SCAD device fabrication
A solution of 25% polystyrene (Sigma-Aldrich, 182,435) in N,N-Dimethylformamide (FUJIFILM Wako, 047–29,191) was prepared and vigorously mixed for 30 min. Then, the mixed solution was vertically rotated at low speed (1 rotation per 25 s) for 10 h. The resultant solution was transferred into a 5-ml plastic syringe (HJ4050-LL, OSAKA CHEMICAL Co., Ltd.) with a 25-gauge needle (490,732, BSA Sakurai) and loaded into the electrospinning apparatus (NANON-04, MECC CO., LTD). Electrospinning of the solution was performed at ambient temperature, with an applied voltage of 9–10 kV DC, injection rate of 1.7 ml/h, tip-to-collector distance of 130 mm, and rotational speed of 2,000 rpm for 13 min once or 6.5 min twice, and fibers were collected throughout. The fibers were collected onto a commercial A4 size paper (0001-PRKA4-BK-01, Etranger di Costarica) and pasted with glue (KE-45-T, Shin-Etsu Chemical Co., Ltd.) to stainless-steel washer with a polycarbonate frame.
SCAD device preparation and cell culture
Before cell seeding, the SCAD device was subjected to 4 min of plasma treatment (PC-40 T, STREX Inc.) followed by 1 min of UV irradiation. Then, the device was coated overnight at 37 °C with 0.02% poly-L-ornithine (P4957, Sigma-Aldrich). After washing with phosphate-buffered saline (PBS), the device was coated with 2.5 µg/mL laminin 511 (381–07,363, Wako) for 2 h at 37 °C.
Cryopreserved hiPSC-derived cortical neurons (XCL-1 Neurons, XCell Science) were thawed and suspended in Neuron Medium (XCS-NM-001-M100-1P, XCell Science). For dispersed culture, 5.4 × 104 cells (6.0 × 105 cells/cm2) in 20 µL Neuron Medium were seeded on a treated SCAD device. Two hours later, devices with cells were transferred into a well of a 24-well plate filled with 1 mL Neuron Medium. After 1 week, the medium was replaced with 1 mL of BrainPhys Neuronal Medium containing SM1 neuronal supplement (ST-05792, STEMCELL technologies), and 20% Astrocyte conditioned medium (1811-sf, ScienCell Reserch Labratories). Afterward, half the volume of the medium was replaced twice per week.
DRG neurons were harvested and cultured as described previously [57] . Briefly, DRG neurons were collected from 10 weeks old male Wistar Rats. The ethical approval for this study was obtained from Tohoku Institute of Technology Animal Care and User Committee. Firstly, rats were asphyxiated with isoflurane and then decapitated. DRGs were harvested from the vertebral column, and the sensory neurons were dissociated by mechanical agitation after incubation for 2 h with collagenase type III (CLS3, Worthington) at 37 °C. Then, cells were washed with Hank’s balanced salt solution and further dissociated with trypsin type I (T8003, Sigma-Aldrich). After cell counting, approximately 5 × 104 cells (6.0 × 105 cells/cm2) in 20 µL BrainPhys Neuron Medium were seeded on a treated SCAD device. One hour later, devices with cells were transferred into a well of a 24-well plate filled with 1 mL BrainPhys Neuron Medium. The next day, the medium was replaced with 1 mL of serum-free medium containing 10 µm uridine and 10 µm 2′-deoxy-5-fluorouridine that was kept for 3 days to suppress the proliferation of glial cells. Afterward, the medium was changed back to 1 mL BrainPhys Neuron Medium, and half the volume of the medium was replaced twice per week.
For spheroid formation, 1.0 × 104 cells suspended in 150 µL of Neuron Medium were transferred in a well of a 96-well plate (MS-9096 M, Sumitomo Bakelite Co., Ltd.). After centrifugation at 200 × g for 2 min, plates were placed in an incubator at 37 °C and 5% CO2. Spheroids were cultured on SCAD devices from day 7. Coated devices equipped with a proprietary seeding jig containing two spheroid mounting holes were immersed into 200 µL of BrainPhys Neuronal Medium in a well of a 24-well plate to prevent devices from drying. Using a 200-µL pipette, a spheroid was transferred into each hole of the jig and settled by centrifugation 300 × g for 3 min. Afterward, half the volume of the medium was replaced twice per week.
Immunocytochemistry
Sample cultures were fixed with 4% paraformaldehyde in PBS on ice (4 °C) for 10 min. Fixed cells were incubated with 0.2% Triton-X-100 in PBS for 5 min, then with preblock buffer (0.05% Triton-X and 5% FBS in PBS) at 4 °C for 1 h, and finally with preblock buffer containing a specific primary antibody (1:1,000) at 4 °C for 24 h. The primary antibodies used were mouse anti-L glutamate (ab9440, Abcam), rabbit anti-gamma-aminobutyric acid (GABA) (A2052, Sigma-Aldrich), rabbit anti-MAP2 (ab281588, abcom), mouse anti-β-tubulin III (T8578, Sigma-Aldrich), and rabbit anti-MBP (ab40390, Abcam), respectively. Then, the samples were incubated with the appropriate secondary antibody (anti-mouse 488 Alexa Fluor, ab150113, Abcam or anti-rabbit 546 Alexa Fluor, A11010, Lifetechnologies, 1:1,000 in preblock buffer) for 1 h at room temperature. Cell nuclei were counterstained with Cellstain DAPI solution for 1 h at room temperature. Stained cultures were washed twice with preblock buffer (5 min/wash) and rinsed twice with PBS. The immunolabeling was visualized using a confocal microscope (Eclipse Ti2-U, Nikon). Image intensity was adjusted using the ImageJ software (NIH). A Cell3 imager Estier system (Screen Holding) was used to acquire 3D images, and the images were adjusted by a Cell Visualizer software provided by the manufacturer.
RNA extraction and analysis
Neurons on SCAD device were lysed directly in the culture well by addition of 500 µL TRIzol™ Reagent (15,596,026, Thermo Fisher Scientific). Total RNA was extracted manually using chloroform and isopropyl alcohol solution following manufacturers protocol. An RNA sequencing analysis was entrusted to Agenta Co., Ltd, Tokyo. Briefly, mRNA was extracted by the poly-A selection method targeting mRNA, and the whole genome sequencing was performed using HiSeq X Ten (Illumina Inc.). Expression levels of all mRNAs were presented by the calculated value of fragments per kilobase of exon per million reads mapped (FPKM). After compared the whole FPKM value between hiPSC-derived cortical neurons cultured on SCAD device and those directly cultured on MEA probe for 5 weeks, several typical mRNAs with different expression level (i.e., VGLUT2, GLUR2, NF160, Synaptophysin, Notch1, Nestin, and MASH1) were manually picked-up as shown in Fig. 2C.
Extracellular recording
Spontaneous extracellular field potentials were acquired at 37 °C under a 5% CO2 atmosphere using either a 24-well MEA system (Presto; Alpha Med Scientific) or a CMOS–MEA system (Maxone; Maxwell) at a sampling rate of 20 kHz/channel.
For the Presto system, neurons or neural spheroids cultured on SCAD devices were transferred to MEA plates and incubated for 30 min at 37 °C under 5% CO2 atmosphere just before measurements. Signals were high-pass filtered at 1 Hz and stored on a personal computer. The spikes in the acquired data were detected using the 100-Hz high-pass filter.
Pharmacological tests
Spontaneous activities were recorded for 10 min before treatment and after the cumulative addition to the culture medium of one of the following convulsant agents or receptor antagonists: 4-AP (0.1, 1, 3, 10, or 30 µm; 016–02,781, Wako), pilocarpine (0.1, 1, 3 10, or 30 µm; P6503, Sigma-Aldrich), picrotoxin (0.1, 0.3, 1, 3, or 10 µm; 2,800,471, Nacalai tesque), and AP-5 (1, 3, 10, 30, and 100 µm; 165,304, Sigma-Aldrich). All chemicals were dissolved in DMSO (0.2%–0.6%), which was used as control.
For frequency analysis experiments, 4-AP (0.3, 3, or 30 µm) was administrated into the culture medium, and spontaneous firing was recorded for 10 min.
For propagation analysis experiments using neural spheroids, spontaneous activities were recorded for 10 min before and 20 min after the addition of one of the following typical convulsant agents or receptor antagonists to the culture medium: 4-AP (0.3, 1, 3, 10, or 30 µm), picrotoxin (0.3, 1, 3, 10, or 30 µm), and CNQX (0.3, 1, 3, 10, or 30 µm; C-140, ALOMONE). All chemicals were dissolved in DMSO (0.2%–0.6%), which was used as control.
During all recordings and drug administration, the cultures were kept at 37 °C under a 5% CO2 atmosphere.
Burst analysis
Electrophysiological activity was first analyzed using the Presto software (Alpha Med Scientific) and MATLAB as described before [32] . Briefly, a spike was counted when the extracellularly recorded signal exceeded a threshold of ± 5 σ, where σ was the standard deviation of the baseline noise during quiescent periods. NBs were detected using the 4-step method, which was described previously. Firstly, spikes separated by interspike intervals of 5–15 ms were attributed to the same NB. Secondly, datasets with a maximum number of spikes in the NB below 50–100 spikes/NB were eliminated from the analysis. Thirdly, NBs separated by inter-NB intervals shorter than 100–200 ms were combined. Finally, an NB was defined when it contained more than 500–1,500 spikes/NB. Appropriate numerical values that can accurately detect bursts with 16 electrodes were used as parameter numerical values. All data were expressed as means ± standard errors.
Frequency analysis
Wavelet analyses were performed using a custom-written program in MATLAB (using function cwt in package “Wavelet Toolbox”) as described before [23] . Briefly, the raw data, f (t), were transformed as follows:
$$\mathrm{W}\left(\mathrm{b},\mathrm{a}\right)=\frac{1}{\sqrt{a}}{\int }_{-\infty }^{\infty }f\left(t\right)G\left(\frac{t-b}{a}\right)dt$$
where a and b were the scaling factor (1/Hz) and the center location (ms) of the mother wavelet function, respectively, and 1/a varied from 0.1 to 250 Hz. G(x) is the complex Morlet function:
$$G\left(x\right)=\frac{1}{\sqrt{\pi {F}_{B}}}\mathrm{exp}\left(-\frac{{x}^{2}}{{F}_{B}}\right)\mathrm{exp}\left(2i\uppi {F}_{C}x\right)$$
where FB = 5 was the frequency bandwidth or wavenumber, and FC = 1 was the center frequency.
The wavelet power spectrum, W (b, a), is shown. The amplitude of this transform was obtained from its absolute value and color-coded. A scalogram was drawn with the Y-axis representing the frequency band as 181 pixels and the X-axis representing time. One pixel on the X-axis was 50 μs.
$${WT}_{A}=\frac{{WT}_{S}}{{N}_{X} \times {N}_{Y}(f)}$$
WTA: Wavelet transform coefficient per pixel in each frequency band.
WTS: Summation of wavelet transform coefficient in each frequency band.
NX: Number of pixels on X-axis.
NY(f): Number of pixels on Y-axis, f is the frequency band.
CMOS–MEA measurements
For the Maxone system, neurons cultured on SCAD devices were transferred to CMOS–MEA plates and incubated for 15 min at 37 °C with 5% CO2 just before measurement. A whole-sample active scan followed by a local active recording were performed for each sample based on the manufacturer’s protocol. Briefly, about 1,020 electrodes that recorded relative high spike amplitude were selected during a whole-sample active scan divided into several blocks. Then, spontaneous firing activities were recorded on these selected electrodes, and the spike amplitude data were outputted for MATLAB analysis. To calculate the velocity of the synaptic propagation between single firing neurons, the location of firing neurons was determined from the magnitude of the spike amplitude, and a raster plot was generated for the identified neurons to detect the propagation delay between neurons. The network propagation velocity was calculated from the distance between neurons, and the delay before the first spike appeared in the NB of each neuron (Fig. 6A). To calculate axon conduction velocity for peripheral neurons, the pathway map of the axon conduction and the axon traces were identified based on the magnitude of the spike amplitude. Briefly, cell bodies of firing neuron were identified based on the magnitude of the spike amplitude after a local active recording, as described above. Then, the average waveform of every electrode during a very short period before and after the firing time point of one certain identified neuron (i.e., 1.5 ms before the firing time point of cell body and 2.5 ms after that, totally 4 ms), was calculated. And electrodes with relatively similar waveform pattern were manually picked up as the signal pathway of axonal conduction. This progress would be repeated for every identified cell body, to find its pertinent axonal pathway. Finally, the axonal conduction velocity was calculated using a linear fit of the interelectrode distance versus the spike-time latency (Fig. 6B).
Statistics
One-way ANOVA followed by Dunnett’s test was used to determine the significance of the differences between 2D cultured neurons and neurons cultured on SCAD devices (Fig. 2B), and the differences between each drug concentration and the vehicle (Figs. 3B, 4A, 5C). The differences between between low-frequency components before TBS and after TBS (Fig. 4B) were analyzed using two-tailed paired Student’s t-test.