Estradiol reverses excitatory synapse loss in a cellular model of neuropsychiatric disorders

Abstract

Loss of glutamatergic synapses is thought to be a key cellular pathology associated with neuropsychiatric disorders including schizophrenia (SCZ) and major depressive disorder (MDD). Genetic and cellular studies of SCZ and MDD using in vivo and in vitro systems have supported a key role for dysfunction of excitatory synapses in the pathophysiology of these disorders. Recent clinical studies have demonstrated that the estrogen, 17β-estradiol can ameliorate many of the symptoms experienced by patients. Yet, to date, our understanding of how 17β-estradiol exerted these beneficial effects is limited. In this study, we have tested the hypothesis that 17β-estradiol can restore dendritic spine number in a cellular model that recapitulates the loss of synapses associated with SCZ and MDD. Ectopic expression of wildtype, mutant or shRNA-mediated knockdown of Disrupted in Schizophrenia 1 (DISC1) reduced dendritic spine density in primary cortical neurons. Acute or chronic treatment with 17β-estradiol increased spine density to control levels in neurons with altered DISC1 levels. In addition, 17β-estradiol reduced the extent to which ectopic wildtype and mutant DISC1 aggregated. Furthermore, 17β-estradiol also caused the enrichment of synaptic proteins at synapses and increased the number of dendritic spines containing PSD-95 or that overlapped with the pre-synaptic marker bassoon. Taken together, our data indicates that estrogens can restore lost excitatory synapses caused by altered DISC1 expression, potentially through the trafficking of DISC1 and its interacting partners. These data highlight the possibility that estrogens exert their beneficial effects in SCZ and MDD in part by modulating dendritic spine number.

Fig. 1. Acute 17β-estradiol (E2) treatment increases dendritic spine linear density in neurons overexpressing HA-DISC1 or HA-DISC1ΔCT and reduces DISC1 aggregation.

a Representative images of cortical neurons (DIV 26) transfected with GFP alone (control), GFP + HA-DISC1 or GFP + HA-DISC1ΔCT and treated with vehicle (Veh) or 17β-estradiol (E2) or not. b Quantification of spine linear density. Treatment of control cells with 17β-estradiol resulted in an increase in spine number. Overexpression of HA-DISC1 or HA-DISC1ΔCT caused a significant reduction in spine linear density. Treatment with 17β-estradiol increased spine density in HA-DISC1 or HA-DISC1ΔCT expressing cells to a level not statistically different to vehicle control cells (dendritic spine linear density/10 µm): control, 5.8 ± 0.21; control + 17β-estradiol, 7.27 ± 0.44; HA-DISC1, 3.9 ± 0.2; HA-DISC1 + 17β-estradiol, 5.22 ± 0.34; HA-DISC1ΔCT, 3.4 ± 0.36; HA-DISC1ΔCT + 17β-estradiol, 5.5 ± 0.37). c Representative images of DIV 26 cortical neurons expressing either HA-DISC1 or HA-DISC1ΔCT were treated with vehicle or 17β-estradiol for 30 min. Pseudo-color scheme is used to highlight regions where DISC1 clustering is occurring. Green arrow heads indicate DISC1 clusters in each condition. d Quantification of DISC1 clustering; both HA-DISC1 or HA-DISC1ΔCT formed large clusters along dendrites. Treatment with 17β-estradiol reduced DISC1 clustering in neurons expressing either HA-DISC1 or HA-DISC1ΔCT. Scale bar = 5 µm.

Fig. 2. Acute 17β-estradiol (E2) treatment restores dendritic spine density following long term DISC1 knockdown.

a, b Western blot of hEK293 cells expressing HA-DISC1 in presence of control shRNA vector (pGSuper) or DISC1-shRNA (a). Quantification of HA-DISC1 expression (b), normalized to β-actin demonstrates that DISC1_shRNA effectivity knockdown exogenous DISC1 as previously reported. c Representative images of cortical neurons expressing control shRNA vector (pGSuper) or shRNA against DISC1 (DISC1_shRNA) for 7 days; neurons were treated with vehicle (Veh) or E2 (10 nM) for 30 min. d Quantification of dendritic spine linear density (per 10 µm). As recently reported, expression of DISC1_shRNA for 7 days resulted in a reduction in spine linear density; as expected 30 min E2 increases spine density in control neurons. Intriguingly, E2 treatment for 30 min increased spine density in DISC1_shRNA expressing cells to control neuron levels (dendritic spine linear density (per 10 µm): control, 5.1 ± 0.3; control + 17β-estradiol, 7.25 ± 0.49; DISC1_shRNA, 3.4 ± 0.35; DISC1_shRNA + 17β-estradiol, 4.95 ± 0.27). Scale bar = 5 µm.

Fig. 3. Bi-directional redistribution of DISC1/PSD-95/kalirin-7 signalosome by 17β-estradiol (E2).

a Whole cell lysate, crude synaptosome (P2) and cytosol (S2) fractions of mixed sex cortical neurons treated with 17β-estradiol (E2) for 0 or 30 min analyzed by western blotting for expression of DISC1, PSD-95 or kalirin-7; β-actin was used as a loading control. Arrows indicate DISC1 and kalirin isoforms measured in analysis. E2 has no effect on overall expression levels of DISC1, kalirin-7 or PSD-95 within 30 min as seen in whole cell lysate. Treatment with E2 causes a reduction of DISC1 in crude synaptosome fraction and an increase in cytosolic fraction. E2 treatment resulted in an enrichment of kalirin-7 and PSD-95 within crude synaptosome fraction and concurrent reduction in cytosolic fraction. b Quantification of DISC1, PSD-95 and kalirin-7 enrichment in crude synaptosome and cytosol fractions. c GFP-expressing and DISC1 (440) stained cortical neurons treated with E2 for 30 min or not. Quantification of DISC1 cluster intensity within spines reveals a reduction in intensity within spines after E2 treatment. d Cortical neurons (DIV 25) fixed before and after treatment with E2 and immunostaining for kalirin-7 and PSD-95. E2 treatment (30 min) increases kalirin-7 puncta density (black bars). Assessment of number of kalirin-7/PSD-95 co-localized puncta revealed and increase in kalirin-7 and PSD-95 positive puncta (red bars), indicating that kalirin-7 is being targeted to synapses (*p < 0.01 (corrected for multiple comparisons) Student t-test; n = 12–15 cells per condition from 3 independent cultures). e GFP-expressing and kalirin-7 stained cortical neurons treated with E2 for 30 min or not. Quantification of kalirin-7 cluster intensity within spines reveals an enrichment within spines following treatment (*p < 0.05, Student t-test; n = 270–341 spines from 6 cells per condition from 3 independent cultures). Scale bar = 5 µm.

Fig. 4. Chronic repeated treatment with 17β-estradiol (E2) rescues DISC1-induced spine deficits and reduces DISC1 aggregation.

a Schematic of chronic repeated 17β-estradiol (E2) treatment of either control or HA-DISC1ΔCT expressing neurons. Cortical neurons were transfected with GFP alone (control) or GFP and HA-DISC1ΔCT at DIV 17. Cells were treated daily with vehicle or 10 nM 17β-estradiol on DIV17–20. Neurons were fixed on DIV21. b Representative confocal images of DIV25 cortical neurons treated as in a. c Quantification of dendritic spine linear density. Chronic repeated treatment with 17β-estradiol increased in spine density in control cells. Ectopic expression of HA-DISC1ΔCT caused a significant reduction in spine linear density. Chronic repeated treatment with 17β-estradiol increased spine density in HA-DISC1ΔCT expressing cells to a level not statistically different to vehicle control cells. d Representative images of DIV 25 cortical neurons expressing either GFP alone or HA-DISC1ΔCT were treated with vehicle or 17β-estradiol daily between DIV17 and 24. Pseudo-color scheme is used to highlight regions where DISC1ΔCT clustering is occurring. e Quantification of DISC1 clustering. Chronic repeated treatment with 17β-estradiol reduced HA-DISC1ΔCT clustering. Scale bar = 5 µm.

Fig. 5. Repeated treatment with 17β-estradiol (E2) increases the number of dendritic spines positive for pre- and post-synaptic markers.

a Representative confocal images of DIV25 mixed sex cortical neurons triple immunostained for GFP, PSD-95 (post-synaptic protein) and bassoon (pre-synaptic protein). Neurons were treated as outlined in Fig. 4a. b Quantification of linear density of total spine density (open bars) or spines containing PSD-95 (gray and blue solid/hatched bars). Chronic repeated treatment with 17β-estradiol increased the number of spines positive for PSD-95 in control and HA-DISC1ΔCT expressing neurons. c Quantification of linear density of total spine density (open bars) or spines with overlap for bassoon (light gray and light magenta solid/hatched bars). Chronic repeated treatment with 17β-estradiol increased the number of spines overlapping with bassoon in control and HA-DISC1ΔCT expressing neurons. d Quantification of linear density of total spine density (open bars) or spines containing PSD-95 and that overlapped with bassoon (dark gray and teal solid/hatched bars). Chronic repeated treatment with 17β-estradiol increased the number of spines positive for PSD-95 and bassoon in control and HA-DISC1ΔCT expressing neurons. Scale bar = 5 µm.