Refine
Has Fulltext
- yes (19)
Is part of the Bibliography
- yes (19)
Year of publication
Document Type
- Doctoral Thesis (19)
Language
- English (19) (remove)
Keywords
- Tissue Engineering (19) (remove)
Institute
- Graduate School of Life Sciences (19) (remove)
Sonstige beteiligte Institutionen
The main function of the small intestine is the absorption of essential nutrients, water and vitamins. Moreover, it constitutes a barrier protecting us from toxic xenobiotics and pathogens. For a better understanding of these processes, the development of intestinal in vitro models is of great interest to the study of pharmacological and pathological issues such as transport mechanisms and barrier function. Depending on the scientific questions, models of different complexity can be applied.
In vitro Transwell® systems based on a porous PET-membrane enable the standardized study of transport mechanisms across the intestinal barrier as well as the investigation of the influence of target substances on barrier integrity. However, this artificial setup reflects only limited aspects of the physiology of the native small intestine and can pose an additional physical barrier. Hence, the applications of this model for tissue engineering are limited.
Previously, tissue models based on a biological decellularized scaffold derived from porcine gut tissue were demonstrated to be a good alternative to the commonly used Transwell® system. This study showed that preserved biological extracellular matrix components like collagen and elastin provide a natural environment for the epithelial cells, promoting cell adhesion and growth. Intestinal epithelial cells such as Caco-2 cultured on such a scaffold showed a confluent, tight monolayer on the apical surface. Additionally, myofibroblasts were able to migrate into the scaffold supporting intestinal barrier formation.
In this thesis, dendritic cells were additionally introduced to this model mimicking an important component of the immune system. This co-culture model was then successfully proven to be suitable for the screening of particle formulations developed as delivery system for cancer antigens in peroral vaccination studies. In particular, nanoparticles based on PLGA, PEG-PAGE-PLGA, Mannose-PEG-PAGE-PLGA and Chitosan were tested. Uptake studies revealed only slight differences in the transcellular transport rate among the different particles. Dendritic cells were shown to phagocytose the particles after they have passed the intestinal barrier. The particles demonstrated to be an effective carrier system to transport peptides across the intestinal barrier and therefore present a useful tool for the development of novel drugs.
Furthermore, to mimic the complex structure and physiology of the gut including the presence of multiple different cell types, the Caco-2 cell line was replaced by primary intestinal cells to set up a de novo tissue model. To that end, intestinal crypts including undifferentiated stem cells and progenitor cells were isolated from human small intestinal tissue samples (jejunum) and expanded in vitro in organoid cultures. Cells were cultured on the decellularized porcine gut matrix in co-culture with intestinal myofibroblasts. These novel tissue models were maintained under either static or dynamic conditions.
Primary intestinal epithelial cells formed a confluent monolayer including the major differentiated cell types positive for mucin (goblet cells), villin (enterocytes), chromogranin A (enteroendocrine cells) and lysozyme (paneth cells). Electron microscopy images depicted essential functional units of an intact epithelium, such as microvilli and tight junctions. FITC-dextran permeability and TEER measurements were used to assess tightness of the cell layer. Models showed characteristic transport activity for several reference substances. Mechanical stimulation of the cells by a dynamic culture system had a great impact on barrier integrity and transporter activity resulting in a tighter barrier and a higher efflux transporter activity.
In Summary, the use of primary human intestinal cells combined with a biological decellularized scaffold offers a new and promising way to setup more physiological intestinal in vitro models. Maintenance of primary intestinal stem cells with their proliferation and differentiation potential together with adjusted culture protocols might help further improve the models. In particular, dynamic culture systems and co culture models proofed to be a first crucial steps towards a more physiological model. Such tissue models might be useful to improve the predictive power of in vitro models and in vitro in vivo correlation (IVIVC) studies. Moreover, these tissue models will be useful tools in preclinical studies to test pharmaceutical substances, probiotic active organisms, human pathogenic germs and could even be used to build up patient-specific tissue model for personalized medicine.
The basement membrane separates the epithelium from the stroma of any given barrier tissue and is essential in regulating cellular behavior, as mechanical barrier and as structural support. It further plays an important role for new tissue formation, homeostasis, and pathological processes, such as diabetes or cancer. Breakdown of the basement membrane is believed to be essential for tumor invasion and metastasization. Since the basement membrane is crucial for many body functions, the development of artificial basement membranes is indispensable for the ultimate formation of engineered functional tissue, however, challenging due to their complex structure.
Electrospinning enables the production of fibers in the nano- or microscale range with morphological similarities to the randomly orientated collagen and elastic fibers in the basement membrane. However, electrospun fibers often lack the functional similarity to guide cells and maintain tissue-specific functions. Hence, their possible applications as matrix structure for tissue engineering are limited.
Herein, the potential of polyester meshes, modified with six armed star-shaped pre-polymers and cell-adhesion-mediating peptides, was evaluated to act as functional isotropic and bipolar artificial basement membranes. Thereby, the meshes were shown to be biocompatible and stable including under dynamic conditions, and the degradation profile to correlate with the rate of new tissue formation. The different peptide sequences did not influence the morphology and integrity of the fibers. The modified membranes exhibited protein-repellent properties over 12 months, indicating the long-term stability of the cross-linked star-polymer surfaces.
Cell culture experiments with primary fibroblasts and a human keratinocyte cell line (HaCaT) revealed that cell adhesion and growth strongly depends on the peptide sequences and their combinations employed. HaCaT cells grew to confluence on membranes modified with a combination of laminin/collagen type IV derived binding sequences and with a combination of fibronectin/laminin/collagen type IV derived peptide sequences. Fibroblasts strongly adhered to the fibronectin derived binding sequence and to membranes containing a combination of fibronectin/laminin/collagen type IV derived peptide sequences. The adhesion and growth of fibroblasts and HaCaT cells were significantly reduced on membranes modified with laminin, as well as collagen IV derived peptide sequences. HaCaT cells and fibroblasts barely adhered onto meshes without peptide sequences.
Co-culture experiments at the air-liquid interface with fibroblasts and HaCaT cells confirmed the possibility of creating biocompatible, biofunctional and biomimetic isotropic and bipolar basement membranes, based on the functionalized fibers. HaCaT cells grew in several layers, differentiating towards the surface and expressing cytokeratin 10 in the suprabasal and cytokeratin 14 in the basal layers. Migration of fibroblasts into the electrospun membrane was shown by vimentin staining. Moreover, specific staining against laminin type V, collagen type I, III, IV and fibronectin illustrated that cells started to remodel the electrospun membrane and produced new extracellular matrix proteins following the adhesion to the synthetic surface structures.
The culturing of primary human skin keratinocytes proved to be difficult on electrospun fibers. Cells attached to the membrane, but failed to form a multilayered, well-stratified, and keratinized epidermal layer. Changing the fiber composition and fixation methods did not promote tissue development. Further investigations of the membrane demonstrated the tremendous influence of the pore size of the membrane on epithelial formation. Furthermore, primary keratinocytes reacted more sensitive to pH changes in the medium than HaCaT cells did.
Since primary keratinocytes did not adequately develop on the functionalized meshes, polycarbonate membranes were used instead of electrospun meshes to establish oral mucosa models. The tissue-engineered models represented important features of native human oral mucosa. They consisted of a multilayered epithelium with stratum basale, stratum spinosum, stratum granulosum, and stratum corneum. The models formed a physical barrier and the expression of characteristic cell markers was comparable with that in native human oral mucosa. The results from the ET-50 assay and the irritation study reflected the reproducibility of the tissue equivalents.
Altogether, electrospinning enables the production of fibers with structural similarity to the basement membrane. Incorporating extracellular matrix components to mimic the functional composition offers a safe and promising way to modify the fibers so that they can be used for different tissue engineering applications. The resultant biomimetic membranes that can be functionalized with binding sequences derived from widely varying proteins can be used as a toolbox to study the influence of isotropic and bipolar basement membranes on tissue formation and matrix remodeling systematically, with regards to the biochemical composition and the influence and importance of mono- and co-culture. The oral mucosa models may be useful for toxicity and permeation studies, to monitor the irritation potential of oral health care products and biomaterials or as a disease model.
Improved treatment options for the degenerative joint disease osteoarthritis (OA) are of major interest, since OA is one of the main sources of disability, pain, and socioeconomic burden worldwide [202]. According to epidemiological data, already 27 million people suffer from OA in the US [23]. Moreover, the WHO expects OA to be the fourth most common cause of disability in 2020 [203], illustrating the need for effective and long-lasting therapy options of severe cartilage defects. Despite numerous clinically available products for the treatment of cartilage defects [62], the development of more cartilage-specific materials is still at the beginning.
Hyaluronic acid (HA) is a major component of the cartilaginous extracellular matrix (ECM) and inherently creates a cell-friendly niche by providing cell attachment and migration sites. Furthermore, it is known that the functional groups of HA are well suited for chemical modification. These characteristics render HA an attractive material for hydrogel-based tissue engineering approaches. Poly(glycidol) (PG) as chemical crosslinker basically features similar chemical characteristics as the widely used poly(ethylene glycol) (PEG), but provides additional side groups at each repeating unit that can be further chemically functionalized. With the introduction of PG as multifunctional crosslinker for HA gels, a higher cross-linking density and, accordingly, a greater potential for biomimetic functionalization may be achieved. However, despite the mentioned potential benefits, PG has not been used for cartilage regeneration approaches so far.
The initial aim of the study was to set up and optimize a HA-based hydrogel for the chondrogenic differentiation of mesenchymal stromal cells (MSCs), using different amounts and variations of cross-linkers. Therefore, the hydrogel composition was optimized by the utilization of different PEG diacrylate (PEGDA) concentrations to cross-link thiol-modified HA (Glycosil, HA-SH) via Michael addition. We aimed to generate volumestable scaffolds that simultaneously enable a maximum of ECM deposition. Histological and biochemical analysis showed 0.4% PEGDA as the most suitable concentration for these requirements (Section 5.1.2).
In order to evaluate the impact of a differently designed cross-linker on MSC chondrogenesis, HA-SH was cross-linked with PEGTA (0.6%) and compared to PEGDA (0.4%) in a next step. Following this, acrylated PG (PG-Acr) as multifunctional cross-linker alternative to acrylated PEG was evaluated. It provides around five times more functional groups when utilized in PG-Acr (0.6%) HA-SH hydrogels compared to PEGTA (0.6%) HA-SH hydrogels, thus enabling higher degrees of biomimetic functionalization. Determination of cartilage-specific ECM components showed no substantial differences between both cross-linkers while the deposition of cartilaginous matrix appeared more homogeneous in HA-SH PG-Acr gels. Taken together, we were able to successfully increase the possibilities for biomimetic functionalization in the developed HA-SH hydrogel system by the introduction of PG-Acr as cross-linker without negatively affecting MSC chondrogenesis (Section 5.1.3).
The next part of this thesis focused extensively on the biomimetic functionalization of PG-Acr (0.6%) cross-linked HA-SH hydrogels. Here, either biomimetic peptides or a chondrogenic growth factor were covalently bound into the hydrogels.
Interestingly, the incorporation of a N-cadherin mimetic (HAV), a collagen type II binding (KLER), or a cell adhesion-mediating peptide (RGD) yielded no improvement of MSC chondrogenesis. For instance, the covalent binding of 2.5mM HAV changed morphology of cell nuclei and reduced GAG production while the incorporation of 1.0mM RGD impaired collagen production. These findings may be attributed to the already supportive conditions of the employed HA-based hydrogels for chondrogenic differentiation. Most of the previous studies reporting positive peptide effects on chondrogenesis have been carried out in less supportive PEG hydrogels or in significantly stiffer MeHA-based hydrogels [99, 101, 160]. Thus, the incorporation of peptides may be more important under unfavorable conditions while inert gel systems may be useful for studying single peptide effects (Section 5.2.1).
The chondrogenic factor transforming growth factor beta 1 (TGF-b1) served as an example for growth factor binding to PG-Acr. The utilization of covalently bound TGF-b1 may thereby help overcome the need for repeated administration of TGF-b1 in in vivo applications, which may be an advantage for potential clinical application. Thus, the effect of covalently incorporated TGF-b1 was compared to the effect of the same amount of TGF-b1 without covalent binding (100nM TGF-b1) on MSC chondrogenesis. It was successfully demonstrated that covalent incorporation of TGF-b1 had a significant positive effect in a dose-dependent manner. Chondrogenesis of MSCs in hydrogels with covalently bound TGF-b1 showed enhanced levels of chondrogenesis compared to hydrogels into which TGF-b1 was merely mixed, as shown by stronger staining for GAGs, total collagen, aggrecan and collagen type II. Biochemical evaluation of GAG and collagen amounts, as well as Western blot analysis confirmed the histological results. Furthermore, the positive effect of covalently bound TGF-b1 was shown by increased expression of chondrogenic marker genes COL2A1, ACAN and SOX9. In summary, covalent growth factor incorporation utilizing PG-Acr as cross-linker demonstrated significant positive effects on chondrogenic differentiation of MSCs (Section 5.2.2).
In general, PG-Acr cross-linked HA hydrogels generated by Michael addition represent a versatile hydrogel platform due to their high degree of acrylate functionality. These hydrogels may further offer the opportunity to combine several biological modifications, such as the incorporation of biomimetic peptides together with growth factors, within one cell carrier.
A proof-of-principle experiment demonstrated the suitability of pure PG gels for studying single peptide effects. Here, the hydrogels were generated by the utilization of thiol-ene-click reaction. In this setting, without the supportive background of hyaluronic acid, MSCs showed enhanced chondrogenic differentiation in response to the incorporation of 1.0mM HAV. This was demonstrated by staining for GAGs, the cartilage-specific ECM molecules aggrecan and type II collagen, and by increased GAG and total collagen amounts shown by biochemical analysis. Thus, pure PG gels exhibit the potential to study the effects and interplay of peptides and growth factors in a highly modifiable, bioinert hydrogel environment.
The last section of the thesis was carried out as part of the EU project HydroZONES that aims to develop and generate zonal constructs. The importance of zonal organization has attracted increased attention in the last years [127, 128], however, it is still underrepresented in tissue engineering approaches so far. Thus, the feasibility of zonal distribution of cells in a scaffold combining two differently composed hydrogels was investigated. A HA-SH(FMZ) containing bottom layer was generated and a pure PG top layer was subsequently cast on top of it, utilizing both times thiol-ene-click reaction. Indeed, stable, hierarchical constructs were generated that allowed encapsulated MSCs to differentiate chondrogenically in both zones as shown by staining for GAGs and collagen type II, and by quantification of GAG amount. Thus, the feasibility of differently composed zonal hydrogels utilizing PG as a main component was successfully demonstrated (Section 5.4).
With the first-time utilization and evaluation of PG-Acr as versatile multifunctional cross-linker for the preparation of Michael addition-generated HA-SH hydrogels in the context of cartilage tissue engineering, a highly modifiable HA-based hydrogel system was introduced. It may be used in future studies as an easily applicable and versatile toolbox for the generation of biomimetically functionalized hydrogels for cell-based cartilage regeneration. The introduction of reinforcement structures to enhance mechanical resistance may thereby further increase the potential of this system for clinical applications.
Additionally, it was also demonstrated that thiol-ene clickable hydrogels can be used for the generation of cell-laden, pure PG gels or for the generation of more complex, coherent zonal constructs. Furthermore, thiol-ene clickable PG hydrogels have already been further modified and successfully been used in 3D bioprinting experiments [204]. 3D bioprinting, as part of the evolving biofabrication field [205], offers the possibilities to generate complex and hierarchical structures, and to exactly position defined layers, yet at the same time alters the requirements for the utilized hydrogels [159, 206–209]. Since a robust chondrogenesis of MSCs was demonstrated in the thiol-ene clickable hydrogel systems, they may serve as a basis for the development of hydrogels as so called bioinks which may be utilized in more sophisticated biofabrication processes.
The knee joint is a complex composite joint containing the C-shaped wedge-like menisci composed of fibrocartilage. Due to their complex composition and structure, they provide mechanical resilience to the knee joint protecting the articular cartilage. Because of the limited repair potential, meniscal injuries do not only affect the meniscus itself but also lead to altered joint homeostasis and inevitably to secondary osteoarthritis.
The meniscus was characterized focusing on its anatomy, structure and meniscal markers such as aggrecan, collagen type I (Col I) and Col II. The components relevant for meniscus tissue engineering, namely cells, Col I scaffolds, biochemical and biomechanical stimuli were studied. Meniscal cells (MCs) were isolated from meniscus, mesenchymal stem cells (MSCs) from bone marrow and dermal microvascular endothelial cells (d-mvECs) from foreskin biopsies. For the human (h) meniscus model, wedge-shape compression of a hMSC-laden Col I gel was successfully established. During three weeks of static culture, the biochemical stimulus transforming growth factor beta-3 (TGF beta-3) led to a compact collagen structure. On day 21, this meniscus model showed high metabolic activity and matrix remodeling as confirmed by matrix metalloproteinases detection. The fibrochondrogenic properties were illustrated by immunohistochemical detection of meniscal markers, significant GAG/DNA increase and increased compressive properties. For further improvement, biomechanical stimulation systems by compression and hydrostatic pressure were designed. As one vascularization approach, direct stimulation with ciclopirox olamine (CPX) significantly increased sprouting of hd-mvEC spheroids even in absence of auxiliary cells such as MSCs. Second, a cell sheet composed of hMSCs and hd-mvECs was fabricated by temperature triggered cell sheet engineering and transferred onto the wedge-shaped meniscus model. Third, a biological vascularized scaffold (BioVaSc-TERM) was re-endothelialized with hd-mvECs providing a viable vascularized network. The vascularized BioVaSc-TERM was suggested as wrapping scaffold of the meniscus model by using two suture techniques, the all-inside-repair (AIR) for the posterior horn, and the outside-in-refixation (OIR) for the anterior horn and the middle part.
This meniscus model for replacing torn menisci is a promising approach to be further optimized regarding vascularization, biochemical and biomechanical stimuli.
Breast cancer is the most common cancer among women worldwide and the second most common cause of cancer death in the developed countries. As the current state of the art in first-line drug screenings is highly ineffective, there is an urgent need for novel test systems that allow for reliable predictions of drug sensitivity.
In this study, a tissue engineering approach was used to successfully establish and standardize a 3-dimensional (3D) mamma carcinoma test system that was optimized for the testing of anti-tumour therapies as well as for the investigation of tumour biological issues. This 3D test system is based on the decellularised scaffold of a porcine small intestinal segment and represents the three molecular subsets of oestrogen receptor-positive, HER2/Neu-overexpressing and triple negative breast cancer (TNBC). The characterization of the test system with respect to morphology as well as the expression of markers for epithelial-mesenchymal transition (EMT) and differentiation indicate that the 3D tumour models cultured under static and dynamic conditions reflect tumour relevant features and have a good correlation with in vivo tumour tissue from the corresponding xenograft models. In this respect, the dynamic culture in a flow bioreactor resulted in the generation of tumour models that exhibited best reflection of the morphology of the xenograft material. Furthermore, the proliferation indices of 3D models were significantly reduced compared to 2-dimensional (2D) cell culture and therefore better reflect the in vivo situation. As this more physiological proliferation index prevents an overestimation of the therapeutic effect of cytostatic compounds, this is a crucial advantage of the test system compared to 2D culture. Moreover, it could be shown that the 3D models can recapitulate different tumour stages with respect to tumour cell invasion. The scaffold SISmuc with the preserved basement membrane structure allowed the investigation of invasion over this barrier which tumour cells of epithelial origin have to cross in in vivo conditions during the process of metastasis formation. Additionally, the data obtained from ultrastructural analysis and in situ zymography indicate that the invasion observed is connected to a tumour cell-associated change in the basement membrane in which matrix metalloproteinases (MMPs) are also involved. This features of the model in combination with the mentioned methods of analysis could be used in the future to mechanistically investigate invasive processes and to test anti-metastatic therapy strategies.
The validation of the 3D models as a test system with respect to the predictability of therapeutic effects was achieved by the clinically relevant targeted therapy with the monoclonal antibody trastuzumab which induces therapeutic response only in patients with HER2/Neu-overexpressing mamma carcinomas due to its specificity for HER2. While neither in 2D nor in 3D models of all molecular subsets a clear reduction of cell viability or an increase in apoptosis could be observed, a distinct increase in antibody-dependent cell-mediated cytotoxicity (ADCC) was detected only in the HER2/NEU-overexpressing 3D model with the help of an ADCC reporter gene assay that had been adapted for the application in the 3D model in the here presented work. This correlates with the clinical observations and underlines the relevance of ADCC as a mechanism of action (MOA) of trastuzumab. In order to measure the effects of ADCC on the tumour cells in a direct way without the indirect measurement via a reporter gene, the introduction of an immunological component into the models was required. This was achieved by the integration of peripheral blood mononuclear cells (PBMCs), thereby allowing the measurement of the induction of tumour cell apoptosis in the HER2/Neu-overexpressing model. Hence, in this study an immunocompetent model could be established that holds the potential for further testing of therapies from the emergent field of cancer immunotherapies.
Subsequently, the established test system was used for the investigation of scientific issues from different areas of application. By the comparison of the sensitivity of the 2D and 3D model of TNBC towards the water-insoluble compound curcumin that was applied in a novel nanoformulation or in a DMSO-based formulation, the 3D test system was successfully applied for the evaluation of an innovative formulation strategy for poorly soluble drugs in order to achieve cancer therapy-relevant concentrations. Moreover, due to the lack of targeted therapies for TNBC, the TNBC model was applied for testing novel treatment strategies. On the one hand, therapy with the WEE1 kinase inhibitor MK 1775 was evaluated as a single agent as well as in combination with the chemotherapeutic agent doxorubicin. This therapy approach did not reveal any distinct benefits in the 3D test system in contrast to testing in 2D culture. On the other hand, a novel therapy approach from the field of cellular immunotherapies was successfully applied in the TNBC 3D model. The treatment with T cells that express a chimeric antigen receptor (CAR) against ROR1 revealed in the static as well as in the dynamic model a migration of T cells into the tumour tissue, an enhanced proliferation of T cells as well as an efficient lysis of the tumour cells via apoptosis and therefore a specific anti-cancer effect of CAR-transduced T cells compared to control T cells. These results illustrate that the therapeutic application of CAR T cells is a promising strategy for the treatment of solid tumours like TNBC and that the here presented 3D models are suitable for the evaluation and optimization of cellular immunotherapies.
In the last part of this work, the 3D models were expanded by components of the tumour stroma for future applications. By coculture with fibroblasts, the natural structures of the intestinal scaffold comprising crypts and villi were remodelled and the tumour cells formed tumour-like structures together with the fibroblasts. This tissue model displayed a strong correlation with xenograft models with respect to morphology, marker expression as well as the activation of dermal fibroblasts towards a cancer-associated fibroblast (CAF) phenotype. For the integration of adipocytes which are an essential component of the breast stroma, a coculture with human adipose-derived stromal/stem cells (hASCs) which could be successfully differentiated along the adipose lineage in 3D static as well as dynamic models was established. These models are suitable especially for the mechanistic analysis of the reciprocal interaction between tumour cells and adipocytes due to the complex differentiation process.
Taken together, in this study a human 3D mamma carcinoma test system for application in the preclinical development and testing of anti-tumour therapies as well as in basic research in the field of tumour biology was successfully established. With the help of this modular test system, relevant data can be obtained concerning the efficacy of therapies in tumours of different molecular subsets and different tumour stages as well as for the optimization of novel therapy strategies like immunotherapies. In the future this can contribute to improve the preclinical screening and thereby to reduce the high attrition rates in pharmaceutical industry as well as the amount of animal experiments.
In reconstructive and plastic surgery, there exists a growing demand of adequate tissue implants, since currently available strategies for autologous transplantation are limited by complications including transplant failure and donor site morbidity. By developing in vitro and in vivo autologous substitutes for defective tissue sites, adipose tissue engineering can address these challenges, although there are several obstacles to overcome. One of the major limitations is the sufficient vascularization of in vitro engineered large constructs that remains crucial and demanding for functional tissues. Decellularized jejunal segments may represent a suitable scaffolding system with preexisting capillary structures that can be repopulated with human microvascular endothelial cells (hMVECs), and a luminal matrix applicable for the adipogenic differentiation of human adipose-derived stem cells (hASCs). Hence, co-culture of these cells in jejunal segments, utilizing a custom-made bioreactor system, was characterized in terms of vascularization and adipose tissue development. Substantial adipogenesis of hASCs was demonstrated within the jejunal lumen in contrast to non-induced controls, and the increase of key adipogenic markers was verified over time upon induction. The development of major extracellular matrix components of mature adipose tissue, such as laminin and collagen IV, was shown within the scaffold in induced samples. Successful reseeding of the vascular network with hMVECs was demonstrated in long-term culture and co-localization of vascular structures and adipogenically differentiated hASCs was observed. Therefore, these results represent a novel approach for in vitro engineering of vascularized adipose tissue constructs that warrants further investigations in preclinical studies.
Another still existing obstacle in adipose tissue engineering is the insufficient knowledge about the applied cells, for instance the understanding of how cells can be optimally expanded and differentiated for successful engineering of tissue transplants. Even though hASCs can be easily isolated from liposuction of abdominal fat depots, yielding low donor site morbidity, huge numbers of cells are required to entirely seed complex and large 3D matrices or scaffolds. Thus, cells need to be large-scale expanded in vitro on the premise of not losing their differentiation capacity caused by replicative aging. Accordingly, an improved differentiation of hASCs in adipose tissue engineering approaches remains still desirable since most engineered constructs exhibit an inhomogeneous differentiation pattern. For mesenchymal stem cells (MSCs), it has been shown that growth factor application can lead to a significant improvement of both proliferation and differentiation capacity. Especially basic fibroblast growth factor (bFGF) represents a potent mitogen for MSCs, while maintaining or even promoting their osteogenic, chondrogenic and adipogenic differentiation potential. As there are currently different contradictory information present in literature about the applied bFGF concentration and the explicit effect of bFGF on ASC differentiation, here, the effect of bFGF on hASC proliferation and differentiation capacity was investigated at different concentrations and time points in 2D culture. Preculture of hASCs with bFGF prior to adipogenic induction showed a remarkable effect, whereas administration of bFGF during culture did not improve adipogenic differentiation capacity. Furthermore, the observations indicated as mode of action an impact of this preculture on cell proliferation capacity, resulting in increased cellular density at the time of adipogenic induction. The difference in cell density at this time point appeared to be pivotal for increased adipogenic capacity of the cells, which was confirmed in a further experiment employing different seeding densities. Interestingly, furthermore, the obtained results suggested a cell-cell contact-mediated mechanism positively influencing adipogenic differentiation. As a consequence, subsequently, studies were conducted focusing on intercellular communication of these cells, which has hardly been investigated to date.
Despite the multitude of literature on the differentiation capacity of ASCs, little is reported about the physiological properties contributing to and controlling the process of lineage differentiation. Direct intercellular communication between adjacent cells via gap junctions has been shown to modulate differentiation processes in other cell types, with connexin 43 (Cx43) being the most abundant isoform of the gap junction-forming connexins. Thus, in the present study we focused on the expression of Cx43 and gap junctional intercellular communication (GJIC) in hASCs, and its significance for adipogenic differentiation of these cells. Cx43 expression in hASCs was demonstrated histologically and on the gene and protein expression level and was shown to be greatly positively influenced by cell seeding density. Functionality of gap junctions was proven by dye transfer analysis in growth medium. Adipogenic differentiation of hASCs was shown to be also distinctly elevated at higher cell seeding densities. Inhibition of GJIC by 18α-glycyrrhetinic acid significantly compromised adipogenic differentiation, as demonstrated by histology, triglyceride quantification, and adipogenic marker gene expression. Flow cytometry analysis showed a lower proportion of cells undergoing adipogenesis when GJIC was inhibited, further indicating the importance of GJIC in the differentiation process. Altogether, these results demonstrate the impact of direct cell-cell communication via gap junctions on the adipogenic differentiation process of hASCs and may contribute to further integrate direct intercellular crosstalk in rationales for tissue engineering approaches.
Articular cartilage damage caused by sports accidents, trauma or gradual wear and tear can lead to degeneration and the development of osteoarthritis because cartilage tissue has only limited capacity for intrinsic healing. Osteoarthritis causes reduction of mobility and chronic pain and is one of the leading causes of disability in the elderly population. Current clinical treatment options can reduce pain and restore mobility for some time, but the formed repair tissue has mostly inferior functionality compared to healthy articular cartilage and does not last long-term. Articular cartilage tissue engineering is a promising approach for the improvement of the quality of cartilage repair tissue and regeneration. In this thesis, a promising new cell type for articular cartilage tissue engineering, the so-called articular cartilage progenitor cell (ACPC), was investigated for the first time in the two different hydrogels agarose and HA-SH/P(AGE-co-G) in comparison to mesenchymal stromal cells (MSCs). In agarose, ACPCs´ and MSCs´ chondrogenic capacity was investigated under normoxic (21 % oxygen) and hypoxic (2 % oxygen) conditions in monoculture constructs and in zonally layered co-culture constructs with ACPCs in the upper layer and MSCs in the lower layer. In the newly developed hyaluronic acid (HA)-based hydrogel HA-SH/P(AGE-co-G), chondrogenesis of ACPCs and MSCs was also evaluated in monoculture constructs and in zonally layered co-culture constructs like in agarose hydrogel. Additionally, the contribution of the bioactive molecule hyaluronic acid to chondrogenic gene expression of MSCs was investigated in 2D monolayer, 3D pellet and HA-SH hydrogel culture. It was shown that both ACPCs and MSCs could chondrogenically differentiate in agarose and HA-SH/P(AGE-co-G) hydrogels. In agarose hydrogel, ACPCs produced a more articular cartilage-like tissue than MSCs that contained more glycosaminoglycan (GAG), less type I collagen and only little alkaline phosphatase (ALP) activity. Hypoxic conditions did not increase extracellular matrix (ECM) production of ACPCs and MSCs significantly but improved the quality of the neo-cartilage tissue produced by MSCs. The creation of zonal agarose constructs with ACPCs in the upper layer and MSCs in the lower layer led to an ECM production in zonal hydrogels that lay in general in between the ECM production of non-zonal ACPC and MSC hydrogels. Even though zonal co-culture of ACPCs and MSCs did not increase ECM production, the two cell types influenced each other and, for example, modulated the staining intensities of type II and type I collagen in comparison to non-zonal constructs under normoxic and hypoxic conditions. In HA-SH/P(AGE-co-G) hydrogel, MSCs produced more ECM than ACPCs, but the ECM was limited to the pericellular region for both cell types. Zonal HASH/P(AGE-co-G) hydrogels resulted in a native-like zonal distribution of ECM as MSCs in the lower zone produced more ECM than ACPCs in the upper zone. It appeared that chondrogenesis of ACPCs was supported by hydrogels without biological attachment sites such as agarose, and that chondrogenesis of MSCs benefited from hydrogels with biological cues like HA. As HA is an attractive material for cartilage tissue engineering, and the HA-based hydrogel HA-SH/P(AGE-co-G) appeared to be beneficial for MSC chondrogenic differentiation, the contribution of HA to chondrogenic gene expression of MSCs was investigated. An upregulation of chondrogenic gene expression was found in 2D monolayer and 3D pellet culture of MSCs in response to HA supplementation, while gene expression of osteogenic and adipogenic transcription factors was not upregulated. MSCs, encapsulated in a HA-based hydrogel, showed upregulation of gene expression for chondrogenic, osteogenic and adipogenic differentiation markers as well as for stemness markers. In a 3D bioprinting process, using the HA-based hydrogel, gene expression levels of MSCs mostly did not change. Nevertheless, expression of three tested genes (COL2A1, SOX2, CD168) was downregulated in printed in comparison to cast constructs, underscoring the importance of closely monitoring cellular behaviour during and after the printing process. In summary, it was confirmed that ACPCs are a promising cell source for articular cartilage engineering with advantages over MSCs when they were cultured in a suitable hydrogel like agarose. The performance of the cells was strongly dependent on the hydrogel environment they were cultured in. The different chondrogenic performance of ACPCs and MSCs in agarose and HA-SH/P(AGE-co-G) hydrogels highlighted the importance of choosing suitable hydrogels for the different cell types used in articular cartilage tissue engineering. Hydrogels with high polymer content, such as the investigated HA-SH/P(AGE-co-G) hydrogels, can limit ECM distribution to the pericellular area and should be developed further towards less polymer content, leading to more homogenous ECM distribution of the cultured cells. The influence of HA on chondrogenic gene expression and on the balance between differentiation and maintenance of stemness in MSCs was demonstrated. More studies should be performed in the future to further elucidate the signalling functions of HA and the effects of 3D bioprinting in HA-based hydrogels. Taken together, the results of this thesis expand the knowledge in the area of articular cartilage engineering with regard to the rational combination of cell types and hydrogel materials and open up new possible approaches to the regeneration of articular cartilage tissue.
The small intestine represents a strong barrier separating the lumen from blood circulation thereby playing a major role in the absorption and the transport of pharmacological agents prior to their arrival on the respective target site. In order to gain more knowledge about specialized uptake mechanisms and risk assessment for the patient after oral admission of drugs, intestinal in vitro models demonstrating a close similarity to the in vivo situation are needed.
In the past, cell line-based in vitro models composed of Caco-2 cells cultured on synthetic cell carriers represented the “gold standard” in the field of intestinal tissue engineering. Expressive advantages of these models are a reproducible, cost-efficient and standardized model set up, but cell function can be negatively influenced by the low porosity or unwanted molecular adhesion effects of the artificial scaffold material. Natural extracellular matrices (ECM) such as the porcine decellularized small intestinal submucosa (SIS) are used as alternative to overcome some common drawbacks; however, the fabrication of these scaffolds is time- and cost-intensive, less well standardized and the 3Rs (replacement, reduction, refinement) principle is not entirely fulfilled. Nowadays, biopolymer-based scaffolds such as the bacterial nanocellulose (BNC) suggest an interesting option of novel intestinal tissue engineered models, as the BNC shows comparable features to the native ECM regarding fiber arrangement and hydrophilic properties. Furthermore, the BNC is of non-animal origin and the manufacturing process is faster as well as well standardized at low costs.
In this context, the first part of this thesis analyzed the BNC as alternative scaffold to derive standardized and functional organ models in vitro. Therefore, Caco-2 cells were cultured on two versions of BNC with respect to their surface topography, the unmodified BNC as rather smooth surface and the surface-structured BNC presenting an aligned fiber arrangement. As controls, Caco-2 in vitro models were set up on PET and SIS matrices. In this study, the BNC-based models demonstrated organ-specific properties comprising typical cellular morphologies, a characteristic tight junction protein expression profile, representative ultrastructural features and the formation of a tight epithelial barrier together with a corresponding transport activity. In summary, these results validated the high quality of the BNC-based Caco-2 models under cost-efficient conditions and their suitability for pre-clinical research purposes. However, the full functional diversity of the human intestine cannot be presented by Caco-2 cells due to their tumorigenic background and their exclusive representation of mature enterocytes.
Next to the scaffold used for the setup of in vitro models, the cellular unit mainly drives functional performance, which demonstrates the crucial importance of mimicking the cellular diversity of the small intestine in vitro. In this context, intestinal primary organoids are of high interest, as they show a close similarity to the native epithelium regarding their cellular diversity comprising enterocytes, goblet cells, enteroendocrine cells, paneth cells, transit amplifying cells and stem cells. In general, such primary organoids grow in a 3D Matrigel® based environment and a medium formulation supplemented with a variety of growth factors to maintain stemness, to inhibit differentiation and to stimulate cell migration supporting long term in vitro culture.
Intestinal primary spheroid/organoid cultures were set up as Transwell®-like models on both BNC variants, which resulted in a fragmentary cell layer and thereby unfavorable properties of these scaffold materials under the applied circumstances. As the BNC manufacturing process is highly flexible, surface properties could be adapted in future studies to enable a good cell adherence and barrier formation for primary intestinal cells, too. However, the application of these organoid cultures in pre-clinical research represents an enormous challenge, as the in vitro culture is complex and additionally time- and cost-intensive.
With regard to the high potential of primary intestinal spheroids/organoids and the necessity of a simplified but predictive model in pre-clinical research purposes, the second part of this thesis addressed the establishment of a primary-derived immortalized intestinal cell line, which enables a standardized and cost-efficient culture (including in 2D), while maintaining the cellular diversity of the organoid in vitro cultures. In this study, immortalization of murine and human intestinal primary organoids was induced by ectopic expression of a 10- (murine) or 12 component (human) pool of genes regulating stemness and the cell cycle, which was performed in cooperation with the InSCREENeX GmbH in a 2D- and 3D-based transduction strategy. In first line, the established cell lines (cell clones) were investigated for their cell culture prerequisites to grow under simplified and cost-efficient conditions. While murine cell clones grew on uncoated plastic in a medium formulation supplemented with EGF, Noggin, Y-27632 and 10% FCS, the human cell clones demonstrated the necessity of a Col I pre coating together with the need for a medium composition commonly used for primary human spheroid/organoid cultures. Furthermore, the preceding analyses resulted in only one human cell clone and three murine cell clones for ongoing characterization. Studies regarding the proliferative properties and the specific gene as well as protein expression profile of the remaining cell clones have shown, that it is likely that transient amplifying cells (TACs) were immortalized instead of the differentiated cell types localized in primary organoids, as 2D, 3D or Transwell®-based cultures resulted in slightly different gene expression profiles and in a dramatically reduced mRNA transcript level for the analyzed marker genes representative for the differentiated cell types of the native epithelium. Further, 3D cultures demonstrated the formation of spheroid-like structures; however without forming organoid-like structures due to prolonged culture, indicating that these cell populations have lost their ability to differentiate into specific intestinal cell types. The Transwell®-based models set up of each clone exhibit organ-specific properties comprising an epithelial-like morphology, a characteristic protein expression profile with an apical mucus-layer covering the villin-1 positive cell layer, thereby representing goblet cells and enterocytes, together with representative tight junction complexes indicating an integer epithelial barrier. The proof of a functional as well as tight epithelial barrier in TEER measurements and in vivo-like transport activities qualified the established cell clones as alternative cell sources for tissue engineered models representing the small intestine to some extent. Additionally, the easy handling and cell expansion under more cost-efficient conditions compared to primary organoid cultures favors the use of these newly generated cell clones in bioavailability studies.
Altogether, this work demonstrated new components, structural and cellular, for the establishment of alternative in vitro models of the small intestinal epithelium, which could be used in pre-clinical screenings for reproducible drug delivery studies.
Malignant melanoma (MM) is the most dangerous type of skin cancer with rising incidences worldwide. Melanoma skin models can help to elucidate its causes and formation or to develop new treatment strategies. However, most of the current skin models lack a vasculature, limiting their functionality and applicability. MM relies on the vascular system for its own supply and for its dissemination to distant body sites via lymphatic and blood vessels. Thus, to accurately study MM progression, a functional vasculature is indispensable. To date, there are no vascularized skin models to study melanoma metastasis in vitro, which is why such studies still rely on animal experimentation.
In the present thesis, two different approaches for the vascularization of skin models are employed with the aim to establish a vascularized 3D in vitro full-thickness skin equivalent (FTSE) that can serve as a test system for the investigation of the progression of MM.
Initially, endothelial cells were incorporated in the dermal part of FTSEs. The optimal seeding density, a spheroid conformation of the cells and the cell culture medium were tested. A high cell density resulted in the formation of lumen-forming shapes distributed in the dermal part of the model. These capillary-like structures were proven to be of endothelial origin by staining for the endothelial cell marker CD31. The established vascularized FTSE (vFTSE) was characterized histologically after 4 weeks of culture, revealing an architecture similar to human skin in vivo with a stratified epidermis, separated from the dermal equivalent by a basement membrane indicated by collagen type IV. However, this random capillary-like network is not functional as it cannot be perfused.
Therefore, the second vascularization approach focused on the generation of a perfusable tissue construct. A channel was molded within a collagen hydrogel and seeded with endothelial cells to mimic a central, perfusable vessel. The generation and the perfusion culture of the collagen hydrogel was enabled by the use of two custom-made, 3D printed bioreactors. Histological assessment of the hydrogels revealed the lining of the channel with a monolayer of endothelial cells, expressing the cell specific marker CD31.
For the investigation of MM progression in vitro, a 3D melanoma skin equivalent was established. Melanoma cells were incorporated in the epidermal part of FTSEs, representing the native microenvironment of the tumor. Melanoma nests grew at the dermo-epidermal junction within the well stratified epidermis and were characterized by the expression of common melanoma markers. First experiments were conducted showing the feasibility of combining the melanoma model with the vFTSE, resulting in skin models with tumors at the dermo-epidermal junction and lumen-like structures in the dermis.
Taken together, the models presented in this thesis provide further steps towards the establishment of a vascularized, perfusable melanoma model to study melanoma progression and metastasis.
Additive manufacturing processes such as 3D printing are booming in the industry due to their high degree of freedom in terms of geometric shapes and available materials. Focusing on patient-specific medicine, 3D printing has also proven useful in the Life Sciences, where it exploits the shape fidelity for individualized tissues in the field of bioprinting. In parallel, the current systems of bioreactor technology have adapted to the new manufacturing technology as well and 3D-printed bioreactors are increasingly being developed. For the first time, this work combines the manufacturing of the tissue and a tailored bioreactor, significantly streamlining the overall process and optimally merging the two processes. This way the production of the tissues can be individualized by customizing the reactor to the tissue and the patient-specific wound geometry. For this reason, a common basis and guideline for the cross-device and cross-material use of 3D printers was created initially. Their applicability was demonstrated by the iterative development of a perfusable bioreactor system, made from polydimethylsiloxane (PDMS) and a lignin-based filament, into which a biological tissue of flexible shape can be bioprinted. Cost-effective bioink-replacements and in silico computational fluid dynamics simulations were used for material sustainability and shape development. Also, nutrient distribution and shear stress could be predicted in this way pre-experimentally.
As a proof of functionality and adaptability of the reactor, tissues made from a nanocellulose-based Cellink® Bioink, as well as an alginate-based ink mixed with Me-PMeOx100-b-PnPrOzi100-EIP (POx) (Alginate-POx bioink) were successfully cultured dynamically in the bioreactor together with C2C12 cell line. Tissue maturation was further demonstrated using hMSC which were successfully induced to adipocyte differentiation. For further standardization, a mobile electrical device for automated media exchange was developed, improving handling in the laboratory and thus reduces the probability of contamination.