Autism Spectrum Disorder Induced Pluripotent Stem Cells Display Dysregulated Calcium Signaling During Neural Differentiation

Stem cells were utilized in accordance with the guidelines and regulations of King Faisal Specialist Hospital and Research Centre (KFSH&RC) and with the approval of the Office of Research Affairs Research Ethics Committee and Basic Research Committee (Project no: 2190006 and approval reference no: C380/345/41). Two ASD iPSC lines (SC119 and SC125) and two healthy iPSC controls (007 and FAM) were included in this study (Table 1). The two idiopathic ASD iPSC lines were generously provided by the University of Melbourne, Melbourne, Australia, under an approved and signed material transfer agreement. Healthy iPSC line 007 was kindly provided by Prof. Alice Pébay (UoM, Melbourne, Australia) [21,22], and iPSC line FAM was kindly provided by Dr. Reem Alhejilan (KFSH&RC, KSA, Riyadh, Saudi Arabia) [23]. Cells were cultured using the feeder free system on vitronectin-coated 60 mm dishes (Stem Cell Technologies, Vancouver, BC, Canada, CAT#: 07180), maintained in TeSR™-E8 defined medium as per the manufacturer's instructions (Stem Cell Technologies, CAT#: 05990). Colonies were mechanically dissected on a weekly basis and replated on freshly coated plates to maintain their pluripotency, and they were regularly assessed for pluripotency by morphology and expression of OCT4. Cells were maintained at 37 °C under 5% CO₂ and medium changes were performed daily.

Fibroblasts of ASD iPSC lines were originally obtained from Dr. Philip Schwartz (Children's Hospital of Orange County Research Institute, Orange, CA, USA). ASD Fibroblasts were reprogrammed in the Dottori laboratory using the LONZA Transfection System (Lonza Group AG, Basel, Switzerland, CAT#: VP100) with non-viral episomal vectors provided by the Epi5 Episomal iPSC Reprogramming Kit (Life Technologies, Carlsbad, CA, USA, CAT#: A15960). The generated ASD iPSC lines were validated for pluripotency through the expression of key markers using the immunofluorescence assay (). They were also assessed for their differentiation potential toward endoderm, mesoderm, and ectoderm derivates (). Figure S1 Figure S2

For endoderm differentiation, embryoid bodies (EBs) were generated from both ASD lines and harvested at day 14 for RT-PCR analysis. Total RNA was extracted using the PureLink RNA Kit according to the manufacturer's instructions (Life Technologies, CAT#:12183025), and one microgram of RNA was used to synthesize first-strand cDNA using the Sensifast cDNA Synthesis Kit (Bioline, Cincinnati, OH, USA, CAT#; BIO-65054), according to the manufacturer's instructions, under the following conditions: 25 °C for 10 min (primer annealing), 42 °C for 15 min (reverse transcription), 85 °C for 5 min (inactivation), and 4 °C hold. Each final reaction contained 1 µg cDNA in a volume of 20 µL. The GATA6 transcript—an endoderm marker—was amplified using the OneTaq RT/PCR Kit (BioLabs, Boston, MA, USA, CAT#: E5310S), following the manufacturer's instructions, and visualized using a FluorChem E blot analyzer (ProteinSimple, San Jose, CA, USA). The following primers sequences were used: GATA6 (F: GCC TCA CTC CAC TCG TGT CT; R: TCA GAT CAG CCA CAC AAT ATG A) and BETA ACTIN (F: CAC CAC ACC TTC TAC AAT GAG C; R: TCG TAG ATG GGC ACA GTG TGG G).

For mesoderm differentiation, both ASD lines were differentiated toward cardiac specification. IPSCs were dissociated into single cells with TrypLE (Thermo Fisher Scientific, Waltham, MA, USA) and seeded onto plates coated with hESC-qualified Matrigel (Corning, Corning, NY, USA) at a density of 1 × 10⁵ cells/cm² in TeSR-E8 medium supplemented with 10 µM Y-27632 (Tocris Bioscience, Bristol, UK). After 2 days, which is referred to as day 0, the medium was replaced with RPMI 1640 basal medium containing B-27 without insulin supplement (Thermo Fisher Scientific), growth factor-reduced Matrigel (1:60 dilution), and 10 µM CHIR99021 (Cayman Chemical, Ann Arbor, MI, USA). After 24 h, cells were initially treated with 5 ng/mL Activin A (Peprotech, Cranbury, NJ, USA) in RPMI 1640 basal medium containing B-27 without insulin supplement for 24 h. On day 2, the medium was changed to RPMI 1640 basal medium containing B-27 without insulin supplement and 5 µM IWP2 (Tocris Bioscience) for 72 h. Starting from day 5, cultures were maintained in RPMI 1640 supplemented with B-27 (Thermo Fisher Scientific) and 200 µg/mL L-ascorbic acid 2-phosphate sesquimagnesium salt hydrate (Sigma-Aldrich, St. Louis, MO, USA). On day 14, the samples were fixed for immunofluorescence analysis.

For ectoderm differentiation, both ASD iPSC lines were differentiated toward neuroectoderm specification as described below (see). Neurospheres (NSPs) at day 28 of differentiation were fixed for the immunofluorescence assay. Section 2.2

For neural induction and differentiation of iPSC lines, the dual SMAD inhibition protocol was adopted as previously described by Denham and Dottori, with some slight modifications (Figure 1A) [24]. Briefly, iPSC lines were mechanically dissected into pieces approximately 0.5 mm in width and transferred onto laminin-coated organ culture plates in N2B27 medium containing 1:1 mix of neurobasal medium with DMEM/F12 medium, supplemented with 1% insulin transferrin selenium, 1% N2, 1% retinol-free B27, 0.3% glucose, 25 U/mL penicillin, and 25 μg/mL streptomycin (Thermo Fisher Scientific, CAT#: 11320033, A3582901, 17502-048, 12587-010, 51300-044, 25030-081, and 15070063, respectively). For induction of cortical neurons, the small molecule inhibitor SB431542 (10 μM, Tocris, CAT#: 1614) and LDN (100 ng/mL, Peprotech, CAT#6053) were added to the medium for the first 7 days, followed by the addition of FGF2 (20 ng/mL, Peprotech, CAT#: 130093841) for the remaining 7 days. Following 2 weeks of neural induction (NI), neural progenitors were mechanically harvested and cultured in suspension in neural basal medium (NBM) supplemented with FGF2 (20 ng/mL, Peprotech, CAT#: 130093841) and EGF (20 ng/mL, Peprotech, CAT#: AF-100-15) to promote NSP formation [25]. For terminal differentiation (Diff), NSPs were mechanically dissociated, seeded onto 25 mm coverslips coated with poly-D-lysine (SIGMA Aldrich, St. Louis, MO, USA, CAT#: P1024) and laminin (Life Technologies, CAT#: 23017015), and maintained in NBM medium without supplements for up to 4 weeks. Cells were maintained at 37 °C in 5% CO₂ incubators and the medium was changed every second day.

Cell monolayers (samples from iPSC, NI, and Diff) and NSPs were fixed in 4% PFA for 20 min at 4 °C and then washed briefly in PBS. NSPs were embedded in Tissue-Tek OCT compound (Labtek, Grand Rapids, MI, USA), cut in 15 μm slices on a cryostat, and sections were placed on superfrost slides. After washing in PBS, sections or culture dishes were permeabilized in 0.1% Triton-X-100 in PBS (PBT) for 5 min and then blocked in 10% fetal calf serum in PBT for 60 min at room temperature. Samples were then incubated with primary antibody (diluted in the block buffer) overnight at 4 °C. The following primary antibodies were used: OCT4A (R&D, Minneapolis, MN, USA, CAT# MAB17591), SOX2 (R&D, CAT# MAB2018), NANOG (R&D, CAT#:AF1997), TRA160(R) (R&D, CAT#MAB4770), ACTC1 (Sigma-Aldrich), PAX6 (Santa Cruz, Dallas, TX, USA, CAT#: sc-81649), TBR2 (Abcam, Cambridge, UK, CAT#: ab23345), TBR1 (Abcam, CAT#: ab31940), CTIP2 (Abcam, CAT#: ab18465), SATP2 (Abcam, CAT#: ab51502), TUBB3 (Millipore, Burlington, MA, USA, CAT#: MAB1637), KI67 (Abcam, CAT#: ab15580), NKX2.1 (Abcam, CAT#: ab72876), and IP3R (Abcam, CAT#: ab5804). Following three 5 min washes in PBT, ALEXA-fluor secondary antibodies (Life Technologies/Invitrogen, Carlsbad, CA, USA) (1:1000 diluted in the block buffer) were applied for 1 h at room temperature. All samples were counterstained with 49,6-diamidino-2-phenylindole (Dapi; 1 μg/mL, SIGMA-Aldrich). Samples were then mounted onto glass slides with Mowiol aqueous mountant, followed by viewing and image capturing under Zeiss spinning wheel confocal microscope Axio Observer 7 (Carl Zeiss Microscopy GmbH, Jena, Germany) using ZEN imaging software version 3.4.

RNA was isolated from samples using the Qiagen RNeasy Mini Kit, as per the manufacturer's instructions. In our study, we used two independent iPSC lines reprogrammed from two ASD donors and two independent iPSC lines reprogrammed from two control donors. Each line represented one clone per donor; thus, the biological replicates were donor-derived iPSC lines (n = 2 per group). All four lines underwent the dual SMAD inhibition differentiation protocol on three independent occasions. At each differentiation run, samples were collected at four developmental stages: (i) the iPSC stage (at least three colonies per dish, 3 dishes per line), (ii) the NI stage (day 14 rosettes derived from at least three colonies, 3 dishes per line), (iii) the NSP stage (2-week-old NSPs, 12–18 NSPs per line), and (iv) the Diff stage (neurons differentiated from ~12 NSPs plated into 3 dishes per line). For RNA-seq, RNA was extracted independently from each replicate dish (or NSP batch) per line and time point. Following extraction, RNA samples from replicate dishes within the same line and time point were pooled prior to library preparation. Thus, the unit of biological replication was the iPSC line (one clone per donor), while technical variability between dishes was averaged by pooling post-extraction.

For each of the four iPSC lines (two ASD and two control lines), we collected RNA samples at four time points along the differentiation trajectory: iPSC, 2-week NI, 2-week NSP, and 4-week Diff. This resulted in a total of 16 RNA samples, representing four independent biological replicates (iPSC lines) per time point. The concentration and overall quality of RNA samples were assessed using a Nanodrop ND-100 spectrophotometer and Agilent 2100 Bioanalyzer, with all samples showing an RNA integrity number (RIN) greater than 8 (). Figure S7

RNA sequencing and bioinformatics analysis were outsourced to BGI Genomics (Hong Kong) using the DNBSEQ platform (MGI Tech Co., Ltd., Shenzhen, China), averagely generating about 6.62G Gb bases per sample. The sequencing data were filtered using SOAPnuke. All of the samples had greater than 80% clean reads after filtering out reads of low quality (base quality ≤ 15), reads with adaptor sequences, and reads with high levels of unknown base N. The clean reads were mapped to human genome reference hg19 GRCh37 using HISAT. The average total mapping ratio with the reference genome was 95.92%, and the average total mapping ratio with genes was 74.28%. A total of 18,119 genes were identified.

For gene expression analysis, the clean reads were mapped to reference transcripts using Bowtie2. Analysis was carried out to compare ASD and control samples at each of the 4 stages of differentiation, including iPSC, NI, NSP and Diff. Gene expression level was determined using RSEM (v1.2.28), which provided read count, FPKM, and TPM values. Differential expression gene (DEG) analysis was carried out through DESeq2 with a Q value of ≤0.05. The resulting DEG analysis findings were visualized in a heatmap using pheatmap. To gain insight into changes in the phenotype, Gene Ontology (GO) enrichment analysis and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis were performed using the phyper function in R software (R Foundation for Statistical Computing, Vienna, Austria). The significance levels of both terms and pathways were subjected to correction by Q values, applying a stringent threshold (Q value ≤ 0.05).

For calcium imaging, we first recorded timelapse images from iPSC colonies. For these experiments, three colonies per line were plated on coverslips, and recordings were obtained from at least three coverslips for each line and condition (ATP and KCl stimulation). This procedure was repeated at three independent times. We then recorded timelapse images from differentiated neurons. Two-week-old NSPs (3–6 per coverslip) were plated on three coverslips per line, and recordings were obtained at both 1 week and 4 weeks post-differentiation. For each time point and condition (ATP, KCl, and DHPG stimulation), at least three coverslips were analyzed per line. This workflow was also repeated across three independent differentiation runs. In both cases, the unit of biological replication was the iPSC line, while multiple coverslips and cells within each line provided technical replication.

IPSC colonies and 1-week-old and 4-week-old cultured differentiated neurons were loaded with Ca²⁺ indicator Fluo-4 AM (Thermo Fisher Scientific) in freshly prepared regular Krebs-HEPES buffer with a PH of 7.4 for a duration of 30 min and then were processed for imaging, as previously described [26]. Krebs-HEPES buffer contained 120 mM NaCl, 1.3 mM CaCl₂, 1.2 mM MgSO₄, 4.8 mM KCl, 1.2 mM KH₂PO₄, 25 mM HEPES, and 0.1% BSA. All salts, HEPES, and BSA were prepared from respective powders dissolved in ultrapure (deionized-distilled) water produced by the Milli-Q^® Integral Water Purification System.

Image acquisition and analyses were performed using Zeiss LSM META 510 laser scanning confocal microscope and CLSM software version 3.2. Serial images were acquired at approximately 1 s intervals. A baseline recording was first taken, then the stimulus was delivered at second 20, whereby cells were either stimulated with 25 mM KCl, 100 µM ATP (Tocris, CAT#3245), or 200 µM DHPG (SIGMA Aldrich, CAT#: D368), followed by 4 μM ionomycin at second 200 and EGTA at second 400. Imaging of cells was performed at 37 °C.

For analysis, we performed background subtraction, then we calculated ratio changes by normalizing intensity changes to the averaged intensity of the first three images before delivering the stimulus. Further analysis was performed to compare ASD and control neurons in terms of rate of rise, maximum receptor mediated response, and maximum ionomycin response. For both statistical analysis and visualization, data from the two ASD cell lines were combined into a single ASD group, and data from the two control cell lines were combined into a single control group for each condition: iPSC, 1-week Diff, and 4-week Diff. Statistical analysis was performed to compare between ASD and control datasets using the two-tailed Mann–Whitney U test in GraphPad Prism version 5.04 (GraphPad Software, San Diego, CA, USA).

Two ASD iPSC lines were generated from fibroblasts originally provided by Dr. Philip Schwartz (Children's Hospital of Orange County Research Institute, California) (Table 1). Cells were validated for pluripotency through the expression of key markers, including OCT4, SOX2, TRA1-60, and NANOG (Figure S1). Their differentiation potential was further confirmed via EB formation or directed in vitro differentiation, demonstrating their ability to generate derivatives of all three germ layers (Figure S2). We employed two male control cell lines for a direct comparison with patient iPSCs.

Cortical neurons play a pivotal role in the brain's higher functions, such as sensory processing, communication, and social behavior, the very functions affected in ASD [27]. We therefore sought to specifically differentiate ASD and control iPSC lines to cortical-like neuronal populations. To achieve this, we adopted the dual SMAD inhibition protocol, which is a well-established method and a well-characterized protocol in the Dottori laboratory to obtain cortical-like neurons [24,25,28]. This protocol involves three sequential steps to convert iPSCs into relatively mature cortical neurons, including NI, NSP, and Diff (Figure 1A).

NI was the first step, whereby iPSCs were treated with the small molecule SB431542 and LDN for 7 days, followed by treatment with FGF2 for another 7 days. Colonies at the end of NI showed rosette-like structures representative of the neuroepithelia of the neural tube and immunostaining analyses revealed the expression of neural progenitor marker SOX2 and dorsal telencephalic neuroepithelial marker PAX6 (Figure S3). Transcriptomic data revealed that samples from the NI stage had higher expression of dorsal telencephalic marker PAX6 and forebrain marker FOXG1 compared to the iPSC stage (Figure S4A,B).

NSP formation was the second step, whereby progenitors were harvested and maintained in suspension in NBM medium supplemented with FGF2 and EGF for two weeks to form NSPs (Figure 1A). This stage also coincided with progenitor differentiation toward deep layer cortical subpopulations. Immunostaining revealed strong expression of PAX6 and the deep layer cortical marker TBR1, suggesting a progression toward post-mitotic cortical neuron fate, with limited the expansion of intermediate progenitors, as indicated by low TBR2 expression. Cells also expressed TUBB3, confirming neuronal identity, and a subset was positive for Ki67, reflecting ongoing proliferation within the culture (Figure 1B). While, as expected, minimal to no expression was observed for ventral telencephalic marker NKX2.1 and intermediate and upper layer markers CTIP2 and BRN1. This suggested that 2-week-old NSPs were mainly enriched with neuronal populations of dorsal telencephalic identity, specifically the deep layer cortical neuronal populations (Figure 1B).

Diff was the third step, whereby 2-week-old NSPs were mechanically dissociated and plated onto poly-D-lysine and laminin coated surface to terminally differentiate neural progenitors to post-mitotic neurons (Figure 1A). Immunostaining analyses of neuronal cultures within the fourth week post-differentiation showed the expression of neuronal markers TUBB3 and TBR1, suggesting enrichment of deep cortical neuronal populations in Diff cultures. Additionally, the presence of Ki67-positive cells suggested that a subset of the culture remained proliferative at this stage (Figure 1C). Complementing these results, our transcriptomic analyses also showed higher levels of TUBB3 in NSP and Diff samples compared to iPSC samples (Figure S4C).

Building on these observations, transcriptomic analyses further revealed the altered expression of genes associated with neuronal differentiation in ASD-derived Diff samples relative to controls, including upregulation of the pyramidal neuron marker EMX1 and the layer 6 cortical neuron marker TBR1 (Figure S5). These molecular changes may reflect underlying alterations in corticogenesis associated with ASD. While these transcriptomic data provide valuable insights into stage-specific gene expression patterns, the limited sample size constrained our ability to perform formal statistical analyses. These findings suggest potential deviations in the neuronal differentiation trajectory of ASD-derived neurons.

Ca²⁺ signaling has been identified by previous transcriptomic studies using iPSC models as a major hub in ASD pathophysiology [15]. Accordingly, we sought to conduct a transcriptomic profile of samples collected from different time points along the course of differentiation to identify the developmental stages affected by dysregulation of Ca²⁺ signaling. To achieve this, we performed bulk RNA sequencing analysis of each cell line (2 ASD and 2 control lines) at four sequential stages of the differentiation—iPSC, 2-week NI, 2-week NSPs, and 4-week Diff. The analysis was carried out to compare the gene expression profiles of ASD and control samples at each of the mentioned 4 stages of differentiation. Using a Q value of <0.05, our data revealed that the number of DEGs was highest in the Diff stage, with a total number of DEGs corresponding to 495, followed by the iPSC stage, corresponding to 387, while the numbers of DEGs were lower in the NI and NSP stages, corresponding to 58 and 56, respectively, with a general trend of more upregulated than downregulated genes (Figure 2A). Tables S1 and S2 illustrate the list of significant DEGs at the iPSC and Diff stages.

DEGs in the iPSC and Diff stages were then subjected to functional classification and GO term enrichment analysis. Results suggested that binding was the most enriched GO term under Molecular Function, corresponding to 360 DEGs in the iPSC stage and 441 DEGs in the Diff stage. Further analyses revealed that, in the iPSC stage, 39 DEGs were annotated as Ca²⁺ ion binding (Q value = 0.00011) and, in the Diff stage, 39 DEGs were also annotated as Ca²⁺ ion binding (Q value = 0.002) (Figure 2B). We also subjected DEGs in the iPSC and Diff stages to KEGG pathway annotation classification. Our results revealed that 25 DEGs in the Diff stage were classified under the term Calcium Signaling Pathway, with a Q value of 0.00000014. On the other hand, 9 DEGs in the iPSC stage were identified under the term Calcium Signaling Pathway but did not reach statistical significance (Q value = 0.79). The KEGG network map and expression heatmap for these DEGs identified under the term Calcium Signaling Pathway in the iPSC and Diff stages are shown in Figure 2C,D. Among the DEGs in the Diff stage, we found upregulated genes that are well recognized for their integral role in maintaining resting Ca²⁺ and facilitating different Ca²⁺ signals, some of which were already reported in previous studies. These include ryanodine receptors (RYR1/3), calcium voltage-gated channel subunit alpha1 S (CACNAS), calcium/calmodulin dependent protein kinase IV (CAMK4), glutamate metabotropic receptor 5 (GRM5), ATPase plasma membrane Ca2+ transporting 3 (ATP2B3), glutamate ionotropic receptor NMDA type subunit 1 (GRIN1), and calcium/calmodulin dependent protein kinase II beta (CAMK2B). Given the crucial role of Ca²⁺ signaling in regulating distinctive processes during different stages of neurodevelopment, we were prompted to validate these findings and investigate the biological implication of such disruptions.

We performed Ca²⁺ imaging studies of ASD and healthy control samples at multiple stages of differentiation, starting with the pluripotency stage. To assess the impact of dysregulation in the Ca²⁺ signaling-related transcriptome on ASD iPSCs, ATP was used in these studies to stimulate purinergic receptors and induce Ca²⁺ release from internal stores. A similar approach and settings to previously published studies from our group were adopted [29]. Briefly, Fluo-4-loaded iPSC colonies from the ASD and control groups were processed for time lapse recordings for a duration of 500 s. A baseline recording was first taken, then the stimulus, 100 µM ATP, was delivered at second 20, followed by 4 µM ionomycin at second 200 and EGTA at second 400 (Figure 3A). Both ASD and control iPSC lines showed homogeneous responses to ATP. Interestingly, ASD iPSCs had higher levels of intracellular Ca²⁺ in response to ATP, reaching an average maximum change of 4.23 ± 0.085 fold (normalized fold increase relative to baseline) compared to control iPSCs, which reached an average maximum change of 3.71 ± 0.085 fold (p value ≤ 0.0001, n = 3) (Figure 3B). On the other hand, ASD iPSCs had slightly lower levels in response to ionomycin, reaching a maximum change of 3.60 ± 0.081 fold compared to control iPSCs, which reached a maximum change of 3.89 ± 0.10 fold (not statistically significant, n = 3) (Figure 3D). The rise in intracellular Ca²⁺ to reach the maximum levels in ASD iPSCs was significantly slower, taking an average time of around 82.59 ± 2.45 s compared to controls, which corresponded to 49.87 ± 1.91 s (p value ≤ 0.0001, n = 3) (Figure 3E). These results confirmed the findings of our transcriptomic analyses, that there are differences in Ca²⁺ homeostasis and dynamics between ASD and control samples at the iPSC stage.

Our transcriptomic data indicated that the highest number of Ca²⁺ signaling-relevant DEGs was in the Diff neuronal stage. Therefore, we sought to investigate functional Ca²⁺ differences at the Diff neuronal stage, and we selected to perform these studies at two time points post-differentiation, early at 1-week Diff and late at 4-week Diff. For studies conducted at the 1-week Diff time point, we characterized the dynamics of intracellular Ca²⁺ upon stimulation with either KCl, to probe the release of Ca²⁺ through activation of voltage-gated Ca²⁺ channels, or ATP, to probe the release of Ca²⁺ through activation of the purinergic receptor and downstream IP₃ pathway signaling.

Accordingly, we firstly sought to determine if there were differences between the ASD and control groups in Ca²⁺ dynamics during the early stage of differentiation upon stimulation with KCl. Our results indicated that the percentage of cells that had doubled their response to KCl was slightly higher for the ASD group, corresponding to 57.37 ± 6.33%, compared to 42.30 ± 6.85% for the control group. Data also showed that the average maximum response to KCl in ASD neurons was significantly higher than in controls, corresponding to an average intensity of 2.55 ± 0.14, compared to 2.20 ± 0.13 in controls (p value ≤ 0.05, n = 3) (Figure 4A–C). The rate of rise to reach the maximum was significantly faster in ASD neurons, corresponding to 15.57 ± 6.88 s, compared to 36.88 ± 3.21 s in the control group (p value ≤ 0.05, n = 3), overall suggesting that ASD neurons may have higher excitability than controls early (week 1) in the differentiation stage (Figure 4A–C). No significant differences were identified in terms of subsequent global Ca²⁺ rise in response to ionomycin in these experiments.

We then assessed whether the kinetics of Ca²⁺ release would be different between ASD 1-week neurons and controls upon activation of purinergic receptors by ATP. Our data showed that the percentage of ASD neurons that doubled their Ca²⁺ response was 13.63 ± 5.233%, which was lower compared to 85 ± 4.35% in controls. In addition, the average maximum values during ATP-mediated response in ASD neurons was lower, corresponding to 1.58 ± 0.12 (p value ≤ 0.0001, n = 3), and slower to reach the maximum, with an average duration of 57 ± 7.78 s compared to controls, which reached an average maximum of about 2.74 ± 0.10 s, with an average duration of 27.62 ± 6.30 s (p value ≤ 0.01, n = 3) (Figure 4D–F). The average maximum Ca²⁺ response to ionomycin was comparable between the two groups, although slightly less in the ASD group (ASD = 2.73 ± 0.16 fold, control = 2.97 ± 0.13 fold, p value ≤ 0.05, n = 3), with a significant longer duration to reach the maximum (ASD = 81.69 ± 7.52 s, control = 37.85 ± 6.10 s, p value ≤ 0.0001, n = 3) compared to the control group (Figure 4D–F). The response demonstrated by ASD and control neurons at the NI stage was different than at the iPSC stage, which was expected given that different machinery is operating to regulate and maintain Ca²⁺ according to different cellular context across different cell types. ATP and downstream purinergic signaling play an important role during neurogenesis and maintenance of neural stem cells. This was consistent with our observation that a homogeneous response to ATP was observed for both ASD and control samples at the iPSC stage. By contrast, in 1-week-old neurons, only a small percentage of ASD neurons responded to ATP compared to controls.

We next sought to assess whether phenotypic differences in Ca²⁺ dynamics between ASD and control samples that were observed during the early phase of differentiation were also maintained at a later and more mature stage of differentiation, specifically week 4. In addition to KCl and ATP, we also used DHPG to assess the dynamics of Ca²⁺ upon activation of metabotropic glutamate receptors (mGLURs) given that neurons at this stage express mGLURs and are mature enough to respond to DHPG, unlike 1-week-old neurons, which showed minimal to no response.

Our data showed that the differential profile of KCl-induced Ca²⁺ dynamics in 4-week-old neurons between ASD and control samples was comparable to the 1-week-old neurons, whereby ASD neurons showed a higher average maximum response (ASD = 3.30 ± 0.13, control = 2.75 ± 0.10, p value ≤ 0.001, n = 3) and significantly shorter duration (ASD = 14.38 s, control = 20.97 s, p value ≤ 0.0005, n = 3) to reach the maximum compared to controls (Figure 5A–C). The percentage of cells doubling their response relative to the resting level was slightly higher in ASD neurons (85.71% ± 3.66) compared to controls (75.22% ± 4.061). Compared to 1-week-old neurons, the response of 4-week-old neurons in both groups (ASD and control) was higher, suggesting that neurons at week 4 had increased excitability compared to 1-week Diff neurons. Overall, ASD neurons had increased responses to KCl at both 1-week and 4-week Diff time points compared to controls, which suggested that ASD neurons are more excitable (Figure 5A–C). Furthermore, differences in global Ca²⁺ dynamics upon the addition of ionomycin was significantly higher in ASD neurons, whereby ASD 4-week-old neurons showed a higher average maximum response (ASD = 3.28 ± 0.12 and control = 2.89 ± 0.10, p value ≤ 0.005, n = 3) and faster duration to reach the maximum (ASD = 36.16 ± 5.01 and control = 51.64 ± 4.46, p value ≤ 0.001, n = 3) compared to controls (Figure 5A–C).

We next assessed the dynamics of Ca²⁺ in ASD neurons upon stimulation with ATP and downstream purinergic signaling. Similar profiles were observed at 1 and 4 weeks, whereby ASD 4-week-old neurons showed a lower average maximum Ca²⁺ response (ASD = 2.35 ± 0.14 fold, Control = 2.88 ± 0.08 fold, p value ≤ 0.0001, n = 3) with a longer duration to reach the maximum (ASD = 38 ± 4.70 s, control = 27.53 ± 2.75 s, p value ≤ 0.05, n = 3) compared to controls (Figure 5D–F). The percentage of cells doubling their response to ATP was lower (58.19 ± 3.71%) in ASD 4-week-old neurons compared to controls (90.78 ± 3.32%). The profile of Ca²⁺ rise during the response to ionomycin in these experiments was also significant in ASD neurons compared to controls, where ASD neurons showed a lower average maximum (ASD = 2.97 ± 0.14, control = 3.60 ± 0.08, p value ≤ 0.005, n = 3) and longer duration (ASD = 45.35 ± 4.52, control = 36.27 ± 3.24, not statistically significant, n = 3) to reach the maximum compared to controls (Figure 5D–F).

We next assessed whether there would be differences between ASD and control neurons in the kinetics of Ca²⁺ release upon stimulation of mGLURs with DHPG. Given that our transcriptomic analyses suggested higher expression of mGLURs in ASD neurons compared to controls, we expected that the rise in Ca²⁺ in response to activation of mGLURs would be higher in ASD neurons compared to controls. Indeed, ASD 4-week-old neurons showed a higher average maximum Ca²⁺ response (ASD = 2.68 ± 0.14, control = 2.14 ± 0.08, p value ≤ 0.05, n = 3) with a shorter duration to reach the maximum (ASD = 14.07 ± 1.60 s, control = 25.86 ± 0.12 s, p value ≤ 0.0001, n = 3) compared to controls (Figure 5G–I). The percentage of cells doubling their response to DHPG was higher (59.57 ± 5.06%) in ASD 4-week-old neurons compared to controls (51.13 ± 5.32%). The response to ionomycin was also similarly different in ASD neurons compared to control neurons, whereby ASD neurons showed a higher average maximum (ASD = 3.22 ± 0.15, control = 2.81 ± 0.12, not statistically significant, n = 3) and a shorter duration to reach the maximum (ASD = 33.31 ± 3.40 s, control = 53.15 ± 4.94 s, p value ≤ 0.0005, n = 3), overall suggesting altered global intracellular Ca²⁺ dynamics in ASD (Figure 5G–I).