Ptchd1 deficiency induces excitatory synaptic and cognitive dysfunctions in mouse

Abstract

Synapse development and neuronal activity represent fundamental processes for the establishment of cognitive function. Structural organization as well as signalling pathways from receptor stimulation to gene expression regulation are mediated by synaptic activity and misregulated in neurodevelopmental disorders such as autism spectrum disorder (ASD) and intellectual disability (ID). Deleterious mutations in the PTCHD1 (Patched domain containing 1) gene have been described in male patients with X-linked ID and/or ASD. The structure of PTCHD1 protein is similar to the Patched (PTCH1) receptor; however, the cellular mechanisms and pathways associated with PTCHD1 in the developing brain are poorly determined. Here we show that PTCHD1 displays a C-terminal PDZ-binding motif that binds to the postsynaptic proteins PSD95 and SAP102. We also report that PTCHD1 is unable to rescue the canonical sonic hedgehog (SHH) pathway in cells depleted of PTCH1, suggesting that both proteins are involved in distinct cellular signalling pathways. We find that Ptchd1 deficiency in male mice (Ptchd1^-/y) induces global changes in synaptic gene expression, affects the expression of the immediate-early expression genes Egr1 and Npas4 and finally impairs excitatory synaptic structure and neuronal excitatory activity in the hippocampus, leading to cognitive dysfunction, motor disabilities and hyperactivity. Thus our results support that PTCHD1 deficiency induces a neurodevelopmental disorder causing excitatory synaptic dysfunction.

Figure 1

Ptchd1 gene expression profile in the mouse developing brain. Ptchd1 mRNA expression level was assessed by quantitative reverse transcriptase–PCR on cDNA arrays (Origene) from 48 mouse cDNA extracted from telencephalon, frontal cortex, posterior cortex, entorhinal cortex, olfactory bulb, hippocampus, striatum, mesencephalon, midbrain, rhombencephalon, pons, medulla, cerebellum, diencephalon, thalamus, hypothalamus and spinal cord. Five developmental stages were investigated: E13, E15, E18, postnatal day 7 (P7), and adult week 5 (P35). Expression of Ptchd1 mRNA is shown relative to the Gapdh reference gene. Data are presented as mean±s.e.m., with median. Black or red dot plots represent embryonic or postnatal stages, respectively. n=3 independent experiments.

Figure 2

PTCHD1 (Patched domain containing 1) is a synaptic receptor interacting with PSD95 and SAP102 but does not modulate the canonical sonic hedgehog signalling pathway. (a) Subcellular neuronal localization of PTCHD1-GFP (green fluorescent protein) in primary cultured hippocampal neurons. Representative images from confocal microscopy showing mature primary neurons expressing PTCHD1-GFP protein displaying punctate fluorescence throughout dendrites and dendritic spines. Hippocampal neurons were transfected at 12 days in vitro with PTCHD1-GFP (wild type (WT)), PTCHD1-GFP lacking the last 39 amino acids (PTCHD1delCter-GFP) or the last 15 amino acids (PTCHD1del874_888-GFP) and visualized 48 h later. Co-transfection with pDsRed was performed with PTCHD1delCter-GFP to visualize the entire cell. PTCHD1-GFP, n=15 neurons from 4 transfections; PTCHD1delCter-GFP, n=11 neurons from 3 transfections; PTCHD1del874_888-GFP, n=6 neurons from 2 transfections. Scale bars: 50 μm and 10 μm for PTCHD1-GFP and PTCHD1del874_888-GFP, 20 μm for PTCHD1delCter-GFP. (b) PTCHD1-GFP co-localizes with PSD95 in dendritic spines. Co-localization assay was performed using three primary neuronal cultures transfected with PTCHD1wtGFP (green) at 11 days in vitro (DIV11) and fixed at DIV14 and then immunostained with a mouse monoclonal PSD-95 antibody and a secondary Donkey anti-Mouse antibody (Alexa Fluor 594, red). Co-localization signal in spines (yellow) is indicated by white arrowheads. Data were collected from 10 transfected neurons with 31 dendrite sections in total (n=3 independent transfections). Scale bar: 10 μm. (c) Identification of a PDZ-binding motif in the C-terminal tail of PTCHD1 allowing interaction with PSD95 and SAP102 proteins. Predicted structure of PTCHD1 and sequence alignment of the 10 C-terminal amino acids of human proteins PTCHD1, PTCHD2, PTCHD3, PTCHD4, PTCH1 and PTCH2. The PDZ-binding motif is shown in red. PTCHD1 and PTCHD4 C-terminal sequences are identical between human and mouse. GST-pulldown assays using the C-teminal tail (39 amino acids) with or without the predicted PDZ-binding site as bait, and the synaptoneurosomal lysates from adult WT mouse cortex lysates as the prey, evidenced for the presence of a specific interaction between the predicted PDZ-binding motif and PSD95 and SAP102. The presence of a faint band in eluates from GST and from GST-Cter-delITTV using SAP102 antibody was considered as the non-specific background. A total of three independent experiments were performed. (d) PTCHD1 does not repress Gli reporter activity in Ptch1^−/− mouse embryonic fibroblasts (MEFs). Luciferase activity was assessed in Ptch1^−/− MEFs co-transfected with PTCHD1-GFP or PTCH1, and an 8 × Gli Luciferase Reporter Plasmid. A constitutively active renilla luciferase plasmid was also co-transfected to normalize for transfection efficiency. Luciferase values were measured using the Dual Luciferase Reporter Assay System (Promega) and normalized to an empty GFP control vector. Data were evaluated using a one-way analysis of variance test followed by a pairwise comparison of means using a Student’s t-test (with Bonferroni correction for multiple comparisons). Data are displayed as a box and whisker plot with a central line for the median, a box comprised of the first–third quartiles and whiskers showing the minimum and maximum values. n=3 independent experiments, with the displayed value for each being the average of duplicate plates of cells. *P<0.05.

Figure 3

Ptchd1^−/y males are hyperactive, show reduced anxiety and exhibit altered cognitive abilities in different learning paradigms. (a) Locomotor activity in the circadian activity over 35 h of testing presented per hour period and as a total during each phase of the cycle (n=12 wild-type (WT) and 11 Ptchd1^−/y mice; results are expressed as mean±s.e.m. for evolution graph or as a scattergram with the median; *P<0.05, **P<0.01, ***P<0.001 from WT, unpaired Student’s t-test); (b) locomotor activity and rears in the open field (n=12 WT and 11 Ptchd1^−/y mice; results are expressed as a scattergrams with the median; *P<0.05, **P<0.01, from WT, unpaired Student’s t-test); (c) the number of visits in the social recognition test (n=12 WT and 11 Ptchd1^−/y mice; results are expressed as a scattergram with the median; ***P<0.001 from WT, unpaired Student’s t-test); (d, e) anxiety-related behaviour as measured by the percentage of time in the centre of the open field arena (d), and by the percentage of entries and time in the open arms in the elevated plus maze (e) (n=12 WT and 11 Ptchd1^−/y mice; results are expressed as scattergrams with the median; *P<0.05, from WT, unpaired Student’s t-test); (f) working memory performance (n=12 WT and 11 Ptchd1^−/y mice; results are expressed as a scattergram with the median; *P<0.05 from WT, unpaired Student’s t-test); (g) duration of object exploration during acquisition (n=12 WT and 11 Ptchd1^−/y mice; results are expressed as a scattergram with the median; unpaired Student’s t-test), and object recognition performance during retention (n=12 WT and 11 Ptchd1^−/y mice; results are expressed as a scattergram with the median; §P<0.05 significantly different from Hasard (50%), one group t-test); (h) duration of immobility during habituation to the conditioning cage and during context in the fear conditioning and contextual and cued freezing performance (n=8 WT and 8 Ptchd1^−/y mice; results are expressed as mean±s.e.m.; repeated analysis of variance (ANOVA) followed by unpaired Student’s t-test for each time point; ANOVA for baseline habituation: gene F(1,14)=8.01, P=0.013, time F(1,14)=2.61, P=0.129, interaction F(1,14)=0.852, P=0.372; ANOVA for contextual fear: gene F(1,14)=10.01, P=0.0069, time F(1,14)=0.897, P=0.419, interaction F(1,14)=0.467, P=0.632; ANOVA for cued fear: gene F(1,14)=3.67, P=0.076, time F(1,14)=7.994, P=0.0134, interaction F(1,14)=0.149, P=0.705).

Figure 4

Analysis of the hippocampal transcriptome of Ptchd1^−/y reveals global increased expression of synaptic genes. (a) Clear segregation of wild-type (WT) and Ptchd1^−/y transcriptomes. Clustering of Pearson correlations (Ward’s linkage), an unbiased method to quantify the degree of similarity between data set, shows clear segregation between genotype and high concordance of biological replicas. (b) Upregulated genes impact the neuronal populations in the hippocampus. Radar plot representing the enrichments (hypergeometric test, Bonferroni corrected) of cell-type-specific markers (see Supplementary Methods) in the lists of upregulated and downregulated genes. Outer rings correspond to higher enrichments (P-values represented in Z-score scale). Red and blue shape occupy largely separated areas, that is, cell types. (c) Ptchd1 inactivation is associated with the upregulation of synaptic genes. Radar plot summarizing the major gene ontology (GO) enrichments for upregulated and downregulated genes. Outer rings correspond to stronger enrichments. Bonferroni-corrected P-values are represented in Z-score scale. Upregulated genes are strongly enriched in synaptic proteins while downregulated genes seem to be implicated, among other pathways, in the mitochondrial respiratory chain and the SRP-dependent co-translational protein targeting to membrane. (d) Detailed overview on the synaptic and neurodevelopmental genes upregulated in Ptchd1^−/y hippocampi. Histogram showing the Bonferroni-corrected P-values for the selected GO enrichments. The numbers indicate the amount of genes annotated with each term in the upregulated gene set and genomewide, respectively. Almost 25% of the PSD genes are upregulated in Ptchd1^−/y hippocampi. (e) Enrichments for experimental binding sites from Cistrome chromatin immunoprecipitation-seq database. Upregulated genes are enriched in the binding sites of Suz12, Acsl1 and Rest, whereas downregulated genes are not. Complete data in Supplementary Table S9.

Figure 5

Impairments of synaptic morphology and activity in Ptchd1^−/y mice. (a) Altered morphology of primary embryonic hippocampal cultures from Ptchd1^−/y embryos. Quantification of dendritic lengths, branching numbers and complexity from green fluorescent protein (GFP)-labelled hippocampal neurons at 18 days in vitro between wild-type (WT) and Ptchd1^−/y littermate embryos. Scale bar: 10 μm. Data are presented as mean±s.e.m., n=21 WT and 20 Ptchd1^−/y neurons. *P<0.05 and **P<0.01, Mann–Whitney (length of branches: P=0.005; and number of branches: P=0.0452) and two-way analysis of variance with Sidak's multiple comparisons (branching complexity: level 1, P>0.999; level 2, P>0.999; level 3, P=0.9668; level 4, P=0.4663; level 5, P=0.0119; level 6, P=0.0004) tests. (b) Reduced synaptic density in hippocampus from Ptchd1^−/y mice. Representative electron micrographs (EM) of synaptic contacts from hippocampal sections. Scale bar: 200 nm. Quantification of postsynaptic densities (PSDs) performed on a defined plane of 65 μm². A total of 3 WT and 3 Ptchd1^−/y brains (31 sections for each genotype representing 8–13 sections per animal). Graphical data are presented as scatter plots representing sections analysed for each animal±s.e.m., **P<0.01 (exact P=0.0085, calculated using the mean values per animal); unpaired two-tailed Student’s t-test. (c) Altered morphology of the excitatory synapses in Ptchd1^−/y mice. Higher magnification of EM is shown (white bars, dotted bars and white arrows indicate synaptic cleft width, PSD length and PSD width, respectively). All analysed excitatory synaptic junctions were defined by an asymmetric structural organization, including well-defined presynaptic (for example, vesicles) and postsynaptic (PSD) structures. Scale bars: 200 nm and 50 nm. (d) Comparison of the cumulative frequency for PSD length and width and for synaptic cleft width in WT and in Ptchd1^−/y hippocampi. A total of 3 WT brains (86 synapses splitted in 19, 30 and 37 structures in the respective brains) and 3 Ptchd1^−/y brains (79 synapses splitted in 17, 26 and 35 in the respective brains) were analysed. Graphical data are presented as mean cumulative frequency curves obtained for each genotype±s.e.m.. Kolmogorov–Smirnov test; PSD length, P=0.4154; PSD width, P=0.0310; synaptic cleft width, P=0.0008. (e–i) Reduced miniature excitatory postsynaptic current (mEPSC) frequency and increased release probability in Ptchd1^−/y mice. Example traces (e) and quantification of mEPSC amplitude (f) and frequency (g). Example traces (h) and quantification (i) of paired pulse ratio. Sample size (n) is indicated in the bars as cells per animals. Data are presented as mean±s.e.m., **P<0.01. Student’s t-test. KO, knockout.