Patterned Neurons in a Dish: In the Seek of Order

Abstract

The brain is one of the most complex organs in the mammalian body. Electrical and chemical interaction of an incredible amount of cells gives rise to mind the way we perceive, remember, learn and act. With such huge complexity in mind, it comes to no surprise that we still are far from understanding the processes behind. In the past years though, different research fields started to cooperate to achieve one common goal. Thus, many new tools were developed which allowed gaining insights on the interactions of individual neurons and systems of nerve cells. In this work, I established and designed tools to manipulate and interact with neurons based on a bottom-up strategy. In contrast to work with the brain as one entity and investigate outputs of the system in terms of behavioral changes of the organism, the bottom-up approach represented in this thesis focuses on a small part of cells. Isolating small amounts of cells, extracted from the brain and cultured in a dish, helps to improve the reproducibility and target-selectivity of stimulations. In an ideal case, the behavior of each cell within this simplified network could be interfaced in a bidirectional way such that each individual cell could be measured and modified. Findings in such abstract systems can later serve as a basis to extrapolate the elementary functions of small circuits to a more complex structure within the brain. The primary aim of this work was to establish and test a platform, which can provide engineered, small and functional neuronal networks with defined topology and connectivity, together with the option of bidirectional interaction with the neuronal cells. In the first part of this thesis, an approach to reach patterns in the scale of single cells is presented. The developed system made it possible to adapt the surface pattern during the culture process. Therefore, it was possible to place the cells at a defined location, but furthermore also define the directionality of the connection between two cell clusters. This approach is based on the fluidic force microscope (FluidFM), which combines microfluidics with an atomic force microscope (AFM) system. Since the closed fluidic system allows the FluidFM to operate in liquid environment, modifications of the surface were also possible in presence of cultured cells. Starting from a non-adhesive background preventing attachment of cells, local modifications of the surface with the FluidFM allowed attachment of cells at specific sites. Since further modifications were possible after the cells already attached, delayed patterning of cell adhesive cues allowed inducing outgrowth of neurites in a predefined direction. In a first approach, we used the exchange of a non-fouling polymer (Poly(L-Lysine)-graft-Poly(Ethylene Glycol); PLL-g-PEG) by an adhesive polymer (Poly(L-Lysine); PLL) to induce such local adhesive sites on the surface. Since the polarity of neurons has shown to be highly influenced by different guidance cues, we further extended the system. Instead of a system based on the unspecific binding of molecules to the PLL layer, adaptation of a system based on the avidin/biotin interaction allowed attaching proteins specifically to a non-fouling surface. The flexibility of such a system could be demonstrated by sequentially patterning two different types of cells. To be able to test the activity of cultured neurons a calcium indicator and extracellular recordings were used in this work. Although the extracellular recordings have the advantage of a better time resolution, the disadvantage is the need of special culture dishes with embedded electrodes. Especially if random influences need to be excluded from the measured activity, a high amount of parallel cultures becomes important. Therefore, we tested the feasibility of cellulose as a culture substrate. Culturing cells on filter paper allowed for a higher amount of parallel cultures. Only during the measurement, the paper were sequentially transferred onto the electrode chip. Therefore, blockage of the electrode chips during the whole culture period could be avoided. In addition, the physical structure of the paper could be easily modified with a laser cutter, allowing macroscopic confined patterns of neuronal cells. The fiber structure of the filter paper furthermore allowed neurons to extend their processes in three dimensions. To allow also interacting with the neurons at the single cell level, the last part of this work focused on the local chemical stimulation of single neurons. In this case, the FluidFM could be utilized to locally deliver a neurotransmitter and therefore stimulate the cell below the cantilever. In combination with extracellular recordings, the system allowed to investigate changes within the whole network induced by such local chemical stimulation. Whereas local delivery of substances can also easily be achieved with glass pipettes, we could show that the stimulation efficiency is depending on the distance to the surface. The force control of the FluidFM in this case allowed for an exact calibration of the distance. Altogether, the presented toolset supports a highly flexible approach to achieve oriented neuronal networks with controlled connectivity. The FluidFM provides a method to extend a surface pattern even when the cells are already in place. This showed to be helpful for active guidance of the processes formed by the neurons, but also for a potential application of patterning of multiple cell types. Paper as a substrate for neuron cultures on the other hand allows for increased throughput instead, but with a limited resolution when it comes to single cell patterns compared to the approach with the FluidFM. In addition, the fiber structure of the filter paper provides the cells with a three dimensional environment to which the cells showed to respond. Laser cutting of the filter paper constitutes a fast solution to increase also the flexibility of this approach and to minimize the development time if a new design needs to be tested. Finally, the free movement of the FluidFM within the dish together with the force feedback ensured the flexibility to interact with the neuronal network at the single cell level, supporting the simulation of external inputs to the network at any time. Within this work, we show first applications of the different methods to present their potential in the field of neuroscience. Once we improve our experience with the patterning process of single cells, the tools developed in this thesis: the positioning of neurons, the directional connectivity, the writing of different chemical guidance cues in combination with the local chemical stimulation at the single cell level will help to tackle bigger questions.