Early maturation and hyperexcitability is a shared phenotype of cortical neurons derived from different ASD-associated mutations

Autism Spectrum Disorder (ASD) is characterized mainly by social and sensory-motor abnormal and repetitive behavior patterns. Over hundreds of genes and thousands of genetic variants were reported to be highly penetrant and causative of ASD. Many of these mutations cause comorbidities such as epilepsy and intellectual disabilities (ID). In this study, we measured cortical neurons derived from induced pluripotent stem cells (iPSCs) of patients with four mutations in the genes GRIN2B, SHANK3, UBTF, as well as chromosomal duplication in the 7q11.23 region and compared them to neurons derived from a first-degree relative without the mutation. Using a whole-cell patch-clamp, we observed that the mutant cortical neurons demonstrated hyperexcitability and early maturation compared to control lines. These changes were characterized by increased sodium currents, increased amplitude and rate of excitatory postsynaptic currents (EPSCs), and more evoked action potentials in response to current stimulation in early-stage cell development (3–5 weeks post differentiation). These changes that appeared in all the different mutant lines, together with previously reported data, indicate that an early maturation and hyperexcitability may be a convergent phenotype of ASD cortical neurons.


INTRODUCTION
Autism Spectrum Disorder (ASD) was first defined by Leo Kanner in 1943 and named "early infantile autism" as an independent disorder from the psychotic disorder of schizophrenia, describing 11 children with social, biological, and emotional abnormalities such as the inability to relate to others or objects in a traditional way, an anxious and obsessive desire for consistency, early eating difficulties and hearing problems 1 . Recently, more symptoms entered the category of autistic-like behaviors such as attention deficit hyperactivity disorder (ADHD), poorly integrated verbal and non-verbal communication, abnormalities in eye contact, hyper or hypo-reactivity to sensory input, repetitive body movements, and more 2,3 . On the genomic aspect, there has been more attention to many variants in which chromosomal subregions are deleted or duplicated in an inherited and de novo manner as well 2 . Various genes and mutations have been reported to be associated with ASD 4 , and there seem to be a most definitive leaders: ADNP, CHD8, and SHANK3 5 . On the neurobiological aspect, no specific brain area nor system has been confirmed to be entirely associated with the disorder, but an overall brain impairment has been shown starting from childhood 3 . The areas in the brain that are thought to be affected include cortical as well as non-cortical regions: the prefrontal cortex (PFC), Brodmann's areas 9, 21, and 22, orbitofrontal cortex (OFC) fusiform gyrus, fronto-insular cortex, cingulate cortex, hippocampus, amygdala, cerebellum, and brainstem 6 . Our study focuses on the following four genetic mutations:

Analysis of electrophysiological recordings
Analysis was performed based on previously described 57 analysis using custom-written MATLAB scripts modified as follows: Synaptic currents analysis. The mean and standard error (SE) of the excitatory postsynaptic currents (EPSCs) amplitudes for each active cell were calculated. The cumulative distribution of EPSCs amplitude was calculated for each group. For each cell, the rate of the events was calculated by dividing the number of events by the time period of the recording (non-active cells were included and had an event rate=0). The mean of all cells' rates and the standard error of the frequencies were computed for the control and mutant groups. Non-parametric statistical tests (Wilcoxon signed rank test) were performed for comparisons. Sodium, fast, and slow potassium currents. Neurons were held in voltage clamp mode at -60 mV, and voltage steps of 400 ms were performed in the -100 to 90 mV range. Currents were typically normalized by the cells' capacitance; The sodium current was computed by subtracting the sodium current after stabilization from the lowest value of the inward sodium current. The fast potassium currents were measured by the maximum outward currents that appeared within a few milliseconds after a depolarization step. The slow potassium currents were measured after the 400 ms depolarization phase. A one-way ANOVA test was performed for the statistical analysis. Evoked action potentials (APs). Neurons were held in current clamp mode at -60 mV with a constant holding current. Following this, current injections were given in 3-pA steps throughout 400 ms, starting 12 pA below the steady-hold current. A total of 38 depolarization steps were given. The total evoked action potential was the total number of action potentials that were counted in the 38 depolarization steps. Non-parametric statistical tests (Wilcoxon signed rank test) were performed for comparisons. Action potential shape analysis. The first evoked action potential generated with the lowest amount of injected current was used for spike shape analysis. The spike threshold was defined as the membrane potential at which the slope of the depolarizing membrane potential increased dramatically, resulting in an AP (the second derivative of the voltage versus time as the initial maximum). The spike height was calculated as the difference between the highest membrane potential during a spike and the threshold. The spike rise time is the time it takes the spike to reach the maximum. The spike width was calculated as the time it took the membrane potential to reach half of the spike amplitude in the rising part.

Code Availability
The custom-written scripts used for the analysis can be shared upon reasonable request.

Dup7 cortical neurons display increased sodium and potassium currents, increased synaptic activity and hyperexcitability early in the differentiation.
We performed whole-cell patch clamp experiments five weeks (day 34) after the start of the differentiation of 16 Dup7-mutant neurons and 13 control neurons derived from a first-degree relative of the same gender. In a voltage clamp mode, EPSC recordings were performed by holding the cell at -60mV. We observed an increase in the rate of EPSCs of the mutant neurons compared to the controls (0.13 + 0.08 Hz in Dup7-mutant neurons and 0.07 + 0.07 Hz in the control neurons, p=0.03), as shown in Fig. 1a-c. Fig. 1a-b presents representative traces, while Fig. 1c represents the average over all the recordings. Additionally, a significant increase in the mean amplitude of the EPSCs was observed. The Dup7-mutant neurons had a larger amplitude compared to the control neurons (12.72 + 2.86 pA for the mutant neurons and 7.12 + 5.09 pA for the control neurons (p=0.002, (Fig. 1d.)). The cumulative distribution of the EPSC amplitudes for Dup7-mutant neurons is slightly right-shifted compared to control neurons indicating larger amplitudes of EPSCs (Fig. 1e). Next, we recorded in voltage clamp mode the sodium and potassium currents. We observed a significantly larger normalized sodium current in the Dup7-mutant neurons compared to the control neurons (F (1,38) = 9.43, p=0.004). Representative traces of the recordings are shown in Fig. 1f (control) and 1g (mutant). The average sodium currents are presented in Fig. 1h. Additionally, we observed increased slow and fast potassium currents (normalized by the capacitance) in the Dup7-mutant neurons compared to controls (F (1,14) = 5.61; p=0.03 for the slow potassium currents and F (1,14) = 8.36; p=0.01 for the fast potassium currents) over the 10-80 mV range ( Fig. 1h-j). We next measured the number of evoked action potentials in a current clamp mode as a measure of the neuronal excitability. We observed a hyperexcitability pattern for the Dup7-mutant neurons compared to control neurons. The total number of evoked potentials (see Methods) for the Dup7-mutant neurons was 47.63 + 49.61; for the control neurons, it was 28.5 + 69.48 (p=0.006). A representative example is presented in Fig. 1k (control) and 1l (Dup7), and the average over all recordings is shown in Fig. 1m. Spike shape analysis (see Methods) is presented in Table S2 .Examples of ICC images for control and mutant lines are shown in Fig. 1n-o. (typical NPC markers PAX6 and NESTIN) and 1p-q. (neuronal markers MAP2 and the cortical marker TBR1).

GRIN2B cortical neurons display increased sodium and potassium currents and hyperexcitability early in the differentiation.
We We next measured the number of evoked action potentials in current clamp mode as a measure of the neuronal excitability. We observed a hyperexcitability pattern for the GRIN2B-mutant neurons compared to control neurons. The total number of evoked potentials (see Methods) in the GRIN2B-mutant neurons was 38.86 + 10.17, and in the control neurons, it was 22.37 + 19.5 (p=0.003). A representative example is presented in Fig. 2k (control) and 2l (GRIN2B), and the average over all recordings is presented in Fig. 2m. Furthermore, we observed a significant increase in GRIN2B-mutant neurons' spike amplitude compared to control neurons (41.2 + 12.5 mV in GRIN2Bmutant neurons and 20.5 + 14.01 mV in control neurons, p =0.008); further spike shape analysis is presented in Table S2. Examples of ICC images for control and mutant lines are shown in Fig. 2n-o. (typical NPC markers PAX6 and NESTIN) and 2p-q. (neuronal markers MAP2, the cortical marker TBR1).

SHANK3 cortical neurons display increased sodium and slow potassium currents, a drastic increase in synaptic activity and hyperexcitability early in the differentiation.
We performed whole-cell patch clamp experiments five weeks (days 29-32) after the start of the differentiation of 21 SHANK3-mutant and 17 control neurons. In a voltage clamp mode, EPSC recordings were performed by holding the cell at -60mV. We observed a significant increase in the rate of EPSCs of the mutant neurons compared to the controls (0.28 + 0.36 Hz in SHANK3-mutant neurons and 0.08 + 0.06 Hz in the control neurons, p0.003) as shown in Fig. 3a-c. Fig. 3a-b presents representative traces, while Fig. 3c represents the average over all the recordings. Additionally, a drastic increase in the mean amplitude of the EPSCs was observed. The SHANK3-mutant neurons had a larger amplitude compared to the control neurons (10.145 + 3.26 pA for the mutant neurons and 5.29 + 1.65 pA for the control neurons, p=1.15e -5 , (Fig. 3d)). The cumulative distribution of the EPSC amplitudes for SHANK3mutant neurons is right shifted compared to control neurons indicating larger amplitudes of EPSCs (Fig. 3e). Next, we recorded in voltage clamp mode the sodium and potassium currents. We observed a significantly larger normalized sodium current in the SHANK3-mutant neurons compared to the control neurons F (1,36) = 4.51, p=0.04. Representative traces of the recordings are shown in Fig. 3f (control) and 3g (mutant). The average sodium currents are presented in Fig. 3h. Additionally, we observed increased slow, but not fast, potassium currents (normalized by the capacitance) in the SHANK3-mutant neurons compared to controls, (F (1,10) = 5.68; p=0.03) over the 40-90 mV range ( Fig. 3h-j). We next measured the number of evoked action potentials in current clamp mode as a measure of the neuronal excitability. We observed a hyperexcitability pattern for the SHANK3-mutant neurons compared to control neurons. The total number of evoked potentials (see Method) for the SHANK3-mutant neurons was 31.32 + 39.8; for the control neurons, it was 18.4 + 22.96 (p=0.03). A representative example is presented in Fig. 3k. (control) and 3l. (SHANK3) and the average over all recordings are presented in Fig. 3m. Furthermore, we observed a more depolarized threshold (higher) in the SHANK3-mutant neurons compared to control neurons (30.4 + 5.6 mV in the SHANK3-mutant neurons and 25.5 + 4.2 mV in control neurons, p =0.02); further spike shape analysis is presented in Table S2. Examples of ICC images of control and mutant lines are shown in Fig. 3n-o. (typical NPC markers PAX6 and NESTIN) and 3p-q. (neuronal markers MAP2, the cortical marker TBR1).

UBTF cortical neurons display increased sodium currents, an increase in synaptic amplitude, an increase in spontaneous activity and hyperexcitability early in the differentiation.
We performed whole-cell patch clamp experiments five to six weeks (days 32-37) after the start of the differentiation of 22 UBTF-mutant neurons and five to six weeks (days 32-37) after the beginning of the differentiation of 18 control neurons. In a voltage clamp mode, EPSC recordings were performed by holding the cell at -60mV. Fig. 4a-b presents representative traces, while Fig. 4c. represents the average EPSC amplitude over all the recordings. The EPSC rate of the UBTF-mutant neurons is slightly similar to the control neurons (Fig. 4c). We observed an increase in the mean amplitude of the EPSCs. The UBTF-mutant neurons had a larger amplitude compared to control neurons it was7.81+ 1.16 pA for the mutant neurons and 6.55 + 2.39 pA for control neurons, p=0.02 (Fig. 4d.). The cumulative distribution of the EPSC amplitudes for UBTF-mutant neurons is slightly rightshifted compared to control neurons indicating larger amplitudes of EPSCs (Fig. 4e.). Next, we recorded in voltage clamp mode the sodium and potassium currents. We observed a significantly larger normalized sodium current in the UBTF-mutant neurons compared to the control neurons F (1,28) = 5.58, p=0.03 over the -50-90-mV range. Representative traces of the recordings are shown in Fig. 4f. (control) and 4g. (mutant). The average sodium currents is presented in Fig. 4h. An ANOVA test indicated no significant differences in the slow and fast potassium currents (Fig. 4h-j.). We next measured the number of evoked action potentials (see Methods). For the UBTF-mutant neurons, it was 62.95 + 31.56 and for the control neurons, it was 31.46+ 22.3 (p=5.56e -4 ). A representative example is presented in Fig. 4k. (control) and 4l. (UBTF) and the average over all recordings are presented in Fig. 4m. Furthermore, we observed a significant increase in UBTF-mutant neurons' spike amplitude compared to control neurons (50+ 17.05 mV in the UBTF-mutant neurons and 36.04 + 18.01 mV in control neurons, p =0.03). Besides, we observed a narrower spike in the UBTF-mutant neurons compared to the control neurons (3.3 + 1.5 ms in the UBTF-mutant neurons and 12.5 +17.8 ms in the control neurons, p = 0.003); further spike shape analysis is presented in Table S2. The spontaneous neuronal activity (spontaneous action potentials) was measured in a holding potential of -45mV. We observed a significant increase in the spontaneous activity rate (p= 0.013) and amplitudes (p= 0.012) in the UBTF-mutant neurons compared to control neurons (Fig. 4n-o.). Examples of ICC images of control and mutant lines are shown in Fig. 4p-q. (typical NPC markers PAX6 and NESTIN) and 4r-s. (neuronal markers MAP2, the cortical marker TBR1).

Mutant Neural Progenitor Cells Exhibit Longer Neurite Lengths Compared to Controls, alongside with decreased GABA-Positive Neurons in ASD-Related Mutant neuronal cultures
We observed that the lengths of neurites of neural progenitor cells (NPCs) in mutant groups were significantly increased compared to control groups (Fig. 5a). This difference was determined by analyzing brightfield images captured at 20X magnification and measuring the length of neurites using a standardized methodology (see Methods). Specifically, the average length of NPC neurites in the Dup7-mutant NPCs (left) was 0.12 ± 0.02 cm compared to control NPCs was 0.07 ± 0.02 cm (p=4.36e-04), in the SHANK3-mutant NPCs (middle) the average length was 0.09 ± 0.03 cm compared to control NPCs 0.07 ± 0.02 cm (p = 6.75e-08) and in the UBTF-mutant NPCs it was 0.09 ± 0.03 cm compared to control NPCs 0.05 ± 0.01 cm (p=0.01), as determined by Wilcoxon signed-rank test. Example images of control (b) and patient (c) are shown in Fig 5b-c. Next, we used GABA and VGluT1 antibodies to confirm the existence of GABA and VGluT1-positive cells in our cultures. Indeed, the control cultures contained 15.08 ±1.13 % GABA-expressing and 77.87± 2% VGluT1-expressing neurons. However, the ASD-related mutant cultures contained only 11.73±1.14 % GABA-expressing and 79.99 ± 2.93% VGluT1-expressing neurons (P GABA = 0.04, P VGluT1 = 0.56); Additionally, we observed a higher ratio of Excitation-Inhibition (E-I ratio) in the Patient group compared to Control (8.183 ±1.05 and 5.51 ± 0.38 respectively;p=0.03) as shown in Fig 5d. Fig 5e-f shows an example image of control (e.) and mutant (f) neuronal cultures that were immunostained for DAPI (blue), VGluT1 (red), and GABA (green).

DISCUSSION
In this study, iPSC technology was used to investigate the physiological features of cortical neurons derived from human patients with different ASD-related mutations: Dup7, GRIN2B, SHANK3, and UBTF. For that purpose, we differentiated patient-derived iPSCs into cortical neurons 55 since cortical alterations and malformations within the brain were identified in these mentioned mutations 14,29,43,50 . A broad range of mutations has been associated with ASD, sharing autistic-like behaviors such as difficulties in social communication and interaction and restricted or repetitive behaviors or interests. The broad spectrum of mutations is characterized by affecting different genes and pathways. Various genes affecting the cytoskeletal [58][59][60][61][62] and microtubule 63 dynamics are altered in ASD-associated genes, disturbing a critical stage in axons and dendrites development 64 ; For Dup7 mutation, for example, a rare genetic syndrome caused by a micro-duplication in section q11.23 of chromosome 7, alterations in the ELN gene such as inherited and de novo deletions were reported, coding for the extracellular matrix protein, elastin, associated with connective-tissue malformations as reported in human patients 12 . SHANK3 is a gene located at the terminal long arm of chromosome 22, coding for a master scaffolding protein found in the body's tissues, and importantly in the brain, and has a critical role in the postsynaptic density of glutamatergic synapses and synaptic functions in rat and human brains 15,19 . The GRIN2B mutation results in a production of a nonfunctional GluN2B protein. A shortage or dysfunction of this protein may cause an extreme reduction in the number of the functional NMDA receptors 35 causing neuronal impairments in both mice and human models 65,66 . UBTF is a gene coding for UBF, a transcription factor in RNA Pol I, critical for rRNA transcripts synthesis from rDNA in the nucleolus 45 . Loss of UBF induces nuclear disruptions, including inhibition of cell proliferation, rapid and synchronous apoptosis, and cell death in mice models 46 . Although these mutations are functionally very different from one another, they cause similar symptoms in the patients. Interestingly, we observed a hyperexcitability pattern in all these mentioned ASD-related mutations in an early-stage cell development (3-5 weeks post differentiation) that we measured by electrophysiological recordings. This hyperexcitability involved different aspects; In terms of sodium-potassium currents and activity, a consistent increase in sodium currents was observed within the four patient-derived neurons, which could, in turn, increase the excitability of the neurons by decreasing the action potential threshold 67 . These alterations of sodium currents can lead to abnormal neuronal activity, a phenomenon that also occurs in epilepsy 68 . It is interesting to note that all these four mutations have a strong association with epilepsy, and many of the patients also suffer from epilepsy 14,27,44,53 . We also observed an increase in the EPSC rate and amplitude, indicating pre and postsynaptic changes that occur in the mutant neurons compared to the controls neurons. Similar findings were reported in mice models of autism [69][70][71][72][73] . Furthermore, we observed more evoked action potentials in response to current stimulation in the mutant-derived neurons, alongside with a decrease in GABA-positive cells and an increase in the E-I ratio, which may contribute to the occurrence of infantile epileptic seizures in affected human subjects. We previously reported such changes also in another ASD model of an IQSEC2 mutation 74 . Furthermore, the NPCs derived from individuals with ASD-related mutations display longer neurite-like branches compared to control groups supporting our conclusions of their early maturation. All these changes can indicate that the ASD mutant neurons develop faster, and at this early stage, when the control neurons are still very immature, they are already spiking and connecting with other neurons. Several previous genetic studies showed a rise in cortical activity by documenting excitation-to-inhibition ratio system alterations [74][75][76] , suggesting that periodic seizures and sensory hyperreactivity in ASD are caused by cortical hyperexcitability 77,78 , which was also previously observed in fragile-X syndrome 79 . Previously, we reported a similar early time point hyperexcitability pattern in another ASD-related mutation -the A350V IQSEC2mutation 74 . In that study, we followed the IQSEC2-mutant neurons that started more active and more connected (5 weeks post differentiation) as they became hypoexcitable with reduced synaptic connections later on in the differentiation process. An additional hyperexcitability pattern in 15q11-q13 Duplication Syndrome, a different ASD-associated mutation, was reported 80 . A reduction in synaptic connections was reported in a long line of ASD-related and IDrelated studies using mice and human models [81][82][83][84] , suggesting a potential shared accelerated aging mechanisms in ASD-associated mutations 62,63,[85][86][87] . We speculate that there may be a connection between this early hyperexcitability and the later synaptic deficits, as perhaps this early hyperexcitability is neurotoxic to the cell at such an early stage of development. Since, especially with mice studies, it is much harder to measure this early developmental stage, perhaps this is a stage that precedes the synaptic degradation in many other ASD mutations. More evidence of this early maturation was presented in a study with ASD patients with macrocephaly where the neurons derived from the patients were more arborized early in the development, and gene expression profiles also suggested an earlier maturation 88 . More evidence for functional hyperactivity in epilepsy and ASD-related mutations in human models were reported; briefly, an engineered iPSC-derived neuron with the homozygous P924L mutation (one of many epilepsy-associated Slack mutations) displayed increased K Na currents and more evoked action potentials in both single neurons and a connected neuronal network 89 . Our Findings present a shared phenotype of early maturation and hyperexcitability in four ASD-related mutations using patient-derived cortical neurons, indicating that there may be a common neurophysiological phenotype in ASD-related variants, sharing similar behavioral phenotypes but a different genotype. iPSC-derived neurons were previously used as a research tool for investigating physiological and cellular alterations characterizing various disorders including autism 74,88,90,91 ,epilepsy 89,92,93 and other neurodegenerative diseases such as Parkinson's [94][95][96][97] and schizophrenia [98][99][100] . Here we concentrated on the early developmental physiological alterations in 4 different ASD and epilepsy-related genes. The enhanced maturation and excitability in such young neurons may be deleterious to the cells and may later result in synaptic degeneration as was previously described in neurons derived from ASD and epilepsy patients 74,84,[88][89][90]92,93 .

Fig. 1. Young (5 weeks post differentiation) Dup7-mutant neurons are hyperexcitable compared to control neurons.
a A representative trace of (EPSCsmeasured in control cortical neurons at five weeks post-differentiation. b A representative trace of EPSCs measured in a dup7-mutant neuron at five weeks post-differentiation. c The mean rate of synaptic events was higher in the dup7-mutant neurons compared to control neurons (p=0.029). d The average amplitude of EPSCs was increased in the dup7-mutant neurons (p =0.002). e The cumulative distribution of the amplitude of EPSCs is slightly right-shifted in the dup7-mutant neurons, indicating an increase in the amplitudes. f A Representative trace of sodium and potassium currents recorded in a voltage-clamp mode in control neurons. g A Representative trace of sodium and potassium currents recorded in a voltage-clamp mode in dup7-mutant neurons. h The average sodium currents in dup7-mutant neurons is increased compared to control neurons (p=0.004). i The average slow potassium currents in dup7-mutant neurons is increased compared to control neurons (p=0.03). j The average fast potassium currents is increased in dup7-mutant compared to control neurons (p=0.01). k A representative recording of evoked action potentials in a current-clamp mode of a control neuron. l A representative recording of evoked action potentials in a current-clamp mode of a dup7-mutant neuron. m The total number of evoked action potentials is larger in dup7-mutant neurons compared to control neurons (p=0.006). n,o A representative image of control n and mutant o NPCs that were immunostained for DAPI, PAX6, and Nestin. p-q A representative image of control p and mutant q neurons that were immunostained for DAPI, MAP2, and TBR1.