Viral particles containing an open reading frame of Neurogenin-2 (NGN2) were produced to differentiate iPSCs into mature neurons as described in Hulme et al. (2020) and Maksour et al. (2024). Briefly, HEK293T cells were transfected with the DNA of lentiviral packaging plasmids vSVG (Addgene, USA, #8454), RSV (Addgene, #12253), pMDL (Addgene, #12251), and either the tetracycline transactivator (TTA) vector, M2rtTA (Addgene, #20342), or the NGN2 overexpression vector, TetO-NGN2-eGFP-Puro plasmid (Addgene, #79823) using Polyethyleneimine (Sigma-Aldrich, USA, #408727). DNA was added in a ratio of 4:2:1:1, transfer vector:pMDL:RSV:vSVG. The cell culture media containing viral particles was collected every 24 h over 3 days. The viral supernatant was concentrated 200× by ultracentrifugation at 66,000 × g for 2 h at 4°C. The viral pellet was resuspended in PBS and stored at −80°C until needed.

This study used a protocol published in Maksour et al. (2024) to generate mature neurons via NGN2 overexpression. Briefly, iPSCs were resuspended as single cells using Accutase for 2–3 min at RT. Single cells were plated at 15,000 cells/cm² onto 10 μg/mL poly-D-lysine (PDL) and laminin (LAM) coated culture plate in mTeSR1 media supplemented with 10 μM Y27632. Cells were allowed to attach for 6–8 h, after which 0.5 μL of viral particles of both NGN2 overexpression and the TTA per 15,000 cells. Virus was removed 16–20 h following transduction with fresh neural media [Neurobasal medium (NBM; Life Technologies, #21103–049) supplemented with 1× N-2 supplement 1× B-27 supplement, 1× Insulin-Transferrin-Selenium-A and 2 mM L-glutamine] supplemented with 1 μg/mL doxycycline (DOX; Sigma-Aldrich, # D9891), 10 μM SB431542 and 0.1 μM LDN193189 to promote a cortical fate. After 24 h of DOX induction, 0.5 μg/mL puromycin was added daily for 3 days for selection of successfully transduced cells, in addition to DOX, SB431542 and LDN193189. Following selection fresh media supplemented with 10 μg/mL BDNF was added. Following selection, BrainPhys media [Brainphys medium (StemCell Technologies, #05790) supplemented with NeuroCult SM1 (without vitamin A; StemCell Technologies, #05731) and N2 supplement-A (StemCell Technologies, #07152)] was subsequently added in at increasing concentrations (25–100% in NM) for each media change to improve maturation. Neurons were assessed by whole-cell patch clamp between 21 and 35 days post viral transduction.

Sterile plastic coverslips, cut into 10 mm slides, were coated with PDL and LAM before plating NGN2 iNs for functional characterization. Whole-cell patch clamp recordings were performed on matured neurons aged 3–5 weeks, following the protocol outlined in Hulme et al. (2020). Recordings were conducted at room temperature (20–22°C) using a MultiClamp 700B Amplifier, digitized with a Digidata 1,440, and controlled via pClamp11 software (Molecular Devices). Whole-cell membrane currents were measured at 100 kHz, with series resistance compensated at 60–80%. Fire-polished borosilicate patch pipettes with a resistance of 2–4 MΩ were employed with an intracellular buffer composed of (in mM) 140 K-gluconate, 10 NaCl, 2 MgCl₂, 10 HEPES, 5 EGTA (pH 7.2, osmolality 295 ± 5 mOsm/kg). The bath solution for current clamp experiments contained (in mM) 135 NaCl, 2 CaCl₂, 2 MgCl₂, 5 KCl, 10 glucose, 10 HEPES (pH 7.4, osmolality 315 ± 5 mOsm/kg).

Under current-clamp conditions, the resting membrane potential (RMP) was measured as the average membrane voltage without any current injection. Total membrane capacitance was estimated by automatically integrating the transient capacitive current produced during voltage-clamp steps, using pClamp11 (Molecular Devices). To analyze the intrinsic properties and excitability of iNs, 1-s current steps ranging from −150 to 140 pA in 10-pA increments were injected, starting from the RMP. Input resistance (Rin) was calculated from the steady-state voltage responses to 1-s hyperpolarizing currents, using the slope of the voltage–current relationship. Hyperpolarizing current induced membrane potential change (ΔMP) was defined as the difference between the minimum membrane potential recorded during the −100 pA current injection and the RMP. Rheobase was determined as the minimal current required to trigger an action potential (AP). AP characteristics were analyzed using Clampfit 11 (Molecular Devices). The AP threshold was defined as the voltage at which dV/dt exceeded 10 mV/ms, and AP amplitude was measured as the difference between the threshold and peak voltage. AP half-width was calculated as the duration at 50% of the AP amplitude. The threshold potential relative to the baseline was estimated by identifying the intersection of the AP’s rising phase and the slope leading up to its initiation, using a three-point tangent slope vector to find where the slope reached or exceeded 10 mV/ms. Under voltage-clamp conditions, total voltage-dependent sodium (INav) and potassium (IKv) currents were quantified at their peak amplitudes. INav was measured at the minimum current inflection at −10 mV, and IKv at the maximum current observed at 20 mV. These current values were then normalized by dividing by the capacitance of the respective iNs and reported as conductances in pA/pF.

Experiments were performed with 3–4 independent experimental differentiations (n = 3–4, biological replicates). Statistical analyses were conducted using GraphPad Prism software, version 10 (GraphPad Software, La Jolla, USA). Data was determined to be normally distributed using a Shapiro–Wilk normality test, where data were normally distributed, data were analyzed by a multiple t-test with Holm-Sidak method, unless stated otherwise. An α of 0.05 (p-value <0.05) was considered statistically significant.

We have previously demonstrated the forced overexpression of NGN2 successfully and robustly generates functional, glutamatergic excitatory induced neurons (iNs) (Maksour et al., 2024). This study generated iNs from iPSCs of fAD patients bearing the PSEN1^S290C (S290C) and PSEN1^A246E (A246E) and their respective CRISPR corrected counterparts, S290^IC and A246^IC, to assess their electrophysiological properties (Figure 1A). Previous studies from our team have shown that iNs derived from patients with S290C and A246E mutations do not exhibit tau or Aβ pathology within 35 days in culture (Targa Dias Anastacio et al., 2024). This model allows us to investigate neuronal excitability alterations in a controlled environment without the confounding effects of disease pathology or the presence of glial cells. Therefore, it provides a unique opportunity to study the intrinsic effects of these mutations on neuronal ion channel function. Neurite analysis of fAD and isogenic corrected control iNs was performed using the Incucyte live-cell imager to assess neurite properties (Supplementary Figure S2). The neurite length and branch points of S290C iNs were significantly decreased compared to the respective controls, further supporting the observation of decreased capacitance and consistent with our previous findings in AD neurons (Balez et al., 2016). Neuronal excitability was evaluated by manual patch-clamp in the whole-cell configuration with representative recordings displayed in Figure 1B with data for all parameters summarized in Supplementary Table S2.

The resting membrane potential (RMP) was evaluated within 2 min of switching to the current-clamp mode. RMP showed no difference between S290C and S290^IC iNs, while A246E iNs had significantly lower RMP compared to their isogenic controls (−63.3 ± 2.7 mV, n = 8 vs. −46.6 ± 2.9 mV, n = 9, respectively; p < 0.001) (Figure 1C). S290C iNs had significantly lower capacitance than S290^IC (13.0 ± 0.8 mV, n = 17 vs. 15.5 ± 0.6 mV, n = 41, respectively, p < 0.05) reflecting the generation of smaller iNs from this fAD line (Figure 1D), while A246E and A246^IC capacitance values were comparable. The iNs were stimulated through to 1 s long step current injections ranging from −150 pA to 140 pA. A comparison of the slopes of the steady-state voltage responses hyperpolarizing currents (−150 pA to −20 pA) revealed no differences in the input resistance (Rin) between the two fAD PSEN1 mutants and their corrected controls (S290C vs. S290^IC p = 0.2397; A246E vs. A246^IC p = 0.1087) (Figure 1E).

Negative current injections elicited hyperpolarizing membrane responses (sags) consistent with the activation of hyperpolarization-gated cyclic nucleotide-activated ion channels (HCN). Quantification of the change in membrane potential (ΔMP) evidenced upon −100 pA current injections revealed smaller hyperpolarizing sags in fAD iNs compared to their isogenic controls (S290C vs. S290^IC p = 0.016257; A246E vs. A246^IC p = 0.015170) (Figure 1F).

The rheobase (the minimum current injection required to induce one action potential, AP) was higher in both fAD iNs (S290C: 41.5 ± 9.8 pA, n = 13; A246E: 66.3 ± 9.8 pA, n = 8) compared to their corrected controls S290^IC (19.5 ± 3.0 pA, n = 21; p < 0.05) and A246^IC (22.2 ± 4.7 pA, n = 10; p < 0.01), respectively (Figure 1G). Notably, positive current injections (up to 140 pA) elicited AP firing in all iNs, but multiple action potentials were only observed in corrected iNs (Figure 1B). Accordingly, the number of APs fired upon current injections equivalent to two times the rheobase (#AP@2xRheo) was significantly lower in both fAD lines compared to corrected iNs (Figure 1H). Thus, iNs derived from PSEN1 S290C iPSCs fired 2.0 ± 0.7 action potentials (n = 12) compared to 6.6 ± 0.8 (n = 21) APs by S290^IC (p < 0.001), and PSEN1 A246E fired 1.4 ± 0.4 APs (n = 8) compared to 4.0 ± 1.0 APs (n = 9) by A246^IC iNs (p < 0.05). Thus, neurons derived from fAD patients with S290C or A246E mutations in PSEN1 exhibited reduced firing capacity and excitability compared to their isogenic controls, indicating a common phenotype of reduced neuronal excitability.

Given the observed inter-group variability in AP waveforms, the first action potentials fired by iNs from each group were aligned at the point of initiation and averaged for illustration purposes (Figure 2A). Comparative analysis of AP shapes (at rheobase) between fAD and corrected iNs unveiled disparities in AP peak amplitude, area, half-width, rise time, rise slope, and maximal decay slope, as outlined in Supplementary Table S1. While AP peak amplitude showed no variance between S290C and S290^IC neurons (Figure 2B), their AP area differed significantly, with S290C iNs exhibiting over a 5-fold smaller area than their isogenic control (2229.0 ± 423.5 mV*ms, n = 13; vs. 12036.2 ± 1390.3 mV*ms, n = 21; p < 0.001) (Figure 2C). Additionally, the AP rise slope was markedly higher in S290^IC-derived neurons (59.3 ± 9.5 mV/ms, n = 21), compared to those harboring the S290C mutation (25.6 ± 4.1 mV/ms, n = 13, p < 0.05) (Figure 2E). Consequently, the AP rise time was more than 2-fold slower in S290C neurons compared to S290C^IC (1.3 ± 0.2 ms, n = 13 vs. 0.6 ± 0.2 ms, n = 20, p < 0.05) (Figure 2F) with no discernible differences in AP maximal decay slope (p = 0.15) (Figure 2G).

In contrast, PSEN1 A246E fAD iNs exhibited higher AP peak amplitudes (99.7 ± 3.3 mV, n = 8) compared to corrected A246^IC iNs (72.9 ± 3.7 mV, n = 9; p < 0.001) (Figure 2B), while showing no significant difference in AP area (p = 0.44) (Figure 2C). Despite similar AP rise parameters (Figures 1C,E,F), A246E neurons displayed narrower APs than their isogenic control (AP half-width 4.2 ± 0.5 ms, n = 8 vs. 8.0 ± 1.2 ms, n = 9, respectively; p < 0.05) (Figure 2D), accompanied by markedly larger maximal decay slopes (−28.7 ± 3.8 mV/ms, n = 8 vs. −11.9 ± 1.4 mV/ms, n = 9, p < 0.01) (Figure 2G). Thus, the action potential shape in fAD iNs appears distinctively affected by specific PSEN1 mutations.

The shape of the action potential is intricately governed by factors, such as ion channel dynamics and the equilibrium between inward and outward currents, primarily mediated by voltage-gated sodium (Nav) and potassium (Kv) channels, respectively. Representative iN whole-cell currents elicited by 250 ms step depolarization (from −80 mV to 50 mV; Vh −80 mV, inset) recorded in voltage-clamp mode are displayed for S290C and S290IC (Figure 3A) and A246E and A246IC neurons (Figure 3B). To assess the ability of iNs to initiate and terminate action potential firing, Nav and Kv currents were analyzed. Nav channels, which drive rapid depolarization, open quickly near the threshold for action potential initiation (~ −40 to −20 mV). Sodium currents were measured at −10 mV to ensure the activation of a large fraction of available Nav channels. Kv channels, responsible for restoring the resting membrane potential and regulating action potential duration and firing frequency, activate more slowly at more depolarized potentials (~ −10 to +20 mV). Kv currents were therefore quantified at 20 mV to capture maximal outward potassium flow. Total Nav and Kv currents (at −10 mV and 20 mV, respectively) were normalized to the cell capacitance and expressed as current density (pA/pF). The disease lines and isogenic corrected controls demonstrated similar AP thresholds (Supplementary Table S2), indicating no significant differences in the voltage required to trigger an action potential. In alignment with these findings, no detectable differences in Nav channel-mediated currents were observed between these groups (Figure 3C). In contrast, Kv channel mediated currents were significantly larger in neurons derived from patients with fAD PSEN1 mutations (S290C: 161.5 ± 28.1 pA/pF, n = 17; A246E: 89.2 ± 7.1 pA/pF, n = 10), compared to their respective CRISPR-corrected isogenic controls (S290^IC: 97.0 ± 8.1 pA/pF, n = 41; p < 0.01; A246^IC: 50.7 ± 5.4, n = 10; p < 0.001) (Figure 3D). Consequently, the ratio between Nav and Kv current densities (INav/IKv) was significantly larger for both isogenic controls, compared to their respective fAD lines (p < 0.05) (Figure 3E). Similarly, the ratio of maximal Nav and Kv conductances (from conductance-voltage relationships) was higher in the corrected lines (Supplementary Table S2), suggesting a potential imbalance between inward and outward currents as a possible cause of reduced excitability in PSEN1 fAD iNs.