Electrical Responses and Spontaneous Activity of Human iPS-Derived Neuronal Networks Characterized for 3-month Culture with 4096-Electrode Arrays

Abstract

The recent availability of human induced pluripotent stem cells (hiPSCs) holds great promise as a novel source of human-derived neurons for cell and tissue therapies as well as for in vitro drug screenings that might replace the use of animal models. However, there is still a considerable lack of knowledge on the functional properties of hiPSC-derived neuronal networks, thus limiting their application. Here, upon optimization of cell culture protocols, we demonstrate that both spontaneous and evoked electrical spiking activities of these networks can be characterized on-chip by taking advantage of the resolution provided by CMOS multielectrode arrays (CMOS-MEAs). These devices feature a large and closely-spaced array of 4096 simultaneously recording electrodes and multi-site on-chip electrical stimulation. Our results show that networks of human-derived neurons can respond to electrical stimulation with a physiological repertoire of spike waveforms after 3 months of cell culture, a period of time during which the network undergoes the expression of developing patterns of spontaneous spiking activity. To achieve this, we have investigated the impact on the network formation and on the emerging network-wide functional properties induced by different biochemical substrates, i.e., poly-dl-ornithine (PDLO), poly-l-ornithine (PLO), and polyethylenimine (PEI), that were used as adhesion promoters for the cell culture. Interestingly, we found that neuronal networks grown on PDLO coated substrates show significantly higher spontaneous firing activity, reliable responses to low-frequency electrical stimuli, and an appropriate level of PSD-95 that may denote a physiological neuronal maturation profile and synapse stabilization. However, our results also suggest that even 3-month culture might not be sufficient for human-derived neuronal network maturation. Taken together, our results highlight the tight relationship existing between substrate coatings and emerging network properties, i.e., spontaneous activity, responsiveness, synapse formation and maturation. Additionally, our results provide a baseline on the functional properties expressed over 3 months of network development for a commercially available line of hiPSC-derived neurons. This is a first step toward the development of functional pre-clinical assays to test pharmaceutical compounds on human-derived neuronal networks with CMOS-MEAs.

Keywords: CMOS-multielectrode arrays; iPSC-derived neurons; neural networks; spontaneous and evoked activities; surface functionalization.

Figure 1

Overview of the experimental platform used in this work for high-resolution electrical and optical read-outs. (A)(from left to right); Real-time hardware for electrophysiological recordings from CMOS multielectrode arrays (CMOS-MEAs); view of a CMOS-MEA that features on-chip recording and stimulation electrodes (3Brain GmbH, Switzerland); scanning electron microscope (SEM) image of an area of the electrode array, with recording and stimulating electrodes. (B) Overview of the experimental read-outs used in this work, combining optical and electrical measures, i.e., left: electrical using CMOS-MEA; right: optical by means of confocal microscopy. (C) Time line of the experiment and time-points for optical and electrical measures.

Figure 2

Optical read-outs allow to track the development of human iCell neuronal networks over 3-month culture and to discriminate the effects of biochemical adhesion molecules on neuronal growth and function. (A) Fluorescence images elucidate differences in neuronal growth after 1 DIV on CMOS-MEAs and glass coverslips coated with PDLO, PEI, and PLO adhesion molecules. Cultures grow homogenously on PDLO and PEI, but they cluster and fascicule on PLO surfaces. Scale bars represent 100 and 50 μm, from left to right, respectively. (B) Confocal micrographs of iCell neurons at 8, 28, and 90 DIV show expression of typical neuronal and synaptic proteins, i.e., Synaptophysin-1, GAD-65, and PSD-95 (blue), V-GAT, and c-FOS (red), in addition to MAP-2 (green). At 28 DIV, neurons grow on PDLO, and PEI show no distinguishable differences between their cellular neuronal activity (c-FOS) and presynaptic expression (synaptophysin-1), whereas at 90 DIV, the expression of post-synaptic protein PSD-95 is prominently higher in cultures grown on PDLO than others grown on PEI, which provides evidence of different maturation profiles. Scale bars represent 50 μm for (8, and 28 DIV), and 30 μm for 90 DIV. (C) Quantification of PSD-95 expression. To quantify the puncta of PSD-95 constructs, images were processed with the granulometric filtering (see Materials and Methods and Supplementary Figure 2). Cross-sections elucidate the intensity and density of PSD-95 constructs. Blue and red cross-sections depict the fluorescence and the corresponding filtered intensities, respectively. Single and double asterisks illustrate the starting position of the cross-section, i.e., 0 μm, and its end, i.e., 40 μm, respectively. The position of PSD-95 puncta was defined above an arbitrary offset, i.e., 0.07, here indicated by gray bars. Scale bar represents 20 μm. (D) Bar graph showing the quantified effect of surface functionalization on the expression of PSD-95 ensembles (density). PSD-95 level significantly reduced 2.5-fold in neurons grown on PEI compared with neurons grown on PDLO-coated substrate, (n = 3, at least 9 image fields per sample, ^**p < 0.01). (E) At 90 DIV, quantifications performed on confocal micrographs (n = 3, at least 9 image fields per sample, ^**p < 0.01, ^***p < 0.001) illustrate lower expression of V-GAT in neurons grown on PDLO than those grown on PEI. Scale bars represent 50 μm (left), and 20 μm (right) for the magnified region. N.S, indicates non-significant.

Figure 3

High-resolution electrical read-outs with CMOS-MEAs unveil changes in spiking activity of iCell neuronal networks during development. (A)Typically recorded extracellular signal traces associated with the detected firing rates during development. The activity develops from single spikes (8 DIV), tonic firing (28 DIV) to bursting, and synchronized spikes (81, and 90 DIV). Red arrows indicate start of propagating burst. (B) Bar graph showing the effect of surface functionalization on the evolution of network-wide spiking activity, i.e., total spontaneous spikes, per 10 min recording phases. Spontaneous firing rates of cultures grown on PDLO coated substrates tend to gradually increase during development to reach a peak at 81 DIV, followed by a plateau at 91 DIV. A similar initial trend (lower rate) is observed for cultures grown on PEI coated substrates cultures, with a peak at 73 DIV (earlier than PDLO cultures). Afterward, firing activity significantly declines at 81, and 90 DIV (n = 3 for each substrate, ^*p < 0.05, one-way ANOVA). Significant differences of firing activity measures between substrate (PDLO vs. PEI) and DIV (8, 14, 28, 61, 73, 81, and 90) were determined by two-way ANOVA followed by Tukey's post-hoc test (^*p < 0.05). (C) Bar graph illustrating the dynamic changes in the total number of active electrodes over 3 months of culture, on PDLO and PEI coated substrates. Consistently with the development of the spiking activity shown in (B), the number of active electrodes indicates an increasing tendency. A higher increase is observed for cultures grown on PDLO than PEI and this tendency becomes significantly higher for cultures on PDLO at 81, and 90 DIV (n = 3 for each substrate, ^*p < 0.05, one-way ANOVA). Significant differences of detected active electrodes between substrate (PDLO vs. PEI) and DIV (8, 14, 28, 61, 73, 81, and 90) were determined by two-way ANOVA followed by Tukey's post-hoc test (^*p < 0.05).

Figure 4

Distributions of the firing frequencies of iCell neuronal networks is lognormal- like and allow to elucidate the dynamic behavior of human iCell neuronal networks during development.(A) Gaussian fits of the firing frequency distributions for three developmental phases of neurons grown on PDLO coated substrates, i.e., at 8, 81, 90 DIV. The means of the distributions at 8 and 81 DIV hold the same frequency position, but a higher peak of the distribution (red) is observed at 81 DIV. At 90 DIV, the mean of the distribution (blue) shifts left toward low-firing rates and the occurrence of active electrodes reaches its peak at this range of low frequency. (B) Neuronal cultures grown on PEI coated substrates hold over development the means of their lognormal-like distributions at the same frequency range (no shift). The peak of these distributions indicates a lower probability of active electrode occurrence during development compared with networks frown on PDLO coated substrates.

Figure 5

Electrical stimulation delivered by electrodes on-chip show evoked responses only at 90 DIV. (A) Sample of evoked signal traces illustrates the confined stimulation artifact to electrodes close to the stimulating electrodes (< 100 μm), e.g., when low-frequency (0.2 Hz) trains of biphasic current stimuli (300 μs per phase, peak-to-peak amplitude of 300 μA) is delivered for 3 min to exemplary electrode S3. Colors of signal traces correspond to the distance from the stimulating site, i.e., correlated with colors of dotted circles in the 64 × 64 array; for instance, the green response occurred at ~6 electrodes distance (~100 μm) from the stimulating site, while blue at ~15 electrode distance (~700 μm), and black trace at ~26 electrode distance (~2100 μm). (B) Exemplary electrical stimuli (S7 and S12) delivered by specific electrodes on the 4096 electrode arrays (left) and their corresponding evoked responses by means of rastergram and PSTH analyses (middle and right, respectively). Two forms of evoked responses were observed and show, fast responses lasting < 100 ms (top), and long-lasting responses up to ~500 ms (bottom). In the middle, rastergrams showing a sequence of evoked spikes that are confined within the green line (top), while evoked spikes are spread along the blue line (bottom). On the right, exemplary PSTHs that are computed from all evoked activities corresponding to their stimulating electrodes (see Materials and Methods). This shows that fast responses are characterized by a higher number of spikes per 25 ms bin size and lasting < 100 ms, in contrast with a lower number of spikes per bin and duration up to ~500 ms in case of long-lasting responses. Note that the negative tough in the PSTH is determined by the adopted artifact removal and smoothing procedures. (C) Average of the first spike timing delays with respect to electrical stimulation computed from two-group responses (fast vs. long-lasting). A delay of about 70 ms is observed from fast responses (green), while the onset of long-lasting responses (blue) occurred significantly over a longer period, i.e., 120 ms (n = 3, ^**p < 0.01). (D) The average of first spike timing is measured with respect to the distance from the stimulation site, by starting from four electrodes (~50 μm) until 30 electrodes (~2827 μm) after stimulus. This indicates that the first spikes in case of fast responses (green) occur within ~60–70 ms after stimulation, regardless of their distance from the stimulus, whilst, first spikes in long-lasting responses (blue) begins at ~85 ms, in close proximity to the stimulus (within four electrodes distance), and delayed to 120–140 ms in the distant ones. Non-significant differences of first spike timing between response (fast vs. long-lasting) and distance in μm (5, 10, 15, 20, 25, and 30) were determined by two-way ANOVA (p > 0.05). Responses of both groups (fast vs. long-lasting) are highly significant over the entire computed distance, i.e., ~2827 μm. (n = 3, p < 2^*10^∧−8, two-way ANOVA).

Figure 6

Categorization of extracellular spike waveforms of iCell neuronal recorded. Waveform classifications of evoked responses detected and sorted at 7.8 kHz sampling frequency. Two main classes were found; belonging to negative spikes namely (monophasic 41%, biphasic 15%, and triphasic 26%), and positive biphasic spikes (18%).