Dongjo Yoon1,Yoonkey Nam1
Korea Advanced Institute of Science and Technology1
Dongjo Yoon1,Yoonkey Nam1
Korea Advanced Institute of Science and Technology1
The In vitro biological neuronal network model has been widely used as a tool to improve understanding of neuronal networks by lowering the complexity of the native brain. Among them, the three-dimensional (3D) neuronal network model embedded in the extracellular-based matrix (ECM) is known to be closer to the gene expression and functional activity of the native brain than the flat cultured neuronal network. With these advantages, the 3D neuronal network model has been developed into models with various structures, such as modular structures and layer-by-layer structures. Microelectrode array (MEA) device has the advantage of being able to track the network activity for a long time with high-temporal resolution and has been used to measure the functional activity of a 3D neuronal network. However, signal measurement of the 3D neuronal network models had a limitation in that most of the neurons could not be measured by electrodes because neurons were randomly distributed in 3D space and far from electrodes. 3D MEAs which have electrodes in 3D have been developed, but there are still many neurons that cannot be measured compared to the total neurons in the network.<br/>In this study, we introduce an engineered 3D neuronal network model in which neurons are located only on electrodes. This model can be used as a 3D neuronal network read-out platform to read most of the neurons in the 3D network. In order to develop this 3D network structure, we used the previously reported alginate patterning technique on flat MEA (electrode diameter: 30 μm, interval: 200 μm). Alginate hydrogel was biocompatible and neuron-repulsive so that neurons can be attached to the desired region (diameter: 100 μm) on microelectrode. Then, collagen type 1 (Col1, 1.0 mg/mL) which is ECM-based hydrogel was covered on the cultured cells at 3 days in vitro (DIV). For stable adhesion of Col1, a polydimethylsiloxane (PDMS) sheet (outer diameter: 7 mm, inner diameter: 4 mm, thickness: 200~250 μm) was attached to the bottom of the MEA, and a nylon mesh (Outer diameter: 4 mm, inner diameter: 50 mm) was inserted in the middle.<br/>We investigated the morphological characterization of our model through immunostaining and confocal microscope (LSM880, Zeiss) imaging. It was confirmed that somas were maintained only in the originally placed area even after 3 weeks. Neurites grew more densely in space over 5, 7, and 14 DIV. The neurite density decreased as the height increased, but on the top interface, the neurites were concentrated. There are also dendrites and synapses in 3D space. We verified that our model was formed as a 3D neuronal network through morphological characterization.<br/>In addition, we measured spontaneous activity for 10 to 28 DIV. The 2D neuronal network produced without covering Col1 was compared as a control group. The ratio of the active channel (firing rate > 0.1 Hz) and the amplitude of spikes were similar without significant differences from the 2D network. It was confirmed that the process of covering Col1 had no adverse effect at the electrode-neuron interface. The array-wide firing rate, burst rate, burst duration, percentage of burst spikes, and network synchronization index increased until 19 DIV and were maintained thereafter. As the result of the peristimulus time histogram (PSTH) through electrical stimulation, the 3D network showed a longer reaction during 200-400 msec compared with the 2D network which had 50 msec. It is inferred that more diverse paths are formed by the synapses in 3D space than the 2D network, so the synaptic reverberation is maintained long.<br/>Our platform was able to successfully measure the activity of biological 3D neuronal networks and alginate patterns covered with Col1 maintained the position of all neurons on the electrode for 4 weeks. It can be used as a 3D neuronal network read-out platform to measure the signals from most neurons in the network for precise network analysis.