runx2 secretion in the co-cultures. Furthermore the effect of a dynamic flow, initiated 2 days after seeding, on osteogenic differentiation of OB and CO12 was investigated. After 7 days of dynamic culture, Runx2 levels of the mono-cultures were significantly higher than the static cultures. OBs were positively affected by forming dense cell layers throughout the entire microfluidic channel. For dynamic co-cultures, no to little cells were present inside the channel after 7 days. Finally we attempted to fabricate a microfluidic device that could "sandwich" the patterned waver inside, so the effect of both the nanopattern and microflow on mono- and co-cultures could be investigated. Cells were seeded inside the channel after 1 day, but unfortunately forces on the PMMA layer of this chip resulted in cracks and failure of this system. Our findings in the design and fabrication of the chip can help in future research towards the osteoimmunodulatory properties of biomaterials.
The study on a direct co-culture model, presented in this work, opens possibilities for future research towards co-cultures inside microfluidic devices. It discusses the possibilities and the drawbacks of both the use of co-cultures as the development of microfluidic devices.

Cholangiocarcinoma (CCA) is a highly aggressive biliary tumor with a poor prognosis and limited treatment options. Various critical aspects of CCA development remain unclear. Lately, the dynamic process of epithelial-mesenchymal transition (EMT) was highlighted to play a crucial role in the fundamental mechanism of metastatic dissemination. However, in vitro EMT studies are mainly limited to conventional 2D models which lack in vivo tumor physiology. Recently, 3D organoid culture methods provide us new model possibilities. Therefore, our research aim was to establish a representative in vitro organoid-based epithelial-mesenchymal transition model in cholangiocarcinoma including its in vivo pathophysiology and tumor microenvironment. This study focusses on three different aspects of the tumor microenvironment regarding EMT activation in CCA organoids: 1.The effects of EMT promoting growth factors and culture media on EMT activation. Various growth factors are highlighted to play a crucial role in EMT activation in different cancer types. To study EMT activation in CCA, various growth factors were added to CCA organoids. We observed round shaped organoid structures consisting of single layered epithelium for all conditions. Additionally, gene expression levels of EMT related markers were increased for specific growth factor conditions. The use of various culture media combined with transforming growth factor β1 (TGF-β1) showed the most substantial differences in mRNA expression of EMT markers in branching medium (BM). Based on these results, specific growth factors and BM were used to find the optimal culture conditions to induce EMT. Although we did not find significant morphological differences between the various conditions, altered gene expression levels of EMT markers were observed after TGF-β1 and tumor necrosis factor alpha (TNF-α) treatment in BM. 2.The effect of the tumor extracellular matrix on EMT activation. The tumor extracellular matrix (ECM) is an important component of the tumor microenvironment that stimulates the malignant behavior of surrounding cells. The combination of CCA organoids and patient-derived ECM resulted in morphological changes. Cells with a polygonal shape adopted an irregular shape after TGF-β1 treatment. On top of that, our data revealed that patient-specific ECM and CCA organoids altered the gene expression of EMT related markers. 3.The effect of the interstitial fluid flow on EMT activation. The use of a microfluidic platform enabled us to study the effect of fluid flow on EMT in vitro. Cells formed organoid structures in the absence and the presence of a fluid flow. However, cells only migrated to the surrounding culture medium due to the presence of a fluid flow. When cells were seeded directly into the chip, cells formed a tube structure with protrusions growing outwards. The protrusions growth was accelerated by the addition of TGF-β1 and led to increased cell migration. Additionally, fluid flow conditions resulted in altered gene expression levels of EMT related markers. Our data suggest that fluid flow could potentially stimulate EMT in CCA organoids. Overall, our findings have demonstrated that growth factors, native ECM and interstitial fluid flow could be used to induce EMT in vitro. Indicating that the tumor microenvironment plays a crucial role in EMT and tumor progression. Our novel in vitro CCA organoid-based EMT model provides a foundation towards a more representative EMT in vitro model to unravel the underlying mechanism of CCA and its progression.