Tunable and cell-responsive 3D poly(ethylene glycol) microenvironments for the development of tissue models

The recapitulation of tissue organization and function in vitro is one of the main objectives of tissue engineering. Proper tissue function depends on the spatial assembly and interactions between key components such as cells, growth factors, signaling cues and extracellular matrix (ECM). The ECM is of utmost importance in the design of engineered tissues, since it not only provides support but also actively interacts with cells and growth factors to regulate tissue function. Tissue engineering has thus pushed towards the development of novel biocompatible and biodegradable materials capable of both supporting and interacting with cells. Such biomaterials are expected to allow the positioning of different cell types and the localization of biological factors, as well as having controlled adhesive properties and stiffness. Yet despite constant improvement on biomaterial design, technological advances and the increasing knowledge gained on tissue functions, the synthesis of artificial tissues and organs still remains far ahead. Current achievements in research have enabled the generation of engineered constructs with either controlled positioning of cells or patterning of growth factors. But the combination of both aspects into a reproducible and high-throughput strategy is still a task at hand. Another limitation is the current inability to recreate stable vascularization in tissue constructs. Vascularization is a compulsory aspect of engineering artificial tissues, since survival of any clinical-sized constructs depends on the availability of oxygen and nutrients normally provided by blood circulation in vivo. Such a shortcoming is directly linked to the necessity of establishing new strategies for the localized control of cellular function in 3D-arranged microenvironments, since vascularization is a complex and well-orchestrated phenomenon involving cell migration and differentiation under the control of specific guidance and signaling cues. To this end, a platform for the synthesis of tunable tissue constructs was developed in this thesis work. Taking advantage of the biocompatible and highly customizable enzymatically cross-linked poly(ethylene glycol) hydrogel platform, a strategy for the controlled incorporation of bioactive factors via peptide linkers was developed. The successful localized immobilization of growth factors and their ability to differentiate cells in a 3D-controlled manner was demonstrated by the osteogenic differentiation of mesenchymal cells by tethered bone morphogenic protein-2. Furthermore, using the same model, the release dynamics of the immobilized growth factors were investigated and a cell-dependant release strategy was devised by designing a proteolytically degradable peptide linker. The availability of immobilization strategies that are cell-responsive and that can mimic ECM properties is a desirable feature that can also contribute to the improvement of growth factor dose-response relationship. Additionally, selected factors modulating angiogenesis were evaluated for their potential role and use in the design of tissue constructs. Pro-angiogenic, anti-angiogenic effects and influence on cell migration were assessed in vitro and using the highly vascularized chicken chorioallantoic membrane. Preliminary experiments support the hypothesis that the selected growth factors could be used for establishing prevascular structures in engineered constructs.

Tissue engineering of the bladder

Catharina Adelöw

Stem cell use in bladder tissue engineering is a recently addressed area of investigation that has generated excitement as a novel way to restore and regenerate lost or damaged urinary bladder tissue. The remodelling of smooth muscle plays a significant role in bladder repair and regeneration. De-differentiation of smooth muscle cells from a contractile phenotype to a synthetic proliferative phenotype is characterised by smooth muscle hypertrophy and fibrosis, and leads to a poorly compliant bladder. To elucidate the underlying mechanisms of such phenomena, recent efforts in bladder tissue engineering have focused on understanding smooth muscle biology and associated mechanisms of bladder diseases. In this thesis we have designed a synthetic poly (ethylene glycol) (PEG) hydrogel for stem cell and drug delivery aimed to enhance control smooth muscle cell phenotype and thus assist in bladder repair and regeneration. We began by developing a method for isolation of cells from the bladder wall. Traditionally urothelium and smooth muscle cells are separated by dissection, and primary cultures are initiated in distinctly different medium formulations to obtain homogenous cultures and limit contaminating cells. However, the bladder wall harbours, apart from urothelial and smooth muscle cells, fibroblasts that have similar nutritional requirements and are likely candidates to contaminate smooth muscle cell cultures. Isolation of these three cell types by conjugation to magnetic beads and repeated separations allowed us to obtain pure cell populations. Furthermore, this technique permitted us to freely choose any specific medium formulation we wished to utilise. We then turned our attentions to engineering a materials system, by chemically and physically tuning a PEG hydrogel in terms of ligand concentration and matrix stiffness, to provide a three-dimensional environment for optimal spreading of human smooth muscle cells and mesenchymal stem cells. Cell viability, proliferation and differentiation were assessed in optimised hydrogels. Compared to cultures on traditional two-dimensional plastic, mesenchymal stem cells cultured within our three-dimensional hydrogels acquired a smooth muscle cell-like phenotype and smooth muscle cells obtained a less de-differentiated phenotype. In addition, we demonstrated cell-demanded gel degradation and deposition of newly synthesised extracellular matrix proteins. In order to modulate cell phenotypes further, we hypothesised that cell differentiation could be directed by the adhesion ligands made available within the matrix. In this manner, cells that were exposed to a matrix presenting certain integrin-binding ligands would begin expressing the corresponding integrins in order to adhere to the matrix. Consequently, we explored the effect of various specific integrin-binding ligands on mesenchymal stem cell integrin expression and differentiation. Fibronectin (FN) fragments engineered to demonstrate specificity to different integrins, namely α5β1 (FNIII910) and αvβ3 (FNIII10), were conjugated to PEG functionalised with vinyl sulfone and compared to fibronectin wild type (FNIIIWT) fragment and the RGD peptide, which both provide binding sites for a range of integrins. We found that assembly of focal adhesions and formation of stress fibres were altered due to the specific fragments provided, and expression of integrin- and smooth muscle cell specific genes varied in mesenchymal stem cells cultured within hydrogels functionalised with different fragments. Moreover, these fibronectin fragments turned out to be especially helpful for improved attachment of urothelial cells on top of hydrogels. With this biomaterial system in hand, we sought to develop a co-culture model of the bladder wall that permitted mesenchymal stem cells and urothelial cells cultured within / on top of PEG hydrogels, and to explore the combinatorial effect of ligand binding and cell-cell interactions on mesenchymal stem cell fate. Interestingly, we found that co-culturing with urothelial cells influenced gene expression of mesenchymal stem cells differently, depending on the protein fragment to which they were adhering. Mesenchymal stem cells cultured within PEG-FNIIIWT hydrogels were shown to up-regulate smooth muscle cell specific genes, while down-regulation was demonstrated in PEG-RGD, PEG-FNIII910 and PEG-FNIII10 hydrogels. Again, mesenchymal stem cells cultured within PEG-FNIIIWT hydrogels were shown to up-regulate integrin expressions α5, αv, β1 and β3, while down-regulation of the same genes was demonstrated for cells cultured within PEG-FNIII9*10 hydrogels, and a more complex pattern of integrin regulation was found for cells cultured within PEG-RGD and PEG-FNIII10 hydrogels. Finally, we investigated a novel approach to growth factor delivery in a controlled release manner, by providing cell- and growth factor binding protein fragments. In this system we show that transforming growth factor beta 1 (TGF-β1) release from PEG hydrogels influence smooth muscle cell specific gene expressions differently depending on the integrin- and growth factor binding proteins available within the hydrogel. Delivery of TGF-β1 within PEG-FNIIIWT hydrogels was shown to up-regulate smooth muscle specific genes in mesenchymal stem cells, while the other conditions displayed more complex patterns of gene expression. Further experiments are required to characterise the conjugation of PEGprotein-growth factor, and the release profiles for the growth factor in the different conditions. In conclusion, this thesis investigates the effect of various chemical and physical signals on mesenchymal stem cell-to-smooth muscle cell differentiation in an in vitro model to further understand cell-cell and cell-extracellular matrix interactions and create improved artificial microenvironments for directing and / or maintaining cell differentiation. The PEG hydrogel that we designed permits the co-culture of cells as well as the incorporation of peptides and proteins such as cell adhesion- and growth factor binding ligands, and growth factors for directing the differentiation of human mesenchymal stem cells into smooth muscle cells. Here, we explore these features and their effects on cell differentiation in vitro. Hence, our PEG hydrogel system can serve both as an in vitro model for studying basic cell biology, and as a cell delivery vehicle for improved bladder tissue regeneration.

EPFL2009

Rat bladder muscle regeneration induced by local delivery of an engineered insulin-like growth factor 1 in fibrin gels

Lirong Yang

Currently, both congenital abnormalities and developmental problems of the bladder in children, and other dysfunctions in adults, require reconstructive surgery. Such correction involves transplant action of native tissues (such as gastrointestinal segments, or mucosa), homologous tissues from a donor, heterologous tissues or substances, or artificial materials to act as a replacement for normal bladder tissue. However, such surgery does not entirely restore the function, as the replacement tissue is either rejected due to immune system, fibrosis, contraction or causes metabolic complications due to a mismatch in different functional parameters, such as gastrointestinal segment for absorbance versus bladder for excretion. Tissue engineering is emerging as a significant alternative potential treatment for bladder dysfunction. To achieve this goal, we have explored fibrin gels as a natural material based scaffold to seed infiltrating smooth muscle cells to mimic the ontogeny of the bladder tissue. Our interest in developing fibrin-based biomaterial technology is based on fibrin's central role in tissue binding and in the initiation of tissue repair and defence. It is well known that fibrinogen/fibrin binds to platelets as well as to different cells, growth factors, and extracellular matrix proteins, which is critical for the wound repair process. Insulin-like growth factor I (IGF1) is well known as a key regulator in carbohydrate metabolism and growth. It can promote smooth muscle cell growth. In this thesis it was also chosen due to its simple active form, namely a single chain of low molecular weight, 10kDa. A novel engineered insulin-like growth factor I (IGF1) – factor XIIIa substrate fusion protein was chosen as the bioactive macromolecule to be released from the scaffold in order to control cellular growth and differentiation. Therefore, we can achieve close to the natural biomechanical environment in the regenerated tissue. In this work, two generations of fibrin matrices for tissue engineering were developed. The first part of the work was devoted to preventing rapid diffusion of the growth factor from the matrix and to controlling its release. IGF1 was modified by inserting a factor XIIIa substrate sequence (denoted TG) based on the α2-plasmin inhibitor to cross-link into the fibrin gel during coagulation. This is so that the variant IGF1 will bind to the fibrin gel itself and will be released upon matrix degradation. In order to produce this recombinant protein TG-IGF1, IGF1 GST tag plasmid DNA was transformed into BL21 competent cells. After protein production, the recombinant TG-IGF1 protein was purified by GST affinity chromatography. The purified TG-IGF1 was confirmed via SDS-PAGE and western blot with an anti-hIGF1 antibody. The sequence was confirmed using MALDI-TOF mass spectrometry. The biological activity of TG-IGF1 was validated in vitro by receptor tyrosine phosphorylation and metabolic assay (MTT assay) with fibroblast 3T3 cells. The cell proliferation profile was similar for both TG-IGF1 and native recombinant IGF1. Although we were able to obtain relatively pure TG-IGF1 protein, the purification protocol may still require some development due to loss of protein via aggregation during protein production. Using a three-dimension in vitro model, neonatal human bladder smooth muscle cells were seeded in the fibrin gel only, in the fibrin gel either with TG-IGF1 or native recombinant IGF1 as a control. The morphology of the seeded cells was analysed using ultra-structural transmission electron microscopy after three days. Secretory vesicles had not been found in the cells without IGF1. The cells with the TG-IGF1 displayed larger and more numerous secretory vesicles than those with native recombinant IGF1, presumably due to the rapid diffusional loss of the native recombinant IGF1 from the gel. The bladder smooth muscle cell in the fibrin gel responses to either TG-IGF1 or native recombinant IGF1 at the genomic level were analyzed by qRT-PCR for extracellular matrix and adhesion molecules after 24 hours incubation. The cells in the fibrin gel only were used as control. We did not observe a difference in gene expression between the two groups. The bioactivity of TG-IGF1 was also investigated in vivo utilizing a rat bladder model. A wound was induced via resection on the rat bladder. Either the TG-IGF1 or native recombinant IGF1 containing fibrin gel was then applied to the wound site, with a second control group receiving only the fibrin gel alone. Three classic histological stainings were done using Hematoxylin & Erythrosine (HE), Masson's Trichrome (TM) and Prussian Blue (PB). It was found that TG-IGF1 greatly enhanced rat bladder muscle layer regeneration compared to native recombinant IGF1 and the control group of fibrin gel according to the histology staining and quantitative analysis of the ratio of detrusor muscle regenerated / normal detrusor muscle (0.27±0.10 for TG-IGF1 and 0.23±0.10 for native recombinant IGF1). The second part of the work was devoted to making a new generation of fibrin matrices. Fibrinogen (Fgn), a soluble plasma protein found in all vertebrates, is a covalent dimer composed of pairs of three polypeptide chains called Aα-, Bβ- and γ-chains. Fibrinogen Aα chain can be truncated to Bβ-and γ-chains while maintaining the capacity to be assembled into a secreted, detectable fibrinogen. It is feasible to produce recombinant chimeric fibrinogen in cell culture, which can then be used to incorporate growth factors into fibrin matrices as a fusion protein with fibrinogen at the genetic level. In this way, fibrin based natural material hydrogels can be formed which can bind and release signalling biomacromolecules for directed cellular growth and tissue regeneration. In order to produce this new fibrinogen recombinant protein, a novel stable cell line of Chinese Hamster Ovary cells, CHOfgn-hIGF1, needed to be developed. This was accomplished via transfection of an FGA-hIGF1 plasmid DNA into CHOβγ cells. After protein production, the recombinant chimeric fibrinogen variant hIGF1 fusion protein was purified by affinity chromatography. The successful production of these cell lines and the resultant variant IGF1 was confirmed via ELISA with fibrinogen and hIGF1 antibodies, and also by fluorescence activated cell sorting (FACS). Due to protein polymerisation in the column during purification, the protein purification step still requires additional development. Although the targeted fusion protein was successfully produced, overall protein yield was so low as to prevent further exploration. By harnessing the powerful technique of protein engineering and using the fibrinogen/fibrin hydrogels developed in our lab, we were able to demonstrate a new type of multifunctional gel capable of regenerating smooth muscle type tissue with the aim to repair bladder function. Here we were able to demonstrate the production of the variant IGF1 factor XIIIa substrate fusion protein, its biological activity in vitro, and its application in an in vivo rat model.

EPFL2010

Photopatterning of Synthetic Hydrogels

Katarzyna Mosiewicz

In recent years sophisticated biomaterials have made a significant impact on our understanding of cellular behavior, particularly in the fields of stem cell research and tissue engineering. Nevertheless, the promising potential of stem cells for regenerative medicine has remained largely unmet. One reason for such limitations could be that there is still incomplete understanding of stem cell behavior due to a lack of advanced in vitro tools suitable to achieve adequate control over cellular processes. By mimicking the dynamic and complex biophysical and biochemical interactions between cells and their natural extracellular matrixes (ECMs), it is possible that we may bridge this gap between current limited in vitro models and complex in vivo organisms. In this thesis, an innovative class of biomaterials is presented that enables the flexible in vitro recapitulation of the dynamic biochemical and biophysical signaling involved in controlling cell fate. Such systems of increasing fidelity to real physiological processes could bring new insights into fundamental cell biology as well as bring us to closer towards cell-based therapeutics applications. In order to achieve a dynamic artificial ECM (aECM), we designed chemical schemes which would allow us to control the biochemical and biophysical properties of the hydrogel in space, time and intensity. The chosen concept is built on a synthetic hydrogel providing static physicochemical support, while spatiotemporal control over a single or multiple signals is achieved by spatiotemporal photoactivation reactions within this hydrogel. Indeed, by modulating the duration, location and intensity of illumination, desired heterogeneities in biochemical and biophysical signal can be introduced in the otherwise homogeneous and static hydrogel network. To implement the designed hydrogel photo-patterning tool, new chemical components were synthesized. Synthetic hydrogels based on poly (ethylene glycol) (PEG) were utilized due to their inert properties, cell biocompatibility and lack of unspecific protein binding. Biochemical signal tethering was rendered spatiotemporally controllable by adding photosensitivity to the existing enzyme-based crosslinking scheme. This was achieved by functionalizing one of enzymatic substrate with a photo-labile “caging” moiety, thereby preventing crosslinking to occur until light illumination. This caged substrate was incorporated into the hydrogel network and served as a light-controllable and specific ligand-binding site. Importantly, it was possible to photo-pattern proteins, which are much complex than peptide-based biomolecules, and their bioactivity was fully preserved with this site-specific immobilization scheme. It was also possible to pattern biophysical cues by making the Michael-type (MT) addition reaction for crosslinking PEG-based hydrogel photosensitive. In this case, the thiol moieties of one of the reactive PEG macromers undergoing crosslinking, were equipped with a caging group which prevented its susceptibility to the counter-reactive functionality, the vinyl sulfone group. Thus, the crosslinking density of the hydrogel backbone could be controlled by light, which was directly translated into differential patterns of hydrogel stiffness. Using this approach, user-defined biochemical and biophysical patterns and even gradients could be obtained in the spatiotemporal manner within the same hydrogel. The crosslinking reaction is the critical parameter in the success of hydrogel photo-patterning, and must rely on a well-controlled and highly selective chemical scheme. Therefore, we also developed a novel enzyme-based cross-linking scheme for synthetic PEG hydrogels. The reaction originates from the naturally occurring post-translational modification of proteins by phosphopantetheine (Ppant) group performed by a family of enzymes known as Ppant transferases (PPTases). By synthesizing biochemical components of this reaction, we implemented a PPTase-catalyzed reaction to form and bio-functionalize PEG hydrogel. The utility of the developed hydrogel platforms was determined in the context of 2D and 3D cell culture. Photo-directed patterns of a cell-adhesive fibronectin fragment triggered adhesion and spreading of muscle progenitor cells, a phenomenon which occurred only within the areas of patterned signaling protein. A photo-patterned elasticity gradient was used to direct mesenchymal stem cell spreading and migration. Finally, the PPTase-based hydrogel was proven to be suitable for 3D culture, as shown by encapsulation of primary human fibroblasts. In this thesis, we demonstrated that a combination of chemical schemes could be assembled to achieve a functional materials toolbox which would begin to reconstruct the complexity of the dynamic in vivo ECM. Such a tool is expected to contribute to the field of cell biology by allowing us to study questions which previously could not be addressed in vitro, and could ultimately contribute to cell-based regenerative medicine.

EPFL2012