Apparatus for observing neural activity in neural circuit tissue and method for screening substances using the same

Abstract

The present invention provides a neural circuit tissue that is induced in vitro and that has at least two nerve cell masses (e.g., organoids) connected via axons. The present invention also provides a device for observing the neural activity of this neural circuit tissue and a method for screening substances using this device. This neural circuit tissue carries out spontaneous firing activity and is characterized in that two nerve cell masses exhibit interrelated activity, and in particular can be used as a brain model.

Description

[Technical Field]

[0001] The present invention relates to an in vitro induced neural circuit tissue that exhibits complex activity and plasticity, an observation device therefor, and a screening method for substances that alter the neural activity of the neural circuit tissue using the observation device.

[0002] Many regions in the brain are connected, and by exchanging and influencing each other's activities, advanced functions are produced. Understanding the connections between regions is thought to shed light on how the brain functions, but the complexity of neural circuits in the brain is a major obstacle to functional analysis, so simplified in vitro neural circuit models that are easy to analyze are needed.

[0003] In recent years, there have been many attempts to understand brain development and diseases using in vitro generated cerebral organoids. Organoid generation technology allows for three-dimensional modeling of animal organs in vitro, and can be used to analyze the function of various organs in conditions closer to in vivo. To date, cerebral organoids including the cerebral cortex, thalamus, cerebellum, hippocampus, and choroid plexus have been reported (see Non-Patent Documents 1-6, etc.), providing a new platform for studying human brain regions in vitro. Although the structural and morphological characteristics of brain organoids reported to date resemble parts of the developing human brain, functional improvements are needed. Because they are believed to be the basis for higher-order functions, there is a particular need for brain organoid model tissues that have the ability to exchange activity between regions and possess the plasticity of neural circuits. In addition, obtaining organoids that exhibit complex activity patterns requires long-term culture (more than three months), and there is a need for a method to obtain organoids that exhibit complex activity in a short culture period.

[0004] Brain organoids connected by axonal bundles have been reported as an in vitro model of neural tissue (Non-Patent Document 7 and Patent Document 1). This "connected organoid" model is constructed by connecting two brain organoids via axonal bundles extending from each organoid. This model reflects the in vivo brain state in which brain regions are connected by axonal bundles, and is considered to be a highly useful model for brain research. On the other hand, a "fused organoid" technology has also been developed, in which two or more brain organoids are directly fused by placing them adjacent to each other (e.g., Non-Patent Document 8). However, it was unclear what activities and functions these in vitro neural circuit models exhibited. [Prior art documents] [Patent documents]

[0005] [Patent Document 1] WO2017 / 187696 publication [Non-patent literature]

[0006] [Non-Patent Document 1] Del Dosso et al., Neuron 107 1014-1028 2020. [Non-patent document 2] Xiang et al., Cell Stem Cell 21 383-398 e7 2017. [Non-patent document 3] Sakaguchi et al., Nat Commun. 6 8896 2015. [Non-patent document 4] Muguruma et al., Cell Rep. 10 537-550 2015. [Non-patent document 5] Xiang et al., Cell stem cell 24 487-497 e7 2019. [Non-patent document 6] Pellegrini et al., Science 369 eaaz5626 2020. [Non-Patent Document 7] Kirihara et al., iScience 14 301-311 2019. [Non-patent document 8] Birey et al., Nature 545 54-59 2017. Summary of the Invention [Problem to be solved by the invention]

[0007] In view of the above circumstances, an objective of the present invention is to obtain an in vitro induced neural circuit tissue that is capable of exchanging neural activity between regions and exhibits complex activity and plasticity. Another objective of the present invention is to provide an apparatus for observing neural activity in neural circuit tissue. Another objective of the present invention is to provide a method for screening substances that alter the neural activity of neural circuit tissue using neural circuit tissue induced in vitro and this observation device. [Means for solving the problem]

[0008] The present inventors cultured multiple brain organoids differentiated from human iPS cells on a multi-electrode array (MEA), created neural tissue in which the brain organoids were connected by axon bundles, and analyzed the neuronal activity occurring in the neural tissue. For the first time, they succeeded in detecting highly complex neuronal activity from organoids connected by axon bundles (hereinafter referred to as "connected (brain) organoids").

[0009] Neuronal activity in the connected brain organoids was detected as spontaneous firing after a relatively short period of culture (approximately 7-8 weeks after iPS cell culture). The connected organoids exhibited burst activity, with a coefficient of variation of burst activity frequency exceeding 0.2. Furthermore, the local field potential (LFP) detected from the neural tissue was characterized by delta (0.5 Hz-4 Hz) and / or theta (4.0 Hz-8.0 Hz) wavelength components.

[0010] Furthermore, optogenetic inhibition of neuronal activity propagating through the axon bundles of the connected organoids significantly suppressed neuronal activity within each organoid. These results suggest that axonal connections between organoids are crucial for generating vibrant and complex neuronal activity. Furthermore, light stimulation of the axon bundles of the connected organoids induced neuronal activity within the organoids in response to the stimulation. This neuronal activity persisted even after light stimulation was discontinued. A time lag was required for neuronal activity in response to the stimulation pattern. However, repeated light stimulation shortened this time lag, and the previous activity pattern was quickly recapitulated after the second stimulation. This suggests that connected organoids are capable of retaining short-term temporal memory.

[0011] The present invention was completed based on the above findings. That is, the features of the present invention are listed as (1) to (24) below. (1) A neural circuit tissue induced in vitro, in which two or more neural cell clusters are connected via axons (also referred to as "connected organoid"). The neural cell cluster may be an organoid. (2) The two or more neuronal masses exhibit correlated activities. The correlated activities may occur with a time difference of 500 milliseconds or less. The correlated activities may be synchronized activities. (3) Spontaneous firing activity occurs, and the spontaneous firing activity may be 50 or more times per minute. (4) Burst activity may be performed, and the coefficient of variation of the frequency of the burst activity may be 0.2 or more. (5) The activity of each neuronal cluster connected via the axons exhibits coherence, and the coherence may differ for each frequency band. (6) The local field potential detected from the neuronal mass contains a delta waveband (0.5 Hz-4.0 Hz) component and / or a gamma waveband (300 Hz-3000 Hz) component, and may further contain a theta waveband (4.0 Hz-8.0 Hz) component. (7) The neural cell mass may be induced to differentiate from pluripotent stem cells, and may be formed by culturing the pluripotent stem cells for 6 weeks or more. The pluripotent stem cells may be iPS cells (induced pluripotent stem cells). (7) The nerve cell mass is obtained by culturing the nerve cell mass for two weeks or more. (8) It exhibits plasticity in response to external stimuli. (9) An apparatus for observing the neural activity of neural circuit tissue induced in vitro in which two or more neuronal clusters are connected via axons, comprising: a substrate; a plurality of wells provided on the surface of the substrate for accommodating the neuronal clusters; and guide grooves provided by connecting the wells for guiding and extending the axons so as to interconnect the neuronal clusters; each of the wells is provided with an electrode; and the apparatus further comprises an analysis unit for analyzing the correlation of electrical signals from the electrodes. (10) The analyzing unit separates each of the electrical signals into frequency bands and analyzes the correlation by phase amplitude coupling. (11) The analysis unit performs analysis using wavelet coherence. (12) The analysis unit classifies the neural activity into one of an action potential, a burst activity, a neural avalanche, and a local field potential based on the pattern of the electrical signal. (13) The analysis unit has in advance a signal pattern of the electrical signal corresponding to the neural activity, and classifies the neural activity by comparing it with this. (14) The neural activity can be generated by externally applying a stimulus to the nerve cell mass and / or the axon, and the resulting signal pattern can be given in advance. (15) The stimulus may be one or more of light irradiation, electrical stimulation, and administration of a compound, and may enhance or suppress the neural activity. (16) The electrode comprises a plurality of electrodes arranged in an array on the bottom of the well. (17) The electrode receives the electrical signal from the nerve cell mass and electrically stimulates the nerve cell mass. (18) The substrate is transparent so that the neuronal cell mass in the well can be optically observed from the bottom. (19) A light irradiation device for optically stimulating the nerve cell cluster and / or the axon is provided facing the surface of the substrate. (20) The light irradiation device includes an irradiation unit that locally irradiates light onto the nerve cell cluster and / or a part of the axons. (21) An apparatus for observing the neural activity of in vitro induced neural circuit tissue in which two or more neuronal clusters are connected via axons, comprising: a substrate; a plurality of wells provided on the surface of the substrate for accommodating the neuronal clusters; and guide grooves provided connecting the wells for guiding and extending the axons so as to connect the neuronal clusters, wherein each of the wells is provided with an optical system for measuring light intensity from the neuronal clusters, and further comprising an analysis unit for acquiring the correlation of light intensity signals from each of the neuronal clusters in the wells. (22) Two wells may be provided in the substrate, and a first objective lens of the optical system may be provided close to one of the wells from one side of the substrate, and a second objective lens may be provided close to the other well from the other side of the substrate. (23) The increase in calcium ion concentration in the nerve cell mass caused by an action potential can be detected by detecting luminescence from a calcium fluorescent probe, wherein a light beam for causing the luminescence can be incident from a side edge surface of the substrate. (24) A method for screening a substance that fluctuates neural activity in neural circuit tissue using an observation device having the above-mentioned characteristics, characterized in that a target neural circuit tissue is set in the observation device, the substance is applied to the neural circuit tissue, and fluctuations in the neural activity are observed. In this specification, the symbol "to" indicates a numerical range including the values on either side of it. [Effects of the Invention]

[0012] The neural circuit tissue of the present invention mimics brain function in vitro, and by combining it with the observation device of the present invention, various useful information can be obtained by observing (analyzing) it. For example, using a neural circuit tissue consisting of a cerebral organoid, one of the neural circuit tissues of the present invention, and an observation device, it is possible to screen for substances that may alter neural activity in the brain. Such findings are useful as a brain disease model, including disorders of higher brain function, in the search for therapeutic agents for psychiatric disorders, neurodegenerative diseases, and other conditions. [Brief explanation of the drawings]

[0013] [Figure 1] Formation and characterization of connected organoids on PDMS-MEA chips. [Figure 2] Analysis of neuronal activity in brain organoids connected by axon bundles. [Figure 3] Genetic analysis of connected brain organoids (connected organoids) using single cell RNA sequencing. [Figure 4] Characterization of axon bundles using photoconvertible fluorescent proteins. [Figure 5] Optogenetic inhibition of burst activity between connected organoids and synchronization between the two organoids. [Figure 6] Complex activity generated in connected organoids analyzed using phase-amplitude coupling and hidden Markov models. [Figure 7] Analysis of short-term memory mechanisms in connected organoids. [Figure 8] Analysis of diverse burst activity patterns supported by CaMKII-dependent signals. [Figure 9] Visualization of elongated axons in connected organoids. [Figure 10] Comparison of neural activity in single, fused, and connected organoids. [Figure 11] Measurements of axonal conduction velocity in organoids connected by axon bundles of various lengths. [Figure 12] Analysis of neuronal activity in organoids in which axon bundles were physically severed. [Figure 13] The results of investigating changes in LFPs as the connected organoids mature and the effects of drug treatment on the neural activity of the connected organoids. [Figure 14] Scalograms of LFP signals and wavelet transforms of connected organoids in the presence of various drugs. [Figure 15] Simultaneous measurements of Ca2+ transients and electrical activity. DETAILED DESCRIPTION OF THE INVENTION

[0014] Hereinafter, an embodiment of the present invention will be described. A first embodiment of the present invention is a neural circuit tissue induced in vitro, in which two or more neural cell masses are connected via axons (neural circuit tissue according to this embodiment). In this embodiment, a "cell cluster" refers to a cell population in which cells adhere to each other to form a three-dimensional structure, similar to the state of existence in a living body. A neural cell cluster is a cell cluster composed of neural cells. An "organoid" is a complex cell cluster consisting of cells derived from organs such as the brain, stomach, liver, and bladder, or organ-specific cells. Organoids can be produced by the self-aggregation of pluripotent stem cells (see, for example, Non-Patent Documents 1 to 7). In this specification, "organoid" refers to a brain organoid, i.e., an organoid containing neural cells, unless otherwise specified.

[0015] The neural circuit tissue according to this embodiment has, as its smallest unit, a tissue in which neuronal cell clusters, such as two brain organoids, are connected by axonal bundles; three or more neuronal cell clusters may be connected to each other by axonal bundles. Here, "neural circuit tissue" refers to neural tissue in which neurons are connected to each other via axons, and is an organization in which the activity of neurons within one brain organoid, or of neurons as a group, can interfere with the activity of other brain organoids. This can be defined, for example, by the spontaneous firing activity of each brain organoid, or the synchrony or asynchronous nature of neural activity between two or more brain organoids, the strength of PAC, etc. These are tissues capable of burst activity and periodic oscillation activity as a result of the coordination of neural activity among multiple organoids. Connected organoids can generate continuous neural avalanches. Furthermore, connected cell clusters exhibit correlated responses to external stimuli (e.g., light irradiation) and are characterized by plasticity. The neural circuit tissue according to this embodiment can be produced, for example, by placing a neural cell mass such as an organoid in a small well connected by a guide (thin groove) for axonal extension, and culturing the organoid. Organoids can be produced by culturing pluripotent stem cells such as iPS (induced pluripotent stem cells) cells under appropriate culture conditions for approximately 4 to 6 weeks (see Non-Patent Document 7 or Examples for details of the method for producing neural tissue according to this embodiment).

[0016] Although the situation varies somewhat depending on the culture conditions, neural tissue prepared as described above will begin to exhibit "spontaneous firing activity" (a phenomenon in which short-duration spikes are generated even in the absence of external stimuli) after 1.5 or 2 weeks from the start of culturing the neural cell mass. In the neural circuit tissue according to this embodiment, the frequency of spontaneous firing activity is, for example, 50 or more times per minute. Here, "activity" refers to changes in nerve cells caused by action potentials generated in the nerve tissue. The action potential generally refers to the diffusion of sodium and potassium ions across the cell membrane via voltage-dependent ion channels due to the concentration difference between the inside and outside of the cell, which starts with a change in membrane potential generated in the cell membrane due to synaptic activity, etc. Although MEA cannot directly measure membrane potential, it can measure (extracellular recording) the change in the weak electrical signal caused by the action potential outside the cell near the neuron. Calcium imaging also allows the measurement of the change in the intracellular calcium ion concentration (Ca 2+ The increase in calcium levels can be observed by detecting the emission of calcium from fluorescent probes using a microscope. The changes caused by neural action potentials recorded using techniques such as these are called the "activity" of neurons and neural circuit tissues.

[0017] Furthermore, the neural circuit tissue according to this embodiment generates "burst activity," in which multiple action potentials fire together at high frequency, in addition to spontaneous firing activity. Burst activity can be caused by the continuous activity of a single neuron. It can also be caused by the continuous activity of multiple neurons. During burst activity, the frequency of observed activity is significantly increased compared to periods when burst activity is not occurring.

[0018] A characteristic of burst activity in neural tissue according to this embodiment is that the coefficient of variance of the frequency of burst activity is, for example, 0.2 or more. Here, the coefficient of variance is the value obtained by dividing the standard deviation of the burst frequency by the mean. A coefficient of variation of burst activity of 0.2 or more is an indicator of the "complexity" of the activity of neural circuit tissue.

[0019] After 1.5 or 2 weeks of neuronal culture, local field potentials (LFPs) can be detected from neuronal clusters. These LFPs contain components in the delta (0.5 Hz-4.0 Hz), gamma (300 Hz-3000 Hz), and / or theta (4.0 Hz-8.0 Hz) wavelength ranges.

[0020] The nerve cell masses in the neural circuit tissue according to this embodiment exhibit mutually related activities. "Mutually correlated activity" refers to neural activity in each connected neural cell mass when some correlation is observed between the neural activity or neuronal activity in each connected neural cell mass. Taking a neural circuit tissue in which organoid 1 and organoid 2 are connected as an example, if activity or burst activity occurs in organoid 1 and activity or burst activity also occurs in organoid 2 at approximately the same time, i.e., when the neural activity in organoid 1 and organoid 2 is synchronized, the neural circuit tissue can be said to have engaged in mutually correlated activity. Furthermore, in addition to synchronized activity, mutually correlated activity can also be observed when neural activity occurs in one organoid and the other generates the same neural activity with a certain time lag, or when the neural activity is repeated with a time lag or a certain phase difference. Furthermore, as for "mutually related activities," it is desirable that the activities of each neuronal mass of two neuronal masses occur simultaneously or within a short time interval. Although not particularly limited, for example, the activities of each neuronal mass (related activities) are activities that occur within 1,000 milliseconds, preferably within 500 milliseconds, and more preferably within 100 milliseconds. Another indicator of "complexity" is the interrelated activities within an organization that have time or phase differences.

[0021] In the neural circuit tissue according to this embodiment, the activities of each neuronal cell mass connected by axon bundles may interfere with each other (coherence), and this interference is one of the characteristics of the neural circuit tissue. This interference may differ depending on the frequency band of the activity pattern. This interference for each frequency band can be expressed as the degree of linear correlation at each frequency by performing wavelet coherence analysis based on the time series data of two activity waves detected from each neuronal cell mass. Wavelet coherence analysis is described in detail in the Examples section, so please refer to that section. The wavelet coherence exhibited by different parts of a neural circuit is also an indicator of "complexity."

[0022] Furthermore, the neural circuit tissue according to this embodiment is characterized by having plasticity in response to external stimuli (e.g., light irradiation, etc.). Here, "plasticity" refers to the property of responding to external stimuli and changing activity patterns after the stimuli. The neural circuit tissue is characterized by having plasticity.

[0023] The neural circuit tissue according to this embodiment may be provided with means for detecting neuronal activity detected from the neural circuit tissue. Here, the term "means" refers to, when detecting neuronal activity as an electrical signal, an electrode can be used as a detection means. The electrical signals of neuronal activity can be analyzed, for example, by placing neuronal cell clusters on a substrate with one or more electrodes and detecting electrical signals of activity waves from each neuronal cell cluster from the electrodes. To connect multiple neuronal cell clusters to each other through axonal bundles, neuronal cell clusters may be cultured on a substrate equipped with a multi-electrode array while simultaneously detecting electrical signals obtained from the neuronal cell clusters. Specifically, the neural circuit tissue of this embodiment may be cultured using a device in which small holes (small holes for containing neuronal cell clusters and culture medium) are arranged on each multi-electrode array, allowing the culture of neuronal cell clusters, and grooves are arranged between each small hole to serve as axonal extension guides (which also serve as a medium flow path) for extending from one neuronal cell cluster to another. The activity of the neural tissue of this embodiment can be manipulated (modulated) by irradiating a portion of the neural tissue (e.g., an axonal bundle) with light. When performing such an operation, the device may be provided with an element for applying an external stimulus such as light irradiation, for example, a groove (which also serves as a medium flow path) for arranging an optical fiber, etc. As an example of such a device, see Figures 1A and 1B described below.

[0024] Furthermore, as shown in Figure 15A (described later), an optical system (such as a microscope) for detecting luminescence emitted from each neuronal cell cluster can be used as a "means" for detecting neuronal activity using calcium imaging. The optical system uses a CCD camera or the like to measure light intensity, and analyzes the light intensity signals from each neuronal cell cluster, including the presence or absence of correlation. Here, the first objective lens (Objective 1) of the optical system is brought close to one of the neuronal cell clusters from one side of the substrate, and the second objective lens (Objective 2) is brought close to the other of the neuronal cell clusters from the other side of the substrate, and the light intensity is measured. In this case, the calcium ion concentration (Ca 2+ To detect changes in calcium levels, it is recommended to preliminarily introduce a fluorescent calcium probe into the cells. Examples of proteinaceous fluorescent calcium probes include cameleon, GCaMP (G-CaMP2, G-CaMP4, G-CaMP6, G-CaMP7, and G-CaMP8, etc.), and R-CaMP (R-CaMP1.07, R-CaMP2, etc.). Although not shown in Figure 15, a light beam for fluorescence emission may be incident from the side edge of a transparent substrate.

[0025] The second embodiment is a method for analyzing neural activity in vitro, which includes detecting activity for each frequency band obtained from a neural cell mass that constitutes the neural circuit tissue of the first embodiment, and analyzing the presence or absence of correlation and the detected activity correlation using a method for analyzing the correlation of neural activity, such as phase-amplitude-coupling (PAC) or wavelet coherence. Phase-amplitude coupling is a well-established method for assessing the relationship between the phase of low-frequency activity and the amplitude of high-frequency spikes (Fell et al., Nat. Rev. Neurosci. 12 105-118 2011; Canolty et al., Trends Cogn. Sci. 14 506-515 2010). For example, delta (phase)-gamma (amplitude) PAC can be used to analyze whether the amplitude strength of frequencies in the gamma wave region influences the phase in the delta wave region. By the method according to the second embodiment, for example, an increase in the correlation ratio (Modulation index) of delta (phase)-gamma (amplitude) PAC or theta (phase)-gamma (amplitude) PAC that accompanies the cultivation of cerebral organoids can be explained from the viewpoint of neural activity as indicating the maturation of the cerebral organoids. It is preferable that the correlation rate between activity in different frequency bands by PAC is 0.05 or higher, which is also an index of the "complexity" of neural circuit organization.

[0026] The third embodiment is a method for modulating the neural activity of the neural circuit tissue according to this embodiment, which includes recording the neural activity of the neural circuit tissue and stimulating the neural circuit tissue in an arbitrary pattern. Here, "neural activity patterns" refer to the spatiotemporal characteristics or correlations of neural circuit tissue or neuronal activity. Examples of neural activity patterns include the frequency of "burst activity" (or "burst-like activity"). "Burst activity" is characterized by a temporal concentration of high-frequency neural activity over a short, fixed period (e.g., 100 ms) (i.e., neural activity is observed at a significantly higher frequency than during non-burst periods). Neural activity patterns can be characterized not only by a constant frequency, but also by inconsistencies (i.e., a small or large coefficient of variation in the intervals between bursts). Similar to burst activity, neural activity patterns can also include neural activity in specific frequency bands and their correlations. In addition to temporal characteristics and correlations, neural activity patterns can also include correlations between electrodes (or between observations at spatially distinct sites or positions). Therefore, neural avalanches are also considered a neural activity pattern. Furthermore, spatiotemporal correlations of neural activity between connected neural tissues can also constitute a neural activity pattern. Furthermore, "stimulating the neural circuit tissue in an arbitrary pattern" means applying stimulation at an arbitrary frequency or location determined in space and time to induce neural activity. For example, this means applying stimulation at an arbitrary frequency or to an arbitrary location. It is also possible to stimulate in a random space and time pattern. The pattern can be determined in advance as desired, or it can be calculated based on observed neural activity patterns. Furthermore, the term "stimulation" is not particularly limited, but may refer to, for example, stimulation provided by light, electricity, administration of a compound, etc. The neural tissue circuit according to this embodiment has a characteristic in that its activity is enhanced or suppressed by these stimuli.

[0027] The fourth embodiment is a screening method for substances that may alter neural activity in the brain, and includes contacting a candidate substance with a neural circuit tissue according to this embodiment and detecting the neural activity of the neural tissue. The neural circuit tissue according to this embodiment exhibits spontaneous firing activity and exhibits an activity pattern similar to neural activity in the brain in vivo. Furthermore, when the neural tissue is repeatedly stimulated externally (e.g., by irradiating the axon bundles with light), the response time to the stimulation shortens with each stimulation. It has also been shown that when the axon bundles of the neural tissue are stimulated with light for a certain period of time (e.g., approximately 20 minutes), the neural tissue exhibits burst-like activity. However, this burst-like activity does not immediately disappear after the stimulation is stopped, but rather maintains an elevated activity frequency for some time. Therefore, the neural circuit tissue according to this embodiment can be interpreted as having the function of storing (memorizing) external stimulus patterns within the tissue. Furthermore, the activity supporting memory in the neural tissue is suppressed by a CaM kinase II inhibitor, suggesting that at least the neural circuit tissue according to the first embodiment exhibits activity similar to the neural activity occurring in the brain during short-term memory.

[0028] As described above, the neural circuit tissue according to this embodiment can be used as a brain model that mimics brain functions. Therefore, for example, by contacting the neural circuit tissue with a desired substance (e.g., an NMDA inhibitor, an AMPA inhibitor, or a psychotropic drug) and detecting its effect on excitatory synaptic transmission or inhibitory synaptic transmission, or by detecting its effect on phase-amplitude coupling (PAC), which is characteristic of certain diseases, it is possible to screen for candidate substances that alter neural activity in the brain. Furthermore, it is also possible to screen for substances (such as K252a and anisomycin) that may alter short-term or long-term memory, particularly substances that may enhance short-term memory. For example, if the duration of neural activity in response to light stimulation of the axon bundles of the neural circuit tissue according to this embodiment (duration of neural activity after light irradiation) is longer in the presence of a certain substance than in the absence of the substance, it can be determined that the substance may effectively function to maintain short-term memory.

[0029] Where this specification is translated into English and includes the singular words "a," "an," and "the," these shall include the plural as well as the singular, unless the context clearly indicates otherwise. The present invention will be further explained below by showing examples, but these examples are merely illustrative of embodiments of the present invention and do not limit the scope of the present invention. [Example]

[0030] 1. Materials and Methods 1-1. Fabrication of PDMS-MEA chip SU-8 master positive pattern models were fabricated using standard photolithography techniques (Bowen et al., Frontiers in Systems Neuroscience 13, doi: 10.3389 / fnsys.2019.00045. eCollection 2019). SU-8 (2100 or 2075) was poured onto a silicon wafer (4 inches) and centrifuged (1200-1500 rpm for 30 seconds) to coat the surface. The wafer was pre-treated by heating it on a hot plate at 65°C for 90 minutes, and then heat-treated at 95°C for 40 minutes. Then, using a photomask, the wafer was exposed to UV (365 nm, 2.5-3.0 mWcm). 2 ) for 60-75 seconds. The wafer was heated on a hot plate at 65°C for 7 minutes, followed by a thermal treatment at 95°C for 13 minutes. After cooling, the SU-8 was developed in SU-8 developer for 15 minutes and washed three times with isopropyl alcohol. The wafer was then heated in an oven at 150°C for 3 minutes. The thickness of the SU-8 was approximately 150 μm. The microfluidic device was fabricated using a polydimethylsiloxane (PDMS) silicone elastomer kit (Sylgard 184, Dow Corning). The silicone elastomer and curing agent were mixed in a 10:1 weight ratio, degassed, poured onto the patterned SU-8 structure, and cured in an oven at 80°C for 6 hours. Holes for the organoids and reference electrode were created using a biopsy punch (1.5 mm and 2 mm, respectively). A glass ring (inner diameter: 22 mm, outer diameter: 25 mm) used to store the medium was glued to the PMDS device. The fabricated PMDS device was sterilized by autoclave, followed by treatment with 70% ethanol and UV light.

[0031] 1-2. Human iPS cells Human iPS cells were obtained from the Riken Cell Bank (409B2, HPS0076) (Okita et al., Nat Methods 8 409-412 2011). Cells were maintained on ESC-qualified Matrigel-coated 6-well plates in mTeSR plus medium (STEMCELL Technologies) supplemented with 10 μM Y-23632 (first 24 hours only, Wako) for the first day. Thereafter, cells were seeded and cultured every 5–7 days using ReLeSR (STEMCELL Technologies).

[0032] 1-3.Creation of brain organoids To generate brain organoids, iPSCs were first dissociated into single cells using TrypLE Express. 20,000 cells were then plated into a round-bottom, low-attachment 96-well plate (Prime surface, Sumitomo bakelite) in mTeSR medium containing 10 μM Y-23632. After 24 hours, the medium was replaced with neural induction medium (DMEM-F12, 15% (v / v) knockout serum replacement, 1% (v / v) MEM-NEAA, 1% (v / v) Glutamax, 100 nM LDN-193189, and 10 mM SB431542), and then replaced every 2 days. On the 10th day of culture, the medium was replaced with a 1:1 mixture of DMEM / F12 and Neurobasal medium supplemented with 0.5% (v / v) N2 supplement, 1% (v / v) B27 supplement (vitamin A-free), 1% (v / v) Glutamax, 0.5% (v / v) MEM-NEAA, 0.25 mg / ml (v / v) human insulin solution, and 1% (v / v) penicillin / streptomycin. The medium was then replaced every two days, and the cells were cultured until the 18th day. On the 18th day of culture, the medium was replaced with a maintenance medium (0.5% (v / v) N2 supplement, 1% (v / v) B27 supplement (vitamin A-free), 1% (v / v) Glutamax, 0.5% (v / v) MEM-NEAA, 0.25 mg / ml (v / v) human insulin solution, 20 ng / ml BDNF, 200 mM ascorbic acid). The medium was then replaced with Neurobasal medium (supplemented with 1% (v / v) penicillin / streptomycin and 1% (v / v) acid). The cerebral organoids were then cultured for 4 weeks and used to generate connected organoids.

[0033] 1-4. Formation of connected organoids in PDMS-MEA Two brain organoids were cultured in a PDMS-MEA and connected by axonal bundles. The culture for connecting the brain organoids with each other by axonal bundles was performed by a modified method disclosed in a previous report (Non-Patent Document 7). The electrodes of a multi-electrode array (MEA) were aligned with the PDMS microchannels (the culture wells for brain organoids) and connected. The microchannels were coated with ESC-qualified Matrigel (Corning) in DMEM / F12 (1:30) for 1 hour at room temperature. The coating solution was replaced with maintenance medium. The brain organoids were then placed in the microchannels and allowed to sink to the bottom by gravity. The maintenance medium was replaced every two days.

[0034] 1-5. Multi-microelectrode array (MEA) Twenty-four hours before multielectrode neuronal activity measurements, the maintenance medium was replaced with Brainphys supplemented with 1% (v / v) B27 supplement (containing vitamin A), 1% (v / v) Glutamax, 20 ng / ml BDNF, and 1% (v / v) penicillin / streptomycin. The PDMS-MEA was placed in an MED64 system (Alpha MED Scientific) and electrical signals from all 64 electrodes were recorded at a sampling rate of 20,000 Hz for 5–30 min at 37°C. Noise during electrical signal recording was removed using a band-pass filter between 0.1–10,000 Hz. The raw signals were further band-pass filtered (300–3,000 Hz) for spike analysis, raster plotting, and spike clustering, or low-frequency filtered (<1,000 Hz) for analysis of local field potentials. All post-hoc analyses were then performed using the Signal Processing Toolbox, Curve fitting Toolbox, Deep learning Toolbox, Parallel Computing Toolbox, and Wavelet Toolbox in MATLAB®. All analyses and calculations were performed using MATLAB software. All scripts for calculations in this example were downloaded from https: / / github.com / TatsuyaOsaki / Matlab_function.

[0035] 1-6. Wavelet Coherence and Wavelet Transform for Frequency Separation Wavelet coherence is a measure of the correlation between two signals at a specific frequency. Wavelet coherence from LFP recordings was calculated using equation (1). f(t) was calculated using the functions cwt() and icwt() in the "Wavelet Toolbox":

number

number

[0036] 1-7. Cross-correlation The cross-correlation R was calculated using the xcorr() function in MATLAB. The two signal series, xn and yn, were calculated as shown in the following equation:

number

number

[0037] 1-8.Neural avalanche Neuronal avalanches are events characterized by a continuous pattern of neuronal activity within the nervous system. The time bin (Δt) was set to 3 ms. The probability was calculated using the following formula:

number

[0038] 1-9. Optogenetic control of connected organoids To control the activity of connected organoids, we used optogenetic tools (see Figure 10, below). AAV-CAG-hChR2H134R-tdTomato was kindly provided by Karel Svoboda (Addgene plasmid # 28017). pAAV-CAG-ArchT-GFP was kindly provided by Edward Boyden (Addgene plasmid # 29777). 5 μL of AAV viral vector was mixed with 500 μL of maintenance medium, which was then replaced with the medium in the PDMS-MEA chip 72 hours before MEA measurements. 2H134R The 470 nm fiber-coupled LED (M470F3 - 470 nm, 17.2 mW (min) Fiber-Coupled LED, 1000 mA, Thorlabs) for the Arch-T and the 565 nm fiber-coupled LED (M565F3, 565 nm, 9.9 mW (min) Fiber-Coupled LED, 700 mA, Thorlabs) for the Arch-T were controlled by a T-cube high-power LED driver (LEDD1B, 1.2A, Thorlabs). Light was delivered through a multimode fiber (0.22 NA, High-OH, φ105 μm core, 250-1200 nm, Thorlabs). TTL pulses were generated by an Arduino and controlled by the LED driver. The source code was downloaded from https: / / github.com / TatsuyaOsaki / Arduino_optogenetics.

[0039] 1-10.scRNA sequencing and data processing Single, fused, and connected organoids cultured on PDMS-MEA chips for 7 weeks were extracted from the PDMS device and centrifuged at 100 x g for 30 seconds. To obtain a single-cell suspension, organoids were dissociated in an AccuMax at 37°C for 10–30 minutes, followed by centrifugation at 200 x g for 5 minutes. Pelleted cells were then resuspended in DMEM containing 10% FBS and subjected to 10x Genomics Chromium single-cell RNA-seq library preparation according to the manufacturer's protocol. Finally, the library was sequenced on DNBSEQ with 150-bp paired-end reads. Sequencing data were processed using the Cell Ranger analysis pipeline v3 with default parameters. Reads were aligned to the human reference genome (GRCh38). The Cell Ranger output, a "filtered gene-barcoded" count matrix, was loaded into Scanpy (Wolf et al., 2018) and then imported into other Python packages (scanpy==1.8.1, anndata==0.7.6, umap==0.5.1, numpy==1.19.5, scipy==1.4.1, pandas==1.1.5, scikit-learn==0.22.2.post1, statsmodels==0.10.2, python-igraph==0.9.6, pynndescent==0.5.4) for downstream analysis. Poor-quality cells were excluded using the following criteria: min_genes > 200, min_cells < 3, mitochondrial gene percentage < 10%, nFeatures < 4000. Cells with a hemoglobin read percentage of 5% or more were also excluded. A total of 17,636 cells were analyzed. Principal component analysis (PCA) was performed using Scanpy to reduce the dimensionality of the data. UMAP was used to visualize the clustered data.

[0040] 1-11. Identification of neurons involved in axon fasciculation using the fluorescent protein Kaede To identify axon bundle-associated neurons in connected organoids, we transfected the photoconvertible fluorescent dye Kaede into connected organoids. First, we constructed the AAV-CAG-Kaede plasmid using AAV-CAG-EGFP (Addgene plasmid #28014) (Mao et al., Neuron 72, 111-123. 10.1016 / j.neuron.2011.07.029. 2011) and CoralHue Kaede (pKaede-S1) (see Figure 4A). AAV-CAG-GFP was kindly provided by Karel Svoboda (Addgene plasmid #28014). AAV was produced using AAVpro 293T (Takara). AAV was purified using the AAVpro Purification kit midi (Takara) according to the manufacturer's protocol. Connected organoids were infected with AAV at weeks 4–6 and phototransduction experiments were performed at week 7 in culture. To visualize axon bundle-associated neurons in connected organoids, a 405 nm laser was irradiated onto the axon bundle region and observed under a Nikon confocal microscope (1.5 mm × 0.5 mm × 0.2 mm, 5x magnification, total time: 60 min, Nikon A1R). The organoids were then dissociated using an AccuMax at 37°C for 10–30 minutes and centrifuged at 200 x g for 5 minutes. For flow cytometry, single cells were resuspended in PBS containing 1% BSA. Using a BD FACS Melody, the Kaede red positive / Kaede green negative and Kaede red negative / Kaede green positive populations were collected as axon bundle-associated and non-axon bundle-associated neurons, respectively. After sorting, total RNA was harvested for RT-PCR analysis.

[0041] 1-12. Knock-in of GFP and mCherry fluorescent proteins using CRISPR-Cas9 by electroporation To visualize axon outgrowth, human iPS cells were transfected with GFP or mCherry fluorescent proteins. The GFP and mCherry fluorescent proteins were inserted into the safe region of the AAVS1 (Adeno-associated virus integration site 1) gene, respectively. iPS cells were harvested by TrypL Eexpress treatment and centrifuged. Then, 5 μg of PX458-AAVS1 plasmid and 5 μg of AAVS1-Pur-CAG-EGFP plasmid (Varley et al., PLOS ONE 15, e0223812 2020) or 5 μg of AAVS1-Pur-CAG-mCherry plasmid (Varley et al., PLOS ONE 15, e0223812 2020) were mixed in 100 μL of Opti-MEM at a concentration of 1 × 10 6The plasmid-cell mixture was transferred to a NEPA cuvette (EC-002S) using a pipette, and electric pulses (Poring pulse: 125 V, pulse length: 5 msec, pulse length: 50 msec; number of pulses: 2, decay rate: 10%) and Transfer pulse: 20 V, pulse length: 50 msec, pulse length: 50 msec; number of pulses: 2, decay rate: 40%) were generated using a NEPA21 electroporator (NEPA gene). Electroporated iPS cells were seeded into four wells of a Matrigel-coated 6-well plate containing mTeSR plus (containing 10 μM Y-23632). After 24 hours, transfected cells were selected by adding 0.75 μg / ml puromycin and treating for 2 hours. The cells were then subcultured and expanded in two Matrigel-coated wells of the 6-well plate. PX458-AAVS1, AAVS1-Pur-CAG-EGFP, and AAVS1-Pur-CAG-mCherry plasmids were kindly provided by Dr. Adam Karpf and Dr. Su-Chun Zhang (Addgene 113194, 80945, and 80946).

[0042] 1-13. Frozen Sections and Immunohistochemistry Brain organoids were fixed with 4% paraformaldehyde (PFA) and 8% sucrose for 15 minutes at 4°C, washed three times with PBS (each wash was incubated for 10 minutes at room temperature), transferred to a 30% sucrose solution, and incubated overnight at 4°C. The sucrose solution was then removed, and the brain organoids were equilibrated with OCT compound for 15 minutes at room temperature. The brain organoids were then embedded in OCT compound on dry ice. The brain organoids were then stored at -80°C or cryosectioned at 20 μm thickness.

[0043] Cells were fixed with 4% paraformaldehyde for 20 minutes and then treated with 0.2% Triton X-100 for 5 minutes to permeabilize the cell membrane. Blocking was performed with 1% bovine serum albumin (BSA) for 2 hours. Cells were then treated with primary antibodies for 2 hours at room temperature. Then, cells were treated with secondary antibodies for 2 hours at room temperature. The primary antibodies used were mouse anti-neuron-specific βIII tubulin (Biolegend 801202, 1:1200), rabbit anti-neuron-specific βIII tubulin (Sigma, ZooMAb, 1:200), mouse anti-human PAX6 (DHSB, 1:100), rabbit anti-human GAD67 (Santa Cruz, 1:100), rabbit anti-human vGluT1 (Sigma, ZooMAb, 1:200), mouse anti-human CTiP2 (Abcam, 1:100), rabbit anti-human SATB2 (Abcam ab51502, 1:100), or rabbit anti-human MAP2 (Sigma, ZBR2290, 1:200). The secondary antibodies used were Alexa Fluor 555 anti-rabbit IgG (H+L), Alexa Fluor 405 anti-rabbit IgG (H+L), Alexa Fluor 488 goat anti-mouse IgG (H+L), Alexa Fluor 488 goat anti-rabbit IgG (H+L), and hAlexa Fluor 647 goat anti-rat IgG (H+L). Nuclei were stained with Hoechst dye for 20 minutes at room temperature, and Ca 2+ and Mg 2+ Dulbecco's Phosphate-Buffered Saline (D-PBS) containing ++ ) and rinsed three times. All cells and samples were observed under a fluorescence microscope (Axio Observer, Zeiss) or a confocal laser scanning microscope (Zeiss).

[0044] 1-14.RT-PCR (Real-time reverse-transcription) To measure the biological activity of brain organoids, total RNA was isolated from tissues using TriPure (Sigma). Reverse transcription was performed using KOD One (Toyobo). Primer sequences are listed in Table 1. RT-PCR was performed at CFX Connect using KAPA SYBR FAST qPCR Master Mix (KAPA Biosystems). In all experiments, the mRNA expression level of glyceraldehyde 3-phosphate dehydrogenase (GAPDH) was used as an internal standard. RT-PCR was performed at least three times using cDNA prepared from different tissues. [Table 1]

[0045] 1-15. Calcium (Ca 2+ ) Imaging To visualize neuronal activity by fluorescence, Ca 2+ The indicator (GCaMP6f driven by the CAG promoter) was transfected with AAV1. pAAV.CAG.GCaMP6f.WPRE.SV4 was a kind gift from Douglas Kim & GENIE Project (Addgene #100836). 5 μL of AAV viral vector was mixed with 500 μL of maintenance medium, which was then replaced with the medium in the PDMS-MEA 3 days before measurement. After 6–12 hours of AAV incubation, the medium was replaced with fresh maintenance medium. Thirty minutes before measurement, the maintenance medium was replaced with Brainphys supplemented with 1% (v / v) B27 supplement (containing vitamin A), 1% (v / v) Glutamax, 20 ng / ml BDNF, and 1% (v / v) penicillin / streptomycin. The device was then placed on the microscope stage (the microscope configuration is shown in Figure 12). Time-lapse images were captured at a frame rate of 20 fps or higher for 10 minutes. Data analysis was performed using MATLAB (MathWorks). Regions of interest (ROIs) were manually drawn around the cell bodies of neurons in the organoids. The baseline fluorescence value for each ROI was calculated as the average of the lowest baseline values for the sample. ΔF / F was calculated as (F - F) / F × 100, where F is the instantaneous fluorescence intensity from the time-series images of the untreated ROI.

[0046] 2.Results First, the results of this embodiment shown in the drawings will be explained.

[0047] Figure 1 shows the formation and characterization of connected organoids on a PDMS-MEA chip. (A) Schematic diagram of connected organoids on a PDMS-MEA chip. Two brain organoids were cultured in two chambers bridged by a microchannel on the chip. (B) An example of a PDMS-MEA chip is shown. The PDMS-MEA chip consists of an MEA probe (MEA), a PDMS (PDMS microfloral layer), a glass reservoir ring, and a PDMS lid (i and ii). Sixteen electrodes (4 x 4 array) made of metal thin films are installed under each brain organoid (iii). (C) Representative internal structures after 4 weeks (top) and 8 weeks (bottom). Scale bar: 150 μm. (D) Gene expression profiles of brain organoids cultured for 2 to 10 weeks. The vertical axis indicates gene names, and the horizontal axis indicates time (weeks) after culture. (E) Axons extended from one organoid to another within 5 weeks, and axon bundles formed by 6 weeks. From top to bottom, fluorescent images of brain organoids are shown after 4, 5, and 6 weeks (axon extension and axon bundle connection) of iPS cell culture. (F) Changes in axon bundle thickness over time. The vertical axis represents axon bundle thickness, and the horizontal axis represents the time in culture from iPS cells (weeks) (n=6). (G) The ratio of excitatory neurons (positive for VGLUT1 antibody staining) to inhibitory neurons (positive for GAD67 antibody staining) in organoids. The vertical axis represents cell density, and the horizontal axis represents the time in culture from iPS cells (weeks) (n=3). (H) Immunohistological analysis of organoids revealed the presence of a layered structure within the connected organoids after 8 weeks of iPS cell culture. Immunostaining for PAX6 and CTIP2 indicates the proliferative layer and cortical sublayer, respectively. I. Schematic diagram of the method for recording neuronal activity from connected organoids on a PDMS-MEA chip. Raw analog signals from the electrodes were amplified and converted to digital signals (16 bits) at a sampling rate of 20 kHz.The signals were then filtered with a 300-3,000 Hz bandpass filter for spike analysis and a 1,000 Hz low-frequency bandpass filter for local field potentials (LFPs). J shows a representative example of a connected organoid at 5 weeks after iPS cell culture. The scale bar is 1 mm (i). Also shown are examples of filtered signals from four representative electrodes under each organoid. LE is the signal from the left brain organoid, and RE is the signal from the right brain organoid (ii). Wavelet coherence between signals from the left and right organoids is shown (iii). K shows a representative example of a connected connectoid at 5.5 weeks after iPS cell culture. The scale bar is 1 mm (i). Dense spikes and synchronized burst-like activity were detected from multiple electrodes in the left and right organoids (ii). Wavelet coherence indicated a strong correlation between the two connected connectoids (iii). L shows the results of analyzing the synchronization of activity between organoids. The synchronization of neuronal activity between the two organoids increased throughout the culture period. The vertical axis represents the synchrony index, and the horizontal axis represents the iPS cell culture time (weeks). M shows the results of measuring the frequency of burst activity. The frequency of burst-like activity significantly increased depending on the culture period. The vertical axis represents the burst frequency, and the horizontal axis represents the iPS cell culture time (weeks). N shows the neuronal activity signals of two connected organoids (middle panel) and an enlarged view (bottom panel). The upper panel shows a raster plot. A time lag was observed between the synchronized burst-like activity of the left and right organoids. O shows the time lag between bursts of the left and right organoids at different culture time points. The vertical axis shows the burst delay, and the horizontal axis shows the iPS cell culture time (weeks). n = 20. *p<0.05, **p<0.01; one-way ANOVA. Error bars indicate SD (standard deviation).

[0048] Figure 2 shows the results of analyzing neuronal activity in axon-connected brain organoids. (A) Local field potential (LFP) signals in the 0.2-0.5 Hz, 0.5-4 Hz (δ), and 30-300 Hz (γ) bands were extracted using inverse continuous wavelet transformation. At 8 weeks after iPS cell culture, connected organoids generated low-frequency oscillations in the 0.5-4 Hz (δ) band. (B) Integrated power results for each frequency band. The vertical axis represents the integrated wave power, and the horizontal axis represents the time (weeks) after culture. (C) Representative examples of neuronal activity signals from connected, single, and fused organoids. (D) Burst frequency for each organoid type. (Single organoid, fused organoid, and connected organoid represent the results for single, fused, and connected organoids, respectively (n=10). E shows the results of inverse continuous wavelet transform in the 0.2-0.5 Hz, 0.5-4 Hz (δ), and 30-300 Hz (γ) bands. δ-band oscillations were detected in connected organoids but not in single or fused organoids. In the right panel, for each wavelength band, the left bar shows the results for single organoids, the middle bar shows the results for fused organoids, and the right bar shows the results for connected organoids. F shows the coefficient of variation of burst frequency for the three types of organoids. The vertical axis shows the coefficient of variation of the burst interval, and the horizontal axis shows the time after culture (weeks). G shows the relationship between organoid volume (horizontal axis) and average burst frequency (vertical axis).

[0049] Figure 3 shows the results of genetic analysis of connected brain organoids (connected organoids) using single-cell RNA-seq. A shows a UMAP plot of 17,636 samples. Leiden clustering was performed on single, fused, and connected organoids. (i) Cell populations were classified into 14 clusters. (ii) UMAP and density plots are shown for each sample. B and D show heatmaps of the top 30 statistically significant genes in each cluster. Based on known markers, the 14 clusters were further divided into four groups: Group 1 "NPC," Group 2 "Intermediate," Group 3 "Neurons," and Group 4 "Other." C shows the normalized ratios of cells in clusters 8, 6, 12, and 4 for single, fused, and connected organoids. The number of cells in cluster 8 of connected organoids was significantly higher than in single and fused organoids, while the number of cells in clusters 6 and 12 (NPC populations) of connected organoids was slightly lower than in single organoids. E shows an annotated UMAP plot of known marker genes. F shows the classification results for cell types using known marker genes. G shows a UMAP plot of vGlut1 for visualizing excitatory neurons and DLX6 and GAL for visualizing inhibitory neurons. H shows a UMAP plot of GRIA1 and GRIA2, which are highly expressed in cluster 8. I and J show plots of GRIA2 expression in cluster 8 of single organoids, fused organoids, and connected organoids. The average expression level of GRIA2 in connected brain organoids was higher than that in single and fused organoids. GRIA2 was particularly highly expressed in DCX-positive cells in cluster 8 (J).

[0050] Figure 4 shows the results of characterizing axon bundles using photoconvertible fluorescent proteins. (A) shows the plasmid map of the pAAV backbone plasmid expressing Kaede, a protein whose fluorescence wavelength is tunable by UV irradiation under the CAG promoter. Kaede green fluorescent protein can be converted to Kaede red fluorescent protein by UV irradiation. Cerebral organoids were infected with AAV-CAG-Kaede one week after being introduced into the microfluidic device. On day 49 (week 7 of culture), the axon bundles were irradiated with UV light (405 nm laser, confocal microscope mounted). Cells were then separated using a cell sorter, and neurons involved in axon bundle formation (Kaede red positive) and neurons not involved in axon bundle formation (Kaede green positive) were identified (see B). (C) shows confocal microscope images of connected organoids before and after UV irradiation. UV irradiation rapidly converts Kaede green to Kaede red. Kaede-red then rapidly diffused and distributed in the anterior-posterior direction within the axon bundle, forming a gradient of Kaede-green and Kaede-red within the axon bundle. (D) shows a cross-section of the axon bundle and the XY cross-sections at the bottom and top. In the center of the axon bundle, most of the Kaede-green was converted to Kaede-red. Furthermore, Kaede-red in the axon bundle was found to diffuse not only to the lower part of the connected brain organoids but also to the cell bodies located at the top. (E) 3D reconstruction shows the distribution of Kaede-red in axon-associated neurons. (F) Quantification of the Kaede-green and Kaede-red fluorescence intensity relative to the distance from the center of the axon bundle before and after UV irradiation. Kaede-red was concentrated on the side closest to the axon bundle (see G). (H) shows the distribution of Kaede-green and Kaede-red in the z-direction of the connected organoids. It was observed that Kaede-red (neurons extending axon bundles) was not localized at the bottom of the glass plate (see I). We also used flow cytometry to separate and quantify axon-extending neurons. The average percentage of axon-associated neurons was 32% and that of non-axon-associated neurons was 68%, which correlated well with the confocal microscopy images (see J).Panel K shows the relative changes in gene expression levels in neurons extending axons compared to neurons not extending axons. Genes such as CFOS, TBR1, vGLUT1, and NEAT1 were highly expressed in neurons derived from axon bundles, while DLX5, GAD1, and BCL11B were expressed. Furthermore, the scatter plot (L) and UMAP plot (M) of scRNA-seq revealed that NEAT1 and GRIA2 co-expressing cells were abundant in cluster 8, suggesting that these are neurons involved in axon bundle formation. Panel N shows a summary of the results obtained from Kaede's experiments and scRNA-seq. *p<0.05, **p<0.01; student's t-test. Error bars indicate standard deviation (SD).

[0051] Figure 5 shows the results of optogenetic inhibition of burst activity between connected organoids and the synchronization between them. A shows the configuration of the optogenetic device for inhibiting axon-mediated synaptic interactions between left and right organoids. (i) A schematic diagram of the microfluidic device for optogenetic control, and (ii) AAV-mediated light irradiation of ArchT expressed in connected organoids. The optical fiber can be moved and positioned via a fiber guide to selectively irradiate organoids and / or axon bundles. In this example, the fiber was positioned perpendicular to the axon bundles, spaced 100 μm apart. A 470 nm or 565 nm LED and a pulse generator (Arduino) were connected to the optical fiber for MEA measurements. The timing of light irradiation and signals from representative channels from the MEA amplifier were recorded using a TTL logger. The TTL signals were synchronized with neural activity during analysis, and then analyzed (see the analysis section) (iii). The curved PDMS lens structure helped to focus the light onto the axon bundles (see B). C shows LFPs and raster plots detected from the left and right organoids of the connected organoids under light illumination and without light illumination. The synchronized burst frequency was approximately 0.65 Hz without light illumination, but decreased to almost zero with light illumination (see D, n = 8). E shows wavelet coherence between the left and right organoids of the connected organoids. Light illumination eliminated low-frequency oscillations, i.e., correlated activity between organoids (n = 8). Furthermore, light illumination reduced interregional synchrony, measured by delta-delta coupling (see F, n = 8). Light illumination completely suppressed these synchronized burst activities (G, n = 5 from three independent samples). H shows the total number of single spikes calculated over a 5-minute period. Light irradiation induced an increase in spike frequency, and optogenetic inhibition of inter-regional interactions resulted in the persistence of neuronal avalanches (see I).*p<0.05, **p<0.01; one-way ANOVA. Error bars indicate SD (standard deviation).

[0052] Figure 6 shows the results of analyzing complex activity in connected organoids using phase-amplitude coupling and hidden Markov models. (A) Raw (unprocessed) LFP plots detected from each connected organoid at week 9 of iPS cell culture. (B) Wavelet coherence between two organoids. (C) Modulation index of phase-amplitude coupling in delta-phase / gamma power and theta-phase / gamma power for connected organoids at weeks 5, 7, and 9 of culture. (D) PAC modulation index of delta-phase / gamma power and theta-phase / gamma power for single organoids ("S"), fused organoids ("F"), and connected organoids ("C"). (E) PAC modulation index within and between connected organoids. (F) Overview of neuronal avalanche analysis for the left and right organoids. The results of extracting neuronal avalanche cascades are also shown. Neuronal avalanches were calculated based on signals from eight electrodes. Single spike cascades were analyzed at a 3 msec scale. G shows a logarithmic plot of neuronal avalanche size and appearance probability at 5, 5.5, and 8.5 weeks after culture. H shows the number of hidden patterns in neuronal avalanches at 5, 5.5, and 8.5 weeks after culture. I, J, K, and L show the results of comparing neuronal activity patterns in connected organoids treated with various neuromodulatory compounds (CNQX, APV, bicculline, baclofen, buprenorphine, clozapine, and diazepam). The average number of spikes (I), burst frequency (J), integrated delta power (K), and PAC modulation index (L) were calculated. *p<0.05, **p<0.01; one-way ANOVA. Error bars indicate SD (standard deviation).

[0053] Figure 7 shows the results of an analysis of short-term memory mechanisms in connected organoids. (A) Schematic diagram of the optogenic stimulation experiment. Optogenic stimulation of axon bundles using a 470 nm laser light source or LED at 0.5, 1, and 1.5 Hz induced synchronized burst activity. The effect of stimulation persisted even after light irradiation was discontinued. (B) and (C) show measurements of burst frequency modulated by light stimulation. There was a time lag between light stimulation and modulation of burst frequency. (D) Logarithmic plots of neuronal avalanche size (horizontal axis) and occurrence probability (vertical axis) before, during, and after stimulation. (E) Time course of burst frequency after stimulation at 1 Hz for 20 minutes (i) and 5 minutes (ii). The vertical axis represents burst frequency, and the horizontal axis represents time (minutes) from the start of stimulation. F shows the results of measuring the duration of burst activity after light stimulation (20 or 5 minutes). Note that the duration is the time it takes for the burst frequency to decrease to 75% of the maximum burst frequency after cessation of light stimulation. G shows the time lag from light stimulation to burst-induced activity. When connected organoids were stimulated for 20 minutes, the time lag from the onset of light stimulation to burst induction was significantly reduced after the second and third stimulations compared to the first stimulation. H shows the time course of burst frequency in the presence of K252a or anisomycin. I shows the duration of burst activity of connected organoids for each stimulation (1st, 2nd, and 3rd) without compound treatment (control) or with compound treatment (K252a or anisomycin). J shows the time lag from the onset of light stimulation to burst induction for each stimulation in the presence of K252a and anisomycin. K shows the relationship between the number of hidden patterns of neuronal avalanches and the size of the hidden patterns. L shows the results of investigating the probability of neuronal avalanches occurring. The results are for each stimulation (1st, 2nd, and 3rd) in the control, K252a, and anisomycin (on: with light stimulation, off: without light stimulation).M indicates the total number of hidden patterns in neuronal avalanches. Results are for the first, second, and third stimulations (1st, 2nd, and 3rd) in the control, K252a, and anisomycin groups. (Before: before light stimulation, on: during light stimulation (left bar for each stimulation), off: after light stimulation (right bar for each stimulation).) N and O indicate the fractal dimension of the LFP signal. Results are for the first, second, and third stimulations (1st, 2nd, and 3rd) in the control, K252a, and anisomycin groups. (Before: before light stimulation, on: during light stimulation (left bar for each stimulation), off: after light stimulation (right bar for each stimulation). *p<0.05, **p<0.01; one-way ANOVA. Error bars indicate standard deviation.)

[0054] Figure 8 shows the results of analyzing the diverse burst activity patterns supported by CaMKII-dependent signaling. (A) Illustrates burst activity induced by optogenic stimulation and the distribution of neuronal potentials in the organoids on the right and left. 891 burst traces are shown. Spikes and burst waves induced by light stimulation continued after light stimulation. Secondary and tertiary waves of burst activity were also observed. (B) The time lag (ms) after the first, second, and third light stimulations until burst activity was induced was measured in control, K252a, or anisomycin. In the control and anisomycin-containing conditions, repeated light stimulation (second and third stimulations) significantly reduced the time lag from the stimulation to burst activity. In contrast, no reduction in the time lag was observed in the presence of K252a (n = 3). C(i) shows histograms of induced bursts overlaid with kernel density estimates (line graphs) in the control, K252a, or anisomycin-treated control. Repeated light stimulation increased the complexity of burst activity. C(ii) shows the ratio of induced bursts with a second peak to those without a second peak. Results are shown for the control, K252a, or anisomycin-treated control. D shows violin plots of peak burst activity (n = 5). E shows a representative example of crosstalk between connected organoids in self-induced burst activity. F shows the results of quantifying diversity by calculating the entropy of burst activity PCA (n = 3). *p < 0.05, **p < 0.01; one-way ANOVA or student's t-test. Error bars indicate standard deviation.

[0055] Figure 9 shows the results of visualization and analysis of the axonal extension of connected organoids. (A) A schematic diagram of the method for knocking in EGFP and mCherry under the control of the CAG promoter into the AAVS1 region of iPSCs (left) and a fluorescent image of the fluorescent proteins expressed in iPSCs (right). To visualize axonal extension on a PDMS chip, GFP- or mCherry-labeled brain organoids were generated. (B) Comparison of maximum axon length between wild-type and GFP- or mCherry-labeled brain organoids (n=3). The vertical axis of the right graph represents the length of the most extended axon (μm), and the horizontal axis represents the iPS cell culture time (days). (C) Images of axonal extension between organoids spaced at different distances (2 mm, 3 mm, 4 mm, and 5 mm) at 5 and 6 weeks after iPS cell culture. (D) shows the results of measuring the thickness of axon bundles extending between organoids placed at different intervals (2 mm, 3 mm, 4 mm, and 5 mm). Axon bundle thickness was measured at the center of the microchannel. Axon bundle thickness of GFP-labeled connected organoids (n=3). (E) shows a 3D confocal microscope image of GFP-labeled connected organoids. GFP- and mCherry-labeled brain organoids were connected in a microfluidic device. After two weeks on the chip, GFP- or mCherry-labeled axons extended to mCherry-labeled brain organoids and GFP-labeled organoids, respectively, via overlapping thick axons (F). (G) shows the results of counting the number of axons reaching other brain organoids. The vertical axis represents the number of axons reaching each organoid, and the horizontal axis represents the iPSC culture time (days). (H) is a graph showing the relationship between axon bundle thickness and the frequency of synchronized burst activity. The frequency of synchronized burst activity increased as the axon bundle thickness increased. (I) Image showing axon extension toward the other organoid. SynI immunostaining revealed synaptic connections between organoids. **p<0.01; one-way ANOVA or student's t-test. Error bars indicate SD (standard deviation).

[0056] Figure 10 compares the neural activity of single, fused, and connected organoids. (A) Schematic diagram of the procedure for generating single, fused, and connected organoids. All brain organoids were generated using the same method up to day 21. To generate fused organoids, two brain organoids were placed in one well of a low-adhesion 96-well plate. On day 28, single and fused organoids were placed on an MEA probe. On day 28, organoids were also placed on a PDMS-MEA probe to generate connected organoids. After two weeks of culture on the MEA probe, neural activity was measured. (B) A representative example of a single organoid placed on an MEA probe (left). Periodic and synchronized neural activity was detected from the single organoid (right). (C) A representative example of a fused organoid placed on an MEA probe (left). Compared to single organoids, fused organoids demonstrated more active and synchronized neuronal activity. (D) Schematic diagram of the method used to detect the time lag between intra- and inter-organoid neuronal activity using electrodes. (E) Quantification of the time lag in neuronal activity detected by electrodes (n = 4). The time lag between interconnected organoids was significantly smaller than that between inter-fused organoids. (F) Measured signaling speed. Signaling between interconnected organoids was faster than that between other two points (n = 4). *p < 0.05, **p < 0.01; one-way ANOVA or student's t-test. Error bars indicate standard deviation.

[0057] Figure 11 shows organoids connected by axonal bundles of various lengths and the results of measuring their axonal velocity. (A) Shows examples of three types of microfluidic chips (5.5 mm, 7.8 mm, and 12 mm) with different inter-channel lengths. The arrows indicate the positions of the organoids. To estimate axonal conduction velocity, the conduction lag time was plotted against distance (B). The delay constant was 65 msec. The axonal conduction velocity was approximately 2 mm / sec (n = 4).

[0058] Figure 12 shows the results of analyzing neuronal activity in organoids in which axon bundles were physically severed. After physically severing the axon bundles between connected organoids (A) (B), neuronal activity detected in each organoid (left organoid and right organoid) was measured. C shows the results of measuring wavelet coherence in disconnected organoids. These results revealed that there was almost no synchronized activity in the disconnected organoids.

[0059] Figure 13 shows the results of drug treatment on connected organoids. (A) LFP signals of connected organoids at different culture periods (5.5, 6.5, 7, 7.5, 8, and 8.5 weeks). (B) Measurement of signal propagation velocity in the presence of drugs. The results showed that drugs did not affect propagation velocity (n = 4). (C) Measurement of spike amplitude in the presence of drugs (n = 4).

[0060] Figure 14 shows scalograms of the LFP signal and wavelet transform of connected organoids in the presence of various drugs.

[0061] Figure 15 shows Ca 2+ The results of simultaneous measurements of transient and electrical activity are shown. (i) in A shows the Ca concentration of the connected organoids. 2+ The optical setup for imaging and MEA recording is shown. Three days before measurement, Ca was injected into the connected organoids using AAV2.2+ A reporter gene (GCaMP6f) was transiently transfected (see ii and iii). Time-lapse images were then captured using a microscope, while LFP activity from the MEA was simultaneously acquired. (B) Trace images of calcium responses were used to show the firing patterns of neurons in connected organoids at 7 weeks after iPS cell culture. Synchronous burst activity was observed between the left and right organoids (see C). The time lag between bursts between the two connected organoids was consistent with the MEA recordings. (D) Ca 2+ Plots of the concentration change and MEA signal recording are shown. The two signals corresponded to each other. Calcium imaging and MEA recording were consistent during burst activity (see E). F and G show the correlation coefficients obtained from 12 neurons in the left organoid (g) and 12 neurons in the right organoid (g), plotted against each other. Ca 2+ Transient activity was correlated not only within each organoid but also between organoids. However, when the signal was shifted by 50 msec (1 frame), the correlation between organoids became stronger, indicating that there was a delay of about 50 msec between the left and right organoids.

[0062] The results of this example will be described in detail below with reference to the results shown in the above drawings.

[0063] 2-1. Neuronal activity of interconnected brain organoids on a PDMS-MEA chip Many brain regions are connected by reciprocal axonal projections. To model the simplest macroscopic neural circuitry, we cultured a pair of brain organoids derived from human iPS cells in a microfluidic culture chip (Figure 1A and B). The PDMS-MEA chip consists of a multi-electrode array (MEA) layer, a polydimethylpolysiloxane (PDMS) microfluidic layer, and a culture medium reservoir ring and lid. The PDMS layer contains a pair of holes for culturing each brain organoid. These two holes are connected at opposite ends by a spatial channel structure that guides the organoid axons. Four weeks after initiating culture to differentiate human iPS cells into brain organoids, the brain organoids expressed neural markers (e.g., DCX and TUBB3) and cortical layer-related genes (e.g., TBR1 and SATB2) (Figure 1C and D), confirming successful differentiation. After 4 weeks of culture, brain organoids were placed on PDMS-MEAs. Two brain organoids were connected by thick axonal bundles within 6 weeks (2 weeks after placement on the chip) (Figure 1E). On the chip, GFP-expressing and mCherry-expressing brain organoids exhibited "handshake"-type connections (Figure 9A-I). Two brain organoids connected to each other with similar numbers of axons (Figure 9G). The thickness of the axonal bundles was approximately 75 μm after 6 weeks and approximately 120 μm after 8 weeks (4 weeks after placement on the chip) (Figure 1F). vGlut1-positive excitatory neurons and GAD67-positive inhibitory neurons accounted for approximately 70% and 5-10% of the cells in the brain organoids, respectively (Figure 1G). A layered structure was observed in the subcortical regions of the axon-connected brain organoids (Figure 1H). As described above, it is thought that when two brain organoids are cultured, they become connected to each other by axons, thereby mimicking the functions of the developing brain.

[0064] Neuronal activity (spontaneous firing) was detected using electrodes attached under each brain organoid (Figure 1B and I). Action potential spikes and local field potentials (LFPs) were extracted using high- and low-frequency filters, respectively. After 5 weeks (1 week after culturing on a chip), neuronal activity was detected from each brain organoid (Figure 1J). At this time point, the activity of the two brain organoids was not synchronized, and axonal connections between the two brain organoids were not sufficient (Figure 1E). After 5.5–6 weeks (1.5–2 weeks after culturing on a chip), synchronized burst-like activity was observed between the two brain organoids (Figure 1K, arrows). The burst-like activity of each brain organoid increased rapidly, and the synchronization of the two burst-like activities coincided with the establishment of axonal connections between the two brain organoids. As the brain organoids were cultured, the frequency of synchronized neuronal activity and the frequency of burst-like activity increased (Figure 1L and M). A positive correlation was observed between the thickness of axon bundles and the frequency of neural activity (Figure 9H). These results suggest that axon bundles facilitate neural activity in connected organoids. Furthermore, signals from two connected brain organoids were detected with a short time delay (less than 100 ms) (Figure 1N and O). These results suggest that spontaneous neural activity originated in one brain organoid and propagated to the other via the axon bundles with a slight time delay. Furthermore, these results demonstrate that the two brain organoids induced signal propagation to each other, demonstrating functional bidirectionality of the connection.

[0065] 2-2. Complex neuronal activity induced by axonal connections between brain organoids Six weeks after iPS cell culture (two weeks after culturing on chip), slow LFP signals were detected in each brain organoid (Figure 2A), which were not detected before axonal connections were established. Clustered action potentials from multiple neurons are known to generate coordinated, low-frequency LFP patterns. After an additional week of culturing on chip (seven weeks after iPS cell culture), the LFP patterns of the connected brain organoids became more complex, with strong action potentials appearing in the delta wavelength band (0.5-4 Hz) (Figures 2A and B). Given previous reports showing that brain organoids only generate delta wavelength band action potentials after several months of culture (Trujillo et al., Cell Stem Cell 25 558-569.e7 2019), the detection of strong, complex action potentials from brain organoids at such an early stage was unexpected. Therefore, we compared the neuronal activity of the "connected" (axon-connected) organoids in this study with that of conventional brain organoids ("single organoids"). Because connected organoids contain nearly twice the number of cells compared with single organoids, we decided to evaluate the neuronal activity of tissues formed by directly fusing two organoids ("fused" organoids) as an additional control (Figure 2C and Figure 10A). After 6 weeks of iPSC culture, oscillatory neuronal activity was detected in all brain organoids (Figure 2C, Figures 10B and C). Compared to single organoids, burst-like activity was observed more frequently in fused organoids, suggesting that the total number of neuronal cells in the organoids influences neuronal activity. The frequency of burst-like activity from connected organoids was significantly higher than that of single or fused organoids, but the number of neuronal cells contained in fused and connected organoids was approximately the same (Figure 2D). These findings suggest that axonal connections between organoids enhance neuronal activity in cerebral organoids.

[0066] To understand the network structure of organoids, we analyzed the signal propagation speed, calculated from the distance between two electrodes and the time lag between signals detected from these electrodes. The signal transmission speed between fused organoids was slower than that within single organoids (Figure 10E and F), whereas the signal transmission speed between connected organoids was faster than that within single organoids (Figure 10E and F). The time lag of signal transmission between organoids was significantly greater in fused organoids than in connected organoids. The transmission speed was so fast that changes in axon length did not affect the signal transmission delay (below the measurement limit) (Figure 11). This result suggests that the configuration of synaptic connections within organoids determines the signal transmission speed. These results suggest that axonal connections between organoids serve as a "highway" for signals between connected organoids, facilitating neuronal activity.

[0067] Compared with single and fused organoids, connected organoids showed activity concentrated in the delta wavelength band (0.5-4 Hz) (Figure 2E). To quantify the temporal complexity of neuronal activity in another way, we assessed the periodicity of the organoids' burst-like patterns. The coefficient of variance (CV, which indicates variability; standard deviation / arithmetic mean) of the frequency of burst-like neuronal activity was significantly higher in connected organoids than in single and fused organoids (Figure 2F). This suggests that burst-like neuronal activity in connected organoids is significantly more irregular and complex than that in single and fused organoids. The CV of connected organoids consistently increased up to 9 weeks after iPSC culture, whereas the CV of single and fused organoids did not show significant fluctuations during this culture period. The burst activity of fused organoids was more frequent than that of single organoids, consistent with their size (Figure 2G), suggesting that this is due to the increased number of neurons and synapses within the tissue. On the other hand, connected organoids exhibited a higher frequency of burst-like neuronal activity than single or fused organoids, regardless of their size. These results suggest that connected organoids form more complex neural circuits than single or fused organoids, and that interconnections between organoids drive complex neuronal activity in brain organoids.

[0068] 2-3. Gene expression analysis in connected organoids using single-cell RNA-seq We evaluated how interconnections between organoids altered gene expression profiles within the organoids using single-cell RNA-seq. We performed single-cell (sc) RNA-seq on single, fused, and connected organoids at week 7 in culture. Principal component analysis (PCA) of 17,636 single cells aggregated from all three conditions was performed and visualized as a UMAP plot (Figure 3A(i)). Gene expression profiles of single, fused, and connected organoids were generally similar, with minor differences (Figure 3A(ii)). Cells were separated into 14 clusters, divided into four groups: "NPC," "intermediate," "neuron," and "other" clusters (Figure 3B; the top 30 statistically significant genes were plotted as a heatmap). According to known markers (e.g., HES1, DCX, TBR1), we determined “NPC” (clusters 6, 12, 13), “Intermediate” (clusters 1, 2, 3, 11), “Neurons” (clusters 4, 5, 7, 8, 9, 10), and “Other” (cluster 0) ( Figure 3C ).

[0069] Density plots of single, fused, and connected organoids revealed variations in cell populations across the three samples (Figure 3A(ii)). In particular, clusters 1, 4, and 8 were more abundant in connected organoids than in single and fused organoids. Furthermore, the number of cells classified as NPC clusters 6 and 12 was lower in connected organoids than in single and fused organoids (Figure 3C). Furthermore, annotating clusters 4 and 8 by differential expression revealed that they may be involved in mature excitatory neurons expressing neurotransmitters (Figure 3B). These clusters were further classified as "excitatory neurons" (clusters 4 and 8) and "inhibitory neurons" (clusters 9 and 10), characterized by the expression of VGLUT1 and DLX6, respectively (Figure 3E and G). The ratio of excitatory to inhibitory neurons in connected organoids was approximately 85:15, almost the same as that in single and fused organoids (Figure 3G). This result was consistent with the immunohistochemistry results of brain organoids (Figure 1H). Furthermore, GRIA2, a glutamate ionotropic receptor AMPA subunit II, was highly expressed in neurons of cluster 8, while GRIA1 was predominantly expressed in other clusters (Figure 3H), indicating that neurons in cluster 8 follow the maturation process that occurs in vivo (Balik et al., 2013). Furthermore, the expression level of GRIA2 in connected organoids of cluster 8 was higher than that in single and fused organoids (Figure 3I). These results suggest that connected organoids have more mature neurons than single and fused organoids, and that these neurons are more mature in connected organoids. Notably, among DCX-expressing neurons, GRIA2 expression levels were elevated in connected organoids (Figure 3J), suggesting that increased neuronal activity may contribute to more mature neurons in connected organoids, as previously suggested (Gordon et al., Nature Neuroscience 24 331-342 2021).Gene-ontology analysis revealed that neurons in cluster 8 express various receptor and voltage-gated channel-related genes (GRIN2B, CANA1A, CANA1E, SCN2A, NTRK2).

[0070] 2-4. Identification of axon-specific cell populations using Kaede To identify neurons extending axons from brain organoids, we used the fluorescent protein Kaede, whose fluorescence wavelength changes with light exposure (Figure 4A). Kaede (Kaede-green) can be readily converted to red-fluorescent Kaede (Kaede-red) by UV irradiation (Ando et al., Proc. Acad. Natl. Sci., 99 12651-12656 2002). At week 5 of culture, connected organoids were infected with a Kaede-expressing virus (AAV-CAG-Kaede). Then, at week 7 of culture, we performed photoconversion by UV irradiation using a confocal microscope (Figure 4B). Before UV irradiation, Kaede-green was distributed throughout the connected organoids, but Kaede-red was almost absent (Figure 4C). Photoconversion was induced by 60 minutes of UV irradiation at the center of the axon bundle. Furthermore, we successfully visualized axon-associated neurons in connected organoids by translocating the converted Kaede red protein intracellularly for 2 hours (Figure 4C). The distribution of green and red Kaede revealed the location of neurons associated with axon bundles within connected organoids (Figure 4D). Red Kaede was distributed equally in both left and right organoids (Figure 4F). Within the organoids, axon-associated neurons were more abundant in regions closer to axon bundles (Figure 4G). Axons in the lowest layer of connected organoids were distributed almost evenly across the z-axis, except for axons in the lowest layer, which did not extend into the bundles very much (Figure 4H). Flow cytometry analysis of dissociated neurons from UV-irradiated connected organoids revealed that approximately 30% of neurons extended their axons into the bundles between organoids (Figure 4J), based on the Kaede green / red ratio (Figure 4I). Furthermore, the expression levels of FOS, TBR1, VGLUT1, GRIA2, and NEAT1 were significantly higher in neurons associated with axon bundles than in neurons not associated with axon bundles (Fig. 4K), which was similar to the gene expression pattern of neurons in cluster 8 in scRNA-seq (Fig. 3C, G, H and Fig. 4L, M). The results of both Kaede-labeling and scRNA-seq suggested that neurons forming axon bundles may have higher expression levels of TBR1, GRIA2, VGLUT1, and NEAT1 and may be more mature neurons.From a developmental and structural perspective, the neurons involved in the formation of these axon bundles may play an important role in performing advanced neural activity (Fig. 4N).

[0071] 2-5. Optogenic inhibition of axon bundles between connected organoids To investigate whether axonal connections play an important role in the complex neuronal activity of connected organoids, we performed an experiment to suppress neurotransmission activity through axonal bundles between organoids. Physically severing the axonal bundles between organoids resulted in the disappearance of burst-like neuronal activity and the LFP pattern in the delta wavelength band. Therefore, macroscopic connections appear to be important for the generation of complex and periodic neuronal activity patterns in brain organoids (Figure 12).

[0072] To further explore the role of axonal connections between organoids, we performed optogenic inhibition of axonal bundles between organoids. To achieve this, we modified the microfluidic chip (Figure 5A). We used an adeno-associated virus (AAV) vector to express ArchT (archirhodopsin, a H receptor that responds to light). + We expressed a mitogen-activated receptor (MTR) in organoids. Using an optical fiber and a PDMS lens, we selectively suppressed axon bundles in the microchannels with light (Figure 5B). Irradiating the axon bundles with orange light (568 nm, 20 ms, 20 Hz for 5 minutes) completely suppressed neuronal activity in each organoid. The highly amplified burst-like activity and low-frequency delta-LFP pattern disappeared under light irradiation (Figure 5C and D). These results suggest that action potentials propagate through axon bundles to induce burst-like activity in connected organoids. Upon cessation of light irradiation, the burst-like activity and delta-LFP pattern immediately recovered. This light-induced response was repeatedly observed after light irradiation. The frequency of spontaneous burst-like activity remained unchanged after light irradiation, suggesting that the frequency of spontaneous burst-like activity is determined by the intrinsic properties of the neural circuitry within the organoids.

[0073] Furthermore, the coherence and synchrony of signals from each organoid in the connected organoids were lost during light irradiation (Figure 5E and F). Optogenically severing the connection between the two organoids significantly suppressed the overall intensity of neuronal activity, resembling the state observed after surgery in patients with intractable epilepsy. Optogenic inhibition of the inter-organoid axons eliminated burst-like activity, but the number of observed action potentials increased during light irradiation (Figure 5G and H). These results suggest that the inter-organoid axons contribute to the induction of burst-like activity by coordinating and integrating the activity of individual neurons within the connected brain organoids. Furthermore, neuronal avalanches were reduced by light irradiation (Figure 5I). These results suggest that the continuous firing of action potentials observed as neuronal avalanches is inhibited by optically inhibiting the inter-organoid axons, indicating that inter-organoid connections are important for the generation of neuronal avalanches. These results demonstrate that activity transmitted via axonal bundles between organoids underlies the complex neuronal activity that occurs in connected organoids, supporting the importance of macroscopic connections in the brain.

[0074] 2-6. Phase-Amplitude Coupling (PAC) of Connected Organoids Culturing neural circuit tissue for more than 8 weeks further increased the LFP frequency and action potential spikes in the connected organoids (Figure 13A). Furthermore, the signal complexity increased, with activity in the theta wavelength band becoming more frequent (Figure 6A). Strong correlations were also observed between the gamma and delta wavelength bands, and between the gamma and theta wavelength bands. To examine the correlation between these wavelength bands, we calculated phase-amplitude coupling (PAC), an index of the relationship between low-frequency activity and the amplitude of high-frequency spikes. PAC is an established method for assessing the relationship between different frequency bands of electroencephalogram (EEG) signals (Fell et al., Nat. Rev. Neurosci. 12 105–118 2011; Canolty et al., Trends Cogn. Sci. 14 506–515 2010). In terms of burst activity, LFP waves in the delta and theta wavelength bands appeared as synchronized bursts coordinated with the gamma wavelength band (Figure 6B). The delta-γPAC modulation of connected organoids increased with continued culture, followed by an increase in theta-γPAC modulation (Figure 6C). Both delta-γPAC and theta-γPAC modulation in connected organoids were significantly higher than PAC modulation in single or fused organoids (Figure 6D). Within each connected organoid, delta-γPAC and theta-γPAC showed high modulation (intra-organoid PAC). The delta-γPAC and theta-γPAC modulation between organoids connected by axon bundles (inter-organoid PAC modulation) were higher than intra-organoid PAC modulation, indicating close communication in the delta and theta frequency bands between the two organoids (Figure 6E).

[0075] MEA measurements have enabled the recording of extracellular and local neuronal activity with high spatiotemporal resolution. However, the low density of electrodes compared to the cell density within organoids can make it difficult to understand the spike activity of action potentials from individual neurons. 2+Although imaging allows for measurements with higher spatial resolution than the electrode-based method, it is difficult to simultaneously obtain high-magnification images from two organoids using a conventional microscope. Therefore, we performed imaging of individual Ca2+ receptors in two organoids located a few mm apart. 2+ To simultaneously acquire images, we set up two microscopes with independent optical systems. This system enabled us to simultaneously capture Ca signals from two organoids with high spatiotemporal resolution, even when combined with MEA recording. 2+ Imaging signals can be acquired (Figure 15A). 2+ The indicator GCaMP6f was expressed in the connected organoids via AAV for 3-7 days (Fig. 15A). Camera image frames for the left and right organoids were synchronized using a trigger signal, and images were acquired at a rate of 20 frames per second. The change in fluorescence intensity (ΔF / F) in neurons within the organoids was calculated. The fluorescent signal displayed burst activity synchronized with the activity of isolated individual neurons in both organoids (Fig. 15B). The interval between bursts was consistent with the burst interval measured by MEA (Fig. 15C). MEA and Ca 2+ Simultaneous imaging of MEA signals and Ca in organoids 2+ The increase in concentration was consistent (Figure 15D). To identify sets of neurons with correlated activity, we created a correlation matrix of 26 neurons (13 neurons in the left organoid and 13 neurons in the right organoid; Figure 15F). During burst activity, neurons within the same organoid exhibited strong correlation, whereas correlation between neurons in different organoids was weak (Figure 15G). When analyzing correlation, we shifted the neuronal activity within a single organoid by 50 msec, and observed a strong correlation between the organoids, indicating a 50 msec lag in the activity of the two organoids, which was consistent with the results observed by MEA (Figure 1N).

[0076] Next, we examined the neuronal avalanches of signals recorded from connected organoids (Figure 6F). Temporally proximal signals obtained from the electrodes were grouped and quantified as neuronal avalanches. The neuronal avalanche index is considered a scale-free index of the critical dynamics of a network (Beggs et al., Journal of Neuroscience 23 11167-11177 2003; Bowen et al., Frontiers in Systems Neuroscience 13 45 2019). The distribution of avalanche sizes increased with continued culture of connected organoids. The neuronal avalanche index of connected organoids showed α = -2.8 at 5 weeks of iPSC culture (Figure 6G). It then increased to α = -2.1 at 5.5 weeks and reached α = -1.6 at 8.5 weeks. This α value theoretically corresponds to an exponent of -3 / 2 in the critical avalanche process. Furthermore, we investigated the spatiotemporal patterns of neuronal avalanches using pattern recognition with a hidden Markov model, and found that the variation in the patterns of neuronal avalanches increased with culture time (Figure 6H).

[0077] 2-7. Dynamic synaptic balance underlying the complex activity of connected organoids To investigate the balance of synaptic receptors underlying the complex neuronal activity of connected organoids, we treated them with the synaptic channel antagonists CNQX, APV, and vinculin. CNQX, APV, and bicuculline are antagonists of NMDA-type glutamate receptors (excitatory receptors), AMPA-type glutamate receptors (excitatory receptors), and GABA-type receptors (inhibitory receptors), respectively (Figures 6I-L, 13A, and 14). Antagonist treatment did not alter signal propagation velocity (Figures 13B and C). However, each antagonist affected neuronal activity, including action potential spikes and burst-like activity (Figures 6I and J). CNQX and APV reduced neuronal activity, i.e., delta-gamma PAC and theta-gamma PAC, in connected organoids, suggesting that excitatory synaptic transmission plays an important role in generating the complex activity of connected organoids (Figure 6K and L). On the other hand, bicuculline treatment significantly increased spike activity but did not alter the number of burst-like activity. However, the burst-like activity was attenuated by bicuculline treatment (Figure 14). These results suggest that inhibitory synapses play an important role in generating the rhythmic neuronal activity of connected connectoids.

[0078] Next, we examined the effects of clinical compounds that alter synaptic properties. Treatment with the GABA receptor agonists baclofen and diazepam reduced spike counts (Figure 6I and J). Treatment with the antipsychotic clozapine modestly reduced neuronal activity but had no effect on PACs (Figure 6I, K, and L). Treatment with the opioid buprenorphine specifically reduced theta-gamma PACs, suggesting that buprenorphine affects the coordination of activity across different frequency bands in connected organoids (Figure 6I, K, and L). These results demonstrate the utility of connected organoids for investigating the effects of compounds on complex neuronal activity.

[0079] 2-8. Optogenic stimulation-induced memory formation in connected organoids To investigate the detailed response of connected organoids to external stimuli, we again used optogenetic techniques. Channelrhodopsin was expressed in connected organoids to elicit action potentials via light irradiation. Axons of connected organoids, which spontaneously fired at 0.5 Hz, were stimulated with pacing light (470 nm, 200 ms) at the same frequency as the spontaneous firing (0.5 Hz) for 5 minutes. The connected organoids were then stimulated at a higher frequency of 1.0 Hz for 5 minutes, followed by another 5 minutes at 1.5 Hz (Figure 7A). Burst-like activity increased in accordance with the temporal stimulation pattern of light irradiation. After cessation of stimulation, the frequency of burst-like activity remained high for more than 10 minutes, after which it returned to the pre-stimulation frequency (Figure 7B and C). Thus, light stimulation induced sustained echo-like neuronal activity, indicating that the temporal activity patterns of connected organoids can be modulated by external stimuli and that connected organoids can maintain temporal information of stimuli, further suggesting that connected organoids may maintain temporal information as a primitive (basic) form of memory. In particular, there was a time lag between the onset of stimulation and the emergence of burst-like activity (Figure 7C), suggesting that multiple stimulations were required to modulate the activity of connected organoids. Furthermore, during this time lag, the neuronal avalanches of connected organoids expanded as they were periodically stimulated (Figure 7D). This result suggests that by the time the activity of connected organoids responded to the temporal stimulation pattern, the neurons comprising each organoid slowly adapted to the stimulation and established excitable local circuits (subcircuits).

[0080] Stimulation of connected organoids at 1 Hz for 20 minutes induced sustained activity even after stimulation was discontinued (Figure 7E and F). Burst activity continued to increase for approximately 20 minutes after stimulation (decay period), after which burst activity returned to its original state. In contrast, burst-like activity in connected organoids stimulated at the same frequency for 5 minutes immediately returned to its pre-stimulation state. Although burst frequency increased during light stimulation, no sustained increase in burst activity was observed. This result indicates that continuous stimulation is necessary for the enhancement of circuit-level activity in connected organoids. Next, we examined the effect of periodic optogenic stimulation on connected organoids. Stimulation was performed three times every 60 min (Fig. 7E and F). The burst frequency returned to normal after each stimulation. In particular, when the stimulation time was 20 min, the latency to respond to external stimuli was significantly shorter after the second or third stimulation compared to the first stimulation (Figure 7G). On the other hand, when the connected organoids were stimulated for 5 min, the latency to respond to external stimuli remained unchanged. These results suggest that the connected organoids respond quickly to the next repeated stimulation through the network-level enhancement of activity by the previous stimulation.

[0081] Neuronal plasticity and memory are regulated by synaptic plasticity via diverse molecular programs, including calcium-dependent signaling pathways and local protein synthesis required for early and late responses. To explore the mechanisms underlying memory formation in connected organoids, we treated connected organoids with K252a, a CaM kinase II inhibitor, and anisomycin, a protein synthesis inhibitor (Figure 7H). Connected organoids responded to light stimulation, and sustained activity was observed after stimulation in the presence of K252a or anisomycin. The decay period of sustained activity after stimulation was significantly reduced in the presence of K252a after the second and third stimulations (Figure 7I). In contrast, anisomycin treatment only slightly shortened sustained activity after the third stimulation. Furthermore, K252a treatment inhibited the shortening of the time lag from the onset of the second stimulation to the response compared to control treatment (Figure 7J). These results suggest that a calcium-dependent signaling pathway underpins the short-term potentiation observed in connected organoids.

[0082] K252a treatment also reduced neuronal avalanches (Figure 7K). During light exposure, the number of hidden patterns in neuronal avalanches was reduced compared to avalanches without light exposure (Figure 7L). This result suggests that during light exposure, organoid activity is fixed to a constant cycle of stimulation, resulting in a decrease in the amount of information contained in neuronal activity. Notably, the number of hidden patterns gradually increased with each stimulation, suggesting that neural circuits within the connected organoids developed and matured through stimulation and the proliferation of neural activity induced by stimulation. Anisomycin treatment inhibited the expansion of hidden patterns in neuronal avalanches during the light withdrawal period (Figure 4M), suggesting that the mechanism of long-term potentiation was partially activated within the connected organoids, although the conventional form of long-term potentiation was not observed.

[0083] Next, we calculated Higuchi's fractal dimension to examine the complexity of connected organoid activity. Fractal dimension (FD) is an index of fine-scale structural complexity and has been used to assess the complexity and temporal changes of EEG signals (Varley et al., PLOS ONE 15 e0223812 2020). In controls, FD increased after the cessation of the first light stimulation, and the increased FD was maintained even in the absence of a second or third light exposure (Figure 7N and O). This result suggests that the entire network acquired complexity after the first light exposure and maintained it thereafter. In the presence of K2532a, FD did not increase with the first light stimulation, but it did increase with the second light exposure, and this increase in FD was maintained with the third light exposure. This result suggests that K252a treatment disrupted the enhancement and complexity of network activity. In contrast, in the presence of anisomycin, FD did not change with light stimulation.

[0084] To better understand the burst-like activity, we sorted and compared the induced activity (Fig. 8A). Detailed analysis revealed that optical stimulation induced multiple EEG waves of neuronal activity in burst-like activity. In the control group, the time lag of the burst-like activity shortened with each repetition of stimulation (Fig. 8B). This shortening of the time lag was also observed with anisomycin treatment, but not with K252a. To further analyze the induced burst-like activity, we calculated the probability of the induced burst (Fig. 8C and D, kernel density estimation). The results showed that a sharp initial peak was followed by secondary and tertiary responses. In the first stimulation, the initial peak was observed near the end of the stimulation, followed by a weaker secondary peak. In the second and third stimulations, the initial peak became stronger and shifted toward the onset of the stimulation. Furthermore, the secondary and tertiary peaks became sharper and stronger with each repetition of stimulation (Fig. 8C and D). Secondary and tertiary waves were also observed under K252a and anisomycin treatment. Notably, left and right organoids connected via axon bundles sometimes responded to light with slightly different kinetics (Figure 8E). Furthermore, the induced second and tertiary burst responses were observed alternating between the two connected organoids (Figure 8E). These results suggest that activity in connected organoids is generated and developed by a complex combination of activities within and between the left and right organoids.

[0085] Next, to quantify the diverse patterns of induced activity, we calculated the entropy of the waveforms of burst-like activity obtained from connected organoids (Figure 8F). Repeated stimulation increased the entropy of induced burst-like activity, indicating that this quantification method can capture the development of complex burst waves. This increase in entropy was inhibited by K252a treatment, but not by anisomycin treatment. Therefore, we suggest that changes in burst waves are also driven by a calcium signaling-dependent mechanism. [Industrial Applicability]

[0086] The present invention provides an in vitro neural circuit tissue that exhibits complex activity and plasticity, particularly neural tissue that mimics brain function. It also provides an apparatus for observing the neural activity and a method for screening substances using the same. Therefore, it is highly anticipated that this brain model will be used in fields such as medicine and pharmacy.

Claims

1. An apparatus for observing neural activity of a neural circuit tissue induced in vitro in which two or more neural cell masses are connected via axons, comprising: a substrate; a plurality of wells provided on the surface of the substrate for accommodating the neuronal clusters; and guide grooves provided by connecting the wells for guiding and extending the axons so as to interconnect the neuronal clusters; An electrode is provided in each of the wells, and the device further includes an analysis unit that analyzes correlations of electrical signals from the electrodes, wherein the analysis unit separates each of the electrical signals into frequency bands and analyzes the correlations by phase amplitude coupling; The substrate comprises a first substrate having the electrodes provided on its surface, and a second substrate having a plurality of wells each consisting of a through-hole and the guide groove formed therein, wherein the first substrate and the second substrate are separably combined so that the electrodes are placed at the bottom of the well and the bottom is closed.

2. An apparatus for observing neural activity of a neural circuit tissue induced in vitro in which two or more neural cell masses are connected via axons, comprising: a substrate; a plurality of wells provided on the surface of the substrate for accommodating the neuronal clusters; and guide grooves provided by connecting the wells for guiding and extending the axons so as to interconnect the neuronal clusters; An electrode is provided in each of the wells, and the device further includes an analysis unit that analyzes a correlation of an electrical signal from the electrodes, the analysis unit analyzing the correlation using wavelet coherence, and The substrate comprises a first substrate having the electrodes provided on its surface, and a second substrate having a plurality of wells each consisting of a through-hole and the guide groove formed therein, wherein the first substrate and the second substrate are separably combined so that the electrodes are placed at the bottom of the well and the bottom is closed.

3. An apparatus for observing neural activity of a neural circuit tissue induced in vitro in which two or more neural cell masses are connected via axons, comprising: a substrate; a plurality of wells provided on the surface of the substrate for accommodating the neuronal clusters; and guide grooves provided by connecting the wells for guiding and extending the axons so as to interconnect the neuronal clusters; An electrode is provided in each of the wells, and the device further includes an analysis unit that analyzes correlations of electrical signals from the electrodes, and the analysis unit classifies the neural activity into an action potential, a burst activity, a neural avalanche, or a local field potential based on a signal pattern of the electrical signal; The substrate comprises a first substrate having the electrodes provided on its surface, and a second substrate having a plurality of wells each consisting of a through-hole and the guide groove formed therein, wherein the first substrate and the second substrate are separably combined so that the electrodes are placed at the bottom of the well and the bottom is closed.

4. 3. The apparatus for observing neural activity of neural circuit tissue according to claim 1, wherein the analysis unit classifies the signal pattern of the electrical signal into one of action potentials, burst activity, neural avalanches, and local field potentials.

5. 5. The apparatus for observing neural activity of neural circuit tissue according to claim 3, wherein the analysis unit has in advance the signal patterns corresponding to the neural activity and classifies the neural activity by comparing the signal patterns with the signal patterns.

6. 6. The apparatus for observing neural activity of neural circuit tissue according to claim 5, wherein the neural activity is generated by applying an external stimulus to the nerve cell cluster and / or the axon, and a corresponding signal pattern is acquired in advance.

7. 7. The apparatus for observing neural activity of neural circuit tissue according to claim 6, wherein the stimulus is one or more of light irradiation, electrical stimulation, and administration of a compound.

8. The apparatus for observing neural activity of neural circuit tissue according to claim 7 , wherein the stimulation enhances or suppresses the neural activity.

9. 9. The observation device for neural activity of neural circuit tissue according to claim 1, wherein the electrode comprises a plurality of electrodes arranged in an array on the bottom of the well.

10. 10. The apparatus for observing neural activity of neural circuit tissue according to claim 9, wherein the electrode receives the electrical signal from the nerve cell mass and electrically stimulates the nerve cell mass.

11. 11. The apparatus for observing neural activity in neural circuit tissue according to claim 9 or 10, wherein the substrate is transparent so that the neural cell mass in the well can be optically observed from the bottom.

12. An apparatus for observing neural activity of a neural circuit tissue induced in vitro in which two or more neural cell masses are connected via axons, comprising: a substrate; a plurality of wells provided on the surface of the substrate for accommodating the neuronal clusters; and guide grooves provided by connecting the wells for guiding and extending the axons so as to interconnect the neuronal clusters; An electrode is provided in each of the wells, and an analysis unit that analyzes the correlation of electrical signals from the electrodes is further included, and a light irradiation device for optically stimulating the nerve cell cluster and / or the axon is provided opposite the surface of the substrate, The substrate comprises a first substrate having the electrodes provided on its surface, and a second substrate having a plurality of wells each consisting of a through-hole and the guide groove formed therein, wherein the first substrate and the second substrate are separably combined so that the electrodes are placed at the bottom of the well and the bottom is closed.

13. An observation device for neural activity of neural circuit tissue described in any one of claims 1 to 11, wherein a light irradiation device for optically stimulating the neural cell cluster and / or the axon is provided opposite the surface of the substrate.

14. The apparatus for observing neural activity of neural circuit tissue according to claim 12 or 13, wherein the light irradiation device includes an irradiation unit that locally irradiates the nerve cell mass and / or a part of the axon with light.

15. A method for screening a substance that alters neural activity of neural circuit tissue using the observation device according to any one of claims 1 to 14, comprising: A substance screening method comprising setting a target neural circuit tissue in the observation device, administering the substance to the neural circuit tissue, and observing fluctuations in the neural activity.

Images