Phenotypic diversity is caused by non-linear genetic interactions between two SNAREopathy genes

Phenotypic diversity is caused by non-linear genetic interactions between two SNAREopathy genes | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Phenotypic diversity is caused by non-linear genetic interactions between two SNAREopathy genes Matthijs Verhage, Jovana Kovacevic, Sébastien Houy, Johny Pires, and 7 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-6440830/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 05 Mar, 2026 Read the published version in Molecular Psychiatry → Version 1 posted 13 You are reading this latest preprint version Abstract Brain disorders caused by large effect mutations in single genes often present unexplained large phenotypic diversity, even among carriers of the same mutation. Here we examined genetic interactions as a possible explanation for this diversity for SNAREopathies, a group of neurodevelopmental disorders caused by de novo genetic variation in genes that together drive secretion of chemical signals in the brain. SNAREopathies are characterized by a striking phenotypic diversity, including different types/degrees or absence of seizures, developmental delay and intellectual disability. First, we present and test a theoretical framework predicting that large phenotypic diversity is caused by non-linear genetic interactions between two or more functionally related genes. Second, we test this prediction in validated SNAREopathy mouse models by analyzing phenotypic diversity at the synaptic, network, system and behavioral level in single versus double mutants for SNAREopathy genes Stxbp1 and Snap25 . Whereas single mutants all showed similar EEG- and motor abnormalities, but no overt seizures, as reported before, double mutants exhibited extreme diversity in seizure phenotypes. Some mice had lethal generalized seizures, frequent and complex epileptiform EEG activity and thalamic hyper-excitability as indicated by increased cFos staining, while other mice of the same genotype showed no detectable abnormalities, no increased cFos staining and a normal life span. The surviving double mutant mice showed phenotypes not more severe than single mutants at the synaptic, network, and behavioral level. Taken together, this study shows that haploinsufficiency at two interacting loci leads to extreme phenotypic diversity at the systems level, but not at the cellular level. These findings provide a proof of concept for how modifying genes in the patient genome may enhance phenotypic diversity. Biological sciences/Genetics Biological sciences/Neuroscience Developmental and epileptic encephalopathy SNARE-opathy epistasis epilepsy cognitive deficit Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 Introduction Developmental and epileptic encephalopathies (DEEs) are a heterogeneous group of severe childhood disorders characterized by different types of seizures and epileptiform activities, intellectual disability (ID) and severe neurodevelopmental delays ( 1 , 2 ). DEEs are characterized by large genetic and phenotypic diversity, with hundreds of genes implicated as causal, and each gene associated with multiple phenotypes ( 3 – 6 ). The genetic diversity implies the involvement of several cellular pathways in the etiology of DEEs. Sub-grouping genes to specific cellular pathways could be used as a strategy to resolve this heterogeneity and elucidate disease mechanisms more efficiently. This led to classification of channelopathies, synaptopathies, and recently also SNAREopathies ( 5 , 7 , 8 ). However, striking phenotypic diversity remains within such DEE sub-groups, even among patients with mutations in the same gene, suggesting that additional explanations are required. Several studies have suggested influences of interacting factors such as modifying genes, epigenetic factors, environmental factors, and stochastic processes ( 9 – 11 ). However, experimental tests of such suggestions in controlled laboratory conditions are scarce. SNAREopathies are caused by mutations in genes encoding the neuronal SNARE (soluble N-ethylmaleimide sensitive factor attachment protein receptor)-complex and its interactors, which together form the essential machinery for the secretion of chemical signals in the brain. Most SNAREopathies are caused by de novo missense or loss-of-function variants in one of the SNARE genes and several molecular disease mechanisms have been suggested including haploinsufficiency, dominant-negative, gain-of-function and recessive mechanisms ( 12 – 17 ). Despite the fact that SNARE-genes work together in a single, integrated molecular machine, a striking phenotypic diversity among SNAREopathy patients is evident ( 8 , 18 , 19 ). However, the incidence of pathogenic mutations in individual SNARE genes in the population is low, with an estimated incidence of one in 30,000 individuals for the most reported SNAREopathy to date, STXBP1 syndrome ( 20 ). This severely limits the power to identify modifying genes and epigenetic or environmental factors in population studies. In the present study we propose a theoretical framework to explain extreme phenotypic diversity for ‘monogenic’ disorders and test this model experimentally in controlled laboratory conditions using validated mouse models for SNAREopathies. We defined three modes of gene interactions that can be experimentally tested: common pathway, additive- and multiplicative- interaction, based on previous models ( 21 , 22 ) and selected two SNARE genes, Stxbp1 and Snap25 , expected to have genetic interactions and confirmed pathogenic potential for DEE. We established a panel of four mouse models for single and combined haploinsufficiency on the same genetic background to minimize other sources of genetic variation and tested the phenotypic consequences of gene interactions on the cellular level by assessing synaptic transmission in cultured neurons, on the network level by assessing spontaneous synchronous network activity (SSA) in acute brain slices and excessive neuronal activation by c-Fos staining in brain sections, and on the system level, using ECoG-video monitoring and cognitive assessments. We found robust support for multiplicative genetic interactions at the highest level of organization, but not at the lower (cellular) levels. Generalized seizures, complex epileptiform activities and brain hyper-excitability were observed in double mutants, but never in single mutants. Materials and methods Study design The study had multilevel design: in silico , in vitro , ex vivo and in vivo . The study followed the ARRIVE guidelines ( https://arriveguidelines.org/arrive-guidelines ). We generated, behaviorally and electrographically characterized Stxbp1 single, Snap25 single and double mutants and compared them to wildtype littermates (controls). We performed standardized battery of behavioral tests for assessment of different aspects of behavior. For behavioral analysis and ECoG recording, we used in total 79 mice. Experiments were performed in several batches and data were pooled. Ex vivo experiments were performed on the brain slices from mice. For Ca 2+ -imaging of the brain slices, 34 mice were used, and between 3–5 replicates per animal were analyzed. C-Fos immunohistochemistry was performed on brain slices from 21 mice. In vitro analysis of electrophysiological properties of neurons was performed in two batches of mice (n = 31 and 23 per batch). All experiments were performed by researcher unaware of the animal’s genotype. Subjects Control ( Stxbp1 +/+ Snap25 +/+ ), Snap25 single mutant ( Stxbp1 +/+ Snap25 +/− ), Stxbp1 single mutant ( Stxbp1 +/− Snap25 +/+ ) and double mutants ( Stxbp1 +/− Snap25 +/− ) mice were generated by mating male congenic C57BL6/J Stxbp1 +/− mice( 14 , 23 ) with female Snap25 +/− mice also bred on C57BL6/J genetic background, backcrossed for 20 generations( 24 ). All animals were kept in standard husbandry conditions on a 12 h light-dark cycle with food and water available ad libitum . Animals aged around 8 weeks were separately housed on sawdust in the standard Makrolon type II cages. All experiments were approved by the local animal research committee and complied with the European Council Directive(86/609/EEC). In total 200 mice were used. Behavioral phenotyping Behavioral experiments were performed on male mice between 8 and 16 weeks of age. In total three batches of Stxbp1Snap25 mice were used: batch 1 and 2 each consisted of 6 controls and 6 double mutants; batch 3 consisted of 8 controls, 8 Stxbp1 single-, 8 Snap25 single- and 8 double- mutant mice. In all three batches, we assessed spontaneous behavior in an automated home-cage environment (PhenoTyper model 3000, Noldus Information Technology, Wageningen, The Netherlands). Spontaneous behavior was automatically monitored for two and a half days in the PhenoTyper. Spontaneous home-cage behavior is highly dimensional aspect of behavior leading to separation of 90 separate parameters divided into 6 different categories( 25 ): kinematics, activity, sheltering, habituation, dark-light index and light-dark phase transition. Kinematic parameters describe specific elements of animal’s behavior related to movement characteristics, such as short and long movement segments and short and long arrest segments. Activity bouts describe mouse behavior on sub-minute scale and cumulatively during the period of days of spontaneous behavior monitoring. Sheltering behavior describes the tendency of mice to sleep inside the shelter( 25 , 26 ). Habituation ratio evaluates changes in activity over days by taking the ratio for respective parameters of dark phase 3 over dark phase 1. Dark-light indexes assess difference in the behavior during the dark and light phase of cycle, by taking the ratio for respective parameters of dark phase over the light phase. The effect of light/dark phase transition on spontaneous behavior was assessed by analysis of change in activity parameters during the periods surrounding phase transition. The observation of spontaneous behavior was followed by the assessment of discrimination- and reversal learning using the CognitionWall task in the same automated home-cage( 27 ). After testing in the PhenoTyper, animals acclimated to the new housing for one week before further testing. The batch 1 and batch 2 consisted of 12 control- and 12 double mutant- mice. We used standard behavioral battery for assessment of vision, muscle strength, motor coordination, anxiety, learning and memory, as previously described( 14 ). All tests were performed during the light phase with the least stressful tests done at the beginning, and at least 1 day apart. Briefly, the vision was assessed using the vision test; muscle strength was assessed measuring grip strength; motor coordination and motor learning were assessed on the rotarod. We used three anxiety-related paradigms: elevated plus maze test, open field and dark-light box. Fear-conditioning experiment evaluated associative learning and memory in mice. With Barnes maze test we tested spatial learning, memory and reversal learning. Short-term (working) memory was measured using T-maze spontaneous alteration task. Detailed protocols for these behavioral tests can be found in ( 14 ) and in Supplementary Material and Methods. The third batch of haploinsufficiency mutant mice was tested for spontaneous behavior in the Phenotyper followed by the fear conditioning and Barnes maze, since results from the first two tested batches of animals showed most phenotypical changes in these two paradigms. Detailed protocols for these behavioral tests can be found in( 14 ) and in Supplementary Material and Methods. Video monitoring, simultaneous radiotelemetric video, ECoG recordings Video monitoring of mice from batch 4, age: between 8 and 10 weeks old, have been performed in their home cages (2 control, 4 Snap25 single- , 6 Stxbp1 single-, and 6 double mutants). Mice were video monitored for at least 24 h. After video monitoring, seven mice from batch 4 and 13 mice from batches 1–3 were implanted with ECoG transmitters. In total: 4 control, 5 Stxbp1 single- , 4 Snap25 single-, and 7 double- mutants were implanted with ECoG transmitters (ETA-F10; specification: https://www.datasci.com/products/implantabletelemetry/mouse-(miniature)/eta-f10 ) as previously described in Kovačević et al. ( 14 ). Briefly, mice were anesthetized with isoflurane (3% isoflurane/oxygene, flow 0.8 l/min) and immobilized in the stereotaxic instrument. After administration of lidocaine (2%, s.c. ), a small incision was made on the skull allowing the recording electrode to be positioned above the motor cortex (2.2 mm anterior, 1 mm lateral) and ground electrode above the cerebellum (6 mm posterior, 1 mm lateral) using the stainless screws. The transmitter was placed in the abdominal subcutaneous pocket. The incision was closed with suture material. All animals received pre- and post-operative analgesic treatment with buprenorphine (0.05 mg/kg, s.c. ). Animals were daily checked during recovery period of at least 7 days. During ECoG recordings (for at least 24 h), implanted animals were placed on the DSI receiver board ( www.datasci.com ) in front of an infrared camera allowing synchronous recording of behavior and ECoG signals. Video and ECoG data were time matched at the beginning and at the end of recording. ECoG data were visually inspected and analyzed using the Event Classifier application within Neuroarchiver tool in LWDAQ software (Open Source Instruments, Inc.). The application classified the 1s-segments of EEG according to several metrics (asymmetry, intermittency, coherence, power, coastline and spikiness) enabling that similar patterns cluster together( 28 ).Video analysis was performed independently of ECoG data analysis and later matched with ECoG data. Sleep states were identified from the video recordings by researcher. ECoG signal that corresponds to the sleep state were selected and divided in one-minute epochs. Epochs were visually checked and those that contained artifacts were excluded from analysis. Spectral power analysis during the sleep episodes was performed using open-source software package Chronux ( http://chronux.org/ ) for signal processing of neural time-series data( 29 ). We used routine mtspectrumc which is applicable for analysis of continuous-valued data using the moving-window with 2s time-width and 1s step-size. Data were sampled at 1kHz and high-pass filtered at 1Hz and low-pass filtered at 200Hz. Average spectral power was calculated as a mean value of relative power (expressed as the ratio of total power) for all 1-min episodes per animal. Power per frequency band was calculated as the average relative power within the frequency ranges: 1-4Hz for delta band, 5-8Hz for theta band, 9–12 Hz for alfa band. Spectrogram were made using our own written script in Matlab. c-Fos staining Mice, around 3 months old, were sacrificed either by cervical dislocation or overdose of avertin (2,2,2-tribromoethyl alcohol). Animals sacrificed by avertin were perfused with 4% paraformaldehyde (PFA) in 0.1M phosphate buffered saline (PBS.) In the present study, 5 Snap25 single mutants, 10 double mutants and 6 controls were used for c-Fos analyses. In addition, we re-analyzed data from 5 Stxbp1 single mutants obtained in our laboratory and compared these to the controls from the same experiment. All brains were removed and post-fixed in 4% PFA in 0.1M PBS. After overnight cryoprotection in 30% sucrose solution in 0.1M PBS, brains were blocked in the coronal plane, frozen on dry ice and sectioned at 50 µm on a cryostat. To reveal c-Fos expression levels in the brain, free-floating sections were incubated in 0.3% H2O2 in 0.1M PBS for 10 min. After three rinses in 0.1M PBS, sections were incubated in 0.1M PBS containing 5% normal goat serum, 0.25% TritonX-100 and a c-Fos antibody raised in rabbit (Santa Cruz, sc-52; 1:800/1:500) or c-Fos antibody raised in rat (SySy, 226017; 1:1000) and left for overnight incubation (up to 96 hours) at 4°C. Sections were washed with 1xPBS and incubated at RT for 1 h in biotinylated goat anti-rabbit (#65-6140, Invitrogen; 1:400) and anti-rat-Ab secondary antibodies (# 31830, ThermoFisher Scientific, 1:400). Sections were washed three times with 0.1M PBS and incubated at room temperature (RT) for one and a half of hour in biotinylated goat anti-rabbit (#65-6140, Invitrogen; 1:400) or anti-rat-Ab secondary antibodies (# 31830, ThermoFisher Scientific, 1:400). After three rinses with 0.1M PBS, the sections were incubated at RT for 1,5 h in avidin-biotin peroxidase complex (Vectastain ABC, Vector Laboratories; 1:800). To visualize the peroxidase labeling, sections were processed with a DAB/nickel substrate working solution (DAB Peroxidase Substrate, SK-4100; Vector Laboratories) for 7 min at RT. After rinsing with 0.1M PBS, sections were mounted on gelatin-coated slides, dehydrated, and put on coverslips.Sections were imaged using a Leica bright-field microscope at 5x and 10x magnification. Several brain regions were selected for analysis: prefrontal cortex (PFC), primary motor cortex and somatosensory cortex, hippocampus (CA1, CA3, and dentate gyrus (DG)) and thalamus (Thal). Each brain region of interest was identified using a standard mouse brain atlas (Paxinos and Franklin). c-Fos immunoreactive nuclei were counted using predefined threshold values in ImageJ software. Labelled cells were counted bilaterally, averaged and normalized to the size of area and expressed as a relative compared to appropriate control. Calcium imaging in developing brain slices Calcium imaging of spontaneous synchronous activity (SSA) in the developing PFC- brain slices was performed according to previously published protocol by Dawitz et al. ( 30 ) and Pires et al. ( 31 ). Briefly this method contains several steps: slice preparation, dye-loading slices, imaging and analysis. In total for the calcium imaging experiment, 9 controls, 8 Stxbp1 single-, 9 Snap25 single- and 8 double- mutants were used. Animals (P14 ± 1) were decapitated and brains were rapidly removed and placed in cold oxygenated aCSF (artificial Cerebrospinal Fluid)( 30 , 31 ). PFC- brain slices were cut (300 µm thick) using Microm HM 650V and transferred into a slice holder containing oxygenated recovery-aCSF (rACSF)( 30 ). After 1h of recovery period, slices were transferred to a staining-chamber filled with 1 ml experimental ACSF (eACSF) and heated to 34°C. Slices were incubated in eACSF containing Fura2-AM dye (ThermoFisher scientific, F1201) between 20 and 45 minutes depending on the age of the animal. After incubation, slices were transferred to coated recording chambers and approximately 1 ml of eACSF was added. Recording chambers with slices were kept in a large humidified interface chamber to recover for at least one hour before recording. Network imaging was performed on a two-photon laser-scanning microscope (Trimscope LaVision Biotec). Slices were heated to 37°C and constantly perfused with oxygenated eACSF. Using a Hamamatsu C9100 EM-CCD camera as a detector, two time-lapse movies (4000 frames each) in PFC-ROIs were acquired with a sampling frequency of 7.65Hz (450µmX525µm, binning 2x2). The first was to evaluate baseline network activity and the second the effect of incubation with Gabazine (10 µM, Hellobio, SR95531) for 10 minutes. For the detection of the somas, a z-stack ± 20µm around the central plane with step size of 1µm thickness was acquired after each recording. To analyze calcium-imaging data, custom-built Matlab® (Mathworks) scripts were used (EvA methodology)( 32 ). Neurons were semi-automatically detected using the z-stacks of the network being imaged. Finally, the events in individual traces were then analyzed individually and on a network level and three different categories of neurons were derived: silent neurons (those neurons where no activity was detected), active neurons and synchronized neurons (active neurons whose activity is synchronized with other neurons in network)( 32 ). Several different parameters were extracted and statistically analyzed: percentage of active cells, frequency of active cells, percentage of synchronously active cells and the frequency of synchronously active cells. Cell culture and electrophysiology Hippocampal microisland cultures were prepared from controls, Stxbp1 single-, Snap25 single-and double mutants at embryonic day 18. Briefly, hippocampi were isolated, collected in ice-cold Hank’s balanced salt solution (HBSS; Sigma, H9394) supplemented with 10mM HEPES and digested with 0.25% trypsin (Life Technologies, 15090-046) at 37°C for 20 min. After trituration, cells were resuspended in Neurobasal medium supplemented with B-27, 1% HEPES, 0.25% GlutaMAX, and 0.1% Penicillin-Streptomycin (all Invitrogen) and plated in a 6-well plate at a density of 7000 cells/well on top of pre-grown mouse glia islands on 30-mm coverslips (first experiment, see below) or in a 12-well plate at a density of 2500 cells/well on top of pre-grown rat glia islands on 18mm coverslips (second experiment) to achieve autaptic cultures. Two independent electrophysiological experiments were performed. In the first experiment, electrophysiological measurements were performed in autaptic cultures grown for 10–14 days in vitro as previously described in Ruiter et al. ( 33 ). The recordings were performed at room temperature (RT) using an EPC10 amplifier (HEKA) with the recording program Patchmaster v2x73.5 (HEKA). Traces were filtered with a 3kHz Bessel low-pass filter and data were acquired at 20 kHz. The series resistance was compensated to 70%. Borosilicate glass pipettes (resistance between 3 to 5 MΩ) were filled with 136 mM KCl, 17.8 mM HEPES, 1 mM EGTA, 4 mM ATP-Na, 4.6 mM MgCl2, 15 mM Creatine Phosphate and 50 U/mL phosphocreatine kinase (pH 7.4). External solution contained the following (in mM): 10 HEPES, 10 glucose, 140 NaCl, 2.4 KCl, 2 MgCl 2 and 2 CaCl 2 (pH = 7.4, 300 mOsmol). Recordings were done in voltage clamp, with the holding potential kept at -70 mV. Evoked excitatory postsynaptic currents (EPSC) were induced by raising the holding voltage to 0 mV for 2 ms. The size of the readily releasable pool (RRP) was assessed by hypertonic sucrose application (500 mM, 3.5 s). Sucrose was dissolved into the external solution. Application to the cells was done using a custom-made barrel system, controlled by SF-77B perfusion fast step (Warner Instruments) via digital output switches from the EPC10. In the second experiment, electrophysiological measurements were performed in autaptic cultures grown for 13–16 days in vitro as explained in Meijer et al. ( 34 ). Whole-cell voltage clamp recordings (Vm = -70 mV) were performed at RT using an Axopatch 200B amplifier, and Digidata 1440A and Clampex 10 for signal acquisition (Molecular Devices). Action potentials were induced by stepping to + 30 mV for 1ms. Data were acquired at 10 kHz with 2 kHz filtering with a low-pass Bessel filter. Series resistance was compensated by 70–80% (bandwidth 7.52 Hz). Borosilicate glass pipettes (resistance of 2-4.5 MΩ) were filled with (in mM): 136 KCl, 17.8 HEPES, 1 EGTA, 0.6 MgCl 2 x6H 2 O, 4 ATP-Mg, 0.3 GTP-Na, 12 phosphocreatine di-potassium-salt and 50 U/mL phosphocreatine kinase. External solution contained the following (in mM): 10 HEPES, 10 glucose, 140 NaCl, 2.4 KCl, 4 MgCl 2 , and 2 CaCl 2 (pH = 7.30, 300 mOsmol). Traces were excluded when series resistance was higher than 15 MΩ or leak current was larger than 300pA. GABAergic traces were recognized by prolonged EPSC decays and excluded. Analysis was performed in MATLAB R2019a (Mathworks) using available scripts (viewEPSC, downloaded from Github user vhuson). The RRP was estimated by back extrapolation from the cumulative EPSC charge during a 40 Hz 100 action potential train as in Meijer et al. ( 35 ). Immunocytochemistry Neurons were fixed on DIV10-DIV14. Coverslips were fixed at RT with 2% PFA diluted in culture medium (Neurobasal with 2% ml B-27, 1 M HEPES, 0.26% Glutamax, 14.3 mM β-mercaptoethanol, 20 U/ml penicillin, 20 µg/ml streptomycin) for 10 min followed by another fixation with 4% PFA diluted in PBS for 10min. After washing with PBS, the neurons were permeabilized for 5 min at RT with a solution of 0.5% Triton X-100 diluted in PBS. The neurons were then blocked with a solution containing 4% goat serum and 0.1% Triton X-100 for 30 min at RT. The coverslips were then incubated for 2 hours at RT with a blocking solution containing α-VGLUT antibody (1/1000; Guinea pig, Millipore AB5905) and α-MAP2 antibody (1/5000; chicken, Abcam Ab5392) or a solution containing α-VGAT antibody (1/500; Rabbit, Synaptic System 131002) and α-MAP2 antibody (1/200; chicken, Abcam Ab5392). Cells were washed with PBS, and then transferred for 1 hour at RT to a blocking solution containing the secondary antibodies Alexa Fluor-568 α-Chicken (1/1000; goat, Invitrogen A11041) and Alexa fluor-647 α-Guinea pig (1/4000; goat, Invitrogen A21450) or a solution containing the secondary antibodies Alexa Fluor-568 α-Chicken (1/1000; goat, Invitrogen A11041) and Alexa fluor-488 α-Rabbit (1/5000; goat, Invitrogen A11008). Cells were washed with PBS before being mounted on glass slides with FluorSave (Calbiochem, 345789). Coverslips were examined on Leica SP8 inverted confocal microscope with 40x oil immersion Objective (NA = 1.4). Survival rate assessment The assessment of survival rate per genotype was performed in twelve E18-litters bred for electrophysiological experiments (caesarian section) and ten nests bred for behavioral experiments (natural birth). All pups from E18-nests were collected and genotyped before culturing and the results analyzed for the expected genotype ratio. In addition, genotyping results of ten nests bred for behavioral experiments were performed after weaning, at 3–4 weeks of age and the survival rate of animals was compared to those at E18. Following behavioral and/or video monitoring experiments, genotypes were conformed at the time of sacrifice. Statistics All statistical analyses were performed using IBM SPSS statistic 24 (IBM corporation, Armonk, NY, USA) and GraphPad Prism®, version 8.3.1. If the normality and homoscedasticity criteria were met, data were analyzed using parametric tests (t-test, ANOVA, repeated measure ANOVA). If the normality and homoscedasticity criteria were not met, nonparametric tests were performed (Mann Whitney U-test) or data were ln-transformed. Preliminary analysis of electrophysiological data was performed to assess the variability between different preparation (weeks of experiment) or different batches by applying nested- and two-way- ANOVA. The analysis revealed high variability between preparations (weeks of experiment). Thus, to reduce the variability between preparations, data obtained from electrophysiological experiment were normalized to the average of control group for the appropriate week. If normalized data were not normally distributed, ln-transformation was applied. There were no significant differences of behavioral measures between batches. Therefore, data from different batches were analyzed together. Outliers were removed from analysis using ROUT method, with Q set to 1%. Tukey’s posthoc tests were performed if ANOVA showed significant effect. The genotype effect in CognitionWall DL/RL task was assessed by performing log-rank test on two Kaplan Meier survival curves. An error probability lower than p < 0.05 was accepted as statistically significant through the study. For all given level of analysis of PhenoTyper spontaneous behavior data, statistical analysis was based on estimated false discovery rate (FDR), P-values were corrected by minimum positive FDR with a threshold set at 5%. Data availability Data are available: http://link.will.be.added (SciStore server @VU University) Results Non-linear gene-gene interactions can explain phenotypic diversity for DEEs We defined three modes of gene-gene interaction to describe how phenotypic differences observed among individuals with a pathogenic variant in one gene depends on their genotypes at other loci. First, we assume that if two genes encode proteins that operate in the same pathway, a variant/mutant has a unique phenotypic spectrum characterized by a mean effect and a variance determined by environmental factors as well as background genetic diversity (Fig. 1 A). If both variants affect the same pathway and are expressed in the same individual this would result in a phenotypic spectrum of the strongest single case, but not more severe or diverse than that (lack of additivity, Fig. 1 , model B.1). In contrast, if two genetic variants from independent pathways lead to common disorders, the combination of variants results in summation of the phenotypic spectra (additive model, Fig. 1 , model B.2.1 ( 36 )). Finally, in cases where genes interact genetically or physically, genetic variation in two such genes may give rise to a multiplicative effect on the phenotypic spectra (‘epistasis’ or ‘super-additivity’ Fig. 1 , model B.2.2, ( 37 )). The last two scenarios can be distinguished by the different distributions and variances of their effects (phenotypes): if the effects of single gene variants follow a normal distribution (Fig. 1 A), the additive model predicts that the effect of the combination of variants also follows a normal distribution, with a mean being the sum of the means of single variants and the variance being the sum of the variances of single variants. However, according to the multiplicative model, the distribution of the effect of the combination of two normally distributed variants is right-skewed, it approaches a lognormal distribution and has markedly larger variance compared to the additive model (Fig. 1 ). As a result of the increased variance, multiplicative interaction results both in more individuals with severe phenotypes, but also more individuals with mild phenotypes, compared to additive interactions (Fig. 1 ). Notably, this conclusion follows from simply combining the background variances (Fig. 1 A) in a multiplicative model, without introducing new sources of variance. Whether genetic variants act in the same pathogenic pathway, or via independent or interacting pathways can be experimentally tested in animal models bearing variants in two genes. The models in Fig. 1 predict that the large phenotypic diversity in DEE-patients can be explained by multiplicative interactions between genes. Stxbp1/Snap25 double mutants show extremely diverse seizure phenotypes Two presynaptic DEE genes, STXBP1 and SNAP25 , were selected to test the consequences of gene-gene interactions, given their close functional relationship in SNARE-complex assembly. We generated three haploinsufficiency mouse models by heterozygous inactivation of 1) single Stxbp1 (‘ Stxbp1 single mutants’), 2) Snap25 (‘ Snap25 single mutants’) gene and 3) combined haploinsufficient inactivation of Stxbp1 and Snap25 (‘double mutants’) mice ( 23 , 24 ). Array-based genetic analysis of 11,000 SNP probes (miniMUGA, ( 38 ) confirmed a homogeneous genomic background for all experimental groups (96.8% of SNPs consistent with C57BL/6J sub-strain, Supplementary Table 3 and Supplementary Fig. 1), with small contributions of 0.1% and 0.4% of flanking regions of 129 strain background around the deletion site, for the Stxbp1 and Snap25 locus, respectively, as reported before for Stxbp1 single mutants.( 14 ). In E18 embryos, obtained by caesarian section for synapse physiology experiments (see below), all four genotypes were obtained, with a trend towards a reduced fraction for double mutants (18% vs expected Mendelian ratio of 25%, p = 0.093, χ 2 , labelled A in Fig. 2 A). At the age of three/four weeks, when naturally born animals bred for behavioral experiments were weaned and genotyped, the distribution of genotypes significantly deviated from expected, with double mutants only 12% of animals, approximately half of the expected number (p = 0.041, χ 2 , Fig. 2 A). Concomitantly, animal care takers reported lethal generalized seizures prior to and during weaning in these nests (Suppl video 1). Following the three/four weeks genotyping, sudden unexpected death of epilepsy (SUDEP) was detected in two additional double mutants older than 6 weeks. Together, this led to an even more significant deviation of the expected number of double mutants (10% instead of 25%, p = 0.031, labelled B in Fig. 2 A). Further breeding of these double mutants was restricted by the local ethical committee regulations. Strikingly, surviving double mutants had no reported seizures, normal vision, normal muscle strength, normal motor coordination and ability to acquire motor skills, but 16% lower body weight compared to their controls (Supplementary Fig. 2 and Supplementary Table 1). These data suggest that at least half of double mutants died before the age of eight weeks due to severe seizures, while the surviving double mutants had no apparent phenotypes, except a lower body weight. Video observations of mice in their home cage confirmed frequent generalized and clonic seizures in some double mutant mice, but not others (Fig. 2 B). The incidence of seizures varied substantially among double mutants (compare the two double mutant individuals in Fig. 2 B). Stxbp1 single mutants and double mutants showed twitches (n = 20.2 ± 2.6 / 12h) and jumps (n = 7.4 ± 1.3 / 12h, Fig. 2 C), as reported earlier for Stxbp1 single mutants ( 13 , 14 ). The incidence of these two types of behaviors was lower in double mutants (n = 4.5 ± 1.9 and n = 1.0 + 0.4 for twitches and jumps per 12 h, respectively, Fig. 2 C). Except for these abnormalities, overall development of single Stxbp1 and Snap25 mice was normal, as reported before ( 14 , 39 – 41 ). Taken together, the reduced incidence of twitches and jumps and the occurrence of clonic, lethal seizures in a subset of double mutant mice suggests large phenotypic diversity regarding behavioral manifestation of epilepsy in double mutant mice. Next, we combined video with electrocorticography (ECoG) monitoring at six to twelve weeks of age using implanted electrodes in 19 mice from new litters. These recordings confirmed generalized seizures in a subset of double mutant mice, but not others (Fig. 2 D-E). The generalized seizures began with clusters of spike-slow waves (interictal spikes) accompanied with clonic seizures; the seizure progressed to a full generalized seizure and stopped with the postictal suppression and behavioral immobility (Fig. 2 E). Strikingly, clonic and generalized seizures were observed only in a subset of double mutant mice, while other double mutant mice showed no generalized seizures during 24 hours of recording (Fig. 2 D). Hence, double mutant mice show extreme phenotype diversity in behavioral and electrographic abnormalities related to seizures. Analysis of ECoG recordings revealed a collection of diverse patterns of epileptiform activity in several experimental groups: slow-wave discharges (SWDs), sharp spikes and spike-slow waves (Fig. 2 G-J). SWDs were observed before in Stxbp1 single mutants ( 14 ) and typically accompanied by behavioral twitches ( 14 ) (Fig. 2 B). The number of SWDs was higher in Stxbp1 single mutants compared to control, Snap25 single mutants and double mutants (F( 3 , 15 ) = 4.501, p = 0.019, post hoc : p = 0.003, p = 0.015 and p = 0.039, respectively, Fig. 2 H). The lower number of SWDs in double mutant mice compared to single Stxbp1 mice confirms the previous conclusion that a higher phenotypic diversity exists for epilepsy-related phenotypes in double mutant mice compared to single mutant and control mice. The number of spike slow waves was significantly higher in double mutant mice compared to control mice, Snap25 single and Stxbp1 single mutants (F( 3 , 15 ) = 4.913, p = 0.014, post hoc : p = 0.004, p = 0.020 and p = 0.026, respectively, Fig. 2 I). When the spike-slow waves occurred during the awake state, they were accompanied by clonic seizures in double mutant mice. Sharp spikes were observed mainly during sleep and the number of sharp spikes did not differ between genotypes (Fig. 2 J). Hence, more severe electrographic seizures were observed in double mutant mice compared to Stxbp1 single- and Snap25 single- mutants. Taken together, the diversity in epileptic phenotype observed in double mutant mice is characterized by higher probability of low incidence SWDs but also with exacerbated seizure activity represented with increased lethality, generalized seizures and increased number of spike-slow wave epileptiform discharges in a subset of double mutants, while others showed mild/no phenotypes. According to our proposed model of gene interaction (Fig. 1 ), the observed phenotypic diversity of double mutant mice is in line with right-skewed distribution and larger variance predicted by the multiplicative model of gene interactions (Fig. 2 F). Finally, power spectral analysis during the sleep episodes (Fig. 2 K) revealed a shift in relative power towards higher frequency bands for all three genotypes, represented as a decrease in the relative power in the delta band (F( 3 , 4 ) = 3.846, p = 0.113) and an increase in the relative power in the alfa band (F( 3 , 4 ) = 26.72, p = 0.0042) for all three mutant groups, independent on the genotype (Fig. 2 K - M). According to our proposed model (Fig. 1 ), this finding suggests that a common pathway underlies shifts in spectral power in single Stxbp1- , single Snap25- and double mutant mice during sleep. c-Fos expression is increased in the thalamus of surviving double mutants To corroborate the observed seizure activity, c-Fos immunoreactivity was used as a marker for excessive neuronal activity in the brain ( 42 ), Fig. 3 . Increased c-Fos expression was observed in cortical brain regions of Stxbp1 single mutants as shown before ( 14 ), in the surviving double mutants (p = 0.026) and a strong trend was observed in Snap25 single mutants (p = 0.058) (Fig. 3 A-B and 3 E). Furthermore, only in double mutants, c-Fos expression was significantly increased in the thalamus (Fig. 3 C and 3 E; p = 0.026) and a strong trend was observed in hippocampal regions (Fig. 3 D and 3 E; p = 0.058). Thus, c-Fos expression data confirms genetic interaction between the two genotypes in thalamus and hippocampus of the double mutants, but not in cortical regions, possibly because double mutants with most excessive cortical excitability had already died (see above) prior to c-Fos expression analysis. Surviving double mutants show impaired cognition, like Stxbp1 mice Different aspects of learning and memory were assessed in surviving animals of the four experimental groups. To assess associative learning and memory, the fear conditioning test was performed (Fig. 4 A). After one pairing session between shock and tone, a contextual memory was assessed by placing animals in the training context during the next day (Fig. 4 B). Snap25 single mutants showed a similar percentage of freezing as controls (p = 0.639). Stxbp1 single mutants showed a significantly lower percentage of freezing in the training context compared to controls (p = 0.001, Fig. 4 C), as shown before 13 and compared to Snap25 single mutants (p = 0.010). Double mutants showed similar effects as Stxbp1 single mutants (p < 0.001 and p = 0.007, respectively, Fig. 4 B-C). Supplementary Table 1 lists all statistical tests. Exposure of mice to the new context resulted in ~ 10% of time freezing in control mice and Snap25 single mutants due to general fear (Fig. 4 B). Stxbp1 single mutants and double mutants showed significantly lower percentage of freezing compared to controls (p = 0.014 and p = 0.020, respectively, Fig. 4 C and Supplementary Table 1) and Snap25 single mutants (p = 0.048 and p = 0.060, respectively, Fig. 4 C). Cued memory was assessed by tone exposure of animals in a new context Fig. 4 B. Stxbp1 single mutants and double mutants showed a significantly lower percentage of freezing compared to controls (p = 0.009 and p = 0.013 Fig. 4 C and Supplementary Table 1) and Snap25 single mutants (p = 0.044 and p = 0.104, respectively, Fig. 4 C). Taken together, fear conditioning experiments suggest a strong and similar impairment in the contextual and cued fear memory in Stxbp1 single mutants and surviving double mutants. No evidence was observed for stronger phenotypes in surviving double mutant mice than for single mutants. The Barnes maze test was used to assess spatial learning, memory and reversal learning (Fig. 4 D). During the acquisition phase, mice were trained to locate the escape hole and the time needed to escape the aversive environment was assessed. No significant effects of genotypes on the escape latency were observed, although latencies tended to be higher for Stxbp1 single mutants and double mutants compared to controls (F(3,208) = 2.613, p = 0.061, Fig. 4 E and Supplementary Table 1). During the probe trial, to assess spatial memory one day after the last acquisition training, Stxbp1 single- and double- mutants showed a higher probability of hole visits in the target octant compared to controls and Snap25 single mutants (p = 0.042 and p = 0.033, respectively, Fig. 4 F and Supplementary Table 1). Impaired behavioral flexibility was already described in Stxbp1 single mutants( 14 ) and a similar phenotype was observed in surviving double mutant mice. During the reversal phase, the location of the escape hole was changed to the opposite side of the maze. Similar to the trend observed during the acquisition phase, double mutants needed more time to find the new escape hole compared to controls and Snap25 single mutants (p = 0.036 and 0.011, respectively Fig. 4 G) and Stxbp1 single mutants showed a trend towards longer escape latencies compared to controls and Snap25 single mutants (p = 0.010 and 0.057, respectively Fig. 4 G). These data show that single Stxbp1 and surviving double mutants preserved the learned response stronger than single Snap25 mutants and control mice, suggesting impaired behavioral flexibility and deficits in the reversal learning in single Stxbp1 and double mutants. Again, no evidence was observed for stronger phenotypes in surviving double mutant mice compared to single mutants. Attention and working memory were assessed using the spontaneous alteration task in the T maze (Supplementary Fig. 3A). This test is based on the natural tendency of mice to visit the previously not visited arm ( 43 ). Double mutants showed a similar percentage of alterations as their controls (t( 22 ) = -0.963, p = 0.346, Supplementary Fig. 3B and Supplementary Table 1), suggesting normal attention and hippocampal-dependent short-term memory. To assess discrimination and reversal learning, we performed a 4-day automated home-cage task, the CognitionWall test ( 27 ) Supplementary Fig. 3C and Supplementary Table 1. During the discrimination-learning phase (DL), animals should learn to earn food rewards by passing through the correct hole of the three holed CognitionWall placed inside the PhenoTyper. During the reversal-learning (RL) task animals should suppress previously learned response and learn that passing through the other hole is rewarded. During the discrimination and reversal learning tasks, all mice showed similar distribution of entries made to reach the criterion of 80% correct entries (p = 0.264 and p = 0.766, Supplementary Fig. 3D-G and Supplementary Table 1). Thus, the results from the Cognition Wall test suggest a normal discrimination and reversal learning in all four experimental groups. Surviving double mutants show anxiety-related behaviors like Stxbp1 mice The anxiety-related phenotypes in double mutant mice were tested using a classical anxiety-related paradigm, the elevated plus maze test (EPM). In this test, double mutants showed anxiety-related behaviors represented by significantly less time spent on the open arms (t( 22 ) = 2.561, p = 0.018) and a lower percentage of visits to the open arms (t( 22 ) = 3.216, p = 0.004) accompanied with a mild, but significant increase of total distance moved compared to their controls (t( 22 ) = -2.813, p = 0.040), (Supplementary Fig. 4A-C and Supplementary Table 1). Double mutants did not show increased anxiety in the open field test and in the dark-light box test (Supplementary Fig. 4D - I and Supplementary Table 1). Taken together, these data show mild anxiety-related behavior accompanied with hyper-activity detected in the elevated plus maze test, comparable to Stxbp1 single mutant mice. Surviving double mutants show spontaneous behaviors like Stxbp1 mice Analysis of spontaneous behavior was performed with surviving animals in all experimental groups in the automated home-cage environment (PhenoTyper) enriched with a shelter ( 25 ). Double mutants showed several behaviors that were altered to a similar extent as the single Stxbp1 mutants, especially in Kinematics (parameters 1–26), but also a few abnormalities in other spontaneous behaviors (Fig. 5 A-F, Supplementary Table 2). Snap25 single mutants showed very few significantly altered (kinematic) behaviors (Fig. 5 A). Taken together, spontaneous behavior analysis indicates that surviving double mutants showed no evidence for super-additivity, but only phenotypes like the strongest single mutant ( Stxbp1 ). GABA inhibition in developing cortex is abnormal in double mutants, like single Stxbp1 mice Spontaneous Synchronous Activity (SSA) is essential for the correct development of neural circuits( 44 ). SSA has been recently characterized in the mPFC and it has been shown that GABA blockade at the end of the second postnatal week can partially restore the SSA, which is present at earlier times ( 31 ). We measured SSA in the developing prefrontal cortex of all experimental groups at 2 weeks of age, by monitoring Ca 2+ -transients using two-photon calcium imaging in acute brain slices. The role of GABA in SSA was assessed by adding gabazine at concentrations that block both the phasic and tonic activity of GABA A receptors (Fig. 6 A-B). On average, approximately 40% of all neurons had spontaneous Ca 2+ -transients with an overall frequency of 0.0063 Hz in all groups, independent of the genotype (percentage of active cells: F(3,90) = 1.630, p = 0.188 and frequency of active cells: F(3,90) = 1.577, p = 0.200, Supplementary Fig. 5A-B). The majority of active cells was synchronously active; the percentage of synchronously active cells and the frequency of their activity did not differ between genotypes during baseline recording (F(3,90) = 0.600, p = 0.617 and F(3,90) = 2.058, p = 0.111, respectively, Supplementary Fig. 5A-B). Blockade of GABA A receptors by gabazine did not significantly affect the percentage of active cells in any of the experimental groups (F(1,90) = 1.614, p = 0.207, Supplementary Fig. 5A), but did show an overall trend towards increased frequency of active cells (F(1,90) = 3.463, p = 0.067, Supplementary Fig. 5B and Supplementary Table 1). Gabazine affected the percentage of SSA-participating cells and the frequency of SSA: a significant increase of the percentage of SSA was observed for control and Snap25 single slices (p = 0.045 and p = 0.014, respectively Fig. 6 C), but not for Stxbp1 single and double mutant slices (p = 0.145 and p = 0.827 Fig. 6 D). The frequency of SSA in control slices was significantly increased after application of gabazine (p = 0.022, Fig. 6 D). On the other hand, the application of gabazine did not affect the frequency of SSA in Stxbp1 single and double mutant slices. Thus, the frequency of SSA in brain slice from Stxbp1 single mutants was significantly lower than in brain slices from control and Snap25 single mutants after application of gabazine (p = 0.016 and p = 0.018 Fig. 6 D). These data show that GABA inhibits SSA in brain slices from control and Snap25 single mutants but that this inhibitory effect of GABA was absent in brain slices from Stxbp1 single- and double- mutants. Double mutants show reduced synaptic transmission like single Stxbp1 neurons To assess effects of Stxbp1 and Snap25 gene interaction on basic synaptic function, we performed two independent electrophysiological experiments in single (autaptic) hippocampal neurons in culture. Experiments were performed over 5–9 independent experimental weeks and data were normalized to the average of control for every week. In the first series of experiments a significant effect of genotype on the amplitude of evoked response in glutamatergic neurons (F(3, 240) = 5.523, p = 0.0011, Fig. 7 A-B) and the frequency of spontaneous release (F(3, 235) = 3.942, p = 0.0090, Supplementary Fig. 6A-B) was observed. A posthoc test revealed a decrease of EPSC amplitude in single Stxbp1 and double mutant neurons (p = 0.007, p = 0.002, respectively, Fig. 7 B). Similarly, a trend of decreased mEPSC frequency in single Stxbp1 and a significant decrease in double mutant neurons were observed (p = 0.063 and p = 0.008, respectively) but no effect on mEPSC amplitude (F(3, 237) = 3.150, p = 0.0257, Supplementary Fig. 6A-C). Snap25 single mutant neurons showed no significant difference on any of the synaptic parameters tested (Fig. 7 A-D). Taken together, these data indicate that glutamatergic transmission is normal in Snap25 single mutant synapses, while equally affected in single Stxbp1 and double mutant synapses. The decreased glutamatergic transmission can be caused by a decreased number of glutamatergic synapses, decreased synaptic efficacy (release probability) and/or by decreased size of the readily releasable pool (RRP). To discriminate between these possibilities, we tested in vitro synapse formation. Morphological analysis revealed no consistent differences in the dendritic length, synaptic size and synaptic density between single Stxbp1 and double mutant VGLUT(+) neurons compared to control neurons (Supplementary Fig. 7 and Supplementary Table 1). The size of RRP was assessed by application of hypertonic sucrose (Fig. 7 C-D) and was significantly decreased in single Stxbp1 and double mutant neurons (p < 0.001 and p = 0.002, respectively, Fig. 7 D). Taken together, these data suggest that the decreased glutamate release in single Stxbp1 and double mutant neurons is mainly due to a decreased size of the RRP. Synaptic transmission was also assessed in Stxbp1 single, Snap25 single and double mutant GABA-ergic neurons. (Fig. 7 E). No significant effect was observed for evoked IPSC, RRP size and mIPSC frequency and amplitude (Fig. 7 F-H, Supplementary Fig. 6F-J and Supplementary Table 1), indicating impaired excitatory synaptic transmission, without affecting inhibitory synaptic function. These findings are consistent with data obtained from slice recordings in the somatosensory cortex of the same mouse line ( 45 ). The second experiment was performed in a different laboratory, on the mouse lines used for the system- and behavioral analyses (Fig. 2 – 7 ), (Fig. 7 I-L). In the second series of experiments, we examined evoked synaptic response after single action potential (EPSC), paired pulse ratio after 20, 50 and 100 ms intervals and synaptic run-down and cumulative discharge after 100 pulses at 40 Hz train. The EPSC amplitudes did not differ between groups (Fig. 7 I) and paired pulse ratios showed paired-pulse facilitation without significant differences between groups (Fig. 7 J, Supplementary Table 1). The cumulative charge released after 100 pulses at 40 Hz showed non-significant trend toward increased value in single Snap25 neurons compared to control neurons (> 60% increase compared to control), while it did not differ in single Stxbp1 neurons and double mutant neurons compared to control neurons (Fig. 7 K, Supplementary Table 1). The RRP was estimated from the cumulative charge at 40Hz, 100 pulses by extrapolation of the last 20 pulses (Fig. 7 K) and showed no significant differences between groups (Supplementary Table 1). Taken together, the results from the second experimental series suggested no overall deficits in basal synaptic transmission in cultured autaptic neurons of single and double mutant mice. To understand the cause of differences found for the electrophysiological phenotypes in two laboratories, we analyzed the samples from single Stxbp1 and single Snap25 mice in both laboratories using miniMUGA ( 38 ). MiniMUGA analysis revealed the presence of 129-strain SNPs in all samples flanking the deletion sites for the two genes (Supplementary Table 3 and Supplementary Fig. 1). The dominant genetic background of mice used in the first electrophysiological experiments was the C57BL/6JBomTac sub-strain (90% consistent SNPs, Supplementary Table 3 and Supplementary Fig. 2), while mice used in the second electrophysiological experiment were original C57BL/6J sub-strain (96.8% consistent SNPs, Supplementary Table 3 and Supplementary Fig. 1). This difference in genomic background, in addition to environmental factors and subtle differences in experimental procedure between laboratories, may contribute to variation in electrophysiological phenotypes between the labs. Discussion Symptoms among SNAREopathy patients are diverse, with different degrees of developmental delay in different domains (language, motor function, cognition), often, but not always, accompanied by seizures and autistic features ( 8 , 16 , 18 , 19 ). Here, we investigated, using a theoretical framework combined with empirical tests, how gene interactions influence phenotypic diversity, i.e., when phenotypic differences observed among individuals with a given (disease-causing) genotype at one locus are influenced by their genotypes at another locus ( 46 ). We found strong evidence for multiplicative (epistatic) interaction between two SNAREopathy loci at the systems level: seizures and epileptiform activities ranged from no detectable or mild abnormalities in single mutants to lethal clonic and generalized seizures in double mutants. C-Fos staining showed a concomitant large variation. However, at the synapse and network level we found no evidence for such interactions. At the behavioral/cognitive level, we were only able to access surviving animals (approximately half the population, the least severely affected half). These remaining double mutant animals showed phenotypes similar to the strongest of the two single mutants ( Stxbp1 single mutants). Epistatic interactions explain diversity in seizures and EEG-abnormalities The incidence of (mild) seizure-like events among individual mice showed a normal distribution in Stxbp1 and Snap25 single mutants with more severe effects in Stxbp1 single mutants. This suggests that epileptic events in these mice are largely mediated by one factor (inactivation of one Stxbp1 or Snap25 allele, Fig. 1 model A), with Stxbp1 haploinsufficiency having a larger impact than Snap25 haploinsufficiency. Stxbp1 and Snap25 are known to work together to regulate neurotransmitter release and synaptic transmission ( 23 , 24 ). The extreme phenotypic diversity observed in double mutants is not consistent with the predicted phenotypic effects of variation in two genes acting in the same pathway (common pathway model, Fig. 1 B1) and suggests multiplicative (epistatic) gene-gene interactions (Fig. 1 , model B2.2.), and the involvement of distinct, interacting deficiencies caused by Stxbp1 and Snap25 haploinsufficiency. Such distinct, interacting deficiencies may be explained in different, not mutually exclusive ways. First, although the two genes work together in neurotransmitter release, and also in the secretion of neuropeptides and neuromodulators ( 47 , 48 ), haploinsufficiency may affect different aspects of these processes, in opposite direction or under different circumstances for the two genes. The current study did not reveal deficits in synaptic transmission for Snap25 haploinsufficiency in single neurons (Fig. 7 ), but previous studies have reported impairments in different aspects of synaptic transmission in Snap25 haploinsufficiency models which are distinct from the deficits for Stxbp1 haploinsufficiency detected in the present and the previous studies, (Fig. 7 and ( 41 , 49 – 51 )). Second, Stxbp1 and Snap25 haploinsufficiency may have distinct, interacting effects on different populations of neurons and/or brain networks. Stxbp1 haploinsufficiency was shown to have different effects on hippocampal GABAergic and glutamatergic neurons (Fig. 7 , Toonen et al. ( 50 )) or deficits specifically in GABAergic interneurons in the cortex ( 13 ), deficits in recruiting these interneurons ( 45 ) and glutamatergic inputs in the striatum ( 52 ). On the other hand, the sporadic epileptiform events in single Snap25 mice were ascribed to increased calcium responsiveness of thalamic neurons and hyper-excitability of thalamo-cortical circuits ( 53 , 54 ). In the present study we found a trend towards an increase in c-Fos expression in the cortical regions of Snap25 single mutants and increased c-Fos expression in double mutants, confirming the cortical hyperexcitability. Interestingly, c-Fos expression was increased in the thalamus of double mutants, but not in the thalamus of both single mutants, suggesting that the thalamus is a critical brain region for genetic interaction. The idea that Stxbp1 and Snap25 haploinsufficiency may have distinct, interacting effects on different populations of neurons and/or brain networks is in line with the observation that multiplicative (epistatic) effects were not observed in individual neurons (Fig. 7 ) or networks (Fig. 6 ), but were pronounced at the system level (generalized seizures, only in the double mutants). Third, in addition to their best characterized cellular functions in the regulated secretion of neurotransmitters and neuromodulators, Stxbp1 and Snap25 haploinsufficiency may affect other functions which may contribute to non-linear genetic interaction, especially during earlier developmental phases. For instance, the data from early postnatal network activity suggests deficits in the GABA shift in Stxbp1 mutants, but not Snap25 mutants (see also below). Furthermore, Snap25 , but not Stxbp1 haploinsufficiency was reported to produce negative modulation of voltage-gated calcium channels ( 51 , 55 , 56 ) and both genes have crucial, not fully overlapping roles in neuronal viability ( 57 ). Finally, epigenetic variation or stochastic processes during development may also contribute to interacting deficiencies detected on the system level. Taken together, we conclude that non-linear (multiplicative) genetic interaction of distinct aspects of regulated secretion, distinct neuronal populations/networks in the brain, with other cellular functions and/or epigenetic/stochastic effects, together explain the broad diversity in phenotypic manifestations in mutant mice. No evidence for epistatic interactions in surviving double mutants Across many behavioral domains, the behavioral phenotypes of single Stxbp1 mutants and surviving double mutants were similar: learning and memory, behavioral flexibility, anxiety and spontaneous behavior, while Snap25 single mutants behaved like control mice in all these domains. Single Stxbp1 mice and double mutant mice showed pronounced impairment in associative learning and memory and behavioral flexibility, lower body weight and normal motor coordination and muscle strength in line with previous studies in Stxbp1 haploinsufficient mice ( 13 , 14 ). Interestingly, subtle increases in activity in a habituated environment during the dark phase in single Stxbp1 mice was alleviated in double mutants. On the other hand, normal cognition, spontaneous behavior, motor coordination, muscle strength and anxiety found in single Snap25 mice are in line with previously reported lack of significant behavioral and cognitive impairments in Snap25 +/− mice ( 39 , 40 ). Behavior was assessed in the surviving mice older than eight-weeks, after approximately 50% of double mutants have died (Fig. 2 A). It is plausible that double mutants with most severe behavioral phenotypes were lost due to early lethality. However, the relationship between early lethality and later performance in behavioral tests of the surviving animals is unknown, precluding strong conclusions on which type of genetic interaction (Fig. 1 ) is most consistent with the behavioral data. A similar conclusion can be reached for sleep phenotypes, with the only variation that phenotypes of the single mutants were similar. A shift of the spectral power to higher frequencies during sleep was found in Stxbp1 single, Snap25 single and double mutants. Hence, while for behavior, the phenotypic spectrum of Stxbp1 mutants is the strongest among the two single mutants, for sleep this spectrum is approximately the same for both single mutants and the double mutants behave similar to both single mutants. A failing GABA-shift as a key determinant in Stxbp1 haploinsufficiency mice The assessment of the spontaneous synchronous activity (SSA) in two-weeks-old mutants indicated that the normal shift from depolarizing to hyperpolarizing GABA action, observed in control mice and Snap25 single mutants, was not observed in Stxbp1 single- and double-mutants. SSA is important for the establishment and maturation of functional neural circuits( 44 , 58 , 59 ) and the emergence of inhibitory GABA action is crucial for the termination of SSA ( 31 ). GABA-shift abnormalities have been observed before in several other neurodevelopmental disorders, such as Fragile-X syndrome and Rett syndrome( 60 , 61 ). The GABA-shift is mediated by several (external and internal) factors, including network activity and neuropeptide release ( 62 ). Both processes depend on SNARE-dependent vesicle fusion. Therefore, it is plausible that initial dysregulation of network activity and/or neuropeptide release is a key determinant in Stxbp1 syndrome and potentially other SNAREopathies. Interestingly, while both Snap25 and Munc18/Stxbp1 are essential for neuropeptide release from dense core vesicles (DCV) in mature neurons ( 47 , 48 ), immature Snap25 null mutant neurons showed substantial remaining DCV exocytosis, similar to immature wild type neurons, which was attributed to redundancy with other Qb/c SNAREs during early development ( 47 ). This finding provides a plausible explanation for the fact that Snap25 single mutants showed a normal GABA-shift and multiplicative (epistatic) effects were not observed for spontaneous synchronous activity in slices. Taken together, our findings support the hypothesis that neurodevelopmental disorders including Stxbp1 syndrome are initiated at early developmental stages and involve abnormalities in GABA-shift and delayed cortical network development( 63 ). Genetic interaction effects may extrapolate to the human population This study provides experimental evidence for the predicted effects of genetic interactions on phenotypic diversity (Fig. 1 ). Phenotypic diversity is a major, unexplained issue for SNAREopathies and other DEE and a major complication for the assessment of therapy success and the evaluation of new candidate therapies, e.g. in clinical trials. To establish proof of concept, the current study took a reductionistic approach, with a single ‘modifying gene’ of large effect size (heterozygous deletion) interacting with a primary genetic deficit under conditions where other genomic variation and environmental factors are radically minimalized (inbred mouse lines and highly standardized laboratory conditions, respectively). In the human population, variation in these factors add complexity and the accumulation of two genetic variants of large effect size in single individuals is very rare. However, the clear evidence for genetic interaction in the current study can be extrapolated to the human population (Fig. 8 ). Interactions between a primary genetic deficit (pathogenic variant) and multiple (common) genetic variation is expected to produce similar effects on phenotypic diversity and may explain the high diversity observed among patients with mutations in the genes that work together. This idea is consistent with the concept of ‘genetic buffering’( 64 ). According to the multiplicative model (Fig. 1 B2.2), a fraction of individuals with pathogenic variants is only mildly affected, or even unaffected, due to the ‘buffering’ effect of certain other genetic variants in the genome. Furthermore, the conclusion that epilepsy and cognitive impairments are mediated by different cellular mechanisms explains the limited efficacy of antiepileptic drugs that act on one specific cellular mechanism and their inability to ameliorate developmental aspects of DEE. Finally, this study associates atypical GABA cortical network development with neurodevelopmental delay in SNAREopathy models, suggesting a novel predictive biomarker for future research, diagnostics and treatment design. Abbreviations EPSC excitatory postsynaptic current IPSC inhibitory postsynaptic current Declarations Competing interests: The authors report no competing interests. At the start of the experiments described in this manuscript, J.K. was a full time employee of Sylics (Synaptologics BV), a private, VU University spin-off company that offers mouse phenotyping services, also on the mice described in this manuscript; during later phases, J.K. was employed by VU University. Sylics had no influence on the scientific decisions made by her and others involved in her work. During the time of this study, M.V. participated in a holding that owned Sylics shares and has received consulting fees from Sylics. The Belgian company InnoSer NV acquired Sylics in 2023. M.V. continued to received consulting fees from Sylics/InnoSer. Funding: The work was supported by the Independent Research Fund Denmark (8020-00228A to JBS) and the Lundbeck Foundation (to JBS and MV, R277-2018-802) and the Novo Nordisk Foundation (to JBS, NNF10OC0058298). Acknowledgements: We thank Sita van der Wal, Natasja Bos and Christian van der Meer for breeding mutant mice. We thank Joost Hoetjes and Dorte Lauritsen for genotyping mutant mice. We thank Rolinka van der Loo for help with organizing perfusions and surgeries. References Hamdan FF, Myers CT, Cossette P, Lemay P, Spiegelman D, Laporte AD et al. High Rate of Recurrent De Novo Mutations in Developmental and Epileptic Encephalopathies. Am J Hum Genet 2017; 101 (5) : 664-685. Scheffer IE, Berkovic S, Capovilla G, Connolly MB, French J, Guilhoto L et al. ILAE classification of the epilepsies: Position paper of the ILAE Commission for Classification and Terminology. Epilepsia 2017; 58 (4) : 512-521. Carvill GL, Heavin SB, Yendle SC, McMahon JM, O'Roak BJ, Cook J et al. Targeted resequencing in epileptic encephalopathies identifies de novo mutations in CHD2 and SYNGAP1. Nat Genet 2013; 45 (7) : 825-830. Euro E-RESC, Epilepsy Phenome/Genome P, Epi KC. De novo mutations in synaptic transmission genes including DNM1 cause epileptic encephalopathies. Am J Hum Genet 2014; 95 (4) : 360-370. Heyne HO, Singh T, Stamberger H, Abou Jamra R, Caglayan H, Craiu D et al. De novo variants in neurodevelopmental disorders with epilepsy. Nat Genet 2018; 50 (7) : 1048-1053. McTague A, Howell KB, Cross JH, Kurian MA, Scheffer IE. The genetic landscape of the epileptic encephalopathies of infancy and childhood. The Lancet Neurology 2016; 15 (3) : 304-316. Kullmann DM. Neurological channelopathies. Annu Rev Neurosci 2010; 33: 151-172. Verhage M, Sørensen JB. SNAREopathies: Diversity in Mechanisms and Symptoms. Neuron 2020; 107 (1) : 22-37. Vissers LE, Gilissen C, Veltman JA. Genetic studies in intellectual disability and related disorders. Nat Rev Genet 2016; 17 (1) : 9-18. McClellan J, King MC. Genetic heterogeneity in human disease. Cell 2010; 141 (2) : 210-217. Genç Ö, An J-Y, Fetter RD, Kulik Y, Zunino G, Sanders SJ et al. Homeostatic plasticity fails at the intersection of autism-gene mutations and a novel class of common genetic modifiers. Elife 2020; 9: e55775. Abramov D, Guiberson NGL, Daab A, Na Y, Petsko GA, Sharma M et al. Targeted stabilization of Munc18-1 function via pharmacological chaperones. EMBO Mol Med 2021; 13 (1) : e12354. Chen W, Cai ZL, Chao ES, Chen H, Longley CM, Hao S et al. Stxbp1/Munc18-1 haploinsufficiency impairs inhibition and mediates key neurological features of STXBP1 encephalopathy. Elife 2020; 9 . Kovacevic J, Maroteaux G, Schut D, Loos M, Dubey M, Pitsch J et al. Protein instability, haploinsufficiency, and cortical hyper-excitability underlie STXBP1 encephalopathy. Brain 2018; 141 (5) : 1350-1374. Lammertse HCA, van Berkel AA, Iacomino M, Toonen RF, Striano P, Gambardella A et al. Homozygous STXBP1 variant causes encephalopathy and gain-of-function in synaptic transmission. Brain 2019. Xian J, Parthasarathy S, McKeown S, Balagura G, Fitch E, Helbig K et al. Assessing the landscape of STXBP1-related disorders in 534 individuals. Brain 2021. Kadkova A, Murach J, Ostergaard M, Malsam A, Malsam J, Lolicato F et al. SNAP25 disease mutations change the energy landscape for synaptic exocytosis due to aberrant SNARE interactions. Elife 2024; 12 . Stamberger H, Nikanorova M, Willemsen MH, Accorsi P, Angriman M, Baier H et al. STXBP1 encephalopathy: A neurodevelopmental disorder including epilepsy. Neurology 2016; 86 (10) : 954-962. Abramov D, Guiberson NGL, Burré J. STXBP1 encephalopathies: Clinical spectrum, disease mechanisms, and therapeutic strategies. J Neurochem 2021; 157 (2) : 165-178. López-Rivera JA, Pérez-Palma E, Symonds J, Lindy AS, McKnight DA, Leu C et al. A catalogue of new incidence estimates of monogenic neurodevelopmental disorders caused by de novo variants. Brain 2020; 143 (4) : 1099-1105. Phillips PC. Epistasis--the essential role of gene interactions in the structure and evolution of genetic systems. Nat Rev Genet 2008; 9 (11) : 855-867. Chou HH, Chiu HC, Delaney NF, Segre D, Marx CJ. Diminishing returns epistasis among beneficial mutations decelerates adaptation. Science 2011; 332 (6034) : 1190-1192. Verhage M, Maia AS, Plomp JJ, Brussaard AB, Heeroma JH, Vermeer H et al. Synaptic assembly of the brain in the absence of neurotransmitter secretion. Science 2000; 287 (5454) : 864-869. Washbourne P, Thompson PM, Carta M, Costa ET, Mathews JR, Lopez-Benditó G et al. Genetic ablation of the t-SNARE SNAP-25 distinguishes mechanisms of neuroexocytosis. Nat Neurosci 2002; 5 (1) : 19-26. Loos M, Koopmans B, Aarts E, Maroteaux G, van der Sluis S, Verhage M et al. Sheltering behavior and locomotor activity in 11 genetically diverse common inbred mouse strains using home-cage monitoring. PLoS One 2014; 9 (9) : e108563. Maroteaux G, Loos M, van der Sluis S, Koopmans B, Aarts E, van Gassen K et al. High-throughput phenotyping of avoidance learning in mice discriminates different genotypes and identifies a novel gene. Genes Brain Behav 2012; 11 (7) : 772-784. Remmelink E, Smit AB, Verhage M, Loos M. Measuring discrimination- and reversal learning in mouse models within 4 days and without prior food deprivation. Learn Mem 2016; 23 (11) : 660-667. Wright S, Hashemi K, Stasiak L, Bartram J, Lang B, Vincent A et al. Epileptogenic effects of NMDAR antibodies in a passive transfer mouse model. Brain 2015; 138 (Pt 11) : 3159-3167. Bokil H, Andrews P, Kulkarni JE, Mehta S, Mitra PP. Chronux: a platform for analyzing neural signals. J Neurosci Methods 2010; 192 (1) : 146-151. Dawitz J, Kroon T, Hjorth JJ, Meredith RM. Functional calcium imaging in developing cortical networks. J Vis Exp 2011; (56). Pires J, Nelissen R, Mansvelder HD, Meredith RM. Spontaneous synchronous network activity in the neonatal development of mPFC in mice. Dev Neurobiol 2021; 81 (2) : 207-225. Hjorth JJ, Dawitz J, Kroon T, Pires J, Dassen VJ, Berkhout JA et al. Detection of silent cells, synchronization and modulatory activity in developing cellular networks. Dev Neurobiol 2016; 76 (4) : 357-374. Ruiter M, Kadkova A, Scheutzow A, Malsam J, Sollner TH, Sorensen JB. An Electrostatic Energy Barrier for SNARE-Dependent Spontaneous and Evoked Synaptic Transmission. Cell Rep 2019; 26 (9) : 2340-2352 e2345. Meijer M, Dörr B, Lammertse HC, Blithikioti C, van Weering JR, Toonen RF et al. Tyrosine phosphorylation of Munc18-1 inhibits synaptic transmission by preventing SNARE assembly. Embo j 2018; 37 (2) : 300-320. Meijer M, Burkhardt P, de Wit H, Toonen RF, Fasshauer D, Verhage M. Munc18-1 mutations that strongly impair SNARE-complex binding support normal synaptic transmission. Embo j 2012; 31 (9) : 2156-2168. Fisher RA. The correlations between relatives on the supposition of Mendelian inheritance. Trans R Soc Edinb 1918; 52: 399-433. Mani R, St Onge RP, Hartman JLt, Giaever G, Roth FP. Defining genetic interaction. Proc Natl Acad Sci U S A 2008; 105 (9) : 3461-3466. Sigmon JS, Blanchard MW, Baric RS, Bell TA, Brennan J, Brockmann GA et al. Content and Performance of the MiniMUGA Genotyping Array: A New Tool To Improve Rigor and Reproducibility in Mouse Research. Genetics 2020; 216 (4) : 905-930. Corradini I, Donzelli A, Antonucci F, Welzl H, Loos M, Martucci R et al. Epileptiform activity and cognitive deficits in SNAP-25(+/-) mice are normalized by antiepileptic drugs. Cereb Cortex 2014; 24 (2) : 364-376. Monteggia LM, Lin PY, Adachi M, Kavalali ET. Behavioral Analysis of SNAP-25 and Synaptobrevin-2 Haploinsufficiency in Mice. Neuroscience 2019; 420: 129-135. Orock A, Logan S, Deak F. Munc18-1 haploinsufficiency impairs learning and memory by reduced synaptic vesicular release in a model of Ohtahara syndrome. Mol Cell Neurosci 2018; 88: 33-42. Herrera DG, Robertson HA. Activation of c-fos in the brain. Prog Neurobiol 1996; 50 (2-3) : 83-107. Deacon RM, Rawlins JN. T-maze alternation in the rodent. Nat Protoc 2006; 1 (1) : 7-12. Blankenship AG, Feller MB. Mechanisms underlying spontaneous patterned activity in developing neural circuits. Nat Rev Neurosci 2010; 11 (1) : 18-29. Dos Santos AB, Larsen SD, Guo L, Barbagallo P, Montalant A, Verhage M et al. Microcircuit failure in STXBP1 encephalopathy leads to hyperexcitability. Cell Rep Med 2023; 4 (12) : 101308. Cheverud JM, Routman EJ. Epistasis and its contribution to genetic variance components. Genetics 1995; 139 (3) : 1455-1461. Arora S, Saarloos I, Kooistra R, van de Bospoort R, Verhage M, Toonen RF. SNAP-25 gene family members differentially support secretory vesicle fusion. J Cell Sci 2017; 130 (11) : 1877-1889. Puntman DC, Arora S, Farina M, Toonen RF, Verhage M. Munc18-1 is essential for neuropeptide secretion in neurons. J Neurosci 2021; 41 (28) : 5980-5993. Patzke C, Han Y, Covy J, Yi F, Maxeiner S, Wernig M et al. Analysis of conditional heterozygous STXBP1 mutations in human neurons. J Clin Invest 2015; 125 (9) : 3560-3571. Toonen RF, Wierda K, Sons MS, de Wit H, Cornelisse LN, Brussaard A et al. Munc18-1 expression levels control synapse recovery by regulating readily releasable pool size. Proc Natl Acad Sci U S A 2006; 103 (48) : 18332-18337. Antonucci F, Corradini I, Morini R, Fossati G, Menna E, Pozzi D et al. Reduced SNAP-25 alters short-term plasticity at developing glutamatergic synapses. EMBO Rep 2013; 14 (7) : 645-651. Miyamoto H, Tatsukawa T, Shimohata A, Yamagata T, Suzuki T, Amano K et al. Impaired cortico-striatal excitatory transmission triggers epilepsy. Nat Commun 2019; 10 (1) : 1917. Watanabe S, Yamamori S, Otsuka S, Saito M, Suzuki E, Kataoka M et al. Epileptogenesis and epileptic maturation in phosphorylation site-specific SNAP-25 mutant mice. Epilepsy Res 2015; 115: 30-44. Zhang Y, Vilaythong AP, Yoshor D, Noebels JL. Elevated thalamic low-voltage-activated currents precede the onset of absence epilepsy in the SNAP25-deficient mouse mutant coloboma. J Neurosci 2004; 24 (22) : 5239-5248. Condliffe SB, Corradini I, Pozzi D, Verderio C, Matteoli M. Endogenous SNAP-25 regulates native voltage-gated calcium channels in glutamatergic neurons. J Biol Chem 2010; 285 (32) : 24968-24976. Toft-Bertelsen TL, Ziomkiewicz I, Houy S, Pinheiro PS, Sørensen JB. Regulation of Ca2+ channels by SNAP-25 via recruitment of syntaxin-1 from plasma membrane clusters. Mol Biol Cell 2016; 27 (21) : 3329-3341. Santos TC, Wierda K, Broeke JH, Toonen RF, Verhage M. Early Golgi Abnormalities and Neurodegeneration upon Loss of Presynaptic Proteins Munc18-1, Syntaxin-1, or SNAP-25. J Neurosci 2017; 37 (17) : 4525-4539. Kirkby LA, Sack GS, Firl A, Feller MB. A role for correlated spontaneous activity in the assembly of neural circuits. Neuron 2013; 80 (5) : 1129-1144. Ben-Ari Y. Excitatory actions of gaba during development: the nature of the nurture. Nat Rev Neurosci 2002; 3 (9) : 728-739. He Q, Nomura T, Xu J, Contractor A. The developmental switch in GABA polarity is delayed in fragile X mice. J Neurosci 2014; 34 (2) : 446-450. Tang X, Jaenisch R, Sur M. The role of GABAergic signalling in neurodevelopmental disorders. Nat Rev Neurosci 2021; 22 (5) : 290-307. Peerboom C, Wierenga CJ. The postnatal GABA shift: A developmental perspective. Neurosci Biobehav Rev 2021; 124: 179-192. Ben-Ari Y. Neuro-archaeology: pre-symptomatic architecture and signature of neurological disorders. Trends Neurosci 2008; 31 (12) : 626-636. Bourgeron T. From the genetic architecture to synaptic plasticity in autism spectrum disorder. Nat Rev Neurosci 2015; 16 (9) : 551-563. Additional Declarations The authors have declared there is NO conflict of interest to disclose Supplementary Files SupplementaryTables250324.xlsx Supplemental Tables Suppl.Video1.mp4 Generalized epileptic attack with accompanied ECoG-detected icta levents Suppl.Figureswithlegend.pdf Supplemental Figures Cite Share Download PDF Status: Published Journal Publication published 05 Mar, 2026 Read the published version in Molecular Psychiatry → Version 1 posted Editorial decision: revise 14 Aug, 2025 Review # 1 received at journal 27 May, 2025 Review # 4 received at journal 27 May, 2025 Reviewer # 5 agreed at journal 16 May, 2025 Reviewer # 4 agreed at journal 14 May, 2025 Reviewer # 3 agreed at journal 12 May, 2025 Reviewer # 2 agreed at journal 30 Apr, 2025 Reviewer # 1 agreed at journal 28 Apr, 2025 Reviewers invited by journal 28 Apr, 2025 Editor assigned by journal 16 Apr, 2025 Submission checks completed at journal 16 Apr, 2025 First submitted to journal 15 Apr, 2025 Unknown event 14 Apr, 2025 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-6440830","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":449191838,"identity":"43186541-3097-48bb-9461-6e72ad018e9a","order_by":0,"name":"Matthijs Verhage","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAsElEQVRIiWNgGAWjYLACxgYgwQ4iGCxI0cJzAMSUIEWLRAKRWnQbeAwfV+6ws+uXfGMmwbiDCC1mB3iMDc+eSU6eOTsHqOUMUVrY0iQb25iTDW6npUkwthGnJf1nY1t9sv3NY0RrYT7G2Nh22M5AgvkYkVoOMx8GOux4gsSZ5MMWiURpOd7Y+LGxrdqev/1g442PbTaEtTAwQ6jEBhCZQIQGOLAnRfEoGAWjYBSMMAAAhjI0oPQHUlcAAAAASUVORK5CYII=","orcid":"","institution":"Vrije Universiteit Amsterdam","correspondingAuthor":true,"prefix":"","firstName":"Matthijs","middleName":"","lastName":"Verhage","suffix":""},{"id":449191839,"identity":"2e154f28-7aee-41c5-bad8-8e62dad3e6aa","order_by":1,"name":"Jovana Kovacevic","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Jovana","middleName":"","lastName":"Kovacevic","suffix":""},{"id":449191840,"identity":"9c4593bb-fba1-492b-ac43-85088b69089a","order_by":2,"name":"Sébastien Houy","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Sébastien","middleName":"","lastName":"Houy","suffix":""},{"id":449191841,"identity":"f51603bb-4ec1-45be-a64d-4fdfc6fd5bbc","order_by":3,"name":"Johny Pires","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Johny","middleName":"","lastName":"Pires","suffix":""},{"id":449191842,"identity":"6f7174db-d277-4cb1-83b7-bc18de8a5f11","order_by":4,"name":"Anna Kádková","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Anna","middleName":"","lastName":"Kádková","suffix":""},{"id":449191843,"identity":"9ae27754-a0cb-4809-ad0e-6eb35a496fb3","order_by":5,"name":"Hanna Lammertse","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Hanna","middleName":"","lastName":"Lammertse","suffix":""},{"id":449191844,"identity":"ffdd160c-1d3a-48ac-b941-61ecec348e86","order_by":6,"name":"Joana Martins","email":"","orcid":"https://orcid.org/0000-0002-6721-2935","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Joana","middleName":"","lastName":"Martins","suffix":""},{"id":449191845,"identity":"2b29cf6f-134f-45ec-a6d9-bd92330b38be","order_by":7,"name":"Miriam Öttl","email":"","orcid":"https://orcid.org/0000-0001-6435-4658","institution":"Vrije Universiteit Amsterdam","correspondingAuthor":false,"prefix":"","firstName":"Miriam","middleName":"","lastName":"Öttl","suffix":""},{"id":449191846,"identity":"899d469d-d08a-46aa-aaee-decf94dfe339","order_by":8,"name":"Keimpe Wierda","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Keimpe","middleName":"","lastName":"Wierda","suffix":""},{"id":449191847,"identity":"7037f2c6-d7db-482d-93c4-d215e35437b9","order_by":9,"name":"Joke Wortel","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Joke","middleName":"","lastName":"Wortel","suffix":""},{"id":449191848,"identity":"8adb141a-40ce-4454-b110-14a9ef701c2b","order_by":10,"name":"Jakob Sørensen","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Jakob","middleName":"","lastName":"Sørensen","suffix":""}],"badges":[],"createdAt":"2025-04-13 19:20:08","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-6440830/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-6440830/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1038/s41380-026-03509-3","type":"published","date":"2026-03-05T05:00:00+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":82148268,"identity":"7a0ea9f2-bffe-434f-853d-dba03069260a","added_by":"auto","created_at":"2025-05-07 07:07:08","extension":"jpg","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":97431,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eModels of gene interaction to explain phenotypic diversity in population. (A) \u003c/strong\u003eEffects of pathogenic variation in two hypothetical genes (A and B) are represented as normally distributed severity of phenotypes that are shifted to the right relative to the normal (healthy) individuals, indicating pathogenicity. (\u003cstrong\u003eB)\u003c/strong\u003e Graphs from B.1 to B2.2. represent the predicted effects of the combination of variants in the two genes in one individual. \u003cstrong\u003eB.1)\u003c/strong\u003e Common pathway model: combinations of two variants in these genes are predicted to lead to phenotypes similar to the strongest single variant (a\u003csub\u003eA\u003c/sub\u003e or a\u003csub\u003eB\u003c/sub\u003e). The resulting distribution is similar, but not identical, to the distribution of the more severe of the two variants (α\u003csub\u003eB\u003c/sub\u003e); the distribution was identified by simulating pairs of normally distributed variables according to a\u003csub\u003eA\u003c/sub\u003e or a\u003csub\u003eB\u003c/sub\u003e, and selecting the maximal number. \u003cstrong\u003eB.2.1)\u003c/strong\u003e Additive interaction model: phenotypic severity in individuals with variants in two genes is represented as the summation of effects of single gene variants (a\u003csub\u003eA\u003c/sub\u003e + a\u003csub\u003eB\u003c/sub\u003e). If the effects of single gene variants follow a normal distribution (as represented in panel A), the effect of a combination of variants in two or more genes follows a normal distribution, according to an additive model of interaction. The vertical dashed line indicates 20% of most severe cases in an additive model with two hypothetical gene variants (a\u003csub\u003eA\u003c/sub\u003e and a\u003csub\u003eB\u003c/sub\u003e), called ‘moderate’ phenotype. The percentage of population with moderate phenotype according to each model is indicated. \u003cstrong\u003eB.2.2\u003c/strong\u003e) Multiplicative interaction model: phenotypic severity in individuals with variants in two genes is represented as the multiplication of the effects of single gene variants (a\u003csub\u003eA\u003c/sub\u003e * a\u003csub\u003eB\u003c/sub\u003e). If the effects of single gene variants follow the normal distribution, according to multiplicative model of interaction the effect of combination of the variants in more genes would be right-skewed, approaching lognormal distribution, with larger variance compared to additive model. The blank (non-filled) curves represent the effects of single gene variants, re-plotted from panel A. Vertical black dash-dash line is drawn to represent 20% of most severe cases of multiplicative model, called ‘severe’ phenotype. Vertical grey dash-dot-dash line is drawn to represent 10% of the least severe cases of multiplicative model, called ‘mild’ phenotype. The percentage of population with mild and severe phenotypes according to other models are represented within the respective graphs. Following parameters were used to generate a model: normal (mean = 0, s.d. = 0.1), variant in gene 1 (mean = 2, s.d. = 0.6), variant in gene 2 (mean = 4, s.d. = 1.4) on 1*10\u003csup\u003e6\u003c/sup\u003e random repetitions. Distribution fit was performed in MATLAB.\u003c/p\u003e","description":"","filename":"Figure1.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6440830/v1/cdaa83afd7ef6094d834ba9a.jpg"},{"id":82148273,"identity":"e7755658-a402-43d9-a7a7-49ee7282f9b3","added_by":"auto","created_at":"2025-05-07 07:07:08","extension":"jpg","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":190313,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eStrong additive effects in double mutant mice on survival, seizure-related behaviors but not ECoG telemetry-recordings. (A) \u003c/strong\u003eNumber of mice in litters with experimental time-line. The number of mice was calculated from 12 litters at E18 and 10 litters at age of three weeks (P3w). The number of mice per genotype in genotyped and weaned litters was counted again at the age of 6 weeks, after video monitoring. Exact number of mice per genotype in 12litters (E18) and 10 litters (P3w/P\u0026gt;6w) is shown in pie-charts. \u003cstrong\u003e(B) \u003c/strong\u003eVideo monitoring of single-housed mice revealed different types of behavioral events: generalized seizure, clonic attack, jump, twitch, myoclonus and body extension. The aberrant behavioral events were analyzed in controls (n = 2), \u003cem\u003eStxbp1\u003c/em\u003e single mutants (n = 5), \u003cem\u003eSnap25\u003c/em\u003e single mutants (n = 4) and double mutants\u003cem\u003e \u003c/em\u003e(n = 6). (\u003cstrong\u003eC)\u003c/strong\u003e Average number of twitches and jumps detected in mice during 12h of video monitoring. (\u003cstrong\u003eD)\u003c/strong\u003e Number of clonic attacks and generalized seizures detected in controls (n = 4), \u003cem\u003eStxbp1\u003c/em\u003e single mutants (n = 4), \u003cem\u003eSnap25\u003c/em\u003e single mutants (n = 4) and double mutants\u003cem\u003e \u003c/em\u003e(n = 7) during 12h of video monitoring. (\u003cstrong\u003eE)\u003c/strong\u003eRepresentative ECoG trace of tonic-clonic generalized seizures detected in double mutant\u003cem\u003e \u003c/em\u003emouse with corresponding behavioral state pictogram. \u003cstrong\u003e(F)\u003c/strong\u003eModel of epileptic diversity represents the diversity of epileptic phenotype according to multiplicative model of interaction. We assume that most severe phenotype corresponds to lethal generalized seizures; moderate phenotype corresponds to clonic-tonic seizures and mild phenotype corresponds to low-incidence of SWDs. (\u003cstrong\u003eG)\u003c/strong\u003e 3D representation of different epileptiform events library. Event Classifier application in Neuroarchiver software distinguished three different types of epileptiform activities: SWDs, sharp spikes and spike-slow waves. (\u003cstrong\u003eH-J)\u003c/strong\u003e Number of different epileptiform events during 24h in controls, \u003cem\u003eStxbp1\u003c/em\u003esingle-, \u003cem\u003eSnap25\u003c/em\u003e single- and double mutants represented as natural logarithm of (number of events +1). \u003cstrong\u003e(H)\u003c/strong\u003e Number of SWDs (\u003cstrong\u003eI)\u003c/strong\u003eNumber of spike-slow waves and (\u003cstrong\u003eJ)\u003c/strong\u003eNumber of sharp spikes.\u003cstrong\u003e (K) \u003c/strong\u003eRepresentation of ECoG traces and corresponding time-frequency spectogram for control, \u003cem\u003eStxbp1\u003c/em\u003e single-, \u003cem\u003eSnap25 \u003c/em\u003esingle- and double mutants. (\u003cstrong\u003eL)\u003c/strong\u003e Average relative power during the sleep episodes for all four genotypes. (\u003cstrong\u003eM)\u003c/strong\u003e Quantification of average relative power during the sleep in two frequency bands: delta band (1-4 Hz) and alfa band (9-12 Hz). Shift of relative power to higher frequency bands for all four genotypes is represented as the significant increase of power in alfa frequency band independent on the genotype. Panels represent mean value ± SEM.\u003c/p\u003e","description":"","filename":"Figure2.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6440830/v1/90ee48cb5268297a7cebb0e9.jpg"},{"id":82154592,"identity":"19e6acab-4ea5-40d4-a629-af4091c731e0","added_by":"auto","created_at":"2025-05-07 07:31:08","extension":"jpg","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":3499194,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003ec-Fos expression and quantification. (A - D) \u003c/strong\u003eRepresentatives of c-Fos expression in controls and double mutants in the prefrontal cortex (PFC), cortex, thalamus (Thal) and hippocampus. \u003cstrong\u003e(E)\u003c/strong\u003e Quantification of c-Fos expression in 5 brain regions expressed as relative number of c-Fos positive cells in \u003cem\u003eStxbp1\u003c/em\u003e\u003csup\u003e\u003cem\u003e+/-\u003c/em\u003e\u003c/sup\u003e single BL6 mutants (n = 5), \u003cem\u003eSnap25\u003c/em\u003e single mutants (n = 6) and double mutants (n = 10) compared to appropriate control.\u003c/p\u003e","description":"","filename":"Figure3.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6440830/v1/7af41565817ccdec2cd683bb.jpg"},{"id":82148276,"identity":"51394925-606b-46ff-8268-830e6089337d","added_by":"auto","created_at":"2025-05-07 07:07:08","extension":"jpg","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":1024456,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eAssessment of different aspects of learning and memory in single \u003c/strong\u003e\u003cem\u003e\u003cstrong\u003eStxbp1\u003c/strong\u003e\u003c/em\u003e\u003cstrong\u003e, single \u003c/strong\u003e\u003cem\u003e\u003cstrong\u003eSnap25 \u003c/strong\u003e\u003c/em\u003e\u003cstrong\u003eand double mutant\u003c/strong\u003e\u003cem\u003e\u003cstrong\u003e \u003c/strong\u003e\u003c/em\u003e\u003cstrong\u003emice. (A)\u003c/strong\u003e Fear conditioning test protocol. (\u003cstrong\u003eB)\u003c/strong\u003e Percentage of freezing duration per trial for control, \u003cem\u003eStxbp1\u003c/em\u003e single-, \u003cem\u003eSnap25\u003c/em\u003e single- and double mutants. (\u003cstrong\u003eC)\u003c/strong\u003e Quantification of percentage of freezing time during exposure to the training context during the second day and exposure to the new context and tone in new context during the third day. \u003cem\u003eStxbp1\u003c/em\u003e single- and double- mutants showed decreased percentage of freezing during those three trials. (\u003cstrong\u003eD)\u003c/strong\u003e Barnes maze test protocol. (\u003cstrong\u003eE)\u003c/strong\u003e Latency to find a platform during the acquisition phase was similar for all tested groups. (\u003cstrong\u003eF)\u003c/strong\u003e Hole visits in target octant during the first probe trial was significantly increased for \u003cem\u003eStxbp1\u003c/em\u003esingle- and double- mutants. (\u003cstrong\u003eG)\u003c/strong\u003eLatency to find target hole during reversal phase was significantly longer for double mutants during reversal phase learning. Panels represent mean value ± SEM.\u003c/p\u003e","description":"","filename":"Figure4.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6440830/v1/c7ecb4cba8668f3367f71e95.jpg"},{"id":82148274,"identity":"9d77316b-0109-4d42-9aaa-9ba3deb1094a","added_by":"auto","created_at":"2025-05-07 07:07:08","extension":"jpg","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":827848,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eAnalysis of spontaneous behavior in automated home-cage environment (PhenoTyper). (A)\u003c/strong\u003e Statistical analysis of 90 parameters of spontaneous behavior separated into 6 categories: kinematics, activity, sheltering, habituation, dark-light index and phase transition. Three annuli show statistical analysis for double mutants, \u003cem\u003eStxbp1\u003c/em\u003e single- and \u003cem\u003eSnap25\u003c/em\u003esingle- mutants compared to their control groups, respectively from periphery to the center. The full list of parameters and statistical analysis can be found in Supplemental Table 2. (\u003cstrong\u003eB-C)\u003c/strong\u003eKinematic parameters: short movement number during the light and long movement number during the light were significantly decreased in \u003cem\u003eStxbp1\u003c/em\u003e single- and double- mutants compared to control group. (\u003cstrong\u003eD) \u003c/strong\u003eActivity parameter: On shelter zone duration during the dark was significantly decreased in \u003cem\u003eStxbp1\u003c/em\u003e single- and double- mutants compared to control groups. (\u003cstrong\u003eE)\u003c/strong\u003e Sheltering parameter: Long shelter visit number during light was significantly decreased in double mutants compared to \u003cem\u003eSnap25\u003c/em\u003e single mutants and controls. (\u003cstrong\u003eF)\u003c/strong\u003e Habituation: On shelter zone habituation ratio was significantly decreased in \u003cem\u003eStxbp1\u003c/em\u003e single- and double- mutants compared to control group. Panels B - F represent mean value ± SEM\u003c/p\u003e","description":"","filename":"Figure5.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6440830/v1/a4b6378969e900b8b7a4daad.jpg"},{"id":82148279,"identity":"e4e6d828-e45d-482c-8a1b-0cd9591bcdf5","added_by":"auto","created_at":"2025-05-07 07:07:08","extension":"jpg","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":1560468,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eCalcium imaging of spontaneous synchronized activity in the prefrontal cortex (PFC).\u003c/strong\u003e (\u003cstrong\u003eA)\u003c/strong\u003e Pictorgram representation of experimental design. PFC-brain slices of mice, age postnatal day 14 (P14) were loaded with Fura-2 AM and imaged under two-photon microscope for 8 mins to assess baseline network activity. The effect of gabazine was assessed during the following 8 minutes of recording. (\u003cstrong\u003eB)\u003c/strong\u003e Contour map of Fura2-AM ester bulk-loaded cells and representative traces of cells in the PFC brain slices. Active neurons are shown in blue, silent neurons are shown in black and synchronized neurons are shown in red. Scale bar is 50 mm. (\u003cstrong\u003eC - D)\u003c/strong\u003e The percentage (C) and frequency (D) of SSA cells at the baseline and after application of gabazine in the PFC- brain slices from control, \u003cem\u003eStxbp1\u003c/em\u003e single-, \u003cem\u003eSnap25\u003c/em\u003e single- and double- mutants (n = 24, n = 27, n = 25, n = 17, respectively). Panels C and D represent mean value ± SEM.\u003c/p\u003e","description":"","filename":"Figure6.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6440830/v1/75845c79cbd0dc4cea81f965.jpg"},{"id":82148300,"identity":"b20416ff-0836-4586-873a-e04267db428c","added_by":"auto","created_at":"2025-05-07 07:07:08","extension":"jpg","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":2497875,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eAssessment of synaptic transmission phenotype in the autaptic hippocampal cultures in two separate experiments. (A)\u003c/strong\u003e Typical evoked responses (EPSC) in control-, \u003cem\u003eStxbp1\u003c/em\u003e single\u003cem\u003e-\u003c/em\u003e, \u003cem\u003eSnap25-\u003c/em\u003e single and double\u003cem\u003e-\u003c/em\u003e mutant glutamatergic neurons in 2 mM / 2 mM Ca\u003csup\u003e2+\u003c/sup\u003e/Mg\u003csup\u003e2+\u003c/sup\u003e concentrations. (\u003cstrong\u003eB)\u003c/strong\u003e Normalized EPSC per week. EPSC was significantly lower in \u003cem\u003eStxbp1\u003c/em\u003e single - and double\u003cem\u003e-\u003c/em\u003e mutant glutamatergic neurons compared to control group. (\u003cstrong\u003eC)\u003c/strong\u003e Representative traces of 500mM sucrose responses. (\u003cstrong\u003eD)\u003c/strong\u003e Normalized readily releasable pool (RRP) of glutamatergic neurons calculated from hypertonic sucrose responses. RRP was significantly lower in \u003cem\u003eStxbp1 \u003c/em\u003esingle- and double\u003cem\u003e-\u003c/em\u003e mutant glutamatergic neurons compared to control group. (\u003cstrong\u003eE)\u003c/strong\u003e Typical inhibitory evoked response (IPSC) in control, \u003cem\u003eStxbp1 \u003c/em\u003esingle- , \u003cem\u003eSnap25 \u003c/em\u003esingle- and double\u003cem\u003e-\u003c/em\u003e mutants GABA-ergic neurons in 2 mM / 2 mM Ca\u003csup\u003e2+\u003c/sup\u003e/Mg\u003csup\u003e2+\u003c/sup\u003e concentrations. (\u003cstrong\u003eF)\u003c/strong\u003e Normalized IPSC per week. (\u003cstrong\u003eG)\u003c/strong\u003e Representative traces of 500mM sucrose responses in GABA-ergic neurons. (\u003cstrong\u003eH)\u003c/strong\u003e Normalized RRP in GABA-ergic neurons. (\u003cstrong\u003eI)\u003c/strong\u003e Normalized EPSC per week in control-, \u003cem\u003eStxbp1 \u003c/em\u003esingle \u003cem\u003e-\u003c/em\u003e, \u003cem\u003eSnap25 \u003c/em\u003esingle\u003cem\u003e-\u003c/em\u003e and double\u003cem\u003e-\u003c/em\u003e mutant glutamatergic neurons in 2 mM / 4 mM Ca\u003csup\u003e2+\u003c/sup\u003e/Mg\u003csup\u003e2+\u003c/sup\u003e concentrations. (\u003cstrong\u003eJ)\u003c/strong\u003e Normalized paired pulse ratio (PPR) per week per group for paired-pulse intervals of 20 ms, 50 ms and 100 ms. (\u003cstrong\u003eK)\u003c/strong\u003e Cumulative charge during train-stimulation (40Hz, 100 pulses) with back-extrapolation line of last 20 pulses used for estimation of RRP. (\u003cstrong\u003eL)\u003c/strong\u003e Normalized synaptic rundown of EPSC during first ten pulses of 40 Hz train-stimulation. ** p \u0026lt; 0.01, *** p \u0026lt; 0.001 compared to control group.\u003c/p\u003e","description":"","filename":"Figure7.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6440830/v1/848c4f402d06a3e7b8ad13f4.jpg"},{"id":82150252,"identity":"518fa3c1-e931-4b77-bfa0-aea4e6ea7ddd","added_by":"auto","created_at":"2025-05-07 07:15:08","extension":"jpg","order_by":8,"title":"Figure 8","display":"","copyAsset":false,"role":"figure","size":234819,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eModel of gene interaction in the human population. \u003c/strong\u003eThe model explains phenotypic diversity in the human population with one pathogenic variant and three common variants that interact genetically, physically with pathogenic variant. \u003cstrong\u003e(A)\u003c/strong\u003e Effects of three common gene variants (cv\u003csub\u003e1\u003c/sub\u003e, cv\u003csub\u003e2\u003c/sub\u003e and cv\u003csub\u003e3\u003c/sub\u003e, green, blue and purple lines, respectively) and one pathogenic variant (pv, yellow curve) on the phenotypic severity are represented as normally distributed severity of phenotypes that are shifted to the right compared to the normal (healthy) individual (grey curve). Green, blue and purple curves represent the predicted effects of the multiplicative (pv*cv\u003csub\u003e1\u003c/sub\u003e, blue curve) and the additive (pv + cv\u003csub\u003e2\u003c/sub\u003e and pv + cv\u003csub\u003e3\u003c/sub\u003e, green and purple curves, respectively) gene interactions between one common variant with pathogenic variant. \u003cstrong\u003e(B)\u003c/strong\u003e Distribution of the phenotypic severity in the human population is represented with pink curve. The pink curve shows the distribution of the phenotypic severity in population containing equal proportion of individuals with cv\u003csub\u003e1\u003c/sub\u003e, cv\u003csub\u003e2\u003c/sub\u003e and cv\u003csub\u003e3\u003c/sub\u003e interacting with pathogenic variant.\u003c/p\u003e","description":"","filename":"Figure8.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6440830/v1/916d72318d76f63af6bdcc81.jpg"},{"id":103973938,"identity":"d9f830b9-d100-4184-8f6f-67a882879316","added_by":"auto","created_at":"2026-03-05 08:13:15","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":9164052,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6440830/v1/8189c8a1-6cfa-4a1c-aa17-f0449227c26c.pdf"},{"id":82148269,"identity":"fd01511a-34f6-473a-8088-015a8f08fa7f","added_by":"auto","created_at":"2025-05-07 07:07:08","extension":"xlsx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":25258,"visible":true,"origin":"","legend":"Supplemental Tables","description":"","filename":"SupplementaryTables250324.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-6440830/v1/8afcbce336e369ef093ed0bb.xlsx"},{"id":82150247,"identity":"f100d3b3-9c2f-419c-af42-6ba723893867","added_by":"auto","created_at":"2025-05-07 07:15:08","extension":"mp4","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":4438443,"visible":true,"origin":"","legend":"Generalized epileptic attack with accompanied ECoG-detected icta levents","description":"","filename":"Suppl.Video1.mp4","url":"https://assets-eu.researchsquare.com/files/rs-6440830/v1/44fcd3de408b82b7ebb74a65.mp4"},{"id":82148288,"identity":"5ebdbf92-04d0-46d1-9965-796dd29a5c70","added_by":"auto","created_at":"2025-05-07 07:07:08","extension":"pdf","order_by":3,"title":"","display":"","copyAsset":false,"role":"supplement","size":10101404,"visible":true,"origin":"","legend":"Supplemental Figures","description":"","filename":"Suppl.Figureswithlegend.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6440830/v1/161760ee469b18db7a593d52.pdf"}],"financialInterests":"The authors have declared there is \u003cb\u003eNO\u003c/b\u003e conflict of interest to disclose","formattedTitle":"Phenotypic diversity is caused by non-linear genetic interactions between two SNAREopathy genes","fulltext":[{"header":"Introduction","content":"\u003cp\u003eDevelopmental and epileptic encephalopathies (DEEs) are a heterogeneous group of severe childhood disorders characterized by different types of seizures and epileptiform activities, intellectual disability (ID) and severe neurodevelopmental delays (\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e, \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e). DEEs are characterized by large genetic and phenotypic diversity, with hundreds of genes implicated as causal, and each gene associated with multiple phenotypes (\u003cspan additionalcitationids=\"CR4 CR5\" citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e). The genetic diversity implies the involvement of several cellular pathways in the etiology of DEEs. Sub-grouping genes to specific cellular pathways could be used as a strategy to resolve this heterogeneity and elucidate disease mechanisms more efficiently. This led to classification of channelopathies, synaptopathies, and recently also SNAREopathies (\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e, \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e, \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e). However, striking phenotypic diversity remains within such DEE sub-groups, even among patients with mutations in the same gene, suggesting that additional explanations are required. Several studies have suggested influences of interacting factors such as modifying genes, epigenetic factors, environmental factors, and stochastic processes (\u003cspan additionalcitationids=\"CR10\" citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e). However, experimental tests of such suggestions in controlled laboratory conditions are scarce.\u003c/p\u003e \u003cp\u003eSNAREopathies are caused by mutations in genes encoding the neuronal SNARE (soluble N-ethylmaleimide sensitive factor attachment protein receptor)-complex and its interactors, which together form the essential machinery for the secretion of chemical signals in the brain. Most SNAREopathies are caused by \u003cem\u003ede novo\u003c/em\u003e missense or loss-of-function variants in one of the SNARE genes and several molecular disease mechanisms have been suggested including haploinsufficiency, dominant-negative, gain-of-function and recessive mechanisms (\u003cspan additionalcitationids=\"CR13 CR14 CR15 CR16\" citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e). Despite the fact that SNARE-genes work together in a single, integrated molecular machine, a striking phenotypic diversity among SNAREopathy patients is evident (\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e, \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e, \u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e). However, the incidence of pathogenic mutations in individual SNARE genes in the population is low, with an estimated incidence of one in 30,000 individuals for the most reported SNAREopathy to date, STXBP1 syndrome (\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e). This severely limits the power to identify modifying genes and epigenetic or environmental factors in population studies.\u003c/p\u003e \u003cp\u003eIn the present study we propose a theoretical framework to explain extreme phenotypic diversity for \u0026lsquo;monogenic\u0026rsquo; disorders and test this model experimentally in controlled laboratory conditions using validated mouse models for SNAREopathies. We defined three modes of gene interactions that can be experimentally tested: common pathway, additive- and multiplicative- interaction, based on previous models (\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e, \u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e) and selected two SNARE genes, \u003cem\u003eStxbp1\u003c/em\u003e and \u003cem\u003eSnap25\u003c/em\u003e, expected to have genetic interactions and confirmed pathogenic potential for DEE. We established a panel of four mouse models for single and combined haploinsufficiency on the same genetic background to minimize other sources of genetic variation and tested the phenotypic consequences of gene interactions on the cellular level by assessing synaptic transmission in cultured neurons, on the network level by assessing spontaneous synchronous network activity (SSA) in acute brain slices and excessive neuronal activation by c-Fos staining in brain sections, and on the system level, using ECoG-video monitoring and cognitive assessments. We found robust support for multiplicative genetic interactions at the highest level of organization, but not at the lower (cellular) levels. Generalized seizures, complex epileptiform activities and brain hyper-excitability were observed in double mutants, but never in single mutants.\u003c/p\u003e"},{"header":"Materials and methods","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003eStudy design\u003c/h2\u003e \u003cp\u003eThe study had multilevel design: \u003cem\u003ein silico\u003c/em\u003e, \u003cem\u003ein vitro\u003c/em\u003e, \u003cem\u003eex vivo\u003c/em\u003e and \u003cem\u003ein vivo\u003c/em\u003e. The study followed the ARRIVE guidelines (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://arriveguidelines.org/arrive-guidelines\u003c/span\u003e\u003cspan address=\"https://arriveguidelines.org/arrive-guidelines\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e). We generated, behaviorally and electrographically characterized \u003cem\u003eStxbp1\u003c/em\u003e single, \u003cem\u003eSnap25\u003c/em\u003e single and double mutants and compared them to wildtype littermates (controls). We performed standardized battery of behavioral tests for assessment of different aspects of behavior. For behavioral analysis and ECoG recording, we used in total 79 mice. Experiments were performed in several batches and data were pooled. \u003cem\u003eEx vivo\u003c/em\u003e experiments were performed on the brain slices from mice. For Ca\u003csup\u003e2+\u003c/sup\u003e-imaging of the brain slices, 34 mice were used, and between 3\u0026ndash;5 replicates per animal were analyzed. C-Fos immunohistochemistry was performed on brain slices from 21 mice. \u003cem\u003eIn vitro\u003c/em\u003e analysis of electrophysiological properties of neurons was performed in two batches of mice (n\u0026thinsp;=\u0026thinsp;31 and 23 per batch). All experiments were performed by researcher unaware of the animal\u0026rsquo;s genotype.\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eSubjects\u003c/h3\u003e\n\u003cp\u003eControl (\u003cem\u003eStxbp1\u003c/em\u003e\u003csup\u003e\u003cem\u003e+/+\u003c/em\u003e\u003c/sup\u003e\u003cem\u003eSnap25\u003c/em\u003e\u003csup\u003e\u003cem\u003e+/+\u003c/em\u003e\u003c/sup\u003e), \u003cem\u003eSnap25\u003c/em\u003e single mutant (\u003cem\u003eStxbp1\u003c/em\u003e\u003csup\u003e\u003cem\u003e+/+\u003c/em\u003e\u003c/sup\u003e\u003cem\u003eSnap25\u003c/em\u003e\u003csup\u003e\u003cem\u003e+/\u0026minus;\u003c/em\u003e\u003c/sup\u003e), \u003cem\u003eStxbp1\u003c/em\u003e single mutant (\u003cem\u003eStxbp1\u003c/em\u003e\u003csup\u003e\u003cem\u003e+/\u0026minus;\u003c/em\u003e\u003c/sup\u003e\u003cem\u003eSnap25\u003c/em\u003e\u003csup\u003e\u003cem\u003e+/+\u003c/em\u003e\u003c/sup\u003e) and double mutants (\u003cem\u003eStxbp1\u003c/em\u003e\u003csup\u003e\u003cem\u003e+/\u0026minus;\u003c/em\u003e\u003c/sup\u003e\u003cem\u003eSnap25\u003c/em\u003e\u003csup\u003e\u003cem\u003e+/\u0026minus;\u003c/em\u003e\u003c/sup\u003e) mice were generated by mating male congenic C57BL6/J \u003cem\u003eStxbp1\u003c/em\u003e\u003csup\u003e\u003cem\u003e+/\u0026minus;\u003c/em\u003e\u003c/sup\u003e mice(\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e, \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e) with female \u003cem\u003eSnap25\u003c/em\u003e\u003csup\u003e\u003cem\u003e+/\u0026minus;\u003c/em\u003e\u003c/sup\u003e mice also bred on C57BL6/J genetic background, backcrossed for 20 generations(\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e). All animals were kept in standard husbandry conditions on a 12 h light-dark cycle with food and water available \u003cem\u003ead libitum\u003c/em\u003e. Animals aged around 8 weeks were separately housed on sawdust in the standard Makrolon type II cages. All experiments were approved by the local animal research committee and complied with the European Council Directive(86/609/EEC). In total 200 mice were used.\u003c/p\u003e\n\u003ch3\u003eBehavioral phenotyping\u003c/h3\u003e\n\u003cp\u003eBehavioral experiments were performed on male mice between 8 and 16 weeks of age. In total three batches of \u003cem\u003eStxbp1Snap25\u003c/em\u003e mice were used: batch 1 and 2 each consisted of 6 controls and 6 double mutants; batch 3 consisted of 8 controls, 8 \u003cem\u003eStxbp1\u003c/em\u003e single-, 8 \u003cem\u003eSnap25\u003c/em\u003e single- and 8 double- mutant mice. In all three batches, we assessed spontaneous behavior in an automated home-cage environment (PhenoTyper model 3000, Noldus Information Technology, Wageningen, The Netherlands). Spontaneous behavior was automatically monitored for two and a half days in the PhenoTyper. Spontaneous home-cage behavior is highly dimensional aspect of behavior leading to separation of 90 separate parameters divided into 6 different categories(\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e): kinematics, activity, sheltering, habituation, dark-light index and light-dark phase transition. Kinematic parameters describe specific elements of animal\u0026rsquo;s behavior related to movement characteristics, such as short and long movement segments and short and long arrest segments. Activity bouts describe mouse behavior on sub-minute scale and cumulatively during the period of days of spontaneous behavior monitoring. Sheltering behavior describes the tendency of mice to sleep inside the shelter(\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e, \u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e). Habituation ratio evaluates changes in activity over days by taking the ratio for respective parameters of dark phase 3 over dark phase 1. Dark-light indexes assess difference in the behavior during the dark and light phase of cycle, by taking the ratio for respective parameters of dark phase over the light phase. The effect of light/dark phase transition on spontaneous behavior was assessed by analysis of change in activity parameters during the periods surrounding phase transition.\u003c/p\u003e \u003cp\u003eThe observation of spontaneous behavior was followed by the assessment of discrimination- and reversal learning using the CognitionWall task in the same automated home-cage(\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e). After testing in the PhenoTyper, animals acclimated to the new housing for one week before further testing.\u003c/p\u003e \u003cp\u003eThe batch 1 and batch 2 consisted of 12 control- and 12 double mutant- mice. We used standard behavioral battery for assessment of vision, muscle strength, motor coordination, anxiety, learning and memory, as previously described(\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e). All tests were performed during the light phase with the least stressful tests done at the beginning, and at least 1 day apart. Briefly, the vision was assessed using the vision test; muscle strength was assessed measuring grip strength; motor coordination and motor learning were assessed on the rotarod. We used three anxiety-related paradigms: elevated plus maze test, open field and dark-light box. Fear-conditioning experiment evaluated associative learning and memory in mice. With Barnes maze test we tested spatial learning, memory and reversal learning. Short-term (working) memory was measured using T-maze spontaneous alteration task. Detailed protocols for these behavioral tests can be found in (\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e) and in Supplementary Material and Methods.\u003c/p\u003e \u003cp\u003e The third batch of haploinsufficiency mutant mice was tested for spontaneous behavior in the Phenotyper followed by the fear conditioning and Barnes maze, since results from the first two tested batches of animals showed most phenotypical changes in these two paradigms. Detailed protocols for these behavioral tests can be found in(\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e) and in Supplementary Material and Methods.\u003c/p\u003e\n\u003ch3\u003eVideo monitoring, simultaneous radiotelemetric video, ECoG recordings\u003c/h3\u003e\n\u003cp\u003eVideo monitoring of mice from batch 4, age: between 8 and 10 weeks old, have been performed in their home cages (2 control, 4 \u003cem\u003eSnap25 single-\u003c/em\u003e, 6 \u003cem\u003eStxbp1\u003c/em\u003e single-, and 6 double mutants). Mice were video monitored for at least 24 h. After video monitoring, seven mice from batch 4 and 13 mice from batches 1\u0026ndash;3 were implanted with ECoG transmitters. In total: 4 control, 5 \u003cem\u003eStxbp1 single-\u003c/em\u003e, 4 \u003cem\u003eSnap25\u003c/em\u003e single-, and 7 double- mutants were implanted with ECoG transmitters (ETA-F10; specification: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.datasci.com/products/implantabletelemetry/mouse-(miniature)/eta-f10\u003c/span\u003e\u003cspan address=\"https://www.datasci.com/products/implantabletelemetry/mouse-(miniature)/eta-f10\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) as previously described in Kovačević \u003cem\u003eet al.\u003c/em\u003e(\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e). Briefly, mice were anesthetized with isoflurane (3% isoflurane/oxygene, flow 0.8 l/min) and immobilized in the stereotaxic instrument. After administration of lidocaine (2%, \u003cem\u003es.c.\u003c/em\u003e), a small incision was made on the skull allowing the recording electrode to be positioned above the motor cortex (2.2 mm anterior, 1 mm lateral) and ground electrode above the cerebellum (6 mm posterior, 1 mm lateral) using the stainless screws. The transmitter was placed in the abdominal subcutaneous pocket. The incision was closed with suture material. All animals received pre- and post-operative analgesic treatment with buprenorphine (0.05 mg/kg, \u003cem\u003es.c.\u003c/em\u003e). Animals were daily checked during recovery period of at least 7 days. During ECoG recordings (for at least 24 h), implanted animals were placed on the DSI receiver board (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ewww.datasci.com\u003c/span\u003e\u003cspan address=\"http://www.datasci.com\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) in front of an infrared camera allowing synchronous recording of behavior and ECoG signals.\u003c/p\u003e \u003cp\u003eVideo and ECoG data were time matched at the beginning and at the end of recording. ECoG data were visually inspected and analyzed using the Event Classifier application within Neuroarchiver tool in LWDAQ software (Open Source Instruments, Inc.). The application classified the 1s-segments of EEG according to several metrics (asymmetry, intermittency, coherence, power, coastline and spikiness) enabling that similar patterns cluster together(\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e).Video analysis was performed independently of ECoG data analysis and later matched with ECoG data. Sleep states were identified from the video recordings by researcher. ECoG signal that corresponds to the sleep state were selected and divided in one-minute epochs. Epochs were visually checked and those that contained artifacts were excluded from analysis. Spectral power analysis during the sleep episodes was performed using open-source software package Chronux (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://chronux.org/\u003c/span\u003e\u003cspan address=\"http://chronux.org/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) for signal processing of neural time-series data(\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e). We used routine mtspectrumc which is applicable for analysis of continuous-valued data using the moving-window with 2s time-width and 1s step-size. Data were sampled at 1kHz and high-pass filtered at 1Hz and low-pass filtered at 200Hz. Average spectral power was calculated as a mean value of relative power (expressed as the ratio of total power) for all 1-min episodes per animal. Power per frequency band was calculated as the average relative power within the frequency ranges: 1-4Hz for delta band, 5-8Hz for theta band, 9\u0026ndash;12 Hz for alfa band. Spectrogram were made using our own written script in Matlab.\u003c/p\u003e\n\u003ch3\u003ec-Fos staining\u003c/h3\u003e\n\u003cp\u003eMice, around 3 months old, were sacrificed either by cervical dislocation or overdose of avertin (2,2,2-tribromoethyl alcohol). Animals sacrificed by avertin were perfused with 4% paraformaldehyde (PFA) in 0.1M phosphate buffered saline (PBS.) In the present study, 5 \u003cem\u003eSnap25\u003c/em\u003e single mutants, 10 double mutants and 6 controls were used for c-Fos analyses. In addition, we re-analyzed data from 5 \u003cem\u003eStxbp1\u003c/em\u003e single mutants obtained in our laboratory and compared these to the controls from the same experiment. All brains were removed and post-fixed in 4% PFA in 0.1M PBS. After overnight cryoprotection in 30% sucrose solution in 0.1M PBS, brains were blocked in the coronal plane, frozen on dry ice and sectioned at 50 \u0026micro;m on a cryostat. To reveal c-Fos expression levels in the brain, free-floating sections were incubated in 0.3% H2O2 in 0.1M PBS for 10 min. After three rinses in 0.1M PBS, sections were incubated in 0.1M PBS containing 5% normal goat serum, 0.25% TritonX-100 and a c-Fos antibody raised in rabbit (Santa Cruz, sc-52; 1:800/1:500) or c-Fos antibody raised in rat (SySy, 226017; 1:1000) and left for overnight incubation (up to 96 hours) at 4\u0026deg;C. Sections were washed with 1xPBS and incubated at RT for 1 h in biotinylated goat anti-rabbit (#65-6140, Invitrogen; 1:400) and anti-rat-Ab secondary antibodies (# 31830, ThermoFisher Scientific, 1:400). Sections were washed three times with 0.1M PBS and incubated at room temperature (RT) for one and a half of hour in biotinylated goat anti-rabbit (#65-6140, Invitrogen; 1:400) or anti-rat-Ab secondary antibodies (# 31830, ThermoFisher Scientific, 1:400). After three rinses with 0.1M PBS, the sections were incubated at RT for 1,5 h in avidin-biotin peroxidase complex (Vectastain ABC, Vector Laboratories; 1:800). To visualize the peroxidase labeling, sections were processed with a DAB/nickel substrate working solution (DAB Peroxidase Substrate, SK-4100; Vector Laboratories) for 7 min at RT. After rinsing with 0.1M PBS, sections were mounted on gelatin-coated slides, dehydrated, and put on coverslips.Sections were imaged using a Leica bright-field microscope at 5x and 10x magnification. Several brain regions were selected for analysis: prefrontal cortex (PFC), primary motor cortex and somatosensory cortex, hippocampus (CA1, CA3, and dentate gyrus (DG)) and thalamus (Thal). Each brain region of interest was identified using a standard mouse brain atlas (Paxinos and Franklin). c-Fos immunoreactive nuclei were counted using predefined threshold values in ImageJ software. Labelled cells were counted bilaterally, averaged and normalized to the size of area and expressed as a relative compared to appropriate control.\u003c/p\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003eCalcium imaging in developing brain slices\u003c/h2\u003e \u003cp\u003eCalcium imaging of spontaneous synchronous activity (SSA) in the developing PFC- brain slices was performed according to previously published protocol by Dawitz \u003cem\u003eet al.\u003c/em\u003e(\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e) and Pires \u003cem\u003eet al.\u003c/em\u003e(\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e). Briefly this method contains several steps: slice preparation, dye-loading slices, imaging and analysis. In total for the calcium imaging experiment, 9 controls, 8 \u003cem\u003eStxbp1\u003c/em\u003e single-, 9 \u003cem\u003eSnap25\u003c/em\u003e single- and 8 double- mutants were used. Animals (P14 \u0026plusmn; 1) were decapitated and brains were rapidly removed and placed in cold oxygenated aCSF (artificial Cerebrospinal Fluid)(\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e, \u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e). PFC- brain slices were cut (300 \u0026micro;m thick) using Microm HM 650V and transferred into a slice holder containing oxygenated recovery-aCSF (rACSF)(\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e). After 1h of recovery period, slices were transferred to a staining-chamber filled with 1 ml experimental ACSF (eACSF) and heated to 34\u0026deg;C. Slices were incubated in eACSF containing Fura2-AM dye (ThermoFisher scientific, F1201) between 20 and 45 minutes depending on the age of the animal. After incubation, slices were transferred to coated recording chambers and approximately 1 ml of eACSF was added. Recording chambers with slices were kept in a large humidified interface chamber to recover for at least one hour before recording. Network imaging was performed on a two-photon laser-scanning microscope (Trimscope LaVision Biotec). Slices were heated to 37\u0026deg;C and constantly perfused with oxygenated eACSF. Using a Hamamatsu C9100 EM-CCD camera as a detector, two time-lapse movies (4000 frames each) in PFC-ROIs were acquired with a sampling frequency of 7.65Hz (450\u0026micro;mX525\u0026micro;m, binning 2x2). The first was to evaluate baseline network activity and the second the effect of incubation with Gabazine (10 \u0026micro;M, Hellobio, SR95531) for 10 minutes. For the detection of the somas, a z-stack\u0026thinsp;\u0026plusmn;\u0026thinsp;20\u0026micro;m around the central plane with step size of 1\u0026micro;m thickness was acquired after each recording.\u003c/p\u003e \u003cp\u003eTo analyze calcium-imaging data, custom-built Matlab\u0026reg; (Mathworks) scripts were used (EvA methodology)(\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e). Neurons were semi-automatically detected using the z-stacks of the network being imaged. Finally, the events in individual traces were then analyzed individually and on a network level and three different categories of neurons were derived: silent neurons (those neurons where no activity was detected), active neurons and synchronized neurons (active neurons whose activity is synchronized with other neurons in network)(\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e). Several different parameters were extracted and statistically analyzed: percentage of active cells, frequency of active cells, percentage of synchronously active cells and the frequency of synchronously active cells.\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eCell culture and electrophysiology\u003c/h3\u003e\n\u003cp\u003eHippocampal microisland cultures were prepared from controls, \u003cem\u003eStxbp1\u003c/em\u003e single-, \u003cem\u003eSnap25\u003c/em\u003e single-and double mutants at embryonic day 18. Briefly, hippocampi were isolated, collected in ice-cold Hank\u0026rsquo;s balanced salt solution (HBSS; Sigma, H9394) supplemented with 10mM HEPES and digested with 0.25% trypsin (Life Technologies, 15090-046) at 37\u0026deg;C for 20 min. After trituration, cells were resuspended in Neurobasal medium supplemented with B-27, 1% HEPES, 0.25% GlutaMAX, and 0.1% Penicillin-Streptomycin (all Invitrogen) and plated in a 6-well plate at a density of 7000 cells/well on top of pre-grown mouse glia islands on 30-mm coverslips (first experiment, see below) or in a 12-well plate at a density of 2500 cells/well on top of pre-grown rat glia islands on 18mm coverslips (second experiment) to achieve autaptic cultures.\u003c/p\u003e \u003cp\u003eTwo independent electrophysiological experiments were performed. In the first experiment, electrophysiological measurements were performed in autaptic cultures grown for 10\u0026ndash;14 days \u003cem\u003ein vitro\u003c/em\u003e as previously described in Ruiter \u003cem\u003eet al.\u003c/em\u003e(\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e). The recordings were performed at room temperature (RT) using an EPC10 amplifier (HEKA) with the recording program Patchmaster v2x73.5 (HEKA). Traces were filtered with a 3kHz Bessel low-pass filter and data were acquired at 20 kHz. The series resistance was compensated to 70%. Borosilicate glass pipettes (resistance between 3 to 5 MΩ) were filled with 136 mM KCl, 17.8 mM HEPES, 1 mM EGTA, 4 mM ATP-Na, 4.6 mM MgCl2, 15 mM Creatine Phosphate and 50 U/mL phosphocreatine kinase (pH 7.4). External solution contained the following (in mM): 10 HEPES, 10 glucose, 140 NaCl, 2.4 KCl, 2 MgCl\u003csub\u003e2\u003c/sub\u003e and 2 CaCl\u003csub\u003e2\u003c/sub\u003e (pH\u0026thinsp;=\u0026thinsp;7.4, 300 mOsmol). Recordings were done in voltage clamp, with the holding potential kept at -70 mV. Evoked excitatory postsynaptic currents (EPSC) were induced by raising the holding voltage to 0 mV for 2 ms. The size of the readily releasable pool (RRP) was assessed by hypertonic sucrose application (500 mM, 3.5 s). Sucrose was dissolved into the external solution. Application to the cells was done using a custom-made barrel system, controlled by SF-77B perfusion fast step (Warner Instruments) via digital output switches from the EPC10.\u003c/p\u003e \u003cp\u003eIn the second experiment, electrophysiological measurements were performed in autaptic cultures grown for 13\u0026ndash;16 days \u003cem\u003ein vitro\u003c/em\u003e as explained in Meijer \u003cem\u003eet al.\u003c/em\u003e(\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e). Whole-cell voltage clamp recordings (Vm = -70 mV) were performed at RT using an Axopatch 200B amplifier, and Digidata 1440A and Clampex 10 for signal acquisition (Molecular Devices). Action potentials were induced by stepping to +\u0026thinsp;30 mV for 1ms. Data were acquired at 10 kHz with 2 kHz filtering with a low-pass Bessel filter. Series resistance was compensated by 70\u0026ndash;80% (bandwidth 7.52 Hz). Borosilicate glass pipettes (resistance of 2-4.5 MΩ) were filled with (in mM): 136 KCl, 17.8 HEPES, 1 EGTA, 0.6 MgCl\u003csub\u003e2\u003c/sub\u003ex6H\u003csub\u003e2\u003c/sub\u003eO, 4 ATP-Mg, 0.3 GTP-Na, 12 phosphocreatine di-potassium-salt and 50 U/mL phosphocreatine kinase. External solution contained the following (in mM): 10 HEPES, 10 glucose, 140 NaCl, 2.4 KCl, 4 MgCl\u003csub\u003e2\u003c/sub\u003e, and 2 CaCl\u003csub\u003e2\u003c/sub\u003e (pH\u0026thinsp;=\u0026thinsp;7.30, 300 mOsmol). Traces were excluded when series resistance was higher than 15 MΩ or leak current was larger than 300pA. GABAergic traces were recognized by prolonged EPSC decays and excluded. Analysis was performed in MATLAB R2019a (Mathworks) using available scripts (viewEPSC, downloaded from Github user vhuson). The RRP was estimated by back extrapolation from the cumulative EPSC charge during a 40 Hz 100 action potential train as in Meijer \u003cem\u003eet al.\u003c/em\u003e(\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e).\u003c/p\u003e\n\u003ch3\u003eImmunocytochemistry\u003c/h3\u003e\n\u003cp\u003eNeurons were fixed on DIV10-DIV14. Coverslips were fixed at RT with 2% PFA diluted in culture medium (Neurobasal with 2% ml B-27, 1 M HEPES, 0.26% Glutamax, 14.3 mM β-mercaptoethanol, 20 U/ml penicillin, 20 \u0026micro;g/ml streptomycin) for 10 min followed by another fixation with 4% PFA diluted in PBS for 10min. After washing with PBS, the neurons were permeabilized for 5 min at RT with a solution of 0.5% Triton X-100 diluted in PBS. The neurons were then blocked with a solution containing 4% goat serum and 0.1% Triton X-100 for 30 min at RT. The coverslips were then incubated for 2 hours at RT with a blocking solution containing α-VGLUT antibody (1/1000; Guinea pig, Millipore AB5905) and α-MAP2 antibody (1/5000; chicken, Abcam Ab5392) or a solution containing α-VGAT antibody (1/500; Rabbit, Synaptic System 131002) and α-MAP2 antibody (1/200; chicken, Abcam Ab5392). Cells were washed with PBS, and then transferred for 1 hour at RT to a blocking solution containing the secondary antibodies Alexa Fluor-568 α-Chicken (1/1000; goat, Invitrogen A11041) and Alexa fluor-647 α-Guinea pig (1/4000; goat, Invitrogen A21450) or a solution containing the secondary antibodies Alexa Fluor-568 α-Chicken (1/1000; goat, Invitrogen A11041) and Alexa fluor-488 α-Rabbit (1/5000; goat, Invitrogen A11008). Cells were washed with PBS before being mounted on glass slides with FluorSave (Calbiochem, 345789). Coverslips were examined on Leica SP8 inverted confocal microscope with 40x oil immersion Objective (NA\u0026thinsp;=\u0026thinsp;1.4).\u003c/p\u003e \u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003eSurvival rate assessment\u003c/h2\u003e \u003cp\u003eThe assessment of survival rate per genotype was performed in twelve E18-litters bred for electrophysiological experiments (caesarian section) and ten nests bred for behavioral experiments (natural birth). All pups from E18-nests were collected and genotyped before culturing and the results analyzed for the expected genotype ratio. In addition, genotyping results of ten nests bred for behavioral experiments were performed after weaning, at 3\u0026ndash;4 weeks of age and the survival rate of animals was compared to those at E18. Following behavioral and/or video monitoring experiments, genotypes were conformed at the time of sacrifice.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec12\" class=\"Section2\"\u003e \u003ch2\u003eStatistics\u003c/h2\u003e \u003cp\u003eAll statistical analyses were performed using IBM SPSS statistic 24 (IBM corporation, Armonk, NY, USA) and GraphPad Prism\u0026reg;, version 8.3.1. If the normality and homoscedasticity criteria were met, data were analyzed using parametric tests (t-test, ANOVA, repeated measure ANOVA). If the normality and homoscedasticity criteria were not met, nonparametric tests were performed (Mann Whitney U-test) or data were ln-transformed. Preliminary analysis of electrophysiological data was performed to assess the variability between different preparation (weeks of experiment) or different batches by applying nested- and two-way- ANOVA. The analysis revealed high variability between preparations (weeks of experiment). Thus, to reduce the variability between preparations, data obtained from electrophysiological experiment were normalized to the average of control group for the appropriate week. If normalized data were not normally distributed, ln-transformation was applied. There were no significant differences of behavioral measures between batches. Therefore, data from different batches were analyzed together. Outliers were removed from analysis using ROUT method, with Q set to 1%. Tukey\u0026rsquo;s posthoc tests were performed if ANOVA showed significant effect. The genotype effect in CognitionWall DL/RL task was assessed by performing log-rank test on two Kaplan Meier survival curves. An error probability lower than p\u0026thinsp;\u0026lt;\u0026thinsp;0.05 was accepted as statistically significant through the study. For all given level of analysis of PhenoTyper spontaneous behavior data, statistical analysis was based on estimated false discovery rate (FDR), P-values were corrected by minimum positive FDR with a threshold set at 5%.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec13\" class=\"Section2\"\u003e \u003ch2\u003eData availability\u003c/h2\u003e \u003cp\u003eData are available: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://link.will.be.added\u003c/span\u003e\u003cspan address=\"http://link.will.be.added\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (SciStore server @VU University)\u003c/p\u003e \u003c/div\u003e"},{"header":"Results","content":"\u003cdiv id=\"Sec15\" class=\"Section2\"\u003e \u003ch2\u003eNon-linear gene-gene interactions can explain phenotypic diversity for DEEs\u003c/h2\u003e \u003cp\u003eWe defined three modes of gene-gene interaction to describe how phenotypic differences observed among individuals with a pathogenic variant in one gene depends on their genotypes at other loci. First, we assume that if two genes encode proteins that operate in the same pathway, a variant/mutant has a unique phenotypic spectrum characterized by a mean effect and a variance determined by environmental factors as well as background genetic diversity (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eA). If both variants affect the same pathway and are expressed in the same individual this would result in a phenotypic spectrum of the strongest single case, but not more severe or diverse than that (lack of additivity, Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e, model B.1). In contrast, if two genetic variants from independent pathways lead to common disorders, the combination of variants results in summation of the phenotypic spectra (additive model, Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e, model B.2.1 (\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e)). Finally, in cases where genes interact genetically or physically, genetic variation in two such genes may give rise to a multiplicative effect on the phenotypic spectra (\u0026lsquo;epistasis\u0026rsquo; or \u0026lsquo;super-additivity\u0026rsquo; Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e, model B.2.2, (\u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e)). The last two scenarios can be distinguished by the different distributions and variances of their effects (phenotypes): if the effects of single gene variants follow a normal distribution (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eA), the additive model predicts that the effect of the combination of variants also follows a normal distribution, with a mean being the sum of the means of single variants and the variance being the sum of the variances of single variants. However, according to the multiplicative model, the distribution of the effect of the combination of two normally distributed variants is right-skewed, it approaches a lognormal distribution and has markedly larger variance compared to the additive model (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). As a result of the increased variance, multiplicative interaction results both in more individuals with severe phenotypes, but also more individuals with mild phenotypes, compared to additive interactions (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). Notably, this conclusion follows from simply combining the background variances (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eA) in a multiplicative model, without introducing new sources of variance. Whether genetic variants act in the same pathogenic pathway, or via independent or interacting pathways can be experimentally tested in animal models bearing variants in two genes. The models in Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e predict that the large phenotypic diversity in DEE-patients can be explained by multiplicative interactions between genes.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec16\" class=\"Section2\"\u003e \u003ch2\u003eStxbp1/Snap25 double mutants show extremely diverse seizure phenotypes\u003c/h2\u003e \u003cp\u003eTwo presynaptic DEE genes, \u003cem\u003eSTXBP1\u003c/em\u003e and \u003cem\u003eSNAP25\u003c/em\u003e, were selected to test the consequences of gene-gene interactions, given their close functional relationship in SNARE-complex assembly. We generated three haploinsufficiency mouse models by heterozygous inactivation of 1) single \u003cem\u003eStxbp1\u003c/em\u003e (\u0026lsquo;\u003cem\u003eStxbp1\u003c/em\u003e single mutants\u0026rsquo;), 2) \u003cem\u003eSnap25\u003c/em\u003e (\u0026lsquo;\u003cem\u003eSnap25\u003c/em\u003e single mutants\u0026rsquo;) gene and 3) combined haploinsufficient inactivation of \u003cem\u003eStxbp1\u003c/em\u003e and \u003cem\u003eSnap25\u003c/em\u003e (\u0026lsquo;double mutants\u0026rsquo;) mice (\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e, \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e). Array-based genetic analysis of 11,000 SNP probes (miniMUGA, (\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e) confirmed a homogeneous genomic background for all experimental groups (96.8% of SNPs consistent with C57BL/6J sub-strain, Supplementary Table\u0026nbsp;3 and Supplementary Fig.\u0026nbsp;1), with small contributions of 0.1% and 0.4% of flanking regions of 129 strain background around the deletion site, for the \u003cem\u003eStxbp1\u003c/em\u003e and \u003cem\u003eSnap25\u003c/em\u003e locus, respectively, as reported before for \u003cem\u003eStxbp1\u003c/em\u003e single mutants.(\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eIn E18 embryos, obtained by caesarian section for synapse physiology experiments (see below), all four genotypes were obtained, with a trend towards a reduced fraction for double mutants (18% vs expected Mendelian ratio of 25%, p\u0026thinsp;=\u0026thinsp;0.093, χ\u003csup\u003e2\u003c/sup\u003e, labelled A in Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eA). At the age of three/four weeks, when naturally born animals bred for behavioral experiments were weaned and genotyped, the distribution of genotypes significantly deviated from expected, with double mutants only 12% of animals, approximately half of the expected number (p\u0026thinsp;=\u0026thinsp;0.041, χ\u003csup\u003e2\u003c/sup\u003e, Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eA). Concomitantly, animal care takers reported lethal generalized seizures prior to and during weaning in these nests (Suppl video 1). Following the three/four weeks genotyping, sudden unexpected death of epilepsy (SUDEP) was detected in two additional double mutants older than 6 weeks. Together, this led to an even more significant deviation of the expected number of double mutants (10% instead of 25%, p\u0026thinsp;=\u0026thinsp;0.031, labelled B in Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eA). Further breeding of these double mutants was restricted by the local ethical committee regulations. Strikingly, surviving double mutants had no reported seizures, normal vision, normal muscle strength, normal motor coordination and ability to acquire motor skills, but 16% lower body weight compared to their controls (Supplementary Fig.\u0026nbsp;2 and Supplementary Table\u0026nbsp;1). These data suggest that at least half of double mutants died before the age of eight weeks due to severe seizures, while the surviving double mutants had no apparent phenotypes, except a lower body weight.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eVideo observations of mice in their home cage confirmed frequent generalized and clonic seizures in some double mutant mice, but not others (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eB). The incidence of seizures varied substantially among double mutants (compare the two double mutant individuals in Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eB). \u003cem\u003eStxbp1\u003c/em\u003e single mutants and double mutants showed twitches (n\u0026thinsp;=\u0026thinsp;20.2\u0026thinsp;\u0026plusmn;\u0026thinsp;2.6 / 12h) and jumps (n\u0026thinsp;=\u0026thinsp;7.4\u0026thinsp;\u0026plusmn;\u0026thinsp;1.3 / 12h, Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eC), as reported earlier for \u003cem\u003eStxbp1\u003c/em\u003e single mutants (\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e, \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e). The incidence of these two types of behaviors was lower in double mutants (n\u0026thinsp;=\u0026thinsp;4.5\u0026thinsp;\u0026plusmn;\u0026thinsp;1.9 and n\u0026thinsp;=\u0026thinsp;1.0\u0026thinsp;+\u0026thinsp;0.4 for twitches and jumps per 12 h, respectively, Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eC). Except for these abnormalities, overall development of single \u003cem\u003eStxbp1\u003c/em\u003e and \u003cem\u003eSnap25\u003c/em\u003e mice was normal, as reported before (\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e, \u003cspan additionalcitationids=\"CR40\" citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e41\u003c/span\u003e). Taken together, the reduced incidence of twitches and jumps and the occurrence of clonic, lethal seizures in a subset of double mutant mice suggests large phenotypic diversity regarding behavioral manifestation of epilepsy in double mutant mice.\u003c/p\u003e \u003cp\u003eNext, we combined video with electrocorticography (ECoG) monitoring at six to twelve weeks of age using implanted electrodes in 19 mice from new litters. These recordings confirmed generalized seizures in a subset of double mutant mice, but not others (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eD-E). The generalized seizures began with clusters of spike-slow waves (interictal spikes) accompanied with clonic seizures; the seizure progressed to a full generalized seizure and stopped with the postictal suppression and behavioral immobility (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eE). Strikingly, clonic and generalized seizures were observed only in a subset of double mutant mice, while other double mutant mice showed no generalized seizures during 24 hours of recording (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eD). Hence, double mutant mice show extreme phenotype diversity in behavioral and electrographic abnormalities related to seizures.\u003c/p\u003e \u003cp\u003eAnalysis of ECoG recordings revealed a collection of diverse patterns of epileptiform activity in several experimental groups: slow-wave discharges (SWDs), sharp spikes and spike-slow waves (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eG-J). SWDs were observed before in \u003cem\u003eStxbp1\u003c/em\u003e single mutants (\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e) and typically accompanied by behavioral twitches (\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e) (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eB). The number of SWDs was higher in \u003cem\u003eStxbp1\u003c/em\u003e single mutants compared to control, \u003cem\u003eSnap25\u003c/em\u003e single mutants and double mutants (F(\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e, \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e)\u0026thinsp;=\u0026thinsp;4.501, p\u0026thinsp;=\u0026thinsp;0.019, \u003cem\u003epost hoc\u003c/em\u003e: p\u0026thinsp;=\u0026thinsp;0.003, p\u0026thinsp;=\u0026thinsp;0.015 and p\u0026thinsp;=\u0026thinsp;0.039, respectively, Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eH). The lower number of SWDs in double mutant mice compared to single \u003cem\u003eStxbp1\u003c/em\u003e mice confirms the previous conclusion that a higher phenotypic diversity exists for epilepsy-related phenotypes in double mutant mice compared to single mutant and control mice.\u003c/p\u003e \u003cp\u003eThe number of spike slow waves was significantly higher in double mutant mice compared to control mice, \u003cem\u003eSnap25\u003c/em\u003e single and \u003cem\u003eStxbp1\u003c/em\u003e single mutants (F(\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e, \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e)\u0026thinsp;=\u0026thinsp;4.913, p\u0026thinsp;=\u0026thinsp;0.014, \u003cem\u003epost hoc\u003c/em\u003e: p\u0026thinsp;=\u0026thinsp;0.004, p\u0026thinsp;=\u0026thinsp;0.020 and p\u0026thinsp;=\u0026thinsp;0.026, respectively, Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eI). When the spike-slow waves occurred during the awake state, they were accompanied by clonic seizures in double mutant mice. Sharp spikes were observed mainly during sleep and the number of sharp spikes did not differ between genotypes (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eJ). Hence, more severe electrographic seizures were observed in double mutant mice compared to \u003cem\u003eStxbp1\u003c/em\u003e single- and \u003cem\u003eSnap25\u003c/em\u003e single- mutants. Taken together, the diversity in epileptic phenotype observed in double mutant mice is characterized by higher probability of low incidence SWDs but also with exacerbated seizure activity represented with increased lethality, generalized seizures and increased number of spike-slow wave epileptiform discharges in a subset of double mutants, while others showed mild/no phenotypes. According to our proposed model of gene interaction (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e), the observed phenotypic diversity of double mutant mice is in line with right-skewed distribution and larger variance predicted by the multiplicative model of gene interactions (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eF).\u003c/p\u003e \u003cp\u003eFinally, power spectral analysis during the sleep episodes (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eK) revealed a shift in relative power towards higher frequency bands for all three genotypes, represented as a decrease in the relative power in the delta band (F(\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e, \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e)\u0026thinsp;=\u0026thinsp;3.846, p\u0026thinsp;=\u0026thinsp;0.113) and an increase in the relative power in the alfa band (F(\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e, \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e)\u0026thinsp;=\u0026thinsp;26.72, p\u0026thinsp;=\u0026thinsp;0.0042) for all three mutant groups, independent on the genotype (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eK - M). According to our proposed model (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e), this finding suggests that a common pathway underlies shifts in spectral power in single \u003cem\u003eStxbp1-\u003c/em\u003e, single \u003cem\u003eSnap25-\u003c/em\u003e and double mutant mice during sleep.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec17\" class=\"Section2\"\u003e \u003ch2\u003ec-Fos expression is increased in the thalamus of surviving double mutants\u003c/h2\u003e \u003cp\u003eTo corroborate the observed seizure activity, c-Fos immunoreactivity was used as a marker for excessive neuronal activity in the brain (\u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e42\u003c/span\u003e), Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e. Increased c-Fos expression was observed in cortical brain regions of \u003cem\u003eStxbp1\u003c/em\u003e single mutants as shown before (\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e), in the surviving double mutants (p\u0026thinsp;=\u0026thinsp;0.026) and a strong trend was observed in \u003cem\u003eSnap25\u003c/em\u003e single mutants (p\u0026thinsp;=\u0026thinsp;0.058) (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eA-B and \u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eE). Furthermore, only in double mutants, c-Fos expression was significantly increased in the thalamus (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eC and \u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eE; p\u0026thinsp;=\u0026thinsp;0.026) and a strong trend was observed in hippocampal regions (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eD and \u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eE; p\u0026thinsp;=\u0026thinsp;0.058). Thus, c-Fos expression data confirms genetic interaction between the two genotypes in thalamus and hippocampus of the double mutants, but not in cortical regions, possibly because double mutants with most excessive cortical excitability had already died (see above) prior to c-Fos expression analysis.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003cb\u003eSurviving double mutants show impaired cognition, like\u003c/b\u003e \u003cb\u003eStxbp1\u003c/b\u003e \u003cb\u003emice\u003c/b\u003e\u003c/p\u003e \u003cp\u003eDifferent aspects of learning and memory were assessed in surviving animals of the four experimental groups. To assess associative learning and memory, the fear conditioning test was performed (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eA). After one pairing session between shock and tone, a contextual memory was assessed by placing animals in the training context during the next day (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eB). \u003cem\u003eSnap25\u003c/em\u003e single mutants showed a similar percentage of freezing as controls (p\u0026thinsp;=\u0026thinsp;0.639). \u003cem\u003eStxbp1\u003c/em\u003e single mutants showed a significantly lower percentage of freezing in the training context compared to controls (p\u0026thinsp;=\u0026thinsp;0.001, Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eC), as shown before\u003csup\u003e13\u003c/sup\u003e and compared to \u003cem\u003eSnap25\u003c/em\u003e single mutants (p\u0026thinsp;=\u0026thinsp;0.010). Double mutants showed similar effects as \u003cem\u003eStxbp1\u003c/em\u003e single mutants (p\u0026thinsp;\u0026lt;\u0026thinsp;0.001 and p\u0026thinsp;=\u0026thinsp;0.007, respectively, Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eB-C). Supplementary Table\u0026nbsp;1 lists all statistical tests. Exposure of mice to the new context resulted in ~\u0026thinsp;10% of time freezing in control mice and \u003cem\u003eSnap25\u003c/em\u003e single mutants due to general fear (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eB). \u003cem\u003eStxbp1\u003c/em\u003e single mutants and double mutants showed significantly lower percentage of freezing compared to controls (p\u0026thinsp;=\u0026thinsp;0.014 and p\u0026thinsp;=\u0026thinsp;0.020, respectively, Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eC and Supplementary Table\u0026nbsp;1) and \u003cem\u003eSnap25\u003c/em\u003e single mutants (p\u0026thinsp;=\u0026thinsp;0.048 and p\u0026thinsp;=\u0026thinsp;0.060, respectively, Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eC). Cued memory was assessed by tone exposure of animals in a new context Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eB. \u003cem\u003eStxbp1\u003c/em\u003e single mutants and double mutants showed a significantly lower percentage of freezing compared to controls (p\u0026thinsp;=\u0026thinsp;0.009 and p\u0026thinsp;=\u0026thinsp;0.013 Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eC and Supplementary Table\u0026nbsp;1) and \u003cem\u003eSnap25\u003c/em\u003e single mutants (p\u0026thinsp;=\u0026thinsp;0.044 and p\u0026thinsp;=\u0026thinsp;0.104, respectively, Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eC). Taken together, fear conditioning experiments suggest a strong and similar impairment in the contextual and cued fear memory in \u003cem\u003eStxbp1\u003c/em\u003e single mutants and surviving double mutants. No evidence was observed for stronger phenotypes in surviving double mutant mice than for single mutants.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eThe Barnes maze test was used to assess spatial learning, memory and reversal learning (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eD). During the acquisition phase, mice were trained to locate the escape hole and the time needed to escape the aversive environment was assessed. No significant effects of genotypes on the escape latency were observed, although latencies tended to be higher for \u003cem\u003eStxbp1\u003c/em\u003e single mutants and double mutants compared to controls (F(3,208)\u0026thinsp;=\u0026thinsp;2.613, p\u0026thinsp;=\u0026thinsp;0.061, Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eE and Supplementary Table\u0026nbsp;1). During the probe trial, to assess spatial memory one day after the last acquisition training, \u003cem\u003eStxbp1\u003c/em\u003e single- and double- mutants showed a higher probability of hole visits in the target octant compared to controls and \u003cem\u003eSnap25\u003c/em\u003e single mutants (p\u0026thinsp;=\u0026thinsp;0.042 and p\u0026thinsp;=\u0026thinsp;0.033, respectively, Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eF and Supplementary Table\u0026nbsp;1). Impaired behavioral flexibility was already described in \u003cem\u003eStxbp1\u003c/em\u003e single mutants(\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e) and a similar phenotype was observed in surviving double mutant mice.\u003c/p\u003e \u003cp\u003eDuring the reversal phase, the location of the escape hole was changed to the opposite side of the maze. Similar to the trend observed during the acquisition phase, double mutants needed more time to find the new escape hole compared to controls and \u003cem\u003eSnap25\u003c/em\u003e single mutants (p\u0026thinsp;=\u0026thinsp;0.036 and 0.011, respectively Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eG) and \u003cem\u003eStxbp1\u003c/em\u003e single mutants showed a trend towards longer escape latencies compared to controls and \u003cem\u003eSnap25\u003c/em\u003e single mutants (p\u0026thinsp;=\u0026thinsp;0.010 and 0.057, respectively Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eG). These data show that single \u003cem\u003eStxbp1\u003c/em\u003e and surviving double mutants preserved the learned response stronger than single \u003cem\u003eSnap25\u003c/em\u003e mutants and control mice, suggesting impaired behavioral flexibility and deficits in the reversal learning in single \u003cem\u003eStxbp1\u003c/em\u003e and double mutants. Again, no evidence was observed for stronger phenotypes in surviving double mutant mice compared to single mutants.\u003c/p\u003e \u003cp\u003eAttention and working memory were assessed using the spontaneous alteration task in the T maze (Supplementary Fig.\u0026nbsp;3A). This test is based on the natural tendency of mice to visit the previously not visited arm (\u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e43\u003c/span\u003e). Double mutants showed a similar percentage of alterations as their controls (t(\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e) = -0.963, p\u0026thinsp;=\u0026thinsp;0.346, Supplementary Fig.\u0026nbsp;3B and Supplementary Table\u0026nbsp;1), suggesting normal attention and hippocampal-dependent short-term memory.\u003c/p\u003e \u003cp\u003eTo assess discrimination and reversal learning, we performed a 4-day automated home-cage task, the CognitionWall test (\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e) Supplementary Fig.\u0026nbsp;3C and Supplementary Table\u0026nbsp;1. During the discrimination-learning phase (DL), animals should learn to earn food rewards by passing through the correct hole of the three holed CognitionWall placed inside the PhenoTyper. During the reversal-learning (RL) task animals should suppress previously learned response and learn that passing through the other hole is rewarded. During the discrimination and reversal learning tasks, all mice showed similar distribution of entries made to reach the criterion of 80% correct entries (p\u0026thinsp;=\u0026thinsp;0.264 and p\u0026thinsp;=\u0026thinsp;0.766, Supplementary Fig.\u0026nbsp;3D-G and Supplementary Table\u0026nbsp;1). Thus, the results from the Cognition Wall test suggest a normal discrimination and reversal learning in all four experimental groups.\u003c/p\u003e \u003cp\u003e \u003cb\u003eSurviving double mutants show anxiety-related behaviors like\u003c/b\u003e \u003cb\u003eStxbp1\u003c/b\u003e \u003cb\u003emice\u003c/b\u003e\u003c/p\u003e \u003cp\u003eThe anxiety-related phenotypes in double mutant mice were tested using a classical anxiety-related paradigm, the elevated plus maze test (EPM). In this test, double mutants showed anxiety-related behaviors represented by significantly less time spent on the open arms (t(\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e)\u0026thinsp;=\u0026thinsp;2.561, p\u0026thinsp;=\u0026thinsp;0.018) and a lower percentage of visits to the open arms (t(\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e)\u0026thinsp;=\u0026thinsp;3.216, p\u0026thinsp;=\u0026thinsp;0.004) accompanied with a mild, but significant increase of total distance moved compared to their controls (t(\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e) = -2.813, p\u0026thinsp;=\u0026thinsp;0.040), (Supplementary Fig.\u0026nbsp;4A-C and Supplementary Table\u0026nbsp;1). Double mutants did not show increased anxiety in the open field test and in the dark-light box test (Supplementary Fig.\u0026nbsp;4D - I and Supplementary Table\u0026nbsp;1). Taken together, these data show mild anxiety-related behavior accompanied with hyper-activity detected in the elevated plus maze test, comparable to \u003cem\u003eStxbp1\u003c/em\u003e single mutant mice.\u003c/p\u003e \u003cp\u003e \u003cb\u003eSurviving double mutants show spontaneous behaviors like\u003c/b\u003e \u003cb\u003eStxbp1\u003c/b\u003e \u003cb\u003emice\u003c/b\u003e\u003c/p\u003e \u003cp\u003eAnalysis of spontaneous behavior was performed with surviving animals in all experimental groups in the automated home-cage environment (PhenoTyper) enriched with a shelter (\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e). Double mutants showed several behaviors that were altered to a similar extent as the single \u003cem\u003eStxbp1\u003c/em\u003e mutants, especially in Kinematics (parameters 1\u0026ndash;26), but also a few abnormalities in other spontaneous behaviors (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eA-F, Supplementary Table\u0026nbsp;2). \u003cem\u003eSnap25\u003c/em\u003e single mutants showed very few significantly altered (kinematic) behaviors (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eA). Taken together, spontaneous behavior analysis indicates that surviving double mutants showed no evidence for super-additivity, but only phenotypes like the strongest single mutant (\u003cem\u003eStxbp1\u003c/em\u003e).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003cb\u003eGABA inhibition in developing cortex is abnormal in double mutants, like single\u003c/b\u003e \u003cb\u003eStxbp1\u003c/b\u003e \u003cb\u003emice\u003c/b\u003e\u003c/p\u003e \u003cp\u003eSpontaneous Synchronous Activity (SSA) is essential for the correct development of neural circuits(\u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e44\u003c/span\u003e). SSA has been recently characterized in the mPFC and it has been shown that GABA blockade at the end of the second postnatal week can partially restore the SSA, which is present at earlier times (\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e). We measured SSA in the developing prefrontal cortex of all experimental groups at 2 weeks of age, by monitoring Ca\u003csup\u003e2+\u003c/sup\u003e-transients using two-photon calcium imaging in acute brain slices. The role of GABA in SSA was assessed by adding gabazine at concentrations that block both the phasic and tonic activity of GABA\u003csub\u003eA\u003c/sub\u003e receptors (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eA-B). On average, approximately 40% of all neurons had spontaneous Ca\u003csup\u003e2+\u003c/sup\u003e-transients with an overall frequency of 0.0063 Hz in all groups, independent of the genotype (percentage of active cells: F(3,90)\u0026thinsp;=\u0026thinsp;1.630, p\u0026thinsp;=\u0026thinsp;0.188 and frequency of active cells: F(3,90)\u0026thinsp;=\u0026thinsp;1.577, p\u0026thinsp;=\u0026thinsp;0.200, Supplementary Fig.\u0026nbsp;5A-B). The majority of active cells was synchronously active; the percentage of synchronously active cells and the frequency of their activity did not differ between genotypes during baseline recording (F(3,90)\u0026thinsp;=\u0026thinsp;0.600, p\u0026thinsp;=\u0026thinsp;0.617 and F(3,90)\u0026thinsp;=\u0026thinsp;2.058, p\u0026thinsp;=\u0026thinsp;0.111, respectively, Supplementary Fig.\u0026nbsp;5A-B). Blockade of GABA\u003csub\u003eA\u003c/sub\u003e receptors by gabazine did not significantly affect the percentage of active cells in any of the experimental groups (F(1,90)\u0026thinsp;=\u0026thinsp;1.614, p\u0026thinsp;=\u0026thinsp;0.207, Supplementary Fig.\u0026nbsp;5A), but did show an overall trend towards increased frequency of active cells (F(1,90)\u0026thinsp;=\u0026thinsp;3.463, p\u0026thinsp;=\u0026thinsp;0.067, Supplementary Fig.\u0026nbsp;5B and Supplementary Table\u0026nbsp;1). Gabazine affected the percentage of SSA-participating cells and the frequency of SSA: a significant increase of the percentage of SSA was observed for control and \u003cem\u003eSnap25\u003c/em\u003e single slices (p\u0026thinsp;=\u0026thinsp;0.045 and p\u0026thinsp;=\u0026thinsp;0.014, respectively Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eC), but not for \u003cem\u003eStxbp1\u003c/em\u003e single and double mutant slices (p\u0026thinsp;=\u0026thinsp;0.145 and p\u0026thinsp;=\u0026thinsp;0.827 Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eD). The frequency of SSA in control slices was significantly increased after application of gabazine (p\u0026thinsp;=\u0026thinsp;0.022, Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eD). On the other hand, the application of gabazine did not affect the frequency of SSA in \u003cem\u003eStxbp1\u003c/em\u003e single and double mutant slices. Thus, the frequency of SSA in brain slice from \u003cem\u003eStxbp1\u003c/em\u003e single mutants was significantly lower than in brain slices from control and \u003cem\u003eSnap25\u003c/em\u003e single mutants after application of gabazine (p\u0026thinsp;=\u0026thinsp;0.016 and p\u0026thinsp;=\u0026thinsp;0.018 Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eD). These data show that GABA inhibits SSA in brain slices from control and \u003cem\u003eSnap25\u003c/em\u003e single mutants but that this inhibitory effect of GABA was absent in brain slices from \u003cem\u003eStxbp1\u003c/em\u003e single- and double- mutants.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003cb\u003eDouble mutants show reduced synaptic transmission like single\u003c/b\u003e \u003cb\u003eStxbp1\u003c/b\u003e \u003cb\u003eneurons\u003c/b\u003e\u003c/p\u003e \u003cp\u003eTo assess effects of \u003cem\u003eStxbp1\u003c/em\u003e and \u003cem\u003eSnap25\u003c/em\u003e gene interaction on basic synaptic function, we performed two independent electrophysiological experiments in single (autaptic) hippocampal neurons in culture. Experiments were performed over 5\u0026ndash;9 independent experimental weeks and data were normalized to the average of control for every week. In the first series of experiments a significant effect of genotype on the amplitude of evoked response in glutamatergic neurons (F(3, 240)\u0026thinsp;=\u0026thinsp;5.523, p\u0026thinsp;=\u0026thinsp;0.0011, Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eA-B) and the frequency of spontaneous release (F(3, 235)\u0026thinsp;=\u0026thinsp;3.942, p\u0026thinsp;=\u0026thinsp;0.0090, Supplementary Fig.\u0026nbsp;6A-B) was observed. A \u003cem\u003eposthoc\u003c/em\u003e test revealed a decrease of EPSC amplitude in single \u003cem\u003eStxbp1\u003c/em\u003e and double mutant neurons (p\u0026thinsp;=\u0026thinsp;0.007, p\u0026thinsp;=\u0026thinsp;0.002, respectively, Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eB). Similarly, a trend of decreased mEPSC frequency in single \u003cem\u003eStxbp1\u003c/em\u003e and a significant decrease in double mutant neurons were observed (p\u0026thinsp;=\u0026thinsp;0.063 and p\u0026thinsp;=\u0026thinsp;0.008, respectively) but no effect on mEPSC amplitude (F(3, 237)\u0026thinsp;=\u0026thinsp;3.150, p\u0026thinsp;=\u0026thinsp;0.0257, Supplementary Fig.\u0026nbsp;6A-C). \u003cem\u003eSnap25\u003c/em\u003e single mutant neurons showed no significant difference on any of the synaptic parameters tested (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eA-D). Taken together, these data indicate that glutamatergic transmission is normal in \u003cem\u003eSnap25\u003c/em\u003e single mutant synapses, while equally affected in single \u003cem\u003eStxbp1\u003c/em\u003e and double mutant synapses.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eThe decreased glutamatergic transmission can be caused by a decreased number of glutamatergic synapses, decreased synaptic efficacy (release probability) and/or by decreased size of the readily releasable pool (RRP). To discriminate between these possibilities, we tested \u003cem\u003ein vitro\u003c/em\u003e synapse formation. Morphological analysis revealed no consistent differences in the dendritic length, synaptic size and synaptic density between single \u003cem\u003eStxbp1\u003c/em\u003e and double mutant VGLUT(+) neurons compared to control neurons (Supplementary Fig.\u0026nbsp;7 and Supplementary Table\u0026nbsp;1). The size of RRP was assessed by application of hypertonic sucrose (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eC-D) and was significantly decreased in single \u003cem\u003eStxbp1\u003c/em\u003e and double mutant neurons (p\u0026thinsp;\u0026lt;\u0026thinsp;0.001 and p\u0026thinsp;=\u0026thinsp;0.002, respectively, Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eD). Taken together, these data suggest that the decreased glutamate release in single \u003cem\u003eStxbp1\u003c/em\u003e and double mutant neurons is mainly due to a decreased size of the RRP.\u003c/p\u003e \u003cp\u003eSynaptic transmission was also assessed in \u003cem\u003eStxbp1\u003c/em\u003e single, \u003cem\u003eSnap25\u003c/em\u003e single and double mutant GABA-ergic neurons. (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eE). No significant effect was observed for evoked IPSC, RRP size and mIPSC frequency and amplitude (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eF-H, Supplementary Fig.\u0026nbsp;6F-J and Supplementary Table\u0026nbsp;1), indicating impaired excitatory synaptic transmission, without affecting inhibitory synaptic function. These findings are consistent with data obtained from slice recordings in the somatosensory cortex of the same mouse line (\u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e45\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eThe second experiment was performed in a different laboratory, on the mouse lines used for the system- and behavioral analyses (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e\u0026ndash;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003e), (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eI-L). In the second series of experiments, we examined evoked synaptic response after single action potential (EPSC), paired pulse ratio after 20, 50 and 100 ms intervals and synaptic run-down and cumulative discharge after 100 pulses at 40 Hz train. The EPSC amplitudes did not differ between groups (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eI) and paired pulse ratios showed paired-pulse facilitation without significant differences between groups (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eJ, Supplementary Table\u0026nbsp;1). The cumulative charge released after 100 pulses at 40 Hz showed non-significant trend toward increased value in single \u003cem\u003eSnap25\u003c/em\u003e neurons compared to control neurons (\u0026gt;\u0026thinsp;60% increase compared to control), while it did not differ in single \u003cem\u003eStxbp1\u003c/em\u003e neurons and double mutant neurons compared to control neurons (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eK, Supplementary Table\u0026nbsp;1). The RRP was estimated from the cumulative charge at 40Hz, 100 pulses by extrapolation of the last 20 pulses (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eK) and showed no significant differences between groups (Supplementary Table\u0026nbsp;1). Taken together, the results from the second experimental series suggested no overall deficits in basal synaptic transmission in cultured autaptic neurons of single and double mutant mice.\u003c/p\u003e \u003cp\u003eTo understand the cause of differences found for the electrophysiological phenotypes in two laboratories, we analyzed the samples from single \u003cem\u003eStxbp1\u003c/em\u003e and single \u003cem\u003eSnap25\u003c/em\u003e mice in both laboratories using miniMUGA (\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e). MiniMUGA analysis revealed the presence of 129-strain SNPs in all samples flanking the deletion sites for the two genes (Supplementary Table\u0026nbsp;3 and Supplementary Fig.\u0026nbsp;1). The dominant genetic background of mice used in the first electrophysiological experiments was the C57BL/6JBomTac sub-strain (90% consistent SNPs, Supplementary Table\u0026nbsp;3 and Supplementary Fig.\u0026nbsp;2), while mice used in the second electrophysiological experiment were original C57BL/6J sub-strain (96.8% consistent SNPs, Supplementary Table\u0026nbsp;3 and Supplementary Fig.\u0026nbsp;1). This difference in genomic background, in addition to environmental factors and subtle differences in experimental procedure between laboratories, may contribute to variation in electrophysiological phenotypes between the labs.\u003c/p\u003e \u003c/div\u003e"},{"header":"Discussion","content":"\u003cp\u003eSymptoms among SNAREopathy patients are diverse, with different degrees of developmental delay in different domains (language, motor function, cognition), often, but not always, accompanied by seizures and autistic features (\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e, \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e, \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e, \u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e). Here, we investigated, using a theoretical framework combined with empirical tests, how gene interactions influence phenotypic diversity, i.e., when phenotypic differences observed among individuals with a given (disease-causing) genotype at one locus are influenced by their genotypes at another locus (\u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e46\u003c/span\u003e). We found strong evidence for multiplicative (epistatic) interaction between two SNAREopathy loci at the systems level: seizures and epileptiform activities ranged from no detectable or mild abnormalities in single mutants to lethal clonic and generalized seizures in double mutants. C-Fos staining showed a concomitant large variation. However, at the synapse and network level we found no evidence for such interactions. At the behavioral/cognitive level, we were only able to access surviving animals (approximately half the population, the least severely affected half). These remaining double mutant animals showed phenotypes similar to the strongest of the two single mutants (\u003cem\u003eStxbp1\u003c/em\u003e single mutants).\u003c/p\u003e \u003cdiv id=\"Sec19\" class=\"Section2\"\u003e \u003ch2\u003eEpistatic interactions explain diversity in seizures and EEG-abnormalities\u003c/h2\u003e \u003cp\u003eThe incidence of (mild) seizure-like events among individual mice showed a normal distribution in \u003cem\u003eStxbp1\u003c/em\u003e and \u003cem\u003eSnap25\u003c/em\u003e single mutants with more severe effects in \u003cem\u003eStxbp1\u003c/em\u003e single mutants. This suggests that epileptic events in these mice are largely mediated by one factor (inactivation of one \u003cem\u003eStxbp1\u003c/em\u003e or \u003cem\u003eSnap25\u003c/em\u003e allele, Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e model A), with \u003cem\u003eStxbp1\u003c/em\u003e haploinsufficiency having a larger impact than \u003cem\u003eSnap25\u003c/em\u003e haploinsufficiency. \u003cem\u003eStxbp1\u003c/em\u003e and \u003cem\u003eSnap25\u003c/em\u003e are known to work together to regulate neurotransmitter release and synaptic transmission (\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e, \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e). The extreme phenotypic diversity observed in double mutants is not consistent with the predicted phenotypic effects of variation in two genes acting in the same pathway (common pathway model, Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eB1) and suggests multiplicative (epistatic) gene-gene interactions (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e, model B2.2.), and the involvement of distinct, interacting deficiencies caused by \u003cem\u003eStxbp1\u003c/em\u003e and \u003cem\u003eSnap25\u003c/em\u003e haploinsufficiency.\u003c/p\u003e \u003cp\u003eSuch distinct, interacting deficiencies may be explained in different, not mutually exclusive ways. First, although the two genes work together in neurotransmitter release, and also in the secretion of neuropeptides and neuromodulators (\u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e47\u003c/span\u003e, \u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e48\u003c/span\u003e), haploinsufficiency may affect different aspects of these processes, in opposite direction or under different circumstances for the two genes. The current study did not reveal deficits in synaptic transmission for \u003cem\u003eSnap25\u003c/em\u003e haploinsufficiency in single neurons (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003e), but previous studies have reported impairments in different aspects of synaptic transmission in \u003cem\u003eSnap25\u003c/em\u003e haploinsufficiency models which are distinct from the deficits for \u003cem\u003eStxbp1\u003c/em\u003e haploinsufficiency detected in the present and the previous studies, (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003e and (\u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e41\u003c/span\u003e, \u003cspan additionalcitationids=\"CR50\" citationid=\"CR49\" class=\"CitationRef\"\u003e49\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR51\" class=\"CitationRef\"\u003e51\u003c/span\u003e)).\u003c/p\u003e \u003cp\u003eSecond, \u003cem\u003eStxbp1\u003c/em\u003e and \u003cem\u003eSnap25\u003c/em\u003e haploinsufficiency may have distinct, interacting effects on different populations of neurons and/or brain networks. \u003cem\u003eStxbp1\u003c/em\u003e haploinsufficiency was shown to have different effects on hippocampal GABAergic and glutamatergic neurons (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003e, Toonen \u003cem\u003eet al.\u003c/em\u003e(\u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e50\u003c/span\u003e)) or deficits specifically in GABAergic interneurons in the cortex (\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e), deficits in recruiting these interneurons (\u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e45\u003c/span\u003e) and glutamatergic inputs in the striatum (\u003cspan citationid=\"CR52\" class=\"CitationRef\"\u003e52\u003c/span\u003e). On the other hand, the sporadic epileptiform events in single \u003cem\u003eSnap25\u003c/em\u003e mice were ascribed to increased calcium responsiveness of thalamic neurons and hyper-excitability of thalamo-cortical circuits (\u003cspan citationid=\"CR53\" class=\"CitationRef\"\u003e53\u003c/span\u003e, \u003cspan citationid=\"CR54\" class=\"CitationRef\"\u003e54\u003c/span\u003e). In the present study we found a trend towards an increase in c-Fos expression in the cortical regions of \u003cem\u003eSnap25\u003c/em\u003e single mutants and increased c-Fos expression in double mutants, confirming the cortical hyperexcitability. Interestingly, c-Fos expression was increased in the thalamus of double mutants, but not in the thalamus of both single mutants, suggesting that the thalamus is a critical brain region for genetic interaction. The idea that \u003cem\u003eStxbp1\u003c/em\u003e and \u003cem\u003eSnap25\u003c/em\u003e haploinsufficiency may have distinct, interacting effects on different populations of neurons and/or brain networks is in line with the observation that multiplicative (epistatic) effects were not observed in individual neurons (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003e) or networks (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003e), but were pronounced at the system level (generalized seizures, only in the double mutants).\u003c/p\u003e \u003cp\u003eThird, in addition to their best characterized cellular functions in the regulated secretion of neurotransmitters and neuromodulators, \u003cem\u003eStxbp1\u003c/em\u003e and \u003cem\u003eSnap25\u003c/em\u003e haploinsufficiency may affect other functions which may contribute to non-linear genetic interaction, especially during earlier developmental phases. For instance, the data from early postnatal network activity suggests deficits in the GABA shift in \u003cem\u003eStxbp1\u003c/em\u003e mutants, but not \u003cem\u003eSnap25\u003c/em\u003e mutants (see also below). Furthermore, \u003cem\u003eSnap25\u003c/em\u003e, but not \u003cem\u003eStxbp1\u003c/em\u003e haploinsufficiency was reported to produce negative modulation of voltage-gated calcium channels (\u003cspan citationid=\"CR51\" class=\"CitationRef\"\u003e51\u003c/span\u003e, \u003cspan citationid=\"CR55\" class=\"CitationRef\"\u003e55\u003c/span\u003e, \u003cspan citationid=\"CR56\" class=\"CitationRef\"\u003e56\u003c/span\u003e) and both genes have crucial, not fully overlapping roles in neuronal viability (\u003cspan citationid=\"CR57\" class=\"CitationRef\"\u003e57\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eFinally, epigenetic variation or stochastic processes during development may also contribute to interacting deficiencies detected on the system level. Taken together, we conclude that non-linear (multiplicative) genetic interaction of distinct aspects of regulated secretion, distinct neuronal populations/networks in the brain, with other cellular functions and/or epigenetic/stochastic effects, together explain the broad diversity in phenotypic manifestations in mutant mice.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec20\" class=\"Section2\"\u003e \u003ch2\u003eNo evidence for epistatic interactions in surviving double mutants\u003c/h2\u003e \u003cp\u003eAcross many behavioral domains, the behavioral phenotypes of single \u003cem\u003eStxbp1\u003c/em\u003e mutants and surviving double mutants were similar: learning and memory, behavioral flexibility, anxiety and spontaneous behavior, while \u003cem\u003eSnap25\u003c/em\u003e single mutants behaved like control mice in all these domains. Single \u003cem\u003eStxbp1\u003c/em\u003e mice and double mutant mice showed pronounced impairment in associative learning and memory and behavioral flexibility, lower body weight and normal motor coordination and muscle strength in line with previous studies in \u003cem\u003eStxbp1\u003c/em\u003e haploinsufficient mice (\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e, \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e). Interestingly, subtle increases in activity in a habituated environment during the dark phase in single \u003cem\u003eStxbp1\u003c/em\u003e mice was alleviated in double mutants. On the other hand, normal cognition, spontaneous behavior, motor coordination, muscle strength and anxiety found in single \u003cem\u003eSnap25\u003c/em\u003e mice are in line with previously reported lack of significant behavioral and cognitive impairments in \u003cem\u003eSnap25\u003c/em\u003e\u003csup\u003e\u003cem\u003e+/\u0026minus;\u003c/em\u003e\u003c/sup\u003e mice (\u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e, \u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e). Behavior was assessed in the surviving mice older than eight-weeks, after approximately 50% of double mutants have died (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eA). It is plausible that double mutants with most severe behavioral phenotypes were lost due to early lethality. However, the relationship between early lethality and later performance in behavioral tests of the surviving animals is unknown, precluding strong conclusions on which type of genetic interaction (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e) is most consistent with the behavioral data.\u003c/p\u003e \u003cp\u003eA similar conclusion can be reached for sleep phenotypes, with the only variation that phenotypes of the single mutants were similar. A shift of the spectral power to higher frequencies during sleep was found in \u003cem\u003eStxbp1\u003c/em\u003e single, \u003cem\u003eSnap25\u003c/em\u003e single and double mutants. Hence, while for behavior, the phenotypic spectrum of \u003cem\u003eStxbp1\u003c/em\u003e mutants is the strongest among the two single mutants, for sleep this spectrum is approximately the same for both single mutants and the double mutants behave similar to both single mutants.\u003c/p\u003e \u003cp\u003e \u003cb\u003eA failing GABA-shift as a key determinant in\u003c/b\u003e \u003cb\u003eStxbp1\u003c/b\u003e \u003cb\u003ehaploinsufficiency mice\u003c/b\u003e\u003c/p\u003e \u003cp\u003eThe assessment of the spontaneous synchronous activity (SSA) in two-weeks-old mutants indicated that the normal shift from depolarizing to hyperpolarizing GABA action, observed in control mice and \u003cem\u003eSnap25\u003c/em\u003e single mutants, was not observed in \u003cem\u003eStxbp1\u003c/em\u003e single- and double-mutants. SSA is important for the establishment and maturation of functional neural circuits(\u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e44\u003c/span\u003e, \u003cspan citationid=\"CR58\" class=\"CitationRef\"\u003e58\u003c/span\u003e, \u003cspan citationid=\"CR59\" class=\"CitationRef\"\u003e59\u003c/span\u003e) and the emergence of inhibitory GABA action is crucial for the termination of SSA (\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e). GABA-shift abnormalities have been observed before in several other neurodevelopmental disorders, such as Fragile-X syndrome and Rett syndrome(\u003cspan citationid=\"CR60\" class=\"CitationRef\"\u003e60\u003c/span\u003e, \u003cspan citationid=\"CR61\" class=\"CitationRef\"\u003e61\u003c/span\u003e). The GABA-shift is mediated by several (external and internal) factors, including network activity and neuropeptide release (\u003cspan citationid=\"CR62\" class=\"CitationRef\"\u003e62\u003c/span\u003e). Both processes depend on SNARE-dependent vesicle fusion. Therefore, it is plausible that initial dysregulation of network activity and/or neuropeptide release is a key determinant in Stxbp1 syndrome and potentially other SNAREopathies. Interestingly, while both Snap25 and Munc18/Stxbp1 are essential for neuropeptide release from dense core vesicles (DCV) in mature neurons (\u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e47\u003c/span\u003e, \u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e48\u003c/span\u003e), immature \u003cem\u003eSnap25\u003c/em\u003e null mutant neurons showed substantial remaining DCV exocytosis, similar to immature wild type neurons, which was attributed to redundancy with other Qb/c SNAREs during early development (\u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e47\u003c/span\u003e). This finding provides a plausible explanation for the fact that \u003cem\u003eSnap25\u003c/em\u003e single mutants showed a normal GABA-shift and multiplicative (epistatic) effects were not observed for spontaneous synchronous activity in slices. Taken together, our findings support the hypothesis that neurodevelopmental disorders including \u003cem\u003eStxbp1\u003c/em\u003e syndrome are initiated at early developmental stages and involve abnormalities in GABA-shift and delayed cortical network development(\u003cspan citationid=\"CR63\" class=\"CitationRef\"\u003e63\u003c/span\u003e).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec21\" class=\"Section2\"\u003e \u003ch2\u003eGenetic interaction effects may extrapolate to the human population\u003c/h2\u003e \u003cp\u003eThis study provides experimental evidence for the predicted effects of genetic interactions on phenotypic diversity (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). Phenotypic diversity is a major, unexplained issue for SNAREopathies and other DEE and a major complication for the assessment of therapy success and the evaluation of new candidate therapies, e.g. in clinical trials. To establish proof of concept, the current study took a reductionistic approach, with a single \u0026lsquo;modifying gene\u0026rsquo; of large effect size (heterozygous deletion) interacting with a primary genetic deficit under conditions where other genomic variation and environmental factors are radically minimalized (inbred mouse lines and highly standardized laboratory conditions, respectively). In the human population, variation in these factors add complexity and the accumulation of two genetic variants of large effect size in single individuals is very rare. However, the clear evidence for genetic interaction in the current study can be extrapolated to the human population (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003e). Interactions between a primary genetic deficit (pathogenic variant) and multiple (common) genetic variation is expected to produce similar effects on phenotypic diversity and may explain the high diversity observed among patients with mutations in the genes that work together. This idea is consistent with the concept of \u0026lsquo;genetic buffering\u0026rsquo;(\u003cspan citationid=\"CR64\" class=\"CitationRef\"\u003e64\u003c/span\u003e). According to the multiplicative model (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eB2.2), a fraction of individuals with pathogenic variants is only mildly affected, or even unaffected, due to the \u0026lsquo;buffering\u0026rsquo; effect of certain other genetic variants in the genome. Furthermore, the conclusion that epilepsy and cognitive impairments are mediated by different cellular mechanisms explains the limited efficacy of antiepileptic drugs that act on one specific cellular mechanism and their inability to ameliorate developmental aspects of DEE. Finally, this study associates atypical GABA cortical network development with neurodevelopmental delay in SNAREopathy models, suggesting a novel predictive biomarker for future research, diagnostics and treatment design.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e"},{"header":"Abbreviations","content":"\u003cdiv class=\"DefinitionList\"\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eEPSC\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eexcitatory postsynaptic current\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eIPSC\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003einhibitory postsynaptic current\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003c/div\u003e"},{"header":"Declarations","content":" \u003ch2\u003eCompeting interests:\u003c/h2\u003e \u003cp\u003eThe authors report no competing interests. At the start of the experiments described in this manuscript, J.K. was a full time employee of Sylics (Synaptologics BV), a private, VU University spin-off company that offers mouse phenotyping services, also on the mice described in this manuscript; during later phases, J.K. was employed by VU University. Sylics had no influence on the scientific decisions made by her and others involved in her work. During the time of this study, M.V. participated in a holding that owned Sylics shares and has received consulting fees from Sylics. The Belgian company InnoSer NV acquired Sylics in 2023. M.V. continued to received consulting fees from Sylics/InnoSer.\u003c/p\u003e \u003c/p\u003e\u003ch2\u003eFunding:\u003c/h2\u003e \u003cp\u003eThe work was supported by the Independent Research Fund Denmark (8020-00228A to JBS) and the Lundbeck Foundation (to JBS and MV, R277-2018-802) and the Novo Nordisk Foundation (to JBS, NNF10OC0058298).\u003c/p\u003e\u003ch2\u003eAcknowledgements:\u003c/h2\u003e \u003cp\u003eWe thank Sita van der Wal, Natasja Bos and Christian van der Meer for breeding mutant mice. We thank Joost Hoetjes and Dorte Lauritsen for genotyping mutant mice. We thank Rolinka van der Loo for help with organizing perfusions and surgeries.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eHamdan FF, Myers CT, Cossette P, Lemay P, Spiegelman D, Laporte AD\u003cem\u003e et al.\u003c/em\u003e High Rate of Recurrent De Novo Mutations in Developmental and Epileptic Encephalopathies. \u003cem\u003eAm J Hum Genet\u003c/em\u003e 2017; \u003cstrong\u003e101\u003c/strong\u003e(5)\u003cstrong\u003e: \u003c/strong\u003e664-685.\u003c/li\u003e\n\u003cli\u003eScheffer IE, Berkovic S, Capovilla G, Connolly MB, French J, Guilhoto L\u003cem\u003e et al.\u003c/em\u003e ILAE classification of the epilepsies: Position paper of the ILAE Commission for Classification and Terminology. \u003cem\u003eEpilepsia\u003c/em\u003e 2017; \u003cstrong\u003e58\u003c/strong\u003e(4)\u003cstrong\u003e: \u003c/strong\u003e512-521.\u003c/li\u003e\n\u003cli\u003eCarvill GL, Heavin SB, Yendle SC, McMahon JM, O\u0026apos;Roak BJ, Cook J\u003cem\u003e et al.\u003c/em\u003e Targeted resequencing in epileptic encephalopathies identifies de novo mutations in CHD2 and SYNGAP1. \u003cem\u003eNat Genet\u003c/em\u003e 2013; \u003cstrong\u003e45\u003c/strong\u003e(7)\u003cstrong\u003e: \u003c/strong\u003e825-830.\u003c/li\u003e\n\u003cli\u003eEuro E-RESC, Epilepsy Phenome/Genome P, Epi KC. De novo mutations in synaptic transmission genes including DNM1 cause epileptic encephalopathies. \u003cem\u003eAm J Hum Genet\u003c/em\u003e 2014; \u003cstrong\u003e95\u003c/strong\u003e(4)\u003cstrong\u003e: \u003c/strong\u003e360-370.\u003c/li\u003e\n\u003cli\u003eHeyne HO, Singh T, Stamberger H, Abou Jamra R, Caglayan H, Craiu D\u003cem\u003e et al.\u003c/em\u003e De novo variants in neurodevelopmental disorders with epilepsy. \u003cem\u003eNat Genet\u003c/em\u003e 2018; \u003cstrong\u003e50\u003c/strong\u003e(7)\u003cstrong\u003e: \u003c/strong\u003e1048-1053.\u003c/li\u003e\n\u003cli\u003eMcTague A, Howell KB, Cross JH, Kurian MA, Scheffer IE. The genetic landscape of the epileptic encephalopathies of infancy and childhood. \u003cem\u003eThe Lancet Neurology\u003c/em\u003e 2016; \u003cstrong\u003e15\u003c/strong\u003e(3)\u003cstrong\u003e: \u003c/strong\u003e304-316.\u003c/li\u003e\n\u003cli\u003eKullmann DM. Neurological channelopathies. \u003cem\u003eAnnu Rev Neurosci\u003c/em\u003e 2010; \u003cstrong\u003e33: \u003c/strong\u003e151-172.\u003c/li\u003e\n\u003cli\u003eVerhage M, S\u0026oslash;rensen JB. SNAREopathies: Diversity in Mechanisms and Symptoms. \u003cem\u003eNeuron\u003c/em\u003e 2020; \u003cstrong\u003e107\u003c/strong\u003e(1)\u003cstrong\u003e: \u003c/strong\u003e22-37.\u003c/li\u003e\n\u003cli\u003eVissers LE, Gilissen C, Veltman JA. Genetic studies in intellectual disability and related disorders. \u003cem\u003eNat Rev Genet\u003c/em\u003e 2016; \u003cstrong\u003e17\u003c/strong\u003e(1)\u003cstrong\u003e: \u003c/strong\u003e9-18.\u003c/li\u003e\n\u003cli\u003eMcClellan J, King MC. Genetic heterogeneity in human disease. \u003cem\u003eCell\u003c/em\u003e 2010; \u003cstrong\u003e141\u003c/strong\u003e(2)\u003cstrong\u003e: \u003c/strong\u003e210-217.\u003c/li\u003e\n\u003cli\u003eGen\u0026ccedil; \u0026Ouml;, An J-Y, Fetter RD, Kulik Y, Zunino G, Sanders SJ\u003cem\u003e et al.\u003c/em\u003e Homeostatic plasticity fails at the intersection of autism-gene mutations and a novel class of common genetic modifiers. \u003cem\u003eElife\u003c/em\u003e 2020; \u003cstrong\u003e9: \u003c/strong\u003ee55775.\u003c/li\u003e\n\u003cli\u003eAbramov D, Guiberson NGL, Daab A, Na Y, Petsko GA, Sharma M\u003cem\u003e et al.\u003c/em\u003e Targeted stabilization of Munc18-1 function via pharmacological chaperones. \u003cem\u003eEMBO Mol Med\u003c/em\u003e 2021; \u003cstrong\u003e13\u003c/strong\u003e(1)\u003cstrong\u003e: \u003c/strong\u003ee12354.\u003c/li\u003e\n\u003cli\u003eChen W, Cai ZL, Chao ES, Chen H, Longley CM, Hao S\u003cem\u003e et al.\u003c/em\u003e Stxbp1/Munc18-1 haploinsufficiency impairs inhibition and mediates key neurological features of STXBP1 encephalopathy. \u003cem\u003eElife\u003c/em\u003e 2020; \u003cstrong\u003e9\u003c/strong\u003e.\u003c/li\u003e\n\u003cli\u003eKovacevic J, Maroteaux G, Schut D, Loos M, Dubey M, Pitsch J\u003cem\u003e et al.\u003c/em\u003e Protein instability, haploinsufficiency, and cortical hyper-excitability underlie STXBP1 encephalopathy. \u003cem\u003eBrain\u003c/em\u003e 2018; \u003cstrong\u003e141\u003c/strong\u003e(5)\u003cstrong\u003e: \u003c/strong\u003e1350-1374.\u003c/li\u003e\n\u003cli\u003eLammertse HCA, van Berkel AA, Iacomino M, Toonen RF, Striano P, Gambardella A\u003cem\u003e et al.\u003c/em\u003e Homozygous STXBP1 variant causes encephalopathy and gain-of-function in synaptic transmission. \u003cem\u003eBrain\u003c/em\u003e 2019.\u003c/li\u003e\n\u003cli\u003eXian J, Parthasarathy S, McKeown S, Balagura G, Fitch E, Helbig K\u003cem\u003e et al.\u003c/em\u003e Assessing the landscape of STXBP1-related disorders in 534 individuals. \u003cem\u003eBrain\u003c/em\u003e 2021.\u003c/li\u003e\n\u003cli\u003eKadkova A, Murach J, Ostergaard M, Malsam A, Malsam J, Lolicato F\u003cem\u003e et al.\u003c/em\u003e SNAP25 disease mutations change the energy landscape for synaptic exocytosis due to aberrant SNARE interactions. \u003cem\u003eElife\u003c/em\u003e 2024; \u003cstrong\u003e12\u003c/strong\u003e.\u003c/li\u003e\n\u003cli\u003eStamberger H, Nikanorova M, Willemsen MH, Accorsi P, Angriman M, Baier H\u003cem\u003e et al.\u003c/em\u003e STXBP1 encephalopathy: A neurodevelopmental disorder including epilepsy. \u003cem\u003eNeurology\u003c/em\u003e 2016; \u003cstrong\u003e86\u003c/strong\u003e(10)\u003cstrong\u003e: \u003c/strong\u003e954-962.\u003c/li\u003e\n\u003cli\u003eAbramov D, Guiberson NGL, Burr\u0026eacute; J. STXBP1 encephalopathies: Clinical spectrum, disease mechanisms, and therapeutic strategies. \u003cem\u003eJ Neurochem\u003c/em\u003e 2021; \u003cstrong\u003e157\u003c/strong\u003e(2)\u003cstrong\u003e: \u003c/strong\u003e165-178.\u003c/li\u003e\n\u003cli\u003eL\u0026oacute;pez-Rivera JA, P\u0026eacute;rez-Palma E, Symonds J, Lindy AS, McKnight DA, Leu C\u003cem\u003e et al.\u003c/em\u003e A catalogue of new incidence estimates of monogenic neurodevelopmental disorders caused by de novo variants. \u003cem\u003eBrain\u003c/em\u003e 2020; \u003cstrong\u003e143\u003c/strong\u003e(4)\u003cstrong\u003e: \u003c/strong\u003e1099-1105.\u003c/li\u003e\n\u003cli\u003ePhillips PC. Epistasis--the essential role of gene interactions in the structure and evolution of genetic systems. \u003cem\u003eNat Rev Genet\u003c/em\u003e 2008; \u003cstrong\u003e9\u003c/strong\u003e(11)\u003cstrong\u003e: \u003c/strong\u003e855-867.\u003c/li\u003e\n\u003cli\u003eChou HH, Chiu HC, Delaney NF, Segre D, Marx CJ. Diminishing returns epistasis among beneficial mutations decelerates adaptation. \u003cem\u003eScience\u003c/em\u003e 2011; \u003cstrong\u003e332\u003c/strong\u003e(6034)\u003cstrong\u003e: \u003c/strong\u003e1190-1192.\u003c/li\u003e\n\u003cli\u003eVerhage M, Maia AS, Plomp JJ, Brussaard AB, Heeroma JH, Vermeer H\u003cem\u003e et al.\u003c/em\u003e Synaptic assembly of the brain in the absence of neurotransmitter secretion. \u003cem\u003eScience\u003c/em\u003e 2000; \u003cstrong\u003e287\u003c/strong\u003e(5454)\u003cstrong\u003e: \u003c/strong\u003e864-869.\u003c/li\u003e\n\u003cli\u003eWashbourne P, Thompson PM, Carta M, Costa ET, Mathews JR, Lopez-Bendit\u0026oacute; G\u003cem\u003e et al.\u003c/em\u003e Genetic ablation of the t-SNARE SNAP-25 distinguishes mechanisms of neuroexocytosis. \u003cem\u003eNat Neurosci\u003c/em\u003e 2002; \u003cstrong\u003e5\u003c/strong\u003e(1)\u003cstrong\u003e: \u003c/strong\u003e19-26.\u003c/li\u003e\n\u003cli\u003eLoos M, Koopmans B, Aarts E, Maroteaux G, van der Sluis S, Verhage M\u003cem\u003e et al.\u003c/em\u003e Sheltering behavior and locomotor activity in 11 genetically diverse common inbred mouse strains using home-cage monitoring. \u003cem\u003ePLoS One\u003c/em\u003e 2014; \u003cstrong\u003e9\u003c/strong\u003e(9)\u003cstrong\u003e: \u003c/strong\u003ee108563.\u003c/li\u003e\n\u003cli\u003eMaroteaux G, Loos M, van der Sluis S, Koopmans B, Aarts E, van Gassen K\u003cem\u003e et al.\u003c/em\u003e High-throughput phenotyping of avoidance learning in mice discriminates different genotypes and identifies a novel gene. \u003cem\u003eGenes Brain Behav\u003c/em\u003e 2012; \u003cstrong\u003e11\u003c/strong\u003e(7)\u003cstrong\u003e: \u003c/strong\u003e772-784.\u003c/li\u003e\n\u003cli\u003eRemmelink E, Smit AB, Verhage M, Loos M. Measuring discrimination- and reversal learning in mouse models within 4 days and without prior food deprivation. \u003cem\u003eLearn Mem\u003c/em\u003e 2016; \u003cstrong\u003e23\u003c/strong\u003e(11)\u003cstrong\u003e: \u003c/strong\u003e660-667.\u003c/li\u003e\n\u003cli\u003eWright S, Hashemi K, Stasiak L, Bartram J, Lang B, Vincent A\u003cem\u003e et al.\u003c/em\u003e Epileptogenic effects of NMDAR antibodies in a passive transfer mouse model. \u003cem\u003eBrain\u003c/em\u003e 2015; \u003cstrong\u003e138\u003c/strong\u003e(Pt 11)\u003cstrong\u003e: \u003c/strong\u003e3159-3167.\u003c/li\u003e\n\u003cli\u003eBokil H, Andrews P, Kulkarni JE, Mehta S, Mitra PP. Chronux: a platform for analyzing neural signals. \u003cem\u003eJ Neurosci Methods\u003c/em\u003e 2010; \u003cstrong\u003e192\u003c/strong\u003e(1)\u003cstrong\u003e: \u003c/strong\u003e146-151.\u003c/li\u003e\n\u003cli\u003eDawitz J, Kroon T, Hjorth JJ, Meredith RM. Functional calcium imaging in developing cortical networks. \u003cem\u003eJ Vis Exp\u003c/em\u003e 2011; (56).\u003c/li\u003e\n\u003cli\u003ePires J, Nelissen R, Mansvelder HD, Meredith RM. Spontaneous synchronous network activity in the neonatal development of mPFC in mice. \u003cem\u003eDev Neurobiol\u003c/em\u003e 2021; \u003cstrong\u003e81\u003c/strong\u003e(2)\u003cstrong\u003e: \u003c/strong\u003e207-225.\u003c/li\u003e\n\u003cli\u003eHjorth JJ, Dawitz J, Kroon T, Pires J, Dassen VJ, Berkhout JA\u003cem\u003e et al.\u003c/em\u003e Detection of silent cells, synchronization and modulatory activity in developing cellular networks. \u003cem\u003eDev Neurobiol\u003c/em\u003e 2016; \u003cstrong\u003e76\u003c/strong\u003e(4)\u003cstrong\u003e: \u003c/strong\u003e357-374.\u003c/li\u003e\n\u003cli\u003eRuiter M, Kadkova A, Scheutzow A, Malsam J, Sollner TH, Sorensen JB. An Electrostatic Energy Barrier for SNARE-Dependent Spontaneous and Evoked Synaptic Transmission. \u003cem\u003eCell Rep\u003c/em\u003e 2019; \u003cstrong\u003e26\u003c/strong\u003e(9)\u003cstrong\u003e: \u003c/strong\u003e2340-2352 e2345.\u003c/li\u003e\n\u003cli\u003eMeijer M, D\u0026ouml;rr B, Lammertse HC, Blithikioti C, van Weering JR, Toonen RF\u003cem\u003e et al.\u003c/em\u003e Tyrosine phosphorylation of Munc18-1 inhibits synaptic transmission by preventing SNARE assembly. \u003cem\u003eEmbo j\u003c/em\u003e 2018; \u003cstrong\u003e37\u003c/strong\u003e(2)\u003cstrong\u003e: \u003c/strong\u003e300-320.\u003c/li\u003e\n\u003cli\u003eMeijer M, Burkhardt P, de Wit H, Toonen RF, Fasshauer D, Verhage M. Munc18-1 mutations that strongly impair SNARE-complex binding support normal synaptic transmission. \u003cem\u003eEmbo j\u003c/em\u003e 2012; \u003cstrong\u003e31\u003c/strong\u003e(9)\u003cstrong\u003e: \u003c/strong\u003e2156-2168.\u003c/li\u003e\n\u003cli\u003eFisher RA. The correlations between relatives on the supposition of Mendelian inheritance. \u003cem\u003eTrans R Soc Edinb\u003c/em\u003e 1918; \u003cstrong\u003e52: \u003c/strong\u003e399-433.\u003c/li\u003e\n\u003cli\u003eMani R, St Onge RP, Hartman JLt, Giaever G, Roth FP. Defining genetic interaction. \u003cem\u003eProc Natl Acad Sci U S A\u003c/em\u003e 2008; \u003cstrong\u003e105\u003c/strong\u003e(9)\u003cstrong\u003e: \u003c/strong\u003e3461-3466.\u003c/li\u003e\n\u003cli\u003eSigmon JS, Blanchard MW, Baric RS, Bell TA, Brennan J, Brockmann GA\u003cem\u003e et al.\u003c/em\u003e Content and Performance of the MiniMUGA Genotyping Array: A New Tool To Improve Rigor and Reproducibility in Mouse Research. \u003cem\u003eGenetics\u003c/em\u003e 2020; \u003cstrong\u003e216\u003c/strong\u003e(4)\u003cstrong\u003e: \u003c/strong\u003e905-930.\u003c/li\u003e\n\u003cli\u003eCorradini I, Donzelli A, Antonucci F, Welzl H, Loos M, Martucci R\u003cem\u003e et al.\u003c/em\u003e Epileptiform activity and cognitive deficits in SNAP-25(+/-) mice are normalized by antiepileptic drugs. \u003cem\u003eCereb Cortex\u003c/em\u003e 2014; \u003cstrong\u003e24\u003c/strong\u003e(2)\u003cstrong\u003e: \u003c/strong\u003e364-376.\u003c/li\u003e\n\u003cli\u003eMonteggia LM, Lin PY, Adachi M, Kavalali ET. Behavioral Analysis of SNAP-25 and Synaptobrevin-2 Haploinsufficiency in Mice. \u003cem\u003eNeuroscience\u003c/em\u003e 2019; \u003cstrong\u003e420: \u003c/strong\u003e129-135.\u003c/li\u003e\n\u003cli\u003eOrock A, Logan S, Deak F. Munc18-1 haploinsufficiency impairs learning and memory by reduced synaptic vesicular release in a model of Ohtahara syndrome. \u003cem\u003eMol Cell Neurosci\u003c/em\u003e 2018; \u003cstrong\u003e88: \u003c/strong\u003e33-42.\u003c/li\u003e\n\u003cli\u003eHerrera DG, Robertson HA. Activation of c-fos in the brain. \u003cem\u003eProg Neurobiol\u003c/em\u003e 1996; \u003cstrong\u003e50\u003c/strong\u003e(2-3)\u003cstrong\u003e: \u003c/strong\u003e83-107.\u003c/li\u003e\n\u003cli\u003eDeacon RM, Rawlins JN. T-maze alternation in the rodent. \u003cem\u003eNat Protoc\u003c/em\u003e 2006; \u003cstrong\u003e1\u003c/strong\u003e(1)\u003cstrong\u003e: \u003c/strong\u003e7-12.\u003c/li\u003e\n\u003cli\u003eBlankenship AG, Feller MB. Mechanisms underlying spontaneous patterned activity in developing neural circuits. \u003cem\u003eNat Rev Neurosci\u003c/em\u003e 2010; \u003cstrong\u003e11\u003c/strong\u003e(1)\u003cstrong\u003e: \u003c/strong\u003e18-29.\u003c/li\u003e\n\u003cli\u003eDos Santos AB, Larsen SD, Guo L, Barbagallo P, Montalant A, Verhage M\u003cem\u003e et al.\u003c/em\u003e Microcircuit failure in STXBP1 encephalopathy leads to hyperexcitability. \u003cem\u003eCell Rep Med\u003c/em\u003e 2023; \u003cstrong\u003e4\u003c/strong\u003e(12)\u003cstrong\u003e: \u003c/strong\u003e101308.\u003c/li\u003e\n\u003cli\u003eCheverud JM, Routman EJ. Epistasis and its contribution to genetic variance components. \u003cem\u003eGenetics\u003c/em\u003e 1995; \u003cstrong\u003e139\u003c/strong\u003e(3)\u003cstrong\u003e: \u003c/strong\u003e1455-1461.\u003c/li\u003e\n\u003cli\u003eArora S, Saarloos I, Kooistra R, van de Bospoort R, Verhage M, Toonen RF. SNAP-25 gene family members differentially support secretory vesicle fusion. \u003cem\u003eJ Cell Sci\u003c/em\u003e 2017; \u003cstrong\u003e130\u003c/strong\u003e(11)\u003cstrong\u003e: \u003c/strong\u003e1877-1889.\u003c/li\u003e\n\u003cli\u003ePuntman DC, Arora S, Farina M, Toonen RF, Verhage M. Munc18-1 is essential for neuropeptide secretion in neurons. \u003cem\u003eJ Neurosci\u003c/em\u003e 2021; \u003cstrong\u003e41\u003c/strong\u003e(28)\u003cstrong\u003e: \u003c/strong\u003e5980-5993.\u003c/li\u003e\n\u003cli\u003ePatzke C, Han Y, Covy J, Yi F, Maxeiner S, Wernig M\u003cem\u003e et al.\u003c/em\u003e Analysis of conditional heterozygous STXBP1 mutations in human neurons. \u003cem\u003eJ Clin Invest\u003c/em\u003e 2015; \u003cstrong\u003e125\u003c/strong\u003e(9)\u003cstrong\u003e: \u003c/strong\u003e3560-3571.\u003c/li\u003e\n\u003cli\u003eToonen RF, Wierda K, Sons MS, de Wit H, Cornelisse LN, Brussaard A\u003cem\u003e et al.\u003c/em\u003e Munc18-1 expression levels control synapse recovery by regulating readily releasable pool size. \u003cem\u003eProc Natl Acad Sci U S A\u003c/em\u003e 2006; \u003cstrong\u003e103\u003c/strong\u003e(48)\u003cstrong\u003e: \u003c/strong\u003e18332-18337.\u003c/li\u003e\n\u003cli\u003eAntonucci F, Corradini I, Morini R, Fossati G, Menna E, Pozzi D\u003cem\u003e et al.\u003c/em\u003e Reduced SNAP-25 alters short-term plasticity at developing glutamatergic synapses. \u003cem\u003eEMBO Rep\u003c/em\u003e 2013; \u003cstrong\u003e14\u003c/strong\u003e(7)\u003cstrong\u003e: \u003c/strong\u003e645-651.\u003c/li\u003e\n\u003cli\u003eMiyamoto H, Tatsukawa T, Shimohata A, Yamagata T, Suzuki T, Amano K\u003cem\u003e et al.\u003c/em\u003e Impaired cortico-striatal excitatory transmission triggers epilepsy. \u003cem\u003eNat Commun\u003c/em\u003e 2019; \u003cstrong\u003e10\u003c/strong\u003e(1)\u003cstrong\u003e: \u003c/strong\u003e1917.\u003c/li\u003e\n\u003cli\u003eWatanabe S, Yamamori S, Otsuka S, Saito M, Suzuki E, Kataoka M\u003cem\u003e et al.\u003c/em\u003e Epileptogenesis and epileptic maturation in phosphorylation site-specific SNAP-25 mutant mice. \u003cem\u003eEpilepsy Res\u003c/em\u003e 2015; \u003cstrong\u003e115: \u003c/strong\u003e30-44.\u003c/li\u003e\n\u003cli\u003eZhang Y, Vilaythong AP, Yoshor D, Noebels JL. Elevated thalamic low-voltage-activated currents precede the onset of absence epilepsy in the SNAP25-deficient mouse mutant coloboma. \u003cem\u003eJ Neurosci\u003c/em\u003e 2004; \u003cstrong\u003e24\u003c/strong\u003e(22)\u003cstrong\u003e: \u003c/strong\u003e5239-5248.\u003c/li\u003e\n\u003cli\u003eCondliffe SB, Corradini I, Pozzi D, Verderio C, Matteoli M. Endogenous SNAP-25 regulates native voltage-gated calcium channels in glutamatergic neurons. \u003cem\u003eJ Biol Chem\u003c/em\u003e 2010; \u003cstrong\u003e285\u003c/strong\u003e(32)\u003cstrong\u003e: \u003c/strong\u003e24968-24976.\u003c/li\u003e\n\u003cli\u003eToft-Bertelsen TL, Ziomkiewicz I, Houy S, Pinheiro PS, S\u0026oslash;rensen JB. Regulation of Ca2+ channels by SNAP-25 via recruitment of syntaxin-1 from plasma membrane clusters. \u003cem\u003eMol Biol Cell\u003c/em\u003e 2016; \u003cstrong\u003e27\u003c/strong\u003e(21)\u003cstrong\u003e: \u003c/strong\u003e3329-3341.\u003c/li\u003e\n\u003cli\u003eSantos TC, Wierda K, Broeke JH, Toonen RF, Verhage M. Early Golgi Abnormalities and Neurodegeneration upon Loss of Presynaptic Proteins Munc18-1, Syntaxin-1, or SNAP-25. \u003cem\u003eJ Neurosci\u003c/em\u003e 2017; \u003cstrong\u003e37\u003c/strong\u003e(17)\u003cstrong\u003e: \u003c/strong\u003e4525-4539.\u003c/li\u003e\n\u003cli\u003eKirkby LA, Sack GS, Firl A, Feller MB. A role for correlated spontaneous activity in the assembly of neural circuits. \u003cem\u003eNeuron\u003c/em\u003e 2013; \u003cstrong\u003e80\u003c/strong\u003e(5)\u003cstrong\u003e: \u003c/strong\u003e1129-1144.\u003c/li\u003e\n\u003cli\u003eBen-Ari Y. Excitatory actions of gaba during development: the nature of the nurture. \u003cem\u003eNat Rev Neurosci\u003c/em\u003e 2002; \u003cstrong\u003e3\u003c/strong\u003e(9)\u003cstrong\u003e: \u003c/strong\u003e728-739.\u003c/li\u003e\n\u003cli\u003eHe Q, Nomura T, Xu J, Contractor A. The developmental switch in GABA polarity is delayed in fragile X mice. \u003cem\u003eJ Neurosci\u003c/em\u003e 2014; \u003cstrong\u003e34\u003c/strong\u003e(2)\u003cstrong\u003e: \u003c/strong\u003e446-450.\u003c/li\u003e\n\u003cli\u003eTang X, Jaenisch R, Sur M. The role of GABAergic signalling in neurodevelopmental disorders. \u003cem\u003eNat Rev Neurosci\u003c/em\u003e 2021; \u003cstrong\u003e22\u003c/strong\u003e(5)\u003cstrong\u003e: \u003c/strong\u003e290-307.\u003c/li\u003e\n\u003cli\u003ePeerboom C, Wierenga CJ. The postnatal GABA shift: A developmental perspective. \u003cem\u003eNeurosci Biobehav Rev\u003c/em\u003e 2021; \u003cstrong\u003e124: \u003c/strong\u003e179-192.\u003c/li\u003e\n\u003cli\u003eBen-Ari Y. Neuro-archaeology: pre-symptomatic architecture and signature of neurological disorders. \u003cem\u003eTrends Neurosci\u003c/em\u003e 2008; \u003cstrong\u003e31\u003c/strong\u003e(12)\u003cstrong\u003e: \u003c/strong\u003e626-636.\u003c/li\u003e\n\u003cli\u003eBourgeron T. From the genetic architecture to synaptic plasticity in autism spectrum disorder. \u003cem\u003eNat Rev Neurosci\u003c/em\u003e 2015; \u003cstrong\u003e16\u003c/strong\u003e(9)\u003cstrong\u003e: \u003c/strong\u003e551-563.\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"molecular-psychiatry","isNatureJournal":false,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"mp","sideBox":"Learn more about [Molecular Psychiatry](http://www.nature.com/mp/)","snPcode":"41380","submissionUrl":"https://mts-mp.nature.com/cgi-bin/main.plex","title":"Molecular Psychiatry","twitterHandle":"@molpsychiatry","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"ejp","reportingPortfolio":"Nature AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"Developmental and epileptic encephalopathy, SNARE-opathy, epistasis, epilepsy, cognitive deficit","lastPublishedDoi":"10.21203/rs.3.rs-6440830/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-6440830/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eBrain disorders caused by large effect mutations in single genes often present unexplained large phenotypic diversity, even among carriers of the same mutation. Here we examined genetic interactions as a possible explanation for this diversity for SNAREopathies, a group of neurodevelopmental disorders caused by \u003cem\u003ede novo\u003c/em\u003e genetic variation in genes that together drive secretion of chemical signals in the brain. SNAREopathies are characterized by a striking phenotypic diversity, including different types/degrees or absence of seizures, developmental delay and intellectual disability. First, we present and test a theoretical framework predicting that large phenotypic diversity is caused by non-linear genetic interactions between two or more functionally related genes. Second, we test this prediction in validated SNAREopathy mouse models by analyzing phenotypic diversity at the synaptic, network, system and behavioral level in single versus double mutants for SNAREopathy genes \u003cem\u003eStxbp1\u003c/em\u003e and \u003cem\u003eSnap25\u003c/em\u003e. Whereas single mutants all showed similar EEG- and motor abnormalities, but no overt seizures, as reported before, double mutants exhibited extreme diversity in seizure phenotypes. Some mice had lethal generalized seizures, frequent and complex epileptiform EEG activity and thalamic hyper-excitability as indicated by increased cFos staining, while other mice of the same genotype showed no detectable abnormalities, no increased cFos staining and a normal life span. The surviving double mutant mice showed phenotypes not more severe than single mutants at the synaptic, network, and behavioral level. Taken together, this study shows that haploinsufficiency at two interacting loci leads to extreme phenotypic diversity at the systems level, but not at the cellular level. These findings provide a proof of concept for how modifying genes in the patient genome may enhance phenotypic diversity.\u003c/p\u003e","manuscriptTitle":"Phenotypic diversity is caused by non-linear genetic interactions between two SNAREopathy genes","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-05-07 07:07:03","doi":"10.21203/rs.3.rs-6440830/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"revise","date":"2025-08-14T07:05:32+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"This content is not available.","date":"2025-05-27T19:33:57+00:00","index":1,"fulltext":"This content is not available."},{"type":"editorInvitedReview","content":"This content is not available.","date":"2025-05-27T12:20:51+00:00","index":4,"fulltext":"This content is not available."},{"type":"reviewerAgreed","content":"This content is not available.","date":"2025-05-16T09:10:58+00:00","index":5,"fulltext":"This content is not available."},{"type":"reviewerAgreed","content":"This content is not available.","date":"2025-05-14T11:34:50+00:00","index":4,"fulltext":"This content is not available."},{"type":"reviewerAgreed","content":"This content is not available.","date":"2025-05-12T16:56:31+00:00","index":3,"fulltext":"This content is not available."},{"type":"reviewerAgreed","content":"This content is not available.","date":"2025-04-30T21:20:24+00:00","index":2,"fulltext":"This content is not available."},{"type":"reviewerAgreed","content":"This content is not available.","date":"2025-04-28T14:53:11+00:00","index":1,"fulltext":"This content is not available."},{"type":"reviewersInvited","content":"","date":"2025-04-28T14:50:46+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2025-04-16T10:27:03+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-04-16T10:23:18+00:00","index":"","fulltext":""},{"type":"submitted","content":"Molecular Psychiatry","date":"2025-04-15T21:07:41+00:00","index":"","fulltext":""},{"type":"checksFailed","content":"","date":"2025-04-14T13:39:31+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"molecular-psychiatry","isNatureJournal":false,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"mp","sideBox":"Learn more about [Molecular Psychiatry](http://www.nature.com/mp/)","snPcode":"41380","submissionUrl":"https://mts-mp.nature.com/cgi-bin/main.plex","title":"Molecular Psychiatry","twitterHandle":"@molpsychiatry","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"ejp","reportingPortfolio":"Nature AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"e93dd59b-0253-4d3d-a2f8-653594699b4a","owner":[],"postedDate":"May 7th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[{"id":47799720,"name":"Biological sciences/Genetics"},{"id":47799721,"name":"Biological sciences/Neuroscience"}],"tags":[],"updatedAt":"2026-03-05T08:13:05+00:00","versionOfRecord":{"articleIdentity":"rs-6440830","link":"https://doi.org/10.1038/s41380-026-03509-3","journal":{"identity":"molecular-psychiatry","isVorOnly":false,"title":"Molecular Psychiatry"},"publishedOn":"2026-03-05 05:00:00","publishedOnDateReadable":"March 5th, 2026"},"versionCreatedAt":"2025-05-07 07:07:03","video":"","vorDoi":"10.1038/s41380-026-03509-3","vorDoiUrl":"https://doi.org/10.1038/s41380-026-03509-3","workflowStages":[]},"version":"v1","identity":"rs-6440830","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-6440830","identity":"rs-6440830","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}