Modeling early phenotypes of Parkinson’s disease by age-induced midbrain-striatum assembloids

iPSCs and NESCs

The iPSCs that were used in this study are described in Supplementary Table 1. The iPSCs were cultured in 6-well plates (Thermo Fisher Scientific, 140675) coated with Geltrex (Life Technologies, A1413302). For the first 24 h the cells were cultured in Essential 8 Basal medium (Thermo Fisher Scientific, A1517001) supplemented with 1% Penicillin/Streptomycin (Invitrogen, 15140122) and 10 µM ROCK Inhibitor (Ri) (Y-27632, Millipore, SCM075). After the 24 h, the cells were cultured in Essential 8 Basal Medium, with daily media changes. Confluent iPSCs (~70–90%) were split using UltraPure™ 0.5 M EDTA, pH 8.0 (Thermo Fisher Scientific, 15575020). Immunofluorescence staining was performed to confirm the pluripotent identity of the iPSCs (Supplementary Fig. 1a–d). NESCs were generated from iPSCs using a previously described protocol⁵⁸. Briefly, NESCs were derived from iPSCs through embryoid body formation and neuroepithelial expansion, facilitated by the small molecules CHIR99021 (CHIR, 3 μM) and purmorphamine (PMA, 0.5 μM). The NESCs were maintained in N2B27 medium on 6-well plates (Thermo Scientific) pre-coated with Geltrex. The N2B27 medium consisted of a 1:1 mixture of DMEM-F12 (Invitrogen) and Neurobasal (Invitrogen), supplemented with 1:200 N2 (Invitrogen), 1:100 B-27 without vitamin A (Invitrogen), 1% Glutamax (Thermo Fisher), and 1% penicillin/streptomycin (Invitrogen). For maintenance, the medium was freshly supplemented with 3 μM CHIR (Axon Medchem), 0.75 μM PMA (Enzo Life Science), and 150 μM ascorbic acid (Sigma). Media changes occurred every other day, and cells were routinely passaged at 80–90% confluence using Accutase (Sigma). Cells were cultured at 37 °C with 5% CO₂. The NESCs’ neural stem cell identity was confirmed with immunofluorescence staining (Supplementary Fig. 1e–g).

Striatum organoids

Striatum organoids (StrOs) were generated using an adapted protocol (C4) from¹⁵ (C3) after comparing it with two other conditions (RA and SR) in two time points of culture (day (D)35 and D50). iPSCs at ~70% confluency were used for the StrOs generation. Before the procedure of spheroids formation, the cells were treated overnight with 1% DMSO (Sigma-Aldrich, D2650) in Essential 8 basal medium. For the spheroids formation (D -1), the iPSCs were first incubated into Accutase at 37 °C for 5 min. The accutase was stopped using 5× DMEM-F12 (Thermo Fisher Scientific, 21331-046). The cells were then resuspended in Essential 8 medium containing 20 μM Ri and counted using the Countless cell counting chambers slides (Invitrogen, C10313). For condition C4, 10,000 cells were added per well in the BIOFLOAT™ 96-well plate U-bottom (faCellitate, F202003), centrifuged in 100 × g for 3 min and then incubated in the normal culture conditions of 37 °C with 5% CO₂. The spheroids were left intact for two days to form properly and at D1 the medium was exchanged with Essential 6 medium (Thermo Fisher Scientific, A1516401) supplemented with 10 μM Ri, 2.5 μM Dorsomorphin (Sigma-Aldrich, P5499) and 10 μM SB-431542 (Abcam, ab120163). The spheroids were cultured in the same medium until D5, with a reduction of the Ri concentration to 10 μΜ on D2, 5 μΜ at D3 until completely removed at D4. To start the differentiation process of the spheroids into StrOs, on D6 the medium was exchanged with a medium containing Neurobasal-A (Thermo Fisher Scientific, 10888022), 2% B-27 without Vitamin A (Thermo Fisher Scientific, 12587010), 1% Penicillin/Streptomycin (Invitrogen, 15140122), 1% GlutaMAX (Thermo Fisher Scientific, 35050061) and supplemented with 2.5 μM IWP-2 (Selleckchem, S7085) and 50 ng/ml Activin A (Thermo Fisher Scientific, PHC9561). From D9 to 17, the media was additionally supplemented with 100 nM SR11237 (Tocris, 3411). Until D17 the medium was exchanged daily. From D17 to 35, the media was changed to promote the neuronal differentiation and it was supplemented with 20 ng/ml BDNF (PeproTech, 450-02), 20 ng/ml NT-3 (Alomone labs, N-260), 200 μM AA (Sigma-Aldrich, A4544) and 100 μM cAMP (Biosynth, ND07996), with medium exchanges every 3 to 4 days. For condition C3, the same protocol as described in Miura and colleagues¹⁵ was followed, without the addition of cis-4,7,10,13,16,1 9-docosahexaenoic acid (DHA) at D22 of differentiation.

For the generation of StrOs from the RA and SR conditions, iPSCs at ~70% confluency were collected using accutase as described above, and 9000 cells were plated in each well of the BIOFLOAT™ 96-well plate U-bottom (D -2). The media used for plating had 80% DMEM F12 (Thermo Fisher Scientific, 21331046), 20% KOSR (Thermo Fisher Scientific, 10828028), 3% FBS (Invitrogen, 16140071), 1% GlutaMAX, 1% NEAA (Thermo Fisher Scientific, 11140-050) and 0.7% 2-Mercaptoethanol 50 mM (Thermo Fisher Scientific, 31350-010). This media was supplemented with 10 μM Ri and 40 ng/ml FGF-basic (PeproTech, 100-18B). After two days (D0 of culture) the media was exchanged with the Neural induction medium (NIM) containing 76.8% DMEM F12, 20% KOSR, 1% NEAA, 1% GlutaMAX, 1% Penicillin/Streptomycin and 0.2% 2-Mercaptoethanol 50 mM. From D0 to 2 the NIM was supplemented with 5 μΜ DM, 10 μM SB and 10 μM Ri, while on D2 Ri was removed. From D2 to D16 the medium was exchanged daily. On D4 and 5 the medium was additionally supplemented with 5 μΜ IWP-2. On D6, NIM was exchanged to the neural differentiation medium (NM) which contained 96% Neurobasal A medium, 2% B-27 without Vitamin A, 1% GlutaMAX and 1% Penicillin/Streptomycin. From D6 to 8, NM was supplemented with 20 ng/ml FGF-basic, 20 ng/ml EGF (PeproTech, AF-100-15) and 5 μΜ IWP-2. From D8 to 16, NM was supplemented with 20 ng/ml FGF-basic, 20 ng/ml EGF, 5 μΜ IWP-2, 50 nM SAG (Merck Millipore, 566660), 50 ng/ml Activin A and 1 mM for condition RA or 100 nM SR11237 for condition SR. On D17 the media was exchanged with a non-supplemented NM. From D18 to D35 or D50, NM was supplemented with 20 ng/ml BDNF and 20 ng/ml NT3, with media changes every 3 to 4 days.

Midbrain organoids

MOs were generated using NESCs as the starting population of cells. The protocol that was used is a slightly altered version of the one described in Monzel and colleagues⁷. At D0, 9000 NESCs were added per well in the BIOFLOAT™ 96-well plate U-bottom and cultured in maintenance medium for 2 days. At D3 the medium was exchanged with the differentiation medium containing 1 µM purmorphamine and at D8 with the final differentiation medium without purmorphamine. The organoids were kept in static conditions and in the 96-well plates U-bottom, until used for the assembloids generation (described below).

Assembloids

Assembloids of midbrain and striatum organoids were generated using MOs at D20 and striatum organoids at D35 of culture. Due to the different media composition of the organoids, an optimisation of the co-culture medium was needed. Four different media were tested. The Neural Medium (NM) was comprised of Neurobasal-A, 2% B-27 without Vitamin A, 1% Penicillin/Streptomycin and 1% GlutaMAX, while the Neural Medium Plus (NMpl) was supplemented with B-27 plus (Thermo Fisher Scientific, A3582801) instead of the B-27 without Vitamin A. In the third condition, the Neural Medium++ (NM++) was tested, which was the NM medium supplemented with 20 ng/ml BDNF (PeproTech, 450-02), 10 ng/ml GDNF (PeproTech, 450-10), 20 ng/ml NT-3 (Alomone labs, N-260), 200 μM AA (Sigma-Aldrich, A4544) and 100 μM cAMP (Biosynth, ND07996). The optimal assembloid co-culture condition for the assembloid model was consisted of the N2B27 medium (described by Monzel and colleagues⁷), which consists of DMEM F12 (Invitrogen)/Neurobasal (Invitrogen) 50:50 with 0.5% N2 supplement (Thermo Fisher Scientific, 17502001), 1% B-27 without Vitamin A, 1% GlutaMAX and 1% Penicillin/Streptomycin. The media was further supplemented with 20 ng/ml BDNF (PeproTech, 450-02), 10 ng/ml GDNF (PeproTech, 450-10), 20 ng/ml NT-3 (Alomone labs, N-260), 200 μM AA (Sigma-Aldrich, A4544) and 100 μM cAMP (Biosynth, ND07996). At D0 of the assembloids generation, each StrO was transferred in each well of the 96-well plate U-bottom that contained the MOs and the medium was exchanged to the assembloid co-culture medium. After 4 days, the two organoids were merged into an assembloid, and they were transferred in 24 well ultra-low attachment plates (Celltreat, 229524). Some of the assembloids were embedded in 30 µl Geltrex (Invitrogen, A1413302), as described before⁷. The assembloids were cultured at 37 °C, 5% CO₂ under static conditions.

For the Progerin overexpression induction, assembloids that were generated from the Progerin-overexpressing cell line were treated with 4 ng/μl doxycycline (Sigma-Aldrich, D9891). Generation and validation of iPSC line expressing Progerin under control of the Tet-ON system is described elsewhere.

ATP and LDH assay

Intracellular ATP in assembloids was measured using luminescence based CellTiter-Glo® 3D Cell Viability Assay (Promega, G9681). Three organoids per cell line and per condition were transferred each in one well of the imaging plate (PerkinElmer, 6055300). Fifty microliters of CellTiter-Glo® reagent were added to each well and the plate was incubated for 30 min on a shaker at room temperature (RT). Luminescence was measured using Cytation5 M cell imaging reader (RRID:SCR_019732). The experiment was repeated for three independent derivations (batches) at D30 assembloids. The mean signal of three assembloids of each cell line and condition was calculated and normalised to the mean size (area) of the assembloids. Brightfield images of random assembloids of the three batches taken with the ZEISS Axio Vert.A1 + Axiocam ICM1 microscope and the area of each assembloid was calculated with the ZEN (blue edition) software (RRID:SCR_013672). The mean area of all assembloids measured per line and conditions from each batch was used for the normalisation.

For determining the cytotoxicity in the assembloids, the LDH-Glo™ Cytotoxicity Assay (Promega, J2381) was used. In the day of assembloids collections (D30 of culture) media from three assembloids per line and condition from three batches was collected and snap frozen in liquid nitrogen. For the LDH assay, the snap frozen media was thawed on ice. Fifty microliters of media from each sample and 50 μl of the enzyme and substrate mix was pipetted in a well of the imaging plate (PerkinElmer, 6055300). The plate was briefly mixed and incubated for 1 h at RT avoiding exposure to light. Luminescence was measured using Cytation5 M cell imaging reader. Similar to the ATP assay, the mean signal of the assembloids was normalised to the mean of the assembloids’ area.

Western blotting – RIPA buffer

For Western blotting, four non-embedded assembloids or six to eight organoids were lysed using RIPA buffer (Abcam, ab156034) supplemented with cOmplete^TM Protease Inhibitor Cocktail (Roche, 11697498001) and Phosphatase Inhibitor Cocktail Set V (Merck Millipore, 524629). The samples were pipetted 10–20 times up and down until dissolved and were incubated on ice for 20 min. For DNA disruption, lysates were sonicated for 10 cycles (30 s on / 30 s off) using the Bioruptor Pico (Diagenode), followed by centrifugation at 4 °C for 20 min at 14,000 × g. The protein concentration was measured using the Pierce™ BCA Protein Assay Kit (Thermo Fisher Scientific, 23225). Samples were adjusted to the same concentration by appropriate dilution with RIPA buffer and boiled at 95 °C for 5 min in denaturing loading buffer. 2.5–10 μg of protein was loaded per sample for every Western blot. Protein separation was achieved using SDS polyacrylamide gel electrophoresis (Bolt™ 4–12% Bis-Tris Plus Gel, Thermo Fisher Scientific) and transferred onto a PVDF membrane using iBlot™ 2 Gel Transfer Device (Thermo Fisher Scientific). After transfer the membrane was dried for 10 min at 37 °C and subsequently activated with 100% Methanol for 30 s. The membranes were washed twice with PBS containing 0.02% Tween and they were blocked for 1 h at RT in 5% skimmed milk powder dissolved in PBS. After blocking, the membranes were washed quickly with PBS containing 0.02% Tween and were incubated overnight at 4 °C with the primary antibodies prepared in 5% BSA and 0.02% Tween in PBS (Supplementary Table 2). The next day, membranes were washed three times for 5 min with PBS containing 0.02% Tween and incubated with DyLight™ secondary antibodies at a dilution of 1:10,000 (anti-rabbit IgG (H + L) 800, Cell Signaling, 5151P or anti-mouse IgG (H + L) 680, Cell Signaling, 5470P) for 1 h. Membranes were revealed in the Odyssey® Fc 2800 Imaging System and exposure time was from 30 s to 4 min, depending on the primary antibody used. Western blots were analysed using ImageJ (RRID:SCR_003070) software.

Western blotting – nuclear and cytoplasmic fractionation

The protocol described by Abcam (https://www.abcam.com/protocols/nuclear-extraction-protocol-nuclear-fractionation-protocol) was used for the nuclear-cytoplasmic fractionation. At the last step of the protocol both nuclear and cytoplasmic samples were sonicated for 10 cycles (30 s on / 30 s off) using the Bioruptor Pico (Diagenode). For the preparation of the nuclear extraction and fractionation buffers the reagents used were HEPES (Sigma-Aldrich, H3375), KCl (AppliChem, 8059), MgCl₂ (Sigma-Aldrich, M8266), EDTA (Sigma-Aldrich, E9884), EGTA (Sigma-Aldrich, E3889), DTT (Thermo Fisher Scientific, R0861), cOmplete^TM Protease Inhibitor Cocktail and Phosphatase Inhibitor Cocktail Set V. The following Western blotting procedure is described in the previous section.

Flow cytometry

Three embedded assembloids per condition were used for GFP+, live cells measurement in BD LSRFortessa flow cytometer (RRID:SCR_019601). Geltrex embedded assembloids were first incubated at 37 °C for 40–50 min on shaker in 500 μl of papain solution containing 20 ml DMEM-F12, 36 mg Papain (Sigma-Aldrich, P4762), 8 mg EDTA (Sigma-Aldrich, E6758) and 8 mg L-Cystein (Sigma-Aldrich, C6852). To start the dissociation process, papain solution was replaced with 500 μl accutase and the assembloids were pipetted with the 1000 pipette, followed by a 10 min incubation shaking. After that, pipetting with the 200 μl pipette and incubation cycles were continued until the complete dissociation of the assembloids. For accutase and papain inhibition, 500 μl papain inhibitor solution containing 5 mg/ml BSA (Carl Roth, 8076.4) and 5 mg/ml Trypsin inhibitor (Sigma-Aldrich/Roche, 10109878001) in PBS was added. After transferring the total volume in 2 ml Eppendorf tube, the dissociated assembloids were centrifuged at 500 × g for 5 min. Supernatant was discarded and the pellet was washed once with PBS. The pellet was resuspended in 300 μl DMEM (Thermo Fisher Scientific, A14430-01) containing 1:1000 concentration live-dead stain Zombie NIR (Biolegend, 423106), followed by incubation at 37 °C for up to 30 min. Cells were then centrifuged at 500 × g for 3 min and pellet was washed twice with PBS and centrifuged again with the same setting. After the final wash and centrifugation, the pellet was resuspended in DMEM. The samples were then run in Becton Dickinson LSRFortessa, with 10,000 events acquisition of GFP+, live-cells. Each sample was run in two technical replicates. The data were analysed using the FlowJo software (v.10.7.2, RRID:SCR_008520).

Rabies virus based retrograde monosynaptic tracing

Lentiviral vector and rabies virus vector productions

The construct pBOB-synP-HTB [gift from Edward Callaway & Liqun Luo (Addgene plasmid # 30195 ; http://n2t.net/addgene:30195 ; RRID:Addgene_30195)]⁵⁹ was used for the production of the replication-deficient LV-GP-TVA-GFP lentiviral vector. High-titer preparations of lentiviral particles were produced, as previously described⁶⁰. The titer of the preparation used was 3 × 10⁸ IU/mL.

For the production of RBV-ΔG-EnvA-RFP rabies viral particles, we followed stages III-VI of the previously established protocol for the amplification, pseudotyping, and concentration of the virus⁶¹. Titer was 4 × 10⁷ IU/mL.

Lentiviral vector and rabies virus vector transduction of assembloids

To assess the connectivity through active synapses between the midbrain and striatum neurons in the assembloid model, striatum organoids at D35 of culture were transduced by adding concentrated LV-GP-TVA-GFP viral particles at 1:500 dilution, in the culture medium. After 7 days the medium containing the lentiviral vector was discarded and the organoids were washed twice with fresh medium. The LV-transduced StrOs were then merged with MOs and the assembloids were infected by the addition of the concentrated RBV-ΔG-EnvA-RFP rabies viral particles at 1:500 dilution in the culture medium. Control experiments, with single infection of assembloids with only one of the vectors at a time, were performed to confirm the specificity of the signal. After 7 days the media was changed and the assembloids were cultured for up to D30. Seventy micrometres-thick sections were either imaged directly for the observation of the RFP and GFP signal from the viral infections or were immunostained using a tyrosine hydroxylase (TH) antibody to assess the colocalization between TH and RFP.

β-galactosidase

Seventy micrometres sections from assembloids were used in the β-galactosidase staining with the Senescence Detection Kit (Abcam, ab65351). One section from two or three assembloids per condition from four batches were used. Images were taken using the colour camera setting with 4× objectives in the Olympus IX83 microscope (RRID:SCR_020344). β-galactosidase positive areas were quantified using ImageJ. Positive areas in each section were summed and normalised to the total area of the section in the image. The normalised value was multiplied by 100 for calculating the % of positive β-galactosidase areas in the image.

Immunofluorescence staining

iPSCs and NESCs

The procedure that was used for the immunofluorescence staining characterisation of the iPSCs is detailed described by Gomez-Giro and colleagues⁶². iPSCs and NESCs were cultured on Geltrex-coated 96-well imaging plates (PerkinElmer, 6055300) until they reached approximately 70% confluency. Next, cells were fixed for 15 min at RT with 4% Paraformaldehyde (PFA), washed 3× for 5 min with PBS and permeabilized with 0.3% Triton X-100 in PBS for 15 min at RT. After permeabilization, cells were washed 3× for 5 min with PBS and then blocked with 10% fetal bovine serum (FBS) in PBS for 1 h at RT. Primary antibodies (Supplementary Table 3) were diluted in 3% FBS in PBS and the cells were incubated overnight at 4 °C. Then cells were washed 3× for 5 min with PBS and incubated for 1 h at RT with secondary antibodies (Supplementary Table 3) and Hoechst 33342 (Invitrogen, 62249). After 3 washes for 5 min with PBS, cells were kept in 0.1% Sodium Azide (NaAz) in PBS until imaging with confocal microscopy.

Organoid and assembloid sections

Assembloids and organoids were fixed with 4% PFA overnight at 4 °C, and then washed with PBS three times for 15 min at RT. At least three organoids/assembloids per line and time point were embedded in 3% low-melting point agarose (Biozym, 840100). 70 μm sections were obtained using the vibratome (Leica VT1000s, RRID:SCR_016495). The sections were permeabilized for 30 min in 0.5% Triton X-100 and blocked for 2 h with blocking buffer containing 2.5% normal goat serum, 2.5% BSA, 0.01% Triton X-100 and 0.1% sodium azide in PBS at RT. Sections were incubated with the primary antibodies (Supplementary Table 3) diluted in blocking buffer for 48–72 h at 4 °C. The sections were washed with 0.01% Triton X-100 for 5 min three times and then incubated with the secondary antibodies (Supplementary Table 3) and Hoechst at 1:1000 dilution for 2 h at RT. The sections were then washed again with 0.01% Triton X-100 for 5 min three times at RT. After the last wash, the sections were kept in MilliQ water and mounted on slides using Fluoromount-G mounting medium as described by Nickels and colleagues¹⁹.

Microscopy

For high-content image analysis, one 70 μm section from three organoids/assembloids of each condition from at least three batches was acquired using the Yokogawa CV8000 high-content screening microscope (RRID:SCR_023270) with a 20X/0.75 numerical aperture (NA) objective.

For qualitative analysis, images were acquired using a confocal laser scanning microscope (Zeiss LSM 710, RRID:SCR_018063) with the 20X/0.8 NA, 40X/1.3 NA or 63X/1.4 NA objective.

Light sheet fluorescence expansion microscopy

Sample preparation

Following Rodriguez-Gatica and colleagues⁶³ expansion protocol for organoids, whole-assembloids were prepared using a methylacrylic acid-NHS linker, washed, and incubated in a monomer solution. Gelling involved 4-hydroxy-TEMPO, TEMED, and ammonium persulfate, with samples placed on ice to prevent premature polymerisation. After gelling, samples underwent digestion, were washed, and stored in PBS, expanding to 1.5 times their original size. Autofluorescent protein preservation facilitated specific immunolabeling (e.g., DANs) and post-digestion nuclear staining with Hoechst 33342 at a concentration of 2.5 μg/ml.

Light sheet microscopy

Post-expansion, whole-assembloids were imaged using light sheet microscopy for meso- and microscopic scale analysis. The Blaze microscope (LaVision-Miltenyi Biotec) was used for whole-assembloid imaging, and a custom-built setup was employed for higher-resolution imaging of selected ROIs. Samples were affixed to coverslips with poly-L-lysine and placed in an imaging chamber filled with PBS solution. For mesoscopic imaging with the Blaze microscope, a single illumination arm with a light-sheet thickness of 6 µm and a LaVision-Miltenyi 4X/0.35 NA, WD 15 mm WI objective lens captured images in a mosaic pattern due to sample size. A 12× (LaVision-Miltenyi 12X/0.53 NA WD 8.5 mm WI) objective was used as an intermediate step for unclear projections. Microscopic imaging employed a high-resolution Nikon CFI75 25x/1.1 NA WI objective, achieving resolutions suitable for subcellular structural characterisation. Complete assembloid imaging at this resolution is often impractical due to data volume (about 500 GB per 1 mm³ per channel), as noted by Rodriguez-Gatica and colleagues⁶³. Therefore, specific ROIs identified in the mesoscale data were selected for detailed microscopic analysis.

Data processing

3D image stacks were processed using MATLAB scripts and spatially deconvolved with Huygens software (Professional version 22.04, Scientific Volume Imaging, The Netherlands). Stitching multiple datasets for 3D representations was conducted in FIJI⁶⁴ using a two-step approach to manage large data sizes. Visualisation was achieved with Imaris (Version 10.0.1, Bitplane Inc.). All processing was performed on a HIVE workstation equipped with dual Intel Xeon Gold 6252 CPUs and an Nvidia RTX A4000 GPU.

Image analysis

Images obtained from the Yokogawa microscope were processed and analyzed in MATLAB (2021a, Mathworks, RRID:SCR_001622) using a previously described image analysis pipeline^34,65.

Electrochemical detection of catecholamines in neuronal tissues

Commercially available Nafion-coated carbon fiber microelectrodes (World Precision Instruments, CF10–50) were employed to measure the presence of catecholaminergic neurotransmitters (e.g., dopamine) within the assembloids. Amperometric measurements were performed according to the manufacturer’s instructions using a potentiostat (Bio-Logic, VMP3) equipped with a low current module (Bio-Logic) and a custom-made platform (Supplementary Fig. 11d). Electrodes were activated by applying a potential of 1.2 V and simultaneous exposure to a 150 mM NaCl solution with a pH of 9.5. All measurements were conducted at a potential of 0.65 V. Prior to the measurements, assembloids were washed three times using PBS. Assembloids were subsequently placed onto the custom-made platform, covered in 100 µl of PBS before electrodes were carefully inserted into the tissue using a laboratory jack. Measurements on StrOs alone and StrOs within D30 assembloids were conducted by introducing the electrode 1 and 2 times into the centre of the neuronal tissue. For the StrOs cultured in pre-used MOs media, the media was in contact with the MOs for 24 h. Then it was transferred into the wells containing StrOs. StrOs were cultured in the pre-used MO media for 48 h prior to the measurements. Due to the occasional occurrence of necrotic cores within assembloids electrochemical measurements on D60 assembloids were performed by introducing the electrodes at the border of the StrO. Measurements were conducted at five different locations within the StrO of the assembloid. To exclude any cross-contaminations, electrodes were carefully washed, and background signals were recorded in between the measurement of each assembloid. Measurement results were averaged and unless stated otherwise data were background subtracted.

Quantitative PCR

The RNeasy Mini Kit (Qiagen, 74106) was used for the total RNA extraction from striatum organoids. The RNA concentration was measured using the Nanodrop 2000c Spectrophotometer (Thermo Fisher Scientific, RRID:SCR_020309). The High-Capacity RNA-to-cDNA^TM Kit (Thermo Fisher Scientific, 4387406) was used for the cDNA synthesis. For the quantitative PCR reaction, the Maxima SYBR Green qPCR Master Mix (Thermo Fisher Scientific, K0221) was used with the primers listed in Supplementary Table 4. The Aria Mx Real-Time PCR system (Agilent) was used and the data were extracted from the AriaMx PC software. For each qPCR reaction, n = 3–5, where n is one sample of 6–8 pooled organoids, from 3–5 batches.

RNA sequencing

RNA was extracted from the assembloids using the RNeasy Mini Kit. Four assembloids were used per condition and from three batches. The RNA samples were shipped with dry ice to Novogene in UK for the RNA sequencing experiment and bioinformatic analysis.

In the process of obtaining clean reads, reads containing adapters, higher than 10% undetermined bases and low quality (Qscore of over 50% bases of the read is <=5) were removed. Poly-T oligo-attached magnetic beads were used to purify the messenger RNA from the total RNA. After fragmentation, the first strand cDNA was synthesised using random hexamer primers, followed by the second strand cDNA synthesis using dUTP for directional library. To quality control the library, Qubit and real time PCR were used for quantification, while for the size distribution detection bioanalyzer was used. The quantified libraries were combined and sequenced on Illumina platforms, taking into consideration the optimal library concentration and desired data volume. The clustering of the index coded samples was performed according to the manufacturer’s instructions. Following the generation of clusters, the library preparations underwent sequencing using an Illumina platform, resulting in the generation of paired-end reads.

RNA sequencing data analysis

Raw data (raw reads) were processed in fastq format utilizing the fastp software (RRID:SCR_016962). This step involved removing reads that contained adapters, reads with poly-N sequences, and low-quality reads from the raw data, resulting in obtaining clean data (clean reads). Additionally, metrics such as Q20, Q30 and GC content were calculated for the clean data. The reference genome (hg38) and gene model annotation files were directly downloaded from the genome website. To enable alignment of the paired-end clean reads, an index of the reference genome was constructed using Hisat2 v2.0.5 (RRID:SCR_015530). Subsequently, Hisat2 v2.0.5 was employed as the mapping tool of choice. To determine the number of reads mapped to each gene, featureCounts v1.5.0-p3 (RRID:SCR_012919) was employed. Subsequently, the Fragments Per Kilobase of transcript sequence per Millions (FPKM) base pairs sequenced for each gene was calculated based on the gene’s length and the count of reads mapped to it. Differential expression analysis was conducted on two conditions/groups using the DESeq2 R package (version 1.20.0, RRID: SCR_015687). The resulting P-values were adjusted using the Benjamini and Hochberg’s method to control the false discovery rate. Genes with an adjusted P-value of less than or equal to 0.05, as determined by DESeq2, were identified as differentially expressed. The read counts for each sequenced library were adjusted using the edgeR package (version 3.22.5, RRID:SCR_012802) by applying a scaling normalisation factor. The differential expression analysis of the two conditions was carried out using the edgeR. The P-values were adjusted using the Benjamini and Hochberg’s method. A corrected P-value threshold of 0.05 and an absolute fold change of 2 were set to determine significantly differential expression. The clusterProfiler (RRID:SCR_016884) R package was used for Gene Ontology (GO) (RRID:SCR_002811) enrichment analysis on the DEGs, with gene length bias being corrected. GO terms with a corrected P-value < 0.05 were deemed significantly enriched by the DEGs. For the analysis of KEGG (RRID:SCR_012773) pathways, the clusterProfiler R package was employed. The statistical enrichment of DEGs in KEGG pathways was assessed.

Single nuclei RNA sequencing

Samples processing

Ten MOs, 10 StrOs and 4 assembloids per batch form 2 batches, generated from the same line (201, Supplementary Table 1) were used. The culture time point was D30 for assembloids, D50 for MOs and D65 for StrOs. MOs and StrOs that were cultured separately after D20 and D35, were cultured in the optimised co-culture medium. Samples from the two batches were pulled together for nuclei extraction and sequencing by Singleron.

Data analysis

Reads were mapped to Homo_sapiens_ensembl_92 genome, and 10× matrices were generated for each sample. To perform the downstream analysis, we used R studio (23.06.1 + 504 version, RRID:SCR_000432) and R 4.2.2 version (RRID:SCR_001905). Using the Seurat package (version 4.3.0, RRID:SCR_016341), Seurat objects were created for each sample and quality control filtering was performed. For all datasets, cells with >5% mitochondrial genes were filtered out. Additional filtering for cell debris and doublets was performed for each dataset. In MOs data cells with <100 and >1000 genes, in StrOs cells with <100 and >1800 genes and in assembloids cells with <100 and >4000 genes were filtered out. Similar to another study, after filtering, ribosomal and mitochondrial genes were removed from all datasets, as they are considered contamination in the single nuclei RNA sequencing experiments⁶⁶. After filtering, using the standard Seurat workflow, we analyzed each dataset separately, for identifying clusters specific to each model. LogNormalisation was performed in each dataset, followed by the identification of the 2000 most variable genes (FindVariableFeatures function). Next, data were scaled with the ScaleData function and linear dimensionality reduction was performed with the RunPCA function. Ten PCs were used for the MO dataset clustering and 15 PCs for the StrO and assembloid datasets. Clusters were identified using the FindNeighbors and FindClusters functions, at 0.5 resolution in all datasets. Determination of cell type identity in each cluster was performed with the evaluation of specific cellular markers expression using the DotPlot visualization method. For identifying differentially expressed genes (DEGs) between assembloids and MOs or StrOs, integration analysis was performed²⁹. After integrating the data, DEG lists using the FindMarkers function were computed, defining the assembloid dataset as ident.1 and the MO or StrO dataset as ident.2. The DEG lists from both comparisons were imported in the Metacore (Clarivate, 2023) online software, were enrichment analysis with FDR threshold >0.25 and adjusted P.value < 0.05 was performed. Enriched pathways from the “Process Networks” category were extracted. Genes related to the enrichment of the “Development_Neurogenesis_Axonal guidance” pathway were exported, and their expression pattern (up or downregulation) was evaluated. DEGs were also used to assess the expression pattern of genes related to neuronal maturity.

Microelectrode array

Basal activity measurement in assembloids

Non-embedded assembloids were used for the electrophysiological analysis using the Axion Microelectrode array (MEA) system. Forty eight-well MEA plates (Axion, M768-tMEA-48B-5) were first coated with 0.1 mg/ml poly-D-lysine (Sigma-Aldrich, P7886) and incubated at 37 °C, 5% CO₂ overnight, followed by 1 h incubation with 1 mg/ml laminin (Sigma-Aldrich, L2020). Laminin coating was removed, and the plates were washed twice with sterile PBS (Thermo Fisher Scientific, 14190250). Each assembloid was placed in the center of the well on the electrodes and after the media was carefully aspirated, it was left for 2 and 3 min to dry. Then, 15 μl of Geltrex was added on top of each assembloid and was left in the incubator for 5 min to polymerise. Five hundred microliters of fresh culture medium was added in each well and the plate was kept in the incubator (37 °C, 5% CO₂) under static conditions. Electrophysiological data were acquired with the Axion Maestro Multiwell 768-channel MEA System (Axion Biosystems) and the Axis software (Axon Biosystems,RRID:SCR_016308). For the analysis of the data the spike lists were exported from the Axis software and the data were processed with MATALB (2021a, Mathworks,RRID:SCR_001622). Plots of the MATLAB exported data were generated using R 4.2.2 vesrion.

Connectivity measurement in assembloids

The assembloid electrical activity was measured by 8 × 8 microelectrode arrays (MEAs) positioned at the bottom of a 12 multiwell. The MEA signals were measured at a sample rate of 12.5 kHz (M768-GLx 12-well plate, Axion Biosystems). A microfluidic system laying on top of the MEA was designed to physically separate the MO and the StrO, allowing their connection only through microtunnels (10 µm width) where neurites can grow along both directions (Fig. 4a).

MEA recordings were pre-processed using the AxIS Navigator 3.9.1 software (Axion Biosystems). The spike events were extracted for each electrode using the Peak Detection Adaptive Threshold method, setting an amplitude threshold of 6 standard deviations (Fig. 4b). The timestamps of the extracted spikes were exported to MATLAB (2019b, Mathworks) for further analysis (https://github.com/dlpigozzi/CoMEA).

The degree of connectivity between different MEA regions was assessed by quantifying the correlation between the recorded spike trains. The STTC⁶⁷ was calculated for each pair of electrodes using the open code repository MEA-NAP⁶⁸. The time interval Δt for the STTC computation was set at 10 ms to privilege the detection of mono-synaptic connections (Fig. 4c).

The cross-correlogram provides an estimate of the time delay and signal directionality between electrode pairs. A 0.5 ms time binning was utilized to plot the probability histogram of the time delay (Fig. 4d), where the histogram maximum corresponded to the (most probable) time t required by the signal to move between the two electrodes, e.g., from A to B. The propagation speed V_AB was computed as the ratio of the distance d_AB (in meters) between A and B and the selected time delay t_AB (Fig. 4d). Only maxima exceeding the histogram mean value by 5 standard deviations were considered as significant. An STTC cut-off of 0.3 was utilized as an additional criterion to accept the existence of a connection between the two electrodes. The signal propagation between A and B was graphically represented as a vector connecting the two electrodes with orientation dependent on the sign (positive or negative) of the time delay. The average signal directionality for a given electrode was computed as the vectorial sum of the vectors connecting it to the electrodes having both significant cross-correlation maximum and STCC (Fig. 4e).

The mean directionality of a given assembloid was computed as the vectorial sum of its electrode vectors. During the assembloid development, if the vertical component (y) of the resulting vector reached a value ≥2 (i.e., the distance between 2 electrodes), the connection between the MO and the StrO was considered as established (Fig. 4f).

For the 92% of the assembloids establishing connection, we calculated the mean vertical (y) and lateral (x) components of their directionality vectors. The vertical component can result positive or negative, depending on the overall direction of the signal from the MO to the StrO or from the StrO to the MO, respectively (Fig. 4g). Instead, the lateral component was calculated considering its absolute value (|x|) without distinction between left and right side of the MEA (Fig. 4h).

Statistics and reproducibility

Data were analysed using GraphPad Prism 9.0.0 or R studio (23.06.1 + 504 version) with R 4.2.2 version. Normality test was performed using the Shapiro test. If not stated otherwise, outlier removal was performed using the ROUT method Q 1% in GraphPad or the Inter-Quartile Range proximity rule in R and data were batch normalised to the mean value of each batch. For not normally distributed data, two-sided Wilcoxon test or Kruskal–Wallis with Dunn’s multiple comparison test and Benjamini–Hochberg correction was implemented. For normally distributed data, Welch’s t-test or one-way ANOVA with Tukey’s multiple comparison test was performed. Significant P value is represented with asterisks in the order p < 0.05 *, p < 0.01 **, p < 0.001 ***, p < 0.0001 ****. Error bars represent mean ± SD.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

• SEO Powered Content & PR Distribution. Get Amplified Today.
• PlatoData.Network Vertical Generative Ai. Empower Yourself. Access Here.
• PlatoAiStream. Web3 Intelligence. Knowledge Amplified. Access Here.
• PlatoESG. Carbon, CleanTech, Energy, Environment, Solar, Waste Management. Access Here.
• PlatoHealth. Biotech and Clinical Trials Intelligence. Access Here.
• Source: https://www.nature.com/articles/s42003-024-07273-4