Long-Term Single Cell Analysis of S. pombe on a Microfluidic Microchemostat Array

Although Schyzosaccharomyces pombe is one of the principal model organisms for studying the cell cycle, surprisingly few methods have characterized S. pombe growth on the single cell level, and no methods exist capable of analyzing thousands of cells and tens of thousands of cell division events. We developed an automated microfluidic platform permitting S. pombe to be grown on-chip for several days under defined and changeable conditions. We developed an image processing pipeline to extract and quantitate several physiological parameters including cell length, time to division, and elongation rate without requiring synchronization of the culture. Over a period of 50 hours our platform analyzed over 100000 cell division events and reconstructed single cell lineages up to 10 generations in length. We characterized cell lengths and division times in a temperature shift experiment in which cells were initially grown at 30 degrees C and transitioned to 25 degrees C. Although cell length was identical at both temperatures at steady-state, we observed transient changes in cell length if the temperature shift took place during a critical phase of the cell cycle. We further show that cells born with normal length do divide over a wide range of cell lengths and that cell length appears to be controlled in the second generation, were large newly born cells have a tendency to divide more rapidly and thus at a normalized cell size. The platform is thus applicable to measure fine-details in cell cycle dynamics, should be a useful tool to decipher the molecular mechanism underlying size homeostasis, and will be generally applicable to study processes on the single cell level that require large numbers of precision measurements and single cell lineages.

Hydrogel Microfluidics to Control Stem Cell Fate

Steffen Cosson

Biomolecular signaling is of utmost importance in governing many biological processes such as morphogenesis during tissue development where biomolecules regulate key cell-fate decisions. In vivo, these factors are presented in a spatiotemporally tightly controlled fashion and in the context of a soft and hydrated microenvironment. Although state-of-the-art microfluidic technologies allow precise biomolecule delivery in time and space, long-term (stem) cell culture at the micro- scale is far from ideal due to issues related to medium evaporation, limited space for cell growth, shear stress and a lack of cell-instructive microenvironments that might adversely impact cell-fate. As a result, microfluidic cell culture systems are not yet suitable to unravel complex multicellular phenomena. This may explain why they are not yet widely used in biology laboratories. Consequently, the overall goal of this thesis was to overcome this technology gap by developing next generation microfluidics through a combination of microfabrication and smart biomaterials. A special emphasis was placed on decoupling biomolecule presentation at micro-scale from macro-scale cell culture in order to enable long-term stem cell culture within user-friendly multiwell plate formats. First, a novel microfluidic concept was invented to rapidly immobilize linear protein gradients on the surface of poly(ethylene glycol) (PEG)-based hydrogels whose biophysical properties are reminiscent of natural extracellular matrices. This method allows efficient capture of steady-state gradients of tagged proteins (e.g. using Fc or biotin tags) in just a few minutes on engineered hydrogels that display the corresponding auxiliary proteins (e.g. ProteinA or NeutrAvidin). The selectivity and orthogonality of the chosen binding schemes enables the formation of parallel and orthogonal overlapping gradients of multiple proteins, which is impossible using existing platforms. After patterning, the microfluidic chip can be readily removed from the gel for cell culture applications. Using quantitative single-cell time-lapse microscopy, this platform was validated here by probing the effect of fibronectin concentration on the directionality and speed of cell migration. Next, the resolution, flexibility and throughput of the protein patterning on hydrogels was substantially enhanced by the introduction of hydrodynamic flow focusing, that is, the generation of user-defined patterns by spatially controlled step- wise deposition of biomolecules. To this end, a microfluidic device was conceived that enabled the generation of arrays of parallel and crossed overlapping gradients. Application of the platform to generate gradients of immobilized leukemia inhibitory factor (LIF) showed an influence of the LIF concentration on ESC self-renewal. This platform should be useful to systematically probe the effect of biomolecule dose, singly or in combinations, on stem cell behavior in vitro; however, it lacks the ability to dynamically vary the presentation of biomolecules that might be crucial to control stem cell fate towards establishing functional in vitro tissue models. To address this limitation, in the last part of the thesis work a hydrogel-based microfluidic chip was developed that decouples macro-scale cell culture on the gel surface from the precise spatiotemporal biomolecule delivery at the micro-scale, i.e. through the gel layer. The hydrogel chip, used here as a simple insert that is compatible with conventional multiwell plate formats, was optimized to support long- term ESC maintenance in both adherent format and as uniformly sized ESC aggregates termed embryoid bodies. The platform was successfully used for the spatially controlled neuronal commitment of ESC via delivery of gradients of the morphogen retinoic acid. Taken together, in this thesis several innovative microengineered cell culture platforms are presented which enable, perhaps for the first time, the long-term culture and manipulation of stem cell fate at the micro-scale. These systems should be useful to systematically probe in vitro the effect of biomolecule dose and delivery dynamics on stem cell behavior, ultimately facilitating the development of complex in vitro tissue models.

EPFL2012

Single-cell dynamics of the chromosome replication and cell division cycles in mycobacteria

Djenet Bousbaine, Neeraj Dhar, John McKinney, Isabella Santi

During the bacterial cell cycle, chromosome replication and cell division must be coordinated with overall cell growth in order to maintain the correct ploidy and cell size. The spatial and temporal coordination of these processes in mycobacteria is not understood. Here we use microfluidics and time-lapse fluorescence microscopy to measure the dynamics of cell growth, division and chromosome replication in single cells of Mycobacterium smegmatis. We find that single-cell growth is size-dependent (large cells grow faster than small cells), which implicates a size-control mechanism in cell-size homoeostasis. Asymmetric division of mother cells gives rise to unequally sized sibling cells that grow at different velocities but show no differential sensitivity to antibiotics. Individual cells are restricted to one round of chromosome replication per cell division cycle, although replication usually initiates in the mother cell before cytokinesis and terminates in the daughter cells after cytokinesis. These studies reveal important differences between cell cycle organization in mycobacteria compared with better-studied model organisms.

Nature Publishing Group2013

Automated long-term single cell analysis of Schizosaccharomyces pombe on a microfluidic microchemostat array

Jean-Bernard Nobs

How cells grow and divide is a question that was raised more than 100 years ago. Despite years of research and in-depth knowledge of several molecular mechanisms, the complete system that regulates and coordinates both of those activities within Schizosaccharomyces pombe is still unknown. A systemic approach might provide some help to decipher them. Fission yeast is one of the most actively used model organisms to study cell-cycle and cell growth. The first single-cell measurements of its growth profile were performed already in the nineteen-sforties. Despite the huge amount of studies performed, only very few methods were provided to scientists to investigate growth profiles. All the proposed approaches are highly time consuming and have technological drawbacks. Furthermore, they do not allow long-term imaging and are time-limited by their design. Recent advances in the microfluidic field allowed to design culture chambers that can achieve long-term and stable culturing of cells in a constant and tightly controlled environment. The surrounding cell growth conditions can be rapidly and easily changed over time. In this thesis, we aimed at coupling microfluidic devices with an automated microscope and an image analysis pipeline to record accurate and numerous growth profiles. Our platform allowed to handle tens of thousands of images and we could record more than 100’000 division events within a single fifty hour experiment. We automatically extracted data regarding basic physiological parameters such as cell size, time to divisions and elongation rate. Our complete analysis pipeline does not require any specific "marker" and relies exclusively on bright-field images. Using our platform, we characterized the evolution of several parameters regarding cell length and time to divisions during a temperature shift experiment in which cells were first grown at 30° followed by a transition to 25°. A similar experiment was performed in an upward transition from 30° to 33°. Cell length was nearly identical at the three temperatures, whereas time related parameters as well as elongation rate underwent changes. We quantified the distributions of each parameter at steady state and were able to detect transient size changes in respect to the single-cell growth phase when facing the transition. Indeed, an upshift of temperature induced an immediate division in cells that are less than halfway through their cell cycle resulting in a shortened cell length at divisions. In the opposite direction, a downshift of temperature delayed the beginning of the division process for cells that accomplished less than half of their cycle, leading to an increased cell length at division. Furthermore, those transient effects impacted the daughter cells (second generation) before the situation was back to normal in the third generation of cells. The involved temperature sensitive processes are highly correlated to the cell-cycle status of single-cells and despite the fact that our culture was unsynchronized it was possible to detect them precisely. Our platform is thus applicable to measure fine details of S.pombe cell-cycle dynamics and could be an interesting tool to help unravel the complex mechanisms maintaining cell homoeostasis. Basically, the study of any cell-cycle related processes that require accurate and numerous numbers of division events can be pushed forward with such a tool.