Complex activity and short-term plasticity of human cerebral organoids reciprocally connected with axons

Abstract

An inter-regional cortical tract is one of the most fundamental architectural motifs that integrates neural circuits to orchestrate and generate complex functions of the human brain. To understand the mechanistic significance of inter-regional projections on development of neural circuits, we investigated an in vitro neural tissue model for inter-regional connections, in which two cerebral organoids are connected with a bundle of reciprocally extended axons. The connected organoids produced more complex and intense oscillatory activity than conventional or directly fused cerebral organoids, suggesting the inter-organoid axonal connections enhance and support the complex network activity. In addition, optogenetic stimulation of the inter-organoid axon bundles could entrain the activity of the organoids and induce robust short-term plasticity of the macroscopic circuit. These results demonstrated that the projection axons could serve as a structural hub that boosts functionality of the organoid-circuits. This model could contribute to further investigation on development and functions of macroscopic neuronal circuits in vitro.

Fig. 1. Formation and characterization of the connected organoids in a PDMS-MEA chip.

A Preparation of the connected cerebral organoids. B PDMS-MEA chip schematic. C Gene expression profiles. Relative abundance of mRNAs was normalized to GAPDH. D Axons extended from one organoid to another organoid and formed axon bundles. Scale bar: 1 mm. E GFP-labeled and mCherry-labeled cerebral organoids were connected. Scale bar: 500 µm. F A time-course plot of axon bundle thickness. n = 7 (week 10) or 9 connected organoids (weeks 4–9). Scale bar: 100 µm. G The proportions of excitatory neurons, inhibitory neurons, and other neurons in the organoids. n = 3 organoids. H Immunohistochemical analyses revealed layers of different cell types within the connected organoids after 8 weeks of culture. Scale bar: 10 µm. I Whole-cell patch-clamp recording in neurons in connected organoids after 8.5 weeks from differentiation when hyperpolarizing and depolarizing square wave current pulses were injected. Action potential frequency responded with the different injected current (0 ± 30 pA, 1 s). J (i) Representative images of connected organoids after 5 weeks of culture. Scale bar: 1 mm. (ii) Filtered signals from four representative electrodes under each organoid. (iii) Wavelet coherence between signals from electrodes of each organoid in the connected organoid. K Representative images, filtered signals, and wavelet coherence of the connected organoids after 5.5 weeks of culture. (ii) Black arrowheads represent synchronized burst activities associated with dense spikes. (iii) The wavelet coherence indicated a strong correlation between the two connected organoids. L Synchronicity of activity in the two connected organoids increased during the culture period. n = 4 organoids. M Burst frequency increased significantly with culture time. n = 12 organoids (week 5.5) or 18 organoids (week 6, 6.3, 6.7 and 7). P = 2.4e−4; 7.7e−6 (6.7 weeks and 7 weeks relative to 5.5 weeks). N Magnified view of the plot of neuronal activity of the two connected organoids. Synchronized burst activity was observed with a delay. O Burst delay after different culture periods. n = 20 bursts. *p < 0.05; one-way ANOVA with Tukey’s multiple comparison. LE left electrode, RE right electrode. Data are presented as mean values ± SD.

Fig. 2. Reciprocal connections through bundled axons generate complex neuronal activity.

A LFP signals were extracted from the 0.2–0.5 Hz, 0.5–4 Hz (delta), and 30–300 Hz (gamma) bands by inverse continuous wavelet transformation. At 8 weeks, the connected organoids generated slow-wave oscillations in the 0.5–4 Hz (delta) band. B Integral of power in frequency bands. n = 3 organoids. C Representative neuronal activity of the single, fused and connected organoids. D Burst frequency of the single, fused and connected organoids. n = 5 organoids (single and fused), 10 organoids (connected). P = 0.0123 (single/connected); 0.0171 (fused/connected). E Inverse continuous wavelet transformation in the 0.2–0.5 Hz, 0.5–4 Hz (delta), and 30–300 Hz (gamma) bands. Delta band oscillations were observed in the connected organoids but not in the single or fused organoids. n = 3 organoids. P = 0.0242 (LFP, single/connected); 0.0054 (delta, single/connected); 0.0208 (fused/connected). F Inter-event interval coefficient of variation among the three types of organoids. n = 3 organoids. P = 0.0019 (6.5 weeks, single/connected); 0.0003 (6.5 weeks, fused/connected); 1.2e−5 (7.8 week, single/connected); 9.1e−6 (7.8 week, fused/connected); 4.4e−5 (9 week single/connected); 3.8e−5 (9 week, fused/connected). *p < 0.05; One-way ANOVA with Tukey’s multiple comparison test. Data are presented as mean values ± SD.

Fig. 3. Photo-convertible fluorescent revealed the important role of axon bundle and their specific population.

A Plasmid construct expressing photo-convertible fluorescent protein Kaede, under CAG promotor in pAAV backbone plasmid. Kaede green fluorescent protein can be converted to Kaede red fluorescent protein by UV exposure. B AAV-CAG-Kaede was infected one week after introducing cerebral organoid into microfluidic device. At day 49 (7 weeks of culture), UV light (405 nm laser equipped with a confocal microscope) was irradiated to the axon bundle. Then, the cells were sorted by cell sorter to identify axon bundle-associated neurons (Kaede-red positive) and non-associated neurons (Kaede-red negative). C Kaede photo-conversion in connected organoids before and after UV light-irradiation. UV exposure rapidly converted Kaede-green to Kaede-red, then, Kaede-red was quickly diffused in the axon bundle to both anterograde and retrograde direction, resulting in the gradation of Kaede-green and Kaede-red in axon bundle. Scale bar: 150 µm. D The ratio of axon bundle-associated neurons and non-associated neurons from two independent samples. The average percentage of axon bundle-associated neurons was 32%, whereas that of non-associated neurons was 68%. E Relative fold change of gene expressions in axon bundle-associated neurons to non-associated neurons. TBR1 and VGLUT1 were highly expressed in axon bundle-associated neurons. n = 3 organoids. P = 0.0004 (TBR1 relative to GAPDH); 0.0168 (VGLUT1); 8.5 e–10 (DLX5); 0.0117 (SATB1). *p < 0.05, Student’s t test (two-sided). Data are presented as mean values ± SD.

Fig. 4. Optogenetics inhibition and synchronicity of burst activity between organoids.

A Optogenetic setup for inhibiting the synaptic interaction between left and right organoids through axon bundle. (i) A microfluidic device for optogenetic control consisting of fiber guides (thin channels) for optical fiber insertion. (ii) ArchT was expressed in the connected organoid by AAV. The light at 565 nm waveband illuminated the axon bundle in microfluidic device. (iii) An optical fiber was connected to 565 nm LED and a pulse generator (Arduino). An optical fiber was positioned perpendicular to the axon bundle with a 100 μm gap. Light exposure timing and a representative channel from MEA amplifier were recorded in TTL logger to synchronize TTL signal and recording. B An axon bundle and an optical fiber. Curved structure serves as a PDMS lens which helps the light to be focused on the axon bundle. C LFP and raster plot from left (pink) and right (blue) organoids of a connected organoid with or without light illumination (orange bar: TTL). D Synchronized burst frequency was around 0.65 Hz in the absence of light illumination and decreased to zero during light on. Then, synchronized burst frequency immediately increased and recovered after stopping the light. E Wavelet coherence by wavelet transformation revealed that slow wave oscillation disappeared during light illumination, indicating the vanishment of any correlative activity during light exposure. F Light illumination completely suppressed these synchronized burst activities. n = 8 from 3 organoids. P = 1.5e−8. G Inter-regional synchronicity measured by delta phase-delta phase coupling showed significant decrease during light illumination. n = 5 from 3 organoids. P = 0.0008 (first light off/light on); 0.0001 (second light off/light on). H Total numbers of single spikes were calculated in a 5 min time slot. Light-on significantly induced the increase in number of spikes. *p < 0.05; Paired-test (two-sided) for F and one-way ANOVA with Tukey’s multiple comparison for G. Data are presented as mean values ± SD.

Fig. 5. Theta-band oscillation, phase-amplitude coupling, and neuronal avalanches indicate complex activity in the connected organoids.

A Raw LFP plot from four electrodes of each of the connected organoids at 9 weeks. B Wavelet coherence between the two organoids exhibited synchronous activity in the theta band frequency. C Modulation index of PAC in delta-phase/gamma-power and theta-phase/gamma-power of the connected organoids cultured for 5, 7, and 9 weeks. n = 3 organoids. P = 0.0004 (delta, 5–9 weeks); 0.0016 (delta, 7–9 weeks); 0.0052 (theta, 5–9 weeks). D Delta-phase/gamma-power and theta-phase/gamma-power PAC modulation index of single, fused, and connected organoids. n = 3 organoids. P = 8e−7 (delta, single/connected); 9e−7 (fused/connected); 7.35e−5 (theta, single/connected); 0.0002 (fused/connected). E Intra- or inter-organoid PAC modulation index in the connected organoids. n = 16 from 4 organoids. F Schematic illustration of analyses of neuronal avalanches including the extraction of neuronal avalanche cascades. Neuronal avalanches were calculated from 8 electrodes. The cascade of single spikes was characterized at 3 msec scale size. G The log plot of neuronal avalanche size and probability at 5, 5.5, and 8.5 weeks of culture. *p < 0.05; One-way ANOVA with Tukey’s multiple comparison test. Data are presented as mean values ± SD.

Fig. 6. Potentiation of the connected organoids by optogenetic stimulation.

A Optogenetic setup for stimulation of the axon bundle. B Optogenetic stimulation drove synchronized burst activity. C Burst frequency was modulated by optical stimulation. The burst frequency followed the stimulation frequency after a significant delay. D Log plot of neuronal avalanche size and probability before, during, and after stimulation. E Time course of burst frequency with 1 Hz stimulation for (i) 20 min or (ii) 5 min every hour. n = 3 organoids. F Duration until the burst frequency decreased to 75% of the maximum burst frequency after cessation of light stimulation. n = 3 organoids. G The delay from the start of light stimulation to the induction of burst frequency was significantly reduced during the second and third attempts when the connected organoids were stimulated for 20 min. n = 4 organoids. P = 2.1e−6, 2.3e−6 (2nd and 3rd relative to 1st, within 20 min stimulation). H Time series of burst frequency in the presence of K252a or anisomycin. n = 3 organoids. I, J Duration of sustained activity (I) (n = 3 organoids. P = 0.0123; 0.0096 (2nd and 3rd respectively, in comparison with 1st within control); 0.0435; 0.0019 (2nd and 3rd respectively, in comparison with 1st within K252a); 0.0153 (3rd, in comparison with 1st within anisomycin)) and the delay of burst frequency induction (J) of the connected organoids in the presence of K252a or anisomycin, as shown in F and G. n = 4 organoids. P = 2.1e−6, 2.3e−6 (control); 3.3e−5, 4.5e−5 (K252a); 2.4e−9, 2.7e−9 (anisomycin). K Probability slope of neuronal avalanches. K252a treatment, but not anisomycin treatment, led to a decreased probability of neuronal avalanches. n = 3 organoids. P = 2.8e−17 (control/K252a before); 3.9e−6 (control/K252a 2nd off); 6.2e−7 (K252a/anisomycin before); 3.7e−5 (K252a/anisomycin 2nd on); 2.1e−8 (K252a/anisomycin 2nd off); P = 0.0025 (control/K252a 1st off); 0.002 (control/K252a 2nd on); 0.0144 (control/K252a 3rd on); 0.0031 (control/K252a 3rd off); 0.0315 (K252a/anisomycin 3rd on); 0.0025 (K252a/anisomycin 3rd off). *p < 0.05; one-way ANOVA with Tukey’s multiple comparison test. Data are presented as mean values ± SD.

Fig. 7. Evoked bursts self-reinforce the diversity of burst activity in a CaM kinase-dependent manner.

A Representative image of sorted bursts evoked by optogenetic stimulation. A total of 891 burst traces are presented. Light stimulation induced evoked spikes, and the burst responses persisted after light stimulation. Secondary and tertiary responses were also observed. B Latency of the evoked bursts. Repeated light stimulation (at second and third attempts) significantly decreased evoked burst latency in control and anisomycin-treated conditions, whereas K252a treatment did not influence the latency. n = 3 organoids. P = 0.0179; 0.0033 (2nd and 3rd, in comparison to 1st within control). C (i) Overlayed power histograms of evoked bursts and kernel density estimation (line) in the presence of K252a and anisomycin. Repeated light stimulation increased burst response complexity. (ii) Percentage of evoked bursts with and without secondary peaks. D Violin plots of burst response peak times. Red line indicates median. n = 50 bursts. E Representative crosstalk between the two connected organoids in self-evoked bursts. Color-coded maps (bottom circles) depict the voltage distribution from neurons in left and right cerebral organoids. *p < 0.05; one-way ANOVA followed by Tukey’s multiple-comparison test. Data are presented as mean values ± SD.