Targeting the postsynaptic scaffolding protein PSD-95 enhances BDNF signaling to mitigate depression-like behaviors in mice

Abstract

Signaling mediated by brain-derived neurotrophic factor (BDNF), which is supported by the postsynaptic scaffolding protein PSD-95, has antidepressant effects. Conversely, clinical depression is associated with reduced BDNF signaling. We found that peptidomimetic compounds that bind to PSD-95 promoted signaling by the BDNF receptor TrkB in the hippocampus and reduced depression-like behaviors in mice. The compounds CN2097 and Syn3 both bind to the PDZ3 domain of PSD-95, and Syn3 also binds to an α-helical region of the protein. Syn3 reduced depression-like behaviors in two mouse models of stress-induced depression; CN2097 had similar but less potent effects. In hippocampal neurons, application of Syn3 enhanced the formation of TrkB-Gα_i1/3-PSD-95 complexes and potentiated downstream PI3K-Akt-mTOR signaling. In mice subjected to chronic mild stress (CMS), systemic administration of Syn3 reversed the CMS-induced, depression-associated changes in PI3K-Akt-mTOR signaling, dendrite complexity, spine density, and autophagy in the hippocampus and reduced depression-like behaviors. Knocking out Gα_i1/3 in hippocampal neurons prevented the therapeutic effects of Syn3, indicating dependence of these effects on the TrkB pathway. The findings suggest that compounds that induce the formation of PSD-95-TrkB complexes have therapeutic potential to alleviate depression.

Figure 1.. CN2097 mitigates depression-like behaviors in mice.

(A) Schematic diagram of the animal protocol. (B to G) Assessment of CN2097 on antidepressant-like behaviors in corticosterone model (B-D) and CMS model (E-G) mice. C57BL/6J mice were subcutaneously (s.c.) injected with corticosterone (CORT, 20 mg/kg) daily for 21 days (B-D) or subjected to chronic mild stress (CMS) for five consecutive weeks (E-G), thereafter CN2097 (1 or 10 mg/kg) or CN5135 (10 mg/kg) were administered i.p.. After 12 hours, tail suspension (TST; B and E), forced swim (FST; C and F) and sucrose preference (SPT; D and G) behavioral tests were performed. Data are presented as mean ± standard deviation (SD) from, n = 10 (B to D) or 14 (E to G) mice per group. **P < 0.01, ***P < 0.001, “n.s.” not significant (P > 0.05) by one-way ANOVA plus Tukey’s post hoc test.

Figure 2.. Syn3 promotes TrkB signaling in hippocampal neurons.

(A) Structure of the active cyclic peptide Syn3*. (B) Comparison of the amino acid sequences of Syn3 and CN2097 and SPR measurements for binding to the PSD-95 PDZ3cα domain. (C) Nuclear magnetic resonance (NMR) showing the binding of Syn3 with the PDZ3cα fragment of PSD-95. (D to F) Effect of Syn3 on TrkB signaling. Primary murine hippocampal neurons were pre-treated with Syn3 (0.2 μM) or PBS vehicle control (“Veh”) for 20 min, followed by BDNF (25 ng/mL) treatment for 1 to 20 min. Immunoblotting and analysis assessing the effect of Syn3 on the relative abundance of the indicated phosphorylated proteins in primary murine hippocampal neurons that were pre-treated with Syn3 (0.2 μM) or PBS vehicle control (“Veh”) for 20 min, followed by BDNF (25 ng/mL) treatment for 1 to 20 min. (G) Role of TrkB in the signaling effects of Syn3. Primary murine hippocampal neurons were treated with a lentiviral TrkB shRNA (“shTrkB”) or lentiviral scramble control non-sense shRNA (“shC”) for five days, after which neurons were stimulated with BDNF (25 ng/mL) for 20 min after a 20-min pretreatment with Syn3 (0.2 μM) (“BDNF+Syn3”). TrkB signaling was analyzed in whole cell lysates by immunoblotting for the indicated proteins. (H) Representative immunoblotting to assess TrkB-Akt-S6K signaling in primary murine hippocampal neurons treated with Syn3 (0.2 μM) alone for 1 to 20 min and were analyzed by the listed proteins in total lysates. (I) Immunoblotting analysis to assess the effect of Syn3 on insulin- or PDGF-induced Akt-S6K signaling. Primary murine hippocampal neurons were pre-treated with Syn3 (0.2 μM) or vehicle for 20 min, followed by treatment with insulin (1 μg/mL) or PDGF-BB (25 ng/mL) for 20 min. Data are presented as mean ± SD; in (D to I), n = 5 biological repeats. ***P < 0.001. “n.s.” not significant (P > 0.05) by one-way ANOVA plus Tukey’s post hoc test (D to F) or Student’s t test (G and I).

Figure 3.. Syn3 enhances BDNF-induced neurite outgrowth and dendritic spine formation in hippocampal neurons.

(A) Representative neuronal morphology of GFP-expressing primary murine hippocampal neurons pre-treated with Syn3 (0.2 μM) or PBS vehicle control (“Veh”) for 20 min, followed by BDNF (25 ng/mL) treatment for 24 hours. Scale bar = 50 μm. (B) Sholl analyses demonstrates that Syn3 increases neuronal complexity and the number of neurite crossings (n = 26 neurons). (C and D) Syn3 increases the number of PSD-95 puncta. Primary hippocampal neurons were pre-treated with Syn3 (0.2 μM) or PBS vehicle control (“Veh”) for 20 min, followed by BDNF (25 ng/mL) treatment for 24h. Representative images of immunofluorescence staining for PSD-95 (green fluorescence) and MAP2 (red fluorescence) are shown (C), scale bars = 10 μm, and the number of PSD-95 puncta per 10 μm dendrite from n = 25 neurons over five independent experiments are quantified in (D). (E to L) Inhibition of TrkB or PI3K-Akt blocks Syn3 effects on Akt-S6K signaling (E), PSD-95 puncta formation (F, quantified per 10 μm dendrite, n = 20 neurons per group over five independent experiments), synaptic density (G to I) and spine formation (J to L). Hippocampal neurons were first treated with either the TrkB inhibitor K252a (200 nM) or the pan PI3K-Akt-mTOR kinase inhibitor LY294002 (1 μM) for 1 hour; then Syn3 (0.2 μM) for 20 min, followed by BDNF (25 ng/mL) for 20 min for the immunoblotting in (E, n = 5 biological repeats) or 24 hours for immunofluorescence analysis (F to L). For synaptic density, representative dendritic staining for PSD-95 (green), synaptophysin (red) and MAP2 (cyan) is shown (G), and PSD-95-synaptophysin contacts per 10 μm dendrite was quantified (H, n = 25 neurons per group over five independent experiments), also in the presence of K252a or LY294002 (I, n = 20 neurons per group over five independent experiments). For spine density, representative F-actin-GFP positive hippocampal dendrites are shown (J); the number of F-actin-GFP positive spines per 10 μm dendrite was quantified (K, n = 20 neurons per group over five independent experiments), also in the presence of K252a or LY294002 (L, n = 15 neurons per group over five independent experiments). In (G and J), scale bars = 10 μm. Data presented as mean ± SEM (B) or SD (D-L). *P < 0.05 vs. “Veh” group (A); ^#P < 0.05 vs. “BDNF” only treatment (A); ***P < 0.001 (D-L), by two-way ANOVA with Sidak’s post hoc test (B) or one-way ANOVA plus Tukey’s post hoc test (D, E, F, H, I, K and L).

Figure 4.. Syn3 mitigates depression-like behaviors in mice.

(A) Schematic showing C57BL/6J mice subjected to 5-week of CMS exposure were i.p.-injected with Syn3 (1 mg/kg). (B to D) Effect of Syn3 depression-like phenotypes following CMS. Behavioral tests—tail suspension test (TST, B), forced swim test (FST, C) and sucrose preference test (SPT, D)—were performed after 12 hours in n = 25 mice per group. (E) Effect of Syn3 on the CMS-induced reduction in dendritic spines. Representative images (scale bar 10 μm) showing Golgi staining of dendritic spines in hippocampal CA1 neurons from mice described in (A). The number of stubby, mushroom, and thin spines were counted from 30-μm-long dendritic segments located ~50–80 μm from the soma of n = 26 randomly selected neurons from N = 7 mice per group. (F) Effect of Syn3 on the CMS-induced reduction in synaptic contacts. Representative low- and high-magnification images (scale bars = 10 and 2 μm) of PSD-95 and synaptophysin puncta in hippocampal CA1 brain sections (stratum radiatum subregion) from mice exposed to CMS. PSD-95-positive clusters, synaptophysin-positive clusters, and PSD-95–synaptophysin contacts were quantified from random views from sections from n = 16 mice per group. (G) Effect of Syn3 on PSD-95 and synaptophysin protein expression in hippocampal CA1 tissues of mice exposed to CMS. Immunoblotting of tissues of n = 6 mice per group, described in (A), were analyzed. (H) C57BL/6J mice with 5 weeks of CMS exposure were injected i.p. with Syn3 (1 mg/kg) or 7,8-DHF (10 mg/kg). After 24 hours, behavioral tests were performed, as in (B to D), in n = 16 mice per group. (I to L) Effect of Syn3 on depression-like phenotypes in the corticosterone model. C57BL/6J mice were injected with corticosterone (s.c., 20 mg/kg) daily for 21 consecutive days. Thereafter, Syn3 (1 mg/kg, i.p.) was administered, and 12 hours later performance on behavioral tests, as described in (B to D), was assessed in n = 9 mice per group. Data are presented as mean ± SD. *P < 0.05, **P < 0.01, and ***P < 0.001; “n.s.” not significant (P > 0.05); by one-way ANOVA plus Tukey’s post hoc test.

Figure 5.. Syn3 restores Akt-mTOR activation in hippocampus of mice subjected to CMS.

(A) Effect of Syn3 on CMS-suppressed Akt and S6K phosphorylation. C57BL/6J mice with 5-week of CMS exposure were i.p. injected with Syn3 (1 mg/kg) for 12 hours, and the expression of Akt and S6K proteins in hippocampal CA1 region (stratum radiatum subregion) tissues examined by immunoblotting, and phosphorylation quantified (n = 7 stands for tissues of seven different mice in each group). (B and C) Effect of Syn3 on p-Akt and p-S6 immunofluorescence staining in CA1 neurons. Representative p-Akt (Ser⁴⁷³) and p-S6 (Ser^235/236) fluorescence images in hippocampal CA1 brain sections were quantified from random views of sections from n = 5 mice in each group) were performed. (D to F) Effect of TrkB or PI3K-Akt-mTOR inhibition on Syn3’s antidepressant-like activity. After 5 weeks of CMS exposure, C57BL/6J mice then underwent bilateral intrahippocampal infusion of K252a (1 ng in 0.5 μL per side) or LY294002 (5 ng/mL in 0.5 μL per side), followed by Syn3 administration (i.p., 1 mg/kg) for 12 hours, and performance in the TST (D), FST (E) and SPT (F) behavioral tests was assessed in n = 16 mice per group. Data are presented as mean ± SD. *P < 0.05, **P < 0. 01, and ***P < 0.001 by one-way ANOVA plus Tukey’s post hoc test.

Figure 6.. Syn3 facilitates BDNF-induced TrkB-PSD-95-Gα_i1/3 complex formation in hippocampal neurons.

(A) Co-immunoprecipitation (Co-IP) assays assessing the effect of Syn3 on TrkB-PSD-95-Gα_i1/3 association. Primary murine hippocampal neurons were treated with BDNF (25 ng/mL, 5 min) alone or after Syn3 (0.2 μM, 20 min pretreatment), followed by co-IP and immunoblotting. Expression of the indicated proteins was assessed in whole lysate “Input” and TrkB-associated PSD-95, Gα_i1 and Gα_i3 were quantified from n = 5 biological repeats. (B) Role of Gα_i1/3 in the effects of Syn3 on BDNF signaling. Primary hippocampal neurons transduced with the lentiviral Gα_i1 shRNA plus lentiviral Gα_i3 shRNA (Gα_i1/3-shRNA) or the lentiviral scramble control shRNA (shC) were treated with BDNF (25 ng/mL) after a pretreatment with Syn3 (0.2 μM, 20 min). The relative Gα_i1/2/3 protein expression and Akt/S6 phosphorylation was quantified in immunoblots from n = 5 biological repeats. (C to E) Role of Gα_i1/3 in the effects of Syn3 on dendritic branching (C, images of GFP-expression and Sholl analyses quantifying the number of crossings), spine formation (D, images and quantification of dendritic F-actin GFP-expression) and synaptic density (E, images and analysis of PSD-95–synaptophysin contacts). Primary hippocampal neurons were infected with lentiviral Gα_i1/3-shRNA or a lentiviral scramble control shRNA (shC), treated with BDNF (25 ng/mL) after Syn3 (0.2 μM, 20 min pretreatment) for 24 hours. Sholl analyses in (C) was performed on 30 neurons per group; the number of spines per 10-μm dendrite in (D) was calculated from 17 neurons per group; and synaptic density in (E) was calculated from 16 neurons per group, each pooled from four independent experiments. (F) Effect of PSD-95 knockdown on the ability of Syn3 to enhance BDNF signaling. Primary hippocampal neurons transduced with the lentiviral PSD-95 shRNA (-s1/-s2) or the lentiviral scramble control shRNA (shC) were treated with BDNF (25 ng/mL) after Syn3 (0.2 μM, 20 min pretreatment), and total cell lysates were analyzed by immunoblotting for the indicated proteins, quantified from n = 5 biological repeats. Data are presented as mean ± SD. *P < 0.05 vs. “shC” (C), ***P < 0.001, “n.s.” not significant (P > 0.05) by one-way ANOVA plus Tukey’s post hoc test (A, D and F), Student’s t test (B and E), or two-way ANOVA with Sidak’s post hoc test (C).

Figure 7.. Syn3 fails to mitigate depression-like phenotypes in mice with neuronal conditional knockout of Gα_i1/3 in the hippocampus.

(A to B) Effect of Gα_i1/3 neuronal conditional knockout (Gα_i1/3-nCKO) on Syn3 signaling. Schematic, top right (A). AAV9-CMV-FLEX-Cas9-U6-sgGα_i1 (0.25 μL/side) plus AAV9-CMV-FLEX-Cas9-U6-sgGα_i3 (0.25 μL/side) were bilaterally injected into the CA1 region of the hippocampus of Syn1-Cre C57BL/6J mice, generating hippocampal Gα_i1/3-nCKO mice after three weeks. Control Syn1-Cre mice were injected with AAV9-CMV-FLEX-Cas9-U6-sgC (0.5 μL, “sgC”). Gα_i1/3-nCKO mice were injected with a single dose of Syn3 (1 mg/kg i.p.; pink bars) or not (red bars). mRNA expression (A) and protein abundance and phosphorylation by immunoblotting (B) 12 hours after Syn3 injection was quantified from n = 5 biological repeats. (C) Images and analysis of p-Akt (Ser⁴⁷³) staining in hippocampal CA1 brain sections from mice described in (A-B), quantified in sections from n = 5 mice in each group. Scale bar, 20 μm. Analysis beside bottom-right merged image. (D to F). Performance of Gα_i1/3-nCKO mice described in (A) on the TST (D), FST (E) and SPT (F) behavioral tests. n = 18 mice per group. (G) Images of Golgi-stained dendritic spines in the hippocampal neurons from Gα_i1/3-nCKO mice (scale bar, 10 μm), with assessment of spine density (below images, right). Dendritic spines were counted from 30-μm-long dendritic segments located 50–80 μm from the soma of n = 26 randomly selected neurons from N = 6 to 7 mice per group. (H to K) Imaging analysis of synaptic density in the hippocampal CA1 stratum radiatum subregion of Gα_i1/3-nCKO mice. The density of PSD-95 puncta (I), synaptophysin puncta (J) and PSD-95–synaptophysin contacts (K) were quantified in randomly-selected sections from n = 11 mice. Data are presented as mean ± SD. ***P < 0.001 vs. “sgC” group, “n.s.” not signficant (P > 0.05) by one-way ANOVA plus Tukey’s post hoc test.

Figure 8.. Syn3 ameliorates autophagy activation in hippocampal neurons following CMS.

(A and B) Autophagosome formation in the hippocampal CA1 region, assessed by electron microscopy (A) and immunofluorescence analysis of AAV-mRFP-GFP-LC3 reporter (B), in mice exposed to 5 weeks of CMS assessed 12 hours after i.p. injection with a single injection of Syn3 (1 mg/kg). Autophagosome quantification in (A) was performed in randomly selected hippocampal neurons from n = 9 mice per group, and from n = 30 randomly selected neurons from six mice each group in (B). (C to E) Western blotting of hippocampal tissues for markers of autophagic flux in mice after CMS exposure (C) and in Gα_i1/3 nKD mice (D) and Gα_i1/3 nCKO mice (E), after treatment with Syn3 (i.p. 1 mg/kg, 12 hours), from n = 5 mice per group. ***P < 0.001 (A to C); ***P < 0.01, **P < 0.01 vs. “shC”/“sgC” tissues (D and E); “n.s.” not significant (P > 0.05; D and E), by one-way ANOVA plus Tukey’s post hoc test. (F) The proposed signaling mechanism derived from this study. PSD-95, a master synaptic scaffold protein, complexes with TrkB to amplify BDNF signaling. In response to BDNF, PSD-95 associates with TrkB to recruit Gα_i1/3 resulting in amplification of PI3K-Akt-mTOR signaling. Syn3 facilitates BDNF-induced TrkB-PSD-95-Gα_i1/3 complex formation to restore Akt-mTOR activation in mice subjected to CMS to increase neurite outgrowth and dendritic spine formation in murine hippocampal neurons.