Fig 2: Subcellular Distribution of EAAT3 Is Dynamically Regulated during Epilepsy in a SorCS2-Dependent Manner(A) Immunostaining of EAAT3 (red) and SorCS2 (green) in cultured WT neurons (mixed cortical-hippocampal culture) non-treated (control) or treated with 200 µM H2O2 for 30 min. Cells were counterstained with DAPI. Scale bar, 10 µm.(B) Extent of colocalization of EAAT3 and SorCS2 in WT neurons under control conditions and after H2O2 treatment as exemplified in (A) (tM1, fraction of EAAT3 overlapping with SorCS2). Mean ± SEM. n = 10 neurons per condition. *p < 0.05 (unpaired t test). This experiment was replicated three times.(C) Immunodetection of SorCS2 (red) and EAAT3 (blue) in CA1 and CA2 regions of the human control hippocampus and of the hippocampus of a patient with TLE and hippocampal sclerosis. Higher magnification insets are shown. Scale bar, 100 µm.(D) Distribution of EAAT3 in subcellular brain fractions from WT and S2KO mice under basal conditions (control) or after PTZ kindling (PTZ). ext, extrasynaptic membranes; syn, synaptosomes; tot, total lysate. In tissue lysates and fractions, only monomeric EAAT3 is present. Synaptophysin and PSD95 were used as fraction markers.(E) Ratio of EAAT3 (monomer) in extrasynaptic membranes (ext) versus P3 fractions from WT and S2KO mouse brains as exemplified in the western blot analysis in (D). A shift of EAAT3 from P3 to extrasynaptic membranes is seen in WT but not S2KO brains upon PTZ kindling. Mean ± SEM. n = 6 samples per group. *p < 0.05 (two-way ANOVA) for interaction.

Fig 3: Conditioned medium from Aß-stimulated astrocytes exaggerates the damaging effects of Aß at synapses. A Dendritic spine density was measured in neurons transfected at 7DIV with GFP and treated with WTCM astro, TGCM astro and TGCM astro-ID or directly with WTCM and TGCM at 13DIV for 24 h. Representative images are shown. Scale bar is 10 µm. B Quantification of spine density in 10–20 neurons per experimental replicate from eight individual neuron preps (n = 8). TGCM astro caused increased dendritic spine loss relative to treatment with TGCM directly, and this is maintained when Aß is immunodepleted from the medium. Box indicates data from neurons directly exposed to TGCM or WTCM for clarity. Data was analysed by ordinary one-way ANOVA and Tukey’s post-hoc tests. C Neuronal lysates were immunoblotted using antibodies against PSD-95 (95 kDa), synapsin-1 (77 kDa), cleaved caspase-3 (17–19 kDa) and ß-actin (42 kDa). Representative blots are shown. Western blot band intensities were quantified and showed that TGCM astro caused reductions in the abundance of the synaptic markers D PSD-95 and E synapsin-1 and an increase in F cleaved (active) caspase-3, that is maintained when Aß was immunodepleted (TGCM astro-ID). Data is from eight individual neuron preps (n = 8), each of which have two experimental replicates. Data was analysed by ordinary one-way ANOVA and Tukey’s multiple comparison post-hoc tests. G Neuronal complexity as a measure of neuron health was analysed for all cells in three wells of five biological replicates (n = 5) using Harmony software. This showed H reduced total neurite length, and fewer number of I nodes and J roots following treatment with TGCM astro and TGCM astro-ID, indicating that astrocytic secretions are damaging to neurons. Data was analysed by ordinary one-way ANOVA and Tukey’s multiple comparison post-hoc tests. K Astrocytes were cultured and treated in cell culture inserts, the medium removed by washing and the astrocytes added to neuron cultures. Representative western blots of neuronal lysate from co-culture experiments are shown. Western blot data was quantified by normalising the protein of interest to ß-actin levels in the same sample and showed reduced L PSD-95 and M synapsin-1, and N increased levels of cleaved caspase-3 when neurons were co-cultured with astrocytes that had previously been exposed to TGCM (n = 6, data is from six independent experiments, each of which had three technical replicates). Data was analysed using an unpaired t test. Data on graphs is mean ± SEM and is shown as percentage average control (WTCM astro). *p < 0.05, **p < 0.01

Fig 4: KOR agonist inhibits Ca2+ overload and improves synaptic plasticity in amyloid precursor protein/presenilin-1 mice. (A) ELISA analysis demonstrated the Glu concentration of mice in each group. (B) Western blot analysis of NMDAR, CaMKII, p-CaMKII, CREB and p-CREB expression levels of mice in each group. (C) Golgi-Cox staining of brain tissue samples. Scale bar, 10 µm. (D) Western blot analysis demonstrating NCAM, NR2B, GluR1, PSD95 and SYN expression levels of mice in each group. *P<0.05 vs. Control group; #P<0.05 vs. AD group. Experiments were performed in triplicate and repeated three times. KOR, ?-opioid receptor; Glu, glutamate; NMDAR, N-methyl-D-aspartate receptor; CaMKII, calcium/calmodulin-dependent protein kinase II; p, phosphorylated; CREB, cyclic adenosine monophosphate response element binding protein; NCAM, neural cell adhesion molecule; NR2B, N-methyl D-aspartate receptor subtype 2B; GluR1, glutamate receptor 1; PSD95, postsynaptic density protein 95; SYN, a-synuclein; AD, Alzheimer's disease.

Fig 5: Interneuron-specific expression of Dact1, Dvl1, or DISC1 rescues excitatory synapse numbers in IDact1-KO interneurons. A–B Infected interneurons were detected by expression of mCherry. Characterization of PSD95 puncta on primary dendrites of cultured GFP+ cortical interneurons from control (A) and I-Dact1KO (B) mice infected with IS-lentiviral constructs expressing (i) mCherry alone, (ii) Dact1, (iii) Dvl1, (iv) RacCA, or (v) DISC1. C Quantification of PSD95 puncta per 10 µm of primary dendrite length in control (open bars) and IDact1-KO mutant neurons (closed bars). Results are presented as mean ± sem from at least 2 independent experiments, collected from at least 2 mice per genotype, 10–15 neurons per animal. *p<0.05; ***p<0.001; n.s., not significant. All p values are relative to either control (IS-lentivirus, none) or IDact1-KO (IS-lentivirus, none). Scale bars = 10 µm.