Structural and functional alterations of neurons derived from sporadic Alzheimer's disease hiPSCs are associated with downregulation of the LIMK1-cofilin axis

Abstract

Background: Alzheimer's Disease (AD) is a neurodegenerative disorder characterized by the accumulation of pathological proteins and synaptic dysfunction. This study aims to investigate the molecular and functional differences between human induced pluripotent stem cells (hiPSCs) derived from patients with sporadic AD (sAD) and age-matched controls (healthy subjects, HS), focusing on their neuronal differentiation and synaptic properties in order to better understand the cellular and molecular mechanisms underlying AD pathology.

Methods: Skin fibroblasts from sAD patients (n = 5) and HS subjects (n = 5) were reprogrammed into hiPSCs using non-integrating Sendai virus vectors. Through karyotyping, we assessed pluripotency markers (OCT4, SOX2, TRA-1-60) and genomic integrity. Neuronal differentiation was evaluated by immunostaining for MAP2 and NEUN. Electrophysiological properties were measured using whole-cell patch-clamp, while protein expression of Aβ, phosphorylated tau, Synapsin-1, Synaptophysin, PSD95, and GluA1 was quantified by western blot. We then focused on PAK1-LIMK1-Cofilin signaling, which plays a key role in regulating synaptic structure and function, both of which are disrupted in neurodegenerative diseases such as AD.

Results: sAD and HS hiPSCs displayed similar stemness features and genomic stability. However, they differed in neuronal differentiation and function. sAD-derived neurons (sAD-hNs) displayed increased levels of AD-related proteins, including Aβ and phosphorylated tau. Electrophysiological analyses revealed that while both sAD- and HS-hNs generated action potentials, sAD-hNs exhibited decreased spontaneous synaptic activity. Significant reductions in the expression of synaptic proteins such as Synapsin-1, Synaptophysin, PSD95, and GluA1 were found in sAD-hNs, which are also characterized by reduced neurite length, indicating impaired differentiation. Notably, sAD-hNs demonstrated a marked reduction in LIMK1 phosphorylation, which could be the underlying cause for the changes in cytoskeletal dynamics that we found, leading to the morphological and functional modifications observed in sAD-hNs. To further investigate the involvement of the LIMK1 pathway in the morphological and functional changes observed in sAD neurons, we conducted perturbation experiments using the specific LIMK1 inhibitor, BMS-5. Neurons obtained from healthy subjects treated with the inhibitor showed similar morphological changes to those observed in sAD neurons, confirming that LIMK1 activity is crucial for maintaining normal neuronal structure. Furthermore, administration of the inhibitor to sAD neurons did not exacerbate the morphological alterations, suggesting that LIMK1 activity is already compromised in these cells.

Conclusion: Our findings demonstrate that although sAD- and HS-hiPSCs are similar in their stemness and genomic stability, sAD-hNs exhibit distinct functional and structural anomalies mirroring AD pathology. These anomalies include synaptic dysfunction, altered cytoskeletal organization, and accumulation of AD-related proteins. Our study underscores the usefulness of hiPSCs in modeling AD and provides insights into the disease's molecular underpinnings, thus highlighting potential therapeutic targets.

Keywords: Alzheimer’s disease; Human neurons; LIMK1; Neurites; Synaptic function; hiPSCs.

Fig. 1

Generation and characterization of hiPSCs from HS and sAD patients. A Schematic representation of the reprogramming process used to generate hiPSCs from HS and sAD fibroblasts. Human somatic cells from both HS (n = 5) and sAD (n = 5) donors were reprogrammed into hiPSCs using Sendai virus. B Table summarizing demographic data and clinical information of the donors used for generating HS and sAD hiPSCs. The table includes information on age, gender, ethnicity, APOE genotype, and Clinical Dementia Rating (CDR). (C) Representative of hiPSCs exhibiting pluripotent stem cell morphology (in contrast phase) and immunofluorescence images showing the expression of pluripotency markers in HS- and sAD-derived hiPSCs. Cells were stained for SOX2 (green), NANOG (green), OCT3/4 (green), SSEA-4 (red), and TRA-1–60 (red). DAPI (blue) was used to label nuclei. Both HS and sAD hiPSCs are positive for these pluripotency markers, confirming their stem cell identity. Scale bars: 50 µm

Fig. 2

Differentiation of hiPSCs into neurons and analysis of neuronal markers in HS and sAD neurons. A Schematic representation of the Ngn2-mediated differentiation process used to convert hiPSCs from HS and sAD donors into neurons. Neurons were induced using doxycycline (Doxy) and selected with puromycin (Puro). B Diagram of the lentiviral constructs used in the differentiation process, including Ngn2, EGFP, and Puromycin resistance genes. C Timeline of neuronal differentiation showing key steps from hiPSC plating (DIV0), doxycycline induction (DIV1), puromycin selection (DIV3), and final analysis at DIV30. D Immunofluorescence staining of neuronal markers in HS and sAD-derived neurons at DIV30. DAPI (blue) stains nuclei, EGFP (green) marks neurons, and NEUN (red) marks mature neurons. Merged images confirm co-localization of neuronal markers. Scale bars: 50 µm. E Immunofluorescence staining of the neuronal marker MAP2 (red) in HS and sAD-derived neurons, with DAPI and EGFP labeling. F Quantification of MAP2 and NEUN median fluorescence intensity (in arbitrary units, a.u.) in HS and sAD neurons, showing no significant difference (n.s.) between groups (n = 5/group). G Representative Western blot images of MAP2 and NEUN in HS and sAD neurons, with GAPDH as the loading control. H Densitometric analysis of MAP2 and NEUN expression in HS and sAD neurons showing no significant differences in fold change(n = 5/group). Each dot represents the average of three independent experiments. Data are presented as mean ± SEM. n.s. p > 0.05, assessed by Mann Whitney Rank Sum test

Fig. 3

Neurons obtained from sAD-hiPSCs recapitulate the typical hallmarks of AD. A Immunofluorescence staining for Aβ using 6E10 antibody (red), DAPI (nuclei, blue), and EGFP (neurons, green) in HS and sAD hiPSC-derived neurons. sAD neurons show strong 6E10 staining, indicating the presence of Aβ pathology. Scale bars: 50 µm. B Quantification of 6E10 median fluorescence intensity (in arbitrary units, a.u.) showing a significant increase in Aβ accumulation in sAD-hNs neurons compared to HS-hNs; (n = 5/group). C Representative dot blot images of Aβ levels in DIV30 HS- and AD-hNs lysates; (D) Bar graphs showing the band intensity analysis of Aβ in HS- and sAD-hNs. Red Ponceau, RP, was used as loading index and used to normalize sample; synthetic oligomeric Aβ42 (200 nM) was used as positive control; (E) Representative WB images of p-TAU (Thr181, Ser199, Thr205, Thr217 and Thr231) levels in DIV30 HS- and sAD-hNs lysates; (F) Densitometric analysis of p-TAU (Thr181, Ser199, Thr205, Thr217 and Thr231), normalized to total human TAU (Ht7) protein levels; n = 5/group. GAPDH was used as a loading control;; norm. normalized protein levels. Each dot represents the average of three independent experiments conducted on n = 5/group. Data are presented as mean ± SEM, n.s. p > 0.05, ** p < 0.01 vs HS-hNs, assessed by Mann Whitney Rank Sum test

Fig. 4

sAD-hNs show altered neurite morphology compared to HS-hNs. A Cell body morphology and neurite length of HS- and sAD-hNs from DIV1 to DIV30 were measured by performing fluorescence-based NeuroTrack analysis (NT) in IncuCyte^® time-lapse microscopy system by exploiting EGFP-expression of hNs; Cell-body cluster area (in yellow), Neurites length (in red). B Dot plot graph showing cell body cluster area quantification of neuronal cells during 30 days of differentiation (n = 5/group); (B) Quantification of neurite length per neuron across different time points (DIV1, DIV14, DIV21, DIV30) shows a significant increase in neurite complexity in sAD neurons at DIV14, DIV21, and DIV30, while there is no significant difference at DIV1 (n = 5/group). C Analysis of cell bodies shows no significant differences between HS and sAD neurons across time points (n = 5/group). D Quantification of the cell viability at different time points, showing no significant difference between HS and sAD neurons (n = 5/group). E Representative camera lucida drawings from HS and sAD neurons at DIV30 with Sholl analysis. F–H Bar graphs showing differences in dendritic length (F), total number of bifurcating nodes (G) and total number of dendritic intersections (H) indicating decreased branching complexity in sAD neurons compared to HS neurons. Each dot in Figs B,C,D represents a triplicate of an experiment conducted on n = 5/group. Each dot in Figs F–H represents a single cell analysis conducted on n = 5/group. Data are presented as mean ± SEM, n.s. p > 0.05, *** p < 0.001, **** p < 0.0001 vs HS-hNs, assessed by assessed by 2way ANOVA-Sidak’s multiple comparisons

Fig. 5

Electrophysiological properties of neurons derived from HS and sAD patients. A Summary graphs of membrane potential values and (B) input resistance values recorded from HS- and sAD-hNs. C Representative traces showing current-evoked action potentials. D Frequency of action-potentials plotted against current pulse values. E Representative traces of spontaneous excitatory postsynaptic currents (sEPSCs) in HS (gray) and sAD (orange) neurons. F, G Quantification of sEPSC amplitude (F) shows no significant differences (n.s., p > 0.05), while frequency (G) is significantly decreased in sAD neurons compared to HS neurons (p < 0.01, n = 12 cells per group). H, I Analysis of rise time (H) and decay time (I) reveals no significant differences between HS and sAD neurons; n = 12 cells per group. Data are presented as mean ± SEM, n.s. p > 0.05, * p < 0.05 assessed by two tailed Student’s t test

Fig. 6

Neurons obtained from sAD-hiPSCs show impairments in synaptic protein expression. A Representative WB images of pre- and postsynaptic proteins in DIV30 HS- and sAD-hNs lysates; Densitometric analyses of presynaptic proteins: SYN-1, SYP and postsynaptic proteins: PSD95,GluA1; GAPDH was used as loading control; n = 5/group. C Immunofluorescence staining of synaptic markers SYN-1, SYP, PSD95, and GluA1 in HS and sAD neurons. Co-localization with EGFP (neuronal marker) is shown. sAD neurons display decreased synaptic marker expression. Scale bars: 50 µm. D Quantification of median fluorescence intensity (MFI) in HS and sAD neurons shows a significant reduction in SYN-1, SYP, PSD95, and GluA1 in sAD neurons compared to HS neurons; n = 5/group. Each dot represents the average of three independent experiments conducted on n = 5/group. Data are presented as mean ± SEM, * p < 0.05, ** p < 0.01 vs HS-hNs, assessed by Mann Whitney Rank Sum test

Fig. 7

sAD-hNs display reduced phosphorylation levels of actin-modulating proteins LIMK1 and cofilin. A Representative WB images of p-PAK1 Ser204, p-LIMK1 Thr508 and p-Cof Ser3 levels in HS- and AD-hNs lysates at DIV30; (B) Bar graphs showing the densitometric analysis of p-PAK1 Ser204, p-LIMK1 Thr508 and p-Cof Ser3 levels in HS- and AD-hNs (n = 5/group). GAPDH was used as a loading control. norm. normalized protein levels. Each dot represents the average of three independent experiments conducted on n = 5/group. Data are presented as mean ± SEM, n.s. p > 0.05, * p < 0.05, ** p < 0.01 vs HS hNs, assessed by Mann Whitney Rank Sum test

Fig. 8

Effect of BMS-5 treatment on neurite outgrowth and morphology in neurons derived from HS and sAD patients. A Representative images of neurons treated with vehicle (Veh) or BMS-5, an inhibitor of LIMK1, in HS and sAD neurons at DIV30. EGFP (green) shows the neuronal processes, and yellow indicates neurite tracing. Neuronal complexity appears reduced in sAD neurons compared to HS, and BMS-5 treatment does not rescue neurite complexity in sAD neurons. Scale bars: 50 µm. B Quantification of neurite length per neuron shows no significant difference between BMS-5-treated HS-hNs and sAD-hNs treated with veh or BMS-5 (n = 5/group). BMS-5-treated HS-hNs showed significantly reduced neurite length than Veh-treated HS-hNs. Quantification of cell bodies (C) and cell viability (D) shows no significant difference between vehicle and BMS-5-treated neurons in both HS and sAD groups (n = 5/group). E Representative images of camera lucida drawings obtained from HS and sAD neurons treated with vehicle or BMS-5. F–H Results of Sholl analysis showing the quantification of neurite length, number of bifurcating nodes, and total number intersections. No significant differences were found between BMS-5-treated HS-hNs and sAD-hNs treated with vehicle (n = 5/groups) or BMS-5 (n = 5/groups). BMS-5-treated HS-hNs showed significantly reduced neurite length than Veh-treated HS-hNs. Each dot in Figs B,C,D represents a triplicate of an experiment conducted on n = 5/group. Each dot in Figs F,G,H represents a single cell analysis conducted on n = 17 neurons from n = 5 for HS + Veh; n = 29 neurons from n = 5 for HS + BMS-5; n = 17 neurons from n = 5 for sAD + Veh; n = 15 neurons from n = 5 for sAD + BMS-5). Data are presented as mean ± SEM, n.s. p > 0.05, ** p < 0.01, ****p < 0.0001 vs HS hNs, assessed by assessed by 2way ANOVA-Sidak’s multiple comparisons