Characteristic Features and Alterations of Sleep Patterns in Pigv^341E Mice.

The most prominent differences that we observed first between Pigv^341E and wild-type (WT) mice were reduced weight (Fig. 1B and SI Appendix, Fig. S1F) and hindlimb clasping behavior (Fig. 1 C and D). Due to the intellectual disability and psychomotor delay that are the key clinical features of IGD, patients are impaired in their everyday lives. Therefore, we sought to determine which spontaneous behaviors our mouse model exhibited while living undisturbed in their home cage (singly or group housed). Using HomeCageScan (HCS), we monitored singly housed Pigv^341E mice for 23 h at two different time points (8 and 16 wk). Among the 19 behaviors accurately detected by analysis of the HCS data, three behaviors were consistently altered in Pigv^341E mice at both time points. Pigv^341E mice hung less often to the top of the cage (total occurrences) and for shorter durations (total duration and duration per hour) than wild-type mice; groomed more often (total occurrences) and for longer durations (total duration and duration per hour); and slept less (total duration and duration per hour) (Fig. 2 A and B and SI Appendix, Fig. S2A). At the earlier time point, Pigv^341E mice spent more time walking (duration per hour) during the dark phase of the day than wild-type mice (Fig. 2B, Top graph). Furthermore, at both time points, hanging behavior was an important variable for differentiating genotypes along the dimensions of a principal component analysis (PCA) (SI Appendix, Figs. S3 A–F and S4 A–F).

Fig. 2.

Pigv^341E mice exhibit a behavioral phenotype in HCS and an elevated activity level in the SAM (second approach). (A) Duration of selected behaviors from the HCS. Pigv^341E mice revealed elevated grooming and reduced hanging and sleeping behavior at both time points. (B) Duration of selected behaviors per hour, as determined by HCS. The dark phase took place between the two dashed vertical lines. Pigv^341E mice exhibited elevated grooming and hanging and reduced sleeping at both time points. Pigv^341E mice exhibited elevated walking at time point 1. (C) The SAM revealed an elevation in total distance traveled by Pigv^341E mice when animals were separated by genotype (nonmixed). When housed in mixed genotypes (mixed), total distance traveled did not significantly differ between genotypes. (D) Pigv^341E mice exhibited an increase in distance traveled per phase when separated by genotype (nonmixed). When housed in mixed genotype (mixed), total distance traveled per phase did not significantly differ between genotypes. (E and F) Pigv^341E mice traveled longer distances per day and per hour when separated by genotype (nonmixed). Pigv^341E = homozygous for Pigv p.Ala341Glu. Animals used for the HCS (WT: female n = 8, male n = 4; Pigv^341E: female n = 4, male n = 2) were phenotyped at two different consecutive time points: tp1 at the age of 8 wk, and tp2 at the age of 16 wk. Animals used for the SAM: WT(female n = 4, male n = 6) Pigv^341E(female n = 2, male n = 5). Data from the HCS (total duration) and SAM (total distance, mean distance traveled per phase and per day) were analyzed by Wilcoxon rank-sum test (nonparametric). Data from the HCS (duration per hour) and SAM (distance traveled per hour) were analyzed with a glmm using a MCMC. *P < 0.05.

We also used the social activity monitor (SAM) to assess spontaneous home-cage activity in Pigv^341E mice while living in a group setting. For this test, we implanted a radio-frequency identification (RFID) transponder into the mice and put the home cage with mixed genotypes on a grid box that could locate individual animals and their position in the cage at all times (continuous 24 h/day recording). Because the animals were undisturbed in their home cage, SAM analysis could be performed several times without the animals noticing. For the first two time points (9 and 17 wk), SAM analysis revealed no difference between genotypes in total distance traveled over 14 d (SI Appendix, Fig. S4G), but an unexpected switch in diurnal/nocturnal activity for both genotypes was observed. The mice were more active during the light (normal sleeping time) than the dark phases of their days (SI Appendix, Fig. S5 A and C; confirmed by Markov chain Monte Carlo generalized linear mixed-effects models [MCMCglmm], pMCMC = 0.001). This observation supported the HCS data that Pigv^341E mice slept less, which is a feature of some patients with IGDs who suffer from sleep disturbances (7, 20). We performed a second experiment in which we evaluated the spontaneous activity of the Pigv^341E and wild-type mice separately (nonmixed vs. mixed-genotype cages). When Pigv^341E and wild-type mice lived together in the same cage (mixed genotype), we reproduced our previous results: Mice of different genotypes exhibited no difference in total distance traveled over 4 d (Fig. 2C, Left graph), per phase (Fig. 2D, Left graph), per day (SI Appendix, Fig. S2C), and per hour (SI Appendix, Fig. S2D), but they still exhibited higher activity levels (larger variability) during the light vs. dark phase (SI Appendix, Fig. S2D, confirmed by MCMCglmm, pMCMC = 0.001). However, in nonmixed housing conditions Pigv^341E mice walked greater distances over 4 d (Fig. 2C, Right graph), per phase (Fig. 2D, Right graph), per day, and per hour (Fig. 2 E and F) than wild-type mice, and again, Pigv^341E mice were more active during the light than the dark cycle on days 1 and 2 (Fig. 2E).

Pigv^341E Mice Exhibit Motor Dysfunction and Alterations in Sociability.

Extensive testing of various motor functions in Pigv^341E mice revealed a clear and elaborated dysfunctional motor phenotype. Pigv^341E mice had reduced balance and motor coordination, reflected by a reduced latency of falling off the rotarod (Fig. 3A). This was confirmed by an elevated latency of traversing an elevated beam (Fig. 3B). In the rope grip test, Pigv^341E mice exhibited an elevated latency of climbing on the rope, and had a lower hanging score than wild-type mice (Fig. 3C and SI Appendix, Fig. S6A, Right graph). Furthermore, Pigv^341E mice had reduced grip strength (Fig. 3D). Next, we evaluated the walking pattern of Pigv^341E mice using the footprint test (Fig. 3E). Pigv^341E mice had a larger distance between forepaw and hindpaw (S) in paw placement of the stride (Fig. 3F); this is remarkable because walking mice usually place their hind paws in the same positions as their forepaws (Fig. 3E, see WT).

Fig. 3.

Pigv^341E mice exhibit motor dysfunctions. (A and B) Pigv^341E mice exhibited a reduction in motor coordination, reflected by a reduced latency of falling off the rotarod and elevated latency of traversal of the beam (diameter, 15 mm). (C) Pigv^341E mice exhibited a reduction in climbing performance, reflected by a decrease in the time spent climbing the rope. (D) Pigv^341E mice exhibited a reduced grip strength, as determined by grip strength meter. [g] was normalized to the weight of the animal. (E) Representative image of walking pattern between genotypes. SL–FL = stride length–forelimb, SL–HL = stride length–hindlimb, FBW = fore–base width, HBW = hind–base width, S = distance between forepaw and hindpaw. (F) Pigv^341E mice had an altered walking pattern, reflected by an increase in the distance between forepaw to hindpaw and a decrease in fore–base width. (G) Pigv^341E mice exhibited enhanced social approach behavior, reflected by a decrease in “rear up” behavior and an increase in the number of nose-to-anogenital contacts in the social proximity test. (H) The three-chamber test (first 5 min) indicated that Pigv^341E mice exhibited an enhanced social approach behavior relative to wild-type mice, reflected by spending more time with the stranger mouse than with the empty. Pigv^341E = homozygous for Pigv p.Ala341Glu. Animals used in the motor tests were 7 wk old: WT(female n = 8, male n = 4) Pigv^341E(female n = 4, male n = 2). Animals used in the social proximity test: WT(female n = 9, male n = 11), Pigv^341E(female n = 4, male n = 7). Animals used in the three-chamber test: WT(female n = 9, male n = 11), Pigv^341E(female n = 4, male n = 6). The data from the motor tests and the social proximity test were analyzed by nonparametric t test (Mann–Whitney). The data from the three-chamber test were analyzed by two-way ANOVA followed by Bonferroni’s multiple comparisons test. *P < 0.05, **P < 0.01, ***P < 0.001. ns, not significant.

We evaluated social behavior in Pigv^341E mice in the social proximity and three-chamber tests. In the social proximity test, Pigv^341E mice exhibited a reduced number of “rear up” behaviors and an elevated number of nose-to-anogenital contacts with the stranger mouse (Fig. 3G). However, no differences in the number of nose-tip-to-nose-tip, nose-to-head-contact, “crawl over,” or “crawl under” behaviors were observed between genotypes (SI Appendix, Fig. S7A). In the three-chamber test, Pigv^341E mice spent more time with the stranger mouse than in the vicinity of the empty cage (Fig. 3H). Furthermore, the discrimination ratio (stranger vs. empty cage) was higher in Pigv^341E than in wild-type mice (SI Appendix, Fig. S7D, Right graph).

Pigv^341E Mice Exhibit Cognitive Deficits in Spatial Long-Term Memory and Species-Specific Hippocampus-Dependent Functions.

To characterize the cognitive and affective profile of Pigv^341E mice, we performed a battery of tests to assess aspects of spatial learning and memory, species-specific functions (Barnes maze, y-maze, marble burying, and nest building behavior), and spontaneous response to novel, open, and elevated or bright environments (open field, elevated plus maze, and dark/light box). In the Barnes maze test, Pigv^341E mice were delayed in spatial learning, as indicated by an elevated latency to escape during days 1 to 3 (Fig. 4A). Despite this delay in spatial learning during days 1 to 3, Pigv^341E mice had learned the location of the escape box by day 4 and exhibited normal short-term spatial memory at day 5. However, Pigv^341E mice had impaired long-term spatial memory at day 12, reflected by an elevated latency to escape (Fig. 4A). Similar to the results that measured latency to escape, Pigv^341E mice had an elevated path length for days 1 to 4 and 12, but not day 5 (Fig. 4B). Furthermore, Pigv^341E mice spent less time in the target quadrant than wild-type mice at day 12, confirming deficits in long-term spatial memory (SI Appendix, Fig. S8B, Right graph). In the y-maze test, similar spontaneous alternation behavior between genotypes suggested normal short-term spatial working memory in Pigv^341E mice (SI Appendix, Fig. S8A). In the species-specific and hippocampus-dependent tests of marble burying and nest construction (21), Pigv^341E mice buried fewer marbles and had lower-quality nests (Fig. 4 C and D).

Fig. 4.

Pigv^341E mice exhibit cognitive deficits in learning and memory. (A and B) Pigv^341E mice exhibited cognitive deficits in learning and long-term memory, reflected by increases in latency to escape and path length on days 1 to 3 (learning) and day 12 (long-term memory in the Barnes maze). (C) Pigv^341E mice exhibited a decrease in burying and nest construction behavior. (D) Representative image of burrowing behavior (Left) and nest construction behavior (Right) of wild-type and Pigv^341E mice. Pigv^341E = homozygous for Pigv p.Ala341Glu. Animals used in the Barnes maze: WT(female n = 8, male n = 11), Pigv^341E(female n = 4, male n = 6). Animals used in the marble-burying test: WT(female n = 9, male n = 11), Pigv^341E(female n = 4, male n = 7). Animals used in the nest construction test were 7 wk old: WT(female n = 8, male n = 4), Pigv^341E(female n = 4, male n = 2). The data from the nest construction and marble-burying tests were analyzed with a nonparametric t test (Mann–Whitney). The data from the Barnes maze (latency to escape, path length) were analyzed by two-way ANOVA followed by Bonferroni’s multiple comparisons test. *P < 0.05 **P < 0.01, ***P < 0.001. ns, not significant.

Pigv^341E mice behaved similarly to wild-type mice in the dark/light box and elevated plus maze (SI Appendix, Fig. S9 A and B). In the open field test, Pigv^341E mice spent less time in the center and more time in the periphery than wild-type mice (SI Appendix, Fig. S9C). Moreover, the path length and number of visits to the center were reduced in Pigv^341E mice (SI Appendix, Fig. S9 D and E).

Pigv^341E Mice Exhibit Defects in Synaptic Transmission.

Because Pigv^341E mice exhibited cognitive impairments in spatial learning and memory, we hypothesized a hippocampal defect, and this idea was supported by the reduced burrowing and nest building behavior (hippocampus-dependent) in Pigv^341E mice. Hence, we decided to analyze the hippocampus in more detail. Nissl staining revealed no morphological abnormalities in the hippocampus of Pigv^341E mice (SI Appendix, Fig. S10C). Because many GPI-linked proteins play crucial roles in synapse formation and plasticity, causing deficits that can be detected in cell culture (22), we performed immunostaining to visualize synaptophysin, a presynaptic marker. We observed a decreased immunoreactivity for synaptophysin in cornu ammonis 1–stratum radiatum (CA1–SR) of Pigv^341E mice (Fig. 5A). By contrast, we observed no significant differences between genotypes in cornu ammonis 3–stratum radiatum (CA3–SR) or cornu ammonis 1–molecular layer of dentate gyrus (CA1–ML) (SI Appendix, Fig. S10 A and B). To determine whether the behavioral abnormalities were accompanied by deficits in synaptic transmission, we conducted electrophysiology recordings in the CA1–SR region, where synaptophysin levels were reduced in Pigv^341E mice. We tested input–output functions and observed lower amplitudes of excitatory postsynaptic potential (EPSPs) at different stimulation intensities in Pigv^341E mice (Fig. 5B). In paired pulse facilitation (PPF), Pigv^341E mice exhibited an elevated paired pulse ratio (PPR) (Fig. 5C). Moreover, posttetanic potentiation (PTP) exhibited elevated facilitation in Pigv^341E mice (Fig. 5D).

Fig. 5.

Pigv^341E exhibit a hippocampal synaptic defect. (A) Representative images of immunofluorescence staining show less synaptophysin immunoreactivity (green) in CA1–SR of Pigv^341E mice. 4′,6-diamidino-2-phenylindole (DAPI) staining is shown in blue. (Scale bar: 10 µm.) Synaptophysin immunoreactivity per µm², quantified using FIJI ImageJ, was significantly reduced in CA1–SR in Pigv^341E mice (Right). (B) Input–output: EPSP amplitude was reduced in Pigv^341E mice at different presynaptic fiber volleys (PFVs). (Scale bar: L indicates 0.2 mV/5 ms.) (C) PPF revealed an elevated PPR in Pigv^341E mice. (Scale bar: L indicates 0.2 mV/10 ms.) (D) PTP was elevated in Pigv^341E mice. (Scale bar: L indicates 0.2 mV/5 ms.) (E) In the PTZ kindling model Pigv^341E mice had a significantly different seizure threshold compared to that of wild-type control littermates (Left graph). In Pigv^341E mice the latency to the first convulsive seizure was significantly reduced (Middle graph). Furthermore, the seizure phenotype appears more severe (Right graph). Pigv^341E = homozygous for Pigv p.Ala341Glu. Animals used for the synaptophysin-staining were 3 wk old: WT(female n = 3, male n = 2) Pigv^341E(female n = 2, male n = 4). Animals used for the electrophysiology recordings: WT(female n = 2, male n = 2) Pigv^341E(female n = 2, male n = 2). Animals used for the PTZ model: WT(female n = 3, male n = 9) Pigv^341E(female n = 5, male n = 6). The data from the synaptophysin immunofluorescence staining and electrophysiology recordings (PPF, PTP) were analyzed by parametric Student’s t test. Furthermore, the data from the input–output curve were analyzed by two-way ANOVA followed by Bonferroni’s multiple comparisons test. In the PTZ model, group comparisons (dose-dependent distribution, severity of seizures) were analyzed by χ² test and latency to first seizure by unpaired Student’s t test with Welch’s correction. *P < 0.05 **P < 0.01, ***P < 0.001.

Increased Susceptibility of Pigv^341E Mice to Chemically Induced Acute Seizures.

Considering the altered electrophysiological properties in Pigv^341E mice, we examined for differences in the seizure threshold between Pigv^341E and corresponding littermate wild-type mice in an acute epilepsy model, the pentylenetetrazole (PTZ)-induced kindling model (23). To this end, mice were repetitively exposed to PTZ (10 mg/kg, i.p.) every 10 min until the occurrence of a first focal-to-bilateral tonic–clonic seizure occurred. Intriguingly, Pigv^341E mice exhibited a significantly lower seizure threshold, manifesting as a reduced latency to first seizure than wild-type mice, which exhibited convulsive seizures after PTZ exposure. Wild-type animals manifested generalized seizures after 93.3 min, whereas in Pigv^341E mice, the first seizure manifested already after 63.3 min (Fig. 5E, Middle graph). Furthermore, the observation that four wild-type animals did not exhibit any seizure after 10 injections, whereas all Pigv^341E mice did, underlines the higher susceptibility of Pigv^341E mice for seizure induction via PTZ (χ² = 105.3, df = 4, P < 0.0001) (Fig. 5E, Left graph). In addition, Pigv^341E mice exhibited more severe seizures than wild-type mice (χ² = 74.21, df = 1, P < 0.0001) (Fig. 5E, Right graph).

Shift in Relative Cell Count of Hippocampal Cellular Subgroups in Pigv^341E Mice.

The synaptic defect in the CA1–SR region of Pigv^341E mice could have been responsible for the observed impairments in spatial learning and memory. To identify the cell types most affected by GPIBD, we performed single-cell RNA sequencing on freshly isolated hippocampal cells after cognitive behavioral tests. The isolation of hippocampal cells was performed as previously described (24). Based on the gene expression profiles of 8,800 single cells from Pigv^341E mice and 7,100 cells from wild-type animals, we defined 17 cellular subgroups (Fig. 6A). Cells from both Pigv^341E and wild-type mice were present in all subgroups, but the distributions differed between genotypes (χ² = 306.49, df = 16, P < 2.2 × 10⁻¹⁶). While the fractions of granule cells, oligodendrocytes, and a microglia subpopulation (microglia 3) were reduced in pooled samples from Pigv^341E mice, the proportions of subicular neurons (neurons subiculum 1), GABAergic (inhibitory) interneurons, and fibroblast-like cells were higher than in the pooled wild-type samples (Fig. 6B).

Fig. 6.

Differences between Pigv^341E and wild-type mice in the transcriptional landscape of single hippocampus cells. (A) Hippocampal cells from both Pigv^341E and wild-type mice clustered into 17 cellular subgroups (B), with a skewed distribution of cells across subgroups in the Pigv^341E sample relative to wild-type. Percentages below 3% are not shown for clarity. (C) Differential expression testing between Pigv^341E and wild-type cells within each cellular subgroup yielded differentially expressed genes with absolute log₂(fold change) ≥0.25 and adjusted P value ≤0.01. The most extensive changes in gene expression were observed for the third subgroup of microglia cells and the first subgroup of subicular neurons. Gene counts are not shown for sets with <20 differentially expressed genes. (D) Global comparison between all Pigv^341E and wild-type cells highlights the hippocampus-wide effect on the genes Abl1, Hdc, Cyp4x1, and Gm14216, for which significant changes in expression were detected within each cellular subgroup. The dashed horizontal line is located at an adjusted P value of 0.01, the dashed vertical lines at an absolute log₂(fold change) of 0.25, and the dotted vertical lines at an absolute log₂(fold change) of 0.7. For genes with an absolute log₂(fold change) >0.7, gene symbols are shown. Within the subgroups of microglia 3, neurons subiculum 1 and pyramidal neurons (CA3), GO enrichment analysis revealed a number of overrepresented biological process (E), cellular component (F), and molecular function terms (G) in the sets of genes differentially expressed between Pigv^341E and wild-type cells within the cellular subgroups. For each set of genes that were up-regulated or down-regulated between genotypes, up to three terms with the lowest adjusted P value after removal of redundant GO terms are shown. The gene ratio, i.e., the number of genes annotated with a term divided by the respective number of differentially expressed genes, is encoded by point size. Cellular subgroups with no significantly enriched terms were omitted. GO terms enriched in differentially expressed genes with elevated or reduced expression within a subgroup are contracted into a single star-shaped symbol. OPC, oligodendrocyte precursor cells; DE genes, differentially expressed genes; FC, fold change; P_adj, adjusted P value. Pooled samples [WT(male = 4), Pigv^341E(male = 4)]. n = 1 per genotype.

Pigv^341E Hippocampal Cells Exhibit a Deregulation in Gene Expression Related to Synapse Organization and Signaling Transduction.

Expression analysis of single-cell RNA sequencing data revealed multiple genes that were differentially expressed between Pigv^341E and wild-type cells within each cellular subgroup (Fig. 6C). In the mutant mice, nonreceptor tyrosine kinase Abl1 was down-regulated, whereas histidine decarboxylase Hdc, cytochrome P450 member Cyp4x1, and predicted lncRNA Gm14216 were up-regulated, both within and across all cellular subgroups (Fig. 6D and SI Appendix, Figs. S16–S18). The most extensive change in gene expression within a cellular subgroup was observed in the first subgroup of subicular neurons and the third subgroup of microglia (Fig. 6C and SI Appendix, Figs. S16 and S17). We also performed Gene Ontology (GO) analysis of differentially expressed genes between genotypes, including terms in three categories: biological process, cellular component, and molecular function (Fig. 6 E–G). Among the differentially expressed genes in subicular neurons, the biological process term “regulation of synapse organization” was enriched in genes up-regulated in Pigv^341E cells, whereas the biological process terms “cell morphogenesis” and “commissural neuron axon guidance” were enriched in down-regulated genes (Fig. 6E). Furthermore, the biological process term “cell–cell adhesion via plasma-membrane adhesion molecules” was enriched in genes with elevated and reduced expression (Fig. 6E).

Single-cell RNA sequencing revealed three hippocampal microglia populations in both genotypes (Fig. 6 A and B, microglia 1 to 3). All three subpopulations expressed the marker genes Csf1r and C1qa (SI Appendix, Fig. S15). Among the most significant differentially expressed genes between the three microglia subgroups and all remaining cells in both genotypes, “cell activation,” “migration,” “phagocytosis,” and “immune responses” were among the top 10 GO biological process terms in microglia 1 and 2 (Dataset S2). By contrast, the top 10 GO biological process terms in microglia 3 cells were “ribosome,” “ribonucleoprotein complex biogenesis,” and “cytoplasmic translation” (Dataset S2). Hence, we considered microglia 1 and 2 cells as potentially more phagocytic and migratory than microglia 3 cells. We identified 326 genes that were differentially expressed between genotypes in microglia 3 cells. Remarkably, 306 of these 326 genes were down-regulated in microglia 3 cells of Pigv^341E mutants (Fig. 6C). GO analysis revealed that the down-regulated genes were enriched for the biological process terms “small GTPase-mediated signal transduction” and “regulation of microtubule cytoskeleton polymerization and depolymerization” (Fig. 6E). In addition, we identified 20 genes differentially expressed between genotypes that were up-regulated in Pigv^341E microglia 3 cells. Among these were Rpl38, Rps21, and Rps28, which encode ribosomal proteins.

CA3 pyramidal neurons project their axons to the CA1 region, where we observed dysfunction in synaptic transmission in Pigv^341E mice. Therefore, we were particularly interested in the cellular subgroup of CA3 pyramidal neurons, in which we identified 79 genes that were differentially expressed between genotypes (Fig. 6C). GO analysis revealed that down-regulated genes were enriched for the terms “adherens junction organization,” “regulation of cell-substrate junction assembly,” and “focal adhesion assembly” in Pigv^341E CA3 pyramidal neurons (Fig. 6E). Interestingly, we also observed enrichment of genes with reduced expression related to “neuron–neuron synaptic transmission,” “regulation of synaptic vesicle exocytosis,” and “synaptic vesicle transport” (Dataset S3).

Animals.

Pigv^341E mice were generated by diploid or tetraploid aggregation (40) (SI Appendix, Fig. S1A) and maintained by crossing with C57BL.6/J mice. Mice were genotyped by PCR using the primers mPigvEx4_fw and mPigvEx4_rv. PCR amplicons were digested with BcuI (Thermo Fisher Scientific) and subjected to agarose gel electrophoresis (SI Appendix, Fig. S1D). All animals were handled according to government regulations as approved by local authorities (LaGeSo Berlin and LANUV Recklinghausen). In addition, all experiments were carried out following the 3R (Replacement, Reduction and Refinement) guidelines for animal welfare. Mice were housed in groups with mixed genotypes in single ventilated cages with an enriched environment. The mice were housed in a pathogen-free animal facility with a 12 h dark/light cycle and had food and water ad libitum unless otherwise indicated. Mice used for experiments were 8 wk to 6 mo old unless otherwise indicated. Pigv^341E and wild-type mice used in a given experiment were the same age. To avoid bias effects, littermates were assigned equally to both experimental groups, according to their genotype and sex. The experimenter was blinded except during behavioral testing, as Pigv^341E mice were physically smaller. Moreover, experiments were randomized with respect to mouse genotype.

Statistical Analysis.

For all experiments, at least four animals per genotype were used, except for the weight curve, for which at least three animals per genotype were used. One animal was defined as one biological replicate and represented one data point, except for electrophysiology recordings (see SI Appendix, Schaffer collateral recordings). Means and SDs were calculated for each genotype group unless otherwise indicated. Data were statistically analyzed with GraphPad Prism (version 7) or R (version 3.6.3) (41), and results are expressed as means ± SDs. Statistical tests were performed for each experiment as indicated in Dataset S1. Results with P value <0.05 were considered significant unless otherwise indicated.