Mutual interaction of neurons and astrocytes derived from iPSCs with APP V717L mutation developed the astrocytic phenotypes of Alzheimer’s disease

Background The development of induced pluripotent stem cells (iPSCs) technology has enabled human cellular disease modeling for inaccessible cell types, such as neural cells in the brain. However, many of the iPSC-derived disease models established to date typically involve only a single cell type. These monoculture models are inadequate for accurately simulating the brain environment, where multiple cell types interact. The limited cell type diversity in monoculture models hinders the accurate recapitulation of disease phenotypes resulting from interactions between different cell types. Therefore, our goal was to create cell models that include multiple interacting cell types to better recapitulate disease phenotypes. Methods To establish a co-culture model of neurons and astrocytes, we individually induced neurons and astrocytes from the same iPSCs using our novel differentiation methods, and then co-cultured them. We evaluated the effects of co-culture on neurons and astrocytes using immunocytochemistry, immuno-electron microscopy, and Ca2+ imaging. We also developed a co-culture model using iPSCs from a patient with familial Alzheimer's disease (AD) patient (APP V717L mutation) to investigate whether this model would manifest disease phenotypes not seen in the monoculture models. Results The co-culture of the neurons and astrocytes increased the branching of astrocyte processes, the number of GFAP-positive cells, neuronal activities, the number of synapses, and the density of presynaptic vesicles. In addition, immuno-electron microscopy confirmed the formation of a tripartite synaptic structure in the co-culture model, and inhibition of glutamate transporters increased neuronal activity. Compared to the co-culture model of the control iPSCs, the co-culture model of familial AD developed astrogliosis-like phenotype, which was not observed in the monoculture model of astrocytes. Conclusions Co-culture of iPSC-derived neurons and astrocytes enhanced the morphological changes mimicking the in vivo condition of both cell types. The formation of the functional tripartite synaptic structures in the co-culture model suggested the mutual interaction between the cells. Furthermore, the co-culture model with the APP V717L mutation expressed in neurons exhibited an astrocytic phenotype reminiscent of AD brain pathology. These results suggest that our co-culture model is a valuable tool for disease modeling of neurodegenerative diseases. Supplementary Information The online version contains supplementary material available at 10.1186/s41232-023-00310-5.


Background
Induced Pluripotent Stem Cells (iPSCs) are stem cells derived from adult somatic cells by introducing specific transcription factors such as Oct3/4, Sox2, Klf4, and c-Myc [1,2].Like embryonic stem cells (ESCs), the iPSCs have proliferative and pluripotent properties.Therefore, iPSCs can be differentiated into various cell types and generate human cell models in vitro [3].Since the iPSCs hold the donor's genetic information, it allows the in vitro study of human diseases with the patient's own genetic background [4].In recent years, especially for neurodegenerative disease research, the human cellular models generated from the iPSCs have been widely employed [5,6].For example, in the studies of Alzheimer's disease (AD), the neurons induced from the iPSCs expressed the disease phenotypes according to the disease characteristic of the donor [7,8].
Regarding the iPSC-derived cellular models used for disease analysis, the conventional approach involves the induction of iPSCs into a two-dimensional model (2D model) consisting of a single cell type.However, the monoculture model often lacks cell-cell interactions between different cell types of cells when compared to the human brain [9].In recent years, the three-dimensional model (3D model) of brain organoids has been developed and applied to the study of neurodegenerative diseases including AD [10][11][12][13].Although the brain organoid provides a platform for many cell types, it requires complicated culture procedures and relatively long culture periods.In particular, a relatively long culture period of more than 300 days is required to obtain specific cell types including astrocytes that exhibit astrocyte-like morphology [14].Therefore, a model that promotes cell maturation and allows the study of cultured cells in a short period of time will be a useful tool for studying disease in a dish.
Furthermore, there was only a limited number of disease phenotypes that have been reported in a study of age-related diseases under the monoculture model of a single cell type.In particular, studies of AD have recapitulated only a few of the phenotypes of the non-neuronal population [15].In recent years, a 2D cell model that includes more than a single cell type, called the coculture model, has been developed.The co-culture model allows the analysis of the cell in a more physiological environment where more than one cell type is present compared to the monoculture model.A study of the coculture model comprising the iPSC-derived neurons, astrocytes, and microglia succeeded in recapitulating neuroinflammation given the secreted complement component 3 (C3) from the microglia under the co-culture environment [16].However, the changes in cellular properties, cell-cell interactions, and the phenotypes of other cell types under the co-culture system were still under investigation.
In the present study, our co-culture model of neurons and astrocytes which were separately differentiated from the iPSCs using the newly reported induction methods, was successfully established.This co-culture model demonstrated the promotion of morphological changes such as in vivo state in both neurons and astrocytes, the formation of functional tripartite synapse structure, and the enhancement of astrocytic disease phenotypes in the model of AD donor model (APP V717L ).Thus, this newly developed model provided a useful platform for modeling neurodegenerative diseases in human cells.

Methods iPSC culture
The human iPSC lines used in this study are a healthy control line (201B7 line, female, 36 years old) [2] and a familial AD line with the APP V717L mutation (APP2E26 line, female, 57 years old) [17].The cell culture method for iPSCs followed the protocol established by the Center for iPS Cell Research and Application (CiRA) [18].The iPSCs were cultured in the feeder-free condition with StemFit ® AK02N medium (Ajinomoto) on culture dishes pre-coated with 3.0 μg/mL iMatrix-511 silk (Matrixome 892021) in a humidified atmosphere with 5% CO 2 .The culture media were exchanged every other day.The iPSCs were cultured for 7 days before passaging using 0.5 × Try-pLE Select (Thermo Fisher Scientific).After passage, cells were seeded onto a 6-well plate coated with 3.0 μg/mL iMatrix-511 silk at a density of 1.3 × 10 4 cells/well in the presence of 10 μM Y-27632 (Nacalai).

Establishment of the co-culture model
Neurons and astrocytes were induced separately from the same iPSC line.After removing DAPT from the neuronal assay on PID 18, the astrocytes were plated into the wells pre-plated with neurons at a ratio of 8:1 (neurons to astrocytes).After mixing neurons and astrocytes, cells were maintained using the neural differentiation medium.The cells were analyzed at 30 days after plating the astrocytes to the pre-plated neurons.

Immunocytochemistry (ICC)
Cells were washed with PBS twice, then fixed with a 4% paraformaldehyde (PFA) solution (Wako 163-20145) for 15 min at room temperature, and subsequently washed three times with PBS.The cells were then permeabilized and blocked using a solution of 5% fetal bovine serum (FBS) and 0.3% Triton X-100 in PBS.Primary antibodies were applied to the cells and incubated overnight at 4 °C.On the following day, the cells were washed three times with PBS and incubated with secondary antibodies for 1 h at room temperature.Then, cell nuclei were stained with Hoechst 33258 (Dojindo Laboratories) diluted in PBS for 15 min.After two washes with PBS and one wash with MilliQ water, the cells were mounted using PermaFluor (Thermo Fisher Scientific TA-030-FM).Images of the cells were captured using a fluorescence microscope (BZ-X810; Keyence, Osaka, Japan and IX73; Olympus, Tokyo, Japan) or a confocal microscope (Zeiss LSM700; ZEISS Group, Oberkochen, Germany).The primary and secondary antibodies, along with their dilution ratios, are summarized in Supplementary Table 1.The PSD-95 intensity was quantified using MetaMorph software (Molecular Devices) as previously described [23].

Reverse transcription-quantitative polymerase chain reaction (RT-qPCR)
RNA from the cultured cells was extracted using the RNeasy Mini Kit (Qiagen).Subsequently, cDNA was synthesized from the extracted RNA using the iScript cDNA synthesis kit (Bio-Rad).RT-qPCR was conducted utilizing 4 ng/μL cDNA, TB Green II, and Rox Dye II (TaKaRa), along with 20 μM of each primer.The amplification was carried out using a ViiA 7 Real-Time PCR System (Thermo Fisher Scientific 4453723) in accordance with the manufacturer's instructions.The detail of the primers used for qPCR was summarized in Supplementary Table 2.

Glutamate reuptake assay
The L-Glutamate concentration was measured by the L-Glutamate measurement kit, YAMASA NEO (YAMASA).iPSC-derived astrocytes and iPSCs were plated at 1 × 10 5 and 1 × 10 4 cells/well on iMatrix precoated 96-well plate respectively.iPSC-derived astrocytes were cultured with iPSC-derived astrocyte media, and iPSCs were cultured with AK02N media for 7 days before starting the L-Glutamate concentration assay.L-Glutamate was dissolved in AK02N media at 150 mg/L and filtered through the 0.22 μm membrane (Millex ® , Merck).For the experiment, iPSC-derived astrocytes and iPSCs were washed with PBS and replaced with AK02N media containing 150 mg/L L-Glutamate.L-Glutamate concentration in the media at 0, 1, 2, and 4 h after the experiment started was measured with the L-Glutamate measurement kit YAMASA NEO according to the manufacturer's protocol.The concentration of L-Glutamate at each time point was normalized by the value at 0 h.

Ca 2+ imaging
The co-culture model in a 96-well format at day 30 after mixing the cells was used for the Ca 2+ imaging assay.Cells were incubated with 100 μL of neuronal induction medium, 2 μL of 500 μM Fluo-8 indicator (AAT Bioquest 21083), and 0.8 μL of 250 μM probenecid water (ThermoFisher Scientific P36400) at 37 °C for 15 min before imaging using an Olympus IX83 microscope.The cells were imaged 161 frames in a duration of 2 min and analyzed using the MetaMorph ® Microscopy Automation and Image Analysis Software.Changes in Ca 2+ oscillations of neurons in the co-culture model after the treatment with 20 μM inhibitors of Excitatory Amino Acid Transporter (EAAT) 1,2 (TFB-TBOA 2532) were measured to assess the function of tripartite synapse structure.

Electron microscopy (EM)
Samples were prepared for EM observation as previously described [24,25].Briefly, the cells were plated onto the 4-well plastic chamber slide (Lab-Tek ® Chamber Slide ™ System 177437) with a density of neurons of 40 × 10 4 cells and astrocytes of 5 × 10 4 cells per well.Cells were fixed with 2.5% Glutaraldehyde (TAAB Laboratories, England, UK) in 0.1 M Cacodylate Buffer for 24 h at 4 °C.On the following day, after washing the cells using 0.1 M Cacodylate Buffer 3 times for 5 min each, the cells were fixed with 1% OsO 4 (TAAB Laboratories, England, UK) in Cacodylate Buffer for 90 min at 4 °C.Cells were dehydrated in 50% Ethanol 2 times for 5 min each and stained with Uranium for 20 min at 4 °C.Then, cells were dehydrated in a series of 70-100% Ethanol and 100% Epon (27.0 g MNA, 51.3 g EPOK-812, 21.9 g DDSA, and 1.1 ml DMP-30 for 100 g Epon) (Oken-shoji Co. Ltd.) for 48 h at 4 °C.After incubating the cells in 100% Epon for 3 h at room temperature, cells were embedded and polymerized in 100% fresh Epon for 72 h at 60 °C.After polymerization, the cells were dissected from the slide glass at 100 °C, fixed on the sectioning stage for 24 h at 60 °C, and dried up in the desiccator for 24 h at room temperature.Ultrathin sections of the cells with a thickness of 80 nm were prepared using an ultramicrotome (Leica UC7; Leica Biosystems, Wetzlar, Germany) with a diamond knife (Ultra, DiATOME).After 24 h of incubation at room temperature, the ultrathin sections were stained with uranyl acetate and lead citrate respectively at room temperature.The sections were examined and imaged using a Transmission Electron Microscope (TEM) (JEM-1400plus; JEOL, Tokyo, Japan) at 100 keV.
For pre-embedding immuno-EM (iEM), cells in the 4-well plastic chamber slide (Lab-Tek ® Chamber Slide ™ System 177437) were fixed with 4% PFA in 0.1 M Phosphate Buffer (PB, pH 7.4) (Muto Pure Chemicals) for 30 min at 4 °C, washed three times for 3 min each in PB, and blocked in a blocking solution (5% Block Ace (KAC) and 0.01% saponin in 0.1 M PB) for 1 h at room temperature.Cells were then incubated with a mouse anti-S100β antibody (1:250) (Supplementary Table 1) in a blocking solution for 72 h at 4 °C.After washing washed ten times for 10 min each in 0.1 M PB with 0.005% saponin, cells were incubated with Alexa Fluor 488 / FluoroNanogoldconjugated goat anti-mouse secondary antibody (1:100) (Thermo Fisher Scientific) and Hoechst 33258 (1:1000) (Sigma-Aldrich) for 24 h at 4 °C.After washing washed ten times for 10 min each in 0.1 M PB with 0.005% saponin, areas of interest were identified utilizing Alexa Fluor 488 signals under a Zeiss Axio Imager M1 Fluorescence Microscope (ZEISS Group, Oberkochen, Germany).The sections were fixed with 2.5% Glutaraldehyde in 0.1 M PB for 1 h at 4 °C, washed with 0.1 M PB three times for 5 min each, washed with 50 mM HEPES (pH 5.8) (Sigma-Aldrich) three times for 10 min each, and washed with diluted water for 1 min.Next, the nanogold signals were enhanced using silver enhancement solution for 2 min and 30 s at 25 °C.After washing with diluted water five times for 1 min each, the cells were fixed with 1% osmium tetroxide (TAAB Laboratories) for 1.5 h at 4 °C, washed with distilled water, and dehydrated with 50% Ethanol.The cells were en bloc stained with 2% Uranium in 50% Ethanol for 20 min at 4 °C.After, samples were dehydrated in a series of 70-100% Ethanol, incubated in 100% Epon (Oken-shoji Co. Ltd.) for 30 min at room temperature followed by several hours at 4 °C, and polymerized in 100% Epon for 72 h at 60 °C.The resin blocks of the samples were trimmed and ultrathin sectioned (thickness of 80 nm) with an ultramicrotome (Leica UC7; Leica Biosystems, Wetzlar, Germany).After 24 h of incubation at room temperature, the sections were stained with uranyl acetate and lead citrate for 10 min.Samples were imaged using TEM (JEM-1400plus; JEOL, Tokyo, Japan) at 100 keV.

Glutamate toxicity assay
The co-culture model and neuronal monoculture model at day 30 after mixing the cells were treated with 15 μg/ mL L-Glutamate (Wako) for one week.To investigate cell viability, the cells were treated with 1 μg/mL Calcein-AM (DOJINDO) in the culture media for 1 h, washed with PBS, and fixed with 4% paraformaldehyde for 16 h at 4 ℃.Then, fluorescent signals of Calcein-AM were detected by In Cell Analyzer 6000 (Cytiva, Marlborough, United States).Calcein-positive and MAP2-positive neurons were counted by In Cell Developer (Cytiva, Marlborough, United States).The surviving neuron ratios were calculated as Calceinpositive and MAP2-positive neurons/Hoechst.

DNA sequencing and genotyping
Genomic DNA from the cells was extracted using the DNeasy Blood and Tissue kit (Qiagen).The extracted DNA was subjected to PCR amplification using primers for the fragment of APP (Supplementary Table 2).The Proflex PCR system (Thermo Fisher Scientific) was used with the following cycling parameters: an initial denaturation at 95 °C for 2 min, followed by 35 cycles of denaturation at 95 °C for 30 s, annealing at 58 °C for 30 s, and extension at 72 °C for 1 min.Gel electrophoresis was performed on each PCR product using 125 mV for 30 min.The DNA band corresponding to the target product size was excised, and the DNA was purified using the QIAquick Gel Extraction Kit following the manufacturer's instructions.Purified DNA (20 ng) and 1 μL of each 10 μM primer were used for Sanger sequencing conducted by Eurofins Genomics K.K., Japan.

Enzyme-linked immunosorbent assay (ELISA)
ELISA was conducted using the conditioned media from both mono-cultured and co-cultured cells to quantify the levels of secreted amyloid-beta (Aβ) 1-42 and Aβ 1-40.The assay was conducted following the manufacturer's instructions with the ELISA kit (Wako 298-64601; Wako 296-64401).Briefly, the Standard Solution was prepared by diluting Aβ peptides with Standard Diluent at several concentrations.Then, 100 μL of the Standard Solution and the conditioned medium was applied to an antibodycoated plate, the plate was sealed and incubated at 4 °C overnight.On the next day, the Standard Solution and the conditioned medium were removed from the wells, and the wells were washed 4-5 times with a washing solution diluted in 1xPBS.Subsequently, 100 μL/well of the Horseradish peroxidase (HRP) antibody solution was added to the wells, and the plate was sealed and incubated at 4 °C (1 h for Aβ 1-42 and 2 h for Aβ 1-40 ).After the incubation, the HRP antibodies were removed, and the wells were washed 4-5 times with the washing solution.Then, 100 μL/well of 3,3' , 5,5;-tetramethylbenzidine (TMB) solution was added, and the plate was placed in a dark environment at room temperature for 30 min.Later, without removing the TMB solution, 100 μL/well of the Stop Solution was added to the wells.The absorbance of the conditioned media from each Aβ 1-42 and Aβ 1-40 plate was measured at 450 nm using the iMarkTM Microplate Reader (Bio-Rad).Subsequently, the Aβ 1-42 levels, Aβ 1-40 levels, and Aβ 1-42 / Aβ 1-40 ratios were calculated.

Statistical analysis
Statistical analysis was performed with GraphPad Prism (Version 9.4.1 (458); GraphPad Software, Boston, MA, USA).All data are presented as the mean ± SEM.The Unpaired t-test and Mann-Whitney test were used for comparison between the 2 groups depending on the normality assessed by the Shapiro-Wilk test.One-way ANOVA and Kruskal-Wallis test were used for group analysis upon the normality test by the Shapiro-Wilk test.Two-way ANOVA was used for the glutamate reuptake and Sholl analyses.Differences were considered significant at *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001.

Human iPSCs were induced into the functional neurons
To induce the functional neurons, the iPSCs were differentiated into neurons by the dual SMAD inhibition using small compounds via neural progenitor cells (NPCs) as previously described [19,20] (Fig. 1a).The qPCR of the induced NPCs at PID 15 confirmed the decreased expression of a pluripotent stem cell marker (OCT4) and increased expression of the cortical forebrain markers (FOXG1 and PAX6) compared to the original iPSCs (Fig. 1b).The induced neurons at PID 45 have normal morphology (Fig. 1c), and the neuronal markers (MAP2 and NeuN) and synaptic markers (NR2B and Syn1) were expressed (Fig. 1d, f ).The iPSCs were differentiated into neurons with high efficacy using our method (Fig. 1e).To assess the function of the induced neurons, Ca 2+ imaging using the Fluo-8 indicator was performed on the neurons at PID 45.In addition to the neuronal activities observed in the induced neurons, the treatment with 10 μM MK-801 (blocker of N-Methyl-D-aspartate (NMDA) receptor) resulted in decreased Ca 2+ oscillations (Fig. 1g, h).Thus, this confirmed the NMDA receptor-dependent activities of the iPSCs-derived neurons.

Human iPSCs were induced into functional astrocytes
iPSCs were differentiated into astrocytes via the formation of embryoid bodies (EBs) and neurospheres (NSs) according to the previous report [21] (Fig. 2a).This induction method allows the derivation of a relatively pure astrocyte population through the passages of the cells several times after the induction of NSs without exposure to serum.The qPCR results of the iPSC-derived astrocytes showed significantly increased expression levels of S100β, GFAP, and CD44 compared to the original iPSCs at mRNA levels (Fig. 2b).The differentiated cells showed characteristic morphology mimicking astrocyte morphology (Fig. 2c).The immunocytochemistry also showed that the induced astrocytes expressed GFAP, EAAT1, EAAT2, S100β, and CD44 (Fig. 2d), which are astrocytic markers.S100β and CD44 were expressed in high efficacy using our method (Fig. 2e).As a functional assay, the L-Glutamate uptake assay was conducted.Conditioned media of iPSC-derived astrocytes and iPSCs were sampled time-dependently.The concentration of L-Glutamate in the conditioned media of the induced astrocytes decreased in a time-dependent manner, while there were no changes in the concentration of L-Glutamate in the conditioned media of the iPSCs even after 4 h (Fig. 2f ).These results suggested that the iPSCs were successfully differentiated into functional astrocytes.

Co-culturing of neurons and astrocytes induced cellular changes resembling the astrocytes in vivo
To enable the mutual interaction between neurons and astrocytes, we aimed to establish a novel co-culture system of the iPSC-derived neurons and astrocytes.The neurons and astrocytes were individually induced from the identical iPSCs as described above.Then, the iPSCderived astrocytes were plated onto the wells pre-plated with iPSC-derived neurons in the neural differentiation medium to establish the co-culture model (Fig. 3a).Bright-field images of the co-culture model revealed the co-existence of two different types of cells even 30 days after co-culturing neurons and astrocytes.The induced astrocytes have larger soma, long and thick cell processes (in the yellow arrow of Fig. 3b), while the induced neurons have smaller cell soma (in the white arrow of Fig. 3b).Immunocytochemistry images of the co-culture model revealed the increased number of cell processes of the induced astrocytes (GFAP-positive cell) (Fig. 3c).
To further confirm the morphological changes in astrocytes, we carried out a Sholl analysis of the cells.Sholl analysis revealed the increased complexity, as measured by the increase in the number of intersections at a distance from the center of the cells, of the induced astrocytes in the co-culture model (Fig. 3d, e).Furthermore, to examine if the induced astrocytes would increase their activation when co-culturing with the neurons, we stained the cells with GFAP.As a result, immunocytochemistry showed that astrocytes in the co-culture model had a higher ratio of GFAP/S100β compared to astrocytes in the monoculture model (Fig. 3f, g).These results suggested that the co-culture with neurons could enhance morphological changes and activation of the iPSC-derived astrocytes.

Neurons in the co-culture model were more mature compared to neurons in the monoculture model
To investigate changes in neurons, we compared the activity and the morphology of neurons in the co-culture model with neurons in the monoculture model.First, Ca 2+ imaging using the Fluo-8 indicator revealed increased neuronal activity in the neurons co-cultured with astrocytes (Fig. 4a) (Supplementary Video 1).The immunocytochemistry of PSD-95, the post-synaptic marker, revealed a higher density of the post-synaptic compartment in the co-culture model compared to the monoculture model (Fig. 4b).In addition, quantification of the number of synapses on transmission electron microscopy (TEM) showed that there were more synapses in the co-culture model than the monoculture model of neurons (Fig. 4c).The TEM images of the synapse of the neurons also revealed the pre-synaptic compartments fulfilled with more presynaptic vesicles and a thicker post-synaptic density structure in the post-synaptic compartments of the neurons in the co-culture model (Fig. 4d, e).These results suggested the enhanced maturation of iPSC-derived neurons in the co-culture model compared to those in the monoculture model.
To further investigate the morphological changes of the neurons in the co-culture and monoculture, the induced neurons were labeled by introducing the pCl-dsRed plasmid to the cells and double-labeled with an anti-RFP antibody to supplement the intrinsic fluorescent signals of RFP (Fig. 4f ).Compared to the monoculture model, neurons in the co-culture model showed increased neurite branching and complexity (Fig. 4g).Together, these results indicated that both the function and the structure of the induced neurons were more mature in the coculture system with astrocytes compared to those of the neurons in the monoculture system.

The co-culture improves neuronal viability compared to the monoculture under a stressed condition
To investigate whether astrocytes would protect neurons from neuronal cell death induced by neurotoxins [26], we examined the glutamate toxicity assay (15 μg/mL L-Glutamate treatment for one week) followed by the staining with Calcein-AM.The surviving neuron ratios (Calceinpositive and MAP2-positive neurons/Hoechst) were not different between the monoculture model and the coculture model when not treated with excessive L-Glutamate.Although the monoculture model of neurons with excessive L-Glutamate treatment showed a lower surviving neuron ratio compared to the monoculture models without excessive amount of L-Glutamate, the surviving ratio of the neurons in the co-culture model with excessive L-Glutamate treatment was not changed (Fig. 5a, b).This indicated that astrocytes in the co-culture model not only enhanced neuronal maturation but also protected neurons from excessive amount of glutamate.

Functional tripartite synapse formation in the co-culture model of neurons and astrocytes
We examined whether the co-culture model of the induced neurons and astrocytes from the iPSCs recapitulate the tripartite synapse structure comprised the presynaptic and postsynaptic terminals of neurons and the attaching process of the astrocytes [27].First, the immunocytochemistry of pre-, post-synaptic markers (VGLUT2 and PSD95), and the astrocyte marker (GFAP) in the co-culture model visualized the close associations of them (Fig. 6a).Further imaging utilizing the immuno-electron microscope (iEM) revealed the neuronal synapse structure and the associated process of astrocyte stained with S100β (Fig. 6b).Treatment with 20 μM TFB-TBOA (Excitatory Amino Acid Transporter (EAAT) 1,2 inhibitor) to the co-culture model increased Ca 2+ oscillations (Spikes/Min) of the neurons (Fig. 6c-e) (Supplementary Video 2).Thus, this suggested functional tripartite synapse formation in our co-culture model.

Co-culturing developed AD-like phenotypes of astrocytes induced from the iPSCs with familial AD mutation (APP V717L )
To examine if our co-culture model can recapitulate the AD-like phenotypes, we co-cultured neurons and astrocytes derived from iPSCs bearing the familial AD (fAD) mutation (APP V717L ) [17,28].The iPSC line (APP2E26 line) from a female donor harboring the mutation of familial AD (APP V717L ) were obtained [17], and we confirmed a single nucleotide substitution (G > C) at APP Exon 17 (Fig. 7a) [29].We co-cultured the neurons and astrocytes from the fAD donor (APP V717L ) and confirmed them by immunostaining of the markers (Fig. 7b).ELISA measurement of Aβ 1-42 and Aβ 1-40 confirmed the increased ratio of Aβ 1-42 /Aβ 1-40 in the APP V717L co-culture model compared to the control co-culture model as previously reported [17,28] (Fig. 7c).The APP V717L iPSCs can be differentiated into both neurons and astrocytes with high efficacy (Supplementary Fig. 1a-d).The astrocytes differentiated from the APP V717L iPSCs also demonstrated the glutamate reuptake function.Then, there is no significant difference between the astrocyte of APP V717L and the control (Supplementary Fig. 1e).Furthermore, after co-culture, an increase in astrocytic complexity and an increase in neurite branching were compatible in the control and APP V717L .This indicated that both control and APP V717L neurons and astrocytes enhanced astrocytic and neuronal maturation, respectively, to a comparable extent (Supplementary Fig. 1f, g).
Since APP is mainly expressed in neurons but not in astrocytes, we hypothesized that our co-culture method developed AD-like phenotypes of astrocytes derived from fAD iPSCs that are hindered in monoculture.The astrocytes in the APP V717L co-culture model showed larger cell bodies and shorter cell processes, which resembled astrocytic hypertrophy found in brains of AD patients and other neurodegenerative diseases [30] (Fig. 8a).The quantification of the GFAP-positive astrocytes revealed that soma areas of astrocytes in the APP V717L were significantly larger compared to the astrocytes in the control in both monoculture model and co-culture model (Fig. 8b).In addition, the co-culture model of the APP V717L donor increased the GFAP/ S100β ratio compared to the co-culture model of a healthy control donor (Fig. 8c).However, the GFAP/ S100β ratio was not altered between the monoculture models of healthy control and APP V717L astrocytes alone (Fig. 8d).
To further investigate the mechanism of the development of astrocytic phenotypes, we performed the crossed co-culture of the control astrocytes with APP V717L neurons (Fig. 8e).The increase in the Aβ 1-42 /Aβ 1-40 ratio in the co-culture model of the control astrocytes and the APP V717L neurons compared to the co-culture model of the control astrocytes and the control neurons was confirmed (Fig. 8f ).The GFAP/S100β ratio was increased in the crossed co-culture model (Fig. 8g).In addition, there was an enlargement of the soma in some of the control astrocytes when crossed co-cultured with the APP V717L neurons (Fig. 8e).Thus, this indicated that an astrogliosis-like phenotype can be induced by secreted Aβ from neurons, which is hindered in astrocytic monocultures.These data suggested that our novel co-culture system could expose astrocytic phenotypes of AD with APP V717L mutation.

Discussion
This study demonstrated that the co-culture model of iPSC-derived neurons and astrocytes enhanced the maturation of both cell types.Because the maturation of the differentiated cells from iPSCs is often insufficient, the insufficient maturation has been one of the main problems in using iPSC-based models to study of neurological diseases [9].In our study, we observed significant changes in both neurons and astrocytes differentiated from the iPSCs after co-culture.The morphologies of astrocytes in the co-culture model were different from those in the monoculture model alone [20], reminiscent of those of astrocytes in vivo [31].The increase in the GFAP/S100β ratio in the co-culture model indicated the increased astrocyte activation [32].
For neurons, both the number of synapses and the number of presynaptic vesicles per synapse were increased in the co-culture model compared to the Fig. 3 Establishment of iPSC-derived neurons and astrocytes co-culture system and cellular changes of the iPSC-derived astrocytes in the co-culture system.a Schematic experimental paradigm of the co-culture using individually induced neurons and astrocytes from the same iPSCs.b Bright-field images of the co-culture model of the iPSC-derived neurons (white arrow) and astrocytes (yellow arrow) at 30 days after co-culturing of the cells.Scale bar; 100 μm.c Representative immunocytochemistry images of the iPSC-derived mono-culture model of astrocytes and co-culture model, at 30 days after co-culturing of the cells, stained with antibodies specific for S100β, GFAP, and MAP2.Scale bar; 100 μm.d Representative images of iPSC-derived astrocytes in the mono-and co-culture models.e Sholl analysis showing the complexity of the iPSC-derived astrocyte of the mono-culture model (n = 30) and co-culture model (n = 30) (three independent experiments).Bars, mean ± SEM. ****p < 0.0001 (Two-way ANOVA).f Immunocytochemistry images of iPSC-derived astrocytes stained with S100β and GFAP in the mono-and co-culture models.Scale bar; 100 μm.g Quantification and comparison of the GFAP/S100β ratios between the mono-culture model monoculture model.The increased number of presynaptic vesicles indicated an increased maturation of the neurons [33,34].Transcriptomic analyses of the co-culture model of neurons and astrocytes induced from the iPSCs-derived cortical progenitor pool indicated that astrocytes express extracellular proteins that interact with pre-and postsynaptic proteins and mediate maturation of neurons [35].These extracellular proteins include ITGAV, EFEMP1, and SIRPA, which were the proteins associated with synaptic development [36][37][38].In our co-culture model, iPSC-derived astrocytes may also promote neuronal maturation through similar mechanisms.Other underlying mechanisms may involve the secretion of neurotrophic factors such as glial cell line-derived neurotrophic factor (GDNF) directly from the co-cultured astrocytes [39].
Furthermore, since most of the disease models have been differentiated to only a single cell type of cells [reviewed in 6], many of the disease phenotypes due to the abnormal interaction of multiple cell types are still hindered when using a monoculture model.By utilizing the cellular model that includes mutual interactions between different cell types, analyses of the human cellular models in physiological environments would become possible and more disease phenotypes could be recapitulated.To date, there have been several reports of ADlike phenotypes that could be observed using in vitro 2D models generated from the iPSCs carrying fAD mutations on APP and PSEN.For example, as shown in Table 1, the iPSC-derived neurons could recapitulate the phenotypes associated with the secretion of Aβ, or pTau in some studies.However, the observed phenotypes are mostly limited to the neurons, with only a few reports on glial [7,8,17,[40][41][42][43][44][45][46][47][48].Fong et al. (2018) demonstrated that iPSC-derived astrocytes from the donors with APP Swe mutations showed impaired Aβ uptake, decreased APP secretion, and reduced lipid endocytosis [48].We also observed the astrocytic hypertrophy in the monoculture model of AD astrocytes because APPs are also expressed in astrocytes even at low levels [49,50].Another astrocytic phenotype, increased astrogliosis due to fAD mutations, was not observed in the iPSC-derived astrocytic monoculture models [48] (Fig. 8c, d).Therefore, the astrocytes co-cultured with neurons exhibited astrogliosis-like phenotypes possibly due to the exposure to high levels of Aβ secreted by neurons [51].
We also showed that our induced neurons and astrocytes interact with each other through the formation of the tripartite synapse structure.Since the tripartite synapse plays an important role in controlling neuronal functions in both physiological and pathological conditions [52], the human cellular model that can recapitulate this structure will be useful for the future studies of several neural diseases, especially those involving the clearance of excessive amounts of glutamate.For example, in AD, recent studies suggest that a soluble oligomeric Aβ causes excitotoxicity through several mechanisms including the stimulation of presynaptic glutamate release, inhibition of glutamate uptake in astrocytes, and alterations of glutamate receptors at the postsynaptic terminal (prolonged activation of S-NMDARs and AMPARs) [53][54][55][56].In some of the frontotemporal temporal dementia spectrum disorders associated with the Tau mutation A152T is associated with neuronal excitotoxicity due to excessive glutamate release from the presynaptic terminal of neurons [57].Stimulation of the glutamate uptake by the glutamate transporter (EAAT) on astrocytes alleviated the pathological effects [57].Our co-culture model, which recapitulates tripartite synapse structure related to these pathologies, will be a useful tool to study these pathogenesis and drug responses in human cells.
There are some previous studies that demonstrated the changes in cellular morphology and function after coculturing the neural cells.Some studies utilized murine cells from animal models.For example, Jones et al. (2012) showed that the embryonic rat hippocampal neurons suspended on the layer of rat astrocytes developed more synapses compared to the culture without astrocytes [58].Another study showed that the co-culturing of cerebellar neurons with astrocytes protected neurons from glutamate toxicity [59].Luchena et al. (2022) developed the triplet co-culture model of the murine rat neurons, astrocytes, and microglia.Similar to our study, the neurons in this triplet co-culture model developed longer neurite branches and expressed more post-synaptic markers, while the co-cultured astrocytes showed ramified morphology compared to the monocultured astrocytes [60].Furthermore, microglia in the triplet co-culture model express less pro-inflammatory markers and more antiinflammatory markers [60].In addition, co-culture of iPSC-derived neurons and rat astrocytes accelerated the maturation of iPSC-derived neurons [61].
The co-culture models of iPSC-derived neural cells have also been reported in recent studies.These models include the co-culture of iPSC-derived neurons and microglia [62], iPSC-derived neurons and astrocytes [63], iPSC-derived motor neurons and microglia [64], and triplet co-culture of neurons, astrocytes, and microglia [16,65].Cell maturation in these culture models accelerated compared to the monoculture models of each cell type alone, and the models have been used to study neurodegenerative diseases, including Parkinson's and Alzheimer's disease, specifically the inflammatory responses of microglia and the disease pathologies observed in neurons.
In our study, the co-culture model of the individually differentiated neurons and astrocytes newly confirmed the utility of co-culturing the iPSC-derived cells together.In particular, while Guttikonda et al. (2021) used dual SMAD inhibition when inducing neurons as in our study, the astrocytes were derived from the direct induction of human pluripotent stem cells (hPSCs) via transient expression of NFIA [66].In the triplet co-culture model of Bassil et al. (2021), human primary astrocytes were utilized.In our study, the astrocytes were differentiated from iPSCs via the formation of embryoid bodies and neurospheres, recapitulating the developmental process of astrocytes.Furthermore, the neurons and astrocytes were differentiated from the identical iPSCs prior to co-culture, which demonstrated the alternative method for establishing the co-culture model of human cells.In addition, the serum-free induction method allowed us to observe not only the enhanced maturation of neurons but also the transition of astrocytes from basal to pathological conditions by co-culture.
Finally, several limitations of our study remain.First, we used only one clone of patient-derived iPSCs and demonstrated the phenotypes of AD with the APP V717L mutation [17].Second, the control line used in this study is the non-AD 201B7 iPSC line [2].An iPSC line, in which the APP V717L mutation was corrected by genome editing, would provide a better comparison to the APP V717L mutation in terms of phenotypic analyses.Nevertheless, it is noteworthy that the mutual interactions of the iPSC-derived neural cells could enhance cellular maturation and exposed astrocytic phenotypes.The tripartite synapse structure formed in our culture model served as a useful platform for the studies of disease pathogenesis involving this structure.
Our study also provides a future perspective to include multiple cell types to study other pathological conditions: for example, the iPSC-derived microglia to assess the inflammatory responses [16,67,68], or the iPSCderived oligodendrocytes to assess the myelination [69,70].Further studies on models that include more cell types are also expected to recapitulate more precise interactions of the cells in a dish.

Conclusions
This present study demonstrated the establishment of the novel co-culture model of neurons and astrocytes that had been individually differentiated from the identical iPSCs.The co-culture system promoted the maturation of both cell types included in the model.Not only the maturation of individual cells, but also the interaction of these cells was confirmed by the formation of the functional tripartite synapse structure.Furthermore, the co-culture model was able to recapitulate astrocytic AD phenotypes including astrocytic hypertrophy and astrogliosis-like phenotype, which have not been reported in other studies using iPSC-based models before.This co-culture system is a powerful tool to study both physiological and pathological conditions of human neural cells.

Fig. 3 (
Fig. 3 Establishment of iPSC-derived neurons and astrocytes co-culture system and cellular changes of the iPSC-derived astrocytes in the co-culture system.a Schematic experimental paradigm of the co-culture using individually induced neurons and astrocytes from the same iPSCs.b Bright-field images of the co-culture model of the iPSC-derived neurons (white arrow) and astrocytes (yellow arrow) at 30 days after co-culturing of the cells.Scale bar; 100 μm.c Representative immunocytochemistry images of the iPSC-derived mono-culture model of astrocytes and co-culture model, at 30 days after co-culturing of the cells, stained with antibodies specific for S100β, GFAP, and MAP2.Scale bar; 100 μm.d Representative images of iPSC-derived astrocytes in the mono-and co-culture models.e Sholl analysis showing the complexity of the iPSC-derived astrocyte of the mono-culture model (n = 30) and co-culture model (n = 30) (three independent experiments).Bars, mean ± SEM. ****p < 0.0001 (Two-way ANOVA).f Immunocytochemistry images of iPSC-derived astrocytes stained with S100β and GFAP in the mono-and co-culture models.Scale bar; 100 μm.g Quantification and comparison of the GFAP/S100β ratios between the mono-culture model (n = 45) and the co-culture model (n = 45).****p < 0.0001 (Mann-Whitney test) (See figure on next page.)

Fig. 5
Fig. 5 Improving neuronal vulnerability against excessive L-Glutamate treatment by co-culturing with astrocytes.a Representative immunocytochemistry images of the co-culture model and neuronal mono-culture with or without excessive L-Glutamate treatment (15 μg/ mL) for one week stained with Calcein-AM reagent and an antibody specific for MAP2.Scale bar; 100 μm.b Quantification and comparison of surviving ratio of iPSC-derived neurons (Calcein-positive and MAP2-positive neurons/Hoechst) among the mono-culture model of neurons and the co-culture model with and without excessive L-Glutamate treatment (15 μg/mL) for one week (n = 50 each) (three independent experiments).Bars, mean ± SEM. ****p < 0.0001 (One-way ANOVA)

Fig. 6
Fig. 6 Tripartite synapse structure in the co-culture system and the cell-cell interaction between the induced neurons and astrocytes.a Immunocytochemistry image showing the neuronal synaptic markers (VGLUT2 and PSD95) and the astrocyte marker (GFAP).Scale bar; 10 μm.b Immuno-electron microscopy (iEM) image showing the neuronal synaptic terminals, and astrocytic process stained with S100β.Scale bar; 200 nm.c Evaluation of neuronal excitability measured by Ca imaging using Fluo-8 indicator in the co-culture model after treatment with 20 μM TFB-TBOA (Excitatory Amino Acid Transporter (EAAT) 1,2 inhibitor).d Representative waveforms of neuronal activity (Ca 2+ oscillations) of the co-culture model before and after treatment with 20 μM TFB-TBOA.e Quantification of neuronal firing (Spikes/Min and ΔF/F 0 ) of the co-culture model after treatment with 20 μM TFB-TBOA (n = 30) compared to the treatment with DMSO (n = 30) (three independent experiments).****p < 0.0001 (Mann-Whitney test)

Table 1
Summary table of the AD phenotypes evaluated using the iPSC-derived models of donors with familial AD mutations