• Automatized Single Cell Assays for Quantitative Time-Lapse Microscopy, 2011
Within this master thesis, a single-cell array based on cellular self-organization has been developed. The idea is to create adhesive sites that favor permanent cell adhesion and repellent areas on which cells can migrate so that the cells distribute themselves according to the surface affinity, transiting from a random to an ordered distribution. In the first part of this work, the appropriate surface chemistry is established. For this purpose, a novel technique is developed that combines micro plasma induced patterning (μPIP) and surface passivation with PLL-g-PEG. This technique enables to create adhesive islands for the cells that can be modified with various proteins such as fibronectin (FN) while the PEGgylation prevents the attachment of the cells in between. In a second step, the parameters such as size of the sites, distance between the sites and initial cell concentration are optimized. To describe such a system, an order parameter is introduced describing the filling of the sites and the misplaced occupations. The time evolution of this order parameter is studied by time-lapse microscopy. The results can be taken to optimize the above described net parameters. In a single cell array, the exact position and shape of each cell are known. This fact allows a fast automatized readout of the experiment and the tracking of single cell traces for long periods of time. For this task, within this thesis, a MATLAB program was implemented. As a proof of principle for time-lapse measurement, the single cell array and the automatic readout are applied to assess the apoptosis induced by nanoparticles (NPs). A two color fluorescent marker and time-lapse microscopy are used to study single cell apoptotic traces of HuH7 cells. Preliminary studies are performed with a standard apoptosis inducing agent. Differences in the onset-times and the distributions of the apoptosis rate are analyzed. The single cell array could be modified if the surface is further functionalized. One example is the adsorption of NPs on the surface prior to cell seeding. It has been shown that a fix more defined presentation of NPs avoids NP aggregation . For this reason, the selective adsorption of NPs on the cell adhesive islands is studied. A further functionalization is the deposition of gene delivery complexes, which can be absorbed on a surface in a controlled way (reverse transfection) resulting in a reduce of external noise. Preliminary studies for the reverse transfection on single cell arrays are performed.
• Spontaneous Symmetry Breaking and Collective Cell Migration on Microstructured Surfaces, 2011
Collective migration of cell assemblies plays an important role in a great variety of processes like wound healing, different phases of the embryogenesis or tumor growth. The objective of this work is to gain a deeper insight into the principles of such migrations and to develop theoretical models which account qualitatively and quantitatively for the observed migration patterns. To do so, it is of great importance to understand what drives cells to break the isotropy of their uncoordinated random walk and to alter into collective directed migration. Micropatterning techniques can be exploited to decrease the number of cells in contact and to provide defined boundary conditions, reducing notably the degrees of freedom of the system and constraining the number of possible cell movements. Here, the migratory behavior of epithelial cells seeded on different circle and ring shaped microstructures, artificially created via micro contact printing (_CP) or microscale plasma initiated patterning (_PIP) are studied. In such an environment, the onset of a cohort circular migration after a few hours of confinement, as it also has been reported in the work of S. Huang et al. , clearly represents a symmetry breaking event. It can be observed that cell assemblies rotate with a common constant angular velocity dependent on the number of confined cells. The angular velocity decreases with increasing number of cells per assembly. Also changes in direction of rotation are observed, which seem yet in most cases to be triggered by disturbances such as cell divisions, but can also appear spontaneously. For further simplification of the setup, also the migration within a ring shaped micropattern was studied. To describe the observed behavior mathematically, models that are based on simple energy minimization are not sufficient since they are not able to account for active processes like migrations within a cell layer. Therefore, the standard cellular Potts model was modified by F. Thüroff from the group of Erwin Frey, in order to account for both energy minimization and processes like cell polarization and persistent migration. Finally, the experimental findings are compared with the theoretical simulations.
• Cell-to-cell variability of gene expression in inducible regulatory networks, 2012
• Flow Profile and Associated Alteration of Short-Scale Noise in Channel-Guided Cell Migration, 2013
Collective cell migration plays an essential role in embryonic development, wound-healing and cancer progression. The arising phenomena cannot be explained sufficiently through the well-known motion behavior of isolated cells and instead is likely a result of their interaction. Confluent cell sheets are well-studied model systems for investigating this emerging collective behavior. Varying phenomena, such as the spontaneous occurrence of swirls spanning several cell diameters and cell density dependent glass-like dynamic arrest have already been reported. In contrast, fluid-like behavior of such systems under confinement has not yet been well understood. In this thesis, short-scale motion and long-scale directional flow of Madin-Darby canine kidney (MDCK) cells invading microchannels is investigated using single cell tracking and particle image velocimetry (PIV) analysis. A defined, stationary plug-flow profile is found in the direction of the channel after averaging over characteristic spatial and temporal scales smooths out perpendicular fluctuations. Collective transport of cells along the channel is found to be well-described by the Fisher-Kolmogorov reaction-diffusion equation. A particular focus was placed on analyzing the effect of a collective flow on short-scale fluctuations. Comparison to a stationary case of resting cell sheets reveals that vortex formation decreases under flow, while a mean squared displacement analysis yields an increase in the individual cell motion within the sheet superimposed on the collective migration. These results suggest that short-scale fluctuations might play a role in the mechanism that allows cells to sense density gradients over long distances and thus leads to directed collective flow. The PIV analysis refined in this work was used in a second project to track the displacement of fluorescent beads in matrigel. Highly motile MDA-MB-231 cancer cells embedded in the matrix generate contractile forces on their environment. This contractility was shown to reduce under treatment with the myxobacterial complex chondramide by quantification of reduced movement.
• Microfluidic Shear Flow Devices for Stretching of the von Willebrand Factor, 2013
The von Willebrand factor (VWF) is a polymeric protein, which contributes essentially to blood coagulation. Better knowledge of its synthesis, degradation and mechanisms helps to treat bleeding disorders and thrombotic diseases caused by quantitative or qualitative deficiencies of VWF. Since VWF’s binding site for ADAMTS13 is only accessible after a shear-activated transition into a stretched conformation, the level of mechanosensitive VWF activity strongly increases with higher shear rates and the VWF polymer size, which is exponentially distributed. The goal of this thesis was the developement of a micro reaction rheometer in order to measure the shear-dependency of VWF proteolysis via ADAMTS13 with fluorescence correlation spectroscopy (FCS). The challanges were: FCS compability, wide range of shear rates (10/s−10000/s), shear times of several hours, small sample volume (200μl) and physiological conditions. For flow measurements a 2-focus FCS was built and three devices for shear-flow generation were implemented: surface acoustic waves (SAWs), membrane pumps and a shear cell. While SAWs provide a circulatory flow shearing the complete small volume, they have difficulties in generating high shear rates and efficient cooling. While the samples in the PDMS channels could succesfully be actuate via SAWs, experiments showed fast temperature increase and evaporation of the samples. Membrane pumps are capable of generating high shear rate, but struggle with shearing the whole sample volume continually. Even with a specifically developed control program and thorough calibrated pumps faulty switchings between the set pressures occured at the necessary frequencies moving the measuring volume out of the channel. Furthermore a flow profiling measurement was performed with a single pump for comparison of single focus and 2fFCS. Thereby the easured velocities for single and 2fFCS corresponded at equal pressures, weheras higher pressures yielded corresponding higher velocities. All requirements are met by a custom-made shear cell, which additionally provides homogeneous shear rates. First control experiments in the shear cell with no shear yielded deviation from constant VWF concentration, which were assumed to be caused by evaporation or leakage of the sample. A measurement with applied shear lead to a rapid decrease of VWF concentration indicating clotting of VWF. To overcome these issues the shear cell was adapted with both a distance cylinder preventing evaporation as well as pressure on the bottom glass and a BSA surface passivation hindering VWF clotting. Following measurements indicate ADAMTS13 cleavage. However, to verify this finding and to investigate the actual shear dependency of the leavage process, further measurements are needed.
• Amyloid beta(1-42) aggregation in vitro quanti_ed by fluorescence correlation spectroscopy, 2013
Amyloid beta (Abeta) is a peptide that is part of the cerebrospinal _uid (CSF) and its small aggregates (oligomers) are suspected to cause Alzheimer's Disease (AD). The molecular mechanism that is responsible for AD is still not understood sufficiently due to the lack of experiments that work under in vivo conditions. Fluorescence correlation spectroscopy (FCS) is a highly sensitive optical technique that can measure larger and smaller particles and their interaction simultaneously in solution. In this work, the aggregation process, especially the development of the oligomers in the subcritical concentration regime (B 20 _M) is observed by FCS. The behaviour of synthetic Abeta(1-42), solved in a standard phosphate buffer, is investigated by varying the total concentration of Abeta(1-42). The size information is extracted from the correlation function by application of two models: a maximum entropy method based standard fitting routine (MEMFCS) and a newly developed Gaussian distribution model (GDM). Both fitting models provide size distributions for the particles in solution that are consistent. GDM is superior to MEMFCS because it works faster and a sharper size distribution can be obtained. The size distributions gained from the fitting process are further analysed to follow the kinetics of the aggregation process of Abeta(1-42). Early stages of aggregation of Abeta(1-42) are put on a quantitative basis by following the development of the peak diffusion times and the fractions of these components. The results obtained with FCS are supported by additional experiments like transmission electron microscopy, gel electrophoresis, Thioflavin T assays, fluorescence and bright field microscopy. Gel electrophoresis of the Abeta(1-42) showed that characteristics of the starting material are in agreement with FCS. Transmission electron microscopy is used to characterise the sample after a finished FCS measurement and to study the structure of amyloid fibrils. With thioflavin T assays the sigmoidal behaviour of the aggregation process is observed. The results of these additional experiments are consistent with the results obtained from FCS data analysis. As an intermediate step towards complex bio fluids like rat-CSF, experiments in artificial cerebrospinal fluid (ACSF) are performed. The observed kinetics is much faster than in phosphate buffer. Furthermore, integration experiments are performed to see if labeled Abeta(1-42) integrates into pre-formed oligomers. With FCS, it was possible to follow the aggregation process of Abeta(1-42) and to distinguish different sizes of aggregates.
• Characterizing Cell Motility and Transmigration on Ring Shaped Micro Patterns, 2014
Cell migration is a multistep process that is very important for our life. It plays a key role during embryo-genesis, wound healing and immune response. Cell migration dysfunctions are a reason for diseases like atherosclerosis, arthritis or cancer metastasis. Thus understanding the molecular mechanism and the resulting migration behavior gives hope to develop new therapeutic solutions to many diseases. In the last years, the key elements of the molecular mechanism have been revealed [18, 26]. Nevertheless cell migration it is a complex process that is influenced by many external cues like chemicals, electric fields, mechanical properties and dimensionality of the environment, and is not yet fully understood . On the way to develop a model  which describes cell migration one needs experiments, that allow a systematic phenomenological analysis of the migration behavior in dependence of defined external influences. To study cell migration with defined external influences artificial micro environments are used. With techniques known as soft lithography it is possible to create substrates with different surface coatings. By presenting cell friendly proteins or cell repelling polymers one can constrict cell adhesion and migration to defined geometries on a 2D surface. Yet, a systematical study of cell behavior at a chemical interface and of transmigration over chemical barriers has not been performed. Here we use ring shaped micro patterns with a defined chemical barrier to guide cells on a one dimensional track and study the effect of the chemical barrier on cell migration behavior. In this thesis I first give a short introduction to the fundamental concepts of cell migration and present the applied methods. Afterwards the main findings are shown in a self contained form of a research article. Further results that are not mentioned in this part are presented and discussed in the following chapter.
• Aufbau eines „Fluorescence Recovery After Photobleaching“-Setups zur Messung von
Diffusionskonstanten in Lipid-Doppelschichten, 2015
• Model selection in deterministic models of mRNA transfection, 2015
• Metabolische Regulation durch einen limitierenden Fluss im bakteriellen Phosphotransferase-System (PTS), 2015
The phosphotransferase system (PTS) comprises a network of sugar uptake systems in bacteria that e_ects and regulates the import of many carbohydrates in parallel, including major sugar sources like glucose, N-Acetylglucosamine (NAG), mannose and fructose. Experimental evidence points at a regulatory mechanism within the PTS that arises because its branches compete for a limited ux of phosphoryl groups, required for substrate translocation. Considering the autocatalytic nature of inducible uptake systems, together with the limited availability of a common precursor, an analogy between the PTS and models of population dynamics with interspecies competition can be made. Thattai and Shraiman (TS) proposed a mathematical model which predicts that such a structure in the PTS could only give rise to one robust pattern of phenotypic switching: winner-take-all outcomes . In E. coli, experimental studies of the single-cell enzymatic expression levels after exposure to mixtures of di_erent concentrations of NAG and sorbitol (srb) showed that the NAGspeci _c PTS, PTSNAG, inhibits the srb-speci_c PTS, PTSsrb, at high NAG concentrations. However, at low NAG concentrations, co-expression of PTSNAG and PTSsrb was found in certain cells. The prediction of the TS model is not consistent with these results. Fitness parameters are crucial in the description of competitive processes. The aim of the present thesis is to determine the _tness landscape of NAG-srb competition by measuring the growth rates sustained by the substrates at di_erent concentrations. Growth measurements in batch at high concentrations of NAG and srb revealed a clear pattern of diauxic growth. NAG was found to sustain faster growth and was identi_ed as the preferred substrate, con_rming the results of expression measurements. A microuidic setup was used to measure growth rates at low concentrations of NAG and at low concentrations of srb. Interestingly, this revealed that below a certain concentration threshold srb utilization yields faster growth than NAG utilization, which is attributed to low NAG uptake rates due to limited availability of substrate. This concentration regime matches the regime in which co-expression of the PTSNAG and PTSsrb had been observed, suggesting that low NAG uptake rates may be responsible for the relief of the inhibition of PTSsrb and showing that the emergence of a co-expression phenotype coincides with a qualitative change in the _tness landscape. Finally, the TS model was modi_ed in two key aspects. First, an alternative description of the induction mechanism was incorporated, taking into account that operon induction is a saturating process. Second, a new expression for the phosphor demand from each competing PTS branch was chosen, based on Michaelis-Menten kinetics. This reects the viii Abstract fact that hierarchy between competitors is established by the e_ciency at which phosphorylation of sugar-speci_c PTS proteins occurs, which is an enzyme-catalyzed reaction. The modi_ed model was analyzed extensively, incorporating the _tness parameters determined by growth measurements and _tting the remaining free parameters by comparison with the available experimental data. The model was found to correctly describe the patterns of enzymatic expression observed in experiment, both the inhibition of the srb system in the presence of high NAG concentrations and the co-expression of the PTSNAG and PTSsrb at low NAG concentrations. This shows that the previously collected experimental evidence can be consistently described by a model based on the competition of di_erent PTS branches for a limited ux of phosphoryl groups. Moreover, the presented model can account for the winner-take-all architecture of the PTS, predicted by the TS model, while exhibiting a greater exibility in the description of heterogeneous responses, which are of great importance in the adaptation of bacterial populations to uctuating environments.
• Occupancy and Stochastic Transitions of Cells on Multi-Well Micropatterns, 2015
Cell migration is crucial for embryogenesis, morphogenesis and during cancer metastasis. In recent years, micropatterned substrates have proved to provide a defined, tunable and reproducible microenvironment for immobilisation of single cells, and to control cell spreading and migration. In this thesis, we have designed dumbbellshaped micropatterns for the study of the migratory behaviour of MDA-MB-231 cells. Pattern geometries were varied as to have (1) two square adhesion sites of equal sizes connected by a stripe of varying length, (2) two adhesion sites of different sizes, (3) differently spaced pairs of unconnected square islands and (4) a row of five connected square adhesion sites. The motion of cells and the concomitant stochastic transmigration between the connected sites was tracked by recording the positions of fluorescently stained nuclei over up to 50 h. Hopping rates are sensitive to all the studied variations in the setup. Thus, hopping rates are seen to depend exponentially on the length of the connecting bridges. The absence of a guidance cue leads to a significant decrease of hopping rates. Moreover, they are sensitive to surface area and ligand spacing on the adhesion sites. The continuous hopping behaviour can be interrupted over time, possibly by a change in surface chemistry due to cell migration. Variations in adhesion site area and structure have direct impact on hopping rates and hopping statistics. This is due to the particular design of our micropatterns, which facilitates the growth of a lamellipodial fan on the unoccupied adhesion site. Apparently, the lamellipodial growth and subsequent force-generation are sensitive to the surface architecture of the adhesive islands. Thus, it seems as though we are able to measure relative cellular affinities for surfaces with our setup. The extension of the dumbbell-shaped micropatterns to micropatterns consisting of five squares in a row provides more space for migration. Thus, label-free tracking of up to two cell generations is possible. In that context, the five-wells-in-a-row patterns promote the separation of many-cell systems into independently migrating ix cells. On the single-cell level, they allow the determination of repolarisation times. Our dumbbell-shaped micropatterns can be used as lab-on-a-chip device for tissue engineering or drug screening applications, as they allow to measure the relative affinity of cells for different surfaces and to quantify the direct response of motility on drug treatment.
• The dynamics of apoptosis on the single-cell basis, 2016
Apoptosis is a heterogeneous and time-dependent chain of events, including lysosomal and mitochondrial membrane permeabilization, caspase activation, phosphatidylserine externalization, and cell membrane permeabilization. These can be measured by single-cell time-lapse microscopy on cells fixed to a microstructured surface. Changes in the fluorescence of apoptosis markers reflect events in the apoptosis pathway of the cells. The heterogeneity of the events makes a large amount of observed cells necessary for statistical analysis, requiring automated analysis tools. In this master’s thesis, an automated method for more accurate and quantitative analysis of cell death signaling is presented. An improved background interpolation for a clustering-based background correction tool reduces the bias of the fluorescence signals read out from microscopy images. Its cell recognition algorithm can be further used for automated readout. The fluorescence signals exhibit parameters like the event times, which were for the first time extracted by a new method called PETI analysis in a more robust and flexible way than by the thresholding used before. The PETI analysis fits marker dependent phenomenological functions to the signals by a maximum-likelihood method. The markers of early cell death events (LysoTracker, CellROX, and TMRM) are described by a parabola which, at the signal breakdown, transitions into a constant value. The fit function for the late cell death markers (Caspase, pSIVA, and PI) starts at a constant level, shows a strong increase at the signal onset, and may then decay exponentially if the marker bleaches. From the fits, onset or breakdown times and the signal amplitudes are extracted. Untypical signals originating from cells that do not undergo apoptosis during the observed time are automatically sorted out by marker-specific filter rules to avoid false positives. The accuracy of the parameters automatically derived by this technique was examined. The comparison to manual event time identification showed that the event times of clear signals are reliably determined to a similar value as obtained manually, and that the amount of false positives is at a low level of 2 %. The evaluation of robustness against noise yielded a good accuracy of both the fitting and the parameter extraction for signals with noise below a certain threshold. Single cell data with noise above the threshold is sorted out. Finally, the advantages of the PETI analysis over cross-correlation analysis are discussed. The parameters extracted by the PETI analysis can be used for analyzing the correlations of events in the apoptosis pathway. Statistical consideration of the event times reveals subpopulations and dependencies among different events. These observations allow for a more detailed study of heterogeneities in the apoptosis signaling cascade. They are made possible by the PETI analysis of high-throughput experiments.
• Kinetics of mRNA Onset in Single Cell Arrays, 2016
Since the first non-viral delivery experiment for mRNA transfection, extensive research has been carried out to study and optimize different delivery agents in terms of their efficiency for controlled delivery. There have been some reports on the study of kinetics of mRNA delivery, however the role of transfection agents on the eGFP expression onset times after mRNA transfection still needs further exploration. Here we analyze the distribution in mRNA onset time as a function of physicochemical properties of different transfection agents and target cell lines. Automated time-lapse microscopy combined with micropatterned arrays allows the generation of single cell fluorescence time traces after mRNA transfection. The single cell timecurves obtained have been modeled and analyzed to study the uptake kinetics followed by different transfection agents.
• Functionalization of DNA Origami Structures with Singe Chain Variable Fragments, 2016
In 2006, Paul W. K. Rothemund developed a technique which enabled the production of various DNA structures with a deﬁned tertiary fold via Watson-Crick base paring, which he called DNA origami. Due to its programmability, DNA origami serves as an optimal experimental platform for studying biochemical processes with nanometer precision. Moreover, DNA origami might provide a novel immuno-stimulatory system if decorated with cell targeting molecules, e.g. antibodies. Such modiﬁed DNA origami structures would bind speciﬁcally to, for example, tumour cells generating a strong immune response against the tumour. For all these applications precise functionalization of DNA origami structures is indispensable. In this thesis a method was developed, which generates a covalent and site-speciﬁc one-to-one connection between a recombinant protein and DNA. First, a peptide is linked via dibenzocyclooctyne click-chemistry to an oligonucleotide. Second, the enzyme sortase ligates the DNA modiﬁed peptide to a protein, in this case a single chain variable fragment. Finally, the Protein-DNA construct hybridizes to a DNA origami structure. With various techniques (Gel-Electrophoresis, TEM, AFM, FCS) the functionalized DNA origami was characterized. Apart from minor improvements necessary, the method evaluated in this thesis proved successful and therefore opens up many possibilities for novel applications of DNA origami structures. Furthermore, FCS was evaluated for binding studies with cells, allowing the determination of Kd values in binding equilibrium. In this thesis ﬁrst evidence is given, that this method enables measurements with cells and hence, with further optimization, FCS might proof as a well suited technique for studying ligand - cell receptor interactions with nano-molar precision.