Clinical phenotypes of these three subjects with 15q13.3 duplication are summarized in Table 1. The AP’s pregnancy was planned and there was no history of in utero exposure to drugs, alcohol, or tobacco. There were no complications during the pregnancy, and the AP was delivered normally at full term, with a birth weight of seven pounds, six ounces. As an infant, the AP was delayed in sleeping through the night, which did not begin to occur until after 1 year of age. He has no history of delay in motor development but has a history of significant delay in language development, producing his first word after the age of 2 years. He started speech therapy at 2 years of age, resulting in rapid improvement in his language development. The AP was 12 years old when this study was initiated, with a history of ADHD, depression, and ASD. Prior to age five, he reportedly did not make eye contact and did not exhibit age-appropriate social reciprocity. While in the third grade, the AP had frequent crying episodes and was overwhelmed by homework, which brought him to psychological evaluation. At that time, he manifested both autistic features and pronounced mood lability on exam, which was manifested on several occasions by the child becoming morose and tearful when told even slightly sad stories that would not have elicited such a reaction in a typical child his age. During periods of depressive symptomatology, his mother reported that he had difficulty falling asleep, low appetite, and decreased interest in his favorite activities, such as sports. In middle school, he manifested poor concentration and difficulty with time management, standardized test-taking, and organizing tasks and activities. These issues were treated by using cognitive behavioral therapy and play therapy. He was subsequently treated with sertraline, followed by escitalopram. On these selective serotonin reuptake inhibitors, the AP’s anxiety was significantly lessened, but he then experienced a residual lack of motivation and his perseverative traits were not improved by treatment. Based on his developmental history, including language delay, impairment in social reciprocity and non-verbal communication, and repetitive thinking, it was clinically determined that a significant contribution to his overall impairment was autistic perseveration and rigidity, for which a trial of risperidone was initiated and resulted in significant clinical improvement over the ensuing years. Ultimately, he was successfully weaned from risperidone and was reasonably well adapted in high school. The AP’s mother reported a history of mild depression, anxiety, and obsessive-compulsive traits, while the AP’s 8-year-old brother had subtle autistic traits that were less pronounced than those of the AP and behavioral features of emotional dysregulation that were more pronounced than those of the AP; he met criteria for disruptive mood dysregulation disorder, ADHD, and mood disorder.
Cytogenetics Microarray (CMA) analysis was performed for research testing by the Washington University Cytogenetics and Molecular (CMP) Pathology Laboratory, using the Affymetrix CytoScan HD array. This array includes 2.6 million copy number markers, 1.9 million non-polymorphic probes, and nearly 750,000 single-nucleotide polymorphism (SNP) probes. Average intragenic marker spacing is equivalent to 1 probe per 880 basepairs. Analysis of these data by the CMP Laboratory, after alignment to hg19, defined a 424 kb gain at 15q13.3 in samples from the affected proband (AP) and unaffected mother (UM) and a 444 kb gain in the same location in the affected brother. This copy number variant was not present in the father.
The Washington University Genome Engineering and Induced Pluripotency Center (GEiC) derived multiple clonal iPSC lines from individuals in this family. Briefly, fresh urine samples were procured from the AP and UM and were used to obtain renal epithelial cells, which were reprogrammed using the CytoTune-iPSC 2.0 Sendai virus-based reprogramming kit (Thermo Fischer Scientific), as per the manufacturer’s instructions. iPSC clones were picked and three clonal cell lines were derived from each study subject. Clones number 1 and 3 from each subject were used for experimentation. These newly derived lines were compared with established male and female iPSC control lines from unrelated individuals (BJFF6 and AN1.1, denoted UC-M and UC-F in this study) provided by the Washington University Genome Engineering and Induced Pluripotency Center.
iPSC cultures and differentiation
iPSC line derivation and directed differentiation were performed by modification of our previously described methods [22, 85]. iPSC lines were grown on Matrigel (Corning) under feeder-free conditions using mTeSR1 (STEMCELL Technologies), authenticated by STR profiling, and tested for mycoplasma contamination regularly during culture. For directed differentiation to generate cortical excitatory neural progenitor cells (cExNPCs), iPSCs were dissociated into single cells with Accutase (Life Technologies) and 40,000 cells were seeded in V-bottom 96 well non-adherent plates (Corning). Embryoid bodies (EBs) were generated by centrifugation of the plate at 200xg for 5 min and were then incubated in 5% CO2 at 37°C, in cExNPC differentiation medium with 10μM Y-27632 (Tocris Biosciences). cExNPC differentiation medium includes Neurobasal-A (Life Technologies), 1X B-27 supplement without Vitamin A (Life Technologies), 10μM SB-431542 (Tocris Biosciences), and 100nM LDN-193189 (Tocris Biosciences). On day 4, EBs were picked with wide bore P1000 tips and were transferred to Poly-l-Ornithine- (20μg/ml) and laminin- (10μg/ml) coated plates. Every other day media without Y-27632 was replenished, and on day 15 Neural Rosette Selection Reagent (STEMCELL Technologies) was used to isolate cExNPCs from rosettes, as per the manufacturer’s instructions. cExNPCs were grown as a monolayer using cExNPC differentiation media up to 15 passages.
Cortical inhibitory neural progenitor cells (cINPCs) were generated by directed differentiation in media which included Neurobasal-A (Life Technologies), 1X B-27 supplement without Vitamin A (Life Technologies), 10μM SB-431542 (Tocris Biosciences), 100nM LDN-193189 (Tocris Biosciences), 1μM Purmorphamine (Calbiochem), and 2μM XAV-939 (Tocris Biosciences). Y-27632 was also included in this media until day eight. For cINPC differentiation, EBs were generated as described above for cExNPC differentiation. On day 4, EBs were transferred to non-adherent plates and were placed on an orbital shaker at 80rpm in an incubator with 5% CO2 and 37°C. cINPC media were replenished every other day, and on day 10, EBs were transferred to Matrigel- and laminin- (5μg/ml) coated plates. On day 15, cINPCs were dissociated with Accutase and were either cryopreserved and/or grown in monolayer culture on Matrigel- and laminin-coated for up to 15 passages. For analysis of both cExNPC and cINPC growth properties, equal numbers of cells for each line were seeded on Matrigel- and laminin- (5μg/ml) coated plates and the total number of cells was counted after 4 days.
For differentiation and maturation of neurons, cortical neuroids were generated by seeding both 2X104 cExNPCs and 2X104 cINPCs into each well of a V-bottom 96 well non-adherent plate in maturation media. Plates were spun at 200xg for 5 min and were incubated in 5% CO2 at 37°C in maturation media, with addition of Y-27632 for the first 4 days of culture. The composition of maturation media includes Neurobasal-A and 1X B-27 supplement without vitamin A, while DAPT (10μM; Tocris) was included in the media from day 7 to day 11, and 200μM ascorbic acid (Sigma Aldrich), 20ng/ml BDNF (Peprotech), and 200μM cAMP were included in the media from day 11 to day 15. On day 4, neuroids were transferred to non-adherent plates and were placed on an orbital shaker at 80rpm in an incubator with 5% CO2 and 37°C. On day 5, neuroids were moved to Matrigel- and laminin- (5μg/ml) coated plates and further incubated in 5% CO2 at 37°C. Media were replenished every other day until day 15.
Immunocytochemistry (ICC) and immunoblotting
ICC and immunoblotting experiments were performed as previously described . In brief, for ICC, cortical neuroids were dissociated after 15 days of maturation and cells were plated in eight-well chamber slides coated with Matrigel and laminin (5μg/ml). After 24 h, cells were washed with PBS without calcium and magnesium and were fixed in 4% paraformaldehyde for 15–20 min. See  for detailed protocol. Primary and secondary antibodies used for these experiments and for immunoblotting are provided in Additional file 4. Images were taken using a spinning-disk confocal microscope (Quorum) and an Olympus inverted microscope using MetaMorph software. ImageJ was used to process images and for quantification: 15–20 random fields were imaged from three to five biological replicate experiments, which included work with two different clones per sample type for the AP and UM models and one clone for the UC-M/F models; total numbers of both immune-positive and all DAPI stained cell nuclei quantified are shown in Additional file 4. For each experimental finding in this manuscript, the number of biological replicate experiments and the clones used for each replicate of each type of experiment is also summarized in Additional file 4.
Approximately one million cExNPCs or cINPCs between passages 4–9 were used for FACS analysis, and experiments were performed as described previously . The plot shows the median value gathered from seven biological replicates, using two different clones per sample type for the AP and UM models and one clone for the UC-M/F models, calculated by using a Kruskal-Wallis non-parametric test. P values: *P < 0.05, **P < 0.01, and ***P < 0.001.
RNA-Sequencing and RT-qPCR
After 15 days of cortical neuroid differentiation as described above, total RNA was collected from the AP, UM, UC-M, and UC-F lines, using the NucleoSpin RNA II kit (Takara) per the manufacturer’s instructions. RNA was quantified using a NanoDrop ND-1000 spectrophotometer (Thermo Scientific) and the Agilent Bioanalyzer 2100 was used to assess RNA integrity, with only samples with an RNA Integrity Number of >8 used for sequencing and analysis. RNA-Sequencing (RNA-Seq) library preparation and Illumina Sequencing were performed by the Genome Technology Access Center at Washington University. The Illumina Hi-Seq3000 was used to obtain single-end 50 base pair reads, with approximately 30 million unique reads per sample obtained after alignment. Four independent biological replicates per cell line were analyzed by RNA-Seq. For RT-qPCR, 1μg total RNA was reverse transcribed using iScript Reverse Transcription Supermix (Bio-Rad) and equal quantities of cDNA were used as a template for RT-qPCR using the Applied Biosystems Fast Real-Time quantitative PCR platform. GAPDH or RPL30 were used as endogenous controls for normalization. Four biological replicate experiments using one clonal line for each sample type were used for RNA-seq analysis (n = 4), while a second clonal line for the UM and AP models was used to generate RNA for RT-qPCR validation of a subset of the RNA-Seq findings. P values for RT-qPCR validation: *P < 0.05, **P < 0.01, and ***P < 0.001 were determined by unpaired t testing.
Bioinformatics and IPA analysis
RNA-Seq data analysis was performed as described in [22, 85] to obtain differentially regulated genes (DEG). Briefly, STAR version 2.5.4b was used to align the RNA-Seq reads to the human genome assembly hg38 . To derive uniquely aligned unambiguous reads, we used Subread:featureCount, version 1.6.3 with GENCODE gene annotation  and gene-level transcripts were imported into the R-bioconductor package . After excluding genes expressed at <1.0 counts per million (CPM), differentially expressed genes (DEG) were curated based upon a Log2 fold change >1 and a Benjamini and Hochberg FDR of <0.05. DEGs were used to perform hierarchical clustering analysis using ClustVis  and to perform Ingenuity Pathway Analysis (IPA) (Qiagen), as described previously . To determine the contribution of different covariates to gene expression, we also performed variancePartition analysis, including the individual sample types (UC-M, UC-F, UM, and AP), age (young and old), and sex (male and female) as variables .
To measure neurite extension, cortical neuroids were generated and cultured to promote differentiation and maturation as described above. On day 6, images were acquired using an inverted light microscope, with data collected for three or more independent biological replicate experiments encompassing work with two clonal lines per subject for the AP and UM, and one clonal line for the UC-F and UC-M. Neurite extension length from adherently plated neuroids was measured using ImageJ, as the distance between two circles drawn at the border of the plated neuroid and at the tips of neurites extending from that neuroid. Since neurites extend in all directions from each plated neuroid, each datapoint is a calculation of neurite length from the EB border to the tip of the neurites, calculated on a per-EB basis and normalized by the number of EBs quantified per sample type (e.g., UC-M). The plot shows the median value calculated from 50–80 data points per sample type, with data points gathered from seven biological replicates.
ICC for MAP2 in adherent neuroids was conducted as described above, with image acquisition using a spinning-disk (Quorum) confocal microscope and an Axiovision inverted microscope. Day 15 neuroids were also dissociated and the neurons plated on Matrigel- and laminin-coated plated and stained with MAP2.
To assess neurite length, images were acquired using a spinning-disk (Quorum) confocal microscope and an Axiovision inverted microscope. Images were processed with Imaris software (Bitplane) and neurite length was measured using the filament tracer application and normalized to the number of nuclei stained with DAPI in the neuroid, which was measured with the particle application in Imaris. The neuronal soma area was measured from MAP2 stained images using ImageJ. The data points are normalized on a per neuron basis, with neurons quantified from at least 20 different images (with at least five neurons measured per image) per biological replicate, with a total of three independent biological replicates from two clones analyzed.
To quantify VGAT- and VGLUT-expressing punctae, dissociated and plated neuroids generated as described above were immunostained for the respective antibodies, and images were taken using a spinning-disk (Quorum) confocal microscope and Axiovision inverted microscope. Punctae were measured using the synaptic counter plugin in ImageJ. Each finding was obtained in three or more independent biological replicate experiments encompassing work with two clonal lines per subject for the AP and UM, and one clonal line for the UC-M.
To study the migration of cExN and cIN neurons, we developed an approach that utilized fused co-culture of two 3D spheres consisting of cExNs and cINs. These 3D spheres were generated by transducing cExNPCs and cINPCs separately with either a lentiviral synapsin-eGFP or a synapsin-RFP expression construct, respectively. 30,000 of these transduced cExNPCs or cINPCs per well were then seeded into separate wells of a V bottom 96 well plate in 100μl of maturation media containing 10μM Y-27632. The V bottom plate was centrifuged at 200xg for 5 min at room temperature and then incubated in 5% CO2 at 37°C. On day 2, 50μl of media was replaced with fresh media without disturbing the spheres. On day 4, cExN and cIN EBs were selected with wide bore P1000 tips and moved to a U bottom plate. One cExN and one cIN sphere were placed side by side in each U bottom well. Placement of spheres in close apposition caused them to undergo fusion without further manipulation, enabling assessment of neuronal migration. On day 6, these fused spheres were moved with wide bore P1000 tips to a coverslip placed in a 3-cm plate and coated with matrigel and laminin (5μg/ml). On day 10, images were acquired using a spinning-disk (Quorum) confocal microscope and Axiovision inverted microscope and image processing was performed using ImageJ. Migration was assessed in such fused co-cultures. The ability of Wnt signaling to rescue migration deficits observed in the AP line was tested by the addition of 10μM CHIR-99021 in DMSO, with parallel treatment of control spheres with equal quantities of DMSO. Three or more independent biological replicate experiments were performed in two clonal lines for the AP and UM, and in one clonal line for the UC-F and UC-M. Each data point is the number of neurons that migrated from the cIN into cExN EB or from the cExN into the cIN EB, normalized to each fused EB (e.g., EB pair). Each data point represents the number of migrating cExNs per fused EB or the number of cINs per fused EB, with the median value calculated from ~40–80 fused EBs per sample type. The plot shows the median value calculated from 40 to 80 data points per sample type, with data points gathered from six biological replicates, using two different clones per sample type. P values were calculated by using a Kruskal-Wallis non-parametric test. *P < 0.05, **P < 0.01, and ***P < 0.001.
ER stress luciferase assay
To test the effects of endoplasmic reticulum (ER) stress on cINPCs, we used an expression construct encoding a stress sensor . This construct encodes a Guassia luciferase protein fusion, with replacement of the first 18 amino acids with the signal peptide from the mesencephalic astrocyte-derived neurotrophic factor (MANF) protein and carboxy-terminal fusion to MANF’s final 5 amino acids, which encode a stress sensor. To detect ER stress, 35,000 cINPCs were seeded on Matrigel- and laminin- (5μg/ml) coated 96 well plates in cIN differentiation media containing Y-27632. After 24 h, the cells were transfected with the stress sensor expression construct using FuGENE 6 (Promega) transfection reagent. 48 h after transfection, 50μl of supernatant was removed and assayed for luciferase activity using the BioLux Gaussia Luciferase Assay Detection System (New England Biolabs). For rescue experiments, small molecules were obtained from Sigma Aldrich or Tocris Biosciences, reconstituted in DMSO or PBS-Ca2+/Mg2+, and added in the medium after 24 h of transfection, at the final concentrations indicated (Tudca-50μM, PBA-500μM, Dantrolene sodium-1μM, and JTV-519-10μM). Luciferase levels were measured after 48 h of small molecule treatment as above. Luciferase data for rescue experiments performed in the presence of small molecules were normalized to the DMSO control, and three or more independent biological replicate experiments were performed, using two clonal lines for the AP and UM and one clonal line for the UC-F and UC-M.
cExNPCs and cINPCs were transduced with Synapsin-GFP and Synapsin-RFP expression constructs, respectively, before performing cortical neuroid maturation. At day 0 of cortical neuroid differentiation, 2X104 cExNPCs and 2X104 cINPCs were mixed in each well of a V-bottom 96-well non-adherent plate in maturation media. Maturation followed the approach described above. At day 15, neuroids were dissociated using Accutase (Life Technologies), were seeded onto a layer of rat cortical astrocytes, prepared as described previously , and were grown for another three weeks using Neurobasal-A, 1XB27 with vitamin A (Life Technologies), and supplementation with BDNF (20ng), cAMP (200μM), and ascorbic acid (200μM). iPSC-derived neuron cultures were perfused at 1 ml/min with room temperature (22°C) Tyrode’s solution (in mM): 150 NaCl, 4 KCl, 2 MgCl2, 2 CaCl2, 10 glucose, 10 HEPES, with pH adjusted to pH 7.4 with NaOH. Recording electrodes had an open-tip resistance of 2–6 MOhm when filled with (in mM): 140 K-glucuronate, 10 NaCl, 5 MgCl2, 0.2 EGTA, 5 Na-ATP, 1 Na-GTP, and 10 HEPES, pH adjusted to 7.4 with KOH. Whole-cell currents and membrane potentials were recorded with an Axopatch 200A amplifier (Molecular Devices). Voltage clamp recordings were used to determine cell capacitance and input resistance as well as peak inward sodium current and steady-state outward potassium currents during depolarizing voltage steps from a holding potential of −80 mV . Current clamp recordings and currents evoked by choline and acetylcholine (ACh) were obtained in a modified extracellular solution (in mM): 120 NaCl, 3 KCl, 10 glucose, 1 NaH2PO4, 4 NaHCO3, 5 HEPES, pH adjusted to 7.4 with NaOH, delivered from an 8-barrelled local perfusion pipette positioned very close to the recorded cell in order to minimize desensitization. Although desensitization can be further reduced by recording from outside-out patches that enable faster solution exchange, the present study employed whole-cell recordings in order to allow for current and voltage-clamp analysis of cellular physiology as well as agonist-gated currents all within the same cells. In future experiments, higher speed agonist delivery would allow for a more detailed kinetic comparison among the genotypes.
All statistical analyses were performed using IBM SPSS Statistics (v.27) or Sigma-STAT. Prior to analyses, data was screened for missing values and fit of distributions with assumptions underlying univariate analysis. This included the Shapiro-Wilk test on z-score-transformed data and qqplot investigations for normality, Levene’s test for homogeneity of variance, and boxplot and z-score (±3.29) investigation for identification of influential outliers. Means and standard errors were computed for each measure. Non-parametric Kruskal-Wallis and Mann-Whitney U tests were used to analyze data. The critical alpha value for all analyses was P <0.05, unless otherwise stated. Multiple pairwise comparisons were subjected to Bonferroni correction, where appropriate. The datasets generated and analyzed during the current study are available from the corresponding author upon reasonable request.