The following sections are expanded upon in Supplementary Methods.
We collected biological and clinical data from nine individuals from nine unrelated families, through to a multicenter and international collaboration through GeneMatcher [12]. Inclusion in the study was based on the identification of a de novo DPSYL5 variant, except for one adopted individual whose biological parents were not assessed (Table 1). Detailed clinical data including medical and developmental history, brain imaging and any other relevant medical information were collected through a standardized clinical form completed by referring physicians. Genetic analyses were performed at each clinical center through routine clinical and genetic diagnostic pipelines of exome or genome sequencing. DNA was extracted either from peripheral blood mononuclear cells (PBMC) or from amniotic cells for fetuses. The allelic frequency of each nucleotide variation was determined using gnomAD (v3.1 non neuro, and v4.1), and predicted deleterious impact on the protein was assessed by in silico prediction softwares, such as CADD, REVEL, ClinPred and alphaMissense [13,14,15,16]. The variants annotations were checked using Variant Validator [17], and are referenced to GRCh38, NM_020134.4 and NP_064519.2 reference sequence numbers (Supplementary Table 1).
Full-length human DPYSL5 cDNA (GenBank AF264015) was amplified by PCR and inserted directionally into the pEGFP-C1 vector (Clontech), as previously described [8]. The six variants were generated by site-directed mutagenesis using the Q5Site-Directed Mutagenesis Kit (New England Biolabs). The oligonucleotides primers sequences for each variant are indicated in Supplementary Table 2. The presence and the correct insertion of each nucleotide substitution on the GFP-DPYSL5 plasmid was confirmed by Sanger sequencing.
Hippocampal and cortical cultures were prepared from embryonic day 17.5 (E17,5) C57BL/6 J WT mouse embryos (Janvier Labs) as previously described [18]. The developing neurons were transfected with pCAG-GFP and different plasmid constructs (WT or mutated forms of GFP-DPYSL5) using Lipofectamine 2000 (Invitrogen) with 1 µg of DNA/1,6 µl of L2000 diluted in Neurobasal medium (Invitrogen), either at 4 days after plating (day in vitro, DIV 4) to analyze neurite growth, or at DIV 11 to study synapses morphology and density. The transfection mix was added to cells and incubated for 4 h (37 °C, 5% CO), then the transfection medium was replaced with a mix 1:1 of new complete culture medium and previously removed medium (conditioned medium) for neurons or only new medium for cell lines. Cells cultured on coverslips were fixed with a solution of paraformaldehyde 4%/sucrose 4% 72 h after transfection. Fixed neurons were blocked and permeabilized with blocking buffer containing donkey serum 10%/Triton X-100 0,2% in DPBS for 1 h and incubated 1 h in a similar solution (DPBS, 0,2% Triton X-100, 3% donkey serum) containing primary antibodies: Rabbit polyclonal anti-Tau Antibody (1/1000; Cat# 314003, Synaptic Systems), Mouse monoclonal anti-MAP2 antibody (1/200; Cat# M9942, Sigma-Aldrich), Mouse monoclonal anti-PSD95 (1/300; Cat# MA1-045, ThermoFisher Scientific). After several washes in DPBS, cells were incubated for 1 h with the following secondary antibodies diluted in same solution (DPBS, 0,2% Triton X-100, 3% donkey serum): Alexa Fluor 594 Donkey anti-mouse IgG (1/300; Cat# A-21203, ThermoFisher Scientific), Alexa Fluor 405 Goat anti-Rabbit IgG (1/300; Cat# A-31556, ThermoFisher Scientific). After 3 washes in DPBS, the stained neurons were mounted in ProLong Diamond Antifade reagent (ThermoFisher scientific).
The control human induced pluripotent stem cell (hiPSC) line used for this study was the ASE-9211 line (Applied Stem Cell, CliniSciences). The cells have been authenticated by the provider and routinely tested for mycoplasma contamination. For culture, plates were precoated with 1X matrigel (ThermoFisher Scientific) diluted in DMEM/F12 (ThermoFisher Scientific) and 1X antibiotic-antimycotic (ThermoFisher Scientific). Cells were cultured in supplemented mTeSR1 medium (StemCell) with a change of medium every day. On days of thawing or passage of hiPSC, the medium was supplemented with 10 µM Rock inhibitor (Y-27632, Sigma-Aldrich). The cells were then genetically modified with CRISPR-Cas9 system (ICV-iPS core facility, Paris Brain Institute, ICM): a total of 1×10 hiPSC were nucleofected with RNP complex (225pmol of each RNA (crRNA, crRNA-ATTO+), 120pmol of Cas9 protein and 2nmol HR-template. The crRNA sequence used is as follows: CGCCAGGGATCATGAGCTcGCGG. The positive clones were then validated by using PCR and Sanger sequencing, to confirm the insertion of the mutation (genotype ), and the heterozygous state of the mutation. The detection of CNVs was performed using ICS-digital method (Stem Genomics).
The hiPSC clones were dissociated into single cells and seeded at a density of 200,000 cells in a matrigel-coated 6-well plate. The following day, mTeSR1 was replaced by neuronal induction medium (Neurobasal Medium with 2% Neural Induction Supplement 50X, Thermo Fisher scientific). On day 7, differentiating cells were detached using accutase and resuspended in neural expansion medium (49% Neurobasal Medium, 49% Advanced DMEM/F12, 2% Neural Induction Supplement 50X, ThermoFisher scientific). Cells were maintained and expanded on pre-coated matrigel plates at a density of 5 × 10 cells per cm. Cells were passaged and treated overnight with 10 µM Rock inhibitor to promote attachment until passage 4, when a pure Neural Stem Cells (NSC) culture was obtained.
The NSC were labelled with MAP2 Antibody to specifically detect the neurites. The stained NSC were mounted in ProLong Diamond Antifade reagent with DAPI (Invitrogen). Neuronal proteins were extracted from NSCs and iPSC using N-Per lysis buffer (ThermoFisher Scientific) supplemented with 1% protease inhibitor (Halt™ Protease Inhibitor Cocktail 100X, EDTA-Free, Thermofisher Scientific).
Protein samples were processed for SDS-PAGE conditions at 110 V for 1 h. SDS-PAGE gels were transferred to a PVDF membrane using the TransBlot Turbo transfer system (Bio-Rad) for 3 min (2.5 A, 25 V). The membrane was blocked for 1 h with 5% milk diluted in TBS-Tween (Tris Buffer Saline, ThermoFisher Scientific with 1% Tween 20, Sigma-Aldrich). The primary antibodies were incubated overnight at 4 °C in TBST buffer with 5% milk. After 3 washes in TBST, the membranes were incubated with the corresponding horseradish peroxidase (HRP)-coupled secondary antibodies for 1 h at room temperature. For normalization of protein levels, we used actin (antibody: anti-βactin-peroxydase, 1/100000; Cat# A3854-200UL, Sigma-Aldrich) as a loading control. Membrane revelation by chemiluminescence was performed using the Clarity Western ECL kit (Bio-Rad) on the ChemiDoc Touch instrument (Bio-Rad). The following antibodies were used: Affinity-purified rabbit polyclonal anti-DPYSL5 (1/2000; kind gift from Pr Honnorat's lab [10]; Goat polyclonal antiserum anti-vGLUT1 (1/1000; Cat# 134307, Synaptic System); Monoclonal Mouse anti-PSD95 (1/100; Cat# MA1-045, ThermoFisher Scientific). The secondary antibodies were: Peroxidase conjugated anti-rabbit (1/2500; Cat# W401B, Promega), peroxidase conjugated anti-mouse (1/2500; Cat# W402B, Promega) and peroxidase conjugated anti-goat (1/2500; Cat# V8051, Promega).
The cellular imaging study was performed using a laser-scanning confocal microscope SP-8 (Leica) and the software Leica Application Suite X (LAS X). Sequential acquisitions were made, and high-resolution z stack images of cells were taken with the X63 or 20X objective with optical section separation (z interval) of 0.2 µm for the images of the dendritic spines (20X objective, z interval 0,3 µm for the complete neuron). The length of NSC neurites was measured on images acquired at 40X, z interval 0,3 µm. The cells included in the analysis had process <2 cell bodies in length [19]. The images were analyzed on maximum projections using with ImageJ Software. Three different portions of dendrites were taken within each neuron, and the number of dendritic spines was normalized to 10 µm of dendrite length. Mature (mushroom and stubby) and immature (thin and filopodia) spines were differentiated according to two criteria: (1) the presence or absence of PSD95 protein labeling, and (2) head diameter > 0.6 µm for mature (mushroom) spines or head diameter <0.6 µm for immature thin spines [20].
Statistical analysis was carried out using the GraphPad Prism 8.0 software (La Jolla, CA, USA). For the neuronal morphological study and the expression of variants, the data were analyzed using Kruskal Wallis Test followed by Dunn's multiple comparisons test. For the density of spines analysis, a Shapiro-Wilk normality tests determined data normality, the data were then analyzed using an ordinary one-way ANOVA followed by Tukey's multiple comparisons. To evaluate the ratio of mature and immature spines, Kruskal Wallis and Dunn's post-hoc tests were done. For the different experiments that compared WT/WT NSC and NSC, Mann Whitney tests were performed. Statistical significance was defined as P < 0.05.