c-Abl Deficiency Provides Synaptic Resiliency Against Aβ-Oligomers

Abstract

Spine pathology has been implicated in the early onset of Alzheimer's disease (AD), where Aβ-Oligomers (AβOs) cause synaptic dysfunction and loss. Previously, we described that pharmacological inhibition of c-Abl prevents AβOs-induced synaptic alterations. Hence, this kinase seems to be a key element in AD progression. Here, we studied the role of c-Abl on dendritic spine morphological changes induced by AβOs using c-Abl null neurons (c-Abl-KO). First, we characterized the effect of c-Abl deficiency on dendritic spine density and found that its absence increases dendritic spine density. While AβOs-treatment reduces the spine number in both wild-type (WT) and c-Abl-KO neurons, AβOs-driven spine density loss was not affected by c-Abl. We then characterized AβOs-induced morphological changes in dendritic spines of c-Abl-KO neurons. AβOs induced a decrease in the number of mushroom spines in c-Abl-KO neurons while preserving the populations of immature stubby, thin, and filopodia spines. Furthermore, synaptic contacts evaluated by PSD95/Piccolo clustering and cell viability were preserved in AβOs-exposed c-Abl-KO neurons. In conclusion, our results indicate that in the presence of AβOs c-Abl participates in synaptic contact removal, increasing susceptibility to AβOs damage. Its deficiency increases the immature spine population reducing AβOs-induced synapse elimination. Therefore, c-Abl signaling could be a relevant actor in the early stages of AD.

Keywords: Alzheimer’s disease; Aβ-oligomers; c-Abl tyrosine kinase; dendritic spines; synapse.

Figure 1

c-Abl is located at the synapse and its mRNA and protein levels increase after Aβ oligomers (AβOs) treatment. (A) c-Abl, PSD95 and SAP102 protein levels increase as the neuronal culture ages (from 7 to 20 DIV). The graph shows quantification of protein levels. n = 3 (B) Immunofluorescence of hippocampal neurons showing c-Abl (in green) from soma to dendrites at 15 DIV. (C) Immunofluorescence showing c-Abl (in green) and PSD95 (in red) in hippocampal neurons at 19 DIV. The graph shows protein colocalization by Pearson’s correlation. Scale bar = 20 μm and 2 μm for magnifications. (D) Hippocampal cultured neurons were treated with 5 μM AβOs at indicated times. mRNA expression for Abl1 gene was assayed by RT-PCR using gapdh as loading control (water negative control). (E) Western-blot of c-Al total protein levels normalized to βIII tubulin, as the graph shows (F). Neurons incubated with 5 μM fibrillary forms of the Aβ-peptide (Aβf) were used as control. Unpaired t-test, *p < 0.5.

Figure 2

Aβ oligomer-induced dendritic spine density decrease is independent of c-Abl. (A) GFP-transfected wild-type (WT) and c-Abl-KO neurons were treated with Aβ oligomers (AβOs) for 5 h, and dendritic spines were counted (PSD95 is shown in red). The sections of secondary dendrites (rectangle) were delimited to show dendritic spine morphology. Complete image scale bar = 20 μm, magnifications scale bar = 2 μm. (B) Quantification of spine density (number of spines/10 μm dendrite) shows that c-Abl-KO neurons show higher spine density (4.22 ± 0.22 spines/10 mm dendrite) than WT neurons (3.36 ± 0.20 spines/10 mm dendrite). On the other hand, AβOs treatment significantly reduces spine density in both, WT and c-Abl-KO neurons (2.45 ± 0.15 and 3.19 ± 0.15 spines/10 mm, respectively; n = 47 WT; n = 49 WT+AβOs; n = 49 KO, and n = 53 dendrites for KO+AβOs). Two-way ANOVA and Tukey’s multiple comparisons. *p < 0.5; **p < 0.01; ***p < 0.001. n = 3 independent cultures, 4–5 mice embryos per condition.

Figure 3

c-Abl modulates dendritic spine morphology: c-Abl deficiency increased formation of immature spines after Aβ oligomers treatment. (A) Dendrite sections of WT and c-Abl-KO neurons treated with AβOs present morphological variations on dendritic spines. Scale bar = 2 μm. Dendritic spines were classified into five categories: mushroom when the head (H) is wider (W) than the width of the neck (WN); stubby for neckless protrusions less than 0.75 μm length (L); thin for protrusions shorter than 2 μm; filopodium for protrusions longer than 2 μm. Finally, two-headed spines were classified as branched. (B) Analysis of spines suggests their distribution per 10 μm dendrite. Control WT and c-Abl-KO neurons have significantly more mushroom than treated neurons (WT: 0.04 ± 0.01 vs. KO: 0.06 ± 0.01 spines/10 μm dendrite), slightly decreased stubby (WT: 0.17 ± 0.02 vs. KO: 0.15 ± 0.02 spines/10 μm dendrite), slightly increased thin (WT: 0.18 ± 0.03 vs. KO: 0.19 ± 0.02 spines/10 μm dendrite) and filopodium (WT: 0.04 ± 0.01 vs. KO: 0.04 ± 0.01 spines/10 μm dendrite) and slightly increased branched spines per dendrite (WT: 0.02 ± 0.01 vs. KO: 0.03 ± 0.01 spines/10 μm dendrite). After AβOs treatment, both, WT and c-Abl-KO neurons showed a significant reduction in the density of mushroom (WT: 0.03 ± 0.01 and KO: 0.03 ± 0.01 spines/10 μm dendrite) and stubby spines (WT: 0.15 ± 0.01 and KO: 0.16 ± 0.03 spines/10 μm). After AβOs treatment WT neurons presented higher density of filopodia spines than c-Abl-KO neurons (0.07 ± 0.02 vs. 0.05 ± 0.01 spines/10 μm dendrite, respectively) and maintenance of thin spines (0.17 ± 0.02 vs. 0.18 ± 0.02 spines/10 μm dendrite). Branched spines showed no changes after AβOs treatment (WT: 0.01 ± 0.002 vs. KO: 0.02 ± 0.01 spines/10 μm dendrite). (C) Spine type relative profile [percentage of each spine type classified in (B) counting 2–5 dendrites per neuron]. In WT and c-Abl-KO neurons, the population of thin spines increases (WT: 37.2 ± 2.8% n = 14 neurons; KO: 41.3 ± 1.5% n = 15 neurons; WT+AβOs: 41 ± 3.9% n = 13 neurons; KO+AβOs: 43.4 ± 1.7%, n = 18 neurons) and mushroom spines decreases after AβOs treatment (WT: 11.3 ± 2.2%; KO: 12.5 ± 1.8%; WT+AβOs: 6.6 ± 1.6%; KO+AβOs: 7.9 ± 2.3%); while filopodia spines increased (WT: 8.6 ± 2%; KO: 8.9 ± 1.4%; WT+AβOs: 13.3 ± 1.9%; KO+AβOs: 11.6 ± 1.9%); Stubby (WT: 38.1 ± 3.5%; KO: 30.9 ± 2.1%; WT+AβOs: 35.4 ± 2.5%; KO+AβOs: 32.9 ± 3.3%); and branched (WT: 4.8 ± 1.0%; KO: 6.5 ± 1.1%; WT+AβOs: 3.3 ± 0.8%; KO+AβOs: 3.9 ± 1.0%). (D) The graph shows spine length for mushroom, branched, thin and filopodia spines. AβOs induced longer filopodia spines (WT+AβOs: 3.22 ± 0.19 μm, KO: 3.83 ± 0.26 μm). Two-way ANOVA, ***p < 0.001. n = 3 independent cultures, 4–5 mice embryos per condition. non-significant: ns; *p < 0.5; **p < 0.01.

Figure 4

c-Abl deficiency protects Piccolo/PSD95 synaptic clustering against Aβ oligomers. (A) Representative example of WT and c-Abl-KO neurons stained for the pre-synaptic marker Piccolo (green) and for the post-synaptic markers PSD95 (red). (B) AβOs treatment does not affect the clustering of Piccolo in c-Abl-KO neurons compared with WT-treated neurons (WT: 5.67 ± 0.19 vs. WT+AβOs: 3.82 ± 0.17 clusters/10 μm dendrite, and KO: 7.45 ± 0.18 vs. KO+AβOs: 6.41 ± 0.19 clusters/10 μm dendrite; WT: n = 75, WT+AβOs: n = 81, KO: n = 57 and KO+AβOs: n = 85 dendrites; C). AβOs induce a significant reduction of PSD95 protein clustering in WT neurons. However, not significant reduction in the number of PSD95 clusters as observed in c-Abl-KO neurons (WT: 5.72 ± 0.18 vs. WT+AβOs: 3.91 ± 0.14 clusters/10 μm dendrite, and KO: 5.27 ± 0.24 vs. KO+AβOs: 4.38 ± 0.15 clusters/10 μm dendrite; WT: n = 83, WT+AβOs: n = 86, KO: n = 63 and KO+AβOs: n = 85 dendrites). (D) AβOs-induced reduction of synaptic contacts between PSD95 and Piccolo affects WT neurons (WT: 5.59 ± 0.46 vs. WT+AβOs: 2.53 ± 0.2 contacts/10 μm dendrite), while c-Abl-KO neurons maintained intact synaptic contacts (KO: 5.04 ± 0.37 vs. KO+AβOs: 4.15 ± 0.27 contacts/10 μm dendrite; WT: n = 23, WT+AβOs: n = 33, KO: n = 33 and KO+AβOs: n = 24 neurons). Two-way ANOVA and Tukey’s multiple comparison test, ***p < 0.001. Scale bar = 20 μm and 2 μm for dendrite magnifications. non-significant: ns; **p < 0.01.

Figure 5

c-Abl ablation protects against Aβ oligomers-induced cell death and preserves Piccolo clusters in apoptotic and non-apoptotic neurons. (A,B) WT and c-Abl-KO neurons were treated with 5 μM AβOs for 5 h and Hoechst stained to label apoptotic nuclei (blue; PSD95 in red) (A); and co-stained with active caspase-3 (red) antibody to confirm apoptosis; cytoskeletal protein β-tubulin is shown in green (B). Graphs show the percentage of apoptotic nuclei (C) and the percentage of active caspase-3 nuclei (D) for each condition. (E–F) WT and c-Abl-KO neurons were classified into apoptotic vs. non-apoptotic neurons and analyzed for Piccolo (E) and PSD95 (F) cluster quantification per 10 μm dendrite. Both were significantly affected in non-apoptotic neurons, especially in WT neurons. While c-Abl-KO neurons display higher number of Piccolo clusters and were significantly preserved after AβOs treatment (WT: 6.52 ± 0.18 vs. WT+AβOs: 4.36 ± 0.21 clusters/10 μm dendrite, and KO: 8.19 ± 0.19 vs. KO+AβOs: 7.1 ± 0.24 clusters/10 μm dendrite; WT: n = 38, WT+AβOs: n = 43, KO: n = 29 and KO+AβOs: n = 42 dendrites). Apoptotic neurons displayed the same tendency for Piccolo clusters (WT: 4.81 ± 0.26 vs. WT+AβOs: 3.22 ± 0.23 clusters/10 μm dendrite, and KO: 6.77 ± 0.23 vs. KO+AβOs: 5.73 ± 0.25 clusters/10 μm dendrite; WT: n = 37, WT+AβOs: n = 38, KO: n = 29 and KO+AβOs: n = 42 dendrites). (F) PSD95 clusters strongly decreased in WT compared with c-Abl-KO non-apoptotic neurons while all conditions in apoptosis display significantly less clusters than healthy conditions (Non-apoptotic: WT: 6.76 ± 0.17 vs. WT+AβOs: 4.29 ± 0.17 clusters/10 μm dendrite, and KO: 5.7 ± 0.33 vs. KO+AβOs: 4.66 ± 0.22 clusters/10 μm dendrite; WT: n = 42, WT+AβOs: n = 44, KO: n = 36 and KO+AβOs: n = 42 dendrites; Apoptotic: WT: 4.65 ± 0.22 vs. WT+AβOs: 3.5 ± 0.21 clusters/10 μm dendrite, and KO: 4.69 ± 0.34 vs. KO+AβOs: 4.11 ± 0.19 clusters/10 μm dendrite; WT: n = 41, WT+AβOs: n = 42, KO: n = 27 and KO+AβOs: n = 43 dendrites; non-apoptotic: WT: n = 9, WT+AβOs: n = 10, KO: n = 7 and KO+AβOs: n = 9 neurons; apoptotic: WT: n = 9, WT+AβOs: n = 11, KO: n = 6 and KO+AβOs: n = 9 neurons). Two-way ANOVA and Tukey’s multiple comparison test, ***p < 0.001. Scale bar = 20 μm and 10 μm for magnifications. n = 2 independent cultures, 3–4 mice embryos per condition. ns: non-significant; *p < 0.5; **p < 0.01; ***p < 0.001.