Skip to main content

Thank you for visiting You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

Defective AMPA-mediated synaptic transmission and morphology in human neurons with hemizygous SHANK3 deletion engrafted in mouse prefrontal cortex


Genetic abnormalities in synaptic proteins are common in individuals with autism; however, our understanding of the cellular and molecular mechanisms disrupted by these abnormalities is limited. SHANK3 is a postsynaptic scaffolding protein of excitatory synapses that has been found mutated or deleted in most patients with 22q13 deletion syndrome and about 2% of individuals with idiopathic autism and intellectual disability. Here, we generated CRISPR/Cas9-engineered human pluripotent stem cells (PSCs) with complete hemizygous SHANK3 deletion (SHANK3+/–), which is the most common genetic abnormality in patients, and investigated the synaptic and morphological properties of SHANK3-deficient PSC-derived cortical neurons engrafted in the mouse prefrontal cortex. We show that human PSC-derived neurons integrate into the mouse cortex by acquiring appropriate cortical layer identities and by receiving and sending anatomical projections from/to multiple different brain regions. We also demonstrate that SHANK3-deficient human neurons have reduced AMPA-, but not NMDA- or GABA-mediated synaptic transmission and exhibit impaired dendritic arbors and spines, as compared to isogenic control neurons co-engrafted in the same brain region. Together, this study reveals specific synaptic and morphological deficits caused by SHANK3 hemizygosity in human cortical neurons at different developmental stages under physiological conditions and validates the use of co-engrafted control and mutant human neurons as a new platform for studying connectivity deficits in genetic neurodevelopmental disorders associated with autism.


Accumulating evidence suggests that synaptic deficits may be responsible for the behavioral abnormalities in autism spectrum disorders (ASD) [1, 2]. Indeed, brain imaging studies have consistently identified altered functional connectivity in the brain of individuals with ASD [3], and well-powered sequencing studies have found multiple genetic abnormalities in various genes encoding synaptic proteins [4, 5]. However, it remains to be determined whether and how the disruption of specific synaptic genes influences the synaptic properties of human neurons under physiological conditions to alter brain connectivity.

SHANK3 is a postsynaptic scaffolding protein of excitatory synapses [6, 7]. It has been found mutated or deleted in about 0.5% of patients with idiopathic ASD, 2% of patients with idiopathic ASD and intellectual disability (ID) [8], and most patients with Phelan–McDermid syndrome (PMDS), a syndromic form of ASD caused by the microdeletion of chromosome 22q13.3 and is characterized by profound ID, developmental delay, severely impaired speech, and autistic behaviors [9, 10]. The most common genetic abnormality found in PMDS patients is a complete hemizygous SHANK3 deletion [11]. However, no overt deficits have been detected in mice with complete heterozygous Shank3 deletion, which were generated to mimic patient-related genetic abnormality [12]. Therefore, it remains unclear whether and how SHANK3 hemizygosity alters the synaptic properties of human neurons.

Synaptic deficits induced by partial homozygous Shank3 deletions or heterozygous mutations have been extensively investigated in animal models [12,13,14,15,16,17,18,19,20,21,22,23,24,25,26]. Although all studies reported deficits, there is some inconsistency between the specific synaptic and behavioral phenotypes observed in different studies, which is likely due to the loss/ectopic gain of particular isoforms of Shank3 and/or differences in the genetic backgrounds, cell types, and developmental stages examined in those studies [27]. As a result, it remains unclear which of the reported deficits are relevant to patients, as human and rodent neurons have very different developmental trajectories as well as genetic and epigenetic profiles [28].

More recent studies on SHANK3-deficient human neurons obtained from patient-induced pluripotent stem cells (iPSCs) and CRISPR/Cas9-engineered embryonic stem cells (ESCs) reported more severe deficits in excitatory synaptic transmission and new deficits in intrinsic excitability and morphological development [29,30,31,32]. However, these studies were conducted on stem cell-derived neurons obtained from PMDS patients with large deletions or rare point mutations or on engineered neurons with partial SHANK3 deletions. In addition, stem cell-derived human neurons were investigated in vitro under non-physiological conditions, which may have influenced the severity and manifestation of the deficits. Therefore, it remains unclear whether and how the synaptic deficits detected in human neurons in vitro are related to those in patients.

In the present study, we developed a novel approach to transplant human stem cell-derived cortical neurons into the mouse prefrontal cortex (PFC) and used a state-of-the-art trans-synaptic tracing technique and patch-clamp electrophysiology to investigate their synaptic and morphological properties. The results show that human neurons survive in the mouse PFC for many months post transplantation, acquire cortical laminar identities, and establish anatomical connections with mouse neurons in multiple brain regions. To characterize the synaptic deficits induced by SHANK3 hemizygosity, we generated a CRISPR/Cas9-engineered stem cell line with complete hemizygous SHANK3 deletion and co-transplanted both isogenic control (iCtrl) and SHANK3+/– human stem cell-derived cortical neurons into the mouse PFC. We found that SHANK3-deficient human neurons have impaired AMPA-, but not NMDA- or GABA-mediated synaptic transmission and deficient dendrites and dendritic spines. This study identifies specific deficits induced by SHANK3 hemizygosity in human cortical neurons and validates the use of the co-xenotransplantation model to study synaptic and connectivity deficits in neurodevelopmental disorders associated with ASD.

Materials and methods

Stem cell culture and neural differentiation

The stem cell lines used in this study (H9, EYQ2-20 [SHANK3+/–], 2242–5, and EP215 [SHANK3–/–]) were cultured in E8 medium (Life Technologies, Cat#A1517001) on 6-well plates (VWR, Cat#29442–036) coated with 1% Matrigel (Corning, Cat#354277) in humidified incubator with 37 °C/5% CO2. Stem cells were passaged when reached 70–80% confluency using Dispase (1 U/mL, STEMCELL Technologies, Cat#07923). During passaging a ROCK Inhibitor, Y-27632 (2–4 µg/mL, Hello Bio, Cat#HB2297) was added to the medium to improve cell survival.

Stem cells were differentiated using a modified version of the dual-SMAD inhibition protocol [33]. Briefly, stem cell colonies were re-plated on Matrigel-coated 6-well plates and allowed to reach approximately 95–100% confluency in E8 medium. On day 0, approximately half of the E8 medium was removed and neuroectodermal fate was induced by a 6–10-day-long incubation in neural differentiation medium (ND) containing of 1:1 mixture of N2 medium (DMEM/F-12 [Invitrogen, Cat#11330–032], 1% N2 Supplement [Thermo Fisher Scientific, Cat#17502048], 1% MEM–NEAA [Invitrogen, Cat#11140050], 2 mg/mL heparin [STEMCELL Technologies, Cat#07980], and 1% Pen/Strep) and B-27 medium (Neurobasal-A [Invitrogen, Cat#10888022], 2% B27 Supplement with vitamin A [Thermo Fisher Scientific, Cat#17504044], 1% GlutaMAX [Invitrogen, Cat#35050061], and 1% Pen/Strep) containing inhibitors of SMAD signaling, 10 µM SB431542 (Selleck Chemicals, Cat#S1067) and 4 µM dorsomorphin (Selleck Chemicals, Cat#S7306). On days 6–10, cells were dissociated into large clumps using dispase and transferred to Matrigel-coated plates in ND medium. The day after re-plating, half the medium was replaced with ND medium containing 20 ng/mL EGF (STEMCELL Technologies, Cat#780006.1) and 20 ng/mL FGF (STEMCELL Technologies, Cat#78003). On days 5–10, cells were cultured with EGF, but without FGF, to allow the formation of neural rosettes. On day 10, clusters of rosettes were dissociated using Dispase, transferred to wells precoated with Matrigel, and cultured in ND medium with EGF. On day 15, neural rosettes were dissociated to single cells with Accutase and propagated on Matrigel coated plates in ND medium without growth factors. On days 35–40, stem cell-derived neurons were transduced with different lentiviruses.

Generation of engineered stem cell lines

For generating stem cells with hemizygous and homozygous deletions, guide RNAs were designed using an online design tool ( to target different regions of the SHANK3 genomic locus. Each individual sgRNA was synthesized as a DNA fragment and cloned into the plasmid pSpCas9(BB)-2A-Puro (Addgene, Cat#48139) which co-expresses the gRNA with the SpCas9 protein in mammalian cells [34]. To assess the cutting efficiency, each of the plasmids expressing the sgRNA was transfected into HEK293 cells and genomic DNA of the cells was collected at 72 h post transfection. Each sgRNA-targeting region was amplified using PCR, and the amplicons were sequenced on the MiSeq instrument (Illumina, San Diego CA) to directly determine the frequency of mutations. Cutting efficiency was calculated as the highest deletion frequency of the base pair positions within the target region. Then, HEK293 cells were transfected with different combinations of gRNAs, two at a time, to test whether they can make any deletions. PCR was used to screen deletions between the two gRNA-binding regions and to select the best pairs of gRNAs.

Hemizygous SHANK3 deletion was introduced into the H9 human ESC line (WiCell Research Institute). Two sgRNA sequences, gRNA #1 (ATACCACAAGAAGGACGTCCGGG) and gRNA #4 (TTTCTCAGGGGTCCCCGGTGGGG), designed to flank the entire SHANK3 gene for creating a 69.5-kb deletion, were introduced using the same technique previously described. A total of 64 single colonies were screened for the deletion using PCR with two pairs of forward and reverse primers. Primers 1 (CAGACACTTCACGAAAGAAGC) and 4 (GGGAGCAAAAAGGAAACCTC) were amplified if deletion was successful. Primers 2 (GGAACTCAGCCCATGCCTTC) and 3 (TGCAGACGCAGAGCCTAACA) were amplified if no deletion occurred. Out of the 64 colonies, 4 were found to have heterozygous SHANK3 deletion. We selected one of the four lines for downstream experiments. The deletion was verified by Sanger sequencing.

Homozygous SHANK3 deletions were introduced into the 2242–5 human iPSC line (obtained from Dolmetsch’s laboratory). Two sgRNA sequences, gRNA #8 (TGAGCCGCTATGACGCTTCAGGG) and gRNA #4 (TTTCTCAGGGGTCCCCGGTGGGG), designed to flank exons 21–22 and the stop codon for creating an ~14-kb deletion, were introduced using electroporation together with the Cas9 sequence under the control of the EF-lα promoter. Upon electroporation of 5 µM of gRNA-Cas9 plasmids into ~3 × 106 cells using the Amaxa Nucleofector System (program A023), iPSCs were resuspended in E8 medium with ROCK inhibitor. On the following day, the cells were exposed to 0.25 µg/mL puromycin for 48 h. Antibiotic-resistant colonies were picked and expanded on Matrigel-coated plates in E8 medium. A total of 82 single colonies were manually picked, propagated, and screened for the deletion using conventional PCR and a set of one forward and two reverse primers (Primer F [yw3]: CTCAGGGCCTGCTTGATGAC; Primer R [yw4]: GTGCGCTCCTGAAGGACAAT; and Primer R [yw7]: GGTCTTGCATCGAGGTGCTC), which recognize different regions in the proximity of exons 21–22. Out of the 82 colonies, one was identified to have homozygous SHANK3 deletion encompassing exons 21–22. The deletion was verified using PCR product purification, subcloning into a TOPO TA cloning vector (Thermo Fisher, Cat# 450030), and sequencing.

Vectors and viruses

The following lenti vectors were used in this study: pBOB-synP-HTB (Addgene, Cat#30195) [35]; CamKIIα:smFP-TVA-RVG (constructed in the lab by subcloning the following fragments into the pLenti_CaMKIIα_GFP vector [Addgene plasmid #96941]; smFP-Flag fragment from the pCAG_smFP-Flag_dark vector [a gift from Megan Williams] [36]; F2A-TVA-T2A fragment from the pBOB_synP-HTB vector (Addgene, Cat#30195); oG fragment from the pCAG_oG vector [Addgene plasmid #74288] [37]); CamKIIα:GFP-TVA-RGV (constructed in the lab using the CamKIIα:smFP-TVA-RGV vector and substituting the smFP-Flag fragment with smFP-Flag_bright fragment from the pCAG_smFP-Flag_membrane_bright vector [a gift from Megan Williams] [36]). The lentiviruses were produced and titrated as previously reported [29]. Rabies ENVAΔG-mCherry virus (Addgene, Cat#32636) was purchased from the Salk Institute Gene Transfer, Targeting and Therapeutics core facility.

Animals, transplantation, and rabies virus injection

All work with animals was performed in accordance with guidance by the University of Utah office of Institutional Animal Care & Use Committee. Both female and male NOD scid gamma mice purchased from The Jackson Laboratory (NSG, Cat#005557) were used for experiments. Neurons were dissociated using Papain enzyme (Worthington Biochemical, Cat# LS003119) and resuspended in ND medium supplemented with DNAseI (0.05 mg/mL, Sigma-Aldrich, Cat# D4527–40KU) to prevent cell clumping. Neurons were concentrated to a final concentration of 10,000–50,000 cells/μL. Cells were kept on ice before transplantation. Human neurons were transplanted into the prelimbic cortical area (AP 3.6, ML 0.3, and DV 0.9) of mouse neonates (P0–P3) using a stereotaxic injection system (Stoelting Co.). Hypothermia was used to anesthetize the animals. For this, pups are placed in a latex glove and immersed up to the neck in crushed ice and water (2–3 °C) for about 10 min. The depth of anesthesia is monitored based on the lack of spinal reflex after toe pinch. Anesthetized pups are placed under a stereoscope. A midline incision is made to open the skin, and a glass pipet is used to penetrate the skull and deliver the human cells into the brain (0.5 μL of dissociated human iPSC-derived neurons were delivered using a thin glass needle with <100 μm tip diameter at a 0.1 μL/min flow rate driven by a micro-syringe pump controller). Upon completion, the incision is closed, and the skin is sutured from side to side with sterile nylon 5–0 sutures. The entire surgical procedure lasted no longer than 10 min. After surgery, pups were kept at room temperature for 5 min and then transferred on pre-warmed Deltaphase isothermal pads for recovery. Once fully recovered, responsive, and mobile, pups were returned to their home cage and then to the animal facility. Animal were removed from the study and euthanized using methods consistent with AVMA guidelines if we saw any signs of pain or distress or rejection of transplanted pups by the dam.

Rabies ENVAΔG-mCherry virus was injected into the mPFC of adult animals in an animal biosafety level 2 environment. Briefly, animals were placed on a stereotaxic frame and kept under anesthesia by 1–3% isoflurane inhalation. A small incision was made in the skin to access the skull. The skull above the injection site was opened with a drill. The virus was loaded in a Hamilton syringe and injected at 0.1 μL/min into the adult mPFC following the stereotactic coordinates AP 1.7, ML 0.45, and DV 1.5. The opened skin was sutured after the surgery, and buprenorphine analgesic was provided to the animal every 12 h for 3 days. Rabies virus (RV) was allowed to spread for 7 days before tissue clearing and immunohistochemical analysis.


Acute slices for electrophysiology experiments were obtained from 2- to 15-month-old transplant recipient mice. Mice were anesthetized using isoflurane and, when non-responsive to a toe pinch, transcardially perfused with ice-cold “cutting” solution (~60 mL, 1 mL/s), containing (in mM/L): 2.5 KCl, 7 MgCl2, 1.25 NaH2PO4, 105 Choline-Cl, 25 NaHCO3, 25 D-glucose, 3 Na+-pyruvate, 11 Na+-L-Ascorbate, and. 0.5 CaCl2. Afterwards, the brain was quickly dissected and placed in the ice-cold “cutting” solution. Coronal slices (300–350 µm) were prepared using a Leica VT1200 vibratome in ice-cold “cutting” solution. Slices were allowed to recover for 30 min at 37 °C in standard ACSF, containing (in mM): 124 NaCl, 2.5 KCl, 2 CaCl2, 1 MgCl2, 1.25 NaH2PO4, 25 NaHCO3, 25 d-glucose, 1 Na+-pyruvate, and 1 Na+-l-Ascorbate and then kept at room temperature for the duration of the experiment (5–7 h). Both “cutting” and ACSF solutions were oxygenated using carbogen gas (95% O2/5% CO2).

Whole-cell voltage-clamp experiments were performed using an upright Ziess Axioskop 2 microscope and Multi-Clamp 700B amplifier (Molecular Devices, Palo Alto, CA, USA) at room temperature. Briefly, cells in slices were visualized under a 40× water-immersion objective using an infrared differential interference contrast filter, a digital camera (Ximea), and a fluorescence LED lamp. Recordings were obtained at room temperature in ACSF using patch pipettes pulled from borosilicate capillary glass (BF150–86–10, Sutter Instruments) with resistances 3–5 MΩ when filled with intracellular solution, containing (in mM): 120 CsMeS, 5 CsCl, 10 HEPES, 0.5 CaCl2, 2 EGTA, 3 Na-Ascorbate, 4 MgCl2, 5 d-glucose, 10 Na2-phosphocreatine, 0.4 Na2GTP, 4 Na2ATP, and 0.3% Biocytin (Osmolarity: 306 mOsm). Signals were filtered using 2 kHz Bessel filter and digitized at 50 kHz using Axon DigiData 1500 A (Molecular Devices). The Axon pClamp 10 (Molecular Devices) software was used to control the experiment. Spontaneous excitatory postsynaptic currents (EPSCs) and miniature excitatory and inhibitory postsynaptic currents (mEPSCs and mIPSCs, respectively) were recorded from fluorescent neurons positioned in the deep cortical layers. mEPSCs and mIPSCs were recorded using bath solution containing with 1 µM tetrodotoxin at −60 mV (the reversal potential for mIPSCs) and 0 mV (the reversal potential for mEPSCs), respectively. Events were collected from 3 min of recordings using a template-based approach in Clampfit 10. Templates were generated from control traces with clearly visible postsynaptic events. Thirty randomly selected events from individual recordings were used for statistical analysis in EXCEL (Microsoft) and Prizm (GraphPad). Evoked α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid receptors (AMPAR)- and N-methyl-D-aspartic acid receptors (NMDAR)-EPSCs were obtained from fluorescent neurons positioned in the deep cortical layers using a bipolar stimulating electrode placed in the layers II/III (at least 200 μm from the recorded cells) in response to 1 ms long current pulses of different intensity (0–0.5 mA) delivered every 5 s using an external stimulator (A365, WPI). In these experiments, the intracellular solution contained 5 mM QX-314 to suppress sodium channels and action potentials in the recorded neuron, and the bath solution contained 100 µM picrotoxin to inhibit IPSCs. In some experiments, 2 µM 2-Chloroadenosine was added to the bath solution to suppress seizure-like activity and population spikes. The amplitudes of the smallest responses (average of two traces), measured at the peak using holding potential −70 mV for AMPA-EPSCs and 100 ms post stimulation using holding potential +40 mV for NMDA–EPSCs, were used for analysis. The analysis was performed using the Clampfit 10 software and custom Python scripts.

Western blot analysis

Neurons were harvested and lysed with Laemmli Buffer (Bio-Rad, 1610737). Samples were loaded to a sodium dodecyl sulfate–polyacrylamide gel electrophoresis chamber, transferred to a polyvinylidene difluoride membrane overnight at 4 °C and blocked with 5% milk in phosphate-buffered saline (PBS) with 0.1% Tween 20 (PBST) for 1–2 h with rocking at room temperature prior to primary antibody incubation. Membranes were then incubated with anti-SHANK3 (1:350, Santa Cruz, [sc-301903]) or anti-TUJ1 (1:38,000, Covance [MMS-435P]) antibody in 5% milk in PBS containing Tween 20 (PBST) and slowly rocked at 4 °C overnight. Membranes were washed three times for 10 min each time with PBST and then incubated with horseradish peroxidase (HRP)-conjugated secondary antibody anti-rabbit or anti-mouse at 1:15,000–1:25,000 in 5% milk in PBST for 1–2 h at room temperature. Chemiluminescent HRP substrate (EMD Millipore, Cat#WBKLS0100) was used for visualizing proteins. Membranes were imaged with the ChemiDoc XRS+ System with Image Lab Software (Bio-Rad, 1708265) and analyzed using ImageJ (NIH).

Tissue clearing, immunohistochemistry, and imaging

The mice were anesthetized by intraperitoneal injection of 10 mg/kg xylazine and 100 mg/kg ketamine (Sigma-Aldrich, Cat#K113–10ML) and perfused with cold 16% paraformaldehyde (PFA, Electron Microscope Science, Cat# 15710) and 4% GA (glutaraldehyde, Electron Microscope Science, Cat# 16220) in 1X PBS, at a steady state rate of 100 mmHg2. After perfusion, the brain was dissected and kept in fixative solution for 24 h at 4 °C. To facilitate image analysis, the mouse brain was sectioned in 2.5–3-mm coronal slices for tissue clearing processing (as described in the SWITCH clearing technique by Murray et al. 2015 [38]).

Thick sections

Immediately after electrophysiology experiments, tissue sections (300 μm thick) were transferred to a 24-well plate containing 4% PFA for fixation. After 12 h of fixation, slices were transferred to 1X PBS for storage prior to free-floating immunostaining. Sections were washed three times for 10 min each in 1X PBS, blocked with 10% blocking solution (0.3% Triton X-100, 3% fetal bovine serum [FBS] in 1X PBS) for 6 h at room temperature. Brain sections were incubated with different combinations of primary antibodies (anti-eGFP, anti-CTIP2, anti-SATB2, anti-mKate, and anti-GAD67) in blocking solution for 24 h at 4 °C. Sections were then incubated with secondary antibodies (Alexa fluor 488-mouse, Alexa fluor 647-rat, Alexa fluor 405-mouse, Alexa fluor 568-rabbit, Alexa fluor 647-mouse, and streptavidin Alexa fluor 647conjugate) in blocking solution for 24 h at 4 °C in the dark. Sections were mounted in an aqueous-based mounting medium (Polysciences, Cat#18606–20) with a glass slide and coverslip. Confocal imaging (ZEISS LSM 880 with Airyscan) was performed with excitations at 405, 488, 594, and 647 nm using 5X, 10X, 20X, and 63X oil objectives.

Thin sections

For both the characterization of engrafted GFP-positive neurons and axon tracing, mice were perfused, and fixed brains were cryosectioned at 50-μm thickness. Brain slices were stained with anti-GFP antibody, and confocal imaging (ZEISS LSM 880 with Airyscan) was performed with an excitation wavelength of 488 nm using 5X, 10X, 20X, and 63X oil objectives.

Neuronal cultures

Cells were fixed in 4% PFA (15 min at room temperature), washed (3×) with PBS 1X, permeabilized, and blocked in blocking solution (10% FBS, 0.1% Triton X-100, and 1% bovine serum albumin) for 1 h at room temperature. Primary antibodies were diluted 1:1000 in blocking solution and applied overnight at 4 °C. The following primary antibodies were used: chicken anti-eGFP (1:1000, Novus Biologicals [NB100–1614]), rabbit anti-mKate (1:1000, Axxora [AB234]), mouse anti-FLAG-m2 (1:500, Millipore Sigma [F1804]), rat anti-HA (1:500, Roche [11867423001]), rabbit anti-TBR1 (1:200, Abcam [ab31940]), rat anti-CTIP2 (1:500, Abcam [ab18465]), mouse anti-SATB2 (1:250, Abcam [ab51502]), mouse anti-GAD67 (1:500, EMD Millipore [MAB5406]), mouse anti-TUJ1 (1:1000, Covance [MMS-435P]), and chicken anti-MAP2 (1:1000, Novus Biologicals [NB300–213]). Secondary antibodies and Hoechst 33342 were diluted 1:1000 in blocking solution and applied for 2 h at room temperature. The following secondary antibodies were used: Alexa-405, -488, -594, and -647. Neurons were imaged with a 20X objective on an Airyscan confocal microscope (Zeiss 880).

Rabies virus tracing in vitro

iCtrl and SHANK3+/– neurons transduced with CamKIIα:GFP-TVA-RGV lentivirus were co-plated with non-transduced human neurons of the same genotype from the same round of differentiation at 1:100 ratio (total 1 × 105 cells/well in 24-well plates) on the bed of rat astrocytes grown on poly-d-lysine precoated 15 mm glass cover slips in ND medium as previously described [29]. One-week post plating, neurons were transduced with ENVAΔG-mCherry RV and fixed 48 h post transduction for immunostaining with anti-GFP and DAPI antibodies. Endogenous mCherry fluorescence was used for quantification. In the experiments with human neurons co-cultured with primary mouse cortical neurons, iCtrl or SHANK3+/– neurons transduced with pBOB-synP-HTB or CamKIIα:smFP-TVA-RGV lentiviruses were plated onto the bed of mouse cortical neurons at 1:100 ratio. One-week post plating, neurons were transduced with ENVAΔG-mCherry RV and fixed 48 h post transduction for immunostaining with anti-GFP, anti-FLAG, and Hoechst 33342. Endogenous mCherry fluorescence was used for quantification. The cultures of primary mouse cortical neurons were prepared from E18.5 mouse as previously described in [39]. Briefly, mouse cortices were dissected, gently triturated, dissociated, and plated at 1 × 105 cells/well on poly-l-lysine (Sigma-Aldrich, cat#P9155) precoated 15 mm glass cover slips in Neurobasal medium (Neurobasal-A [Invitrogen, Cat#10888022], 2% B27 Supplement with vitamin A [Thermo Fisher Scientific, Cat#17504044], 1% GlutaMAX [Invitrogen, Cat#35050061], 25 μM 2-Mercaptoethanol [Gibco, cat#21985–023], and 1% Pen/Strep). Two days later, human neurons were added to the rodent cultures. Neurons were visualized using a Nikon Widefield microscope equipped with 405/488/543 lasers using Nikon NIS-Elements multi-platform acquisition software. For each coverslip, a wide-field-of-view image was acquired using a 20X objective and Perfect Focus System.

Neurite tracing and spine analysis

For analysis of neurite morphology, the 300-μm thick coronal sections previously used in the electrophysiology experiment and filled with biocytin were imaged with a 20X objective on an Airyscan confocal microscope (Zeiss 880). Basilar dendritic arbors of selected neurons were traced using Neurolucida (MBF Bioscience, Williston, VT, USA), and arbor complexity was quantified by Sholl analysis using Neurolucida Explorer software. The outcome measures were total dendritic length, number of primary dendrites, maximum number of dendritic intersections within Sholl radii, and branching points (calculated as maximum number of dendritic intersections/number of primary dendrites). Apical dendrites (100 μm in length, located ~10 μm away from the cell soma) were imaged using a 60X objective and 5X electrical zoom on Airyscan confocal microscope (Zeiss 880). For dendritic spine quantification, confocal microscopy images were analyzed using the function Autospine in Neurolucida software.

Data acquisition and statistical analysis

All experiments were performed using human stem cell-derived neurons generated in at least three independent differentiations. Pairs of isogenic lines were simultaneously differentiated and used for the experiments. Both male and female mice were used for transplantations and pooled together for analysis. Animals and cells were randomly selected for the experiments. Investigators were blinded for the genotypes of transplanted cells in animals that were used for the experiments and during analysis. The sample sizes were estimated using 3 G*Power software and the results obtained in our previous study and preliminary experiments. The sample size and statistical tests used for comparisons are reported in the figure legends and tables. Samples were excluded from analysis only if they were clear outliers (identified visually and confirmed using the ROUT method [Q = 2%]). The exclusion criteria were pre-established. We used a bootstrap test of the difference between the means to confirm that the experimentally determined differences between control and SHANK3-deficient cells were not stochastic due to an increased experimental variability.


Connectivity deficits in human stem cell-derived cortical neurons with complete hemizygous SHANK3 deletion in vitro

To investigate synaptic deficits elicited by SHANK3 hemizygosity in human neurons, we generated engineered human PSCs with complete hemizygous SHANK3 deletion (SHANK3+/–) using CRISPR/Cas9 technology and widely used H9 human ESCs (Fig. 1A–C). We employed two single guide RNAs (sgRNAs), one targeting the first exon and the second one directed against the stop codon, to achieve a complete deletion of the genomic region spanning the entire SHANK3 coding sequence (Fig. 1A). Out of 64 selected colonies, a complete hemizygous SHANK3 deletion (69.5 kB) was confirmed in one line using conventional gel electrophoresis and Sanger sequencing (Fig. 1B, C).

Fig. 1: SHANK3 hemizygosity in human ESC-derived neurons caused synaptic connectivity deficits in vitro.

A CRISPR/Cas9 strategy for introducing complete hemizygous SHANK3 deletion in control human H9 ESCs. The positions of sgRNAs and primer pairs used for PCR and sequencing depicted with arrows. B Confirmation of hemizygous SHANK3 deletion using PCR gel electrophoresis (top) and Sanger sequencing (bottom). The band from the primer pair 2–3 shows an intact copy of SHANK3 at 3’ cutting sites, and the band from the primer pair 1–4 is only present when SHANK3 is successfully deleted from at least one copy. The Sanger sequencing pinpoints the exact cutting sites. C Images of iCtrl and SHANK3+/– stem cell colonies. D Images of human neurons immunostained with anti-MAP2, TBR1, CTIP2, and SATB2 antibodies. E Percentages of MAP2-expressing neurons co-immunostained with TBR1 (n = 2 iCtrl and 2 SHANK3+/– cover slips/580 cells), CTIP2 (n = 2 iCtrl and 2 SHANK3+/– cover slips/468 cells) and SATB2 (n = 2 iCtrl and 2 SHANK3+/– cover slips/ 256 cells). F Immunoblot images of lysates from iCtrl or SHANK3+/– hESC-derived neurons stained using anti-SHANK3 (top) and anti-TUJ1 (bottom) antibodies. G Quantification of expression of different isoforms of SHANK3 in iCtrl or SHANK3+/– neurons (n = 3 biological replicates). H Experimental design to assess synaptic connectivity using RV tracing in vitro. iCtrl and SHANK3+/– neurons were transduced with lentiviruses carrying GFP, TVA, and RVG under the control of the CaMKIIα promoter. Only human neurons expressing TVA could be infected with G-deleted EnvA-pseudotyped RV carrying mCherry (starter neurons). Assembled RV can propagate one step to immediate presynaptic neurons. As a result, starter cells express both GFP and mCherry (yellow) and presynaptic neurons express mCherry (red). Connectivity index is calculated as a ratio of red to yellow cells. I Representative images of starter (GFP- and mCherry-expressing cells, marked with arrows) and presynaptic traced cells (mCherry-expressing cells, marked with arrow heads) in cultures of iCtrl and SHANK3+/– neurons. J Quantifications of starter cells (left), traced cells (middle), and connectivity index (right) in cultures of control and SHANK3–/+ neurons (n = 3 iCtrl and 3 SHANK3+/– cover slips). Data presented as means ± standard error of the mean (s.e.m.); *P < 0.05, unpaired t-test. Scale bars = 30 μm.

We next differentiated iCtrl and SHANK3+/– stem cells into neurons using a modified dual-SMAD inhibition protocol (Supplementary Fig. 1A), which allows the generation of human neurons with telencephalic identities (Supplementary Fig. 1B, C). We confirmed that at 6 weeks post induction, MAP2-expressing cells expressed typical deep-layer cortical markers TBR1 (~40%) and CTIP2 (~30%) and the superficial layer marker SATB2 (~15%) (Fig. 1D, E). We observed very similar percentages of these cells among both iCtrl and SHANK3+/– neurons (Fig. 1E), suggesting that SHANK3 hemizygosity produces no major effects on the specification of cortical excitatory neurons in vitro.

We performed western blot analysis to characterize the expression levels of SHANK3 proteins in iCtrl and SHANK3+/– cultures (Fig. 1F, G). Interestingly, we found that complete hemizygous SHANK3 deletion caused an approximately 50% reduction in the expression levels of the longest (~250 kDa) and shortest (~160 KDa) isoforms of SHANK3 proteins, but not in the expression levels of two intermediate isoforms (~200 and 180 kDa) (Fig. 1G). This result suggests that SHANK3 hemizygosity in human neurons triggers an ectopic SHANK3 expression from alternative internal promoters. This is consistent with the results of previous studies on Shank3-deficient mice and monkeys [19, 23, 40], suggesting the presence of conservative compensatory feedback mechanisms for regulating the level of SHANK3 expression [41, 42].

We next investigated whether SHANK3 hemizygosity in human neurons affects their synaptic connectivity, as was previously reported for SHANK3-deficient neurons from PMDS patients with large 22q13 microdeletions and engineered ESCs with partial SHANK3 deletion [29, 31]. To assess the levels of synaptic connectivity in the cultures of iCtrl and SHANK3+/– neurons, we used trans-synaptic RV tracing [37] (Fig. 1H). This approach has been widely used for mapping synaptic connectivity in the mouse brain [43] and in human stem cell-derived neurons [44, 45]. In contrast to conventional synaptic electrophysiology or immunostaining, this approach allows to assess the level of synaptic connectivity based on the proportions of starter cells (transduced to express the avian tumor virus receptor A (TVA), which is required for RV entry into the cells) and presynaptically connected traced cells (acquired RV through synaptic connections with the starter cells) in a robust and high-throughput fashion. Using RV tracing, we found that SHANK3+/– neurons establish significantly fewer connections than iCtrl neurons in vitro (Fig. 1I, J).

To determine whether connectivity deficits in SHANK3+/– cultures arise cell-autonomously as a result of deficits in postsynaptic neurons as we observed in our previous study on PMDS neurons [29], we co-cultured iCtrl and SHANK3+/– neurons with dissociated mouse cortical neurons to provide the same presynaptic inputs to both iCtrl and SHANK3+/– neuron (Supplementary Fig. 2A). We observed that SHANK3+/– neurons established significantly fewer connections with co-cultured mouse cortical neurons than did iCtrl neurons (Supplementary Fig. 2B, C). This result indicates that hemizygous deletion of SHANK3 in the postsynaptic neurons impairs synaptic connectivity.

To further validate RV tracing for studying synaptic connectivity in vitro and connectivity deficits detected in the cultures with SHANK3+/– neurons, we differentiated human neurons from another pair of iCtrl and SHANK3-deficient stem cell lines with homozygous deletions of the last two exons and the stop codon of SHANK3 (Supplementary Fig. 3A, B). We observed no SHANK3 expression in neurons with homozygous SHANK3 deletion (Wang et al., in review). We co-cultured iCtrl and homozygous SHANK3-deficient neurons with dissociated mouse cortical neurons and assessed synaptic connectivity using RV tracing (Supplementary Fig. 3C). Consistent with the results of our experiments with hemizygous SHANK3-deficient neurons, we found significantly fewer connections to homozygous SHANK3-deficient neurons, as compared to iCtrl neurons (Supplementary Fig. 3E). As expected, the connectivity deficits in homozygous SHANK3-deficient neurons were more pronounced than those in hemizygous SHANK3-deficient neurons. Finally, the transduction of mouse cortical neurons with EnvA-pseudotyped RV alone resulted in no mCherry-expressing cells (negative control, Supplementary Fig. 3F), confirming that RV transduction relies exclusively on the presence of starter cells expressing TVA receptors [46]. Together, these results indicate that SHANK3 hemizygosity in human neurons leads to impaired excitatory synaptic connectivity, likely due to cell-autonomous deficits in the postsynaptic neurons induced by the loss of the longest and/or shortest isoforms of SHANK3.

Engraftment of human stem cell-derived neurons into the mouse PFC for studying connectivity

The main caveat of connectivity studies on human iPSC-derived neurons in vitro is the lack of a physiologically relevant brain environment. Engrafted human stem cell-derived neurons have been used for studying the maturation of human neurons and cellular deficits in human neurological and neurodegenerative disorders [47,48,49,50,51,52]. We investigated whether human stem cell-derived cortical neurons can be transplanted into the mouse PFC for studying synaptic deficits caused by SHANK3 hemizygosity. We chose the PFC region for transplantation because the expression pattern of high-confidence ASD-associated genes converges on the PFC [53, 54], and because altered functional connectivity in this region has frequently been identified in individuals with ASD [55,56,57].

We transduced control human stem cell-derived neurons with lentiviruses carrying GFP under the control of the CamKIIα promoter, which primarily labels functionally mature cortical excitatory neurons [29], and transplanted them into the PFC of neonatal immunocompromised mice (P0–P5) to promote their survival and integration into developing mouse neural circuits (Supplementary Fig. 1). We observed that the majority of transplanted animals (65%) survived and possessed GFP-expressing cells in the PFC (Fig. 2A, B and Supplementary Table 1). Two weeks post transplantation, all GFP-expressing cells were found localized within a 300-μm range surrounding the center of the graft (Fig. 2C), and exhibited relatively immature neuronal morphology with processes primarily oriented towards the midline (Fig. 2B), which is similar to the orientation of mouse pyramidal cortical neurons in the corresponding brain region. However, we also noticed that the numbers and distributions of GFP-expressing cells varied from animal to animal (Fig. 2D).

Fig. 2: Anatomical integration of human neurons transplanted into the mouse PFC.

A Images of PFC slices from five different mice engrafted with control ESC-derived human neurons expressing CaMKIIαːGFP. B Magnified image of engrafted human cells in mouse 4. C Distributions of GFP-expressing human neurons along the medial-lateral (ML), dorso-ventral (DV), and anterior-posterior (AP) axes in the mouse brain (n = 5 mice). D Graft coordinates and number of cells in different animals (n = 5 mice). E Experimental schematic for assessing the distribution of anatomical inputs on transplanted human neurons using trans-synaptic RV tracing. Human stem cell-derived neurons, transduced with lentiviruses carrying a “spaghetti monster” dark fluorescent protein (smFP), TVA, RVG under the control of the CaMKIIα promoter in vitro, were transplanted into the PFC of neonatal immunocompromised mice. Only engrafted human neurons expressing TVA could be infected with G-deleted EnvA-pseudotyped RV carrying mCherry (starter neurons). Assembled RV can propagate one step to immediate presynaptic neurons. As a result, both starter and presynaptic neurons are labeled with mCherry. F Images of mCherry-expressing cells in different regions of the mouse brain. G Images of GFP-expressing processes originating from hESCs-derived neurons expressing GFP under the control of the CaMKIIα promoter found in different brain regions. H Quantification of anatomical inputs on human neurons transplanted into the mouse PFC (n = 8 mice). I Quantification of anatomical outputs from human neurons transplanted into the mouse PFC (n = 4 mice). Data presented as means ± s.e.m. Scale bars = 500 (A, Fa, Fb, Fc, Fd, Fe, Ff, Ga, Gb, and Gc), 200 (B), 100 (zoom-in images in F), 50 (B), and 10 μm (zoom-in images in G). Age of investigated animals: 4 months (AD), 6–8 months (F, H), 12 months (G, I). MO motor cortex, MOp primary motor cortex, MOs secondary motor cortex, mPFC medial prefrontal cortex, ACC anterior cingulate cortex, SSp primary somatosensory cortex, TH thalamus, cSSp contralateral primary somatosensory cortex, AI agranular insula, cAI contralateral agranular insula, BLA basolateral amygdala, HI hippocampus, cmPFC contralateral medial prefrontal cortex, cAI contralateral agranular insula, PAL pallidum, cACC contralateral anterior cingulate cortex, MB midbrain brain, ORB orbital cortex, cmPFC contralateral mPFC, TT taenia tecta of the olfactory cortex, STR striatum, CC corpus callosum.

We next investigated how transplanted human neurons become anatomically integrated in the mouse brain. To map neuronal inputs onto the transplanted human neurons, we used RV-mediated retrograde trans-synaptic tracing [43, 58] and 3-mm thick cleared brain slices using the SWITCH approach [38]. For this, hESC-derived neurons expressing TVA receptor and rabies virus glycoprotein were transplanted into the PFC of immunocompromised neonatal mice and then transduced with RV for connectivity tracing 4–8 months post transplantation (Fig. 2E). Interestingly, we found that RV-transduced neurons expressing mCherry were distributed to multiple brain regions proximal and distal to the transplantation site, including the secondary motor area (MOs: ~56% of all mCherry-expressing cells, Fig. 2Fa, H), medial PFC (mPFC: ~10%, Fig. 2Fb, H), anterior cingulate cortex (ACC: ~10%, Fig. 2H), primary somatosensory cortex (SSp: ~4.5%, Fig. 2H), thalamus (TH: ~3.5%, Fig. 2Fd, H), contralateral primary somatosensory cortex (cSSp: ~3%, Fig. 2H), agranular insula (AI: ~2%, Fig. 2Fc, H), contralateral agranular insula (cAI: ~2%, Fig. 2Fe, H), basolateral amygdala (BLA: ~1.7%, Fig. 2Fd, H), hippocampus (HI: ~1.7%, Fig. 2H), pallidum (PAL: ~1.4%, Fig. 2Ff, H), primary motor area (MOp: ~1.2%, Fig. 2Fa, H), contralateral anterior cingulate cortex (cACC: ~1.2%, Fig. 2H), midbrain brain (MB: ~0.5%, Fig. 2H), orbital cortex (ORB: ~0.3%, Fig. 2H), contralateral mPFC (cmPFC: ~0.1%, Fig. 2Fe, H), and taenia tecta of the olfactory cortex (TT: ~0.1%, Fig. 2H). These results suggest that human neurons transplanted into the mouse neonatal cortex are capable of receiving anatomical inputs from multiple regions in the mouse brain, which is consistent with the connectivity pattern observed for excitatory neurons in the mouse PFC [43, 59]. A caveat of this experiment, however, is that we were unable to use immunostaining with the cleared mouse brain tissue to map the starter cells and to identify the subtypes of traced presynaptic neurons in different brain regions.

To map outputs from the transplanted human neurons expressing GFP under the control of the CamKIIα promoter (CamKIIα:GFP), we counted the numbers of fluorescent processes in different mouse brain regions that are known to receive anatomical projection from the PFC [60]. We found GFP-expressing processes in the thalamus (TH: ~38%, Fig. 2Gb, I), primary and secondary motor area (MOs: ~32%, Fig. 2Ga, I), striatum (STR: ~20%, Fig. 2Ga, I), primary somatosensory cortex (SSp: ~5%, Fig. 2Ga, I), and corpus callosum (CC: ~5%, Fig. 2Gc, I). This distribution pattern of GFP-expressing processes is similar to that observed for mouse PFC neurons [60].

Together, these results indicate that transplanted human neurons become anatomically integrated in the mouse PFC and to a large extent mimic the connectivity pattern of the host mouse neurons in the corresponding brain region. However, the transplantation experiments also revealed that the connectivity pattern and distribution of transplanted human neurons vary substantially from animal to animal (coefficient of variability: 57–283%, Supplementary Table 2), which may prohibit the use of this approach for identifying specific anatomical connections disrupted in ASD. To overcome this limitation, we adapted the dual-color labeling scheme [29] and co-transplantation to deliver human iCtrl and SHANK3+/– neurons into the same brain region for co-development under the same experimental conditions (Fig. 3).

Fig. 3: AMPAR-mediated excitatory synaptic deficits in developing SHANK3+/− neurons co-engrafted with iCtrl neurons in the mouse PFC.

A Experimental schematic for co-transplantation of iCtrl and SHANK3+/– neurons into the mouse PFC. B Quantifications of fluorescent cells expressing deep-layer marker CTIP (n = 3 iCtrl and 3 SHANK3+/– sections) and superficial layer marker SATB2 (n = 3 iCtrl and 3 SHANK3+/– sections). C Image of slice with co-engrafted iCtrl (red) and SHANK3+/– (green) neurons. D, E Magnified images of CTIP2- (D) and SATB2-expressing (E) iCtrl (red) and SHANK3+/– (green) neurons. F Area in PFC slices targeted for synaptic electrophysiology (top) and image of co-engrafted iCtrl (green) and SHANK3+/– (red) neurons in that area (bottom). G Traces of EPSCs obtained from iCtrl (black) and SHANK3+/– (red) neurons at holding potential of −70 mV (J). H, I Cumulative distributions and quantifications of EPSC amplitude (H) and frequency (I) (n = 8 iCtrl and 7 SHANK3+/– cells/7 animals). J Images of PFC slice used for electrophysiology with positions of stimulating electrode, patch pipette, and patched GFP-expressing human neuron. K,Traces of evoked AMPAR- and NMDAR-mediated postsynaptic currents measured at different holding potentials (−70 and +40 mV, respectively) in iCtrl (black) and SHANK3+/– (red) human neurons. Stimulus artifacts are blanked for clarity. L, M Quantifications of amplitudes of AMPA (left, measured at peak) and NMDA (right, measured at 100 ms post stimulation) currents (L) and their ratios (M) (n = 5 iCtrl and 5 SHANK3+/– cells/6 animals). Data presented as means ± s.e.m.; *P < 0.05, unpaired t-test. Scale bars = 500 (C), 50 (DF), 100 (J), and 5 (J, zoom-in) μm. Age of investigated animals: 12–17 (CE) and 3–8 (FM) months.

Impaired AMPAR-mediated excitatory synaptic transmission in developing SHANK3 +/− human neurons

To determine whether hemizygous SHANK3 human neurons develop excitatory synaptic deficits in vivo, we transduced iCtrl and SHANK3+/– neurons with lentiviruses carrying RFP or GFP, respectively (or vice versa), under the control of the CamKIIα promoter and co-transplanted them together into the PFC of neonatal immunocompromised mice (Fig. 3A). Consistent with the results of our experiments in vitro, we observed that iCtrl and SHANK3+/– neurons acquired similar cortical laminar identities in the mouse brain (Fig. 3B–E). Specifically, we found that about 55% of engrafted human neurons expressed the deep-layer cortical marker CTIP2 (Fig. 3B, D) and approximately 30% expressed the superficial layer cortical marker SATB2 (Fig. 3B, E). These results suggest that SHANK3 hemizygosity in human neurons produces no overt effects on the specification of deep and superficial cortical neuronal identities in vivo.

We next investigated the functional synaptic properties of co-transplanted iCtrl and SHANK3+/– neurons in the PFC slices using patch-clamp electrophysiology (Fig. 3F–M). To determine whether basal excitatory synaptic transmissions was altered in SHANK3+/– neurons, we measured the amplitudes and frequencies of spontaneous EPSC at −70 mV in the presence of picrotoxin, an antagonist of GABAA receptors 3–8 months post transplantation (Fig. 3G and Supplementary Fig. 4). We found that amplitude but not frequency of spontaneous EPSC was significantly reduced SHANK3+/– neurons, as compared to co-transplanted iCtrl neurons (Fig. 3H, I). The reduced EPSC amplitude in SHANK3+/– neurons was observed at both 3 and 8 months post transplantation (Supplementary Fig. 4). This result is consistent with the result of a previous study on induced stem cell-derived human neurons with heterozygous SHANK3 deletion in vitro [31]. It suggests that SHANK3 deficiency in developing human neurons affects the strength of excitatory synapses.

Excitatory synapses in the mammalian brain express two major subtypes of glutamatergic receptors: AMPARs and NMDARs [61]. We investigated whether the impaired excitatory synaptic transmission detected in SHANK3+/– neurons is associated with reductions in both AMPAR- and NMDAR-mediated excitatory postsynaptic currents (AMPAR-EPSCs and NMDAR-EPSCs, respectively) (Fig. 3J–M). To assess AMPAR- and NMDAR-EPSCs in transplanted human neurons, we recorded evoked postsynaptic currents from co-transplanted iCtrl and SHANK3+/– human neurons in layer 5 induced by electrical stimulations with an electrode positioned in layers 2/3 to activate both local and distal inputs (Fig. 3J). Interestingly, we detected a significantly reduced amplitude of AMPAR-EPSCs, but not NMDAR-EPSCs (Fig. 3K, L), and, as a result, a significantly reduced AMPA to NMDA ratio in SHANK3+/– human neurons as compared to co-transplanted iCtrl neurons (Fig. 3M). This result suggests that excitatory synaptic deficits in SHANK3-deficient human neurons are associated with reduced expression of functional AMPARs, but not NMDARs, at excitatory synapses. This is consistent with the results of our previous study [29], in which we observed that only the AMPAR-EPSCs were completely rescued by exogenous SHANK3 in iPSC-derived neurons from PMDS patients.

Reduced excitatory synapses, spines, and dendrites in mature SHANK3 +/− human neurons

An advantage of the xenotransplantation model is that it supports prolonged development of xenografted human neurons to allow their functional and morphological maturation [51, 52]. It has been shown that the properties of xenografted stem cell-derived human cortical neurons resemble those of mature human cortical pyramidal neurons only after 10–11 months post transplantation [52]. Therefore, we investigated the synaptic and morphological properties of co-transplanted iCtrl and SHANK3+/– neurons at 13–17 months post transplantation (Fig. 4). To assess the basal excitatory and inhibitory synaptic transmissions, we measured the amplitudes and frequencies of miniature excitatory and inhibitory postsynaptic currents (mEPSC and mIPSC, respectively) in the presence of TTX at different holding membrane potentials in co-transplanted iCtrl and SHANK3+/– neurons (Fig. 4A–F). Surprisingly, we found a similar mEPSC amplitude but significantly reduced mEPSC frequency in SHANK3+/– neurons, as compared to iCtrl neurons (Fig. 4A–C). Assuming that co-transplanted iCtrl and SHANK3+/– neurons interact with the same presynaptic inputs in the brain, this result suggests that SHANK3+/– neurons have reduced numbers of excitatory synapses, as compared to co-transplanted iCtrl.

Fig. 4: Decreased excitatory synaptic transmission, spines, and dendritic arbors in mature SHANK3+/− neurons co-engrafted with iCtrl neurons in the mouse PFC.

A, D Traces of mEPSCs (A) and mIPSCs (D) obtained from iCtrl (black) and SHANK3+/– (red) neurons at holding potentials of −60 (A) and 0 mV (D), respectively. B, C and E, F Cumulative distributions and quantifications of mEPSC amplitude (B) and frequency (C) and mIPSC amplitude (E) and frequency (F) (n = 10 iCtrl and 12 SHANK3+/– cells/4 animals). G Images of co-transplanted iCtrl (red) and SHANK3+/– (green) human neurons in the mouse PFC at different magnifications (Ga–Gc). H Images of spines on apical dendrites in iCtrl (left) and SHANK3+/– neurons (right). I Quantification of spine densities on apical dendrites (n = 7 iCtrl and 6 SHANK3+/– cells/3 animals; average spine density iCtrl: 0.4 spines/μm and SHANK3+/–: 0.2 spines/μm). J Images of reconstructed iCtrl and SHANK3+/– neurons from three distinct animals (TP1–3). K–M, Sholl analysis (K) and quantifications of total dendritic length (L) and branching points (M) (n = 7 iCtrl and 6 SHANK3+/– cells/4 animals; average total dendritic length iCtrl: 4800 μm, SHANK3+/–: 1100 μm). Data presented as absolute values ± s.e.m.; *P < 0.05, unpaired t-test. Scale bars = 1000 (G), 500 (Ga), 100 (GbGc, and J), and 2 μm (H). Age of investigated animals: 13–17 months.

We observed no differences in either the amplitude or frequency of mIPSCs (Fig. 4D–F), suggesting that co-transplanted iCtrl and SHANK3+/– neurons receive similar inhibitory synaptic inputs. This result is consistent with the results of our previous in vitro study on patient-derived SHANK3+/– neurons [29]. It indicates that a balance of excitatory and inhibitory synaptic inputs on the individual human neurons may be disrupted as a result of SHANK3 deficiency.

Mature AMPAR-containing excitatory synapses predominantly reside on dendritic spines in cortical excitatory neurons [62,63,64,65]. We investigated whether reduced mEPSC frequency detected in SHANK3+/– neurons is associated with defects in dendritic spines. We reconstructed the dendritic morphologies of most recorded neurons filled with biocytin (Fig. 4G) and analyzed the density and morphology of spines on apical dendrites (Fig. 4H, I). We found that the density of dendritic spines on iCtrl neurons was similar to that detected in another study on transplanted stem cell-derived human neurons [52] and comparable to that observed in neonatal human brain tissue [66,67,68,69,70]. Interestingly, SHANK3+/– neurons exhibited significantly reduced spine density as compared to iCtrl neurons (Fig. 4I), particularly those with mushroom and thin morphologies that have been associated with more mature synaptic contacts [58, 59]. This result is consistent with the idea that SHANK3 is an important regulator of spinogenesis in neurons [71] and that reduced number of dendritic spines could be responsible for excitatory synaptic deficits in more mature SHANK3+/– human neurons.

In addition, we found that SHANK3+/– neurons had dramatically underdeveloped dendritic arbors, including significantly reduced dendritic length and number of branches, as compared to iCtrl neurons (Fig. 4B, C). This result is consistent with the results of the previous studies on SHANK3-deficient human neurons that detected morphological deficits in vitro [31]. It suggests that SHANK3 is an important regulator of the overall morphological development and maturation in human neurons.


In this study, we established a co-xenotransplantation model for studying synaptic deficits in genetically engineered human neurons. We demonstrated that a complete hemizygous SHANK3 deletion, which is the most common genetic abnormality identified in PMDS patients, causes impaired AMPAR-mediated excitatory synaptic transmission, reduced dendritic spine density, and underdeveloped dendritic arbors in human cortical neurons engrafted in the mouse PFC. Interestingly, we also found that SHANK3 deficiency predominately affects the strength of excitatory synapses at early developmental time points and leads to the losses of excitatory synapses and spines later in development. These results are consistent with the idea that SHANK3 hemizygosity causes AMPAR-mediated excitatory synaptic deficits and morphological abnormalities in human neurons.

The xenotransplantation approach has been previously used to investigate the maturation of human neurons under physiological conditions [52, 72,73,74] and to develop cell-based therapies for patient with neurological and neurodegenerative disorders [47,48,49,50,51, 58, 75]. Our study presents a new direction for using this approach to study synaptic connectivity deficits in neurodevelopmental disorders associated with ASD. To ensure that the phenotypes observed in human neurons are physiologically relevant and caused by the genetic abnormality and not other factors that are difficult to control in transplantation studies, including different host brain environments and shifts in transplantation coordinates, we developed a co-xenotransplantation model. In this model, both wild-type and mutant human neurons reside in the same brain environment throughout development and interact with the same presynaptic inputs from both human and rodent neurons. Therefore, the synaptic and morphological deficits detected in this study are likely to be postsynaptic cell-autonomous deficits caused by SHANK3 hemizygosity.

Multiple synaptic deficits have been detected in rodent neurons in association with Shank3 deficiency, including an impaired amplitude and/or frequency of AMPAR-mediated excitatory synaptic transmission, impaired NMDAR-mediated synaptic responses, impaired long-term potentiation (LTP) and/or long-term depression (LTD), impaired mGluR signaling and mGluR-dependent LTD, and reduced dendritic spine density. This has led to the idea that synaptic function may be disrupted in association with SHANK3 hemizygosity in humans. However, the specific deficits have been incongruent across studies. For example, five studies found a reduced frequency of EPSCs in Shank3-deficient neurons [12, 13, 19, 20, 22], whereas two studies observed no such reduction [16, 24] and two other studies found an increased frequency of EPSCs [18, 26]. Similarly, four studies reported impaired LTP in CA1 pyramidal neurons of Shank3-deficient mice [12, 13, 24, 26], but three studies found no LTP deficits [16, 19, 20]. Furthermore, two studies reported reduced mGluR5 expression [15, 76], while three studies reported either no reduction or increased mGluR5 expression [12, 13, 22]. Moreover, no overt synaptic deficits have been detected in Shank3-deficient mice with a complete heterozygous Shank3 deletion [12]. These phenotypic inconsistencies are likely due to differences in the deleted isoforms of Shank3, the genetic backgrounds, and the developmental stages and types of investigated neurons. As a result, it remains unclear which of the detected synaptic deficits can be targeted for therapy development in the preclinical studies.

Studies on human stem cell-derived neurons obtained from patients and CRISPR/Cas9-engineered stem cell lines have provided further insights into early patient- and mutation-specific deficits induced by SHANK3 deficiency. Interestingly, it has been demonstrated that even heterozygous SHANK3 deletions in human neurons lead to profound excitatory synaptic deficits and morphological abnormalities in vitro [31]. These results suggest either that SHANK3 is critically involved in the development of excitatory synapses in human neurons so that the loss of even a single copy of SHANK3 leads to excitatory synaptic deficits, or that synaptic deficits detected in vitro are compensated in the brain environment. Our results on co-transplanted control and SHANK3-deficient human neurons grown in the mouse PFC support the notion that SHANK3 is essential for excitatory synaptic development in human neurons.

Interestingly, we found that human neurons with hemizygous SHANK3 deletion exhibit different excitatory synaptic deficits at different developmental stages. At 3–8 months post transplantation, SHANK3 deficiency is associated with reduced amplitude but not frequency of spontaneous EPSCs and impaired amplitude of evoked AMPAR-mediated synaptic responses; while at 13–17 months post transplantation, it is associated with reduced mEPSC frequency, but not amplitude, and decreased number of dendritic spines. Reduced amplitude of spontaneous EPSCs and evoked AMPAR-mediated postsynaptic current can be attributable to reduced expression of functional AMPAR in the postsynaptic density. Reduced mEPSC frequency can be caused by presynaptic deficits and/or decreased number of excitatory synapses. In the co-xenotransplantation system, both iCtrl and SHANK3+/– neurons interact with the same pool of presynaptic inputs coming from both mouse and human neurons, suggesting that reduced mEPSC frequency is likely associated with decreased number of excitatory synapses on SHANK3+/– neurons as compared to iCtrl neurons. AMPAR-containing excitatory synapses are predominantly localized on spines [63]. Consistent with this idea, we observed that mature SHANK3+/– neurons have fewer dendritic spines as compared to iCtrl neurons. Together, these results suggest that SHANK3 hemizygosity in human neurons grown under physiological conditions leads to defective expression of functional postsynaptic AMPA receptors early in development and decreased number of excitatory synapses and spines at later developmental stages.

It remains to be determined whether the deficits detected during different developmental stages are interrelated or part of distinct mechanisms of synapse development and/or function mediated by SHANK3. In addition, in future studies, it is important to investigate the time course for the development of functional and morphological deficits in SHANK3+/– human neurons and determine the relationship between defective synaptic connectivity and the intrinsic excitability deficits observed in vitro [31]. Finally, it remains unclear how reduced SHANK3 expression causes these deficits in human neurons and whether the same deficits could be detected in patient-derived neurons with different genetic abnormalities or different lines of isogenic iPSC-derived neurons with hemizygous SHANK3 deletion.

In summary, this study identifies specific postsynaptic deficits induced by SHANK3 hemizygosity in human stem cell-derived cortical neurons transplanted in the mouse PFC and validates the use of the co-xenotransplantation model for detecting synaptic and morphological deficits in genetic neurodevelopmental disorders associated with ASD.


  1. 1.

    Bourgeron T. From the genetic architecture to synaptic plasticity in autism spectrum disorder. Nat Rev Neurosci. 2015;16:551–63.

    CAS  Article  Google Scholar 

  2. 2.

    Zoghbi HY, Bear MF. Synaptic dysfunction in neurodevelopmental disorders associated with autism and intellectual disabilities. Cold Spring Harb Perspect Biol. 2012;4:a009886.

    Article  Google Scholar 

  3. 3.

    Khan S, Gramfort A, Shetty NR, Kitzbichler MG, Ganesan S, Moran JM, et al. Local and long-range functional connectivity is reduced in concert in autism spectrum disorders. Proc Natl Acad Sci. 2013.

  4. 4.

    De Rubeis S, He X, Goldberg AP, Poultney CS, Samocha K, Cicek AE, et al. Synaptic, transcriptional and chromatin genes disrupted in autism. Nature. 2014.

  5. 5.

    Satterstrom FK, Kosmicki JA, Wang J, Breen MS, De Rubeis S, An JY, et al. Large-scale exome sequencing study implicates both developmental and functional changes in the neurobiology of autism. Cell. 2020.

  6. 6.

    Boeckers TM, Inter C, Smalla KH, Kreutz MR, Bockmann J, Seidenbecher C, et al. Proline-rich synapse-associated proteins ProSAP1 and ProSAP2 interact with synaptic proteins of the SAPAP/GKAP family. Biochem Biophys Res Commun. 1999.

  7. 7.

    Naisbitt S, Kim E, Tu JC, Xiao B, Sala C, Valtschanoff J, et al. Shank, a novel family of postsynaptic density proteins that binds to the NMDA receptor/PSD-95/GKAP complex and cortactin. Neuron. 1999;23:569–82.

    CAS  Article  Google Scholar 

  8. 8.

    Leblond CS, Nava C, Polge A, Gauthier J, Huguet G, Lumbroso S, et al. Meta-analysis of SHANK mutations in autism spectrum disorders: a gradient of severity in cognitive impairments. PLoS Genet. 2014;10:e1004580.

    Article  Google Scholar 

  9. 9.

    Phelan MC, Rogers RC, Saul RA, Stapleton GA, Sweet K, McDermid H, et al. 22q13 deletion syndrome. Am J Med Genet. 2001.<91::aid-ajmg1340>;2-c.

  10. 10.

    Mitz AR, Philyaw TJ, Boccuto L, Shcheglovitov A, Sarasua SM, Kaufmann WE, et al. Identification of 22q13 genes most likely to contribute to Phelan McDermid syndrome. Eur J Hum Genet. 2018;26:293–302.

    CAS  Article  Google Scholar 

  11. 11.

    Wilson HL, Wong ACC, Shaw SR, Tse W-Y, Stapleton GA, Phelan MC, et al. Molecular characterisation of the 22q13 deletion syndrome supports the role of haploinsufficiency of SHANK3/PROSAP2 in the major neurological symptoms. J Med Genet. 2003;40:575–84.

    CAS  Article  Google Scholar 

  12. 12.

    Wang X, Bey AL, Katz BM, Badea A, Kim N, David LK, et al. Altered mGluR5-Homer scaffolds and corticostriatal connectivity in a Shank3 complete knockout model of autism. Nat Commun. 2016.

  13. 13.

    Kouser M, Speed HE, Dewey CM, Reimers JM, Widman AJ, Gupta N, et al. Loss of predominant Shank3 isoforms results in hippocampus-dependent impairments in behavior and synaptic transmission. J Neurosci. 2013;33:18448–68.

    CAS  Article  Google Scholar 

  14. 14.

    Duffney LJ, Zhong P, Wei J, Matas E, Cheng J, Qin L, et al. Autism-like deficits in Shank3-deficient mice are rescued by targeting actin regulators. Cell Rep. 2015.

  15. 15.

    Verpelli C, Dvoretskova E, Vicidomini C, Rossi F, Chiappalone M, Schoen M, et al. Importance of Shank3 protein in regulating metabotropic glutamate receptor 5 (mGluR5) expression and signaling at synapses. J Biol Chem. 2011;286:34839–50.

    CAS  Article  Google Scholar 

  16. 16.

    Lee J, Chung C, Ha S, Lee D, Kim D-Y, Kim H, et al. Shank3-mutant mice lacking exon 9 show altered excitation/inhibition balance, enhanced rearing, and spatial memory deficit. Front Cell Neurosci. 2015;9:94.

    PubMed  PubMed Central  Google Scholar 

  17. 17.

    Peixoto RT, Chantranupong L, Hakim R, Levasseur J, Wang W, Merchant T, et al. Abnormal striatal development underlies the early onset of behavioral deficits in Shank3B−/− mice. Cell Rep. 2019.

  18. 18.

    Peixoto RT, Wang W, Croney DM, Kozorovitskiy Y, Sabatini BL. Early hyperactivity and precocious maturation of corticostriatal circuits in Shank3B(–/–) mice. Nat Neurosci. 2016.

  19. 19.

    Speed HE, Kouser M, Xuan Z, Reimers JM, Ochoa CF, Gupta N, et al. Autism-associated insertion mutation (InsG) of Shank3 Exon 21 causes impaired synaptic transmission and behavioral deficits. J Neurosci. 2015;35:9648–65.

    CAS  Article  Google Scholar 

  20. 20.

    Peça J, Feliciano C, Ting JT, Wang W, Wells MF, Venkatraman TN, et al. Shank3 mutant mice display autistic-like behaviours and striatal dysfunction. Nature. 2011;472:437–42.

    Article  Google Scholar 

  21. 21.

    Zhou Y, Sharma J, Ke Q, Landman R, Yuan J, Chen H, et al. Atypical behaviour and connectivity in SHANK3-mutant macaques. Nature. 2019.

  22. 22.

    Mei Y, Monteiro P, Zhou Y, Kim J-A, Gao X, Fu Z, et al. Adult restoration of Shank3 expression rescues selective autistic-like phenotypes. Nature. 2016;530:481–4.

    CAS  Article  Google Scholar 

  23. 23.

    Zhao H, Tu Z, Xu H, Yan S, Yan H, Zheng Y, et al. Altered neurogenesis and disrupted expression of synaptic proteins in prefrontal cortex of SHANK3-deficient non-human primate. Cell Res. 2017;27:1293–7.

    CAS  Article  Google Scholar 

  24. 24.

    Wang X, McCoy PA, Rodriguiz RM, Pan Y, Je HS, Roberts AC, et al. Synaptic dysfunction and abnormal behaviors in mice lacking major isoforms of Shank3. Hum Mol Genet. 2011;20:3093–108.

    CAS  Article  Google Scholar 

  25. 25.

    Bey AL, Wang X, Yan H, Kim N, Passman RL, Yang Y, et al. Brain region-specific disruption of Shank3 in mice reveals a dissociation for cortical and striatal circuits in autism-related behaviors. Transl Psychiatry. 2018.

  26. 26.

    Bozdagi O, Sakurai T, Papapetrou D, Wang X, Dickstein DL, Takahashi N, et al. Haploinsufficiency of the autism-associated Shank3 gene leads to deficits in synaptic function, social interaction, and social communication. Mol Autism. 2010;1:15.

    CAS  Article  Google Scholar 

  27. 27.

    Jiang Yhui, Ehlers MD. Modeling autism by SHANK gene mutations in mice. Neuron. 2013;78:8–27.

    CAS  Article  Google Scholar 

  28. 28.

    Silbereis JC, Pochareddy S, Zhu Y, Li M, Sestan N. The cellular and molecular landscapes of the developing human central nervous system. Neuron. 2016;89:248–68.

    CAS  Article  Google Scholar 

  29. 29.

    Shcheglovitov A, Shcheglovitova O, Yazawa M, Portmann T, Shu R, Sebastiano V, et al. SHANK3 and IGF1 restore synaptic deficits in neurons from 22q13 deletion syndrome patients. Nature. 2013.

  30. 30.

    Kathuria A, Nowosiad P, Jagasia R, Aigner S, Taylor RD, Andreae LC, et al. Stem cell-derived neurons from autistic individuals with SHANK3 mutation show morphogenetic abnormalities during early development. Mol Psychiatry. 2018.

  31. 31.

    Yi F, Danko T, Botelho SC, Patzke C, Pak C, Wernig M, et al. Autism-associated SHANK3 haploinsufficiency causes Ih channelopathy in human neurons. Science. 2016.

  32. 32.

    Gouder L, Vitrac A, Goubran-Botros H, Danckaert A, Tinevez JY, André-Leroux G, et al. Altered spinogenesis in iPSC-derived cortical neurons from patients with autism carrying de novo SHANK3 mutations. Sci Rep. 2019.

  33. 33.

    Chambers SM, Fasano CA, Papapetrou EP, Tomishima M, Sadelain M, Studer L. Highly efficient neural conversion of human ES and iPS cells by dual inhibition of SMAD signaling. Nat Biotechnol. 2009.

  34. 34.

    Ran FA, Hsu PD, Wright J, Agarwala V, Scott DA, Zhang F. Genome engineering using the CRISPR-Cas9 system. Nat Protoc. 2013.

  35. 35.

    Miyamichi K, Amat F, Moussavi F, Wang C, Wickersham I, Wall NR, et al. Cortical representations of olfactory input by trans-synaptic tracing. Nature. 2011.

  36. 36.

    Viswanathan S, Williams ME, Bloss EB, Stasevich TJ, Speer CM, Nern A, et al. High-performance probes for light and electron microscopy. Nat Methods. 2015.

  37. 37.

    Kim EJ, Jacobs MW, Ito-Cole T, Callaway EM. Improved monosynaptic neural circuit tracing using engineered rabies virus glycoproteins. Cell Rep. 2016.

  38. 38.

    Murray E, Cho JH, Goodwin D, Ku T, Swaney J, Kim SY, et al. Simple, scalable proteomic imaging for high-dimensional profiling of intact systems. Cell. 2015.

  39. 39.

    Chiola S, Do MD, Centrone L, Mallamaci A. Foxg1 overexpression in neocortical pyramids stimulates dendrite elongation via Hes1 and pCreb1 upregulation. Cereb Cortex. 2018.

  40. 40.

    Jin C, Kang HR, Kang H, Zhang Y, Lee Y, Kim Y, et al. Unexpected compensatory increase in Shank3 transcripts in Shank3 knock-out mice having partial deletions of exons. Front Mol Neurosci. 2019.

  41. 41.

    Ching TT, Maunakea AK, Jun P, Hong C, Zardo G, Pinkel D, et al. Epigenome analyses using BAG microarrays identify evolutionary conservation of tissue-specific methylation of SHANK3. Nat Genet. 2005.

  42. 42.

    Maunakea AK, Nagarajan RP, Bilenky M, Ballinger TJ, Dsouza C, Fouse SD, et al. Conserved role of intragenic DNA methylation in regulating alternative promoters. Nature. 2010.

  43. 43.

    Ährlund-Richter S, Xuan Y, van Lunteren JA, Kim H, Ortiz C, Pollak Dorocic I, et al. A whole-brain atlas of monosynaptic input targeting four different cell types in the medial prefrontal cortex of the mouse. Nat Neurosci. 2019.

  44. 44.

    Brennand KJ, Simone A, Jou J, Gelboin-Burkhart C, Tran N, Sangar S, et al. Modelling schizophrenia using human induced pluripotent stem cells. Nature. 2011.

  45. 45.

    Sarkar A, Mei A, Paquola ACM, Stern S, Bardy C, Klug JR, et al. Efficient generation of CA3 neurons from human pluripotent stem cells enables modeling of hippocampal connectivity in vitro. Cell Stem Cell. 2018.

  46. 46.

    Osakada F, Callaway EM. Design and generation of recombinant rabies virus vectors. Nat Protoc. 2013.

  47. 47.

    Maisano X, Litvina E, Tagliatela S, Aaron GB, Grabel LB, Naegele JR. Differentiation and functional incorporation of embryonic stem cell-derived GABAergic interneurons in the dentate gyrus of mice with temporal lobe epilepsy. J Neurosci. 2012.

  48. 48.

    Cunningham M, Cho JH, Leung A, Savvidis G, Ahn S, Moon M, et al. hPSC-derived maturing GABAergic interneurons ameliorate seizures and abnormal behavior in epileptic mice. Cell Stem Cell. 2014.

  49. 49.

    Steinbeck JA, Choi SJ, Mrejeru A, Ganat Y, Deisseroth K, Sulzer D, et al. Optogenetics enables functional analysis of human embryonic stem cell-derived grafts in a Parkinson’s disease model. Nat Biotechnol. 2015.

  50. 50.

    Adler AF, Cardoso T, Nolbrant S, Mattsson B, Hoban DB, Jarl U, et al. hESC-derived dopaminergic transplants integrate into basal ganglia circuitry in a preclinical model of Parkinson’s disease. Cell Rep. 2019.

  51. 51.

    Espuny-Camacho I, Arranz AM, Fiers M, Snellinx A, Ando K, Munck S, et al. Hallmarks of Alzheimer’s disease in stem-cell-derived human neurons transplanted into mouse brain. Neuron. 2017.

  52. 52.

    Linaro D, Vermaercke B, Iwata R, Ramaswamy A, Libé-Philippot B, Boubakar L, et al. Xenotransplanted human cortical neurons reveal species-specific development and functional integration into mouse visual circuits. Neuron. 2019;104:972.e6.

    CAS  Article  Google Scholar 

  53. 53.

    Willsey AJ, Sanders SJ, Li M, Dong S, Tebbenkamp AT, Muhle RA, et al. XCoexpression networks implicate human midfetal deep cortical projection neurons in the pathogenesis of autism. Cell. 2013.

  54. 54.

    Parikshak NN, Luo R, Zhang A, Won H, Lowe JK, Chandran V, et al. Integrative functional genomic analyses implicate specific molecular pathways and circuits in autism. Cell. 2013;155:1008–21.

    CAS  Article  Google Scholar 

  55. 55.

    Koshino H, Kana RK, Keller TA, Cherkassky VL, Minshew NJ, Just MA. fMRI investigation of working memory for faces in autism: visual coding and underconnectivity with frontal areas. Cereb Cortex. 2008.

  56. 56.

    Weng SJ, Wiggins JL, Peltier SJ, Carrasco M, Risi S, Lord C, et al. Alterations of resting state functional connectivity in the default network in adolescents with autism spectrum disorders. Brain Res. 2010.

  57. 57.

    Ibrahim K, Eilbott JA, Ventola P, He G, Pelphrey KA, McCarthy G, et al. Reduced amygdala–prefrontal functional connectivity in children with autism spectrum disorder and co-occurring disruptive behavior. Biol Psychiatry Cogn Neurosci Neuroimaging. 2019.

  58. 58.

    Grealish S, Heuer A, Cardoso T, Kirkeby A, Jönsson M, Johansson J, et al. Monosynaptic tracing using modified rabies virus reveals early and extensive circuit integration of human embryonic stem cell-derived neurons. Stem Cell Reports. 2015.

  59. 59.

    Denardo LA, Berns DS, Deloach K, Luo L. Connectivity of mouse somatosensory and prefrontal cortex examined with trans-synaptic tracing. Nat Neurosci. 2015.

  60. 60.

    Feng D, Lau C, Ng L, Li Y, Kuan L, Sunkin SM, et al. Exploration and visualization of connectivity in the adult mouse brain. Methods. 2015.

  61. 61.

    Scannevin RH, Huganir RL. Postsynaptic organisation and regulation of excitatory synapses. Nat Rev Neurosci. 2000.

  62. 62.

    Nusser Z, Lujan R, Laube G, Roberts JDB, Molnar E, Somogyi P. Cell type and pathway dependence of synaptic AMPA receptor number and variability in the hippocampus. Neuron. 1998.

  63. 63.

    Matsuzaki M, Ellis-Davies GCR, Nemoto T, Miyashita Y, Iino M, Kasai H. Dendritic spine geometry is critical for AMPA receptor expression in hippocampal CA1 pyramidal neurons. Nat Neurosci. 2001.

  64. 64.

    Yuste R, Bonhoeffer T. Genesis of dendritic spines: Insights from ultrastructural and imaging studies. Nat Rev Neurosci. 2004;5:24–34.

    CAS  Article  Google Scholar 

  65. 65.

    Li Z, Sheng M. Some assembly required: the development of neuronal synapses. Nat Rev Mol Cell Biol. 2003;4:833–41.

    CAS  Article  Google Scholar 

  66. 66.

    Kolluri N, Sun Z, Sampson AR, Lewis DA Lamina-specific reductions in dendritic spine density in the prefrontal cortex of subjects with schizophrenia. Am J Psychiatry. 2005.

  67. 67.

    Lewis DA, González-Burgos G. Neuroplasticity of neocortical circuits in schizophrenia. Neuropsychopharmacology. 2008;33:141–65.

    Article  Google Scholar 

  68. 68.

    Glantz LA, Lewis DA. Decreased dendritic spine density on prefrontal cortical pyramidal neurons in schizophrenia. Arch Gen Psychiatry. 2000.

  69. 69.

    Jacobs B, Schall M, Prather M, Kapler E, Driscoll L, Baca S, et al. Regional dendritic and spine variation in human cerebral cortex: a quantitative golgi study. Cereb Cortex. 2001.

  70. 70.

    Petanjek Z, Judaš M, Šimić G, Rašin MR, Uylings HBM, Rakic P, et al. Extraordinary neoteny of synaptic spines in the human prefrontal cortex. Proc Natl Acad Sci USA. 2011.

  71. 71.

    Roussignol G, Ango F, Romorini S, Tu JC, Sala C, Worley PF, et al. Shank expression is sufficient to induce functional dendritic spine synapses in aspiny neurons. J Neurosci. 2005.

  72. 72.

    Pina-Crespo JC, Talantova M, Cho E-G, Soussou W, Dolatabadi N, Ryan SD, et al. High-frequency hippocampal oscillations activated by optogenetic stimulation of transplanted human ESC-derived neurons. J Neurosci. 2012.

  73. 73.

    Nagashima F, Suzuki IK, Shitamukai A, Sakaguchi H, Iwashita M, Kobayashi T, et al. Novel and robust transplantation reveals the acquisition of polarized processes by cortical cells derived from mouse and human pluripotent stem cells. Stem Cells Dev. 2013.

  74. 74.

    Espuny-Camacho I, Michelsen KA, Gall D, Linaro D, Hasche A, Bonnefont J, et al. Pyramidal neurons derived from human pluripotent stem cells integrate efficiently into mouse brain circuits in vivo. Neuron. 2013.

  75. 75.

    Oki K, Tatarishvili J, Wood J, Koch P, Wattananit S, Mine Y, et al. Human-induced pluripotent stem cells form functional neurons and improve recovery after grafting in stroke-damaged brain. Stem Cells. 2012.

  76. 76.

    Vicidomini C, Ponzoni L, Lim D, Schmeisser M, Reim D, Morello N, et al. Pharmacological enhancement of mGlu5 receptors rescues behavioral deficits in SHANK3 knock-out mice. Mol Psychiatry. 2017;22:689–702.

    CAS  Article  Google Scholar 

Download references


We thank Ricardo Dolmetsch and Theo Palmer for providing iPSC lines; Yuanyuan Wu for helping with generating homozygous CRISPR/Cas9-engineered stem cells; Travis Philyaw for helping with mapping deletions; Brittney Nhem and Nico Edgar for helping with image analysis; Megan Williams, Edward Callaway, Liqun Luo, and John Naughton for plasmids; Chris Rodesch and the University of Utah Cell Imaging Core for assisting with microscopy, Jay Spampanato and Peter West for advising on intracranial injections and slice physiology; Monica Vetter, Richard Dorsky, and Sungjin Park for commenting on the manuscript. This work was supported by grants from the NIMH, NINDS, Brain Research Foundation, Brain and Behavior Research Foundation, Whitehall Foundation, Alan B. Slifka Foundation, and Utah Neuroscience and Genome Project Initiatives.

Author information



Corresponding author

Correspondence to Aleksandr Shcheglovitov.

Ethics declarations

Conflict of interest

The authors declare that they have no conflict of interest.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Chiola, S., Napan, K.L., Wang, Y. et al. Defective AMPA-mediated synaptic transmission and morphology in human neurons with hemizygous SHANK3 deletion engrafted in mouse prefrontal cortex. Mol Psychiatry 26, 4670–4686 (2021).

Download citation

Further reading


Quick links