CDKL5 Deficiency Augments Inhibitory Input into the Dentate Gyrus That Can Be Reversed by Deep Brain Stimulation

Shuang Hao; Qi Wang; Bin Tang; Zhenyu Wu; Tingting Yang; Jianrong Tang

doi:10.1523/JNEUROSCI.1010-21.2021

Abstract

Cognitive impairment is a core feature of cyclin-dependent kinase-like 5 (CDKL5) deficiency, a neurodevelopmental disorder characterized by early epileptic seizures, intellectual disability, and autistic behaviors. Although loss of CDKL5 affects a number of molecular pathways, very little has been discovered about the physiological effects of these changes on the neural circuitry. We therefore studied synaptic plasticity and local circuit activity in the dentate gyrus of both Cdkl5^−/y and Cdkl5^+/− mutant mice. We found that CDKL5 haploinsufficiency in both male and female mice impairs hippocampus-dependent learning and memory in multiple tasks. In vivo, loss of CDKL5 reduced LTP of the perforant path to the dentate gyrus and augmented feedforward inhibition in this pathway; ex vivo experiments confirmed that excitatory/inhibitory input into the dentate gyrus is skewed toward inhibition. Injecting the GABAergic antagonist gabazine into the dentate improved contextual fear memory in Cdkl5^−/y mice. Finally, chronic forniceal deep brain stimulation rescued hippocampal memory deficits, restored synaptic plasticity, and relieved feedforward inhibition in Cdkl5^+/− mice. These results indicate that CDKL5 is important for maintaining proper dentate excitatory/inhibitory balance, with consequences for hippocampal memory.

SIGNIFICANCE STATEMENT Cognitive impairment is a core feature of cyclin-dependent kinase-like 5 (CDKL5) deficiency disorder. Although CDKL5 deficiency has been found to affect a number of molecular pathways, little is known about its physiological effects on the neural circuitry. We find that CDKL5 loss reduces hippocampal synaptic plasticity and augments feedforward inhibition in the perforant path to the dentate gyrus in vivo in Cdkl5 mutant mice. Chronic forniceal deep brain stimulation rescued hippocampal memory deficits, restored synaptic plasticity, and relieved feedforward inhibition in Cdkl5^+/− mice, as it had previously done with Rett syndrome mice, suggesting that such stimulation may be useful for other neurodevelopmental disorders.

Introduction

CDKL5 deficiency disorder (CDD) is a neurodevelopmental disease caused by mutation of the X-linked gene cyclin-dependent kinase-like 5 (CDKL5). CDD is characterized by severe intellectual disability, gross motor impairment, early-onset epilepsy, and autistic features (Kalscheuer et al., 2003; Weaving et al., 2004). Mouse models of CDD faithfully recapitulate the cognitive deficits and other characteristics of this condition (Wang et al., 2012; Fuchs et al., 2015; Okuda et al., 2018). Global KO/knock-in mutation of CDKL5 or conditional CDKL5 loss in forebrain excitatory neurons impairs hippocampus-dependent memory in mice (Wang et al., 2012; Trazzi et al., 2016; Tang et al., 2017; Okuda et al., 2018; Yennawar et al., 2019), but it is not known how CDKL5 loss affects the functioning of the hippocampal circuit.

Only a few studies have addressed the influence of CDKL5 on synaptic function in the CA1 region of the hippocampus. Deleting CDKL5 from forebrain glutamatergic or GABAergic neurons affects synaptic transmission and the overall excitatory/inhibitory (E/I) balance in the CA1 pyramidal neurons (Tang et al., 2017, 2019). Yet in slice preparations, global CDKL5 loss enhances LTP of synaptic transmission in the Shaffer collateral to CA1 pathway (Okuda et al., 2017; Yennawar et al., 2019). Another key component of the hippocampal formation for learning and memory is the dentate gyrus (DG) (Liu et al., 2012; Hainmueller and Bartos, 2020). The DG is crucial for gating/processing contextual, place, and spatial information from the entorhinal cortex to the hippocampus (Hargreaves et al., 2005). It is responsible for important cognitive functions, such as pattern separation (Leutgeb et al., 2007; McHugh et al., 2007), pattern completion (Nakashiba et al., 2012), and spatial and working memory (Lee and Jung, 2017; Sasaki et al., 2018). Because synaptic transmission and plasticity of the perforant path (PP) to the DG play an important role in hippocampus-dependent learning and memory (I. W. Jones et al., 2001; McHugh et al., 2007), we hypothesized that loss of CDKL5 causes memory deficits by specifically impairing this pathway.

We therefore set out to delineate the precise impairments of the hippocampal circuit in a mouse model of CDD. We used behavioral, in vivo and ex vivo approaches to discover that increased feedforward inhibition, mediated by molecular layer perforant pathway (MOPP) cells (Li et al., 2013), is the chief neural circuit mechanism that underlies memory dysfunction. Given that deep brain stimulation (DBS) effectively modulates neural circuit activity (Bouthour et al., 2019; Lozano et al., 2019), we also tested whether forniceal DBS can relieve the feedforward inhibition and rescue hippocampal memory and synaptic plasticity deficits in Cdkl5 mutant mice.

Materials and Methods

Animals

The CDKL5 mouse line was obtained from The Jackson Laboratory and maintained on the C57BL/6J background (B6.129(FVB)-Cdkl5^tm1.1Joez/J, stock #021967) (Wang et al., 2012). Adult male Cdkl5^−/y (4 months of age) and female Cdkl5^+/− (4 or 6 months of age) and age-/gender-matched WT littermates (Cdkl5^+/y and Cdkl5^+/+) were housed in a 14 h light/10 h dark cycle (light on at 6:00 A.M.) with standard mouse chow and water ad libitum in our on-site American Association for Laboratory Animal Science-accredited facility. The longer light cycle was used to mimic the Spring and Summer daylight hours to help with mouse breeding. Mice were group-housed up to 5 per cage without/before surgery but housed individually with nesting material in the cage after surgery in a room maintained at 22°C. All the experimental procedures and tests were conducted during the light phase. Behavioral, electrophysiological, and immunohistochemical assays were performed and analyzed blind to genotypes and/or treatments. All research and animal care procedures were approved by the Baylor College of Medicine Institutional Animal Care and Use Committee.

Behavioral assays

All behavioral studies were conducted blinded to genotype. Mice were allowed to habituate to the testing room for at least 30 min before any test, and testing was performed around the same time of day. Open field test was conducted 1 week before fear conditioning or water maze tasks. The light intensity of 150 lux and the background white noise at 60 dB were consistently presented during the habituation and throughout the testing periods as we previously reported (Hao et al., 2015). All animal behaviors were performed in male (4 months old) and female (4 or 6 months old) mice, respectively.

Open field

The open field apparatus consists of a clear, open Plexiglas box (40 × 40 × 30 cm, Stoelting) with overhead camera and photograph beams to record horizontal and vertical movements. Activity was tracked and quantified over a-30 min period by ANY-maze (Stoelting).

Fear conditioning

Fear conditioning was performed to evaluate hippocampus-dependent contextual fear memory and hippocampus-independent cued fear memory (Hao et al., 2015). Briefly, on day 0 animals were trained in a chamber with a grid floor that could deliver an electric shock (Med Associates). This enclosure was put in a sound-attenuating box that contained a digital camera, a loudspeaker, and a house light. Each mouse was placed in the chamber for 2 min during which mice were actively exploring the new environment, after which a tone (30 s, 5 kHz, 80 dB) coincided with a scrambled foot shock (2 s, 0.7 mA). The tone/foot-shock stimuli were repeated once after 1 min. After an additional 1 min, the mouse was removed and returned to its home cage. Fear memory retention was evaluated 24 h after training. The mice were first recorded for 5 min in the same chamber (fresh with 70% ethanol) without tone. Then, they were tested in a novel cage (cleaned with 1% acetic acid) where a 5-min tone was presented after the animals had acclimated to the cage for 5 min. Mouse behavior was recorded and scored automatically by ANY-maze tracking system (Stoelting). Freezing, defined as an absence of all movement except for respiration (Maren et al., 1994), was scored only if the animal was immobile for at least 1 s (Corcoran and Maren, 2001). The percentage of freezing time during the tests serves as an index of fear memory. Cued fear memory was the subtraction of freezing time between the tone phase and the no-tone phase. The preshock immobilization level was normalized from the 2-min baseline before the first foot shock. The postshock freezing level was calculated from the 1 min immediately after the second foot shock (see Fig. 1A). One week after the fear conditioning training, animals were placed back to the fear conditioning chamber for the test of pain threshold of the electric foot shock. Every 30 s, a 2-s scrambled electric foot shock with 0.05 mA increments (starting from 0 mA) was applied. The shock current thresholds of flinch, vocalization, and jumping responses were each recorded (Hao et al., 2015).

Passive avoidance

The passive avoidance assay evaluates associative learning and memory (Samaco et al., 2013). The test apparatus consists of a light and a dark compartment with a guillotine door, across which mice can pass. During the training session, mice were placed into the light compartment and their latency to entry the dark compartment was timed. Immediately after the mice entered the dark compartment with all four paws, the door was shut and an unavoidable foot shock (0.4 mA, 2 s) was delivered. Then the mice were removed to home cage after 10 s. After 24 h, each mouse was placed into the light compartment again and the latency to enter the dark compartment was recorded to measure the retention memory. The maximum cutoff latency was 300 s.

Morris water maze

The Morris water maze was used to assess spatial learning and memory (Moy et al., 2007; Hao et al., 2015). The pool (120 cm in diameter) was filled with water (50 cm deep, 22°C-24°C) made opaque with nontoxic white tempera paint. Visual cues were set on the wall of the testing room, at least 1 m from the pool edge. The ANY-maze tracking system (Stoelting) was used to track and analyze animal swimming. Mice were trained for their ability to find an escape platform (10 cm in diameter) on three different components of the test in turn: hidden platform acquisition, subsequent probe trial in the absence of the platform, and visible platform acquisition. In the hidden platform acquisition training, each animal was given four trials of swimming, a maximum of 60 s for each trial, to find a hidden escape platform cued by visual cues for 9 d. Twenty-four hours after the last training trial of day 9, all the mice were given a probe trial. During the probe trial, the platform was removed, and each animal was allowed 60 s to search in the pool. The amount of time that each animal spent in each quadrant was recorded, as well as the number of times a mouse crossed the imaging location of the platform during training (platform crossings). Another 24 h after the probe test, all the mice received 1-d visible platform training to swim to an escape platform with a visible cue on the platform (above the surface of the water).

Surgery and DBS

Similar to our previous report (Hao et al., 2015), female Cdkl5^+/− mice and WT littermates (4.5 months old) were secured on a stereotaxic frame (David Kopf) under 1%-2% isoflurane anesthesia. Bipolar DBS electrodes were constructed with Teflon-coated tungsten wire (bare diameter 50 µm, A-M Systems) and the two tips were horizontally separated by 0.1-0.15 mm. The electrodes were targeted unilaterally to the fimbria-fornix (0.20 mm posterior, 0.24 mm lateral, and 2.3 mm below the bregma) under the guidance of evoked potentials recorded in the ipsilateral DG (1.8-2.0 mm posterior, 1.4-1.6 mm lateral of bregma, 2.2-2.3 mm below the skull) (Paxinos and Franklin, 2001). All the electrodes together with the attached connector sockets were fixed on the skull by dental cement. Animals were given at least 2 weeks to recover.

After recovery, animals were assigned into four groups, respectively: Cdkl5^+/+-sham, Cdkl5^+/+-DBS, Cdkl5^+/−-sham, and Cdkl5^+/−-DBS. Mice were randomly assigned to DBS or sham groups within the same genotype. Animals in the 2 DBS groups received 1 h DBS daily for 14 consecutive days. The chronic DBS was biphasic rectangular pulses (130 Hz, 60 µs pulse duration), and the stimulus intensities were individually optimized to 80% of the threshold that elicits an after discharge in the hippocampus. Animals in the sham groups experienced the same procedures as those in the DBS groups except for DBS treatment. There was a 3-week interval between the last DBS administration and the beginning of fear conditioning, in vivo LTP, or in vivo feedforward/feedback inhibition test.

After finishing all experiments, mice were killed with an overdose of isoflurane. An anodal current (30 µA, 10 s) was passed through the electrode wire to verify the electrode placements. Frozen 30 µm coronal sections were cut and stained with cresyl violet for histologic examination.

In vivo electrophysiology

Surgery

Mice were secured on a stereotaxic frame (David Kopf) under 1%-2% isoflurane anesthesia. The stimulating and recording electrodes were surgically implanted as previously described (Tang and Dani, 2009; Hao et al., 2015) with the following modification. Under the guidance of evoked potentials recorded in the DG (1.8-2.0 mm posterior, 1.4-1.6 mm lateral of bregma, 2.2-2.3 mm below the skull), a concentric stimulating electrode was implanted ipsilaterally in the medial PP (0.2 mm posterior and 2.8-3.0 mm lateral of lambda, 1.0-1.3 mm below the dura). The final depth of the electrodes was determined by electrophysiological guidance. A cortical silver ball, placed contralaterally, served as a recording reference as well as ground. Dental cement was used to anchor the electrode assembly that was connected to a 10× gain preamplifier and then the connecting device during chronic recordings. Animals were given at least 2 weeks to recover.

Feedforward inhibition

Monophasic square pulses (0.1 ms duration) were delivered to the medial PP. Signals were amplified (100×), filtered (bandpass, 0.1-5 kHz), digitized at 10 kHz, and stored on disk for offline analysis (pClamp10 and Digidata 1440A, Molecular Devices). The frequency-dependent inhibition (FDI) was determined as described previously (Sloviter, 1991; Rosenblum et al., 1999; Zhang et al., 2010). Briefly, 10 stimuli at 0.1 Hz followed by 10 stimuli at 1.0 Hz were applied at 40% of the maximal population spike amplitude, and this pattern was repeated once. The population spike amplitude of responses 2-4 at 0.1 Hz stimulation was averaged and compared with the averaged amplitude of last three responses at 1.0 Hz stimulation in each set (see Fig. 3B). Then the results of the two sets were averaged to determine the FDI index.

Feedback inhibition

Following previous studies (Andersen et al., 1966; Sloviter, 1991; Zhang et al., 2010), we delivered paired pulses (conditioning and test) with interstimulus interval of 15, 30, 50, 100, or 150 ms at 0.025 Hz to the medial PP (see Fig. 3D). First, the intensity of the test stimulus was adjusted to evoke a population spike with 70%-80% amplitude of the maximum (see Fig. 3D, Test alone). Then the strength of the conditioning pulse was adjusted to evoke 70%-80% inhibition of the test pop spike when using 15 ms interstimulus interval (i.e., 20%-30% of the test response alone). By delivering pairs of the conditioning pulse and test pulse at various interstimulus intervals (15, 30, 50, 100, or 150 ms), the feedback inhibition index was normalized as the percentage of the test pop spike amplitude in each paired response over the amplitude of the test pop spike alone.

In vivo LTP

After recovery from surgical implantation, mice were habituated to the recording system for 4 d before starting the LTP test. For LTP evaluation, test responses elicited by 0.033 Hz monophasic pulses (0.1 ms duration) were recorded for 20 min on consecutive days at an intensity that evoked 30%–40% of the maximal population spike. Following 2 d of stable baseline, a tetanus (theta burst stimulation) was delivered to the PP for LTP induction. Pulse duration was doubled during tetani, which consisted of six series of six trains of six stimuli at 400 Hz, 200 ms between trains, 20 s between series. Responses were measured for 60 min right after tetanus and again for 20 min at 24, 48, and 120 h after tetanus. Since the latency of the population spike usually decreases following LTP induction, it is impractical to compare the initial slope of the fEPSP before and after LTP induction in awake animals (M. W. Jones et al., 2001; Malleret et al., 2001). Accordingly, we quantified the amplitude of the population spikes to evaluate LTP (Tang and Dani, 2009; Hao et al., 2015). Data were averaged every 5 min and normalized to the baseline measured over the 10 min before tetanic stimulation.

Ex vivo electrophysiology

Acute brain slice preparation

Acute hippocampal slices were prepared from 4-month-old male Cdkl5^−/y mice and Cdkl5^+/y littermates. Animals were deeply anesthetized with isoflurane and decapitated. The brains were quickly removed and transferred to the cold (0°C-4°C) oxygenated cutting solution containing the following (in mm): 93 NMDG, 93 HCl, 2.5 KCl, 1.2 NaH₂PO₄, 30 NaHCO₃, 20 HEPES, 25 glucose, 5 sodium ascorbate, 2 thiourea, 3 sodium pyruvate, 10 MgSO₄, and 0.5 CaCl₂, pH 7.4. Each brain was separated into halves by a sagittal cut at the midline. Then, coronal slices (300 µm thickness) containing the dorsal hippocampus were cut with a vibratome (Leica Microsystems) at a speed of 0.18 mm/s. Slices were then transferred to ACSF containing the following (in mm): 125 NaCl, 2.5 KCl, 1.5 MgCl₂, 2 CaCl₂, 25 D-glucose, 25 NaHCO₃, and 1.25 NaH₂PO₄, pH 7.4, incubated at 37.0 ± 0.5°C for 15-30 min and then at room temperature (20°C-22°C) for 0.5-1 h before recordings.

Whole-cell recordings

Individual slices were transferred to a submerged chamber that was perfused with oxygenated ACSF at a rate of 2 ml/min. The whole-cell patch-clamp recordings were performed under room temperature from the soma of DG granule cells (DGCs) and MOPP cells in the molecular layer of the DG. Signals were acquired with a Multiclamp 700B amplifier (Molecular Device), filtered between 1 Hz and 5 kHz, and digitized at 20 kHz with a Digidata-1550 system (Molecular Devices). Data acquisition and offline analysis were performed with Clampfit 10.6 software (Molecular Devices). Recording pipettes (6-9 mΩ, filled with intracellular solution) were pulled from borosilicate glass capillaries (B200-116-10, Sutter Instrument) using a horizontal glass electrode puller (P-1000, Sutter Instrument). Cell capacitance and membrane input resistance were determined in whole-cell voltage-clamp mode with a depolarizing step voltage command (10 mV) by using the membrane test function integrated in the Clampex software after the whole-cell patch-clamp configuration was achieved. The access resistance ranged from 10 to 30 mΩ and was periodically monitored and only neurons within a change of access resistance <15% during the experiments were used for further analysis.

Intrinsic membrane/firing properties were acquired in current-clamp mode with an intracellular recording pipette solution containing the following (in mm): 120 potassium gluconate, 10 HEPES, 4 KCl, 4 MgATP, 0.4 Na₃GTP, and 10 sodium phosphocreatine, pH 7.25 (adjusted with KOH). A series of current steps (1000 ms) were presented in ascending order (50 pA step) from −200 pA to 500 pA every 10 s to analyze the membrane and firing properties of neurons.

sEPSCs and sIPSCs were recorded in voltage-clamp with intracellular solution containing the following (in mm): 125 Cs-methanesulfonate, 1 EGTA, 2.5 MgCl₂, 4 MgATP, 0.4 Na₃GTP, 10 sodium phosphocreatine, 10 HEPES, and 2 QX-314, pH 7.25 (adjust with CsOH). To assess sEPSCs, cells were recorded in the presence of 25 µm bicuculline. sIPSCs were collected in the presence of CNQX (20 µm) and D-AP5 (25 µm). sEPSCs and sIPSCs were isolated by clamping neurons at −60 mV and 10 mV, which are in very close proximity to the reversal potential for GABA_ARs and AMPARs, respectively (Allene et al., 2012). In some experiments, biocytin (0.5%, Sigma) was added to the intracellular solution to identify the location of recorded cells.

We measured evoked EPSCs (eEPSCs) in the DGCs and MOPP cells received from the PP by dual voltage-clamp. A concentric bipolar stimulating microelectrode with a cone tip 25 µm in diameter (30202, FHC) was placed within the middle one-third of the DG molecular layer, ∼50-100 µm to the target neurons. The position of the stimulating microelectrode was carefully adjusted to evoke responses simultaneously in DGCs and MOPP cells. This setting avoided the direct activation of the hilar commissural pathway from the contralateral mossy cells that innervate the inner one-third of the DG molecular layer. A dual voltage-clamp technique allowed us to directly compare the synaptic input on the DGCs and MOPP cells induced by the same group of PP fibers from the entorhinal cortex. Stimuli were delivered every 15 s, with intensity ranging from 10 to 110 µA (increasing in increments of 5 µA), and two trials for each intensity. To estimate presynaptic neurotransmitter release, paired-pulse stimuli with 50 ms interstimulus interval were delivered to the medial PP every 15 s for 5 trials. The amplitude of the second response was divided by the amplitude of the first response within a pair to get the paired-pulse ratio (PPR). PPRs in MOPP cells and DGCs were calculated as the mean ratio of the 5 paired stimulus trials, respectively. To measure the size of readily releasable pool (RRP) and the vesicle release probability (P_ves), cells were held at −60 mV and 60 stimuli were delivered at 20 Hz for 3 s with an intensity adjusted to evoke 70% of the maximal eEPSC amplitude. Peak amplitudes of eEPSCs are measured and cumulated. A straight line was fitted to the final 15 points of the cumulative eEPSCs and back-extrapolated to the y axis. The Y intercept corresponds to RRP size, and P_ves equals the amplitude of the first eEPSC divided by RRP size, as previously described (Schneggenburger et al., 1999).

To measure AMPAR and NMDAR-mediated current ratio, stimulus intensity was set to result in 90% of maximal eEPSC. Both DGCs and MOPP cells were first held at a holding potential of −60 mV to record evoked AMPAR-mediated EPSCs (AMPA-EPSCs). The holding potential was then changed to 40 mV for granule cells and 60 mV for MOPP cells to record eEPSCs mediated by AMPAR and NMDAR. The amplitude of eEPSC 50 ms after the stimulus was considered to be the NMDAR-mediated component of EPSC (NMDA-EPSCs), since the AMPA EPSCs fade because of their unique channel property (Umemiya et al., 1999; Etherton et al., 2011). For each cell, at least 10 trials were averaged to obtain a representative response at each holding potential. To calculate the AMPA/NMDA ratio in DGCs and MOPP cells, the peak amplitude of the AMPA-EPSC at −60 mV was divided by the NMDA-EPSC amplitude at 40 or 60 mV, respectively.

Immunohistochemistry

The whole-cell patch recorded slices were fixed in 4% PFA for 3-5 h and transferred into 30% sucrose overnight. These slices were washed 3 times in PBS (10 min each) before being treated with Cy5 conjugate of streptavidin (1:1000, Thermo Fisher Scientific) diluted in PBS containing 5% serum and 1% Triton X-100 for 3 h. Slices were then counterstained with DAPI (900 nm, D1036, Invitrogen) for 15 min at room temperature in darkness. After washing the slices with 0.1 m PBS 3 times, we mounted them on glass slides with Fluoromount-G (Southern Biotechnology). Images were obtained with Carl Zeiss 880 confocal microscope using a 20× dry objective to identify the recorded cells.

Intracranial drug infusion

Under aseptic conditions, custom-made 26G guide cannulas were implanted bilaterally (2.0 mm posterior, 1.0 mm lateral, and 0.9 mm below the bregma) in Cdkl5^−/y mice and littermate Cdkl5^+/y controls anesthetized with isoflurane. After 2 weeks of recovery, mice of each genotype were randomly assigned into drug- or vehicle-treated groups. Gabazine hydrobromide (Sigma, SR-95531) was dissolved in PBS, pH 7.4, and the solution was back-filled into a 33G injector that was lowered 1.0 mm below the guide cannula tip while in place. The solution of 0.5 µl drug (1.0 ng gabazine, based on literature [Lu et al., 2016] and our pilot tests) or PBS vehicle per side was microinfused over 1 min into the dorsal hippocampus through a pump (Harvard Apparatus) (Tang and Dani, 2009; Hao et al., 2015). The injection sites in the hippocampi were aimed to the molecular layer of the DG where the MOPP cells locate. Following each injection, the injector was left in place for an additional minute to allow drug diffusion. To determine the effect of gabazine on hippocampus-dependent memory, animals that received gabazine or vehicle were transferred into a fear conditioning chamber ∼15 min after the injection. The fear conditioning training and 24 h retention tests of contextual and cued fear memory were conducted as described under Behavioral assays. For histologic verification of the injection sites, 4% methylene blue in PBS (∼ 0.2 µl) was injected into each injection site at the end of the experiment. The mice were killed with an overdose of isoflurane and their brains dissected. Frozen 30 µm coronal sections were cut and costained with cresyl violet and eosin. One mouse with injection sites outside of the targeted brain regions was excluded from data analysis.

Experimental design and statistical analysis

For behavioral assays, we determined sample sizes based on previously published results in this CDD model (Wang et al., 2012). For in vivo electrophysiology, we chose sample sizes and power based on our previous publications (Zhang et al., 2010; Hao et al., 2015). Mice within the same genotype were randomly assigned to receive DBS or drug treatment. For ex vivo electrophysiological experiments, sample sizes were determined based on previous publications (Wang et al., 2017a,b) and reported as the numbers of cells/mouse.

All data were analyzed using the Shapiro–Wilk test for normality. For two-sample comparisons, datasets with normal distributions were analyzed for significance using two-tailed unpaired or paired t test, whereas datasets with non-normal distributions were analyzed using the Mann–Whitney test. One-way or two-way repeated-measures ANOVA and two-way ANOVA were conducted as appropriate. If any of the main effects were significant, the Tukey post hoc test was used for all pairwise multiple comparisons. Significance was defined as p < 0.05. Data in line plots and in the text are presented as mean ± SEM. In box plots, horizontal lines are the median and the 25% and 75% range values; whiskers present the extreme data points from the box edges. Sample sizes and statistical details of each experiment can be found in Results. GraphPad Prism 8 was used to create all the plots and the statistical tests in this study.

Results

Loss of CDKL5 impairs hippocampus-dependent learning and memory

To explore the effects of CDKL5 deficiency on hippocampal memory in the mouse model used in this study (B6.129(FVB)-Cdkl5^tm1.1Joez/J), we performed multiple behavioral assays in both male (Cdkl5^−/y) and female (Cdkl5^+/−) mutant mice and their sex- and age-matched WT littermate controls (WT, Cdkl5^+/y or Cdkl5^+/+). We first conducted the classic fear conditioning task (Fig. 1A). Because CDKL5-deficient mice tend to show more motor activity (Extended Data Fig. 1-1), we first established that, during fear conditioning training, 4-month-old Cdkl5^−/y mice showed similar levels of exploration before the foot shock as their WT littermates (Cdkl5^−/y 0.38 ± 0.21% vs Cdkl5^+/y 0.34 ± 0.21%, Mann–Whitney test, U = 67.5, p = 0.745, preshock data in Fig. 1B). The mutant animals froze less, however, immediately after the foot shock (Cdkl5^−/y 22.40 ± 3.64% vs Cdkl5^+/y 37.54 ± 3.80%, two-tailed unpaired t test, t₍₂₂₎ = 3.800, p = 0.0088) compared with WT littermates (postshock data in Fig. 1B). Consistent with previous fear conditioning study results (Wang et al., 2012), Cdkl5^−/y mice froze less in both the conditioning context and the cued novel environment than did controls when tested 24 h after fear conditioning training (contextual: Cdkl5^−/y 17.50 ± 2.70% vs Cdkl5^+/y 29.37 ± 4.44%, two-tailed unpaired t test, t₍₂₂₎ = 2.284, p = 0.0324; cued: Cdkl5^−/y 32.89 ± 3.86% vs Cdkl5^+/y 57.25 ± 5.93%, Mann–Whitney test, U = 29.0, p = 0.0140; Fig. 1C,D). During a passive avoidance task, in which we trained the mice to avoid a dark compartment by giving a foot shock (Samaco et al., 2013), the male mutant mice entered the shock compartment more quickly than controls (Cdkl5^−/y 22.48 ± 4.46 vs Cdkl5^+/y 62.87 ± 12.16, two-tailed unpaired t test, t₍₂₃₎ = 3.218, p = 0.0038; Fig. 1E), indicating impaired hippocampal memory. There was no difference in pain threshold between genotypes (current intensity for flinch: Cdkl5^−/y 0.11 ± 0.006 vs Cdkl5^+/y 0.11 ± 0.007, Mann–Whitney test, U = 66.0, p = 0.6520; vocalization: Cdkl5^−/y 0.21 ± 0.02 vs Cdkl5^+/y 0.18 ± 0.01, two-tailed unpaired t test, t₍₂₂₎ = 1.410, p = 0.1720; jump: Cdkl5^−/y 0.35 ± 0.03 vs Cdkl5^+/y 0.28 ± 0.02, Mann–Whitney test, U = 39.0, p = 0.0550; Fig. 1F).

Figure 1.

Loss of CDKL5 impairs hippocampus-dependent learning and memory in male Cdkl5^−/y and female Cdkl5^+/− mice. A-D, Learning and memory in the fear conditioning task in 4-month-old male Cdkl5^−/y mice and 6-month-old female Cdkl5^+/− mice (dark blue and red, respectively), together with their age- and sex-matched WT littermates (Cdkl5^+/y, Cdkl5^+/+; light blue and orange, respectively) (n = 12 mice per group). A, Schematic representation of fear conditioning paradigm. B, During fear conditioning training, both male and female mutant mice showed comparable exploration levels before the first foot shock compared with their WT littermates, respectively. However, both male and female mutant mice froze less than their WT controls immediately after the second foot shock. C, Both male and female mutant mice showed less freezing in response to the contextual fear conditioning than their WT littermates. D, Cdkl5^−/y mice showed less freezing in response to the cued fear conditioning than their WT littermates, whereas Cdkl5^+/− mice showed comparable cued freezing level as their WT littermates. E, Both male Cdkl5^−/y mice (n = 12) at 4 months of age and female Cdkl5^+/− mice (n = 13) at 6 months of age showed less hesitance to enter the dark compartment in the passive avoidance task than their age- and sex-matched WT littermates (Cdkl5^+/y, n = 13; Cdkl5^+/+, n = 13). F, There was no difference in pain threshold intensities to evoke flinch, vocalization, or jumping between male Cdkl5^−/y mice and Cdkl5^+/y littermates or between female Cdkl5^+/− mice and Cdkl5^+/+ controls (n = 12 per group). G-J, Learning and memory in the water maze task in male Cdkl5^−/y mice (n = 15) at 4 months of age and female Cdkl5^+/− mice (n = 13) at 6 months of age, together with their age- and sex-matched WT littermates (Cdkl5^+/y, n = 13; Cdkl5^+/+, n = 13). G, Escape latency during acquisition training. The mutant mice did not show much improvement during the training. H, Percentage of time in the target zone and averaged time in other zones during probe test. I, Number of platform area crossings in the probe test. J, Escape latency in the water maze task with a flagged platform. *p < 0.05, **p < 0.01, ***p < 0.001; ^#p < 0.05, ^##p < 0.01, Mann–Whitney test. For details on locomotor activity in both male and female Cdkl5 mutant mice as well as memory performances in 4-month-old female Cdkl5^+/− mice and Cdkl5^+/+ controls, see Extended Data Figures 1-1 and 1-2.

Figure 1-1

Locomotor assay. Travel distances were plotted in 5-min bins over a 30 min test in the open field task. A, Male Cdkl5^-/y mice (n = 12) at 4 months of age in comparison with their Cdkl5^+/y controls (n = 12). There was no genotypic difference during the first 5 min of locomotion, but by ten minutes the mutant mice were more active than their littermates (Cdkl5^-/y 13.65 ± 1.60 vs Cdkl5^+/y 16.23 ± 1.43, p = 0.090). B, Female Cdkl5^+/- mice (n = 11) at 4 months of age in comparison with their Cdkl5^+/+ controls (n = 12). As with the male mice, a genotypic difference in activity level became apparent only after the first five minutes (Cdkl5^+/- 21.18 ± 1.65 vs Cdkl5^+/+ 17.34 ± 1.80, p = 0.072). C, Female Cdkl5^+/- mice (n = 12) at 6 months of age in comparison with their Cdkl5^+/+ controls (n = 12). Again, genotypic differences became apparent only after five minutes (Cdkl5^+/- 19.97 ± 2.29 vs Cdkl5^+/+ 22.60 ± 1.55, p = 0.328). Two-way repeated-measures ANOVA followed by Tukey posthoc analysis, *p < 0.05, **p < 0.01, ***p < 0.001; Download Figure 1-1, TIF file.

Figure 1-2

Female Cdkl5^+/- mice show normal learning and memory at the age of 4 months. A-C, Learning and memory in the fear conditioning task. (A) During fear conditioning training, Cdkl5^+/- mice (n = 11) showed comparable exploration level as Cdkl5^+/+ littermates (n = 12) before the first foot shock (Cdkl5^+/- 0.74 ± 0.35% vs Cdkl5^+/+ 0.60 ± 0.27%, Mann–Whitney test, U = 63.0, p = 0.856), but froze less than Cdkl5^+/+ controls immediately after the second shock (Cdkl5^+/- 32.65 ± 3.98% vs Cdkl5^+/+ 50.07 ± 3.87%, two-tailed unpaired t test, t₍₂₁₎ = 3.138, p = 0.0050). There was no difference of contextual (B, Cdkl5^+/- 24.12 ± 4.15% vs Cdkl5^+/+ 32.29 ± 2.94%, two-tailed unpaired t test, t₍₂₁₎ = 1.627, p = 0.1190) and cued (C, Cdkl5^+/- 46.28 ± 6.83% vs Cdkl5^+/+ 54.91 ± 7.24%, two-tailed unpaired t test, t₍₂₁₎ = 0.864, p = 0.3980) fear memory between Cdkl5^+/+ littermates and Cdkl5^+/- mice. D, In the passive avoidance task, Cdkl5^+/- mice (n = 10) had similar latency as Cdkl5^+/+ mice (n = 10) (Cdkl5^+/- 61.56 ± 28.68 vs Cdkl5^+/+ 115.97 ± 36.67, Mann–Whitney test, U = 40.0, p = 0.4720) to enter the dark compartment when tested 24 h after training. Download Figure 1-2, TIF file.

We next used a water maze task to test spatial learning and memory (Morris et al., 1982; Hao et al., 2015). The Cdkl5^−/y mice took longer to locate the platform during training than their WT littermates (two-way repeated-measures ANOVA followed by Tukey's post hoc analysis: genotype, F_(1,28) = 29.66, p < 0.001; day, F_(8,224) = 8.35, p < 0.001; genotype × day interaction, F_(8,224) = 2.23, p = 0.0260; Fig. 1G). When we removed the platform during the probe trial, the mutant mice could not remember the platform location and spent less time in the target quadrant (Cdkl5^+/y target 36.63 ± 3.90 vs Cdkl5^+/y others 21.12 ± 1.30, two-tailed paired t test, t₍₁₄₎ = 2.981, p = 0.0099; Cdkl5^−/y target 13.86 ± 3.29 vs Cdkl5^−/y others 28.73 ± 1.10, two-tailed paired t test, t₍₁₄₎ = 3.386, p = 0.0044; Cdkl5^−/y target 13.86 ± 3.29 vs Cdkl5^+/y target 36.63 ± 3.90, two-tailed unpaired t test, t₍₂₈₎ = 4.460, p = 0.0001; Fig. 1H). As a result, Cdkl5^−/y mice swam through the area where the platform had been less frequently (Cdkl5^−/y 0.73 ± 0.27 vs Cdkl5^+/y 1.80 ± 0.34, Mann–Whitney test, U = 60.0, p = 0.0240; Fig. 1I). This poor performance was not because of differences in visual or sensorimotor skills (escape latency during visible training: Cdkl5^−/y 20.99 ± 1.83 vs Cdkl5^+/y 23.93 ± 2.54, two-tailed unpaired t test, t₍₂₈₎ = 0.939, p = 0.3560; Fig. 1J).

Although CDD primarily affects human females, there have been fewer studies done with female Cdkl5^+/− mice because CDKL5 is on the X chromosome and thus is expressed mosaically (so the phenotype would be milder and later-onset than in males); skewed X-chromosome inactivation could also make the results less predictable. Nevertheless, because female Cdkl5^+/− mice are more clinically relevant, we subjected them to the same behavioral tests to evaluate learning and memory. During fear conditioning training, 4-month-old female mice showed similar levels of exploratory behavior before the foot shock but froze less immediately after the foot shock compared with WT littermates (Extended Data Fig. 1-2A). These female mutant mice did not show memory deficits in either contextual or cued fear conditioning (Extended Data Fig. 1-2B,C) or the passive avoidance task (Extended Data Fig. 1-2D) when tested 24 h later.

To determine whether the fear-conditioned memory impairment in female Cdkl5 heterozygous mice is affected by aging, we tested hippocampal memory in an older cohort of mice. Similar to the younger mice, 6-month-old female mutant mice showed comparable exploratory behavior before the foot shock (Cdkl5^+/− 0.51 ± 0.25% vs Cdkl5^+/+ 0.84 ± 0.41%, Mann–Whitney test, U = 59.5, p = 0.664) and froze less immediately following the foot shock (Cdkl5^+/− 37.17 ± 4.38% vs Cdkl5^+/+ 52.64 ± 4.25%, two-tailed unpaired t test, t₍₂₁₎ = 2.536, p = 0.0196) compared with WT littermates (Fig. 1B). In contrast to the younger female mutant mice, 6-month-old female heterozygous mice did display contextual memory deficits (contextual: Cdkl5^+/− 24.81 ± 3.73% vs Cdkl5^+/+ 43.68 ± 5.00%, two-tailed unpaired t test, t₍₂₂₎ = 3.024, p = 0.0062; cued: Cdkl5^+/− 59.86 ± 4.84% vs Cdkl5^+/+ 65.68 ± 4.42%, two-tailed unpaired t test, t₍₂₂₎ = 0.887, p = 0.3840; Fig. 1C,D). In the passive avoidance task, the 6-month-old female mutant mice entered the shock compartment much earlier than did the WT controls (Cdkl5^+/− 18.87 ± 2.33 vs Cdkl5^+/+ 214.23 ± 35.59, two-tailed unpaired t test, t₍₂₄₎ = 5.478, p = 0.00,001; Fig. 1E). Again, there was no difference of pain threshold between genotypes (current intensity for flinch: Cdkl5^+/− 0.13 ± 0.007 vs Cdkl5^+/+ 0.13 ± 0.007, Mann–Whitney test, U = 72.0, p = 1.0000; vocalization: Cdkl5^+/− 0.25 ± 0.02 vs Cdkl5^+/+ 0.21 ± 0.02, Mann–Whitney test, U = 45.5, p = 0.1210; jump: Cdkl5^+/− 0.35 ± 0.02 vs Cdkl5^+/+ 0.33 ± 0.02, two-tailed unpaired t test, t₍₂₂₎ = 0.692, p = 0.4960; Fig. 1F).

Like the male Cdkl5 null mice in the water maze task, female Cdkl5^+/− mice took longer to find the platform over the acquisition sessions than did their WT littermates (two-way repeated-measures ANOVA followed by Tukey's post hoc analysis: genotype, F_(1,24) = 27.32, p < 0.001; day, F_(8,192) = 5.38, p < 0.001) and did not show a preference for the target quadrant (Fig. 1G). As a result, Cdkl5^+/− mice spent less time in the target quadrant (Cdkl5^+/+ target 31.76 ± 2.28 vs Cdkl5^+/+ others 22.74 ± 0.76, two-tailed paired t test, t₍₁₂₎ = 2.973, p = 0.0116; Cdkl5^+/− target 18.78 ± 4.70 vs Cdkl5^+/− others 27.08 ± 1.57, two-tailed paired t test, t₍₁₂₎ = 1.323, p = 0.2100; Cdkl5^+/− target 18.78 ± 4.70 vs Cdkl5^+/+ target 31.76 ± 2.28, Mann–Whitney test, U =38.0, p = 0.0180; Fig. 1H) and swam over the platform area fewer times than did their WT counterparts (Cdkl5^+/− 0.85 ± 0.39 vs Cdkl5^+/+ 1.69 ± 0.13, Mann–Whitney test, U = 29.5, p = 0.0030; Fig. 1I) during the probe trial. Their visual and sensorimotor skills were normal (escape latency during visible training: Cdkl5^+/− 27.20 ± 3.14 vs Cdkl5^+/+ 24.22 ± 2.56, two-tailed unpaired t test, t₍₂₄₎ = 0.735, p = 0.4690; Fig. 1J).

CDKL5 loss impairs hippocampal synaptic plasticity and inhibitory activity in freely moving mice

The clear hippocampus-dependent memory impairment in Cdkl5 mutant mice led us to examine synaptic plasticity of the PP to the DG. Given that LTP of synaptic transmission serves as a neural substrate for learning and memory (Malenka and Bear, 2004; Whitlock et al., 2006), we expected that LTP would be impaired. We therefore implanted stimulation electrodes in the medial PP and recording electrodes in the ipsilateral DG in both male Cdkl5^−/y and female Cdkl5^+/− mice and their WT littermates at the ages they displayed impaired memory. After 2 d of baseline recordings, we used theta burst stimulation to induce LTP and monitored the responses for 5 d while they moved freely, as previously described (Hao et al., 2015). Both Cdkl5^−/y mice and their Cdkl5^+/y littermates showed LTP (one-way repeated-measures ANOVA on day 0: Cdkl5^+/y, F_(15,135) = 29.74, p < 0.001; Cdkl5^−/y, F_(15,90) = 26.32, p < 0.001; Fig. 2A,B). LTP in Cdkl5^−/y mice, however, had lower magnitude than in controls. Two-way repeated-measures ANOVA revealed a significant genotype effect last for 3 d after induction (day 0, F_(1,14) = 13.41, p = 0.0030; day 1, F_(1,13) = 7.63, p = 0.0160; day 2, F_(1,13) = 5.94, p = 0.0300; Fig. 2B).

Figure 2.

Freely moving Cdkl5^−/y and Cdkl5^+/− mice show impaired hippocampal synaptic plasticity in the DG. A, Superimposed representative traces of evoked response of the PP recorded in the DG 5 min before (gray) and 55 min after tetani. B, Summary of in vivo LTP in 4-month-old male Cdkl5^−/y mice (n = 6–7) and Cdkl5^+/y littermates (n = 9). LTP was induced by theta burst stimulation (arrow) on day 0 and followed up for 5 more days. C, Summary of in vivo LTP in 6-month-old female Cdkl5^+/− mice (n = 7) and Cdkl5^+/+ controls (n = 7). LTP was induced by theta burst stimulation (arrow) on day 0 and followed up for 5 more days. *p < 0.05, **p < 0.01.

Similarly, theta burst stimulation induced LTP in both Cdkl5^+/− mice (one-way ANOVA on day 0, F_(15,90) = 13.24, p < 0.001) and WT controls (F_(15,90) = 26.50, p < 0.001; Fig. 2A,C). The heterozygous female mice had impaired LTP for at least 2 d after induction (two-way repeated-measures ANOVA, genotype effect: day 0, F_(1,12) = 5.42, p = 0.0380; day 1, F_(1,12) = 8.37, p = 0.014; day 2, F_(1,12) = 3.95, p = 0.070; Fig. 2C). Therefore, complete loss of CDKL5 in the male nulls or partial loss of CDKL5 in the female heterozygous mice severely impairs synaptic plasticity in the PP-DG pathway. In vivo synaptic plasticity in the neural pathway gating cortical information to the hippocampus thus requires CDKL5 protein.

Altered local circuit activity may impair hippocampal memory and synaptic plasticity. Inhibitory interneurons exert powerful control over local circuit activity (Freund and Buzsáki, 1996). GABAergic interneurons in the DG integrate into the PP-initiated local feedforward and feedback inhibition circuits (Fig. 3A) and play major roles not only in regulating single-cell excitability but in gating, filtering, and processing cortical inputs to the hippocampus (Hsu et al., 2016; Hainmueller and Bartos, 2020). Therefore, we hypothesized that CDKL5 might be critical specifically to the inhibitory circuits that mediate synaptic plasticity and memory function.

Figure 3.

Freely moving Cdkl5^−/y and Cdkl5^+/− mice show altered inhibitory circuit activity in the DG. A, Diagram of the feedforward and feedback inhibition circuitry in the DG. IN, Interneuron; +, excitatory; -, inhibitory. B, The stimulus protocol of FDI was used to determine the feedforward inhibition in the PP to DG pathway. Ten pulses were delivered to the medial PP at 0.1 Hz followed by 10 pulses at 1.0 Hz. This pattern was repeated once. The evoked responses in the ipsilateral dentate were recorded. The population spike amplitudes from the responses 2-4 at 0.1 Hz (a1, a2) were averaged, respectively, and divided the averaged amplitude of the last three responses at 1.0 Hz (b1, b2) in each set. The results of the two sets were averaged to give the FDI index. Representative traces were evoked by 0.1 and 1.0 Hz pulses, respectively. C, Summary of the FDI index in male Cdkl5^−/y mice (n = 7) and Cdkl5^+/y littermates (n = 10) as well as female Cdkl5^+/− mice (n = 7) and Cdkl5^+/+ littermates (n = 7). Loss of CDKL5 decreased FDI index (for increased feedforward inhibition). D, Determination of feedback inhibition. Field potential recordings (average of 10 traces collected at 0.025 Hz) were recorded from the dentate in response to the ipsilateral paired stimuli of the medial PP. The examples showed the responses evoked by conditioning stimulus alone, test stimulus alone, and the test responses evoked by paired stimuli (conditioning + test) at different interstimulus intervals (15, 30, 50, 100, and 150 ms). E, Normalized feedback inhibition index versus interstimulus interval plot in male Cdkl5^-/y mice (n = 7) and Cdkl5^+/y littermates (n = 10) as well as female Cdkl5^+/− mice (n = 5) and Cdkl5^+/+ littermates (n = 6). *p < 0.05.

To test feedforward inhibition, we applied 10 stimuli at 0.1 Hz followed by 10 stimuli at 1.0 Hz to the medial PP and repeated this pattern once. We normalized the pop spike amplitudes of the resulted responses in the DG to get the FDI index (Rosenblum et al., 1999; Zhang et al., 2010) (Fig. 3B). Compared with their WT littermates, both Cdkl5^−/y and Cdkl5^+/− mice had a lower FDI index, indicating increased feedforward inhibition after CDKL5 loss (Cdkl5^−/y 0.69 ± 0.03 vs Cdkl5^+/y 0.80 ± 0.03, two-tailed unpaired t test, t₍₁₅₎ = 2.636, p = 0.0187; Cdkl5^+/− 0.58 ± 0.05 vs Cdkl5^+/+ 0.74 ± 0.04, two-tailed unpaired t test, t₍₁₂₎ = 2.700, p = 0.0193; Fig. 3C). To evaluate feedback inhibition, we delivered paired-pulse stimuli (conditioning and test) at varied interstimulus intervals to the medial PP, and then compared the pop spike amplitude of the test responses to the amplitude of the response evoked by single test pulse stimulation (Andersen et al., 1966; Sloviter, 1991; Zhang et al., 2010) (Fig. 3D). As shown in Figure 3E, there was no difference in feedback inhibition index between mutants and their WT littermates across the interstimulus intervals in both Cdkl5^−/y mice (two-way repeated-measures ANOVA, genotype effect, F_(1,15) = 0.87, p = 0.3660) and Cdkl5^+/− animals (F_(1,9) = 0.68, p = 0.4320). These results indicate that CDKL5 loss specifically increases feedforward inhibition in the PP to the DG.

CDKL5 loss alters excitatory and inhibitory synaptic transmission in the Cdkl5^−/y DG

Several studies have shown altered E/I balance at play in multiple psychiatric and neurodevelopmental disorders (Nelson and Valakh, 2015; Foss-Feig et al., 2017). The DGCs receive both excitatory glutamatergic input from the PP and inhibitory input from local inhibitory GABAergic interneurons. We therefore tested the effect of CDKL5 loss on synaptic transmission of DGCs by recording sEPSCs and sIPSCs in hippocampal slices. Given the confounding effects of mosaic CDKL5 expression in female heterozygous mice from random X-chromosome inactivation, we compared male Cdkl5 null (Cdkl5^−/y) mice with their male WT littermates (Cdkl5^+/y). Both the frequency and amplitude of sEPSCs were lower in Cdkl5^−/y mice than in WT controls (frequency: Cdkl5^−/y 0.65 ± 0.05 vs Cdkl5^+/y 0.98 ± 0.08, Mann–Whitney test, U = 252.0, p = 0.0020; amplitude: Cdkl5^−/y 7.32 ± 0.20 vs Cdkl5^+/y 8.13 ± 0.24, two-tailed unpaired t test, t₍₅₉₎ = 2.564, p = 0.0129; Fig. 4A). In contrast, the sIPSC frequency was higher (Cdkl5^−/y 1.56 ± 0.13 vs Cdkl5^+/y 0.79 ± 0.0730, Mann–Whitney test, U = 89.0, p ≤ 0.001), whereas sIPSC amplitude remained the same (Cdkl5^−/y 13.98 ± 0.92 vs Cdkl5^+/y 15.10 ± 0.79, two-tailed unpaired t test, t₍₄₇₎ = 0.925, p = 0.3600; Fig. 4B). These data suggest that DGCs in Cdkl5^−/y mice receive less excitatory and more inhibitory input, skewing the E/I balance toward inhibition.

Figure 4.

Male Cdkl5^-/y mice have impaired synaptic transmission in the PP to the DG. A, Top, Representative sEPSC traces recorded from the DGCs in Cdkl5^−/y mice and Cdkl5^+/y littermates. Calibration: 1 s, 10 pA. Bottom, Frequency and amplitude of sEPSCs measured from the DGCs in Cdkl5^−/y mice (n_{cells/animals} = 31/5) and Cdkl5^+/y littermates (n_{cells/animals} = 30/5). B, Top, Representative sIPSC traces of DGCs in Cdkl5^−/y mice and Cdkl5^+/y littermates. Calibration: 1 s, 15 pA. Bottom, Frequency and amplitude of sIPSCs measured from DGCs in Cdkl5^−/y mice (n_{cells/animals} = 24/3) and Cdkl5^+/y littermates (n_{cells/animals} = 25/3). C, Top, Representative sEPSC traces of MOPP cells in Cdkl5^−/y mice and Cdkl5^+/y littermates. Calibration: 1 s, 5 pA. Bottom, Frequency and amplitude of sEPSCs measured from MOPP cells in Cdkl5^−/y mice (n_{cells/animals} = 31/5) and Cdkl5^+/y littermates (n_{cells/animals} = 30/5). D, Schematic diagram of dual whole-cell recording configuration in MOPP cells (Recording 1) and DGCs (Recording 2). The evoked responses to the same stimulation were measured in pairs. E, Immunofluorescent image of dual-patched, biocytin-filled MOPP cells and DGCs in the dentate molecular layer and dentate granule cell layer, respectively. White and yellow arrowheads indicate the paired recorded cells. Mol, Molecular layer; GCL, granule cell layer. Scale bar, 100 μm. F, Representative recording traces of dual patch, physiologically demonstrating the direct connection from a MOPP cell to a DGC. Excitation of MOPP cell by current injection (3 nA, 100 Hz for 1 s) evoked an IPSP in the DGC. Black arrow indicates the beginning of activation train. Red arrow indicates the IPSP initiation with ∼10 ms latency in the DGC. G, Pairwise analysis of eEPSC amplitude versus stimulus intensity in MOPP cells and DGCs of both Cdkl5^−/y and Cdkl5^+/y mice (n_{cells/animals} = 29 pairs/5 for each group). H, Representative traces of paired-pulse evoked responses with 50 ms interval recorded in MOPP cells and DGCs of Cdkl5^+/y and Cdkl5^−/y mice. I, Pairwise analysis of PPR in four groups (Cdkl5^+/y-MOPP and Cdkl5^+/y-DGC, n_{cells/animals} = 24/5 per group; Cdkl5^−/y-MOPP and Cdkl5^-/y-DGC, n_{cells/animals} = 26/5 per group). J, Representative traces of evoked AMPAR-mediated EPSCs (negative going) recorded at −60 mV and compound EPSCs (positive going) recorded at 60 mV in MOPP cells and 40 mV in DGCs. Red dotted lines indicate 50 ms after stimulus, a time point at which the AMPAR-mediated contribution is considered negligible. K-M, Pairwise analysis of AMPAR- and NMDAR-mediated EPSCs in Cdkl5^−/y mice (n_{cells/animals} = 26 pairs/4) and Cdkl5^+/y littermates (n_{cells/animals} = 28 pairs/4). K, AMPAR-mediated component of EPSC. L, NMDAR-mediated component of EPSCs. M, AMPA/NMDA ratio. *p < 0.05, **p < 0.01, ***p < 0.001; ^#p < 0.05, ^##p < 0.01, ^###p < 0.001, Mann–Whitney test. For details on presynaptic readily releasable size and release probability of MOPP cells as well as the intrinsic properties of DGCs and MOPP cells, see Extended Data Figures 4-1 and 4-2.

Figure 4-1

Loss of CDKL5 does not alter the presynaptic function in the PP to the MOPP cells. A, Representative evoked responses in MOPP cells induced by 60 stimuli at 20 Hz for 3 s to the medial PP in Cdkl5^+/y and Cdkl5^-/y mice. Stimulation artifacts were removed for clarity. B-D, Analysis of the presynaptic properties in Cdkl5^-/y mice (n_{cells/animals} = 25/5) and Cdkl5^+/y littermates (n_{cells/animals} = 24/5). (B) Amplitudes of the first eEPSCs (Cdkl5^+/y 90.99 ± 8.94 vs Cdkl5^-/y 118.97 ± 11.58; Mann–Whitney test, U = 208.0, p = 0.067); (C) RRP sizes (Cdkl5^+/y 1717.05 ± 152.12 vs Cdkl5^-/y 1942.57 ± 137.52; two-tailed unpaired t test, t₍₄₇₎ = 1.102, p = 0.276); (D) Vesicle release probability (P_ves) (Cdkl5^+/y 0.06 ± 0.004 vs Cdkl5^-/y 0.07 ± 0.009, Mann–Whitney test, U = 260.0, p = 0.430). Download Figure 4-1, TIF file.

Figure 4-2

Loss of CDKL5 does not alter intrinsic membrane properties of MOPP cells and DGCs in Cdkl5^-/y mice. A, Intrinsic membrane properties of MOPP cells. Compared to Cdkl5^+/y controls, MOPP cells in Cdkl5^-/y mice showed similar resting membrane potential (RMP) (left; Cdkl5^-/y MOPP 63.49 ± 1.29 vs Cdkl5^+/y MOPP 65.91 ± 3.62, Mann–Whitney test, U = 227.0, p = 0.7200; Cdkl5^+/y, n_{cells/animals} = 18/3; Cdkl5^-/y, n_{cells/animals} = 18/3) and normal rheobase (right; Cdkl5^-/y MOPP 183.33 ± 15.13 vs Cdkl5^+/y MOPP 222.22 ± 18.18, two-tailed unpaired t test, t₍₃₄₎ = 0.109, p = 0.1090; Cdkl5^+/y, n_{cells/animals} = 18/3; Cdkl5^-/y, n_{cells/animals} = 18/3). B, Intrinsic membrane properties of DGCs. Compared to Cdkl5^+/y controls, DGCs in Cdkl5^-/y mice showed similar resting membrane potential (left; Cdkl5^-/y DGC 72.59 ± 1.32 vs Cdkl5^+/y DGC 74.29 ± 1.20, Mann–Whitney test, U = 395.0, p = 0.093; Cdkl5^+/y, n_{cells/animals} = 26/3; Cdkl5^-/y, n_{cells/animals} = 25/3) and normal rheobase (right; Cdkl5^-/y DGC 120.00 ± 7.64 vs Cdkl5^+/y DGC 136.54 ± 9.01, Mann–Whitney test, U = 264.5, p = 0.2010; Cdkl5^+/y, n_{cells/animals} = 26/3; Cdkl5^-/y, n_{cells/animals} = 25/3). Download Figure 4-2, TIF file.

We reasoned that this imbalance might originate from interneurons. We were particularly interested in MOPP cells, which contribute the most to feedforward inhibition of DGCs (Li et al., 2013). To determine whether CDKL5 loss enhances the excitatory synaptic transmission from the PP to MOPP cells, which would then increase inhibitory input to the DGCs, we recorded their sEPSCs. As expected, MOPP cells in Cdkl5^−/y mice displayed higher sEPSC frequency than in Cdkl5^+/y controls (Cdkl5^−/y 3.96 ± 0.26 vs Cdkl5^+/y 2.68 ± 0.19, Mann–Whitney test, U = 206.0, p ≤ 0.001), but similar amplitude (Cdkl5^−/y 7.75 ± 0.25 vs Cdkl5^+/y 7.39 ± 0.24, two-tailed unpaired t test, t₍₅₉₎ = 1.064, p = 0.2920; Fig. 4C). This supports the notion that increased excitatory input onto MOPP cells led to the increased DGC inhibition we observed in Cdkl5^−/y mice.

To further probe this model, we performed simultaneous dual patch recordings in MOPP cells and DGCs to get PP-induced evoked responses. Placing a fine-tipped (25 µm in diameter) concentric bipolar stimulating electrode in the middle one-third of the DG molecular layer (Fig. 4D) allowed us to stimulate only the medial PP and avoid activating other synaptic input to the MOPP cells and DGCs (Hsu et al., 2016). In a few cases, the dual patch proved a direct connection between MOPP cells and DGCs. Activation of the MOPP cell evoked an inhibitory postsynaptic potential in the paired DGC with a ∼10 ms latency (Fig. 4E,F), confirming that the recorded MOPP cell mediated feedforward inhibition from the PP to the DGC. We conducted pairwise analysis on evoked responses over incremental stimulus intensities (input-output curve). The eEPSC amplitude in MOPP cells was significantly larger than that of DGCs in Cdkl5^+/y mice, indicating that the MOPP cells do receive stronger PP input than DGCs at baseline (Fig. 4G). Cdkl5^−/y mice showed greater eEPSC amplitude in MOPP cells, but lower amplitude in DGCs, than did Cdkl5^+/y controls (Fig. 4G). Within-genotype analysis indicated a significant cell type effect in both Cdkl5^+/y (cell type, F_(1,56) = 30.15, p < 0.001; cell type × stimulus intensity interaction, F_(20,1120) = 12.43, p < 0.001) and Cdkl5^−/y mice (cell type, F_(1,56) = 106.26, p < 0.001; cell type × stimulus intensity interaction, F_(20,1120) = 61.90, p < 0.001). Within-cell type analysis revealed a significant CDKL5 effect in both MOPP cells (genotype, F_(1,56) = 7.39, p = 0.0090; genotype × stimulus intensity interaction, F_(20,1120) = 3.07, p < 0.001) and DGCs (genotype, F_(1,56) = 20.21, p < 0.001; genotype × stimulus intensity interaction, F_(20,1120) = 12.86, p < 0.001). These data suggest that loss of CDKL5 amplifies the difference between MOPP cells and DGCs in excitatory synaptic transmission from the PP.

Because changes in synaptic transmission could arise on either the presynaptic or postsynaptic side, we next tested paired-pulse facilitation, which is inversely related to presynaptic neurotransmitter release, using the same dual patch settings. There was a significant interaction of the PPR between genotype and cell type (two-way ANOVA: F_(1,96) = 5.05, p = 0.0270) (Fig. 4H,I). Tukey's post hoc analysis revealed that Cdkl5^−/y mice had a greater PPR of DGCs than did Cdkl5^+/y controls (Cdkl5^−/y DGC 1.64 ± 0.12 vs Cdkl5^+/y DGC 1.31 ± 0.08, p = 0.0160), indicating that CDKL5 loss decreased glutamate release from the PP onto these cells. Notably, the PPR in DGCs was higher than in MOPP cells in Cdkl5^−/y mice (Cdkl5^−/y MOPP 1.31 ± 0.08 vs Cdkl5^−/y DGC 1.64 ± 0.12, p = 0.0130), whereas Cdkl5^+/y littermates showed no cell type difference (Cdkl5^+/y MOPP 1.41 ± 0.10 vs Cdkl5^+/y DGC 1.31 ± 0.08, p = 0.4910). CDKL5 did not affect PPR in MOPP cells (Cdkl5^−/y MOPP 1.31 ± 0.08 vs Cdkl5^+/y MOPP 1.41 ± 0.10, p = 0.4680), which is further confirmed by the unchanged RRP size and presynaptic release probability in MOPP cells of Cdkl5^−/y mice (Extended Data Fig. 4-1), indicating that CDKL5 is not involved in the presynaptic release in MOPP cells. This suggests that there was less excitatory neurotransmitter release from the PP to the DGCs than to the MOPP cells in Cdkl5^−/y mice, which is consistent with the data in Figure 4G.

A change in sEPSC amplitude is routinely attributed to a postsynaptic alteration in receptor function, number, or both. A change in sEPSC frequency could be because of either an altered probability of neurotransmitter release that was not detected using the paired-pulse analysis or a postsynaptic change in the number of functional AMPARs at synapses that were previously electrophysiologically “silent” (Isaac et al., 1995). Because CDKL5 loss increased MOPP cell sEPSC frequency but did not change the PPR and presynaptic release probability, we tested postsynaptic glutamatergic receptor function and number by recording AMPAR- and NMDAR-dependent currents in both MOPP cells and DGCs (Fig. 4J) (Umemiya et al., 1999; Etherton et al., 2011). Two-way ANOVA followed by Tukey's post hoc analysis revealed the significant main effects of AMPAR-mediated current (AMPA-EPSC; genotype, F_(1,104) = 0.75, p = 0.3890; cell type, F_(1,104) = 76.90, p < 0.001; genotype × cell type, F_(1,104) = 20.06, p < 0.001) and NMDAR-mediated current (NMDA-EPSC; genotype, F_(1,104) = 0.04, p = 0.8370; cell type, F_(1,104) = 34.73, p < 0.001; genotype × cell type, F_(1,104) = 1.99, p = 0.1610) (Fig. 4K,L). AMPA-EPSC and NMDA-EPSC in MOPP cells were both larger than in DGCs in both genotypes (AMPA-EPSC: Cdkl5^+/y MOPP 108.34 ± 8.55 vs Cdkl5^+/y DGC 74.83 ± 5.73, p = 0.0030; Cdkl5^−/y MOPP 150.09 ± 10.09 vs Cdkl5^−/y DGC 46.61 ± 6.18, p < 0.001; NMDA-EPSC: Cdkl5^+/y MOPP 111.08 ± 14.72 vs Cdkl5^+/y DGC 53.52 ± 5.20, p = 0.0020; Cdkl5^−/y MOPP 131.83 ± 20.27 vs Cdkl5^−/y DGC 38.05 ± 4.15, p < 0.001). CDKL5 loss increased AMPA-EPSC in MOPP cells (Cdkl5^−/y MOPP 150.09 ± 10.09 vs Cdkl5^+/y MOPP 108.34 ± 8.55, p < 0.001) but decreased AMPA-EPSC (Cdkl5^−/y DGC 46.61 ± 6.18 vs Cdkl5^+/y DGC 74.83 ± 5.73, p = 0.012) and NMDA-EPSC (Cdkl5^−/y DGC 38.05 ± 4.15 vs Cdkl5^+/y DGC 53.52 ± 5.20, Mann–Whitney test, U = 227.0, p = 0.018) in DGCs of Cdkl5^−/y mice, which suggests that CDKL5 is critical to the function of postsynaptic glutamatergic receptors in the DG. The AMPA/NMDA ratios did not differ between the two genotypes for either cell type (Cdkl5^+/y MOPP 1.71 ± 0.38; Cdkl5^+/y DGC 1.55 ± 0.12; Cdkl5^−/y MOPP 2.53 ± 0.63; Cdkl5^−/y DGC 1.32 ± 0.14; two-way ANOVA: genotype × cell type interaction, F_(1,104) = 2.00, p = 0.161; Fig. 4M). Together, these results suggest that, although CDKL5 loss did not alter the intrinsic properties of either MOPP cells or DGCs (Extended Data Fig. 4-2), it did alter the postsynaptic glutamatergic receptor functions in the PP to MOPP cells and DGCs.

Blocking GABAergic transmission in the DG enhances hippocampal memory in Cdkl5^−/y mice

We next asked whether we could rescue the hippocampus-associated memory deficits in Cdkl5^−/y mice by relieving this overinhibition in the DG circuitry. We intracranially infused the GABA_A receptor antagonist gabazine (1.0 ng per side) bilaterally into the DG of the dorsal hippocampus (Fig. 5A,D). The dose was selected based on previous publications (Chee et al., 2015; Lu et al., 2016) and a few pilot tests. We trained the mice with a fear conditioning task 13-15 min after infusing the gabazine or vehicle and tested for contextual and cued fear memory 24 h later. During the acquisition phase, both saline- and gabazine-injected Cdkl5^−/y mice showed similar levels of exploratory behavior compared with Cdkl5^+/y littermates before the foot shock (Cdkl5^−/y-saline 4.36 ± 1.55% vs Cdkl5^+/y-saline 2.40 ± 2.07%, Mann–Whitney test, U = 49.5, p = 0.1640; Cdkl5^−/y-gabazine 0.82 ± 0.27% vs Cdkl5^+/y-gabazine 0.39 ± 0.22%, Mann–Whitney test, U = 42.0, p = 0.1850). Immediately after the foot shock, saline-treated Cdkl5^−/y mice froze less (Cdkl5^-/y-saline 36.33 ± 4.49% vs Cdkl5^+/y-saline 56.72 ± 4.10%, two-tailed unpaired t test, t₍₂₂₎ = 3.356, p = 0.0029) while gabazine-treated Cdkl5^-/y mice showed a level of freezing similar to Cdkl5^+/y mice (Cdkl5^-/y-gabazine 38.67 ± 6.77% vs Cdkl5^+/y-gabazine 40.53 ± 5.16%, two-tailed unpaired t test, t₍₂₀₎ = 0.219, p = 0.8290) (Fig. 5A). Twenty-four hours after the training, saline-treated Cdkl5^-/y mice froze less in the conditioning context than did their saline-treated Cdkl5^+/y littermates, thereby replicating the contextual memory impairment in the Cdkl5 nulls. Gabazine treatment significantly increased the freezing time of Cdkl5^-/y mice in the conditioning context but did not alter the contextual freezing level in WT mice compared with their saline-treated controls (two-way ANOVA followed by Tukey's post hoc analysis: genotype, F_(1,41) = 99.16, p < 0.001; treatment, F_(1,41) = 0.23, p = 0.638; Cdkl5^-/y saline 6.44 ± 1.31% vs Cdkl5^+/y saline 51.65 ± 5.90%, p < 0.001; Cdkl5^+/y saline 51.65 ± 5.90% vs Cdkl5^+/y gabazine 49.28 ± 4.26%, p = 0.6760; Cdkl5^-/y saline 6.44 ± 1.31% vs Cdkl5^-/y gabazine 12.70 ± 1.11%, two-tailed unpaired t test, t₍₁₉₎ = 3.622, p = 0.0017; Fig. 5B). As expected, local hippocampal gabazine infusion did not affect hippocampus-independent cued fear memory in either Cdkl5^-/y mice or their Cdkl5^+/y littermates (two-way ANOVA: genotype, F_(1,41) = 3.40, p = 0.0730, treatment, F_(1,41) = 0.004, p = 0.953; Cdkl5^+/y saline 21.30 ± 4.10%; Cdkl5^+/y gabazine 13.86 ± 4.58%; Cdkl5^-/y saline 6.40 ± 2.77%; Cdkl5^−/y gabazine 14.30 ± 3.61%; Fig. 5C). That the effect of gabazine was exclusive to the Cdkl5^-/y mice confirms the functional consequences of the excessive GABAergic inhibitory transmission to the DG.

Figure 5.

Blocking GABAergic transmission in the DG enhances hippocampal memory in Cdkl5^−/y mice. A, During fear conditioning training, saline-treated Cdkl5^−/y mice (n = 10) versus saline-treated Cdkl5^+/y mice (n = 13) as well as gabazine-treated Cdkl5^−/y mice (n = 11) versus gabazine-treated Cdkl5^+/y mice (n = 11) showed comparable exploration levels before the first foot shock. Saline-treated Cdkl5^−/y mice froze less than saline-treated Cdkl5^+/y mice immediately after the second foot shock, whereas gabazine-treated Cdkl5^–/y mice showed a similar level of freezing as gabazine-treated Cdkl5^+/y mice after the foot shock. Inset, Diagram showing the placement of bilateral microinjection of gabazine or saline into the DG of dorsal hippocampi. The numbers on the upper left represent the posterior coordinate from the bregma. B, Contextual freezing responses of the fear memory 24 h after training. C, Cued freezing responses of the fear memory 24 h after training. D, Schematic representation of the dorsal hippocampi at six rostral-caudal planes for the microinfusion sites of gabazine or saline in Cdkl5^–/y mice and Cdkl5^+/y littermates. The numbers on the left represent the posterior coordinate from the bregma. **p < 0.01, ***p < 0.001.

Forniceal DBS improves fear memory in female Cdkl5^+/− mice

We previously showed that forniceal DBS can rescue hippocampal memory in a Rett syndrome mouse model (Hao et al., 2015). We decided to test whether forniceal DBS might prove beneficial in this intellectual disability disorder model as well. We implanted Cdkl5^+/− mice and WT female littermates at 4.5 months with unilateral DBS electrodes to the fimbria-fornix. After recovery, both Cdkl5^+/− mice and WT controls received 2 weeks of DBS or sham treatment 3 weeks before the fear conditioning task. During the acquisition phase, both sham- and DBS-treated Cdkl5^+/− mice showed similar levels of exploratory behavior before the foot shock (Cdkl5^+/−-sham 0.55 ± 0.26% vs Cdkl5^+/+-sham 0.99 ± 0.27%, Mann–Whitney test, U = 104.5, p = 0.0790; Cdkl5^+/−-DBS 1.56 ± 0.54% vs Cdkl5^+/+-DBS 0.86 ± 0.25%, Mann–Whitney test, U = 99.0, p = 0.4020) and comparable freezing immediately after the foot shock (Cdkl5^+/−-sham 15.94 ± 3.39% vs Cdkl5^+/+-sham 19.17 ± 4.15%, Mann–Whitney test, U = 143.0, p = 0.7530; Cdkl5^+/−-DBS 21.76 ± 4.74% vs Cdkl5^+/+-DBS 18.64 ± 2.97%, two-tailed unpaired t test, t₍₂₉₎ = 0.548, p = 0.5880) compared with Cdkl5^+/+ littermates (Fig. 6A). Twenty-four hours after the training, two-way ANOVA followed by Tukey's post hoc analysis revealed the significant effects of genotype (F_(1,61) = 9.57, p = 0.003) and genotype × treatment interaction (F_(1,61) = 8.51, p = 0.005) on contextual memory. Sham-treated female Cdkl5^+/− mice (21.49 ± 3.19%) replicated the contextual fear memory impairment compared with the Cdkl5^+/+-sham group (45.97 ± 5.19%) (p < 0.0010). Forniceal DBS rescued hippocampus-dependent contextual fear memory in Cdkl5^+/− mice compared with sham-treated Cdkl5^+/− mice (Cdkl5^+/− sham 21.49 ± 3.19% vs Cdkl5^+/− DBS 34.93 ± 3.40%, p = 0.0220) (Fig. 6B). As a result, there is no difference between DBS-treated Cdkl5^+/− mice (34.93 ± 3.40%) and sham-treated Cdkl5^+/+ animals (45.97 ± 5.19%) (two-tailed unpaired t test, t₍₃₁₎ = 1.756, p = 0.0890). Notably, forniceal DBS also restored hippocampus-independent cued fear memory (two-way ANOVA followed by Tukey's post hoc analysis: genotype, F_(1,61) = 5.53, p = 0.022; treatment, F_(1,61) = 9.14, p = 0.004; Cdkl5^+/− sham 24.78 ± 5.28% vs Cdkl5^+/+ sham 43.63 ± 5.99%, p = 0.0350; Cdkl5^+/− sham 24.78 ± 5.28% vs Cdkl5^+/− DBS 47.86 ± 8.41%, p = 0.0120; Cdkl5^+/− DBS 47.86 ± 8.41% vs Cdkl5^+/+ sham 43.63 ± 5.99%, Mann–Whitney test, U = 129.0, p = 0.8150; Fig. 6C). DBS did not alter contextual (Cdkl5^+/+ sham 45.97 ± 5.19 vs Cdkl5^+/+ DBS 35.65 ± 4.10%, p = 0.0810) or cued (Cdkl5^+/+ sham 43.63 ± 5.99% vs Cdkl5^+/+ DBS 58.68 ± 4.89%, p = 0.0990) fear memory in female WT mice, however, likely because of a ceiling effect of this model under these particular experimental conditions.

Figure 6.

Forniceal DBS improves fear memory in 6-month-old female Cdkl5^+/− mice. A, During fear conditioning training, sham-treated Cdkl5^+/+ mice (n = 17) versus sham-treated Cdkl5^+/− mice (n = 17) as well as DBS-treated Cdkl5^+/+ mice (n = 15) versus DBS-treated Cdkl5^+/− mice (n = 16) showed comparable exploration levels both before the first foot shock and froze at similar amount immediately after the second foot shock. B, Forniceal DBS improved contextual fear memory in Cdkl5^+/− mice but not in WT littermates. C, Forniceal DBS improved cued fear memory in Cdkl5^+/− mice but not in WT littermates. *p < 0.05, ***p < 0.001.

Forniceal DBS rescues hippocampal LTP and normalizes feedforward inhibition in the DG of freely moving Cdkl5^+/− mice

Given the DBS-induced benefits of memory, we next examined whether forniceal DBS improves hippocampal LTP in female heterozygous animals. Using similar procedures, we implanted Cdkl5^+/− mice and WT littermates at 4.5 months with unilateral DBS electrodes to the fimbria-fornix, stimulation electrodes to the PP, and ipsilateral recording electrodes in the DG (Hao et al., 2015). After recovery, we assigned both Cdkl5^+/− mice and WT controls to DBS or sham groups in which they received 2 weeks of DBS/sham treatment as in the fear memory experiment. Three weeks after the DBS/sham session, LTP induction potentiated population spikes for multiple days in all groups (one-way repeated-measures ANOVA, day 0, p < 0.001) (Fig. 7A,B). Two-way repeated-measures ANOVA revealed that LTP decreased in sham-treated Cdkl5^+/− mice compared with sham-treated Cdkl5^+/+ group on day 0 (F_(1,12) = 9.04, p = 0.011) and day 1 (F_(1,12) = 7.02, p = 0.021), which replicated the LTP impairment in naive Cdkl5^+/− mice (Fig. 2C). Notably, forniceal DBS completely rescued LTP in Cdkl5^+/− mice (Cdkl5^+/−-DBS vs Cdkl5^+/−-sham: day 0, F_(1,12) = 10.46, p = 0.007; day 1, F_(1,12) = 4.87, p = 0.048; day 2, F_(1,12) = 7.31, p = 0.019) so that it equaled the LTP in sham-treated Cdkl5^+/+ animals (Cdkl5^+/−-DBS vs Cdkl5^+/+-sham, p > 0.05 for all the test days). DBS did not alter LTP levels (Cdkl5^+/+-sham vs Cdkl5^+/+-DBS, p > 0.05 for all the test days) or hippocampal memory in WT animals, because of the same ceiling effect we observed before (Fig. 6B).

Figure 7.

Forniceal DBS rescues hippocampal synaptic plasticity and local inhibitory circuit activity in the DG of freely moving Cdkl5^+/− mice. A, Superimposed traces of the PP recorded in the DG 5 min before (gray) and 55 min after tetani. B, Summary of in vivo LTP in Cdkl5^+/+-sham (n = 7), Cdkl5^+/+-DBS (n = 9), Cdkl5^+/−-sham (n = 7), and Cdkl5^+/−-DBS (n = 7) mice. LTP was induced by theta burst stimulation (arrow) on day 0 and followed up for 5 more days. C, Summary of FDI index in Cdkl5^+/+-sham (n = 8), Cdkl5^+/+-DBS (n = 9), Cdkl5^+/−-sham (n = 7), and Cdkl5^+/−-DBS (n = 7) mice. D, Summary of feedback inhibition index in Cdkl5^+/+-sham (n = 8), Cdkl5^+/+-DBS (n = 9), Cdkl5^+/−-sham (n = 7), and Cdkl5^+/−-DBS (n = 7) mice. *p < 0.05, **p < 0.01.

Last, we tested whether forniceal DBS modulates feedforward and feedback inhibition in the DG in freely moving Cdkl5^+/− mice. We performed the same surgical preparations used for the LTP test. Three weeks after the 2-week DBS/sham treatment, we tested all 4 groups of mice for both feedforward and feedback inhibition (Fig. 3). For feedforward inhibition, two-way ANOVA revealed the significant main effects of genotype (F_(1,27) = 4.60, p = 0.0410) and genotype × treatment interaction (F_(1,27) = 4.68, p = 0.0400) (Fig. 7C). Tukey's post hoc test indicated that sham-treated Cdkl5^+/− mice had lower FDI indices (indicating stronger feedforward inhibition) than sham-treated Cdkl5^+/+ animals (Cdkl5^+/− sham 0.72 ± 0.06 vs Cdkl5^+/+ sham 0.89 ± 0.04, p = 0.0060). Forniceal DBS restored the feedforward inhibition in Cdkl5^+/− mice (Cdkl5^+/− sham 0.72 ± 0.06 vs Cdkl5^+/− DBS 0.85 ± 0.02, p = 0.0320; Cdkl5^+/− DBS 0.85 ± 0.02 vs Cdkl5^+/+ sham 0.89 ± 0.04, two-tailed unpaired t test, t₍₁₃₎ = 0.758, p = 0.4620) without altering it in WT animals (Cdkl5^+/+ sham 0.89 ± 0.04 vs Cdkl5^+/+ DBS 0.85 ± 0.03, p = 0.4740). Forniceal DBS had no effect on feedback inhibition in either Cdkl5^+/− mice or WT littermates (two-way repeated-measures ANOVA: genotype, F_(3,27) = 0.68, p = 0.5720; genotype × interval interaction, F_(12,108) = 0.53, p = 0.8880; Fig. 7D). These data show that forniceal DBS rescues hippocampal LTP and normalizes feedforward inhibition in female Cdkl5^+/− mice.

Discussion

Hippocampus-dependent learning and memory deficits have been consistently demonstrated across CDD models (Wang et al., 2012; Fuchs et al., 2015; Okuda et al., 2018). Although several studies have addressed the molecular derangements involved (Fuchs et al., 2014; Trazzi et al., 2016; Okuda et al., 2017; Tang et al., 2019; Zhu and Xiong, 2019; Fuchs et al., 2020), the underlying neural circuit mechanisms of memory dysfunction have not been investigated. We therefore sought to understand how the molecular changes that follow CDKL5 loss affect physiology at the circuit, neuronal, and synaptic level. Our study highlights a novel circuit mechanism whereby loss of CDKL5 augments feedforward inhibition into the DG from overactivated MOPP cells, decreases excitatory activity from the PP, and suppresses the plasticity necessary for learning and memory formation. Remarkably, forniceal DBS rescued hippocampal memory, synaptic plasticity, and feedforward inhibition in Cdkl5 mutant mice (Fig. 8).

Figure 8.

Forniceal DBS rescues fear memory and hippocampal circuit activity in Cdkl5 mutant mice. The PP originating in the entorhinal cortex innervates both MOPP cells and DGCs in the DG. The MOPP cells form synaptic connections with the DGCs and mediate feedforward inhibition of the perforant-to-DG pathway. In WT mice, the PP provides greater synaptic input to the MOPP cells than to the granule cells and maintains a balanced E/I in DGCs. Loss of CDKL5, however, further amplifies the difference in PP input between MOPP cells and DGCs: glutamate-mediated PP excitatory synaptic transmission increases to the MOPP cells but decreases to the DGCs cells in Cdkl5 mutant mice. As a result, the stronger GABA-mediated inhibition from overactivated MOPP cells and decreased excitatory input directly from the PP shift the E/I of the granule cells to inhibition. This overinhibition of the DGCs contributes to hippocampal learning and memory deficits in Cdkl5 mutant mice. Chronic forniceal DBS rescues the feedforward inhibition of this pathway and the hippocampal memory in Cdkl5^+/− mice.

A few studies have examined CDKL5-mediated synaptic plasticity in different brain regions using brain slice preparations. CDKL5 loss in male nulls decreases LTP in the somatosensory cortex (Della Sala et al., 2016) and perirhinal cortex (Ren et al., 2019) but increases LTP in the Schaffer collateral to CA1 pathway of the hippocampus (Okuda et al., 2017). The CDKL5 knock-in mice also show increased early phase LTP and normal long-term depression in the Schaffer collateral to CA1 pathway (Yennawar et al., 2019). Interpretation of such ex vivo data, however, is complicated by mechanical damage to neural connections, changes in extracellular glutamate concentration, and alteration in immediate early gene expressions (Dolphin et al., 1982; French et al., 2001). Since the DG is crucial in gating contextual and spatial information from the entorhinal cortex to the hippocampus (Hargreaves et al., 2005) and in hippocampal memory (Sasaki et al., 2018), we examined LTP in the PP-DG pathway in vivo, which to our knowledge has never been studied in any CDD model. Loss of CDKL5 impairs LTP in this pathway in both male Cdkl5 null and female Cdkl5 heterozygous mice. Therefore, CDKL5 plays a role in regulating LTP in the DG and, by extension, hippocampal memory.

Impaired E/I balance is commonly found in mouse models of neurodevelopmental disorders and is thought to underlie a range of pathologies (Yizhar et al., 2011; Bateup et al., 2013; Han et al., 2013; Nelson and Valakh, 2015). Conditional KO of CDKL5 in forebrain glutamatergic neurons increases both excitatory and inhibitory synaptic transmission in CA1 pyramidal neurons and induces hippocampal circuit hypoexcitability that is linked to the memory deficits (Tang et al., 2017). In contrast, CDKL5 loss in the forebrain GABAergic neurons increases excitatory but not inhibitory transmission in CA1 pyramidal neurons and induces circuit hyperexcitability, contributing to autistic features (Tang et al., 2019). We found that CDKL5 loss increased feedforward inhibition and reduced the E/I ratio in the DGCs. The E/I ratio difference between the projection neurons in DG and CA1 regions might stem from differences in the intrinsic properties of DGCs and CA1 pyramidal neurons or differences in overall network activity in the conventional KO model we used and the conditional KO of CDKL5 in excitatory or inhibitory neurons. Although conditional KO models may provide valuable insight into the function of CDKL5 in different types of neurons, our results showcase the collective hippocampal circuit defects that CDKL5 loss causes and demonstrate that feedforward inhibition in the PP to DG contributes to memory deficits.

Of all the DG inhibitory interneurons, the function of the MOPP cells is least understood, partly because they were discovered only fairly recently (Amaral et al., 2007). We found that MOPP cells mediate the feedforward inhibition from the PP to the DGCs, which is consistent with a previous report (Li et al., 2013). Our observation that PP-initiated excitatory synaptic transmission is stronger in MOPP cells than in DGCs in WT mice indicates that the MOPP cell-mediated tonic feedforward inhibition helps enable normal circuit activity in the DG. We also found that CDKL5 is vital to MOPP cell-mediated inhibition.

Previously, only impairments in the fear conditioning task had been reported in male nulls from the model used in this study (Wang et al., 2012). We show here that loss of CDKL5 impaired hippocampal learning and memory in multiple tasks, in both male and female mutant mice. In particular, the clinically relevant female heterozygous mice have behavioral deficits comparable to those in the male nulls, although these deficits appear 2 months later because of mosaicism. These new findings are critical to understand the neurophysiological functions of CDKL5 and validate the fact that female Cdkl5^+/− mice are a valuable model for testing therapeutic interventions.

This study also suggests that DBS's ability to rescue hippocampus-dependent learning and memory is not limited to particular molecular defects. We previously used DBS to rescue hippocampal circuit function and memory in a mouse model of Rett syndrome, which is caused by loss of function of MeCP2 (Hao et al., 2015), a protein whose function is completely unrelated to CDKL5. Consistent with our new finding showing that increased feedforward inhibition contributes to the behavioral deficits in Cdkl5 mutant mice, DBS again rescued both memory deficits and hippocampal circuit dysfunction in parallel. Interestingly, forniceal DBS also enhanced hippocampus-independent cued fear memory in Cdkl5 heterozygous mice, something that we did not observe in Rett syndrome mice (Hao et al., 2015). We have yet to pinpoint why forniceal DBS enhances cued fear memory. The amygdala and hippocampus act synergistically to form long-term memories of emotional events (Richter-Levin and Akirav, 2000; Phelps, 2004); it seems likely that either the amygdala-hippocampal circuit itself, or the effect of DBS on the circuit, differs between CDKL5- and MeCP2-deficient mice. Further studies will be needed to decipher such differences.

It was proposed nearly a decade ago that interventions targeted at relevant neural circuits could correct the neural circuit defects in the CDD brain (Zoghbi and Bear, 2012). Our results validate this hypothesis and demonstrate that forniceal DBS can relieve the feedforward inhibition in the PP and rescue hippocampus-dependent learning and memory deficits. Our results provide an important circuit-based understanding of CDKL5-mediated cognitive dysfunction. Forniceal DBS may constitute a promising translational intervention for treating CDD and, possibly, other intellectual disability disorders.

Footnotes

This work was supported in part by the National Institute of Neurological Disorders and Stroke R01NS100738 to J.T.; the LouLou Foundation to J.T.; the Eunice Kennedy Shriver National Institute of Child Health and Human Development (U54HD083092, P50HD103555 to Baylor College of Medicine Intellectual and Developmental Disabilities Research Center, Neuroconnectivity Core, Circuit Modulation Core, Neurovisualization Core); the In Vivo Neurophysiology Core of Jan and Dan Duncan Neurological Research Institute at Texas Children's Hospital; the Chao Family Foundation; and the Cockrell Family Foundation. We thank Dr. Huda Y. Zoghbi, Dr. Jin Xu, Vicky Brandt, and Emily W. Schaffer for critical reading of the manuscript.
The authors declare no competing financial interests.
Correspondence should be addressed to Jianrong Tang at jtang1{at}bcm.edu

SfN exclusive license.

References

↵
2. Allene C,
3. Picardo MA,
4. Becq H,
5. Miyoshi G,
6. Fishell G,
7. Cossart R
(2012) Dynamic changes in interneuron morphophysiological properties mark the maturation of hippocampal network activity. J Neurosci 32:6688–6698. doi:10.1523/JNEUROSCI.0081-12.2012 pmid:22573691
OpenUrl Abstract/FREE Full Text
↵
2. Amaral DG,
3. Scharfman HE,
4. Lavenex P
(2007) The dentate gyrus: fundamental neuroanatomical organization (dentate gyrus for dummies). Prog Brain Res 163:3–22. doi:10.1016/S0079-6123(07)63001-5 pmid:17765709
OpenUrl CrossRef PubMed
↵
2. Andersen P,
3. Holmqvist B,
4. Voorhoeve PE
(1966) Entorhinal activation of dentate granule cells. Acta Physiol Scand 66:448–460. doi:10.1111/j.1748-1716.1966.tb03223.x pmid:5927271
OpenUrl CrossRef PubMed
↵
2. Bateup HS,
3. Johnson CA,
4. Denefrio CL,
5. Saulnier JL,
6. Kornacker K,
7. Sabatini BL
(2013) Excitatory/inhibitory synaptic imbalance leads to hippocampal hyperexcitability in mouse models of tuberous sclerosis. Neuron 78:510–522. doi:10.1016/j.neuron.2013.03.017 pmid:23664616
OpenUrl CrossRef PubMed
↵
2. Bouthour W,
3. Megevand P,
4. Donoghue J,
5. Luscher C,
6. Birbaumer N,
7. Krack P
(2019) Biomarkers for closed-loop deep brain stimulation in Parkinson disease and beyond. Nat Rev Neurol 15:343–352. doi:10.1038/s41582-019-0166-4 pmid:30936569
OpenUrl CrossRef PubMed
↵
2. Chee SS,
3. Menard JL,
4. Dringenberg HC
(2015) The lateral septum as a regulator of hippocampal theta oscillations and defensive behavior in rats. J Neurophysiol 113:1831–1841. doi:10.1152/jn.00806.2014 pmid:25589592
OpenUrl CrossRef PubMed
↵
2. Corcoran KA,
3. Maren S
(2001) Hippocampal inactivation disrupts contextual retrieval of fear memory after extinction. J Neurosci 21:1720–1726. doi:10.1523/JNEUROSCI.21-05-01720.2001
OpenUrl Abstract/FREE Full Text
↵
2. Della Sala G,
3. Putignano E,
4. Chelini G,
5. Melani R,
6. Calcagno E,
7. Ratto GM,
8. Amendola E,
9. Gross CT,
10. Giustetto M,
11. Pizzorusso T
(2016) Dendritic spine instability in a mouse model of CDKL5 disorder is rescued by insulin-like growth factor 1. Biol Psychiatry 80:302–311. doi:10.1016/j.biopsych.2015.08.028 pmid:26452614
OpenUrl CrossRef PubMed
↵
2. Dolphin AC,
3. Errington ML,
4. Bliss TV
(1982) Long-term potentiation of the perforant path in vivo is associated with increased glutamate release. Nature 297:496–498. doi:10.1038/297496a0 pmid:6123949
OpenUrl CrossRef PubMed
↵
2. Etherton M,
3. Foldy C,
4. Sharma M,
5. Tabuchi K,
6. Liu X,
7. Shamloo M,
8. Malenka RC,
9. Sudhof TC
(2011) Autism-linked neuroligin-3 R451C mutation differentially alters hippocampal and cortical synaptic function. Proc Natl Acad Sci USA 108:13764–13769. doi:10.1073/pnas.1111093108 pmid:21808020
OpenUrl Abstract/FREE Full Text
↵
2. Foss-Feig JH,
3. Adkinson BD,
4. Ji JL,
5. Yang G,
6. Srihari VH,
7. McPartland JC,
8. Krystal JH,
9. Murray JD,
10. Anticevic A
(2017) Searching for cross-diagnostic convergence: neural mechanisms governing excitation and inhibition balance in schizophrenia and autism spectrum disorders. Biol Psychiatry 81:848–861. doi:10.1016/j.biopsych.2017.03.005 pmid:28434615
OpenUrl CrossRef PubMed
↵
2. French PJ,
3. O'Connor V,
4. Jones MW,
5. Davis S,
6. Errington ML,
7. Voss K,
8. Truchet B,
9. Wotjak C,
10. Stean T,
11. Doyere V,
12. Maroun M,
13. Laroche S,
14. Bliss TV
(2001) Subfield-specific immediate early gene expression associated with hippocampal long-term potentiation in vivo. Eur J Neurosci 13:968–976. doi:10.1046/j.0953-816x.2001.01467.x pmid:11264669
OpenUrl CrossRef PubMed
↵
2. Freund TF,
3. Buzsáki G
(1996) Interneurons of the hippocampus. Hippocampus 6:347–470. doi:10.1002/(SICI)1098-1063(1996)6:4<347::AID-HIPO1>3.0.CO;2-I
OpenUrl CrossRef PubMed
↵
2. Fuchs C,
3. Trazzi S,
4. Torricella R,
5. Viggiano R,
6. De Franceschi M,
7. Amendola E,
8. Gross C,
9. Calza L,
10. Bartesaghi R,
11. Ciani E
(2014) Loss of CDKL5 impairs survival and dendritic growth of newborn neurons by altering AKT/GSK-3beta signaling. Neurobiol Dis 70:53–68. doi:10.1016/j.nbd.2014.06.006 pmid:24952363
OpenUrl CrossRef PubMed
↵
2. Fuchs C,
3. Rimondini R,
4. Viggiano R,
5. Trazzi S,
6. De Franceschi M,
7. Bartesaghi R,
8. Ciani E
(2015) Inhibition of GSK3beta rescues hippocampal development and learning in a mouse model of CDKL5 disorder. Neurobiol Dis 82:298–310. doi:10.1016/j.nbd.2015.06.018 pmid:26143616
OpenUrl CrossRef PubMed
↵
2. Fuchs C,
3. Gennaccaro L,
4. Ren E,
5. Galvani G,
6. Trazzi S,
7. Medici G,
8. Loi M,
9. Conway E,
10. Devinsky O,
11. Rimondini R,
12. Ciani E
(2020) Pharmacotherapy with sertraline rescues brain development and behavior in a mouse model of CDKL5 deficiency disorder. Neuropharmacology 167:107746. doi:10.1016/j.neuropharm.2019.107746 pmid:31469994
OpenUrl CrossRef PubMed
↵
2. Hainmueller T,
3. Bartos M
(2020) Dentate gyrus circuits for encoding, retrieval and discrimination of episodic memories. Nat Rev Neurosci 21:153–168. doi:10.1038/s41583-019-0260-z pmid:32042144
OpenUrl CrossRef PubMed
↵
2. Han K,
3. Holder JL Jr.,
4. Schaaf CP,
5. Lu H,
6. Chen H,
7. Kang H,
8. Tang J,
9. Wu Z,
10. Hao S,
11. Cheung SW,
12. Yu P,
13. Sun H,
14. Breman AM,
15. Patel A,
16. Lu HC,
17. Zoghbi HY
(2013) SHANK3 overexpression causes manic-like behaviour with unique pharmacogenetic properties. Nature 503:72–77. doi:10.1038/nature12630 pmid:24153177
OpenUrl CrossRef PubMed
↵
2. Hao S,
3. Tang B,
4. Wu Z,
5. Ure K,
6. Sun Y,
7. Tao H,
8. Gao Y,
9. Patel AJ,
10. Curry DJ,
11. Samaco RC,
12. Zoghbi HY,
13. Tang J
(2015) Forniceal deep brain stimulation rescues hippocampal memory in Rett syndrome mice. Nature 526:430–434. doi:10.1038/nature15694 pmid:26469053
OpenUrl CrossRef PubMed
↵
2. Hargreaves EL,
3. Rao G,
4. Lee I,
5. Knierim JJ
(2005) Major dissociation between medial and lateral entorhinal input to dorsal hippocampus. Science 308:1792–1794. doi:10.1126/science.1110449 pmid:15961670
OpenUrl Abstract/FREE Full Text
↵
2. Hsu TT,
3. Lee CT,
4. Tai MH,
5. Lien CC
(2016) Differential recruitment of dentate gyrus interneuron types by commissural versus perforant pathways. Cereb Cortex 26:2715–2727. doi:10.1093/cercor/bhv127 pmid:26045570
OpenUrl CrossRef PubMed
↵
2. Isaac JT,
3. Nicoll RA,
4. Malenka RC
(1995) Evidence for silent synapses: implications for the expression of LTP. Neuron 15:427–434. doi:10.1016/0896-6273(95)90046-2 pmid:7646894
OpenUrl CrossRef PubMed
↵
2. Jones IW,
3. Bolam JP,
4. Wonnacott S
(2001) Presynaptic localisation of the nicotinic acetylcholine receptor beta2 subunit immunoreactivity in rat nigrostriatal dopaminergic neurones. J Comp Neurol 439:235–247. doi:10.1002/cne.1345 pmid:11596051
OpenUrl CrossRef PubMed
↵
2. Jones MW,
3. Errington ML,
4. French PJ,
5. Fine A,
6. Bliss TV,
7. Garel S,
8. Charnay P,
9. Bozon B,
10. Laroche S,
11. Davis S
(2001) A requirement for the immediate early gene Zif268 in the expression of late LTP and long-term memories. Nat Neurosci 4:289–296. doi:10.1038/85138 pmid:11224546
OpenUrl CrossRef PubMed
↵
2. Kalscheuer VM,
3. Tao J,
4. Donnelly A,
5. Hollway G,
6. Schwinger E,
7. Kubart S,
8. Menzel C,
9. Hoeltzenbein M,
10. Tommerup N,
11. Eyre H,
12. Harbord M,
13. Haan E,
14. Sutherland GR,
15. Ropers HH,
16. Gecz J
(2003) Disruption of the serine/threonine kinase 9 gene causes severe X-linked infantile spasms and mental retardation. Am J Hum Genet 72:1401–1411. doi:10.1086/375538 pmid:12736870
OpenUrl CrossRef PubMed
↵
2. Lee JW,
3. Jung MW
(2017) Separation or binding? Role of the dentate gyrus in hippocampal mnemonic processing. Neurosci Biobehav Rev 75:183–194. doi:10.1016/j.neubiorev.2017.01.049 pmid:28174077
OpenUrl CrossRef PubMed
↵
2. Leutgeb JK,
3. Leutgeb S,
4. Moser MB,
5. Moser EI
(2007) Pattern separation in the dentate gyrus and CA3 of the hippocampus. Science 315:961–966. doi:10.1126/science.1135801 pmid:17303747
OpenUrl Abstract/FREE Full Text
↵
2. Li Y,
3. Stam FJ,
4. Aimone JB,
5. Goulding M,
6. Callaway EM,
7. Gage FH
(2013) Molecular layer perforant path-associated cells contribute to feed-forward inhibition in the adult dentate gyrus. Proc Natl Acad Sci USA 110:9106–9111. doi:10.1073/pnas.1306912110 pmid:23671081
OpenUrl Abstract/FREE Full Text
↵
2. Liu X,
3. Ramirez S,
4. Pang PT,
5. Puryear CB,
6. Govindarajan A,
7. Deisseroth K,
8. Tonegawa S
(2012) Optogenetic stimulation of a hippocampal engram activates fear memory recall. Nature 484:381–385. doi:10.1038/nature11028 pmid:22441246
OpenUrl CrossRef PubMed
↵
2. Lozano AM,
3. Lipsman N,
4. Bergman H,
5. Brown P,
6. Chabardes S,
7. Chang JW,
8. Matthews K,
9. McIntyre CC,
10. Schlaepfer TE,
11. Schulder M,
12. Temel Y,
13. Volkmann J,
14. Krauss JK
(2019) Deep brain stimulation: current challenges and future directions. Nat Rev Neurol 15:148–160. doi:10.1038/s41582-018-0128-2 pmid:30683913
OpenUrl CrossRef PubMed
↵
2. Lu H,
3. Ash RT,
4. He L,
5. Kee SE,
6. Wang W,
7. Yu D,
8. Hao S,
9. Meng X,
10. Ure K,
11. Ito-Ishida A,
12. Tang B,
13. Sun Y,
14. Ji D,
15. Tang J,
16. Arenkiel BR,
17. Smirnakis SM,
18. Zoghbi HY
(2016) Loss and gain of MeCP2 cause similar hippocampal circuit dysfunction that is rescued by deep brain stimulation in a Rett syndrome mouse model. Neuron 91:739–747. doi:10.1016/j.neuron.2016.07.018 pmid:27499081
OpenUrl CrossRef PubMed
↵
2. Malenka RC,
3. Bear MF
(2004) LTP and LTD: an embarrassment of riches. Neuron 44:5–21. doi:10.1016/j.neuron.2004.09.012 pmid:15450156
OpenUrl CrossRef PubMed
↵
2. Malleret G,
3. Haditsch U,
4. Genoux D,
5. Jones MW,
6. Bliss TV,
7. Vanhoose AM,
8. Weitlauf C,
9. Kandel ER,
10. Winder DG,
11. Mansuy IM
(2001) Inducible and reversible enhancement of learning, memory, and long-term potentiation by genetic inhibition of calcineurin. Cell 104:675–686. doi:10.1016/s0092-8674(01)00264-1 pmid:11257222
OpenUrl CrossRef PubMed
↵
2. Maren S,
3. De Oca B,
4. Fanselow MS
(1994) Sex differences in hippocampal long-term potentiation (LTP) and Pavlovian fear conditioning in rats: positive correlation between LTP and contextual learning. Brain Res 661:25–34. doi:10.1016/0006-8993(94)91176-2 pmid:7834376
OpenUrl CrossRef PubMed
↵
2. McHugh TJ,
3. Jones MW,
4. Quinn JJ,
5. Balthasar N,
6. Coppari R,
7. Elmquist JK,
8. Lowell BB,
9. Fanselow MS,
10. Wilson MA,
11. Tonegawa S
(2007) Dentate gyrus NMDA receptors mediate rapid pattern separation in the hippocampal network. Science 317:94–99. doi:10.1126/science.1140263 pmid:17556551
OpenUrl Abstract/FREE Full Text
↵
2. Morris RG,
3. Garrud P,
4. Rawlins JN,
5. O'Keefe J
(1982) Place navigation impaired in rats with hippocampal lesions. Nature 297:681–683. doi:10.1038/297681a0 pmid:7088155
OpenUrl CrossRef PubMed
↵
2. Moy SS,
3. Nadler JJ,
4. Young NB,
5. Perez A,
6. Holloway LP,
7. Barbaro RP,
8. Barbaro JR,
9. Wilson LM,
10. Threadgill DW,
11. Lauder JM,
12. Magnuson TR,
13. Crawley JN
(2007) Mouse behavioral tasks relevant to autism: phenotypes of 10 inbred strains. Behav Brain Res 176:4–20. doi:10.1016/j.bbr.2006.07.030 pmid:16971002
OpenUrl CrossRef PubMed
↵
2. Nakashiba T,
3. Cushman JD,
4. Pelkey KA,
5. Renaudineau S,
6. Buhl DL,
7. McHugh TJ,
8. Rodriguez Barrera V,
9. Chittajallu R,
10. Iwamoto KS,
11. McBain CJ,
12. Fanselow MS,
13. Tonegawa S
(2012) Young dentate granule cells mediate pattern separation, whereas old granule cells facilitate pattern completion. Cell 149:188–201. doi:10.1016/j.cell.2012.01.046 pmid:22365813
OpenUrl CrossRef PubMed
↵
2. Nelson SB,
3. Valakh V
(2015) Excitatory/inhibitory balance and circuit homeostasis in autism spectrum disorders. Neuron 87:684–698. doi:10.1016/j.neuron.2015.07.033 pmid:26291155
OpenUrl CrossRef PubMed
↵
2. Okuda K,
3. Kobayashi S,
4. Fukaya M,
5. Watanabe A,
6. Murakami T,
7. Hagiwara M,
8. Sato T,
9. Ueno H,
10. Ogonuki N,
11. Komano-Inoue S,
12. Manabe H,
13. Yamaguchi M,
14. Ogura A,
15. Asahara H,
16. Sakagami H,
17. Mizuguchi M,
18. Manabe T,
19. Tanaka T
(2017) CDKL5 controls postsynaptic localization of GluN2B-containing NMDA receptors in the hippocampus and regulates seizure susceptibility. Neurobiol Dis 106:158–170. doi:10.1016/j.nbd.2017.07.002 pmid:28688852
OpenUrl CrossRef PubMed
↵
2. Okuda K,
3. Takao K,
4. Watanabe A,
5. Miyakawa T,
6. Mizuguchi M,
7. Tanaka T
(2018) Comprehensive behavioral analysis of the Cdkl5 knockout mice revealed significant enhancement in anxiety- and fear-related behaviors and impairment in both acquisition and long-term retention of spatial reference memory. PLoS One 13:e0196587. doi:10.1371/journal.pone.0196587 pmid:29702698
OpenUrl CrossRef PubMed
↵
2. Paxinos G,
3. Franklin KB
(2001) The mouse brain in stereotaxic coordinates. San Diego: Academic.
↵
2. Phelps EA
(2004) Human emotion and memory: interactions of the amygdala and hippocampal complex. Curr Opin Neurobiol 14:198–202. doi:10.1016/j.conb.2004.03.015 pmid:15082325
OpenUrl CrossRef PubMed
↵
2. Ren E,
3. Roncace V,
4. Trazzi S,
5. Fuchs C,
6. Medici G,
7. Gennaccaro L,
8. Loi M,
9. Galvani G,
10. Ye K,
11. Rimondini R,
12. Aicardi G,
13. Ciani E
(2019) Functional and structural impairments in the perirhinal cortex of a mouse model of CDKL5 deficiency disorder are rescued by a Trkb agonist. Front Cell Neurosci 13:169. doi:10.3389/fncel.2019.00169 pmid:31114483
OpenUrl CrossRef PubMed
↵
2. Richter-Levin G,
3. Akirav I
(2000) Amygdala-hippocampus dynamic interaction in relation to memory. Mol Neurobiol 22:11–20. doi:10.1385/MN:22:1-3:011 pmid:11414274
OpenUrl CrossRef PubMed
↵
2. Rosenblum K,
3. Maroun M,
4. Richter-Levin G
(1999) Frequency-dependent inhibition in the dentate gyrus is attenuated by the NMDA receptor blocker MK-801 at doses that do not yet affect long-term potentiation. Hippocampus 9:491–494. doi:10.1002/(SICI)1098-1063(1999)9:5<491::AID-HIPO1>3.0.CO;2-V
OpenUrl CrossRef PubMed
↵
2. Samaco RC,
3. McGraw CM,
4. Ward CS,
5. Sun Y,
6. Neul JL,
7. Zoghbi HY
(2013) Female Mecp2(+/-) mice display robust behavioral deficits on two different genetic backgrounds providing a framework for pre-clinical studies. Hum Mol Genet 22:96–109. doi:10.1093/hmg/dds406 pmid:23026749
OpenUrl CrossRef PubMed
↵
2. Sasaki T,
3. Piatti VC,
4. Hwaun E,
5. Ahmadi S,
6. Lisman JE,
7. Leutgeb S,
8. Leutgeb JK
(2018) Dentate network activity is necessary for spatial working memory by supporting CA3 sharp-wave ripple generation and prospective firing of CA3 neurons. Nat Neurosci 21:258–269. doi:10.1038/s41593-017-0061-5 pmid:29335604
OpenUrl CrossRef PubMed
↵
2. Schneggenburger R,
3. Meyer AC,
4. Neher E
(1999) Released fraction and total size of a pool of immediately available transmitter quanta at a calyx synapse. Neuron 23:399–409. doi:10.1016/S0896-6273(00)80789-8
OpenUrl CrossRef PubMed
↵
2. Sloviter RS
(1991) Feedforward and feedback inhibition of hippocampal principal cell activity evoked by perforant path stimulation: GABA-mediated mechanisms that regulate excitability in vivo. Hippocampus 1:31–40. doi:10.1002/hipo.450010105 pmid:1669342
OpenUrl CrossRef PubMed
↵
2. Tang J,
3. Dani JA
(2009) Dopamine enables in vivo synaptic plasticity associated with the addictive drug nicotine. Neuron 63:673–682. doi:10.1016/j.neuron.2009.07.025 pmid:19755109
OpenUrl CrossRef PubMed
↵
2. Tang S,
3. Wang IJ,
4. Yue C,
5. Takano H,
6. Terzic B,
7. Pance K,
8. Lee JY,
9. Cui Y,
10. Coulter DA,
11. Zhou Z
(2017) Loss of CDKL5 in glutamatergic neurons disrupts hippocampal microcircuitry and leads to memory impairment in mice. J Neurosci 37:7420–7437. doi:10.1523/JNEUROSCI.0539-17.2017 pmid:28674172
OpenUrl Abstract/FREE Full Text
↵
2. Tang S,
3. Terzic B,
4. Wang IJ,
5. Sarmiento N,
6. Sizov K,
7. Cui Y,
8. Takano H,
9. Marsh ED,
10. Zhou Z,
11. Coulter DA
(2019) Altered NMDAR signaling underlies autistic-like features in mouse models of CDKL5 deficiency disorder. Nat Commun 10:2655. doi:10.1038/s41467-019-10689-w pmid:31201320
OpenUrl CrossRef PubMed
↵
2. Trazzi S,
3. Fuchs C,
4. Viggiano R,
5. De Franceschi M,
6. Valli E,
7. Jedynak P,
8. Hansen FK,
9. Perini G,
10. Rimondini R,
11. Kurz T,
12. Bartesaghi R,
13. Ciani E
(2016) HDAC4: a key factor underlying brain developmental alterations in CDKL5 disorder. Hum Mol Genet 25:3887–3907. doi:10.1093/hmg/ddw231 pmid:27466189
OpenUrl CrossRef PubMed
↵
2. Umemiya M,
3. Senda M,
4. Murphy TH
(1999) Behaviour of NMDA and AMPA receptor-mediated miniature EPSCs at rat cortical neuron synapses identified by calcium imaging. J Physiol 521:113–122. doi:10.1111/j.1469-7793.1999.00113.x pmid:10562338
OpenUrl CrossRef PubMed
↵
2. Wang IT,
3. Allen M,
4. Goffin D,
5. Zhu X,
6. Fairless AH,
7. Brodkin ES,
8. Siegel SJ,
9. Marsh ED,
10. Blendy JA,
11. Zhou Z
(2012) Loss of CDKL5 disrupts kinome profile and event-related potentials leading to autistic-like phenotypes in mice. Proc Natl Acad Sci USA 109:21516–21521. doi:10.1073/pnas.1216988110 pmid:23236174
OpenUrl Abstract/FREE Full Text
↵
2. Wang Q,
3. Mergia E,
4. Koesling D,
5. Mittmann T
(2017a) Nitric oxide/cGMP signaling via guanylyl cyclase isoform 1 modulates glutamate and GABA release in somatosensory cortex of mice. Neuroscience 360:180–189. doi:10.1016/j.neuroscience.2017.07.063 pmid:28782641
OpenUrl CrossRef PubMed
↵
2. Wang Q,
3. Yin P,
4. Yu B,
5. Zhao Z,
6. Richter-Levin G,
7. Yu L,
8. Cao X
(2017b) Down-regulation of dorsal striatal alphaCaMKII causes striatum-related cognitive and synaptic disorders. Exp Neurol 298:112–121. doi:10.1016/j.expneurol.2017.09.004 pmid:28890075
OpenUrl CrossRef PubMed
↵
2. Weaving LS,
3. Christodoulou J,
4. Williamson SL,
5. Friend KL,
6. McKenzie OL,
7. Archer H,
8. Evans J,
9. Clarke A,
10. Pelka GJ,
11. Tam PP,
12. Watson C,
13. Lahooti H,
14. Ellaway CJ,
15. Bennetts B,
16. Leonard H,
17. Gecz J
(2004) Mutations of CDKL5 cause a severe neurodevelopmental disorder with infantile spasms and mental retardation. Am J Hum Genet 75:1079–1093. doi:10.1086/426462 pmid:15492925
OpenUrl CrossRef PubMed
↵
2. Whitlock JR,
3. Heynen AJ,
4. Shuler MG,
5. Bear MF
(2006) Learning induces long-term potentiation in the hippocampus. Science 313:1093–1097. doi:10.1126/science.1128134 pmid:16931756
OpenUrl Abstract/FREE Full Text
↵
2. Yennawar M,
3. White RS,
4. Jensen FE
(2019) AMPA receptor dysregulation and therapeutic interventions in a mouse model of CDKL5 deficiency disorder. J Neurosci 39:4814–4828. doi:10.1523/JNEUROSCI.2041-18.2019 pmid:30952813
OpenUrl Abstract/FREE Full Text
↵
2. Yizhar O,
3. Fenno LE,
4. Prigge M,
5. Schneider F,
6. Davidson TJ,
7. O'Shea DJ,
8. Sohal VS,
9. Goshen I,
10. Finkelstein J,
11. Paz JT,
12. Stehfest K,
13. Fudim R,
14. Ramakrishnan C,
15. Huguenard JR,
16. Hegemann P,
17. Deisseroth K
(2011) Neocortical excitation/inhibition balance in information processing and social dysfunction. Nature 477:171–178. doi:10.1038/nature10360 pmid:21796121
OpenUrl CrossRef PubMed
↵
2. Zhang TA,
3. Tang J,
4. Pidoplichko VI,
5. Dani JA
(2010) Addictive nicotine alters local circuit inhibition during the induction of in vivo hippocampal synaptic potentiation. J Neurosci 30:6443–6453. doi:10.1523/JNEUROSCI.0458-10.2010 pmid:20445070
OpenUrl Abstract/FREE Full Text
↵
2. Zhu YC,
3. Xiong ZQ
(2019) Molecular and synaptic bases of CDKL5 disorder. Dev Neurobiol 79:8–19. doi:10.1002/dneu.22639 pmid:30246934
OpenUrl CrossRef PubMed
↵
2. Zoghbi HY,
3. Bear MF
(2012) Synaptic dysfunction in neurodevelopmental disorders associated with autism and intellectual disabilities. Cold Spring Harb Perspect Biol 4:a009886. doi:10.1101/cshperspect.a009886
OpenUrl Abstract/FREE Full Text

In this issue

View Full Page PDF

Citation Tools

Respond to this article

Request Permissions

Keywords

Cited By...

Research Articles

Show more Research Articles

Neurobiology of Disease

Show more Neurobiology of Disease

[1] ↵

Allene C,
Picardo MA,
Becq H,
Miyoshi G,
Fishell G,
Cossart R
(2012) Dynamic changes in interneuron morphophysiological properties mark the maturation of hippocampal network activity. J Neurosci 32:6688–6698. doi:10.1523/JNEUROSCI.0081-12.2012 pmid:22573691
OpenUrl Abstract/FREE Full Text

[3] Allene C,

[4] Picardo MA,

[5] Becq H,

[6] Miyoshi G,

[7] Fishell G,

[8] Cossart R

[9] ↵

Amaral DG,
Scharfman HE,
Lavenex P
(2007) The dentate gyrus: fundamental neuroanatomical organization (dentate gyrus for dummies). Prog Brain Res 163:3–22. doi:10.1016/S0079-6123(07)63001-5 pmid:17765709
OpenUrl CrossRef PubMed

[11] Amaral DG,

[12] Scharfman HE,

[13] Lavenex P

[14] ↵

Andersen P,
Holmqvist B,
Voorhoeve PE
(1966) Entorhinal activation of dentate granule cells. Acta Physiol Scand 66:448–460. doi:10.1111/j.1748-1716.1966.tb03223.x pmid:5927271
OpenUrl CrossRef PubMed

[16] Andersen P,

[17] Holmqvist B,

[18] Voorhoeve PE

[19] ↵

Bateup HS,
Johnson CA,
Denefrio CL,
Saulnier JL,
Kornacker K,
Sabatini BL
(2013) Excitatory/inhibitory synaptic imbalance leads to hippocampal hyperexcitability in mouse models of tuberous sclerosis. Neuron 78:510–522. doi:10.1016/j.neuron.2013.03.017 pmid:23664616
OpenUrl CrossRef PubMed

[21] Bateup HS,

[22] Johnson CA,

[23] Denefrio CL,

[24] Saulnier JL,

[25] Kornacker K,

[26] Sabatini BL

[27] ↵

Bouthour W,
Megevand P,
Donoghue J,
Luscher C,
Birbaumer N,
Krack P
(2019) Biomarkers for closed-loop deep brain stimulation in Parkinson disease and beyond. Nat Rev Neurol 15:343–352. doi:10.1038/s41582-019-0166-4 pmid:30936569
OpenUrl CrossRef PubMed

[29] Bouthour W,

[30] Megevand P,

[31] Donoghue J,

[32] Luscher C,

[33] Birbaumer N,

[34] Krack P

[35] ↵

Chee SS,
Menard JL,
Dringenberg HC
(2015) The lateral septum as a regulator of hippocampal theta oscillations and defensive behavior in rats. J Neurophysiol 113:1831–1841. doi:10.1152/jn.00806.2014 pmid:25589592
OpenUrl CrossRef PubMed

[37] Chee SS,

[38] Menard JL,

[39] Dringenberg HC

[40] ↵

Corcoran KA,
Maren S
(2001) Hippocampal inactivation disrupts contextual retrieval of fear memory after extinction. J Neurosci 21:1720–1726. doi:10.1523/JNEUROSCI.21-05-01720.2001
OpenUrl Abstract/FREE Full Text

[42] Corcoran KA,

[43] Maren S

[44] ↵

Della Sala G,
Putignano E,
Chelini G,
Melani R,
Calcagno E,
Ratto GM,
Amendola E,
Gross CT,
Giustetto M,
Pizzorusso T
(2016) Dendritic spine instability in a mouse model of CDKL5 disorder is rescued by insulin-like growth factor 1. Biol Psychiatry 80:302–311. doi:10.1016/j.biopsych.2015.08.028 pmid:26452614
OpenUrl CrossRef PubMed

[46] Della Sala G,

[47] Putignano E,

[48] Chelini G,

[49] Melani R,

[50] Calcagno E,

[51] Ratto GM,

[52] Amendola E,

[53] Gross CT,

[54] Giustetto M,

[55] Pizzorusso T

[56] ↵

Dolphin AC,
Errington ML,
Bliss TV
(1982) Long-term potentiation of the perforant path in vivo is associated with increased glutamate release. Nature 297:496–498. doi:10.1038/297496a0 pmid:6123949
OpenUrl CrossRef PubMed

[58] Dolphin AC,

[59] Errington ML,

[60] Bliss TV

[61] ↵

Etherton M,
Foldy C,
Sharma M,
Tabuchi K,
Liu X,
Shamloo M,
Malenka RC,
Sudhof TC
(2011) Autism-linked neuroligin-3 R451C mutation differentially alters hippocampal and cortical synaptic function. Proc Natl Acad Sci USA 108:13764–13769. doi:10.1073/pnas.1111093108 pmid:21808020
OpenUrl Abstract/FREE Full Text

[63] Etherton M,

[64] Foldy C,

[65] Sharma M,

[66] Tabuchi K,

[67] Liu X,

[68] Shamloo M,

[69] Malenka RC,

[70] Sudhof TC

[71] ↵

Foss-Feig JH,
Adkinson BD,
Ji JL,
Yang G,
Srihari VH,
McPartland JC,
Krystal JH,
Murray JD,
Anticevic A
(2017) Searching for cross-diagnostic convergence: neural mechanisms governing excitation and inhibition balance in schizophrenia and autism spectrum disorders. Biol Psychiatry 81:848–861. doi:10.1016/j.biopsych.2017.03.005 pmid:28434615
OpenUrl CrossRef PubMed

[73] Foss-Feig JH,

[74] Adkinson BD,

[75] Ji JL,

[76] Yang G,

[77] Srihari VH,

[78] McPartland JC,

[79] Krystal JH,

[80] Murray JD,

[81] Anticevic A

[82] ↵

French PJ,
O'Connor V,
Jones MW,
Davis S,
Errington ML,
Voss K,
Truchet B,
Wotjak C,
Stean T,
Doyere V,
Maroun M,
Laroche S,
Bliss TV
(2001) Subfield-specific immediate early gene expression associated with hippocampal long-term potentiation in vivo. Eur J Neurosci 13:968–976. doi:10.1046/j.0953-816x.2001.01467.x pmid:11264669
OpenUrl CrossRef PubMed

[84] French PJ,

[85] O'Connor V,

[86] Jones MW,

[87] Davis S,

[88] Errington ML,

[89] Voss K,

[90] Truchet B,

[91] Wotjak C,

[92] Stean T,

[93] Doyere V,

[94] Maroun M,

[95] Laroche S,

[96] Bliss TV

[97] ↵

Freund TF,
Buzsáki G
(1996) Interneurons of the hippocampus. Hippocampus 6:347–470. doi:10.1002/(SICI)1098-1063(1996)6:4<347::AID-HIPO1>3.0.CO;2-I
OpenUrl CrossRef PubMed

[99] Freund TF,

[100] Buzsáki G

[101] ↵

Fuchs C,
Trazzi S,
Torricella R,
Viggiano R,
De Franceschi M,
Amendola E,
Gross C,
Calza L,
Bartesaghi R,
Ciani E
(2014) Loss of CDKL5 impairs survival and dendritic growth of newborn neurons by altering AKT/GSK-3beta signaling. Neurobiol Dis 70:53–68. doi:10.1016/j.nbd.2014.06.006 pmid:24952363
OpenUrl CrossRef PubMed

[103] Fuchs C,

[104] Trazzi S,

[105] Torricella R,

[106] Viggiano R,

[107] De Franceschi M,

[108] Amendola E,

[109] Gross C,

[110] Calza L,

[111] Bartesaghi R,

[112] Ciani E

[113] ↵

Fuchs C,
Rimondini R,
Viggiano R,
Trazzi S,
De Franceschi M,
Bartesaghi R,
Ciani E
(2015) Inhibition of GSK3beta rescues hippocampal development and learning in a mouse model of CDKL5 disorder. Neurobiol Dis 82:298–310. doi:10.1016/j.nbd.2015.06.018 pmid:26143616
OpenUrl CrossRef PubMed

[115] Fuchs C,

[116] Rimondini R,

[117] Viggiano R,

[118] Trazzi S,

[119] De Franceschi M,

[120] Bartesaghi R,

[121] Ciani E

[122] ↵

Fuchs C,
Gennaccaro L,
Ren E,
Galvani G,
Trazzi S,
Medici G,
Loi M,
Conway E,
Devinsky O,
Rimondini R,
Ciani E
(2020) Pharmacotherapy with sertraline rescues brain development and behavior in a mouse model of CDKL5 deficiency disorder. Neuropharmacology 167:107746. doi:10.1016/j.neuropharm.2019.107746 pmid:31469994
OpenUrl CrossRef PubMed

[124] Fuchs C,

[125] Gennaccaro L,

[126] Ren E,

[127] Galvani G,

[128] Trazzi S,

[129] Medici G,

[130] Loi M,

[131] Conway E,

[132] Devinsky O,

[133] Rimondini R,

[134] Ciani E

[135] ↵

Hainmueller T,
Bartos M
(2020) Dentate gyrus circuits for encoding, retrieval and discrimination of episodic memories. Nat Rev Neurosci 21:153–168. doi:10.1038/s41583-019-0260-z pmid:32042144
OpenUrl CrossRef PubMed

[137] Hainmueller T,

[138] Bartos M

[139] ↵

Han K,
Holder JL Jr.,
Schaaf CP,
Lu H,
Chen H,
Kang H,
Tang J,
Wu Z,
Hao S,
Cheung SW,
Yu P,
Sun H,
Breman AM,
Patel A,
Lu HC,
Zoghbi HY
(2013) SHANK3 overexpression causes manic-like behaviour with unique pharmacogenetic properties. Nature 503:72–77. doi:10.1038/nature12630 pmid:24153177
OpenUrl CrossRef PubMed

[141] Han K,

[142] Holder JL Jr.,

[143] Schaaf CP,

[144] Lu H,

[145] Chen H,

[146] Kang H,

[147] Tang J,

[148] Wu Z,

[149] Hao S,

[150] Cheung SW,

[151] Yu P,

[152] Sun H,

[153] Breman AM,

[154] Patel A,

[155] Lu HC,

[156] Zoghbi HY

[157] ↵

Hao S,
Tang B,
Wu Z,
Ure K,
Sun Y,
Tao H,
Gao Y,
Patel AJ,
Curry DJ,
Samaco RC,
Zoghbi HY,
Tang J
(2015) Forniceal deep brain stimulation rescues hippocampal memory in Rett syndrome mice. Nature 526:430–434. doi:10.1038/nature15694 pmid:26469053
OpenUrl CrossRef PubMed

[159] Hao S,

[160] Tang B,

[161] Wu Z,

[162] Ure K,

[163] Sun Y,

[164] Tao H,

[165] Gao Y,

[166] Patel AJ,

[167] Curry DJ,

[168] Samaco RC,

[169] Zoghbi HY,

[170] Tang J

[171] ↵

Hargreaves EL,
Rao G,
Lee I,
Knierim JJ
(2005) Major dissociation between medial and lateral entorhinal input to dorsal hippocampus. Science 308:1792–1794. doi:10.1126/science.1110449 pmid:15961670
OpenUrl Abstract/FREE Full Text

[173] Hargreaves EL,

[174] Rao G,

[175] Lee I,

[176] Knierim JJ

[177] ↵

Hsu TT,
Lee CT,
Tai MH,
Lien CC
(2016) Differential recruitment of dentate gyrus interneuron types by commissural versus perforant pathways. Cereb Cortex 26:2715–2727. doi:10.1093/cercor/bhv127 pmid:26045570
OpenUrl CrossRef PubMed

[179] Hsu TT,

[180] Lee CT,

[181] Tai MH,

[182] Lien CC

[183] ↵

Isaac JT,
Nicoll RA,
Malenka RC
(1995) Evidence for silent synapses: implications for the expression of LTP. Neuron 15:427–434. doi:10.1016/0896-6273(95)90046-2 pmid:7646894
OpenUrl CrossRef PubMed

[185] Isaac JT,

[186] Nicoll RA,

[187] Malenka RC

[188] ↵

Jones IW,
Bolam JP,
Wonnacott S
(2001) Presynaptic localisation of the nicotinic acetylcholine receptor beta2 subunit immunoreactivity in rat nigrostriatal dopaminergic neurones. J Comp Neurol 439:235–247. doi:10.1002/cne.1345 pmid:11596051
OpenUrl CrossRef PubMed

[190] Jones IW,

[191] Bolam JP,

[192] Wonnacott S

[193] ↵

Jones MW,
Errington ML,
French PJ,
Fine A,
Bliss TV,
Garel S,
Charnay P,
Bozon B,
Laroche S,
Davis S
(2001) A requirement for the immediate early gene Zif268 in the expression of late LTP and long-term memories. Nat Neurosci 4:289–296. doi:10.1038/85138 pmid:11224546
OpenUrl CrossRef PubMed

[195] Jones MW,

[196] Errington ML,

[197] French PJ,

[198] Fine A,

[199] Bliss TV,

[200] Garel S,

[201] Charnay P,

[202] Bozon B,

[203] Laroche S,

[204] Davis S

[205] ↵

Kalscheuer VM,
Tao J,
Donnelly A,
Hollway G,
Schwinger E,
Kubart S,
Menzel C,
Hoeltzenbein M,
Tommerup N,
Eyre H,
Harbord M,
Haan E,
Sutherland GR,
Ropers HH,
Gecz J
(2003) Disruption of the serine/threonine kinase 9 gene causes severe X-linked infantile spasms and mental retardation. Am J Hum Genet 72:1401–1411. doi:10.1086/375538 pmid:12736870
OpenUrl CrossRef PubMed

[207] Kalscheuer VM,

[208] Tao J,

[209] Donnelly A,

[210] Hollway G,

[211] Schwinger E,

[212] Kubart S,

[213] Menzel C,

[214] Hoeltzenbein M,

[215] Tommerup N,

[216] Eyre H,

[217] Harbord M,

[218] Haan E,

[219] Sutherland GR,

[220] Ropers HH,

[221] Gecz J

[222] ↵

Lee JW,
Jung MW
(2017) Separation or binding? Role of the dentate gyrus in hippocampal mnemonic processing. Neurosci Biobehav Rev 75:183–194. doi:10.1016/j.neubiorev.2017.01.049 pmid:28174077
OpenUrl CrossRef PubMed

[224] Lee JW,

[225] Jung MW

[226] ↵

Leutgeb JK,
Leutgeb S,
Moser MB,
Moser EI
(2007) Pattern separation in the dentate gyrus and CA3 of the hippocampus. Science 315:961–966. doi:10.1126/science.1135801 pmid:17303747
OpenUrl Abstract/FREE Full Text

[228] Leutgeb JK,

[229] Leutgeb S,

[230] Moser MB,

[231] Moser EI

[232] ↵

Li Y,
Stam FJ,
Aimone JB,
Goulding M,
Callaway EM,
Gage FH
(2013) Molecular layer perforant path-associated cells contribute to feed-forward inhibition in the adult dentate gyrus. Proc Natl Acad Sci USA 110:9106–9111. doi:10.1073/pnas.1306912110 pmid:23671081
OpenUrl Abstract/FREE Full Text

[234] Li Y,

[235] Stam FJ,

[236] Aimone JB,

[237] Goulding M,

[238] Callaway EM,

[239] Gage FH

[240] ↵

Liu X,
Ramirez S,
Pang PT,
Puryear CB,
Govindarajan A,
Deisseroth K,
Tonegawa S
(2012) Optogenetic stimulation of a hippocampal engram activates fear memory recall. Nature 484:381–385. doi:10.1038/nature11028 pmid:22441246
OpenUrl CrossRef PubMed

[242] Liu X,

[243] Ramirez S,

[244] Pang PT,

[245] Puryear CB,

[246] Govindarajan A,

[247] Deisseroth K,

[248] Tonegawa S

[249] ↵

Lozano AM,
Lipsman N,
Bergman H,
Brown P,
Chabardes S,
Chang JW,
Matthews K,
McIntyre CC,
Schlaepfer TE,
Schulder M,
Temel Y,
Volkmann J,
Krauss JK
(2019) Deep brain stimulation: current challenges and future directions. Nat Rev Neurol 15:148–160. doi:10.1038/s41582-018-0128-2 pmid:30683913
OpenUrl CrossRef PubMed

[251] Lozano AM,

[252] Lipsman N,

[253] Bergman H,

[254] Brown P,

[255] Chabardes S,

[256] Chang JW,

[257] Matthews K,

[258] McIntyre CC,

[259] Schlaepfer TE,

[260] Schulder M,

[261] Temel Y,

[262] Volkmann J,

[263] Krauss JK

[264] ↵

Lu H,
Ash RT,
He L,
Kee SE,
Wang W,
Yu D,
Hao S,
Meng X,
Ure K,
Ito-Ishida A,
Tang B,
Sun Y,
Ji D,
Tang J,
Arenkiel BR,
Smirnakis SM,
Zoghbi HY
(2016) Loss and gain of MeCP2 cause similar hippocampal circuit dysfunction that is rescued by deep brain stimulation in a Rett syndrome mouse model. Neuron 91:739–747. doi:10.1016/j.neuron.2016.07.018 pmid:27499081
OpenUrl CrossRef PubMed

[266] Lu H,

[267] Ash RT,

[268] He L,

[269] Kee SE,

[270] Wang W,

[271] Yu D,

[272] Hao S,

[273] Meng X,

[274] Ure K,

[275] Ito-Ishida A,

[276] Tang B,

[277] Sun Y,

[278] Ji D,

[279] Tang J,

[280] Arenkiel BR,

[281] Smirnakis SM,

[282] Zoghbi HY

[283] ↵

Malenka RC,
Bear MF
(2004) LTP and LTD: an embarrassment of riches. Neuron 44:5–21. doi:10.1016/j.neuron.2004.09.012 pmid:15450156
OpenUrl CrossRef PubMed

[285] Malenka RC,

[286] Bear MF

[287] ↵

Malleret G,
Haditsch U,
Genoux D,
Jones MW,
Bliss TV,
Vanhoose AM,
Weitlauf C,
Kandel ER,
Winder DG,
Mansuy IM
(2001) Inducible and reversible enhancement of learning, memory, and long-term potentiation by genetic inhibition of calcineurin. Cell 104:675–686. doi:10.1016/s0092-8674(01)00264-1 pmid:11257222
OpenUrl CrossRef PubMed

[289] Malleret G,

[290] Haditsch U,

[291] Genoux D,

[292] Jones MW,

[293] Bliss TV,

[294] Vanhoose AM,

[295] Weitlauf C,

[296] Kandel ER,

[297] Winder DG,

[298] Mansuy IM

[299] ↵

Maren S,
De Oca B,
Fanselow MS
(1994) Sex differences in hippocampal long-term potentiation (LTP) and Pavlovian fear conditioning in rats: positive correlation between LTP and contextual learning. Brain Res 661:25–34. doi:10.1016/0006-8993(94)91176-2 pmid:7834376
OpenUrl CrossRef PubMed

[301] Maren S,

[302] De Oca B,

[303] Fanselow MS

[304] ↵

McHugh TJ,
Jones MW,
Quinn JJ,
Balthasar N,
Coppari R,
Elmquist JK,
Lowell BB,
Fanselow MS,
Wilson MA,
Tonegawa S
(2007) Dentate gyrus NMDA receptors mediate rapid pattern separation in the hippocampal network. Science 317:94–99. doi:10.1126/science.1140263 pmid:17556551
OpenUrl Abstract/FREE Full Text

[306] McHugh TJ,

[307] Jones MW,

[308] Quinn JJ,

[309] Balthasar N,

[310] Coppari R,

[311] Elmquist JK,

[312] Lowell BB,

[313] Fanselow MS,

[314] Wilson MA,

[315] Tonegawa S

[316] ↵

Morris RG,
Garrud P,
Rawlins JN,
O'Keefe J
(1982) Place navigation impaired in rats with hippocampal lesions. Nature 297:681–683. doi:10.1038/297681a0 pmid:7088155
OpenUrl CrossRef PubMed

[318] Morris RG,

[319] Garrud P,

[320] Rawlins JN,

[321] O'Keefe J

[322] ↵

Moy SS,
Nadler JJ,
Young NB,
Perez A,
Holloway LP,
Barbaro RP,
Barbaro JR,
Wilson LM,
Threadgill DW,
Lauder JM,
Magnuson TR,
Crawley JN
(2007) Mouse behavioral tasks relevant to autism: phenotypes of 10 inbred strains. Behav Brain Res 176:4–20. doi:10.1016/j.bbr.2006.07.030 pmid:16971002
OpenUrl CrossRef PubMed

[324] Moy SS,

[325] Nadler JJ,

[326] Young NB,

[327] Perez A,

[328] Holloway LP,

[329] Barbaro RP,

[330] Barbaro JR,

[331] Wilson LM,

[332] Threadgill DW,

[333] Lauder JM,

[334] Magnuson TR,

[335] Crawley JN

[336] ↵

Nakashiba T,
Cushman JD,
Pelkey KA,
Renaudineau S,
Buhl DL,
McHugh TJ,
Rodriguez Barrera V,
Chittajallu R,
Iwamoto KS,
McBain CJ,
Fanselow MS,
Tonegawa S
(2012) Young dentate granule cells mediate pattern separation, whereas old granule cells facilitate pattern completion. Cell 149:188–201. doi:10.1016/j.cell.2012.01.046 pmid:22365813
OpenUrl CrossRef PubMed

[338] Nakashiba T,

[339] Cushman JD,

[340] Pelkey KA,

[341] Renaudineau S,

[342] Buhl DL,

[343] McHugh TJ,

[344] Rodriguez Barrera V,

[345] Chittajallu R,

[346] Iwamoto KS,

[347] McBain CJ,

[348] Fanselow MS,

[349] Tonegawa S

[350] ↵

Nelson SB,
Valakh V
(2015) Excitatory/inhibitory balance and circuit homeostasis in autism spectrum disorders. Neuron 87:684–698. doi:10.1016/j.neuron.2015.07.033 pmid:26291155
OpenUrl CrossRef PubMed

[352] Nelson SB,

[353] Valakh V

[354] ↵

Okuda K,
Kobayashi S,
Fukaya M,
Watanabe A,
Murakami T,
Hagiwara M,
Sato T,
Ueno H,
Ogonuki N,
Komano-Inoue S,
Manabe H,
Yamaguchi M,
Ogura A,
Asahara H,
Sakagami H,
Mizuguchi M,
Manabe T,
Tanaka T
(2017) CDKL5 controls postsynaptic localization of GluN2B-containing NMDA receptors in the hippocampus and regulates seizure susceptibility. Neurobiol Dis 106:158–170. doi:10.1016/j.nbd.2017.07.002 pmid:28688852
OpenUrl CrossRef PubMed

[356] Okuda K,

[357] Kobayashi S,

[358] Fukaya M,

[359] Watanabe A,

[360] Murakami T,

[361] Hagiwara M,

[362] Sato T,

[363] Ueno H,

[364] Ogonuki N,

[365] Komano-Inoue S,

[366] Manabe H,

[367] Yamaguchi M,

[368] Ogura A,

[369] Asahara H,

[370] Sakagami H,

[371] Mizuguchi M,

[372] Manabe T,

[373] Tanaka T

[374] ↵

Okuda K,
Takao K,
Watanabe A,
Miyakawa T,
Mizuguchi M,
Tanaka T
(2018) Comprehensive behavioral analysis of the Cdkl5 knockout mice revealed significant enhancement in anxiety- and fear-related behaviors and impairment in both acquisition and long-term retention of spatial reference memory. PLoS One 13:e0196587. doi:10.1371/journal.pone.0196587 pmid:29702698
OpenUrl CrossRef PubMed

[376] Okuda K,

[377] Takao K,

[378] Watanabe A,

[379] Miyakawa T,

[380] Mizuguchi M,

[381] Tanaka T

[382] ↵

Paxinos G,
Franklin KB
(2001) The mouse brain in stereotaxic coordinates. San Diego: Academic.

[384] Paxinos G,

[385] Franklin KB

[386] ↵

Phelps EA
(2004) Human emotion and memory: interactions of the amygdala and hippocampal complex. Curr Opin Neurobiol 14:198–202. doi:10.1016/j.conb.2004.03.015 pmid:15082325
OpenUrl CrossRef PubMed

[388] Phelps EA

[389] ↵

Ren E,
Roncace V,
Trazzi S,
Fuchs C,
Medici G,
Gennaccaro L,
Loi M,
Galvani G,
Ye K,
Rimondini R,
Aicardi G,
Ciani E
(2019) Functional and structural impairments in the perirhinal cortex of a mouse model of CDKL5 deficiency disorder are rescued by a Trkb agonist. Front Cell Neurosci 13:169. doi:10.3389/fncel.2019.00169 pmid:31114483
OpenUrl CrossRef PubMed

[391] Ren E,

[392] Roncace V,

[393] Trazzi S,

[394] Fuchs C,

[395] Medici G,

[396] Gennaccaro L,

[397] Loi M,

[398] Galvani G,

[399] Ye K,

[400] Rimondini R,

[401] Aicardi G,

[402] Ciani E

[403] ↵

Richter-Levin G,
Akirav I
(2000) Amygdala-hippocampus dynamic interaction in relation to memory. Mol Neurobiol 22:11–20. doi:10.1385/MN:22:1-3:011 pmid:11414274
OpenUrl CrossRef PubMed

[405] Richter-Levin G,

[406] Akirav I

[407] ↵

Rosenblum K,
Maroun M,
Richter-Levin G
(1999) Frequency-dependent inhibition in the dentate gyrus is attenuated by the NMDA receptor blocker MK-801 at doses that do not yet affect long-term potentiation. Hippocampus 9:491–494. doi:10.1002/(SICI)1098-1063(1999)9:5<491::AID-HIPO1>3.0.CO;2-V
OpenUrl CrossRef PubMed

[409] Rosenblum K,

[410] Maroun M,

[411] Richter-Levin G

[412] ↵

Samaco RC,
McGraw CM,
Ward CS,
Sun Y,
Neul JL,
Zoghbi HY
(2013) Female Mecp2(+/-) mice display robust behavioral deficits on two different genetic backgrounds providing a framework for pre-clinical studies. Hum Mol Genet 22:96–109. doi:10.1093/hmg/dds406 pmid:23026749
OpenUrl CrossRef PubMed

[414] Samaco RC,

[415] McGraw CM,

[416] Ward CS,

[417] Sun Y,

[418] Neul JL,

[419] Zoghbi HY

[420] ↵

Sasaki T,
Piatti VC,
Hwaun E,
Ahmadi S,
Lisman JE,
Leutgeb S,
Leutgeb JK
(2018) Dentate network activity is necessary for spatial working memory by supporting CA3 sharp-wave ripple generation and prospective firing of CA3 neurons. Nat Neurosci 21:258–269. doi:10.1038/s41593-017-0061-5 pmid:29335604
OpenUrl CrossRef PubMed

[422] Sasaki T,

[423] Piatti VC,

[424] Hwaun E,

[425] Ahmadi S,

[426] Lisman JE,

[427] Leutgeb S,

[428] Leutgeb JK

[429] ↵

Schneggenburger R,
Meyer AC,
Neher E
(1999) Released fraction and total size of a pool of immediately available transmitter quanta at a calyx synapse. Neuron 23:399–409. doi:10.1016/S0896-6273(00)80789-8
OpenUrl CrossRef PubMed

[431] Schneggenburger R,

[432] Meyer AC,

[433] Neher E

[434] ↵

Sloviter RS
(1991) Feedforward and feedback inhibition of hippocampal principal cell activity evoked by perforant path stimulation: GABA-mediated mechanisms that regulate excitability in vivo. Hippocampus 1:31–40. doi:10.1002/hipo.450010105 pmid:1669342
OpenUrl CrossRef PubMed

[436] Sloviter RS

[437] ↵

Tang J,
Dani JA
(2009) Dopamine enables in vivo synaptic plasticity associated with the addictive drug nicotine. Neuron 63:673–682. doi:10.1016/j.neuron.2009.07.025 pmid:19755109
OpenUrl CrossRef PubMed

[439] Tang J,

[440] Dani JA

[441] ↵

Tang S,
Wang IJ,
Yue C,
Takano H,
Terzic B,
Pance K,
Lee JY,
Cui Y,
Coulter DA,
Zhou Z
(2017) Loss of CDKL5 in glutamatergic neurons disrupts hippocampal microcircuitry and leads to memory impairment in mice. J Neurosci 37:7420–7437. doi:10.1523/JNEUROSCI.0539-17.2017 pmid:28674172
OpenUrl Abstract/FREE Full Text

[443] Tang S,

[444] Wang IJ,

[445] Yue C,

[446] Takano H,

[447] Terzic B,

[448] Pance K,

[449] Lee JY,

[450] Cui Y,

[451] Coulter DA,

[452] Zhou Z

[453] ↵

Tang S,
Terzic B,
Wang IJ,
Sarmiento N,
Sizov K,
Cui Y,
Takano H,
Marsh ED,
Zhou Z,
Coulter DA
(2019) Altered NMDAR signaling underlies autistic-like features in mouse models of CDKL5 deficiency disorder. Nat Commun 10:2655. doi:10.1038/s41467-019-10689-w pmid:31201320
OpenUrl CrossRef PubMed

[455] Tang S,

[456] Terzic B,

[457] Wang IJ,

[458] Sarmiento N,

[459] Sizov K,

[460] Cui Y,

[461] Takano H,

[462] Marsh ED,

[463] Zhou Z,

[464] Coulter DA

[465] ↵

Trazzi S,
Fuchs C,
Viggiano R,
De Franceschi M,
Valli E,
Jedynak P,
Hansen FK,
Perini G,
Rimondini R,
Kurz T,
Bartesaghi R,
Ciani E
(2016) HDAC4: a key factor underlying brain developmental alterations in CDKL5 disorder. Hum Mol Genet 25:3887–3907. doi:10.1093/hmg/ddw231 pmid:27466189
OpenUrl CrossRef PubMed

[467] Trazzi S,

[468] Fuchs C,

[469] Viggiano R,

[470] De Franceschi M,

[471] Valli E,

[472] Jedynak P,

[473] Hansen FK,

[474] Perini G,

[475] Rimondini R,

[476] Kurz T,

[477] Bartesaghi R,

[478] Ciani E

[479] ↵

Umemiya M,
Senda M,
Murphy TH
(1999) Behaviour of NMDA and AMPA receptor-mediated miniature EPSCs at rat cortical neuron synapses identified by calcium imaging. J Physiol 521:113–122. doi:10.1111/j.1469-7793.1999.00113.x pmid:10562338
OpenUrl CrossRef PubMed

[481] Umemiya M,

[482] Senda M,

[483] Murphy TH

[484] ↵

Wang IT,
Allen M,
Goffin D,
Zhu X,
Fairless AH,
Brodkin ES,
Siegel SJ,
Marsh ED,
Blendy JA,
Zhou Z
(2012) Loss of CDKL5 disrupts kinome profile and event-related potentials leading to autistic-like phenotypes in mice. Proc Natl Acad Sci USA 109:21516–21521. doi:10.1073/pnas.1216988110 pmid:23236174
OpenUrl Abstract/FREE Full Text

[486] Wang IT,

[487] Allen M,

[488] Goffin D,

[489] Zhu X,

[490] Fairless AH,

[491] Brodkin ES,

[492] Siegel SJ,

[493] Marsh ED,

[494] Blendy JA,

[495] Zhou Z

[496] ↵

Wang Q,
Mergia E,
Koesling D,
Mittmann T
(2017a) Nitric oxide/cGMP signaling via guanylyl cyclase isoform 1 modulates glutamate and GABA release in somatosensory cortex of mice. Neuroscience 360:180–189. doi:10.1016/j.neuroscience.2017.07.063 pmid:28782641
OpenUrl CrossRef PubMed

[498] Wang Q,

[499] Mergia E,

[500] Koesling D,

[501] Mittmann T

Main menu

User menu

Search