Early disruption of the CREB pathway drives dendritic morphological alterations in FTD/ALS cortical neurons

Abstract

Synaptic loss and dendritic degeneration are common pathologies in several neurodegenerative diseases characterized by progressive cognitive and/or motor decline, such as Alzheimer's disease (AD) and frontotemporal dementia/amyotrophic lateral sclerosis (FTD/ALS). An essential regulator of neuronal health, the cAMP-dependent transcription factor CREB positively regulates synaptic growth, learning, and memory. Phosphorylation of CREB by protein kinase A (PKA) and other cellular kinases promotes neuronal survival and maturation via transcriptional activation of a wide range of downstream target genes. CREB pathway dysfunction has been strongly implicated in AD pathogenesis, and recent data suggest that impaired CREB activation may contribute to disease phenotypes in FTD/ALS as well. However, the mechanisms behind reduced CREB activity in FTD/ALS pathology are not clear. In this study, we found that cortical-like neurons derived from iPSC lines carrying the hexanucleotide repeat expansion in the C9ORF72 gene, a common genetic cause of FTD/ALS, displayed a diminished activation of CREB, resulting in decreased dendritic and synaptic health. Importantly, we determined such impairments to be mechanistically linked to an imbalance in the ratio of regulatory and catalytic subunits of the CREB activator PKA and to be conserved in C9-ALS patient's postmortem tissue. Modulation of cAMP upstream of this impairment allowed for a rescue of CREB activity and an amelioration of dendritic morphology and synaptic protein levels. Our data elucidate the mechanism behind early CREB pathway dysfunction and discern a feasible therapeutic target for the treatment of FTD/ALS and possibly other neurodegenerative diseases.

Keywords: ALS; CREB; FTD; PKA; dendrites.

Fig. 1.

HRE cortical neurons display a reduction in dendritic branching and synaptic protein expression. (A). Representative images of HRE and ISO i³CNs neurons stained with MAP2 (magenta). The yellow boxes identify the neurons enlarged in the right panels. (B–C). Sholl analysis of dendritic crosses per concentric shells across the cell radius (5 μm increments, (B). Statistical analysis of the area under the curves (AUC, C) revealed a significant difference (paired t test, n = 4 biological replicates). (D–E). Skeleton analysis of HRE compared to ISO neurons revealed a significantly lower number of junctions (D) and branches (E). The frequency distribution graphs is shown on the Right in D and E (Fisher’s exact test, n = 36 for both HRE and ISO from 4 biological replicates). (F). Quantification of maximum branch length in ISO and HRE neurons. Data are shown as scattered dot plots of individual cells (Left; unpaired t test, n = 36 ISO and HRE) or as the average for each biological replicate (Right; paired t test, n = 4 biological replicates). (G–H). Representative images (G) and quantification (H) of dendritic surface expression of GluN2A (green). Data are shown as scattered dot plots of individual cells (Left; Mann–Whitney test, n = 49 ISO and 52 HRE neurons) or as the average for each biological replicate (Right; paired t test, n = 6 biological replicates). DAPI (blue) was used to label the cell’s nucleus in (A) and (G). Scale bars: 100 µm in A; 5 µm in G. Bars and lines are mean and SEM in all but D and E, where they indicate the median; *P < 0.05, ***P < 0.001.

Fig. 2.

pCREB levels are significantly lower in C9 HRE neurons. (A) Representative images of phosphorylated CREB (pCREB, green) in ISO and HRE neurons. A 16-bit heat map of the staining is shown on the right. Scale bars: 5 µm. (B) Quantification of the mean fluorescence intensity (mFI) demonstrates a significant reduction in pCREB levels in HRE neurons relative to ISO controls. Data are shown as scattered dot plots of individual cells (Left; unpaired t test, n = 88 ISO and 93 HRE from 6 independent replicates) or as the average for each biological replicate (Right; paired t test, n = 6 biological replicates). (C–F) Representative western blot (C) of pCREB and CREB expression compared to GAPDH, used as a loading control. Quantification of band intensities for CREB (D), pCREB (E), and pCREB to CREB ratio (pCREB/CREB, F) is shown (paired t test, n = 4 biological replicates). (G) qPCR analysis of BDNF mRNA levels reveals a significant decrease in C9 HRE neurons (paired t test, n = 7 biological replicates). For all graphs, lines indicate mean and SEM; *P < 0.05, **P < 0.01, ns: not significant.

Fig. 3.

PKA-dependent activation of CREB is selectively impaired in HRE cortical neurons. (A) Schematics of the experimental timeline. (B and C) Representative images (B) and quantification of mFI of pCREB (red, C) and of PKA catalytic subunit (Cα, green) in the whole cell (D) or nucleus (E) in ISO and HRE neurons before and after stimulation with FSK. (F) Representative western blot of pCREB in HRE and ISO neurons treated with 5 µM FSK for 30 min as shown in A. GAPDH was used as leading control. (G and H) Quantification of pCREB band intensities (E) or fold change of pCREB levels in FSK-treated neurons over TTX-only (Untr.) (F) in HRE and ISO neurons. Data are shown as scattered dot plots (n = 102-100-101 ISO and 90-100-104 HRE, left graphs in C–E, n = 4 in G) or as the average of the FSK/Untr fold change for each biological replicate (n = 6, right graphs in C-E, n = 4 in H). Statistical analyses were performed via two-way ANOVAs with Šídák’s post hoc test in C, D, E, and G (Left graphs), repeated measures (RM) two-way ANOVA with Fisher’s LSD post hoc test in C, D, and E (Right graphs) and G, and with the paired t test in H. For all, bars or lines indicate mean and SEM; *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001, ns: not significant.

Fig. 4.

PKA inhibition via H-89 treatment leads to a decrease in pCREB levels and dendritic arborization. (A and B) Representative images (A) and quantification (B) of pCREB levels (green) in ISO neurons treated chronically with 10 μM H-89 or vehicle alone (Untr.) for 14 d. MAP2 (magenta) and DAPI (blue) labeled the cell soma and nucleus, respectively. Data are shown as scattered dot plots (n = 35 Untr. and 38 H-89 treated ISO neurons, unpaired t test) or as the average for each biological replicate (n = 3, paired t test). (Scale bar, 10 µm.) (C) Representative black and white images of MAP2 stained ISO neurons treated with H-89 or vehicle alone (Untr.). (D–E) Sholl analysis (D) and quantification (E) of the area under the curve (paired t test, n = 3 biological replicates). (F–G) Skeleton-based quantification of the number of junctions (F) and branches (G) in H-89 treated neurons. The frequency distribution of the number of branches and junctions is shown (Fisher’s exact test, n = 41 Untr. and 50 H-89 ISO neurons from 3 biological replicates). (H) Quantification of maximum branch length in ISO neurons treated with H-89 versus untreated controls. Data are shown as scattered dot plots of individual cells (Left; unpaired t test, n = 41 Untr. and 50 H-89 ISO neurons) or as the average for each biological replicate (Right; paired t test, n = 3 biological replicates, P = 0.18). Bars and lines represent mean and SEM for all; *P < 0.05, **P < 0.01, ****P < 0.0001.

Fig. 5.

Elevated levels of PKA regulatory subunits contribute to decreased pCREB levels in HRE neurons. (A) Representative images of active PKA Cα (green) in untreated HRE and ISO neurons. A 16-bit heat map of the staining is shown on the right. (B and C) Quantification of total (B) or nuclear (C) Cα mFI revealed a significant reduction in HRE neurons compared to ISO controls. Data are shown as scattered dot plots (Left graphs, unpaired t test, n = 78 ISO and 77 HRE neurons) or as the average for each biological replicate (right graphs, paired t test, n = 5). (D–F) Representative blots (D) and quantification of Cα (E), R1β (F), and the ratio of R1β to Cα (G) (paired t test, n = 5). DAPI (blue) was used to label the nucleus. (Scale bars, 5 µm.) For all, lines represent mean and SEM; *P < 0.05, ****P < 0.0001, ns: not significant.

Fig. 6.

PKA dysregulation and pCREB reduction are conserved in C9-ALS postmortem brain tissue. (A and B) Representative images of postmortem brain tissue from control or C9-ALS individuals labeled for pCREB (red) and PKA Cα (green). White boxes indicate the area enlarged on the right, dashed lines identify the nucleus, while asterisks indicate the presence of lipofuscin accumulation. (C–E) Quantification of PKA Cα cytoplasmic (C) and nuclear levels (D) shows a significant reduction in C9-ALS patients versus control. (A) similar downregulation in the nuclear levels of pCREB (E) was also observed (unpaired t test, n = 5 C9-ALS patients and 5 controls). (F) The nuclear levels of PKA Cα and pCREB in control and ALS patients show a high degree of correlation (linear regression, n = 117 neurons for both controls and C9-ALS, R²= 0.2666). (G–H). Representative images of control (G) and C9-ALS tissue (H) labeled for PKA R1β (green). White boxes indicate the area enlarged on the Right. (I) Quantification of R1β levels shows significantly higher levels in C9-ALS patients compared to controls (unpaired t test, n = 5 C9-ALS patients and 5 controls). For all, bars indicate mean and SEM; *P < 0.05, **P < 0.01. MAP2 (grays) and DAPI (blue) were used to identify neurons and nuclei, respectively. (Scale bar, 50 μm.)

Fig. 7.

ROL treatment rescues PKA subunit homeostasis and pCREB levels in HRE neurons. (A) Representative western blot showing PKA Cα and R1β levels in ISO and HRE i³CNs treated with ROL or vehicle alone (Untr.) for 2 wk. (B) Quantification of R1β/Cα ratio (RM two-way ANOVA with Fisher’s LSD post hoc test, n = 6 biological replicates). (C and D). Representative images (C) and quantification (D) of ROL effect on pCREB levels (green) in ISO and HRE neurons. Data are shown as scattered dot plots (two-way ANOVA with Fisher’s LSD post hoc test, n = 45 to 54 ISO cells and 43 to 58 HRE neurons) or as average per biological replicate (RM two-way ANOVA with Fisher’s LSD post hoc test, n = 3). DAPI (blue) was used to identify the cell’s nucleus. (Scale bar, 5 µm.) (E and F) Representative western blot (E) and quantification (F) of pCREB levels in HRE neurons treated with ROL for 2 wk (paired t test, n = 5). Histone H3 (H3) was used as loading control in A and E. For all, lines represent mean and SEM; *P < 0.05, ***P < 0.001, ****P < 0.0001, ns: not significant.

Fig. 8.

Dendritic branching and synaptic proteins are rescued in C9 HRE via cAMP modulation. (A) Representative images HRE neurons treated with ROL or vehicle (Untr.). The images were inverted to highlight the MAP2-positive dendrites. (B and C) Sholl analysis (B) and quantification of the area under the curve (AUC, C) of HRE neurons treated with ROL (paired t test, n = 6 biological replicates). (D) Quantification of maximum branch length in Untr. and ROL-treated HRE neurons. Data are shown as scattered dot plots of individual cells (Left; unpaired t test, n = 95 Untr. and 90 ROL) or as the average for each biological replicate (Right; paired t test, n = 6 biological replicates). (E–F) Skeleton-based analysis of dendritic junctions (E) and dendrites (F). The binned frequency distribution is shown (Fisher’s exact test, n = 95 Untr. and 90 ROL HRE neurons from 6 biological replicates). (G–H) Representative images (G) and quantification (H) of GluN2A (green) surface levels in HRE neurons treated with ROL compared to untreated HRE controls. DAPI (blue) labels the cell’s nucleus. Data are shown as scattered dot plots of individual cells (Left, Mann–Whitney test, n = 75 Untr. and 68 ROL HRE neurons) or as the average for each biological replicate (Right, paired t test, n = 5). Lines represent mean and SEM; *P < 0.05, ***p <0.001, ****P < 0.0001. (Scale bars, 100 µm) in A, 5 µm in G.